738 files changed, 127034 insertions, 31487 deletions

diff --git a/crypto/LPdir_nyi.c b/crypto/LPdir_nyi.c
deleted file mode 100644
index 6c1a50e6a8e2..000000000000
--- a/crypto/LPdir_nyi.c
+++ /dev/null
@@ -1,42 +0,0 @@
-/* $LP: LPlib/source/LPdir_win.c,v 1.1 2004/06/14 10:07:56 _cvs_levitte Exp $ */
-/*
- * Copyright (c) 2004, Richard Levitte <richard@levitte.org>
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- *
- * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- */
-
-#ifndef LPDIR_H
-#include "LPdir.h"
-#endif
-
-struct LP_dir_context_st { void *dummy; };
-const char *LP_find_file(LP_DIR_CTX **ctx, const char *directory)
-	{
-	errno = EINVAL;
-	return 0;
-	}
-int LP_find_file_end(LP_DIR_CTX **ctx)
-	{
-	errno = EINVAL;
-	return 0;
-	}
diff --git a/crypto/LPdir_vms.c b/crypto/LPdir_vms.c
deleted file mode 100644
index 85b427a623b6..000000000000
--- a/crypto/LPdir_vms.c
+++ /dev/null
@@ -1,199 +0,0 @@
-/* $LP: LPlib/source/LPdir_vms.c,v 1.20 2004/08/26 13:36:05 _cvs_levitte Exp $ */
-/*
- * Copyright (c) 2004, Richard Levitte <richard@levitte.org>
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#include <stddef.h>
-#include <stdlib.h>
-#include <string.h>
-#include <errno.h>
-#include <descrip.h>
-#include <namdef.h>
-#include <rmsdef.h>
-#include <libfildef.h>
-#include <lib$routines.h>
-#include <strdef.h>
-#include <str$routines.h>
-#include <stsdef.h>
-#ifndef LPDIR_H
-#include "LPdir.h"
-#endif
-
-/* Because some compiler options hide this macor */
-#ifndef EVMSERR
-#define EVMSERR		65535  /* error for non-translatable VMS errors */
-#endif
-
-struct LP_dir_context_st
-{
-  unsigned long VMS_context;
-#ifdef NAML$C_MAXRSS
-  char filespec[NAML$C_MAXRSS+1];
-  char result[NAML$C_MAXRSS+1];
-#else
-  char filespec[256];
-  char result[256];
-#endif
-  struct dsc$descriptor_d filespec_dsc;
-  struct dsc$descriptor_d result_dsc;
-};
-
-const char *LP_find_file(LP_DIR_CTX **ctx, const char *directory)
-{
-  int status;
-  char *p, *r;
-  size_t l;
-  unsigned long flags = 0;
-#ifdef NAML$C_MAXRSS
-  flags |= LIB$M_FIL_LONG_NAMES;
-#endif
-
-  if (ctx == NULL || directory == NULL)
-    {
-      errno = EINVAL;
-      return 0;
-    }
-
-  errno = 0;
-  if (*ctx == NULL)
-    {
-      size_t filespeclen = strlen(directory);
-      char *filespec = NULL;
-
-      /* MUST be a VMS directory specification!  Let's estimate if it is. */
-      if (directory[filespeclen-1] != ']'
-	  && directory[filespeclen-1] != '>'
-	  && directory[filespeclen-1] != ':')
-	{
-	  errno = EINVAL;
-	  return 0;
-	}
-
-      filespeclen += 4;		/* "*.*;" */
-
-      if (filespeclen >
-#ifdef NAML$C_MAXRSS
-	  NAML$C_MAXRSS
-#else
-	  255
-#endif
-	  )
-	{
-	  errno = ENAMETOOLONG;
-	  return 0;
-	}
-
-      *ctx = (LP_DIR_CTX *)malloc(sizeof(LP_DIR_CTX));
-      if (*ctx == NULL)
-	{
-	  errno = ENOMEM;
-	  return 0;
-	}
-      memset(*ctx, '\0', sizeof(LP_DIR_CTX));
-
-      strcpy((*ctx)->filespec,directory);
-      strcat((*ctx)->filespec,"*.*;");
-      (*ctx)->filespec_dsc.dsc$w_length = filespeclen;
-      (*ctx)->filespec_dsc.dsc$b_dtype = DSC$K_DTYPE_T;
-      (*ctx)->filespec_dsc.dsc$b_class = DSC$K_CLASS_S;
-      (*ctx)->filespec_dsc.dsc$a_pointer = (*ctx)->filespec;
-      (*ctx)->result_dsc.dsc$w_length = 0;
-      (*ctx)->result_dsc.dsc$b_dtype = DSC$K_DTYPE_T;
-      (*ctx)->result_dsc.dsc$b_class = DSC$K_CLASS_D;
-      (*ctx)->result_dsc.dsc$a_pointer = 0;
-    }
-
-  (*ctx)->result_dsc.dsc$w_length = 0;
-  (*ctx)->result_dsc.dsc$b_dtype = DSC$K_DTYPE_T;
-  (*ctx)->result_dsc.dsc$b_class = DSC$K_CLASS_D;
-  (*ctx)->result_dsc.dsc$a_pointer = 0;
-
-  status = lib$find_file(&(*ctx)->filespec_dsc, &(*ctx)->result_dsc,
-			 &(*ctx)->VMS_context, 0, 0, 0, &flags);
-
-  if (status == RMS$_NMF)
-    {
-      errno = 0;
-      vaxc$errno = status;
-      return NULL;
-    }
-
-  if(!$VMS_STATUS_SUCCESS(status))
-    {
-      errno = EVMSERR;
-      vaxc$errno = status;
-      return NULL;
-    }
-
-  /* Quick, cheap and dirty way to discard any device and directory,
-     since we only want file names */
-  l = (*ctx)->result_dsc.dsc$w_length;
-  p = (*ctx)->result_dsc.dsc$a_pointer;
-  r = p;
-  for (; *p; p++)
-    {
-      if (*p == '^' && p[1] != '\0') /* Take care of ODS-5 escapes */
-	{
-	  p++;
-	}
-      else if (*p == ':' || *p == '>' || *p == ']')
-	{
-	  l -= p + 1 - r;
-	  r = p + 1;
-	}
-      else if (*p == ';')
-	{
-	  l = p - r;
-	  break;
-	}
-    }
-
-  strncpy((*ctx)->result, r, l);
-  (*ctx)->result[l] = '\0';
-  str$free1_dx(&(*ctx)->result_dsc);
-
-  return (*ctx)->result;
-}
-
-int LP_find_file_end(LP_DIR_CTX **ctx)
-{
-  if (ctx != NULL && *ctx != NULL)
-    {
-      int status = lib$find_file_end(&(*ctx)->VMS_context);
-
-      free(*ctx);
-
-      if(!$VMS_STATUS_SUCCESS(status))
-	{
-	  errno = EVMSERR;
-	  vaxc$errno = status;
-	  return 0;
-	}
-      return 1;
-    }
-  errno = EINVAL;
-  return 0;
-}
-
diff --git a/crypto/LPdir_win.c b/crypto/LPdir_win.c
deleted file mode 100644
index 09b475beed17..000000000000
--- a/crypto/LPdir_win.c
+++ /dev/null
@@ -1,155 +0,0 @@
-/* $LP: LPlib/source/LPdir_win.c,v 1.10 2004/08/26 13:36:05 _cvs_levitte Exp $ */
-/*
- * Copyright (c) 2004, Richard Levitte <richard@levitte.org>
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-#include <windows.h>
-#include <tchar.h>
-#ifndef LPDIR_H
-#include "LPdir.h"
-#endif
-
-/* We're most likely overcautious here, but let's reserve for
-    broken WinCE headers and explicitly opt for UNICODE call.
-    Keep in mind that our WinCE builds are compiled with -DUNICODE
-    [as well as -D_UNICODE]. */
-#if defined(LP_SYS_WINCE) && !defined(FindFirstFile)
-# define FindFirstFile FindFirstFileW
-#endif
-#if defined(LP_SYS_WINCE) && !defined(FindFirstFile)
-# define FindNextFile FindNextFileW
-#endif
-
-#ifndef NAME_MAX
-#define NAME_MAX 255
-#endif
-
-struct LP_dir_context_st
-{
-  WIN32_FIND_DATA ctx;
-  HANDLE handle;
-  char entry_name[NAME_MAX+1];
-};
-
-const char *LP_find_file(LP_DIR_CTX **ctx, const char *directory)
-{
-  struct dirent *direntry = NULL;
-
-  if (ctx == NULL || directory == NULL)
-    {
-      errno = EINVAL;
-      return 0;
-    }
-
-  errno = 0;
-  if (*ctx == NULL)
-    {
-      *ctx = (LP_DIR_CTX *)malloc(sizeof(LP_DIR_CTX));
-      if (*ctx == NULL)
-	{
-	  errno = ENOMEM;
-	  return 0;
-	}
-      memset(*ctx, '\0', sizeof(LP_DIR_CTX));
-
-      if (sizeof(TCHAR) != sizeof(char))
-	{
-	  TCHAR *wdir = NULL;
-	  /* len_0 denotes string length *with* trailing 0 */ 
-	  size_t index = 0,len_0 = strlen(directory) + 1;
-
-	  wdir = (TCHAR *)malloc(len_0 * sizeof(TCHAR));
-	  if (wdir == NULL)
-	    {
-	      free(*ctx);
-	      *ctx = NULL;
-	      errno = ENOMEM;
-	      return 0;
-	    }
-
-#ifdef LP_MULTIBYTE_AVAILABLE
-	  if (!MultiByteToWideChar(CP_ACP, 0, directory, len_0, (WCHAR *)wdir, len_0))
-#endif
-	    for (index = 0; index < len_0; index++)
-	      wdir[index] = (TCHAR)directory[index];
-
-	  (*ctx)->handle = FindFirstFile(wdir, &(*ctx)->ctx);
-
-	  free(wdir);
-	}
-      else
-	(*ctx)->handle = FindFirstFile((TCHAR *)directory, &(*ctx)->ctx);
-
-      if ((*ctx)->handle == INVALID_HANDLE_VALUE)
-	{
-	  free(*ctx);
-	  *ctx = NULL;
-	  errno = EINVAL;
-	  return 0;
-	}
-    }
-  else
-    {
-      if (FindNextFile((*ctx)->handle, &(*ctx)->ctx) == FALSE)
-	{
-	  return 0;
-	}
-    }
-
-  if (sizeof(TCHAR) != sizeof(char))
-    {
-      TCHAR *wdir = (*ctx)->ctx.cFileName;
-      size_t index, len_0 = 0;
-
-      while (wdir[len_0] && len_0 < (sizeof((*ctx)->entry_name) - 1)) len_0++;
-      len_0++;
-
-#ifdef LP_MULTIBYTE_AVAILABLE
-      if (!WideCharToMultiByte(CP_ACP, 0, (WCHAR *)wdir, len_0, (*ctx)->entry_name,
-			       sizeof((*ctx)->entry_name), NULL, 0))
-#endif
-	for (index = 0; index < len_0; index++)
-	  (*ctx)->entry_name[index] = (char)wdir[index];
-    }
-  else
-    strncpy((*ctx)->entry_name, (const char *)(*ctx)->ctx.cFileName,
-	    sizeof((*ctx)->entry_name)-1);
-
-  (*ctx)->entry_name[sizeof((*ctx)->entry_name)-1] = '\0';
-
-  return (*ctx)->entry_name;
-}
-
-int LP_find_file_end(LP_DIR_CTX **ctx)
-{
-  if (ctx != NULL && *ctx != NULL)
-    {
-      FindClose((*ctx)->handle);
-      free(*ctx);
-      *ctx = NULL;
-      return 1;
-    }
-  errno = EINVAL;
-  return 0;
-}
diff --git a/crypto/LPdir_win32.c b/crypto/LPdir_win32.c
deleted file mode 100644
index e39872da5259..000000000000
--- a/crypto/LPdir_win32.c
+++ /dev/null
@@ -1,30 +0,0 @@
-/* $LP: LPlib/source/LPdir_win32.c,v 1.3 2004/08/26 13:36:05 _cvs_levitte Exp $ */
-/*
- * Copyright (c) 2004, Richard Levitte <richard@levitte.org>
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#define LP_SYS_WIN32
-#define LP_MULTIBYTE_AVAILABLE
-#include "LPdir_win.c"
diff --git a/crypto/LPdir_wince.c b/crypto/LPdir_wince.c
deleted file mode 100644
index ab0e1e6f4f8d..000000000000
--- a/crypto/LPdir_wince.c
+++ /dev/null
@@ -1,31 +0,0 @@
-/* $LP: LPlib/source/LPdir_wince.c,v 1.3 2004/08/26 13:36:05 _cvs_levitte Exp $ */
-/*
- * Copyright (c) 2004, Richard Levitte <richard@levitte.org>
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#define LP_SYS_WINCE
-/* We might want to define LP_MULTIBYTE_AVAILABLE here.  It's currently
-   under investigation what the exact conditions would be */
-#include "LPdir_win.c"
diff --git a/crypto/Makefile b/crypto/Makefile
index 6557f2b4e1dc..947dd5d44ec5 100644
--- a/crypto/Makefile
+++ b/crypto/Makefile
@@ -5,9 +5,9 @@
 DIR=		crypto
 TOP=		..
 CC=		cc
-INCLUDE=	-I. -I$(TOP) -I../include
+INCLUDE=	-I. -I$(TOP) -I../include $(ZLIB_INCLUDE)
 # INCLUDES targets sudbirs!
-INCLUDES=	-I.. -I../.. -I../../include
+INCLUDES=	-I.. -I../.. -I../modes -I../asn1 -I../evp -I../../include $(ZLIB_INCLUDE)
 CFLAG=		-g
 MAKEDEPPROG=	makedepend
 MAKEDEPEND=	$(TOP)/util/domd $(TOP) -MD $(MAKEDEPPROG)
@@ -17,7 +17,7 @@ AR=		ar r
 
 RECURSIVE_MAKE=	[ -n "$(SDIRS)" ] && for i in $(SDIRS) ; do \
 		    (cd $$i && echo "making $$target in $(DIR)/$$i..." && \
-		    $(MAKE) -e TOP=../.. DIR=$$i INCLUDES='${INCLUDES}' $$target ) || exit 1; \
+		    $(MAKE) -e TOP=../.. DIR=$$i INCLUDES='$(INCLUDES)' $$target ) || exit 1; \
 		done;
 
 PEX_LIBS=
@@ -26,6 +26,7 @@ EX_LIBS=
 CFLAGS= $(INCLUDE) $(CFLAG)
 ASFLAGS= $(INCLUDE) $(ASFLAG)
 AFLAGS=$(ASFLAGS)
+CPUID_OBJ=mem_clr.o
 
 LIBS=
 
@@ -33,12 +34,14 @@ GENERAL=Makefile README crypto-lib.com install.com
 
 LIB= $(TOP)/libcrypto.a
 SHARED_LIB= libcrypto$(SHLIB_EXT)
-LIBSRC=	cryptlib.c dyn_lck.c mem.c mem_clr.c mem_dbg.c cversion.c ex_data.c tmdiff.c cpt_err.c ebcdic.c uid.c o_time.c o_str.c o_dir.c o_init.c fips_err.c
-LIBOBJ= cryptlib.o dyn_lck.o mem.o mem_clr.o mem_dbg.o cversion.o ex_data.o tmdiff.o cpt_err.o ebcdic.o uid.o o_time.o o_str.o o_dir.o o_init.o fips_err.o $(CPUID_OBJ)
+LIBSRC=	cryptlib.c mem.c mem_clr.c mem_dbg.c cversion.c ex_data.c cpt_err.c \
+	ebcdic.c uid.c o_time.c o_str.c o_dir.c o_fips.c o_init.c fips_ers.c
+LIBOBJ= cryptlib.o mem.o mem_dbg.o cversion.o ex_data.o cpt_err.o ebcdic.o \
+	uid.o o_time.o o_str.o o_dir.o o_fips.o o_init.o fips_ers.o $(CPUID_OBJ)
 
 SRC= $(LIBSRC)
 
-EXHEADER= crypto.h tmdiff.h opensslv.h opensslconf.h ebcdic.h symhacks.h \
+EXHEADER= crypto.h opensslv.h opensslconf.h ebcdic.h symhacks.h \
 	ossl_typ.h
 HEADER=	cryptlib.h buildinf.h md32_common.h o_time.h o_str.h o_dir.h $(EXHEADER)
 
@@ -47,7 +50,7 @@ ALL=    $(GENERAL) $(SRC) $(HEADER)
 top:
 	@(cd ..; $(MAKE) DIRS=$(DIR) all)
 
-all: lib
+all: shared
 
 buildinf.h: ../Makefile
 	( echo "#ifndef MK1MF_BUILD"; \
@@ -57,26 +60,27 @@ buildinf.h: ../Makefile
 	echo "  #define DATE \"`LC_ALL=C LC_TIME=C date`\""; \
 	echo '#endif' ) >buildinf.h
 
-x86cpuid-elf.s:	x86cpuid.pl perlasm/x86asm.pl
-	$(PERL) x86cpuid.pl elf $(CFLAGS) $(PROCESSOR) > $@
-x86cpuid-cof.s: x86cpuid.pl perlasm/x86asm.pl
-	$(PERL) x86cpuid.pl coff $(CFLAGS) $(PROCESSOR) > $@
-x86cpuid-out.s: x86cpuid.pl perlasm/x86asm.pl
-	$(PERL) x86cpuid.pl a.out $(CFLAGS) $(PROCESSOR) > $@
+x86cpuid.s:	x86cpuid.pl perlasm/x86asm.pl
+	$(PERL) x86cpuid.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
 
-uplink.o:	../ms/uplink.c
-	$(CC) $(CFLAGS) -c -o $@ ../ms/uplink.c
+applink.o:	$(TOP)/ms/applink.c
+	$(CC) $(CFLAGS) -c -o $@ $(TOP)/ms/applink.c
 
-uplink-cof.s:	../ms/uplink.pl
-	$(PERL) ../ms/uplink.pl coff > $@
+uplink.o:	$(TOP)/ms/uplink.c applink.o
+	$(CC) $(CFLAGS) -c -o $@ $(TOP)/ms/uplink.c
 
-x86_64cpuid.s: x86_64cpuid.pl
-	$(PERL) x86_64cpuid.pl $@
-ia64cpuid.s: ia64cpuid.S
-	$(CC) $(CFLAGS) -E ia64cpuid.S > $@
+uplink-x86.s:	$(TOP)/ms/uplink-x86.pl
+	$(PERL) $(TOP)/ms/uplink-x86.pl $(PERLASM_SCHEME) > $@
+
+x86_64cpuid.s: x86_64cpuid.pl;	$(PERL) x86_64cpuid.pl $(PERLASM_SCHEME) > $@
+ia64cpuid.s: ia64cpuid.S;	$(CC) $(CFLAGS) -E ia64cpuid.S > $@
+ppccpuid.s:	ppccpuid.pl;	$(PERL) ppccpuid.pl $(PERLASM_SCHEME) $@
+pariscid.s:	pariscid.pl;	$(PERL) pariscid.pl $(PERLASM_SCHEME) $@
+alphacpuid.s:	alphacpuid.pl
+	$(PERL) $< | $(CC) -E - | tee $@ > /dev/null
 
 testapps:
-	[ -z "$(THIS)" ] || (	if echo ${SDIRS} | fgrep ' des '; \
+	[ -z "$(THIS)" ] || (	if echo $(SDIRS) | fgrep ' des '; \
 				then cd des && $(MAKE) -e des; fi )
 	[ -z "$(THIS)" ] || ( cd pkcs7 && $(MAKE) -e testapps );
 	@if [ -z "$(THIS)" ]; then $(MAKE) -f $(TOP)/Makefile reflect THIS=$@; fi
@@ -94,11 +98,12 @@ links:
 	@$(PERL) $(TOP)/util/mklink.pl ../apps $(APPS)
 	@target=links; $(RECURSIVE_MAKE)
 
-# lib: and $(LIB): are splitted to avoid end-less loop
-lib:	buildinf.h $(LIB) subdirs
+# lib: $(LIB): are splitted to avoid end-less loop
+lib:	$(LIB)
 	@touch lib
 $(LIB):	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
+	[ -z "$(FIPSLIBDIR)" ] || $(AR) $(LIB) $(FIPSLIBDIR)fipscanister.o
 	$(RANLIB) $(LIB) || echo Never mind.
 
 shared: buildinf.h lib subdirs
@@ -135,6 +140,7 @@ clean:
 dclean:
 	$(PERL) -pe 'if (/^# DO NOT DELETE THIS LINE/) {print; exit(0);}' $(MAKEFILE) >Makefile.new
 	mv -f Makefile.new $(MAKEFILE)
+	rm -f opensslconf.h
 	@target=dclean; $(RECURSIVE_MAKE)
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
@@ -159,13 +165,6 @@ cversion.o: ../include/openssl/opensslconf.h ../include/openssl/opensslv.h
 cversion.o: ../include/openssl/ossl_typ.h ../include/openssl/safestack.h
 cversion.o: ../include/openssl/stack.h ../include/openssl/symhacks.h buildinf.h
 cversion.o: cryptlib.h cversion.c
-dyn_lck.o: ../e_os.h ../include/openssl/bio.h ../include/openssl/buffer.h
-dyn_lck.o: ../include/openssl/crypto.h ../include/openssl/e_os2.h
-dyn_lck.o: ../include/openssl/err.h ../include/openssl/lhash.h
-dyn_lck.o: ../include/openssl/opensslconf.h ../include/openssl/opensslv.h
-dyn_lck.o: ../include/openssl/ossl_typ.h ../include/openssl/safestack.h
-dyn_lck.o: ../include/openssl/stack.h ../include/openssl/symhacks.h cryptlib.h
-dyn_lck.o: dyn_lck.c
 ebcdic.o: ../include/openssl/e_os2.h ../include/openssl/opensslconf.h ebcdic.c
 ex_data.o: ../e_os.h ../include/openssl/bio.h ../include/openssl/buffer.h
 ex_data.o: ../include/openssl/crypto.h ../include/openssl/e_os2.h
@@ -174,13 +173,7 @@ ex_data.o: ../include/openssl/opensslconf.h ../include/openssl/opensslv.h
 ex_data.o: ../include/openssl/ossl_typ.h ../include/openssl/safestack.h
 ex_data.o: ../include/openssl/stack.h ../include/openssl/symhacks.h cryptlib.h
 ex_data.o: ex_data.c
-fips_err.o: ../include/openssl/bio.h ../include/openssl/crypto.h
-fips_err.o: ../include/openssl/e_os2.h ../include/openssl/err.h
-fips_err.o: ../include/openssl/fips.h ../include/openssl/lhash.h
-fips_err.o: ../include/openssl/opensslconf.h ../include/openssl/opensslv.h
-fips_err.o: ../include/openssl/ossl_typ.h ../include/openssl/safestack.h
-fips_err.o: ../include/openssl/stack.h ../include/openssl/symhacks.h fips_err.c
-fips_err.o: fips_err.h
+fips_ers.o: ../include/openssl/opensslconf.h fips_ers.c
 mem.o: ../e_os.h ../include/openssl/bio.h ../include/openssl/buffer.h
 mem.o: ../include/openssl/crypto.h ../include/openssl/e_os2.h
 mem.o: ../include/openssl/err.h ../include/openssl/lhash.h
@@ -201,6 +194,13 @@ mem_dbg.o: ../include/openssl/stack.h ../include/openssl/symhacks.h cryptlib.h
 mem_dbg.o: mem_dbg.c
 o_dir.o: ../e_os.h ../include/openssl/e_os2.h ../include/openssl/opensslconf.h
 o_dir.o: LPdir_unix.c o_dir.c o_dir.h
+o_fips.o: ../e_os.h ../include/openssl/bio.h ../include/openssl/buffer.h
+o_fips.o: ../include/openssl/crypto.h ../include/openssl/e_os2.h
+o_fips.o: ../include/openssl/err.h ../include/openssl/lhash.h
+o_fips.o: ../include/openssl/opensslconf.h ../include/openssl/opensslv.h
+o_fips.o: ../include/openssl/ossl_typ.h ../include/openssl/safestack.h
+o_fips.o: ../include/openssl/stack.h ../include/openssl/symhacks.h cryptlib.h
+o_fips.o: o_fips.c
 o_init.o: ../e_os.h ../include/openssl/bio.h ../include/openssl/crypto.h
 o_init.o: ../include/openssl/e_os2.h ../include/openssl/err.h
 o_init.o: ../include/openssl/lhash.h ../include/openssl/opensslconf.h
@@ -211,13 +211,6 @@ o_str.o: ../e_os.h ../include/openssl/e_os2.h ../include/openssl/opensslconf.h
 o_str.o: o_str.c o_str.h
 o_time.o: ../include/openssl/e_os2.h ../include/openssl/opensslconf.h o_time.c
 o_time.o: o_time.h
-tmdiff.o: ../e_os.h ../include/openssl/bio.h ../include/openssl/buffer.h
-tmdiff.o: ../include/openssl/crypto.h ../include/openssl/e_os2.h
-tmdiff.o: ../include/openssl/err.h ../include/openssl/lhash.h
-tmdiff.o: ../include/openssl/opensslconf.h ../include/openssl/opensslv.h
-tmdiff.o: ../include/openssl/ossl_typ.h ../include/openssl/safestack.h
-tmdiff.o: ../include/openssl/stack.h ../include/openssl/symhacks.h
-tmdiff.o: ../include/openssl/tmdiff.h cryptlib.h tmdiff.c
 uid.o: ../include/openssl/crypto.h ../include/openssl/e_os2.h
 uid.o: ../include/openssl/opensslconf.h ../include/openssl/opensslv.h
 uid.o: ../include/openssl/ossl_typ.h ../include/openssl/safestack.h
diff --git a/crypto/aes/Makefile b/crypto/aes/Makefile
index 9d174f4c3ef8..45ede0a0b459 100644
--- a/crypto/aes/Makefile
+++ b/crypto/aes/Makefile
@@ -11,7 +11,7 @@ CFLAG=-g
 MAKEFILE=	Makefile
 AR=		ar r
 
-AES_ASM_OBJ=aes_core.o aes_cbc.o
+AES_ENC=aes_core.o aes_cbc.o
 
 CFLAGS= $(INCLUDES) $(CFLAG)
 ASFLAGS= $(INCLUDES) $(ASFLAG)
@@ -26,7 +26,7 @@ LIB=$(TOP)/libcrypto.a
 LIBSRC=aes_core.c aes_misc.c aes_ecb.c aes_cbc.c aes_cfb.c aes_ofb.c \
        aes_ctr.c aes_ige.c aes_wrap.c
 LIBOBJ=aes_misc.o aes_ecb.o aes_cfb.o aes_ofb.o aes_ctr.o aes_ige.o aes_wrap.o \
-       $(AES_ASM_OBJ)
+       $(AES_ENC)
 
 SRC= $(LIBSRC)
 
@@ -41,24 +41,46 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-$(LIBOBJ): $(LIBSRC)
-
 aes-ia64.s: asm/aes-ia64.S
 	$(CC) $(CFLAGS) -E asm/aes-ia64.S > $@
 
-ax86-elf.s: asm/aes-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) aes-586.pl elf $(CFLAGS) $(PROCESSOR) > ../$@)
-ax86-cof.s: asm/aes-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) aes-586.pl coff $(CFLAGS) $(PROCESSOR) > ../$@)
-ax86-out.s: asm/aes-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) aes-586.pl a.out $(CFLAGS) $(PROCESSOR) > ../$@)
+aes-586.s:	asm/aes-586.pl ../perlasm/x86asm.pl
+	$(PERL) asm/aes-586.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+vpaes-x86.s:	asm/vpaes-x86.pl ../perlasm/x86asm.pl
+	$(PERL) asm/vpaes-x86.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+aesni-x86.s:	asm/aesni-x86.pl ../perlasm/x86asm.pl
+	$(PERL) asm/aesni-x86.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
 
 aes-x86_64.s: asm/aes-x86_64.pl
-	$(PERL) asm/aes-x86_64.pl $@
+	$(PERL) asm/aes-x86_64.pl $(PERLASM_SCHEME) > $@
+vpaes-x86_64.s:	asm/vpaes-x86_64.pl
+	$(PERL) asm/vpaes-x86_64.pl $(PERLASM_SCHEME) > $@
+bsaes-x86_64.s:	asm/bsaes-x86_64.pl
+	$(PERL) asm/bsaes-x86_64.pl $(PERLASM_SCHEME) > $@
+aesni-x86_64.s: asm/aesni-x86_64.pl
+	$(PERL) asm/aesni-x86_64.pl $(PERLASM_SCHEME) > $@
+aesni-sha1-x86_64.s:	asm/aesni-sha1-x86_64.pl
+	$(PERL) asm/aesni-sha1-x86_64.pl $(PERLASM_SCHEME) > $@
+
+aes-sparcv9.s: asm/aes-sparcv9.pl
+	$(PERL) asm/aes-sparcv9.pl $(CFLAGS) > $@
+
+aes-ppc.s:	asm/aes-ppc.pl
+	$(PERL) asm/aes-ppc.pl $(PERLASM_SCHEME) $@
+
+aes-parisc.s:	asm/aes-parisc.pl
+	$(PERL) asm/aes-parisc.pl $(PERLASM_SCHEME) $@
+
+aes-mips.S:	asm/aes-mips.pl
+	$(PERL) asm/aes-mips.pl $(PERLASM_SCHEME) $@
+
+# GNU make "catch all"
+aes-%.S:	asm/aes-%.pl;	$(PERL) $< $(PERLASM_SCHEME) > $@
+aes-armv4.o:	aes-armv4.S
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
@@ -97,16 +119,14 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
-aes_cbc.o: ../../include/openssl/aes.h ../../include/openssl/e_os2.h
-aes_cbc.o: ../../include/openssl/opensslconf.h aes_cbc.c aes_locl.h
-aes_cfb.o: ../../e_os.h ../../include/openssl/aes.h
-aes_cfb.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-aes_cfb.o: aes_cfb.c aes_locl.h
+aes_cbc.o: ../../include/openssl/aes.h ../../include/openssl/modes.h
+aes_cbc.o: ../../include/openssl/opensslconf.h aes_cbc.c
+aes_cfb.o: ../../include/openssl/aes.h ../../include/openssl/modes.h
+aes_cfb.o: ../../include/openssl/opensslconf.h aes_cfb.c
 aes_core.o: ../../include/openssl/aes.h ../../include/openssl/e_os2.h
-aes_core.o: ../../include/openssl/fips.h ../../include/openssl/opensslconf.h
-aes_core.o: aes_core.c aes_locl.h
-aes_ctr.o: ../../include/openssl/aes.h ../../include/openssl/e_os2.h
-aes_ctr.o: ../../include/openssl/opensslconf.h aes_ctr.c aes_locl.h
+aes_core.o: ../../include/openssl/opensslconf.h aes_core.c aes_locl.h
+aes_ctr.o: ../../include/openssl/aes.h ../../include/openssl/modes.h
+aes_ctr.o: ../../include/openssl/opensslconf.h aes_ctr.c
 aes_ecb.o: ../../include/openssl/aes.h ../../include/openssl/e_os2.h
 aes_ecb.o: ../../include/openssl/opensslconf.h aes_ecb.c aes_locl.h
 aes_ige.o: ../../e_os.h ../../include/openssl/aes.h ../../include/openssl/bio.h
@@ -116,11 +136,13 @@ aes_ige.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 aes_ige.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 aes_ige.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 aes_ige.o: ../../include/openssl/symhacks.h ../cryptlib.h aes_ige.c aes_locl.h
-aes_misc.o: ../../include/openssl/aes.h ../../include/openssl/e_os2.h
-aes_misc.o: ../../include/openssl/opensslconf.h
-aes_misc.o: ../../include/openssl/opensslv.h aes_locl.h aes_misc.c
-aes_ofb.o: ../../include/openssl/aes.h ../../include/openssl/e_os2.h
-aes_ofb.o: ../../include/openssl/opensslconf.h aes_locl.h aes_ofb.c
+aes_misc.o: ../../include/openssl/aes.h ../../include/openssl/crypto.h
+aes_misc.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
+aes_misc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+aes_misc.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+aes_misc.o: ../../include/openssl/symhacks.h aes_locl.h aes_misc.c
+aes_ofb.o: ../../include/openssl/aes.h ../../include/openssl/modes.h
+aes_ofb.o: ../../include/openssl/opensslconf.h aes_ofb.c
 aes_wrap.o: ../../e_os.h ../../include/openssl/aes.h
 aes_wrap.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 aes_wrap.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
diff --git a/crypto/aes/aes.h b/crypto/aes/aes.h
index 450f2b4051b2..031abf01b50c 100644
--- a/crypto/aes/aes.h
+++ b/crypto/aes/aes.h
@@ -58,6 +58,8 @@
 #error AES is disabled.
 #endif
 
+#include <stddef.h>
+
 #define AES_ENCRYPT	1
 #define AES_DECRYPT	0
 
@@ -66,10 +68,6 @@
 #define AES_MAXNR 14
 #define AES_BLOCK_SIZE 16
 
-#ifdef OPENSSL_FIPS
-#define FIPS_AES_SIZE_T	int
-#endif
-
 #ifdef  __cplusplus
 extern "C" {
 #endif
@@ -92,6 +90,11 @@ int AES_set_encrypt_key(const unsigned char *userKey, const int bits,
 int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
 	AES_KEY *key);
 
+int private_AES_set_encrypt_key(const unsigned char *userKey, const int bits,
+	AES_KEY *key);
+int private_AES_set_decrypt_key(const unsigned char *userKey, const int bits,
+	AES_KEY *key);
+
 void AES_encrypt(const unsigned char *in, unsigned char *out,
 	const AES_KEY *key);
 void AES_decrypt(const unsigned char *in, unsigned char *out,
@@ -100,37 +103,32 @@ void AES_decrypt(const unsigned char *in, unsigned char *out,
 void AES_ecb_encrypt(const unsigned char *in, unsigned char *out,
 	const AES_KEY *key, const int enc);
 void AES_cbc_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const AES_KEY *key,
+	size_t length, const AES_KEY *key,
 	unsigned char *ivec, const int enc);
 void AES_cfb128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const AES_KEY *key,
+	size_t length, const AES_KEY *key,
 	unsigned char *ivec, int *num, const int enc);
 void AES_cfb1_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const AES_KEY *key,
+	size_t length, const AES_KEY *key,
 	unsigned char *ivec, int *num, const int enc);
 void AES_cfb8_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const AES_KEY *key,
+	size_t length, const AES_KEY *key,
 	unsigned char *ivec, int *num, const int enc);
-void AES_cfbr_encrypt_block(const unsigned char *in,unsigned char *out,
-			    const int nbits,const AES_KEY *key,
-			    unsigned char *ivec,const int enc);
 void AES_ofb128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const AES_KEY *key,
+	size_t length, const AES_KEY *key,
 	unsigned char *ivec, int *num);
 void AES_ctr128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const AES_KEY *key,
+	size_t length, const AES_KEY *key,
 	unsigned char ivec[AES_BLOCK_SIZE],
 	unsigned char ecount_buf[AES_BLOCK_SIZE],
 	unsigned int *num);
-
-/* For IGE, see also http://www.links.org/files/openssl-ige.pdf */
 /* NB: the IV is _two_ blocks long */
 void AES_ige_encrypt(const unsigned char *in, unsigned char *out,
-		     const unsigned long length, const AES_KEY *key,
+		     size_t length, const AES_KEY *key,
 		     unsigned char *ivec, const int enc);
 /* NB: the IV is _four_ blocks long */
 void AES_bi_ige_encrypt(const unsigned char *in, unsigned char *out,
-			const unsigned long length, const AES_KEY *key,
+			size_t length, const AES_KEY *key,
 			const AES_KEY *key2, const unsigned char *ivec,
 			const int enc);
 
@@ -141,6 +139,7 @@ int AES_unwrap_key(AES_KEY *key, const unsigned char *iv,
 		unsigned char *out,
 		const unsigned char *in, unsigned int inlen);
 
+
 #ifdef  __cplusplus
 }
 #endif
diff --git a/crypto/aes/aes_cbc.c b/crypto/aes/aes_cbc.c
index 373864cd4bdb..227f75625dba 100644
--- a/crypto/aes/aes_cbc.c
+++ b/crypto/aes/aes_cbc.c
@@ -49,85 +49,15 @@
  *
  */
 
-#ifndef AES_DEBUG
-# ifndef NDEBUG
-#  define NDEBUG
-# endif
-#endif
-#include <assert.h>
-
 #include <openssl/aes.h>
-#include "aes_locl.h"
+#include <openssl/modes.h>
 
-#if !defined(OPENSSL_FIPS_AES_ASM)
 void AES_cbc_encrypt(const unsigned char *in, unsigned char *out,
-		     const unsigned long length, const AES_KEY *key,
+		     size_t len, const AES_KEY *key,
 		     unsigned char *ivec, const int enc) {
 
-	unsigned long n;
-	unsigned long len = length;
-	unsigned char tmp[AES_BLOCK_SIZE];
-	const unsigned char *iv = ivec;
-
-	assert(in && out && key && ivec);
-	assert((AES_ENCRYPT == enc)||(AES_DECRYPT == enc));
-
-	if (AES_ENCRYPT == enc) {
-		while (len >= AES_BLOCK_SIZE) {
-			for(n=0; n < AES_BLOCK_SIZE; ++n)
-				out[n] = in[n] ^ iv[n];
-			AES_encrypt(out, out, key);
-			iv = out;
-			len -= AES_BLOCK_SIZE;
-			in += AES_BLOCK_SIZE;
-			out += AES_BLOCK_SIZE;
-		}
-		if (len) {
-			for(n=0; n < len; ++n)
-				out[n] = in[n] ^ iv[n];
-			for(n=len; n < AES_BLOCK_SIZE; ++n)
-				out[n] = iv[n];
-			AES_encrypt(out, out, key);
-			iv = out;
-		}
-		memcpy(ivec,iv,AES_BLOCK_SIZE);
-	} else if (in != out) {
-		while (len >= AES_BLOCK_SIZE) {
-			AES_decrypt(in, out, key);
-			for(n=0; n < AES_BLOCK_SIZE; ++n)
-				out[n] ^= iv[n];
-			iv = in;
-			len -= AES_BLOCK_SIZE;
-			in  += AES_BLOCK_SIZE;
-			out += AES_BLOCK_SIZE;
-		}
-		if (len) {
-			AES_decrypt(in,tmp,key);
-			for(n=0; n < len; ++n)
-				out[n] = tmp[n] ^ iv[n];
-			iv = in;
-		}
-		memcpy(ivec,iv,AES_BLOCK_SIZE);
-	} else {
-		while (len >= AES_BLOCK_SIZE) {
-			memcpy(tmp, in, AES_BLOCK_SIZE);
-			AES_decrypt(in, out, key);
-			for(n=0; n < AES_BLOCK_SIZE; ++n)
-				out[n] ^= ivec[n];
-			memcpy(ivec, tmp, AES_BLOCK_SIZE);
-			len -= AES_BLOCK_SIZE;
-			in += AES_BLOCK_SIZE;
-			out += AES_BLOCK_SIZE;
-		}
-		if (len) {
-			memcpy(tmp, in, AES_BLOCK_SIZE);
-			AES_decrypt(tmp, out, key);
-			for(n=0; n < len; ++n)
-				out[n] ^= ivec[n];
-			for(n=len; n < AES_BLOCK_SIZE; ++n)
-				out[n] = tmp[n];
-			memcpy(ivec, tmp, AES_BLOCK_SIZE);
-		}
-	}
+	if (enc)
+		CRYPTO_cbc128_encrypt(in,out,len,key,ivec,(block128_f)AES_encrypt);
+	else
+		CRYPTO_cbc128_decrypt(in,out,len,key,ivec,(block128_f)AES_decrypt);
 }
-#endif
diff --git a/crypto/aes/aes_cfb.c b/crypto/aes/aes_cfb.c
index 9384ba67f7f9..0c6d058ce729 100644
--- a/crypto/aes/aes_cfb.c
+++ b/crypto/aes/aes_cfb.c
@@ -1,6 +1,6 @@
 /* crypto/aes/aes_cfb.c -*- mode:C; c-file-style: "eay" -*- */
 /* ====================================================================
- * Copyright (c) 1998-2002 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2002-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -48,73 +48,9 @@
  * ====================================================================
  *
  */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#ifndef AES_DEBUG
-# ifndef NDEBUG
-#  define NDEBUG
-# endif
-#endif
-#include <assert.h>
 
 #include <openssl/aes.h>
-#include "aes_locl.h"
-#include "e_os.h"
+#include <openssl/modes.h>
 
 /* The input and output encrypted as though 128bit cfb mode is being
  * used.  The extra state information to record how much of the
@@ -122,103 +58,24 @@
  */
 
 void AES_cfb128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const AES_KEY *key,
+	size_t length, const AES_KEY *key,
 	unsigned char *ivec, int *num, const int enc) {
 
-	unsigned int n;
-	unsigned long l = length;
-	unsigned char c;
-
-	assert(in && out && key && ivec && num);
-
-	n = *num;
-
-	if (enc) {
-		while (l--) {
-			if (n == 0) {
-				AES_encrypt(ivec, ivec, key);
-			}
-			ivec[n] = *(out++) = *(in++) ^ ivec[n];
-			n = (n+1) % AES_BLOCK_SIZE;
-		}
-	} else {
-		while (l--) {
-			if (n == 0) {
-				AES_encrypt(ivec, ivec, key);
-			}
-			c = *(in);
-			*(out++) = *(in++) ^ ivec[n];
-			ivec[n] = c;
-			n = (n+1) % AES_BLOCK_SIZE;
-		}
-	}
-
-	*num=n;
+	CRYPTO_cfb128_encrypt(in,out,length,key,ivec,num,enc,(block128_f)AES_encrypt);
 }
 
-/* This expects a single block of size nbits for both in and out. Note that
-   it corrupts any extra bits in the last byte of out */
-void AES_cfbr_encrypt_block(const unsigned char *in,unsigned char *out,
-			    const int nbits,const AES_KEY *key,
-			    unsigned char *ivec,const int enc)
-    {
-    int n,rem,num;
-    unsigned char ovec[AES_BLOCK_SIZE*2];
-
-    if (nbits<=0 || nbits>128) return;
-
-	/* fill in the first half of the new IV with the current IV */
-	memcpy(ovec,ivec,AES_BLOCK_SIZE);
-	/* construct the new IV */
-	AES_encrypt(ivec,ivec,key);
-	num = (nbits+7)/8;
-	if (enc)	/* encrypt the input */
-	    for(n=0 ; n < num ; ++n)
-		out[n] = (ovec[AES_BLOCK_SIZE+n] = in[n] ^ ivec[n]);
-	else		/* decrypt the input */
-	    for(n=0 ; n < num ; ++n)
-		out[n] = (ovec[AES_BLOCK_SIZE+n] = in[n]) ^ ivec[n];
-	/* shift ovec left... */
-	rem = nbits%8;
-	num = nbits/8;
-	if(rem==0)
-	    memcpy(ivec,ovec+num,AES_BLOCK_SIZE);
-	else
-	    for(n=0 ; n < AES_BLOCK_SIZE ; ++n)
-		ivec[n] = ovec[n+num]<<rem | ovec[n+num+1]>>(8-rem);
-
-    /* it is not necessary to cleanse ovec, since the IV is not secret */
-    }
-
 /* N.B. This expects the input to be packed, MS bit first */
 void AES_cfb1_encrypt(const unsigned char *in, unsigned char *out,
-		      const unsigned long length, const AES_KEY *key,
+		      size_t length, const AES_KEY *key,
 		      unsigned char *ivec, int *num, const int enc)
     {
-    unsigned int n;
-    unsigned char c[1],d[1];
-
-    assert(in && out && key && ivec && num);
-    assert(*num == 0);
-
-    for(n=0 ; n < length ; ++n)
-	{
-	c[0]=(in[n/8]&(1 << (7-n%8))) ? 0x80 : 0;
-	AES_cfbr_encrypt_block(c,d,1,key,ivec,enc);
-	out[n/8]=(out[n/8]&~(1 << (7-n%8)))|((d[0]&0x80) >> (n%8));
-	}
+    CRYPTO_cfb128_1_encrypt(in,out,length,key,ivec,num,enc,(block128_f)AES_encrypt);
     }
 
 void AES_cfb8_encrypt(const unsigned char *in, unsigned char *out,
-		      const unsigned long length, const AES_KEY *key,
+		      size_t length, const AES_KEY *key,
 		      unsigned char *ivec, int *num, const int enc)
     {
-    unsigned int n;
-
-    assert(in && out && key && ivec && num);
-    assert(*num == 0);
-
-    for(n=0 ; n < length ; ++n)
-	AES_cfbr_encrypt_block(&in[n],&out[n],8,key,ivec,enc);
+    CRYPTO_cfb128_8_encrypt(in,out,length,key,ivec,num,enc,(block128_f)AES_encrypt);
     }
 
diff --git a/crypto/aes/aes_core.c b/crypto/aes/aes_core.c
index cffdd4daec4d..8f5210ac70a9 100644
--- a/crypto/aes/aes_core.c
+++ b/crypto/aes/aes_core.c
@@ -37,12 +37,9 @@
 
 #include <stdlib.h>
 #include <openssl/aes.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 #include "aes_locl.h"
 
+#ifndef AES_ASM
 /*
 Te0[x] = S [x].[02, 01, 01, 03];
 Te1[x] = S [x].[03, 02, 01, 01];
@@ -628,17 +625,13 @@ static const u32 rcon[] = {
 /**
  * Expand the cipher key into the encryption key schedule.
  */
-int AES_set_encrypt_key(const unsigned char *userKey, const int bits,
+int private_AES_set_encrypt_key(const unsigned char *userKey, const int bits,
 			AES_KEY *key) {
 
 	u32 *rk;
    	int i = 0;
 	u32 temp;
 
-#ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
-
 	if (!userKey || !key)
 		return -1;
 	if (bits != 128 && bits != 192 && bits != 256)
@@ -733,7 +726,7 @@ int AES_set_encrypt_key(const unsigned char *userKey, const int bits,
 /**
  * Expand the cipher key into the decryption key schedule.
  */
-int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
+int private_AES_set_decrypt_key(const unsigned char *userKey, const int bits,
 			 AES_KEY *key) {
 
         u32 *rk;
@@ -741,7 +734,7 @@ int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
 	u32 temp;
 
 	/* first, start with an encryption schedule */
-	status = AES_set_encrypt_key(userKey, bits, key);
+	status = private_AES_set_encrypt_key(userKey, bits, key);
 	if (status < 0)
 		return status;
 
@@ -781,7 +774,6 @@ int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
 	return 0;
 }
 
-#ifndef AES_ASM
 /*
  * Encrypt a single block
  * in and out can overlap
@@ -1164,4 +1156,203 @@ void AES_decrypt(const unsigned char *in, unsigned char *out,
 	PUTU32(out + 12, s3);
 }
 
+#else /* AES_ASM */
+
+static const u8 Te4[256] = {
+    0x63U, 0x7cU, 0x77U, 0x7bU, 0xf2U, 0x6bU, 0x6fU, 0xc5U,
+    0x30U, 0x01U, 0x67U, 0x2bU, 0xfeU, 0xd7U, 0xabU, 0x76U,
+    0xcaU, 0x82U, 0xc9U, 0x7dU, 0xfaU, 0x59U, 0x47U, 0xf0U,
+    0xadU, 0xd4U, 0xa2U, 0xafU, 0x9cU, 0xa4U, 0x72U, 0xc0U,
+    0xb7U, 0xfdU, 0x93U, 0x26U, 0x36U, 0x3fU, 0xf7U, 0xccU,
+    0x34U, 0xa5U, 0xe5U, 0xf1U, 0x71U, 0xd8U, 0x31U, 0x15U,
+    0x04U, 0xc7U, 0x23U, 0xc3U, 0x18U, 0x96U, 0x05U, 0x9aU,
+    0x07U, 0x12U, 0x80U, 0xe2U, 0xebU, 0x27U, 0xb2U, 0x75U,
+    0x09U, 0x83U, 0x2cU, 0x1aU, 0x1bU, 0x6eU, 0x5aU, 0xa0U,
+    0x52U, 0x3bU, 0xd6U, 0xb3U, 0x29U, 0xe3U, 0x2fU, 0x84U,
+    0x53U, 0xd1U, 0x00U, 0xedU, 0x20U, 0xfcU, 0xb1U, 0x5bU,
+    0x6aU, 0xcbU, 0xbeU, 0x39U, 0x4aU, 0x4cU, 0x58U, 0xcfU,
+    0xd0U, 0xefU, 0xaaU, 0xfbU, 0x43U, 0x4dU, 0x33U, 0x85U,
+    0x45U, 0xf9U, 0x02U, 0x7fU, 0x50U, 0x3cU, 0x9fU, 0xa8U,
+    0x51U, 0xa3U, 0x40U, 0x8fU, 0x92U, 0x9dU, 0x38U, 0xf5U,
+    0xbcU, 0xb6U, 0xdaU, 0x21U, 0x10U, 0xffU, 0xf3U, 0xd2U,
+    0xcdU, 0x0cU, 0x13U, 0xecU, 0x5fU, 0x97U, 0x44U, 0x17U,
+    0xc4U, 0xa7U, 0x7eU, 0x3dU, 0x64U, 0x5dU, 0x19U, 0x73U,
+    0x60U, 0x81U, 0x4fU, 0xdcU, 0x22U, 0x2aU, 0x90U, 0x88U,
+    0x46U, 0xeeU, 0xb8U, 0x14U, 0xdeU, 0x5eU, 0x0bU, 0xdbU,
+    0xe0U, 0x32U, 0x3aU, 0x0aU, 0x49U, 0x06U, 0x24U, 0x5cU,
+    0xc2U, 0xd3U, 0xacU, 0x62U, 0x91U, 0x95U, 0xe4U, 0x79U,
+    0xe7U, 0xc8U, 0x37U, 0x6dU, 0x8dU, 0xd5U, 0x4eU, 0xa9U,
+    0x6cU, 0x56U, 0xf4U, 0xeaU, 0x65U, 0x7aU, 0xaeU, 0x08U,
+    0xbaU, 0x78U, 0x25U, 0x2eU, 0x1cU, 0xa6U, 0xb4U, 0xc6U,
+    0xe8U, 0xddU, 0x74U, 0x1fU, 0x4bU, 0xbdU, 0x8bU, 0x8aU,
+    0x70U, 0x3eU, 0xb5U, 0x66U, 0x48U, 0x03U, 0xf6U, 0x0eU,
+    0x61U, 0x35U, 0x57U, 0xb9U, 0x86U, 0xc1U, 0x1dU, 0x9eU,
+    0xe1U, 0xf8U, 0x98U, 0x11U, 0x69U, 0xd9U, 0x8eU, 0x94U,
+    0x9bU, 0x1eU, 0x87U, 0xe9U, 0xceU, 0x55U, 0x28U, 0xdfU,
+    0x8cU, 0xa1U, 0x89U, 0x0dU, 0xbfU, 0xe6U, 0x42U, 0x68U,
+    0x41U, 0x99U, 0x2dU, 0x0fU, 0xb0U, 0x54U, 0xbbU, 0x16U
+};
+static const u32 rcon[] = {
+	0x01000000, 0x02000000, 0x04000000, 0x08000000,
+	0x10000000, 0x20000000, 0x40000000, 0x80000000,
+	0x1B000000, 0x36000000, /* for 128-bit blocks, Rijndael never uses more than 10 rcon values */
+};
+
+/**
+ * Expand the cipher key into the encryption key schedule.
+ */
+int private_AES_set_encrypt_key(const unsigned char *userKey, const int bits,
+			AES_KEY *key) {
+	u32 *rk;
+   	int i = 0;
+	u32 temp;
+
+	if (!userKey || !key)
+		return -1;
+	if (bits != 128 && bits != 192 && bits != 256)
+		return -2;
+
+	rk = key->rd_key;
+
+	if (bits==128)
+		key->rounds = 10;
+	else if (bits==192)
+		key->rounds = 12;
+	else
+		key->rounds = 14;
+
+	rk[0] = GETU32(userKey     );
+	rk[1] = GETU32(userKey +  4);
+	rk[2] = GETU32(userKey +  8);
+	rk[3] = GETU32(userKey + 12);
+	if (bits == 128) {
+		while (1) {
+			temp  = rk[3];
+			rk[4] = rk[0] ^
+				(Te4[(temp >> 16) & 0xff] << 24) ^
+				(Te4[(temp >>  8) & 0xff] << 16) ^
+				(Te4[(temp      ) & 0xff] << 8) ^
+				(Te4[(temp >> 24)       ]) ^
+				rcon[i];
+			rk[5] = rk[1] ^ rk[4];
+			rk[6] = rk[2] ^ rk[5];
+			rk[7] = rk[3] ^ rk[6];
+			if (++i == 10) {
+				return 0;
+			}
+			rk += 4;
+		}
+	}
+	rk[4] = GETU32(userKey + 16);
+	rk[5] = GETU32(userKey + 20);
+	if (bits == 192) {
+		while (1) {
+			temp = rk[ 5];
+			rk[ 6] = rk[ 0] ^
+				(Te4[(temp >> 16) & 0xff] << 24) ^
+				(Te4[(temp >>  8) & 0xff] << 16) ^
+				(Te4[(temp      ) & 0xff] << 8) ^
+				(Te4[(temp >> 24)       ]) ^
+				rcon[i];
+			rk[ 7] = rk[ 1] ^ rk[ 6];
+			rk[ 8] = rk[ 2] ^ rk[ 7];
+			rk[ 9] = rk[ 3] ^ rk[ 8];
+			if (++i == 8) {
+				return 0;
+			}
+			rk[10] = rk[ 4] ^ rk[ 9];
+			rk[11] = rk[ 5] ^ rk[10];
+			rk += 6;
+		}
+	}
+	rk[6] = GETU32(userKey + 24);
+	rk[7] = GETU32(userKey + 28);
+	if (bits == 256) {
+		while (1) {
+			temp = rk[ 7];
+			rk[ 8] = rk[ 0] ^
+				(Te4[(temp >> 16) & 0xff] << 24) ^
+				(Te4[(temp >>  8) & 0xff] << 16) ^
+				(Te4[(temp      ) & 0xff] << 8) ^
+				(Te4[(temp >> 24)       ]) ^
+				rcon[i];
+			rk[ 9] = rk[ 1] ^ rk[ 8];
+			rk[10] = rk[ 2] ^ rk[ 9];
+			rk[11] = rk[ 3] ^ rk[10];
+			if (++i == 7) {
+				return 0;
+			}
+			temp = rk[11];
+			rk[12] = rk[ 4] ^
+				(Te4[(temp >> 24)       ] << 24) ^
+				(Te4[(temp >> 16) & 0xff] << 16) ^
+				(Te4[(temp >>  8) & 0xff] << 8) ^
+				(Te4[(temp      ) & 0xff]);
+			rk[13] = rk[ 5] ^ rk[12];
+			rk[14] = rk[ 6] ^ rk[13];
+			rk[15] = rk[ 7] ^ rk[14];
+
+			rk += 8;
+        	}
+	}
+	return 0;
+}
+
+/**
+ * Expand the cipher key into the decryption key schedule.
+ */
+int private_AES_set_decrypt_key(const unsigned char *userKey, const int bits,
+			 AES_KEY *key) {
+
+        u32 *rk;
+	int i, j, status;
+	u32 temp;
+
+	/* first, start with an encryption schedule */
+	status = private_AES_set_encrypt_key(userKey, bits, key);
+	if (status < 0)
+		return status;
+
+	rk = key->rd_key;
+
+	/* invert the order of the round keys: */
+	for (i = 0, j = 4*(key->rounds); i < j; i += 4, j -= 4) {
+		temp = rk[i    ]; rk[i    ] = rk[j    ]; rk[j    ] = temp;
+		temp = rk[i + 1]; rk[i + 1] = rk[j + 1]; rk[j + 1] = temp;
+		temp = rk[i + 2]; rk[i + 2] = rk[j + 2]; rk[j + 2] = temp;
+		temp = rk[i + 3]; rk[i + 3] = rk[j + 3]; rk[j + 3] = temp;
+	}
+	/* apply the inverse MixColumn transform to all round keys but the first and the last: */
+	for (i = 1; i < (key->rounds); i++) {
+		rk += 4;
+		for (j = 0; j < 4; j++) {
+			u32 tp1, tp2, tp4, tp8, tp9, tpb, tpd, tpe, m;
+
+			tp1 = rk[j];
+			m = tp1 & 0x80808080;
+			tp2 = ((tp1 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp2 & 0x80808080;
+			tp4 = ((tp2 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp4 & 0x80808080;
+			tp8 = ((tp4 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			tp9 = tp8 ^ tp1;
+			tpb = tp9 ^ tp2;
+			tpd = tp9 ^ tp4;
+			tpe = tp8 ^ tp4 ^ tp2;
+#if defined(ROTATE)
+			rk[j] = tpe ^ ROTATE(tpd,16) ^
+				ROTATE(tp9,24) ^ ROTATE(tpb,8);
+#else
+			rk[j] = tpe ^ (tpd >> 16) ^ (tpd << 16) ^ 
+				(tp9 >> 8) ^ (tp9 << 24) ^
+				(tpb >> 24) ^ (tpb << 8);
+#endif
+		}
+	}
+	return 0;
+}
+
 #endif /* AES_ASM */
diff --git a/crypto/aes/aes_ctr.c b/crypto/aes/aes_ctr.c
index f36982be1e26..7c9d165d8adb 100644
--- a/crypto/aes/aes_ctr.c
+++ b/crypto/aes/aes_ctr.c
@@ -49,91 +49,13 @@
  *
  */
 
-#ifndef AES_DEBUG
-# ifndef NDEBUG
-#  define NDEBUG
-# endif
-#endif
-#include <assert.h>
-
 #include <openssl/aes.h>
-#include "aes_locl.h"
-
-/* NOTE: the IV/counter CTR mode is big-endian.  The rest of the AES code
- * is endian-neutral. */
-
-/* increment counter (128-bit int) by 1 */
-static void AES_ctr128_inc(unsigned char *counter) {
-	unsigned long c;
-
-	/* Grab bottom dword of counter and increment */
-	c = GETU32(counter + 12);
-	c++;	c &= 0xFFFFFFFF;
-	PUTU32(counter + 12, c);
-
-	/* if no overflow, we're done */
-	if (c)
-		return;
-
-	/* Grab 1st dword of counter and increment */
-	c = GETU32(counter +  8);
-	c++;	c &= 0xFFFFFFFF;
-	PUTU32(counter +  8, c);
-
-	/* if no overflow, we're done */
-	if (c)
-		return;
-
-	/* Grab 2nd dword of counter and increment */
-	c = GETU32(counter +  4);
-	c++;	c &= 0xFFFFFFFF;
-	PUTU32(counter +  4, c);
-
-	/* if no overflow, we're done */
-	if (c)
-		return;
+#include <openssl/modes.h>
 
-	/* Grab top dword of counter and increment */
-	c = GETU32(counter +  0);
-	c++;	c &= 0xFFFFFFFF;
-	PUTU32(counter +  0, c);
-}
-
-/* The input encrypted as though 128bit counter mode is being
- * used.  The extra state information to record how much of the
- * 128bit block we have used is contained in *num, and the
- * encrypted counter is kept in ecount_buf.  Both *num and
- * ecount_buf must be initialised with zeros before the first
- * call to AES_ctr128_encrypt().
- *
- * This algorithm assumes that the counter is in the x lower bits
- * of the IV (ivec), and that the application has full control over
- * overflow and the rest of the IV.  This implementation takes NO
- * responsability for checking that the counter doesn't overflow
- * into the rest of the IV when incremented.
- */
 void AES_ctr128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const AES_KEY *key,
-	unsigned char ivec[AES_BLOCK_SIZE],
-	unsigned char ecount_buf[AES_BLOCK_SIZE],
-	unsigned int *num) {
-
-	unsigned int n;
-	unsigned long l=length;
-
-	assert(in && out && key && counter && num);
-	assert(*num < AES_BLOCK_SIZE);
-
-	n = *num;
-
-	while (l--) {
-		if (n == 0) {
-			AES_encrypt(ivec, ecount_buf, key);
- 			AES_ctr128_inc(ivec);
-		}
-		*(out++) = *(in++) ^ ecount_buf[n];
-		n = (n+1) % AES_BLOCK_SIZE;
-	}
-
-	*num=n;
+			size_t length, const AES_KEY *key,
+			unsigned char ivec[AES_BLOCK_SIZE],
+			unsigned char ecount_buf[AES_BLOCK_SIZE],
+			unsigned int *num) {
+	CRYPTO_ctr128_encrypt(in,out,length,key,ivec,ecount_buf,num,(block128_f)AES_encrypt);
 }
diff --git a/crypto/aes/aes_ige.c b/crypto/aes/aes_ige.c
index 45d709618187..c161351e6545 100644
--- a/crypto/aes/aes_ige.c
+++ b/crypto/aes/aes_ige.c
@@ -77,11 +77,11 @@ typedef struct {
 /* N.B. The IV for this mode is _twice_ the block size */
 
 void AES_ige_encrypt(const unsigned char *in, unsigned char *out,
-					 const unsigned long length, const AES_KEY *key,
+					 size_t length, const AES_KEY *key,
 					 unsigned char *ivec, const int enc)
 	{
-	unsigned long n;
-	unsigned long len;
+	size_t n;
+	size_t len = length;
 
 	OPENSSL_assert(in && out && key && ivec);
 	OPENSSL_assert((AES_ENCRYPT == enc)||(AES_DECRYPT == enc));
@@ -211,12 +211,12 @@ void AES_ige_encrypt(const unsigned char *in, unsigned char *out,
 /* N.B. The IV for this mode is _four times_ the block size */
 
 void AES_bi_ige_encrypt(const unsigned char *in, unsigned char *out,
-						const unsigned long length, const AES_KEY *key,
+						size_t length, const AES_KEY *key,
 						const AES_KEY *key2, const unsigned char *ivec,
 						const int enc)
 	{
-	unsigned long n;
-	unsigned long len = length;
+	size_t n;
+	size_t len = length;
 	unsigned char tmp[AES_BLOCK_SIZE];
 	unsigned char tmp2[AES_BLOCK_SIZE];
 	unsigned char tmp3[AES_BLOCK_SIZE];
diff --git a/crypto/aes/aes_misc.c b/crypto/aes/aes_misc.c
index 4fead1b4c7c5..f083488ecb34 100644
--- a/crypto/aes/aes_misc.c
+++ b/crypto/aes/aes_misc.c
@@ -50,6 +50,7 @@
  */
 
 #include <openssl/opensslv.h>
+#include <openssl/crypto.h>
 #include <openssl/aes.h>
 #include "aes_locl.h"
 
@@ -62,3 +63,23 @@ const char *AES_options(void) {
         return "aes(partial)";
 #endif
 }
+
+/* FIPS wrapper functions to block low level AES calls in FIPS mode */
+
+int AES_set_encrypt_key(const unsigned char *userKey, const int bits,
+			AES_KEY *key)
+	{
+#ifdef OPENSSL_FIPS
+	fips_cipher_abort(AES);
+#endif
+	return private_AES_set_encrypt_key(userKey, bits, key);
+	}
+
+int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
+			AES_KEY *key)
+	{
+#ifdef OPENSSL_FIPS
+	fips_cipher_abort(AES);
+#endif
+	return private_AES_set_decrypt_key(userKey, bits, key);
+	}
diff --git a/crypto/aes/aes_ofb.c b/crypto/aes/aes_ofb.c
index f358bb39e27c..50bf0b832504 100644
--- a/crypto/aes/aes_ofb.c
+++ b/crypto/aes/aes_ofb.c
@@ -1,6 +1,6 @@
 /* crypto/aes/aes_ofb.c -*- mode:C; c-file-style: "eay" -*- */
 /* ====================================================================
- * Copyright (c) 1998-2002 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2002-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -48,95 +48,13 @@
  * ====================================================================
  *
  */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#ifndef AES_DEBUG
-# ifndef NDEBUG
-#  define NDEBUG
-# endif
-#endif
-#include <assert.h>
 
 #include <openssl/aes.h>
-#include "aes_locl.h"
+#include <openssl/modes.h>
 
-/* The input and output encrypted as though 128bit ofb mode is being
- * used.  The extra state information to record how much of the
- * 128bit block we have used is contained in *num;
- */
 void AES_ofb128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const AES_KEY *key,
-	unsigned char *ivec, int *num) {
-
-	unsigned int n;
-	unsigned long l=length;
-
-	assert(in && out && key && ivec && num);
-
-	n = *num;
-
-	while (l--) {
-		if (n == 0) {
-			AES_encrypt(ivec, ivec, key);
-		}
-		*(out++) = *(in++) ^ ivec[n];
-		n = (n+1) % AES_BLOCK_SIZE;
-	}
-
-	*num=n;
+	size_t length, const AES_KEY *key,
+	unsigned char *ivec, int *num)
+{
+	CRYPTO_ofb128_encrypt(in,out,length,key,ivec,num,(block128_f)AES_encrypt);
 }
diff --git a/crypto/aes/aes_x86core.c b/crypto/aes/aes_x86core.c
new file mode 100644
index 000000000000..d323e265c05a
--- /dev/null
+++ b/crypto/aes/aes_x86core.c
@@ -0,0 +1,1063 @@
+/* crypto/aes/aes_core.c -*- mode:C; c-file-style: "eay" -*- */
+/**
+ * rijndael-alg-fst.c
+ *
+ * @version 3.0 (December 2000)
+ *
+ * Optimised ANSI C code for the Rijndael cipher (now AES)
+ *
+ * @author Vincent Rijmen <vincent.rijmen@esat.kuleuven.ac.be>
+ * @author Antoon Bosselaers <antoon.bosselaers@esat.kuleuven.ac.be>
+ * @author Paulo Barreto <paulo.barreto@terra.com.br>
+ *
+ * This code is hereby placed in the public domain.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ''AS IS'' AND ANY EXPRESS
+ * OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
+ * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+ * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/*
+ * This is experimental x86[_64] derivative. It assumes little-endian
+ * byte order and expects CPU to sustain unaligned memory references.
+ * It is used as playground for cache-time attack mitigations and
+ * serves as reference C implementation for x86[_64] assembler.
+ *
+ *					<appro@fy.chalmers.se>
+ */
+
+
+#ifndef AES_DEBUG
+# ifndef NDEBUG
+#  define NDEBUG
+# endif
+#endif
+#include <assert.h>
+
+#include <stdlib.h>
+#include <openssl/aes.h>
+#include "aes_locl.h"
+
+/*
+ * These two parameters control which table, 256-byte or 2KB, is
+ * referenced in outer and respectively inner rounds.
+ */
+#define AES_COMPACT_IN_OUTER_ROUNDS
+#ifdef  AES_COMPACT_IN_OUTER_ROUNDS
+/* AES_COMPACT_IN_OUTER_ROUNDS costs ~30% in performance, while
+ * adding AES_COMPACT_IN_INNER_ROUNDS reduces benchmark *further*
+ * by factor of ~2. */
+# undef  AES_COMPACT_IN_INNER_ROUNDS
+#endif
+
+#if 1
+static void prefetch256(const void *table)
+{
+	volatile unsigned long *t=(void *)table,ret;
+	unsigned long sum;
+	int i;
+
+	/* 32 is common least cache-line size */
+	for (sum=0,i=0;i<256/sizeof(t[0]);i+=32/sizeof(t[0]))	sum ^= t[i];
+
+	ret = sum;
+}
+#else
+# define prefetch256(t)
+#endif
+
+#undef GETU32
+#define GETU32(p) (*((u32*)(p)))
+
+#if (defined(_WIN32) || defined(_WIN64)) && !defined(__MINGW32__)
+typedef unsigned __int64 u64;
+#define U64(C)	C##UI64
+#elif defined(__arch64__)
+typedef unsigned long u64;
+#define U64(C)	C##UL
+#else
+typedef unsigned long long u64;
+#define U64(C)	C##ULL
+#endif
+
+#undef ROTATE
+#if defined(_MSC_VER) || defined(__ICC)
+# define ROTATE(a,n)	_lrotl(a,n)
+#elif defined(__GNUC__) && __GNUC__>=2
+# if defined(__i386) || defined(__i386__) || defined(__x86_64) || defined(__x86_64__)
+#   define ROTATE(a,n)	({ register unsigned int ret;	\
+				asm (			\
+				"roll %1,%0"		\
+				: "=r"(ret)		\
+				: "I"(n), "0"(a)	\
+				: "cc");		\
+			   ret;				\
+			})
+# endif
+#endif
+/*
+Te [x] = S [x].[02, 01, 01, 03, 02, 01, 01, 03];
+Te0[x] = S [x].[02, 01, 01, 03];
+Te1[x] = S [x].[03, 02, 01, 01];
+Te2[x] = S [x].[01, 03, 02, 01];
+Te3[x] = S [x].[01, 01, 03, 02];
+*/
+#define Te0 (u32)((u64*)((u8*)Te+0))
+#define Te1 (u32)((u64*)((u8*)Te+3))
+#define Te2 (u32)((u64*)((u8*)Te+2))
+#define Te3 (u32)((u64*)((u8*)Te+1))
+/*
+Td [x] = Si[x].[0e, 09, 0d, 0b, 0e, 09, 0d, 0b];
+Td0[x] = Si[x].[0e, 09, 0d, 0b];
+Td1[x] = Si[x].[0b, 0e, 09, 0d];
+Td2[x] = Si[x].[0d, 0b, 0e, 09];
+Td3[x] = Si[x].[09, 0d, 0b, 0e];
+Td4[x] = Si[x].[01];
+*/
+#define Td0 (u32)((u64*)((u8*)Td+0))
+#define Td1 (u32)((u64*)((u8*)Td+3))
+#define Td2 (u32)((u64*)((u8*)Td+2))
+#define Td3 (u32)((u64*)((u8*)Td+1))
+
+static const u64 Te[256] = {
+    U64(0xa56363c6a56363c6), U64(0x847c7cf8847c7cf8),
+    U64(0x997777ee997777ee), U64(0x8d7b7bf68d7b7bf6),
+    U64(0x0df2f2ff0df2f2ff), U64(0xbd6b6bd6bd6b6bd6),
+    U64(0xb16f6fdeb16f6fde), U64(0x54c5c59154c5c591),
+    U64(0x5030306050303060), U64(0x0301010203010102),
+    U64(0xa96767cea96767ce), U64(0x7d2b2b567d2b2b56),
+    U64(0x19fefee719fefee7), U64(0x62d7d7b562d7d7b5),
+    U64(0xe6abab4de6abab4d), U64(0x9a7676ec9a7676ec),
+    U64(0x45caca8f45caca8f), U64(0x9d82821f9d82821f),
+    U64(0x40c9c98940c9c989), U64(0x877d7dfa877d7dfa),
+    U64(0x15fafaef15fafaef), U64(0xeb5959b2eb5959b2),
+    U64(0xc947478ec947478e), U64(0x0bf0f0fb0bf0f0fb),
+    U64(0xecadad41ecadad41), U64(0x67d4d4b367d4d4b3),
+    U64(0xfda2a25ffda2a25f), U64(0xeaafaf45eaafaf45),
+    U64(0xbf9c9c23bf9c9c23), U64(0xf7a4a453f7a4a453),
+    U64(0x967272e4967272e4), U64(0x5bc0c09b5bc0c09b),
+    U64(0xc2b7b775c2b7b775), U64(0x1cfdfde11cfdfde1),
+    U64(0xae93933dae93933d), U64(0x6a26264c6a26264c),
+    U64(0x5a36366c5a36366c), U64(0x413f3f7e413f3f7e),
+    U64(0x02f7f7f502f7f7f5), U64(0x4fcccc834fcccc83),
+    U64(0x5c3434685c343468), U64(0xf4a5a551f4a5a551),
+    U64(0x34e5e5d134e5e5d1), U64(0x08f1f1f908f1f1f9),
+    U64(0x937171e2937171e2), U64(0x73d8d8ab73d8d8ab),
+    U64(0x5331316253313162), U64(0x3f15152a3f15152a),
+    U64(0x0c0404080c040408), U64(0x52c7c79552c7c795),
+    U64(0x6523234665232346), U64(0x5ec3c39d5ec3c39d),
+    U64(0x2818183028181830), U64(0xa1969637a1969637),
+    U64(0x0f05050a0f05050a), U64(0xb59a9a2fb59a9a2f),
+    U64(0x0907070e0907070e), U64(0x3612122436121224),
+    U64(0x9b80801b9b80801b), U64(0x3de2e2df3de2e2df),
+    U64(0x26ebebcd26ebebcd), U64(0x6927274e6927274e),
+    U64(0xcdb2b27fcdb2b27f), U64(0x9f7575ea9f7575ea),
+    U64(0x1b0909121b090912), U64(0x9e83831d9e83831d),
+    U64(0x742c2c58742c2c58), U64(0x2e1a1a342e1a1a34),
+    U64(0x2d1b1b362d1b1b36), U64(0xb26e6edcb26e6edc),
+    U64(0xee5a5ab4ee5a5ab4), U64(0xfba0a05bfba0a05b),
+    U64(0xf65252a4f65252a4), U64(0x4d3b3b764d3b3b76),
+    U64(0x61d6d6b761d6d6b7), U64(0xceb3b37dceb3b37d),
+    U64(0x7b2929527b292952), U64(0x3ee3e3dd3ee3e3dd),
+    U64(0x712f2f5e712f2f5e), U64(0x9784841397848413),
+    U64(0xf55353a6f55353a6), U64(0x68d1d1b968d1d1b9),
+    U64(0x0000000000000000), U64(0x2cededc12cededc1),
+    U64(0x6020204060202040), U64(0x1ffcfce31ffcfce3),
+    U64(0xc8b1b179c8b1b179), U64(0xed5b5bb6ed5b5bb6),
+    U64(0xbe6a6ad4be6a6ad4), U64(0x46cbcb8d46cbcb8d),
+    U64(0xd9bebe67d9bebe67), U64(0x4b3939724b393972),
+    U64(0xde4a4a94de4a4a94), U64(0xd44c4c98d44c4c98),
+    U64(0xe85858b0e85858b0), U64(0x4acfcf854acfcf85),
+    U64(0x6bd0d0bb6bd0d0bb), U64(0x2aefefc52aefefc5),
+    U64(0xe5aaaa4fe5aaaa4f), U64(0x16fbfbed16fbfbed),
+    U64(0xc5434386c5434386), U64(0xd74d4d9ad74d4d9a),
+    U64(0x5533336655333366), U64(0x9485851194858511),
+    U64(0xcf45458acf45458a), U64(0x10f9f9e910f9f9e9),
+    U64(0x0602020406020204), U64(0x817f7ffe817f7ffe),
+    U64(0xf05050a0f05050a0), U64(0x443c3c78443c3c78),
+    U64(0xba9f9f25ba9f9f25), U64(0xe3a8a84be3a8a84b),
+    U64(0xf35151a2f35151a2), U64(0xfea3a35dfea3a35d),
+    U64(0xc0404080c0404080), U64(0x8a8f8f058a8f8f05),
+    U64(0xad92923fad92923f), U64(0xbc9d9d21bc9d9d21),
+    U64(0x4838387048383870), U64(0x04f5f5f104f5f5f1),
+    U64(0xdfbcbc63dfbcbc63), U64(0xc1b6b677c1b6b677),
+    U64(0x75dadaaf75dadaaf), U64(0x6321214263212142),
+    U64(0x3010102030101020), U64(0x1affffe51affffe5),
+    U64(0x0ef3f3fd0ef3f3fd), U64(0x6dd2d2bf6dd2d2bf),
+    U64(0x4ccdcd814ccdcd81), U64(0x140c0c18140c0c18),
+    U64(0x3513132635131326), U64(0x2fececc32fececc3),
+    U64(0xe15f5fbee15f5fbe), U64(0xa2979735a2979735),
+    U64(0xcc444488cc444488), U64(0x3917172e3917172e),
+    U64(0x57c4c49357c4c493), U64(0xf2a7a755f2a7a755),
+    U64(0x827e7efc827e7efc), U64(0x473d3d7a473d3d7a),
+    U64(0xac6464c8ac6464c8), U64(0xe75d5dbae75d5dba),
+    U64(0x2b1919322b191932), U64(0x957373e6957373e6),
+    U64(0xa06060c0a06060c0), U64(0x9881811998818119),
+    U64(0xd14f4f9ed14f4f9e), U64(0x7fdcdca37fdcdca3),
+    U64(0x6622224466222244), U64(0x7e2a2a547e2a2a54),
+    U64(0xab90903bab90903b), U64(0x8388880b8388880b),
+    U64(0xca46468cca46468c), U64(0x29eeeec729eeeec7),
+    U64(0xd3b8b86bd3b8b86b), U64(0x3c1414283c141428),
+    U64(0x79dedea779dedea7), U64(0xe25e5ebce25e5ebc),
+    U64(0x1d0b0b161d0b0b16), U64(0x76dbdbad76dbdbad),
+    U64(0x3be0e0db3be0e0db), U64(0x5632326456323264),
+    U64(0x4e3a3a744e3a3a74), U64(0x1e0a0a141e0a0a14),
+    U64(0xdb494992db494992), U64(0x0a06060c0a06060c),
+    U64(0x6c2424486c242448), U64(0xe45c5cb8e45c5cb8),
+    U64(0x5dc2c29f5dc2c29f), U64(0x6ed3d3bd6ed3d3bd),
+    U64(0xefacac43efacac43), U64(0xa66262c4a66262c4),
+    U64(0xa8919139a8919139), U64(0xa4959531a4959531),
+    U64(0x37e4e4d337e4e4d3), U64(0x8b7979f28b7979f2),
+    U64(0x32e7e7d532e7e7d5), U64(0x43c8c88b43c8c88b),
+    U64(0x5937376e5937376e), U64(0xb76d6ddab76d6dda),
+    U64(0x8c8d8d018c8d8d01), U64(0x64d5d5b164d5d5b1),
+    U64(0xd24e4e9cd24e4e9c), U64(0xe0a9a949e0a9a949),
+    U64(0xb46c6cd8b46c6cd8), U64(0xfa5656acfa5656ac),
+    U64(0x07f4f4f307f4f4f3), U64(0x25eaeacf25eaeacf),
+    U64(0xaf6565caaf6565ca), U64(0x8e7a7af48e7a7af4),
+    U64(0xe9aeae47e9aeae47), U64(0x1808081018080810),
+    U64(0xd5baba6fd5baba6f), U64(0x887878f0887878f0),
+    U64(0x6f25254a6f25254a), U64(0x722e2e5c722e2e5c),
+    U64(0x241c1c38241c1c38), U64(0xf1a6a657f1a6a657),
+    U64(0xc7b4b473c7b4b473), U64(0x51c6c69751c6c697),
+    U64(0x23e8e8cb23e8e8cb), U64(0x7cdddda17cdddda1),
+    U64(0x9c7474e89c7474e8), U64(0x211f1f3e211f1f3e),
+    U64(0xdd4b4b96dd4b4b96), U64(0xdcbdbd61dcbdbd61),
+    U64(0x868b8b0d868b8b0d), U64(0x858a8a0f858a8a0f),
+    U64(0x907070e0907070e0), U64(0x423e3e7c423e3e7c),
+    U64(0xc4b5b571c4b5b571), U64(0xaa6666ccaa6666cc),
+    U64(0xd8484890d8484890), U64(0x0503030605030306),
+    U64(0x01f6f6f701f6f6f7), U64(0x120e0e1c120e0e1c),
+    U64(0xa36161c2a36161c2), U64(0x5f35356a5f35356a),
+    U64(0xf95757aef95757ae), U64(0xd0b9b969d0b9b969),
+    U64(0x9186861791868617), U64(0x58c1c19958c1c199),
+    U64(0x271d1d3a271d1d3a), U64(0xb99e9e27b99e9e27),
+    U64(0x38e1e1d938e1e1d9), U64(0x13f8f8eb13f8f8eb),
+    U64(0xb398982bb398982b), U64(0x3311112233111122),
+    U64(0xbb6969d2bb6969d2), U64(0x70d9d9a970d9d9a9),
+    U64(0x898e8e07898e8e07), U64(0xa7949433a7949433),
+    U64(0xb69b9b2db69b9b2d), U64(0x221e1e3c221e1e3c),
+    U64(0x9287871592878715), U64(0x20e9e9c920e9e9c9),
+    U64(0x49cece8749cece87), U64(0xff5555aaff5555aa),
+    U64(0x7828285078282850), U64(0x7adfdfa57adfdfa5),
+    U64(0x8f8c8c038f8c8c03), U64(0xf8a1a159f8a1a159),
+    U64(0x8089890980898909), U64(0x170d0d1a170d0d1a),
+    U64(0xdabfbf65dabfbf65), U64(0x31e6e6d731e6e6d7),
+    U64(0xc6424284c6424284), U64(0xb86868d0b86868d0),
+    U64(0xc3414182c3414182), U64(0xb0999929b0999929),
+    U64(0x772d2d5a772d2d5a), U64(0x110f0f1e110f0f1e),
+    U64(0xcbb0b07bcbb0b07b), U64(0xfc5454a8fc5454a8),
+    U64(0xd6bbbb6dd6bbbb6d), U64(0x3a16162c3a16162c)
+};
+
+static const u8 Te4[256] = {
+    0x63U, 0x7cU, 0x77U, 0x7bU, 0xf2U, 0x6bU, 0x6fU, 0xc5U,
+    0x30U, 0x01U, 0x67U, 0x2bU, 0xfeU, 0xd7U, 0xabU, 0x76U,
+    0xcaU, 0x82U, 0xc9U, 0x7dU, 0xfaU, 0x59U, 0x47U, 0xf0U,
+    0xadU, 0xd4U, 0xa2U, 0xafU, 0x9cU, 0xa4U, 0x72U, 0xc0U,
+    0xb7U, 0xfdU, 0x93U, 0x26U, 0x36U, 0x3fU, 0xf7U, 0xccU,
+    0x34U, 0xa5U, 0xe5U, 0xf1U, 0x71U, 0xd8U, 0x31U, 0x15U,
+    0x04U, 0xc7U, 0x23U, 0xc3U, 0x18U, 0x96U, 0x05U, 0x9aU,
+    0x07U, 0x12U, 0x80U, 0xe2U, 0xebU, 0x27U, 0xb2U, 0x75U,
+    0x09U, 0x83U, 0x2cU, 0x1aU, 0x1bU, 0x6eU, 0x5aU, 0xa0U,
+    0x52U, 0x3bU, 0xd6U, 0xb3U, 0x29U, 0xe3U, 0x2fU, 0x84U,
+    0x53U, 0xd1U, 0x00U, 0xedU, 0x20U, 0xfcU, 0xb1U, 0x5bU,
+    0x6aU, 0xcbU, 0xbeU, 0x39U, 0x4aU, 0x4cU, 0x58U, 0xcfU,
+    0xd0U, 0xefU, 0xaaU, 0xfbU, 0x43U, 0x4dU, 0x33U, 0x85U,
+    0x45U, 0xf9U, 0x02U, 0x7fU, 0x50U, 0x3cU, 0x9fU, 0xa8U,
+    0x51U, 0xa3U, 0x40U, 0x8fU, 0x92U, 0x9dU, 0x38U, 0xf5U,
+    0xbcU, 0xb6U, 0xdaU, 0x21U, 0x10U, 0xffU, 0xf3U, 0xd2U,
+    0xcdU, 0x0cU, 0x13U, 0xecU, 0x5fU, 0x97U, 0x44U, 0x17U,
+    0xc4U, 0xa7U, 0x7eU, 0x3dU, 0x64U, 0x5dU, 0x19U, 0x73U,
+    0x60U, 0x81U, 0x4fU, 0xdcU, 0x22U, 0x2aU, 0x90U, 0x88U,
+    0x46U, 0xeeU, 0xb8U, 0x14U, 0xdeU, 0x5eU, 0x0bU, 0xdbU,
+    0xe0U, 0x32U, 0x3aU, 0x0aU, 0x49U, 0x06U, 0x24U, 0x5cU,
+    0xc2U, 0xd3U, 0xacU, 0x62U, 0x91U, 0x95U, 0xe4U, 0x79U,
+    0xe7U, 0xc8U, 0x37U, 0x6dU, 0x8dU, 0xd5U, 0x4eU, 0xa9U,
+    0x6cU, 0x56U, 0xf4U, 0xeaU, 0x65U, 0x7aU, 0xaeU, 0x08U,
+    0xbaU, 0x78U, 0x25U, 0x2eU, 0x1cU, 0xa6U, 0xb4U, 0xc6U,
+    0xe8U, 0xddU, 0x74U, 0x1fU, 0x4bU, 0xbdU, 0x8bU, 0x8aU,
+    0x70U, 0x3eU, 0xb5U, 0x66U, 0x48U, 0x03U, 0xf6U, 0x0eU,
+    0x61U, 0x35U, 0x57U, 0xb9U, 0x86U, 0xc1U, 0x1dU, 0x9eU,
+    0xe1U, 0xf8U, 0x98U, 0x11U, 0x69U, 0xd9U, 0x8eU, 0x94U,
+    0x9bU, 0x1eU, 0x87U, 0xe9U, 0xceU, 0x55U, 0x28U, 0xdfU,
+    0x8cU, 0xa1U, 0x89U, 0x0dU, 0xbfU, 0xe6U, 0x42U, 0x68U,
+    0x41U, 0x99U, 0x2dU, 0x0fU, 0xb0U, 0x54U, 0xbbU, 0x16U
+};
+
+static const u64 Td[256] = {
+    U64(0x50a7f45150a7f451), U64(0x5365417e5365417e),
+    U64(0xc3a4171ac3a4171a), U64(0x965e273a965e273a),
+    U64(0xcb6bab3bcb6bab3b), U64(0xf1459d1ff1459d1f),
+    U64(0xab58faacab58faac), U64(0x9303e34b9303e34b),
+    U64(0x55fa302055fa3020), U64(0xf66d76adf66d76ad),
+    U64(0x9176cc889176cc88), U64(0x254c02f5254c02f5),
+    U64(0xfcd7e54ffcd7e54f), U64(0xd7cb2ac5d7cb2ac5),
+    U64(0x8044352680443526), U64(0x8fa362b58fa362b5),
+    U64(0x495ab1de495ab1de), U64(0x671bba25671bba25),
+    U64(0x980eea45980eea45), U64(0xe1c0fe5de1c0fe5d),
+    U64(0x02752fc302752fc3), U64(0x12f04c8112f04c81),
+    U64(0xa397468da397468d), U64(0xc6f9d36bc6f9d36b),
+    U64(0xe75f8f03e75f8f03), U64(0x959c9215959c9215),
+    U64(0xeb7a6dbfeb7a6dbf), U64(0xda595295da595295),
+    U64(0x2d83bed42d83bed4), U64(0xd3217458d3217458),
+    U64(0x2969e0492969e049), U64(0x44c8c98e44c8c98e),
+    U64(0x6a89c2756a89c275), U64(0x78798ef478798ef4),
+    U64(0x6b3e58996b3e5899), U64(0xdd71b927dd71b927),
+    U64(0xb64fe1beb64fe1be), U64(0x17ad88f017ad88f0),
+    U64(0x66ac20c966ac20c9), U64(0xb43ace7db43ace7d),
+    U64(0x184adf63184adf63), U64(0x82311ae582311ae5),
+    U64(0x6033519760335197), U64(0x457f5362457f5362),
+    U64(0xe07764b1e07764b1), U64(0x84ae6bbb84ae6bbb),
+    U64(0x1ca081fe1ca081fe), U64(0x942b08f9942b08f9),
+    U64(0x5868487058684870), U64(0x19fd458f19fd458f),
+    U64(0x876cde94876cde94), U64(0xb7f87b52b7f87b52),
+    U64(0x23d373ab23d373ab), U64(0xe2024b72e2024b72),
+    U64(0x578f1fe3578f1fe3), U64(0x2aab55662aab5566),
+    U64(0x0728ebb20728ebb2), U64(0x03c2b52f03c2b52f),
+    U64(0x9a7bc5869a7bc586), U64(0xa50837d3a50837d3),
+    U64(0xf2872830f2872830), U64(0xb2a5bf23b2a5bf23),
+    U64(0xba6a0302ba6a0302), U64(0x5c8216ed5c8216ed),
+    U64(0x2b1ccf8a2b1ccf8a), U64(0x92b479a792b479a7),
+    U64(0xf0f207f3f0f207f3), U64(0xa1e2694ea1e2694e),
+    U64(0xcdf4da65cdf4da65), U64(0xd5be0506d5be0506),
+    U64(0x1f6234d11f6234d1), U64(0x8afea6c48afea6c4),
+    U64(0x9d532e349d532e34), U64(0xa055f3a2a055f3a2),
+    U64(0x32e18a0532e18a05), U64(0x75ebf6a475ebf6a4),
+    U64(0x39ec830b39ec830b), U64(0xaaef6040aaef6040),
+    U64(0x069f715e069f715e), U64(0x51106ebd51106ebd),
+    U64(0xf98a213ef98a213e), U64(0x3d06dd963d06dd96),
+    U64(0xae053eddae053edd), U64(0x46bde64d46bde64d),
+    U64(0xb58d5491b58d5491), U64(0x055dc471055dc471),
+    U64(0x6fd406046fd40604), U64(0xff155060ff155060),
+    U64(0x24fb981924fb9819), U64(0x97e9bdd697e9bdd6),
+    U64(0xcc434089cc434089), U64(0x779ed967779ed967),
+    U64(0xbd42e8b0bd42e8b0), U64(0x888b8907888b8907),
+    U64(0x385b19e7385b19e7), U64(0xdbeec879dbeec879),
+    U64(0x470a7ca1470a7ca1), U64(0xe90f427ce90f427c),
+    U64(0xc91e84f8c91e84f8), U64(0x0000000000000000),
+    U64(0x8386800983868009), U64(0x48ed2b3248ed2b32),
+    U64(0xac70111eac70111e), U64(0x4e725a6c4e725a6c),
+    U64(0xfbff0efdfbff0efd), U64(0x5638850f5638850f),
+    U64(0x1ed5ae3d1ed5ae3d), U64(0x27392d3627392d36),
+    U64(0x64d90f0a64d90f0a), U64(0x21a65c6821a65c68),
+    U64(0xd1545b9bd1545b9b), U64(0x3a2e36243a2e3624),
+    U64(0xb1670a0cb1670a0c), U64(0x0fe757930fe75793),
+    U64(0xd296eeb4d296eeb4), U64(0x9e919b1b9e919b1b),
+    U64(0x4fc5c0804fc5c080), U64(0xa220dc61a220dc61),
+    U64(0x694b775a694b775a), U64(0x161a121c161a121c),
+    U64(0x0aba93e20aba93e2), U64(0xe52aa0c0e52aa0c0),
+    U64(0x43e0223c43e0223c), U64(0x1d171b121d171b12),
+    U64(0x0b0d090e0b0d090e), U64(0xadc78bf2adc78bf2),
+    U64(0xb9a8b62db9a8b62d), U64(0xc8a91e14c8a91e14),
+    U64(0x8519f1578519f157), U64(0x4c0775af4c0775af),
+    U64(0xbbdd99eebbdd99ee), U64(0xfd607fa3fd607fa3),
+    U64(0x9f2601f79f2601f7), U64(0xbcf5725cbcf5725c),
+    U64(0xc53b6644c53b6644), U64(0x347efb5b347efb5b),
+    U64(0x7629438b7629438b), U64(0xdcc623cbdcc623cb),
+    U64(0x68fcedb668fcedb6), U64(0x63f1e4b863f1e4b8),
+    U64(0xcadc31d7cadc31d7), U64(0x1085634210856342),
+    U64(0x4022971340229713), U64(0x2011c6842011c684),
+    U64(0x7d244a857d244a85), U64(0xf83dbbd2f83dbbd2),
+    U64(0x1132f9ae1132f9ae), U64(0x6da129c76da129c7),
+    U64(0x4b2f9e1d4b2f9e1d), U64(0xf330b2dcf330b2dc),
+    U64(0xec52860dec52860d), U64(0xd0e3c177d0e3c177),
+    U64(0x6c16b32b6c16b32b), U64(0x99b970a999b970a9),
+    U64(0xfa489411fa489411), U64(0x2264e9472264e947),
+    U64(0xc48cfca8c48cfca8), U64(0x1a3ff0a01a3ff0a0),
+    U64(0xd82c7d56d82c7d56), U64(0xef903322ef903322),
+    U64(0xc74e4987c74e4987), U64(0xc1d138d9c1d138d9),
+    U64(0xfea2ca8cfea2ca8c), U64(0x360bd498360bd498),
+    U64(0xcf81f5a6cf81f5a6), U64(0x28de7aa528de7aa5),
+    U64(0x268eb7da268eb7da), U64(0xa4bfad3fa4bfad3f),
+    U64(0xe49d3a2ce49d3a2c), U64(0x0d9278500d927850),
+    U64(0x9bcc5f6a9bcc5f6a), U64(0x62467e5462467e54),
+    U64(0xc2138df6c2138df6), U64(0xe8b8d890e8b8d890),
+    U64(0x5ef7392e5ef7392e), U64(0xf5afc382f5afc382),
+    U64(0xbe805d9fbe805d9f), U64(0x7c93d0697c93d069),
+    U64(0xa92dd56fa92dd56f), U64(0xb31225cfb31225cf),
+    U64(0x3b99acc83b99acc8), U64(0xa77d1810a77d1810),
+    U64(0x6e639ce86e639ce8), U64(0x7bbb3bdb7bbb3bdb),
+    U64(0x097826cd097826cd), U64(0xf418596ef418596e),
+    U64(0x01b79aec01b79aec), U64(0xa89a4f83a89a4f83),
+    U64(0x656e95e6656e95e6), U64(0x7ee6ffaa7ee6ffaa),
+    U64(0x08cfbc2108cfbc21), U64(0xe6e815efe6e815ef),
+    U64(0xd99be7bad99be7ba), U64(0xce366f4ace366f4a),
+    U64(0xd4099fead4099fea), U64(0xd67cb029d67cb029),
+    U64(0xafb2a431afb2a431), U64(0x31233f2a31233f2a),
+    U64(0x3094a5c63094a5c6), U64(0xc066a235c066a235),
+    U64(0x37bc4e7437bc4e74), U64(0xa6ca82fca6ca82fc),
+    U64(0xb0d090e0b0d090e0), U64(0x15d8a73315d8a733),
+    U64(0x4a9804f14a9804f1), U64(0xf7daec41f7daec41),
+    U64(0x0e50cd7f0e50cd7f), U64(0x2ff691172ff69117),
+    U64(0x8dd64d768dd64d76), U64(0x4db0ef434db0ef43),
+    U64(0x544daacc544daacc), U64(0xdf0496e4df0496e4),
+    U64(0xe3b5d19ee3b5d19e), U64(0x1b886a4c1b886a4c),
+    U64(0xb81f2cc1b81f2cc1), U64(0x7f5165467f516546),
+    U64(0x04ea5e9d04ea5e9d), U64(0x5d358c015d358c01),
+    U64(0x737487fa737487fa), U64(0x2e410bfb2e410bfb),
+    U64(0x5a1d67b35a1d67b3), U64(0x52d2db9252d2db92),
+    U64(0x335610e9335610e9), U64(0x1347d66d1347d66d),
+    U64(0x8c61d79a8c61d79a), U64(0x7a0ca1377a0ca137),
+    U64(0x8e14f8598e14f859), U64(0x893c13eb893c13eb),
+    U64(0xee27a9ceee27a9ce), U64(0x35c961b735c961b7),
+    U64(0xede51ce1ede51ce1), U64(0x3cb1477a3cb1477a),
+    U64(0x59dfd29c59dfd29c), U64(0x3f73f2553f73f255),
+    U64(0x79ce141879ce1418), U64(0xbf37c773bf37c773),
+    U64(0xeacdf753eacdf753), U64(0x5baafd5f5baafd5f),
+    U64(0x146f3ddf146f3ddf), U64(0x86db447886db4478),
+    U64(0x81f3afca81f3afca), U64(0x3ec468b93ec468b9),
+    U64(0x2c3424382c342438), U64(0x5f40a3c25f40a3c2),
+    U64(0x72c31d1672c31d16), U64(0x0c25e2bc0c25e2bc),
+    U64(0x8b493c288b493c28), U64(0x41950dff41950dff),
+    U64(0x7101a8397101a839), U64(0xdeb30c08deb30c08),
+    U64(0x9ce4b4d89ce4b4d8), U64(0x90c1566490c15664),
+    U64(0x6184cb7b6184cb7b), U64(0x70b632d570b632d5),
+    U64(0x745c6c48745c6c48), U64(0x4257b8d04257b8d0)
+};
+static const u8 Td4[256] = {
+    0x52U, 0x09U, 0x6aU, 0xd5U, 0x30U, 0x36U, 0xa5U, 0x38U,
+    0xbfU, 0x40U, 0xa3U, 0x9eU, 0x81U, 0xf3U, 0xd7U, 0xfbU,
+    0x7cU, 0xe3U, 0x39U, 0x82U, 0x9bU, 0x2fU, 0xffU, 0x87U,
+    0x34U, 0x8eU, 0x43U, 0x44U, 0xc4U, 0xdeU, 0xe9U, 0xcbU,
+    0x54U, 0x7bU, 0x94U, 0x32U, 0xa6U, 0xc2U, 0x23U, 0x3dU,
+    0xeeU, 0x4cU, 0x95U, 0x0bU, 0x42U, 0xfaU, 0xc3U, 0x4eU,
+    0x08U, 0x2eU, 0xa1U, 0x66U, 0x28U, 0xd9U, 0x24U, 0xb2U,
+    0x76U, 0x5bU, 0xa2U, 0x49U, 0x6dU, 0x8bU, 0xd1U, 0x25U,
+    0x72U, 0xf8U, 0xf6U, 0x64U, 0x86U, 0x68U, 0x98U, 0x16U,
+    0xd4U, 0xa4U, 0x5cU, 0xccU, 0x5dU, 0x65U, 0xb6U, 0x92U,
+    0x6cU, 0x70U, 0x48U, 0x50U, 0xfdU, 0xedU, 0xb9U, 0xdaU,
+    0x5eU, 0x15U, 0x46U, 0x57U, 0xa7U, 0x8dU, 0x9dU, 0x84U,
+    0x90U, 0xd8U, 0xabU, 0x00U, 0x8cU, 0xbcU, 0xd3U, 0x0aU,
+    0xf7U, 0xe4U, 0x58U, 0x05U, 0xb8U, 0xb3U, 0x45U, 0x06U,
+    0xd0U, 0x2cU, 0x1eU, 0x8fU, 0xcaU, 0x3fU, 0x0fU, 0x02U,
+    0xc1U, 0xafU, 0xbdU, 0x03U, 0x01U, 0x13U, 0x8aU, 0x6bU,
+    0x3aU, 0x91U, 0x11U, 0x41U, 0x4fU, 0x67U, 0xdcU, 0xeaU,
+    0x97U, 0xf2U, 0xcfU, 0xceU, 0xf0U, 0xb4U, 0xe6U, 0x73U,
+    0x96U, 0xacU, 0x74U, 0x22U, 0xe7U, 0xadU, 0x35U, 0x85U,
+    0xe2U, 0xf9U, 0x37U, 0xe8U, 0x1cU, 0x75U, 0xdfU, 0x6eU,
+    0x47U, 0xf1U, 0x1aU, 0x71U, 0x1dU, 0x29U, 0xc5U, 0x89U,
+    0x6fU, 0xb7U, 0x62U, 0x0eU, 0xaaU, 0x18U, 0xbeU, 0x1bU,
+    0xfcU, 0x56U, 0x3eU, 0x4bU, 0xc6U, 0xd2U, 0x79U, 0x20U,
+    0x9aU, 0xdbU, 0xc0U, 0xfeU, 0x78U, 0xcdU, 0x5aU, 0xf4U,
+    0x1fU, 0xddU, 0xa8U, 0x33U, 0x88U, 0x07U, 0xc7U, 0x31U,
+    0xb1U, 0x12U, 0x10U, 0x59U, 0x27U, 0x80U, 0xecU, 0x5fU,
+    0x60U, 0x51U, 0x7fU, 0xa9U, 0x19U, 0xb5U, 0x4aU, 0x0dU,
+    0x2dU, 0xe5U, 0x7aU, 0x9fU, 0x93U, 0xc9U, 0x9cU, 0xefU,
+    0xa0U, 0xe0U, 0x3bU, 0x4dU, 0xaeU, 0x2aU, 0xf5U, 0xb0U,
+    0xc8U, 0xebU, 0xbbU, 0x3cU, 0x83U, 0x53U, 0x99U, 0x61U,
+    0x17U, 0x2bU, 0x04U, 0x7eU, 0xbaU, 0x77U, 0xd6U, 0x26U,
+    0xe1U, 0x69U, 0x14U, 0x63U, 0x55U, 0x21U, 0x0cU, 0x7dU
+};
+
+static const u32 rcon[] = {
+    0x00000001U, 0x00000002U, 0x00000004U, 0x00000008U,
+    0x00000010U, 0x00000020U, 0x00000040U, 0x00000080U,
+    0x0000001bU, 0x00000036U, /* for 128-bit blocks, Rijndael never uses more than 10 rcon values */
+};
+
+/**
+ * Expand the cipher key into the encryption key schedule.
+ */
+int AES_set_encrypt_key(const unsigned char *userKey, const int bits,
+			AES_KEY *key) {
+
+	u32 *rk;
+   	int i = 0;
+	u32 temp;
+
+	if (!userKey || !key)
+		return -1;
+	if (bits != 128 && bits != 192 && bits != 256)
+		return -2;
+
+	rk = key->rd_key;
+
+	if (bits==128)
+		key->rounds = 10;
+	else if (bits==192)
+		key->rounds = 12;
+	else
+		key->rounds = 14;
+
+	rk[0] = GETU32(userKey     );
+	rk[1] = GETU32(userKey +  4);
+	rk[2] = GETU32(userKey +  8);
+	rk[3] = GETU32(userKey + 12);
+	if (bits == 128) {
+		while (1) {
+			temp  = rk[3];
+			rk[4] = rk[0] ^
+				(Te4[(temp >>  8) & 0xff]      ) ^
+				(Te4[(temp >> 16) & 0xff] <<  8) ^
+				(Te4[(temp >> 24)       ] << 16) ^
+				(Te4[(temp      ) & 0xff] << 24) ^
+				rcon[i];
+			rk[5] = rk[1] ^ rk[4];
+			rk[6] = rk[2] ^ rk[5];
+			rk[7] = rk[3] ^ rk[6];
+			if (++i == 10) {
+				return 0;
+			}
+			rk += 4;
+		}
+	}
+	rk[4] = GETU32(userKey + 16);
+	rk[5] = GETU32(userKey + 20);
+	if (bits == 192) {
+		while (1) {
+			temp = rk[ 5];
+			rk[ 6] = rk[ 0] ^
+				(Te4[(temp >>  8) & 0xff]      ) ^
+				(Te4[(temp >> 16) & 0xff] <<  8) ^
+				(Te4[(temp >> 24)       ] << 16) ^
+				(Te4[(temp      ) & 0xff] << 24) ^
+				rcon[i];
+			rk[ 7] = rk[ 1] ^ rk[ 6];
+			rk[ 8] = rk[ 2] ^ rk[ 7];
+			rk[ 9] = rk[ 3] ^ rk[ 8];
+			if (++i == 8) {
+				return 0;
+			}
+			rk[10] = rk[ 4] ^ rk[ 9];
+			rk[11] = rk[ 5] ^ rk[10];
+			rk += 6;
+		}
+	}
+	rk[6] = GETU32(userKey + 24);
+	rk[7] = GETU32(userKey + 28);
+	if (bits == 256) {
+		while (1) {
+			temp = rk[ 7];
+			rk[ 8] = rk[ 0] ^
+				(Te4[(temp >>  8) & 0xff]      ) ^
+				(Te4[(temp >> 16) & 0xff] <<  8) ^
+				(Te4[(temp >> 24)       ] << 16) ^
+				(Te4[(temp      ) & 0xff] << 24) ^
+				rcon[i];
+			rk[ 9] = rk[ 1] ^ rk[ 8];
+			rk[10] = rk[ 2] ^ rk[ 9];
+			rk[11] = rk[ 3] ^ rk[10];
+			if (++i == 7) {
+				return 0;
+			}
+			temp = rk[11];
+			rk[12] = rk[ 4] ^
+				(Te4[(temp      ) & 0xff]      ) ^
+				(Te4[(temp >>  8) & 0xff] <<  8) ^
+				(Te4[(temp >> 16) & 0xff] << 16) ^
+				(Te4[(temp >> 24)       ] << 24);
+			rk[13] = rk[ 5] ^ rk[12];
+			rk[14] = rk[ 6] ^ rk[13];
+			rk[15] = rk[ 7] ^ rk[14];
+
+			rk += 8;
+        	}
+	}
+	return 0;
+}
+
+/**
+ * Expand the cipher key into the decryption key schedule.
+ */
+int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
+			 AES_KEY *key) {
+
+        u32 *rk;
+	int i, j, status;
+	u32 temp;
+
+	/* first, start with an encryption schedule */
+	status = AES_set_encrypt_key(userKey, bits, key);
+	if (status < 0)
+		return status;
+
+	rk = key->rd_key;
+
+	/* invert the order of the round keys: */
+	for (i = 0, j = 4*(key->rounds); i < j; i += 4, j -= 4) {
+		temp = rk[i    ]; rk[i    ] = rk[j    ]; rk[j    ] = temp;
+		temp = rk[i + 1]; rk[i + 1] = rk[j + 1]; rk[j + 1] = temp;
+		temp = rk[i + 2]; rk[i + 2] = rk[j + 2]; rk[j + 2] = temp;
+		temp = rk[i + 3]; rk[i + 3] = rk[j + 3]; rk[j + 3] = temp;
+	}
+	/* apply the inverse MixColumn transform to all round keys but the first and the last: */
+	for (i = 1; i < (key->rounds); i++) {
+		rk += 4;
+#if 1
+		for (j = 0; j < 4; j++) {
+			u32 tp1, tp2, tp4, tp8, tp9, tpb, tpd, tpe, m;
+
+			tp1 = rk[j];
+			m = tp1 & 0x80808080;
+			tp2 = ((tp1 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp2 & 0x80808080;
+			tp4 = ((tp2 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp4 & 0x80808080;
+			tp8 = ((tp4 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			tp9 = tp8 ^ tp1;
+			tpb = tp9 ^ tp2;
+			tpd = tp9 ^ tp4;
+			tpe = tp8 ^ tp4 ^ tp2;
+#if defined(ROTATE)
+			rk[j] = tpe ^ ROTATE(tpd,16) ^
+				ROTATE(tp9,8) ^ ROTATE(tpb,24);
+#else
+			rk[j] = tpe ^ (tpd >> 16) ^ (tpd << 16) ^ 
+				(tp9 >> 24) ^ (tp9 << 8) ^
+				(tpb >> 8) ^ (tpb << 24);
+#endif
+		}
+#else
+		rk[0] =
+			Td0[Te2[(rk[0]      ) & 0xff] & 0xff] ^
+			Td1[Te2[(rk[0] >>  8) & 0xff] & 0xff] ^
+			Td2[Te2[(rk[0] >> 16) & 0xff] & 0xff] ^
+			Td3[Te2[(rk[0] >> 24)       ] & 0xff];
+		rk[1] =
+			Td0[Te2[(rk[1]      ) & 0xff] & 0xff] ^
+			Td1[Te2[(rk[1] >>  8) & 0xff] & 0xff] ^
+			Td2[Te2[(rk[1] >> 16) & 0xff] & 0xff] ^
+			Td3[Te2[(rk[1] >> 24)       ] & 0xff];
+		rk[2] =
+			Td0[Te2[(rk[2]      ) & 0xff] & 0xff] ^
+			Td1[Te2[(rk[2] >>  8) & 0xff] & 0xff] ^
+			Td2[Te2[(rk[2] >> 16) & 0xff] & 0xff] ^
+			Td3[Te2[(rk[2] >> 24)       ] & 0xff];
+		rk[3] =
+			Td0[Te2[(rk[3]      ) & 0xff] & 0xff] ^
+			Td1[Te2[(rk[3] >>  8) & 0xff] & 0xff] ^
+			Td2[Te2[(rk[3] >> 16) & 0xff] & 0xff] ^
+			Td3[Te2[(rk[3] >> 24)       ] & 0xff];
+#endif
+	}
+	return 0;
+}
+
+/*
+ * Encrypt a single block
+ * in and out can overlap
+ */
+void AES_encrypt(const unsigned char *in, unsigned char *out,
+		 const AES_KEY *key) {
+
+	const u32 *rk;
+	u32 s0, s1, s2, s3, t[4];
+	int r;
+
+	assert(in && out && key);
+	rk = key->rd_key;
+
+	/*
+	 * map byte array block to cipher state
+	 * and add initial round key:
+	 */
+	s0 = GETU32(in     ) ^ rk[0];
+	s1 = GETU32(in +  4) ^ rk[1];
+	s2 = GETU32(in +  8) ^ rk[2];
+	s3 = GETU32(in + 12) ^ rk[3];
+
+#if defined(AES_COMPACT_IN_OUTER_ROUNDS)
+	prefetch256(Te4);
+
+	t[0] =	Te4[(s0      ) & 0xff]       ^
+		Te4[(s1 >>  8) & 0xff] <<  8 ^
+		Te4[(s2 >> 16) & 0xff] << 16 ^
+		Te4[(s3 >> 24)       ] << 24;
+	t[1] =	Te4[(s1      ) & 0xff]       ^
+		Te4[(s2 >>  8) & 0xff] <<  8 ^
+		Te4[(s3 >> 16) & 0xff] << 16 ^
+		Te4[(s0 >> 24)       ] << 24;
+	t[2] =	Te4[(s2      ) & 0xff]       ^
+		Te4[(s3 >>  8) & 0xff] <<  8 ^
+		Te4[(s0 >> 16) & 0xff] << 16 ^
+		Te4[(s1 >> 24)       ] << 24;
+	t[3] =	Te4[(s3      ) & 0xff]       ^
+		Te4[(s0 >>  8) & 0xff] <<  8 ^
+		Te4[(s1 >> 16) & 0xff] << 16 ^
+		Te4[(s2 >> 24)       ] << 24;
+
+	/* now do the linear transform using words */
+	{	int i;
+		u32 r0, r1, r2;
+
+		for (i = 0; i < 4; i++) {
+			r0 = t[i];
+			r1 = r0 & 0x80808080;
+			r2 = ((r0 & 0x7f7f7f7f) << 1) ^
+				((r1 - (r1 >> 7)) & 0x1b1b1b1b);
+#if defined(ROTATE)
+			t[i] = r2 ^ ROTATE(r2,24) ^ ROTATE(r0,24) ^
+				ROTATE(r0,16) ^ ROTATE(r0,8);
+#else
+			t[i] = r2 ^ ((r2 ^ r0) << 24) ^ ((r2 ^ r0) >> 8) ^
+				(r0 << 16) ^ (r0 >> 16) ^
+				(r0 << 8) ^ (r0 >> 24);
+#endif
+			t[i] ^= rk[4+i];
+		}
+	}
+#else
+	t[0] =	Te0[(s0      ) & 0xff] ^
+		Te1[(s1 >>  8) & 0xff] ^
+		Te2[(s2 >> 16) & 0xff] ^
+		Te3[(s3 >> 24)       ] ^
+		rk[4];
+	t[1] =	Te0[(s1      ) & 0xff] ^
+		Te1[(s2 >>  8) & 0xff] ^
+		Te2[(s3 >> 16) & 0xff] ^
+		Te3[(s0 >> 24)       ] ^
+		rk[5];
+	t[2] =	Te0[(s2      ) & 0xff] ^
+		Te1[(s3 >>  8) & 0xff] ^
+		Te2[(s0 >> 16) & 0xff] ^
+		Te3[(s1 >> 24)       ] ^
+		rk[6];
+	t[3] =	Te0[(s3      ) & 0xff] ^
+		Te1[(s0 >>  8) & 0xff] ^
+		Te2[(s1 >> 16) & 0xff] ^
+		Te3[(s2 >> 24)       ] ^
+		rk[7];
+#endif
+	s0 = t[0]; s1 = t[1]; s2 = t[2]; s3 = t[3];
+
+    /*
+     * Nr - 2 full rounds:
+     */
+    for (rk+=8,r=key->rounds-2; r>0; rk+=4,r--) {
+#if defined(AES_COMPACT_IN_INNER_ROUNDS)
+	t[0] =	Te4[(s0      ) & 0xff]       ^
+		Te4[(s1 >>  8) & 0xff] <<  8 ^
+		Te4[(s2 >> 16) & 0xff] << 16 ^
+		Te4[(s3 >> 24)       ] << 24;
+	t[1] =	Te4[(s1      ) & 0xff]       ^
+		Te4[(s2 >>  8) & 0xff] <<  8 ^
+		Te4[(s3 >> 16) & 0xff] << 16 ^
+		Te4[(s0 >> 24)       ] << 24;
+	t[2] =	Te4[(s2      ) & 0xff]       ^
+		Te4[(s3 >>  8) & 0xff] <<  8 ^
+		Te4[(s0 >> 16) & 0xff] << 16 ^
+		Te4[(s1 >> 24)       ] << 24;
+	t[3] =	Te4[(s3      ) & 0xff]       ^
+		Te4[(s0 >>  8) & 0xff] <<  8 ^
+		Te4[(s1 >> 16) & 0xff] << 16 ^
+		Te4[(s2 >> 24)       ] << 24;
+
+	/* now do the linear transform using words */
+	{	int i;
+		u32 r0, r1, r2;
+
+		for (i = 0; i < 4; i++) {
+			r0 = t[i];
+			r1 = r0 & 0x80808080;
+			r2 = ((r0 & 0x7f7f7f7f) << 1) ^
+				((r1 - (r1 >> 7)) & 0x1b1b1b1b);
+#if defined(ROTATE)
+			t[i] = r2 ^ ROTATE(r2,24) ^ ROTATE(r0,24) ^
+				ROTATE(r0,16) ^ ROTATE(r0,8);
+#else
+			t[i] = r2 ^ ((r2 ^ r0) << 24) ^ ((r2 ^ r0) >> 8) ^
+				(r0 << 16) ^ (r0 >> 16) ^
+				(r0 << 8) ^ (r0 >> 24);
+#endif
+			t[i] ^= rk[i];
+		}
+	}
+#else
+	t[0] =	Te0[(s0      ) & 0xff] ^
+		Te1[(s1 >>  8) & 0xff] ^
+		Te2[(s2 >> 16) & 0xff] ^
+		Te3[(s3 >> 24)       ] ^
+		rk[0];
+	t[1] =	Te0[(s1      ) & 0xff] ^
+		Te1[(s2 >>  8) & 0xff] ^
+		Te2[(s3 >> 16) & 0xff] ^
+		Te3[(s0 >> 24)       ] ^
+		rk[1];
+	t[2] =	Te0[(s2      ) & 0xff] ^
+		Te1[(s3 >>  8) & 0xff] ^
+		Te2[(s0 >> 16) & 0xff] ^
+		Te3[(s1 >> 24)       ] ^
+		rk[2];
+	t[3] =	Te0[(s3      ) & 0xff] ^
+		Te1[(s0 >>  8) & 0xff] ^
+		Te2[(s1 >> 16) & 0xff] ^
+		Te3[(s2 >> 24)       ] ^
+		rk[3];
+#endif
+	s0 = t[0]; s1 = t[1]; s2 = t[2]; s3 = t[3];
+    }
+    /*
+	 * apply last round and
+	 * map cipher state to byte array block:
+	 */
+#if defined(AES_COMPACT_IN_OUTER_ROUNDS)
+	prefetch256(Te4);
+
+	*(u32*)(out+0) =
+		Te4[(s0      ) & 0xff]       ^
+		Te4[(s1 >>  8) & 0xff] <<  8 ^
+		Te4[(s2 >> 16) & 0xff] << 16 ^
+		Te4[(s3 >> 24)       ] << 24 ^
+		rk[0];
+	*(u32*)(out+4) =
+		Te4[(s1      ) & 0xff]       ^
+		Te4[(s2 >>  8) & 0xff] <<  8 ^
+		Te4[(s3 >> 16) & 0xff] << 16 ^
+		Te4[(s0 >> 24)       ] << 24 ^
+		rk[1];
+	*(u32*)(out+8) =
+		Te4[(s2      ) & 0xff]       ^
+		Te4[(s3 >>  8) & 0xff] <<  8 ^
+		Te4[(s0 >> 16) & 0xff] << 16 ^
+		Te4[(s1 >> 24)       ] << 24 ^
+		rk[2];
+	*(u32*)(out+12) =
+		Te4[(s3      ) & 0xff]       ^
+		Te4[(s0 >>  8) & 0xff] <<  8 ^
+		Te4[(s1 >> 16) & 0xff] << 16 ^
+		Te4[(s2 >> 24)       ] << 24 ^
+		rk[3];
+#else
+	*(u32*)(out+0) =
+		(Te2[(s0      ) & 0xff] & 0x000000ffU) ^
+		(Te3[(s1 >>  8) & 0xff] & 0x0000ff00U) ^
+		(Te0[(s2 >> 16) & 0xff] & 0x00ff0000U) ^
+		(Te1[(s3 >> 24)       ] & 0xff000000U) ^
+		rk[0];
+	*(u32*)(out+4) =
+		(Te2[(s1      ) & 0xff] & 0x000000ffU) ^
+		(Te3[(s2 >>  8) & 0xff] & 0x0000ff00U) ^
+		(Te0[(s3 >> 16) & 0xff] & 0x00ff0000U) ^
+		(Te1[(s0 >> 24)       ] & 0xff000000U) ^
+		rk[1];
+	*(u32*)(out+8) =
+		(Te2[(s2      ) & 0xff] & 0x000000ffU) ^
+		(Te3[(s3 >>  8) & 0xff] & 0x0000ff00U) ^
+		(Te0[(s0 >> 16) & 0xff] & 0x00ff0000U) ^
+		(Te1[(s1 >> 24)       ] & 0xff000000U) ^
+		rk[2];
+	*(u32*)(out+12) =
+		(Te2[(s3      ) & 0xff] & 0x000000ffU) ^
+		(Te3[(s0 >>  8) & 0xff] & 0x0000ff00U) ^
+		(Te0[(s1 >> 16) & 0xff] & 0x00ff0000U) ^
+		(Te1[(s2 >> 24)       ] & 0xff000000U) ^
+		rk[3];
+#endif
+}
+
+/*
+ * Decrypt a single block
+ * in and out can overlap
+ */
+void AES_decrypt(const unsigned char *in, unsigned char *out,
+		 const AES_KEY *key) {
+
+	const u32 *rk;
+	u32 s0, s1, s2, s3, t[4];
+	int r;
+
+	assert(in && out && key);
+	rk = key->rd_key;
+
+	/*
+	 * map byte array block to cipher state
+	 * and add initial round key:
+	 */
+	s0 = GETU32(in     ) ^ rk[0];
+	s1 = GETU32(in +  4) ^ rk[1];
+	s2 = GETU32(in +  8) ^ rk[2];
+	s3 = GETU32(in + 12) ^ rk[3];
+
+#if defined(AES_COMPACT_IN_OUTER_ROUNDS)
+	prefetch256(Td4);
+
+        t[0] =	Td4[(s0      ) & 0xff]       ^
+		Td4[(s3 >>  8) & 0xff] <<  8 ^
+		Td4[(s2 >> 16) & 0xff] << 16 ^
+		Td4[(s1 >> 24)       ] << 24;
+        t[1] =	Td4[(s1      ) & 0xff]       ^
+		Td4[(s0 >>  8) & 0xff] <<  8 ^
+		Td4[(s3 >> 16) & 0xff] << 16 ^
+		Td4[(s2 >> 24)       ] << 24;
+        t[2] =	Td4[(s2      ) & 0xff]       ^
+		Td4[(s1 >>  8) & 0xff] <<  8 ^
+		Td4[(s0 >> 16) & 0xff] << 16 ^
+		Td4[(s3 >> 24)       ] << 24;
+        t[3] =	Td4[(s3      ) & 0xff]       ^
+		Td4[(s2 >>  8) & 0xff] <<  8 ^
+		Td4[(s1 >> 16) & 0xff] << 16 ^
+		Td4[(s0 >> 24)       ] << 24;
+
+	/* now do the linear transform using words */ 
+	{	int i;
+		u32 tp1, tp2, tp4, tp8, tp9, tpb, tpd, tpe, m;
+
+		for (i = 0; i < 4; i++) {
+			tp1 = t[i];
+			m = tp1 & 0x80808080;
+			tp2 = ((tp1 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp2 & 0x80808080;
+			tp4 = ((tp2 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp4 & 0x80808080;
+			tp8 = ((tp4 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			tp9 = tp8 ^ tp1;
+			tpb = tp9 ^ tp2;
+			tpd = tp9 ^ tp4;
+			tpe = tp8 ^ tp4 ^ tp2;
+#if defined(ROTATE)
+			t[i] = tpe ^ ROTATE(tpd,16) ^
+				ROTATE(tp9,8) ^ ROTATE(tpb,24);
+#else
+			t[i] = tpe ^ (tpd >> 16) ^ (tpd << 16) ^ 
+				(tp9 >> 24) ^ (tp9 << 8) ^
+				(tpb >> 8) ^ (tpb << 24);
+#endif
+			t[i] ^= rk[4+i];
+		}
+	}
+#else
+	t[0] =	Td0[(s0      ) & 0xff] ^
+		Td1[(s3 >>  8) & 0xff] ^
+		Td2[(s2 >> 16) & 0xff] ^
+		Td3[(s1 >> 24)       ] ^
+		rk[4];
+	t[1] =	Td0[(s1      ) & 0xff] ^
+		Td1[(s0 >>  8) & 0xff] ^
+		Td2[(s3 >> 16) & 0xff] ^
+		Td3[(s2 >> 24)       ] ^
+		rk[5];
+	t[2] =	Td0[(s2      ) & 0xff] ^
+		Td1[(s1 >>  8) & 0xff] ^
+		Td2[(s0 >> 16) & 0xff] ^
+		Td3[(s3 >> 24)       ] ^
+		rk[6];
+	t[3] =	Td0[(s3      ) & 0xff] ^
+		Td1[(s2 >>  8) & 0xff] ^
+		Td2[(s1 >> 16) & 0xff] ^
+		Td3[(s0 >> 24)       ] ^
+		rk[7];
+#endif
+	s0 = t[0]; s1 = t[1]; s2 = t[2]; s3 = t[3];
+
+    /*
+     * Nr - 2 full rounds:
+     */
+    for (rk+=8,r=key->rounds-2; r>0; rk+=4,r--) {
+#if defined(AES_COMPACT_IN_INNER_ROUNDS)
+        t[0] =	Td4[(s0      ) & 0xff]       ^
+		Td4[(s3 >>  8) & 0xff] <<  8 ^
+		Td4[(s2 >> 16) & 0xff] << 16 ^
+		Td4[(s1 >> 24)       ] << 24;
+        t[1] =	Td4[(s1      ) & 0xff]       ^
+		Td4[(s0 >>  8) & 0xff] <<  8 ^
+		Td4[(s3 >> 16) & 0xff] << 16 ^
+		Td4[(s2 >> 24)       ] << 24;
+        t[2] =	Td4[(s2      ) & 0xff]       ^
+		Td4[(s1 >>  8) & 0xff] <<  8 ^
+		Td4[(s0 >> 16) & 0xff] << 16 ^
+		Td4[(s3 >> 24)       ] << 24;
+        t[3] =	Td4[(s3      ) & 0xff]       ^
+		Td4[(s2 >>  8) & 0xff] <<  8 ^
+		Td4[(s1 >> 16) & 0xff] << 16 ^
+		Td4[(s0 >> 24)       ] << 24;
+
+	/* now do the linear transform using words */ 
+	{	int i;
+		u32 tp1, tp2, tp4, tp8, tp9, tpb, tpd, tpe, m;
+
+		for (i = 0; i < 4; i++) {
+			tp1 = t[i];
+			m = tp1 & 0x80808080;
+			tp2 = ((tp1 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp2 & 0x80808080;
+			tp4 = ((tp2 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp4 & 0x80808080;
+			tp8 = ((tp4 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			tp9 = tp8 ^ tp1;
+			tpb = tp9 ^ tp2;
+			tpd = tp9 ^ tp4;
+			tpe = tp8 ^ tp4 ^ tp2;
+#if defined(ROTATE)
+			t[i] = tpe ^ ROTATE(tpd,16) ^
+				ROTATE(tp9,8) ^ ROTATE(tpb,24);
+#else
+			t[i] = tpe ^ (tpd >> 16) ^ (tpd << 16) ^ 
+				(tp9 >> 24) ^ (tp9 << 8) ^
+				(tpb >> 8) ^ (tpb << 24);
+#endif
+			t[i] ^= rk[i];
+		}
+	}
+#else
+	t[0] =	Td0[(s0      ) & 0xff] ^
+		Td1[(s3 >>  8) & 0xff] ^
+		Td2[(s2 >> 16) & 0xff] ^
+		Td3[(s1 >> 24)       ] ^
+		rk[0];
+	t[1] =	Td0[(s1      ) & 0xff] ^
+		Td1[(s0 >>  8) & 0xff] ^
+		Td2[(s3 >> 16) & 0xff] ^
+		Td3[(s2 >> 24)       ] ^
+		rk[1];
+	t[2] =	Td0[(s2      ) & 0xff] ^
+		Td1[(s1 >>  8) & 0xff] ^
+		Td2[(s0 >> 16) & 0xff] ^
+		Td3[(s3 >> 24)       ] ^
+		rk[2];
+	t[3] =	Td0[(s3      ) & 0xff] ^
+		Td1[(s2 >>  8) & 0xff] ^
+		Td2[(s1 >> 16) & 0xff] ^
+		Td3[(s0 >> 24)       ] ^
+		rk[3];
+#endif
+	s0 = t[0]; s1 = t[1]; s2 = t[2]; s3 = t[3];
+    }
+    /*
+	 * apply last round and
+	 * map cipher state to byte array block:
+	 */
+	prefetch256(Td4);
+
+	*(u32*)(out+0) =
+		(Td4[(s0      ) & 0xff])	^
+		(Td4[(s3 >>  8) & 0xff] <<  8) ^
+		(Td4[(s2 >> 16) & 0xff] << 16) ^
+		(Td4[(s1 >> 24)       ] << 24) ^
+		rk[0];
+	*(u32*)(out+4) =
+		(Td4[(s1      ) & 0xff])	 ^
+		(Td4[(s0 >>  8) & 0xff] <<  8) ^
+		(Td4[(s3 >> 16) & 0xff] << 16) ^
+		(Td4[(s2 >> 24)       ] << 24) ^
+		rk[1];
+	*(u32*)(out+8) =
+		(Td4[(s2      ) & 0xff])	 ^
+		(Td4[(s1 >>  8) & 0xff] <<  8) ^
+		(Td4[(s0 >> 16) & 0xff] << 16) ^
+		(Td4[(s3 >> 24)       ] << 24) ^
+		rk[2];
+	*(u32*)(out+12) =
+		(Td4[(s3      ) & 0xff])	 ^
+		(Td4[(s2 >>  8) & 0xff] <<  8) ^
+		(Td4[(s1 >> 16) & 0xff] << 16) ^
+		(Td4[(s0 >> 24)       ] << 24) ^
+		rk[3];
+}
diff --git a/crypto/aes/asm/aes-586.pl b/crypto/aes/asm/aes-586.pl
index 3bc46a968e37..687ed811be47 100755
--- a/crypto/aes/asm/aes-586.pl
+++ b/crypto/aes/asm/aes-586.pl
@@ -2,11 +2,12 @@
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
-# project. Rights for redistribution and usage in source and binary
-# forms are granted according to the OpenSSL license.
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
 # ====================================================================
 #
-# Version 3.6.
+# Version 4.3.
 #
 # You might fail to appreciate this module performance from the first
 # try. If compared to "vanilla" linux-ia32-icc target, i.e. considered
@@ -38,7 +39,7 @@
 # but exhibits up to 10% improvement on other cores.
 #
 # Second version is "monolithic" replacement for aes_core.c, which in
-# addition to AES_[de|en]crypt implements AES_set_[de|en]cryption_key.
+# addition to AES_[de|en]crypt implements private_AES_set_[de|en]cryption_key.
 # This made it possible to implement little-endian variant of the
 # algorithm without modifying the base C code. Motivating factor for
 # the undertaken effort was that it appeared that in tight IA-32
@@ -81,11 +82,117 @@
 # AMD K8	20			19
 # PIII		25			23
 # Pentium	81			78
-
-push(@INC,"perlasm","../../perlasm");
+#
+# Version 3.7 reimplements outer rounds as "compact." Meaning that
+# first and last rounds reference compact 256 bytes S-box. This means
+# that first round consumes a lot more CPU cycles and that encrypt
+# and decrypt performance becomes asymmetric. Encrypt performance
+# drops by 10-12%, while decrypt - by 20-25%:-( 256 bytes S-box is
+# aggressively pre-fetched.
+#
+# Version 4.0 effectively rolls back to 3.6 and instead implements
+# additional set of functions, _[x86|sse]_AES_[en|de]crypt_compact,
+# which use exclusively 256 byte S-box. These functions are to be
+# called in modes not concealing plain text, such as ECB, or when
+# we're asked to process smaller amount of data [or unconditionally
+# on hyper-threading CPU]. Currently it's called unconditionally from
+# AES_[en|de]crypt, which affects all modes, but CBC. CBC routine
+# still needs to be modified to switch between slower and faster
+# mode when appropriate... But in either case benchmark landscape
+# changes dramatically and below numbers are CPU cycles per processed
+# byte for 128-bit key.
+#
+#		ECB encrypt	ECB decrypt	CBC large chunk
+# P4		56[60]		84[100]		23
+# AMD K8	48[44]		70[79]		18
+# PIII		41[50]		61[91]		24
+# Core 2	32[38]		45[70]		18.5
+# Pentium	120		160		77
+#
+# Version 4.1 switches to compact S-box even in key schedule setup.
+#
+# Version 4.2 prefetches compact S-box in every SSE round or in other
+# words every cache-line is *guaranteed* to be accessed within ~50
+# cycles window. Why just SSE? Because it's needed on hyper-threading
+# CPU! Which is also why it's prefetched with 64 byte stride. Best
+# part is that it has no negative effect on performance:-)  
+#
+# Version 4.3 implements switch between compact and non-compact block
+# functions in AES_cbc_encrypt depending on how much data was asked
+# to be processed in one stroke.
+#
+######################################################################
+# Timing attacks are classified in two classes: synchronous when
+# attacker consciously initiates cryptographic operation and collects
+# timing data of various character afterwards, and asynchronous when
+# malicious code is executed on same CPU simultaneously with AES,
+# instruments itself and performs statistical analysis of this data.
+#
+# As far as synchronous attacks go the root to the AES timing
+# vulnerability is twofold. Firstly, of 256 S-box elements at most 160
+# are referred to in single 128-bit block operation. Well, in C
+# implementation with 4 distinct tables it's actually as little as 40
+# references per 256 elements table, but anyway... Secondly, even
+# though S-box elements are clustered into smaller amount of cache-
+# lines, smaller than 160 and even 40, it turned out that for certain
+# plain-text pattern[s] or simply put chosen plain-text and given key
+# few cache-lines remain unaccessed during block operation. Now, if
+# attacker can figure out this access pattern, he can deduct the key
+# [or at least part of it]. The natural way to mitigate this kind of
+# attacks is to minimize the amount of cache-lines in S-box and/or
+# prefetch them to ensure that every one is accessed for more uniform
+# timing. But note that *if* plain-text was concealed in such way that
+# input to block function is distributed *uniformly*, then attack
+# wouldn't apply. Now note that some encryption modes, most notably
+# CBC, do mask the plain-text in this exact way [secure cipher output
+# is distributed uniformly]. Yes, one still might find input that
+# would reveal the information about given key, but if amount of
+# candidate inputs to be tried is larger than amount of possible key
+# combinations then attack becomes infeasible. This is why revised
+# AES_cbc_encrypt "dares" to switch to larger S-box when larger chunk
+# of data is to be processed in one stroke. The current size limit of
+# 512 bytes is chosen to provide same [diminishigly low] probability
+# for cache-line to remain untouched in large chunk operation with
+# large S-box as for single block operation with compact S-box and
+# surely needs more careful consideration...
+#
+# As for asynchronous attacks. There are two flavours: attacker code
+# being interleaved with AES on hyper-threading CPU at *instruction*
+# level, and two processes time sharing single core. As for latter.
+# Two vectors. 1. Given that attacker process has higher priority,
+# yield execution to process performing AES just before timer fires
+# off the scheduler, immediately regain control of CPU and analyze the
+# cache state. For this attack to be efficient attacker would have to
+# effectively slow down the operation by several *orders* of magnitute,
+# by ratio of time slice to duration of handful of AES rounds, which
+# unlikely to remain unnoticed. Not to mention that this also means
+# that he would spend correspondigly more time to collect enough
+# statistical data to mount the attack. It's probably appropriate to
+# say that if adeversary reckons that this attack is beneficial and
+# risks to be noticed, you probably have larger problems having him
+# mere opportunity. In other words suggested code design expects you
+# to preclude/mitigate this attack by overall system security design.
+# 2. Attacker manages to make his code interrupt driven. In order for
+# this kind of attack to be feasible, interrupt rate has to be high
+# enough, again comparable to duration of handful of AES rounds. But
+# is there interrupt source of such rate? Hardly, not even 1Gbps NIC
+# generates interrupts at such raging rate...
+#
+# And now back to the former, hyper-threading CPU or more specifically
+# Intel P4. Recall that asynchronous attack implies that malicious
+# code instruments itself. And naturally instrumentation granularity
+# has be noticeably lower than duration of codepath accessing S-box.
+# Given that all cache-lines are accessed during that time that is.
+# Current implementation accesses *all* cache-lines within ~50 cycles
+# window, which is actually *less* than RDTSC latency on Intel P4!
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 
-&asm_init($ARGV[0],"aes-586.pl",$ARGV[$#ARGV] eq "386");
+&asm_init($ARGV[0],"aes-586.pl",$x86only = $ARGV[$#ARGV] eq "386");
+&static_label("AES_Te");
+&static_label("AES_Td");
 
 $s0="eax";
 $s1="ebx";
@@ -93,21 +200,36 @@ $s2="ecx";
 $s3="edx";
 $key="edi";
 $acc="esi";
+$tbl="ebp";
+
+# stack frame layout in _[x86|sse]_AES_* routines, frame is allocated
+# by caller
+$__ra=&DWP(0,"esp");	# return address
+$__s0=&DWP(4,"esp");	# s0 backing store
+$__s1=&DWP(8,"esp");	# s1 backing store
+$__s2=&DWP(12,"esp");	# s2 backing store
+$__s3=&DWP(16,"esp");	# s3 backing store
+$__key=&DWP(20,"esp");	# pointer to key schedule
+$__end=&DWP(24,"esp");	# pointer to end of key schedule
+$__tbl=&DWP(28,"esp");	# %ebp backing store
+
+# stack frame layout in AES_[en|crypt] routines, which differs from
+# above by 4 and overlaps by %ebp backing store
+$_tbl=&DWP(24,"esp");
+$_esp=&DWP(28,"esp");
 
-$compromise=0;		# $compromise=128 abstains from copying key
-			# schedule to stack when encrypting inputs
-			# shorter than 128 bytes at the cost of
-			# risksing aliasing with S-boxes. In return
-			# you get way better, up to +70%, small block
-			# performance.
+sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
+
+$speed_limit=512;	# chunks smaller than $speed_limit are
+			# processed with compact routine in CBC mode
 $small_footprint=1;	# $small_footprint=1 code is ~5% slower [on
 			# recent µ-archs], but ~5 times smaller!
 			# I favor compact code to minimize cache
 			# contention and in hope to "collect" 5% back
 			# in real-life applications...
+
 $vertical_spin=0;	# shift "verticaly" defaults to 0, because of
 			# its proof-of-concept status...
-
 # Note that there is no decvert(), as well as last encryption round is
 # performed with "horizontal" shifts. This is because this "vertical"
 # implementation [one which groups shifts on a given $s[i] to form a
@@ -170,17 +292,484 @@ sub encvert()
 	&movz	($v0,&HB($v1));
 	&and	($v1,0xFF);
 	&xor	($s[3],&DWP(2,$te,$v1,8));		# s1>>16
-	 &mov	($key,&DWP(12,"esp"));			# reincarnate v1 as key
+	 &mov	($key,$__key);				# reincarnate v1 as key
 	&xor	($s[2],&DWP(1,$te,$v0,8));		# s1>>24
 }
 
+# Another experimental routine, which features "horizontal spin," but
+# eliminates one reference to stack. Strangely enough runs slower...
+sub enchoriz()
+{ my $v0 = $key, $v1 = $acc;
+
+	&movz	($v0,&LB($s0));			#  3, 2, 1, 0*
+	&rotr	($s2,8);			#  8,11,10, 9
+	&mov	($v1,&DWP(0,$te,$v0,8));	#  0
+	&movz	($v0,&HB($s1));			#  7, 6, 5*, 4
+	&rotr	($s3,16);			# 13,12,15,14
+	&xor	($v1,&DWP(3,$te,$v0,8));	#  5
+	&movz	($v0,&HB($s2));			#  8,11,10*, 9
+	&rotr	($s0,16);			#  1, 0, 3, 2
+	&xor	($v1,&DWP(2,$te,$v0,8));	# 10
+	&movz	($v0,&HB($s3));			# 13,12,15*,14
+	&xor	($v1,&DWP(1,$te,$v0,8));	# 15, t[0] collected
+	&mov	($__s0,$v1);			# t[0] saved
+
+	&movz	($v0,&LB($s1));			#  7, 6, 5, 4*
+	&shr	($s1,16);			#  -, -, 7, 6
+	&mov	($v1,&DWP(0,$te,$v0,8));	#  4
+	&movz	($v0,&LB($s3));			# 13,12,15,14*
+	&xor	($v1,&DWP(2,$te,$v0,8));	# 14
+	&movz	($v0,&HB($s0));			#  1, 0, 3*, 2
+	&and	($s3,0xffff0000);		# 13,12, -, -
+	&xor	($v1,&DWP(1,$te,$v0,8));	#  3
+	&movz	($v0,&LB($s2));			#  8,11,10, 9*
+	&or	($s3,$s1);			# 13,12, 7, 6
+	&xor	($v1,&DWP(3,$te,$v0,8));	#  9, t[1] collected
+	&mov	($s1,$v1);			#  s[1]=t[1]
+
+	&movz	($v0,&LB($s0));			#  1, 0, 3, 2*
+	&shr	($s2,16);			#  -, -, 8,11
+	&mov	($v1,&DWP(2,$te,$v0,8));	#  2
+	&movz	($v0,&HB($s3));			# 13,12, 7*, 6
+	&xor	($v1,&DWP(1,$te,$v0,8));	#  7
+	&movz	($v0,&HB($s2));			#  -, -, 8*,11
+	&xor	($v1,&DWP(0,$te,$v0,8));	#  8
+	&mov	($v0,$s3);
+	&shr	($v0,24);			# 13
+	&xor	($v1,&DWP(3,$te,$v0,8));	# 13, t[2] collected
+
+	&movz	($v0,&LB($s2));			#  -, -, 8,11*
+	&shr	($s0,24);			#  1*
+	&mov	($s2,&DWP(1,$te,$v0,8));	# 11
+	&xor	($s2,&DWP(3,$te,$s0,8));	#  1
+	&mov	($s0,$__s0);			# s[0]=t[0]
+	&movz	($v0,&LB($s3));			# 13,12, 7, 6*
+	&shr	($s3,16);			#   ,  ,13,12
+	&xor	($s2,&DWP(2,$te,$v0,8));	#  6
+	&mov	($key,$__key);			# reincarnate v0 as key
+	&and	($s3,0xff);			#   ,  ,13,12*
+	&mov	($s3,&DWP(0,$te,$s3,8));	# 12
+	&xor	($s3,$s2);			# s[2]=t[3] collected
+	&mov	($s2,$v1);			# s[2]=t[2]
+}
+
+# More experimental code... SSE one... Even though this one eliminates
+# *all* references to stack, it's not faster...
+sub sse_encbody()
+{
+	&movz	($acc,&LB("eax"));		#  0
+	&mov	("ecx",&DWP(0,$tbl,$acc,8));	#  0
+	&pshufw	("mm2","mm0",0x0d);		#  7, 6, 3, 2
+	&movz	("edx",&HB("eax"));		#  1
+	&mov	("edx",&DWP(3,$tbl,"edx",8));	#  1
+	&shr	("eax",16);			#  5, 4
+
+	&movz	($acc,&LB("ebx"));		# 10
+	&xor	("ecx",&DWP(2,$tbl,$acc,8));	# 10
+	&pshufw	("mm6","mm4",0x08);		# 13,12, 9, 8
+	&movz	($acc,&HB("ebx"));		# 11
+	&xor	("edx",&DWP(1,$tbl,$acc,8));	# 11
+	&shr	("ebx",16);			# 15,14
+
+	&movz	($acc,&HB("eax"));		#  5
+	&xor	("ecx",&DWP(3,$tbl,$acc,8));	#  5
+	&movq	("mm3",QWP(16,$key));
+	&movz	($acc,&HB("ebx"));		# 15
+	&xor	("ecx",&DWP(1,$tbl,$acc,8));	# 15
+	&movd	("mm0","ecx");			# t[0] collected
+
+	&movz	($acc,&LB("eax"));		#  4
+	&mov	("ecx",&DWP(0,$tbl,$acc,8));	#  4
+	&movd	("eax","mm2");			#  7, 6, 3, 2
+	&movz	($acc,&LB("ebx"));		# 14
+	&xor	("ecx",&DWP(2,$tbl,$acc,8));	# 14
+	&movd	("ebx","mm6");			# 13,12, 9, 8
+
+	&movz	($acc,&HB("eax"));		#  3
+	&xor	("ecx",&DWP(1,$tbl,$acc,8));	#  3
+	&movz	($acc,&HB("ebx"));		#  9
+	&xor	("ecx",&DWP(3,$tbl,$acc,8));	#  9
+	&movd	("mm1","ecx");			# t[1] collected
+
+	&movz	($acc,&LB("eax"));		#  2
+	&mov	("ecx",&DWP(2,$tbl,$acc,8));	#  2
+	&shr	("eax",16);			#  7, 6
+	&punpckldq	("mm0","mm1");		# t[0,1] collected
+	&movz	($acc,&LB("ebx"));		#  8
+	&xor	("ecx",&DWP(0,$tbl,$acc,8));	#  8
+	&shr	("ebx",16);			# 13,12
+
+	&movz	($acc,&HB("eax"));		#  7
+	&xor	("ecx",&DWP(1,$tbl,$acc,8));	#  7
+	&pxor	("mm0","mm3");
+	&movz	("eax",&LB("eax"));		#  6
+	&xor	("edx",&DWP(2,$tbl,"eax",8));	#  6
+	&pshufw	("mm1","mm0",0x08);		#  5, 4, 1, 0
+	&movz	($acc,&HB("ebx"));		# 13
+	&xor	("ecx",&DWP(3,$tbl,$acc,8));	# 13
+	&xor	("ecx",&DWP(24,$key));		# t[2]
+	&movd	("mm4","ecx");			# t[2] collected
+	&movz	("ebx",&LB("ebx"));		# 12
+	&xor	("edx",&DWP(0,$tbl,"ebx",8));	# 12
+	&shr	("ecx",16);
+	&movd	("eax","mm1");			#  5, 4, 1, 0
+	&mov	("ebx",&DWP(28,$key));		# t[3]
+	&xor	("ebx","edx");
+	&movd	("mm5","ebx");			# t[3] collected
+	&and	("ebx",0xffff0000);
+	&or	("ebx","ecx");
+
+	&punpckldq	("mm4","mm5");		# t[2,3] collected
+}
+
+######################################################################
+# "Compact" block function
+######################################################################
+
+sub enccompact()
+{ my $Fn = mov;
+  while ($#_>5) { pop(@_); $Fn=sub{}; }
+  my ($i,$te,@s)=@_;
+  my $tmp = $key;
+  my $out = $i==3?$s[0]:$acc;
+
+	# $Fn is used in first compact round and its purpose is to
+	# void restoration of some values from stack, so that after
+	# 4xenccompact with extra argument $key value is left there...
+	if ($i==3)  {	&$Fn	($key,$__key);			}##%edx
+	else        {	&mov	($out,$s[0]);			}
+			&and	($out,0xFF);
+	if ($i==1)  {	&shr	($s[0],16);			}#%ebx[1]
+	if ($i==2)  {	&shr	($s[0],24);			}#%ecx[2]
+			&movz	($out,&BP(-128,$te,$out,1));
+
+	if ($i==3)  {	$tmp=$s[1];				}##%eax
+			&movz	($tmp,&HB($s[1]));
+			&movz	($tmp,&BP(-128,$te,$tmp,1));
+			&shl	($tmp,8);
+			&xor	($out,$tmp);
+
+	if ($i==3)  {	$tmp=$s[2]; &mov ($s[1],$__s0);		}##%ebx
+	else        {	&mov	($tmp,$s[2]);
+			&shr	($tmp,16);			}
+	if ($i==2)  {	&and	($s[1],0xFF);			}#%edx[2]
+			&and	($tmp,0xFF);
+			&movz	($tmp,&BP(-128,$te,$tmp,1));
+			&shl	($tmp,16);
+			&xor	($out,$tmp);
+
+	if ($i==3)  {	$tmp=$s[3]; &mov ($s[2],$__s1);		}##%ecx
+	elsif($i==2){	&movz	($tmp,&HB($s[3]));		}#%ebx[2]
+	else        {	&mov	($tmp,$s[3]);
+			&shr	($tmp,24);			}
+			&movz	($tmp,&BP(-128,$te,$tmp,1));
+			&shl	($tmp,24);
+			&xor	($out,$tmp);
+	if ($i<2)   {	&mov	(&DWP(4+4*$i,"esp"),$out);	}
+	if ($i==3)  {	&mov	($s[3],$acc);			}
+	&comment();
+}
+
+sub enctransform()
+{ my @s = ($s0,$s1,$s2,$s3);
+  my $i = shift;
+  my $tmp = $tbl;
+  my $r2  = $key ;
+
+	&mov	($acc,$s[$i]);
+	&and	($acc,0x80808080);
+	&mov	($tmp,$acc);
+	&shr	($tmp,7);
+	&lea	($r2,&DWP(0,$s[$i],$s[$i]));
+	&sub	($acc,$tmp);
+	&and	($r2,0xfefefefe);
+	&and	($acc,0x1b1b1b1b);
+	&mov	($tmp,$s[$i]);
+	&xor	($acc,$r2);	# r2
+
+	&xor	($s[$i],$acc);	# r0 ^ r2
+	&rotl	($s[$i],24);
+	&xor	($s[$i],$acc)	# ROTATE(r2^r0,24) ^ r2
+	&rotr	($tmp,16);
+	&xor	($s[$i],$tmp);
+	&rotr	($tmp,8);
+	&xor	($s[$i],$tmp);
+}
+
+&function_begin_B("_x86_AES_encrypt_compact");
+	# note that caller is expected to allocate stack frame for me!
+	&mov	($__key,$key);			# save key
+
+	&xor	($s0,&DWP(0,$key));		# xor with key
+	&xor	($s1,&DWP(4,$key));
+	&xor	($s2,&DWP(8,$key));
+	&xor	($s3,&DWP(12,$key));
+
+	&mov	($acc,&DWP(240,$key));		# load key->rounds
+	&lea	($acc,&DWP(-2,$acc,$acc));
+	&lea	($acc,&DWP(0,$key,$acc,8));
+	&mov	($__end,$acc);			# end of key schedule
+
+	# prefetch Te4
+	&mov	($key,&DWP(0-128,$tbl));
+	&mov	($acc,&DWP(32-128,$tbl));
+	&mov	($key,&DWP(64-128,$tbl));
+	&mov	($acc,&DWP(96-128,$tbl));
+	&mov	($key,&DWP(128-128,$tbl));
+	&mov	($acc,&DWP(160-128,$tbl));
+	&mov	($key,&DWP(192-128,$tbl));
+	&mov	($acc,&DWP(224-128,$tbl));
+
+	&set_label("loop",16);
+
+		&enccompact(0,$tbl,$s0,$s1,$s2,$s3,1);
+		&enccompact(1,$tbl,$s1,$s2,$s3,$s0,1);
+		&enccompact(2,$tbl,$s2,$s3,$s0,$s1,1);
+		&enccompact(3,$tbl,$s3,$s0,$s1,$s2,1);
+		&enctransform(2);
+		&enctransform(3);
+		&enctransform(0);
+		&enctransform(1);
+		&mov 	($key,$__key);
+		&mov	($tbl,$__tbl);
+		&add	($key,16);		# advance rd_key
+		&xor	($s0,&DWP(0,$key));
+		&xor	($s1,&DWP(4,$key));
+		&xor	($s2,&DWP(8,$key));
+		&xor	($s3,&DWP(12,$key));
+
+	&cmp	($key,$__end);
+	&mov	($__key,$key);
+	&jb	(&label("loop"));
+
+	&enccompact(0,$tbl,$s0,$s1,$s2,$s3);
+	&enccompact(1,$tbl,$s1,$s2,$s3,$s0);
+	&enccompact(2,$tbl,$s2,$s3,$s0,$s1);
+	&enccompact(3,$tbl,$s3,$s0,$s1,$s2);
+
+	&xor	($s0,&DWP(16,$key));
+	&xor	($s1,&DWP(20,$key));
+	&xor	($s2,&DWP(24,$key));
+	&xor	($s3,&DWP(28,$key));
+
+	&ret	();
+&function_end_B("_x86_AES_encrypt_compact");
+
+######################################################################
+# "Compact" SSE block function.
+######################################################################
+#
+# Performance is not actually extraordinary in comparison to pure
+# x86 code. In particular encrypt performance is virtually the same.
+# Decrypt performance on the other hand is 15-20% better on newer
+# µ-archs [but we're thankful for *any* improvement here], and ~50%
+# better on PIII:-) And additionally on the pros side this code
+# eliminates redundant references to stack and thus relieves/
+# minimizes the pressure on the memory bus.
+#
+# MMX register layout                           lsb
+# +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
+# |          mm4          |          mm0          |
+# +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
+# |     s3    |     s2    |     s1    |     s0    |    
+# +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
+# |15|14|13|12|11|10| 9| 8| 7| 6| 5| 4| 3| 2| 1| 0|
+# +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
+#
+# Indexes translate as s[N/4]>>(8*(N%4)), e.g. 5 means s1>>8.
+# In this terms encryption and decryption "compact" permutation
+# matrices can be depicted as following:
+#
+# encryption              lsb	# decryption              lsb
+# +----++----+----+----+----+	# +----++----+----+----+----+
+# | t0 || 15 | 10 |  5 |  0 |	# | t0 ||  7 | 10 | 13 |  0 |
+# +----++----+----+----+----+	# +----++----+----+----+----+
+# | t1 ||  3 | 14 |  9 |  4 |	# | t1 || 11 | 14 |  1 |  4 |
+# +----++----+----+----+----+	# +----++----+----+----+----+
+# | t2 ||  7 |  2 | 13 |  8 |	# | t2 || 15 |  2 |  5 |  8 |
+# +----++----+----+----+----+	# +----++----+----+----+----+
+# | t3 || 11 |  6 |  1 | 12 |	# | t3 ||  3 |  6 |  9 | 12 |
+# +----++----+----+----+----+	# +----++----+----+----+----+
+#
+######################################################################
+# Why not xmm registers? Short answer. It was actually tested and
+# was not any faster, but *contrary*, most notably on Intel CPUs.
+# Longer answer. Main advantage of using mm registers is that movd
+# latency is lower, especially on Intel P4. While arithmetic
+# instructions are twice as many, they can be scheduled every cycle
+# and not every second one when they are operating on xmm register,
+# so that "arithmetic throughput" remains virtually the same. And
+# finally the code can be executed even on elder SSE-only CPUs:-)
+
+sub sse_enccompact()
+{
+	&pshufw	("mm1","mm0",0x08);		#  5, 4, 1, 0
+	&pshufw	("mm5","mm4",0x0d);		# 15,14,11,10
+	&movd	("eax","mm1");			#  5, 4, 1, 0
+	&movd	("ebx","mm5");			# 15,14,11,10
+
+	&movz	($acc,&LB("eax"));		#  0
+	&movz	("ecx",&BP(-128,$tbl,$acc,1));	#  0
+	&pshufw	("mm2","mm0",0x0d);		#  7, 6, 3, 2
+	&movz	("edx",&HB("eax"));		#  1
+	&movz	("edx",&BP(-128,$tbl,"edx",1));	#  1
+	&shl	("edx",8);			#  1
+	&shr	("eax",16);			#  5, 4
+
+	&movz	($acc,&LB("ebx"));		# 10
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 10
+	&shl	($acc,16);			# 10
+	&or	("ecx",$acc);			# 10
+	&pshufw	("mm6","mm4",0x08);		# 13,12, 9, 8
+	&movz	($acc,&HB("ebx"));		# 11
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 11
+	&shl	($acc,24);			# 11
+	&or	("edx",$acc);			# 11
+	&shr	("ebx",16);			# 15,14
+
+	&movz	($acc,&HB("eax"));		#  5
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	#  5
+	&shl	($acc,8);			#  5
+	&or	("ecx",$acc);			#  5
+	&movz	($acc,&HB("ebx"));		# 15
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 15
+	&shl	($acc,24);			# 15
+	&or	("ecx",$acc);			# 15
+	&movd	("mm0","ecx");			# t[0] collected
+
+	&movz	($acc,&LB("eax"));		#  4
+	&movz	("ecx",&BP(-128,$tbl,$acc,1));	#  4
+	&movd	("eax","mm2");			#  7, 6, 3, 2
+	&movz	($acc,&LB("ebx"));		# 14
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 14
+	&shl	($acc,16);			# 14
+	&or	("ecx",$acc);			# 14
+
+	&movd	("ebx","mm6");			# 13,12, 9, 8
+	&movz	($acc,&HB("eax"));		#  3
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	#  3
+	&shl	($acc,24);			#  3
+	&or	("ecx",$acc);			#  3
+	&movz	($acc,&HB("ebx"));		#  9
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	#  9
+	&shl	($acc,8);			#  9
+	&or	("ecx",$acc);			#  9
+	&movd	("mm1","ecx");			# t[1] collected
+
+	&movz	($acc,&LB("ebx"));		#  8
+	&movz	("ecx",&BP(-128,$tbl,$acc,1));	#  8
+	&shr	("ebx",16);			# 13,12
+	&movz	($acc,&LB("eax"));		#  2
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	#  2
+	&shl	($acc,16);			#  2
+	&or	("ecx",$acc);			#  2
+	&shr	("eax",16);			#  7, 6
+
+	&punpckldq	("mm0","mm1");		# t[0,1] collected
+
+	&movz	($acc,&HB("eax"));		#  7
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	#  7
+	&shl	($acc,24);			#  7
+	&or	("ecx",$acc);			#  7
+	&and	("eax",0xff);			#  6
+	&movz	("eax",&BP(-128,$tbl,"eax",1));	#  6
+	&shl	("eax",16);			#  6
+	&or	("edx","eax");			#  6
+	&movz	($acc,&HB("ebx"));		# 13
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 13
+	&shl	($acc,8);			# 13
+	&or	("ecx",$acc);			# 13
+	&movd	("mm4","ecx");			# t[2] collected
+	&and	("ebx",0xff);			# 12
+	&movz	("ebx",&BP(-128,$tbl,"ebx",1));	# 12
+	&or	("edx","ebx");			# 12
+	&movd	("mm5","edx");			# t[3] collected
+
+	&punpckldq	("mm4","mm5");		# t[2,3] collected
+}
+
+					if (!$x86only) {
+&function_begin_B("_sse_AES_encrypt_compact");
+	&pxor	("mm0",&QWP(0,$key));	#  7, 6, 5, 4, 3, 2, 1, 0
+	&pxor	("mm4",&QWP(8,$key));	# 15,14,13,12,11,10, 9, 8
+
+	# note that caller is expected to allocate stack frame for me!
+	&mov	($acc,&DWP(240,$key));		# load key->rounds
+	&lea	($acc,&DWP(-2,$acc,$acc));
+	&lea	($acc,&DWP(0,$key,$acc,8));
+	&mov	($__end,$acc);			# end of key schedule
+
+	&mov	($s0,0x1b1b1b1b);		# magic constant
+	&mov	(&DWP(8,"esp"),$s0);
+	&mov	(&DWP(12,"esp"),$s0);
+
+	# prefetch Te4
+	&mov	($s0,&DWP(0-128,$tbl));
+	&mov	($s1,&DWP(32-128,$tbl));
+	&mov	($s2,&DWP(64-128,$tbl));
+	&mov	($s3,&DWP(96-128,$tbl));
+	&mov	($s0,&DWP(128-128,$tbl));
+	&mov	($s1,&DWP(160-128,$tbl));
+	&mov	($s2,&DWP(192-128,$tbl));
+	&mov	($s3,&DWP(224-128,$tbl));
+
+	&set_label("loop",16);
+		&sse_enccompact();
+		&add	($key,16);
+		&cmp	($key,$__end);
+		&ja	(&label("out"));
+
+		&movq	("mm2",&QWP(8,"esp"));
+		&pxor	("mm3","mm3");		&pxor	("mm7","mm7");
+		&movq	("mm1","mm0");		&movq	("mm5","mm4");	# r0
+		&pcmpgtb("mm3","mm0");		&pcmpgtb("mm7","mm4");
+		&pand	("mm3","mm2");		&pand	("mm7","mm2");
+		&pshufw	("mm2","mm0",0xb1);	&pshufw	("mm6","mm4",0xb1);# ROTATE(r0,16)
+		&paddb	("mm0","mm0");		&paddb	("mm4","mm4");
+		&pxor	("mm0","mm3");		&pxor	("mm4","mm7");	# = r2
+		&pshufw	("mm3","mm2",0xb1);	&pshufw	("mm7","mm6",0xb1);# r0
+		&pxor	("mm1","mm0");		&pxor	("mm5","mm4");	# r0^r2
+		&pxor	("mm0","mm2");		&pxor	("mm4","mm6");	# ^= ROTATE(r0,16)
+
+		&movq	("mm2","mm3");		&movq	("mm6","mm7");
+		&pslld	("mm3",8);		&pslld	("mm7",8);
+		&psrld	("mm2",24);		&psrld	("mm6",24);
+		&pxor	("mm0","mm3");		&pxor	("mm4","mm7");	# ^= r0<<8
+		&pxor	("mm0","mm2");		&pxor	("mm4","mm6");	# ^= r0>>24
+
+		&movq	("mm3","mm1");		&movq	("mm7","mm5");
+		&movq	("mm2",&QWP(0,$key));	&movq	("mm6",&QWP(8,$key));
+		&psrld	("mm1",8);		&psrld	("mm5",8);
+		&mov	($s0,&DWP(0-128,$tbl));
+		&pslld	("mm3",24);		&pslld	("mm7",24);
+		&mov	($s1,&DWP(64-128,$tbl));
+		&pxor	("mm0","mm1");		&pxor	("mm4","mm5");	# ^= (r2^r0)<<8
+		&mov	($s2,&DWP(128-128,$tbl));
+		&pxor	("mm0","mm3");		&pxor	("mm4","mm7");	# ^= (r2^r0)>>24
+		&mov	($s3,&DWP(192-128,$tbl));
+
+		&pxor	("mm0","mm2");		&pxor	("mm4","mm6");
+	&jmp	(&label("loop"));
+
+	&set_label("out",16);
+	&pxor	("mm0",&QWP(0,$key));
+	&pxor	("mm4",&QWP(8,$key));
+
+	&ret	();
+&function_end_B("_sse_AES_encrypt_compact");
+					}
+
+######################################################################
+# Vanilla block function.
+######################################################################
+
 sub encstep()
 { my ($i,$te,@s) = @_;
   my $tmp = $key;
   my $out = $i==3?$s[0]:$acc;
 
 	# lines marked with #%e?x[i] denote "reordered" instructions...
-	if ($i==3)  {	&mov	($key,&DWP(12,"esp"));		}##%edx
+	if ($i==3)  {	&mov	($key,$__key);			}##%edx
 	else        {	&mov	($out,$s[0]);
 			&and	($out,0xFF);			}
 	if ($i==1)  {	&shr	($s[0],16);			}#%ebx[1]
@@ -191,14 +780,14 @@ sub encstep()
 			&movz	($tmp,&HB($s[1]));
 			&xor	($out,&DWP(3,$te,$tmp,8));
 
-	if ($i==3)  {	$tmp=$s[2]; &mov ($s[1],&DWP(4,"esp"));	}##%ebx
+	if ($i==3)  {	$tmp=$s[2]; &mov ($s[1],$__s0);		}##%ebx
 	else        {	&mov	($tmp,$s[2]);
 			&shr	($tmp,16);			}
 	if ($i==2)  {	&and	($s[1],0xFF);			}#%edx[2]
 			&and	($tmp,0xFF);
 			&xor	($out,&DWP(2,$te,$tmp,8));
 
-	if ($i==3)  {	$tmp=$s[3]; &mov ($s[2],&DWP(8,"esp"));	}##%ecx
+	if ($i==3)  {	$tmp=$s[3]; &mov ($s[2],$__s1);		}##%ecx
 	elsif($i==2){	&movz	($tmp,&HB($s[3]));		}#%ebx[2]
 	else        {	&mov	($tmp,$s[3]); 
 			&shr	($tmp,24)			}
@@ -213,7 +802,7 @@ sub enclast()
   my $tmp = $key;
   my $out = $i==3?$s[0]:$acc;
 
-	if ($i==3)  {	&mov	($key,&DWP(12,"esp"));		}##%edx
+	if ($i==3)  {	&mov	($key,$__key);			}##%edx
 	else        {	&mov	($out,$s[0]);			}
 			&and	($out,0xFF);
 	if ($i==1)  {	&shr	($s[0],16);			}#%ebx[1]
@@ -227,8 +816,8 @@ sub enclast()
 			&and	($tmp,0x0000ff00);
 			&xor	($out,$tmp);
 
-	if ($i==3)  {	$tmp=$s[2]; &mov ($s[1],&DWP(4,"esp"));	}##%ebx
-	else        {	mov	($tmp,$s[2]);
+	if ($i==3)  {	$tmp=$s[2]; &mov ($s[1],$__s0);		}##%ebx
+	else        {	&mov	($tmp,$s[2]);
 			&shr	($tmp,16);			}
 	if ($i==2)  {	&and	($s[1],0xFF);			}#%edx[2]
 			&and	($tmp,0xFF);
@@ -236,7 +825,7 @@ sub enclast()
 			&and	($tmp,0x00ff0000);
 			&xor	($out,$tmp);
 
-	if ($i==3)  {	$tmp=$s[3]; &mov ($s[2],&DWP(8,"esp"));	}##%ecx
+	if ($i==3)  {	$tmp=$s[3]; &mov ($s[2],$__s1);		}##%ecx
 	elsif($i==2){	&movz	($tmp,&HB($s[3]));		}#%ebx[2]
 	else        {	&mov	($tmp,$s[3]);
 			&shr	($tmp,24);			}
@@ -247,9 +836,6 @@ sub enclast()
 	if ($i==3)  {	&mov	($s[3],$acc);			}
 }
 
-sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
-
-&public_label("AES_Te");
 &function_begin_B("_x86_AES_encrypt");
 	if ($vertical_spin) {
 		# I need high parts of volatile registers to be accessible...
@@ -258,7 +844,7 @@ sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
 	}
 
 	# note that caller is expected to allocate stack frame for me!
-	&mov	(&DWP(12,"esp"),$key);		# save key
+	&mov	($__key,$key);			# save key
 
 	&xor	($s0,&DWP(0,$key));		# xor with key
 	&xor	($s1,&DWP(4,$key));
@@ -270,24 +856,24 @@ sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
 	if ($small_footprint) {
 	    &lea	($acc,&DWP(-2,$acc,$acc));
 	    &lea	($acc,&DWP(0,$key,$acc,8));
-	    &mov	(&DWP(16,"esp"),$acc);	# end of key schedule
-	    &align	(4);
-	    &set_label("loop");
+	    &mov	($__end,$acc);		# end of key schedule
+
+	    &set_label("loop",16);
 		if ($vertical_spin) {
-		    &encvert("ebp",$s0,$s1,$s2,$s3);
+		    &encvert($tbl,$s0,$s1,$s2,$s3);
 		} else {
-		    &encstep(0,"ebp",$s0,$s1,$s2,$s3);
-		    &encstep(1,"ebp",$s1,$s2,$s3,$s0);
-		    &encstep(2,"ebp",$s2,$s3,$s0,$s1);
-		    &encstep(3,"ebp",$s3,$s0,$s1,$s2);
+		    &encstep(0,$tbl,$s0,$s1,$s2,$s3);
+		    &encstep(1,$tbl,$s1,$s2,$s3,$s0);
+		    &encstep(2,$tbl,$s2,$s3,$s0,$s1);
+		    &encstep(3,$tbl,$s3,$s0,$s1,$s2);
 		}
 		&add	($key,16);		# advance rd_key
 		&xor	($s0,&DWP(0,$key));
 		&xor	($s1,&DWP(4,$key));
 		&xor	($s2,&DWP(8,$key));
 		&xor	($s3,&DWP(12,$key));
-	    &cmp	($key,&DWP(16,"esp"));
-	    &mov	(&DWP(12,"esp"),$key);
+	    &cmp	($key,$__end);
+	    &mov	($__key,$key);
 	    &jb		(&label("loop"));
 	}
 	else {
@@ -296,15 +882,15 @@ sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
 	    &cmp	($acc,12);
 	    &jle	(&label("12rounds"));
 
-	&set_label("14rounds");
+	&set_label("14rounds",4);
 	    for ($i=1;$i<3;$i++) {
 		if ($vertical_spin) {
-		    &encvert("ebp",$s0,$s1,$s2,$s3);
+		    &encvert($tbl,$s0,$s1,$s2,$s3);
 		} else {
-		    &encstep(0,"ebp",$s0,$s1,$s2,$s3);
-		    &encstep(1,"ebp",$s1,$s2,$s3,$s0);
-		    &encstep(2,"ebp",$s2,$s3,$s0,$s1);
-		    &encstep(3,"ebp",$s3,$s0,$s1,$s2);
+		    &encstep(0,$tbl,$s0,$s1,$s2,$s3);
+		    &encstep(1,$tbl,$s1,$s2,$s3,$s0);
+		    &encstep(2,$tbl,$s2,$s3,$s0,$s1);
+		    &encstep(3,$tbl,$s3,$s0,$s1,$s2);
 		}
 		&xor	($s0,&DWP(16*$i+0,$key));
 		&xor	($s1,&DWP(16*$i+4,$key));
@@ -312,16 +898,16 @@ sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
 		&xor	($s3,&DWP(16*$i+12,$key));
 	    }
 	    &add	($key,32);
-	    &mov	(&DWP(12,"esp"),$key);	# advance rd_key
-	&set_label("12rounds");
+	    &mov	($__key,$key);		# advance rd_key
+	&set_label("12rounds",4);
 	    for ($i=1;$i<3;$i++) {
 		if ($vertical_spin) {
-		    &encvert("ebp",$s0,$s1,$s2,$s3);
+		    &encvert($tbl,$s0,$s1,$s2,$s3);
 		} else {
-		    &encstep(0,"ebp",$s0,$s1,$s2,$s3);
-		    &encstep(1,"ebp",$s1,$s2,$s3,$s0);
-		    &encstep(2,"ebp",$s2,$s3,$s0,$s1);
-		    &encstep(3,"ebp",$s3,$s0,$s1,$s2);
+		    &encstep(0,$tbl,$s0,$s1,$s2,$s3);
+		    &encstep(1,$tbl,$s1,$s2,$s3,$s0);
+		    &encstep(2,$tbl,$s2,$s3,$s0,$s1);
+		    &encstep(3,$tbl,$s3,$s0,$s1,$s2);
 		}
 		&xor	($s0,&DWP(16*$i+0,$key));
 		&xor	($s1,&DWP(16*$i+4,$key));
@@ -329,16 +915,16 @@ sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
 		&xor	($s3,&DWP(16*$i+12,$key));
 	    }
 	    &add	($key,32);
-	    &mov	(&DWP(12,"esp"),$key);	# advance rd_key
-	&set_label("10rounds");
+	    &mov	($__key,$key);		# advance rd_key
+	&set_label("10rounds",4);
 	    for ($i=1;$i<10;$i++) {
 		if ($vertical_spin) {
-		    &encvert("ebp",$s0,$s1,$s2,$s3);
+		    &encvert($tbl,$s0,$s1,$s2,$s3);
 		} else {
-		    &encstep(0,"ebp",$s0,$s1,$s2,$s3);
-		    &encstep(1,"ebp",$s1,$s2,$s3,$s0);
-		    &encstep(2,"ebp",$s2,$s3,$s0,$s1);
-		    &encstep(3,"ebp",$s3,$s0,$s1,$s2);
+		    &encstep(0,$tbl,$s0,$s1,$s2,$s3);
+		    &encstep(1,$tbl,$s1,$s2,$s3,$s0);
+		    &encstep(2,$tbl,$s2,$s3,$s0,$s1);
+		    &encstep(3,$tbl,$s3,$s0,$s1,$s2);
 		}
 		&xor	($s0,&DWP(16*$i+0,$key));
 		&xor	($s1,&DWP(16*$i+4,$key));
@@ -352,10 +938,10 @@ sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
 	    &mov	($s1="ebx",$key="edi");
 	    &mov	($s2="ecx",$acc="esi");
 	}
-	&enclast(0,"ebp",$s0,$s1,$s2,$s3);
-	&enclast(1,"ebp",$s1,$s2,$s3,$s0);
-	&enclast(2,"ebp",$s2,$s3,$s0,$s1);
-	&enclast(3,"ebp",$s3,$s0,$s1,$s2);
+	&enclast(0,$tbl,$s0,$s1,$s2,$s3);
+	&enclast(1,$tbl,$s1,$s2,$s3,$s0);
+	&enclast(2,$tbl,$s2,$s3,$s0,$s1);
+	&enclast(3,$tbl,$s3,$s0,$s1,$s2);
 
 	&add	($key,$small_footprint?16:160);
 	&xor	($s0,&DWP(0,$key));
@@ -430,38 +1016,198 @@ sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
 	&_data_word(0xdabfbf65, 0x31e6e6d7, 0xc6424284, 0xb86868d0);
 	&_data_word(0xc3414182, 0xb0999929, 0x772d2d5a, 0x110f0f1e);
 	&_data_word(0xcbb0b07b, 0xfc5454a8, 0xd6bbbb6d, 0x3a16162c);
+
+#Te4	# four copies of Te4 to choose from to avoid L1 aliasing
+	&data_byte(0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5);
+	&data_byte(0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76);
+	&data_byte(0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0);
+	&data_byte(0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0);
+	&data_byte(0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc);
+	&data_byte(0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15);
+	&data_byte(0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a);
+	&data_byte(0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75);
+	&data_byte(0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0);
+	&data_byte(0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84);
+	&data_byte(0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b);
+	&data_byte(0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf);
+	&data_byte(0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85);
+	&data_byte(0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8);
+	&data_byte(0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5);
+	&data_byte(0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2);
+	&data_byte(0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17);
+	&data_byte(0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73);
+	&data_byte(0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88);
+	&data_byte(0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb);
+	&data_byte(0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c);
+	&data_byte(0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79);
+	&data_byte(0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9);
+	&data_byte(0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08);
+	&data_byte(0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6);
+	&data_byte(0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a);
+	&data_byte(0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e);
+	&data_byte(0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e);
+	&data_byte(0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94);
+	&data_byte(0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf);
+	&data_byte(0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68);
+	&data_byte(0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16);
+
+	&data_byte(0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5);
+	&data_byte(0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76);
+	&data_byte(0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0);
+	&data_byte(0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0);
+	&data_byte(0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc);
+	&data_byte(0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15);
+	&data_byte(0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a);
+	&data_byte(0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75);
+	&data_byte(0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0);
+	&data_byte(0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84);
+	&data_byte(0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b);
+	&data_byte(0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf);
+	&data_byte(0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85);
+	&data_byte(0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8);
+	&data_byte(0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5);
+	&data_byte(0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2);
+	&data_byte(0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17);
+	&data_byte(0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73);
+	&data_byte(0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88);
+	&data_byte(0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb);
+	&data_byte(0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c);
+	&data_byte(0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79);
+	&data_byte(0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9);
+	&data_byte(0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08);
+	&data_byte(0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6);
+	&data_byte(0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a);
+	&data_byte(0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e);
+	&data_byte(0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e);
+	&data_byte(0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94);
+	&data_byte(0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf);
+	&data_byte(0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68);
+	&data_byte(0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16);
+
+	&data_byte(0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5);
+	&data_byte(0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76);
+	&data_byte(0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0);
+	&data_byte(0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0);
+	&data_byte(0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc);
+	&data_byte(0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15);
+	&data_byte(0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a);
+	&data_byte(0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75);
+	&data_byte(0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0);
+	&data_byte(0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84);
+	&data_byte(0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b);
+	&data_byte(0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf);
+	&data_byte(0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85);
+	&data_byte(0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8);
+	&data_byte(0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5);
+	&data_byte(0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2);
+	&data_byte(0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17);
+	&data_byte(0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73);
+	&data_byte(0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88);
+	&data_byte(0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb);
+	&data_byte(0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c);
+	&data_byte(0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79);
+	&data_byte(0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9);
+	&data_byte(0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08);
+	&data_byte(0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6);
+	&data_byte(0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a);
+	&data_byte(0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e);
+	&data_byte(0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e);
+	&data_byte(0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94);
+	&data_byte(0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf);
+	&data_byte(0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68);
+	&data_byte(0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16);
+
+	&data_byte(0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5);
+	&data_byte(0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76);
+	&data_byte(0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0);
+	&data_byte(0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0);
+	&data_byte(0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc);
+	&data_byte(0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15);
+	&data_byte(0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a);
+	&data_byte(0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75);
+	&data_byte(0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0);
+	&data_byte(0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84);
+	&data_byte(0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b);
+	&data_byte(0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf);
+	&data_byte(0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85);
+	&data_byte(0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8);
+	&data_byte(0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5);
+	&data_byte(0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2);
+	&data_byte(0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17);
+	&data_byte(0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73);
+	&data_byte(0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88);
+	&data_byte(0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb);
+	&data_byte(0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c);
+	&data_byte(0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79);
+	&data_byte(0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9);
+	&data_byte(0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08);
+	&data_byte(0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6);
+	&data_byte(0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a);
+	&data_byte(0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e);
+	&data_byte(0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e);
+	&data_byte(0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94);
+	&data_byte(0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf);
+	&data_byte(0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68);
+	&data_byte(0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16);
 #rcon:
 	&data_word(0x00000001, 0x00000002, 0x00000004, 0x00000008);
 	&data_word(0x00000010, 0x00000020, 0x00000040, 0x00000080);
-	&data_word(0x0000001b, 0x00000036, 0, 0, 0, 0, 0, 0);
+	&data_word(0x0000001b, 0x00000036, 0x00000000, 0x00000000);
+	&data_word(0x00000000, 0x00000000, 0x00000000, 0x00000000);
 &function_end_B("_x86_AES_encrypt");
 
 # void AES_encrypt (const void *inp,void *out,const AES_KEY *key);
-&public_label("AES_Te");
 &function_begin("AES_encrypt");
 	&mov	($acc,&wparam(0));		# load inp
 	&mov	($key,&wparam(2));		# load key
 
 	&mov	($s0,"esp");
-	&sub	("esp",24);
-	&and	("esp",-64);
-	&add	("esp",4);
-	&mov	(&DWP(16,"esp"),$s0);
+	&sub	("esp",36);
+	&and	("esp",-64);			# align to cache-line
+
+	# place stack frame just "above" the key schedule
+	&lea	($s1,&DWP(-64-63,$key));
+	&sub	($s1,"esp");
+	&neg	($s1);
+	&and	($s1,0x3C0);	# modulo 1024, but aligned to cache-line
+	&sub	("esp",$s1);
+	&add	("esp",4);	# 4 is reserved for caller's return address
+	&mov	($_esp,$s0);			# save stack pointer
 
 	&call   (&label("pic_point"));          # make it PIC!
 	&set_label("pic_point");
-	&blindpop("ebp");
-	&lea    ("ebp",&DWP(&label("AES_Te")."-".&label("pic_point"),"ebp"));
-
+	&blindpop($tbl);
+	&picmeup($s0,"OPENSSL_ia32cap_P",$tbl,&label("pic_point")) if (!$x86only);
+	&lea    ($tbl,&DWP(&label("AES_Te")."-".&label("pic_point"),$tbl));
+
+	# pick Te4 copy which can't "overlap" with stack frame or key schedule
+	&lea	($s1,&DWP(768-4,"esp"));
+	&sub	($s1,$tbl);
+	&and	($s1,0x300);
+	&lea	($tbl,&DWP(2048+128,$tbl,$s1));
+
+					if (!$x86only) {
+	&bt	(&DWP(0,$s0),25);	# check for SSE bit
+	&jnc	(&label("x86"));
+
+	&movq	("mm0",&QWP(0,$acc));
+	&movq	("mm4",&QWP(8,$acc));
+	&call	("_sse_AES_encrypt_compact");
+	&mov	("esp",$_esp);			# restore stack pointer
+	&mov	($acc,&wparam(1));		# load out
+	&movq	(&QWP(0,$acc),"mm0");		# write output data
+	&movq	(&QWP(8,$acc),"mm4");
+	&emms	();
+	&function_end_A();
+					}
+	&set_label("x86",16);
+	&mov	($_tbl,$tbl);
 	&mov	($s0,&DWP(0,$acc));		# load input data
 	&mov	($s1,&DWP(4,$acc));
 	&mov	($s2,&DWP(8,$acc));
 	&mov	($s3,&DWP(12,$acc));
-
-	&call	("_x86_AES_encrypt");
-
-	&mov	("esp",&DWP(16,"esp"));
-
+	&call	("_x86_AES_encrypt_compact");
+	&mov	("esp",$_esp);			# restore stack pointer
 	&mov	($acc,&wparam(1));		# load out
 	&mov	(&DWP(0,$acc),$s0);		# write output data
 	&mov	(&DWP(4,$acc),$s1);
@@ -469,7 +1215,370 @@ sub _data_word() { my $i; while(defined($i=shift)) { &data_word($i,$i); } }
 	&mov	(&DWP(12,$acc),$s3);
 &function_end("AES_encrypt");
 
-#------------------------------------------------------------------#
+#--------------------------------------------------------------------#
+
+######################################################################
+# "Compact" block function
+######################################################################
+
+sub deccompact()
+{ my $Fn = mov;
+  while ($#_>5) { pop(@_); $Fn=sub{}; }
+  my ($i,$td,@s)=@_;
+  my $tmp = $key;
+  my $out = $i==3?$s[0]:$acc;
+
+	# $Fn is used in first compact round and its purpose is to
+	# void restoration of some values from stack, so that after
+	# 4xdeccompact with extra argument $key, $s0 and $s1 values
+	# are left there...
+	if($i==3)   {	&$Fn	($key,$__key);			}
+	else        {	&mov	($out,$s[0]);			}
+			&and	($out,0xFF);
+			&movz	($out,&BP(-128,$td,$out,1));
+
+	if ($i==3)  {	$tmp=$s[1];				}
+			&movz	($tmp,&HB($s[1]));
+			&movz	($tmp,&BP(-128,$td,$tmp,1));
+			&shl	($tmp,8);
+			&xor	($out,$tmp);
+
+	if ($i==3)  {	$tmp=$s[2]; &mov ($s[1],$acc);		}
+	else        {	mov	($tmp,$s[2]);			}
+			&shr	($tmp,16);
+			&and	($tmp,0xFF);
+			&movz	($tmp,&BP(-128,$td,$tmp,1));
+			&shl	($tmp,16);
+			&xor	($out,$tmp);
+
+	if ($i==3)  {	$tmp=$s[3]; &$Fn ($s[2],$__s1);		}
+	else        {	&mov	($tmp,$s[3]);			}
+			&shr	($tmp,24);
+			&movz	($tmp,&BP(-128,$td,$tmp,1));
+			&shl	($tmp,24);
+			&xor	($out,$tmp);
+	if ($i<2)   {	&mov	(&DWP(4+4*$i,"esp"),$out);	}
+	if ($i==3)  {	&$Fn	($s[3],$__s0);			}
+}
+
+# must be called with 2,3,0,1 as argument sequence!!!
+sub dectransform()
+{ my @s = ($s0,$s1,$s2,$s3);
+  my $i = shift;
+  my $tmp = $key;
+  my $tp2 = @s[($i+2)%4]; $tp2 = @s[2] if ($i==1);
+  my $tp4 = @s[($i+3)%4]; $tp4 = @s[3] if ($i==1);
+  my $tp8 = $tbl;
+
+	&mov	($acc,$s[$i]);
+	&and	($acc,0x80808080);
+	&mov	($tmp,$acc);
+	&shr	($tmp,7);
+	&lea	($tp2,&DWP(0,$s[$i],$s[$i]));
+	&sub	($acc,$tmp);
+	&and	($tp2,0xfefefefe);
+	&and	($acc,0x1b1b1b1b);
+	&xor	($acc,$tp2);
+	&mov	($tp2,$acc);
+
+	&and	($acc,0x80808080);
+	&mov	($tmp,$acc);
+	&shr	($tmp,7);
+	&lea	($tp4,&DWP(0,$tp2,$tp2));
+	&sub	($acc,$tmp);
+	&and	($tp4,0xfefefefe);
+	&and	($acc,0x1b1b1b1b);
+	 &xor	($tp2,$s[$i]);	# tp2^tp1
+	&xor	($acc,$tp4);
+	&mov	($tp4,$acc);
+
+	&and	($acc,0x80808080);
+	&mov	($tmp,$acc);
+	&shr	($tmp,7);
+	&lea	($tp8,&DWP(0,$tp4,$tp4));
+	&sub	($acc,$tmp);
+	&and	($tp8,0xfefefefe);
+	&and	($acc,0x1b1b1b1b);
+	 &xor	($tp4,$s[$i]);	# tp4^tp1
+	 &rotl	($s[$i],8);	# = ROTATE(tp1,8)
+	&xor	($tp8,$acc);
+
+	&xor	($s[$i],$tp2);
+	&xor	($tp2,$tp8);
+	&rotl	($tp2,24);
+	&xor	($s[$i],$tp4);
+	&xor	($tp4,$tp8);
+	&rotl	($tp4,16);
+	&xor	($s[$i],$tp8);	# ^= tp8^(tp4^tp1)^(tp2^tp1)
+	&rotl	($tp8,8);
+	&xor	($s[$i],$tp2);	# ^= ROTATE(tp8^tp2^tp1,24)
+	&xor	($s[$i],$tp4);	# ^= ROTATE(tp8^tp4^tp1,16)
+	 &mov	($s[0],$__s0)			if($i==2); #prefetch $s0
+	 &mov	($s[1],$__s1)			if($i==3); #prefetch $s1
+	 &mov	($s[2],$__s2)			if($i==1);
+	&xor	($s[$i],$tp8);	# ^= ROTATE(tp8,8)
+
+	&mov	($s[3],$__s3)			if($i==1);
+	&mov	(&DWP(4+4*$i,"esp"),$s[$i])	if($i>=2);
+}
+
+&function_begin_B("_x86_AES_decrypt_compact");
+	# note that caller is expected to allocate stack frame for me!
+	&mov	($__key,$key);			# save key
+
+	&xor	($s0,&DWP(0,$key));		# xor with key
+	&xor	($s1,&DWP(4,$key));
+	&xor	($s2,&DWP(8,$key));
+	&xor	($s3,&DWP(12,$key));
+
+	&mov	($acc,&DWP(240,$key));		# load key->rounds
+
+	&lea	($acc,&DWP(-2,$acc,$acc));
+	&lea	($acc,&DWP(0,$key,$acc,8));
+	&mov	($__end,$acc);			# end of key schedule
+
+	# prefetch Td4
+	&mov	($key,&DWP(0-128,$tbl));
+	&mov	($acc,&DWP(32-128,$tbl));
+	&mov	($key,&DWP(64-128,$tbl));
+	&mov	($acc,&DWP(96-128,$tbl));
+	&mov	($key,&DWP(128-128,$tbl));
+	&mov	($acc,&DWP(160-128,$tbl));
+	&mov	($key,&DWP(192-128,$tbl));
+	&mov	($acc,&DWP(224-128,$tbl));
+
+	&set_label("loop",16);
+
+		&deccompact(0,$tbl,$s0,$s3,$s2,$s1,1);
+		&deccompact(1,$tbl,$s1,$s0,$s3,$s2,1);
+		&deccompact(2,$tbl,$s2,$s1,$s0,$s3,1);
+		&deccompact(3,$tbl,$s3,$s2,$s1,$s0,1);
+		&dectransform(2);
+		&dectransform(3);
+		&dectransform(0);
+		&dectransform(1);
+		&mov 	($key,$__key);
+		&mov	($tbl,$__tbl);
+		&add	($key,16);		# advance rd_key
+		&xor	($s0,&DWP(0,$key));
+		&xor	($s1,&DWP(4,$key));
+		&xor	($s2,&DWP(8,$key));
+		&xor	($s3,&DWP(12,$key));
+
+	&cmp	($key,$__end);
+	&mov	($__key,$key);
+	&jb	(&label("loop"));
+
+	&deccompact(0,$tbl,$s0,$s3,$s2,$s1);
+	&deccompact(1,$tbl,$s1,$s0,$s3,$s2);
+	&deccompact(2,$tbl,$s2,$s1,$s0,$s3);
+	&deccompact(3,$tbl,$s3,$s2,$s1,$s0);
+
+	&xor	($s0,&DWP(16,$key));
+	&xor	($s1,&DWP(20,$key));
+	&xor	($s2,&DWP(24,$key));
+	&xor	($s3,&DWP(28,$key));
+
+	&ret	();
+&function_end_B("_x86_AES_decrypt_compact");
+
+######################################################################
+# "Compact" SSE block function.
+######################################################################
+
+sub sse_deccompact()
+{
+	&pshufw	("mm1","mm0",0x0c);		#  7, 6, 1, 0
+	&movd	("eax","mm1");			#  7, 6, 1, 0
+
+	&pshufw	("mm5","mm4",0x09);		# 13,12,11,10
+	&movz	($acc,&LB("eax"));		#  0
+	&movz	("ecx",&BP(-128,$tbl,$acc,1));	#  0
+	&movd	("ebx","mm5");			# 13,12,11,10
+	&movz	("edx",&HB("eax"));		#  1
+	&movz	("edx",&BP(-128,$tbl,"edx",1));	#  1
+	&shl	("edx",8);			#  1
+
+	&pshufw	("mm2","mm0",0x06);		#  3, 2, 5, 4
+	&movz	($acc,&LB("ebx"));		# 10
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 10
+	&shl	($acc,16);			# 10
+	&or	("ecx",$acc);			# 10
+	&shr	("eax",16);			#  7, 6
+	&movz	($acc,&HB("ebx"));		# 11
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 11
+	&shl	($acc,24);			# 11
+	&or	("edx",$acc);			# 11
+	&shr	("ebx",16);			# 13,12
+
+	&pshufw	("mm6","mm4",0x03);		# 9, 8,15,14
+	&movz	($acc,&HB("eax"));		#  7
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	#  7
+	&shl	($acc,24);			#  7
+	&or	("ecx",$acc);			#  7
+	&movz	($acc,&HB("ebx"));		# 13
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 13
+	&shl	($acc,8);			# 13
+	&or	("ecx",$acc);			# 13
+	&movd	("mm0","ecx");			# t[0] collected
+
+	&movz	($acc,&LB("eax"));		#  6
+	&movd	("eax","mm2");			#  3, 2, 5, 4
+	&movz	("ecx",&BP(-128,$tbl,$acc,1));	#  6
+	&shl	("ecx",16);			#  6
+	&movz	($acc,&LB("ebx"));		# 12
+	&movd	("ebx","mm6");			#  9, 8,15,14
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 12
+	&or	("ecx",$acc);			# 12
+
+	&movz	($acc,&LB("eax"));		#  4
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	#  4
+	&or	("edx",$acc);			#  4
+	&movz	($acc,&LB("ebx"));		# 14
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 14
+	&shl	($acc,16);			# 14
+	&or	("edx",$acc);			# 14
+	&movd	("mm1","edx");			# t[1] collected
+
+	&movz	($acc,&HB("eax"));		#  5
+	&movz	("edx",&BP(-128,$tbl,$acc,1));	#  5
+	&shl	("edx",8);			#  5
+	&movz	($acc,&HB("ebx"));		# 15
+	&shr	("eax",16);			#  3, 2
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	# 15
+	&shl	($acc,24);			# 15
+	&or	("edx",$acc);			# 15
+	&shr	("ebx",16);			#  9, 8
+
+	&punpckldq	("mm0","mm1");		# t[0,1] collected
+
+	&movz	($acc,&HB("ebx"));		#  9
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	#  9
+	&shl	($acc,8);			#  9
+	&or	("ecx",$acc);			#  9
+	&and	("ebx",0xff);			#  8
+	&movz	("ebx",&BP(-128,$tbl,"ebx",1));	#  8
+	&or	("edx","ebx");			#  8
+	&movz	($acc,&LB("eax"));		#  2
+	&movz	($acc,&BP(-128,$tbl,$acc,1));	#  2
+	&shl	($acc,16);			#  2
+	&or	("edx",$acc);			#  2
+	&movd	("mm4","edx");			# t[2] collected
+	&movz	("eax",&HB("eax"));		#  3
+	&movz	("eax",&BP(-128,$tbl,"eax",1));	#  3
+	&shl	("eax",24);			#  3
+	&or	("ecx","eax");			#  3
+	&movd	("mm5","ecx");			# t[3] collected
+
+	&punpckldq	("mm4","mm5");		# t[2,3] collected
+}
+
+					if (!$x86only) {
+&function_begin_B("_sse_AES_decrypt_compact");
+	&pxor	("mm0",&QWP(0,$key));	#  7, 6, 5, 4, 3, 2, 1, 0
+	&pxor	("mm4",&QWP(8,$key));	# 15,14,13,12,11,10, 9, 8
+
+	# note that caller is expected to allocate stack frame for me!
+	&mov	($acc,&DWP(240,$key));		# load key->rounds
+	&lea	($acc,&DWP(-2,$acc,$acc));
+	&lea	($acc,&DWP(0,$key,$acc,8));
+	&mov	($__end,$acc);			# end of key schedule
+
+	&mov	($s0,0x1b1b1b1b);		# magic constant
+	&mov	(&DWP(8,"esp"),$s0);
+	&mov	(&DWP(12,"esp"),$s0);
+
+	# prefetch Td4
+	&mov	($s0,&DWP(0-128,$tbl));
+	&mov	($s1,&DWP(32-128,$tbl));
+	&mov	($s2,&DWP(64-128,$tbl));
+	&mov	($s3,&DWP(96-128,$tbl));
+	&mov	($s0,&DWP(128-128,$tbl));
+	&mov	($s1,&DWP(160-128,$tbl));
+	&mov	($s2,&DWP(192-128,$tbl));
+	&mov	($s3,&DWP(224-128,$tbl));
+
+	&set_label("loop",16);
+		&sse_deccompact();
+		&add	($key,16);
+		&cmp	($key,$__end);
+		&ja	(&label("out"));
+
+		# ROTATE(x^y,N) == ROTATE(x,N)^ROTATE(y,N)
+		&movq	("mm3","mm0");		&movq	("mm7","mm4");
+		&movq	("mm2","mm0",1);	&movq	("mm6","mm4",1);
+		&movq	("mm1","mm0");		&movq	("mm5","mm4");
+		&pshufw	("mm0","mm0",0xb1);	&pshufw	("mm4","mm4",0xb1);# = ROTATE(tp0,16)
+		&pslld	("mm2",8);		&pslld	("mm6",8);
+		&psrld	("mm3",8);		&psrld	("mm7",8);
+		&pxor	("mm0","mm2");		&pxor	("mm4","mm6");	# ^= tp0<<8
+		&pxor	("mm0","mm3");		&pxor	("mm4","mm7");	# ^= tp0>>8
+		&pslld	("mm2",16);		&pslld	("mm6",16);
+		&psrld	("mm3",16);		&psrld	("mm7",16);
+		&pxor	("mm0","mm2");		&pxor	("mm4","mm6");	# ^= tp0<<24
+		&pxor	("mm0","mm3");		&pxor	("mm4","mm7");	# ^= tp0>>24
+
+		&movq	("mm3",&QWP(8,"esp"));
+		&pxor	("mm2","mm2");		&pxor	("mm6","mm6");
+		&pcmpgtb("mm2","mm1");		&pcmpgtb("mm6","mm5");
+		&pand	("mm2","mm3");		&pand	("mm6","mm3");
+		&paddb	("mm1","mm1");		&paddb	("mm5","mm5");
+		&pxor	("mm1","mm2");		&pxor	("mm5","mm6");	# tp2
+		&movq	("mm3","mm1");		&movq	("mm7","mm5");
+		&movq	("mm2","mm1");		&movq	("mm6","mm5");
+		&pxor	("mm0","mm1");		&pxor	("mm4","mm5");	# ^= tp2
+		&pslld	("mm3",24);		&pslld	("mm7",24);
+		&psrld	("mm2",8);		&psrld	("mm6",8);
+		&pxor	("mm0","mm3");		&pxor	("mm4","mm7");	# ^= tp2<<24
+		&pxor	("mm0","mm2");		&pxor	("mm4","mm6");	# ^= tp2>>8
+
+		&movq	("mm2",&QWP(8,"esp"));
+		&pxor	("mm3","mm3");		&pxor	("mm7","mm7");
+		&pcmpgtb("mm3","mm1");		&pcmpgtb("mm7","mm5");
+		&pand	("mm3","mm2");		&pand	("mm7","mm2");
+		&paddb	("mm1","mm1");		&paddb	("mm5","mm5");
+		&pxor	("mm1","mm3");		&pxor	("mm5","mm7");	# tp4
+		&pshufw	("mm3","mm1",0xb1);	&pshufw	("mm7","mm5",0xb1);
+		&pxor	("mm0","mm1");		&pxor	("mm4","mm5");	# ^= tp4
+		&pxor	("mm0","mm3");		&pxor	("mm4","mm7");	# ^= ROTATE(tp4,16)	
+
+		&pxor	("mm3","mm3");		&pxor	("mm7","mm7");
+		&pcmpgtb("mm3","mm1");		&pcmpgtb("mm7","mm5");
+		&pand	("mm3","mm2");		&pand	("mm7","mm2");
+		&paddb	("mm1","mm1");		&paddb	("mm5","mm5");
+		&pxor	("mm1","mm3");		&pxor	("mm5","mm7");	# tp8
+		&pxor	("mm0","mm1");		&pxor	("mm4","mm5");	# ^= tp8
+		&movq	("mm3","mm1");		&movq	("mm7","mm5");
+		&pshufw	("mm2","mm1",0xb1);	&pshufw	("mm6","mm5",0xb1);
+		&pxor	("mm0","mm2");		&pxor	("mm4","mm6");	# ^= ROTATE(tp8,16)
+		&pslld	("mm1",8);		&pslld	("mm5",8);
+		&psrld	("mm3",8);		&psrld	("mm7",8);
+		&movq	("mm2",&QWP(0,$key));	&movq	("mm6",&QWP(8,$key));
+		&pxor	("mm0","mm1");		&pxor	("mm4","mm5");	# ^= tp8<<8
+		&pxor	("mm0","mm3");		&pxor	("mm4","mm7");	# ^= tp8>>8
+		&mov	($s0,&DWP(0-128,$tbl));
+		&pslld	("mm1",16);		&pslld	("mm5",16);
+		&mov	($s1,&DWP(64-128,$tbl));
+		&psrld	("mm3",16);		&psrld	("mm7",16);
+		&mov	($s2,&DWP(128-128,$tbl));
+		&pxor	("mm0","mm1");		&pxor	("mm4","mm5");	# ^= tp8<<24
+		&mov	($s3,&DWP(192-128,$tbl));
+		&pxor	("mm0","mm3");		&pxor	("mm4","mm7");	# ^= tp8>>24
+
+		&pxor	("mm0","mm2");		&pxor	("mm4","mm6");
+	&jmp	(&label("loop"));
+
+	&set_label("out",16);
+	&pxor	("mm0",&QWP(0,$key));
+	&pxor	("mm4",&QWP(8,$key));
+
+	&ret	();
+&function_end_B("_sse_AES_decrypt_compact");
+					}
+
+######################################################################
+# Vanilla block function.
+######################################################################
 
 sub decstep()
 { my ($i,$td,@s) = @_;
@@ -480,7 +1589,7 @@ sub decstep()
 	# optimal... or rather that all attempts to reorder didn't
 	# result in better performance [which by the way is not a
 	# bit lower than ecryption].
-	if($i==3)   {	&mov	($key,&DWP(12,"esp"));		}
+	if($i==3)   {	&mov	($key,$__key);			}
 	else        {	&mov	($out,$s[0]);			}
 			&and	($out,0xFF);
 			&mov	($out,&DWP(0,$td,$out,8));
@@ -495,12 +1604,12 @@ sub decstep()
 			&and	($tmp,0xFF);
 			&xor	($out,&DWP(2,$td,$tmp,8));
 
-	if ($i==3)  {	$tmp=$s[3]; &mov ($s[2],&DWP(8,"esp"));	}
+	if ($i==3)  {	$tmp=$s[3]; &mov ($s[2],$__s1);		}
 	else        {	&mov	($tmp,$s[3]);			}
 			&shr	($tmp,24);
 			&xor	($out,&DWP(1,$td,$tmp,8));
 	if ($i<2)   {	&mov	(&DWP(4+4*$i,"esp"),$out);	}
-	if ($i==3)  {	&mov	($s[3],&DWP(4,"esp"));		}
+	if ($i==3)  {	&mov	($s[3],$__s0);			}
 			&comment();
 }
 
@@ -509,14 +1618,24 @@ sub declast()
   my $tmp = $key;
   my $out = $i==3?$s[0]:$acc;
 
-	if($i==3)   {	&mov	($key,&DWP(12,"esp"));		}
+	if($i==0)   {	&lea	($td,&DWP(2048+128,$td));
+			&mov	($tmp,&DWP(0-128,$td));
+			&mov	($acc,&DWP(32-128,$td));
+			&mov	($tmp,&DWP(64-128,$td));
+			&mov	($acc,&DWP(96-128,$td));
+			&mov	($tmp,&DWP(128-128,$td));
+			&mov	($acc,&DWP(160-128,$td));
+			&mov	($tmp,&DWP(192-128,$td));
+			&mov	($acc,&DWP(224-128,$td));
+			&lea	($td,&DWP(-128,$td));		}
+	if($i==3)   {	&mov	($key,$__key);			}
 	else        {	&mov	($out,$s[0]);			}
 			&and	($out,0xFF);
-			&movz	($out,&BP(2048,$td,$out,1));
+			&movz	($out,&BP(0,$td,$out,1));
 
 	if ($i==3)  {	$tmp=$s[1];				}
 			&movz	($tmp,&HB($s[1]));
-			&movz	($tmp,&BP(2048,$td,$tmp,1));
+			&movz	($tmp,&BP(0,$td,$tmp,1));
 			&shl	($tmp,8);
 			&xor	($out,$tmp);
 
@@ -524,24 +1643,24 @@ sub declast()
 	else        {	mov	($tmp,$s[2]);			}
 			&shr	($tmp,16);
 			&and	($tmp,0xFF);
-			&movz	($tmp,&BP(2048,$td,$tmp,1));
+			&movz	($tmp,&BP(0,$td,$tmp,1));
 			&shl	($tmp,16);
 			&xor	($out,$tmp);
 
-	if ($i==3)  {	$tmp=$s[3]; &mov ($s[2],&DWP(8,"esp"));	}
+	if ($i==3)  {	$tmp=$s[3]; &mov ($s[2],$__s1);		}
 	else        {	&mov	($tmp,$s[3]);			}
 			&shr	($tmp,24);
-			&movz	($tmp,&BP(2048,$td,$tmp,1));
+			&movz	($tmp,&BP(0,$td,$tmp,1));
 			&shl	($tmp,24);
 			&xor	($out,$tmp);
 	if ($i<2)   {	&mov	(&DWP(4+4*$i,"esp"),$out);	}
-	if ($i==3)  {	&mov	($s[3],&DWP(4,"esp"));		}
+	if ($i==3)  {	&mov	($s[3],$__s0);
+			&lea	($td,&DWP(-2048,$td));		}
 }
 
-&public_label("AES_Td");
 &function_begin_B("_x86_AES_decrypt");
 	# note that caller is expected to allocate stack frame for me!
-	&mov	(&DWP(12,"esp"),$key);		# save key
+	&mov	($__key,$key);			# save key
 
 	&xor	($s0,&DWP(0,$key));		# xor with key
 	&xor	($s1,&DWP(4,$key));
@@ -553,20 +1672,19 @@ sub declast()
 	if ($small_footprint) {
 	    &lea	($acc,&DWP(-2,$acc,$acc));
 	    &lea	($acc,&DWP(0,$key,$acc,8));
-	    &mov	(&DWP(16,"esp"),$acc);	# end of key schedule
-	    &align	(4);
-	    &set_label("loop");
-		&decstep(0,"ebp",$s0,$s3,$s2,$s1);
-		&decstep(1,"ebp",$s1,$s0,$s3,$s2);
-		&decstep(2,"ebp",$s2,$s1,$s0,$s3);
-		&decstep(3,"ebp",$s3,$s2,$s1,$s0);
+	    &mov	($__end,$acc);		# end of key schedule
+	    &set_label("loop",16);
+		&decstep(0,$tbl,$s0,$s3,$s2,$s1);
+		&decstep(1,$tbl,$s1,$s0,$s3,$s2);
+		&decstep(2,$tbl,$s2,$s1,$s0,$s3);
+		&decstep(3,$tbl,$s3,$s2,$s1,$s0);
 		&add	($key,16);		# advance rd_key
 		&xor	($s0,&DWP(0,$key));
 		&xor	($s1,&DWP(4,$key));
 		&xor	($s2,&DWP(8,$key));
 		&xor	($s3,&DWP(12,$key));
-	    &cmp	($key,&DWP(16,"esp"));
-	    &mov	(&DWP(12,"esp"),$key);
+	    &cmp	($key,$__end);
+	    &mov	($__key,$key);
 	    &jb		(&label("loop"));
 	}
 	else {
@@ -575,38 +1693,38 @@ sub declast()
 	    &cmp	($acc,12);
 	    &jle	(&label("12rounds"));
 
-	&set_label("14rounds");
+	&set_label("14rounds",4);
 	    for ($i=1;$i<3;$i++) {
-		&decstep(0,"ebp",$s0,$s3,$s2,$s1);
-		&decstep(1,"ebp",$s1,$s0,$s3,$s2);
-		&decstep(2,"ebp",$s2,$s1,$s0,$s3);
-		&decstep(3,"ebp",$s3,$s2,$s1,$s0);
+		&decstep(0,$tbl,$s0,$s3,$s2,$s1);
+		&decstep(1,$tbl,$s1,$s0,$s3,$s2);
+		&decstep(2,$tbl,$s2,$s1,$s0,$s3);
+		&decstep(3,$tbl,$s3,$s2,$s1,$s0);
 		&xor	($s0,&DWP(16*$i+0,$key));
 		&xor	($s1,&DWP(16*$i+4,$key));
 		&xor	($s2,&DWP(16*$i+8,$key));
 		&xor	($s3,&DWP(16*$i+12,$key));
 	    }
 	    &add	($key,32);
-	    &mov	(&DWP(12,"esp"),$key);	# advance rd_key
-	&set_label("12rounds");
+	    &mov	($__key,$key);		# advance rd_key
+	&set_label("12rounds",4);
 	    for ($i=1;$i<3;$i++) {
-		&decstep(0,"ebp",$s0,$s3,$s2,$s1);
-		&decstep(1,"ebp",$s1,$s0,$s3,$s2);
-		&decstep(2,"ebp",$s2,$s1,$s0,$s3);
-		&decstep(3,"ebp",$s3,$s2,$s1,$s0);
+		&decstep(0,$tbl,$s0,$s3,$s2,$s1);
+		&decstep(1,$tbl,$s1,$s0,$s3,$s2);
+		&decstep(2,$tbl,$s2,$s1,$s0,$s3);
+		&decstep(3,$tbl,$s3,$s2,$s1,$s0);
 		&xor	($s0,&DWP(16*$i+0,$key));
 		&xor	($s1,&DWP(16*$i+4,$key));
 		&xor	($s2,&DWP(16*$i+8,$key));
 		&xor	($s3,&DWP(16*$i+12,$key));
 	    }
 	    &add	($key,32);
-	    &mov	(&DWP(12,"esp"),$key);	# advance rd_key
-	&set_label("10rounds");
+	    &mov	($__key,$key);		# advance rd_key
+	&set_label("10rounds",4);
 	    for ($i=1;$i<10;$i++) {
-		&decstep(0,"ebp",$s0,$s3,$s2,$s1);
-		&decstep(1,"ebp",$s1,$s0,$s3,$s2);
-		&decstep(2,"ebp",$s2,$s1,$s0,$s3);
-		&decstep(3,"ebp",$s3,$s2,$s1,$s0);
+		&decstep(0,$tbl,$s0,$s3,$s2,$s1);
+		&decstep(1,$tbl,$s1,$s0,$s3,$s2);
+		&decstep(2,$tbl,$s2,$s1,$s0,$s3);
+		&decstep(3,$tbl,$s3,$s2,$s1,$s0);
 		&xor	($s0,&DWP(16*$i+0,$key));
 		&xor	($s1,&DWP(16*$i+4,$key));
 		&xor	($s2,&DWP(16*$i+8,$key));
@@ -614,10 +1732,10 @@ sub declast()
 	    }
 	}
 
-	&declast(0,"ebp",$s0,$s3,$s2,$s1);
-	&declast(1,"ebp",$s1,$s0,$s3,$s2);
-	&declast(2,"ebp",$s2,$s1,$s0,$s3);
-	&declast(3,"ebp",$s3,$s2,$s1,$s0);
+	&declast(0,$tbl,$s0,$s3,$s2,$s1);
+	&declast(1,$tbl,$s1,$s0,$s3,$s2);
+	&declast(2,$tbl,$s2,$s1,$s0,$s3);
+	&declast(3,$tbl,$s3,$s2,$s1,$s0);
 
 	&add	($key,$small_footprint?16:160);
 	&xor	($s0,&DWP(0,$key));
@@ -692,7 +1810,107 @@ sub declast()
 	&_data_word(0x72c31d16, 0x0c25e2bc, 0x8b493c28, 0x41950dff);
 	&_data_word(0x7101a839, 0xdeb30c08, 0x9ce4b4d8, 0x90c15664);
 	&_data_word(0x6184cb7b, 0x70b632d5, 0x745c6c48, 0x4257b8d0);
-#Td4:
+
+#Td4:	# four copies of Td4 to choose from to avoid L1 aliasing
+	&data_byte(0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38);
+	&data_byte(0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb);
+	&data_byte(0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87);
+	&data_byte(0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb);
+	&data_byte(0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d);
+	&data_byte(0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e);
+	&data_byte(0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2);
+	&data_byte(0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25);
+	&data_byte(0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16);
+	&data_byte(0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92);
+	&data_byte(0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda);
+	&data_byte(0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84);
+	&data_byte(0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a);
+	&data_byte(0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06);
+	&data_byte(0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02);
+	&data_byte(0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b);
+	&data_byte(0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea);
+	&data_byte(0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73);
+	&data_byte(0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85);
+	&data_byte(0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e);
+	&data_byte(0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89);
+	&data_byte(0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b);
+	&data_byte(0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20);
+	&data_byte(0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4);
+	&data_byte(0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31);
+	&data_byte(0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f);
+	&data_byte(0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d);
+	&data_byte(0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef);
+	&data_byte(0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0);
+	&data_byte(0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61);
+	&data_byte(0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26);
+	&data_byte(0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d);
+
+	&data_byte(0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38);
+	&data_byte(0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb);
+	&data_byte(0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87);
+	&data_byte(0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb);
+	&data_byte(0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d);
+	&data_byte(0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e);
+	&data_byte(0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2);
+	&data_byte(0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25);
+	&data_byte(0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16);
+	&data_byte(0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92);
+	&data_byte(0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda);
+	&data_byte(0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84);
+	&data_byte(0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a);
+	&data_byte(0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06);
+	&data_byte(0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02);
+	&data_byte(0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b);
+	&data_byte(0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea);
+	&data_byte(0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73);
+	&data_byte(0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85);
+	&data_byte(0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e);
+	&data_byte(0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89);
+	&data_byte(0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b);
+	&data_byte(0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20);
+	&data_byte(0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4);
+	&data_byte(0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31);
+	&data_byte(0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f);
+	&data_byte(0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d);
+	&data_byte(0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef);
+	&data_byte(0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0);
+	&data_byte(0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61);
+	&data_byte(0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26);
+	&data_byte(0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d);
+
+	&data_byte(0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38);
+	&data_byte(0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb);
+	&data_byte(0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87);
+	&data_byte(0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb);
+	&data_byte(0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d);
+	&data_byte(0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e);
+	&data_byte(0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2);
+	&data_byte(0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25);
+	&data_byte(0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16);
+	&data_byte(0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92);
+	&data_byte(0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda);
+	&data_byte(0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84);
+	&data_byte(0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a);
+	&data_byte(0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06);
+	&data_byte(0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02);
+	&data_byte(0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b);
+	&data_byte(0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea);
+	&data_byte(0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73);
+	&data_byte(0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85);
+	&data_byte(0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e);
+	&data_byte(0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89);
+	&data_byte(0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b);
+	&data_byte(0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20);
+	&data_byte(0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4);
+	&data_byte(0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31);
+	&data_byte(0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f);
+	&data_byte(0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d);
+	&data_byte(0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef);
+	&data_byte(0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0);
+	&data_byte(0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61);
+	&data_byte(0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26);
+	&data_byte(0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d);
+
 	&data_byte(0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38);
 	&data_byte(0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb);
 	&data_byte(0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87);
@@ -728,43 +1946,57 @@ sub declast()
 &function_end_B("_x86_AES_decrypt");
 
 # void AES_decrypt (const void *inp,void *out,const AES_KEY *key);
-&public_label("AES_Td");
 &function_begin("AES_decrypt");
 	&mov	($acc,&wparam(0));		# load inp
 	&mov	($key,&wparam(2));		# load key
 
 	&mov	($s0,"esp");
-	&sub	("esp",24);
-	&and	("esp",-64);
-	&add	("esp",4);
-	&mov	(&DWP(16,"esp"),$s0);
+	&sub	("esp",36);
+	&and	("esp",-64);			# align to cache-line
+
+	# place stack frame just "above" the key schedule
+	&lea	($s1,&DWP(-64-63,$key));
+	&sub	($s1,"esp");
+	&neg	($s1);
+	&and	($s1,0x3C0);	# modulo 1024, but aligned to cache-line
+	&sub	("esp",$s1);
+	&add	("esp",4);	# 4 is reserved for caller's return address
+	&mov	($_esp,$s0);	# save stack pointer
 
 	&call   (&label("pic_point"));          # make it PIC!
 	&set_label("pic_point");
-	&blindpop("ebp");
-	&lea    ("ebp",&DWP(&label("AES_Td")."-".&label("pic_point"),"ebp"));
-
-	# prefetch Td4
-	&lea	("ebp",&DWP(2048+128,"ebp"));
-	&mov	($s0,&DWP(0-128,"ebp"));
-	&mov	($s1,&DWP(32-128,"ebp"));
-	&mov	($s2,&DWP(64-128,"ebp"));
-	&mov	($s3,&DWP(96-128,"ebp"));
-	&mov	($s0,&DWP(128-128,"ebp"));
-	&mov	($s1,&DWP(160-128,"ebp"));
-	&mov	($s2,&DWP(192-128,"ebp"));
-	&mov	($s3,&DWP(224-128,"ebp"));
-	&lea	("ebp",&DWP(-2048-128,"ebp"));
-
+	&blindpop($tbl);
+	&picmeup($s0,"OPENSSL_ia32cap_P",$tbl,&label("pic_point")) if(!$x86only);
+	&lea    ($tbl,&DWP(&label("AES_Td")."-".&label("pic_point"),$tbl));
+
+	# pick Td4 copy which can't "overlap" with stack frame or key schedule
+	&lea	($s1,&DWP(768-4,"esp"));
+	&sub	($s1,$tbl);
+	&and	($s1,0x300);
+	&lea	($tbl,&DWP(2048+128,$tbl,$s1));
+
+					if (!$x86only) {
+	&bt	(&DWP(0,$s0),25);	# check for SSE bit
+	&jnc	(&label("x86"));
+
+	&movq	("mm0",&QWP(0,$acc));
+	&movq	("mm4",&QWP(8,$acc));
+	&call	("_sse_AES_decrypt_compact");
+	&mov	("esp",$_esp);			# restore stack pointer
+	&mov	($acc,&wparam(1));		# load out
+	&movq	(&QWP(0,$acc),"mm0");		# write output data
+	&movq	(&QWP(8,$acc),"mm4");
+	&emms	();
+	&function_end_A();
+					}
+	&set_label("x86",16);
+	&mov	($_tbl,$tbl);
 	&mov	($s0,&DWP(0,$acc));		# load input data
 	&mov	($s1,&DWP(4,$acc));
 	&mov	($s2,&DWP(8,$acc));
 	&mov	($s3,&DWP(12,$acc));
-
-	&call	("_x86_AES_decrypt");
-
-	&mov	("esp",&DWP(16,"esp"));
-
+	&call	("_x86_AES_decrypt_compact");
+	&mov	("esp",$_esp);			# restore stack pointer
 	&mov	($acc,&wparam(1));		# load out
 	&mov	(&DWP(0,$acc),$s0);		# write output data
 	&mov	(&DWP(4,$acc),$s1);
@@ -777,126 +2009,136 @@ sub declast()
 #			unsigned char *ivp,const int enc);
 {
 # stack frame layout
-# -4(%esp)	0(%esp)		return address
-# 0(%esp)	4(%esp)		tmp1
-# 4(%esp)	8(%esp)		tmp2
-# 8(%esp)	12(%esp)	key
-# 12(%esp)	16(%esp)	end of key schedule
-my $_esp=&DWP(16,"esp");	#saved %esp
-my $_inp=&DWP(20,"esp");	#copy of wparam(0)
-my $_out=&DWP(24,"esp");	#copy of wparam(1)
-my $_len=&DWP(28,"esp");	#copy of wparam(2)
-my $_key=&DWP(32,"esp");	#copy of wparam(3)
-my $_ivp=&DWP(36,"esp");	#copy of wparam(4)
-my $_tmp=&DWP(40,"esp");	#volatile variable
-my $ivec=&DWP(44,"esp");	#ivec[16]
-my $aes_key=&DWP(60,"esp");	#copy of aes_key
-my $mark=&DWP(60+240,"esp");	#copy of aes_key->rounds
-
-&public_label("AES_Te");
-&public_label("AES_Td");
+#             -4(%esp)		# return address	 0(%esp)
+#              0(%esp)		# s0 backing store	 4(%esp)	
+#              4(%esp)		# s1 backing store	 8(%esp)
+#              8(%esp)		# s2 backing store	12(%esp)
+#             12(%esp)		# s3 backing store	16(%esp)
+#             16(%esp)		# key backup		20(%esp)
+#             20(%esp)		# end of key schedule	24(%esp)
+#             24(%esp)		# %ebp backup		28(%esp)
+#             28(%esp)		# %esp backup
+my $_inp=&DWP(32,"esp");	# copy of wparam(0)
+my $_out=&DWP(36,"esp");	# copy of wparam(1)
+my $_len=&DWP(40,"esp");	# copy of wparam(2)
+my $_key=&DWP(44,"esp");	# copy of wparam(3)
+my $_ivp=&DWP(48,"esp");	# copy of wparam(4)
+my $_tmp=&DWP(52,"esp");	# volatile variable
+#
+my $ivec=&DWP(60,"esp");	# ivec[16]
+my $aes_key=&DWP(76,"esp");	# copy of aes_key
+my $mark=&DWP(76+240,"esp");	# copy of aes_key->rounds
+
 &function_begin("AES_cbc_encrypt");
 	&mov	($s2 eq "ecx"? $s2 : "",&wparam(2));	# load len
 	&cmp	($s2,0);
-	&je	(&label("enc_out"));
+	&je	(&label("drop_out"));
 
 	&call   (&label("pic_point"));		# make it PIC!
 	&set_label("pic_point");
-	&blindpop("ebp");
-
-	&pushf	();
-	&cld	();
+	&blindpop($tbl);
+	&picmeup($s0,"OPENSSL_ia32cap_P",$tbl,&label("pic_point")) if(!$x86only);
 
 	&cmp	(&wparam(5),0);
-	&je	(&label("DECRYPT"));
+	&lea    ($tbl,&DWP(&label("AES_Te")."-".&label("pic_point"),$tbl));
+	&jne	(&label("picked_te"));
+	&lea	($tbl,&DWP(&label("AES_Td")."-".&label("AES_Te"),$tbl));
+	&set_label("picked_te");
 
-	&lea    ("ebp",&DWP(&label("AES_Te")."-".&label("pic_point"),"ebp"));
-
-	# allocate aligned stack frame...
-	&lea	($key,&DWP(-64-244,"esp"));
-	&and	($key,-64);
+	# one can argue if this is required
+	&pushf	();
+	&cld	();
 
-	# ... and make sure it doesn't alias with AES_Te modulo 4096
-	&mov	($s0,"ebp");
-	&lea	($s1,&DWP(2048,"ebp"));
-	&mov	($s3,$key);
+	&cmp	($s2,$speed_limit);
+	&jb	(&label("slow_way"));
+	&test	($s2,15);
+	&jnz	(&label("slow_way"));
+					if (!$x86only) {
+	&bt	(&DWP(0,$s0),28);	# check for hyper-threading bit
+	&jc	(&label("slow_way"));
+					}
+	# pre-allocate aligned stack frame...
+	&lea	($acc,&DWP(-80-244,"esp"));
+	&and	($acc,-64);
+
+	# ... and make sure it doesn't alias with $tbl modulo 4096
+	&mov	($s0,$tbl);
+	&lea	($s1,&DWP(2048+256,$tbl));
+	&mov	($s3,$acc);
 	&and	($s0,0xfff);		# s = %ebp&0xfff
-	&and	($s1,0xfff);		# e = (%ebp+2048)&0xfff
+	&and	($s1,0xfff);		# e = (%ebp+2048+256)&0xfff
 	&and	($s3,0xfff);		# p = %esp&0xfff
 
 	&cmp	($s3,$s1);		# if (p>=e) %esp =- (p-e);
-	&jb	(&label("te_break_out"));
+	&jb	(&label("tbl_break_out"));
 	&sub	($s3,$s1);
-	&sub	($key,$s3);
-	&jmp	(&label("te_ok"));
-	&set_label("te_break_out");	# else %esp -= (p-s)&0xfff + framesz;
+	&sub	($acc,$s3);
+	&jmp	(&label("tbl_ok"));
+	&set_label("tbl_break_out",4);	# else %esp -= (p-s)&0xfff + framesz;
 	&sub	($s3,$s0);
 	&and	($s3,0xfff);
-	&add	($s3,64+256);
-	&sub	($key,$s3);
-	&align	(4);
-	&set_label("te_ok");
+	&add	($s3,384);
+	&sub	($acc,$s3);
+	&set_label("tbl_ok",4);
 
-	&mov	($s0,&wparam(0));	# load inp
-	&mov	($s1,&wparam(1));	# load out
-	&mov	($s3,&wparam(3));	# load key
-	&mov	($acc,&wparam(4));	# load ivp
-
-	&exch	("esp",$key);
+	&lea	($s3,&wparam(0));	# obtain pointer to parameter block
+	&exch	("esp",$acc);		# allocate stack frame
 	&add	("esp",4);		# reserve for return address!
-	&mov	($_esp,$key);		# save %esp
+	&mov	($_tbl,$tbl);		# save %ebp
+	&mov	($_esp,$acc);		# save %esp
+
+	&mov	($s0,&DWP(0,$s3));	# load inp
+	&mov	($s1,&DWP(4,$s3));	# load out
+	#&mov	($s2,&DWP(8,$s3));	# load len
+	&mov	($key,&DWP(12,$s3));	# load key
+	&mov	($acc,&DWP(16,$s3));	# load ivp
+	&mov	($s3,&DWP(20,$s3));	# load enc flag
 
 	&mov	($_inp,$s0);		# save copy of inp
 	&mov	($_out,$s1);		# save copy of out
 	&mov	($_len,$s2);		# save copy of len
-	&mov	($_key,$s3);		# save copy of key
+	&mov	($_key,$key);		# save copy of key
 	&mov	($_ivp,$acc);		# save copy of ivp
 
 	&mov	($mark,0);		# copy of aes_key->rounds = 0;
-	if ($compromise) {
-		&cmp	($s2,$compromise);
-		&jb	(&label("skip_ecopy"));
-	}
 	# do we copy key schedule to stack?
-	&mov	($s1 eq "ebx" ? $s1 : "",$s3);
+	&mov	($s1 eq "ebx" ? $s1 : "",$key);
 	&mov	($s2 eq "ecx" ? $s2 : "",244/4);
-	&sub	($s1,"ebp");
-	&mov	("esi",$s3);
+	&sub	($s1,$tbl);
+	&mov	("esi",$key);
 	&and	($s1,0xfff);
 	&lea	("edi",$aes_key);
-	&cmp	($s1,2048);
-	&jb	(&label("do_ecopy"));
+	&cmp	($s1,2048+256);
+	&jb	(&label("do_copy"));
 	&cmp	($s1,4096-244);
-	&jb	(&label("skip_ecopy"));
-	&align	(4);
-	&set_label("do_ecopy");
+	&jb	(&label("skip_copy"));
+	&set_label("do_copy",4);
 		&mov	($_key,"edi");
 		&data_word(0xA5F3F689);	# rep movsd
-	&set_label("skip_ecopy");
+	&set_label("skip_copy");
 
-	&mov	($acc,$s0);
 	&mov	($key,16);
-	&align	(4);
-	&set_label("prefetch_te");
-		&mov	($s0,&DWP(0,"ebp"));
-		&mov	($s1,&DWP(32,"ebp"));
-		&mov	($s2,&DWP(64,"ebp"));
-		&mov	($s3,&DWP(96,"ebp"));
-		&lea	("ebp",&DWP(128,"ebp"));
-		&dec	($key);
-	&jnz	(&label("prefetch_te"));
-	&sub	("ebp",2048);
-
-	&mov	($s2,$_len);
+	&set_label("prefetch_tbl",4);
+		&mov	($s0,&DWP(0,$tbl));
+		&mov	($s1,&DWP(32,$tbl));
+		&mov	($s2,&DWP(64,$tbl));
+		&mov	($acc,&DWP(96,$tbl));
+		&lea	($tbl,&DWP(128,$tbl));
+		&sub	($key,1);
+	&jnz	(&label("prefetch_tbl"));
+	&sub	($tbl,2048);
+
+	&mov	($acc,$_inp);
 	&mov	($key,$_ivp);
-	&test	($s2,0xFFFFFFF0);
-	&jz	(&label("enc_tail"));		# short input...
 
+	&cmp	($s3,0);
+	&je	(&label("fast_decrypt"));
+
+#----------------------------- ENCRYPT -----------------------------#
 	&mov	($s0,&DWP(0,$key));		# load iv
 	&mov	($s1,&DWP(4,$key));
 
-	&align	(4);
-	&set_label("enc_loop");
+	&set_label("fast_enc_loop",16);
 		&mov	($s2,&DWP(8,$key));
 		&mov	($s3,&DWP(12,$key));
 
@@ -916,22 +2158,16 @@ my $mark=&DWP(60+240,"esp");	#copy of aes_key->rounds
 		&mov	(&DWP(8,$key),$s2);
 		&mov	(&DWP(12,$key),$s3);
 
+		&lea	($acc,&DWP(16,$acc));	# advance inp
 		&mov	($s2,$_len);		# load len
-
-		&lea	($acc,&DWP(16,$acc));
 		&mov	($_inp,$acc);		# save inp
-
-		&lea	($s3,&DWP(16,$key));
+		&lea	($s3,&DWP(16,$key));	# advance out
 		&mov	($_out,$s3);		# save out
-
-		&sub	($s2,16);
-		&test	($s2,0xFFFFFFF0);
+		&sub	($s2,16);		# decrease len
 		&mov	($_len,$s2);		# save len
-	&jnz	(&label("enc_loop"));
-	&test	($s2,15);
-	&jnz	(&label("enc_tail"));
+	&jnz	(&label("fast_enc_loop"));
 	&mov	($acc,$_ivp);		# load ivp
-	&mov	($s2,&DWP(8,$key));	# restore last dwords
+	&mov	($s2,&DWP(8,$key));	# restore last 2 dwords
 	&mov	($s3,&DWP(12,$key));
 	&mov	(&DWP(0,$acc),$s0);	# save ivec
 	&mov	(&DWP(4,$acc),$s1);
@@ -949,125 +2185,20 @@ my $mark=&DWP(60+240,"esp");	#copy of aes_key->rounds
 	&set_label("skip_ezero")
 	&mov	("esp",$_esp);
 	&popf	();
-    &set_label("enc_out");
+    &set_label("drop_out");
 	&function_end_A();
 	&pushf	();			# kludge, never executed
 
-    &align	(4);
-    &set_label("enc_tail");
-	&mov	($s0,$key eq "edi" ? $key : "");
-	&mov	($key,$_out);			# load out
-	&push	($s0);				# push ivp
-	&mov	($s1,16);
-	&sub	($s1,$s2);
-	&cmp	($key,$acc);			# compare with inp
-	&je	(&label("enc_in_place"));
-	&align	(4);
-	&data_word(0xA4F3F689);	# rep movsb	# copy input
-	&jmp	(&label("enc_skip_in_place"));
-    &set_label("enc_in_place");
-	&lea	($key,&DWP(0,$key,$s2));
-    &set_label("enc_skip_in_place");
-	&mov	($s2,$s1);
-	&xor	($s0,$s0);
-	&align	(4);
-	&data_word(0xAAF3F689);	# rep stosb	# zero tail
-	&pop	($key);				# pop ivp
-
-	&mov	($acc,$_out);			# output as input
-	&mov	($s0,&DWP(0,$key));
-	&mov	($s1,&DWP(4,$key));
-	&mov	($_len,16);			# len=16
-	&jmp	(&label("enc_loop"));		# one more spin...
-
 #----------------------------- DECRYPT -----------------------------#
-&align	(4);
-&set_label("DECRYPT");
-	&lea    ("ebp",&DWP(&label("AES_Td")."-".&label("pic_point"),"ebp"));
-
-	# allocate aligned stack frame...
-	&lea	($key,&DWP(-64-244,"esp"));
-	&and	($key,-64);
-
-	# ... and make sure it doesn't alias with AES_Td modulo 4096
-	&mov	($s0,"ebp");
-	&lea	($s1,&DWP(2048+256,"ebp"));
-	&mov	($s3,$key);
-	&and	($s0,0xfff);		# s = %ebp&0xfff
-	&and	($s1,0xfff);		# e = (%ebp+2048+256)&0xfff
-	&and	($s3,0xfff);		# p = %esp&0xfff
-
-	&cmp	($s3,$s1);		# if (p>=e) %esp =- (p-e);
-	&jb	(&label("td_break_out"));
-	&sub	($s3,$s1);
-	&sub	($key,$s3);
-	&jmp	(&label("td_ok"));
-	&set_label("td_break_out");	# else %esp -= (p-s)&0xfff + framesz;
-	&sub	($s3,$s0);
-	&and	($s3,0xfff);
-	&add	($s3,64+256);
-	&sub	($key,$s3);
-	&align	(4);
-	&set_label("td_ok");
-
-	&mov	($s0,&wparam(0));	# load inp
-	&mov	($s1,&wparam(1));	# load out
-	&mov	($s3,&wparam(3));	# load key
-	&mov	($acc,&wparam(4));	# load ivp
-
-	&exch	("esp",$key);
-	&add	("esp",4);		# reserve for return address!
-	&mov	($_esp,$key);		# save %esp
-
-	&mov	($_inp,$s0);		# save copy of inp
-	&mov	($_out,$s1);		# save copy of out
-	&mov	($_len,$s2);		# save copy of len
-	&mov	($_key,$s3);		# save copy of key
-	&mov	($_ivp,$acc);		# save copy of ivp
-
-	&mov	($mark,0);		# copy of aes_key->rounds = 0;
-	if ($compromise) {
-		&cmp	($s2,$compromise);
-		&jb	(&label("skip_dcopy"));
-	}
-	# do we copy key schedule to stack?
-	&mov	($s1 eq "ebx" ? $s1 : "",$s3);
-	&mov	($s2 eq "ecx" ? $s2 : "",244/4);
-	&sub	($s1,"ebp");
-	&mov	("esi",$s3);
-	&and	($s1,0xfff);
-	&lea	("edi",$aes_key);
-	&cmp	($s1,2048+256);
-	&jb	(&label("do_dcopy"));
-	&cmp	($s1,4096-244);
-	&jb	(&label("skip_dcopy"));
-	&align	(4);
-	&set_label("do_dcopy");
-		&mov	($_key,"edi");
-		&data_word(0xA5F3F689);	# rep movsd
-	&set_label("skip_dcopy");
-
-	&mov	($acc,$s0);
-	&mov	($key,18);
-	&align	(4);
-	&set_label("prefetch_td");
-		&mov	($s0,&DWP(0,"ebp"));
-		&mov	($s1,&DWP(32,"ebp"));
-		&mov	($s2,&DWP(64,"ebp"));
-		&mov	($s3,&DWP(96,"ebp"));
-		&lea	("ebp",&DWP(128,"ebp"));
-		&dec	($key);
-	&jnz	(&label("prefetch_td"));
-	&sub	("ebp",2048+256);
+&set_label("fast_decrypt",16);
 
 	&cmp	($acc,$_out);
-	&je	(&label("dec_in_place"));	# in-place processing...
+	&je	(&label("fast_dec_in_place"));	# in-place processing...
 
-	&mov	($key,$_ivp);		# load ivp
 	&mov	($_tmp,$key);
 
 	&align	(4);
-	&set_label("dec_loop");
+	&set_label("fast_dec_loop",16);
 		&mov	($s0,&DWP(0,$acc));	# read input
 		&mov	($s1,&DWP(4,$acc));
 		&mov	($s2,&DWP(8,$acc));
@@ -1083,27 +2214,24 @@ my $mark=&DWP(60+240,"esp");	#copy of aes_key->rounds
 		&xor	($s2,&DWP(8,$key));
 		&xor	($s3,&DWP(12,$key));
 
-		&sub	($acc,16);
-		&jc	(&label("dec_partial"));
-		&mov	($_len,$acc);		# save len
-		&mov	($acc,$_inp);		# load inp
 		&mov	($key,$_out);		# load out
+		&mov	($acc,$_inp);		# load inp
 
 		&mov	(&DWP(0,$key),$s0);	# write output
 		&mov	(&DWP(4,$key),$s1);
 		&mov	(&DWP(8,$key),$s2);
 		&mov	(&DWP(12,$key),$s3);
 
+		&mov	($s2,$_len);		# load len
 		&mov	($_tmp,$acc);		# save ivp
-		&lea	($acc,&DWP(16,$acc));
+		&lea	($acc,&DWP(16,$acc));	# advance inp
 		&mov	($_inp,$acc);		# save inp
-
-		&lea	($key,&DWP(16,$key));
+		&lea	($key,&DWP(16,$key));	# advance out
 		&mov	($_out,$key);		# save out
-
-	&jnz	(&label("dec_loop"));
+		&sub	($s2,16);		# decrease len
+		&mov	($_len,$s2);		# save len
+	&jnz	(&label("fast_dec_loop"));
 	&mov	($key,$_tmp);		# load temp ivp
-    &set_label("dec_end");
 	&mov	($acc,$_ivp);		# load user ivp
 	&mov	($s0,&DWP(0,$key));	# load iv
 	&mov	($s1,&DWP(4,$key));
@@ -1113,31 +2241,16 @@ my $mark=&DWP(60+240,"esp");	#copy of aes_key->rounds
 	&mov	(&DWP(4,$acc),$s1);
 	&mov	(&DWP(8,$acc),$s2);
 	&mov	(&DWP(12,$acc),$s3);
-	&jmp	(&label("dec_out"));
+	&jmp	(&label("fast_dec_out"));
 
-    &align	(4);
-    &set_label("dec_partial");
-	&lea	($key,$ivec);
-	&mov	(&DWP(0,$key),$s0);	# dump output to stack
-	&mov	(&DWP(4,$key),$s1);
-	&mov	(&DWP(8,$key),$s2);
-	&mov	(&DWP(12,$key),$s3);
-	&lea	($s2 eq "ecx" ? $s2 : "",&DWP(16,$acc));
-	&mov	($acc eq "esi" ? $acc : "",$key);
-	&mov	($key eq "edi" ? $key : "",$_out);	# load out
-	&data_word(0xA4F3F689);	# rep movsb		# copy output
-	&mov	($key,$_inp);				# use inp as temp ivp
-	&jmp	(&label("dec_end"));
-
-    &align	(4);
-    &set_label("dec_in_place");
-	&set_label("dec_in_place_loop");
-		&lea	($key,$ivec);
+    &set_label("fast_dec_in_place",16);
+	&set_label("fast_dec_in_place_loop");
 		&mov	($s0,&DWP(0,$acc));	# read input
 		&mov	($s1,&DWP(4,$acc));
 		&mov	($s2,&DWP(8,$acc));
 		&mov	($s3,&DWP(12,$acc));
 
+		&lea	($key,$ivec);
 		&mov	(&DWP(0,$key),$s0);	# copy to temp
 		&mov	(&DWP(4,$key),$s1);
 		&mov	(&DWP(8,$key),$s2);
@@ -1158,7 +2271,7 @@ my $mark=&DWP(60+240,"esp");	#copy of aes_key->rounds
 		&mov	(&DWP(8,$acc),$s2);
 		&mov	(&DWP(12,$acc),$s3);
 
-		&lea	($acc,&DWP(16,$acc));
+		&lea	($acc,&DWP(16,$acc));	# advance out
 		&mov	($_out,$acc);		# save out
 
 		&lea	($acc,$ivec);
@@ -1173,40 +2286,340 @@ my $mark=&DWP(60+240,"esp");	#copy of aes_key->rounds
 		&mov	(&DWP(12,$key),$s3);
 
 		&mov	($acc,$_inp);		# load inp
+		&mov	($s2,$_len);		# load len
+		&lea	($acc,&DWP(16,$acc));	# advance inp
+		&mov	($_inp,$acc);		# save inp
+		&sub	($s2,16);		# decrease len
+		&mov	($_len,$s2);		# save len
+	&jnz	(&label("fast_dec_in_place_loop"));
+
+    &set_label("fast_dec_out",4);
+	&cmp	($mark,0);		# was the key schedule copied?
+	&mov	("edi",$_key);
+	&je	(&label("skip_dzero"));
+	# zero copy of key schedule
+	&mov	("ecx",240/4);
+	&xor	("eax","eax");
+	&align	(4);
+	&data_word(0xABF3F689);	# rep stosd
+	&set_label("skip_dzero")
+	&mov	("esp",$_esp);
+	&popf	();
+	&function_end_A();
+	&pushf	();			# kludge, never executed
+
+#--------------------------- SLOW ROUTINE ---------------------------#
+&set_label("slow_way",16);
+
+	&mov	($s0,&DWP(0,$s0)) if (!$x86only);# load OPENSSL_ia32cap
+	&mov	($key,&wparam(3));	# load key
+
+	# pre-allocate aligned stack frame...
+	&lea	($acc,&DWP(-80,"esp"));
+	&and	($acc,-64);
+
+	# ... and make sure it doesn't alias with $key modulo 1024
+	&lea	($s1,&DWP(-80-63,$key));
+	&sub	($s1,$acc);
+	&neg	($s1);
+	&and	($s1,0x3C0);	# modulo 1024, but aligned to cache-line
+	&sub	($acc,$s1);
+
+	# pick S-box copy which can't overlap with stack frame or $key
+	&lea	($s1,&DWP(768,$acc));
+	&sub	($s1,$tbl);
+	&and	($s1,0x300);
+	&lea	($tbl,&DWP(2048+128,$tbl,$s1));
+
+	&lea	($s3,&wparam(0));	# pointer to parameter block
+
+	&exch	("esp",$acc);
+	&add	("esp",4);		# reserve for return address!
+	&mov	($_tbl,$tbl);		# save %ebp
+	&mov	($_esp,$acc);		# save %esp
+	&mov	($_tmp,$s0);		# save OPENSSL_ia32cap
+
+	&mov	($s0,&DWP(0,$s3));	# load inp
+	&mov	($s1,&DWP(4,$s3));	# load out
+	#&mov	($s2,&DWP(8,$s3));	# load len
+	#&mov	($key,&DWP(12,$s3));	# load key
+	&mov	($acc,&DWP(16,$s3));	# load ivp
+	&mov	($s3,&DWP(20,$s3));	# load enc flag
+
+	&mov	($_inp,$s0);		# save copy of inp
+	&mov	($_out,$s1);		# save copy of out
+	&mov	($_len,$s2);		# save copy of len
+	&mov	($_key,$key);		# save copy of key
+	&mov	($_ivp,$acc);		# save copy of ivp
+
+	&mov	($key,$acc);
+	&mov	($acc,$s0);
+
+	&cmp	($s3,0);
+	&je	(&label("slow_decrypt"));
 
-		&lea	($acc,&DWP(16,$acc));
+#--------------------------- SLOW ENCRYPT ---------------------------#
+	&cmp	($s2,16);
+	&mov	($s3,$s1);
+	&jb	(&label("slow_enc_tail"));
+
+					if (!$x86only) {
+	&bt	($_tmp,25);		# check for SSE bit
+	&jnc	(&label("slow_enc_x86"));
+
+	&movq	("mm0",&QWP(0,$key));	# load iv
+	&movq	("mm4",&QWP(8,$key));
+
+	&set_label("slow_enc_loop_sse",16);
+		&pxor	("mm0",&QWP(0,$acc));	# xor input data
+		&pxor	("mm4",&QWP(8,$acc));
+
+		&mov	($key,$_key);
+		&call	("_sse_AES_encrypt_compact");
+
+		&mov	($acc,$_inp);		# load inp
+		&mov	($key,$_out);		# load out
+		&mov	($s2,$_len);		# load len
+
+		&movq	(&QWP(0,$key),"mm0");	# save output data
+		&movq	(&QWP(8,$key),"mm4");
+
+		&lea	($acc,&DWP(16,$acc));	# advance inp
 		&mov	($_inp,$acc);		# save inp
+		&lea	($s3,&DWP(16,$key));	# advance out
+		&mov	($_out,$s3);		# save out
+		&sub	($s2,16);		# decrease len
+		&cmp	($s2,16);
+		&mov	($_len,$s2);		# save len
+	&jae	(&label("slow_enc_loop_sse"));
+	&test	($s2,15);
+	&jnz	(&label("slow_enc_tail"));
+	&mov	($acc,$_ivp);		# load ivp
+	&movq	(&QWP(0,$acc),"mm0");	# save ivec
+	&movq	(&QWP(8,$acc),"mm4");
+	&emms	();
+	&mov	("esp",$_esp);
+	&popf	();
+	&function_end_A();
+	&pushf	();			# kludge, never executed
+					}
+    &set_label("slow_enc_x86",16);
+	&mov	($s0,&DWP(0,$key));	# load iv
+	&mov	($s1,&DWP(4,$key));
+
+	&set_label("slow_enc_loop_x86",4);
+		&mov	($s2,&DWP(8,$key));
+		&mov	($s3,&DWP(12,$key));
+
+		&xor	($s0,&DWP(0,$acc));	# xor input data
+		&xor	($s1,&DWP(4,$acc));
+		&xor	($s2,&DWP(8,$acc));
+		&xor	($s3,&DWP(12,$acc));
+
+		&mov	($key,$_key);		# load key
+		&call	("_x86_AES_encrypt_compact");
+
+		&mov	($acc,$_inp);		# load inp
+		&mov	($key,$_out);		# load out
+
+		&mov	(&DWP(0,$key),$s0);	# save output data
+		&mov	(&DWP(4,$key),$s1);
+		&mov	(&DWP(8,$key),$s2);
+		&mov	(&DWP(12,$key),$s3);
 
 		&mov	($s2,$_len);		# load len
-		&sub	($s2,16);
-		&jc	(&label("dec_in_place_partial"));
+		&lea	($acc,&DWP(16,$acc));	# advance inp
+		&mov	($_inp,$acc);		# save inp
+		&lea	($s3,&DWP(16,$key));	# advance out
+		&mov	($_out,$s3);		# save out
+		&sub	($s2,16);		# decrease len
+		&cmp	($s2,16);
 		&mov	($_len,$s2);		# save len
-	&jnz	(&label("dec_in_place_loop"));
-	&jmp	(&label("dec_out"));
-
-    &align	(4);
-    &set_label("dec_in_place_partial");
-	# one can argue if this is actually required...
-	&mov	($key eq "edi" ? $key : "",$_out);
-	&lea	($acc eq "esi" ? $acc : "",$ivec);
+	&jae	(&label("slow_enc_loop_x86"));
+	&test	($s2,15);
+	&jnz	(&label("slow_enc_tail"));
+	&mov	($acc,$_ivp);		# load ivp
+	&mov	($s2,&DWP(8,$key));	# restore last dwords
+	&mov	($s3,&DWP(12,$key));
+	&mov	(&DWP(0,$acc),$s0);	# save ivec
+	&mov	(&DWP(4,$acc),$s1);
+	&mov	(&DWP(8,$acc),$s2);
+	&mov	(&DWP(12,$acc),$s3);
+
+	&mov	("esp",$_esp);
+	&popf	();
+	&function_end_A();
+	&pushf	();			# kludge, never executed
+
+    &set_label("slow_enc_tail",16);
+	&emms	()	if (!$x86only);
+	&mov	($key eq "edi"? $key:"",$s3);	# load out to edi
+	&mov	($s1,16);
+	&sub	($s1,$s2);
+	&cmp	($key,$acc eq "esi"? $acc:"");	# compare with inp
+	&je	(&label("enc_in_place"));
+	&align	(4);
+	&data_word(0xA4F3F689);	# rep movsb	# copy input
+	&jmp	(&label("enc_skip_in_place"));
+    &set_label("enc_in_place");
 	&lea	($key,&DWP(0,$key,$s2));
-	&lea	($acc,&DWP(16,$acc,$s2));
-	&neg	($s2 eq "ecx" ? $s2 : "");
-	&data_word(0xA4F3F689);	# rep movsb	# restore tail
-
-    &align	(4);
-    &set_label("dec_out");
-    &cmp	($mark,0);		# was the key schedule copied?
-    &mov	("edi",$_key);
-    &je		(&label("skip_dzero"));
-    # zero copy of key schedule
-    &mov	("ecx",240/4);
-    &xor	("eax","eax");
-    &align	(4);
-    &data_word(0xABF3F689);	# rep stosd
-    &set_label("skip_dzero")
-    &mov	("esp",$_esp);
-    &popf	();
+    &set_label("enc_skip_in_place");
+	&mov	($s2,$s1);
+	&xor	($s0,$s0);
+	&align	(4);
+	&data_word(0xAAF3F689);	# rep stosb	# zero tail
+
+	&mov	($key,$_ivp);			# restore ivp
+	&mov	($acc,$s3);			# output as input
+	&mov	($s0,&DWP(0,$key));
+	&mov	($s1,&DWP(4,$key));
+	&mov	($_len,16);			# len=16
+	&jmp	(&label("slow_enc_loop_x86"));	# one more spin...
+
+#--------------------------- SLOW DECRYPT ---------------------------#
+&set_label("slow_decrypt",16);
+					if (!$x86only) {
+	&bt	($_tmp,25);		# check for SSE bit
+	&jnc	(&label("slow_dec_loop_x86"));
+
+	&set_label("slow_dec_loop_sse",4);
+		&movq	("mm0",&QWP(0,$acc));	# read input
+		&movq	("mm4",&QWP(8,$acc));
+
+		&mov	($key,$_key);
+		&call	("_sse_AES_decrypt_compact");
+
+		&mov	($acc,$_inp);		# load inp
+		&lea	($s0,$ivec);
+		&mov	($s1,$_out);		# load out
+		&mov	($s2,$_len);		# load len
+		&mov	($key,$_ivp);		# load ivp
+
+		&movq	("mm1",&QWP(0,$acc));	# re-read input
+		&movq	("mm5",&QWP(8,$acc));
+
+		&pxor	("mm0",&QWP(0,$key));	# xor iv
+		&pxor	("mm4",&QWP(8,$key));
+
+		&movq	(&QWP(0,$key),"mm1");	# copy input to iv
+		&movq	(&QWP(8,$key),"mm5");
+
+		&sub	($s2,16);		# decrease len
+		&jc	(&label("slow_dec_partial_sse"));
+
+		&movq	(&QWP(0,$s1),"mm0");	# write output
+		&movq	(&QWP(8,$s1),"mm4");
+
+		&lea	($s1,&DWP(16,$s1));	# advance out
+		&mov	($_out,$s1);		# save out
+		&lea	($acc,&DWP(16,$acc));	# advance inp
+		&mov	($_inp,$acc);		# save inp
+		&mov	($_len,$s2);		# save len
+	&jnz	(&label("slow_dec_loop_sse"));
+	&emms	();
+	&mov	("esp",$_esp);
+	&popf	();
+	&function_end_A();
+	&pushf	();			# kludge, never executed
+
+    &set_label("slow_dec_partial_sse",16);
+	&movq	(&QWP(0,$s0),"mm0");	# save output to temp
+	&movq	(&QWP(8,$s0),"mm4");
+	&emms	();
+
+	&add	($s2 eq "ecx" ? "ecx":"",16);
+	&mov	("edi",$s1);		# out
+	&mov	("esi",$s0);		# temp
+	&align	(4);
+	&data_word(0xA4F3F689);		# rep movsb # copy partial output
+
+	&mov	("esp",$_esp);
+	&popf	();
+	&function_end_A();
+	&pushf	();			# kludge, never executed
+					}
+	&set_label("slow_dec_loop_x86",16);
+		&mov	($s0,&DWP(0,$acc));	# read input
+		&mov	($s1,&DWP(4,$acc));
+		&mov	($s2,&DWP(8,$acc));
+		&mov	($s3,&DWP(12,$acc));
+
+		&lea	($key,$ivec);
+		&mov	(&DWP(0,$key),$s0);	# copy to temp
+		&mov	(&DWP(4,$key),$s1);
+		&mov	(&DWP(8,$key),$s2);
+		&mov	(&DWP(12,$key),$s3);
+
+		&mov	($key,$_key);		# load key
+		&call	("_x86_AES_decrypt_compact");
+
+		&mov	($key,$_ivp);		# load ivp
+		&mov	($acc,$_len);		# load len
+		&xor	($s0,&DWP(0,$key));	# xor iv
+		&xor	($s1,&DWP(4,$key));
+		&xor	($s2,&DWP(8,$key));
+		&xor	($s3,&DWP(12,$key));
+
+		&sub	($acc,16);
+		&jc	(&label("slow_dec_partial_x86"));
+
+		&mov	($_len,$acc);		# save len
+		&mov	($acc,$_out);		# load out
+
+		&mov	(&DWP(0,$acc),$s0);	# write output
+		&mov	(&DWP(4,$acc),$s1);
+		&mov	(&DWP(8,$acc),$s2);
+		&mov	(&DWP(12,$acc),$s3);
+
+		&lea	($acc,&DWP(16,$acc));	# advance out
+		&mov	($_out,$acc);		# save out
+
+		&lea	($acc,$ivec);
+		&mov	($s0,&DWP(0,$acc));	# read temp
+		&mov	($s1,&DWP(4,$acc));
+		&mov	($s2,&DWP(8,$acc));
+		&mov	($s3,&DWP(12,$acc));
+
+		&mov	(&DWP(0,$key),$s0);	# copy it to iv
+		&mov	(&DWP(4,$key),$s1);
+		&mov	(&DWP(8,$key),$s2);
+		&mov	(&DWP(12,$key),$s3);
+
+		&mov	($acc,$_inp);		# load inp
+		&lea	($acc,&DWP(16,$acc));	# advance inp
+		&mov	($_inp,$acc);		# save inp
+	&jnz	(&label("slow_dec_loop_x86"));
+	&mov	("esp",$_esp);
+	&popf	();
+	&function_end_A();
+	&pushf	();			# kludge, never executed
+
+    &set_label("slow_dec_partial_x86",16);
+	&lea	($acc,$ivec);
+	&mov	(&DWP(0,$acc),$s0);	# save output to temp
+	&mov	(&DWP(4,$acc),$s1);
+	&mov	(&DWP(8,$acc),$s2);
+	&mov	(&DWP(12,$acc),$s3);
+
+	&mov	($acc,$_inp);
+	&mov	($s0,&DWP(0,$acc));	# re-read input
+	&mov	($s1,&DWP(4,$acc));
+	&mov	($s2,&DWP(8,$acc));
+	&mov	($s3,&DWP(12,$acc));
+
+	&mov	(&DWP(0,$key),$s0);	# copy it to iv
+	&mov	(&DWP(4,$key),$s1);
+	&mov	(&DWP(8,$key),$s2);
+	&mov	(&DWP(12,$key),$s3);
+
+	&mov	("ecx",$_len);
+	&mov	("edi",$_out);
+	&lea	("esi",$ivec);
+	&align	(4);
+	&data_word(0xA4F3F689);		# rep movsb # copy partial output
+
+	&mov	("esp",$_esp);
+	&popf	();
 &function_end("AES_cbc_encrypt");
 }
 
@@ -1215,35 +2628,31 @@ my $mark=&DWP(60+240,"esp");	#copy of aes_key->rounds
 sub enckey()
 {
 	&movz	("esi",&LB("edx"));		# rk[i]>>0
-	&mov	("ebx",&DWP(2,"ebp","esi",8));
+	&movz	("ebx",&BP(-128,$tbl,"esi",1));
 	&movz	("esi",&HB("edx"));		# rk[i]>>8
-	&and	("ebx",0xFF000000);
+	&shl	("ebx",24);
 	&xor	("eax","ebx");
 
-	&mov	("ebx",&DWP(2,"ebp","esi",8));
+	&movz	("ebx",&BP(-128,$tbl,"esi",1));
 	&shr	("edx",16);
-	&and	("ebx",0x000000FF);
 	&movz	("esi",&LB("edx"));		# rk[i]>>16
 	&xor	("eax","ebx");
 
-	&mov	("ebx",&DWP(0,"ebp","esi",8));
+	&movz	("ebx",&BP(-128,$tbl,"esi",1));
 	&movz	("esi",&HB("edx"));		# rk[i]>>24
-	&and	("ebx",0x0000FF00);
+	&shl	("ebx",8);
 	&xor	("eax","ebx");
 
-	&mov	("ebx",&DWP(0,"ebp","esi",8));
-	&and	("ebx",0x00FF0000);
+	&movz	("ebx",&BP(-128,$tbl,"esi",1));
+	&shl	("ebx",16);
 	&xor	("eax","ebx");
 
-	&xor	("eax",&DWP(2048,"ebp","ecx",4));	# rcon
+	&xor	("eax",&DWP(1024-128,$tbl,"ecx",4));	# rcon
 }
 
-# int AES_set_encrypt_key(const unsigned char *userKey, const int bits,
-#                        AES_KEY *key)
-&public_label("AES_Te");
-&function_begin("AES_set_encrypt_key");
-	&mov	("esi",&wparam(0));		# user supplied key
-	&mov	("edi",&wparam(2));		# private key schedule
+&function_begin("_x86_AES_set_encrypt_key");
+	&mov	("esi",&wparam(1));		# user supplied key
+	&mov	("edi",&wparam(3));		# private key schedule
 
 	&test	("esi",-1);
 	&jz	(&label("badpointer"));
@@ -1252,10 +2661,21 @@ sub enckey()
 
 	&call	(&label("pic_point"));
 	&set_label("pic_point");
-	&blindpop("ebp");
-	&lea	("ebp",&DWP(&label("AES_Te")."-".&label("pic_point"),"ebp"));
-
-	&mov	("ecx",&wparam(1));		# number of bits in key
+	&blindpop($tbl);
+	&lea	($tbl,&DWP(&label("AES_Te")."-".&label("pic_point"),$tbl));
+	&lea	($tbl,&DWP(2048+128,$tbl));
+
+	# prefetch Te4
+	&mov	("eax",&DWP(0-128,$tbl));
+	&mov	("ebx",&DWP(32-128,$tbl));
+	&mov	("ecx",&DWP(64-128,$tbl));
+	&mov	("edx",&DWP(96-128,$tbl));
+	&mov	("eax",&DWP(128-128,$tbl));
+	&mov	("ebx",&DWP(160-128,$tbl));
+	&mov	("ecx",&DWP(192-128,$tbl));
+	&mov	("edx",&DWP(224-128,$tbl));
+
+	&mov	("ecx",&wparam(2));		# number of bits in key
 	&cmp	("ecx",128);
 	&je	(&label("10rounds"));
 	&cmp	("ecx",192);
@@ -1394,24 +2814,23 @@ sub enckey()
 		&mov	("edx","eax");
 		&mov	("eax",&DWP(16,"edi"));		# rk[4]
 		&movz	("esi",&LB("edx"));		# rk[11]>>0
-		&mov	("ebx",&DWP(2,"ebp","esi",8));
+		&movz	("ebx",&BP(-128,$tbl,"esi",1));
 		&movz	("esi",&HB("edx"));		# rk[11]>>8
-		&and	("ebx",0x000000FF);
 		&xor	("eax","ebx");
 
-		&mov	("ebx",&DWP(0,"ebp","esi",8));
+		&movz	("ebx",&BP(-128,$tbl,"esi",1));
 		&shr	("edx",16);
-		&and	("ebx",0x0000FF00);
+		&shl	("ebx",8);
 		&movz	("esi",&LB("edx"));		# rk[11]>>16
 		&xor	("eax","ebx");
 
-		&mov	("ebx",&DWP(0,"ebp","esi",8));
+		&movz	("ebx",&BP(-128,$tbl,"esi",1));
 		&movz	("esi",&HB("edx"));		# rk[11]>>24
-		&and	("ebx",0x00FF0000);
+		&shl	("ebx",16);
 		&xor	("eax","ebx");
 
-		&mov	("ebx",&DWP(2,"ebp","esi",8));
-		&and	("ebx",0xFF000000);
+		&movz	("ebx",&BP(-128,$tbl,"esi",1));
+		&shl	("ebx",24);
 		&xor	("eax","ebx");
 
 		&mov	(&DWP(48,"edi"),"eax");		# rk[12]
@@ -1433,43 +2852,74 @@ sub enckey()
     &set_label("badpointer");
 	&mov	("eax",-1);
     &set_label("exit");
-&function_end("AES_set_encrypt_key");
+&function_end("_x86_AES_set_encrypt_key");
 
-sub deckey()
-{ my ($i,$ptr,$te,$td) = @_;
+# int private_AES_set_encrypt_key(const unsigned char *userKey, const int bits,
+#                        AES_KEY *key)
+&function_begin_B("private_AES_set_encrypt_key");
+	&call	("_x86_AES_set_encrypt_key");
+	&ret	();
+&function_end_B("private_AES_set_encrypt_key");
 
-	&mov	("eax",&DWP($i,$ptr));
-	&mov	("edx","eax");
-	&movz	("ebx",&HB("eax"));
-	&shr	("edx",16);
-	&and	("eax",0xFF);
-	&movz	("eax",&BP(2,$te,"eax",8));
-	&movz	("ebx",&BP(2,$te,"ebx",8));
-	&mov	("eax",&DWP(0,$td,"eax",8));
-	&xor	("eax",&DWP(3,$td,"ebx",8));
-	&movz	("ebx",&HB("edx"));
-	&and	("edx",0xFF);
-	&movz	("edx",&BP(2,$te,"edx",8));
-	&movz	("ebx",&BP(2,$te,"ebx",8));
-	&xor	("eax",&DWP(2,$td,"edx",8));
-	&xor	("eax",&DWP(1,$td,"ebx",8));
-	&mov	(&DWP($i,$ptr),"eax");
+sub deckey()
+{ my ($i,$key,$tp1,$tp2,$tp4,$tp8) = @_;
+  my $tmp = $tbl;
+
+	&mov	($acc,$tp1);
+	&and	($acc,0x80808080);
+	&mov	($tmp,$acc);
+	&shr	($tmp,7);
+	&lea	($tp2,&DWP(0,$tp1,$tp1));
+	&sub	($acc,$tmp);
+	&and	($tp2,0xfefefefe);
+	&and	($acc,0x1b1b1b1b);
+	&xor	($acc,$tp2);
+	&mov	($tp2,$acc);
+
+	&and	($acc,0x80808080);
+	&mov	($tmp,$acc);
+	&shr	($tmp,7);
+	&lea	($tp4,&DWP(0,$tp2,$tp2));
+	&sub	($acc,$tmp);
+	&and	($tp4,0xfefefefe);
+	&and	($acc,0x1b1b1b1b);
+	 &xor	($tp2,$tp1);	# tp2^tp1
+	&xor	($acc,$tp4);
+	&mov	($tp4,$acc);
+
+	&and	($acc,0x80808080);
+	&mov	($tmp,$acc);
+	&shr	($tmp,7);
+	&lea	($tp8,&DWP(0,$tp4,$tp4));
+	 &xor	($tp4,$tp1);	# tp4^tp1
+	&sub	($acc,$tmp);
+	&and	($tp8,0xfefefefe);
+	&and	($acc,0x1b1b1b1b);
+	 &rotl	($tp1,8);	# = ROTATE(tp1,8)
+	&xor	($tp8,$acc);
+
+	&mov	($tmp,&DWP(4*($i+1),$key));	# modulo-scheduled load
+
+	&xor	($tp1,$tp2);
+	&xor	($tp2,$tp8);
+	&xor	($tp1,$tp4);
+	&rotl	($tp2,24);
+	&xor	($tp4,$tp8);
+	&xor	($tp1,$tp8);	# ^= tp8^(tp4^tp1)^(tp2^tp1)
+	&rotl	($tp4,16);
+	&xor	($tp1,$tp2);	# ^= ROTATE(tp8^tp2^tp1,24)
+	&rotl	($tp8,8);
+	&xor	($tp1,$tp4);	# ^= ROTATE(tp8^tp4^tp1,16)
+	&mov	($tp2,$tmp);
+	&xor	($tp1,$tp8);	# ^= ROTATE(tp8,8)
+
+	&mov	(&DWP(4*$i,$key),$tp1);
 }
 
-# int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
+# int private_AES_set_decrypt_key(const unsigned char *userKey, const int bits,
 #                        AES_KEY *key)
-&public_label("AES_Td");
-&public_label("AES_Te");
-&function_begin_B("AES_set_decrypt_key");
-	&mov	("eax",&wparam(0));
-	&mov	("ecx",&wparam(1));
-	&mov	("edx",&wparam(2));
-	&sub	("esp",12);
-	&mov	(&DWP(0,"esp"),"eax");
-	&mov	(&DWP(4,"esp"),"ecx");
-	&mov	(&DWP(8,"esp"),"edx");
-	&call	("AES_set_encrypt_key");
-	&add	("esp",12);
+&function_begin_B("private_AES_set_decrypt_key");
+	&call	("_x86_AES_set_encrypt_key");
 	&cmp	("eax",0);
 	&je	(&label("proceed"));
 	&ret	();
@@ -1485,8 +2935,7 @@ sub deckey()
 	&lea	("ecx",&DWP(0,"","ecx",4));
 	&lea	("edi",&DWP(0,"esi","ecx",4));	# pointer to last chunk
 
-	&align	(4);
-	&set_label("invert");			# invert order of chunks
+	&set_label("invert",4);			# invert order of chunks
 		&mov	("eax",&DWP(0,"esi"));
 		&mov	("ebx",&DWP(4,"esi"));
 		&mov	("ecx",&DWP(0,"edi"));
@@ -1508,26 +2957,24 @@ sub deckey()
 		&cmp	("esi","edi");
 	&jne	(&label("invert"));
 
-	&call	(&label("pic_point"));
-	&set_label("pic_point");
-	blindpop("ebp");
-	&lea	("edi",&DWP(&label("AES_Td")."-".&label("pic_point"),"ebp"));
-	&lea	("ebp",&DWP(&label("AES_Te")."-".&label("pic_point"),"ebp"));
-
-	&mov	("esi",&wparam(2));
-	&mov	("ecx",&DWP(240,"esi"));	# pull number of rounds
-	&dec	("ecx");
-	&align	(4);
-	&set_label("permute");			# permute the key schedule
-		&add	("esi",16);
-		&deckey	(0,"esi","ebp","edi");
-		&deckey	(4,"esi","ebp","edi");
-		&deckey	(8,"esi","ebp","edi");
-		&deckey	(12,"esi","ebp","edi");
-		&dec	("ecx");
-	&jnz	(&label("permute"));
+	&mov	($key,&wparam(2));
+	&mov	($acc,&DWP(240,$key));		# pull number of rounds
+	&lea	($acc,&DWP(-2,$acc,$acc));
+	&lea	($acc,&DWP(0,$key,$acc,8));
+	&mov	(&wparam(2),$acc);
+
+	&mov	($s0,&DWP(16,$key));		# modulo-scheduled load
+	&set_label("permute",4);		# permute the key schedule
+		&add	($key,16);
+		&deckey	(0,$key,$s0,$s1,$s2,$s3);
+		&deckey	(1,$key,$s1,$s2,$s3,$s0);
+		&deckey	(2,$key,$s2,$s3,$s0,$s1);
+		&deckey	(3,$key,$s3,$s0,$s1,$s2);
+		&cmp	($key,&wparam(2));
+	&jb	(&label("permute"));
 
 	&xor	("eax","eax");			# return success
-&function_end("AES_set_decrypt_key");
+&function_end("private_AES_set_decrypt_key");
+&asciz("AES for x86, CRYPTOGAMS by <appro\@openssl.org>");
 
 &asm_finish();
diff --git a/crypto/aes/asm/aes-armv4.pl b/crypto/aes/asm/aes-armv4.pl
new file mode 100755
index 000000000000..86b86c4a0fbd
--- /dev/null
+++ b/crypto/aes/asm/aes-armv4.pl
@@ -0,0 +1,1134 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# AES for ARMv4
+
+# January 2007.
+#
+# Code uses single 1K S-box and is >2 times faster than code generated
+# by gcc-3.4.1. This is thanks to unique feature of ARMv4 ISA, which
+# allows to merge logical or arithmetic operation with shift or rotate
+# in one instruction and emit combined result every cycle. The module
+# is endian-neutral. The performance is ~42 cycles/byte for 128-bit
+# key [on single-issue Xscale PXA250 core].
+
+# May 2007.
+#
+# AES_set_[en|de]crypt_key is added.
+
+# July 2010.
+#
+# Rescheduling for dual-issue pipeline resulted in 12% improvement on
+# Cortex A8 core and ~25 cycles per byte processed with 128-bit key.
+
+# February 2011.
+#
+# Profiler-assisted and platform-specific optimization resulted in 16%
+# improvement on Cortex A8 core and ~21.5 cycles per byte.
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$s0="r0";
+$s1="r1";
+$s2="r2";
+$s3="r3";
+$t1="r4";
+$t2="r5";
+$t3="r6";
+$i1="r7";
+$i2="r8";
+$i3="r9";
+
+$tbl="r10";
+$key="r11";
+$rounds="r12";
+
+$code=<<___;
+#include "arm_arch.h"
+.text
+.code	32
+
+.type	AES_Te,%object
+.align	5
+AES_Te:
+.word	0xc66363a5, 0xf87c7c84, 0xee777799, 0xf67b7b8d
+.word	0xfff2f20d, 0xd66b6bbd, 0xde6f6fb1, 0x91c5c554
+.word	0x60303050, 0x02010103, 0xce6767a9, 0x562b2b7d
+.word	0xe7fefe19, 0xb5d7d762, 0x4dababe6, 0xec76769a
+.word	0x8fcaca45, 0x1f82829d, 0x89c9c940, 0xfa7d7d87
+.word	0xeffafa15, 0xb25959eb, 0x8e4747c9, 0xfbf0f00b
+.word	0x41adadec, 0xb3d4d467, 0x5fa2a2fd, 0x45afafea
+.word	0x239c9cbf, 0x53a4a4f7, 0xe4727296, 0x9bc0c05b
+.word	0x75b7b7c2, 0xe1fdfd1c, 0x3d9393ae, 0x4c26266a
+.word	0x6c36365a, 0x7e3f3f41, 0xf5f7f702, 0x83cccc4f
+.word	0x6834345c, 0x51a5a5f4, 0xd1e5e534, 0xf9f1f108
+.word	0xe2717193, 0xabd8d873, 0x62313153, 0x2a15153f
+.word	0x0804040c, 0x95c7c752, 0x46232365, 0x9dc3c35e
+.word	0x30181828, 0x379696a1, 0x0a05050f, 0x2f9a9ab5
+.word	0x0e070709, 0x24121236, 0x1b80809b, 0xdfe2e23d
+.word	0xcdebeb26, 0x4e272769, 0x7fb2b2cd, 0xea75759f
+.word	0x1209091b, 0x1d83839e, 0x582c2c74, 0x341a1a2e
+.word	0x361b1b2d, 0xdc6e6eb2, 0xb45a5aee, 0x5ba0a0fb
+.word	0xa45252f6, 0x763b3b4d, 0xb7d6d661, 0x7db3b3ce
+.word	0x5229297b, 0xdde3e33e, 0x5e2f2f71, 0x13848497
+.word	0xa65353f5, 0xb9d1d168, 0x00000000, 0xc1eded2c
+.word	0x40202060, 0xe3fcfc1f, 0x79b1b1c8, 0xb65b5bed
+.word	0xd46a6abe, 0x8dcbcb46, 0x67bebed9, 0x7239394b
+.word	0x944a4ade, 0x984c4cd4, 0xb05858e8, 0x85cfcf4a
+.word	0xbbd0d06b, 0xc5efef2a, 0x4faaaae5, 0xedfbfb16
+.word	0x864343c5, 0x9a4d4dd7, 0x66333355, 0x11858594
+.word	0x8a4545cf, 0xe9f9f910, 0x04020206, 0xfe7f7f81
+.word	0xa05050f0, 0x783c3c44, 0x259f9fba, 0x4ba8a8e3
+.word	0xa25151f3, 0x5da3a3fe, 0x804040c0, 0x058f8f8a
+.word	0x3f9292ad, 0x219d9dbc, 0x70383848, 0xf1f5f504
+.word	0x63bcbcdf, 0x77b6b6c1, 0xafdada75, 0x42212163
+.word	0x20101030, 0xe5ffff1a, 0xfdf3f30e, 0xbfd2d26d
+.word	0x81cdcd4c, 0x180c0c14, 0x26131335, 0xc3ecec2f
+.word	0xbe5f5fe1, 0x359797a2, 0x884444cc, 0x2e171739
+.word	0x93c4c457, 0x55a7a7f2, 0xfc7e7e82, 0x7a3d3d47
+.word	0xc86464ac, 0xba5d5de7, 0x3219192b, 0xe6737395
+.word	0xc06060a0, 0x19818198, 0x9e4f4fd1, 0xa3dcdc7f
+.word	0x44222266, 0x542a2a7e, 0x3b9090ab, 0x0b888883
+.word	0x8c4646ca, 0xc7eeee29, 0x6bb8b8d3, 0x2814143c
+.word	0xa7dede79, 0xbc5e5ee2, 0x160b0b1d, 0xaddbdb76
+.word	0xdbe0e03b, 0x64323256, 0x743a3a4e, 0x140a0a1e
+.word	0x924949db, 0x0c06060a, 0x4824246c, 0xb85c5ce4
+.word	0x9fc2c25d, 0xbdd3d36e, 0x43acacef, 0xc46262a6
+.word	0x399191a8, 0x319595a4, 0xd3e4e437, 0xf279798b
+.word	0xd5e7e732, 0x8bc8c843, 0x6e373759, 0xda6d6db7
+.word	0x018d8d8c, 0xb1d5d564, 0x9c4e4ed2, 0x49a9a9e0
+.word	0xd86c6cb4, 0xac5656fa, 0xf3f4f407, 0xcfeaea25
+.word	0xca6565af, 0xf47a7a8e, 0x47aeaee9, 0x10080818
+.word	0x6fbabad5, 0xf0787888, 0x4a25256f, 0x5c2e2e72
+.word	0x381c1c24, 0x57a6a6f1, 0x73b4b4c7, 0x97c6c651
+.word	0xcbe8e823, 0xa1dddd7c, 0xe874749c, 0x3e1f1f21
+.word	0x964b4bdd, 0x61bdbddc, 0x0d8b8b86, 0x0f8a8a85
+.word	0xe0707090, 0x7c3e3e42, 0x71b5b5c4, 0xcc6666aa
+.word	0x904848d8, 0x06030305, 0xf7f6f601, 0x1c0e0e12
+.word	0xc26161a3, 0x6a35355f, 0xae5757f9, 0x69b9b9d0
+.word	0x17868691, 0x99c1c158, 0x3a1d1d27, 0x279e9eb9
+.word	0xd9e1e138, 0xebf8f813, 0x2b9898b3, 0x22111133
+.word	0xd26969bb, 0xa9d9d970, 0x078e8e89, 0x339494a7
+.word	0x2d9b9bb6, 0x3c1e1e22, 0x15878792, 0xc9e9e920
+.word	0x87cece49, 0xaa5555ff, 0x50282878, 0xa5dfdf7a
+.word	0x038c8c8f, 0x59a1a1f8, 0x09898980, 0x1a0d0d17
+.word	0x65bfbfda, 0xd7e6e631, 0x844242c6, 0xd06868b8
+.word	0x824141c3, 0x299999b0, 0x5a2d2d77, 0x1e0f0f11
+.word	0x7bb0b0cb, 0xa85454fc, 0x6dbbbbd6, 0x2c16163a
+@ Te4[256]
+.byte	0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5
+.byte	0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76
+.byte	0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0
+.byte	0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0
+.byte	0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc
+.byte	0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15
+.byte	0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a
+.byte	0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75
+.byte	0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0
+.byte	0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84
+.byte	0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b
+.byte	0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf
+.byte	0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85
+.byte	0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8
+.byte	0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5
+.byte	0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2
+.byte	0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17
+.byte	0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73
+.byte	0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88
+.byte	0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb
+.byte	0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c
+.byte	0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79
+.byte	0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9
+.byte	0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08
+.byte	0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6
+.byte	0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a
+.byte	0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e
+.byte	0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e
+.byte	0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94
+.byte	0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf
+.byte	0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68
+.byte	0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16
+@ rcon[]
+.word	0x01000000, 0x02000000, 0x04000000, 0x08000000
+.word	0x10000000, 0x20000000, 0x40000000, 0x80000000
+.word	0x1B000000, 0x36000000, 0, 0, 0, 0, 0, 0
+.size	AES_Te,.-AES_Te
+
+@ void AES_encrypt(const unsigned char *in, unsigned char *out,
+@ 		 const AES_KEY *key) {
+.global AES_encrypt
+.type   AES_encrypt,%function
+.align	5
+AES_encrypt:
+	sub	r3,pc,#8		@ AES_encrypt
+	stmdb   sp!,{r1,r4-r12,lr}
+	mov	$rounds,r0		@ inp
+	mov	$key,r2
+	sub	$tbl,r3,#AES_encrypt-AES_Te	@ Te
+#if __ARM_ARCH__<7
+	ldrb	$s0,[$rounds,#3]	@ load input data in endian-neutral
+	ldrb	$t1,[$rounds,#2]	@ manner...
+	ldrb	$t2,[$rounds,#1]
+	ldrb	$t3,[$rounds,#0]
+	orr	$s0,$s0,$t1,lsl#8
+	ldrb	$s1,[$rounds,#7]
+	orr	$s0,$s0,$t2,lsl#16
+	ldrb	$t1,[$rounds,#6]
+	orr	$s0,$s0,$t3,lsl#24
+	ldrb	$t2,[$rounds,#5]
+	ldrb	$t3,[$rounds,#4]
+	orr	$s1,$s1,$t1,lsl#8
+	ldrb	$s2,[$rounds,#11]
+	orr	$s1,$s1,$t2,lsl#16
+	ldrb	$t1,[$rounds,#10]
+	orr	$s1,$s1,$t3,lsl#24
+	ldrb	$t2,[$rounds,#9]
+	ldrb	$t3,[$rounds,#8]
+	orr	$s2,$s2,$t1,lsl#8
+	ldrb	$s3,[$rounds,#15]
+	orr	$s2,$s2,$t2,lsl#16
+	ldrb	$t1,[$rounds,#14]
+	orr	$s2,$s2,$t3,lsl#24
+	ldrb	$t2,[$rounds,#13]
+	ldrb	$t3,[$rounds,#12]
+	orr	$s3,$s3,$t1,lsl#8
+	orr	$s3,$s3,$t2,lsl#16
+	orr	$s3,$s3,$t3,lsl#24
+#else
+	ldr	$s0,[$rounds,#0]
+	ldr	$s1,[$rounds,#4]
+	ldr	$s2,[$rounds,#8]
+	ldr	$s3,[$rounds,#12]
+#ifdef __ARMEL__
+	rev	$s0,$s0
+	rev	$s1,$s1
+	rev	$s2,$s2
+	rev	$s3,$s3
+#endif
+#endif
+	bl	_armv4_AES_encrypt
+
+	ldr	$rounds,[sp],#4		@ pop out
+#if __ARM_ARCH__>=7
+#ifdef __ARMEL__
+	rev	$s0,$s0
+	rev	$s1,$s1
+	rev	$s2,$s2
+	rev	$s3,$s3
+#endif
+	str	$s0,[$rounds,#0]
+	str	$s1,[$rounds,#4]
+	str	$s2,[$rounds,#8]
+	str	$s3,[$rounds,#12]
+#else
+	mov	$t1,$s0,lsr#24		@ write output in endian-neutral
+	mov	$t2,$s0,lsr#16		@ manner...
+	mov	$t3,$s0,lsr#8
+	strb	$t1,[$rounds,#0]
+	strb	$t2,[$rounds,#1]
+	mov	$t1,$s1,lsr#24
+	strb	$t3,[$rounds,#2]
+	mov	$t2,$s1,lsr#16
+	strb	$s0,[$rounds,#3]
+	mov	$t3,$s1,lsr#8
+	strb	$t1,[$rounds,#4]
+	strb	$t2,[$rounds,#5]
+	mov	$t1,$s2,lsr#24
+	strb	$t3,[$rounds,#6]
+	mov	$t2,$s2,lsr#16
+	strb	$s1,[$rounds,#7]
+	mov	$t3,$s2,lsr#8
+	strb	$t1,[$rounds,#8]
+	strb	$t2,[$rounds,#9]
+	mov	$t1,$s3,lsr#24
+	strb	$t3,[$rounds,#10]
+	mov	$t2,$s3,lsr#16
+	strb	$s2,[$rounds,#11]
+	mov	$t3,$s3,lsr#8
+	strb	$t1,[$rounds,#12]
+	strb	$t2,[$rounds,#13]
+	strb	$t3,[$rounds,#14]
+	strb	$s3,[$rounds,#15]
+#endif
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r12,pc}
+#else
+	ldmia   sp!,{r4-r12,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.size	AES_encrypt,.-AES_encrypt
+
+.type   _armv4_AES_encrypt,%function
+.align	2
+_armv4_AES_encrypt:
+	str	lr,[sp,#-4]!		@ push lr
+	ldmia	$key!,{$t1-$i1}
+	eor	$s0,$s0,$t1
+	ldr	$rounds,[$key,#240-16]
+	eor	$s1,$s1,$t2
+	eor	$s2,$s2,$t3
+	eor	$s3,$s3,$i1
+	sub	$rounds,$rounds,#1
+	mov	lr,#255
+
+	and	$i1,lr,$s0
+	and	$i2,lr,$s0,lsr#8
+	and	$i3,lr,$s0,lsr#16
+	mov	$s0,$s0,lsr#24
+.Lenc_loop:
+	ldr	$t1,[$tbl,$i1,lsl#2]	@ Te3[s0>>0]
+	and	$i1,lr,$s1,lsr#16	@ i0
+	ldr	$t2,[$tbl,$i2,lsl#2]	@ Te2[s0>>8]
+	and	$i2,lr,$s1
+	ldr	$t3,[$tbl,$i3,lsl#2]	@ Te1[s0>>16]
+	and	$i3,lr,$s1,lsr#8
+	ldr	$s0,[$tbl,$s0,lsl#2]	@ Te0[s0>>24]
+	mov	$s1,$s1,lsr#24
+
+	ldr	$i1,[$tbl,$i1,lsl#2]	@ Te1[s1>>16]
+	ldr	$i2,[$tbl,$i2,lsl#2]	@ Te3[s1>>0]
+	ldr	$i3,[$tbl,$i3,lsl#2]	@ Te2[s1>>8]
+	eor	$s0,$s0,$i1,ror#8
+	ldr	$s1,[$tbl,$s1,lsl#2]	@ Te0[s1>>24]
+	and	$i1,lr,$s2,lsr#8	@ i0
+	eor	$t2,$t2,$i2,ror#8
+	and	$i2,lr,$s2,lsr#16	@ i1
+	eor	$t3,$t3,$i3,ror#8
+	and	$i3,lr,$s2
+	ldr	$i1,[$tbl,$i1,lsl#2]	@ Te2[s2>>8]
+	eor	$s1,$s1,$t1,ror#24
+	ldr	$i2,[$tbl,$i2,lsl#2]	@ Te1[s2>>16]
+	mov	$s2,$s2,lsr#24
+
+	ldr	$i3,[$tbl,$i3,lsl#2]	@ Te3[s2>>0]
+	eor	$s0,$s0,$i1,ror#16
+	ldr	$s2,[$tbl,$s2,lsl#2]	@ Te0[s2>>24]
+	and	$i1,lr,$s3		@ i0
+	eor	$s1,$s1,$i2,ror#8
+	and	$i2,lr,$s3,lsr#8	@ i1
+	eor	$t3,$t3,$i3,ror#16
+	and	$i3,lr,$s3,lsr#16	@ i2
+	ldr	$i1,[$tbl,$i1,lsl#2]	@ Te3[s3>>0]
+	eor	$s2,$s2,$t2,ror#16
+	ldr	$i2,[$tbl,$i2,lsl#2]	@ Te2[s3>>8]
+	mov	$s3,$s3,lsr#24
+
+	ldr	$i3,[$tbl,$i3,lsl#2]	@ Te1[s3>>16]
+	eor	$s0,$s0,$i1,ror#24
+	ldr	$i1,[$key],#16
+	eor	$s1,$s1,$i2,ror#16
+	ldr	$s3,[$tbl,$s3,lsl#2]	@ Te0[s3>>24]
+	eor	$s2,$s2,$i3,ror#8
+	ldr	$t1,[$key,#-12]
+	eor	$s3,$s3,$t3,ror#8
+
+	ldr	$t2,[$key,#-8]
+	eor	$s0,$s0,$i1
+	ldr	$t3,[$key,#-4]
+	and	$i1,lr,$s0
+	eor	$s1,$s1,$t1
+	and	$i2,lr,$s0,lsr#8
+	eor	$s2,$s2,$t2
+	and	$i3,lr,$s0,lsr#16
+	eor	$s3,$s3,$t3
+	mov	$s0,$s0,lsr#24
+
+	subs	$rounds,$rounds,#1
+	bne	.Lenc_loop
+
+	add	$tbl,$tbl,#2
+
+	ldrb	$t1,[$tbl,$i1,lsl#2]	@ Te4[s0>>0]
+	and	$i1,lr,$s1,lsr#16	@ i0
+	ldrb	$t2,[$tbl,$i2,lsl#2]	@ Te4[s0>>8]
+	and	$i2,lr,$s1
+	ldrb	$t3,[$tbl,$i3,lsl#2]	@ Te4[s0>>16]
+	and	$i3,lr,$s1,lsr#8
+	ldrb	$s0,[$tbl,$s0,lsl#2]	@ Te4[s0>>24]
+	mov	$s1,$s1,lsr#24
+
+	ldrb	$i1,[$tbl,$i1,lsl#2]	@ Te4[s1>>16]
+	ldrb	$i2,[$tbl,$i2,lsl#2]	@ Te4[s1>>0]
+	ldrb	$i3,[$tbl,$i3,lsl#2]	@ Te4[s1>>8]
+	eor	$s0,$i1,$s0,lsl#8
+	ldrb	$s1,[$tbl,$s1,lsl#2]	@ Te4[s1>>24]
+	and	$i1,lr,$s2,lsr#8	@ i0
+	eor	$t2,$i2,$t2,lsl#8
+	and	$i2,lr,$s2,lsr#16	@ i1
+	eor	$t3,$i3,$t3,lsl#8
+	and	$i3,lr,$s2
+	ldrb	$i1,[$tbl,$i1,lsl#2]	@ Te4[s2>>8]
+	eor	$s1,$t1,$s1,lsl#24
+	ldrb	$i2,[$tbl,$i2,lsl#2]	@ Te4[s2>>16]
+	mov	$s2,$s2,lsr#24
+
+	ldrb	$i3,[$tbl,$i3,lsl#2]	@ Te4[s2>>0]
+	eor	$s0,$i1,$s0,lsl#8
+	ldrb	$s2,[$tbl,$s2,lsl#2]	@ Te4[s2>>24]
+	and	$i1,lr,$s3		@ i0
+	eor	$s1,$s1,$i2,lsl#16
+	and	$i2,lr,$s3,lsr#8	@ i1
+	eor	$t3,$i3,$t3,lsl#8
+	and	$i3,lr,$s3,lsr#16	@ i2
+	ldrb	$i1,[$tbl,$i1,lsl#2]	@ Te4[s3>>0]
+	eor	$s2,$t2,$s2,lsl#24
+	ldrb	$i2,[$tbl,$i2,lsl#2]	@ Te4[s3>>8]
+	mov	$s3,$s3,lsr#24
+
+	ldrb	$i3,[$tbl,$i3,lsl#2]	@ Te4[s3>>16]
+	eor	$s0,$i1,$s0,lsl#8
+	ldr	$i1,[$key,#0]
+	ldrb	$s3,[$tbl,$s3,lsl#2]	@ Te4[s3>>24]
+	eor	$s1,$s1,$i2,lsl#8
+	ldr	$t1,[$key,#4]
+	eor	$s2,$s2,$i3,lsl#16
+	ldr	$t2,[$key,#8]
+	eor	$s3,$t3,$s3,lsl#24
+	ldr	$t3,[$key,#12]
+
+	eor	$s0,$s0,$i1
+	eor	$s1,$s1,$t1
+	eor	$s2,$s2,$t2
+	eor	$s3,$s3,$t3
+
+	sub	$tbl,$tbl,#2
+	ldr	pc,[sp],#4		@ pop and return
+.size	_armv4_AES_encrypt,.-_armv4_AES_encrypt
+
+.global private_AES_set_encrypt_key
+.type   private_AES_set_encrypt_key,%function
+.align	5
+private_AES_set_encrypt_key:
+_armv4_AES_set_encrypt_key:
+	sub	r3,pc,#8		@ AES_set_encrypt_key
+	teq	r0,#0
+	moveq	r0,#-1
+	beq	.Labrt
+	teq	r2,#0
+	moveq	r0,#-1
+	beq	.Labrt
+
+	teq	r1,#128
+	beq	.Lok
+	teq	r1,#192
+	beq	.Lok
+	teq	r1,#256
+	movne	r0,#-1
+	bne	.Labrt
+
+.Lok:	stmdb   sp!,{r4-r12,lr}
+	sub	$tbl,r3,#_armv4_AES_set_encrypt_key-AES_Te-1024	@ Te4
+
+	mov	$rounds,r0		@ inp
+	mov	lr,r1			@ bits
+	mov	$key,r2			@ key
+
+#if __ARM_ARCH__<7
+	ldrb	$s0,[$rounds,#3]	@ load input data in endian-neutral
+	ldrb	$t1,[$rounds,#2]	@ manner...
+	ldrb	$t2,[$rounds,#1]
+	ldrb	$t3,[$rounds,#0]
+	orr	$s0,$s0,$t1,lsl#8
+	ldrb	$s1,[$rounds,#7]
+	orr	$s0,$s0,$t2,lsl#16
+	ldrb	$t1,[$rounds,#6]
+	orr	$s0,$s0,$t3,lsl#24
+	ldrb	$t2,[$rounds,#5]
+	ldrb	$t3,[$rounds,#4]
+	orr	$s1,$s1,$t1,lsl#8
+	ldrb	$s2,[$rounds,#11]
+	orr	$s1,$s1,$t2,lsl#16
+	ldrb	$t1,[$rounds,#10]
+	orr	$s1,$s1,$t3,lsl#24
+	ldrb	$t2,[$rounds,#9]
+	ldrb	$t3,[$rounds,#8]
+	orr	$s2,$s2,$t1,lsl#8
+	ldrb	$s3,[$rounds,#15]
+	orr	$s2,$s2,$t2,lsl#16
+	ldrb	$t1,[$rounds,#14]
+	orr	$s2,$s2,$t3,lsl#24
+	ldrb	$t2,[$rounds,#13]
+	ldrb	$t3,[$rounds,#12]
+	orr	$s3,$s3,$t1,lsl#8
+	str	$s0,[$key],#16
+	orr	$s3,$s3,$t2,lsl#16
+	str	$s1,[$key,#-12]
+	orr	$s3,$s3,$t3,lsl#24
+	str	$s2,[$key,#-8]
+	str	$s3,[$key,#-4]
+#else
+	ldr	$s0,[$rounds,#0]
+	ldr	$s1,[$rounds,#4]
+	ldr	$s2,[$rounds,#8]
+	ldr	$s3,[$rounds,#12]
+#ifdef __ARMEL__
+	rev	$s0,$s0
+	rev	$s1,$s1
+	rev	$s2,$s2
+	rev	$s3,$s3
+#endif
+	str	$s0,[$key],#16
+	str	$s1,[$key,#-12]
+	str	$s2,[$key,#-8]
+	str	$s3,[$key,#-4]
+#endif
+
+	teq	lr,#128
+	bne	.Lnot128
+	mov	$rounds,#10
+	str	$rounds,[$key,#240-16]
+	add	$t3,$tbl,#256			@ rcon
+	mov	lr,#255
+
+.L128_loop:
+	and	$t2,lr,$s3,lsr#24
+	and	$i1,lr,$s3,lsr#16
+	ldrb	$t2,[$tbl,$t2]
+	and	$i2,lr,$s3,lsr#8
+	ldrb	$i1,[$tbl,$i1]
+	and	$i3,lr,$s3
+	ldrb	$i2,[$tbl,$i2]
+	orr	$t2,$t2,$i1,lsl#24
+	ldrb	$i3,[$tbl,$i3]
+	orr	$t2,$t2,$i2,lsl#16
+	ldr	$t1,[$t3],#4			@ rcon[i++]
+	orr	$t2,$t2,$i3,lsl#8
+	eor	$t2,$t2,$t1
+	eor	$s0,$s0,$t2			@ rk[4]=rk[0]^...
+	eor	$s1,$s1,$s0			@ rk[5]=rk[1]^rk[4]
+	str	$s0,[$key],#16
+	eor	$s2,$s2,$s1			@ rk[6]=rk[2]^rk[5]
+	str	$s1,[$key,#-12]
+	eor	$s3,$s3,$s2			@ rk[7]=rk[3]^rk[6]
+	str	$s2,[$key,#-8]
+	subs	$rounds,$rounds,#1
+	str	$s3,[$key,#-4]
+	bne	.L128_loop
+	sub	r2,$key,#176
+	b	.Ldone
+
+.Lnot128:
+#if __ARM_ARCH__<7
+	ldrb	$i2,[$rounds,#19]
+	ldrb	$t1,[$rounds,#18]
+	ldrb	$t2,[$rounds,#17]
+	ldrb	$t3,[$rounds,#16]
+	orr	$i2,$i2,$t1,lsl#8
+	ldrb	$i3,[$rounds,#23]
+	orr	$i2,$i2,$t2,lsl#16
+	ldrb	$t1,[$rounds,#22]
+	orr	$i2,$i2,$t3,lsl#24
+	ldrb	$t2,[$rounds,#21]
+	ldrb	$t3,[$rounds,#20]
+	orr	$i3,$i3,$t1,lsl#8
+	orr	$i3,$i3,$t2,lsl#16
+	str	$i2,[$key],#8
+	orr	$i3,$i3,$t3,lsl#24
+	str	$i3,[$key,#-4]
+#else
+	ldr	$i2,[$rounds,#16]
+	ldr	$i3,[$rounds,#20]
+#ifdef __ARMEL__
+	rev	$i2,$i2
+	rev	$i3,$i3
+#endif
+	str	$i2,[$key],#8
+	str	$i3,[$key,#-4]
+#endif
+
+	teq	lr,#192
+	bne	.Lnot192
+	mov	$rounds,#12
+	str	$rounds,[$key,#240-24]
+	add	$t3,$tbl,#256			@ rcon
+	mov	lr,#255
+	mov	$rounds,#8
+
+.L192_loop:
+	and	$t2,lr,$i3,lsr#24
+	and	$i1,lr,$i3,lsr#16
+	ldrb	$t2,[$tbl,$t2]
+	and	$i2,lr,$i3,lsr#8
+	ldrb	$i1,[$tbl,$i1]
+	and	$i3,lr,$i3
+	ldrb	$i2,[$tbl,$i2]
+	orr	$t2,$t2,$i1,lsl#24
+	ldrb	$i3,[$tbl,$i3]
+	orr	$t2,$t2,$i2,lsl#16
+	ldr	$t1,[$t3],#4			@ rcon[i++]
+	orr	$t2,$t2,$i3,lsl#8
+	eor	$i3,$t2,$t1
+	eor	$s0,$s0,$i3			@ rk[6]=rk[0]^...
+	eor	$s1,$s1,$s0			@ rk[7]=rk[1]^rk[6]
+	str	$s0,[$key],#24
+	eor	$s2,$s2,$s1			@ rk[8]=rk[2]^rk[7]
+	str	$s1,[$key,#-20]
+	eor	$s3,$s3,$s2			@ rk[9]=rk[3]^rk[8]
+	str	$s2,[$key,#-16]
+	subs	$rounds,$rounds,#1
+	str	$s3,[$key,#-12]
+	subeq	r2,$key,#216
+	beq	.Ldone
+
+	ldr	$i1,[$key,#-32]
+	ldr	$i2,[$key,#-28]
+	eor	$i1,$i1,$s3			@ rk[10]=rk[4]^rk[9]
+	eor	$i3,$i2,$i1			@ rk[11]=rk[5]^rk[10]
+	str	$i1,[$key,#-8]
+	str	$i3,[$key,#-4]
+	b	.L192_loop
+
+.Lnot192:
+#if __ARM_ARCH__<7
+	ldrb	$i2,[$rounds,#27]
+	ldrb	$t1,[$rounds,#26]
+	ldrb	$t2,[$rounds,#25]
+	ldrb	$t3,[$rounds,#24]
+	orr	$i2,$i2,$t1,lsl#8
+	ldrb	$i3,[$rounds,#31]
+	orr	$i2,$i2,$t2,lsl#16
+	ldrb	$t1,[$rounds,#30]
+	orr	$i2,$i2,$t3,lsl#24
+	ldrb	$t2,[$rounds,#29]
+	ldrb	$t3,[$rounds,#28]
+	orr	$i3,$i3,$t1,lsl#8
+	orr	$i3,$i3,$t2,lsl#16
+	str	$i2,[$key],#8
+	orr	$i3,$i3,$t3,lsl#24
+	str	$i3,[$key,#-4]
+#else
+	ldr	$i2,[$rounds,#24]
+	ldr	$i3,[$rounds,#28]
+#ifdef __ARMEL__
+	rev	$i2,$i2
+	rev	$i3,$i3
+#endif
+	str	$i2,[$key],#8
+	str	$i3,[$key,#-4]
+#endif
+
+	mov	$rounds,#14
+	str	$rounds,[$key,#240-32]
+	add	$t3,$tbl,#256			@ rcon
+	mov	lr,#255
+	mov	$rounds,#7
+
+.L256_loop:
+	and	$t2,lr,$i3,lsr#24
+	and	$i1,lr,$i3,lsr#16
+	ldrb	$t2,[$tbl,$t2]
+	and	$i2,lr,$i3,lsr#8
+	ldrb	$i1,[$tbl,$i1]
+	and	$i3,lr,$i3
+	ldrb	$i2,[$tbl,$i2]
+	orr	$t2,$t2,$i1,lsl#24
+	ldrb	$i3,[$tbl,$i3]
+	orr	$t2,$t2,$i2,lsl#16
+	ldr	$t1,[$t3],#4			@ rcon[i++]
+	orr	$t2,$t2,$i3,lsl#8
+	eor	$i3,$t2,$t1
+	eor	$s0,$s0,$i3			@ rk[8]=rk[0]^...
+	eor	$s1,$s1,$s0			@ rk[9]=rk[1]^rk[8]
+	str	$s0,[$key],#32
+	eor	$s2,$s2,$s1			@ rk[10]=rk[2]^rk[9]
+	str	$s1,[$key,#-28]
+	eor	$s3,$s3,$s2			@ rk[11]=rk[3]^rk[10]
+	str	$s2,[$key,#-24]
+	subs	$rounds,$rounds,#1
+	str	$s3,[$key,#-20]
+	subeq	r2,$key,#256
+	beq	.Ldone
+
+	and	$t2,lr,$s3
+	and	$i1,lr,$s3,lsr#8
+	ldrb	$t2,[$tbl,$t2]
+	and	$i2,lr,$s3,lsr#16
+	ldrb	$i1,[$tbl,$i1]
+	and	$i3,lr,$s3,lsr#24
+	ldrb	$i2,[$tbl,$i2]
+	orr	$t2,$t2,$i1,lsl#8
+	ldrb	$i3,[$tbl,$i3]
+	orr	$t2,$t2,$i2,lsl#16
+	ldr	$t1,[$key,#-48]
+	orr	$t2,$t2,$i3,lsl#24
+
+	ldr	$i1,[$key,#-44]
+	ldr	$i2,[$key,#-40]
+	eor	$t1,$t1,$t2			@ rk[12]=rk[4]^...
+	ldr	$i3,[$key,#-36]
+	eor	$i1,$i1,$t1			@ rk[13]=rk[5]^rk[12]
+	str	$t1,[$key,#-16]
+	eor	$i2,$i2,$i1			@ rk[14]=rk[6]^rk[13]
+	str	$i1,[$key,#-12]
+	eor	$i3,$i3,$i2			@ rk[15]=rk[7]^rk[14]
+	str	$i2,[$key,#-8]
+	str	$i3,[$key,#-4]
+	b	.L256_loop
+
+.Ldone:	mov	r0,#0
+	ldmia   sp!,{r4-r12,lr}
+.Labrt:	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+.size	private_AES_set_encrypt_key,.-private_AES_set_encrypt_key
+
+.global private_AES_set_decrypt_key
+.type   private_AES_set_decrypt_key,%function
+.align	5
+private_AES_set_decrypt_key:
+	str	lr,[sp,#-4]!            @ push lr
+	bl	_armv4_AES_set_encrypt_key
+	teq	r0,#0
+	ldrne	lr,[sp],#4              @ pop lr
+	bne	.Labrt
+
+	stmdb   sp!,{r4-r12}
+
+	ldr	$rounds,[r2,#240]	@ AES_set_encrypt_key preserves r2,
+	mov	$key,r2			@ which is AES_KEY *key
+	mov	$i1,r2
+	add	$i2,r2,$rounds,lsl#4
+
+.Linv:	ldr	$s0,[$i1]
+	ldr	$s1,[$i1,#4]
+	ldr	$s2,[$i1,#8]
+	ldr	$s3,[$i1,#12]
+	ldr	$t1,[$i2]
+	ldr	$t2,[$i2,#4]
+	ldr	$t3,[$i2,#8]
+	ldr	$i3,[$i2,#12]
+	str	$s0,[$i2],#-16
+	str	$s1,[$i2,#16+4]
+	str	$s2,[$i2,#16+8]
+	str	$s3,[$i2,#16+12]
+	str	$t1,[$i1],#16
+	str	$t2,[$i1,#-12]
+	str	$t3,[$i1,#-8]
+	str	$i3,[$i1,#-4]
+	teq	$i1,$i2
+	bne	.Linv
+___
+$mask80=$i1;
+$mask1b=$i2;
+$mask7f=$i3;
+$code.=<<___;
+	ldr	$s0,[$key,#16]!		@ prefetch tp1
+	mov	$mask80,#0x80
+	mov	$mask1b,#0x1b
+	orr	$mask80,$mask80,#0x8000
+	orr	$mask1b,$mask1b,#0x1b00
+	orr	$mask80,$mask80,$mask80,lsl#16
+	orr	$mask1b,$mask1b,$mask1b,lsl#16
+	sub	$rounds,$rounds,#1
+	mvn	$mask7f,$mask80
+	mov	$rounds,$rounds,lsl#2	@ (rounds-1)*4
+
+.Lmix:	and	$t1,$s0,$mask80
+	and	$s1,$s0,$mask7f
+	sub	$t1,$t1,$t1,lsr#7
+	and	$t1,$t1,$mask1b
+	eor	$s1,$t1,$s1,lsl#1	@ tp2
+
+	and	$t1,$s1,$mask80
+	and	$s2,$s1,$mask7f
+	sub	$t1,$t1,$t1,lsr#7
+	and	$t1,$t1,$mask1b
+	eor	$s2,$t1,$s2,lsl#1	@ tp4
+
+	and	$t1,$s2,$mask80
+	and	$s3,$s2,$mask7f
+	sub	$t1,$t1,$t1,lsr#7
+	and	$t1,$t1,$mask1b
+	eor	$s3,$t1,$s3,lsl#1	@ tp8
+
+	eor	$t1,$s1,$s2
+	eor	$t2,$s0,$s3		@ tp9
+	eor	$t1,$t1,$s3		@ tpe
+	eor	$t1,$t1,$s1,ror#24
+	eor	$t1,$t1,$t2,ror#24	@ ^= ROTATE(tpb=tp9^tp2,8)
+	eor	$t1,$t1,$s2,ror#16
+	eor	$t1,$t1,$t2,ror#16	@ ^= ROTATE(tpd=tp9^tp4,16)
+	eor	$t1,$t1,$t2,ror#8	@ ^= ROTATE(tp9,24)
+
+	ldr	$s0,[$key,#4]		@ prefetch tp1
+	str	$t1,[$key],#4
+	subs	$rounds,$rounds,#1
+	bne	.Lmix
+
+	mov	r0,#0
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r12,pc}
+#else
+	ldmia   sp!,{r4-r12,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.size	private_AES_set_decrypt_key,.-private_AES_set_decrypt_key
+
+.type	AES_Td,%object
+.align	5
+AES_Td:
+.word	0x51f4a750, 0x7e416553, 0x1a17a4c3, 0x3a275e96
+.word	0x3bab6bcb, 0x1f9d45f1, 0xacfa58ab, 0x4be30393
+.word	0x2030fa55, 0xad766df6, 0x88cc7691, 0xf5024c25
+.word	0x4fe5d7fc, 0xc52acbd7, 0x26354480, 0xb562a38f
+.word	0xdeb15a49, 0x25ba1b67, 0x45ea0e98, 0x5dfec0e1
+.word	0xc32f7502, 0x814cf012, 0x8d4697a3, 0x6bd3f9c6
+.word	0x038f5fe7, 0x15929c95, 0xbf6d7aeb, 0x955259da
+.word	0xd4be832d, 0x587421d3, 0x49e06929, 0x8ec9c844
+.word	0x75c2896a, 0xf48e7978, 0x99583e6b, 0x27b971dd
+.word	0xbee14fb6, 0xf088ad17, 0xc920ac66, 0x7dce3ab4
+.word	0x63df4a18, 0xe51a3182, 0x97513360, 0x62537f45
+.word	0xb16477e0, 0xbb6bae84, 0xfe81a01c, 0xf9082b94
+.word	0x70486858, 0x8f45fd19, 0x94de6c87, 0x527bf8b7
+.word	0xab73d323, 0x724b02e2, 0xe31f8f57, 0x6655ab2a
+.word	0xb2eb2807, 0x2fb5c203, 0x86c57b9a, 0xd33708a5
+.word	0x302887f2, 0x23bfa5b2, 0x02036aba, 0xed16825c
+.word	0x8acf1c2b, 0xa779b492, 0xf307f2f0, 0x4e69e2a1
+.word	0x65daf4cd, 0x0605bed5, 0xd134621f, 0xc4a6fe8a
+.word	0x342e539d, 0xa2f355a0, 0x058ae132, 0xa4f6eb75
+.word	0x0b83ec39, 0x4060efaa, 0x5e719f06, 0xbd6e1051
+.word	0x3e218af9, 0x96dd063d, 0xdd3e05ae, 0x4de6bd46
+.word	0x91548db5, 0x71c45d05, 0x0406d46f, 0x605015ff
+.word	0x1998fb24, 0xd6bde997, 0x894043cc, 0x67d99e77
+.word	0xb0e842bd, 0x07898b88, 0xe7195b38, 0x79c8eedb
+.word	0xa17c0a47, 0x7c420fe9, 0xf8841ec9, 0x00000000
+.word	0x09808683, 0x322bed48, 0x1e1170ac, 0x6c5a724e
+.word	0xfd0efffb, 0x0f853856, 0x3daed51e, 0x362d3927
+.word	0x0a0fd964, 0x685ca621, 0x9b5b54d1, 0x24362e3a
+.word	0x0c0a67b1, 0x9357e70f, 0xb4ee96d2, 0x1b9b919e
+.word	0x80c0c54f, 0x61dc20a2, 0x5a774b69, 0x1c121a16
+.word	0xe293ba0a, 0xc0a02ae5, 0x3c22e043, 0x121b171d
+.word	0x0e090d0b, 0xf28bc7ad, 0x2db6a8b9, 0x141ea9c8
+.word	0x57f11985, 0xaf75074c, 0xee99ddbb, 0xa37f60fd
+.word	0xf701269f, 0x5c72f5bc, 0x44663bc5, 0x5bfb7e34
+.word	0x8b432976, 0xcb23c6dc, 0xb6edfc68, 0xb8e4f163
+.word	0xd731dcca, 0x42638510, 0x13972240, 0x84c61120
+.word	0x854a247d, 0xd2bb3df8, 0xaef93211, 0xc729a16d
+.word	0x1d9e2f4b, 0xdcb230f3, 0x0d8652ec, 0x77c1e3d0
+.word	0x2bb3166c, 0xa970b999, 0x119448fa, 0x47e96422
+.word	0xa8fc8cc4, 0xa0f03f1a, 0x567d2cd8, 0x223390ef
+.word	0x87494ec7, 0xd938d1c1, 0x8ccaa2fe, 0x98d40b36
+.word	0xa6f581cf, 0xa57ade28, 0xdab78e26, 0x3fadbfa4
+.word	0x2c3a9de4, 0x5078920d, 0x6a5fcc9b, 0x547e4662
+.word	0xf68d13c2, 0x90d8b8e8, 0x2e39f75e, 0x82c3aff5
+.word	0x9f5d80be, 0x69d0937c, 0x6fd52da9, 0xcf2512b3
+.word	0xc8ac993b, 0x10187da7, 0xe89c636e, 0xdb3bbb7b
+.word	0xcd267809, 0x6e5918f4, 0xec9ab701, 0x834f9aa8
+.word	0xe6956e65, 0xaaffe67e, 0x21bccf08, 0xef15e8e6
+.word	0xbae79bd9, 0x4a6f36ce, 0xea9f09d4, 0x29b07cd6
+.word	0x31a4b2af, 0x2a3f2331, 0xc6a59430, 0x35a266c0
+.word	0x744ebc37, 0xfc82caa6, 0xe090d0b0, 0x33a7d815
+.word	0xf104984a, 0x41ecdaf7, 0x7fcd500e, 0x1791f62f
+.word	0x764dd68d, 0x43efb04d, 0xccaa4d54, 0xe49604df
+.word	0x9ed1b5e3, 0x4c6a881b, 0xc12c1fb8, 0x4665517f
+.word	0x9d5eea04, 0x018c355d, 0xfa877473, 0xfb0b412e
+.word	0xb3671d5a, 0x92dbd252, 0xe9105633, 0x6dd64713
+.word	0x9ad7618c, 0x37a10c7a, 0x59f8148e, 0xeb133c89
+.word	0xcea927ee, 0xb761c935, 0xe11ce5ed, 0x7a47b13c
+.word	0x9cd2df59, 0x55f2733f, 0x1814ce79, 0x73c737bf
+.word	0x53f7cdea, 0x5ffdaa5b, 0xdf3d6f14, 0x7844db86
+.word	0xcaaff381, 0xb968c43e, 0x3824342c, 0xc2a3405f
+.word	0x161dc372, 0xbce2250c, 0x283c498b, 0xff0d9541
+.word	0x39a80171, 0x080cb3de, 0xd8b4e49c, 0x6456c190
+.word	0x7bcb8461, 0xd532b670, 0x486c5c74, 0xd0b85742
+@ Td4[256]
+.byte	0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38
+.byte	0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb
+.byte	0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87
+.byte	0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb
+.byte	0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d
+.byte	0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e
+.byte	0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2
+.byte	0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25
+.byte	0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16
+.byte	0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92
+.byte	0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda
+.byte	0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84
+.byte	0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a
+.byte	0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06
+.byte	0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02
+.byte	0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b
+.byte	0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea
+.byte	0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73
+.byte	0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85
+.byte	0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e
+.byte	0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89
+.byte	0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b
+.byte	0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20
+.byte	0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4
+.byte	0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31
+.byte	0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f
+.byte	0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d
+.byte	0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef
+.byte	0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0
+.byte	0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61
+.byte	0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26
+.byte	0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
+.size	AES_Td,.-AES_Td
+
+@ void AES_decrypt(const unsigned char *in, unsigned char *out,
+@ 		 const AES_KEY *key) {
+.global AES_decrypt
+.type   AES_decrypt,%function
+.align	5
+AES_decrypt:
+	sub	r3,pc,#8		@ AES_decrypt
+	stmdb   sp!,{r1,r4-r12,lr}
+	mov	$rounds,r0		@ inp
+	mov	$key,r2
+	sub	$tbl,r3,#AES_decrypt-AES_Td		@ Td
+#if __ARM_ARCH__<7
+	ldrb	$s0,[$rounds,#3]	@ load input data in endian-neutral
+	ldrb	$t1,[$rounds,#2]	@ manner...
+	ldrb	$t2,[$rounds,#1]
+	ldrb	$t3,[$rounds,#0]
+	orr	$s0,$s0,$t1,lsl#8
+	ldrb	$s1,[$rounds,#7]
+	orr	$s0,$s0,$t2,lsl#16
+	ldrb	$t1,[$rounds,#6]
+	orr	$s0,$s0,$t3,lsl#24
+	ldrb	$t2,[$rounds,#5]
+	ldrb	$t3,[$rounds,#4]
+	orr	$s1,$s1,$t1,lsl#8
+	ldrb	$s2,[$rounds,#11]
+	orr	$s1,$s1,$t2,lsl#16
+	ldrb	$t1,[$rounds,#10]
+	orr	$s1,$s1,$t3,lsl#24
+	ldrb	$t2,[$rounds,#9]
+	ldrb	$t3,[$rounds,#8]
+	orr	$s2,$s2,$t1,lsl#8
+	ldrb	$s3,[$rounds,#15]
+	orr	$s2,$s2,$t2,lsl#16
+	ldrb	$t1,[$rounds,#14]
+	orr	$s2,$s2,$t3,lsl#24
+	ldrb	$t2,[$rounds,#13]
+	ldrb	$t3,[$rounds,#12]
+	orr	$s3,$s3,$t1,lsl#8
+	orr	$s3,$s3,$t2,lsl#16
+	orr	$s3,$s3,$t3,lsl#24
+#else
+	ldr	$s0,[$rounds,#0]
+	ldr	$s1,[$rounds,#4]
+	ldr	$s2,[$rounds,#8]
+	ldr	$s3,[$rounds,#12]
+#ifdef __ARMEL__
+	rev	$s0,$s0
+	rev	$s1,$s1
+	rev	$s2,$s2
+	rev	$s3,$s3
+#endif
+#endif
+	bl	_armv4_AES_decrypt
+
+	ldr	$rounds,[sp],#4		@ pop out
+#if __ARM_ARCH__>=7
+#ifdef __ARMEL__
+	rev	$s0,$s0
+	rev	$s1,$s1
+	rev	$s2,$s2
+	rev	$s3,$s3
+#endif
+	str	$s0,[$rounds,#0]
+	str	$s1,[$rounds,#4]
+	str	$s2,[$rounds,#8]
+	str	$s3,[$rounds,#12]
+#else
+	mov	$t1,$s0,lsr#24		@ write output in endian-neutral
+	mov	$t2,$s0,lsr#16		@ manner...
+	mov	$t3,$s0,lsr#8
+	strb	$t1,[$rounds,#0]
+	strb	$t2,[$rounds,#1]
+	mov	$t1,$s1,lsr#24
+	strb	$t3,[$rounds,#2]
+	mov	$t2,$s1,lsr#16
+	strb	$s0,[$rounds,#3]
+	mov	$t3,$s1,lsr#8
+	strb	$t1,[$rounds,#4]
+	strb	$t2,[$rounds,#5]
+	mov	$t1,$s2,lsr#24
+	strb	$t3,[$rounds,#6]
+	mov	$t2,$s2,lsr#16
+	strb	$s1,[$rounds,#7]
+	mov	$t3,$s2,lsr#8
+	strb	$t1,[$rounds,#8]
+	strb	$t2,[$rounds,#9]
+	mov	$t1,$s3,lsr#24
+	strb	$t3,[$rounds,#10]
+	mov	$t2,$s3,lsr#16
+	strb	$s2,[$rounds,#11]
+	mov	$t3,$s3,lsr#8
+	strb	$t1,[$rounds,#12]
+	strb	$t2,[$rounds,#13]
+	strb	$t3,[$rounds,#14]
+	strb	$s3,[$rounds,#15]
+#endif
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r12,pc}
+#else
+	ldmia   sp!,{r4-r12,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.size	AES_decrypt,.-AES_decrypt
+
+.type   _armv4_AES_decrypt,%function
+.align	2
+_armv4_AES_decrypt:
+	str	lr,[sp,#-4]!		@ push lr
+	ldmia	$key!,{$t1-$i1}
+	eor	$s0,$s0,$t1
+	ldr	$rounds,[$key,#240-16]
+	eor	$s1,$s1,$t2
+	eor	$s2,$s2,$t3
+	eor	$s3,$s3,$i1
+	sub	$rounds,$rounds,#1
+	mov	lr,#255
+
+	and	$i1,lr,$s0,lsr#16
+	and	$i2,lr,$s0,lsr#8
+	and	$i3,lr,$s0
+	mov	$s0,$s0,lsr#24
+.Ldec_loop:
+	ldr	$t1,[$tbl,$i1,lsl#2]	@ Td1[s0>>16]
+	and	$i1,lr,$s1		@ i0
+	ldr	$t2,[$tbl,$i2,lsl#2]	@ Td2[s0>>8]
+	and	$i2,lr,$s1,lsr#16
+	ldr	$t3,[$tbl,$i3,lsl#2]	@ Td3[s0>>0]
+	and	$i3,lr,$s1,lsr#8
+	ldr	$s0,[$tbl,$s0,lsl#2]	@ Td0[s0>>24]
+	mov	$s1,$s1,lsr#24
+
+	ldr	$i1,[$tbl,$i1,lsl#2]	@ Td3[s1>>0]
+	ldr	$i2,[$tbl,$i2,lsl#2]	@ Td1[s1>>16]
+	ldr	$i3,[$tbl,$i3,lsl#2]	@ Td2[s1>>8]
+	eor	$s0,$s0,$i1,ror#24
+	ldr	$s1,[$tbl,$s1,lsl#2]	@ Td0[s1>>24]
+	and	$i1,lr,$s2,lsr#8	@ i0
+	eor	$t2,$i2,$t2,ror#8
+	and	$i2,lr,$s2		@ i1
+	eor	$t3,$i3,$t3,ror#8
+	and	$i3,lr,$s2,lsr#16
+	ldr	$i1,[$tbl,$i1,lsl#2]	@ Td2[s2>>8]
+	eor	$s1,$s1,$t1,ror#8
+	ldr	$i2,[$tbl,$i2,lsl#2]	@ Td3[s2>>0]
+	mov	$s2,$s2,lsr#24
+
+	ldr	$i3,[$tbl,$i3,lsl#2]	@ Td1[s2>>16]
+	eor	$s0,$s0,$i1,ror#16
+	ldr	$s2,[$tbl,$s2,lsl#2]	@ Td0[s2>>24]
+	and	$i1,lr,$s3,lsr#16	@ i0
+	eor	$s1,$s1,$i2,ror#24
+	and	$i2,lr,$s3,lsr#8	@ i1
+	eor	$t3,$i3,$t3,ror#8
+	and	$i3,lr,$s3		@ i2
+	ldr	$i1,[$tbl,$i1,lsl#2]	@ Td1[s3>>16]
+	eor	$s2,$s2,$t2,ror#8
+	ldr	$i2,[$tbl,$i2,lsl#2]	@ Td2[s3>>8]
+	mov	$s3,$s3,lsr#24
+
+	ldr	$i3,[$tbl,$i3,lsl#2]	@ Td3[s3>>0]
+	eor	$s0,$s0,$i1,ror#8
+	ldr	$i1,[$key],#16
+	eor	$s1,$s1,$i2,ror#16
+	ldr	$s3,[$tbl,$s3,lsl#2]	@ Td0[s3>>24]
+	eor	$s2,$s2,$i3,ror#24
+
+	ldr	$t1,[$key,#-12]
+	eor	$s0,$s0,$i1
+	ldr	$t2,[$key,#-8]
+	eor	$s3,$s3,$t3,ror#8
+	ldr	$t3,[$key,#-4]
+	and	$i1,lr,$s0,lsr#16
+	eor	$s1,$s1,$t1
+	and	$i2,lr,$s0,lsr#8
+	eor	$s2,$s2,$t2
+	and	$i3,lr,$s0
+	eor	$s3,$s3,$t3
+	mov	$s0,$s0,lsr#24
+
+	subs	$rounds,$rounds,#1
+	bne	.Ldec_loop
+
+	add	$tbl,$tbl,#1024
+
+	ldr	$t2,[$tbl,#0]		@ prefetch Td4
+	ldr	$t3,[$tbl,#32]
+	ldr	$t1,[$tbl,#64]
+	ldr	$t2,[$tbl,#96]
+	ldr	$t3,[$tbl,#128]
+	ldr	$t1,[$tbl,#160]
+	ldr	$t2,[$tbl,#192]
+	ldr	$t3,[$tbl,#224]
+
+	ldrb	$s0,[$tbl,$s0]		@ Td4[s0>>24]
+	ldrb	$t1,[$tbl,$i1]		@ Td4[s0>>16]
+	and	$i1,lr,$s1		@ i0
+	ldrb	$t2,[$tbl,$i2]		@ Td4[s0>>8]
+	and	$i2,lr,$s1,lsr#16
+	ldrb	$t3,[$tbl,$i3]		@ Td4[s0>>0]
+	and	$i3,lr,$s1,lsr#8
+
+	ldrb	$i1,[$tbl,$i1]		@ Td4[s1>>0]
+	ldrb	$s1,[$tbl,$s1,lsr#24]	@ Td4[s1>>24]
+	ldrb	$i2,[$tbl,$i2]		@ Td4[s1>>16]
+	eor	$s0,$i1,$s0,lsl#24
+	ldrb	$i3,[$tbl,$i3]		@ Td4[s1>>8]
+	eor	$s1,$t1,$s1,lsl#8
+	and	$i1,lr,$s2,lsr#8	@ i0
+	eor	$t2,$t2,$i2,lsl#8
+	and	$i2,lr,$s2		@ i1
+	ldrb	$i1,[$tbl,$i1]		@ Td4[s2>>8]
+	eor	$t3,$t3,$i3,lsl#8
+	ldrb	$i2,[$tbl,$i2]		@ Td4[s2>>0]
+	and	$i3,lr,$s2,lsr#16
+
+	ldrb	$s2,[$tbl,$s2,lsr#24]	@ Td4[s2>>24]
+	eor	$s0,$s0,$i1,lsl#8
+	ldrb	$i3,[$tbl,$i3]		@ Td4[s2>>16]
+	eor	$s1,$i2,$s1,lsl#16
+	and	$i1,lr,$s3,lsr#16	@ i0
+	eor	$s2,$t2,$s2,lsl#16
+	and	$i2,lr,$s3,lsr#8	@ i1
+	ldrb	$i1,[$tbl,$i1]		@ Td4[s3>>16]
+	eor	$t3,$t3,$i3,lsl#16
+	ldrb	$i2,[$tbl,$i2]		@ Td4[s3>>8]
+	and	$i3,lr,$s3		@ i2
+
+	ldrb	$i3,[$tbl,$i3]		@ Td4[s3>>0]
+	ldrb	$s3,[$tbl,$s3,lsr#24]	@ Td4[s3>>24]
+	eor	$s0,$s0,$i1,lsl#16
+	ldr	$i1,[$key,#0]
+	eor	$s1,$s1,$i2,lsl#8
+	ldr	$t1,[$key,#4]
+	eor	$s2,$i3,$s2,lsl#8
+	ldr	$t2,[$key,#8]
+	eor	$s3,$t3,$s3,lsl#24
+	ldr	$t3,[$key,#12]
+
+	eor	$s0,$s0,$i1
+	eor	$s1,$s1,$t1
+	eor	$s2,$s2,$t2
+	eor	$s3,$s3,$t3
+
+	sub	$tbl,$tbl,#1024
+	ldr	pc,[sp],#4		@ pop and return
+.size	_armv4_AES_decrypt,.-_armv4_AES_decrypt
+.asciz	"AES for ARMv4, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+___
+
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
+print $code;
+close STDOUT;	# enforce flush
diff --git a/crypto/aes/asm/aes-mips.pl b/crypto/aes/asm/aes-mips.pl
new file mode 100755
index 000000000000..2ce6deffc88c
--- /dev/null
+++ b/crypto/aes/asm/aes-mips.pl
@@ -0,0 +1,1611 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# AES for MIPS
+
+# October 2010
+#
+# Code uses 1K[+256B] S-box and on single-issue core [such as R5000]
+# spends ~68 cycles per byte processed with 128-bit key. This is ~16%
+# faster than gcc-generated code, which is not very impressive. But
+# recall that compressed S-box requires extra processing, namely
+# additional rotations. Rotations are implemented with lwl/lwr pairs,
+# which is normally used for loading unaligned data. Another cool
+# thing about this module is its endian neutrality, which means that
+# it processes data without ever changing byte order...
+
+######################################################################
+# There is a number of MIPS ABI in use, O32 and N32/64 are most
+# widely used. Then there is a new contender: NUBI. It appears that if
+# one picks the latter, it's possible to arrange code in ABI neutral
+# manner. Therefore let's stick to NUBI register layout:
+#
+($zero,$at,$t0,$t1,$t2)=map("\$$_",(0..2,24,25));
+($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7,$s8,$s9,$s10,$s11)=map("\$$_",(12..23));
+($gp,$tp,$sp,$fp,$ra)=map("\$$_",(3,28..31));
+#
+# The return value is placed in $a0. Following coding rules facilitate
+# interoperability:
+#
+# - never ever touch $tp, "thread pointer", former $gp;
+# - copy return value to $t0, former $v0 [or to $a0 if you're adapting
+#   old code];
+# - on O32 populate $a4-$a7 with 'lw $aN,4*N($sp)' if necessary;
+#
+# For reference here is register layout for N32/64 MIPS ABIs:
+#
+# ($zero,$at,$v0,$v1)=map("\$$_",(0..3));
+# ($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+# ($t0,$t1,$t2,$t3,$t8,$t9)=map("\$$_",(12..15,24,25));
+# ($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7)=map("\$$_",(16..23));
+# ($gp,$sp,$fp,$ra)=map("\$$_",(28..31));
+#
+$flavour = shift; # supported flavours are o32,n32,64,nubi32,nubi64
+
+if ($flavour =~ /64|n32/i) {
+	$PTR_ADD="dadd";	# incidentally works even on n32
+	$PTR_SUB="dsub";	# incidentally works even on n32
+	$REG_S="sd";
+	$REG_L="ld";
+	$PTR_SLL="dsll";	# incidentally works even on n32
+	$SZREG=8;
+} else {
+	$PTR_ADD="add";
+	$PTR_SUB="sub";
+	$REG_S="sw";
+	$REG_L="lw";
+	$PTR_SLL="sll";
+	$SZREG=4;
+}
+$pf = ($flavour =~ /nubi/i) ? $t0 : $t2;
+#
+# <appro@openssl.org>
+#
+######################################################################
+
+$big_endian=(`echo MIPSEL | $ENV{CC} -E -P -`=~/MIPSEL/)?1:0;
+
+for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);	}
+open STDOUT,">$output";
+
+if (!defined($big_endian))
+{    $big_endian=(unpack('L',pack('N',1))==1);   }
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+my ($MSB,$LSB)=(0,3);	# automatically converted to little-endian
+
+$code.=<<___;
+.text
+#ifdef OPENSSL_FIPSCANISTER
+# include <openssl/fipssyms.h>
+#endif
+
+#if !defined(__vxworks) || defined(__pic__)
+.option	pic2
+#endif
+.set	noat
+___
+
+{{{
+my $FRAMESIZE=16*$SZREG;
+my $SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0xc0fff008 : 0xc0ff0000;
+
+my ($inp,$out,$key,$Tbl,$s0,$s1,$s2,$s3)=($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7);
+my ($i0,$i1,$i2,$i3)=($at,$t0,$t1,$t2);
+my ($t0,$t1,$t2,$t3,$t4,$t5,$t6,$t7,$t8,$t9,$t10,$t11) = map("\$$_",(12..23));
+my ($key0,$cnt)=($gp,$fp);
+
+# instuction ordering is "stolen" from output from MIPSpro assembler
+# invoked with -mips3 -O3 arguments...
+$code.=<<___;
+.align	5
+.ent	_mips_AES_encrypt
+_mips_AES_encrypt:
+	.frame	$sp,0,$ra
+	.set	reorder
+	lw	$t0,0($key)
+	lw	$t1,4($key)
+	lw	$t2,8($key)
+	lw	$t3,12($key)
+	lw	$cnt,240($key)
+	$PTR_ADD $key0,$key,16
+
+	xor	$s0,$t0
+	xor	$s1,$t1
+	xor	$s2,$t2
+	xor	$s3,$t3
+
+	sub	$cnt,1
+	_xtr	$i0,$s1,16-2
+.Loop_enc:
+	_xtr	$i1,$s2,16-2
+	_xtr	$i2,$s3,16-2
+	_xtr	$i3,$s0,16-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lwl	$t0,3($i0)		# Te1[s1>>16]
+	lwl	$t1,3($i1)		# Te1[s2>>16]
+	lwl	$t2,3($i2)		# Te1[s3>>16]
+	lwl	$t3,3($i3)		# Te1[s0>>16]
+	lwr	$t0,2($i0)		# Te1[s1>>16]
+	lwr	$t1,2($i1)		# Te1[s2>>16]
+	lwr	$t2,2($i2)		# Te1[s3>>16]
+	lwr	$t3,2($i3)		# Te1[s0>>16]
+
+	_xtr	$i0,$s2,8-2
+	_xtr	$i1,$s3,8-2
+	_xtr	$i2,$s0,8-2
+	_xtr	$i3,$s1,8-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lwl	$t4,2($i0)		# Te2[s2>>8]
+	lwl	$t5,2($i1)		# Te2[s3>>8]
+	lwl	$t6,2($i2)		# Te2[s0>>8]
+	lwl	$t7,2($i3)		# Te2[s1>>8]
+	lwr	$t4,1($i0)		# Te2[s2>>8]
+	lwr	$t5,1($i1)		# Te2[s3>>8]
+	lwr	$t6,1($i2)		# Te2[s0>>8]
+	lwr	$t7,1($i3)		# Te2[s1>>8]
+
+	_xtr	$i0,$s3,0-2
+	_xtr	$i1,$s0,0-2
+	_xtr	$i2,$s1,0-2
+	_xtr	$i3,$s2,0-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lwl	$t8,1($i0)		# Te3[s3]
+	lwl	$t9,1($i1)		# Te3[s0]
+	lwl	$t10,1($i2)		# Te3[s1]
+	lwl	$t11,1($i3)		# Te3[s2]
+	lwr	$t8,0($i0)		# Te3[s3]
+	lwr	$t9,0($i1)		# Te3[s0]
+	lwr	$t10,0($i2)		# Te3[s1]
+	lwr	$t11,0($i3)		# Te3[s2]
+
+	_xtr	$i0,$s0,24-2
+	_xtr	$i1,$s1,24-2
+	_xtr	$i2,$s2,24-2
+	_xtr	$i3,$s3,24-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	xor	$t0,$t4
+	xor	$t1,$t5
+	xor	$t2,$t6
+	xor	$t3,$t7
+	lw	$t4,0($i0)		# Te0[s0>>24]
+	lw	$t5,0($i1)		# Te0[s1>>24]
+	lw	$t6,0($i2)		# Te0[s2>>24]
+	lw	$t7,0($i3)		# Te0[s3>>24]
+
+	lw	$s0,0($key0)
+	lw	$s1,4($key0)
+	lw	$s2,8($key0)
+	lw	$s3,12($key0)
+
+	xor	$t0,$t8
+	xor	$t1,$t9
+	xor	$t2,$t10
+	xor	$t3,$t11
+
+	xor	$t0,$t4
+	xor	$t1,$t5
+	xor	$t2,$t6
+	xor	$t3,$t7
+
+	sub	$cnt,1
+	$PTR_ADD $key0,16
+	xor	$s0,$t0
+	xor	$s1,$t1
+	xor	$s2,$t2
+	xor	$s3,$t3
+	.set	noreorder
+	bnez	$cnt,.Loop_enc
+	_xtr	$i0,$s1,16-2
+
+	.set	reorder
+	_xtr	$i1,$s2,16-2
+	_xtr	$i2,$s3,16-2
+	_xtr	$i3,$s0,16-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$t0,2($i0)		# Te4[s1>>16]
+	lbu	$t1,2($i1)		# Te4[s2>>16]
+	lbu	$t2,2($i2)		# Te4[s3>>16]
+	lbu	$t3,2($i3)		# Te4[s0>>16]
+
+	_xtr	$i0,$s2,8-2
+	_xtr	$i1,$s3,8-2
+	_xtr	$i2,$s0,8-2
+	_xtr	$i3,$s1,8-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$t4,2($i0)		# Te4[s2>>8]
+	lbu	$t5,2($i1)		# Te4[s3>>8]
+	lbu	$t6,2($i2)		# Te4[s0>>8]
+	lbu	$t7,2($i3)		# Te4[s1>>8]
+
+	_xtr	$i0,$s0,24-2
+	_xtr	$i1,$s1,24-2
+	_xtr	$i2,$s2,24-2
+	_xtr	$i3,$s3,24-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$t8,2($i0)		# Te4[s0>>24]
+	lbu	$t9,2($i1)		# Te4[s1>>24]
+	lbu	$t10,2($i2)		# Te4[s2>>24]
+	lbu	$t11,2($i3)		# Te4[s3>>24]
+
+	_xtr	$i0,$s3,0-2
+	_xtr	$i1,$s0,0-2
+	_xtr	$i2,$s1,0-2
+	_xtr	$i3,$s2,0-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+
+	_ins	$t0,16
+	_ins	$t1,16
+	_ins	$t2,16
+	_ins	$t3,16
+
+	_ins	$t4,8
+	_ins	$t5,8
+	_ins	$t6,8
+	_ins	$t7,8
+
+	xor	$t0,$t4
+	xor	$t1,$t5
+	xor	$t2,$t6
+	xor	$t3,$t7
+
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$t4,2($i0)		# Te4[s3]
+	lbu	$t5,2($i1)		# Te4[s0]
+	lbu	$t6,2($i2)		# Te4[s1]
+	lbu	$t7,2($i3)		# Te4[s2]
+
+	_ins	$t8,24
+	_ins	$t9,24
+	_ins	$t10,24
+	_ins	$t11,24
+
+	lw	$s0,0($key0)
+	lw	$s1,4($key0)
+	lw	$s2,8($key0)
+	lw	$s3,12($key0)
+
+	xor	$t0,$t8
+	xor	$t1,$t9
+	xor	$t2,$t10
+	xor	$t3,$t11
+
+	_ins	$t4,0
+	_ins	$t5,0
+	_ins	$t6,0
+	_ins	$t7,0
+
+	xor	$t0,$t4
+	xor	$t1,$t5
+	xor	$t2,$t6
+	xor	$t3,$t7
+
+	xor	$s0,$t0
+	xor	$s1,$t1
+	xor	$s2,$t2
+	xor	$s3,$t3
+
+	jr	$ra
+.end	_mips_AES_encrypt
+
+.align	5
+.globl	AES_encrypt
+.ent	AES_encrypt
+AES_encrypt:
+	.frame	$sp,$FRAMESIZE,$ra
+	.mask	$SAVED_REGS_MASK,-$SZREG
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /o32/i);	# o32 PIC-ification
+	.cpload	$pf
+___
+$code.=<<___;
+	$PTR_SUB $sp,$FRAMESIZE
+	$REG_S	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_S	$fp,$FRAMESIZE-2*$SZREG($sp)
+	$REG_S	$s11,$FRAMESIZE-3*$SZREG($sp)
+	$REG_S	$s10,$FRAMESIZE-4*$SZREG($sp)
+	$REG_S	$s9,$FRAMESIZE-5*$SZREG($sp)
+	$REG_S	$s8,$FRAMESIZE-6*$SZREG($sp)
+	$REG_S	$s7,$FRAMESIZE-7*$SZREG($sp)
+	$REG_S	$s6,$FRAMESIZE-8*$SZREG($sp)
+	$REG_S	$s5,$FRAMESIZE-9*$SZREG($sp)
+	$REG_S	$s4,$FRAMESIZE-10*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);	# optimize non-nubi prologue
+	$REG_S	\$15,$FRAMESIZE-11*$SZREG($sp)
+	$REG_S	\$14,$FRAMESIZE-12*$SZREG($sp)
+	$REG_S	\$13,$FRAMESIZE-13*$SZREG($sp)
+	$REG_S	\$12,$FRAMESIZE-14*$SZREG($sp)
+	$REG_S	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___ if ($flavour !~ /o32/i);	# non-o32 PIC-ification
+	.cplocal	$Tbl
+	.cpsetup	$pf,$zero,AES_encrypt
+___
+$code.=<<___;
+	.set	reorder
+	la	$Tbl,AES_Te		# PIC-ified 'load address'
+
+	lwl	$s0,0+$MSB($inp)
+	lwl	$s1,4+$MSB($inp)
+	lwl	$s2,8+$MSB($inp)
+	lwl	$s3,12+$MSB($inp)
+	lwr	$s0,0+$LSB($inp)
+	lwr	$s1,4+$LSB($inp)
+	lwr	$s2,8+$LSB($inp)
+	lwr	$s3,12+$LSB($inp)
+
+	bal	_mips_AES_encrypt
+
+	swr	$s0,0+$LSB($out)
+	swr	$s1,4+$LSB($out)
+	swr	$s2,8+$LSB($out)
+	swr	$s3,12+$LSB($out)
+	swl	$s0,0+$MSB($out)
+	swl	$s1,4+$MSB($out)
+	swl	$s2,8+$MSB($out)
+	swl	$s3,12+$MSB($out)
+
+	.set	noreorder
+	$REG_L	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_L	$fp,$FRAMESIZE-2*$SZREG($sp)
+	$REG_L	$s11,$FRAMESIZE-3*$SZREG($sp)
+	$REG_L	$s10,$FRAMESIZE-4*$SZREG($sp)
+	$REG_L	$s9,$FRAMESIZE-5*$SZREG($sp)
+	$REG_L	$s8,$FRAMESIZE-6*$SZREG($sp)
+	$REG_L	$s7,$FRAMESIZE-7*$SZREG($sp)
+	$REG_L	$s6,$FRAMESIZE-8*$SZREG($sp)
+	$REG_L	$s5,$FRAMESIZE-9*$SZREG($sp)
+	$REG_L	$s4,$FRAMESIZE-10*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	\$15,$FRAMESIZE-11*$SZREG($sp)
+	$REG_L	\$14,$FRAMESIZE-12*$SZREG($sp)
+	$REG_L	\$13,$FRAMESIZE-13*$SZREG($sp)
+	$REG_L	\$12,$FRAMESIZE-14*$SZREG($sp)
+	$REG_L	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___;
+	jr	$ra
+	$PTR_ADD $sp,$FRAMESIZE
+.end	AES_encrypt
+___
+
+$code.=<<___;
+.align	5
+.ent	_mips_AES_decrypt
+_mips_AES_decrypt:
+	.frame	$sp,0,$ra
+	.set	reorder
+	lw	$t0,0($key)
+	lw	$t1,4($key)
+	lw	$t2,8($key)
+	lw	$t3,12($key)
+	lw	$cnt,240($key)
+	$PTR_ADD $key0,$key,16
+
+	xor	$s0,$t0
+	xor	$s1,$t1
+	xor	$s2,$t2
+	xor	$s3,$t3
+
+	sub	$cnt,1
+	_xtr	$i0,$s3,16-2
+.Loop_dec:
+	_xtr	$i1,$s0,16-2
+	_xtr	$i2,$s1,16-2
+	_xtr	$i3,$s2,16-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lwl	$t0,3($i0)		# Td1[s3>>16]
+	lwl	$t1,3($i1)		# Td1[s0>>16]
+	lwl	$t2,3($i2)		# Td1[s1>>16]
+	lwl	$t3,3($i3)		# Td1[s2>>16]
+	lwr	$t0,2($i0)		# Td1[s3>>16]
+	lwr	$t1,2($i1)		# Td1[s0>>16]
+	lwr	$t2,2($i2)		# Td1[s1>>16]
+	lwr	$t3,2($i3)		# Td1[s2>>16]
+
+	_xtr	$i0,$s2,8-2
+	_xtr	$i1,$s3,8-2
+	_xtr	$i2,$s0,8-2
+	_xtr	$i3,$s1,8-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lwl	$t4,2($i0)		# Td2[s2>>8]
+	lwl	$t5,2($i1)		# Td2[s3>>8]
+	lwl	$t6,2($i2)		# Td2[s0>>8]
+	lwl	$t7,2($i3)		# Td2[s1>>8]
+	lwr	$t4,1($i0)		# Td2[s2>>8]
+	lwr	$t5,1($i1)		# Td2[s3>>8]
+	lwr	$t6,1($i2)		# Td2[s0>>8]
+	lwr	$t7,1($i3)		# Td2[s1>>8]
+
+	_xtr	$i0,$s1,0-2
+	_xtr	$i1,$s2,0-2
+	_xtr	$i2,$s3,0-2
+	_xtr	$i3,$s0,0-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lwl	$t8,1($i0)		# Td3[s1]
+	lwl	$t9,1($i1)		# Td3[s2]
+	lwl	$t10,1($i2)		# Td3[s3]
+	lwl	$t11,1($i3)		# Td3[s0]
+	lwr	$t8,0($i0)		# Td3[s1]
+	lwr	$t9,0($i1)		# Td3[s2]
+	lwr	$t10,0($i2)		# Td3[s3]
+	lwr	$t11,0($i3)		# Td3[s0]
+
+	_xtr	$i0,$s0,24-2
+	_xtr	$i1,$s1,24-2
+	_xtr	$i2,$s2,24-2
+	_xtr	$i3,$s3,24-2
+	and	$i0,0x3fc
+	and	$i1,0x3fc
+	and	$i2,0x3fc
+	and	$i3,0x3fc
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+
+	xor	$t0,$t4
+	xor	$t1,$t5
+	xor	$t2,$t6
+	xor	$t3,$t7
+
+
+	lw	$t4,0($i0)		# Td0[s0>>24]
+	lw	$t5,0($i1)		# Td0[s1>>24]
+	lw	$t6,0($i2)		# Td0[s2>>24]
+	lw	$t7,0($i3)		# Td0[s3>>24]
+
+	lw	$s0,0($key0)
+	lw	$s1,4($key0)
+	lw	$s2,8($key0)
+	lw	$s3,12($key0)
+
+	xor	$t0,$t8
+	xor	$t1,$t9
+	xor	$t2,$t10
+	xor	$t3,$t11
+
+	xor	$t0,$t4
+	xor	$t1,$t5
+	xor	$t2,$t6
+	xor	$t3,$t7
+
+	sub	$cnt,1
+	$PTR_ADD $key0,16
+	xor	$s0,$t0
+	xor	$s1,$t1
+	xor	$s2,$t2
+	xor	$s3,$t3
+	.set	noreorder
+	bnez	$cnt,.Loop_dec
+	_xtr	$i0,$s3,16-2
+
+	.set	reorder
+	lw	$t4,1024($Tbl)		# prefetch Td4
+	lw	$t5,1024+32($Tbl)
+	lw	$t6,1024+64($Tbl)
+	lw	$t7,1024+96($Tbl)
+	lw	$t8,1024+128($Tbl)
+	lw	$t9,1024+160($Tbl)
+	lw	$t10,1024+192($Tbl)
+	lw	$t11,1024+224($Tbl)
+
+	_xtr	$i0,$s3,16
+	_xtr	$i1,$s0,16
+	_xtr	$i2,$s1,16
+	_xtr	$i3,$s2,16
+	and	$i0,0xff
+	and	$i1,0xff
+	and	$i2,0xff
+	and	$i3,0xff
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$t0,1024($i0)		# Td4[s3>>16]
+	lbu	$t1,1024($i1)		# Td4[s0>>16]
+	lbu	$t2,1024($i2)		# Td4[s1>>16]
+	lbu	$t3,1024($i3)		# Td4[s2>>16]
+
+	_xtr	$i0,$s2,8
+	_xtr	$i1,$s3,8
+	_xtr	$i2,$s0,8
+	_xtr	$i3,$s1,8
+	and	$i0,0xff
+	and	$i1,0xff
+	and	$i2,0xff
+	and	$i3,0xff
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$t4,1024($i0)		# Td4[s2>>8]
+	lbu	$t5,1024($i1)		# Td4[s3>>8]
+	lbu	$t6,1024($i2)		# Td4[s0>>8]
+	lbu	$t7,1024($i3)		# Td4[s1>>8]
+
+	_xtr	$i0,$s0,24
+	_xtr	$i1,$s1,24
+	_xtr	$i2,$s2,24
+	_xtr	$i3,$s3,24
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$t8,1024($i0)		# Td4[s0>>24]
+	lbu	$t9,1024($i1)		# Td4[s1>>24]
+	lbu	$t10,1024($i2)		# Td4[s2>>24]
+	lbu	$t11,1024($i3)		# Td4[s3>>24]
+
+	_xtr	$i0,$s1,0
+	_xtr	$i1,$s2,0
+	_xtr	$i2,$s3,0
+	_xtr	$i3,$s0,0
+
+	_ins	$t0,16
+	_ins	$t1,16
+	_ins	$t2,16
+	_ins	$t3,16
+
+	_ins	$t4,8
+	_ins	$t5,8
+	_ins	$t6,8
+	_ins	$t7,8
+
+	xor	$t0,$t4
+	xor	$t1,$t5
+	xor	$t2,$t6
+	xor	$t3,$t7
+
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$t4,1024($i0)		# Td4[s1]
+	lbu	$t5,1024($i1)		# Td4[s2]
+	lbu	$t6,1024($i2)		# Td4[s3]
+	lbu	$t7,1024($i3)		# Td4[s0]
+
+	_ins	$t8,24
+	_ins	$t9,24
+	_ins	$t10,24
+	_ins	$t11,24
+
+	lw	$s0,0($key0)
+	lw	$s1,4($key0)
+	lw	$s2,8($key0)
+	lw	$s3,12($key0)
+
+	_ins	$t4,0
+	_ins	$t5,0
+	_ins	$t6,0
+	_ins	$t7,0
+
+
+	xor	$t0,$t8
+	xor	$t1,$t9
+	xor	$t2,$t10
+	xor	$t3,$t11
+
+	xor	$t0,$t4
+	xor	$t1,$t5
+	xor	$t2,$t6
+	xor	$t3,$t7
+
+	xor	$s0,$t0
+	xor	$s1,$t1
+	xor	$s2,$t2
+	xor	$s3,$t3
+
+	jr	$ra
+.end	_mips_AES_decrypt
+
+.align	5
+.globl	AES_decrypt
+.ent	AES_decrypt
+AES_decrypt:
+	.frame	$sp,$FRAMESIZE,$ra
+	.mask	$SAVED_REGS_MASK,-$SZREG
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /o32/i);	# o32 PIC-ification
+	.cpload	$pf
+___
+$code.=<<___;
+	$PTR_SUB $sp,$FRAMESIZE
+	$REG_S	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_S	$fp,$FRAMESIZE-2*$SZREG($sp)
+	$REG_S	$s11,$FRAMESIZE-3*$SZREG($sp)
+	$REG_S	$s10,$FRAMESIZE-4*$SZREG($sp)
+	$REG_S	$s9,$FRAMESIZE-5*$SZREG($sp)
+	$REG_S	$s8,$FRAMESIZE-6*$SZREG($sp)
+	$REG_S	$s7,$FRAMESIZE-7*$SZREG($sp)
+	$REG_S	$s6,$FRAMESIZE-8*$SZREG($sp)
+	$REG_S	$s5,$FRAMESIZE-9*$SZREG($sp)
+	$REG_S	$s4,$FRAMESIZE-10*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);	# optimize non-nubi prologue
+	$REG_S	\$15,$FRAMESIZE-11*$SZREG($sp)
+	$REG_S	\$14,$FRAMESIZE-12*$SZREG($sp)
+	$REG_S	\$13,$FRAMESIZE-13*$SZREG($sp)
+	$REG_S	\$12,$FRAMESIZE-14*$SZREG($sp)
+	$REG_S	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___ if ($flavour !~ /o32/i);	# non-o32 PIC-ification
+	.cplocal	$Tbl
+	.cpsetup	$pf,$zero,AES_decrypt
+___
+$code.=<<___;
+	.set	reorder
+	la	$Tbl,AES_Td		# PIC-ified 'load address'
+
+	lwl	$s0,0+$MSB($inp)
+	lwl	$s1,4+$MSB($inp)
+	lwl	$s2,8+$MSB($inp)
+	lwl	$s3,12+$MSB($inp)
+	lwr	$s0,0+$LSB($inp)
+	lwr	$s1,4+$LSB($inp)
+	lwr	$s2,8+$LSB($inp)
+	lwr	$s3,12+$LSB($inp)
+
+	bal	_mips_AES_decrypt
+
+	swr	$s0,0+$LSB($out)
+	swr	$s1,4+$LSB($out)
+	swr	$s2,8+$LSB($out)
+	swr	$s3,12+$LSB($out)
+	swl	$s0,0+$MSB($out)
+	swl	$s1,4+$MSB($out)
+	swl	$s2,8+$MSB($out)
+	swl	$s3,12+$MSB($out)
+
+	.set	noreorder
+	$REG_L	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_L	$fp,$FRAMESIZE-2*$SZREG($sp)
+	$REG_L	$s11,$FRAMESIZE-3*$SZREG($sp)
+	$REG_L	$s10,$FRAMESIZE-4*$SZREG($sp)
+	$REG_L	$s9,$FRAMESIZE-5*$SZREG($sp)
+	$REG_L	$s8,$FRAMESIZE-6*$SZREG($sp)
+	$REG_L	$s7,$FRAMESIZE-7*$SZREG($sp)
+	$REG_L	$s6,$FRAMESIZE-8*$SZREG($sp)
+	$REG_L	$s5,$FRAMESIZE-9*$SZREG($sp)
+	$REG_L	$s4,$FRAMESIZE-10*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	\$15,$FRAMESIZE-11*$SZREG($sp)
+	$REG_L	\$14,$FRAMESIZE-12*$SZREG($sp)
+	$REG_L	\$13,$FRAMESIZE-13*$SZREG($sp)
+	$REG_L	\$12,$FRAMESIZE-14*$SZREG($sp)
+	$REG_L	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___;
+	jr	$ra
+	$PTR_ADD $sp,$FRAMESIZE
+.end	AES_decrypt
+___
+}}}
+
+{{{
+my $FRAMESIZE=8*$SZREG;
+my $SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0xc000f008 : 0xc0000000;
+
+my ($inp,$bits,$key,$Tbl)=($a0,$a1,$a2,$a3);
+my ($rk0,$rk1,$rk2,$rk3,$rk4,$rk5,$rk6,$rk7)=($a4,$a5,$a6,$a7,$s0,$s1,$s2,$s3);
+my ($i0,$i1,$i2,$i3)=($at,$t0,$t1,$t2);
+my ($rcon,$cnt)=($gp,$fp);
+
+$code.=<<___;
+.align	5
+.ent	_mips_AES_set_encrypt_key
+_mips_AES_set_encrypt_key:
+	.frame	$sp,0,$ra
+	.set	noreorder
+	beqz	$inp,.Lekey_done
+	li	$t0,-1
+	beqz	$key,.Lekey_done
+	$PTR_ADD $rcon,$Tbl,1024+256
+
+	.set	reorder
+	lwl	$rk0,0+$MSB($inp)	# load 128 bits
+	lwl	$rk1,4+$MSB($inp)
+	lwl	$rk2,8+$MSB($inp)
+	lwl	$rk3,12+$MSB($inp)
+	li	$at,128
+	lwr	$rk0,0+$LSB($inp)
+	lwr	$rk1,4+$LSB($inp)
+	lwr	$rk2,8+$LSB($inp)
+	lwr	$rk3,12+$LSB($inp)
+	.set	noreorder
+	beq	$bits,$at,.L128bits
+	li	$cnt,10
+
+	.set	reorder
+	lwl	$rk4,16+$MSB($inp)	# load 192 bits
+	lwl	$rk5,20+$MSB($inp)
+	li	$at,192
+	lwr	$rk4,16+$LSB($inp)
+	lwr	$rk5,20+$LSB($inp)
+	.set	noreorder
+	beq	$bits,$at,.L192bits
+	li	$cnt,8
+
+	.set	reorder
+	lwl	$rk6,24+$MSB($inp)	# load 256 bits
+	lwl	$rk7,28+$MSB($inp)
+	li	$at,256
+	lwr	$rk6,24+$LSB($inp)
+	lwr	$rk7,28+$LSB($inp)
+	.set	noreorder
+	beq	$bits,$at,.L256bits
+	li	$cnt,7
+
+	b	.Lekey_done
+	li	$t0,-2
+
+.align	4
+.L128bits:
+	.set	reorder
+	srl	$i0,$rk3,16
+	srl	$i1,$rk3,8
+	and	$i0,0xff
+	and	$i1,0xff
+	and	$i2,$rk3,0xff
+	srl	$i3,$rk3,24
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$i0,1024($i0)
+	lbu	$i1,1024($i1)
+	lbu	$i2,1024($i2)
+	lbu	$i3,1024($i3)
+
+	sw	$rk0,0($key)
+	sw	$rk1,4($key)
+	sw	$rk2,8($key)
+	sw	$rk3,12($key)
+	sub	$cnt,1
+	$PTR_ADD $key,16
+
+	_bias	$i0,24
+	_bias	$i1,16
+	_bias	$i2,8
+	_bias	$i3,0
+
+	xor	$rk0,$i0
+	lw	$i0,0($rcon)
+	xor	$rk0,$i1
+	xor	$rk0,$i2
+	xor	$rk0,$i3
+	xor	$rk0,$i0
+
+	xor	$rk1,$rk0
+	xor	$rk2,$rk1
+	xor	$rk3,$rk2
+
+	.set	noreorder
+	bnez	$cnt,.L128bits
+	$PTR_ADD $rcon,4
+
+	sw	$rk0,0($key)
+	sw	$rk1,4($key)
+	sw	$rk2,8($key)
+	li	$cnt,10
+	sw	$rk3,12($key)
+	li	$t0,0
+	sw	$cnt,80($key)
+	b	.Lekey_done
+	$PTR_SUB $key,10*16
+
+.align	4
+.L192bits:
+	.set	reorder
+	srl	$i0,$rk5,16
+	srl	$i1,$rk5,8
+	and	$i0,0xff
+	and	$i1,0xff
+	and	$i2,$rk5,0xff
+	srl	$i3,$rk5,24
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$i0,1024($i0)
+	lbu	$i1,1024($i1)
+	lbu	$i2,1024($i2)
+	lbu	$i3,1024($i3)
+
+	sw	$rk0,0($key)
+	sw	$rk1,4($key)
+	sw	$rk2,8($key)
+	sw	$rk3,12($key)
+	sw	$rk4,16($key)
+	sw	$rk5,20($key)
+	sub	$cnt,1
+	$PTR_ADD $key,24
+
+	_bias	$i0,24
+	_bias	$i1,16
+	_bias	$i2,8
+	_bias	$i3,0
+
+	xor	$rk0,$i0
+	lw	$i0,0($rcon)
+	xor	$rk0,$i1
+	xor	$rk0,$i2
+	xor	$rk0,$i3
+	xor	$rk0,$i0
+
+	xor	$rk1,$rk0
+	xor	$rk2,$rk1
+	xor	$rk3,$rk2
+	xor	$rk4,$rk3
+	xor	$rk5,$rk4
+
+	.set	noreorder
+	bnez	$cnt,.L192bits
+	$PTR_ADD $rcon,4
+
+	sw	$rk0,0($key)
+	sw	$rk1,4($key)
+	sw	$rk2,8($key)
+	li	$cnt,12
+	sw	$rk3,12($key)
+	li	$t0,0
+	sw	$cnt,48($key)
+	b	.Lekey_done
+	$PTR_SUB $key,12*16
+
+.align	4
+.L256bits:
+	.set	reorder
+	srl	$i0,$rk7,16
+	srl	$i1,$rk7,8
+	and	$i0,0xff
+	and	$i1,0xff
+	and	$i2,$rk7,0xff
+	srl	$i3,$rk7,24
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$i0,1024($i0)
+	lbu	$i1,1024($i1)
+	lbu	$i2,1024($i2)
+	lbu	$i3,1024($i3)
+
+	sw	$rk0,0($key)
+	sw	$rk1,4($key)
+	sw	$rk2,8($key)
+	sw	$rk3,12($key)
+	sw	$rk4,16($key)
+	sw	$rk5,20($key)
+	sw	$rk6,24($key)
+	sw	$rk7,28($key)
+	sub	$cnt,1
+
+	_bias	$i0,24
+	_bias	$i1,16
+	_bias	$i2,8
+	_bias	$i3,0
+
+	xor	$rk0,$i0
+	lw	$i0,0($rcon)
+	xor	$rk0,$i1
+	xor	$rk0,$i2
+	xor	$rk0,$i3
+	xor	$rk0,$i0
+
+	xor	$rk1,$rk0
+	xor	$rk2,$rk1
+	xor	$rk3,$rk2
+	beqz	$cnt,.L256bits_done
+
+	srl	$i0,$rk3,24
+	srl	$i1,$rk3,16
+	srl	$i2,$rk3,8
+	and	$i3,$rk3,0xff
+	and	$i1,0xff
+	and	$i2,0xff
+	$PTR_ADD $i0,$Tbl
+	$PTR_ADD $i1,$Tbl
+	$PTR_ADD $i2,$Tbl
+	$PTR_ADD $i3,$Tbl
+	lbu	$i0,1024($i0)
+	lbu	$i1,1024($i1)
+	lbu	$i2,1024($i2)
+	lbu	$i3,1024($i3)
+	sll	$i0,24
+	sll	$i1,16
+	sll	$i2,8
+
+	xor	$rk4,$i0
+	xor	$rk4,$i1
+	xor	$rk4,$i2
+	xor	$rk4,$i3
+
+	xor	$rk5,$rk4
+	xor	$rk6,$rk5
+	xor	$rk7,$rk6
+
+	$PTR_ADD $key,32
+	.set	noreorder
+	b	.L256bits
+	$PTR_ADD $rcon,4
+
+.L256bits_done:
+	sw	$rk0,32($key)
+	sw	$rk1,36($key)
+	sw	$rk2,40($key)
+	li	$cnt,14
+	sw	$rk3,44($key)
+	li	$t0,0
+	sw	$cnt,48($key)
+	$PTR_SUB $key,12*16
+
+.Lekey_done:
+	jr	$ra
+	nop
+.end	_mips_AES_set_encrypt_key
+
+.globl	AES_set_encrypt_key
+.ent	AES_set_encrypt_key
+AES_set_encrypt_key:
+	.frame	$sp,$FRAMESIZE,$ra
+	.mask	$SAVED_REGS_MASK,-$SZREG
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /o32/i);	# o32 PIC-ification
+	.cpload	$pf
+___
+$code.=<<___;
+	$PTR_SUB $sp,$FRAMESIZE
+	$REG_S	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_S	$fp,$FRAMESIZE-2*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);	# optimize non-nubi prologue
+	$REG_S	$s3,$FRAMESIZE-3*$SZREG($sp)
+	$REG_S	$s2,$FRAMESIZE-4*$SZREG($sp)
+	$REG_S	$s1,$FRAMESIZE-5*$SZREG($sp)
+	$REG_S	$s0,$FRAMESIZE-6*$SZREG($sp)
+	$REG_S	$gp,$FRAMESIZE-7*$SZREG($sp)
+___
+$code.=<<___ if ($flavour !~ /o32/i);	# non-o32 PIC-ification
+	.cplocal	$Tbl
+	.cpsetup	$pf,$zero,AES_set_encrypt_key
+___
+$code.=<<___;
+	.set	reorder
+	la	$Tbl,AES_Te		# PIC-ified 'load address'
+
+	bal	_mips_AES_set_encrypt_key
+
+	.set	noreorder
+	move	$a0,$t0
+	$REG_L	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_L	$fp,$FRAMESIZE-2*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$s3,$FRAMESIZE-11*$SZREG($sp)
+	$REG_L	$s2,$FRAMESIZE-12*$SZREG($sp)
+	$REG_L	$s1,$FRAMESIZE-13*$SZREG($sp)
+	$REG_L	$s0,$FRAMESIZE-14*$SZREG($sp)
+	$REG_L	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___;
+	jr	$ra
+	$PTR_ADD $sp,$FRAMESIZE
+.end	AES_set_encrypt_key
+___
+
+my ($head,$tail)=($inp,$bits);
+my ($tp1,$tp2,$tp4,$tp8,$tp9,$tpb,$tpd,$tpe)=($a4,$a5,$a6,$a7,$s0,$s1,$s2,$s3);
+my ($m,$x80808080,$x7f7f7f7f,$x1b1b1b1b)=($at,$t0,$t1,$t2);
+$code.=<<___;
+.align	5
+.globl	AES_set_decrypt_key
+.ent	AES_set_decrypt_key
+AES_set_decrypt_key:
+	.frame	$sp,$FRAMESIZE,$ra
+	.mask	$SAVED_REGS_MASK,-$SZREG
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /o32/i);	# o32 PIC-ification
+	.cpload	$pf
+___
+$code.=<<___;
+	$PTR_SUB $sp,$FRAMESIZE
+	$REG_S	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_S	$fp,$FRAMESIZE-2*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);	# optimize non-nubi prologue
+	$REG_S	$s3,$FRAMESIZE-3*$SZREG($sp)
+	$REG_S	$s2,$FRAMESIZE-4*$SZREG($sp)
+	$REG_S	$s1,$FRAMESIZE-5*$SZREG($sp)
+	$REG_S	$s0,$FRAMESIZE-6*$SZREG($sp)
+	$REG_S	$gp,$FRAMESIZE-7*$SZREG($sp)
+___
+$code.=<<___ if ($flavour !~ /o32/i);	# non-o32 PIC-ification
+	.cplocal	$Tbl
+	.cpsetup	$pf,$zero,AES_set_decrypt_key
+___
+$code.=<<___;
+	.set	reorder
+	la	$Tbl,AES_Te		# PIC-ified 'load address'
+
+	bal	_mips_AES_set_encrypt_key
+
+	bltz	$t0,.Ldkey_done
+
+	sll	$at,$cnt,4
+	$PTR_ADD $head,$key,0
+	$PTR_ADD $tail,$key,$at
+.align	4
+.Lswap:
+	lw	$rk0,0($head)
+	lw	$rk1,4($head)
+	lw	$rk2,8($head)
+	lw	$rk3,12($head)
+	lw	$rk4,0($tail)
+	lw	$rk5,4($tail)
+	lw	$rk6,8($tail)
+	lw	$rk7,12($tail)
+	sw	$rk0,0($tail)
+	sw	$rk1,4($tail)
+	sw	$rk2,8($tail)
+	sw	$rk3,12($tail)
+	$PTR_ADD $head,16
+	$PTR_SUB $tail,16
+	sw	$rk4,-16($head)
+	sw	$rk5,-12($head)
+	sw	$rk6,-8($head)
+	sw	$rk7,-4($head)
+	bne	$head,$tail,.Lswap
+
+	lw	$tp1,16($key)		# modulo-scheduled
+	lui	$x80808080,0x8080
+	sub	$cnt,1
+	or	$x80808080,0x8080
+	sll	$cnt,2
+	$PTR_ADD $key,16
+	lui	$x1b1b1b1b,0x1b1b
+	nor	$x7f7f7f7f,$zero,$x80808080
+	or	$x1b1b1b1b,0x1b1b
+.align	4
+.Lmix:
+	and	$m,$tp1,$x80808080
+	and	$tp2,$tp1,$x7f7f7f7f
+	srl	$tp4,$m,7
+	addu	$tp2,$tp2		# tp2<<1
+	subu	$m,$tp4
+	and	$m,$x1b1b1b1b
+	xor	$tp2,$m
+
+	and	$m,$tp2,$x80808080
+	and	$tp4,$tp2,$x7f7f7f7f
+	srl	$tp8,$m,7
+	addu	$tp4,$tp4		# tp4<<1
+	subu	$m,$tp8
+	and	$m,$x1b1b1b1b
+	xor	$tp4,$m
+
+	and	$m,$tp4,$x80808080
+	and	$tp8,$tp4,$x7f7f7f7f
+	srl	$tp9,$m,7
+	addu	$tp8,$tp8		# tp8<<1
+	subu	$m,$tp9
+	and	$m,$x1b1b1b1b
+	xor	$tp8,$m
+
+	xor	$tp9,$tp8,$tp1
+	xor	$tpe,$tp8,$tp4
+	xor	$tpb,$tp9,$tp2
+	xor	$tpd,$tp9,$tp4
+
+	_ror	$tp1,$tpd,16
+	 xor	$tpe,$tp2
+	_ror	$tp2,$tpd,-16
+	xor	$tpe,$tp1
+	_ror	$tp1,$tp9,8
+	xor	$tpe,$tp2
+	_ror	$tp2,$tp9,-24
+	xor	$tpe,$tp1
+	_ror	$tp1,$tpb,24
+	xor	$tpe,$tp2
+	_ror	$tp2,$tpb,-8
+	xor	$tpe,$tp1
+	lw	$tp1,4($key)		# modulo-scheduled
+	xor	$tpe,$tp2
+	sub	$cnt,1
+	sw	$tpe,0($key)
+	$PTR_ADD $key,4
+	bnez	$cnt,.Lmix
+
+	li	$t0,0
+.Ldkey_done:
+	.set	noreorder
+	move	$a0,$t0
+	$REG_L	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_L	$fp,$FRAMESIZE-2*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$s3,$FRAMESIZE-11*$SZREG($sp)
+	$REG_L	$s2,$FRAMESIZE-12*$SZREG($sp)
+	$REG_L	$s1,$FRAMESIZE-13*$SZREG($sp)
+	$REG_L	$s0,$FRAMESIZE-14*$SZREG($sp)
+	$REG_L	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___;
+	jr	$ra
+	$PTR_ADD $sp,$FRAMESIZE
+.end	AES_set_decrypt_key
+___
+}}}
+
+######################################################################
+# Tables are kept in endian-neutral manner
+$code.=<<___;
+.rdata
+.align	6
+AES_Te:
+.byte	0xc6,0x63,0x63,0xa5,	0xf8,0x7c,0x7c,0x84	# Te0
+.byte	0xee,0x77,0x77,0x99,	0xf6,0x7b,0x7b,0x8d
+.byte	0xff,0xf2,0xf2,0x0d,	0xd6,0x6b,0x6b,0xbd
+.byte	0xde,0x6f,0x6f,0xb1,	0x91,0xc5,0xc5,0x54
+.byte	0x60,0x30,0x30,0x50,	0x02,0x01,0x01,0x03
+.byte	0xce,0x67,0x67,0xa9,	0x56,0x2b,0x2b,0x7d
+.byte	0xe7,0xfe,0xfe,0x19,	0xb5,0xd7,0xd7,0x62
+.byte	0x4d,0xab,0xab,0xe6,	0xec,0x76,0x76,0x9a
+.byte	0x8f,0xca,0xca,0x45,	0x1f,0x82,0x82,0x9d
+.byte	0x89,0xc9,0xc9,0x40,	0xfa,0x7d,0x7d,0x87
+.byte	0xef,0xfa,0xfa,0x15,	0xb2,0x59,0x59,0xeb
+.byte	0x8e,0x47,0x47,0xc9,	0xfb,0xf0,0xf0,0x0b
+.byte	0x41,0xad,0xad,0xec,	0xb3,0xd4,0xd4,0x67
+.byte	0x5f,0xa2,0xa2,0xfd,	0x45,0xaf,0xaf,0xea
+.byte	0x23,0x9c,0x9c,0xbf,	0x53,0xa4,0xa4,0xf7
+.byte	0xe4,0x72,0x72,0x96,	0x9b,0xc0,0xc0,0x5b
+.byte	0x75,0xb7,0xb7,0xc2,	0xe1,0xfd,0xfd,0x1c
+.byte	0x3d,0x93,0x93,0xae,	0x4c,0x26,0x26,0x6a
+.byte	0x6c,0x36,0x36,0x5a,	0x7e,0x3f,0x3f,0x41
+.byte	0xf5,0xf7,0xf7,0x02,	0x83,0xcc,0xcc,0x4f
+.byte	0x68,0x34,0x34,0x5c,	0x51,0xa5,0xa5,0xf4
+.byte	0xd1,0xe5,0xe5,0x34,	0xf9,0xf1,0xf1,0x08
+.byte	0xe2,0x71,0x71,0x93,	0xab,0xd8,0xd8,0x73
+.byte	0x62,0x31,0x31,0x53,	0x2a,0x15,0x15,0x3f
+.byte	0x08,0x04,0x04,0x0c,	0x95,0xc7,0xc7,0x52
+.byte	0x46,0x23,0x23,0x65,	0x9d,0xc3,0xc3,0x5e
+.byte	0x30,0x18,0x18,0x28,	0x37,0x96,0x96,0xa1
+.byte	0x0a,0x05,0x05,0x0f,	0x2f,0x9a,0x9a,0xb5
+.byte	0x0e,0x07,0x07,0x09,	0x24,0x12,0x12,0x36
+.byte	0x1b,0x80,0x80,0x9b,	0xdf,0xe2,0xe2,0x3d
+.byte	0xcd,0xeb,0xeb,0x26,	0x4e,0x27,0x27,0x69
+.byte	0x7f,0xb2,0xb2,0xcd,	0xea,0x75,0x75,0x9f
+.byte	0x12,0x09,0x09,0x1b,	0x1d,0x83,0x83,0x9e
+.byte	0x58,0x2c,0x2c,0x74,	0x34,0x1a,0x1a,0x2e
+.byte	0x36,0x1b,0x1b,0x2d,	0xdc,0x6e,0x6e,0xb2
+.byte	0xb4,0x5a,0x5a,0xee,	0x5b,0xa0,0xa0,0xfb
+.byte	0xa4,0x52,0x52,0xf6,	0x76,0x3b,0x3b,0x4d
+.byte	0xb7,0xd6,0xd6,0x61,	0x7d,0xb3,0xb3,0xce
+.byte	0x52,0x29,0x29,0x7b,	0xdd,0xe3,0xe3,0x3e
+.byte	0x5e,0x2f,0x2f,0x71,	0x13,0x84,0x84,0x97
+.byte	0xa6,0x53,0x53,0xf5,	0xb9,0xd1,0xd1,0x68
+.byte	0x00,0x00,0x00,0x00,	0xc1,0xed,0xed,0x2c
+.byte	0x40,0x20,0x20,0x60,	0xe3,0xfc,0xfc,0x1f
+.byte	0x79,0xb1,0xb1,0xc8,	0xb6,0x5b,0x5b,0xed
+.byte	0xd4,0x6a,0x6a,0xbe,	0x8d,0xcb,0xcb,0x46
+.byte	0x67,0xbe,0xbe,0xd9,	0x72,0x39,0x39,0x4b
+.byte	0x94,0x4a,0x4a,0xde,	0x98,0x4c,0x4c,0xd4
+.byte	0xb0,0x58,0x58,0xe8,	0x85,0xcf,0xcf,0x4a
+.byte	0xbb,0xd0,0xd0,0x6b,	0xc5,0xef,0xef,0x2a
+.byte	0x4f,0xaa,0xaa,0xe5,	0xed,0xfb,0xfb,0x16
+.byte	0x86,0x43,0x43,0xc5,	0x9a,0x4d,0x4d,0xd7
+.byte	0x66,0x33,0x33,0x55,	0x11,0x85,0x85,0x94
+.byte	0x8a,0x45,0x45,0xcf,	0xe9,0xf9,0xf9,0x10
+.byte	0x04,0x02,0x02,0x06,	0xfe,0x7f,0x7f,0x81
+.byte	0xa0,0x50,0x50,0xf0,	0x78,0x3c,0x3c,0x44
+.byte	0x25,0x9f,0x9f,0xba,	0x4b,0xa8,0xa8,0xe3
+.byte	0xa2,0x51,0x51,0xf3,	0x5d,0xa3,0xa3,0xfe
+.byte	0x80,0x40,0x40,0xc0,	0x05,0x8f,0x8f,0x8a
+.byte	0x3f,0x92,0x92,0xad,	0x21,0x9d,0x9d,0xbc
+.byte	0x70,0x38,0x38,0x48,	0xf1,0xf5,0xf5,0x04
+.byte	0x63,0xbc,0xbc,0xdf,	0x77,0xb6,0xb6,0xc1
+.byte	0xaf,0xda,0xda,0x75,	0x42,0x21,0x21,0x63
+.byte	0x20,0x10,0x10,0x30,	0xe5,0xff,0xff,0x1a
+.byte	0xfd,0xf3,0xf3,0x0e,	0xbf,0xd2,0xd2,0x6d
+.byte	0x81,0xcd,0xcd,0x4c,	0x18,0x0c,0x0c,0x14
+.byte	0x26,0x13,0x13,0x35,	0xc3,0xec,0xec,0x2f
+.byte	0xbe,0x5f,0x5f,0xe1,	0x35,0x97,0x97,0xa2
+.byte	0x88,0x44,0x44,0xcc,	0x2e,0x17,0x17,0x39
+.byte	0x93,0xc4,0xc4,0x57,	0x55,0xa7,0xa7,0xf2
+.byte	0xfc,0x7e,0x7e,0x82,	0x7a,0x3d,0x3d,0x47
+.byte	0xc8,0x64,0x64,0xac,	0xba,0x5d,0x5d,0xe7
+.byte	0x32,0x19,0x19,0x2b,	0xe6,0x73,0x73,0x95
+.byte	0xc0,0x60,0x60,0xa0,	0x19,0x81,0x81,0x98
+.byte	0x9e,0x4f,0x4f,0xd1,	0xa3,0xdc,0xdc,0x7f
+.byte	0x44,0x22,0x22,0x66,	0x54,0x2a,0x2a,0x7e
+.byte	0x3b,0x90,0x90,0xab,	0x0b,0x88,0x88,0x83
+.byte	0x8c,0x46,0x46,0xca,	0xc7,0xee,0xee,0x29
+.byte	0x6b,0xb8,0xb8,0xd3,	0x28,0x14,0x14,0x3c
+.byte	0xa7,0xde,0xde,0x79,	0xbc,0x5e,0x5e,0xe2
+.byte	0x16,0x0b,0x0b,0x1d,	0xad,0xdb,0xdb,0x76
+.byte	0xdb,0xe0,0xe0,0x3b,	0x64,0x32,0x32,0x56
+.byte	0x74,0x3a,0x3a,0x4e,	0x14,0x0a,0x0a,0x1e
+.byte	0x92,0x49,0x49,0xdb,	0x0c,0x06,0x06,0x0a
+.byte	0x48,0x24,0x24,0x6c,	0xb8,0x5c,0x5c,0xe4
+.byte	0x9f,0xc2,0xc2,0x5d,	0xbd,0xd3,0xd3,0x6e
+.byte	0x43,0xac,0xac,0xef,	0xc4,0x62,0x62,0xa6
+.byte	0x39,0x91,0x91,0xa8,	0x31,0x95,0x95,0xa4
+.byte	0xd3,0xe4,0xe4,0x37,	0xf2,0x79,0x79,0x8b
+.byte	0xd5,0xe7,0xe7,0x32,	0x8b,0xc8,0xc8,0x43
+.byte	0x6e,0x37,0x37,0x59,	0xda,0x6d,0x6d,0xb7
+.byte	0x01,0x8d,0x8d,0x8c,	0xb1,0xd5,0xd5,0x64
+.byte	0x9c,0x4e,0x4e,0xd2,	0x49,0xa9,0xa9,0xe0
+.byte	0xd8,0x6c,0x6c,0xb4,	0xac,0x56,0x56,0xfa
+.byte	0xf3,0xf4,0xf4,0x07,	0xcf,0xea,0xea,0x25
+.byte	0xca,0x65,0x65,0xaf,	0xf4,0x7a,0x7a,0x8e
+.byte	0x47,0xae,0xae,0xe9,	0x10,0x08,0x08,0x18
+.byte	0x6f,0xba,0xba,0xd5,	0xf0,0x78,0x78,0x88
+.byte	0x4a,0x25,0x25,0x6f,	0x5c,0x2e,0x2e,0x72
+.byte	0x38,0x1c,0x1c,0x24,	0x57,0xa6,0xa6,0xf1
+.byte	0x73,0xb4,0xb4,0xc7,	0x97,0xc6,0xc6,0x51
+.byte	0xcb,0xe8,0xe8,0x23,	0xa1,0xdd,0xdd,0x7c
+.byte	0xe8,0x74,0x74,0x9c,	0x3e,0x1f,0x1f,0x21
+.byte	0x96,0x4b,0x4b,0xdd,	0x61,0xbd,0xbd,0xdc
+.byte	0x0d,0x8b,0x8b,0x86,	0x0f,0x8a,0x8a,0x85
+.byte	0xe0,0x70,0x70,0x90,	0x7c,0x3e,0x3e,0x42
+.byte	0x71,0xb5,0xb5,0xc4,	0xcc,0x66,0x66,0xaa
+.byte	0x90,0x48,0x48,0xd8,	0x06,0x03,0x03,0x05
+.byte	0xf7,0xf6,0xf6,0x01,	0x1c,0x0e,0x0e,0x12
+.byte	0xc2,0x61,0x61,0xa3,	0x6a,0x35,0x35,0x5f
+.byte	0xae,0x57,0x57,0xf9,	0x69,0xb9,0xb9,0xd0
+.byte	0x17,0x86,0x86,0x91,	0x99,0xc1,0xc1,0x58
+.byte	0x3a,0x1d,0x1d,0x27,	0x27,0x9e,0x9e,0xb9
+.byte	0xd9,0xe1,0xe1,0x38,	0xeb,0xf8,0xf8,0x13
+.byte	0x2b,0x98,0x98,0xb3,	0x22,0x11,0x11,0x33
+.byte	0xd2,0x69,0x69,0xbb,	0xa9,0xd9,0xd9,0x70
+.byte	0x07,0x8e,0x8e,0x89,	0x33,0x94,0x94,0xa7
+.byte	0x2d,0x9b,0x9b,0xb6,	0x3c,0x1e,0x1e,0x22
+.byte	0x15,0x87,0x87,0x92,	0xc9,0xe9,0xe9,0x20
+.byte	0x87,0xce,0xce,0x49,	0xaa,0x55,0x55,0xff
+.byte	0x50,0x28,0x28,0x78,	0xa5,0xdf,0xdf,0x7a
+.byte	0x03,0x8c,0x8c,0x8f,	0x59,0xa1,0xa1,0xf8
+.byte	0x09,0x89,0x89,0x80,	0x1a,0x0d,0x0d,0x17
+.byte	0x65,0xbf,0xbf,0xda,	0xd7,0xe6,0xe6,0x31
+.byte	0x84,0x42,0x42,0xc6,	0xd0,0x68,0x68,0xb8
+.byte	0x82,0x41,0x41,0xc3,	0x29,0x99,0x99,0xb0
+.byte	0x5a,0x2d,0x2d,0x77,	0x1e,0x0f,0x0f,0x11
+.byte	0x7b,0xb0,0xb0,0xcb,	0xa8,0x54,0x54,0xfc
+.byte	0x6d,0xbb,0xbb,0xd6,	0x2c,0x16,0x16,0x3a
+
+.byte	0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5	# Te4
+.byte	0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76
+.byte	0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0
+.byte	0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0
+.byte	0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc
+.byte	0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15
+.byte	0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a
+.byte	0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75
+.byte	0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0
+.byte	0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84
+.byte	0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b
+.byte	0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf
+.byte	0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85
+.byte	0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8
+.byte	0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5
+.byte	0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2
+.byte	0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17
+.byte	0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73
+.byte	0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88
+.byte	0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb
+.byte	0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c
+.byte	0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79
+.byte	0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9
+.byte	0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08
+.byte	0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6
+.byte	0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a
+.byte	0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e
+.byte	0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e
+.byte	0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94
+.byte	0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf
+.byte	0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68
+.byte	0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16
+
+.byte	0x01,0x00,0x00,0x00,	0x02,0x00,0x00,0x00	# rcon
+.byte	0x04,0x00,0x00,0x00,	0x08,0x00,0x00,0x00
+.byte	0x10,0x00,0x00,0x00,	0x20,0x00,0x00,0x00
+.byte	0x40,0x00,0x00,0x00,	0x80,0x00,0x00,0x00
+.byte	0x1B,0x00,0x00,0x00,	0x36,0x00,0x00,0x00
+
+.align	6
+AES_Td:
+.byte	0x51,0xf4,0xa7,0x50,	0x7e,0x41,0x65,0x53	# Td0
+.byte	0x1a,0x17,0xa4,0xc3,	0x3a,0x27,0x5e,0x96
+.byte	0x3b,0xab,0x6b,0xcb,	0x1f,0x9d,0x45,0xf1
+.byte	0xac,0xfa,0x58,0xab,	0x4b,0xe3,0x03,0x93
+.byte	0x20,0x30,0xfa,0x55,	0xad,0x76,0x6d,0xf6
+.byte	0x88,0xcc,0x76,0x91,	0xf5,0x02,0x4c,0x25
+.byte	0x4f,0xe5,0xd7,0xfc,	0xc5,0x2a,0xcb,0xd7
+.byte	0x26,0x35,0x44,0x80,	0xb5,0x62,0xa3,0x8f
+.byte	0xde,0xb1,0x5a,0x49,	0x25,0xba,0x1b,0x67
+.byte	0x45,0xea,0x0e,0x98,	0x5d,0xfe,0xc0,0xe1
+.byte	0xc3,0x2f,0x75,0x02,	0x81,0x4c,0xf0,0x12
+.byte	0x8d,0x46,0x97,0xa3,	0x6b,0xd3,0xf9,0xc6
+.byte	0x03,0x8f,0x5f,0xe7,	0x15,0x92,0x9c,0x95
+.byte	0xbf,0x6d,0x7a,0xeb,	0x95,0x52,0x59,0xda
+.byte	0xd4,0xbe,0x83,0x2d,	0x58,0x74,0x21,0xd3
+.byte	0x49,0xe0,0x69,0x29,	0x8e,0xc9,0xc8,0x44
+.byte	0x75,0xc2,0x89,0x6a,	0xf4,0x8e,0x79,0x78
+.byte	0x99,0x58,0x3e,0x6b,	0x27,0xb9,0x71,0xdd
+.byte	0xbe,0xe1,0x4f,0xb6,	0xf0,0x88,0xad,0x17
+.byte	0xc9,0x20,0xac,0x66,	0x7d,0xce,0x3a,0xb4
+.byte	0x63,0xdf,0x4a,0x18,	0xe5,0x1a,0x31,0x82
+.byte	0x97,0x51,0x33,0x60,	0x62,0x53,0x7f,0x45
+.byte	0xb1,0x64,0x77,0xe0,	0xbb,0x6b,0xae,0x84
+.byte	0xfe,0x81,0xa0,0x1c,	0xf9,0x08,0x2b,0x94
+.byte	0x70,0x48,0x68,0x58,	0x8f,0x45,0xfd,0x19
+.byte	0x94,0xde,0x6c,0x87,	0x52,0x7b,0xf8,0xb7
+.byte	0xab,0x73,0xd3,0x23,	0x72,0x4b,0x02,0xe2
+.byte	0xe3,0x1f,0x8f,0x57,	0x66,0x55,0xab,0x2a
+.byte	0xb2,0xeb,0x28,0x07,	0x2f,0xb5,0xc2,0x03
+.byte	0x86,0xc5,0x7b,0x9a,	0xd3,0x37,0x08,0xa5
+.byte	0x30,0x28,0x87,0xf2,	0x23,0xbf,0xa5,0xb2
+.byte	0x02,0x03,0x6a,0xba,	0xed,0x16,0x82,0x5c
+.byte	0x8a,0xcf,0x1c,0x2b,	0xa7,0x79,0xb4,0x92
+.byte	0xf3,0x07,0xf2,0xf0,	0x4e,0x69,0xe2,0xa1
+.byte	0x65,0xda,0xf4,0xcd,	0x06,0x05,0xbe,0xd5
+.byte	0xd1,0x34,0x62,0x1f,	0xc4,0xa6,0xfe,0x8a
+.byte	0x34,0x2e,0x53,0x9d,	0xa2,0xf3,0x55,0xa0
+.byte	0x05,0x8a,0xe1,0x32,	0xa4,0xf6,0xeb,0x75
+.byte	0x0b,0x83,0xec,0x39,	0x40,0x60,0xef,0xaa
+.byte	0x5e,0x71,0x9f,0x06,	0xbd,0x6e,0x10,0x51
+.byte	0x3e,0x21,0x8a,0xf9,	0x96,0xdd,0x06,0x3d
+.byte	0xdd,0x3e,0x05,0xae,	0x4d,0xe6,0xbd,0x46
+.byte	0x91,0x54,0x8d,0xb5,	0x71,0xc4,0x5d,0x05
+.byte	0x04,0x06,0xd4,0x6f,	0x60,0x50,0x15,0xff
+.byte	0x19,0x98,0xfb,0x24,	0xd6,0xbd,0xe9,0x97
+.byte	0x89,0x40,0x43,0xcc,	0x67,0xd9,0x9e,0x77
+.byte	0xb0,0xe8,0x42,0xbd,	0x07,0x89,0x8b,0x88
+.byte	0xe7,0x19,0x5b,0x38,	0x79,0xc8,0xee,0xdb
+.byte	0xa1,0x7c,0x0a,0x47,	0x7c,0x42,0x0f,0xe9
+.byte	0xf8,0x84,0x1e,0xc9,	0x00,0x00,0x00,0x00
+.byte	0x09,0x80,0x86,0x83,	0x32,0x2b,0xed,0x48
+.byte	0x1e,0x11,0x70,0xac,	0x6c,0x5a,0x72,0x4e
+.byte	0xfd,0x0e,0xff,0xfb,	0x0f,0x85,0x38,0x56
+.byte	0x3d,0xae,0xd5,0x1e,	0x36,0x2d,0x39,0x27
+.byte	0x0a,0x0f,0xd9,0x64,	0x68,0x5c,0xa6,0x21
+.byte	0x9b,0x5b,0x54,0xd1,	0x24,0x36,0x2e,0x3a
+.byte	0x0c,0x0a,0x67,0xb1,	0x93,0x57,0xe7,0x0f
+.byte	0xb4,0xee,0x96,0xd2,	0x1b,0x9b,0x91,0x9e
+.byte	0x80,0xc0,0xc5,0x4f,	0x61,0xdc,0x20,0xa2
+.byte	0x5a,0x77,0x4b,0x69,	0x1c,0x12,0x1a,0x16
+.byte	0xe2,0x93,0xba,0x0a,	0xc0,0xa0,0x2a,0xe5
+.byte	0x3c,0x22,0xe0,0x43,	0x12,0x1b,0x17,0x1d
+.byte	0x0e,0x09,0x0d,0x0b,	0xf2,0x8b,0xc7,0xad
+.byte	0x2d,0xb6,0xa8,0xb9,	0x14,0x1e,0xa9,0xc8
+.byte	0x57,0xf1,0x19,0x85,	0xaf,0x75,0x07,0x4c
+.byte	0xee,0x99,0xdd,0xbb,	0xa3,0x7f,0x60,0xfd
+.byte	0xf7,0x01,0x26,0x9f,	0x5c,0x72,0xf5,0xbc
+.byte	0x44,0x66,0x3b,0xc5,	0x5b,0xfb,0x7e,0x34
+.byte	0x8b,0x43,0x29,0x76,	0xcb,0x23,0xc6,0xdc
+.byte	0xb6,0xed,0xfc,0x68,	0xb8,0xe4,0xf1,0x63
+.byte	0xd7,0x31,0xdc,0xca,	0x42,0x63,0x85,0x10
+.byte	0x13,0x97,0x22,0x40,	0x84,0xc6,0x11,0x20
+.byte	0x85,0x4a,0x24,0x7d,	0xd2,0xbb,0x3d,0xf8
+.byte	0xae,0xf9,0x32,0x11,	0xc7,0x29,0xa1,0x6d
+.byte	0x1d,0x9e,0x2f,0x4b,	0xdc,0xb2,0x30,0xf3
+.byte	0x0d,0x86,0x52,0xec,	0x77,0xc1,0xe3,0xd0
+.byte	0x2b,0xb3,0x16,0x6c,	0xa9,0x70,0xb9,0x99
+.byte	0x11,0x94,0x48,0xfa,	0x47,0xe9,0x64,0x22
+.byte	0xa8,0xfc,0x8c,0xc4,	0xa0,0xf0,0x3f,0x1a
+.byte	0x56,0x7d,0x2c,0xd8,	0x22,0x33,0x90,0xef
+.byte	0x87,0x49,0x4e,0xc7,	0xd9,0x38,0xd1,0xc1
+.byte	0x8c,0xca,0xa2,0xfe,	0x98,0xd4,0x0b,0x36
+.byte	0xa6,0xf5,0x81,0xcf,	0xa5,0x7a,0xde,0x28
+.byte	0xda,0xb7,0x8e,0x26,	0x3f,0xad,0xbf,0xa4
+.byte	0x2c,0x3a,0x9d,0xe4,	0x50,0x78,0x92,0x0d
+.byte	0x6a,0x5f,0xcc,0x9b,	0x54,0x7e,0x46,0x62
+.byte	0xf6,0x8d,0x13,0xc2,	0x90,0xd8,0xb8,0xe8
+.byte	0x2e,0x39,0xf7,0x5e,	0x82,0xc3,0xaf,0xf5
+.byte	0x9f,0x5d,0x80,0xbe,	0x69,0xd0,0x93,0x7c
+.byte	0x6f,0xd5,0x2d,0xa9,	0xcf,0x25,0x12,0xb3
+.byte	0xc8,0xac,0x99,0x3b,	0x10,0x18,0x7d,0xa7
+.byte	0xe8,0x9c,0x63,0x6e,	0xdb,0x3b,0xbb,0x7b
+.byte	0xcd,0x26,0x78,0x09,	0x6e,0x59,0x18,0xf4
+.byte	0xec,0x9a,0xb7,0x01,	0x83,0x4f,0x9a,0xa8
+.byte	0xe6,0x95,0x6e,0x65,	0xaa,0xff,0xe6,0x7e
+.byte	0x21,0xbc,0xcf,0x08,	0xef,0x15,0xe8,0xe6
+.byte	0xba,0xe7,0x9b,0xd9,	0x4a,0x6f,0x36,0xce
+.byte	0xea,0x9f,0x09,0xd4,	0x29,0xb0,0x7c,0xd6
+.byte	0x31,0xa4,0xb2,0xaf,	0x2a,0x3f,0x23,0x31
+.byte	0xc6,0xa5,0x94,0x30,	0x35,0xa2,0x66,0xc0
+.byte	0x74,0x4e,0xbc,0x37,	0xfc,0x82,0xca,0xa6
+.byte	0xe0,0x90,0xd0,0xb0,	0x33,0xa7,0xd8,0x15
+.byte	0xf1,0x04,0x98,0x4a,	0x41,0xec,0xda,0xf7
+.byte	0x7f,0xcd,0x50,0x0e,	0x17,0x91,0xf6,0x2f
+.byte	0x76,0x4d,0xd6,0x8d,	0x43,0xef,0xb0,0x4d
+.byte	0xcc,0xaa,0x4d,0x54,	0xe4,0x96,0x04,0xdf
+.byte	0x9e,0xd1,0xb5,0xe3,	0x4c,0x6a,0x88,0x1b
+.byte	0xc1,0x2c,0x1f,0xb8,	0x46,0x65,0x51,0x7f
+.byte	0x9d,0x5e,0xea,0x04,	0x01,0x8c,0x35,0x5d
+.byte	0xfa,0x87,0x74,0x73,	0xfb,0x0b,0x41,0x2e
+.byte	0xb3,0x67,0x1d,0x5a,	0x92,0xdb,0xd2,0x52
+.byte	0xe9,0x10,0x56,0x33,	0x6d,0xd6,0x47,0x13
+.byte	0x9a,0xd7,0x61,0x8c,	0x37,0xa1,0x0c,0x7a
+.byte	0x59,0xf8,0x14,0x8e,	0xeb,0x13,0x3c,0x89
+.byte	0xce,0xa9,0x27,0xee,	0xb7,0x61,0xc9,0x35
+.byte	0xe1,0x1c,0xe5,0xed,	0x7a,0x47,0xb1,0x3c
+.byte	0x9c,0xd2,0xdf,0x59,	0x55,0xf2,0x73,0x3f
+.byte	0x18,0x14,0xce,0x79,	0x73,0xc7,0x37,0xbf
+.byte	0x53,0xf7,0xcd,0xea,	0x5f,0xfd,0xaa,0x5b
+.byte	0xdf,0x3d,0x6f,0x14,	0x78,0x44,0xdb,0x86
+.byte	0xca,0xaf,0xf3,0x81,	0xb9,0x68,0xc4,0x3e
+.byte	0x38,0x24,0x34,0x2c,	0xc2,0xa3,0x40,0x5f
+.byte	0x16,0x1d,0xc3,0x72,	0xbc,0xe2,0x25,0x0c
+.byte	0x28,0x3c,0x49,0x8b,	0xff,0x0d,0x95,0x41
+.byte	0x39,0xa8,0x01,0x71,	0x08,0x0c,0xb3,0xde
+.byte	0xd8,0xb4,0xe4,0x9c,	0x64,0x56,0xc1,0x90
+.byte	0x7b,0xcb,0x84,0x61,	0xd5,0x32,0xb6,0x70
+.byte	0x48,0x6c,0x5c,0x74,	0xd0,0xb8,0x57,0x42
+
+.byte	0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38	# Td4
+.byte	0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb
+.byte	0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87
+.byte	0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb
+.byte	0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d
+.byte	0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e
+.byte	0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2
+.byte	0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25
+.byte	0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16
+.byte	0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92
+.byte	0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda
+.byte	0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84
+.byte	0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a
+.byte	0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06
+.byte	0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02
+.byte	0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b
+.byte	0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea
+.byte	0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73
+.byte	0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85
+.byte	0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e
+.byte	0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89
+.byte	0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b
+.byte	0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20
+.byte	0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4
+.byte	0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31
+.byte	0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f
+.byte	0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d
+.byte	0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef
+.byte	0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0
+.byte	0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61
+.byte	0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26
+.byte	0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
+___
+
+foreach (split("\n",$code)) {
+	s/\`([^\`]*)\`/eval $1/ge;
+
+	# made-up _instructions, _xtr, _ins, _ror and _bias, cope
+	# with byte order dependencies...
+	if (/^\s+_/) {
+	    s/(_[a-z]+\s+)(\$[0-9]+),([^,]+)(#.*)*$/$1$2,$2,$3/;
+
+	    s/_xtr\s+(\$[0-9]+),(\$[0-9]+),([0-9]+(\-2)*)/
+		sprintf("srl\t$1,$2,%d",$big_endian ?	eval($3)
+					:		eval("24-$3"))/e or
+	    s/_ins\s+(\$[0-9]+),(\$[0-9]+),([0-9]+)/
+		sprintf("sll\t$1,$2,%d",$big_endian ?	eval($3)
+					:		eval("24-$3"))/e or
+	    s/_ror\s+(\$[0-9]+),(\$[0-9]+),(\-?[0-9]+)/
+		sprintf("srl\t$1,$2,%d",$big_endian ?	eval($3)
+					:		eval("$3*-1"))/e or
+	    s/_bias\s+(\$[0-9]+),(\$[0-9]+),([0-9]+)/
+		sprintf("sll\t$1,$2,%d",$big_endian ?	eval($3)
+					:		eval("($3-16)&31"))/e;
+
+	    s/srl\s+(\$[0-9]+),(\$[0-9]+),\-([0-9]+)/
+		sprintf("sll\t$1,$2,$3")/e				or
+	    s/srl\s+(\$[0-9]+),(\$[0-9]+),0/
+		sprintf("and\t$1,$2,0xff")/e				or
+	    s/(sll\s+\$[0-9]+,\$[0-9]+,0)/#$1/;
+	}
+
+	# convert lwl/lwr and swr/swl to little-endian order
+	if (!$big_endian && /^\s+[sl]w[lr]\s+/) {
+	    s/([sl]wl.*)([0-9]+)\((\$[0-9]+)\)/
+		sprintf("$1%d($3)",eval("$2-$2%4+($2%4-1)&3"))/e	or
+	    s/([sl]wr.*)([0-9]+)\((\$[0-9]+)\)/
+		sprintf("$1%d($3)",eval("$2-$2%4+($2%4+1)&3"))/e;
+	}
+
+	print $_,"\n";
+}
+
+close STDOUT;
diff --git a/crypto/aes/asm/aes-parisc.pl b/crypto/aes/asm/aes-parisc.pl
new file mode 100755
index 000000000000..c36b6a22705c
--- /dev/null
+++ b/crypto/aes/asm/aes-parisc.pl
@@ -0,0 +1,1021 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# AES for PA-RISC.
+#
+# June 2009.
+#
+# The module is mechanical transliteration of aes-sparcv9.pl, but with
+# a twist: S-boxes are compressed even further down to 1K+256B. On
+# PA-7100LC performance is ~40% better than gcc 3.2 generated code and
+# is about 33 cycles per byte processed with 128-bit key. Newer CPUs
+# perform at 16 cycles per byte. It's not faster than code generated
+# by vendor compiler, but recall that it has compressed S-boxes, which
+# requires extra processing.
+#
+# Special thanks to polarhome.com for providing HP-UX account.
+
+$flavour = shift;
+$output = shift;
+open STDOUT,">$output";
+
+if ($flavour =~ /64/) {
+	$LEVEL		="2.0W";
+	$SIZE_T		=8;
+	$FRAME_MARKER	=80;
+	$SAVED_RP	=16;
+	$PUSH		="std";
+	$PUSHMA		="std,ma";
+	$POP		="ldd";
+	$POPMB		="ldd,mb";
+} else {
+	$LEVEL		="1.0";
+	$SIZE_T		=4;
+	$FRAME_MARKER	=48;
+	$SAVED_RP	=20;
+	$PUSH		="stw";
+	$PUSHMA		="stwm";
+	$POP		="ldw";
+	$POPMB		="ldwm";
+}
+
+$FRAME=16*$SIZE_T+$FRAME_MARKER;# 16 saved regs + frame marker
+				#                 [+ argument transfer]
+$inp="%r26";	# arg0
+$out="%r25";	# arg1
+$key="%r24";	# arg2
+
+($s0,$s1,$s2,$s3) = ("%r1","%r2","%r3","%r4");
+($t0,$t1,$t2,$t3) = ("%r5","%r6","%r7","%r8");
+
+($acc0, $acc1, $acc2, $acc3, $acc4, $acc5, $acc6, $acc7,
+ $acc8, $acc9,$acc10,$acc11,$acc12,$acc13,$acc14,$acc15) =
+("%r9","%r10","%r11","%r12","%r13","%r14","%r15","%r16",
+"%r17","%r18","%r19","%r20","%r21","%r22","%r23","%r26");
+
+$tbl="%r28";
+$rounds="%r29";
+
+$code=<<___;
+	.LEVEL	$LEVEL
+	.SPACE	\$TEXT\$
+	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY
+
+	.EXPORT	AES_encrypt,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR
+	.ALIGN	64
+AES_encrypt
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-16*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=18
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
+	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
+	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
+	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
+	$PUSH	%r11,`-$FRAME+8*$SIZE_T`(%sp)
+	$PUSH	%r12,`-$FRAME+9*$SIZE_T`(%sp)
+	$PUSH	%r13,`-$FRAME+10*$SIZE_T`(%sp)
+	$PUSH	%r14,`-$FRAME+11*$SIZE_T`(%sp)
+	$PUSH	%r15,`-$FRAME+12*$SIZE_T`(%sp)
+	$PUSH	%r16,`-$FRAME+13*$SIZE_T`(%sp)
+	$PUSH	%r17,`-$FRAME+14*$SIZE_T`(%sp)
+	$PUSH	%r18,`-$FRAME+15*$SIZE_T`(%sp)
+
+	blr	%r0,$tbl
+	ldi	3,$t0
+L\$enc_pic
+	andcm	$tbl,$t0,$tbl
+	ldo	L\$AES_Te-L\$enc_pic($tbl),$tbl
+
+	and	$inp,$t0,$t0
+	sub	$inp,$t0,$inp
+	ldw	0($inp),$s0
+	ldw	4($inp),$s1
+	ldw	8($inp),$s2
+	comib,=	0,$t0,L\$enc_inp_aligned
+	ldw	12($inp),$s3
+
+	sh3addl	$t0,%r0,$t0
+	subi	32,$t0,$t0
+	mtctl	$t0,%cr11
+	ldw	16($inp),$t1
+	vshd	$s0,$s1,$s0
+	vshd	$s1,$s2,$s1
+	vshd	$s2,$s3,$s2
+	vshd	$s3,$t1,$s3
+
+L\$enc_inp_aligned
+	bl	_parisc_AES_encrypt,%r31
+	nop
+
+	extru,<> $out,31,2,%r0
+	b	L\$enc_out_aligned
+	nop
+
+	_srm	$s0,24,$acc0
+	_srm	$s0,16,$acc1
+	stb	$acc0,0($out)
+	_srm	$s0,8,$acc2
+	stb	$acc1,1($out)
+	_srm	$s1,24,$acc4
+	stb	$acc2,2($out)
+	_srm	$s1,16,$acc5
+	stb	$s0,3($out)
+	_srm	$s1,8,$acc6
+	stb	$acc4,4($out)
+	_srm	$s2,24,$acc0
+	stb	$acc5,5($out)
+	_srm	$s2,16,$acc1
+	stb	$acc6,6($out)
+	_srm	$s2,8,$acc2
+	stb	$s1,7($out)
+	_srm	$s3,24,$acc4
+	stb	$acc0,8($out)
+	_srm	$s3,16,$acc5
+	stb	$acc1,9($out)
+	_srm	$s3,8,$acc6
+	stb	$acc2,10($out)
+	stb	$s2,11($out)
+	stb	$acc4,12($out)
+	stb	$acc5,13($out)
+	stb	$acc6,14($out)
+	b	L\$enc_done
+	stb	$s3,15($out)
+
+L\$enc_out_aligned
+	stw	$s0,0($out)
+	stw	$s1,4($out)
+	stw	$s2,8($out)
+	stw	$s3,12($out)
+
+L\$enc_done
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2	; standard epilogue
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
+	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
+	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
+	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
+	$POP	`-$FRAME+8*$SIZE_T`(%sp),%r11
+	$POP	`-$FRAME+9*$SIZE_T`(%sp),%r12
+	$POP	`-$FRAME+10*$SIZE_T`(%sp),%r13
+	$POP	`-$FRAME+11*$SIZE_T`(%sp),%r14
+	$POP	`-$FRAME+12*$SIZE_T`(%sp),%r15
+	$POP	`-$FRAME+13*$SIZE_T`(%sp),%r16
+	$POP	`-$FRAME+14*$SIZE_T`(%sp),%r17
+	$POP	`-$FRAME+15*$SIZE_T`(%sp),%r18
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+
+	.ALIGN	16
+_parisc_AES_encrypt
+	.PROC
+	.CALLINFO	MILLICODE
+	.ENTRY
+	ldw	240($key),$rounds
+	ldw	0($key),$t0
+	ldw	4($key),$t1
+	ldw	8($key),$t2
+	_srm	$rounds,1,$rounds
+	xor	$t0,$s0,$s0
+	ldw	12($key),$t3
+	_srm	$s0,24,$acc0
+	xor	$t1,$s1,$s1
+	ldw	16($key),$t0
+	_srm	$s1,16,$acc1
+	xor	$t2,$s2,$s2
+	ldw	20($key),$t1
+	xor	$t3,$s3,$s3
+	ldw	24($key),$t2
+	ldw	28($key),$t3
+L\$enc_loop
+	_srm	$s2,8,$acc2
+	ldwx,s	$acc0($tbl),$acc0
+	_srm	$s3,0,$acc3
+	ldwx,s	$acc1($tbl),$acc1
+	_srm	$s1,24,$acc4
+	ldwx,s	$acc2($tbl),$acc2
+	_srm	$s2,16,$acc5
+	ldwx,s	$acc3($tbl),$acc3
+	_srm	$s3,8,$acc6
+	ldwx,s	$acc4($tbl),$acc4
+	_srm	$s0,0,$acc7
+	ldwx,s	$acc5($tbl),$acc5
+	_srm	$s2,24,$acc8
+	ldwx,s	$acc6($tbl),$acc6
+	_srm	$s3,16,$acc9
+	ldwx,s	$acc7($tbl),$acc7
+	_srm	$s0,8,$acc10
+	ldwx,s	$acc8($tbl),$acc8
+	_srm	$s1,0,$acc11
+	ldwx,s	$acc9($tbl),$acc9
+	_srm	$s3,24,$acc12
+	ldwx,s	$acc10($tbl),$acc10
+	_srm	$s0,16,$acc13
+	ldwx,s	$acc11($tbl),$acc11
+	_srm	$s1,8,$acc14
+	ldwx,s	$acc12($tbl),$acc12
+	_srm	$s2,0,$acc15
+	ldwx,s	$acc13($tbl),$acc13
+	ldwx,s	$acc14($tbl),$acc14
+	ldwx,s	$acc15($tbl),$acc15
+	addib,= -1,$rounds,L\$enc_last
+	ldo	32($key),$key
+
+		_ror	$acc1,8,$acc1
+		xor	$acc0,$t0,$t0
+	ldw	0($key),$s0
+		_ror	$acc2,16,$acc2
+		xor	$acc1,$t0,$t0
+	ldw	4($key),$s1
+		_ror	$acc3,24,$acc3
+		xor	$acc2,$t0,$t0
+	ldw	8($key),$s2
+		_ror	$acc5,8,$acc5
+		xor	$acc3,$t0,$t0
+	ldw	12($key),$s3
+		_ror	$acc6,16,$acc6
+		xor	$acc4,$t1,$t1
+		_ror	$acc7,24,$acc7
+		xor	$acc5,$t1,$t1
+		_ror	$acc9,8,$acc9
+		xor	$acc6,$t1,$t1
+		_ror	$acc10,16,$acc10
+		xor	$acc7,$t1,$t1
+		_ror	$acc11,24,$acc11
+		xor	$acc8,$t2,$t2
+		_ror	$acc13,8,$acc13
+		xor	$acc9,$t2,$t2
+		_ror	$acc14,16,$acc14
+		xor	$acc10,$t2,$t2
+		_ror	$acc15,24,$acc15
+		xor	$acc11,$t2,$t2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$t3,$t3
+	_srm	$t0,24,$acc0
+		xor	$acc14,$t3,$t3
+	_srm	$t1,16,$acc1
+		xor	$acc15,$t3,$t3
+
+	_srm	$t2,8,$acc2
+	ldwx,s	$acc0($tbl),$acc0
+	_srm	$t3,0,$acc3
+	ldwx,s	$acc1($tbl),$acc1
+	_srm	$t1,24,$acc4
+	ldwx,s	$acc2($tbl),$acc2
+	_srm	$t2,16,$acc5
+	ldwx,s	$acc3($tbl),$acc3
+	_srm	$t3,8,$acc6
+	ldwx,s	$acc4($tbl),$acc4
+	_srm	$t0,0,$acc7
+	ldwx,s	$acc5($tbl),$acc5
+	_srm	$t2,24,$acc8
+	ldwx,s	$acc6($tbl),$acc6
+	_srm	$t3,16,$acc9
+	ldwx,s	$acc7($tbl),$acc7
+	_srm	$t0,8,$acc10
+	ldwx,s	$acc8($tbl),$acc8
+	_srm	$t1,0,$acc11
+	ldwx,s	$acc9($tbl),$acc9
+	_srm	$t3,24,$acc12
+	ldwx,s	$acc10($tbl),$acc10
+	_srm	$t0,16,$acc13
+	ldwx,s	$acc11($tbl),$acc11
+	_srm	$t1,8,$acc14
+	ldwx,s	$acc12($tbl),$acc12
+	_srm	$t2,0,$acc15
+	ldwx,s	$acc13($tbl),$acc13
+		_ror	$acc1,8,$acc1
+	ldwx,s	$acc14($tbl),$acc14
+
+		_ror	$acc2,16,$acc2
+		xor	$acc0,$s0,$s0
+	ldwx,s	$acc15($tbl),$acc15
+		_ror	$acc3,24,$acc3
+		xor	$acc1,$s0,$s0
+	ldw	16($key),$t0
+		_ror	$acc5,8,$acc5
+		xor	$acc2,$s0,$s0
+	ldw	20($key),$t1
+		_ror	$acc6,16,$acc6
+		xor	$acc3,$s0,$s0
+	ldw	24($key),$t2
+		_ror	$acc7,24,$acc7
+		xor	$acc4,$s1,$s1
+	ldw	28($key),$t3
+		_ror	$acc9,8,$acc9
+		xor	$acc5,$s1,$s1
+	ldw	1024+0($tbl),%r0		; prefetch te4
+		_ror	$acc10,16,$acc10
+		xor	$acc6,$s1,$s1
+	ldw	1024+32($tbl),%r0		; prefetch te4
+		_ror	$acc11,24,$acc11
+		xor	$acc7,$s1,$s1
+	ldw	1024+64($tbl),%r0		; prefetch te4
+		_ror	$acc13,8,$acc13
+		xor	$acc8,$s2,$s2
+	ldw	1024+96($tbl),%r0		; prefetch te4
+		_ror	$acc14,16,$acc14
+		xor	$acc9,$s2,$s2
+	ldw	1024+128($tbl),%r0		; prefetch te4
+		_ror	$acc15,24,$acc15
+		xor	$acc10,$s2,$s2
+	ldw	1024+160($tbl),%r0		; prefetch te4
+	_srm	$s0,24,$acc0
+		xor	$acc11,$s2,$s2
+	ldw	1024+192($tbl),%r0		; prefetch te4
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$s3,$s3
+	ldw	1024+224($tbl),%r0		; prefetch te4
+	_srm	$s1,16,$acc1
+		xor	$acc14,$s3,$s3
+	b	L\$enc_loop
+		xor	$acc15,$s3,$s3
+
+	.ALIGN	16
+L\$enc_last
+	ldo	1024($tbl),$rounds
+		_ror	$acc1,8,$acc1
+		xor	$acc0,$t0,$t0
+	ldw	0($key),$s0
+		_ror	$acc2,16,$acc2
+		xor	$acc1,$t0,$t0
+	ldw	4($key),$s1
+		_ror	$acc3,24,$acc3
+		xor	$acc2,$t0,$t0
+	ldw	8($key),$s2
+		_ror	$acc5,8,$acc5
+		xor	$acc3,$t0,$t0
+	ldw	12($key),$s3
+		_ror	$acc6,16,$acc6
+		xor	$acc4,$t1,$t1
+		_ror	$acc7,24,$acc7
+		xor	$acc5,$t1,$t1
+		_ror	$acc9,8,$acc9
+		xor	$acc6,$t1,$t1
+		_ror	$acc10,16,$acc10
+		xor	$acc7,$t1,$t1
+		_ror	$acc11,24,$acc11
+		xor	$acc8,$t2,$t2
+		_ror	$acc13,8,$acc13
+		xor	$acc9,$t2,$t2
+		_ror	$acc14,16,$acc14
+		xor	$acc10,$t2,$t2
+		_ror	$acc15,24,$acc15
+		xor	$acc11,$t2,$t2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$t3,$t3
+	_srm	$t0,24,$acc0
+		xor	$acc14,$t3,$t3
+	_srm	$t1,16,$acc1
+		xor	$acc15,$t3,$t3
+
+	_srm	$t2,8,$acc2
+	ldbx	$acc0($rounds),$acc0
+	_srm	$t1,24,$acc4
+	ldbx	$acc1($rounds),$acc1
+	_srm	$t2,16,$acc5
+	_srm	$t3,0,$acc3
+	ldbx	$acc2($rounds),$acc2
+	ldbx	$acc3($rounds),$acc3
+	_srm	$t3,8,$acc6
+	ldbx	$acc4($rounds),$acc4
+	_srm	$t2,24,$acc8
+	ldbx	$acc5($rounds),$acc5
+	_srm	$t3,16,$acc9
+	_srm	$t0,0,$acc7
+	ldbx	$acc6($rounds),$acc6
+	ldbx	$acc7($rounds),$acc7
+	_srm	$t0,8,$acc10
+	ldbx	$acc8($rounds),$acc8
+	_srm	$t3,24,$acc12
+	ldbx	$acc9($rounds),$acc9
+	_srm	$t0,16,$acc13
+	_srm	$t1,0,$acc11
+	ldbx	$acc10($rounds),$acc10
+	_srm	$t1,8,$acc14
+	ldbx	$acc11($rounds),$acc11
+	ldbx	$acc12($rounds),$acc12
+	ldbx	$acc13($rounds),$acc13
+	_srm	$t2,0,$acc15
+	ldbx	$acc14($rounds),$acc14
+
+		dep	$acc0,7,8,$acc3
+	ldbx	$acc15($rounds),$acc15
+		dep	$acc4,7,8,$acc7
+		dep	$acc1,15,8,$acc3
+		dep	$acc5,15,8,$acc7
+		dep	$acc2,23,8,$acc3
+		dep	$acc6,23,8,$acc7
+		xor	$acc3,$s0,$s0
+		xor	$acc7,$s1,$s1
+		dep	$acc8,7,8,$acc11
+		dep	$acc12,7,8,$acc15
+		dep	$acc9,15,8,$acc11
+		dep	$acc13,15,8,$acc15
+		dep	$acc10,23,8,$acc11
+		dep	$acc14,23,8,$acc15
+		xor	$acc11,$s2,$s2
+
+	bv	(%r31)
+	.EXIT
+		xor	$acc15,$s3,$s3
+	.PROCEND
+
+	.ALIGN	64
+L\$AES_Te
+	.WORD	0xc66363a5, 0xf87c7c84, 0xee777799, 0xf67b7b8d
+	.WORD	0xfff2f20d, 0xd66b6bbd, 0xde6f6fb1, 0x91c5c554
+	.WORD	0x60303050, 0x02010103, 0xce6767a9, 0x562b2b7d
+	.WORD	0xe7fefe19, 0xb5d7d762, 0x4dababe6, 0xec76769a
+	.WORD	0x8fcaca45, 0x1f82829d, 0x89c9c940, 0xfa7d7d87
+	.WORD	0xeffafa15, 0xb25959eb, 0x8e4747c9, 0xfbf0f00b
+	.WORD	0x41adadec, 0xb3d4d467, 0x5fa2a2fd, 0x45afafea
+	.WORD	0x239c9cbf, 0x53a4a4f7, 0xe4727296, 0x9bc0c05b
+	.WORD	0x75b7b7c2, 0xe1fdfd1c, 0x3d9393ae, 0x4c26266a
+	.WORD	0x6c36365a, 0x7e3f3f41, 0xf5f7f702, 0x83cccc4f
+	.WORD	0x6834345c, 0x51a5a5f4, 0xd1e5e534, 0xf9f1f108
+	.WORD	0xe2717193, 0xabd8d873, 0x62313153, 0x2a15153f
+	.WORD	0x0804040c, 0x95c7c752, 0x46232365, 0x9dc3c35e
+	.WORD	0x30181828, 0x379696a1, 0x0a05050f, 0x2f9a9ab5
+	.WORD	0x0e070709, 0x24121236, 0x1b80809b, 0xdfe2e23d
+	.WORD	0xcdebeb26, 0x4e272769, 0x7fb2b2cd, 0xea75759f
+	.WORD	0x1209091b, 0x1d83839e, 0x582c2c74, 0x341a1a2e
+	.WORD	0x361b1b2d, 0xdc6e6eb2, 0xb45a5aee, 0x5ba0a0fb
+	.WORD	0xa45252f6, 0x763b3b4d, 0xb7d6d661, 0x7db3b3ce
+	.WORD	0x5229297b, 0xdde3e33e, 0x5e2f2f71, 0x13848497
+	.WORD	0xa65353f5, 0xb9d1d168, 0x00000000, 0xc1eded2c
+	.WORD	0x40202060, 0xe3fcfc1f, 0x79b1b1c8, 0xb65b5bed
+	.WORD	0xd46a6abe, 0x8dcbcb46, 0x67bebed9, 0x7239394b
+	.WORD	0x944a4ade, 0x984c4cd4, 0xb05858e8, 0x85cfcf4a
+	.WORD	0xbbd0d06b, 0xc5efef2a, 0x4faaaae5, 0xedfbfb16
+	.WORD	0x864343c5, 0x9a4d4dd7, 0x66333355, 0x11858594
+	.WORD	0x8a4545cf, 0xe9f9f910, 0x04020206, 0xfe7f7f81
+	.WORD	0xa05050f0, 0x783c3c44, 0x259f9fba, 0x4ba8a8e3
+	.WORD	0xa25151f3, 0x5da3a3fe, 0x804040c0, 0x058f8f8a
+	.WORD	0x3f9292ad, 0x219d9dbc, 0x70383848, 0xf1f5f504
+	.WORD	0x63bcbcdf, 0x77b6b6c1, 0xafdada75, 0x42212163
+	.WORD	0x20101030, 0xe5ffff1a, 0xfdf3f30e, 0xbfd2d26d
+	.WORD	0x81cdcd4c, 0x180c0c14, 0x26131335, 0xc3ecec2f
+	.WORD	0xbe5f5fe1, 0x359797a2, 0x884444cc, 0x2e171739
+	.WORD	0x93c4c457, 0x55a7a7f2, 0xfc7e7e82, 0x7a3d3d47
+	.WORD	0xc86464ac, 0xba5d5de7, 0x3219192b, 0xe6737395
+	.WORD	0xc06060a0, 0x19818198, 0x9e4f4fd1, 0xa3dcdc7f
+	.WORD	0x44222266, 0x542a2a7e, 0x3b9090ab, 0x0b888883
+	.WORD	0x8c4646ca, 0xc7eeee29, 0x6bb8b8d3, 0x2814143c
+	.WORD	0xa7dede79, 0xbc5e5ee2, 0x160b0b1d, 0xaddbdb76
+	.WORD	0xdbe0e03b, 0x64323256, 0x743a3a4e, 0x140a0a1e
+	.WORD	0x924949db, 0x0c06060a, 0x4824246c, 0xb85c5ce4
+	.WORD	0x9fc2c25d, 0xbdd3d36e, 0x43acacef, 0xc46262a6
+	.WORD	0x399191a8, 0x319595a4, 0xd3e4e437, 0xf279798b
+	.WORD	0xd5e7e732, 0x8bc8c843, 0x6e373759, 0xda6d6db7
+	.WORD	0x018d8d8c, 0xb1d5d564, 0x9c4e4ed2, 0x49a9a9e0
+	.WORD	0xd86c6cb4, 0xac5656fa, 0xf3f4f407, 0xcfeaea25
+	.WORD	0xca6565af, 0xf47a7a8e, 0x47aeaee9, 0x10080818
+	.WORD	0x6fbabad5, 0xf0787888, 0x4a25256f, 0x5c2e2e72
+	.WORD	0x381c1c24, 0x57a6a6f1, 0x73b4b4c7, 0x97c6c651
+	.WORD	0xcbe8e823, 0xa1dddd7c, 0xe874749c, 0x3e1f1f21
+	.WORD	0x964b4bdd, 0x61bdbddc, 0x0d8b8b86, 0x0f8a8a85
+	.WORD	0xe0707090, 0x7c3e3e42, 0x71b5b5c4, 0xcc6666aa
+	.WORD	0x904848d8, 0x06030305, 0xf7f6f601, 0x1c0e0e12
+	.WORD	0xc26161a3, 0x6a35355f, 0xae5757f9, 0x69b9b9d0
+	.WORD	0x17868691, 0x99c1c158, 0x3a1d1d27, 0x279e9eb9
+	.WORD	0xd9e1e138, 0xebf8f813, 0x2b9898b3, 0x22111133
+	.WORD	0xd26969bb, 0xa9d9d970, 0x078e8e89, 0x339494a7
+	.WORD	0x2d9b9bb6, 0x3c1e1e22, 0x15878792, 0xc9e9e920
+	.WORD	0x87cece49, 0xaa5555ff, 0x50282878, 0xa5dfdf7a
+	.WORD	0x038c8c8f, 0x59a1a1f8, 0x09898980, 0x1a0d0d17
+	.WORD	0x65bfbfda, 0xd7e6e631, 0x844242c6, 0xd06868b8
+	.WORD	0x824141c3, 0x299999b0, 0x5a2d2d77, 0x1e0f0f11
+	.WORD	0x7bb0b0cb, 0xa85454fc, 0x6dbbbbd6, 0x2c16163a
+	.BYTE	0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5
+	.BYTE	0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76
+	.BYTE	0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0
+	.BYTE	0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0
+	.BYTE	0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc
+	.BYTE	0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15
+	.BYTE	0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a
+	.BYTE	0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75
+	.BYTE	0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0
+	.BYTE	0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84
+	.BYTE	0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b
+	.BYTE	0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf
+	.BYTE	0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85
+	.BYTE	0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8
+	.BYTE	0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5
+	.BYTE	0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2
+	.BYTE	0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17
+	.BYTE	0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73
+	.BYTE	0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88
+	.BYTE	0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb
+	.BYTE	0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c
+	.BYTE	0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79
+	.BYTE	0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9
+	.BYTE	0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08
+	.BYTE	0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6
+	.BYTE	0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a
+	.BYTE	0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e
+	.BYTE	0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e
+	.BYTE	0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94
+	.BYTE	0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf
+	.BYTE	0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68
+	.BYTE	0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16
+___
+
+$code.=<<___;
+	.EXPORT	AES_decrypt,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR
+	.ALIGN	16
+AES_decrypt
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-16*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=18
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
+	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
+	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
+	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
+	$PUSH	%r11,`-$FRAME+8*$SIZE_T`(%sp)
+	$PUSH	%r12,`-$FRAME+9*$SIZE_T`(%sp)
+	$PUSH	%r13,`-$FRAME+10*$SIZE_T`(%sp)
+	$PUSH	%r14,`-$FRAME+11*$SIZE_T`(%sp)
+	$PUSH	%r15,`-$FRAME+12*$SIZE_T`(%sp)
+	$PUSH	%r16,`-$FRAME+13*$SIZE_T`(%sp)
+	$PUSH	%r17,`-$FRAME+14*$SIZE_T`(%sp)
+	$PUSH	%r18,`-$FRAME+15*$SIZE_T`(%sp)
+
+	blr	%r0,$tbl
+	ldi	3,$t0
+L\$dec_pic
+	andcm	$tbl,$t0,$tbl
+	ldo	L\$AES_Td-L\$dec_pic($tbl),$tbl
+
+	and	$inp,$t0,$t0
+	sub	$inp,$t0,$inp
+	ldw	0($inp),$s0
+	ldw	4($inp),$s1
+	ldw	8($inp),$s2
+	comib,=	0,$t0,L\$dec_inp_aligned
+	ldw	12($inp),$s3
+
+	sh3addl	$t0,%r0,$t0
+	subi	32,$t0,$t0
+	mtctl	$t0,%cr11
+	ldw	16($inp),$t1
+	vshd	$s0,$s1,$s0
+	vshd	$s1,$s2,$s1
+	vshd	$s2,$s3,$s2
+	vshd	$s3,$t1,$s3
+
+L\$dec_inp_aligned
+	bl	_parisc_AES_decrypt,%r31
+	nop
+
+	extru,<> $out,31,2,%r0
+	b	L\$dec_out_aligned
+	nop
+
+	_srm	$s0,24,$acc0
+	_srm	$s0,16,$acc1
+	stb	$acc0,0($out)
+	_srm	$s0,8,$acc2
+	stb	$acc1,1($out)
+	_srm	$s1,24,$acc4
+	stb	$acc2,2($out)
+	_srm	$s1,16,$acc5
+	stb	$s0,3($out)
+	_srm	$s1,8,$acc6
+	stb	$acc4,4($out)
+	_srm	$s2,24,$acc0
+	stb	$acc5,5($out)
+	_srm	$s2,16,$acc1
+	stb	$acc6,6($out)
+	_srm	$s2,8,$acc2
+	stb	$s1,7($out)
+	_srm	$s3,24,$acc4
+	stb	$acc0,8($out)
+	_srm	$s3,16,$acc5
+	stb	$acc1,9($out)
+	_srm	$s3,8,$acc6
+	stb	$acc2,10($out)
+	stb	$s2,11($out)
+	stb	$acc4,12($out)
+	stb	$acc5,13($out)
+	stb	$acc6,14($out)
+	b	L\$dec_done
+	stb	$s3,15($out)
+
+L\$dec_out_aligned
+	stw	$s0,0($out)
+	stw	$s1,4($out)
+	stw	$s2,8($out)
+	stw	$s3,12($out)
+
+L\$dec_done
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2	; standard epilogue
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
+	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
+	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
+	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
+	$POP	`-$FRAME+8*$SIZE_T`(%sp),%r11
+	$POP	`-$FRAME+9*$SIZE_T`(%sp),%r12
+	$POP	`-$FRAME+10*$SIZE_T`(%sp),%r13
+	$POP	`-$FRAME+11*$SIZE_T`(%sp),%r14
+	$POP	`-$FRAME+12*$SIZE_T`(%sp),%r15
+	$POP	`-$FRAME+13*$SIZE_T`(%sp),%r16
+	$POP	`-$FRAME+14*$SIZE_T`(%sp),%r17
+	$POP	`-$FRAME+15*$SIZE_T`(%sp),%r18
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+
+	.ALIGN	16
+_parisc_AES_decrypt
+	.PROC
+	.CALLINFO	MILLICODE
+	.ENTRY
+	ldw	240($key),$rounds
+	ldw	0($key),$t0
+	ldw	4($key),$t1
+	ldw	8($key),$t2
+	ldw	12($key),$t3
+	_srm	$rounds,1,$rounds
+	xor	$t0,$s0,$s0
+	ldw	16($key),$t0
+	xor	$t1,$s1,$s1
+	ldw	20($key),$t1
+	_srm	$s0,24,$acc0
+	xor	$t2,$s2,$s2
+	ldw	24($key),$t2
+	xor	$t3,$s3,$s3
+	ldw	28($key),$t3
+	_srm	$s3,16,$acc1
+L\$dec_loop
+	_srm	$s2,8,$acc2
+	ldwx,s	$acc0($tbl),$acc0
+	_srm	$s1,0,$acc3
+	ldwx,s	$acc1($tbl),$acc1
+	_srm	$s1,24,$acc4
+	ldwx,s	$acc2($tbl),$acc2
+	_srm	$s0,16,$acc5
+	ldwx,s	$acc3($tbl),$acc3
+	_srm	$s3,8,$acc6
+	ldwx,s	$acc4($tbl),$acc4
+	_srm	$s2,0,$acc7
+	ldwx,s	$acc5($tbl),$acc5
+	_srm	$s2,24,$acc8
+	ldwx,s	$acc6($tbl),$acc6
+	_srm	$s1,16,$acc9
+	ldwx,s	$acc7($tbl),$acc7
+	_srm	$s0,8,$acc10
+	ldwx,s	$acc8($tbl),$acc8
+	_srm	$s3,0,$acc11
+	ldwx,s	$acc9($tbl),$acc9
+	_srm	$s3,24,$acc12
+	ldwx,s	$acc10($tbl),$acc10
+	_srm	$s2,16,$acc13
+	ldwx,s	$acc11($tbl),$acc11
+	_srm	$s1,8,$acc14
+	ldwx,s	$acc12($tbl),$acc12
+	_srm	$s0,0,$acc15
+	ldwx,s	$acc13($tbl),$acc13
+	ldwx,s	$acc14($tbl),$acc14
+	ldwx,s	$acc15($tbl),$acc15
+	addib,= -1,$rounds,L\$dec_last
+	ldo	32($key),$key
+
+		_ror	$acc1,8,$acc1
+		xor	$acc0,$t0,$t0
+	ldw	0($key),$s0
+		_ror	$acc2,16,$acc2
+		xor	$acc1,$t0,$t0
+	ldw	4($key),$s1
+		_ror	$acc3,24,$acc3
+		xor	$acc2,$t0,$t0
+	ldw	8($key),$s2
+		_ror	$acc5,8,$acc5
+		xor	$acc3,$t0,$t0
+	ldw	12($key),$s3
+		_ror	$acc6,16,$acc6
+		xor	$acc4,$t1,$t1
+		_ror	$acc7,24,$acc7
+		xor	$acc5,$t1,$t1
+		_ror	$acc9,8,$acc9
+		xor	$acc6,$t1,$t1
+		_ror	$acc10,16,$acc10
+		xor	$acc7,$t1,$t1
+		_ror	$acc11,24,$acc11
+		xor	$acc8,$t2,$t2
+		_ror	$acc13,8,$acc13
+		xor	$acc9,$t2,$t2
+		_ror	$acc14,16,$acc14
+		xor	$acc10,$t2,$t2
+		_ror	$acc15,24,$acc15
+		xor	$acc11,$t2,$t2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$t3,$t3
+	_srm	$t0,24,$acc0
+		xor	$acc14,$t3,$t3
+		xor	$acc15,$t3,$t3
+	_srm	$t3,16,$acc1
+
+	_srm	$t2,8,$acc2
+	ldwx,s	$acc0($tbl),$acc0
+	_srm	$t1,0,$acc3
+	ldwx,s	$acc1($tbl),$acc1
+	_srm	$t1,24,$acc4
+	ldwx,s	$acc2($tbl),$acc2
+	_srm	$t0,16,$acc5
+	ldwx,s	$acc3($tbl),$acc3
+	_srm	$t3,8,$acc6
+	ldwx,s	$acc4($tbl),$acc4
+	_srm	$t2,0,$acc7
+	ldwx,s	$acc5($tbl),$acc5
+	_srm	$t2,24,$acc8
+	ldwx,s	$acc6($tbl),$acc6
+	_srm	$t1,16,$acc9
+	ldwx,s	$acc7($tbl),$acc7
+	_srm	$t0,8,$acc10
+	ldwx,s	$acc8($tbl),$acc8
+	_srm	$t3,0,$acc11
+	ldwx,s	$acc9($tbl),$acc9
+	_srm	$t3,24,$acc12
+	ldwx,s	$acc10($tbl),$acc10
+	_srm	$t2,16,$acc13
+	ldwx,s	$acc11($tbl),$acc11
+	_srm	$t1,8,$acc14
+	ldwx,s	$acc12($tbl),$acc12
+	_srm	$t0,0,$acc15
+	ldwx,s	$acc13($tbl),$acc13
+		_ror	$acc1,8,$acc1
+	ldwx,s	$acc14($tbl),$acc14
+
+		_ror	$acc2,16,$acc2
+		xor	$acc0,$s0,$s0
+	ldwx,s	$acc15($tbl),$acc15
+		_ror	$acc3,24,$acc3
+		xor	$acc1,$s0,$s0
+	ldw	16($key),$t0
+		_ror	$acc5,8,$acc5
+		xor	$acc2,$s0,$s0
+	ldw	20($key),$t1
+		_ror	$acc6,16,$acc6
+		xor	$acc3,$s0,$s0
+	ldw	24($key),$t2
+		_ror	$acc7,24,$acc7
+		xor	$acc4,$s1,$s1
+	ldw	28($key),$t3
+		_ror	$acc9,8,$acc9
+		xor	$acc5,$s1,$s1
+	ldw	1024+0($tbl),%r0		; prefetch td4
+		_ror	$acc10,16,$acc10
+		xor	$acc6,$s1,$s1
+	ldw	1024+32($tbl),%r0		; prefetch td4
+		_ror	$acc11,24,$acc11
+		xor	$acc7,$s1,$s1
+	ldw	1024+64($tbl),%r0		; prefetch td4
+		_ror	$acc13,8,$acc13
+		xor	$acc8,$s2,$s2
+	ldw	1024+96($tbl),%r0		; prefetch td4
+		_ror	$acc14,16,$acc14
+		xor	$acc9,$s2,$s2
+	ldw	1024+128($tbl),%r0		; prefetch td4
+		_ror	$acc15,24,$acc15
+		xor	$acc10,$s2,$s2
+	ldw	1024+160($tbl),%r0		; prefetch td4
+	_srm	$s0,24,$acc0
+		xor	$acc11,$s2,$s2
+	ldw	1024+192($tbl),%r0		; prefetch td4
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$s3,$s3
+	ldw	1024+224($tbl),%r0		; prefetch td4
+		xor	$acc14,$s3,$s3
+		xor	$acc15,$s3,$s3
+	b	L\$dec_loop
+	_srm	$s3,16,$acc1
+
+	.ALIGN	16
+L\$dec_last
+	ldo	1024($tbl),$rounds
+		_ror	$acc1,8,$acc1
+		xor	$acc0,$t0,$t0
+	ldw	0($key),$s0
+		_ror	$acc2,16,$acc2
+		xor	$acc1,$t0,$t0
+	ldw	4($key),$s1
+		_ror	$acc3,24,$acc3
+		xor	$acc2,$t0,$t0
+	ldw	8($key),$s2
+		_ror	$acc5,8,$acc5
+		xor	$acc3,$t0,$t0
+	ldw	12($key),$s3
+		_ror	$acc6,16,$acc6
+		xor	$acc4,$t1,$t1
+		_ror	$acc7,24,$acc7
+		xor	$acc5,$t1,$t1
+		_ror	$acc9,8,$acc9
+		xor	$acc6,$t1,$t1
+		_ror	$acc10,16,$acc10
+		xor	$acc7,$t1,$t1
+		_ror	$acc11,24,$acc11
+		xor	$acc8,$t2,$t2
+		_ror	$acc13,8,$acc13
+		xor	$acc9,$t2,$t2
+		_ror	$acc14,16,$acc14
+		xor	$acc10,$t2,$t2
+		_ror	$acc15,24,$acc15
+		xor	$acc11,$t2,$t2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$t3,$t3
+	_srm	$t0,24,$acc0
+		xor	$acc14,$t3,$t3
+		xor	$acc15,$t3,$t3
+	_srm	$t3,16,$acc1
+
+	_srm	$t2,8,$acc2
+	ldbx	$acc0($rounds),$acc0
+	_srm	$t1,24,$acc4
+	ldbx	$acc1($rounds),$acc1
+	_srm	$t0,16,$acc5
+	_srm	$t1,0,$acc3
+	ldbx	$acc2($rounds),$acc2
+	ldbx	$acc3($rounds),$acc3
+	_srm	$t3,8,$acc6
+	ldbx	$acc4($rounds),$acc4
+	_srm	$t2,24,$acc8
+	ldbx	$acc5($rounds),$acc5
+	_srm	$t1,16,$acc9
+	_srm	$t2,0,$acc7
+	ldbx	$acc6($rounds),$acc6
+	ldbx	$acc7($rounds),$acc7
+	_srm	$t0,8,$acc10
+	ldbx	$acc8($rounds),$acc8
+	_srm	$t3,24,$acc12
+	ldbx	$acc9($rounds),$acc9
+	_srm	$t2,16,$acc13
+	_srm	$t3,0,$acc11
+	ldbx	$acc10($rounds),$acc10
+	_srm	$t1,8,$acc14
+	ldbx	$acc11($rounds),$acc11
+	ldbx	$acc12($rounds),$acc12
+	ldbx	$acc13($rounds),$acc13
+	_srm	$t0,0,$acc15
+	ldbx	$acc14($rounds),$acc14
+
+		dep	$acc0,7,8,$acc3
+	ldbx	$acc15($rounds),$acc15
+		dep	$acc4,7,8,$acc7
+		dep	$acc1,15,8,$acc3
+		dep	$acc5,15,8,$acc7
+		dep	$acc2,23,8,$acc3
+		dep	$acc6,23,8,$acc7
+		xor	$acc3,$s0,$s0
+		xor	$acc7,$s1,$s1
+		dep	$acc8,7,8,$acc11
+		dep	$acc12,7,8,$acc15
+		dep	$acc9,15,8,$acc11
+		dep	$acc13,15,8,$acc15
+		dep	$acc10,23,8,$acc11
+		dep	$acc14,23,8,$acc15
+		xor	$acc11,$s2,$s2
+
+	bv	(%r31)
+	.EXIT
+		xor	$acc15,$s3,$s3
+	.PROCEND
+
+	.ALIGN	64
+L\$AES_Td
+	.WORD	0x51f4a750, 0x7e416553, 0x1a17a4c3, 0x3a275e96
+	.WORD	0x3bab6bcb, 0x1f9d45f1, 0xacfa58ab, 0x4be30393
+	.WORD	0x2030fa55, 0xad766df6, 0x88cc7691, 0xf5024c25
+	.WORD	0x4fe5d7fc, 0xc52acbd7, 0x26354480, 0xb562a38f
+	.WORD	0xdeb15a49, 0x25ba1b67, 0x45ea0e98, 0x5dfec0e1
+	.WORD	0xc32f7502, 0x814cf012, 0x8d4697a3, 0x6bd3f9c6
+	.WORD	0x038f5fe7, 0x15929c95, 0xbf6d7aeb, 0x955259da
+	.WORD	0xd4be832d, 0x587421d3, 0x49e06929, 0x8ec9c844
+	.WORD	0x75c2896a, 0xf48e7978, 0x99583e6b, 0x27b971dd
+	.WORD	0xbee14fb6, 0xf088ad17, 0xc920ac66, 0x7dce3ab4
+	.WORD	0x63df4a18, 0xe51a3182, 0x97513360, 0x62537f45
+	.WORD	0xb16477e0, 0xbb6bae84, 0xfe81a01c, 0xf9082b94
+	.WORD	0x70486858, 0x8f45fd19, 0x94de6c87, 0x527bf8b7
+	.WORD	0xab73d323, 0x724b02e2, 0xe31f8f57, 0x6655ab2a
+	.WORD	0xb2eb2807, 0x2fb5c203, 0x86c57b9a, 0xd33708a5
+	.WORD	0x302887f2, 0x23bfa5b2, 0x02036aba, 0xed16825c
+	.WORD	0x8acf1c2b, 0xa779b492, 0xf307f2f0, 0x4e69e2a1
+	.WORD	0x65daf4cd, 0x0605bed5, 0xd134621f, 0xc4a6fe8a
+	.WORD	0x342e539d, 0xa2f355a0, 0x058ae132, 0xa4f6eb75
+	.WORD	0x0b83ec39, 0x4060efaa, 0x5e719f06, 0xbd6e1051
+	.WORD	0x3e218af9, 0x96dd063d, 0xdd3e05ae, 0x4de6bd46
+	.WORD	0x91548db5, 0x71c45d05, 0x0406d46f, 0x605015ff
+	.WORD	0x1998fb24, 0xd6bde997, 0x894043cc, 0x67d99e77
+	.WORD	0xb0e842bd, 0x07898b88, 0xe7195b38, 0x79c8eedb
+	.WORD	0xa17c0a47, 0x7c420fe9, 0xf8841ec9, 0x00000000
+	.WORD	0x09808683, 0x322bed48, 0x1e1170ac, 0x6c5a724e
+	.WORD	0xfd0efffb, 0x0f853856, 0x3daed51e, 0x362d3927
+	.WORD	0x0a0fd964, 0x685ca621, 0x9b5b54d1, 0x24362e3a
+	.WORD	0x0c0a67b1, 0x9357e70f, 0xb4ee96d2, 0x1b9b919e
+	.WORD	0x80c0c54f, 0x61dc20a2, 0x5a774b69, 0x1c121a16
+	.WORD	0xe293ba0a, 0xc0a02ae5, 0x3c22e043, 0x121b171d
+	.WORD	0x0e090d0b, 0xf28bc7ad, 0x2db6a8b9, 0x141ea9c8
+	.WORD	0x57f11985, 0xaf75074c, 0xee99ddbb, 0xa37f60fd
+	.WORD	0xf701269f, 0x5c72f5bc, 0x44663bc5, 0x5bfb7e34
+	.WORD	0x8b432976, 0xcb23c6dc, 0xb6edfc68, 0xb8e4f163
+	.WORD	0xd731dcca, 0x42638510, 0x13972240, 0x84c61120
+	.WORD	0x854a247d, 0xd2bb3df8, 0xaef93211, 0xc729a16d
+	.WORD	0x1d9e2f4b, 0xdcb230f3, 0x0d8652ec, 0x77c1e3d0
+	.WORD	0x2bb3166c, 0xa970b999, 0x119448fa, 0x47e96422
+	.WORD	0xa8fc8cc4, 0xa0f03f1a, 0x567d2cd8, 0x223390ef
+	.WORD	0x87494ec7, 0xd938d1c1, 0x8ccaa2fe, 0x98d40b36
+	.WORD	0xa6f581cf, 0xa57ade28, 0xdab78e26, 0x3fadbfa4
+	.WORD	0x2c3a9de4, 0x5078920d, 0x6a5fcc9b, 0x547e4662
+	.WORD	0xf68d13c2, 0x90d8b8e8, 0x2e39f75e, 0x82c3aff5
+	.WORD	0x9f5d80be, 0x69d0937c, 0x6fd52da9, 0xcf2512b3
+	.WORD	0xc8ac993b, 0x10187da7, 0xe89c636e, 0xdb3bbb7b
+	.WORD	0xcd267809, 0x6e5918f4, 0xec9ab701, 0x834f9aa8
+	.WORD	0xe6956e65, 0xaaffe67e, 0x21bccf08, 0xef15e8e6
+	.WORD	0xbae79bd9, 0x4a6f36ce, 0xea9f09d4, 0x29b07cd6
+	.WORD	0x31a4b2af, 0x2a3f2331, 0xc6a59430, 0x35a266c0
+	.WORD	0x744ebc37, 0xfc82caa6, 0xe090d0b0, 0x33a7d815
+	.WORD	0xf104984a, 0x41ecdaf7, 0x7fcd500e, 0x1791f62f
+	.WORD	0x764dd68d, 0x43efb04d, 0xccaa4d54, 0xe49604df
+	.WORD	0x9ed1b5e3, 0x4c6a881b, 0xc12c1fb8, 0x4665517f
+	.WORD	0x9d5eea04, 0x018c355d, 0xfa877473, 0xfb0b412e
+	.WORD	0xb3671d5a, 0x92dbd252, 0xe9105633, 0x6dd64713
+	.WORD	0x9ad7618c, 0x37a10c7a, 0x59f8148e, 0xeb133c89
+	.WORD	0xcea927ee, 0xb761c935, 0xe11ce5ed, 0x7a47b13c
+	.WORD	0x9cd2df59, 0x55f2733f, 0x1814ce79, 0x73c737bf
+	.WORD	0x53f7cdea, 0x5ffdaa5b, 0xdf3d6f14, 0x7844db86
+	.WORD	0xcaaff381, 0xb968c43e, 0x3824342c, 0xc2a3405f
+	.WORD	0x161dc372, 0xbce2250c, 0x283c498b, 0xff0d9541
+	.WORD	0x39a80171, 0x080cb3de, 0xd8b4e49c, 0x6456c190
+	.WORD	0x7bcb8461, 0xd532b670, 0x486c5c74, 0xd0b85742
+	.BYTE	0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38
+	.BYTE	0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb
+	.BYTE	0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87
+	.BYTE	0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb
+	.BYTE	0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d
+	.BYTE	0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e
+	.BYTE	0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2
+	.BYTE	0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25
+	.BYTE	0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16
+	.BYTE	0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92
+	.BYTE	0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda
+	.BYTE	0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84
+	.BYTE	0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a
+	.BYTE	0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06
+	.BYTE	0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02
+	.BYTE	0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b
+	.BYTE	0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea
+	.BYTE	0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73
+	.BYTE	0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85
+	.BYTE	0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e
+	.BYTE	0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89
+	.BYTE	0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b
+	.BYTE	0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20
+	.BYTE	0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4
+	.BYTE	0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31
+	.BYTE	0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f
+	.BYTE	0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d
+	.BYTE	0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef
+	.BYTE	0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0
+	.BYTE	0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61
+	.BYTE	0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26
+	.BYTE	0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
+	.STRINGZ "AES for PA-RISC, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+foreach (split("\n",$code)) {
+	s/\`([^\`]*)\`/eval $1/ge;
+
+	# translate made up instructons: _ror, _srm
+	s/_ror(\s+)(%r[0-9]+),/shd$1$2,$2,/				or
+
+	s/_srm(\s+%r[0-9]+),([0-9]+),/
+		$SIZE_T==4 ? sprintf("extru%s,%d,8,",$1,31-$2)
+		:            sprintf("extrd,u%s,%d,8,",$1,63-$2)/e;
+
+	s/,\*/,/ if ($SIZE_T==4);
+	print $_,"\n";
+}
+close STDOUT;
diff --git a/crypto/aes/asm/aes-ppc.pl b/crypto/aes/asm/aes-ppc.pl
new file mode 100755
index 000000000000..7c52cbe5f9fa
--- /dev/null
+++ b/crypto/aes/asm/aes-ppc.pl
@@ -0,0 +1,1365 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# Needs more work: key setup, CBC routine...
+#
+# ppc_AES_[en|de]crypt perform at 18 cycles per byte processed with
+# 128-bit key, which is ~40% better than 64-bit code generated by gcc
+# 4.0. But these are not the ones currently used! Their "compact"
+# counterparts are, for security reason. ppc_AES_encrypt_compact runs
+# at 1/2 of ppc_AES_encrypt speed, while ppc_AES_decrypt_compact -
+# at 1/3 of ppc_AES_decrypt.
+
+# February 2010
+#
+# Rescheduling instructions to favour Power6 pipeline gave 10%
+# performance improvement on the platfrom in question (and marginal
+# improvement even on others). It should be noted that Power6 fails
+# to process byte in 18 cycles, only in 23, because it fails to issue
+# 4 load instructions in two cycles, only in 3. As result non-compact
+# block subroutines are 25% slower than one would expect. Compact
+# functions scale better, because they have pure computational part,
+# which scales perfectly with clock frequency. To be specific
+# ppc_AES_encrypt_compact operates at 42 cycles per byte, while
+# ppc_AES_decrypt_compact - at 55 (in 64-bit build).
+
+$flavour = shift;
+
+if ($flavour =~ /64/) {
+	$SIZE_T	=8;
+	$LRSAVE	=2*$SIZE_T;
+	$STU	="stdu";
+	$POP	="ld";
+	$PUSH	="std";
+} elsif ($flavour =~ /32/) {
+	$SIZE_T	=4;
+	$LRSAVE	=$SIZE_T;
+	$STU	="stwu";
+	$POP	="lwz";
+	$PUSH	="stw";
+} else { die "nonsense $flavour"; }
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour ".shift || die "can't call $xlate: $!";
+
+$FRAME=32*$SIZE_T;
+
+sub _data_word()
+{ my $i;
+    while(defined($i=shift)) { $code.=sprintf"\t.long\t0x%08x,0x%08x\n",$i,$i; }
+}
+
+$sp="r1";
+$toc="r2";
+$inp="r3";
+$out="r4";
+$key="r5";
+
+$Tbl0="r3";
+$Tbl1="r6";
+$Tbl2="r7";
+$Tbl3="r2";
+
+$s0="r8";
+$s1="r9";
+$s2="r10";
+$s3="r11";
+
+$t0="r12";
+$t1="r13";
+$t2="r14";
+$t3="r15";
+
+$acc00="r16";
+$acc01="r17";
+$acc02="r18";
+$acc03="r19";
+
+$acc04="r20";
+$acc05="r21";
+$acc06="r22";
+$acc07="r23";
+
+$acc08="r24";
+$acc09="r25";
+$acc10="r26";
+$acc11="r27";
+
+$acc12="r28";
+$acc13="r29";
+$acc14="r30";
+$acc15="r31";
+
+# stay away from TLS pointer
+if ($SIZE_T==8)	{ die if ($t1 ne "r13");  $t1="r0";		}
+else		{ die if ($Tbl3 ne "r2"); $Tbl3=$t0; $t0="r0";	}
+$mask80=$Tbl2;
+$mask1b=$Tbl3;
+
+$code.=<<___;
+.machine	"any"
+.text
+
+.align	7
+LAES_Te:
+	mflr	r0
+	bcl	20,31,\$+4
+	mflr	$Tbl0	;    vvvvv "distance" between . and 1st data entry
+	addi	$Tbl0,$Tbl0,`128-8`
+	mtlr	r0
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+	.space	`64-9*4`
+LAES_Td:
+	mflr	r0
+	bcl	20,31,\$+4
+	mflr	$Tbl0	;    vvvvvvvv "distance" between . and 1st data entry
+	addi	$Tbl0,$Tbl0,`128-64-8+2048+256`
+	mtlr	r0
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+	.space	`128-64-9*4`
+___
+&_data_word(
+	0xc66363a5, 0xf87c7c84, 0xee777799, 0xf67b7b8d,
+	0xfff2f20d, 0xd66b6bbd, 0xde6f6fb1, 0x91c5c554,
+	0x60303050, 0x02010103, 0xce6767a9, 0x562b2b7d,
+	0xe7fefe19, 0xb5d7d762, 0x4dababe6, 0xec76769a,
+	0x8fcaca45, 0x1f82829d, 0x89c9c940, 0xfa7d7d87,
+	0xeffafa15, 0xb25959eb, 0x8e4747c9, 0xfbf0f00b,
+	0x41adadec, 0xb3d4d467, 0x5fa2a2fd, 0x45afafea,
+	0x239c9cbf, 0x53a4a4f7, 0xe4727296, 0x9bc0c05b,
+	0x75b7b7c2, 0xe1fdfd1c, 0x3d9393ae, 0x4c26266a,
+	0x6c36365a, 0x7e3f3f41, 0xf5f7f702, 0x83cccc4f,
+	0x6834345c, 0x51a5a5f4, 0xd1e5e534, 0xf9f1f108,
+	0xe2717193, 0xabd8d873, 0x62313153, 0x2a15153f,
+	0x0804040c, 0x95c7c752, 0x46232365, 0x9dc3c35e,
+	0x30181828, 0x379696a1, 0x0a05050f, 0x2f9a9ab5,
+	0x0e070709, 0x24121236, 0x1b80809b, 0xdfe2e23d,
+	0xcdebeb26, 0x4e272769, 0x7fb2b2cd, 0xea75759f,
+	0x1209091b, 0x1d83839e, 0x582c2c74, 0x341a1a2e,
+	0x361b1b2d, 0xdc6e6eb2, 0xb45a5aee, 0x5ba0a0fb,
+	0xa45252f6, 0x763b3b4d, 0xb7d6d661, 0x7db3b3ce,
+	0x5229297b, 0xdde3e33e, 0x5e2f2f71, 0x13848497,
+	0xa65353f5, 0xb9d1d168, 0x00000000, 0xc1eded2c,
+	0x40202060, 0xe3fcfc1f, 0x79b1b1c8, 0xb65b5bed,
+	0xd46a6abe, 0x8dcbcb46, 0x67bebed9, 0x7239394b,
+	0x944a4ade, 0x984c4cd4, 0xb05858e8, 0x85cfcf4a,
+	0xbbd0d06b, 0xc5efef2a, 0x4faaaae5, 0xedfbfb16,
+	0x864343c5, 0x9a4d4dd7, 0x66333355, 0x11858594,
+	0x8a4545cf, 0xe9f9f910, 0x04020206, 0xfe7f7f81,
+	0xa05050f0, 0x783c3c44, 0x259f9fba, 0x4ba8a8e3,
+	0xa25151f3, 0x5da3a3fe, 0x804040c0, 0x058f8f8a,
+	0x3f9292ad, 0x219d9dbc, 0x70383848, 0xf1f5f504,
+	0x63bcbcdf, 0x77b6b6c1, 0xafdada75, 0x42212163,
+	0x20101030, 0xe5ffff1a, 0xfdf3f30e, 0xbfd2d26d,
+	0x81cdcd4c, 0x180c0c14, 0x26131335, 0xc3ecec2f,
+	0xbe5f5fe1, 0x359797a2, 0x884444cc, 0x2e171739,
+	0x93c4c457, 0x55a7a7f2, 0xfc7e7e82, 0x7a3d3d47,
+	0xc86464ac, 0xba5d5de7, 0x3219192b, 0xe6737395,
+	0xc06060a0, 0x19818198, 0x9e4f4fd1, 0xa3dcdc7f,
+	0x44222266, 0x542a2a7e, 0x3b9090ab, 0x0b888883,
+	0x8c4646ca, 0xc7eeee29, 0x6bb8b8d3, 0x2814143c,
+	0xa7dede79, 0xbc5e5ee2, 0x160b0b1d, 0xaddbdb76,
+	0xdbe0e03b, 0x64323256, 0x743a3a4e, 0x140a0a1e,
+	0x924949db, 0x0c06060a, 0x4824246c, 0xb85c5ce4,
+	0x9fc2c25d, 0xbdd3d36e, 0x43acacef, 0xc46262a6,
+	0x399191a8, 0x319595a4, 0xd3e4e437, 0xf279798b,
+	0xd5e7e732, 0x8bc8c843, 0x6e373759, 0xda6d6db7,
+	0x018d8d8c, 0xb1d5d564, 0x9c4e4ed2, 0x49a9a9e0,
+	0xd86c6cb4, 0xac5656fa, 0xf3f4f407, 0xcfeaea25,
+	0xca6565af, 0xf47a7a8e, 0x47aeaee9, 0x10080818,
+	0x6fbabad5, 0xf0787888, 0x4a25256f, 0x5c2e2e72,
+	0x381c1c24, 0x57a6a6f1, 0x73b4b4c7, 0x97c6c651,
+	0xcbe8e823, 0xa1dddd7c, 0xe874749c, 0x3e1f1f21,
+	0x964b4bdd, 0x61bdbddc, 0x0d8b8b86, 0x0f8a8a85,
+	0xe0707090, 0x7c3e3e42, 0x71b5b5c4, 0xcc6666aa,
+	0x904848d8, 0x06030305, 0xf7f6f601, 0x1c0e0e12,
+	0xc26161a3, 0x6a35355f, 0xae5757f9, 0x69b9b9d0,
+	0x17868691, 0x99c1c158, 0x3a1d1d27, 0x279e9eb9,
+	0xd9e1e138, 0xebf8f813, 0x2b9898b3, 0x22111133,
+	0xd26969bb, 0xa9d9d970, 0x078e8e89, 0x339494a7,
+	0x2d9b9bb6, 0x3c1e1e22, 0x15878792, 0xc9e9e920,
+	0x87cece49, 0xaa5555ff, 0x50282878, 0xa5dfdf7a,
+	0x038c8c8f, 0x59a1a1f8, 0x09898980, 0x1a0d0d17,
+	0x65bfbfda, 0xd7e6e631, 0x844242c6, 0xd06868b8,
+	0x824141c3, 0x299999b0, 0x5a2d2d77, 0x1e0f0f11,
+	0x7bb0b0cb, 0xa85454fc, 0x6dbbbbd6, 0x2c16163a);
+$code.=<<___;
+.byte	0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5
+.byte	0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76
+.byte	0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0
+.byte	0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0
+.byte	0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc
+.byte	0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15
+.byte	0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a
+.byte	0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75
+.byte	0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0
+.byte	0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84
+.byte	0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b
+.byte	0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf
+.byte	0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85
+.byte	0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8
+.byte	0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5
+.byte	0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2
+.byte	0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17
+.byte	0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73
+.byte	0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88
+.byte	0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb
+.byte	0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c
+.byte	0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79
+.byte	0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9
+.byte	0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08
+.byte	0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6
+.byte	0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a
+.byte	0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e
+.byte	0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e
+.byte	0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94
+.byte	0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf
+.byte	0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68
+.byte	0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16
+___
+&_data_word(
+	0x51f4a750, 0x7e416553, 0x1a17a4c3, 0x3a275e96,
+	0x3bab6bcb, 0x1f9d45f1, 0xacfa58ab, 0x4be30393,
+	0x2030fa55, 0xad766df6, 0x88cc7691, 0xf5024c25,
+	0x4fe5d7fc, 0xc52acbd7, 0x26354480, 0xb562a38f,
+	0xdeb15a49, 0x25ba1b67, 0x45ea0e98, 0x5dfec0e1,
+	0xc32f7502, 0x814cf012, 0x8d4697a3, 0x6bd3f9c6,
+	0x038f5fe7, 0x15929c95, 0xbf6d7aeb, 0x955259da,
+	0xd4be832d, 0x587421d3, 0x49e06929, 0x8ec9c844,
+	0x75c2896a, 0xf48e7978, 0x99583e6b, 0x27b971dd,
+	0xbee14fb6, 0xf088ad17, 0xc920ac66, 0x7dce3ab4,
+	0x63df4a18, 0xe51a3182, 0x97513360, 0x62537f45,
+	0xb16477e0, 0xbb6bae84, 0xfe81a01c, 0xf9082b94,
+	0x70486858, 0x8f45fd19, 0x94de6c87, 0x527bf8b7,
+	0xab73d323, 0x724b02e2, 0xe31f8f57, 0x6655ab2a,
+	0xb2eb2807, 0x2fb5c203, 0x86c57b9a, 0xd33708a5,
+	0x302887f2, 0x23bfa5b2, 0x02036aba, 0xed16825c,
+	0x8acf1c2b, 0xa779b492, 0xf307f2f0, 0x4e69e2a1,
+	0x65daf4cd, 0x0605bed5, 0xd134621f, 0xc4a6fe8a,
+	0x342e539d, 0xa2f355a0, 0x058ae132, 0xa4f6eb75,
+	0x0b83ec39, 0x4060efaa, 0x5e719f06, 0xbd6e1051,
+	0x3e218af9, 0x96dd063d, 0xdd3e05ae, 0x4de6bd46,
+	0x91548db5, 0x71c45d05, 0x0406d46f, 0x605015ff,
+	0x1998fb24, 0xd6bde997, 0x894043cc, 0x67d99e77,
+	0xb0e842bd, 0x07898b88, 0xe7195b38, 0x79c8eedb,
+	0xa17c0a47, 0x7c420fe9, 0xf8841ec9, 0x00000000,
+	0x09808683, 0x322bed48, 0x1e1170ac, 0x6c5a724e,
+	0xfd0efffb, 0x0f853856, 0x3daed51e, 0x362d3927,
+	0x0a0fd964, 0x685ca621, 0x9b5b54d1, 0x24362e3a,
+	0x0c0a67b1, 0x9357e70f, 0xb4ee96d2, 0x1b9b919e,
+	0x80c0c54f, 0x61dc20a2, 0x5a774b69, 0x1c121a16,
+	0xe293ba0a, 0xc0a02ae5, 0x3c22e043, 0x121b171d,
+	0x0e090d0b, 0xf28bc7ad, 0x2db6a8b9, 0x141ea9c8,
+	0x57f11985, 0xaf75074c, 0xee99ddbb, 0xa37f60fd,
+	0xf701269f, 0x5c72f5bc, 0x44663bc5, 0x5bfb7e34,
+	0x8b432976, 0xcb23c6dc, 0xb6edfc68, 0xb8e4f163,
+	0xd731dcca, 0x42638510, 0x13972240, 0x84c61120,
+	0x854a247d, 0xd2bb3df8, 0xaef93211, 0xc729a16d,
+	0x1d9e2f4b, 0xdcb230f3, 0x0d8652ec, 0x77c1e3d0,
+	0x2bb3166c, 0xa970b999, 0x119448fa, 0x47e96422,
+	0xa8fc8cc4, 0xa0f03f1a, 0x567d2cd8, 0x223390ef,
+	0x87494ec7, 0xd938d1c1, 0x8ccaa2fe, 0x98d40b36,
+	0xa6f581cf, 0xa57ade28, 0xdab78e26, 0x3fadbfa4,
+	0x2c3a9de4, 0x5078920d, 0x6a5fcc9b, 0x547e4662,
+	0xf68d13c2, 0x90d8b8e8, 0x2e39f75e, 0x82c3aff5,
+	0x9f5d80be, 0x69d0937c, 0x6fd52da9, 0xcf2512b3,
+	0xc8ac993b, 0x10187da7, 0xe89c636e, 0xdb3bbb7b,
+	0xcd267809, 0x6e5918f4, 0xec9ab701, 0x834f9aa8,
+	0xe6956e65, 0xaaffe67e, 0x21bccf08, 0xef15e8e6,
+	0xbae79bd9, 0x4a6f36ce, 0xea9f09d4, 0x29b07cd6,
+	0x31a4b2af, 0x2a3f2331, 0xc6a59430, 0x35a266c0,
+	0x744ebc37, 0xfc82caa6, 0xe090d0b0, 0x33a7d815,
+	0xf104984a, 0x41ecdaf7, 0x7fcd500e, 0x1791f62f,
+	0x764dd68d, 0x43efb04d, 0xccaa4d54, 0xe49604df,
+	0x9ed1b5e3, 0x4c6a881b, 0xc12c1fb8, 0x4665517f,
+	0x9d5eea04, 0x018c355d, 0xfa877473, 0xfb0b412e,
+	0xb3671d5a, 0x92dbd252, 0xe9105633, 0x6dd64713,
+	0x9ad7618c, 0x37a10c7a, 0x59f8148e, 0xeb133c89,
+	0xcea927ee, 0xb761c935, 0xe11ce5ed, 0x7a47b13c,
+	0x9cd2df59, 0x55f2733f, 0x1814ce79, 0x73c737bf,
+	0x53f7cdea, 0x5ffdaa5b, 0xdf3d6f14, 0x7844db86,
+	0xcaaff381, 0xb968c43e, 0x3824342c, 0xc2a3405f,
+	0x161dc372, 0xbce2250c, 0x283c498b, 0xff0d9541,
+	0x39a80171, 0x080cb3de, 0xd8b4e49c, 0x6456c190,
+	0x7bcb8461, 0xd532b670, 0x486c5c74, 0xd0b85742);
+$code.=<<___;
+.byte	0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38
+.byte	0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb
+.byte	0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87
+.byte	0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb
+.byte	0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d
+.byte	0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e
+.byte	0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2
+.byte	0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25
+.byte	0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16
+.byte	0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92
+.byte	0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda
+.byte	0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84
+.byte	0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a
+.byte	0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06
+.byte	0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02
+.byte	0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b
+.byte	0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea
+.byte	0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73
+.byte	0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85
+.byte	0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e
+.byte	0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89
+.byte	0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b
+.byte	0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20
+.byte	0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4
+.byte	0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31
+.byte	0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f
+.byte	0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d
+.byte	0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef
+.byte	0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0
+.byte	0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61
+.byte	0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26
+.byte	0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
+
+
+.globl	.AES_encrypt
+.align	7
+.AES_encrypt:
+	$STU	$sp,-$FRAME($sp)
+	mflr	r0
+
+	$PUSH	$toc,`$FRAME-$SIZE_T*20`($sp)
+	$PUSH	r13,`$FRAME-$SIZE_T*19`($sp)
+	$PUSH	r14,`$FRAME-$SIZE_T*18`($sp)
+	$PUSH	r15,`$FRAME-$SIZE_T*17`($sp)
+	$PUSH	r16,`$FRAME-$SIZE_T*16`($sp)
+	$PUSH	r17,`$FRAME-$SIZE_T*15`($sp)
+	$PUSH	r18,`$FRAME-$SIZE_T*14`($sp)
+	$PUSH	r19,`$FRAME-$SIZE_T*13`($sp)
+	$PUSH	r20,`$FRAME-$SIZE_T*12`($sp)
+	$PUSH	r21,`$FRAME-$SIZE_T*11`($sp)
+	$PUSH	r22,`$FRAME-$SIZE_T*10`($sp)
+	$PUSH	r23,`$FRAME-$SIZE_T*9`($sp)
+	$PUSH	r24,`$FRAME-$SIZE_T*8`($sp)
+	$PUSH	r25,`$FRAME-$SIZE_T*7`($sp)
+	$PUSH	r26,`$FRAME-$SIZE_T*6`($sp)
+	$PUSH	r27,`$FRAME-$SIZE_T*5`($sp)
+	$PUSH	r28,`$FRAME-$SIZE_T*4`($sp)
+	$PUSH	r29,`$FRAME-$SIZE_T*3`($sp)
+	$PUSH	r30,`$FRAME-$SIZE_T*2`($sp)
+	$PUSH	r31,`$FRAME-$SIZE_T*1`($sp)
+	$PUSH	r0,`$FRAME+$LRSAVE`($sp)
+
+	andi.	$t0,$inp,3
+	andi.	$t1,$out,3
+	or.	$t0,$t0,$t1
+	bne	Lenc_unaligned
+
+Lenc_unaligned_ok:
+	lwz	$s0,0($inp)
+	lwz	$s1,4($inp)
+	lwz	$s2,8($inp)
+	lwz	$s3,12($inp)
+	bl	LAES_Te
+	bl	Lppc_AES_encrypt_compact
+	stw	$s0,0($out)
+	stw	$s1,4($out)
+	stw	$s2,8($out)
+	stw	$s3,12($out)
+	b	Lenc_done
+
+Lenc_unaligned:
+	subfic	$t0,$inp,4096
+	subfic	$t1,$out,4096
+	andi.	$t0,$t0,4096-16
+	beq	Lenc_xpage
+	andi.	$t1,$t1,4096-16
+	bne	Lenc_unaligned_ok
+
+Lenc_xpage:
+	lbz	$acc00,0($inp)
+	lbz	$acc01,1($inp)
+	lbz	$acc02,2($inp)
+	lbz	$s0,3($inp)
+	lbz	$acc04,4($inp)
+	lbz	$acc05,5($inp)
+	lbz	$acc06,6($inp)
+	lbz	$s1,7($inp)
+	lbz	$acc08,8($inp)
+	lbz	$acc09,9($inp)
+	lbz	$acc10,10($inp)
+	insrwi	$s0,$acc00,8,0
+	lbz	$s2,11($inp)
+	insrwi	$s1,$acc04,8,0
+	lbz	$acc12,12($inp)
+	insrwi	$s0,$acc01,8,8
+	lbz	$acc13,13($inp)
+	insrwi	$s1,$acc05,8,8
+	lbz	$acc14,14($inp)
+	insrwi	$s0,$acc02,8,16
+	lbz	$s3,15($inp)
+	insrwi	$s1,$acc06,8,16
+	insrwi	$s2,$acc08,8,0
+	insrwi	$s3,$acc12,8,0
+	insrwi	$s2,$acc09,8,8
+	insrwi	$s3,$acc13,8,8
+	insrwi	$s2,$acc10,8,16
+	insrwi	$s3,$acc14,8,16
+
+	bl	LAES_Te
+	bl	Lppc_AES_encrypt_compact
+
+	extrwi	$acc00,$s0,8,0
+	extrwi	$acc01,$s0,8,8
+	stb	$acc00,0($out)
+	extrwi	$acc02,$s0,8,16
+	stb	$acc01,1($out)
+	stb	$acc02,2($out)
+	extrwi	$acc04,$s1,8,0
+	stb	$s0,3($out)
+	extrwi	$acc05,$s1,8,8
+	stb	$acc04,4($out)
+	extrwi	$acc06,$s1,8,16
+	stb	$acc05,5($out)
+	stb	$acc06,6($out)
+	extrwi	$acc08,$s2,8,0
+	stb	$s1,7($out)
+	extrwi	$acc09,$s2,8,8
+	stb	$acc08,8($out)
+	extrwi	$acc10,$s2,8,16
+	stb	$acc09,9($out)
+	stb	$acc10,10($out)
+	extrwi	$acc12,$s3,8,0
+	stb	$s2,11($out)
+	extrwi	$acc13,$s3,8,8
+	stb	$acc12,12($out)
+	extrwi	$acc14,$s3,8,16
+	stb	$acc13,13($out)
+	stb	$acc14,14($out)
+	stb	$s3,15($out)
+
+Lenc_done:
+	$POP	r0,`$FRAME+$LRSAVE`($sp)
+	$POP	$toc,`$FRAME-$SIZE_T*20`($sp)
+	$POP	r13,`$FRAME-$SIZE_T*19`($sp)
+	$POP	r14,`$FRAME-$SIZE_T*18`($sp)
+	$POP	r15,`$FRAME-$SIZE_T*17`($sp)
+	$POP	r16,`$FRAME-$SIZE_T*16`($sp)
+	$POP	r17,`$FRAME-$SIZE_T*15`($sp)
+	$POP	r18,`$FRAME-$SIZE_T*14`($sp)
+	$POP	r19,`$FRAME-$SIZE_T*13`($sp)
+	$POP	r20,`$FRAME-$SIZE_T*12`($sp)
+	$POP	r21,`$FRAME-$SIZE_T*11`($sp)
+	$POP	r22,`$FRAME-$SIZE_T*10`($sp)
+	$POP	r23,`$FRAME-$SIZE_T*9`($sp)
+	$POP	r24,`$FRAME-$SIZE_T*8`($sp)
+	$POP	r25,`$FRAME-$SIZE_T*7`($sp)
+	$POP	r26,`$FRAME-$SIZE_T*6`($sp)
+	$POP	r27,`$FRAME-$SIZE_T*5`($sp)
+	$POP	r28,`$FRAME-$SIZE_T*4`($sp)
+	$POP	r29,`$FRAME-$SIZE_T*3`($sp)
+	$POP	r30,`$FRAME-$SIZE_T*2`($sp)
+	$POP	r31,`$FRAME-$SIZE_T*1`($sp)
+	mtlr	r0
+	addi	$sp,$sp,$FRAME
+	blr
+	.long	0
+	.byte	0,12,4,1,0x80,18,3,0
+	.long	0
+
+.align	5
+Lppc_AES_encrypt:
+	lwz	$acc00,240($key)
+	addi	$Tbl1,$Tbl0,3
+	lwz	$t0,0($key)
+	addi	$Tbl2,$Tbl0,2
+	lwz	$t1,4($key)
+	addi	$Tbl3,$Tbl0,1
+	lwz	$t2,8($key)
+	addi	$acc00,$acc00,-1
+	lwz	$t3,12($key)
+	addi	$key,$key,16
+	xor	$s0,$s0,$t0
+	xor	$s1,$s1,$t1
+	xor	$s2,$s2,$t2
+	xor	$s3,$s3,$t3
+	mtctr	$acc00
+.align	4
+Lenc_loop:
+	rlwinm	$acc00,$s0,`32-24+3`,21,28
+	rlwinm	$acc01,$s1,`32-24+3`,21,28
+	rlwinm	$acc02,$s2,`32-24+3`,21,28
+	rlwinm	$acc03,$s3,`32-24+3`,21,28
+	lwz	$t0,0($key)
+	rlwinm	$acc04,$s1,`32-16+3`,21,28
+	lwz	$t1,4($key)
+	rlwinm	$acc05,$s2,`32-16+3`,21,28
+	lwz	$t2,8($key)
+	rlwinm	$acc06,$s3,`32-16+3`,21,28
+	lwz	$t3,12($key)
+	rlwinm	$acc07,$s0,`32-16+3`,21,28
+	lwzx	$acc00,$Tbl0,$acc00
+	rlwinm	$acc08,$s2,`32-8+3`,21,28
+	lwzx	$acc01,$Tbl0,$acc01
+	rlwinm	$acc09,$s3,`32-8+3`,21,28
+	lwzx	$acc02,$Tbl0,$acc02
+	rlwinm	$acc10,$s0,`32-8+3`,21,28
+	lwzx	$acc03,$Tbl0,$acc03
+	rlwinm	$acc11,$s1,`32-8+3`,21,28
+	lwzx	$acc04,$Tbl1,$acc04
+	rlwinm	$acc12,$s3,`0+3`,21,28
+	lwzx	$acc05,$Tbl1,$acc05
+	rlwinm	$acc13,$s0,`0+3`,21,28
+	lwzx	$acc06,$Tbl1,$acc06
+	rlwinm	$acc14,$s1,`0+3`,21,28
+	lwzx	$acc07,$Tbl1,$acc07
+	rlwinm	$acc15,$s2,`0+3`,21,28
+	lwzx	$acc08,$Tbl2,$acc08
+	xor	$t0,$t0,$acc00
+	lwzx	$acc09,$Tbl2,$acc09
+	xor	$t1,$t1,$acc01
+	lwzx	$acc10,$Tbl2,$acc10
+	xor	$t2,$t2,$acc02
+	lwzx	$acc11,$Tbl2,$acc11
+	xor	$t3,$t3,$acc03
+	lwzx	$acc12,$Tbl3,$acc12
+	xor	$t0,$t0,$acc04
+	lwzx	$acc13,$Tbl3,$acc13
+	xor	$t1,$t1,$acc05
+	lwzx	$acc14,$Tbl3,$acc14
+	xor	$t2,$t2,$acc06
+	lwzx	$acc15,$Tbl3,$acc15
+	xor	$t3,$t3,$acc07
+	xor	$t0,$t0,$acc08
+	xor	$t1,$t1,$acc09
+	xor	$t2,$t2,$acc10
+	xor	$t3,$t3,$acc11
+	xor	$s0,$t0,$acc12
+	xor	$s1,$t1,$acc13
+	xor	$s2,$t2,$acc14
+	xor	$s3,$t3,$acc15
+	addi	$key,$key,16
+	bdnz-	Lenc_loop
+
+	addi	$Tbl2,$Tbl0,2048
+	nop
+	lwz	$t0,0($key)
+	rlwinm	$acc00,$s0,`32-24`,24,31
+	lwz	$t1,4($key)
+	rlwinm	$acc01,$s1,`32-24`,24,31
+	lwz	$t2,8($key)
+	rlwinm	$acc02,$s2,`32-24`,24,31
+	lwz	$t3,12($key)
+	rlwinm	$acc03,$s3,`32-24`,24,31
+	lwz	$acc08,`2048+0`($Tbl0)	! prefetch Te4
+	rlwinm	$acc04,$s1,`32-16`,24,31
+	lwz	$acc09,`2048+32`($Tbl0)
+	rlwinm	$acc05,$s2,`32-16`,24,31
+	lwz	$acc10,`2048+64`($Tbl0)
+	rlwinm	$acc06,$s3,`32-16`,24,31
+	lwz	$acc11,`2048+96`($Tbl0)
+	rlwinm	$acc07,$s0,`32-16`,24,31
+	lwz	$acc12,`2048+128`($Tbl0)
+	rlwinm	$acc08,$s2,`32-8`,24,31
+	lwz	$acc13,`2048+160`($Tbl0)
+	rlwinm	$acc09,$s3,`32-8`,24,31
+	lwz	$acc14,`2048+192`($Tbl0)
+	rlwinm	$acc10,$s0,`32-8`,24,31
+	lwz	$acc15,`2048+224`($Tbl0)
+	rlwinm	$acc11,$s1,`32-8`,24,31
+	lbzx	$acc00,$Tbl2,$acc00
+	rlwinm	$acc12,$s3,`0`,24,31
+	lbzx	$acc01,$Tbl2,$acc01
+	rlwinm	$acc13,$s0,`0`,24,31
+	lbzx	$acc02,$Tbl2,$acc02
+	rlwinm	$acc14,$s1,`0`,24,31
+	lbzx	$acc03,$Tbl2,$acc03
+	rlwinm	$acc15,$s2,`0`,24,31
+	lbzx	$acc04,$Tbl2,$acc04
+	rlwinm	$s0,$acc00,24,0,7
+	lbzx	$acc05,$Tbl2,$acc05
+	rlwinm	$s1,$acc01,24,0,7
+	lbzx	$acc06,$Tbl2,$acc06
+	rlwinm	$s2,$acc02,24,0,7
+	lbzx	$acc07,$Tbl2,$acc07
+	rlwinm	$s3,$acc03,24,0,7
+	lbzx	$acc08,$Tbl2,$acc08
+	rlwimi	$s0,$acc04,16,8,15
+	lbzx	$acc09,$Tbl2,$acc09
+	rlwimi	$s1,$acc05,16,8,15
+	lbzx	$acc10,$Tbl2,$acc10
+	rlwimi	$s2,$acc06,16,8,15
+	lbzx	$acc11,$Tbl2,$acc11
+	rlwimi	$s3,$acc07,16,8,15
+	lbzx	$acc12,$Tbl2,$acc12
+	rlwimi	$s0,$acc08,8,16,23
+	lbzx	$acc13,$Tbl2,$acc13
+	rlwimi	$s1,$acc09,8,16,23
+	lbzx	$acc14,$Tbl2,$acc14
+	rlwimi	$s2,$acc10,8,16,23
+	lbzx	$acc15,$Tbl2,$acc15
+	rlwimi	$s3,$acc11,8,16,23
+	or	$s0,$s0,$acc12
+	or	$s1,$s1,$acc13
+	or	$s2,$s2,$acc14
+	or	$s3,$s3,$acc15
+	xor	$s0,$s0,$t0
+	xor	$s1,$s1,$t1
+	xor	$s2,$s2,$t2
+	xor	$s3,$s3,$t3
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+
+.align	4
+Lppc_AES_encrypt_compact:
+	lwz	$acc00,240($key)
+	addi	$Tbl1,$Tbl0,2048
+	lwz	$t0,0($key)
+	lis	$mask80,0x8080
+	lwz	$t1,4($key)
+	lis	$mask1b,0x1b1b
+	lwz	$t2,8($key)
+	ori	$mask80,$mask80,0x8080
+	lwz	$t3,12($key)
+	ori	$mask1b,$mask1b,0x1b1b
+	addi	$key,$key,16
+	mtctr	$acc00
+.align	4
+Lenc_compact_loop:
+	xor	$s0,$s0,$t0
+	xor	$s1,$s1,$t1
+	rlwinm	$acc00,$s0,`32-24`,24,31
+	xor	$s2,$s2,$t2
+	rlwinm	$acc01,$s1,`32-24`,24,31
+	xor	$s3,$s3,$t3
+	rlwinm	$acc02,$s2,`32-24`,24,31
+	rlwinm	$acc03,$s3,`32-24`,24,31
+	rlwinm	$acc04,$s1,`32-16`,24,31
+	rlwinm	$acc05,$s2,`32-16`,24,31
+	rlwinm	$acc06,$s3,`32-16`,24,31
+	rlwinm	$acc07,$s0,`32-16`,24,31
+	lbzx	$acc00,$Tbl1,$acc00
+	rlwinm	$acc08,$s2,`32-8`,24,31
+	lbzx	$acc01,$Tbl1,$acc01
+	rlwinm	$acc09,$s3,`32-8`,24,31
+	lbzx	$acc02,$Tbl1,$acc02
+	rlwinm	$acc10,$s0,`32-8`,24,31
+	lbzx	$acc03,$Tbl1,$acc03
+	rlwinm	$acc11,$s1,`32-8`,24,31
+	lbzx	$acc04,$Tbl1,$acc04
+	rlwinm	$acc12,$s3,`0`,24,31
+	lbzx	$acc05,$Tbl1,$acc05
+	rlwinm	$acc13,$s0,`0`,24,31
+	lbzx	$acc06,$Tbl1,$acc06
+	rlwinm	$acc14,$s1,`0`,24,31
+	lbzx	$acc07,$Tbl1,$acc07
+	rlwinm	$acc15,$s2,`0`,24,31
+	lbzx	$acc08,$Tbl1,$acc08
+	rlwinm	$s0,$acc00,24,0,7
+	lbzx	$acc09,$Tbl1,$acc09
+	rlwinm	$s1,$acc01,24,0,7
+	lbzx	$acc10,$Tbl1,$acc10
+	rlwinm	$s2,$acc02,24,0,7
+	lbzx	$acc11,$Tbl1,$acc11
+	rlwinm	$s3,$acc03,24,0,7
+	lbzx	$acc12,$Tbl1,$acc12
+	rlwimi	$s0,$acc04,16,8,15
+	lbzx	$acc13,$Tbl1,$acc13
+	rlwimi	$s1,$acc05,16,8,15
+	lbzx	$acc14,$Tbl1,$acc14
+	rlwimi	$s2,$acc06,16,8,15
+	lbzx	$acc15,$Tbl1,$acc15
+	rlwimi	$s3,$acc07,16,8,15
+	rlwimi	$s0,$acc08,8,16,23
+	rlwimi	$s1,$acc09,8,16,23
+	rlwimi	$s2,$acc10,8,16,23
+	rlwimi	$s3,$acc11,8,16,23
+	lwz	$t0,0($key)
+	or	$s0,$s0,$acc12
+	lwz	$t1,4($key)
+	or	$s1,$s1,$acc13
+	lwz	$t2,8($key)
+	or	$s2,$s2,$acc14
+	lwz	$t3,12($key)
+	or	$s3,$s3,$acc15
+
+	addi	$key,$key,16
+	bdz	Lenc_compact_done
+
+	and	$acc00,$s0,$mask80	# r1=r0&0x80808080
+	and	$acc01,$s1,$mask80
+	and	$acc02,$s2,$mask80
+	and	$acc03,$s3,$mask80
+	srwi	$acc04,$acc00,7		# r1>>7
+	andc	$acc08,$s0,$mask80	# r0&0x7f7f7f7f
+	srwi	$acc05,$acc01,7
+	andc	$acc09,$s1,$mask80
+	srwi	$acc06,$acc02,7
+	andc	$acc10,$s2,$mask80
+	srwi	$acc07,$acc03,7
+	andc	$acc11,$s3,$mask80
+	sub	$acc00,$acc00,$acc04	# r1-(r1>>7)
+	sub	$acc01,$acc01,$acc05
+	sub	$acc02,$acc02,$acc06
+	sub	$acc03,$acc03,$acc07
+	add	$acc08,$acc08,$acc08	# (r0&0x7f7f7f7f)<<1
+	add	$acc09,$acc09,$acc09
+	add	$acc10,$acc10,$acc10
+	add	$acc11,$acc11,$acc11
+	and	$acc00,$acc00,$mask1b	# (r1-(r1>>7))&0x1b1b1b1b
+	and	$acc01,$acc01,$mask1b
+	and	$acc02,$acc02,$mask1b
+	and	$acc03,$acc03,$mask1b
+	xor	$acc00,$acc00,$acc08	# r2
+	xor	$acc01,$acc01,$acc09
+	 rotlwi	$acc12,$s0,16		# ROTATE(r0,16)
+	xor	$acc02,$acc02,$acc10
+	 rotlwi	$acc13,$s1,16
+	xor	$acc03,$acc03,$acc11
+	 rotlwi	$acc14,$s2,16
+
+	xor	$s0,$s0,$acc00		# r0^r2
+	rotlwi	$acc15,$s3,16
+	xor	$s1,$s1,$acc01
+	rotrwi	$s0,$s0,24		# ROTATE(r2^r0,24)
+	xor	$s2,$s2,$acc02
+	rotrwi	$s1,$s1,24
+	xor	$s3,$s3,$acc03
+	rotrwi	$s2,$s2,24
+	xor	$s0,$s0,$acc00		# ROTATE(r2^r0,24)^r2
+	rotrwi	$s3,$s3,24
+	xor	$s1,$s1,$acc01
+	xor	$s2,$s2,$acc02
+	xor	$s3,$s3,$acc03
+	rotlwi	$acc08,$acc12,8		# ROTATE(r0,24)
+	xor	$s0,$s0,$acc12		#
+	rotlwi	$acc09,$acc13,8
+	xor	$s1,$s1,$acc13
+	rotlwi	$acc10,$acc14,8
+	xor	$s2,$s2,$acc14
+	rotlwi	$acc11,$acc15,8
+	xor	$s3,$s3,$acc15
+	xor	$s0,$s0,$acc08		#
+	xor	$s1,$s1,$acc09
+	xor	$s2,$s2,$acc10
+	xor	$s3,$s3,$acc11
+
+	b	Lenc_compact_loop
+.align	4
+Lenc_compact_done:
+	xor	$s0,$s0,$t0
+	xor	$s1,$s1,$t1
+	xor	$s2,$s2,$t2
+	xor	$s3,$s3,$t3
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+
+.globl	.AES_decrypt
+.align	7
+.AES_decrypt:
+	$STU	$sp,-$FRAME($sp)
+	mflr	r0
+
+	$PUSH	$toc,`$FRAME-$SIZE_T*20`($sp)
+	$PUSH	r13,`$FRAME-$SIZE_T*19`($sp)
+	$PUSH	r14,`$FRAME-$SIZE_T*18`($sp)
+	$PUSH	r15,`$FRAME-$SIZE_T*17`($sp)
+	$PUSH	r16,`$FRAME-$SIZE_T*16`($sp)
+	$PUSH	r17,`$FRAME-$SIZE_T*15`($sp)
+	$PUSH	r18,`$FRAME-$SIZE_T*14`($sp)
+	$PUSH	r19,`$FRAME-$SIZE_T*13`($sp)
+	$PUSH	r20,`$FRAME-$SIZE_T*12`($sp)
+	$PUSH	r21,`$FRAME-$SIZE_T*11`($sp)
+	$PUSH	r22,`$FRAME-$SIZE_T*10`($sp)
+	$PUSH	r23,`$FRAME-$SIZE_T*9`($sp)
+	$PUSH	r24,`$FRAME-$SIZE_T*8`($sp)
+	$PUSH	r25,`$FRAME-$SIZE_T*7`($sp)
+	$PUSH	r26,`$FRAME-$SIZE_T*6`($sp)
+	$PUSH	r27,`$FRAME-$SIZE_T*5`($sp)
+	$PUSH	r28,`$FRAME-$SIZE_T*4`($sp)
+	$PUSH	r29,`$FRAME-$SIZE_T*3`($sp)
+	$PUSH	r30,`$FRAME-$SIZE_T*2`($sp)
+	$PUSH	r31,`$FRAME-$SIZE_T*1`($sp)
+	$PUSH	r0,`$FRAME+$LRSAVE`($sp)
+
+	andi.	$t0,$inp,3
+	andi.	$t1,$out,3
+	or.	$t0,$t0,$t1
+	bne	Ldec_unaligned
+
+Ldec_unaligned_ok:
+	lwz	$s0,0($inp)
+	lwz	$s1,4($inp)
+	lwz	$s2,8($inp)
+	lwz	$s3,12($inp)
+	bl	LAES_Td
+	bl	Lppc_AES_decrypt_compact
+	stw	$s0,0($out)
+	stw	$s1,4($out)
+	stw	$s2,8($out)
+	stw	$s3,12($out)
+	b	Ldec_done
+
+Ldec_unaligned:
+	subfic	$t0,$inp,4096
+	subfic	$t1,$out,4096
+	andi.	$t0,$t0,4096-16
+	beq	Ldec_xpage
+	andi.	$t1,$t1,4096-16
+	bne	Ldec_unaligned_ok
+
+Ldec_xpage:
+	lbz	$acc00,0($inp)
+	lbz	$acc01,1($inp)
+	lbz	$acc02,2($inp)
+	lbz	$s0,3($inp)
+	lbz	$acc04,4($inp)
+	lbz	$acc05,5($inp)
+	lbz	$acc06,6($inp)
+	lbz	$s1,7($inp)
+	lbz	$acc08,8($inp)
+	lbz	$acc09,9($inp)
+	lbz	$acc10,10($inp)
+	insrwi	$s0,$acc00,8,0
+	lbz	$s2,11($inp)
+	insrwi	$s1,$acc04,8,0
+	lbz	$acc12,12($inp)
+	insrwi	$s0,$acc01,8,8
+	lbz	$acc13,13($inp)
+	insrwi	$s1,$acc05,8,8
+	lbz	$acc14,14($inp)
+	insrwi	$s0,$acc02,8,16
+	lbz	$s3,15($inp)
+	insrwi	$s1,$acc06,8,16
+	insrwi	$s2,$acc08,8,0
+	insrwi	$s3,$acc12,8,0
+	insrwi	$s2,$acc09,8,8
+	insrwi	$s3,$acc13,8,8
+	insrwi	$s2,$acc10,8,16
+	insrwi	$s3,$acc14,8,16
+
+	bl	LAES_Td
+	bl	Lppc_AES_decrypt_compact
+
+	extrwi	$acc00,$s0,8,0
+	extrwi	$acc01,$s0,8,8
+	stb	$acc00,0($out)
+	extrwi	$acc02,$s0,8,16
+	stb	$acc01,1($out)
+	stb	$acc02,2($out)
+	extrwi	$acc04,$s1,8,0
+	stb	$s0,3($out)
+	extrwi	$acc05,$s1,8,8
+	stb	$acc04,4($out)
+	extrwi	$acc06,$s1,8,16
+	stb	$acc05,5($out)
+	stb	$acc06,6($out)
+	extrwi	$acc08,$s2,8,0
+	stb	$s1,7($out)
+	extrwi	$acc09,$s2,8,8
+	stb	$acc08,8($out)
+	extrwi	$acc10,$s2,8,16
+	stb	$acc09,9($out)
+	stb	$acc10,10($out)
+	extrwi	$acc12,$s3,8,0
+	stb	$s2,11($out)
+	extrwi	$acc13,$s3,8,8
+	stb	$acc12,12($out)
+	extrwi	$acc14,$s3,8,16
+	stb	$acc13,13($out)
+	stb	$acc14,14($out)
+	stb	$s3,15($out)
+
+Ldec_done:
+	$POP	r0,`$FRAME+$LRSAVE`($sp)
+	$POP	$toc,`$FRAME-$SIZE_T*20`($sp)
+	$POP	r13,`$FRAME-$SIZE_T*19`($sp)
+	$POP	r14,`$FRAME-$SIZE_T*18`($sp)
+	$POP	r15,`$FRAME-$SIZE_T*17`($sp)
+	$POP	r16,`$FRAME-$SIZE_T*16`($sp)
+	$POP	r17,`$FRAME-$SIZE_T*15`($sp)
+	$POP	r18,`$FRAME-$SIZE_T*14`($sp)
+	$POP	r19,`$FRAME-$SIZE_T*13`($sp)
+	$POP	r20,`$FRAME-$SIZE_T*12`($sp)
+	$POP	r21,`$FRAME-$SIZE_T*11`($sp)
+	$POP	r22,`$FRAME-$SIZE_T*10`($sp)
+	$POP	r23,`$FRAME-$SIZE_T*9`($sp)
+	$POP	r24,`$FRAME-$SIZE_T*8`($sp)
+	$POP	r25,`$FRAME-$SIZE_T*7`($sp)
+	$POP	r26,`$FRAME-$SIZE_T*6`($sp)
+	$POP	r27,`$FRAME-$SIZE_T*5`($sp)
+	$POP	r28,`$FRAME-$SIZE_T*4`($sp)
+	$POP	r29,`$FRAME-$SIZE_T*3`($sp)
+	$POP	r30,`$FRAME-$SIZE_T*2`($sp)
+	$POP	r31,`$FRAME-$SIZE_T*1`($sp)
+	mtlr	r0
+	addi	$sp,$sp,$FRAME
+	blr
+	.long	0
+	.byte	0,12,4,1,0x80,18,3,0
+	.long	0
+
+.align	5
+Lppc_AES_decrypt:
+	lwz	$acc00,240($key)
+	addi	$Tbl1,$Tbl0,3
+	lwz	$t0,0($key)
+	addi	$Tbl2,$Tbl0,2
+	lwz	$t1,4($key)
+	addi	$Tbl3,$Tbl0,1
+	lwz	$t2,8($key)
+	addi	$acc00,$acc00,-1
+	lwz	$t3,12($key)
+	addi	$key,$key,16
+	xor	$s0,$s0,$t0
+	xor	$s1,$s1,$t1
+	xor	$s2,$s2,$t2
+	xor	$s3,$s3,$t3
+	mtctr	$acc00
+.align	4
+Ldec_loop:
+	rlwinm	$acc00,$s0,`32-24+3`,21,28
+	rlwinm	$acc01,$s1,`32-24+3`,21,28
+	rlwinm	$acc02,$s2,`32-24+3`,21,28
+	rlwinm	$acc03,$s3,`32-24+3`,21,28
+	lwz	$t0,0($key)
+	rlwinm	$acc04,$s3,`32-16+3`,21,28
+	lwz	$t1,4($key)
+	rlwinm	$acc05,$s0,`32-16+3`,21,28
+	lwz	$t2,8($key)
+	rlwinm	$acc06,$s1,`32-16+3`,21,28
+	lwz	$t3,12($key)
+	rlwinm	$acc07,$s2,`32-16+3`,21,28
+	lwzx	$acc00,$Tbl0,$acc00
+	rlwinm	$acc08,$s2,`32-8+3`,21,28
+	lwzx	$acc01,$Tbl0,$acc01
+	rlwinm	$acc09,$s3,`32-8+3`,21,28
+	lwzx	$acc02,$Tbl0,$acc02
+	rlwinm	$acc10,$s0,`32-8+3`,21,28
+	lwzx	$acc03,$Tbl0,$acc03
+	rlwinm	$acc11,$s1,`32-8+3`,21,28
+	lwzx	$acc04,$Tbl1,$acc04
+	rlwinm	$acc12,$s1,`0+3`,21,28
+	lwzx	$acc05,$Tbl1,$acc05
+	rlwinm	$acc13,$s2,`0+3`,21,28
+	lwzx	$acc06,$Tbl1,$acc06
+	rlwinm	$acc14,$s3,`0+3`,21,28
+	lwzx	$acc07,$Tbl1,$acc07
+	rlwinm	$acc15,$s0,`0+3`,21,28
+	lwzx	$acc08,$Tbl2,$acc08
+	xor	$t0,$t0,$acc00
+	lwzx	$acc09,$Tbl2,$acc09
+	xor	$t1,$t1,$acc01
+	lwzx	$acc10,$Tbl2,$acc10
+	xor	$t2,$t2,$acc02
+	lwzx	$acc11,$Tbl2,$acc11
+	xor	$t3,$t3,$acc03
+	lwzx	$acc12,$Tbl3,$acc12
+	xor	$t0,$t0,$acc04
+	lwzx	$acc13,$Tbl3,$acc13
+	xor	$t1,$t1,$acc05
+	lwzx	$acc14,$Tbl3,$acc14
+	xor	$t2,$t2,$acc06
+	lwzx	$acc15,$Tbl3,$acc15
+	xor	$t3,$t3,$acc07
+	xor	$t0,$t0,$acc08
+	xor	$t1,$t1,$acc09
+	xor	$t2,$t2,$acc10
+	xor	$t3,$t3,$acc11
+	xor	$s0,$t0,$acc12
+	xor	$s1,$t1,$acc13
+	xor	$s2,$t2,$acc14
+	xor	$s3,$t3,$acc15
+	addi	$key,$key,16
+	bdnz-	Ldec_loop
+
+	addi	$Tbl2,$Tbl0,2048
+	nop
+	lwz	$t0,0($key)
+	rlwinm	$acc00,$s0,`32-24`,24,31
+	lwz	$t1,4($key)
+	rlwinm	$acc01,$s1,`32-24`,24,31
+	lwz	$t2,8($key)
+	rlwinm	$acc02,$s2,`32-24`,24,31
+	lwz	$t3,12($key)
+	rlwinm	$acc03,$s3,`32-24`,24,31
+	lwz	$acc08,`2048+0`($Tbl0)	! prefetch Td4
+	rlwinm	$acc04,$s3,`32-16`,24,31
+	lwz	$acc09,`2048+32`($Tbl0)
+	rlwinm	$acc05,$s0,`32-16`,24,31
+	lwz	$acc10,`2048+64`($Tbl0)
+	lbzx	$acc00,$Tbl2,$acc00
+	lwz	$acc11,`2048+96`($Tbl0)
+	lbzx	$acc01,$Tbl2,$acc01
+	lwz	$acc12,`2048+128`($Tbl0)
+	rlwinm	$acc06,$s1,`32-16`,24,31
+	lwz	$acc13,`2048+160`($Tbl0)
+	rlwinm	$acc07,$s2,`32-16`,24,31
+	lwz	$acc14,`2048+192`($Tbl0)
+	rlwinm	$acc08,$s2,`32-8`,24,31
+	lwz	$acc15,`2048+224`($Tbl0)
+	rlwinm	$acc09,$s3,`32-8`,24,31
+	lbzx	$acc02,$Tbl2,$acc02
+	rlwinm	$acc10,$s0,`32-8`,24,31
+	lbzx	$acc03,$Tbl2,$acc03
+	rlwinm	$acc11,$s1,`32-8`,24,31
+	lbzx	$acc04,$Tbl2,$acc04
+	rlwinm	$acc12,$s1,`0`,24,31
+	lbzx	$acc05,$Tbl2,$acc05
+	rlwinm	$acc13,$s2,`0`,24,31
+	lbzx	$acc06,$Tbl2,$acc06
+	rlwinm	$acc14,$s3,`0`,24,31
+	lbzx	$acc07,$Tbl2,$acc07
+	rlwinm	$acc15,$s0,`0`,24,31
+	lbzx	$acc08,$Tbl2,$acc08
+	rlwinm	$s0,$acc00,24,0,7
+	lbzx	$acc09,$Tbl2,$acc09
+	rlwinm	$s1,$acc01,24,0,7
+	lbzx	$acc10,$Tbl2,$acc10
+	rlwinm	$s2,$acc02,24,0,7
+	lbzx	$acc11,$Tbl2,$acc11
+	rlwinm	$s3,$acc03,24,0,7
+	lbzx	$acc12,$Tbl2,$acc12
+	rlwimi	$s0,$acc04,16,8,15
+	lbzx	$acc13,$Tbl2,$acc13
+	rlwimi	$s1,$acc05,16,8,15
+	lbzx	$acc14,$Tbl2,$acc14
+	rlwimi	$s2,$acc06,16,8,15
+	lbzx	$acc15,$Tbl2,$acc15
+	rlwimi	$s3,$acc07,16,8,15
+	rlwimi	$s0,$acc08,8,16,23
+	rlwimi	$s1,$acc09,8,16,23
+	rlwimi	$s2,$acc10,8,16,23
+	rlwimi	$s3,$acc11,8,16,23
+	or	$s0,$s0,$acc12
+	or	$s1,$s1,$acc13
+	or	$s2,$s2,$acc14
+	or	$s3,$s3,$acc15
+	xor	$s0,$s0,$t0
+	xor	$s1,$s1,$t1
+	xor	$s2,$s2,$t2
+	xor	$s3,$s3,$t3
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+
+.align	4
+Lppc_AES_decrypt_compact:
+	lwz	$acc00,240($key)
+	addi	$Tbl1,$Tbl0,2048
+	lwz	$t0,0($key)
+	lis	$mask80,0x8080
+	lwz	$t1,4($key)
+	lis	$mask1b,0x1b1b
+	lwz	$t2,8($key)
+	ori	$mask80,$mask80,0x8080
+	lwz	$t3,12($key)
+	ori	$mask1b,$mask1b,0x1b1b
+	addi	$key,$key,16
+___
+$code.=<<___ if ($SIZE_T==8);
+	insrdi	$mask80,$mask80,32,0
+	insrdi	$mask1b,$mask1b,32,0
+___
+$code.=<<___;
+	mtctr	$acc00
+.align	4
+Ldec_compact_loop:
+	xor	$s0,$s0,$t0
+	xor	$s1,$s1,$t1
+	rlwinm	$acc00,$s0,`32-24`,24,31
+	xor	$s2,$s2,$t2
+	rlwinm	$acc01,$s1,`32-24`,24,31
+	xor	$s3,$s3,$t3
+	rlwinm	$acc02,$s2,`32-24`,24,31
+	rlwinm	$acc03,$s3,`32-24`,24,31
+	rlwinm	$acc04,$s3,`32-16`,24,31
+	rlwinm	$acc05,$s0,`32-16`,24,31
+	rlwinm	$acc06,$s1,`32-16`,24,31
+	rlwinm	$acc07,$s2,`32-16`,24,31
+	lbzx	$acc00,$Tbl1,$acc00
+	rlwinm	$acc08,$s2,`32-8`,24,31
+	lbzx	$acc01,$Tbl1,$acc01
+	rlwinm	$acc09,$s3,`32-8`,24,31
+	lbzx	$acc02,$Tbl1,$acc02
+	rlwinm	$acc10,$s0,`32-8`,24,31
+	lbzx	$acc03,$Tbl1,$acc03
+	rlwinm	$acc11,$s1,`32-8`,24,31
+	lbzx	$acc04,$Tbl1,$acc04
+	rlwinm	$acc12,$s1,`0`,24,31
+	lbzx	$acc05,$Tbl1,$acc05
+	rlwinm	$acc13,$s2,`0`,24,31
+	lbzx	$acc06,$Tbl1,$acc06
+	rlwinm	$acc14,$s3,`0`,24,31
+	lbzx	$acc07,$Tbl1,$acc07
+	rlwinm	$acc15,$s0,`0`,24,31
+	lbzx	$acc08,$Tbl1,$acc08
+	rlwinm	$s0,$acc00,24,0,7
+	lbzx	$acc09,$Tbl1,$acc09
+	rlwinm	$s1,$acc01,24,0,7
+	lbzx	$acc10,$Tbl1,$acc10
+	rlwinm	$s2,$acc02,24,0,7
+	lbzx	$acc11,$Tbl1,$acc11
+	rlwinm	$s3,$acc03,24,0,7
+	lbzx	$acc12,$Tbl1,$acc12
+	rlwimi	$s0,$acc04,16,8,15
+	lbzx	$acc13,$Tbl1,$acc13
+	rlwimi	$s1,$acc05,16,8,15
+	lbzx	$acc14,$Tbl1,$acc14
+	rlwimi	$s2,$acc06,16,8,15
+	lbzx	$acc15,$Tbl1,$acc15
+	rlwimi	$s3,$acc07,16,8,15
+	rlwimi	$s0,$acc08,8,16,23
+	rlwimi	$s1,$acc09,8,16,23
+	rlwimi	$s2,$acc10,8,16,23
+	rlwimi	$s3,$acc11,8,16,23
+	lwz	$t0,0($key)
+	or	$s0,$s0,$acc12
+	lwz	$t1,4($key)
+	or	$s1,$s1,$acc13
+	lwz	$t2,8($key)
+	or	$s2,$s2,$acc14
+	lwz	$t3,12($key)
+	or	$s3,$s3,$acc15
+
+	addi	$key,$key,16
+	bdz	Ldec_compact_done
+___
+$code.=<<___ if ($SIZE_T==8);
+	# vectorized permutation improves decrypt performance by 10%
+	insrdi	$s0,$s1,32,0
+	insrdi	$s2,$s3,32,0
+
+	and	$acc00,$s0,$mask80	# r1=r0&0x80808080
+	and	$acc02,$s2,$mask80
+	srdi	$acc04,$acc00,7		# r1>>7
+	srdi	$acc06,$acc02,7
+	andc	$acc08,$s0,$mask80	# r0&0x7f7f7f7f
+	andc	$acc10,$s2,$mask80
+	sub	$acc00,$acc00,$acc04	# r1-(r1>>7)
+	sub	$acc02,$acc02,$acc06
+	add	$acc08,$acc08,$acc08	# (r0&0x7f7f7f7f)<<1
+	add	$acc10,$acc10,$acc10
+	and	$acc00,$acc00,$mask1b	# (r1-(r1>>7))&0x1b1b1b1b
+	and	$acc02,$acc02,$mask1b
+	xor	$acc00,$acc00,$acc08	# r2
+	xor	$acc02,$acc02,$acc10
+
+	and	$acc04,$acc00,$mask80	# r1=r2&0x80808080
+	and	$acc06,$acc02,$mask80
+	srdi	$acc08,$acc04,7		# r1>>7
+	srdi	$acc10,$acc06,7
+	andc	$acc12,$acc00,$mask80	# r2&0x7f7f7f7f
+	andc	$acc14,$acc02,$mask80
+	sub	$acc04,$acc04,$acc08	# r1-(r1>>7)
+	sub	$acc06,$acc06,$acc10
+	add	$acc12,$acc12,$acc12	# (r2&0x7f7f7f7f)<<1
+	add	$acc14,$acc14,$acc14
+	and	$acc04,$acc04,$mask1b	# (r1-(r1>>7))&0x1b1b1b1b
+	and	$acc06,$acc06,$mask1b
+	xor	$acc04,$acc04,$acc12	# r4
+	xor	$acc06,$acc06,$acc14
+
+	and	$acc08,$acc04,$mask80	# r1=r4&0x80808080
+	and	$acc10,$acc06,$mask80
+	srdi	$acc12,$acc08,7		# r1>>7
+	srdi	$acc14,$acc10,7
+	sub	$acc08,$acc08,$acc12	# r1-(r1>>7)
+	sub	$acc10,$acc10,$acc14
+	andc	$acc12,$acc04,$mask80	# r4&0x7f7f7f7f
+	andc	$acc14,$acc06,$mask80
+	add	$acc12,$acc12,$acc12	# (r4&0x7f7f7f7f)<<1
+	add	$acc14,$acc14,$acc14
+	and	$acc08,$acc08,$mask1b	# (r1-(r1>>7))&0x1b1b1b1b
+	and	$acc10,$acc10,$mask1b
+	xor	$acc08,$acc08,$acc12	# r8
+	xor	$acc10,$acc10,$acc14
+
+	xor	$acc00,$acc00,$s0	# r2^r0
+	xor	$acc02,$acc02,$s2
+	xor	$acc04,$acc04,$s0	# r4^r0
+	xor	$acc06,$acc06,$s2
+
+	extrdi	$acc01,$acc00,32,0
+	extrdi	$acc03,$acc02,32,0
+	extrdi	$acc05,$acc04,32,0
+	extrdi	$acc07,$acc06,32,0
+	extrdi	$acc09,$acc08,32,0
+	extrdi	$acc11,$acc10,32,0
+___
+$code.=<<___ if ($SIZE_T==4);
+	and	$acc00,$s0,$mask80	# r1=r0&0x80808080
+	and	$acc01,$s1,$mask80
+	and	$acc02,$s2,$mask80
+	and	$acc03,$s3,$mask80
+	srwi	$acc04,$acc00,7		# r1>>7
+	andc	$acc08,$s0,$mask80	# r0&0x7f7f7f7f
+	srwi	$acc05,$acc01,7
+	andc	$acc09,$s1,$mask80
+	srwi	$acc06,$acc02,7
+	andc	$acc10,$s2,$mask80
+	srwi	$acc07,$acc03,7
+	andc	$acc11,$s3,$mask80
+	sub	$acc00,$acc00,$acc04	# r1-(r1>>7)
+	sub	$acc01,$acc01,$acc05
+	sub	$acc02,$acc02,$acc06
+	sub	$acc03,$acc03,$acc07
+	add	$acc08,$acc08,$acc08	# (r0&0x7f7f7f7f)<<1
+	add	$acc09,$acc09,$acc09
+	add	$acc10,$acc10,$acc10
+	add	$acc11,$acc11,$acc11
+	and	$acc00,$acc00,$mask1b	# (r1-(r1>>7))&0x1b1b1b1b
+	and	$acc01,$acc01,$mask1b
+	and	$acc02,$acc02,$mask1b
+	and	$acc03,$acc03,$mask1b
+	xor	$acc00,$acc00,$acc08	# r2
+	xor	$acc01,$acc01,$acc09
+	xor	$acc02,$acc02,$acc10
+	xor	$acc03,$acc03,$acc11
+
+	and	$acc04,$acc00,$mask80	# r1=r2&0x80808080
+	and	$acc05,$acc01,$mask80
+	and	$acc06,$acc02,$mask80
+	and	$acc07,$acc03,$mask80
+	srwi	$acc08,$acc04,7		# r1>>7
+	andc	$acc12,$acc00,$mask80	# r2&0x7f7f7f7f
+	srwi	$acc09,$acc05,7
+	andc	$acc13,$acc01,$mask80
+	srwi	$acc10,$acc06,7
+	andc	$acc14,$acc02,$mask80
+	srwi	$acc11,$acc07,7
+	andc	$acc15,$acc03,$mask80
+	sub	$acc04,$acc04,$acc08	# r1-(r1>>7)
+	sub	$acc05,$acc05,$acc09
+	sub	$acc06,$acc06,$acc10
+	sub	$acc07,$acc07,$acc11
+	add	$acc12,$acc12,$acc12	# (r2&0x7f7f7f7f)<<1
+	add	$acc13,$acc13,$acc13
+	add	$acc14,$acc14,$acc14
+	add	$acc15,$acc15,$acc15
+	and	$acc04,$acc04,$mask1b	# (r1-(r1>>7))&0x1b1b1b1b
+	and	$acc05,$acc05,$mask1b
+	and	$acc06,$acc06,$mask1b
+	and	$acc07,$acc07,$mask1b
+	xor	$acc04,$acc04,$acc12	# r4
+	xor	$acc05,$acc05,$acc13
+	xor	$acc06,$acc06,$acc14
+	xor	$acc07,$acc07,$acc15
+
+	and	$acc08,$acc04,$mask80	# r1=r4&0x80808080
+	and	$acc09,$acc05,$mask80
+	srwi	$acc12,$acc08,7		# r1>>7
+	and	$acc10,$acc06,$mask80
+	srwi	$acc13,$acc09,7
+	and	$acc11,$acc07,$mask80
+	srwi	$acc14,$acc10,7
+	sub	$acc08,$acc08,$acc12	# r1-(r1>>7)
+	srwi	$acc15,$acc11,7
+	sub	$acc09,$acc09,$acc13
+	sub	$acc10,$acc10,$acc14
+	sub	$acc11,$acc11,$acc15
+	andc	$acc12,$acc04,$mask80	# r4&0x7f7f7f7f
+	andc	$acc13,$acc05,$mask80
+	andc	$acc14,$acc06,$mask80
+	andc	$acc15,$acc07,$mask80
+	add	$acc12,$acc12,$acc12	# (r4&0x7f7f7f7f)<<1
+	add	$acc13,$acc13,$acc13
+	add	$acc14,$acc14,$acc14
+	add	$acc15,$acc15,$acc15
+	and	$acc08,$acc08,$mask1b	# (r1-(r1>>7))&0x1b1b1b1b
+	and	$acc09,$acc09,$mask1b
+	and	$acc10,$acc10,$mask1b
+	and	$acc11,$acc11,$mask1b
+	xor	$acc08,$acc08,$acc12	# r8
+	xor	$acc09,$acc09,$acc13
+	xor	$acc10,$acc10,$acc14
+	xor	$acc11,$acc11,$acc15
+
+	xor	$acc00,$acc00,$s0	# r2^r0
+	xor	$acc01,$acc01,$s1
+	xor	$acc02,$acc02,$s2
+	xor	$acc03,$acc03,$s3
+	xor	$acc04,$acc04,$s0	# r4^r0
+	xor	$acc05,$acc05,$s1
+	xor	$acc06,$acc06,$s2
+	xor	$acc07,$acc07,$s3
+___
+$code.=<<___;
+	rotrwi	$s0,$s0,8		# = ROTATE(r0,8)
+	rotrwi	$s1,$s1,8
+	xor	$s0,$s0,$acc00		# ^= r2^r0
+	rotrwi	$s2,$s2,8
+	xor	$s1,$s1,$acc01
+	rotrwi	$s3,$s3,8
+	xor	$s2,$s2,$acc02
+	xor	$s3,$s3,$acc03
+	xor	$acc00,$acc00,$acc08
+	xor	$acc01,$acc01,$acc09
+	xor	$acc02,$acc02,$acc10
+	xor	$acc03,$acc03,$acc11
+	xor	$s0,$s0,$acc04		# ^= r4^r0
+	rotrwi	$acc00,$acc00,24
+	xor	$s1,$s1,$acc05
+	rotrwi	$acc01,$acc01,24
+	xor	$s2,$s2,$acc06
+	rotrwi	$acc02,$acc02,24
+	xor	$s3,$s3,$acc07
+	rotrwi	$acc03,$acc03,24
+	xor	$acc04,$acc04,$acc08
+	xor	$acc05,$acc05,$acc09
+	xor	$acc06,$acc06,$acc10
+	xor	$acc07,$acc07,$acc11
+	xor	$s0,$s0,$acc08		# ^= r8 [^((r4^r0)^(r2^r0)=r4^r2)]
+	rotrwi	$acc04,$acc04,16
+	xor	$s1,$s1,$acc09
+	rotrwi	$acc05,$acc05,16
+	xor	$s2,$s2,$acc10
+	rotrwi	$acc06,$acc06,16
+	xor	$s3,$s3,$acc11
+	rotrwi	$acc07,$acc07,16
+	xor	$s0,$s0,$acc00		# ^= ROTATE(r8^r2^r0,24)
+	rotrwi	$acc08,$acc08,8
+	xor	$s1,$s1,$acc01
+	rotrwi	$acc09,$acc09,8
+	xor	$s2,$s2,$acc02
+	rotrwi	$acc10,$acc10,8
+	xor	$s3,$s3,$acc03
+	rotrwi	$acc11,$acc11,8
+	xor	$s0,$s0,$acc04		# ^= ROTATE(r8^r4^r0,16)
+	xor	$s1,$s1,$acc05
+	xor	$s2,$s2,$acc06
+	xor	$s3,$s3,$acc07
+	xor	$s0,$s0,$acc08		# ^= ROTATE(r8,8)	
+	xor	$s1,$s1,$acc09	
+	xor	$s2,$s2,$acc10	
+	xor	$s3,$s3,$acc11	
+
+	b	Ldec_compact_loop
+.align	4
+Ldec_compact_done:
+	xor	$s0,$s0,$t0
+	xor	$s1,$s1,$t1
+	xor	$s2,$s2,$t2
+	xor	$s3,$s3,$t3
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+
+.asciz	"AES for PPC, CRYPTOGAMS by <appro\@openssl.org>"
+.align	7
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/aes/asm/aes-s390x.pl b/crypto/aes/asm/aes-s390x.pl
new file mode 100755
index 000000000000..445a1e67620e
--- /dev/null
+++ b/crypto/aes/asm/aes-s390x.pl
@@ -0,0 +1,2254 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# AES for s390x.
+
+# April 2007.
+#
+# Software performance improvement over gcc-generated code is ~70% and
+# in absolute terms is ~73 cycles per byte processed with 128-bit key.
+# You're likely to exclaim "why so slow?" Keep in mind that z-CPUs are
+# *strictly* in-order execution and issued instruction [in this case
+# load value from memory is critical] has to complete before execution
+# flow proceeds. S-boxes are compressed to 2KB[+256B].
+#
+# As for hardware acceleration support. It's basically a "teaser," as
+# it can and should be improved in several ways. Most notably support
+# for CBC is not utilized, nor multiple blocks are ever processed.
+# Then software key schedule can be postponed till hardware support
+# detection... Performance improvement over assembler is reportedly
+# ~2.5x, but can reach >8x [naturally on larger chunks] if proper
+# support is implemented.
+
+# May 2007.
+#
+# Implement AES_set_[en|de]crypt_key. Key schedule setup is avoided
+# for 128-bit keys, if hardware support is detected.
+
+# Januray 2009.
+#
+# Add support for hardware AES192/256 and reschedule instructions to
+# minimize/avoid Address Generation Interlock hazard and to favour
+# dual-issue z10 pipeline. This gave ~25% improvement on z10 and
+# almost 50% on z9. The gain is smaller on z10, because being dual-
+# issue z10 makes it improssible to eliminate the interlock condition:
+# critial path is not long enough. Yet it spends ~24 cycles per byte
+# processed with 128-bit key.
+#
+# Unlike previous version hardware support detection takes place only
+# at the moment of key schedule setup, which is denoted in key->rounds.
+# This is done, because deferred key setup can't be made MT-safe, not
+# for keys longer than 128 bits.
+#
+# Add AES_cbc_encrypt, which gives incredible performance improvement,
+# it was measured to be ~6.6x. It's less than previously mentioned 8x,
+# because software implementation was optimized.
+
+# May 2010.
+#
+# Add AES_ctr32_encrypt. If hardware-assisted, it provides up to 4.3x
+# performance improvement over "generic" counter mode routine relying
+# on single-block, also hardware-assisted, AES_encrypt. "Up to" refers
+# to the fact that exact throughput value depends on current stack
+# frame alignment within 4KB page. In worst case you get ~75% of the
+# maximum, but *on average* it would be as much as ~98%. Meaning that
+# worst case is unlike, it's like hitting ravine on plateau.
+
+# November 2010.
+#
+# Adapt for -m31 build. If kernel supports what's called "highgprs"
+# feature on Linux [see /proc/cpuinfo], it's possible to use 64-bit
+# instructions and achieve "64-bit" performance even in 31-bit legacy
+# application context. The feature is not specific to any particular
+# processor, as long as it's "z-CPU". Latter implies that the code
+# remains z/Architecture specific. On z990 it was measured to perform
+# 2x better than code generated by gcc 4.3.
+
+# December 2010.
+#
+# Add support for z196 "cipher message with counter" instruction.
+# Note however that it's disengaged, because it was measured to
+# perform ~12% worse than vanilla km-based code...
+
+# February 2011.
+#
+# Add AES_xts_[en|de]crypt. This includes support for z196 km-xts-aes
+# instructions, which deliver ~70% improvement at 8KB block size over
+# vanilla km-based code, 37% - at most like 512-bytes block size.
+
+$flavour = shift;
+
+if ($flavour =~ /3[12]/) {
+	$SIZE_T=4;
+	$g="";
+} else {
+	$SIZE_T=8;
+	$g="g";
+}
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$softonly=0;	# allow hardware support
+
+$t0="%r0";	$mask="%r0";
+$t1="%r1";
+$t2="%r2";	$inp="%r2";
+$t3="%r3";	$out="%r3";	$bits="%r3";
+$key="%r4";
+$i1="%r5";
+$i2="%r6";
+$i3="%r7";
+$s0="%r8";
+$s1="%r9";
+$s2="%r10";
+$s3="%r11";
+$tbl="%r12";
+$rounds="%r13";
+$ra="%r14";
+$sp="%r15";
+
+$stdframe=16*$SIZE_T+4*8;
+
+sub _data_word()
+{ my $i;
+    while(defined($i=shift)) { $code.=sprintf".long\t0x%08x,0x%08x\n",$i,$i; }
+}
+
+$code=<<___;
+.text
+
+.type	AES_Te,\@object
+.align	256
+AES_Te:
+___
+&_data_word(
+	0xc66363a5, 0xf87c7c84, 0xee777799, 0xf67b7b8d,
+	0xfff2f20d, 0xd66b6bbd, 0xde6f6fb1, 0x91c5c554,
+	0x60303050, 0x02010103, 0xce6767a9, 0x562b2b7d,
+	0xe7fefe19, 0xb5d7d762, 0x4dababe6, 0xec76769a,
+	0x8fcaca45, 0x1f82829d, 0x89c9c940, 0xfa7d7d87,
+	0xeffafa15, 0xb25959eb, 0x8e4747c9, 0xfbf0f00b,
+	0x41adadec, 0xb3d4d467, 0x5fa2a2fd, 0x45afafea,
+	0x239c9cbf, 0x53a4a4f7, 0xe4727296, 0x9bc0c05b,
+	0x75b7b7c2, 0xe1fdfd1c, 0x3d9393ae, 0x4c26266a,
+	0x6c36365a, 0x7e3f3f41, 0xf5f7f702, 0x83cccc4f,
+	0x6834345c, 0x51a5a5f4, 0xd1e5e534, 0xf9f1f108,
+	0xe2717193, 0xabd8d873, 0x62313153, 0x2a15153f,
+	0x0804040c, 0x95c7c752, 0x46232365, 0x9dc3c35e,
+	0x30181828, 0x379696a1, 0x0a05050f, 0x2f9a9ab5,
+	0x0e070709, 0x24121236, 0x1b80809b, 0xdfe2e23d,
+	0xcdebeb26, 0x4e272769, 0x7fb2b2cd, 0xea75759f,
+	0x1209091b, 0x1d83839e, 0x582c2c74, 0x341a1a2e,
+	0x361b1b2d, 0xdc6e6eb2, 0xb45a5aee, 0x5ba0a0fb,
+	0xa45252f6, 0x763b3b4d, 0xb7d6d661, 0x7db3b3ce,
+	0x5229297b, 0xdde3e33e, 0x5e2f2f71, 0x13848497,
+	0xa65353f5, 0xb9d1d168, 0x00000000, 0xc1eded2c,
+	0x40202060, 0xe3fcfc1f, 0x79b1b1c8, 0xb65b5bed,
+	0xd46a6abe, 0x8dcbcb46, 0x67bebed9, 0x7239394b,
+	0x944a4ade, 0x984c4cd4, 0xb05858e8, 0x85cfcf4a,
+	0xbbd0d06b, 0xc5efef2a, 0x4faaaae5, 0xedfbfb16,
+	0x864343c5, 0x9a4d4dd7, 0x66333355, 0x11858594,
+	0x8a4545cf, 0xe9f9f910, 0x04020206, 0xfe7f7f81,
+	0xa05050f0, 0x783c3c44, 0x259f9fba, 0x4ba8a8e3,
+	0xa25151f3, 0x5da3a3fe, 0x804040c0, 0x058f8f8a,
+	0x3f9292ad, 0x219d9dbc, 0x70383848, 0xf1f5f504,
+	0x63bcbcdf, 0x77b6b6c1, 0xafdada75, 0x42212163,
+	0x20101030, 0xe5ffff1a, 0xfdf3f30e, 0xbfd2d26d,
+	0x81cdcd4c, 0x180c0c14, 0x26131335, 0xc3ecec2f,
+	0xbe5f5fe1, 0x359797a2, 0x884444cc, 0x2e171739,
+	0x93c4c457, 0x55a7a7f2, 0xfc7e7e82, 0x7a3d3d47,
+	0xc86464ac, 0xba5d5de7, 0x3219192b, 0xe6737395,
+	0xc06060a0, 0x19818198, 0x9e4f4fd1, 0xa3dcdc7f,
+	0x44222266, 0x542a2a7e, 0x3b9090ab, 0x0b888883,
+	0x8c4646ca, 0xc7eeee29, 0x6bb8b8d3, 0x2814143c,
+	0xa7dede79, 0xbc5e5ee2, 0x160b0b1d, 0xaddbdb76,
+	0xdbe0e03b, 0x64323256, 0x743a3a4e, 0x140a0a1e,
+	0x924949db, 0x0c06060a, 0x4824246c, 0xb85c5ce4,
+	0x9fc2c25d, 0xbdd3d36e, 0x43acacef, 0xc46262a6,
+	0x399191a8, 0x319595a4, 0xd3e4e437, 0xf279798b,
+	0xd5e7e732, 0x8bc8c843, 0x6e373759, 0xda6d6db7,
+	0x018d8d8c, 0xb1d5d564, 0x9c4e4ed2, 0x49a9a9e0,
+	0xd86c6cb4, 0xac5656fa, 0xf3f4f407, 0xcfeaea25,
+	0xca6565af, 0xf47a7a8e, 0x47aeaee9, 0x10080818,
+	0x6fbabad5, 0xf0787888, 0x4a25256f, 0x5c2e2e72,
+	0x381c1c24, 0x57a6a6f1, 0x73b4b4c7, 0x97c6c651,
+	0xcbe8e823, 0xa1dddd7c, 0xe874749c, 0x3e1f1f21,
+	0x964b4bdd, 0x61bdbddc, 0x0d8b8b86, 0x0f8a8a85,
+	0xe0707090, 0x7c3e3e42, 0x71b5b5c4, 0xcc6666aa,
+	0x904848d8, 0x06030305, 0xf7f6f601, 0x1c0e0e12,
+	0xc26161a3, 0x6a35355f, 0xae5757f9, 0x69b9b9d0,
+	0x17868691, 0x99c1c158, 0x3a1d1d27, 0x279e9eb9,
+	0xd9e1e138, 0xebf8f813, 0x2b9898b3, 0x22111133,
+	0xd26969bb, 0xa9d9d970, 0x078e8e89, 0x339494a7,
+	0x2d9b9bb6, 0x3c1e1e22, 0x15878792, 0xc9e9e920,
+	0x87cece49, 0xaa5555ff, 0x50282878, 0xa5dfdf7a,
+	0x038c8c8f, 0x59a1a1f8, 0x09898980, 0x1a0d0d17,
+	0x65bfbfda, 0xd7e6e631, 0x844242c6, 0xd06868b8,
+	0x824141c3, 0x299999b0, 0x5a2d2d77, 0x1e0f0f11,
+	0x7bb0b0cb, 0xa85454fc, 0x6dbbbbd6, 0x2c16163a);
+$code.=<<___;
+# Te4[256]
+.byte	0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5
+.byte	0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76
+.byte	0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0
+.byte	0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0
+.byte	0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc
+.byte	0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15
+.byte	0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a
+.byte	0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75
+.byte	0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0
+.byte	0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84
+.byte	0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b
+.byte	0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf
+.byte	0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85
+.byte	0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8
+.byte	0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5
+.byte	0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2
+.byte	0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17
+.byte	0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73
+.byte	0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88
+.byte	0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb
+.byte	0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c
+.byte	0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79
+.byte	0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9
+.byte	0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08
+.byte	0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6
+.byte	0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a
+.byte	0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e
+.byte	0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e
+.byte	0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94
+.byte	0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf
+.byte	0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68
+.byte	0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16
+# rcon[]
+.long	0x01000000, 0x02000000, 0x04000000, 0x08000000
+.long	0x10000000, 0x20000000, 0x40000000, 0x80000000
+.long	0x1B000000, 0x36000000, 0, 0, 0, 0, 0, 0
+.align	256
+.size	AES_Te,.-AES_Te
+
+# void AES_encrypt(const unsigned char *inp, unsigned char *out,
+# 		 const AES_KEY *key) {
+.globl	AES_encrypt
+.type	AES_encrypt,\@function
+AES_encrypt:
+___
+$code.=<<___ if (!$softonly);
+	l	%r0,240($key)
+	lhi	%r1,16
+	clr	%r0,%r1
+	jl	.Lesoft
+
+	la	%r1,0($key)
+	#la	%r2,0($inp)
+	la	%r4,0($out)
+	lghi	%r3,16		# single block length
+	.long	0xb92e0042	# km %r4,%r2
+	brc	1,.-4		# can this happen?
+	br	%r14
+.align	64
+.Lesoft:
+___
+$code.=<<___;
+	stm${g}	%r3,$ra,3*$SIZE_T($sp)
+
+	llgf	$s0,0($inp)
+	llgf	$s1,4($inp)
+	llgf	$s2,8($inp)
+	llgf	$s3,12($inp)
+
+	larl	$tbl,AES_Te
+	bras	$ra,_s390x_AES_encrypt
+
+	l${g}	$out,3*$SIZE_T($sp)
+	st	$s0,0($out)
+	st	$s1,4($out)
+	st	$s2,8($out)
+	st	$s3,12($out)
+
+	lm${g}	%r6,$ra,6*$SIZE_T($sp)
+	br	$ra
+.size	AES_encrypt,.-AES_encrypt
+
+.type   _s390x_AES_encrypt,\@function
+.align	16
+_s390x_AES_encrypt:
+	st${g}	$ra,15*$SIZE_T($sp)
+	x	$s0,0($key)
+	x	$s1,4($key)
+	x	$s2,8($key)
+	x	$s3,12($key)
+	l	$rounds,240($key)
+	llill	$mask,`0xff<<3`
+	aghi	$rounds,-1
+	j	.Lenc_loop
+.align	16
+.Lenc_loop:
+	sllg	$t1,$s0,`0+3`
+	srlg	$t2,$s0,`8-3`
+	srlg	$t3,$s0,`16-3`
+	srl	$s0,`24-3`
+	nr	$s0,$mask
+	ngr	$t1,$mask
+	nr	$t2,$mask
+	nr	$t3,$mask
+
+	srlg	$i1,$s1,`16-3`	# i0
+	sllg	$i2,$s1,`0+3`
+	srlg	$i3,$s1,`8-3`
+	srl	$s1,`24-3`
+	nr	$i1,$mask
+	nr	$s1,$mask
+	ngr	$i2,$mask
+	nr	$i3,$mask
+
+	l	$s0,0($s0,$tbl)	# Te0[s0>>24]
+	l	$t1,1($t1,$tbl)	# Te3[s0>>0]
+	l	$t2,2($t2,$tbl) # Te2[s0>>8]
+	l	$t3,3($t3,$tbl)	# Te1[s0>>16]
+
+	x	$s0,3($i1,$tbl)	# Te1[s1>>16]
+	l	$s1,0($s1,$tbl)	# Te0[s1>>24]
+	x	$t2,1($i2,$tbl)	# Te3[s1>>0]
+	x	$t3,2($i3,$tbl)	# Te2[s1>>8]
+
+	srlg	$i1,$s2,`8-3`	# i0
+	srlg	$i2,$s2,`16-3`	# i1
+	nr	$i1,$mask
+	nr	$i2,$mask
+	sllg	$i3,$s2,`0+3`
+	srl	$s2,`24-3`
+	nr	$s2,$mask
+	ngr	$i3,$mask
+
+	xr	$s1,$t1
+	srlg	$ra,$s3,`8-3`	# i1
+	sllg	$t1,$s3,`0+3`	# i0
+	nr	$ra,$mask
+	la	$key,16($key)
+	ngr	$t1,$mask
+
+	x	$s0,2($i1,$tbl)	# Te2[s2>>8]
+	x	$s1,3($i2,$tbl)	# Te1[s2>>16]
+	l	$s2,0($s2,$tbl)	# Te0[s2>>24]
+	x	$t3,1($i3,$tbl)	# Te3[s2>>0]
+
+	srlg	$i3,$s3,`16-3`	# i2
+	xr	$s2,$t2
+	srl	$s3,`24-3`
+	nr	$i3,$mask
+	nr	$s3,$mask
+
+	x	$s0,0($key)
+	x	$s1,4($key)
+	x	$s2,8($key)
+	x	$t3,12($key)
+
+	x	$s0,1($t1,$tbl)	# Te3[s3>>0]
+	x	$s1,2($ra,$tbl)	# Te2[s3>>8]
+	x	$s2,3($i3,$tbl)	# Te1[s3>>16]
+	l	$s3,0($s3,$tbl)	# Te0[s3>>24]
+	xr	$s3,$t3
+
+	brct	$rounds,.Lenc_loop
+	.align	16
+
+	sllg	$t1,$s0,`0+3`
+	srlg	$t2,$s0,`8-3`
+	ngr	$t1,$mask
+	srlg	$t3,$s0,`16-3`
+	srl	$s0,`24-3`
+	nr	$s0,$mask
+	nr	$t2,$mask
+	nr	$t3,$mask
+
+	srlg	$i1,$s1,`16-3`	# i0
+	sllg	$i2,$s1,`0+3`
+	ngr	$i2,$mask
+	srlg	$i3,$s1,`8-3`
+	srl	$s1,`24-3`
+	nr	$i1,$mask
+	nr	$s1,$mask
+	nr	$i3,$mask
+
+	llgc	$s0,2($s0,$tbl)	# Te4[s0>>24]
+	llgc	$t1,2($t1,$tbl)	# Te4[s0>>0]
+	sll	$s0,24
+	llgc	$t2,2($t2,$tbl)	# Te4[s0>>8]
+	llgc	$t3,2($t3,$tbl)	# Te4[s0>>16]
+	sll	$t2,8
+	sll	$t3,16
+
+	llgc	$i1,2($i1,$tbl)	# Te4[s1>>16]
+	llgc	$s1,2($s1,$tbl)	# Te4[s1>>24]
+	llgc	$i2,2($i2,$tbl)	# Te4[s1>>0]
+	llgc	$i3,2($i3,$tbl)	# Te4[s1>>8]
+	sll	$i1,16
+	sll	$s1,24
+	sll	$i3,8
+	or	$s0,$i1
+	or	$s1,$t1
+	or	$t2,$i2
+	or	$t3,$i3
+	
+	srlg	$i1,$s2,`8-3`	# i0
+	srlg	$i2,$s2,`16-3`	# i1
+	nr	$i1,$mask
+	nr	$i2,$mask
+	sllg	$i3,$s2,`0+3`
+	srl	$s2,`24-3`
+	ngr	$i3,$mask
+	nr	$s2,$mask
+
+	sllg	$t1,$s3,`0+3`	# i0
+	srlg	$ra,$s3,`8-3`	# i1
+	ngr	$t1,$mask
+
+	llgc	$i1,2($i1,$tbl)	# Te4[s2>>8]
+	llgc	$i2,2($i2,$tbl)	# Te4[s2>>16]
+	sll	$i1,8
+	llgc	$s2,2($s2,$tbl)	# Te4[s2>>24]
+	llgc	$i3,2($i3,$tbl)	# Te4[s2>>0]
+	sll	$i2,16
+	nr	$ra,$mask
+	sll	$s2,24
+	or	$s0,$i1
+	or	$s1,$i2
+	or	$s2,$t2
+	or	$t3,$i3
+
+	srlg	$i3,$s3,`16-3`	# i2
+	srl	$s3,`24-3`
+	nr	$i3,$mask
+	nr	$s3,$mask
+
+	l	$t0,16($key)
+	l	$t2,20($key)
+
+	llgc	$i1,2($t1,$tbl)	# Te4[s3>>0]
+	llgc	$i2,2($ra,$tbl)	# Te4[s3>>8]
+	llgc	$i3,2($i3,$tbl)	# Te4[s3>>16]
+	llgc	$s3,2($s3,$tbl)	# Te4[s3>>24]
+	sll	$i2,8
+	sll	$i3,16
+	sll	$s3,24
+	or	$s0,$i1
+	or	$s1,$i2
+	or	$s2,$i3
+	or	$s3,$t3
+
+	l${g}	$ra,15*$SIZE_T($sp)
+	xr	$s0,$t0
+	xr	$s1,$t2
+	x	$s2,24($key)
+	x	$s3,28($key)
+
+	br	$ra	
+.size	_s390x_AES_encrypt,.-_s390x_AES_encrypt
+___
+
+$code.=<<___;
+.type	AES_Td,\@object
+.align	256
+AES_Td:
+___
+&_data_word(
+	0x51f4a750, 0x7e416553, 0x1a17a4c3, 0x3a275e96,
+	0x3bab6bcb, 0x1f9d45f1, 0xacfa58ab, 0x4be30393,
+	0x2030fa55, 0xad766df6, 0x88cc7691, 0xf5024c25,
+	0x4fe5d7fc, 0xc52acbd7, 0x26354480, 0xb562a38f,
+	0xdeb15a49, 0x25ba1b67, 0x45ea0e98, 0x5dfec0e1,
+	0xc32f7502, 0x814cf012, 0x8d4697a3, 0x6bd3f9c6,
+	0x038f5fe7, 0x15929c95, 0xbf6d7aeb, 0x955259da,
+	0xd4be832d, 0x587421d3, 0x49e06929, 0x8ec9c844,
+	0x75c2896a, 0xf48e7978, 0x99583e6b, 0x27b971dd,
+	0xbee14fb6, 0xf088ad17, 0xc920ac66, 0x7dce3ab4,
+	0x63df4a18, 0xe51a3182, 0x97513360, 0x62537f45,
+	0xb16477e0, 0xbb6bae84, 0xfe81a01c, 0xf9082b94,
+	0x70486858, 0x8f45fd19, 0x94de6c87, 0x527bf8b7,
+	0xab73d323, 0x724b02e2, 0xe31f8f57, 0x6655ab2a,
+	0xb2eb2807, 0x2fb5c203, 0x86c57b9a, 0xd33708a5,
+	0x302887f2, 0x23bfa5b2, 0x02036aba, 0xed16825c,
+	0x8acf1c2b, 0xa779b492, 0xf307f2f0, 0x4e69e2a1,
+	0x65daf4cd, 0x0605bed5, 0xd134621f, 0xc4a6fe8a,
+	0x342e539d, 0xa2f355a0, 0x058ae132, 0xa4f6eb75,
+	0x0b83ec39, 0x4060efaa, 0x5e719f06, 0xbd6e1051,
+	0x3e218af9, 0x96dd063d, 0xdd3e05ae, 0x4de6bd46,
+	0x91548db5, 0x71c45d05, 0x0406d46f, 0x605015ff,
+	0x1998fb24, 0xd6bde997, 0x894043cc, 0x67d99e77,
+	0xb0e842bd, 0x07898b88, 0xe7195b38, 0x79c8eedb,
+	0xa17c0a47, 0x7c420fe9, 0xf8841ec9, 0x00000000,
+	0x09808683, 0x322bed48, 0x1e1170ac, 0x6c5a724e,
+	0xfd0efffb, 0x0f853856, 0x3daed51e, 0x362d3927,
+	0x0a0fd964, 0x685ca621, 0x9b5b54d1, 0x24362e3a,
+	0x0c0a67b1, 0x9357e70f, 0xb4ee96d2, 0x1b9b919e,
+	0x80c0c54f, 0x61dc20a2, 0x5a774b69, 0x1c121a16,
+	0xe293ba0a, 0xc0a02ae5, 0x3c22e043, 0x121b171d,
+	0x0e090d0b, 0xf28bc7ad, 0x2db6a8b9, 0x141ea9c8,
+	0x57f11985, 0xaf75074c, 0xee99ddbb, 0xa37f60fd,
+	0xf701269f, 0x5c72f5bc, 0x44663bc5, 0x5bfb7e34,
+	0x8b432976, 0xcb23c6dc, 0xb6edfc68, 0xb8e4f163,
+	0xd731dcca, 0x42638510, 0x13972240, 0x84c61120,
+	0x854a247d, 0xd2bb3df8, 0xaef93211, 0xc729a16d,
+	0x1d9e2f4b, 0xdcb230f3, 0x0d8652ec, 0x77c1e3d0,
+	0x2bb3166c, 0xa970b999, 0x119448fa, 0x47e96422,
+	0xa8fc8cc4, 0xa0f03f1a, 0x567d2cd8, 0x223390ef,
+	0x87494ec7, 0xd938d1c1, 0x8ccaa2fe, 0x98d40b36,
+	0xa6f581cf, 0xa57ade28, 0xdab78e26, 0x3fadbfa4,
+	0x2c3a9de4, 0x5078920d, 0x6a5fcc9b, 0x547e4662,
+	0xf68d13c2, 0x90d8b8e8, 0x2e39f75e, 0x82c3aff5,
+	0x9f5d80be, 0x69d0937c, 0x6fd52da9, 0xcf2512b3,
+	0xc8ac993b, 0x10187da7, 0xe89c636e, 0xdb3bbb7b,
+	0xcd267809, 0x6e5918f4, 0xec9ab701, 0x834f9aa8,
+	0xe6956e65, 0xaaffe67e, 0x21bccf08, 0xef15e8e6,
+	0xbae79bd9, 0x4a6f36ce, 0xea9f09d4, 0x29b07cd6,
+	0x31a4b2af, 0x2a3f2331, 0xc6a59430, 0x35a266c0,
+	0x744ebc37, 0xfc82caa6, 0xe090d0b0, 0x33a7d815,
+	0xf104984a, 0x41ecdaf7, 0x7fcd500e, 0x1791f62f,
+	0x764dd68d, 0x43efb04d, 0xccaa4d54, 0xe49604df,
+	0x9ed1b5e3, 0x4c6a881b, 0xc12c1fb8, 0x4665517f,
+	0x9d5eea04, 0x018c355d, 0xfa877473, 0xfb0b412e,
+	0xb3671d5a, 0x92dbd252, 0xe9105633, 0x6dd64713,
+	0x9ad7618c, 0x37a10c7a, 0x59f8148e, 0xeb133c89,
+	0xcea927ee, 0xb761c935, 0xe11ce5ed, 0x7a47b13c,
+	0x9cd2df59, 0x55f2733f, 0x1814ce79, 0x73c737bf,
+	0x53f7cdea, 0x5ffdaa5b, 0xdf3d6f14, 0x7844db86,
+	0xcaaff381, 0xb968c43e, 0x3824342c, 0xc2a3405f,
+	0x161dc372, 0xbce2250c, 0x283c498b, 0xff0d9541,
+	0x39a80171, 0x080cb3de, 0xd8b4e49c, 0x6456c190,
+	0x7bcb8461, 0xd532b670, 0x486c5c74, 0xd0b85742);
+$code.=<<___;
+# Td4[256]
+.byte	0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38
+.byte	0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb
+.byte	0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87
+.byte	0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb
+.byte	0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d
+.byte	0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e
+.byte	0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2
+.byte	0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25
+.byte	0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16
+.byte	0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92
+.byte	0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda
+.byte	0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84
+.byte	0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a
+.byte	0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06
+.byte	0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02
+.byte	0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b
+.byte	0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea
+.byte	0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73
+.byte	0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85
+.byte	0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e
+.byte	0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89
+.byte	0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b
+.byte	0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20
+.byte	0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4
+.byte	0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31
+.byte	0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f
+.byte	0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d
+.byte	0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef
+.byte	0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0
+.byte	0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61
+.byte	0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26
+.byte	0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
+.size	AES_Td,.-AES_Td
+
+# void AES_decrypt(const unsigned char *inp, unsigned char *out,
+# 		 const AES_KEY *key) {
+.globl	AES_decrypt
+.type	AES_decrypt,\@function
+AES_decrypt:
+___
+$code.=<<___ if (!$softonly);
+	l	%r0,240($key)
+	lhi	%r1,16
+	clr	%r0,%r1
+	jl	.Ldsoft
+
+	la	%r1,0($key)
+	#la	%r2,0($inp)
+	la	%r4,0($out)
+	lghi	%r3,16		# single block length
+	.long	0xb92e0042	# km %r4,%r2
+	brc	1,.-4		# can this happen?
+	br	%r14
+.align	64
+.Ldsoft:
+___
+$code.=<<___;
+	stm${g}	%r3,$ra,3*$SIZE_T($sp)
+
+	llgf	$s0,0($inp)
+	llgf	$s1,4($inp)
+	llgf	$s2,8($inp)
+	llgf	$s3,12($inp)
+
+	larl	$tbl,AES_Td
+	bras	$ra,_s390x_AES_decrypt
+
+	l${g}	$out,3*$SIZE_T($sp)
+	st	$s0,0($out)
+	st	$s1,4($out)
+	st	$s2,8($out)
+	st	$s3,12($out)
+
+	lm${g}	%r6,$ra,6*$SIZE_T($sp)
+	br	$ra
+.size	AES_decrypt,.-AES_decrypt
+
+.type   _s390x_AES_decrypt,\@function
+.align	16
+_s390x_AES_decrypt:
+	st${g}	$ra,15*$SIZE_T($sp)
+	x	$s0,0($key)
+	x	$s1,4($key)
+	x	$s2,8($key)
+	x	$s3,12($key)
+	l	$rounds,240($key)
+	llill	$mask,`0xff<<3`
+	aghi	$rounds,-1
+	j	.Ldec_loop
+.align	16
+.Ldec_loop:
+	srlg	$t1,$s0,`16-3`
+	srlg	$t2,$s0,`8-3`
+	sllg	$t3,$s0,`0+3`
+	srl	$s0,`24-3`
+	nr	$s0,$mask
+	nr	$t1,$mask
+	nr	$t2,$mask
+	ngr	$t3,$mask
+
+	sllg	$i1,$s1,`0+3`	# i0
+	srlg	$i2,$s1,`16-3`
+	srlg	$i3,$s1,`8-3`
+	srl	$s1,`24-3`
+	ngr	$i1,$mask
+	nr	$s1,$mask
+	nr	$i2,$mask
+	nr	$i3,$mask
+
+	l	$s0,0($s0,$tbl)	# Td0[s0>>24]
+	l	$t1,3($t1,$tbl)	# Td1[s0>>16]
+	l	$t2,2($t2,$tbl)	# Td2[s0>>8]
+	l	$t3,1($t3,$tbl)	# Td3[s0>>0]
+
+	x	$s0,1($i1,$tbl)	# Td3[s1>>0]
+	l	$s1,0($s1,$tbl)	# Td0[s1>>24]
+	x	$t2,3($i2,$tbl)	# Td1[s1>>16]
+	x	$t3,2($i3,$tbl)	# Td2[s1>>8]
+
+	srlg	$i1,$s2,`8-3`	# i0
+	sllg	$i2,$s2,`0+3`	# i1
+	srlg	$i3,$s2,`16-3`
+	srl	$s2,`24-3`
+	nr	$i1,$mask
+	ngr	$i2,$mask
+	nr	$s2,$mask
+	nr	$i3,$mask
+
+	xr	$s1,$t1
+	srlg	$ra,$s3,`8-3`	# i1
+	srlg	$t1,$s3,`16-3`	# i0
+	nr	$ra,$mask
+	la	$key,16($key)
+	nr	$t1,$mask
+
+	x	$s0,2($i1,$tbl)	# Td2[s2>>8]
+	x	$s1,1($i2,$tbl)	# Td3[s2>>0]
+	l	$s2,0($s2,$tbl)	# Td0[s2>>24]
+	x	$t3,3($i3,$tbl)	# Td1[s2>>16]
+
+	sllg	$i3,$s3,`0+3`	# i2
+	srl	$s3,`24-3`
+	ngr	$i3,$mask
+	nr	$s3,$mask
+
+	xr	$s2,$t2
+	x	$s0,0($key)
+	x	$s1,4($key)
+	x	$s2,8($key)
+	x	$t3,12($key)
+
+	x	$s0,3($t1,$tbl)	# Td1[s3>>16]
+	x	$s1,2($ra,$tbl)	# Td2[s3>>8]
+	x	$s2,1($i3,$tbl)	# Td3[s3>>0]
+	l	$s3,0($s3,$tbl)	# Td0[s3>>24]
+	xr	$s3,$t3
+
+	brct	$rounds,.Ldec_loop
+	.align	16
+
+	l	$t1,`2048+0`($tbl)	# prefetch Td4
+	l	$t2,`2048+64`($tbl)
+	l	$t3,`2048+128`($tbl)
+	l	$i1,`2048+192`($tbl)
+	llill	$mask,0xff
+
+	srlg	$i3,$s0,24	# i0
+	srlg	$t1,$s0,16
+	srlg	$t2,$s0,8
+	nr	$s0,$mask	# i3
+	nr	$t1,$mask
+
+	srlg	$i1,$s1,24
+	nr	$t2,$mask
+	srlg	$i2,$s1,16
+	srlg	$ra,$s1,8
+	nr	$s1,$mask	# i0
+	nr	$i2,$mask
+	nr	$ra,$mask
+
+	llgc	$i3,2048($i3,$tbl)	# Td4[s0>>24]
+	llgc	$t1,2048($t1,$tbl)	# Td4[s0>>16]
+	llgc	$t2,2048($t2,$tbl)	# Td4[s0>>8]
+	sll	$t1,16
+	llgc	$t3,2048($s0,$tbl)	# Td4[s0>>0]
+	sllg	$s0,$i3,24
+	sll	$t2,8
+
+	llgc	$s1,2048($s1,$tbl)	# Td4[s1>>0]
+	llgc	$i1,2048($i1,$tbl)	# Td4[s1>>24]
+	llgc	$i2,2048($i2,$tbl)	# Td4[s1>>16]
+	sll	$i1,24
+	llgc	$i3,2048($ra,$tbl)	# Td4[s1>>8]
+	sll	$i2,16
+	sll	$i3,8
+	or	$s0,$s1
+	or	$t1,$i1
+	or	$t2,$i2
+	or	$t3,$i3
+
+	srlg	$i1,$s2,8	# i0
+	srlg	$i2,$s2,24
+	srlg	$i3,$s2,16
+	nr	$s2,$mask	# i1
+	nr	$i1,$mask
+	nr	$i3,$mask
+	llgc	$i1,2048($i1,$tbl)	# Td4[s2>>8]
+	llgc	$s1,2048($s2,$tbl)	# Td4[s2>>0]
+	llgc	$i2,2048($i2,$tbl)	# Td4[s2>>24]
+	llgc	$i3,2048($i3,$tbl)	# Td4[s2>>16]
+	sll	$i1,8
+	sll	$i2,24
+	or	$s0,$i1
+	sll	$i3,16
+	or	$t2,$i2
+	or	$t3,$i3
+
+	srlg	$i1,$s3,16	# i0
+	srlg	$i2,$s3,8	# i1
+	srlg	$i3,$s3,24
+	nr	$s3,$mask	# i2
+	nr	$i1,$mask
+	nr	$i2,$mask
+
+	l${g}	$ra,15*$SIZE_T($sp)
+	or	$s1,$t1
+	l	$t0,16($key)
+	l	$t1,20($key)
+
+	llgc	$i1,2048($i1,$tbl)	# Td4[s3>>16]
+	llgc	$i2,2048($i2,$tbl)	# Td4[s3>>8]
+	sll	$i1,16
+	llgc	$s2,2048($s3,$tbl)	# Td4[s3>>0]
+	llgc	$s3,2048($i3,$tbl)	# Td4[s3>>24]
+	sll	$i2,8
+	sll	$s3,24
+	or	$s0,$i1
+	or	$s1,$i2
+	or	$s2,$t2
+	or	$s3,$t3
+
+	xr	$s0,$t0
+	xr	$s1,$t1
+	x	$s2,24($key)
+	x	$s3,28($key)
+
+	br	$ra	
+.size	_s390x_AES_decrypt,.-_s390x_AES_decrypt
+___
+
+$code.=<<___;
+# void AES_set_encrypt_key(const unsigned char *in, int bits,
+# 		 AES_KEY *key) {
+.globl	private_AES_set_encrypt_key
+.type	private_AES_set_encrypt_key,\@function
+.align	16
+private_AES_set_encrypt_key:
+_s390x_AES_set_encrypt_key:
+	lghi	$t0,0
+	cl${g}r	$inp,$t0
+	je	.Lminus1
+	cl${g}r	$key,$t0
+	je	.Lminus1
+
+	lghi	$t0,128
+	clr	$bits,$t0
+	je	.Lproceed
+	lghi	$t0,192
+	clr	$bits,$t0
+	je	.Lproceed
+	lghi	$t0,256
+	clr	$bits,$t0
+	je	.Lproceed
+	lghi	%r2,-2
+	br	%r14
+
+.align	16
+.Lproceed:
+___
+$code.=<<___ if (!$softonly);
+	# convert bits to km code, [128,192,256]->[18,19,20]
+	lhi	%r5,-128
+	lhi	%r0,18
+	ar	%r5,$bits
+	srl	%r5,6
+	ar	%r5,%r0
+
+	larl	%r1,OPENSSL_s390xcap_P
+	lg	%r0,0(%r1)
+	tmhl	%r0,0x4000	# check for message-security assist
+	jz	.Lekey_internal
+
+	lghi	%r0,0		# query capability vector
+	la	%r1,16($sp)
+	.long	0xb92f0042	# kmc %r4,%r2
+
+	llihh	%r1,0x8000
+	srlg	%r1,%r1,0(%r5)
+	ng	%r1,16($sp)
+	jz	.Lekey_internal
+
+	lmg	%r0,%r1,0($inp)	# just copy 128 bits...
+	stmg	%r0,%r1,0($key)
+	lhi	%r0,192
+	cr	$bits,%r0
+	jl	1f
+	lg	%r1,16($inp)
+	stg	%r1,16($key)
+	je	1f
+	lg	%r1,24($inp)
+	stg	%r1,24($key)
+1:	st	$bits,236($key)	# save bits [for debugging purposes]
+	lgr	$t0,%r5
+	st	%r5,240($key)	# save km code
+	lghi	%r2,0
+	br	%r14
+___
+$code.=<<___;
+.align	16
+.Lekey_internal:
+	stm${g}	%r4,%r13,4*$SIZE_T($sp)	# all non-volatile regs and $key
+
+	larl	$tbl,AES_Te+2048
+
+	llgf	$s0,0($inp)
+	llgf	$s1,4($inp)
+	llgf	$s2,8($inp)
+	llgf	$s3,12($inp)
+	st	$s0,0($key)
+	st	$s1,4($key)
+	st	$s2,8($key)
+	st	$s3,12($key)
+	lghi	$t0,128
+	cr	$bits,$t0
+	jne	.Lnot128
+
+	llill	$mask,0xff
+	lghi	$t3,0			# i=0
+	lghi	$rounds,10
+	st	$rounds,240($key)
+
+	llgfr	$t2,$s3			# temp=rk[3]
+	srlg	$i1,$s3,8
+	srlg	$i2,$s3,16
+	srlg	$i3,$s3,24
+	nr	$t2,$mask
+	nr	$i1,$mask
+	nr	$i2,$mask
+
+.align	16
+.L128_loop:
+	la	$t2,0($t2,$tbl)
+	la	$i1,0($i1,$tbl)
+	la	$i2,0($i2,$tbl)
+	la	$i3,0($i3,$tbl)
+	icm	$t2,2,0($t2)		# Te4[rk[3]>>0]<<8
+	icm	$t2,4,0($i1)		# Te4[rk[3]>>8]<<16
+	icm	$t2,8,0($i2)		# Te4[rk[3]>>16]<<24
+	icm	$t2,1,0($i3)		# Te4[rk[3]>>24]
+	x	$t2,256($t3,$tbl)	# rcon[i]
+	xr	$s0,$t2			# rk[4]=rk[0]^...
+	xr	$s1,$s0			# rk[5]=rk[1]^rk[4]
+	xr	$s2,$s1			# rk[6]=rk[2]^rk[5]
+	xr	$s3,$s2			# rk[7]=rk[3]^rk[6]
+
+	llgfr	$t2,$s3			# temp=rk[3]
+	srlg	$i1,$s3,8
+	srlg	$i2,$s3,16
+	nr	$t2,$mask
+	nr	$i1,$mask
+	srlg	$i3,$s3,24
+	nr	$i2,$mask
+
+	st	$s0,16($key)
+	st	$s1,20($key)
+	st	$s2,24($key)
+	st	$s3,28($key)
+	la	$key,16($key)		# key+=4
+	la	$t3,4($t3)		# i++
+	brct	$rounds,.L128_loop
+	lghi	$t0,10
+	lghi	%r2,0
+	lm${g}	%r4,%r13,4*$SIZE_T($sp)
+	br	$ra
+
+.align	16
+.Lnot128:
+	llgf	$t0,16($inp)
+	llgf	$t1,20($inp)
+	st	$t0,16($key)
+	st	$t1,20($key)
+	lghi	$t0,192
+	cr	$bits,$t0
+	jne	.Lnot192
+
+	llill	$mask,0xff
+	lghi	$t3,0			# i=0
+	lghi	$rounds,12
+	st	$rounds,240($key)
+	lghi	$rounds,8
+
+	srlg	$i1,$t1,8
+	srlg	$i2,$t1,16
+	srlg	$i3,$t1,24
+	nr	$t1,$mask
+	nr	$i1,$mask
+	nr	$i2,$mask
+
+.align	16
+.L192_loop:
+	la	$t1,0($t1,$tbl)
+	la	$i1,0($i1,$tbl)
+	la	$i2,0($i2,$tbl)
+	la	$i3,0($i3,$tbl)
+	icm	$t1,2,0($t1)		# Te4[rk[5]>>0]<<8
+	icm	$t1,4,0($i1)		# Te4[rk[5]>>8]<<16
+	icm	$t1,8,0($i2)		# Te4[rk[5]>>16]<<24
+	icm	$t1,1,0($i3)		# Te4[rk[5]>>24]
+	x	$t1,256($t3,$tbl)	# rcon[i]
+	xr	$s0,$t1			# rk[6]=rk[0]^...
+	xr	$s1,$s0			# rk[7]=rk[1]^rk[6]
+	xr	$s2,$s1			# rk[8]=rk[2]^rk[7]
+	xr	$s3,$s2			# rk[9]=rk[3]^rk[8]
+
+	st	$s0,24($key)
+	st	$s1,28($key)
+	st	$s2,32($key)
+	st	$s3,36($key)
+	brct	$rounds,.L192_continue
+	lghi	$t0,12
+	lghi	%r2,0
+	lm${g}	%r4,%r13,4*$SIZE_T($sp)
+	br	$ra
+
+.align	16
+.L192_continue:
+	lgr	$t1,$s3
+	x	$t1,16($key)		# rk[10]=rk[4]^rk[9]
+	st	$t1,40($key)
+	x	$t1,20($key)		# rk[11]=rk[5]^rk[10]
+	st	$t1,44($key)
+
+	srlg	$i1,$t1,8
+	srlg	$i2,$t1,16
+	srlg	$i3,$t1,24
+	nr	$t1,$mask
+	nr	$i1,$mask
+	nr	$i2,$mask
+
+	la	$key,24($key)		# key+=6
+	la	$t3,4($t3)		# i++
+	j	.L192_loop
+
+.align	16
+.Lnot192:
+	llgf	$t0,24($inp)
+	llgf	$t1,28($inp)
+	st	$t0,24($key)
+	st	$t1,28($key)
+	llill	$mask,0xff
+	lghi	$t3,0			# i=0
+	lghi	$rounds,14
+	st	$rounds,240($key)
+	lghi	$rounds,7
+
+	srlg	$i1,$t1,8
+	srlg	$i2,$t1,16
+	srlg	$i3,$t1,24
+	nr	$t1,$mask
+	nr	$i1,$mask
+	nr	$i2,$mask
+
+.align	16
+.L256_loop:
+	la	$t1,0($t1,$tbl)
+	la	$i1,0($i1,$tbl)
+	la	$i2,0($i2,$tbl)
+	la	$i3,0($i3,$tbl)
+	icm	$t1,2,0($t1)		# Te4[rk[7]>>0]<<8
+	icm	$t1,4,0($i1)		# Te4[rk[7]>>8]<<16
+	icm	$t1,8,0($i2)		# Te4[rk[7]>>16]<<24
+	icm	$t1,1,0($i3)		# Te4[rk[7]>>24]
+	x	$t1,256($t3,$tbl)	# rcon[i]
+	xr	$s0,$t1			# rk[8]=rk[0]^...
+	xr	$s1,$s0			# rk[9]=rk[1]^rk[8]
+	xr	$s2,$s1			# rk[10]=rk[2]^rk[9]
+	xr	$s3,$s2			# rk[11]=rk[3]^rk[10]
+	st	$s0,32($key)
+	st	$s1,36($key)
+	st	$s2,40($key)
+	st	$s3,44($key)
+	brct	$rounds,.L256_continue
+	lghi	$t0,14
+	lghi	%r2,0
+	lm${g}	%r4,%r13,4*$SIZE_T($sp)
+	br	$ra
+
+.align	16
+.L256_continue:
+	lgr	$t1,$s3			# temp=rk[11]
+	srlg	$i1,$s3,8
+	srlg	$i2,$s3,16
+	srlg	$i3,$s3,24
+	nr	$t1,$mask
+	nr	$i1,$mask
+	nr	$i2,$mask
+	la	$t1,0($t1,$tbl)
+	la	$i1,0($i1,$tbl)
+	la	$i2,0($i2,$tbl)
+	la	$i3,0($i3,$tbl)
+	llgc	$t1,0($t1)		# Te4[rk[11]>>0]
+	icm	$t1,2,0($i1)		# Te4[rk[11]>>8]<<8
+	icm	$t1,4,0($i2)		# Te4[rk[11]>>16]<<16
+	icm	$t1,8,0($i3)		# Te4[rk[11]>>24]<<24
+	x	$t1,16($key)		# rk[12]=rk[4]^...
+	st	$t1,48($key)
+	x	$t1,20($key)		# rk[13]=rk[5]^rk[12]
+	st	$t1,52($key)
+	x	$t1,24($key)		# rk[14]=rk[6]^rk[13]
+	st	$t1,56($key)
+	x	$t1,28($key)		# rk[15]=rk[7]^rk[14]
+	st	$t1,60($key)
+
+	srlg	$i1,$t1,8
+	srlg	$i2,$t1,16
+	srlg	$i3,$t1,24
+	nr	$t1,$mask
+	nr	$i1,$mask
+	nr	$i2,$mask
+
+	la	$key,32($key)		# key+=8
+	la	$t3,4($t3)		# i++
+	j	.L256_loop
+
+.Lminus1:
+	lghi	%r2,-1
+	br	$ra
+.size	private_AES_set_encrypt_key,.-private_AES_set_encrypt_key
+
+# void AES_set_decrypt_key(const unsigned char *in, int bits,
+# 		 AES_KEY *key) {
+.globl	private_AES_set_decrypt_key
+.type	private_AES_set_decrypt_key,\@function
+.align	16
+private_AES_set_decrypt_key:
+	#st${g}	$key,4*$SIZE_T($sp)	# I rely on AES_set_encrypt_key to
+	st${g}	$ra,14*$SIZE_T($sp)	# save non-volatile registers and $key!
+	bras	$ra,_s390x_AES_set_encrypt_key
+	#l${g}	$key,4*$SIZE_T($sp)
+	l${g}	$ra,14*$SIZE_T($sp)
+	ltgr	%r2,%r2
+	bnzr	$ra
+___
+$code.=<<___ if (!$softonly);
+	#l	$t0,240($key)
+	lhi	$t1,16
+	cr	$t0,$t1
+	jl	.Lgo
+	oill	$t0,0x80	# set "decrypt" bit
+	st	$t0,240($key)
+	br	$ra
+___
+$code.=<<___;
+.align	16
+.Lgo:	lgr	$rounds,$t0	#llgf	$rounds,240($key)
+	la	$i1,0($key)
+	sllg	$i2,$rounds,4
+	la	$i2,0($i2,$key)
+	srl	$rounds,1
+	lghi	$t1,-16
+
+.align	16
+.Linv:	lmg	$s0,$s1,0($i1)
+	lmg	$s2,$s3,0($i2)
+	stmg	$s0,$s1,0($i2)
+	stmg	$s2,$s3,0($i1)
+	la	$i1,16($i1)
+	la	$i2,0($t1,$i2)
+	brct	$rounds,.Linv
+___
+$mask80=$i1;
+$mask1b=$i2;
+$maskfe=$i3;
+$code.=<<___;
+	llgf	$rounds,240($key)
+	aghi	$rounds,-1
+	sll	$rounds,2	# (rounds-1)*4
+	llilh	$mask80,0x8080
+	llilh	$mask1b,0x1b1b
+	llilh	$maskfe,0xfefe
+	oill	$mask80,0x8080
+	oill	$mask1b,0x1b1b
+	oill	$maskfe,0xfefe
+
+.align	16
+.Lmix:	l	$s0,16($key)	# tp1
+	lr	$s1,$s0
+	ngr	$s1,$mask80
+	srlg	$t1,$s1,7
+	slr	$s1,$t1
+	nr	$s1,$mask1b
+	sllg	$t1,$s0,1
+	nr	$t1,$maskfe
+	xr	$s1,$t1		# tp2
+
+	lr	$s2,$s1
+	ngr	$s2,$mask80
+	srlg	$t1,$s2,7
+	slr	$s2,$t1
+	nr	$s2,$mask1b
+	sllg	$t1,$s1,1
+	nr	$t1,$maskfe
+	xr	$s2,$t1		# tp4
+
+	lr	$s3,$s2
+	ngr	$s3,$mask80
+	srlg	$t1,$s3,7
+	slr	$s3,$t1
+	nr	$s3,$mask1b
+	sllg	$t1,$s2,1
+	nr	$t1,$maskfe
+	xr	$s3,$t1		# tp8
+
+	xr	$s1,$s0		# tp2^tp1
+	xr	$s2,$s0		# tp4^tp1
+	rll	$s0,$s0,24	# = ROTATE(tp1,8)
+	xr	$s2,$s3		# ^=tp8
+	xr	$s0,$s1		# ^=tp2^tp1
+	xr	$s1,$s3		# tp2^tp1^tp8
+	xr	$s0,$s2		# ^=tp4^tp1^tp8
+	rll	$s1,$s1,8
+	rll	$s2,$s2,16
+	xr	$s0,$s1		# ^= ROTATE(tp8^tp2^tp1,24)
+	rll	$s3,$s3,24
+	xr	$s0,$s2    	# ^= ROTATE(tp8^tp4^tp1,16)
+	xr	$s0,$s3		# ^= ROTATE(tp8,8)
+
+	st	$s0,16($key)
+	la	$key,4($key)
+	brct	$rounds,.Lmix
+
+	lm${g}	%r6,%r13,6*$SIZE_T($sp)# as was saved by AES_set_encrypt_key!
+	lghi	%r2,0
+	br	$ra
+.size	private_AES_set_decrypt_key,.-private_AES_set_decrypt_key
+___
+
+########################################################################
+# void AES_cbc_encrypt(const unsigned char *in, unsigned char *out,
+#                     size_t length, const AES_KEY *key,
+#                     unsigned char *ivec, const int enc)
+{
+my $inp="%r2";
+my $out="%r4";	# length and out are swapped
+my $len="%r3";
+my $key="%r5";
+my $ivp="%r6";
+
+$code.=<<___;
+.globl	AES_cbc_encrypt
+.type	AES_cbc_encrypt,\@function
+.align	16
+AES_cbc_encrypt:
+	xgr	%r3,%r4		# flip %r3 and %r4, out and len
+	xgr	%r4,%r3
+	xgr	%r3,%r4
+___
+$code.=<<___ if (!$softonly);
+	lhi	%r0,16
+	cl	%r0,240($key)
+	jh	.Lcbc_software
+
+	lg	%r0,0($ivp)	# copy ivec
+	lg	%r1,8($ivp)
+	stmg	%r0,%r1,16($sp)
+	lmg	%r0,%r1,0($key)	# copy key, cover 256 bit
+	stmg	%r0,%r1,32($sp)
+	lmg	%r0,%r1,16($key)
+	stmg	%r0,%r1,48($sp)
+	l	%r0,240($key)	# load kmc code
+	lghi	$key,15		# res=len%16, len-=res;
+	ngr	$key,$len
+	sl${g}r	$len,$key
+	la	%r1,16($sp)	# parameter block - ivec || key
+	jz	.Lkmc_truncated
+	.long	0xb92f0042	# kmc %r4,%r2
+	brc	1,.-4		# pay attention to "partial completion"
+	ltr	$key,$key
+	jnz	.Lkmc_truncated
+.Lkmc_done:
+	lmg	%r0,%r1,16($sp)	# copy ivec to caller
+	stg	%r0,0($ivp)
+	stg	%r1,8($ivp)
+	br	$ra
+.align	16
+.Lkmc_truncated:
+	ahi	$key,-1		# it's the way it's encoded in mvc
+	tmll	%r0,0x80
+	jnz	.Lkmc_truncated_dec
+	lghi	%r1,0
+	stg	%r1,16*$SIZE_T($sp)
+	stg	%r1,16*$SIZE_T+8($sp)
+	bras	%r1,1f
+	mvc	16*$SIZE_T(1,$sp),0($inp)
+1:	ex	$key,0(%r1)
+	la	%r1,16($sp)	# restore parameter block
+	la	$inp,16*$SIZE_T($sp)
+	lghi	$len,16
+	.long	0xb92f0042	# kmc %r4,%r2
+	j	.Lkmc_done
+.align	16
+.Lkmc_truncated_dec:
+	st${g}	$out,4*$SIZE_T($sp)
+	la	$out,16*$SIZE_T($sp)
+	lghi	$len,16
+	.long	0xb92f0042	# kmc %r4,%r2
+	l${g}	$out,4*$SIZE_T($sp)
+	bras	%r1,2f
+	mvc	0(1,$out),16*$SIZE_T($sp)
+2:	ex	$key,0(%r1)
+	j	.Lkmc_done
+.align	16
+.Lcbc_software:
+___
+$code.=<<___;
+	stm${g}	$key,$ra,5*$SIZE_T($sp)
+	lhi	%r0,0
+	cl	%r0,`$stdframe+$SIZE_T-4`($sp)
+	je	.Lcbc_decrypt
+
+	larl	$tbl,AES_Te
+
+	llgf	$s0,0($ivp)
+	llgf	$s1,4($ivp)
+	llgf	$s2,8($ivp)
+	llgf	$s3,12($ivp)
+
+	lghi	$t0,16
+	sl${g}r	$len,$t0
+	brc	4,.Lcbc_enc_tail	# if borrow
+.Lcbc_enc_loop:
+	stm${g}	$inp,$out,2*$SIZE_T($sp)
+	x	$s0,0($inp)
+	x	$s1,4($inp)
+	x	$s2,8($inp)
+	x	$s3,12($inp)
+	lgr	%r4,$key
+
+	bras	$ra,_s390x_AES_encrypt
+
+	lm${g}	$inp,$key,2*$SIZE_T($sp)
+	st	$s0,0($out)
+	st	$s1,4($out)
+	st	$s2,8($out)
+	st	$s3,12($out)
+
+	la	$inp,16($inp)
+	la	$out,16($out)
+	lghi	$t0,16
+	lt${g}r	$len,$len
+	jz	.Lcbc_enc_done
+	sl${g}r	$len,$t0
+	brc	4,.Lcbc_enc_tail	# if borrow
+	j	.Lcbc_enc_loop
+.align	16
+.Lcbc_enc_done:
+	l${g}	$ivp,6*$SIZE_T($sp)
+	st	$s0,0($ivp)
+	st	$s1,4($ivp)	
+	st	$s2,8($ivp)
+	st	$s3,12($ivp)
+
+	lm${g}	%r7,$ra,7*$SIZE_T($sp)
+	br	$ra
+
+.align	16
+.Lcbc_enc_tail:
+	aghi	$len,15
+	lghi	$t0,0
+	stg	$t0,16*$SIZE_T($sp)
+	stg	$t0,16*$SIZE_T+8($sp)
+	bras	$t1,3f
+	mvc	16*$SIZE_T(1,$sp),0($inp)
+3:	ex	$len,0($t1)
+	lghi	$len,0
+	la	$inp,16*$SIZE_T($sp)
+	j	.Lcbc_enc_loop
+
+.align	16
+.Lcbc_decrypt:
+	larl	$tbl,AES_Td
+
+	lg	$t0,0($ivp)
+	lg	$t1,8($ivp)
+	stmg	$t0,$t1,16*$SIZE_T($sp)
+
+.Lcbc_dec_loop:
+	stm${g}	$inp,$out,2*$SIZE_T($sp)
+	llgf	$s0,0($inp)
+	llgf	$s1,4($inp)
+	llgf	$s2,8($inp)
+	llgf	$s3,12($inp)
+	lgr	%r4,$key
+
+	bras	$ra,_s390x_AES_decrypt
+
+	lm${g}	$inp,$key,2*$SIZE_T($sp)
+	sllg	$s0,$s0,32
+	sllg	$s2,$s2,32
+	lr	$s0,$s1
+	lr	$s2,$s3
+
+	lg	$t0,0($inp)
+	lg	$t1,8($inp)
+	xg	$s0,16*$SIZE_T($sp)
+	xg	$s2,16*$SIZE_T+8($sp)
+	lghi	$s1,16
+	sl${g}r	$len,$s1
+	brc	4,.Lcbc_dec_tail	# if borrow
+	brc	2,.Lcbc_dec_done	# if zero
+	stg	$s0,0($out)
+	stg	$s2,8($out)
+	stmg	$t0,$t1,16*$SIZE_T($sp)
+
+	la	$inp,16($inp)
+	la	$out,16($out)
+	j	.Lcbc_dec_loop
+
+.Lcbc_dec_done:
+	stg	$s0,0($out)
+	stg	$s2,8($out)
+.Lcbc_dec_exit:
+	lm${g}	%r6,$ra,6*$SIZE_T($sp)
+	stmg	$t0,$t1,0($ivp)
+
+	br	$ra
+
+.align	16
+.Lcbc_dec_tail:
+	aghi	$len,15
+	stg	$s0,16*$SIZE_T($sp)
+	stg	$s2,16*$SIZE_T+8($sp)
+	bras	$s1,4f
+	mvc	0(1,$out),16*$SIZE_T($sp)
+4:	ex	$len,0($s1)
+	j	.Lcbc_dec_exit
+.size	AES_cbc_encrypt,.-AES_cbc_encrypt
+___
+}
+########################################################################
+# void AES_ctr32_encrypt(const unsigned char *in, unsigned char *out,
+#                     size_t blocks, const AES_KEY *key,
+#                     const unsigned char *ivec)
+{
+my $inp="%r2";
+my $out="%r4";	# blocks and out are swapped
+my $len="%r3";
+my $key="%r5";	my $iv0="%r5";
+my $ivp="%r6";
+my $fp ="%r7";
+
+$code.=<<___;
+.globl	AES_ctr32_encrypt
+.type	AES_ctr32_encrypt,\@function
+.align	16
+AES_ctr32_encrypt:
+	xgr	%r3,%r4		# flip %r3 and %r4, $out and $len
+	xgr	%r4,%r3
+	xgr	%r3,%r4
+	llgfr	$len,$len	# safe in ctr32 subroutine even in 64-bit case
+___
+$code.=<<___ if (!$softonly);
+	l	%r0,240($key)
+	lhi	%r1,16
+	clr	%r0,%r1
+	jl	.Lctr32_software
+
+	stm${g}	%r6,$s3,6*$SIZE_T($sp)
+
+	slgr	$out,$inp
+	la	%r1,0($key)	# %r1 is permanent copy of $key
+	lg	$iv0,0($ivp)	# load ivec
+	lg	$ivp,8($ivp)
+
+	# prepare and allocate stack frame at the top of 4K page
+	# with 1K reserved for eventual signal handling
+	lghi	$s0,-1024-256-16# guarantee at least 256-bytes buffer
+	lghi	$s1,-4096
+	algr	$s0,$sp
+	lgr	$fp,$sp
+	ngr	$s0,$s1		# align at page boundary
+	slgr	$fp,$s0		# total buffer size
+	lgr	$s2,$sp
+	lghi	$s1,1024+16	# sl[g]fi is extended-immediate facility
+	slgr	$fp,$s1		# deduct reservation to get usable buffer size
+	# buffer size is at lest 256 and at most 3072+256-16
+
+	la	$sp,1024($s0)	# alloca
+	srlg	$fp,$fp,4	# convert bytes to blocks, minimum 16
+	st${g}	$s2,0($sp)	# back-chain
+	st${g}	$fp,$SIZE_T($sp)
+
+	slgr	$len,$fp
+	brc	1,.Lctr32_hw_switch	# not zero, no borrow
+	algr	$fp,$len	# input is shorter than allocated buffer
+	lghi	$len,0
+	st${g}	$fp,$SIZE_T($sp)
+
+.Lctr32_hw_switch:
+___
+$code.=<<___ if (0);	######### kmctr code was measured to be ~12% slower
+	larl	$s0,OPENSSL_s390xcap_P
+	lg	$s0,8($s0)
+	tmhh	$s0,0x0004	# check for message_security-assist-4
+	jz	.Lctr32_km_loop
+
+	llgfr	$s0,%r0
+	lgr	$s1,%r1
+	lghi	%r0,0
+	la	%r1,16($sp)
+	.long	0xb92d2042	# kmctr %r4,%r2,%r2
+
+	llihh	%r0,0x8000	# check if kmctr supports the function code
+	srlg	%r0,%r0,0($s0)
+	ng	%r0,16($sp)
+	lgr	%r0,$s0
+	lgr	%r1,$s1
+	jz	.Lctr32_km_loop
+
+####### kmctr code
+	algr	$out,$inp	# restore $out
+	lgr	$s1,$len	# $s1 undertakes $len
+	j	.Lctr32_kmctr_loop
+.align	16
+.Lctr32_kmctr_loop:
+	la	$s2,16($sp)
+	lgr	$s3,$fp
+.Lctr32_kmctr_prepare:
+	stg	$iv0,0($s2)
+	stg	$ivp,8($s2)
+	la	$s2,16($s2)
+	ahi	$ivp,1		# 32-bit increment, preserves upper half
+	brct	$s3,.Lctr32_kmctr_prepare
+
+	#la	$inp,0($inp)	# inp
+	sllg	$len,$fp,4	# len
+	#la	$out,0($out)	# out
+	la	$s2,16($sp)	# iv
+	.long	0xb92da042	# kmctr $out,$s2,$inp
+	brc	1,.-4		# pay attention to "partial completion"
+
+	slgr	$s1,$fp
+	brc	1,.Lctr32_kmctr_loop	# not zero, no borrow
+	algr	$fp,$s1
+	lghi	$s1,0
+	brc	4+1,.Lctr32_kmctr_loop	# not zero
+
+	l${g}	$sp,0($sp)
+	lm${g}	%r6,$s3,6*$SIZE_T($sp)
+	br	$ra
+.align	16
+___
+$code.=<<___;
+.Lctr32_km_loop:
+	la	$s2,16($sp)
+	lgr	$s3,$fp
+.Lctr32_km_prepare:
+	stg	$iv0,0($s2)
+	stg	$ivp,8($s2)
+	la	$s2,16($s2)
+	ahi	$ivp,1		# 32-bit increment, preserves upper half
+	brct	$s3,.Lctr32_km_prepare
+
+	la	$s0,16($sp)	# inp
+	sllg	$s1,$fp,4	# len
+	la	$s2,16($sp)	# out
+	.long	0xb92e00a8	# km %r10,%r8
+	brc	1,.-4		# pay attention to "partial completion"
+
+	la	$s2,16($sp)
+	lgr	$s3,$fp
+	slgr	$s2,$inp
+.Lctr32_km_xor:
+	lg	$s0,0($inp)
+	lg	$s1,8($inp)
+	xg	$s0,0($s2,$inp)
+	xg	$s1,8($s2,$inp)
+	stg	$s0,0($out,$inp)
+	stg	$s1,8($out,$inp)
+	la	$inp,16($inp)
+	brct	$s3,.Lctr32_km_xor
+
+	slgr	$len,$fp
+	brc	1,.Lctr32_km_loop	# not zero, no borrow
+	algr	$fp,$len
+	lghi	$len,0
+	brc	4+1,.Lctr32_km_loop	# not zero
+
+	l${g}	$s0,0($sp)
+	l${g}	$s1,$SIZE_T($sp)
+	la	$s2,16($sp)
+.Lctr32_km_zap:
+	stg	$s0,0($s2)
+	stg	$s0,8($s2)
+	la	$s2,16($s2)
+	brct	$s1,.Lctr32_km_zap
+
+	la	$sp,0($s0)
+	lm${g}	%r6,$s3,6*$SIZE_T($sp)
+	br	$ra
+.align	16
+.Lctr32_software:
+___
+$code.=<<___;
+	stm${g}	$key,$ra,5*$SIZE_T($sp)
+	sl${g}r	$inp,$out
+	larl	$tbl,AES_Te
+	llgf	$t1,12($ivp)
+
+.Lctr32_loop:
+	stm${g}	$inp,$out,2*$SIZE_T($sp)
+	llgf	$s0,0($ivp)
+	llgf	$s1,4($ivp)
+	llgf	$s2,8($ivp)
+	lgr	$s3,$t1
+	st	$t1,16*$SIZE_T($sp)
+	lgr	%r4,$key
+
+	bras	$ra,_s390x_AES_encrypt
+
+	lm${g}	$inp,$ivp,2*$SIZE_T($sp)
+	llgf	$t1,16*$SIZE_T($sp)
+	x	$s0,0($inp,$out)
+	x	$s1,4($inp,$out)
+	x	$s2,8($inp,$out)
+	x	$s3,12($inp,$out)
+	stm	$s0,$s3,0($out)
+
+	la	$out,16($out)
+	ahi	$t1,1		# 32-bit increment
+	brct	$len,.Lctr32_loop
+
+	lm${g}	%r6,$ra,6*$SIZE_T($sp)
+	br	$ra
+.size	AES_ctr32_encrypt,.-AES_ctr32_encrypt
+___
+}
+
+########################################################################
+# void AES_xts_encrypt(const char *inp,char *out,size_t len,
+#	const AES_KEY *key1, const AES_KEY *key2,
+#	const unsigned char iv[16]);
+#
+{
+my $inp="%r2";
+my $out="%r4";	# len and out are swapped
+my $len="%r3";
+my $key1="%r5";	# $i1
+my $key2="%r6";	# $i2
+my $fp="%r7";	# $i3
+my $tweak=16*$SIZE_T+16;	# or $stdframe-16, bottom of the frame...
+
+$code.=<<___;
+.type	_s390x_xts_km,\@function
+.align	16
+_s390x_xts_km:
+___
+$code.=<<___ if(1);
+	llgfr	$s0,%r0			# put aside the function code
+	lghi	$s1,0x7f
+	nr	$s1,%r0
+	lghi	%r0,0			# query capability vector
+	la	%r1,2*$SIZE_T($sp)
+	.long	0xb92e0042		# km %r4,%r2
+	llihh	%r1,0x8000
+	srlg	%r1,%r1,32($s1)		# check for 32+function code
+	ng	%r1,2*$SIZE_T($sp)
+	lgr	%r0,$s0			# restore the function code
+	la	%r1,0($key1)		# restore $key1
+	jz	.Lxts_km_vanilla
+
+	lmg	$i2,$i3,$tweak($sp)	# put aside the tweak value
+	algr	$out,$inp
+
+	oill	%r0,32			# switch to xts function code
+	aghi	$s1,-18			#
+	sllg	$s1,$s1,3		# (function code - 18)*8, 0 or 16
+	la	%r1,$tweak-16($sp)
+	slgr	%r1,$s1			# parameter block position
+	lmg	$s0,$s3,0($key1)	# load 256 bits of key material,
+	stmg	$s0,$s3,0(%r1)		# and copy it to parameter block.
+					# yes, it contains junk and overlaps
+					# with the tweak in 128-bit case.
+					# it's done to avoid conditional
+					# branch.
+	stmg	$i2,$i3,$tweak($sp)	# "re-seat" the tweak value
+
+	.long	0xb92e0042		# km %r4,%r2
+	brc	1,.-4			# pay attention to "partial completion"
+
+	lrvg	$s0,$tweak+0($sp)	# load the last tweak
+	lrvg	$s1,$tweak+8($sp)
+	stmg	%r0,%r3,$tweak-32(%r1)	# wipe copy of the key
+
+	nill	%r0,0xffdf		# switch back to original function code
+	la	%r1,0($key1)		# restore pointer to $key1
+	slgr	$out,$inp
+
+	llgc	$len,2*$SIZE_T-1($sp)
+	nill	$len,0x0f		# $len%=16
+	br	$ra
+	
+.align	16
+.Lxts_km_vanilla:
+___
+$code.=<<___;
+	# prepare and allocate stack frame at the top of 4K page
+	# with 1K reserved for eventual signal handling
+	lghi	$s0,-1024-256-16# guarantee at least 256-bytes buffer
+	lghi	$s1,-4096
+	algr	$s0,$sp
+	lgr	$fp,$sp
+	ngr	$s0,$s1		# align at page boundary
+	slgr	$fp,$s0		# total buffer size
+	lgr	$s2,$sp
+	lghi	$s1,1024+16	# sl[g]fi is extended-immediate facility
+	slgr	$fp,$s1		# deduct reservation to get usable buffer size
+	# buffer size is at lest 256 and at most 3072+256-16
+
+	la	$sp,1024($s0)	# alloca
+	nill	$fp,0xfff0	# round to 16*n
+	st${g}	$s2,0($sp)	# back-chain
+	nill	$len,0xfff0	# redundant
+	st${g}	$fp,$SIZE_T($sp)
+
+	slgr	$len,$fp
+	brc	1,.Lxts_km_go	# not zero, no borrow
+	algr	$fp,$len	# input is shorter than allocated buffer
+	lghi	$len,0
+	st${g}	$fp,$SIZE_T($sp)
+
+.Lxts_km_go:
+	lrvg	$s0,$tweak+0($s2)	# load the tweak value in little-endian
+	lrvg	$s1,$tweak+8($s2)
+
+	la	$s2,16($sp)		# vector of ascending tweak values
+	slgr	$s2,$inp
+	srlg	$s3,$fp,4
+	j	.Lxts_km_start
+
+.Lxts_km_loop:
+	la	$s2,16($sp)
+	slgr	$s2,$inp
+	srlg	$s3,$fp,4
+.Lxts_km_prepare:
+	lghi	$i1,0x87
+	srag	$i2,$s1,63		# broadcast upper bit
+	ngr	$i1,$i2			# rem
+	srlg	$i2,$s0,63		# carry bit from lower half
+	sllg	$s0,$s0,1
+	sllg	$s1,$s1,1
+	xgr	$s0,$i1
+	ogr	$s1,$i2
+.Lxts_km_start:
+	lrvgr	$i1,$s0			# flip byte order
+	lrvgr	$i2,$s1
+	stg	$i1,0($s2,$inp)
+	stg	$i2,8($s2,$inp)
+	xg	$i1,0($inp)
+	xg	$i2,8($inp)
+	stg	$i1,0($out,$inp)
+	stg	$i2,8($out,$inp)
+	la	$inp,16($inp)
+	brct	$s3,.Lxts_km_prepare
+
+	slgr	$inp,$fp		# rewind $inp
+	la	$s2,0($out,$inp)
+	lgr	$s3,$fp
+	.long	0xb92e00aa		# km $s2,$s2
+	brc	1,.-4			# pay attention to "partial completion"
+
+	la	$s2,16($sp)
+	slgr	$s2,$inp
+	srlg	$s3,$fp,4
+.Lxts_km_xor:
+	lg	$i1,0($out,$inp)
+	lg	$i2,8($out,$inp)
+	xg	$i1,0($s2,$inp)
+	xg	$i2,8($s2,$inp)
+	stg	$i1,0($out,$inp)
+	stg	$i2,8($out,$inp)
+	la	$inp,16($inp)
+	brct	$s3,.Lxts_km_xor
+
+	slgr	$len,$fp
+	brc	1,.Lxts_km_loop		# not zero, no borrow
+	algr	$fp,$len
+	lghi	$len,0
+	brc	4+1,.Lxts_km_loop	# not zero
+
+	l${g}	$i1,0($sp)		# back-chain
+	llgf	$fp,`2*$SIZE_T-4`($sp)	# bytes used
+	la	$i2,16($sp)
+	srlg	$fp,$fp,4
+.Lxts_km_zap:
+	stg	$i1,0($i2)
+	stg	$i1,8($i2)
+	la	$i2,16($i2)
+	brct	$fp,.Lxts_km_zap
+
+	la	$sp,0($i1)
+	llgc	$len,2*$SIZE_T-1($i1)
+	nill	$len,0x0f		# $len%=16
+	bzr	$ra
+
+	# generate one more tweak...
+	lghi	$i1,0x87
+	srag	$i2,$s1,63		# broadcast upper bit
+	ngr	$i1,$i2			# rem
+	srlg	$i2,$s0,63		# carry bit from lower half
+	sllg	$s0,$s0,1
+	sllg	$s1,$s1,1
+	xgr	$s0,$i1
+	ogr	$s1,$i2
+
+	ltr	$len,$len		# clear zero flag
+	br	$ra
+.size	_s390x_xts_km,.-_s390x_xts_km
+
+.globl	AES_xts_encrypt
+.type	AES_xts_encrypt,\@function
+.align	16
+AES_xts_encrypt:
+	xgr	%r3,%r4			# flip %r3 and %r4, $out and $len
+	xgr	%r4,%r3
+	xgr	%r3,%r4
+___
+$code.=<<___ if ($SIZE_T==4);
+	llgfr	$len,$len
+___
+$code.=<<___;
+	st${g}	$len,1*$SIZE_T($sp)	# save copy of $len
+	srag	$len,$len,4		# formally wrong, because it expands
+					# sign byte, but who can afford asking
+					# to process more than 2^63-1 bytes?
+					# I use it, because it sets condition
+					# code...
+	bcr	8,$ra			# abort if zero (i.e. less than 16)
+___
+$code.=<<___ if (!$softonly);
+	llgf	%r0,240($key2)
+	lhi	%r1,16
+	clr	%r0,%r1
+	jl	.Lxts_enc_software
+
+	stm${g}	%r6,$s3,6*$SIZE_T($sp)
+	st${g}	$ra,14*$SIZE_T($sp)
+
+	sllg	$len,$len,4		# $len&=~15
+	slgr	$out,$inp
+
+	# generate the tweak value
+	l${g}	$s3,$stdframe($sp)	# pointer to iv
+	la	$s2,$tweak($sp)
+	lmg	$s0,$s1,0($s3)
+	lghi	$s3,16
+	stmg	$s0,$s1,0($s2)
+	la	%r1,0($key2)		# $key2 is not needed anymore
+	.long	0xb92e00aa		# km $s2,$s2, generate the tweak
+	brc	1,.-4			# can this happen?
+
+	l	%r0,240($key1)
+	la	%r1,0($key1)		# $key1 is not needed anymore
+	bras	$ra,_s390x_xts_km
+	jz	.Lxts_enc_km_done
+
+	aghi	$inp,-16		# take one step back
+	la	$i3,0($out,$inp)	# put aside real $out
+.Lxts_enc_km_steal:
+	llgc	$i1,16($inp)
+	llgc	$i2,0($out,$inp)
+	stc	$i1,0($out,$inp)
+	stc	$i2,16($out,$inp)
+	la	$inp,1($inp)
+	brct	$len,.Lxts_enc_km_steal
+
+	la	$s2,0($i3)
+	lghi	$s3,16
+	lrvgr	$i1,$s0			# flip byte order
+	lrvgr	$i2,$s1
+	xg	$i1,0($s2)
+	xg	$i2,8($s2)
+	stg	$i1,0($s2)
+	stg	$i2,8($s2)
+	.long	0xb92e00aa		# km $s2,$s2
+	brc	1,.-4			# can this happen?
+	lrvgr	$i1,$s0			# flip byte order
+	lrvgr	$i2,$s1
+	xg	$i1,0($i3)
+	xg	$i2,8($i3)
+	stg	$i1,0($i3)
+	stg	$i2,8($i3)
+
+.Lxts_enc_km_done:
+	l${g}	$ra,14*$SIZE_T($sp)
+	st${g}	$sp,$tweak($sp)		# wipe tweak
+	st${g}	$sp,$tweak($sp)
+	lm${g}	%r6,$s3,6*$SIZE_T($sp)
+	br	$ra
+.align	16
+.Lxts_enc_software:
+___
+$code.=<<___;
+	stm${g}	%r6,$ra,6*$SIZE_T($sp)
+
+	slgr	$out,$inp
+
+	xgr	$s0,$s0			# clear upper half
+	xgr	$s1,$s1
+	lrv	$s0,$stdframe+4($sp)	# load secno
+	lrv	$s1,$stdframe+0($sp)
+	xgr	$s2,$s2
+	xgr	$s3,$s3
+	stm${g}	%r2,%r5,2*$SIZE_T($sp)
+	la	$key,0($key2)
+	larl	$tbl,AES_Te
+	bras	$ra,_s390x_AES_encrypt	# generate the tweak
+	lm${g}	%r2,%r5,2*$SIZE_T($sp)
+	stm	$s0,$s3,$tweak($sp)	# save the tweak
+	j	.Lxts_enc_enter
+
+.align	16
+.Lxts_enc_loop:
+	lrvg	$s1,$tweak+0($sp)	# load the tweak in little-endian
+	lrvg	$s3,$tweak+8($sp)
+	lghi	%r1,0x87
+	srag	%r0,$s3,63		# broadcast upper bit
+	ngr	%r1,%r0			# rem
+	srlg	%r0,$s1,63		# carry bit from lower half
+	sllg	$s1,$s1,1
+	sllg	$s3,$s3,1
+	xgr	$s1,%r1
+	ogr	$s3,%r0
+	lrvgr	$s1,$s1			# flip byte order
+	lrvgr	$s3,$s3
+	srlg	$s0,$s1,32		# smash the tweak to 4x32-bits 
+	stg	$s1,$tweak+0($sp)	# save the tweak
+	llgfr	$s1,$s1
+	srlg	$s2,$s3,32
+	stg	$s3,$tweak+8($sp)
+	llgfr	$s3,$s3
+	la	$inp,16($inp)		# $inp+=16
+.Lxts_enc_enter:
+	x	$s0,0($inp)		# ^=*($inp)
+	x	$s1,4($inp)
+	x	$s2,8($inp)
+	x	$s3,12($inp)
+	stm${g}	%r2,%r3,2*$SIZE_T($sp)	# only two registers are changing
+	la	$key,0($key1)
+	bras	$ra,_s390x_AES_encrypt
+	lm${g}	%r2,%r5,2*$SIZE_T($sp)
+	x	$s0,$tweak+0($sp)	# ^=tweak
+	x	$s1,$tweak+4($sp)
+	x	$s2,$tweak+8($sp)
+	x	$s3,$tweak+12($sp)
+	st	$s0,0($out,$inp)
+	st	$s1,4($out,$inp)
+	st	$s2,8($out,$inp)
+	st	$s3,12($out,$inp)
+	brct${g}	$len,.Lxts_enc_loop
+
+	llgc	$len,`2*$SIZE_T-1`($sp)
+	nill	$len,0x0f		# $len%16
+	jz	.Lxts_enc_done
+
+	la	$i3,0($inp,$out)	# put aside real $out
+.Lxts_enc_steal:
+	llgc	%r0,16($inp)
+	llgc	%r1,0($out,$inp)
+	stc	%r0,0($out,$inp)
+	stc	%r1,16($out,$inp)
+	la	$inp,1($inp)
+	brct	$len,.Lxts_enc_steal
+	la	$out,0($i3)		# restore real $out
+
+	# generate last tweak...
+	lrvg	$s1,$tweak+0($sp)	# load the tweak in little-endian
+	lrvg	$s3,$tweak+8($sp)
+	lghi	%r1,0x87
+	srag	%r0,$s3,63		# broadcast upper bit
+	ngr	%r1,%r0			# rem
+	srlg	%r0,$s1,63		# carry bit from lower half
+	sllg	$s1,$s1,1
+	sllg	$s3,$s3,1
+	xgr	$s1,%r1
+	ogr	$s3,%r0
+	lrvgr	$s1,$s1			# flip byte order
+	lrvgr	$s3,$s3
+	srlg	$s0,$s1,32		# smash the tweak to 4x32-bits 
+	stg	$s1,$tweak+0($sp)	# save the tweak
+	llgfr	$s1,$s1
+	srlg	$s2,$s3,32
+	stg	$s3,$tweak+8($sp)
+	llgfr	$s3,$s3
+
+	x	$s0,0($out)		# ^=*(inp)|stolen cipther-text
+	x	$s1,4($out)
+	x	$s2,8($out)
+	x	$s3,12($out)
+	st${g}	$out,4*$SIZE_T($sp)
+	la	$key,0($key1)
+	bras	$ra,_s390x_AES_encrypt
+	l${g}	$out,4*$SIZE_T($sp)
+	x	$s0,`$tweak+0`($sp)	# ^=tweak
+	x	$s1,`$tweak+4`($sp)
+	x	$s2,`$tweak+8`($sp)
+	x	$s3,`$tweak+12`($sp)
+	st	$s0,0($out)
+	st	$s1,4($out)
+	st	$s2,8($out)
+	st	$s3,12($out)
+
+.Lxts_enc_done:
+	stg	$sp,$tweak+0($sp)	# wipe tweak
+	stg	$sp,$twesk+8($sp)
+	lm${g}	%r6,$ra,6*$SIZE_T($sp)
+	br	$ra
+.size	AES_xts_encrypt,.-AES_xts_encrypt
+___
+# void AES_xts_decrypt(const char *inp,char *out,size_t len,
+#	const AES_KEY *key1, const AES_KEY *key2,u64 secno);
+#
+$code.=<<___;
+.globl	AES_xts_decrypt
+.type	AES_xts_decrypt,\@function
+.align	16
+AES_xts_decrypt:
+	xgr	%r3,%r4			# flip %r3 and %r4, $out and $len
+	xgr	%r4,%r3
+	xgr	%r3,%r4
+___
+$code.=<<___ if ($SIZE_T==4);
+	llgfr	$len,$len
+___
+$code.=<<___;
+	st${g}	$len,1*$SIZE_T($sp)	# save copy of $len
+	aghi	$len,-16
+	bcr	4,$ra			# abort if less than zero. formally
+					# wrong, because $len is unsigned,
+					# but who can afford asking to
+					# process more than 2^63-1 bytes?
+	tmll	$len,0x0f
+	jnz	.Lxts_dec_proceed
+	aghi	$len,16
+.Lxts_dec_proceed:
+___
+$code.=<<___ if (!$softonly);
+	llgf	%r0,240($key2)
+	lhi	%r1,16
+	clr	%r0,%r1
+	jl	.Lxts_dec_software
+
+	stm${g}	%r6,$s3,6*$SIZE_T($sp)
+	st${g}	$ra,14*$SIZE_T($sp)
+
+	nill	$len,0xfff0		# $len&=~15
+	slgr	$out,$inp
+
+	# generate the tweak value
+	l${g}	$s3,$stdframe($sp)	# pointer to iv
+	la	$s2,$tweak($sp)
+	lmg	$s0,$s1,0($s3)
+	lghi	$s3,16
+	stmg	$s0,$s1,0($s2)
+	la	%r1,0($key2)		# $key2 is not needed past this point
+	.long	0xb92e00aa		# km $s2,$s2, generate the tweak
+	brc	1,.-4			# can this happen?
+
+	l	%r0,240($key1)
+	la	%r1,0($key1)		# $key1 is not needed anymore
+
+	ltgr	$len,$len
+	jz	.Lxts_dec_km_short
+	bras	$ra,_s390x_xts_km
+	jz	.Lxts_dec_km_done
+
+	lrvgr	$s2,$s0			# make copy in reverse byte order
+	lrvgr	$s3,$s1
+	j	.Lxts_dec_km_2ndtweak
+
+.Lxts_dec_km_short:
+	llgc	$len,`2*$SIZE_T-1`($sp)
+	nill	$len,0x0f		# $len%=16
+	lrvg	$s0,$tweak+0($sp)	# load the tweak
+	lrvg	$s1,$tweak+8($sp)
+	lrvgr	$s2,$s0			# make copy in reverse byte order
+	lrvgr	$s3,$s1
+
+.Lxts_dec_km_2ndtweak:
+	lghi	$i1,0x87
+	srag	$i2,$s1,63		# broadcast upper bit
+	ngr	$i1,$i2			# rem
+	srlg	$i2,$s0,63		# carry bit from lower half
+	sllg	$s0,$s0,1
+	sllg	$s1,$s1,1
+	xgr	$s0,$i1
+	ogr	$s1,$i2
+	lrvgr	$i1,$s0			# flip byte order
+	lrvgr	$i2,$s1
+
+	xg	$i1,0($inp)
+	xg	$i2,8($inp)
+	stg	$i1,0($out,$inp)
+	stg	$i2,8($out,$inp)
+	la	$i2,0($out,$inp)
+	lghi	$i3,16
+	.long	0xb92e0066		# km $i2,$i2
+	brc	1,.-4			# can this happen?
+	lrvgr	$i1,$s0
+	lrvgr	$i2,$s1
+	xg	$i1,0($out,$inp)
+	xg	$i2,8($out,$inp)
+	stg	$i1,0($out,$inp)
+	stg	$i2,8($out,$inp)
+
+	la	$i3,0($out,$inp)	# put aside real $out
+.Lxts_dec_km_steal:
+	llgc	$i1,16($inp)
+	llgc	$i2,0($out,$inp)
+	stc	$i1,0($out,$inp)
+	stc	$i2,16($out,$inp)
+	la	$inp,1($inp)
+	brct	$len,.Lxts_dec_km_steal
+
+	lgr	$s0,$s2
+	lgr	$s1,$s3
+	xg	$s0,0($i3)
+	xg	$s1,8($i3)
+	stg	$s0,0($i3)
+	stg	$s1,8($i3)
+	la	$s0,0($i3)
+	lghi	$s1,16
+	.long	0xb92e0088		# km $s0,$s0
+	brc	1,.-4			# can this happen?
+	xg	$s2,0($i3)
+	xg	$s3,8($i3)
+	stg	$s2,0($i3)
+	stg	$s3,8($i3)
+.Lxts_dec_km_done:
+	l${g}	$ra,14*$SIZE_T($sp)
+	st${g}	$sp,$tweak($sp)		# wipe tweak
+	st${g}	$sp,$tweak($sp)
+	lm${g}	%r6,$s3,6*$SIZE_T($sp)
+	br	$ra
+.align	16
+.Lxts_dec_software:
+___
+$code.=<<___;
+	stm${g}	%r6,$ra,6*$SIZE_T($sp)
+
+	srlg	$len,$len,4
+	slgr	$out,$inp
+
+	xgr	$s0,$s0			# clear upper half
+	xgr	$s1,$s1
+	lrv	$s0,$stdframe+4($sp)	# load secno
+	lrv	$s1,$stdframe+0($sp)
+	xgr	$s2,$s2
+	xgr	$s3,$s3
+	stm${g}	%r2,%r5,2*$SIZE_T($sp)
+	la	$key,0($key2)
+	larl	$tbl,AES_Te
+	bras	$ra,_s390x_AES_encrypt	# generate the tweak
+	lm${g}	%r2,%r5,2*$SIZE_T($sp)
+	larl	$tbl,AES_Td
+	lt${g}r	$len,$len
+	stm	$s0,$s3,$tweak($sp)	# save the tweak
+	jz	.Lxts_dec_short
+	j	.Lxts_dec_enter
+
+.align	16
+.Lxts_dec_loop:
+	lrvg	$s1,$tweak+0($sp)	# load the tweak in little-endian
+	lrvg	$s3,$tweak+8($sp)
+	lghi	%r1,0x87
+	srag	%r0,$s3,63		# broadcast upper bit
+	ngr	%r1,%r0			# rem
+	srlg	%r0,$s1,63		# carry bit from lower half
+	sllg	$s1,$s1,1
+	sllg	$s3,$s3,1
+	xgr	$s1,%r1
+	ogr	$s3,%r0
+	lrvgr	$s1,$s1			# flip byte order
+	lrvgr	$s3,$s3
+	srlg	$s0,$s1,32		# smash the tweak to 4x32-bits 
+	stg	$s1,$tweak+0($sp)	# save the tweak
+	llgfr	$s1,$s1
+	srlg	$s2,$s3,32
+	stg	$s3,$tweak+8($sp)
+	llgfr	$s3,$s3
+.Lxts_dec_enter:
+	x	$s0,0($inp)		# tweak^=*(inp)
+	x	$s1,4($inp)
+	x	$s2,8($inp)
+	x	$s3,12($inp)
+	stm${g}	%r2,%r3,2*$SIZE_T($sp)	# only two registers are changing
+	la	$key,0($key1)
+	bras	$ra,_s390x_AES_decrypt
+	lm${g}	%r2,%r5,2*$SIZE_T($sp)
+	x	$s0,$tweak+0($sp)	# ^=tweak
+	x	$s1,$tweak+4($sp)
+	x	$s2,$tweak+8($sp)
+	x	$s3,$tweak+12($sp)
+	st	$s0,0($out,$inp)
+	st	$s1,4($out,$inp)
+	st	$s2,8($out,$inp)
+	st	$s3,12($out,$inp)
+	la	$inp,16($inp)
+	brct${g}	$len,.Lxts_dec_loop
+
+	llgc	$len,`2*$SIZE_T-1`($sp)
+	nill	$len,0x0f		# $len%16
+	jz	.Lxts_dec_done
+
+	# generate pair of tweaks...
+	lrvg	$s1,$tweak+0($sp)	# load the tweak in little-endian
+	lrvg	$s3,$tweak+8($sp)
+	lghi	%r1,0x87
+	srag	%r0,$s3,63		# broadcast upper bit
+	ngr	%r1,%r0			# rem
+	srlg	%r0,$s1,63		# carry bit from lower half
+	sllg	$s1,$s1,1
+	sllg	$s3,$s3,1
+	xgr	$s1,%r1
+	ogr	$s3,%r0
+	lrvgr	$i2,$s1			# flip byte order
+	lrvgr	$i3,$s3
+	stmg	$i2,$i3,$tweak($sp)	# save the 1st tweak
+	j	.Lxts_dec_2ndtweak
+
+.align	16
+.Lxts_dec_short:
+	llgc	$len,`2*$SIZE_T-1`($sp)
+	nill	$len,0x0f		# $len%16
+	lrvg	$s1,$tweak+0($sp)	# load the tweak in little-endian
+	lrvg	$s3,$tweak+8($sp)
+.Lxts_dec_2ndtweak:
+	lghi	%r1,0x87
+	srag	%r0,$s3,63		# broadcast upper bit
+	ngr	%r1,%r0			# rem
+	srlg	%r0,$s1,63		# carry bit from lower half
+	sllg	$s1,$s1,1
+	sllg	$s3,$s3,1
+	xgr	$s1,%r1
+	ogr	$s3,%r0
+	lrvgr	$s1,$s1			# flip byte order
+	lrvgr	$s3,$s3
+	srlg	$s0,$s1,32		# smash the tweak to 4x32-bits
+	stg	$s1,$tweak-16+0($sp)	# save the 2nd tweak
+	llgfr	$s1,$s1
+	srlg	$s2,$s3,32
+	stg	$s3,$tweak-16+8($sp)
+	llgfr	$s3,$s3
+
+	x	$s0,0($inp)		# tweak_the_2nd^=*(inp)
+	x	$s1,4($inp)
+	x	$s2,8($inp)
+	x	$s3,12($inp)
+	stm${g}	%r2,%r3,2*$SIZE_T($sp)
+	la	$key,0($key1)
+	bras	$ra,_s390x_AES_decrypt
+	lm${g}	%r2,%r5,2*$SIZE_T($sp)
+	x	$s0,$tweak-16+0($sp)	# ^=tweak_the_2nd
+	x	$s1,$tweak-16+4($sp)
+	x	$s2,$tweak-16+8($sp)
+	x	$s3,$tweak-16+12($sp)
+	st	$s0,0($out,$inp)
+	st	$s1,4($out,$inp)
+	st	$s2,8($out,$inp)
+	st	$s3,12($out,$inp)
+
+	la	$i3,0($out,$inp)	# put aside real $out
+.Lxts_dec_steal:
+	llgc	%r0,16($inp)
+	llgc	%r1,0($out,$inp)
+	stc	%r0,0($out,$inp)
+	stc	%r1,16($out,$inp)
+	la	$inp,1($inp)
+	brct	$len,.Lxts_dec_steal
+	la	$out,0($i3)		# restore real $out
+
+	lm	$s0,$s3,$tweak($sp)	# load the 1st tweak
+	x	$s0,0($out)		# tweak^=*(inp)|stolen cipher-text
+	x	$s1,4($out)
+	x	$s2,8($out)
+	x	$s3,12($out)
+	st${g}	$out,4*$SIZE_T($sp)
+	la	$key,0($key1)
+	bras	$ra,_s390x_AES_decrypt
+	l${g}	$out,4*$SIZE_T($sp)
+	x	$s0,$tweak+0($sp)	# ^=tweak
+	x	$s1,$tweak+4($sp)
+	x	$s2,$tweak+8($sp)
+	x	$s3,$tweak+12($sp)
+	st	$s0,0($out)
+	st	$s1,4($out)
+	st	$s2,8($out)
+	st	$s3,12($out)
+	stg	$sp,$tweak-16+0($sp)	# wipe 2nd tweak
+	stg	$sp,$tweak-16+8($sp)
+.Lxts_dec_done:
+	stg	$sp,$tweak+0($sp)	# wipe tweak
+	stg	$sp,$twesk+8($sp)
+	lm${g}	%r6,$ra,6*$SIZE_T($sp)
+	br	$ra
+.size	AES_xts_decrypt,.-AES_xts_decrypt
+___
+}
+$code.=<<___;
+.string	"AES for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+.comm	OPENSSL_s390xcap_P,16,8
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;	# force flush
diff --git a/crypto/aes/asm/aes-sparcv9.pl b/crypto/aes/asm/aes-sparcv9.pl
new file mode 100755
index 000000000000..403c4d129048
--- /dev/null
+++ b/crypto/aes/asm/aes-sparcv9.pl
@@ -0,0 +1,1182 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. Rights for redistribution and usage in source and binary
+# forms are granted according to the OpenSSL license.
+# ====================================================================
+#
+# Version 1.1
+#
+# The major reason for undertaken effort was to mitigate the hazard of
+# cache-timing attack. This is [currently and initially!] addressed in
+# two ways. 1. S-boxes are compressed from 5KB to 2KB+256B size each.
+# 2. References to them are scheduled for L2 cache latency, meaning
+# that the tables don't have to reside in L1 cache. Once again, this
+# is an initial draft and one should expect more countermeasures to
+# be implemented...
+#
+# Version 1.1 prefetches T[ed]4 in order to mitigate attack on last
+# round.
+#
+# Even though performance was not the primary goal [on the contrary,
+# extra shifts "induced" by compressed S-box and longer loop epilogue
+# "induced" by scheduling for L2 have negative effect on performance],
+# the code turned out to run in ~23 cycles per processed byte en-/
+# decrypted with 128-bit key. This is pretty good result for code
+# with mentioned qualities and UltraSPARC core. Compared to Sun C
+# generated code my encrypt procedure runs just few percents faster,
+# while decrypt one - whole 50% faster [yes, Sun C failed to generate
+# optimal decrypt procedure]. Compared to GNU C generated code both
+# procedures are more than 60% faster:-)
+
+$bits=32;
+for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+if ($bits==64)	{ $bias=2047; $frame=192; }
+else		{ $bias=0;    $frame=112; }
+$locals=16;
+
+$acc0="%l0";
+$acc1="%o0";
+$acc2="%o1";
+$acc3="%o2";
+
+$acc4="%l1";
+$acc5="%o3";
+$acc6="%o4";
+$acc7="%o5";
+
+$acc8="%l2";
+$acc9="%o7";
+$acc10="%g1";
+$acc11="%g2";
+
+$acc12="%l3";
+$acc13="%g3";
+$acc14="%g4";
+$acc15="%g5";
+
+$t0="%l4";
+$t1="%l5";
+$t2="%l6";
+$t3="%l7";
+
+$s0="%i0";
+$s1="%i1";
+$s2="%i2";
+$s3="%i3";
+$tbl="%i4";
+$key="%i5";
+$rounds="%i7";	# aliases with return address, which is off-loaded to stack
+
+sub _data_word()
+{ my $i;
+    while(defined($i=shift)) { $code.=sprintf"\t.long\t0x%08x,0x%08x\n",$i,$i; }
+}
+
+$code.=<<___ if ($bits==64);
+.register	%g2,#scratch
+.register	%g3,#scratch
+___
+$code.=<<___;
+.section	".text",#alloc,#execinstr
+
+.align	256
+AES_Te:
+___
+&_data_word(
+	0xc66363a5, 0xf87c7c84, 0xee777799, 0xf67b7b8d,
+	0xfff2f20d, 0xd66b6bbd, 0xde6f6fb1, 0x91c5c554,
+	0x60303050, 0x02010103, 0xce6767a9, 0x562b2b7d,
+	0xe7fefe19, 0xb5d7d762, 0x4dababe6, 0xec76769a,
+	0x8fcaca45, 0x1f82829d, 0x89c9c940, 0xfa7d7d87,
+	0xeffafa15, 0xb25959eb, 0x8e4747c9, 0xfbf0f00b,
+	0x41adadec, 0xb3d4d467, 0x5fa2a2fd, 0x45afafea,
+	0x239c9cbf, 0x53a4a4f7, 0xe4727296, 0x9bc0c05b,
+	0x75b7b7c2, 0xe1fdfd1c, 0x3d9393ae, 0x4c26266a,
+	0x6c36365a, 0x7e3f3f41, 0xf5f7f702, 0x83cccc4f,
+	0x6834345c, 0x51a5a5f4, 0xd1e5e534, 0xf9f1f108,
+	0xe2717193, 0xabd8d873, 0x62313153, 0x2a15153f,
+	0x0804040c, 0x95c7c752, 0x46232365, 0x9dc3c35e,
+	0x30181828, 0x379696a1, 0x0a05050f, 0x2f9a9ab5,
+	0x0e070709, 0x24121236, 0x1b80809b, 0xdfe2e23d,
+	0xcdebeb26, 0x4e272769, 0x7fb2b2cd, 0xea75759f,
+	0x1209091b, 0x1d83839e, 0x582c2c74, 0x341a1a2e,
+	0x361b1b2d, 0xdc6e6eb2, 0xb45a5aee, 0x5ba0a0fb,
+	0xa45252f6, 0x763b3b4d, 0xb7d6d661, 0x7db3b3ce,
+	0x5229297b, 0xdde3e33e, 0x5e2f2f71, 0x13848497,
+	0xa65353f5, 0xb9d1d168, 0x00000000, 0xc1eded2c,
+	0x40202060, 0xe3fcfc1f, 0x79b1b1c8, 0xb65b5bed,
+	0xd46a6abe, 0x8dcbcb46, 0x67bebed9, 0x7239394b,
+	0x944a4ade, 0x984c4cd4, 0xb05858e8, 0x85cfcf4a,
+	0xbbd0d06b, 0xc5efef2a, 0x4faaaae5, 0xedfbfb16,
+	0x864343c5, 0x9a4d4dd7, 0x66333355, 0x11858594,
+	0x8a4545cf, 0xe9f9f910, 0x04020206, 0xfe7f7f81,
+	0xa05050f0, 0x783c3c44, 0x259f9fba, 0x4ba8a8e3,
+	0xa25151f3, 0x5da3a3fe, 0x804040c0, 0x058f8f8a,
+	0x3f9292ad, 0x219d9dbc, 0x70383848, 0xf1f5f504,
+	0x63bcbcdf, 0x77b6b6c1, 0xafdada75, 0x42212163,
+	0x20101030, 0xe5ffff1a, 0xfdf3f30e, 0xbfd2d26d,
+	0x81cdcd4c, 0x180c0c14, 0x26131335, 0xc3ecec2f,
+	0xbe5f5fe1, 0x359797a2, 0x884444cc, 0x2e171739,
+	0x93c4c457, 0x55a7a7f2, 0xfc7e7e82, 0x7a3d3d47,
+	0xc86464ac, 0xba5d5de7, 0x3219192b, 0xe6737395,
+	0xc06060a0, 0x19818198, 0x9e4f4fd1, 0xa3dcdc7f,
+	0x44222266, 0x542a2a7e, 0x3b9090ab, 0x0b888883,
+	0x8c4646ca, 0xc7eeee29, 0x6bb8b8d3, 0x2814143c,
+	0xa7dede79, 0xbc5e5ee2, 0x160b0b1d, 0xaddbdb76,
+	0xdbe0e03b, 0x64323256, 0x743a3a4e, 0x140a0a1e,
+	0x924949db, 0x0c06060a, 0x4824246c, 0xb85c5ce4,
+	0x9fc2c25d, 0xbdd3d36e, 0x43acacef, 0xc46262a6,
+	0x399191a8, 0x319595a4, 0xd3e4e437, 0xf279798b,
+	0xd5e7e732, 0x8bc8c843, 0x6e373759, 0xda6d6db7,
+	0x018d8d8c, 0xb1d5d564, 0x9c4e4ed2, 0x49a9a9e0,
+	0xd86c6cb4, 0xac5656fa, 0xf3f4f407, 0xcfeaea25,
+	0xca6565af, 0xf47a7a8e, 0x47aeaee9, 0x10080818,
+	0x6fbabad5, 0xf0787888, 0x4a25256f, 0x5c2e2e72,
+	0x381c1c24, 0x57a6a6f1, 0x73b4b4c7, 0x97c6c651,
+	0xcbe8e823, 0xa1dddd7c, 0xe874749c, 0x3e1f1f21,
+	0x964b4bdd, 0x61bdbddc, 0x0d8b8b86, 0x0f8a8a85,
+	0xe0707090, 0x7c3e3e42, 0x71b5b5c4, 0xcc6666aa,
+	0x904848d8, 0x06030305, 0xf7f6f601, 0x1c0e0e12,
+	0xc26161a3, 0x6a35355f, 0xae5757f9, 0x69b9b9d0,
+	0x17868691, 0x99c1c158, 0x3a1d1d27, 0x279e9eb9,
+	0xd9e1e138, 0xebf8f813, 0x2b9898b3, 0x22111133,
+	0xd26969bb, 0xa9d9d970, 0x078e8e89, 0x339494a7,
+	0x2d9b9bb6, 0x3c1e1e22, 0x15878792, 0xc9e9e920,
+	0x87cece49, 0xaa5555ff, 0x50282878, 0xa5dfdf7a,
+	0x038c8c8f, 0x59a1a1f8, 0x09898980, 0x1a0d0d17,
+	0x65bfbfda, 0xd7e6e631, 0x844242c6, 0xd06868b8,
+	0x824141c3, 0x299999b0, 0x5a2d2d77, 0x1e0f0f11,
+	0x7bb0b0cb, 0xa85454fc, 0x6dbbbbd6, 0x2c16163a);
+$code.=<<___;
+	.byte	0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5
+	.byte	0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76
+	.byte	0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0
+	.byte	0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0
+	.byte	0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc
+	.byte	0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15
+	.byte	0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a
+	.byte	0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75
+	.byte	0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0
+	.byte	0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84
+	.byte	0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b
+	.byte	0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf
+	.byte	0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85
+	.byte	0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8
+	.byte	0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5
+	.byte	0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2
+	.byte	0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17
+	.byte	0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73
+	.byte	0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88
+	.byte	0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb
+	.byte	0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c
+	.byte	0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79
+	.byte	0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9
+	.byte	0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08
+	.byte	0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6
+	.byte	0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a
+	.byte	0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e
+	.byte	0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e
+	.byte	0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94
+	.byte	0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf
+	.byte	0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68
+	.byte	0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16
+.type	AES_Te,#object
+.size	AES_Te,(.-AES_Te)
+
+.align	64
+.skip	16
+_sparcv9_AES_encrypt:
+	save	%sp,-$frame-$locals,%sp
+	stx	%i7,[%sp+$bias+$frame+0]	! off-load return address
+	ld	[$key+240],$rounds
+	ld	[$key+0],$t0
+	ld	[$key+4],$t1			!
+	ld	[$key+8],$t2
+	srl	$rounds,1,$rounds
+	xor	$t0,$s0,$s0
+	ld	[$key+12],$t3
+	srl	$s0,21,$acc0
+	xor	$t1,$s1,$s1
+	ld	[$key+16],$t0
+	srl	$s1,13,$acc1			!
+	xor	$t2,$s2,$s2
+	ld	[$key+20],$t1
+	xor	$t3,$s3,$s3
+	ld	[$key+24],$t2
+	and	$acc0,2040,$acc0
+	ld	[$key+28],$t3
+	nop
+.Lenc_loop:
+	srl	$s2,5,$acc2			!
+	and	$acc1,2040,$acc1
+	ldx	[$tbl+$acc0],$acc0
+	sll	$s3,3,$acc3
+	and	$acc2,2040,$acc2
+	ldx	[$tbl+$acc1],$acc1
+	srl	$s1,21,$acc4
+	and	$acc3,2040,$acc3
+	ldx	[$tbl+$acc2],$acc2		!
+	srl	$s2,13,$acc5
+	and	$acc4,2040,$acc4
+	ldx	[$tbl+$acc3],$acc3
+	srl	$s3,5,$acc6
+	and	$acc5,2040,$acc5
+	ldx	[$tbl+$acc4],$acc4
+	fmovs	%f0,%f0
+	sll	$s0,3,$acc7			!
+	and	$acc6,2040,$acc6
+	ldx	[$tbl+$acc5],$acc5
+	srl	$s2,21,$acc8
+	and	$acc7,2040,$acc7
+	ldx	[$tbl+$acc6],$acc6
+	srl	$s3,13,$acc9
+	and	$acc8,2040,$acc8
+	ldx	[$tbl+$acc7],$acc7		!
+	srl	$s0,5,$acc10
+	and	$acc9,2040,$acc9
+	ldx	[$tbl+$acc8],$acc8
+	sll	$s1,3,$acc11
+	and	$acc10,2040,$acc10
+	ldx	[$tbl+$acc9],$acc9
+	fmovs	%f0,%f0
+	srl	$s3,21,$acc12			!
+	and	$acc11,2040,$acc11
+	ldx	[$tbl+$acc10],$acc10
+	srl	$s0,13,$acc13
+	and	$acc12,2040,$acc12
+	ldx	[$tbl+$acc11],$acc11
+	srl	$s1,5,$acc14
+	and	$acc13,2040,$acc13
+	ldx	[$tbl+$acc12],$acc12		!
+	sll	$s2,3,$acc15
+	and	$acc14,2040,$acc14
+	ldx	[$tbl+$acc13],$acc13
+	and	$acc15,2040,$acc15
+	add	$key,32,$key
+	ldx	[$tbl+$acc14],$acc14
+	fmovs	%f0,%f0
+	subcc	$rounds,1,$rounds		!
+	ldx	[$tbl+$acc15],$acc15
+	bz,a,pn	%icc,.Lenc_last
+	add	$tbl,2048,$rounds
+
+		srlx	$acc1,8,$acc1
+		xor	$acc0,$t0,$t0
+	ld	[$key+0],$s0
+	fmovs	%f0,%f0
+		srlx	$acc2,16,$acc2		!
+		xor	$acc1,$t0,$t0
+	ld	[$key+4],$s1
+		srlx	$acc3,24,$acc3
+		xor	$acc2,$t0,$t0
+	ld	[$key+8],$s2
+		srlx	$acc5,8,$acc5
+		xor	$acc3,$t0,$t0
+	ld	[$key+12],$s3			!
+		srlx	$acc6,16,$acc6
+		xor	$acc4,$t1,$t1
+	fmovs	%f0,%f0
+		srlx	$acc7,24,$acc7
+		xor	$acc5,$t1,$t1
+		srlx	$acc9,8,$acc9
+		xor	$acc6,$t1,$t1
+		srlx	$acc10,16,$acc10	!
+		xor	$acc7,$t1,$t1
+		srlx	$acc11,24,$acc11
+		xor	$acc8,$t2,$t2
+		srlx	$acc13,8,$acc13
+		xor	$acc9,$t2,$t2
+		srlx	$acc14,16,$acc14
+		xor	$acc10,$t2,$t2
+		srlx	$acc15,24,$acc15	!
+		xor	$acc11,$t2,$t2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$t3,$t3
+	srl	$t0,21,$acc0
+		xor	$acc14,$t3,$t3
+	srl	$t1,13,$acc1
+		xor	$acc15,$t3,$t3
+
+	and	$acc0,2040,$acc0		!
+	srl	$t2,5,$acc2
+	and	$acc1,2040,$acc1
+	ldx	[$tbl+$acc0],$acc0
+	sll	$t3,3,$acc3
+	and	$acc2,2040,$acc2
+	ldx	[$tbl+$acc1],$acc1
+	fmovs	%f0,%f0
+	srl	$t1,21,$acc4			!
+	and	$acc3,2040,$acc3
+	ldx	[$tbl+$acc2],$acc2
+	srl	$t2,13,$acc5
+	and	$acc4,2040,$acc4
+	ldx	[$tbl+$acc3],$acc3
+	srl	$t3,5,$acc6
+	and	$acc5,2040,$acc5
+	ldx	[$tbl+$acc4],$acc4		!
+	sll	$t0,3,$acc7
+	and	$acc6,2040,$acc6
+	ldx	[$tbl+$acc5],$acc5
+	srl	$t2,21,$acc8
+	and	$acc7,2040,$acc7
+	ldx	[$tbl+$acc6],$acc6
+	fmovs	%f0,%f0
+	srl	$t3,13,$acc9			!
+	and	$acc8,2040,$acc8
+	ldx	[$tbl+$acc7],$acc7
+	srl	$t0,5,$acc10
+	and	$acc9,2040,$acc9
+	ldx	[$tbl+$acc8],$acc8
+	sll	$t1,3,$acc11
+	and	$acc10,2040,$acc10
+	ldx	[$tbl+$acc9],$acc9		!
+	srl	$t3,21,$acc12
+	and	$acc11,2040,$acc11
+	ldx	[$tbl+$acc10],$acc10
+	srl	$t0,13,$acc13
+	and	$acc12,2040,$acc12
+	ldx	[$tbl+$acc11],$acc11
+	fmovs	%f0,%f0
+	srl	$t1,5,$acc14			!
+	and	$acc13,2040,$acc13
+	ldx	[$tbl+$acc12],$acc12
+	sll	$t2,3,$acc15
+	and	$acc14,2040,$acc14
+	ldx	[$tbl+$acc13],$acc13
+		srlx	$acc1,8,$acc1
+	and	$acc15,2040,$acc15
+	ldx	[$tbl+$acc14],$acc14		!
+
+		srlx	$acc2,16,$acc2
+		xor	$acc0,$s0,$s0
+	ldx	[$tbl+$acc15],$acc15
+		srlx	$acc3,24,$acc3
+		xor	$acc1,$s0,$s0
+	ld	[$key+16],$t0
+	fmovs	%f0,%f0
+		srlx	$acc5,8,$acc5		!
+		xor	$acc2,$s0,$s0
+	ld	[$key+20],$t1
+		srlx	$acc6,16,$acc6
+		xor	$acc3,$s0,$s0
+	ld	[$key+24],$t2
+		srlx	$acc7,24,$acc7
+		xor	$acc4,$s1,$s1
+	ld	[$key+28],$t3			!
+		srlx	$acc9,8,$acc9
+		xor	$acc5,$s1,$s1
+	ldx	[$tbl+2048+0],%g0		! prefetch te4
+		srlx	$acc10,16,$acc10
+		xor	$acc6,$s1,$s1
+	ldx	[$tbl+2048+32],%g0		! prefetch te4
+		srlx	$acc11,24,$acc11
+		xor	$acc7,$s1,$s1
+	ldx	[$tbl+2048+64],%g0		! prefetch te4
+		srlx	$acc13,8,$acc13
+		xor	$acc8,$s2,$s2
+	ldx	[$tbl+2048+96],%g0		! prefetch te4
+		srlx	$acc14,16,$acc14	!
+		xor	$acc9,$s2,$s2
+	ldx	[$tbl+2048+128],%g0		! prefetch te4
+		srlx	$acc15,24,$acc15
+		xor	$acc10,$s2,$s2
+	ldx	[$tbl+2048+160],%g0		! prefetch te4
+	srl	$s0,21,$acc0
+		xor	$acc11,$s2,$s2
+	ldx	[$tbl+2048+192],%g0		! prefetch te4
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$s3,$s3
+	ldx	[$tbl+2048+224],%g0		! prefetch te4
+	srl	$s1,13,$acc1			!
+		xor	$acc14,$s3,$s3
+		xor	$acc15,$s3,$s3
+	ba	.Lenc_loop
+	and	$acc0,2040,$acc0
+
+.align	32
+.Lenc_last:
+		srlx	$acc1,8,$acc1		!
+		xor	$acc0,$t0,$t0
+	ld	[$key+0],$s0
+		srlx	$acc2,16,$acc2
+		xor	$acc1,$t0,$t0
+	ld	[$key+4],$s1
+		srlx	$acc3,24,$acc3
+		xor	$acc2,$t0,$t0
+	ld	[$key+8],$s2			!
+		srlx	$acc5,8,$acc5
+		xor	$acc3,$t0,$t0
+	ld	[$key+12],$s3
+		srlx	$acc6,16,$acc6
+		xor	$acc4,$t1,$t1
+		srlx	$acc7,24,$acc7
+		xor	$acc5,$t1,$t1
+		srlx	$acc9,8,$acc9		!
+		xor	$acc6,$t1,$t1
+		srlx	$acc10,16,$acc10
+		xor	$acc7,$t1,$t1
+		srlx	$acc11,24,$acc11
+		xor	$acc8,$t2,$t2
+		srlx	$acc13,8,$acc13
+		xor	$acc9,$t2,$t2
+		srlx	$acc14,16,$acc14	!
+		xor	$acc10,$t2,$t2
+		srlx	$acc15,24,$acc15
+		xor	$acc11,$t2,$t2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$t3,$t3
+	srl	$t0,24,$acc0
+		xor	$acc14,$t3,$t3
+	srl	$t1,16,$acc1			!
+		xor	$acc15,$t3,$t3
+
+	srl	$t2,8,$acc2
+	and	$acc1,255,$acc1
+	ldub	[$rounds+$acc0],$acc0
+	srl	$t1,24,$acc4
+	and	$acc2,255,$acc2
+	ldub	[$rounds+$acc1],$acc1
+	srl	$t2,16,$acc5			!
+	and	$t3,255,$acc3
+	ldub	[$rounds+$acc2],$acc2
+	ldub	[$rounds+$acc3],$acc3
+	srl	$t3,8,$acc6
+	and	$acc5,255,$acc5
+	ldub	[$rounds+$acc4],$acc4
+	fmovs	%f0,%f0
+	srl	$t2,24,$acc8			!
+	and	$acc6,255,$acc6
+	ldub	[$rounds+$acc5],$acc5
+	srl	$t3,16,$acc9
+	and	$t0,255,$acc7
+	ldub	[$rounds+$acc6],$acc6
+	ldub	[$rounds+$acc7],$acc7
+	fmovs	%f0,%f0
+	srl	$t0,8,$acc10			!
+	and	$acc9,255,$acc9
+	ldub	[$rounds+$acc8],$acc8
+	srl	$t3,24,$acc12
+	and	$acc10,255,$acc10
+	ldub	[$rounds+$acc9],$acc9
+	srl	$t0,16,$acc13
+	and	$t1,255,$acc11
+	ldub	[$rounds+$acc10],$acc10		!
+	srl	$t1,8,$acc14
+	and	$acc13,255,$acc13
+	ldub	[$rounds+$acc11],$acc11
+	ldub	[$rounds+$acc12],$acc12
+	and	$acc14,255,$acc14
+	ldub	[$rounds+$acc13],$acc13
+	and	$t2,255,$acc15
+	ldub	[$rounds+$acc14],$acc14		!
+
+		sll	$acc0,24,$acc0
+		xor	$acc3,$s0,$s0
+	ldub	[$rounds+$acc15],$acc15
+		sll	$acc1,16,$acc1
+		xor	$acc0,$s0,$s0
+	ldx	[%sp+$bias+$frame+0],%i7	! restore return address
+	fmovs	%f0,%f0
+		sll	$acc2,8,$acc2		!
+		xor	$acc1,$s0,$s0
+		sll	$acc4,24,$acc4
+		xor	$acc2,$s0,$s0
+		sll	$acc5,16,$acc5
+		xor	$acc7,$s1,$s1
+		sll	$acc6,8,$acc6
+		xor	$acc4,$s1,$s1
+		sll	$acc8,24,$acc8		!
+		xor	$acc5,$s1,$s1
+		sll	$acc9,16,$acc9
+		xor	$acc11,$s2,$s2
+		sll	$acc10,8,$acc10
+		xor	$acc6,$s1,$s1
+		sll	$acc12,24,$acc12
+		xor	$acc8,$s2,$s2
+		sll	$acc13,16,$acc13	!
+		xor	$acc9,$s2,$s2
+		sll	$acc14,8,$acc14
+		xor	$acc10,$s2,$s2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$s3,$s3
+		xor	$acc14,$s3,$s3
+		xor	$acc15,$s3,$s3
+
+	ret
+	restore
+.type	_sparcv9_AES_encrypt,#function
+.size	_sparcv9_AES_encrypt,(.-_sparcv9_AES_encrypt)
+
+.align	32
+.globl	AES_encrypt
+AES_encrypt:
+	or	%o0,%o1,%g1
+	andcc	%g1,3,%g0
+	bnz,pn	%xcc,.Lunaligned_enc
+	save	%sp,-$frame,%sp
+
+	ld	[%i0+0],%o0
+	ld	[%i0+4],%o1
+	ld	[%i0+8],%o2
+	ld	[%i0+12],%o3
+
+1:	call	.+8
+	add	%o7,AES_Te-1b,%o4
+	call	_sparcv9_AES_encrypt
+	mov	%i2,%o5
+
+	st	%o0,[%i1+0]
+	st	%o1,[%i1+4]
+	st	%o2,[%i1+8]
+	st	%o3,[%i1+12]
+
+	ret
+	restore
+
+.align	32
+.Lunaligned_enc:
+	ldub	[%i0+0],%l0
+	ldub	[%i0+1],%l1
+	ldub	[%i0+2],%l2
+
+	sll	%l0,24,%l0
+	ldub	[%i0+3],%l3
+	sll	%l1,16,%l1
+	ldub	[%i0+4],%l4
+	sll	%l2,8,%l2
+	or	%l1,%l0,%l0
+	ldub	[%i0+5],%l5
+	sll	%l4,24,%l4
+	or	%l3,%l2,%l2
+	ldub	[%i0+6],%l6
+	sll	%l5,16,%l5
+	or	%l0,%l2,%o0
+	ldub	[%i0+7],%l7
+
+	sll	%l6,8,%l6
+	or	%l5,%l4,%l4
+	ldub	[%i0+8],%l0
+	or	%l7,%l6,%l6
+	ldub	[%i0+9],%l1
+	or	%l4,%l6,%o1
+	ldub	[%i0+10],%l2
+
+	sll	%l0,24,%l0
+	ldub	[%i0+11],%l3
+	sll	%l1,16,%l1
+	ldub	[%i0+12],%l4
+	sll	%l2,8,%l2
+	or	%l1,%l0,%l0
+	ldub	[%i0+13],%l5
+	sll	%l4,24,%l4
+	or	%l3,%l2,%l2
+	ldub	[%i0+14],%l6
+	sll	%l5,16,%l5
+	or	%l0,%l2,%o2
+	ldub	[%i0+15],%l7
+
+	sll	%l6,8,%l6
+	or	%l5,%l4,%l4
+	or	%l7,%l6,%l6
+	or	%l4,%l6,%o3
+
+1:	call	.+8
+	add	%o7,AES_Te-1b,%o4
+	call	_sparcv9_AES_encrypt
+	mov	%i2,%o5
+
+	srl	%o0,24,%l0
+	srl	%o0,16,%l1
+	stb	%l0,[%i1+0]
+	srl	%o0,8,%l2
+	stb	%l1,[%i1+1]
+	stb	%l2,[%i1+2]
+	srl	%o1,24,%l4
+	stb	%o0,[%i1+3]
+
+	srl	%o1,16,%l5
+	stb	%l4,[%i1+4]
+	srl	%o1,8,%l6
+	stb	%l5,[%i1+5]
+	stb	%l6,[%i1+6]
+	srl	%o2,24,%l0
+	stb	%o1,[%i1+7]
+
+	srl	%o2,16,%l1
+	stb	%l0,[%i1+8]
+	srl	%o2,8,%l2
+	stb	%l1,[%i1+9]
+	stb	%l2,[%i1+10]
+	srl	%o3,24,%l4
+	stb	%o2,[%i1+11]
+
+	srl	%o3,16,%l5
+	stb	%l4,[%i1+12]
+	srl	%o3,8,%l6
+	stb	%l5,[%i1+13]
+	stb	%l6,[%i1+14]
+	stb	%o3,[%i1+15]
+
+	ret
+	restore
+.type	AES_encrypt,#function
+.size	AES_encrypt,(.-AES_encrypt)
+
+___
+
+$code.=<<___;
+.align	256
+AES_Td:
+___
+&_data_word(
+	0x51f4a750, 0x7e416553, 0x1a17a4c3, 0x3a275e96,
+	0x3bab6bcb, 0x1f9d45f1, 0xacfa58ab, 0x4be30393,
+	0x2030fa55, 0xad766df6, 0x88cc7691, 0xf5024c25,
+	0x4fe5d7fc, 0xc52acbd7, 0x26354480, 0xb562a38f,
+	0xdeb15a49, 0x25ba1b67, 0x45ea0e98, 0x5dfec0e1,
+	0xc32f7502, 0x814cf012, 0x8d4697a3, 0x6bd3f9c6,
+	0x038f5fe7, 0x15929c95, 0xbf6d7aeb, 0x955259da,
+	0xd4be832d, 0x587421d3, 0x49e06929, 0x8ec9c844,
+	0x75c2896a, 0xf48e7978, 0x99583e6b, 0x27b971dd,
+	0xbee14fb6, 0xf088ad17, 0xc920ac66, 0x7dce3ab4,
+	0x63df4a18, 0xe51a3182, 0x97513360, 0x62537f45,
+	0xb16477e0, 0xbb6bae84, 0xfe81a01c, 0xf9082b94,
+	0x70486858, 0x8f45fd19, 0x94de6c87, 0x527bf8b7,
+	0xab73d323, 0x724b02e2, 0xe31f8f57, 0x6655ab2a,
+	0xb2eb2807, 0x2fb5c203, 0x86c57b9a, 0xd33708a5,
+	0x302887f2, 0x23bfa5b2, 0x02036aba, 0xed16825c,
+	0x8acf1c2b, 0xa779b492, 0xf307f2f0, 0x4e69e2a1,
+	0x65daf4cd, 0x0605bed5, 0xd134621f, 0xc4a6fe8a,
+	0x342e539d, 0xa2f355a0, 0x058ae132, 0xa4f6eb75,
+	0x0b83ec39, 0x4060efaa, 0x5e719f06, 0xbd6e1051,
+	0x3e218af9, 0x96dd063d, 0xdd3e05ae, 0x4de6bd46,
+	0x91548db5, 0x71c45d05, 0x0406d46f, 0x605015ff,
+	0x1998fb24, 0xd6bde997, 0x894043cc, 0x67d99e77,
+	0xb0e842bd, 0x07898b88, 0xe7195b38, 0x79c8eedb,
+	0xa17c0a47, 0x7c420fe9, 0xf8841ec9, 0x00000000,
+	0x09808683, 0x322bed48, 0x1e1170ac, 0x6c5a724e,
+	0xfd0efffb, 0x0f853856, 0x3daed51e, 0x362d3927,
+	0x0a0fd964, 0x685ca621, 0x9b5b54d1, 0x24362e3a,
+	0x0c0a67b1, 0x9357e70f, 0xb4ee96d2, 0x1b9b919e,
+	0x80c0c54f, 0x61dc20a2, 0x5a774b69, 0x1c121a16,
+	0xe293ba0a, 0xc0a02ae5, 0x3c22e043, 0x121b171d,
+	0x0e090d0b, 0xf28bc7ad, 0x2db6a8b9, 0x141ea9c8,
+	0x57f11985, 0xaf75074c, 0xee99ddbb, 0xa37f60fd,
+	0xf701269f, 0x5c72f5bc, 0x44663bc5, 0x5bfb7e34,
+	0x8b432976, 0xcb23c6dc, 0xb6edfc68, 0xb8e4f163,
+	0xd731dcca, 0x42638510, 0x13972240, 0x84c61120,
+	0x854a247d, 0xd2bb3df8, 0xaef93211, 0xc729a16d,
+	0x1d9e2f4b, 0xdcb230f3, 0x0d8652ec, 0x77c1e3d0,
+	0x2bb3166c, 0xa970b999, 0x119448fa, 0x47e96422,
+	0xa8fc8cc4, 0xa0f03f1a, 0x567d2cd8, 0x223390ef,
+	0x87494ec7, 0xd938d1c1, 0x8ccaa2fe, 0x98d40b36,
+	0xa6f581cf, 0xa57ade28, 0xdab78e26, 0x3fadbfa4,
+	0x2c3a9de4, 0x5078920d, 0x6a5fcc9b, 0x547e4662,
+	0xf68d13c2, 0x90d8b8e8, 0x2e39f75e, 0x82c3aff5,
+	0x9f5d80be, 0x69d0937c, 0x6fd52da9, 0xcf2512b3,
+	0xc8ac993b, 0x10187da7, 0xe89c636e, 0xdb3bbb7b,
+	0xcd267809, 0x6e5918f4, 0xec9ab701, 0x834f9aa8,
+	0xe6956e65, 0xaaffe67e, 0x21bccf08, 0xef15e8e6,
+	0xbae79bd9, 0x4a6f36ce, 0xea9f09d4, 0x29b07cd6,
+	0x31a4b2af, 0x2a3f2331, 0xc6a59430, 0x35a266c0,
+	0x744ebc37, 0xfc82caa6, 0xe090d0b0, 0x33a7d815,
+	0xf104984a, 0x41ecdaf7, 0x7fcd500e, 0x1791f62f,
+	0x764dd68d, 0x43efb04d, 0xccaa4d54, 0xe49604df,
+	0x9ed1b5e3, 0x4c6a881b, 0xc12c1fb8, 0x4665517f,
+	0x9d5eea04, 0x018c355d, 0xfa877473, 0xfb0b412e,
+	0xb3671d5a, 0x92dbd252, 0xe9105633, 0x6dd64713,
+	0x9ad7618c, 0x37a10c7a, 0x59f8148e, 0xeb133c89,
+	0xcea927ee, 0xb761c935, 0xe11ce5ed, 0x7a47b13c,
+	0x9cd2df59, 0x55f2733f, 0x1814ce79, 0x73c737bf,
+	0x53f7cdea, 0x5ffdaa5b, 0xdf3d6f14, 0x7844db86,
+	0xcaaff381, 0xb968c43e, 0x3824342c, 0xc2a3405f,
+	0x161dc372, 0xbce2250c, 0x283c498b, 0xff0d9541,
+	0x39a80171, 0x080cb3de, 0xd8b4e49c, 0x6456c190,
+	0x7bcb8461, 0xd532b670, 0x486c5c74, 0xd0b85742);
+$code.=<<___;
+	.byte	0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38
+	.byte	0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb
+	.byte	0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87
+	.byte	0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb
+	.byte	0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d
+	.byte	0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e
+	.byte	0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2
+	.byte	0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25
+	.byte	0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16
+	.byte	0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92
+	.byte	0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda
+	.byte	0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84
+	.byte	0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a
+	.byte	0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06
+	.byte	0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02
+	.byte	0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b
+	.byte	0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea
+	.byte	0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73
+	.byte	0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85
+	.byte	0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e
+	.byte	0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89
+	.byte	0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b
+	.byte	0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20
+	.byte	0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4
+	.byte	0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31
+	.byte	0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f
+	.byte	0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d
+	.byte	0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef
+	.byte	0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0
+	.byte	0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61
+	.byte	0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26
+	.byte	0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
+.type	AES_Td,#object
+.size	AES_Td,(.-AES_Td)
+
+.align	64
+.skip	16
+_sparcv9_AES_decrypt:
+	save	%sp,-$frame-$locals,%sp
+	stx	%i7,[%sp+$bias+$frame+0]	! off-load return address
+	ld	[$key+240],$rounds
+	ld	[$key+0],$t0
+	ld	[$key+4],$t1			!
+	ld	[$key+8],$t2
+	ld	[$key+12],$t3
+	srl	$rounds,1,$rounds
+	xor	$t0,$s0,$s0
+	ld	[$key+16],$t0
+	xor	$t1,$s1,$s1
+	ld	[$key+20],$t1
+	srl	$s0,21,$acc0			!
+	xor	$t2,$s2,$s2
+	ld	[$key+24],$t2
+	xor	$t3,$s3,$s3
+	and	$acc0,2040,$acc0
+	ld	[$key+28],$t3
+	srl	$s3,13,$acc1
+	nop
+.Ldec_loop:
+	srl	$s2,5,$acc2			!
+	and	$acc1,2040,$acc1
+	ldx	[$tbl+$acc0],$acc0
+	sll	$s1,3,$acc3
+	and	$acc2,2040,$acc2
+	ldx	[$tbl+$acc1],$acc1
+	srl	$s1,21,$acc4
+	and	$acc3,2040,$acc3
+	ldx	[$tbl+$acc2],$acc2		!
+	srl	$s0,13,$acc5
+	and	$acc4,2040,$acc4
+	ldx	[$tbl+$acc3],$acc3
+	srl	$s3,5,$acc6
+	and	$acc5,2040,$acc5
+	ldx	[$tbl+$acc4],$acc4
+	fmovs	%f0,%f0
+	sll	$s2,3,$acc7			!
+	and	$acc6,2040,$acc6
+	ldx	[$tbl+$acc5],$acc5
+	srl	$s2,21,$acc8
+	and	$acc7,2040,$acc7
+	ldx	[$tbl+$acc6],$acc6
+	srl	$s1,13,$acc9
+	and	$acc8,2040,$acc8
+	ldx	[$tbl+$acc7],$acc7		!
+	srl	$s0,5,$acc10
+	and	$acc9,2040,$acc9
+	ldx	[$tbl+$acc8],$acc8
+	sll	$s3,3,$acc11
+	and	$acc10,2040,$acc10
+	ldx	[$tbl+$acc9],$acc9
+	fmovs	%f0,%f0
+	srl	$s3,21,$acc12			!
+	and	$acc11,2040,$acc11
+	ldx	[$tbl+$acc10],$acc10
+	srl	$s2,13,$acc13
+	and	$acc12,2040,$acc12
+	ldx	[$tbl+$acc11],$acc11
+	srl	$s1,5,$acc14
+	and	$acc13,2040,$acc13
+	ldx	[$tbl+$acc12],$acc12		!
+	sll	$s0,3,$acc15
+	and	$acc14,2040,$acc14
+	ldx	[$tbl+$acc13],$acc13
+	and	$acc15,2040,$acc15
+	add	$key,32,$key
+	ldx	[$tbl+$acc14],$acc14
+	fmovs	%f0,%f0
+	subcc	$rounds,1,$rounds		!
+	ldx	[$tbl+$acc15],$acc15
+	bz,a,pn	%icc,.Ldec_last
+	add	$tbl,2048,$rounds
+
+		srlx	$acc1,8,$acc1
+		xor	$acc0,$t0,$t0
+	ld	[$key+0],$s0
+	fmovs	%f0,%f0
+		srlx	$acc2,16,$acc2		!
+		xor	$acc1,$t0,$t0
+	ld	[$key+4],$s1
+		srlx	$acc3,24,$acc3
+		xor	$acc2,$t0,$t0
+	ld	[$key+8],$s2
+		srlx	$acc5,8,$acc5
+		xor	$acc3,$t0,$t0
+	ld	[$key+12],$s3			!
+		srlx	$acc6,16,$acc6
+		xor	$acc4,$t1,$t1
+	fmovs	%f0,%f0
+		srlx	$acc7,24,$acc7
+		xor	$acc5,$t1,$t1
+		srlx	$acc9,8,$acc9
+		xor	$acc6,$t1,$t1
+		srlx	$acc10,16,$acc10	!
+		xor	$acc7,$t1,$t1
+		srlx	$acc11,24,$acc11
+		xor	$acc8,$t2,$t2
+		srlx	$acc13,8,$acc13
+		xor	$acc9,$t2,$t2
+		srlx	$acc14,16,$acc14
+		xor	$acc10,$t2,$t2
+		srlx	$acc15,24,$acc15	!
+		xor	$acc11,$t2,$t2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$t3,$t3
+	srl	$t0,21,$acc0
+		xor	$acc14,$t3,$t3
+		xor	$acc15,$t3,$t3
+	srl	$t3,13,$acc1
+
+	and	$acc0,2040,$acc0		!
+	srl	$t2,5,$acc2
+	and	$acc1,2040,$acc1
+	ldx	[$tbl+$acc0],$acc0
+	sll	$t1,3,$acc3
+	and	$acc2,2040,$acc2
+	ldx	[$tbl+$acc1],$acc1
+	fmovs	%f0,%f0
+	srl	$t1,21,$acc4			!
+	and	$acc3,2040,$acc3
+	ldx	[$tbl+$acc2],$acc2
+	srl	$t0,13,$acc5
+	and	$acc4,2040,$acc4
+	ldx	[$tbl+$acc3],$acc3
+	srl	$t3,5,$acc6
+	and	$acc5,2040,$acc5
+	ldx	[$tbl+$acc4],$acc4		!
+	sll	$t2,3,$acc7
+	and	$acc6,2040,$acc6
+	ldx	[$tbl+$acc5],$acc5
+	srl	$t2,21,$acc8
+	and	$acc7,2040,$acc7
+	ldx	[$tbl+$acc6],$acc6
+	fmovs	%f0,%f0
+	srl	$t1,13,$acc9			!
+	and	$acc8,2040,$acc8
+	ldx	[$tbl+$acc7],$acc7
+	srl	$t0,5,$acc10
+	and	$acc9,2040,$acc9
+	ldx	[$tbl+$acc8],$acc8
+	sll	$t3,3,$acc11
+	and	$acc10,2040,$acc10
+	ldx	[$tbl+$acc9],$acc9		!
+	srl	$t3,21,$acc12
+	and	$acc11,2040,$acc11
+	ldx	[$tbl+$acc10],$acc10
+	srl	$t2,13,$acc13
+	and	$acc12,2040,$acc12
+	ldx	[$tbl+$acc11],$acc11
+	fmovs	%f0,%f0
+	srl	$t1,5,$acc14			!
+	and	$acc13,2040,$acc13
+	ldx	[$tbl+$acc12],$acc12
+	sll	$t0,3,$acc15
+	and	$acc14,2040,$acc14
+	ldx	[$tbl+$acc13],$acc13
+		srlx	$acc1,8,$acc1
+	and	$acc15,2040,$acc15
+	ldx	[$tbl+$acc14],$acc14		!
+
+		srlx	$acc2,16,$acc2
+		xor	$acc0,$s0,$s0
+	ldx	[$tbl+$acc15],$acc15
+		srlx	$acc3,24,$acc3
+		xor	$acc1,$s0,$s0
+	ld	[$key+16],$t0
+	fmovs	%f0,%f0
+		srlx	$acc5,8,$acc5		!
+		xor	$acc2,$s0,$s0
+	ld	[$key+20],$t1
+		srlx	$acc6,16,$acc6
+		xor	$acc3,$s0,$s0
+	ld	[$key+24],$t2
+		srlx	$acc7,24,$acc7
+		xor	$acc4,$s1,$s1
+	ld	[$key+28],$t3			!
+		srlx	$acc9,8,$acc9
+		xor	$acc5,$s1,$s1
+	ldx	[$tbl+2048+0],%g0		! prefetch td4
+		srlx	$acc10,16,$acc10
+		xor	$acc6,$s1,$s1
+	ldx	[$tbl+2048+32],%g0		! prefetch td4
+		srlx	$acc11,24,$acc11
+		xor	$acc7,$s1,$s1
+	ldx	[$tbl+2048+64],%g0		! prefetch td4
+		srlx	$acc13,8,$acc13
+		xor	$acc8,$s2,$s2
+	ldx	[$tbl+2048+96],%g0		! prefetch td4
+		srlx	$acc14,16,$acc14	!
+		xor	$acc9,$s2,$s2
+	ldx	[$tbl+2048+128],%g0		! prefetch td4
+		srlx	$acc15,24,$acc15
+		xor	$acc10,$s2,$s2
+	ldx	[$tbl+2048+160],%g0		! prefetch td4
+	srl	$s0,21,$acc0
+		xor	$acc11,$s2,$s2
+	ldx	[$tbl+2048+192],%g0		! prefetch td4
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$s3,$s3
+	ldx	[$tbl+2048+224],%g0		! prefetch td4
+	and	$acc0,2040,$acc0		!
+		xor	$acc14,$s3,$s3
+		xor	$acc15,$s3,$s3
+	ba	.Ldec_loop
+	srl	$s3,13,$acc1
+
+.align	32
+.Ldec_last:
+		srlx	$acc1,8,$acc1		!
+		xor	$acc0,$t0,$t0
+	ld	[$key+0],$s0
+		srlx	$acc2,16,$acc2
+		xor	$acc1,$t0,$t0
+	ld	[$key+4],$s1
+		srlx	$acc3,24,$acc3
+		xor	$acc2,$t0,$t0
+	ld	[$key+8],$s2			!
+		srlx	$acc5,8,$acc5
+		xor	$acc3,$t0,$t0
+	ld	[$key+12],$s3
+		srlx	$acc6,16,$acc6
+		xor	$acc4,$t1,$t1
+		srlx	$acc7,24,$acc7
+		xor	$acc5,$t1,$t1
+		srlx	$acc9,8,$acc9		!
+		xor	$acc6,$t1,$t1
+		srlx	$acc10,16,$acc10
+		xor	$acc7,$t1,$t1
+		srlx	$acc11,24,$acc11
+		xor	$acc8,$t2,$t2
+		srlx	$acc13,8,$acc13
+		xor	$acc9,$t2,$t2
+		srlx	$acc14,16,$acc14	!
+		xor	$acc10,$t2,$t2
+		srlx	$acc15,24,$acc15
+		xor	$acc11,$t2,$t2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$t3,$t3
+	srl	$t0,24,$acc0
+		xor	$acc14,$t3,$t3
+		xor	$acc15,$t3,$t3		!
+	srl	$t3,16,$acc1
+
+	srl	$t2,8,$acc2
+	and	$acc1,255,$acc1
+	ldub	[$rounds+$acc0],$acc0
+	srl	$t1,24,$acc4
+	and	$acc2,255,$acc2
+	ldub	[$rounds+$acc1],$acc1
+	srl	$t0,16,$acc5			!
+	and	$t1,255,$acc3
+	ldub	[$rounds+$acc2],$acc2
+	ldub	[$rounds+$acc3],$acc3
+	srl	$t3,8,$acc6
+	and	$acc5,255,$acc5
+	ldub	[$rounds+$acc4],$acc4
+	fmovs	%f0,%f0
+	srl	$t2,24,$acc8			!
+	and	$acc6,255,$acc6
+	ldub	[$rounds+$acc5],$acc5
+	srl	$t1,16,$acc9
+	and	$t2,255,$acc7
+	ldub	[$rounds+$acc6],$acc6
+	ldub	[$rounds+$acc7],$acc7
+	fmovs	%f0,%f0
+	srl	$t0,8,$acc10			!
+	and	$acc9,255,$acc9
+	ldub	[$rounds+$acc8],$acc8
+	srl	$t3,24,$acc12
+	and	$acc10,255,$acc10
+	ldub	[$rounds+$acc9],$acc9
+	srl	$t2,16,$acc13
+	and	$t3,255,$acc11
+	ldub	[$rounds+$acc10],$acc10		!
+	srl	$t1,8,$acc14
+	and	$acc13,255,$acc13
+	ldub	[$rounds+$acc11],$acc11
+	ldub	[$rounds+$acc12],$acc12
+	and	$acc14,255,$acc14
+	ldub	[$rounds+$acc13],$acc13
+	and	$t0,255,$acc15
+	ldub	[$rounds+$acc14],$acc14		!
+
+		sll	$acc0,24,$acc0
+		xor	$acc3,$s0,$s0
+	ldub	[$rounds+$acc15],$acc15
+		sll	$acc1,16,$acc1
+		xor	$acc0,$s0,$s0
+	ldx	[%sp+$bias+$frame+0],%i7	! restore return address
+	fmovs	%f0,%f0
+		sll	$acc2,8,$acc2		!
+		xor	$acc1,$s0,$s0
+		sll	$acc4,24,$acc4
+		xor	$acc2,$s0,$s0
+		sll	$acc5,16,$acc5
+		xor	$acc7,$s1,$s1
+		sll	$acc6,8,$acc6
+		xor	$acc4,$s1,$s1
+		sll	$acc8,24,$acc8		!
+		xor	$acc5,$s1,$s1
+		sll	$acc9,16,$acc9
+		xor	$acc11,$s2,$s2
+		sll	$acc10,8,$acc10
+		xor	$acc6,$s1,$s1
+		sll	$acc12,24,$acc12
+		xor	$acc8,$s2,$s2
+		sll	$acc13,16,$acc13	!
+		xor	$acc9,$s2,$s2
+		sll	$acc14,8,$acc14
+		xor	$acc10,$s2,$s2
+		xor	$acc12,$acc14,$acc14
+		xor	$acc13,$s3,$s3
+		xor	$acc14,$s3,$s3
+		xor	$acc15,$s3,$s3
+
+	ret
+	restore
+.type	_sparcv9_AES_decrypt,#function
+.size	_sparcv9_AES_decrypt,(.-_sparcv9_AES_decrypt)
+
+.align	32
+.globl	AES_decrypt
+AES_decrypt:
+	or	%o0,%o1,%g1
+	andcc	%g1,3,%g0
+	bnz,pn	%xcc,.Lunaligned_dec
+	save	%sp,-$frame,%sp
+
+	ld	[%i0+0],%o0
+	ld	[%i0+4],%o1
+	ld	[%i0+8],%o2
+	ld	[%i0+12],%o3
+
+1:	call	.+8
+	add	%o7,AES_Td-1b,%o4
+	call	_sparcv9_AES_decrypt
+	mov	%i2,%o5
+
+	st	%o0,[%i1+0]
+	st	%o1,[%i1+4]
+	st	%o2,[%i1+8]
+	st	%o3,[%i1+12]
+
+	ret
+	restore
+
+.align	32
+.Lunaligned_dec:
+	ldub	[%i0+0],%l0
+	ldub	[%i0+1],%l1
+	ldub	[%i0+2],%l2
+
+	sll	%l0,24,%l0
+	ldub	[%i0+3],%l3
+	sll	%l1,16,%l1
+	ldub	[%i0+4],%l4
+	sll	%l2,8,%l2
+	or	%l1,%l0,%l0
+	ldub	[%i0+5],%l5
+	sll	%l4,24,%l4
+	or	%l3,%l2,%l2
+	ldub	[%i0+6],%l6
+	sll	%l5,16,%l5
+	or	%l0,%l2,%o0
+	ldub	[%i0+7],%l7
+
+	sll	%l6,8,%l6
+	or	%l5,%l4,%l4
+	ldub	[%i0+8],%l0
+	or	%l7,%l6,%l6
+	ldub	[%i0+9],%l1
+	or	%l4,%l6,%o1
+	ldub	[%i0+10],%l2
+
+	sll	%l0,24,%l0
+	ldub	[%i0+11],%l3
+	sll	%l1,16,%l1
+	ldub	[%i0+12],%l4
+	sll	%l2,8,%l2
+	or	%l1,%l0,%l0
+	ldub	[%i0+13],%l5
+	sll	%l4,24,%l4
+	or	%l3,%l2,%l2
+	ldub	[%i0+14],%l6
+	sll	%l5,16,%l5
+	or	%l0,%l2,%o2
+	ldub	[%i0+15],%l7
+
+	sll	%l6,8,%l6
+	or	%l5,%l4,%l4
+	or	%l7,%l6,%l6
+	or	%l4,%l6,%o3
+
+1:	call	.+8
+	add	%o7,AES_Td-1b,%o4
+	call	_sparcv9_AES_decrypt
+	mov	%i2,%o5
+
+	srl	%o0,24,%l0
+	srl	%o0,16,%l1
+	stb	%l0,[%i1+0]
+	srl	%o0,8,%l2
+	stb	%l1,[%i1+1]
+	stb	%l2,[%i1+2]
+	srl	%o1,24,%l4
+	stb	%o0,[%i1+3]
+
+	srl	%o1,16,%l5
+	stb	%l4,[%i1+4]
+	srl	%o1,8,%l6
+	stb	%l5,[%i1+5]
+	stb	%l6,[%i1+6]
+	srl	%o2,24,%l0
+	stb	%o1,[%i1+7]
+
+	srl	%o2,16,%l1
+	stb	%l0,[%i1+8]
+	srl	%o2,8,%l2
+	stb	%l1,[%i1+9]
+	stb	%l2,[%i1+10]
+	srl	%o3,24,%l4
+	stb	%o2,[%i1+11]
+
+	srl	%o3,16,%l5
+	stb	%l4,[%i1+12]
+	srl	%o3,8,%l6
+	stb	%l5,[%i1+13]
+	stb	%l6,[%i1+14]
+	stb	%o3,[%i1+15]
+
+	ret
+	restore
+.type	AES_decrypt,#function
+.size	AES_decrypt,(.-AES_decrypt)
+___
+
+# fmovs instructions substituting for FP nops were originally added
+# to meet specific instruction alignment requirements to maximize ILP.
+# As UltraSPARC T1, a.k.a. Niagara, has shared FPU, FP nops can have
+# undesired effect, so just omit them and sacrifice some portion of
+# percent in performance...
+$code =~ s/fmovs.*$//gm;
+
+print $code;
+close STDOUT;	# ensure flush
diff --git a/crypto/aes/asm/aes-x86_64.pl b/crypto/aes/asm/aes-x86_64.pl
index b510057d166c..48fa857d5b1b 100755
--- a/crypto/aes/asm/aes-x86_64.pl
+++ b/crypto/aes/asm/aes-x86_64.pl
@@ -2,11 +2,12 @@
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
-# project. Rights for redistribution and usage in source and binary
-# forms are granted according to the OpenSSL license.
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
 # ====================================================================
 #
-# Version 1.2.
+# Version 2.1.
 #
 # aes-*-cbc benchmarks are improved by >70% [compared to gcc 3.3.2 on
 # Opteron 240 CPU] plus all the bells-n-whistles from 32-bit version
@@ -17,17 +18,29 @@
 #
 # Performance in number of cycles per processed byte for 128-bit key:
 #
-#		ECB		CBC encrypt
-# AMD64		13.7		13.0(*)
-# EM64T		20.2		18.6(*)
+#		ECB encrypt	ECB decrypt	CBC large chunk
+# AMD64		33		41		13.0
+# EM64T		38		59		18.6(*)
+# Core 2	30		43		14.5(*)
 #
-# (*)	CBC benchmarks are better than ECB thanks to custom ABI used
-#	by the private block encryption function.
+# (*) with hyper-threading off
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
 
 $verticalspin=1;	# unlike 32-bit version $verticalspin performs
 			# ~15% better on both AMD and Intel cores
-$output=shift;
-open STDOUT,"| $^X ../perlasm/x86_64-xlate.pl $output";
+$speed_limit=512;	# see aes-586.pl for details
 
 $code=".text\n";
 
@@ -35,9 +48,9 @@ $s0="%eax";
 $s1="%ebx";
 $s2="%ecx";
 $s3="%edx";
-$acc0="%esi";
-$acc1="%edi";
-$acc2="%ebp";
+$acc0="%esi";	$mask80="%rsi";
+$acc1="%edi";	$maskfe="%rdi";
+$acc2="%ebp";	$mask1b="%rbp";
 $inp="%r8";
 $out="%r9";
 $t0="%r10d";
@@ -51,6 +64,8 @@ sub hi() { my $r=shift;	$r =~ s/%[er]([a-d])x/%\1h/;	$r; }
 sub lo() { my $r=shift;	$r =~ s/%[er]([a-d])x/%\1l/;
 			$r =~ s/%[er]([sd]i)/%\1l/;
 			$r =~ s/%(r[0-9]+)[d]?/%\1b/;	$r; }
+sub LO() { my $r=shift; $r =~ s/%r([a-z]+)/%e\1/;
+			$r =~ s/%r([0-9]+)/%r\1d/;	$r; }
 sub _data_word()
 { my $i;
     while(defined($i=shift)) { $code.=sprintf".long\t0x%08x,0x%08x\n",$i,$i; }
@@ -138,22 +153,17 @@ $code.=<<___;
 	movzb	`&lo("$s0")`,$acc0
 	movzb	`&lo("$s1")`,$acc1
 	movzb	`&lo("$s2")`,$acc2
-	mov	2($sbox,$acc0,8),$t0
-	mov	2($sbox,$acc1,8),$t1
-	mov	2($sbox,$acc2,8),$t2
-
-	and	\$0x000000ff,$t0
-	and	\$0x000000ff,$t1
-	and	\$0x000000ff,$t2
+	movzb	2($sbox,$acc0,8),$t0
+	movzb	2($sbox,$acc1,8),$t1
+	movzb	2($sbox,$acc2,8),$t2
 
 	movzb	`&lo("$s3")`,$acc0
 	movzb	`&hi("$s1")`,$acc1
 	movzb	`&hi("$s2")`,$acc2
-	mov	2($sbox,$acc0,8),$t3
+	movzb	2($sbox,$acc0,8),$t3
 	mov	0($sbox,$acc1,8),$acc1	#$t0
 	mov	0($sbox,$acc2,8),$acc2	#$t1
 
-	and	\$0x000000ff,$t3
 	and	\$0x0000ff00,$acc1
 	and	\$0x0000ff00,$acc2
 
@@ -345,11 +355,242 @@ $code.=<<___;
 .size	_x86_64_AES_encrypt,.-_x86_64_AES_encrypt
 ___
 
+# it's possible to implement this by shifting tN by 8, filling least
+# significant byte with byte load and finally bswap-ing at the end,
+# but such partial register load kills Core 2...
+sub enccompactvert()
+{ my ($t3,$t4,$t5)=("%r8d","%r9d","%r13d");
+
+$code.=<<___;
+	movzb	`&lo("$s0")`,$t0
+	movzb	`&lo("$s1")`,$t1
+	movzb	`&lo("$s2")`,$t2
+	movzb	($sbox,$t0,1),$t0
+	movzb	($sbox,$t1,1),$t1
+	movzb	($sbox,$t2,1),$t2
+
+	movzb	`&lo("$s3")`,$t3
+	movzb	`&hi("$s1")`,$acc0
+	movzb	`&hi("$s2")`,$acc1
+	movzb	($sbox,$t3,1),$t3
+	movzb	($sbox,$acc0,1),$t4	#$t0
+	movzb	($sbox,$acc1,1),$t5	#$t1
+
+	movzb	`&hi("$s3")`,$acc2
+	movzb	`&hi("$s0")`,$acc0
+	shr	\$16,$s2
+	movzb	($sbox,$acc2,1),$acc2	#$t2
+	movzb	($sbox,$acc0,1),$acc0	#$t3
+	shr	\$16,$s3
+
+	movzb	`&lo("$s2")`,$acc1
+	shl	\$8,$t4
+	shl	\$8,$t5
+	movzb	($sbox,$acc1,1),$acc1	#$t0
+	xor	$t4,$t0
+	xor	$t5,$t1
+
+	movzb	`&lo("$s3")`,$t4
+	shr	\$16,$s0
+	shr	\$16,$s1
+	movzb	`&lo("$s0")`,$t5
+	shl	\$8,$acc2
+	shl	\$8,$acc0
+	movzb	($sbox,$t4,1),$t4	#$t1
+	movzb	($sbox,$t5,1),$t5	#$t2
+	xor	$acc2,$t2
+	xor	$acc0,$t3
+
+	movzb	`&lo("$s1")`,$acc2
+	movzb	`&hi("$s3")`,$acc0
+	shl	\$16,$acc1
+	movzb	($sbox,$acc2,1),$acc2	#$t3
+	movzb	($sbox,$acc0,1),$acc0	#$t0
+	xor	$acc1,$t0
+
+	movzb	`&hi("$s0")`,$acc1
+	shr	\$8,$s2
+	shr	\$8,$s1
+	movzb	($sbox,$acc1,1),$acc1	#$t1
+	movzb	($sbox,$s2,1),$s3	#$t3
+	movzb	($sbox,$s1,1),$s2	#$t2
+	shl	\$16,$t4
+	shl	\$16,$t5
+	shl	\$16,$acc2
+	xor	$t4,$t1
+	xor	$t5,$t2
+	xor	$acc2,$t3
+
+	shl	\$24,$acc0
+	shl	\$24,$acc1
+	shl	\$24,$s3
+	xor	$acc0,$t0
+	shl	\$24,$s2
+	xor	$acc1,$t1
+	mov	$t0,$s0
+	mov	$t1,$s1
+	xor	$t2,$s2
+	xor	$t3,$s3
+___
+}
+
+sub enctransform_ref()
+{ my $sn = shift;
+  my ($acc,$r2,$tmp)=("%r8d","%r9d","%r13d");
+
+$code.=<<___;
+	mov	$sn,$acc
+	and	\$0x80808080,$acc
+	mov	$acc,$tmp
+	shr	\$7,$tmp
+	lea	($sn,$sn),$r2
+	sub	$tmp,$acc
+	and	\$0xfefefefe,$r2
+	and	\$0x1b1b1b1b,$acc
+	mov	$sn,$tmp
+	xor	$acc,$r2
+
+	xor	$r2,$sn
+	rol	\$24,$sn
+	xor	$r2,$sn
+	ror	\$16,$tmp
+	xor	$tmp,$sn
+	ror	\$8,$tmp
+	xor	$tmp,$sn
+___
+}
+
+# unlike decrypt case it does not pay off to parallelize enctransform
+sub enctransform()
+{ my ($t3,$r20,$r21)=($acc2,"%r8d","%r9d");
+
+$code.=<<___;
+	mov	$s0,$acc0
+	mov	$s1,$acc1
+	and	\$0x80808080,$acc0
+	and	\$0x80808080,$acc1
+	mov	$acc0,$t0
+	mov	$acc1,$t1
+	shr	\$7,$t0
+	lea	($s0,$s0),$r20
+	shr	\$7,$t1
+	lea	($s1,$s1),$r21
+	sub	$t0,$acc0
+	sub	$t1,$acc1
+	and	\$0xfefefefe,$r20
+	and	\$0xfefefefe,$r21
+	and	\$0x1b1b1b1b,$acc0
+	and	\$0x1b1b1b1b,$acc1
+	mov	$s0,$t0
+	mov	$s1,$t1
+	xor	$acc0,$r20
+	xor	$acc1,$r21
+
+	xor	$r20,$s0
+	xor	$r21,$s1
+	 mov	$s2,$acc0
+	 mov	$s3,$acc1
+	rol	\$24,$s0
+	rol	\$24,$s1
+	 and	\$0x80808080,$acc0
+	 and	\$0x80808080,$acc1
+	xor	$r20,$s0
+	xor	$r21,$s1
+	 mov	$acc0,$t2
+	 mov	$acc1,$t3
+	ror	\$16,$t0
+	ror	\$16,$t1
+	 shr	\$7,$t2
+	 lea	($s2,$s2),$r20
+	xor	$t0,$s0
+	xor	$t1,$s1
+	 shr	\$7,$t3
+	 lea	($s3,$s3),$r21
+	ror	\$8,$t0
+	ror	\$8,$t1
+	 sub	$t2,$acc0
+	 sub	$t3,$acc1
+	xor	$t0,$s0
+	xor	$t1,$s1
+
+	and	\$0xfefefefe,$r20
+	and	\$0xfefefefe,$r21
+	and	\$0x1b1b1b1b,$acc0
+	and	\$0x1b1b1b1b,$acc1
+	mov	$s2,$t2
+	mov	$s3,$t3
+	xor	$acc0,$r20
+	xor	$acc1,$r21
+
+	xor	$r20,$s2
+	xor	$r21,$s3
+	rol	\$24,$s2
+	rol	\$24,$s3
+	xor	$r20,$s2
+	xor	$r21,$s3
+	mov	0($sbox),$acc0			# prefetch Te4
+	ror	\$16,$t2
+	ror	\$16,$t3
+	mov	64($sbox),$acc1
+	xor	$t2,$s2
+	xor	$t3,$s3
+	mov	128($sbox),$r20
+	ror	\$8,$t2
+	ror	\$8,$t3
+	mov	192($sbox),$r21
+	xor	$t2,$s2
+	xor	$t3,$s3
+___
+}
+
+$code.=<<___;
+.type	_x86_64_AES_encrypt_compact,\@abi-omnipotent
+.align	16
+_x86_64_AES_encrypt_compact:
+	lea	128($sbox),$inp			# size optimization
+	mov	0-128($inp),$acc1		# prefetch Te4
+	mov	32-128($inp),$acc2
+	mov	64-128($inp),$t0
+	mov	96-128($inp),$t1
+	mov	128-128($inp),$acc1
+	mov	160-128($inp),$acc2
+	mov	192-128($inp),$t0
+	mov	224-128($inp),$t1
+	jmp	.Lenc_loop_compact
+.align	16
+.Lenc_loop_compact:
+		xor	0($key),$s0		# xor with key
+		xor	4($key),$s1
+		xor	8($key),$s2
+		xor	12($key),$s3
+		lea	16($key),$key
+___
+		&enccompactvert();
+$code.=<<___;
+		cmp	16(%rsp),$key
+		je	.Lenc_compact_done
+___
+		&enctransform();
+$code.=<<___;
+	jmp	.Lenc_loop_compact
+.align	16
+.Lenc_compact_done:
+	xor	0($key),$s0
+	xor	4($key),$s1
+	xor	8($key),$s2
+	xor	12($key),$s3
+	.byte	0xf3,0xc3			# rep ret
+.size	_x86_64_AES_encrypt_compact,.-_x86_64_AES_encrypt_compact
+___
+
 # void AES_encrypt (const void *inp,void *out,const AES_KEY *key);
 $code.=<<___;
 .globl	AES_encrypt
 .type	AES_encrypt,\@function,3
 .align	16
+.globl	asm_AES_encrypt
+.hidden	asm_AES_encrypt
+asm_AES_encrypt:
 AES_encrypt:
 	push	%rbx
 	push	%rbp
@@ -358,31 +599,57 @@ AES_encrypt:
 	push	%r14
 	push	%r15
 
-	mov	%rdx,$key
-	mov	%rdi,$inp
-	mov	%rsi,$out
-
-	.picmeup	$sbox
-	lea	AES_Te-.($sbox),$sbox
-
-	mov	0($inp),$s0
-	mov	4($inp),$s1
-	mov	8($inp),$s2
-	mov	12($inp),$s3
+	# allocate frame "above" key schedule
+	mov	%rsp,%r10
+	lea	-63(%rdx),%rcx	# %rdx is key argument
+	and	\$-64,%rsp
+	sub	%rsp,%rcx
+	neg	%rcx
+	and	\$0x3c0,%rcx
+	sub	%rcx,%rsp
+	sub	\$32,%rsp
 
-	call	_x86_64_AES_encrypt
+	mov	%rsi,16(%rsp)	# save out
+	mov	%r10,24(%rsp)	# save real stack pointer
+.Lenc_prologue:
 
-	mov	$s0,0($out)
+	mov	%rdx,$key
+	mov	240($key),$rnds	# load rounds
+
+	mov	0(%rdi),$s0	# load input vector
+	mov	4(%rdi),$s1
+	mov	8(%rdi),$s2
+	mov	12(%rdi),$s3
+
+	shl	\$4,$rnds
+	lea	($key,$rnds),%rbp
+	mov	$key,(%rsp)	# key schedule
+	mov	%rbp,8(%rsp)	# end of key schedule
+
+	# pick Te4 copy which can't "overlap" with stack frame or key schedule
+	lea	.LAES_Te+2048(%rip),$sbox
+	lea	768(%rsp),%rbp
+	sub	$sbox,%rbp
+	and	\$0x300,%rbp
+	lea	($sbox,%rbp),$sbox
+
+	call	_x86_64_AES_encrypt_compact
+
+	mov	16(%rsp),$out	# restore out
+	mov	24(%rsp),%rsi	# restore saved stack pointer
+	mov	$s0,0($out)	# write output vector
 	mov	$s1,4($out)
 	mov	$s2,8($out)
 	mov	$s3,12($out)
 
-	pop	%r15
-	pop	%r14
-	pop	%r13
-	pop	%r12
-	pop	%rbp
-	pop	%rbx
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lenc_epilogue:
 	ret
 .size	AES_encrypt,.-AES_encrypt
 ___
@@ -453,19 +720,20 @@ sub declastvert()
 { my $t3="%r8d";	# zaps $inp!
 
 $code.=<<___;
+	lea	2048($sbox),$sbox	# size optimization
 	movzb	`&lo("$s0")`,$acc0
 	movzb	`&lo("$s1")`,$acc1
 	movzb	`&lo("$s2")`,$acc2
-	movzb	2048($sbox,$acc0,1),$t0
-	movzb	2048($sbox,$acc1,1),$t1
-	movzb	2048($sbox,$acc2,1),$t2
+	movzb	($sbox,$acc0,1),$t0
+	movzb	($sbox,$acc1,1),$t1
+	movzb	($sbox,$acc2,1),$t2
 
 	movzb	`&lo("$s3")`,$acc0
 	movzb	`&hi("$s3")`,$acc1
 	movzb	`&hi("$s0")`,$acc2
-	movzb	2048($sbox,$acc0,1),$t3
-	movzb	2048($sbox,$acc1,1),$acc1	#$t0
-	movzb	2048($sbox,$acc2,1),$acc2	#$t1
+	movzb	($sbox,$acc0,1),$t3
+	movzb	($sbox,$acc1,1),$acc1	#$t0
+	movzb	($sbox,$acc2,1),$acc2	#$t1
 
 	shl	\$8,$acc1
 	shl	\$8,$acc2
@@ -477,8 +745,8 @@ $code.=<<___;
 	movzb	`&hi("$s1")`,$acc0
 	movzb	`&hi("$s2")`,$acc1
 	shr	\$16,$s0
-	movzb	2048($sbox,$acc0,1),$acc0	#$t2
-	movzb	2048($sbox,$acc1,1),$acc1	#$t3
+	movzb	($sbox,$acc0,1),$acc0	#$t2
+	movzb	($sbox,$acc1,1),$acc1	#$t3
 
 	shl	\$8,$acc0
 	shl	\$8,$acc1
@@ -490,9 +758,9 @@ $code.=<<___;
 	movzb	`&lo("$s2")`,$acc0
 	movzb	`&lo("$s3")`,$acc1
 	movzb	`&lo("$s0")`,$acc2
-	movzb	2048($sbox,$acc0,1),$acc0	#$t0
-	movzb	2048($sbox,$acc1,1),$acc1	#$t1
-	movzb	2048($sbox,$acc2,1),$acc2	#$t2
+	movzb	($sbox,$acc0,1),$acc0	#$t0
+	movzb	($sbox,$acc1,1),$acc1	#$t1
+	movzb	($sbox,$acc2,1),$acc2	#$t2
 
 	shl	\$16,$acc0
 	shl	\$16,$acc1
@@ -505,9 +773,9 @@ $code.=<<___;
 	movzb	`&lo("$s1")`,$acc0
 	movzb	`&hi("$s1")`,$acc1
 	movzb	`&hi("$s2")`,$acc2
-	movzb	2048($sbox,$acc0,1),$acc0	#$t3
-	movzb	2048($sbox,$acc1,1),$acc1	#$t0
-	movzb	2048($sbox,$acc2,1),$acc2	#$t1
+	movzb	($sbox,$acc0,1),$acc0	#$t3
+	movzb	($sbox,$acc1,1),$acc1	#$t0
+	movzb	($sbox,$acc2,1),$acc2	#$t1
 
 	shl	\$16,$acc0
 	shl	\$24,$acc1
@@ -520,8 +788,8 @@ $code.=<<___;
 	movzb	`&hi("$s3")`,$acc0
 	movzb	`&hi("$s0")`,$acc1
 	mov	16+12($key),$s3
-	movzb	2048($sbox,$acc0,1),$acc0	#$t2
-	movzb	2048($sbox,$acc1,1),$acc1	#$t3
+	movzb	($sbox,$acc0,1),$acc0	#$t2
+	movzb	($sbox,$acc1,1),$acc1	#$t3
 	mov	16+0($key),$s0
 
 	shl	\$24,$acc0
@@ -532,6 +800,7 @@ $code.=<<___;
 
 	mov	16+4($key),$s1
 	mov	16+8($key),$s2
+	lea	-2048($sbox),$sbox
 	xor	$t0,$s0
 	xor	$t1,$s1
 	xor	$t2,$s2
@@ -659,11 +928,268 @@ $code.=<<___;
 .size	_x86_64_AES_decrypt,.-_x86_64_AES_decrypt
 ___
 
+sub deccompactvert()
+{ my ($t3,$t4,$t5)=("%r8d","%r9d","%r13d");
+
+$code.=<<___;
+	movzb	`&lo("$s0")`,$t0
+	movzb	`&lo("$s1")`,$t1
+	movzb	`&lo("$s2")`,$t2
+	movzb	($sbox,$t0,1),$t0
+	movzb	($sbox,$t1,1),$t1
+	movzb	($sbox,$t2,1),$t2
+
+	movzb	`&lo("$s3")`,$t3
+	movzb	`&hi("$s3")`,$acc0
+	movzb	`&hi("$s0")`,$acc1
+	movzb	($sbox,$t3,1),$t3
+	movzb	($sbox,$acc0,1),$t4	#$t0
+	movzb	($sbox,$acc1,1),$t5	#$t1
+
+	movzb	`&hi("$s1")`,$acc2
+	movzb	`&hi("$s2")`,$acc0
+	shr	\$16,$s2
+	movzb	($sbox,$acc2,1),$acc2	#$t2
+	movzb	($sbox,$acc0,1),$acc0	#$t3
+	shr	\$16,$s3
+
+	movzb	`&lo("$s2")`,$acc1
+	shl	\$8,$t4
+	shl	\$8,$t5
+	movzb	($sbox,$acc1,1),$acc1	#$t0
+	xor	$t4,$t0
+	xor	$t5,$t1
+
+	movzb	`&lo("$s3")`,$t4
+	shr	\$16,$s0
+	shr	\$16,$s1
+	movzb	`&lo("$s0")`,$t5
+	shl	\$8,$acc2
+	shl	\$8,$acc0
+	movzb	($sbox,$t4,1),$t4	#$t1
+	movzb	($sbox,$t5,1),$t5	#$t2
+	xor	$acc2,$t2
+	xor	$acc0,$t3
+
+	movzb	`&lo("$s1")`,$acc2
+	movzb	`&hi("$s1")`,$acc0
+	shl	\$16,$acc1
+	movzb	($sbox,$acc2,1),$acc2	#$t3
+	movzb	($sbox,$acc0,1),$acc0	#$t0
+	xor	$acc1,$t0
+
+	movzb	`&hi("$s2")`,$acc1
+	shl	\$16,$t4
+	shl	\$16,$t5
+	movzb	($sbox,$acc1,1),$s1	#$t1
+	xor	$t4,$t1
+	xor	$t5,$t2
+
+	movzb	`&hi("$s3")`,$acc1
+	shr	\$8,$s0
+	shl	\$16,$acc2
+	movzb	($sbox,$acc1,1),$s2	#$t2
+	movzb	($sbox,$s0,1),$s3	#$t3
+	xor	$acc2,$t3
+
+	shl	\$24,$acc0
+	shl	\$24,$s1
+	shl	\$24,$s2
+	xor	$acc0,$t0
+	shl	\$24,$s3
+	xor	$t1,$s1
+	mov	$t0,$s0
+	xor	$t2,$s2
+	xor	$t3,$s3
+___
+}
+
+# parallelized version! input is pair of 64-bit values: %rax=s1.s0
+# and %rcx=s3.s2, output is four 32-bit values in %eax=s0, %ebx=s1,
+# %ecx=s2 and %edx=s3.
+sub dectransform()
+{ my ($tp10,$tp20,$tp40,$tp80,$acc0)=("%rax","%r8", "%r9", "%r10","%rbx");
+  my ($tp18,$tp28,$tp48,$tp88,$acc8)=("%rcx","%r11","%r12","%r13","%rdx");
+  my $prefetch = shift;
+
+$code.=<<___;
+	mov	$tp10,$acc0
+	mov	$tp18,$acc8
+	and	$mask80,$acc0
+	and	$mask80,$acc8
+	mov	$acc0,$tp40
+	mov	$acc8,$tp48
+	shr	\$7,$tp40
+	lea	($tp10,$tp10),$tp20
+	shr	\$7,$tp48
+	lea	($tp18,$tp18),$tp28
+	sub	$tp40,$acc0
+	sub	$tp48,$acc8
+	and	$maskfe,$tp20
+	and	$maskfe,$tp28
+	and	$mask1b,$acc0
+	and	$mask1b,$acc8
+	xor	$tp20,$acc0
+	xor	$tp28,$acc8
+	mov	$acc0,$tp20
+	mov	$acc8,$tp28
+
+	and	$mask80,$acc0
+	and	$mask80,$acc8
+	mov	$acc0,$tp80
+	mov	$acc8,$tp88
+	shr	\$7,$tp80
+	lea	($tp20,$tp20),$tp40
+	shr	\$7,$tp88
+	lea	($tp28,$tp28),$tp48
+	sub	$tp80,$acc0
+	sub	$tp88,$acc8
+	and	$maskfe,$tp40
+	and	$maskfe,$tp48
+	and	$mask1b,$acc0
+	and	$mask1b,$acc8
+	xor	$tp40,$acc0
+	xor	$tp48,$acc8
+	mov	$acc0,$tp40
+	mov	$acc8,$tp48
+
+	and	$mask80,$acc0
+	and	$mask80,$acc8
+	mov	$acc0,$tp80
+	mov	$acc8,$tp88
+	shr	\$7,$tp80
+	 xor	$tp10,$tp20		# tp2^=tp1
+	shr	\$7,$tp88
+	 xor	$tp18,$tp28		# tp2^=tp1
+	sub	$tp80,$acc0
+	sub	$tp88,$acc8
+	lea	($tp40,$tp40),$tp80
+	lea	($tp48,$tp48),$tp88
+	 xor	$tp10,$tp40		# tp4^=tp1
+	 xor	$tp18,$tp48		# tp4^=tp1
+	and	$maskfe,$tp80
+	and	$maskfe,$tp88
+	and	$mask1b,$acc0
+	and	$mask1b,$acc8
+	xor	$acc0,$tp80
+	xor	$acc8,$tp88
+
+	xor	$tp80,$tp10		# tp1^=tp8
+	xor	$tp88,$tp18		# tp1^=tp8
+	xor	$tp80,$tp20		# tp2^tp1^=tp8
+	xor	$tp88,$tp28		# tp2^tp1^=tp8
+	mov	$tp10,$acc0
+	mov	$tp18,$acc8
+	xor	$tp80,$tp40		# tp4^tp1^=tp8
+	xor	$tp88,$tp48		# tp4^tp1^=tp8
+	shr	\$32,$acc0
+	shr	\$32,$acc8
+	xor	$tp20,$tp80		# tp8^=tp8^tp2^tp1=tp2^tp1
+	xor	$tp28,$tp88		# tp8^=tp8^tp2^tp1=tp2^tp1
+	rol	\$8,`&LO("$tp10")`	# ROTATE(tp1^tp8,8)
+	rol	\$8,`&LO("$tp18")`	# ROTATE(tp1^tp8,8)
+	xor	$tp40,$tp80		# tp2^tp1^=tp8^tp4^tp1=tp8^tp4^tp2
+	xor	$tp48,$tp88		# tp2^tp1^=tp8^tp4^tp1=tp8^tp4^tp2
+
+	rol	\$8,`&LO("$acc0")`	# ROTATE(tp1^tp8,8)
+	rol	\$8,`&LO("$acc8")`	# ROTATE(tp1^tp8,8)
+	xor	`&LO("$tp80")`,`&LO("$tp10")`
+	xor	`&LO("$tp88")`,`&LO("$tp18")`
+	shr	\$32,$tp80
+	shr	\$32,$tp88
+	xor	`&LO("$tp80")`,`&LO("$acc0")`
+	xor	`&LO("$tp88")`,`&LO("$acc8")`
+
+	mov	$tp20,$tp80
+	mov	$tp28,$tp88
+	shr	\$32,$tp80
+	shr	\$32,$tp88
+	rol	\$24,`&LO("$tp20")`	# ROTATE(tp2^tp1^tp8,24)
+	rol	\$24,`&LO("$tp28")`	# ROTATE(tp2^tp1^tp8,24)
+	rol	\$24,`&LO("$tp80")`	# ROTATE(tp2^tp1^tp8,24)
+	rol	\$24,`&LO("$tp88")`	# ROTATE(tp2^tp1^tp8,24)
+	xor	`&LO("$tp20")`,`&LO("$tp10")`
+	xor	`&LO("$tp28")`,`&LO("$tp18")`
+	mov	$tp40,$tp20
+	mov	$tp48,$tp28
+	xor	`&LO("$tp80")`,`&LO("$acc0")`
+	xor	`&LO("$tp88")`,`&LO("$acc8")`
+
+	`"mov	0($sbox),$mask80"	if ($prefetch)`
+	shr	\$32,$tp20
+	shr	\$32,$tp28
+	`"mov	64($sbox),$maskfe"	if ($prefetch)`
+	rol	\$16,`&LO("$tp40")`	# ROTATE(tp4^tp1^tp8,16)
+	rol	\$16,`&LO("$tp48")`	# ROTATE(tp4^tp1^tp8,16)
+	`"mov	128($sbox),$mask1b"	if ($prefetch)`
+	rol	\$16,`&LO("$tp20")`	# ROTATE(tp4^tp1^tp8,16)
+	rol	\$16,`&LO("$tp28")`	# ROTATE(tp4^tp1^tp8,16)
+	`"mov	192($sbox),$tp80"	if ($prefetch)`
+	xor	`&LO("$tp40")`,`&LO("$tp10")`
+	xor	`&LO("$tp48")`,`&LO("$tp18")`
+	`"mov	256($sbox),$tp88"	if ($prefetch)`
+	xor	`&LO("$tp20")`,`&LO("$acc0")`
+	xor	`&LO("$tp28")`,`&LO("$acc8")`
+___
+}
+
+$code.=<<___;
+.type	_x86_64_AES_decrypt_compact,\@abi-omnipotent
+.align	16
+_x86_64_AES_decrypt_compact:
+	lea	128($sbox),$inp			# size optimization
+	mov	0-128($inp),$acc1		# prefetch Td4
+	mov	32-128($inp),$acc2
+	mov	64-128($inp),$t0
+	mov	96-128($inp),$t1
+	mov	128-128($inp),$acc1
+	mov	160-128($inp),$acc2
+	mov	192-128($inp),$t0
+	mov	224-128($inp),$t1
+	jmp	.Ldec_loop_compact
+
+.align	16
+.Ldec_loop_compact:
+		xor	0($key),$s0		# xor with key
+		xor	4($key),$s1
+		xor	8($key),$s2
+		xor	12($key),$s3
+		lea	16($key),$key
+___
+		&deccompactvert();
+$code.=<<___;
+		cmp	16(%rsp),$key
+		je	.Ldec_compact_done
+
+		mov	256+0($sbox),$mask80
+		shl	\$32,%rbx
+		shl	\$32,%rdx
+		mov	256+8($sbox),$maskfe
+		or	%rbx,%rax
+		or	%rdx,%rcx
+		mov	256+16($sbox),$mask1b
+___
+		&dectransform(1);
+$code.=<<___;
+	jmp	.Ldec_loop_compact
+.align	16
+.Ldec_compact_done:
+	xor	0($key),$s0
+	xor	4($key),$s1
+	xor	8($key),$s2
+	xor	12($key),$s3
+	.byte	0xf3,0xc3			# rep ret
+.size	_x86_64_AES_decrypt_compact,.-_x86_64_AES_decrypt_compact
+___
+
 # void AES_decrypt (const void *inp,void *out,const AES_KEY *key);
 $code.=<<___;
 .globl	AES_decrypt
 .type	AES_decrypt,\@function,3
 .align	16
+.globl	asm_AES_decrypt
+.hidden	asm_AES_decrypt
+asm_AES_decrypt:
 AES_decrypt:
 	push	%rbx
 	push	%rbp
@@ -672,43 +1198,59 @@ AES_decrypt:
 	push	%r14
 	push	%r15
 
-	mov	%rdx,$key
-	mov	%rdi,$inp
-	mov	%rsi,$out
+	# allocate frame "above" key schedule
+	mov	%rsp,%r10
+	lea	-63(%rdx),%rcx	# %rdx is key argument
+	and	\$-64,%rsp
+	sub	%rsp,%rcx
+	neg	%rcx
+	and	\$0x3c0,%rcx
+	sub	%rcx,%rsp
+	sub	\$32,%rsp
 
-	.picmeup	$sbox
-	lea	AES_Td-.($sbox),$sbox
-
-	# prefetch Td4
-	lea	2048+128($sbox),$sbox;
-	mov	0-128($sbox),$s0
-	mov	32-128($sbox),$s1
-	mov	64-128($sbox),$s2
-	mov	96-128($sbox),$s3
-	mov	128-128($sbox),$s0
-	mov	160-128($sbox),$s1
-	mov	192-128($sbox),$s2
-	mov	224-128($sbox),$s3
-	lea	-2048-128($sbox),$sbox;
-
-	mov	0($inp),$s0
-	mov	4($inp),$s1
-	mov	8($inp),$s2
-	mov	12($inp),$s3
-
-	call	_x86_64_AES_decrypt
-
-	mov	$s0,0($out)
+	mov	%rsi,16(%rsp)	# save out
+	mov	%r10,24(%rsp)	# save real stack pointer
+.Ldec_prologue:
+
+	mov	%rdx,$key
+	mov	240($key),$rnds	# load rounds
+
+	mov	0(%rdi),$s0	# load input vector
+	mov	4(%rdi),$s1
+	mov	8(%rdi),$s2
+	mov	12(%rdi),$s3
+
+	shl	\$4,$rnds
+	lea	($key,$rnds),%rbp
+	mov	$key,(%rsp)	# key schedule
+	mov	%rbp,8(%rsp)	# end of key schedule
+
+	# pick Td4 copy which can't "overlap" with stack frame or key schedule
+	lea	.LAES_Td+2048(%rip),$sbox
+	lea	768(%rsp),%rbp
+	sub	$sbox,%rbp
+	and	\$0x300,%rbp
+	lea	($sbox,%rbp),$sbox
+	shr	\$3,%rbp	# recall "magic" constants!
+	add	%rbp,$sbox
+
+	call	_x86_64_AES_decrypt_compact
+
+	mov	16(%rsp),$out	# restore out
+	mov	24(%rsp),%rsi	# restore saved stack pointer
+	mov	$s0,0($out)	# write output vector
 	mov	$s1,4($out)
 	mov	$s2,8($out)
 	mov	$s3,12($out)
 
-	pop	%r15
-	pop	%r14
-	pop	%r13
-	pop	%r12
-	pop	%rbp
-	pop	%rbx
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Ldec_epilogue:
 	ret
 .size	AES_decrypt,.-AES_decrypt
 ___
@@ -718,48 +1260,57 @@ sub enckey()
 {
 $code.=<<___;
 	movz	%dl,%esi		# rk[i]>>0
-	mov	2(%rbp,%rsi,8),%ebx
+	movzb	-128(%rbp,%rsi),%ebx
 	movz	%dh,%esi		# rk[i]>>8
-	and	\$0xFF000000,%ebx
+	shl	\$24,%ebx
 	xor	%ebx,%eax
 
-	mov	2(%rbp,%rsi,8),%ebx
+	movzb	-128(%rbp,%rsi),%ebx
 	shr	\$16,%edx
-	and	\$0x000000FF,%ebx
 	movz	%dl,%esi		# rk[i]>>16
 	xor	%ebx,%eax
 
-	mov	0(%rbp,%rsi,8),%ebx
+	movzb	-128(%rbp,%rsi),%ebx
 	movz	%dh,%esi		# rk[i]>>24
-	and	\$0x0000FF00,%ebx
+	shl	\$8,%ebx
 	xor	%ebx,%eax
 
-	mov	0(%rbp,%rsi,8),%ebx
-	and	\$0x00FF0000,%ebx
+	movzb	-128(%rbp,%rsi),%ebx
+	shl	\$16,%ebx
 	xor	%ebx,%eax
 
-	xor	2048(%rbp,%rcx,4),%eax		# rcon
+	xor	1024-128(%rbp,%rcx,4),%eax		# rcon
 ___
 }
 
-# int AES_set_encrypt_key(const unsigned char *userKey, const int bits,
+# int private_AES_set_encrypt_key(const unsigned char *userKey, const int bits,
 #                        AES_KEY *key)
 $code.=<<___;
-.globl	AES_set_encrypt_key
-.type	AES_set_encrypt_key,\@function,3
+.globl	private_AES_set_encrypt_key
+.type	private_AES_set_encrypt_key,\@function,3
 .align	16
-AES_set_encrypt_key:
+private_AES_set_encrypt_key:
 	push	%rbx
 	push	%rbp
+	push	%r12			# redundant, but allows to share 
+	push	%r13			# exception handler...
+	push	%r14
+	push	%r15
 	sub	\$8,%rsp
+.Lenc_key_prologue:
 
 	call	_x86_64_AES_set_encrypt_key
 
-	mov	8(%rsp),%rbp
-	mov	16(%rsp),%rbx
-	add	\$24,%rsp
+	mov	8(%rsp),%r15
+	mov	16(%rsp),%r14
+	mov	24(%rsp),%r13
+	mov	32(%rsp),%r12
+	mov	40(%rsp),%rbp
+	mov	48(%rsp),%rbx
+	add	\$56,%rsp
+.Lenc_key_epilogue:
 	ret
-.size	AES_set_encrypt_key,.-AES_set_encrypt_key
+.size	private_AES_set_encrypt_key,.-private_AES_set_encrypt_key
 
 .type	_x86_64_AES_set_encrypt_key,\@abi-omnipotent
 .align	16
@@ -773,8 +1324,18 @@ _x86_64_AES_set_encrypt_key:
 	test	\$-1,%rdi
 	jz	.Lbadpointer
 
-	.picmeup %rbp
-	lea	AES_Te-.(%rbp),%rbp
+	lea	.LAES_Te(%rip),%rbp
+	lea	2048+128(%rbp),%rbp
+
+	# prefetch Te4
+	mov	0-128(%rbp),%eax
+	mov	32-128(%rbp),%ebx
+	mov	64-128(%rbp),%r8d
+	mov	96-128(%rbp),%edx
+	mov	128-128(%rbp),%eax
+	mov	160-128(%rbp),%ebx
+	mov	192-128(%rbp),%r8d
+	mov	224-128(%rbp),%edx
 
 	cmp	\$128,%ecx
 	je	.L10rounds
@@ -786,15 +1347,12 @@ _x86_64_AES_set_encrypt_key:
 	jmp	.Lexit
 
 .L10rounds:
-	mov	0(%rsi),%eax			# copy first 4 dwords
-	mov	4(%rsi),%ebx
-	mov	8(%rsi),%ecx
-	mov	12(%rsi),%edx
-	mov	%eax,0(%rdi)
-	mov	%ebx,4(%rdi)
-	mov	%ecx,8(%rdi)
-	mov	%edx,12(%rdi)
+	mov	0(%rsi),%rax			# copy first 4 dwords
+	mov	8(%rsi),%rdx
+	mov	%rax,0(%rdi)
+	mov	%rdx,8(%rdi)
 
+	shr	\$32,%rdx
 	xor	%ecx,%ecx
 	jmp	.L10shortcut
 .align	4
@@ -822,19 +1380,14 @@ $code.=<<___;
 	jmp	.Lexit
 
 .L12rounds:
-	mov	0(%rsi),%eax			# copy first 6 dwords
-	mov	4(%rsi),%ebx
-	mov	8(%rsi),%ecx
-	mov	12(%rsi),%edx
-	mov	%eax,0(%rdi)
-	mov	%ebx,4(%rdi)
-	mov	%ecx,8(%rdi)
-	mov	%edx,12(%rdi)
-	mov	16(%rsi),%ecx
-	mov	20(%rsi),%edx
-	mov	%ecx,16(%rdi)
-	mov	%edx,20(%rdi)
-
+	mov	0(%rsi),%rax			# copy first 6 dwords
+	mov	8(%rsi),%rbx
+	mov	16(%rsi),%rdx
+	mov	%rax,0(%rdi)
+	mov	%rbx,8(%rdi)
+	mov	%rdx,16(%rdi)
+
+	shr	\$32,%rdx
 	xor	%ecx,%ecx
 	jmp	.L12shortcut
 .align	4
@@ -870,30 +1423,23 @@ $code.=<<___;
 	jmp	.Lexit
 
 .L14rounds:		
-	mov	0(%rsi),%eax			# copy first 8 dwords
-	mov	4(%rsi),%ebx
-	mov	8(%rsi),%ecx
-	mov	12(%rsi),%edx
-	mov	%eax,0(%rdi)
-	mov	%ebx,4(%rdi)
-	mov	%ecx,8(%rdi)
-	mov	%edx,12(%rdi)
-	mov	16(%rsi),%eax
-	mov	20(%rsi),%ebx
-	mov	24(%rsi),%ecx
-	mov	28(%rsi),%edx
-	mov	%eax,16(%rdi)
-	mov	%ebx,20(%rdi)
-	mov	%ecx,24(%rdi)
-	mov	%edx,28(%rdi)
-
+	mov	0(%rsi),%rax			# copy first 8 dwords
+	mov	8(%rsi),%rbx
+	mov	16(%rsi),%rcx
+	mov	24(%rsi),%rdx
+	mov	%rax,0(%rdi)
+	mov	%rbx,8(%rdi)
+	mov	%rcx,16(%rdi)
+	mov	%rdx,24(%rdi)
+
+	shr	\$32,%rdx
 	xor	%ecx,%ecx
 	jmp	.L14shortcut
 .align	4
 .L14loop:
+		mov	0(%rdi),%eax			# rk[0]
 		mov	28(%rdi),%edx			# rk[4]
 .L14shortcut:
-		mov	0(%rdi),%eax			# rk[0]
 ___
 		&enckey	();
 $code.=<<___;
@@ -912,24 +1458,23 @@ $code.=<<___;
 		mov	%eax,%edx
 		mov	16(%rdi),%eax			# rk[4]
 		movz	%dl,%esi			# rk[11]>>0
-		mov	2(%rbp,%rsi,8),%ebx
+		movzb	-128(%rbp,%rsi),%ebx
 		movz	%dh,%esi			# rk[11]>>8
-		and	\$0x000000FF,%ebx
 		xor	%ebx,%eax
 
-		mov	0(%rbp,%rsi,8),%ebx
+		movzb	-128(%rbp,%rsi),%ebx
 		shr	\$16,%edx
-		and	\$0x0000FF00,%ebx
+		shl	\$8,%ebx
 		movz	%dl,%esi			# rk[11]>>16
 		xor	%ebx,%eax
 
-		mov	0(%rbp,%rsi,8),%ebx
+		movzb	-128(%rbp,%rsi),%ebx
 		movz	%dh,%esi			# rk[11]>>24
-		and	\$0x00FF0000,%ebx
+		shl	\$16,%ebx
 		xor	%ebx,%eax
 
-		mov	2(%rbp,%rsi,8),%ebx
-		and	\$0xFF000000,%ebx
+		movzb	-128(%rbp,%rsi),%ebx
+		shl	\$24,%ebx
 		xor	%ebx,%eax
 
 		mov	%eax,48(%rdi)			# rk[12]
@@ -950,51 +1495,88 @@ $code.=<<___;
 .Lbadpointer:
 	mov	\$-1,%rax
 .Lexit:
-	.byte	0xf3,0xc3		# rep ret
+	.byte	0xf3,0xc3			# rep ret
 .size	_x86_64_AES_set_encrypt_key,.-_x86_64_AES_set_encrypt_key
 ___
 
-sub deckey()
+sub deckey_ref()
 { my ($i,$ptr,$te,$td) = @_;
+  my ($tp1,$tp2,$tp4,$tp8,$acc)=("%eax","%ebx","%edi","%edx","%r8d");
 $code.=<<___;
-	mov	$i($ptr),%eax
-	mov	%eax,%edx
-	movz	%ah,%ebx
-	shr	\$16,%edx
-	and	\$0xFF,%eax
-	movzb	2($te,%rax,8),%rax
-	movzb	2($te,%rbx,8),%rbx
-	mov	0($td,%rax,8),%eax
-	xor	3($td,%rbx,8),%eax
-	movzb	%dh,%ebx
-	and	\$0xFF,%edx
-	movzb	2($te,%rdx,8),%rdx
-	movzb	2($te,%rbx,8),%rbx
-	xor	2($td,%rdx,8),%eax
-	xor	1($td,%rbx,8),%eax
-	mov	%eax,$i($ptr)
+	mov	$i($ptr),$tp1
+	mov	$tp1,$acc
+	and	\$0x80808080,$acc
+	mov	$acc,$tp4
+	shr	\$7,$tp4
+	lea	0($tp1,$tp1),$tp2
+	sub	$tp4,$acc
+	and	\$0xfefefefe,$tp2
+	and	\$0x1b1b1b1b,$acc
+	xor	$tp2,$acc
+	mov	$acc,$tp2
+
+	and	\$0x80808080,$acc
+	mov	$acc,$tp8
+	shr	\$7,$tp8
+	lea	0($tp2,$tp2),$tp4
+	sub	$tp8,$acc
+	and	\$0xfefefefe,$tp4
+	and	\$0x1b1b1b1b,$acc
+	 xor	$tp1,$tp2		# tp2^tp1
+	xor	$tp4,$acc
+	mov	$acc,$tp4
+
+	and	\$0x80808080,$acc
+	mov	$acc,$tp8
+	shr	\$7,$tp8
+	sub	$tp8,$acc
+	lea	0($tp4,$tp4),$tp8
+	 xor	$tp1,$tp4		# tp4^tp1
+	and	\$0xfefefefe,$tp8
+	and	\$0x1b1b1b1b,$acc
+	xor	$acc,$tp8
+
+	xor	$tp8,$tp1		# tp1^tp8
+	rol	\$8,$tp1		# ROTATE(tp1^tp8,8)
+	xor	$tp8,$tp2		# tp2^tp1^tp8
+	xor	$tp8,$tp4		# tp4^tp1^tp8
+	xor	$tp2,$tp8
+	xor	$tp4,$tp8		# tp8^(tp8^tp4^tp1)^(tp8^tp2^tp1)=tp8^tp4^tp2
+
+	xor	$tp8,$tp1
+	rol	\$24,$tp2		# ROTATE(tp2^tp1^tp8,24)
+	xor	$tp2,$tp1
+	rol	\$16,$tp4		# ROTATE(tp4^tp1^tp8,16)
+	xor	$tp4,$tp1
+
+	mov	$tp1,$i($ptr)
 ___
 }
 
-# int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
+# int private_AES_set_decrypt_key(const unsigned char *userKey, const int bits,
 #                        AES_KEY *key)
 $code.=<<___;
-.globl	AES_set_decrypt_key
-.type	AES_set_decrypt_key,\@function,3
+.globl	private_AES_set_decrypt_key
+.type	private_AES_set_decrypt_key,\@function,3
 .align	16
-AES_set_decrypt_key:
+private_AES_set_decrypt_key:
 	push	%rbx
 	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
 	push	%rdx			# save key schedule
+.Ldec_key_prologue:
 
 	call	_x86_64_AES_set_encrypt_key
 	mov	(%rsp),%r8		# restore key schedule
 	cmp	\$0,%eax
 	jne	.Labort
 
-	mov	240(%r8),%ecx		# pull number of rounds
+	mov	240(%r8),%r14d		# pull number of rounds
 	xor	%rdi,%rdi
-	lea	(%rdi,%rcx,4),%rcx
+	lea	(%rdi,%r14d,4),%rcx
 	mov	%r8,%rsi
 	lea	(%r8,%rcx,4),%rdi	# pointer to last chunk
 .align	4
@@ -1012,32 +1594,41 @@ AES_set_decrypt_key:
 		cmp	%rsi,%rdi
 	jne	.Linvert
 
-	.picmeup %r9
-	lea	AES_Td-.(%r9),%rdi
-	lea	AES_Te-AES_Td(%rdi),%r9
+	lea	.LAES_Te+2048+1024(%rip),%rax	# rcon
 
-	mov	%r8,%rsi
-	mov	240(%r8),%ecx		# pull number of rounds
-	sub	\$1,%ecx
+	mov	40(%rax),$mask80
+	mov	48(%rax),$maskfe
+	mov	56(%rax),$mask1b
+
+	mov	%r8,$key
+	sub	\$1,%r14d
 .align	4
 .Lpermute:
-		lea	16(%rsi),%rsi
+		lea	16($key),$key
+		mov	0($key),%rax
+		mov	8($key),%rcx
 ___
-		&deckey	(0,"%rsi","%r9","%rdi");
-		&deckey	(4,"%rsi","%r9","%rdi");
-		&deckey	(8,"%rsi","%r9","%rdi");
-		&deckey	(12,"%rsi","%r9","%rdi");
+		&dectransform ();
 $code.=<<___;
-		sub	\$1,%ecx
+		mov	%eax,0($key)
+		mov	%ebx,4($key)
+		mov	%ecx,8($key)
+		mov	%edx,12($key)
+		sub	\$1,%r14d
 	jnz	.Lpermute
 
 	xor	%rax,%rax
 .Labort:
-	mov	8(%rsp),%rbp
-	mov	16(%rsp),%rbx
-	add	\$24,%rsp
+	mov	8(%rsp),%r15
+	mov	16(%rsp),%r14
+	mov	24(%rsp),%r13
+	mov	32(%rsp),%r12
+	mov	40(%rsp),%rbp
+	mov	48(%rsp),%rbx
+	add	\$56,%rsp
+.Ldec_key_epilogue:
 	ret
-.size	AES_set_decrypt_key,.-AES_set_decrypt_key
+.size	private_AES_set_decrypt_key,.-private_AES_set_decrypt_key
 ___
 
 # void AES_cbc_encrypt (const void char *inp, unsigned char *out,
@@ -1046,47 +1637,62 @@ ___
 {
 # stack frame layout
 # -8(%rsp)		return address
-my $_rsp="0(%rsp)";		# saved %rsp
-my $_len="8(%rsp)";		# copy of 3rd parameter, length
-my $_key="16(%rsp)";		# copy of 4th parameter, key
-my $_ivp="24(%rsp)";		# copy of 5th parameter, ivp
-my $keyp="32(%rsp)";		# one to pass as $key
-my $ivec="40(%rsp)";		# ivec[16]
-my $aes_key="56(%rsp)";		# copy of aes_key
-my $mark="56+240(%rsp)";	# copy of aes_key->rounds
+my $keyp="0(%rsp)";		# one to pass as $key
+my $keyend="8(%rsp)";		# &(keyp->rd_key[4*keyp->rounds])
+my $_rsp="16(%rsp)";		# saved %rsp
+my $_inp="24(%rsp)";		# copy of 1st parameter, inp
+my $_out="32(%rsp)";		# copy of 2nd parameter, out
+my $_len="40(%rsp)";		# copy of 3rd parameter, length
+my $_key="48(%rsp)";		# copy of 4th parameter, key
+my $_ivp="56(%rsp)";		# copy of 5th parameter, ivp
+my $ivec="64(%rsp)";		# ivec[16]
+my $aes_key="80(%rsp)";		# copy of aes_key
+my $mark="80+240(%rsp)";	# copy of aes_key->rounds
 
 $code.=<<___;
 .globl	AES_cbc_encrypt
 .type	AES_cbc_encrypt,\@function,6
 .align	16
+.extern	OPENSSL_ia32cap_P
+.globl	asm_AES_cbc_encrypt
+.hidden	asm_AES_cbc_encrypt
+asm_AES_cbc_encrypt:
 AES_cbc_encrypt:
 	cmp	\$0,%rdx	# check length
-	je	.Lcbc_just_ret
+	je	.Lcbc_epilogue
+	pushfq
 	push	%rbx
 	push	%rbp
 	push	%r12
 	push	%r13
 	push	%r14
 	push	%r15
-	pushfq
+.Lcbc_prologue:
+
 	cld
 	mov	%r9d,%r9d	# clear upper half of enc
 
-	.picmeup $sbox
-.Lcbc_pic_point:
-
+	lea	.LAES_Te(%rip),$sbox
 	cmp	\$0,%r9
-	je	.LDECRYPT
-
-	lea	AES_Te-.Lcbc_pic_point($sbox),$sbox
+	jne	.Lcbc_picked_te
+	lea	.LAES_Td(%rip),$sbox
+.Lcbc_picked_te:
+
+	mov	OPENSSL_ia32cap_P(%rip),%r10d
+	cmp	\$$speed_limit,%rdx
+	jb	.Lcbc_slow_prologue
+	test	\$15,%rdx
+	jnz	.Lcbc_slow_prologue
+	bt	\$28,%r10d
+	jc	.Lcbc_slow_prologue
 
 	# allocate aligned stack frame...
-	lea	-64-248(%rsp),$key
+	lea	-88-248(%rsp),$key
 	and	\$-64,$key
 
-	# ... and make it doesn't alias with AES_Te modulo 4096
+	# ... and make sure it doesn't alias with AES_T[ed] modulo 4096
 	mov	$sbox,%r10
-	lea	2048($sbox),%r11
+	lea	2304($sbox),%r11
 	mov	$key,%r12
 	and	\$0xFFF,%r10	# s = $sbox&0xfff
 	and	\$0xFFF,%r11	# e = ($sbox+2048)&0xfff
@@ -1106,22 +1712,27 @@ AES_cbc_encrypt:
 .Lcbc_te_ok:
 
 	xchg	%rsp,$key
-	add	\$8,%rsp	# reserve for return address!
+	#add	\$8,%rsp	# reserve for return address!
 	mov	$key,$_rsp	# save %rsp
+.Lcbc_fast_body:
+	mov	%rdi,$_inp	# save copy of inp
+	mov	%rsi,$_out	# save copy of out
 	mov	%rdx,$_len	# save copy of len
 	mov	%rcx,$_key	# save copy of key
 	mov	%r8,$_ivp	# save copy of ivp
 	movl	\$0,$mark	# copy of aes_key->rounds = 0;
 	mov	%r8,%rbp	# rearrange input arguments
+	mov	%r9,%rbx
 	mov	%rsi,$out
 	mov	%rdi,$inp
 	mov	%rcx,$key
 
+	mov	240($key),%eax		# key->rounds
 	# do we copy key schedule to stack?
 	mov	$key,%r10
 	sub	$sbox,%r10
 	and	\$0xfff,%r10
-	cmp	\$2048,%r10
+	cmp	\$2304,%r10
 	jb	.Lcbc_do_ecopy
 	cmp	\$4096-248,%r10
 	jb	.Lcbc_skip_ecopy
@@ -1132,12 +1743,11 @@ AES_cbc_encrypt:
 		lea	$aes_key,$key
 		mov	\$240/8,%ecx
 		.long	0x90A548F3	# rep movsq
-		mov	(%rsi),%eax	# copy aes_key->rounds
-		mov	%eax,(%rdi)
+		mov	%eax,(%rdi)	# copy aes_key->rounds
 .Lcbc_skip_ecopy:
 	mov	$key,$keyp	# save key pointer
 
-	mov	\$16,%ecx
+	mov	\$18,%ecx
 .align	4
 .Lcbc_prefetch_te:
 		mov	0($sbox),%r10
@@ -1147,184 +1757,77 @@ AES_cbc_encrypt:
 		lea	128($sbox),$sbox
 		sub	\$1,%ecx
 	jnz	.Lcbc_prefetch_te
-	sub	\$2048,$sbox
+	lea	-2304($sbox),$sbox
 
-	test	\$-16,%rdx		# check upon length
-	mov	%rdx,%r10
+	cmp	\$0,%rbx
+	je	.LFAST_DECRYPT
+
+#----------------------------- ENCRYPT -----------------------------#
 	mov	0(%rbp),$s0		# load iv
 	mov	4(%rbp),$s1
 	mov	8(%rbp),$s2
 	mov	12(%rbp),$s3
-	jz	.Lcbc_enc_tail		# short input...
 
 .align	4
-.Lcbc_enc_loop:
+.Lcbc_fast_enc_loop:
 		xor	0($inp),$s0
 		xor	4($inp),$s1
 		xor	8($inp),$s2
 		xor	12($inp),$s3
-		mov	$inp,$ivec	# if ($verticalspin) save inp
-
 		mov	$keyp,$key	# restore key
+		mov	$inp,$_inp	# if ($verticalspin) save inp
+
 		call	_x86_64_AES_encrypt
 
-		mov	$ivec,$inp	# if ($verticalspin) restore inp
+		mov	$_inp,$inp	# if ($verticalspin) restore inp
+		mov	$_len,%r10
 		mov	$s0,0($out)
 		mov	$s1,4($out)
 		mov	$s2,8($out)
 		mov	$s3,12($out)
 
-		mov	$_len,%r10
 		lea	16($inp),$inp
 		lea	16($out),$out
 		sub	\$16,%r10
 		test	\$-16,%r10
 		mov	%r10,$_len
-	jnz	.Lcbc_enc_loop
-	test	\$15,%r10
-	jnz	.Lcbc_enc_tail
+	jnz	.Lcbc_fast_enc_loop
 	mov	$_ivp,%rbp	# restore ivp
 	mov	$s0,0(%rbp)	# save ivec
 	mov	$s1,4(%rbp)
 	mov	$s2,8(%rbp)
 	mov	$s3,12(%rbp)
 
-.align	4
-.Lcbc_cleanup:
-	cmpl	\$0,$mark	# was the key schedule copied?
-	lea	$aes_key,%rdi
-	je	.Lcbc_exit
-		mov	\$240/8,%ecx
-		xor	%rax,%rax
-		.long	0x90AB48F3	# rep stosq
-.Lcbc_exit:
-	mov	$_rsp,%rsp
-	popfq
-	pop	%r15
-	pop	%r14
-	pop	%r13
-	pop	%r12
-	pop	%rbp
-	pop	%rbx
-.Lcbc_just_ret:
-	ret
-.align	4
-.Lcbc_enc_tail:
-	mov	%rax,%r11
-	mov	%rcx,%r12
-	mov	%r10,%rcx
-	mov	$inp,%rsi
-	mov	$out,%rdi
-	.long	0xF689A4F3		# rep movsb
-	mov	\$16,%rcx		# zero tail
-	sub	%r10,%rcx
-	xor	%rax,%rax
-	.long	0xF689AAF3		# rep stosb
-	mov	$out,$inp		# this is not a mistake!
-	movq	\$16,$_len		# len=16
-	mov	%r11,%rax
-	mov	%r12,%rcx
-	jmp	.Lcbc_enc_loop		# one more spin...
+	jmp	.Lcbc_fast_cleanup
+
 #----------------------------- DECRYPT -----------------------------#
 .align	16
-.LDECRYPT:
-	lea	AES_Td-.Lcbc_pic_point($sbox),$sbox
-
-	# allocate aligned stack frame...
-	lea	-64-248(%rsp),$key
-	and	\$-64,$key
-
-	# ... and make it doesn't alias with AES_Td modulo 4096
-	mov	$sbox,%r10
-	lea	2304($sbox),%r11
-	mov	$key,%r12
-	and	\$0xFFF,%r10	# s = $sbox&0xfff
-	and	\$0xFFF,%r11	# e = ($sbox+2048+256)&0xfff
-	and	\$0xFFF,%r12	# p = %rsp&0xfff
-
-	cmp	%r11,%r12	# if (p=>e) %rsp =- (p-e);
-	jb	.Lcbc_td_break_out
-	sub	%r11,%r12
-	sub	%r12,$key
-	jmp	.Lcbc_td_ok
-.Lcbc_td_break_out:		# else %rsp -= (p-s)&0xfff + framesz
-	sub	%r10,%r12
-	and	\$0xFFF,%r12
-	add	\$320,%r12
-	sub	%r12,$key
-.align	4
-.Lcbc_td_ok:
-
-	xchg	%rsp,$key
-	add	\$8,%rsp	# reserve for return address!
-	mov	$key,$_rsp	# save %rsp
-	mov	%rdx,$_len	# save copy of len
-	mov	%rcx,$_key	# save copy of key
-	mov	%r8,$_ivp	# save copy of ivp
-	movl	\$0,$mark	# copy of aes_key->rounds = 0;
-	mov	%r8,%rbp	# rearrange input arguments
-	mov	%rsi,$out
-	mov	%rdi,$inp
-	mov	%rcx,$key
-
-	# do we copy key schedule to stack?
-	mov	$key,%r10
-	sub	$sbox,%r10
-	and	\$0xfff,%r10
-	cmp	\$2304,%r10
-	jb	.Lcbc_do_dcopy
-	cmp	\$4096-248,%r10
-	jb	.Lcbc_skip_dcopy
-.align	4
-.Lcbc_do_dcopy:
-		mov	$key,%rsi
-		lea	$aes_key,%rdi
-		lea	$aes_key,$key
-		mov	\$240/8,%ecx
-		.long	0x90A548F3	# rep movsq
-		mov	(%rsi),%eax	# copy aes_key->rounds
-		mov	%eax,(%rdi)
-.Lcbc_skip_dcopy:
-	mov	$key,$keyp	# save key pointer
-
-	mov	\$18,%ecx
-.align	4
-.Lcbc_prefetch_td:
-		mov	0($sbox),%r10
-		mov	32($sbox),%r11
-		mov	64($sbox),%r12
-		mov	96($sbox),%r13
-		lea	128($sbox),$sbox
-		sub	\$1,%ecx
-	jnz	.Lcbc_prefetch_td
-	sub	\$2304,$sbox
-
+.LFAST_DECRYPT:
 	cmp	$inp,$out
-	je	.Lcbc_dec_in_place
+	je	.Lcbc_fast_dec_in_place
 
 	mov	%rbp,$ivec
 .align	4
-.Lcbc_dec_loop:
-		mov	0($inp),$s0		# read input
+.Lcbc_fast_dec_loop:
+		mov	0($inp),$s0	# read input
 		mov	4($inp),$s1
 		mov	8($inp),$s2
 		mov	12($inp),$s3
-		mov	$inp,8+$ivec	# if ($verticalspin) save inp
-
 		mov	$keyp,$key	# restore key
+		mov	$inp,$_inp	# if ($verticalspin) save inp
+
 		call	_x86_64_AES_decrypt
 
 		mov	$ivec,%rbp	# load ivp
-		mov	8+$ivec,$inp	# if ($verticalspin) restore inp
+		mov	$_inp,$inp	# if ($verticalspin) restore inp
+		mov	$_len,%r10	# load len
 		xor	0(%rbp),$s0	# xor iv
 		xor	4(%rbp),$s1
 		xor	8(%rbp),$s2
 		xor	12(%rbp),$s3
 		mov	$inp,%rbp	# current input, next iv
 
-		mov	$_len,%r10	# load len
 		sub	\$16,%r10
-		jc	.Lcbc_dec_partial
 		mov	%r10,$_len	# update len
 		mov	%rbp,$ivec	# update ivp
 
@@ -1335,81 +1838,281 @@ AES_cbc_encrypt:
 
 		lea	16($inp),$inp
 		lea	16($out),$out
-	jnz	.Lcbc_dec_loop
-.Lcbc_dec_end:
+	jnz	.Lcbc_fast_dec_loop
 	mov	$_ivp,%r12		# load user ivp
 	mov	0(%rbp),%r10		# load iv
 	mov	8(%rbp),%r11
 	mov	%r10,0(%r12)		# copy back to user
 	mov	%r11,8(%r12)
-	jmp	.Lcbc_cleanup
-
-.align	4
-.Lcbc_dec_partial:
-	mov	$s0,0+$ivec		# dump output to stack
-	mov	$s1,4+$ivec
-	mov	$s2,8+$ivec
-	mov	$s3,12+$ivec
-	mov	$out,%rdi
-	lea	$ivec,%rsi
-	mov	\$16,%rcx
-	add	%r10,%rcx		# number of bytes to copy
-	.long	0xF689A4F3		# rep movsb
-	jmp	.Lcbc_dec_end
+	jmp	.Lcbc_fast_cleanup
 
 .align	16
-.Lcbc_dec_in_place:
+.Lcbc_fast_dec_in_place:
+	mov	0(%rbp),%r10		# copy iv to stack
+	mov	8(%rbp),%r11
+	mov	%r10,0+$ivec
+	mov	%r11,8+$ivec
+.align	4
+.Lcbc_fast_dec_in_place_loop:
 		mov	0($inp),$s0	# load input
 		mov	4($inp),$s1
 		mov	8($inp),$s2
 		mov	12($inp),$s3
+		mov	$keyp,$key	# restore key
+		mov	$inp,$_inp	# if ($verticalspin) save inp
 
-		mov	$inp,$ivec	# if ($verticalspin) save inp
-		mov	$keyp,$key
 		call	_x86_64_AES_decrypt
 
-		mov	$ivec,$inp	# if ($verticalspin) restore inp
-		mov	$_ivp,%rbp
-		xor	0(%rbp),$s0
-		xor	4(%rbp),$s1
-		xor	8(%rbp),$s2
-		xor	12(%rbp),$s3
+		mov	$_inp,$inp	# if ($verticalspin) restore inp
+		mov	$_len,%r10
+		xor	0+$ivec,$s0
+		xor	4+$ivec,$s1
+		xor	8+$ivec,$s2
+		xor	12+$ivec,$s3
 
-		mov	0($inp),%r10	# copy input to iv
-		mov	8($inp),%r11
-		mov	%r10,0(%rbp)
-		mov	%r11,8(%rbp)
+		mov	0($inp),%r11	# load input
+		mov	8($inp),%r12
+		sub	\$16,%r10
+		jz	.Lcbc_fast_dec_in_place_done
+
+		mov	%r11,0+$ivec	# copy input to iv
+		mov	%r12,8+$ivec
 
 		mov	$s0,0($out)	# save output [zaps input]
 		mov	$s1,4($out)
 		mov	$s2,8($out)
 		mov	$s3,12($out)
 
-		mov	$_len,%rcx
 		lea	16($inp),$inp
 		lea	16($out),$out
-		sub	\$16,%rcx
-		jc	.Lcbc_dec_in_place_partial
-		mov	%rcx,$_len
-	jnz	.Lcbc_dec_in_place
-	jmp	.Lcbc_cleanup
+		mov	%r10,$_len
+	jmp	.Lcbc_fast_dec_in_place_loop
+.Lcbc_fast_dec_in_place_done:
+	mov	$_ivp,%rdi
+	mov	%r11,0(%rdi)	# copy iv back to user
+	mov	%r12,8(%rdi)
+
+	mov	$s0,0($out)	# save output [zaps input]
+	mov	$s1,4($out)
+	mov	$s2,8($out)
+	mov	$s3,12($out)
 
 .align	4
-.Lcbc_dec_in_place_partial:
-	# one can argue if this is actually required
-	lea	($out,%rcx),%rdi
-	lea	(%rbp,%rcx),%rsi
-	neg	%rcx
-	.long	0xF689A4F3	# rep movsb	# restore tail
-	jmp	.Lcbc_cleanup
+.Lcbc_fast_cleanup:
+	cmpl	\$0,$mark	# was the key schedule copied?
+	lea	$aes_key,%rdi
+	je	.Lcbc_exit
+		mov	\$240/8,%ecx
+		xor	%rax,%rax
+		.long	0x90AB48F3	# rep stosq
+
+	jmp	.Lcbc_exit
+
+#--------------------------- SLOW ROUTINE ---------------------------#
+.align	16
+.Lcbc_slow_prologue:
+	# allocate aligned stack frame...
+	lea	-88(%rsp),%rbp
+	and	\$-64,%rbp
+	# ... just "above" key schedule
+	lea	-88-63(%rcx),%r10
+	sub	%rbp,%r10
+	neg	%r10
+	and	\$0x3c0,%r10
+	sub	%r10,%rbp
+
+	xchg	%rsp,%rbp
+	#add	\$8,%rsp	# reserve for return address!
+	mov	%rbp,$_rsp	# save %rsp
+.Lcbc_slow_body:
+	#mov	%rdi,$_inp	# save copy of inp
+	#mov	%rsi,$_out	# save copy of out
+	#mov	%rdx,$_len	# save copy of len
+	#mov	%rcx,$_key	# save copy of key
+	mov	%r8,$_ivp	# save copy of ivp
+	mov	%r8,%rbp	# rearrange input arguments
+	mov	%r9,%rbx
+	mov	%rsi,$out
+	mov	%rdi,$inp
+	mov	%rcx,$key
+	mov	%rdx,%r10
+
+	mov	240($key),%eax
+	mov	$key,$keyp	# save key pointer
+	shl	\$4,%eax
+	lea	($key,%rax),%rax
+	mov	%rax,$keyend
+
+	# pick Te4 copy which can't "overlap" with stack frame or key scdedule
+	lea	2048($sbox),$sbox
+	lea	768-8(%rsp),%rax
+	sub	$sbox,%rax
+	and	\$0x300,%rax
+	lea	($sbox,%rax),$sbox
+
+	cmp	\$0,%rbx
+	je	.LSLOW_DECRYPT
+
+#--------------------------- SLOW ENCRYPT ---------------------------#
+	test	\$-16,%r10		# check upon length
+	mov	0(%rbp),$s0		# load iv
+	mov	4(%rbp),$s1
+	mov	8(%rbp),$s2
+	mov	12(%rbp),$s3
+	jz	.Lcbc_slow_enc_tail	# short input...
+
+.align	4
+.Lcbc_slow_enc_loop:
+		xor	0($inp),$s0
+		xor	4($inp),$s1
+		xor	8($inp),$s2
+		xor	12($inp),$s3
+		mov	$keyp,$key	# restore key
+		mov	$inp,$_inp	# save inp
+		mov	$out,$_out	# save out
+		mov	%r10,$_len	# save len
+
+		call	_x86_64_AES_encrypt_compact
+
+		mov	$_inp,$inp	# restore inp
+		mov	$_out,$out	# restore out
+		mov	$_len,%r10	# restore len
+		mov	$s0,0($out)
+		mov	$s1,4($out)
+		mov	$s2,8($out)
+		mov	$s3,12($out)
+
+		lea	16($inp),$inp
+		lea	16($out),$out
+		sub	\$16,%r10
+		test	\$-16,%r10
+	jnz	.Lcbc_slow_enc_loop
+	test	\$15,%r10
+	jnz	.Lcbc_slow_enc_tail
+	mov	$_ivp,%rbp	# restore ivp
+	mov	$s0,0(%rbp)	# save ivec
+	mov	$s1,4(%rbp)
+	mov	$s2,8(%rbp)
+	mov	$s3,12(%rbp)
+
+	jmp	.Lcbc_exit
+
+.align	4
+.Lcbc_slow_enc_tail:
+	mov	%rax,%r11
+	mov	%rcx,%r12
+	mov	%r10,%rcx
+	mov	$inp,%rsi
+	mov	$out,%rdi
+	.long	0x9066A4F3		# rep movsb
+	mov	\$16,%rcx		# zero tail
+	sub	%r10,%rcx
+	xor	%rax,%rax
+	.long	0x9066AAF3		# rep stosb
+	mov	$out,$inp		# this is not a mistake!
+	mov	\$16,%r10		# len=16
+	mov	%r11,%rax
+	mov	%r12,%rcx
+	jmp	.Lcbc_slow_enc_loop	# one more spin...
+#--------------------------- SLOW DECRYPT ---------------------------#
+.align	16
+.LSLOW_DECRYPT:
+	shr	\$3,%rax
+	add	%rax,$sbox		# recall "magic" constants!
+
+	mov	0(%rbp),%r11		# copy iv to stack
+	mov	8(%rbp),%r12
+	mov	%r11,0+$ivec
+	mov	%r12,8+$ivec
+
+.align	4
+.Lcbc_slow_dec_loop:
+		mov	0($inp),$s0	# load input
+		mov	4($inp),$s1
+		mov	8($inp),$s2
+		mov	12($inp),$s3
+		mov	$keyp,$key	# restore key
+		mov	$inp,$_inp	# save inp
+		mov	$out,$_out	# save out
+		mov	%r10,$_len	# save len
+
+		call	_x86_64_AES_decrypt_compact
+
+		mov	$_inp,$inp	# restore inp
+		mov	$_out,$out	# restore out
+		mov	$_len,%r10
+		xor	0+$ivec,$s0
+		xor	4+$ivec,$s1
+		xor	8+$ivec,$s2
+		xor	12+$ivec,$s3
+
+		mov	0($inp),%r11	# load input
+		mov	8($inp),%r12
+		sub	\$16,%r10
+		jc	.Lcbc_slow_dec_partial
+		jz	.Lcbc_slow_dec_done
+
+		mov	%r11,0+$ivec	# copy input to iv
+		mov	%r12,8+$ivec
+
+		mov	$s0,0($out)	# save output [can zap input]
+		mov	$s1,4($out)
+		mov	$s2,8($out)
+		mov	$s3,12($out)
+
+		lea	16($inp),$inp
+		lea	16($out),$out
+	jmp	.Lcbc_slow_dec_loop
+.Lcbc_slow_dec_done:
+	mov	$_ivp,%rdi
+	mov	%r11,0(%rdi)		# copy iv back to user
+	mov	%r12,8(%rdi)
+
+	mov	$s0,0($out)		# save output [can zap input]
+	mov	$s1,4($out)
+	mov	$s2,8($out)
+	mov	$s3,12($out)
+
+	jmp	.Lcbc_exit
+
+.align	4
+.Lcbc_slow_dec_partial:
+	mov	$_ivp,%rdi
+	mov	%r11,0(%rdi)		# copy iv back to user
+	mov	%r12,8(%rdi)
+
+	mov	$s0,0+$ivec		# save output to stack
+	mov	$s1,4+$ivec
+	mov	$s2,8+$ivec
+	mov	$s3,12+$ivec
+
+	mov	$out,%rdi
+	lea	$ivec,%rsi
+	lea	16(%r10),%rcx
+	.long	0x9066A4F3	# rep movsb
+	jmp	.Lcbc_exit
+
+.align	16
+.Lcbc_exit:
+	mov	$_rsp,%rsi
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lcbc_popfq:
+	popfq
+.Lcbc_epilogue:
+	ret
 .size	AES_cbc_encrypt,.-AES_cbc_encrypt
 ___
 }
 
 $code.=<<___;
-.globl	AES_Te
 .align	64
-AES_Te:
+.LAES_Te:
 ___
 	&_data_word(0xa56363c6, 0x847c7cf8, 0x997777ee, 0x8d7b7bf6);
 	&_data_word(0x0df2f2ff, 0xbd6b6bd6, 0xb16f6fde, 0x54c5c591);
@@ -1475,16 +2178,149 @@ ___
 	&_data_word(0xdabfbf65, 0x31e6e6d7, 0xc6424284, 0xb86868d0);
 	&_data_word(0xc3414182, 0xb0999929, 0x772d2d5a, 0x110f0f1e);
 	&_data_word(0xcbb0b07b, 0xfc5454a8, 0xd6bbbb6d, 0x3a16162c);
+
+#Te4	# four copies of Te4 to choose from to avoid L1 aliasing
+	&data_byte(0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5);
+	&data_byte(0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76);
+	&data_byte(0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0);
+	&data_byte(0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0);
+	&data_byte(0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc);
+	&data_byte(0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15);
+	&data_byte(0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a);
+	&data_byte(0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75);
+	&data_byte(0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0);
+	&data_byte(0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84);
+	&data_byte(0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b);
+	&data_byte(0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf);
+	&data_byte(0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85);
+	&data_byte(0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8);
+	&data_byte(0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5);
+	&data_byte(0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2);
+	&data_byte(0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17);
+	&data_byte(0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73);
+	&data_byte(0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88);
+	&data_byte(0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb);
+	&data_byte(0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c);
+	&data_byte(0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79);
+	&data_byte(0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9);
+	&data_byte(0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08);
+	&data_byte(0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6);
+	&data_byte(0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a);
+	&data_byte(0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e);
+	&data_byte(0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e);
+	&data_byte(0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94);
+	&data_byte(0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf);
+	&data_byte(0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68);
+	&data_byte(0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16);
+
+	&data_byte(0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5);
+	&data_byte(0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76);
+	&data_byte(0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0);
+	&data_byte(0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0);
+	&data_byte(0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc);
+	&data_byte(0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15);
+	&data_byte(0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a);
+	&data_byte(0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75);
+	&data_byte(0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0);
+	&data_byte(0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84);
+	&data_byte(0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b);
+	&data_byte(0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf);
+	&data_byte(0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85);
+	&data_byte(0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8);
+	&data_byte(0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5);
+	&data_byte(0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2);
+	&data_byte(0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17);
+	&data_byte(0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73);
+	&data_byte(0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88);
+	&data_byte(0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb);
+	&data_byte(0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c);
+	&data_byte(0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79);
+	&data_byte(0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9);
+	&data_byte(0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08);
+	&data_byte(0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6);
+	&data_byte(0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a);
+	&data_byte(0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e);
+	&data_byte(0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e);
+	&data_byte(0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94);
+	&data_byte(0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf);
+	&data_byte(0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68);
+	&data_byte(0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16);
+
+	&data_byte(0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5);
+	&data_byte(0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76);
+	&data_byte(0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0);
+	&data_byte(0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0);
+	&data_byte(0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc);
+	&data_byte(0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15);
+	&data_byte(0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a);
+	&data_byte(0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75);
+	&data_byte(0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0);
+	&data_byte(0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84);
+	&data_byte(0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b);
+	&data_byte(0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf);
+	&data_byte(0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85);
+	&data_byte(0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8);
+	&data_byte(0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5);
+	&data_byte(0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2);
+	&data_byte(0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17);
+	&data_byte(0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73);
+	&data_byte(0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88);
+	&data_byte(0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb);
+	&data_byte(0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c);
+	&data_byte(0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79);
+	&data_byte(0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9);
+	&data_byte(0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08);
+	&data_byte(0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6);
+	&data_byte(0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a);
+	&data_byte(0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e);
+	&data_byte(0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e);
+	&data_byte(0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94);
+	&data_byte(0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf);
+	&data_byte(0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68);
+	&data_byte(0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16);
+
+	&data_byte(0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5);
+	&data_byte(0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76);
+	&data_byte(0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0);
+	&data_byte(0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0);
+	&data_byte(0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc);
+	&data_byte(0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15);
+	&data_byte(0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a);
+	&data_byte(0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75);
+	&data_byte(0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0);
+	&data_byte(0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84);
+	&data_byte(0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b);
+	&data_byte(0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf);
+	&data_byte(0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85);
+	&data_byte(0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8);
+	&data_byte(0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5);
+	&data_byte(0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2);
+	&data_byte(0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17);
+	&data_byte(0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73);
+	&data_byte(0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88);
+	&data_byte(0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb);
+	&data_byte(0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c);
+	&data_byte(0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79);
+	&data_byte(0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9);
+	&data_byte(0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08);
+	&data_byte(0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6);
+	&data_byte(0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a);
+	&data_byte(0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e);
+	&data_byte(0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e);
+	&data_byte(0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94);
+	&data_byte(0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf);
+	&data_byte(0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68);
+	&data_byte(0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16);
 #rcon:
 $code.=<<___;
 	.long	0x00000001, 0x00000002, 0x00000004, 0x00000008
 	.long	0x00000010, 0x00000020, 0x00000040, 0x00000080
-	.long	0x0000001b, 0x00000036, 0, 0, 0, 0, 0, 0
+	.long	0x0000001b, 0x00000036, 0x80808080, 0x80808080
+	.long	0xfefefefe, 0xfefefefe, 0x1b1b1b1b, 0x1b1b1b1b
 ___
 $code.=<<___;
-.globl	AES_Td
 .align	64
-AES_Td:
+.LAES_Td:
 ___
 	&_data_word(0x50a7f451, 0x5365417e, 0xc3a4171a, 0x965e273a);
 	&_data_word(0xcb6bab3b, 0xf1459d1f, 0xab58faac, 0x9303e34b);
@@ -1550,7 +2386,8 @@ ___
 	&_data_word(0x72c31d16, 0x0c25e2bc, 0x8b493c28, 0x41950dff);
 	&_data_word(0x7101a839, 0xdeb30c08, 0x9ce4b4d8, 0x90c15664);
 	&_data_word(0x6184cb7b, 0x70b632d5, 0x745c6c48, 0x4257b8d0);
-#Td4:
+
+#Td4:	# four copies of Td4 to choose from to avoid L1 aliasing
 	&data_byte(0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38);
 	&data_byte(0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb);
 	&data_byte(0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87);
@@ -1583,6 +2420,396 @@ ___
 	&data_byte(0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61);
 	&data_byte(0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26);
 	&data_byte(0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d);
+$code.=<<___;
+	.long	0x80808080, 0x80808080, 0xfefefefe, 0xfefefefe
+	.long	0x1b1b1b1b, 0x1b1b1b1b, 0, 0
+___
+	&data_byte(0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38);
+	&data_byte(0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb);
+	&data_byte(0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87);
+	&data_byte(0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb);
+	&data_byte(0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d);
+	&data_byte(0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e);
+	&data_byte(0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2);
+	&data_byte(0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25);
+	&data_byte(0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16);
+	&data_byte(0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92);
+	&data_byte(0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda);
+	&data_byte(0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84);
+	&data_byte(0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a);
+	&data_byte(0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06);
+	&data_byte(0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02);
+	&data_byte(0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b);
+	&data_byte(0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea);
+	&data_byte(0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73);
+	&data_byte(0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85);
+	&data_byte(0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e);
+	&data_byte(0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89);
+	&data_byte(0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b);
+	&data_byte(0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20);
+	&data_byte(0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4);
+	&data_byte(0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31);
+	&data_byte(0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f);
+	&data_byte(0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d);
+	&data_byte(0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef);
+	&data_byte(0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0);
+	&data_byte(0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61);
+	&data_byte(0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26);
+	&data_byte(0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d);
+$code.=<<___;
+	.long	0x80808080, 0x80808080, 0xfefefefe, 0xfefefefe
+	.long	0x1b1b1b1b, 0x1b1b1b1b, 0, 0
+___
+	&data_byte(0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38);
+	&data_byte(0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb);
+	&data_byte(0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87);
+	&data_byte(0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb);
+	&data_byte(0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d);
+	&data_byte(0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e);
+	&data_byte(0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2);
+	&data_byte(0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25);
+	&data_byte(0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16);
+	&data_byte(0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92);
+	&data_byte(0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda);
+	&data_byte(0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84);
+	&data_byte(0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a);
+	&data_byte(0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06);
+	&data_byte(0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02);
+	&data_byte(0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b);
+	&data_byte(0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea);
+	&data_byte(0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73);
+	&data_byte(0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85);
+	&data_byte(0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e);
+	&data_byte(0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89);
+	&data_byte(0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b);
+	&data_byte(0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20);
+	&data_byte(0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4);
+	&data_byte(0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31);
+	&data_byte(0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f);
+	&data_byte(0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d);
+	&data_byte(0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef);
+	&data_byte(0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0);
+	&data_byte(0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61);
+	&data_byte(0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26);
+	&data_byte(0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d);
+$code.=<<___;
+	.long	0x80808080, 0x80808080, 0xfefefefe, 0xfefefefe
+	.long	0x1b1b1b1b, 0x1b1b1b1b, 0, 0
+___
+	&data_byte(0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38);
+	&data_byte(0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb);
+	&data_byte(0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87);
+	&data_byte(0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb);
+	&data_byte(0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d);
+	&data_byte(0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e);
+	&data_byte(0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2);
+	&data_byte(0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25);
+	&data_byte(0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16);
+	&data_byte(0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92);
+	&data_byte(0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda);
+	&data_byte(0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84);
+	&data_byte(0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a);
+	&data_byte(0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06);
+	&data_byte(0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02);
+	&data_byte(0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b);
+	&data_byte(0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea);
+	&data_byte(0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73);
+	&data_byte(0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85);
+	&data_byte(0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e);
+	&data_byte(0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89);
+	&data_byte(0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b);
+	&data_byte(0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20);
+	&data_byte(0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4);
+	&data_byte(0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31);
+	&data_byte(0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f);
+	&data_byte(0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d);
+	&data_byte(0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef);
+	&data_byte(0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0);
+	&data_byte(0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61);
+	&data_byte(0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26);
+	&data_byte(0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d);
+$code.=<<___;
+	.long	0x80808080, 0x80808080, 0xfefefefe, 0xfefefefe
+	.long	0x1b1b1b1b, 0x1b1b1b1b, 0, 0
+.asciz  "AES for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+.align	64
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	block_se_handler,\@abi-omnipotent
+.align	16
+block_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lin_block_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lin_block_prologue
+
+	mov	24(%rax),%rax		# pull saved real stack pointer
+	lea	48(%rax),%rax		# adjust...
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	-40(%rax),%r14
+	mov	-48(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_block_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	jmp	.Lcommon_seh_exit
+.size	block_se_handler,.-block_se_handler
+
+.type	key_se_handler,\@abi-omnipotent
+.align	16
+key_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lin_key_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lin_key_prologue
+
+	lea	56(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	-40(%rax),%r14
+	mov	-48(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_key_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	jmp	.Lcommon_seh_exit
+.size	key_se_handler,.-key_se_handler
+
+.type	cbc_se_handler,\@abi-omnipotent
+.align	16
+cbc_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lcbc_prologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lcbc_prologue
+	jb	.Lin_cbc_prologue
+
+	lea	.Lcbc_fast_body(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lcbc_fast_body
+	jb	.Lin_cbc_frame_setup
+
+	lea	.Lcbc_slow_prologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lcbc_slow_prologue
+	jb	.Lin_cbc_body
+
+	lea	.Lcbc_slow_body(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lcbc_slow_body
+	jb	.Lin_cbc_frame_setup
+
+.Lin_cbc_body:
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lcbc_epilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lcbc_epilogue
+	jae	.Lin_cbc_prologue
+
+	lea	8(%rax),%rax
+
+	lea	.Lcbc_popfq(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lcbc_popfq
+	jae	.Lin_cbc_prologue
+
+	mov	`16-8`(%rax),%rax	# biased $_rsp
+	lea	56(%rax),%rax
+
+.Lin_cbc_frame_setup:
+	mov	-16(%rax),%rbx
+	mov	-24(%rax),%rbp
+	mov	-32(%rax),%r12
+	mov	-40(%rax),%r13
+	mov	-48(%rax),%r14
+	mov	-56(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_cbc_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+.Lcommon_seh_exit:
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$`1232/8`,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	cbc_se_handler,.-cbc_se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_AES_encrypt
+	.rva	.LSEH_end_AES_encrypt
+	.rva	.LSEH_info_AES_encrypt
+
+	.rva	.LSEH_begin_AES_decrypt
+	.rva	.LSEH_end_AES_decrypt
+	.rva	.LSEH_info_AES_decrypt
+
+	.rva	.LSEH_begin_private_AES_set_encrypt_key
+	.rva	.LSEH_end_private_AES_set_encrypt_key
+	.rva	.LSEH_info_private_AES_set_encrypt_key
+
+	.rva	.LSEH_begin_private_AES_set_decrypt_key
+	.rva	.LSEH_end_private_AES_set_decrypt_key
+	.rva	.LSEH_info_private_AES_set_decrypt_key
+
+	.rva	.LSEH_begin_AES_cbc_encrypt
+	.rva	.LSEH_end_AES_cbc_encrypt
+	.rva	.LSEH_info_AES_cbc_encrypt
+
+.section	.xdata
+.align	8
+.LSEH_info_AES_encrypt:
+	.byte	9,0,0,0
+	.rva	block_se_handler
+	.rva	.Lenc_prologue,.Lenc_epilogue	# HandlerData[]
+.LSEH_info_AES_decrypt:
+	.byte	9,0,0,0
+	.rva	block_se_handler
+	.rva	.Ldec_prologue,.Ldec_epilogue	# HandlerData[]
+.LSEH_info_private_AES_set_encrypt_key:
+	.byte	9,0,0,0
+	.rva	key_se_handler
+	.rva	.Lenc_key_prologue,.Lenc_key_epilogue	# HandlerData[]
+.LSEH_info_private_AES_set_decrypt_key:
+	.byte	9,0,0,0
+	.rva	key_se_handler
+	.rva	.Ldec_key_prologue,.Ldec_key_epilogue	# HandlerData[]
+.LSEH_info_AES_cbc_encrypt:
+	.byte	9,0,0,0
+	.rva	cbc_se_handler
+___
+}
 
 $code =~ s/\`([^\`]*)\`/eval($1)/gem;
 
diff --git a/crypto/aes/asm/aesni-sha1-x86_64.pl b/crypto/aes/asm/aesni-sha1-x86_64.pl
new file mode 100755
index 000000000000..c6f6b3334afe
--- /dev/null
+++ b/crypto/aes/asm/aesni-sha1-x86_64.pl
@@ -0,0 +1,1249 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# June 2011
+#
+# This is AESNI-CBC+SHA1 "stitch" implementation. The idea, as spelled
+# in http://download.intel.com/design/intarch/papers/323686.pdf, is
+# that since AESNI-CBC encrypt exhibit *very* low instruction-level
+# parallelism, interleaving it with another algorithm would allow to
+# utilize processor resources better and achieve better performance.
+# SHA1 instruction sequences(*) are taken from sha1-x86_64.pl and
+# AESNI code is weaved into it. Below are performance numbers in
+# cycles per processed byte, less is better, for standalone AESNI-CBC
+# encrypt, sum of the latter and standalone SHA1, and "stitched"
+# subroutine:
+#
+#		AES-128-CBC	+SHA1		stitch      gain
+# Westmere	3.77[+5.6]	9.37		6.65	    +41%
+# Sandy Bridge	5.05[+5.2(6.3)]	10.25(11.35)	6.16(7.08)  +67%(+60%)
+#
+#		AES-192-CBC
+# Westmere	4.51		10.11		6.97	    +45%
+# Sandy Bridge	6.05		11.25(12.35)	6.34(7.27)  +77%(+70%)
+#
+#		AES-256-CBC
+# Westmere	5.25		10.85		7.25	    +50%
+# Sandy Bridge	7.05		12.25(13.35)	7.06(7.70)  +74%(+73%)
+#
+# (*)	There are two code paths: SSSE3 and AVX. See sha1-568.pl for
+#	background information. Above numbers in parentheses are SSSE3
+#	results collected on AVX-capable CPU, i.e. apply on OSes that
+#	don't support AVX.
+#
+# Needless to mention that it makes no sense to implement "stitched"
+# *decrypt* subroutine. Because *both* AESNI-CBC decrypt and SHA1
+# fully utilize parallelism, so stitching would not give any gain
+# anyway. Well, there might be some, e.g. because of better cache
+# locality... For reference, here are performance results for
+# standalone AESNI-CBC decrypt:
+#
+#		AES-128-CBC	AES-192-CBC	AES-256-CBC
+# Westmere	1.31		1.55		1.80
+# Sandy Bridge	0.93		1.06		1.22
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+$avx=1 if (`$ENV{CC} -Wa,-v -c -o /dev/null -x assembler /dev/null 2>&1`
+		=~ /GNU assembler version ([2-9]\.[0-9]+)/ &&
+	   $1>=2.19);
+$avx=1 if (!$avx && $win64 && ($flavour =~ /nasm/ || $ENV{ASM} =~ /nasm/) &&
+	   `nasm -v 2>&1` =~ /NASM version ([2-9]\.[0-9]+)/ &&
+	   $1>=2.09);
+$avx=1 if (!$avx && $win64 && ($flavour =~ /masm/ || $ENV{ASM} =~ /ml64/) &&
+	   `ml64 2>&1` =~ /Version ([0-9]+)\./ &&
+	   $1>=10);
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+# void aesni_cbc_sha1_enc(const void *inp,
+#			void *out,
+#			size_t length,
+#			const AES_KEY *key,
+#			unsigned char *iv,
+#			SHA_CTX *ctx,
+#			const void *in0);
+
+$code.=<<___;
+.text
+.extern	OPENSSL_ia32cap_P
+
+.globl	aesni_cbc_sha1_enc
+.type	aesni_cbc_sha1_enc,\@abi-omnipotent
+.align	16
+aesni_cbc_sha1_enc:
+	# caller should check for SSSE3 and AES-NI bits
+	mov	OPENSSL_ia32cap_P+0(%rip),%r10d
+	mov	OPENSSL_ia32cap_P+4(%rip),%r11d
+___
+$code.=<<___ if ($avx);
+	and	\$`1<<28`,%r11d		# mask AVX bit
+	and	\$`1<<30`,%r10d		# mask "Intel CPU" bit
+	or	%r11d,%r10d
+	cmp	\$`1<<28|1<<30`,%r10d
+	je	aesni_cbc_sha1_enc_avx
+___
+$code.=<<___;
+	jmp	aesni_cbc_sha1_enc_ssse3
+	ret
+.size	aesni_cbc_sha1_enc,.-aesni_cbc_sha1_enc
+___
+
+my ($in0,$out,$len,$key,$ivp,$ctx,$inp)=("%rdi","%rsi","%rdx","%rcx","%r8","%r9","%r10");
+
+my $Xi=4;
+my @X=map("%xmm$_",(4..7,0..3));
+my @Tx=map("%xmm$_",(8..10));
+my @V=($A,$B,$C,$D,$E)=("%eax","%ebx","%ecx","%edx","%ebp");	# size optimization
+my @T=("%esi","%edi");
+my $j=0; my $jj=0; my $r=0; my $sn=0;
+my $K_XX_XX="%r11";
+my ($iv,$in,$rndkey0)=map("%xmm$_",(11..13));
+my @rndkey=("%xmm14","%xmm15");
+
+sub AUTOLOAD()		# thunk [simplified] 32-bit style perlasm
+{ my $opcode = $AUTOLOAD; $opcode =~ s/.*:://;
+  my $arg = pop;
+    $arg = "\$$arg" if ($arg*1 eq $arg);
+    $code .= "\t$opcode\t".join(',',$arg,reverse @_)."\n";
+}
+
+my $_rol=sub { &rol(@_) };
+my $_ror=sub { &ror(@_) };
+
+$code.=<<___;
+.type	aesni_cbc_sha1_enc_ssse3,\@function,6
+.align	16
+aesni_cbc_sha1_enc_ssse3:
+	mov	`($win64?56:8)`(%rsp),$inp	# load 7th argument
+	#shr	\$6,$len			# debugging artefact
+	#jz	.Lepilogue_ssse3		# debugging artefact
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	lea	`-104-($win64?10*16:0)`(%rsp),%rsp
+	#mov	$in0,$inp			# debugging artefact
+	#lea	64(%rsp),$ctx			# debugging artefact
+___
+$code.=<<___ if ($win64);
+	movaps	%xmm6,96+0(%rsp)
+	movaps	%xmm7,96+16(%rsp)
+	movaps	%xmm8,96+32(%rsp)
+	movaps	%xmm9,96+48(%rsp)
+	movaps	%xmm10,96+64(%rsp)
+	movaps	%xmm11,96+80(%rsp)
+	movaps	%xmm12,96+96(%rsp)
+	movaps	%xmm13,96+112(%rsp)
+	movaps	%xmm14,96+128(%rsp)
+	movaps	%xmm15,96+144(%rsp)
+.Lprologue_ssse3:
+___
+$code.=<<___;
+	mov	$in0,%r12			# reassign arguments
+	mov	$out,%r13
+	mov	$len,%r14
+	mov	$key,%r15
+	movdqu	($ivp),$iv			# load IV
+	mov	$ivp,88(%rsp)			# save $ivp
+___
+my ($in0,$out,$len,$key)=map("%r$_",(12..15));	# reassign arguments
+my $rounds="${ivp}d";
+$code.=<<___;
+	shl	\$6,$len
+	sub	$in0,$out
+	mov	240($key),$rounds
+	add	$inp,$len		# end of input
+
+	lea	K_XX_XX(%rip),$K_XX_XX
+	mov	0($ctx),$A		# load context
+	mov	4($ctx),$B
+	mov	8($ctx),$C
+	mov	12($ctx),$D
+	mov	$B,@T[0]		# magic seed
+	mov	16($ctx),$E
+
+	movdqa	64($K_XX_XX),@X[2]	# pbswap mask
+	movdqa	0($K_XX_XX),@Tx[1]	# K_00_19
+	movdqu	0($inp),@X[-4&7]	# load input to %xmm[0-3]
+	movdqu	16($inp),@X[-3&7]
+	movdqu	32($inp),@X[-2&7]
+	movdqu	48($inp),@X[-1&7]
+	pshufb	@X[2],@X[-4&7]		# byte swap
+	add	\$64,$inp
+	pshufb	@X[2],@X[-3&7]
+	pshufb	@X[2],@X[-2&7]
+	pshufb	@X[2],@X[-1&7]
+	paddd	@Tx[1],@X[-4&7]		# add K_00_19
+	paddd	@Tx[1],@X[-3&7]
+	paddd	@Tx[1],@X[-2&7]
+	movdqa	@X[-4&7],0(%rsp)	# X[]+K xfer to IALU
+	psubd	@Tx[1],@X[-4&7]		# restore X[]
+	movdqa	@X[-3&7],16(%rsp)
+	psubd	@Tx[1],@X[-3&7]
+	movdqa	@X[-2&7],32(%rsp)
+	psubd	@Tx[1],@X[-2&7]
+	movups	($key),$rndkey0		# $key[0]
+	movups	16($key),$rndkey[0]	# forward reference
+	jmp	.Loop_ssse3
+___
+
+my $aesenc=sub {
+  use integer;
+  my ($n,$k)=($r/10,$r%10);
+    if ($k==0) {
+      $code.=<<___;
+	movups		`16*$n`($in0),$in		# load input
+	xorps		$rndkey0,$in
+___
+      $code.=<<___ if ($n);
+	movups		$iv,`16*($n-1)`($out,$in0)	# write output
+___
+      $code.=<<___;
+	xorps		$in,$iv
+	aesenc		$rndkey[0],$iv
+	movups		`32+16*$k`($key),$rndkey[1]
+___
+    } elsif ($k==9) {
+      $sn++;
+      $code.=<<___;
+	cmp		\$11,$rounds
+	jb		.Laesenclast$sn
+	movups		`32+16*($k+0)`($key),$rndkey[1]
+	aesenc		$rndkey[0],$iv
+	movups		`32+16*($k+1)`($key),$rndkey[0]
+	aesenc		$rndkey[1],$iv
+	je		.Laesenclast$sn
+	movups		`32+16*($k+2)`($key),$rndkey[1]
+	aesenc		$rndkey[0],$iv
+	movups		`32+16*($k+3)`($key),$rndkey[0]
+	aesenc		$rndkey[1],$iv
+.Laesenclast$sn:
+	aesenclast	$rndkey[0],$iv
+	movups		16($key),$rndkey[1]		# forward reference
+___
+    } else {
+      $code.=<<___;
+	aesenc		$rndkey[0],$iv
+	movups		`32+16*$k`($key),$rndkey[1]
+___
+    }
+    $r++;	unshift(@rndkey,pop(@rndkey));
+};
+
+sub Xupdate_ssse3_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&movdqa	(@X[0],@X[-3&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(@Tx[0],@X[-1&7]);
+	&palignr(@X[0],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &paddd	(@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psrldq	(@Tx[0],4);		# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@Tx[0],@X[-2&7]);	# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@Tx[0]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&movdqa	(@Tx[2],@X[0]);
+	&movdqa	(@Tx[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslldq	(@Tx[2],12);		# "X[0]"<<96, extract one dword
+	&paddd	(@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@Tx[0],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(@Tx[1],@Tx[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@Tx[2],30);
+	&por	(@X[0],@Tx[0]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslld	(@Tx[1],2);
+	&pxor	(@X[0],@Tx[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(@Tx[2],eval(16*(($Xi)/5))."($K_XX_XX)");	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@Tx[1]);		# "X[0]"^=("X[0]">>96)<<<2
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xupdate_ssse3_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&movdqa	(@Tx[0],@X[-1&7])	if ($Xi==8);
+	 eval(shift(@insns));		# body_20_39
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"="X[-32]"^"X[-16]"
+	&palignr(@Tx[0],@X[-2&7],8);	# compose "X[-6]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&pxor	(@X[0],@X[-7&7]);	# "X[0]"^="X[-28]"
+	 eval(shift(@insns));
+	 eval(shift(@insns))	if (@insns[0] !~ /&ro[rl]/);
+	if ($Xi%5) {
+	  &movdqa	(@Tx[2],@Tx[1]);# "perpetuate" K_XX_XX...
+	} else {			# ... or load next one
+	  &movdqa	(@Tx[2],eval(16*($Xi/5))."($K_XX_XX)");
+	}
+	  &paddd	(@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@Tx[0]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&movdqa	(@Tx[0],@X[0]);
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pslld	(@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	&psrld	(@Tx[0],30);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&por	(@X[0],@Tx[0]);		# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &movdqa	(@Tx[1],@X[0])	if ($Xi<19);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xuplast_ssse3_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &paddd	(@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&cmp	($inp,$len);
+	&je	(".Ldone_ssse3");
+
+	unshift(@Tx,pop(@Tx));
+
+	&movdqa	(@X[2],"64($K_XX_XX)");		# pbswap mask
+	&movdqa	(@Tx[1],"0($K_XX_XX)");		# K_00_19
+	&movdqu	(@X[-4&7],"0($inp)");		# load input
+	&movdqu	(@X[-3&7],"16($inp)");
+	&movdqu	(@X[-2&7],"32($inp)");
+	&movdqu	(@X[-1&7],"48($inp)");
+	&pshufb	(@X[-4&7],@X[2]);		# byte swap
+	&add	($inp,64);
+
+  $Xi=0;
+}
+
+sub Xloop_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pshufb	(@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&paddd	(@X[($Xi-4)&7],@Tx[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(eval(16*$Xi)."(%rsp)",@X[($Xi-4)&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psubd	(@X[($Xi-4)&7],@Tx[1]);
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+sub body_00_19 () {
+  use integer;
+  my ($k,$n);
+  my @r=(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,eval(4*($j&15))."(%rsp)");',	# X[]+K xfer
+	'&xor	($c,$d);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&xor	($c,$d);',	# restore $c
+	'&xor	(@T[0],$d);',
+	'&add	($e,$a);',
+	'&$_ror	($b,$j?7:2);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'$j++; unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+	$n = scalar(@r);
+	$k = (($jj+1)*12/20)*20*$n/12;	# 12 aesencs per these 20 rounds
+	@r[$k%$n].='&$aesenc();'	if ($jj==$k/$n);
+	$jj++;
+    return @r;
+}
+
+sub body_20_39 () {
+  use integer;
+  my ($k,$n);
+  my @r=(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,eval(4*($j++&15))."(%rsp)");',	# X[]+K xfer
+	'&xor	(@T[0],$d);',	# ($b^$d)
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&xor	(@T[0],$c);',	# ($b^$d^$c)
+	'&add	($e,$a);',
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+	$n = scalar(@r);
+	$k = (($jj+1)*8/20)*20*$n/8;	# 8 aesencs per these 20 rounds
+	@r[$k%$n].='&$aesenc();'	if ($jj==$k/$n);
+	$jj++;
+    return @r;
+}
+
+sub body_40_59 () {
+  use integer;
+  my ($k,$n);
+  my @r=(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&mov	(@T[1],$c);',
+	'&xor	($c,$d);',
+	'&add	($e,eval(4*($j++&15))."(%rsp)");',	# X[]+K xfer
+	'&and	(@T[1],$d);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[1]);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&add	($e,@T[0]);',
+	'&xor	($c,$d);',	# restore $c
+	'&add	($e,$a);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+	$n = scalar(@r);
+	$k=(($jj+1)*12/20)*20*$n/12;	# 12 aesencs per these 20 rounds
+	@r[$k%$n].='&$aesenc();'	if ($jj==$k/$n);
+	$jj++;
+    return @r;
+}
+$code.=<<___;
+.align	16
+.Loop_ssse3:
+___
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xuplast_ssse3_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+				$saved_r=$r; @saved_rndkey=@rndkey;
+
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+
+$code.=<<___;
+	movups	$iv,48($out,$in0)		# write output
+	lea	64($in0),$in0
+
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	add	12($ctx),$D
+	mov	$A,0($ctx)
+	add	16($ctx),$E
+	mov	@T[0],4($ctx)
+	mov	@T[0],$B			# magic seed
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+	jmp	.Loop_ssse3
+
+.align	16
+.Ldone_ssse3:
+___
+				$jj=$j=$saved_j; @V=@saved_V;
+				$r=$saved_r;     @rndkey=@saved_rndkey;
+
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+
+$code.=<<___;
+	movups	$iv,48($out,$in0)		# write output
+	mov	88(%rsp),$ivp			# restore $ivp
+
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	mov	$A,0($ctx)
+	add	12($ctx),$D
+	mov	@T[0],4($ctx)
+	add	16($ctx),$E
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+	movups	$iv,($ivp)			# write IV
+___
+$code.=<<___ if ($win64);
+	movaps	96+0(%rsp),%xmm6
+	movaps	96+16(%rsp),%xmm7
+	movaps	96+32(%rsp),%xmm8
+	movaps	96+48(%rsp),%xmm9
+	movaps	96+64(%rsp),%xmm10
+	movaps	96+80(%rsp),%xmm11
+	movaps	96+96(%rsp),%xmm12
+	movaps	96+112(%rsp),%xmm13
+	movaps	96+128(%rsp),%xmm14
+	movaps	96+144(%rsp),%xmm15
+___
+$code.=<<___;
+	lea	`104+($win64?10*16:0)`(%rsp),%rsi
+	mov	0(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lepilogue_ssse3:
+	ret
+.size	aesni_cbc_sha1_enc_ssse3,.-aesni_cbc_sha1_enc_ssse3
+___
+
+$j=$jj=$r=$sn=0;
+
+if ($avx) {
+my ($in0,$out,$len,$key,$ivp,$ctx,$inp)=("%rdi","%rsi","%rdx","%rcx","%r8","%r9","%r10");
+
+my $Xi=4;
+my @X=map("%xmm$_",(4..7,0..3));
+my @Tx=map("%xmm$_",(8..10));
+my @V=($A,$B,$C,$D,$E)=("%eax","%ebx","%ecx","%edx","%ebp");	# size optimization
+my @T=("%esi","%edi");
+
+my $_rol=sub { &shld(@_[0],@_) };
+my $_ror=sub { &shrd(@_[0],@_) };
+
+$code.=<<___;
+.type	aesni_cbc_sha1_enc_avx,\@function,6
+.align	16
+aesni_cbc_sha1_enc_avx:
+	mov	`($win64?56:8)`(%rsp),$inp	# load 7th argument
+	#shr	\$6,$len			# debugging artefact
+	#jz	.Lepilogue_avx			# debugging artefact
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	lea	`-104-($win64?10*16:0)`(%rsp),%rsp
+	#mov	$in0,$inp			# debugging artefact
+	#lea	64(%rsp),$ctx			# debugging artefact
+___
+$code.=<<___ if ($win64);
+	movaps	%xmm6,96+0(%rsp)
+	movaps	%xmm7,96+16(%rsp)
+	movaps	%xmm8,96+32(%rsp)
+	movaps	%xmm9,96+48(%rsp)
+	movaps	%xmm10,96+64(%rsp)
+	movaps	%xmm11,96+80(%rsp)
+	movaps	%xmm12,96+96(%rsp)
+	movaps	%xmm13,96+112(%rsp)
+	movaps	%xmm14,96+128(%rsp)
+	movaps	%xmm15,96+144(%rsp)
+.Lprologue_avx:
+___
+$code.=<<___;
+	vzeroall
+	mov	$in0,%r12			# reassign arguments
+	mov	$out,%r13
+	mov	$len,%r14
+	mov	$key,%r15
+	vmovdqu	($ivp),$iv			# load IV
+	mov	$ivp,88(%rsp)			# save $ivp
+___
+my ($in0,$out,$len,$key)=map("%r$_",(12..15));	# reassign arguments
+my $rounds="${ivp}d";
+$code.=<<___;
+	shl	\$6,$len
+	sub	$in0,$out
+	mov	240($key),$rounds
+	add	\$112,$key		# size optimization
+	add	$inp,$len		# end of input
+
+	lea	K_XX_XX(%rip),$K_XX_XX
+	mov	0($ctx),$A		# load context
+	mov	4($ctx),$B
+	mov	8($ctx),$C
+	mov	12($ctx),$D
+	mov	$B,@T[0]		# magic seed
+	mov	16($ctx),$E
+
+	vmovdqa	64($K_XX_XX),@X[2]	# pbswap mask
+	vmovdqa	0($K_XX_XX),@Tx[1]	# K_00_19
+	vmovdqu	0($inp),@X[-4&7]	# load input to %xmm[0-3]
+	vmovdqu	16($inp),@X[-3&7]
+	vmovdqu	32($inp),@X[-2&7]
+	vmovdqu	48($inp),@X[-1&7]
+	vpshufb	@X[2],@X[-4&7],@X[-4&7]	# byte swap
+	add	\$64,$inp
+	vpshufb	@X[2],@X[-3&7],@X[-3&7]
+	vpshufb	@X[2],@X[-2&7],@X[-2&7]
+	vpshufb	@X[2],@X[-1&7],@X[-1&7]
+	vpaddd	@Tx[1],@X[-4&7],@X[0]	# add K_00_19
+	vpaddd	@Tx[1],@X[-3&7],@X[1]
+	vpaddd	@Tx[1],@X[-2&7],@X[2]
+	vmovdqa	@X[0],0(%rsp)		# X[]+K xfer to IALU
+	vmovdqa	@X[1],16(%rsp)
+	vmovdqa	@X[2],32(%rsp)
+	vmovups	-112($key),$rndkey0	# $key[0]
+	vmovups	16-112($key),$rndkey[0]	# forward reference
+	jmp	.Loop_avx
+___
+
+my $aesenc=sub {
+  use integer;
+  my ($n,$k)=($r/10,$r%10);
+    if ($k==0) {
+      $code.=<<___;
+	vmovups		`16*$n`($in0),$in		# load input
+	vxorps		$rndkey0,$in,$in
+___
+      $code.=<<___ if ($n);
+	vmovups		$iv,`16*($n-1)`($out,$in0)	# write output
+___
+      $code.=<<___;
+	vxorps		$in,$iv,$iv
+	vaesenc		$rndkey[0],$iv,$iv
+	vmovups		`32+16*$k-112`($key),$rndkey[1]
+___
+    } elsif ($k==9) {
+      $sn++;
+      $code.=<<___;
+	cmp		\$11,$rounds
+	jb		.Lvaesenclast$sn
+	vaesenc		$rndkey[0],$iv,$iv
+	vmovups		`32+16*($k+0)-112`($key),$rndkey[1]
+	vaesenc		$rndkey[1],$iv,$iv
+	vmovups		`32+16*($k+1)-112`($key),$rndkey[0]
+	je		.Lvaesenclast$sn
+	vaesenc		$rndkey[0],$iv,$iv
+	vmovups		`32+16*($k+2)-112`($key),$rndkey[1]
+	vaesenc		$rndkey[1],$iv,$iv
+	vmovups		`32+16*($k+3)-112`($key),$rndkey[0]
+.Lvaesenclast$sn:
+	vaesenclast	$rndkey[0],$iv,$iv
+	vmovups		16-112($key),$rndkey[1]		# forward reference
+___
+    } else {
+      $code.=<<___;
+	vaesenc		$rndkey[0],$iv,$iv
+	vmovups		`32+16*$k-112`($key),$rndkey[1]
+___
+    }
+    $r++;	unshift(@rndkey,pop(@rndkey));
+};
+
+sub Xupdate_avx_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpalignr(@X[0],@X[-3&7],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &vpaddd	(@Tx[1],@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpsrldq(@Tx[0],@X[-1&7],4);	# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpxor	(@X[0],@X[0],@X[-4&7]);		# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@Tx[0],@Tx[0],@X[-2&7]);	# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@Tx[0]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@Tx[0],@X[0],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslldq(@Tx[2],@X[0],12);		# "X[0]"<<96, extract one dword
+	&vpaddd	(@X[0],@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@Tx[1],@Tx[2],30);
+	&vpor	(@X[0],@X[0],@Tx[0]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslld	(@Tx[2],@Tx[2],2);
+	&vpxor	(@X[0],@X[0],@Tx[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@Tx[2]);		# "X[0]"^=("X[0]">>96)<<<2
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(@Tx[2],eval(16*(($Xi)/5))."($K_XX_XX)");	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xupdate_avx_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&vpalignr(@Tx[0],@X[-1&7],@X[-2&7],8);	# compose "X[-6]"
+	&vpxor	(@X[0],@X[0],@X[-4&7]);		# "X[0]"="X[-32]"^"X[-16]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpxor	(@X[0],@X[0],@X[-7&7]);		# "X[0]"^="X[-28]"
+	 eval(shift(@insns));
+	 eval(shift(@insns))	if (@insns[0] !~ /&ro[rl]/);
+	if ($Xi%5) {
+	  &vmovdqa	(@Tx[2],@Tx[1]);# "perpetuate" K_XX_XX...
+	} else {			# ... or load next one
+	  &vmovdqa	(@Tx[2],eval(16*($Xi/5))."($K_XX_XX)");
+	}
+	  &vpaddd	(@Tx[1],@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@Tx[0]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpsrld	(@Tx[0],@X[0],30);
+	  &vmovdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpslld	(@X[0],@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpor	(@X[0],@X[0],@Tx[0]);		# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &vmovdqa	(@Tx[1],@X[0])	if ($Xi<19);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xuplast_avx_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &vpaddd	(@Tx[1],@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&cmp	($inp,$len);
+	&je	(".Ldone_avx");
+
+	unshift(@Tx,pop(@Tx));
+
+	&vmovdqa(@X[2],"64($K_XX_XX)");		# pbswap mask
+	&vmovdqa(@Tx[1],"0($K_XX_XX)");		# K_00_19
+	&vmovdqu(@X[-4&7],"0($inp)");		# load input
+	&vmovdqu(@X[-3&7],"16($inp)");
+	&vmovdqu(@X[-2&7],"32($inp)");
+	&vmovdqu(@X[-1&7],"48($inp)");
+	&vpshufb(@X[-4&7],@X[-4&7],@X[2]);	# byte swap
+	&add	($inp,64);
+
+  $Xi=0;
+}
+
+sub Xloop_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpshufb(@X[($Xi-3)&7],@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpaddd	(@X[$Xi&7],@X[($Xi-4)&7],@Tx[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vmovdqa(eval(16*$Xi)."(%rsp)",@X[$Xi&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+$code.=<<___;
+.align	16
+.Loop_avx:
+___
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xuplast_avx_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+				$saved_r=$r; @saved_rndkey=@rndkey;
+
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+
+$code.=<<___;
+	vmovups	$iv,48($out,$in0)		# write output
+	lea	64($in0),$in0
+
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	add	12($ctx),$D
+	mov	$A,0($ctx)
+	add	16($ctx),$E
+	mov	@T[0],4($ctx)
+	mov	@T[0],$B			# magic seed
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+	jmp	.Loop_avx
+
+.align	16
+.Ldone_avx:
+___
+				$jj=$j=$saved_j; @V=@saved_V;
+				$r=$saved_r;     @rndkey=@saved_rndkey;
+
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+
+$code.=<<___;
+	vmovups	$iv,48($out,$in0)		# write output
+	mov	88(%rsp),$ivp			# restore $ivp
+
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	mov	$A,0($ctx)
+	add	12($ctx),$D
+	mov	@T[0],4($ctx)
+	add	16($ctx),$E
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+	vmovups	$iv,($ivp)			# write IV
+	vzeroall
+___
+$code.=<<___ if ($win64);
+	movaps	96+0(%rsp),%xmm6
+	movaps	96+16(%rsp),%xmm7
+	movaps	96+32(%rsp),%xmm8
+	movaps	96+48(%rsp),%xmm9
+	movaps	96+64(%rsp),%xmm10
+	movaps	96+80(%rsp),%xmm11
+	movaps	96+96(%rsp),%xmm12
+	movaps	96+112(%rsp),%xmm13
+	movaps	96+128(%rsp),%xmm14
+	movaps	96+144(%rsp),%xmm15
+___
+$code.=<<___;
+	lea	`104+($win64?10*16:0)`(%rsp),%rsi
+	mov	0(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lepilogue_avx:
+	ret
+.size	aesni_cbc_sha1_enc_avx,.-aesni_cbc_sha1_enc_avx
+___
+}
+$code.=<<___;
+.align	64
+K_XX_XX:
+.long	0x5a827999,0x5a827999,0x5a827999,0x5a827999	# K_00_19
+.long	0x6ed9eba1,0x6ed9eba1,0x6ed9eba1,0x6ed9eba1	# K_20_39
+.long	0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc	# K_40_59
+.long	0xca62c1d6,0xca62c1d6,0xca62c1d6,0xca62c1d6	# K_60_79
+.long	0x00010203,0x04050607,0x08090a0b,0x0c0d0e0f	# pbswap mask
+
+.asciz	"AESNI-CBC+SHA1 stitch for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+.align	64
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	ssse3_handler,\@abi-omnipotent
+.align	16
+ssse3_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lcommon_seh_tail
+
+	lea	96(%rax),%rsi
+	lea	512($context),%rdi	# &context.Xmm6
+	mov	\$20,%ecx
+	.long	0xa548f3fc		# cld; rep movsq
+	lea	`104+10*16`(%rax),%rax	# adjust stack pointer
+
+	mov	0(%rax),%r15
+	mov	8(%rax),%r14
+	mov	16(%rax),%r13
+	mov	24(%rax),%r12
+	mov	32(%rax),%rbp
+	mov	40(%rax),%rbx
+	lea	48(%rax),%rax
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lcommon_seh_tail:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	ssse3_handler,.-ssse3_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_aesni_cbc_sha1_enc_ssse3
+	.rva	.LSEH_end_aesni_cbc_sha1_enc_ssse3
+	.rva	.LSEH_info_aesni_cbc_sha1_enc_ssse3
+___
+$code.=<<___ if ($avx);
+	.rva	.LSEH_begin_aesni_cbc_sha1_enc_avx
+	.rva	.LSEH_end_aesni_cbc_sha1_enc_avx
+	.rva	.LSEH_info_aesni_cbc_sha1_enc_avx
+___
+$code.=<<___;
+.section	.xdata
+.align	8
+.LSEH_info_aesni_cbc_sha1_enc_ssse3:
+	.byte	9,0,0,0
+	.rva	ssse3_handler
+	.rva	.Lprologue_ssse3,.Lepilogue_ssse3	# HandlerData[]
+___
+$code.=<<___ if ($avx);
+.LSEH_info_aesni_cbc_sha1_enc_avx:
+	.byte	9,0,0,0
+	.rva	ssse3_handler
+	.rva	.Lprologue_avx,.Lepilogue_avx		# HandlerData[]
+___
+}
+
+####################################################################
+sub rex {
+  local *opcode=shift;
+  my ($dst,$src)=@_;
+  my $rex=0;
+
+    $rex|=0x04			if($dst>=8);
+    $rex|=0x01			if($src>=8);
+    push @opcode,$rex|0x40	if($rex);
+}
+
+sub aesni {
+  my $line=shift;
+  my @opcode=(0x66);
+
+    if ($line=~/(aes[a-z]+)\s+%xmm([0-9]+),\s*%xmm([0-9]+)/) {
+	my %opcodelet = (
+		"aesenc" => 0xdc,	"aesenclast" => 0xdd
+	);
+	return undef if (!defined($opcodelet{$1}));
+	rex(\@opcode,$3,$2);
+	push @opcode,0x0f,0x38,$opcodelet{$1};
+	push @opcode,0xc0|($2&7)|(($3&7)<<3);	# ModR/M
+	return ".byte\t".join(',',@opcode);
+    }
+    return $line;
+}
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+$code =~ s/\b(aes.*%xmm[0-9]+).*$/aesni($1)/gem;
+
+print $code;
+close STDOUT;
diff --git a/crypto/aes/asm/aesni-x86.pl b/crypto/aes/asm/aesni-x86.pl
new file mode 100755
index 000000000000..3dc345b585f6
--- /dev/null
+++ b/crypto/aes/asm/aesni-x86.pl
@@ -0,0 +1,2189 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# This module implements support for Intel AES-NI extension. In
+# OpenSSL context it's used with Intel engine, but can also be used as
+# drop-in replacement for crypto/aes/asm/aes-586.pl [see below for
+# details].
+#
+# Performance.
+#
+# To start with see corresponding paragraph in aesni-x86_64.pl...
+# Instead of filling table similar to one found there I've chosen to
+# summarize *comparison* results for raw ECB, CTR and CBC benchmarks.
+# The simplified table below represents 32-bit performance relative
+# to 64-bit one in every given point. Ratios vary for different
+# encryption modes, therefore interval values.
+#
+#	16-byte     64-byte     256-byte    1-KB        8-KB
+#	53-67%      67-84%      91-94%      95-98%      97-99.5%
+#
+# Lower ratios for smaller block sizes are perfectly understandable,
+# because function call overhead is higher in 32-bit mode. Largest
+# 8-KB block performance is virtually same: 32-bit code is less than
+# 1% slower for ECB, CBC and CCM, and ~3% slower otherwise.
+
+# January 2011
+#
+# See aesni-x86_64.pl for details. Unlike x86_64 version this module
+# interleaves at most 6 aes[enc|dec] instructions, because there are
+# not enough registers for 8x interleave [which should be optimal for
+# Sandy Bridge]. Actually, performance results for 6x interleave
+# factor presented in aesni-x86_64.pl (except for CTR) are for this
+# module.
+
+# April 2011
+#
+# Add aesni_xts_[en|de]crypt. Westmere spends 1.50 cycles processing
+# one byte out of 8KB with 128-bit key, Sandy Bridge - 1.09.
+
+$PREFIX="aesni";	# if $PREFIX is set to "AES", the script
+			# generates drop-in replacement for
+			# crypto/aes/asm/aes-586.pl:-)
+$inline=1;		# inline _aesni_[en|de]crypt
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],$0);
+
+if ($PREFIX eq "aesni")	{ $movekey=*movups; }
+else			{ $movekey=*movups; }
+
+$len="eax";
+$rounds="ecx";
+$key="edx";
+$inp="esi";
+$out="edi";
+$rounds_="ebx";	# backup copy for $rounds
+$key_="ebp";	# backup copy for $key
+
+$rndkey0="xmm0";
+$rndkey1="xmm1";
+$inout0="xmm2";
+$inout1="xmm3";
+$inout2="xmm4";
+$inout3="xmm5";	$in1="xmm5";
+$inout4="xmm6";	$in0="xmm6";
+$inout5="xmm7";	$ivec="xmm7";
+
+# AESNI extenstion
+sub aeskeygenassist
+{ my($dst,$src,$imm)=@_;
+    if ("$dst:$src" =~ /xmm([0-7]):xmm([0-7])/)
+    {	&data_byte(0x66,0x0f,0x3a,0xdf,0xc0|($1<<3)|$2,$imm);	}
+}
+sub aescommon
+{ my($opcodelet,$dst,$src)=@_;
+    if ("$dst:$src" =~ /xmm([0-7]):xmm([0-7])/)
+    {	&data_byte(0x66,0x0f,0x38,$opcodelet,0xc0|($1<<3)|$2);}
+}
+sub aesimc	{ aescommon(0xdb,@_); }
+sub aesenc	{ aescommon(0xdc,@_); }
+sub aesenclast	{ aescommon(0xdd,@_); }
+sub aesdec	{ aescommon(0xde,@_); }
+sub aesdeclast	{ aescommon(0xdf,@_); }
+
+# Inline version of internal aesni_[en|de]crypt1
+{ my $sn;
+sub aesni_inline_generate1
+{ my ($p,$inout,$ivec)=@_; $inout=$inout0 if (!defined($inout));
+  $sn++;
+
+    &$movekey		($rndkey0,&QWP(0,$key));
+    &$movekey		($rndkey1,&QWP(16,$key));
+    &xorps		($ivec,$rndkey0)	if (defined($ivec));
+    &lea		($key,&DWP(32,$key));
+    &xorps		($inout,$ivec)		if (defined($ivec));
+    &xorps		($inout,$rndkey0)	if (!defined($ivec));
+    &set_label("${p}1_loop_$sn");
+	eval"&aes${p}	($inout,$rndkey1)";
+	&dec		($rounds);
+	&$movekey	($rndkey1,&QWP(0,$key));
+	&lea		($key,&DWP(16,$key));
+    &jnz		(&label("${p}1_loop_$sn"));
+    eval"&aes${p}last	($inout,$rndkey1)";
+}}
+
+sub aesni_generate1	# fully unrolled loop
+{ my ($p,$inout)=@_; $inout=$inout0 if (!defined($inout));
+
+    &function_begin_B("_aesni_${p}rypt1");
+	&movups		($rndkey0,&QWP(0,$key));
+	&$movekey	($rndkey1,&QWP(0x10,$key));
+	&xorps		($inout,$rndkey0);
+	&$movekey	($rndkey0,&QWP(0x20,$key));
+	&lea		($key,&DWP(0x30,$key));
+	&cmp		($rounds,11);
+	&jb		(&label("${p}128"));
+	&lea		($key,&DWP(0x20,$key));
+	&je		(&label("${p}192"));
+	&lea		($key,&DWP(0x20,$key));
+	eval"&aes${p}	($inout,$rndkey1)";
+	&$movekey	($rndkey1,&QWP(-0x40,$key));
+	eval"&aes${p}	($inout,$rndkey0)";
+	&$movekey	($rndkey0,&QWP(-0x30,$key));
+    &set_label("${p}192");
+	eval"&aes${p}	($inout,$rndkey1)";
+	&$movekey	($rndkey1,&QWP(-0x20,$key));
+	eval"&aes${p}	($inout,$rndkey0)";
+	&$movekey	($rndkey0,&QWP(-0x10,$key));
+    &set_label("${p}128");
+	eval"&aes${p}	($inout,$rndkey1)";
+	&$movekey	($rndkey1,&QWP(0,$key));
+	eval"&aes${p}	($inout,$rndkey0)";
+	&$movekey	($rndkey0,&QWP(0x10,$key));
+	eval"&aes${p}	($inout,$rndkey1)";
+	&$movekey	($rndkey1,&QWP(0x20,$key));
+	eval"&aes${p}	($inout,$rndkey0)";
+	&$movekey	($rndkey0,&QWP(0x30,$key));
+	eval"&aes${p}	($inout,$rndkey1)";
+	&$movekey	($rndkey1,&QWP(0x40,$key));
+	eval"&aes${p}	($inout,$rndkey0)";
+	&$movekey	($rndkey0,&QWP(0x50,$key));
+	eval"&aes${p}	($inout,$rndkey1)";
+	&$movekey	($rndkey1,&QWP(0x60,$key));
+	eval"&aes${p}	($inout,$rndkey0)";
+	&$movekey	($rndkey0,&QWP(0x70,$key));
+	eval"&aes${p}	($inout,$rndkey1)";
+    eval"&aes${p}last	($inout,$rndkey0)";
+    &ret();
+    &function_end_B("_aesni_${p}rypt1");
+}
+
+# void $PREFIX_encrypt (const void *inp,void *out,const AES_KEY *key);
+&aesni_generate1("enc") if (!$inline);
+&function_begin_B("${PREFIX}_encrypt");
+	&mov	("eax",&wparam(0));
+	&mov	($key,&wparam(2));
+	&movups	($inout0,&QWP(0,"eax"));
+	&mov	($rounds,&DWP(240,$key));
+	&mov	("eax",&wparam(1));
+	if ($inline)
+	{   &aesni_inline_generate1("enc");	}
+	else
+	{   &call	("_aesni_encrypt1");	}
+	&movups	(&QWP(0,"eax"),$inout0);
+	&ret	();
+&function_end_B("${PREFIX}_encrypt");
+
+# void $PREFIX_decrypt (const void *inp,void *out,const AES_KEY *key);
+&aesni_generate1("dec") if(!$inline);
+&function_begin_B("${PREFIX}_decrypt");
+	&mov	("eax",&wparam(0));
+	&mov	($key,&wparam(2));
+	&movups	($inout0,&QWP(0,"eax"));
+	&mov	($rounds,&DWP(240,$key));
+	&mov	("eax",&wparam(1));
+	if ($inline)
+	{   &aesni_inline_generate1("dec");	}
+	else
+	{   &call	("_aesni_decrypt1");	}
+	&movups	(&QWP(0,"eax"),$inout0);
+	&ret	();
+&function_end_B("${PREFIX}_decrypt");
+
+# _aesni_[en|de]cryptN are private interfaces, N denotes interleave
+# factor. Why 3x subroutine were originally used in loops? Even though
+# aes[enc|dec] latency was originally 6, it could be scheduled only
+# every *2nd* cycle. Thus 3x interleave was the one providing optimal
+# utilization, i.e. when subroutine's throughput is virtually same as
+# of non-interleaved subroutine [for number of input blocks up to 3].
+# This is why it makes no sense to implement 2x subroutine.
+# aes[enc|dec] latency in next processor generation is 8, but the
+# instructions can be scheduled every cycle. Optimal interleave for
+# new processor is therefore 8x, but it's unfeasible to accommodate it
+# in XMM registers addreassable in 32-bit mode and therefore 6x is
+# used instead...
+
+sub aesni_generate3
+{ my $p=shift;
+
+    &function_begin_B("_aesni_${p}rypt3");
+	&$movekey	($rndkey0,&QWP(0,$key));
+	&shr		($rounds,1);
+	&$movekey	($rndkey1,&QWP(16,$key));
+	&lea		($key,&DWP(32,$key));
+	&xorps		($inout0,$rndkey0);
+	&pxor		($inout1,$rndkey0);
+	&pxor		($inout2,$rndkey0);
+	&$movekey	($rndkey0,&QWP(0,$key));
+
+    &set_label("${p}3_loop");
+	eval"&aes${p}	($inout0,$rndkey1)";
+	eval"&aes${p}	($inout1,$rndkey1)";
+	&dec		($rounds);
+	eval"&aes${p}	($inout2,$rndkey1)";
+	&$movekey	($rndkey1,&QWP(16,$key));
+	eval"&aes${p}	($inout0,$rndkey0)";
+	eval"&aes${p}	($inout1,$rndkey0)";
+	&lea		($key,&DWP(32,$key));
+	eval"&aes${p}	($inout2,$rndkey0)";
+	&$movekey	($rndkey0,&QWP(0,$key));
+	&jnz		(&label("${p}3_loop"));
+    eval"&aes${p}	($inout0,$rndkey1)";
+    eval"&aes${p}	($inout1,$rndkey1)";
+    eval"&aes${p}	($inout2,$rndkey1)";
+    eval"&aes${p}last	($inout0,$rndkey0)";
+    eval"&aes${p}last	($inout1,$rndkey0)";
+    eval"&aes${p}last	($inout2,$rndkey0)";
+    &ret();
+    &function_end_B("_aesni_${p}rypt3");
+}
+
+# 4x interleave is implemented to improve small block performance,
+# most notably [and naturally] 4 block by ~30%. One can argue that one
+# should have implemented 5x as well, but improvement  would be <20%,
+# so it's not worth it...
+sub aesni_generate4
+{ my $p=shift;
+
+    &function_begin_B("_aesni_${p}rypt4");
+	&$movekey	($rndkey0,&QWP(0,$key));
+	&$movekey	($rndkey1,&QWP(16,$key));
+	&shr		($rounds,1);
+	&lea		($key,&DWP(32,$key));
+	&xorps		($inout0,$rndkey0);
+	&pxor		($inout1,$rndkey0);
+	&pxor		($inout2,$rndkey0);
+	&pxor		($inout3,$rndkey0);
+	&$movekey	($rndkey0,&QWP(0,$key));
+
+    &set_label("${p}4_loop");
+	eval"&aes${p}	($inout0,$rndkey1)";
+	eval"&aes${p}	($inout1,$rndkey1)";
+	&dec		($rounds);
+	eval"&aes${p}	($inout2,$rndkey1)";
+	eval"&aes${p}	($inout3,$rndkey1)";
+	&$movekey	($rndkey1,&QWP(16,$key));
+	eval"&aes${p}	($inout0,$rndkey0)";
+	eval"&aes${p}	($inout1,$rndkey0)";
+	&lea		($key,&DWP(32,$key));
+	eval"&aes${p}	($inout2,$rndkey0)";
+	eval"&aes${p}	($inout3,$rndkey0)";
+	&$movekey	($rndkey0,&QWP(0,$key));
+    &jnz		(&label("${p}4_loop"));
+
+    eval"&aes${p}	($inout0,$rndkey1)";
+    eval"&aes${p}	($inout1,$rndkey1)";
+    eval"&aes${p}	($inout2,$rndkey1)";
+    eval"&aes${p}	($inout3,$rndkey1)";
+    eval"&aes${p}last	($inout0,$rndkey0)";
+    eval"&aes${p}last	($inout1,$rndkey0)";
+    eval"&aes${p}last	($inout2,$rndkey0)";
+    eval"&aes${p}last	($inout3,$rndkey0)";
+    &ret();
+    &function_end_B("_aesni_${p}rypt4");
+}
+
+sub aesni_generate6
+{ my $p=shift;
+
+    &function_begin_B("_aesni_${p}rypt6");
+    &static_label("_aesni_${p}rypt6_enter");
+	&$movekey	($rndkey0,&QWP(0,$key));
+	&shr		($rounds,1);
+	&$movekey	($rndkey1,&QWP(16,$key));
+	&lea		($key,&DWP(32,$key));
+	&xorps		($inout0,$rndkey0);
+	&pxor		($inout1,$rndkey0);	# pxor does better here
+	eval"&aes${p}	($inout0,$rndkey1)";
+	&pxor		($inout2,$rndkey0);
+	eval"&aes${p}	($inout1,$rndkey1)";
+	&pxor		($inout3,$rndkey0);
+	&dec		($rounds);
+	eval"&aes${p}	($inout2,$rndkey1)";
+	&pxor		($inout4,$rndkey0);
+	eval"&aes${p}	($inout3,$rndkey1)";
+	&pxor		($inout5,$rndkey0);
+	eval"&aes${p}	($inout4,$rndkey1)";
+	&$movekey	($rndkey0,&QWP(0,$key));
+	eval"&aes${p}	($inout5,$rndkey1)";
+	&jmp		(&label("_aesni_${p}rypt6_enter"));
+
+    &set_label("${p}6_loop",16);
+	eval"&aes${p}	($inout0,$rndkey1)";
+	eval"&aes${p}	($inout1,$rndkey1)";
+	&dec		($rounds);
+	eval"&aes${p}	($inout2,$rndkey1)";
+	eval"&aes${p}	($inout3,$rndkey1)";
+	eval"&aes${p}	($inout4,$rndkey1)";
+	eval"&aes${p}	($inout5,$rndkey1)";
+    &set_label("_aesni_${p}rypt6_enter",16);
+	&$movekey	($rndkey1,&QWP(16,$key));
+	eval"&aes${p}	($inout0,$rndkey0)";
+	eval"&aes${p}	($inout1,$rndkey0)";
+	&lea		($key,&DWP(32,$key));
+	eval"&aes${p}	($inout2,$rndkey0)";
+	eval"&aes${p}	($inout3,$rndkey0)";
+	eval"&aes${p}	($inout4,$rndkey0)";
+	eval"&aes${p}	($inout5,$rndkey0)";
+	&$movekey	($rndkey0,&QWP(0,$key));
+    &jnz		(&label("${p}6_loop"));
+
+    eval"&aes${p}	($inout0,$rndkey1)";
+    eval"&aes${p}	($inout1,$rndkey1)";
+    eval"&aes${p}	($inout2,$rndkey1)";
+    eval"&aes${p}	($inout3,$rndkey1)";
+    eval"&aes${p}	($inout4,$rndkey1)";
+    eval"&aes${p}	($inout5,$rndkey1)";
+    eval"&aes${p}last	($inout0,$rndkey0)";
+    eval"&aes${p}last	($inout1,$rndkey0)";
+    eval"&aes${p}last	($inout2,$rndkey0)";
+    eval"&aes${p}last	($inout3,$rndkey0)";
+    eval"&aes${p}last	($inout4,$rndkey0)";
+    eval"&aes${p}last	($inout5,$rndkey0)";
+    &ret();
+    &function_end_B("_aesni_${p}rypt6");
+}
+&aesni_generate3("enc") if ($PREFIX eq "aesni");
+&aesni_generate3("dec");
+&aesni_generate4("enc") if ($PREFIX eq "aesni");
+&aesni_generate4("dec");
+&aesni_generate6("enc") if ($PREFIX eq "aesni");
+&aesni_generate6("dec");
+
+if ($PREFIX eq "aesni") {
+######################################################################
+# void aesni_ecb_encrypt (const void *in, void *out,
+#                         size_t length, const AES_KEY *key,
+#                         int enc);
+&function_begin("aesni_ecb_encrypt");
+	&mov	($inp,&wparam(0));
+	&mov	($out,&wparam(1));
+	&mov	($len,&wparam(2));
+	&mov	($key,&wparam(3));
+	&mov	($rounds_,&wparam(4));
+	&and	($len,-16);
+	&jz	(&label("ecb_ret"));
+	&mov	($rounds,&DWP(240,$key));
+	&test	($rounds_,$rounds_);
+	&jz	(&label("ecb_decrypt"));
+
+	&mov	($key_,$key);		# backup $key
+	&mov	($rounds_,$rounds);	# backup $rounds
+	&cmp	($len,0x60);
+	&jb	(&label("ecb_enc_tail"));
+
+	&movdqu	($inout0,&QWP(0,$inp));
+	&movdqu	($inout1,&QWP(0x10,$inp));
+	&movdqu	($inout2,&QWP(0x20,$inp));
+	&movdqu	($inout3,&QWP(0x30,$inp));
+	&movdqu	($inout4,&QWP(0x40,$inp));
+	&movdqu	($inout5,&QWP(0x50,$inp));
+	&lea	($inp,&DWP(0x60,$inp));
+	&sub	($len,0x60);
+	&jmp	(&label("ecb_enc_loop6_enter"));
+
+&set_label("ecb_enc_loop6",16);
+	&movups	(&QWP(0,$out),$inout0);
+	&movdqu	($inout0,&QWP(0,$inp));
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movdqu	($inout1,&QWP(0x10,$inp));
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movdqu	($inout2,&QWP(0x20,$inp));
+	&movups	(&QWP(0x30,$out),$inout3);
+	&movdqu	($inout3,&QWP(0x30,$inp));
+	&movups	(&QWP(0x40,$out),$inout4);
+	&movdqu	($inout4,&QWP(0x40,$inp));
+	&movups	(&QWP(0x50,$out),$inout5);
+	&lea	($out,&DWP(0x60,$out));
+	&movdqu	($inout5,&QWP(0x50,$inp));
+	&lea	($inp,&DWP(0x60,$inp));
+&set_label("ecb_enc_loop6_enter");
+
+	&call	("_aesni_encrypt6");
+
+	&mov	($key,$key_);		# restore $key
+	&mov	($rounds,$rounds_);	# restore $rounds
+	&sub	($len,0x60);
+	&jnc	(&label("ecb_enc_loop6"));
+
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+	&movups	(&QWP(0x40,$out),$inout4);
+	&movups	(&QWP(0x50,$out),$inout5);
+	&lea	($out,&DWP(0x60,$out));
+	&add	($len,0x60);
+	&jz	(&label("ecb_ret"));
+
+&set_label("ecb_enc_tail");
+	&movups	($inout0,&QWP(0,$inp));
+	&cmp	($len,0x20);
+	&jb	(&label("ecb_enc_one"));
+	&movups	($inout1,&QWP(0x10,$inp));
+	&je	(&label("ecb_enc_two"));
+	&movups	($inout2,&QWP(0x20,$inp));
+	&cmp	($len,0x40);
+	&jb	(&label("ecb_enc_three"));
+	&movups	($inout3,&QWP(0x30,$inp));
+	&je	(&label("ecb_enc_four"));
+	&movups	($inout4,&QWP(0x40,$inp));
+	&xorps	($inout5,$inout5);
+	&call	("_aesni_encrypt6");
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+	&movups	(&QWP(0x40,$out),$inout4);
+	jmp	(&label("ecb_ret"));
+
+&set_label("ecb_enc_one",16);
+	if ($inline)
+	{   &aesni_inline_generate1("enc");	}
+	else
+	{   &call	("_aesni_encrypt1");	}
+	&movups	(&QWP(0,$out),$inout0);
+	&jmp	(&label("ecb_ret"));
+
+&set_label("ecb_enc_two",16);
+	&xorps	($inout2,$inout2);
+	&call	("_aesni_encrypt3");
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&jmp	(&label("ecb_ret"));
+
+&set_label("ecb_enc_three",16);
+	&call	("_aesni_encrypt3");
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&jmp	(&label("ecb_ret"));
+
+&set_label("ecb_enc_four",16);
+	&call	("_aesni_encrypt4");
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+	&jmp	(&label("ecb_ret"));
+######################################################################
+&set_label("ecb_decrypt",16);
+	&mov	($key_,$key);		# backup $key
+	&mov	($rounds_,$rounds);	# backup $rounds
+	&cmp	($len,0x60);
+	&jb	(&label("ecb_dec_tail"));
+
+	&movdqu	($inout0,&QWP(0,$inp));
+	&movdqu	($inout1,&QWP(0x10,$inp));
+	&movdqu	($inout2,&QWP(0x20,$inp));
+	&movdqu	($inout3,&QWP(0x30,$inp));
+	&movdqu	($inout4,&QWP(0x40,$inp));
+	&movdqu	($inout5,&QWP(0x50,$inp));
+	&lea	($inp,&DWP(0x60,$inp));
+	&sub	($len,0x60);
+	&jmp	(&label("ecb_dec_loop6_enter"));
+
+&set_label("ecb_dec_loop6",16);
+	&movups	(&QWP(0,$out),$inout0);
+	&movdqu	($inout0,&QWP(0,$inp));
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movdqu	($inout1,&QWP(0x10,$inp));
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movdqu	($inout2,&QWP(0x20,$inp));
+	&movups	(&QWP(0x30,$out),$inout3);
+	&movdqu	($inout3,&QWP(0x30,$inp));
+	&movups	(&QWP(0x40,$out),$inout4);
+	&movdqu	($inout4,&QWP(0x40,$inp));
+	&movups	(&QWP(0x50,$out),$inout5);
+	&lea	($out,&DWP(0x60,$out));
+	&movdqu	($inout5,&QWP(0x50,$inp));
+	&lea	($inp,&DWP(0x60,$inp));
+&set_label("ecb_dec_loop6_enter");
+
+	&call	("_aesni_decrypt6");
+
+	&mov	($key,$key_);		# restore $key
+	&mov	($rounds,$rounds_);	# restore $rounds
+	&sub	($len,0x60);
+	&jnc	(&label("ecb_dec_loop6"));
+
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+	&movups	(&QWP(0x40,$out),$inout4);
+	&movups	(&QWP(0x50,$out),$inout5);
+	&lea	($out,&DWP(0x60,$out));
+	&add	($len,0x60);
+	&jz	(&label("ecb_ret"));
+
+&set_label("ecb_dec_tail");
+	&movups	($inout0,&QWP(0,$inp));
+	&cmp	($len,0x20);
+	&jb	(&label("ecb_dec_one"));
+	&movups	($inout1,&QWP(0x10,$inp));
+	&je	(&label("ecb_dec_two"));
+	&movups	($inout2,&QWP(0x20,$inp));
+	&cmp	($len,0x40);
+	&jb	(&label("ecb_dec_three"));
+	&movups	($inout3,&QWP(0x30,$inp));
+	&je	(&label("ecb_dec_four"));
+	&movups	($inout4,&QWP(0x40,$inp));
+	&xorps	($inout5,$inout5);
+	&call	("_aesni_decrypt6");
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+	&movups	(&QWP(0x40,$out),$inout4);
+	&jmp	(&label("ecb_ret"));
+
+&set_label("ecb_dec_one",16);
+	if ($inline)
+	{   &aesni_inline_generate1("dec");	}
+	else
+	{   &call	("_aesni_decrypt1");	}
+	&movups	(&QWP(0,$out),$inout0);
+	&jmp	(&label("ecb_ret"));
+
+&set_label("ecb_dec_two",16);
+	&xorps	($inout2,$inout2);
+	&call	("_aesni_decrypt3");
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&jmp	(&label("ecb_ret"));
+
+&set_label("ecb_dec_three",16);
+	&call	("_aesni_decrypt3");
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&jmp	(&label("ecb_ret"));
+
+&set_label("ecb_dec_four",16);
+	&call	("_aesni_decrypt4");
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+
+&set_label("ecb_ret");
+&function_end("aesni_ecb_encrypt");
+
+######################################################################
+# void aesni_ccm64_[en|de]crypt_blocks (const void *in, void *out,
+#                         size_t blocks, const AES_KEY *key,
+#                         const char *ivec,char *cmac);
+#
+# Handles only complete blocks, operates on 64-bit counter and
+# does not update *ivec! Nor does it finalize CMAC value
+# (see engine/eng_aesni.c for details)
+#
+{ my $cmac=$inout1;
+&function_begin("aesni_ccm64_encrypt_blocks");
+	&mov	($inp,&wparam(0));
+	&mov	($out,&wparam(1));
+	&mov	($len,&wparam(2));
+	&mov	($key,&wparam(3));
+	&mov	($rounds_,&wparam(4));
+	&mov	($rounds,&wparam(5));
+	&mov	($key_,"esp");
+	&sub	("esp",60);
+	&and	("esp",-16);			# align stack
+	&mov	(&DWP(48,"esp"),$key_);
+
+	&movdqu	($ivec,&QWP(0,$rounds_));	# load ivec
+	&movdqu	($cmac,&QWP(0,$rounds));	# load cmac
+	&mov	($rounds,&DWP(240,$key));
+
+	# compose byte-swap control mask for pshufb on stack
+	&mov	(&DWP(0,"esp"),0x0c0d0e0f);
+	&mov	(&DWP(4,"esp"),0x08090a0b);
+	&mov	(&DWP(8,"esp"),0x04050607);
+	&mov	(&DWP(12,"esp"),0x00010203);
+
+	# compose counter increment vector on stack
+	&mov	($rounds_,1);
+	&xor	($key_,$key_);
+	&mov	(&DWP(16,"esp"),$rounds_);
+	&mov	(&DWP(20,"esp"),$key_);
+	&mov	(&DWP(24,"esp"),$key_);
+	&mov	(&DWP(28,"esp"),$key_);
+
+	&shr	($rounds,1);
+	&lea	($key_,&DWP(0,$key));
+	&movdqa	($inout3,&QWP(0,"esp"));
+	&movdqa	($inout0,$ivec);
+	&mov	($rounds_,$rounds);
+	&pshufb	($ivec,$inout3);
+
+&set_label("ccm64_enc_outer");
+	&$movekey	($rndkey0,&QWP(0,$key_));
+	&mov		($rounds,$rounds_);
+	&movups		($in0,&QWP(0,$inp));
+
+	&xorps		($inout0,$rndkey0);
+	&$movekey	($rndkey1,&QWP(16,$key_));
+	&xorps		($rndkey0,$in0);
+	&lea		($key,&DWP(32,$key_));
+	&xorps		($cmac,$rndkey0);		# cmac^=inp
+	&$movekey	($rndkey0,&QWP(0,$key));
+
+&set_label("ccm64_enc2_loop");
+	&aesenc		($inout0,$rndkey1);
+	&dec		($rounds);
+	&aesenc		($cmac,$rndkey1);
+	&$movekey	($rndkey1,&QWP(16,$key));
+	&aesenc		($inout0,$rndkey0);
+	&lea		($key,&DWP(32,$key));
+	&aesenc		($cmac,$rndkey0);
+	&$movekey	($rndkey0,&QWP(0,$key));
+	&jnz		(&label("ccm64_enc2_loop"));
+	&aesenc		($inout0,$rndkey1);
+	&aesenc		($cmac,$rndkey1);
+	&paddq		($ivec,&QWP(16,"esp"));
+	&aesenclast	($inout0,$rndkey0);
+	&aesenclast	($cmac,$rndkey0);
+
+	&dec	($len);
+	&lea	($inp,&DWP(16,$inp));
+	&xorps	($in0,$inout0);			# inp^=E(ivec)
+	&movdqa	($inout0,$ivec);
+	&movups	(&QWP(0,$out),$in0);		# save output
+	&lea	($out,&DWP(16,$out));
+	&pshufb	($inout0,$inout3);
+	&jnz	(&label("ccm64_enc_outer"));
+
+	&mov	("esp",&DWP(48,"esp"));
+	&mov	($out,&wparam(5));
+	&movups	(&QWP(0,$out),$cmac);
+&function_end("aesni_ccm64_encrypt_blocks");
+
+&function_begin("aesni_ccm64_decrypt_blocks");
+	&mov	($inp,&wparam(0));
+	&mov	($out,&wparam(1));
+	&mov	($len,&wparam(2));
+	&mov	($key,&wparam(3));
+	&mov	($rounds_,&wparam(4));
+	&mov	($rounds,&wparam(5));
+	&mov	($key_,"esp");
+	&sub	("esp",60);
+	&and	("esp",-16);			# align stack
+	&mov	(&DWP(48,"esp"),$key_);
+
+	&movdqu	($ivec,&QWP(0,$rounds_));	# load ivec
+	&movdqu	($cmac,&QWP(0,$rounds));	# load cmac
+	&mov	($rounds,&DWP(240,$key));
+
+	# compose byte-swap control mask for pshufb on stack
+	&mov	(&DWP(0,"esp"),0x0c0d0e0f);
+	&mov	(&DWP(4,"esp"),0x08090a0b);
+	&mov	(&DWP(8,"esp"),0x04050607);
+	&mov	(&DWP(12,"esp"),0x00010203);
+
+	# compose counter increment vector on stack
+	&mov	($rounds_,1);
+	&xor	($key_,$key_);
+	&mov	(&DWP(16,"esp"),$rounds_);
+	&mov	(&DWP(20,"esp"),$key_);
+	&mov	(&DWP(24,"esp"),$key_);
+	&mov	(&DWP(28,"esp"),$key_);
+
+	&movdqa	($inout3,&QWP(0,"esp"));	# bswap mask
+	&movdqa	($inout0,$ivec);
+
+	&mov	($key_,$key);
+	&mov	($rounds_,$rounds);
+
+	&pshufb	($ivec,$inout3);
+	if ($inline)
+	{   &aesni_inline_generate1("enc");	}
+	else
+	{   &call	("_aesni_encrypt1");	}
+	&movups	($in0,&QWP(0,$inp));		# load inp
+	&paddq	($ivec,&QWP(16,"esp"));
+	&lea	($inp,&QWP(16,$inp));
+	&jmp	(&label("ccm64_dec_outer"));
+
+&set_label("ccm64_dec_outer",16);
+	&xorps	($in0,$inout0);			# inp ^= E(ivec)
+	&movdqa	($inout0,$ivec);
+	&mov	($rounds,$rounds_);
+	&movups	(&QWP(0,$out),$in0);		# save output
+	&lea	($out,&DWP(16,$out));
+	&pshufb	($inout0,$inout3);
+
+	&sub	($len,1);
+	&jz	(&label("ccm64_dec_break"));
+
+	&$movekey	($rndkey0,&QWP(0,$key_));
+	&shr		($rounds,1);
+	&$movekey	($rndkey1,&QWP(16,$key_));
+	&xorps		($in0,$rndkey0);
+	&lea		($key,&DWP(32,$key_));
+	&xorps		($inout0,$rndkey0);
+	&xorps		($cmac,$in0);		# cmac^=out
+	&$movekey	($rndkey0,&QWP(0,$key));
+
+&set_label("ccm64_dec2_loop");
+	&aesenc		($inout0,$rndkey1);
+	&dec		($rounds);
+	&aesenc		($cmac,$rndkey1);
+	&$movekey	($rndkey1,&QWP(16,$key));
+	&aesenc		($inout0,$rndkey0);
+	&lea		($key,&DWP(32,$key));
+	&aesenc		($cmac,$rndkey0);
+	&$movekey	($rndkey0,&QWP(0,$key));
+	&jnz		(&label("ccm64_dec2_loop"));
+	&movups		($in0,&QWP(0,$inp));	# load inp
+	&paddq		($ivec,&QWP(16,"esp"));
+	&aesenc		($inout0,$rndkey1);
+	&aesenc		($cmac,$rndkey1);
+	&lea		($inp,&QWP(16,$inp));
+	&aesenclast	($inout0,$rndkey0);
+	&aesenclast	($cmac,$rndkey0);
+	&jmp	(&label("ccm64_dec_outer"));
+
+&set_label("ccm64_dec_break",16);
+	&mov	($key,$key_);
+	if ($inline)
+	{   &aesni_inline_generate1("enc",$cmac,$in0);	}
+	else
+	{   &call	("_aesni_encrypt1",$cmac);	}
+
+	&mov	("esp",&DWP(48,"esp"));
+	&mov	($out,&wparam(5));
+	&movups	(&QWP(0,$out),$cmac);
+&function_end("aesni_ccm64_decrypt_blocks");
+}
+
+######################################################################
+# void aesni_ctr32_encrypt_blocks (const void *in, void *out,
+#                         size_t blocks, const AES_KEY *key,
+#                         const char *ivec);
+#
+# Handles only complete blocks, operates on 32-bit counter and
+# does not update *ivec! (see engine/eng_aesni.c for details)
+#
+# stack layout:
+#	0	pshufb mask
+#	16	vector addend: 0,6,6,6
+# 	32	counter-less ivec
+#	48	1st triplet of counter vector
+#	64	2nd triplet of counter vector
+#	80	saved %esp
+
+&function_begin("aesni_ctr32_encrypt_blocks");
+	&mov	($inp,&wparam(0));
+	&mov	($out,&wparam(1));
+	&mov	($len,&wparam(2));
+	&mov	($key,&wparam(3));
+	&mov	($rounds_,&wparam(4));
+	&mov	($key_,"esp");
+	&sub	("esp",88);
+	&and	("esp",-16);			# align stack
+	&mov	(&DWP(80,"esp"),$key_);
+
+	&cmp	($len,1);
+	&je	(&label("ctr32_one_shortcut"));
+
+	&movdqu	($inout5,&QWP(0,$rounds_));	# load ivec
+
+	# compose byte-swap control mask for pshufb on stack
+	&mov	(&DWP(0,"esp"),0x0c0d0e0f);
+	&mov	(&DWP(4,"esp"),0x08090a0b);
+	&mov	(&DWP(8,"esp"),0x04050607);
+	&mov	(&DWP(12,"esp"),0x00010203);
+
+	# compose counter increment vector on stack
+	&mov	($rounds,6);
+	&xor	($key_,$key_);
+	&mov	(&DWP(16,"esp"),$rounds);
+	&mov	(&DWP(20,"esp"),$rounds);
+	&mov	(&DWP(24,"esp"),$rounds);
+	&mov	(&DWP(28,"esp"),$key_);
+
+	&pextrd	($rounds_,$inout5,3);		# pull 32-bit counter
+	&pinsrd	($inout5,$key_,3);		# wipe 32-bit counter
+
+	&mov	($rounds,&DWP(240,$key));	# key->rounds
+
+	# compose 2 vectors of 3x32-bit counters
+	&bswap	($rounds_);
+	&pxor	($rndkey1,$rndkey1);
+	&pxor	($rndkey0,$rndkey0);
+	&movdqa	($inout0,&QWP(0,"esp"));	# load byte-swap mask
+	&pinsrd	($rndkey1,$rounds_,0);
+	&lea	($key_,&DWP(3,$rounds_));
+	&pinsrd	($rndkey0,$key_,0);
+	&inc	($rounds_);
+	&pinsrd	($rndkey1,$rounds_,1);
+	&inc	($key_);
+	&pinsrd	($rndkey0,$key_,1);
+	&inc	($rounds_);
+	&pinsrd	($rndkey1,$rounds_,2);
+	&inc	($key_);
+	&pinsrd	($rndkey0,$key_,2);
+	&movdqa	(&QWP(48,"esp"),$rndkey1);	# save 1st triplet
+	&pshufb	($rndkey1,$inout0);		# byte swap
+	&movdqa	(&QWP(64,"esp"),$rndkey0);	# save 2nd triplet
+	&pshufb	($rndkey0,$inout0);		# byte swap
+
+	&pshufd	($inout0,$rndkey1,3<<6);	# place counter to upper dword
+	&pshufd	($inout1,$rndkey1,2<<6);
+	&cmp	($len,6);
+	&jb	(&label("ctr32_tail"));
+	&movdqa	(&QWP(32,"esp"),$inout5);	# save counter-less ivec
+	&shr	($rounds,1);
+	&mov	($key_,$key);			# backup $key
+	&mov	($rounds_,$rounds);		# backup $rounds
+	&sub	($len,6);
+	&jmp	(&label("ctr32_loop6"));
+
+&set_label("ctr32_loop6",16);
+	&pshufd	($inout2,$rndkey1,1<<6);
+	&movdqa	($rndkey1,&QWP(32,"esp"));	# pull counter-less ivec
+	&pshufd	($inout3,$rndkey0,3<<6);
+	&por	($inout0,$rndkey1);		# merge counter-less ivec
+	&pshufd	($inout4,$rndkey0,2<<6);
+	&por	($inout1,$rndkey1);
+	&pshufd	($inout5,$rndkey0,1<<6);
+	&por	($inout2,$rndkey1);
+	&por	($inout3,$rndkey1);
+	&por	($inout4,$rndkey1);
+	&por	($inout5,$rndkey1);
+
+	# inlining _aesni_encrypt6's prologue gives ~4% improvement...
+	&$movekey	($rndkey0,&QWP(0,$key_));
+	&$movekey	($rndkey1,&QWP(16,$key_));
+	&lea		($key,&DWP(32,$key_));
+	&dec		($rounds);
+	&pxor		($inout0,$rndkey0);
+	&pxor		($inout1,$rndkey0);
+	&aesenc		($inout0,$rndkey1);
+	&pxor		($inout2,$rndkey0);
+	&aesenc		($inout1,$rndkey1);
+	&pxor		($inout3,$rndkey0);
+	&aesenc		($inout2,$rndkey1);
+	&pxor		($inout4,$rndkey0);
+	&aesenc		($inout3,$rndkey1);
+	&pxor		($inout5,$rndkey0);
+	&aesenc		($inout4,$rndkey1);
+	&$movekey	($rndkey0,&QWP(0,$key));
+	&aesenc		($inout5,$rndkey1);
+
+	&call		(&label("_aesni_encrypt6_enter"));
+
+	&movups	($rndkey1,&QWP(0,$inp));
+	&movups	($rndkey0,&QWP(0x10,$inp));
+	&xorps	($inout0,$rndkey1);
+	&movups	($rndkey1,&QWP(0x20,$inp));
+	&xorps	($inout1,$rndkey0);
+	&movups	(&QWP(0,$out),$inout0);
+	&movdqa	($rndkey0,&QWP(16,"esp"));	# load increment
+	&xorps	($inout2,$rndkey1);
+	&movdqa	($rndkey1,&QWP(48,"esp"));	# load 1st triplet
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+
+	&paddd	($rndkey1,$rndkey0);		# 1st triplet increment
+	&paddd	($rndkey0,&QWP(64,"esp"));	# 2nd triplet increment
+	&movdqa	($inout0,&QWP(0,"esp"));	# load byte swap mask
+
+	&movups	($inout1,&QWP(0x30,$inp));
+	&movups	($inout2,&QWP(0x40,$inp));
+	&xorps	($inout3,$inout1);
+	&movups	($inout1,&QWP(0x50,$inp));
+	&lea	($inp,&DWP(0x60,$inp));
+	&movdqa	(&QWP(48,"esp"),$rndkey1);	# save 1st triplet
+	&pshufb	($rndkey1,$inout0);		# byte swap
+	&xorps	($inout4,$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+	&xorps	($inout5,$inout1);
+	&movdqa	(&QWP(64,"esp"),$rndkey0);	# save 2nd triplet
+	&pshufb	($rndkey0,$inout0);		# byte swap
+	&movups	(&QWP(0x40,$out),$inout4);
+	&pshufd	($inout0,$rndkey1,3<<6);
+	&movups	(&QWP(0x50,$out),$inout5);
+	&lea	($out,&DWP(0x60,$out));
+
+	&mov	($rounds,$rounds_);
+	&pshufd	($inout1,$rndkey1,2<<6);
+	&sub	($len,6);
+	&jnc	(&label("ctr32_loop6"));
+
+	&add	($len,6);
+	&jz	(&label("ctr32_ret"));
+	&mov	($key,$key_);
+	&lea	($rounds,&DWP(1,"",$rounds,2));	# restore $rounds
+	&movdqa	($inout5,&QWP(32,"esp"));	# pull count-less ivec
+
+&set_label("ctr32_tail");
+	&por	($inout0,$inout5);
+	&cmp	($len,2);
+	&jb	(&label("ctr32_one"));
+
+	&pshufd	($inout2,$rndkey1,1<<6);
+	&por	($inout1,$inout5);
+	&je	(&label("ctr32_two"));
+
+	&pshufd	($inout3,$rndkey0,3<<6);
+	&por	($inout2,$inout5);
+	&cmp	($len,4);
+	&jb	(&label("ctr32_three"));
+
+	&pshufd	($inout4,$rndkey0,2<<6);
+	&por	($inout3,$inout5);
+	&je	(&label("ctr32_four"));
+
+	&por	($inout4,$inout5);
+	&call	("_aesni_encrypt6");
+	&movups	($rndkey1,&QWP(0,$inp));
+	&movups	($rndkey0,&QWP(0x10,$inp));
+	&xorps	($inout0,$rndkey1);
+	&movups	($rndkey1,&QWP(0x20,$inp));
+	&xorps	($inout1,$rndkey0);
+	&movups	($rndkey0,&QWP(0x30,$inp));
+	&xorps	($inout2,$rndkey1);
+	&movups	($rndkey1,&QWP(0x40,$inp));
+	&xorps	($inout3,$rndkey0);
+	&movups	(&QWP(0,$out),$inout0);
+	&xorps	($inout4,$rndkey1);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+	&movups	(&QWP(0x40,$out),$inout4);
+	&jmp	(&label("ctr32_ret"));
+
+&set_label("ctr32_one_shortcut",16);
+	&movups	($inout0,&QWP(0,$rounds_));	# load ivec
+	&mov	($rounds,&DWP(240,$key));
+	
+&set_label("ctr32_one");
+	if ($inline)
+	{   &aesni_inline_generate1("enc");	}
+	else
+	{   &call	("_aesni_encrypt1");	}
+	&movups	($in0,&QWP(0,$inp));
+	&xorps	($in0,$inout0);
+	&movups	(&QWP(0,$out),$in0);
+	&jmp	(&label("ctr32_ret"));
+
+&set_label("ctr32_two",16);
+	&call	("_aesni_encrypt3");
+	&movups	($inout3,&QWP(0,$inp));
+	&movups	($inout4,&QWP(0x10,$inp));
+	&xorps	($inout0,$inout3);
+	&xorps	($inout1,$inout4);
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&jmp	(&label("ctr32_ret"));
+
+&set_label("ctr32_three",16);
+	&call	("_aesni_encrypt3");
+	&movups	($inout3,&QWP(0,$inp));
+	&movups	($inout4,&QWP(0x10,$inp));
+	&xorps	($inout0,$inout3);
+	&movups	($inout5,&QWP(0x20,$inp));
+	&xorps	($inout1,$inout4);
+	&movups	(&QWP(0,$out),$inout0);
+	&xorps	($inout2,$inout5);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&jmp	(&label("ctr32_ret"));
+
+&set_label("ctr32_four",16);
+	&call	("_aesni_encrypt4");
+	&movups	($inout4,&QWP(0,$inp));
+	&movups	($inout5,&QWP(0x10,$inp));
+	&movups	($rndkey1,&QWP(0x20,$inp));
+	&xorps	($inout0,$inout4);
+	&movups	($rndkey0,&QWP(0x30,$inp));
+	&xorps	($inout1,$inout5);
+	&movups	(&QWP(0,$out),$inout0);
+	&xorps	($inout2,$rndkey1);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&xorps	($inout3,$rndkey0);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+
+&set_label("ctr32_ret");
+	&mov	("esp",&DWP(80,"esp"));
+&function_end("aesni_ctr32_encrypt_blocks");
+
+######################################################################
+# void aesni_xts_[en|de]crypt(const char *inp,char *out,size_t len,
+#	const AES_KEY *key1, const AES_KEY *key2
+#	const unsigned char iv[16]);
+#
+{ my ($tweak,$twtmp,$twres,$twmask)=($rndkey1,$rndkey0,$inout0,$inout1);
+
+&function_begin("aesni_xts_encrypt");
+	&mov	($key,&wparam(4));		# key2
+	&mov	($inp,&wparam(5));		# clear-text tweak
+
+	&mov	($rounds,&DWP(240,$key));	# key2->rounds
+	&movups	($inout0,&QWP(0,$inp));
+	if ($inline)
+	{   &aesni_inline_generate1("enc");	}
+	else
+	{   &call	("_aesni_encrypt1");	}
+
+	&mov	($inp,&wparam(0));
+	&mov	($out,&wparam(1));
+	&mov	($len,&wparam(2));
+	&mov	($key,&wparam(3));		# key1
+
+	&mov	($key_,"esp");
+	&sub	("esp",16*7+8);
+	&mov	($rounds,&DWP(240,$key));	# key1->rounds
+	&and	("esp",-16);			# align stack
+
+	&mov	(&DWP(16*6+0,"esp"),0x87);	# compose the magic constant
+	&mov	(&DWP(16*6+4,"esp"),0);
+	&mov	(&DWP(16*6+8,"esp"),1);
+	&mov	(&DWP(16*6+12,"esp"),0);
+	&mov	(&DWP(16*7+0,"esp"),$len);	# save original $len
+	&mov	(&DWP(16*7+4,"esp"),$key_);	# save original %esp
+
+	&movdqa	($tweak,$inout0);
+	&pxor	($twtmp,$twtmp);
+	&movdqa	($twmask,&QWP(6*16,"esp"));	# 0x0...010...87
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+
+	&and	($len,-16);
+	&mov	($key_,$key);			# backup $key
+	&mov	($rounds_,$rounds);		# backup $rounds
+	&sub	($len,16*6);
+	&jc	(&label("xts_enc_short"));
+
+	&shr	($rounds,1);
+	&mov	($rounds_,$rounds);
+	&jmp	(&label("xts_enc_loop6"));
+
+&set_label("xts_enc_loop6",16);
+	for ($i=0;$i<4;$i++) {
+	    &pshufd	($twres,$twtmp,0x13);
+	    &pxor	($twtmp,$twtmp);
+	    &movdqa	(&QWP(16*$i,"esp"),$tweak);
+	    &paddq	($tweak,$tweak);	# &psllq($tweak,1);
+	    &pand	($twres,$twmask);	# isolate carry and residue
+	    &pcmpgtd	($twtmp,$tweak);	# broadcast upper bits
+	    &pxor	($tweak,$twres);
+	}
+	&pshufd	($inout5,$twtmp,0x13);
+	&movdqa	(&QWP(16*$i++,"esp"),$tweak);
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	 &$movekey	($rndkey0,&QWP(0,$key_));
+	&pand	($inout5,$twmask);		# isolate carry and residue
+	 &movups	($inout0,&QWP(0,$inp));	# load input
+	&pxor	($inout5,$tweak);
+
+	# inline _aesni_encrypt6 prologue and flip xor with tweak and key[0]
+	&movdqu	($inout1,&QWP(16*1,$inp));
+	 &xorps		($inout0,$rndkey0);	# input^=rndkey[0]
+	&movdqu	($inout2,&QWP(16*2,$inp));
+	 &pxor		($inout1,$rndkey0);
+	&movdqu	($inout3,&QWP(16*3,$inp));
+	 &pxor		($inout2,$rndkey0);
+	&movdqu	($inout4,&QWP(16*4,$inp));
+	 &pxor		($inout3,$rndkey0);
+	&movdqu	($rndkey1,&QWP(16*5,$inp));
+	 &pxor		($inout4,$rndkey0);
+	&lea	($inp,&DWP(16*6,$inp));
+	&pxor	($inout0,&QWP(16*0,"esp"));	# input^=tweak
+	&movdqa	(&QWP(16*$i,"esp"),$inout5);	# save last tweak
+	&pxor	($inout5,$rndkey1);
+
+	 &$movekey	($rndkey1,&QWP(16,$key_));
+	 &lea		($key,&DWP(32,$key_));
+	&pxor	($inout1,&QWP(16*1,"esp"));
+	 &aesenc	($inout0,$rndkey1);
+	&pxor	($inout2,&QWP(16*2,"esp"));
+	 &aesenc	($inout1,$rndkey1);
+	&pxor	($inout3,&QWP(16*3,"esp"));
+	 &dec		($rounds);
+	 &aesenc	($inout2,$rndkey1);
+	&pxor	($inout4,&QWP(16*4,"esp"));
+	 &aesenc	($inout3,$rndkey1);
+	&pxor		($inout5,$rndkey0);
+	 &aesenc	($inout4,$rndkey1);
+	 &$movekey	($rndkey0,&QWP(0,$key));
+	 &aesenc	($inout5,$rndkey1);
+	&call		(&label("_aesni_encrypt6_enter"));
+
+	&movdqa	($tweak,&QWP(16*5,"esp"));	# last tweak
+       &pxor	($twtmp,$twtmp);
+	&xorps	($inout0,&QWP(16*0,"esp"));	# output^=tweak
+       &pcmpgtd	($twtmp,$tweak);		# broadcast upper bits
+	&xorps	($inout1,&QWP(16*1,"esp"));
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&xorps	($inout2,&QWP(16*2,"esp"));
+	&movups	(&QWP(16*1,$out),$inout1);
+	&xorps	($inout3,&QWP(16*3,"esp"));
+	&movups	(&QWP(16*2,$out),$inout2);
+	&xorps	($inout4,&QWP(16*4,"esp"));
+	&movups	(&QWP(16*3,$out),$inout3);
+	&xorps	($inout5,$tweak);
+	&movups	(&QWP(16*4,$out),$inout4);
+       &pshufd	($twres,$twtmp,0x13);
+	&movups	(&QWP(16*5,$out),$inout5);
+	&lea	($out,&DWP(16*6,$out));
+       &movdqa	($twmask,&QWP(16*6,"esp"));	# 0x0...010...87
+
+	&pxor	($twtmp,$twtmp);
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($twres,$twmask);		# isolate carry and residue
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&mov	($rounds,$rounds_);		# restore $rounds
+	&pxor	($tweak,$twres);
+
+	&sub	($len,16*6);
+	&jnc	(&label("xts_enc_loop6"));
+
+	&lea	($rounds,&DWP(1,"",$rounds,2));	# restore $rounds
+	&mov	($key,$key_);			# restore $key
+	&mov	($rounds_,$rounds);
+
+&set_label("xts_enc_short");
+	&add	($len,16*6);
+	&jz	(&label("xts_enc_done6x"));
+
+	&movdqa	($inout3,$tweak);		# put aside previous tweak
+	&cmp	($len,0x20);
+	&jb	(&label("xts_enc_one"));
+
+	&pshufd	($twres,$twtmp,0x13);
+	&pxor	($twtmp,$twtmp);
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($twres,$twmask);		# isolate carry and residue
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&pxor	($tweak,$twres);
+	&je	(&label("xts_enc_two"));
+
+	&pshufd	($twres,$twtmp,0x13);
+	&pxor	($twtmp,$twtmp);
+	&movdqa	($inout4,$tweak);		# put aside previous tweak
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($twres,$twmask);		# isolate carry and residue
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&pxor	($tweak,$twres);
+	&cmp	($len,0x40);
+	&jb	(&label("xts_enc_three"));
+
+	&pshufd	($twres,$twtmp,0x13);
+	&pxor	($twtmp,$twtmp);
+	&movdqa	($inout5,$tweak);		# put aside previous tweak
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($twres,$twmask);		# isolate carry and residue
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&pxor	($tweak,$twres);
+	&movdqa	(&QWP(16*0,"esp"),$inout3);
+	&movdqa	(&QWP(16*1,"esp"),$inout4);
+	&je	(&label("xts_enc_four"));
+
+	&movdqa	(&QWP(16*2,"esp"),$inout5);
+	&pshufd	($inout5,$twtmp,0x13);
+	&movdqa	(&QWP(16*3,"esp"),$tweak);
+	&paddq	($tweak,$tweak);		# &psllq($inout0,1);
+	&pand	($inout5,$twmask);		# isolate carry and residue
+	&pxor	($inout5,$tweak);
+
+	&movdqu	($inout0,&QWP(16*0,$inp));	# load input
+	&movdqu	($inout1,&QWP(16*1,$inp));
+	&movdqu	($inout2,&QWP(16*2,$inp));
+	&pxor	($inout0,&QWP(16*0,"esp"));	# input^=tweak
+	&movdqu	($inout3,&QWP(16*3,$inp));
+	&pxor	($inout1,&QWP(16*1,"esp"));
+	&movdqu	($inout4,&QWP(16*4,$inp));
+	&pxor	($inout2,&QWP(16*2,"esp"));
+	&lea	($inp,&DWP(16*5,$inp));
+	&pxor	($inout3,&QWP(16*3,"esp"));
+	&movdqa	(&QWP(16*4,"esp"),$inout5);	# save last tweak
+	&pxor	($inout4,$inout5);
+
+	&call	("_aesni_encrypt6");
+
+	&movaps	($tweak,&QWP(16*4,"esp"));	# last tweak
+	&xorps	($inout0,&QWP(16*0,"esp"));	# output^=tweak
+	&xorps	($inout1,&QWP(16*1,"esp"));
+	&xorps	($inout2,&QWP(16*2,"esp"));
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&xorps	($inout3,&QWP(16*3,"esp"));
+	&movups	(&QWP(16*1,$out),$inout1);
+	&xorps	($inout4,$tweak);
+	&movups	(&QWP(16*2,$out),$inout2);
+	&movups	(&QWP(16*3,$out),$inout3);
+	&movups	(&QWP(16*4,$out),$inout4);
+	&lea	($out,&DWP(16*5,$out));
+	&jmp	(&label("xts_enc_done"));
+
+&set_label("xts_enc_one",16);
+	&movups	($inout0,&QWP(16*0,$inp));	# load input
+	&lea	($inp,&DWP(16*1,$inp));
+	&xorps	($inout0,$inout3);		# input^=tweak
+	if ($inline)
+	{   &aesni_inline_generate1("enc");	}
+	else
+	{   &call	("_aesni_encrypt1");	}
+	&xorps	($inout0,$inout3);		# output^=tweak
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&lea	($out,&DWP(16*1,$out));
+
+	&movdqa	($tweak,$inout3);		# last tweak
+	&jmp	(&label("xts_enc_done"));
+
+&set_label("xts_enc_two",16);
+	&movaps	($inout4,$tweak);		# put aside last tweak
+
+	&movups	($inout0,&QWP(16*0,$inp));	# load input
+	&movups	($inout1,&QWP(16*1,$inp));
+	&lea	($inp,&DWP(16*2,$inp));
+	&xorps	($inout0,$inout3);		# input^=tweak
+	&xorps	($inout1,$inout4);
+	&xorps	($inout2,$inout2);
+
+	&call	("_aesni_encrypt3");
+
+	&xorps	($inout0,$inout3);		# output^=tweak
+	&xorps	($inout1,$inout4);
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&movups	(&QWP(16*1,$out),$inout1);
+	&lea	($out,&DWP(16*2,$out));
+
+	&movdqa	($tweak,$inout4);		# last tweak
+	&jmp	(&label("xts_enc_done"));
+
+&set_label("xts_enc_three",16);
+	&movaps	($inout5,$tweak);		# put aside last tweak
+	&movups	($inout0,&QWP(16*0,$inp));	# load input
+	&movups	($inout1,&QWP(16*1,$inp));
+	&movups	($inout2,&QWP(16*2,$inp));
+	&lea	($inp,&DWP(16*3,$inp));
+	&xorps	($inout0,$inout3);		# input^=tweak
+	&xorps	($inout1,$inout4);
+	&xorps	($inout2,$inout5);
+
+	&call	("_aesni_encrypt3");
+
+	&xorps	($inout0,$inout3);		# output^=tweak
+	&xorps	($inout1,$inout4);
+	&xorps	($inout2,$inout5);
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&movups	(&QWP(16*1,$out),$inout1);
+	&movups	(&QWP(16*2,$out),$inout2);
+	&lea	($out,&DWP(16*3,$out));
+
+	&movdqa	($tweak,$inout5);		# last tweak
+	&jmp	(&label("xts_enc_done"));
+
+&set_label("xts_enc_four",16);
+	&movaps	($inout4,$tweak);		# put aside last tweak
+
+	&movups	($inout0,&QWP(16*0,$inp));	# load input
+	&movups	($inout1,&QWP(16*1,$inp));
+	&movups	($inout2,&QWP(16*2,$inp));
+	&xorps	($inout0,&QWP(16*0,"esp"));	# input^=tweak
+	&movups	($inout3,&QWP(16*3,$inp));
+	&lea	($inp,&DWP(16*4,$inp));
+	&xorps	($inout1,&QWP(16*1,"esp"));
+	&xorps	($inout2,$inout5);
+	&xorps	($inout3,$inout4);
+
+	&call	("_aesni_encrypt4");
+
+	&xorps	($inout0,&QWP(16*0,"esp"));	# output^=tweak
+	&xorps	($inout1,&QWP(16*1,"esp"));
+	&xorps	($inout2,$inout5);
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&xorps	($inout3,$inout4);
+	&movups	(&QWP(16*1,$out),$inout1);
+	&movups	(&QWP(16*2,$out),$inout2);
+	&movups	(&QWP(16*3,$out),$inout3);
+	&lea	($out,&DWP(16*4,$out));
+
+	&movdqa	($tweak,$inout4);		# last tweak
+	&jmp	(&label("xts_enc_done"));
+
+&set_label("xts_enc_done6x",16);		# $tweak is pre-calculated
+	&mov	($len,&DWP(16*7+0,"esp"));	# restore original $len
+	&and	($len,15);
+	&jz	(&label("xts_enc_ret"));
+	&movdqa	($inout3,$tweak);
+	&mov	(&DWP(16*7+0,"esp"),$len);	# save $len%16
+	&jmp	(&label("xts_enc_steal"));
+
+&set_label("xts_enc_done",16);
+	&mov	($len,&DWP(16*7+0,"esp"));	# restore original $len
+	&pxor	($twtmp,$twtmp);
+	&and	($len,15);
+	&jz	(&label("xts_enc_ret"));
+
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&mov	(&DWP(16*7+0,"esp"),$len);	# save $len%16
+	&pshufd	($inout3,$twtmp,0x13);
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($inout3,&QWP(16*6,"esp"));	# isolate carry and residue
+	&pxor	($inout3,$tweak);
+
+&set_label("xts_enc_steal");
+	&movz	($rounds,&BP(0,$inp));
+	&movz	($key,&BP(-16,$out));
+	&lea	($inp,&DWP(1,$inp));
+	&mov	(&BP(-16,$out),&LB($rounds));
+	&mov	(&BP(0,$out),&LB($key));
+	&lea	($out,&DWP(1,$out));
+	&sub	($len,1);
+	&jnz	(&label("xts_enc_steal"));
+
+	&sub	($out,&DWP(16*7+0,"esp"));	# rewind $out
+	&mov	($key,$key_);			# restore $key
+	&mov	($rounds,$rounds_);		# restore $rounds
+
+	&movups	($inout0,&QWP(-16,$out));	# load input
+	&xorps	($inout0,$inout3);		# input^=tweak
+	if ($inline)
+	{   &aesni_inline_generate1("enc");	}
+	else
+	{   &call	("_aesni_encrypt1");	}
+	&xorps	($inout0,$inout3);		# output^=tweak
+	&movups	(&QWP(-16,$out),$inout0);	# write output
+
+&set_label("xts_enc_ret");
+	&mov	("esp",&DWP(16*7+4,"esp"));	# restore %esp
+&function_end("aesni_xts_encrypt");
+
+&function_begin("aesni_xts_decrypt");
+	&mov	($key,&wparam(4));		# key2
+	&mov	($inp,&wparam(5));		# clear-text tweak
+
+	&mov	($rounds,&DWP(240,$key));	# key2->rounds
+	&movups	($inout0,&QWP(0,$inp));
+	if ($inline)
+	{   &aesni_inline_generate1("enc");	}
+	else
+	{   &call	("_aesni_encrypt1");	}
+
+	&mov	($inp,&wparam(0));
+	&mov	($out,&wparam(1));
+	&mov	($len,&wparam(2));
+	&mov	($key,&wparam(3));		# key1
+
+	&mov	($key_,"esp");
+	&sub	("esp",16*7+8);
+	&and	("esp",-16);			# align stack
+
+	&xor	($rounds_,$rounds_);		# if(len%16) len-=16;
+	&test	($len,15);
+	&setnz	(&LB($rounds_));
+	&shl	($rounds_,4);
+	&sub	($len,$rounds_);
+
+	&mov	(&DWP(16*6+0,"esp"),0x87);	# compose the magic constant
+	&mov	(&DWP(16*6+4,"esp"),0);
+	&mov	(&DWP(16*6+8,"esp"),1);
+	&mov	(&DWP(16*6+12,"esp"),0);
+	&mov	(&DWP(16*7+0,"esp"),$len);	# save original $len
+	&mov	(&DWP(16*7+4,"esp"),$key_);	# save original %esp
+
+	&mov	($rounds,&DWP(240,$key));	# key1->rounds
+	&mov	($key_,$key);			# backup $key
+	&mov	($rounds_,$rounds);		# backup $rounds
+
+	&movdqa	($tweak,$inout0);
+	&pxor	($twtmp,$twtmp);
+	&movdqa	($twmask,&QWP(6*16,"esp"));	# 0x0...010...87
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+
+	&and	($len,-16);
+	&sub	($len,16*6);
+	&jc	(&label("xts_dec_short"));
+
+	&shr	($rounds,1);
+	&mov	($rounds_,$rounds);
+	&jmp	(&label("xts_dec_loop6"));
+
+&set_label("xts_dec_loop6",16);
+	for ($i=0;$i<4;$i++) {
+	    &pshufd	($twres,$twtmp,0x13);
+	    &pxor	($twtmp,$twtmp);
+	    &movdqa	(&QWP(16*$i,"esp"),$tweak);
+	    &paddq	($tweak,$tweak);	# &psllq($tweak,1);
+	    &pand	($twres,$twmask);	# isolate carry and residue
+	    &pcmpgtd	($twtmp,$tweak);	# broadcast upper bits
+	    &pxor	($tweak,$twres);
+	}
+	&pshufd	($inout5,$twtmp,0x13);
+	&movdqa	(&QWP(16*$i++,"esp"),$tweak);
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	 &$movekey	($rndkey0,&QWP(0,$key_));
+	&pand	($inout5,$twmask);		# isolate carry and residue
+	 &movups	($inout0,&QWP(0,$inp));	# load input
+	&pxor	($inout5,$tweak);
+
+	# inline _aesni_encrypt6 prologue and flip xor with tweak and key[0]
+	&movdqu	($inout1,&QWP(16*1,$inp));
+	 &xorps		($inout0,$rndkey0);	# input^=rndkey[0]
+	&movdqu	($inout2,&QWP(16*2,$inp));
+	 &pxor		($inout1,$rndkey0);
+	&movdqu	($inout3,&QWP(16*3,$inp));
+	 &pxor		($inout2,$rndkey0);
+	&movdqu	($inout4,&QWP(16*4,$inp));
+	 &pxor		($inout3,$rndkey0);
+	&movdqu	($rndkey1,&QWP(16*5,$inp));
+	 &pxor		($inout4,$rndkey0);
+	&lea	($inp,&DWP(16*6,$inp));
+	&pxor	($inout0,&QWP(16*0,"esp"));	# input^=tweak
+	&movdqa	(&QWP(16*$i,"esp"),$inout5);	# save last tweak
+	&pxor	($inout5,$rndkey1);
+
+	 &$movekey	($rndkey1,&QWP(16,$key_));
+	 &lea		($key,&DWP(32,$key_));
+	&pxor	($inout1,&QWP(16*1,"esp"));
+	 &aesdec	($inout0,$rndkey1);
+	&pxor	($inout2,&QWP(16*2,"esp"));
+	 &aesdec	($inout1,$rndkey1);
+	&pxor	($inout3,&QWP(16*3,"esp"));
+	 &dec		($rounds);
+	 &aesdec	($inout2,$rndkey1);
+	&pxor	($inout4,&QWP(16*4,"esp"));
+	 &aesdec	($inout3,$rndkey1);
+	&pxor		($inout5,$rndkey0);
+	 &aesdec	($inout4,$rndkey1);
+	 &$movekey	($rndkey0,&QWP(0,$key));
+	 &aesdec	($inout5,$rndkey1);
+	&call		(&label("_aesni_decrypt6_enter"));
+
+	&movdqa	($tweak,&QWP(16*5,"esp"));	# last tweak
+       &pxor	($twtmp,$twtmp);
+	&xorps	($inout0,&QWP(16*0,"esp"));	# output^=tweak
+       &pcmpgtd	($twtmp,$tweak);		# broadcast upper bits
+	&xorps	($inout1,&QWP(16*1,"esp"));
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&xorps	($inout2,&QWP(16*2,"esp"));
+	&movups	(&QWP(16*1,$out),$inout1);
+	&xorps	($inout3,&QWP(16*3,"esp"));
+	&movups	(&QWP(16*2,$out),$inout2);
+	&xorps	($inout4,&QWP(16*4,"esp"));
+	&movups	(&QWP(16*3,$out),$inout3);
+	&xorps	($inout5,$tweak);
+	&movups	(&QWP(16*4,$out),$inout4);
+       &pshufd	($twres,$twtmp,0x13);
+	&movups	(&QWP(16*5,$out),$inout5);
+	&lea	($out,&DWP(16*6,$out));
+       &movdqa	($twmask,&QWP(16*6,"esp"));	# 0x0...010...87
+
+	&pxor	($twtmp,$twtmp);
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($twres,$twmask);		# isolate carry and residue
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&mov	($rounds,$rounds_);		# restore $rounds
+	&pxor	($tweak,$twres);
+
+	&sub	($len,16*6);
+	&jnc	(&label("xts_dec_loop6"));
+
+	&lea	($rounds,&DWP(1,"",$rounds,2));	# restore $rounds
+	&mov	($key,$key_);			# restore $key
+	&mov	($rounds_,$rounds);
+
+&set_label("xts_dec_short");
+	&add	($len,16*6);
+	&jz	(&label("xts_dec_done6x"));
+
+	&movdqa	($inout3,$tweak);		# put aside previous tweak
+	&cmp	($len,0x20);
+	&jb	(&label("xts_dec_one"));
+
+	&pshufd	($twres,$twtmp,0x13);
+	&pxor	($twtmp,$twtmp);
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($twres,$twmask);		# isolate carry and residue
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&pxor	($tweak,$twres);
+	&je	(&label("xts_dec_two"));
+
+	&pshufd	($twres,$twtmp,0x13);
+	&pxor	($twtmp,$twtmp);
+	&movdqa	($inout4,$tweak);		# put aside previous tweak
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($twres,$twmask);		# isolate carry and residue
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&pxor	($tweak,$twres);
+	&cmp	($len,0x40);
+	&jb	(&label("xts_dec_three"));
+
+	&pshufd	($twres,$twtmp,0x13);
+	&pxor	($twtmp,$twtmp);
+	&movdqa	($inout5,$tweak);		# put aside previous tweak
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($twres,$twmask);		# isolate carry and residue
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&pxor	($tweak,$twres);
+	&movdqa	(&QWP(16*0,"esp"),$inout3);
+	&movdqa	(&QWP(16*1,"esp"),$inout4);
+	&je	(&label("xts_dec_four"));
+
+	&movdqa	(&QWP(16*2,"esp"),$inout5);
+	&pshufd	($inout5,$twtmp,0x13);
+	&movdqa	(&QWP(16*3,"esp"),$tweak);
+	&paddq	($tweak,$tweak);		# &psllq($inout0,1);
+	&pand	($inout5,$twmask);		# isolate carry and residue
+	&pxor	($inout5,$tweak);
+
+	&movdqu	($inout0,&QWP(16*0,$inp));	# load input
+	&movdqu	($inout1,&QWP(16*1,$inp));
+	&movdqu	($inout2,&QWP(16*2,$inp));
+	&pxor	($inout0,&QWP(16*0,"esp"));	# input^=tweak
+	&movdqu	($inout3,&QWP(16*3,$inp));
+	&pxor	($inout1,&QWP(16*1,"esp"));
+	&movdqu	($inout4,&QWP(16*4,$inp));
+	&pxor	($inout2,&QWP(16*2,"esp"));
+	&lea	($inp,&DWP(16*5,$inp));
+	&pxor	($inout3,&QWP(16*3,"esp"));
+	&movdqa	(&QWP(16*4,"esp"),$inout5);	# save last tweak
+	&pxor	($inout4,$inout5);
+
+	&call	("_aesni_decrypt6");
+
+	&movaps	($tweak,&QWP(16*4,"esp"));	# last tweak
+	&xorps	($inout0,&QWP(16*0,"esp"));	# output^=tweak
+	&xorps	($inout1,&QWP(16*1,"esp"));
+	&xorps	($inout2,&QWP(16*2,"esp"));
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&xorps	($inout3,&QWP(16*3,"esp"));
+	&movups	(&QWP(16*1,$out),$inout1);
+	&xorps	($inout4,$tweak);
+	&movups	(&QWP(16*2,$out),$inout2);
+	&movups	(&QWP(16*3,$out),$inout3);
+	&movups	(&QWP(16*4,$out),$inout4);
+	&lea	($out,&DWP(16*5,$out));
+	&jmp	(&label("xts_dec_done"));
+
+&set_label("xts_dec_one",16);
+	&movups	($inout0,&QWP(16*0,$inp));	# load input
+	&lea	($inp,&DWP(16*1,$inp));
+	&xorps	($inout0,$inout3);		# input^=tweak
+	if ($inline)
+	{   &aesni_inline_generate1("dec");	}
+	else
+	{   &call	("_aesni_decrypt1");	}
+	&xorps	($inout0,$inout3);		# output^=tweak
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&lea	($out,&DWP(16*1,$out));
+
+	&movdqa	($tweak,$inout3);		# last tweak
+	&jmp	(&label("xts_dec_done"));
+
+&set_label("xts_dec_two",16);
+	&movaps	($inout4,$tweak);		# put aside last tweak
+
+	&movups	($inout0,&QWP(16*0,$inp));	# load input
+	&movups	($inout1,&QWP(16*1,$inp));
+	&lea	($inp,&DWP(16*2,$inp));
+	&xorps	($inout0,$inout3);		# input^=tweak
+	&xorps	($inout1,$inout4);
+
+	&call	("_aesni_decrypt3");
+
+	&xorps	($inout0,$inout3);		# output^=tweak
+	&xorps	($inout1,$inout4);
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&movups	(&QWP(16*1,$out),$inout1);
+	&lea	($out,&DWP(16*2,$out));
+
+	&movdqa	($tweak,$inout4);		# last tweak
+	&jmp	(&label("xts_dec_done"));
+
+&set_label("xts_dec_three",16);
+	&movaps	($inout5,$tweak);		# put aside last tweak
+	&movups	($inout0,&QWP(16*0,$inp));	# load input
+	&movups	($inout1,&QWP(16*1,$inp));
+	&movups	($inout2,&QWP(16*2,$inp));
+	&lea	($inp,&DWP(16*3,$inp));
+	&xorps	($inout0,$inout3);		# input^=tweak
+	&xorps	($inout1,$inout4);
+	&xorps	($inout2,$inout5);
+
+	&call	("_aesni_decrypt3");
+
+	&xorps	($inout0,$inout3);		# output^=tweak
+	&xorps	($inout1,$inout4);
+	&xorps	($inout2,$inout5);
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&movups	(&QWP(16*1,$out),$inout1);
+	&movups	(&QWP(16*2,$out),$inout2);
+	&lea	($out,&DWP(16*3,$out));
+
+	&movdqa	($tweak,$inout5);		# last tweak
+	&jmp	(&label("xts_dec_done"));
+
+&set_label("xts_dec_four",16);
+	&movaps	($inout4,$tweak);		# put aside last tweak
+
+	&movups	($inout0,&QWP(16*0,$inp));	# load input
+	&movups	($inout1,&QWP(16*1,$inp));
+	&movups	($inout2,&QWP(16*2,$inp));
+	&xorps	($inout0,&QWP(16*0,"esp"));	# input^=tweak
+	&movups	($inout3,&QWP(16*3,$inp));
+	&lea	($inp,&DWP(16*4,$inp));
+	&xorps	($inout1,&QWP(16*1,"esp"));
+	&xorps	($inout2,$inout5);
+	&xorps	($inout3,$inout4);
+
+	&call	("_aesni_decrypt4");
+
+	&xorps	($inout0,&QWP(16*0,"esp"));	# output^=tweak
+	&xorps	($inout1,&QWP(16*1,"esp"));
+	&xorps	($inout2,$inout5);
+	&movups	(&QWP(16*0,$out),$inout0);	# write output
+	&xorps	($inout3,$inout4);
+	&movups	(&QWP(16*1,$out),$inout1);
+	&movups	(&QWP(16*2,$out),$inout2);
+	&movups	(&QWP(16*3,$out),$inout3);
+	&lea	($out,&DWP(16*4,$out));
+
+	&movdqa	($tweak,$inout4);		# last tweak
+	&jmp	(&label("xts_dec_done"));
+
+&set_label("xts_dec_done6x",16);		# $tweak is pre-calculated
+	&mov	($len,&DWP(16*7+0,"esp"));	# restore original $len
+	&and	($len,15);
+	&jz	(&label("xts_dec_ret"));
+	&mov	(&DWP(16*7+0,"esp"),$len);	# save $len%16
+	&jmp	(&label("xts_dec_only_one_more"));
+
+&set_label("xts_dec_done",16);
+	&mov	($len,&DWP(16*7+0,"esp"));	# restore original $len
+	&pxor	($twtmp,$twtmp);
+	&and	($len,15);
+	&jz	(&label("xts_dec_ret"));
+
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&mov	(&DWP(16*7+0,"esp"),$len);	# save $len%16
+	&pshufd	($twres,$twtmp,0x13);
+	&pxor	($twtmp,$twtmp);
+	&movdqa	($twmask,&QWP(16*6,"esp"));
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($twres,$twmask);		# isolate carry and residue
+	&pcmpgtd($twtmp,$tweak);		# broadcast upper bits
+	&pxor	($tweak,$twres);
+
+&set_label("xts_dec_only_one_more");
+	&pshufd	($inout3,$twtmp,0x13);
+	&movdqa	($inout4,$tweak);		# put aside previous tweak
+	&paddq	($tweak,$tweak);		# &psllq($tweak,1);
+	&pand	($inout3,$twmask);		# isolate carry and residue
+	&pxor	($inout3,$tweak);
+
+	&mov	($key,$key_);			# restore $key
+	&mov	($rounds,$rounds_);		# restore $rounds
+
+	&movups	($inout0,&QWP(0,$inp));		# load input
+	&xorps	($inout0,$inout3);		# input^=tweak
+	if ($inline)
+	{   &aesni_inline_generate1("dec");	}
+	else
+	{   &call	("_aesni_decrypt1");	}
+	&xorps	($inout0,$inout3);		# output^=tweak
+	&movups	(&QWP(0,$out),$inout0);		# write output
+
+&set_label("xts_dec_steal");
+	&movz	($rounds,&BP(16,$inp));
+	&movz	($key,&BP(0,$out));
+	&lea	($inp,&DWP(1,$inp));
+	&mov	(&BP(0,$out),&LB($rounds));
+	&mov	(&BP(16,$out),&LB($key));
+	&lea	($out,&DWP(1,$out));
+	&sub	($len,1);
+	&jnz	(&label("xts_dec_steal"));
+
+	&sub	($out,&DWP(16*7+0,"esp"));	# rewind $out
+	&mov	($key,$key_);			# restore $key
+	&mov	($rounds,$rounds_);		# restore $rounds
+
+	&movups	($inout0,&QWP(0,$out));		# load input
+	&xorps	($inout0,$inout4);		# input^=tweak
+	if ($inline)
+	{   &aesni_inline_generate1("dec");	}
+	else
+	{   &call	("_aesni_decrypt1");	}
+	&xorps	($inout0,$inout4);		# output^=tweak
+	&movups	(&QWP(0,$out),$inout0);		# write output
+
+&set_label("xts_dec_ret");
+	&mov	("esp",&DWP(16*7+4,"esp"));	# restore %esp
+&function_end("aesni_xts_decrypt");
+}
+}
+
+######################################################################
+# void $PREFIX_cbc_encrypt (const void *inp, void *out,
+#                           size_t length, const AES_KEY *key,
+#                           unsigned char *ivp,const int enc);
+&function_begin("${PREFIX}_cbc_encrypt");
+	&mov	($inp,&wparam(0));
+	&mov	($rounds_,"esp");
+	&mov	($out,&wparam(1));
+	&sub	($rounds_,24);
+	&mov	($len,&wparam(2));
+	&and	($rounds_,-16);
+	&mov	($key,&wparam(3));
+	&mov	($key_,&wparam(4));
+	&test	($len,$len);
+	&jz	(&label("cbc_abort"));
+
+	&cmp	(&wparam(5),0);
+	&xchg	($rounds_,"esp");		# alloca
+	&movups	($ivec,&QWP(0,$key_));		# load IV
+	&mov	($rounds,&DWP(240,$key));
+	&mov	($key_,$key);			# backup $key
+	&mov	(&DWP(16,"esp"),$rounds_);	# save original %esp
+	&mov	($rounds_,$rounds);		# backup $rounds
+	&je	(&label("cbc_decrypt"));
+
+	&movaps	($inout0,$ivec);
+	&cmp	($len,16);
+	&jb	(&label("cbc_enc_tail"));
+	&sub	($len,16);
+	&jmp	(&label("cbc_enc_loop"));
+
+&set_label("cbc_enc_loop",16);
+	&movups	($ivec,&QWP(0,$inp));		# input actually
+	&lea	($inp,&DWP(16,$inp));
+	if ($inline)
+	{   &aesni_inline_generate1("enc",$inout0,$ivec);	}
+	else
+	{   &xorps($inout0,$ivec); &call("_aesni_encrypt1");	}
+	&mov	($rounds,$rounds_);	# restore $rounds
+	&mov	($key,$key_);		# restore $key
+	&movups	(&QWP(0,$out),$inout0);	# store output
+	&lea	($out,&DWP(16,$out));
+	&sub	($len,16);
+	&jnc	(&label("cbc_enc_loop"));
+	&add	($len,16);
+	&jnz	(&label("cbc_enc_tail"));
+	&movaps	($ivec,$inout0);
+	&jmp	(&label("cbc_ret"));
+
+&set_label("cbc_enc_tail");
+	&mov	("ecx",$len);		# zaps $rounds
+	&data_word(0xA4F3F689);		# rep movsb
+	&mov	("ecx",16);		# zero tail
+	&sub	("ecx",$len);
+	&xor	("eax","eax");		# zaps $len
+	&data_word(0xAAF3F689);		# rep stosb
+	&lea	($out,&DWP(-16,$out));	# rewind $out by 1 block
+	&mov	($rounds,$rounds_);	# restore $rounds
+	&mov	($inp,$out);		# $inp and $out are the same
+	&mov	($key,$key_);		# restore $key
+	&jmp	(&label("cbc_enc_loop"));
+######################################################################
+&set_label("cbc_decrypt",16);
+	&cmp	($len,0x50);
+	&jbe	(&label("cbc_dec_tail"));
+	&movaps	(&QWP(0,"esp"),$ivec);		# save IV
+	&sub	($len,0x50);
+	&jmp	(&label("cbc_dec_loop6_enter"));
+
+&set_label("cbc_dec_loop6",16);
+	&movaps	(&QWP(0,"esp"),$rndkey0);	# save IV
+	&movups	(&QWP(0,$out),$inout5);
+	&lea	($out,&DWP(0x10,$out));
+&set_label("cbc_dec_loop6_enter");
+	&movdqu	($inout0,&QWP(0,$inp));
+	&movdqu	($inout1,&QWP(0x10,$inp));
+	&movdqu	($inout2,&QWP(0x20,$inp));
+	&movdqu	($inout3,&QWP(0x30,$inp));
+	&movdqu	($inout4,&QWP(0x40,$inp));
+	&movdqu	($inout5,&QWP(0x50,$inp));
+
+	&call	("_aesni_decrypt6");
+
+	&movups	($rndkey1,&QWP(0,$inp));
+	&movups	($rndkey0,&QWP(0x10,$inp));
+	&xorps	($inout0,&QWP(0,"esp"));	# ^=IV
+	&xorps	($inout1,$rndkey1);
+	&movups	($rndkey1,&QWP(0x20,$inp));
+	&xorps	($inout2,$rndkey0);
+	&movups	($rndkey0,&QWP(0x30,$inp));
+	&xorps	($inout3,$rndkey1);
+	&movups	($rndkey1,&QWP(0x40,$inp));
+	&xorps	($inout4,$rndkey0);
+	&movups	($rndkey0,&QWP(0x50,$inp));	# IV
+	&xorps	($inout5,$rndkey1);
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&lea	($inp,&DWP(0x60,$inp));
+	&movups	(&QWP(0x20,$out),$inout2);
+	&mov	($rounds,$rounds_)		# restore $rounds
+	&movups	(&QWP(0x30,$out),$inout3);
+	&mov	($key,$key_);			# restore $key
+	&movups	(&QWP(0x40,$out),$inout4);
+	&lea	($out,&DWP(0x50,$out));
+	&sub	($len,0x60);
+	&ja	(&label("cbc_dec_loop6"));
+
+	&movaps	($inout0,$inout5);
+	&movaps	($ivec,$rndkey0);
+	&add	($len,0x50);
+	&jle	(&label("cbc_dec_tail_collected"));
+	&movups	(&QWP(0,$out),$inout0);
+	&lea	($out,&DWP(0x10,$out));
+&set_label("cbc_dec_tail");
+	&movups	($inout0,&QWP(0,$inp));
+	&movaps	($in0,$inout0);
+	&cmp	($len,0x10);
+	&jbe	(&label("cbc_dec_one"));
+
+	&movups	($inout1,&QWP(0x10,$inp));
+	&movaps	($in1,$inout1);
+	&cmp	($len,0x20);
+	&jbe	(&label("cbc_dec_two"));
+
+	&movups	($inout2,&QWP(0x20,$inp));
+	&cmp	($len,0x30);
+	&jbe	(&label("cbc_dec_three"));
+
+	&movups	($inout3,&QWP(0x30,$inp));
+	&cmp	($len,0x40);
+	&jbe	(&label("cbc_dec_four"));
+
+	&movups	($inout4,&QWP(0x40,$inp));
+	&movaps	(&QWP(0,"esp"),$ivec);		# save IV
+	&movups	($inout0,&QWP(0,$inp));
+	&xorps	($inout5,$inout5);
+	&call	("_aesni_decrypt6");
+	&movups	($rndkey1,&QWP(0,$inp));
+	&movups	($rndkey0,&QWP(0x10,$inp));
+	&xorps	($inout0,&QWP(0,"esp"));	# ^= IV
+	&xorps	($inout1,$rndkey1);
+	&movups	($rndkey1,&QWP(0x20,$inp));
+	&xorps	($inout2,$rndkey0);
+	&movups	($rndkey0,&QWP(0x30,$inp));
+	&xorps	($inout3,$rndkey1);
+	&movups	($ivec,&QWP(0x40,$inp));	# IV
+	&xorps	($inout4,$rndkey0);
+	&movups	(&QWP(0,$out),$inout0);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&movups	(&QWP(0x30,$out),$inout3);
+	&lea	($out,&DWP(0x40,$out));
+	&movaps	($inout0,$inout4);
+	&sub	($len,0x50);
+	&jmp	(&label("cbc_dec_tail_collected"));
+
+&set_label("cbc_dec_one",16);
+	if ($inline)
+	{   &aesni_inline_generate1("dec");	}
+	else
+	{   &call	("_aesni_decrypt1");	}
+	&xorps	($inout0,$ivec);
+	&movaps	($ivec,$in0);
+	&sub	($len,0x10);
+	&jmp	(&label("cbc_dec_tail_collected"));
+
+&set_label("cbc_dec_two",16);
+	&xorps	($inout2,$inout2);
+	&call	("_aesni_decrypt3");
+	&xorps	($inout0,$ivec);
+	&xorps	($inout1,$in0);
+	&movups	(&QWP(0,$out),$inout0);
+	&movaps	($inout0,$inout1);
+	&lea	($out,&DWP(0x10,$out));
+	&movaps	($ivec,$in1);
+	&sub	($len,0x20);
+	&jmp	(&label("cbc_dec_tail_collected"));
+
+&set_label("cbc_dec_three",16);
+	&call	("_aesni_decrypt3");
+	&xorps	($inout0,$ivec);
+	&xorps	($inout1,$in0);
+	&xorps	($inout2,$in1);
+	&movups	(&QWP(0,$out),$inout0);
+	&movaps	($inout0,$inout2);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&lea	($out,&DWP(0x20,$out));
+	&movups	($ivec,&QWP(0x20,$inp));
+	&sub	($len,0x30);
+	&jmp	(&label("cbc_dec_tail_collected"));
+
+&set_label("cbc_dec_four",16);
+	&call	("_aesni_decrypt4");
+	&movups	($rndkey1,&QWP(0x10,$inp));
+	&movups	($rndkey0,&QWP(0x20,$inp));
+	&xorps	($inout0,$ivec);
+	&movups	($ivec,&QWP(0x30,$inp));
+	&xorps	($inout1,$in0);
+	&movups	(&QWP(0,$out),$inout0);
+	&xorps	($inout2,$rndkey1);
+	&movups	(&QWP(0x10,$out),$inout1);
+	&xorps	($inout3,$rndkey0);
+	&movups	(&QWP(0x20,$out),$inout2);
+	&lea	($out,&DWP(0x30,$out));
+	&movaps	($inout0,$inout3);
+	&sub	($len,0x40);
+
+&set_label("cbc_dec_tail_collected");
+	&and	($len,15);
+	&jnz	(&label("cbc_dec_tail_partial"));
+	&movups	(&QWP(0,$out),$inout0);
+	&jmp	(&label("cbc_ret"));
+
+&set_label("cbc_dec_tail_partial",16);
+	&movaps	(&QWP(0,"esp"),$inout0);
+	&mov	("ecx",16);
+	&mov	($inp,"esp");
+	&sub	("ecx",$len);
+	&data_word(0xA4F3F689);		# rep movsb
+
+&set_label("cbc_ret");
+	&mov	("esp",&DWP(16,"esp"));	# pull original %esp
+	&mov	($key_,&wparam(4));
+	&movups	(&QWP(0,$key_),$ivec);	# output IV
+&set_label("cbc_abort");
+&function_end("${PREFIX}_cbc_encrypt");
+
+######################################################################
+# Mechanical port from aesni-x86_64.pl.
+#
+# _aesni_set_encrypt_key is private interface,
+# input:
+#	"eax"	const unsigned char *userKey
+#	$rounds	int bits
+#	$key	AES_KEY *key
+# output:
+#	"eax"	return code
+#	$round	rounds
+
+&function_begin_B("_aesni_set_encrypt_key");
+	&test	("eax","eax");
+	&jz	(&label("bad_pointer"));
+	&test	($key,$key);
+	&jz	(&label("bad_pointer"));
+
+	&movups	("xmm0",&QWP(0,"eax"));	# pull first 128 bits of *userKey
+	&xorps	("xmm4","xmm4");	# low dword of xmm4 is assumed 0
+	&lea	($key,&DWP(16,$key));
+	&cmp	($rounds,256);
+	&je	(&label("14rounds"));
+	&cmp	($rounds,192);
+	&je	(&label("12rounds"));
+	&cmp	($rounds,128);
+	&jne	(&label("bad_keybits"));
+
+&set_label("10rounds",16);
+	&mov		($rounds,9);
+	&$movekey	(&QWP(-16,$key),"xmm0");	# round 0
+	&aeskeygenassist("xmm1","xmm0",0x01);		# round 1
+	&call		(&label("key_128_cold"));
+	&aeskeygenassist("xmm1","xmm0",0x2);		# round 2
+	&call		(&label("key_128"));
+	&aeskeygenassist("xmm1","xmm0",0x04);		# round 3
+	&call		(&label("key_128"));
+	&aeskeygenassist("xmm1","xmm0",0x08);		# round 4
+	&call		(&label("key_128"));
+	&aeskeygenassist("xmm1","xmm0",0x10);		# round 5
+	&call		(&label("key_128"));
+	&aeskeygenassist("xmm1","xmm0",0x20);		# round 6
+	&call		(&label("key_128"));
+	&aeskeygenassist("xmm1","xmm0",0x40);		# round 7
+	&call		(&label("key_128"));
+	&aeskeygenassist("xmm1","xmm0",0x80);		# round 8
+	&call		(&label("key_128"));
+	&aeskeygenassist("xmm1","xmm0",0x1b);		# round 9
+	&call		(&label("key_128"));
+	&aeskeygenassist("xmm1","xmm0",0x36);		# round 10
+	&call		(&label("key_128"));
+	&$movekey	(&QWP(0,$key),"xmm0");
+	&mov		(&DWP(80,$key),$rounds);
+	&xor		("eax","eax");
+	&ret();
+
+&set_label("key_128",16);
+	&$movekey	(&QWP(0,$key),"xmm0");
+	&lea		($key,&DWP(16,$key));
+&set_label("key_128_cold");
+	&shufps		("xmm4","xmm0",0b00010000);
+	&xorps		("xmm0","xmm4");
+	&shufps		("xmm4","xmm0",0b10001100);
+	&xorps		("xmm0","xmm4");
+	&shufps		("xmm1","xmm1",0b11111111);	# critical path
+	&xorps		("xmm0","xmm1");
+	&ret();
+
+&set_label("12rounds",16);
+	&movq		("xmm2",&QWP(16,"eax"));	# remaining 1/3 of *userKey
+	&mov		($rounds,11);
+	&$movekey	(&QWP(-16,$key),"xmm0")		# round 0
+	&aeskeygenassist("xmm1","xmm2",0x01);		# round 1,2
+	&call		(&label("key_192a_cold"));
+	&aeskeygenassist("xmm1","xmm2",0x02);		# round 2,3
+	&call		(&label("key_192b"));
+	&aeskeygenassist("xmm1","xmm2",0x04);		# round 4,5
+	&call		(&label("key_192a"));
+	&aeskeygenassist("xmm1","xmm2",0x08);		# round 5,6
+	&call		(&label("key_192b"));
+	&aeskeygenassist("xmm1","xmm2",0x10);		# round 7,8
+	&call		(&label("key_192a"));
+	&aeskeygenassist("xmm1","xmm2",0x20);		# round 8,9
+	&call		(&label("key_192b"));
+	&aeskeygenassist("xmm1","xmm2",0x40);		# round 10,11
+	&call		(&label("key_192a"));
+	&aeskeygenassist("xmm1","xmm2",0x80);		# round 11,12
+	&call		(&label("key_192b"));
+	&$movekey	(&QWP(0,$key),"xmm0");
+	&mov		(&DWP(48,$key),$rounds);
+	&xor		("eax","eax");
+	&ret();
+
+&set_label("key_192a",16);
+	&$movekey	(&QWP(0,$key),"xmm0");
+	&lea		($key,&DWP(16,$key));
+&set_label("key_192a_cold",16);
+	&movaps		("xmm5","xmm2");
+&set_label("key_192b_warm");
+	&shufps		("xmm4","xmm0",0b00010000);
+	&movdqa		("xmm3","xmm2");
+	&xorps		("xmm0","xmm4");
+	&shufps		("xmm4","xmm0",0b10001100);
+	&pslldq		("xmm3",4);
+	&xorps		("xmm0","xmm4");
+	&pshufd		("xmm1","xmm1",0b01010101);	# critical path
+	&pxor		("xmm2","xmm3");
+	&pxor		("xmm0","xmm1");
+	&pshufd		("xmm3","xmm0",0b11111111);
+	&pxor		("xmm2","xmm3");
+	&ret();
+
+&set_label("key_192b",16);
+	&movaps		("xmm3","xmm0");
+	&shufps		("xmm5","xmm0",0b01000100);
+	&$movekey	(&QWP(0,$key),"xmm5");
+	&shufps		("xmm3","xmm2",0b01001110);
+	&$movekey	(&QWP(16,$key),"xmm3");
+	&lea		($key,&DWP(32,$key));
+	&jmp		(&label("key_192b_warm"));
+
+&set_label("14rounds",16);
+	&movups		("xmm2",&QWP(16,"eax"));	# remaining half of *userKey
+	&mov		($rounds,13);
+	&lea		($key,&DWP(16,$key));
+	&$movekey	(&QWP(-32,$key),"xmm0");	# round 0
+	&$movekey	(&QWP(-16,$key),"xmm2");	# round 1
+	&aeskeygenassist("xmm1","xmm2",0x01);		# round 2
+	&call		(&label("key_256a_cold"));
+	&aeskeygenassist("xmm1","xmm0",0x01);		# round 3
+	&call		(&label("key_256b"));
+	&aeskeygenassist("xmm1","xmm2",0x02);		# round 4
+	&call		(&label("key_256a"));
+	&aeskeygenassist("xmm1","xmm0",0x02);		# round 5
+	&call		(&label("key_256b"));
+	&aeskeygenassist("xmm1","xmm2",0x04);		# round 6
+	&call		(&label("key_256a"));
+	&aeskeygenassist("xmm1","xmm0",0x04);		# round 7
+	&call		(&label("key_256b"));
+	&aeskeygenassist("xmm1","xmm2",0x08);		# round 8
+	&call		(&label("key_256a"));
+	&aeskeygenassist("xmm1","xmm0",0x08);		# round 9
+	&call		(&label("key_256b"));
+	&aeskeygenassist("xmm1","xmm2",0x10);		# round 10
+	&call		(&label("key_256a"));
+	&aeskeygenassist("xmm1","xmm0",0x10);		# round 11
+	&call		(&label("key_256b"));
+	&aeskeygenassist("xmm1","xmm2",0x20);		# round 12
+	&call		(&label("key_256a"));
+	&aeskeygenassist("xmm1","xmm0",0x20);		# round 13
+	&call		(&label("key_256b"));
+	&aeskeygenassist("xmm1","xmm2",0x40);		# round 14
+	&call		(&label("key_256a"));
+	&$movekey	(&QWP(0,$key),"xmm0");
+	&mov		(&DWP(16,$key),$rounds);
+	&xor		("eax","eax");
+	&ret();
+
+&set_label("key_256a",16);
+	&$movekey	(&QWP(0,$key),"xmm2");
+	&lea		($key,&DWP(16,$key));
+&set_label("key_256a_cold");
+	&shufps		("xmm4","xmm0",0b00010000);
+	&xorps		("xmm0","xmm4");
+	&shufps		("xmm4","xmm0",0b10001100);
+	&xorps		("xmm0","xmm4");
+	&shufps		("xmm1","xmm1",0b11111111);	# critical path
+	&xorps		("xmm0","xmm1");
+	&ret();
+
+&set_label("key_256b",16);
+	&$movekey	(&QWP(0,$key),"xmm0");
+	&lea		($key,&DWP(16,$key));
+
+	&shufps		("xmm4","xmm2",0b00010000);
+	&xorps		("xmm2","xmm4");
+	&shufps		("xmm4","xmm2",0b10001100);
+	&xorps		("xmm2","xmm4");
+	&shufps		("xmm1","xmm1",0b10101010);	# critical path
+	&xorps		("xmm2","xmm1");
+	&ret();
+
+&set_label("bad_pointer",4);
+	&mov	("eax",-1);
+	&ret	();
+&set_label("bad_keybits",4);
+	&mov	("eax",-2);
+	&ret	();
+&function_end_B("_aesni_set_encrypt_key");
+
+# int $PREFIX_set_encrypt_key (const unsigned char *userKey, int bits,
+#                              AES_KEY *key)
+&function_begin_B("${PREFIX}_set_encrypt_key");
+	&mov	("eax",&wparam(0));
+	&mov	($rounds,&wparam(1));
+	&mov	($key,&wparam(2));
+	&call	("_aesni_set_encrypt_key");
+	&ret	();
+&function_end_B("${PREFIX}_set_encrypt_key");
+
+# int $PREFIX_set_decrypt_key (const unsigned char *userKey, int bits,
+#                              AES_KEY *key)
+&function_begin_B("${PREFIX}_set_decrypt_key");
+	&mov	("eax",&wparam(0));
+	&mov	($rounds,&wparam(1));
+	&mov	($key,&wparam(2));
+	&call	("_aesni_set_encrypt_key");
+	&mov	($key,&wparam(2));
+	&shl	($rounds,4)	# rounds-1 after _aesni_set_encrypt_key
+	&test	("eax","eax");
+	&jnz	(&label("dec_key_ret"));
+	&lea	("eax",&DWP(16,$key,$rounds));	# end of key schedule
+
+	&$movekey	("xmm0",&QWP(0,$key));	# just swap
+	&$movekey	("xmm1",&QWP(0,"eax"));
+	&$movekey	(&QWP(0,"eax"),"xmm0");
+	&$movekey	(&QWP(0,$key),"xmm1");
+	&lea		($key,&DWP(16,$key));
+	&lea		("eax",&DWP(-16,"eax"));
+
+&set_label("dec_key_inverse");
+	&$movekey	("xmm0",&QWP(0,$key));	# swap and inverse
+	&$movekey	("xmm1",&QWP(0,"eax"));
+	&aesimc		("xmm0","xmm0");
+	&aesimc		("xmm1","xmm1");
+	&lea		($key,&DWP(16,$key));
+	&lea		("eax",&DWP(-16,"eax"));
+	&$movekey	(&QWP(16,"eax"),"xmm0");
+	&$movekey	(&QWP(-16,$key),"xmm1");
+	&cmp		("eax",$key);
+	&ja		(&label("dec_key_inverse"));
+
+	&$movekey	("xmm0",&QWP(0,$key));	# inverse middle
+	&aesimc		("xmm0","xmm0");
+	&$movekey	(&QWP(0,$key),"xmm0");
+
+	&xor		("eax","eax");		# return success
+&set_label("dec_key_ret");
+	&ret	();
+&function_end_B("${PREFIX}_set_decrypt_key");
+&asciz("AES for Intel AES-NI, CRYPTOGAMS by <appro\@openssl.org>");
+
+&asm_finish();
diff --git a/crypto/aes/asm/aesni-x86_64.pl b/crypto/aes/asm/aesni-x86_64.pl
new file mode 100755
index 000000000000..499f3b3f42f8
--- /dev/null
+++ b/crypto/aes/asm/aesni-x86_64.pl
@@ -0,0 +1,3068 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# This module implements support for Intel AES-NI extension. In
+# OpenSSL context it's used with Intel engine, but can also be used as
+# drop-in replacement for crypto/aes/asm/aes-x86_64.pl [see below for
+# details].
+#
+# Performance.
+#
+# Given aes(enc|dec) instructions' latency asymptotic performance for
+# non-parallelizable modes such as CBC encrypt is 3.75 cycles per byte
+# processed with 128-bit key. And given their throughput asymptotic
+# performance for parallelizable modes is 1.25 cycles per byte. Being
+# asymptotic limit it's not something you commonly achieve in reality,
+# but how close does one get? Below are results collected for
+# different modes and block sized. Pairs of numbers are for en-/
+# decryption.
+#
+#	16-byte     64-byte     256-byte    1-KB        8-KB
+# ECB	4.25/4.25   1.38/1.38   1.28/1.28   1.26/1.26	1.26/1.26
+# CTR	5.42/5.42   1.92/1.92   1.44/1.44   1.28/1.28   1.26/1.26
+# CBC	4.38/4.43   4.15/1.43   4.07/1.32   4.07/1.29   4.06/1.28
+# CCM	5.66/9.42   4.42/5.41   4.16/4.40   4.09/4.15   4.06/4.07   
+# OFB	5.42/5.42   4.64/4.64   4.44/4.44   4.39/4.39   4.38/4.38
+# CFB	5.73/5.85   5.56/5.62   5.48/5.56   5.47/5.55   5.47/5.55
+#
+# ECB, CTR, CBC and CCM results are free from EVP overhead. This means
+# that otherwise used 'openssl speed -evp aes-128-??? -engine aesni
+# [-decrypt]' will exhibit 10-15% worse results for smaller blocks.
+# The results were collected with specially crafted speed.c benchmark
+# in order to compare them with results reported in "Intel Advanced
+# Encryption Standard (AES) New Instruction Set" White Paper Revision
+# 3.0 dated May 2010. All above results are consistently better. This
+# module also provides better performance for block sizes smaller than
+# 128 bytes in points *not* represented in the above table.
+#
+# Looking at the results for 8-KB buffer.
+#
+# CFB and OFB results are far from the limit, because implementation
+# uses "generic" CRYPTO_[c|o]fb128_encrypt interfaces relying on
+# single-block aesni_encrypt, which is not the most optimal way to go.
+# CBC encrypt result is unexpectedly high and there is no documented
+# explanation for it. Seemingly there is a small penalty for feeding
+# the result back to AES unit the way it's done in CBC mode. There is
+# nothing one can do and the result appears optimal. CCM result is
+# identical to CBC, because CBC-MAC is essentially CBC encrypt without
+# saving output. CCM CTR "stays invisible," because it's neatly
+# interleaved wih CBC-MAC. This provides ~30% improvement over
+# "straghtforward" CCM implementation with CTR and CBC-MAC performed
+# disjointly. Parallelizable modes practically achieve the theoretical
+# limit.
+#
+# Looking at how results vary with buffer size.
+#
+# Curves are practically saturated at 1-KB buffer size. In most cases
+# "256-byte" performance is >95%, and "64-byte" is ~90% of "8-KB" one.
+# CTR curve doesn't follow this pattern and is "slowest" changing one
+# with "256-byte" result being 87% of "8-KB." This is because overhead
+# in CTR mode is most computationally intensive. Small-block CCM
+# decrypt is slower than encrypt, because first CTR and last CBC-MAC
+# iterations can't be interleaved.
+#
+# Results for 192- and 256-bit keys.
+#
+# EVP-free results were observed to scale perfectly with number of
+# rounds for larger block sizes, i.e. 192-bit result being 10/12 times
+# lower and 256-bit one - 10/14. Well, in CBC encrypt case differences
+# are a tad smaller, because the above mentioned penalty biases all
+# results by same constant value. In similar way function call
+# overhead affects small-block performance, as well as OFB and CFB
+# results. Differences are not large, most common coefficients are
+# 10/11.7 and 10/13.4 (as opposite to 10/12.0 and 10/14.0), but one
+# observe even 10/11.2 and 10/12.4 (CTR, OFB, CFB)...
+
+# January 2011
+#
+# While Westmere processor features 6 cycles latency for aes[enc|dec]
+# instructions, which can be scheduled every second cycle, Sandy
+# Bridge spends 8 cycles per instruction, but it can schedule them
+# every cycle. This means that code targeting Westmere would perform
+# suboptimally on Sandy Bridge. Therefore this update.
+#
+# In addition, non-parallelizable CBC encrypt (as well as CCM) is
+# optimized. Relative improvement might appear modest, 8% on Westmere,
+# but in absolute terms it's 3.77 cycles per byte encrypted with
+# 128-bit key on Westmere, and 5.07 - on Sandy Bridge. These numbers
+# should be compared to asymptotic limits of 3.75 for Westmere and
+# 5.00 for Sandy Bridge. Actually, the fact that they get this close
+# to asymptotic limits is quite amazing. Indeed, the limit is
+# calculated as latency times number of rounds, 10 for 128-bit key,
+# and divided by 16, the number of bytes in block, or in other words
+# it accounts *solely* for aesenc instructions. But there are extra
+# instructions, and numbers so close to the asymptotic limits mean
+# that it's as if it takes as little as *one* additional cycle to
+# execute all of them. How is it possible? It is possible thanks to
+# out-of-order execution logic, which manages to overlap post-
+# processing of previous block, things like saving the output, with
+# actual encryption of current block, as well as pre-processing of
+# current block, things like fetching input and xor-ing it with
+# 0-round element of the key schedule, with actual encryption of
+# previous block. Keep this in mind...
+#
+# For parallelizable modes, such as ECB, CBC decrypt, CTR, higher
+# performance is achieved by interleaving instructions working on
+# independent blocks. In which case asymptotic limit for such modes
+# can be obtained by dividing above mentioned numbers by AES
+# instructions' interleave factor. Westmere can execute at most 3 
+# instructions at a time, meaning that optimal interleave factor is 3,
+# and that's where the "magic" number of 1.25 come from. "Optimal
+# interleave factor" means that increase of interleave factor does
+# not improve performance. The formula has proven to reflect reality
+# pretty well on Westmere... Sandy Bridge on the other hand can
+# execute up to 8 AES instructions at a time, so how does varying
+# interleave factor affect the performance? Here is table for ECB
+# (numbers are cycles per byte processed with 128-bit key):
+#
+# instruction interleave factor		3x	6x	8x
+# theoretical asymptotic limit		1.67	0.83	0.625
+# measured performance for 8KB block	1.05	0.86	0.84
+#
+# "as if" interleave factor		4.7x	5.8x	6.0x
+#
+# Further data for other parallelizable modes:
+#
+# CBC decrypt				1.16	0.93	0.93
+# CTR					1.14	0.91	n/a
+#
+# Well, given 3x column it's probably inappropriate to call the limit
+# asymptotic, if it can be surpassed, isn't it? What happens there?
+# Rewind to CBC paragraph for the answer. Yes, out-of-order execution
+# magic is responsible for this. Processor overlaps not only the
+# additional instructions with AES ones, but even AES instuctions
+# processing adjacent triplets of independent blocks. In the 6x case
+# additional instructions  still claim disproportionally small amount
+# of additional cycles, but in 8x case number of instructions must be
+# a tad too high for out-of-order logic to cope with, and AES unit
+# remains underutilized... As you can see 8x interleave is hardly
+# justifiable, so there no need to feel bad that 32-bit aesni-x86.pl
+# utilizies 6x interleave because of limited register bank capacity.
+#
+# Higher interleave factors do have negative impact on Westmere
+# performance. While for ECB mode it's negligible ~1.5%, other
+# parallelizables perform ~5% worse, which is outweighed by ~25%
+# improvement on Sandy Bridge. To balance regression on Westmere
+# CTR mode was implemented with 6x aesenc interleave factor.
+
+# April 2011
+#
+# Add aesni_xts_[en|de]crypt. Westmere spends 1.33 cycles processing
+# one byte out of 8KB with 128-bit key, Sandy Bridge - 0.97. Just like
+# in CTR mode AES instruction interleave factor was chosen to be 6x.
+
+$PREFIX="aesni";	# if $PREFIX is set to "AES", the script
+			# generates drop-in replacement for
+			# crypto/aes/asm/aes-x86_64.pl:-)
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+$movkey = $PREFIX eq "aesni" ? "movups" : "movups";
+@_4args=$win64?	("%rcx","%rdx","%r8", "%r9") :	# Win64 order
+		("%rdi","%rsi","%rdx","%rcx");	# Unix order
+
+$code=".text\n";
+
+$rounds="%eax";	# input to and changed by aesni_[en|de]cryptN !!!
+# this is natural Unix argument order for public $PREFIX_[ecb|cbc]_encrypt ...
+$inp="%rdi";
+$out="%rsi";
+$len="%rdx";
+$key="%rcx";	# input to and changed by aesni_[en|de]cryptN !!!
+$ivp="%r8";	# cbc, ctr, ...
+
+$rnds_="%r10d";	# backup copy for $rounds
+$key_="%r11";	# backup copy for $key
+
+# %xmm register layout
+$rndkey0="%xmm0";	$rndkey1="%xmm1";
+$inout0="%xmm2";	$inout1="%xmm3";
+$inout2="%xmm4";	$inout3="%xmm5";
+$inout4="%xmm6";	$inout5="%xmm7";
+$inout6="%xmm8";	$inout7="%xmm9";
+
+$in2="%xmm6";		$in1="%xmm7";	# used in CBC decrypt, CTR, ...
+$in0="%xmm8";		$iv="%xmm9";
+
+# Inline version of internal aesni_[en|de]crypt1.
+#
+# Why folded loop? Because aes[enc|dec] is slow enough to accommodate
+# cycles which take care of loop variables...
+{ my $sn;
+sub aesni_generate1 {
+my ($p,$key,$rounds,$inout,$ivec)=@_;	$inout=$inout0 if (!defined($inout));
+++$sn;
+$code.=<<___;
+	$movkey	($key),$rndkey0
+	$movkey	16($key),$rndkey1
+___
+$code.=<<___ if (defined($ivec));
+	xorps	$rndkey0,$ivec
+	lea	32($key),$key
+	xorps	$ivec,$inout
+___
+$code.=<<___ if (!defined($ivec));
+	lea	32($key),$key
+	xorps	$rndkey0,$inout
+___
+$code.=<<___;
+.Loop_${p}1_$sn:
+	aes${p}	$rndkey1,$inout
+	dec	$rounds
+	$movkey	($key),$rndkey1
+	lea	16($key),$key
+	jnz	.Loop_${p}1_$sn	# loop body is 16 bytes
+	aes${p}last	$rndkey1,$inout
+___
+}}
+# void $PREFIX_[en|de]crypt (const void *inp,void *out,const AES_KEY *key);
+#
+{ my ($inp,$out,$key) = @_4args;
+
+$code.=<<___;
+.globl	${PREFIX}_encrypt
+.type	${PREFIX}_encrypt,\@abi-omnipotent
+.align	16
+${PREFIX}_encrypt:
+	movups	($inp),$inout0		# load input
+	mov	240($key),$rounds	# key->rounds
+___
+	&aesni_generate1("enc",$key,$rounds);
+$code.=<<___;
+	movups	$inout0,($out)		# output
+	ret
+.size	${PREFIX}_encrypt,.-${PREFIX}_encrypt
+
+.globl	${PREFIX}_decrypt
+.type	${PREFIX}_decrypt,\@abi-omnipotent
+.align	16
+${PREFIX}_decrypt:
+	movups	($inp),$inout0		# load input
+	mov	240($key),$rounds	# key->rounds
+___
+	&aesni_generate1("dec",$key,$rounds);
+$code.=<<___;
+	movups	$inout0,($out)		# output
+	ret
+.size	${PREFIX}_decrypt, .-${PREFIX}_decrypt
+___
+}
+
+# _aesni_[en|de]cryptN are private interfaces, N denotes interleave
+# factor. Why 3x subroutine were originally used in loops? Even though
+# aes[enc|dec] latency was originally 6, it could be scheduled only
+# every *2nd* cycle. Thus 3x interleave was the one providing optimal
+# utilization, i.e. when subroutine's throughput is virtually same as
+# of non-interleaved subroutine [for number of input blocks up to 3].
+# This is why it makes no sense to implement 2x subroutine.
+# aes[enc|dec] latency in next processor generation is 8, but the
+# instructions can be scheduled every cycle. Optimal interleave for
+# new processor is therefore 8x...
+sub aesni_generate3 {
+my $dir=shift;
+# As already mentioned it takes in $key and $rounds, which are *not*
+# preserved. $inout[0-2] is cipher/clear text...
+$code.=<<___;
+.type	_aesni_${dir}rypt3,\@abi-omnipotent
+.align	16
+_aesni_${dir}rypt3:
+	$movkey	($key),$rndkey0
+	shr	\$1,$rounds
+	$movkey	16($key),$rndkey1
+	lea	32($key),$key
+	xorps	$rndkey0,$inout0
+	xorps	$rndkey0,$inout1
+	xorps	$rndkey0,$inout2
+	$movkey		($key),$rndkey0
+
+.L${dir}_loop3:
+	aes${dir}	$rndkey1,$inout0
+	aes${dir}	$rndkey1,$inout1
+	dec		$rounds
+	aes${dir}	$rndkey1,$inout2
+	$movkey		16($key),$rndkey1
+	aes${dir}	$rndkey0,$inout0
+	aes${dir}	$rndkey0,$inout1
+	lea		32($key),$key
+	aes${dir}	$rndkey0,$inout2
+	$movkey		($key),$rndkey0
+	jnz		.L${dir}_loop3
+
+	aes${dir}	$rndkey1,$inout0
+	aes${dir}	$rndkey1,$inout1
+	aes${dir}	$rndkey1,$inout2
+	aes${dir}last	$rndkey0,$inout0
+	aes${dir}last	$rndkey0,$inout1
+	aes${dir}last	$rndkey0,$inout2
+	ret
+.size	_aesni_${dir}rypt3,.-_aesni_${dir}rypt3
+___
+}
+# 4x interleave is implemented to improve small block performance,
+# most notably [and naturally] 4 block by ~30%. One can argue that one
+# should have implemented 5x as well, but improvement would be <20%,
+# so it's not worth it...
+sub aesni_generate4 {
+my $dir=shift;
+# As already mentioned it takes in $key and $rounds, which are *not*
+# preserved. $inout[0-3] is cipher/clear text...
+$code.=<<___;
+.type	_aesni_${dir}rypt4,\@abi-omnipotent
+.align	16
+_aesni_${dir}rypt4:
+	$movkey	($key),$rndkey0
+	shr	\$1,$rounds
+	$movkey	16($key),$rndkey1
+	lea	32($key),$key
+	xorps	$rndkey0,$inout0
+	xorps	$rndkey0,$inout1
+	xorps	$rndkey0,$inout2
+	xorps	$rndkey0,$inout3
+	$movkey	($key),$rndkey0
+
+.L${dir}_loop4:
+	aes${dir}	$rndkey1,$inout0
+	aes${dir}	$rndkey1,$inout1
+	dec		$rounds
+	aes${dir}	$rndkey1,$inout2
+	aes${dir}	$rndkey1,$inout3
+	$movkey		16($key),$rndkey1
+	aes${dir}	$rndkey0,$inout0
+	aes${dir}	$rndkey0,$inout1
+	lea		32($key),$key
+	aes${dir}	$rndkey0,$inout2
+	aes${dir}	$rndkey0,$inout3
+	$movkey		($key),$rndkey0
+	jnz		.L${dir}_loop4
+
+	aes${dir}	$rndkey1,$inout0
+	aes${dir}	$rndkey1,$inout1
+	aes${dir}	$rndkey1,$inout2
+	aes${dir}	$rndkey1,$inout3
+	aes${dir}last	$rndkey0,$inout0
+	aes${dir}last	$rndkey0,$inout1
+	aes${dir}last	$rndkey0,$inout2
+	aes${dir}last	$rndkey0,$inout3
+	ret
+.size	_aesni_${dir}rypt4,.-_aesni_${dir}rypt4
+___
+}
+sub aesni_generate6 {
+my $dir=shift;
+# As already mentioned it takes in $key and $rounds, which are *not*
+# preserved. $inout[0-5] is cipher/clear text...
+$code.=<<___;
+.type	_aesni_${dir}rypt6,\@abi-omnipotent
+.align	16
+_aesni_${dir}rypt6:
+	$movkey		($key),$rndkey0
+	shr		\$1,$rounds
+	$movkey		16($key),$rndkey1
+	lea		32($key),$key
+	xorps		$rndkey0,$inout0
+	pxor		$rndkey0,$inout1
+	aes${dir}	$rndkey1,$inout0
+	pxor		$rndkey0,$inout2
+	aes${dir}	$rndkey1,$inout1
+	pxor		$rndkey0,$inout3
+	aes${dir}	$rndkey1,$inout2
+	pxor		$rndkey0,$inout4
+	aes${dir}	$rndkey1,$inout3
+	pxor		$rndkey0,$inout5
+	dec		$rounds
+	aes${dir}	$rndkey1,$inout4
+	$movkey		($key),$rndkey0
+	aes${dir}	$rndkey1,$inout5
+	jmp		.L${dir}_loop6_enter
+.align	16
+.L${dir}_loop6:
+	aes${dir}	$rndkey1,$inout0
+	aes${dir}	$rndkey1,$inout1
+	dec		$rounds
+	aes${dir}	$rndkey1,$inout2
+	aes${dir}	$rndkey1,$inout3
+	aes${dir}	$rndkey1,$inout4
+	aes${dir}	$rndkey1,$inout5
+.L${dir}_loop6_enter:				# happens to be 16-byte aligned
+	$movkey		16($key),$rndkey1
+	aes${dir}	$rndkey0,$inout0
+	aes${dir}	$rndkey0,$inout1
+	lea		32($key),$key
+	aes${dir}	$rndkey0,$inout2
+	aes${dir}	$rndkey0,$inout3
+	aes${dir}	$rndkey0,$inout4
+	aes${dir}	$rndkey0,$inout5
+	$movkey		($key),$rndkey0
+	jnz		.L${dir}_loop6
+
+	aes${dir}	$rndkey1,$inout0
+	aes${dir}	$rndkey1,$inout1
+	aes${dir}	$rndkey1,$inout2
+	aes${dir}	$rndkey1,$inout3
+	aes${dir}	$rndkey1,$inout4
+	aes${dir}	$rndkey1,$inout5
+	aes${dir}last	$rndkey0,$inout0
+	aes${dir}last	$rndkey0,$inout1
+	aes${dir}last	$rndkey0,$inout2
+	aes${dir}last	$rndkey0,$inout3
+	aes${dir}last	$rndkey0,$inout4
+	aes${dir}last	$rndkey0,$inout5
+	ret
+.size	_aesni_${dir}rypt6,.-_aesni_${dir}rypt6
+___
+}
+sub aesni_generate8 {
+my $dir=shift;
+# As already mentioned it takes in $key and $rounds, which are *not*
+# preserved. $inout[0-7] is cipher/clear text...
+$code.=<<___;
+.type	_aesni_${dir}rypt8,\@abi-omnipotent
+.align	16
+_aesni_${dir}rypt8:
+	$movkey		($key),$rndkey0
+	shr		\$1,$rounds
+	$movkey		16($key),$rndkey1
+	lea		32($key),$key
+	xorps		$rndkey0,$inout0
+	xorps		$rndkey0,$inout1
+	aes${dir}	$rndkey1,$inout0
+	pxor		$rndkey0,$inout2
+	aes${dir}	$rndkey1,$inout1
+	pxor		$rndkey0,$inout3
+	aes${dir}	$rndkey1,$inout2
+	pxor		$rndkey0,$inout4
+	aes${dir}	$rndkey1,$inout3
+	pxor		$rndkey0,$inout5
+	dec		$rounds
+	aes${dir}	$rndkey1,$inout4
+	pxor		$rndkey0,$inout6
+	aes${dir}	$rndkey1,$inout5
+	pxor		$rndkey0,$inout7
+	$movkey		($key),$rndkey0
+	aes${dir}	$rndkey1,$inout6
+	aes${dir}	$rndkey1,$inout7
+	$movkey		16($key),$rndkey1
+	jmp		.L${dir}_loop8_enter
+.align	16
+.L${dir}_loop8:
+	aes${dir}	$rndkey1,$inout0
+	aes${dir}	$rndkey1,$inout1
+	dec		$rounds
+	aes${dir}	$rndkey1,$inout2
+	aes${dir}	$rndkey1,$inout3
+	aes${dir}	$rndkey1,$inout4
+	aes${dir}	$rndkey1,$inout5
+	aes${dir}	$rndkey1,$inout6
+	aes${dir}	$rndkey1,$inout7
+	$movkey		16($key),$rndkey1
+.L${dir}_loop8_enter:				# happens to be 16-byte aligned
+	aes${dir}	$rndkey0,$inout0
+	aes${dir}	$rndkey0,$inout1
+	lea		32($key),$key
+	aes${dir}	$rndkey0,$inout2
+	aes${dir}	$rndkey0,$inout3
+	aes${dir}	$rndkey0,$inout4
+	aes${dir}	$rndkey0,$inout5
+	aes${dir}	$rndkey0,$inout6
+	aes${dir}	$rndkey0,$inout7
+	$movkey		($key),$rndkey0
+	jnz		.L${dir}_loop8
+
+	aes${dir}	$rndkey1,$inout0
+	aes${dir}	$rndkey1,$inout1
+	aes${dir}	$rndkey1,$inout2
+	aes${dir}	$rndkey1,$inout3
+	aes${dir}	$rndkey1,$inout4
+	aes${dir}	$rndkey1,$inout5
+	aes${dir}	$rndkey1,$inout6
+	aes${dir}	$rndkey1,$inout7
+	aes${dir}last	$rndkey0,$inout0
+	aes${dir}last	$rndkey0,$inout1
+	aes${dir}last	$rndkey0,$inout2
+	aes${dir}last	$rndkey0,$inout3
+	aes${dir}last	$rndkey0,$inout4
+	aes${dir}last	$rndkey0,$inout5
+	aes${dir}last	$rndkey0,$inout6
+	aes${dir}last	$rndkey0,$inout7
+	ret
+.size	_aesni_${dir}rypt8,.-_aesni_${dir}rypt8
+___
+}
+&aesni_generate3("enc") if ($PREFIX eq "aesni");
+&aesni_generate3("dec");
+&aesni_generate4("enc") if ($PREFIX eq "aesni");
+&aesni_generate4("dec");
+&aesni_generate6("enc") if ($PREFIX eq "aesni");
+&aesni_generate6("dec");
+&aesni_generate8("enc") if ($PREFIX eq "aesni");
+&aesni_generate8("dec");
+
+if ($PREFIX eq "aesni") {
+########################################################################
+# void aesni_ecb_encrypt (const void *in, void *out,
+#			  size_t length, const AES_KEY *key,
+#			  int enc);
+$code.=<<___;
+.globl	aesni_ecb_encrypt
+.type	aesni_ecb_encrypt,\@function,5
+.align	16
+aesni_ecb_encrypt:
+	and	\$-16,$len
+	jz	.Lecb_ret
+
+	mov	240($key),$rounds	# key->rounds
+	$movkey	($key),$rndkey0
+	mov	$key,$key_		# backup $key
+	mov	$rounds,$rnds_		# backup $rounds
+	test	%r8d,%r8d		# 5th argument
+	jz	.Lecb_decrypt
+#--------------------------- ECB ENCRYPT ------------------------------#
+	cmp	\$0x80,$len
+	jb	.Lecb_enc_tail
+
+	movdqu	($inp),$inout0
+	movdqu	0x10($inp),$inout1
+	movdqu	0x20($inp),$inout2
+	movdqu	0x30($inp),$inout3
+	movdqu	0x40($inp),$inout4
+	movdqu	0x50($inp),$inout5
+	movdqu	0x60($inp),$inout6
+	movdqu	0x70($inp),$inout7
+	lea	0x80($inp),$inp
+	sub	\$0x80,$len
+	jmp	.Lecb_enc_loop8_enter
+.align 16
+.Lecb_enc_loop8:
+	movups	$inout0,($out)
+	mov	$key_,$key		# restore $key
+	movdqu	($inp),$inout0
+	mov	$rnds_,$rounds		# restore $rounds
+	movups	$inout1,0x10($out)
+	movdqu	0x10($inp),$inout1
+	movups	$inout2,0x20($out)
+	movdqu	0x20($inp),$inout2
+	movups	$inout3,0x30($out)
+	movdqu	0x30($inp),$inout3
+	movups	$inout4,0x40($out)
+	movdqu	0x40($inp),$inout4
+	movups	$inout5,0x50($out)
+	movdqu	0x50($inp),$inout5
+	movups	$inout6,0x60($out)
+	movdqu	0x60($inp),$inout6
+	movups	$inout7,0x70($out)
+	lea	0x80($out),$out
+	movdqu	0x70($inp),$inout7
+	lea	0x80($inp),$inp
+.Lecb_enc_loop8_enter:
+
+	call	_aesni_encrypt8
+
+	sub	\$0x80,$len
+	jnc	.Lecb_enc_loop8
+
+	movups	$inout0,($out)
+	mov	$key_,$key		# restore $key
+	movups	$inout1,0x10($out)
+	mov	$rnds_,$rounds		# restore $rounds
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	movups	$inout5,0x50($out)
+	movups	$inout6,0x60($out)
+	movups	$inout7,0x70($out)
+	lea	0x80($out),$out
+	add	\$0x80,$len
+	jz	.Lecb_ret
+
+.Lecb_enc_tail:
+	movups	($inp),$inout0
+	cmp	\$0x20,$len
+	jb	.Lecb_enc_one
+	movups	0x10($inp),$inout1
+	je	.Lecb_enc_two
+	movups	0x20($inp),$inout2
+	cmp	\$0x40,$len
+	jb	.Lecb_enc_three
+	movups	0x30($inp),$inout3
+	je	.Lecb_enc_four
+	movups	0x40($inp),$inout4
+	cmp	\$0x60,$len
+	jb	.Lecb_enc_five
+	movups	0x50($inp),$inout5
+	je	.Lecb_enc_six
+	movdqu	0x60($inp),$inout6
+	call	_aesni_encrypt8
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	movups	$inout5,0x50($out)
+	movups	$inout6,0x60($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_enc_one:
+___
+	&aesni_generate1("enc",$key,$rounds);
+$code.=<<___;
+	movups	$inout0,($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_enc_two:
+	xorps	$inout2,$inout2
+	call	_aesni_encrypt3
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_enc_three:
+	call	_aesni_encrypt3
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_enc_four:
+	call	_aesni_encrypt4
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_enc_five:
+	xorps	$inout5,$inout5
+	call	_aesni_encrypt6
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_enc_six:
+	call	_aesni_encrypt6
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	movups	$inout5,0x50($out)
+	jmp	.Lecb_ret
+#--------------------------- ECB DECRYPT ------------------------------#
+.align	16
+.Lecb_decrypt:
+	cmp	\$0x80,$len
+	jb	.Lecb_dec_tail
+
+	movdqu	($inp),$inout0
+	movdqu	0x10($inp),$inout1
+	movdqu	0x20($inp),$inout2
+	movdqu	0x30($inp),$inout3
+	movdqu	0x40($inp),$inout4
+	movdqu	0x50($inp),$inout5
+	movdqu	0x60($inp),$inout6
+	movdqu	0x70($inp),$inout7
+	lea	0x80($inp),$inp
+	sub	\$0x80,$len
+	jmp	.Lecb_dec_loop8_enter
+.align 16
+.Lecb_dec_loop8:
+	movups	$inout0,($out)
+	mov	$key_,$key		# restore $key
+	movdqu	($inp),$inout0
+	mov	$rnds_,$rounds		# restore $rounds
+	movups	$inout1,0x10($out)
+	movdqu	0x10($inp),$inout1
+	movups	$inout2,0x20($out)
+	movdqu	0x20($inp),$inout2
+	movups	$inout3,0x30($out)
+	movdqu	0x30($inp),$inout3
+	movups	$inout4,0x40($out)
+	movdqu	0x40($inp),$inout4
+	movups	$inout5,0x50($out)
+	movdqu	0x50($inp),$inout5
+	movups	$inout6,0x60($out)
+	movdqu	0x60($inp),$inout6
+	movups	$inout7,0x70($out)
+	lea	0x80($out),$out
+	movdqu	0x70($inp),$inout7
+	lea	0x80($inp),$inp
+.Lecb_dec_loop8_enter:
+
+	call	_aesni_decrypt8
+
+	$movkey	($key_),$rndkey0
+	sub	\$0x80,$len
+	jnc	.Lecb_dec_loop8
+
+	movups	$inout0,($out)
+	mov	$key_,$key		# restore $key
+	movups	$inout1,0x10($out)
+	mov	$rnds_,$rounds		# restore $rounds
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	movups	$inout5,0x50($out)
+	movups	$inout6,0x60($out)
+	movups	$inout7,0x70($out)
+	lea	0x80($out),$out
+	add	\$0x80,$len
+	jz	.Lecb_ret
+
+.Lecb_dec_tail:
+	movups	($inp),$inout0
+	cmp	\$0x20,$len
+	jb	.Lecb_dec_one
+	movups	0x10($inp),$inout1
+	je	.Lecb_dec_two
+	movups	0x20($inp),$inout2
+	cmp	\$0x40,$len
+	jb	.Lecb_dec_three
+	movups	0x30($inp),$inout3
+	je	.Lecb_dec_four
+	movups	0x40($inp),$inout4
+	cmp	\$0x60,$len
+	jb	.Lecb_dec_five
+	movups	0x50($inp),$inout5
+	je	.Lecb_dec_six
+	movups	0x60($inp),$inout6
+	$movkey	($key),$rndkey0
+	call	_aesni_decrypt8
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	movups	$inout5,0x50($out)
+	movups	$inout6,0x60($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_dec_one:
+___
+	&aesni_generate1("dec",$key,$rounds);
+$code.=<<___;
+	movups	$inout0,($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_dec_two:
+	xorps	$inout2,$inout2
+	call	_aesni_decrypt3
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_dec_three:
+	call	_aesni_decrypt3
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_dec_four:
+	call	_aesni_decrypt4
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_dec_five:
+	xorps	$inout5,$inout5
+	call	_aesni_decrypt6
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	jmp	.Lecb_ret
+.align	16
+.Lecb_dec_six:
+	call	_aesni_decrypt6
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	movups	$inout5,0x50($out)
+
+.Lecb_ret:
+	ret
+.size	aesni_ecb_encrypt,.-aesni_ecb_encrypt
+___
+
+{
+######################################################################
+# void aesni_ccm64_[en|de]crypt_blocks (const void *in, void *out,
+#                         size_t blocks, const AES_KEY *key,
+#                         const char *ivec,char *cmac);
+#
+# Handles only complete blocks, operates on 64-bit counter and
+# does not update *ivec! Nor does it finalize CMAC value
+# (see engine/eng_aesni.c for details)
+#
+{
+my $cmac="%r9";	# 6th argument
+
+my $increment="%xmm6";
+my $bswap_mask="%xmm7";
+
+$code.=<<___;
+.globl	aesni_ccm64_encrypt_blocks
+.type	aesni_ccm64_encrypt_blocks,\@function,6
+.align	16
+aesni_ccm64_encrypt_blocks:
+___
+$code.=<<___ if ($win64);
+	lea	-0x58(%rsp),%rsp
+	movaps	%xmm6,(%rsp)
+	movaps	%xmm7,0x10(%rsp)
+	movaps	%xmm8,0x20(%rsp)
+	movaps	%xmm9,0x30(%rsp)
+.Lccm64_enc_body:
+___
+$code.=<<___;
+	mov	240($key),$rounds		# key->rounds
+	movdqu	($ivp),$iv
+	movdqa	.Lincrement64(%rip),$increment
+	movdqa	.Lbswap_mask(%rip),$bswap_mask
+
+	shr	\$1,$rounds
+	lea	0($key),$key_
+	movdqu	($cmac),$inout1
+	movdqa	$iv,$inout0
+	mov	$rounds,$rnds_
+	pshufb	$bswap_mask,$iv
+	jmp	.Lccm64_enc_outer
+.align	16
+.Lccm64_enc_outer:
+	$movkey	($key_),$rndkey0
+	mov	$rnds_,$rounds
+	movups	($inp),$in0			# load inp
+
+	xorps	$rndkey0,$inout0		# counter
+	$movkey	16($key_),$rndkey1
+	xorps	$in0,$rndkey0
+	lea	32($key_),$key
+	xorps	$rndkey0,$inout1		# cmac^=inp
+	$movkey	($key),$rndkey0
+
+.Lccm64_enc2_loop:
+	aesenc	$rndkey1,$inout0
+	dec	$rounds
+	aesenc	$rndkey1,$inout1
+	$movkey	16($key),$rndkey1
+	aesenc	$rndkey0,$inout0
+	lea	32($key),$key
+	aesenc	$rndkey0,$inout1
+	$movkey	0($key),$rndkey0
+	jnz	.Lccm64_enc2_loop
+	aesenc	$rndkey1,$inout0
+	aesenc	$rndkey1,$inout1
+	paddq	$increment,$iv
+	aesenclast	$rndkey0,$inout0
+	aesenclast	$rndkey0,$inout1
+
+	dec	$len
+	lea	16($inp),$inp
+	xorps	$inout0,$in0			# inp ^= E(iv)
+	movdqa	$iv,$inout0
+	movups	$in0,($out)			# save output
+	lea	16($out),$out
+	pshufb	$bswap_mask,$inout0
+	jnz	.Lccm64_enc_outer
+
+	movups	$inout1,($cmac)
+___
+$code.=<<___ if ($win64);
+	movaps	(%rsp),%xmm6
+	movaps	0x10(%rsp),%xmm7
+	movaps	0x20(%rsp),%xmm8
+	movaps	0x30(%rsp),%xmm9
+	lea	0x58(%rsp),%rsp
+.Lccm64_enc_ret:
+___
+$code.=<<___;
+	ret
+.size	aesni_ccm64_encrypt_blocks,.-aesni_ccm64_encrypt_blocks
+___
+######################################################################
+$code.=<<___;
+.globl	aesni_ccm64_decrypt_blocks
+.type	aesni_ccm64_decrypt_blocks,\@function,6
+.align	16
+aesni_ccm64_decrypt_blocks:
+___
+$code.=<<___ if ($win64);
+	lea	-0x58(%rsp),%rsp
+	movaps	%xmm6,(%rsp)
+	movaps	%xmm7,0x10(%rsp)
+	movaps	%xmm8,0x20(%rsp)
+	movaps	%xmm9,0x30(%rsp)
+.Lccm64_dec_body:
+___
+$code.=<<___;
+	mov	240($key),$rounds		# key->rounds
+	movups	($ivp),$iv
+	movdqu	($cmac),$inout1
+	movdqa	.Lincrement64(%rip),$increment
+	movdqa	.Lbswap_mask(%rip),$bswap_mask
+
+	movaps	$iv,$inout0
+	mov	$rounds,$rnds_
+	mov	$key,$key_
+	pshufb	$bswap_mask,$iv
+___
+	&aesni_generate1("enc",$key,$rounds);
+$code.=<<___;
+	movups	($inp),$in0			# load inp
+	paddq	$increment,$iv
+	lea	16($inp),$inp
+	jmp	.Lccm64_dec_outer
+.align	16
+.Lccm64_dec_outer:
+	xorps	$inout0,$in0			# inp ^= E(iv)
+	movdqa	$iv,$inout0
+	mov	$rnds_,$rounds
+	movups	$in0,($out)			# save output
+	lea	16($out),$out
+	pshufb	$bswap_mask,$inout0
+
+	sub	\$1,$len
+	jz	.Lccm64_dec_break
+
+	$movkey	($key_),$rndkey0
+	shr	\$1,$rounds
+	$movkey	16($key_),$rndkey1
+	xorps	$rndkey0,$in0
+	lea	32($key_),$key
+	xorps	$rndkey0,$inout0
+	xorps	$in0,$inout1			# cmac^=out
+	$movkey	($key),$rndkey0
+
+.Lccm64_dec2_loop:
+	aesenc	$rndkey1,$inout0
+	dec	$rounds
+	aesenc	$rndkey1,$inout1
+	$movkey	16($key),$rndkey1
+	aesenc	$rndkey0,$inout0
+	lea	32($key),$key
+	aesenc	$rndkey0,$inout1
+	$movkey	0($key),$rndkey0
+	jnz	.Lccm64_dec2_loop
+	movups	($inp),$in0			# load inp
+	paddq	$increment,$iv
+	aesenc	$rndkey1,$inout0
+	aesenc	$rndkey1,$inout1
+	lea	16($inp),$inp
+	aesenclast	$rndkey0,$inout0
+	aesenclast	$rndkey0,$inout1
+	jmp	.Lccm64_dec_outer
+
+.align	16
+.Lccm64_dec_break:
+	#xorps	$in0,$inout1			# cmac^=out
+___
+	&aesni_generate1("enc",$key_,$rounds,$inout1,$in0);
+$code.=<<___;
+	movups	$inout1,($cmac)
+___
+$code.=<<___ if ($win64);
+	movaps	(%rsp),%xmm6
+	movaps	0x10(%rsp),%xmm7
+	movaps	0x20(%rsp),%xmm8
+	movaps	0x30(%rsp),%xmm9
+	lea	0x58(%rsp),%rsp
+.Lccm64_dec_ret:
+___
+$code.=<<___;
+	ret
+.size	aesni_ccm64_decrypt_blocks,.-aesni_ccm64_decrypt_blocks
+___
+}
+######################################################################
+# void aesni_ctr32_encrypt_blocks (const void *in, void *out,
+#                         size_t blocks, const AES_KEY *key,
+#                         const char *ivec);
+#
+# Handles only complete blocks, operates on 32-bit counter and
+# does not update *ivec! (see engine/eng_aesni.c for details)
+#
+{
+my $reserved = $win64?0:-0x28;
+my ($in0,$in1,$in2,$in3)=map("%xmm$_",(8..11));
+my ($iv0,$iv1,$ivec)=("%xmm12","%xmm13","%xmm14");
+my $bswap_mask="%xmm15";
+
+$code.=<<___;
+.globl	aesni_ctr32_encrypt_blocks
+.type	aesni_ctr32_encrypt_blocks,\@function,5
+.align	16
+aesni_ctr32_encrypt_blocks:
+___
+$code.=<<___ if ($win64);
+	lea	-0xc8(%rsp),%rsp
+	movaps	%xmm6,0x20(%rsp)
+	movaps	%xmm7,0x30(%rsp)
+	movaps	%xmm8,0x40(%rsp)
+	movaps	%xmm9,0x50(%rsp)
+	movaps	%xmm10,0x60(%rsp)
+	movaps	%xmm11,0x70(%rsp)
+	movaps	%xmm12,0x80(%rsp)
+	movaps	%xmm13,0x90(%rsp)
+	movaps	%xmm14,0xa0(%rsp)
+	movaps	%xmm15,0xb0(%rsp)
+.Lctr32_body:
+___
+$code.=<<___;
+	cmp	\$1,$len
+	je	.Lctr32_one_shortcut
+
+	movdqu	($ivp),$ivec
+	movdqa	.Lbswap_mask(%rip),$bswap_mask
+	xor	$rounds,$rounds
+	pextrd	\$3,$ivec,$rnds_		# pull 32-bit counter
+	pinsrd	\$3,$rounds,$ivec		# wipe 32-bit counter
+
+	mov	240($key),$rounds		# key->rounds
+	bswap	$rnds_
+	pxor	$iv0,$iv0			# vector of 3 32-bit counters
+	pxor	$iv1,$iv1			# vector of 3 32-bit counters
+	pinsrd	\$0,$rnds_,$iv0
+	lea	3($rnds_),$key_
+	pinsrd	\$0,$key_,$iv1
+	inc	$rnds_
+	pinsrd	\$1,$rnds_,$iv0
+	inc	$key_
+	pinsrd	\$1,$key_,$iv1
+	inc	$rnds_
+	pinsrd	\$2,$rnds_,$iv0
+	inc	$key_
+	pinsrd	\$2,$key_,$iv1
+	movdqa	$iv0,$reserved(%rsp)
+	pshufb	$bswap_mask,$iv0
+	movdqa	$iv1,`$reserved+0x10`(%rsp)
+	pshufb	$bswap_mask,$iv1
+
+	pshufd	\$`3<<6`,$iv0,$inout0		# place counter to upper dword
+	pshufd	\$`2<<6`,$iv0,$inout1
+	pshufd	\$`1<<6`,$iv0,$inout2
+	cmp	\$6,$len
+	jb	.Lctr32_tail
+	shr	\$1,$rounds
+	mov	$key,$key_			# backup $key
+	mov	$rounds,$rnds_			# backup $rounds
+	sub	\$6,$len
+	jmp	.Lctr32_loop6
+
+.align	16
+.Lctr32_loop6:
+	pshufd	\$`3<<6`,$iv1,$inout3
+	por	$ivec,$inout0			# merge counter-less ivec
+	 $movkey	($key_),$rndkey0
+	pshufd	\$`2<<6`,$iv1,$inout4
+	por	$ivec,$inout1
+	 $movkey	16($key_),$rndkey1
+	pshufd	\$`1<<6`,$iv1,$inout5
+	por	$ivec,$inout2
+	por	$ivec,$inout3
+	 xorps		$rndkey0,$inout0
+	por	$ivec,$inout4
+	por	$ivec,$inout5
+
+	# inline _aesni_encrypt6 and interleave last rounds
+	# with own code...
+
+	pxor		$rndkey0,$inout1
+	aesenc		$rndkey1,$inout0
+	lea		32($key_),$key
+	pxor		$rndkey0,$inout2
+	aesenc		$rndkey1,$inout1
+	 movdqa		.Lincrement32(%rip),$iv1
+	pxor		$rndkey0,$inout3
+	aesenc		$rndkey1,$inout2
+	 movdqa		$reserved(%rsp),$iv0
+	pxor		$rndkey0,$inout4
+	aesenc		$rndkey1,$inout3
+	pxor		$rndkey0,$inout5
+	$movkey		($key),$rndkey0
+	dec		$rounds
+	aesenc		$rndkey1,$inout4
+	aesenc		$rndkey1,$inout5
+	jmp		.Lctr32_enc_loop6_enter
+.align	16
+.Lctr32_enc_loop6:
+	aesenc		$rndkey1,$inout0
+	aesenc		$rndkey1,$inout1
+	dec		$rounds
+	aesenc		$rndkey1,$inout2
+	aesenc		$rndkey1,$inout3
+	aesenc		$rndkey1,$inout4
+	aesenc		$rndkey1,$inout5
+.Lctr32_enc_loop6_enter:
+	$movkey		16($key),$rndkey1
+	aesenc		$rndkey0,$inout0
+	aesenc		$rndkey0,$inout1
+	lea		32($key),$key
+	aesenc		$rndkey0,$inout2
+	aesenc		$rndkey0,$inout3
+	aesenc		$rndkey0,$inout4
+	aesenc		$rndkey0,$inout5
+	$movkey		($key),$rndkey0
+	jnz		.Lctr32_enc_loop6
+
+	aesenc		$rndkey1,$inout0
+	 paddd		$iv1,$iv0		# increment counter vector
+	aesenc		$rndkey1,$inout1
+	 paddd		`$reserved+0x10`(%rsp),$iv1
+	aesenc		$rndkey1,$inout2
+	 movdqa		$iv0,$reserved(%rsp)	# save counter vector
+	aesenc		$rndkey1,$inout3
+	 movdqa		$iv1,`$reserved+0x10`(%rsp)
+	aesenc		$rndkey1,$inout4
+	 pshufb		$bswap_mask,$iv0	# byte swap
+	aesenc		$rndkey1,$inout5
+	 pshufb		$bswap_mask,$iv1
+
+	aesenclast	$rndkey0,$inout0
+	 movups		($inp),$in0		# load input
+	aesenclast	$rndkey0,$inout1
+	 movups		0x10($inp),$in1
+	aesenclast	$rndkey0,$inout2
+	 movups		0x20($inp),$in2
+	aesenclast	$rndkey0,$inout3
+	 movups		0x30($inp),$in3
+	aesenclast	$rndkey0,$inout4
+	 movups		0x40($inp),$rndkey1
+	aesenclast	$rndkey0,$inout5
+	 movups		0x50($inp),$rndkey0
+	 lea	0x60($inp),$inp
+
+	xorps	$inout0,$in0			# xor
+	 pshufd	\$`3<<6`,$iv0,$inout0
+	xorps	$inout1,$in1
+	 pshufd	\$`2<<6`,$iv0,$inout1
+	movups	$in0,($out)			# store output
+	xorps	$inout2,$in2
+	 pshufd	\$`1<<6`,$iv0,$inout2
+	movups	$in1,0x10($out)
+	xorps	$inout3,$in3
+	movups	$in2,0x20($out)
+	xorps	$inout4,$rndkey1
+	movups	$in3,0x30($out)
+	xorps	$inout5,$rndkey0
+	movups	$rndkey1,0x40($out)
+	movups	$rndkey0,0x50($out)
+	lea	0x60($out),$out
+	mov	$rnds_,$rounds
+	sub	\$6,$len
+	jnc	.Lctr32_loop6
+
+	add	\$6,$len
+	jz	.Lctr32_done
+	mov	$key_,$key			# restore $key
+	lea	1($rounds,$rounds),$rounds	# restore original value
+
+.Lctr32_tail:
+	por	$ivec,$inout0
+	movups	($inp),$in0
+	cmp	\$2,$len
+	jb	.Lctr32_one
+
+	por	$ivec,$inout1
+	movups	0x10($inp),$in1
+	je	.Lctr32_two
+
+	pshufd	\$`3<<6`,$iv1,$inout3
+	por	$ivec,$inout2
+	movups	0x20($inp),$in2
+	cmp	\$4,$len
+	jb	.Lctr32_three
+
+	pshufd	\$`2<<6`,$iv1,$inout4
+	por	$ivec,$inout3
+	movups	0x30($inp),$in3
+	je	.Lctr32_four
+
+	por	$ivec,$inout4
+	xorps	$inout5,$inout5
+
+	call	_aesni_encrypt6
+
+	movups	0x40($inp),$rndkey1
+	xorps	$inout0,$in0
+	xorps	$inout1,$in1
+	movups	$in0,($out)
+	xorps	$inout2,$in2
+	movups	$in1,0x10($out)
+	xorps	$inout3,$in3
+	movups	$in2,0x20($out)
+	xorps	$inout4,$rndkey1
+	movups	$in3,0x30($out)
+	movups	$rndkey1,0x40($out)
+	jmp	.Lctr32_done
+
+.align	16
+.Lctr32_one_shortcut:
+	movups	($ivp),$inout0
+	movups	($inp),$in0
+	mov	240($key),$rounds		# key->rounds
+.Lctr32_one:
+___
+	&aesni_generate1("enc",$key,$rounds);
+$code.=<<___;
+	xorps	$inout0,$in0
+	movups	$in0,($out)
+	jmp	.Lctr32_done
+
+.align	16
+.Lctr32_two:
+	xorps	$inout2,$inout2
+	call	_aesni_encrypt3
+	xorps	$inout0,$in0
+	xorps	$inout1,$in1
+	movups	$in0,($out)
+	movups	$in1,0x10($out)
+	jmp	.Lctr32_done
+
+.align	16
+.Lctr32_three:
+	call	_aesni_encrypt3
+	xorps	$inout0,$in0
+	xorps	$inout1,$in1
+	movups	$in0,($out)
+	xorps	$inout2,$in2
+	movups	$in1,0x10($out)
+	movups	$in2,0x20($out)
+	jmp	.Lctr32_done
+
+.align	16
+.Lctr32_four:
+	call	_aesni_encrypt4
+	xorps	$inout0,$in0
+	xorps	$inout1,$in1
+	movups	$in0,($out)
+	xorps	$inout2,$in2
+	movups	$in1,0x10($out)
+	xorps	$inout3,$in3
+	movups	$in2,0x20($out)
+	movups	$in3,0x30($out)
+
+.Lctr32_done:
+___
+$code.=<<___ if ($win64);
+	movaps	0x20(%rsp),%xmm6
+	movaps	0x30(%rsp),%xmm7
+	movaps	0x40(%rsp),%xmm8
+	movaps	0x50(%rsp),%xmm9
+	movaps	0x60(%rsp),%xmm10
+	movaps	0x70(%rsp),%xmm11
+	movaps	0x80(%rsp),%xmm12
+	movaps	0x90(%rsp),%xmm13
+	movaps	0xa0(%rsp),%xmm14
+	movaps	0xb0(%rsp),%xmm15
+	lea	0xc8(%rsp),%rsp
+.Lctr32_ret:
+___
+$code.=<<___;
+	ret
+.size	aesni_ctr32_encrypt_blocks,.-aesni_ctr32_encrypt_blocks
+___
+}
+
+######################################################################
+# void aesni_xts_[en|de]crypt(const char *inp,char *out,size_t len,
+#	const AES_KEY *key1, const AES_KEY *key2
+#	const unsigned char iv[16]);
+#
+{
+my @tweak=map("%xmm$_",(10..15));
+my ($twmask,$twres,$twtmp)=("%xmm8","%xmm9",@tweak[4]);
+my ($key2,$ivp,$len_)=("%r8","%r9","%r9");
+my $frame_size = 0x68 + ($win64?160:0);
+
+$code.=<<___;
+.globl	aesni_xts_encrypt
+.type	aesni_xts_encrypt,\@function,6
+.align	16
+aesni_xts_encrypt:
+	lea	-$frame_size(%rsp),%rsp
+___
+$code.=<<___ if ($win64);
+	movaps	%xmm6,0x60(%rsp)
+	movaps	%xmm7,0x70(%rsp)
+	movaps	%xmm8,0x80(%rsp)
+	movaps	%xmm9,0x90(%rsp)
+	movaps	%xmm10,0xa0(%rsp)
+	movaps	%xmm11,0xb0(%rsp)
+	movaps	%xmm12,0xc0(%rsp)
+	movaps	%xmm13,0xd0(%rsp)
+	movaps	%xmm14,0xe0(%rsp)
+	movaps	%xmm15,0xf0(%rsp)
+.Lxts_enc_body:
+___
+$code.=<<___;
+	movups	($ivp),@tweak[5]		# load clear-text tweak
+	mov	240(%r8),$rounds		# key2->rounds
+	mov	240($key),$rnds_		# key1->rounds
+___
+	# generate the tweak
+	&aesni_generate1("enc",$key2,$rounds,@tweak[5]);
+$code.=<<___;
+	mov	$key,$key_			# backup $key
+	mov	$rnds_,$rounds			# backup $rounds
+	mov	$len,$len_			# backup $len
+	and	\$-16,$len
+
+	movdqa	.Lxts_magic(%rip),$twmask
+	pxor	$twtmp,$twtmp
+	pcmpgtd	@tweak[5],$twtmp		# broadcast upper bits
+___
+    for ($i=0;$i<4;$i++) {
+    $code.=<<___;
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[$i]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	pand	$twmask,$twres			# isolate carry and residue
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	pxor	$twres,@tweak[5]
+___
+    }
+$code.=<<___;
+	sub	\$16*6,$len
+	jc	.Lxts_enc_short
+
+	shr	\$1,$rounds
+	sub	\$1,$rounds
+	mov	$rounds,$rnds_
+	jmp	.Lxts_enc_grandloop
+
+.align	16
+.Lxts_enc_grandloop:
+	pshufd	\$0x13,$twtmp,$twres
+	movdqa	@tweak[5],@tweak[4]
+	paddq	@tweak[5],@tweak[5]		# psllq 1,$tweak
+	movdqu	`16*0`($inp),$inout0		# load input
+	pand	$twmask,$twres			# isolate carry and residue
+	movdqu	`16*1`($inp),$inout1
+	pxor	$twres,@tweak[5]
+
+	movdqu	`16*2`($inp),$inout2
+	pxor	@tweak[0],$inout0		# input^=tweak
+	movdqu	`16*3`($inp),$inout3
+	pxor	@tweak[1],$inout1
+	movdqu	`16*4`($inp),$inout4
+	pxor	@tweak[2],$inout2
+	movdqu	`16*5`($inp),$inout5
+	lea	`16*6`($inp),$inp
+	pxor	@tweak[3],$inout3
+	$movkey		($key_),$rndkey0
+	pxor	@tweak[4],$inout4
+	pxor	@tweak[5],$inout5
+
+	# inline _aesni_encrypt6 and interleave first and last rounds
+	# with own code...
+	$movkey		16($key_),$rndkey1
+	pxor		$rndkey0,$inout0
+	pxor		$rndkey0,$inout1
+	 movdqa	@tweak[0],`16*0`(%rsp)		# put aside tweaks
+	aesenc		$rndkey1,$inout0
+	lea		32($key_),$key
+	pxor		$rndkey0,$inout2
+	 movdqa	@tweak[1],`16*1`(%rsp)
+	aesenc		$rndkey1,$inout1
+	pxor		$rndkey0,$inout3
+	 movdqa	@tweak[2],`16*2`(%rsp)
+	aesenc		$rndkey1,$inout2
+	pxor		$rndkey0,$inout4
+	 movdqa	@tweak[3],`16*3`(%rsp)
+	aesenc		$rndkey1,$inout3
+	pxor		$rndkey0,$inout5
+	$movkey		($key),$rndkey0
+	dec		$rounds
+	 movdqa	@tweak[4],`16*4`(%rsp)
+	aesenc		$rndkey1,$inout4
+	 movdqa	@tweak[5],`16*5`(%rsp)
+	aesenc		$rndkey1,$inout5
+	pxor	$twtmp,$twtmp
+	pcmpgtd	@tweak[5],$twtmp
+	jmp		.Lxts_enc_loop6_enter
+
+.align	16
+.Lxts_enc_loop6:
+	aesenc		$rndkey1,$inout0
+	aesenc		$rndkey1,$inout1
+	dec		$rounds
+	aesenc		$rndkey1,$inout2
+	aesenc		$rndkey1,$inout3
+	aesenc		$rndkey1,$inout4
+	aesenc		$rndkey1,$inout5
+.Lxts_enc_loop6_enter:
+	$movkey		16($key),$rndkey1
+	aesenc		$rndkey0,$inout0
+	aesenc		$rndkey0,$inout1
+	lea		32($key),$key
+	aesenc		$rndkey0,$inout2
+	aesenc		$rndkey0,$inout3
+	aesenc		$rndkey0,$inout4
+	aesenc		$rndkey0,$inout5
+	$movkey		($key),$rndkey0
+	jnz		.Lxts_enc_loop6
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 aesenc		$rndkey1,$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 aesenc		$rndkey1,$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcast upper bits
+	 aesenc		$rndkey1,$inout2
+	pxor	$twres,@tweak[5]
+	 aesenc		$rndkey1,$inout3
+	 aesenc		$rndkey1,$inout4
+	 aesenc		$rndkey1,$inout5
+	 $movkey	16($key),$rndkey1
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[0]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 aesenc		$rndkey0,$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 aesenc		$rndkey0,$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	 aesenc		$rndkey0,$inout2
+	pxor	$twres,@tweak[5]
+	 aesenc		$rndkey0,$inout3
+	 aesenc		$rndkey0,$inout4
+	 aesenc		$rndkey0,$inout5
+	 $movkey	32($key),$rndkey0
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[1]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 aesenc		$rndkey1,$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 aesenc		$rndkey1,$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	 aesenc		$rndkey1,$inout2
+	pxor	$twres,@tweak[5]
+	 aesenc		$rndkey1,$inout3
+	 aesenc		$rndkey1,$inout4
+	 aesenc		$rndkey1,$inout5
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[2]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 aesenclast	$rndkey0,$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 aesenclast	$rndkey0,$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	 aesenclast	$rndkey0,$inout2
+	pxor	$twres,@tweak[5]
+	 aesenclast	$rndkey0,$inout3
+	 aesenclast	$rndkey0,$inout4
+	 aesenclast	$rndkey0,$inout5
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[3]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 xorps	`16*0`(%rsp),$inout0		# output^=tweak
+	pand	$twmask,$twres			# isolate carry and residue
+	 xorps	`16*1`(%rsp),$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	pxor	$twres,@tweak[5]
+
+	xorps	`16*2`(%rsp),$inout2
+	movups	$inout0,`16*0`($out)		# write output
+	xorps	`16*3`(%rsp),$inout3
+	movups	$inout1,`16*1`($out)
+	xorps	`16*4`(%rsp),$inout4
+	movups	$inout2,`16*2`($out)
+	xorps	`16*5`(%rsp),$inout5
+	movups	$inout3,`16*3`($out)
+	mov	$rnds_,$rounds			# restore $rounds
+	movups	$inout4,`16*4`($out)
+	movups	$inout5,`16*5`($out)
+	lea	`16*6`($out),$out
+	sub	\$16*6,$len
+	jnc	.Lxts_enc_grandloop
+
+	lea	3($rounds,$rounds),$rounds	# restore original value
+	mov	$key_,$key			# restore $key
+	mov	$rounds,$rnds_			# backup $rounds
+
+.Lxts_enc_short:
+	add	\$16*6,$len
+	jz	.Lxts_enc_done
+
+	cmp	\$0x20,$len
+	jb	.Lxts_enc_one
+	je	.Lxts_enc_two
+
+	cmp	\$0x40,$len
+	jb	.Lxts_enc_three
+	je	.Lxts_enc_four
+
+	pshufd	\$0x13,$twtmp,$twres
+	movdqa	@tweak[5],@tweak[4]
+	paddq	@tweak[5],@tweak[5]		# psllq 1,$tweak
+	 movdqu	($inp),$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 movdqu	16*1($inp),$inout1
+	pxor	$twres,@tweak[5]
+
+	movdqu	16*2($inp),$inout2
+	pxor	@tweak[0],$inout0
+	movdqu	16*3($inp),$inout3
+	pxor	@tweak[1],$inout1
+	movdqu	16*4($inp),$inout4
+	lea	16*5($inp),$inp
+	pxor	@tweak[2],$inout2
+	pxor	@tweak[3],$inout3
+	pxor	@tweak[4],$inout4
+
+	call	_aesni_encrypt6
+
+	xorps	@tweak[0],$inout0
+	movdqa	@tweak[5],@tweak[0]
+	xorps	@tweak[1],$inout1
+	xorps	@tweak[2],$inout2
+	movdqu	$inout0,($out)
+	xorps	@tweak[3],$inout3
+	movdqu	$inout1,16*1($out)
+	xorps	@tweak[4],$inout4
+	movdqu	$inout2,16*2($out)
+	movdqu	$inout3,16*3($out)
+	movdqu	$inout4,16*4($out)
+	lea	16*5($out),$out
+	jmp	.Lxts_enc_done
+
+.align	16
+.Lxts_enc_one:
+	movups	($inp),$inout0
+	lea	16*1($inp),$inp
+	xorps	@tweak[0],$inout0
+___
+	&aesni_generate1("enc",$key,$rounds);
+$code.=<<___;
+	xorps	@tweak[0],$inout0
+	movdqa	@tweak[1],@tweak[0]
+	movups	$inout0,($out)
+	lea	16*1($out),$out
+	jmp	.Lxts_enc_done
+
+.align	16
+.Lxts_enc_two:
+	movups	($inp),$inout0
+	movups	16($inp),$inout1
+	lea	32($inp),$inp
+	xorps	@tweak[0],$inout0
+	xorps	@tweak[1],$inout1
+
+	call	_aesni_encrypt3
+
+	xorps	@tweak[0],$inout0
+	movdqa	@tweak[2],@tweak[0]
+	xorps	@tweak[1],$inout1
+	movups	$inout0,($out)
+	movups	$inout1,16*1($out)
+	lea	16*2($out),$out
+	jmp	.Lxts_enc_done
+
+.align	16
+.Lxts_enc_three:
+	movups	($inp),$inout0
+	movups	16*1($inp),$inout1
+	movups	16*2($inp),$inout2
+	lea	16*3($inp),$inp
+	xorps	@tweak[0],$inout0
+	xorps	@tweak[1],$inout1
+	xorps	@tweak[2],$inout2
+
+	call	_aesni_encrypt3
+
+	xorps	@tweak[0],$inout0
+	movdqa	@tweak[3],@tweak[0]
+	xorps	@tweak[1],$inout1
+	xorps	@tweak[2],$inout2
+	movups	$inout0,($out)
+	movups	$inout1,16*1($out)
+	movups	$inout2,16*2($out)
+	lea	16*3($out),$out
+	jmp	.Lxts_enc_done
+
+.align	16
+.Lxts_enc_four:
+	movups	($inp),$inout0
+	movups	16*1($inp),$inout1
+	movups	16*2($inp),$inout2
+	xorps	@tweak[0],$inout0
+	movups	16*3($inp),$inout3
+	lea	16*4($inp),$inp
+	xorps	@tweak[1],$inout1
+	xorps	@tweak[2],$inout2
+	xorps	@tweak[3],$inout3
+
+	call	_aesni_encrypt4
+
+	xorps	@tweak[0],$inout0
+	movdqa	@tweak[5],@tweak[0]
+	xorps	@tweak[1],$inout1
+	xorps	@tweak[2],$inout2
+	movups	$inout0,($out)
+	xorps	@tweak[3],$inout3
+	movups	$inout1,16*1($out)
+	movups	$inout2,16*2($out)
+	movups	$inout3,16*3($out)
+	lea	16*4($out),$out
+	jmp	.Lxts_enc_done
+
+.align	16
+.Lxts_enc_done:
+	and	\$15,$len_
+	jz	.Lxts_enc_ret
+	mov	$len_,$len
+
+.Lxts_enc_steal:
+	movzb	($inp),%eax			# borrow $rounds ...
+	movzb	-16($out),%ecx			# ... and $key
+	lea	1($inp),$inp
+	mov	%al,-16($out)
+	mov	%cl,0($out)
+	lea	1($out),$out
+	sub	\$1,$len
+	jnz	.Lxts_enc_steal
+
+	sub	$len_,$out			# rewind $out
+	mov	$key_,$key			# restore $key
+	mov	$rnds_,$rounds			# restore $rounds
+
+	movups	-16($out),$inout0
+	xorps	@tweak[0],$inout0
+___
+	&aesni_generate1("enc",$key,$rounds);
+$code.=<<___;
+	xorps	@tweak[0],$inout0
+	movups	$inout0,-16($out)
+
+.Lxts_enc_ret:
+___
+$code.=<<___ if ($win64);
+	movaps	0x60(%rsp),%xmm6
+	movaps	0x70(%rsp),%xmm7
+	movaps	0x80(%rsp),%xmm8
+	movaps	0x90(%rsp),%xmm9
+	movaps	0xa0(%rsp),%xmm10
+	movaps	0xb0(%rsp),%xmm11
+	movaps	0xc0(%rsp),%xmm12
+	movaps	0xd0(%rsp),%xmm13
+	movaps	0xe0(%rsp),%xmm14
+	movaps	0xf0(%rsp),%xmm15
+___
+$code.=<<___;
+	lea	$frame_size(%rsp),%rsp
+.Lxts_enc_epilogue:
+	ret
+.size	aesni_xts_encrypt,.-aesni_xts_encrypt
+___
+
+$code.=<<___;
+.globl	aesni_xts_decrypt
+.type	aesni_xts_decrypt,\@function,6
+.align	16
+aesni_xts_decrypt:
+	lea	-$frame_size(%rsp),%rsp
+___
+$code.=<<___ if ($win64);
+	movaps	%xmm6,0x60(%rsp)
+	movaps	%xmm7,0x70(%rsp)
+	movaps	%xmm8,0x80(%rsp)
+	movaps	%xmm9,0x90(%rsp)
+	movaps	%xmm10,0xa0(%rsp)
+	movaps	%xmm11,0xb0(%rsp)
+	movaps	%xmm12,0xc0(%rsp)
+	movaps	%xmm13,0xd0(%rsp)
+	movaps	%xmm14,0xe0(%rsp)
+	movaps	%xmm15,0xf0(%rsp)
+.Lxts_dec_body:
+___
+$code.=<<___;
+	movups	($ivp),@tweak[5]		# load clear-text tweak
+	mov	240($key2),$rounds		# key2->rounds
+	mov	240($key),$rnds_		# key1->rounds
+___
+	# generate the tweak
+	&aesni_generate1("enc",$key2,$rounds,@tweak[5]);
+$code.=<<___;
+	xor	%eax,%eax			# if ($len%16) len-=16;
+	test	\$15,$len
+	setnz	%al
+	shl	\$4,%rax
+	sub	%rax,$len
+
+	mov	$key,$key_			# backup $key
+	mov	$rnds_,$rounds			# backup $rounds
+	mov	$len,$len_			# backup $len
+	and	\$-16,$len
+
+	movdqa	.Lxts_magic(%rip),$twmask
+	pxor	$twtmp,$twtmp
+	pcmpgtd	@tweak[5],$twtmp		# broadcast upper bits
+___
+    for ($i=0;$i<4;$i++) {
+    $code.=<<___;
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[$i]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	pand	$twmask,$twres			# isolate carry and residue
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	pxor	$twres,@tweak[5]
+___
+    }
+$code.=<<___;
+	sub	\$16*6,$len
+	jc	.Lxts_dec_short
+
+	shr	\$1,$rounds
+	sub	\$1,$rounds
+	mov	$rounds,$rnds_
+	jmp	.Lxts_dec_grandloop
+
+.align	16
+.Lxts_dec_grandloop:
+	pshufd	\$0x13,$twtmp,$twres
+	movdqa	@tweak[5],@tweak[4]
+	paddq	@tweak[5],@tweak[5]		# psllq 1,$tweak
+	movdqu	`16*0`($inp),$inout0		# load input
+	pand	$twmask,$twres			# isolate carry and residue
+	movdqu	`16*1`($inp),$inout1
+	pxor	$twres,@tweak[5]
+
+	movdqu	`16*2`($inp),$inout2
+	pxor	@tweak[0],$inout0		# input^=tweak
+	movdqu	`16*3`($inp),$inout3
+	pxor	@tweak[1],$inout1
+	movdqu	`16*4`($inp),$inout4
+	pxor	@tweak[2],$inout2
+	movdqu	`16*5`($inp),$inout5
+	lea	`16*6`($inp),$inp
+	pxor	@tweak[3],$inout3
+	$movkey		($key_),$rndkey0
+	pxor	@tweak[4],$inout4
+	pxor	@tweak[5],$inout5
+
+	# inline _aesni_decrypt6 and interleave first and last rounds
+	# with own code...
+	$movkey		16($key_),$rndkey1
+	pxor		$rndkey0,$inout0
+	pxor		$rndkey0,$inout1
+	 movdqa	@tweak[0],`16*0`(%rsp)		# put aside tweaks
+	aesdec		$rndkey1,$inout0
+	lea		32($key_),$key
+	pxor		$rndkey0,$inout2
+	 movdqa	@tweak[1],`16*1`(%rsp)
+	aesdec		$rndkey1,$inout1
+	pxor		$rndkey0,$inout3
+	 movdqa	@tweak[2],`16*2`(%rsp)
+	aesdec		$rndkey1,$inout2
+	pxor		$rndkey0,$inout4
+	 movdqa	@tweak[3],`16*3`(%rsp)
+	aesdec		$rndkey1,$inout3
+	pxor		$rndkey0,$inout5
+	$movkey		($key),$rndkey0
+	dec		$rounds
+	 movdqa	@tweak[4],`16*4`(%rsp)
+	aesdec		$rndkey1,$inout4
+	 movdqa	@tweak[5],`16*5`(%rsp)
+	aesdec		$rndkey1,$inout5
+	pxor	$twtmp,$twtmp
+	pcmpgtd	@tweak[5],$twtmp
+	jmp		.Lxts_dec_loop6_enter
+
+.align	16
+.Lxts_dec_loop6:
+	aesdec		$rndkey1,$inout0
+	aesdec		$rndkey1,$inout1
+	dec		$rounds
+	aesdec		$rndkey1,$inout2
+	aesdec		$rndkey1,$inout3
+	aesdec		$rndkey1,$inout4
+	aesdec		$rndkey1,$inout5
+.Lxts_dec_loop6_enter:
+	$movkey		16($key),$rndkey1
+	aesdec		$rndkey0,$inout0
+	aesdec		$rndkey0,$inout1
+	lea		32($key),$key
+	aesdec		$rndkey0,$inout2
+	aesdec		$rndkey0,$inout3
+	aesdec		$rndkey0,$inout4
+	aesdec		$rndkey0,$inout5
+	$movkey		($key),$rndkey0
+	jnz		.Lxts_dec_loop6
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 aesdec		$rndkey1,$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 aesdec		$rndkey1,$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcast upper bits
+	 aesdec		$rndkey1,$inout2
+	pxor	$twres,@tweak[5]
+	 aesdec		$rndkey1,$inout3
+	 aesdec		$rndkey1,$inout4
+	 aesdec		$rndkey1,$inout5
+	 $movkey	16($key),$rndkey1
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[0]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 aesdec		$rndkey0,$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 aesdec		$rndkey0,$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	 aesdec		$rndkey0,$inout2
+	pxor	$twres,@tweak[5]
+	 aesdec		$rndkey0,$inout3
+	 aesdec		$rndkey0,$inout4
+	 aesdec		$rndkey0,$inout5
+	 $movkey	32($key),$rndkey0
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[1]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 aesdec		$rndkey1,$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 aesdec		$rndkey1,$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	 aesdec		$rndkey1,$inout2
+	pxor	$twres,@tweak[5]
+	 aesdec		$rndkey1,$inout3
+	 aesdec		$rndkey1,$inout4
+	 aesdec		$rndkey1,$inout5
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[2]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 aesdeclast	$rndkey0,$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 aesdeclast	$rndkey0,$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	 aesdeclast	$rndkey0,$inout2
+	pxor	$twres,@tweak[5]
+	 aesdeclast	$rndkey0,$inout3
+	 aesdeclast	$rndkey0,$inout4
+	 aesdeclast	$rndkey0,$inout5
+
+	pshufd	\$0x13,$twtmp,$twres
+	pxor	$twtmp,$twtmp
+	movdqa	@tweak[5],@tweak[3]
+	paddq	@tweak[5],@tweak[5]		# psllq	1,$tweak
+	 xorps	`16*0`(%rsp),$inout0		# output^=tweak
+	pand	$twmask,$twres			# isolate carry and residue
+	 xorps	`16*1`(%rsp),$inout1
+	pcmpgtd	@tweak[5],$twtmp		# broadcat upper bits
+	pxor	$twres,@tweak[5]
+
+	xorps	`16*2`(%rsp),$inout2
+	movups	$inout0,`16*0`($out)		# write output
+	xorps	`16*3`(%rsp),$inout3
+	movups	$inout1,`16*1`($out)
+	xorps	`16*4`(%rsp),$inout4
+	movups	$inout2,`16*2`($out)
+	xorps	`16*5`(%rsp),$inout5
+	movups	$inout3,`16*3`($out)
+	mov	$rnds_,$rounds			# restore $rounds
+	movups	$inout4,`16*4`($out)
+	movups	$inout5,`16*5`($out)
+	lea	`16*6`($out),$out
+	sub	\$16*6,$len
+	jnc	.Lxts_dec_grandloop
+
+	lea	3($rounds,$rounds),$rounds	# restore original value
+	mov	$key_,$key			# restore $key
+	mov	$rounds,$rnds_			# backup $rounds
+
+.Lxts_dec_short:
+	add	\$16*6,$len
+	jz	.Lxts_dec_done
+
+	cmp	\$0x20,$len
+	jb	.Lxts_dec_one
+	je	.Lxts_dec_two
+
+	cmp	\$0x40,$len
+	jb	.Lxts_dec_three
+	je	.Lxts_dec_four
+
+	pshufd	\$0x13,$twtmp,$twres
+	movdqa	@tweak[5],@tweak[4]
+	paddq	@tweak[5],@tweak[5]		# psllq 1,$tweak
+	 movdqu	($inp),$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 movdqu	16*1($inp),$inout1
+	pxor	$twres,@tweak[5]
+
+	movdqu	16*2($inp),$inout2
+	pxor	@tweak[0],$inout0
+	movdqu	16*3($inp),$inout3
+	pxor	@tweak[1],$inout1
+	movdqu	16*4($inp),$inout4
+	lea	16*5($inp),$inp
+	pxor	@tweak[2],$inout2
+	pxor	@tweak[3],$inout3
+	pxor	@tweak[4],$inout4
+
+	call	_aesni_decrypt6
+
+	xorps	@tweak[0],$inout0
+	xorps	@tweak[1],$inout1
+	xorps	@tweak[2],$inout2
+	movdqu	$inout0,($out)
+	xorps	@tweak[3],$inout3
+	movdqu	$inout1,16*1($out)
+	xorps	@tweak[4],$inout4
+	movdqu	$inout2,16*2($out)
+	 pxor		$twtmp,$twtmp
+	movdqu	$inout3,16*3($out)
+	 pcmpgtd	@tweak[5],$twtmp
+	movdqu	$inout4,16*4($out)
+	lea	16*5($out),$out
+	 pshufd		\$0x13,$twtmp,@tweak[1]	# $twres
+	and	\$15,$len_
+	jz	.Lxts_dec_ret
+
+	movdqa	@tweak[5],@tweak[0]
+	paddq	@tweak[5],@tweak[5]		# psllq 1,$tweak
+	pand	$twmask,@tweak[1]		# isolate carry and residue
+	pxor	@tweak[5],@tweak[1]
+	jmp	.Lxts_dec_done2
+
+.align	16
+.Lxts_dec_one:
+	movups	($inp),$inout0
+	lea	16*1($inp),$inp
+	xorps	@tweak[0],$inout0
+___
+	&aesni_generate1("dec",$key,$rounds);
+$code.=<<___;
+	xorps	@tweak[0],$inout0
+	movdqa	@tweak[1],@tweak[0]
+	movups	$inout0,($out)
+	movdqa	@tweak[2],@tweak[1]
+	lea	16*1($out),$out
+	jmp	.Lxts_dec_done
+
+.align	16
+.Lxts_dec_two:
+	movups	($inp),$inout0
+	movups	16($inp),$inout1
+	lea	32($inp),$inp
+	xorps	@tweak[0],$inout0
+	xorps	@tweak[1],$inout1
+
+	call	_aesni_decrypt3
+
+	xorps	@tweak[0],$inout0
+	movdqa	@tweak[2],@tweak[0]
+	xorps	@tweak[1],$inout1
+	movdqa	@tweak[3],@tweak[1]
+	movups	$inout0,($out)
+	movups	$inout1,16*1($out)
+	lea	16*2($out),$out
+	jmp	.Lxts_dec_done
+
+.align	16
+.Lxts_dec_three:
+	movups	($inp),$inout0
+	movups	16*1($inp),$inout1
+	movups	16*2($inp),$inout2
+	lea	16*3($inp),$inp
+	xorps	@tweak[0],$inout0
+	xorps	@tweak[1],$inout1
+	xorps	@tweak[2],$inout2
+
+	call	_aesni_decrypt3
+
+	xorps	@tweak[0],$inout0
+	movdqa	@tweak[3],@tweak[0]
+	xorps	@tweak[1],$inout1
+	movdqa	@tweak[5],@tweak[1]
+	xorps	@tweak[2],$inout2
+	movups	$inout0,($out)
+	movups	$inout1,16*1($out)
+	movups	$inout2,16*2($out)
+	lea	16*3($out),$out
+	jmp	.Lxts_dec_done
+
+.align	16
+.Lxts_dec_four:
+	pshufd	\$0x13,$twtmp,$twres
+	movdqa	@tweak[5],@tweak[4]
+	paddq	@tweak[5],@tweak[5]		# psllq 1,$tweak
+	 movups	($inp),$inout0
+	pand	$twmask,$twres			# isolate carry and residue
+	 movups	16*1($inp),$inout1
+	pxor	$twres,@tweak[5]
+
+	movups	16*2($inp),$inout2
+	xorps	@tweak[0],$inout0
+	movups	16*3($inp),$inout3
+	lea	16*4($inp),$inp
+	xorps	@tweak[1],$inout1
+	xorps	@tweak[2],$inout2
+	xorps	@tweak[3],$inout3
+
+	call	_aesni_decrypt4
+
+	xorps	@tweak[0],$inout0
+	movdqa	@tweak[4],@tweak[0]
+	xorps	@tweak[1],$inout1
+	movdqa	@tweak[5],@tweak[1]
+	xorps	@tweak[2],$inout2
+	movups	$inout0,($out)
+	xorps	@tweak[3],$inout3
+	movups	$inout1,16*1($out)
+	movups	$inout2,16*2($out)
+	movups	$inout3,16*3($out)
+	lea	16*4($out),$out
+	jmp	.Lxts_dec_done
+
+.align	16
+.Lxts_dec_done:
+	and	\$15,$len_
+	jz	.Lxts_dec_ret
+.Lxts_dec_done2:
+	mov	$len_,$len
+	mov	$key_,$key			# restore $key
+	mov	$rnds_,$rounds			# restore $rounds
+
+	movups	($inp),$inout0
+	xorps	@tweak[1],$inout0
+___
+	&aesni_generate1("dec",$key,$rounds);
+$code.=<<___;
+	xorps	@tweak[1],$inout0
+	movups	$inout0,($out)
+
+.Lxts_dec_steal:
+	movzb	16($inp),%eax			# borrow $rounds ...
+	movzb	($out),%ecx			# ... and $key
+	lea	1($inp),$inp
+	mov	%al,($out)
+	mov	%cl,16($out)
+	lea	1($out),$out
+	sub	\$1,$len
+	jnz	.Lxts_dec_steal
+
+	sub	$len_,$out			# rewind $out
+	mov	$key_,$key			# restore $key
+	mov	$rnds_,$rounds			# restore $rounds
+
+	movups	($out),$inout0
+	xorps	@tweak[0],$inout0
+___
+	&aesni_generate1("dec",$key,$rounds);
+$code.=<<___;
+	xorps	@tweak[0],$inout0
+	movups	$inout0,($out)
+
+.Lxts_dec_ret:
+___
+$code.=<<___ if ($win64);
+	movaps	0x60(%rsp),%xmm6
+	movaps	0x70(%rsp),%xmm7
+	movaps	0x80(%rsp),%xmm8
+	movaps	0x90(%rsp),%xmm9
+	movaps	0xa0(%rsp),%xmm10
+	movaps	0xb0(%rsp),%xmm11
+	movaps	0xc0(%rsp),%xmm12
+	movaps	0xd0(%rsp),%xmm13
+	movaps	0xe0(%rsp),%xmm14
+	movaps	0xf0(%rsp),%xmm15
+___
+$code.=<<___;
+	lea	$frame_size(%rsp),%rsp
+.Lxts_dec_epilogue:
+	ret
+.size	aesni_xts_decrypt,.-aesni_xts_decrypt
+___
+} }}
+
+########################################################################
+# void $PREFIX_cbc_encrypt (const void *inp, void *out,
+#			    size_t length, const AES_KEY *key,
+#			    unsigned char *ivp,const int enc);
+{
+my $reserved = $win64?0x40:-0x18;	# used in decrypt
+$code.=<<___;
+.globl	${PREFIX}_cbc_encrypt
+.type	${PREFIX}_cbc_encrypt,\@function,6
+.align	16
+${PREFIX}_cbc_encrypt:
+	test	$len,$len		# check length
+	jz	.Lcbc_ret
+
+	mov	240($key),$rnds_	# key->rounds
+	mov	$key,$key_		# backup $key
+	test	%r9d,%r9d		# 6th argument
+	jz	.Lcbc_decrypt
+#--------------------------- CBC ENCRYPT ------------------------------#
+	movups	($ivp),$inout0		# load iv as initial state
+	mov	$rnds_,$rounds
+	cmp	\$16,$len
+	jb	.Lcbc_enc_tail
+	sub	\$16,$len
+	jmp	.Lcbc_enc_loop
+.align	16
+.Lcbc_enc_loop:
+	movups	($inp),$inout1		# load input
+	lea	16($inp),$inp
+	#xorps	$inout1,$inout0
+___
+	&aesni_generate1("enc",$key,$rounds,$inout0,$inout1);
+$code.=<<___;
+	mov	$rnds_,$rounds		# restore $rounds
+	mov	$key_,$key		# restore $key
+	movups	$inout0,0($out)		# store output
+	lea	16($out),$out
+	sub	\$16,$len
+	jnc	.Lcbc_enc_loop
+	add	\$16,$len
+	jnz	.Lcbc_enc_tail
+	movups	$inout0,($ivp)
+	jmp	.Lcbc_ret
+
+.Lcbc_enc_tail:
+	mov	$len,%rcx	# zaps $key
+	xchg	$inp,$out	# $inp is %rsi and $out is %rdi now
+	.long	0x9066A4F3	# rep movsb
+	mov	\$16,%ecx	# zero tail
+	sub	$len,%rcx
+	xor	%eax,%eax
+	.long	0x9066AAF3	# rep stosb
+	lea	-16(%rdi),%rdi	# rewind $out by 1 block
+	mov	$rnds_,$rounds	# restore $rounds
+	mov	%rdi,%rsi	# $inp and $out are the same
+	mov	$key_,$key	# restore $key
+	xor	$len,$len	# len=16
+	jmp	.Lcbc_enc_loop	# one more spin
+#--------------------------- CBC DECRYPT ------------------------------#
+.align	16
+.Lcbc_decrypt:
+___
+$code.=<<___ if ($win64);
+	lea	-0x58(%rsp),%rsp
+	movaps	%xmm6,(%rsp)
+	movaps	%xmm7,0x10(%rsp)
+	movaps	%xmm8,0x20(%rsp)
+	movaps	%xmm9,0x30(%rsp)
+.Lcbc_decrypt_body:
+___
+$code.=<<___;
+	movups	($ivp),$iv
+	mov	$rnds_,$rounds
+	cmp	\$0x70,$len
+	jbe	.Lcbc_dec_tail
+	shr	\$1,$rnds_
+	sub	\$0x70,$len
+	mov	$rnds_,$rounds
+	movaps	$iv,$reserved(%rsp)
+	jmp	.Lcbc_dec_loop8_enter
+.align	16
+.Lcbc_dec_loop8:
+	movaps	$rndkey0,$reserved(%rsp)	# save IV
+	movups	$inout7,($out)
+	lea	0x10($out),$out
+.Lcbc_dec_loop8_enter:
+	$movkey		($key),$rndkey0
+	movups	($inp),$inout0			# load input
+	movups	0x10($inp),$inout1
+	$movkey		16($key),$rndkey1
+
+	lea		32($key),$key
+	movdqu	0x20($inp),$inout2
+	xorps		$rndkey0,$inout0
+	movdqu	0x30($inp),$inout3
+	xorps		$rndkey0,$inout1
+	movdqu	0x40($inp),$inout4
+	aesdec		$rndkey1,$inout0
+	pxor		$rndkey0,$inout2
+	movdqu	0x50($inp),$inout5
+	aesdec		$rndkey1,$inout1
+	pxor		$rndkey0,$inout3
+	movdqu	0x60($inp),$inout6
+	aesdec		$rndkey1,$inout2
+	pxor		$rndkey0,$inout4
+	movdqu	0x70($inp),$inout7
+	aesdec		$rndkey1,$inout3
+	pxor		$rndkey0,$inout5
+	dec		$rounds
+	aesdec		$rndkey1,$inout4
+	pxor		$rndkey0,$inout6
+	aesdec		$rndkey1,$inout5
+	pxor		$rndkey0,$inout7
+	$movkey		($key),$rndkey0
+	aesdec		$rndkey1,$inout6
+	aesdec		$rndkey1,$inout7
+	$movkey		16($key),$rndkey1
+
+	call		.Ldec_loop8_enter
+
+	movups	($inp),$rndkey1		# re-load input
+	movups	0x10($inp),$rndkey0
+	xorps	$reserved(%rsp),$inout0	# ^= IV
+	xorps	$rndkey1,$inout1
+	movups	0x20($inp),$rndkey1
+	xorps	$rndkey0,$inout2
+	movups	0x30($inp),$rndkey0
+	xorps	$rndkey1,$inout3
+	movups	0x40($inp),$rndkey1
+	xorps	$rndkey0,$inout4
+	movups	0x50($inp),$rndkey0
+	xorps	$rndkey1,$inout5
+	movups	0x60($inp),$rndkey1
+	xorps	$rndkey0,$inout6
+	movups	0x70($inp),$rndkey0	# IV
+	xorps	$rndkey1,$inout7
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	mov	$rnds_,$rounds		# restore $rounds
+	movups	$inout4,0x40($out)
+	mov	$key_,$key		# restore $key
+	movups	$inout5,0x50($out)
+	lea	0x80($inp),$inp
+	movups	$inout6,0x60($out)
+	lea	0x70($out),$out
+	sub	\$0x80,$len
+	ja	.Lcbc_dec_loop8
+
+	movaps	$inout7,$inout0
+	movaps	$rndkey0,$iv
+	add	\$0x70,$len
+	jle	.Lcbc_dec_tail_collected
+	movups	$inout0,($out)
+	lea	1($rnds_,$rnds_),$rounds
+	lea	0x10($out),$out
+.Lcbc_dec_tail:
+	movups	($inp),$inout0
+	movaps	$inout0,$in0
+	cmp	\$0x10,$len
+	jbe	.Lcbc_dec_one
+
+	movups	0x10($inp),$inout1
+	movaps	$inout1,$in1
+	cmp	\$0x20,$len
+	jbe	.Lcbc_dec_two
+
+	movups	0x20($inp),$inout2
+	movaps	$inout2,$in2
+	cmp	\$0x30,$len
+	jbe	.Lcbc_dec_three
+
+	movups	0x30($inp),$inout3
+	cmp	\$0x40,$len
+	jbe	.Lcbc_dec_four
+
+	movups	0x40($inp),$inout4
+	cmp	\$0x50,$len
+	jbe	.Lcbc_dec_five
+
+	movups	0x50($inp),$inout5
+	cmp	\$0x60,$len
+	jbe	.Lcbc_dec_six
+
+	movups	0x60($inp),$inout6
+	movaps	$iv,$reserved(%rsp)	# save IV
+	call	_aesni_decrypt8
+	movups	($inp),$rndkey1
+	movups	0x10($inp),$rndkey0
+	xorps	$reserved(%rsp),$inout0	# ^= IV
+	xorps	$rndkey1,$inout1
+	movups	0x20($inp),$rndkey1
+	xorps	$rndkey0,$inout2
+	movups	0x30($inp),$rndkey0
+	xorps	$rndkey1,$inout3
+	movups	0x40($inp),$rndkey1
+	xorps	$rndkey0,$inout4
+	movups	0x50($inp),$rndkey0
+	xorps	$rndkey1,$inout5
+	movups	0x60($inp),$iv		# IV
+	xorps	$rndkey0,$inout6
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	movups	$inout5,0x50($out)
+	lea	0x60($out),$out
+	movaps	$inout6,$inout0
+	sub	\$0x70,$len
+	jmp	.Lcbc_dec_tail_collected
+.align	16
+.Lcbc_dec_one:
+___
+	&aesni_generate1("dec",$key,$rounds);
+$code.=<<___;
+	xorps	$iv,$inout0
+	movaps	$in0,$iv
+	sub	\$0x10,$len
+	jmp	.Lcbc_dec_tail_collected
+.align	16
+.Lcbc_dec_two:
+	xorps	$inout2,$inout2
+	call	_aesni_decrypt3
+	xorps	$iv,$inout0
+	xorps	$in0,$inout1
+	movups	$inout0,($out)
+	movaps	$in1,$iv
+	movaps	$inout1,$inout0
+	lea	0x10($out),$out
+	sub	\$0x20,$len
+	jmp	.Lcbc_dec_tail_collected
+.align	16
+.Lcbc_dec_three:
+	call	_aesni_decrypt3
+	xorps	$iv,$inout0
+	xorps	$in0,$inout1
+	movups	$inout0,($out)
+	xorps	$in1,$inout2
+	movups	$inout1,0x10($out)
+	movaps	$in2,$iv
+	movaps	$inout2,$inout0
+	lea	0x20($out),$out
+	sub	\$0x30,$len
+	jmp	.Lcbc_dec_tail_collected
+.align	16
+.Lcbc_dec_four:
+	call	_aesni_decrypt4
+	xorps	$iv,$inout0
+	movups	0x30($inp),$iv
+	xorps	$in0,$inout1
+	movups	$inout0,($out)
+	xorps	$in1,$inout2
+	movups	$inout1,0x10($out)
+	xorps	$in2,$inout3
+	movups	$inout2,0x20($out)
+	movaps	$inout3,$inout0
+	lea	0x30($out),$out
+	sub	\$0x40,$len
+	jmp	.Lcbc_dec_tail_collected
+.align	16
+.Lcbc_dec_five:
+	xorps	$inout5,$inout5
+	call	_aesni_decrypt6
+	movups	0x10($inp),$rndkey1
+	movups	0x20($inp),$rndkey0
+	xorps	$iv,$inout0
+	xorps	$in0,$inout1
+	xorps	$rndkey1,$inout2
+	movups	0x30($inp),$rndkey1
+	xorps	$rndkey0,$inout3
+	movups	0x40($inp),$iv
+	xorps	$rndkey1,$inout4
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	lea	0x40($out),$out
+	movaps	$inout4,$inout0
+	sub	\$0x50,$len
+	jmp	.Lcbc_dec_tail_collected
+.align	16
+.Lcbc_dec_six:
+	call	_aesni_decrypt6
+	movups	0x10($inp),$rndkey1
+	movups	0x20($inp),$rndkey0
+	xorps	$iv,$inout0
+	xorps	$in0,$inout1
+	xorps	$rndkey1,$inout2
+	movups	0x30($inp),$rndkey1
+	xorps	$rndkey0,$inout3
+	movups	0x40($inp),$rndkey0
+	xorps	$rndkey1,$inout4
+	movups	0x50($inp),$iv
+	xorps	$rndkey0,$inout5
+	movups	$inout0,($out)
+	movups	$inout1,0x10($out)
+	movups	$inout2,0x20($out)
+	movups	$inout3,0x30($out)
+	movups	$inout4,0x40($out)
+	lea	0x50($out),$out
+	movaps	$inout5,$inout0
+	sub	\$0x60,$len
+	jmp	.Lcbc_dec_tail_collected
+.align	16
+.Lcbc_dec_tail_collected:
+	and	\$15,$len
+	movups	$iv,($ivp)
+	jnz	.Lcbc_dec_tail_partial
+	movups	$inout0,($out)
+	jmp	.Lcbc_dec_ret
+.align	16
+.Lcbc_dec_tail_partial:
+	movaps	$inout0,$reserved(%rsp)
+	mov	\$16,%rcx
+	mov	$out,%rdi
+	sub	$len,%rcx
+	lea	$reserved(%rsp),%rsi
+	.long	0x9066A4F3	# rep movsb
+
+.Lcbc_dec_ret:
+___
+$code.=<<___ if ($win64);
+	movaps	(%rsp),%xmm6
+	movaps	0x10(%rsp),%xmm7
+	movaps	0x20(%rsp),%xmm8
+	movaps	0x30(%rsp),%xmm9
+	lea	0x58(%rsp),%rsp
+___
+$code.=<<___;
+.Lcbc_ret:
+	ret
+.size	${PREFIX}_cbc_encrypt,.-${PREFIX}_cbc_encrypt
+___
+} 
+# int $PREFIX_set_[en|de]crypt_key (const unsigned char *userKey,
+#				int bits, AES_KEY *key)
+{ my ($inp,$bits,$key) = @_4args;
+  $bits =~ s/%r/%e/;
+
+$code.=<<___;
+.globl	${PREFIX}_set_decrypt_key
+.type	${PREFIX}_set_decrypt_key,\@abi-omnipotent
+.align	16
+${PREFIX}_set_decrypt_key:
+	.byte	0x48,0x83,0xEC,0x08	# sub rsp,8
+	call	__aesni_set_encrypt_key
+	shl	\$4,$bits		# rounds-1 after _aesni_set_encrypt_key
+	test	%eax,%eax
+	jnz	.Ldec_key_ret
+	lea	16($key,$bits),$inp	# points at the end of key schedule
+
+	$movkey	($key),%xmm0		# just swap
+	$movkey	($inp),%xmm1
+	$movkey	%xmm0,($inp)
+	$movkey	%xmm1,($key)
+	lea	16($key),$key
+	lea	-16($inp),$inp
+
+.Ldec_key_inverse:
+	$movkey	($key),%xmm0		# swap and inverse
+	$movkey	($inp),%xmm1
+	aesimc	%xmm0,%xmm0
+	aesimc	%xmm1,%xmm1
+	lea	16($key),$key
+	lea	-16($inp),$inp
+	$movkey	%xmm0,16($inp)
+	$movkey	%xmm1,-16($key)
+	cmp	$key,$inp
+	ja	.Ldec_key_inverse
+
+	$movkey	($key),%xmm0		# inverse middle
+	aesimc	%xmm0,%xmm0
+	$movkey	%xmm0,($inp)
+.Ldec_key_ret:
+	add	\$8,%rsp
+	ret
+.LSEH_end_set_decrypt_key:
+.size	${PREFIX}_set_decrypt_key,.-${PREFIX}_set_decrypt_key
+___
+
+# This is based on submission by
+#
+#	Huang Ying <ying.huang@intel.com>
+#	Vinodh Gopal <vinodh.gopal@intel.com>
+#	Kahraman Akdemir
+#
+# Agressively optimized in respect to aeskeygenassist's critical path
+# and is contained in %xmm0-5 to meet Win64 ABI requirement.
+#
+$code.=<<___;
+.globl	${PREFIX}_set_encrypt_key
+.type	${PREFIX}_set_encrypt_key,\@abi-omnipotent
+.align	16
+${PREFIX}_set_encrypt_key:
+__aesni_set_encrypt_key:
+	.byte	0x48,0x83,0xEC,0x08	# sub rsp,8
+	mov	\$-1,%rax
+	test	$inp,$inp
+	jz	.Lenc_key_ret
+	test	$key,$key
+	jz	.Lenc_key_ret
+
+	movups	($inp),%xmm0		# pull first 128 bits of *userKey
+	xorps	%xmm4,%xmm4		# low dword of xmm4 is assumed 0
+	lea	16($key),%rax
+	cmp	\$256,$bits
+	je	.L14rounds
+	cmp	\$192,$bits
+	je	.L12rounds
+	cmp	\$128,$bits
+	jne	.Lbad_keybits
+
+.L10rounds:
+	mov	\$9,$bits			# 10 rounds for 128-bit key
+	$movkey	%xmm0,($key)			# round 0
+	aeskeygenassist	\$0x1,%xmm0,%xmm1	# round 1
+	call		.Lkey_expansion_128_cold
+	aeskeygenassist	\$0x2,%xmm0,%xmm1	# round 2
+	call		.Lkey_expansion_128
+	aeskeygenassist	\$0x4,%xmm0,%xmm1	# round 3
+	call		.Lkey_expansion_128
+	aeskeygenassist	\$0x8,%xmm0,%xmm1	# round 4
+	call		.Lkey_expansion_128
+	aeskeygenassist	\$0x10,%xmm0,%xmm1	# round 5
+	call		.Lkey_expansion_128
+	aeskeygenassist	\$0x20,%xmm0,%xmm1	# round 6
+	call		.Lkey_expansion_128
+	aeskeygenassist	\$0x40,%xmm0,%xmm1	# round 7
+	call		.Lkey_expansion_128
+	aeskeygenassist	\$0x80,%xmm0,%xmm1	# round 8
+	call		.Lkey_expansion_128
+	aeskeygenassist	\$0x1b,%xmm0,%xmm1	# round 9
+	call		.Lkey_expansion_128
+	aeskeygenassist	\$0x36,%xmm0,%xmm1	# round 10
+	call		.Lkey_expansion_128
+	$movkey	%xmm0,(%rax)
+	mov	$bits,80(%rax)	# 240(%rdx)
+	xor	%eax,%eax
+	jmp	.Lenc_key_ret
+
+.align	16
+.L12rounds:
+	movq	16($inp),%xmm2			# remaining 1/3 of *userKey
+	mov	\$11,$bits			# 12 rounds for 192
+	$movkey	%xmm0,($key)			# round 0
+	aeskeygenassist	\$0x1,%xmm2,%xmm1	# round 1,2
+	call		.Lkey_expansion_192a_cold
+	aeskeygenassist	\$0x2,%xmm2,%xmm1	# round 2,3
+	call		.Lkey_expansion_192b
+	aeskeygenassist	\$0x4,%xmm2,%xmm1	# round 4,5
+	call		.Lkey_expansion_192a
+	aeskeygenassist	\$0x8,%xmm2,%xmm1	# round 5,6
+	call		.Lkey_expansion_192b
+	aeskeygenassist	\$0x10,%xmm2,%xmm1	# round 7,8
+	call		.Lkey_expansion_192a
+	aeskeygenassist	\$0x20,%xmm2,%xmm1	# round 8,9
+	call		.Lkey_expansion_192b
+	aeskeygenassist	\$0x40,%xmm2,%xmm1	# round 10,11
+	call		.Lkey_expansion_192a
+	aeskeygenassist	\$0x80,%xmm2,%xmm1	# round 11,12
+	call		.Lkey_expansion_192b
+	$movkey	%xmm0,(%rax)
+	mov	$bits,48(%rax)	# 240(%rdx)
+	xor	%rax, %rax
+	jmp	.Lenc_key_ret
+
+.align	16
+.L14rounds:
+	movups	16($inp),%xmm2			# remaning half of *userKey
+	mov	\$13,$bits			# 14 rounds for 256
+	lea	16(%rax),%rax
+	$movkey	%xmm0,($key)			# round 0
+	$movkey	%xmm2,16($key)			# round 1
+	aeskeygenassist	\$0x1,%xmm2,%xmm1	# round 2
+	call		.Lkey_expansion_256a_cold
+	aeskeygenassist	\$0x1,%xmm0,%xmm1	# round 3
+	call		.Lkey_expansion_256b
+	aeskeygenassist	\$0x2,%xmm2,%xmm1	# round 4
+	call		.Lkey_expansion_256a
+	aeskeygenassist	\$0x2,%xmm0,%xmm1	# round 5
+	call		.Lkey_expansion_256b
+	aeskeygenassist	\$0x4,%xmm2,%xmm1	# round 6
+	call		.Lkey_expansion_256a
+	aeskeygenassist	\$0x4,%xmm0,%xmm1	# round 7
+	call		.Lkey_expansion_256b
+	aeskeygenassist	\$0x8,%xmm2,%xmm1	# round 8
+	call		.Lkey_expansion_256a
+	aeskeygenassist	\$0x8,%xmm0,%xmm1	# round 9
+	call		.Lkey_expansion_256b
+	aeskeygenassist	\$0x10,%xmm2,%xmm1	# round 10
+	call		.Lkey_expansion_256a
+	aeskeygenassist	\$0x10,%xmm0,%xmm1	# round 11
+	call		.Lkey_expansion_256b
+	aeskeygenassist	\$0x20,%xmm2,%xmm1	# round 12
+	call		.Lkey_expansion_256a
+	aeskeygenassist	\$0x20,%xmm0,%xmm1	# round 13
+	call		.Lkey_expansion_256b
+	aeskeygenassist	\$0x40,%xmm2,%xmm1	# round 14
+	call		.Lkey_expansion_256a
+	$movkey	%xmm0,(%rax)
+	mov	$bits,16(%rax)	# 240(%rdx)
+	xor	%rax,%rax
+	jmp	.Lenc_key_ret
+
+.align	16
+.Lbad_keybits:
+	mov	\$-2,%rax
+.Lenc_key_ret:
+	add	\$8,%rsp
+	ret
+.LSEH_end_set_encrypt_key:
+
+.align	16
+.Lkey_expansion_128:
+	$movkey	%xmm0,(%rax)
+	lea	16(%rax),%rax
+.Lkey_expansion_128_cold:
+	shufps	\$0b00010000,%xmm0,%xmm4
+	xorps	%xmm4, %xmm0
+	shufps	\$0b10001100,%xmm0,%xmm4
+	xorps	%xmm4, %xmm0
+	shufps	\$0b11111111,%xmm1,%xmm1	# critical path
+	xorps	%xmm1,%xmm0
+	ret
+
+.align 16
+.Lkey_expansion_192a:
+	$movkey	%xmm0,(%rax)
+	lea	16(%rax),%rax
+.Lkey_expansion_192a_cold:
+	movaps	%xmm2, %xmm5
+.Lkey_expansion_192b_warm:
+	shufps	\$0b00010000,%xmm0,%xmm4
+	movdqa	%xmm2,%xmm3
+	xorps	%xmm4,%xmm0
+	shufps	\$0b10001100,%xmm0,%xmm4
+	pslldq	\$4,%xmm3
+	xorps	%xmm4,%xmm0
+	pshufd	\$0b01010101,%xmm1,%xmm1	# critical path
+	pxor	%xmm3,%xmm2
+	pxor	%xmm1,%xmm0
+	pshufd	\$0b11111111,%xmm0,%xmm3
+	pxor	%xmm3,%xmm2
+	ret
+
+.align 16
+.Lkey_expansion_192b:
+	movaps	%xmm0,%xmm3
+	shufps	\$0b01000100,%xmm0,%xmm5
+	$movkey	%xmm5,(%rax)
+	shufps	\$0b01001110,%xmm2,%xmm3
+	$movkey	%xmm3,16(%rax)
+	lea	32(%rax),%rax
+	jmp	.Lkey_expansion_192b_warm
+
+.align	16
+.Lkey_expansion_256a:
+	$movkey	%xmm2,(%rax)
+	lea	16(%rax),%rax
+.Lkey_expansion_256a_cold:
+	shufps	\$0b00010000,%xmm0,%xmm4
+	xorps	%xmm4,%xmm0
+	shufps	\$0b10001100,%xmm0,%xmm4
+	xorps	%xmm4,%xmm0
+	shufps	\$0b11111111,%xmm1,%xmm1	# critical path
+	xorps	%xmm1,%xmm0
+	ret
+
+.align 16
+.Lkey_expansion_256b:
+	$movkey	%xmm0,(%rax)
+	lea	16(%rax),%rax
+
+	shufps	\$0b00010000,%xmm2,%xmm4
+	xorps	%xmm4,%xmm2
+	shufps	\$0b10001100,%xmm2,%xmm4
+	xorps	%xmm4,%xmm2
+	shufps	\$0b10101010,%xmm1,%xmm1	# critical path
+	xorps	%xmm1,%xmm2
+	ret
+.size	${PREFIX}_set_encrypt_key,.-${PREFIX}_set_encrypt_key
+.size	__aesni_set_encrypt_key,.-__aesni_set_encrypt_key
+___
+}
+
+$code.=<<___;
+.align	64
+.Lbswap_mask:
+	.byte	15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0
+.Lincrement32:
+	.long	6,6,6,0
+.Lincrement64:
+	.long	1,0,0,0
+.Lxts_magic:
+	.long	0x87,0,1,0
+
+.asciz  "AES for Intel AES-NI, CRYPTOGAMS by <appro\@openssl.org>"
+.align	64
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+___
+$code.=<<___ if ($PREFIX eq "aesni");
+.type	ecb_se_handler,\@abi-omnipotent
+.align	16
+ecb_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	jmp	.Lcommon_seh_tail
+.size	ecb_se_handler,.-ecb_se_handler
+
+.type	ccm64_se_handler,\@abi-omnipotent
+.align	16
+ccm64_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lcommon_seh_tail
+
+	lea	0(%rax),%rsi		# %xmm save area
+	lea	512($context),%rdi	# &context.Xmm6
+	mov	\$8,%ecx		# 4*sizeof(%xmm0)/sizeof(%rax)
+	.long	0xa548f3fc		# cld; rep movsq
+	lea	0x58(%rax),%rax		# adjust stack pointer
+
+	jmp	.Lcommon_seh_tail
+.size	ccm64_se_handler,.-ccm64_se_handler
+
+.type	ctr32_se_handler,\@abi-omnipotent
+.align	16
+ctr32_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lctr32_body(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<"prologue" label
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lctr32_ret(%rip),%r10
+	cmp	%r10,%rbx
+	jae	.Lcommon_seh_tail
+
+	lea	0x20(%rax),%rsi		# %xmm save area
+	lea	512($context),%rdi	# &context.Xmm6
+	mov	\$20,%ecx		# 10*sizeof(%xmm0)/sizeof(%rax)
+	.long	0xa548f3fc		# cld; rep movsq
+	lea	0xc8(%rax),%rax		# adjust stack pointer
+
+	jmp	.Lcommon_seh_tail
+.size	ctr32_se_handler,.-ctr32_se_handler
+
+.type	xts_se_handler,\@abi-omnipotent
+.align	16
+xts_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue lable
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lcommon_seh_tail
+
+	lea	0x60(%rax),%rsi		# %xmm save area
+	lea	512($context),%rdi	# & context.Xmm6
+	mov	\$20,%ecx		# 10*sizeof(%xmm0)/sizeof(%rax)
+	.long	0xa548f3fc		# cld; rep movsq
+	lea	0x68+160(%rax),%rax	# adjust stack pointer
+
+	jmp	.Lcommon_seh_tail
+.size	xts_se_handler,.-xts_se_handler
+___
+$code.=<<___;
+.type	cbc_se_handler,\@abi-omnipotent
+.align	16
+cbc_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	152($context),%rax	# pull context->Rsp
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lcbc_decrypt(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<"prologue" label
+	jb	.Lcommon_seh_tail
+
+	lea	.Lcbc_decrypt_body(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<cbc_decrypt_body
+	jb	.Lrestore_cbc_rax
+
+	lea	.Lcbc_ret(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>="epilogue" label
+	jae	.Lcommon_seh_tail
+
+	lea	0(%rax),%rsi		# top of stack
+	lea	512($context),%rdi	# &context.Xmm6
+	mov	\$8,%ecx		# 4*sizeof(%xmm0)/sizeof(%rax)
+	.long	0xa548f3fc		# cld; rep movsq
+	lea	0x58(%rax),%rax		# adjust stack pointer
+	jmp	.Lcommon_seh_tail
+
+.Lrestore_cbc_rax:
+	mov	120($context),%rax
+
+.Lcommon_seh_tail:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	cbc_se_handler,.-cbc_se_handler
+
+.section	.pdata
+.align	4
+___
+$code.=<<___ if ($PREFIX eq "aesni");
+	.rva	.LSEH_begin_aesni_ecb_encrypt
+	.rva	.LSEH_end_aesni_ecb_encrypt
+	.rva	.LSEH_info_ecb
+
+	.rva	.LSEH_begin_aesni_ccm64_encrypt_blocks
+	.rva	.LSEH_end_aesni_ccm64_encrypt_blocks
+	.rva	.LSEH_info_ccm64_enc
+
+	.rva	.LSEH_begin_aesni_ccm64_decrypt_blocks
+	.rva	.LSEH_end_aesni_ccm64_decrypt_blocks
+	.rva	.LSEH_info_ccm64_dec
+
+	.rva	.LSEH_begin_aesni_ctr32_encrypt_blocks
+	.rva	.LSEH_end_aesni_ctr32_encrypt_blocks
+	.rva	.LSEH_info_ctr32
+
+	.rva	.LSEH_begin_aesni_xts_encrypt
+	.rva	.LSEH_end_aesni_xts_encrypt
+	.rva	.LSEH_info_xts_enc
+
+	.rva	.LSEH_begin_aesni_xts_decrypt
+	.rva	.LSEH_end_aesni_xts_decrypt
+	.rva	.LSEH_info_xts_dec
+___
+$code.=<<___;
+	.rva	.LSEH_begin_${PREFIX}_cbc_encrypt
+	.rva	.LSEH_end_${PREFIX}_cbc_encrypt
+	.rva	.LSEH_info_cbc
+
+	.rva	${PREFIX}_set_decrypt_key
+	.rva	.LSEH_end_set_decrypt_key
+	.rva	.LSEH_info_key
+
+	.rva	${PREFIX}_set_encrypt_key
+	.rva	.LSEH_end_set_encrypt_key
+	.rva	.LSEH_info_key
+.section	.xdata
+.align	8
+___
+$code.=<<___ if ($PREFIX eq "aesni");
+.LSEH_info_ecb:
+	.byte	9,0,0,0
+	.rva	ecb_se_handler
+.LSEH_info_ccm64_enc:
+	.byte	9,0,0,0
+	.rva	ccm64_se_handler
+	.rva	.Lccm64_enc_body,.Lccm64_enc_ret	# HandlerData[]
+.LSEH_info_ccm64_dec:
+	.byte	9,0,0,0
+	.rva	ccm64_se_handler
+	.rva	.Lccm64_dec_body,.Lccm64_dec_ret	# HandlerData[]
+.LSEH_info_ctr32:
+	.byte	9,0,0,0
+	.rva	ctr32_se_handler
+.LSEH_info_xts_enc:
+	.byte	9,0,0,0
+	.rva	xts_se_handler
+	.rva	.Lxts_enc_body,.Lxts_enc_epilogue	# HandlerData[]
+.LSEH_info_xts_dec:
+	.byte	9,0,0,0
+	.rva	xts_se_handler
+	.rva	.Lxts_dec_body,.Lxts_dec_epilogue	# HandlerData[]
+___
+$code.=<<___;
+.LSEH_info_cbc:
+	.byte	9,0,0,0
+	.rva	cbc_se_handler
+.LSEH_info_key:
+	.byte	0x01,0x04,0x01,0x00
+	.byte	0x04,0x02,0x00,0x00	# sub rsp,8
+___
+}
+
+sub rex {
+  local *opcode=shift;
+  my ($dst,$src)=@_;
+  my $rex=0;
+
+    $rex|=0x04			if($dst>=8);
+    $rex|=0x01			if($src>=8);
+    push @opcode,$rex|0x40	if($rex);
+}
+
+sub aesni {
+  my $line=shift;
+  my @opcode=(0x66);
+
+    if ($line=~/(aeskeygenassist)\s+\$([x0-9a-f]+),\s*%xmm([0-9]+),\s*%xmm([0-9]+)/) {
+	rex(\@opcode,$4,$3);
+	push @opcode,0x0f,0x3a,0xdf;
+	push @opcode,0xc0|($3&7)|(($4&7)<<3);	# ModR/M
+	my $c=$2;
+	push @opcode,$c=~/^0/?oct($c):$c;
+	return ".byte\t".join(',',@opcode);
+    }
+    elsif ($line=~/(aes[a-z]+)\s+%xmm([0-9]+),\s*%xmm([0-9]+)/) {
+	my %opcodelet = (
+		"aesimc" => 0xdb,
+		"aesenc" => 0xdc,	"aesenclast" => 0xdd,
+		"aesdec" => 0xde,	"aesdeclast" => 0xdf
+	);
+	return undef if (!defined($opcodelet{$1}));
+	rex(\@opcode,$3,$2);
+	push @opcode,0x0f,0x38,$opcodelet{$1};
+	push @opcode,0xc0|($2&7)|(($3&7)<<3);	# ModR/M
+	return ".byte\t".join(',',@opcode);
+    }
+    return $line;
+}
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+$code =~ s/\b(aes.*%xmm[0-9]+).*$/aesni($1)/gem;
+
+print $code;
+
+close STDOUT;
diff --git a/crypto/aes/asm/bsaes-x86_64.pl b/crypto/aes/asm/bsaes-x86_64.pl
new file mode 100755
index 000000000000..c9c6312fa74a
--- /dev/null
+++ b/crypto/aes/asm/bsaes-x86_64.pl
@@ -0,0 +1,3044 @@
+#!/usr/bin/env perl
+
+###################################################################
+### AES-128 [originally in CTR mode]				###
+### bitsliced implementation for Intel Core 2 processors	###
+### requires support of SSE extensions up to SSSE3		###
+### Author: Emilia Käsper and Peter Schwabe			###
+### Date: 2009-03-19						###
+### Public domain						###
+###								###
+### See http://homes.esat.kuleuven.be/~ekasper/#software for	###
+### further information.					###
+###################################################################
+#
+# September 2011.
+#
+# Started as transliteration to "perlasm" the original code has
+# undergone following changes:
+#
+# - code was made position-independent;
+# - rounds were folded into a loop resulting in >5x size reduction
+#   from 12.5KB to 2.2KB;
+# - above was possibile thanks to mixcolumns() modification that
+#   allowed to feed its output back to aesenc[last], this was
+#   achieved at cost of two additional inter-registers moves;
+# - some instruction reordering and interleaving;
+# - this module doesn't implement key setup subroutine, instead it
+#   relies on conversion of "conventional" key schedule as returned
+#   by AES_set_encrypt_key (see discussion below);
+# - first and last round keys are treated differently, which allowed
+#   to skip one shiftrows(), reduce bit-sliced key schedule and
+#   speed-up conversion by 22%;
+# - support for 192- and 256-bit keys was added;
+#
+# Resulting performance in CPU cycles spent to encrypt one byte out
+# of 4096-byte buffer with 128-bit key is:
+#
+#		Emilia's	this(*)		difference
+#
+# Core 2    	9.30		8.69		+7%
+# Nehalem(**) 	7.63		6.98		+9%
+# Atom	    	17.1		17.4		-2%(***)
+#
+# (*)	Comparison is not completely fair, because "this" is ECB,
+#	i.e. no extra processing such as counter values calculation
+#	and xor-ing input as in Emilia's CTR implementation is
+#	performed. However, the CTR calculations stand for not more
+#	than 1% of total time, so comparison is *rather* fair.
+#
+# (**)	Results were collected on Westmere, which is considered to
+#	be equivalent to Nehalem for this code.
+#
+# (***)	Slowdown on Atom is rather strange per se, because original
+#	implementation has a number of 9+-bytes instructions, which
+#	are bad for Atom front-end, and which I eliminated completely.
+#	In attempt to address deterioration sbox() was tested in FP
+#	SIMD "domain" (movaps instead of movdqa, xorps instead of
+#	pxor, etc.). While it resulted in nominal 4% improvement on
+#	Atom, it hurted Westmere by more than 2x factor.
+#
+# As for key schedule conversion subroutine. Interface to OpenSSL
+# relies on per-invocation on-the-fly conversion. This naturally
+# has impact on performance, especially for short inputs. Conversion
+# time in CPU cycles and its ratio to CPU cycles spent in 8x block
+# function is:
+#
+# 		conversion	conversion/8x block
+# Core 2	240		0.22
+# Nehalem	180		0.20
+# Atom		430		0.19
+#
+# The ratio values mean that 128-byte blocks will be processed
+# 16-18% slower, 256-byte blocks - 9-10%, 384-byte blocks - 6-7%,
+# etc. Then keep in mind that input sizes not divisible by 128 are
+# *effectively* slower, especially shortest ones, e.g. consecutive
+# 144-byte blocks are processed 44% slower than one would expect,
+# 272 - 29%, 400 - 22%, etc. Yet, despite all these "shortcomings"
+# it's still faster than ["hyper-threading-safe" code path in]
+# aes-x86_64.pl on all lengths above 64 bytes...
+#
+# October 2011.
+#
+# Add decryption procedure. Performance in CPU cycles spent to decrypt
+# one byte out of 4096-byte buffer with 128-bit key is:
+#
+# Core 2	11.0
+# Nehalem	9.16
+# Atom		20.9
+#
+# November 2011.
+#
+# Add bsaes_xts_[en|de]crypt. Less-than-80-bytes-block performance is
+# suboptimal, but XTS is meant to be used with larger blocks...
+#
+#						<appro@openssl.org>
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+my ($inp,$out,$len,$key,$ivp)=("%rdi","%rsi","%rdx","%rcx");
+my @XMM=map("%xmm$_",(15,0..14));	# best on Atom, +10% over (0..15)
+my $ecb=0;	# suppress unreferenced ECB subroutines, spare some space...
+
+{
+my ($key,$rounds,$const)=("%rax","%r10d","%r11");
+
+sub Sbox {
+# input in  lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b0, b1, b4, b6, b3, b7, b2, b5] < msb
+my @b=@_[0..7];
+my @t=@_[8..11];
+my @s=@_[12..15];
+	&InBasisChange	(@b);
+	&Inv_GF256	(@b[6,5,0,3,7,1,4,2],@t,@s);
+	&OutBasisChange	(@b[7,1,4,2,6,5,0,3]);
+}
+
+sub InBasisChange {
+# input in  lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b6, b5, b0, b3, b7, b1, b4, b2] < msb 
+my @b=@_[0..7];
+$code.=<<___;
+	pxor	@b[6], @b[5]
+	pxor	@b[1], @b[2]
+	pxor	@b[0], @b[3]
+	pxor	@b[2], @b[6]
+	pxor 	@b[0], @b[5]
+
+	pxor	@b[3], @b[6]
+	pxor	@b[7], @b[3]
+	pxor	@b[5], @b[7]
+	pxor	@b[4], @b[3]
+	pxor	@b[5], @b[4]
+	pxor	@b[1], @b[3]
+
+	pxor	@b[7], @b[2]
+	pxor	@b[5], @b[1]
+___
+}
+
+sub OutBasisChange {
+# input in  lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b6, b1, b2, b4, b7, b0, b3, b5] < msb
+my @b=@_[0..7];
+$code.=<<___;
+	pxor	@b[6], @b[0]
+	pxor	@b[4], @b[1]
+	pxor	@b[0], @b[2]
+	pxor	@b[6], @b[4]
+	pxor	@b[1], @b[6]
+
+	pxor	@b[5], @b[1]
+	pxor	@b[3], @b[5]
+	pxor	@b[7], @b[3]
+	pxor	@b[5], @b[7]
+	pxor	@b[5], @b[2]
+
+	pxor	@b[7], @b[4]
+___
+}
+
+sub InvSbox {
+# input in lsb 	> [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb	> [b0, b1, b6, b4, b2, b7, b3, b5] < msb
+my @b=@_[0..7];
+my @t=@_[8..11];
+my @s=@_[12..15];
+	&InvInBasisChange	(@b);
+	&Inv_GF256		(@b[5,1,2,6,3,7,0,4],@t,@s);
+	&InvOutBasisChange	(@b[3,7,0,4,5,1,2,6]);
+}
+
+sub InvInBasisChange {		# OutBasisChange in reverse
+my @b=@_[5,1,2,6,3,7,0,4];
+$code.=<<___
+	pxor	@b[7], @b[4]
+
+	pxor	@b[5], @b[7]
+	pxor	@b[5], @b[2]
+	pxor	@b[7], @b[3]
+	pxor	@b[3], @b[5]
+	pxor	@b[5], @b[1]
+
+	pxor	@b[1], @b[6]
+	pxor	@b[0], @b[2]
+	pxor	@b[6], @b[4]
+	pxor	@b[6], @b[0]
+	pxor	@b[4], @b[1]
+___
+}
+
+sub InvOutBasisChange {		# InBasisChange in reverse
+my @b=@_[2,5,7,3,6,1,0,4];
+$code.=<<___;
+	pxor	@b[5], @b[1]
+	pxor	@b[7], @b[2]
+
+	pxor	@b[1], @b[3]
+	pxor	@b[5], @b[4]
+	pxor	@b[5], @b[7]
+	pxor	@b[4], @b[3]
+	 pxor 	@b[0], @b[5]
+	pxor	@b[7], @b[3]
+	 pxor	@b[2], @b[6]
+	 pxor	@b[1], @b[2]
+	pxor	@b[3], @b[6]
+
+	pxor	@b[0], @b[3]
+	pxor	@b[6], @b[5]
+___
+}
+
+sub Mul_GF4 {
+#;*************************************************************
+#;* Mul_GF4: Input x0-x1,y0-y1 Output x0-x1 Temp t0 (8) *
+#;*************************************************************
+my ($x0,$x1,$y0,$y1,$t0)=@_;
+$code.=<<___;
+	movdqa	$y0, $t0
+	pxor 	$y1, $t0
+	pand	$x0, $t0
+	pxor	$x1, $x0
+	pand	$y0, $x1
+	pand	$y1, $x0
+	pxor	$x1, $x0
+	pxor	$t0, $x1
+___
+}
+
+sub Mul_GF4_N {				# not used, see next subroutine
+# multiply and scale by N
+my ($x0,$x1,$y0,$y1,$t0)=@_;
+$code.=<<___;
+	movdqa	$y0, $t0
+	pxor	$y1, $t0
+	pand	$x0, $t0
+	pxor	$x1, $x0
+	pand	$y0, $x1
+	pand	$y1, $x0
+	pxor	$x0, $x1
+	pxor	$t0, $x0
+___
+}
+
+sub Mul_GF4_N_GF4 {
+# interleaved Mul_GF4_N and Mul_GF4
+my ($x0,$x1,$y0,$y1,$t0,
+    $x2,$x3,$y2,$y3,$t1)=@_;
+$code.=<<___;
+	movdqa	$y0, $t0
+	 movdqa	$y2, $t1
+	pxor	$y1, $t0
+	 pxor 	$y3, $t1
+	pand	$x0, $t0
+	 pand	$x2, $t1
+	pxor	$x1, $x0
+	 pxor	$x3, $x2
+	pand	$y0, $x1
+	 pand	$y2, $x3
+	pand	$y1, $x0
+	 pand	$y3, $x2
+	pxor	$x0, $x1
+	 pxor	$x3, $x2
+	pxor	$t0, $x0
+	 pxor	$t1, $x3
+___
+}
+sub Mul_GF16_2 {
+my @x=@_[0..7];
+my @y=@_[8..11];
+my @t=@_[12..15];
+$code.=<<___;
+	movdqa	@x[0], @t[0]
+	movdqa	@x[1], @t[1]
+___
+	&Mul_GF4  	(@x[0], @x[1], @y[0], @y[1], @t[2]);
+$code.=<<___;
+	pxor	@x[2], @t[0]
+	pxor	@x[3], @t[1]
+	pxor	@y[2], @y[0]
+	pxor	@y[3], @y[1]
+___
+	Mul_GF4_N_GF4	(@t[0], @t[1], @y[0], @y[1], @t[3],
+			 @x[2], @x[3], @y[2], @y[3], @t[2]);
+$code.=<<___;
+	pxor	@t[0], @x[0]
+	pxor	@t[0], @x[2]
+	pxor	@t[1], @x[1]
+	pxor	@t[1], @x[3]
+
+	movdqa	@x[4], @t[0]
+	movdqa	@x[5], @t[1]
+	pxor	@x[6], @t[0]
+	pxor	@x[7], @t[1]
+___
+	&Mul_GF4_N_GF4	(@t[0], @t[1], @y[0], @y[1], @t[3],
+			 @x[6], @x[7], @y[2], @y[3], @t[2]);
+$code.=<<___;
+	pxor	@y[2], @y[0]
+	pxor	@y[3], @y[1]
+___
+	&Mul_GF4  	(@x[4], @x[5], @y[0], @y[1], @t[3]);
+$code.=<<___;
+	pxor	@t[0], @x[4]
+	pxor	@t[0], @x[6]
+	pxor	@t[1], @x[5]
+	pxor	@t[1], @x[7]
+___
+}
+sub Inv_GF256 {
+#;********************************************************************
+#;* Inv_GF256: Input x0-x7 Output x0-x7 Temp t0-t3,s0-s3 (144)       *
+#;********************************************************************
+my @x=@_[0..7];
+my @t=@_[8..11];
+my @s=@_[12..15];
+# direct optimizations from hardware
+$code.=<<___;
+	movdqa	@x[4], @t[3]
+	movdqa	@x[5], @t[2]
+	movdqa	@x[1], @t[1]
+	movdqa	@x[7], @s[1]
+	movdqa	@x[0], @s[0]
+
+	pxor	@x[6], @t[3]
+	pxor	@x[7], @t[2]
+	pxor	@x[3], @t[1]
+	 movdqa	@t[3], @s[2]
+	pxor	@x[6], @s[1]
+	 movdqa	@t[2], @t[0]
+	pxor	@x[2], @s[0]
+	 movdqa	@t[3], @s[3]
+
+	por	@t[1], @t[2]
+	por	@s[0], @t[3]
+	pxor	@t[0], @s[3]
+	pand	@s[0], @s[2]
+	pxor	@t[1], @s[0]
+	pand	@t[1], @t[0]
+	pand	@s[0], @s[3]
+	movdqa	@x[3], @s[0]
+	pxor	@x[2], @s[0]
+	pand	@s[0], @s[1]
+	pxor	@s[1], @t[3]
+	pxor	@s[1], @t[2]
+	movdqa	@x[4], @s[1]
+	movdqa	@x[1], @s[0]
+	pxor	@x[5], @s[1]
+	pxor	@x[0], @s[0]
+	movdqa	@s[1], @t[1]
+	pand	@s[0], @s[1]
+	por	@s[0], @t[1]
+	pxor	@s[1], @t[0]
+	pxor	@s[3], @t[3]
+	pxor	@s[2], @t[2]
+	pxor	@s[3], @t[1]
+	movdqa	@x[7], @s[0]
+	pxor	@s[2], @t[0]
+	movdqa	@x[6], @s[1]
+	pxor	@s[2], @t[1]
+	movdqa	@x[5], @s[2]
+	pand	@x[3], @s[0]
+	movdqa	@x[4], @s[3]
+	pand	@x[2], @s[1]
+	pand	@x[1], @s[2]
+	por	@x[0], @s[3]
+	pxor	@s[0], @t[3]
+	pxor	@s[1], @t[2]
+	pxor	@s[2], @t[1]
+	pxor	@s[3], @t[0] 
+
+	#Inv_GF16 \t0, \t1, \t2, \t3, \s0, \s1, \s2, \s3
+
+	# new smaller inversion
+
+	movdqa	@t[3], @s[0]
+	pand	@t[1], @t[3]
+	pxor	@t[2], @s[0]
+
+	movdqa	@t[0], @s[2]
+	movdqa	@s[0], @s[3]
+	pxor	@t[3], @s[2]
+	pand	@s[2], @s[3]
+
+	movdqa	@t[1], @s[1]
+	pxor	@t[2], @s[3]
+	pxor	@t[0], @s[1]
+
+	pxor	@t[2], @t[3]
+
+	pand	@t[3], @s[1]
+
+	movdqa	@s[2], @t[2]
+	pxor	@t[0], @s[1]
+
+	pxor	@s[1], @t[2]
+	pxor	@s[1], @t[1]
+
+	pand	@t[0], @t[2]
+
+	pxor	@t[2], @s[2]
+	pxor	@t[2], @t[1]
+
+	pand	@s[3], @s[2]
+
+	pxor	@s[0], @s[2]
+___
+# output in s3, s2, s1, t1
+
+# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \t2, \t3, \t0, \t1, \s0, \s1, \s2, \s3
+
+# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \s3, \s2, \s1, \t1, \s0, \t0, \t2, \t3
+	&Mul_GF16_2(@x,@s[3,2,1],@t[1],@s[0],@t[0,2,3]);
+
+### output msb > [x3,x2,x1,x0,x7,x6,x5,x4] < lsb
+}
+
+# AES linear components
+
+sub ShiftRows {
+my @x=@_[0..7];
+my $mask=pop;
+$code.=<<___;
+	pxor	0x00($key),@x[0]
+	pxor	0x10($key),@x[1]
+	pshufb	$mask,@x[0]
+	pxor	0x20($key),@x[2]
+	pshufb	$mask,@x[1]
+	pxor	0x30($key),@x[3]
+	pshufb	$mask,@x[2]
+	pxor	0x40($key),@x[4]
+	pshufb	$mask,@x[3]
+	pxor	0x50($key),@x[5]
+	pshufb	$mask,@x[4]
+	pxor	0x60($key),@x[6]
+	pshufb	$mask,@x[5]
+	pxor	0x70($key),@x[7]
+	pshufb	$mask,@x[6]
+	lea	0x80($key),$key
+	pshufb	$mask,@x[7]
+___
+}
+
+sub MixColumns {
+# modified to emit output in order suitable for feeding back to aesenc[last]
+my @x=@_[0..7];
+my @t=@_[8..15];
+$code.=<<___;
+	pshufd	\$0x93, @x[0], @t[0]	# x0 <<< 32
+	pshufd	\$0x93, @x[1], @t[1]
+	 pxor	@t[0], @x[0]		# x0 ^ (x0 <<< 32)
+	pshufd	\$0x93, @x[2], @t[2]
+	 pxor	@t[1], @x[1]
+	pshufd	\$0x93, @x[3], @t[3]
+	 pxor	@t[2], @x[2]
+	pshufd	\$0x93, @x[4], @t[4]
+	 pxor	@t[3], @x[3]
+	pshufd	\$0x93, @x[5], @t[5]
+	 pxor	@t[4], @x[4]
+	pshufd	\$0x93, @x[6], @t[6]
+	 pxor	@t[5], @x[5]
+	pshufd	\$0x93, @x[7], @t[7]
+	 pxor	@t[6], @x[6]
+	 pxor	@t[7], @x[7]
+
+	pxor	@x[0], @t[1]
+	pxor	@x[7], @t[0]
+	pxor	@x[7], @t[1]
+	 pshufd	\$0x4E, @x[0], @x[0] 	# (x0 ^ (x0 <<< 32)) <<< 64)
+	pxor	@x[1], @t[2]
+	 pshufd	\$0x4E, @x[1], @x[1]
+	pxor	@x[4], @t[5]
+	 pxor	@t[0], @x[0]
+	pxor	@x[5], @t[6]
+	 pxor	@t[1], @x[1]
+	pxor	@x[3], @t[4]
+	 pshufd	\$0x4E, @x[4], @t[0]
+	pxor	@x[6], @t[7]
+	 pshufd	\$0x4E, @x[5], @t[1]
+	pxor	@x[2], @t[3]
+	 pshufd	\$0x4E, @x[3], @x[4]
+	pxor	@x[7], @t[3]
+	 pshufd	\$0x4E, @x[7], @x[5]
+	pxor	@x[7], @t[4]
+	 pshufd	\$0x4E, @x[6], @x[3]
+	pxor	@t[4], @t[0]
+	 pshufd	\$0x4E, @x[2], @x[6]
+	pxor	@t[5], @t[1]
+
+	pxor	@t[3], @x[4]
+	pxor	@t[7], @x[5]
+	pxor	@t[6], @x[3]
+	 movdqa	@t[0], @x[2]
+	pxor	@t[2], @x[6]
+	 movdqa	@t[1], @x[7]
+___
+}
+
+sub InvMixColumns {
+my @x=@_[0..7];
+my @t=@_[8..15];
+
+$code.=<<___;
+	# multiplication by 0x0e
+	pshufd	\$0x93, @x[7], @t[7]
+	movdqa	@x[2], @t[2]
+	pxor	@x[5], @x[7]		# 7 5
+	pxor	@x[5], @x[2]		# 2 5
+	pshufd	\$0x93, @x[0], @t[0]
+	movdqa	@x[5], @t[5]
+	pxor	@x[0], @x[5]		# 5 0		[1]
+	pxor	@x[1], @x[0]		# 0 1
+	pshufd	\$0x93, @x[1], @t[1]
+	pxor	@x[2], @x[1]		# 1 25
+	pxor	@x[6], @x[0]		# 01 6		[2]
+	pxor	@x[3], @x[1]		# 125 3		[4]
+	pshufd	\$0x93, @x[3], @t[3]
+	pxor	@x[0], @x[2]		# 25 016	[3]
+	pxor	@x[7], @x[3]		# 3 75
+	pxor	@x[6], @x[7]		# 75 6		[0]
+	pshufd	\$0x93, @x[6], @t[6]
+	movdqa	@x[4], @t[4]
+	pxor	@x[4], @x[6]		# 6 4
+	pxor	@x[3], @x[4]		# 4 375		[6]
+	pxor	@x[7], @x[3]		# 375 756=36
+	pxor	@t[5], @x[6]		# 64 5		[7]
+	pxor	@t[2], @x[3]		# 36 2
+	pxor	@t[4], @x[3]		# 362 4		[5]
+	pshufd	\$0x93, @t[5], @t[5]
+___
+					my @y = @x[7,5,0,2,1,3,4,6];
+$code.=<<___;
+	# multiplication by 0x0b
+	pxor	@y[0], @y[1]
+	pxor	@t[0], @y[0]
+	pxor	@t[1], @y[1]
+	pshufd	\$0x93, @t[2], @t[2]
+	pxor	@t[5], @y[0]
+	pxor	@t[6], @y[1]
+	pxor	@t[7], @y[0]
+	pshufd	\$0x93, @t[4], @t[4]
+	pxor	@t[6], @t[7]		# clobber t[7]
+	pxor	@y[0], @y[1]
+
+	pxor	@t[0], @y[3]
+	pshufd	\$0x93, @t[0], @t[0]
+	pxor	@t[1], @y[2]
+	pxor	@t[1], @y[4]
+	pxor	@t[2], @y[2]
+	pshufd	\$0x93, @t[1], @t[1]
+	pxor	@t[2], @y[3]
+	pxor	@t[2], @y[5]
+	pxor	@t[7], @y[2]
+	pshufd	\$0x93, @t[2], @t[2]
+	pxor	@t[3], @y[3]
+	pxor	@t[3], @y[6]
+	pxor	@t[3], @y[4]
+	pshufd	\$0x93, @t[3], @t[3]
+	pxor	@t[4], @y[7]
+	pxor	@t[4], @y[5]
+	pxor	@t[7], @y[7]
+	pxor	@t[5], @y[3]
+	pxor	@t[4], @y[4]
+	pxor	@t[5], @t[7]		# clobber t[7] even more
+
+	pxor	@t[7], @y[5]
+	pshufd	\$0x93, @t[4], @t[4]
+	pxor	@t[7], @y[6]
+	pxor	@t[7], @y[4]
+
+	pxor	@t[5], @t[7]
+	pshufd	\$0x93, @t[5], @t[5]
+	pxor	@t[6], @t[7]		# restore t[7]
+
+	# multiplication by 0x0d
+	pxor	@y[7], @y[4]
+	pxor	@t[4], @y[7]
+	pshufd	\$0x93, @t[6], @t[6]
+	pxor	@t[0], @y[2]
+	pxor	@t[5], @y[7]
+	pxor	@t[2], @y[2]
+	pshufd	\$0x93, @t[7], @t[7]
+
+	pxor	@y[1], @y[3]
+	pxor	@t[1], @y[1]
+	pxor	@t[0], @y[0]
+	pxor	@t[0], @y[3]
+	pxor	@t[5], @y[1]
+	pxor	@t[5], @y[0]
+	pxor	@t[7], @y[1]
+	pshufd	\$0x93, @t[0], @t[0]
+	pxor	@t[6], @y[0]
+	pxor	@y[1], @y[3]
+	pxor	@t[1], @y[4]
+	pshufd	\$0x93, @t[1], @t[1]
+
+	pxor	@t[7], @y[7]
+	pxor	@t[2], @y[4]
+	pxor	@t[2], @y[5]
+	pshufd	\$0x93, @t[2], @t[2]
+	pxor	@t[6], @y[2]
+	pxor	@t[3], @t[6]		# clobber t[6]
+	pxor	@y[7], @y[4]
+	pxor	@t[6], @y[3]
+
+	pxor	@t[6], @y[6]
+	pxor	@t[5], @y[5]
+	pxor	@t[4], @y[6]
+	pshufd	\$0x93, @t[4], @t[4]
+	pxor	@t[6], @y[5]
+	pxor	@t[7], @y[6]
+	pxor	@t[3], @t[6]		# restore t[6]
+
+	pshufd	\$0x93, @t[5], @t[5]
+	pshufd	\$0x93, @t[6], @t[6]
+	pshufd	\$0x93, @t[7], @t[7]
+	pshufd	\$0x93, @t[3], @t[3]
+
+	# multiplication by 0x09
+	pxor	@y[1], @y[4]
+	pxor	@y[1], @t[1]		# t[1]=y[1]
+	pxor	@t[5], @t[0]		# clobber t[0]
+	pxor	@t[5], @t[1]
+	pxor	@t[0], @y[3]
+	pxor	@y[0], @t[0]		# t[0]=y[0]
+	pxor	@t[6], @t[1]
+	pxor	@t[7], @t[6]		# clobber t[6]
+	pxor	@t[1], @y[4]
+	pxor	@t[4], @y[7]
+	pxor	@y[4], @t[4]		# t[4]=y[4]
+	pxor	@t[3], @y[6]
+	pxor	@y[3], @t[3]		# t[3]=y[3]
+	pxor	@t[2], @y[5]
+	pxor	@y[2], @t[2]		# t[2]=y[2]
+	pxor	@t[7], @t[3]
+	pxor	@y[5], @t[5]		# t[5]=y[5]
+	pxor	@t[6], @t[2]
+	pxor	@t[6], @t[5]
+	pxor	@y[6], @t[6]		# t[6]=y[6]
+	pxor	@y[7], @t[7]		# t[7]=y[7]
+
+	movdqa	@t[0],@XMM[0]
+	movdqa	@t[1],@XMM[1]
+	movdqa	@t[2],@XMM[2]
+	movdqa	@t[3],@XMM[3]
+	movdqa	@t[4],@XMM[4]
+	movdqa	@t[5],@XMM[5]
+	movdqa	@t[6],@XMM[6]
+	movdqa	@t[7],@XMM[7]
+___
+}
+
+sub aesenc {				# not used
+my @b=@_[0..7];
+my @t=@_[8..15];
+$code.=<<___;
+	movdqa	0x30($const),@t[0]	# .LSR
+___
+	&ShiftRows	(@b,@t[0]);
+	&Sbox		(@b,@t);
+	&MixColumns	(@b[0,1,4,6,3,7,2,5],@t);
+}
+
+sub aesenclast {			# not used
+my @b=@_[0..7];
+my @t=@_[8..15];
+$code.=<<___;
+	movdqa	0x40($const),@t[0]	# .LSRM0
+___
+	&ShiftRows	(@b,@t[0]);
+	&Sbox		(@b,@t);
+$code.=<<___
+	pxor	0x00($key),@b[0]
+	pxor	0x10($key),@b[1]
+	pxor	0x20($key),@b[4]
+	pxor	0x30($key),@b[6]
+	pxor	0x40($key),@b[3]
+	pxor	0x50($key),@b[7]
+	pxor	0x60($key),@b[2]
+	pxor	0x70($key),@b[5]
+___
+}
+
+sub swapmove {
+my ($a,$b,$n,$mask,$t)=@_;
+$code.=<<___;
+	movdqa	$b,$t
+	psrlq	\$$n,$b
+	pxor  	$a,$b
+	pand	$mask,$b
+	pxor	$b,$a
+	psllq	\$$n,$b
+	pxor	$t,$b
+___
+}
+sub swapmove2x {
+my ($a0,$b0,$a1,$b1,$n,$mask,$t0,$t1)=@_;
+$code.=<<___;
+	movdqa	$b0,$t0
+	psrlq	\$$n,$b0
+	 movdqa	$b1,$t1
+	 psrlq	\$$n,$b1
+	pxor  	$a0,$b0
+	 pxor  	$a1,$b1
+	pand	$mask,$b0
+	 pand	$mask,$b1
+	pxor	$b0,$a0
+	psllq	\$$n,$b0
+	 pxor	$b1,$a1
+	 psllq	\$$n,$b1
+	pxor	$t0,$b0
+	 pxor	$t1,$b1
+___
+}
+
+sub bitslice {
+my @x=reverse(@_[0..7]);
+my ($t0,$t1,$t2,$t3)=@_[8..11];
+$code.=<<___;
+	movdqa	0x00($const),$t0	# .LBS0
+	movdqa	0x10($const),$t1	# .LBS1
+___
+	&swapmove2x(@x[0,1,2,3],1,$t0,$t2,$t3);
+	&swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3);
+$code.=<<___;
+	movdqa	0x20($const),$t0	# .LBS2
+___
+	&swapmove2x(@x[0,2,1,3],2,$t1,$t2,$t3);
+	&swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3);
+
+	&swapmove2x(@x[0,4,1,5],4,$t0,$t2,$t3);
+	&swapmove2x(@x[2,6,3,7],4,$t0,$t2,$t3);
+}
+
+$code.=<<___;
+.text
+
+.extern	asm_AES_encrypt
+.extern	asm_AES_decrypt
+
+.type	_bsaes_encrypt8,\@abi-omnipotent
+.align	64
+_bsaes_encrypt8:
+	lea	.LBS0(%rip), $const	# constants table
+
+	movdqa	($key), @XMM[9]		# round 0 key
+	lea	0x10($key), $key
+	movdqa	0x50($const), @XMM[8]	# .LM0SR
+	pxor	@XMM[9], @XMM[0]	# xor with round0 key
+	pxor	@XMM[9], @XMM[1]
+	 pshufb	@XMM[8], @XMM[0]
+	pxor	@XMM[9], @XMM[2]
+	 pshufb	@XMM[8], @XMM[1]
+	pxor	@XMM[9], @XMM[3]
+	 pshufb	@XMM[8], @XMM[2]
+	pxor	@XMM[9], @XMM[4]
+	 pshufb	@XMM[8], @XMM[3]
+	pxor	@XMM[9], @XMM[5]
+	 pshufb	@XMM[8], @XMM[4]
+	pxor	@XMM[9], @XMM[6]
+	 pshufb	@XMM[8], @XMM[5]
+	pxor	@XMM[9], @XMM[7]
+	 pshufb	@XMM[8], @XMM[6]
+	 pshufb	@XMM[8], @XMM[7]
+_bsaes_encrypt8_bitslice:
+___
+	&bitslice	(@XMM[0..7, 8..11]);
+$code.=<<___;
+	dec	$rounds
+	jmp	.Lenc_sbox
+.align	16
+.Lenc_loop:
+___
+	&ShiftRows	(@XMM[0..7, 8]);
+$code.=".Lenc_sbox:\n";
+	&Sbox		(@XMM[0..7, 8..15]);
+$code.=<<___;
+	dec	$rounds
+	jl	.Lenc_done
+___
+	&MixColumns	(@XMM[0,1,4,6,3,7,2,5, 8..15]);
+$code.=<<___;
+	movdqa	0x30($const), @XMM[8]	# .LSR
+	jnz	.Lenc_loop
+	movdqa	0x40($const), @XMM[8]	# .LSRM0
+	jmp	.Lenc_loop
+.align	16
+.Lenc_done:
+___
+	# output in lsb > [t0, t1, t4, t6, t3, t7, t2, t5] < msb
+	&bitslice	(@XMM[0,1,4,6,3,7,2,5, 8..11]);
+$code.=<<___;
+	movdqa	($key), @XMM[8]		# last round key
+	pxor	@XMM[8], @XMM[4]
+	pxor	@XMM[8], @XMM[6]
+	pxor	@XMM[8], @XMM[3]
+	pxor	@XMM[8], @XMM[7]
+	pxor	@XMM[8], @XMM[2]
+	pxor	@XMM[8], @XMM[5]
+	pxor	@XMM[8], @XMM[0]
+	pxor	@XMM[8], @XMM[1]
+	ret
+.size	_bsaes_encrypt8,.-_bsaes_encrypt8
+
+.type	_bsaes_decrypt8,\@abi-omnipotent
+.align	64
+_bsaes_decrypt8:
+	lea	.LBS0(%rip), $const	# constants table
+
+	movdqa	($key), @XMM[9]		# round 0 key
+	lea	0x10($key), $key
+	movdqa	-0x30($const), @XMM[8]	# .LM0ISR
+	pxor	@XMM[9], @XMM[0]	# xor with round0 key
+	pxor	@XMM[9], @XMM[1]
+	 pshufb	@XMM[8], @XMM[0]
+	pxor	@XMM[9], @XMM[2]
+	 pshufb	@XMM[8], @XMM[1]
+	pxor	@XMM[9], @XMM[3]
+	 pshufb	@XMM[8], @XMM[2]
+	pxor	@XMM[9], @XMM[4]
+	 pshufb	@XMM[8], @XMM[3]
+	pxor	@XMM[9], @XMM[5]
+	 pshufb	@XMM[8], @XMM[4]
+	pxor	@XMM[9], @XMM[6]
+	 pshufb	@XMM[8], @XMM[5]
+	pxor	@XMM[9], @XMM[7]
+	 pshufb	@XMM[8], @XMM[6]
+	 pshufb	@XMM[8], @XMM[7]
+___
+	&bitslice	(@XMM[0..7, 8..11]);
+$code.=<<___;
+	dec	$rounds
+	jmp	.Ldec_sbox
+.align	16
+.Ldec_loop:
+___
+	&ShiftRows	(@XMM[0..7, 8]);
+$code.=".Ldec_sbox:\n";
+	&InvSbox	(@XMM[0..7, 8..15]);
+$code.=<<___;
+	dec	$rounds
+	jl	.Ldec_done
+___
+	&InvMixColumns	(@XMM[0,1,6,4,2,7,3,5, 8..15]);
+$code.=<<___;
+	movdqa	-0x10($const), @XMM[8]	# .LISR
+	jnz	.Ldec_loop
+	movdqa	-0x20($const), @XMM[8]	# .LISRM0
+	jmp	.Ldec_loop
+.align	16
+.Ldec_done:
+___
+	&bitslice	(@XMM[0,1,6,4,2,7,3,5, 8..11]);
+$code.=<<___;
+	movdqa	($key), @XMM[8]		# last round key
+	pxor	@XMM[8], @XMM[6]
+	pxor	@XMM[8], @XMM[4]
+	pxor	@XMM[8], @XMM[2]
+	pxor	@XMM[8], @XMM[7]
+	pxor	@XMM[8], @XMM[3]
+	pxor	@XMM[8], @XMM[5]
+	pxor	@XMM[8], @XMM[0]
+	pxor	@XMM[8], @XMM[1]
+	ret
+.size	_bsaes_decrypt8,.-_bsaes_decrypt8
+___
+}
+{
+my ($out,$inp,$rounds,$const)=("%rax","%rcx","%r10d","%r11");
+
+sub bitslice_key {
+my @x=reverse(@_[0..7]);
+my ($bs0,$bs1,$bs2,$t2,$t3)=@_[8..12];
+
+	&swapmove	(@x[0,1],1,$bs0,$t2,$t3);
+$code.=<<___;
+	#&swapmove(@x[2,3],1,$t0,$t2,$t3);
+	movdqa	@x[0], @x[2]
+	movdqa	@x[1], @x[3]
+___
+	#&swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3);
+
+	&swapmove2x	(@x[0,2,1,3],2,$bs1,$t2,$t3);
+$code.=<<___;
+	#&swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3);
+	movdqa	@x[0], @x[4]
+	movdqa	@x[2], @x[6]
+	movdqa	@x[1], @x[5]
+	movdqa	@x[3], @x[7]
+___
+	&swapmove2x	(@x[0,4,1,5],4,$bs2,$t2,$t3);
+	&swapmove2x	(@x[2,6,3,7],4,$bs2,$t2,$t3);
+}
+
+$code.=<<___;
+.type	_bsaes_key_convert,\@abi-omnipotent
+.align	16
+_bsaes_key_convert:
+	lea	.Lmasks(%rip), $const
+	movdqu	($inp), %xmm7		# load round 0 key
+	lea	0x10($inp), $inp
+	movdqa	0x00($const), %xmm0	# 0x01...
+	movdqa	0x10($const), %xmm1	# 0x02...
+	movdqa	0x20($const), %xmm2	# 0x04...
+	movdqa	0x30($const), %xmm3	# 0x08...
+	movdqa	0x40($const), %xmm4	# .LM0
+	pcmpeqd	%xmm5, %xmm5		# .LNOT
+
+	movdqu	($inp), %xmm6		# load round 1 key
+	movdqa	%xmm7, ($out)		# save round 0 key
+	lea	0x10($out), $out
+	dec	$rounds
+	jmp	.Lkey_loop
+.align	16
+.Lkey_loop:
+	pshufb	%xmm4, %xmm6		# .LM0
+
+	movdqa	%xmm0,	%xmm8
+	movdqa	%xmm1,	%xmm9
+
+	pand	%xmm6,	%xmm8
+	pand	%xmm6,	%xmm9
+	movdqa	%xmm2,	%xmm10
+	pcmpeqb	%xmm0,	%xmm8
+	psllq	\$4,	%xmm0		# 0x10...
+	movdqa	%xmm3,	%xmm11
+	pcmpeqb	%xmm1,	%xmm9
+	psllq	\$4,	%xmm1		# 0x20...
+
+	pand	%xmm6,	%xmm10
+	pand	%xmm6,	%xmm11
+	movdqa	%xmm0,	%xmm12
+	pcmpeqb	%xmm2,	%xmm10
+	psllq	\$4,	%xmm2		# 0x40...
+	movdqa	%xmm1,	%xmm13
+	pcmpeqb	%xmm3,	%xmm11
+	psllq	\$4,	%xmm3		# 0x80...
+
+	movdqa	%xmm2,	%xmm14
+	movdqa	%xmm3,	%xmm15
+	 pxor	%xmm5,	%xmm8		# "pnot"
+	 pxor	%xmm5,	%xmm9
+
+	pand	%xmm6,	%xmm12
+	pand	%xmm6,	%xmm13
+	 movdqa	%xmm8, 0x00($out)	# write bit-sliced round key
+	pcmpeqb	%xmm0,	%xmm12
+	psrlq	\$4,	%xmm0		# 0x01...
+	 movdqa	%xmm9, 0x10($out)
+	pcmpeqb	%xmm1,	%xmm13
+	psrlq	\$4,	%xmm1		# 0x02...
+	 lea	0x10($inp), $inp
+
+	pand	%xmm6,	%xmm14
+	pand	%xmm6,	%xmm15
+	 movdqa	%xmm10, 0x20($out)
+	pcmpeqb	%xmm2,	%xmm14
+	psrlq	\$4,	%xmm2		# 0x04...
+	 movdqa	%xmm11, 0x30($out)
+	pcmpeqb	%xmm3,	%xmm15
+	psrlq	\$4,	%xmm3		# 0x08...
+	 movdqu	($inp), %xmm6		# load next round key
+
+	pxor	%xmm5, %xmm13		# "pnot"
+	pxor	%xmm5, %xmm14
+	movdqa	%xmm12, 0x40($out)
+	movdqa	%xmm13, 0x50($out)
+	movdqa	%xmm14, 0x60($out)
+	movdqa	%xmm15, 0x70($out)
+	lea	0x80($out),$out
+	dec	$rounds
+	jnz	.Lkey_loop
+
+	movdqa	0x50($const), %xmm7	# .L63
+	#movdqa	%xmm6, ($out)		# don't save last round key
+	ret
+.size	_bsaes_key_convert,.-_bsaes_key_convert
+___
+}
+
+if (0 && !$win64) {	# following four functions are unsupported interface
+			# used for benchmarking...
+$code.=<<___;
+.globl	bsaes_enc_key_convert
+.type	bsaes_enc_key_convert,\@function,2
+.align	16
+bsaes_enc_key_convert:
+	mov	240($inp),%r10d		# pass rounds
+	mov	$inp,%rcx		# pass key
+	mov	$out,%rax		# pass key schedule
+	call	_bsaes_key_convert
+	pxor	%xmm6,%xmm7		# fix up last round key
+	movdqa	%xmm7,(%rax)		# save last round key
+	ret
+.size	bsaes_enc_key_convert,.-bsaes_enc_key_convert
+
+.globl	bsaes_encrypt_128
+.type	bsaes_encrypt_128,\@function,4
+.align	16
+bsaes_encrypt_128:
+.Lenc128_loop:
+	movdqu	0x00($inp), @XMM[0]	# load input
+	movdqu	0x10($inp), @XMM[1]
+	movdqu	0x20($inp), @XMM[2]
+	movdqu	0x30($inp), @XMM[3]
+	movdqu	0x40($inp), @XMM[4]
+	movdqu	0x50($inp), @XMM[5]
+	movdqu	0x60($inp), @XMM[6]
+	movdqu	0x70($inp), @XMM[7]
+	mov	$key, %rax		# pass the $key
+	lea	0x80($inp), $inp
+	mov	\$10,%r10d
+
+	call	_bsaes_encrypt8
+
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[4], 0x20($out)
+	movdqu	@XMM[6], 0x30($out)
+	movdqu	@XMM[3], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[2], 0x60($out)
+	movdqu	@XMM[5], 0x70($out)
+	lea	0x80($out), $out
+	sub	\$0x80,$len
+	ja	.Lenc128_loop
+	ret
+.size	bsaes_encrypt_128,.-bsaes_encrypt_128
+
+.globl	bsaes_dec_key_convert
+.type	bsaes_dec_key_convert,\@function,2
+.align	16
+bsaes_dec_key_convert:
+	mov	240($inp),%r10d		# pass rounds
+	mov	$inp,%rcx		# pass key
+	mov	$out,%rax		# pass key schedule
+	call	_bsaes_key_convert
+	pxor	($out),%xmm7		# fix up round 0 key
+	movdqa	%xmm6,(%rax)		# save last round key
+	movdqa	%xmm7,($out)
+	ret
+.size	bsaes_dec_key_convert,.-bsaes_dec_key_convert
+
+.globl	bsaes_decrypt_128
+.type	bsaes_decrypt_128,\@function,4
+.align	16
+bsaes_decrypt_128:
+.Ldec128_loop:
+	movdqu	0x00($inp), @XMM[0]	# load input
+	movdqu	0x10($inp), @XMM[1]
+	movdqu	0x20($inp), @XMM[2]
+	movdqu	0x30($inp), @XMM[3]
+	movdqu	0x40($inp), @XMM[4]
+	movdqu	0x50($inp), @XMM[5]
+	movdqu	0x60($inp), @XMM[6]
+	movdqu	0x70($inp), @XMM[7]
+	mov	$key, %rax		# pass the $key
+	lea	0x80($inp), $inp
+	mov	\$10,%r10d
+
+	call	_bsaes_decrypt8
+
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[3], 0x60($out)
+	movdqu	@XMM[5], 0x70($out)
+	lea	0x80($out), $out
+	sub	\$0x80,$len
+	ja	.Ldec128_loop
+	ret
+.size	bsaes_decrypt_128,.-bsaes_decrypt_128
+___
+}
+{
+######################################################################
+#
+# OpenSSL interface
+#
+my ($arg1,$arg2,$arg3,$arg4,$arg5,$arg6)=$win64	? ("%rcx","%rdx","%r8","%r9","%r10","%r11d")
+						: ("%rdi","%rsi","%rdx","%rcx","%r8","%r9d");
+my ($inp,$out,$len,$key)=("%r12","%r13","%r14","%r15");
+
+if ($ecb) {
+$code.=<<___;
+.globl	bsaes_ecb_encrypt_blocks
+.type	bsaes_ecb_encrypt_blocks,\@abi-omnipotent
+.align	16
+bsaes_ecb_encrypt_blocks:
+	mov	%rsp, %rax
+.Lecb_enc_prologue:
+	push	%rbp
+	push	%rbx
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	lea	-0x48(%rsp),%rsp
+___
+$code.=<<___ if ($win64);
+	lea	-0xa0(%rsp), %rsp
+	movaps	%xmm6, 0x40(%rsp)
+	movaps	%xmm7, 0x50(%rsp)
+	movaps	%xmm8, 0x60(%rsp)
+	movaps	%xmm9, 0x70(%rsp)
+	movaps	%xmm10, 0x80(%rsp)
+	movaps	%xmm11, 0x90(%rsp)
+	movaps	%xmm12, 0xa0(%rsp)
+	movaps	%xmm13, 0xb0(%rsp)
+	movaps	%xmm14, 0xc0(%rsp)
+	movaps	%xmm15, 0xd0(%rsp)
+.Lecb_enc_body:
+___
+$code.=<<___;
+	mov	%rsp,%rbp		# backup %rsp
+	mov	240($arg4),%eax		# rounds
+	mov	$arg1,$inp		# backup arguments
+	mov	$arg2,$out
+	mov	$arg3,$len
+	mov	$arg4,$key
+	cmp	\$8,$arg3
+	jb	.Lecb_enc_short
+
+	mov	%eax,%ebx		# backup rounds
+	shl	\$7,%rax		# 128 bytes per inner round key
+	sub	\$`128-32`,%rax		# size of bit-sliced key schedule
+	sub	%rax,%rsp
+	mov	%rsp,%rax		# pass key schedule
+	mov	$key,%rcx		# pass key
+	mov	%ebx,%r10d		# pass rounds
+	call	_bsaes_key_convert
+	pxor	%xmm6,%xmm7		# fix up last round key
+	movdqa	%xmm7,(%rax)		# save last round key
+
+	sub	\$8,$len
+.Lecb_enc_loop:
+	movdqu	0x00($inp), @XMM[0]	# load input
+	movdqu	0x10($inp), @XMM[1]
+	movdqu	0x20($inp), @XMM[2]
+	movdqu	0x30($inp), @XMM[3]
+	movdqu	0x40($inp), @XMM[4]
+	movdqu	0x50($inp), @XMM[5]
+	mov	%rsp, %rax		# pass key schedule
+	movdqu	0x60($inp), @XMM[6]
+	mov	%ebx,%r10d		# pass rounds
+	movdqu	0x70($inp), @XMM[7]
+	lea	0x80($inp), $inp
+
+	call	_bsaes_encrypt8
+
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[4], 0x20($out)
+	movdqu	@XMM[6], 0x30($out)
+	movdqu	@XMM[3], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[2], 0x60($out)
+	movdqu	@XMM[5], 0x70($out)
+	lea	0x80($out), $out
+	sub	\$8,$len
+	jnc	.Lecb_enc_loop
+
+	add	\$8,$len
+	jz	.Lecb_enc_done
+
+	movdqu	0x00($inp), @XMM[0]	# load input
+	mov	%rsp, %rax		# pass key schedule
+	mov	%ebx,%r10d		# pass rounds
+	cmp	\$2,$len
+	jb	.Lecb_enc_one
+	movdqu	0x10($inp), @XMM[1]
+	je	.Lecb_enc_two
+	movdqu	0x20($inp), @XMM[2]
+	cmp	\$4,$len
+	jb	.Lecb_enc_three
+	movdqu	0x30($inp), @XMM[3]
+	je	.Lecb_enc_four
+	movdqu	0x40($inp), @XMM[4]
+	cmp	\$6,$len
+	jb	.Lecb_enc_five
+	movdqu	0x50($inp), @XMM[5]
+	je	.Lecb_enc_six
+	movdqu	0x60($inp), @XMM[6]
+	call	_bsaes_encrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[4], 0x20($out)
+	movdqu	@XMM[6], 0x30($out)
+	movdqu	@XMM[3], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[2], 0x60($out)
+	jmp	.Lecb_enc_done
+.align	16
+.Lecb_enc_six:
+	call	_bsaes_encrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[4], 0x20($out)
+	movdqu	@XMM[6], 0x30($out)
+	movdqu	@XMM[3], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	jmp	.Lecb_enc_done
+.align	16
+.Lecb_enc_five:
+	call	_bsaes_encrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[4], 0x20($out)
+	movdqu	@XMM[6], 0x30($out)
+	movdqu	@XMM[3], 0x40($out)
+	jmp	.Lecb_enc_done
+.align	16
+.Lecb_enc_four:
+	call	_bsaes_encrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[4], 0x20($out)
+	movdqu	@XMM[6], 0x30($out)
+	jmp	.Lecb_enc_done
+.align	16
+.Lecb_enc_three:
+	call	_bsaes_encrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[4], 0x20($out)
+	jmp	.Lecb_enc_done
+.align	16
+.Lecb_enc_two:
+	call	_bsaes_encrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	jmp	.Lecb_enc_done
+.align	16
+.Lecb_enc_one:
+	call	_bsaes_encrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	jmp	.Lecb_enc_done
+.align	16
+.Lecb_enc_short:
+	lea	($inp), $arg1
+	lea	($out), $arg2
+	lea	($key), $arg3
+	call	asm_AES_encrypt
+	lea	16($inp), $inp
+	lea	16($out), $out
+	dec	$len
+	jnz	.Lecb_enc_short
+
+.Lecb_enc_done:
+	lea	(%rsp),%rax
+	pxor	%xmm0, %xmm0
+.Lecb_enc_bzero:			# wipe key schedule [if any]
+	movdqa	%xmm0, 0x00(%rax)
+	movdqa	%xmm0, 0x10(%rax)
+	lea	0x20(%rax), %rax
+	cmp	%rax, %rbp
+	jb	.Lecb_enc_bzero
+
+	lea	(%rbp),%rsp		# restore %rsp
+___
+$code.=<<___ if ($win64);
+	movaps	0x40(%rbp), %xmm6
+	movaps	0x50(%rbp), %xmm7
+	movaps	0x60(%rbp), %xmm8
+	movaps	0x70(%rbp), %xmm9
+	movaps	0x80(%rbp), %xmm10
+	movaps	0x90(%rbp), %xmm11
+	movaps	0xa0(%rbp), %xmm12
+	movaps	0xb0(%rbp), %xmm13
+	movaps	0xc0(%rbp), %xmm14
+	movaps	0xd0(%rbp), %xmm15
+	lea	0xa0(%rbp), %rsp
+___
+$code.=<<___;
+	mov	0x48(%rsp), %r15
+	mov	0x50(%rsp), %r14
+	mov	0x58(%rsp), %r13
+	mov	0x60(%rsp), %r12
+	mov	0x68(%rsp), %rbx
+	mov	0x70(%rsp), %rax
+	lea	0x78(%rsp), %rsp
+	mov	%rax, %rbp
+.Lecb_enc_epilogue:
+	ret
+.size	bsaes_ecb_encrypt_blocks,.-bsaes_ecb_encrypt_blocks
+
+.globl	bsaes_ecb_decrypt_blocks
+.type	bsaes_ecb_decrypt_blocks,\@abi-omnipotent
+.align	16
+bsaes_ecb_decrypt_blocks:
+	mov	%rsp, %rax
+.Lecb_dec_prologue:
+	push	%rbp
+	push	%rbx
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	lea	-0x48(%rsp),%rsp
+___
+$code.=<<___ if ($win64);
+	lea	-0xa0(%rsp), %rsp
+	movaps	%xmm6, 0x40(%rsp)
+	movaps	%xmm7, 0x50(%rsp)
+	movaps	%xmm8, 0x60(%rsp)
+	movaps	%xmm9, 0x70(%rsp)
+	movaps	%xmm10, 0x80(%rsp)
+	movaps	%xmm11, 0x90(%rsp)
+	movaps	%xmm12, 0xa0(%rsp)
+	movaps	%xmm13, 0xb0(%rsp)
+	movaps	%xmm14, 0xc0(%rsp)
+	movaps	%xmm15, 0xd0(%rsp)
+.Lecb_dec_body:
+___
+$code.=<<___;
+	mov	%rsp,%rbp		# backup %rsp
+	mov	240($arg4),%eax		# rounds
+	mov	$arg1,$inp		# backup arguments
+	mov	$arg2,$out
+	mov	$arg3,$len
+	mov	$arg4,$key
+	cmp	\$8,$arg3
+	jb	.Lecb_dec_short
+
+	mov	%eax,%ebx		# backup rounds
+	shl	\$7,%rax		# 128 bytes per inner round key
+	sub	\$`128-32`,%rax		# size of bit-sliced key schedule
+	sub	%rax,%rsp
+	mov	%rsp,%rax		# pass key schedule
+	mov	$key,%rcx		# pass key
+	mov	%ebx,%r10d		# pass rounds
+	call	_bsaes_key_convert
+	pxor	(%rsp),%xmm7		# fix up 0 round key
+	movdqa	%xmm6,(%rax)		# save last round key
+	movdqa	%xmm7,(%rsp)
+
+	sub	\$8,$len
+.Lecb_dec_loop:
+	movdqu	0x00($inp), @XMM[0]	# load input
+	movdqu	0x10($inp), @XMM[1]
+	movdqu	0x20($inp), @XMM[2]
+	movdqu	0x30($inp), @XMM[3]
+	movdqu	0x40($inp), @XMM[4]
+	movdqu	0x50($inp), @XMM[5]
+	mov	%rsp, %rax		# pass key schedule
+	movdqu	0x60($inp), @XMM[6]
+	mov	%ebx,%r10d		# pass rounds
+	movdqu	0x70($inp), @XMM[7]
+	lea	0x80($inp), $inp
+
+	call	_bsaes_decrypt8
+
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[3], 0x60($out)
+	movdqu	@XMM[5], 0x70($out)
+	lea	0x80($out), $out
+	sub	\$8,$len
+	jnc	.Lecb_dec_loop
+
+	add	\$8,$len
+	jz	.Lecb_dec_done
+
+	movdqu	0x00($inp), @XMM[0]	# load input
+	mov	%rsp, %rax		# pass key schedule
+	mov	%ebx,%r10d		# pass rounds
+	cmp	\$2,$len
+	jb	.Lecb_dec_one
+	movdqu	0x10($inp), @XMM[1]
+	je	.Lecb_dec_two
+	movdqu	0x20($inp), @XMM[2]
+	cmp	\$4,$len
+	jb	.Lecb_dec_three
+	movdqu	0x30($inp), @XMM[3]
+	je	.Lecb_dec_four
+	movdqu	0x40($inp), @XMM[4]
+	cmp	\$6,$len
+	jb	.Lecb_dec_five
+	movdqu	0x50($inp), @XMM[5]
+	je	.Lecb_dec_six
+	movdqu	0x60($inp), @XMM[6]
+	call	_bsaes_decrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[3], 0x60($out)
+	jmp	.Lecb_dec_done
+.align	16
+.Lecb_dec_six:
+	call	_bsaes_decrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	jmp	.Lecb_dec_done
+.align	16
+.Lecb_dec_five:
+	call	_bsaes_decrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	jmp	.Lecb_dec_done
+.align	16
+.Lecb_dec_four:
+	call	_bsaes_decrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	jmp	.Lecb_dec_done
+.align	16
+.Lecb_dec_three:
+	call	_bsaes_decrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	jmp	.Lecb_dec_done
+.align	16
+.Lecb_dec_two:
+	call	_bsaes_decrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	jmp	.Lecb_dec_done
+.align	16
+.Lecb_dec_one:
+	call	_bsaes_decrypt8
+	movdqu	@XMM[0], 0x00($out)	# write output
+	jmp	.Lecb_dec_done
+.align	16
+.Lecb_dec_short:
+	lea	($inp), $arg1
+	lea	($out), $arg2
+	lea	($key), $arg3
+	call	asm_AES_decrypt
+	lea	16($inp), $inp
+	lea	16($out), $out
+	dec	$len
+	jnz	.Lecb_dec_short
+
+.Lecb_dec_done:
+	lea	(%rsp),%rax
+	pxor	%xmm0, %xmm0
+.Lecb_dec_bzero:			# wipe key schedule [if any]
+	movdqa	%xmm0, 0x00(%rax)
+	movdqa	%xmm0, 0x10(%rax)
+	lea	0x20(%rax), %rax
+	cmp	%rax, %rbp
+	jb	.Lecb_dec_bzero
+
+	lea	(%rbp),%rsp		# restore %rsp
+___
+$code.=<<___ if ($win64);
+	movaps	0x40(%rbp), %xmm6
+	movaps	0x50(%rbp), %xmm7
+	movaps	0x60(%rbp), %xmm8
+	movaps	0x70(%rbp), %xmm9
+	movaps	0x80(%rbp), %xmm10
+	movaps	0x90(%rbp), %xmm11
+	movaps	0xa0(%rbp), %xmm12
+	movaps	0xb0(%rbp), %xmm13
+	movaps	0xc0(%rbp), %xmm14
+	movaps	0xd0(%rbp), %xmm15
+	lea	0xa0(%rbp), %rsp
+___
+$code.=<<___;
+	mov	0x48(%rsp), %r15
+	mov	0x50(%rsp), %r14
+	mov	0x58(%rsp), %r13
+	mov	0x60(%rsp), %r12
+	mov	0x68(%rsp), %rbx
+	mov	0x70(%rsp), %rax
+	lea	0x78(%rsp), %rsp
+	mov	%rax, %rbp
+.Lecb_dec_epilogue:
+	ret
+.size	bsaes_ecb_decrypt_blocks,.-bsaes_ecb_decrypt_blocks
+___
+}
+$code.=<<___;
+.extern	asm_AES_cbc_encrypt
+.globl	bsaes_cbc_encrypt
+.type	bsaes_cbc_encrypt,\@abi-omnipotent
+.align	16
+bsaes_cbc_encrypt:
+___
+$code.=<<___ if ($win64);
+	mov	48(%rsp),$arg6		# pull direction flag
+___
+$code.=<<___;
+	cmp	\$0,$arg6
+	jne	asm_AES_cbc_encrypt
+	cmp	\$128,$arg3
+	jb	asm_AES_cbc_encrypt
+
+	mov	%rsp, %rax
+.Lcbc_dec_prologue:
+	push	%rbp
+	push	%rbx
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	lea	-0x48(%rsp), %rsp
+___
+$code.=<<___ if ($win64);
+	mov	0xa0(%rsp),$arg5	# pull ivp
+	lea	-0xa0(%rsp), %rsp
+	movaps	%xmm6, 0x40(%rsp)
+	movaps	%xmm7, 0x50(%rsp)
+	movaps	%xmm8, 0x60(%rsp)
+	movaps	%xmm9, 0x70(%rsp)
+	movaps	%xmm10, 0x80(%rsp)
+	movaps	%xmm11, 0x90(%rsp)
+	movaps	%xmm12, 0xa0(%rsp)
+	movaps	%xmm13, 0xb0(%rsp)
+	movaps	%xmm14, 0xc0(%rsp)
+	movaps	%xmm15, 0xd0(%rsp)
+.Lcbc_dec_body:
+___
+$code.=<<___;
+	mov	%rsp, %rbp		# backup %rsp
+	mov	240($arg4), %eax	# rounds
+	mov	$arg1, $inp		# backup arguments
+	mov	$arg2, $out
+	mov	$arg3, $len
+	mov	$arg4, $key
+	mov	$arg5, %rbx
+	shr	\$4, $len		# bytes to blocks
+
+	mov	%eax, %edx		# rounds
+	shl	\$7, %rax		# 128 bytes per inner round key
+	sub	\$`128-32`, %rax	# size of bit-sliced key schedule
+	sub	%rax, %rsp
+
+	mov	%rsp, %rax		# pass key schedule
+	mov	$key, %rcx		# pass key
+	mov	%edx, %r10d		# pass rounds
+	call	_bsaes_key_convert
+	pxor	(%rsp),%xmm7		# fix up 0 round key
+	movdqa	%xmm6,(%rax)		# save last round key
+	movdqa	%xmm7,(%rsp)
+
+	movdqu	(%rbx), @XMM[15]	# load IV
+	sub	\$8,$len
+.Lcbc_dec_loop:
+	movdqu	0x00($inp), @XMM[0]	# load input
+	movdqu	0x10($inp), @XMM[1]
+	movdqu	0x20($inp), @XMM[2]
+	movdqu	0x30($inp), @XMM[3]
+	movdqu	0x40($inp), @XMM[4]
+	movdqu	0x50($inp), @XMM[5]
+	mov	%rsp, %rax		# pass key schedule
+	movdqu	0x60($inp), @XMM[6]
+	mov	%edx,%r10d		# pass rounds
+	movdqu	0x70($inp), @XMM[7]
+	movdqa	@XMM[15], 0x20(%rbp)	# put aside IV
+
+	call	_bsaes_decrypt8
+
+	pxor	0x20(%rbp), @XMM[0]	# ^= IV
+	movdqu	0x00($inp), @XMM[8]	# re-load input
+	movdqu	0x10($inp), @XMM[9]
+	pxor	@XMM[8], @XMM[1]
+	movdqu	0x20($inp), @XMM[10]
+	pxor	@XMM[9], @XMM[6]
+	movdqu	0x30($inp), @XMM[11]
+	pxor	@XMM[10], @XMM[4]
+	movdqu	0x40($inp), @XMM[12]
+	pxor	@XMM[11], @XMM[2]
+	movdqu	0x50($inp), @XMM[13]
+	pxor	@XMM[12], @XMM[7]
+	movdqu	0x60($inp), @XMM[14]
+	pxor	@XMM[13], @XMM[3]
+	movdqu	0x70($inp), @XMM[15]	# IV
+	pxor	@XMM[14], @XMM[5]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	lea	0x80($inp), $inp
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[3], 0x60($out)
+	movdqu	@XMM[5], 0x70($out)
+	lea	0x80($out), $out
+	sub	\$8,$len
+	jnc	.Lcbc_dec_loop
+
+	add	\$8,$len
+	jz	.Lcbc_dec_done
+
+	movdqu	0x00($inp), @XMM[0]	# load input
+	mov	%rsp, %rax		# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+	cmp	\$2,$len
+	jb	.Lcbc_dec_one
+	movdqu	0x10($inp), @XMM[1]
+	je	.Lcbc_dec_two
+	movdqu	0x20($inp), @XMM[2]
+	cmp	\$4,$len
+	jb	.Lcbc_dec_three
+	movdqu	0x30($inp), @XMM[3]
+	je	.Lcbc_dec_four
+	movdqu	0x40($inp), @XMM[4]
+	cmp	\$6,$len
+	jb	.Lcbc_dec_five
+	movdqu	0x50($inp), @XMM[5]
+	je	.Lcbc_dec_six
+	movdqu	0x60($inp), @XMM[6]
+	movdqa	@XMM[15], 0x20(%rbp)	# put aside IV
+	call	_bsaes_decrypt8
+	pxor	0x20(%rbp), @XMM[0]	# ^= IV
+	movdqu	0x00($inp), @XMM[8]	# re-load input
+	movdqu	0x10($inp), @XMM[9]
+	pxor	@XMM[8], @XMM[1]
+	movdqu	0x20($inp), @XMM[10]
+	pxor	@XMM[9], @XMM[6]
+	movdqu	0x30($inp), @XMM[11]
+	pxor	@XMM[10], @XMM[4]
+	movdqu	0x40($inp), @XMM[12]
+	pxor	@XMM[11], @XMM[2]
+	movdqu	0x50($inp), @XMM[13]
+	pxor	@XMM[12], @XMM[7]
+	movdqu	0x60($inp), @XMM[15]	# IV
+	pxor	@XMM[13], @XMM[3]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[3], 0x60($out)
+	jmp	.Lcbc_dec_done
+.align	16
+.Lcbc_dec_six:
+	movdqa	@XMM[15], 0x20(%rbp)	# put aside IV
+	call	_bsaes_decrypt8
+	pxor	0x20(%rbp), @XMM[0]	# ^= IV
+	movdqu	0x00($inp), @XMM[8]	# re-load input
+	movdqu	0x10($inp), @XMM[9]
+	pxor	@XMM[8], @XMM[1]
+	movdqu	0x20($inp), @XMM[10]
+	pxor	@XMM[9], @XMM[6]
+	movdqu	0x30($inp), @XMM[11]
+	pxor	@XMM[10], @XMM[4]
+	movdqu	0x40($inp), @XMM[12]
+	pxor	@XMM[11], @XMM[2]
+	movdqu	0x50($inp), @XMM[15]	# IV
+	pxor	@XMM[12], @XMM[7]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	jmp	.Lcbc_dec_done
+.align	16
+.Lcbc_dec_five:
+	movdqa	@XMM[15], 0x20(%rbp)	# put aside IV
+	call	_bsaes_decrypt8
+	pxor	0x20(%rbp), @XMM[0]	# ^= IV
+	movdqu	0x00($inp), @XMM[8]	# re-load input
+	movdqu	0x10($inp), @XMM[9]
+	pxor	@XMM[8], @XMM[1]
+	movdqu	0x20($inp), @XMM[10]
+	pxor	@XMM[9], @XMM[6]
+	movdqu	0x30($inp), @XMM[11]
+	pxor	@XMM[10], @XMM[4]
+	movdqu	0x40($inp), @XMM[15]	# IV
+	pxor	@XMM[11], @XMM[2]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	jmp	.Lcbc_dec_done
+.align	16
+.Lcbc_dec_four:
+	movdqa	@XMM[15], 0x20(%rbp)	# put aside IV
+	call	_bsaes_decrypt8
+	pxor	0x20(%rbp), @XMM[0]	# ^= IV
+	movdqu	0x00($inp), @XMM[8]	# re-load input
+	movdqu	0x10($inp), @XMM[9]
+	pxor	@XMM[8], @XMM[1]
+	movdqu	0x20($inp), @XMM[10]
+	pxor	@XMM[9], @XMM[6]
+	movdqu	0x30($inp), @XMM[15]	# IV
+	pxor	@XMM[10], @XMM[4]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	jmp	.Lcbc_dec_done
+.align	16
+.Lcbc_dec_three:
+	movdqa	@XMM[15], 0x20(%rbp)	# put aside IV
+	call	_bsaes_decrypt8
+	pxor	0x20(%rbp), @XMM[0]	# ^= IV
+	movdqu	0x00($inp), @XMM[8]	# re-load input
+	movdqu	0x10($inp), @XMM[9]
+	pxor	@XMM[8], @XMM[1]
+	movdqu	0x20($inp), @XMM[15]	# IV
+	pxor	@XMM[9], @XMM[6]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	jmp	.Lcbc_dec_done
+.align	16
+.Lcbc_dec_two:
+	movdqa	@XMM[15], 0x20(%rbp)	# put aside IV
+	call	_bsaes_decrypt8
+	pxor	0x20(%rbp), @XMM[0]	# ^= IV
+	movdqu	0x00($inp), @XMM[8]	# re-load input
+	movdqu	0x10($inp), @XMM[15]	# IV
+	pxor	@XMM[8], @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	jmp	.Lcbc_dec_done
+.align	16
+.Lcbc_dec_one:
+	lea	($inp), $arg1
+	lea	0x20(%rbp), $arg2	# buffer output
+	lea	($key), $arg3
+	call	asm_AES_decrypt		# doesn't touch %xmm
+	pxor	0x20(%rbp), @XMM[15]	# ^= IV
+	movdqu	@XMM[15], ($out)	# write output
+	movdqa	@XMM[0], @XMM[15]	# IV
+
+.Lcbc_dec_done:
+	movdqu	@XMM[15], (%rbx)	# return IV
+	lea	(%rsp), %rax
+	pxor	%xmm0, %xmm0
+.Lcbc_dec_bzero:			# wipe key schedule [if any]
+	movdqa	%xmm0, 0x00(%rax)
+	movdqa	%xmm0, 0x10(%rax)
+	lea	0x20(%rax), %rax
+	cmp	%rax, %rbp
+	ja	.Lcbc_dec_bzero
+
+	lea	(%rbp),%rsp		# restore %rsp
+___
+$code.=<<___ if ($win64);
+	movaps	0x40(%rbp), %xmm6
+	movaps	0x50(%rbp), %xmm7
+	movaps	0x60(%rbp), %xmm8
+	movaps	0x70(%rbp), %xmm9
+	movaps	0x80(%rbp), %xmm10
+	movaps	0x90(%rbp), %xmm11
+	movaps	0xa0(%rbp), %xmm12
+	movaps	0xb0(%rbp), %xmm13
+	movaps	0xc0(%rbp), %xmm14
+	movaps	0xd0(%rbp), %xmm15
+	lea	0xa0(%rbp), %rsp
+___
+$code.=<<___;
+	mov	0x48(%rsp), %r15
+	mov	0x50(%rsp), %r14
+	mov	0x58(%rsp), %r13
+	mov	0x60(%rsp), %r12
+	mov	0x68(%rsp), %rbx
+	mov	0x70(%rsp), %rax
+	lea	0x78(%rsp), %rsp
+	mov	%rax, %rbp
+.Lcbc_dec_epilogue:
+	ret
+.size	bsaes_cbc_encrypt,.-bsaes_cbc_encrypt
+
+.globl	bsaes_ctr32_encrypt_blocks
+.type	bsaes_ctr32_encrypt_blocks,\@abi-omnipotent
+.align	16
+bsaes_ctr32_encrypt_blocks:
+	mov	%rsp, %rax
+.Lctr_enc_prologue:
+	push	%rbp
+	push	%rbx
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	lea	-0x48(%rsp), %rsp
+___
+$code.=<<___ if ($win64);
+	mov	0xa0(%rsp),$arg5	# pull ivp
+	lea	-0xa0(%rsp), %rsp
+	movaps	%xmm6, 0x40(%rsp)
+	movaps	%xmm7, 0x50(%rsp)
+	movaps	%xmm8, 0x60(%rsp)
+	movaps	%xmm9, 0x70(%rsp)
+	movaps	%xmm10, 0x80(%rsp)
+	movaps	%xmm11, 0x90(%rsp)
+	movaps	%xmm12, 0xa0(%rsp)
+	movaps	%xmm13, 0xb0(%rsp)
+	movaps	%xmm14, 0xc0(%rsp)
+	movaps	%xmm15, 0xd0(%rsp)
+.Lctr_enc_body:
+___
+$code.=<<___;
+	mov	%rsp, %rbp		# backup %rsp
+	movdqu	($arg5), %xmm0		# load counter
+	mov	240($arg4), %eax	# rounds
+	mov	$arg1, $inp		# backup arguments
+	mov	$arg2, $out
+	mov	$arg3, $len
+	mov	$arg4, $key
+	movdqa	%xmm0, 0x20(%rbp)	# copy counter
+	cmp	\$8, $arg3
+	jb	.Lctr_enc_short
+
+	mov	%eax, %ebx		# rounds
+	shl	\$7, %rax		# 128 bytes per inner round key
+	sub	\$`128-32`, %rax	# size of bit-sliced key schedule
+	sub	%rax, %rsp
+
+	mov	%rsp, %rax		# pass key schedule
+	mov	$key, %rcx		# pass key
+	mov	%ebx, %r10d		# pass rounds
+	call	_bsaes_key_convert
+	pxor	%xmm6,%xmm7		# fix up last round key
+	movdqa	%xmm7,(%rax)		# save last round key
+
+	movdqa	(%rsp), @XMM[9]		# load round0 key
+	lea	.LADD1(%rip), %r11
+	movdqa	0x20(%rbp), @XMM[0]	# counter copy
+	movdqa	-0x20(%r11), @XMM[8]	# .LSWPUP
+	pshufb	@XMM[8], @XMM[9]	# byte swap upper part
+	pshufb	@XMM[8], @XMM[0]
+	movdqa	@XMM[9], (%rsp)		# save adjusted round0 key
+	jmp	.Lctr_enc_loop
+.align	16
+.Lctr_enc_loop:
+	movdqa	@XMM[0], 0x20(%rbp)	# save counter
+	movdqa	@XMM[0], @XMM[1]	# prepare 8 counter values
+	movdqa	@XMM[0], @XMM[2]
+	paddd	0x00(%r11), @XMM[1]	# .LADD1
+	movdqa	@XMM[0], @XMM[3]
+	paddd	0x10(%r11), @XMM[2]	# .LADD2
+	movdqa	@XMM[0], @XMM[4]
+	paddd	0x20(%r11), @XMM[3]	# .LADD3
+	movdqa	@XMM[0], @XMM[5]
+	paddd	0x30(%r11), @XMM[4]	# .LADD4
+	movdqa	@XMM[0], @XMM[6]
+	paddd	0x40(%r11), @XMM[5]	# .LADD5
+	movdqa	@XMM[0], @XMM[7]
+	paddd	0x50(%r11), @XMM[6]	# .LADD6
+	paddd	0x60(%r11), @XMM[7]	# .LADD7
+
+	# Borrow prologue from _bsaes_encrypt8 to use the opportunity
+	# to flip byte order in 32-bit counter
+	movdqa	(%rsp), @XMM[9]		# round 0 key
+	lea	0x10(%rsp), %rax	# pass key schedule
+	movdqa	-0x10(%r11), @XMM[8]	# .LSWPUPM0SR
+	pxor	@XMM[9], @XMM[0]	# xor with round0 key
+	pxor	@XMM[9], @XMM[1]
+	 pshufb	@XMM[8], @XMM[0]
+	pxor	@XMM[9], @XMM[2]
+	 pshufb	@XMM[8], @XMM[1]
+	pxor	@XMM[9], @XMM[3]
+	 pshufb	@XMM[8], @XMM[2]
+	pxor	@XMM[9], @XMM[4]
+	 pshufb	@XMM[8], @XMM[3]
+	pxor	@XMM[9], @XMM[5]
+	 pshufb	@XMM[8], @XMM[4]
+	pxor	@XMM[9], @XMM[6]
+	 pshufb	@XMM[8], @XMM[5]
+	pxor	@XMM[9], @XMM[7]
+	 pshufb	@XMM[8], @XMM[6]
+	lea	.LBS0(%rip), %r11	# constants table
+	 pshufb	@XMM[8], @XMM[7]
+	mov	%ebx,%r10d		# pass rounds
+
+	call	_bsaes_encrypt8_bitslice
+
+	sub	\$8,$len
+	jc	.Lctr_enc_loop_done
+
+	movdqu	0x00($inp), @XMM[8]	# load input
+	movdqu	0x10($inp), @XMM[9]
+	movdqu	0x20($inp), @XMM[10]
+	movdqu	0x30($inp), @XMM[11]
+	movdqu	0x40($inp), @XMM[12]
+	movdqu	0x50($inp), @XMM[13]
+	movdqu	0x60($inp), @XMM[14]
+	movdqu	0x70($inp), @XMM[15]
+	lea	0x80($inp),$inp
+	pxor	@XMM[0], @XMM[8]
+	movdqa	0x20(%rbp), @XMM[0]	# load counter
+	pxor	@XMM[9], @XMM[1]
+	movdqu	@XMM[8], 0x00($out)	# write output
+	pxor	@XMM[10], @XMM[4]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	@XMM[11], @XMM[6]
+	movdqu	@XMM[4], 0x20($out)
+	pxor	@XMM[12], @XMM[3]
+	movdqu	@XMM[6], 0x30($out)
+	pxor	@XMM[13], @XMM[7]
+	movdqu	@XMM[3], 0x40($out)
+	pxor	@XMM[14], @XMM[2]
+	movdqu	@XMM[7], 0x50($out)
+	pxor	@XMM[15], @XMM[5]
+	movdqu	@XMM[2], 0x60($out)
+	lea	.LADD1(%rip), %r11
+	movdqu	@XMM[5], 0x70($out)
+	lea	0x80($out), $out
+	paddd	0x70(%r11), @XMM[0]	# .LADD8
+	jnz	.Lctr_enc_loop
+
+	jmp	.Lctr_enc_done
+.align	16
+.Lctr_enc_loop_done:
+	add	\$8, $len
+	movdqu	0x00($inp), @XMM[8]	# load input
+	pxor	@XMM[8], @XMM[0]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	cmp	\$2,$len
+	jb	.Lctr_enc_done
+	movdqu	0x10($inp), @XMM[9]
+	pxor	@XMM[9], @XMM[1]
+	movdqu	@XMM[1], 0x10($out)
+	je	.Lctr_enc_done
+	movdqu	0x20($inp), @XMM[10]
+	pxor	@XMM[10], @XMM[4]
+	movdqu	@XMM[4], 0x20($out)
+	cmp	\$4,$len
+	jb	.Lctr_enc_done
+	movdqu	0x30($inp), @XMM[11]
+	pxor	@XMM[11], @XMM[6]
+	movdqu	@XMM[6], 0x30($out)
+	je	.Lctr_enc_done
+	movdqu	0x40($inp), @XMM[12]
+	pxor	@XMM[12], @XMM[3]
+	movdqu	@XMM[3], 0x40($out)
+	cmp	\$6,$len
+	jb	.Lctr_enc_done
+	movdqu	0x50($inp), @XMM[13]
+	pxor	@XMM[13], @XMM[7]
+	movdqu	@XMM[7], 0x50($out)
+	je	.Lctr_enc_done
+	movdqu	0x60($inp), @XMM[14]
+	pxor	@XMM[14], @XMM[2]
+	movdqu	@XMM[2], 0x60($out)
+	jmp	.Lctr_enc_done
+
+.align	16
+.Lctr_enc_short:
+	lea	0x20(%rbp), $arg1
+	lea	0x30(%rbp), $arg2
+	lea	($key), $arg3
+	call	asm_AES_encrypt
+	movdqu	($inp), @XMM[1]
+	lea	16($inp), $inp
+	mov	0x2c(%rbp), %eax	# load 32-bit counter
+	bswap	%eax
+	pxor	0x30(%rbp), @XMM[1]
+	inc	%eax			# increment
+	movdqu	@XMM[1], ($out)
+	bswap	%eax
+	lea	16($out), $out
+	mov	%eax, 0x2c(%rsp)	# save 32-bit counter
+	dec	$len
+	jnz	.Lctr_enc_short
+
+.Lctr_enc_done:
+	lea	(%rsp), %rax
+	pxor	%xmm0, %xmm0
+.Lctr_enc_bzero:			# wipe key schedule [if any]
+	movdqa	%xmm0, 0x00(%rax)
+	movdqa	%xmm0, 0x10(%rax)
+	lea	0x20(%rax), %rax
+	cmp	%rax, %rbp
+	ja	.Lctr_enc_bzero
+
+	lea	(%rbp),%rsp		# restore %rsp
+___
+$code.=<<___ if ($win64);
+	movaps	0x40(%rbp), %xmm6
+	movaps	0x50(%rbp), %xmm7
+	movaps	0x60(%rbp), %xmm8
+	movaps	0x70(%rbp), %xmm9
+	movaps	0x80(%rbp), %xmm10
+	movaps	0x90(%rbp), %xmm11
+	movaps	0xa0(%rbp), %xmm12
+	movaps	0xb0(%rbp), %xmm13
+	movaps	0xc0(%rbp), %xmm14
+	movaps	0xd0(%rbp), %xmm15
+	lea	0xa0(%rbp), %rsp
+___
+$code.=<<___;
+	mov	0x48(%rsp), %r15
+	mov	0x50(%rsp), %r14
+	mov	0x58(%rsp), %r13
+	mov	0x60(%rsp), %r12
+	mov	0x68(%rsp), %rbx
+	mov	0x70(%rsp), %rax
+	lea	0x78(%rsp), %rsp
+	mov	%rax, %rbp
+.Lctr_enc_epilogue:
+	ret
+.size	bsaes_ctr32_encrypt_blocks,.-bsaes_ctr32_encrypt_blocks
+___
+######################################################################
+# void bsaes_xts_[en|de]crypt(const char *inp,char *out,size_t len,
+#	const AES_KEY *key1, const AES_KEY *key2,
+#	const unsigned char iv[16]);
+#
+my ($twmask,$twres,$twtmp)=@XMM[13..15];
+$code.=<<___;
+.globl	bsaes_xts_encrypt
+.type	bsaes_xts_encrypt,\@abi-omnipotent
+.align	16
+bsaes_xts_encrypt:
+	mov	%rsp, %rax
+.Lxts_enc_prologue:
+	push	%rbp
+	push	%rbx
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	lea	-0x48(%rsp), %rsp
+___
+$code.=<<___ if ($win64);
+	mov	0xa0(%rsp),$arg5	# pull key2
+	mov	0xa8(%rsp),$arg6	# pull ivp
+	lea	-0xa0(%rsp), %rsp
+	movaps	%xmm6, 0x40(%rsp)
+	movaps	%xmm7, 0x50(%rsp)
+	movaps	%xmm8, 0x60(%rsp)
+	movaps	%xmm9, 0x70(%rsp)
+	movaps	%xmm10, 0x80(%rsp)
+	movaps	%xmm11, 0x90(%rsp)
+	movaps	%xmm12, 0xa0(%rsp)
+	movaps	%xmm13, 0xb0(%rsp)
+	movaps	%xmm14, 0xc0(%rsp)
+	movaps	%xmm15, 0xd0(%rsp)
+.Lxts_enc_body:
+___
+$code.=<<___;
+	mov	%rsp, %rbp		# backup %rsp
+	mov	$arg1, $inp		# backup arguments
+	mov	$arg2, $out
+	mov	$arg3, $len
+	mov	$arg4, $key
+
+	lea	($arg6), $arg1
+	lea	0x20(%rbp), $arg2
+	lea	($arg5), $arg3
+	call	asm_AES_encrypt		# generate initial tweak
+
+	mov	240($key), %eax		# rounds
+	mov	$len, %rbx		# backup $len
+
+	mov	%eax, %edx		# rounds
+	shl	\$7, %rax		# 128 bytes per inner round key
+	sub	\$`128-32`, %rax	# size of bit-sliced key schedule
+	sub	%rax, %rsp
+
+	mov	%rsp, %rax		# pass key schedule
+	mov	$key, %rcx		# pass key
+	mov	%edx, %r10d		# pass rounds
+	call	_bsaes_key_convert
+	pxor	%xmm6, %xmm7		# fix up last round key
+	movdqa	%xmm7, (%rax)		# save last round key
+
+	and	\$-16, $len
+	sub	\$0x80, %rsp		# place for tweak[8]
+	movdqa	0x20(%rbp), @XMM[7]	# initial tweak
+
+	pxor	$twtmp, $twtmp
+	movdqa	.Lxts_magic(%rip), $twmask
+	pcmpgtd	@XMM[7], $twtmp		# broadcast upper bits
+
+	sub	\$0x80, $len
+	jc	.Lxts_enc_short
+	jmp	.Lxts_enc_loop
+
+.align	16
+.Lxts_enc_loop:
+___
+    for ($i=0;$i<7;$i++) {
+    $code.=<<___;
+	pshufd	\$0x13, $twtmp, $twres
+	pxor	$twtmp, $twtmp
+	movdqa	@XMM[7], @XMM[$i]
+	movdqa	@XMM[7], `0x10*$i`(%rsp)# save tweak[$i]
+	paddq	@XMM[7], @XMM[7]	# psllq	1,$tweak
+	pand	$twmask, $twres		# isolate carry and residue
+	pcmpgtd	@XMM[7], $twtmp		# broadcast upper bits
+	pxor	$twres, @XMM[7]
+___
+    $code.=<<___ if ($i>=1);
+	movdqu	`0x10*($i-1)`($inp), @XMM[8+$i-1]
+___
+    $code.=<<___ if ($i>=2);
+	pxor	@XMM[8+$i-2], @XMM[$i-2]# input[] ^ tweak[]
+___
+    }
+$code.=<<___;
+	movdqu	0x60($inp), @XMM[8+6]
+	pxor	@XMM[8+5], @XMM[5]
+	movdqu	0x70($inp), @XMM[8+7]
+	lea	0x80($inp), $inp
+	movdqa	@XMM[7], 0x70(%rsp)
+	pxor	@XMM[8+6], @XMM[6]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	pxor	@XMM[8+7], @XMM[7]
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_encrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[4]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[6]
+	movdqu	@XMM[4], 0x20($out)
+	pxor	0x40(%rsp), @XMM[3]
+	movdqu	@XMM[6], 0x30($out)
+	pxor	0x50(%rsp), @XMM[7]
+	movdqu	@XMM[3], 0x40($out)
+	pxor	0x60(%rsp), @XMM[2]
+	movdqu	@XMM[7], 0x50($out)
+	pxor	0x70(%rsp), @XMM[5]
+	movdqu	@XMM[2], 0x60($out)
+	movdqu	@XMM[5], 0x70($out)
+	lea	0x80($out), $out
+
+	movdqa	0x70(%rsp), @XMM[7]	# prepare next iteration tweak
+	pxor	$twtmp, $twtmp
+	movdqa	.Lxts_magic(%rip), $twmask
+	pcmpgtd	@XMM[7], $twtmp
+	pshufd	\$0x13, $twtmp, $twres
+	pxor	$twtmp, $twtmp
+	paddq	@XMM[7], @XMM[7]	# psllq	1,$tweak
+	pand	$twmask, $twres		# isolate carry and residue
+	pcmpgtd	@XMM[7], $twtmp		# broadcast upper bits
+	pxor	$twres, @XMM[7]
+
+	sub	\$0x80,$len
+	jnc	.Lxts_enc_loop
+
+.Lxts_enc_short:
+	add	\$0x80, $len
+	jz	.Lxts_enc_done
+___
+    for ($i=0;$i<7;$i++) {
+    $code.=<<___;
+	pshufd	\$0x13, $twtmp, $twres
+	pxor	$twtmp, $twtmp
+	movdqa	@XMM[7], @XMM[$i]
+	movdqa	@XMM[7], `0x10*$i`(%rsp)# save tweak[$i]
+	paddq	@XMM[7], @XMM[7]	# psllq	1,$tweak
+	pand	$twmask, $twres		# isolate carry and residue
+	pcmpgtd	@XMM[7], $twtmp		# broadcast upper bits
+	pxor	$twres, @XMM[7]
+___
+    $code.=<<___ if ($i>=1);
+	movdqu	`0x10*($i-1)`($inp), @XMM[8+$i-1]
+	cmp	\$`0x10*$i`,$len
+	je	.Lxts_enc_$i
+___
+    $code.=<<___ if ($i>=2);
+	pxor	@XMM[8+$i-2], @XMM[$i-2]# input[] ^ tweak[]
+___
+    }
+$code.=<<___;
+	movdqu	0x60($inp), @XMM[8+6]
+	pxor	@XMM[8+5], @XMM[5]
+	movdqa	@XMM[7], 0x70(%rsp)
+	lea	0x70($inp), $inp
+	pxor	@XMM[8+6], @XMM[6]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_encrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[4]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[6]
+	movdqu	@XMM[4], 0x20($out)
+	pxor	0x40(%rsp), @XMM[3]
+	movdqu	@XMM[6], 0x30($out)
+	pxor	0x50(%rsp), @XMM[7]
+	movdqu	@XMM[3], 0x40($out)
+	pxor	0x60(%rsp), @XMM[2]
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[2], 0x60($out)
+	lea	0x70($out), $out
+
+	movdqa	0x70(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_enc_done
+.align	16
+.Lxts_enc_6:
+	pxor	@XMM[8+4], @XMM[4]
+	lea	0x60($inp), $inp
+	pxor	@XMM[8+5], @XMM[5]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_encrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[4]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[6]
+	movdqu	@XMM[4], 0x20($out)
+	pxor	0x40(%rsp), @XMM[3]
+	movdqu	@XMM[6], 0x30($out)
+	pxor	0x50(%rsp), @XMM[7]
+	movdqu	@XMM[3], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	lea	0x60($out), $out
+
+	movdqa	0x60(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_enc_done
+.align	16
+.Lxts_enc_5:
+	pxor	@XMM[8+3], @XMM[3]
+	lea	0x50($inp), $inp
+	pxor	@XMM[8+4], @XMM[4]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_encrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[4]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[6]
+	movdqu	@XMM[4], 0x20($out)
+	pxor	0x40(%rsp), @XMM[3]
+	movdqu	@XMM[6], 0x30($out)
+	movdqu	@XMM[3], 0x40($out)
+	lea	0x50($out), $out
+
+	movdqa	0x50(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_enc_done
+.align	16
+.Lxts_enc_4:
+	pxor	@XMM[8+2], @XMM[2]
+	lea	0x40($inp), $inp
+	pxor	@XMM[8+3], @XMM[3]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_encrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[4]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[6]
+	movdqu	@XMM[4], 0x20($out)
+	movdqu	@XMM[6], 0x30($out)
+	lea	0x40($out), $out
+
+	movdqa	0x40(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_enc_done
+.align	16
+.Lxts_enc_3:
+	pxor	@XMM[8+1], @XMM[1]
+	lea	0x30($inp), $inp
+	pxor	@XMM[8+2], @XMM[2]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_encrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[4]
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[4], 0x20($out)
+	lea	0x30($out), $out
+
+	movdqa	0x30(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_enc_done
+.align	16
+.Lxts_enc_2:
+	pxor	@XMM[8+0], @XMM[0]
+	lea	0x20($inp), $inp
+	pxor	@XMM[8+1], @XMM[1]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_encrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	lea	0x20($out), $out
+
+	movdqa	0x20(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_enc_done
+.align	16
+.Lxts_enc_1:
+	pxor	@XMM[0], @XMM[8]
+	lea	0x10($inp), $inp
+	movdqa	@XMM[8], 0x20(%rbp)
+	lea	0x20(%rbp), $arg1
+	lea	0x20(%rbp), $arg2
+	lea	($key), $arg3
+	call	asm_AES_encrypt		# doesn't touch %xmm
+	pxor	0x20(%rbp), @XMM[0]	# ^= tweak[]
+	#pxor	@XMM[8], @XMM[0]
+	#lea	0x80(%rsp), %rax	# pass key schedule
+	#mov	%edx, %r10d		# pass rounds
+	#call	_bsaes_encrypt8
+	#pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	lea	0x10($out), $out
+
+	movdqa	0x10(%rsp), @XMM[7]	# next iteration tweak
+
+.Lxts_enc_done:
+	and	\$15, %ebx
+	jz	.Lxts_enc_ret
+	mov	$out, %rdx
+
+.Lxts_enc_steal:
+	movzb	($inp), %eax
+	movzb	-16(%rdx), %ecx
+	lea	1($inp), $inp
+	mov	%al, -16(%rdx)
+	mov	%cl, 0(%rdx)
+	lea	1(%rdx), %rdx
+	sub	\$1,%ebx
+	jnz	.Lxts_enc_steal
+
+	movdqu	-16($out), @XMM[0]
+	lea	0x20(%rbp), $arg1
+	pxor	@XMM[7], @XMM[0]
+	lea	0x20(%rbp), $arg2
+	movdqa	@XMM[0], 0x20(%rbp)
+	lea	($key), $arg3
+	call	asm_AES_encrypt		# doesn't touch %xmm
+	pxor	0x20(%rbp), @XMM[7]
+	movdqu	@XMM[7], -16($out)
+
+.Lxts_enc_ret:
+	lea	(%rsp), %rax
+	pxor	%xmm0, %xmm0
+.Lxts_enc_bzero:			# wipe key schedule [if any]
+	movdqa	%xmm0, 0x00(%rax)
+	movdqa	%xmm0, 0x10(%rax)
+	lea	0x20(%rax), %rax
+	cmp	%rax, %rbp
+	ja	.Lxts_enc_bzero
+
+	lea	(%rbp),%rsp		# restore %rsp
+___
+$code.=<<___ if ($win64);
+	movaps	0x40(%rbp), %xmm6
+	movaps	0x50(%rbp), %xmm7
+	movaps	0x60(%rbp), %xmm8
+	movaps	0x70(%rbp), %xmm9
+	movaps	0x80(%rbp), %xmm10
+	movaps	0x90(%rbp), %xmm11
+	movaps	0xa0(%rbp), %xmm12
+	movaps	0xb0(%rbp), %xmm13
+	movaps	0xc0(%rbp), %xmm14
+	movaps	0xd0(%rbp), %xmm15
+	lea	0xa0(%rbp), %rsp
+___
+$code.=<<___;
+	mov	0x48(%rsp), %r15
+	mov	0x50(%rsp), %r14
+	mov	0x58(%rsp), %r13
+	mov	0x60(%rsp), %r12
+	mov	0x68(%rsp), %rbx
+	mov	0x70(%rsp), %rax
+	lea	0x78(%rsp), %rsp
+	mov	%rax, %rbp
+.Lxts_enc_epilogue:
+	ret
+.size	bsaes_xts_encrypt,.-bsaes_xts_encrypt
+
+.globl	bsaes_xts_decrypt
+.type	bsaes_xts_decrypt,\@abi-omnipotent
+.align	16
+bsaes_xts_decrypt:
+	mov	%rsp, %rax
+.Lxts_dec_prologue:
+	push	%rbp
+	push	%rbx
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	lea	-0x48(%rsp), %rsp
+___
+$code.=<<___ if ($win64);
+	mov	0xa0(%rsp),$arg5	# pull key2
+	mov	0xa8(%rsp),$arg6	# pull ivp
+	lea	-0xa0(%rsp), %rsp
+	movaps	%xmm6, 0x40(%rsp)
+	movaps	%xmm7, 0x50(%rsp)
+	movaps	%xmm8, 0x60(%rsp)
+	movaps	%xmm9, 0x70(%rsp)
+	movaps	%xmm10, 0x80(%rsp)
+	movaps	%xmm11, 0x90(%rsp)
+	movaps	%xmm12, 0xa0(%rsp)
+	movaps	%xmm13, 0xb0(%rsp)
+	movaps	%xmm14, 0xc0(%rsp)
+	movaps	%xmm15, 0xd0(%rsp)
+.Lxts_dec_body:
+___
+$code.=<<___;
+	mov	%rsp, %rbp		# backup %rsp
+	mov	$arg1, $inp		# backup arguments
+	mov	$arg2, $out
+	mov	$arg3, $len
+	mov	$arg4, $key
+
+	lea	($arg6), $arg1
+	lea	0x20(%rbp), $arg2
+	lea	($arg5), $arg3
+	call	asm_AES_encrypt		# generate initial tweak
+
+	mov	240($key), %eax		# rounds
+	mov	$len, %rbx		# backup $len
+
+	mov	%eax, %edx		# rounds
+	shl	\$7, %rax		# 128 bytes per inner round key
+	sub	\$`128-32`, %rax	# size of bit-sliced key schedule
+	sub	%rax, %rsp
+
+	mov	%rsp, %rax		# pass key schedule
+	mov	$key, %rcx		# pass key
+	mov	%edx, %r10d		# pass rounds
+	call	_bsaes_key_convert
+	pxor	(%rsp), %xmm7		# fix up round 0 key
+	movdqa	%xmm6, (%rax)		# save last round key
+	movdqa	%xmm7, (%rsp)
+
+	xor	%eax, %eax		# if ($len%16) len-=16;
+	and	\$-16, $len
+	test	\$15, %ebx
+	setnz	%al
+	shl	\$4, %rax
+	sub	%rax, $len
+
+	sub	\$0x80, %rsp		# place for tweak[8]
+	movdqa	0x20(%rbp), @XMM[7]	# initial tweak
+
+	pxor	$twtmp, $twtmp
+	movdqa	.Lxts_magic(%rip), $twmask
+	pcmpgtd	@XMM[7], $twtmp		# broadcast upper bits
+
+	sub	\$0x80, $len
+	jc	.Lxts_dec_short
+	jmp	.Lxts_dec_loop
+
+.align	16
+.Lxts_dec_loop:
+___
+    for ($i=0;$i<7;$i++) {
+    $code.=<<___;
+	pshufd	\$0x13, $twtmp, $twres
+	pxor	$twtmp, $twtmp
+	movdqa	@XMM[7], @XMM[$i]
+	movdqa	@XMM[7], `0x10*$i`(%rsp)# save tweak[$i]
+	paddq	@XMM[7], @XMM[7]	# psllq	1,$tweak
+	pand	$twmask, $twres		# isolate carry and residue
+	pcmpgtd	@XMM[7], $twtmp		# broadcast upper bits
+	pxor	$twres, @XMM[7]
+___
+    $code.=<<___ if ($i>=1);
+	movdqu	`0x10*($i-1)`($inp), @XMM[8+$i-1]
+___
+    $code.=<<___ if ($i>=2);
+	pxor	@XMM[8+$i-2], @XMM[$i-2]# input[] ^ tweak[]
+___
+    }
+$code.=<<___;
+	movdqu	0x60($inp), @XMM[8+6]
+	pxor	@XMM[8+5], @XMM[5]
+	movdqu	0x70($inp), @XMM[8+7]
+	lea	0x80($inp), $inp
+	movdqa	@XMM[7], 0x70(%rsp)
+	pxor	@XMM[8+6], @XMM[6]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	pxor	@XMM[8+7], @XMM[7]
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_decrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[6]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[4]
+	movdqu	@XMM[6], 0x20($out)
+	pxor	0x40(%rsp), @XMM[2]
+	movdqu	@XMM[4], 0x30($out)
+	pxor	0x50(%rsp), @XMM[7]
+	movdqu	@XMM[2], 0x40($out)
+	pxor	0x60(%rsp), @XMM[3]
+	movdqu	@XMM[7], 0x50($out)
+	pxor	0x70(%rsp), @XMM[5]
+	movdqu	@XMM[3], 0x60($out)
+	movdqu	@XMM[5], 0x70($out)
+	lea	0x80($out), $out
+
+	movdqa	0x70(%rsp), @XMM[7]	# prepare next iteration tweak
+	pxor	$twtmp, $twtmp
+	movdqa	.Lxts_magic(%rip), $twmask
+	pcmpgtd	@XMM[7], $twtmp
+	pshufd	\$0x13, $twtmp, $twres
+	pxor	$twtmp, $twtmp
+	paddq	@XMM[7], @XMM[7]	# psllq	1,$tweak
+	pand	$twmask, $twres		# isolate carry and residue
+	pcmpgtd	@XMM[7], $twtmp		# broadcast upper bits
+	pxor	$twres, @XMM[7]
+
+	sub	\$0x80,$len
+	jnc	.Lxts_dec_loop
+
+.Lxts_dec_short:
+	add	\$0x80, $len
+	jz	.Lxts_dec_done
+___
+    for ($i=0;$i<7;$i++) {
+    $code.=<<___;
+	pshufd	\$0x13, $twtmp, $twres
+	pxor	$twtmp, $twtmp
+	movdqa	@XMM[7], @XMM[$i]
+	movdqa	@XMM[7], `0x10*$i`(%rsp)# save tweak[$i]
+	paddq	@XMM[7], @XMM[7]	# psllq	1,$tweak
+	pand	$twmask, $twres		# isolate carry and residue
+	pcmpgtd	@XMM[7], $twtmp		# broadcast upper bits
+	pxor	$twres, @XMM[7]
+___
+    $code.=<<___ if ($i>=1);
+	movdqu	`0x10*($i-1)`($inp), @XMM[8+$i-1]
+	cmp	\$`0x10*$i`,$len
+	je	.Lxts_dec_$i
+___
+    $code.=<<___ if ($i>=2);
+	pxor	@XMM[8+$i-2], @XMM[$i-2]# input[] ^ tweak[]
+___
+    }
+$code.=<<___;
+	movdqu	0x60($inp), @XMM[8+6]
+	pxor	@XMM[8+5], @XMM[5]
+	movdqa	@XMM[7], 0x70(%rsp)
+	lea	0x70($inp), $inp
+	pxor	@XMM[8+6], @XMM[6]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_decrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[6]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[4]
+	movdqu	@XMM[6], 0x20($out)
+	pxor	0x40(%rsp), @XMM[2]
+	movdqu	@XMM[4], 0x30($out)
+	pxor	0x50(%rsp), @XMM[7]
+	movdqu	@XMM[2], 0x40($out)
+	pxor	0x60(%rsp), @XMM[3]
+	movdqu	@XMM[7], 0x50($out)
+	movdqu	@XMM[3], 0x60($out)
+	lea	0x70($out), $out
+
+	movdqa	0x70(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_dec_done
+.align	16
+.Lxts_dec_6:
+	pxor	@XMM[8+4], @XMM[4]
+	lea	0x60($inp), $inp
+	pxor	@XMM[8+5], @XMM[5]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_decrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[6]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[4]
+	movdqu	@XMM[6], 0x20($out)
+	pxor	0x40(%rsp), @XMM[2]
+	movdqu	@XMM[4], 0x30($out)
+	pxor	0x50(%rsp), @XMM[7]
+	movdqu	@XMM[2], 0x40($out)
+	movdqu	@XMM[7], 0x50($out)
+	lea	0x60($out), $out
+
+	movdqa	0x60(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_dec_done
+.align	16
+.Lxts_dec_5:
+	pxor	@XMM[8+3], @XMM[3]
+	lea	0x50($inp), $inp
+	pxor	@XMM[8+4], @XMM[4]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_decrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[6]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[4]
+	movdqu	@XMM[6], 0x20($out)
+	pxor	0x40(%rsp), @XMM[2]
+	movdqu	@XMM[4], 0x30($out)
+	movdqu	@XMM[2], 0x40($out)
+	lea	0x50($out), $out
+
+	movdqa	0x50(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_dec_done
+.align	16
+.Lxts_dec_4:
+	pxor	@XMM[8+2], @XMM[2]
+	lea	0x40($inp), $inp
+	pxor	@XMM[8+3], @XMM[3]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_decrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[6]
+	movdqu	@XMM[1], 0x10($out)
+	pxor	0x30(%rsp), @XMM[4]
+	movdqu	@XMM[6], 0x20($out)
+	movdqu	@XMM[4], 0x30($out)
+	lea	0x40($out), $out
+
+	movdqa	0x40(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_dec_done
+.align	16
+.Lxts_dec_3:
+	pxor	@XMM[8+1], @XMM[1]
+	lea	0x30($inp), $inp
+	pxor	@XMM[8+2], @XMM[2]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_decrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	pxor	0x20(%rsp), @XMM[6]
+	movdqu	@XMM[1], 0x10($out)
+	movdqu	@XMM[6], 0x20($out)
+	lea	0x30($out), $out
+
+	movdqa	0x30(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_dec_done
+.align	16
+.Lxts_dec_2:
+	pxor	@XMM[8+0], @XMM[0]
+	lea	0x20($inp), $inp
+	pxor	@XMM[8+1], @XMM[1]
+	lea	0x80(%rsp), %rax	# pass key schedule
+	mov	%edx, %r10d		# pass rounds
+
+	call	_bsaes_decrypt8
+
+	pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	pxor	0x10(%rsp), @XMM[1]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	movdqu	@XMM[1], 0x10($out)
+	lea	0x20($out), $out
+
+	movdqa	0x20(%rsp), @XMM[7]	# next iteration tweak
+	jmp	.Lxts_dec_done
+.align	16
+.Lxts_dec_1:
+	pxor	@XMM[0], @XMM[8]
+	lea	0x10($inp), $inp
+	movdqa	@XMM[8], 0x20(%rbp)
+	lea	0x20(%rbp), $arg1
+	lea	0x20(%rbp), $arg2
+	lea	($key), $arg3
+	call	asm_AES_decrypt		# doesn't touch %xmm
+	pxor	0x20(%rbp), @XMM[0]	# ^= tweak[]
+	#pxor	@XMM[8], @XMM[0]
+	#lea	0x80(%rsp), %rax	# pass key schedule
+	#mov	%edx, %r10d		# pass rounds
+	#call	_bsaes_decrypt8
+	#pxor	0x00(%rsp), @XMM[0]	# ^= tweak[]
+	movdqu	@XMM[0], 0x00($out)	# write output
+	lea	0x10($out), $out
+
+	movdqa	0x10(%rsp), @XMM[7]	# next iteration tweak
+
+.Lxts_dec_done:
+	and	\$15, %ebx
+	jz	.Lxts_dec_ret
+
+	pxor	$twtmp, $twtmp
+	movdqa	.Lxts_magic(%rip), $twmask
+	pcmpgtd	@XMM[7], $twtmp
+	pshufd	\$0x13, $twtmp, $twres
+	movdqa	@XMM[7], @XMM[6]
+	paddq	@XMM[7], @XMM[7]	# psllq 1,$tweak
+	pand	$twmask, $twres		# isolate carry and residue
+	movdqu	($inp), @XMM[0]
+	pxor	$twres, @XMM[7]
+
+	lea	0x20(%rbp), $arg1
+	pxor	@XMM[7], @XMM[0]
+	lea	0x20(%rbp), $arg2
+	movdqa	@XMM[0], 0x20(%rbp)
+	lea	($key), $arg3
+	call	asm_AES_decrypt		# doesn't touch %xmm
+	pxor	0x20(%rbp), @XMM[7]
+	mov	$out, %rdx
+	movdqu	@XMM[7], ($out)
+
+.Lxts_dec_steal:
+	movzb	16($inp), %eax
+	movzb	(%rdx), %ecx
+	lea	1($inp), $inp
+	mov	%al, (%rdx)
+	mov	%cl, 16(%rdx)
+	lea	1(%rdx), %rdx
+	sub	\$1,%ebx
+	jnz	.Lxts_dec_steal
+
+	movdqu	($out), @XMM[0]
+	lea	0x20(%rbp), $arg1
+	pxor	@XMM[6], @XMM[0]
+	lea	0x20(%rbp), $arg2
+	movdqa	@XMM[0], 0x20(%rbp)
+	lea	($key), $arg3
+	call	asm_AES_decrypt		# doesn't touch %xmm
+	pxor	0x20(%rbp), @XMM[6]
+	movdqu	@XMM[6], ($out)
+
+.Lxts_dec_ret:
+	lea	(%rsp), %rax
+	pxor	%xmm0, %xmm0
+.Lxts_dec_bzero:			# wipe key schedule [if any]
+	movdqa	%xmm0, 0x00(%rax)
+	movdqa	%xmm0, 0x10(%rax)
+	lea	0x20(%rax), %rax
+	cmp	%rax, %rbp
+	ja	.Lxts_dec_bzero
+
+	lea	(%rbp),%rsp		# restore %rsp
+___
+$code.=<<___ if ($win64);
+	movaps	0x40(%rbp), %xmm6
+	movaps	0x50(%rbp), %xmm7
+	movaps	0x60(%rbp), %xmm8
+	movaps	0x70(%rbp), %xmm9
+	movaps	0x80(%rbp), %xmm10
+	movaps	0x90(%rbp), %xmm11
+	movaps	0xa0(%rbp), %xmm12
+	movaps	0xb0(%rbp), %xmm13
+	movaps	0xc0(%rbp), %xmm14
+	movaps	0xd0(%rbp), %xmm15
+	lea	0xa0(%rbp), %rsp
+___
+$code.=<<___;
+	mov	0x48(%rsp), %r15
+	mov	0x50(%rsp), %r14
+	mov	0x58(%rsp), %r13
+	mov	0x60(%rsp), %r12
+	mov	0x68(%rsp), %rbx
+	mov	0x70(%rsp), %rax
+	lea	0x78(%rsp), %rsp
+	mov	%rax, %rbp
+.Lxts_dec_epilogue:
+	ret
+.size	bsaes_xts_decrypt,.-bsaes_xts_decrypt
+___
+}
+$code.=<<___;
+.type	_bsaes_const,\@object
+.align	64
+_bsaes_const:
+.LM0ISR:	# InvShiftRows constants
+	.quad	0x0a0e0206070b0f03, 0x0004080c0d010509
+.LISRM0:
+	.quad	0x01040b0e0205080f, 0x0306090c00070a0d
+.LISR:
+	.quad	0x0504070602010003, 0x0f0e0d0c080b0a09
+.LBS0:		# bit-slice constants
+	.quad	0x5555555555555555, 0x5555555555555555
+.LBS1:
+	.quad	0x3333333333333333, 0x3333333333333333
+.LBS2:
+	.quad	0x0f0f0f0f0f0f0f0f, 0x0f0f0f0f0f0f0f0f
+.LSR:		# shiftrows constants
+	.quad	0x0504070600030201, 0x0f0e0d0c0a09080b
+.LSRM0:
+	.quad	0x0304090e00050a0f, 0x01060b0c0207080d
+.LM0SR:
+	.quad	0x0a0e02060f03070b, 0x0004080c05090d01
+.LSWPUP:	# byte-swap upper dword
+	.quad	0x0706050403020100, 0x0c0d0e0f0b0a0908
+.LSWPUPM0SR:
+	.quad	0x0a0d02060c03070b, 0x0004080f05090e01
+.LADD1:		# counter increment constants
+	.quad	0x0000000000000000, 0x0000000100000000
+.LADD2:
+	.quad	0x0000000000000000, 0x0000000200000000
+.LADD3:
+	.quad	0x0000000000000000, 0x0000000300000000
+.LADD4:
+	.quad	0x0000000000000000, 0x0000000400000000
+.LADD5:
+	.quad	0x0000000000000000, 0x0000000500000000
+.LADD6:
+	.quad	0x0000000000000000, 0x0000000600000000
+.LADD7:
+	.quad	0x0000000000000000, 0x0000000700000000
+.LADD8:
+	.quad	0x0000000000000000, 0x0000000800000000
+.Lxts_magic:
+	.long	0x87,0,1,0
+.Lmasks:
+	.quad	0x0101010101010101, 0x0101010101010101
+	.quad	0x0202020202020202, 0x0202020202020202
+	.quad	0x0404040404040404, 0x0404040404040404
+	.quad	0x0808080808080808, 0x0808080808080808
+.LM0:
+	.quad	0x02060a0e03070b0f, 0x0004080c0105090d
+.L63:
+	.quad	0x6363636363636363, 0x6363636363636363
+.asciz	"Bit-sliced AES for x86_64/SSSE3, Emilia Käsper, Peter Schwabe, Andy Polyakov"
+.align	64
+.size	_bsaes_const,.-_bsaes_const
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lin_prologue
+
+	mov	160($context),%rax	# pull context->Rbp
+
+	lea	0x40(%rax),%rsi		# %xmm save area
+	lea	512($context),%rdi	# &context.Xmm6
+	mov	\$20,%ecx		# 10*sizeof(%xmm0)/sizeof(%rax)
+	.long	0xa548f3fc		# cld; rep movsq
+	lea	0xa0(%rax),%rax		# adjust stack pointer
+
+	mov	0x70(%rax),%rbp
+	mov	0x68(%rax),%rbx
+	mov	0x60(%rax),%r12
+	mov	0x58(%rax),%r13
+	mov	0x50(%rax),%r14
+	mov	0x48(%rax),%r15
+	lea	0x78(%rax),%rax		# adjust stack pointer
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_prologue:
+	mov	%rax,152($context)	# restore context->Rsp
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$`1232/8`,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	se_handler,.-se_handler
+
+.section	.pdata
+.align	4
+___
+$code.=<<___ if ($ecb);
+	.rva	.Lecb_enc_prologue
+	.rva	.Lecb_enc_epilogue
+	.rva	.Lecb_enc_info
+
+	.rva	.Lecb_dec_prologue
+	.rva	.Lecb_dec_epilogue
+	.rva	.Lecb_dec_info
+___
+$code.=<<___;
+	.rva	.Lcbc_dec_prologue
+	.rva	.Lcbc_dec_epilogue
+	.rva	.Lcbc_dec_info
+
+	.rva	.Lctr_enc_prologue
+	.rva	.Lctr_enc_epilogue
+	.rva	.Lctr_enc_info
+
+	.rva	.Lxts_enc_prologue
+	.rva	.Lxts_enc_epilogue
+	.rva	.Lxts_enc_info
+
+	.rva	.Lxts_dec_prologue
+	.rva	.Lxts_dec_epilogue
+	.rva	.Lxts_dec_info
+
+.section	.xdata
+.align	8
+___
+$code.=<<___ if ($ecb);
+.Lecb_enc_info:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lecb_enc_body,.Lecb_enc_epilogue	# HandlerData[]
+.Lecb_dec_info:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lecb_dec_body,.Lecb_dec_epilogue	# HandlerData[]
+___
+$code.=<<___;
+.Lcbc_dec_info:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lcbc_dec_body,.Lcbc_dec_epilogue	# HandlerData[]
+.Lctr_enc_info:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lctr_enc_body,.Lctr_enc_epilogue	# HandlerData[]
+.Lxts_enc_info:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lxts_enc_body,.Lxts_enc_epilogue	# HandlerData[]
+.Lxts_dec_info:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lxts_dec_body,.Lxts_dec_epilogue	# HandlerData[]
+___
+}
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+
+print $code;
+
+close STDOUT;
diff --git a/crypto/aes/asm/vpaes-x86.pl b/crypto/aes/asm/vpaes-x86.pl
new file mode 100755
index 000000000000..1533e2c3042c
--- /dev/null
+++ b/crypto/aes/asm/vpaes-x86.pl
@@ -0,0 +1,903 @@
+#!/usr/bin/env perl
+
+######################################################################
+## Constant-time SSSE3 AES core implementation.
+## version 0.1
+##
+## By Mike Hamburg (Stanford University), 2009
+## Public domain.
+##
+## For details see http://shiftleft.org/papers/vector_aes/ and
+## http://crypto.stanford.edu/vpaes/.
+
+######################################################################
+# September 2011.
+#
+# Port vpaes-x86_64.pl as 32-bit "almost" drop-in replacement for
+# aes-586.pl. "Almost" refers to the fact that AES_cbc_encrypt
+# doesn't handle partial vectors (doesn't have to if called from
+# EVP only). "Drop-in" implies that this module doesn't share key
+# schedule structure with the original nor does it make assumption
+# about its alignment...
+#
+# Performance summary. aes-586.pl column lists large-block CBC
+# encrypt/decrypt/with-hyper-threading-off(*) results in cycles per
+# byte processed with 128-bit key, and vpaes-x86.pl column - [also
+# large-block CBC] encrypt/decrypt.
+#
+#		aes-586.pl		vpaes-x86.pl
+#
+# Core 2(**)	29.1/42.3/18.3		22.0/25.6(***)
+# Nehalem	27.9/40.4/18.1		10.3/12.0
+# Atom		102./119./60.1		64.5/85.3(***)
+#
+# (*)	"Hyper-threading" in the context refers rather to cache shared
+#	among multiple cores, than to specifically Intel HTT. As vast
+#	majority of contemporary cores share cache, slower code path
+#	is common place. In other words "with-hyper-threading-off"
+#	results are presented mostly for reference purposes.
+#
+# (**)	"Core 2" refers to initial 65nm design, a.k.a. Conroe.
+#
+# (***)	Less impressive improvement on Core 2 and Atom is due to slow
+#	pshufb,	yet it's respectable +32%/65%  improvement on Core 2
+#	and +58%/40% on Atom (as implied, over "hyper-threading-safe"
+#	code path).
+#
+#						<appro@openssl.org>
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"vpaes-x86.pl",$x86only = $ARGV[$#ARGV] eq "386");
+
+$PREFIX="vpaes";
+
+my  ($round, $base, $magic, $key, $const, $inp, $out)=
+    ("eax",  "ebx", "ecx",  "edx","ebp",  "esi","edi");
+
+&static_label("_vpaes_consts");
+&static_label("_vpaes_schedule_low_round");
+
+&set_label("_vpaes_consts",64);
+$k_inv=-0x30;		# inv, inva
+	&data_word(0x0D080180,0x0E05060F,0x0A0B0C02,0x04070309);
+	&data_word(0x0F0B0780,0x01040A06,0x02050809,0x030D0E0C);
+
+$k_s0F=-0x10;		# s0F
+	&data_word(0x0F0F0F0F,0x0F0F0F0F,0x0F0F0F0F,0x0F0F0F0F);
+
+$k_ipt=0x00;		# input transform (lo, hi)
+	&data_word(0x5A2A7000,0xC2B2E898,0x52227808,0xCABAE090);
+	&data_word(0x317C4D00,0x4C01307D,0xB0FDCC81,0xCD80B1FC);
+
+$k_sb1=0x20;		# sb1u, sb1t
+	&data_word(0xCB503E00,0xB19BE18F,0x142AF544,0xA5DF7A6E);
+	&data_word(0xFAE22300,0x3618D415,0x0D2ED9EF,0x3BF7CCC1);
+$k_sb2=0x40;		# sb2u, sb2t
+	&data_word(0x0B712400,0xE27A93C6,0xBC982FCD,0x5EB7E955);
+	&data_word(0x0AE12900,0x69EB8840,0xAB82234A,0xC2A163C8);
+$k_sbo=0x60;		# sbou, sbot
+	&data_word(0x6FBDC700,0xD0D26D17,0xC502A878,0x15AABF7A);
+	&data_word(0x5FBB6A00,0xCFE474A5,0x412B35FA,0x8E1E90D1);
+
+$k_mc_forward=0x80;	# mc_forward
+	&data_word(0x00030201,0x04070605,0x080B0A09,0x0C0F0E0D);
+	&data_word(0x04070605,0x080B0A09,0x0C0F0E0D,0x00030201);
+	&data_word(0x080B0A09,0x0C0F0E0D,0x00030201,0x04070605);
+	&data_word(0x0C0F0E0D,0x00030201,0x04070605,0x080B0A09);
+
+$k_mc_backward=0xc0;	# mc_backward
+	&data_word(0x02010003,0x06050407,0x0A09080B,0x0E0D0C0F);
+	&data_word(0x0E0D0C0F,0x02010003,0x06050407,0x0A09080B);
+	&data_word(0x0A09080B,0x0E0D0C0F,0x02010003,0x06050407);
+	&data_word(0x06050407,0x0A09080B,0x0E0D0C0F,0x02010003);
+
+$k_sr=0x100;		# sr
+	&data_word(0x03020100,0x07060504,0x0B0A0908,0x0F0E0D0C);
+	&data_word(0x0F0A0500,0x030E0904,0x07020D08,0x0B06010C);
+	&data_word(0x0B020900,0x0F060D04,0x030A0108,0x070E050C);
+	&data_word(0x070A0D00,0x0B0E0104,0x0F020508,0x0306090C);
+
+$k_rcon=0x140;		# rcon
+	&data_word(0xAF9DEEB6,0x1F8391B9,0x4D7C7D81,0x702A9808);
+
+$k_s63=0x150;		# s63: all equal to 0x63 transformed
+	&data_word(0x5B5B5B5B,0x5B5B5B5B,0x5B5B5B5B,0x5B5B5B5B);
+
+$k_opt=0x160;		# output transform
+	&data_word(0xD6B66000,0xFF9F4929,0xDEBE6808,0xF7974121);
+	&data_word(0x50BCEC00,0x01EDBD51,0xB05C0CE0,0xE10D5DB1);
+
+$k_deskew=0x180;	# deskew tables: inverts the sbox's "skew"
+	&data_word(0x47A4E300,0x07E4A340,0x5DBEF91A,0x1DFEB95A);
+	&data_word(0x83EA6900,0x5F36B5DC,0xF49D1E77,0x2841C2AB);
+##
+##  Decryption stuff
+##  Key schedule constants
+##
+$k_dksd=0x1a0;		# decryption key schedule: invskew x*D
+	&data_word(0xA3E44700,0xFEB91A5D,0x5A1DBEF9,0x0740E3A4);
+	&data_word(0xB5368300,0x41C277F4,0xAB289D1E,0x5FDC69EA);
+$k_dksb=0x1c0;		# decryption key schedule: invskew x*B
+	&data_word(0x8550D500,0x9A4FCA1F,0x1CC94C99,0x03D65386);
+	&data_word(0xB6FC4A00,0x115BEDA7,0x7E3482C8,0xD993256F);
+$k_dkse=0x1e0;		# decryption key schedule: invskew x*E + 0x63
+	&data_word(0x1FC9D600,0xD5031CCA,0x994F5086,0x53859A4C);
+	&data_word(0x4FDC7BE8,0xA2319605,0x20B31487,0xCD5EF96A);
+$k_dks9=0x200;		# decryption key schedule: invskew x*9
+	&data_word(0x7ED9A700,0xB6116FC8,0x82255BFC,0x4AED9334);
+	&data_word(0x27143300,0x45765162,0xE9DAFDCE,0x8BB89FAC);
+
+##
+##  Decryption stuff
+##  Round function constants
+##
+$k_dipt=0x220;		# decryption input transform
+	&data_word(0x0B545F00,0x0F505B04,0x114E451A,0x154A411E);
+	&data_word(0x60056500,0x86E383E6,0xF491F194,0x12771772);
+
+$k_dsb9=0x240;		# decryption sbox output *9*u, *9*t
+	&data_word(0x9A86D600,0x851C0353,0x4F994CC9,0xCAD51F50);
+	&data_word(0xECD74900,0xC03B1789,0xB2FBA565,0x725E2C9E);
+$k_dsbd=0x260;		# decryption sbox output *D*u, *D*t
+	&data_word(0xE6B1A200,0x7D57CCDF,0x882A4439,0xF56E9B13);
+	&data_word(0x24C6CB00,0x3CE2FAF7,0x15DEEFD3,0x2931180D);
+$k_dsbb=0x280;		# decryption sbox output *B*u, *B*t
+	&data_word(0x96B44200,0xD0226492,0xB0F2D404,0x602646F6);
+	&data_word(0xCD596700,0xC19498A6,0x3255AA6B,0xF3FF0C3E);
+$k_dsbe=0x2a0;		# decryption sbox output *E*u, *E*t
+	&data_word(0x26D4D000,0x46F29296,0x64B4F6B0,0x22426004);
+	&data_word(0xFFAAC100,0x0C55A6CD,0x98593E32,0x9467F36B);
+$k_dsbo=0x2c0;		# decryption sbox final output
+	&data_word(0x7EF94000,0x1387EA53,0xD4943E2D,0xC7AA6DB9);
+	&data_word(0x93441D00,0x12D7560F,0xD8C58E9C,0xCA4B8159);
+&asciz	("Vector Permutation AES for x86/SSSE3, Mike Hamburg (Stanford University)");
+&align	(64);
+
+&function_begin_B("_vpaes_preheat");
+	&add	($const,&DWP(0,"esp"));
+	&movdqa	("xmm7",&QWP($k_inv,$const));
+	&movdqa	("xmm6",&QWP($k_s0F,$const));
+	&ret	();
+&function_end_B("_vpaes_preheat");
+
+##
+##  _aes_encrypt_core
+##
+##  AES-encrypt %xmm0.
+##
+##  Inputs:
+##     %xmm0 = input
+##     %xmm6-%xmm7 as in _vpaes_preheat
+##    (%edx) = scheduled keys
+##
+##  Output in %xmm0
+##  Clobbers  %xmm1-%xmm5, %eax, %ebx, %ecx, %edx
+##
+##
+&function_begin_B("_vpaes_encrypt_core");
+	&mov	($magic,16);
+	&mov	($round,&DWP(240,$key));
+	&movdqa	("xmm1","xmm6")
+	&movdqa	("xmm2",&QWP($k_ipt,$const));
+	&pandn	("xmm1","xmm0");
+	&movdqu	("xmm5",&QWP(0,$key));
+	&psrld	("xmm1",4);
+	&pand	("xmm0","xmm6");
+	&pshufb	("xmm2","xmm0");
+	&movdqa	("xmm0",&QWP($k_ipt+16,$const));
+	&pshufb	("xmm0","xmm1");
+	&pxor	("xmm2","xmm5");
+	&pxor	("xmm0","xmm2");
+	&add	($key,16);
+	&lea	($base,&DWP($k_mc_backward,$const));
+	&jmp	(&label("enc_entry"));
+
+
+&set_label("enc_loop",16);
+	# middle of middle round
+	&movdqa	("xmm4",&QWP($k_sb1,$const));	# 4 : sb1u
+	&pshufb	("xmm4","xmm2");		# 4 = sb1u
+	&pxor	("xmm4","xmm5");		# 4 = sb1u + k
+	&movdqa	("xmm0",&QWP($k_sb1+16,$const));# 0 : sb1t
+	&pshufb	("xmm0","xmm3");		# 0 = sb1t
+	&pxor	("xmm0","xmm4");		# 0 = A
+	&movdqa	("xmm5",&QWP($k_sb2,$const));	# 4 : sb2u
+	&pshufb	("xmm5","xmm2");		# 4 = sb2u
+	&movdqa	("xmm1",&QWP(-0x40,$base,$magic));# .Lk_mc_forward[]
+	&movdqa	("xmm2",&QWP($k_sb2+16,$const));# 2 : sb2t
+	&pshufb	("xmm2","xmm3");		# 2 = sb2t
+	&pxor	("xmm2","xmm5");		# 2 = 2A
+	&movdqa	("xmm4",&QWP(0,$base,$magic));	# .Lk_mc_backward[]
+	&movdqa	("xmm3","xmm0");		# 3 = A
+	&pshufb	("xmm0","xmm1");		# 0 = B
+	&add	($key,16);			# next key
+	&pxor	("xmm0","xmm2");		# 0 = 2A+B
+	&pshufb	("xmm3","xmm4");		# 3 = D
+	&add	($magic,16);			# next mc
+	&pxor	("xmm3","xmm0");		# 3 = 2A+B+D
+	&pshufb	("xmm0","xmm1");		# 0 = 2B+C
+	&and	($magic,0x30);			# ... mod 4
+	&pxor	("xmm0","xmm3");		# 0 = 2A+3B+C+D
+	&sub	($round,1);			# nr--
+
+&set_label("enc_entry");
+	# top of round
+	&movdqa	("xmm1","xmm6");		# 1 : i
+	&pandn	("xmm1","xmm0");		# 1 = i<<4
+	&psrld	("xmm1",4);			# 1 = i
+	&pand	("xmm0","xmm6");		# 0 = k
+	&movdqa	("xmm5",&QWP($k_inv+16,$const));# 2 : a/k
+	&pshufb	("xmm5","xmm0");		# 2 = a/k
+	&pxor	("xmm0","xmm1");		# 0 = j
+	&movdqa	("xmm3","xmm7");		# 3 : 1/i
+	&pshufb	("xmm3","xmm1");		# 3 = 1/i
+	&pxor	("xmm3","xmm5");		# 3 = iak = 1/i + a/k
+	&movdqa	("xmm4","xmm7");		# 4 : 1/j
+	&pshufb	("xmm4","xmm0");		# 4 = 1/j
+	&pxor	("xmm4","xmm5");		# 4 = jak = 1/j + a/k
+	&movdqa	("xmm2","xmm7");		# 2 : 1/iak
+	&pshufb	("xmm2","xmm3");		# 2 = 1/iak
+	&pxor	("xmm2","xmm0");		# 2 = io
+	&movdqa	("xmm3","xmm7");		# 3 : 1/jak
+	&movdqu	("xmm5",&QWP(0,$key));
+	&pshufb	("xmm3","xmm4");		# 3 = 1/jak
+	&pxor	("xmm3","xmm1");		# 3 = jo
+	&jnz	(&label("enc_loop"));
+
+	# middle of last round
+	&movdqa	("xmm4",&QWP($k_sbo,$const));	# 3 : sbou      .Lk_sbo
+	&movdqa	("xmm0",&QWP($k_sbo+16,$const));# 3 : sbot      .Lk_sbo+16
+	&pshufb	("xmm4","xmm2");		# 4 = sbou
+	&pxor	("xmm4","xmm5");		# 4 = sb1u + k
+	&pshufb	("xmm0","xmm3");		# 0 = sb1t
+	&movdqa	("xmm1",&QWP(0x40,$base,$magic));# .Lk_sr[]
+	&pxor	("xmm0","xmm4");		# 0 = A
+	&pshufb	("xmm0","xmm1");
+	&ret	();
+&function_end_B("_vpaes_encrypt_core");
+
+##
+##  Decryption core
+##
+##  Same API as encryption core.
+##
+&function_begin_B("_vpaes_decrypt_core");
+	&mov	($round,&DWP(240,$key));
+	&lea	($base,&DWP($k_dsbd,$const));
+	&movdqa	("xmm1","xmm6");
+	&movdqa	("xmm2",&QWP($k_dipt-$k_dsbd,$base));
+	&pandn	("xmm1","xmm0");
+	&mov	($magic,$round);
+	&psrld	("xmm1",4)
+	&movdqu	("xmm5",&QWP(0,$key));
+	&shl	($magic,4);
+	&pand	("xmm0","xmm6");
+	&pshufb	("xmm2","xmm0");
+	&movdqa	("xmm0",&QWP($k_dipt-$k_dsbd+16,$base));
+	&xor	($magic,0x30);
+	&pshufb	("xmm0","xmm1");
+	&and	($magic,0x30);
+	&pxor	("xmm2","xmm5");
+	&movdqa	("xmm5",&QWP($k_mc_forward+48,$const));
+	&pxor	("xmm0","xmm2");
+	&add	($key,16);
+	&lea	($magic,&DWP($k_sr-$k_dsbd,$base,$magic));
+	&jmp	(&label("dec_entry"));
+
+&set_label("dec_loop",16);
+##
+##  Inverse mix columns
+##
+	&movdqa	("xmm4",&QWP(-0x20,$base));	# 4 : sb9u
+	&pshufb	("xmm4","xmm2");		# 4 = sb9u
+	&pxor	("xmm4","xmm0");
+	&movdqa	("xmm0",&QWP(-0x10,$base));	# 0 : sb9t
+	&pshufb	("xmm0","xmm3");		# 0 = sb9t
+	&pxor	("xmm0","xmm4");		# 0 = ch
+	&add	($key,16);			# next round key
+
+	&pshufb	("xmm0","xmm5");		# MC ch
+	&movdqa	("xmm4",&QWP(0,$base));		# 4 : sbdu
+	&pshufb	("xmm4","xmm2");		# 4 = sbdu
+	&pxor	("xmm4","xmm0");		# 4 = ch
+	&movdqa	("xmm0",&QWP(0x10,$base));	# 0 : sbdt
+	&pshufb	("xmm0","xmm3");		# 0 = sbdt
+	&pxor	("xmm0","xmm4");		# 0 = ch
+	&sub	($round,1);			# nr--
+
+	&pshufb	("xmm0","xmm5");		# MC ch
+	&movdqa	("xmm4",&QWP(0x20,$base));	# 4 : sbbu
+	&pshufb	("xmm4","xmm2");		# 4 = sbbu
+	&pxor	("xmm4","xmm0");		# 4 = ch
+	&movdqa	("xmm0",&QWP(0x30,$base));	# 0 : sbbt
+	&pshufb	("xmm0","xmm3");		# 0 = sbbt
+	&pxor	("xmm0","xmm4");		# 0 = ch
+
+	&pshufb	("xmm0","xmm5");		# MC ch
+	&movdqa	("xmm4",&QWP(0x40,$base));	# 4 : sbeu
+	&pshufb	("xmm4","xmm2");		# 4 = sbeu
+	&pxor	("xmm4","xmm0");		# 4 = ch
+	&movdqa	("xmm0",&QWP(0x50,$base));	# 0 : sbet
+	&pshufb	("xmm0","xmm3");		# 0 = sbet
+	&pxor	("xmm0","xmm4");		# 0 = ch
+
+	&palignr("xmm5","xmm5",12);
+
+&set_label("dec_entry");
+	# top of round
+	&movdqa	("xmm1","xmm6");		# 1 : i
+	&pandn	("xmm1","xmm0");		# 1 = i<<4
+	&psrld	("xmm1",4);			# 1 = i
+	&pand	("xmm0","xmm6");		# 0 = k
+	&movdqa	("xmm2",&QWP($k_inv+16,$const));# 2 : a/k
+	&pshufb	("xmm2","xmm0");		# 2 = a/k
+	&pxor	("xmm0","xmm1");		# 0 = j
+	&movdqa	("xmm3","xmm7");		# 3 : 1/i
+	&pshufb	("xmm3","xmm1");		# 3 = 1/i
+	&pxor	("xmm3","xmm2");		# 3 = iak = 1/i + a/k
+	&movdqa	("xmm4","xmm7");		# 4 : 1/j
+	&pshufb	("xmm4","xmm0");		# 4 = 1/j
+	&pxor	("xmm4","xmm2");		# 4 = jak = 1/j + a/k
+	&movdqa	("xmm2","xmm7");		# 2 : 1/iak
+	&pshufb	("xmm2","xmm3");		# 2 = 1/iak
+	&pxor	("xmm2","xmm0");		# 2 = io
+	&movdqa	("xmm3","xmm7");		# 3 : 1/jak
+	&pshufb	("xmm3","xmm4");		# 3 = 1/jak
+	&pxor	("xmm3","xmm1");		# 3 = jo
+	&movdqu	("xmm0",&QWP(0,$key));
+	&jnz	(&label("dec_loop"));
+
+	# middle of last round
+	&movdqa	("xmm4",&QWP(0x60,$base));	# 3 : sbou
+	&pshufb	("xmm4","xmm2");		# 4 = sbou
+	&pxor	("xmm4","xmm0");		# 4 = sb1u + k
+	&movdqa	("xmm0",&QWP(0x70,$base));	# 0 : sbot
+	&movdqa	("xmm2",&QWP(0,$magic));
+	&pshufb	("xmm0","xmm3");		# 0 = sb1t
+	&pxor	("xmm0","xmm4");		# 0 = A
+	&pshufb	("xmm0","xmm2");
+	&ret	();
+&function_end_B("_vpaes_decrypt_core");
+
+########################################################
+##                                                    ##
+##                  AES key schedule                  ##
+##                                                    ##
+########################################################
+&function_begin_B("_vpaes_schedule_core");
+	&add	($const,&DWP(0,"esp"));
+	&movdqu	("xmm0",&QWP(0,$inp));		# load key (unaligned)
+	&movdqa	("xmm2",&QWP($k_rcon,$const));	# load rcon
+
+	# input transform
+	&movdqa	("xmm3","xmm0");
+	&lea	($base,&DWP($k_ipt,$const));
+	&movdqa	(&QWP(4,"esp"),"xmm2");		# xmm8
+	&call	("_vpaes_schedule_transform");
+	&movdqa	("xmm7","xmm0");
+
+	&test	($out,$out);
+	&jnz	(&label("schedule_am_decrypting"));
+
+	# encrypting, output zeroth round key after transform
+	&movdqu	(&QWP(0,$key),"xmm0");
+	&jmp	(&label("schedule_go"));
+
+&set_label("schedule_am_decrypting");
+	# decrypting, output zeroth round key after shiftrows
+	&movdqa	("xmm1",&QWP($k_sr,$const,$magic));
+	&pshufb	("xmm3","xmm1");
+	&movdqu	(&QWP(0,$key),"xmm3");
+	&xor	($magic,0x30);
+
+&set_label("schedule_go");
+	&cmp	($round,192);
+	&ja	(&label("schedule_256"));
+	&je	(&label("schedule_192"));
+	# 128: fall though
+
+##
+##  .schedule_128
+##
+##  128-bit specific part of key schedule.
+##
+##  This schedule is really simple, because all its parts
+##  are accomplished by the subroutines.
+##
+&set_label("schedule_128");
+	&mov	($round,10);
+
+&set_label("loop_schedule_128");
+	&call	("_vpaes_schedule_round");
+	&dec	($round);
+	&jz	(&label("schedule_mangle_last"));
+	&call	("_vpaes_schedule_mangle");	# write output
+	&jmp	(&label("loop_schedule_128"));
+
+##
+##  .aes_schedule_192
+##
+##  192-bit specific part of key schedule.
+##
+##  The main body of this schedule is the same as the 128-bit
+##  schedule, but with more smearing.  The long, high side is
+##  stored in %xmm7 as before, and the short, low side is in
+##  the high bits of %xmm6.
+##
+##  This schedule is somewhat nastier, however, because each
+##  round produces 192 bits of key material, or 1.5 round keys.
+##  Therefore, on each cycle we do 2 rounds and produce 3 round
+##  keys.
+##
+&set_label("schedule_192",16);
+	&movdqu	("xmm0",&QWP(8,$inp));		# load key part 2 (very unaligned)
+	&call	("_vpaes_schedule_transform");	# input transform	
+	&movdqa	("xmm6","xmm0");		# save short part
+	&pxor	("xmm4","xmm4");		# clear 4
+	&movhlps("xmm6","xmm4");		# clobber low side with zeros
+	&mov	($round,4);
+
+&set_label("loop_schedule_192");
+	&call	("_vpaes_schedule_round");
+	&palignr("xmm0","xmm6",8);
+	&call	("_vpaes_schedule_mangle");	# save key n
+	&call	("_vpaes_schedule_192_smear");
+	&call	("_vpaes_schedule_mangle");	# save key n+1
+	&call	("_vpaes_schedule_round");
+	&dec	($round);
+	&jz	(&label("schedule_mangle_last"));
+	&call	("_vpaes_schedule_mangle");	# save key n+2
+	&call	("_vpaes_schedule_192_smear");
+	&jmp	(&label("loop_schedule_192"));
+
+##
+##  .aes_schedule_256
+##
+##  256-bit specific part of key schedule.
+##
+##  The structure here is very similar to the 128-bit
+##  schedule, but with an additional "low side" in
+##  %xmm6.  The low side's rounds are the same as the
+##  high side's, except no rcon and no rotation.
+##
+&set_label("schedule_256",16);
+	&movdqu	("xmm0",&QWP(16,$inp));		# load key part 2 (unaligned)
+	&call	("_vpaes_schedule_transform");	# input transform	
+	&mov	($round,7);
+
+&set_label("loop_schedule_256");
+	&call	("_vpaes_schedule_mangle");	# output low result
+	&movdqa	("xmm6","xmm0");		# save cur_lo in xmm6
+
+	# high round
+	&call	("_vpaes_schedule_round");
+	&dec	($round);
+	&jz	(&label("schedule_mangle_last"));
+	&call	("_vpaes_schedule_mangle");	
+
+	# low round. swap xmm7 and xmm6
+	&pshufd	("xmm0","xmm0",0xFF);
+	&movdqa	(&QWP(20,"esp"),"xmm7");
+	&movdqa	("xmm7","xmm6");
+	&call	("_vpaes_schedule_low_round");
+	&movdqa	("xmm7",&QWP(20,"esp"));
+
+	&jmp	(&label("loop_schedule_256"));
+
+##
+##  .aes_schedule_mangle_last
+##
+##  Mangler for last round of key schedule
+##  Mangles %xmm0
+##    when encrypting, outputs out(%xmm0) ^ 63
+##    when decrypting, outputs unskew(%xmm0)
+##
+##  Always called right before return... jumps to cleanup and exits
+##
+&set_label("schedule_mangle_last",16);
+	# schedule last round key from xmm0
+	&lea	($base,&DWP($k_deskew,$const));
+	&test	($out,$out);
+	&jnz	(&label("schedule_mangle_last_dec"));
+
+	# encrypting
+	&movdqa	("xmm1",&QWP($k_sr,$const,$magic));
+	&pshufb	("xmm0","xmm1");		# output permute
+	&lea	($base,&DWP($k_opt,$const));	# prepare to output transform
+	&add	($key,32);
+
+&set_label("schedule_mangle_last_dec");
+	&add	($key,-16);
+	&pxor	("xmm0",&QWP($k_s63,$const));
+	&call	("_vpaes_schedule_transform");	# output transform
+	&movdqu	(&QWP(0,$key),"xmm0");		# save last key
+
+	# cleanup
+	&pxor	("xmm0","xmm0");
+	&pxor	("xmm1","xmm1");
+	&pxor	("xmm2","xmm2");
+	&pxor	("xmm3","xmm3");
+	&pxor	("xmm4","xmm4");
+	&pxor	("xmm5","xmm5");
+	&pxor	("xmm6","xmm6");
+	&pxor	("xmm7","xmm7");
+	&ret	();
+&function_end_B("_vpaes_schedule_core");
+
+##
+##  .aes_schedule_192_smear
+##
+##  Smear the short, low side in the 192-bit key schedule.
+##
+##  Inputs:
+##    %xmm7: high side, b  a  x  y
+##    %xmm6:  low side, d  c  0  0
+##    %xmm13: 0
+##
+##  Outputs:
+##    %xmm6: b+c+d  b+c  0  0
+##    %xmm0: b+c+d  b+c  b  a
+##
+&function_begin_B("_vpaes_schedule_192_smear");
+	&pshufd	("xmm0","xmm6",0x80);		# d c 0 0 -> c 0 0 0
+	&pxor	("xmm6","xmm0");		# -> c+d c 0 0
+	&pshufd	("xmm0","xmm7",0xFE);		# b a _ _ -> b b b a
+	&pxor	("xmm6","xmm0");		# -> b+c+d b+c b a
+	&movdqa	("xmm0","xmm6");
+	&pxor	("xmm1","xmm1");
+	&movhlps("xmm6","xmm1");		# clobber low side with zeros
+	&ret	();
+&function_end_B("_vpaes_schedule_192_smear");
+
+##
+##  .aes_schedule_round
+##
+##  Runs one main round of the key schedule on %xmm0, %xmm7
+##
+##  Specifically, runs subbytes on the high dword of %xmm0
+##  then rotates it by one byte and xors into the low dword of
+##  %xmm7.
+##
+##  Adds rcon from low byte of %xmm8, then rotates %xmm8 for
+##  next rcon.
+##
+##  Smears the dwords of %xmm7 by xoring the low into the
+##  second low, result into third, result into highest.
+##
+##  Returns results in %xmm7 = %xmm0.
+##  Clobbers %xmm1-%xmm5.
+##
+&function_begin_B("_vpaes_schedule_round");
+	# extract rcon from xmm8
+	&movdqa	("xmm2",&QWP(8,"esp"));		# xmm8
+	&pxor	("xmm1","xmm1");
+	&palignr("xmm1","xmm2",15);
+	&palignr("xmm2","xmm2",15);
+	&pxor	("xmm7","xmm1");
+
+	# rotate
+	&pshufd	("xmm0","xmm0",0xFF);
+	&palignr("xmm0","xmm0",1);
+
+	# fall through...
+	&movdqa	(&QWP(8,"esp"),"xmm2");		# xmm8
+
+	# low round: same as high round, but no rotation and no rcon.
+&set_label("_vpaes_schedule_low_round");
+	# smear xmm7
+	&movdqa	("xmm1","xmm7");
+	&pslldq	("xmm7",4);
+	&pxor	("xmm7","xmm1");
+	&movdqa	("xmm1","xmm7");
+	&pslldq	("xmm7",8);
+	&pxor	("xmm7","xmm1");
+	&pxor	("xmm7",&QWP($k_s63,$const));
+
+	# subbyte
+	&movdqa	("xmm4",&QWP($k_s0F,$const));
+	&movdqa	("xmm5",&QWP($k_inv,$const));	# 4 : 1/j
+	&movdqa	("xmm1","xmm4");	
+	&pandn	("xmm1","xmm0");
+	&psrld	("xmm1",4);			# 1 = i
+	&pand	("xmm0","xmm4");		# 0 = k
+	&movdqa	("xmm2",&QWP($k_inv+16,$const));# 2 : a/k
+	&pshufb	("xmm2","xmm0");		# 2 = a/k
+	&pxor	("xmm0","xmm1");		# 0 = j
+	&movdqa	("xmm3","xmm5");		# 3 : 1/i
+	&pshufb	("xmm3","xmm1");		# 3 = 1/i
+	&pxor	("xmm3","xmm2");		# 3 = iak = 1/i + a/k
+	&movdqa	("xmm4","xmm5");		# 4 : 1/j
+	&pshufb	("xmm4","xmm0");		# 4 = 1/j
+	&pxor	("xmm4","xmm2");		# 4 = jak = 1/j + a/k
+	&movdqa	("xmm2","xmm5");		# 2 : 1/iak
+	&pshufb	("xmm2","xmm3");		# 2 = 1/iak
+	&pxor	("xmm2","xmm0");		# 2 = io
+	&movdqa	("xmm3","xmm5");		# 3 : 1/jak
+	&pshufb	("xmm3","xmm4");		# 3 = 1/jak
+	&pxor	("xmm3","xmm1");		# 3 = jo
+	&movdqa	("xmm4",&QWP($k_sb1,$const));	# 4 : sbou
+	&pshufb	("xmm4","xmm2");		# 4 = sbou
+	&movdqa	("xmm0",&QWP($k_sb1+16,$const));# 0 : sbot
+	&pshufb	("xmm0","xmm3");		# 0 = sb1t
+	&pxor	("xmm0","xmm4");		# 0 = sbox output
+
+	# add in smeared stuff
+	&pxor	("xmm0","xmm7");
+	&movdqa	("xmm7","xmm0");
+	&ret	();
+&function_end_B("_vpaes_schedule_round");
+
+##
+##  .aes_schedule_transform
+##
+##  Linear-transform %xmm0 according to tables at (%ebx)
+##
+##  Output in %xmm0
+##  Clobbers %xmm1, %xmm2
+##
+&function_begin_B("_vpaes_schedule_transform");
+	&movdqa	("xmm2",&QWP($k_s0F,$const));
+	&movdqa	("xmm1","xmm2");
+	&pandn	("xmm1","xmm0");
+	&psrld	("xmm1",4);
+	&pand	("xmm0","xmm2");
+	&movdqa	("xmm2",&QWP(0,$base));
+	&pshufb	("xmm2","xmm0");
+	&movdqa	("xmm0",&QWP(16,$base));
+	&pshufb	("xmm0","xmm1");
+	&pxor	("xmm0","xmm2");
+	&ret	();
+&function_end_B("_vpaes_schedule_transform");
+
+##
+##  .aes_schedule_mangle
+##
+##  Mangle xmm0 from (basis-transformed) standard version
+##  to our version.
+##
+##  On encrypt,
+##    xor with 0x63
+##    multiply by circulant 0,1,1,1
+##    apply shiftrows transform
+##
+##  On decrypt,
+##    xor with 0x63
+##    multiply by "inverse mixcolumns" circulant E,B,D,9
+##    deskew
+##    apply shiftrows transform
+##
+##
+##  Writes out to (%edx), and increments or decrements it
+##  Keeps track of round number mod 4 in %ecx
+##  Preserves xmm0
+##  Clobbers xmm1-xmm5
+##
+&function_begin_B("_vpaes_schedule_mangle");
+	&movdqa	("xmm4","xmm0");	# save xmm0 for later
+	&movdqa	("xmm5",&QWP($k_mc_forward,$const));
+	&test	($out,$out);
+	&jnz	(&label("schedule_mangle_dec"));
+
+	# encrypting
+	&add	($key,16);
+	&pxor	("xmm4",&QWP($k_s63,$const));
+	&pshufb	("xmm4","xmm5");
+	&movdqa	("xmm3","xmm4");
+	&pshufb	("xmm4","xmm5");
+	&pxor	("xmm3","xmm4");
+	&pshufb	("xmm4","xmm5");
+	&pxor	("xmm3","xmm4");
+
+	&jmp	(&label("schedule_mangle_both"));
+
+&set_label("schedule_mangle_dec",16);
+	# inverse mix columns
+	&movdqa	("xmm2",&QWP($k_s0F,$const));
+	&lea	($inp,&DWP($k_dksd,$const));
+	&movdqa	("xmm1","xmm2");
+	&pandn	("xmm1","xmm4");
+	&psrld	("xmm1",4);			# 1 = hi
+	&pand	("xmm4","xmm2");		# 4 = lo
+
+	&movdqa	("xmm2",&QWP(0,$inp));
+	&pshufb	("xmm2","xmm4");
+	&movdqa	("xmm3",&QWP(0x10,$inp));
+	&pshufb	("xmm3","xmm1");
+	&pxor	("xmm3","xmm2");
+	&pshufb	("xmm3","xmm5");
+
+	&movdqa	("xmm2",&QWP(0x20,$inp));
+	&pshufb	("xmm2","xmm4");
+	&pxor	("xmm2","xmm3");
+	&movdqa	("xmm3",&QWP(0x30,$inp));
+	&pshufb	("xmm3","xmm1");
+	&pxor	("xmm3","xmm2");
+	&pshufb	("xmm3","xmm5");
+
+	&movdqa	("xmm2",&QWP(0x40,$inp));
+	&pshufb	("xmm2","xmm4");
+	&pxor	("xmm2","xmm3");
+	&movdqa	("xmm3",&QWP(0x50,$inp));
+	&pshufb	("xmm3","xmm1");
+	&pxor	("xmm3","xmm2");
+	&pshufb	("xmm3","xmm5");
+
+	&movdqa	("xmm2",&QWP(0x60,$inp));
+	&pshufb	("xmm2","xmm4");
+	&pxor	("xmm2","xmm3");
+	&movdqa	("xmm3",&QWP(0x70,$inp));
+	&pshufb	("xmm3","xmm1");
+	&pxor	("xmm3","xmm2");
+
+	&add	($key,-16);
+
+&set_label("schedule_mangle_both");
+	&movdqa	("xmm1",&QWP($k_sr,$const,$magic));
+	&pshufb	("xmm3","xmm1");
+	&add	($magic,-16);
+	&and	($magic,0x30);
+	&movdqu	(&QWP(0,$key),"xmm3");
+	&ret	();
+&function_end_B("_vpaes_schedule_mangle");
+
+#
+# Interface to OpenSSL
+#
+&function_begin("${PREFIX}_set_encrypt_key");
+	&mov	($inp,&wparam(0));		# inp
+	&lea	($base,&DWP(-56,"esp"));
+	&mov	($round,&wparam(1));		# bits
+	&and	($base,-16);
+	&mov	($key,&wparam(2));		# key
+	&xchg	($base,"esp");			# alloca
+	&mov	(&DWP(48,"esp"),$base);
+
+	&mov	($base,$round);
+	&shr	($base,5);
+	&add	($base,5);
+	&mov	(&DWP(240,$key),$base);		# AES_KEY->rounds = nbits/32+5;
+	&mov	($magic,0x30);
+	&mov	($out,0);
+
+	&lea	($const,&DWP(&label("_vpaes_consts")."+0x30-".&label("pic_point")));
+	&call	("_vpaes_schedule_core");
+&set_label("pic_point");
+
+	&mov	("esp",&DWP(48,"esp"));
+	&xor	("eax","eax");
+&function_end("${PREFIX}_set_encrypt_key");
+
+&function_begin("${PREFIX}_set_decrypt_key");
+	&mov	($inp,&wparam(0));		# inp
+	&lea	($base,&DWP(-56,"esp"));
+	&mov	($round,&wparam(1));		# bits
+	&and	($base,-16);
+	&mov	($key,&wparam(2));		# key
+	&xchg	($base,"esp");			# alloca
+	&mov	(&DWP(48,"esp"),$base);
+
+	&mov	($base,$round);
+	&shr	($base,5);
+	&add	($base,5);
+	&mov	(&DWP(240,$key),$base);	# AES_KEY->rounds = nbits/32+5;
+	&shl	($base,4);
+	&lea	($key,&DWP(16,$key,$base));
+
+	&mov	($out,1);
+	&mov	($magic,$round);
+	&shr	($magic,1);
+	&and	($magic,32);
+	&xor	($magic,32);			# nbist==192?0:32;
+
+	&lea	($const,&DWP(&label("_vpaes_consts")."+0x30-".&label("pic_point")));
+	&call	("_vpaes_schedule_core");
+&set_label("pic_point");
+
+	&mov	("esp",&DWP(48,"esp"));
+	&xor	("eax","eax");
+&function_end("${PREFIX}_set_decrypt_key");
+
+&function_begin("${PREFIX}_encrypt");
+	&lea	($const,&DWP(&label("_vpaes_consts")."+0x30-".&label("pic_point")));
+	&call	("_vpaes_preheat");
+&set_label("pic_point");
+	&mov	($inp,&wparam(0));		# inp
+	&lea	($base,&DWP(-56,"esp"));
+	&mov	($out,&wparam(1));		# out
+	&and	($base,-16);
+	&mov	($key,&wparam(2));		# key
+	&xchg	($base,"esp");			# alloca
+	&mov	(&DWP(48,"esp"),$base);
+
+	&movdqu	("xmm0",&QWP(0,$inp));
+	&call	("_vpaes_encrypt_core");
+	&movdqu	(&QWP(0,$out),"xmm0");
+
+	&mov	("esp",&DWP(48,"esp"));
+&function_end("${PREFIX}_encrypt");
+
+&function_begin("${PREFIX}_decrypt");
+	&lea	($const,&DWP(&label("_vpaes_consts")."+0x30-".&label("pic_point")));
+	&call	("_vpaes_preheat");
+&set_label("pic_point");
+	&mov	($inp,&wparam(0));		# inp
+	&lea	($base,&DWP(-56,"esp"));
+	&mov	($out,&wparam(1));		# out
+	&and	($base,-16);
+	&mov	($key,&wparam(2));		# key
+	&xchg	($base,"esp");			# alloca
+	&mov	(&DWP(48,"esp"),$base);
+
+	&movdqu	("xmm0",&QWP(0,$inp));
+	&call	("_vpaes_decrypt_core");
+	&movdqu	(&QWP(0,$out),"xmm0");
+
+	&mov	("esp",&DWP(48,"esp"));
+&function_end("${PREFIX}_decrypt");
+
+&function_begin("${PREFIX}_cbc_encrypt");
+	&mov	($inp,&wparam(0));		# inp
+	&mov	($out,&wparam(1));		# out
+	&mov	($round,&wparam(2));		# len
+	&mov	($key,&wparam(3));		# key
+	&sub	($round,16);
+	&jc	(&label("cbc_abort"));
+	&lea	($base,&DWP(-56,"esp"));
+	&mov	($const,&wparam(4));		# ivp
+	&and	($base,-16);
+	&mov	($magic,&wparam(5));		# enc
+	&xchg	($base,"esp");			# alloca
+	&movdqu	("xmm1",&QWP(0,$const));	# load IV
+	&sub	($out,$inp);
+	&mov	(&DWP(48,"esp"),$base);
+
+	&mov	(&DWP(0,"esp"),$out);		# save out
+	&mov	(&DWP(4,"esp"),$key)		# save key
+	&mov	(&DWP(8,"esp"),$const);		# save ivp
+	&mov	($out,$round);			# $out works as $len
+
+	&lea	($const,&DWP(&label("_vpaes_consts")."+0x30-".&label("pic_point")));
+	&call	("_vpaes_preheat");
+&set_label("pic_point");
+	&cmp	($magic,0);
+	&je	(&label("cbc_dec_loop"));
+	&jmp	(&label("cbc_enc_loop"));
+
+&set_label("cbc_enc_loop",16);
+	&movdqu	("xmm0",&QWP(0,$inp));		# load input
+	&pxor	("xmm0","xmm1");		# inp^=iv
+	&call	("_vpaes_encrypt_core");
+	&mov	($base,&DWP(0,"esp"));		# restore out
+	&mov	($key,&DWP(4,"esp"));		# restore key
+	&movdqa	("xmm1","xmm0");
+	&movdqu	(&QWP(0,$base,$inp),"xmm0");	# write output
+	&lea	($inp,&DWP(16,$inp));
+	&sub	($out,16);
+	&jnc	(&label("cbc_enc_loop"));
+	&jmp	(&label("cbc_done"));
+
+&set_label("cbc_dec_loop",16);
+	&movdqu	("xmm0",&QWP(0,$inp));		# load input
+	&movdqa	(&QWP(16,"esp"),"xmm1");	# save IV
+	&movdqa	(&QWP(32,"esp"),"xmm0");	# save future IV
+	&call	("_vpaes_decrypt_core");
+	&mov	($base,&DWP(0,"esp"));		# restore out
+	&mov	($key,&DWP(4,"esp"));		# restore key
+	&pxor	("xmm0",&QWP(16,"esp"));	# out^=iv
+	&movdqa	("xmm1",&QWP(32,"esp"));	# load next IV
+	&movdqu	(&QWP(0,$base,$inp),"xmm0");	# write output
+	&lea	($inp,&DWP(16,$inp));
+	&sub	($out,16);
+	&jnc	(&label("cbc_dec_loop"));
+
+&set_label("cbc_done");
+	&mov	($base,&DWP(8,"esp"));		# restore ivp
+	&mov	("esp",&DWP(48,"esp"));
+	&movdqu	(&QWP(0,$base),"xmm1");		# write IV
+&set_label("cbc_abort");
+&function_end("${PREFIX}_cbc_encrypt");
+
+&asm_finish();
diff --git a/crypto/aes/asm/vpaes-x86_64.pl b/crypto/aes/asm/vpaes-x86_64.pl
new file mode 100755
index 000000000000..37998db5e139
--- /dev/null
+++ b/crypto/aes/asm/vpaes-x86_64.pl
@@ -0,0 +1,1206 @@
+#!/usr/bin/env perl
+
+######################################################################
+## Constant-time SSSE3 AES core implementation.
+## version 0.1
+##
+## By Mike Hamburg (Stanford University), 2009
+## Public domain.
+##
+## For details see http://shiftleft.org/papers/vector_aes/ and
+## http://crypto.stanford.edu/vpaes/.
+
+######################################################################
+# September 2011.
+#
+# Interface to OpenSSL as "almost" drop-in replacement for
+# aes-x86_64.pl. "Almost" refers to the fact that AES_cbc_encrypt
+# doesn't handle partial vectors (doesn't have to if called from
+# EVP only). "Drop-in" implies that this module doesn't share key
+# schedule structure with the original nor does it make assumption
+# about its alignment...
+#
+# Performance summary. aes-x86_64.pl column lists large-block CBC
+# encrypt/decrypt/with-hyper-threading-off(*) results in cycles per
+# byte processed with 128-bit key, and vpaes-x86_64.pl column -
+# [also large-block CBC] encrypt/decrypt.
+#
+#		aes-x86_64.pl		vpaes-x86_64.pl
+#
+# Core 2(**)	30.5/43.7/14.3		21.8/25.7(***)
+# Nehalem	30.5/42.2/14.6		 9.8/11.8
+# Atom		63.9/79.0/32.1		64.0/84.8(***)
+#
+# (*)	"Hyper-threading" in the context refers rather to cache shared
+#	among multiple cores, than to specifically Intel HTT. As vast
+#	majority of contemporary cores share cache, slower code path
+#	is common place. In other words "with-hyper-threading-off"
+#	results are presented mostly for reference purposes.
+#
+# (**)	"Core 2" refers to initial 65nm design, a.k.a. Conroe.
+#
+# (***)	Less impressive improvement on Core 2 and Atom is due to slow
+#	pshufb,	yet it's respectable +40%/78% improvement on Core 2
+#	(as implied, over "hyper-threading-safe" code path).
+#
+#						<appro@openssl.org>
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+$PREFIX="vpaes";
+
+$code.=<<___;
+.text
+
+##
+##  _aes_encrypt_core
+##
+##  AES-encrypt %xmm0.
+##
+##  Inputs:
+##     %xmm0 = input
+##     %xmm9-%xmm15 as in _vpaes_preheat
+##    (%rdx) = scheduled keys
+##
+##  Output in %xmm0
+##  Clobbers  %xmm1-%xmm5, %r9, %r10, %r11, %rax
+##  Preserves %xmm6 - %xmm8 so you get some local vectors
+##
+##
+.type	_vpaes_encrypt_core,\@abi-omnipotent
+.align 16
+_vpaes_encrypt_core:
+	mov	%rdx,	%r9
+	mov	\$16,	%r11
+	mov	240(%rdx),%eax
+	movdqa	%xmm9,	%xmm1
+	movdqa	.Lk_ipt(%rip), %xmm2	# iptlo
+	pandn	%xmm0,	%xmm1
+	movdqu	(%r9),	%xmm5		# round0 key
+	psrld	\$4,	%xmm1
+	pand	%xmm9,	%xmm0
+	pshufb	%xmm0,	%xmm2
+	movdqa	.Lk_ipt+16(%rip), %xmm0	# ipthi
+	pshufb	%xmm1,	%xmm0
+	pxor	%xmm5,	%xmm2
+	pxor	%xmm2,	%xmm0
+	add	\$16,	%r9
+	lea	.Lk_mc_backward(%rip),%r10
+	jmp	.Lenc_entry
+
+.align 16
+.Lenc_loop:
+	# middle of middle round
+	movdqa  %xmm13,	%xmm4	# 4 : sb1u
+	pshufb  %xmm2,	%xmm4	# 4 = sb1u
+	pxor	%xmm5,	%xmm4	# 4 = sb1u + k
+	movdqa  %xmm12,	%xmm0	# 0 : sb1t
+	pshufb  %xmm3,	%xmm0	# 0 = sb1t
+	pxor	%xmm4,	%xmm0	# 0 = A
+	movdqa  %xmm15,	%xmm5	# 4 : sb2u
+	pshufb	%xmm2,	%xmm5	# 4 = sb2u
+	movdqa	-0x40(%r11,%r10), %xmm1		# .Lk_mc_forward[]
+	movdqa	%xmm14, %xmm2	# 2 : sb2t
+	pshufb	%xmm3,  %xmm2	# 2 = sb2t
+	pxor	%xmm5,	%xmm2	# 2 = 2A
+	movdqa	(%r11,%r10), %xmm4		# .Lk_mc_backward[]
+	movdqa	%xmm0,  %xmm3	# 3 = A
+	pshufb  %xmm1,  %xmm0	# 0 = B
+	add	\$16,	%r9	# next key
+	pxor	%xmm2,  %xmm0	# 0 = 2A+B
+	pshufb	%xmm4,	%xmm3	# 3 = D
+	add	\$16,	%r11	# next mc
+	pxor	%xmm0,	%xmm3	# 3 = 2A+B+D
+	pshufb  %xmm1,	%xmm0	# 0 = 2B+C
+	and	\$0x30,	%r11	# ... mod 4
+	pxor	%xmm3,	%xmm0	# 0 = 2A+3B+C+D
+	sub	\$1,%rax	# nr--
+
+.Lenc_entry:
+	# top of round
+	movdqa  %xmm9, 	%xmm1	# 1 : i
+	pandn	%xmm0, 	%xmm1	# 1 = i<<4
+	psrld	\$4,   	%xmm1   # 1 = i
+	pand	%xmm9, 	%xmm0   # 0 = k
+	movdqa	%xmm11, %xmm5	# 2 : a/k
+	pshufb  %xmm0,  %xmm5	# 2 = a/k
+	pxor	%xmm1,	%xmm0	# 0 = j
+	movdqa	%xmm10,	%xmm3  	# 3 : 1/i
+	pshufb  %xmm1, 	%xmm3  	# 3 = 1/i
+	pxor	%xmm5, 	%xmm3  	# 3 = iak = 1/i + a/k
+	movdqa	%xmm10,	%xmm4  	# 4 : 1/j
+	pshufb	%xmm0, 	%xmm4  	# 4 = 1/j
+	pxor	%xmm5, 	%xmm4  	# 4 = jak = 1/j + a/k
+	movdqa	%xmm10,	%xmm2  	# 2 : 1/iak
+	pshufb  %xmm3,	%xmm2  	# 2 = 1/iak
+	pxor	%xmm0, 	%xmm2  	# 2 = io
+	movdqa	%xmm10, %xmm3   # 3 : 1/jak
+	movdqu	(%r9),	%xmm5
+	pshufb  %xmm4,  %xmm3   # 3 = 1/jak
+	pxor	%xmm1,  %xmm3   # 3 = jo
+	jnz	.Lenc_loop
+
+	# middle of last round
+	movdqa	-0x60(%r10), %xmm4	# 3 : sbou	.Lk_sbo
+	movdqa	-0x50(%r10), %xmm0	# 0 : sbot	.Lk_sbo+16
+	pshufb  %xmm2,  %xmm4	# 4 = sbou
+	pxor	%xmm5,  %xmm4	# 4 = sb1u + k
+	pshufb  %xmm3,	%xmm0	# 0 = sb1t
+	movdqa	0x40(%r11,%r10), %xmm1		# .Lk_sr[]
+	pxor	%xmm4,	%xmm0	# 0 = A
+	pshufb	%xmm1,	%xmm0
+	ret
+.size	_vpaes_encrypt_core,.-_vpaes_encrypt_core
+	
+##
+##  Decryption core
+##
+##  Same API as encryption core.
+##
+.type	_vpaes_decrypt_core,\@abi-omnipotent
+.align	16
+_vpaes_decrypt_core:
+	mov	%rdx,	%r9		# load key
+	mov	240(%rdx),%eax
+	movdqa	%xmm9,	%xmm1
+	movdqa	.Lk_dipt(%rip), %xmm2	# iptlo
+	pandn	%xmm0,	%xmm1
+	mov	%rax,	%r11
+	psrld	\$4,	%xmm1
+	movdqu	(%r9),	%xmm5		# round0 key
+	shl	\$4,	%r11
+	pand	%xmm9,	%xmm0
+	pshufb	%xmm0,	%xmm2
+	movdqa	.Lk_dipt+16(%rip), %xmm0 # ipthi
+	xor	\$0x30,	%r11
+	lea	.Lk_dsbd(%rip),%r10
+	pshufb	%xmm1,	%xmm0
+	and	\$0x30,	%r11
+	pxor	%xmm5,	%xmm2
+	movdqa	.Lk_mc_forward+48(%rip), %xmm5
+	pxor	%xmm2,	%xmm0
+	add	\$16,	%r9
+	add	%r10,	%r11
+	jmp	.Ldec_entry
+
+.align 16
+.Ldec_loop:
+##
+##  Inverse mix columns
+##
+	movdqa  -0x20(%r10),%xmm4	# 4 : sb9u
+	pshufb	%xmm2,	%xmm4		# 4 = sb9u
+	pxor	%xmm0,	%xmm4
+	movdqa  -0x10(%r10),%xmm0	# 0 : sb9t
+	pshufb	%xmm3,	%xmm0		# 0 = sb9t
+	pxor	%xmm4,	%xmm0		# 0 = ch
+	add	\$16, %r9		# next round key
+
+	pshufb	%xmm5,	%xmm0		# MC ch
+	movdqa  0x00(%r10),%xmm4	# 4 : sbdu
+	pshufb	%xmm2,	%xmm4		# 4 = sbdu
+	pxor	%xmm0,	%xmm4		# 4 = ch
+	movdqa  0x10(%r10),%xmm0	# 0 : sbdt
+	pshufb	%xmm3,	%xmm0		# 0 = sbdt
+	pxor	%xmm4,	%xmm0		# 0 = ch
+	sub	\$1,%rax		# nr--
+	
+	pshufb	%xmm5,	%xmm0		# MC ch
+	movdqa  0x20(%r10),%xmm4	# 4 : sbbu
+	pshufb	%xmm2,	%xmm4		# 4 = sbbu
+	pxor	%xmm0,	%xmm4		# 4 = ch
+	movdqa  0x30(%r10),%xmm0	# 0 : sbbt
+	pshufb	%xmm3,	%xmm0		# 0 = sbbt
+	pxor	%xmm4,	%xmm0		# 0 = ch
+	
+	pshufb	%xmm5,	%xmm0		# MC ch
+	movdqa  0x40(%r10),%xmm4	# 4 : sbeu
+	pshufb	%xmm2,	%xmm4		# 4 = sbeu
+	pxor	%xmm0,	%xmm4		# 4 = ch
+	movdqa  0x50(%r10),%xmm0	# 0 : sbet
+	pshufb	%xmm3,	%xmm0		# 0 = sbet
+	pxor	%xmm4,	%xmm0		# 0 = ch
+
+	palignr	\$12,	%xmm5,	%xmm5
+	
+.Ldec_entry:
+	# top of round
+	movdqa  %xmm9, 	%xmm1	# 1 : i
+	pandn	%xmm0, 	%xmm1	# 1 = i<<4
+	psrld	\$4,    %xmm1	# 1 = i
+	pand	%xmm9, 	%xmm0	# 0 = k
+	movdqa	%xmm11, %xmm2	# 2 : a/k
+	pshufb  %xmm0,  %xmm2	# 2 = a/k
+	pxor	%xmm1,	%xmm0	# 0 = j
+	movdqa	%xmm10,	%xmm3	# 3 : 1/i
+	pshufb  %xmm1, 	%xmm3	# 3 = 1/i
+	pxor	%xmm2, 	%xmm3	# 3 = iak = 1/i + a/k
+	movdqa	%xmm10,	%xmm4	# 4 : 1/j
+	pshufb	%xmm0, 	%xmm4	# 4 = 1/j
+	pxor	%xmm2, 	%xmm4	# 4 = jak = 1/j + a/k
+	movdqa	%xmm10,	%xmm2	# 2 : 1/iak
+	pshufb  %xmm3,	%xmm2	# 2 = 1/iak
+	pxor	%xmm0, 	%xmm2	# 2 = io
+	movdqa	%xmm10, %xmm3	# 3 : 1/jak
+	pshufb  %xmm4,  %xmm3	# 3 = 1/jak
+	pxor	%xmm1,  %xmm3	# 3 = jo
+	movdqu	(%r9),	%xmm0
+	jnz	.Ldec_loop
+
+	# middle of last round
+	movdqa	0x60(%r10), %xmm4	# 3 : sbou
+	pshufb  %xmm2,  %xmm4	# 4 = sbou
+	pxor	%xmm0,  %xmm4	# 4 = sb1u + k
+	movdqa	0x70(%r10), %xmm0	# 0 : sbot
+	movdqa	-0x160(%r11), %xmm2	# .Lk_sr-.Lk_dsbd=-0x160
+	pshufb  %xmm3,	%xmm0	# 0 = sb1t
+	pxor	%xmm4,	%xmm0	# 0 = A
+	pshufb	%xmm2,	%xmm0
+	ret
+.size	_vpaes_decrypt_core,.-_vpaes_decrypt_core
+
+########################################################
+##                                                    ##
+##                  AES key schedule                  ##
+##                                                    ##
+########################################################
+.type	_vpaes_schedule_core,\@abi-omnipotent
+.align	16
+_vpaes_schedule_core:
+	# rdi = key
+	# rsi = size in bits
+	# rdx = buffer
+	# rcx = direction.  0=encrypt, 1=decrypt
+
+	call	_vpaes_preheat		# load the tables
+	movdqa	.Lk_rcon(%rip), %xmm8	# load rcon
+	movdqu	(%rdi),	%xmm0		# load key (unaligned)
+
+	# input transform
+	movdqa	%xmm0,	%xmm3
+	lea	.Lk_ipt(%rip), %r11
+	call	_vpaes_schedule_transform
+	movdqa	%xmm0,	%xmm7
+
+	lea	.Lk_sr(%rip),%r10
+	test	%rcx,	%rcx
+	jnz	.Lschedule_am_decrypting
+
+	# encrypting, output zeroth round key after transform
+	movdqu	%xmm0,	(%rdx)
+	jmp	.Lschedule_go
+
+.Lschedule_am_decrypting:
+	# decrypting, output zeroth round key after shiftrows
+	movdqa	(%r8,%r10),%xmm1
+	pshufb  %xmm1,	%xmm3
+	movdqu	%xmm3,	(%rdx)
+	xor	\$0x30, %r8
+
+.Lschedule_go:
+	cmp	\$192,	%esi
+	ja	.Lschedule_256
+	je	.Lschedule_192
+	# 128: fall though
+
+##
+##  .schedule_128
+##
+##  128-bit specific part of key schedule.
+##
+##  This schedule is really simple, because all its parts
+##  are accomplished by the subroutines.
+##
+.Lschedule_128:
+	mov	\$10, %esi
+	
+.Loop_schedule_128:
+	call 	_vpaes_schedule_round
+	dec	%rsi
+	jz 	.Lschedule_mangle_last
+	call	_vpaes_schedule_mangle	# write output
+	jmp 	.Loop_schedule_128
+
+##
+##  .aes_schedule_192
+##
+##  192-bit specific part of key schedule.
+##
+##  The main body of this schedule is the same as the 128-bit
+##  schedule, but with more smearing.  The long, high side is
+##  stored in %xmm7 as before, and the short, low side is in
+##  the high bits of %xmm6.
+##
+##  This schedule is somewhat nastier, however, because each
+##  round produces 192 bits of key material, or 1.5 round keys.
+##  Therefore, on each cycle we do 2 rounds and produce 3 round
+##  keys.
+##
+.align	16
+.Lschedule_192:
+	movdqu	8(%rdi),%xmm0		# load key part 2 (very unaligned)
+	call	_vpaes_schedule_transform	# input transform
+	movdqa	%xmm0,	%xmm6		# save short part
+	pxor	%xmm4,	%xmm4		# clear 4
+	movhlps	%xmm4,	%xmm6		# clobber low side with zeros
+	mov	\$4,	%esi
+
+.Loop_schedule_192:
+	call	_vpaes_schedule_round
+	palignr	\$8,%xmm6,%xmm0	
+	call	_vpaes_schedule_mangle	# save key n
+	call	_vpaes_schedule_192_smear
+	call	_vpaes_schedule_mangle	# save key n+1
+	call	_vpaes_schedule_round
+	dec	%rsi
+	jz 	.Lschedule_mangle_last
+	call	_vpaes_schedule_mangle	# save key n+2
+	call	_vpaes_schedule_192_smear
+	jmp	.Loop_schedule_192
+
+##
+##  .aes_schedule_256
+##
+##  256-bit specific part of key schedule.
+##
+##  The structure here is very similar to the 128-bit
+##  schedule, but with an additional "low side" in
+##  %xmm6.  The low side's rounds are the same as the
+##  high side's, except no rcon and no rotation.
+##
+.align	16
+.Lschedule_256:
+	movdqu	16(%rdi),%xmm0		# load key part 2 (unaligned)
+	call	_vpaes_schedule_transform	# input transform
+	mov	\$7, %esi
+	
+.Loop_schedule_256:
+	call	_vpaes_schedule_mangle	# output low result
+	movdqa	%xmm0,	%xmm6		# save cur_lo in xmm6
+
+	# high round
+	call	_vpaes_schedule_round
+	dec	%rsi
+	jz 	.Lschedule_mangle_last
+	call	_vpaes_schedule_mangle	
+
+	# low round. swap xmm7 and xmm6
+	pshufd	\$0xFF,	%xmm0,	%xmm0
+	movdqa	%xmm7,	%xmm5
+	movdqa	%xmm6,	%xmm7
+	call	_vpaes_schedule_low_round
+	movdqa	%xmm5,	%xmm7
+	
+	jmp	.Loop_schedule_256
+
+	
+##
+##  .aes_schedule_mangle_last
+##
+##  Mangler for last round of key schedule
+##  Mangles %xmm0
+##    when encrypting, outputs out(%xmm0) ^ 63
+##    when decrypting, outputs unskew(%xmm0)
+##
+##  Always called right before return... jumps to cleanup and exits
+##
+.align	16
+.Lschedule_mangle_last:
+	# schedule last round key from xmm0
+	lea	.Lk_deskew(%rip),%r11	# prepare to deskew
+	test	%rcx, 	%rcx
+	jnz	.Lschedule_mangle_last_dec
+
+	# encrypting
+	movdqa	(%r8,%r10),%xmm1
+	pshufb	%xmm1,	%xmm0		# output permute
+	lea	.Lk_opt(%rip),	%r11	# prepare to output transform
+	add	\$32,	%rdx
+
+.Lschedule_mangle_last_dec:
+	add	\$-16,	%rdx
+	pxor	.Lk_s63(%rip),	%xmm0
+	call	_vpaes_schedule_transform # output transform
+	movdqu	%xmm0,	(%rdx)		# save last key
+
+	# cleanup
+	pxor	%xmm0,  %xmm0
+	pxor	%xmm1,  %xmm1
+	pxor	%xmm2,  %xmm2
+	pxor	%xmm3,  %xmm3
+	pxor	%xmm4,  %xmm4
+	pxor	%xmm5,  %xmm5
+	pxor	%xmm6,  %xmm6
+	pxor	%xmm7,  %xmm7
+	ret
+.size	_vpaes_schedule_core,.-_vpaes_schedule_core
+
+##
+##  .aes_schedule_192_smear
+##
+##  Smear the short, low side in the 192-bit key schedule.
+##
+##  Inputs:
+##    %xmm7: high side, b  a  x  y
+##    %xmm6:  low side, d  c  0  0
+##    %xmm13: 0
+##
+##  Outputs:
+##    %xmm6: b+c+d  b+c  0  0
+##    %xmm0: b+c+d  b+c  b  a
+##
+.type	_vpaes_schedule_192_smear,\@abi-omnipotent
+.align	16
+_vpaes_schedule_192_smear:
+	pshufd	\$0x80,	%xmm6,	%xmm0	# d c 0 0 -> c 0 0 0
+	pxor	%xmm0,	%xmm6		# -> c+d c 0 0
+	pshufd	\$0xFE,	%xmm7,	%xmm0	# b a _ _ -> b b b a
+	pxor	%xmm0,	%xmm6		# -> b+c+d b+c b a
+	movdqa	%xmm6,	%xmm0
+	pxor	%xmm1,	%xmm1
+	movhlps	%xmm1,	%xmm6		# clobber low side with zeros
+	ret
+.size	_vpaes_schedule_192_smear,.-_vpaes_schedule_192_smear
+
+##
+##  .aes_schedule_round
+##
+##  Runs one main round of the key schedule on %xmm0, %xmm7
+##
+##  Specifically, runs subbytes on the high dword of %xmm0
+##  then rotates it by one byte and xors into the low dword of
+##  %xmm7.
+##
+##  Adds rcon from low byte of %xmm8, then rotates %xmm8 for
+##  next rcon.
+##
+##  Smears the dwords of %xmm7 by xoring the low into the
+##  second low, result into third, result into highest.
+##
+##  Returns results in %xmm7 = %xmm0.
+##  Clobbers %xmm1-%xmm4, %r11.
+##
+.type	_vpaes_schedule_round,\@abi-omnipotent
+.align	16
+_vpaes_schedule_round:
+	# extract rcon from xmm8
+	pxor	%xmm1,	%xmm1
+	palignr	\$15,	%xmm8,	%xmm1
+	palignr	\$15,	%xmm8,	%xmm8
+	pxor	%xmm1,	%xmm7
+
+	# rotate
+	pshufd	\$0xFF,	%xmm0,	%xmm0
+	palignr	\$1,	%xmm0,	%xmm0
+	
+	# fall through...
+	
+	# low round: same as high round, but no rotation and no rcon.
+_vpaes_schedule_low_round:
+	# smear xmm7
+	movdqa	%xmm7,	%xmm1
+	pslldq	\$4,	%xmm7
+	pxor	%xmm1,	%xmm7
+	movdqa	%xmm7,	%xmm1
+	pslldq	\$8,	%xmm7
+	pxor	%xmm1,	%xmm7
+	pxor	.Lk_s63(%rip), %xmm7
+
+	# subbytes
+	movdqa  %xmm9, 	%xmm1
+	pandn	%xmm0, 	%xmm1
+	psrld	\$4,    %xmm1		# 1 = i
+	pand	%xmm9, 	%xmm0		# 0 = k
+	movdqa	%xmm11, %xmm2		# 2 : a/k
+	pshufb  %xmm0,  %xmm2		# 2 = a/k
+	pxor	%xmm1,	%xmm0		# 0 = j
+	movdqa	%xmm10,	%xmm3		# 3 : 1/i
+	pshufb  %xmm1, 	%xmm3		# 3 = 1/i
+	pxor	%xmm2, 	%xmm3		# 3 = iak = 1/i + a/k
+	movdqa	%xmm10,	%xmm4		# 4 : 1/j
+	pshufb	%xmm0, 	%xmm4		# 4 = 1/j
+	pxor	%xmm2, 	%xmm4		# 4 = jak = 1/j + a/k
+	movdqa	%xmm10,	%xmm2		# 2 : 1/iak
+	pshufb  %xmm3,	%xmm2		# 2 = 1/iak
+	pxor	%xmm0, 	%xmm2		# 2 = io
+	movdqa	%xmm10, %xmm3		# 3 : 1/jak
+	pshufb  %xmm4,  %xmm3		# 3 = 1/jak
+	pxor	%xmm1,  %xmm3		# 3 = jo
+	movdqa	%xmm13, %xmm4		# 4 : sbou
+	pshufb  %xmm2,  %xmm4		# 4 = sbou
+	movdqa	%xmm12, %xmm0		# 0 : sbot
+	pshufb  %xmm3,	%xmm0		# 0 = sb1t
+	pxor	%xmm4, 	%xmm0		# 0 = sbox output
+
+	# add in smeared stuff
+	pxor	%xmm7,	%xmm0	
+	movdqa	%xmm0,	%xmm7
+	ret
+.size	_vpaes_schedule_round,.-_vpaes_schedule_round
+
+##
+##  .aes_schedule_transform
+##
+##  Linear-transform %xmm0 according to tables at (%r11)
+##
+##  Requires that %xmm9 = 0x0F0F... as in preheat
+##  Output in %xmm0
+##  Clobbers %xmm1, %xmm2
+##
+.type	_vpaes_schedule_transform,\@abi-omnipotent
+.align	16
+_vpaes_schedule_transform:
+	movdqa	%xmm9,	%xmm1
+	pandn	%xmm0,	%xmm1
+	psrld	\$4,	%xmm1
+	pand	%xmm9,	%xmm0
+	movdqa	(%r11), %xmm2 	# lo
+	pshufb	%xmm0,	%xmm2
+	movdqa	16(%r11), %xmm0 # hi
+	pshufb	%xmm1,	%xmm0
+	pxor	%xmm2,	%xmm0
+	ret
+.size	_vpaes_schedule_transform,.-_vpaes_schedule_transform
+
+##
+##  .aes_schedule_mangle
+##
+##  Mangle xmm0 from (basis-transformed) standard version
+##  to our version.
+##
+##  On encrypt,
+##    xor with 0x63
+##    multiply by circulant 0,1,1,1
+##    apply shiftrows transform
+##
+##  On decrypt,
+##    xor with 0x63
+##    multiply by "inverse mixcolumns" circulant E,B,D,9
+##    deskew
+##    apply shiftrows transform
+##
+##
+##  Writes out to (%rdx), and increments or decrements it
+##  Keeps track of round number mod 4 in %r8
+##  Preserves xmm0
+##  Clobbers xmm1-xmm5
+##
+.type	_vpaes_schedule_mangle,\@abi-omnipotent
+.align	16
+_vpaes_schedule_mangle:
+	movdqa	%xmm0,	%xmm4	# save xmm0 for later
+	movdqa	.Lk_mc_forward(%rip),%xmm5
+	test	%rcx, 	%rcx
+	jnz	.Lschedule_mangle_dec
+
+	# encrypting
+	add	\$16,	%rdx
+	pxor	.Lk_s63(%rip),%xmm4
+	pshufb	%xmm5,	%xmm4
+	movdqa	%xmm4,	%xmm3
+	pshufb	%xmm5,	%xmm4
+	pxor	%xmm4,	%xmm3
+	pshufb	%xmm5,	%xmm4
+	pxor	%xmm4,	%xmm3
+
+	jmp	.Lschedule_mangle_both
+.align	16
+.Lschedule_mangle_dec:
+	# inverse mix columns
+	lea	.Lk_dksd(%rip),%r11
+	movdqa	%xmm9,	%xmm1
+	pandn	%xmm4,	%xmm1
+	psrld	\$4,	%xmm1	# 1 = hi
+	pand	%xmm9,	%xmm4	# 4 = lo
+
+	movdqa	0x00(%r11), %xmm2
+	pshufb	%xmm4,	%xmm2
+	movdqa	0x10(%r11), %xmm3
+	pshufb	%xmm1,	%xmm3
+	pxor	%xmm2,	%xmm3
+	pshufb	%xmm5,	%xmm3
+
+	movdqa	0x20(%r11), %xmm2
+	pshufb	%xmm4,	%xmm2
+	pxor	%xmm3,	%xmm2
+	movdqa	0x30(%r11), %xmm3
+	pshufb	%xmm1,	%xmm3
+	pxor	%xmm2,	%xmm3
+	pshufb	%xmm5,	%xmm3
+
+	movdqa	0x40(%r11), %xmm2
+	pshufb	%xmm4,	%xmm2
+	pxor	%xmm3,	%xmm2
+	movdqa	0x50(%r11), %xmm3
+	pshufb	%xmm1,	%xmm3
+	pxor	%xmm2,	%xmm3
+	pshufb	%xmm5,	%xmm3
+
+	movdqa	0x60(%r11), %xmm2
+	pshufb	%xmm4,	%xmm2
+	pxor	%xmm3,	%xmm2
+	movdqa	0x70(%r11), %xmm3
+	pshufb	%xmm1,	%xmm3
+	pxor	%xmm2,	%xmm3
+
+	add	\$-16,	%rdx
+
+.Lschedule_mangle_both:
+	movdqa	(%r8,%r10),%xmm1
+	pshufb	%xmm1,%xmm3
+	add	\$-16,	%r8
+	and	\$0x30,	%r8
+	movdqu	%xmm3,	(%rdx)
+	ret
+.size	_vpaes_schedule_mangle,.-_vpaes_schedule_mangle
+
+#
+# Interface to OpenSSL
+#
+.globl	${PREFIX}_set_encrypt_key
+.type	${PREFIX}_set_encrypt_key,\@function,3
+.align	16
+${PREFIX}_set_encrypt_key:
+___
+$code.=<<___ if ($win64);
+	lea	-0xb8(%rsp),%rsp
+	movaps	%xmm6,0x10(%rsp)
+	movaps	%xmm7,0x20(%rsp)
+	movaps	%xmm8,0x30(%rsp)
+	movaps	%xmm9,0x40(%rsp)
+	movaps	%xmm10,0x50(%rsp)
+	movaps	%xmm11,0x60(%rsp)
+	movaps	%xmm12,0x70(%rsp)
+	movaps	%xmm13,0x80(%rsp)
+	movaps	%xmm14,0x90(%rsp)
+	movaps	%xmm15,0xa0(%rsp)
+.Lenc_key_body:
+___
+$code.=<<___;
+	mov	%esi,%eax
+	shr	\$5,%eax
+	add	\$5,%eax
+	mov	%eax,240(%rdx)	# AES_KEY->rounds = nbits/32+5;
+
+	mov	\$0,%ecx
+	mov	\$0x30,%r8d
+	call	_vpaes_schedule_core
+___
+$code.=<<___ if ($win64);
+	movaps	0x10(%rsp),%xmm6
+	movaps	0x20(%rsp),%xmm7
+	movaps	0x30(%rsp),%xmm8
+	movaps	0x40(%rsp),%xmm9
+	movaps	0x50(%rsp),%xmm10
+	movaps	0x60(%rsp),%xmm11
+	movaps	0x70(%rsp),%xmm12
+	movaps	0x80(%rsp),%xmm13
+	movaps	0x90(%rsp),%xmm14
+	movaps	0xa0(%rsp),%xmm15
+	lea	0xb8(%rsp),%rsp
+.Lenc_key_epilogue:
+___
+$code.=<<___;
+	xor	%eax,%eax
+	ret
+.size	${PREFIX}_set_encrypt_key,.-${PREFIX}_set_encrypt_key
+
+.globl	${PREFIX}_set_decrypt_key
+.type	${PREFIX}_set_decrypt_key,\@function,3
+.align	16
+${PREFIX}_set_decrypt_key:
+___
+$code.=<<___ if ($win64);
+	lea	-0xb8(%rsp),%rsp
+	movaps	%xmm6,0x10(%rsp)
+	movaps	%xmm7,0x20(%rsp)
+	movaps	%xmm8,0x30(%rsp)
+	movaps	%xmm9,0x40(%rsp)
+	movaps	%xmm10,0x50(%rsp)
+	movaps	%xmm11,0x60(%rsp)
+	movaps	%xmm12,0x70(%rsp)
+	movaps	%xmm13,0x80(%rsp)
+	movaps	%xmm14,0x90(%rsp)
+	movaps	%xmm15,0xa0(%rsp)
+.Ldec_key_body:
+___
+$code.=<<___;
+	mov	%esi,%eax
+	shr	\$5,%eax
+	add	\$5,%eax
+	mov	%eax,240(%rdx)	# AES_KEY->rounds = nbits/32+5;
+	shl	\$4,%eax
+	lea	16(%rdx,%rax),%rdx
+
+	mov	\$1,%ecx
+	mov	%esi,%r8d
+	shr	\$1,%r8d
+	and	\$32,%r8d
+	xor	\$32,%r8d	# nbits==192?0:32
+	call	_vpaes_schedule_core
+___
+$code.=<<___ if ($win64);
+	movaps	0x10(%rsp),%xmm6
+	movaps	0x20(%rsp),%xmm7
+	movaps	0x30(%rsp),%xmm8
+	movaps	0x40(%rsp),%xmm9
+	movaps	0x50(%rsp),%xmm10
+	movaps	0x60(%rsp),%xmm11
+	movaps	0x70(%rsp),%xmm12
+	movaps	0x80(%rsp),%xmm13
+	movaps	0x90(%rsp),%xmm14
+	movaps	0xa0(%rsp),%xmm15
+	lea	0xb8(%rsp),%rsp
+.Ldec_key_epilogue:
+___
+$code.=<<___;
+	xor	%eax,%eax
+	ret
+.size	${PREFIX}_set_decrypt_key,.-${PREFIX}_set_decrypt_key
+
+.globl	${PREFIX}_encrypt
+.type	${PREFIX}_encrypt,\@function,3
+.align	16
+${PREFIX}_encrypt:
+___
+$code.=<<___ if ($win64);
+	lea	-0xb8(%rsp),%rsp
+	movaps	%xmm6,0x10(%rsp)
+	movaps	%xmm7,0x20(%rsp)
+	movaps	%xmm8,0x30(%rsp)
+	movaps	%xmm9,0x40(%rsp)
+	movaps	%xmm10,0x50(%rsp)
+	movaps	%xmm11,0x60(%rsp)
+	movaps	%xmm12,0x70(%rsp)
+	movaps	%xmm13,0x80(%rsp)
+	movaps	%xmm14,0x90(%rsp)
+	movaps	%xmm15,0xa0(%rsp)
+.Lenc_body:
+___
+$code.=<<___;
+	movdqu	(%rdi),%xmm0
+	call	_vpaes_preheat
+	call	_vpaes_encrypt_core
+	movdqu	%xmm0,(%rsi)
+___
+$code.=<<___ if ($win64);
+	movaps	0x10(%rsp),%xmm6
+	movaps	0x20(%rsp),%xmm7
+	movaps	0x30(%rsp),%xmm8
+	movaps	0x40(%rsp),%xmm9
+	movaps	0x50(%rsp),%xmm10
+	movaps	0x60(%rsp),%xmm11
+	movaps	0x70(%rsp),%xmm12
+	movaps	0x80(%rsp),%xmm13
+	movaps	0x90(%rsp),%xmm14
+	movaps	0xa0(%rsp),%xmm15
+	lea	0xb8(%rsp),%rsp
+.Lenc_epilogue:
+___
+$code.=<<___;
+	ret
+.size	${PREFIX}_encrypt,.-${PREFIX}_encrypt
+
+.globl	${PREFIX}_decrypt
+.type	${PREFIX}_decrypt,\@function,3
+.align	16
+${PREFIX}_decrypt:
+___
+$code.=<<___ if ($win64);
+	lea	-0xb8(%rsp),%rsp
+	movaps	%xmm6,0x10(%rsp)
+	movaps	%xmm7,0x20(%rsp)
+	movaps	%xmm8,0x30(%rsp)
+	movaps	%xmm9,0x40(%rsp)
+	movaps	%xmm10,0x50(%rsp)
+	movaps	%xmm11,0x60(%rsp)
+	movaps	%xmm12,0x70(%rsp)
+	movaps	%xmm13,0x80(%rsp)
+	movaps	%xmm14,0x90(%rsp)
+	movaps	%xmm15,0xa0(%rsp)
+.Ldec_body:
+___
+$code.=<<___;
+	movdqu	(%rdi),%xmm0
+	call	_vpaes_preheat
+	call	_vpaes_decrypt_core
+	movdqu	%xmm0,(%rsi)
+___
+$code.=<<___ if ($win64);
+	movaps	0x10(%rsp),%xmm6
+	movaps	0x20(%rsp),%xmm7
+	movaps	0x30(%rsp),%xmm8
+	movaps	0x40(%rsp),%xmm9
+	movaps	0x50(%rsp),%xmm10
+	movaps	0x60(%rsp),%xmm11
+	movaps	0x70(%rsp),%xmm12
+	movaps	0x80(%rsp),%xmm13
+	movaps	0x90(%rsp),%xmm14
+	movaps	0xa0(%rsp),%xmm15
+	lea	0xb8(%rsp),%rsp
+.Ldec_epilogue:
+___
+$code.=<<___;
+	ret
+.size	${PREFIX}_decrypt,.-${PREFIX}_decrypt
+___
+{
+my ($inp,$out,$len,$key,$ivp,$enc)=("%rdi","%rsi","%rdx","%rcx","%r8","%r9");
+# void AES_cbc_encrypt (const void char *inp, unsigned char *out,
+#                       size_t length, const AES_KEY *key,
+#                       unsigned char *ivp,const int enc);
+$code.=<<___;
+.globl	${PREFIX}_cbc_encrypt
+.type	${PREFIX}_cbc_encrypt,\@function,6
+.align	16
+${PREFIX}_cbc_encrypt:
+	xchg	$key,$len
+___
+($len,$key)=($key,$len);
+$code.=<<___;
+	sub	\$16,$len
+	jc	.Lcbc_abort
+___
+$code.=<<___ if ($win64);
+	lea	-0xb8(%rsp),%rsp
+	movaps	%xmm6,0x10(%rsp)
+	movaps	%xmm7,0x20(%rsp)
+	movaps	%xmm8,0x30(%rsp)
+	movaps	%xmm9,0x40(%rsp)
+	movaps	%xmm10,0x50(%rsp)
+	movaps	%xmm11,0x60(%rsp)
+	movaps	%xmm12,0x70(%rsp)
+	movaps	%xmm13,0x80(%rsp)
+	movaps	%xmm14,0x90(%rsp)
+	movaps	%xmm15,0xa0(%rsp)
+.Lcbc_body:
+___
+$code.=<<___;
+	movdqu	($ivp),%xmm6		# load IV
+	sub	$inp,$out
+	call	_vpaes_preheat
+	cmp	\$0,${enc}d
+	je	.Lcbc_dec_loop
+	jmp	.Lcbc_enc_loop
+.align	16
+.Lcbc_enc_loop:
+	movdqu	($inp),%xmm0
+	pxor	%xmm6,%xmm0
+	call	_vpaes_encrypt_core
+	movdqa	%xmm0,%xmm6
+	movdqu	%xmm0,($out,$inp)
+	lea	16($inp),$inp
+	sub	\$16,$len
+	jnc	.Lcbc_enc_loop
+	jmp	.Lcbc_done
+.align	16
+.Lcbc_dec_loop:
+	movdqu	($inp),%xmm0
+	movdqa	%xmm0,%xmm7
+	call	_vpaes_decrypt_core
+	pxor	%xmm6,%xmm0
+	movdqa	%xmm7,%xmm6
+	movdqu	%xmm0,($out,$inp)
+	lea	16($inp),$inp
+	sub	\$16,$len
+	jnc	.Lcbc_dec_loop
+.Lcbc_done:
+	movdqu	%xmm6,($ivp)		# save IV
+___
+$code.=<<___ if ($win64);
+	movaps	0x10(%rsp),%xmm6
+	movaps	0x20(%rsp),%xmm7
+	movaps	0x30(%rsp),%xmm8
+	movaps	0x40(%rsp),%xmm9
+	movaps	0x50(%rsp),%xmm10
+	movaps	0x60(%rsp),%xmm11
+	movaps	0x70(%rsp),%xmm12
+	movaps	0x80(%rsp),%xmm13
+	movaps	0x90(%rsp),%xmm14
+	movaps	0xa0(%rsp),%xmm15
+	lea	0xb8(%rsp),%rsp
+.Lcbc_epilogue:
+___
+$code.=<<___;
+.Lcbc_abort:
+	ret
+.size	${PREFIX}_cbc_encrypt,.-${PREFIX}_cbc_encrypt
+___
+}
+$code.=<<___;
+##
+##  _aes_preheat
+##
+##  Fills register %r10 -> .aes_consts (so you can -fPIC)
+##  and %xmm9-%xmm15 as specified below.
+##
+.type	_vpaes_preheat,\@abi-omnipotent
+.align	16
+_vpaes_preheat:
+	lea	.Lk_s0F(%rip), %r10
+	movdqa	-0x20(%r10), %xmm10	# .Lk_inv
+	movdqa	-0x10(%r10), %xmm11	# .Lk_inv+16
+	movdqa	0x00(%r10), %xmm9	# .Lk_s0F
+	movdqa	0x30(%r10), %xmm13	# .Lk_sb1
+	movdqa	0x40(%r10), %xmm12	# .Lk_sb1+16
+	movdqa	0x50(%r10), %xmm15	# .Lk_sb2
+	movdqa	0x60(%r10), %xmm14	# .Lk_sb2+16
+	ret
+.size	_vpaes_preheat,.-_vpaes_preheat
+########################################################
+##                                                    ##
+##                     Constants                      ##
+##                                                    ##
+########################################################
+.type	_vpaes_consts,\@object
+.align	64
+_vpaes_consts:
+.Lk_inv:	# inv, inva
+	.quad	0x0E05060F0D080180, 0x040703090A0B0C02
+	.quad	0x01040A060F0B0780, 0x030D0E0C02050809
+
+.Lk_s0F:	# s0F
+	.quad	0x0F0F0F0F0F0F0F0F, 0x0F0F0F0F0F0F0F0F
+
+.Lk_ipt:	# input transform (lo, hi)
+	.quad	0xC2B2E8985A2A7000, 0xCABAE09052227808
+	.quad	0x4C01307D317C4D00, 0xCD80B1FCB0FDCC81
+
+.Lk_sb1:	# sb1u, sb1t
+	.quad	0xB19BE18FCB503E00, 0xA5DF7A6E142AF544
+	.quad	0x3618D415FAE22300, 0x3BF7CCC10D2ED9EF
+.Lk_sb2:	# sb2u, sb2t
+	.quad	0xE27A93C60B712400, 0x5EB7E955BC982FCD
+	.quad	0x69EB88400AE12900, 0xC2A163C8AB82234A
+.Lk_sbo:	# sbou, sbot
+	.quad	0xD0D26D176FBDC700, 0x15AABF7AC502A878
+	.quad	0xCFE474A55FBB6A00, 0x8E1E90D1412B35FA
+
+.Lk_mc_forward:	# mc_forward
+	.quad	0x0407060500030201, 0x0C0F0E0D080B0A09
+	.quad	0x080B0A0904070605, 0x000302010C0F0E0D
+	.quad	0x0C0F0E0D080B0A09, 0x0407060500030201
+	.quad	0x000302010C0F0E0D, 0x080B0A0904070605
+
+.Lk_mc_backward:# mc_backward
+	.quad	0x0605040702010003, 0x0E0D0C0F0A09080B
+	.quad	0x020100030E0D0C0F, 0x0A09080B06050407
+	.quad	0x0E0D0C0F0A09080B, 0x0605040702010003
+	.quad	0x0A09080B06050407, 0x020100030E0D0C0F
+
+.Lk_sr:		# sr
+	.quad	0x0706050403020100, 0x0F0E0D0C0B0A0908
+	.quad	0x030E09040F0A0500, 0x0B06010C07020D08
+	.quad	0x0F060D040B020900, 0x070E050C030A0108
+	.quad	0x0B0E0104070A0D00, 0x0306090C0F020508
+
+.Lk_rcon:	# rcon
+	.quad	0x1F8391B9AF9DEEB6, 0x702A98084D7C7D81
+
+.Lk_s63:	# s63: all equal to 0x63 transformed
+	.quad	0x5B5B5B5B5B5B5B5B, 0x5B5B5B5B5B5B5B5B
+
+.Lk_opt:	# output transform
+	.quad	0xFF9F4929D6B66000, 0xF7974121DEBE6808
+	.quad	0x01EDBD5150BCEC00, 0xE10D5DB1B05C0CE0
+
+.Lk_deskew:	# deskew tables: inverts the sbox's "skew"
+	.quad	0x07E4A34047A4E300, 0x1DFEB95A5DBEF91A
+	.quad	0x5F36B5DC83EA6900, 0x2841C2ABF49D1E77
+
+##
+##  Decryption stuff
+##  Key schedule constants
+##
+.Lk_dksd:	# decryption key schedule: invskew x*D
+	.quad	0xFEB91A5DA3E44700, 0x0740E3A45A1DBEF9
+	.quad	0x41C277F4B5368300, 0x5FDC69EAAB289D1E
+.Lk_dksb:	# decryption key schedule: invskew x*B
+	.quad	0x9A4FCA1F8550D500, 0x03D653861CC94C99
+	.quad	0x115BEDA7B6FC4A00, 0xD993256F7E3482C8
+.Lk_dkse:	# decryption key schedule: invskew x*E + 0x63
+	.quad	0xD5031CCA1FC9D600, 0x53859A4C994F5086
+	.quad	0xA23196054FDC7BE8, 0xCD5EF96A20B31487
+.Lk_dks9:	# decryption key schedule: invskew x*9
+	.quad	0xB6116FC87ED9A700, 0x4AED933482255BFC
+	.quad	0x4576516227143300, 0x8BB89FACE9DAFDCE
+
+##
+##  Decryption stuff
+##  Round function constants
+##
+.Lk_dipt:	# decryption input transform
+	.quad	0x0F505B040B545F00, 0x154A411E114E451A
+	.quad	0x86E383E660056500, 0x12771772F491F194
+
+.Lk_dsb9:	# decryption sbox output *9*u, *9*t
+	.quad	0x851C03539A86D600, 0xCAD51F504F994CC9
+	.quad	0xC03B1789ECD74900, 0x725E2C9EB2FBA565
+.Lk_dsbd:	# decryption sbox output *D*u, *D*t
+	.quad	0x7D57CCDFE6B1A200, 0xF56E9B13882A4439
+	.quad	0x3CE2FAF724C6CB00, 0x2931180D15DEEFD3
+.Lk_dsbb:	# decryption sbox output *B*u, *B*t
+	.quad	0xD022649296B44200, 0x602646F6B0F2D404
+	.quad	0xC19498A6CD596700, 0xF3FF0C3E3255AA6B
+.Lk_dsbe:	# decryption sbox output *E*u, *E*t
+	.quad	0x46F2929626D4D000, 0x2242600464B4F6B0
+	.quad	0x0C55A6CDFFAAC100, 0x9467F36B98593E32
+.Lk_dsbo:	# decryption sbox final output
+	.quad	0x1387EA537EF94000, 0xC7AA6DB9D4943E2D
+	.quad	0x12D7560F93441D00, 0xCA4B8159D8C58E9C
+.asciz	"Vector Permutaion AES for x86_64/SSSE3, Mike Hamburg (Stanford University)"
+.align	64
+.size	_vpaes_consts,.-_vpaes_consts
+___
+
+if ($win64) {
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lin_prologue
+
+	lea	16(%rax),%rsi		# %xmm save area
+	lea	512($context),%rdi	# &context.Xmm6
+	mov	\$20,%ecx		# 10*sizeof(%xmm0)/sizeof(%rax)
+	.long	0xa548f3fc		# cld; rep movsq
+	lea	0xb8(%rax),%rax		# adjust stack pointer
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$`1232/8`,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	se_handler,.-se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_${PREFIX}_set_encrypt_key
+	.rva	.LSEH_end_${PREFIX}_set_encrypt_key
+	.rva	.LSEH_info_${PREFIX}_set_encrypt_key
+
+	.rva	.LSEH_begin_${PREFIX}_set_decrypt_key
+	.rva	.LSEH_end_${PREFIX}_set_decrypt_key
+	.rva	.LSEH_info_${PREFIX}_set_decrypt_key
+
+	.rva	.LSEH_begin_${PREFIX}_encrypt
+	.rva	.LSEH_end_${PREFIX}_encrypt
+	.rva	.LSEH_info_${PREFIX}_encrypt
+
+	.rva	.LSEH_begin_${PREFIX}_decrypt
+	.rva	.LSEH_end_${PREFIX}_decrypt
+	.rva	.LSEH_info_${PREFIX}_decrypt
+
+	.rva	.LSEH_begin_${PREFIX}_cbc_encrypt
+	.rva	.LSEH_end_${PREFIX}_cbc_encrypt
+	.rva	.LSEH_info_${PREFIX}_cbc_encrypt
+
+.section	.xdata
+.align	8
+.LSEH_info_${PREFIX}_set_encrypt_key:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lenc_key_body,.Lenc_key_epilogue	# HandlerData[]
+.LSEH_info_${PREFIX}_set_decrypt_key:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Ldec_key_body,.Ldec_key_epilogue	# HandlerData[]
+.LSEH_info_${PREFIX}_encrypt:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lenc_body,.Lenc_epilogue		# HandlerData[]
+.LSEH_info_${PREFIX}_decrypt:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Ldec_body,.Ldec_epilogue		# HandlerData[]
+.LSEH_info_${PREFIX}_cbc_encrypt:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lcbc_body,.Lcbc_epilogue		# HandlerData[]
+___
+}
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+
+print $code;
+
+close STDOUT;
diff --git a/crypto/arm_arch.h b/crypto/arm_arch.h
new file mode 100644
index 000000000000..5a8310768003
--- /dev/null
+++ b/crypto/arm_arch.h
@@ -0,0 +1,51 @@
+#ifndef __ARM_ARCH_H__
+#define __ARM_ARCH_H__
+
+#if !defined(__ARM_ARCH__)
+# if defined(__CC_ARM)
+#  define __ARM_ARCH__ __TARGET_ARCH_ARM
+#  if defined(__BIG_ENDIAN)
+#   define __ARMEB__
+#  else
+#   define __ARMEL__
+#  endif
+# elif defined(__GNUC__)
+  /*
+   * Why doesn't gcc define __ARM_ARCH__? Instead it defines
+   * bunch of below macros. See all_architectires[] table in
+   * gcc/config/arm/arm.c. On a side note it defines
+   * __ARMEL__/__ARMEB__ for little-/big-endian.
+   */
+#  if	defined(__ARM_ARCH_7__)	|| defined(__ARM_ARCH_7A__)	|| \
+	defined(__ARM_ARCH_7R__)|| defined(__ARM_ARCH_7M__)	|| \
+	defined(__ARM_ARCH_7EM__)
+#   define __ARM_ARCH__ 7
+#  elif	defined(__ARM_ARCH_6__)	|| defined(__ARM_ARCH_6J__)	|| \
+	defined(__ARM_ARCH_6K__)|| defined(__ARM_ARCH_6M__)	|| \
+	defined(__ARM_ARCH_6Z__)|| defined(__ARM_ARCH_6ZK__)	|| \
+	defined(__ARM_ARCH_6T2__)
+#   define __ARM_ARCH__ 6
+#  elif	defined(__ARM_ARCH_5__)	|| defined(__ARM_ARCH_5T__)	|| \
+	defined(__ARM_ARCH_5E__)|| defined(__ARM_ARCH_5TE__)	|| \
+	defined(__ARM_ARCH_5TEJ__)
+#   define __ARM_ARCH__ 5
+#  elif	defined(__ARM_ARCH_4__)	|| defined(__ARM_ARCH_4T__)
+#   define __ARM_ARCH__ 4
+#  else
+#   error "unsupported ARM architecture"
+#  endif
+# endif
+#endif
+
+#ifdef OPENSSL_FIPSCANISTER
+#include <openssl/fipssyms.h>
+#endif
+
+#if !__ASSEMBLER__
+extern unsigned int OPENSSL_armcap_P;
+                                     
+#define ARMV7_NEON      (1<<0)
+#define ARMV7_TICK      (1<<1)
+#endif
+
+#endif
diff --git a/crypto/armcap.c b/crypto/armcap.c
new file mode 100644
index 000000000000..5258d2fbddfa
--- /dev/null
+++ b/crypto/armcap.c
@@ -0,0 +1,80 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <setjmp.h>
+#include <signal.h>
+#include <crypto.h>
+
+#include "arm_arch.h"
+
+unsigned int OPENSSL_armcap_P;
+
+static sigset_t all_masked;
+
+static sigjmp_buf ill_jmp;
+static void ill_handler (int sig) { siglongjmp(ill_jmp,sig); }
+
+/*
+ * Following subroutines could have been inlined, but it's not all
+ * ARM compilers support inline assembler...
+ */
+void _armv7_neon_probe(void);
+unsigned int _armv7_tick(void);
+
+unsigned int OPENSSL_rdtsc(void)
+	{
+	if (OPENSSL_armcap_P|ARMV7_TICK)
+		return _armv7_tick();
+	else
+		return 0;
+	}
+
+#if defined(__GNUC__) && __GNUC__>=2
+void OPENSSL_cpuid_setup(void) __attribute__((constructor));
+#endif
+void OPENSSL_cpuid_setup(void)
+	{
+	char *e;
+	struct sigaction	ill_oact,ill_act;
+	sigset_t		oset;
+	static int trigger=0;
+
+	if (trigger) return;
+	trigger=1;
+ 
+	if ((e=getenv("OPENSSL_armcap")))
+		{
+		OPENSSL_armcap_P=strtoul(e,NULL,0);
+		return;
+		}
+
+	sigfillset(&all_masked);
+	sigdelset(&all_masked,SIGILL);
+	sigdelset(&all_masked,SIGTRAP);
+	sigdelset(&all_masked,SIGFPE);
+	sigdelset(&all_masked,SIGBUS);
+	sigdelset(&all_masked,SIGSEGV);
+
+	OPENSSL_armcap_P = 0;
+
+	memset(&ill_act,0,sizeof(ill_act));
+	ill_act.sa_handler = ill_handler;
+	ill_act.sa_mask    = all_masked;
+
+	sigprocmask(SIG_SETMASK,&ill_act.sa_mask,&oset);
+	sigaction(SIGILL,&ill_act,&ill_oact);
+
+	if (sigsetjmp(ill_jmp,1) == 0)
+		{
+		_armv7_neon_probe();
+		OPENSSL_armcap_P |= ARMV7_NEON;
+		}
+	if (sigsetjmp(ill_jmp,1) == 0)
+		{
+		_armv7_tick();
+		OPENSSL_armcap_P |= ARMV7_TICK;
+		}
+
+	sigaction (SIGILL,&ill_oact,NULL);
+	sigprocmask(SIG_SETMASK,&oset,NULL);
+	}
diff --git a/crypto/armv4cpuid.S b/crypto/armv4cpuid.S
new file mode 100644
index 000000000000..2d618deaa43e
--- /dev/null
+++ b/crypto/armv4cpuid.S
@@ -0,0 +1,154 @@
+#include "arm_arch.h"
+
+.text
+.code	32
+
+.align	5
+.global	_armv7_neon_probe
+.type	_armv7_neon_probe,%function
+_armv7_neon_probe:
+	.word	0xf26ee1fe	@ vorr	q15,q15,q15
+	.word	0xe12fff1e	@ bx	lr
+.size	_armv7_neon_probe,.-_armv7_neon_probe
+
+.global	_armv7_tick
+.type	_armv7_tick,%function
+_armv7_tick:
+	mrc	p15,0,r0,c9,c13,0
+	.word	0xe12fff1e	@ bx	lr
+.size	_armv7_tick,.-_armv7_tick
+
+.global	OPENSSL_atomic_add
+.type	OPENSSL_atomic_add,%function
+OPENSSL_atomic_add:
+#if __ARM_ARCH__>=6
+.Ladd:	ldrex	r2,[r0]
+	add	r3,r2,r1
+	strex	r2,r3,[r0]
+	cmp	r2,#0
+	bne	.Ladd
+	mov	r0,r3
+	.word	0xe12fff1e	@ bx	lr
+#else
+	stmdb	sp!,{r4-r6,lr}
+	ldr	r2,.Lspinlock
+	adr	r3,.Lspinlock
+	mov	r4,r0
+	mov	r5,r1
+	add	r6,r3,r2	@ &spinlock
+	b	.+8
+.Lspin:	bl	sched_yield
+	mov	r0,#-1
+	swp	r0,r0,[r6]
+	cmp	r0,#0
+	bne	.Lspin
+
+	ldr	r2,[r4]
+	add	r2,r2,r5
+	str	r2,[r4]
+	str	r0,[r6]		@ release spinlock
+	ldmia	sp!,{r4-r6,lr}
+	tst	lr,#1
+	moveq	pc,lr
+	.word	0xe12fff1e	@ bx	lr
+#endif
+.size	OPENSSL_atomic_add,.-OPENSSL_atomic_add
+
+.global	OPENSSL_cleanse
+.type	OPENSSL_cleanse,%function
+OPENSSL_cleanse:
+	eor	ip,ip,ip
+	cmp	r1,#7
+	subhs	r1,r1,#4
+	bhs	.Lot
+	cmp	r1,#0
+	beq	.Lcleanse_done
+.Little:
+	strb	ip,[r0],#1
+	subs	r1,r1,#1
+	bhi	.Little
+	b	.Lcleanse_done
+
+.Lot:	tst	r0,#3
+	beq	.Laligned
+	strb	ip,[r0],#1
+	sub	r1,r1,#1
+	b	.Lot
+.Laligned:
+	str	ip,[r0],#4
+	subs	r1,r1,#4
+	bhs	.Laligned
+	adds	r1,r1,#4
+	bne	.Little
+.Lcleanse_done:
+	tst	lr,#1
+	moveq	pc,lr
+	.word	0xe12fff1e	@ bx	lr
+.size	OPENSSL_cleanse,.-OPENSSL_cleanse
+
+.global	OPENSSL_wipe_cpu
+.type	OPENSSL_wipe_cpu,%function
+OPENSSL_wipe_cpu:
+	ldr	r0,.LOPENSSL_armcap
+	adr	r1,.LOPENSSL_armcap
+	ldr	r0,[r1,r0]
+	eor	r2,r2,r2
+	eor	r3,r3,r3
+	eor	ip,ip,ip
+	tst	r0,#1
+	beq	.Lwipe_done
+	.word	0xf3000150	@ veor    q0, q0, q0
+	.word	0xf3022152	@ veor    q1, q1, q1
+	.word	0xf3044154	@ veor    q2, q2, q2
+	.word	0xf3066156	@ veor    q3, q3, q3
+	.word	0xf34001f0	@ veor    q8, q8, q8
+	.word	0xf34221f2	@ veor    q9, q9, q9
+	.word	0xf34441f4	@ veor    q10, q10, q10
+	.word	0xf34661f6	@ veor    q11, q11, q11
+	.word	0xf34881f8	@ veor    q12, q12, q12
+	.word	0xf34aa1fa	@ veor    q13, q13, q13
+	.word	0xf34cc1fc	@ veor    q14, q14, q14
+	.word	0xf34ee1fe	@ veor    q15, q15, q15
+.Lwipe_done:
+	mov	r0,sp
+	tst	lr,#1
+	moveq	pc,lr
+	.word	0xe12fff1e	@ bx	lr
+.size	OPENSSL_wipe_cpu,.-OPENSSL_wipe_cpu
+
+.global	OPENSSL_instrument_bus
+.type	OPENSSL_instrument_bus,%function
+OPENSSL_instrument_bus:
+	eor	r0,r0,r0
+	tst	lr,#1
+	moveq	pc,lr
+	.word	0xe12fff1e	@ bx	lr
+.size	OPENSSL_instrument_bus,.-OPENSSL_instrument_bus
+
+.global	OPENSSL_instrument_bus2
+.type	OPENSSL_instrument_bus2,%function
+OPENSSL_instrument_bus2:
+	eor	r0,r0,r0
+	tst	lr,#1
+	moveq	pc,lr
+	.word	0xe12fff1e	@ bx	lr
+.size	OPENSSL_instrument_bus2,.-OPENSSL_instrument_bus2
+
+.align	5
+.LOPENSSL_armcap:
+.word	OPENSSL_armcap_P-.LOPENSSL_armcap
+#if __ARM_ARCH__>=6
+.align	5
+#else
+.Lspinlock:
+.word	atomic_add_spinlock-.Lspinlock
+.align	5
+
+.data
+.align	2
+atomic_add_spinlock:
+.word	0
+#endif
+
+.comm	OPENSSL_armcap_P,4,4
+.hidden	OPENSSL_armcap_P
diff --git a/crypto/asn1/Makefile b/crypto/asn1/Makefile
index 94a6885804ff..f7787005d454 100644
--- a/crypto/asn1/Makefile
+++ b/crypto/asn1/Makefile
@@ -22,30 +22,32 @@ LIBSRC=	a_object.c a_bitstr.c a_utctm.c a_gentm.c a_time.c a_int.c a_octet.c \
 	a_enum.c a_utf8.c a_sign.c a_digest.c a_verify.c a_mbstr.c a_strex.c \
 	x_algor.c x_val.c x_pubkey.c x_sig.c x_req.c x_attrib.c x_bignum.c \
 	x_long.c x_name.c x_x509.c x_x509a.c x_crl.c x_info.c x_spki.c nsseq.c \
-	d2i_pu.c d2i_pr.c i2d_pu.c i2d_pr.c\
+	x_nx509.c d2i_pu.c d2i_pr.c i2d_pu.c i2d_pr.c\
 	t_req.c t_x509.c t_x509a.c t_crl.c t_pkey.c t_spki.c t_bitst.c \
 	tasn_new.c tasn_fre.c tasn_enc.c tasn_dec.c tasn_utl.c tasn_typ.c \
+	tasn_prn.c ameth_lib.c \
 	f_int.c f_string.c n_pkey.c \
-	f_enum.c a_hdr.c x_pkey.c a_bool.c x_exten.c asn_mime.c \
-	asn1_gen.c asn1_par.c asn1_lib.c asn1_err.c a_meth.c a_bytes.c a_strnid.c \
+	f_enum.c x_pkey.c a_bool.c x_exten.c bio_asn1.c bio_ndef.c asn_mime.c \
+	asn1_gen.c asn1_par.c asn1_lib.c asn1_err.c a_bytes.c a_strnid.c \
 	evp_asn1.c asn_pack.c p5_pbe.c p5_pbev2.c p8_pkey.c asn_moid.c
 LIBOBJ= a_object.o a_bitstr.o a_utctm.o a_gentm.o a_time.o a_int.o a_octet.o \
 	a_print.o a_type.o a_set.o a_dup.o a_d2i_fp.o a_i2d_fp.o \
 	a_enum.o a_utf8.o a_sign.o a_digest.o a_verify.o a_mbstr.o a_strex.o \
 	x_algor.o x_val.o x_pubkey.o x_sig.o x_req.o x_attrib.o x_bignum.o \
 	x_long.o x_name.o x_x509.o x_x509a.o x_crl.o x_info.o x_spki.o nsseq.o \
-	d2i_pu.o d2i_pr.o i2d_pu.o i2d_pr.o \
+	x_nx509.o d2i_pu.o d2i_pr.o i2d_pu.o i2d_pr.o \
 	t_req.o t_x509.o t_x509a.o t_crl.o t_pkey.o t_spki.o t_bitst.o \
 	tasn_new.o tasn_fre.o tasn_enc.o tasn_dec.o tasn_utl.o tasn_typ.o \
+	tasn_prn.o ameth_lib.o \
 	f_int.o f_string.o n_pkey.o \
-	f_enum.o a_hdr.o x_pkey.o a_bool.o x_exten.o asn_mime.o \
-	asn1_gen.o asn1_par.o asn1_lib.o asn1_err.o a_meth.o a_bytes.o a_strnid.o \
+	f_enum.o x_pkey.o a_bool.o x_exten.o bio_asn1.o bio_ndef.o asn_mime.o \
+	asn1_gen.o asn1_par.o asn1_lib.o asn1_err.o a_bytes.o a_strnid.o \
 	evp_asn1.o asn_pack.o p5_pbe.o p5_pbev2.o p8_pkey.o asn_moid.o
 
 SRC= $(LIBSRC)
 
 EXHEADER=  asn1.h asn1_mac.h asn1t.h
-HEADER=	$(EXHEADER)
+HEADER=	$(EXHEADER) asn1_locl.h
 
 ALL=    $(GENERAL) $(SRC) $(HEADER)
 
@@ -63,7 +65,7 @@ pk:	pk.c
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -142,9 +144,9 @@ a_digest.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 a_digest.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 a_digest.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 a_digest.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-a_digest.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-a_digest.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-a_digest.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+a_digest.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+a_digest.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+a_digest.o: ../../include/openssl/opensslconf.h
 a_digest.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 a_digest.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 a_digest.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -173,14 +175,6 @@ a_gentm.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
 a_gentm.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
 a_gentm.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 a_gentm.o: ../cryptlib.h ../o_time.h a_gentm.c
-a_hdr.o: ../../e_os.h ../../include/openssl/asn1.h
-a_hdr.o: ../../include/openssl/asn1_mac.h ../../include/openssl/bio.h
-a_hdr.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-a_hdr.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-a_hdr.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
-a_hdr.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-a_hdr.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-a_hdr.o: ../../include/openssl/symhacks.h ../cryptlib.h a_hdr.c
 a_i2d_fp.o: ../../e_os.h ../../include/openssl/asn1.h
 a_i2d_fp.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 a_i2d_fp.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
@@ -205,13 +199,6 @@ a_mbstr.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
 a_mbstr.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
 a_mbstr.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 a_mbstr.o: ../cryptlib.h a_mbstr.c
-a_meth.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
-a_meth.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-a_meth.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-a_meth.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
-a_meth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-a_meth.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-a_meth.o: ../../include/openssl/symhacks.h ../cryptlib.h a_meth.c
 a_object.o: ../../e_os.h ../../include/openssl/asn1.h
 a_object.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 a_object.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
@@ -250,27 +237,27 @@ a_sign.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
 a_sign.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 a_sign.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 a_sign.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-a_sign.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-a_sign.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-a_sign.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-a_sign.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-a_sign.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-a_sign.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-a_sign.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-a_sign.o: ../../include/openssl/x509_vfy.h ../cryptlib.h a_sign.c
+a_sign.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+a_sign.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+a_sign.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+a_sign.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+a_sign.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+a_sign.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+a_sign.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+a_sign.o: ../cryptlib.h a_sign.c asn1_locl.h
 a_strex.o: ../../e_os.h ../../include/openssl/asn1.h
 a_strex.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 a_strex.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 a_strex.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 a_strex.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-a_strex.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-a_strex.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-a_strex.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-a_strex.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-a_strex.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-a_strex.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-a_strex.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-a_strex.o: ../../include/openssl/x509_vfy.h ../cryptlib.h a_strex.c charmap.h
+a_strex.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+a_strex.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+a_strex.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+a_strex.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+a_strex.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+a_strex.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+a_strex.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+a_strex.o: ../cryptlib.h a_strex.c charmap.h
 a_strnid.o: ../../e_os.h ../../include/openssl/asn1.h
 a_strnid.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 a_strnid.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
@@ -318,14 +305,29 @@ a_verify.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 a_verify.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 a_verify.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 a_verify.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-a_verify.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-a_verify.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-a_verify.o: ../../include/openssl/opensslconf.h
+a_verify.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+a_verify.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 a_verify.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 a_verify.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 a_verify.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 a_verify.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 a_verify.o: ../../include/openssl/x509_vfy.h ../cryptlib.h a_verify.c
+a_verify.o: asn1_locl.h
+ameth_lib.o: ../../e_os.h ../../include/openssl/asn1.h
+ameth_lib.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+ameth_lib.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+ameth_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ameth_lib.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ameth_lib.o: ../../include/openssl/engine.h ../../include/openssl/err.h
+ameth_lib.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ameth_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ameth_lib.o: ../../include/openssl/opensslconf.h
+ameth_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ameth_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+ameth_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ameth_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+ameth_lib.o: ../../include/openssl/x509_vfy.h ../cryptlib.h ameth_lib.c
+ameth_lib.o: asn1_locl.h
 asn1_err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 asn1_err.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 asn1_err.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
@@ -339,9 +341,8 @@ asn1_gen.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 asn1_gen.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 asn1_gen.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 asn1_gen.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-asn1_gen.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-asn1_gen.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-asn1_gen.o: ../../include/openssl/opensslconf.h
+asn1_gen.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+asn1_gen.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 asn1_gen.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 asn1_gen.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 asn1_gen.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -371,24 +372,23 @@ asn_mime.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 asn_mime.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 asn_mime.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 asn_mime.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-asn_mime.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-asn_mime.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-asn_mime.o: ../../include/openssl/opensslconf.h
+asn_mime.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+asn_mime.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 asn_mime.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 asn_mime.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
 asn_mime.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 asn_mime.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 asn_mime.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-asn_mime.o: ../cryptlib.h asn_mime.c
+asn_mime.o: ../cryptlib.h asn1_locl.h asn_mime.c
 asn_moid.o: ../../e_os.h ../../include/openssl/asn1.h
 asn_moid.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 asn_moid.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 asn_moid.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
 asn_moid.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 asn_moid.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-asn_moid.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-asn_moid.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-asn_moid.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+asn_moid.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+asn_moid.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+asn_moid.o: ../../include/openssl/opensslconf.h
 asn_moid.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 asn_moid.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 asn_moid.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -402,28 +402,43 @@ asn_pack.o: ../../include/openssl/opensslconf.h
 asn_pack.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 asn_pack.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 asn_pack.o: ../../include/openssl/symhacks.h ../cryptlib.h asn_pack.c
+bio_asn1.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+bio_asn1.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+bio_asn1.o: ../../include/openssl/opensslconf.h
+bio_asn1.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+bio_asn1.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+bio_asn1.o: ../../include/openssl/symhacks.h bio_asn1.c
+bio_ndef.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
+bio_ndef.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
+bio_ndef.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
+bio_ndef.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
+bio_ndef.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+bio_ndef.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+bio_ndef.o: ../../include/openssl/symhacks.h bio_ndef.c
 d2i_pr.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 d2i_pr.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
-d2i_pr.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
-d2i_pr.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+d2i_pr.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+d2i_pr.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+d2i_pr.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 d2i_pr.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-d2i_pr.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-d2i_pr.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-d2i_pr.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-d2i_pr.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rsa.h
-d2i_pr.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-d2i_pr.o: ../../include/openssl/symhacks.h ../cryptlib.h d2i_pr.c
+d2i_pr.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+d2i_pr.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+d2i_pr.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+d2i_pr.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+d2i_pr.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+d2i_pr.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+d2i_pr.o: ../../include/openssl/x509_vfy.h ../cryptlib.h asn1_locl.h d2i_pr.c
 d2i_pu.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 d2i_pu.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
 d2i_pu.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
 d2i_pu.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 d2i_pu.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-d2i_pu.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-d2i_pu.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-d2i_pu.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-d2i_pu.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rsa.h
-d2i_pu.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-d2i_pu.o: ../../include/openssl/symhacks.h ../cryptlib.h d2i_pu.c
+d2i_pu.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+d2i_pu.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+d2i_pu.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+d2i_pu.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+d2i_pu.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+d2i_pu.o: ../cryptlib.h d2i_pu.c
 evp_asn1.o: ../../e_os.h ../../include/openssl/asn1.h
 evp_asn1.o: ../../include/openssl/asn1_mac.h ../../include/openssl/bio.h
 evp_asn1.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
@@ -455,77 +470,76 @@ f_string.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 f_string.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 f_string.o: ../../include/openssl/symhacks.h ../cryptlib.h f_string.c
 i2d_pr.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
-i2d_pr.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
-i2d_pr.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
+i2d_pr.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 i2d_pr.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+i2d_pr.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 i2d_pr.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-i2d_pr.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-i2d_pr.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-i2d_pr.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-i2d_pr.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rsa.h
-i2d_pr.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-i2d_pr.o: ../../include/openssl/symhacks.h ../cryptlib.h i2d_pr.c
+i2d_pr.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+i2d_pr.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+i2d_pr.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+i2d_pr.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+i2d_pr.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+i2d_pr.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+i2d_pr.o: ../../include/openssl/x509_vfy.h ../cryptlib.h asn1_locl.h i2d_pr.c
 i2d_pu.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 i2d_pu.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
 i2d_pu.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
 i2d_pu.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 i2d_pu.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-i2d_pu.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-i2d_pu.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-i2d_pu.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-i2d_pu.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rsa.h
-i2d_pu.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-i2d_pu.o: ../../include/openssl/symhacks.h ../cryptlib.h i2d_pu.c
+i2d_pu.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+i2d_pu.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+i2d_pu.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+i2d_pu.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+i2d_pu.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+i2d_pu.o: ../cryptlib.h i2d_pu.c
 n_pkey.o: ../../e_os.h ../../include/openssl/asn1.h
 n_pkey.o: ../../include/openssl/asn1_mac.h ../../include/openssl/asn1t.h
 n_pkey.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 n_pkey.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 n_pkey.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 n_pkey.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-n_pkey.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-n_pkey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-n_pkey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-n_pkey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-n_pkey.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
-n_pkey.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-n_pkey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-n_pkey.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-n_pkey.o: ../cryptlib.h n_pkey.c
+n_pkey.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+n_pkey.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+n_pkey.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+n_pkey.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+n_pkey.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+n_pkey.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+n_pkey.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+n_pkey.o: ../../include/openssl/x509_vfy.h ../cryptlib.h n_pkey.c
 nsseq.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
 nsseq.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 nsseq.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 nsseq.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 nsseq.o: ../../include/openssl/ecdsa.h ../../include/openssl/evp.h
-nsseq.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-nsseq.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-nsseq.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-nsseq.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-nsseq.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-nsseq.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-nsseq.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h nsseq.c
+nsseq.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+nsseq.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+nsseq.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+nsseq.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+nsseq.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+nsseq.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+nsseq.o: ../../include/openssl/x509_vfy.h nsseq.c
 p5_pbe.o: ../../e_os.h ../../include/openssl/asn1.h
 p5_pbe.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 p5_pbe.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p5_pbe.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p5_pbe.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p5_pbe.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p5_pbe.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p5_pbe.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p5_pbe.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-p5_pbe.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-p5_pbe.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-p5_pbe.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-p5_pbe.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-p5_pbe.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p5_pbe.c
+p5_pbe.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p5_pbe.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p5_pbe.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+p5_pbe.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+p5_pbe.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+p5_pbe.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+p5_pbe.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+p5_pbe.o: ../cryptlib.h p5_pbe.c
 p5_pbev2.o: ../../e_os.h ../../include/openssl/asn1.h
 p5_pbev2.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 p5_pbev2.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p5_pbev2.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p5_pbev2.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p5_pbev2.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p5_pbev2.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p5_pbev2.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p5_pbev2.o: ../../include/openssl/opensslconf.h
+p5_pbev2.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p5_pbev2.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 p5_pbev2.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p5_pbev2.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
 p5_pbev2.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -538,51 +552,48 @@ p8_pkey.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p8_pkey.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p8_pkey.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p8_pkey.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p8_pkey.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p8_pkey.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p8_pkey.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-p8_pkey.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-p8_pkey.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p8_pkey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p8_pkey.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p8_pkey.o: ../cryptlib.h p8_pkey.c
+p8_pkey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p8_pkey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p8_pkey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+p8_pkey.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+p8_pkey.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p8_pkey.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p8_pkey.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p8_pkey.c
 t_bitst.o: ../../e_os.h ../../include/openssl/asn1.h
 t_bitst.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 t_bitst.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 t_bitst.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 t_bitst.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 t_bitst.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-t_bitst.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-t_bitst.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-t_bitst.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-t_bitst.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-t_bitst.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-t_bitst.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-t_bitst.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-t_bitst.o: ../../include/openssl/x509v3.h ../cryptlib.h t_bitst.c
+t_bitst.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+t_bitst.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+t_bitst.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+t_bitst.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+t_bitst.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+t_bitst.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+t_bitst.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+t_bitst.o: ../cryptlib.h t_bitst.c
 t_crl.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 t_crl.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
 t_crl.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 t_crl.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 t_crl.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 t_crl.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-t_crl.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-t_crl.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-t_crl.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-t_crl.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-t_crl.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-t_crl.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-t_crl.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-t_crl.o: ../../include/openssl/x509v3.h ../cryptlib.h t_crl.c
+t_crl.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+t_crl.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+t_crl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+t_crl.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+t_crl.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+t_crl.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+t_crl.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+t_crl.o: ../cryptlib.h t_crl.c
 t_pkey.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 t_pkey.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
-t_pkey.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
-t_pkey.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
-t_pkey.o: ../../include/openssl/ec.h ../../include/openssl/err.h
-t_pkey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-t_pkey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-t_pkey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-t_pkey.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+t_pkey.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+t_pkey.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
+t_pkey.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+t_pkey.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+t_pkey.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
 t_pkey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 t_pkey.o: ../cryptlib.h t_pkey.c
 t_req.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
@@ -591,57 +602,57 @@ t_req.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 t_req.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 t_req.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 t_req.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-t_req.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-t_req.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-t_req.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-t_req.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-t_req.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
-t_req.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-t_req.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-t_req.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-t_req.o: ../../include/openssl/x509v3.h ../cryptlib.h t_req.c
+t_req.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+t_req.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+t_req.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+t_req.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+t_req.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+t_req.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+t_req.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+t_req.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+t_req.o: ../cryptlib.h t_req.c
 t_spki.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 t_spki.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
 t_spki.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
 t_spki.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 t_spki.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 t_spki.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-t_spki.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-t_spki.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-t_spki.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-t_spki.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-t_spki.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
-t_spki.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-t_spki.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-t_spki.o: ../../include/openssl/x509_vfy.h ../cryptlib.h t_spki.c
+t_spki.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+t_spki.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+t_spki.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+t_spki.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+t_spki.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+t_spki.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+t_spki.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+t_spki.o: ../cryptlib.h t_spki.c
 t_x509.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 t_x509.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
 t_x509.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 t_x509.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 t_x509.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 t_x509.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-t_x509.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-t_x509.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-t_x509.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-t_x509.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-t_x509.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
-t_x509.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-t_x509.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-t_x509.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-t_x509.o: ../../include/openssl/x509v3.h ../cryptlib.h t_x509.c
+t_x509.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+t_x509.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+t_x509.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+t_x509.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+t_x509.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+t_x509.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+t_x509.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+t_x509.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+t_x509.o: ../cryptlib.h asn1_locl.h t_x509.c
 t_x509a.o: ../../e_os.h ../../include/openssl/asn1.h
 t_x509a.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 t_x509a.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 t_x509a.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 t_x509a.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-t_x509a.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-t_x509a.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-t_x509a.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-t_x509a.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-t_x509a.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-t_x509a.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-t_x509a.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-t_x509a.o: ../../include/openssl/x509_vfy.h ../cryptlib.h t_x509a.c
+t_x509a.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+t_x509a.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+t_x509a.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+t_x509a.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+t_x509a.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+t_x509a.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+t_x509a.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+t_x509a.o: ../cryptlib.h t_x509a.c
 tasn_dec.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
 tasn_dec.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 tasn_dec.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
@@ -675,6 +686,21 @@ tasn_new.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 tasn_new.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 tasn_new.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 tasn_new.o: ../../include/openssl/symhacks.h tasn_new.c
+tasn_prn.o: ../../e_os.h ../../include/openssl/asn1.h
+tasn_prn.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+tasn_prn.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
+tasn_prn.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+tasn_prn.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+tasn_prn.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+tasn_prn.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+tasn_prn.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+tasn_prn.o: ../../include/openssl/opensslconf.h
+tasn_prn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+tasn_prn.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+tasn_prn.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+tasn_prn.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+tasn_prn.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+tasn_prn.o: ../cryptlib.h asn1_locl.h tasn_prn.c
 tasn_typ.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
 tasn_typ.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
 tasn_typ.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
@@ -694,23 +720,21 @@ x_algor.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x_algor.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x_algor.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x_algor.o: ../../include/openssl/ecdsa.h ../../include/openssl/evp.h
-x_algor.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_algor.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_algor.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_algor.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_algor.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_algor.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_algor.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_algor.o: x_algor.c
+x_algor.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_algor.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_algor.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_algor.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_algor.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_algor.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_algor.o: ../../include/openssl/x509_vfy.h x_algor.c
 x_attrib.o: ../../e_os.h ../../include/openssl/asn1.h
 x_attrib.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 x_attrib.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_attrib.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x_attrib.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x_attrib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_attrib.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_attrib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_attrib.o: ../../include/openssl/opensslconf.h
+x_attrib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_attrib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 x_attrib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x_attrib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x_attrib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -727,44 +751,42 @@ x_bignum.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 x_bignum.o: ../../include/openssl/symhacks.h ../cryptlib.h x_bignum.c
 x_crl.o: ../../e_os.h ../../include/openssl/asn1.h
 x_crl.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
-x_crl.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-x_crl.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
-x_crl.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-x_crl.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_crl.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+x_crl.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
+x_crl.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+x_crl.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+x_crl.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+x_crl.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
 x_crl.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
 x_crl.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
 x_crl.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
 x_crl.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 x_crl.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 x_crl.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_crl.o: ../cryptlib.h x_crl.c
+x_crl.o: ../../include/openssl/x509v3.h ../cryptlib.h asn1_locl.h x_crl.c
 x_exten.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
 x_exten.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x_exten.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x_exten.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x_exten.o: ../../include/openssl/ecdsa.h ../../include/openssl/evp.h
-x_exten.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_exten.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_exten.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_exten.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_exten.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_exten.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_exten.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_exten.o: x_exten.c
+x_exten.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_exten.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_exten.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_exten.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_exten.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_exten.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_exten.o: ../../include/openssl/x509_vfy.h x_exten.c
 x_info.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 x_info.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_info.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x_info.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x_info.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_info.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_info.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_info.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_info.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_info.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_info.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_info.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_info.o: ../cryptlib.h x_info.c
+x_info.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_info.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_info.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_info.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_info.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_info.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_info.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x_info.c
 x_long.o: ../../e_os.h ../../include/openssl/asn1.h
 x_long.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 x_long.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
@@ -780,125 +802,129 @@ x_name.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_name.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x_name.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x_name.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_name.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_name.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_name.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_name.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_name.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_name.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_name.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_name.o: ../cryptlib.h x_name.c
+x_name.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_name.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_name.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_name.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_name.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_name.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_name.o: ../../include/openssl/x509_vfy.h ../cryptlib.h asn1_locl.h x_name.c
+x_nx509.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
+x_nx509.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+x_nx509.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+x_nx509.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+x_nx509.o: ../../include/openssl/ecdsa.h ../../include/openssl/evp.h
+x_nx509.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_nx509.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_nx509.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_nx509.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_nx509.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_nx509.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_nx509.o: ../../include/openssl/x509_vfy.h x_nx509.c
 x_pkey.o: ../../e_os.h ../../include/openssl/asn1.h
 x_pkey.o: ../../include/openssl/asn1_mac.h ../../include/openssl/bio.h
 x_pkey.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_pkey.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x_pkey.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x_pkey.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_pkey.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_pkey.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_pkey.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_pkey.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_pkey.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_pkey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_pkey.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_pkey.o: ../cryptlib.h x_pkey.c
+x_pkey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_pkey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_pkey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_pkey.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_pkey.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_pkey.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_pkey.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x_pkey.c
 x_pubkey.o: ../../e_os.h ../../include/openssl/asn1.h
 x_pubkey.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 x_pubkey.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_pubkey.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 x_pubkey.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x_pubkey.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x_pubkey.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x_pubkey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x_pubkey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_pubkey.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x_pubkey.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x_pubkey.o: ../../include/openssl/opensslconf.h
 x_pubkey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x_pubkey.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
 x_pubkey.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 x_pubkey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 x_pubkey.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_pubkey.o: ../cryptlib.h x_pubkey.c
+x_pubkey.o: ../cryptlib.h asn1_locl.h x_pubkey.c
 x_req.o: ../../e_os.h ../../include/openssl/asn1.h
 x_req.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 x_req.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_req.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x_req.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x_req.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_req.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_req.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_req.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_req.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_req.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_req.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_req.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_req.o: ../cryptlib.h x_req.c
+x_req.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_req.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_req.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_req.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_req.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_req.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_req.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x_req.c
 x_sig.o: ../../e_os.h ../../include/openssl/asn1.h
 x_sig.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 x_sig.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_sig.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x_sig.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x_sig.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_sig.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_sig.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_sig.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_sig.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_sig.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_sig.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_sig.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_sig.o: ../cryptlib.h x_sig.c
+x_sig.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_sig.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_sig.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_sig.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_sig.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_sig.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_sig.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x_sig.c
 x_spki.o: ../../e_os.h ../../include/openssl/asn1.h
 x_spki.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 x_spki.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_spki.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x_spki.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x_spki.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_spki.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_spki.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_spki.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_spki.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_spki.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_spki.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_spki.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_spki.o: ../cryptlib.h x_spki.c
+x_spki.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_spki.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_spki.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_spki.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_spki.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_spki.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_spki.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x_spki.c
 x_val.o: ../../e_os.h ../../include/openssl/asn1.h
 x_val.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 x_val.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_val.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x_val.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x_val.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_val.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_val.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_val.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_val.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_val.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_val.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_val.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_val.o: ../cryptlib.h x_val.c
+x_val.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_val.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_val.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_val.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_val.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_val.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_val.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x_val.c
 x_x509.o: ../../e_os.h ../../include/openssl/asn1.h
 x_x509.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 x_x509.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 x_x509.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x_x509.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x_x509.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x_x509.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x_x509.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x_x509.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-x_x509.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-x_x509.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-x_x509.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-x_x509.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-x_x509.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-x_x509.o: ../cryptlib.h x_x509.c
+x_x509.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x_x509.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x_x509.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+x_x509.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+x_x509.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+x_x509.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+x_x509.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+x_x509.o: ../../include/openssl/x509v3.h ../cryptlib.h x_x509.c
 x_x509a.o: ../../e_os.h ../../include/openssl/asn1.h
 x_x509a.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 x_x509a.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_x509a.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x_x509a.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x_x509a.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x_x509a.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x_x509a.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x_x509a.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x_x509a.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x_x509a.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_x509a.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_x509a.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_x509a.o: ../cryptlib.h x_x509a.c
+x_x509a.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x_x509a.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x_x509a.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x_x509a.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x_x509a.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_x509a.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_x509a.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x_x509a.c
diff --git a/crypto/asn1/a_bitstr.c b/crypto/asn1/a_bitstr.c
index 0fb9ce0c2aea..34179960b87a 100644
--- a/crypto/asn1/a_bitstr.c
+++ b/crypto/asn1/a_bitstr.c
@@ -223,3 +223,26 @@ int ASN1_BIT_STRING_get_bit(ASN1_BIT_STRING *a, int n)
 	return((a->data[w]&v) != 0);
 	}
 
+/*
+ * Checks if the given bit string contains only bits specified by 
+ * the flags vector. Returns 0 if there is at least one bit set in 'a'
+ * which is not specified in 'flags', 1 otherwise.
+ * 'len' is the length of 'flags'.
+ */
+int ASN1_BIT_STRING_check(ASN1_BIT_STRING *a,
+			  unsigned char *flags, int flags_len)
+	{
+	int i, ok;
+	/* Check if there is one bit set at all. */
+	if (!a || !a->data) return 1;
+
+	/* Check each byte of the internal representation of the bit string. */
+	ok = 1;
+	for (i = 0; i < a->length && ok; ++i)
+		{
+		unsigned char mask = i < flags_len ? ~flags[i] : 0xff;
+		/* We are done if there is an unneeded bit set. */
+		ok = (a->data[i] & mask) == 0;
+		}
+	return ok;
+	}
diff --git a/crypto/asn1/a_digest.c b/crypto/asn1/a_digest.c
index d00d9e22b188..cbdeea6ac07d 100644
--- a/crypto/asn1/a_digest.c
+++ b/crypto/asn1/a_digest.c
@@ -87,7 +87,8 @@ int ASN1_digest(i2d_of_void *i2d, const EVP_MD *type, char *data,
 	p=str;
 	i2d(data,&p);
 
-	EVP_Digest(str, i, md, len, type, NULL);
+	if (!EVP_Digest(str, i, md, len, type, NULL))
+		return 0;
 	OPENSSL_free(str);
 	return(1);
 	}
@@ -104,7 +105,8 @@ int ASN1_item_digest(const ASN1_ITEM *it, const EVP_MD *type, void *asn,
 	i=ASN1_item_i2d(asn,&str, it);
 	if (!str) return(0);
 
-	EVP_Digest(str, i, md, len, type, NULL);
+	if (!EVP_Digest(str, i, md, len, type, NULL))
+		return 0;
 	OPENSSL_free(str);
 	return(1);
 	}
diff --git a/crypto/asn1/a_dup.c b/crypto/asn1/a_dup.c
index 199d50f521cd..d98992548ab1 100644
--- a/crypto/asn1/a_dup.c
+++ b/crypto/asn1/a_dup.c
@@ -62,7 +62,7 @@
 
 #ifndef NO_OLD_ASN1
 
-void *ASN1_dup(i2d_of_void *i2d, d2i_of_void *d2i, char *x)
+void *ASN1_dup(i2d_of_void *i2d, d2i_of_void *d2i, void *x)
 	{
 	unsigned char *b,*p;
 	const unsigned char *p2;
diff --git a/crypto/asn1/a_gentm.c b/crypto/asn1/a_gentm.c
index def79062a57d..c79c6f538cdc 100644
--- a/crypto/asn1/a_gentm.c
+++ b/crypto/asn1/a_gentm.c
@@ -117,8 +117,8 @@ err:
 
 int ASN1_GENERALIZEDTIME_check(ASN1_GENERALIZEDTIME *d)
 	{
-	static int min[9]={ 0, 0, 1, 1, 0, 0, 0, 0, 0};
-	static int max[9]={99, 99,12,31,23,59,59,12,59};
+	static const int min[9]={ 0, 0, 1, 1, 0, 0, 0, 0, 0};
+	static const int max[9]={99, 99,12,31,23,59,59,12,59};
 	char *a;
 	int n,i,l,o;
 
@@ -176,6 +176,11 @@ int ASN1_GENERALIZEDTIME_check(ASN1_GENERALIZEDTIME *d)
 			o++;
 			}
 		}
+	else
+		{
+		/* Missing time zone information. */
+		goto err;
+		}
 	return(o == l);
 err:
 	return(0);
@@ -206,6 +211,12 @@ int ASN1_GENERALIZEDTIME_set_string(ASN1_GENERALIZEDTIME *s, const char *str)
 ASN1_GENERALIZEDTIME *ASN1_GENERALIZEDTIME_set(ASN1_GENERALIZEDTIME *s,
 	     time_t t)
 	{
+		return ASN1_GENERALIZEDTIME_adj(s, t, 0, 0);
+	}
+
+ASN1_GENERALIZEDTIME *ASN1_GENERALIZEDTIME_adj(ASN1_GENERALIZEDTIME *s,
+	     time_t t, int offset_day, long offset_sec)
+	{
 	char *p;
 	struct tm *ts;
 	struct tm data;
@@ -220,13 +231,19 @@ ASN1_GENERALIZEDTIME *ASN1_GENERALIZEDTIME_set(ASN1_GENERALIZEDTIME *s,
 	if (ts == NULL)
 		return(NULL);
 
+	if (offset_day || offset_sec)
+		{ 
+		if (!OPENSSL_gmtime_adj(ts, offset_day, offset_sec))
+			return NULL;
+		}
+
 	p=(char *)s->data;
 	if ((p == NULL) || ((size_t)s->length < len))
 		{
 		p=OPENSSL_malloc(len);
 		if (p == NULL)
 			{
-			ASN1err(ASN1_F_ASN1_GENERALIZEDTIME_SET,
+			ASN1err(ASN1_F_ASN1_GENERALIZEDTIME_ADJ,
 				ERR_R_MALLOC_FAILURE);
 			return(NULL);
 			}
diff --git a/crypto/asn1/a_hdr.c b/crypto/asn1/a_hdr.c
deleted file mode 100644
index d1c2a7b9e3a0..000000000000
--- a/crypto/asn1/a_hdr.c
+++ /dev/null
@@ -1,119 +0,0 @@
-/* crypto/asn1/a_hdr.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#include <stdio.h>
-#include "cryptlib.h"
-#include <openssl/asn1_mac.h>
-#include <openssl/asn1.h>
-
-int i2d_ASN1_HEADER(ASN1_HEADER *a, unsigned char **pp)
-	{
-	M_ASN1_I2D_vars(a);
-
-	M_ASN1_I2D_len(a->header,	i2d_ASN1_OCTET_STRING);
-	M_ASN1_I2D_len(a->data,		a->meth->i2d);
-
-	M_ASN1_I2D_seq_total();
-
-	M_ASN1_I2D_put(a->header,	i2d_ASN1_OCTET_STRING);
-	M_ASN1_I2D_put(a->data,		a->meth->i2d);
-
-	M_ASN1_I2D_finish();
-	}
-
-ASN1_HEADER *d2i_ASN1_HEADER(ASN1_HEADER **a, const unsigned char **pp,
-	     long length)
-	{
-	M_ASN1_D2I_vars(a,ASN1_HEADER *,ASN1_HEADER_new);
-
-	M_ASN1_D2I_Init();
-        M_ASN1_D2I_start_sequence();
-        M_ASN1_D2I_get_x(ASN1_OCTET_STRING,ret->header,d2i_ASN1_OCTET_STRING);
-	if (ret->meth != NULL)
-		{
-		M_ASN1_D2I_get_x(void,ret->data,ret->meth->d2i);
-		}
-	else
-		{
-		if (a != NULL) (*a)=ret;
-		return(ret);
-		}
-        M_ASN1_D2I_Finish(a,ASN1_HEADER_free,ASN1_F_D2I_ASN1_HEADER);
-	}
-
-ASN1_HEADER *ASN1_HEADER_new(void)
-	{
-	ASN1_HEADER *ret=NULL;
-	ASN1_CTX c;
-
-	M_ASN1_New_Malloc(ret,ASN1_HEADER);
-	M_ASN1_New(ret->header,M_ASN1_OCTET_STRING_new);
-	ret->meth=NULL;
-	ret->data=NULL;
-	return(ret);
-        M_ASN1_New_Error(ASN1_F_ASN1_HEADER_NEW);
-	}
-
-void ASN1_HEADER_free(ASN1_HEADER *a)
-	{
-	if (a == NULL) return;
-	M_ASN1_OCTET_STRING_free(a->header);
-	if (a->meth != NULL)
-		a->meth->destroy(a->data);
-	OPENSSL_free(a);
-	}
diff --git a/crypto/asn1/a_int.c b/crypto/asn1/a_int.c
index f551bdbaadd3..ad0d2506f632 100644
--- a/crypto/asn1/a_int.c
+++ b/crypto/asn1/a_int.c
@@ -61,10 +61,10 @@
 #include <openssl/asn1.h>
 #include <openssl/bn.h>
 
-ASN1_INTEGER *ASN1_INTEGER_dup(ASN1_INTEGER *x)
+ASN1_INTEGER *ASN1_INTEGER_dup(const ASN1_INTEGER *x)
 { return M_ASN1_INTEGER_dup(x);}
 
-int ASN1_INTEGER_cmp(ASN1_INTEGER *x, ASN1_INTEGER *y)
+int ASN1_INTEGER_cmp(const ASN1_INTEGER *x, const ASN1_INTEGER *y)
 	{ 
 	int neg, ret;
 	/* Compare signs */
@@ -372,7 +372,7 @@ int ASN1_INTEGER_set(ASN1_INTEGER *a, long v)
 	return(1);
 	}
 
-long ASN1_INTEGER_get(ASN1_INTEGER *a)
+long ASN1_INTEGER_get(const ASN1_INTEGER *a)
 	{
 	int neg=0,i;
 	long r=0;
@@ -386,8 +386,8 @@ long ASN1_INTEGER_get(ASN1_INTEGER *a)
 	
 	if (a->length > (int)sizeof(long))
 		{
-		/* hmm... a bit ugly */
-		return(0xffffffffL);
+		/* hmm... a bit ugly, return all ones */
+		return -1;
 		}
 	if (a->data == NULL)
 		return 0;
@@ -401,7 +401,7 @@ long ASN1_INTEGER_get(ASN1_INTEGER *a)
 	return(r);
 	}
 
-ASN1_INTEGER *BN_to_ASN1_INTEGER(BIGNUM *bn, ASN1_INTEGER *ai)
+ASN1_INTEGER *BN_to_ASN1_INTEGER(const BIGNUM *bn, ASN1_INTEGER *ai)
 	{
 	ASN1_INTEGER *ret;
 	int len,j;
@@ -443,7 +443,7 @@ err:
 	return(NULL);
 	}
 
-BIGNUM *ASN1_INTEGER_to_BN(ASN1_INTEGER *ai, BIGNUM *bn)
+BIGNUM *ASN1_INTEGER_to_BN(const ASN1_INTEGER *ai, BIGNUM *bn)
 	{
 	BIGNUM *ret;
 
diff --git a/crypto/asn1/a_meth.c b/crypto/asn1/a_meth.c
deleted file mode 100644
index 50bea917e38b..000000000000
--- a/crypto/asn1/a_meth.c
+++ /dev/null
@@ -1,84 +0,0 @@
-/* crypto/asn1/a_meth.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#include <stdio.h>
-#include "cryptlib.h"
-#include <openssl/buffer.h>
-#include <openssl/asn1.h>
-
-static  ASN1_METHOD ia5string_meth={
-	(I2D_OF(void))	i2d_ASN1_IA5STRING,
-	(D2I_OF(void))	d2i_ASN1_IA5STRING,
-	(void *(*)(void))ASN1_STRING_new,
-	(void (*)(void *))ASN1_STRING_free};
-
-static  ASN1_METHOD bit_string_meth={
-	(I2D_OF(void))	i2d_ASN1_BIT_STRING,
-	(D2I_OF(void))	d2i_ASN1_BIT_STRING,
-	(void *(*)(void))ASN1_STRING_new,
-	(void (*)(void *))ASN1_STRING_free};
-
-ASN1_METHOD *ASN1_IA5STRING_asn1_meth(void)
-	{
-	return(&ia5string_meth);
-	}
-
-ASN1_METHOD *ASN1_BIT_STRING_asn1_meth(void)
-	{
-	return(&bit_string_meth);
-	}
diff --git a/crypto/asn1/a_object.c b/crypto/asn1/a_object.c
index 3ac2bc2a01a1..3978c9150d8c 100644
--- a/crypto/asn1/a_object.c
+++ b/crypto/asn1/a_object.c
@@ -281,8 +281,6 @@ ASN1_OBJECT *d2i_ASN1_OBJECT(ASN1_OBJECT **a, const unsigned char **pp,
 	return ret;
 err:
 	ASN1err(ASN1_F_D2I_ASN1_OBJECT,i);
-	if ((ret != NULL) && ((a == NULL) || (*a != ret)))
-		ASN1_OBJECT_free(ret);
 	return(NULL);
 }
 ASN1_OBJECT *c2i_ASN1_OBJECT(ASN1_OBJECT **a, const unsigned char **pp,
@@ -290,6 +288,7 @@ ASN1_OBJECT *c2i_ASN1_OBJECT(ASN1_OBJECT **a, const unsigned char **pp,
 	{
 	ASN1_OBJECT *ret=NULL;
 	const unsigned char *p;
+	unsigned char *data;
 	int i;
 	/* Sanity check OID encoding: can't have leading 0x80 in
 	 * subidentifiers, see: X.690 8.19.2
@@ -313,15 +312,22 @@ ASN1_OBJECT *c2i_ASN1_OBJECT(ASN1_OBJECT **a, const unsigned char **pp,
 	else	ret=(*a);
 
 	p= *pp;
-	if ((ret->data == NULL) || (ret->length < len))
+	/* detach data from object */
+	data = (unsigned char *)ret->data;
+	ret->data = NULL;
+	/* once detached we can change it */
+	if ((data == NULL) || (ret->length < len))
 		{
-		if (ret->data != NULL) OPENSSL_free(ret->data);
-		ret->data=(unsigned char *)OPENSSL_malloc(len ? (int)len : 1);
-		ret->flags|=ASN1_OBJECT_FLAG_DYNAMIC_DATA;
-		if (ret->data == NULL)
+		ret->length=0;
+		if (data != NULL) OPENSSL_free(data);
+		data=(unsigned char *)OPENSSL_malloc(len ? (int)len : 1);
+		if (data == NULL)
 			{ i=ERR_R_MALLOC_FAILURE; goto err; }
+		ret->flags|=ASN1_OBJECT_FLAG_DYNAMIC_DATA;
 		}
-	memcpy(ret->data,p,(int)len);
+	memcpy(data,p,(int)len);
+	/* reattach data to object, after which it remains const */
+	ret->data  =data;
 	ret->length=(int)len;
 	ret->sn=NULL;
 	ret->ln=NULL;
@@ -370,7 +376,7 @@ void ASN1_OBJECT_free(ASN1_OBJECT *a)
 		}
 	if (a->flags & ASN1_OBJECT_FLAG_DYNAMIC_DATA)
 		{
-		if (a->data != NULL) OPENSSL_free(a->data);
+		if (a->data != NULL) OPENSSL_free((void *)a->data);
 		a->data=NULL;
 		a->length=0;
 		}
diff --git a/crypto/asn1/a_octet.c b/crypto/asn1/a_octet.c
index 24fd0f8e5a76..e8725e44f1f5 100644
--- a/crypto/asn1/a_octet.c
+++ b/crypto/asn1/a_octet.c
@@ -60,10 +60,10 @@
 #include "cryptlib.h"
 #include <openssl/asn1.h>
 
-ASN1_OCTET_STRING *ASN1_OCTET_STRING_dup(ASN1_OCTET_STRING *x)
+ASN1_OCTET_STRING *ASN1_OCTET_STRING_dup(const ASN1_OCTET_STRING *x)
 { return M_ASN1_OCTET_STRING_dup(x); }
 
-int ASN1_OCTET_STRING_cmp(ASN1_OCTET_STRING *a, ASN1_OCTET_STRING *b)
+int ASN1_OCTET_STRING_cmp(const ASN1_OCTET_STRING *a, const ASN1_OCTET_STRING *b)
 { return M_ASN1_OCTET_STRING_cmp(a, b); }
 
 int ASN1_OCTET_STRING_set(ASN1_OCTET_STRING *x, const unsigned char *d, int len)
diff --git a/crypto/asn1/a_set.c b/crypto/asn1/a_set.c
index 958558c204dd..d726c8d3a87c 100644
--- a/crypto/asn1/a_set.c
+++ b/crypto/asn1/a_set.c
@@ -85,8 +85,9 @@ static int SetBlobCmp(const void *elem1, const void *elem2 )
     }
 
 /* int is_set:  if TRUE, then sort the contents (i.e. it isn't a SEQUENCE)    */
-int i2d_ASN1_SET(STACK *a, unsigned char **pp, i2d_of_void *i2d, int ex_tag,
-		 int ex_class, int is_set)
+int i2d_ASN1_SET(STACK_OF(OPENSSL_BLOCK) *a, unsigned char **pp,
+		 i2d_of_void *i2d, int ex_tag, int ex_class,
+		 int is_set)
 	{
 	int ret=0,r;
 	int i;
@@ -96,8 +97,8 @@ int i2d_ASN1_SET(STACK *a, unsigned char **pp, i2d_of_void *i2d, int ex_tag,
         int totSize;
 
 	if (a == NULL) return(0);
-	for (i=sk_num(a)-1; i>=0; i--)
-		ret+=i2d(sk_value(a,i),NULL);
+	for (i=sk_OPENSSL_BLOCK_num(a)-1; i>=0; i--)
+		ret+=i2d(sk_OPENSSL_BLOCK_value(a,i),NULL);
 	r=ASN1_object_size(1,ret,ex_tag);
 	if (pp == NULL) return(r);
 
@@ -108,10 +109,10 @@ int i2d_ASN1_SET(STACK *a, unsigned char **pp, i2d_of_void *i2d, int ex_tag,
 	/* And then again by Ben */
 	/* And again by Steve */
 
-	if(!is_set || (sk_num(a) < 2))
+	if(!is_set || (sk_OPENSSL_BLOCK_num(a) < 2))
 		{
-		for (i=0; i<sk_num(a); i++)
-                	i2d(sk_value(a,i),&p);
+		for (i=0; i<sk_OPENSSL_BLOCK_num(a); i++)
+                	i2d(sk_OPENSSL_BLOCK_value(a,i),&p);
 
 		*pp=p;
 		return(r);
@@ -119,17 +120,17 @@ int i2d_ASN1_SET(STACK *a, unsigned char **pp, i2d_of_void *i2d, int ex_tag,
 
         pStart  = p; /* Catch the beg of Setblobs*/
 		/* In this array we will store the SET blobs */
-		rgSetBlob = (MYBLOB *)OPENSSL_malloc(sk_num(a) * sizeof(MYBLOB));
+		rgSetBlob = OPENSSL_malloc(sk_OPENSSL_BLOCK_num(a) * sizeof(MYBLOB));
 		if (rgSetBlob == NULL)
 			{
 			ASN1err(ASN1_F_I2D_ASN1_SET,ERR_R_MALLOC_FAILURE);
 			return(0);
 			}
 
-        for (i=0; i<sk_num(a); i++)
+        for (i=0; i<sk_OPENSSL_BLOCK_num(a); i++)
 	        {
                 rgSetBlob[i].pbData = p;  /* catch each set encode blob */
-                i2d(sk_value(a,i),&p);
+                i2d(sk_OPENSSL_BLOCK_value(a,i),&p);
                 rgSetBlob[i].cbData = p - rgSetBlob[i].pbData; /* Length of this
 SetBlob
 */
@@ -139,7 +140,7 @@ SetBlob
 
  /* Now we have to sort the blobs. I am using a simple algo.
     *Sort ptrs *Copy to temp-mem *Copy from temp-mem to user-mem*/
-        qsort( rgSetBlob, sk_num(a), sizeof(MYBLOB), SetBlobCmp);
+        qsort( rgSetBlob, sk_OPENSSL_BLOCK_num(a), sizeof(MYBLOB), SetBlobCmp);
 		if (!(pTempMem = OPENSSL_malloc(totSize)))
 			{
 			ASN1err(ASN1_F_I2D_ASN1_SET,ERR_R_MALLOC_FAILURE);
@@ -148,7 +149,7 @@ SetBlob
 
 /* Copy to temp mem */
         p = pTempMem;
-        for(i=0; i<sk_num(a); ++i)
+        for(i=0; i<sk_OPENSSL_BLOCK_num(a); ++i)
 		{
                 memcpy(p, rgSetBlob[i].pbData, rgSetBlob[i].cbData);
                 p += rgSetBlob[i].cbData;
@@ -162,16 +163,18 @@ SetBlob
         return(r);
         }
 
-STACK *d2i_ASN1_SET(STACK **a, const unsigned char **pp, long length,
-		    d2i_of_void *d2i, void (*free_func)(void *), int ex_tag,
-		    int ex_class)
+STACK_OF(OPENSSL_BLOCK) *d2i_ASN1_SET(STACK_OF(OPENSSL_BLOCK) **a,
+			      const unsigned char **pp,
+			      long length, d2i_of_void *d2i,
+			      void (*free_func)(OPENSSL_BLOCK), int ex_tag,
+			      int ex_class)
 	{
 	ASN1_const_CTX c;
-	STACK *ret=NULL;
+	STACK_OF(OPENSSL_BLOCK) *ret=NULL;
 
 	if ((a == NULL) || ((*a) == NULL))
 		{
-		if ((ret=sk_new_null()) == NULL)
+		if ((ret=sk_OPENSSL_BLOCK_new_null()) == NULL)
 			{
 			ASN1err(ASN1_F_D2I_ASN1_SET,ERR_R_MALLOC_FAILURE);
 			goto err;
@@ -216,10 +219,10 @@ STACK *d2i_ASN1_SET(STACK **a, const unsigned char **pp, long length,
 		if ((s=d2i(NULL,&c.p,c.slen)) == NULL)
 			{
 			ASN1err(ASN1_F_D2I_ASN1_SET,ASN1_R_ERROR_PARSING_SET_ELEMENT);
-			asn1_add_error(*pp,(int)(c.q- *pp));
+			asn1_add_error(*pp,(int)(c.p- *pp));
 			goto err;
 			}
-		if (!sk_push(ret,s)) goto err;
+		if (!sk_OPENSSL_BLOCK_push(ret,s)) goto err;
 		}
 	if (a != NULL) (*a)=ret;
 	*pp=c.p;
@@ -228,9 +231,9 @@ err:
 	if ((ret != NULL) && ((a == NULL) || (*a != ret)))
 		{
 		if (free_func != NULL)
-			sk_pop_free(ret,free_func);
+			sk_OPENSSL_BLOCK_pop_free(ret,free_func);
 		else
-			sk_free(ret);
+			sk_OPENSSL_BLOCK_free(ret);
 		}
 	return(NULL);
 	}
diff --git a/crypto/asn1/a_sign.c b/crypto/asn1/a_sign.c
index 4dee45fbb83a..7b4a193d6b15 100644
--- a/crypto/asn1/a_sign.c
+++ b/crypto/asn1/a_sign.c
@@ -123,6 +123,7 @@
 #include <openssl/x509.h>
 #include <openssl/objects.h>
 #include <openssl/buffer.h>
+#include "asn1_locl.h"
 
 #ifndef NO_ASN1_OLD
 
@@ -183,9 +184,9 @@ int ASN1_sign(i2d_of_void *i2d, X509_ALGOR *algor1, X509_ALGOR *algor2,
 	p=buf_in;
 
 	i2d(data,&p);
-	EVP_SignInit_ex(&ctx,type, NULL);
-	EVP_SignUpdate(&ctx,(unsigned char *)buf_in,inl);
-	if (!EVP_SignFinal(&ctx,(unsigned char *)buf_out,
+	if (!EVP_SignInit_ex(&ctx,type, NULL)
+		|| !EVP_SignUpdate(&ctx,(unsigned char *)buf_in,inl)
+		|| !EVP_SignFinal(&ctx,(unsigned char *)buf_out,
 			(unsigned int *)&outl,pkey))
 		{
 		outl=0;
@@ -217,68 +218,100 @@ int ASN1_item_sign(const ASN1_ITEM *it, X509_ALGOR *algor1, X509_ALGOR *algor2,
 	     const EVP_MD *type)
 	{
 	EVP_MD_CTX ctx;
+	EVP_MD_CTX_init(&ctx);
+	if (!EVP_DigestSignInit(&ctx, NULL, type, NULL, pkey))
+		{
+		EVP_MD_CTX_cleanup(&ctx);
+		return 0;
+		}
+	return ASN1_item_sign_ctx(it, algor1, algor2, signature, asn, &ctx);
+	}
+		
+
+int ASN1_item_sign_ctx(const ASN1_ITEM *it,
+		X509_ALGOR *algor1, X509_ALGOR *algor2,
+	     	ASN1_BIT_STRING *signature, void *asn, EVP_MD_CTX *ctx)
+	{
+	const EVP_MD *type;
+	EVP_PKEY *pkey;
 	unsigned char *buf_in=NULL,*buf_out=NULL;
-	int i,inl=0,outl=0,outll=0;
-	X509_ALGOR *a;
+	size_t inl=0,outl=0,outll=0;
+	int signid, paramtype;
+	int rv;
 
-	EVP_MD_CTX_init(&ctx);
-	for (i=0; i<2; i++)
+	type = EVP_MD_CTX_md(ctx);
+	pkey = EVP_PKEY_CTX_get0_pkey(ctx->pctx);
+
+	if (!type || !pkey)
 		{
-		if (i == 0)
-			a=algor1;
-		else
-			a=algor2;
-		if (a == NULL) continue;
-                if (type->pkey_type == NID_dsaWithSHA1 ||
-			type->pkey_type == NID_ecdsa_with_SHA1)
-			{
-			/* special case: RFC 3279 tells us to omit 'parameters'
-			 * with id-dsa-with-sha1 and ecdsa-with-SHA1 */
-			ASN1_TYPE_free(a->parameter);
-			a->parameter = NULL;
-			}
-		else if ((a->parameter == NULL) || 
-			(a->parameter->type != V_ASN1_NULL))
-			{
-			ASN1_TYPE_free(a->parameter);
-			if ((a->parameter=ASN1_TYPE_new()) == NULL) goto err;
-			a->parameter->type=V_ASN1_NULL;
-			}
-		ASN1_OBJECT_free(a->algorithm);
-		a->algorithm=OBJ_nid2obj(type->pkey_type);
-		if (a->algorithm == NULL)
-			{
-			ASN1err(ASN1_F_ASN1_ITEM_SIGN,ASN1_R_UNKNOWN_OBJECT_TYPE);
+		ASN1err(ASN1_F_ASN1_ITEM_SIGN_CTX, ASN1_R_CONTEXT_NOT_INITIALISED);
+		return 0;
+		}
+
+	if (pkey->ameth->item_sign)
+		{
+		rv = pkey->ameth->item_sign(ctx, it, asn, algor1, algor2,
+						signature);
+		if (rv == 1)
+			outl = signature->length;
+		/* Return value meanings:
+		 * <=0: error.
+		 *   1: method does everything.
+		 *   2: carry on as normal.
+		 *   3: ASN1 method sets algorithm identifiers: just sign.
+		 */
+		if (rv <= 0)
+			ASN1err(ASN1_F_ASN1_ITEM_SIGN_CTX, ERR_R_EVP_LIB);
+		if (rv <= 1)
 			goto err;
-			}
-		if (a->algorithm->length == 0)
+		}
+	else
+		rv = 2;
+
+	if (rv == 2)
+		{
+		if (type->flags & EVP_MD_FLAG_PKEY_METHOD_SIGNATURE)
 			{
-			ASN1err(ASN1_F_ASN1_ITEM_SIGN,ASN1_R_THE_ASN1_OBJECT_IDENTIFIER_IS_NOT_KNOWN_FOR_THIS_MD);
-			goto err;
+			if (!pkey->ameth ||
+				!OBJ_find_sigid_by_algs(&signid,
+							EVP_MD_nid(type),
+							pkey->ameth->pkey_id))
+				{
+				ASN1err(ASN1_F_ASN1_ITEM_SIGN_CTX,
+					ASN1_R_DIGEST_AND_KEY_TYPE_NOT_SUPPORTED);
+				return 0;
+				}
 			}
+		else
+			signid = type->pkey_type;
+
+		if (pkey->ameth->pkey_flags & ASN1_PKEY_SIGPARAM_NULL)
+			paramtype = V_ASN1_NULL;
+		else
+			paramtype = V_ASN1_UNDEF;
+
+		if (algor1)
+			X509_ALGOR_set0(algor1, OBJ_nid2obj(signid), paramtype, NULL);
+		if (algor2)
+			X509_ALGOR_set0(algor2, OBJ_nid2obj(signid), paramtype, NULL);
+
 		}
+
 	inl=ASN1_item_i2d(asn,&buf_in, it);
 	outll=outl=EVP_PKEY_size(pkey);
-	buf_out=(unsigned char *)OPENSSL_malloc((unsigned int)outl);
+	buf_out=OPENSSL_malloc((unsigned int)outl);
 	if ((buf_in == NULL) || (buf_out == NULL))
 		{
 		outl=0;
-		ASN1err(ASN1_F_ASN1_ITEM_SIGN,ERR_R_MALLOC_FAILURE);
+		ASN1err(ASN1_F_ASN1_ITEM_SIGN_CTX,ERR_R_MALLOC_FAILURE);
 		goto err;
 		}
 
-	if (!EVP_SignInit_ex(&ctx,type, NULL))
+	if (!EVP_DigestSignUpdate(ctx, buf_in, inl)
+		|| !EVP_DigestSignFinal(ctx, buf_out, &outl))
 		{
 		outl=0;
-		ASN1err(ASN1_F_ASN1_ITEM_SIGN,ERR_R_EVP_LIB);
-		goto err;
-		}
-	EVP_SignUpdate(&ctx,(unsigned char *)buf_in,inl);
-	if (!EVP_SignFinal(&ctx,(unsigned char *)buf_out,
-			(unsigned int *)&outl,pkey))
-		{
-		outl=0;
-		ASN1err(ASN1_F_ASN1_ITEM_SIGN,ERR_R_EVP_LIB);
+		ASN1err(ASN1_F_ASN1_ITEM_SIGN_CTX,ERR_R_EVP_LIB);
 		goto err;
 		}
 	if (signature->data != NULL) OPENSSL_free(signature->data);
@@ -291,7 +324,7 @@ int ASN1_item_sign(const ASN1_ITEM *it, X509_ALGOR *algor1, X509_ALGOR *algor2,
 	signature->flags&= ~(ASN1_STRING_FLAG_BITS_LEFT|0x07);
 	signature->flags|=ASN1_STRING_FLAG_BITS_LEFT;
 err:
-	EVP_MD_CTX_cleanup(&ctx);
+	EVP_MD_CTX_cleanup(ctx);
 	if (buf_in != NULL)
 		{ OPENSSL_cleanse((char *)buf_in,(unsigned int)inl); OPENSSL_free(buf_in); }
 	if (buf_out != NULL)
diff --git a/crypto/asn1/a_strnid.c b/crypto/asn1/a_strnid.c
index b68ae433d4ea..2fc48c1551a5 100644
--- a/crypto/asn1/a_strnid.c
+++ b/crypto/asn1/a_strnid.c
@@ -67,7 +67,6 @@ static STACK_OF(ASN1_STRING_TABLE) *stable = NULL;
 static void st_free(ASN1_STRING_TABLE *tbl);
 static int sk_table_cmp(const ASN1_STRING_TABLE * const *a,
 			const ASN1_STRING_TABLE * const *b);
-static int table_cmp(const void *a, const void *b);
 
 
 /* This is the global mask for the mbstring functions: this is use to
@@ -158,7 +157,7 @@ ASN1_STRING *ASN1_STRING_set_by_NID(ASN1_STRING **out, const unsigned char *in,
 
 /* This table must be kept in NID order */
 
-static ASN1_STRING_TABLE tbl_standard[] = {
+static const ASN1_STRING_TABLE tbl_standard[] = {
 {NID_commonName,		1, ub_common_name, DIRSTRING_TYPE, 0},
 {NID_countryName,		2, 2, B_ASN1_PRINTABLESTRING, STABLE_NO_MASK},
 {NID_localityName,		1, ub_locality_name, DIRSTRING_TYPE, 0},
@@ -186,22 +185,23 @@ static int sk_table_cmp(const ASN1_STRING_TABLE * const *a,
 	return (*a)->nid - (*b)->nid;
 }
 
-static int table_cmp(const void *a, const void *b)
+DECLARE_OBJ_BSEARCH_CMP_FN(ASN1_STRING_TABLE, ASN1_STRING_TABLE, table);
+
+static int table_cmp(const ASN1_STRING_TABLE *a, const ASN1_STRING_TABLE *b)
 {
-	const ASN1_STRING_TABLE *sa = a, *sb = b;
-	return sa->nid - sb->nid;
+	return a->nid - b->nid;
 }
 
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(ASN1_STRING_TABLE, ASN1_STRING_TABLE, table);
+
 ASN1_STRING_TABLE *ASN1_STRING_TABLE_get(int nid)
 {
 	int idx;
 	ASN1_STRING_TABLE *ttmp;
 	ASN1_STRING_TABLE fnd;
 	fnd.nid = nid;
-	ttmp = (ASN1_STRING_TABLE *) OBJ_bsearch((char *)&fnd,
-					(char *)tbl_standard, 
-			sizeof(tbl_standard)/sizeof(ASN1_STRING_TABLE),
-			sizeof(ASN1_STRING_TABLE), table_cmp);
+	ttmp = OBJ_bsearch_table(&fnd, tbl_standard, 
+			   sizeof(tbl_standard)/sizeof(ASN1_STRING_TABLE));
 	if(ttmp) return ttmp;
 	if(!stable) return NULL;
 	idx = sk_ASN1_STRING_TABLE_find(stable, &fnd);
diff --git a/crypto/asn1/a_time.c b/crypto/asn1/a_time.c
index 159681fbcb06..e2eb9b243efb 100644
--- a/crypto/asn1/a_time.c
+++ b/crypto/asn1/a_time.c
@@ -100,18 +100,29 @@ int i2d_ASN1_TIME(ASN1_TIME *a, unsigned char **pp)
 
 ASN1_TIME *ASN1_TIME_set(ASN1_TIME *s, time_t t)
 	{
+	return ASN1_TIME_adj(s, t, 0, 0);
+	}
+
+ASN1_TIME *ASN1_TIME_adj(ASN1_TIME *s, time_t t,
+				int offset_day, long offset_sec)
+	{
 	struct tm *ts;
 	struct tm data;
 
 	ts=OPENSSL_gmtime(&t,&data);
 	if (ts == NULL)
 		{
-		ASN1err(ASN1_F_ASN1_TIME_SET, ASN1_R_ERROR_GETTING_TIME);
+		ASN1err(ASN1_F_ASN1_TIME_ADJ, ASN1_R_ERROR_GETTING_TIME);
 		return NULL;
 		}
+	if (offset_day || offset_sec)
+		{ 
+		if (!OPENSSL_gmtime_adj(ts, offset_day, offset_sec))
+			return NULL;
+		}
 	if((ts->tm_year >= 50) && (ts->tm_year < 150))
-					return ASN1_UTCTIME_set(s, t);
-	return ASN1_GENERALIZEDTIME_set(s,t);
+			return ASN1_UTCTIME_adj(s, t, offset_day, offset_sec);
+	return ASN1_GENERALIZEDTIME_adj(s, t, offset_day, offset_sec);
 	}
 
 int ASN1_TIME_check(ASN1_TIME *t)
@@ -162,3 +173,26 @@ ASN1_GENERALIZEDTIME *ASN1_TIME_to_generalizedtime(ASN1_TIME *t, ASN1_GENERALIZE
 
 	return ret;
 	}
+
+int ASN1_TIME_set_string(ASN1_TIME *s, const char *str)
+	{
+	ASN1_TIME t;
+
+	t.length = strlen(str);
+	t.data = (unsigned char *)str;
+	t.flags = 0;
+	
+	t.type = V_ASN1_UTCTIME;
+
+	if (!ASN1_TIME_check(&t))
+		{
+		t.type = V_ASN1_GENERALIZEDTIME;
+		if (!ASN1_TIME_check(&t))
+			return 0;
+		}
+	
+	if (s && !ASN1_STRING_copy((ASN1_STRING *)s, (ASN1_STRING *)&t))
+			return 0;
+
+	return 1;
+	}
diff --git a/crypto/asn1/a_type.c b/crypto/asn1/a_type.c
index 36beceacdb07..a45d2f9d12dd 100644
--- a/crypto/asn1/a_type.c
+++ b/crypto/asn1/a_type.c
@@ -77,7 +77,10 @@ void ASN1_TYPE_set(ASN1_TYPE *a, int type, void *value)
 		ASN1_primitive_free((ASN1_VALUE **)tmp_a, NULL);
 		}
 	a->type=type;
-	a->value.ptr=value;
+	if (type == V_ASN1_BOOLEAN)
+		a->value.boolean = value ? 0xff : 0;
+	else
+		a->value.ptr=value;
 	}
 
 int ASN1_TYPE_set1(ASN1_TYPE *a, int type, const void *value)
@@ -98,7 +101,7 @@ int ASN1_TYPE_set1(ASN1_TYPE *a, int type, const void *value)
 	else
 		{
 		ASN1_STRING *sdup;
-		sdup = ASN1_STRING_dup((ASN1_STRING *)value);
+		sdup = ASN1_STRING_dup(value);
 		if (!sdup)
 			return 0;
 		ASN1_TYPE_set(a, type, sdup);
@@ -108,3 +111,49 @@ int ASN1_TYPE_set1(ASN1_TYPE *a, int type, const void *value)
 
 IMPLEMENT_STACK_OF(ASN1_TYPE)
 IMPLEMENT_ASN1_SET_OF(ASN1_TYPE)
+
+/* Returns 0 if they are equal, != 0 otherwise. */
+int ASN1_TYPE_cmp(ASN1_TYPE *a, ASN1_TYPE *b)
+	{
+	int result = -1;
+
+	if (!a || !b || a->type != b->type) return -1;
+
+	switch (a->type)
+		{
+	case V_ASN1_OBJECT:
+		result = OBJ_cmp(a->value.object, b->value.object);
+		break;
+	case V_ASN1_NULL:
+		result = 0;	/* They do not have content. */
+		break;
+	case V_ASN1_INTEGER:
+	case V_ASN1_NEG_INTEGER:
+	case V_ASN1_ENUMERATED:
+	case V_ASN1_NEG_ENUMERATED:
+	case V_ASN1_BIT_STRING:
+	case V_ASN1_OCTET_STRING:
+	case V_ASN1_SEQUENCE:
+	case V_ASN1_SET:
+	case V_ASN1_NUMERICSTRING:
+	case V_ASN1_PRINTABLESTRING:
+	case V_ASN1_T61STRING:
+	case V_ASN1_VIDEOTEXSTRING:
+	case V_ASN1_IA5STRING:
+	case V_ASN1_UTCTIME:
+	case V_ASN1_GENERALIZEDTIME:
+	case V_ASN1_GRAPHICSTRING:
+	case V_ASN1_VISIBLESTRING:
+	case V_ASN1_GENERALSTRING:
+	case V_ASN1_UNIVERSALSTRING:
+	case V_ASN1_BMPSTRING:
+	case V_ASN1_UTF8STRING:
+	case V_ASN1_OTHER:
+	default:
+		result = ASN1_STRING_cmp((ASN1_STRING *) a->value.ptr,
+					 (ASN1_STRING *) b->value.ptr);
+		break;
+		}
+
+	return result;
+	}
diff --git a/crypto/asn1/a_utctm.c b/crypto/asn1/a_utctm.c
index d31c0281930b..072e23659233 100644
--- a/crypto/asn1/a_utctm.c
+++ b/crypto/asn1/a_utctm.c
@@ -114,8 +114,8 @@ err:
 
 int ASN1_UTCTIME_check(ASN1_UTCTIME *d)
 	{
-	static int min[8]={ 0, 1, 1, 0, 0, 0, 0, 0};
-	static int max[8]={99,12,31,23,59,59,12,59};
+	static const int min[8]={ 0, 1, 1, 0, 0, 0, 0, 0};
+	static const int max[8]={99,12,31,23,59,59,12,59};
 	char *a;
 	int n,i,l,o;
 
@@ -186,6 +186,12 @@ int ASN1_UTCTIME_set_string(ASN1_UTCTIME *s, const char *str)
 
 ASN1_UTCTIME *ASN1_UTCTIME_set(ASN1_UTCTIME *s, time_t t)
 	{
+	return ASN1_UTCTIME_adj(s, t, 0, 0);
+	}
+
+ASN1_UTCTIME *ASN1_UTCTIME_adj(ASN1_UTCTIME *s, time_t t,
+				int offset_day, long offset_sec)
+	{
 	char *p;
 	struct tm *ts;
 	struct tm data;
@@ -200,13 +206,22 @@ ASN1_UTCTIME *ASN1_UTCTIME_set(ASN1_UTCTIME *s, time_t t)
 	if (ts == NULL)
 		return(NULL);
 
+	if (offset_day || offset_sec)
+		{ 
+		if (!OPENSSL_gmtime_adj(ts, offset_day, offset_sec))
+			return NULL;
+		}
+
+	if((ts->tm_year < 50) || (ts->tm_year >= 150))
+		return NULL;
+
 	p=(char *)s->data;
 	if ((p == NULL) || ((size_t)s->length < len))
 		{
 		p=OPENSSL_malloc(len);
 		if (p == NULL)
 			{
-			ASN1err(ASN1_F_ASN1_UTCTIME_SET,ERR_R_MALLOC_FAILURE);
+			ASN1err(ASN1_F_ASN1_UTCTIME_ADJ,ERR_R_MALLOC_FAILURE);
 			return(NULL);
 			}
 		if (s->data != NULL)
diff --git a/crypto/asn1/a_verify.c b/crypto/asn1/a_verify.c
index da3efaaf8de5..432722e4099e 100644
--- a/crypto/asn1/a_verify.c
+++ b/crypto/asn1/a_verify.c
@@ -60,6 +60,7 @@
 #include <time.h>
 
 #include "cryptlib.h"
+#include "asn1_locl.h"
 
 #ifndef NO_SYS_TYPES_H
 # include <sys/types.h>
@@ -100,13 +101,13 @@ int ASN1_verify(i2d_of_void *i2d, X509_ALGOR *a, ASN1_BIT_STRING *signature,
 	p=buf_in;
 
 	i2d(data,&p);
-	if (!EVP_VerifyInit_ex(&ctx,type, NULL))
+	if (!EVP_VerifyInit_ex(&ctx,type, NULL)
+		|| !EVP_VerifyUpdate(&ctx,(unsigned char *)buf_in,inl))
 		{
 		ASN1err(ASN1_F_ASN1_VERIFY,ERR_R_EVP_LIB);
 		ret=0;
 		goto err;
 		}
-	EVP_VerifyUpdate(&ctx,(unsigned char *)buf_in,inl);
 
 	OPENSSL_cleanse(buf_in,(unsigned int)inl);
 	OPENSSL_free(buf_in);
@@ -130,28 +131,64 @@ err:
 #endif
 
 
-int ASN1_item_verify(const ASN1_ITEM *it, X509_ALGOR *a, ASN1_BIT_STRING *signature,
-	     void *asn, EVP_PKEY *pkey)
+int ASN1_item_verify(const ASN1_ITEM *it, X509_ALGOR *a,
+		ASN1_BIT_STRING *signature, void *asn, EVP_PKEY *pkey)
 	{
 	EVP_MD_CTX ctx;
-	const EVP_MD *type;
 	unsigned char *buf_in=NULL;
-	int ret= -1,i,inl;
+	int ret= -1,inl;
+
+	int mdnid, pknid;
 
 	EVP_MD_CTX_init(&ctx);
-	i=OBJ_obj2nid(a->algorithm);
-	type=EVP_get_digestbyname(OBJ_nid2sn(i));
-	if (type == NULL)
+
+	/* Convert signature OID into digest and public key OIDs */
+	if (!OBJ_find_sigid_algs(OBJ_obj2nid(a->algorithm), &mdnid, &pknid))
 		{
-		ASN1err(ASN1_F_ASN1_ITEM_VERIFY,ASN1_R_UNKNOWN_MESSAGE_DIGEST_ALGORITHM);
+		ASN1err(ASN1_F_ASN1_ITEM_VERIFY,ASN1_R_UNKNOWN_SIGNATURE_ALGORITHM);
 		goto err;
 		}
-
-	if (!EVP_VerifyInit_ex(&ctx,type, NULL))
+	if (mdnid == NID_undef)
 		{
-		ASN1err(ASN1_F_ASN1_ITEM_VERIFY,ERR_R_EVP_LIB);
-		ret=0;
-		goto err;
+		if (!pkey->ameth || !pkey->ameth->item_verify)
+			{
+			ASN1err(ASN1_F_ASN1_ITEM_VERIFY,ASN1_R_UNKNOWN_SIGNATURE_ALGORITHM);
+			goto err;
+			}
+		ret = pkey->ameth->item_verify(&ctx, it, asn, a,
+							signature, pkey);
+		/* Return value of 2 means carry on, anything else means we
+		 * exit straight away: either a fatal error of the underlying
+		 * verification routine handles all verification.
+		 */
+		if (ret != 2)
+			goto err;
+		ret = -1;
+		}
+	else
+		{
+		const EVP_MD *type;
+		type=EVP_get_digestbynid(mdnid);
+		if (type == NULL)
+			{
+			ASN1err(ASN1_F_ASN1_ITEM_VERIFY,ASN1_R_UNKNOWN_MESSAGE_DIGEST_ALGORITHM);
+			goto err;
+			}
+
+		/* Check public key OID matches public key type */
+		if (EVP_PKEY_type(pknid) != pkey->ameth->pkey_id)
+			{
+			ASN1err(ASN1_F_ASN1_ITEM_VERIFY,ASN1_R_WRONG_PUBLIC_KEY_TYPE);
+			goto err;
+			}
+
+		if (!EVP_DigestVerifyInit(&ctx, NULL, type, NULL, pkey))
+			{
+			ASN1err(ASN1_F_ASN1_ITEM_VERIFY,ERR_R_EVP_LIB);
+			ret=0;
+			goto err;
+			}
+
 		}
 
 	inl = ASN1_item_i2d(asn, &buf_in, it);
@@ -162,13 +199,18 @@ int ASN1_item_verify(const ASN1_ITEM *it, X509_ALGOR *a, ASN1_BIT_STRING *signat
 		goto err;
 		}
 
-	EVP_VerifyUpdate(&ctx,(unsigned char *)buf_in,inl);
+	if (!EVP_DigestVerifyUpdate(&ctx,buf_in,inl))
+		{
+		ASN1err(ASN1_F_ASN1_ITEM_VERIFY,ERR_R_EVP_LIB);
+		ret=0;
+		goto err;
+		}
 
 	OPENSSL_cleanse(buf_in,(unsigned int)inl);
 	OPENSSL_free(buf_in);
 
-	if (EVP_VerifyFinal(&ctx,(unsigned char *)signature->data,
-			(unsigned int)signature->length,pkey) <= 0)
+	if (EVP_DigestVerifyFinal(&ctx,signature->data,
+			(size_t)signature->length) <= 0)
 		{
 		ASN1err(ASN1_F_ASN1_ITEM_VERIFY,ERR_R_EVP_LIB);
 		ret=0;
diff --git a/crypto/asn1/ameth_lib.c b/crypto/asn1/ameth_lib.c
new file mode 100644
index 000000000000..a19e058fca69
--- /dev/null
+++ b/crypto/asn1/ameth_lib.c
@@ -0,0 +1,460 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/asn1t.h>
+#include <openssl/x509.h>
+#ifndef OPENSSL_NO_ENGINE
+#include <openssl/engine.h>
+#endif
+#include "asn1_locl.h"
+
+extern const EVP_PKEY_ASN1_METHOD rsa_asn1_meths[];
+extern const EVP_PKEY_ASN1_METHOD dsa_asn1_meths[];
+extern const EVP_PKEY_ASN1_METHOD dh_asn1_meth;
+extern const EVP_PKEY_ASN1_METHOD eckey_asn1_meth;
+extern const EVP_PKEY_ASN1_METHOD hmac_asn1_meth;
+extern const EVP_PKEY_ASN1_METHOD cmac_asn1_meth;
+
+/* Keep this sorted in type order !! */
+static const EVP_PKEY_ASN1_METHOD *standard_methods[] = 
+	{
+#ifndef OPENSSL_NO_RSA
+	&rsa_asn1_meths[0],
+	&rsa_asn1_meths[1],
+#endif
+#ifndef OPENSSL_NO_DH
+	&dh_asn1_meth,
+#endif
+#ifndef OPENSSL_NO_DSA
+	&dsa_asn1_meths[0],
+	&dsa_asn1_meths[1],
+	&dsa_asn1_meths[2],
+	&dsa_asn1_meths[3],
+	&dsa_asn1_meths[4],
+#endif
+#ifndef OPENSSL_NO_EC
+	&eckey_asn1_meth,
+#endif
+	&hmac_asn1_meth,
+	&cmac_asn1_meth
+	};
+
+typedef int sk_cmp_fn_type(const char * const *a, const char * const *b);
+DECLARE_STACK_OF(EVP_PKEY_ASN1_METHOD)
+static STACK_OF(EVP_PKEY_ASN1_METHOD) *app_methods = NULL;
+
+
+
+#ifdef TEST
+void main()
+	{
+	int i;
+	for (i = 0;
+		i < sizeof(standard_methods)/sizeof(EVP_PKEY_ASN1_METHOD *);
+		i++)
+		fprintf(stderr, "Number %d id=%d (%s)\n", i,
+			standard_methods[i]->pkey_id,
+			OBJ_nid2sn(standard_methods[i]->pkey_id));
+	}
+#endif
+
+DECLARE_OBJ_BSEARCH_CMP_FN(const EVP_PKEY_ASN1_METHOD *,
+			   const EVP_PKEY_ASN1_METHOD *, ameth);
+
+static int ameth_cmp(const EVP_PKEY_ASN1_METHOD * const *a,
+		     const EVP_PKEY_ASN1_METHOD * const *b)
+	{
+        return ((*a)->pkey_id - (*b)->pkey_id);
+	}
+
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(const EVP_PKEY_ASN1_METHOD *,
+			     const EVP_PKEY_ASN1_METHOD *, ameth);
+
+int EVP_PKEY_asn1_get_count(void)
+	{
+	int num = sizeof(standard_methods)/sizeof(EVP_PKEY_ASN1_METHOD *);
+	if (app_methods)
+		num += sk_EVP_PKEY_ASN1_METHOD_num(app_methods);
+	return num;
+	}
+
+const EVP_PKEY_ASN1_METHOD *EVP_PKEY_asn1_get0(int idx)
+	{
+	int num = sizeof(standard_methods)/sizeof(EVP_PKEY_ASN1_METHOD *);
+	if (idx < 0)
+		return NULL; 
+	if (idx < num)
+		return standard_methods[idx];
+	idx -= num;
+	return sk_EVP_PKEY_ASN1_METHOD_value(app_methods, idx);
+	}
+
+static const EVP_PKEY_ASN1_METHOD *pkey_asn1_find(int type)
+	{
+	EVP_PKEY_ASN1_METHOD tmp;
+	const EVP_PKEY_ASN1_METHOD *t = &tmp, **ret;
+	tmp.pkey_id = type;
+	if (app_methods)
+		{
+		int idx;
+		idx = sk_EVP_PKEY_ASN1_METHOD_find(app_methods, &tmp);
+		if (idx >= 0)
+			return sk_EVP_PKEY_ASN1_METHOD_value(app_methods, idx);
+		}
+	ret = OBJ_bsearch_ameth(&t, standard_methods,
+			  sizeof(standard_methods)
+			  /sizeof(EVP_PKEY_ASN1_METHOD *));
+	if (!ret || !*ret)
+		return NULL;
+	return *ret;
+	}
+
+/* Find an implementation of an ASN1 algorithm. If 'pe' is not NULL
+ * also search through engines and set *pe to a functional reference
+ * to the engine implementing 'type' or NULL if no engine implements 
+ * it.
+ */
+
+const EVP_PKEY_ASN1_METHOD *EVP_PKEY_asn1_find(ENGINE **pe, int type)
+	{
+	const EVP_PKEY_ASN1_METHOD *t;
+
+	for (;;)
+		{
+		t = pkey_asn1_find(type);
+		if (!t || !(t->pkey_flags & ASN1_PKEY_ALIAS))
+			break;
+		type = t->pkey_base_id;
+		}
+	if (pe)
+		{
+#ifndef OPENSSL_NO_ENGINE
+		ENGINE *e;
+		/* type will contain the final unaliased type */
+		e = ENGINE_get_pkey_asn1_meth_engine(type);
+		if (e)
+			{
+			*pe = e;
+			return ENGINE_get_pkey_asn1_meth(e, type);
+			}
+#endif
+		*pe = NULL;
+		}
+	return t;
+	}
+
+const EVP_PKEY_ASN1_METHOD *EVP_PKEY_asn1_find_str(ENGINE **pe,
+					const char *str, int len)
+	{
+	int i;
+	const EVP_PKEY_ASN1_METHOD *ameth;
+	if (len == -1)
+		len = strlen(str);
+	if (pe)
+		{
+#ifndef OPENSSL_NO_ENGINE
+		ENGINE *e;
+		ameth = ENGINE_pkey_asn1_find_str(&e, str, len);
+		if (ameth)
+			{
+			/* Convert structural into
+			 * functional reference
+			 */
+			if (!ENGINE_init(e))
+				ameth = NULL;
+			ENGINE_free(e);
+			*pe = e;
+			return ameth;
+			}
+#endif
+		*pe = NULL;
+		}
+	for (i = 0; i < EVP_PKEY_asn1_get_count(); i++)
+		{
+		ameth = EVP_PKEY_asn1_get0(i);
+		if (ameth->pkey_flags & ASN1_PKEY_ALIAS)
+			continue;
+		if (((int)strlen(ameth->pem_str) == len) && 
+			!strncasecmp(ameth->pem_str, str, len))
+			return ameth;
+		}
+	return NULL;
+	}
+
+int EVP_PKEY_asn1_add0(const EVP_PKEY_ASN1_METHOD *ameth)
+	{
+	if (app_methods == NULL)
+		{
+		app_methods = sk_EVP_PKEY_ASN1_METHOD_new(ameth_cmp);
+		if (!app_methods)
+			return 0;
+		}
+	if (!sk_EVP_PKEY_ASN1_METHOD_push(app_methods, ameth))
+		return 0;
+	sk_EVP_PKEY_ASN1_METHOD_sort(app_methods);
+	return 1;
+	}
+
+int EVP_PKEY_asn1_add_alias(int to, int from)
+	{
+	EVP_PKEY_ASN1_METHOD *ameth;
+	ameth = EVP_PKEY_asn1_new(from, ASN1_PKEY_ALIAS, NULL, NULL);
+	if (!ameth)
+		return 0;
+	ameth->pkey_base_id = to;
+	return EVP_PKEY_asn1_add0(ameth);
+	}
+
+int EVP_PKEY_asn1_get0_info(int *ppkey_id, int *ppkey_base_id, int *ppkey_flags,
+				const char **pinfo, const char **ppem_str,
+					const EVP_PKEY_ASN1_METHOD *ameth)
+	{
+	if (!ameth)
+		return 0;
+	if (ppkey_id)
+		*ppkey_id = ameth->pkey_id;
+	if (ppkey_base_id)
+		*ppkey_base_id = ameth->pkey_base_id;
+	if (ppkey_flags)
+		*ppkey_flags = ameth->pkey_flags;
+	if (pinfo)
+		*pinfo = ameth->info;
+	if (ppem_str)
+		*ppem_str = ameth->pem_str;
+	return 1;
+	}
+
+const EVP_PKEY_ASN1_METHOD* EVP_PKEY_get0_asn1(EVP_PKEY *pkey)
+	{
+	return pkey->ameth;
+	}
+
+EVP_PKEY_ASN1_METHOD* EVP_PKEY_asn1_new(int id, int flags,
+					const char *pem_str, const char *info)
+	{
+	EVP_PKEY_ASN1_METHOD *ameth;
+	ameth = OPENSSL_malloc(sizeof(EVP_PKEY_ASN1_METHOD));
+	if (!ameth)
+		return NULL;
+
+	memset(ameth, 0, sizeof(EVP_PKEY_ASN1_METHOD));
+
+	ameth->pkey_id = id;
+	ameth->pkey_base_id = id;
+	ameth->pkey_flags = flags | ASN1_PKEY_DYNAMIC;
+
+	if (info)
+		{
+		ameth->info = BUF_strdup(info);
+		if (!ameth->info)
+			goto err;
+		}
+	else
+		ameth->info = NULL;
+
+	if (pem_str)
+		{
+		ameth->pem_str = BUF_strdup(pem_str);
+		if (!ameth->pem_str)
+			goto err;
+		}
+	else
+		ameth->pem_str = NULL;
+
+	ameth->pub_decode = 0;
+	ameth->pub_encode = 0;
+	ameth->pub_cmp = 0;
+	ameth->pub_print = 0;
+
+	ameth->priv_decode = 0;
+	ameth->priv_encode = 0;
+	ameth->priv_print = 0;
+
+	ameth->old_priv_encode = 0;
+	ameth->old_priv_decode = 0;
+
+	ameth->item_verify = 0;
+	ameth->item_sign = 0;
+
+	ameth->pkey_size = 0;
+	ameth->pkey_bits = 0;
+
+	ameth->param_decode = 0;
+	ameth->param_encode = 0;
+	ameth->param_missing = 0;
+	ameth->param_copy = 0;
+	ameth->param_cmp = 0;
+	ameth->param_print = 0;
+
+	ameth->pkey_free = 0;
+	ameth->pkey_ctrl = 0;
+
+	return ameth;
+
+	err:
+
+	EVP_PKEY_asn1_free(ameth);
+	return NULL;
+
+	}
+
+void EVP_PKEY_asn1_copy(EVP_PKEY_ASN1_METHOD *dst, 
+			const EVP_PKEY_ASN1_METHOD *src)
+	{
+
+	dst->pub_decode = src->pub_decode;
+	dst->pub_encode = src->pub_encode;
+	dst->pub_cmp = src->pub_cmp;
+	dst->pub_print = src->pub_print;
+
+	dst->priv_decode = src->priv_decode;
+	dst->priv_encode = src->priv_encode;
+	dst->priv_print = src->priv_print;
+
+	dst->old_priv_encode = src->old_priv_encode;
+	dst->old_priv_decode = src->old_priv_decode;
+
+	dst->pkey_size = src->pkey_size;
+	dst->pkey_bits = src->pkey_bits;
+
+	dst->param_decode = src->param_decode;
+	dst->param_encode = src->param_encode;
+	dst->param_missing = src->param_missing;
+	dst->param_copy = src->param_copy;
+	dst->param_cmp = src->param_cmp;
+	dst->param_print = src->param_print;
+
+	dst->pkey_free = src->pkey_free;
+	dst->pkey_ctrl = src->pkey_ctrl;
+
+	dst->item_sign = src->item_sign;
+	dst->item_verify = src->item_verify;
+
+	}
+
+void EVP_PKEY_asn1_free(EVP_PKEY_ASN1_METHOD *ameth)
+	{
+	if (ameth && (ameth->pkey_flags & ASN1_PKEY_DYNAMIC))
+		{
+		if (ameth->pem_str)
+			OPENSSL_free(ameth->pem_str);
+		if (ameth->info)
+			OPENSSL_free(ameth->info);
+		OPENSSL_free(ameth);
+		}
+	}
+
+void EVP_PKEY_asn1_set_public(EVP_PKEY_ASN1_METHOD *ameth,
+		int (*pub_decode)(EVP_PKEY *pk, X509_PUBKEY *pub),
+		int (*pub_encode)(X509_PUBKEY *pub, const EVP_PKEY *pk),
+		int (*pub_cmp)(const EVP_PKEY *a, const EVP_PKEY *b),
+		int (*pub_print)(BIO *out, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *pctx),
+		int (*pkey_size)(const EVP_PKEY *pk),
+		int (*pkey_bits)(const EVP_PKEY *pk))
+	{
+	ameth->pub_decode = pub_decode;
+	ameth->pub_encode = pub_encode;
+	ameth->pub_cmp = pub_cmp;
+	ameth->pub_print = pub_print;
+	ameth->pkey_size = pkey_size;
+	ameth->pkey_bits = pkey_bits;
+	}
+
+void EVP_PKEY_asn1_set_private(EVP_PKEY_ASN1_METHOD *ameth,
+		int (*priv_decode)(EVP_PKEY *pk, PKCS8_PRIV_KEY_INFO *p8inf),
+		int (*priv_encode)(PKCS8_PRIV_KEY_INFO *p8, const EVP_PKEY *pk),
+		int (*priv_print)(BIO *out, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *pctx))
+	{
+	ameth->priv_decode = priv_decode;
+	ameth->priv_encode = priv_encode;
+	ameth->priv_print = priv_print;
+	}
+
+void EVP_PKEY_asn1_set_param(EVP_PKEY_ASN1_METHOD *ameth,
+		int (*param_decode)(EVP_PKEY *pkey,
+				const unsigned char **pder, int derlen),
+		int (*param_encode)(const EVP_PKEY *pkey, unsigned char **pder),
+		int (*param_missing)(const EVP_PKEY *pk),
+		int (*param_copy)(EVP_PKEY *to, const EVP_PKEY *from),
+		int (*param_cmp)(const EVP_PKEY *a, const EVP_PKEY *b),
+		int (*param_print)(BIO *out, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *pctx))
+	{
+	ameth->param_decode = param_decode;
+	ameth->param_encode = param_encode;
+	ameth->param_missing = param_missing;
+	ameth->param_copy = param_copy;
+	ameth->param_cmp = param_cmp;
+	ameth->param_print = param_print;
+	}
+
+void EVP_PKEY_asn1_set_free(EVP_PKEY_ASN1_METHOD *ameth,
+		void (*pkey_free)(EVP_PKEY *pkey))
+	{
+	ameth->pkey_free = pkey_free;
+	}
+
+void EVP_PKEY_asn1_set_ctrl(EVP_PKEY_ASN1_METHOD *ameth,
+		int (*pkey_ctrl)(EVP_PKEY *pkey, int op,
+							long arg1, void *arg2))
+	{
+	ameth->pkey_ctrl = pkey_ctrl;
+	}
diff --git a/crypto/asn1/asn1.h b/crypto/asn1/asn1.h
index d9d5443a33ca..220a0c8c637e 100644
--- a/crypto/asn1/asn1.h
+++ b/crypto/asn1/asn1.h
@@ -213,7 +213,7 @@ typedef struct asn1_object_st
 	const char *sn,*ln;
 	int nid;
 	int length;
-	unsigned char *data;
+	const unsigned char *data;	/* data remains const after init */
 	int flags;	/* Should we free this one */
 	} ASN1_OBJECT;
 
@@ -228,10 +228,14 @@ typedef struct asn1_object_st
  * complete and is a place holder for content when it had all been 
  * accessed. The flag will be reset when content has been written to it.
  */
-#define ASN1_STRING_FLAG_CONT 0x020 
 
+#define ASN1_STRING_FLAG_CONT 0x020 
+/* This flag is used by ASN1 code to indicate an ASN1_STRING is an MSTRING
+ * type.
+ */
+#define ASN1_STRING_FLAG_MSTRING 0x040 
 /* This is the base type that holds just about everything :-) */
-typedef struct asn1_string_st
+struct asn1_string_st
 	{
 	int length;
 	int type;
@@ -241,7 +245,7 @@ typedef struct asn1_string_st
 	 * input data has a non-zero 'unused bits' value, it will be
 	 * handled correctly */
 	long flags;
-	} ASN1_STRING;
+	};
 
 /* ASN1_ENCODING structure: this is used to save the received
  * encoding of an ASN1 type. This is useful to get round
@@ -289,7 +293,6 @@ DECLARE_STACK_OF(ASN1_STRING_TABLE)
  * see asn1t.h
  */
 typedef struct ASN1_TEMPLATE_st ASN1_TEMPLATE;
-typedef struct ASN1_ITEM_st ASN1_ITEM;
 typedef struct ASN1_TLC_st ASN1_TLC;
 /* This is just an opaque pointer */
 typedef struct ASN1_VALUE_st ASN1_VALUE;
@@ -330,6 +333,13 @@ typedef struct ASN1_VALUE_st ASN1_VALUE;
 	type *name##_new(void); \
 	void name##_free(type *a);
 
+#define DECLARE_ASN1_PRINT_FUNCTION(stname) \
+	DECLARE_ASN1_PRINT_FUNCTION_fname(stname, stname)
+
+#define DECLARE_ASN1_PRINT_FUNCTION_fname(stname, fname) \
+	int fname##_print_ctx(BIO *out, stname *x, int indent, \
+					 const ASN1_PCTX *pctx);
+
 #define D2I_OF(type) type *(*)(type **,const unsigned char **,long)
 #define I2D_OF(type) int (*)(type *,unsigned char **)
 #define I2D_OF_const(type) int (*)(const type *,unsigned char **)
@@ -344,8 +354,6 @@ typedef struct ASN1_VALUE_st ASN1_VALUE;
     ((void*) (1 ? p : (type*)0))
 #define CHECKED_PPTR_OF(type, p) \
     ((void**) (1 ? p : (type**)0))
-#define CHECKED_PTR_OF_TO_CHAR(type, p) \
-    ((char*) (1 ? p : (type*)0))
 
 #define TYPEDEF_D2I_OF(type) typedef type *d2i_of_##type(type **,const unsigned char **,long)
 #define TYPEDEF_I2D_OF(type) typedef int i2d_of_##type(type *,unsigned char **)
@@ -536,28 +544,23 @@ typedef struct asn1_type_st
 		 * contain the set or sequence bytes */
 		ASN1_STRING *		set;
 		ASN1_STRING *		sequence;
-		ASN1_VALUE  *		asn1_value;
+		ASN1_VALUE *		asn1_value;
 		} value;
 	} ASN1_TYPE;
 
 DECLARE_STACK_OF(ASN1_TYPE)
 DECLARE_ASN1_SET_OF(ASN1_TYPE)
 
-typedef struct asn1_method_st
-	{
-	i2d_of_void *i2d;
-	d2i_of_void *d2i;
-	void *(*create)(void);
-	void (*destroy)(void *);
-	} ASN1_METHOD;
-
-/* This is used when parsing some Netscape objects */
-typedef struct asn1_header_st
+typedef STACK_OF(ASN1_TYPE) ASN1_SEQUENCE_ANY;
+
+DECLARE_ASN1_ENCODE_FUNCTIONS_const(ASN1_SEQUENCE_ANY, ASN1_SEQUENCE_ANY)
+DECLARE_ASN1_ENCODE_FUNCTIONS_const(ASN1_SEQUENCE_ANY, ASN1_SET_ANY)
+
+typedef struct NETSCAPE_X509_st
 	{
 	ASN1_OCTET_STRING *header;
-	void *data;
-	ASN1_METHOD *meth;
-	} ASN1_HEADER;
+	X509 *cert;
+	} NETSCAPE_X509;
 
 /* This is used to contain a list of bit names */
 typedef struct BIT_STRING_BITNAME_st {
@@ -577,32 +580,34 @@ typedef struct BIT_STRING_BITNAME_st {
 		ASN1_STRING_type_new(V_ASN1_BIT_STRING)
 #define M_ASN1_BIT_STRING_free(a)	ASN1_STRING_free((ASN1_STRING *)a)
 #define M_ASN1_BIT_STRING_dup(a) (ASN1_BIT_STRING *)\
-		ASN1_STRING_dup((ASN1_STRING *)a)
+		ASN1_STRING_dup((const ASN1_STRING *)a)
 #define M_ASN1_BIT_STRING_cmp(a,b) ASN1_STRING_cmp(\
-		(ASN1_STRING *)a,(ASN1_STRING *)b)
+		(const ASN1_STRING *)a,(const ASN1_STRING *)b)
 #define M_ASN1_BIT_STRING_set(a,b,c) ASN1_STRING_set((ASN1_STRING *)a,b,c)
 
 #define M_ASN1_INTEGER_new()	(ASN1_INTEGER *)\
 		ASN1_STRING_type_new(V_ASN1_INTEGER)
 #define M_ASN1_INTEGER_free(a)		ASN1_STRING_free((ASN1_STRING *)a)
-#define M_ASN1_INTEGER_dup(a) (ASN1_INTEGER *)ASN1_STRING_dup((ASN1_STRING *)a)
+#define M_ASN1_INTEGER_dup(a) (ASN1_INTEGER *)\
+		ASN1_STRING_dup((const ASN1_STRING *)a)
 #define M_ASN1_INTEGER_cmp(a,b)	ASN1_STRING_cmp(\
-		(ASN1_STRING *)a,(ASN1_STRING *)b)
+		(const ASN1_STRING *)a,(const ASN1_STRING *)b)
 
 #define M_ASN1_ENUMERATED_new()	(ASN1_ENUMERATED *)\
 		ASN1_STRING_type_new(V_ASN1_ENUMERATED)
 #define M_ASN1_ENUMERATED_free(a)	ASN1_STRING_free((ASN1_STRING *)a)
-#define M_ASN1_ENUMERATED_dup(a) (ASN1_ENUMERATED *)ASN1_STRING_dup((ASN1_STRING *)a)
+#define M_ASN1_ENUMERATED_dup(a) (ASN1_ENUMERATED *)\
+		ASN1_STRING_dup((const ASN1_STRING *)a)
 #define M_ASN1_ENUMERATED_cmp(a,b)	ASN1_STRING_cmp(\
-		(ASN1_STRING *)a,(ASN1_STRING *)b)
+		(const ASN1_STRING *)a,(const ASN1_STRING *)b)
 
 #define M_ASN1_OCTET_STRING_new()	(ASN1_OCTET_STRING *)\
 		ASN1_STRING_type_new(V_ASN1_OCTET_STRING)
 #define M_ASN1_OCTET_STRING_free(a)	ASN1_STRING_free((ASN1_STRING *)a)
 #define M_ASN1_OCTET_STRING_dup(a) (ASN1_OCTET_STRING *)\
-		ASN1_STRING_dup((ASN1_STRING *)a)
+		ASN1_STRING_dup((const ASN1_STRING *)a)
 #define M_ASN1_OCTET_STRING_cmp(a,b) ASN1_STRING_cmp(\
-		(ASN1_STRING *)a,(ASN1_STRING *)b)
+		(const ASN1_STRING *)a,(const ASN1_STRING *)b)
 #define M_ASN1_OCTET_STRING_set(a,b,c)	ASN1_STRING_set((ASN1_STRING *)a,b,c)
 #define M_ASN1_OCTET_STRING_print(a,b)	ASN1_STRING_print(a,(ASN1_STRING *)b)
 #define M_i2d_ASN1_OCTET_STRING(a,pp) \
@@ -686,7 +691,7 @@ typedef struct BIT_STRING_BITNAME_st {
 		ASN1_STRING_type_new(V_ASN1_IA5STRING)
 #define M_ASN1_IA5STRING_free(a)	ASN1_STRING_free((ASN1_STRING *)a)
 #define M_ASN1_IA5STRING_dup(a)	\
-			(ASN1_IA5STRING *)ASN1_STRING_dup((ASN1_STRING *)a)
+		(ASN1_IA5STRING *)ASN1_STRING_dup((const ASN1_STRING *)a)
 #define M_i2d_ASN1_IA5STRING(a,pp) \
 		i2d_ASN1_bytes((ASN1_STRING *)a,pp,V_ASN1_IA5STRING,\
 			V_ASN1_UNIVERSAL)
@@ -697,18 +702,20 @@ typedef struct BIT_STRING_BITNAME_st {
 #define M_ASN1_UTCTIME_new()	(ASN1_UTCTIME *)\
 		ASN1_STRING_type_new(V_ASN1_UTCTIME)
 #define M_ASN1_UTCTIME_free(a)	ASN1_STRING_free((ASN1_STRING *)a)
-#define M_ASN1_UTCTIME_dup(a) (ASN1_UTCTIME *)ASN1_STRING_dup((ASN1_STRING *)a)
+#define M_ASN1_UTCTIME_dup(a) (ASN1_UTCTIME *)\
+		ASN1_STRING_dup((const ASN1_STRING *)a)
 
 #define M_ASN1_GENERALIZEDTIME_new()	(ASN1_GENERALIZEDTIME *)\
 		ASN1_STRING_type_new(V_ASN1_GENERALIZEDTIME)
 #define M_ASN1_GENERALIZEDTIME_free(a)	ASN1_STRING_free((ASN1_STRING *)a)
 #define M_ASN1_GENERALIZEDTIME_dup(a) (ASN1_GENERALIZEDTIME *)ASN1_STRING_dup(\
-	(ASN1_STRING *)a)
+	(const ASN1_STRING *)a)
 
 #define M_ASN1_TIME_new()	(ASN1_TIME *)\
 		ASN1_STRING_type_new(V_ASN1_UTCTIME)
 #define M_ASN1_TIME_free(a)	ASN1_STRING_free((ASN1_STRING *)a)
-#define M_ASN1_TIME_dup(a) (ASN1_TIME *)ASN1_STRING_dup((ASN1_STRING *)a)
+#define M_ASN1_TIME_dup(a) (ASN1_TIME *)\
+	ASN1_STRING_dup((const ASN1_STRING *)a)
 
 #define M_ASN1_GENERALSTRING_new()	(ASN1_GENERALSTRING *)\
 		ASN1_STRING_type_new(V_ASN1_GENERALSTRING)
@@ -769,6 +776,7 @@ DECLARE_ASN1_FUNCTIONS_fname(ASN1_TYPE, ASN1_ANY, ASN1_TYPE)
 int ASN1_TYPE_get(ASN1_TYPE *a);
 void ASN1_TYPE_set(ASN1_TYPE *a, int type, void *value);
 int ASN1_TYPE_set1(ASN1_TYPE *a, int type, const void *value);
+int            ASN1_TYPE_cmp(ASN1_TYPE *a, ASN1_TYPE *b);
 
 ASN1_OBJECT *	ASN1_OBJECT_new(void );
 void		ASN1_OBJECT_free(ASN1_OBJECT *a);
@@ -785,14 +793,15 @@ DECLARE_ASN1_SET_OF(ASN1_OBJECT)
 
 ASN1_STRING *	ASN1_STRING_new(void);
 void		ASN1_STRING_free(ASN1_STRING *a);
-ASN1_STRING *	ASN1_STRING_dup(ASN1_STRING *a);
+int		ASN1_STRING_copy(ASN1_STRING *dst, const ASN1_STRING *str);
+ASN1_STRING *	ASN1_STRING_dup(const ASN1_STRING *a);
 ASN1_STRING *	ASN1_STRING_type_new(int type );
-int 		ASN1_STRING_cmp(ASN1_STRING *a, ASN1_STRING *b);
+int 		ASN1_STRING_cmp(const ASN1_STRING *a, const ASN1_STRING *b);
   /* Since this is used to store all sorts of things, via macros, for now, make
      its data void * */
 int 		ASN1_STRING_set(ASN1_STRING *str, const void *data, int len);
 void		ASN1_STRING_set0(ASN1_STRING *str, void *data, int len);
-int ASN1_STRING_length(ASN1_STRING *x);
+int ASN1_STRING_length(const ASN1_STRING *x);
 void ASN1_STRING_length_set(ASN1_STRING *x, int n);
 int ASN1_STRING_type(ASN1_STRING *x);
 unsigned char * ASN1_STRING_data(ASN1_STRING *x);
@@ -805,6 +814,8 @@ int		ASN1_BIT_STRING_set(ASN1_BIT_STRING *a, unsigned char *d,
 			int length );
 int		ASN1_BIT_STRING_set_bit(ASN1_BIT_STRING *a, int n, int value);
 int		ASN1_BIT_STRING_get_bit(ASN1_BIT_STRING *a, int n);
+int            ASN1_BIT_STRING_check(ASN1_BIT_STRING *a,
+                                     unsigned char *flags, int flags_len);
 
 #ifndef OPENSSL_NO_BIO
 int ASN1_BIT_STRING_name_print(BIO *out, ASN1_BIT_STRING *bs,
@@ -823,13 +834,15 @@ ASN1_INTEGER *c2i_ASN1_INTEGER(ASN1_INTEGER **a,const unsigned char **pp,
 			long length);
 ASN1_INTEGER *d2i_ASN1_UINTEGER(ASN1_INTEGER **a,const unsigned char **pp,
 			long length);
-ASN1_INTEGER *	ASN1_INTEGER_dup(ASN1_INTEGER *x);
-int ASN1_INTEGER_cmp(ASN1_INTEGER *x, ASN1_INTEGER *y);
+ASN1_INTEGER *	ASN1_INTEGER_dup(const ASN1_INTEGER *x);
+int ASN1_INTEGER_cmp(const ASN1_INTEGER *x, const ASN1_INTEGER *y);
 
 DECLARE_ASN1_FUNCTIONS(ASN1_ENUMERATED)
 
 int ASN1_UTCTIME_check(ASN1_UTCTIME *a);
 ASN1_UTCTIME *ASN1_UTCTIME_set(ASN1_UTCTIME *s,time_t t);
+ASN1_UTCTIME *ASN1_UTCTIME_adj(ASN1_UTCTIME *s, time_t t,
+				int offset_day, long offset_sec);
 int ASN1_UTCTIME_set_string(ASN1_UTCTIME *s, const char *str);
 int ASN1_UTCTIME_cmp_time_t(const ASN1_UTCTIME *s, time_t t);
 #if 0
@@ -838,11 +851,13 @@ time_t ASN1_UTCTIME_get(const ASN1_UTCTIME *s);
 
 int ASN1_GENERALIZEDTIME_check(ASN1_GENERALIZEDTIME *a);
 ASN1_GENERALIZEDTIME *ASN1_GENERALIZEDTIME_set(ASN1_GENERALIZEDTIME *s,time_t t);
+ASN1_GENERALIZEDTIME *ASN1_GENERALIZEDTIME_adj(ASN1_GENERALIZEDTIME *s,
+	     time_t t, int offset_day, long offset_sec);
 int ASN1_GENERALIZEDTIME_set_string(ASN1_GENERALIZEDTIME *s, const char *str);
 
 DECLARE_ASN1_FUNCTIONS(ASN1_OCTET_STRING)
-ASN1_OCTET_STRING *	ASN1_OCTET_STRING_dup(ASN1_OCTET_STRING *a);
-int 	ASN1_OCTET_STRING_cmp(ASN1_OCTET_STRING *a, ASN1_OCTET_STRING *b);
+ASN1_OCTET_STRING *	ASN1_OCTET_STRING_dup(const ASN1_OCTET_STRING *a);
+int 	ASN1_OCTET_STRING_cmp(const ASN1_OCTET_STRING *a, const ASN1_OCTET_STRING *b);
 int 	ASN1_OCTET_STRING_set(ASN1_OCTET_STRING *str, const unsigned char *data, int len);
 
 DECLARE_ASN1_FUNCTIONS(ASN1_VISIBLESTRING)
@@ -869,14 +884,20 @@ DECLARE_ASN1_FUNCTIONS(ASN1_TIME)
 DECLARE_ASN1_ITEM(ASN1_OCTET_STRING_NDEF)
 
 ASN1_TIME *ASN1_TIME_set(ASN1_TIME *s,time_t t);
+ASN1_TIME *ASN1_TIME_adj(ASN1_TIME *s,time_t t,
+				int offset_day, long offset_sec);
 int ASN1_TIME_check(ASN1_TIME *t);
 ASN1_GENERALIZEDTIME *ASN1_TIME_to_generalizedtime(ASN1_TIME *t, ASN1_GENERALIZEDTIME **out);
+int ASN1_TIME_set_string(ASN1_TIME *s, const char *str);
 
-int i2d_ASN1_SET(STACK *a, unsigned char **pp,
-		 i2d_of_void *i2d, int ex_tag, int ex_class, int is_set);
-STACK *	d2i_ASN1_SET(STACK **a, const unsigned char **pp, long length,
-		     d2i_of_void *d2i, void (*free_func)(void *),
-		     int ex_tag, int ex_class);
+int i2d_ASN1_SET(STACK_OF(OPENSSL_BLOCK) *a, unsigned char **pp,
+		 i2d_of_void *i2d, int ex_tag, int ex_class,
+		 int is_set);
+STACK_OF(OPENSSL_BLOCK) *d2i_ASN1_SET(STACK_OF(OPENSSL_BLOCK) **a,
+			      const unsigned char **pp,
+			      long length, d2i_of_void *d2i,
+			      void (*free_func)(OPENSSL_BLOCK), int ex_tag,
+			      int ex_class);
 
 #ifndef OPENSSL_NO_BIO
 int i2a_ASN1_INTEGER(BIO *bp, ASN1_INTEGER *a);
@@ -894,9 +915,9 @@ ASN1_OBJECT *ASN1_OBJECT_create(int nid, unsigned char *data,int len,
 	const char *sn, const char *ln);
 
 int ASN1_INTEGER_set(ASN1_INTEGER *a, long v);
-long ASN1_INTEGER_get(ASN1_INTEGER *a);
-ASN1_INTEGER *BN_to_ASN1_INTEGER(BIGNUM *bn, ASN1_INTEGER *ai);
-BIGNUM *ASN1_INTEGER_to_BN(ASN1_INTEGER *ai,BIGNUM *bn);
+long ASN1_INTEGER_get(const ASN1_INTEGER *a);
+ASN1_INTEGER *BN_to_ASN1_INTEGER(const BIGNUM *bn, ASN1_INTEGER *ai);
+BIGNUM *ASN1_INTEGER_to_BN(const ASN1_INTEGER *ai,BIGNUM *bn);
 
 int ASN1_ENUMERATED_set(ASN1_ENUMERATED *a, long v);
 long ASN1_ENUMERATED_get(ASN1_ENUMERATED *a);
@@ -930,17 +951,17 @@ int ASN1_put_eoc(unsigned char **pp);
 int ASN1_object_size(int constructed, int length, int tag);
 
 /* Used to implement other functions */
-void *ASN1_dup(i2d_of_void *i2d, d2i_of_void *d2i, char *x);
+void *ASN1_dup(i2d_of_void *i2d, d2i_of_void *d2i, void *x);
 
 #define ASN1_dup_of(type,i2d,d2i,x) \
     ((type*)ASN1_dup(CHECKED_I2D_OF(type, i2d), \
 		     CHECKED_D2I_OF(type, d2i), \
-		     CHECKED_PTR_OF_TO_CHAR(type, x)))
+		     CHECKED_PTR_OF(type, x)))
 
 #define ASN1_dup_of_const(type,i2d,d2i,x) \
     ((type*)ASN1_dup(CHECKED_I2D_OF(const type, i2d), \
 		     CHECKED_D2I_OF(type, d2i), \
-		     CHECKED_PTR_OF_TO_CHAR(const type, x)))
+		     CHECKED_PTR_OF(const type, x)))
 
 void *ASN1_item_dup(const ASN1_ITEM *it, void *x);
 
@@ -1001,29 +1022,23 @@ int ASN1_i2d_bio(i2d_of_void *i2d,BIO *out, unsigned char *x);
 		  CHECKED_PTR_OF(const type, x)))
 
 int ASN1_item_i2d_bio(const ASN1_ITEM *it, BIO *out, void *x);
-int ASN1_UTCTIME_print(BIO *fp,ASN1_UTCTIME *a);
-int ASN1_GENERALIZEDTIME_print(BIO *fp,ASN1_GENERALIZEDTIME *a);
-int ASN1_TIME_print(BIO *fp,ASN1_TIME *a);
-int ASN1_STRING_print(BIO *bp,ASN1_STRING *v);
+int ASN1_UTCTIME_print(BIO *fp, const ASN1_UTCTIME *a);
+int ASN1_GENERALIZEDTIME_print(BIO *fp, const ASN1_GENERALIZEDTIME *a);
+int ASN1_TIME_print(BIO *fp, const ASN1_TIME *a);
+int ASN1_STRING_print(BIO *bp, const ASN1_STRING *v);
 int ASN1_STRING_print_ex(BIO *out, ASN1_STRING *str, unsigned long flags);
+int ASN1_bn_print(BIO *bp, const char *number, const BIGNUM *num,
+				unsigned char *buf, int off);
 int ASN1_parse(BIO *bp,const unsigned char *pp,long len,int indent);
 int ASN1_parse_dump(BIO *bp,const unsigned char *pp,long len,int indent,int dump);
 #endif
 const char *ASN1_tag2str(int tag);
 
-/* Used to load and write netscape format cert/key */
-int i2d_ASN1_HEADER(ASN1_HEADER *a,unsigned char **pp);
-ASN1_HEADER *d2i_ASN1_HEADER(ASN1_HEADER **a,const unsigned char **pp, long length);
-ASN1_HEADER *ASN1_HEADER_new(void );
-void ASN1_HEADER_free(ASN1_HEADER *a);
+/* Used to load and write netscape format cert */
 
-int ASN1_UNIVERSALSTRING_to_string(ASN1_UNIVERSALSTRING *s);
+DECLARE_ASN1_FUNCTIONS(NETSCAPE_X509)
 
-/* Not used that much at this point, except for the first two */
-ASN1_METHOD *X509_asn1_meth(void);
-ASN1_METHOD *RSAPrivateKey_asn1_meth(void);
-ASN1_METHOD *ASN1_IA5STRING_asn1_meth(void);
-ASN1_METHOD *ASN1_BIT_STRING_asn1_meth(void);
+int ASN1_UNIVERSALSTRING_to_string(ASN1_UNIVERSALSTRING *s);
 
 int ASN1_TYPE_set_octetstring(ASN1_TYPE *a,
 	unsigned char *data, int len);
@@ -1034,9 +1049,9 @@ int ASN1_TYPE_set_int_octetstring(ASN1_TYPE *a, long num,
 int ASN1_TYPE_get_int_octetstring(ASN1_TYPE *a,long *num,
 	unsigned char *data, int max_len);
 
-STACK *ASN1_seq_unpack(const unsigned char *buf, int len,
-		       d2i_of_void *d2i, void (*free_func)(void *));
-unsigned char *ASN1_seq_pack(STACK *safes, i2d_of_void *i2d,
+STACK_OF(OPENSSL_BLOCK) *ASN1_seq_unpack(const unsigned char *buf, int len,
+				 d2i_of_void *d2i, void (*free_func)(OPENSSL_BLOCK));
+unsigned char *ASN1_seq_pack(STACK_OF(OPENSSL_BLOCK) *safes, i2d_of_void *i2d,
 			     unsigned char **buf, int *len );
 void *ASN1_unpack_string(ASN1_STRING *oct, d2i_of_void *d2i);
 void *ASN1_item_unpack(ASN1_STRING *oct, const ASN1_ITEM *it);
@@ -1079,15 +1094,58 @@ void ASN1_add_oid_module(void);
 ASN1_TYPE *ASN1_generate_nconf(char *str, CONF *nconf);
 ASN1_TYPE *ASN1_generate_v3(char *str, X509V3_CTX *cnf);
 
-typedef int asn1_output_data_fn(BIO *out, BIO *data, ASN1_VALUE *val, int flags,
-					const ASN1_ITEM *it);
-
-int int_smime_write_ASN1(BIO *bio, ASN1_VALUE *val, BIO *data, int flags,
+/* ASN1 Print flags */
+
+/* Indicate missing OPTIONAL fields */
+#define ASN1_PCTX_FLAGS_SHOW_ABSENT		0x001	
+/* Mark start and end of SEQUENCE */
+#define ASN1_PCTX_FLAGS_SHOW_SEQUENCE		0x002
+/* Mark start and end of SEQUENCE/SET OF */
+#define ASN1_PCTX_FLAGS_SHOW_SSOF		0x004
+/* Show the ASN1 type of primitives */
+#define ASN1_PCTX_FLAGS_SHOW_TYPE		0x008
+/* Don't show ASN1 type of ANY */
+#define ASN1_PCTX_FLAGS_NO_ANY_TYPE		0x010
+/* Don't show ASN1 type of MSTRINGs */
+#define ASN1_PCTX_FLAGS_NO_MSTRING_TYPE		0x020
+/* Don't show field names in SEQUENCE */
+#define ASN1_PCTX_FLAGS_NO_FIELD_NAME		0x040
+/* Show structure names of each SEQUENCE field */
+#define ASN1_PCTX_FLAGS_SHOW_FIELD_STRUCT_NAME	0x080
+/* Don't show structure name even at top level */
+#define ASN1_PCTX_FLAGS_NO_STRUCT_NAME		0x100
+
+int ASN1_item_print(BIO *out, ASN1_VALUE *ifld, int indent,
+				const ASN1_ITEM *it, const ASN1_PCTX *pctx);
+ASN1_PCTX *ASN1_PCTX_new(void);
+void ASN1_PCTX_free(ASN1_PCTX *p);
+unsigned long ASN1_PCTX_get_flags(ASN1_PCTX *p);
+void ASN1_PCTX_set_flags(ASN1_PCTX *p, unsigned long flags);
+unsigned long ASN1_PCTX_get_nm_flags(ASN1_PCTX *p);
+void ASN1_PCTX_set_nm_flags(ASN1_PCTX *p, unsigned long flags);
+unsigned long ASN1_PCTX_get_cert_flags(ASN1_PCTX *p);
+void ASN1_PCTX_set_cert_flags(ASN1_PCTX *p, unsigned long flags);
+unsigned long ASN1_PCTX_get_oid_flags(ASN1_PCTX *p);
+void ASN1_PCTX_set_oid_flags(ASN1_PCTX *p, unsigned long flags);
+unsigned long ASN1_PCTX_get_str_flags(ASN1_PCTX *p);
+void ASN1_PCTX_set_str_flags(ASN1_PCTX *p, unsigned long flags);
+
+BIO_METHOD *BIO_f_asn1(void);
+
+BIO *BIO_new_NDEF(BIO *out, ASN1_VALUE *val, const ASN1_ITEM *it);
+
+int i2d_ASN1_bio_stream(BIO *out, ASN1_VALUE *val, BIO *in, int flags,
+				const ASN1_ITEM *it);
+int PEM_write_bio_ASN1_stream(BIO *out, ASN1_VALUE *val, BIO *in, int flags,
+				const char *hdr,
+				const ASN1_ITEM *it);
+int SMIME_write_ASN1(BIO *bio, ASN1_VALUE *val, BIO *data, int flags,
 				int ctype_nid, int econt_nid,
 				STACK_OF(X509_ALGOR) *mdalgs,
-				asn1_output_data_fn *data_fn,
 				const ASN1_ITEM *it);
 ASN1_VALUE *SMIME_read_ASN1(BIO *bio, BIO **bcont, const ASN1_ITEM *it);
+int SMIME_crlf_copy(BIO *in, BIO *out, int flags);
+int SMIME_text(BIO *in, BIO *out);
 
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
@@ -1118,6 +1176,7 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_F_ASN1_ENUMERATED_TO_BN			 113
 #define ASN1_F_ASN1_EX_C2I				 204
 #define ASN1_F_ASN1_FIND_END				 190
+#define ASN1_F_ASN1_GENERALIZEDTIME_ADJ			 216
 #define ASN1_F_ASN1_GENERALIZEDTIME_SET			 185
 #define ASN1_F_ASN1_GENERATE_V3				 178
 #define ASN1_F_ASN1_GET_OBJECT				 114
@@ -1134,11 +1193,12 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_F_ASN1_ITEM_I2D_FP				 193
 #define ASN1_F_ASN1_ITEM_PACK				 198
 #define ASN1_F_ASN1_ITEM_SIGN				 195
+#define ASN1_F_ASN1_ITEM_SIGN_CTX			 220
 #define ASN1_F_ASN1_ITEM_UNPACK				 199
 #define ASN1_F_ASN1_ITEM_VERIFY				 197
 #define ASN1_F_ASN1_MBSTRING_NCOPY			 122
 #define ASN1_F_ASN1_OBJECT_NEW				 123
-#define ASN1_F_ASN1_OUTPUT_DATA				 207
+#define ASN1_F_ASN1_OUTPUT_DATA				 214
 #define ASN1_F_ASN1_PACK_STRING				 124
 #define ASN1_F_ASN1_PCTX_NEW				 205
 #define ASN1_F_ASN1_PKCS5_PBE_SET			 125
@@ -1152,14 +1212,17 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_F_ASN1_TEMPLATE_EX_D2I			 132
 #define ASN1_F_ASN1_TEMPLATE_NEW			 133
 #define ASN1_F_ASN1_TEMPLATE_NOEXP_D2I			 131
+#define ASN1_F_ASN1_TIME_ADJ				 217
 #define ASN1_F_ASN1_TIME_SET				 175
 #define ASN1_F_ASN1_TYPE_GET_INT_OCTETSTRING		 134
 #define ASN1_F_ASN1_TYPE_GET_OCTETSTRING		 135
 #define ASN1_F_ASN1_UNPACK_STRING			 136
+#define ASN1_F_ASN1_UTCTIME_ADJ				 218
 #define ASN1_F_ASN1_UTCTIME_SET				 187
 #define ASN1_F_ASN1_VERIFY				 137
-#define ASN1_F_B64_READ_ASN1				 208
-#define ASN1_F_B64_WRITE_ASN1				 209
+#define ASN1_F_B64_READ_ASN1				 209
+#define ASN1_F_B64_WRITE_ASN1				 210
+#define ASN1_F_BIO_NEW_NDEF				 208
 #define ASN1_F_BITSTR_CB				 180
 #define ASN1_F_BN_TO_ASN1_ENUMERATED			 138
 #define ASN1_F_BN_TO_ASN1_INTEGER			 139
@@ -1178,6 +1241,7 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_F_D2I_ASN1_TYPE_BYTES			 149
 #define ASN1_F_D2I_ASN1_UINTEGER			 150
 #define ASN1_F_D2I_ASN1_UTCTIME				 151
+#define ASN1_F_D2I_AUTOPRIVATEKEY			 207
 #define ASN1_F_D2I_NETSCAPE_RSA				 152
 #define ASN1_F_D2I_NETSCAPE_RSA_2			 153
 #define ASN1_F_D2I_PRIVATEKEY				 154
@@ -1187,6 +1251,7 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_F_D2I_X509					 156
 #define ASN1_F_D2I_X509_CINF				 157
 #define ASN1_F_D2I_X509_PKEY				 159
+#define ASN1_F_I2D_ASN1_BIO_STREAM			 211
 #define ASN1_F_I2D_ASN1_SET				 188
 #define ASN1_F_I2D_ASN1_TIME				 160
 #define ASN1_F_I2D_DSA_PUBKEY				 161
@@ -1198,10 +1263,12 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_F_LONG_C2I					 166
 #define ASN1_F_OID_MODULE_INIT				 174
 #define ASN1_F_PARSE_TAGGING				 182
-#define ASN1_F_PKCS5_PBE2_SET				 167
+#define ASN1_F_PKCS5_PBE2_SET_IV			 167
 #define ASN1_F_PKCS5_PBE_SET				 202
-#define ASN1_F_SMIME_READ_ASN1				 210
-#define ASN1_F_SMIME_TEXT				 211
+#define ASN1_F_PKCS5_PBE_SET0_ALGOR			 215
+#define ASN1_F_PKCS5_PBKDF2_SET				 219
+#define ASN1_F_SMIME_READ_ASN1				 212
+#define ASN1_F_SMIME_TEXT				 213
 #define ASN1_F_X509_CINF_NEW				 168
 #define ASN1_F_X509_CRL_ADD0_REVOKED			 169
 #define ASN1_F_X509_INFO_NEW				 170
@@ -1213,22 +1280,24 @@ void ERR_load_ASN1_strings(void);
 
 /* Reason codes. */
 #define ASN1_R_ADDING_OBJECT				 171
-#define ASN1_R_ASN1_PARSE_ERROR				 198
-#define ASN1_R_ASN1_SIG_PARSE_ERROR			 199
+#define ASN1_R_ASN1_PARSE_ERROR				 203
+#define ASN1_R_ASN1_SIG_PARSE_ERROR			 204
 #define ASN1_R_AUX_ERROR				 100
 #define ASN1_R_BAD_CLASS				 101
 #define ASN1_R_BAD_OBJECT_HEADER			 102
 #define ASN1_R_BAD_PASSWORD_READ			 103
 #define ASN1_R_BAD_TAG					 104
-#define ASN1_R_BMPSTRING_IS_WRONG_LENGTH		 210
+#define ASN1_R_BMPSTRING_IS_WRONG_LENGTH		 214
 #define ASN1_R_BN_LIB					 105
 #define ASN1_R_BOOLEAN_IS_WRONG_LENGTH			 106
 #define ASN1_R_BUFFER_TOO_SMALL				 107
 #define ASN1_R_CIPHER_HAS_NO_OBJECT_IDENTIFIER		 108
+#define ASN1_R_CONTEXT_NOT_INITIALISED			 217
 #define ASN1_R_DATA_IS_WRONG				 109
 #define ASN1_R_DECODE_ERROR				 110
 #define ASN1_R_DECODING_ERROR				 111
 #define ASN1_R_DEPTH_EXCEEDED				 174
+#define ASN1_R_DIGEST_AND_KEY_TYPE_NOT_SUPPORTED	 198
 #define ASN1_R_ENCODE_ERROR				 112
 #define ASN1_R_ERROR_GETTING_TIME			 173
 #define ASN1_R_ERROR_LOADING_SECTION			 172
@@ -1262,10 +1331,10 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_R_INTEGER_TOO_LARGE_FOR_LONG		 128
 #define ASN1_R_INVALID_BMPSTRING_LENGTH			 129
 #define ASN1_R_INVALID_DIGIT				 130
-#define ASN1_R_INVALID_MIME_TYPE			 200
+#define ASN1_R_INVALID_MIME_TYPE			 205
 #define ASN1_R_INVALID_MODIFIER				 186
 #define ASN1_R_INVALID_NUMBER				 187
-#define ASN1_R_INVALID_OBJECT_ENCODING			 212
+#define ASN1_R_INVALID_OBJECT_ENCODING			 216
 #define ASN1_R_INVALID_SEPARATOR			 131
 #define ASN1_R_INVALID_TIME_FORMAT			 132
 #define ASN1_R_INVALID_UNIVERSALSTRING_LENGTH		 133
@@ -1273,9 +1342,9 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_R_IV_TOO_LARGE				 135
 #define ASN1_R_LENGTH_ERROR				 136
 #define ASN1_R_LIST_ERROR				 188
-#define ASN1_R_MIME_NO_CONTENT_TYPE			 201
-#define ASN1_R_MIME_PARSE_ERROR				 202
-#define ASN1_R_MIME_SIG_PARSE_ERROR			 203
+#define ASN1_R_MIME_NO_CONTENT_TYPE			 206
+#define ASN1_R_MIME_PARSE_ERROR				 207
+#define ASN1_R_MIME_SIG_PARSE_ERROR			 208
 #define ASN1_R_MISSING_EOC				 137
 #define ASN1_R_MISSING_SECOND_NUMBER			 138
 #define ASN1_R_MISSING_VALUE				 189
@@ -1285,11 +1354,12 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_R_NON_HEX_CHARACTERS			 141
 #define ASN1_R_NOT_ASCII_FORMAT				 190
 #define ASN1_R_NOT_ENOUGH_DATA				 142
-#define ASN1_R_NO_CONTENT_TYPE				 204
+#define ASN1_R_NO_CONTENT_TYPE				 209
+#define ASN1_R_NO_DEFAULT_DIGEST			 201
 #define ASN1_R_NO_MATCHING_CHOICE_TYPE			 143
-#define ASN1_R_NO_MULTIPART_BODY_FAILURE		 205
-#define ASN1_R_NO_MULTIPART_BOUNDARY			 206
-#define ASN1_R_NO_SIG_CONTENT_TYPE			 207
+#define ASN1_R_NO_MULTIPART_BODY_FAILURE		 210
+#define ASN1_R_NO_MULTIPART_BOUNDARY			 211
+#define ASN1_R_NO_SIG_CONTENT_TYPE			 212
 #define ASN1_R_NULL_IS_WRONG_LENGTH			 144
 #define ASN1_R_OBJECT_NOT_ASCII_FORMAT			 191
 #define ASN1_R_ODD_NUMBER_OF_CHARS			 145
@@ -1299,8 +1369,8 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_R_SEQUENCE_NOT_CONSTRUCTED			 149
 #define ASN1_R_SEQUENCE_OR_SET_NEEDS_CONFIG		 192
 #define ASN1_R_SHORT_LINE				 150
-#define ASN1_R_SIG_INVALID_MIME_TYPE			 208
-#define ASN1_R_STREAMING_NOT_SUPPORTED			 209
+#define ASN1_R_SIG_INVALID_MIME_TYPE			 213
+#define ASN1_R_STREAMING_NOT_SUPPORTED			 202
 #define ASN1_R_STRING_TOO_LONG				 151
 #define ASN1_R_STRING_TOO_SHORT				 152
 #define ASN1_R_TAG_VALUE_TOO_HIGH			 153
@@ -1311,11 +1381,12 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_R_UNABLE_TO_DECODE_RSA_KEY			 157
 #define ASN1_R_UNABLE_TO_DECODE_RSA_PRIVATE_KEY		 158
 #define ASN1_R_UNEXPECTED_EOC				 159
-#define ASN1_R_UNIVERSALSTRING_IS_WRONG_LENGTH		 211
+#define ASN1_R_UNIVERSALSTRING_IS_WRONG_LENGTH		 215
 #define ASN1_R_UNKNOWN_FORMAT				 160
 #define ASN1_R_UNKNOWN_MESSAGE_DIGEST_ALGORITHM		 161
 #define ASN1_R_UNKNOWN_OBJECT_TYPE			 162
 #define ASN1_R_UNKNOWN_PUBLIC_KEY_TYPE			 163
+#define ASN1_R_UNKNOWN_SIGNATURE_ALGORITHM		 199
 #define ASN1_R_UNKNOWN_TAG				 194
 #define ASN1_R_UNKOWN_FORMAT				 195
 #define ASN1_R_UNSUPPORTED_ANY_DEFINED_BY_TYPE		 164
@@ -1323,6 +1394,7 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_R_UNSUPPORTED_ENCRYPTION_ALGORITHM		 166
 #define ASN1_R_UNSUPPORTED_PUBLIC_KEY_TYPE		 167
 #define ASN1_R_UNSUPPORTED_TYPE				 196
+#define ASN1_R_WRONG_PUBLIC_KEY_TYPE			 200
 #define ASN1_R_WRONG_TAG				 168
 #define ASN1_R_WRONG_TYPE				 169
 
diff --git a/crypto/asn1/asn1_err.c b/crypto/asn1/asn1_err.c
index ba88eb318be3..1a30bf119bd2 100644
--- a/crypto/asn1/asn1_err.c
+++ b/crypto/asn1/asn1_err.c
@@ -1,6 +1,6 @@
 /* crypto/asn1/asn1_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2008 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -90,10 +90,11 @@ static ERR_STRING_DATA ASN1_str_functs[]=
 {ERR_FUNC(ASN1_F_ASN1_ENUMERATED_TO_BN),	"ASN1_ENUMERATED_to_BN"},
 {ERR_FUNC(ASN1_F_ASN1_EX_C2I),	"ASN1_EX_C2I"},
 {ERR_FUNC(ASN1_F_ASN1_FIND_END),	"ASN1_FIND_END"},
+{ERR_FUNC(ASN1_F_ASN1_GENERALIZEDTIME_ADJ),	"ASN1_GENERALIZEDTIME_adj"},
 {ERR_FUNC(ASN1_F_ASN1_GENERALIZEDTIME_SET),	"ASN1_GENERALIZEDTIME_set"},
 {ERR_FUNC(ASN1_F_ASN1_GENERATE_V3),	"ASN1_generate_v3"},
 {ERR_FUNC(ASN1_F_ASN1_GET_OBJECT),	"ASN1_get_object"},
-{ERR_FUNC(ASN1_F_ASN1_HEADER_NEW),	"ASN1_HEADER_new"},
+{ERR_FUNC(ASN1_F_ASN1_HEADER_NEW),	"ASN1_HEADER_NEW"},
 {ERR_FUNC(ASN1_F_ASN1_I2D_BIO),	"ASN1_i2d_bio"},
 {ERR_FUNC(ASN1_F_ASN1_I2D_FP),	"ASN1_i2d_fp"},
 {ERR_FUNC(ASN1_F_ASN1_INTEGER_SET),	"ASN1_INTEGER_set"},
@@ -106,13 +107,14 @@ static ERR_STRING_DATA ASN1_str_functs[]=
 {ERR_FUNC(ASN1_F_ASN1_ITEM_I2D_FP),	"ASN1_item_i2d_fp"},
 {ERR_FUNC(ASN1_F_ASN1_ITEM_PACK),	"ASN1_item_pack"},
 {ERR_FUNC(ASN1_F_ASN1_ITEM_SIGN),	"ASN1_item_sign"},
+{ERR_FUNC(ASN1_F_ASN1_ITEM_SIGN_CTX),	"ASN1_item_sign_ctx"},
 {ERR_FUNC(ASN1_F_ASN1_ITEM_UNPACK),	"ASN1_item_unpack"},
 {ERR_FUNC(ASN1_F_ASN1_ITEM_VERIFY),	"ASN1_item_verify"},
 {ERR_FUNC(ASN1_F_ASN1_MBSTRING_NCOPY),	"ASN1_mbstring_ncopy"},
 {ERR_FUNC(ASN1_F_ASN1_OBJECT_NEW),	"ASN1_OBJECT_new"},
 {ERR_FUNC(ASN1_F_ASN1_OUTPUT_DATA),	"ASN1_OUTPUT_DATA"},
 {ERR_FUNC(ASN1_F_ASN1_PACK_STRING),	"ASN1_pack_string"},
-{ERR_FUNC(ASN1_F_ASN1_PCTX_NEW),	"ASN1_PCTX_NEW"},
+{ERR_FUNC(ASN1_F_ASN1_PCTX_NEW),	"ASN1_PCTX_new"},
 {ERR_FUNC(ASN1_F_ASN1_PKCS5_PBE_SET),	"ASN1_PKCS5_PBE_SET"},
 {ERR_FUNC(ASN1_F_ASN1_SEQ_PACK),	"ASN1_seq_pack"},
 {ERR_FUNC(ASN1_F_ASN1_SEQ_UNPACK),	"ASN1_seq_unpack"},
@@ -124,14 +126,17 @@ static ERR_STRING_DATA ASN1_str_functs[]=
 {ERR_FUNC(ASN1_F_ASN1_TEMPLATE_EX_D2I),	"ASN1_TEMPLATE_EX_D2I"},
 {ERR_FUNC(ASN1_F_ASN1_TEMPLATE_NEW),	"ASN1_TEMPLATE_NEW"},
 {ERR_FUNC(ASN1_F_ASN1_TEMPLATE_NOEXP_D2I),	"ASN1_TEMPLATE_NOEXP_D2I"},
+{ERR_FUNC(ASN1_F_ASN1_TIME_ADJ),	"ASN1_TIME_adj"},
 {ERR_FUNC(ASN1_F_ASN1_TIME_SET),	"ASN1_TIME_set"},
 {ERR_FUNC(ASN1_F_ASN1_TYPE_GET_INT_OCTETSTRING),	"ASN1_TYPE_get_int_octetstring"},
 {ERR_FUNC(ASN1_F_ASN1_TYPE_GET_OCTETSTRING),	"ASN1_TYPE_get_octetstring"},
 {ERR_FUNC(ASN1_F_ASN1_UNPACK_STRING),	"ASN1_unpack_string"},
+{ERR_FUNC(ASN1_F_ASN1_UTCTIME_ADJ),	"ASN1_UTCTIME_adj"},
 {ERR_FUNC(ASN1_F_ASN1_UTCTIME_SET),	"ASN1_UTCTIME_set"},
 {ERR_FUNC(ASN1_F_ASN1_VERIFY),	"ASN1_verify"},
 {ERR_FUNC(ASN1_F_B64_READ_ASN1),	"B64_READ_ASN1"},
 {ERR_FUNC(ASN1_F_B64_WRITE_ASN1),	"B64_WRITE_ASN1"},
+{ERR_FUNC(ASN1_F_BIO_NEW_NDEF),	"BIO_new_NDEF"},
 {ERR_FUNC(ASN1_F_BITSTR_CB),	"BITSTR_CB"},
 {ERR_FUNC(ASN1_F_BN_TO_ASN1_ENUMERATED),	"BN_to_ASN1_ENUMERATED"},
 {ERR_FUNC(ASN1_F_BN_TO_ASN1_INTEGER),	"BN_to_ASN1_INTEGER"},
@@ -143,13 +148,14 @@ static ERR_STRING_DATA ASN1_str_functs[]=
 {ERR_FUNC(ASN1_F_D2I_ASN1_BOOLEAN),	"d2i_ASN1_BOOLEAN"},
 {ERR_FUNC(ASN1_F_D2I_ASN1_BYTES),	"d2i_ASN1_bytes"},
 {ERR_FUNC(ASN1_F_D2I_ASN1_GENERALIZEDTIME),	"D2I_ASN1_GENERALIZEDTIME"},
-{ERR_FUNC(ASN1_F_D2I_ASN1_HEADER),	"d2i_ASN1_HEADER"},
+{ERR_FUNC(ASN1_F_D2I_ASN1_HEADER),	"D2I_ASN1_HEADER"},
 {ERR_FUNC(ASN1_F_D2I_ASN1_INTEGER),	"D2I_ASN1_INTEGER"},
 {ERR_FUNC(ASN1_F_D2I_ASN1_OBJECT),	"d2i_ASN1_OBJECT"},
 {ERR_FUNC(ASN1_F_D2I_ASN1_SET),	"d2i_ASN1_SET"},
 {ERR_FUNC(ASN1_F_D2I_ASN1_TYPE_BYTES),	"d2i_ASN1_type_bytes"},
 {ERR_FUNC(ASN1_F_D2I_ASN1_UINTEGER),	"d2i_ASN1_UINTEGER"},
 {ERR_FUNC(ASN1_F_D2I_ASN1_UTCTIME),	"D2I_ASN1_UTCTIME"},
+{ERR_FUNC(ASN1_F_D2I_AUTOPRIVATEKEY),	"d2i_AutoPrivateKey"},
 {ERR_FUNC(ASN1_F_D2I_NETSCAPE_RSA),	"d2i_Netscape_RSA"},
 {ERR_FUNC(ASN1_F_D2I_NETSCAPE_RSA_2),	"D2I_NETSCAPE_RSA_2"},
 {ERR_FUNC(ASN1_F_D2I_PRIVATEKEY),	"d2i_PrivateKey"},
@@ -159,6 +165,7 @@ static ERR_STRING_DATA ASN1_str_functs[]=
 {ERR_FUNC(ASN1_F_D2I_X509),	"D2I_X509"},
 {ERR_FUNC(ASN1_F_D2I_X509_CINF),	"D2I_X509_CINF"},
 {ERR_FUNC(ASN1_F_D2I_X509_PKEY),	"d2i_X509_PKEY"},
+{ERR_FUNC(ASN1_F_I2D_ASN1_BIO_STREAM),	"i2d_ASN1_bio_stream"},
 {ERR_FUNC(ASN1_F_I2D_ASN1_SET),	"i2d_ASN1_SET"},
 {ERR_FUNC(ASN1_F_I2D_ASN1_TIME),	"I2D_ASN1_TIME"},
 {ERR_FUNC(ASN1_F_I2D_DSA_PUBKEY),	"i2d_DSA_PUBKEY"},
@@ -170,8 +177,10 @@ static ERR_STRING_DATA ASN1_str_functs[]=
 {ERR_FUNC(ASN1_F_LONG_C2I),	"LONG_C2I"},
 {ERR_FUNC(ASN1_F_OID_MODULE_INIT),	"OID_MODULE_INIT"},
 {ERR_FUNC(ASN1_F_PARSE_TAGGING),	"PARSE_TAGGING"},
-{ERR_FUNC(ASN1_F_PKCS5_PBE2_SET),	"PKCS5_pbe2_set"},
+{ERR_FUNC(ASN1_F_PKCS5_PBE2_SET_IV),	"PKCS5_pbe2_set_iv"},
 {ERR_FUNC(ASN1_F_PKCS5_PBE_SET),	"PKCS5_pbe_set"},
+{ERR_FUNC(ASN1_F_PKCS5_PBE_SET0_ALGOR),	"PKCS5_pbe_set0_algor"},
+{ERR_FUNC(ASN1_F_PKCS5_PBKDF2_SET),	"PKCS5_pbkdf2_set"},
 {ERR_FUNC(ASN1_F_SMIME_READ_ASN1),	"SMIME_read_ASN1"},
 {ERR_FUNC(ASN1_F_SMIME_TEXT),	"SMIME_text"},
 {ERR_FUNC(ASN1_F_X509_CINF_NEW),	"X509_CINF_NEW"},
@@ -200,10 +209,12 @@ static ERR_STRING_DATA ASN1_str_reasons[]=
 {ERR_REASON(ASN1_R_BOOLEAN_IS_WRONG_LENGTH),"boolean is wrong length"},
 {ERR_REASON(ASN1_R_BUFFER_TOO_SMALL)     ,"buffer too small"},
 {ERR_REASON(ASN1_R_CIPHER_HAS_NO_OBJECT_IDENTIFIER),"cipher has no object identifier"},
+{ERR_REASON(ASN1_R_CONTEXT_NOT_INITIALISED),"context not initialised"},
 {ERR_REASON(ASN1_R_DATA_IS_WRONG)        ,"data is wrong"},
 {ERR_REASON(ASN1_R_DECODE_ERROR)         ,"decode error"},
 {ERR_REASON(ASN1_R_DECODING_ERROR)       ,"decoding error"},
 {ERR_REASON(ASN1_R_DEPTH_EXCEEDED)       ,"depth exceeded"},
+{ERR_REASON(ASN1_R_DIGEST_AND_KEY_TYPE_NOT_SUPPORTED),"digest and key type not supported"},
 {ERR_REASON(ASN1_R_ENCODE_ERROR)         ,"encode error"},
 {ERR_REASON(ASN1_R_ERROR_GETTING_TIME)   ,"error getting time"},
 {ERR_REASON(ASN1_R_ERROR_LOADING_SECTION),"error loading section"},
@@ -261,6 +272,7 @@ static ERR_STRING_DATA ASN1_str_reasons[]=
 {ERR_REASON(ASN1_R_NOT_ASCII_FORMAT)     ,"not ascii format"},
 {ERR_REASON(ASN1_R_NOT_ENOUGH_DATA)      ,"not enough data"},
 {ERR_REASON(ASN1_R_NO_CONTENT_TYPE)      ,"no content type"},
+{ERR_REASON(ASN1_R_NO_DEFAULT_DIGEST)    ,"no default digest"},
 {ERR_REASON(ASN1_R_NO_MATCHING_CHOICE_TYPE),"no matching choice type"},
 {ERR_REASON(ASN1_R_NO_MULTIPART_BODY_FAILURE),"no multipart body failure"},
 {ERR_REASON(ASN1_R_NO_MULTIPART_BOUNDARY),"no multipart boundary"},
@@ -291,6 +303,7 @@ static ERR_STRING_DATA ASN1_str_reasons[]=
 {ERR_REASON(ASN1_R_UNKNOWN_MESSAGE_DIGEST_ALGORITHM),"unknown message digest algorithm"},
 {ERR_REASON(ASN1_R_UNKNOWN_OBJECT_TYPE)  ,"unknown object type"},
 {ERR_REASON(ASN1_R_UNKNOWN_PUBLIC_KEY_TYPE),"unknown public key type"},
+{ERR_REASON(ASN1_R_UNKNOWN_SIGNATURE_ALGORITHM),"unknown signature algorithm"},
 {ERR_REASON(ASN1_R_UNKNOWN_TAG)          ,"unknown tag"},
 {ERR_REASON(ASN1_R_UNKOWN_FORMAT)        ,"unkown format"},
 {ERR_REASON(ASN1_R_UNSUPPORTED_ANY_DEFINED_BY_TYPE),"unsupported any defined by type"},
@@ -298,6 +311,7 @@ static ERR_STRING_DATA ASN1_str_reasons[]=
 {ERR_REASON(ASN1_R_UNSUPPORTED_ENCRYPTION_ALGORITHM),"unsupported encryption algorithm"},
 {ERR_REASON(ASN1_R_UNSUPPORTED_PUBLIC_KEY_TYPE),"unsupported public key type"},
 {ERR_REASON(ASN1_R_UNSUPPORTED_TYPE)     ,"unsupported type"},
+{ERR_REASON(ASN1_R_WRONG_PUBLIC_KEY_TYPE),"wrong public key type"},
 {ERR_REASON(ASN1_R_WRONG_TAG)            ,"wrong tag"},
 {ERR_REASON(ASN1_R_WRONG_TYPE)           ,"wrong type"},
 {0,NULL}
diff --git a/crypto/asn1/asn1_gen.c b/crypto/asn1/asn1_gen.c
index 213a8e986c1b..4fc241908f69 100644
--- a/crypto/asn1/asn1_gen.c
+++ b/crypto/asn1/asn1_gen.c
@@ -247,8 +247,14 @@ ASN1_TYPE *ASN1_generate_v3(char *str, X509V3_CTX *cnf)
 	/* If IMPLICIT, output tag */
 
 	if (asn1_tags.imp_tag != -1)
+		{
+		if (asn1_tags.imp_class == V_ASN1_UNIVERSAL 
+		    && (asn1_tags.imp_tag == V_ASN1_SEQUENCE
+		     || asn1_tags.imp_tag == V_ASN1_SET) )
+			hdr_constructed = V_ASN1_CONSTRUCTED;
 		ASN1_put_object(&p, hdr_constructed, hdr_len,
 					asn1_tags.imp_tag, asn1_tags.imp_class);
+		}
 
 	/* Copy across original encoding */
 	memcpy(p, cpy_start, cpy_len);
@@ -441,12 +447,12 @@ static int parse_tagging(const char *vstart, int vlen, int *ptag, int *pclass)
 
 static ASN1_TYPE *asn1_multi(int utype, const char *section, X509V3_CTX *cnf)
 	{
-	ASN1_TYPE *ret = NULL, *typ = NULL;
+	ASN1_TYPE *ret = NULL;
 	STACK_OF(ASN1_TYPE) *sk = NULL;
 	STACK_OF(CONF_VALUE) *sect = NULL;
-	unsigned char *der = NULL, *p;
+	unsigned char *der = NULL;
 	int derlen;
-	int i, is_set;
+	int i;
 	sk = sk_ASN1_TYPE_new_null();
 	if (!sk)
 		goto bad;
@@ -459,31 +465,23 @@ static ASN1_TYPE *asn1_multi(int utype, const char *section, X509V3_CTX *cnf)
 			goto bad;
 		for (i = 0; i < sk_CONF_VALUE_num(sect); i++)
 			{
-			typ = ASN1_generate_v3(sk_CONF_VALUE_value(sect, i)->value, cnf);
+			ASN1_TYPE *typ = ASN1_generate_v3(sk_CONF_VALUE_value(sect, i)->value, cnf);
 			if (!typ)
 				goto bad;
 			if (!sk_ASN1_TYPE_push(sk, typ))
 				goto bad;
-			typ = NULL;
 			}
 		}
 
 	/* Now we has a STACK of the components, convert to the correct form */
 
 	if (utype == V_ASN1_SET)
-		is_set = 1;
+		derlen = i2d_ASN1_SET_ANY(sk, &der);
 	else
-		is_set = 0;
-
+		derlen = i2d_ASN1_SEQUENCE_ANY(sk, &der);
 
-	derlen = i2d_ASN1_SET_OF_ASN1_TYPE(sk, NULL, i2d_ASN1_TYPE, utype,
-					   V_ASN1_UNIVERSAL, is_set);
-	der = OPENSSL_malloc(derlen);
-	if (!der)
+	if (derlen < 0)
 		goto bad;
-	p = der;
-	i2d_ASN1_SET_OF_ASN1_TYPE(sk, &p, i2d_ASN1_TYPE, utype,
-				  V_ASN1_UNIVERSAL, is_set);
 
 	if (!(ret = ASN1_TYPE_new()))
 		goto bad;
@@ -505,8 +503,6 @@ static ASN1_TYPE *asn1_multi(int utype, const char *section, X509V3_CTX *cnf)
 
 	if (sk)
 		sk_ASN1_TYPE_pop_free(sk, ASN1_TYPE_free);
-	if (typ)
-		ASN1_TYPE_free(typ);
 	if (sect)
 		X509V3_section_free(cnf, sect);
 
@@ -556,7 +552,7 @@ static int append_exp(tag_exp_arg *arg, int exp_tag, int exp_class, int exp_cons
 static int asn1_str2tag(const char *tagstr, int len)
 	{
 	unsigned int i;
-	static struct tag_name_st *tntmp, tnst [] = {
+	static const struct tag_name_st *tntmp, tnst [] = {
 		ASN1_GEN_STR("BOOL", V_ASN1_BOOLEAN),
 		ASN1_GEN_STR("BOOLEAN", V_ASN1_BOOLEAN),
 		ASN1_GEN_STR("NULL", V_ASN1_NULL),
@@ -591,6 +587,8 @@ static int asn1_str2tag(const char *tagstr, int len)
 		ASN1_GEN_STR("TELETEXSTRING", V_ASN1_T61STRING),
 		ASN1_GEN_STR("GeneralString", V_ASN1_GENERALSTRING),
 		ASN1_GEN_STR("GENSTR", V_ASN1_GENERALSTRING),
+		ASN1_GEN_STR("NUMERIC", V_ASN1_NUMERICSTRING),
+		ASN1_GEN_STR("NUMERICSTRING", V_ASN1_NUMERICSTRING),
 
 		/* Special cases */
 		ASN1_GEN_STR("SEQUENCE", V_ASN1_SEQUENCE),
@@ -736,6 +734,7 @@ static ASN1_TYPE *asn1_str2type(const char *str, int format, int utype)
 		case V_ASN1_VISIBLESTRING:
 		case V_ASN1_UNIVERSALSTRING:
 		case V_ASN1_GENERALSTRING:
+		case V_ASN1_NUMERICSTRING:
 
 		if (format == ASN1_GEN_FORMAT_ASCII)
 			format = MBSTRING_ASC;
diff --git a/crypto/asn1/asn1_lib.c b/crypto/asn1/asn1_lib.c
index 5af559ef8da7..1bcb44aee203 100644
--- a/crypto/asn1/asn1_lib.c
+++ b/crypto/asn1/asn1_lib.c
@@ -340,20 +340,31 @@ int asn1_GetSequence(ASN1_const_CTX *c, long *length)
 	return(1);
 	}
 
-ASN1_STRING *ASN1_STRING_dup(ASN1_STRING *str)
+int ASN1_STRING_copy(ASN1_STRING *dst, const ASN1_STRING *str)
 	{
-	ASN1_STRING *ret;
+	if (str == NULL)
+		return 0;
+	dst->type = str->type;
+	if (!ASN1_STRING_set(dst,str->data,str->length))
+		return 0;
+	dst->flags = str->flags;
+	return 1;
+	}
 
-	if (str == NULL) return(NULL);
-	if ((ret=ASN1_STRING_type_new(str->type)) == NULL)
-		return(NULL);
-	if (!ASN1_STRING_set(ret,str->data,str->length))
+ASN1_STRING *ASN1_STRING_dup(const ASN1_STRING *str)
+	{
+	ASN1_STRING *ret;
+	if (!str)
+		 return NULL;
+	ret=ASN1_STRING_new();
+	if (!ret)
+		return NULL;
+	if (!ASN1_STRING_copy(ret,str))
 		{
 		ASN1_STRING_free(ret);
-		return(NULL);
+		return NULL;
 		}
-	ret->flags = str->flags;
-	return(ret);
+	return ret;
 	}
 
 int ASN1_STRING_set(ASN1_STRING *str, const void *_data, int len)
@@ -427,11 +438,12 @@ ASN1_STRING *ASN1_STRING_type_new(int type)
 void ASN1_STRING_free(ASN1_STRING *a)
 	{
 	if (a == NULL) return;
-	if (a->data != NULL) OPENSSL_free(a->data);
+	if (a->data && !(a->flags & ASN1_STRING_FLAG_NDEF))
+		OPENSSL_free(a->data);
 	OPENSSL_free(a);
 	}
 
-int ASN1_STRING_cmp(ASN1_STRING *a, ASN1_STRING *b)
+int ASN1_STRING_cmp(const ASN1_STRING *a, const ASN1_STRING *b)
 	{
 	int i;
 
@@ -457,7 +469,7 @@ void asn1_add_error(const unsigned char *address, int offset)
 	ERR_add_error_data(4,"address=",buf1," offset=",buf2);
 	}
 
-int ASN1_STRING_length(ASN1_STRING *x)
+int ASN1_STRING_length(const ASN1_STRING *x)
 { return M_ASN1_STRING_length(x); }
 
 void ASN1_STRING_length_set(ASN1_STRING *x, int len)
diff --git a/crypto/asn1/asn1_locl.h b/crypto/asn1/asn1_locl.h
new file mode 100644
index 000000000000..9fcf0d9530ff
--- /dev/null
+++ b/crypto/asn1/asn1_locl.h
@@ -0,0 +1,145 @@
+/* asn1t.h */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+/* Internal ASN1 structures and functions: not for application use */
+
+/* ASN1 print context structure */
+
+struct asn1_pctx_st
+	{
+	unsigned long flags;
+	unsigned long nm_flags;
+	unsigned long cert_flags;
+	unsigned long oid_flags;
+	unsigned long str_flags;
+	} /* ASN1_PCTX */;
+
+/* ASN1 public key method structure */
+
+struct evp_pkey_asn1_method_st
+	{
+	int pkey_id;
+	int pkey_base_id;
+	unsigned long pkey_flags;
+
+	char *pem_str;
+	char *info;
+
+	int (*pub_decode)(EVP_PKEY *pk, X509_PUBKEY *pub);
+	int (*pub_encode)(X509_PUBKEY *pub, const EVP_PKEY *pk);
+	int (*pub_cmp)(const EVP_PKEY *a, const EVP_PKEY *b);
+	int (*pub_print)(BIO *out, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *pctx);
+
+	int (*priv_decode)(EVP_PKEY *pk, PKCS8_PRIV_KEY_INFO *p8inf);
+	int (*priv_encode)(PKCS8_PRIV_KEY_INFO *p8, const EVP_PKEY *pk);
+	int (*priv_print)(BIO *out, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *pctx);
+
+	int (*pkey_size)(const EVP_PKEY *pk);
+	int (*pkey_bits)(const EVP_PKEY *pk);
+
+	int (*param_decode)(EVP_PKEY *pkey,
+				const unsigned char **pder, int derlen);
+	int (*param_encode)(const EVP_PKEY *pkey, unsigned char **pder);
+	int (*param_missing)(const EVP_PKEY *pk);
+	int (*param_copy)(EVP_PKEY *to, const EVP_PKEY *from);
+	int (*param_cmp)(const EVP_PKEY *a, const EVP_PKEY *b);
+	int (*param_print)(BIO *out, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *pctx);
+	int (*sig_print)(BIO *out,
+			 const X509_ALGOR *sigalg, const ASN1_STRING *sig,
+					 int indent, ASN1_PCTX *pctx);
+
+
+	void (*pkey_free)(EVP_PKEY *pkey);
+	int (*pkey_ctrl)(EVP_PKEY *pkey, int op, long arg1, void *arg2);
+
+	/* Legacy functions for old PEM */
+
+	int (*old_priv_decode)(EVP_PKEY *pkey,
+				const unsigned char **pder, int derlen);
+	int (*old_priv_encode)(const EVP_PKEY *pkey, unsigned char **pder);
+	/* Custom ASN1 signature verification */
+	int (*item_verify)(EVP_MD_CTX *ctx, const ASN1_ITEM *it, void *asn,
+				X509_ALGOR *a, ASN1_BIT_STRING *sig,
+				EVP_PKEY *pkey);
+	int (*item_sign)(EVP_MD_CTX *ctx, const ASN1_ITEM *it, void *asn,
+				X509_ALGOR *alg1, X509_ALGOR *alg2, 
+				ASN1_BIT_STRING *sig);
+
+	} /* EVP_PKEY_ASN1_METHOD */;
+
+/* Method to handle CRL access.
+ * In general a CRL could be very large (several Mb) and can consume large
+ * amounts of resources if stored in memory by multiple processes.
+ * This method allows general CRL operations to be redirected to more
+ * efficient callbacks: for example a CRL entry database.
+ */
+
+#define X509_CRL_METHOD_DYNAMIC		1
+
+struct x509_crl_method_st
+	{
+	int flags;
+	int (*crl_init)(X509_CRL *crl);
+	int (*crl_free)(X509_CRL *crl);
+	int (*crl_lookup)(X509_CRL *crl, X509_REVOKED **ret,
+				ASN1_INTEGER *ser, X509_NAME *issuer);
+	int (*crl_verify)(X509_CRL *crl, EVP_PKEY *pk);
+	};
diff --git a/crypto/asn1/asn1_mac.h b/crypto/asn1/asn1_mac.h
index d958ca60d9dc..87bd0e9e1d81 100644
--- a/crypto/asn1/asn1_mac.h
+++ b/crypto/asn1/asn1_mac.h
@@ -153,6 +153,13 @@ err:\
 		M_ASN1_D2I_get(b,func); \
 		}
 
+#define M_ASN1_D2I_get_int_opt(b,func,type) \
+	if ((c.slen != 0) && ((M_ASN1_next & (~V_ASN1_CONSTRUCTED)) \
+		== (V_ASN1_UNIVERSAL|(type)))) \
+		{ \
+		M_ASN1_D2I_get_int(b,func); \
+		}
+
 #define M_ASN1_D2I_get_imp(b,func, type) \
 	M_ASN1_next=(_tmp& V_ASN1_CONSTRUCTED)|type; \
 	c.q=c.p; \
diff --git a/crypto/asn1/asn1_par.c b/crypto/asn1/asn1_par.c
index cb08e154888f..aaca69aebd94 100644
--- a/crypto/asn1/asn1_par.c
+++ b/crypto/asn1/asn1_par.c
@@ -70,9 +70,8 @@ static int asn1_print_info(BIO *bp, int tag, int xclass, int constructed,
 	     int indent)
 	{
 	static const char fmt[]="%-18s";
-	static const char fmt2[]="%2d %-15s";
 	char str[128];
-	const char *p,*p2=NULL;
+	const char *p;
 
 	if (constructed & V_ASN1_CONSTRUCTED)
 		p="cons: ";
@@ -93,14 +92,8 @@ static int asn1_print_info(BIO *bp, int tag, int xclass, int constructed,
 	else
 		p = ASN1_tag2str(tag);
 
-	if (p2 != NULL)
-		{
-		if (BIO_printf(bp,fmt2,tag,p2) <= 0) goto err;
-		}
-	else
-		{
-		if (BIO_printf(bp,fmt,p) <= 0) goto err;
-		}
+	if (BIO_printf(bp,fmt,p) <= 0)
+		goto err;
 	return(1);
 err:
 	return(0);
@@ -424,7 +417,7 @@ end:
 
 const char *ASN1_tag2str(int tag)
 {
-	static const char *tag2str[] = {
+	static const char * const tag2str[] = {
 	 "EOC", "BOOLEAN", "INTEGER", "BIT STRING", "OCTET STRING", /* 0-4 */
 	 "NULL", "OBJECT", "OBJECT DESCRIPTOR", "EXTERNAL", "REAL", /* 5-9 */
 	 "ENUMERATED", "<ASN1 11>", "UTF8STRING", "<ASN1 13>", 	    /* 10-13 */
diff --git a/crypto/asn1/asn1t.h b/crypto/asn1/asn1t.h
index ac14f9415b84..d230e4bf7003 100644
--- a/crypto/asn1/asn1t.h
+++ b/crypto/asn1/asn1t.h
@@ -3,7 +3,7 @@
  * project 2000.
  */
 /* ====================================================================
- * Copyright (c) 2000 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2000-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -218,6 +218,18 @@ extern "C" {
 		#stname \
 	ASN1_ITEM_end(tname)
 
+#define ASN1_NDEF_SEQUENCE_END_cb(stname, tname) \
+	;\
+	ASN1_ITEM_start(tname) \
+		ASN1_ITYPE_NDEF_SEQUENCE,\
+		V_ASN1_SEQUENCE,\
+		tname##_seq_tt,\
+		sizeof(tname##_seq_tt) / sizeof(ASN1_TEMPLATE),\
+		&tname##_aux,\
+		sizeof(stname),\
+		#stname \
+	ASN1_ITEM_end(tname)
+
 
 /* This pair helps declare a CHOICE type. We can do:
  *
@@ -651,8 +663,13 @@ typedef int ASN1_ex_i2d(ASN1_VALUE **pval, unsigned char **out, const ASN1_ITEM
 typedef int ASN1_ex_new_func(ASN1_VALUE **pval, const ASN1_ITEM *it);
 typedef void ASN1_ex_free_func(ASN1_VALUE **pval, const ASN1_ITEM *it);
 
+typedef int ASN1_ex_print_func(BIO *out, ASN1_VALUE **pval, 
+						int indent, const char *fname, 
+						const ASN1_PCTX *pctx);
+
 typedef int ASN1_primitive_i2c(ASN1_VALUE **pval, unsigned char *cont, int *putype, const ASN1_ITEM *it);
 typedef int ASN1_primitive_c2i(ASN1_VALUE **pval, const unsigned char *cont, int len, int utype, char *free_cont, const ASN1_ITEM *it);
+typedef int ASN1_primitive_print(BIO *out, ASN1_VALUE **pval, const ASN1_ITEM *it, int indent, const ASN1_PCTX *pctx);
 
 typedef struct ASN1_COMPAT_FUNCS_st {
 	ASN1_new_func *asn1_new;
@@ -668,6 +685,7 @@ typedef struct ASN1_EXTERN_FUNCS_st {
 	ASN1_ex_free_func *asn1_ex_clear;
 	ASN1_ex_d2i *asn1_ex_d2i;
 	ASN1_ex_i2d *asn1_ex_i2d;
+	ASN1_ex_print_func *asn1_ex_print;
 } ASN1_EXTERN_FUNCS;
 
 typedef struct ASN1_PRIMITIVE_FUNCS_st {
@@ -678,6 +696,7 @@ typedef struct ASN1_PRIMITIVE_FUNCS_st {
 	ASN1_ex_free_func *prim_clear;
 	ASN1_primitive_c2i *prim_c2i;
 	ASN1_primitive_i2c *prim_i2c;
+	ASN1_primitive_print *prim_print;
 } ASN1_PRIMITIVE_FUNCS;
 
 /* This is the ASN1_AUX structure: it handles various
@@ -697,7 +716,8 @@ typedef struct ASN1_PRIMITIVE_FUNCS_st {
  * then an external type is more appropriate.
  */
 
-typedef int ASN1_aux_cb(int operation, ASN1_VALUE **in, const ASN1_ITEM *it);
+typedef int ASN1_aux_cb(int operation, ASN1_VALUE **in, const ASN1_ITEM *it,
+				void *exarg);
 
 typedef struct ASN1_AUX_st {
 	void *app_data;
@@ -708,6 +728,23 @@ typedef struct ASN1_AUX_st {
 	int enc_offset;		/* Offset of ASN1_ENCODING structure */
 } ASN1_AUX;
 
+/* For print related callbacks exarg points to this structure */
+typedef struct ASN1_PRINT_ARG_st {
+	BIO *out;
+	int indent;
+	const ASN1_PCTX *pctx;
+} ASN1_PRINT_ARG;
+
+/* For streaming related callbacks exarg points to this structure */
+typedef struct ASN1_STREAM_ARG_st {
+	/* BIO to stream through */
+	BIO *out;
+	/* BIO with filters appended */
+	BIO *ndef_bio;
+	/* Streaming I/O boundary */
+	unsigned char **boundary;
+} ASN1_STREAM_ARG;
+
 /* Flags in ASN1_AUX */
 
 /* Use a reference count */
@@ -727,6 +764,12 @@ typedef struct ASN1_AUX_st {
 #define ASN1_OP_D2I_POST	5
 #define ASN1_OP_I2D_PRE		6
 #define ASN1_OP_I2D_POST	7
+#define ASN1_OP_PRINT_PRE	8
+#define ASN1_OP_PRINT_POST	9
+#define ASN1_OP_STREAM_PRE	10
+#define ASN1_OP_STREAM_POST	11
+#define ASN1_OP_DETACHED_PRE	12
+#define ASN1_OP_DETACHED_POST	13
 
 /* Macro to implement a primitive type */
 #define IMPLEMENT_ASN1_TYPE(stname) IMPLEMENT_ASN1_TYPE_ex(stname, stname, 0)
@@ -782,9 +825,22 @@ typedef struct ASN1_AUX_st {
 #define IMPLEMENT_ASN1_FUNCTIONS_ENCODE_name(stname, itname) \
 			IMPLEMENT_ASN1_FUNCTIONS_ENCODE_fname(stname, itname, itname)
 
+#define IMPLEMENT_STATIC_ASN1_ALLOC_FUNCTIONS(stname) \
+		IMPLEMENT_ASN1_ALLOC_FUNCTIONS_pfname(static, stname, stname, stname)
+
 #define IMPLEMENT_ASN1_ALLOC_FUNCTIONS(stname) \
 		IMPLEMENT_ASN1_ALLOC_FUNCTIONS_fname(stname, stname, stname)
 
+#define IMPLEMENT_ASN1_ALLOC_FUNCTIONS_pfname(pre, stname, itname, fname) \
+	pre stname *fname##_new(void) \
+	{ \
+		return (stname *)ASN1_item_new(ASN1_ITEM_rptr(itname)); \
+	} \
+	pre void fname##_free(stname *a) \
+	{ \
+		ASN1_item_free((ASN1_VALUE *)a, ASN1_ITEM_rptr(itname)); \
+	}
+
 #define IMPLEMENT_ASN1_ALLOC_FUNCTIONS_fname(stname, itname, fname) \
 	stname *fname##_new(void) \
 	{ \
@@ -834,6 +890,17 @@ typedef struct ASN1_AUX_st {
         return ASN1_item_dup(ASN1_ITEM_rptr(stname), x); \
         }
 
+#define IMPLEMENT_ASN1_PRINT_FUNCTION(stname) \
+	IMPLEMENT_ASN1_PRINT_FUNCTION_fname(stname, stname, stname)
+
+#define IMPLEMENT_ASN1_PRINT_FUNCTION_fname(stname, itname, fname) \
+	int fname##_print_ctx(BIO *out, stname *x, int indent, \
+						const ASN1_PCTX *pctx) \
+	{ \
+		return ASN1_item_print(out, (ASN1_VALUE *)x, indent, \
+			ASN1_ITEM_rptr(itname), pctx); \
+	} 
+
 #define IMPLEMENT_ASN1_FUNCTIONS_const(name) \
 		IMPLEMENT_ASN1_FUNCTIONS_const_fname(name, name, name)
 
diff --git a/crypto/asn1/asn_mime.c b/crypto/asn1/asn_mime.c
index ad8fbed90725..54a704a969ea 100644
--- a/crypto/asn1/asn_mime.c
+++ b/crypto/asn1/asn_mime.c
@@ -59,6 +59,7 @@
 #include <openssl/x509.h>
 #include <openssl/asn1.h>
 #include <openssl/asn1t.h>
+#include "asn1_locl.h"
 
 /* Generalised MIME like utilities for streaming ASN1. Although many
  * have a PKCS7/CMS like flavour others are more general purpose.
@@ -86,6 +87,8 @@ STACK_OF(MIME_PARAM) *params;		/* Zero or more parameters */
 DECLARE_STACK_OF(MIME_HEADER)
 IMPLEMENT_STACK_OF(MIME_HEADER)
 
+static int asn1_output_data(BIO *out, BIO *data, ASN1_VALUE *val, int flags,
+					const ASN1_ITEM *it);
 static char * strip_ends(char *name);
 static char * strip_start(char *name);
 static char * strip_end(char *name);
@@ -107,6 +110,39 @@ static void mime_hdr_free(MIME_HEADER *hdr);
 #define MAX_SMLEN 1024
 #define mime_debug(x) /* x */
 
+/* Output an ASN1 structure in BER format streaming if necessary */
+
+int i2d_ASN1_bio_stream(BIO *out, ASN1_VALUE *val, BIO *in, int flags,
+				const ASN1_ITEM *it)
+	{
+	/* If streaming create stream BIO and copy all content through it */
+	if (flags & SMIME_STREAM)
+		{
+		BIO *bio, *tbio;
+		bio = BIO_new_NDEF(out, val, it);
+		if (!bio)
+			{
+			ASN1err(ASN1_F_I2D_ASN1_BIO_STREAM,ERR_R_MALLOC_FAILURE);
+			return 0;
+			}
+		SMIME_crlf_copy(in, bio, flags);
+		(void)BIO_flush(bio);
+		/* Free up successive BIOs until we hit the old output BIO */
+		do
+			{
+			tbio = BIO_pop(bio);
+			BIO_free(bio);
+			bio = tbio;
+			} while (bio != out);
+		}
+	/* else just write out ASN1 structure which will have all content
+	 * stored internally
+	 */
+	else
+		ASN1_item_i2d_bio(it, out, val);
+	return 1;
+	}
+
 /* Base 64 read and write of ASN1 structure */
 
 static int B64_write_ASN1(BIO *out, ASN1_VALUE *val, BIO *in, int flags,
@@ -123,13 +159,26 @@ static int B64_write_ASN1(BIO *out, ASN1_VALUE *val, BIO *in, int flags,
 	/* prepend the b64 BIO so all data is base64 encoded.
 	 */
 	out = BIO_push(b64, out);
-	r = ASN1_item_i2d_bio(it, out, val);
+	r = i2d_ASN1_bio_stream(out, val, in, flags, it);
 	(void)BIO_flush(out);
 	BIO_pop(out);
 	BIO_free(b64);
 	return r;
 	}
 
+/* Streaming ASN1 PEM write */
+
+int PEM_write_bio_ASN1_stream(BIO *out, ASN1_VALUE *val, BIO *in, int flags,
+				const char *hdr,
+				const ASN1_ITEM *it)
+	{
+	int r;
+	BIO_printf(out, "-----BEGIN %s-----\n", hdr);
+	r = B64_write_ASN1(out, val, in, flags, it);
+	BIO_printf(out, "-----END %s-----\n", hdr);
+	return r;
+	}
+
 static ASN1_VALUE *b64_read_asn1(BIO *bio, const ASN1_ITEM *it)
 {
 	BIO *b64;
@@ -152,7 +201,8 @@ static ASN1_VALUE *b64_read_asn1(BIO *bio, const ASN1_ITEM *it)
 
 static int asn1_write_micalg(BIO *out, STACK_OF(X509_ALGOR) *mdalgs)
 	{
-	int i, have_unknown = 0, write_comma, md_nid;
+	const EVP_MD *md;
+	int i, have_unknown = 0, write_comma, ret = 0, md_nid;
 	have_unknown = 0;
 	write_comma = 0;
 	for (i = 0; i < sk_X509_ALGOR_num(mdalgs); i++)
@@ -161,6 +211,21 @@ static int asn1_write_micalg(BIO *out, STACK_OF(X509_ALGOR) *mdalgs)
 			BIO_write(out, ",", 1);
 		write_comma = 1;
 		md_nid = OBJ_obj2nid(sk_X509_ALGOR_value(mdalgs, i)->algorithm);
+		md = EVP_get_digestbynid(md_nid);
+		if (md && md->md_ctrl)
+			{
+			int rv;
+			char *micstr;
+			rv = md->md_ctrl(NULL, EVP_MD_CTRL_MICALG, 0, &micstr);
+			if (rv > 0)
+				{
+				BIO_puts(out, micstr);
+				OPENSSL_free(micstr);
+				continue;
+				}
+			if (rv != -2)
+				goto err;
+			}
 		switch(md_nid)
 			{
 			case NID_sha1:
@@ -183,6 +248,11 @@ static int asn1_write_micalg(BIO *out, STACK_OF(X509_ALGOR) *mdalgs)
 			BIO_puts(out, "sha-512");
 			break;
 
+			case NID_id_GostR3411_94:
+			BIO_puts(out, "gostr3411-94");
+				goto err;
+			break;
+
 			default:
 			if (have_unknown)
 				write_comma = 0;
@@ -196,16 +266,18 @@ static int asn1_write_micalg(BIO *out, STACK_OF(X509_ALGOR) *mdalgs)
 			}
 		}
 
-	return 1;
+	ret = 1;
+	err:
+
+	return ret;
 
 	}
 
 /* SMIME sender */
 
-int int_smime_write_ASN1(BIO *bio, ASN1_VALUE *val, BIO *data, int flags,
+int SMIME_write_ASN1(BIO *bio, ASN1_VALUE *val, BIO *data, int flags,
 				int ctype_nid, int econt_nid,
 				STACK_OF(X509_ALGOR) *mdalgs,
-				asn1_output_data_fn *data_fn,
 				const ASN1_ITEM *it)
 {
 	char bound[33], c;
@@ -243,7 +315,7 @@ int int_smime_write_ASN1(BIO *bio, ASN1_VALUE *val, BIO *data, int flags,
 						mime_eol, mime_eol);
 		/* Now write out the first part */
 		BIO_printf(bio, "------%s%s", bound, mime_eol);
-		if (!data_fn(bio, data, val, flags, it))
+		if (!asn1_output_data(bio, data, val, flags, it))
 			return 0;
 		BIO_printf(bio, "%s------%s%s", mime_eol, bound, mime_eol);
 
@@ -296,8 +368,6 @@ int int_smime_write_ASN1(BIO *bio, ASN1_VALUE *val, BIO *data, int flags,
 	return 1;
 }
 
-#if 0
-
 /* Handle output of ASN1 data */
 
 
@@ -307,8 +377,12 @@ static int asn1_output_data(BIO *out, BIO *data, ASN1_VALUE *val, int flags,
 	BIO *tmpbio;
 	const ASN1_AUX *aux = it->funcs;
 	ASN1_STREAM_ARG sarg;
+	int rv = 1;
 
-	if (!(flags & SMIME_DETACHED))
+	/* If data is not deteched or resigning then the output BIO is
+	 * already set up to finalise when it is written through.
+	 */
+	if (!(flags & SMIME_DETACHED) || (flags & PKCS7_REUSE_DIGEST))
 		{
 		SMIME_crlf_copy(data, out, flags);
 		return 1;
@@ -335,7 +409,7 @@ static int asn1_output_data(BIO *out, BIO *data, ASN1_VALUE *val, int flags,
 
 	/* Finalize structure */
 	if (aux->asn1_cb(ASN1_OP_DETACHED_POST, &val, it, &sarg) <= 0)
-		return 0;
+		rv = 0;
 
 	/* Now remove any digests prepended to the BIO */
 
@@ -346,12 +420,10 @@ static int asn1_output_data(BIO *out, BIO *data, ASN1_VALUE *val, int flags,
 		sarg.ndef_bio = tmpbio;
 		}
 
-	return 1;
+	return rv;
 
 	}
 
-#endif
-
 /* SMIME reader: handle multipart/signed and opaque signing.
  * in multipart case the content is placed in a memory BIO
  * pointed to by "bcont". In opaque this is set to NULL
@@ -733,7 +805,7 @@ static MIME_HEADER *mime_hdr_new(char *name, char *value)
 	if(name) {
 		if(!(tmpname = BUF_strdup(name))) return NULL;
 		for(p = tmpname ; *p; p++) {
-			c = *p;
+			c = (unsigned char)*p;
 			if(isupper(c)) {
 				c = tolower(c);
 				*p = c;
@@ -743,7 +815,7 @@ static MIME_HEADER *mime_hdr_new(char *name, char *value)
 	if(value) {
 		if(!(tmpval = BUF_strdup(value))) return NULL;
 		for(p = tmpval ; *p; p++) {
-			c = *p;
+			c = (unsigned char)*p;
 			if(isupper(c)) {
 				c = tolower(c);
 				*p = c;
@@ -767,7 +839,7 @@ static int mime_hdr_addparam(MIME_HEADER *mhdr, char *name, char *value)
 		tmpname = BUF_strdup(name);
 		if(!tmpname) return 0;
 		for(p = tmpname ; *p; p++) {
-			c = *p;
+			c = (unsigned char)*p;
 			if(isupper(c)) {
 				c = tolower(c);
 				*p = c;
diff --git a/crypto/asn1/asn_pack.c b/crypto/asn1/asn_pack.c
index f1a5a0563227..ad738217d778 100644
--- a/crypto/asn1/asn_pack.c
+++ b/crypto/asn1/asn_pack.c
@@ -66,10 +66,10 @@
 
 /* Turn an ASN1 encoded SEQUENCE OF into a STACK of structures */
 
-STACK *ASN1_seq_unpack(const unsigned char *buf, int len,
-		       d2i_of_void *d2i,void (*free_func)(void *))
+STACK_OF(OPENSSL_BLOCK) *ASN1_seq_unpack(const unsigned char *buf, int len,
+			 d2i_of_void *d2i, void (*free_func)(OPENSSL_BLOCK))
 {
-    STACK *sk;
+    STACK_OF(OPENSSL_BLOCK) *sk;
     const unsigned char *pbuf;
     pbuf =  buf;
     if (!(sk = d2i_ASN1_SET(NULL, &pbuf, len, d2i, free_func,
@@ -82,7 +82,7 @@ STACK *ASN1_seq_unpack(const unsigned char *buf, int len,
  * OPENSSL_malloc'ed buffer
  */
 
-unsigned char *ASN1_seq_pack(STACK *safes, i2d_of_void *i2d,
+unsigned char *ASN1_seq_pack(STACK_OF(OPENSSL_BLOCK) *safes, i2d_of_void *i2d,
 			     unsigned char **buf, int *len)
 {
 	int safelen;
diff --git a/crypto/asn1/bio_asn1.c b/crypto/asn1/bio_asn1.c
new file mode 100644
index 000000000000..dc7efd551c05
--- /dev/null
+++ b/crypto/asn1/bio_asn1.c
@@ -0,0 +1,495 @@
+/* bio_asn1.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+/* Experimental ASN1 BIO. When written through the data is converted
+ * to an ASN1 string type: default is OCTET STRING. Additional functions
+ * can be provided to add prefix and suffix data.
+ */
+
+#include <string.h>
+#include <openssl/bio.h>
+#include <openssl/asn1.h>
+
+/* Must be large enough for biggest tag+length */
+#define DEFAULT_ASN1_BUF_SIZE 20
+
+typedef enum 
+	{
+	ASN1_STATE_START,
+	ASN1_STATE_PRE_COPY,
+	ASN1_STATE_HEADER,
+	ASN1_STATE_HEADER_COPY,
+	ASN1_STATE_DATA_COPY,
+	ASN1_STATE_POST_COPY,
+	ASN1_STATE_DONE
+	} asn1_bio_state_t;
+
+typedef struct BIO_ASN1_EX_FUNCS_st
+	{
+	asn1_ps_func	*ex_func;
+	asn1_ps_func	*ex_free_func;
+	} BIO_ASN1_EX_FUNCS;
+
+typedef struct BIO_ASN1_BUF_CTX_t
+	{
+	/* Internal state */
+	asn1_bio_state_t state;
+	/* Internal buffer */
+	unsigned char *buf;
+	/* Size of buffer */
+	int bufsize;
+	/* Current position in buffer */
+	int bufpos;
+	/* Current buffer length */
+	int buflen;
+	/* Amount of data to copy */
+	int copylen;
+	/* Class and tag to use */
+	int asn1_class, asn1_tag;
+	asn1_ps_func *prefix, *prefix_free, *suffix, *suffix_free;
+	/* Extra buffer for prefix and suffix data */
+	unsigned char *ex_buf;
+	int ex_len;
+	int ex_pos;
+	void *ex_arg;
+	} BIO_ASN1_BUF_CTX;
+
+
+static int asn1_bio_write(BIO *h, const char *buf,int num);
+static int asn1_bio_read(BIO *h, char *buf, int size);
+static int asn1_bio_puts(BIO *h, const char *str);
+static int asn1_bio_gets(BIO *h, char *str, int size);
+static long asn1_bio_ctrl(BIO *h, int cmd, long arg1, void *arg2);
+static int asn1_bio_new(BIO *h);
+static int asn1_bio_free(BIO *data);
+static long asn1_bio_callback_ctrl(BIO *h, int cmd, bio_info_cb *fp);
+
+static int asn1_bio_init(BIO_ASN1_BUF_CTX *ctx, int size);
+static int asn1_bio_flush_ex(BIO *b, BIO_ASN1_BUF_CTX *ctx,
+				asn1_ps_func *cleanup, asn1_bio_state_t next);
+static int asn1_bio_setup_ex(BIO *b, BIO_ASN1_BUF_CTX *ctx,
+				asn1_ps_func *setup, 
+				asn1_bio_state_t ex_state,
+				asn1_bio_state_t other_state);
+
+static BIO_METHOD methods_asn1=
+	{
+	BIO_TYPE_ASN1,
+	"asn1",
+	asn1_bio_write,
+	asn1_bio_read,
+	asn1_bio_puts,
+	asn1_bio_gets,
+	asn1_bio_ctrl,
+	asn1_bio_new,
+	asn1_bio_free,
+	asn1_bio_callback_ctrl,
+	};
+
+BIO_METHOD *BIO_f_asn1(void)
+	{
+	return(&methods_asn1);
+	}
+
+
+static int asn1_bio_new(BIO *b)
+	{
+	BIO_ASN1_BUF_CTX *ctx;
+	ctx = OPENSSL_malloc(sizeof(BIO_ASN1_BUF_CTX));
+	if (!ctx)
+		return 0;
+	if (!asn1_bio_init(ctx, DEFAULT_ASN1_BUF_SIZE))
+		return 0;
+	b->init = 1;
+	b->ptr = (char *)ctx;
+	b->flags = 0;
+	return 1;
+	}
+
+static int asn1_bio_init(BIO_ASN1_BUF_CTX *ctx, int size)
+	{
+	ctx->buf = OPENSSL_malloc(size);
+	if (!ctx->buf)
+		return 0;
+	ctx->bufsize = size;
+	ctx->bufpos = 0;
+	ctx->buflen = 0;
+	ctx->copylen = 0;
+	ctx->asn1_class = V_ASN1_UNIVERSAL;
+	ctx->asn1_tag = V_ASN1_OCTET_STRING;
+	ctx->ex_buf = 0;
+	ctx->ex_pos = 0;
+	ctx->ex_len = 0;
+	ctx->state = ASN1_STATE_START;
+	return 1;
+	}
+
+static int asn1_bio_free(BIO *b)
+	{
+	BIO_ASN1_BUF_CTX *ctx;
+	ctx = (BIO_ASN1_BUF_CTX *) b->ptr;
+	if (ctx == NULL)
+		return 0;
+	if (ctx->buf)
+		OPENSSL_free(ctx->buf);
+	OPENSSL_free(ctx);
+	b->init = 0;
+	b->ptr = NULL;
+	b->flags = 0;
+	return 1;
+	}
+
+static int asn1_bio_write(BIO *b, const char *in , int inl)
+	{
+	BIO_ASN1_BUF_CTX *ctx;
+	int wrmax, wrlen, ret;
+	unsigned char *p;
+	if (!in || (inl < 0) || (b->next_bio == NULL))
+		return 0;
+	ctx = (BIO_ASN1_BUF_CTX *) b->ptr;
+	if (ctx == NULL)
+		return 0;
+
+	wrlen = 0;
+	ret = -1;
+
+	for(;;)
+		{
+		switch (ctx->state)
+			{
+
+			/* Setup prefix data, call it */
+			case ASN1_STATE_START:
+			if (!asn1_bio_setup_ex(b, ctx, ctx->prefix,
+				ASN1_STATE_PRE_COPY, ASN1_STATE_HEADER))
+				return 0;
+			break;
+
+			/* Copy any pre data first */
+			case ASN1_STATE_PRE_COPY:
+
+			ret = asn1_bio_flush_ex(b, ctx, ctx->prefix_free,
+							ASN1_STATE_HEADER);
+
+			if (ret <= 0)
+				goto done;
+
+			break;
+
+			case ASN1_STATE_HEADER:
+			ctx->buflen =
+				ASN1_object_size(0, inl, ctx->asn1_tag) - inl;
+			OPENSSL_assert(ctx->buflen <= ctx->bufsize);
+			p = ctx->buf;
+			ASN1_put_object(&p, 0, inl,
+					ctx->asn1_tag, ctx->asn1_class);
+			ctx->copylen = inl;
+			ctx->state = ASN1_STATE_HEADER_COPY;
+
+			break;
+
+			case ASN1_STATE_HEADER_COPY:	
+			ret = BIO_write(b->next_bio,
+					ctx->buf + ctx->bufpos, ctx->buflen);
+			if (ret <= 0)
+				goto done;
+
+			ctx->buflen -= ret;
+			if (ctx->buflen)
+				ctx->bufpos += ret;
+			else
+				{
+				ctx->bufpos = 0;
+				ctx->state = ASN1_STATE_DATA_COPY;
+				}
+
+			break;
+
+			case ASN1_STATE_DATA_COPY:
+
+			if (inl > ctx->copylen)
+				wrmax = ctx->copylen;
+			else
+				wrmax = inl;
+			ret = BIO_write(b->next_bio, in, wrmax);
+			if (ret <= 0)
+				break;
+			wrlen += ret;
+			ctx->copylen -= ret;
+			in += ret;
+			inl -= ret;
+
+			if (ctx->copylen == 0)
+				ctx->state = ASN1_STATE_HEADER;
+
+			if (inl == 0)
+				goto done;
+
+			break;
+
+			default:
+			BIO_clear_retry_flags(b);
+			return 0;
+
+			}
+
+		}
+
+	done:
+	BIO_clear_retry_flags(b);
+	BIO_copy_next_retry(b);
+
+	return (wrlen > 0) ? wrlen : ret;
+
+	}
+
+static int asn1_bio_flush_ex(BIO *b, BIO_ASN1_BUF_CTX *ctx,
+				asn1_ps_func *cleanup, asn1_bio_state_t next)
+	{
+	int ret;
+	if (ctx->ex_len <= 0)
+		return 1;
+	for(;;)
+		{
+		ret = BIO_write(b->next_bio, ctx->ex_buf + ctx->ex_pos,
+								ctx->ex_len);
+		if (ret <= 0)
+			break;
+		ctx->ex_len -= ret;
+		if (ctx->ex_len > 0)
+			ctx->ex_pos += ret;
+		else
+			{
+			if(cleanup)
+				cleanup(b, &ctx->ex_buf, &ctx->ex_len,
+								&ctx->ex_arg);
+			ctx->state = next;
+			ctx->ex_pos = 0;
+			break;
+			}
+		}
+	return ret;
+	}
+
+static int asn1_bio_setup_ex(BIO *b, BIO_ASN1_BUF_CTX *ctx,
+				asn1_ps_func *setup, 
+				asn1_bio_state_t ex_state,
+				asn1_bio_state_t other_state)
+	{
+	if (setup && !setup(b, &ctx->ex_buf, &ctx->ex_len, &ctx->ex_arg))
+		{
+		BIO_clear_retry_flags(b);
+		return 0;
+		}
+	if (ctx->ex_len > 0)
+		ctx->state = ex_state;
+	else
+		ctx->state = other_state;
+	return 1;
+	}
+
+static int asn1_bio_read(BIO *b, char *in , int inl)
+	{
+	if (!b->next_bio)
+		return 0;
+	return BIO_read(b->next_bio, in , inl);
+	}
+
+static int asn1_bio_puts(BIO *b, const char *str)
+	{
+	return asn1_bio_write(b, str, strlen(str));
+	}
+
+static int asn1_bio_gets(BIO *b, char *str, int size)
+	{
+	if (!b->next_bio)
+		return 0;
+	return BIO_gets(b->next_bio, str , size);
+	}
+
+static long asn1_bio_callback_ctrl(BIO *b, int cmd, bio_info_cb *fp)
+	{
+	if (b->next_bio == NULL) return(0);
+	return BIO_callback_ctrl(b->next_bio,cmd,fp);
+	}
+
+static long asn1_bio_ctrl(BIO *b, int cmd, long arg1, void *arg2)
+	{
+	BIO_ASN1_BUF_CTX *ctx;
+	BIO_ASN1_EX_FUNCS *ex_func;
+	long ret = 1;
+	ctx = (BIO_ASN1_BUF_CTX *) b->ptr;
+	if (ctx == NULL)
+		return 0;
+	switch(cmd)
+		{
+
+		case BIO_C_SET_PREFIX:
+		ex_func = arg2;
+		ctx->prefix  = ex_func->ex_func;
+		ctx->prefix_free  = ex_func->ex_free_func;
+		break;
+
+		case BIO_C_GET_PREFIX:
+		ex_func = arg2;
+		ex_func->ex_func = ctx->prefix;
+		ex_func->ex_free_func = ctx->prefix_free;
+		break;
+
+		case BIO_C_SET_SUFFIX:
+		ex_func = arg2;
+		ctx->suffix  = ex_func->ex_func;
+		ctx->suffix_free  = ex_func->ex_free_func;
+		break;
+
+		case BIO_C_GET_SUFFIX:
+		ex_func = arg2;
+		ex_func->ex_func = ctx->suffix;
+		ex_func->ex_free_func = ctx->suffix_free;
+		break;
+
+		case BIO_C_SET_EX_ARG:
+		ctx->ex_arg = arg2;
+		break;
+
+		case BIO_C_GET_EX_ARG:
+		*(void **)arg2 = ctx->ex_arg;
+		break;
+
+		case BIO_CTRL_FLUSH:
+		if (!b->next_bio)
+			return 0;
+
+		/* Call post function if possible */
+		if (ctx->state == ASN1_STATE_HEADER)
+			{
+			if (!asn1_bio_setup_ex(b, ctx, ctx->suffix,
+				ASN1_STATE_POST_COPY, ASN1_STATE_DONE))
+				return 0;
+			}
+
+		if (ctx->state == ASN1_STATE_POST_COPY)
+			{
+			ret = asn1_bio_flush_ex(b, ctx, ctx->suffix_free,
+							ASN1_STATE_DONE);
+			if (ret <= 0)
+				return ret;
+			}
+
+		if (ctx->state == ASN1_STATE_DONE)
+			return BIO_ctrl(b->next_bio, cmd, arg1, arg2);
+		else
+			{
+			BIO_clear_retry_flags(b);
+			return 0;
+			}
+		break;
+
+
+		default:
+		if (!b->next_bio)
+			return 0;
+		return BIO_ctrl(b->next_bio, cmd, arg1, arg2);
+
+		}
+
+	return ret;
+	}
+
+static int asn1_bio_set_ex(BIO *b, int cmd,
+		asn1_ps_func *ex_func, asn1_ps_func *ex_free_func)
+	{
+	BIO_ASN1_EX_FUNCS extmp;
+	extmp.ex_func = ex_func;
+	extmp.ex_free_func = ex_free_func;
+	return BIO_ctrl(b, cmd, 0, &extmp);
+	}
+
+static int asn1_bio_get_ex(BIO *b, int cmd,
+		asn1_ps_func **ex_func, asn1_ps_func **ex_free_func)
+	{
+	BIO_ASN1_EX_FUNCS extmp;
+	int ret;
+	ret = BIO_ctrl(b, cmd, 0, &extmp);
+	if (ret > 0)
+		{
+		*ex_func = extmp.ex_func;
+		*ex_free_func = extmp.ex_free_func;
+		}
+	return ret;
+	}
+
+int BIO_asn1_set_prefix(BIO *b, asn1_ps_func *prefix, asn1_ps_func *prefix_free)
+	{
+	return asn1_bio_set_ex(b, BIO_C_SET_PREFIX, prefix, prefix_free);
+	}
+
+int BIO_asn1_get_prefix(BIO *b, asn1_ps_func **pprefix, asn1_ps_func **pprefix_free)
+	{
+	return asn1_bio_get_ex(b, BIO_C_GET_PREFIX, pprefix, pprefix_free);
+	}
+
+int BIO_asn1_set_suffix(BIO *b, asn1_ps_func *suffix, asn1_ps_func *suffix_free)
+	{
+	return asn1_bio_set_ex(b, BIO_C_SET_SUFFIX, suffix, suffix_free);
+	}
+
+int BIO_asn1_get_suffix(BIO *b, asn1_ps_func **psuffix, asn1_ps_func **psuffix_free)
+	{
+	return asn1_bio_get_ex(b, BIO_C_GET_SUFFIX, psuffix, psuffix_free);
+	}
diff --git a/crypto/asn1/bio_ndef.c b/crypto/asn1/bio_ndef.c
new file mode 100644
index 000000000000..b91f97a1b139
--- /dev/null
+++ b/crypto/asn1/bio_ndef.c
@@ -0,0 +1,243 @@
+/* bio_ndef.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project.
+ */
+/* ====================================================================
+ * Copyright (c) 2008 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ */
+
+#include <openssl/asn1.h>
+#include <openssl/asn1t.h>
+#include <openssl/bio.h>
+#include <openssl/err.h>
+
+#include <stdio.h>
+
+/* Experimental NDEF ASN1 BIO support routines */
+
+/* The usage is quite simple, initialize an ASN1 structure,
+ * get a BIO from it then any data written through the BIO
+ * will end up translated to approptiate format on the fly.
+ * The data is streamed out and does *not* need to be
+ * all held in memory at once.
+ *
+ * When the BIO is flushed the output is finalized and any
+ * signatures etc written out.
+ *
+ * The BIO is a 'proper' BIO and can handle non blocking I/O
+ * correctly.
+ *
+ * The usage is simple. The implementation is *not*...
+ */
+
+/* BIO support data stored in the ASN1 BIO ex_arg */
+
+typedef struct ndef_aux_st
+	{
+	/* ASN1 structure this BIO refers to */
+	ASN1_VALUE *val;
+	const ASN1_ITEM *it;
+	/* Top of the BIO chain */
+	BIO *ndef_bio;
+	/* Output BIO */
+	BIO *out;
+	/* Boundary where content is inserted */
+	unsigned char **boundary;
+	/* DER buffer start */
+	unsigned char *derbuf;
+	} NDEF_SUPPORT;
+
+static int ndef_prefix(BIO *b, unsigned char **pbuf, int *plen, void *parg);
+static int ndef_prefix_free(BIO *b, unsigned char **pbuf, int *plen, void *parg);
+static int ndef_suffix(BIO *b, unsigned char **pbuf, int *plen, void *parg);
+static int ndef_suffix_free(BIO *b, unsigned char **pbuf, int *plen, void *parg);
+
+BIO *BIO_new_NDEF(BIO *out, ASN1_VALUE *val, const ASN1_ITEM *it)
+	{
+	NDEF_SUPPORT *ndef_aux = NULL;
+	BIO *asn_bio = NULL;
+	const ASN1_AUX *aux = it->funcs;
+	ASN1_STREAM_ARG sarg;
+
+	if (!aux || !aux->asn1_cb)
+		{
+		ASN1err(ASN1_F_BIO_NEW_NDEF, ASN1_R_STREAMING_NOT_SUPPORTED);
+		return NULL;
+		}
+	ndef_aux = OPENSSL_malloc(sizeof(NDEF_SUPPORT));
+	asn_bio = BIO_new(BIO_f_asn1());
+
+	/* ASN1 bio needs to be next to output BIO */
+
+	out = BIO_push(asn_bio, out);
+
+	if (!ndef_aux || !asn_bio || !out)
+		goto err;
+
+	BIO_asn1_set_prefix(asn_bio, ndef_prefix, ndef_prefix_free);
+	BIO_asn1_set_suffix(asn_bio, ndef_suffix, ndef_suffix_free);
+
+	/* Now let callback prepend any digest, cipher etc BIOs
+	 * ASN1 structure needs.
+	 */
+
+	sarg.out = out;
+	sarg.ndef_bio = NULL;
+	sarg.boundary = NULL;
+
+	if (aux->asn1_cb(ASN1_OP_STREAM_PRE, &val, it, &sarg) <= 0)
+		goto err;
+
+	ndef_aux->val = val;
+	ndef_aux->it = it;
+	ndef_aux->ndef_bio = sarg.ndef_bio;
+	ndef_aux->boundary = sarg.boundary;
+	ndef_aux->out = out;
+
+	BIO_ctrl(asn_bio, BIO_C_SET_EX_ARG, 0, ndef_aux);
+
+	return sarg.ndef_bio;
+
+	err:
+	if (asn_bio)
+		BIO_free(asn_bio);
+	if (ndef_aux)
+		OPENSSL_free(ndef_aux);
+	return NULL;
+	}
+
+static int ndef_prefix(BIO *b, unsigned char **pbuf, int *plen, void *parg)
+	{
+	NDEF_SUPPORT *ndef_aux;
+	unsigned char *p;
+	int derlen;
+
+	if (!parg)
+		return 0;
+
+	ndef_aux = *(NDEF_SUPPORT **)parg;
+
+	derlen = ASN1_item_ndef_i2d(ndef_aux->val, NULL, ndef_aux->it);
+	p = OPENSSL_malloc(derlen);
+	ndef_aux->derbuf = p;
+	*pbuf = p;
+	derlen = ASN1_item_ndef_i2d(ndef_aux->val, &p, ndef_aux->it);
+
+	if (!*ndef_aux->boundary)
+		return 0;
+
+	*plen = *ndef_aux->boundary - *pbuf;
+
+	return 1;
+	}
+
+static int ndef_prefix_free(BIO *b, unsigned char **pbuf, int *plen, void *parg)
+	{
+	NDEF_SUPPORT *ndef_aux;
+
+	if (!parg)
+		return 0;
+
+	ndef_aux = *(NDEF_SUPPORT **)parg;
+
+	if (ndef_aux->derbuf)
+		OPENSSL_free(ndef_aux->derbuf);
+
+	ndef_aux->derbuf = NULL;
+	*pbuf = NULL;
+	*plen = 0;
+	return 1;
+	}
+
+static int ndef_suffix_free(BIO *b, unsigned char **pbuf, int *plen, void *parg)
+	{
+	NDEF_SUPPORT **pndef_aux = (NDEF_SUPPORT **)parg;
+	if (!ndef_prefix_free(b, pbuf, plen, parg))
+		return 0;
+	OPENSSL_free(*pndef_aux);
+	*pndef_aux = NULL;
+	return 1;
+	}
+
+static int ndef_suffix(BIO *b, unsigned char **pbuf, int *plen, void *parg)
+	{
+	NDEF_SUPPORT *ndef_aux;
+	unsigned char *p;
+	int derlen;
+	const ASN1_AUX *aux;
+	ASN1_STREAM_ARG sarg;
+
+	if (!parg)
+		return 0;
+
+	ndef_aux = *(NDEF_SUPPORT **)parg;
+
+	aux = ndef_aux->it->funcs;
+
+	/* Finalize structures */
+	sarg.ndef_bio = ndef_aux->ndef_bio;
+	sarg.out = ndef_aux->out;
+	sarg.boundary = ndef_aux->boundary;
+	if (aux->asn1_cb(ASN1_OP_STREAM_POST,
+				&ndef_aux->val, ndef_aux->it, &sarg) <= 0)
+		return 0;
+
+	derlen = ASN1_item_ndef_i2d(ndef_aux->val, NULL, ndef_aux->it);
+	p = OPENSSL_malloc(derlen);
+	ndef_aux->derbuf = p;
+	*pbuf = p;
+	derlen = ASN1_item_ndef_i2d(ndef_aux->val, &p, ndef_aux->it);
+
+	if (!*ndef_aux->boundary)
+		return 0;
+	*pbuf = *ndef_aux->boundary;
+	*plen = derlen - (*ndef_aux->boundary - ndef_aux->derbuf);
+
+	return 1;
+	}
diff --git a/crypto/asn1/charmap.h b/crypto/asn1/charmap.h
index bd020a9562fd..b55e6387259c 100644
--- a/crypto/asn1/charmap.h
+++ b/crypto/asn1/charmap.h
@@ -2,7 +2,7 @@
  * Mask of various character properties
  */
 
-static unsigned char char_type[] = {
+static const unsigned char char_type[] = {
  2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
  2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
 120, 0, 1,40, 0, 0, 0,16,16,16, 0,25,25,16,16,16,
diff --git a/crypto/asn1/d2i_pr.c b/crypto/asn1/d2i_pr.c
index 207ccda5ac95..28289447772c 100644
--- a/crypto/asn1/d2i_pr.c
+++ b/crypto/asn1/d2i_pr.c
@@ -61,16 +61,12 @@
 #include <openssl/bn.h>
 #include <openssl/evp.h>
 #include <openssl/objects.h>
-#include <openssl/asn1.h>
-#ifndef OPENSSL_NO_RSA
-#include <openssl/rsa.h>
-#endif
-#ifndef OPENSSL_NO_DSA
-#include <openssl/dsa.h>
-#endif
-#ifndef OPENSSL_NO_EC
-#include <openssl/ec.h>
+#ifndef OPENSSL_NO_ENGINE
+#include <openssl/engine.h>
 #endif
+#include <openssl/x509.h>
+#include <openssl/asn1.h>
+#include "asn1_locl.h"
 
 EVP_PKEY *d2i_PrivateKey(int type, EVP_PKEY **a, const unsigned char **pp,
 	     long length)
@@ -85,47 +81,43 @@ EVP_PKEY *d2i_PrivateKey(int type, EVP_PKEY **a, const unsigned char **pp,
 			return(NULL);
 			}
 		}
-	else	ret= *a;
-
-	ret->save_type=type;
-	ret->type=EVP_PKEY_type(type);
-	switch (ret->type)
+	else
 		{
-#ifndef OPENSSL_NO_RSA
-	case EVP_PKEY_RSA:
-		if ((ret->pkey.rsa=d2i_RSAPrivateKey(NULL,
-			(const unsigned char **)pp,length)) == NULL) /* TMP UGLY CAST */
+		ret= *a;
+#ifndef OPENSSL_NO_ENGINE
+		if (ret->engine)
 			{
-			ASN1err(ASN1_F_D2I_PRIVATEKEY,ERR_R_ASN1_LIB);
-			goto err;
+			ENGINE_finish(ret->engine);
+			ret->engine = NULL;
 			}
-		break;
 #endif
-#ifndef OPENSSL_NO_DSA
-	case EVP_PKEY_DSA:
-		if ((ret->pkey.dsa=d2i_DSAPrivateKey(NULL,
-			(const unsigned char **)pp,length)) == NULL) /* TMP UGLY CAST */
+		}
+
+	if (!EVP_PKEY_set_type(ret, type))
+		{
+		ASN1err(ASN1_F_D2I_PRIVATEKEY,ASN1_R_UNKNOWN_PUBLIC_KEY_TYPE);
+		goto err;
+		}
+
+	if (!ret->ameth->old_priv_decode ||
+			!ret->ameth->old_priv_decode(ret, pp, length))
+		{
+		if (ret->ameth->priv_decode) 
 			{
-			ASN1err(ASN1_F_D2I_PRIVATEKEY,ERR_R_ASN1_LIB);
-			goto err;
-			}
-		break;
-#endif
-#ifndef OPENSSL_NO_EC
-	case EVP_PKEY_EC:
-		if ((ret->pkey.ec = d2i_ECPrivateKey(NULL, 
-			(const unsigned char **)pp, length)) == NULL)
+			PKCS8_PRIV_KEY_INFO *p8=NULL;
+			p8=d2i_PKCS8_PRIV_KEY_INFO(NULL,pp,length);
+			if (!p8) goto err;
+			EVP_PKEY_free(ret);
+			ret = EVP_PKCS82PKEY(p8);
+			PKCS8_PRIV_KEY_INFO_free(p8);
+
+			} 
+		else 
 			{
-			ASN1err(ASN1_F_D2I_PRIVATEKEY, ERR_R_ASN1_LIB);
+			ASN1err(ASN1_F_D2I_PRIVATEKEY,ERR_R_ASN1_LIB);
 			goto err;
 			}
-		break;
-#endif
-	default:
-		ASN1err(ASN1_F_D2I_PRIVATEKEY,ASN1_R_UNKNOWN_PUBLIC_KEY_TYPE);
-		goto err;
-		/* break; */
-		}
+		}	
 	if (a != NULL) (*a)=ret;
 	return(ret);
 err:
@@ -146,8 +138,7 @@ EVP_PKEY *d2i_AutoPrivateKey(EVP_PKEY **a, const unsigned char **pp,
 	 * by analyzing it we can determine the passed structure: this
 	 * assumes the input is surrounded by an ASN1 SEQUENCE.
 	 */
-	inkey = d2i_ASN1_SET_OF_ASN1_TYPE(NULL, &p, length, d2i_ASN1_TYPE, 
-			ASN1_TYPE_free, V_ASN1_SEQUENCE, V_ASN1_UNIVERSAL);
+	inkey = d2i_ASN1_SEQUENCE_ANY(NULL, &p, length);
 	/* Since we only need to discern "traditional format" RSA and DSA
 	 * keys we can just count the elements.
          */
@@ -155,6 +146,24 @@ EVP_PKEY *d2i_AutoPrivateKey(EVP_PKEY **a, const unsigned char **pp,
 		keytype = EVP_PKEY_DSA;
 	else if (sk_ASN1_TYPE_num(inkey) == 4)
 		keytype = EVP_PKEY_EC;
+	else if (sk_ASN1_TYPE_num(inkey) == 3)  
+		{ /* This seems to be PKCS8, not traditional format */
+			PKCS8_PRIV_KEY_INFO *p8 = d2i_PKCS8_PRIV_KEY_INFO(NULL,pp,length);
+			EVP_PKEY *ret;
+
+			sk_ASN1_TYPE_pop_free(inkey, ASN1_TYPE_free);
+			if (!p8) 
+				{
+				ASN1err(ASN1_F_D2I_AUTOPRIVATEKEY,ASN1_R_UNSUPPORTED_PUBLIC_KEY_TYPE);
+				return NULL;
+				}
+			ret = EVP_PKCS82PKEY(p8);
+			PKCS8_PRIV_KEY_INFO_free(p8);
+			if (a) {
+				*a = ret;
+			}	
+			return ret;
+		}
 	else keytype = EVP_PKEY_RSA;
 	sk_ASN1_TYPE_pop_free(inkey, ASN1_TYPE_free);
 	return d2i_PrivateKey(keytype, a, pp, length);
diff --git a/crypto/asn1/d2i_pu.c b/crypto/asn1/d2i_pu.c
index 3694f51a8c0a..c8f39ceb037b 100644
--- a/crypto/asn1/d2i_pu.c
+++ b/crypto/asn1/d2i_pu.c
@@ -87,9 +87,13 @@ EVP_PKEY *d2i_PublicKey(int type, EVP_PKEY **a, const unsigned char **pp,
 		}
 	else	ret= *a;
 
-	ret->save_type=type;
-	ret->type=EVP_PKEY_type(type);
-	switch (ret->type)
+	if (!EVP_PKEY_set_type(ret, type))
+		{
+		ASN1err(ASN1_F_D2I_PUBLICKEY,ERR_R_EVP_LIB);
+		goto err;
+		}
+
+	switch (EVP_PKEY_id(ret))
 		{
 #ifndef OPENSSL_NO_RSA
 	case EVP_PKEY_RSA:
diff --git a/crypto/asn1/i2d_pr.c b/crypto/asn1/i2d_pr.c
index 0be52c5b76ac..e398b6266612 100644
--- a/crypto/asn1/i2d_pr.c
+++ b/crypto/asn1/i2d_pr.c
@@ -58,41 +58,22 @@
 
 #include <stdio.h>
 #include "cryptlib.h"
-#include <openssl/bn.h>
 #include <openssl/evp.h>
-#include <openssl/objects.h>
-#ifndef OPENSSL_NO_RSA
-#include <openssl/rsa.h>
-#endif
-#ifndef OPENSSL_NO_DSA
-#include <openssl/dsa.h>
-#endif
-#ifndef OPENSSL_NO_EC
-#include <openssl/ec.h>
-#endif
+#include <openssl/x509.h>
+#include "asn1_locl.h"
 
 int i2d_PrivateKey(EVP_PKEY *a, unsigned char **pp)
 	{
-#ifndef OPENSSL_NO_RSA
-	if (a->type == EVP_PKEY_RSA)
+	if (a->ameth && a->ameth->old_priv_encode)
 		{
-		return(i2d_RSAPrivateKey(a->pkey.rsa,pp));
+		return a->ameth->old_priv_encode(a, pp);
 		}
-	else
-#endif
-#ifndef OPENSSL_NO_DSA
-	if (a->type == EVP_PKEY_DSA)
-		{
-		return(i2d_DSAPrivateKey(a->pkey.dsa,pp));
-		}
-#endif
-#ifndef OPENSSL_NO_EC
-	if (a->type == EVP_PKEY_EC)
-		{
-		return(i2d_ECPrivateKey(a->pkey.ec, pp));
-		}
-#endif
-
+	if (a->ameth && a->ameth->priv_encode) {
+		PKCS8_PRIV_KEY_INFO *p8 = EVP_PKEY2PKCS8(a);
+		int ret = i2d_PKCS8_PRIV_KEY_INFO(p8,pp);
+		PKCS8_PRIV_KEY_INFO_free(p8);
+		return ret;
+	}	
 	ASN1err(ASN1_F_I2D_PRIVATEKEY,ASN1_R_UNSUPPORTED_PUBLIC_KEY_TYPE);
 	return(-1);
 	}
diff --git a/crypto/asn1/n_pkey.c b/crypto/asn1/n_pkey.c
index e7d04390625b..e2517399335c 100644
--- a/crypto/asn1/n_pkey.c
+++ b/crypto/asn1/n_pkey.c
@@ -129,6 +129,7 @@ int i2d_RSA_NET(const RSA *a, unsigned char **pp,
 	unsigned char buf[256],*zz;
 	unsigned char key[EVP_MAX_KEY_LENGTH];
 	EVP_CIPHER_CTX ctx;
+	EVP_CIPHER_CTX_init(&ctx);
 
 	if (a == NULL) return(0);
 
@@ -206,24 +207,28 @@ int i2d_RSA_NET(const RSA *a, unsigned char **pp,
 	i = strlen((char *)buf);
 	/* If the key is used for SGC the algorithm is modified a little. */
 	if(sgckey) {
-		EVP_Digest(buf, i, buf, NULL, EVP_md5(), NULL);
+		if (!EVP_Digest(buf, i, buf, NULL, EVP_md5(), NULL))
+			goto err;
 		memcpy(buf + 16, "SGCKEYSALT", 10);
 		i = 26;
 	}
 
-	EVP_BytesToKey(EVP_rc4(),EVP_md5(),NULL,buf,i,1,key,NULL);
+	if (!EVP_BytesToKey(EVP_rc4(),EVP_md5(),NULL,buf,i,1,key,NULL))
+		goto err;
 	OPENSSL_cleanse(buf,256);
 
 	/* Encrypt private key in place */
 	zz = enckey->enckey->digest->data;
-	EVP_CIPHER_CTX_init(&ctx);
-	EVP_EncryptInit_ex(&ctx,EVP_rc4(),NULL,key,NULL);
-	EVP_EncryptUpdate(&ctx,zz,&i,zz,pkeylen);
-	EVP_EncryptFinal_ex(&ctx,zz + i,&j);
-	EVP_CIPHER_CTX_cleanup(&ctx);
+	if (!EVP_EncryptInit_ex(&ctx,EVP_rc4(),NULL,key,NULL))
+		goto err;
+	if (!EVP_EncryptUpdate(&ctx,zz,&i,zz,pkeylen))
+		goto err;
+	if (!EVP_EncryptFinal_ex(&ctx,zz + i,&j))
+		goto err;
 
 	ret = i2d_NETSCAPE_ENCRYPTED_PKEY(enckey, pp);
 err:
+	EVP_CIPHER_CTX_cleanup(&ctx);
 	NETSCAPE_ENCRYPTED_PKEY_free(enckey);
 	NETSCAPE_PKEY_free(pkey);
 	return(ret);
@@ -288,6 +293,7 @@ static RSA *d2i_RSA_NET_2(RSA **a, ASN1_OCTET_STRING *os,
 	const unsigned char *zz;
 	unsigned char key[EVP_MAX_KEY_LENGTH];
 	EVP_CIPHER_CTX ctx;
+	EVP_CIPHER_CTX_init(&ctx);
 
 	i=cb((char *)buf,256,"Enter Private Key password:",0);
 	if (i != 0)
@@ -298,19 +304,22 @@ static RSA *d2i_RSA_NET_2(RSA **a, ASN1_OCTET_STRING *os,
 
 	i = strlen((char *)buf);
 	if(sgckey){
-		EVP_Digest(buf, i, buf, NULL, EVP_md5(), NULL);
+		if (!EVP_Digest(buf, i, buf, NULL, EVP_md5(), NULL))
+			goto err;
 		memcpy(buf + 16, "SGCKEYSALT", 10);
 		i = 26;
 	}
 		
-	EVP_BytesToKey(EVP_rc4(),EVP_md5(),NULL,buf,i,1,key,NULL);
+	if (!EVP_BytesToKey(EVP_rc4(),EVP_md5(),NULL,buf,i,1,key,NULL))
+		goto err;
 	OPENSSL_cleanse(buf,256);
 
-	EVP_CIPHER_CTX_init(&ctx);
-	EVP_DecryptInit_ex(&ctx,EVP_rc4(),NULL, key,NULL);
-	EVP_DecryptUpdate(&ctx,os->data,&i,os->data,os->length);
-	EVP_DecryptFinal_ex(&ctx,&(os->data[i]),&j);
-	EVP_CIPHER_CTX_cleanup(&ctx);
+	if (!EVP_DecryptInit_ex(&ctx,EVP_rc4(),NULL, key,NULL))
+		goto err;
+	if (!EVP_DecryptUpdate(&ctx,os->data,&i,os->data,os->length))
+		goto err;
+	if (!EVP_DecryptFinal_ex(&ctx,&(os->data[i]),&j))
+		goto err;
 	os->length=i+j;
 
 	zz=os->data;
@@ -328,6 +337,7 @@ static RSA *d2i_RSA_NET_2(RSA **a, ASN1_OCTET_STRING *os,
 		goto err;
 		}
 err:
+	EVP_CIPHER_CTX_cleanup(&ctx);
 	NETSCAPE_PKEY_free(pkey);
 	return(ret);
 	}
diff --git a/crypto/asn1/nsseq.c b/crypto/asn1/nsseq.c
index e551c57d59d0..b8c420223090 100644
--- a/crypto/asn1/nsseq.c
+++ b/crypto/asn1/nsseq.c
@@ -3,7 +3,7 @@
  * project 1999.
  */
 /* ====================================================================
- * Copyright (c) 1999 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -62,7 +62,8 @@
 #include <openssl/x509.h>
 #include <openssl/objects.h>
 
-static int nsseq_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int nsseq_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+							void *exarg)
 {
 	if(operation == ASN1_OP_NEW_POST) {
 		NETSCAPE_CERT_SEQUENCE *nsseq;
diff --git a/crypto/asn1/p5_pbe.c b/crypto/asn1/p5_pbe.c
index c4582f8041e4..94bc38b99f93 100644
--- a/crypto/asn1/p5_pbe.c
+++ b/crypto/asn1/p5_pbe.c
@@ -71,61 +71,78 @@ ASN1_SEQUENCE(PBEPARAM) = {
 
 IMPLEMENT_ASN1_FUNCTIONS(PBEPARAM)
 
-/* Return an algorithm identifier for a PKCS#5 PBE algorithm */
 
-X509_ALGOR *PKCS5_pbe_set(int alg, int iter, unsigned char *salt,
-	     int saltlen)
-{
+/* Set an algorithm identifier for a PKCS#5 PBE algorithm */
+
+int PKCS5_pbe_set0_algor(X509_ALGOR *algor, int alg, int iter,
+				const unsigned char *salt, int saltlen)
+	{
 	PBEPARAM *pbe=NULL;
-	ASN1_OBJECT *al;
-	X509_ALGOR *algor;
-	ASN1_TYPE *astype=NULL;
+	ASN1_STRING *pbe_str=NULL;
+	unsigned char *sstr;
 
-	if (!(pbe = PBEPARAM_new ())) {
-		ASN1err(ASN1_F_PKCS5_PBE_SET,ERR_R_MALLOC_FAILURE);
+	pbe = PBEPARAM_new();
+	if (!pbe)
+		{
+		ASN1err(ASN1_F_PKCS5_PBE_SET0_ALGOR,ERR_R_MALLOC_FAILURE);
 		goto err;
-	}
-	if(iter <= 0) iter = PKCS5_DEFAULT_ITER;
-	if (!ASN1_INTEGER_set(pbe->iter, iter)) {
-		ASN1err(ASN1_F_PKCS5_PBE_SET,ERR_R_MALLOC_FAILURE);
+		}
+	if(iter <= 0)
+		iter = PKCS5_DEFAULT_ITER;
+	if (!ASN1_INTEGER_set(pbe->iter, iter))
+		{
+		ASN1err(ASN1_F_PKCS5_PBE_SET0_ALGOR,ERR_R_MALLOC_FAILURE);
 		goto err;
-	}
-	if (!saltlen) saltlen = PKCS5_SALT_LEN;
-	if (!(pbe->salt->data = OPENSSL_malloc (saltlen))) {
-		ASN1err(ASN1_F_PKCS5_PBE_SET,ERR_R_MALLOC_FAILURE);
+		}
+	if (!saltlen)
+		saltlen = PKCS5_SALT_LEN;
+	if (!ASN1_STRING_set(pbe->salt, NULL, saltlen))
+		{
+		ASN1err(ASN1_F_PKCS5_PBE_SET0_ALGOR,ERR_R_MALLOC_FAILURE);
 		goto err;
-	}
-	pbe->salt->length = saltlen;
-	if (salt) memcpy (pbe->salt->data, salt, saltlen);
-	else if (RAND_pseudo_bytes (pbe->salt->data, saltlen) < 0)
+		}
+	sstr = ASN1_STRING_data(pbe->salt);
+	if (salt)
+		memcpy(sstr, salt, saltlen);
+	else if (RAND_pseudo_bytes(sstr, saltlen) < 0)
 		goto err;
 
-	if (!(astype = ASN1_TYPE_new())) {
-		ASN1err(ASN1_F_PKCS5_PBE_SET,ERR_R_MALLOC_FAILURE);
+	if(!ASN1_item_pack(pbe, ASN1_ITEM_rptr(PBEPARAM), &pbe_str))
+		{
+		ASN1err(ASN1_F_PKCS5_PBE_SET0_ALGOR,ERR_R_MALLOC_FAILURE);
 		goto err;
-	}
+		}
 
-	astype->type = V_ASN1_SEQUENCE;
-	if(!ASN1_pack_string_of(PBEPARAM, pbe, i2d_PBEPARAM,
-				&astype->value.sequence)) {
-		ASN1err(ASN1_F_PKCS5_PBE_SET,ERR_R_MALLOC_FAILURE);
-		goto err;
-	}
-	PBEPARAM_free (pbe);
+	PBEPARAM_free(pbe);
 	pbe = NULL;
-	
-	al = OBJ_nid2obj(alg); /* never need to free al */
-	if (!(algor = X509_ALGOR_new())) {
-		ASN1err(ASN1_F_PKCS5_PBE_SET,ERR_R_MALLOC_FAILURE);
-		goto err;
-	}
-	ASN1_OBJECT_free(algor->algorithm);
-	algor->algorithm = al;
-	algor->parameter = astype;
 
-	return (algor);
+	if (X509_ALGOR_set0(algor, OBJ_nid2obj(alg), V_ASN1_SEQUENCE, pbe_str))
+		return 1;
+
 err:
-	if (pbe != NULL) PBEPARAM_free(pbe);
-	if (astype != NULL) ASN1_TYPE_free(astype);
+	if (pbe != NULL)
+		PBEPARAM_free(pbe);
+	if (pbe_str != NULL)
+		ASN1_STRING_free(pbe_str);
+	return 0;
+	}
+
+/* Return an algorithm identifier for a PKCS#5 PBE algorithm */
+
+X509_ALGOR *PKCS5_pbe_set(int alg, int iter,
+				const unsigned char *salt, int saltlen)
+	{
+	X509_ALGOR *ret;
+	ret = X509_ALGOR_new();
+	if (!ret)
+		{
+		ASN1err(ASN1_F_PKCS5_PBE_SET,ERR_R_MALLOC_FAILURE);
+		return NULL;
+		}
+
+	if (PKCS5_pbe_set0_algor(ret, alg, iter, salt, saltlen)) 
+		return ret;
+
+	X509_ALGOR_free(ret);
 	return NULL;
-}
+	}
diff --git a/crypto/asn1/p5_pbev2.c b/crypto/asn1/p5_pbev2.c
index 2b0516afeed3..4ea683036b17 100644
--- a/crypto/asn1/p5_pbev2.c
+++ b/crypto/asn1/p5_pbev2.c
@@ -82,23 +82,24 @@ IMPLEMENT_ASN1_FUNCTIONS(PBKDF2PARAM)
 
 /* Return an algorithm identifier for a PKCS#5 v2.0 PBE algorithm:
  * yes I know this is horrible!
+ *
+ * Extended version to allow application supplied PRF NID and IV.
  */
 
-X509_ALGOR *PKCS5_pbe2_set(const EVP_CIPHER *cipher, int iter,
-				 unsigned char *salt, int saltlen)
+X509_ALGOR *PKCS5_pbe2_set_iv(const EVP_CIPHER *cipher, int iter,
+				 unsigned char *salt, int saltlen,
+				 unsigned char *aiv, int prf_nid)
 {
 	X509_ALGOR *scheme = NULL, *kalg = NULL, *ret = NULL;
-	int alg_nid;
+	int alg_nid, keylen;
 	EVP_CIPHER_CTX ctx;
 	unsigned char iv[EVP_MAX_IV_LENGTH];
-	PBKDF2PARAM *kdf = NULL;
 	PBE2PARAM *pbe2 = NULL;
-	ASN1_OCTET_STRING *osalt = NULL;
 	ASN1_OBJECT *obj;
 
 	alg_nid = EVP_CIPHER_type(cipher);
 	if(alg_nid == NID_undef) {
-		ASN1err(ASN1_F_PKCS5_PBE2_SET,
+		ASN1err(ASN1_F_PKCS5_PBE2_SET_IV,
 				ASN1_R_CIPHER_HAS_NO_OBJECT_IDENTIFIER);
 		goto err;
 	}
@@ -113,63 +114,51 @@ X509_ALGOR *PKCS5_pbe2_set(const EVP_CIPHER *cipher, int iter,
 	if(!(scheme->parameter = ASN1_TYPE_new())) goto merr;
 
 	/* Create random IV */
-	if (EVP_CIPHER_iv_length(cipher) &&
-		RAND_pseudo_bytes(iv, EVP_CIPHER_iv_length(cipher)) < 0)
-  		goto err;
+	if (EVP_CIPHER_iv_length(cipher))
+		{
+		if (aiv)
+			memcpy(iv, aiv, EVP_CIPHER_iv_length(cipher));
+		else if (RAND_pseudo_bytes(iv, EVP_CIPHER_iv_length(cipher)) < 0)
+  			goto err;
+		}
 
 	EVP_CIPHER_CTX_init(&ctx);
 
-	/* Dummy cipherinit to just setup the IV */
-	EVP_CipherInit_ex(&ctx, cipher, NULL, NULL, iv, 0);
+	/* Dummy cipherinit to just setup the IV, and PRF */
+	if (!EVP_CipherInit_ex(&ctx, cipher, NULL, NULL, iv, 0))
+		goto err;
 	if(EVP_CIPHER_param_to_asn1(&ctx, scheme->parameter) < 0) {
-		ASN1err(ASN1_F_PKCS5_PBE2_SET,
+		ASN1err(ASN1_F_PKCS5_PBE2_SET_IV,
 					ASN1_R_ERROR_SETTING_CIPHER_PARAMS);
 		EVP_CIPHER_CTX_cleanup(&ctx);
 		goto err;
 	}
+	/* If prf NID unspecified see if cipher has a preference.
+	 * An error is OK here: just means use default PRF.
+	 */
+	if ((prf_nid == -1) && 
+	EVP_CIPHER_CTX_ctrl(&ctx, EVP_CTRL_PBE_PRF_NID, 0, &prf_nid) <= 0)
+		{
+		ERR_clear_error();
+		prf_nid = NID_hmacWithSHA1;
+		}
 	EVP_CIPHER_CTX_cleanup(&ctx);
 
-	if(!(kdf = PBKDF2PARAM_new())) goto merr;
-	if(!(osalt = M_ASN1_OCTET_STRING_new())) goto merr;
-
-	if (!saltlen) saltlen = PKCS5_SALT_LEN;
-	if (!(osalt->data = OPENSSL_malloc (saltlen))) goto merr;
-	osalt->length = saltlen;
-	if (salt) memcpy (osalt->data, salt, saltlen);
-	else if (RAND_pseudo_bytes (osalt->data, saltlen) < 0) goto merr;
-
-	if(iter <= 0) iter = PKCS5_DEFAULT_ITER;
-	if(!ASN1_INTEGER_set(kdf->iter, iter)) goto merr;
-
-	/* Now include salt in kdf structure */
-	kdf->salt->value.octet_string = osalt;
-	kdf->salt->type = V_ASN1_OCTET_STRING;
-	osalt = NULL;
-
 	/* If its RC2 then we'd better setup the key length */
 
-	if(alg_nid == NID_rc2_cbc) {
-		if(!(kdf->keylength = M_ASN1_INTEGER_new())) goto merr;
-		if(!ASN1_INTEGER_set (kdf->keylength,
-				 EVP_CIPHER_key_length(cipher))) goto merr;
-	}
-
-	/* prf can stay NULL because we are using hmacWithSHA1 */
-
-	/* Now setup the PBE2PARAM keyfunc structure */
+	if(alg_nid == NID_rc2_cbc)
+		keylen = EVP_CIPHER_key_length(cipher);
+	else
+		keylen = -1;
 
-	pbe2->keyfunc->algorithm = OBJ_nid2obj(NID_id_pbkdf2);
+	/* Setup keyfunc */
 
-	/* Encode PBKDF2PARAM into parameter of pbe2 */
+	X509_ALGOR_free(pbe2->keyfunc);
 
-	if(!(pbe2->keyfunc->parameter = ASN1_TYPE_new())) goto merr;
+	pbe2->keyfunc = PKCS5_pbkdf2_set(iter, salt, saltlen, prf_nid, keylen);
 
-	if(!ASN1_pack_string_of(PBKDF2PARAM, kdf, i2d_PBKDF2PARAM,
-			 &pbe2->keyfunc->parameter->value.sequence)) goto merr;
-	pbe2->keyfunc->parameter->type = V_ASN1_SEQUENCE;
-
-	PBKDF2PARAM_free(kdf);
-	kdf = NULL;
+	if (!pbe2->keyfunc)
+		goto merr;
 
 	/* Now set up top level AlgorithmIdentifier */
 
@@ -180,7 +169,7 @@ X509_ALGOR *PKCS5_pbe2_set(const EVP_CIPHER *cipher, int iter,
 
 	/* Encode PBE2PARAM into parameter */
 
-	if(!ASN1_pack_string_of(PBE2PARAM, pbe2, i2d_PBE2PARAM,
+	if(!ASN1_item_pack(pbe2, ASN1_ITEM_rptr(PBE2PARAM),
 				 &ret->parameter->value.sequence)) goto merr;
 	ret->parameter->type = V_ASN1_SEQUENCE;
 
@@ -190,16 +179,102 @@ X509_ALGOR *PKCS5_pbe2_set(const EVP_CIPHER *cipher, int iter,
 	return ret;
 
 	merr:
-	ASN1err(ASN1_F_PKCS5_PBE2_SET,ERR_R_MALLOC_FAILURE);
+	ASN1err(ASN1_F_PKCS5_PBE2_SET_IV,ERR_R_MALLOC_FAILURE);
 
 	err:
 	PBE2PARAM_free(pbe2);
 	/* Note 'scheme' is freed as part of pbe2 */
-	M_ASN1_OCTET_STRING_free(osalt);
-	PBKDF2PARAM_free(kdf);
 	X509_ALGOR_free(kalg);
 	X509_ALGOR_free(ret);
 
 	return NULL;
 
 }
+
+X509_ALGOR *PKCS5_pbe2_set(const EVP_CIPHER *cipher, int iter,
+				 unsigned char *salt, int saltlen)
+	{
+	return PKCS5_pbe2_set_iv(cipher, iter, salt, saltlen, NULL, -1);
+	}
+
+X509_ALGOR *PKCS5_pbkdf2_set(int iter, unsigned char *salt, int saltlen,
+				int prf_nid, int keylen)
+	{
+	X509_ALGOR *keyfunc = NULL;
+	PBKDF2PARAM *kdf = NULL;
+	ASN1_OCTET_STRING *osalt = NULL;
+
+	if(!(kdf = PBKDF2PARAM_new()))
+		goto merr;
+	if(!(osalt = M_ASN1_OCTET_STRING_new()))
+		goto merr;
+
+	kdf->salt->value.octet_string = osalt;
+	kdf->salt->type = V_ASN1_OCTET_STRING;
+
+	if (!saltlen)
+		saltlen = PKCS5_SALT_LEN;
+	if (!(osalt->data = OPENSSL_malloc (saltlen)))
+		goto merr;
+
+	osalt->length = saltlen;
+
+	if (salt)
+		memcpy (osalt->data, salt, saltlen);
+	else if (RAND_pseudo_bytes (osalt->data, saltlen) < 0)
+		goto merr;
+
+	if(iter <= 0)
+		iter = PKCS5_DEFAULT_ITER;
+
+	if(!ASN1_INTEGER_set(kdf->iter, iter))
+		goto merr;
+
+	/* If have a key len set it up */
+
+	if(keylen > 0) 
+		{
+		if(!(kdf->keylength = M_ASN1_INTEGER_new()))
+			goto merr;
+		if(!ASN1_INTEGER_set (kdf->keylength, keylen))
+			goto merr;
+		}
+
+	/* prf can stay NULL if we are using hmacWithSHA1 */
+	if (prf_nid > 0 && prf_nid != NID_hmacWithSHA1)
+		{
+		kdf->prf = X509_ALGOR_new();
+		if (!kdf->prf)
+			goto merr;
+		X509_ALGOR_set0(kdf->prf, OBJ_nid2obj(prf_nid),
+					V_ASN1_NULL, NULL);
+		}
+
+	/* Finally setup the keyfunc structure */
+
+	keyfunc = X509_ALGOR_new();
+	if (!keyfunc)
+		goto merr;
+
+	keyfunc->algorithm = OBJ_nid2obj(NID_id_pbkdf2);
+
+	/* Encode PBKDF2PARAM into parameter of pbe2 */
+
+	if(!(keyfunc->parameter = ASN1_TYPE_new()))
+		goto merr;
+
+	if(!ASN1_item_pack(kdf, ASN1_ITEM_rptr(PBKDF2PARAM),
+			 &keyfunc->parameter->value.sequence))
+		goto merr;
+	keyfunc->parameter->type = V_ASN1_SEQUENCE;
+
+	PBKDF2PARAM_free(kdf);
+	return keyfunc;
+
+	merr:
+	ASN1err(ASN1_F_PKCS5_PBKDF2_SET,ERR_R_MALLOC_FAILURE);
+	PBKDF2PARAM_free(kdf);
+	X509_ALGOR_free(keyfunc);
+	return NULL;
+	}
+
diff --git a/crypto/asn1/p8_key.c b/crypto/asn1/p8_key.c
deleted file mode 100644
index 3a31248e14be..000000000000
--- a/crypto/asn1/p8_key.c
+++ /dev/null
@@ -1,131 +0,0 @@
-/* crypto/asn1/p8_key.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#include <stdio.h>
-#include "cryptlib.h"
-#include <openssl/asn1_mac.h>
-#include <openssl/objects.h>
-
-int i2d_X509_KEY(X509 *a, unsigned char **pp)
-	{
-	M_ASN1_I2D_vars(a);
-
-	M_ASN1_I2D_len(a->cert_info,	i2d_X509_CINF);
-	M_ASN1_I2D_len(a->sig_alg,	i2d_X509_ALGOR);
-	M_ASN1_I2D_len(a->signature,	i2d_ASN1_BIT_STRING);
-
-	M_ASN1_I2D_seq_total();
-
-	M_ASN1_I2D_put(a->cert_info,	i2d_X509_CINF);
-	M_ASN1_I2D_put(a->sig_alg,	i2d_X509_ALGOR);
-	M_ASN1_I2D_put(a->signature,	i2d_ASN1_BIT_STRING);
-
-	M_ASN1_I2D_finish();
-	}
-
-X509 *d2i_X509_KEY(X509 **a, unsigned char **pp, long length)
-	{
-	M_ASN1_D2I_vars(a,X509 *,X509_new);
-
-	M_ASN1_D2I_Init();
-	M_ASN1_D2I_start_sequence();
-	M_ASN1_D2I_get(ret->cert_info,d2i_X509_CINF);
-	M_ASN1_D2I_get(ret->sig_alg,d2i_X509_ALGOR);
-	M_ASN1_D2I_get(ret->signature,d2i_ASN1_BIT_STRING);
-	M_ASN1_D2I_Finish(a,X509_free,ASN1_F_D2I_X509);
-	}
-
-X509 *X509_KEY_new(void)
-	{
-	X509_KEY *ret=NULL;
-
-	M_ASN1_New_OPENSSL_malloc(ret,X509_KEY);
-	ret->references=1;
-	ret->type=NID
-	M_ASN1_New(ret->cert_info,X509_CINF_new);
-	M_ASN1_New(ret->sig_alg,X509_ALGOR_new);
-	M_ASN1_New(ret->signature,ASN1_BIT_STRING_new);
-	return(ret);
-	M_ASN1_New_Error(ASN1_F_X509_NEW);
-	}
-
-void X509_KEY_free(X509 *a)
-	{
-	int i;
-
-	if (a == NULL) return;
-
-	i=CRYPTO_add_lock(&a->references,-1,CRYPTO_LOCK_X509_KEY);
-#ifdef REF_PRINT
-	REF_PRINT("X509_KEY",a);
-#endif
-	if (i > 0) return;
-#ifdef REF_CHECK
-	if (i < 0)
-		{
-		fprintf(stderr,"X509_KEY_free, bad reference count\n");
-		abort();
-		}
-#endif
-
-	X509_CINF_free(a->cert_info);
-	X509_ALGOR_free(a->sig_alg);
-	ASN1_BIT_STRING_free(a->signature);
-	OPENSSL_free(a);
-	}
-
diff --git a/crypto/asn1/p8_pkey.c b/crypto/asn1/p8_pkey.c
index 0a1957556e54..17b68d386d85 100644
--- a/crypto/asn1/p8_pkey.c
+++ b/crypto/asn1/p8_pkey.c
@@ -3,7 +3,7 @@
  * project 1999.
  */
 /* ====================================================================
- * Copyright (c) 1999 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -62,7 +62,8 @@
 #include <openssl/x509.h>
 
 /* Minor tweak to operation: zero private key data */
-static int pkey_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int pkey_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+							void *exarg)
 {
 	/* Since the structure must still be valid use ASN1_OP_FREE_PRE */
 	if(operation == ASN1_OP_FREE_PRE) {
@@ -82,3 +83,73 @@ ASN1_SEQUENCE_cb(PKCS8_PRIV_KEY_INFO, pkey_cb) = {
 } ASN1_SEQUENCE_END_cb(PKCS8_PRIV_KEY_INFO, PKCS8_PRIV_KEY_INFO)
 
 IMPLEMENT_ASN1_FUNCTIONS(PKCS8_PRIV_KEY_INFO)
+
+int PKCS8_pkey_set0(PKCS8_PRIV_KEY_INFO *priv, ASN1_OBJECT *aobj,
+					int version,
+					int ptype, void *pval,
+					unsigned char *penc, int penclen)
+	{
+	unsigned char **ppenc = NULL;
+	if (version >= 0)
+		{
+		if (!ASN1_INTEGER_set(priv->version, version))
+			return 0;
+		}
+	if (penc)
+		{
+		int pmtype;
+		ASN1_OCTET_STRING *oct;
+		oct = ASN1_OCTET_STRING_new();
+		if (!oct)
+			return 0;
+		oct->data = penc;
+		ppenc = &oct->data;
+		oct->length = penclen;
+		if (priv->broken == PKCS8_NO_OCTET)
+			pmtype = V_ASN1_SEQUENCE;
+		else
+			pmtype = V_ASN1_OCTET_STRING;
+		ASN1_TYPE_set(priv->pkey, pmtype, oct);
+		}
+	if (!X509_ALGOR_set0(priv->pkeyalg, aobj, ptype, pval))
+		{
+		/* If call fails do not swallow 'enc' */
+		if (ppenc)
+			*ppenc = NULL;
+		return 0;
+		}
+	return 1;
+	}
+
+int PKCS8_pkey_get0(ASN1_OBJECT **ppkalg,
+		const unsigned char **pk, int *ppklen,
+		X509_ALGOR **pa,
+		PKCS8_PRIV_KEY_INFO *p8)
+	{
+	if (ppkalg)
+		*ppkalg = p8->pkeyalg->algorithm;
+	if(p8->pkey->type == V_ASN1_OCTET_STRING)
+		{
+		p8->broken = PKCS8_OK;
+		if (pk)
+			{
+			*pk = p8->pkey->value.octet_string->data;
+			*ppklen = p8->pkey->value.octet_string->length;
+			}
+		}
+	else if (p8->pkey->type == V_ASN1_SEQUENCE)
+		{
+		p8->broken = PKCS8_NO_OCTET;
+		if (pk)
+			{
+			*pk = p8->pkey->value.sequence->data;
+			*ppklen = p8->pkey->value.sequence->length;
+			}
+		}
+	else
+		return 0;
+	if (pa)
+		*pa = p8->pkeyalg;
+	return 1;
+	}
+
diff --git a/crypto/asn1/t_crl.c b/crypto/asn1/t_crl.c
index ee5a687ce83a..c61169208a34 100644
--- a/crypto/asn1/t_crl.c
+++ b/crypto/asn1/t_crl.c
@@ -94,8 +94,7 @@ int X509_CRL_print(BIO *out, X509_CRL *x)
 	l = X509_CRL_get_version(x);
 	BIO_printf(out, "%8sVersion %lu (0x%lx)\n", "", l+1, l);
 	i = OBJ_obj2nid(x->sig_alg->algorithm);
-	BIO_printf(out, "%8sSignature Algorithm: %s\n", "",
-				 (i == NID_undef) ? "NONE" : OBJ_nid2ln(i));
+	X509_signature_print(out, x->sig_alg, NULL);
 	p=X509_NAME_oneline(X509_CRL_get_issuer(x),NULL,0);
 	BIO_printf(out,"%8sIssuer: %s\n","",p);
 	OPENSSL_free(p);
diff --git a/crypto/asn1/t_pkey.c b/crypto/asn1/t_pkey.c
index afb95d67121a..9dd18f657937 100644
--- a/crypto/asn1/t_pkey.c
+++ b/crypto/asn1/t_pkey.c
@@ -55,520 +55,15 @@
  * copied and put under another distribution licence
  * [including the GNU Public Licence.]
  */
-/* ====================================================================
- * Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
- * Binary polynomial ECC support in OpenSSL originally developed by 
- * SUN MICROSYSTEMS, INC., and contributed to the OpenSSL project.
- */
 
 #include <stdio.h>
 #include "cryptlib.h"
 #include <openssl/objects.h>
 #include <openssl/buffer.h>
 #include <openssl/bn.h>
-#ifndef OPENSSL_NO_RSA
-#include <openssl/rsa.h>
-#endif
-#ifndef OPENSSL_NO_DH
-#include <openssl/dh.h>
-#endif
-#ifndef OPENSSL_NO_DSA
-#include <openssl/dsa.h>
-#endif
-#ifndef OPENSSL_NO_EC
-#include <openssl/ec.h>
-#endif
-
-static int print(BIO *fp,const char *str, const BIGNUM *num,
-		unsigned char *buf,int off);
-#ifndef OPENSSL_NO_EC
-static int print_bin(BIO *fp, const char *str, const unsigned char *num,
-		size_t len, int off);
-#endif
-#ifndef OPENSSL_NO_RSA
-#ifndef OPENSSL_NO_FP_API
-int RSA_print_fp(FILE *fp, const RSA *x, int off)
-	{
-	BIO *b;
-	int ret;
-
-	if ((b=BIO_new(BIO_s_file())) == NULL)
-		{
-		RSAerr(RSA_F_RSA_PRINT_FP,ERR_R_BUF_LIB);
-		return(0);
-		}
-	BIO_set_fp(b,fp,BIO_NOCLOSE);
-	ret=RSA_print(b,x,off);
-	BIO_free(b);
-	return(ret);
-	}
-#endif
-
-int RSA_print(BIO *bp, const RSA *x, int off)
-	{
-	char str[128];
-	const char *s;
-	unsigned char *m=NULL;
-	int ret=0, mod_len = 0;
-	size_t buf_len=0, i;
-
-	if (x->n)
-		buf_len = (size_t)BN_num_bytes(x->n);
-	if (x->e)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->e)))
-			buf_len = i;
-	if (x->d)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->d)))
-			buf_len = i;
-	if (x->p)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->p)))
-			buf_len = i;
-	if (x->q)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->q)))
-			buf_len = i;
-	if (x->dmp1)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->dmp1)))
-			buf_len = i;
-	if (x->dmq1)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->dmq1)))
-			buf_len = i;
-	if (x->iqmp)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->iqmp)))
-			buf_len = i;
-
-	m=(unsigned char *)OPENSSL_malloc(buf_len+10);
-	if (m == NULL)
-		{
-		RSAerr(RSA_F_RSA_PRINT,ERR_R_MALLOC_FAILURE);
-		goto err;
-		}
-
-	if (x->n != NULL)
-		mod_len = BN_num_bits(x->n);
-
-	if (x->d != NULL)
-		{
-		if(!BIO_indent(bp,off,128))
-		   goto err;
-		if (BIO_printf(bp,"Private-Key: (%d bit)\n", mod_len)
-			<= 0) goto err;
-		}
-
-	if (x->d == NULL)
-		BIO_snprintf(str,sizeof str,"Modulus (%d bit):", mod_len);
-	else
-		BUF_strlcpy(str,"modulus:",sizeof str);
-	if (!print(bp,str,x->n,m,off)) goto err;
-	s=(x->d == NULL)?"Exponent:":"publicExponent:";
-	if ((x->e != NULL) && !print(bp,s,x->e,m,off))
-		goto err;
-	if ((x->d != NULL) && !print(bp,"privateExponent:",x->d,m,off))
-		goto err;
-	if ((x->p != NULL) && !print(bp,"prime1:",x->p,m,off))
-		goto err;
-	if ((x->q != NULL) && !print(bp,"prime2:",x->q,m,off))
-		goto err;
-	if ((x->dmp1 != NULL) && !print(bp,"exponent1:",x->dmp1,m,off))
-		goto err;
-	if ((x->dmq1 != NULL) && !print(bp,"exponent2:",x->dmq1,m,off))
-		goto err;
-	if ((x->iqmp != NULL) && !print(bp,"coefficient:",x->iqmp,m,off))
-		goto err;
-	ret=1;
-err:
-	if (m != NULL) OPENSSL_free(m);
-	return(ret);
-	}
-#endif /* OPENSSL_NO_RSA */
-
-#ifndef OPENSSL_NO_DSA
-#ifndef OPENSSL_NO_FP_API
-int DSA_print_fp(FILE *fp, const DSA *x, int off)
-	{
-	BIO *b;
-	int ret;
-
-	if ((b=BIO_new(BIO_s_file())) == NULL)
-		{
-		DSAerr(DSA_F_DSA_PRINT_FP,ERR_R_BUF_LIB);
-		return(0);
-		}
-	BIO_set_fp(b,fp,BIO_NOCLOSE);
-	ret=DSA_print(b,x,off);
-	BIO_free(b);
-	return(ret);
-	}
-#endif
-
-int DSA_print(BIO *bp, const DSA *x, int off)
-	{
-	unsigned char *m=NULL;
-	int ret=0;
-	size_t buf_len=0,i;
-
-	if (x->p)
-		buf_len = (size_t)BN_num_bytes(x->p);
-	else
-		{
-		DSAerr(DSA_F_DSA_PRINT,DSA_R_MISSING_PARAMETERS);
-		goto err;
-		}
-	if (x->q)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->q)))
-			buf_len = i;
-	if (x->g)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->g)))
-			buf_len = i;
-	if (x->priv_key)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->priv_key)))
-			buf_len = i;
-	if (x->pub_key)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->pub_key)))
-			buf_len = i;
-
-	m=(unsigned char *)OPENSSL_malloc(buf_len+10);
-	if (m == NULL)
-		{
-		DSAerr(DSA_F_DSA_PRINT,ERR_R_MALLOC_FAILURE);
-		goto err;
-		}
-
-	if (x->priv_key != NULL)
-		{
-		if(!BIO_indent(bp,off,128))
-		   goto err;
-		if (BIO_printf(bp,"Private-Key: (%d bit)\n",BN_num_bits(x->p))
-			<= 0) goto err;
-		}
-
-	if ((x->priv_key != NULL) && !print(bp,"priv:",x->priv_key,m,off))
-		goto err;
-	if ((x->pub_key  != NULL) && !print(bp,"pub: ",x->pub_key,m,off))
-		goto err;
-	if ((x->p != NULL) && !print(bp,"P:   ",x->p,m,off)) goto err;
-	if ((x->q != NULL) && !print(bp,"Q:   ",x->q,m,off)) goto err;
-	if ((x->g != NULL) && !print(bp,"G:   ",x->g,m,off)) goto err;
-	ret=1;
-err:
-	if (m != NULL) OPENSSL_free(m);
-	return(ret);
-	}
-#endif /* !OPENSSL_NO_DSA */
-
-#ifndef OPENSSL_NO_EC
-#ifndef OPENSSL_NO_FP_API
-int ECPKParameters_print_fp(FILE *fp, const EC_GROUP *x, int off)
-	{
-	BIO *b;
-	int ret;
-
-	if ((b=BIO_new(BIO_s_file())) == NULL)
-		{
-		ECerr(EC_F_ECPKPARAMETERS_PRINT_FP,ERR_R_BUF_LIB);
-		return(0);
-		}
-	BIO_set_fp(b, fp, BIO_NOCLOSE);
-	ret = ECPKParameters_print(b, x, off);
-	BIO_free(b);
-	return(ret);
-	}
-
-int EC_KEY_print_fp(FILE *fp, const EC_KEY *x, int off)
-	{
-	BIO *b;
-	int ret;
- 
-	if ((b=BIO_new(BIO_s_file())) == NULL)
-		{
-		ECerr(EC_F_EC_KEY_PRINT_FP, ERR_R_BIO_LIB);
-		return(0);
-		}
-	BIO_set_fp(b, fp, BIO_NOCLOSE);
-	ret = EC_KEY_print(b, x, off);
-	BIO_free(b);
-	return(ret);
-	}
-#endif
-
-int ECPKParameters_print(BIO *bp, const EC_GROUP *x, int off)
-	{
-	unsigned char *buffer=NULL;
-	size_t	buf_len=0, i;
-	int     ret=0, reason=ERR_R_BIO_LIB;
-	BN_CTX  *ctx=NULL;
-	const EC_POINT *point=NULL;
-	BIGNUM	*p=NULL, *a=NULL, *b=NULL, *gen=NULL,
-		*order=NULL, *cofactor=NULL;
-	const unsigned char *seed;
-	size_t	seed_len=0;
-	
-	static const char *gen_compressed = "Generator (compressed):";
-	static const char *gen_uncompressed = "Generator (uncompressed):";
-	static const char *gen_hybrid = "Generator (hybrid):";
- 
-	if (!x)
-		{
-		reason = ERR_R_PASSED_NULL_PARAMETER;
-		goto err;
-		}
-
-	if (EC_GROUP_get_asn1_flag(x))
-		{
-		/* the curve parameter are given by an asn1 OID */
-		int nid;
-
-		if (!BIO_indent(bp, off, 128))
-			goto err;
-
-		nid = EC_GROUP_get_curve_name(x);
-		if (nid == 0)
-			goto err;
-
-		if (BIO_printf(bp, "ASN1 OID: %s", OBJ_nid2sn(nid)) <= 0)
-			goto err;
-		if (BIO_printf(bp, "\n") <= 0)
-			goto err;
-		}
-	else
-		{
-		/* explicit parameters */
-		int is_char_two = 0;
-		point_conversion_form_t form;
-		int tmp_nid = EC_METHOD_get_field_type(EC_GROUP_method_of(x));
-
-		if (tmp_nid == NID_X9_62_characteristic_two_field)
-			is_char_two = 1;
-
-		if ((p = BN_new()) == NULL || (a = BN_new()) == NULL ||
-			(b = BN_new()) == NULL || (order = BN_new()) == NULL ||
-			(cofactor = BN_new()) == NULL)
-			{
-			reason = ERR_R_MALLOC_FAILURE;
-			goto err;
-			}
-
-		if (is_char_two)
-			{
-			if (!EC_GROUP_get_curve_GF2m(x, p, a, b, ctx))
-				{
-				reason = ERR_R_EC_LIB;
-				goto err;
-				}
-			}
-		else /* prime field */
-			{
-			if (!EC_GROUP_get_curve_GFp(x, p, a, b, ctx))
-				{
-				reason = ERR_R_EC_LIB;
-				goto err;
-				}
-			}
-
-		if ((point = EC_GROUP_get0_generator(x)) == NULL)
-			{
-			reason = ERR_R_EC_LIB;
-			goto err;
-			}
-		if (!EC_GROUP_get_order(x, order, NULL) || 
-            		!EC_GROUP_get_cofactor(x, cofactor, NULL))
-			{
-			reason = ERR_R_EC_LIB;
-			goto err;
-			}
-		
-		form = EC_GROUP_get_point_conversion_form(x);
-
-		if ((gen = EC_POINT_point2bn(x, point, 
-				form, NULL, ctx)) == NULL)
-			{
-			reason = ERR_R_EC_LIB;
-			goto err;
-			}
-
-		buf_len = (size_t)BN_num_bytes(p);
-		if (buf_len < (i = (size_t)BN_num_bytes(a)))
-			buf_len = i;
-		if (buf_len < (i = (size_t)BN_num_bytes(b)))
-			buf_len = i;
-		if (buf_len < (i = (size_t)BN_num_bytes(gen)))
-			buf_len = i;
-		if (buf_len < (i = (size_t)BN_num_bytes(order)))
-			buf_len = i;
-		if (buf_len < (i = (size_t)BN_num_bytes(cofactor))) 
-			buf_len = i;
-
-		if ((seed = EC_GROUP_get0_seed(x)) != NULL)
-			seed_len = EC_GROUP_get_seed_len(x);
-
-		buf_len += 10;
-		if ((buffer = OPENSSL_malloc(buf_len)) == NULL)
-			{
-			reason = ERR_R_MALLOC_FAILURE;
-			goto err;
-			}
-
-		if (!BIO_indent(bp, off, 128))
-			goto err;
-
-		/* print the 'short name' of the field type */
-		if (BIO_printf(bp, "Field Type: %s\n", OBJ_nid2sn(tmp_nid))
-			<= 0)
-			goto err;  
-
-		if (is_char_two)
-			{
-			/* print the 'short name' of the base type OID */
-			int basis_type = EC_GROUP_get_basis_type(x);
-			if (basis_type == 0)
-				goto err;
-
-			if (!BIO_indent(bp, off, 128))
-				goto err;
-
-			if (BIO_printf(bp, "Basis Type: %s\n", 
-				OBJ_nid2sn(basis_type)) <= 0)
-				goto err;
-
-			/* print the polynomial */
-			if ((p != NULL) && !print(bp, "Polynomial:", p, buffer,
-				off))
-				goto err;
-			}
-		else
-			{
-			if ((p != NULL) && !print(bp, "Prime:", p, buffer,off))
-				goto err;
-			}
-		if ((a != NULL) && !print(bp, "A:   ", a, buffer, off)) 
-			goto err;
-		if ((b != NULL) && !print(bp, "B:   ", b, buffer, off))
-			goto err;
-		if (form == POINT_CONVERSION_COMPRESSED)
-			{
-			if ((gen != NULL) && !print(bp, gen_compressed, gen,
-				buffer, off))
-				goto err;
-			}
-		else if (form == POINT_CONVERSION_UNCOMPRESSED)
-			{
-			if ((gen != NULL) && !print(bp, gen_uncompressed, gen,
-				buffer, off))
-				goto err;
-			}
-		else /* form == POINT_CONVERSION_HYBRID */
-			{
-			if ((gen != NULL) && !print(bp, gen_hybrid, gen,
-				buffer, off))
-				goto err;
-			}
-		if ((order != NULL) && !print(bp, "Order: ", order, 
-			buffer, off)) goto err;
-		if ((cofactor != NULL) && !print(bp, "Cofactor: ", cofactor, 
-			buffer, off)) goto err;
-		if (seed && !print_bin(bp, "Seed:", seed, seed_len, off))
-			goto err;
-		}
-	ret=1;
-err:
-	if (!ret)
- 		ECerr(EC_F_ECPKPARAMETERS_PRINT, reason);
-	if (p) 
-		BN_free(p);
-	if (a) 
-		BN_free(a);
-	if (b)
-		BN_free(b);
-	if (gen)
-		BN_free(gen);
-	if (order)
-		BN_free(order);
-	if (cofactor)
-		BN_free(cofactor);
-	if (ctx)
-		BN_CTX_free(ctx);
-	if (buffer != NULL) 
-		OPENSSL_free(buffer);
-	return(ret);	
-	}
 
-int EC_KEY_print(BIO *bp, const EC_KEY *x, int off)
-	{
-	unsigned char *buffer=NULL;
-	size_t	buf_len=0, i;
-	int     ret=0, reason=ERR_R_BIO_LIB;
-	BIGNUM  *pub_key=NULL, *order=NULL;
-	BN_CTX  *ctx=NULL;
-	const EC_GROUP *group;
-	const EC_POINT *public_key;
-	const BIGNUM *priv_key;
- 
-	if (x == NULL || (group = EC_KEY_get0_group(x)) == NULL)
-		{
-		reason = ERR_R_PASSED_NULL_PARAMETER;
-		goto err;
-		}
-
-	public_key = EC_KEY_get0_public_key(x);
-	if ((pub_key = EC_POINT_point2bn(group, public_key,
-		EC_KEY_get_conv_form(x), NULL, ctx)) == NULL)
-		{
-		reason = ERR_R_EC_LIB;
-		goto err;
-		}
-
-	buf_len = (size_t)BN_num_bytes(pub_key);
-	priv_key = EC_KEY_get0_private_key(x);
-	if (priv_key != NULL)
-		{
-		if ((i = (size_t)BN_num_bytes(priv_key)) > buf_len)
-			buf_len = i;
-		}
-
-	buf_len += 10;
-	if ((buffer = OPENSSL_malloc(buf_len)) == NULL)
-		{
-		reason = ERR_R_MALLOC_FAILURE;
-		goto err;
-		}
-
-	if (priv_key != NULL)
-		{
-		if (!BIO_indent(bp, off, 128))
-			goto err;
-		if ((order = BN_new()) == NULL)
-			goto err;
-		if (!EC_GROUP_get_order(group, order, NULL))
-			goto err;
-		if (BIO_printf(bp, "Private-Key: (%d bit)\n", 
-			BN_num_bits(order)) <= 0) goto err;
-		}
-  
-	if ((priv_key != NULL) && !print(bp, "priv:", priv_key, 
-		buffer, off))
-		goto err;
-	if ((pub_key != NULL) && !print(bp, "pub: ", pub_key,
-		buffer, off))
-		goto err;
-	if (!ECPKParameters_print(bp, group, off))
-		goto err;
-	ret=1;
-err:
-	if (!ret)
- 		ECerr(EC_F_EC_KEY_PRINT, reason);
-	if (pub_key) 
-		BN_free(pub_key);
-	if (order)
-		BN_free(order);
-	if (ctx)
-		BN_CTX_free(ctx);
-	if (buffer != NULL)
-		OPENSSL_free(buffer);
-	return(ret);
-	}
-#endif /* OPENSSL_NO_EC */
-
-static int print(BIO *bp, const char *number, const BIGNUM *num, unsigned char *buf,
-	     int off)
+int ASN1_bn_print(BIO *bp, const char *number, const BIGNUM *num,
+			unsigned char *buf, int off)
 	{
 	int n,i;
 	const char *neg;
@@ -617,223 +112,3 @@ static int print(BIO *bp, const char *number, const BIGNUM *num, unsigned char *
 		}
 	return(1);
 	}
-
-#ifndef OPENSSL_NO_EC
-static int print_bin(BIO *fp, const char *name, const unsigned char *buf,
-		size_t len, int off)
-	{
-	size_t i;
-	char str[128];
-
-	if (buf == NULL)
-		return 1;
-	if (off)
-		{
-		if (off > 128)
-			off=128;
-		memset(str,' ',off);
-		if (BIO_write(fp, str, off) <= 0)
-			return 0;
-		}
-
-	if (BIO_printf(fp,"%s", name) <= 0)
-		return 0;
-
-	for (i=0; i<len; i++)
-		{
-		if ((i%15) == 0)
-			{
-			str[0]='\n';
-			memset(&(str[1]),' ',off+4);
-			if (BIO_write(fp, str, off+1+4) <= 0)
-				return 0;
-			}
-		if (BIO_printf(fp,"%02x%s",buf[i],((i+1) == len)?"":":") <= 0)
-			return 0;
-		}
-	if (BIO_write(fp,"\n",1) <= 0)
-		return 0;
-
-	return 1;
-	}
-#endif
-
-#ifndef OPENSSL_NO_DH
-#ifndef OPENSSL_NO_FP_API
-int DHparams_print_fp(FILE *fp, const DH *x)
-	{
-	BIO *b;
-	int ret;
-
-	if ((b=BIO_new(BIO_s_file())) == NULL)
-		{
-		DHerr(DH_F_DHPARAMS_PRINT_FP,ERR_R_BUF_LIB);
-		return(0);
-		}
-	BIO_set_fp(b,fp,BIO_NOCLOSE);
-	ret=DHparams_print(b, x);
-	BIO_free(b);
-	return(ret);
-	}
-#endif
-
-int DHparams_print(BIO *bp, const DH *x)
-	{
-	unsigned char *m=NULL;
-	int reason=ERR_R_BUF_LIB,ret=0;
-	size_t buf_len=0, i;
-
-	if (x->p)
-		buf_len = (size_t)BN_num_bytes(x->p);
-	else
-		{
-		reason = ERR_R_PASSED_NULL_PARAMETER;
-		goto err;
-		}
-	if (x->g)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->g)))
-			buf_len = i;
-	m=(unsigned char *)OPENSSL_malloc(buf_len+10);
-	if (m == NULL)
-		{
-		reason=ERR_R_MALLOC_FAILURE;
-		goto err;
-		}
-
-	if (BIO_printf(bp,"Diffie-Hellman-Parameters: (%d bit)\n",
-		BN_num_bits(x->p)) <= 0)
-		goto err;
-	if (!print(bp,"prime:",x->p,m,4)) goto err;
-	if (!print(bp,"generator:",x->g,m,4)) goto err;
-	if (x->length != 0)
-		{
-		if (BIO_printf(bp,"    recommended-private-length: %d bits\n",
-			(int)x->length) <= 0) goto err;
-		}
-	ret=1;
-	if (0)
-		{
-err:
-		DHerr(DH_F_DHPARAMS_PRINT,reason);
-		}
-	if (m != NULL) OPENSSL_free(m);
-	return(ret);
-	}
-#endif
-
-#ifndef OPENSSL_NO_DSA
-#ifndef OPENSSL_NO_FP_API
-int DSAparams_print_fp(FILE *fp, const DSA *x)
-	{
-	BIO *b;
-	int ret;
-
-	if ((b=BIO_new(BIO_s_file())) == NULL)
-		{
-		DSAerr(DSA_F_DSAPARAMS_PRINT_FP,ERR_R_BUF_LIB);
-		return(0);
-		}
-	BIO_set_fp(b,fp,BIO_NOCLOSE);
-	ret=DSAparams_print(b, x);
-	BIO_free(b);
-	return(ret);
-	}
-#endif
-
-int DSAparams_print(BIO *bp, const DSA *x)
-	{
-	unsigned char *m=NULL;
-	int ret=0;
-	size_t buf_len=0,i;
-
-	if (x->p)
-		buf_len = (size_t)BN_num_bytes(x->p);
-	else
-		{
-		DSAerr(DSA_F_DSAPARAMS_PRINT,DSA_R_MISSING_PARAMETERS);
-		goto err;
-		}
-	if (x->q)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->q)))
-			buf_len = i;
-	if (x->g)
-		if (buf_len < (i = (size_t)BN_num_bytes(x->g)))
-			buf_len = i;
-	m=(unsigned char *)OPENSSL_malloc(buf_len+10);
-	if (m == NULL)
-		{
-		DSAerr(DSA_F_DSAPARAMS_PRINT,ERR_R_MALLOC_FAILURE);
-		goto err;
-		}
-
-	if (BIO_printf(bp,"DSA-Parameters: (%d bit)\n",
-		BN_num_bits(x->p)) <= 0)
-		goto err;
-	if (!print(bp,"p:",x->p,m,4)) goto err;
-	if ((x->q != NULL) && !print(bp,"q:",x->q,m,4)) goto err;
-	if ((x->g != NULL) && !print(bp,"g:",x->g,m,4)) goto err;
-	ret=1;
-err:
-	if (m != NULL) OPENSSL_free(m);
-	return(ret);
-	}
-
-#endif /* !OPENSSL_NO_DSA */
-
-#ifndef OPENSSL_NO_EC
-#ifndef OPENSSL_NO_FP_API
-int ECParameters_print_fp(FILE *fp, const EC_KEY *x)
-	{
-	BIO *b;
-	int ret;
- 
-	if ((b=BIO_new(BIO_s_file())) == NULL)
-		{
-		ECerr(EC_F_ECPARAMETERS_PRINT_FP, ERR_R_BIO_LIB);
-		return(0);
-		}
-	BIO_set_fp(b, fp, BIO_NOCLOSE);
-	ret = ECParameters_print(b, x);
-	BIO_free(b);
-	return(ret);
-	}
-#endif
-
-int ECParameters_print(BIO *bp, const EC_KEY *x)
-	{
-	int     reason=ERR_R_EC_LIB, ret=0;
-	BIGNUM	*order=NULL;
-	const EC_GROUP *group;
- 
-	if (x == NULL || (group = EC_KEY_get0_group(x)) == NULL)
-		{
-		reason = ERR_R_PASSED_NULL_PARAMETER;;
-		goto err;
-		}
-
-	if ((order = BN_new()) == NULL)
-		{
-		reason = ERR_R_MALLOC_FAILURE;
-		goto err;
-		}
-
-	if (!EC_GROUP_get_order(group, order, NULL))
-		{
-		reason = ERR_R_EC_LIB;
-		goto err;
-		}
- 
-	if (BIO_printf(bp, "ECDSA-Parameters: (%d bit)\n", 
-		BN_num_bits(order)) <= 0)
-		goto err;
-	if (!ECPKParameters_print(bp, group, 4))
-		goto err;
-	ret=1;
-err:
-	if (order)
-		BN_free(order);
-	ECerr(EC_F_ECPARAMETERS_PRINT, reason);
-	return(ret);
-	}
-  
-#endif
diff --git a/crypto/asn1/t_req.c b/crypto/asn1/t_req.c
index 5557e0658448..ea1794e3e082 100644
--- a/crypto/asn1/t_req.c
+++ b/crypto/asn1/t_req.c
@@ -149,34 +149,10 @@ int X509_REQ_print_ex(BIO *bp, X509_REQ *x, unsigned long nmflags, unsigned long
 			ERR_print_errors(bp);
 			}
 		else
-#ifndef OPENSSL_NO_RSA
-		if (pkey->type == EVP_PKEY_RSA)
-			{
-			BIO_printf(bp,"%12sRSA Public Key: (%d bit)\n","",
-			BN_num_bits(pkey->pkey.rsa->n));
-			RSA_print(bp,pkey->pkey.rsa,16);
-			}
-		else
-#endif
-#ifndef OPENSSL_NO_DSA
-		if (pkey->type == EVP_PKEY_DSA)
 			{
-			BIO_printf(bp,"%12sDSA Public Key:\n","");
-			DSA_print(bp,pkey->pkey.dsa,16);
+			EVP_PKEY_print_public(bp, pkey, 16, NULL);
+			EVP_PKEY_free(pkey);
 			}
-		else
-#endif
-#ifndef OPENSSL_NO_EC
-		if (pkey->type == EVP_PKEY_EC)
-		{
-			BIO_printf(bp, "%12sEC Public Key: \n","");
-			EC_KEY_print(bp, pkey->pkey.ec, 16);
-		}
-	else
-#endif
-			BIO_printf(bp,"%12sUnknown Public Key:\n","");
-
-		EVP_PKEY_free(pkey);
 		}
 
 	if(!(cflag & X509_FLAG_NO_ATTRIBUTES))
diff --git a/crypto/asn1/t_spki.c b/crypto/asn1/t_spki.c
index a73369b94920..079c081a81c7 100644
--- a/crypto/asn1/t_spki.c
+++ b/crypto/asn1/t_spki.c
@@ -82,36 +82,11 @@ int NETSCAPE_SPKI_print(BIO *out, NETSCAPE_SPKI *spki)
 				(i == NID_undef)?"UNKNOWN":OBJ_nid2ln(i));
 	pkey = X509_PUBKEY_get(spki->spkac->pubkey);
 	if(!pkey) BIO_printf(out, "  Unable to load public key\n");
-	else {
-#ifndef OPENSSL_NO_RSA
-		if (pkey->type == EVP_PKEY_RSA)
-			{
-			BIO_printf(out,"  RSA Public Key: (%d bit)\n",
-				BN_num_bits(pkey->pkey.rsa->n));
-			RSA_print(out,pkey->pkey.rsa,2);
-			}
-		else 
-#endif
-#ifndef OPENSSL_NO_DSA
-		if (pkey->type == EVP_PKEY_DSA)
-		{
-		BIO_printf(out,"  DSA Public Key:\n");
-		DSA_print(out,pkey->pkey.dsa,2);
-		}
-		else
-#endif
-#ifndef OPENSSL_NO_EC
-		if (pkey->type == EVP_PKEY_EC)
+	else
 		{
-			BIO_printf(out, "  EC Public Key:\n");
-			EC_KEY_print(out, pkey->pkey.ec,2);
-		}
-		else
-#endif
-
-			BIO_printf(out,"  Unknown Public Key:\n");
+		EVP_PKEY_print_public(out, pkey, 4, NULL);
 		EVP_PKEY_free(pkey);
-	}
+		}
 	chal = spki->spkac->challenge;
 	if(chal->length)
 		BIO_printf(out, "  Challenge String: %s\n", chal->data);
diff --git a/crypto/asn1/t_x509.c b/crypto/asn1/t_x509.c
index 6f295b4e145c..edbb39a02f7a 100644
--- a/crypto/asn1/t_x509.c
+++ b/crypto/asn1/t_x509.c
@@ -72,6 +72,7 @@
 #include <openssl/objects.h>
 #include <openssl/x509.h>
 #include <openssl/x509v3.h>
+#include "asn1_locl.h"
 
 #ifndef OPENSSL_NO_FP_API
 int X509_print_fp(FILE *fp, X509 *x)
@@ -111,7 +112,6 @@ int X509_print_ex(BIO *bp, X509 *x, unsigned long nmflags, unsigned long cflag)
 	ASN1_INTEGER *bs;
 	EVP_PKEY *pkey=NULL;
 	const char *neg;
-	ASN1_STRING *str=NULL;
 
 	if((nmflags & XN_FLAG_SEP_MASK) == XN_FLAG_SEP_MULTILINE) {
 			mlch = '\n';
@@ -138,10 +138,10 @@ int X509_print_ex(BIO *bp, X509 *x, unsigned long nmflags, unsigned long cflag)
 		if (BIO_write(bp,"        Serial Number:",22) <= 0) goto err;
 
 		bs=X509_get_serialNumber(x);
-		if (bs->length <= 4)
+		if (bs->length <= (int)sizeof(long))
 			{
 			l=ASN1_INTEGER_get(bs);
-			if (l < 0)
+			if (bs->type == V_ASN1_NEG_INTEGER)
 				{
 				l= -l;
 				neg="-";
@@ -168,12 +168,16 @@ int X509_print_ex(BIO *bp, X509 *x, unsigned long nmflags, unsigned long cflag)
 
 	if(!(cflag & X509_FLAG_NO_SIGNAME))
 		{
+		if(X509_signature_print(bp, x->sig_alg, NULL) <= 0)
+			goto err;
+#if 0
 		if (BIO_printf(bp,"%8sSignature Algorithm: ","") <= 0) 
 			goto err;
 		if (i2a_ASN1_OBJECT(bp, ci->signature->algorithm) <= 0)
 			goto err;
 		if (BIO_puts(bp, "\n") <= 0)
 			goto err;
+#endif
 		}
 
 	if(!(cflag & X509_FLAG_NO_ISSUER))
@@ -215,34 +219,10 @@ int X509_print_ex(BIO *bp, X509 *x, unsigned long nmflags, unsigned long cflag)
 			ERR_print_errors(bp);
 			}
 		else
-#ifndef OPENSSL_NO_RSA
-		if (pkey->type == EVP_PKEY_RSA)
-			{
-			BIO_printf(bp,"%12sRSA Public Key: (%d bit)\n","",
-			BN_num_bits(pkey->pkey.rsa->n));
-			RSA_print(bp,pkey->pkey.rsa,16);
-			}
-		else
-#endif
-#ifndef OPENSSL_NO_DSA
-		if (pkey->type == EVP_PKEY_DSA)
 			{
-			BIO_printf(bp,"%12sDSA Public Key:\n","");
-			DSA_print(bp,pkey->pkey.dsa,16);
+			EVP_PKEY_print_public(bp, pkey, 16, NULL);
+			EVP_PKEY_free(pkey);
 			}
-		else
-#endif
-#ifndef OPENSSL_NO_EC
-		if (pkey->type == EVP_PKEY_EC)
-			{
-			BIO_printf(bp, "%12sEC Public Key:\n","");
-			EC_KEY_print(bp, pkey->pkey.ec, 16);
-			}
-		else
-#endif
-			BIO_printf(bp,"%12sUnknown Public Key:\n","");
-
-		EVP_PKEY_free(pkey);
 		}
 
 	if (!(cflag & X509_FLAG_NO_EXTENSIONS))
@@ -259,7 +239,6 @@ int X509_print_ex(BIO *bp, X509 *x, unsigned long nmflags, unsigned long cflag)
 		}
 	ret=1;
 err:
-	if (str != NULL) ASN1_STRING_free(str);
 	if (m != NULL) OPENSSL_free(m);
 	return(ret);
 	}
@@ -281,7 +260,8 @@ int X509_ocspid_print (BIO *bp, X509 *x)
 		goto err;
 	i2d_X509_NAME(x->cert_info->subject, &dertmp);
 
-	EVP_Digest(der, derlen, SHA1md, NULL, EVP_sha1(), NULL);
+	if (!EVP_Digest(der, derlen, SHA1md, NULL, EVP_sha1(), NULL))
+		goto err;
 	for (i=0; i < SHA_DIGEST_LENGTH; i++)
 		{
 		if (BIO_printf(bp,"%02X",SHA1md[i]) <= 0) goto err;
@@ -294,8 +274,10 @@ int X509_ocspid_print (BIO *bp, X509 *x)
 	if (BIO_printf(bp,"\n        Public key OCSP hash: ") <= 0)
 		goto err;
 
-	EVP_Digest(x->cert_info->key->public_key->data,
-		x->cert_info->key->public_key->length, SHA1md, NULL, EVP_sha1(), NULL);
+	if (!EVP_Digest(x->cert_info->key->public_key->data,
+			x->cert_info->key->public_key->length,
+			SHA1md, NULL, EVP_sha1(), NULL))
+		goto err;
 	for (i=0; i < SHA_DIGEST_LENGTH; i++)
 		{
 		if (BIO_printf(bp,"%02X",SHA1md[i]) <= 0)
@@ -309,34 +291,62 @@ err:
 	return(0);
 	}
 
-int X509_signature_print(BIO *bp, X509_ALGOR *sigalg, ASN1_STRING *sig)
+int X509_signature_dump(BIO *bp, const ASN1_STRING *sig, int indent)
 {
-	unsigned char *s;
+	const unsigned char *s;
 	int i, n;
-	if (BIO_puts(bp,"    Signature Algorithm: ") <= 0) return 0;
-	if (i2a_ASN1_OBJECT(bp, sigalg->algorithm) <= 0) return 0;
 
 	n=sig->length;
 	s=sig->data;
 	for (i=0; i<n; i++)
 		{
 		if ((i%18) == 0)
-			if (BIO_write(bp,"\n        ",9) <= 0) return 0;
+			{
+			if (BIO_write(bp,"\n",1) <= 0) return 0;
+			if (BIO_indent(bp, indent, indent) <= 0) return 0;
+			}
 			if (BIO_printf(bp,"%02x%s",s[i],
 				((i+1) == n)?"":":") <= 0) return 0;
 		}
 	if (BIO_write(bp,"\n",1) != 1) return 0;
+
 	return 1;
 }
 
-int ASN1_STRING_print(BIO *bp, ASN1_STRING *v)
+int X509_signature_print(BIO *bp, X509_ALGOR *sigalg, ASN1_STRING *sig)
+{
+	int sig_nid;
+	if (BIO_puts(bp,"    Signature Algorithm: ") <= 0) return 0;
+	if (i2a_ASN1_OBJECT(bp, sigalg->algorithm) <= 0) return 0;
+
+	sig_nid = OBJ_obj2nid(sigalg->algorithm);
+	if (sig_nid != NID_undef)
+		{
+		int pkey_nid, dig_nid;
+		const EVP_PKEY_ASN1_METHOD *ameth;
+		if (OBJ_find_sigid_algs(sig_nid, &dig_nid, &pkey_nid))
+			{
+			ameth = EVP_PKEY_asn1_find(NULL, pkey_nid);
+			if (ameth && ameth->sig_print)
+				return ameth->sig_print(bp, sigalg, sig, 9, 0);
+			}
+		}
+	if (sig)
+		return X509_signature_dump(bp, sig, 9);
+	else if (BIO_puts(bp, "\n") <= 0)
+		return 0;
+	return 1;
+}
+
+int ASN1_STRING_print(BIO *bp, const ASN1_STRING *v)
 	{
 	int i,n;
-	char buf[80],*p;
+	char buf[80];
+	const char *p;
 
 	if (v == NULL) return(0);
 	n=0;
-	p=(char *)v->data;
+	p=(const char *)v->data;
 	for (i=0; i<v->length; i++)
 		{
 		if ((p[i] > '~') || ((p[i] < ' ') &&
@@ -358,7 +368,7 @@ int ASN1_STRING_print(BIO *bp, ASN1_STRING *v)
 	return(1);
 	}
 
-int ASN1_TIME_print(BIO *bp, ASN1_TIME *tm)
+int ASN1_TIME_print(BIO *bp, const ASN1_TIME *tm)
 {
 	if(tm->type == V_ASN1_UTCTIME) return ASN1_UTCTIME_print(bp, tm);
 	if(tm->type == V_ASN1_GENERALIZEDTIME)
@@ -373,7 +383,7 @@ static const char *mon[12]=
     "Jul","Aug","Sep","Oct","Nov","Dec"
     };
 
-int ASN1_GENERALIZEDTIME_print(BIO *bp, ASN1_GENERALIZEDTIME *tm)
+int ASN1_GENERALIZEDTIME_print(BIO *bp, const ASN1_GENERALIZEDTIME *tm)
 	{
 	char *v;
 	int gmt=0;
@@ -421,15 +431,15 @@ err:
 	return(0);
 	}
 
-int ASN1_UTCTIME_print(BIO *bp, ASN1_UTCTIME *tm)
+int ASN1_UTCTIME_print(BIO *bp, const ASN1_UTCTIME *tm)
 	{
-	char *v;
+	const char *v;
 	int gmt=0;
 	int i;
 	int y=0,M=0,d=0,h=0,m=0,s=0;
 
 	i=tm->length;
-	v=(char *)tm->data;
+	v=(const char *)tm->data;
 
 	if (i < 10) goto err;
 	if (v[i-1] == 'Z') gmt=1;
diff --git a/crypto/asn1/tasn_dec.c b/crypto/asn1/tasn_dec.c
index a228c0d6cd16..87d7dfdf5c37 100644
--- a/crypto/asn1/tasn_dec.c
+++ b/crypto/asn1/tasn_dec.c
@@ -114,6 +114,8 @@ unsigned long ASN1_tag2bit(int tag)
 /* Macro to initialize and invalidate the cache */
 
 #define asn1_tlc_clear(c)	if (c) (c)->valid = 0
+/* Version to avoid compiler warning about 'c' always non-NULL */
+#define asn1_tlc_clear_nc(c)	(c)->valid = 0
 
 /* Decode an ASN1 item, this currently behaves just 
  * like a standard 'd2i' function. 'in' points to 
@@ -130,7 +132,7 @@ ASN1_VALUE *ASN1_item_d2i(ASN1_VALUE **pval,
 	ASN1_VALUE *ptmpval = NULL;
 	if (!pval)
 		pval = &ptmpval;
-	c.valid = 0;
+	asn1_tlc_clear_nc(&c);
 	if (ASN1_item_ex_d2i(pval, in, len, it, -1, 0, 0, &c) > 0) 
 		return *pval;
 	return NULL;
@@ -140,7 +142,7 @@ int ASN1_template_d2i(ASN1_VALUE **pval,
 		const unsigned char **in, long len, const ASN1_TEMPLATE *tt)
 	{
 	ASN1_TLC c;
-	c.valid = 0;
+	asn1_tlc_clear_nc(&c);
 	return asn1_template_ex_d2i(pval, in, len, tt, 0, &c);
 	}
 
@@ -306,7 +308,7 @@ int ASN1_item_ex_d2i(ASN1_VALUE **pval, const unsigned char **in, long len,
 
 
 		case ASN1_ITYPE_CHOICE:
-		if (asn1_cb && !asn1_cb(ASN1_OP_D2I_PRE, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_D2I_PRE, pval, it, NULL))
 				goto auxerr;
 
 		/* Allocate structure */
@@ -355,7 +357,7 @@ int ASN1_item_ex_d2i(ASN1_VALUE **pval, const unsigned char **in, long len,
 
 		asn1_set_choice_selector(pval, i, it);
 		*in = p;
-		if (asn1_cb && !asn1_cb(ASN1_OP_D2I_POST, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_D2I_POST, pval, it, NULL))
 				goto auxerr;
 		return 1;
 
@@ -402,7 +404,7 @@ int ASN1_item_ex_d2i(ASN1_VALUE **pval, const unsigned char **in, long len,
 			goto err;
 			}
 
-		if (asn1_cb && !asn1_cb(ASN1_OP_D2I_PRE, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_D2I_PRE, pval, it, NULL))
 				goto auxerr;
 
 		/* Get each field entry */
@@ -504,7 +506,7 @@ int ASN1_item_ex_d2i(ASN1_VALUE **pval, const unsigned char **in, long len,
 		if (!asn1_enc_save(pval, *in, p - *in, it))
 			goto auxerr;
 		*in = p;
-		if (asn1_cb && !asn1_cb(ASN1_OP_D2I_POST, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_D2I_POST, pval, it, NULL))
 				goto auxerr;
 		return 1;
 
@@ -664,11 +666,12 @@ static int asn1_template_noexp_d2i(ASN1_VALUE **val,
 		else
 			{
 			/* We've got a valid STACK: free up any items present */
-			STACK *sktmp = (STACK *)*val;
+			STACK_OF(ASN1_VALUE) *sktmp
+			    = (STACK_OF(ASN1_VALUE) *)*val;
 			ASN1_VALUE *vtmp;
-			while(sk_num(sktmp) > 0)
+			while(sk_ASN1_VALUE_num(sktmp) > 0)
 				{
-				vtmp = (ASN1_VALUE *)sk_pop(sktmp);
+				vtmp = sk_ASN1_VALUE_pop(sktmp);
 				ASN1_item_ex_free(&vtmp,
 						ASN1_ITEM_ptr(tt->item));
 				}
@@ -709,7 +712,8 @@ static int asn1_template_noexp_d2i(ASN1_VALUE **val,
 				goto err;
 				}
 			len -= p - q;
-			if (!sk_push((STACK *)*val, (char *)skfield))
+			if (!sk_ASN1_VALUE_push((STACK_OF(ASN1_VALUE) *)*val,
+						skfield))
 				{
 				ASN1err(ASN1_F_ASN1_TEMPLATE_NOEXP_D2I,
 						ERR_R_MALLOC_FAILURE);
diff --git a/crypto/asn1/tasn_enc.c b/crypto/asn1/tasn_enc.c
index 2721f904a633..936ad1f767c6 100644
--- a/crypto/asn1/tasn_enc.c
+++ b/crypto/asn1/tasn_enc.c
@@ -158,7 +158,7 @@ int ASN1_item_ex_i2d(ASN1_VALUE **pval, unsigned char **out,
 		return asn1_i2d_ex_primitive(pval, out, it, -1, aclass);
 
 		case ASN1_ITYPE_CHOICE:
-		if (asn1_cb && !asn1_cb(ASN1_OP_I2D_PRE, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_I2D_PRE, pval, it, NULL))
 				return 0;
 		i = asn1_get_choice_selector(pval, it);
 		if ((i >= 0) && (i < it->tcount))
@@ -171,7 +171,7 @@ int ASN1_item_ex_i2d(ASN1_VALUE **pval, unsigned char **out,
 								-1, aclass);
 			}
 		/* Fixme: error condition if selector out of range */
-		if (asn1_cb && !asn1_cb(ASN1_OP_I2D_POST, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_I2D_POST, pval, it, NULL))
 				return 0;
 		break;
 
@@ -216,7 +216,7 @@ int ASN1_item_ex_i2d(ASN1_VALUE **pval, unsigned char **out,
 			aclass = (aclass & ~ASN1_TFLG_TAG_CLASS)
 					| V_ASN1_UNIVERSAL;
 			}
-		if (asn1_cb && !asn1_cb(ASN1_OP_I2D_PRE, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_I2D_PRE, pval, it, NULL))
 				return 0;
 		/* First work out sequence content length */
 		for (i = 0, tt = it->templates; i < it->tcount; tt++, i++)
@@ -250,7 +250,7 @@ int ASN1_item_ex_i2d(ASN1_VALUE **pval, unsigned char **out,
 			}
 		if (ndef == 2)
 			ASN1_put_eoc(out);
-		if (asn1_cb  && !asn1_cb(ASN1_OP_I2D_POST, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_I2D_POST, pval, it, NULL))
 				return 0;
 		return seqlen;
 
@@ -569,7 +569,8 @@ int asn1_ex_i2c(ASN1_VALUE **pval, unsigned char *cout, int *putype,
 	ASN1_STRING *strtmp;
 	ASN1_OBJECT *otmp;
 	int utype;
-	unsigned char *cont, c;
+	const unsigned char *cont;
+	unsigned char c;
 	int len;
 	const ASN1_PRIMITIVE_FUNCS *pf;
 	pf = it->funcs;
diff --git a/crypto/asn1/tasn_fre.c b/crypto/asn1/tasn_fre.c
index d7c017fa1d42..77d3092d31ec 100644
--- a/crypto/asn1/tasn_fre.c
+++ b/crypto/asn1/tasn_fre.c
@@ -110,7 +110,7 @@ static void asn1_item_combine_free(ASN1_VALUE **pval, const ASN1_ITEM *it, int c
 		case ASN1_ITYPE_CHOICE:
 		if (asn1_cb)
 			{
-			i = asn1_cb(ASN1_OP_FREE_PRE, pval, it);
+			i = asn1_cb(ASN1_OP_FREE_PRE, pval, it, NULL);
 			if (i == 2)
 				return;
 			}
@@ -123,7 +123,7 @@ static void asn1_item_combine_free(ASN1_VALUE **pval, const ASN1_ITEM *it, int c
 			ASN1_template_free(pchval, tt);
 			}
 		if (asn1_cb)
-			asn1_cb(ASN1_OP_FREE_POST, pval, it);
+			asn1_cb(ASN1_OP_FREE_POST, pval, it, NULL);
 		if (!combine)
 			{
 			OPENSSL_free(*pval);
@@ -149,7 +149,7 @@ static void asn1_item_combine_free(ASN1_VALUE **pval, const ASN1_ITEM *it, int c
 			return;
 		if (asn1_cb)
 			{
-			i = asn1_cb(ASN1_OP_FREE_PRE, pval, it);
+			i = asn1_cb(ASN1_OP_FREE_PRE, pval, it, NULL);
 			if (i == 2)
 				return;
 			}		
@@ -170,7 +170,7 @@ static void asn1_item_combine_free(ASN1_VALUE **pval, const ASN1_ITEM *it, int c
 			ASN1_template_free(pseqval, seqtt);
 			}
 		if (asn1_cb)
-			asn1_cb(ASN1_OP_FREE_POST, pval, it);
+			asn1_cb(ASN1_OP_FREE_POST, pval, it, NULL);
 		if (!combine)
 			{
 			OPENSSL_free(*pval);
diff --git a/crypto/asn1/tasn_new.c b/crypto/asn1/tasn_new.c
index 5c6a2ebd4d04..0d9e78cc7cd0 100644
--- a/crypto/asn1/tasn_new.c
+++ b/crypto/asn1/tasn_new.c
@@ -68,7 +68,7 @@ static int asn1_item_ex_combine_new(ASN1_VALUE **pval, const ASN1_ITEM *it,
 								int combine);
 static void asn1_item_clear(ASN1_VALUE **pval, const ASN1_ITEM *it);
 static void asn1_template_clear(ASN1_VALUE **pval, const ASN1_TEMPLATE *tt);
-void asn1_primitive_clear(ASN1_VALUE **pval, const ASN1_ITEM *it);
+static void asn1_primitive_clear(ASN1_VALUE **pval, const ASN1_ITEM *it);
 
 ASN1_VALUE *ASN1_item_new(const ASN1_ITEM *it)
 	{
@@ -146,7 +146,7 @@ static int asn1_item_ex_combine_new(ASN1_VALUE **pval, const ASN1_ITEM *it,
 		case ASN1_ITYPE_CHOICE:
 		if (asn1_cb)
 			{
-			i = asn1_cb(ASN1_OP_NEW_PRE, pval, it);
+			i = asn1_cb(ASN1_OP_NEW_PRE, pval, it, NULL);
 			if (!i)
 				goto auxerr;
 			if (i==2)
@@ -166,7 +166,7 @@ static int asn1_item_ex_combine_new(ASN1_VALUE **pval, const ASN1_ITEM *it,
 			memset(*pval, 0, it->size);
 			}
 		asn1_set_choice_selector(pval, -1, it);
-		if (asn1_cb && !asn1_cb(ASN1_OP_NEW_POST, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_NEW_POST, pval, it, NULL))
 				goto auxerr;
 		break;
 
@@ -174,7 +174,7 @@ static int asn1_item_ex_combine_new(ASN1_VALUE **pval, const ASN1_ITEM *it,
 		case ASN1_ITYPE_SEQUENCE:
 		if (asn1_cb)
 			{
-			i = asn1_cb(ASN1_OP_NEW_PRE, pval, it);
+			i = asn1_cb(ASN1_OP_NEW_PRE, pval, it, NULL);
 			if (!i)
 				goto auxerr;
 			if (i==2)
@@ -201,7 +201,7 @@ static int asn1_item_ex_combine_new(ASN1_VALUE **pval, const ASN1_ITEM *it,
 			if (!ASN1_template_new(pseqval, tt))
 				goto memerr;
 			}
-		if (asn1_cb && !asn1_cb(ASN1_OP_NEW_POST, pval, it))
+		if (asn1_cb && !asn1_cb(ASN1_OP_NEW_POST, pval, it, NULL))
 				goto auxerr;
 		break;
 	}
@@ -325,6 +325,7 @@ static void asn1_template_clear(ASN1_VALUE **pval, const ASN1_TEMPLATE *tt)
 int ASN1_primitive_new(ASN1_VALUE **pval, const ASN1_ITEM *it)
 	{
 	ASN1_TYPE *typ;
+	ASN1_STRING *str;
 	int utype;
 
 	if (it && it->funcs)
@@ -345,10 +346,7 @@ int ASN1_primitive_new(ASN1_VALUE **pval, const ASN1_ITEM *it)
 		return 1;
 
 		case V_ASN1_BOOLEAN:
-		if (it)
-			*(ASN1_BOOLEAN *)pval = it->size;
-		else
-			*(ASN1_BOOLEAN *)pval = -1;
+		*(ASN1_BOOLEAN *)pval = it->size;
 		return 1;
 
 		case V_ASN1_NULL:
@@ -365,7 +363,10 @@ int ASN1_primitive_new(ASN1_VALUE **pval, const ASN1_ITEM *it)
 		break;
 
 		default:
-		*pval = (ASN1_VALUE *)ASN1_STRING_type_new(utype);
+		str = ASN1_STRING_type_new(utype);
+		if (it->itype == ASN1_ITYPE_MSTRING && str)
+			str->flags |= ASN1_STRING_FLAG_MSTRING;
+		*pval = (ASN1_VALUE *)str;
 		break;
 		}
 	if (*pval)
@@ -373,7 +374,7 @@ int ASN1_primitive_new(ASN1_VALUE **pval, const ASN1_ITEM *it)
 	return 0;
 	}
 
-void asn1_primitive_clear(ASN1_VALUE **pval, const ASN1_ITEM *it)
+static void asn1_primitive_clear(ASN1_VALUE **pval, const ASN1_ITEM *it)
 	{
 	int utype;
 	if (it && it->funcs)
diff --git a/crypto/asn1/tasn_prn.c b/crypto/asn1/tasn_prn.c
index b9c96a6dbe1d..542a091a6676 100644
--- a/crypto/asn1/tasn_prn.c
+++ b/crypto/asn1/tasn_prn.c
@@ -3,7 +3,7 @@
  * project 2000.
  */
 /* ====================================================================
- * Copyright (c) 2000 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2000,2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -58,141 +58,570 @@
 
 
 #include <stddef.h>
+#include "cryptlib.h"
 #include <openssl/asn1.h>
+#include <openssl/asn1t.h>
 #include <openssl/objects.h>
 #include <openssl/buffer.h>
 #include <openssl/err.h>
-#include <openssl/nasn.h>
+#include <openssl/x509v3.h>
+#include "asn1_locl.h"
 
-/* Print routines. Print out a whole structure from a template.
+/* Print routines.
  */
 
-static int asn1_item_print_nm(BIO *out, void *fld, int indent, const ASN1_ITEM *it, const char *name);
+/* ASN1_PCTX routines */
 
-int ASN1_item_print(BIO *out, void *fld, int indent, const ASN1_ITEM *it)
-{
-	return asn1_item_print_nm(out, fld, indent, it, it->sname);
-}
+ASN1_PCTX default_pctx = 
+	{
+	ASN1_PCTX_FLAGS_SHOW_ABSENT,	/* flags */
+	0,	/* nm_flags */
+	0,	/* cert_flags */
+	0,	/* oid_flags */
+	0	/* str_flags */
+	};
+	
 
-static int asn1_item_print_nm(BIO *out, void *fld, int indent, const ASN1_ITEM *it, const char *name)
-{
-	ASN1_STRING *str;
+ASN1_PCTX *ASN1_PCTX_new(void)
+	{
+	ASN1_PCTX *ret;
+	ret = OPENSSL_malloc(sizeof(ASN1_PCTX));
+	if (ret == NULL)
+		{
+		ASN1err(ASN1_F_ASN1_PCTX_NEW, ERR_R_MALLOC_FAILURE);
+		return NULL;
+		}
+	ret->flags = 0;
+	ret->nm_flags = 0;
+	ret->cert_flags = 0;
+	ret->oid_flags = 0;
+	ret->str_flags = 0;
+	return ret;
+	}
+
+void ASN1_PCTX_free(ASN1_PCTX *p)
+	{
+	OPENSSL_free(p);
+	}
+
+unsigned long ASN1_PCTX_get_flags(ASN1_PCTX *p)
+	{
+	return p->flags;
+	}
+
+void ASN1_PCTX_set_flags(ASN1_PCTX *p, unsigned long flags)
+	{
+	p->flags = flags;
+	}
+
+unsigned long ASN1_PCTX_get_nm_flags(ASN1_PCTX *p)
+	{
+	return p->nm_flags;
+	}
+
+void ASN1_PCTX_set_nm_flags(ASN1_PCTX *p, unsigned long flags)
+	{
+	p->nm_flags = flags;
+	}
+
+unsigned long ASN1_PCTX_get_cert_flags(ASN1_PCTX *p)
+	{
+	return p->cert_flags;
+	}
+
+void ASN1_PCTX_set_cert_flags(ASN1_PCTX *p, unsigned long flags)
+	{
+	p->cert_flags = flags;
+	}
+
+unsigned long ASN1_PCTX_get_oid_flags(ASN1_PCTX *p)
+	{
+	return p->oid_flags;
+	}
+
+void ASN1_PCTX_set_oid_flags(ASN1_PCTX *p, unsigned long flags)
+	{
+	p->oid_flags = flags;
+	}
+
+unsigned long ASN1_PCTX_get_str_flags(ASN1_PCTX *p)
+	{
+	return p->str_flags;
+	}
+
+void ASN1_PCTX_set_str_flags(ASN1_PCTX *p, unsigned long flags)
+	{
+	p->str_flags = flags;
+	}
+
+/* Main print routines */
+
+static int asn1_item_print_ctx(BIO *out, ASN1_VALUE **fld, int indent,
+				const ASN1_ITEM *it,
+				const char *fname, const char *sname,
+				int nohdr, const ASN1_PCTX *pctx);
+
+int asn1_template_print_ctx(BIO *out, ASN1_VALUE **fld, int indent,
+				const ASN1_TEMPLATE *tt, const ASN1_PCTX *pctx);
+
+static int asn1_primitive_print(BIO *out, ASN1_VALUE **fld,
+				const ASN1_ITEM *it, int indent,
+				const char *fname, const char *sname,
+				const ASN1_PCTX *pctx);
+
+static int asn1_print_fsname(BIO *out, int indent,
+			const char *fname, const char *sname,
+			const ASN1_PCTX *pctx);
+
+int ASN1_item_print(BIO *out, ASN1_VALUE *ifld, int indent,
+				const ASN1_ITEM *it, const ASN1_PCTX *pctx)
+	{
+	const char *sname;
+	if (pctx == NULL)
+		pctx = &default_pctx;
+	if (pctx->flags & ASN1_PCTX_FLAGS_NO_STRUCT_NAME)
+		sname = NULL;
+	else
+		sname = it->sname;
+	return asn1_item_print_ctx(out, &ifld, indent, it,
+							NULL, sname, 0, pctx);
+	}
+
+static int asn1_item_print_ctx(BIO *out, ASN1_VALUE **fld, int indent,
+				const ASN1_ITEM *it,
+				const char *fname, const char *sname,
+				int nohdr, const ASN1_PCTX *pctx)
+	{
 	const ASN1_TEMPLATE *tt;
-	void *tmpfld;
+	const ASN1_EXTERN_FUNCS *ef;
+	ASN1_VALUE **tmpfld;
+	const ASN1_AUX *aux = it->funcs;
+	ASN1_aux_cb *asn1_cb;
+	ASN1_PRINT_ARG parg;
 	int i;
-	if(!fld) {
-		BIO_printf(out, "%*s%s ABSENT\n", indent, "", name);
+	if (aux && aux->asn1_cb)
+		{
+		parg.out = out;
+		parg.indent = indent;
+		parg.pctx = pctx;
+		asn1_cb = aux->asn1_cb;
+		}
+	else asn1_cb = 0;
+
+	if(*fld == NULL)
+		{
+		if (pctx->flags & ASN1_PCTX_FLAGS_SHOW_ABSENT)
+			{
+			if (!nohdr && !asn1_print_fsname(out, indent,
+							fname, sname, pctx))
+				return 0;
+			if (BIO_puts(out, "<ABSENT>\n") <= 0)
+				return 0;
+			}
 		return 1;
-	}
-	switch(it->itype) {
+		}
 
+	switch(it->itype)
+		{
 		case ASN1_ITYPE_PRIMITIVE:
 		if(it->templates)
-			return ASN1_template_print(out, fld, indent, it->templates);
-		return asn1_primitive_print(out, fld, it->utype, indent, name);
-		break;
-
+			{
+			if (!asn1_template_print_ctx(out, fld, indent,
+							it->templates, pctx))
+				return 0;
+			}
+		/* fall thru */
 		case ASN1_ITYPE_MSTRING:
-		str = fld;
-		return asn1_primitive_print(out, fld, str->type, indent, name);
+		if (!asn1_primitive_print(out, fld, it,
+				indent, fname, sname,pctx))
+			return 0;
+		break;
 
 		case ASN1_ITYPE_EXTERN:
-		BIO_printf(out, "%*s%s:EXTERNAL TYPE %s %s\n", indent, "", name, it->sname, fld ? "" : "ABSENT");
-		return 1;
-		case ASN1_ITYPE_COMPAT:
-		BIO_printf(out, "%*s%s:COMPATIBLE TYPE %s %s\n", indent, "", name, it->sname, fld ? "" : "ABSENT");
-		return 1;
-
+		if (!nohdr && !asn1_print_fsname(out, indent, fname, sname, pctx))
+			return 0;
+		/* Use new style print routine if possible */
+		ef = it->funcs;
+		if (ef && ef->asn1_ex_print)
+			{
+			i = ef->asn1_ex_print(out, fld, indent, "", pctx);
+			if (!i)
+				return 0;
+			if ((i == 2) && (BIO_puts(out, "\n") <= 0))
+				return 0;
+			return 1;
+			}
+		else if (sname && 
+			BIO_printf(out, ":EXTERNAL TYPE %s\n", sname) <= 0)
+			return 0;
+		break;
 
 		case ASN1_ITYPE_CHOICE:
+#if 0
+		if (!nohdr && !asn1_print_fsname(out, indent, fname, sname, pctx))
+			return 0;
+#endif
 		/* CHOICE type, get selector */
 		i = asn1_get_choice_selector(fld, it);
 		/* This should never happen... */
-		if((i < 0) || (i >= it->tcount)) {
-			BIO_printf(out, "%s selector [%d] out of range\n", it->sname, i);
+		if((i < 0) || (i >= it->tcount))
+			{
+			if (BIO_printf(out,
+				"ERROR: selector [%d] invalid\n", i) <= 0)
+				return 0;
 			return 1;
-		}
+			}
 		tt = it->templates + i;
-		tmpfld = asn1_get_field(fld, tt);
-		return ASN1_template_print(out, tmpfld, indent, tt);
+		tmpfld = asn1_get_field_ptr(fld, tt);
+		if (!asn1_template_print_ctx(out, tmpfld, indent, tt, pctx))
+			return 0;
+		break;
 
 		case ASN1_ITYPE_SEQUENCE:
-		BIO_printf(out, "%*s%s {\n", indent, "", name);
-		/* Get each field entry */
-		for(i = 0, tt = it->templates; i < it->tcount; i++, tt++) {
-			tmpfld = asn1_get_field(fld, tt);
-			ASN1_template_print(out, tmpfld, indent + 2, tt);
-		}
-		BIO_printf(out, "%*s}\n", indent, "");
-		return 1;
+		case ASN1_ITYPE_NDEF_SEQUENCE:
+		if (!nohdr && !asn1_print_fsname(out, indent, fname, sname, pctx))
+			return 0;
+		if (fname || sname)
+			{
+			if (pctx->flags & ASN1_PCTX_FLAGS_SHOW_SEQUENCE)
+				{
+				if (BIO_puts(out, " {\n") <= 0)
+					return 0;
+				}
+			else
+				{
+				if (BIO_puts(out, "\n") <= 0)
+					return 0;
+				}
+			}
+
+		if (asn1_cb)
+			{
+			i = asn1_cb(ASN1_OP_PRINT_PRE, fld, it, &parg);
+			if (i == 0)
+				return 0;
+			if (i == 2)
+				return 1;
+			}
+
+		/* Print each field entry */
+		for(i = 0, tt = it->templates; i < it->tcount; i++, tt++)
+			{
+			const ASN1_TEMPLATE *seqtt;
+			seqtt = asn1_do_adb(fld, tt, 1);
+			tmpfld = asn1_get_field_ptr(fld, seqtt);
+			if (!asn1_template_print_ctx(out, tmpfld,
+						indent + 2, seqtt, pctx))
+				return 0;
+			}
+		if (pctx->flags & ASN1_PCTX_FLAGS_SHOW_SEQUENCE)
+			{
+			if (BIO_printf(out, "%*s}\n", indent, "") < 0)
+				return 0;
+			}
+
+		if (asn1_cb)
+			{
+			i = asn1_cb(ASN1_OP_PRINT_POST, fld, it, &parg);
+			if (i == 0)
+				return 0;
+			}
+		break;
 
 		default:
+		BIO_printf(out, "Unprocessed type %d\n", it->itype);
 		return 0;
+		}
+
+	return 1;
 	}
-}
 
-int ASN1_template_print(BIO *out, void *fld, int indent, const ASN1_TEMPLATE *tt)
-{
+int asn1_template_print_ctx(BIO *out, ASN1_VALUE **fld, int indent,
+				const ASN1_TEMPLATE *tt, const ASN1_PCTX *pctx)
+	{
 	int i, flags;
-#if 0
-	if(!fld) return 0; 
-#endif
+	const char *sname, *fname;
 	flags = tt->flags;
-	if(flags & ASN1_TFLG_SK_MASK) {
+	if(pctx->flags & ASN1_PCTX_FLAGS_SHOW_FIELD_STRUCT_NAME)
+		sname = ASN1_ITEM_ptr(tt->item)->sname;
+	else
+		sname = NULL;
+	if(pctx->flags & ASN1_PCTX_FLAGS_NO_FIELD_NAME)
+		fname = NULL;
+	else
+		fname = tt->field_name;
+	if(flags & ASN1_TFLG_SK_MASK)
+		{
 		char *tname;
-		void *skitem;
+		ASN1_VALUE *skitem;
+		STACK_OF(ASN1_VALUE) *stack;
+
 		/* SET OF, SEQUENCE OF */
-		if(flags & ASN1_TFLG_SET_OF) tname = "SET";
-		else tname = "SEQUENCE";
-		if(fld) {
-			BIO_printf(out, "%*s%s OF %s {\n", indent, "", tname, tt->field_name);
-			for(i = 0; i < sk_num(fld); i++) {
-				skitem = sk_value(fld, i);
-				asn1_item_print_nm(out, skitem, indent + 2, tt->item, "");
+		if (fname)
+			{
+			if(pctx->flags & ASN1_PCTX_FLAGS_SHOW_SSOF)
+				{
+				if(flags & ASN1_TFLG_SET_OF)
+					tname = "SET";
+				else
+					tname = "SEQUENCE";
+				if (BIO_printf(out, "%*s%s OF %s {\n",
+					indent, "", tname, tt->field_name) <= 0)
+					return 0;
+				}
+			else if (BIO_printf(out, "%*s%s:\n", indent, "",
+					fname) <= 0)
+				return 0;
+			}
+		stack = (STACK_OF(ASN1_VALUE) *)*fld;
+		for(i = 0; i < sk_ASN1_VALUE_num(stack); i++)
+			{
+			if ((i > 0) && (BIO_puts(out, "\n") <= 0))
+				return 0;
+
+			skitem = sk_ASN1_VALUE_value(stack, i);
+			if (!asn1_item_print_ctx(out, &skitem, indent + 2,
+				ASN1_ITEM_ptr(tt->item), NULL, NULL, 1, pctx))
+				return 0;
+			}
+		if (!i && BIO_printf(out, "%*s<EMPTY>\n", indent + 2, "") <= 0)
+				return 0;
+		if(pctx->flags & ASN1_PCTX_FLAGS_SHOW_SEQUENCE)
+			{
+			if (BIO_printf(out, "%*s}\n", indent, "") <= 0)
+				return 0;
 			}
-			BIO_printf(out, "%*s}\n", indent, "");
-		} else 
-			BIO_printf(out, "%*s%s OF %s ABSENT\n", indent, "", tname, tt->field_name);
 		return 1;
+		}
+	return asn1_item_print_ctx(out, fld, indent, ASN1_ITEM_ptr(tt->item),
+							fname, sname, 0, pctx);
 	}
-	return asn1_item_print_nm(out, fld, indent, tt->item, tt->field_name);
-}
-
-static int asn1_primitive_print(BIO *out, void *fld, long utype, int indent, const char *name)
-{
-	ASN1_STRING *str = fld;
-	if(fld) {
-		if(utype == V_ASN1_BOOLEAN) {
-			int *bool = fld;
-if(*bool == -1) printf("BOOL MISSING\n");
-			BIO_printf(out, "%*s%s:%s", indent, "", "BOOLEAN", *bool ? "TRUE" : "FALSE");
-		} else if((utype == V_ASN1_INTEGER) 
-			  || (utype == V_ASN1_ENUMERATED)) {
-			char *s, *nm;
-			s = i2s_ASN1_INTEGER(NULL, fld);
-			if(utype == V_ASN1_INTEGER) nm = "INTEGER";
-			else nm = "ENUMERATED";
-			BIO_printf(out, "%*s%s:%s", indent, "", nm, s);
-			OPENSSL_free(s);
-		} else if(utype == V_ASN1_NULL) {
-			BIO_printf(out, "%*s%s", indent, "", "NULL");
-		} else if(utype == V_ASN1_UTCTIME) {
-			BIO_printf(out, "%*s%s:%s:", indent, "", name, "UTCTIME");
-			ASN1_UTCTIME_print(out, str);
-		} else if(utype == V_ASN1_GENERALIZEDTIME) {
-			BIO_printf(out, "%*s%s:%s:", indent, "", name, "GENERALIZEDTIME");
-			ASN1_GENERALIZEDTIME_print(out, str);
-		} else if(utype == V_ASN1_OBJECT) {
-			char objbuf[80], *ln;
-			ln = OBJ_nid2ln(OBJ_obj2nid(fld));
-			if(!ln) ln = "";
-			OBJ_obj2txt(objbuf, sizeof objbuf, fld, 1);
-			BIO_printf(out, "%*s%s:%s (%s)", indent, "", "OBJECT", ln, objbuf);
-		} else {
-			BIO_printf(out, "%*s%s:", indent, "", name);
-			ASN1_STRING_print_ex(out, str, ASN1_STRFLGS_DUMP_UNKNOWN|ASN1_STRFLGS_SHOW_TYPE);
+
+static int asn1_print_fsname(BIO *out, int indent,
+			const char *fname, const char *sname,
+			const ASN1_PCTX *pctx)
+	{
+	static char spaces[] = "                    ";
+	const int nspaces = sizeof(spaces) - 1;
+
+#if 0
+	if (!sname && !fname)
+		return 1;
+#endif
+
+	while (indent > nspaces)
+		{
+		if (BIO_write(out, spaces, nspaces) != nspaces)
+			return 0;
+		indent -= nspaces;
+		}
+	if (BIO_write(out, spaces, indent) != indent)
+		return 0;
+	if (pctx->flags & ASN1_PCTX_FLAGS_NO_STRUCT_NAME)
+		sname = NULL;
+	if (pctx->flags & ASN1_PCTX_FLAGS_NO_FIELD_NAME)
+		fname = NULL;
+	if (!sname && !fname)
+		return 1;
+	if (fname)
+		{
+		if (BIO_puts(out, fname) <= 0)
+			return 0;
 		}
-		BIO_printf(out, "\n");
-	} else BIO_printf(out, "%*s%s [ABSENT]\n", indent, "", name);
+	if (sname)
+		{
+		if (fname)
+			{
+			if (BIO_printf(out, " (%s)", sname) <= 0)
+				return 0;
+			}
+		else
+			{
+			if (BIO_puts(out, sname) <= 0)
+				return 0;
+			}
+		}
+	if (BIO_write(out, ": ", 2) != 2)
+		return 0;
 	return 1;
-}
+	}
+
+static int asn1_print_boolean_ctx(BIO *out, int boolval,
+							const ASN1_PCTX *pctx)
+	{
+	const char *str;
+	switch (boolval)
+		{
+		case -1:
+		str = "BOOL ABSENT";
+		break;
+
+		case 0:
+		str = "FALSE";
+		break;
+
+		default:
+		str = "TRUE";
+		break;
+
+		}
+
+	if (BIO_puts(out, str) <= 0)
+		return 0;
+	return 1;
+
+	}
+
+static int asn1_print_integer_ctx(BIO *out, ASN1_INTEGER *str,
+						const ASN1_PCTX *pctx)
+	{
+	char *s;
+	int ret = 1;
+	s = i2s_ASN1_INTEGER(NULL, str);
+	if (BIO_puts(out, s) <= 0)
+		ret = 0;
+	OPENSSL_free(s);
+	return ret;
+	}
+
+static int asn1_print_oid_ctx(BIO *out, const ASN1_OBJECT *oid,
+						const ASN1_PCTX *pctx)
+	{
+	char objbuf[80];
+	const char *ln;
+	ln = OBJ_nid2ln(OBJ_obj2nid(oid));
+	if(!ln)
+		ln = "";
+	OBJ_obj2txt(objbuf, sizeof objbuf, oid, 1);
+	if (BIO_printf(out, "%s (%s)", ln, objbuf) <= 0)
+		return 0;
+	return 1;
+	}
+
+static int asn1_print_obstring_ctx(BIO *out, ASN1_STRING *str, int indent,
+						const ASN1_PCTX *pctx)
+	{
+	if (str->type == V_ASN1_BIT_STRING)
+		{
+		if (BIO_printf(out, " (%ld unused bits)\n",
+					str->flags & 0x7) <= 0)
+				return 0;
+		}
+	else if (BIO_puts(out, "\n") <= 0)
+		return 0;
+	if ((str->length > 0)
+		&& BIO_dump_indent(out, (char *)str->data, str->length,
+				indent + 2) <= 0)
+		return 0;
+	return 1;
+	}
+
+static int asn1_primitive_print(BIO *out, ASN1_VALUE **fld,
+				const ASN1_ITEM *it, int indent,
+				const char *fname, const char *sname,
+				const ASN1_PCTX *pctx)
+	{
+	long utype;
+	ASN1_STRING *str;
+	int ret = 1, needlf = 1;
+	const char *pname;
+	const ASN1_PRIMITIVE_FUNCS *pf;
+	pf = it->funcs;
+	if (!asn1_print_fsname(out, indent, fname, sname, pctx))
+			return 0;
+	if (pf && pf->prim_print)
+		return pf->prim_print(out, fld, it, indent, pctx);
+	str = (ASN1_STRING *)*fld;
+	if (it->itype == ASN1_ITYPE_MSTRING)
+		utype = str->type & ~V_ASN1_NEG;
+	else
+		utype = it->utype;
+	if (utype == V_ASN1_ANY)
+		{
+		ASN1_TYPE *atype = (ASN1_TYPE *)*fld;
+		utype = atype->type;
+		fld = &atype->value.asn1_value;
+		str = (ASN1_STRING *)*fld;
+		if (pctx->flags & ASN1_PCTX_FLAGS_NO_ANY_TYPE)
+			pname = NULL;
+		else 
+			pname = ASN1_tag2str(utype);
+		}
+	else
+		{
+		if (pctx->flags & ASN1_PCTX_FLAGS_SHOW_TYPE)
+			pname = ASN1_tag2str(utype);
+		else 
+			pname = NULL;
+		}
+
+	if (utype == V_ASN1_NULL)
+		{
+		if (BIO_puts(out, "NULL\n") <= 0)
+			return 0;
+		return 1;
+		}
+
+	if (pname)
+		{
+		if (BIO_puts(out, pname) <= 0)
+			return 0;
+		if (BIO_puts(out, ":") <= 0)
+			return 0;
+		}
+
+	switch (utype)
+		{
+		case V_ASN1_BOOLEAN:
+			{
+			int boolval = *(int *)fld;
+			if (boolval == -1)
+				boolval = it->size;
+			ret = asn1_print_boolean_ctx(out, boolval, pctx);
+			}
+		break;
+
+		case V_ASN1_INTEGER:
+		case V_ASN1_ENUMERATED:
+		ret = asn1_print_integer_ctx(out, str, pctx);
+		break;
+
+		case V_ASN1_UTCTIME:
+		ret = ASN1_UTCTIME_print(out, str);
+		break;
+
+		case V_ASN1_GENERALIZEDTIME:
+		ret = ASN1_GENERALIZEDTIME_print(out, str);
+		break;
+
+		case V_ASN1_OBJECT:
+		ret = asn1_print_oid_ctx(out, (const ASN1_OBJECT *)*fld, pctx);
+		break;
+
+		case V_ASN1_OCTET_STRING:
+		case V_ASN1_BIT_STRING:
+		ret = asn1_print_obstring_ctx(out, str, indent, pctx);
+		needlf = 0;
+		break;
+
+		case V_ASN1_SEQUENCE:
+		case V_ASN1_SET:
+		case V_ASN1_OTHER:
+		if (BIO_puts(out, "\n") <= 0)
+			return 0;
+		if (ASN1_parse_dump(out, str->data, str->length,
+						indent, 0) <= 0)
+			ret = 0;
+		needlf = 0;
+		break;
+
+		default:
+		ret = ASN1_STRING_print_ex(out, str, pctx->str_flags);
+
+		}
+	if (!ret)
+		return 0;
+	if (needlf && BIO_puts(out, "\n") <= 0)
+		return 0;
+	return 1;
+	}
diff --git a/crypto/asn1/tasn_typ.c b/crypto/asn1/tasn_typ.c
index 6252213d15f0..6fb1c372dab1 100644
--- a/crypto/asn1/tasn_typ.c
+++ b/crypto/asn1/tasn_typ.c
@@ -135,3 +135,14 @@ IMPLEMENT_ASN1_TYPE_ex(ASN1_FBOOLEAN, ASN1_BOOLEAN, 0)
 /* Special, OCTET STRING with indefinite length constructed support */
 
 IMPLEMENT_ASN1_TYPE_ex(ASN1_OCTET_STRING_NDEF, ASN1_OCTET_STRING, ASN1_TFLG_NDEF)
+
+ASN1_ITEM_TEMPLATE(ASN1_SEQUENCE_ANY) = 
+	ASN1_EX_TEMPLATE_TYPE(ASN1_TFLG_SEQUENCE_OF, 0, ASN1_SEQUENCE_ANY, ASN1_ANY)
+ASN1_ITEM_TEMPLATE_END(ASN1_SEQUENCE_ANY)
+
+ASN1_ITEM_TEMPLATE(ASN1_SET_ANY) = 
+	ASN1_EX_TEMPLATE_TYPE(ASN1_TFLG_SET_OF, 0, ASN1_SET_ANY, ASN1_ANY)
+ASN1_ITEM_TEMPLATE_END(ASN1_SET_ANY)
+
+IMPLEMENT_ASN1_ENCODE_FUNCTIONS_const_fname(ASN1_SEQUENCE_ANY, ASN1_SEQUENCE_ANY, ASN1_SEQUENCE_ANY)
+IMPLEMENT_ASN1_ENCODE_FUNCTIONS_const_fname(ASN1_SEQUENCE_ANY, ASN1_SET_ANY, ASN1_SET_ANY)
diff --git a/crypto/asn1/x_algor.c b/crypto/asn1/x_algor.c
index 99e53429b797..274e456c7382 100644
--- a/crypto/asn1/x_algor.c
+++ b/crypto/asn1/x_algor.c
@@ -128,3 +128,17 @@ void X509_ALGOR_get0(ASN1_OBJECT **paobj, int *pptype, void **ppval,
 		}
 	}
 
+/* Set up an X509_ALGOR DigestAlgorithmIdentifier from an EVP_MD */
+
+void X509_ALGOR_set_md(X509_ALGOR *alg, const EVP_MD *md)
+	{
+	int param_type;
+
+	if (md->flags & EVP_MD_FLAG_DIGALGID_ABSENT)
+		param_type = V_ASN1_UNDEF;
+	else
+		param_type = V_ASN1_NULL;
+
+	X509_ALGOR_set0(alg, OBJ_nid2obj(EVP_MD_type(md)), param_type, NULL);
+
+	}
diff --git a/crypto/asn1/x_crl.c b/crypto/asn1/x_crl.c
index 70d56a67f26a..c51c690ba9d3 100644
--- a/crypto/asn1/x_crl.c
+++ b/crypto/asn1/x_crl.c
@@ -58,11 +58,14 @@
 
 #include <stdio.h>
 #include "cryptlib.h"
+#include "asn1_locl.h"
 #include <openssl/asn1t.h>
 #include <openssl/x509.h>
+#include <openssl/x509v3.h>
 
 static int X509_REVOKED_cmp(const X509_REVOKED * const *a,
 				const X509_REVOKED * const *b);
+static void setup_idp(X509_CRL *crl, ISSUING_DIST_POINT *idp);
 
 ASN1_SEQUENCE(X509_REVOKED) = {
 	ASN1_SIMPLE(X509_REVOKED,serialNumber, ASN1_INTEGER),
@@ -70,11 +73,26 @@ ASN1_SEQUENCE(X509_REVOKED) = {
 	ASN1_SEQUENCE_OF_OPT(X509_REVOKED,extensions, X509_EXTENSION)
 } ASN1_SEQUENCE_END(X509_REVOKED)
 
+static int def_crl_verify(X509_CRL *crl, EVP_PKEY *r);
+static int def_crl_lookup(X509_CRL *crl,
+		X509_REVOKED **ret, ASN1_INTEGER *serial, X509_NAME *issuer);
+
+static X509_CRL_METHOD int_crl_meth =
+	{
+	0,
+	0,0,
+	def_crl_lookup,
+	def_crl_verify
+	};
+
+static const X509_CRL_METHOD *default_crl_method = &int_crl_meth;
+
 /* The X509_CRL_INFO structure needs a bit of customisation.
  * Since we cache the original encoding the signature wont be affected by
  * reordering of the revoked field.
  */
-static int crl_inf_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int crl_inf_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+								void *exarg)
 {
 	X509_CRL_INFO *a = (X509_CRL_INFO *)*pval;
 
@@ -101,7 +119,237 @@ ASN1_SEQUENCE_enc(X509_CRL_INFO, enc, crl_inf_cb) = {
 	ASN1_EXP_SEQUENCE_OF_OPT(X509_CRL_INFO, extensions, X509_EXTENSION, 0)
 } ASN1_SEQUENCE_END_enc(X509_CRL_INFO, X509_CRL_INFO)
 
-ASN1_SEQUENCE_ref(X509_CRL, 0, CRYPTO_LOCK_X509_CRL) = {
+/* Set CRL entry issuer according to CRL certificate issuer extension.
+ * Check for unhandled critical CRL entry extensions.
+ */
+
+static int crl_set_issuers(X509_CRL *crl)
+	{
+
+	int i, j;
+	GENERAL_NAMES *gens, *gtmp;
+	STACK_OF(X509_REVOKED) *revoked;
+
+	revoked = X509_CRL_get_REVOKED(crl);
+
+	gens = NULL;
+	for (i = 0; i < sk_X509_REVOKED_num(revoked); i++)
+		{
+		X509_REVOKED *rev = sk_X509_REVOKED_value(revoked, i);
+		STACK_OF(X509_EXTENSION) *exts;
+		ASN1_ENUMERATED *reason;
+		X509_EXTENSION *ext;
+		gtmp = X509_REVOKED_get_ext_d2i(rev, 
+						NID_certificate_issuer,
+						&j, NULL);
+		if (!gtmp && (j != -1))
+			{
+			crl->flags |= EXFLAG_INVALID;
+			return 1;
+			}
+
+		if (gtmp)
+			{
+			gens = gtmp;
+			if (!crl->issuers)
+				{
+				crl->issuers = sk_GENERAL_NAMES_new_null();
+				if (!crl->issuers)
+					return 0;
+				}
+			if (!sk_GENERAL_NAMES_push(crl->issuers, gtmp))
+				return 0;
+			}
+		rev->issuer = gens;
+
+		reason = X509_REVOKED_get_ext_d2i(rev, NID_crl_reason,
+								&j, NULL);
+		if (!reason && (j != -1))
+			{
+			crl->flags |= EXFLAG_INVALID;
+			return 1;
+			}
+
+		if (reason)
+			{
+			rev->reason = ASN1_ENUMERATED_get(reason);
+			ASN1_ENUMERATED_free(reason);
+			}
+		else
+			rev->reason = CRL_REASON_NONE;	
+
+		/* Check for critical CRL entry extensions */
+
+		exts = rev->extensions;
+
+		for (j = 0; j < sk_X509_EXTENSION_num(exts); j++)
+			{
+			ext = sk_X509_EXTENSION_value(exts, j);
+			if (ext->critical > 0)
+				{
+				if (OBJ_obj2nid(ext->object) ==
+					NID_certificate_issuer)
+					continue;
+				crl->flags |= EXFLAG_CRITICAL;
+				break;
+				}
+			}
+
+
+		}
+
+	return 1;
+
+	}
+
+/* The X509_CRL structure needs a bit of customisation. Cache some extensions
+ * and hash of the whole CRL.
+ */
+static int crl_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+								void *exarg)
+	{
+	X509_CRL *crl = (X509_CRL *)*pval;
+	STACK_OF(X509_EXTENSION) *exts;
+	X509_EXTENSION *ext;
+	int idx;
+
+	switch(operation)
+		{
+		case ASN1_OP_NEW_POST:
+		crl->idp = NULL;
+		crl->akid = NULL;
+		crl->flags = 0;
+		crl->idp_flags = 0;
+		crl->idp_reasons = CRLDP_ALL_REASONS;
+		crl->meth = default_crl_method;
+		crl->meth_data = NULL;
+		crl->issuers = NULL;
+		crl->crl_number = NULL;
+		crl->base_crl_number = NULL;
+		break;
+
+		case ASN1_OP_D2I_POST:
+#ifndef OPENSSL_NO_SHA
+		X509_CRL_digest(crl, EVP_sha1(), crl->sha1_hash, NULL);
+#endif
+		crl->idp = X509_CRL_get_ext_d2i(crl,
+				NID_issuing_distribution_point, NULL, NULL);
+		if (crl->idp)
+			setup_idp(crl, crl->idp);
+
+		crl->akid = X509_CRL_get_ext_d2i(crl,
+				NID_authority_key_identifier, NULL, NULL);	
+
+		crl->crl_number = X509_CRL_get_ext_d2i(crl,
+				NID_crl_number, NULL, NULL);	
+
+		crl->base_crl_number = X509_CRL_get_ext_d2i(crl,
+				NID_delta_crl, NULL, NULL);	
+		/* Delta CRLs must have CRL number */
+		if (crl->base_crl_number && !crl->crl_number)
+			crl->flags |= EXFLAG_INVALID;
+
+		/* See if we have any unhandled critical CRL extensions and 
+		 * indicate this in a flag. We only currently handle IDP so
+		 * anything else critical sets the flag.
+		 *
+		 * This code accesses the X509_CRL structure directly:
+		 * applications shouldn't do this.
+		 */
+
+		exts = crl->crl->extensions;
+
+		for (idx = 0; idx < sk_X509_EXTENSION_num(exts); idx++)
+			{
+			int nid;
+			ext = sk_X509_EXTENSION_value(exts, idx);
+			nid = OBJ_obj2nid(ext->object);
+			if (nid == NID_freshest_crl)
+				crl->flags |= EXFLAG_FRESHEST;
+			if (ext->critical > 0)
+				{
+				/* We handle IDP and deltas */
+				if ((nid == NID_issuing_distribution_point)
+					|| (nid == NID_delta_crl))
+					break;;
+				crl->flags |= EXFLAG_CRITICAL;
+				break;
+				}
+			}
+
+
+		if (!crl_set_issuers(crl))
+			return 0;
+
+		if (crl->meth->crl_init)
+			{
+			if (crl->meth->crl_init(crl) == 0)
+				return 0;
+			}
+		break;
+
+		case ASN1_OP_FREE_POST:
+		if (crl->meth->crl_free)
+			{
+			if (!crl->meth->crl_free(crl))
+				return 0;
+			}
+		if (crl->akid)
+			AUTHORITY_KEYID_free(crl->akid);
+		if (crl->idp)
+			ISSUING_DIST_POINT_free(crl->idp);
+		ASN1_INTEGER_free(crl->crl_number);
+		ASN1_INTEGER_free(crl->base_crl_number);
+		sk_GENERAL_NAMES_pop_free(crl->issuers, GENERAL_NAMES_free);
+		break;
+		}
+	return 1;
+	}
+
+/* Convert IDP into a more convenient form */
+
+static void setup_idp(X509_CRL *crl, ISSUING_DIST_POINT *idp)
+	{
+	int idp_only = 0;
+	/* Set various flags according to IDP */
+	crl->idp_flags |= IDP_PRESENT;
+	if (idp->onlyuser > 0)
+		{
+		idp_only++;
+		crl->idp_flags |= IDP_ONLYUSER;
+		}
+	if (idp->onlyCA > 0)
+		{
+		idp_only++;
+		crl->idp_flags |= IDP_ONLYCA;
+		}
+	if (idp->onlyattr > 0)
+		{
+		idp_only++;
+		crl->idp_flags |= IDP_ONLYATTR;
+		}
+
+	if (idp_only > 1)
+		crl->idp_flags |= IDP_INVALID;
+
+	if (idp->indirectCRL > 0)
+		crl->idp_flags |= IDP_INDIRECT;
+
+	if (idp->onlysomereasons)
+		{
+		crl->idp_flags |= IDP_REASONS;
+		if (idp->onlysomereasons->length > 0)
+			crl->idp_reasons = idp->onlysomereasons->data[0];
+		if (idp->onlysomereasons->length > 1)
+			crl->idp_reasons |=
+				(idp->onlysomereasons->data[1] << 8);
+		crl->idp_reasons &= CRLDP_ALL_REASONS;
+		}
+
+	DIST_POINT_set_dpname(idp->distpoint, X509_CRL_get_issuer(crl));
+	}
+
+ASN1_SEQUENCE_ref(X509_CRL, crl_cb, CRYPTO_LOCK_X509_CRL) = {
 	ASN1_SIMPLE(X509_CRL, crl, X509_CRL_INFO),
 	ASN1_SIMPLE(X509_CRL, sig_alg, X509_ALGOR),
 	ASN1_SIMPLE(X509_CRL, signature, ASN1_BIT_STRING)
@@ -134,6 +382,145 @@ int X509_CRL_add0_revoked(X509_CRL *crl, X509_REVOKED *rev)
 	return 1;
 }
 
+int X509_CRL_verify(X509_CRL *crl, EVP_PKEY *r)
+	{
+	if (crl->meth->crl_verify)
+		return crl->meth->crl_verify(crl, r);
+	return 0;
+	}
+
+int X509_CRL_get0_by_serial(X509_CRL *crl,
+		X509_REVOKED **ret, ASN1_INTEGER *serial)
+	{
+	if (crl->meth->crl_lookup)
+		return crl->meth->crl_lookup(crl, ret, serial, NULL);
+	return 0;
+	}
+
+int X509_CRL_get0_by_cert(X509_CRL *crl, X509_REVOKED **ret, X509 *x)
+	{
+	if (crl->meth->crl_lookup)
+		return crl->meth->crl_lookup(crl, ret,
+						X509_get_serialNumber(x),
+						X509_get_issuer_name(x));
+	return 0;
+	}
+
+static int def_crl_verify(X509_CRL *crl, EVP_PKEY *r)
+	{
+	return(ASN1_item_verify(ASN1_ITEM_rptr(X509_CRL_INFO),
+		crl->sig_alg, crl->signature,crl->crl,r));
+	}
+
+static int crl_revoked_issuer_match(X509_CRL *crl, X509_NAME *nm,
+						X509_REVOKED *rev)
+	{
+	int i;
+
+	if (!rev->issuer)
+		{
+		if (!nm)
+			return 1;
+		if (!X509_NAME_cmp(nm, X509_CRL_get_issuer(crl)))
+			return 1;
+		return 0;
+		}
+
+	if (!nm)
+		nm = X509_CRL_get_issuer(crl);
+
+	for (i = 0; i < sk_GENERAL_NAME_num(rev->issuer); i++)
+		{
+		GENERAL_NAME *gen = sk_GENERAL_NAME_value(rev->issuer, i);
+		if (gen->type != GEN_DIRNAME)
+			continue;
+		if (!X509_NAME_cmp(nm, gen->d.directoryName))
+			return 1;
+		}
+	return 0;
+
+	}
+
+static int def_crl_lookup(X509_CRL *crl,
+		X509_REVOKED **ret, ASN1_INTEGER *serial, X509_NAME *issuer)
+	{
+	X509_REVOKED rtmp, *rev;
+	int idx;
+	rtmp.serialNumber = serial;
+	/* Sort revoked into serial number order if not already sorted.
+	 * Do this under a lock to avoid race condition.
+ 	 */
+	if (!sk_X509_REVOKED_is_sorted(crl->crl->revoked))
+		{
+		CRYPTO_w_lock(CRYPTO_LOCK_X509_CRL);
+		sk_X509_REVOKED_sort(crl->crl->revoked);
+		CRYPTO_w_unlock(CRYPTO_LOCK_X509_CRL);
+		}
+	idx = sk_X509_REVOKED_find(crl->crl->revoked, &rtmp);
+	if(idx < 0)
+		return 0;
+	/* Need to look for matching name */
+	for(;idx < sk_X509_REVOKED_num(crl->crl->revoked); idx++)
+		{
+		rev = sk_X509_REVOKED_value(crl->crl->revoked, idx);
+		if (ASN1_INTEGER_cmp(rev->serialNumber, serial))
+			return 0;
+		if (crl_revoked_issuer_match(crl, issuer, rev))
+			{
+			if (ret)
+				*ret = rev;
+			if (rev->reason == CRL_REASON_REMOVE_FROM_CRL)
+				return 2;
+			return 1;
+			}
+		}
+	return 0;
+	}
+
+void X509_CRL_set_default_method(const X509_CRL_METHOD *meth)
+	{
+	if (meth == NULL)
+		default_crl_method = &int_crl_meth;
+	else 
+		default_crl_method = meth;
+	}
+
+X509_CRL_METHOD *X509_CRL_METHOD_new(
+	int (*crl_init)(X509_CRL *crl),
+	int (*crl_free)(X509_CRL *crl),
+	int (*crl_lookup)(X509_CRL *crl, X509_REVOKED **ret,
+				ASN1_INTEGER *ser, X509_NAME *issuer),
+	int (*crl_verify)(X509_CRL *crl, EVP_PKEY *pk))
+	{
+	X509_CRL_METHOD *m;
+	m = OPENSSL_malloc(sizeof(X509_CRL_METHOD));
+	if (!m)
+		return NULL;
+	m->crl_init = crl_init;
+	m->crl_free = crl_free;
+	m->crl_lookup = crl_lookup;
+	m->crl_verify = crl_verify;
+	m->flags = X509_CRL_METHOD_DYNAMIC;
+	return m;
+	}
+
+void X509_CRL_METHOD_free(X509_CRL_METHOD *m)
+	{
+	if (!(m->flags & X509_CRL_METHOD_DYNAMIC))
+		return;
+	OPENSSL_free(m);
+	}
+
+void X509_CRL_set_meth_data(X509_CRL *crl, void *dat)
+	{
+	crl->meth_data = dat;
+	}
+
+void *X509_CRL_get_meth_data(X509_CRL *crl)
+	{
+	return crl->meth_data;
+	}
+
 IMPLEMENT_STACK_OF(X509_REVOKED)
 IMPLEMENT_ASN1_SET_OF(X509_REVOKED)
 IMPLEMENT_STACK_OF(X509_CRL)
diff --git a/crypto/asn1/x_long.c b/crypto/asn1/x_long.c
index bf35457c1f74..75317418e1d8 100644
--- a/crypto/asn1/x_long.c
+++ b/crypto/asn1/x_long.c
@@ -71,6 +71,7 @@ static void long_free(ASN1_VALUE **pval, const ASN1_ITEM *it);
 
 static int long_i2c(ASN1_VALUE **pval, unsigned char *cont, int *putype, const ASN1_ITEM *it);
 static int long_c2i(ASN1_VALUE **pval, const unsigned char *cont, int len, int utype, char *free_cont, const ASN1_ITEM *it);
+static int long_print(BIO *out, ASN1_VALUE **pval, const ASN1_ITEM *it, int indent, const ASN1_PCTX *pctx);
 
 static ASN1_PRIMITIVE_FUNCS long_pf = {
 	NULL, 0,
@@ -78,7 +79,8 @@ static ASN1_PRIMITIVE_FUNCS long_pf = {
 	long_free,
 	long_free,	/* Clear should set to initial value */
 	long_c2i,
-	long_i2c
+	long_i2c,
+	long_print
 };
 
 ASN1_ITEM_start(LONG)
@@ -169,3 +171,9 @@ static int long_c2i(ASN1_VALUE **pval, const unsigned char *cont, int len,
 	memcpy(cp, &ltmp, sizeof(long));
 	return 1;
 }
+
+static int long_print(BIO *out, ASN1_VALUE **pval, const ASN1_ITEM *it,
+			int indent, const ASN1_PCTX *pctx)
+	{
+	return BIO_printf(out, "%ld\n", *(long *)pval);
+	}
diff --git a/crypto/asn1/x_name.c b/crypto/asn1/x_name.c
index 9a1a9f415ff3..d7c2318693f5 100644
--- a/crypto/asn1/x_name.c
+++ b/crypto/asn1/x_name.c
@@ -57,18 +57,36 @@
  */
 
 #include <stdio.h>
+#include <ctype.h>
 #include "cryptlib.h"
 #include <openssl/asn1t.h>
 #include <openssl/x509.h>
+#include "asn1_locl.h"
 
-static int x509_name_ex_d2i(ASN1_VALUE **val, const unsigned char **in, long len, const ASN1_ITEM *it,
-					int tag, int aclass, char opt, ASN1_TLC *ctx);
+typedef STACK_OF(X509_NAME_ENTRY) STACK_OF_X509_NAME_ENTRY;
+DECLARE_STACK_OF(STACK_OF_X509_NAME_ENTRY)
 
-static int x509_name_ex_i2d(ASN1_VALUE **val, unsigned char **out, const ASN1_ITEM *it, int tag, int aclass);
+static int x509_name_ex_d2i(ASN1_VALUE **val,
+				const unsigned char **in, long len,
+				const ASN1_ITEM *it,
+				int tag, int aclass, char opt, ASN1_TLC *ctx);
+
+static int x509_name_ex_i2d(ASN1_VALUE **val, unsigned char **out,
+				const ASN1_ITEM *it, int tag, int aclass);
 static int x509_name_ex_new(ASN1_VALUE **val, const ASN1_ITEM *it);
 static void x509_name_ex_free(ASN1_VALUE **val, const ASN1_ITEM *it);
 
 static int x509_name_encode(X509_NAME *a);
+static int x509_name_canon(X509_NAME *a);
+static int asn1_string_canon(ASN1_STRING *out, ASN1_STRING *in);
+static int i2d_name_canon(STACK_OF(STACK_OF_X509_NAME_ENTRY) *intname,
+			  unsigned char **in);
+
+
+static int x509_name_ex_print(BIO *out, ASN1_VALUE **pval,
+						int indent,
+						const char *fname, 
+						const ASN1_PCTX *pctx);
 
 ASN1_SEQUENCE(X509_NAME_ENTRY) = {
 	ASN1_SIMPLE(X509_NAME_ENTRY, object, ASN1_OBJECT),
@@ -102,7 +120,8 @@ const ASN1_EXTERN_FUNCS x509_name_ff = {
 	x509_name_ex_free,
 	0,	/* Default clear behaviour is OK */
 	x509_name_ex_d2i,
-	x509_name_ex_i2d
+	x509_name_ex_i2d,
+	x509_name_ex_print
 };
 
 IMPLEMENT_EXTERN_ASN1(X509_NAME, V_ASN1_SEQUENCE, x509_name_ff) 
@@ -118,6 +137,8 @@ static int x509_name_ex_new(ASN1_VALUE **val, const ASN1_ITEM *it)
 	if ((ret->entries=sk_X509_NAME_ENTRY_new_null()) == NULL)
 		goto memerr;
 	if((ret->bytes = BUF_MEM_new()) == NULL) goto memerr;
+	ret->canon_enc = NULL;
+	ret->canon_enclen = 0;
 	ret->modified=1;
 	*val = (ASN1_VALUE *)ret;
 	return 1;
@@ -142,25 +163,19 @@ static void x509_name_ex_free(ASN1_VALUE **pval, const ASN1_ITEM *it)
 
 	BUF_MEM_free(a->bytes);
 	sk_X509_NAME_ENTRY_pop_free(a->entries,X509_NAME_ENTRY_free);
+	if (a->canon_enc)
+		OPENSSL_free(a->canon_enc);
 	OPENSSL_free(a);
 	*pval = NULL;
 }
 
-/* Used with sk_pop_free() to free up the internal representation.
- * NB: we only free the STACK and not its contents because it is
- * already present in the X509_NAME structure.
- */
-
-static void sk_internal_free(void *a)
-{
-	sk_free(a);
-}
-
-static int x509_name_ex_d2i(ASN1_VALUE **val, const unsigned char **in, long len, const ASN1_ITEM *it,
-					int tag, int aclass, char opt, ASN1_TLC *ctx)
+static int x509_name_ex_d2i(ASN1_VALUE **val,
+			const unsigned char **in, long len, const ASN1_ITEM *it,
+				int tag, int aclass, char opt, ASN1_TLC *ctx)
 {
 	const unsigned char *p = *in, *q;
-	union { STACK *s; ASN1_VALUE *a; } intname = {NULL};
+	union { STACK_OF(STACK_OF_X509_NAME_ENTRY) *s;
+		ASN1_VALUE *a; } intname = {NULL};
 	union { X509_NAME *x; ASN1_VALUE *a; } nm = {NULL};
 	int i, j, ret;
 	STACK_OF(X509_NAME_ENTRY) *entries;
@@ -181,8 +196,8 @@ static int x509_name_ex_d2i(ASN1_VALUE **val, const unsigned char **in, long len
 	memcpy(nm.x->bytes->data, q, p - q);
 
 	/* Convert internal representation to X509_NAME structure */
-	for(i = 0; i < sk_num(intname.s); i++) {
-		entries = (STACK_OF(X509_NAME_ENTRY) *)sk_value(intname.s, i);
+	for(i = 0; i < sk_STACK_OF_X509_NAME_ENTRY_num(intname.s); i++) {
+		entries = sk_STACK_OF_X509_NAME_ENTRY_value(intname.s, i);
 		for(j = 0; j < sk_X509_NAME_ENTRY_num(entries); j++) {
 			entry = sk_X509_NAME_ENTRY_value(entries, j);
 			entry->set = i;
@@ -191,7 +206,10 @@ static int x509_name_ex_d2i(ASN1_VALUE **val, const unsigned char **in, long len
 		}
 		sk_X509_NAME_ENTRY_free(entries);
 	}
-	sk_free(intname.s);
+	sk_STACK_OF_X509_NAME_ENTRY_free(intname.s);
+	ret = x509_name_canon(nm.x);
+	if (!ret)
+		goto err;
 	nm.x->modified = 0;
 	*val = nm.a;
 	*in = p;
@@ -208,8 +226,12 @@ static int x509_name_ex_i2d(ASN1_VALUE **val, unsigned char **out, const ASN1_IT
 	int ret;
 	X509_NAME *a = (X509_NAME *)*val;
 	if(a->modified) {
-		ret = x509_name_encode((X509_NAME *)a);
-		if(ret < 0) return ret;
+		ret = x509_name_encode(a);
+		if(ret < 0)
+			return ret;
+		ret = x509_name_canon(a);
+		if(ret < 0)
+			return ret;
 	}
 	ret = a->bytes->length;
 	if(out != NULL) {
@@ -219,22 +241,35 @@ static int x509_name_ex_i2d(ASN1_VALUE **val, unsigned char **out, const ASN1_IT
 	return ret;
 }
 
+static void local_sk_X509_NAME_ENTRY_free(STACK_OF(X509_NAME_ENTRY) *ne)
+	{
+	sk_X509_NAME_ENTRY_free(ne);
+	}
+
+static void local_sk_X509_NAME_ENTRY_pop_free(STACK_OF(X509_NAME_ENTRY) *ne)
+	{
+	sk_X509_NAME_ENTRY_pop_free(ne, X509_NAME_ENTRY_free);
+	}
+
 static int x509_name_encode(X509_NAME *a)
 {
-	union { STACK *s; ASN1_VALUE *a; } intname = {NULL};
+	union { STACK_OF(STACK_OF_X509_NAME_ENTRY) *s;
+		ASN1_VALUE *a; } intname = {NULL};
 	int len;
 	unsigned char *p;
 	STACK_OF(X509_NAME_ENTRY) *entries = NULL;
 	X509_NAME_ENTRY *entry;
 	int i, set = -1;
-	intname.s = sk_new_null();
+	intname.s = sk_STACK_OF_X509_NAME_ENTRY_new_null();
 	if(!intname.s) goto memerr;
 	for(i = 0; i < sk_X509_NAME_ENTRY_num(a->entries); i++) {
 		entry = sk_X509_NAME_ENTRY_value(a->entries, i);
 		if(entry->set != set) {
 			entries = sk_X509_NAME_ENTRY_new_null();
 			if(!entries) goto memerr;
-			if(!sk_push(intname.s, (char *)entries)) goto memerr;
+			if(!sk_STACK_OF_X509_NAME_ENTRY_push(intname.s,
+							     entries))
+				goto memerr;
 			set = entry->set;
 		}
 		if(!sk_X509_NAME_ENTRY_push(entries, entry)) goto memerr;
@@ -245,15 +280,222 @@ static int x509_name_encode(X509_NAME *a)
 	p=(unsigned char *)a->bytes->data;
 	ASN1_item_ex_i2d(&intname.a,
 			 &p, ASN1_ITEM_rptr(X509_NAME_INTERNAL), -1, -1);
-	sk_pop_free(intname.s, sk_internal_free);
+	sk_STACK_OF_X509_NAME_ENTRY_pop_free(intname.s,
+					     local_sk_X509_NAME_ENTRY_free);
 	a->modified = 0;
 	return len;
-	memerr:
-	sk_pop_free(intname.s, sk_internal_free);
+memerr:
+	sk_STACK_OF_X509_NAME_ENTRY_pop_free(intname.s,
+					     local_sk_X509_NAME_ENTRY_free);
 	ASN1err(ASN1_F_X509_NAME_ENCODE, ERR_R_MALLOC_FAILURE);
 	return -1;
 }
 
+static int x509_name_ex_print(BIO *out, ASN1_VALUE **pval,
+						int indent,
+						const char *fname, 
+						const ASN1_PCTX *pctx)
+	{
+	if (X509_NAME_print_ex(out, (X509_NAME *)*pval,
+					indent, pctx->nm_flags) <= 0)
+		return 0;
+	return 2;
+	}
+
+/* This function generates the canonical encoding of the Name structure.
+ * In it all strings are converted to UTF8, leading, trailing and
+ * multiple spaces collapsed, converted to lower case and the leading
+ * SEQUENCE header removed.
+ *
+ * In future we could also normalize the UTF8 too.
+ *
+ * By doing this comparison of Name structures can be rapidly
+ * perfomed by just using memcmp() of the canonical encoding.
+ * By omitting the leading SEQUENCE name constraints of type
+ * dirName can also be checked with a simple memcmp().
+ */
+
+static int x509_name_canon(X509_NAME *a)
+	{
+	unsigned char *p;
+	STACK_OF(STACK_OF_X509_NAME_ENTRY) *intname = NULL;
+	STACK_OF(X509_NAME_ENTRY) *entries = NULL;
+	X509_NAME_ENTRY *entry, *tmpentry = NULL;
+	int i, set = -1, ret = 0;
+
+	if (a->canon_enc)
+		{
+		OPENSSL_free(a->canon_enc);
+		a->canon_enc = NULL;
+		}
+	/* Special case: empty X509_NAME => null encoding */
+	if (sk_X509_NAME_ENTRY_num(a->entries) == 0)
+		{
+		a->canon_enclen = 0;
+		return 1;
+		}
+	intname = sk_STACK_OF_X509_NAME_ENTRY_new_null();
+	if(!intname)
+		goto err;
+	for(i = 0; i < sk_X509_NAME_ENTRY_num(a->entries); i++)
+		{
+		entry = sk_X509_NAME_ENTRY_value(a->entries, i);
+		if(entry->set != set)
+			{
+			entries = sk_X509_NAME_ENTRY_new_null();
+			if(!entries)
+				goto err;
+			if(!sk_STACK_OF_X509_NAME_ENTRY_push(intname, entries))
+				goto err;
+			set = entry->set;
+			}
+		tmpentry = X509_NAME_ENTRY_new();
+		tmpentry->object = OBJ_dup(entry->object);
+		if (!asn1_string_canon(tmpentry->value, entry->value))
+			goto err;
+		if(!sk_X509_NAME_ENTRY_push(entries, tmpentry))
+			goto err;
+		tmpentry = NULL;
+		}
+
+	/* Finally generate encoding */
+
+	a->canon_enclen = i2d_name_canon(intname, NULL);
+
+	p = OPENSSL_malloc(a->canon_enclen);
+
+	if (!p)
+		goto err;
+
+	a->canon_enc = p;
+
+	i2d_name_canon(intname, &p);
+
+	ret = 1;
+
+	err:
+
+	if (tmpentry)
+		X509_NAME_ENTRY_free(tmpentry);
+	if (intname)
+		sk_STACK_OF_X509_NAME_ENTRY_pop_free(intname,
+					local_sk_X509_NAME_ENTRY_pop_free);
+	return ret;
+	}
+
+/* Bitmap of all the types of string that will be canonicalized. */
+
+#define ASN1_MASK_CANON	\
+	(B_ASN1_UTF8STRING | B_ASN1_BMPSTRING | B_ASN1_UNIVERSALSTRING \
+	| B_ASN1_PRINTABLESTRING | B_ASN1_T61STRING | B_ASN1_IA5STRING \
+	| B_ASN1_VISIBLESTRING)
+	
+
+static int asn1_string_canon(ASN1_STRING *out, ASN1_STRING *in)
+	{
+	unsigned char *to, *from;
+	int len, i;
+
+	/* If type not in bitmask just copy string across */
+	if (!(ASN1_tag2bit(in->type) & ASN1_MASK_CANON))
+		{
+		if (!ASN1_STRING_copy(out, in))
+			return 0;
+		return 1;
+		}
+
+	out->type = V_ASN1_UTF8STRING;
+	out->length = ASN1_STRING_to_UTF8(&out->data, in);
+	if (out->length == -1)
+		return 0;
+
+	to = out->data;
+	from = to;
+
+	len = out->length;
+
+	/* Convert string in place to canonical form.
+	 * Ultimately we may need to handle a wider range of characters
+	 * but for now ignore anything with MSB set and rely on the
+	 * isspace() and tolower() functions.
+	 */
+
+	/* Ignore leading spaces */
+	while((len > 0) && !(*from & 0x80) && isspace(*from))
+		{
+		from++;
+		len--;
+		}
+
+	to = from + len - 1;
+
+	/* Ignore trailing spaces */
+	while ((len > 0) && !(*to & 0x80) && isspace(*to))
+		{
+		to--;
+		len--;
+		}
+
+	to = out->data;
+
+	i = 0;
+	while(i < len)
+		{
+		/* If MSB set just copy across */
+		if (*from & 0x80)
+			{
+			*to++ = *from++;
+			i++;
+			}
+		/* Collapse multiple spaces */
+		else if (isspace(*from))
+			{
+			/* Copy one space across */
+			*to++ = ' ';
+			/* Ignore subsequent spaces. Note: don't need to
+			 * check len here because we know the last 
+			 * character is a non-space so we can't overflow.
+			 */
+			do
+				{
+				from++;
+				i++;
+				}
+			while(!(*from & 0x80) && isspace(*from));
+			}
+		else
+			{
+			*to++ = tolower(*from);
+			from++;
+			i++;
+			}
+		}
+
+	out->length = to - out->data;
+
+	return 1;
+
+	}
+
+static int i2d_name_canon(STACK_OF(STACK_OF_X509_NAME_ENTRY) *_intname,
+			  unsigned char **in)
+	{
+	int i, len, ltmp;
+	ASN1_VALUE *v;
+	STACK_OF(ASN1_VALUE) *intname = (STACK_OF(ASN1_VALUE) *)_intname;
+
+	len = 0;
+	for (i = 0; i < sk_ASN1_VALUE_num(intname); i++)
+		{
+		v = sk_ASN1_VALUE_value(intname, i);
+		ltmp = ASN1_item_ex_i2d(&v, in,
+			ASN1_ITEM_rptr(X509_NAME_ENTRIES), -1, -1);
+		if (ltmp < 0)
+			return ltmp;
+		len += ltmp;
+		}
+	return len;
+	}
 
 int X509_NAME_set(X509_NAME **xn, X509_NAME *name)
 	{
diff --git a/crypto/asn1/x_nx509.c b/crypto/asn1/x_nx509.c
new file mode 100644
index 000000000000..fbd9a22db345
--- /dev/null
+++ b/crypto/asn1/x_nx509.c
@@ -0,0 +1,72 @@
+/* x_nx509.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2005.
+ */
+/* ====================================================================
+ * Copyright (c) 2005 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stddef.h>
+#include <openssl/x509.h>
+#include <openssl/asn1.h>
+#include <openssl/asn1t.h>
+
+/* Old netscape certificate wrapper format */
+
+ASN1_SEQUENCE(NETSCAPE_X509) = {
+	ASN1_SIMPLE(NETSCAPE_X509, header, ASN1_OCTET_STRING),
+	ASN1_OPT(NETSCAPE_X509, cert, X509)
+} ASN1_SEQUENCE_END(NETSCAPE_X509)
+
+IMPLEMENT_ASN1_FUNCTIONS(NETSCAPE_X509)
+
diff --git a/crypto/asn1/x_pubkey.c b/crypto/asn1/x_pubkey.c
index 94d9f7ebab98..627ec87f9f89 100644
--- a/crypto/asn1/x_pubkey.c
+++ b/crypto/asn1/x_pubkey.c
@@ -60,6 +60,7 @@
 #include "cryptlib.h"
 #include <openssl/asn1t.h>
 #include <openssl/x509.h>
+#include "asn1_locl.h"
 #ifndef OPENSSL_NO_RSA
 #include <openssl/rsa.h>
 #endif
@@ -68,7 +69,8 @@
 #endif
 
 /* Minor tweak to operation: free up EVP_PKEY */
-static int pubkey_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int pubkey_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+			void *exarg)
 	{
 	if (operation == ASN1_OP_FREE_POST)
 		{
@@ -88,169 +90,42 @@ IMPLEMENT_ASN1_FUNCTIONS(X509_PUBKEY)
 int X509_PUBKEY_set(X509_PUBKEY **x, EVP_PKEY *pkey)
 	{
 	X509_PUBKEY *pk=NULL;
-	X509_ALGOR *a;
-	ASN1_OBJECT *o;
-	unsigned char *s,*p = NULL;
-	int i;
 
 	if (x == NULL) return(0);
 
-	if ((pk=X509_PUBKEY_new()) == NULL) goto err;
-	a=pk->algor;
+	if ((pk=X509_PUBKEY_new()) == NULL) goto error;
 
-	/* set the algorithm id */
-	if ((o=OBJ_nid2obj(pkey->type)) == NULL) goto err;
-	ASN1_OBJECT_free(a->algorithm);
-	a->algorithm=o;
-
-	/* Set the parameter list */
-	if (!pkey->save_parameters || (pkey->type == EVP_PKEY_RSA))
+	if (pkey->ameth)
 		{
-		if ((a->parameter == NULL) ||
-			(a->parameter->type != V_ASN1_NULL))
+		if (pkey->ameth->pub_encode)
 			{
-			ASN1_TYPE_free(a->parameter);
-			if (!(a->parameter=ASN1_TYPE_new()))
+			if (!pkey->ameth->pub_encode(pk, pkey))
 				{
-				X509err(X509_F_X509_PUBKEY_SET,ERR_R_MALLOC_FAILURE);
-				goto err;
+				X509err(X509_F_X509_PUBKEY_SET,
+					X509_R_PUBLIC_KEY_ENCODE_ERROR);
+				goto error;
 				}
-			a->parameter->type=V_ASN1_NULL;
-			}
-		}
-#ifndef OPENSSL_NO_DSA
-	else if (pkey->type == EVP_PKEY_DSA)
-		{
-		unsigned char *pp;
-		DSA *dsa;
-		
-		dsa=pkey->pkey.dsa;
-		dsa->write_params=0;
-		ASN1_TYPE_free(a->parameter);
-		if ((i=i2d_DSAparams(dsa,NULL)) <= 0)
-			goto err;
-		if (!(p=(unsigned char *)OPENSSL_malloc(i)))
-			{
-			X509err(X509_F_X509_PUBKEY_SET,ERR_R_MALLOC_FAILURE);
-			goto err;
-			}
-		pp=p;
-		i2d_DSAparams(dsa,&pp);
-		if (!(a->parameter=ASN1_TYPE_new()))
-			{
-			OPENSSL_free(p);
-			X509err(X509_F_X509_PUBKEY_SET,ERR_R_MALLOC_FAILURE);
-			goto err;
-			}
-		a->parameter->type=V_ASN1_SEQUENCE;
-		if (!(a->parameter->value.sequence=ASN1_STRING_new()))
-			{
-			OPENSSL_free(p);
-			X509err(X509_F_X509_PUBKEY_SET,ERR_R_MALLOC_FAILURE);
-			goto err;
-			}
-		if (!ASN1_STRING_set(a->parameter->value.sequence,p,i))
-			{
-			OPENSSL_free(p);
-			X509err(X509_F_X509_PUBKEY_SET,ERR_R_MALLOC_FAILURE);
-			goto err;
-			}
-		OPENSSL_free(p);
-		}
-#endif
-#ifndef OPENSSL_NO_EC
-	else if (pkey->type == EVP_PKEY_EC)
-		{
-		int nid=0;
-		unsigned char *pp;
-		EC_KEY *ec_key;
-		const EC_GROUP *group;
-		
-		ec_key = pkey->pkey.ec;
-		ASN1_TYPE_free(a->parameter);
-
-		if ((a->parameter = ASN1_TYPE_new()) == NULL)
-			{
-			X509err(X509_F_X509_PUBKEY_SET, ERR_R_ASN1_LIB);
-			goto err;
-			}
-
-		group = EC_KEY_get0_group(ec_key);
-		if (EC_GROUP_get_asn1_flag(group)
-                     && (nid = EC_GROUP_get_curve_name(group)))
-			{
-			/* just set the OID */
-			a->parameter->type = V_ASN1_OBJECT;
-			a->parameter->value.object = OBJ_nid2obj(nid);
 			}
-		else /* explicit parameters */
+		else
 			{
-			if ((i = i2d_ECParameters(ec_key, NULL)) == 0)
-				{
-				X509err(X509_F_X509_PUBKEY_SET, ERR_R_EC_LIB);
-				goto err;
-				}
-			if ((p = (unsigned char *) OPENSSL_malloc(i)) == NULL)
-				{
-				X509err(X509_F_X509_PUBKEY_SET, ERR_R_MALLOC_FAILURE);
-				goto err;
-				}	
-			pp = p;
-			if (!i2d_ECParameters(ec_key, &pp))
-				{
-				X509err(X509_F_X509_PUBKEY_SET, ERR_R_EC_LIB);
-				OPENSSL_free(p);
-				goto err;
-				}
-			a->parameter->type = V_ASN1_SEQUENCE;
-			if ((a->parameter->value.sequence = ASN1_STRING_new()) == NULL)
-				{
-				X509err(X509_F_X509_PUBKEY_SET, ERR_R_ASN1_LIB);
-				OPENSSL_free(p);
-				goto err;
-				}
-			ASN1_STRING_set(a->parameter->value.sequence, p, i);
-			OPENSSL_free(p);
+			X509err(X509_F_X509_PUBKEY_SET,
+				X509_R_METHOD_NOT_SUPPORTED);
+			goto error;
 			}
 		}
-#endif
-	else if (1)
+	else
 		{
 		X509err(X509_F_X509_PUBKEY_SET,X509_R_UNSUPPORTED_ALGORITHM);
-		goto err;
+		goto error;
 		}
 
-	if ((i=i2d_PublicKey(pkey,NULL)) <= 0) goto err;
-	if ((s=(unsigned char *)OPENSSL_malloc(i+1)) == NULL)
-		{
-		X509err(X509_F_X509_PUBKEY_SET,ERR_R_MALLOC_FAILURE);
-		goto err;
-		}
-	p=s;
-	i2d_PublicKey(pkey,&p);
-	if (!M_ASN1_BIT_STRING_set(pk->public_key,s,i))
-		{
-		X509err(X509_F_X509_PUBKEY_SET,ERR_R_MALLOC_FAILURE);
-		goto err;
-		}
-  	/* Set number of unused bits to zero */
-	pk->public_key->flags&= ~(ASN1_STRING_FLAG_BITS_LEFT|0x07);
-	pk->public_key->flags|=ASN1_STRING_FLAG_BITS_LEFT;
-
-	OPENSSL_free(s);
-
-#if 0
-	CRYPTO_add(&pkey->references,1,CRYPTO_LOCK_EVP_PKEY);
-	pk->pkey=pkey;
-#endif
-
 	if (*x != NULL)
 		X509_PUBKEY_free(*x);
 
 	*x=pk;
 
 	return 1;
-err:
+error:
 	if (pk != NULL) X509_PUBKEY_free(pk);
 	return 0;
 	}
@@ -258,113 +133,42 @@ err:
 EVP_PKEY *X509_PUBKEY_get(X509_PUBKEY *key)
 	{
 	EVP_PKEY *ret=NULL;
-	long j;
-	int type;
-	const unsigned char *p;
-#if !defined(OPENSSL_NO_DSA) || !defined(OPENSSL_NO_ECDSA)
-	const unsigned char *cp;
-	X509_ALGOR *a;
-#endif
 
-	if (key == NULL) goto err;
+	if (key == NULL) goto error;
 
 	if (key->pkey != NULL)
 		{
 		CRYPTO_add(&key->pkey->references, 1, CRYPTO_LOCK_EVP_PKEY);
-		return(key->pkey);
+		return key->pkey;
 		}
 
-	if (key->public_key == NULL) goto err;
+	if (key->public_key == NULL) goto error;
 
-	type=OBJ_obj2nid(key->algor->algorithm);
 	if ((ret = EVP_PKEY_new()) == NULL)
 		{
 		X509err(X509_F_X509_PUBKEY_GET, ERR_R_MALLOC_FAILURE);
-		goto err;
+		goto error;
 		}
-	ret->type = EVP_PKEY_type(type);
-
-	/* the parameters must be extracted before the public key (ECDSA!) */
-	
-#if !defined(OPENSSL_NO_DSA) || !defined(OPENSSL_NO_ECDSA)
-	a=key->algor;
-#endif
 
-	if (0)
-		;
-#ifndef OPENSSL_NO_DSA
-	else if (ret->type == EVP_PKEY_DSA)
+	if (!EVP_PKEY_set_type(ret, OBJ_obj2nid(key->algor->algorithm)))
 		{
-		if (a->parameter && (a->parameter->type == V_ASN1_SEQUENCE))
-			{
-			if ((ret->pkey.dsa = DSA_new()) == NULL)
-				{
-				X509err(X509_F_X509_PUBKEY_GET, ERR_R_MALLOC_FAILURE);
-				goto err;
-				}
-			ret->pkey.dsa->write_params=0;
-			cp=p=a->parameter->value.sequence->data;
-			j=a->parameter->value.sequence->length;
-			if (!d2i_DSAparams(&ret->pkey.dsa, &cp, (long)j))
-				goto err;
-			}
-		ret->save_parameters=1;
+		X509err(X509_F_X509_PUBKEY_GET,X509_R_UNSUPPORTED_ALGORITHM);
+		goto error;
 		}
-#endif
-#ifndef OPENSSL_NO_EC
-	else if (ret->type == EVP_PKEY_EC)
+
+	if (ret->ameth->pub_decode)
 		{
-		if (a->parameter && (a->parameter->type == V_ASN1_SEQUENCE))
+		if (!ret->ameth->pub_decode(ret, key))
 			{
-			/* type == V_ASN1_SEQUENCE => we have explicit parameters
-                         * (e.g. parameters in the X9_62_EC_PARAMETERS-structure )
-			 */
-			if ((ret->pkey.ec= EC_KEY_new()) == NULL)
-				{
-				X509err(X509_F_X509_PUBKEY_GET, 
-					ERR_R_MALLOC_FAILURE);
-				goto err;
-				}
-			cp = p = a->parameter->value.sequence->data;
-			j = a->parameter->value.sequence->length;
-			if (!d2i_ECParameters(&ret->pkey.ec, &cp, (long)j))
-				{
-				X509err(X509_F_X509_PUBKEY_GET, ERR_R_EC_LIB);
-				goto err;
-				}
+			X509err(X509_F_X509_PUBKEY_GET,
+						X509_R_PUBLIC_KEY_DECODE_ERROR);
+			goto error;
 			}
-		else if (a->parameter && (a->parameter->type == V_ASN1_OBJECT))
-			{
-			/* type == V_ASN1_OBJECT => the parameters are given
-			 * by an asn1 OID
-			 */
-			EC_KEY   *ec_key;
-			EC_GROUP *group;
-
-			if (ret->pkey.ec == NULL)
-				ret->pkey.ec = EC_KEY_new();
-			ec_key = ret->pkey.ec;
-			if (ec_key == NULL)
-				goto err;
-			group = EC_GROUP_new_by_curve_name(OBJ_obj2nid(a->parameter->value.object));
-			if (group == NULL)
-				goto err;
-			EC_GROUP_set_asn1_flag(group, OPENSSL_EC_NAMED_CURVE);
-			if (EC_KEY_set_group(ec_key, group) == 0)
-				goto err;
-			EC_GROUP_free(group);
-			}
-			/* the case implicitlyCA is currently not implemented */
-		ret->save_parameters = 1;
 		}
-#endif
-
-	p=key->public_key->data;
-        j=key->public_key->length;
-        if (!d2i_PublicKey(type, &ret, &p, (long)j))
+	else
 		{
-		X509err(X509_F_X509_PUBKEY_GET, X509_R_ERR_ASN1_LIB);
-		goto err;
+		X509err(X509_F_X509_PUBKEY_GET, X509_R_METHOD_NOT_SUPPORTED);
+		goto error;
 		}
 
 	/* Check to see if another thread set key->pkey first */
@@ -378,8 +182,10 @@ EVP_PKEY *X509_PUBKEY_get(X509_PUBKEY *key)
 		key->pkey = ret;
 	CRYPTO_w_unlock(CRYPTO_LOCK_EVP_PKEY);
 	CRYPTO_add(&ret->references, 1, CRYPTO_LOCK_EVP_PKEY);
-	return(ret);
-err:
+
+	return ret;
+
+	error:
 	if (ret != NULL)
 		EVP_PKEY_free(ret);
 	return(NULL);
@@ -538,3 +344,39 @@ int i2d_EC_PUBKEY(EC_KEY *a, unsigned char **pp)
 	return(ret);
 	}
 #endif
+
+int X509_PUBKEY_set0_param(X509_PUBKEY *pub, ASN1_OBJECT *aobj,
+					int ptype, void *pval,
+					unsigned char *penc, int penclen)
+	{
+	if (!X509_ALGOR_set0(pub->algor, aobj, ptype, pval))
+		return 0;
+	if (penc)
+		{
+		if (pub->public_key->data)
+			OPENSSL_free(pub->public_key->data);
+		pub->public_key->data = penc;
+		pub->public_key->length = penclen;
+  		/* Set number of unused bits to zero */
+		pub->public_key->flags&= ~(ASN1_STRING_FLAG_BITS_LEFT|0x07);
+		pub->public_key->flags|=ASN1_STRING_FLAG_BITS_LEFT;
+		}
+	return 1;
+	}
+
+int X509_PUBKEY_get0_param(ASN1_OBJECT **ppkalg,
+		const unsigned char **pk, int *ppklen,
+		X509_ALGOR **pa,
+		X509_PUBKEY *pub)
+	{
+	if (ppkalg)
+		*ppkalg = pub->algor->algorithm;
+	if (pk)
+		{
+		*pk = pub->public_key->data;
+		*ppklen = pub->public_key->length;
+		}
+	if (pa)
+		*pa = pub->algor;
+	return 1;
+	}
diff --git a/crypto/asn1/x_req.c b/crypto/asn1/x_req.c
index 59ca8ce32937..d57555827cca 100644
--- a/crypto/asn1/x_req.c
+++ b/crypto/asn1/x_req.c
@@ -79,7 +79,8 @@
  *
  */
 
-static int rinf_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int rinf_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+							void *exarg)
 {
 	X509_REQ_INFO *rinf = (X509_REQ_INFO *)*pval;
 
diff --git a/crypto/asn1/x_x509.c b/crypto/asn1/x_x509.c
index 088d5507480e..de3df9eb51c1 100644
--- a/crypto/asn1/x_x509.c
+++ b/crypto/asn1/x_x509.c
@@ -81,7 +81,8 @@ IMPLEMENT_ASN1_FUNCTIONS(X509_CINF)
 
 extern void policy_cache_free(X509_POLICY_CACHE *cache);
 
-static int x509_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int x509_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+								void *exarg)
 {
 	X509 *ret = (X509 *)*pval;
 
@@ -99,6 +100,7 @@ static int x509_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
 		ret->rfc3779_asid = NULL;
 #endif
 		ret->aux = NULL;
+		ret->crldp = NULL;
 		CRYPTO_new_ex_data(CRYPTO_EX_INDEX_X509, ret, &ret->ex_data);
 		break;
 
@@ -112,7 +114,10 @@ static int x509_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
 		X509_CERT_AUX_free(ret->aux);
 		ASN1_OCTET_STRING_free(ret->skid);
 		AUTHORITY_KEYID_free(ret->akid);
+		CRL_DIST_POINTS_free(ret->crldp);
 		policy_cache_free(ret->policy_cache);
+		GENERAL_NAMES_free(ret->altname);
+		NAME_CONSTRAINTS_free(ret->nc);
 #ifndef OPENSSL_NO_RFC3779
 		sk_IPAddressFamily_pop_free(ret->rfc3779_addr, IPAddressFamily_free);
 		ASIdentifiers_free(ret->rfc3779_asid);
@@ -136,19 +141,6 @@ ASN1_SEQUENCE_ref(X509, x509_cb, CRYPTO_LOCK_X509) = {
 IMPLEMENT_ASN1_FUNCTIONS(X509)
 IMPLEMENT_ASN1_DUP_FUNCTION(X509)
 
-static ASN1_METHOD meth=
-    {
-    (I2D_OF(void))  i2d_X509,
-    (D2I_OF(void)) d2i_X509,
-    (void *(*)(void))X509_new,
-    (void (*)(void *)) X509_free
-    };
-
-ASN1_METHOD *X509_asn1_meth(void)
-	{
-	return(&meth);
-	}
-
 int X509_get_ex_new_index(long argl, void *argp, CRYPTO_EX_new *new_func,
 	     CRYPTO_EX_dup *dup_func, CRYPTO_EX_free *free_func)
         {
diff --git a/crypto/bf/Makefile b/crypto/bf/Makefile
index 7f4f03eb821c..d01bfaa31558 100644
--- a/crypto/bf/Makefile
+++ b/crypto/bf/Makefile
@@ -12,8 +12,6 @@ MAKEFILE=	Makefile
 AR=		ar r
 
 BF_ENC=		bf_enc.o
-# or use
-#DES_ENC=	bx86-elf.o
 
 CFLAGS= $(INCLUDES) $(CFLAG)
 ASFLAGS= $(INCLUDES) $(ASFLAG)
@@ -40,19 +38,12 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-# ELF
-bx86-elf.s: asm/bf-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) bf-586.pl elf $(CFLAGS) $(PROCESSOR) > ../$@)
-# COFF
-bx86-cof.s: asm/bf-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) bf-586.pl coff $(CFLAGS) $(PROCESSOR) > ../$@)
-# a.out
-bx86-out.s: asm/bf-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) bf-586.pl a.out $(CFLAGS) $(PROCESSOR) > ../$@)
+bf-586.s:	asm/bf-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
+	$(PERL) asm/bf-586.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
@@ -104,8 +95,7 @@ bf_enc.o: ../../include/openssl/opensslconf.h bf_enc.c bf_locl.h
 bf_ofb64.o: ../../include/openssl/blowfish.h ../../include/openssl/e_os2.h
 bf_ofb64.o: ../../include/openssl/opensslconf.h bf_locl.h bf_ofb64.c
 bf_skey.o: ../../include/openssl/blowfish.h ../../include/openssl/crypto.h
-bf_skey.o: ../../include/openssl/e_os2.h ../../include/openssl/fips.h
-bf_skey.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-bf_skey.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-bf_skey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-bf_skey.o: bf_locl.h bf_pi.h bf_skey.c
+bf_skey.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
+bf_skey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+bf_skey.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+bf_skey.o: ../../include/openssl/symhacks.h bf_locl.h bf_pi.h bf_skey.c
diff --git a/crypto/bf/asm/bf-586.pl b/crypto/bf/asm/bf-586.pl
index b556642c949a..b74cfbafd4b0 100644
--- a/crypto/bf/asm/bf-586.pl
+++ b/crypto/bf/asm/bf-586.pl
@@ -1,6 +1,7 @@
 #!/usr/local/bin/perl
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 require "cbc.pl";
 
diff --git a/crypto/bf/bf_skey.c b/crypto/bf/bf_skey.c
index 6ac2aeb27959..3b0bca41aece 100644
--- a/crypto/bf/bf_skey.c
+++ b/crypto/bf/bf_skey.c
@@ -58,16 +58,19 @@
 
 #include <stdio.h>
 #include <string.h>
-#include <openssl/blowfish.h>
 #include <openssl/crypto.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/blowfish.h>
 #include "bf_locl.h"
 #include "bf_pi.h"
 
-FIPS_NON_FIPS_VCIPHER_Init(BF)
+void BF_set_key(BF_KEY *key, int len, const unsigned char *data)
+#ifdef OPENSSL_FIPS
+	{
+	fips_cipher_abort(BLOWFISH);
+	private_BF_set_key(key, len, data);
+	}
+void private_BF_set_key(BF_KEY *key, int len, const unsigned char *data)
+#endif
 	{
 	int i;
 	BF_LONG *p,ri,in[2];
diff --git a/crypto/bf/bfs.cpp b/crypto/bf/bfs.cpp
deleted file mode 100644
index d74c45776078..000000000000
--- a/crypto/bf/bfs.cpp
+++ /dev/null
@@ -1,67 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/blowfish.h>
-
-void main(int argc,char *argv[])
-	{
-	BF_KEY key;
-	unsigned long s1,s2,e1,e2;
-	unsigned long data[2];
-	int i,j;
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<1000; i++) /**/
-			{
-			BF_encrypt(&data[0],&key);
-			GetTSC(s1);
-			BF_encrypt(&data[0],&key);
-			BF_encrypt(&data[0],&key);
-			BF_encrypt(&data[0],&key);
-			GetTSC(e1);
-			GetTSC(s2);
-			BF_encrypt(&data[0],&key);
-			BF_encrypt(&data[0],&key);
-			BF_encrypt(&data[0],&key);
-			BF_encrypt(&data[0],&key);
-			GetTSC(e2);
-			BF_encrypt(&data[0],&key);
-			}
-
-		printf("blowfish %d %d (%d)\n",
-			e1-s1,e2-s2,((e2-s2)-(e1-s1)));
-		}
-	}
-
diff --git a/crypto/bf/blowfish.h b/crypto/bf/blowfish.h
index d24ffccb65f6..4b6c8920a4ad 100644
--- a/crypto/bf/blowfish.h
+++ b/crypto/bf/blowfish.h
@@ -79,7 +79,7 @@ extern "C" {
  * !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  */
 
-#if defined(OPENSSL_SYS_WIN16) || defined(__LP32__)
+#if defined(__LP32__)
 #define BF_LONG unsigned long
 #elif defined(OPENSSL_SYS_CRAY) || defined(__ILP64__)
 #define BF_LONG unsigned long
diff --git a/crypto/bio/Makefile b/crypto/bio/Makefile
index 1cd76ce7a2fb..c395d8049681 100644
--- a/crypto/bio/Makefile
+++ b/crypto/bio/Makefile
@@ -45,7 +45,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -102,11 +102,12 @@ b_print.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 b_print.o: ../../include/openssl/symhacks.h ../cryptlib.h b_print.c
 b_sock.o: ../../e_os.h ../../include/openssl/bio.h
 b_sock.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-b_sock.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-b_sock.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
-b_sock.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-b_sock.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-b_sock.o: ../../include/openssl/symhacks.h ../cryptlib.h b_sock.c
+b_sock.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
+b_sock.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
+b_sock.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+b_sock.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+b_sock.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+b_sock.o: ../cryptlib.h b_sock.c
 bf_buff.o: ../../e_os.h ../../include/openssl/bio.h
 bf_buff.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 bf_buff.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
diff --git a/crypto/bio/b_print.c b/crypto/bio/b_print.c
index 3a87b0ec0b2e..143a7cfefa37 100644
--- a/crypto/bio/b_print.c
+++ b/crypto/bio/b_print.c
@@ -115,8 +115,8 @@
 #define LDOUBLE double
 #endif
 
-#if HAVE_LONG_LONG
-# if defined(OPENSSL_SYS_WIN32) && !defined(__GNUC__)
+#ifdef HAVE_LONG_LONG
+# if defined(_WIN32) && !defined(__GNUC__)
 # define LLONG __int64
 # else
 # define LLONG long long
diff --git a/crypto/bio/b_sock.c b/crypto/bio/b_sock.c
index f0d9cfca67cc..41f958be7106 100644
--- a/crypto/bio/b_sock.c
+++ b/crypto/bio/b_sock.c
@@ -72,11 +72,9 @@ NETDB_DEFINE_CONTEXT
 
 #ifndef OPENSSL_NO_SOCK
 
-#ifdef OPENSSL_SYS_WIN16
-#define SOCKET_PROTOCOL 0 /* more microsoft stupidity */
-#else
+#include <openssl/dso.h>
+
 #define SOCKET_PROTOCOL IPPROTO_TCP
-#endif
 
 #ifdef SO_MAXCONN
 #define MAX_LISTEN  SO_MAXCONN
@@ -90,6 +88,17 @@ NETDB_DEFINE_CONTEXT
 static int wsa_init_done=0;
 #endif
 
+/*
+ * WSAAPI specifier is required to make indirect calls to run-time
+ * linked WinSock 2 functions used in this module, to be specific
+ * [get|free]addrinfo and getnameinfo. This is because WinSock uses
+ * uses non-C calling convention, __stdcall vs. __cdecl, on x86
+ * Windows. On non-WinSock platforms WSAAPI needs to be void.
+ */
+#ifndef WSAAPI
+#define WSAAPI
+#endif
+
 #if 0
 static unsigned long BIO_ghbn_hits=0L;
 static unsigned long BIO_ghbn_miss=0L;
@@ -226,6 +235,10 @@ int BIO_sock_error(int sock)
 	int j,i;
 	int size;
 		 
+#if defined(OPENSSL_SYS_BEOS_R5)
+	return 0;
+#endif
+		 
 	size=sizeof(int);
 	/* Note: under Windows the third parameter is of type (char *)
 	 * whereas under other systems it is (void *) if you don't have
@@ -466,7 +479,12 @@ int BIO_sock_init(void)
 	  
 		wsa_init_done=1;
 		memset(&wsa_state,0,sizeof(wsa_state));
-		if (WSAStartup(0x0101,&wsa_state)!=0)
+		/* Not making wsa_state available to the rest of the
+		 * code is formally wrong. But the structures we use
+		 * are [beleived to be] invariable among Winsock DLLs,
+		 * while API availability is [expected to be] probed
+		 * at run-time with DSO_global_lookup. */
+		if (WSAStartup(0x0202,&wsa_state)!=0)
 			{
 			err=WSAGetLastError();
 			SYSerr(SYS_F_WSASTARTUP,err);
@@ -510,8 +528,8 @@ void BIO_sock_cleanup(void)
 	if (wsa_init_done)
 		{
 		wsa_init_done=0;
-#ifndef OPENSSL_SYS_WINCE
-		WSACancelBlockingCall();	/* Winsock 1.1 specific */
+#if 0		/* this call is claimed to be non-present in Winsock2 */
+		WSACancelBlockingCall();
 #endif
 		WSACleanup();
 		}
@@ -533,7 +551,30 @@ int BIO_socket_ioctl(int fd, long type, void *arg)
 #ifdef __DJGPP__
 	i=ioctlsocket(fd,type,(char *)arg);
 #else
-	i=ioctlsocket(fd,type,arg);
+# if defined(OPENSSL_SYS_VMS)
+	/* 2011-02-18 SMS.
+	 * VMS ioctl() can't tolerate a 64-bit "void *arg", but we
+	 * observe that all the consumers pass in an "unsigned long *",
+	 * so we arrange a local copy with a short pointer, and use
+	 * that, instead.
+	 */
+#  if __INITIAL_POINTER_SIZE == 64
+#   define ARG arg_32p
+#   pragma pointer_size save
+#   pragma pointer_size 32
+	unsigned long arg_32;
+	unsigned long *arg_32p;
+#   pragma pointer_size restore
+	arg_32p = &arg_32;
+	arg_32 = *((unsigned long *) arg);
+#  else /* __INITIAL_POINTER_SIZE == 64 */
+#   define ARG arg
+#  endif /* __INITIAL_POINTER_SIZE == 64 [else] */
+# else /* defined(OPENSSL_SYS_VMS) */
+#  define ARG arg
+# endif /* defined(OPENSSL_SYS_VMS) [else] */
+
+	i=ioctlsocket(fd,type,ARG);
 #endif /* __DJGPP__ */
 	if (i < 0)
 		SYSerr(SYS_F_IOCTLSOCKET,get_last_socket_error());
@@ -581,12 +622,18 @@ static int get_ip(const char *str, unsigned char ip[4])
 int BIO_get_accept_socket(char *host, int bind_mode)
 	{
 	int ret=0;
-	struct sockaddr_in server,client;
-	int s=INVALID_SOCKET,cs;
+	union {
+		struct sockaddr sa;
+		struct sockaddr_in sa_in;
+#if OPENSSL_USE_IPV6
+		struct sockaddr_in6 sa_in6;
+#endif
+	} server,client;
+	int s=INVALID_SOCKET,cs,addrlen;
 	unsigned char ip[4];
 	unsigned short port;
 	char *str=NULL,*e;
-	const char *h,*p;
+	char *h,*p;
 	unsigned long l;
 	int err_num;
 
@@ -600,8 +647,7 @@ int BIO_get_accept_socket(char *host, int bind_mode)
 		{
 		if (*e == ':')
 			{
-			p= &(e[1]);
-			*e='\0';
+			p=e;
 			}
 		else if (*e == '/')
 			{
@@ -609,21 +655,74 @@ int BIO_get_accept_socket(char *host, int bind_mode)
 			break;
 			}
 		}
-
-	if (p == NULL)
+	if (p)	*p++='\0';	/* points at last ':', '::port' is special [see below] */
+	else	p=h,h=NULL;
+
+#ifdef EAI_FAMILY
+	do {
+	static union {	void *p;
+			int (WSAAPI *f)(const char *,const char *,
+				 const struct addrinfo *,
+				 struct addrinfo **);
+			} p_getaddrinfo = {NULL};
+	static union {	void *p;
+			void (WSAAPI *f)(struct addrinfo *);
+			} p_freeaddrinfo = {NULL};
+	struct addrinfo *res,hint;
+
+	if (p_getaddrinfo.p==NULL)
+		{
+		if ((p_getaddrinfo.p=DSO_global_lookup("getaddrinfo"))==NULL ||
+		    (p_freeaddrinfo.p=DSO_global_lookup("freeaddrinfo"))==NULL)
+			p_getaddrinfo.p=(void*)-1;
+		}
+	if (p_getaddrinfo.p==(void *)-1) break;
+
+	/* '::port' enforces IPv6 wildcard listener. Some OSes,
+	 * e.g. Solaris, default to IPv6 without any hint. Also
+	 * note that commonly IPv6 wildchard socket can service
+	 * IPv4 connections just as well...  */
+	memset(&hint,0,sizeof(hint));
+	hint.ai_flags = AI_PASSIVE;
+	if (h)
 		{
-		p=h;
-		h="*";
+		if (strchr(h,':'))
+			{
+			if (h[1]=='\0') h=NULL;
+#if OPENSSL_USE_IPV6
+			hint.ai_family = AF_INET6;
+#else
+			h=NULL;
+#endif
+			}
+	    	else if (h[0]=='*' && h[1]=='\0')
+			{
+			hint.ai_family = AF_INET;
+			h=NULL;
+			}
 		}
 
+	if ((*p_getaddrinfo.f)(h,p,&hint,&res)) break;
+
+	addrlen = res->ai_addrlen<=sizeof(server) ?
+			res->ai_addrlen :
+			sizeof(server);
+	memcpy(&server, res->ai_addr, addrlen);
+
+	(*p_freeaddrinfo.f)(res);
+	goto again;
+	} while (0);
+#endif
+
 	if (!BIO_get_port(p,&port)) goto err;
 
 	memset((char *)&server,0,sizeof(server));
-	server.sin_family=AF_INET;
-	server.sin_port=htons(port);
+	server.sa_in.sin_family=AF_INET;
+	server.sa_in.sin_port=htons(port);
+	addrlen = sizeof(server.sa_in);
 
-	if (strcmp(h,"*") == 0)
-		server.sin_addr.s_addr=INADDR_ANY;
+	if (h == NULL || strcmp(h,"*") == 0)
+		server.sa_in.sin_addr.s_addr=INADDR_ANY;
 	else
 		{
                 if (!BIO_get_host_ip(h,&(ip[0]))) goto err;
@@ -632,11 +731,11 @@ int BIO_get_accept_socket(char *host, int bind_mode)
 			((unsigned long)ip[1]<<16L)|
 			((unsigned long)ip[2]<< 8L)|
 			((unsigned long)ip[3]);
-		server.sin_addr.s_addr=htonl(l);
+		server.sa_in.sin_addr.s_addr=htonl(l);
 		}
 
 again:
-	s=socket(AF_INET,SOCK_STREAM,SOCKET_PROTOCOL);
+	s=socket(server.sa.sa_family,SOCK_STREAM,SOCKET_PROTOCOL);
 	if (s == INVALID_SOCKET)
 		{
 		SYSerr(SYS_F_SOCKET,get_last_socket_error());
@@ -654,7 +753,7 @@ again:
 		bind_mode=BIO_BIND_NORMAL;
 		}
 #endif
-	if (bind(s,(struct sockaddr *)&server,sizeof(server)) == -1)
+	if (bind(s,&server.sa,addrlen) == -1)
 		{
 #ifdef SO_REUSEADDR
 		err_num=get_last_socket_error();
@@ -668,15 +767,28 @@ again:
 			(err_num == EADDRINUSE))
 #endif
 			{
-			memcpy((char *)&client,(char *)&server,sizeof(server));
-			if (strcmp(h,"*") == 0)
-				client.sin_addr.s_addr=htonl(0x7F000001);
-			cs=socket(AF_INET,SOCK_STREAM,SOCKET_PROTOCOL);
+			client = server;
+			if (h == NULL || strcmp(h,"*") == 0)
+				{
+#if OPENSSL_USE_IPV6
+				if (client.sa.sa_family == AF_INET6)
+					{
+					memset(&client.sa_in6.sin6_addr,0,sizeof(client.sa_in6.sin6_addr));
+					client.sa_in6.sin6_addr.s6_addr[15]=1;
+					}
+				else
+#endif
+				if (client.sa.sa_family == AF_INET)
+					{
+					client.sa_in.sin_addr.s_addr=htonl(0x7F000001);
+					}
+				else	goto err;
+				}
+			cs=socket(client.sa.sa_family,SOCK_STREAM,SOCKET_PROTOCOL);
 			if (cs != INVALID_SOCKET)
 				{
 				int ii;
-				ii=connect(cs,(struct sockaddr *)&client,
-					sizeof(client));
+				ii=connect(cs,&client.sa,addrlen);
 				closesocket(cs);
 				if (ii == INVALID_SOCKET)
 					{
@@ -715,20 +827,52 @@ err:
 int BIO_accept(int sock, char **addr)
 	{
 	int ret=INVALID_SOCKET;
-	static struct sockaddr_in from;
 	unsigned long l;
 	unsigned short port;
-	int len;
 	char *p;
 
-	memset((char *)&from,0,sizeof(from));
-	len=sizeof(from);
-	/* Note: under VMS with SOCKETSHR the fourth parameter is currently
-	 * of type (int *) whereas under other systems it is (void *) if
-	 * you don't have a cast it will choke the compiler: if you do
-	 * have a cast then you can either go for (int *) or (void *).
+	struct {
+	/*
+	 * As for following union. Trouble is that there are platforms
+	 * that have socklen_t and there are platforms that don't, on
+	 * some platforms socklen_t is int and on some size_t. So what
+	 * one can do? One can cook #ifdef spaghetti, which is nothing
+	 * but masochistic. Or one can do union between int and size_t.
+	 * One naturally does it primarily for 64-bit platforms where
+	 * sizeof(int) != sizeof(size_t). But would it work? Note that
+	 * if size_t member is initialized to 0, then later int member
+	 * assignment naturally does the job on little-endian platforms
+	 * regardless accept's expectations! What about big-endians?
+	 * If accept expects int*, then it works, and if size_t*, then
+	 * length value would appear as unreasonably large. But this
+	 * won't prevent it from filling in the address structure. The
+	 * trouble of course would be if accept returns more data than
+	 * actual buffer can accomodate and overwrite stack... That's
+	 * where early OPENSSL_assert comes into picture. Besides, the
+	 * only 64-bit big-endian platform found so far that expects
+	 * size_t* is HP-UX, where stack grows towards higher address.
+	 * <appro>
 	 */
-	ret=accept(sock,(struct sockaddr *)&from,(void *)&len);
+	union { size_t s; int i; } len;
+	union {
+		struct sockaddr sa;
+		struct sockaddr_in sa_in;
+#if OPENSSL_USE_IPV6
+		struct sockaddr_in6 sa_in6;
+#endif
+		} from;
+	} sa;
+
+	sa.len.s=0;
+	sa.len.i=sizeof(sa.from);
+	memset(&sa.from,0,sizeof(sa.from));
+	ret=accept(sock,&sa.from.sa,(void *)&sa.len);
+	if (sizeof(sa.len.i)!=sizeof(sa.len.s) && sa.len.i==0)
+		{
+		OPENSSL_assert(sa.len.s<=sizeof(sa.from));
+		sa.len.i = (int)sa.len.s;
+		/* use sa.len.i from this point */
+		}
 	if (ret == INVALID_SOCKET)
 		{
 		if(BIO_sock_should_retry(ret)) return -2;
@@ -739,8 +883,46 @@ int BIO_accept(int sock, char **addr)
 
 	if (addr == NULL) goto end;
 
-	l=ntohl(from.sin_addr.s_addr);
-	port=ntohs(from.sin_port);
+#ifdef EAI_FAMILY
+	do {
+	char   h[NI_MAXHOST],s[NI_MAXSERV];
+	size_t nl;
+	static union {	void *p;
+			int (WSAAPI *f)(const struct sockaddr *,size_t/*socklen_t*/,
+				 char *,size_t,char *,size_t,int);
+			} p_getnameinfo = {NULL};
+			/* 2nd argument to getnameinfo is specified to
+			 * be socklen_t. Unfortunately there is a number
+			 * of environments where socklen_t is not defined.
+			 * As it's passed by value, it's safe to pass it
+			 * as size_t... <appro> */
+
+	if (p_getnameinfo.p==NULL)
+		{
+		if ((p_getnameinfo.p=DSO_global_lookup("getnameinfo"))==NULL)
+			p_getnameinfo.p=(void*)-1;
+		}
+	if (p_getnameinfo.p==(void *)-1) break;
+
+	if ((*p_getnameinfo.f)(&sa.from.sa,sa.len.i,h,sizeof(h),s,sizeof(s),
+	    NI_NUMERICHOST|NI_NUMERICSERV)) break;
+	nl = strlen(h)+strlen(s)+2;
+	p = *addr;
+	if (p)	{ *p = '\0'; p = OPENSSL_realloc(p,nl);	}
+	else	{ p = OPENSSL_malloc(nl);		}
+	if (p==NULL)
+		{
+		BIOerr(BIO_F_BIO_ACCEPT,ERR_R_MALLOC_FAILURE);
+		goto end;
+		}
+	*addr = p;
+	BIO_snprintf(*addr,nl,"%s:%s",h,s);
+	goto end;
+	} while(0);
+#endif
+	if (sa.from.sa.sa_family != AF_INET) goto end;
+	l=ntohl(sa.from.sa_in.sin_addr.s_addr);
+	port=ntohs(sa.from.sa_in.sin_port);
 	if (*addr == NULL)
 		{
 		if ((p=OPENSSL_malloc(24)) == NULL)
@@ -778,7 +960,6 @@ int BIO_set_tcp_ndelay(int s, int on)
 #endif
 	return(ret == 0);
 	}
-#endif
 
 int BIO_socket_nbio(int s, int mode)
 	{
@@ -791,3 +972,4 @@ int BIO_socket_nbio(int s, int mode)
 #endif
 	return(ret == 0);
 	}
+#endif
diff --git a/crypto/bio/bio.h b/crypto/bio/bio.h
index 03bd3b28758f..05699ab21210 100644
--- a/crypto/bio/bio.h
+++ b/crypto/bio/bio.h
@@ -68,6 +68,14 @@
 
 #include <openssl/crypto.h>
 
+#ifndef OPENSSL_NO_SCTP
+# ifndef OPENSSL_SYS_VMS
+# include <stdint.h>
+# else
+# include <inttypes.h>
+# endif
+#endif
+
 #ifdef  __cplusplus
 extern "C" {
 #endif
@@ -95,6 +103,10 @@ extern "C" {
 #define BIO_TYPE_BIO		(19|0x0400)		/* (half a) BIO pair */
 #define BIO_TYPE_LINEBUFFER	(20|0x0200)		/* filter */
 #define BIO_TYPE_DGRAM		(21|0x0400|0x0100)
+#ifndef OPENSSL_NO_SCTP
+#define BIO_TYPE_DGRAM_SCTP	(24|0x0400|0x0100)
+#endif
+#define BIO_TYPE_ASN1 		(22|0x0200)		/* filter */
 #define BIO_TYPE_COMP 		(23|0x0200)		/* filter */
 
 #define BIO_TYPE_DESCRIPTOR	0x0100	/* socket, fd, connect or accept */
@@ -161,7 +173,22 @@ extern "C" {
 #define BIO_CTRL_DGRAM_SET_PEER           44 /* Destination for the data */
 
 #define BIO_CTRL_DGRAM_SET_NEXT_TIMEOUT   45 /* Next DTLS handshake timeout to
-											  * adjust socket timeouts */
+                                              * adjust socket timeouts */
+
+#ifndef OPENSSL_NO_SCTP
+/* SCTP stuff */
+#define BIO_CTRL_DGRAM_SCTP_SET_IN_HANDSHAKE	50
+#define BIO_CTRL_DGRAM_SCTP_ADD_AUTH_KEY		51
+#define BIO_CTRL_DGRAM_SCTP_NEXT_AUTH_KEY		52
+#define BIO_CTRL_DGRAM_SCTP_AUTH_CCS_RCVD		53
+#define BIO_CTRL_DGRAM_SCTP_GET_SNDINFO		60
+#define BIO_CTRL_DGRAM_SCTP_SET_SNDINFO		61
+#define BIO_CTRL_DGRAM_SCTP_GET_RCVINFO		62
+#define BIO_CTRL_DGRAM_SCTP_SET_RCVINFO		63
+#define BIO_CTRL_DGRAM_SCTP_GET_PRINFO			64
+#define BIO_CTRL_DGRAM_SCTP_SET_PRINFO			65
+#define BIO_CTRL_DGRAM_SCTP_SAVE_SHUTDOWN		70
+#endif
 
 /* modifiers */
 #define BIO_FP_READ		0x02
@@ -266,7 +293,6 @@ int BIO_method_type(const BIO *b);
 
 typedef void bio_info_cb(struct bio_st *, int, const char *, int, long, long);
 
-#ifndef OPENSSL_SYS_WIN16
 typedef struct bio_method_st
 	{
 	int type;
@@ -280,21 +306,6 @@ typedef struct bio_method_st
 	int (*destroy)(BIO *);
         long (*callback_ctrl)(BIO *, int, bio_info_cb *);
 	} BIO_METHOD;
-#else
-typedef struct bio_method_st
-	{
-	int type;
-	const char *name;
-	int (_far *bwrite)();
-	int (_far *bread)();
-	int (_far *bputs)();
-	int (_far *bgets)();
-	long (_far *ctrl)();
-	int (_far *create)();
-	int (_far *destroy)();
-	long (_far *callback_ctrl)();
-	} BIO_METHOD;
-#endif
 
 struct bio_st
 	{
@@ -344,6 +355,37 @@ typedef struct bio_f_buffer_ctx_struct
 	int obuf_off;		/* write/read offset */
 	} BIO_F_BUFFER_CTX;
 
+/* Prefix and suffix callback in ASN1 BIO */
+typedef int asn1_ps_func(BIO *b, unsigned char **pbuf, int *plen, void *parg);
+
+#ifndef OPENSSL_NO_SCTP
+/* SCTP parameter structs */
+struct bio_dgram_sctp_sndinfo
+	{
+	uint16_t snd_sid;
+	uint16_t snd_flags;
+	uint32_t snd_ppid;
+	uint32_t snd_context;
+	};
+
+struct bio_dgram_sctp_rcvinfo
+	{
+	uint16_t rcv_sid;
+	uint16_t rcv_ssn;
+	uint16_t rcv_flags;
+	uint32_t rcv_ppid;
+	uint32_t rcv_tsn;
+	uint32_t rcv_cumtsn;
+	uint32_t rcv_context;
+	};
+
+struct bio_dgram_sctp_prinfo
+	{
+	uint16_t pr_policy;
+	uint32_t pr_value;
+	};
+#endif
+
 /* connect BIO stuff */
 #define BIO_CONN_S_BEFORE		1
 #define BIO_CONN_S_GET_IP		2
@@ -406,6 +448,13 @@ typedef struct bio_f_buffer_ctx_struct
 #define BIO_C_RESET_READ_REQUEST		147
 #define BIO_C_SET_MD_CTX			148
 
+#define BIO_C_SET_PREFIX			149
+#define BIO_C_GET_PREFIX			150
+#define BIO_C_SET_SUFFIX			151
+#define BIO_C_GET_SUFFIX			152
+
+#define BIO_C_SET_EX_ARG			153
+#define BIO_C_GET_EX_ARG			154
 
 #define BIO_set_app_data(s,arg)		BIO_set_ex_data(s,0,arg)
 #define BIO_get_app_data(s)		BIO_get_ex_data(s,0)
@@ -569,22 +618,21 @@ int BIO_get_ex_new_index(long argl, void *argp, CRYPTO_EX_new *new_func,
 unsigned long BIO_number_read(BIO *bio);
 unsigned long BIO_number_written(BIO *bio);
 
+/* For BIO_f_asn1() */
+int BIO_asn1_set_prefix(BIO *b, asn1_ps_func *prefix,
+					asn1_ps_func *prefix_free);
+int BIO_asn1_get_prefix(BIO *b, asn1_ps_func **pprefix,
+					asn1_ps_func **pprefix_free);
+int BIO_asn1_set_suffix(BIO *b, asn1_ps_func *suffix,
+					asn1_ps_func *suffix_free);
+int BIO_asn1_get_suffix(BIO *b, asn1_ps_func **psuffix,
+					asn1_ps_func **psuffix_free);
+
 # ifndef OPENSSL_NO_FP_API
-#  if defined(OPENSSL_SYS_WIN16) && defined(_WINDLL)
-BIO_METHOD *BIO_s_file_internal(void);
-BIO *BIO_new_file_internal(char *filename, char *mode);
-BIO *BIO_new_fp_internal(FILE *stream, int close_flag);
-#    define BIO_s_file	BIO_s_file_internal
-#    define BIO_new_file	BIO_new_file_internal
-#    define BIO_new_fp	BIO_new_fp_internal
-#  else /* FP_API */
 BIO_METHOD *BIO_s_file(void );
 BIO *BIO_new_file(const char *filename, const char *mode);
 BIO *BIO_new_fp(FILE *stream, int close_flag);
-#    define BIO_s_file_internal		BIO_s_file
-#    define BIO_new_file_internal	BIO_new_file
-#    define BIO_new_fp_internal		BIO_s_file
-#  endif /* FP_API */
+# define BIO_s_file_internal	BIO_s_file
 # endif
 BIO *	BIO_new(BIO_METHOD *type);
 int	BIO_set(BIO *a,BIO_METHOD *type);
@@ -613,13 +661,8 @@ int BIO_nread(BIO *bio, char **buf, int num);
 int BIO_nwrite0(BIO *bio, char **buf);
 int BIO_nwrite(BIO *bio, char **buf, int num);
 
-#ifndef OPENSSL_SYS_WIN16
 long BIO_debug_callback(BIO *bio,int cmd,const char *argp,int argi,
 	long argl,long ret);
-#else
-long _far _loadds BIO_debug_callback(BIO *bio,int cmd,const char *argp,int argi,
-	long argl,long ret);
-#endif
 
 BIO_METHOD *BIO_s_mem(void);
 BIO *BIO_new_mem_buf(void *buf, int len);
@@ -640,6 +683,9 @@ BIO_METHOD *BIO_f_linebuffer(void);
 BIO_METHOD *BIO_f_nbio_test(void);
 #ifndef OPENSSL_NO_DGRAM
 BIO_METHOD *BIO_s_datagram(void);
+#ifndef OPENSSL_NO_SCTP
+BIO_METHOD *BIO_s_datagram_sctp(void);
+#endif
 #endif
 
 /* BIO_METHOD *BIO_f_ber(void); */
@@ -682,6 +728,15 @@ int BIO_set_tcp_ndelay(int sock,int turn_on);
 
 BIO *BIO_new_socket(int sock, int close_flag);
 BIO *BIO_new_dgram(int fd, int close_flag);
+#ifndef OPENSSL_NO_SCTP
+BIO *BIO_new_dgram_sctp(int fd, int close_flag);
+int BIO_dgram_is_sctp(BIO *bio);
+int BIO_dgram_sctp_notification_cb(BIO *b,
+                                   void (*handle_notifications)(BIO *bio, void *context, void *buf),
+                                   void *context);
+int BIO_dgram_sctp_wait_for_dry(BIO *b);
+int BIO_dgram_sctp_msg_waiting(BIO *b);
+#endif
 BIO *BIO_new_fd(int fd, int close_flag);
 BIO *BIO_new_connect(char *host_port);
 BIO *BIO_new_accept(char *host_port);
@@ -746,6 +801,7 @@ void ERR_load_BIO_strings(void);
 #define BIO_F_BUFFER_CTRL				 114
 #define BIO_F_CONN_CTRL					 127
 #define BIO_F_CONN_STATE				 115
+#define BIO_F_DGRAM_SCTP_READ				 132
 #define BIO_F_FILE_CTRL					 116
 #define BIO_F_FILE_READ					 130
 #define BIO_F_LINEBUFFER_CTRL				 129
diff --git a/crypto/bio/bio_cb.c b/crypto/bio/bio_cb.c
index 6f4254a1141a..9bcbc321d94f 100644
--- a/crypto/bio/bio_cb.c
+++ b/crypto/bio/bio_cb.c
@@ -85,28 +85,32 @@ long MS_CALLBACK BIO_debug_callback(BIO *bio, int cmd, const char *argp,
 		break;
 	case BIO_CB_READ:
 		if (bio->method->type & BIO_TYPE_DESCRIPTOR)
-			BIO_snprintf(p,p_maxlen,"read(%d,%d) - %s fd=%d\n",
-				 bio->num,argi,bio->method->name,bio->num);
+			BIO_snprintf(p,p_maxlen,"read(%d,%lu) - %s fd=%d\n",
+				 bio->num,(unsigned long)argi,
+				 bio->method->name,bio->num);
 		else
-			BIO_snprintf(p,p_maxlen,"read(%d,%d) - %s\n",
-				 bio->num,argi,bio->method->name);
+			BIO_snprintf(p,p_maxlen,"read(%d,%lu) - %s\n",
+				 bio->num,(unsigned long)argi,
+				 bio->method->name);
 		break;
 	case BIO_CB_WRITE:
 		if (bio->method->type & BIO_TYPE_DESCRIPTOR)
-			BIO_snprintf(p,p_maxlen,"write(%d,%d) - %s fd=%d\n",
-				 bio->num,argi,bio->method->name,bio->num);
+			BIO_snprintf(p,p_maxlen,"write(%d,%lu) - %s fd=%d\n",
+				 bio->num,(unsigned long)argi,
+				 bio->method->name,bio->num);
 		else
-			BIO_snprintf(p,p_maxlen,"write(%d,%d) - %s\n",
-				 bio->num,argi,bio->method->name);
+			BIO_snprintf(p,p_maxlen,"write(%d,%lu) - %s\n",
+				 bio->num,(unsigned long)argi,
+				 bio->method->name);
 		break;
 	case BIO_CB_PUTS:
 		BIO_snprintf(p,p_maxlen,"puts() - %s\n",bio->method->name);
 		break;
 	case BIO_CB_GETS:
-		BIO_snprintf(p,p_maxlen,"gets(%d) - %s\n",argi,bio->method->name);
+		BIO_snprintf(p,p_maxlen,"gets(%lu) - %s\n",(unsigned long)argi,bio->method->name);
 		break;
 	case BIO_CB_CTRL:
-		BIO_snprintf(p,p_maxlen,"ctrl(%d) - %s\n",argi,bio->method->name);
+		BIO_snprintf(p,p_maxlen,"ctrl(%lu) - %s\n",(unsigned long)argi,bio->method->name);
 		break;
 	case BIO_CB_RETURN|BIO_CB_READ:
 		BIO_snprintf(p,p_maxlen,"read return %ld\n",ret);
diff --git a/crypto/bio/bio_err.c b/crypto/bio/bio_err.c
index 6603f1c74dc8..0dbfbd80d30c 100644
--- a/crypto/bio/bio_err.c
+++ b/crypto/bio/bio_err.c
@@ -1,6 +1,6 @@
 /* crypto/bio/bio_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -95,6 +95,7 @@ static ERR_STRING_DATA BIO_str_functs[]=
 {ERR_FUNC(BIO_F_BUFFER_CTRL),	"BUFFER_CTRL"},
 {ERR_FUNC(BIO_F_CONN_CTRL),	"CONN_CTRL"},
 {ERR_FUNC(BIO_F_CONN_STATE),	"CONN_STATE"},
+{ERR_FUNC(BIO_F_DGRAM_SCTP_READ),	"DGRAM_SCTP_READ"},
 {ERR_FUNC(BIO_F_FILE_CTRL),	"FILE_CTRL"},
 {ERR_FUNC(BIO_F_FILE_READ),	"FILE_READ"},
 {ERR_FUNC(BIO_F_LINEBUFFER_CTRL),	"LINEBUFFER_CTRL"},
diff --git a/crypto/bio/bio_lcl.h b/crypto/bio/bio_lcl.h
index dba2919d430c..e7f7ec8d8bf0 100644
--- a/crypto/bio/bio_lcl.h
+++ b/crypto/bio/bio_lcl.h
@@ -18,11 +18,19 @@
 #define	UP_ftell	ftell
 #define	UP_fflush	fflush
 #define	UP_ferror	ferror
+#ifdef _WIN32
+#define	UP_fileno	_fileno
+#define	UP_open		_open
+#define	UP_read		_read
+#define	UP_write	_write
+#define	UP_lseek	_lseek
+#define	UP_close	_close
+#else
 #define	UP_fileno	fileno
-
 #define	UP_open		open
 #define	UP_read		read
 #define	UP_write	write
 #define	UP_lseek	lseek
 #define	UP_close	close
 #endif
+#endif
diff --git a/crypto/bio/bio_lib.c b/crypto/bio/bio_lib.c
index 371cdf52cb06..9c9646afa8a2 100644
--- a/crypto/bio/bio_lib.c
+++ b/crypto/bio/bio_lib.c
@@ -429,7 +429,7 @@ BIO *BIO_push(BIO *b, BIO *bio)
 	if (bio != NULL)
 		bio->prev_bio=lb;
 	/* called to do internal processing */
-	BIO_ctrl(b,BIO_CTRL_PUSH,0,NULL);
+	BIO_ctrl(b,BIO_CTRL_PUSH,0,lb);
 	return(b);
 	}
 
@@ -441,7 +441,7 @@ BIO *BIO_pop(BIO *b)
 	if (b == NULL) return(NULL);
 	ret=b->next_bio;
 
-	BIO_ctrl(b,BIO_CTRL_POP,0,NULL);
+	BIO_ctrl(b,BIO_CTRL_POP,0,b);
 
 	if (b->prev_bio != NULL)
 		b->prev_bio->next_bio=b->next_bio;
@@ -521,40 +521,40 @@ void BIO_free_all(BIO *bio)
 
 BIO *BIO_dup_chain(BIO *in)
 	{
-	BIO *ret=NULL,*eoc=NULL,*bio,*new;
+	BIO *ret=NULL,*eoc=NULL,*bio,*new_bio;
 
 	for (bio=in; bio != NULL; bio=bio->next_bio)
 		{
-		if ((new=BIO_new(bio->method)) == NULL) goto err;
-		new->callback=bio->callback;
-		new->cb_arg=bio->cb_arg;
-		new->init=bio->init;
-		new->shutdown=bio->shutdown;
-		new->flags=bio->flags;
+		if ((new_bio=BIO_new(bio->method)) == NULL) goto err;
+		new_bio->callback=bio->callback;
+		new_bio->cb_arg=bio->cb_arg;
+		new_bio->init=bio->init;
+		new_bio->shutdown=bio->shutdown;
+		new_bio->flags=bio->flags;
 
 		/* This will let SSL_s_sock() work with stdin/stdout */
-		new->num=bio->num;
+		new_bio->num=bio->num;
 
-		if (!BIO_dup_state(bio,(char *)new))
+		if (!BIO_dup_state(bio,(char *)new_bio))
 			{
-			BIO_free(new);
+			BIO_free(new_bio);
 			goto err;
 			}
 
 		/* copy app data */
-		if (!CRYPTO_dup_ex_data(CRYPTO_EX_INDEX_BIO, &new->ex_data,
+		if (!CRYPTO_dup_ex_data(CRYPTO_EX_INDEX_BIO, &new_bio->ex_data,
 					&bio->ex_data))
 			goto err;
 
 		if (ret == NULL)
 			{
-			eoc=new;
+			eoc=new_bio;
 			ret=eoc;
 			}
 		else
 			{
-			BIO_push(eoc,new);
-			eoc=new;
+			BIO_push(eoc,new_bio);
+			eoc=new_bio;
 			}
 		}
 	return(ret);
diff --git a/crypto/bio/bss_acpt.c b/crypto/bio/bss_acpt.c
index e7fb892eaa2b..5d49e1a72bcc 100644
--- a/crypto/bio/bss_acpt.c
+++ b/crypto/bio/bss_acpt.c
@@ -100,8 +100,8 @@ static int acpt_new(BIO *h);
 static int acpt_free(BIO *data);
 static int acpt_state(BIO *b, BIO_ACCEPT *c);
 static void acpt_close_socket(BIO *data);
-BIO_ACCEPT *BIO_ACCEPT_new(void );
-void BIO_ACCEPT_free(BIO_ACCEPT *a);
+static BIO_ACCEPT *BIO_ACCEPT_new(void );
+static void BIO_ACCEPT_free(BIO_ACCEPT *a);
 
 #define ACPT_S_BEFORE			1
 #define ACPT_S_GET_ACCEPT_SOCKET	2
@@ -141,7 +141,7 @@ static int acpt_new(BIO *bi)
 	return(1);
 	}
 
-BIO_ACCEPT *BIO_ACCEPT_new(void)
+static BIO_ACCEPT *BIO_ACCEPT_new(void)
 	{
 	BIO_ACCEPT *ret;
 
@@ -154,7 +154,7 @@ BIO_ACCEPT *BIO_ACCEPT_new(void)
 	return(ret);
 	}
 
-void BIO_ACCEPT_free(BIO_ACCEPT *a)
+static void BIO_ACCEPT_free(BIO_ACCEPT *a)
 	{
 	if(a == NULL)
 	    return;
diff --git a/crypto/bio/bss_bio.c b/crypto/bio/bss_bio.c
index 76bd48e7679b..52ef0ebcb3e5 100644
--- a/crypto/bio/bss_bio.c
+++ b/crypto/bio/bss_bio.c
@@ -277,10 +277,10 @@ static int bio_read(BIO *bio, char *buf, int size_)
  */
 /* WARNING: The non-copying interface is largely untested as of yet
  * and may contain bugs. */
-static ssize_t bio_nread0(BIO *bio, char **buf)
+static ossl_ssize_t bio_nread0(BIO *bio, char **buf)
 	{
 	struct bio_bio_st *b, *peer_b;
-	ssize_t num;
+	ossl_ssize_t num;
 	
 	BIO_clear_retry_flags(bio);
 
@@ -315,15 +315,15 @@ static ssize_t bio_nread0(BIO *bio, char **buf)
 	return num;
 	}
 
-static ssize_t bio_nread(BIO *bio, char **buf, size_t num_)
+static ossl_ssize_t bio_nread(BIO *bio, char **buf, size_t num_)
 	{
 	struct bio_bio_st *b, *peer_b;
-	ssize_t num, available;
+	ossl_ssize_t num, available;
 
 	if (num_ > SSIZE_MAX)
 		num = SSIZE_MAX;
 	else
-		num = (ssize_t)num_;
+		num = (ossl_ssize_t)num_;
 
 	available = bio_nread0(bio, buf);
 	if (num > available)
@@ -428,7 +428,7 @@ static int bio_write(BIO *bio, const char *buf, int num_)
  * (example usage:  bio_nwrite0(), write to buffer, bio_nwrite()
  *  or just         bio_nwrite(), write to buffer)
  */
-static ssize_t bio_nwrite0(BIO *bio, char **buf)
+static ossl_ssize_t bio_nwrite0(BIO *bio, char **buf)
 	{
 	struct bio_bio_st *b;
 	size_t num;
@@ -476,15 +476,15 @@ static ssize_t bio_nwrite0(BIO *bio, char **buf)
 	return num;
 	}
 
-static ssize_t bio_nwrite(BIO *bio, char **buf, size_t num_)
+static ossl_ssize_t bio_nwrite(BIO *bio, char **buf, size_t num_)
 	{
 	struct bio_bio_st *b;
-	ssize_t num, space;
+	ossl_ssize_t num, space;
 
 	if (num_ > SSIZE_MAX)
 		num = SSIZE_MAX;
 	else
-		num = (ssize_t)num_;
+		num = (ossl_ssize_t)num_;
 
 	space = bio_nwrite0(bio, buf);
 	if (num > space)
diff --git a/crypto/bio/bss_dgram.c b/crypto/bio/bss_dgram.c
index e0327bdea672..1b1e4bec8121 100644
--- a/crypto/bio/bss_dgram.c
+++ b/crypto/bio/bss_dgram.c
@@ -70,6 +70,13 @@
 #include <sys/timeb.h>
 #endif
 
+#ifndef OPENSSL_NO_SCTP
+#include <netinet/sctp.h>
+#include <fcntl.h>
+#define OPENSSL_SCTP_DATA_CHUNK_TYPE            0x00
+#define OPENSSL_SCTP_FORWARD_CUM_TSN_CHUNK_TYPE 0xc0
+#endif
+
 #ifdef OPENSSL_SYS_LINUX
 #define IP_MTU      14 /* linux is lame */
 #endif
@@ -88,6 +95,18 @@ static int dgram_new(BIO *h);
 static int dgram_free(BIO *data);
 static int dgram_clear(BIO *bio);
 
+#ifndef OPENSSL_NO_SCTP
+static int dgram_sctp_write(BIO *h, const char *buf, int num);
+static int dgram_sctp_read(BIO *h, char *buf, int size);
+static int dgram_sctp_puts(BIO *h, const char *str);
+static long dgram_sctp_ctrl(BIO *h, int cmd, long arg1, void *arg2);
+static int dgram_sctp_new(BIO *h);
+static int dgram_sctp_free(BIO *data);
+#ifdef SCTP_AUTHENTICATION_EVENT
+static void dgram_sctp_handle_auth_free_key_event(BIO *b, union sctp_notification *snp);
+#endif
+#endif
+
 static int BIO_dgram_should_retry(int s);
 
 static void get_current_time(struct timeval *t);
@@ -106,9 +125,31 @@ static BIO_METHOD methods_dgramp=
 	NULL,
 	};
 
+#ifndef OPENSSL_NO_SCTP
+static BIO_METHOD methods_dgramp_sctp=
+	{
+	BIO_TYPE_DGRAM_SCTP,
+	"datagram sctp socket",
+	dgram_sctp_write,
+	dgram_sctp_read,
+	dgram_sctp_puts,
+	NULL, /* dgram_gets, */
+	dgram_sctp_ctrl,
+	dgram_sctp_new,
+	dgram_sctp_free,
+	NULL,
+	};
+#endif
+
 typedef struct bio_dgram_data_st
 	{
-	struct sockaddr peer;
+	union {
+		struct sockaddr sa;
+		struct sockaddr_in sa_in;
+#if OPENSSL_USE_IPV6
+		struct sockaddr_in6 sa_in6;
+#endif
+	} peer;
 	unsigned int connected;
 	unsigned int _errno;
 	unsigned int mtu;
@@ -116,6 +157,40 @@ typedef struct bio_dgram_data_st
 	struct timeval socket_timeout;
 	} bio_dgram_data;
 
+#ifndef OPENSSL_NO_SCTP
+typedef struct bio_dgram_sctp_save_message_st
+	{
+        BIO *bio;
+        char *data;
+        int length;
+	} bio_dgram_sctp_save_message;
+
+typedef struct bio_dgram_sctp_data_st
+	{
+	union {
+		struct sockaddr sa;
+		struct sockaddr_in sa_in;
+#if OPENSSL_USE_IPV6
+		struct sockaddr_in6 sa_in6;
+#endif
+	} peer;
+	unsigned int connected;
+	unsigned int _errno;
+	unsigned int mtu;
+	struct bio_dgram_sctp_sndinfo sndinfo;
+	struct bio_dgram_sctp_rcvinfo rcvinfo;
+	struct bio_dgram_sctp_prinfo prinfo;
+	void (*handle_notifications)(BIO *bio, void *context, void *buf);
+	void* notification_context;
+	int in_handshake;
+	int ccs_rcvd;
+	int ccs_sent;
+	int save_shutdown;
+	int peer_auth_tested;
+	bio_dgram_sctp_save_message saved_message;
+	} bio_dgram_sctp_data;
+#endif
+
 BIO_METHOD *BIO_s_datagram(void)
 	{
 	return(&methods_dgramp);
@@ -274,23 +349,37 @@ static int dgram_read(BIO *b, char *out, int outl)
 	int ret=0;
 	bio_dgram_data *data = (bio_dgram_data *)b->ptr;
 
-	struct sockaddr peer;
-	int peerlen = sizeof(peer);
+	struct	{
+	/*
+	 * See commentary in b_sock.c. <appro>
+	 */
+	union	{ size_t s; int i; } len;
+	union	{
+		struct sockaddr sa;
+		struct sockaddr_in sa_in;
+#if OPENSSL_USE_IPV6
+		struct sockaddr_in6 sa_in6;
+#endif
+		} peer;
+	} sa;
+
+	sa.len.s=0;
+	sa.len.i=sizeof(sa.peer);
 
 	if (out != NULL)
 		{
 		clear_socket_error();
-		memset(&peer, 0x00, peerlen);
-		/* Last arg in recvfrom is signed on some platforms and
-		 * unsigned on others. It is of type socklen_t on some
-		 * but this is not universal. Cast to (void *) to avoid
-		 * compiler warnings.
-		 */
+		memset(&sa.peer, 0x00, sizeof(sa.peer));
 		dgram_adjust_rcv_timeout(b);
-		ret=recvfrom(b->num,out,outl,0,&peer,(void *)&peerlen);
+		ret=recvfrom(b->num,out,outl,0,&sa.peer.sa,(void *)&sa.len);
+		if (sizeof(sa.len.i)!=sizeof(sa.len.s) && sa.len.i==0)
+			{
+			OPENSSL_assert(sa.len.s<=sizeof(sa.peer));
+			sa.len.i = (int)sa.len.s;
+			}
 
 		if ( ! data->connected  && ret >= 0)
-			BIO_ctrl(b, BIO_CTRL_DGRAM_SET_PEER, 0, &peer);
+			BIO_ctrl(b, BIO_CTRL_DGRAM_SET_PEER, 0, &sa.peer);
 
 		BIO_clear_retry_flags(b);
 		if (ret < 0)
@@ -313,14 +402,24 @@ static int dgram_write(BIO *b, const char *in, int inl)
 	bio_dgram_data *data = (bio_dgram_data *)b->ptr;
 	clear_socket_error();
 
-    if ( data->connected )
-        ret=writesocket(b->num,in,inl);
-    else
+	if ( data->connected )
+		ret=writesocket(b->num,in,inl);
+	else
+		{
+		int peerlen = sizeof(data->peer);
+
+		if (data->peer.sa.sa_family == AF_INET)
+			peerlen = sizeof(data->peer.sa_in);
+#if OPENSSL_USE_IPV6
+		else if (data->peer.sa.sa_family == AF_INET6)
+			peerlen = sizeof(data->peer.sa_in6);
+#endif
 #if defined(NETWARE_CLIB) && defined(NETWARE_BSDSOCK)
-        ret=sendto(b->num, (char *)in, inl, 0, &data->peer, sizeof(data->peer));
+		ret=sendto(b->num, (char *)in, inl, 0, &data->peer.sa, peerlen);
 #else
-        ret=sendto(b->num, in, inl, 0, &data->peer, sizeof(data->peer));
+		ret=sendto(b->num, in, inl, 0, &data->peer.sa, peerlen);
 #endif
+		}
 
 	BIO_clear_retry_flags(b);
 	if (ret <= 0)
@@ -352,7 +451,13 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 #endif
 #ifdef OPENSSL_SYS_LINUX
 	socklen_t addr_len;
-	struct sockaddr_storage addr;
+	union	{
+		struct sockaddr	sa;
+		struct sockaddr_in s4;
+#if OPENSSL_USE_IPV6
+		struct sockaddr_in6 s6;
+#endif
+		} addr;
 #endif
 
 	data = (bio_dgram_data *)b->ptr;
@@ -406,7 +511,20 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 		else
 			{
 #endif
-			memcpy(&(data->peer),to, sizeof(struct sockaddr));
+			switch (to->sa_family)
+				{
+				case AF_INET:
+					memcpy(&data->peer,to,sizeof(data->peer.sa_in));
+					break;
+#if OPENSSL_USE_IPV6
+				case AF_INET6:
+					memcpy(&data->peer,to,sizeof(data->peer.sa_in6));
+					break;
+#endif
+				default:
+					memcpy(&data->peer,to,sizeof(data->peer.sa));
+					break;
+				}
 #if 0
 			}
 #endif
@@ -414,15 +532,15 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 		/* (Linux)kernel sets DF bit on outgoing IP packets */
 	case BIO_CTRL_DGRAM_MTU_DISCOVER:
 #ifdef OPENSSL_SYS_LINUX
-		addr_len = (socklen_t)sizeof(struct sockaddr_storage);
-		memset((void *)&addr, 0, sizeof(struct sockaddr_storage));
-		if (getsockname(b->num, (void *)&addr, &addr_len) < 0)
+		addr_len = (socklen_t)sizeof(addr);
+		memset((void *)&addr, 0, sizeof(addr));
+		if (getsockname(b->num, &addr.sa, &addr_len) < 0)
 			{
 			ret = 0;
 			break;
 			}
 		sockopt_len = sizeof(sockopt_val);
-		switch (addr.ss_family)
+		switch (addr.sa.sa_family)
 			{
 		case AF_INET:
 			sockopt_val = IP_PMTUDISC_DO;
@@ -430,12 +548,14 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 				&sockopt_val, sizeof(sockopt_val))) < 0)
 				perror("setsockopt");
 			break;
+#if OPENSSL_USE_IPV6 && defined(IPV6_MTU_DISCOVER)
 		case AF_INET6:
 			sockopt_val = IPV6_PMTUDISC_DO;
 			if ((ret = setsockopt(b->num, IPPROTO_IPV6, IPV6_MTU_DISCOVER,
 				&sockopt_val, sizeof(sockopt_val))) < 0)
 				perror("setsockopt");
 			break;
+#endif
 		default:
 			ret = -1;
 			break;
@@ -446,15 +566,15 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 #endif
 	case BIO_CTRL_DGRAM_QUERY_MTU:
 #ifdef OPENSSL_SYS_LINUX
-		addr_len = (socklen_t)sizeof(struct sockaddr_storage);
-		memset((void *)&addr, 0, sizeof(struct sockaddr_storage));
-		if (getsockname(b->num, (void *)&addr, &addr_len) < 0)
+		addr_len = (socklen_t)sizeof(addr);
+		memset((void *)&addr, 0, sizeof(addr));
+		if (getsockname(b->num, &addr.sa, &addr_len) < 0)
 			{
 			ret = 0;
 			break;
 			}
 		sockopt_len = sizeof(sockopt_val);
-		switch (addr.ss_family)
+		switch (addr.sa.sa_family)
 			{
 		case AF_INET:
 			if ((ret = getsockopt(b->num, IPPROTO_IP, IP_MTU, (void *)&sockopt_val,
@@ -471,6 +591,7 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 				ret = data->mtu;
 				}
 			break;
+#if OPENSSL_USE_IPV6 && defined(IPV6_MTU)
 		case AF_INET6:
 			if ((ret = getsockopt(b->num, IPPROTO_IPV6, IPV6_MTU, (void *)&sockopt_val,
 				&sockopt_len)) < 0 || sockopt_val < 0)
@@ -486,6 +607,7 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 				ret = data->mtu;
 				}
 			break;
+#endif
 		default:
 			ret = 0;
 			break;
@@ -495,7 +617,25 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 #endif
 		break;
 	case BIO_CTRL_DGRAM_GET_FALLBACK_MTU:
-		ret = 576 - 20 - 8;
+		switch (data->peer.sa.sa_family)
+			{
+			case AF_INET:
+				ret = 576 - 20 - 8;
+				break;
+#if OPENSSL_USE_IPV6
+			case AF_INET6:
+#ifdef IN6_IS_ADDR_V4MAPPED
+				if (IN6_IS_ADDR_V4MAPPED(&data->peer.sa_in6.sin6_addr))
+					ret = 576 - 20 - 8;
+				else
+#endif
+					ret = 1280 - 40 - 8;
+				break;
+#endif
+			default:
+				ret = 576 - 20 - 8;
+				break;
+			}
 		break;
 	case BIO_CTRL_DGRAM_GET_MTU:
 		return data->mtu;
@@ -510,27 +650,65 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 		if ( to != NULL)
 			{
 			data->connected = 1;
-			memcpy(&(data->peer),to, sizeof(struct sockaddr));
+			switch (to->sa_family)
+				{
+				case AF_INET:
+					memcpy(&data->peer,to,sizeof(data->peer.sa_in));
+					break;
+#if OPENSSL_USE_IPV6
+				case AF_INET6:
+					memcpy(&data->peer,to,sizeof(data->peer.sa_in6));
+					break;
+#endif
+				default:
+					memcpy(&data->peer,to,sizeof(data->peer.sa));
+					break;
+				}
 			}
 		else
 			{
 			data->connected = 0;
-			memset(&(data->peer), 0x00, sizeof(struct sockaddr));
+			memset(&(data->peer), 0x00, sizeof(data->peer));
+			}
+		break;
+	case BIO_CTRL_DGRAM_GET_PEER:
+		switch (data->peer.sa.sa_family)
+			{
+			case AF_INET:
+				ret=sizeof(data->peer.sa_in);
+				break;
+#if OPENSSL_USE_IPV6
+			case AF_INET6:
+				ret=sizeof(data->peer.sa_in6);
+				break;
+#endif
+			default:
+				ret=sizeof(data->peer.sa);
+				break;
+			}
+		if (num==0 || num>ret)
+			num=ret;
+		memcpy(ptr,&data->peer,(ret=num));
+		break;
+	case BIO_CTRL_DGRAM_SET_PEER:
+		to = (struct sockaddr *) ptr;
+		switch (to->sa_family)
+			{
+			case AF_INET:
+				memcpy(&data->peer,to,sizeof(data->peer.sa_in));
+				break;
+#if OPENSSL_USE_IPV6
+			case AF_INET6:
+				memcpy(&data->peer,to,sizeof(data->peer.sa_in6));
+				break;
+#endif
+			default:
+				memcpy(&data->peer,to,sizeof(data->peer.sa));
+				break;
 			}
 		break;
-    case BIO_CTRL_DGRAM_GET_PEER:
-        to = (struct sockaddr *) ptr;
-
-        memcpy(to, &(data->peer), sizeof(struct sockaddr));
-		ret = sizeof(struct sockaddr);
-        break;
-    case BIO_CTRL_DGRAM_SET_PEER:
-        to = (struct sockaddr *) ptr;
-
-        memcpy(&(data->peer), to, sizeof(struct sockaddr));
-        break;
 	case BIO_CTRL_DGRAM_SET_NEXT_TIMEOUT:
-		memcpy(&(data->next_timeout), ptr, sizeof(struct timeval));		
+		memcpy(&(data->next_timeout), ptr, sizeof(struct timeval));
 		break;
 #if defined(SO_RCVTIMEO)
 	case BIO_CTRL_DGRAM_SET_RECV_TIMEOUT:
@@ -650,6 +828,912 @@ static int dgram_puts(BIO *bp, const char *str)
 	return(ret);
 	}
 
+#ifndef OPENSSL_NO_SCTP
+BIO_METHOD *BIO_s_datagram_sctp(void)
+	{
+	return(&methods_dgramp_sctp);
+	}
+
+BIO *BIO_new_dgram_sctp(int fd, int close_flag)
+	{
+	BIO *bio;
+	int ret, optval = 20000;
+	int auth_data = 0, auth_forward = 0;
+	unsigned char *p;
+	struct sctp_authchunk auth;
+	struct sctp_authchunks *authchunks;
+	socklen_t sockopt_len;
+#ifdef SCTP_AUTHENTICATION_EVENT
+#ifdef SCTP_EVENT
+	struct sctp_event event;
+#else
+	struct sctp_event_subscribe event;
+#endif
+#endif
+
+	bio=BIO_new(BIO_s_datagram_sctp());
+	if (bio == NULL) return(NULL);
+	BIO_set_fd(bio,fd,close_flag);
+
+	/* Activate SCTP-AUTH for DATA and FORWARD-TSN chunks */
+	auth.sauth_chunk = OPENSSL_SCTP_DATA_CHUNK_TYPE;
+	ret = setsockopt(fd, IPPROTO_SCTP, SCTP_AUTH_CHUNK, &auth, sizeof(struct sctp_authchunk));
+	OPENSSL_assert(ret >= 0);
+	auth.sauth_chunk = OPENSSL_SCTP_FORWARD_CUM_TSN_CHUNK_TYPE;
+	ret = setsockopt(fd, IPPROTO_SCTP, SCTP_AUTH_CHUNK, &auth, sizeof(struct sctp_authchunk));
+	OPENSSL_assert(ret >= 0);
+
+	/* Test if activation was successful. When using accept(),
+	 * SCTP-AUTH has to be activated for the listening socket
+	 * already, otherwise the connected socket won't use it. */
+	sockopt_len = (socklen_t)(sizeof(sctp_assoc_t) + 256 * sizeof(uint8_t));
+	authchunks = OPENSSL_malloc(sockopt_len);
+	memset(authchunks, 0, sizeof(sockopt_len));
+	ret = getsockopt(fd, IPPROTO_SCTP, SCTP_LOCAL_AUTH_CHUNKS, authchunks, &sockopt_len);
+	OPENSSL_assert(ret >= 0);
+	
+	for (p = (unsigned char*) authchunks + sizeof(sctp_assoc_t);
+	     p < (unsigned char*) authchunks + sockopt_len;
+	     p += sizeof(uint8_t))
+		{
+		if (*p == OPENSSL_SCTP_DATA_CHUNK_TYPE) auth_data = 1;
+		if (*p == OPENSSL_SCTP_FORWARD_CUM_TSN_CHUNK_TYPE) auth_forward = 1;
+		}
+		
+	OPENSSL_free(authchunks);
+
+	OPENSSL_assert(auth_data);
+	OPENSSL_assert(auth_forward);
+
+#ifdef SCTP_AUTHENTICATION_EVENT
+#ifdef SCTP_EVENT
+	memset(&event, 0, sizeof(struct sctp_event));
+	event.se_assoc_id = 0;
+	event.se_type = SCTP_AUTHENTICATION_EVENT;
+	event.se_on = 1;
+	ret = setsockopt(fd, IPPROTO_SCTP, SCTP_EVENT, &event, sizeof(struct sctp_event));
+	OPENSSL_assert(ret >= 0);
+#else
+	sockopt_len = (socklen_t) sizeof(struct sctp_event_subscribe);
+	ret = getsockopt(fd, IPPROTO_SCTP, SCTP_EVENTS, &event, &sockopt_len);
+	OPENSSL_assert(ret >= 0);
+
+	event.sctp_authentication_event = 1;
+
+	ret = setsockopt(fd, IPPROTO_SCTP, SCTP_EVENTS, &event, sizeof(struct sctp_event_subscribe));
+	OPENSSL_assert(ret >= 0);
+#endif
+#endif
+
+	/* Disable partial delivery by setting the min size
+	 * larger than the max record size of 2^14 + 2048 + 13
+	 */
+	ret = setsockopt(fd, IPPROTO_SCTP, SCTP_PARTIAL_DELIVERY_POINT, &optval, sizeof(optval));
+	OPENSSL_assert(ret >= 0);
+
+	return(bio);
+	}
+
+int BIO_dgram_is_sctp(BIO *bio)
+	{
+	return (BIO_method_type(bio) == BIO_TYPE_DGRAM_SCTP);
+	}
+
+static int dgram_sctp_new(BIO *bi)
+	{
+	bio_dgram_sctp_data *data = NULL;
+
+	bi->init=0;
+	bi->num=0;
+	data = OPENSSL_malloc(sizeof(bio_dgram_sctp_data));
+	if (data == NULL)
+		return 0;
+	memset(data, 0x00, sizeof(bio_dgram_sctp_data));
+#ifdef SCTP_PR_SCTP_NONE
+	data->prinfo.pr_policy = SCTP_PR_SCTP_NONE;
+#endif
+    bi->ptr = data;
+
+	bi->flags=0;
+	return(1);
+	}
+
+static int dgram_sctp_free(BIO *a)
+	{
+	bio_dgram_sctp_data *data;
+
+	if (a == NULL) return(0);
+	if ( ! dgram_clear(a))
+		return 0;
+
+	data = (bio_dgram_sctp_data *)a->ptr;
+	if(data != NULL) OPENSSL_free(data);
+
+	return(1);
+	}
+
+#ifdef SCTP_AUTHENTICATION_EVENT
+void dgram_sctp_handle_auth_free_key_event(BIO *b, union sctp_notification *snp)
+	{
+	unsigned int sockopt_len = 0;
+	int ret;
+	struct sctp_authkey_event* authkeyevent = &snp->sn_auth_event;
+
+	if (authkeyevent->auth_indication == SCTP_AUTH_FREE_KEY)
+		{
+		struct sctp_authkeyid authkeyid;
+
+		/* delete key */
+		authkeyid.scact_keynumber = authkeyevent->auth_keynumber;
+		sockopt_len = sizeof(struct sctp_authkeyid);
+		ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_AUTH_DELETE_KEY,
+		      &authkeyid, sockopt_len);
+		}
+	}
+#endif
+
+static int dgram_sctp_read(BIO *b, char *out, int outl)
+	{
+	int ret = 0, n = 0, i, optval;
+	socklen_t optlen;
+	bio_dgram_sctp_data *data = (bio_dgram_sctp_data *)b->ptr;
+	union sctp_notification *snp;
+	struct msghdr msg;
+	struct iovec iov;
+	struct cmsghdr *cmsg;
+	char cmsgbuf[512];
+
+	if (out != NULL)
+		{
+		clear_socket_error();
+
+		do
+			{
+			memset(&data->rcvinfo, 0x00, sizeof(struct bio_dgram_sctp_rcvinfo));
+			iov.iov_base = out;
+			iov.iov_len = outl;
+			msg.msg_name = NULL;
+			msg.msg_namelen = 0;
+			msg.msg_iov = &iov;
+			msg.msg_iovlen = 1;
+			msg.msg_control = cmsgbuf;
+			msg.msg_controllen = 512;
+			msg.msg_flags = 0;
+			n = recvmsg(b->num, &msg, 0);
+
+			if (msg.msg_controllen > 0)
+				{
+				for (cmsg = CMSG_FIRSTHDR(&msg); cmsg; cmsg = CMSG_NXTHDR(&msg, cmsg))
+					{
+					if (cmsg->cmsg_level != IPPROTO_SCTP)
+						continue;
+#ifdef SCTP_RCVINFO
+					if (cmsg->cmsg_type == SCTP_RCVINFO)
+						{
+						struct sctp_rcvinfo *rcvinfo;
+
+						rcvinfo = (struct sctp_rcvinfo *)CMSG_DATA(cmsg);
+						data->rcvinfo.rcv_sid = rcvinfo->rcv_sid;
+						data->rcvinfo.rcv_ssn = rcvinfo->rcv_ssn;
+						data->rcvinfo.rcv_flags = rcvinfo->rcv_flags;
+						data->rcvinfo.rcv_ppid = rcvinfo->rcv_ppid;
+						data->rcvinfo.rcv_tsn = rcvinfo->rcv_tsn;
+						data->rcvinfo.rcv_cumtsn = rcvinfo->rcv_cumtsn;
+						data->rcvinfo.rcv_context = rcvinfo->rcv_context;
+						}
+#endif
+#ifdef SCTP_SNDRCV
+					if (cmsg->cmsg_type == SCTP_SNDRCV)
+						{
+						struct sctp_sndrcvinfo *sndrcvinfo;
+
+						sndrcvinfo = (struct sctp_sndrcvinfo *)CMSG_DATA(cmsg);
+						data->rcvinfo.rcv_sid = sndrcvinfo->sinfo_stream;
+						data->rcvinfo.rcv_ssn = sndrcvinfo->sinfo_ssn;
+						data->rcvinfo.rcv_flags = sndrcvinfo->sinfo_flags;
+						data->rcvinfo.rcv_ppid = sndrcvinfo->sinfo_ppid;
+						data->rcvinfo.rcv_tsn = sndrcvinfo->sinfo_tsn;
+						data->rcvinfo.rcv_cumtsn = sndrcvinfo->sinfo_cumtsn;
+						data->rcvinfo.rcv_context = sndrcvinfo->sinfo_context;
+						}
+#endif
+					}
+				}
+
+			if (n <= 0)
+				{
+				if (n < 0)
+					ret = n;
+				break;
+				}
+
+			if (msg.msg_flags & MSG_NOTIFICATION)
+				{
+				snp = (union sctp_notification*) out;
+				if (snp->sn_header.sn_type == SCTP_SENDER_DRY_EVENT)
+					{
+#ifdef SCTP_EVENT
+					struct sctp_event event;
+#else
+					struct sctp_event_subscribe event;
+					socklen_t eventsize;
+#endif
+					/* If a message has been delayed until the socket
+					 * is dry, it can be sent now.
+					 */
+					if (data->saved_message.length > 0)
+						{
+						dgram_sctp_write(data->saved_message.bio, data->saved_message.data,
+						                 data->saved_message.length);
+						OPENSSL_free(data->saved_message.data);
+						data->saved_message.length = 0;
+						}
+
+					/* disable sender dry event */
+#ifdef SCTP_EVENT
+					memset(&event, 0, sizeof(struct sctp_event));
+					event.se_assoc_id = 0;
+					event.se_type = SCTP_SENDER_DRY_EVENT;
+					event.se_on = 0;
+					i = setsockopt(b->num, IPPROTO_SCTP, SCTP_EVENT, &event, sizeof(struct sctp_event));
+					OPENSSL_assert(i >= 0);
+#else
+					eventsize = sizeof(struct sctp_event_subscribe);
+					i = getsockopt(b->num, IPPROTO_SCTP, SCTP_EVENTS, &event, &eventsize);
+					OPENSSL_assert(i >= 0);
+
+					event.sctp_sender_dry_event = 0;
+
+					i = setsockopt(b->num, IPPROTO_SCTP, SCTP_EVENTS, &event, sizeof(struct sctp_event_subscribe));
+					OPENSSL_assert(i >= 0);
+#endif
+					}
+
+#ifdef SCTP_AUTHENTICATION_EVENT
+				if (snp->sn_header.sn_type == SCTP_AUTHENTICATION_EVENT)
+					dgram_sctp_handle_auth_free_key_event(b, snp);
+#endif
+
+				if (data->handle_notifications != NULL)
+					data->handle_notifications(b, data->notification_context, (void*) out);
+
+				memset(out, 0, outl);
+				}
+			else
+				ret += n;
+			}
+		while ((msg.msg_flags & MSG_NOTIFICATION) && (msg.msg_flags & MSG_EOR) && (ret < outl));
+
+		if (ret > 0 && !(msg.msg_flags & MSG_EOR))
+			{
+			/* Partial message read, this should never happen! */
+
+			/* The buffer was too small, this means the peer sent
+			 * a message that was larger than allowed. */
+			if (ret == outl)
+				return -1;
+
+			/* Test if socket buffer can handle max record
+			 * size (2^14 + 2048 + 13)
+			 */
+			optlen = (socklen_t) sizeof(int);
+			ret = getsockopt(b->num, SOL_SOCKET, SO_RCVBUF, &optval, &optlen);
+			OPENSSL_assert(ret >= 0);
+			OPENSSL_assert(optval >= 18445);
+
+			/* Test if SCTP doesn't partially deliver below
+			 * max record size (2^14 + 2048 + 13)
+			 */
+			optlen = (socklen_t) sizeof(int);
+			ret = getsockopt(b->num, IPPROTO_SCTP, SCTP_PARTIAL_DELIVERY_POINT,
+			                 &optval, &optlen);
+			OPENSSL_assert(ret >= 0);
+			OPENSSL_assert(optval >= 18445);
+
+			/* Partially delivered notification??? Probably a bug.... */
+			OPENSSL_assert(!(msg.msg_flags & MSG_NOTIFICATION));
+
+			/* Everything seems ok till now, so it's most likely
+			 * a message dropped by PR-SCTP.
+			 */
+			memset(out, 0, outl);
+			BIO_set_retry_read(b);
+			return -1;
+			}
+
+		BIO_clear_retry_flags(b);
+		if (ret < 0)
+			{
+			if (BIO_dgram_should_retry(ret))
+				{
+				BIO_set_retry_read(b);
+				data->_errno = get_last_socket_error();
+				}
+			}
+
+		/* Test if peer uses SCTP-AUTH before continuing */
+		if (!data->peer_auth_tested)
+			{
+			int ii, auth_data = 0, auth_forward = 0;
+			unsigned char *p;
+			struct sctp_authchunks *authchunks;
+
+			optlen = (socklen_t)(sizeof(sctp_assoc_t) + 256 * sizeof(uint8_t));
+			authchunks = OPENSSL_malloc(optlen);
+			memset(authchunks, 0, sizeof(optlen));
+			ii = getsockopt(b->num, IPPROTO_SCTP, SCTP_PEER_AUTH_CHUNKS, authchunks, &optlen);
+			OPENSSL_assert(ii >= 0);
+
+			for (p = (unsigned char*) authchunks + sizeof(sctp_assoc_t);
+				 p < (unsigned char*) authchunks + optlen;
+				 p += sizeof(uint8_t))
+				{
+				if (*p == OPENSSL_SCTP_DATA_CHUNK_TYPE) auth_data = 1;
+				if (*p == OPENSSL_SCTP_FORWARD_CUM_TSN_CHUNK_TYPE) auth_forward = 1;
+				}
+
+			OPENSSL_free(authchunks);
+
+			if (!auth_data || !auth_forward)
+				{
+				BIOerr(BIO_F_DGRAM_SCTP_READ,BIO_R_CONNECT_ERROR);
+				return -1;
+				}
+
+			data->peer_auth_tested = 1;
+			}
+		}
+	return(ret);
+	}
+
+static int dgram_sctp_write(BIO *b, const char *in, int inl)
+	{
+	int ret;
+	bio_dgram_sctp_data *data = (bio_dgram_sctp_data *)b->ptr;
+	struct bio_dgram_sctp_sndinfo *sinfo = &(data->sndinfo);
+	struct bio_dgram_sctp_prinfo *pinfo = &(data->prinfo);
+	struct bio_dgram_sctp_sndinfo handshake_sinfo;
+	struct iovec iov[1];
+	struct msghdr msg;
+	struct cmsghdr *cmsg;
+#if defined(SCTP_SNDINFO) && defined(SCTP_PRINFO)
+	char cmsgbuf[CMSG_SPACE(sizeof(struct sctp_sndinfo)) + CMSG_SPACE(sizeof(struct sctp_prinfo))];
+	struct sctp_sndinfo *sndinfo;
+	struct sctp_prinfo *prinfo;
+#else
+	char cmsgbuf[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];
+	struct sctp_sndrcvinfo *sndrcvinfo;
+#endif
+
+	clear_socket_error();
+
+	/* If we're send anything else than application data,
+	 * disable all user parameters and flags.
+	 */
+	if (in[0] != 23) {
+		memset(&handshake_sinfo, 0x00, sizeof(struct bio_dgram_sctp_sndinfo));
+#ifdef SCTP_SACK_IMMEDIATELY
+		handshake_sinfo.snd_flags = SCTP_SACK_IMMEDIATELY;
+#endif
+		sinfo = &handshake_sinfo;
+	}
+
+	/* If we have to send a shutdown alert message and the
+	 * socket is not dry yet, we have to save it and send it
+	 * as soon as the socket gets dry.
+	 */
+	if (data->save_shutdown && !BIO_dgram_sctp_wait_for_dry(b))
+	{
+		data->saved_message.bio = b;
+		data->saved_message.length = inl;
+		data->saved_message.data = OPENSSL_malloc(inl);
+		memcpy(data->saved_message.data, in, inl);
+		return inl;
+	}
+
+	iov[0].iov_base = (char *)in;
+	iov[0].iov_len = inl;
+	msg.msg_name = NULL;
+	msg.msg_namelen = 0;
+	msg.msg_iov = iov;
+	msg.msg_iovlen = 1;
+	msg.msg_control = (caddr_t)cmsgbuf;
+	msg.msg_controllen = 0;
+	msg.msg_flags = 0;
+#if defined(SCTP_SNDINFO) && defined(SCTP_PRINFO)
+	cmsg = (struct cmsghdr *)cmsgbuf;
+	cmsg->cmsg_level = IPPROTO_SCTP;
+	cmsg->cmsg_type = SCTP_SNDINFO;
+	cmsg->cmsg_len = CMSG_LEN(sizeof(struct sctp_sndinfo));
+	sndinfo = (struct sctp_sndinfo *)CMSG_DATA(cmsg);
+	memset(sndinfo, 0, sizeof(struct sctp_sndinfo));
+	sndinfo->snd_sid = sinfo->snd_sid;
+	sndinfo->snd_flags = sinfo->snd_flags;
+	sndinfo->snd_ppid = sinfo->snd_ppid;
+	sndinfo->snd_context = sinfo->snd_context;
+	msg.msg_controllen += CMSG_SPACE(sizeof(struct sctp_sndinfo));
+
+	cmsg = (struct cmsghdr *)&cmsgbuf[CMSG_SPACE(sizeof(struct sctp_sndinfo))];
+	cmsg->cmsg_level = IPPROTO_SCTP;
+	cmsg->cmsg_type = SCTP_PRINFO;
+	cmsg->cmsg_len = CMSG_LEN(sizeof(struct sctp_prinfo));
+	prinfo = (struct sctp_prinfo *)CMSG_DATA(cmsg);
+	memset(prinfo, 0, sizeof(struct sctp_prinfo));
+	prinfo->pr_policy = pinfo->pr_policy;
+	prinfo->pr_value = pinfo->pr_value;
+	msg.msg_controllen += CMSG_SPACE(sizeof(struct sctp_prinfo));
+#else
+	cmsg = (struct cmsghdr *)cmsgbuf;
+	cmsg->cmsg_level = IPPROTO_SCTP;
+	cmsg->cmsg_type = SCTP_SNDRCV;
+	cmsg->cmsg_len = CMSG_LEN(sizeof(struct sctp_sndrcvinfo));
+	sndrcvinfo = (struct sctp_sndrcvinfo *)CMSG_DATA(cmsg);
+	memset(sndrcvinfo, 0, sizeof(struct sctp_sndrcvinfo));
+	sndrcvinfo->sinfo_stream = sinfo->snd_sid;
+	sndrcvinfo->sinfo_flags = sinfo->snd_flags;
+#ifdef __FreeBSD__
+	sndrcvinfo->sinfo_flags |= pinfo->pr_policy;
+#endif
+	sndrcvinfo->sinfo_ppid = sinfo->snd_ppid;
+	sndrcvinfo->sinfo_context = sinfo->snd_context;
+	sndrcvinfo->sinfo_timetolive = pinfo->pr_value;
+	msg.msg_controllen += CMSG_SPACE(sizeof(struct sctp_sndrcvinfo));
+#endif
+
+	ret = sendmsg(b->num, &msg, 0);
+
+	BIO_clear_retry_flags(b);
+	if (ret <= 0)
+		{
+		if (BIO_dgram_should_retry(ret))
+			{
+			BIO_set_retry_write(b);  
+			data->_errno = get_last_socket_error();
+			}
+		}
+	return(ret);
+	}
+
+static long dgram_sctp_ctrl(BIO *b, int cmd, long num, void *ptr)
+	{
+	long ret=1;
+	bio_dgram_sctp_data *data = NULL;
+	unsigned int sockopt_len = 0;
+	struct sctp_authkeyid authkeyid;
+	struct sctp_authkey *authkey;
+
+	data = (bio_dgram_sctp_data *)b->ptr;
+
+	switch (cmd)
+		{
+	case BIO_CTRL_DGRAM_QUERY_MTU:
+		/* Set to maximum (2^14)
+		 * and ignore user input to enable transport
+		 * protocol fragmentation.
+		 * Returns always 2^14.
+		 */
+		data->mtu = 16384;
+		ret = data->mtu;
+		break;
+	case BIO_CTRL_DGRAM_SET_MTU:
+		/* Set to maximum (2^14)
+		 * and ignore input to enable transport
+		 * protocol fragmentation.
+		 * Returns always 2^14.
+		 */
+		data->mtu = 16384;
+		ret = data->mtu;
+		break;
+	case BIO_CTRL_DGRAM_SET_CONNECTED:
+	case BIO_CTRL_DGRAM_CONNECT:
+		/* Returns always -1. */
+		ret = -1;
+		break;
+	case BIO_CTRL_DGRAM_SET_NEXT_TIMEOUT:
+		/* SCTP doesn't need the DTLS timer
+		 * Returns always 1.
+		 */
+		break;
+	case BIO_CTRL_DGRAM_SCTP_SET_IN_HANDSHAKE:
+		if (num > 0)
+			data->in_handshake = 1;
+		else
+			data->in_handshake = 0;
+
+		ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_NODELAY, &data->in_handshake, sizeof(int));
+		break;
+	case BIO_CTRL_DGRAM_SCTP_ADD_AUTH_KEY:
+		/* New shared key for SCTP AUTH.
+		 * Returns 0 on success, -1 otherwise.
+		 */
+
+		/* Get active key */
+		sockopt_len = sizeof(struct sctp_authkeyid);
+		ret = getsockopt(b->num, IPPROTO_SCTP, SCTP_AUTH_ACTIVE_KEY, &authkeyid, &sockopt_len);
+		if (ret < 0) break;
+
+		/* Add new key */
+		sockopt_len = sizeof(struct sctp_authkey) + 64 * sizeof(uint8_t);
+		authkey = OPENSSL_malloc(sockopt_len);
+		memset(authkey, 0x00, sockopt_len);
+		authkey->sca_keynumber = authkeyid.scact_keynumber + 1;
+#ifndef __FreeBSD__
+		/* This field is missing in FreeBSD 8.2 and earlier,
+		 * and FreeBSD 8.3 and higher work without it.
+		 */
+		authkey->sca_keylength = 64;
+#endif
+		memcpy(&authkey->sca_key[0], ptr, 64 * sizeof(uint8_t));
+
+		ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_AUTH_KEY, authkey, sockopt_len);
+		if (ret < 0) break;
+
+		/* Reset active key */
+		ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_AUTH_ACTIVE_KEY,
+		      &authkeyid, sizeof(struct sctp_authkeyid));
+		if (ret < 0) break;
+
+		break;
+	case BIO_CTRL_DGRAM_SCTP_NEXT_AUTH_KEY:
+		/* Returns 0 on success, -1 otherwise. */
+
+		/* Get active key */
+		sockopt_len = sizeof(struct sctp_authkeyid);
+		ret = getsockopt(b->num, IPPROTO_SCTP, SCTP_AUTH_ACTIVE_KEY, &authkeyid, &sockopt_len);
+		if (ret < 0) break;
+
+		/* Set active key */
+		authkeyid.scact_keynumber = authkeyid.scact_keynumber + 1;
+		ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_AUTH_ACTIVE_KEY,
+		      &authkeyid, sizeof(struct sctp_authkeyid));
+		if (ret < 0) break;
+
+		/* CCS has been sent, so remember that and fall through
+		 * to check if we need to deactivate an old key
+		 */
+		data->ccs_sent = 1;
+
+	case BIO_CTRL_DGRAM_SCTP_AUTH_CCS_RCVD:
+		/* Returns 0 on success, -1 otherwise. */
+
+		/* Has this command really been called or is this just a fall-through? */
+		if (cmd == BIO_CTRL_DGRAM_SCTP_AUTH_CCS_RCVD)
+			data->ccs_rcvd = 1;
+
+		/* CSS has been both, received and sent, so deactivate an old key */
+		if (data->ccs_rcvd == 1 && data->ccs_sent == 1)
+			{
+			/* Get active key */
+			sockopt_len = sizeof(struct sctp_authkeyid);
+			ret = getsockopt(b->num, IPPROTO_SCTP, SCTP_AUTH_ACTIVE_KEY, &authkeyid, &sockopt_len);
+			if (ret < 0) break;
+
+			/* Deactivate key or delete second last key if
+			 * SCTP_AUTHENTICATION_EVENT is not available.
+			 */
+			authkeyid.scact_keynumber = authkeyid.scact_keynumber - 1;
+#ifdef SCTP_AUTH_DEACTIVATE_KEY
+			sockopt_len = sizeof(struct sctp_authkeyid);
+			ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_AUTH_DEACTIVATE_KEY,
+			      &authkeyid, sockopt_len);
+			if (ret < 0) break;
+#endif
+#ifndef SCTP_AUTHENTICATION_EVENT
+			if (authkeyid.scact_keynumber > 0)
+				{
+				authkeyid.scact_keynumber = authkeyid.scact_keynumber - 1;
+				ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_AUTH_DELETE_KEY,
+					  &authkeyid, sizeof(struct sctp_authkeyid));
+				if (ret < 0) break;
+				}
+#endif
+
+			data->ccs_rcvd = 0;
+			data->ccs_sent = 0;
+			}
+		break;
+	case BIO_CTRL_DGRAM_SCTP_GET_SNDINFO:
+		/* Returns the size of the copied struct. */
+		if (num > (long) sizeof(struct bio_dgram_sctp_sndinfo))
+			num = sizeof(struct bio_dgram_sctp_sndinfo);
+
+		memcpy(ptr, &(data->sndinfo), num);
+		ret = num;
+		break;
+	case BIO_CTRL_DGRAM_SCTP_SET_SNDINFO:
+		/* Returns the size of the copied struct. */
+		if (num > (long) sizeof(struct bio_dgram_sctp_sndinfo))
+			num = sizeof(struct bio_dgram_sctp_sndinfo);
+
+		memcpy(&(data->sndinfo), ptr, num);
+		break;
+	case BIO_CTRL_DGRAM_SCTP_GET_RCVINFO:
+		/* Returns the size of the copied struct. */
+		if (num > (long) sizeof(struct bio_dgram_sctp_rcvinfo))
+			num = sizeof(struct bio_dgram_sctp_rcvinfo);
+
+		memcpy(ptr, &data->rcvinfo, num);
+
+		ret = num;
+		break;
+	case BIO_CTRL_DGRAM_SCTP_SET_RCVINFO:
+		/* Returns the size of the copied struct. */
+		if (num > (long) sizeof(struct bio_dgram_sctp_rcvinfo))
+			num = sizeof(struct bio_dgram_sctp_rcvinfo);
+
+		memcpy(&(data->rcvinfo), ptr, num);
+		break;
+	case BIO_CTRL_DGRAM_SCTP_GET_PRINFO:
+		/* Returns the size of the copied struct. */
+		if (num > (long) sizeof(struct bio_dgram_sctp_prinfo))
+			num = sizeof(struct bio_dgram_sctp_prinfo);
+
+		memcpy(ptr, &(data->prinfo), num);
+		ret = num;
+		break;
+	case BIO_CTRL_DGRAM_SCTP_SET_PRINFO:
+		/* Returns the size of the copied struct. */
+		if (num > (long) sizeof(struct bio_dgram_sctp_prinfo))
+			num = sizeof(struct bio_dgram_sctp_prinfo);
+
+		memcpy(&(data->prinfo), ptr, num);
+		break;
+	case BIO_CTRL_DGRAM_SCTP_SAVE_SHUTDOWN:
+		/* Returns always 1. */
+		if (num > 0)
+			data->save_shutdown = 1;
+		else
+			data->save_shutdown = 0;
+		break;
+
+	default:
+		/* Pass to default ctrl function to
+		 * process SCTP unspecific commands
+		 */
+		ret=dgram_ctrl(b, cmd, num, ptr);
+		break;
+		}
+	return(ret);
+	}
+
+int BIO_dgram_sctp_notification_cb(BIO *b,
+                                   void (*handle_notifications)(BIO *bio, void *context, void *buf),
+                                   void *context)
+	{
+	bio_dgram_sctp_data *data = (bio_dgram_sctp_data *) b->ptr;
+
+	if (handle_notifications != NULL)
+		{
+		data->handle_notifications = handle_notifications;
+		data->notification_context = context;
+		}
+	else
+		return -1;
+
+	return 0;
+	}
+
+int BIO_dgram_sctp_wait_for_dry(BIO *b)
+{
+	int is_dry = 0;
+	int n, sockflags, ret;
+	union sctp_notification snp;
+	struct msghdr msg;
+	struct iovec iov;
+#ifdef SCTP_EVENT
+	struct sctp_event event;
+#else
+	struct sctp_event_subscribe event;
+	socklen_t eventsize;
+#endif
+	bio_dgram_sctp_data *data = (bio_dgram_sctp_data *)b->ptr;
+
+	/* set sender dry event */
+#ifdef SCTP_EVENT
+	memset(&event, 0, sizeof(struct sctp_event));
+	event.se_assoc_id = 0;
+	event.se_type = SCTP_SENDER_DRY_EVENT;
+	event.se_on = 1;
+	ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_EVENT, &event, sizeof(struct sctp_event));
+#else
+	eventsize = sizeof(struct sctp_event_subscribe);
+	ret = getsockopt(b->num, IPPROTO_SCTP, SCTP_EVENTS, &event, &eventsize);
+	if (ret < 0)
+		return -1;
+	
+	event.sctp_sender_dry_event = 1;
+	
+	ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_EVENTS, &event, sizeof(struct sctp_event_subscribe));
+#endif
+	if (ret < 0)
+		return -1;
+
+	/* peek for notification */
+	memset(&snp, 0x00, sizeof(union sctp_notification));
+	iov.iov_base = (char *)&snp;
+	iov.iov_len = sizeof(union sctp_notification);
+	msg.msg_name = NULL;
+	msg.msg_namelen = 0;
+	msg.msg_iov = &iov;
+	msg.msg_iovlen = 1;
+	msg.msg_control = NULL;
+	msg.msg_controllen = 0;
+	msg.msg_flags = 0;
+
+	n = recvmsg(b->num, &msg, MSG_PEEK);
+	if (n <= 0)
+		{
+		if ((n < 0) && (get_last_socket_error() != EAGAIN) && (get_last_socket_error() != EWOULDBLOCK))
+			return -1;
+		else
+			return 0;
+		}
+
+	/* if we find a notification, process it and try again if necessary */
+	while (msg.msg_flags & MSG_NOTIFICATION)
+		{
+		memset(&snp, 0x00, sizeof(union sctp_notification));
+		iov.iov_base = (char *)&snp;
+		iov.iov_len = sizeof(union sctp_notification);
+		msg.msg_name = NULL;
+		msg.msg_namelen = 0;
+		msg.msg_iov = &iov;
+		msg.msg_iovlen = 1;
+		msg.msg_control = NULL;
+		msg.msg_controllen = 0;
+		msg.msg_flags = 0;
+
+		n = recvmsg(b->num, &msg, 0);
+		if (n <= 0)
+			{
+			if ((n < 0) && (get_last_socket_error() != EAGAIN) && (get_last_socket_error() != EWOULDBLOCK))
+				return -1;
+			else
+				return is_dry;
+			}
+		
+		if (snp.sn_header.sn_type == SCTP_SENDER_DRY_EVENT)
+			{
+			is_dry = 1;
+
+			/* disable sender dry event */
+#ifdef SCTP_EVENT
+			memset(&event, 0, sizeof(struct sctp_event));
+			event.se_assoc_id = 0;
+			event.se_type = SCTP_SENDER_DRY_EVENT;
+			event.se_on = 0;
+			ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_EVENT, &event, sizeof(struct sctp_event));
+#else
+			eventsize = (socklen_t) sizeof(struct sctp_event_subscribe);
+			ret = getsockopt(b->num, IPPROTO_SCTP, SCTP_EVENTS, &event, &eventsize);
+			if (ret < 0)
+				return -1;
+
+			event.sctp_sender_dry_event = 0;
+
+			ret = setsockopt(b->num, IPPROTO_SCTP, SCTP_EVENTS, &event, sizeof(struct sctp_event_subscribe));
+#endif
+			if (ret < 0)
+				return -1;
+			}
+
+#ifdef SCTP_AUTHENTICATION_EVENT
+		if (snp.sn_header.sn_type == SCTP_AUTHENTICATION_EVENT)
+			dgram_sctp_handle_auth_free_key_event(b, &snp);
+#endif
+
+		if (data->handle_notifications != NULL)
+			data->handle_notifications(b, data->notification_context, (void*) &snp);
+
+		/* found notification, peek again */
+		memset(&snp, 0x00, sizeof(union sctp_notification));
+		iov.iov_base = (char *)&snp;
+		iov.iov_len = sizeof(union sctp_notification);
+		msg.msg_name = NULL;
+		msg.msg_namelen = 0;
+		msg.msg_iov = &iov;
+		msg.msg_iovlen = 1;
+		msg.msg_control = NULL;
+		msg.msg_controllen = 0;
+		msg.msg_flags = 0;
+
+		/* if we have seen the dry already, don't wait */
+		if (is_dry)
+			{
+			sockflags = fcntl(b->num, F_GETFL, 0);
+			fcntl(b->num, F_SETFL, O_NONBLOCK);
+			}
+
+		n = recvmsg(b->num, &msg, MSG_PEEK);
+
+		if (is_dry)
+			{
+			fcntl(b->num, F_SETFL, sockflags);
+			}
+
+		if (n <= 0)
+			{
+			if ((n < 0) && (get_last_socket_error() != EAGAIN) && (get_last_socket_error() != EWOULDBLOCK))
+				return -1;
+			else
+				return is_dry;
+			}
+		}
+
+	/* read anything else */
+	return is_dry;
+}
+
+int BIO_dgram_sctp_msg_waiting(BIO *b)
+	{
+	int n, sockflags;
+	union sctp_notification snp;
+	struct msghdr msg;
+	struct iovec iov;
+	bio_dgram_sctp_data *data = (bio_dgram_sctp_data *)b->ptr;
+
+	/* Check if there are any messages waiting to be read */
+	do
+		{
+		memset(&snp, 0x00, sizeof(union sctp_notification));
+		iov.iov_base = (char *)&snp;
+		iov.iov_len = sizeof(union sctp_notification);
+		msg.msg_name = NULL;
+		msg.msg_namelen = 0;
+		msg.msg_iov = &iov;
+		msg.msg_iovlen = 1;
+		msg.msg_control = NULL;
+		msg.msg_controllen = 0;
+		msg.msg_flags = 0;
+
+		sockflags = fcntl(b->num, F_GETFL, 0);
+		fcntl(b->num, F_SETFL, O_NONBLOCK);
+		n = recvmsg(b->num, &msg, MSG_PEEK);
+		fcntl(b->num, F_SETFL, sockflags);
+
+		/* if notification, process and try again */
+		if (n > 0 && (msg.msg_flags & MSG_NOTIFICATION))
+			{
+#ifdef SCTP_AUTHENTICATION_EVENT
+			if (snp.sn_header.sn_type == SCTP_AUTHENTICATION_EVENT)
+				dgram_sctp_handle_auth_free_key_event(b, &snp);
+#endif
+
+			memset(&snp, 0x00, sizeof(union sctp_notification));
+			iov.iov_base = (char *)&snp;
+			iov.iov_len = sizeof(union sctp_notification);
+			msg.msg_name = NULL;
+			msg.msg_namelen = 0;
+			msg.msg_iov = &iov;
+			msg.msg_iovlen = 1;
+			msg.msg_control = NULL;
+			msg.msg_controllen = 0;
+			msg.msg_flags = 0;
+			n = recvmsg(b->num, &msg, 0);
+
+			if (data->handle_notifications != NULL)
+				data->handle_notifications(b, data->notification_context, (void*) &snp);
+			}
+
+		} while (n > 0 && (msg.msg_flags & MSG_NOTIFICATION));
+
+	/* Return 1 if there is a message to be read, return 0 otherwise. */
+	if (n > 0)
+		return 1;
+	else
+		return 0;
+	}
+
+static int dgram_sctp_puts(BIO *bp, const char *str)
+	{
+	int n,ret;
+
+	n=strlen(str);
+	ret=dgram_sctp_write(bp,str,n);
+	return(ret);
+	}
+#endif
+
 static int BIO_dgram_should_retry(int i)
 	{
 	int err;
diff --git a/crypto/bio/bss_fd.c b/crypto/bio/bss_fd.c
index 4c229bf64103..d1bf85aae175 100644
--- a/crypto/bio/bss_fd.c
+++ b/crypto/bio/bss_fd.c
@@ -60,6 +60,13 @@
 #include <errno.h>
 #define USE_SOCKETS
 #include "cryptlib.h"
+
+#if defined(OPENSSL_NO_POSIX_IO)
+/*
+ * One can argue that one should implement dummy placeholder for
+ * BIO_s_fd here...
+ */
+#else
 /*
  * As for unconditional usage of "UPLINK" interface in this module.
  * Trouble is that unlike Unix file descriptors [which are indexes
@@ -77,6 +84,7 @@
 static int fd_write(BIO *h, const char *buf, int num);
 static int fd_read(BIO *h, char *buf, int size);
 static int fd_puts(BIO *h, const char *str);
+static int fd_gets(BIO *h, char *buf, int size);
 static long fd_ctrl(BIO *h, int cmd, long arg1, void *arg2);
 static int fd_new(BIO *h);
 static int fd_free(BIO *data);
@@ -88,7 +96,7 @@ static BIO_METHOD methods_fdp=
 	fd_write,
 	fd_read,
 	fd_puts,
-	NULL, /* fd_gets, */
+	fd_gets,
 	fd_ctrl,
 	fd_new,
 	fd_free,
@@ -227,6 +235,22 @@ static int fd_puts(BIO *bp, const char *str)
 	return(ret);
 	}
 
+static int fd_gets(BIO *bp, char *buf, int size)
+        {
+	int ret=0;
+	char *ptr=buf;
+	char *end=buf+size-1;
+
+	while ( (ptr < end) && (fd_read(bp, ptr, 1) > 0) && (ptr[0] != '\n') )
+		ptr++;
+
+	ptr[0]='\0';
+
+	if (buf[0] != '\0')
+		ret=strlen(buf);
+	return(ret);
+        }
+
 int BIO_fd_should_retry(int i)
 	{
 	int err;
@@ -292,3 +316,4 @@ int BIO_fd_non_fatal_error(int err)
 		}
 	return(0);
 	}
+#endif
diff --git a/crypto/bio/bss_file.c b/crypto/bio/bss_file.c
index 47fa266dfc48..b954fe7ebc8d 100644
--- a/crypto/bio/bss_file.c
+++ b/crypto/bio/bss_file.c
@@ -118,10 +118,53 @@ static BIO_METHOD methods_filep=
 
 BIO *BIO_new_file(const char *filename, const char *mode)
 	{
-	BIO *ret;
-	FILE *file;
+	BIO  *ret;
+	FILE *file=NULL;
+
+#if defined(_WIN32) && defined(CP_UTF8)
+	int sz, len_0 = (int)strlen(filename)+1;
+	DWORD flags;
 
-	if ((file=fopen(filename,mode)) == NULL)
+	/*
+	 * Basically there are three cases to cover: a) filename is
+	 * pure ASCII string; b) actual UTF-8 encoded string and
+	 * c) locale-ized string, i.e. one containing 8-bit
+	 * characters that are meaningful in current system locale.
+	 * If filename is pure ASCII or real UTF-8 encoded string,
+	 * MultiByteToWideChar succeeds and _wfopen works. If
+	 * filename is locale-ized string, chances are that
+	 * MultiByteToWideChar fails reporting
+	 * ERROR_NO_UNICODE_TRANSLATION, in which case we fall
+	 * back to fopen...
+	 */
+	if ((sz=MultiByteToWideChar(CP_UTF8,(flags=MB_ERR_INVALID_CHARS),
+					filename,len_0,NULL,0))>0 ||
+	    (GetLastError()==ERROR_INVALID_FLAGS &&
+	     (sz=MultiByteToWideChar(CP_UTF8,(flags=0),
+					filename,len_0,NULL,0))>0)
+	   )
+		{
+		WCHAR  wmode[8];
+		WCHAR *wfilename = _alloca(sz*sizeof(WCHAR));
+
+		if (MultiByteToWideChar(CP_UTF8,flags,
+					filename,len_0,wfilename,sz) &&
+		    MultiByteToWideChar(CP_UTF8,0,mode,strlen(mode)+1,
+			    		wmode,sizeof(wmode)/sizeof(wmode[0])) &&
+		    (file=_wfopen(wfilename,wmode))==NULL &&
+		    (errno==ENOENT || errno==EBADF)
+		   )	/* UTF-8 decode succeeded, but no file, filename
+			 * could still have been locale-ized... */
+			file = fopen(filename,mode);
+		}
+	else if (GetLastError()==ERROR_NO_UNICODE_TRANSLATION)
+		{
+		file = fopen(filename,mode);
+		}
+#else
+	file=fopen(filename,mode);	
+#endif
+	if (file == NULL)
 		{
 		SYSerr(SYS_F_FOPEN,get_last_sys_error());
 		ERR_add_error_data(5,"fopen('",filename,"','",mode,"')");
@@ -131,7 +174,7 @@ BIO *BIO_new_file(const char *filename, const char *mode)
 			BIOerr(BIO_F_BIO_NEW_FILE,ERR_R_SYS_LIB);
 		return(NULL);
 		}
-	if ((ret=BIO_new(BIO_s_file_internal())) == NULL)
+	if ((ret=BIO_new(BIO_s_file())) == NULL)
 		{
 		fclose(file);
 		return(NULL);
@@ -286,8 +329,7 @@ static long MS_CALLBACK file_ctrl(BIO *b, int cmd, long num, void *ptr)
 			_setmode(fd,_O_BINARY);
 #elif defined(OPENSSL_SYS_NETWARE) && defined(NETWARE_CLIB)
 		int fd = fileno((FILE*)ptr);
-         /* Under CLib there are differences in file modes
-         */
+		/* Under CLib there are differences in file modes */
 		if (num & BIO_FP_TEXT)
 			setmode(fd,O_TEXT);
 		else
@@ -308,7 +350,7 @@ static long MS_CALLBACK file_ctrl(BIO *b, int cmd, long num, void *ptr)
 			else
 				_setmode(fd,_O_BINARY);
 			}
-#elif defined(OPENSSL_SYS_OS2)
+#elif defined(OPENSSL_SYS_OS2) || defined(OPENSSL_SYS_WIN32_CYGWIN)
 		int fd = fileno((FILE*)ptr);
 		if (num & BIO_FP_TEXT)
 			setmode(fd, O_TEXT);
diff --git a/crypto/bio/bss_log.c b/crypto/bio/bss_log.c
index 6360dbc820b4..b7dce5c1a27b 100644
--- a/crypto/bio/bss_log.c
+++ b/crypto/bio/bss_log.c
@@ -70,12 +70,20 @@
 
 #if defined(OPENSSL_SYS_WINCE)
 #elif defined(OPENSSL_SYS_WIN32)
-#  include <process.h>
 #elif defined(OPENSSL_SYS_VMS)
 #  include <opcdef.h>
 #  include <descrip.h>
 #  include <lib$routines.h>
 #  include <starlet.h>
+/* Some compiler options may mask the declaration of "_malloc32". */
+#  if __INITIAL_POINTER_SIZE && defined _ANSI_C_SOURCE
+#    if __INITIAL_POINTER_SIZE == 64
+#      pragma pointer_size save
+#      pragma pointer_size 32
+    void * _malloc32  (__size_t);
+#      pragma pointer_size restore
+#    endif /* __INITIAL_POINTER_SIZE == 64 */
+#  endif /* __INITIAL_POINTER_SIZE && defined _ANSI_C_SOURCE */
 #elif defined(__ultrix)
 #  include <sys/syslog.h>
 #elif defined(OPENSSL_SYS_NETWARE)
@@ -122,18 +130,6 @@ static int MS_CALLBACK slg_free(BIO *data);
 static void xopenlog(BIO* bp, char* name, int level);
 static void xsyslog(BIO* bp, int priority, const char* string);
 static void xcloselog(BIO* bp);
-#ifdef OPENSSL_SYS_WIN32
-LONG	(WINAPI *go_for_advapi)()	= RegOpenKeyEx;
-HANDLE	(WINAPI *register_event_source)()	= NULL;
-BOOL	(WINAPI *deregister_event_source)()	= NULL;
-BOOL	(WINAPI *report_event)()	= NULL;
-#define DL_PROC(m,f)	(GetProcAddress( m, f ))
-#ifdef UNICODE
-#define DL_PROC_X(m,f) DL_PROC( m, f "W" )
-#else
-#define DL_PROC_X(m,f) DL_PROC( m, f "A" )
-#endif
-#endif
 
 static BIO_METHOD methods_slg=
 	{
@@ -175,7 +171,7 @@ static int MS_CALLBACK slg_write(BIO *b, const char *in, int inl)
 	char* buf;
 	char* pp;
 	int priority, i;
-	static struct
+	static const struct
 		{
 		int strl;
 		char str[10];
@@ -249,35 +245,20 @@ static int MS_CALLBACK slg_puts(BIO *bp, const char *str)
 
 static void xopenlog(BIO* bp, char* name, int level)
 {
-	if ( !register_event_source )
-		{
-		HANDLE	advapi;
-		if ( !(advapi = GetModuleHandle("advapi32")) )
-			return;
-		register_event_source = (HANDLE (WINAPI *)())DL_PROC_X(advapi,
-			"RegisterEventSource" );
-		deregister_event_source = (BOOL (WINAPI *)())DL_PROC(advapi,
-			"DeregisterEventSource");
-		report_event = (BOOL (WINAPI *)())DL_PROC_X(advapi,
-			"ReportEvent" );
-		if ( !(register_event_source && deregister_event_source &&
-				report_event) )
-			{
-			register_event_source = NULL;
-			deregister_event_source = NULL;
-			report_event = NULL;
-			return;
-			}
-		}
-	bp->ptr= (char *)register_event_source(NULL, name);
+	if (GetVersion() < 0x80000000)
+		bp->ptr = RegisterEventSourceA(NULL,name);
+	else
+		bp->ptr = NULL;
 }
 
 static void xsyslog(BIO *bp, int priority, const char *string)
 {
 	LPCSTR lpszStrings[2];
 	WORD evtype= EVENTLOG_ERROR_TYPE;
-	int pid = _getpid();
-	char pidbuf[DECIMAL_SIZE(pid)+4];
+	char pidbuf[DECIMAL_SIZE(DWORD)+4];
+
+	if (bp->ptr == NULL)
+		return;
 
 	switch (priority)
 		{
@@ -301,19 +282,18 @@ static void xsyslog(BIO *bp, int priority, const char *string)
 		break;
 		}
 
-	sprintf(pidbuf, "[%d] ", pid);
+	sprintf(pidbuf, "[%u] ", GetCurrentProcessId());
 	lpszStrings[0] = pidbuf;
 	lpszStrings[1] = string;
 
-	if(report_event && bp->ptr)
-		report_event(bp->ptr, evtype, 0, 1024, NULL, 2, 0,
+	ReportEventA(bp->ptr, evtype, 0, 1024, NULL, 2, 0,
 				lpszStrings, NULL);
 }
 	
 static void xcloselog(BIO* bp)
 {
-	if(deregister_event_source && bp->ptr)
-		deregister_event_source((HANDLE)(bp->ptr));
+	if(bp->ptr)
+		DeregisterEventSource((HANDLE)(bp->ptr));
 	bp->ptr= NULL;
 }
 
@@ -329,7 +309,24 @@ static void xopenlog(BIO* bp, char* name, int level)
 static void xsyslog(BIO *bp, int priority, const char *string)
 {
 	struct dsc$descriptor_s opc_dsc;
+
+/* Arrange 32-bit pointer to opcdef buffer and malloc(), if needed. */
+#if __INITIAL_POINTER_SIZE == 64
+# pragma pointer_size save
+# pragma pointer_size 32
+# define OPCDEF_TYPE __char_ptr32
+# define OPCDEF_MALLOC _malloc32
+#else /* __INITIAL_POINTER_SIZE == 64 */
+# define OPCDEF_TYPE char *
+# define OPCDEF_MALLOC OPENSSL_malloc
+#endif /* __INITIAL_POINTER_SIZE == 64 [else] */
+
 	struct opcdef *opcdef_p;
+
+#if __INITIAL_POINTER_SIZE == 64
+# pragma pointer_size restore
+#endif /* __INITIAL_POINTER_SIZE == 64 */
+
 	char buf[10240];
 	unsigned int len;
         struct dsc$descriptor_s buf_dsc;
@@ -355,8 +352,8 @@ static void xsyslog(BIO *bp, int priority, const char *string)
 
 	lib$sys_fao(&fao_cmd, &len, &buf_dsc, priority_tag, string);
 
-	/* we know there's an 8 byte header.  That's documented */
-	opcdef_p = (struct opcdef *) OPENSSL_malloc(8 + len);
+	/* We know there's an 8-byte header.  That's documented. */
+	opcdef_p = OPCDEF_MALLOC( 8+ len);
 	opcdef_p->opc$b_ms_type = OPC$_RQ_RQST;
 	memcpy(opcdef_p->opc$z_ms_target_classes, &VMS_OPC_target, 3);
 	opcdef_p->opc$l_ms_rqstid = 0;
@@ -364,7 +361,7 @@ static void xsyslog(BIO *bp, int priority, const char *string)
 
 	opc_dsc.dsc$b_dtype = DSC$K_DTYPE_T;
 	opc_dsc.dsc$b_class = DSC$K_CLASS_S;
-	opc_dsc.dsc$a_pointer = (char *)opcdef_p;
+	opc_dsc.dsc$a_pointer = (OPCDEF_TYPE) opcdef_p;
 	opc_dsc.dsc$w_length = len + 8;
 
 	sys$sndopr(opc_dsc, 0);
diff --git a/crypto/bio/bss_mem.c b/crypto/bio/bss_mem.c
index e7ab9cb3a3f8..37d4194e4bba 100644
--- a/crypto/bio/bss_mem.c
+++ b/crypto/bio/bss_mem.c
@@ -94,16 +94,18 @@ BIO *BIO_new_mem_buf(void *buf, int len)
 {
 	BIO *ret;
 	BUF_MEM *b;
+	size_t sz;
+
 	if (!buf) {
 		BIOerr(BIO_F_BIO_NEW_MEM_BUF,BIO_R_NULL_PARAMETER);
 		return NULL;
 	}
-	if(len == -1) len = strlen(buf);
+	sz = (len<0) ? strlen(buf) : (size_t)len;
 	if(!(ret = BIO_new(BIO_s_mem())) ) return NULL;
 	b = (BUF_MEM *)ret->ptr;
 	b->data = buf;
-	b->length = len;
-	b->max = len;
+	b->length = sz;
+	b->max = sz;
 	ret->flags |= BIO_FLAGS_MEM_RDONLY;
 	/* Since this is static data retrying wont help */
 	ret->num = 0;
@@ -144,22 +146,16 @@ static int mem_read(BIO *b, char *out, int outl)
 	{
 	int ret= -1;
 	BUF_MEM *bm;
-	int i;
-	char *from,*to;
 
 	bm=(BUF_MEM *)b->ptr;
 	BIO_clear_retry_flags(b);
-	ret=(outl > bm->length)?bm->length:outl;
+	ret=(outl >=0 && (size_t)outl > bm->length)?(int)bm->length:outl;
 	if ((out != NULL) && (ret > 0)) {
 		memcpy(out,bm->data,ret);
 		bm->length-=ret;
-		/* memmove(&(bm->data[0]),&(bm->data[ret]), bm->length); */
 		if(b->flags & BIO_FLAGS_MEM_RDONLY) bm->data += ret;
 		else {
-			from=(char *)&(bm->data[ret]);
-			to=(char *)&(bm->data[0]);
-			for (i=0; i<bm->length; i++)
-				to[i]=from[i];
+			memmove(&(bm->data[0]),&(bm->data[ret]),bm->length);
 		}
 	} else if (bm->length == 0)
 		{
diff --git a/crypto/bn/Makefile b/crypto/bn/Makefile
index f5e8f65a4698..672773454cfd 100644
--- a/crypto/bn/Makefile
+++ b/crypto/bn/Makefile
@@ -12,8 +12,6 @@ MAKEFILE=	Makefile
 AR=		ar r
 
 BN_ASM=		bn_asm.o
-# or use
-#BN_ASM=	bn86-elf.o
 
 CFLAGS= $(INCLUDES) $(CFLAG)
 ASFLAGS= $(INCLUDES) $(ASFLAG)
@@ -28,13 +26,13 @@ LIBSRC=	bn_add.c bn_div.c bn_exp.c bn_lib.c bn_ctx.c bn_mul.c bn_mod.c \
 	bn_print.c bn_rand.c bn_shift.c bn_word.c bn_blind.c \
 	bn_kron.c bn_sqrt.c bn_gcd.c bn_prime.c bn_err.c bn_sqr.c bn_asm.c \
 	bn_recp.c bn_mont.c bn_mpi.c bn_exp2.c bn_gf2m.c bn_nist.c \
-	bn_depr.c bn_x931p.c bn_const.c bn_opt.c
+	bn_depr.c bn_const.c bn_x931p.c
 
 LIBOBJ=	bn_add.o bn_div.o bn_exp.o bn_lib.o bn_ctx.o bn_mul.o bn_mod.o \
 	bn_print.o bn_rand.o bn_shift.o bn_word.o bn_blind.o \
 	bn_kron.o bn_sqrt.o bn_gcd.o bn_prime.o bn_err.o bn_sqr.o $(BN_ASM) \
 	bn_recp.o bn_mont.o bn_mpi.o bn_exp2.o bn_gf2m.o bn_nist.o \
-	bn_depr.o bn_x931p.o bn_const.o bn_opt.o
+	bn_depr.o bn_const.o bn_x931p.o
 
 SRC= $(LIBSRC)
 
@@ -58,36 +56,27 @@ bnbug: bnbug.c ../../libcrypto.a top
 	cc -g -I../../include bnbug.c -o bnbug ../../libcrypto.a
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-# ELF
-bn86-elf.s:	asm/bn-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) bn-586.pl elf $(CFLAGS) > ../$@)
-co86-elf.s:	asm/co-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) co-586.pl elf $(CFLAGS) > ../$@)
-mo86-elf.s:	asm/mo-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) mo-586.pl elf $(CFLAGS) > ../$@)
-# COFF
-bn86-cof.s: asm/bn-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) bn-586.pl coff $(CFLAGS) > ../$@)
-co86-cof.s: asm/co-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) co-586.pl coff $(CFLAGS) > ../$@)
-mo86-cof.s: asm/mo-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) mo-586.pl coff $(CFLAGS) > ../$@)
-# a.out
-bn86-out.s: asm/bn-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) bn-586.pl a.out $(CFLAGS) > ../$@)
-co86-out.s: asm/co-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) co-586.pl a.out $(CFLAGS) > ../$@)
-mo86-out.s: asm/mo-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) mo-586.pl a.out $(CFLAGS) > ../$@)
+bn-586.s:	asm/bn-586.pl ../perlasm/x86asm.pl
+	$(PERL) asm/bn-586.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+co-586.s:	asm/co-586.pl ../perlasm/x86asm.pl
+	$(PERL) asm/co-586.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+x86-mont.s:	asm/x86-mont.pl ../perlasm/x86asm.pl
+	$(PERL) asm/x86-mont.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+x86-gf2m.s:	asm/x86-gf2m.pl ../perlasm/x86asm.pl
+	$(PERL) asm/x86-gf2m.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
 
 sparcv8.o:	asm/sparcv8.S
 	$(CC) $(CFLAGS) -c asm/sparcv8.S
-sparcv8plus.o:	asm/sparcv8plus.S
-	$(CC) $(CFLAGS) -c asm/sparcv8plus.S
+bn-sparcv9.o:	asm/sparcv8plus.S
+	$(CC) $(CFLAGS) -c -o $@ asm/sparcv8plus.S
+sparcv9a-mont.s:	asm/sparcv9a-mont.pl
+	$(PERL) asm/sparcv9a-mont.pl $(CFLAGS) > $@
+sparcv9-mont.s:		asm/sparcv9-mont.pl
+	$(PERL) asm/sparcv9-mont.pl $(CFLAGS) > $@
 
 bn-mips3.o:	asm/mips3.s
 	@if [ "$(CC)" = "gcc" ]; then \
@@ -95,13 +84,31 @@ bn-mips3.o:	asm/mips3.s
 		as -$$ABI -O -o $@ asm/mips3.s; \
 	else	$(CC) -c $(CFLAGS) -o $@ asm/mips3.s; fi
 
+bn-mips.s:	asm/mips.pl
+	$(PERL) asm/mips.pl $(PERLASM_SCHEME) $@
+mips-mont.s:	asm/mips-mont.pl
+	$(PERL)	asm/mips-mont.pl $(PERLASM_SCHEME) $@
+
+bn-s390x.o:	asm/s390x.S
+	$(CC) $(CFLAGS) -c -o $@ asm/s390x.S
+s390x-gf2m.s:	asm/s390x-gf2m.pl
+	$(PERL) asm/s390x-gf2m.pl $(PERLASM_SCHEME) $@
+
 x86_64-gcc.o:	asm/x86_64-gcc.c
 	$(CC) $(CFLAGS) -c -o $@ asm/x86_64-gcc.c
 x86_64-mont.s:	asm/x86_64-mont.pl
-	$(PERL) asm/x86_64-mont.pl $@
+	$(PERL) asm/x86_64-mont.pl $(PERLASM_SCHEME) > $@
+x86_64-mont5.s:	asm/x86_64-mont5.pl
+	$(PERL) asm/x86_64-mont5.pl $(PERLASM_SCHEME) > $@
+x86_64-gf2m.s:	asm/x86_64-gf2m.pl
+	$(PERL) asm/x86_64-gf2m.pl $(PERLASM_SCHEME) > $@
+modexp512-x86_64.s:	asm/modexp512-x86_64.pl
+	$(PERL) asm/modexp512-x86_64.pl $(PERLASM_SCHEME) > $@
 
 bn-ia64.s:	asm/ia64.S
 	$(CC) $(CFLAGS) -E asm/ia64.S > $@
+ia64-mont.s:	asm/ia64-mont.pl
+	$(PERL) asm/ia64-mont.pl $@ $(CFLAGS)
 
 # GNU assembler fails to compile PA-RISC2 modules, insist on calling
 # vendor assembler...
@@ -109,14 +116,22 @@ pa-risc2W.o: asm/pa-risc2W.s
 	/usr/ccs/bin/as -o pa-risc2W.o asm/pa-risc2W.s
 pa-risc2.o: asm/pa-risc2.s
 	/usr/ccs/bin/as -o pa-risc2.o asm/pa-risc2.s
+parisc-mont.s:	asm/parisc-mont.pl
+	$(PERL) asm/parisc-mont.pl $(PERLASM_SCHEME) $@
 
 # ppc - AIX, Linux, MacOS X...
-linux_ppc32.s: asm/ppc.pl;	$(PERL) $< $@
-linux_ppc64.s: asm/ppc.pl;	$(PERL) $< $@
-aix_ppc32.s: asm/ppc.pl;	$(PERL) asm/ppc.pl $@
-aix_ppc64.s: asm/ppc.pl;	$(PERL) asm/ppc.pl $@
-osx_ppc32.s: asm/ppc.pl;	$(PERL) $< $@
-osx_ppc64.s: asm/ppc.pl;	$(PERL) $< $@
+bn-ppc.s:	asm/ppc.pl;	$(PERL) asm/ppc.pl $(PERLASM_SCHEME) $@
+ppc-mont.s:	asm/ppc-mont.pl;$(PERL) asm/ppc-mont.pl $(PERLASM_SCHEME) $@
+ppc64-mont.s:	asm/ppc64-mont.pl;$(PERL) asm/ppc64-mont.pl $(PERLASM_SCHEME) $@
+
+alpha-mont.s:	asm/alpha-mont.pl
+	$(PERL) $< | $(CC) -E - | tee $@ > /dev/null
+
+# GNU make "catch all"
+%-mont.s:	asm/%-mont.pl;	$(PERL) $< $(PERLASM_SCHEME) $@
+%-gf2m.S:	asm/%-gf2m.pl;	$(PERL) $< $(PERLASM_SCHEME) $@
+
+armv4-gf2m.o:	armv4-gf2m.S
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
@@ -184,8 +199,11 @@ bn_blind.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 bn_blind.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 bn_blind.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 bn_blind.o: ../../include/openssl/symhacks.h ../cryptlib.h bn_blind.c bn_lcl.h
-bn_const.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-bn_const.o: ../../include/openssl/ossl_typ.h bn.h bn_const.c
+bn_const.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+bn_const.o: ../../include/openssl/opensslconf.h
+bn_const.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+bn_const.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+bn_const.o: ../../include/openssl/symhacks.h bn.h bn_const.c
 bn_ctx.o: ../../e_os.h ../../include/openssl/bio.h ../../include/openssl/bn.h
 bn_ctx.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 bn_ctx.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
@@ -292,13 +310,6 @@ bn_nist.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 bn_nist.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 bn_nist.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 bn_nist.o: ../../include/openssl/symhacks.h ../cryptlib.h bn_lcl.h bn_nist.c
-bn_opt.o: ../../e_os.h ../../include/openssl/bio.h ../../include/openssl/bn.h
-bn_opt.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-bn_opt.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-bn_opt.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
-bn_opt.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-bn_opt.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-bn_opt.o: ../../include/openssl/symhacks.h ../cryptlib.h bn_lcl.h bn_opt.c
 bn_prime.o: ../../e_os.h ../../include/openssl/bio.h ../../include/openssl/bn.h
 bn_prime.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 bn_prime.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
@@ -357,6 +368,8 @@ bn_word.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 bn_word.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 bn_word.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 bn_word.o: ../../include/openssl/symhacks.h ../cryptlib.h bn_lcl.h bn_word.c
-bn_x931p.o: ../../include/openssl/bn.h ../../include/openssl/e_os2.h
-bn_x931p.o: ../../include/openssl/opensslconf.h
-bn_x931p.o: ../../include/openssl/ossl_typ.h bn_x931p.c
+bn_x931p.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
+bn_x931p.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
+bn_x931p.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+bn_x931p.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+bn_x931p.o: ../../include/openssl/symhacks.h bn_x931p.c
diff --git a/crypto/bn/asm/armv4-gf2m.pl b/crypto/bn/asm/armv4-gf2m.pl
new file mode 100755
index 000000000000..c52e0b75b5b6
--- /dev/null
+++ b/crypto/bn/asm/armv4-gf2m.pl
@@ -0,0 +1,278 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# May 2011
+#
+# The module implements bn_GF2m_mul_2x2 polynomial multiplication
+# used in bn_gf2m.c. It's kind of low-hanging mechanical port from
+# C for the time being... Except that it has two code paths: pure
+# integer code suitable for any ARMv4 and later CPU and NEON code
+# suitable for ARMv7. Pure integer 1x1 multiplication subroutine runs
+# in ~45 cycles on dual-issue core such as Cortex A8, which is ~50%
+# faster than compiler-generated code. For ECDH and ECDSA verify (but
+# not for ECDSA sign) it means 25%-45% improvement depending on key
+# length, more for longer keys. Even though NEON 1x1 multiplication
+# runs in even less cycles, ~30, improvement is measurable only on
+# longer keys. One has to optimize code elsewhere to get NEON glow...
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+sub Dlo()   { shift=~m|q([1]?[0-9])|?"d".($1*2):"";     }
+sub Dhi()   { shift=~m|q([1]?[0-9])|?"d".($1*2+1):"";   }
+sub Q()     { shift=~m|d([1-3]?[02468])|?"q".($1/2):""; }
+
+$code=<<___;
+#include "arm_arch.h"
+
+.text
+.code	32
+
+#if __ARM_ARCH__>=7
+.fpu	neon
+
+.type	mul_1x1_neon,%function
+.align	5
+mul_1x1_neon:
+	vshl.u64	`&Dlo("q1")`,d16,#8	@ q1-q3 are slided $a
+	vmull.p8	`&Q("d0")`,d16,d17	@ a·bb
+	vshl.u64	`&Dlo("q2")`,d16,#16
+	vmull.p8	q1,`&Dlo("q1")`,d17	@ a<<8·bb
+	vshl.u64	`&Dlo("q3")`,d16,#24
+	vmull.p8	q2,`&Dlo("q2")`,d17	@ a<<16·bb
+	vshr.u64	`&Dlo("q1")`,#8
+	vmull.p8	q3,`&Dlo("q3")`,d17	@ a<<24·bb
+	vshl.u64	`&Dhi("q1")`,#24
+	veor		d0,`&Dlo("q1")`
+	vshr.u64	`&Dlo("q2")`,#16
+	veor		d0,`&Dhi("q1")`
+	vshl.u64	`&Dhi("q2")`,#16
+	veor		d0,`&Dlo("q2")`
+	vshr.u64	`&Dlo("q3")`,#24
+	veor		d0,`&Dhi("q2")`
+	vshl.u64	`&Dhi("q3")`,#8
+	veor		d0,`&Dlo("q3")`
+	veor		d0,`&Dhi("q3")`
+	bx	lr
+.size	mul_1x1_neon,.-mul_1x1_neon
+#endif
+___
+################
+# private interface to mul_1x1_ialu
+#
+$a="r1";
+$b="r0";
+
+($a0,$a1,$a2,$a12,$a4,$a14)=
+($hi,$lo,$t0,$t1, $i0,$i1 )=map("r$_",(4..9),12);
+
+$mask="r12";
+
+$code.=<<___;
+.type	mul_1x1_ialu,%function
+.align	5
+mul_1x1_ialu:
+	mov	$a0,#0
+	bic	$a1,$a,#3<<30		@ a1=a&0x3fffffff
+	str	$a0,[sp,#0]		@ tab[0]=0
+	add	$a2,$a1,$a1		@ a2=a1<<1
+	str	$a1,[sp,#4]		@ tab[1]=a1
+	eor	$a12,$a1,$a2		@ a1^a2
+	str	$a2,[sp,#8]		@ tab[2]=a2
+	mov	$a4,$a1,lsl#2		@ a4=a1<<2
+	str	$a12,[sp,#12]		@ tab[3]=a1^a2
+	eor	$a14,$a1,$a4		@ a1^a4
+	str	$a4,[sp,#16]		@ tab[4]=a4
+	eor	$a0,$a2,$a4		@ a2^a4
+	str	$a14,[sp,#20]		@ tab[5]=a1^a4
+	eor	$a12,$a12,$a4		@ a1^a2^a4
+	str	$a0,[sp,#24]		@ tab[6]=a2^a4
+	and	$i0,$mask,$b,lsl#2
+	str	$a12,[sp,#28]		@ tab[7]=a1^a2^a4
+
+	and	$i1,$mask,$b,lsr#1
+	ldr	$lo,[sp,$i0]		@ tab[b       & 0x7]
+	and	$i0,$mask,$b,lsr#4
+	ldr	$t1,[sp,$i1]		@ tab[b >>  3 & 0x7]
+	and	$i1,$mask,$b,lsr#7
+	ldr	$t0,[sp,$i0]		@ tab[b >>  6 & 0x7]
+	eor	$lo,$lo,$t1,lsl#3	@ stall
+	mov	$hi,$t1,lsr#29
+	ldr	$t1,[sp,$i1]		@ tab[b >>  9 & 0x7]
+
+	and	$i0,$mask,$b,lsr#10
+	eor	$lo,$lo,$t0,lsl#6
+	eor	$hi,$hi,$t0,lsr#26
+	ldr	$t0,[sp,$i0]		@ tab[b >> 12 & 0x7]
+
+	and	$i1,$mask,$b,lsr#13
+	eor	$lo,$lo,$t1,lsl#9
+	eor	$hi,$hi,$t1,lsr#23
+	ldr	$t1,[sp,$i1]		@ tab[b >> 15 & 0x7]
+
+	and	$i0,$mask,$b,lsr#16
+	eor	$lo,$lo,$t0,lsl#12
+	eor	$hi,$hi,$t0,lsr#20
+	ldr	$t0,[sp,$i0]		@ tab[b >> 18 & 0x7]
+
+	and	$i1,$mask,$b,lsr#19
+	eor	$lo,$lo,$t1,lsl#15
+	eor	$hi,$hi,$t1,lsr#17
+	ldr	$t1,[sp,$i1]		@ tab[b >> 21 & 0x7]
+
+	and	$i0,$mask,$b,lsr#22
+	eor	$lo,$lo,$t0,lsl#18
+	eor	$hi,$hi,$t0,lsr#14
+	ldr	$t0,[sp,$i0]		@ tab[b >> 24 & 0x7]
+
+	and	$i1,$mask,$b,lsr#25
+	eor	$lo,$lo,$t1,lsl#21
+	eor	$hi,$hi,$t1,lsr#11
+	ldr	$t1,[sp,$i1]		@ tab[b >> 27 & 0x7]
+
+	tst	$a,#1<<30
+	and	$i0,$mask,$b,lsr#28
+	eor	$lo,$lo,$t0,lsl#24
+	eor	$hi,$hi,$t0,lsr#8
+	ldr	$t0,[sp,$i0]		@ tab[b >> 30      ]
+
+	eorne	$lo,$lo,$b,lsl#30
+	eorne	$hi,$hi,$b,lsr#2
+	tst	$a,#1<<31
+	eor	$lo,$lo,$t1,lsl#27
+	eor	$hi,$hi,$t1,lsr#5
+	eorne	$lo,$lo,$b,lsl#31
+	eorne	$hi,$hi,$b,lsr#1
+	eor	$lo,$lo,$t0,lsl#30
+	eor	$hi,$hi,$t0,lsr#2
+
+	mov	pc,lr
+.size	mul_1x1_ialu,.-mul_1x1_ialu
+___
+################
+# void	bn_GF2m_mul_2x2(BN_ULONG *r,
+#	BN_ULONG a1,BN_ULONG a0,
+#	BN_ULONG b1,BN_ULONG b0);	# r[3..0]=a1a0·b1b0
+
+($A1,$B1,$A0,$B0,$A1B1,$A0B0)=map("d$_",(18..23));
+
+$code.=<<___;
+.global	bn_GF2m_mul_2x2
+.type	bn_GF2m_mul_2x2,%function
+.align	5
+bn_GF2m_mul_2x2:
+#if __ARM_ARCH__>=7
+	ldr	r12,.LOPENSSL_armcap
+.Lpic:	ldr	r12,[pc,r12]
+	tst	r12,#1
+	beq	.Lialu
+
+	veor	$A1,$A1
+	vmov.32	$B1,r3,r3		@ two copies of b1
+	vmov.32	${A1}[0],r1		@ a1
+
+	veor	$A0,$A0
+	vld1.32	${B0}[],[sp,:32]	@ two copies of b0
+	vmov.32	${A0}[0],r2		@ a0
+	mov	r12,lr
+
+	vmov	d16,$A1
+	vmov	d17,$B1
+	bl	mul_1x1_neon		@ a1·b1
+	vmov	$A1B1,d0
+
+	vmov	d16,$A0
+	vmov	d17,$B0
+	bl	mul_1x1_neon		@ a0·b0
+	vmov	$A0B0,d0
+
+	veor	d16,$A0,$A1
+	veor	d17,$B0,$B1
+	veor	$A0,$A0B0,$A1B1
+	bl	mul_1x1_neon		@ (a0+a1)·(b0+b1)
+
+	veor	d0,$A0			@ (a0+a1)·(b0+b1)-a0·b0-a1·b1
+	vshl.u64 d1,d0,#32
+	vshr.u64 d0,d0,#32
+	veor	$A0B0,d1
+	veor	$A1B1,d0
+	vst1.32	{${A0B0}[0]},[r0,:32]!
+	vst1.32	{${A0B0}[1]},[r0,:32]!
+	vst1.32	{${A1B1}[0]},[r0,:32]!
+	vst1.32	{${A1B1}[1]},[r0,:32]
+	bx	r12
+.align	4
+.Lialu:
+#endif
+___
+$ret="r10";	# reassigned 1st argument
+$code.=<<___;
+	stmdb	sp!,{r4-r10,lr}
+	mov	$ret,r0			@ reassign 1st argument
+	mov	$b,r3			@ $b=b1
+	ldr	r3,[sp,#32]		@ load b0
+	mov	$mask,#7<<2
+	sub	sp,sp,#32		@ allocate tab[8]
+
+	bl	mul_1x1_ialu		@ a1·b1
+	str	$lo,[$ret,#8]
+	str	$hi,[$ret,#12]
+
+	eor	$b,$b,r3		@ flip b0 and b1
+	 eor	$a,$a,r2		@ flip a0 and a1
+	eor	r3,r3,$b
+	 eor	r2,r2,$a
+	eor	$b,$b,r3
+	 eor	$a,$a,r2
+	bl	mul_1x1_ialu		@ a0·b0
+	str	$lo,[$ret]
+	str	$hi,[$ret,#4]
+
+	eor	$a,$a,r2
+	eor	$b,$b,r3
+	bl	mul_1x1_ialu		@ (a1+a0)·(b1+b0)
+___
+@r=map("r$_",(6..9));
+$code.=<<___;
+	ldmia	$ret,{@r[0]-@r[3]}
+	eor	$lo,$lo,$hi
+	eor	$hi,$hi,@r[1]
+	eor	$lo,$lo,@r[0]
+	eor	$hi,$hi,@r[2]
+	eor	$lo,$lo,@r[3]
+	eor	$hi,$hi,@r[3]
+	str	$hi,[$ret,#8]
+	eor	$lo,$lo,$hi
+	add	sp,sp,#32		@ destroy tab[8]
+	str	$lo,[$ret,#4]
+
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r10,pc}
+#else
+	ldmia	sp!,{r4-r10,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.size	bn_GF2m_mul_2x2,.-bn_GF2m_mul_2x2
+#if __ARM_ARCH__>=7
+.align	5
+.LOPENSSL_armcap:
+.word	OPENSSL_armcap_P-(.Lpic+8)
+#endif
+.asciz	"GF(2^m) Multiplication for ARMv4/NEON, CRYPTOGAMS by <appro\@openssl.org>"
+.align	5
+
+.comm	OPENSSL_armcap_P,4,4
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;    # make it possible to compile with -march=armv4
+print $code;
+close STDOUT;   # enforce flush
diff --git a/crypto/bn/asm/armv4-mont.pl b/crypto/bn/asm/armv4-mont.pl
new file mode 100755
index 000000000000..f78a8b5f0f55
--- /dev/null
+++ b/crypto/bn/asm/armv4-mont.pl
@@ -0,0 +1,204 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# January 2007.
+
+# Montgomery multiplication for ARMv4.
+#
+# Performance improvement naturally varies among CPU implementations
+# and compilers. The code was observed to provide +65-35% improvement
+# [depending on key length, less for longer keys] on ARM920T, and
+# +115-80% on Intel IXP425. This is compared to pre-bn_mul_mont code
+# base and compiler generated code with in-lined umull and even umlal
+# instructions. The latter means that this code didn't really have an 
+# "advantage" of utilizing some "secret" instruction.
+#
+# The code is interoperable with Thumb ISA and is rather compact, less
+# than 1/2KB. Windows CE port would be trivial, as it's exclusively
+# about decorations, ABI and instruction syntax are identical.
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$num="r0";	# starts as num argument, but holds &tp[num-1]
+$ap="r1";
+$bp="r2"; $bi="r2"; $rp="r2";
+$np="r3";
+$tp="r4";
+$aj="r5";
+$nj="r6";
+$tj="r7";
+$n0="r8";
+###########	# r9 is reserved by ELF as platform specific, e.g. TLS pointer
+$alo="r10";	# sl, gcc uses it to keep @GOT
+$ahi="r11";	# fp
+$nlo="r12";	# ip
+###########	# r13 is stack pointer
+$nhi="r14";	# lr
+###########	# r15 is program counter
+
+#### argument block layout relative to &tp[num-1], a.k.a. $num
+$_rp="$num,#12*4";
+# ap permanently resides in r1
+$_bp="$num,#13*4";
+# np permanently resides in r3
+$_n0="$num,#14*4";
+$_num="$num,#15*4";	$_bpend=$_num;
+
+$code=<<___;
+.text
+
+.global	bn_mul_mont
+.type	bn_mul_mont,%function
+
+.align	2
+bn_mul_mont:
+	stmdb	sp!,{r0,r2}		@ sp points at argument block
+	ldr	$num,[sp,#3*4]		@ load num
+	cmp	$num,#2
+	movlt	r0,#0
+	addlt	sp,sp,#2*4
+	blt	.Labrt
+
+	stmdb	sp!,{r4-r12,lr}		@ save 10 registers
+
+	mov	$num,$num,lsl#2		@ rescale $num for byte count
+	sub	sp,sp,$num		@ alloca(4*num)
+	sub	sp,sp,#4		@ +extra dword
+	sub	$num,$num,#4		@ "num=num-1"
+	add	$tp,$bp,$num		@ &bp[num-1]
+
+	add	$num,sp,$num		@ $num to point at &tp[num-1]
+	ldr	$n0,[$_n0]		@ &n0
+	ldr	$bi,[$bp]		@ bp[0]
+	ldr	$aj,[$ap],#4		@ ap[0],ap++
+	ldr	$nj,[$np],#4		@ np[0],np++
+	ldr	$n0,[$n0]		@ *n0
+	str	$tp,[$_bpend]		@ save &bp[num]
+
+	umull	$alo,$ahi,$aj,$bi	@ ap[0]*bp[0]
+	str	$n0,[$_n0]		@ save n0 value
+	mul	$n0,$alo,$n0		@ "tp[0]"*n0
+	mov	$nlo,#0
+	umlal	$alo,$nlo,$nj,$n0	@ np[0]*n0+"t[0]"
+	mov	$tp,sp
+
+.L1st:
+	ldr	$aj,[$ap],#4		@ ap[j],ap++
+	mov	$alo,$ahi
+	ldr	$nj,[$np],#4		@ np[j],np++
+	mov	$ahi,#0
+	umlal	$alo,$ahi,$aj,$bi	@ ap[j]*bp[0]
+	mov	$nhi,#0
+	umlal	$nlo,$nhi,$nj,$n0	@ np[j]*n0
+	adds	$nlo,$nlo,$alo
+	str	$nlo,[$tp],#4		@ tp[j-1]=,tp++
+	adc	$nlo,$nhi,#0
+	cmp	$tp,$num
+	bne	.L1st
+
+	adds	$nlo,$nlo,$ahi
+	ldr	$tp,[$_bp]		@ restore bp
+	mov	$nhi,#0
+	ldr	$n0,[$_n0]		@ restore n0
+	adc	$nhi,$nhi,#0
+	str	$nlo,[$num]		@ tp[num-1]=
+	str	$nhi,[$num,#4]		@ tp[num]=
+
+.Louter:
+	sub	$tj,$num,sp		@ "original" $num-1 value
+	sub	$ap,$ap,$tj		@ "rewind" ap to &ap[1]
+	ldr	$bi,[$tp,#4]!		@ *(++bp)
+	sub	$np,$np,$tj		@ "rewind" np to &np[1]
+	ldr	$aj,[$ap,#-4]		@ ap[0]
+	ldr	$alo,[sp]		@ tp[0]
+	ldr	$nj,[$np,#-4]		@ np[0]
+	ldr	$tj,[sp,#4]		@ tp[1]
+
+	mov	$ahi,#0
+	umlal	$alo,$ahi,$aj,$bi	@ ap[0]*bp[i]+tp[0]
+	str	$tp,[$_bp]		@ save bp
+	mul	$n0,$alo,$n0
+	mov	$nlo,#0
+	umlal	$alo,$nlo,$nj,$n0	@ np[0]*n0+"tp[0]"
+	mov	$tp,sp
+
+.Linner:
+	ldr	$aj,[$ap],#4		@ ap[j],ap++
+	adds	$alo,$ahi,$tj		@ +=tp[j]
+	ldr	$nj,[$np],#4		@ np[j],np++
+	mov	$ahi,#0
+	umlal	$alo,$ahi,$aj,$bi	@ ap[j]*bp[i]
+	mov	$nhi,#0
+	umlal	$nlo,$nhi,$nj,$n0	@ np[j]*n0
+	adc	$ahi,$ahi,#0
+	ldr	$tj,[$tp,#8]		@ tp[j+1]
+	adds	$nlo,$nlo,$alo
+	str	$nlo,[$tp],#4		@ tp[j-1]=,tp++
+	adc	$nlo,$nhi,#0
+	cmp	$tp,$num
+	bne	.Linner
+
+	adds	$nlo,$nlo,$ahi
+	mov	$nhi,#0
+	ldr	$tp,[$_bp]		@ restore bp
+	adc	$nhi,$nhi,#0
+	ldr	$n0,[$_n0]		@ restore n0
+	adds	$nlo,$nlo,$tj
+	ldr	$tj,[$_bpend]		@ restore &bp[num]
+	adc	$nhi,$nhi,#0
+	str	$nlo,[$num]		@ tp[num-1]=
+	str	$nhi,[$num,#4]		@ tp[num]=
+
+	cmp	$tp,$tj
+	bne	.Louter
+
+	ldr	$rp,[$_rp]		@ pull rp
+	add	$num,$num,#4		@ $num to point at &tp[num]
+	sub	$aj,$num,sp		@ "original" num value
+	mov	$tp,sp			@ "rewind" $tp
+	mov	$ap,$tp			@ "borrow" $ap
+	sub	$np,$np,$aj		@ "rewind" $np to &np[0]
+
+	subs	$tj,$tj,$tj		@ "clear" carry flag
+.Lsub:	ldr	$tj,[$tp],#4
+	ldr	$nj,[$np],#4
+	sbcs	$tj,$tj,$nj		@ tp[j]-np[j]
+	str	$tj,[$rp],#4		@ rp[j]=
+	teq	$tp,$num		@ preserve carry
+	bne	.Lsub
+	sbcs	$nhi,$nhi,#0		@ upmost carry
+	mov	$tp,sp			@ "rewind" $tp
+	sub	$rp,$rp,$aj		@ "rewind" $rp
+
+	and	$ap,$tp,$nhi
+	bic	$np,$rp,$nhi
+	orr	$ap,$ap,$np		@ ap=borrow?tp:rp
+
+.Lcopy:	ldr	$tj,[$ap],#4		@ copy or in-place refresh
+	str	sp,[$tp],#4		@ zap tp
+	str	$tj,[$rp],#4
+	cmp	$tp,$num
+	bne	.Lcopy
+
+	add	sp,$num,#4		@ skip over tp[num+1]
+	ldmia	sp!,{r4-r12,lr}		@ restore registers
+	add	sp,sp,#2*4		@ skip over {r0,r2}
+	mov	r0,#1
+.Labrt:	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+.size	bn_mul_mont,.-bn_mul_mont
+.asciz	"Montgomery multiplication for ARMv4, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+___
+
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/bn-586.pl b/crypto/bn/asm/bn-586.pl
index 26c2685a726e..332ef3e91d62 100644
--- a/crypto/bn/asm/bn-586.pl
+++ b/crypto/bn/asm/bn-586.pl
@@ -1,6 +1,7 @@
 #!/usr/local/bin/perl
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],$0);
@@ -24,38 +25,25 @@ sub bn_mul_add_words
 	{
 	local($name)=@_;
 
-	&function_begin($name,$sse2?"EXTRN\t_OPENSSL_ia32cap_P:DWORD":"");
+	&function_begin_B($name,$sse2?"EXTRN\t_OPENSSL_ia32cap_P:DWORD":"");
 
-	&comment("");
-	$Low="eax";
-	$High="edx";
-	$a="ebx";
-	$w="ebp";
-	$r="edi";
-	$c="esi";
-
-	&xor($c,$c);		# clear carry
-	&mov($r,&wparam(0));	#
-
-	&mov("ecx",&wparam(2));	#
-	&mov($a,&wparam(1));	#
-
-	&and("ecx",0xfffffff8);	# num / 8
-	&mov($w,&wparam(3));	#
-
-	&push("ecx");		# Up the stack for a tmp variable
-
-	&jz(&label("maw_finish"));
+	$r="eax";
+	$a="edx";
+	$c="ecx";
 
 	if ($sse2) {
 		&picmeup("eax","OPENSSL_ia32cap_P");
 		&bt(&DWP(0,"eax"),26);
-		&jnc(&label("maw_loop"));
+		&jnc(&label("maw_non_sse2"));
 
-		&movd("mm0",$w);		# mm0 = w
+		&mov($r,&wparam(0));
+		&mov($a,&wparam(1));
+		&mov($c,&wparam(2));
+		&movd("mm0",&wparam(3));	# mm0 = w
 		&pxor("mm1","mm1");		# mm1 = carry_in
-
-		&set_label("maw_sse2_loop",0);
+		&jmp(&label("maw_sse2_entry"));
+		
+	&set_label("maw_sse2_unrolled",16);
 		&movd("mm3",&DWP(0,$r,"",0));	# mm3 = r[0]
 		&paddq("mm1","mm3");		# mm1 = carry_in + r[0]
 		&movd("mm2",&DWP(0,$a,"",0));	# mm2 = a[0]
@@ -112,42 +100,82 @@ sub bn_mul_add_words
 		&psrlq("mm1",32);		# mm1 = carry6
 		&paddq("mm1","mm3");		# mm1 = carry6 + r[7] + w*a[7]
 		&movd(&DWP(28,$r,"",0),"mm1");
-		&add($r,32);
+		&lea($r,&DWP(32,$r));
 		&psrlq("mm1",32);		# mm1 = carry_out
 
-		&sub("ecx",8);
+		&sub($c,8);
+		&jz(&label("maw_sse2_exit"));
+	&set_label("maw_sse2_entry");
+		&test($c,0xfffffff8);
+		&jnz(&label("maw_sse2_unrolled"));
+
+	&set_label("maw_sse2_loop",4);
+		&movd("mm2",&DWP(0,$a));	# mm2 = a[i]
+		&movd("mm3",&DWP(0,$r));	# mm3 = r[i]
+		&pmuludq("mm2","mm0");		# a[i] *= w
+		&lea($a,&DWP(4,$a));
+		&paddq("mm1","mm3");		# carry += r[i]
+		&paddq("mm1","mm2");		# carry += a[i]*w
+		&movd(&DWP(0,$r),"mm1");	# r[i] = carry_low
+		&sub($c,1);
+		&psrlq("mm1",32);		# carry = carry_high
+		&lea($r,&DWP(4,$r));
 		&jnz(&label("maw_sse2_loop"));
-
-		&movd($c,"mm1");		# c = carry_out
+	&set_label("maw_sse2_exit");
+		&movd("eax","mm1");		# c = carry_out
 		&emms();
+		&ret();
 
-		&jmp(&label("maw_finish"));
+	&set_label("maw_non_sse2",16);
 	}
 
-	&set_label("maw_loop",0);
+	# function_begin prologue
+	&push("ebp");
+	&push("ebx");
+	&push("esi");
+	&push("edi");
+
+	&comment("");
+	$Low="eax";
+	$High="edx";
+	$a="ebx";
+	$w="ebp";
+	$r="edi";
+	$c="esi";
+
+	&xor($c,$c);		# clear carry
+	&mov($r,&wparam(0));	#
+
+	&mov("ecx",&wparam(2));	#
+	&mov($a,&wparam(1));	#
+
+	&and("ecx",0xfffffff8);	# num / 8
+	&mov($w,&wparam(3));	#
 
-	&mov(&swtmp(0),"ecx");	#
+	&push("ecx");		# Up the stack for a tmp variable
+
+	&jz(&label("maw_finish"));
+
+	&set_label("maw_loop",16);
 
 	for ($i=0; $i<32; $i+=4)
 		{
 		&comment("Round $i");
 
-		 &mov("eax",&DWP($i,$a,"",0)); 	# *a
+		 &mov("eax",&DWP($i,$a)); 	# *a
 		&mul($w);			# *a * w
-		&add("eax",$c);		# L(t)+= *r
-		 &mov($c,&DWP($i,$r,"",0));	# L(t)+= *r
+		&add("eax",$c);			# L(t)+= c
 		&adc("edx",0);			# H(t)+=carry
-		 &add("eax",$c);		# L(t)+=c
+		 &add("eax",&DWP($i,$r));	# L(t)+= *r
 		&adc("edx",0);			# H(t)+=carry
-		 &mov(&DWP($i,$r,"",0),"eax");	# *r= L(t);
+		 &mov(&DWP($i,$r),"eax");	# *r= L(t);
 		&mov($c,"edx");			# c=  H(t);
 		}
 
 	&comment("");
-	&mov("ecx",&swtmp(0));	#
-	&add($a,32);
-	&add($r,32);
 	&sub("ecx",8);
+	&lea($a,&DWP(32,$a));
+	&lea($r,&DWP(32,$r));
 	&jnz(&label("maw_loop"));
 
 	&set_label("maw_finish",0);
@@ -160,16 +188,15 @@ sub bn_mul_add_words
 	for ($i=0; $i<7; $i++)
 		{
 		&comment("Tail Round $i");
-		 &mov("eax",&DWP($i*4,$a,"",0));# *a
+		 &mov("eax",&DWP($i*4,$a));	# *a
 		&mul($w);			# *a * w
 		&add("eax",$c);			# L(t)+=c
-		 &mov($c,&DWP($i*4,$r,"",0));	# L(t)+= *r
 		&adc("edx",0);			# H(t)+=carry
-		 &add("eax",$c);
+		 &add("eax",&DWP($i*4,$r));	# L(t)+= *r
 		&adc("edx",0);			# H(t)+=carry
 		 &dec("ecx") if ($i != 7-1);
-		&mov(&DWP($i*4,$r,"",0),"eax");	# *r= L(t);
-		 &mov($c,"edx");			# c=  H(t);
+		&mov(&DWP($i*4,$r),"eax");	# *r= L(t);
+		 &mov($c,"edx");		# c=  H(t);
 		&jz(&label("maw_end")) if ($i != 7-1);
 		}
 	&set_label("maw_end",0);
@@ -184,7 +211,45 @@ sub bn_mul_words
 	{
 	local($name)=@_;
 
-	&function_begin($name,"");
+	&function_begin_B($name,$sse2?"EXTRN\t_OPENSSL_ia32cap_P:DWORD":"");
+
+	$r="eax";
+	$a="edx";
+	$c="ecx";
+
+	if ($sse2) {
+		&picmeup("eax","OPENSSL_ia32cap_P");
+		&bt(&DWP(0,"eax"),26);
+		&jnc(&label("mw_non_sse2"));
+
+		&mov($r,&wparam(0));
+		&mov($a,&wparam(1));
+		&mov($c,&wparam(2));
+		&movd("mm0",&wparam(3));	# mm0 = w
+		&pxor("mm1","mm1");		# mm1 = carry = 0
+
+	&set_label("mw_sse2_loop",16);
+		&movd("mm2",&DWP(0,$a));	# mm2 = a[i]
+		&pmuludq("mm2","mm0");		# a[i] *= w
+		&lea($a,&DWP(4,$a));
+		&paddq("mm1","mm2");		# carry += a[i]*w
+		&movd(&DWP(0,$r),"mm1");	# r[i] = carry_low
+		&sub($c,1);
+		&psrlq("mm1",32);		# carry = carry_high
+		&lea($r,&DWP(4,$r));
+		&jnz(&label("mw_sse2_loop"));
+
+		&movd("eax","mm1");		# return carry
+		&emms();
+		&ret();
+	&set_label("mw_non_sse2",16);
+	}
+
+	# function_begin prologue
+	&push("ebp");
+	&push("ebx");
+	&push("esi");
+	&push("edi");
 
 	&comment("");
 	$Low="eax";
@@ -257,7 +322,40 @@ sub bn_sqr_words
 	{
 	local($name)=@_;
 
-	&function_begin($name,"");
+	&function_begin_B($name,$sse2?"EXTRN\t_OPENSSL_ia32cap_P:DWORD":"");
+
+	$r="eax";
+	$a="edx";
+	$c="ecx";
+
+	if ($sse2) {
+		&picmeup("eax","OPENSSL_ia32cap_P");
+		&bt(&DWP(0,"eax"),26);
+		&jnc(&label("sqr_non_sse2"));
+
+		&mov($r,&wparam(0));
+		&mov($a,&wparam(1));
+		&mov($c,&wparam(2));
+
+	&set_label("sqr_sse2_loop",16);
+		&movd("mm0",&DWP(0,$a));	# mm0 = a[i]
+		&pmuludq("mm0","mm0");		# a[i] *= a[i]
+		&lea($a,&DWP(4,$a));		# a++
+		&movq(&QWP(0,$r),"mm0");	# r[i] = a[i]*a[i]
+		&sub($c,1);
+		&lea($r,&DWP(8,$r));		# r += 2
+		&jnz(&label("sqr_sse2_loop"));
+
+		&emms();
+		&ret();
+	&set_label("sqr_non_sse2",16);
+	}
+
+	# function_begin prologue
+	&push("ebp");
+	&push("ebx");
+	&push("esi");
+	&push("edi");
 
 	&comment("");
 	$r="esi";
@@ -313,12 +411,13 @@ sub bn_div_words
 	{
 	local($name)=@_;
 
-	&function_begin($name,"");
+	&function_begin_B($name,"");
 	&mov("edx",&wparam(0));	#
 	&mov("eax",&wparam(1));	#
-	&mov("ebx",&wparam(2));	#
-	&div("ebx");
-	&function_end($name);
+	&mov("ecx",&wparam(2));	#
+	&div("ecx");
+	&ret();
+	&function_end_B($name);
 	}
 
 sub bn_add_words
diff --git a/crypto/bn/asm/co-586.pl b/crypto/bn/asm/co-586.pl
index 5d962cb957d3..57101a6bd775 100644
--- a/crypto/bn/asm/co-586.pl
+++ b/crypto/bn/asm/co-586.pl
@@ -1,6 +1,7 @@
 #!/usr/local/bin/perl
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],$0);
diff --git a/crypto/bn/asm/ia64-mont.pl b/crypto/bn/asm/ia64-mont.pl
new file mode 100755
index 000000000000..e258658428a3
--- /dev/null
+++ b/crypto/bn/asm/ia64-mont.pl
@@ -0,0 +1,851 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# January 2010
+#
+# "Teaser" Montgomery multiplication module for IA-64. There are
+# several possibilities for improvement:
+#
+# - modulo-scheduling outer loop would eliminate quite a number of
+#   stalls after ldf8, xma and getf.sig outside inner loop and
+#   improve shorter key performance;
+# - shorter vector support [with input vectors being fetched only
+#   once] should be added;
+# - 2x unroll with help of n0[1] would make the code scalable on
+#   "wider" IA-64, "wider" than Itanium 2 that is, which is not of
+#   acute interest, because upcoming Tukwila's individual cores are
+#   reportedly based on Itanium 2 design;
+# - dedicated squaring procedure(?);
+#
+# January 2010
+#
+# Shorter vector support is implemented by zero-padding ap and np
+# vectors up to 8 elements, or 512 bits. This means that 256-bit
+# inputs will be processed only 2 times faster than 512-bit inputs,
+# not 4 [as one would expect, because algorithm complexity is n^2].
+# The reason for padding is that inputs shorter than 512 bits won't
+# be processed faster anyway, because minimal critical path of the
+# core loop happens to match 512-bit timing. Either way, it resulted
+# in >100% improvement of 512-bit RSA sign benchmark and 50% - of
+# 1024-bit one [in comparison to original version of *this* module].
+#
+# So far 'openssl speed rsa dsa' output on 900MHz Itanium 2 *with*
+# this module is:
+#                   sign    verify    sign/s verify/s
+# rsa  512 bits 0.000290s 0.000024s   3452.8  42031.4
+# rsa 1024 bits 0.000793s 0.000058s   1261.7  17172.0
+# rsa 2048 bits 0.005908s 0.000148s    169.3   6754.0
+# rsa 4096 bits 0.033456s 0.000469s     29.9   2133.6
+# dsa  512 bits 0.000253s 0.000198s   3949.9   5057.0
+# dsa 1024 bits 0.000585s 0.000607s   1708.4   1647.4
+# dsa 2048 bits 0.001453s 0.001703s    688.1    587.4
+#
+# ... and *without* (but still with ia64.S):
+#
+# rsa  512 bits 0.000670s 0.000041s   1491.8  24145.5
+# rsa 1024 bits 0.001988s 0.000080s    502.9  12499.3
+# rsa 2048 bits 0.008702s 0.000189s    114.9   5293.9
+# rsa 4096 bits 0.043860s 0.000533s     22.8   1875.9
+# dsa  512 bits 0.000441s 0.000427s   2265.3   2340.6
+# dsa 1024 bits 0.000823s 0.000867s   1215.6   1153.2
+# dsa 2048 bits 0.001894s 0.002179s    528.1    458.9
+#
+# As it can be seen, RSA sign performance improves by 130-30%,
+# hereafter less for longer keys, while verify - by 74-13%.
+# DSA performance improves by 115-30%.
+
+if ($^O eq "hpux") {
+    $ADDP="addp4";
+    for (@ARGV) { $ADDP="add" if (/[\+DD|\-mlp]64/); }
+} else { $ADDP="add"; }
+
+$code=<<___;
+.explicit
+.text
+
+// int bn_mul_mont (BN_ULONG *rp,const BN_ULONG *ap,
+//		    const BN_ULONG *bp,const BN_ULONG *np,
+//		    const BN_ULONG *n0p,int num);			
+.align	64
+.global	bn_mul_mont#
+.proc	bn_mul_mont#
+bn_mul_mont:
+	.prologue
+	.body
+{ .mmi;	cmp4.le		p6,p7=2,r37;;
+(p6)	cmp4.lt.unc	p8,p9=8,r37
+	mov		ret0=r0		};;
+{ .bbb;
+(p9)	br.cond.dptk.many	bn_mul_mont_8
+(p8)	br.cond.dpnt.many	bn_mul_mont_general
+(p7)	br.ret.spnt.many	b0	};;
+.endp	bn_mul_mont#
+
+prevfs=r2;	prevpr=r3;	prevlc=r10;	prevsp=r11;
+
+rptr=r8;	aptr=r9;	bptr=r14;	nptr=r15;
+tptr=r16;	// &tp[0]
+tp_1=r17;	// &tp[-1]
+num=r18;	len=r19;	lc=r20;
+topbit=r21;	// carry bit from tmp[num]
+
+n0=f6;
+m0=f7;
+bi=f8;
+
+.align	64
+.local	bn_mul_mont_general#
+.proc	bn_mul_mont_general#
+bn_mul_mont_general:
+	.prologue
+{ .mmi;	.save	ar.pfs,prevfs
+	alloc	prevfs=ar.pfs,6,2,0,8
+	$ADDP	aptr=0,in1
+	.save	ar.lc,prevlc
+	mov	prevlc=ar.lc		}
+{ .mmi;	.vframe	prevsp
+	mov	prevsp=sp
+	$ADDP	bptr=0,in2
+	.save	pr,prevpr
+	mov	prevpr=pr		};;
+
+	.body
+	.rotf		alo[6],nlo[4],ahi[8],nhi[6]
+	.rotr		a[3],n[3],t[2]
+
+{ .mmi;	ldf8		bi=[bptr],8		// (*bp++)
+	ldf8		alo[4]=[aptr],16	// ap[0]
+	$ADDP		r30=8,in1	};;
+{ .mmi;	ldf8		alo[3]=[r30],16		// ap[1]
+	ldf8		alo[2]=[aptr],16	// ap[2]
+	$ADDP		in4=0,in4	};;
+{ .mmi;	ldf8		alo[1]=[r30]		// ap[3]
+	ldf8		n0=[in4]		// n0
+	$ADDP		rptr=0,in0		}
+{ .mmi;	$ADDP		nptr=0,in3
+	mov		r31=16
+	zxt4		num=in5		};;
+{ .mmi;	ldf8		nlo[2]=[nptr],8		// np[0]
+	shladd		len=num,3,r0
+	shladd		r31=num,3,r31	};;
+{ .mmi;	ldf8		nlo[1]=[nptr],8		// np[1]
+	add		lc=-5,num
+	sub		r31=sp,r31	};;
+{ .mfb;	and		sp=-16,r31		// alloca
+	xmpy.hu		ahi[2]=alo[4],bi	// ap[0]*bp[0]
+	nop.b		0		}
+{ .mfb;	nop.m		0
+	xmpy.lu		alo[4]=alo[4],bi
+	brp.loop.imp	.L1st_ctop,.L1st_cend-16
+					};;
+{ .mfi;	nop.m		0
+	xma.hu		ahi[1]=alo[3],bi,ahi[2]	// ap[1]*bp[0]
+	add		tp_1=8,sp	}
+{ .mfi;	nop.m		0
+	xma.lu		alo[3]=alo[3],bi,ahi[2]
+	mov		pr.rot=0x20001f<<16
+			// ------^----- (p40) at first (p23)
+			// ----------^^ p[16:20]=1
+					};;
+{ .mfi;	nop.m		0
+	xmpy.lu		m0=alo[4],n0		// (ap[0]*bp[0])*n0
+	mov		ar.lc=lc	}
+{ .mfi;	nop.m		0
+	fcvt.fxu.s1	nhi[1]=f0
+	mov		ar.ec=8		};;
+
+.align	32
+.L1st_ctop:
+.pred.rel	"mutex",p40,p42
+{ .mfi;	(p16)	ldf8		alo[0]=[aptr],8		    // *(aptr++)
+	(p18)	xma.hu		ahi[0]=alo[2],bi,ahi[1]
+	(p40)	add		n[2]=n[2],a[2]		}   // (p23)					}
+{ .mfi;	(p18)	ldf8		nlo[0]=[nptr],8		    // *(nptr++)(p16)
+	(p18)	xma.lu		alo[2]=alo[2],bi,ahi[1]
+	(p42)	add		n[2]=n[2],a[2],1	};; // (p23)
+{ .mfi;	(p21)	getf.sig	a[0]=alo[5]
+	(p20)	xma.hu		nhi[0]=nlo[2],m0,nhi[1]
+	(p42)	cmp.leu		p41,p39=n[2],a[2]   	}   // (p23)
+{ .mfi;	(p23)	st8		[tp_1]=n[2],8
+	(p20)	xma.lu		nlo[2]=nlo[2],m0,nhi[1]
+	(p40)	cmp.ltu		p41,p39=n[2],a[2]	}   // (p23)
+{ .mmb;	(p21)	getf.sig	n[0]=nlo[3]
+	(p16)	nop.m		0
+	br.ctop.sptk	.L1st_ctop			};;
+.L1st_cend:
+
+{ .mmi;	getf.sig	a[0]=ahi[6]		// (p24)
+	getf.sig	n[0]=nhi[4]
+	add		num=-1,num	};;	// num--
+{ .mmi;	.pred.rel	"mutex",p40,p42
+(p40)	add		n[0]=n[0],a[0]
+(p42)	add		n[0]=n[0],a[0],1
+	sub		aptr=aptr,len	};;	// rewind
+{ .mmi;	.pred.rel	"mutex",p40,p42
+(p40)	cmp.ltu		p41,p39=n[0],a[0]
+(p42)	cmp.leu		p41,p39=n[0],a[0]
+	sub		nptr=nptr,len	};;
+{ .mmi;	.pred.rel	"mutex",p39,p41
+(p39)	add		topbit=r0,r0
+(p41)	add		topbit=r0,r0,1
+	nop.i		0		}	
+{ .mmi;	st8		[tp_1]=n[0]
+	add		tptr=16,sp
+	add		tp_1=8,sp	};;
+
+.Louter:
+{ .mmi;	ldf8		bi=[bptr],8		// (*bp++)
+	ldf8		ahi[3]=[tptr]		// tp[0]
+	add		r30=8,aptr	};;
+{ .mmi;	ldf8		alo[4]=[aptr],16	// ap[0]
+	ldf8		alo[3]=[r30],16		// ap[1]
+	add		r31=8,nptr	};;
+{ .mfb;	ldf8		alo[2]=[aptr],16	// ap[2]
+	xma.hu		ahi[2]=alo[4],bi,ahi[3]	// ap[0]*bp[i]+tp[0]
+	brp.loop.imp	.Linner_ctop,.Linner_cend-16
+					}
+{ .mfb;	ldf8		alo[1]=[r30]		// ap[3]
+	xma.lu		alo[4]=alo[4],bi,ahi[3]
+	clrrrb.pr			};;
+{ .mfi;	ldf8		nlo[2]=[nptr],16	// np[0]
+	xma.hu		ahi[1]=alo[3],bi,ahi[2]	// ap[1]*bp[i]
+	nop.i		0		}
+{ .mfi;	ldf8		nlo[1]=[r31]		// np[1]
+	xma.lu		alo[3]=alo[3],bi,ahi[2]
+	mov		pr.rot=0x20101f<<16
+			// ------^----- (p40) at first (p23)
+			// --------^--- (p30) at first (p22)
+			// ----------^^ p[16:20]=1
+					};;
+{ .mfi;	st8		[tptr]=r0		// tp[0] is already accounted
+	xmpy.lu		m0=alo[4],n0		// (ap[0]*bp[i]+tp[0])*n0
+	mov		ar.lc=lc	}
+{ .mfi;
+	fcvt.fxu.s1	nhi[1]=f0
+	mov		ar.ec=8		};;
+
+// This loop spins in 4*(n+7) ticks on Itanium 2 and should spin in
+// 7*(n+7) ticks on Itanium (the one codenamed Merced). Factor of 7
+// in latter case accounts for two-tick pipeline stall, which means
+// that its performance would be ~20% lower than optimal one. No
+// attempt was made to address this, because original Itanium is
+// hardly represented out in the wild...
+.align	32
+.Linner_ctop:
+.pred.rel	"mutex",p40,p42
+.pred.rel	"mutex",p30,p32
+{ .mfi;	(p16)	ldf8		alo[0]=[aptr],8		    // *(aptr++)
+	(p18)	xma.hu		ahi[0]=alo[2],bi,ahi[1]
+	(p40)	add		n[2]=n[2],a[2]		}   // (p23)
+{ .mfi;	(p16)	nop.m		0
+	(p18)	xma.lu		alo[2]=alo[2],bi,ahi[1]
+	(p42)	add		n[2]=n[2],a[2],1	};; // (p23)
+{ .mfi;	(p21)	getf.sig	a[0]=alo[5]
+	(p16)	nop.f		0
+	(p40)	cmp.ltu		p41,p39=n[2],a[2]	}   // (p23)
+{ .mfi;	(p21)	ld8		t[0]=[tptr],8
+	(p16)	nop.f		0
+	(p42)	cmp.leu		p41,p39=n[2],a[2]	};; // (p23)
+{ .mfi;	(p18)	ldf8		nlo[0]=[nptr],8		    // *(nptr++)
+	(p20)	xma.hu		nhi[0]=nlo[2],m0,nhi[1]
+	(p30)	add		a[1]=a[1],t[1]		}   // (p22)
+{ .mfi;	(p16)	nop.m		0
+	(p20)	xma.lu		nlo[2]=nlo[2],m0,nhi[1]
+	(p32)	add		a[1]=a[1],t[1],1	};; // (p22)
+{ .mmi;	(p21)	getf.sig	n[0]=nlo[3]
+	(p16)	nop.m		0
+	(p30)	cmp.ltu		p31,p29=a[1],t[1]	}   // (p22)
+{ .mmb;	(p23)	st8		[tp_1]=n[2],8
+	(p32)	cmp.leu		p31,p29=a[1],t[1]	    // (p22)
+	br.ctop.sptk	.Linner_ctop			};;
+.Linner_cend:
+
+{ .mmi;	getf.sig	a[0]=ahi[6]		// (p24)
+	getf.sig	n[0]=nhi[4]
+	nop.i		0		};;
+
+{ .mmi;	.pred.rel	"mutex",p31,p33
+(p31)	add		a[0]=a[0],topbit
+(p33)	add		a[0]=a[0],topbit,1
+	mov		topbit=r0	};;
+{ .mfi; .pred.rel	"mutex",p31,p33
+(p31)	cmp.ltu		p32,p30=a[0],topbit
+(p33)	cmp.leu		p32,p30=a[0],topbit
+					}
+{ .mfi;	.pred.rel	"mutex",p40,p42
+(p40)	add		n[0]=n[0],a[0]
+(p42)	add		n[0]=n[0],a[0],1
+					};;
+{ .mmi;	.pred.rel	"mutex",p44,p46
+(p40)	cmp.ltu		p41,p39=n[0],a[0]
+(p42)	cmp.leu		p41,p39=n[0],a[0]
+(p32)	add		topbit=r0,r0,1	}
+
+{ .mmi;	st8		[tp_1]=n[0],8
+	cmp4.ne		p6,p0=1,num
+	sub		aptr=aptr,len	};;	// rewind
+{ .mmi;	sub		nptr=nptr,len
+(p41)	add		topbit=r0,r0,1
+	add		tptr=16,sp	}
+{ .mmb;	add		tp_1=8,sp
+	add		num=-1,num		// num--
+(p6)	br.cond.sptk.many	.Louter	};;
+
+{ .mbb;	add		lc=4,lc
+	brp.loop.imp	.Lsub_ctop,.Lsub_cend-16
+	clrrrb.pr			};;
+{ .mii;	nop.m		0
+	mov		pr.rot=0x10001<<16
+			// ------^---- (p33) at first (p17)
+	mov		ar.lc=lc	}
+{ .mii;	nop.m		0
+	mov		ar.ec=3
+	nop.i		0		};;
+
+.Lsub_ctop:
+.pred.rel	"mutex",p33,p35
+{ .mfi;	(p16)	ld8		t[0]=[tptr],8		    // t=*(tp++)
+	(p16)	nop.f		0
+	(p33)	sub		n[1]=t[1],n[1]		}   // (p17)
+{ .mfi;	(p16)	ld8		n[0]=[nptr],8		    // n=*(np++)
+	(p16)	nop.f		0
+	(p35)	sub		n[1]=t[1],n[1],1	};; // (p17)
+{ .mib;	(p18)	st8		[rptr]=n[2],8		    // *(rp++)=r
+	(p33)	cmp.gtu		p34,p32=n[1],t[1]	    // (p17)
+	(p18)	nop.b		0			}
+{ .mib;	(p18)	nop.m		0
+	(p35)	cmp.geu		p34,p32=n[1],t[1]	    // (p17)
+	br.ctop.sptk	.Lsub_ctop			};;
+.Lsub_cend:
+
+{ .mmb;	.pred.rel	"mutex",p34,p36
+(p34)	sub	topbit=topbit,r0	// (p19)
+(p36)	sub	topbit=topbit,r0,1
+	brp.loop.imp	.Lcopy_ctop,.Lcopy_cend-16
+					}
+{ .mmb;	sub	rptr=rptr,len		// rewind
+	sub	tptr=tptr,len
+	clrrrb.pr			};;
+{ .mmi;	and	aptr=tptr,topbit
+	andcm	bptr=rptr,topbit
+	mov	pr.rot=1<<16		};;
+{ .mii;	or	nptr=aptr,bptr
+	mov	ar.lc=lc
+	mov	ar.ec=3			};;
+
+.Lcopy_ctop:
+{ .mmb;	(p16)	ld8	n[0]=[nptr],8
+	(p18)	st8	[tptr]=r0,8
+	(p16)	nop.b	0		}
+{ .mmb;	(p16)	nop.m	0
+	(p18)	st8	[rptr]=n[2],8
+	br.ctop.sptk	.Lcopy_ctop	};;
+.Lcopy_cend:
+
+{ .mmi;	mov		ret0=1			// signal "handled"
+	rum		1<<5			// clear um.mfh
+	mov		ar.lc=prevlc	}
+{ .mib;	.restore	sp
+	mov		sp=prevsp
+	mov		pr=prevpr,0x1ffff
+	br.ret.sptk.many	b0	};;
+.endp	bn_mul_mont_general#
+
+a1=r16;  a2=r17;  a3=r18;  a4=r19;  a5=r20;  a6=r21;  a7=r22;  a8=r23;
+n1=r24;  n2=r25;  n3=r26;  n4=r27;  n5=r28;  n6=r29;  n7=r30;  n8=r31;
+t0=r15;
+
+ai0=f8;  ai1=f9;  ai2=f10; ai3=f11; ai4=f12; ai5=f13; ai6=f14; ai7=f15;
+ni0=f16; ni1=f17; ni2=f18; ni3=f19; ni4=f20; ni5=f21; ni6=f22; ni7=f23;
+
+.align	64
+.skip	48		// aligns loop body
+.local	bn_mul_mont_8#
+.proc	bn_mul_mont_8#
+bn_mul_mont_8:
+	.prologue
+{ .mmi;	.save		ar.pfs,prevfs
+	alloc		prevfs=ar.pfs,6,2,0,8
+	.vframe		prevsp
+	mov		prevsp=sp
+	.save		ar.lc,prevlc
+	mov		prevlc=ar.lc	}
+{ .mmi;	add		r17=-6*16,sp
+	add		sp=-7*16,sp
+	.save		pr,prevpr
+	mov		prevpr=pr	};;
+
+{ .mmi;	.save.gf	0,0x10
+	stf.spill	[sp]=f16,-16
+	.save.gf	0,0x20
+	stf.spill	[r17]=f17,32
+	add		r16=-5*16,prevsp};;
+{ .mmi;	.save.gf	0,0x40
+	stf.spill	[r16]=f18,32
+	.save.gf	0,0x80
+	stf.spill	[r17]=f19,32
+	$ADDP		aptr=0,in1	};;
+{ .mmi;	.save.gf	0,0x100
+	stf.spill	[r16]=f20,32
+	.save.gf	0,0x200
+	stf.spill	[r17]=f21,32
+	$ADDP		r29=8,in1	};;
+{ .mmi;	.save.gf	0,0x400
+	stf.spill	[r16]=f22
+	.save.gf	0,0x800
+	stf.spill	[r17]=f23
+	$ADDP		rptr=0,in0	};;
+
+	.body
+	.rotf		bj[8],mj[2],tf[2],alo[10],ahi[10],nlo[10],nhi[10]
+	.rotr		t[8]
+
+// load input vectors padding them to 8 elements
+{ .mmi;	ldf8		ai0=[aptr],16		// ap[0]
+	ldf8		ai1=[r29],16		// ap[1]
+	$ADDP		bptr=0,in2	}
+{ .mmi;	$ADDP		r30=8,in2
+	$ADDP		nptr=0,in3
+	$ADDP		r31=8,in3	};;
+{ .mmi;	ldf8		bj[7]=[bptr],16		// bp[0]
+	ldf8		bj[6]=[r30],16		// bp[1]
+	cmp4.le		p4,p5=3,in5	}
+{ .mmi;	ldf8		ni0=[nptr],16		// np[0]
+	ldf8		ni1=[r31],16		// np[1]
+	cmp4.le		p6,p7=4,in5	};;
+
+{ .mfi;	(p4)ldf8	ai2=[aptr],16		// ap[2]
+	(p5)fcvt.fxu	ai2=f0
+	cmp4.le		p8,p9=5,in5	}
+{ .mfi;	(p6)ldf8	ai3=[r29],16		// ap[3]
+	(p7)fcvt.fxu	ai3=f0
+	cmp4.le		p10,p11=6,in5	}
+{ .mfi;	(p4)ldf8	bj[5]=[bptr],16		// bp[2]
+	(p5)fcvt.fxu	bj[5]=f0
+	cmp4.le		p12,p13=7,in5	}
+{ .mfi;	(p6)ldf8	bj[4]=[r30],16		// bp[3]
+	(p7)fcvt.fxu	bj[4]=f0
+	cmp4.le		p14,p15=8,in5	}
+{ .mfi;	(p4)ldf8	ni2=[nptr],16		// np[2]
+	(p5)fcvt.fxu	ni2=f0
+	addp4		r28=-1,in5	}
+{ .mfi;	(p6)ldf8	ni3=[r31],16		// np[3]
+	(p7)fcvt.fxu	ni3=f0
+	$ADDP		in4=0,in4	};;
+
+{ .mfi;	ldf8		n0=[in4]
+	fcvt.fxu	tf[1]=f0
+	nop.i		0		}
+
+{ .mfi;	(p8)ldf8	ai4=[aptr],16		// ap[4]
+	(p9)fcvt.fxu	ai4=f0
+	mov		t[0]=r0		}
+{ .mfi;	(p10)ldf8	ai5=[r29],16		// ap[5]
+	(p11)fcvt.fxu	ai5=f0
+	mov		t[1]=r0		}
+{ .mfi;	(p8)ldf8	bj[3]=[bptr],16		// bp[4]
+	(p9)fcvt.fxu	bj[3]=f0
+	mov		t[2]=r0		}
+{ .mfi;	(p10)ldf8	bj[2]=[r30],16		// bp[5]
+	(p11)fcvt.fxu	bj[2]=f0
+	mov		t[3]=r0		}
+{ .mfi;	(p8)ldf8	ni4=[nptr],16		// np[4]
+	(p9)fcvt.fxu	ni4=f0
+	mov		t[4]=r0		}
+{ .mfi;	(p10)ldf8	ni5=[r31],16		// np[5]
+	(p11)fcvt.fxu	ni5=f0
+	mov		t[5]=r0		};;
+
+{ .mfi;	(p12)ldf8	ai6=[aptr],16		// ap[6]
+	(p13)fcvt.fxu	ai6=f0
+	mov		t[6]=r0		}
+{ .mfi;	(p14)ldf8	ai7=[r29],16		// ap[7]
+	(p15)fcvt.fxu	ai7=f0
+	mov		t[7]=r0		}
+{ .mfi;	(p12)ldf8	bj[1]=[bptr],16		// bp[6]
+	(p13)fcvt.fxu	bj[1]=f0
+	mov		ar.lc=r28	}
+{ .mfi;	(p14)ldf8	bj[0]=[r30],16		// bp[7]
+	(p15)fcvt.fxu	bj[0]=f0
+	mov		ar.ec=1		}
+{ .mfi;	(p12)ldf8	ni6=[nptr],16		// np[6]
+	(p13)fcvt.fxu	ni6=f0
+	mov		pr.rot=1<<16	}
+{ .mfb;	(p14)ldf8	ni7=[r31],16		// np[7]
+	(p15)fcvt.fxu	ni7=f0
+	brp.loop.imp	.Louter_8_ctop,.Louter_8_cend-16
+					};;
+
+// The loop is scheduled for 32*n ticks on Itanium 2. Actual attempt
+// to measure with help of Interval Time Counter indicated that the
+// factor is a tad higher: 33 or 34, if not 35. Exact measurement and
+// addressing the issue is problematic, because I don't have access
+// to platform-specific instruction-level profiler. On Itanium it
+// should run in 56*n ticks, because of higher xma latency...
+.Louter_8_ctop:
+	.pred.rel		"mutex",p40,p42
+	.pred.rel		"mutex",p48,p50
+{ .mfi;	(p16)	nop.m		0			// 0:
+	(p16)	xma.hu		ahi[0]=ai0,bj[7],tf[1]	//	ap[0]*b[i]+t[0]
+	(p40)	add		a3=a3,n3	}	//	(p17) a3+=n3
+{ .mfi;	(p42)	add		a3=a3,n3,1
+	(p16)	xma.lu		alo[0]=ai0,bj[7],tf[1]
+	(p16)	nop.i		0		};;
+{ .mii;	(p17)	getf.sig	a7=alo[8]		// 1:
+	(p48)	add		t[6]=t[6],a3		//	(p17) t[6]+=a3
+	(p50)	add		t[6]=t[6],a3,1	};;
+{ .mfi;	(p17)	getf.sig	a8=ahi[8]		// 2:
+	(p17)	xma.hu		nhi[7]=ni6,mj[1],nhi[6]	//	np[6]*m0
+	(p40)	cmp.ltu		p43,p41=a3,n3	}
+{ .mfi;	(p42)	cmp.leu		p43,p41=a3,n3
+	(p17)	xma.lu		nlo[7]=ni6,mj[1],nhi[6]
+	(p16)	nop.i		0		};;
+{ .mii;	(p17)	getf.sig	n5=nlo[6]		// 3:
+	(p48)	cmp.ltu		p51,p49=t[6],a3
+	(p50)	cmp.leu		p51,p49=t[6],a3	};;
+	.pred.rel		"mutex",p41,p43
+	.pred.rel		"mutex",p49,p51
+{ .mfi;	(p16)	nop.m		0			// 4:
+	(p16)	xma.hu		ahi[1]=ai1,bj[7],ahi[0]	//	ap[1]*b[i]
+	(p41)	add		a4=a4,n4	}	//	(p17) a4+=n4
+{ .mfi;	(p43)	add		a4=a4,n4,1
+	(p16)	xma.lu		alo[1]=ai1,bj[7],ahi[0]
+	(p16)	nop.i		0		};;
+{ .mfi;	(p49)	add		t[5]=t[5],a4		// 5:	(p17) t[5]+=a4
+	(p16)	xmpy.lu		mj[0]=alo[0],n0		//	(ap[0]*b[i]+t[0])*n0
+	(p51)	add		t[5]=t[5],a4,1	};;
+{ .mfi;	(p16)	nop.m		0			// 6:
+	(p17)	xma.hu		nhi[8]=ni7,mj[1],nhi[7]	//	np[7]*m0
+	(p41)	cmp.ltu		p42,p40=a4,n4	}
+{ .mfi;	(p43)	cmp.leu		p42,p40=a4,n4
+	(p17)	xma.lu		nlo[8]=ni7,mj[1],nhi[7]
+	(p16)	nop.i		0		};;
+{ .mii;	(p17)	getf.sig	n6=nlo[7]		// 7:
+	(p49)	cmp.ltu		p50,p48=t[5],a4
+	(p51)	cmp.leu		p50,p48=t[5],a4	};;
+	.pred.rel		"mutex",p40,p42
+	.pred.rel		"mutex",p48,p50
+{ .mfi;	(p16)	nop.m		0			// 8:
+	(p16)	xma.hu		ahi[2]=ai2,bj[7],ahi[1]	//	ap[2]*b[i]
+	(p40)	add		a5=a5,n5	}	//	(p17) a5+=n5
+{ .mfi;	(p42)	add		a5=a5,n5,1
+	(p16)	xma.lu		alo[2]=ai2,bj[7],ahi[1]
+	(p16)	nop.i		0		};;
+{ .mii;	(p16)	getf.sig	a1=alo[1]		// 9:
+	(p48)	add		t[4]=t[4],a5		//	p(17) t[4]+=a5
+	(p50)	add		t[4]=t[4],a5,1	};;
+{ .mfi;	(p16)	nop.m		0			// 10:
+	(p16)	xma.hu		nhi[0]=ni0,mj[0],alo[0]	//	np[0]*m0
+	(p40)	cmp.ltu		p43,p41=a5,n5	}
+{ .mfi;	(p42)	cmp.leu		p43,p41=a5,n5
+	(p16)	xma.lu		nlo[0]=ni0,mj[0],alo[0]
+	(p16)	nop.i		0		};;
+{ .mii;	(p17)	getf.sig	n7=nlo[8]		// 11:
+	(p48)	cmp.ltu		p51,p49=t[4],a5
+	(p50)	cmp.leu		p51,p49=t[4],a5	};;
+	.pred.rel		"mutex",p41,p43
+	.pred.rel		"mutex",p49,p51
+{ .mfi;	(p17)	getf.sig	n8=nhi[8]		// 12:
+	(p16)	xma.hu		ahi[3]=ai3,bj[7],ahi[2]	//	ap[3]*b[i]
+	(p41)	add		a6=a6,n6	}	//	(p17) a6+=n6
+{ .mfi;	(p43)	add		a6=a6,n6,1
+	(p16)	xma.lu		alo[3]=ai3,bj[7],ahi[2]
+	(p16)	nop.i		0		};;
+{ .mii;	(p16)	getf.sig	a2=alo[2]		// 13:
+	(p49)	add		t[3]=t[3],a6		//	(p17) t[3]+=a6
+	(p51)	add		t[3]=t[3],a6,1	};;
+{ .mfi;	(p16)	nop.m		0			// 14:
+	(p16)	xma.hu		nhi[1]=ni1,mj[0],nhi[0]	//	np[1]*m0
+	(p41)	cmp.ltu		p42,p40=a6,n6	}
+{ .mfi;	(p43)	cmp.leu		p42,p40=a6,n6
+	(p16)	xma.lu		nlo[1]=ni1,mj[0],nhi[0]
+	(p16)	nop.i		0		};;
+{ .mii;	(p16)	nop.m		0			// 15:
+	(p49)	cmp.ltu		p50,p48=t[3],a6
+	(p51)	cmp.leu		p50,p48=t[3],a6	};;
+	.pred.rel		"mutex",p40,p42
+	.pred.rel		"mutex",p48,p50
+{ .mfi;	(p16)	nop.m		0			// 16:
+	(p16)	xma.hu		ahi[4]=ai4,bj[7],ahi[3]	//	ap[4]*b[i]
+	(p40)	add		a7=a7,n7	}	//	(p17) a7+=n7
+{ .mfi;	(p42)	add		a7=a7,n7,1
+	(p16)	xma.lu		alo[4]=ai4,bj[7],ahi[3]
+	(p16)	nop.i		0		};;
+{ .mii;	(p16)	getf.sig	a3=alo[3]		// 17:
+	(p48)	add		t[2]=t[2],a7		//	(p17) t[2]+=a7
+	(p50)	add		t[2]=t[2],a7,1	};;
+{ .mfi;	(p16)	nop.m		0			// 18:
+	(p16)	xma.hu		nhi[2]=ni2,mj[0],nhi[1]	//	np[2]*m0
+	(p40)	cmp.ltu		p43,p41=a7,n7	}
+{ .mfi;	(p42)	cmp.leu		p43,p41=a7,n7
+	(p16)	xma.lu		nlo[2]=ni2,mj[0],nhi[1]
+	(p16)	nop.i		0		};;
+{ .mii;	(p16)	getf.sig	n1=nlo[1]		// 19:
+	(p48)	cmp.ltu		p51,p49=t[2],a7
+	(p50)	cmp.leu		p51,p49=t[2],a7	};;
+	.pred.rel		"mutex",p41,p43
+	.pred.rel		"mutex",p49,p51
+{ .mfi;	(p16)	nop.m		0			// 20:
+	(p16)	xma.hu		ahi[5]=ai5,bj[7],ahi[4]	//	ap[5]*b[i]
+	(p41)	add		a8=a8,n8	}	//	(p17) a8+=n8
+{ .mfi;	(p43)	add		a8=a8,n8,1
+	(p16)	xma.lu		alo[5]=ai5,bj[7],ahi[4]
+	(p16)	nop.i		0		};;
+{ .mii;	(p16)	getf.sig	a4=alo[4]		// 21:
+	(p49)	add		t[1]=t[1],a8		//	(p17) t[1]+=a8
+	(p51)	add		t[1]=t[1],a8,1	};;
+{ .mfi;	(p16)	nop.m		0			// 22:
+	(p16)	xma.hu		nhi[3]=ni3,mj[0],nhi[2]	//	np[3]*m0
+	(p41)	cmp.ltu		p42,p40=a8,n8	}
+{ .mfi;	(p43)	cmp.leu		p42,p40=a8,n8
+	(p16)	xma.lu		nlo[3]=ni3,mj[0],nhi[2]
+	(p16)	nop.i		0		};;
+{ .mii;	(p16)	getf.sig	n2=nlo[2]		// 23:
+	(p49)	cmp.ltu		p50,p48=t[1],a8
+	(p51)	cmp.leu		p50,p48=t[1],a8	};;
+{ .mfi;	(p16)	nop.m		0			// 24:
+	(p16)	xma.hu		ahi[6]=ai6,bj[7],ahi[5]	//	ap[6]*b[i]
+	(p16)	add		a1=a1,n1	}	//	(p16) a1+=n1
+{ .mfi;	(p16)	nop.m		0
+	(p16)	xma.lu		alo[6]=ai6,bj[7],ahi[5]
+	(p17)	mov		t[0]=r0		};;
+{ .mii;	(p16)	getf.sig	a5=alo[5]		// 25:
+	(p16)	add		t0=t[7],a1		//	(p16) t[7]+=a1
+	(p42)	add		t[0]=t[0],r0,1	};;
+{ .mfi;	(p16)	setf.sig	tf[0]=t0		// 26:
+	(p16)	xma.hu		nhi[4]=ni4,mj[0],nhi[3]	//	np[4]*m0
+	(p50)	add		t[0]=t[0],r0,1	}
+{ .mfi;	(p16)	cmp.ltu.unc	p42,p40=a1,n1
+	(p16)	xma.lu		nlo[4]=ni4,mj[0],nhi[3]
+	(p16)	nop.i		0		};;
+{ .mii;	(p16)	getf.sig	n3=nlo[3]		// 27:
+	(p16)	cmp.ltu.unc	p50,p48=t0,a1
+	(p16)	nop.i		0		};;
+	.pred.rel		"mutex",p40,p42
+	.pred.rel		"mutex",p48,p50
+{ .mfi;	(p16)	nop.m		0			// 28:
+	(p16)	xma.hu		ahi[7]=ai7,bj[7],ahi[6]	//	ap[7]*b[i]
+	(p40)	add		a2=a2,n2	}	//	(p16) a2+=n2
+{ .mfi;	(p42)	add		a2=a2,n2,1
+	(p16)	xma.lu		alo[7]=ai7,bj[7],ahi[6]
+	(p16)	nop.i		0		};;
+{ .mii;	(p16)	getf.sig	a6=alo[6]		// 29:
+	(p48)	add		t[6]=t[6],a2		//	(p16) t[6]+=a2
+	(p50)	add		t[6]=t[6],a2,1	};;
+{ .mfi;	(p16)	nop.m		0			// 30:
+	(p16)	xma.hu		nhi[5]=ni5,mj[0],nhi[4]	//	np[5]*m0
+	(p40)	cmp.ltu		p41,p39=a2,n2	}
+{ .mfi;	(p42)	cmp.leu		p41,p39=a2,n2
+	(p16)	xma.lu		nlo[5]=ni5,mj[0],nhi[4]
+	(p16)	nop.i		0		};;
+{ .mfi;	(p16)	getf.sig	n4=nlo[4]		// 31:
+	(p16)	nop.f		0
+	(p48)	cmp.ltu		p49,p47=t[6],a2	}
+{ .mfb;	(p50)	cmp.leu		p49,p47=t[6],a2
+	(p16)	nop.f		0
+	br.ctop.sptk.many	.Louter_8_ctop	};;
+.Louter_8_cend:
+
+// above loop has to execute one more time, without (p16), which is
+// replaced with merged move of np[8] to GPR bank
+	.pred.rel		"mutex",p40,p42
+	.pred.rel		"mutex",p48,p50
+{ .mmi;	(p0)	getf.sig	n1=ni0			// 0:
+	(p40)	add		a3=a3,n3		//	(p17) a3+=n3
+	(p42)	add		a3=a3,n3,1	};;
+{ .mii;	(p17)	getf.sig	a7=alo[8]		// 1:
+	(p48)	add		t[6]=t[6],a3		//	(p17) t[6]+=a3
+	(p50)	add		t[6]=t[6],a3,1	};;
+{ .mfi;	(p17)	getf.sig	a8=ahi[8]		// 2:
+	(p17)	xma.hu		nhi[7]=ni6,mj[1],nhi[6]	//	np[6]*m0
+	(p40)	cmp.ltu		p43,p41=a3,n3	}
+{ .mfi;	(p42)	cmp.leu		p43,p41=a3,n3
+	(p17)	xma.lu		nlo[7]=ni6,mj[1],nhi[6]
+	(p0)	nop.i		0		};;
+{ .mii;	(p17)	getf.sig	n5=nlo[6]		// 3:
+	(p48)	cmp.ltu		p51,p49=t[6],a3
+	(p50)	cmp.leu		p51,p49=t[6],a3	};;
+	.pred.rel		"mutex",p41,p43
+	.pred.rel		"mutex",p49,p51
+{ .mmi;	(p0)	getf.sig	n2=ni1			// 4:
+	(p41)	add		a4=a4,n4		//	(p17) a4+=n4
+	(p43)	add		a4=a4,n4,1	};;
+{ .mfi;	(p49)	add		t[5]=t[5],a4		// 5:	(p17) t[5]+=a4
+	(p0)	nop.f		0
+	(p51)	add		t[5]=t[5],a4,1	};;
+{ .mfi;	(p0)	getf.sig	n3=ni2			// 6:
+	(p17)	xma.hu		nhi[8]=ni7,mj[1],nhi[7]	//	np[7]*m0
+	(p41)	cmp.ltu		p42,p40=a4,n4	}
+{ .mfi;	(p43)	cmp.leu		p42,p40=a4,n4
+	(p17)	xma.lu		nlo[8]=ni7,mj[1],nhi[7]
+	(p0)	nop.i		0		};;
+{ .mii;	(p17)	getf.sig	n6=nlo[7]		// 7:
+	(p49)	cmp.ltu		p50,p48=t[5],a4
+	(p51)	cmp.leu		p50,p48=t[5],a4	};;
+	.pred.rel		"mutex",p40,p42
+	.pred.rel		"mutex",p48,p50
+{ .mii;	(p0)	getf.sig	n4=ni3			// 8:
+	(p40)	add		a5=a5,n5		//	(p17) a5+=n5
+	(p42)	add		a5=a5,n5,1	};;
+{ .mii;	(p0)	nop.m		0			// 9:
+	(p48)	add		t[4]=t[4],a5		//	p(17) t[4]+=a5
+	(p50)	add		t[4]=t[4],a5,1	};;
+{ .mii;	(p0)	nop.m		0			// 10:
+	(p40)	cmp.ltu		p43,p41=a5,n5
+	(p42)	cmp.leu		p43,p41=a5,n5	};;
+{ .mii;	(p17)	getf.sig	n7=nlo[8]		// 11:
+	(p48)	cmp.ltu		p51,p49=t[4],a5
+	(p50)	cmp.leu		p51,p49=t[4],a5	};;
+	.pred.rel		"mutex",p41,p43
+	.pred.rel		"mutex",p49,p51
+{ .mii;	(p17)	getf.sig	n8=nhi[8]		// 12:
+	(p41)	add		a6=a6,n6		//	(p17) a6+=n6
+	(p43)	add		a6=a6,n6,1	};;
+{ .mii;	(p0)	getf.sig	n5=ni4			// 13:
+	(p49)	add		t[3]=t[3],a6		//	(p17) t[3]+=a6
+	(p51)	add		t[3]=t[3],a6,1	};;
+{ .mii;	(p0)	nop.m		0			// 14:
+	(p41)	cmp.ltu		p42,p40=a6,n6
+	(p43)	cmp.leu		p42,p40=a6,n6	};;
+{ .mii;	(p0)	getf.sig	n6=ni5			// 15:
+	(p49)	cmp.ltu		p50,p48=t[3],a6
+	(p51)	cmp.leu		p50,p48=t[3],a6	};;
+	.pred.rel		"mutex",p40,p42
+	.pred.rel		"mutex",p48,p50
+{ .mii;	(p0)	nop.m		0			// 16:
+	(p40)	add		a7=a7,n7		//	(p17) a7+=n7
+	(p42)	add		a7=a7,n7,1	};;
+{ .mii;	(p0)	nop.m		0			// 17:
+	(p48)	add		t[2]=t[2],a7		//	(p17) t[2]+=a7
+	(p50)	add		t[2]=t[2],a7,1	};;
+{ .mii;	(p0)	nop.m		0			// 18:
+	(p40)	cmp.ltu		p43,p41=a7,n7
+	(p42)	cmp.leu		p43,p41=a7,n7	};;
+{ .mii;	(p0)	getf.sig	n7=ni6			// 19:
+	(p48)	cmp.ltu		p51,p49=t[2],a7
+	(p50)	cmp.leu		p51,p49=t[2],a7	};;
+	.pred.rel		"mutex",p41,p43
+	.pred.rel		"mutex",p49,p51
+{ .mii;	(p0)	nop.m		0			// 20:
+	(p41)	add		a8=a8,n8		//	(p17) a8+=n8
+	(p43)	add		a8=a8,n8,1	};;
+{ .mmi;	(p0)	nop.m		0			// 21:
+	(p49)	add		t[1]=t[1],a8		//	(p17) t[1]+=a8
+	(p51)	add		t[1]=t[1],a8,1	}
+{ .mmi;	(p17)	mov		t[0]=r0
+	(p41)	cmp.ltu		p42,p40=a8,n8
+	(p43)	cmp.leu		p42,p40=a8,n8	};;
+{ .mmi;	(p0)	getf.sig	n8=ni7			// 22:
+	(p49)	cmp.ltu		p50,p48=t[1],a8
+	(p51)	cmp.leu		p50,p48=t[1],a8	}
+{ .mmi;	(p42)	add		t[0]=t[0],r0,1
+	(p0)	add		r16=-7*16,prevsp
+	(p0)	add		r17=-6*16,prevsp	};;
+
+// subtract np[8] from carrybit|tmp[8]
+// carrybit|tmp[8] layout upon exit from above loop is:
+//	t[0]|t[1]|t[2]|t[3]|t[4]|t[5]|t[6]|t[7]|t0 (least significant)
+{ .mmi;	(p50)add	t[0]=t[0],r0,1
+	add		r18=-5*16,prevsp
+	sub		n1=t0,n1	};;
+{ .mmi;	cmp.gtu		p34,p32=n1,t0;;
+	.pred.rel	"mutex",p32,p34
+	(p32)sub	n2=t[7],n2
+	(p34)sub	n2=t[7],n2,1	};;
+{ .mii;	(p32)cmp.gtu	p35,p33=n2,t[7]
+	(p34)cmp.geu	p35,p33=n2,t[7];;
+	.pred.rel	"mutex",p33,p35
+	(p33)sub	n3=t[6],n3	}
+{ .mmi;	(p35)sub	n3=t[6],n3,1;;
+	(p33)cmp.gtu	p34,p32=n3,t[6]
+	(p35)cmp.geu	p34,p32=n3,t[6]	};;
+	.pred.rel	"mutex",p32,p34
+{ .mii;	(p32)sub	n4=t[5],n4
+	(p34)sub	n4=t[5],n4,1;;
+	(p32)cmp.gtu	p35,p33=n4,t[5]	}
+{ .mmi;	(p34)cmp.geu	p35,p33=n4,t[5];;
+	.pred.rel	"mutex",p33,p35
+	(p33)sub	n5=t[4],n5
+	(p35)sub	n5=t[4],n5,1	};;
+{ .mii;	(p33)cmp.gtu	p34,p32=n5,t[4]
+	(p35)cmp.geu	p34,p32=n5,t[4];;
+	.pred.rel	"mutex",p32,p34
+	(p32)sub	n6=t[3],n6	}
+{ .mmi;	(p34)sub	n6=t[3],n6,1;;
+	(p32)cmp.gtu	p35,p33=n6,t[3]
+	(p34)cmp.geu	p35,p33=n6,t[3]	};;
+	.pred.rel	"mutex",p33,p35
+{ .mii;	(p33)sub	n7=t[2],n7
+	(p35)sub	n7=t[2],n7,1;;
+	(p33)cmp.gtu	p34,p32=n7,t[2]	}
+{ .mmi;	(p35)cmp.geu	p34,p32=n7,t[2];;
+	.pred.rel	"mutex",p32,p34
+	(p32)sub	n8=t[1],n8
+	(p34)sub	n8=t[1],n8,1	};;
+{ .mii;	(p32)cmp.gtu	p35,p33=n8,t[1]
+	(p34)cmp.geu	p35,p33=n8,t[1];;
+	.pred.rel	"mutex",p33,p35
+	(p33)sub	a8=t[0],r0	}
+{ .mmi;	(p35)sub	a8=t[0],r0,1;;
+	(p33)cmp.gtu	p34,p32=a8,t[0]
+	(p35)cmp.geu	p34,p32=a8,t[0]	};;
+
+// save the result, either tmp[num] or tmp[num]-np[num]
+	.pred.rel	"mutex",p32,p34
+{ .mmi;	(p32)st8	[rptr]=n1,8
+	(p34)st8	[rptr]=t0,8
+	add		r19=-4*16,prevsp};;
+{ .mmb;	(p32)st8	[rptr]=n2,8
+	(p34)st8	[rptr]=t[7],8
+	(p5)br.cond.dpnt.few	.Ldone	};;
+{ .mmb;	(p32)st8	[rptr]=n3,8
+	(p34)st8	[rptr]=t[6],8
+	(p7)br.cond.dpnt.few	.Ldone	};;
+{ .mmb;	(p32)st8	[rptr]=n4,8
+	(p34)st8	[rptr]=t[5],8
+	(p9)br.cond.dpnt.few	.Ldone	};;
+{ .mmb;	(p32)st8	[rptr]=n5,8
+	(p34)st8	[rptr]=t[4],8
+	(p11)br.cond.dpnt.few	.Ldone	};;
+{ .mmb;	(p32)st8	[rptr]=n6,8
+	(p34)st8	[rptr]=t[3],8
+	(p13)br.cond.dpnt.few	.Ldone	};;
+{ .mmb;	(p32)st8	[rptr]=n7,8
+	(p34)st8	[rptr]=t[2],8
+	(p15)br.cond.dpnt.few	.Ldone	};;
+{ .mmb;	(p32)st8	[rptr]=n8,8
+	(p34)st8	[rptr]=t[1],8
+	nop.b		0		};;
+.Ldone:						// epilogue
+{ .mmi;	ldf.fill	f16=[r16],64
+	ldf.fill	f17=[r17],64
+	nop.i		0		}
+{ .mmi;	ldf.fill	f18=[r18],64
+	ldf.fill	f19=[r19],64
+	mov		pr=prevpr,0x1ffff	};;
+{ .mmi;	ldf.fill	f20=[r16]
+	ldf.fill	f21=[r17]
+	mov		ar.lc=prevlc	}
+{ .mmi;	ldf.fill	f22=[r18]
+	ldf.fill	f23=[r19]
+	mov		ret0=1		}	// signal "handled"
+{ .mib;	rum		1<<5
+	.restore	sp
+	mov		sp=prevsp
+	br.ret.sptk.many	b0	};;
+.endp	bn_mul_mont_8#
+
+.type	copyright#,\@object
+copyright:
+stringz	"Montgomery multiplication for IA-64, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$output=shift and open STDOUT,">$output";
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/mips-mont.pl b/crypto/bn/asm/mips-mont.pl
new file mode 100755
index 000000000000..b944a12b8e29
--- /dev/null
+++ b/crypto/bn/asm/mips-mont.pl
@@ -0,0 +1,426 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# This module doesn't present direct interest for OpenSSL, because it
+# doesn't provide better performance for longer keys, at least not on
+# in-order-execution cores. While 512-bit RSA sign operations can be
+# 65% faster in 64-bit mode, 1024-bit ones are only 15% faster, and
+# 4096-bit ones are up to 15% slower. In 32-bit mode it varies from
+# 16% improvement for 512-bit RSA sign to -33% for 4096-bit RSA
+# verify:-( All comparisons are against bn_mul_mont-free assembler.
+# The module might be of interest to embedded system developers, as
+# the code is smaller than 1KB, yet offers >3x improvement on MIPS64
+# and 75-30% [less for longer keys] on MIPS32 over compiler-generated
+# code.
+
+######################################################################
+# There is a number of MIPS ABI in use, O32 and N32/64 are most
+# widely used. Then there is a new contender: NUBI. It appears that if
+# one picks the latter, it's possible to arrange code in ABI neutral
+# manner. Therefore let's stick to NUBI register layout:
+#
+($zero,$at,$t0,$t1,$t2)=map("\$$_",(0..2,24,25));
+($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7,$s8,$s9,$s10,$s11)=map("\$$_",(12..23));
+($gp,$tp,$sp,$fp,$ra)=map("\$$_",(3,28..31));
+#
+# The return value is placed in $a0. Following coding rules facilitate
+# interoperability:
+#
+# - never ever touch $tp, "thread pointer", former $gp;
+# - copy return value to $t0, former $v0 [or to $a0 if you're adapting
+#   old code];
+# - on O32 populate $a4-$a7 with 'lw $aN,4*N($sp)' if necessary;
+#
+# For reference here is register layout for N32/64 MIPS ABIs:
+#
+# ($zero,$at,$v0,$v1)=map("\$$_",(0..3));
+# ($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+# ($t0,$t1,$t2,$t3,$t8,$t9)=map("\$$_",(12..15,24,25));
+# ($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7)=map("\$$_",(16..23));
+# ($gp,$sp,$fp,$ra)=map("\$$_",(28..31));
+#
+$flavour = shift; # supported flavours are o32,n32,64,nubi32,nubi64
+
+if ($flavour =~ /64|n32/i) {
+	$PTR_ADD="dadd";	# incidentally works even on n32
+	$PTR_SUB="dsub";	# incidentally works even on n32
+	$REG_S="sd";
+	$REG_L="ld";
+	$SZREG=8;
+} else {
+	$PTR_ADD="add";
+	$PTR_SUB="sub";
+	$REG_S="sw";
+	$REG_L="lw";
+	$SZREG=4;
+}
+$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0x00fff000 : 0x00ff0000;
+#
+# <appro@openssl.org>
+#
+######################################################################
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+if ($flavour =~ /64|n32/i) {
+	$LD="ld";
+	$ST="sd";
+	$MULTU="dmultu";
+	$ADDU="daddu";
+	$SUBU="dsubu";
+	$BNSZ=8;
+} else {
+	$LD="lw";
+	$ST="sw";
+	$MULTU="multu";
+	$ADDU="addu";
+	$SUBU="subu";
+	$BNSZ=4;
+}
+
+# int bn_mul_mont(
+$rp=$a0;	# BN_ULONG *rp,
+$ap=$a1;	# const BN_ULONG *ap,
+$bp=$a2;	# const BN_ULONG *bp,
+$np=$a3;	# const BN_ULONG *np,
+$n0=$a4;	# const BN_ULONG *n0,
+$num=$a5;	# int num);
+
+$lo0=$a6;
+$hi0=$a7;
+$lo1=$t1;
+$hi1=$t2;
+$aj=$s0;
+$bi=$s1;
+$nj=$s2;
+$tp=$s3;
+$alo=$s4;
+$ahi=$s5;
+$nlo=$s6;
+$nhi=$s7;
+$tj=$s8;
+$i=$s9;
+$j=$s10;
+$m1=$s11;
+
+$FRAMESIZE=14;
+
+$code=<<___;
+.text
+
+.set	noat
+.set	noreorder
+
+.align	5
+.globl	bn_mul_mont
+.ent	bn_mul_mont
+bn_mul_mont:
+___
+$code.=<<___ if ($flavour =~ /o32/i);
+	lw	$n0,16($sp)
+	lw	$num,20($sp)
+___
+$code.=<<___;
+	slt	$at,$num,4
+	bnez	$at,1f
+	li	$t0,0
+	slt	$at,$num,17	# on in-order CPU
+	bnezl	$at,bn_mul_mont_internal
+	nop
+1:	jr	$ra
+	li	$a0,0
+.end	bn_mul_mont
+
+.align	5
+.ent	bn_mul_mont_internal
+bn_mul_mont_internal:
+	.frame	$fp,$FRAMESIZE*$SZREG,$ra
+	.mask	0x40000000|$SAVED_REGS_MASK,-$SZREG
+	$PTR_SUB $sp,$FRAMESIZE*$SZREG
+	$REG_S	$fp,($FRAMESIZE-1)*$SZREG($sp)
+	$REG_S	$s11,($FRAMESIZE-2)*$SZREG($sp)
+	$REG_S	$s10,($FRAMESIZE-3)*$SZREG($sp)
+	$REG_S	$s9,($FRAMESIZE-4)*$SZREG($sp)
+	$REG_S	$s8,($FRAMESIZE-5)*$SZREG($sp)
+	$REG_S	$s7,($FRAMESIZE-6)*$SZREG($sp)
+	$REG_S	$s6,($FRAMESIZE-7)*$SZREG($sp)
+	$REG_S	$s5,($FRAMESIZE-8)*$SZREG($sp)
+	$REG_S	$s4,($FRAMESIZE-9)*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_S	$s3,($FRAMESIZE-10)*$SZREG($sp)
+	$REG_S	$s2,($FRAMESIZE-11)*$SZREG($sp)
+	$REG_S	$s1,($FRAMESIZE-12)*$SZREG($sp)
+	$REG_S	$s0,($FRAMESIZE-13)*$SZREG($sp)
+___
+$code.=<<___;
+	move	$fp,$sp
+
+	.set	reorder
+	$LD	$n0,0($n0)
+	$LD	$bi,0($bp)	# bp[0]
+	$LD	$aj,0($ap)	# ap[0]
+	$LD	$nj,0($np)	# np[0]
+
+	$PTR_SUB $sp,2*$BNSZ	# place for two extra words
+	sll	$num,`log($BNSZ)/log(2)`
+	li	$at,-4096
+	$PTR_SUB $sp,$num
+	and	$sp,$at
+
+	$MULTU	$aj,$bi
+	$LD	$alo,$BNSZ($ap)
+	$LD	$nlo,$BNSZ($np)
+	mflo	$lo0
+	mfhi	$hi0
+	$MULTU	$lo0,$n0
+	mflo	$m1
+
+	$MULTU	$alo,$bi
+	mflo	$alo
+	mfhi	$ahi
+
+	$MULTU	$nj,$m1
+	mflo	$lo1
+	mfhi	$hi1
+	$MULTU	$nlo,$m1
+	$ADDU	$lo1,$lo0
+	sltu	$at,$lo1,$lo0
+	$ADDU	$hi1,$at
+	mflo	$nlo
+	mfhi	$nhi
+
+	move	$tp,$sp
+	li	$j,2*$BNSZ
+.align	4
+.L1st:
+	.set	noreorder
+	$PTR_ADD $aj,$ap,$j
+	$PTR_ADD $nj,$np,$j
+	$LD	$aj,($aj)
+	$LD	$nj,($nj)
+
+	$MULTU	$aj,$bi
+	$ADDU	$lo0,$alo,$hi0
+	$ADDU	$lo1,$nlo,$hi1
+	sltu	$at,$lo0,$hi0
+	sltu	$t0,$lo1,$hi1
+	$ADDU	$hi0,$ahi,$at
+	$ADDU	$hi1,$nhi,$t0
+	mflo	$alo
+	mfhi	$ahi
+
+	$ADDU	$lo1,$lo0
+	sltu	$at,$lo1,$lo0
+	$MULTU	$nj,$m1
+	$ADDU	$hi1,$at
+	addu	$j,$BNSZ
+	$ST	$lo1,($tp)
+	sltu	$t0,$j,$num
+	mflo	$nlo
+	mfhi	$nhi
+
+	bnez	$t0,.L1st
+	$PTR_ADD $tp,$BNSZ
+	.set	reorder
+
+	$ADDU	$lo0,$alo,$hi0
+	sltu	$at,$lo0,$hi0
+	$ADDU	$hi0,$ahi,$at
+
+	$ADDU	$lo1,$nlo,$hi1
+	sltu	$t0,$lo1,$hi1
+	$ADDU	$hi1,$nhi,$t0
+	$ADDU	$lo1,$lo0
+	sltu	$at,$lo1,$lo0
+	$ADDU	$hi1,$at
+
+	$ST	$lo1,($tp)
+
+	$ADDU	$hi1,$hi0
+	sltu	$at,$hi1,$hi0
+	$ST	$hi1,$BNSZ($tp)
+	$ST	$at,2*$BNSZ($tp)
+
+	li	$i,$BNSZ
+.align	4
+.Louter:
+	$PTR_ADD $bi,$bp,$i
+	$LD	$bi,($bi)
+	$LD	$aj,($ap)
+	$LD	$alo,$BNSZ($ap)
+	$LD	$tj,($sp)
+
+	$MULTU	$aj,$bi
+	$LD	$nj,($np)
+	$LD	$nlo,$BNSZ($np)
+	mflo	$lo0
+	mfhi	$hi0
+	$ADDU	$lo0,$tj
+	$MULTU	$lo0,$n0
+	sltu	$at,$lo0,$tj
+	$ADDU	$hi0,$at
+	mflo	$m1
+
+	$MULTU	$alo,$bi
+	mflo	$alo
+	mfhi	$ahi
+
+	$MULTU	$nj,$m1
+	mflo	$lo1
+	mfhi	$hi1
+
+	$MULTU	$nlo,$m1
+	$ADDU	$lo1,$lo0
+	sltu	$at,$lo1,$lo0
+	$ADDU	$hi1,$at
+	mflo	$nlo
+	mfhi	$nhi
+
+	move	$tp,$sp
+	li	$j,2*$BNSZ
+	$LD	$tj,$BNSZ($tp)
+.align	4
+.Linner:
+	.set	noreorder
+	$PTR_ADD $aj,$ap,$j
+	$PTR_ADD $nj,$np,$j
+	$LD	$aj,($aj)
+	$LD	$nj,($nj)
+
+	$MULTU	$aj,$bi
+	$ADDU	$lo0,$alo,$hi0
+	$ADDU	$lo1,$nlo,$hi1
+	sltu	$at,$lo0,$hi0
+	sltu	$t0,$lo1,$hi1
+	$ADDU	$hi0,$ahi,$at
+	$ADDU	$hi1,$nhi,$t0
+	mflo	$alo
+	mfhi	$ahi
+
+	$ADDU	$lo0,$tj
+	addu	$j,$BNSZ
+	$MULTU	$nj,$m1
+	sltu	$at,$lo0,$tj
+	$ADDU	$lo1,$lo0
+	$ADDU	$hi0,$at
+	sltu	$t0,$lo1,$lo0
+	$LD	$tj,2*$BNSZ($tp)
+	$ADDU	$hi1,$t0
+	sltu	$at,$j,$num
+	mflo	$nlo
+	mfhi	$nhi
+	$ST	$lo1,($tp)
+	bnez	$at,.Linner
+	$PTR_ADD $tp,$BNSZ
+	.set	reorder
+
+	$ADDU	$lo0,$alo,$hi0
+	sltu	$at,$lo0,$hi0
+	$ADDU	$hi0,$ahi,$at
+	$ADDU	$lo0,$tj
+	sltu	$t0,$lo0,$tj
+	$ADDU	$hi0,$t0
+
+	$LD	$tj,2*$BNSZ($tp)
+	$ADDU	$lo1,$nlo,$hi1
+	sltu	$at,$lo1,$hi1
+	$ADDU	$hi1,$nhi,$at
+	$ADDU	$lo1,$lo0
+	sltu	$t0,$lo1,$lo0
+	$ADDU	$hi1,$t0
+	$ST	$lo1,($tp)
+
+	$ADDU	$lo1,$hi1,$hi0
+	sltu	$hi1,$lo1,$hi0
+	$ADDU	$lo1,$tj
+	sltu	$at,$lo1,$tj
+	$ADDU	$hi1,$at
+	$ST	$lo1,$BNSZ($tp)
+	$ST	$hi1,2*$BNSZ($tp)
+
+	addu	$i,$BNSZ
+	sltu	$t0,$i,$num
+	bnez	$t0,.Louter
+
+	.set	noreorder
+	$PTR_ADD $tj,$sp,$num	# &tp[num]
+	move	$tp,$sp
+	move	$ap,$sp
+	li	$hi0,0		# clear borrow bit
+
+.align	4
+.Lsub:	$LD	$lo0,($tp)
+	$LD	$lo1,($np)
+	$PTR_ADD $tp,$BNSZ
+	$PTR_ADD $np,$BNSZ
+	$SUBU	$lo1,$lo0,$lo1	# tp[i]-np[i]
+	sgtu	$at,$lo1,$lo0
+	$SUBU	$lo0,$lo1,$hi0
+	sgtu	$hi0,$lo0,$lo1
+	$ST	$lo0,($rp)
+	or	$hi0,$at
+	sltu	$at,$tp,$tj
+	bnez	$at,.Lsub
+	$PTR_ADD $rp,$BNSZ
+
+	$SUBU	$hi0,$hi1,$hi0	# handle upmost overflow bit
+	move	$tp,$sp
+	$PTR_SUB $rp,$num	# restore rp
+	not	$hi1,$hi0
+
+	and	$ap,$hi0,$sp
+	and	$bp,$hi1,$rp
+	or	$ap,$ap,$bp	# ap=borrow?tp:rp
+
+.align	4
+.Lcopy:	$LD	$aj,($ap)
+	$PTR_ADD $ap,$BNSZ
+	$ST	$zero,($tp)
+	$PTR_ADD $tp,$BNSZ
+	sltu	$at,$tp,$tj
+	$ST	$aj,($rp)
+	bnez	$at,.Lcopy
+	$PTR_ADD $rp,$BNSZ
+
+	li	$a0,1
+	li	$t0,1
+
+	.set	noreorder
+	move	$sp,$fp
+	$REG_L	$fp,($FRAMESIZE-1)*$SZREG($sp)
+	$REG_L	$s11,($FRAMESIZE-2)*$SZREG($sp)
+	$REG_L	$s10,($FRAMESIZE-3)*$SZREG($sp)
+	$REG_L	$s9,($FRAMESIZE-4)*$SZREG($sp)
+	$REG_L	$s8,($FRAMESIZE-5)*$SZREG($sp)
+	$REG_L	$s7,($FRAMESIZE-6)*$SZREG($sp)
+	$REG_L	$s6,($FRAMESIZE-7)*$SZREG($sp)
+	$REG_L	$s5,($FRAMESIZE-8)*$SZREG($sp)
+	$REG_L	$s4,($FRAMESIZE-9)*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$s3,($FRAMESIZE-10)*$SZREG($sp)
+	$REG_L	$s2,($FRAMESIZE-11)*$SZREG($sp)
+	$REG_L	$s1,($FRAMESIZE-12)*$SZREG($sp)
+	$REG_L	$s0,($FRAMESIZE-13)*$SZREG($sp)
+___
+$code.=<<___;
+	jr	$ra
+	$PTR_ADD $sp,$FRAMESIZE*$SZREG
+.end	bn_mul_mont_internal
+.rdata
+.asciiz	"Montgomery Multiplication for MIPS, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/mips.pl b/crypto/bn/asm/mips.pl
new file mode 100755
index 000000000000..c162a3ec2304
--- /dev/null
+++ b/crypto/bn/asm/mips.pl
@@ -0,0 +1,2585 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project.
+#
+# Rights for redistribution and usage in source and binary forms are
+# granted according to the OpenSSL license. Warranty of any kind is
+# disclaimed.
+# ====================================================================
+
+
+# July 1999
+#
+# This is drop-in MIPS III/IV ISA replacement for crypto/bn/bn_asm.c.
+#
+# The module is designed to work with either of the "new" MIPS ABI(5),
+# namely N32 or N64, offered by IRIX 6.x. It's not ment to work under
+# IRIX 5.x not only because it doesn't support new ABIs but also
+# because 5.x kernels put R4x00 CPU into 32-bit mode and all those
+# 64-bit instructions (daddu, dmultu, etc.) found below gonna only
+# cause illegal instruction exception:-(
+#
+# In addition the code depends on preprocessor flags set up by MIPSpro
+# compiler driver (either as or cc) and therefore (probably?) can't be
+# compiled by the GNU assembler. GNU C driver manages fine though...
+# I mean as long as -mmips-as is specified or is the default option,
+# because then it simply invokes /usr/bin/as which in turn takes
+# perfect care of the preprocessor definitions. Another neat feature
+# offered by the MIPSpro assembler is an optimization pass. This gave
+# me the opportunity to have the code looking more regular as all those
+# architecture dependent instruction rescheduling details were left to
+# the assembler. Cool, huh?
+#
+# Performance improvement is astonishing! 'apps/openssl speed rsa dsa'
+# goes way over 3 times faster!
+#
+#					<appro@fy.chalmers.se>
+
+# October 2010
+#
+# Adapt the module even for 32-bit ABIs and other OSes. The former was
+# achieved by mechanical replacement of 64-bit arithmetic instructions
+# such as dmultu, daddu, etc. with their 32-bit counterparts and
+# adjusting offsets denoting multiples of BN_ULONG. Above mentioned
+# >3x performance improvement naturally does not apply to 32-bit code
+# [because there is no instruction 32-bit compiler can't use], one
+# has to content with 40-85% improvement depending on benchmark and
+# key length, more for longer keys.
+
+$flavour = shift;
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+if ($flavour =~ /64|n32/i) {
+	$LD="ld";
+	$ST="sd";
+	$MULTU="dmultu";
+	$DIVU="ddivu";
+	$ADDU="daddu";
+	$SUBU="dsubu";
+	$SRL="dsrl";
+	$SLL="dsll";
+	$BNSZ=8;
+	$PTR_ADD="daddu";
+	$PTR_SUB="dsubu";
+	$SZREG=8;
+	$REG_S="sd";
+	$REG_L="ld";
+} else {
+	$LD="lw";
+	$ST="sw";
+	$MULTU="multu";
+	$DIVU="divu";
+	$ADDU="addu";
+	$SUBU="subu";
+	$SRL="srl";
+	$SLL="sll";
+	$BNSZ=4;
+	$PTR_ADD="addu";
+	$PTR_SUB="subu";
+	$SZREG=4;
+	$REG_S="sw";
+	$REG_L="lw";
+	$code=".set	mips2\n";
+}
+
+# Below is N32/64 register layout used in the original module.
+#
+($zero,$at,$v0,$v1)=map("\$$_",(0..3));
+($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+($t0,$t1,$t2,$t3,$t8,$t9)=map("\$$_",(12..15,24,25));
+($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7)=map("\$$_",(16..23));
+($gp,$sp,$fp,$ra)=map("\$$_",(28..31));
+($ta0,$ta1,$ta2,$ta3)=($a4,$a5,$a6,$a7);
+#
+# No special adaptation is required for O32. NUBI on the other hand
+# is treated by saving/restoring ($v1,$t0..$t3).
+
+$gp=$v1 if ($flavour =~ /nubi/i);
+
+$minus4=$v1;
+
+$code.=<<___;
+.rdata
+.asciiz	"mips3.s, Version 1.2"
+.asciiz	"MIPS II/III/IV ISA artwork by Andy Polyakov <appro\@fy.chalmers.se>"
+
+.text
+.set	noat
+
+.align	5
+.globl	bn_mul_add_words
+.ent	bn_mul_add_words
+bn_mul_add_words:
+	.set	noreorder
+	bgtz	$a2,bn_mul_add_words_internal
+	move	$v0,$zero
+	jr	$ra
+	move	$a0,$v0
+.end	bn_mul_add_words
+
+.align	5
+.ent	bn_mul_add_words_internal
+bn_mul_add_words_internal:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	.set	reorder
+	li	$minus4,-4
+	and	$ta0,$a2,$minus4
+	$LD	$t0,0($a1)
+	beqz	$ta0,.L_bn_mul_add_words_tail
+
+.L_bn_mul_add_words_loop:
+	$MULTU	$t0,$a3
+	$LD	$t1,0($a0)
+	$LD	$t2,$BNSZ($a1)
+	$LD	$t3,$BNSZ($a0)
+	$LD	$ta0,2*$BNSZ($a1)
+	$LD	$ta1,2*$BNSZ($a0)
+	$ADDU	$t1,$v0
+	sltu	$v0,$t1,$v0	# All manuals say it "compares 32-bit
+				# values", but it seems to work fine
+				# even on 64-bit registers.
+	mflo	$at
+	mfhi	$t0
+	$ADDU	$t1,$at
+	$ADDU	$v0,$t0
+	 $MULTU	$t2,$a3
+	sltu	$at,$t1,$at
+	$ST	$t1,0($a0)
+	$ADDU	$v0,$at
+
+	$LD	$ta2,3*$BNSZ($a1)
+	$LD	$ta3,3*$BNSZ($a0)
+	$ADDU	$t3,$v0
+	sltu	$v0,$t3,$v0
+	mflo	$at
+	mfhi	$t2
+	$ADDU	$t3,$at
+	$ADDU	$v0,$t2
+	 $MULTU	$ta0,$a3
+	sltu	$at,$t3,$at
+	$ST	$t3,$BNSZ($a0)
+	$ADDU	$v0,$at
+
+	subu	$a2,4
+	$PTR_ADD $a0,4*$BNSZ
+	$PTR_ADD $a1,4*$BNSZ
+	$ADDU	$ta1,$v0
+	sltu	$v0,$ta1,$v0
+	mflo	$at
+	mfhi	$ta0
+	$ADDU	$ta1,$at
+	$ADDU	$v0,$ta0
+	 $MULTU	$ta2,$a3
+	sltu	$at,$ta1,$at
+	$ST	$ta1,-2*$BNSZ($a0)
+	$ADDU	$v0,$at
+
+
+	and	$ta0,$a2,$minus4
+	$ADDU	$ta3,$v0
+	sltu	$v0,$ta3,$v0
+	mflo	$at
+	mfhi	$ta2
+	$ADDU	$ta3,$at
+	$ADDU	$v0,$ta2
+	sltu	$at,$ta3,$at
+	$ST	$ta3,-$BNSZ($a0)
+	$ADDU	$v0,$at
+	.set	noreorder
+	bgtzl	$ta0,.L_bn_mul_add_words_loop
+	$LD	$t0,0($a1)
+
+	beqz	$a2,.L_bn_mul_add_words_return
+	nop
+
+.L_bn_mul_add_words_tail:
+	.set	reorder
+	$LD	$t0,0($a1)
+	$MULTU	$t0,$a3
+	$LD	$t1,0($a0)
+	subu	$a2,1
+	$ADDU	$t1,$v0
+	sltu	$v0,$t1,$v0
+	mflo	$at
+	mfhi	$t0
+	$ADDU	$t1,$at
+	$ADDU	$v0,$t0
+	sltu	$at,$t1,$at
+	$ST	$t1,0($a0)
+	$ADDU	$v0,$at
+	beqz	$a2,.L_bn_mul_add_words_return
+
+	$LD	$t0,$BNSZ($a1)
+	$MULTU	$t0,$a3
+	$LD	$t1,$BNSZ($a0)
+	subu	$a2,1
+	$ADDU	$t1,$v0
+	sltu	$v0,$t1,$v0
+	mflo	$at
+	mfhi	$t0
+	$ADDU	$t1,$at
+	$ADDU	$v0,$t0
+	sltu	$at,$t1,$at
+	$ST	$t1,$BNSZ($a0)
+	$ADDU	$v0,$at
+	beqz	$a2,.L_bn_mul_add_words_return
+
+	$LD	$t0,2*$BNSZ($a1)
+	$MULTU	$t0,$a3
+	$LD	$t1,2*$BNSZ($a0)
+	$ADDU	$t1,$v0
+	sltu	$v0,$t1,$v0
+	mflo	$at
+	mfhi	$t0
+	$ADDU	$t1,$at
+	$ADDU	$v0,$t0
+	sltu	$at,$t1,$at
+	$ST	$t1,2*$BNSZ($a0)
+	$ADDU	$v0,$at
+
+.L_bn_mul_add_words_return:
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	move	$a0,$v0
+.end	bn_mul_add_words_internal
+
+.align	5
+.globl	bn_mul_words
+.ent	bn_mul_words
+bn_mul_words:
+	.set	noreorder
+	bgtz	$a2,bn_mul_words_internal
+	move	$v0,$zero
+	jr	$ra
+	move	$a0,$v0
+.end	bn_mul_words
+
+.align	5
+.ent	bn_mul_words_internal
+bn_mul_words_internal:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	.set	reorder
+	li	$minus4,-4
+	and	$ta0,$a2,$minus4
+	$LD	$t0,0($a1)
+	beqz	$ta0,.L_bn_mul_words_tail
+
+.L_bn_mul_words_loop:
+	$MULTU	$t0,$a3
+	$LD	$t2,$BNSZ($a1)
+	$LD	$ta0,2*$BNSZ($a1)
+	$LD	$ta2,3*$BNSZ($a1)
+	mflo	$at
+	mfhi	$t0
+	$ADDU	$v0,$at
+	sltu	$t1,$v0,$at
+	 $MULTU	$t2,$a3
+	$ST	$v0,0($a0)
+	$ADDU	$v0,$t1,$t0
+
+	subu	$a2,4
+	$PTR_ADD $a0,4*$BNSZ
+	$PTR_ADD $a1,4*$BNSZ
+	mflo	$at
+	mfhi	$t2
+	$ADDU	$v0,$at
+	sltu	$t3,$v0,$at
+	 $MULTU	$ta0,$a3
+	$ST	$v0,-3*$BNSZ($a0)
+	$ADDU	$v0,$t3,$t2
+
+	mflo	$at
+	mfhi	$ta0
+	$ADDU	$v0,$at
+	sltu	$ta1,$v0,$at
+	 $MULTU	$ta2,$a3
+	$ST	$v0,-2*$BNSZ($a0)
+	$ADDU	$v0,$ta1,$ta0
+
+	and	$ta0,$a2,$minus4
+	mflo	$at
+	mfhi	$ta2
+	$ADDU	$v0,$at
+	sltu	$ta3,$v0,$at
+	$ST	$v0,-$BNSZ($a0)
+	$ADDU	$v0,$ta3,$ta2
+	.set	noreorder
+	bgtzl	$ta0,.L_bn_mul_words_loop
+	$LD	$t0,0($a1)
+
+	beqz	$a2,.L_bn_mul_words_return
+	nop
+
+.L_bn_mul_words_tail:
+	.set	reorder
+	$LD	$t0,0($a1)
+	$MULTU	$t0,$a3
+	subu	$a2,1
+	mflo	$at
+	mfhi	$t0
+	$ADDU	$v0,$at
+	sltu	$t1,$v0,$at
+	$ST	$v0,0($a0)
+	$ADDU	$v0,$t1,$t0
+	beqz	$a2,.L_bn_mul_words_return
+
+	$LD	$t0,$BNSZ($a1)
+	$MULTU	$t0,$a3
+	subu	$a2,1
+	mflo	$at
+	mfhi	$t0
+	$ADDU	$v0,$at
+	sltu	$t1,$v0,$at
+	$ST	$v0,$BNSZ($a0)
+	$ADDU	$v0,$t1,$t0
+	beqz	$a2,.L_bn_mul_words_return
+
+	$LD	$t0,2*$BNSZ($a1)
+	$MULTU	$t0,$a3
+	mflo	$at
+	mfhi	$t0
+	$ADDU	$v0,$at
+	sltu	$t1,$v0,$at
+	$ST	$v0,2*$BNSZ($a0)
+	$ADDU	$v0,$t1,$t0
+
+.L_bn_mul_words_return:
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	move	$a0,$v0
+.end	bn_mul_words_internal
+
+.align	5
+.globl	bn_sqr_words
+.ent	bn_sqr_words
+bn_sqr_words:
+	.set	noreorder
+	bgtz	$a2,bn_sqr_words_internal
+	move	$v0,$zero
+	jr	$ra
+	move	$a0,$v0
+.end	bn_sqr_words
+
+.align	5
+.ent	bn_sqr_words_internal
+bn_sqr_words_internal:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	.set	reorder
+	li	$minus4,-4
+	and	$ta0,$a2,$minus4
+	$LD	$t0,0($a1)
+	beqz	$ta0,.L_bn_sqr_words_tail
+
+.L_bn_sqr_words_loop:
+	$MULTU	$t0,$t0
+	$LD	$t2,$BNSZ($a1)
+	$LD	$ta0,2*$BNSZ($a1)
+	$LD	$ta2,3*$BNSZ($a1)
+	mflo	$t1
+	mfhi	$t0
+	$ST	$t1,0($a0)
+	$ST	$t0,$BNSZ($a0)
+
+	$MULTU	$t2,$t2
+	subu	$a2,4
+	$PTR_ADD $a0,8*$BNSZ
+	$PTR_ADD $a1,4*$BNSZ
+	mflo	$t3
+	mfhi	$t2
+	$ST	$t3,-6*$BNSZ($a0)
+	$ST	$t2,-5*$BNSZ($a0)
+
+	$MULTU	$ta0,$ta0
+	mflo	$ta1
+	mfhi	$ta0
+	$ST	$ta1,-4*$BNSZ($a0)
+	$ST	$ta0,-3*$BNSZ($a0)
+
+
+	$MULTU	$ta2,$ta2
+	and	$ta0,$a2,$minus4
+	mflo	$ta3
+	mfhi	$ta2
+	$ST	$ta3,-2*$BNSZ($a0)
+	$ST	$ta2,-$BNSZ($a0)
+
+	.set	noreorder
+	bgtzl	$ta0,.L_bn_sqr_words_loop
+	$LD	$t0,0($a1)
+
+	beqz	$a2,.L_bn_sqr_words_return
+	nop
+
+.L_bn_sqr_words_tail:
+	.set	reorder
+	$LD	$t0,0($a1)
+	$MULTU	$t0,$t0
+	subu	$a2,1
+	mflo	$t1
+	mfhi	$t0
+	$ST	$t1,0($a0)
+	$ST	$t0,$BNSZ($a0)
+	beqz	$a2,.L_bn_sqr_words_return
+
+	$LD	$t0,$BNSZ($a1)
+	$MULTU	$t0,$t0
+	subu	$a2,1
+	mflo	$t1
+	mfhi	$t0
+	$ST	$t1,2*$BNSZ($a0)
+	$ST	$t0,3*$BNSZ($a0)
+	beqz	$a2,.L_bn_sqr_words_return
+
+	$LD	$t0,2*$BNSZ($a1)
+	$MULTU	$t0,$t0
+	mflo	$t1
+	mfhi	$t0
+	$ST	$t1,4*$BNSZ($a0)
+	$ST	$t0,5*$BNSZ($a0)
+
+.L_bn_sqr_words_return:
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	move	$a0,$v0
+
+.end	bn_sqr_words_internal
+
+.align	5
+.globl	bn_add_words
+.ent	bn_add_words
+bn_add_words:
+	.set	noreorder
+	bgtz	$a3,bn_add_words_internal
+	move	$v0,$zero
+	jr	$ra
+	move	$a0,$v0
+.end	bn_add_words
+
+.align	5
+.ent	bn_add_words_internal
+bn_add_words_internal:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	.set	reorder
+	li	$minus4,-4
+	and	$at,$a3,$minus4
+	$LD	$t0,0($a1)
+	beqz	$at,.L_bn_add_words_tail
+
+.L_bn_add_words_loop:
+	$LD	$ta0,0($a2)
+	subu	$a3,4
+	$LD	$t1,$BNSZ($a1)
+	and	$at,$a3,$minus4
+	$LD	$t2,2*$BNSZ($a1)
+	$PTR_ADD $a2,4*$BNSZ
+	$LD	$t3,3*$BNSZ($a1)
+	$PTR_ADD $a0,4*$BNSZ
+	$LD	$ta1,-3*$BNSZ($a2)
+	$PTR_ADD $a1,4*$BNSZ
+	$LD	$ta2,-2*$BNSZ($a2)
+	$LD	$ta3,-$BNSZ($a2)
+	$ADDU	$ta0,$t0
+	sltu	$t8,$ta0,$t0
+	$ADDU	$t0,$ta0,$v0
+	sltu	$v0,$t0,$ta0
+	$ST	$t0,-4*$BNSZ($a0)
+	$ADDU	$v0,$t8
+
+	$ADDU	$ta1,$t1
+	sltu	$t9,$ta1,$t1
+	$ADDU	$t1,$ta1,$v0
+	sltu	$v0,$t1,$ta1
+	$ST	$t1,-3*$BNSZ($a0)
+	$ADDU	$v0,$t9
+
+	$ADDU	$ta2,$t2
+	sltu	$t8,$ta2,$t2
+	$ADDU	$t2,$ta2,$v0
+	sltu	$v0,$t2,$ta2
+	$ST	$t2,-2*$BNSZ($a0)
+	$ADDU	$v0,$t8
+	
+	$ADDU	$ta3,$t3
+	sltu	$t9,$ta3,$t3
+	$ADDU	$t3,$ta3,$v0
+	sltu	$v0,$t3,$ta3
+	$ST	$t3,-$BNSZ($a0)
+	$ADDU	$v0,$t9
+	
+	.set	noreorder
+	bgtzl	$at,.L_bn_add_words_loop
+	$LD	$t0,0($a1)
+
+	beqz	$a3,.L_bn_add_words_return
+	nop
+
+.L_bn_add_words_tail:
+	.set	reorder
+	$LD	$t0,0($a1)
+	$LD	$ta0,0($a2)
+	$ADDU	$ta0,$t0
+	subu	$a3,1
+	sltu	$t8,$ta0,$t0
+	$ADDU	$t0,$ta0,$v0
+	sltu	$v0,$t0,$ta0
+	$ST	$t0,0($a0)
+	$ADDU	$v0,$t8
+	beqz	$a3,.L_bn_add_words_return
+
+	$LD	$t1,$BNSZ($a1)
+	$LD	$ta1,$BNSZ($a2)
+	$ADDU	$ta1,$t1
+	subu	$a3,1
+	sltu	$t9,$ta1,$t1
+	$ADDU	$t1,$ta1,$v0
+	sltu	$v0,$t1,$ta1
+	$ST	$t1,$BNSZ($a0)
+	$ADDU	$v0,$t9
+	beqz	$a3,.L_bn_add_words_return
+
+	$LD	$t2,2*$BNSZ($a1)
+	$LD	$ta2,2*$BNSZ($a2)
+	$ADDU	$ta2,$t2
+	sltu	$t8,$ta2,$t2
+	$ADDU	$t2,$ta2,$v0
+	sltu	$v0,$t2,$ta2
+	$ST	$t2,2*$BNSZ($a0)
+	$ADDU	$v0,$t8
+
+.L_bn_add_words_return:
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	move	$a0,$v0
+
+.end	bn_add_words_internal
+
+.align	5
+.globl	bn_sub_words
+.ent	bn_sub_words
+bn_sub_words:
+	.set	noreorder
+	bgtz	$a3,bn_sub_words_internal
+	move	$v0,$zero
+	jr	$ra
+	move	$a0,$zero
+.end	bn_sub_words
+
+.align	5
+.ent	bn_sub_words_internal
+bn_sub_words_internal:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	.set	reorder
+	li	$minus4,-4
+	and	$at,$a3,$minus4
+	$LD	$t0,0($a1)
+	beqz	$at,.L_bn_sub_words_tail
+
+.L_bn_sub_words_loop:
+	$LD	$ta0,0($a2)
+	subu	$a3,4
+	$LD	$t1,$BNSZ($a1)
+	and	$at,$a3,$minus4
+	$LD	$t2,2*$BNSZ($a1)
+	$PTR_ADD $a2,4*$BNSZ
+	$LD	$t3,3*$BNSZ($a1)
+	$PTR_ADD $a0,4*$BNSZ
+	$LD	$ta1,-3*$BNSZ($a2)
+	$PTR_ADD $a1,4*$BNSZ
+	$LD	$ta2,-2*$BNSZ($a2)
+	$LD	$ta3,-$BNSZ($a2)
+	sltu	$t8,$t0,$ta0
+	$SUBU	$ta0,$t0,$ta0
+	$SUBU	$t0,$ta0,$v0
+	sgtu	$v0,$t0,$ta0
+	$ST	$t0,-4*$BNSZ($a0)
+	$ADDU	$v0,$t8
+
+	sltu	$t9,$t1,$ta1
+	$SUBU	$ta1,$t1,$ta1
+	$SUBU	$t1,$ta1,$v0
+	sgtu	$v0,$t1,$ta1
+	$ST	$t1,-3*$BNSZ($a0)
+	$ADDU	$v0,$t9
+
+
+	sltu	$t8,$t2,$ta2
+	$SUBU	$ta2,$t2,$ta2
+	$SUBU	$t2,$ta2,$v0
+	sgtu	$v0,$t2,$ta2
+	$ST	$t2,-2*$BNSZ($a0)
+	$ADDU	$v0,$t8
+
+	sltu	$t9,$t3,$ta3
+	$SUBU	$ta3,$t3,$ta3
+	$SUBU	$t3,$ta3,$v0
+	sgtu	$v0,$t3,$ta3
+	$ST	$t3,-$BNSZ($a0)
+	$ADDU	$v0,$t9
+
+	.set	noreorder
+	bgtzl	$at,.L_bn_sub_words_loop
+	$LD	$t0,0($a1)
+
+	beqz	$a3,.L_bn_sub_words_return
+	nop
+
+.L_bn_sub_words_tail:
+	.set	reorder
+	$LD	$t0,0($a1)
+	$LD	$ta0,0($a2)
+	subu	$a3,1
+	sltu	$t8,$t0,$ta0
+	$SUBU	$ta0,$t0,$ta0
+	$SUBU	$t0,$ta0,$v0
+	sgtu	$v0,$t0,$ta0
+	$ST	$t0,0($a0)
+	$ADDU	$v0,$t8
+	beqz	$a3,.L_bn_sub_words_return
+
+	$LD	$t1,$BNSZ($a1)
+	subu	$a3,1
+	$LD	$ta1,$BNSZ($a2)
+	sltu	$t9,$t1,$ta1
+	$SUBU	$ta1,$t1,$ta1
+	$SUBU	$t1,$ta1,$v0
+	sgtu	$v0,$t1,$ta1
+	$ST	$t1,$BNSZ($a0)
+	$ADDU	$v0,$t9
+	beqz	$a3,.L_bn_sub_words_return
+
+	$LD	$t2,2*$BNSZ($a1)
+	$LD	$ta2,2*$BNSZ($a2)
+	sltu	$t8,$t2,$ta2
+	$SUBU	$ta2,$t2,$ta2
+	$SUBU	$t2,$ta2,$v0
+	sgtu	$v0,$t2,$ta2
+	$ST	$t2,2*$BNSZ($a0)
+	$ADDU	$v0,$t8
+
+.L_bn_sub_words_return:
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	move	$a0,$v0
+.end	bn_sub_words_internal
+
+.align 5
+.globl	bn_div_3_words
+.ent	bn_div_3_words
+bn_div_3_words:
+	.set	noreorder
+	move	$a3,$a0		# we know that bn_div_words does not
+				# touch $a3, $ta2, $ta3 and preserves $a2
+				# so that we can save two arguments
+				# and return address in registers
+				# instead of stack:-)
+				
+	$LD	$a0,($a3)
+	move	$ta2,$a1
+	bne	$a0,$a2,bn_div_3_words_internal
+	$LD	$a1,-$BNSZ($a3)
+	li	$v0,-1
+	jr	$ra
+	move	$a0,$v0
+.end	bn_div_3_words
+
+.align	5
+.ent	bn_div_3_words_internal
+bn_div_3_words_internal:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	.set	reorder
+	move	$ta3,$ra
+	bal	bn_div_words
+	move	$ra,$ta3
+	$MULTU	$ta2,$v0
+	$LD	$t2,-2*$BNSZ($a3)
+	move	$ta0,$zero
+	mfhi	$t1
+	mflo	$t0
+	sltu	$t8,$t1,$a1
+.L_bn_div_3_words_inner_loop:
+	bnez	$t8,.L_bn_div_3_words_inner_loop_done
+	sgeu	$at,$t2,$t0
+	seq	$t9,$t1,$a1
+	and	$at,$t9
+	sltu	$t3,$t0,$ta2
+	$ADDU	$a1,$a2
+	$SUBU	$t1,$t3
+	$SUBU	$t0,$ta2
+	sltu	$t8,$t1,$a1
+	sltu	$ta0,$a1,$a2
+	or	$t8,$ta0
+	.set	noreorder
+	beqzl	$at,.L_bn_div_3_words_inner_loop
+	$SUBU	$v0,1
+	.set	reorder
+.L_bn_div_3_words_inner_loop_done:
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	move	$a0,$v0
+.end	bn_div_3_words_internal
+
+.align	5
+.globl	bn_div_words
+.ent	bn_div_words
+bn_div_words:
+	.set	noreorder
+	bnez	$a2,bn_div_words_internal
+	li	$v0,-1		# I would rather signal div-by-zero
+				# which can be done with 'break 7'
+	jr	$ra
+	move	$a0,$v0
+.end	bn_div_words
+
+.align	5
+.ent	bn_div_words_internal
+bn_div_words_internal:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	move	$v1,$zero
+	bltz	$a2,.L_bn_div_words_body
+	move	$t9,$v1
+	$SLL	$a2,1
+	bgtz	$a2,.-4
+	addu	$t9,1
+
+	.set	reorder
+	negu	$t1,$t9
+	li	$t2,-1
+	$SLL	$t2,$t1
+	and	$t2,$a0
+	$SRL	$at,$a1,$t1
+	.set	noreorder
+	bnezl	$t2,.+8
+	break	6		# signal overflow
+	.set	reorder
+	$SLL	$a0,$t9
+	$SLL	$a1,$t9
+	or	$a0,$at
+___
+$QT=$ta0;
+$HH=$ta1;
+$DH=$v1;
+$code.=<<___;
+.L_bn_div_words_body:
+	$SRL	$DH,$a2,4*$BNSZ	# bits
+	sgeu	$at,$a0,$a2
+	.set	noreorder
+	bnezl	$at,.+8
+	$SUBU	$a0,$a2
+	.set	reorder
+
+	li	$QT,-1
+	$SRL	$HH,$a0,4*$BNSZ	# bits
+	$SRL	$QT,4*$BNSZ	# q=0xffffffff
+	beq	$DH,$HH,.L_bn_div_words_skip_div1
+	$DIVU	$zero,$a0,$DH
+	mflo	$QT
+.L_bn_div_words_skip_div1:
+	$MULTU	$a2,$QT
+	$SLL	$t3,$a0,4*$BNSZ	# bits
+	$SRL	$at,$a1,4*$BNSZ	# bits
+	or	$t3,$at
+	mflo	$t0
+	mfhi	$t1
+.L_bn_div_words_inner_loop1:
+	sltu	$t2,$t3,$t0
+	seq	$t8,$HH,$t1
+	sltu	$at,$HH,$t1
+	and	$t2,$t8
+	sltu	$v0,$t0,$a2
+	or	$at,$t2
+	.set	noreorder
+	beqz	$at,.L_bn_div_words_inner_loop1_done
+	$SUBU	$t1,$v0
+	$SUBU	$t0,$a2
+	b	.L_bn_div_words_inner_loop1
+	$SUBU	$QT,1
+	.set	reorder
+.L_bn_div_words_inner_loop1_done:
+
+	$SLL	$a1,4*$BNSZ	# bits
+	$SUBU	$a0,$t3,$t0
+	$SLL	$v0,$QT,4*$BNSZ	# bits
+
+	li	$QT,-1
+	$SRL	$HH,$a0,4*$BNSZ	# bits
+	$SRL	$QT,4*$BNSZ	# q=0xffffffff
+	beq	$DH,$HH,.L_bn_div_words_skip_div2
+	$DIVU	$zero,$a0,$DH
+	mflo	$QT
+.L_bn_div_words_skip_div2:
+	$MULTU	$a2,$QT
+	$SLL	$t3,$a0,4*$BNSZ	# bits
+	$SRL	$at,$a1,4*$BNSZ	# bits
+	or	$t3,$at
+	mflo	$t0
+	mfhi	$t1
+.L_bn_div_words_inner_loop2:
+	sltu	$t2,$t3,$t0
+	seq	$t8,$HH,$t1
+	sltu	$at,$HH,$t1
+	and	$t2,$t8
+	sltu	$v1,$t0,$a2
+	or	$at,$t2
+	.set	noreorder
+	beqz	$at,.L_bn_div_words_inner_loop2_done
+	$SUBU	$t1,$v1
+	$SUBU	$t0,$a2
+	b	.L_bn_div_words_inner_loop2
+	$SUBU	$QT,1
+	.set	reorder
+.L_bn_div_words_inner_loop2_done:
+
+	$SUBU	$a0,$t3,$t0
+	or	$v0,$QT
+	$SRL	$v1,$a0,$t9	# $v1 contains remainder if anybody wants it
+	$SRL	$a2,$t9		# restore $a2
+
+	.set	noreorder
+	move	$a1,$v1
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	move	$a0,$v0
+.end	bn_div_words_internal
+___
+undef $HH; undef $QT; undef $DH;
+
+($a_0,$a_1,$a_2,$a_3)=($t0,$t1,$t2,$t3);
+($b_0,$b_1,$b_2,$b_3)=($ta0,$ta1,$ta2,$ta3);
+
+($a_4,$a_5,$a_6,$a_7)=($s0,$s2,$s4,$a1); # once we load a[7], no use for $a1
+($b_4,$b_5,$b_6,$b_7)=($s1,$s3,$s5,$a2); # once we load b[7], no use for $a2
+
+($t_1,$t_2,$c_1,$c_2,$c_3)=($t8,$t9,$v0,$v1,$a3);
+
+$code.=<<___;
+
+.align	5
+.globl	bn_mul_comba8
+.ent	bn_mul_comba8
+bn_mul_comba8:
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,12*$SZREG,$ra
+	.mask	0x803ff008,-$SZREG
+	$PTR_SUB $sp,12*$SZREG
+	$REG_S	$ra,11*$SZREG($sp)
+	$REG_S	$s5,10*$SZREG($sp)
+	$REG_S	$s4,9*$SZREG($sp)
+	$REG_S	$s3,8*$SZREG($sp)
+	$REG_S	$s2,7*$SZREG($sp)
+	$REG_S	$s1,6*$SZREG($sp)
+	$REG_S	$s0,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___ if ($flavour !~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x003f0000,-$SZREG
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$s5,5*$SZREG($sp)
+	$REG_S	$s4,4*$SZREG($sp)
+	$REG_S	$s3,3*$SZREG($sp)
+	$REG_S	$s2,2*$SZREG($sp)
+	$REG_S	$s1,1*$SZREG($sp)
+	$REG_S	$s0,0*$SZREG($sp)
+___
+$code.=<<___;
+
+	.set	reorder
+	$LD	$a_0,0($a1)	# If compiled with -mips3 option on
+				# R5000 box assembler barks on this
+				# 1ine with "should not have mult/div
+				# as last instruction in bb (R10K
+				# bug)" warning. If anybody out there
+				# has a clue about how to circumvent
+				# this do send me a note.
+				#		<appro\@fy.chalmers.se>
+
+	$LD	$b_0,0($a2)
+	$LD	$a_1,$BNSZ($a1)
+	$LD	$a_2,2*$BNSZ($a1)
+	$MULTU	$a_0,$b_0		# mul_add_c(a[0],b[0],c1,c2,c3);
+	$LD	$a_3,3*$BNSZ($a1)
+	$LD	$b_1,$BNSZ($a2)
+	$LD	$b_2,2*$BNSZ($a2)
+	$LD	$b_3,3*$BNSZ($a2)
+	mflo	$c_1
+	mfhi	$c_2
+
+	$LD	$a_4,4*$BNSZ($a1)
+	$LD	$a_5,5*$BNSZ($a1)
+	$MULTU	$a_0,$b_1		# mul_add_c(a[0],b[1],c2,c3,c1);
+	$LD	$a_6,6*$BNSZ($a1)
+	$LD	$a_7,7*$BNSZ($a1)
+	$LD	$b_4,4*$BNSZ($a2)
+	$LD	$b_5,5*$BNSZ($a2)
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_1,$b_0		# mul_add_c(a[1],b[0],c2,c3,c1);
+	$ADDU	$c_3,$t_2,$at
+	$LD	$b_6,6*$BNSZ($a2)
+	$LD	$b_7,7*$BNSZ($a2)
+	$ST	$c_1,0($a0)	# r[0]=c1;
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	 $MULTU	$a_2,$b_0		# mul_add_c(a[2],b[0],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$c_1,$c_3,$t_2
+	$ST	$c_2,$BNSZ($a0)	# r[1]=c2;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_1,$b_1		# mul_add_c(a[1],b[1],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_0,$b_2		# mul_add_c(a[0],b[2],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$c_2,$c_1,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	 $MULTU	$a_0,$b_3		# mul_add_c(a[0],b[3],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,2*$BNSZ($a0)	# r[2]=c3;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_1,$b_2		# mul_add_c(a[1],b[2],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$c_3,$c_2,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_2,$b_1		# mul_add_c(a[2],b[1],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_3,$b_0		# mul_add_c(a[3],b[0],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	 $MULTU	$a_4,$b_0		# mul_add_c(a[4],b[0],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,3*$BNSZ($a0)	# r[3]=c1;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_3,$b_1		# mul_add_c(a[3],b[1],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$c_1,$c_3,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_2,$b_2		# mul_add_c(a[2],b[2],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_1,$b_3		# mul_add_c(a[1],b[3],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_0,$b_4		# mul_add_c(a[0],b[4],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	 $MULTU	$a_0,$b_5		# mul_add_c(a[0],b[5],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,4*$BNSZ($a0)	# r[4]=c2;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_1,$b_4		# mul_add_c(a[1],b[4],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$c_2,$c_1,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_2,$b_3		# mul_add_c(a[2],b[3],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_3,$b_2		# mul_add_c(a[3],b[2],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_4,$b_1		# mul_add_c(a[4],b[1],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_5,$b_0		# mul_add_c(a[5],b[0],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	 $MULTU	$a_6,$b_0		# mul_add_c(a[6],b[0],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,5*$BNSZ($a0)	# r[5]=c3;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_5,$b_1		# mul_add_c(a[5],b[1],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$c_3,$c_2,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_4,$b_2		# mul_add_c(a[4],b[2],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_3,$b_3		# mul_add_c(a[3],b[3],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_2,$b_4		# mul_add_c(a[2],b[4],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_1,$b_5		# mul_add_c(a[1],b[5],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_0,$b_6		# mul_add_c(a[0],b[6],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	 $MULTU	$a_0,$b_7		# mul_add_c(a[0],b[7],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,6*$BNSZ($a0)	# r[6]=c1;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_1,$b_6		# mul_add_c(a[1],b[6],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$c_1,$c_3,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_2,$b_5		# mul_add_c(a[2],b[5],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_3,$b_4		# mul_add_c(a[3],b[4],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_4,$b_3		# mul_add_c(a[4],b[3],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_5,$b_2		# mul_add_c(a[5],b[2],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_6,$b_1		# mul_add_c(a[6],b[1],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_7,$b_0		# mul_add_c(a[7],b[0],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	 $MULTU	$a_7,$b_1		# mul_add_c(a[7],b[1],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,7*$BNSZ($a0)	# r[7]=c2;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_6,$b_2		# mul_add_c(a[6],b[2],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$c_2,$c_1,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_5,$b_3		# mul_add_c(a[5],b[3],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_4,$b_4		# mul_add_c(a[4],b[4],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_3,$b_5		# mul_add_c(a[3],b[5],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_2,$b_6		# mul_add_c(a[2],b[6],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_1,$b_7		# mul_add_c(a[1],b[7],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	 $MULTU	$a_2,$b_7		# mul_add_c(a[2],b[7],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,8*$BNSZ($a0)	# r[8]=c3;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_3,$b_6		# mul_add_c(a[3],b[6],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$c_3,$c_2,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_4,$b_5		# mul_add_c(a[4],b[5],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_5,$b_4		# mul_add_c(a[5],b[4],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_6,$b_3		# mul_add_c(a[6],b[3],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_7,$b_2		# mul_add_c(a[7],b[2],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	 $MULTU	$a_7,$b_3		# mul_add_c(a[7],b[3],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,9*$BNSZ($a0)	# r[9]=c1;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_6,$b_4		# mul_add_c(a[6],b[4],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$c_1,$c_3,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_5,$b_5		# mul_add_c(a[5],b[5],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_4,$b_6		# mul_add_c(a[4],b[6],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_3,$b_7		# mul_add_c(a[3],b[7],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_4,$b_7		# mul_add_c(a[4],b[7],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,10*$BNSZ($a0)	# r[10]=c2;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_5,$b_6		# mul_add_c(a[5],b[6],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$c_2,$c_1,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_6,$b_5		# mul_add_c(a[6],b[5],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_7,$b_4		# mul_add_c(a[7],b[4],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	 $MULTU	$a_7,$b_5		# mul_add_c(a[7],b[5],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,11*$BNSZ($a0)	# r[11]=c3;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_6,$b_6		# mul_add_c(a[6],b[6],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$c_3,$c_2,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_5,$b_7		# mul_add_c(a[5],b[7],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	 $MULTU	$a_6,$b_7		# mul_add_c(a[6],b[7],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,12*$BNSZ($a0)	# r[12]=c1;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_7,$b_6		# mul_add_c(a[7],b[6],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$c_1,$c_3,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_7,$b_7		# mul_add_c(a[7],b[7],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,13*$BNSZ($a0)	# r[13]=c2;
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	$ST	$c_3,14*$BNSZ($a0)	# r[14]=c3;
+	$ST	$c_1,15*$BNSZ($a0)	# r[15]=c1;
+
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$s5,10*$SZREG($sp)
+	$REG_L	$s4,9*$SZREG($sp)
+	$REG_L	$s3,8*$SZREG($sp)
+	$REG_L	$s2,7*$SZREG($sp)
+	$REG_L	$s1,6*$SZREG($sp)
+	$REG_L	$s0,5*$SZREG($sp)
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	jr	$ra
+	$PTR_ADD $sp,12*$SZREG
+___
+$code.=<<___ if ($flavour !~ /nubi/i);
+	$REG_L	$s5,5*$SZREG($sp)
+	$REG_L	$s4,4*$SZREG($sp)
+	$REG_L	$s3,3*$SZREG($sp)
+	$REG_L	$s2,2*$SZREG($sp)
+	$REG_L	$s1,1*$SZREG($sp)
+	$REG_L	$s0,0*$SZREG($sp)
+	jr	$ra
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+.end	bn_mul_comba8
+
+.align	5
+.globl	bn_mul_comba4
+.ent	bn_mul_comba4
+bn_mul_comba4:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	.set	reorder
+	$LD	$a_0,0($a1)
+	$LD	$b_0,0($a2)
+	$LD	$a_1,$BNSZ($a1)
+	$LD	$a_2,2*$BNSZ($a1)
+	$MULTU	$a_0,$b_0		# mul_add_c(a[0],b[0],c1,c2,c3);
+	$LD	$a_3,3*$BNSZ($a1)
+	$LD	$b_1,$BNSZ($a2)
+	$LD	$b_2,2*$BNSZ($a2)
+	$LD	$b_3,3*$BNSZ($a2)
+	mflo	$c_1
+	mfhi	$c_2
+	$ST	$c_1,0($a0)
+
+	$MULTU	$a_0,$b_1		# mul_add_c(a[0],b[1],c2,c3,c1);
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_1,$b_0		# mul_add_c(a[1],b[0],c2,c3,c1);
+	$ADDU	$c_3,$t_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	 $MULTU	$a_2,$b_0		# mul_add_c(a[2],b[0],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$c_1,$c_3,$t_2
+	$ST	$c_2,$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_1,$b_1		# mul_add_c(a[1],b[1],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_0,$b_2		# mul_add_c(a[0],b[2],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$c_2,$c_1,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	 $MULTU	$a_0,$b_3		# mul_add_c(a[0],b[3],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,2*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_1,$b_2		# mul_add_c(a[1],b[2],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$c_3,$c_2,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_2,$b_1		# mul_add_c(a[2],b[1],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$MULTU	$a_3,$b_0		# mul_add_c(a[3],b[0],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	 $MULTU	$a_3,$b_1		# mul_add_c(a[3],b[1],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,3*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_2,$b_2		# mul_add_c(a[2],b[2],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$c_1,$c_3,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$MULTU	$a_1,$b_3		# mul_add_c(a[1],b[3],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	 $MULTU	$a_2,$b_3		# mul_add_c(a[2],b[3],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,4*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$MULTU	$a_3,$b_2		# mul_add_c(a[3],b[2],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$c_2,$c_1,$t_2
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	 $MULTU	$a_3,$b_3		# mul_add_c(a[3],b[3],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,5*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	$ST	$c_1,6*$BNSZ($a0)
+	$ST	$c_2,7*$BNSZ($a0)
+
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	nop
+.end	bn_mul_comba4
+___
+
+($a_4,$a_5,$a_6,$a_7)=($b_0,$b_1,$b_2,$b_3);
+
+$code.=<<___;
+
+.align	5
+.globl	bn_sqr_comba8
+.ent	bn_sqr_comba8
+bn_sqr_comba8:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	.set	reorder
+	$LD	$a_0,0($a1)
+	$LD	$a_1,$BNSZ($a1)
+	$LD	$a_2,2*$BNSZ($a1)
+	$LD	$a_3,3*$BNSZ($a1)
+
+	$MULTU	$a_0,$a_0		# mul_add_c(a[0],b[0],c1,c2,c3);
+	$LD	$a_4,4*$BNSZ($a1)
+	$LD	$a_5,5*$BNSZ($a1)
+	$LD	$a_6,6*$BNSZ($a1)
+	$LD	$a_7,7*$BNSZ($a1)
+	mflo	$c_1
+	mfhi	$c_2
+	$ST	$c_1,0($a0)
+
+	$MULTU	$a_0,$a_1		# mul_add_c2(a[0],b[1],c2,c3,c1);
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_1,$t_2,$zero
+	$SLL	$t_2,1
+	 $MULTU	$a_2,$a_0		# mul_add_c2(a[2],b[0],c3,c1,c2);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$c_3,$t_2,$at
+	$ST	$c_2,$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_2,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_1,$a_1		# mul_add_c(a[1],b[1],c3,c1,c2);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	 $MULTU	$a_0,$a_3		# mul_add_c2(a[0],b[3],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,2*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_3,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_1,$a_2		# mul_add_c2(a[1],b[2],c1,c2,c3);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_3,$at
+	 $MULTU	$a_4,$a_0		# mul_add_c2(a[4],b[0],c2,c3,c1);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,3*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_1,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_3,$a_1		# mul_add_c2(a[3],b[1],c2,c3,c1);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_1,$at
+	$MULTU	$a_2,$a_2		# mul_add_c(a[2],b[2],c2,c3,c1);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	 $MULTU	$a_0,$a_5		# mul_add_c2(a[0],b[5],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,4*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_2,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_1,$a_4		# mul_add_c2(a[1],b[4],c3,c1,c2);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_2,$at
+	$MULTU	$a_2,$a_3		# mul_add_c2(a[2],b[3],c3,c1,c2);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	 $MULTU	$a_6,$a_0		# mul_add_c2(a[6],b[0],c1,c2,c3);
+	$ADDU	$c_2,$at
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,5*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_3,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_5,$a_1		# mul_add_c2(a[5],b[1],c1,c2,c3);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_3,$at
+	$MULTU	$a_4,$a_2		# mul_add_c2(a[4],b[2],c1,c2,c3);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_3,$at
+	$MULTU	$a_3,$a_3		# mul_add_c(a[3],b[3],c1,c2,c3);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	 $MULTU	$a_0,$a_7		# mul_add_c2(a[0],b[7],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,6*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_1,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_1,$a_6		# mul_add_c2(a[1],b[6],c2,c3,c1);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_1,$at
+	$MULTU	$a_2,$a_5		# mul_add_c2(a[2],b[5],c2,c3,c1);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_1,$at
+	$MULTU	$a_3,$a_4		# mul_add_c2(a[3],b[4],c2,c3,c1);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_1,$at
+	 $MULTU	$a_7,$a_1		# mul_add_c2(a[7],b[1],c3,c1,c2);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,7*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_2,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_6,$a_2		# mul_add_c2(a[6],b[2],c3,c1,c2);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_2,$at
+	$MULTU	$a_5,$a_3		# mul_add_c2(a[5],b[3],c3,c1,c2);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_2,$at
+	$MULTU	$a_4,$a_4		# mul_add_c(a[4],b[4],c3,c1,c2);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	 $MULTU	$a_2,$a_7		# mul_add_c2(a[2],b[7],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,8*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_3,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_3,$a_6		# mul_add_c2(a[3],b[6],c1,c2,c3);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_3,$at
+	$MULTU	$a_4,$a_5		# mul_add_c2(a[4],b[5],c1,c2,c3);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_3,$at
+	 $MULTU	$a_7,$a_3		# mul_add_c2(a[7],b[3],c2,c3,c1);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,9*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_1,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_6,$a_4		# mul_add_c2(a[6],b[4],c2,c3,c1);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_1,$at
+	$MULTU	$a_5,$a_5		# mul_add_c(a[5],b[5],c2,c3,c1);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	 $MULTU	$a_4,$a_7		# mul_add_c2(a[4],b[7],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,10*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_2,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_5,$a_6		# mul_add_c2(a[5],b[6],c3,c1,c2);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_2,$at
+	 $MULTU	$a_7,$a_5		# mul_add_c2(a[7],b[5],c1,c2,c3);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,11*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_3,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_6,$a_6		# mul_add_c(a[6],b[6],c1,c2,c3);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	 $MULTU	$a_6,$a_7		# mul_add_c2(a[6],b[7],c2,c3,c1);
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,12*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_1,$t_2,$zero
+	$SLL	$t_2,1
+	 $MULTU	$a_7,$a_7		# mul_add_c(a[7],b[7],c3,c1,c2);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,13*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	$ST	$c_3,14*$BNSZ($a0)
+	$ST	$c_1,15*$BNSZ($a0)
+
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	nop
+.end	bn_sqr_comba8
+
+.align	5
+.globl	bn_sqr_comba4
+.ent	bn_sqr_comba4
+bn_sqr_comba4:
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	.frame	$sp,6*$SZREG,$ra
+	.mask	0x8000f008,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,6*$SZREG
+	$REG_S	$ra,5*$SZREG($sp)
+	$REG_S	$t3,4*$SZREG($sp)
+	$REG_S	$t2,3*$SZREG($sp)
+	$REG_S	$t1,2*$SZREG($sp)
+	$REG_S	$t0,1*$SZREG($sp)
+	$REG_S	$gp,0*$SZREG($sp)
+___
+$code.=<<___;
+	.set	reorder
+	$LD	$a_0,0($a1)
+	$LD	$a_1,$BNSZ($a1)
+	$MULTU	$a_0,$a_0		# mul_add_c(a[0],b[0],c1,c2,c3);
+	$LD	$a_2,2*$BNSZ($a1)
+	$LD	$a_3,3*$BNSZ($a1)
+	mflo	$c_1
+	mfhi	$c_2
+	$ST	$c_1,0($a0)
+
+	$MULTU	$a_0,$a_1		# mul_add_c2(a[0],b[1],c2,c3,c1);
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_1,$t_2,$zero
+	$SLL	$t_2,1
+	 $MULTU	$a_2,$a_0		# mul_add_c2(a[2],b[0],c3,c1,c2);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$c_3,$t_2,$at
+	$ST	$c_2,$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_2,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_1,$a_1		# mul_add_c(a[1],b[1],c3,c1,c2);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	 $MULTU	$a_0,$a_3		# mul_add_c2(a[0],b[3],c1,c2,c3);
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,2*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_3,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_1,$a_2		# mul_add_c(a2[1],b[2],c1,c2,c3);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$at,$t_2,$zero
+	$ADDU	$c_3,$at
+	 $MULTU	$a_3,$a_1		# mul_add_c2(a[3],b[1],c2,c3,c1);
+	$SLL	$t_2,1
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	sltu	$at,$c_2,$t_2
+	$ADDU	$c_3,$at
+	$ST	$c_1,3*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_1,$t_2,$zero
+	$SLL	$t_2,1
+	$MULTU	$a_2,$a_2		# mul_add_c(a[2],b[2],c2,c3,c1);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_2,$t_1
+	sltu	$at,$c_2,$t_1
+	 $MULTU	$a_2,$a_3		# mul_add_c2(a[2],b[3],c3,c1,c2);
+	$ADDU	$t_2,$at
+	$ADDU	$c_3,$t_2
+	sltu	$at,$c_3,$t_2
+	$ADDU	$c_1,$at
+	$ST	$c_2,4*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	slt	$c_2,$t_2,$zero
+	$SLL	$t_2,1
+	 $MULTU	$a_3,$a_3		# mul_add_c(a[3],b[3],c1,c2,c3);
+	slt	$a2,$t_1,$zero
+	$ADDU	$t_2,$a2
+	$SLL	$t_1,1
+	$ADDU	$c_3,$t_1
+	sltu	$at,$c_3,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_1,$t_2
+	sltu	$at,$c_1,$t_2
+	$ADDU	$c_2,$at
+	$ST	$c_3,5*$BNSZ($a0)
+
+	mflo	$t_1
+	mfhi	$t_2
+	$ADDU	$c_1,$t_1
+	sltu	$at,$c_1,$t_1
+	$ADDU	$t_2,$at
+	$ADDU	$c_2,$t_2
+	$ST	$c_1,6*$BNSZ($a0)
+	$ST	$c_2,7*$BNSZ($a0)
+
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$t3,4*$SZREG($sp)
+	$REG_L	$t2,3*$SZREG($sp)
+	$REG_L	$t1,2*$SZREG($sp)
+	$REG_L	$t0,1*$SZREG($sp)
+	$REG_L	$gp,0*$SZREG($sp)
+	$PTR_ADD $sp,6*$SZREG
+___
+$code.=<<___;
+	jr	$ra
+	nop
+.end	bn_sqr_comba4
+___
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/mips3-mont.pl b/crypto/bn/asm/mips3-mont.pl
new file mode 100755
index 000000000000..8f9156e02af3
--- /dev/null
+++ b/crypto/bn/asm/mips3-mont.pl
@@ -0,0 +1,327 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# This module doesn't present direct interest for OpenSSL, because it
+# doesn't provide better performance for longer keys. While 512-bit
+# RSA private key operations are 40% faster, 1024-bit ones are hardly
+# faster at all, while longer key operations are slower by up to 20%.
+# It might be of interest to embedded system developers though, as
+# it's smaller than 1KB, yet offers ~3x improvement over compiler
+# generated code.
+#
+# The module targets N32 and N64 MIPS ABIs and currently is a bit
+# IRIX-centric, i.e. is likely to require adaptation for other OSes.
+
+# int bn_mul_mont(
+$rp="a0";	# BN_ULONG *rp,
+$ap="a1";	# const BN_ULONG *ap,
+$bp="a2";	# const BN_ULONG *bp,
+$np="a3";	# const BN_ULONG *np,
+$n0="a4";	# const BN_ULONG *n0,
+$num="a5";	# int num);
+
+$lo0="a6";
+$hi0="a7";
+$lo1="v0";
+$hi1="v1";
+$aj="t0";
+$bi="t1";
+$nj="t2";
+$tp="t3";
+$alo="s0";
+$ahi="s1";
+$nlo="s2";
+$nhi="s3";
+$tj="s4";
+$i="s5";
+$j="s6";
+$fp="t8";
+$m1="t9";
+
+$FRAME=8*(2+8);
+
+$code=<<___;
+#include <asm.h>
+#include <regdef.h>
+
+.text
+
+.set	noat
+.set	reorder
+
+.align	5
+.globl	bn_mul_mont
+.ent	bn_mul_mont
+bn_mul_mont:
+	.set	noreorder
+	PTR_SUB	sp,64
+	move	$fp,sp
+	.frame	$fp,64,ra
+	slt	AT,$num,4
+	li	v0,0
+	beqzl	AT,.Lproceed
+	nop
+	jr	ra
+	PTR_ADD	sp,$fp,64
+	.set	reorder
+.align	5
+.Lproceed:
+	ld	$n0,0($n0)
+	ld	$bi,0($bp)	# bp[0]
+	ld	$aj,0($ap)	# ap[0]
+	ld	$nj,0($np)	# np[0]
+	PTR_SUB	sp,16		# place for two extra words
+	sll	$num,3
+	li	AT,-4096
+	PTR_SUB	sp,$num
+	and	sp,AT
+
+	sd	s0,0($fp)
+	sd	s1,8($fp)
+	sd	s2,16($fp)
+	sd	s3,24($fp)
+	sd	s4,32($fp)
+	sd	s5,40($fp)
+	sd	s6,48($fp)
+	sd	s7,56($fp)
+
+	dmultu	$aj,$bi
+	ld	$alo,8($ap)
+	ld	$nlo,8($np)
+	mflo	$lo0
+	mfhi	$hi0
+	dmultu	$lo0,$n0
+	mflo	$m1
+
+	dmultu	$alo,$bi
+	mflo	$alo
+	mfhi	$ahi
+
+	dmultu	$nj,$m1
+	mflo	$lo1
+	mfhi	$hi1
+	dmultu	$nlo,$m1
+	daddu	$lo1,$lo0
+	sltu	AT,$lo1,$lo0
+	daddu	$hi1,AT
+	mflo	$nlo
+	mfhi	$nhi
+
+	move	$tp,sp
+	li	$j,16
+.align	4
+.L1st:
+	.set	noreorder
+	PTR_ADD	$aj,$ap,$j
+	ld	$aj,($aj)
+	PTR_ADD	$nj,$np,$j
+	ld	$nj,($nj)
+
+	dmultu	$aj,$bi
+	daddu	$lo0,$alo,$hi0
+	daddu	$lo1,$nlo,$hi1
+	sltu	AT,$lo0,$hi0
+	sltu	s7,$lo1,$hi1
+	daddu	$hi0,$ahi,AT
+	daddu	$hi1,$nhi,s7
+	mflo	$alo
+	mfhi	$ahi
+
+	daddu	$lo1,$lo0
+	sltu	AT,$lo1,$lo0
+	dmultu	$nj,$m1
+	daddu	$hi1,AT
+	addu	$j,8
+	sd	$lo1,($tp)
+	sltu	s7,$j,$num
+	mflo	$nlo
+	mfhi	$nhi
+
+	bnez	s7,.L1st
+	PTR_ADD	$tp,8
+	.set	reorder
+
+	daddu	$lo0,$alo,$hi0
+	sltu	AT,$lo0,$hi0
+	daddu	$hi0,$ahi,AT
+
+	daddu	$lo1,$nlo,$hi1
+	sltu	s7,$lo1,$hi1
+	daddu	$hi1,$nhi,s7
+	daddu	$lo1,$lo0
+	sltu	AT,$lo1,$lo0
+	daddu	$hi1,AT
+
+	sd	$lo1,($tp)
+
+	daddu	$hi1,$hi0
+	sltu	AT,$hi1,$hi0
+	sd	$hi1,8($tp)
+	sd	AT,16($tp)
+
+	li	$i,8
+.align	4
+.Louter:
+	PTR_ADD	$bi,$bp,$i
+	ld	$bi,($bi)
+	ld	$aj,($ap)
+	ld	$alo,8($ap)
+	ld	$tj,(sp)
+
+	dmultu	$aj,$bi
+	ld	$nj,($np)
+	ld	$nlo,8($np)
+	mflo	$lo0
+	mfhi	$hi0
+	daddu	$lo0,$tj
+	dmultu	$lo0,$n0
+	sltu	AT,$lo0,$tj
+	daddu	$hi0,AT
+	mflo	$m1
+
+	dmultu	$alo,$bi
+	mflo	$alo
+	mfhi	$ahi
+
+	dmultu	$nj,$m1
+	mflo	$lo1
+	mfhi	$hi1
+
+	dmultu	$nlo,$m1
+	daddu	$lo1,$lo0
+	sltu	AT,$lo1,$lo0
+	daddu	$hi1,AT
+	mflo	$nlo
+	mfhi	$nhi
+
+	move	$tp,sp
+	li	$j,16
+	ld	$tj,8($tp)
+.align	4
+.Linner:
+	.set	noreorder
+	PTR_ADD	$aj,$ap,$j
+	ld	$aj,($aj)
+	PTR_ADD	$nj,$np,$j
+	ld	$nj,($nj)
+
+	dmultu	$aj,$bi
+	daddu	$lo0,$alo,$hi0
+	daddu	$lo1,$nlo,$hi1
+	sltu	AT,$lo0,$hi0
+	sltu	s7,$lo1,$hi1
+	daddu	$hi0,$ahi,AT
+	daddu	$hi1,$nhi,s7
+	mflo	$alo
+	mfhi	$ahi
+
+	daddu	$lo0,$tj
+	addu	$j,8
+	dmultu	$nj,$m1
+	sltu	AT,$lo0,$tj
+	daddu	$lo1,$lo0
+	daddu	$hi0,AT
+	sltu	s7,$lo1,$lo0
+	ld	$tj,16($tp)
+	daddu	$hi1,s7
+	sltu	AT,$j,$num
+	mflo	$nlo
+	mfhi	$nhi
+	sd	$lo1,($tp)
+	bnez	AT,.Linner
+	PTR_ADD	$tp,8
+	.set	reorder
+
+	daddu	$lo0,$alo,$hi0
+	sltu	AT,$lo0,$hi0
+	daddu	$hi0,$ahi,AT
+	daddu	$lo0,$tj
+	sltu	s7,$lo0,$tj
+	daddu	$hi0,s7
+
+	ld	$tj,16($tp)
+	daddu	$lo1,$nlo,$hi1
+	sltu	AT,$lo1,$hi1
+	daddu	$hi1,$nhi,AT
+	daddu	$lo1,$lo0
+	sltu	s7,$lo1,$lo0
+	daddu	$hi1,s7
+	sd	$lo1,($tp)
+
+	daddu	$lo1,$hi1,$hi0
+	sltu	$hi1,$lo1,$hi0
+	daddu	$lo1,$tj
+	sltu	AT,$lo1,$tj
+	daddu	$hi1,AT
+	sd	$lo1,8($tp)
+	sd	$hi1,16($tp)
+
+	addu	$i,8
+	sltu	s7,$i,$num
+	bnez	s7,.Louter
+
+	.set	noreorder
+	PTR_ADD	$tj,sp,$num	# &tp[num]
+	move	$tp,sp
+	move	$ap,sp
+	li	$hi0,0		# clear borrow bit
+
+.align	4
+.Lsub:	ld	$lo0,($tp)
+	ld	$lo1,($np)
+	PTR_ADD	$tp,8
+	PTR_ADD	$np,8
+	dsubu	$lo1,$lo0,$lo1	# tp[i]-np[i]
+	sgtu	AT,$lo1,$lo0
+	dsubu	$lo0,$lo1,$hi0
+	sgtu	$hi0,$lo0,$lo1
+	sd	$lo0,($rp)
+	or	$hi0,AT
+	sltu	AT,$tp,$tj
+	bnez	AT,.Lsub
+	PTR_ADD	$rp,8
+
+	dsubu	$hi0,$hi1,$hi0	# handle upmost overflow bit
+	move	$tp,sp
+	PTR_SUB	$rp,$num	# restore rp
+	not	$hi1,$hi0
+
+	and	$ap,$hi0,sp
+	and	$bp,$hi1,$rp
+	or	$ap,$ap,$bp	# ap=borrow?tp:rp
+
+.align	4
+.Lcopy:	ld	$aj,($ap)
+	PTR_ADD	$ap,8
+	PTR_ADD	$tp,8
+	sd	zero,-8($tp)
+	sltu	AT,$tp,$tj
+	sd	$aj,($rp)
+	bnez	AT,.Lcopy
+	PTR_ADD	$rp,8
+
+	ld	s0,0($fp)
+	ld	s1,8($fp)
+	ld	s2,16($fp)
+	ld	s3,24($fp)
+	ld	s4,32($fp)
+	ld	s5,40($fp)
+	ld	s6,48($fp)
+	ld	s7,56($fp)
+	li	v0,1
+	jr	ra
+	PTR_ADD	sp,$fp,64
+	.set	reorder
+END(bn_mul_mont)
+.rdata
+.asciiz	"Montgomery Multiplication for MIPS III/IV, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/modexp512-x86_64.pl b/crypto/bn/asm/modexp512-x86_64.pl
new file mode 100755
index 000000000000..54aeb01921e3
--- /dev/null
+++ b/crypto/bn/asm/modexp512-x86_64.pl
@@ -0,0 +1,1496 @@
+#!/usr/bin/env perl
+#
+# Copyright (c) 2010-2011 Intel Corp.
+#   Author: Vinodh.Gopal@intel.com
+#           Jim Guilford
+#           Erdinc.Ozturk@intel.com
+#           Maxim.Perminov@intel.com
+#
+# More information about algorithm used can be found at:
+#   http://www.cse.buffalo.edu/srds2009/escs2009_submission_Gopal.pdf
+#
+# ====================================================================
+# Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions
+# are met:
+#
+# 1. Redistributions of source code must retain the above copyright
+#    notice, this list of conditions and the following disclaimer.
+#
+# 2. Redistributions in binary form must reproduce the above copyright
+#    notice, this list of conditions and the following disclaimer in
+#    the documentation and/or other materials provided with the
+#    distribution.
+#
+# 3. All advertising materials mentioning features or use of this
+#    software must display the following acknowledgment:
+#    "This product includes software developed by the OpenSSL Project
+#    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+#
+# 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+#    endorse or promote products derived from this software without
+#    prior written permission. For written permission, please contact
+#    licensing@OpenSSL.org.
+#
+# 5. Products derived from this software may not be called "OpenSSL"
+#    nor may "OpenSSL" appear in their names without prior written
+#    permission of the OpenSSL Project.
+#
+# 6. Redistributions of any form whatsoever must retain the following
+#    acknowledgment:
+#    "This product includes software developed by the OpenSSL Project
+#    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+#
+# THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+# EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+# PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+# ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+# NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+# HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+# STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+# OF THE POSSIBILITY OF SUCH DAMAGE.
+# ====================================================================
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+my $win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+use strict;
+my $code=".text\n\n";
+my $m=0;
+
+#
+# Define x512 macros
+#
+
+#MULSTEP_512_ADD	MACRO	x7, x6, x5, x4, x3, x2, x1, x0, dst, src1, src2, add_src, tmp1, tmp2
+#
+# uses rax, rdx, and args
+sub MULSTEP_512_ADD
+{
+ my ($x, $DST, $SRC2, $ASRC, $OP, $TMP)=@_;
+ my @X=@$x;	# make a copy
+$code.=<<___;
+	 mov	(+8*0)($SRC2), %rax
+	 mul	$OP			# rdx:rax = %OP * [0]
+	 mov	($ASRC), $X[0]
+	 add	%rax, $X[0]
+	 adc	\$0, %rdx
+	 mov	$X[0], $DST
+___
+for(my $i=1;$i<8;$i++) {
+$code.=<<___;
+	 mov	%rdx, $TMP
+
+	 mov	(+8*$i)($SRC2), %rax
+	 mul	$OP			# rdx:rax = %OP * [$i]
+	 mov	(+8*$i)($ASRC), $X[$i]
+	 add	%rax, $X[$i]
+	 adc	\$0, %rdx
+	 add	$TMP, $X[$i]
+	 adc	\$0, %rdx
+___
+}
+$code.=<<___;
+	 mov	%rdx, $X[0]
+___
+}
+
+#MULSTEP_512	MACRO	x7, x6, x5, x4, x3, x2, x1, x0, dst, src2, src1_val, tmp
+#
+# uses rax, rdx, and args
+sub MULSTEP_512
+{
+ my ($x, $DST, $SRC2, $OP, $TMP)=@_;
+ my @X=@$x;	# make a copy
+$code.=<<___;
+	 mov	(+8*0)($SRC2), %rax
+	 mul	$OP			# rdx:rax = %OP * [0]
+	 add	%rax, $X[0]
+	 adc	\$0, %rdx
+	 mov	$X[0], $DST
+___
+for(my $i=1;$i<8;$i++) {
+$code.=<<___;
+	 mov	%rdx, $TMP
+
+	 mov	(+8*$i)($SRC2), %rax
+	 mul	$OP			# rdx:rax = %OP * [$i]
+	 add	%rax, $X[$i]
+	 adc	\$0, %rdx
+	 add	$TMP, $X[$i]
+	 adc	\$0, %rdx
+___
+}
+$code.=<<___;
+	 mov	%rdx, $X[0]
+___
+}
+
+#
+# Swizzle Macros
+#
+
+# macro to copy data from flat space to swizzled table
+#MACRO swizzle	pDst, pSrc, tmp1, tmp2
+# pDst and pSrc are modified
+sub swizzle
+{
+ my ($pDst, $pSrc, $cnt, $d0)=@_;
+$code.=<<___;
+	 mov	\$8, $cnt
+loop_$m:
+	 mov	($pSrc), $d0
+	 mov	$d0#w, ($pDst)
+	 shr	\$16, $d0
+	 mov	$d0#w, (+64*1)($pDst)
+	 shr	\$16, $d0
+	 mov	$d0#w, (+64*2)($pDst)
+	 shr	\$16, $d0
+	 mov	$d0#w, (+64*3)($pDst)
+	 lea	8($pSrc), $pSrc
+	 lea	64*4($pDst), $pDst
+	 dec	$cnt
+	 jnz	loop_$m
+___
+
+ $m++;
+}
+
+# macro to copy data from swizzled table to  flat space
+#MACRO unswizzle	pDst, pSrc, tmp*3
+sub unswizzle
+{
+ my ($pDst, $pSrc, $cnt, $d0, $d1)=@_;
+$code.=<<___;
+	 mov	\$4, $cnt
+loop_$m:
+	 movzxw	(+64*3+256*0)($pSrc), $d0
+	 movzxw	(+64*3+256*1)($pSrc), $d1
+	 shl	\$16, $d0
+	 shl	\$16, $d1
+	 mov	(+64*2+256*0)($pSrc), $d0#w
+	 mov	(+64*2+256*1)($pSrc), $d1#w
+	 shl	\$16, $d0
+	 shl	\$16, $d1
+	 mov	(+64*1+256*0)($pSrc), $d0#w
+	 mov	(+64*1+256*1)($pSrc), $d1#w
+	 shl	\$16, $d0
+	 shl	\$16, $d1
+	 mov	(+64*0+256*0)($pSrc), $d0#w
+	 mov	(+64*0+256*1)($pSrc), $d1#w
+	 mov	$d0, (+8*0)($pDst)
+	 mov	$d1, (+8*1)($pDst)
+	 lea	256*2($pSrc), $pSrc
+	 lea	8*2($pDst), $pDst
+	 sub	\$1, $cnt
+	 jnz	loop_$m
+___
+
+ $m++;
+}
+
+#
+# Data Structures
+#
+
+# Reduce Data
+#
+#
+# Offset  Value
+# 0C0     Carries
+# 0B8     X2[10]
+# 0B0     X2[9]
+# 0A8     X2[8]
+# 0A0     X2[7]
+# 098     X2[6]
+# 090     X2[5]
+# 088     X2[4]
+# 080     X2[3]
+# 078     X2[2]
+# 070     X2[1]
+# 068     X2[0]
+# 060     X1[12]  P[10]
+# 058     X1[11]  P[9]  Z[8]
+# 050     X1[10]  P[8]  Z[7]
+# 048     X1[9]   P[7]  Z[6]
+# 040     X1[8]   P[6]  Z[5]
+# 038     X1[7]   P[5]  Z[4]
+# 030     X1[6]   P[4]  Z[3]
+# 028     X1[5]   P[3]  Z[2]
+# 020     X1[4]   P[2]  Z[1]
+# 018     X1[3]   P[1]  Z[0]
+# 010     X1[2]   P[0]  Y[2]
+# 008     X1[1]   Q[1]  Y[1]
+# 000     X1[0]   Q[0]  Y[0]
+
+my $X1_offset           =  0;			# 13 qwords
+my $X2_offset           =  $X1_offset + 13*8;			# 11 qwords
+my $Carries_offset      =  $X2_offset + 11*8;			# 1 qword
+my $Q_offset            =  0;			# 2 qwords
+my $P_offset            =  $Q_offset + 2*8;			# 11 qwords
+my $Y_offset            =  0;			# 3 qwords
+my $Z_offset            =  $Y_offset + 3*8;			# 9 qwords
+
+my $Red_Data_Size       =  $Carries_offset + 1*8;			# (25 qwords)
+
+#
+# Stack Frame
+#
+#
+# offset	value
+# ...		<old stack contents>
+# ...
+# 280		Garray
+
+# 278		tmp16[15]
+# ...		...
+# 200		tmp16[0]
+
+# 1F8		tmp[7]
+# ...		...
+# 1C0		tmp[0]
+
+# 1B8		GT[7]
+# ...		...
+# 180		GT[0]
+
+# 178		Reduce Data
+# ...		...
+# 0B8		Reduce Data
+# 0B0		reserved
+# 0A8		reserved
+# 0A0		reserved
+# 098		reserved
+# 090		reserved
+# 088		reduce result addr
+# 080		exp[8]
+
+# ...
+# 048		exp[1]
+# 040		exp[0]
+
+# 038		reserved
+# 030		loop_idx
+# 028		pg
+# 020		i
+# 018		pData	; arg 4
+# 010		pG	; arg 2
+# 008		pResult	; arg 1
+# 000		rsp	; stack pointer before subtract
+
+my $rsp_offset          =  0;
+my $pResult_offset      =  8*1 + $rsp_offset;
+my $pG_offset           =  8*1 + $pResult_offset;
+my $pData_offset        =  8*1 + $pG_offset;
+my $i_offset            =  8*1 + $pData_offset;
+my $pg_offset           =  8*1 + $i_offset;
+my $loop_idx_offset     =  8*1 + $pg_offset;
+my $reserved1_offset    =  8*1 + $loop_idx_offset;
+my $exp_offset          =  8*1 + $reserved1_offset;
+my $red_result_addr_offset=  8*9 + $exp_offset;
+my $reserved2_offset    =  8*1 + $red_result_addr_offset;
+my $Reduce_Data_offset  =  8*5 + $reserved2_offset;
+my $GT_offset           =  $Red_Data_Size + $Reduce_Data_offset;
+my $tmp_offset          =  8*8 + $GT_offset;
+my $tmp16_offset        =  8*8 + $tmp_offset;
+my $garray_offset       =  8*16 + $tmp16_offset;
+my $mem_size            =  8*8*32 + $garray_offset;
+
+#
+# Offsets within Reduce Data
+#
+#
+#	struct MODF_2FOLD_MONT_512_C1_DATA {
+#	UINT64 t[8][8];
+#	UINT64 m[8];
+#	UINT64 m1[8]; /* 2^768 % m */
+#	UINT64 m2[8]; /* 2^640 % m */
+#	UINT64 k1[2]; /* (- 1/m) % 2^128 */
+#	};
+
+my $T                   =  0;
+my $M                   =  512;			# = 8 * 8 * 8
+my $M1                  =  576;			# = 8 * 8 * 9 /* += 8 * 8 */
+my $M2                  =  640;			# = 8 * 8 * 10 /* += 8 * 8 */
+my $K1                  =  704;			# = 8 * 8 * 11 /* += 8 * 8 */
+
+#
+#   FUNCTIONS
+#
+
+{{{
+#
+# MULADD_128x512 : Function to multiply 128-bits (2 qwords) by 512-bits (8 qwords)
+#                       and add 512-bits (8 qwords)
+#                       to get 640 bits (10 qwords)
+# Input: 128-bit mul source: [rdi+8*1], rbp
+#        512-bit mul source: [rsi+8*n]
+#        512-bit add source: r15, r14, ..., r9, r8
+# Output: r9, r8, r15, r14, r13, r12, r11, r10, [rcx+8*1], [rcx+8*0]
+# Clobbers all regs except: rcx, rsi, rdi
+$code.=<<___;
+.type	MULADD_128x512,\@abi-omnipotent
+.align	16
+MULADD_128x512:
+___
+	&MULSTEP_512([map("%r$_",(8..15))], "(+8*0)(%rcx)", "%rsi", "%rbp", "%rbx");
+$code.=<<___;
+	 mov	(+8*1)(%rdi), %rbp
+___
+	&MULSTEP_512([map("%r$_",(9..15,8))], "(+8*1)(%rcx)", "%rsi", "%rbp", "%rbx");
+$code.=<<___;
+	 ret
+.size	MULADD_128x512,.-MULADD_128x512
+___
+}}}
+
+{{{
+#MULADD_256x512	MACRO	pDst, pA, pB, OP, TMP, X7, X6, X5, X4, X3, X2, X1, X0
+#
+# Inputs: pDst: Destination  (768 bits, 12 qwords)
+#         pA:   Multiplicand (1024 bits, 16 qwords)
+#         pB:   Multiplicand (512 bits, 8 qwords)
+# Dst = Ah * B + Al
+# where Ah is (in qwords) A[15:12] (256 bits) and Al is A[7:0] (512 bits)
+# Results in X3 X2 X1 X0 X7 X6 X5 X4 Dst[3:0]
+# Uses registers: arguments, RAX, RDX
+sub MULADD_256x512
+{
+ my ($pDst, $pA, $pB, $OP, $TMP, $X)=@_;
+$code.=<<___;
+	mov	(+8*12)($pA), $OP
+___
+	&MULSTEP_512_ADD($X, "(+8*0)($pDst)", $pB, $pA, $OP, $TMP);
+	push(@$X,shift(@$X));
+
+$code.=<<___;
+	 mov	(+8*13)($pA), $OP
+___
+	&MULSTEP_512($X, "(+8*1)($pDst)", $pB, $OP, $TMP);
+	push(@$X,shift(@$X));
+
+$code.=<<___;
+	 mov	(+8*14)($pA), $OP
+___
+	&MULSTEP_512($X, "(+8*2)($pDst)", $pB, $OP, $TMP);
+	push(@$X,shift(@$X));
+
+$code.=<<___;
+	 mov	(+8*15)($pA), $OP
+___
+	&MULSTEP_512($X, "(+8*3)($pDst)", $pB, $OP, $TMP);
+	push(@$X,shift(@$X));
+}
+
+#
+# mont_reduce(UINT64 *x,  /* 1024 bits, 16 qwords */
+#	       UINT64 *m,  /*  512 bits,  8 qwords */
+#	       MODF_2FOLD_MONT_512_C1_DATA *data,
+#             UINT64 *r)  /*  512 bits,  8 qwords */
+# Input:  x (number to be reduced): tmp16 (Implicit)
+#         m (modulus):              [pM]  (Implicit)
+#         data (reduce data):       [pData] (Implicit)
+# Output: r (result):		     Address in [red_res_addr]
+#         result also in: r9, r8, r15, r14, r13, r12, r11, r10
+
+my @X=map("%r$_",(8..15));
+
+$code.=<<___;
+.type	mont_reduce,\@abi-omnipotent
+.align	16
+mont_reduce:
+___
+
+my $STACK_DEPTH         =  8;
+	#
+	# X1 = Xh * M1 + Xl
+$code.=<<___;
+	 lea	(+$Reduce_Data_offset+$X1_offset+$STACK_DEPTH)(%rsp), %rdi			# pX1 (Dst) 769 bits, 13 qwords
+	 mov	(+$pData_offset+$STACK_DEPTH)(%rsp), %rsi			# pM1 (Bsrc) 512 bits, 8 qwords
+	 add	\$$M1, %rsi
+	 lea	(+$tmp16_offset+$STACK_DEPTH)(%rsp), %rcx			# X (Asrc) 1024 bits, 16 qwords
+
+___
+
+	&MULADD_256x512("%rdi", "%rcx", "%rsi", "%rbp", "%rbx", \@X);	# rotates @X 4 times
+	# results in r11, r10, r9, r8, r15, r14, r13, r12, X1[3:0]
+
+$code.=<<___;
+	 xor	%rax, %rax
+	# X1 += xl
+	 add	(+8*8)(%rcx), $X[4]
+	 adc	(+8*9)(%rcx), $X[5]
+	 adc	(+8*10)(%rcx), $X[6]
+	 adc	(+8*11)(%rcx), $X[7]
+	 adc	\$0, %rax
+	# X1 is now rax, r11-r8, r15-r12, tmp16[3:0]
+
+	#
+	# check for carry ;; carry stored in rax
+	 mov	$X[4], (+8*8)(%rdi)			# rdi points to X1
+	 mov	$X[5], (+8*9)(%rdi)
+	 mov	$X[6], %rbp
+	 mov	$X[7], (+8*11)(%rdi)
+
+	 mov	%rax, (+$Reduce_Data_offset+$Carries_offset+$STACK_DEPTH)(%rsp)
+
+	 mov	(+8*0)(%rdi), $X[4]
+	 mov	(+8*1)(%rdi), $X[5]
+	 mov	(+8*2)(%rdi), $X[6]
+	 mov	(+8*3)(%rdi), $X[7]
+
+	# X1 is now stored in: X1[11], rbp, X1[9:8], r15-r8
+	# rdi -> X1
+	# rsi -> M1
+
+	#
+	# X2 = Xh * M2 + Xl
+	# do first part (X2 = Xh * M2)
+	 add	\$8*10, %rdi			# rdi -> pXh ; 128 bits, 2 qwords
+				#        Xh is actually { [rdi+8*1], rbp }
+	 add	\$`$M2-$M1`, %rsi			# rsi -> M2
+	 lea	(+$Reduce_Data_offset+$X2_offset+$STACK_DEPTH)(%rsp), %rcx			# rcx -> pX2 ; 641 bits, 11 qwords
+___
+	unshift(@X,pop(@X));	unshift(@X,pop(@X));
+$code.=<<___;
+
+	 call	MULADD_128x512			# args in rcx, rdi / rbp, rsi, r15-r8
+	# result in r9, r8, r15, r14, r13, r12, r11, r10, X2[1:0]
+	 mov	(+$Reduce_Data_offset+$Carries_offset+$STACK_DEPTH)(%rsp), %rax
+
+	# X2 += Xl
+	 add	(+8*8-8*10)(%rdi), $X[6]		# (-8*10) is to adjust rdi -> Xh to Xl
+	 adc	(+8*9-8*10)(%rdi), $X[7]
+	 mov	$X[6], (+8*8)(%rcx)
+	 mov	$X[7], (+8*9)(%rcx)
+
+	 adc	%rax, %rax
+	 mov	%rax, (+$Reduce_Data_offset+$Carries_offset+$STACK_DEPTH)(%rsp)
+
+	 lea	(+$Reduce_Data_offset+$Q_offset+$STACK_DEPTH)(%rsp), %rdi			# rdi -> pQ ; 128 bits, 2 qwords
+	 add	\$`$K1-$M2`, %rsi			# rsi -> pK1 ; 128 bits, 2 qwords
+
+	# MUL_128x128t128	rdi, rcx, rsi	; Q = X2 * K1 (bottom half)
+	# B1:B0 = rsi[1:0] = K1[1:0]
+	# A1:A0 = rcx[1:0] = X2[1:0]
+	# Result = rdi[1],rbp = Q[1],rbp
+	 mov	(%rsi), %r8			# B0
+	 mov	(+8*1)(%rsi), %rbx			# B1
+
+	 mov	(%rcx), %rax			# A0
+	 mul	%r8			# B0
+	 mov	%rax, %rbp
+	 mov	%rdx, %r9
+
+	 mov	(+8*1)(%rcx), %rax			# A1
+	 mul	%r8			# B0
+	 add	%rax, %r9
+
+	 mov	(%rcx), %rax			# A0
+	 mul	%rbx			# B1
+	 add	%rax, %r9
+
+	 mov	%r9, (+8*1)(%rdi)
+	# end MUL_128x128t128
+
+	 sub	\$`$K1-$M`, %rsi
+
+	 mov	(%rcx), $X[6]
+	 mov	(+8*1)(%rcx), $X[7]			# r9:r8 = X2[1:0]
+
+	 call	MULADD_128x512			# args in rcx, rdi / rbp, rsi, r15-r8
+	# result in r9, r8, r15, r14, r13, r12, r11, r10, X2[1:0]
+
+	# load first half of m to rdx, rdi, rbx, rax
+	# moved this here for efficiency
+	 mov	(+8*0)(%rsi), %rax
+	 mov	(+8*1)(%rsi), %rbx
+	 mov	(+8*2)(%rsi), %rdi
+	 mov	(+8*3)(%rsi), %rdx
+
+	# continue with reduction
+	 mov	(+$Reduce_Data_offset+$Carries_offset+$STACK_DEPTH)(%rsp), %rbp
+
+	 add	(+8*8)(%rcx), $X[6]
+	 adc	(+8*9)(%rcx), $X[7]
+
+	#accumulate the final carry to rbp
+	 adc	%rbp, %rbp
+
+	# Add in overflow corrections: R = (X2>>128) += T[overflow]
+	# R = {r9, r8, r15, r14, ..., r10}
+	 shl	\$3, %rbp
+	 mov	(+$pData_offset+$STACK_DEPTH)(%rsp), %rcx			# rsi -> Data (and points to T)
+	 add	%rcx, %rbp			# pT ; 512 bits, 8 qwords, spread out
+
+	# rsi will be used to generate a mask after the addition
+	 xor	%rsi, %rsi
+
+	 add	(+8*8*0)(%rbp), $X[0]
+	 adc	(+8*8*1)(%rbp), $X[1]
+	 adc	(+8*8*2)(%rbp), $X[2]
+	 adc	(+8*8*3)(%rbp), $X[3]
+	 adc	(+8*8*4)(%rbp), $X[4]
+	 adc	(+8*8*5)(%rbp), $X[5]
+	 adc	(+8*8*6)(%rbp), $X[6]
+	 adc	(+8*8*7)(%rbp), $X[7]
+
+	# if there is a carry:	rsi = 0xFFFFFFFFFFFFFFFF
+	# if carry is clear:	rsi = 0x0000000000000000
+	 sbb	\$0, %rsi
+
+	# if carry is clear, subtract 0. Otherwise, subtract 256 bits of m
+	 and	%rsi, %rax
+	 and	%rsi, %rbx
+	 and	%rsi, %rdi
+	 and	%rsi, %rdx
+
+	 mov	\$1, %rbp
+	 sub	%rax, $X[0]
+	 sbb	%rbx, $X[1]
+	 sbb	%rdi, $X[2]
+	 sbb	%rdx, $X[3]
+
+	# if there is a borrow:		rbp = 0
+	# if there is no borrow:	rbp = 1
+	# this is used to save the borrows in between the first half and the 2nd half of the subtraction of m
+	 sbb	\$0, %rbp
+
+	#load second half of m to rdx, rdi, rbx, rax
+
+	 add	\$$M, %rcx
+	 mov	(+8*4)(%rcx), %rax
+	 mov	(+8*5)(%rcx), %rbx
+	 mov	(+8*6)(%rcx), %rdi
+	 mov	(+8*7)(%rcx), %rdx
+
+	# use the rsi mask as before
+	# if carry is clear, subtract 0. Otherwise, subtract 256 bits of m
+	 and	%rsi, %rax
+	 and	%rsi, %rbx
+	 and	%rsi, %rdi
+	 and	%rsi, %rdx
+
+	# if rbp = 0, there was a borrow before, it is moved to the carry flag
+	# if rbp = 1, there was not a borrow before, carry flag is cleared
+	 sub	\$1, %rbp
+
+	 sbb	%rax, $X[4]
+	 sbb	%rbx, $X[5]
+	 sbb	%rdi, $X[6]
+	 sbb	%rdx, $X[7]
+
+	# write R back to memory
+
+	 mov	(+$red_result_addr_offset+$STACK_DEPTH)(%rsp), %rsi
+	 mov	$X[0], (+8*0)(%rsi)
+	 mov	$X[1], (+8*1)(%rsi)
+	 mov	$X[2], (+8*2)(%rsi)
+	 mov	$X[3], (+8*3)(%rsi)
+	 mov	$X[4], (+8*4)(%rsi)
+	 mov	$X[5], (+8*5)(%rsi)
+	 mov	$X[6], (+8*6)(%rsi)
+	 mov	$X[7], (+8*7)(%rsi)
+
+	 ret
+.size	mont_reduce,.-mont_reduce
+___
+}}}
+
+{{{
+#MUL_512x512	MACRO	pDst, pA, pB, x7, x6, x5, x4, x3, x2, x1, x0, tmp*2
+#
+# Inputs: pDst: Destination  (1024 bits, 16 qwords)
+#         pA:   Multiplicand (512 bits, 8 qwords)
+#         pB:   Multiplicand (512 bits, 8 qwords)
+# Uses registers rax, rdx, args
+#   B operand in [pB] and also in x7...x0
+sub MUL_512x512
+{
+ my ($pDst, $pA, $pB, $x, $OP, $TMP, $pDst_o)=@_;
+ my ($pDst,  $pDst_o) = ($pDst =~ m/([^+]*)\+?(.*)?/);
+ my @X=@$x;	# make a copy
+
+$code.=<<___;
+	 mov	(+8*0)($pA), $OP
+
+	 mov	$X[0], %rax
+	 mul	$OP			# rdx:rax = %OP * [0]
+	 mov	%rax, (+$pDst_o+8*0)($pDst)
+	 mov	%rdx, $X[0]
+___
+for(my $i=1;$i<8;$i++) {
+$code.=<<___;
+	 mov	$X[$i], %rax
+	 mul	$OP			# rdx:rax = %OP * [$i]
+	 add	%rax, $X[$i-1]
+	 adc	\$0, %rdx
+	 mov	%rdx, $X[$i]
+___
+}
+
+for(my $i=1;$i<8;$i++) {
+$code.=<<___;
+	 mov	(+8*$i)($pA), $OP
+___
+
+	&MULSTEP_512(\@X, "(+$pDst_o+8*$i)($pDst)", $pB, $OP, $TMP);
+	push(@X,shift(@X));
+}
+
+$code.=<<___;
+	 mov	$X[0], (+$pDst_o+8*8)($pDst)
+	 mov	$X[1], (+$pDst_o+8*9)($pDst)
+	 mov	$X[2], (+$pDst_o+8*10)($pDst)
+	 mov	$X[3], (+$pDst_o+8*11)($pDst)
+	 mov	$X[4], (+$pDst_o+8*12)($pDst)
+	 mov	$X[5], (+$pDst_o+8*13)($pDst)
+	 mov	$X[6], (+$pDst_o+8*14)($pDst)
+	 mov	$X[7], (+$pDst_o+8*15)($pDst)
+___
+}
+
+#
+# mont_mul_a3b : subroutine to compute (Src1 * Src2) % M (all 512-bits)
+# Input:  src1: Address of source 1: rdi
+#         src2: Address of source 2: rsi
+# Output: dst:  Address of destination: [red_res_addr]
+#    src2 and result also in: r9, r8, r15, r14, r13, r12, r11, r10
+# Temp:   Clobbers [tmp16], all registers
+$code.=<<___;
+.type	mont_mul_a3b,\@abi-omnipotent
+.align	16
+mont_mul_a3b:
+	#
+	# multiply tmp = src1 * src2
+	# For multiply: dst = rcx, src1 = rdi, src2 = rsi
+	# stack depth is extra 8 from call
+___
+	&MUL_512x512("%rsp+$tmp16_offset+8", "%rdi", "%rsi", [map("%r$_",(10..15,8..9))], "%rbp", "%rbx");
+$code.=<<___;
+	#
+	# Dst = tmp % m
+	# Call reduce(tmp, m, data, dst)
+
+	# tail recursion optimization: jmp to mont_reduce and return from there
+	 jmp	mont_reduce
+	# call	mont_reduce
+	# ret
+.size	mont_mul_a3b,.-mont_mul_a3b
+___
+}}}
+
+{{{
+#SQR_512 MACRO pDest, pA, x7, x6, x5, x4, x3, x2, x1, x0, tmp*4
+#
+# Input in memory [pA] and also in x7...x0
+# Uses all argument registers plus rax and rdx
+#
+# This version computes all of the off-diagonal terms into memory,
+# and then it adds in the diagonal terms
+
+sub SQR_512
+{
+ my ($pDst, $pA, $x, $A, $tmp, $x7, $x6, $pDst_o)=@_;
+ my ($pDst,  $pDst_o) = ($pDst =~ m/([^+]*)\+?(.*)?/);
+ my @X=@$x;	# make a copy
+$code.=<<___;
+	# ------------------
+	# first pass 01...07
+	# ------------------
+	 mov	$X[0], $A
+
+	 mov	$X[1],%rax
+	 mul	$A
+	 mov	%rax, (+$pDst_o+8*1)($pDst)
+___
+for(my $i=2;$i<8;$i++) {
+$code.=<<___;
+	 mov	%rdx, $X[$i-2]
+	 mov	$X[$i],%rax
+	 mul	$A
+	 add	%rax, $X[$i-2]
+	 adc	\$0, %rdx
+___
+}
+$code.=<<___;
+	 mov	%rdx, $x7
+
+	 mov	$X[0], (+$pDst_o+8*2)($pDst)
+
+	# ------------------
+	# second pass 12...17
+	# ------------------
+
+	 mov	(+8*1)($pA), $A
+
+	 mov	(+8*2)($pA),%rax
+	 mul	$A
+	 add	%rax, $X[1]
+	 adc	\$0, %rdx
+	 mov	$X[1], (+$pDst_o+8*3)($pDst)
+
+	 mov	%rdx, $X[0]
+	 mov	(+8*3)($pA),%rax
+	 mul	$A
+	 add	%rax, $X[2]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[2]
+	 adc	\$0, %rdx
+	 mov	$X[2], (+$pDst_o+8*4)($pDst)
+
+	 mov	%rdx, $X[0]
+	 mov	(+8*4)($pA),%rax
+	 mul	$A
+	 add	%rax, $X[3]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[3]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[0]
+	 mov	(+8*5)($pA),%rax
+	 mul	$A
+	 add	%rax, $X[4]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[4]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[0]
+	 mov	$X[6],%rax
+	 mul	$A
+	 add	%rax, $X[5]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[5]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[0]
+	 mov	$X[7],%rax
+	 mul	$A
+	 add	%rax, $x7
+	 adc	\$0, %rdx
+	 add	$X[0], $x7
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[1]
+
+	# ------------------
+	# third pass 23...27
+	# ------------------
+	 mov	(+8*2)($pA), $A
+
+	 mov	(+8*3)($pA),%rax
+	 mul	$A
+	 add	%rax, $X[3]
+	 adc	\$0, %rdx
+	 mov	$X[3], (+$pDst_o+8*5)($pDst)
+
+	 mov	%rdx, $X[0]
+	 mov	(+8*4)($pA),%rax
+	 mul	$A
+	 add	%rax, $X[4]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[4]
+	 adc	\$0, %rdx
+	 mov	$X[4], (+$pDst_o+8*6)($pDst)
+
+	 mov	%rdx, $X[0]
+	 mov	(+8*5)($pA),%rax
+	 mul	$A
+	 add	%rax, $X[5]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[5]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[0]
+	 mov	$X[6],%rax
+	 mul	$A
+	 add	%rax, $x7
+	 adc	\$0, %rdx
+	 add	$X[0], $x7
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[0]
+	 mov	$X[7],%rax
+	 mul	$A
+	 add	%rax, $X[1]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[1]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[2]
+
+	# ------------------
+	# fourth pass 34...37
+	# ------------------
+
+	 mov	(+8*3)($pA), $A
+
+	 mov	(+8*4)($pA),%rax
+	 mul	$A
+	 add	%rax, $X[5]
+	 adc	\$0, %rdx
+	 mov	$X[5], (+$pDst_o+8*7)($pDst)
+
+	 mov	%rdx, $X[0]
+	 mov	(+8*5)($pA),%rax
+	 mul	$A
+	 add	%rax, $x7
+	 adc	\$0, %rdx
+	 add	$X[0], $x7
+	 adc	\$0, %rdx
+	 mov	$x7, (+$pDst_o+8*8)($pDst)
+
+	 mov	%rdx, $X[0]
+	 mov	$X[6],%rax
+	 mul	$A
+	 add	%rax, $X[1]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[1]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[0]
+	 mov	$X[7],%rax
+	 mul	$A
+	 add	%rax, $X[2]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[2]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[5]
+
+	# ------------------
+	# fifth pass 45...47
+	# ------------------
+	 mov	(+8*4)($pA), $A
+
+	 mov	(+8*5)($pA),%rax
+	 mul	$A
+	 add	%rax, $X[1]
+	 adc	\$0, %rdx
+	 mov	$X[1], (+$pDst_o+8*9)($pDst)
+
+	 mov	%rdx, $X[0]
+	 mov	$X[6],%rax
+	 mul	$A
+	 add	%rax, $X[2]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[2]
+	 adc	\$0, %rdx
+	 mov	$X[2], (+$pDst_o+8*10)($pDst)
+
+	 mov	%rdx, $X[0]
+	 mov	$X[7],%rax
+	 mul	$A
+	 add	%rax, $X[5]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[5]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $X[1]
+
+	# ------------------
+	# sixth pass 56...57
+	# ------------------
+	 mov	(+8*5)($pA), $A
+
+	 mov	$X[6],%rax
+	 mul	$A
+	 add	%rax, $X[5]
+	 adc	\$0, %rdx
+	 mov	$X[5], (+$pDst_o+8*11)($pDst)
+
+	 mov	%rdx, $X[0]
+	 mov	$X[7],%rax
+	 mul	$A
+	 add	%rax, $X[1]
+	 adc	\$0, %rdx
+	 add	$X[0], $X[1]
+	 adc	\$0, %rdx
+	 mov	$X[1], (+$pDst_o+8*12)($pDst)
+
+	 mov	%rdx, $X[2]
+
+	# ------------------
+	# seventh pass 67
+	# ------------------
+	 mov	$X[6], $A
+
+	 mov	$X[7],%rax
+	 mul	$A
+	 add	%rax, $X[2]
+	 adc	\$0, %rdx
+	 mov	$X[2], (+$pDst_o+8*13)($pDst)
+
+	 mov	%rdx, (+$pDst_o+8*14)($pDst)
+
+	# start finalize (add	in squares, and double off-terms)
+	 mov	(+$pDst_o+8*1)($pDst), $X[0]
+	 mov	(+$pDst_o+8*2)($pDst), $X[1]
+	 mov	(+$pDst_o+8*3)($pDst), $X[2]
+	 mov	(+$pDst_o+8*4)($pDst), $X[3]
+	 mov	(+$pDst_o+8*5)($pDst), $X[4]
+	 mov	(+$pDst_o+8*6)($pDst), $X[5]
+
+	 mov	(+8*3)($pA), %rax
+	 mul	%rax
+	 mov	%rax, $x6
+	 mov	%rdx, $X[6]
+
+	 add	$X[0], $X[0]
+	 adc	$X[1], $X[1]
+	 adc	$X[2], $X[2]
+	 adc	$X[3], $X[3]
+	 adc	$X[4], $X[4]
+	 adc	$X[5], $X[5]
+	 adc	\$0, $X[6]
+
+	 mov	(+8*0)($pA), %rax
+	 mul	%rax
+	 mov	%rax, (+$pDst_o+8*0)($pDst)
+	 mov	%rdx, $A
+
+	 mov	(+8*1)($pA), %rax
+	 mul	%rax
+
+	 add	$A, $X[0]
+	 adc	%rax, $X[1]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $A
+	 mov	$X[0], (+$pDst_o+8*1)($pDst)
+	 mov	$X[1], (+$pDst_o+8*2)($pDst)
+
+	 mov	(+8*2)($pA), %rax
+	 mul	%rax
+
+	 add	$A, $X[2]
+	 adc	%rax, $X[3]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $A
+
+	 mov	$X[2], (+$pDst_o+8*3)($pDst)
+	 mov	$X[3], (+$pDst_o+8*4)($pDst)
+
+	 xor	$tmp, $tmp
+	 add	$A, $X[4]
+	 adc	$x6, $X[5]
+	 adc	\$0, $tmp
+
+	 mov	$X[4], (+$pDst_o+8*5)($pDst)
+	 mov	$X[5], (+$pDst_o+8*6)($pDst)
+
+	# %%tmp has 0/1 in column 7
+	# %%A6 has a full value in column 7
+
+	 mov	(+$pDst_o+8*7)($pDst), $X[0]
+	 mov	(+$pDst_o+8*8)($pDst), $X[1]
+	 mov	(+$pDst_o+8*9)($pDst), $X[2]
+	 mov	(+$pDst_o+8*10)($pDst), $X[3]
+	 mov	(+$pDst_o+8*11)($pDst), $X[4]
+	 mov	(+$pDst_o+8*12)($pDst), $X[5]
+	 mov	(+$pDst_o+8*13)($pDst), $x6
+	 mov	(+$pDst_o+8*14)($pDst), $x7
+
+	 mov	$X[7], %rax
+	 mul	%rax
+	 mov	%rax, $X[7]
+	 mov	%rdx, $A
+
+	 add	$X[0], $X[0]
+	 adc	$X[1], $X[1]
+	 adc	$X[2], $X[2]
+	 adc	$X[3], $X[3]
+	 adc	$X[4], $X[4]
+	 adc	$X[5], $X[5]
+	 adc	$x6, $x6
+	 adc	$x7, $x7
+	 adc	\$0, $A
+
+	 add	$tmp, $X[0]
+
+	 mov	(+8*4)($pA), %rax
+	 mul	%rax
+
+	 add	$X[6], $X[0]
+	 adc	%rax, $X[1]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $tmp
+
+	 mov	$X[0], (+$pDst_o+8*7)($pDst)
+	 mov	$X[1], (+$pDst_o+8*8)($pDst)
+
+	 mov	(+8*5)($pA), %rax
+	 mul	%rax
+
+	 add	$tmp, $X[2]
+	 adc	%rax, $X[3]
+	 adc	\$0, %rdx
+
+	 mov	%rdx, $tmp
+
+	 mov	$X[2], (+$pDst_o+8*9)($pDst)
+	 mov	$X[3], (+$pDst_o+8*10)($pDst)
+
+	 mov	(+8*6)($pA), %rax
+	 mul	%rax
+
+	 add	$tmp, $X[4]
+	 adc	%rax, $X[5]
+	 adc	\$0, %rdx
+
+	 mov	$X[4], (+$pDst_o+8*11)($pDst)
+	 mov	$X[5], (+$pDst_o+8*12)($pDst)
+
+	 add	%rdx, $x6
+	 adc	$X[7], $x7
+	 adc	\$0, $A
+
+	 mov	$x6, (+$pDst_o+8*13)($pDst)
+	 mov	$x7, (+$pDst_o+8*14)($pDst)
+	 mov	$A, (+$pDst_o+8*15)($pDst)
+___
+}
+
+#
+# sqr_reduce: subroutine to compute Result = reduce(Result * Result)
+#
+# input and result also in: r9, r8, r15, r14, r13, r12, r11, r10
+#
+$code.=<<___;
+.type	sqr_reduce,\@abi-omnipotent
+.align	16
+sqr_reduce:
+	 mov	(+$pResult_offset+8)(%rsp), %rcx
+___
+	&SQR_512("%rsp+$tmp16_offset+8", "%rcx", [map("%r$_",(10..15,8..9))], "%rbx", "%rbp", "%rsi", "%rdi");
+$code.=<<___;
+	# tail recursion optimization: jmp to mont_reduce and return from there
+	 jmp	mont_reduce
+	# call	mont_reduce
+	# ret
+.size	sqr_reduce,.-sqr_reduce
+___
+}}}
+
+#
+# MAIN FUNCTION
+#
+
+#mod_exp_512(UINT64 *result, /* 512 bits, 8 qwords */
+#           UINT64 *g,   /* 512 bits, 8 qwords */
+#           UINT64 *exp, /* 512 bits, 8 qwords */
+#           struct mod_ctx_512 *data)
+
+# window size = 5
+# table size = 2^5 = 32
+#table_entries	equ	32
+#table_size	equ	table_entries * 8
+$code.=<<___;
+.globl	mod_exp_512
+.type	mod_exp_512,\@function,4
+mod_exp_512:
+	 push	%rbp
+	 push	%rbx
+	 push	%r12
+	 push	%r13
+	 push	%r14
+	 push	%r15
+
+	# adjust stack down and then align it with cache boundary
+	 mov	%rsp, %r8
+	 sub	\$$mem_size, %rsp
+	 and	\$-64, %rsp
+
+	# store previous stack pointer and arguments
+	 mov	%r8, (+$rsp_offset)(%rsp)
+	 mov	%rdi, (+$pResult_offset)(%rsp)
+	 mov	%rsi, (+$pG_offset)(%rsp)
+	 mov	%rcx, (+$pData_offset)(%rsp)
+.Lbody:
+	# transform g into montgomery space
+	# GT = reduce(g * C2) = reduce(g * (2^256))
+	# reduce expects to have the input in [tmp16]
+	 pxor	%xmm4, %xmm4
+	 movdqu	(+16*0)(%rsi), %xmm0
+	 movdqu	(+16*1)(%rsi), %xmm1
+	 movdqu	(+16*2)(%rsi), %xmm2
+	 movdqu	(+16*3)(%rsi), %xmm3
+	 movdqa	%xmm4, (+$tmp16_offset+16*0)(%rsp)
+	 movdqa	%xmm4, (+$tmp16_offset+16*1)(%rsp)
+	 movdqa	%xmm4, (+$tmp16_offset+16*6)(%rsp)
+	 movdqa	%xmm4, (+$tmp16_offset+16*7)(%rsp)
+	 movdqa	%xmm0, (+$tmp16_offset+16*2)(%rsp)
+	 movdqa	%xmm1, (+$tmp16_offset+16*3)(%rsp)
+	 movdqa	%xmm2, (+$tmp16_offset+16*4)(%rsp)
+	 movdqa	%xmm3, (+$tmp16_offset+16*5)(%rsp)
+
+	# load pExp before rdx gets blown away
+	 movdqu	(+16*0)(%rdx), %xmm0
+	 movdqu	(+16*1)(%rdx), %xmm1
+	 movdqu	(+16*2)(%rdx), %xmm2
+	 movdqu	(+16*3)(%rdx), %xmm3
+
+	 lea	(+$GT_offset)(%rsp), %rbx
+	 mov	%rbx, (+$red_result_addr_offset)(%rsp)
+	 call	mont_reduce
+
+	# Initialize tmp = C
+	 lea	(+$tmp_offset)(%rsp), %rcx
+	 xor	%rax, %rax
+	 mov	%rax, (+8*0)(%rcx)
+	 mov	%rax, (+8*1)(%rcx)
+	 mov	%rax, (+8*3)(%rcx)
+	 mov	%rax, (+8*4)(%rcx)
+	 mov	%rax, (+8*5)(%rcx)
+	 mov	%rax, (+8*6)(%rcx)
+	 mov	%rax, (+8*7)(%rcx)
+	 mov	%rax, (+$exp_offset+8*8)(%rsp)
+	 movq	\$1, (+8*2)(%rcx)
+
+	 lea	(+$garray_offset)(%rsp), %rbp
+	 mov	%rcx, %rsi			# pTmp
+	 mov	%rbp, %rdi			# Garray[][0]
+___
+
+	&swizzle("%rdi", "%rcx", "%rax", "%rbx");
+
+	# for (rax = 31; rax != 0; rax--) {
+	#     tmp = reduce(tmp * G)
+	#     swizzle(pg, tmp);
+	#     pg += 2; }
+$code.=<<___;
+	 mov	\$31, %rax
+	 mov	%rax, (+$i_offset)(%rsp)
+	 mov	%rbp, (+$pg_offset)(%rsp)
+	# rsi -> pTmp
+	 mov	%rsi, (+$red_result_addr_offset)(%rsp)
+	 mov	(+8*0)(%rsi), %r10
+	 mov	(+8*1)(%rsi), %r11
+	 mov	(+8*2)(%rsi), %r12
+	 mov	(+8*3)(%rsi), %r13
+	 mov	(+8*4)(%rsi), %r14
+	 mov	(+8*5)(%rsi), %r15
+	 mov	(+8*6)(%rsi), %r8
+	 mov	(+8*7)(%rsi), %r9
+init_loop:
+	 lea	(+$GT_offset)(%rsp), %rdi
+	 call	mont_mul_a3b
+	 lea	(+$tmp_offset)(%rsp), %rsi
+	 mov	(+$pg_offset)(%rsp), %rbp
+	 add	\$2, %rbp
+	 mov	%rbp, (+$pg_offset)(%rsp)
+	 mov	%rsi, %rcx			# rcx = rsi = addr of tmp
+___
+
+	&swizzle("%rbp", "%rcx", "%rax", "%rbx");
+$code.=<<___;
+	 mov	(+$i_offset)(%rsp), %rax
+	 sub	\$1, %rax
+	 mov	%rax, (+$i_offset)(%rsp)
+	 jne	init_loop
+
+	#
+	# Copy exponent onto stack
+	 movdqa	%xmm0, (+$exp_offset+16*0)(%rsp)
+	 movdqa	%xmm1, (+$exp_offset+16*1)(%rsp)
+	 movdqa	%xmm2, (+$exp_offset+16*2)(%rsp)
+	 movdqa	%xmm3, (+$exp_offset+16*3)(%rsp)
+
+
+	#
+	# Do exponentiation
+	# Initialize result to G[exp{511:507}]
+	 mov	(+$exp_offset+62)(%rsp), %eax
+	 mov	%rax, %rdx
+	 shr	\$11, %rax
+	 and	\$0x07FF, %edx
+	 mov	%edx, (+$exp_offset+62)(%rsp)
+	 lea	(+$garray_offset)(%rsp,%rax,2), %rsi
+	 mov	(+$pResult_offset)(%rsp), %rdx
+___
+
+	&unswizzle("%rdx", "%rsi", "%rbp", "%rbx", "%rax");
+
+	#
+	# Loop variables
+	# rcx = [loop_idx] = index: 510-5 to 0 by 5
+$code.=<<___;
+	 movq	\$505, (+$loop_idx_offset)(%rsp)
+
+	 mov	(+$pResult_offset)(%rsp), %rcx
+	 mov	%rcx, (+$red_result_addr_offset)(%rsp)
+	 mov	(+8*0)(%rcx), %r10
+	 mov	(+8*1)(%rcx), %r11
+	 mov	(+8*2)(%rcx), %r12
+	 mov	(+8*3)(%rcx), %r13
+	 mov	(+8*4)(%rcx), %r14
+	 mov	(+8*5)(%rcx), %r15
+	 mov	(+8*6)(%rcx), %r8
+	 mov	(+8*7)(%rcx), %r9
+	 jmp	sqr_2
+
+main_loop_a3b:
+	 call	sqr_reduce
+	 call	sqr_reduce
+	 call	sqr_reduce
+sqr_2:
+	 call	sqr_reduce
+	 call	sqr_reduce
+
+	#
+	# Do multiply, first look up proper value in Garray
+	 mov	(+$loop_idx_offset)(%rsp), %rcx			# bit index
+	 mov	%rcx, %rax
+	 shr	\$4, %rax			# rax is word pointer
+	 mov	(+$exp_offset)(%rsp,%rax,2), %edx
+	 and	\$15, %rcx
+	 shrq	%cl, %rdx
+	 and	\$0x1F, %rdx
+
+	 lea	(+$garray_offset)(%rsp,%rdx,2), %rsi
+	 lea	(+$tmp_offset)(%rsp), %rdx
+	 mov	%rdx, %rdi
+___
+
+	&unswizzle("%rdx", "%rsi", "%rbp", "%rbx", "%rax");
+	# rdi = tmp = pG
+
+	#
+	# Call mod_mul_a1(pDst,  pSrc1, pSrc2, pM, pData)
+	#                 result result pG     M   Data
+$code.=<<___;
+	 mov	(+$pResult_offset)(%rsp), %rsi
+	 call	mont_mul_a3b
+
+	#
+	# finish loop
+	 mov	(+$loop_idx_offset)(%rsp), %rcx
+	 sub	\$5, %rcx
+	 mov	%rcx, (+$loop_idx_offset)(%rsp)
+	 jge	main_loop_a3b
+
+	#
+
+end_main_loop_a3b:
+	# transform result out of Montgomery space
+	# result = reduce(result)
+	 mov	(+$pResult_offset)(%rsp), %rdx
+	 pxor	%xmm4, %xmm4
+	 movdqu	(+16*0)(%rdx), %xmm0
+	 movdqu	(+16*1)(%rdx), %xmm1
+	 movdqu	(+16*2)(%rdx), %xmm2
+	 movdqu	(+16*3)(%rdx), %xmm3
+	 movdqa	%xmm4, (+$tmp16_offset+16*4)(%rsp)
+	 movdqa	%xmm4, (+$tmp16_offset+16*5)(%rsp)
+	 movdqa	%xmm4, (+$tmp16_offset+16*6)(%rsp)
+	 movdqa	%xmm4, (+$tmp16_offset+16*7)(%rsp)
+	 movdqa	%xmm0, (+$tmp16_offset+16*0)(%rsp)
+	 movdqa	%xmm1, (+$tmp16_offset+16*1)(%rsp)
+	 movdqa	%xmm2, (+$tmp16_offset+16*2)(%rsp)
+	 movdqa	%xmm3, (+$tmp16_offset+16*3)(%rsp)
+	 call	mont_reduce
+
+	# If result > m, subract m
+	# load result into r15:r8
+	 mov	(+$pResult_offset)(%rsp), %rax
+	 mov	(+8*0)(%rax), %r8
+	 mov	(+8*1)(%rax), %r9
+	 mov	(+8*2)(%rax), %r10
+	 mov	(+8*3)(%rax), %r11
+	 mov	(+8*4)(%rax), %r12
+	 mov	(+8*5)(%rax), %r13
+	 mov	(+8*6)(%rax), %r14
+	 mov	(+8*7)(%rax), %r15
+
+	# subtract m
+	 mov	(+$pData_offset)(%rsp), %rbx
+	 add	\$$M, %rbx
+
+	 sub	(+8*0)(%rbx), %r8
+	 sbb	(+8*1)(%rbx), %r9
+	 sbb	(+8*2)(%rbx), %r10
+	 sbb	(+8*3)(%rbx), %r11
+	 sbb	(+8*4)(%rbx), %r12
+	 sbb	(+8*5)(%rbx), %r13
+	 sbb	(+8*6)(%rbx), %r14
+	 sbb	(+8*7)(%rbx), %r15
+
+	# if Carry is clear, replace result with difference
+	 mov	(+8*0)(%rax), %rsi
+	 mov	(+8*1)(%rax), %rdi
+	 mov	(+8*2)(%rax), %rcx
+	 mov	(+8*3)(%rax), %rdx
+	 cmovnc	%r8, %rsi
+	 cmovnc	%r9, %rdi
+	 cmovnc	%r10, %rcx
+	 cmovnc	%r11, %rdx
+	 mov	%rsi, (+8*0)(%rax)
+	 mov	%rdi, (+8*1)(%rax)
+	 mov	%rcx, (+8*2)(%rax)
+	 mov	%rdx, (+8*3)(%rax)
+
+	 mov	(+8*4)(%rax), %rsi
+	 mov	(+8*5)(%rax), %rdi
+	 mov	(+8*6)(%rax), %rcx
+	 mov	(+8*7)(%rax), %rdx
+	 cmovnc	%r12, %rsi
+	 cmovnc	%r13, %rdi
+	 cmovnc	%r14, %rcx
+	 cmovnc	%r15, %rdx
+	 mov	%rsi, (+8*4)(%rax)
+	 mov	%rdi, (+8*5)(%rax)
+	 mov	%rcx, (+8*6)(%rax)
+	 mov	%rdx, (+8*7)(%rax)
+
+	 mov	(+$rsp_offset)(%rsp), %rsi
+	 mov	0(%rsi),%r15
+	 mov	8(%rsi),%r14
+	 mov	16(%rsi),%r13
+	 mov	24(%rsi),%r12
+	 mov	32(%rsi),%rbx
+	 mov	40(%rsi),%rbp
+	 lea	48(%rsi),%rsp
+.Lepilogue:
+	 ret
+.size mod_exp_512, . - mod_exp_512
+___
+
+if ($win64) {
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+my $rec="%rcx";
+my $frame="%rdx";
+my $context="%r8";
+my $disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	mod_exp_512_se_handler,\@abi-omnipotent
+.align	16
+mod_exp_512_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lbody(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lepilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lin_prologue
+
+	mov	$rsp_offset(%rax),%rax	# pull saved Rsp
+
+	mov	32(%rax),%rbx
+	mov	40(%rax),%rbp
+	mov	24(%rax),%r12
+	mov	16(%rax),%r13
+	mov	8(%rax),%r14
+	mov	0(%rax),%r15
+	lea	48(%rax),%rax
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	mod_exp_512_se_handler,.-mod_exp_512_se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_mod_exp_512
+	.rva	.LSEH_end_mod_exp_512
+	.rva	.LSEH_info_mod_exp_512
+
+.section	.xdata
+.align	8
+.LSEH_info_mod_exp_512:
+	.byte	9,0,0,0
+	.rva	mod_exp_512_se_handler
+___
+}
+
+sub reg_part {
+my ($reg,$conv)=@_;
+    if ($reg =~ /%r[0-9]+/)	{ $reg .= $conv; }
+    elsif ($conv eq "b")	{ $reg =~ s/%[er]([^x]+)x?/%$1l/;	}
+    elsif ($conv eq "w")	{ $reg =~ s/%[er](.+)/%$1/;		}
+    elsif ($conv eq "d")	{ $reg =~ s/%[er](.+)/%e$1/;		}
+    return $reg;
+}
+
+$code =~ s/(%[a-z0-9]+)#([bwd])/reg_part($1,$2)/gem;
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/(\(\+[^)]+\))/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/parisc-mont.pl b/crypto/bn/asm/parisc-mont.pl
new file mode 100755
index 000000000000..4a766a87fb2e
--- /dev/null
+++ b/crypto/bn/asm/parisc-mont.pl
@@ -0,0 +1,993 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# On PA-7100LC this module performs ~90-50% better, less for longer
+# keys, than code generated by gcc 3.2 for PA-RISC 1.1. Latter means
+# that compiler utilized xmpyu instruction to perform 32x32=64-bit
+# multiplication, which in turn means that "baseline" performance was
+# optimal in respect to instruction set capabilities. Fair comparison
+# with vendor compiler is problematic, because OpenSSL doesn't define
+# BN_LLONG [presumably] for historical reasons, which drives compiler
+# toward 4 times 16x16=32-bit multiplicatons [plus complementary
+# shifts and additions] instead. This means that you should observe
+# several times improvement over code generated by vendor compiler
+# for PA-RISC 1.1, but the "baseline" is far from optimal. The actual
+# improvement coefficient was never collected on PA-7100LC, or any
+# other 1.1 CPU, because I don't have access to such machine with
+# vendor compiler. But to give you a taste, PA-RISC 1.1 code path
+# reportedly outperformed code generated by cc +DA1.1 +O3 by factor
+# of ~5x on PA-8600.
+#
+# On PA-RISC 2.0 it has to compete with pa-risc2[W].s, which is
+# reportedly ~2x faster than vendor compiler generated code [according
+# to comment in pa-risc2[W].s]. Here comes a catch. Execution core of
+# this implementation is actually 32-bit one, in the sense that it
+# operates on 32-bit values. But pa-risc2[W].s operates on arrays of
+# 64-bit BN_LONGs... How do they interoperate then? No problem. This
+# module picks halves of 64-bit values in reverse order and pretends
+# they were 32-bit BN_LONGs. But can 32-bit core compete with "pure"
+# 64-bit code such as pa-risc2[W].s then? Well, the thing is that
+# 32x32=64-bit multiplication is the best even PA-RISC 2.0 can do,
+# i.e. there is no "wider" multiplication like on most other 64-bit
+# platforms. This means that even being effectively 32-bit, this
+# implementation performs "64-bit" computational task in same amount
+# of arithmetic operations, most notably multiplications. It requires
+# more memory references, most notably to tp[num], but this doesn't
+# seem to exhaust memory port capacity. And indeed, dedicated PA-RISC
+# 2.0 code path, provides virtually same performance as pa-risc2[W].s:
+# it's ~10% better for shortest key length and ~10% worse for longest
+# one.
+#
+# In case it wasn't clear. The module has two distinct code paths:
+# PA-RISC 1.1 and PA-RISC 2.0 ones. Latter features carry-free 64-bit
+# additions and 64-bit integer loads, not to mention specific
+# instruction scheduling. In 64-bit build naturally only 2.0 code path
+# is assembled. In 32-bit application context both code paths are
+# assembled, PA-RISC 2.0 CPU is detected at run-time and proper path
+# is taken automatically. Also, in 32-bit build the module imposes
+# couple of limitations: vector lengths has to be even and vector
+# addresses has to be 64-bit aligned. Normally neither is a problem:
+# most common key lengths are even and vectors are commonly malloc-ed,
+# which ensures alignment.
+#
+# Special thanks to polarhome.com for providing HP-UX account on
+# PA-RISC 1.1 machine, and to correspondent who chose to remain
+# anonymous for testing the code on PA-RISC 2.0 machine.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+
+$flavour = shift;
+$output = shift;
+
+open STDOUT,">$output";
+
+if ($flavour =~ /64/) {
+	$LEVEL		="2.0W";
+	$SIZE_T		=8;
+	$FRAME_MARKER	=80;
+	$SAVED_RP	=16;
+	$PUSH		="std";
+	$PUSHMA		="std,ma";
+	$POP		="ldd";
+	$POPMB		="ldd,mb";
+	$BN_SZ		=$SIZE_T;
+} else {
+	$LEVEL		="1.1";	#$LEVEL.="\n\t.ALLOW\t2.0";
+	$SIZE_T		=4;
+	$FRAME_MARKER	=48;
+	$SAVED_RP	=20;
+	$PUSH		="stw";
+	$PUSHMA		="stwm";
+	$POP		="ldw";
+	$POPMB		="ldwm";
+	$BN_SZ		=$SIZE_T;
+	if (open CONF,"<${dir}../../opensslconf.h") {
+	    while(<CONF>) {
+		if (m/#\s*define\s+SIXTY_FOUR_BIT/) {
+		    $BN_SZ=8;
+		    $LEVEL="2.0";
+		    last;
+		}
+	    }
+	    close CONF;
+	}
+}
+
+$FRAME=8*$SIZE_T+$FRAME_MARKER;	# 8 saved regs + frame marker
+				#                [+ argument transfer]
+$LOCALS=$FRAME-$FRAME_MARKER;
+$FRAME+=32;			# local variables
+
+$tp="%r31";
+$ti1="%r29";
+$ti0="%r28";
+
+$rp="%r26";
+$ap="%r25";
+$bp="%r24";
+$np="%r23";
+$n0="%r22";	# passed through stack in 32-bit
+$num="%r21";	# passed through stack in 32-bit
+$idx="%r20";
+$arrsz="%r19";
+
+$nm1="%r7";
+$nm0="%r6";
+$ab1="%r5";
+$ab0="%r4";
+
+$fp="%r3";
+$hi1="%r2";
+$hi0="%r1";
+
+$xfer=$n0;	# accomodates [-16..15] offset in fld[dw]s
+
+$fm0="%fr4";	$fti=$fm0;
+$fbi="%fr5L";
+$fn0="%fr5R";
+$fai="%fr6";	$fab0="%fr7";	$fab1="%fr8";
+$fni="%fr9";	$fnm0="%fr10";	$fnm1="%fr11";
+
+$code=<<___;
+	.LEVEL	$LEVEL
+	.SPACE	\$TEXT\$
+	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY
+
+	.EXPORT	bn_mul_mont,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR,ARGW3=GR
+	.ALIGN	64
+bn_mul_mont
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-8*$SIZE_T`,NO_CALLS,SAVE_RP,SAVE_SP,ENTRY_GR=6
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)		; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
+	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
+	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
+	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
+	ldo	-$FRAME(%sp),$fp
+___
+$code.=<<___ if ($SIZE_T==4);
+	ldw	`-$FRAME_MARKER-4`($fp),$n0
+	ldw	`-$FRAME_MARKER-8`($fp),$num
+	nop
+	nop					; alignment
+___
+$code.=<<___ if ($BN_SZ==4);
+	comiclr,<=	6,$num,%r0		; are vectors long enough?
+	b		L\$abort
+	ldi		0,%r28			; signal "unhandled"
+	add,ev		%r0,$num,$num		; is $num even?
+	b		L\$abort
+	nop
+	or		$ap,$np,$ti1
+	extru,=		$ti1,31,3,%r0		; are ap and np 64-bit aligned?
+	b		L\$abort
+	nop
+	nop					; alignment
+	nop
+
+	fldws		0($n0),${fn0}
+	fldws,ma	4($bp),${fbi}		; bp[0]
+___
+$code.=<<___ if ($BN_SZ==8);
+	comib,>		3,$num,L\$abort		; are vectors long enough?
+	ldi		0,%r28			; signal "unhandled"
+	addl		$num,$num,$num		; I operate on 32-bit values
+
+	fldws		4($n0),${fn0}		; only low part of n0
+	fldws		4($bp),${fbi}		; bp[0] in flipped word order
+___
+$code.=<<___;
+	fldds		0($ap),${fai}		; ap[0,1]
+	fldds		0($np),${fni}		; np[0,1]
+
+	sh2addl		$num,%r0,$arrsz
+	ldi		31,$hi0
+	ldo		36($arrsz),$hi1		; space for tp[num+1]
+	andcm		$hi1,$hi0,$hi1		; align
+	addl		$hi1,%sp,%sp
+	$PUSH		$fp,-$SIZE_T(%sp)
+
+	ldo		`$LOCALS+16`($fp),$xfer
+	ldo		`$LOCALS+32+4`($fp),$tp
+
+	xmpyu		${fai}L,${fbi},${fab0}	; ap[0]*bp[0]
+	xmpyu		${fai}R,${fbi},${fab1}	; ap[1]*bp[0]
+	xmpyu		${fn0},${fab0}R,${fm0}
+
+	addl		$arrsz,$ap,$ap		; point at the end
+	addl		$arrsz,$np,$np
+	subi		0,$arrsz,$idx		; j=0
+	ldo		8($idx),$idx		; j++++
+
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[0]*m
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[1]*m
+	fstds		${fab0},-16($xfer)
+	fstds		${fnm0},-8($xfer)
+	fstds		${fab1},0($xfer)
+	fstds		${fnm1},8($xfer)
+	 flddx		$idx($ap),${fai}	; ap[2,3]
+	 flddx		$idx($np),${fni}	; np[2,3]
+___
+$code.=<<___ if ($BN_SZ==4);
+	mtctl		$hi0,%cr11		; $hi0 still holds 31
+	extrd,u,*=	$hi0,%sar,1,$hi0	; executes on PA-RISC 1.0
+	b		L\$parisc11
+	nop
+___
+$code.=<<___;					# PA-RISC 2.0 code-path
+	xmpyu		${fai}L,${fbi},${fab0}	; ap[j]*bp[0]
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[j]*m
+	ldd		-16($xfer),$ab0
+	fstds		${fab0},-16($xfer)
+
+	extrd,u		$ab0,31,32,$hi0
+	extrd,u		$ab0,63,32,$ab0
+	ldd		-8($xfer),$nm0
+	fstds		${fnm0},-8($xfer)
+	 ldo		8($idx),$idx		; j++++
+	 addl		$ab0,$nm0,$nm0		; low part is discarded
+	 extrd,u	$nm0,31,32,$hi1
+
+L\$1st
+	xmpyu		${fai}R,${fbi},${fab1}	; ap[j+1]*bp[0]
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[j+1]*m
+	ldd		0($xfer),$ab1
+	fstds		${fab1},0($xfer)
+	 addl		$hi0,$ab1,$ab1
+	 extrd,u	$ab1,31,32,$hi0
+	ldd		8($xfer),$nm1
+	fstds		${fnm1},8($xfer)
+	 extrd,u	$ab1,63,32,$ab1
+	 addl		$hi1,$nm1,$nm1
+	flddx		$idx($ap),${fai}	; ap[j,j+1]
+	flddx		$idx($np),${fni}	; np[j,j+1]
+	 addl		$ab1,$nm1,$nm1
+	 extrd,u	$nm1,31,32,$hi1
+
+	xmpyu		${fai}L,${fbi},${fab0}	; ap[j]*bp[0]
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[j]*m
+	ldd		-16($xfer),$ab0
+	fstds		${fab0},-16($xfer)
+	 addl		$hi0,$ab0,$ab0
+	 extrd,u	$ab0,31,32,$hi0
+	ldd		-8($xfer),$nm0
+	fstds		${fnm0},-8($xfer)
+	 extrd,u	$ab0,63,32,$ab0
+	 addl		$hi1,$nm0,$nm0
+	stw		$nm1,-4($tp)		; tp[j-1]
+	 addl		$ab0,$nm0,$nm0
+	 stw,ma		$nm0,8($tp)		; tp[j-1]
+	addib,<>	8,$idx,L\$1st		; j++++
+	 extrd,u	$nm0,31,32,$hi1
+
+	xmpyu		${fai}R,${fbi},${fab1}	; ap[j]*bp[0]
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[j]*m
+	ldd		0($xfer),$ab1
+	fstds		${fab1},0($xfer)
+	 addl		$hi0,$ab1,$ab1
+	 extrd,u	$ab1,31,32,$hi0
+	ldd		8($xfer),$nm1
+	fstds		${fnm1},8($xfer)
+	 extrd,u	$ab1,63,32,$ab1
+	 addl		$hi1,$nm1,$nm1
+	ldd		-16($xfer),$ab0
+	 addl		$ab1,$nm1,$nm1
+	ldd		-8($xfer),$nm0
+	 extrd,u	$nm1,31,32,$hi1
+
+	 addl		$hi0,$ab0,$ab0
+	 extrd,u	$ab0,31,32,$hi0
+	stw		$nm1,-4($tp)		; tp[j-1]
+	 extrd,u	$ab0,63,32,$ab0
+	 addl		$hi1,$nm0,$nm0
+	ldd		0($xfer),$ab1
+	 addl		$ab0,$nm0,$nm0
+	ldd,mb		8($xfer),$nm1
+	 extrd,u	$nm0,31,32,$hi1
+	stw,ma		$nm0,8($tp)		; tp[j-1]
+
+	ldo		-1($num),$num		; i--
+	subi		0,$arrsz,$idx		; j=0
+___
+$code.=<<___ if ($BN_SZ==4);
+	fldws,ma	4($bp),${fbi}		; bp[1]
+___
+$code.=<<___ if ($BN_SZ==8);
+	fldws		0($bp),${fbi}		; bp[1] in flipped word order
+___
+$code.=<<___;
+	 flddx		$idx($ap),${fai}	; ap[0,1]
+	 flddx		$idx($np),${fni}	; np[0,1]
+	 fldws		8($xfer),${fti}R	; tp[0]
+	addl		$hi0,$ab1,$ab1
+	 extrd,u	$ab1,31,32,$hi0
+	 extrd,u	$ab1,63,32,$ab1
+	 ldo		8($idx),$idx		; j++++
+	 xmpyu		${fai}L,${fbi},${fab0}	; ap[0]*bp[1]
+	 xmpyu		${fai}R,${fbi},${fab1}	; ap[1]*bp[1]
+	addl		$hi1,$nm1,$nm1
+	addl		$ab1,$nm1,$nm1
+	extrd,u		$nm1,31,32,$hi1
+	 fstws,mb	${fab0}L,-8($xfer)	; save high part
+	stw		$nm1,-4($tp)		; tp[j-1]
+
+	 fcpy,sgl	%fr0,${fti}L		; zero high part
+	 fcpy,sgl	%fr0,${fab0}L
+	addl		$hi1,$hi0,$hi0
+	extrd,u		$hi0,31,32,$hi1
+	 fcnvxf,dbl,dbl	${fti},${fti}		; 32-bit unsigned int -> double
+	 fcnvxf,dbl,dbl	${fab0},${fab0}
+	stw		$hi0,0($tp)
+	stw		$hi1,4($tp)
+
+	fadd,dbl	${fti},${fab0},${fab0}	; add tp[0]
+	fcnvfx,dbl,dbl	${fab0},${fab0}		; double -> 33-bit unsigned int
+	xmpyu		${fn0},${fab0}R,${fm0}
+	ldo		`$LOCALS+32+4`($fp),$tp
+L\$outer
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[0]*m
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[1]*m
+	fstds		${fab0},-16($xfer)	; 33-bit value
+	fstds		${fnm0},-8($xfer)
+	 flddx		$idx($ap),${fai}	; ap[2]
+	 flddx		$idx($np),${fni}	; np[2]
+	 ldo		8($idx),$idx		; j++++
+	ldd		-16($xfer),$ab0		; 33-bit value
+	ldd		-8($xfer),$nm0
+	ldw		0($xfer),$hi0		; high part
+
+	xmpyu		${fai}L,${fbi},${fab0}	; ap[j]*bp[i]
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[j]*m
+	 extrd,u	$ab0,31,32,$ti0		; carry bit
+	 extrd,u	$ab0,63,32,$ab0
+	fstds		${fab1},0($xfer)
+	 addl		$ti0,$hi0,$hi0		; account carry bit
+	fstds		${fnm1},8($xfer)
+	 addl		$ab0,$nm0,$nm0		; low part is discarded
+	ldw		0($tp),$ti1		; tp[1]
+	 extrd,u	$nm0,31,32,$hi1
+	fstds		${fab0},-16($xfer)
+	fstds		${fnm0},-8($xfer)
+
+L\$inner
+	xmpyu		${fai}R,${fbi},${fab1}	; ap[j+1]*bp[i]
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[j+1]*m
+	ldd		0($xfer),$ab1
+	fstds		${fab1},0($xfer)
+	 addl		$hi0,$ti1,$ti1
+	 addl		$ti1,$ab1,$ab1
+	ldd		8($xfer),$nm1
+	fstds		${fnm1},8($xfer)
+	 extrd,u	$ab1,31,32,$hi0
+	 extrd,u	$ab1,63,32,$ab1
+	flddx		$idx($ap),${fai}	; ap[j,j+1]
+	flddx		$idx($np),${fni}	; np[j,j+1]
+	 addl		$hi1,$nm1,$nm1
+	 addl		$ab1,$nm1,$nm1
+	ldw		4($tp),$ti0		; tp[j]
+	stw		$nm1,-4($tp)		; tp[j-1]
+
+	xmpyu		${fai}L,${fbi},${fab0}	; ap[j]*bp[i]
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[j]*m
+	ldd		-16($xfer),$ab0
+	fstds		${fab0},-16($xfer)
+	 addl		$hi0,$ti0,$ti0
+	 addl		$ti0,$ab0,$ab0
+	ldd		-8($xfer),$nm0
+	fstds		${fnm0},-8($xfer)
+	 extrd,u	$ab0,31,32,$hi0
+	 extrd,u	$nm1,31,32,$hi1
+	ldw		8($tp),$ti1		; tp[j]
+	 extrd,u	$ab0,63,32,$ab0
+	 addl		$hi1,$nm0,$nm0
+	 addl		$ab0,$nm0,$nm0
+	 stw,ma		$nm0,8($tp)		; tp[j-1]
+	addib,<>	8,$idx,L\$inner		; j++++
+	 extrd,u	$nm0,31,32,$hi1
+
+	xmpyu		${fai}R,${fbi},${fab1}	; ap[j]*bp[i]
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[j]*m
+	ldd		0($xfer),$ab1
+	fstds		${fab1},0($xfer)
+	 addl		$hi0,$ti1,$ti1
+	 addl		$ti1,$ab1,$ab1
+	ldd		8($xfer),$nm1
+	fstds		${fnm1},8($xfer)
+	 extrd,u	$ab1,31,32,$hi0
+	 extrd,u	$ab1,63,32,$ab1
+	ldw		4($tp),$ti0		; tp[j]
+	 addl		$hi1,$nm1,$nm1
+	 addl		$ab1,$nm1,$nm1
+	ldd		-16($xfer),$ab0
+	ldd		-8($xfer),$nm0
+	 extrd,u	$nm1,31,32,$hi1
+
+	addl		$hi0,$ab0,$ab0
+	 addl		$ti0,$ab0,$ab0
+	 stw		$nm1,-4($tp)		; tp[j-1]
+	 extrd,u	$ab0,31,32,$hi0
+	ldw		8($tp),$ti1		; tp[j]
+	 extrd,u	$ab0,63,32,$ab0
+	 addl		$hi1,$nm0,$nm0
+	ldd		0($xfer),$ab1
+	 addl		$ab0,$nm0,$nm0
+	ldd,mb		8($xfer),$nm1
+	 extrd,u	$nm0,31,32,$hi1
+	 stw,ma		$nm0,8($tp)		; tp[j-1]
+
+	addib,=		-1,$num,L\$outerdone	; i--
+	subi		0,$arrsz,$idx		; j=0
+___
+$code.=<<___ if ($BN_SZ==4);
+	fldws,ma	4($bp),${fbi}		; bp[i]
+___
+$code.=<<___ if ($BN_SZ==8);
+	ldi		12,$ti0			; bp[i] in flipped word order
+	addl,ev		%r0,$num,$num
+	ldi		-4,$ti0
+	addl		$ti0,$bp,$bp
+	fldws		0($bp),${fbi}
+___
+$code.=<<___;
+	 flddx		$idx($ap),${fai}	; ap[0]
+	addl		$hi0,$ab1,$ab1
+	 flddx		$idx($np),${fni}	; np[0]
+	 fldws		8($xfer),${fti}R	; tp[0]
+	addl		$ti1,$ab1,$ab1
+	extrd,u		$ab1,31,32,$hi0
+	extrd,u		$ab1,63,32,$ab1
+
+	 ldo		8($idx),$idx		; j++++
+	 xmpyu		${fai}L,${fbi},${fab0}	; ap[0]*bp[i]
+	 xmpyu		${fai}R,${fbi},${fab1}	; ap[1]*bp[i]
+	ldw		4($tp),$ti0		; tp[j]
+
+	addl		$hi1,$nm1,$nm1
+	 fstws,mb	${fab0}L,-8($xfer)	; save high part
+	addl		$ab1,$nm1,$nm1
+	extrd,u		$nm1,31,32,$hi1
+	 fcpy,sgl	%fr0,${fti}L		; zero high part
+	 fcpy,sgl	%fr0,${fab0}L
+	stw		$nm1,-4($tp)		; tp[j-1]
+
+	 fcnvxf,dbl,dbl	${fti},${fti}		; 32-bit unsigned int -> double
+	 fcnvxf,dbl,dbl	${fab0},${fab0}
+	addl		$hi1,$hi0,$hi0
+	 fadd,dbl	${fti},${fab0},${fab0}	; add tp[0]
+	addl		$ti0,$hi0,$hi0
+	extrd,u		$hi0,31,32,$hi1
+	 fcnvfx,dbl,dbl	${fab0},${fab0}		; double -> 33-bit unsigned int
+	stw		$hi0,0($tp)
+	stw		$hi1,4($tp)
+	 xmpyu		${fn0},${fab0}R,${fm0}
+
+	b		L\$outer
+	ldo		`$LOCALS+32+4`($fp),$tp
+
+L\$outerdone
+	addl		$hi0,$ab1,$ab1
+	addl		$ti1,$ab1,$ab1
+	extrd,u		$ab1,31,32,$hi0
+	extrd,u		$ab1,63,32,$ab1
+
+	ldw		4($tp),$ti0		; tp[j]
+
+	addl		$hi1,$nm1,$nm1
+	addl		$ab1,$nm1,$nm1
+	extrd,u		$nm1,31,32,$hi1
+	stw		$nm1,-4($tp)		; tp[j-1]
+
+	addl		$hi1,$hi0,$hi0
+	addl		$ti0,$hi0,$hi0
+	extrd,u		$hi0,31,32,$hi1
+	stw		$hi0,0($tp)
+	stw		$hi1,4($tp)
+
+	ldo		`$LOCALS+32`($fp),$tp
+	sub		%r0,%r0,%r0		; clear borrow
+___
+$code.=<<___ if ($BN_SZ==4);
+	ldws,ma		4($tp),$ti0
+	extru,=		$rp,31,3,%r0		; is rp 64-bit aligned?
+	b		L\$sub_pa11
+	addl		$tp,$arrsz,$tp
+L\$sub
+	ldwx		$idx($np),$hi0
+	subb		$ti0,$hi0,$hi1
+	ldwx		$idx($tp),$ti0
+	addib,<>	4,$idx,L\$sub
+	stws,ma		$hi1,4($rp)
+
+	subb		$ti0,%r0,$hi1
+	ldo		-4($tp),$tp
+___
+$code.=<<___ if ($BN_SZ==8);
+	ldd,ma		8($tp),$ti0
+L\$sub
+	ldd		$idx($np),$hi0
+	shrpd		$ti0,$ti0,32,$ti0	; flip word order
+	std		$ti0,-8($tp)		; save flipped value
+	sub,db		$ti0,$hi0,$hi1
+	ldd,ma		8($tp),$ti0
+	addib,<>	8,$idx,L\$sub
+	std,ma		$hi1,8($rp)
+
+	extrd,u		$ti0,31,32,$ti0		; carry in flipped word order
+	sub,db		$ti0,%r0,$hi1
+	ldo		-8($tp),$tp
+___
+$code.=<<___;
+	and		$tp,$hi1,$ap
+	andcm		$rp,$hi1,$bp
+	or		$ap,$bp,$np
+
+	sub		$rp,$arrsz,$rp		; rewind rp
+	subi		0,$arrsz,$idx
+	ldo		`$LOCALS+32`($fp),$tp
+L\$copy
+	ldd		$idx($np),$hi0
+	std,ma		%r0,8($tp)
+	addib,<>	8,$idx,.-8		; L\$copy
+	std,ma		$hi0,8($rp)	
+___
+
+if ($BN_SZ==4) {				# PA-RISC 1.1 code-path
+$ablo=$ab0;
+$abhi=$ab1;
+$nmlo0=$nm0;
+$nmhi0=$nm1;
+$nmlo1="%r9";
+$nmhi1="%r8";
+
+$code.=<<___;
+	b		L\$done
+	nop
+
+	.ALIGN		8
+L\$parisc11
+	xmpyu		${fai}L,${fbi},${fab0}	; ap[j]*bp[0]
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[j]*m
+	ldw		-12($xfer),$ablo
+	ldw		-16($xfer),$hi0
+	ldw		-4($xfer),$nmlo0
+	ldw		-8($xfer),$nmhi0
+	fstds		${fab0},-16($xfer)
+	fstds		${fnm0},-8($xfer)
+
+	 ldo		8($idx),$idx		; j++++
+	 add		$ablo,$nmlo0,$nmlo0	; discarded
+	 addc		%r0,$nmhi0,$hi1
+	ldw		4($xfer),$ablo
+	ldw		0($xfer),$abhi
+	nop
+
+L\$1st_pa11
+	xmpyu		${fai}R,${fbi},${fab1}	; ap[j+1]*bp[0]
+	flddx		$idx($ap),${fai}	; ap[j,j+1]
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[j+1]*m
+	flddx		$idx($np),${fni}	; np[j,j+1]
+	 add		$hi0,$ablo,$ablo
+	ldw		12($xfer),$nmlo1
+	 addc		%r0,$abhi,$hi0
+	ldw		8($xfer),$nmhi1
+	 add		$ablo,$nmlo1,$nmlo1
+	fstds		${fab1},0($xfer)
+	 addc		%r0,$nmhi1,$nmhi1
+	fstds		${fnm1},8($xfer)
+	 add		$hi1,$nmlo1,$nmlo1
+	ldw		-12($xfer),$ablo
+	 addc		%r0,$nmhi1,$hi1
+	ldw		-16($xfer),$abhi
+
+	xmpyu		${fai}L,${fbi},${fab0}	; ap[j]*bp[0]
+	ldw		-4($xfer),$nmlo0
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[j]*m
+	ldw		-8($xfer),$nmhi0
+	 add		$hi0,$ablo,$ablo
+	stw		$nmlo1,-4($tp)		; tp[j-1]
+	 addc		%r0,$abhi,$hi0
+	fstds		${fab0},-16($xfer)
+	 add		$ablo,$nmlo0,$nmlo0
+	fstds		${fnm0},-8($xfer)
+	 addc		%r0,$nmhi0,$nmhi0
+	ldw		0($xfer),$abhi
+	 add		$hi1,$nmlo0,$nmlo0
+	ldw		4($xfer),$ablo
+	 stws,ma	$nmlo0,8($tp)		; tp[j-1]
+	addib,<>	8,$idx,L\$1st_pa11	; j++++
+	 addc		%r0,$nmhi0,$hi1
+
+	 ldw		8($xfer),$nmhi1
+	 ldw		12($xfer),$nmlo1
+	xmpyu		${fai}R,${fbi},${fab1}	; ap[j]*bp[0]
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[j]*m
+	 add		$hi0,$ablo,$ablo
+	fstds		${fab1},0($xfer)
+	 addc		%r0,$abhi,$hi0
+	fstds		${fnm1},8($xfer)
+	 add		$ablo,$nmlo1,$nmlo1
+	ldw		-16($xfer),$abhi
+	 addc		%r0,$nmhi1,$nmhi1
+	ldw		-12($xfer),$ablo
+	 add		$hi1,$nmlo1,$nmlo1
+	ldw		-8($xfer),$nmhi0
+	 addc		%r0,$nmhi1,$hi1
+	ldw		-4($xfer),$nmlo0
+
+	 add		$hi0,$ablo,$ablo
+	stw		$nmlo1,-4($tp)		; tp[j-1]
+	 addc		%r0,$abhi,$hi0
+	ldw		0($xfer),$abhi
+	 add		$ablo,$nmlo0,$nmlo0
+	ldw		4($xfer),$ablo
+	 addc		%r0,$nmhi0,$nmhi0
+	ldws,mb		8($xfer),$nmhi1
+	 add		$hi1,$nmlo0,$nmlo0
+	ldw		4($xfer),$nmlo1
+	 addc		%r0,$nmhi0,$hi1
+	stws,ma		$nmlo0,8($tp)		; tp[j-1]
+
+	ldo		-1($num),$num		; i--
+	subi		0,$arrsz,$idx		; j=0
+
+	 fldws,ma	4($bp),${fbi}		; bp[1]
+	 flddx		$idx($ap),${fai}	; ap[0,1]
+	 flddx		$idx($np),${fni}	; np[0,1]
+	 fldws		8($xfer),${fti}R	; tp[0]
+	add		$hi0,$ablo,$ablo
+	addc		%r0,$abhi,$hi0
+	 ldo		8($idx),$idx		; j++++
+	 xmpyu		${fai}L,${fbi},${fab0}	; ap[0]*bp[1]
+	 xmpyu		${fai}R,${fbi},${fab1}	; ap[1]*bp[1]
+	add		$hi1,$nmlo1,$nmlo1
+	addc		%r0,$nmhi1,$nmhi1
+	add		$ablo,$nmlo1,$nmlo1
+	addc		%r0,$nmhi1,$hi1
+	 fstws,mb	${fab0}L,-8($xfer)	; save high part
+	stw		$nmlo1,-4($tp)		; tp[j-1]
+
+	 fcpy,sgl	%fr0,${fti}L		; zero high part
+	 fcpy,sgl	%fr0,${fab0}L
+	add		$hi1,$hi0,$hi0
+	addc		%r0,%r0,$hi1
+	 fcnvxf,dbl,dbl	${fti},${fti}		; 32-bit unsigned int -> double
+	 fcnvxf,dbl,dbl	${fab0},${fab0}
+	stw		$hi0,0($tp)
+	stw		$hi1,4($tp)
+
+	fadd,dbl	${fti},${fab0},${fab0}	; add tp[0]
+	fcnvfx,dbl,dbl	${fab0},${fab0}		; double -> 33-bit unsigned int
+	xmpyu		${fn0},${fab0}R,${fm0}
+	ldo		`$LOCALS+32+4`($fp),$tp
+L\$outer_pa11
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[0]*m
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[1]*m
+	fstds		${fab0},-16($xfer)	; 33-bit value
+	fstds		${fnm0},-8($xfer)
+	 flddx		$idx($ap),${fai}	; ap[2,3]
+	 flddx		$idx($np),${fni}	; np[2,3]
+	ldw		-16($xfer),$abhi	; carry bit actually
+	 ldo		8($idx),$idx		; j++++
+	ldw		-12($xfer),$ablo
+	ldw		-8($xfer),$nmhi0
+	ldw		-4($xfer),$nmlo0
+	ldw		0($xfer),$hi0		; high part
+
+	xmpyu		${fai}L,${fbi},${fab0}	; ap[j]*bp[i]
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[j]*m
+	fstds		${fab1},0($xfer)
+	 addl		$abhi,$hi0,$hi0		; account carry bit
+	fstds		${fnm1},8($xfer)
+	 add		$ablo,$nmlo0,$nmlo0	; discarded
+	ldw		0($tp),$ti1		; tp[1]
+	 addc		%r0,$nmhi0,$hi1
+	fstds		${fab0},-16($xfer)
+	fstds		${fnm0},-8($xfer)
+	ldw		4($xfer),$ablo
+	ldw		0($xfer),$abhi
+
+L\$inner_pa11
+	xmpyu		${fai}R,${fbi},${fab1}	; ap[j+1]*bp[i]
+	flddx		$idx($ap),${fai}	; ap[j,j+1]
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[j+1]*m
+	flddx		$idx($np),${fni}	; np[j,j+1]
+	 add		$hi0,$ablo,$ablo
+	ldw		4($tp),$ti0		; tp[j]
+	 addc		%r0,$abhi,$abhi
+	ldw		12($xfer),$nmlo1
+	 add		$ti1,$ablo,$ablo
+	ldw		8($xfer),$nmhi1
+	 addc		%r0,$abhi,$hi0
+	fstds		${fab1},0($xfer)
+	 add		$ablo,$nmlo1,$nmlo1
+	fstds		${fnm1},8($xfer)
+	 addc		%r0,$nmhi1,$nmhi1
+	ldw		-12($xfer),$ablo
+	 add		$hi1,$nmlo1,$nmlo1
+	ldw		-16($xfer),$abhi
+	 addc		%r0,$nmhi1,$hi1
+
+	xmpyu		${fai}L,${fbi},${fab0}	; ap[j]*bp[i]
+	ldw		8($tp),$ti1		; tp[j]
+	xmpyu		${fni}L,${fm0}R,${fnm0}	; np[j]*m
+	ldw		-4($xfer),$nmlo0
+	 add		$hi0,$ablo,$ablo
+	ldw		-8($xfer),$nmhi0
+	 addc		%r0,$abhi,$abhi
+	stw		$nmlo1,-4($tp)		; tp[j-1]
+	 add		$ti0,$ablo,$ablo
+	fstds		${fab0},-16($xfer)
+	 addc		%r0,$abhi,$hi0
+	fstds		${fnm0},-8($xfer)
+	 add		$ablo,$nmlo0,$nmlo0
+	ldw		4($xfer),$ablo
+	 addc		%r0,$nmhi0,$nmhi0
+	ldw		0($xfer),$abhi
+	 add		$hi1,$nmlo0,$nmlo0
+	 stws,ma	$nmlo0,8($tp)		; tp[j-1]
+	addib,<>	8,$idx,L\$inner_pa11	; j++++
+	 addc		%r0,$nmhi0,$hi1
+
+	xmpyu		${fai}R,${fbi},${fab1}	; ap[j]*bp[i]
+	ldw		12($xfer),$nmlo1
+	xmpyu		${fni}R,${fm0}R,${fnm1}	; np[j]*m
+	ldw		8($xfer),$nmhi1
+	 add		$hi0,$ablo,$ablo
+	ldw		4($tp),$ti0		; tp[j]
+	 addc		%r0,$abhi,$abhi
+	fstds		${fab1},0($xfer)
+	 add		$ti1,$ablo,$ablo
+	fstds		${fnm1},8($xfer)
+	 addc		%r0,$abhi,$hi0
+	ldw		-16($xfer),$abhi
+	 add		$ablo,$nmlo1,$nmlo1
+	ldw		-12($xfer),$ablo
+	 addc		%r0,$nmhi1,$nmhi1
+	ldw		-8($xfer),$nmhi0
+	 add		$hi1,$nmlo1,$nmlo1
+	ldw		-4($xfer),$nmlo0
+	 addc		%r0,$nmhi1,$hi1
+
+	add		$hi0,$ablo,$ablo
+	 stw		$nmlo1,-4($tp)		; tp[j-1]
+	addc		%r0,$abhi,$abhi
+	 add		$ti0,$ablo,$ablo
+	ldw		8($tp),$ti1		; tp[j]
+	 addc		%r0,$abhi,$hi0
+	ldw		0($xfer),$abhi
+	 add		$ablo,$nmlo0,$nmlo0
+	ldw		4($xfer),$ablo
+	 addc		%r0,$nmhi0,$nmhi0
+	ldws,mb		8($xfer),$nmhi1
+	 add		$hi1,$nmlo0,$nmlo0
+	ldw		4($xfer),$nmlo1
+	 addc		%r0,$nmhi0,$hi1
+	 stws,ma	$nmlo0,8($tp)		; tp[j-1]
+
+	addib,=		-1,$num,L\$outerdone_pa11; i--
+	subi		0,$arrsz,$idx		; j=0
+
+	 fldws,ma	4($bp),${fbi}		; bp[i]
+	 flddx		$idx($ap),${fai}	; ap[0]
+	add		$hi0,$ablo,$ablo
+	addc		%r0,$abhi,$abhi
+	 flddx		$idx($np),${fni}	; np[0]
+	 fldws		8($xfer),${fti}R	; tp[0]
+	add		$ti1,$ablo,$ablo
+	addc		%r0,$abhi,$hi0
+
+	 ldo		8($idx),$idx		; j++++
+	 xmpyu		${fai}L,${fbi},${fab0}	; ap[0]*bp[i]
+	 xmpyu		${fai}R,${fbi},${fab1}	; ap[1]*bp[i]
+	ldw		4($tp),$ti0		; tp[j]
+
+	add		$hi1,$nmlo1,$nmlo1
+	addc		%r0,$nmhi1,$nmhi1
+	 fstws,mb	${fab0}L,-8($xfer)	; save high part
+	add		$ablo,$nmlo1,$nmlo1
+	addc		%r0,$nmhi1,$hi1
+	 fcpy,sgl	%fr0,${fti}L		; zero high part
+	 fcpy,sgl	%fr0,${fab0}L
+	stw		$nmlo1,-4($tp)		; tp[j-1]
+
+	 fcnvxf,dbl,dbl	${fti},${fti}		; 32-bit unsigned int -> double
+	 fcnvxf,dbl,dbl	${fab0},${fab0}
+	add		$hi1,$hi0,$hi0
+	addc		%r0,%r0,$hi1
+	 fadd,dbl	${fti},${fab0},${fab0}	; add tp[0]
+	add		$ti0,$hi0,$hi0
+	addc		%r0,$hi1,$hi1
+	 fcnvfx,dbl,dbl	${fab0},${fab0}		; double -> 33-bit unsigned int
+	stw		$hi0,0($tp)
+	stw		$hi1,4($tp)
+	 xmpyu		${fn0},${fab0}R,${fm0}
+
+	b		L\$outer_pa11
+	ldo		`$LOCALS+32+4`($fp),$tp
+
+L\$outerdone_pa11
+	add		$hi0,$ablo,$ablo
+	addc		%r0,$abhi,$abhi
+	add		$ti1,$ablo,$ablo
+	addc		%r0,$abhi,$hi0
+
+	ldw		4($tp),$ti0		; tp[j]
+
+	add		$hi1,$nmlo1,$nmlo1
+	addc		%r0,$nmhi1,$nmhi1
+	add		$ablo,$nmlo1,$nmlo1
+	addc		%r0,$nmhi1,$hi1
+	stw		$nmlo1,-4($tp)		; tp[j-1]
+
+	add		$hi1,$hi0,$hi0
+	addc		%r0,%r0,$hi1
+	add		$ti0,$hi0,$hi0
+	addc		%r0,$hi1,$hi1
+	stw		$hi0,0($tp)
+	stw		$hi1,4($tp)
+
+	ldo		`$LOCALS+32+4`($fp),$tp
+	sub		%r0,%r0,%r0		; clear borrow
+	ldw		-4($tp),$ti0
+	addl		$tp,$arrsz,$tp
+L\$sub_pa11
+	ldwx		$idx($np),$hi0
+	subb		$ti0,$hi0,$hi1
+	ldwx		$idx($tp),$ti0
+	addib,<>	4,$idx,L\$sub_pa11
+	stws,ma		$hi1,4($rp)
+
+	subb		$ti0,%r0,$hi1
+	ldo		-4($tp),$tp
+	and		$tp,$hi1,$ap
+	andcm		$rp,$hi1,$bp
+	or		$ap,$bp,$np
+
+	sub		$rp,$arrsz,$rp		; rewind rp
+	subi		0,$arrsz,$idx
+	ldo		`$LOCALS+32`($fp),$tp
+L\$copy_pa11
+	ldwx		$idx($np),$hi0
+	stws,ma		%r0,4($tp)
+	addib,<>	4,$idx,L\$copy_pa11
+	stws,ma		$hi0,4($rp)	
+
+	nop					; alignment
+L\$done
+___
+}
+
+$code.=<<___;
+	ldi		1,%r28			; signal "handled"
+	ldo		$FRAME($fp),%sp		; destroy tp[num+1]
+
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2	; standard epilogue
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
+	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
+	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
+	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
+L\$abort
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+	.STRINGZ "Montgomery Multiplication for PA-RISC, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+# Explicitly encode PA-RISC 2.0 instructions used in this module, so
+# that it can be compiled with .LEVEL 1.0. It should be noted that I
+# wouldn't have to do this, if GNU assembler understood .ALLOW 2.0
+# directive...
+
+my $ldd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "ldd$mod\t$args";
+
+    if ($args =~ /%r([0-9]+)\(%r([0-9]+)\),%r([0-9]+)/)		# format 4
+    {	my $opcode=(0x03<<26)|($2<<21)|($1<<16)|(3<<6)|$3;
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    elsif ($args =~ /(\-?[0-9]+)\(%r([0-9]+)\),%r([0-9]+)/)	# format 5
+    {	my $opcode=(0x03<<26)|($2<<21)|(1<<12)|(3<<6)|$3;
+	$opcode|=(($1&0xF)<<17)|(($1&0x10)<<12);		# encode offset
+	$opcode|=(1<<5)  if ($mod =~ /^,m/);
+	$opcode|=(1<<13) if ($mod =~ /^,mb/);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $std = sub {
+  my ($mod,$args) = @_;
+  my $orig = "std$mod\t$args";
+
+    if ($args =~ /%r([0-9]+),(\-?[0-9]+)\(%r([0-9]+)\)/)	# format 6
+    {	my $opcode=(0x03<<26)|($3<<21)|($1<<16)|(1<<12)|(0xB<<6);
+	$opcode|=(($2&0xF)<<1)|(($2&0x10)>>4);			# encode offset
+	$opcode|=(1<<5)  if ($mod =~ /^,m/);
+	$opcode|=(1<<13) if ($mod =~ /^,mb/);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $extrd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "extrd$mod\t$args";
+
+    # I only have ",u" completer, it's implicitly encoded...
+    if ($args =~ /%r([0-9]+),([0-9]+),([0-9]+),%r([0-9]+)/)	# format 15
+    {	my $opcode=(0x36<<26)|($1<<21)|($4<<16);
+	my $len=32-$3;
+	$opcode |= (($2&0x20)<<6)|(($2&0x1f)<<5);		# encode pos
+	$opcode |= (($len&0x20)<<7)|($len&0x1f);		# encode len
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    elsif ($args =~ /%r([0-9]+),%sar,([0-9]+),%r([0-9]+)/)	# format 12
+    {	my $opcode=(0x34<<26)|($1<<21)|($3<<16)|(2<<11)|(1<<9);
+	my $len=32-$2;
+	$opcode |= (($len&0x20)<<3)|($len&0x1f);		# encode len
+	$opcode |= (1<<13) if ($mod =~ /,\**=/);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $shrpd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "shrpd$mod\t$args";
+
+    if ($args =~ /%r([0-9]+),%r([0-9]+),([0-9]+),%r([0-9]+)/)	# format 14
+    {	my $opcode=(0x34<<26)|($2<<21)|($1<<16)|(1<<10)|$4;
+	my $cpos=63-$3;
+	$opcode |= (($cpos&0x20)<<6)|(($cpos&0x1f)<<5);		# encode sa
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $sub = sub {
+  my ($mod,$args) = @_;
+  my $orig = "sub$mod\t$args";
+
+    if ($mod eq ",db" && $args =~ /%r([0-9]+),%r([0-9]+),%r([0-9]+)/) {
+	my $opcode=(0x02<<26)|($2<<21)|($1<<16)|$3;
+	$opcode|=(1<<10);	# e1
+	$opcode|=(1<<8);	# e2
+	$opcode|=(1<<5);	# d
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig
+    }
+    else { "\t".$orig; }
+};
+
+sub assemble {
+  my ($mnemonic,$mod,$args)=@_;
+  my $opcode = eval("\$$mnemonic");
+
+    ref($opcode) eq 'CODE' ? &$opcode($mod,$args) : "\t$mnemonic$mod\t$args";
+}
+
+foreach (split("\n",$code)) {
+	s/\`([^\`]*)\`/eval $1/ge;
+	# flip word order in 64-bit mode...
+	s/(xmpyu\s+)($fai|$fni)([LR])/$1.$2.($3 eq "L"?"R":"L")/e if ($BN_SZ==8);
+	# assemble 2.0 instructions in 32-bit mode...
+	s/^\s+([a-z]+)([\S]*)\s+([\S]*)/&assemble($1,$2,$3)/e if ($BN_SZ==4);
+
+	print $_,"\n";
+}
+close STDOUT;
diff --git a/crypto/bn/asm/ppc-mont.pl b/crypto/bn/asm/ppc-mont.pl
new file mode 100755
index 000000000000..f9b6992ccc82
--- /dev/null
+++ b/crypto/bn/asm/ppc-mont.pl
@@ -0,0 +1,334 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# April 2006
+
+# "Teaser" Montgomery multiplication module for PowerPC. It's possible
+# to gain a bit more by modulo-scheduling outer loop, then dedicated
+# squaring procedure should give further 20% and code can be adapted
+# for 32-bit application running on 64-bit CPU. As for the latter.
+# It won't be able to achieve "native" 64-bit performance, because in
+# 32-bit application context every addc instruction will have to be
+# expanded as addc, twice right shift by 32 and finally adde, etc.
+# So far RSA *sign* performance improvement over pre-bn_mul_mont asm
+# for 64-bit application running on PPC970/G5 is:
+#
+# 512-bit	+65%	
+# 1024-bit	+35%
+# 2048-bit	+18%
+# 4096-bit	+4%
+
+$flavour = shift;
+
+if ($flavour =~ /32/) {
+	$BITS=	32;
+	$BNSZ=	$BITS/8;
+	$SIZE_T=4;
+	$RZONE=	224;
+
+	$LD=	"lwz";		# load
+	$LDU=	"lwzu";		# load and update
+	$LDX=	"lwzx";		# load indexed
+	$ST=	"stw";		# store
+	$STU=	"stwu";		# store and update
+	$STX=	"stwx";		# store indexed
+	$STUX=	"stwux";	# store indexed and update
+	$UMULL=	"mullw";	# unsigned multiply low
+	$UMULH=	"mulhwu";	# unsigned multiply high
+	$UCMP=	"cmplw";	# unsigned compare
+	$SHRI=	"srwi";		# unsigned shift right by immediate	
+	$PUSH=	$ST;
+	$POP=	$LD;
+} elsif ($flavour =~ /64/) {
+	$BITS=	64;
+	$BNSZ=	$BITS/8;
+	$SIZE_T=8;
+	$RZONE=	288;
+
+	# same as above, but 64-bit mnemonics...
+	$LD=	"ld";		# load
+	$LDU=	"ldu";		# load and update
+	$LDX=	"ldx";		# load indexed
+	$ST=	"std";		# store
+	$STU=	"stdu";		# store and update
+	$STX=	"stdx";		# store indexed
+	$STUX=	"stdux";	# store indexed and update
+	$UMULL=	"mulld";	# unsigned multiply low
+	$UMULH=	"mulhdu";	# unsigned multiply high
+	$UCMP=	"cmpld";	# unsigned compare
+	$SHRI=	"srdi";		# unsigned shift right by immediate	
+	$PUSH=	$ST;
+	$POP=	$LD;
+} else { die "nonsense $flavour"; }
+
+$FRAME=8*$SIZE_T+$RZONE;
+$LOCALS=8*$SIZE_T;
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour ".shift || die "can't call $xlate: $!";
+
+$sp="r1";
+$toc="r2";
+$rp="r3";	$ovf="r3";
+$ap="r4";
+$bp="r5";
+$np="r6";
+$n0="r7";
+$num="r8";
+$rp="r9";	# $rp is reassigned
+$aj="r10";
+$nj="r11";
+$tj="r12";
+# non-volatile registers
+$i="r20";
+$j="r21";
+$tp="r22";
+$m0="r23";
+$m1="r24";
+$lo0="r25";
+$hi0="r26";
+$lo1="r27";
+$hi1="r28";
+$alo="r29";
+$ahi="r30";
+$nlo="r31";
+#
+$nhi="r0";
+
+$code=<<___;
+.machine "any"
+.text
+
+.globl	.bn_mul_mont_int
+.align	4
+.bn_mul_mont_int:
+	cmpwi	$num,4
+	mr	$rp,r3		; $rp is reassigned
+	li	r3,0
+	bltlr
+___
+$code.=<<___ if ($BNSZ==4);
+	cmpwi	$num,32		; longer key performance is not better
+	bgelr
+___
+$code.=<<___;
+	slwi	$num,$num,`log($BNSZ)/log(2)`
+	li	$tj,-4096
+	addi	$ovf,$num,$FRAME
+	subf	$ovf,$ovf,$sp	; $sp-$ovf
+	and	$ovf,$ovf,$tj	; minimize TLB usage
+	subf	$ovf,$sp,$ovf	; $ovf-$sp
+	mr	$tj,$sp
+	srwi	$num,$num,`log($BNSZ)/log(2)`
+	$STUX	$sp,$sp,$ovf
+
+	$PUSH	r20,`-12*$SIZE_T`($tj)
+	$PUSH	r21,`-11*$SIZE_T`($tj)
+	$PUSH	r22,`-10*$SIZE_T`($tj)
+	$PUSH	r23,`-9*$SIZE_T`($tj)
+	$PUSH	r24,`-8*$SIZE_T`($tj)
+	$PUSH	r25,`-7*$SIZE_T`($tj)
+	$PUSH	r26,`-6*$SIZE_T`($tj)
+	$PUSH	r27,`-5*$SIZE_T`($tj)
+	$PUSH	r28,`-4*$SIZE_T`($tj)
+	$PUSH	r29,`-3*$SIZE_T`($tj)
+	$PUSH	r30,`-2*$SIZE_T`($tj)
+	$PUSH	r31,`-1*$SIZE_T`($tj)
+
+	$LD	$n0,0($n0)	; pull n0[0] value
+	addi	$num,$num,-2	; adjust $num for counter register
+
+	$LD	$m0,0($bp)	; m0=bp[0]
+	$LD	$aj,0($ap)	; ap[0]
+	addi	$tp,$sp,$LOCALS
+	$UMULL	$lo0,$aj,$m0	; ap[0]*bp[0]
+	$UMULH	$hi0,$aj,$m0
+
+	$LD	$aj,$BNSZ($ap)	; ap[1]
+	$LD	$nj,0($np)	; np[0]
+
+	$UMULL	$m1,$lo0,$n0	; "tp[0]"*n0
+
+	$UMULL	$alo,$aj,$m0	; ap[1]*bp[0]
+	$UMULH	$ahi,$aj,$m0
+
+	$UMULL	$lo1,$nj,$m1	; np[0]*m1
+	$UMULH	$hi1,$nj,$m1
+	$LD	$nj,$BNSZ($np)	; np[1]
+	addc	$lo1,$lo1,$lo0
+	addze	$hi1,$hi1
+
+	$UMULL	$nlo,$nj,$m1	; np[1]*m1
+	$UMULH	$nhi,$nj,$m1
+
+	mtctr	$num
+	li	$j,`2*$BNSZ`
+.align	4
+L1st:
+	$LDX	$aj,$ap,$j	; ap[j]
+	addc	$lo0,$alo,$hi0
+	$LDX	$nj,$np,$j	; np[j]
+	addze	$hi0,$ahi
+	$UMULL	$alo,$aj,$m0	; ap[j]*bp[0]
+	addc	$lo1,$nlo,$hi1
+	$UMULH	$ahi,$aj,$m0
+	addze	$hi1,$nhi
+	$UMULL	$nlo,$nj,$m1	; np[j]*m1
+	addc	$lo1,$lo1,$lo0	; np[j]*m1+ap[j]*bp[0]
+	$UMULH	$nhi,$nj,$m1
+	addze	$hi1,$hi1
+	$ST	$lo1,0($tp)	; tp[j-1]
+
+	addi	$j,$j,$BNSZ	; j++
+	addi	$tp,$tp,$BNSZ	; tp++
+	bdnz-	L1st
+;L1st
+	addc	$lo0,$alo,$hi0
+	addze	$hi0,$ahi
+
+	addc	$lo1,$nlo,$hi1
+	addze	$hi1,$nhi
+	addc	$lo1,$lo1,$lo0	; np[j]*m1+ap[j]*bp[0]
+	addze	$hi1,$hi1
+	$ST	$lo1,0($tp)	; tp[j-1]
+
+	li	$ovf,0
+	addc	$hi1,$hi1,$hi0
+	addze	$ovf,$ovf	; upmost overflow bit
+	$ST	$hi1,$BNSZ($tp)
+
+	li	$i,$BNSZ
+.align	4
+Louter:
+	$LDX	$m0,$bp,$i	; m0=bp[i]
+	$LD	$aj,0($ap)	; ap[0]
+	addi	$tp,$sp,$LOCALS
+	$LD	$tj,$LOCALS($sp); tp[0]
+	$UMULL	$lo0,$aj,$m0	; ap[0]*bp[i]
+	$UMULH	$hi0,$aj,$m0
+	$LD	$aj,$BNSZ($ap)	; ap[1]
+	$LD	$nj,0($np)	; np[0]
+	addc	$lo0,$lo0,$tj	; ap[0]*bp[i]+tp[0]
+	$UMULL	$alo,$aj,$m0	; ap[j]*bp[i]
+	addze	$hi0,$hi0
+	$UMULL	$m1,$lo0,$n0	; tp[0]*n0
+	$UMULH	$ahi,$aj,$m0
+	$UMULL	$lo1,$nj,$m1	; np[0]*m1
+	$UMULH	$hi1,$nj,$m1
+	$LD	$nj,$BNSZ($np)	; np[1]
+	addc	$lo1,$lo1,$lo0
+	$UMULL	$nlo,$nj,$m1	; np[1]*m1
+	addze	$hi1,$hi1
+	$UMULH	$nhi,$nj,$m1
+
+	mtctr	$num
+	li	$j,`2*$BNSZ`
+.align	4
+Linner:
+	$LDX	$aj,$ap,$j	; ap[j]
+	addc	$lo0,$alo,$hi0
+	$LD	$tj,$BNSZ($tp)	; tp[j]
+	addze	$hi0,$ahi
+	$LDX	$nj,$np,$j	; np[j]
+	addc	$lo1,$nlo,$hi1
+	$UMULL	$alo,$aj,$m0	; ap[j]*bp[i]
+	addze	$hi1,$nhi
+	$UMULH	$ahi,$aj,$m0
+	addc	$lo0,$lo0,$tj	; ap[j]*bp[i]+tp[j]
+	$UMULL	$nlo,$nj,$m1	; np[j]*m1
+	addze	$hi0,$hi0
+	$UMULH	$nhi,$nj,$m1
+	addc	$lo1,$lo1,$lo0	; np[j]*m1+ap[j]*bp[i]+tp[j]
+	addi	$j,$j,$BNSZ	; j++
+	addze	$hi1,$hi1
+	$ST	$lo1,0($tp)	; tp[j-1]
+	addi	$tp,$tp,$BNSZ	; tp++
+	bdnz-	Linner
+;Linner
+	$LD	$tj,$BNSZ($tp)	; tp[j]
+	addc	$lo0,$alo,$hi0
+	addze	$hi0,$ahi
+	addc	$lo0,$lo0,$tj	; ap[j]*bp[i]+tp[j]
+	addze	$hi0,$hi0
+
+	addc	$lo1,$nlo,$hi1
+	addze	$hi1,$nhi
+	addc	$lo1,$lo1,$lo0	; np[j]*m1+ap[j]*bp[i]+tp[j]
+	addze	$hi1,$hi1
+	$ST	$lo1,0($tp)	; tp[j-1]
+
+	addic	$ovf,$ovf,-1	; move upmost overflow to XER[CA]
+	li	$ovf,0
+	adde	$hi1,$hi1,$hi0
+	addze	$ovf,$ovf
+	$ST	$hi1,$BNSZ($tp)
+;
+	slwi	$tj,$num,`log($BNSZ)/log(2)`
+	$UCMP	$i,$tj
+	addi	$i,$i,$BNSZ
+	ble-	Louter
+
+	addi	$num,$num,2	; restore $num
+	subfc	$j,$j,$j	; j=0 and "clear" XER[CA]
+	addi	$tp,$sp,$LOCALS
+	mtctr	$num
+
+.align	4
+Lsub:	$LDX	$tj,$tp,$j
+	$LDX	$nj,$np,$j
+	subfe	$aj,$nj,$tj	; tp[j]-np[j]
+	$STX	$aj,$rp,$j
+	addi	$j,$j,$BNSZ
+	bdnz-	Lsub
+
+	li	$j,0
+	mtctr	$num
+	subfe	$ovf,$j,$ovf	; handle upmost overflow bit
+	and	$ap,$tp,$ovf
+	andc	$np,$rp,$ovf
+	or	$ap,$ap,$np	; ap=borrow?tp:rp
+
+.align	4
+Lcopy:				; copy or in-place refresh
+	$LDX	$tj,$ap,$j
+	$STX	$tj,$rp,$j
+	$STX	$j,$tp,$j	; zap at once
+	addi	$j,$j,$BNSZ
+	bdnz-	Lcopy
+
+	$POP	$tj,0($sp)
+	li	r3,1
+	$POP	r20,`-12*$SIZE_T`($tj)
+	$POP	r21,`-11*$SIZE_T`($tj)
+	$POP	r22,`-10*$SIZE_T`($tj)
+	$POP	r23,`-9*$SIZE_T`($tj)
+	$POP	r24,`-8*$SIZE_T`($tj)
+	$POP	r25,`-7*$SIZE_T`($tj)
+	$POP	r26,`-6*$SIZE_T`($tj)
+	$POP	r27,`-5*$SIZE_T`($tj)
+	$POP	r28,`-4*$SIZE_T`($tj)
+	$POP	r29,`-3*$SIZE_T`($tj)
+	$POP	r30,`-2*$SIZE_T`($tj)
+	$POP	r31,`-1*$SIZE_T`($tj)
+	mr	$sp,$tj
+	blr
+	.long	0
+	.byte	0,12,4,0,0x80,12,6,0
+	.long	0
+
+.asciz  "Montgomery Multiplication for PPC, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/ppc.pl b/crypto/bn/asm/ppc.pl
index 84448836e3dd..1249ce229988 100644
--- a/crypto/bn/asm/ppc.pl
+++ b/crypto/bn/asm/ppc.pl
@@ -100,9 +100,9 @@
 #	me a note at schari@us.ibm.com
 #
 
-$opf = shift;
+$flavour = shift;
 
-if ($opf =~ /32\.s/) {
+if ($flavour =~ /32/) {
 	$BITS=	32;
 	$BNSZ=	$BITS/8;
 	$ISA=	"\"ppc\"";
@@ -125,7 +125,7 @@ if ($opf =~ /32\.s/) {
 	$INSR=	"insrwi";	# insert right
 	$ROTL=	"rotlwi";	# rotate left by immediate
 	$TR=	"tw";		# conditional trap
-} elsif ($opf =~ /64\.s/) {
+} elsif ($flavour =~ /64/) {
 	$BITS=	64;
 	$BNSZ=	$BITS/8;
 	$ISA=	"\"ppc64\"";
@@ -149,93 +149,16 @@ if ($opf =~ /32\.s/) {
 	$INSR=	"insrdi";	# insert right 
 	$ROTL=	"rotldi";	# rotate left by immediate
 	$TR=	"td";		# conditional trap
-} else { die "nonsense $opf"; }
+} else { die "nonsense $flavour"; }
 
-( defined shift || open STDOUT,">$opf" ) || die "can't open $opf: $!";
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
 
-# function entry points from the AIX code
-#
-# There are other, more elegant, ways to handle this. We (IBM) chose
-# this approach as it plays well with scripts we run to 'namespace'
-# OpenSSL .i.e. we add a prefix to all the public symbols so we can
-# co-exist in the same process with other implementations of OpenSSL.
-# 'cleverer' ways of doing these substitutions tend to hide data we
-# need to be obvious.
-#
-my @items = ("bn_sqr_comba4",
-	     "bn_sqr_comba8",
-	     "bn_mul_comba4",
-	     "bn_mul_comba8",
-	     "bn_sub_words",
-	     "bn_add_words",
-	     "bn_div_words",
-	     "bn_sqr_words",
-	     "bn_mul_words",
-	     "bn_mul_add_words");
+open STDOUT,"| $^X $xlate $flavour ".shift || die "can't call $xlate: $!";
 
-if    ($opf =~ /linux/)	{  do_linux();	}
-elsif ($opf =~ /aix/)	{  do_aix();	}
-elsif ($opf =~ /osx/)	{  do_osx();	}
-else			{  do_bsd();	}
-
-sub do_linux {
-    $d=&data();
-
-    if ($BITS==64) {
-      foreach $t (@items) {
-        $d =~ s/\.$t:/\
-\t.section\t".opd","aw"\
-\t.align\t3\
-\t.globl\t$t\
-$t:\
-\t.quad\t.$t,.TOC.\@tocbase,0\
-\t.size\t$t,24\
-\t.previous\n\
-\t.type\t.$t,\@function\
-\t.globl\t.$t\
-.$t:/g;
-      }
-    }
-    else {
-      foreach $t (@items) {
-        $d=~s/\.$t/$t/g;
-      }
-    }
-    # hide internal labels to avoid pollution of name table...
-    $d=~s/Lppcasm_/.Lppcasm_/gm;
-    print $d;
-}
-
-sub do_aix {
-    # AIX assembler is smart enough to please the linker without
-    # making us do something special...
-    print &data();
-}
-
-# MacOSX 32 bit
-sub do_osx {
-    $d=&data();
-    # Change the bn symbol prefix from '.' to '_'
-    foreach $t (@items) {
-      $d=~s/\.$t/_$t/g;
-    }
-    # Change .machine to something OS X asm will accept
-    $d=~s/\.machine.*/.text/g;
-    $d=~s/\#/;/g; # change comment from '#' to ';'
-    print $d;
-}
-
-# BSD (Untested)
-sub do_bsd {
-    $d=&data();
-    foreach $t (@items) {
-      $d=~s/\.$t/_$t/g;
-    }
-    print $d;
-}
-
-sub data {
-	local($data)=<<EOF;
+$data=<<EOF;
 #--------------------------------------------------------------------
 #
 #
@@ -297,33 +220,20 @@ sub data {
 #
 #	Defines to be used in the assembly code.
 #	
-.set r0,0	# we use it as storage for value of 0
-.set SP,1	# preserved
-.set RTOC,2	# preserved 
-.set r3,3	# 1st argument/return value
-.set r4,4	# 2nd argument/volatile register
-.set r5,5	# 3rd argument/volatile register
-.set r6,6	# ...
-.set r7,7
-.set r8,8
-.set r9,9
-.set r10,10
-.set r11,11
-.set r12,12
-.set r13,13	# not used, nor any other "below" it...
-
-.set BO_IF_NOT,4
-.set BO_IF,12
-.set BO_dCTR_NZERO,16
-.set BO_dCTR_ZERO,18
-.set BO_ALWAYS,20
-.set CR0_LT,0;
-.set CR0_GT,1;
-.set CR0_EQ,2
-.set CR1_FX,4;
-.set CR1_FEX,5;
-.set CR1_VX,6
-.set LR,8
+#.set r0,0	# we use it as storage for value of 0
+#.set SP,1	# preserved
+#.set RTOC,2	# preserved 
+#.set r3,3	# 1st argument/return value
+#.set r4,4	# 2nd argument/volatile register
+#.set r5,5	# 3rd argument/volatile register
+#.set r6,6	# ...
+#.set r7,7
+#.set r8,8
+#.set r9,9
+#.set r10,10
+#.set r11,11
+#.set r12,12
+#.set r13,13	# not used, nor any other "below" it...
 
 #	Declare function names to be global
 #	NOTE:	For gcc these names MUST be changed to remove
@@ -344,7 +254,7 @@ sub data {
 	
 # .text section
 	
-	.machine	$ISA
+	.machine	"any"
 
 #
 #	NOTE:	The following label name should be changed to
@@ -478,8 +388,10 @@ sub data {
 
 	$ST		r9,`6*$BNSZ`(r3)	#r[6]=c1
 	$ST		r10,`7*$BNSZ`(r3)	#r[7]=c2
-	bclr	BO_ALWAYS,CR0_LT
-	.long	0x00000000
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,2,0
+	.long	0
 
 #
 #	NOTE:	The following label name should be changed to
@@ -903,9 +815,10 @@ sub data {
 	$ST		r9, `15*$BNSZ`(r3)	#r[15]=c1;
 
 
-	bclr	BO_ALWAYS,CR0_LT
-
-	.long	0x00000000
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,2,0
+	.long	0
 
 #
 #	NOTE:	The following label name should be changed to
@@ -1055,8 +968,10 @@ sub data {
 
 	$ST	r10,`6*$BNSZ`(r3)	#r[6]=c1
 	$ST	r11,`7*$BNSZ`(r3)	#r[7]=c2
-	bclr	BO_ALWAYS,CR0_LT
-	.long	0x00000000
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,3,0
+	.long	0
 
 #
 #	NOTE:	The following label name should be changed to
@@ -1591,8 +1506,10 @@ sub data {
 	adde	r10,r10,r9
 	$ST	r12,`14*$BNSZ`(r3)	#r[14]=c3;
 	$ST	r10,`15*$BNSZ`(r3)	#r[15]=c1;
-	bclr	BO_ALWAYS,CR0_LT
-	.long	0x00000000
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,3,0
+	.long	0
 
 #
 #	NOTE:	The following label name should be changed to
@@ -1623,7 +1540,7 @@ sub data {
 	subfc.	r7,r0,r6        # If r6 is 0 then result is 0.
 				# if r6 > 0 then result !=0
 				# In either case carry bit is set.
-	bc	BO_IF,CR0_EQ,Lppcasm_sub_adios
+	beq	Lppcasm_sub_adios
 	addi	r4,r4,-$BNSZ
 	addi	r3,r3,-$BNSZ
 	addi	r5,r5,-$BNSZ
@@ -1635,13 +1552,14 @@ Lppcasm_sub_mainloop:
 				# if carry = 1 this is r7-r8. Else it
 				# is r7-r8 -1 as we need.
 	$STU	r6,$BNSZ(r3)
-	bc	BO_dCTR_NZERO,CR0_EQ,Lppcasm_sub_mainloop
+	bdnz-	Lppcasm_sub_mainloop
 Lppcasm_sub_adios:	
 	subfze	r3,r0		# if carry bit is set then r3 = 0 else -1
 	andi.	r3,r3,1         # keep only last bit.
-	bclr	BO_ALWAYS,CR0_LT
-	.long	0x00000000
-
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,4,0
+	.long	0
 
 #
 #	NOTE:	The following label name should be changed to
@@ -1670,7 +1588,7 @@ Lppcasm_sub_adios:
 #	check for r6 = 0. Is this needed?
 #
 	addic.	r6,r6,0		#test r6 and clear carry bit.
-	bc	BO_IF,CR0_EQ,Lppcasm_add_adios
+	beq	Lppcasm_add_adios
 	addi	r4,r4,-$BNSZ
 	addi	r3,r3,-$BNSZ
 	addi	r5,r5,-$BNSZ
@@ -1680,11 +1598,13 @@ Lppcasm_add_mainloop:
 	$LDU	r8,$BNSZ(r5)
 	adde	r8,r7,r8
 	$STU	r8,$BNSZ(r3)
-	bc	BO_dCTR_NZERO,CR0_EQ,Lppcasm_add_mainloop
+	bdnz-	Lppcasm_add_mainloop
 Lppcasm_add_adios:	
 	addze	r3,r0			#return carry bit.
-	bclr	BO_ALWAYS,CR0_LT
-	.long	0x00000000
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,4,0
+	.long	0
 
 #
 #	NOTE:	The following label name should be changed to
@@ -1707,24 +1627,24 @@ Lppcasm_add_adios:
 #	r5 = d
 	
 	$UCMPI	0,r5,0			# compare r5 and 0
-	bc	BO_IF_NOT,CR0_EQ,Lppcasm_div1	# proceed if d!=0
+	bne	Lppcasm_div1		# proceed if d!=0
 	li	r3,-1			# d=0 return -1
-	bclr	BO_ALWAYS,CR0_LT	
+	blr
 Lppcasm_div1:
 	xor	r0,r0,r0		#r0=0
 	li	r8,$BITS
 	$CNTLZ.	r7,r5			#r7 = num leading 0s in d.
-	bc	BO_IF,CR0_EQ,Lppcasm_div2	#proceed if no leading zeros
+	beq	Lppcasm_div2		#proceed if no leading zeros
 	subf	r8,r7,r8		#r8 = BN_num_bits_word(d)
 	$SHR.	r9,r3,r8		#are there any bits above r8'th?
 	$TR	16,r9,r0		#if there're, signal to dump core...
 Lppcasm_div2:
 	$UCMP	0,r3,r5			#h>=d?
-	bc	BO_IF,CR0_LT,Lppcasm_div3	#goto Lppcasm_div3 if not
+	blt	Lppcasm_div3		#goto Lppcasm_div3 if not
 	subf	r3,r5,r3		#h-=d ; 
 Lppcasm_div3:				#r7 = BN_BITS2-i. so r7=i
 	cmpi	0,0,r7,0		# is (i == 0)?
-	bc	BO_IF,CR0_EQ,Lppcasm_div4
+	beq	Lppcasm_div4
 	$SHL	r3,r3,r7		# h = (h<< i)
 	$SHR	r8,r4,r8		# r8 = (l >> BN_BITS2 -i)
 	$SHL	r5,r5,r7		# d<<=i
@@ -1741,7 +1661,7 @@ Lppcasm_divouterloop:
 	$SHRI	r11,r4,`$BITS/2`	#r11= (l&BN_MASK2h)>>BN_BITS4
 					# compute here for innerloop.
 	$UCMP	0,r8,r9			# is (h>>BN_BITS4)==dh
-	bc	BO_IF_NOT,CR0_EQ,Lppcasm_div5	# goto Lppcasm_div5 if not
+	bne	Lppcasm_div5		# goto Lppcasm_div5 if not
 
 	li	r8,-1
 	$CLRU	r8,r8,`$BITS/2`		#q = BN_MASK2l 
@@ -1762,9 +1682,9 @@ Lppcasm_divinnerloop:
 					# the following 2 instructions do that
 	$SHLI	r7,r10,`$BITS/2`	# r7 = (t<<BN_BITS4)
 	or	r7,r7,r11		# r7|=((l&BN_MASK2h)>>BN_BITS4)
-	$UCMP	1,r6,r7			# compare (tl <= r7)
-	bc	BO_IF_NOT,CR0_EQ,Lppcasm_divinnerexit
-	bc	BO_IF_NOT,CR1_FEX,Lppcasm_divinnerexit
+	$UCMP	cr1,r6,r7		# compare (tl <= r7)
+	bne	Lppcasm_divinnerexit
+	ble	cr1,Lppcasm_divinnerexit
 	addi	r8,r8,-1		#q--
 	subf	r12,r9,r12		#th -=dh
 	$CLRU	r10,r5,`$BITS/2`	#r10=dl. t is no longer needed in loop.
@@ -1773,14 +1693,14 @@ Lppcasm_divinnerloop:
 Lppcasm_divinnerexit:
 	$SHRI	r10,r6,`$BITS/2`	#t=(tl>>BN_BITS4)
 	$SHLI	r11,r6,`$BITS/2`	#tl=(tl<<BN_BITS4)&BN_MASK2h;
-	$UCMP	1,r4,r11		# compare l and tl
+	$UCMP	cr1,r4,r11		# compare l and tl
 	add	r12,r12,r10		# th+=t
-	bc	BO_IF_NOT,CR1_FX,Lppcasm_div7  # if (l>=tl) goto Lppcasm_div7
+	bge	cr1,Lppcasm_div7	# if (l>=tl) goto Lppcasm_div7
 	addi	r12,r12,1		# th++
 Lppcasm_div7:
 	subf	r11,r11,r4		#r11=l-tl
-	$UCMP	1,r3,r12		#compare h and th
-	bc	BO_IF_NOT,CR1_FX,Lppcasm_div8	#if (h>=th) goto Lppcasm_div8
+	$UCMP	cr1,r3,r12		#compare h and th
+	bge	cr1,Lppcasm_div8	#if (h>=th) goto Lppcasm_div8
 	addi	r8,r8,-1		# q--
 	add	r3,r5,r3		# h+=d
 Lppcasm_div8:
@@ -1791,13 +1711,15 @@ Lppcasm_div8:
 					# the following 2 instructions will do this.
 	$INSR	r11,r12,`$BITS/2`,`$BITS/2`	# r11 is the value we want rotated $BITS/2.
 	$ROTL	r3,r11,`$BITS/2`	# rotate by $BITS/2 and store in r3
-	bc	BO_dCTR_ZERO,CR0_EQ,Lppcasm_div9#if (count==0) break ;
+	bdz	Lppcasm_div9		#if (count==0) break ;
 	$SHLI	r0,r8,`$BITS/2`		#ret =q<<BN_BITS4
 	b	Lppcasm_divouterloop
 Lppcasm_div9:
 	or	r3,r8,r0
-	bclr	BO_ALWAYS,CR0_LT
-	.long	0x00000000
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,3,0
+	.long	0
 
 #
 #	NOTE:	The following label name should be changed to
@@ -1822,7 +1744,7 @@ Lppcasm_div9:
 #	No unrolling done here. Not performance critical.
 
 	addic.	r5,r5,0			#test r5.
-	bc	BO_IF,CR0_EQ,Lppcasm_sqr_adios
+	beq	Lppcasm_sqr_adios
 	addi	r4,r4,-$BNSZ
 	addi	r3,r3,-$BNSZ
 	mtctr	r5
@@ -1833,11 +1755,12 @@ Lppcasm_sqr_mainloop:
 	$UMULH  r8,r6,r6
 	$STU	r7,$BNSZ(r3)
 	$STU	r8,$BNSZ(r3)
-	bc	BO_dCTR_NZERO,CR0_EQ,Lppcasm_sqr_mainloop
+	bdnz-	Lppcasm_sqr_mainloop
 Lppcasm_sqr_adios:	
-	bclr	BO_ALWAYS,CR0_LT
-	.long	0x00000000
-
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,3,0
+	.long	0
 
 #
 #	NOTE:	The following label name should be changed to
@@ -1858,7 +1781,7 @@ Lppcasm_sqr_adios:
 	xor	r0,r0,r0
 	xor	r12,r12,r12		# used for carry
 	rlwinm.	r7,r5,30,2,31		# num >> 2
-	bc	BO_IF,CR0_EQ,Lppcasm_mw_REM
+	beq	Lppcasm_mw_REM
 	mtctr	r7
 Lppcasm_mw_LOOP:	
 					#mul(rp[0],ap[0],w,c1);
@@ -1896,11 +1819,11 @@ Lppcasm_mw_LOOP:
 	
 	addi	r3,r3,`4*$BNSZ`
 	addi	r4,r4,`4*$BNSZ`
-	bc	BO_dCTR_NZERO,CR0_EQ,Lppcasm_mw_LOOP
+	bdnz-	Lppcasm_mw_LOOP
 
 Lppcasm_mw_REM:
 	andi.	r5,r5,0x3
-	bc	BO_IF,CR0_EQ,Lppcasm_mw_OVER
+	beq	Lppcasm_mw_OVER
 					#mul(rp[0],ap[0],w,c1);
 	$LD	r8,`0*$BNSZ`(r4)
 	$UMULL	r9,r6,r8
@@ -1912,7 +1835,7 @@ Lppcasm_mw_REM:
 	
 	addi	r5,r5,-1
 	cmpli	0,0,r5,0
-	bc	BO_IF,CR0_EQ,Lppcasm_mw_OVER
+	beq	Lppcasm_mw_OVER
 
 	
 					#mul(rp[1],ap[1],w,c1);
@@ -1926,7 +1849,7 @@ Lppcasm_mw_REM:
 	
 	addi	r5,r5,-1
 	cmpli	0,0,r5,0
-	bc	BO_IF,CR0_EQ,Lppcasm_mw_OVER
+	beq	Lppcasm_mw_OVER
 	
 					#mul_add(rp[2],ap[2],w,c1);
 	$LD	r8,`2*$BNSZ`(r4)
@@ -1939,8 +1862,10 @@ Lppcasm_mw_REM:
 		
 Lppcasm_mw_OVER:	
 	addi	r3,r12,0
-	bclr	BO_ALWAYS,CR0_LT
-	.long	0x00000000
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,4,0
+	.long	0
 
 #
 #	NOTE:	The following label name should be changed to
@@ -1964,7 +1889,7 @@ Lppcasm_mw_OVER:
 	xor	r0,r0,r0		#r0 = 0
 	xor	r12,r12,r12  		#r12 = 0 . used for carry		
 	rlwinm.	r7,r5,30,2,31		# num >> 2
-	bc	BO_IF,CR0_EQ,Lppcasm_maw_leftover	# if (num < 4) go LPPCASM_maw_leftover
+	beq	Lppcasm_maw_leftover	# if (num < 4) go LPPCASM_maw_leftover
 	mtctr	r7
 Lppcasm_maw_mainloop:	
 					#mul_add(rp[0],ap[0],w,c1);
@@ -2017,11 +1942,11 @@ Lppcasm_maw_mainloop:
 	$ST	r11,`3*$BNSZ`(r3)
 	addi	r3,r3,`4*$BNSZ`
 	addi	r4,r4,`4*$BNSZ`
-	bc	BO_dCTR_NZERO,CR0_EQ,Lppcasm_maw_mainloop
+	bdnz-	Lppcasm_maw_mainloop
 	
 Lppcasm_maw_leftover:
 	andi.	r5,r5,0x3
-	bc	BO_IF,CR0_EQ,Lppcasm_maw_adios
+	beq	Lppcasm_maw_adios
 	addi	r3,r3,-$BNSZ
 	addi	r4,r4,-$BNSZ
 					#mul_add(rp[0],ap[0],w,c1);
@@ -2036,7 +1961,7 @@ Lppcasm_maw_leftover:
 	addze	r12,r10
 	$ST	r9,0(r3)
 	
-	bc	BO_dCTR_ZERO,CR0_EQ,Lppcasm_maw_adios
+	bdz	Lppcasm_maw_adios
 					#mul_add(rp[1],ap[1],w,c1);
 	$LDU	r8,$BNSZ(r4)	
 	$UMULL	r9,r6,r8
@@ -2048,7 +1973,7 @@ Lppcasm_maw_leftover:
 	addze	r12,r10
 	$ST	r9,0(r3)
 	
-	bc	BO_dCTR_ZERO,CR0_EQ,Lppcasm_maw_adios
+	bdz	Lppcasm_maw_adios
 					#mul_add(rp[2],ap[2],w,c1);
 	$LDU	r8,$BNSZ(r4)
 	$UMULL	r9,r6,r8
@@ -2062,19 +1987,12 @@ Lppcasm_maw_leftover:
 		
 Lppcasm_maw_adios:	
 	addi	r3,r12,0
-	bclr	BO_ALWAYS,CR0_LT
-	.long	0x00000000
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,4,0
+	.long	0
 	.align	4
 EOF
-	$data =~ s/\`([^\`]*)\`/eval $1/gem;
-
-	# if some assembler chokes on some simplified mnemonic,
-	# this is the spot to fix it up, e.g.:
-	# GNU as doesn't seem to accept cmplw, 32-bit unsigned compare
-	$data =~ s/^(\s*)cmplw(\s+)([^,]+),(.*)/$1cmpl$2$3,0,$4/gm;
-	# assembler X doesn't accept li, load immediate value
-	#$data =~ s/^(\s*)li(\s+)([^,]+),(.*)/$1addi$2$3,0,$4/gm;
-	# assembler Y chokes on apostrophes in comments
-	$data =~ s/'//gm;
-	return($data);
-}
+$data =~ s/\`([^\`]*)\`/eval $1/gem;
+print $data;
+close STDOUT;
diff --git a/crypto/bn/asm/ppc64-mont.pl b/crypto/bn/asm/ppc64-mont.pl
new file mode 100755
index 000000000000..a14e769ad055
--- /dev/null
+++ b/crypto/bn/asm/ppc64-mont.pl
@@ -0,0 +1,1088 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# December 2007
+
+# The reason for undertaken effort is basically following. Even though
+# Power 6 CPU operates at incredible 4.7GHz clock frequency, its PKI
+# performance was observed to be less than impressive, essentially as
+# fast as 1.8GHz PPC970, or 2.6 times(!) slower than one would hope.
+# Well, it's not surprising that IBM had to make some sacrifices to
+# boost the clock frequency that much, but no overall improvement?
+# Having observed how much difference did switching to FPU make on
+# UltraSPARC, playing same stunt on Power 6 appeared appropriate...
+# Unfortunately the resulting performance improvement is not as
+# impressive, ~30%, and in absolute terms is still very far from what
+# one would expect from 4.7GHz CPU. There is a chance that I'm doing
+# something wrong, but in the lack of assembler level micro-profiling
+# data or at least decent platform guide I can't tell... Or better
+# results might be achieved with VMX... Anyway, this module provides
+# *worse* performance on other PowerPC implementations, ~40-15% slower
+# on PPC970 depending on key length and ~40% slower on Power 5 for all
+# key lengths. As it's obviously inappropriate as "best all-round"
+# alternative, it has to be complemented with run-time CPU family
+# detection. Oh! It should also be noted that unlike other PowerPC
+# implementation IALU ppc-mont.pl module performs *suboptimaly* on
+# >=1024-bit key lengths on Power 6. It should also be noted that
+# *everything* said so far applies to 64-bit builds! As far as 32-bit
+# application executed on 64-bit CPU goes, this module is likely to
+# become preferred choice, because it's easy to adapt it for such
+# case and *is* faster than 32-bit ppc-mont.pl on *all* processors.
+
+# February 2008
+
+# Micro-profiling assisted optimization results in ~15% improvement
+# over original ppc64-mont.pl version, or overall ~50% improvement
+# over ppc.pl module on Power 6. If compared to ppc-mont.pl on same
+# Power 6 CPU, this module is 5-150% faster depending on key length,
+# [hereafter] more for longer keys. But if compared to ppc-mont.pl
+# on 1.8GHz PPC970, it's only 5-55% faster. Still far from impressive
+# in absolute terms, but it's apparently the way Power 6 is...
+
+# December 2009
+
+# Adapted for 32-bit build this module delivers 25-120%, yes, more
+# than *twice* for longer keys, performance improvement over 32-bit
+# ppc-mont.pl on 1.8GHz PPC970. However! This implementation utilizes
+# even 64-bit integer operations and the trouble is that most PPC
+# operating systems don't preserve upper halves of general purpose
+# registers upon 32-bit signal delivery. They do preserve them upon
+# context switch, but not signalling:-( This means that asynchronous
+# signals have to be blocked upon entry to this subroutine. Signal
+# masking (and of course complementary unmasking) has quite an impact
+# on performance, naturally larger for shorter keys. It's so severe
+# that 512-bit key performance can be as low as 1/3 of expected one.
+# This is why this routine can be engaged for longer key operations
+# only on these OSes, see crypto/ppccap.c for further details. MacOS X
+# is an exception from this and doesn't require signal masking, and
+# that's where above improvement coefficients were collected. For
+# others alternative would be to break dependence on upper halves of
+# GPRs by sticking to 32-bit integer operations...
+
+$flavour = shift;
+
+if ($flavour =~ /32/) {
+	$SIZE_T=4;
+	$RZONE=	224;
+	$fname=	"bn_mul_mont_fpu64";
+
+	$STUX=	"stwux";	# store indexed and update
+	$PUSH=	"stw";
+	$POP=	"lwz";
+} elsif ($flavour =~ /64/) {
+	$SIZE_T=8;
+	$RZONE=	288;
+	$fname=	"bn_mul_mont_fpu64";
+
+	# same as above, but 64-bit mnemonics...
+	$STUX=	"stdux";	# store indexed and update
+	$PUSH=	"std";
+	$POP=	"ld";
+} else { die "nonsense $flavour"; }
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour ".shift || die "can't call $xlate: $!";
+
+$FRAME=64;	# padded frame header
+$TRANSFER=16*8;
+
+$carry="r0";
+$sp="r1";
+$toc="r2";
+$rp="r3";	$ovf="r3";
+$ap="r4";
+$bp="r5";
+$np="r6";
+$n0="r7";
+$num="r8";
+$rp="r9";	# $rp is reassigned
+$tp="r10";
+$j="r11";
+$i="r12";
+# non-volatile registers
+$nap_d="r22";	# interleaved ap and np in double format
+$a0="r23";	# ap[0]
+$t0="r24";	# temporary registers
+$t1="r25";
+$t2="r26";
+$t3="r27";
+$t4="r28";
+$t5="r29";
+$t6="r30";
+$t7="r31";
+
+# PPC offers enough register bank capacity to unroll inner loops twice
+#
+#     ..A3A2A1A0
+#           dcba
+#    -----------
+#            A0a
+#           A0b
+#          A0c
+#         A0d
+#          A1a
+#         A1b
+#        A1c
+#       A1d
+#        A2a
+#       A2b
+#      A2c
+#     A2d
+#      A3a
+#     A3b
+#    A3c
+#   A3d
+#    ..a
+#   ..b
+#
+$ba="f0";	$bb="f1";	$bc="f2";	$bd="f3";
+$na="f4";	$nb="f5";	$nc="f6";	$nd="f7";
+$dota="f8";	$dotb="f9";
+$A0="f10";	$A1="f11";	$A2="f12";	$A3="f13";
+$N0="f20";	$N1="f21";	$N2="f22";	$N3="f23";
+$T0a="f24";	$T0b="f25";
+$T1a="f26";	$T1b="f27";
+$T2a="f28";	$T2b="f29";
+$T3a="f30";	$T3b="f31";
+
+# sp----------->+-------------------------------+
+#		| saved sp			|
+#		+-------------------------------+
+#		.				.
+#   +64		+-------------------------------+
+#		| 16 gpr<->fpr transfer zone	|
+#		.				.
+#		.				.
+#   +16*8	+-------------------------------+
+#		| __int64 tmp[-1]		|
+#		+-------------------------------+
+#		| __int64 tmp[num]		|
+#		.				.
+#		.				.
+#		.				.
+#   +(num+1)*8	+-------------------------------+
+#		| padding to 64 byte boundary	|
+#		.				.
+#   +X		+-------------------------------+
+#		| double nap_d[4*num]		|
+#		.				.
+#		.				.
+#		.				.
+#		+-------------------------------+
+#		.				.
+#   -12*size_t	+-------------------------------+
+#		| 10 saved gpr, r22-r31		|
+#		.				.
+#		.				.
+#   -12*8	+-------------------------------+
+#		| 12 saved fpr, f20-f31		|
+#		.				.
+#		.				.
+#		+-------------------------------+
+
+$code=<<___;
+.machine "any"
+.text
+
+.globl	.$fname
+.align	5
+.$fname:
+	cmpwi	$num,`3*8/$SIZE_T`
+	mr	$rp,r3		; $rp is reassigned
+	li	r3,0		; possible "not handled" return code
+	bltlr-
+	andi.	r0,$num,`16/$SIZE_T-1`		; $num has to be "even"
+	bnelr-
+
+	slwi	$num,$num,`log($SIZE_T)/log(2)`	; num*=sizeof(BN_LONG)
+	li	$i,-4096
+	slwi	$tp,$num,2	; place for {an}p_{lh}[num], i.e. 4*num
+	add	$tp,$tp,$num	; place for tp[num+1]
+	addi	$tp,$tp,`$FRAME+$TRANSFER+8+64+$RZONE`
+	subf	$tp,$tp,$sp	; $sp-$tp
+	and	$tp,$tp,$i	; minimize TLB usage
+	subf	$tp,$sp,$tp	; $tp-$sp
+	mr	$i,$sp
+	$STUX	$sp,$sp,$tp	; alloca
+
+	$PUSH	r22,`-12*8-10*$SIZE_T`($i)
+	$PUSH	r23,`-12*8-9*$SIZE_T`($i)
+	$PUSH	r24,`-12*8-8*$SIZE_T`($i)
+	$PUSH	r25,`-12*8-7*$SIZE_T`($i)
+	$PUSH	r26,`-12*8-6*$SIZE_T`($i)
+	$PUSH	r27,`-12*8-5*$SIZE_T`($i)
+	$PUSH	r28,`-12*8-4*$SIZE_T`($i)
+	$PUSH	r29,`-12*8-3*$SIZE_T`($i)
+	$PUSH	r30,`-12*8-2*$SIZE_T`($i)
+	$PUSH	r31,`-12*8-1*$SIZE_T`($i)
+	stfd	f20,`-12*8`($i)
+	stfd	f21,`-11*8`($i)
+	stfd	f22,`-10*8`($i)
+	stfd	f23,`-9*8`($i)
+	stfd	f24,`-8*8`($i)
+	stfd	f25,`-7*8`($i)
+	stfd	f26,`-6*8`($i)
+	stfd	f27,`-5*8`($i)
+	stfd	f28,`-4*8`($i)
+	stfd	f29,`-3*8`($i)
+	stfd	f30,`-2*8`($i)
+	stfd	f31,`-1*8`($i)
+___
+$code.=<<___ if ($SIZE_T==8);
+	ld	$a0,0($ap)	; pull ap[0] value
+	ld	$n0,0($n0)	; pull n0[0] value
+	ld	$t3,0($bp)	; bp[0]
+___
+$code.=<<___ if ($SIZE_T==4);
+	mr	$t1,$n0
+	lwz	$a0,0($ap)	; pull ap[0,1] value
+	lwz	$t0,4($ap)
+	lwz	$n0,0($t1)	; pull n0[0,1] value
+	lwz	$t1,4($t1)
+	lwz	$t3,0($bp)	; bp[0,1]
+	lwz	$t2,4($bp)
+	insrdi	$a0,$t0,32,0
+	insrdi	$n0,$t1,32,0
+	insrdi	$t3,$t2,32,0
+___
+$code.=<<___;
+	addi	$tp,$sp,`$FRAME+$TRANSFER+8+64`
+	li	$i,-64
+	add	$nap_d,$tp,$num
+	and	$nap_d,$nap_d,$i	; align to 64 bytes
+
+	mulld	$t7,$a0,$t3	; ap[0]*bp[0]
+	; nap_d is off by 1, because it's used with stfdu/lfdu
+	addi	$nap_d,$nap_d,-8
+	srwi	$j,$num,`3+1`	; counter register, num/2
+	mulld	$t7,$t7,$n0	; tp[0]*n0
+	addi	$j,$j,-1
+	addi	$tp,$sp,`$FRAME+$TRANSFER-8`
+	li	$carry,0
+	mtctr	$j
+
+	; transfer bp[0] to FPU as 4x16-bit values
+	extrdi	$t0,$t3,16,48
+	extrdi	$t1,$t3,16,32
+	extrdi	$t2,$t3,16,16
+	extrdi	$t3,$t3,16,0
+	std	$t0,`$FRAME+0`($sp)
+	std	$t1,`$FRAME+8`($sp)
+	std	$t2,`$FRAME+16`($sp)
+	std	$t3,`$FRAME+24`($sp)
+	; transfer (ap[0]*bp[0])*n0 to FPU as 4x16-bit values
+	extrdi	$t4,$t7,16,48
+	extrdi	$t5,$t7,16,32
+	extrdi	$t6,$t7,16,16
+	extrdi	$t7,$t7,16,0
+	std	$t4,`$FRAME+32`($sp)
+	std	$t5,`$FRAME+40`($sp)
+	std	$t6,`$FRAME+48`($sp)
+	std	$t7,`$FRAME+56`($sp)
+___
+$code.=<<___ if ($SIZE_T==8);
+	lwz	$t0,4($ap)		; load a[j] as 32-bit word pair
+	lwz	$t1,0($ap)
+	lwz	$t2,12($ap)		; load a[j+1] as 32-bit word pair
+	lwz	$t3,8($ap)
+	lwz	$t4,4($np)		; load n[j] as 32-bit word pair
+	lwz	$t5,0($np)
+	lwz	$t6,12($np)		; load n[j+1] as 32-bit word pair
+	lwz	$t7,8($np)
+___
+$code.=<<___ if ($SIZE_T==4);
+	lwz	$t0,0($ap)		; load a[j..j+3] as 32-bit word pairs
+	lwz	$t1,4($ap)
+	lwz	$t2,8($ap)
+	lwz	$t3,12($ap)
+	lwz	$t4,0($np)		; load n[j..j+3] as 32-bit word pairs
+	lwz	$t5,4($np)
+	lwz	$t6,8($np)
+	lwz	$t7,12($np)
+___
+$code.=<<___;
+	lfd	$ba,`$FRAME+0`($sp)
+	lfd	$bb,`$FRAME+8`($sp)
+	lfd	$bc,`$FRAME+16`($sp)
+	lfd	$bd,`$FRAME+24`($sp)
+	lfd	$na,`$FRAME+32`($sp)
+	lfd	$nb,`$FRAME+40`($sp)
+	lfd	$nc,`$FRAME+48`($sp)
+	lfd	$nd,`$FRAME+56`($sp)
+	std	$t0,`$FRAME+64`($sp)
+	std	$t1,`$FRAME+72`($sp)
+	std	$t2,`$FRAME+80`($sp)
+	std	$t3,`$FRAME+88`($sp)
+	std	$t4,`$FRAME+96`($sp)
+	std	$t5,`$FRAME+104`($sp)
+	std	$t6,`$FRAME+112`($sp)
+	std	$t7,`$FRAME+120`($sp)
+	fcfid	$ba,$ba
+	fcfid	$bb,$bb
+	fcfid	$bc,$bc
+	fcfid	$bd,$bd
+	fcfid	$na,$na
+	fcfid	$nb,$nb
+	fcfid	$nc,$nc
+	fcfid	$nd,$nd
+
+	lfd	$A0,`$FRAME+64`($sp)
+	lfd	$A1,`$FRAME+72`($sp)
+	lfd	$A2,`$FRAME+80`($sp)
+	lfd	$A3,`$FRAME+88`($sp)
+	lfd	$N0,`$FRAME+96`($sp)
+	lfd	$N1,`$FRAME+104`($sp)
+	lfd	$N2,`$FRAME+112`($sp)
+	lfd	$N3,`$FRAME+120`($sp)
+	fcfid	$A0,$A0
+	fcfid	$A1,$A1
+	fcfid	$A2,$A2
+	fcfid	$A3,$A3
+	fcfid	$N0,$N0
+	fcfid	$N1,$N1
+	fcfid	$N2,$N2
+	fcfid	$N3,$N3
+	addi	$ap,$ap,16
+	addi	$np,$np,16
+
+	fmul	$T1a,$A1,$ba
+	fmul	$T1b,$A1,$bb
+	stfd	$A0,8($nap_d)		; save a[j] in double format
+	stfd	$A1,16($nap_d)
+	fmul	$T2a,$A2,$ba
+	fmul	$T2b,$A2,$bb
+	stfd	$A2,24($nap_d)		; save a[j+1] in double format
+	stfd	$A3,32($nap_d)
+	fmul	$T3a,$A3,$ba
+	fmul	$T3b,$A3,$bb
+	stfd	$N0,40($nap_d)		; save n[j] in double format
+	stfd	$N1,48($nap_d)
+	fmul	$T0a,$A0,$ba
+	fmul	$T0b,$A0,$bb
+	stfd	$N2,56($nap_d)		; save n[j+1] in double format
+	stfdu	$N3,64($nap_d)
+
+	fmadd	$T1a,$A0,$bc,$T1a
+	fmadd	$T1b,$A0,$bd,$T1b
+	fmadd	$T2a,$A1,$bc,$T2a
+	fmadd	$T2b,$A1,$bd,$T2b
+	fmadd	$T3a,$A2,$bc,$T3a
+	fmadd	$T3b,$A2,$bd,$T3b
+	fmul	$dota,$A3,$bc
+	fmul	$dotb,$A3,$bd
+
+	fmadd	$T1a,$N1,$na,$T1a
+	fmadd	$T1b,$N1,$nb,$T1b
+	fmadd	$T2a,$N2,$na,$T2a
+	fmadd	$T2b,$N2,$nb,$T2b
+	fmadd	$T3a,$N3,$na,$T3a
+	fmadd	$T3b,$N3,$nb,$T3b
+	fmadd	$T0a,$N0,$na,$T0a
+	fmadd	$T0b,$N0,$nb,$T0b
+
+	fmadd	$T1a,$N0,$nc,$T1a
+	fmadd	$T1b,$N0,$nd,$T1b
+	fmadd	$T2a,$N1,$nc,$T2a
+	fmadd	$T2b,$N1,$nd,$T2b
+	fmadd	$T3a,$N2,$nc,$T3a
+	fmadd	$T3b,$N2,$nd,$T3b
+	fmadd	$dota,$N3,$nc,$dota
+	fmadd	$dotb,$N3,$nd,$dotb
+
+	fctid	$T0a,$T0a
+	fctid	$T0b,$T0b
+	fctid	$T1a,$T1a
+	fctid	$T1b,$T1b
+	fctid	$T2a,$T2a
+	fctid	$T2b,$T2b
+	fctid	$T3a,$T3a
+	fctid	$T3b,$T3b
+
+	stfd	$T0a,`$FRAME+0`($sp)
+	stfd	$T0b,`$FRAME+8`($sp)
+	stfd	$T1a,`$FRAME+16`($sp)
+	stfd	$T1b,`$FRAME+24`($sp)
+	stfd	$T2a,`$FRAME+32`($sp)
+	stfd	$T2b,`$FRAME+40`($sp)
+	stfd	$T3a,`$FRAME+48`($sp)
+	stfd	$T3b,`$FRAME+56`($sp)
+
+.align	5
+L1st:
+___
+$code.=<<___ if ($SIZE_T==8);
+	lwz	$t0,4($ap)		; load a[j] as 32-bit word pair
+	lwz	$t1,0($ap)
+	lwz	$t2,12($ap)		; load a[j+1] as 32-bit word pair
+	lwz	$t3,8($ap)
+	lwz	$t4,4($np)		; load n[j] as 32-bit word pair
+	lwz	$t5,0($np)
+	lwz	$t6,12($np)		; load n[j+1] as 32-bit word pair
+	lwz	$t7,8($np)
+___
+$code.=<<___ if ($SIZE_T==4);
+	lwz	$t0,0($ap)		; load a[j..j+3] as 32-bit word pairs
+	lwz	$t1,4($ap)
+	lwz	$t2,8($ap)
+	lwz	$t3,12($ap)
+	lwz	$t4,0($np)		; load n[j..j+3] as 32-bit word pairs
+	lwz	$t5,4($np)
+	lwz	$t6,8($np)
+	lwz	$t7,12($np)
+___
+$code.=<<___;
+	std	$t0,`$FRAME+64`($sp)
+	std	$t1,`$FRAME+72`($sp)
+	std	$t2,`$FRAME+80`($sp)
+	std	$t3,`$FRAME+88`($sp)
+	std	$t4,`$FRAME+96`($sp)
+	std	$t5,`$FRAME+104`($sp)
+	std	$t6,`$FRAME+112`($sp)
+	std	$t7,`$FRAME+120`($sp)
+	ld	$t0,`$FRAME+0`($sp)
+	ld	$t1,`$FRAME+8`($sp)
+	ld	$t2,`$FRAME+16`($sp)
+	ld	$t3,`$FRAME+24`($sp)
+	ld	$t4,`$FRAME+32`($sp)
+	ld	$t5,`$FRAME+40`($sp)
+	ld	$t6,`$FRAME+48`($sp)
+	ld	$t7,`$FRAME+56`($sp)
+	lfd	$A0,`$FRAME+64`($sp)
+	lfd	$A1,`$FRAME+72`($sp)
+	lfd	$A2,`$FRAME+80`($sp)
+	lfd	$A3,`$FRAME+88`($sp)
+	lfd	$N0,`$FRAME+96`($sp)
+	lfd	$N1,`$FRAME+104`($sp)
+	lfd	$N2,`$FRAME+112`($sp)
+	lfd	$N3,`$FRAME+120`($sp)
+	fcfid	$A0,$A0
+	fcfid	$A1,$A1
+	fcfid	$A2,$A2
+	fcfid	$A3,$A3
+	fcfid	$N0,$N0
+	fcfid	$N1,$N1
+	fcfid	$N2,$N2
+	fcfid	$N3,$N3
+	addi	$ap,$ap,16
+	addi	$np,$np,16
+
+	fmul	$T1a,$A1,$ba
+	fmul	$T1b,$A1,$bb
+	fmul	$T2a,$A2,$ba
+	fmul	$T2b,$A2,$bb
+	stfd	$A0,8($nap_d)		; save a[j] in double format
+	stfd	$A1,16($nap_d)
+	fmul	$T3a,$A3,$ba
+	fmul	$T3b,$A3,$bb
+	fmadd	$T0a,$A0,$ba,$dota
+	fmadd	$T0b,$A0,$bb,$dotb
+	stfd	$A2,24($nap_d)		; save a[j+1] in double format
+	stfd	$A3,32($nap_d)
+
+	fmadd	$T1a,$A0,$bc,$T1a
+	fmadd	$T1b,$A0,$bd,$T1b
+	fmadd	$T2a,$A1,$bc,$T2a
+	fmadd	$T2b,$A1,$bd,$T2b
+	stfd	$N0,40($nap_d)		; save n[j] in double format
+	stfd	$N1,48($nap_d)
+	fmadd	$T3a,$A2,$bc,$T3a
+	fmadd	$T3b,$A2,$bd,$T3b
+	 add	$t0,$t0,$carry		; can not overflow
+	fmul	$dota,$A3,$bc
+	fmul	$dotb,$A3,$bd
+	stfd	$N2,56($nap_d)		; save n[j+1] in double format
+	stfdu	$N3,64($nap_d)
+	 srdi	$carry,$t0,16
+	 add	$t1,$t1,$carry
+	 srdi	$carry,$t1,16
+
+	fmadd	$T1a,$N1,$na,$T1a
+	fmadd	$T1b,$N1,$nb,$T1b
+	 insrdi	$t0,$t1,16,32
+	fmadd	$T2a,$N2,$na,$T2a
+	fmadd	$T2b,$N2,$nb,$T2b
+	 add	$t2,$t2,$carry
+	fmadd	$T3a,$N3,$na,$T3a
+	fmadd	$T3b,$N3,$nb,$T3b
+	 srdi	$carry,$t2,16
+	fmadd	$T0a,$N0,$na,$T0a
+	fmadd	$T0b,$N0,$nb,$T0b
+	 insrdi	$t0,$t2,16,16
+	 add	$t3,$t3,$carry
+	 srdi	$carry,$t3,16
+
+	fmadd	$T1a,$N0,$nc,$T1a
+	fmadd	$T1b,$N0,$nd,$T1b
+	 insrdi	$t0,$t3,16,0		; 0..63 bits
+	fmadd	$T2a,$N1,$nc,$T2a
+	fmadd	$T2b,$N1,$nd,$T2b
+	 add	$t4,$t4,$carry
+	fmadd	$T3a,$N2,$nc,$T3a
+	fmadd	$T3b,$N2,$nd,$T3b
+	 srdi	$carry,$t4,16
+	fmadd	$dota,$N3,$nc,$dota
+	fmadd	$dotb,$N3,$nd,$dotb
+	 add	$t5,$t5,$carry
+	 srdi	$carry,$t5,16
+	 insrdi	$t4,$t5,16,32
+
+	fctid	$T0a,$T0a
+	fctid	$T0b,$T0b
+	 add	$t6,$t6,$carry
+	fctid	$T1a,$T1a
+	fctid	$T1b,$T1b
+	 srdi	$carry,$t6,16
+	fctid	$T2a,$T2a
+	fctid	$T2b,$T2b
+	 insrdi	$t4,$t6,16,16
+	fctid	$T3a,$T3a
+	fctid	$T3b,$T3b
+	 add	$t7,$t7,$carry
+	 insrdi	$t4,$t7,16,0		; 64..127 bits
+	 srdi	$carry,$t7,16		; upper 33 bits
+
+	stfd	$T0a,`$FRAME+0`($sp)
+	stfd	$T0b,`$FRAME+8`($sp)
+	stfd	$T1a,`$FRAME+16`($sp)
+	stfd	$T1b,`$FRAME+24`($sp)
+	stfd	$T2a,`$FRAME+32`($sp)
+	stfd	$T2b,`$FRAME+40`($sp)
+	stfd	$T3a,`$FRAME+48`($sp)
+	stfd	$T3b,`$FRAME+56`($sp)
+	 std	$t0,8($tp)		; tp[j-1]
+	 stdu	$t4,16($tp)		; tp[j]
+	bdnz-	L1st
+
+	fctid	$dota,$dota
+	fctid	$dotb,$dotb
+
+	ld	$t0,`$FRAME+0`($sp)
+	ld	$t1,`$FRAME+8`($sp)
+	ld	$t2,`$FRAME+16`($sp)
+	ld	$t3,`$FRAME+24`($sp)
+	ld	$t4,`$FRAME+32`($sp)
+	ld	$t5,`$FRAME+40`($sp)
+	ld	$t6,`$FRAME+48`($sp)
+	ld	$t7,`$FRAME+56`($sp)
+	stfd	$dota,`$FRAME+64`($sp)
+	stfd	$dotb,`$FRAME+72`($sp)
+
+	add	$t0,$t0,$carry		; can not overflow
+	srdi	$carry,$t0,16
+	add	$t1,$t1,$carry
+	srdi	$carry,$t1,16
+	insrdi	$t0,$t1,16,32
+	add	$t2,$t2,$carry
+	srdi	$carry,$t2,16
+	insrdi	$t0,$t2,16,16
+	add	$t3,$t3,$carry
+	srdi	$carry,$t3,16
+	insrdi	$t0,$t3,16,0		; 0..63 bits
+	add	$t4,$t4,$carry
+	srdi	$carry,$t4,16
+	add	$t5,$t5,$carry
+	srdi	$carry,$t5,16
+	insrdi	$t4,$t5,16,32
+	add	$t6,$t6,$carry
+	srdi	$carry,$t6,16
+	insrdi	$t4,$t6,16,16
+	add	$t7,$t7,$carry
+	insrdi	$t4,$t7,16,0		; 64..127 bits
+	srdi	$carry,$t7,16		; upper 33 bits
+	ld	$t6,`$FRAME+64`($sp)
+	ld	$t7,`$FRAME+72`($sp)
+
+	std	$t0,8($tp)		; tp[j-1]
+	stdu	$t4,16($tp)		; tp[j]
+
+	add	$t6,$t6,$carry		; can not overflow
+	srdi	$carry,$t6,16
+	add	$t7,$t7,$carry
+	insrdi	$t6,$t7,48,0
+	srdi	$ovf,$t7,48
+	std	$t6,8($tp)		; tp[num-1]
+
+	slwi	$t7,$num,2
+	subf	$nap_d,$t7,$nap_d	; rewind pointer
+
+	li	$i,8			; i=1
+.align	5
+Louter:
+___
+$code.=<<___ if ($SIZE_T==8);
+	ldx	$t3,$bp,$i	; bp[i]
+___
+$code.=<<___ if ($SIZE_T==4);
+	add	$t0,$bp,$i
+	lwz	$t3,0($t0)		; bp[i,i+1]
+	lwz	$t0,4($t0)
+	insrdi	$t3,$t0,32,0
+___
+$code.=<<___;
+	ld	$t6,`$FRAME+$TRANSFER+8`($sp)	; tp[0]
+	mulld	$t7,$a0,$t3	; ap[0]*bp[i]
+
+	addi	$tp,$sp,`$FRAME+$TRANSFER`
+	add	$t7,$t7,$t6	; ap[0]*bp[i]+tp[0]
+	li	$carry,0
+	mulld	$t7,$t7,$n0	; tp[0]*n0
+	mtctr	$j
+
+	; transfer bp[i] to FPU as 4x16-bit values
+	extrdi	$t0,$t3,16,48
+	extrdi	$t1,$t3,16,32
+	extrdi	$t2,$t3,16,16
+	extrdi	$t3,$t3,16,0
+	std	$t0,`$FRAME+0`($sp)
+	std	$t1,`$FRAME+8`($sp)
+	std	$t2,`$FRAME+16`($sp)
+	std	$t3,`$FRAME+24`($sp)
+	; transfer (ap[0]*bp[i]+tp[0])*n0 to FPU as 4x16-bit values
+	extrdi	$t4,$t7,16,48
+	extrdi	$t5,$t7,16,32
+	extrdi	$t6,$t7,16,16
+	extrdi	$t7,$t7,16,0
+	std	$t4,`$FRAME+32`($sp)
+	std	$t5,`$FRAME+40`($sp)
+	std	$t6,`$FRAME+48`($sp)
+	std	$t7,`$FRAME+56`($sp)
+
+	lfd	$A0,8($nap_d)		; load a[j] in double format
+	lfd	$A1,16($nap_d)
+	lfd	$A2,24($nap_d)		; load a[j+1] in double format
+	lfd	$A3,32($nap_d)
+	lfd	$N0,40($nap_d)		; load n[j] in double format
+	lfd	$N1,48($nap_d)
+	lfd	$N2,56($nap_d)		; load n[j+1] in double format
+	lfdu	$N3,64($nap_d)
+
+	lfd	$ba,`$FRAME+0`($sp)
+	lfd	$bb,`$FRAME+8`($sp)
+	lfd	$bc,`$FRAME+16`($sp)
+	lfd	$bd,`$FRAME+24`($sp)
+	lfd	$na,`$FRAME+32`($sp)
+	lfd	$nb,`$FRAME+40`($sp)
+	lfd	$nc,`$FRAME+48`($sp)
+	lfd	$nd,`$FRAME+56`($sp)
+
+	fcfid	$ba,$ba
+	fcfid	$bb,$bb
+	fcfid	$bc,$bc
+	fcfid	$bd,$bd
+	fcfid	$na,$na
+	fcfid	$nb,$nb
+	fcfid	$nc,$nc
+	fcfid	$nd,$nd
+
+	fmul	$T1a,$A1,$ba
+	fmul	$T1b,$A1,$bb
+	fmul	$T2a,$A2,$ba
+	fmul	$T2b,$A2,$bb
+	fmul	$T3a,$A3,$ba
+	fmul	$T3b,$A3,$bb
+	fmul	$T0a,$A0,$ba
+	fmul	$T0b,$A0,$bb
+
+	fmadd	$T1a,$A0,$bc,$T1a
+	fmadd	$T1b,$A0,$bd,$T1b
+	fmadd	$T2a,$A1,$bc,$T2a
+	fmadd	$T2b,$A1,$bd,$T2b
+	fmadd	$T3a,$A2,$bc,$T3a
+	fmadd	$T3b,$A2,$bd,$T3b
+	fmul	$dota,$A3,$bc
+	fmul	$dotb,$A3,$bd
+
+	fmadd	$T1a,$N1,$na,$T1a
+	fmadd	$T1b,$N1,$nb,$T1b
+	 lfd	$A0,8($nap_d)		; load a[j] in double format
+	 lfd	$A1,16($nap_d)
+	fmadd	$T2a,$N2,$na,$T2a
+	fmadd	$T2b,$N2,$nb,$T2b
+	 lfd	$A2,24($nap_d)		; load a[j+1] in double format
+	 lfd	$A3,32($nap_d)
+	fmadd	$T3a,$N3,$na,$T3a
+	fmadd	$T3b,$N3,$nb,$T3b
+	fmadd	$T0a,$N0,$na,$T0a
+	fmadd	$T0b,$N0,$nb,$T0b
+
+	fmadd	$T1a,$N0,$nc,$T1a
+	fmadd	$T1b,$N0,$nd,$T1b
+	fmadd	$T2a,$N1,$nc,$T2a
+	fmadd	$T2b,$N1,$nd,$T2b
+	fmadd	$T3a,$N2,$nc,$T3a
+	fmadd	$T3b,$N2,$nd,$T3b
+	fmadd	$dota,$N3,$nc,$dota
+	fmadd	$dotb,$N3,$nd,$dotb
+
+	fctid	$T0a,$T0a
+	fctid	$T0b,$T0b
+	fctid	$T1a,$T1a
+	fctid	$T1b,$T1b
+	fctid	$T2a,$T2a
+	fctid	$T2b,$T2b
+	fctid	$T3a,$T3a
+	fctid	$T3b,$T3b
+
+	stfd	$T0a,`$FRAME+0`($sp)
+	stfd	$T0b,`$FRAME+8`($sp)
+	stfd	$T1a,`$FRAME+16`($sp)
+	stfd	$T1b,`$FRAME+24`($sp)
+	stfd	$T2a,`$FRAME+32`($sp)
+	stfd	$T2b,`$FRAME+40`($sp)
+	stfd	$T3a,`$FRAME+48`($sp)
+	stfd	$T3b,`$FRAME+56`($sp)
+
+.align	5
+Linner:
+	fmul	$T1a,$A1,$ba
+	fmul	$T1b,$A1,$bb
+	fmul	$T2a,$A2,$ba
+	fmul	$T2b,$A2,$bb
+	lfd	$N0,40($nap_d)		; load n[j] in double format
+	lfd	$N1,48($nap_d)
+	fmul	$T3a,$A3,$ba
+	fmul	$T3b,$A3,$bb
+	fmadd	$T0a,$A0,$ba,$dota
+	fmadd	$T0b,$A0,$bb,$dotb
+	lfd	$N2,56($nap_d)		; load n[j+1] in double format
+	lfdu	$N3,64($nap_d)
+
+	fmadd	$T1a,$A0,$bc,$T1a
+	fmadd	$T1b,$A0,$bd,$T1b
+	fmadd	$T2a,$A1,$bc,$T2a
+	fmadd	$T2b,$A1,$bd,$T2b
+	 lfd	$A0,8($nap_d)		; load a[j] in double format
+	 lfd	$A1,16($nap_d)
+	fmadd	$T3a,$A2,$bc,$T3a
+	fmadd	$T3b,$A2,$bd,$T3b
+	fmul	$dota,$A3,$bc
+	fmul	$dotb,$A3,$bd
+	 lfd	$A2,24($nap_d)		; load a[j+1] in double format
+	 lfd	$A3,32($nap_d)
+
+	fmadd	$T1a,$N1,$na,$T1a
+	fmadd	$T1b,$N1,$nb,$T1b
+	 ld	$t0,`$FRAME+0`($sp)
+	 ld	$t1,`$FRAME+8`($sp)
+	fmadd	$T2a,$N2,$na,$T2a
+	fmadd	$T2b,$N2,$nb,$T2b
+	 ld	$t2,`$FRAME+16`($sp)
+	 ld	$t3,`$FRAME+24`($sp)
+	fmadd	$T3a,$N3,$na,$T3a
+	fmadd	$T3b,$N3,$nb,$T3b
+	 add	$t0,$t0,$carry		; can not overflow
+	 ld	$t4,`$FRAME+32`($sp)
+	 ld	$t5,`$FRAME+40`($sp)
+	fmadd	$T0a,$N0,$na,$T0a
+	fmadd	$T0b,$N0,$nb,$T0b
+	 srdi	$carry,$t0,16
+	 add	$t1,$t1,$carry
+	 srdi	$carry,$t1,16
+	 ld	$t6,`$FRAME+48`($sp)
+	 ld	$t7,`$FRAME+56`($sp)
+
+	fmadd	$T1a,$N0,$nc,$T1a
+	fmadd	$T1b,$N0,$nd,$T1b
+	 insrdi	$t0,$t1,16,32
+	 ld	$t1,8($tp)		; tp[j]
+	fmadd	$T2a,$N1,$nc,$T2a
+	fmadd	$T2b,$N1,$nd,$T2b
+	 add	$t2,$t2,$carry
+	fmadd	$T3a,$N2,$nc,$T3a
+	fmadd	$T3b,$N2,$nd,$T3b
+	 srdi	$carry,$t2,16
+	 insrdi	$t0,$t2,16,16
+	fmadd	$dota,$N3,$nc,$dota
+	fmadd	$dotb,$N3,$nd,$dotb
+	 add	$t3,$t3,$carry
+	 ldu	$t2,16($tp)		; tp[j+1]
+	 srdi	$carry,$t3,16
+	 insrdi	$t0,$t3,16,0		; 0..63 bits
+	 add	$t4,$t4,$carry
+
+	fctid	$T0a,$T0a
+	fctid	$T0b,$T0b
+	 srdi	$carry,$t4,16
+	fctid	$T1a,$T1a
+	fctid	$T1b,$T1b
+	 add	$t5,$t5,$carry
+	fctid	$T2a,$T2a
+	fctid	$T2b,$T2b
+	 srdi	$carry,$t5,16
+	 insrdi	$t4,$t5,16,32
+	fctid	$T3a,$T3a
+	fctid	$T3b,$T3b
+	 add	$t6,$t6,$carry
+	 srdi	$carry,$t6,16
+	 insrdi	$t4,$t6,16,16
+
+	stfd	$T0a,`$FRAME+0`($sp)
+	stfd	$T0b,`$FRAME+8`($sp)
+	 add	$t7,$t7,$carry
+	 addc	$t3,$t0,$t1
+___
+$code.=<<___ if ($SIZE_T==4);		# adjust XER[CA]
+	extrdi	$t0,$t0,32,0
+	extrdi	$t1,$t1,32,0
+	adde	$t0,$t0,$t1
+___
+$code.=<<___;
+	stfd	$T1a,`$FRAME+16`($sp)
+	stfd	$T1b,`$FRAME+24`($sp)
+	 insrdi	$t4,$t7,16,0		; 64..127 bits
+	 srdi	$carry,$t7,16		; upper 33 bits
+	stfd	$T2a,`$FRAME+32`($sp)
+	stfd	$T2b,`$FRAME+40`($sp)
+	 adde	$t5,$t4,$t2
+___
+$code.=<<___ if ($SIZE_T==4);		# adjust XER[CA]
+	extrdi	$t4,$t4,32,0
+	extrdi	$t2,$t2,32,0
+	adde	$t4,$t4,$t2
+___
+$code.=<<___;
+	stfd	$T3a,`$FRAME+48`($sp)
+	stfd	$T3b,`$FRAME+56`($sp)
+	 addze	$carry,$carry
+	 std	$t3,-16($tp)		; tp[j-1]
+	 std	$t5,-8($tp)		; tp[j]
+	bdnz-	Linner
+
+	fctid	$dota,$dota
+	fctid	$dotb,$dotb
+	ld	$t0,`$FRAME+0`($sp)
+	ld	$t1,`$FRAME+8`($sp)
+	ld	$t2,`$FRAME+16`($sp)
+	ld	$t3,`$FRAME+24`($sp)
+	ld	$t4,`$FRAME+32`($sp)
+	ld	$t5,`$FRAME+40`($sp)
+	ld	$t6,`$FRAME+48`($sp)
+	ld	$t7,`$FRAME+56`($sp)
+	stfd	$dota,`$FRAME+64`($sp)
+	stfd	$dotb,`$FRAME+72`($sp)
+
+	add	$t0,$t0,$carry		; can not overflow
+	srdi	$carry,$t0,16
+	add	$t1,$t1,$carry
+	srdi	$carry,$t1,16
+	insrdi	$t0,$t1,16,32
+	add	$t2,$t2,$carry
+	ld	$t1,8($tp)		; tp[j]
+	srdi	$carry,$t2,16
+	insrdi	$t0,$t2,16,16
+	add	$t3,$t3,$carry
+	ldu	$t2,16($tp)		; tp[j+1]
+	srdi	$carry,$t3,16
+	insrdi	$t0,$t3,16,0		; 0..63 bits
+	add	$t4,$t4,$carry
+	srdi	$carry,$t4,16
+	add	$t5,$t5,$carry
+	srdi	$carry,$t5,16
+	insrdi	$t4,$t5,16,32
+	add	$t6,$t6,$carry
+	srdi	$carry,$t6,16
+	insrdi	$t4,$t6,16,16
+	add	$t7,$t7,$carry
+	insrdi	$t4,$t7,16,0		; 64..127 bits
+	srdi	$carry,$t7,16		; upper 33 bits
+	ld	$t6,`$FRAME+64`($sp)
+	ld	$t7,`$FRAME+72`($sp)
+
+	addc	$t3,$t0,$t1
+___
+$code.=<<___ if ($SIZE_T==4);		# adjust XER[CA]
+	extrdi	$t0,$t0,32,0
+	extrdi	$t1,$t1,32,0
+	adde	$t0,$t0,$t1
+___
+$code.=<<___;
+	adde	$t5,$t4,$t2
+___
+$code.=<<___ if ($SIZE_T==4);		# adjust XER[CA]
+	extrdi	$t4,$t4,32,0
+	extrdi	$t2,$t2,32,0
+	adde	$t4,$t4,$t2
+___
+$code.=<<___;
+	addze	$carry,$carry
+
+	std	$t3,-16($tp)		; tp[j-1]
+	std	$t5,-8($tp)		; tp[j]
+
+	add	$carry,$carry,$ovf	; comsume upmost overflow
+	add	$t6,$t6,$carry		; can not overflow
+	srdi	$carry,$t6,16
+	add	$t7,$t7,$carry
+	insrdi	$t6,$t7,48,0
+	srdi	$ovf,$t7,48
+	std	$t6,0($tp)		; tp[num-1]
+
+	slwi	$t7,$num,2
+	addi	$i,$i,8
+	subf	$nap_d,$t7,$nap_d	; rewind pointer
+	cmpw	$i,$num
+	blt-	Louter
+___
+
+$code.=<<___ if ($SIZE_T==8);
+	subf	$np,$num,$np	; rewind np
+	addi	$j,$j,1		; restore counter
+	subfc	$i,$i,$i	; j=0 and "clear" XER[CA]
+	addi	$tp,$sp,`$FRAME+$TRANSFER+8`
+	addi	$t4,$sp,`$FRAME+$TRANSFER+16`
+	addi	$t5,$np,8
+	addi	$t6,$rp,8
+	mtctr	$j
+
+.align	4
+Lsub:	ldx	$t0,$tp,$i
+	ldx	$t1,$np,$i
+	ldx	$t2,$t4,$i
+	ldx	$t3,$t5,$i
+	subfe	$t0,$t1,$t0	; tp[j]-np[j]
+	subfe	$t2,$t3,$t2	; tp[j+1]-np[j+1]
+	stdx	$t0,$rp,$i
+	stdx	$t2,$t6,$i
+	addi	$i,$i,16
+	bdnz-	Lsub
+
+	li	$i,0
+	subfe	$ovf,$i,$ovf	; handle upmost overflow bit
+	and	$ap,$tp,$ovf
+	andc	$np,$rp,$ovf
+	or	$ap,$ap,$np	; ap=borrow?tp:rp
+	addi	$t7,$ap,8
+	mtctr	$j
+
+.align	4
+Lcopy:				; copy or in-place refresh
+	ldx	$t0,$ap,$i
+	ldx	$t1,$t7,$i
+	std	$i,8($nap_d)	; zap nap_d
+	std	$i,16($nap_d)
+	std	$i,24($nap_d)
+	std	$i,32($nap_d)
+	std	$i,40($nap_d)
+	std	$i,48($nap_d)
+	std	$i,56($nap_d)
+	stdu	$i,64($nap_d)
+	stdx	$t0,$rp,$i
+	stdx	$t1,$t6,$i
+	stdx	$i,$tp,$i	; zap tp at once
+	stdx	$i,$t4,$i
+	addi	$i,$i,16
+	bdnz-	Lcopy
+___
+$code.=<<___ if ($SIZE_T==4);
+	subf	$np,$num,$np	; rewind np
+	addi	$j,$j,1		; restore counter
+	subfc	$i,$i,$i	; j=0 and "clear" XER[CA]
+	addi	$tp,$sp,`$FRAME+$TRANSFER`
+	addi	$np,$np,-4
+	addi	$rp,$rp,-4
+	addi	$ap,$sp,`$FRAME+$TRANSFER+4`
+	mtctr	$j
+
+.align	4
+Lsub:	ld	$t0,8($tp)	; load tp[j..j+3] in 64-bit word order
+	ldu	$t2,16($tp)
+	lwz	$t4,4($np)	; load np[j..j+3] in 32-bit word order
+	lwz	$t5,8($np)
+	lwz	$t6,12($np)
+	lwzu	$t7,16($np)
+	extrdi	$t1,$t0,32,0
+	extrdi	$t3,$t2,32,0
+	subfe	$t4,$t4,$t0	; tp[j]-np[j]
+	 stw	$t0,4($ap)	; save tp[j..j+3] in 32-bit word order
+	subfe	$t5,$t5,$t1	; tp[j+1]-np[j+1]
+	 stw	$t1,8($ap)
+	subfe	$t6,$t6,$t2	; tp[j+2]-np[j+2]
+	 stw	$t2,12($ap)
+	subfe	$t7,$t7,$t3	; tp[j+3]-np[j+3]
+	 stwu	$t3,16($ap)
+	stw	$t4,4($rp)
+	stw	$t5,8($rp)
+	stw	$t6,12($rp)
+	stwu	$t7,16($rp)
+	bdnz-	Lsub
+
+	li	$i,0
+	subfe	$ovf,$i,$ovf	; handle upmost overflow bit
+	addi	$tp,$sp,`$FRAME+$TRANSFER+4`
+	subf	$rp,$num,$rp	; rewind rp
+	and	$ap,$tp,$ovf
+	andc	$np,$rp,$ovf
+	or	$ap,$ap,$np	; ap=borrow?tp:rp
+	addi	$tp,$sp,`$FRAME+$TRANSFER`
+	mtctr	$j
+
+.align	4
+Lcopy:				; copy or in-place refresh
+	lwz	$t0,4($ap)
+	lwz	$t1,8($ap)
+	lwz	$t2,12($ap)
+	lwzu	$t3,16($ap)
+	std	$i,8($nap_d)	; zap nap_d
+	std	$i,16($nap_d)
+	std	$i,24($nap_d)
+	std	$i,32($nap_d)
+	std	$i,40($nap_d)
+	std	$i,48($nap_d)
+	std	$i,56($nap_d)
+	stdu	$i,64($nap_d)
+	stw	$t0,4($rp)
+	stw	$t1,8($rp)
+	stw	$t2,12($rp)
+	stwu	$t3,16($rp)
+	std	$i,8($tp)	; zap tp at once
+	stdu	$i,16($tp)
+	bdnz-	Lcopy
+___
+
+$code.=<<___;
+	$POP	$i,0($sp)
+	li	r3,1	; signal "handled"
+	$POP	r22,`-12*8-10*$SIZE_T`($i)
+	$POP	r23,`-12*8-9*$SIZE_T`($i)
+	$POP	r24,`-12*8-8*$SIZE_T`($i)
+	$POP	r25,`-12*8-7*$SIZE_T`($i)
+	$POP	r26,`-12*8-6*$SIZE_T`($i)
+	$POP	r27,`-12*8-5*$SIZE_T`($i)
+	$POP	r28,`-12*8-4*$SIZE_T`($i)
+	$POP	r29,`-12*8-3*$SIZE_T`($i)
+	$POP	r30,`-12*8-2*$SIZE_T`($i)
+	$POP	r31,`-12*8-1*$SIZE_T`($i)
+	lfd	f20,`-12*8`($i)
+	lfd	f21,`-11*8`($i)
+	lfd	f22,`-10*8`($i)
+	lfd	f23,`-9*8`($i)
+	lfd	f24,`-8*8`($i)
+	lfd	f25,`-7*8`($i)
+	lfd	f26,`-6*8`($i)
+	lfd	f27,`-5*8`($i)
+	lfd	f28,`-4*8`($i)
+	lfd	f29,`-3*8`($i)
+	lfd	f30,`-2*8`($i)
+	lfd	f31,`-1*8`($i)
+	mr	$sp,$i
+	blr
+	.long	0
+	.byte	0,12,4,0,0x8c,10,6,0
+	.long	0
+
+.asciz  "Montgomery Multiplication for PPC64, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/s390x-gf2m.pl b/crypto/bn/asm/s390x-gf2m.pl
new file mode 100755
index 000000000000..cd9f13eca292
--- /dev/null
+++ b/crypto/bn/asm/s390x-gf2m.pl
@@ -0,0 +1,221 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# May 2011
+#
+# The module implements bn_GF2m_mul_2x2 polynomial multiplication used
+# in bn_gf2m.c. It's kind of low-hanging mechanical port from C for
+# the time being... gcc 4.3 appeared to generate poor code, therefore
+# the effort. And indeed, the module delivers 55%-90%(*) improvement
+# on haviest ECDSA verify and ECDH benchmarks for 163- and 571-bit
+# key lengths on z990, 30%-55%(*) - on z10, and 70%-110%(*) - on z196.
+# This is for 64-bit build. In 32-bit "highgprs" case improvement is
+# even higher, for example on z990 it was measured 80%-150%. ECDSA
+# sign is modest 9%-12% faster. Keep in mind that these coefficients
+# are not ones for bn_GF2m_mul_2x2 itself, as not all CPU time is
+# burnt in it...
+#
+# (*)	gcc 4.1 was observed to deliver better results than gcc 4.3,
+#	so that improvement coefficients can vary from one specific
+#	setup to another.
+
+$flavour = shift;
+
+if ($flavour =~ /3[12]/) {
+        $SIZE_T=4;
+        $g="";
+} else {
+        $SIZE_T=8;
+        $g="g";
+}
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$stdframe=16*$SIZE_T+4*8;
+
+$rp="%r2";
+$a1="%r3";
+$a0="%r4";
+$b1="%r5";
+$b0="%r6";
+
+$ra="%r14";
+$sp="%r15";
+
+@T=("%r0","%r1");
+@i=("%r12","%r13");
+
+($a1,$a2,$a4,$a8,$a12,$a48)=map("%r$_",(6..11));
+($lo,$hi,$b)=map("%r$_",(3..5)); $a=$lo; $mask=$a8;
+
+$code.=<<___;
+.text
+
+.type	_mul_1x1,\@function
+.align	16
+_mul_1x1:
+	lgr	$a1,$a
+	sllg	$a2,$a,1
+	sllg	$a4,$a,2
+	sllg	$a8,$a,3
+
+	srag	$lo,$a1,63			# broadcast 63rd bit
+	nihh	$a1,0x1fff
+	srag	@i[0],$a2,63			# broadcast 62nd bit
+	nihh	$a2,0x3fff
+	srag	@i[1],$a4,63			# broadcast 61st bit
+	nihh	$a4,0x7fff
+	ngr	$lo,$b
+	ngr	@i[0],$b
+	ngr	@i[1],$b
+
+	lghi	@T[0],0
+	lgr	$a12,$a1
+	stg	@T[0],`$stdframe+0*8`($sp)	# tab[0]=0
+	xgr	$a12,$a2
+	stg	$a1,`$stdframe+1*8`($sp)	# tab[1]=a1
+	 lgr	$a48,$a4
+	stg	$a2,`$stdframe+2*8`($sp)	# tab[2]=a2
+	 xgr	$a48,$a8
+	stg	$a12,`$stdframe+3*8`($sp)	# tab[3]=a1^a2
+	 xgr	$a1,$a4
+
+	stg	$a4,`$stdframe+4*8`($sp)	# tab[4]=a4
+	xgr	$a2,$a4
+	stg	$a1,`$stdframe+5*8`($sp)	# tab[5]=a1^a4
+	xgr	$a12,$a4
+	stg	$a2,`$stdframe+6*8`($sp)	# tab[6]=a2^a4
+	 xgr	$a1,$a48
+	stg	$a12,`$stdframe+7*8`($sp)	# tab[7]=a1^a2^a4
+	 xgr	$a2,$a48
+
+	stg	$a8,`$stdframe+8*8`($sp)	# tab[8]=a8
+	xgr	$a12,$a48
+	stg	$a1,`$stdframe+9*8`($sp)	# tab[9]=a1^a8
+	 xgr	$a1,$a4
+	stg	$a2,`$stdframe+10*8`($sp)	# tab[10]=a2^a8
+	 xgr	$a2,$a4
+	stg	$a12,`$stdframe+11*8`($sp)	# tab[11]=a1^a2^a8
+
+	xgr	$a12,$a4
+	stg	$a48,`$stdframe+12*8`($sp)	# tab[12]=a4^a8
+	 srlg	$hi,$lo,1
+	stg	$a1,`$stdframe+13*8`($sp)	# tab[13]=a1^a4^a8
+	 sllg	$lo,$lo,63
+	stg	$a2,`$stdframe+14*8`($sp)	# tab[14]=a2^a4^a8
+	 srlg	@T[0],@i[0],2
+	stg	$a12,`$stdframe+15*8`($sp)	# tab[15]=a1^a2^a4^a8
+
+	lghi	$mask,`0xf<<3`
+	sllg	$a1,@i[0],62
+	 sllg	@i[0],$b,3
+	srlg	@T[1],@i[1],3
+	 ngr	@i[0],$mask
+	sllg	$a2,@i[1],61
+	 srlg	@i[1],$b,4-3
+	xgr	$hi,@T[0]
+	 ngr	@i[1],$mask
+	xgr	$lo,$a1
+	xgr	$hi,@T[1]
+	xgr	$lo,$a2
+
+	xg	$lo,$stdframe(@i[0],$sp)
+	srlg	@i[0],$b,8-3
+	ngr	@i[0],$mask
+___
+for($n=1;$n<14;$n++) {
+$code.=<<___;
+	lg	@T[1],$stdframe(@i[1],$sp)
+	srlg	@i[1],$b,`($n+2)*4`-3
+	sllg	@T[0],@T[1],`$n*4`
+	ngr	@i[1],$mask
+	srlg	@T[1],@T[1],`64-$n*4`
+	xgr	$lo,@T[0]
+	xgr	$hi,@T[1]
+___
+	push(@i,shift(@i)); push(@T,shift(@T));
+}
+$code.=<<___;
+	lg	@T[1],$stdframe(@i[1],$sp)
+	sllg	@T[0],@T[1],`$n*4`
+	srlg	@T[1],@T[1],`64-$n*4`
+	xgr	$lo,@T[0]
+	xgr	$hi,@T[1]
+
+	lg	@T[0],$stdframe(@i[0],$sp)
+	sllg	@T[1],@T[0],`($n+1)*4`
+	srlg	@T[0],@T[0],`64-($n+1)*4`
+	xgr	$lo,@T[1]
+	xgr	$hi,@T[0]
+
+	br	$ra
+.size	_mul_1x1,.-_mul_1x1
+
+.globl	bn_GF2m_mul_2x2
+.type	bn_GF2m_mul_2x2,\@function
+.align	16
+bn_GF2m_mul_2x2:
+	stm${g}	%r3,%r15,3*$SIZE_T($sp)
+
+	lghi	%r1,-$stdframe-128
+	la	%r0,0($sp)
+	la	$sp,0(%r1,$sp)			# alloca
+	st${g}	%r0,0($sp)			# back chain
+___
+if ($SIZE_T==8) {
+my @r=map("%r$_",(6..9));
+$code.=<<___;
+	bras	$ra,_mul_1x1			# a1·b1
+	stmg	$lo,$hi,16($rp)
+
+	lg	$a,`$stdframe+128+4*$SIZE_T`($sp)
+	lg	$b,`$stdframe+128+6*$SIZE_T`($sp)
+	bras	$ra,_mul_1x1			# a0·b0
+	stmg	$lo,$hi,0($rp)
+
+	lg	$a,`$stdframe+128+3*$SIZE_T`($sp)
+	lg	$b,`$stdframe+128+5*$SIZE_T`($sp)
+	xg	$a,`$stdframe+128+4*$SIZE_T`($sp)
+	xg	$b,`$stdframe+128+6*$SIZE_T`($sp)
+	bras	$ra,_mul_1x1			# (a0+a1)·(b0+b1)
+	lmg	@r[0],@r[3],0($rp)
+
+	xgr	$lo,$hi
+	xgr	$hi,@r[1]
+	xgr	$lo,@r[0]
+	xgr	$hi,@r[2]
+	xgr	$lo,@r[3]	
+	xgr	$hi,@r[3]
+	xgr	$lo,$hi
+	stg	$hi,16($rp)
+	stg	$lo,8($rp)
+___
+} else {
+$code.=<<___;
+	sllg	%r3,%r3,32
+	sllg	%r5,%r5,32
+	or	%r3,%r4
+	or	%r5,%r6
+	bras	$ra,_mul_1x1
+	rllg	$lo,$lo,32
+	rllg	$hi,$hi,32
+	stmg	$lo,$hi,0($rp)
+___
+}
+$code.=<<___;
+	lm${g}	%r6,%r15,`$stdframe+128+6*$SIZE_T`($sp)
+	br	$ra
+.size	bn_GF2m_mul_2x2,.-bn_GF2m_mul_2x2
+.string	"GF(2^m) Multiplication for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/s390x-mont.pl b/crypto/bn/asm/s390x-mont.pl
new file mode 100755
index 000000000000..9fd64e81eef3
--- /dev/null
+++ b/crypto/bn/asm/s390x-mont.pl
@@ -0,0 +1,277 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# April 2007.
+#
+# Performance improvement over vanilla C code varies from 85% to 45%
+# depending on key length and benchmark. Unfortunately in this context
+# these are not very impressive results [for code that utilizes "wide"
+# 64x64=128-bit multiplication, which is not commonly available to C
+# programmers], at least hand-coded bn_asm.c replacement is known to
+# provide 30-40% better results for longest keys. Well, on a second
+# thought it's not very surprising, because z-CPUs are single-issue
+# and _strictly_ in-order execution, while bn_mul_mont is more or less
+# dependent on CPU ability to pipe-line instructions and have several
+# of them "in-flight" at the same time. I mean while other methods,
+# for example Karatsuba, aim to minimize amount of multiplications at
+# the cost of other operations increase, bn_mul_mont aim to neatly
+# "overlap" multiplications and the other operations [and on most
+# platforms even minimize the amount of the other operations, in
+# particular references to memory]. But it's possible to improve this
+# module performance by implementing dedicated squaring code-path and
+# possibly by unrolling loops...
+
+# January 2009.
+#
+# Reschedule to minimize/avoid Address Generation Interlock hazard,
+# make inner loops counter-based.
+
+# November 2010.
+#
+# Adapt for -m31 build. If kernel supports what's called "highgprs"
+# feature on Linux [see /proc/cpuinfo], it's possible to use 64-bit
+# instructions and achieve "64-bit" performance even in 31-bit legacy
+# application context. The feature is not specific to any particular
+# processor, as long as it's "z-CPU". Latter implies that the code
+# remains z/Architecture specific. Compatibility with 32-bit BN_ULONG
+# is achieved by swapping words after 64-bit loads, follow _dswap-s.
+# On z990 it was measured to perform 2.6-2.2 times better than
+# compiler-generated code, less for longer keys...
+
+$flavour = shift;
+
+if ($flavour =~ /3[12]/) {
+	$SIZE_T=4;
+	$g="";
+} else {
+	$SIZE_T=8;
+	$g="g";
+}
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$stdframe=16*$SIZE_T+4*8;
+
+$mn0="%r0";
+$num="%r1";
+
+# int bn_mul_mont(
+$rp="%r2";		# BN_ULONG *rp,
+$ap="%r3";		# const BN_ULONG *ap,
+$bp="%r4";		# const BN_ULONG *bp,
+$np="%r5";		# const BN_ULONG *np,
+$n0="%r6";		# const BN_ULONG *n0,
+#$num="160(%r15)"	# int num);
+
+$bi="%r2";	# zaps rp
+$j="%r7";
+
+$ahi="%r8";
+$alo="%r9";
+$nhi="%r10";
+$nlo="%r11";
+$AHI="%r12";
+$NHI="%r13";
+$count="%r14";
+$sp="%r15";
+
+$code.=<<___;
+.text
+.globl	bn_mul_mont
+.type	bn_mul_mont,\@function
+bn_mul_mont:
+	lgf	$num,`$stdframe+$SIZE_T-4`($sp)	# pull $num
+	sla	$num,`log($SIZE_T)/log(2)`	# $num to enumerate bytes
+	la	$bp,0($num,$bp)
+
+	st${g}	%r2,2*$SIZE_T($sp)
+
+	cghi	$num,16		#
+	lghi	%r2,0		#
+	blr	%r14		# if($num<16) return 0;
+___
+$code.=<<___ if ($flavour =~ /3[12]/);
+	tmll	$num,4
+	bnzr	%r14		# if ($num&1) return 0;
+___
+$code.=<<___ if ($flavour !~ /3[12]/);
+	cghi	$num,96		#
+	bhr	%r14		# if($num>96) return 0;
+___
+$code.=<<___;
+	stm${g}	%r3,%r15,3*$SIZE_T($sp)
+
+	lghi	$rp,-$stdframe-8	# leave room for carry bit
+	lcgr	$j,$num		# -$num
+	lgr	%r0,$sp
+	la	$rp,0($rp,$sp)
+	la	$sp,0($j,$rp)	# alloca
+	st${g}	%r0,0($sp)	# back chain
+
+	sra	$num,3		# restore $num
+	la	$bp,0($j,$bp)	# restore $bp
+	ahi	$num,-1		# adjust $num for inner loop
+	lg	$n0,0($n0)	# pull n0
+	_dswap	$n0
+
+	lg	$bi,0($bp)
+	_dswap	$bi
+	lg	$alo,0($ap)
+	_dswap	$alo
+	mlgr	$ahi,$bi	# ap[0]*bp[0]
+	lgr	$AHI,$ahi
+
+	lgr	$mn0,$alo	# "tp[0]"*n0
+	msgr	$mn0,$n0
+
+	lg	$nlo,0($np)	#
+	_dswap	$nlo
+	mlgr	$nhi,$mn0	# np[0]*m1
+	algr	$nlo,$alo	# +="tp[0]"
+	lghi	$NHI,0
+	alcgr	$NHI,$nhi
+
+	la	$j,8(%r0)	# j=1
+	lr	$count,$num
+
+.align	16
+.L1st:
+	lg	$alo,0($j,$ap)
+	_dswap	$alo
+	mlgr	$ahi,$bi	# ap[j]*bp[0]
+	algr	$alo,$AHI
+	lghi	$AHI,0
+	alcgr	$AHI,$ahi
+
+	lg	$nlo,0($j,$np)
+	_dswap	$nlo
+	mlgr	$nhi,$mn0	# np[j]*m1
+	algr	$nlo,$NHI
+	lghi	$NHI,0
+	alcgr	$nhi,$NHI	# +="tp[j]"
+	algr	$nlo,$alo
+	alcgr	$NHI,$nhi
+
+	stg	$nlo,$stdframe-8($j,$sp)	# tp[j-1]=
+	la	$j,8($j)	# j++
+	brct	$count,.L1st
+
+	algr	$NHI,$AHI
+	lghi	$AHI,0
+	alcgr	$AHI,$AHI	# upmost overflow bit
+	stg	$NHI,$stdframe-8($j,$sp)
+	stg	$AHI,$stdframe($j,$sp)
+	la	$bp,8($bp)	# bp++
+
+.Louter:
+	lg	$bi,0($bp)	# bp[i]
+	_dswap	$bi
+	lg	$alo,0($ap)
+	_dswap	$alo
+	mlgr	$ahi,$bi	# ap[0]*bp[i]
+	alg	$alo,$stdframe($sp)	# +=tp[0]
+	lghi	$AHI,0
+	alcgr	$AHI,$ahi
+
+	lgr	$mn0,$alo
+	msgr	$mn0,$n0	# tp[0]*n0
+
+	lg	$nlo,0($np)	# np[0]
+	_dswap	$nlo
+	mlgr	$nhi,$mn0	# np[0]*m1
+	algr	$nlo,$alo	# +="tp[0]"
+	lghi	$NHI,0
+	alcgr	$NHI,$nhi
+
+	la	$j,8(%r0)	# j=1
+	lr	$count,$num
+
+.align	16
+.Linner:
+	lg	$alo,0($j,$ap)
+	_dswap	$alo
+	mlgr	$ahi,$bi	# ap[j]*bp[i]
+	algr	$alo,$AHI
+	lghi	$AHI,0
+	alcgr	$ahi,$AHI
+	alg	$alo,$stdframe($j,$sp)# +=tp[j]
+	alcgr	$AHI,$ahi
+
+	lg	$nlo,0($j,$np)
+	_dswap	$nlo
+	mlgr	$nhi,$mn0	# np[j]*m1
+	algr	$nlo,$NHI
+	lghi	$NHI,0
+	alcgr	$nhi,$NHI
+	algr	$nlo,$alo	# +="tp[j]"
+	alcgr	$NHI,$nhi
+
+	stg	$nlo,$stdframe-8($j,$sp)	# tp[j-1]=
+	la	$j,8($j)	# j++
+	brct	$count,.Linner
+
+	algr	$NHI,$AHI
+	lghi	$AHI,0
+	alcgr	$AHI,$AHI
+	alg	$NHI,$stdframe($j,$sp)# accumulate previous upmost overflow bit
+	lghi	$ahi,0
+	alcgr	$AHI,$ahi	# new upmost overflow bit
+	stg	$NHI,$stdframe-8($j,$sp)
+	stg	$AHI,$stdframe($j,$sp)
+
+	la	$bp,8($bp)	# bp++
+	cl${g}	$bp,`$stdframe+8+4*$SIZE_T`($j,$sp)	# compare to &bp[num]
+	jne	.Louter
+
+	l${g}	$rp,`$stdframe+8+2*$SIZE_T`($j,$sp)	# reincarnate rp
+	la	$ap,$stdframe($sp)
+	ahi	$num,1		# restore $num, incidentally clears "borrow"
+
+	la	$j,0(%r0)
+	lr	$count,$num
+.Lsub:	lg	$alo,0($j,$ap)
+	lg	$nlo,0($j,$np)
+	_dswap	$nlo
+	slbgr	$alo,$nlo
+	stg	$alo,0($j,$rp)
+	la	$j,8($j)
+	brct	$count,.Lsub
+	lghi	$ahi,0
+	slbgr	$AHI,$ahi	# handle upmost carry
+
+	ngr	$ap,$AHI
+	lghi	$np,-1
+	xgr	$np,$AHI
+	ngr	$np,$rp
+	ogr	$ap,$np		# ap=borrow?tp:rp
+
+	la	$j,0(%r0)
+	lgr	$count,$num
+.Lcopy:	lg	$alo,0($j,$ap)		# copy or in-place refresh
+	_dswap	$alo
+	stg	$j,$stdframe($j,$sp)	# zap tp
+	stg	$alo,0($j,$rp)
+	la	$j,8($j)
+	brct	$count,.Lcopy
+
+	la	%r1,`$stdframe+8+6*$SIZE_T`($j,$sp)
+	lm${g}	%r6,%r15,0(%r1)
+	lghi	%r2,1		# signal "processed"
+	br	%r14
+.size	bn_mul_mont,.-bn_mul_mont
+.string	"Montgomery Multiplication for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+foreach (split("\n",$code)) {
+	s/\`([^\`]*)\`/eval $1/ge;
+	s/_dswap\s+(%r[0-9]+)/sprintf("rllg\t%s,%s,32",$1,$1) if($SIZE_T==4)/e;
+	print $_,"\n";
+}
+close STDOUT;
diff --git a/crypto/bn/asm/s390x.S b/crypto/bn/asm/s390x.S
new file mode 100755
index 000000000000..43fcb79bc011
--- /dev/null
+++ b/crypto/bn/asm/s390x.S
@@ -0,0 +1,678 @@
+.ident "s390x.S, version 1.1"
+// ====================================================================
+// Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+// project.
+//
+// Rights for redistribution and usage in source and binary forms are
+// granted according to the OpenSSL license. Warranty of any kind is
+// disclaimed.
+// ====================================================================
+
+.text
+
+#define zero	%r0
+
+// BN_ULONG bn_mul_add_words(BN_ULONG *r2,BN_ULONG *r3,int r4,BN_ULONG r5);
+.globl	bn_mul_add_words
+.type	bn_mul_add_words,@function
+.align	4
+bn_mul_add_words:
+	lghi	zero,0		// zero = 0
+	la	%r1,0(%r2)	// put rp aside
+	lghi	%r2,0		// i=0;
+	ltgfr	%r4,%r4
+	bler	%r14		// if (len<=0) return 0;
+
+	stmg	%r6,%r10,48(%r15)
+	lghi	%r10,3
+	lghi	%r8,0		// carry = 0
+	nr	%r10,%r4	// len%4
+	sra	%r4,2		// cnt=len/4
+	jz	.Loop1_madd	// carry is incidentally cleared if branch taken
+	algr	zero,zero	// clear carry
+
+.Loop4_madd:
+	lg	%r7,0(%r2,%r3)	// ap[i]
+	mlgr	%r6,%r5		// *=w
+	alcgr	%r7,%r8		// +=carry
+	alcgr	%r6,zero
+	alg	%r7,0(%r2,%r1)	// +=rp[i]
+	stg	%r7,0(%r2,%r1)	// rp[i]=
+
+	lg	%r9,8(%r2,%r3)
+	mlgr	%r8,%r5
+	alcgr	%r9,%r6
+	alcgr	%r8,zero
+	alg	%r9,8(%r2,%r1)
+	stg	%r9,8(%r2,%r1)
+
+	lg	%r7,16(%r2,%r3)
+	mlgr	%r6,%r5
+	alcgr	%r7,%r8
+	alcgr	%r6,zero
+	alg	%r7,16(%r2,%r1)
+	stg	%r7,16(%r2,%r1)
+
+	lg	%r9,24(%r2,%r3)
+	mlgr	%r8,%r5
+	alcgr	%r9,%r6
+	alcgr	%r8,zero
+	alg	%r9,24(%r2,%r1)
+	stg	%r9,24(%r2,%r1)
+
+	la	%r2,32(%r2)	// i+=4
+	brct	%r4,.Loop4_madd
+
+	la	%r10,1(%r10)		// see if len%4 is zero ...
+	brct	%r10,.Loop1_madd	// without touching condition code:-)
+
+.Lend_madd:
+	alcgr	%r8,zero	// collect carry bit
+	lgr	%r2,%r8
+	lmg	%r6,%r10,48(%r15)
+	br	%r14
+
+.Loop1_madd:
+	lg	%r7,0(%r2,%r3)	// ap[i]
+	mlgr	%r6,%r5		// *=w
+	alcgr	%r7,%r8		// +=carry
+	alcgr	%r6,zero
+	alg	%r7,0(%r2,%r1)	// +=rp[i]
+	stg	%r7,0(%r2,%r1)	// rp[i]=
+
+	lgr	%r8,%r6
+	la	%r2,8(%r2)	// i++
+	brct	%r10,.Loop1_madd
+
+	j	.Lend_madd
+.size	bn_mul_add_words,.-bn_mul_add_words
+
+// BN_ULONG bn_mul_words(BN_ULONG *r2,BN_ULONG *r3,int r4,BN_ULONG r5);
+.globl	bn_mul_words
+.type	bn_mul_words,@function
+.align	4
+bn_mul_words:
+	lghi	zero,0		// zero = 0
+	la	%r1,0(%r2)	// put rp aside
+	lghi	%r2,0		// i=0;
+	ltgfr	%r4,%r4
+	bler	%r14		// if (len<=0) return 0;
+
+	stmg	%r6,%r10,48(%r15)
+	lghi	%r10,3
+	lghi	%r8,0		// carry = 0
+	nr	%r10,%r4	// len%4
+	sra	%r4,2		// cnt=len/4
+	jz	.Loop1_mul	// carry is incidentally cleared if branch taken
+	algr	zero,zero	// clear carry
+
+.Loop4_mul:
+	lg	%r7,0(%r2,%r3)	// ap[i]
+	mlgr	%r6,%r5		// *=w
+	alcgr	%r7,%r8		// +=carry
+	stg	%r7,0(%r2,%r1)	// rp[i]=
+
+	lg	%r9,8(%r2,%r3)
+	mlgr	%r8,%r5
+	alcgr	%r9,%r6
+	stg	%r9,8(%r2,%r1)
+
+	lg	%r7,16(%r2,%r3)
+	mlgr	%r6,%r5
+	alcgr	%r7,%r8
+	stg	%r7,16(%r2,%r1)
+
+	lg	%r9,24(%r2,%r3)
+	mlgr	%r8,%r5
+	alcgr	%r9,%r6
+	stg	%r9,24(%r2,%r1)
+
+	la	%r2,32(%r2)	// i+=4
+	brct	%r4,.Loop4_mul
+
+	la	%r10,1(%r10)		// see if len%4 is zero ...
+	brct	%r10,.Loop1_mul		// without touching condition code:-)
+
+.Lend_mul:
+	alcgr	%r8,zero	// collect carry bit
+	lgr	%r2,%r8
+	lmg	%r6,%r10,48(%r15)
+	br	%r14
+
+.Loop1_mul:
+	lg	%r7,0(%r2,%r3)	// ap[i]
+	mlgr	%r6,%r5		// *=w
+	alcgr	%r7,%r8		// +=carry
+	stg	%r7,0(%r2,%r1)	// rp[i]=
+
+	lgr	%r8,%r6
+	la	%r2,8(%r2)	// i++
+	brct	%r10,.Loop1_mul
+
+	j	.Lend_mul
+.size	bn_mul_words,.-bn_mul_words
+
+// void bn_sqr_words(BN_ULONG *r2,BN_ULONG *r2,int r4)
+.globl	bn_sqr_words
+.type	bn_sqr_words,@function
+.align	4
+bn_sqr_words:
+	ltgfr	%r4,%r4
+	bler	%r14
+
+	stmg	%r6,%r7,48(%r15)
+	srag	%r1,%r4,2	// cnt=len/4
+	jz	.Loop1_sqr
+
+.Loop4_sqr:
+	lg	%r7,0(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,0(%r2)
+	stg	%r6,8(%r2)
+
+	lg	%r7,8(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,16(%r2)
+	stg	%r6,24(%r2)
+
+	lg	%r7,16(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,32(%r2)
+	stg	%r6,40(%r2)
+
+	lg	%r7,24(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,48(%r2)
+	stg	%r6,56(%r2)
+
+	la	%r3,32(%r3)
+	la	%r2,64(%r2)
+	brct	%r1,.Loop4_sqr
+
+	lghi	%r1,3
+	nr	%r4,%r1		// cnt=len%4
+	jz	.Lend_sqr
+
+.Loop1_sqr:
+	lg	%r7,0(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,0(%r2)
+	stg	%r6,8(%r2)
+
+	la	%r3,8(%r3)
+	la	%r2,16(%r2)
+	brct	%r4,.Loop1_sqr
+
+.Lend_sqr:
+	lmg	%r6,%r7,48(%r15)
+	br	%r14
+.size	bn_sqr_words,.-bn_sqr_words
+
+// BN_ULONG bn_div_words(BN_ULONG h,BN_ULONG l,BN_ULONG d);
+.globl	bn_div_words
+.type	bn_div_words,@function
+.align	4
+bn_div_words:
+	dlgr	%r2,%r4
+	lgr	%r2,%r3
+	br	%r14
+.size	bn_div_words,.-bn_div_words
+
+// BN_ULONG bn_add_words(BN_ULONG *r2,BN_ULONG *r3,BN_ULONG *r4,int r5);
+.globl	bn_add_words
+.type	bn_add_words,@function
+.align	4
+bn_add_words:
+	la	%r1,0(%r2)	// put rp aside
+	lghi	%r2,0		// i=0
+	ltgfr	%r5,%r5
+	bler	%r14		// if (len<=0) return 0;
+
+	stg	%r6,48(%r15)
+	lghi	%r6,3
+	nr	%r6,%r5		// len%4
+	sra	%r5,2		// len/4, use sra because it sets condition code
+	jz	.Loop1_add	// carry is incidentally cleared if branch taken
+	algr	%r2,%r2		// clear carry
+
+.Loop4_add:
+	lg	%r0,0(%r2,%r3)
+	alcg	%r0,0(%r2,%r4)
+	stg	%r0,0(%r2,%r1)
+	lg	%r0,8(%r2,%r3)
+	alcg	%r0,8(%r2,%r4)
+	stg	%r0,8(%r2,%r1)
+	lg	%r0,16(%r2,%r3)
+	alcg	%r0,16(%r2,%r4)
+	stg	%r0,16(%r2,%r1)
+	lg	%r0,24(%r2,%r3)
+	alcg	%r0,24(%r2,%r4)
+	stg	%r0,24(%r2,%r1)
+
+	la	%r2,32(%r2)	// i+=4
+	brct	%r5,.Loop4_add
+
+	la	%r6,1(%r6)	// see if len%4 is zero ...
+	brct	%r6,.Loop1_add	// without touching condition code:-)
+
+.Lexit_add:
+	lghi	%r2,0
+	alcgr	%r2,%r2
+	lg	%r6,48(%r15)
+	br	%r14
+
+.Loop1_add:
+	lg	%r0,0(%r2,%r3)
+	alcg	%r0,0(%r2,%r4)
+	stg	%r0,0(%r2,%r1)
+
+	la	%r2,8(%r2)	// i++
+	brct	%r6,.Loop1_add
+
+	j	.Lexit_add
+.size	bn_add_words,.-bn_add_words
+
+// BN_ULONG bn_sub_words(BN_ULONG *r2,BN_ULONG *r3,BN_ULONG *r4,int r5);
+.globl	bn_sub_words
+.type	bn_sub_words,@function
+.align	4
+bn_sub_words:
+	la	%r1,0(%r2)	// put rp aside
+	lghi	%r2,0		// i=0
+	ltgfr	%r5,%r5
+	bler	%r14		// if (len<=0) return 0;
+
+	stg	%r6,48(%r15)
+	lghi	%r6,3
+	nr	%r6,%r5		// len%4
+	sra	%r5,2		// len/4, use sra because it sets condition code
+	jnz	.Loop4_sub	// borrow is incidentally cleared if branch taken
+	slgr	%r2,%r2		// clear borrow
+
+.Loop1_sub:
+	lg	%r0,0(%r2,%r3)
+	slbg	%r0,0(%r2,%r4)
+	stg	%r0,0(%r2,%r1)
+
+	la	%r2,8(%r2)	// i++
+	brct	%r6,.Loop1_sub
+	j	.Lexit_sub
+
+.Loop4_sub:
+	lg	%r0,0(%r2,%r3)
+	slbg	%r0,0(%r2,%r4)
+	stg	%r0,0(%r2,%r1)
+	lg	%r0,8(%r2,%r3)
+	slbg	%r0,8(%r2,%r4)
+	stg	%r0,8(%r2,%r1)
+	lg	%r0,16(%r2,%r3)
+	slbg	%r0,16(%r2,%r4)
+	stg	%r0,16(%r2,%r1)
+	lg	%r0,24(%r2,%r3)
+	slbg	%r0,24(%r2,%r4)
+	stg	%r0,24(%r2,%r1)
+
+	la	%r2,32(%r2)	// i+=4
+	brct	%r5,.Loop4_sub
+
+	la	%r6,1(%r6)	// see if len%4 is zero ...
+	brct	%r6,.Loop1_sub	// without touching condition code:-)
+
+.Lexit_sub:
+	lghi	%r2,0
+	slbgr	%r2,%r2
+	lcgr	%r2,%r2
+	lg	%r6,48(%r15)
+	br	%r14
+.size	bn_sub_words,.-bn_sub_words
+
+#define c1	%r1
+#define c2	%r5
+#define c3	%r8
+
+#define mul_add_c(ai,bi,c1,c2,c3)	\
+	lg	%r7,ai*8(%r3);		\
+	mlg	%r6,bi*8(%r4);		\
+	algr	c1,%r7;			\
+	alcgr	c2,%r6;			\
+	alcgr	c3,zero
+
+// void bn_mul_comba8(BN_ULONG *r2,BN_ULONG *r3,BN_ULONG *r4);
+.globl	bn_mul_comba8
+.type	bn_mul_comba8,@function
+.align	4
+bn_mul_comba8:
+	stmg	%r6,%r8,48(%r15)
+
+	lghi	c1,0
+	lghi	c2,0
+	lghi	c3,0
+	lghi	zero,0
+
+	mul_add_c(0,0,c1,c2,c3);
+	stg	c1,0*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(0,1,c2,c3,c1);
+	mul_add_c(1,0,c2,c3,c1);
+	stg	c2,1*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(2,0,c3,c1,c2);
+	mul_add_c(1,1,c3,c1,c2);
+	mul_add_c(0,2,c3,c1,c2);
+	stg	c3,2*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(0,3,c1,c2,c3);
+	mul_add_c(1,2,c1,c2,c3);
+	mul_add_c(2,1,c1,c2,c3);
+	mul_add_c(3,0,c1,c2,c3);
+	stg	c1,3*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(4,0,c2,c3,c1);
+	mul_add_c(3,1,c2,c3,c1);
+	mul_add_c(2,2,c2,c3,c1);
+	mul_add_c(1,3,c2,c3,c1);
+	mul_add_c(0,4,c2,c3,c1);
+	stg	c2,4*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(0,5,c3,c1,c2);
+	mul_add_c(1,4,c3,c1,c2);
+	mul_add_c(2,3,c3,c1,c2);
+	mul_add_c(3,2,c3,c1,c2);
+	mul_add_c(4,1,c3,c1,c2);
+	mul_add_c(5,0,c3,c1,c2);
+	stg	c3,5*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(6,0,c1,c2,c3);
+	mul_add_c(5,1,c1,c2,c3);
+	mul_add_c(4,2,c1,c2,c3);
+	mul_add_c(3,3,c1,c2,c3);
+	mul_add_c(2,4,c1,c2,c3);
+	mul_add_c(1,5,c1,c2,c3);
+	mul_add_c(0,6,c1,c2,c3);
+	stg	c1,6*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(0,7,c2,c3,c1);
+	mul_add_c(1,6,c2,c3,c1);
+	mul_add_c(2,5,c2,c3,c1);
+	mul_add_c(3,4,c2,c3,c1);
+	mul_add_c(4,3,c2,c3,c1);
+	mul_add_c(5,2,c2,c3,c1);
+	mul_add_c(6,1,c2,c3,c1);
+	mul_add_c(7,0,c2,c3,c1);
+	stg	c2,7*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(7,1,c3,c1,c2);
+	mul_add_c(6,2,c3,c1,c2);
+	mul_add_c(5,3,c3,c1,c2);
+	mul_add_c(4,4,c3,c1,c2);
+	mul_add_c(3,5,c3,c1,c2);
+	mul_add_c(2,6,c3,c1,c2);
+	mul_add_c(1,7,c3,c1,c2);
+	stg	c3,8*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(2,7,c1,c2,c3);
+	mul_add_c(3,6,c1,c2,c3);
+	mul_add_c(4,5,c1,c2,c3);
+	mul_add_c(5,4,c1,c2,c3);
+	mul_add_c(6,3,c1,c2,c3);
+	mul_add_c(7,2,c1,c2,c3);
+	stg	c1,9*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(7,3,c2,c3,c1);
+	mul_add_c(6,4,c2,c3,c1);
+	mul_add_c(5,5,c2,c3,c1);
+	mul_add_c(4,6,c2,c3,c1);
+	mul_add_c(3,7,c2,c3,c1);
+	stg	c2,10*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(4,7,c3,c1,c2);
+	mul_add_c(5,6,c3,c1,c2);
+	mul_add_c(6,5,c3,c1,c2);
+	mul_add_c(7,4,c3,c1,c2);
+	stg	c3,11*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(7,5,c1,c2,c3);
+	mul_add_c(6,6,c1,c2,c3);
+	mul_add_c(5,7,c1,c2,c3);
+	stg	c1,12*8(%r2)
+	lghi	c1,0
+
+
+	mul_add_c(6,7,c2,c3,c1);
+	mul_add_c(7,6,c2,c3,c1);
+	stg	c2,13*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(7,7,c3,c1,c2);
+	stg	c3,14*8(%r2)
+	stg	c1,15*8(%r2)
+
+	lmg	%r6,%r8,48(%r15)
+	br	%r14
+.size	bn_mul_comba8,.-bn_mul_comba8
+
+// void bn_mul_comba4(BN_ULONG *r2,BN_ULONG *r3,BN_ULONG *r4);
+.globl	bn_mul_comba4
+.type	bn_mul_comba4,@function
+.align	4
+bn_mul_comba4:
+	stmg	%r6,%r8,48(%r15)
+
+	lghi	c1,0
+	lghi	c2,0
+	lghi	c3,0
+	lghi	zero,0
+
+	mul_add_c(0,0,c1,c2,c3);
+	stg	c1,0*8(%r3)
+	lghi	c1,0
+
+	mul_add_c(0,1,c2,c3,c1);
+	mul_add_c(1,0,c2,c3,c1);
+	stg	c2,1*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(2,0,c3,c1,c2);
+	mul_add_c(1,1,c3,c1,c2);
+	mul_add_c(0,2,c3,c1,c2);
+	stg	c3,2*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(0,3,c1,c2,c3);
+	mul_add_c(1,2,c1,c2,c3);
+	mul_add_c(2,1,c1,c2,c3);
+	mul_add_c(3,0,c1,c2,c3);
+	stg	c1,3*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(3,1,c2,c3,c1);
+	mul_add_c(2,2,c2,c3,c1);
+	mul_add_c(1,3,c2,c3,c1);
+	stg	c2,4*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(2,3,c3,c1,c2);
+	mul_add_c(3,2,c3,c1,c2);
+	stg	c3,5*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(3,3,c1,c2,c3);
+	stg	c1,6*8(%r2)
+	stg	c2,7*8(%r2)
+
+	stmg	%r6,%r8,48(%r15)
+	br	%r14
+.size	bn_mul_comba4,.-bn_mul_comba4
+
+#define sqr_add_c(ai,c1,c2,c3)		\
+	lg	%r7,ai*8(%r3);		\
+	mlgr	%r6,%r7;		\
+	algr	c1,%r7;			\
+	alcgr	c2,%r6;			\
+	alcgr	c3,zero
+
+#define sqr_add_c2(ai,aj,c1,c2,c3)	\
+	lg	%r7,ai*8(%r3);		\
+	mlg	%r6,aj*8(%r3);		\
+	algr	c1,%r7;			\
+	alcgr	c2,%r6;			\
+	alcgr	c3,zero;		\
+	algr	c1,%r7;			\
+	alcgr	c2,%r6;			\
+	alcgr	c3,zero
+
+// void bn_sqr_comba8(BN_ULONG *r2,BN_ULONG *r3);
+.globl	bn_sqr_comba8
+.type	bn_sqr_comba8,@function
+.align	4
+bn_sqr_comba8:
+	stmg	%r6,%r8,48(%r15)
+
+	lghi	c1,0
+	lghi	c2,0
+	lghi	c3,0
+	lghi	zero,0
+
+	sqr_add_c(0,c1,c2,c3);
+	stg	c1,0*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c2(1,0,c2,c3,c1);
+	stg	c2,1*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c(1,c3,c1,c2);
+	sqr_add_c2(2,0,c3,c1,c2);
+	stg	c3,2*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c2(3,0,c1,c2,c3);
+	sqr_add_c2(2,1,c1,c2,c3);
+	stg	c1,3*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c(2,c2,c3,c1);
+	sqr_add_c2(3,1,c2,c3,c1);
+	sqr_add_c2(4,0,c2,c3,c1);
+	stg	c2,4*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c2(5,0,c3,c1,c2);
+	sqr_add_c2(4,1,c3,c1,c2);
+	sqr_add_c2(3,2,c3,c1,c2);
+	stg	c3,5*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c(3,c1,c2,c3);
+	sqr_add_c2(4,2,c1,c2,c3);
+	sqr_add_c2(5,1,c1,c2,c3);
+	sqr_add_c2(6,0,c1,c2,c3);
+	stg	c1,6*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c2(7,0,c2,c3,c1);
+	sqr_add_c2(6,1,c2,c3,c1);
+	sqr_add_c2(5,2,c2,c3,c1);
+	sqr_add_c2(4,3,c2,c3,c1);
+	stg	c2,7*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c(4,c3,c1,c2);
+	sqr_add_c2(5,3,c3,c1,c2);
+	sqr_add_c2(6,2,c3,c1,c2);
+	sqr_add_c2(7,1,c3,c1,c2);
+	stg	c3,8*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c2(7,2,c1,c2,c3);
+	sqr_add_c2(6,3,c1,c2,c3);
+	sqr_add_c2(5,4,c1,c2,c3);
+	stg	c1,9*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c(5,c2,c3,c1);
+	sqr_add_c2(6,4,c2,c3,c1);
+	sqr_add_c2(7,3,c2,c3,c1);
+	stg	c2,10*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c2(7,4,c3,c1,c2);
+	sqr_add_c2(6,5,c3,c1,c2);
+	stg	c3,11*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c(6,c1,c2,c3);
+	sqr_add_c2(7,5,c1,c2,c3);
+	stg	c1,12*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c2(7,6,c2,c3,c1);
+	stg	c2,13*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c(7,c3,c1,c2);
+	stg	c3,14*8(%r2)
+	stg	c1,15*8(%r2)
+
+	lmg	%r6,%r8,48(%r15)
+	br	%r14
+.size	bn_sqr_comba8,.-bn_sqr_comba8
+
+// void bn_sqr_comba4(BN_ULONG *r2,BN_ULONG *r3);
+.globl bn_sqr_comba4
+.type	bn_sqr_comba4,@function
+.align	4
+bn_sqr_comba4:
+	stmg	%r6,%r8,48(%r15)
+
+	lghi	c1,0
+	lghi	c2,0
+	lghi	c3,0
+	lghi	zero,0
+
+	sqr_add_c(0,c1,c2,c3);
+	stg	c1,0*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c2(1,0,c2,c3,c1);
+	stg	c2,1*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c(1,c3,c1,c2);
+	sqr_add_c2(2,0,c3,c1,c2);
+	stg	c3,2*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c2(3,0,c1,c2,c3);
+	sqr_add_c2(2,1,c1,c2,c3);
+	stg	c1,3*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c(2,c2,c3,c1);
+	sqr_add_c2(3,1,c2,c3,c1);
+	stg	c2,4*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c2(3,2,c3,c1,c2);
+	stg	c3,5*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c(3,c1,c2,c3);
+	stg	c1,6*8(%r2)
+	stg	c2,7*8(%r2)
+
+	lmg	%r6,%r8,48(%r15)
+	br	%r14
+.size	bn_sqr_comba4,.-bn_sqr_comba4
diff --git a/crypto/bn/asm/sparcv8plus.S b/crypto/bn/asm/sparcv8plus.S
index 8c56e2e7e7cb..63de1860f285 100644
--- a/crypto/bn/asm/sparcv8plus.S
+++ b/crypto/bn/asm/sparcv8plus.S
@@ -144,6 +144,19 @@
  *	    }
  */
 
+#if defined(__SUNPRO_C) && defined(__sparcv9)
+  /* They've said -xarch=v9 at command line */
+  .register	%g2,#scratch
+  .register	%g3,#scratch
+# define	FRAME_SIZE	-192
+#elif defined(__GNUC__) && defined(__arch64__)
+  /* They've said -m64 at command line */
+  .register	%g2,#scratch
+  .register	%g3,#scratch
+# define	FRAME_SIZE	-192
+#else 
+# define	FRAME_SIZE	-96
+#endif 
 /*
  * GNU assembler can't stand stuw:-(
  */
@@ -619,8 +632,6 @@ bn_sub_words:
  *							Andy.
  */
 
-#define FRAME_SIZE	-96
-
 /*
  * Here is register usage map for *all* routines below.
  */
diff --git a/crypto/bn/asm/sparcv9-mont.pl b/crypto/bn/asm/sparcv9-mont.pl
new file mode 100755
index 000000000000..b8fb1e8a25dc
--- /dev/null
+++ b/crypto/bn/asm/sparcv9-mont.pl
@@ -0,0 +1,606 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# December 2005
+#
+# Pure SPARCv9/8+ and IALU-only bn_mul_mont implementation. The reasons
+# for undertaken effort are multiple. First of all, UltraSPARC is not
+# the whole SPARCv9 universe and other VIS-free implementations deserve
+# optimized code as much. Secondly, newly introduced UltraSPARC T1,
+# a.k.a. Niagara, has shared FPU and concurrent FPU-intensive pathes,
+# such as sparcv9a-mont, will simply sink it. Yes, T1 is equipped with
+# several integrated RSA/DSA accelerator circuits accessible through
+# kernel driver [only(*)], but having decent user-land software
+# implementation is important too. Finally, reasons like desire to
+# experiment with dedicated squaring procedure. Yes, this module
+# implements one, because it was easiest to draft it in SPARCv9
+# instructions...
+
+# (*)	Engine accessing the driver in question is on my TODO list.
+#	For reference, acceleator is estimated to give 6 to 10 times
+#	improvement on single-threaded RSA sign. It should be noted
+#	that 6-10x improvement coefficient does not actually mean
+#	something extraordinary in terms of absolute [single-threaded]
+#	performance, as SPARCv9 instruction set is by all means least
+#	suitable for high performance crypto among other 64 bit
+#	platforms. 6-10x factor simply places T1 in same performance
+#	domain as say AMD64 and IA-64. Improvement of RSA verify don't
+#	appear impressive at all, but it's the sign operation which is
+#	far more critical/interesting.
+
+# You might notice that inner loops are modulo-scheduled:-) This has
+# essentially negligible impact on UltraSPARC performance, it's
+# Fujitsu SPARC64 V users who should notice and hopefully appreciate
+# the advantage... Currently this module surpasses sparcv9a-mont.pl
+# by ~20% on UltraSPARC-III and later cores, but recall that sparcv9a
+# module still have hidden potential [see TODO list there], which is
+# estimated to be larger than 20%...
+
+# int bn_mul_mont(
+$rp="%i0";	# BN_ULONG *rp,
+$ap="%i1";	# const BN_ULONG *ap,
+$bp="%i2";	# const BN_ULONG *bp,
+$np="%i3";	# const BN_ULONG *np,
+$n0="%i4";	# const BN_ULONG *n0,
+$num="%i5";	# int num);
+
+$bits=32;
+for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+if ($bits==64)	{ $bias=2047; $frame=192; }
+else		{ $bias=0;    $frame=128; }
+
+$car0="%o0";
+$car1="%o1";
+$car2="%o2";	# 1 bit
+$acc0="%o3";
+$acc1="%o4";
+$mask="%g1";	# 32 bits, what a waste...
+$tmp0="%g4";
+$tmp1="%g5";
+
+$i="%l0";
+$j="%l1";
+$mul0="%l2";
+$mul1="%l3";
+$tp="%l4";
+$apj="%l5";
+$npj="%l6";
+$tpj="%l7";
+
+$fname="bn_mul_mont_int";
+
+$code=<<___;
+.section	".text",#alloc,#execinstr
+
+.global	$fname
+.align	32
+$fname:
+	cmp	%o5,4			! 128 bits minimum
+	bge,pt	%icc,.Lenter
+	sethi	%hi(0xffffffff),$mask
+	retl
+	clr	%o0
+.align	32
+.Lenter:
+	save	%sp,-$frame,%sp
+	sll	$num,2,$num		! num*=4
+	or	$mask,%lo(0xffffffff),$mask
+	ld	[$n0],$n0
+	cmp	$ap,$bp
+	and	$num,$mask,$num
+	ld	[$bp],$mul0		! bp[0]
+	nop
+
+	add	%sp,$bias,%o7		! real top of stack
+	ld	[$ap],$car0		! ap[0] ! redundant in squaring context
+	sub	%o7,$num,%o7
+	ld	[$ap+4],$apj		! ap[1]
+	and	%o7,-1024,%o7
+	ld	[$np],$car1		! np[0]
+	sub	%o7,$bias,%sp		! alloca
+	ld	[$np+4],$npj		! np[1]
+	be,pt	`$bits==32?"%icc":"%xcc"`,.Lbn_sqr_mont
+	mov	12,$j
+
+	mulx	$car0,$mul0,$car0	! ap[0]*bp[0]
+	mulx	$apj,$mul0,$tmp0	!prologue! ap[1]*bp[0]
+	and	$car0,$mask,$acc0
+	add	%sp,$bias+$frame,$tp
+	ld	[$ap+8],$apj		!prologue!
+
+	mulx	$n0,$acc0,$mul1		! "t[0]"*n0
+	and	$mul1,$mask,$mul1
+
+	mulx	$car1,$mul1,$car1	! np[0]*"t[0]"*n0
+	mulx	$npj,$mul1,$acc1	!prologue! np[1]*"t[0]"*n0
+	srlx	$car0,32,$car0
+	add	$acc0,$car1,$car1
+	ld	[$np+8],$npj		!prologue!
+	srlx	$car1,32,$car1
+	mov	$tmp0,$acc0		!prologue!
+
+.L1st:
+	mulx	$apj,$mul0,$tmp0
+	mulx	$npj,$mul1,$tmp1
+	add	$acc0,$car0,$car0
+	ld	[$ap+$j],$apj		! ap[j]
+	and	$car0,$mask,$acc0
+	add	$acc1,$car1,$car1
+	ld	[$np+$j],$npj		! np[j]
+	srlx	$car0,32,$car0
+	add	$acc0,$car1,$car1
+	add	$j,4,$j			! j++
+	mov	$tmp0,$acc0
+	st	$car1,[$tp]
+	cmp	$j,$num
+	mov	$tmp1,$acc1
+	srlx	$car1,32,$car1
+	bl	%icc,.L1st
+	add	$tp,4,$tp		! tp++
+!.L1st
+
+	mulx	$apj,$mul0,$tmp0	!epilogue!
+	mulx	$npj,$mul1,$tmp1
+	add	$acc0,$car0,$car0
+	and	$car0,$mask,$acc0
+	add	$acc1,$car1,$car1
+	srlx	$car0,32,$car0
+	add	$acc0,$car1,$car1
+	st	$car1,[$tp]
+	srlx	$car1,32,$car1
+
+	add	$tmp0,$car0,$car0
+	and	$car0,$mask,$acc0
+	add	$tmp1,$car1,$car1
+	srlx	$car0,32,$car0
+	add	$acc0,$car1,$car1
+	st	$car1,[$tp+4]
+	srlx	$car1,32,$car1
+
+	add	$car0,$car1,$car1
+	st	$car1,[$tp+8]
+	srlx	$car1,32,$car2
+
+	mov	4,$i			! i++
+	ld	[$bp+4],$mul0		! bp[1]
+.Louter:
+	add	%sp,$bias+$frame,$tp
+	ld	[$ap],$car0		! ap[0]
+	ld	[$ap+4],$apj		! ap[1]
+	ld	[$np],$car1		! np[0]
+	ld	[$np+4],$npj		! np[1]
+	ld	[$tp],$tmp1		! tp[0]
+	ld	[$tp+4],$tpj		! tp[1]
+	mov	12,$j
+
+	mulx	$car0,$mul0,$car0
+	mulx	$apj,$mul0,$tmp0	!prologue!
+	add	$tmp1,$car0,$car0
+	ld	[$ap+8],$apj		!prologue!
+	and	$car0,$mask,$acc0
+
+	mulx	$n0,$acc0,$mul1
+	and	$mul1,$mask,$mul1
+
+	mulx	$car1,$mul1,$car1
+	mulx	$npj,$mul1,$acc1	!prologue!
+	srlx	$car0,32,$car0
+	add	$acc0,$car1,$car1
+	ld	[$np+8],$npj		!prologue!
+	srlx	$car1,32,$car1
+	mov	$tmp0,$acc0		!prologue!
+
+.Linner:
+	mulx	$apj,$mul0,$tmp0
+	mulx	$npj,$mul1,$tmp1
+	add	$tpj,$car0,$car0
+	ld	[$ap+$j],$apj		! ap[j]
+	add	$acc0,$car0,$car0
+	add	$acc1,$car1,$car1
+	ld	[$np+$j],$npj		! np[j]
+	and	$car0,$mask,$acc0
+	ld	[$tp+8],$tpj		! tp[j]
+	srlx	$car0,32,$car0
+	add	$acc0,$car1,$car1
+	add	$j,4,$j			! j++
+	mov	$tmp0,$acc0
+	st	$car1,[$tp]		! tp[j-1]
+	srlx	$car1,32,$car1
+	mov	$tmp1,$acc1
+	cmp	$j,$num
+	bl	%icc,.Linner
+	add	$tp,4,$tp		! tp++
+!.Linner
+
+	mulx	$apj,$mul0,$tmp0	!epilogue!
+	mulx	$npj,$mul1,$tmp1
+	add	$tpj,$car0,$car0
+	add	$acc0,$car0,$car0
+	ld	[$tp+8],$tpj		! tp[j]
+	and	$car0,$mask,$acc0
+	add	$acc1,$car1,$car1
+	srlx	$car0,32,$car0
+	add	$acc0,$car1,$car1
+	st	$car1,[$tp]		! tp[j-1]
+	srlx	$car1,32,$car1
+
+	add	$tpj,$car0,$car0
+	add	$tmp0,$car0,$car0
+	and	$car0,$mask,$acc0
+	add	$tmp1,$car1,$car1
+	add	$acc0,$car1,$car1
+	st	$car1,[$tp+4]		! tp[j-1]
+	srlx	$car0,32,$car0
+	add	$i,4,$i			! i++
+	srlx	$car1,32,$car1
+
+	add	$car0,$car1,$car1
+	cmp	$i,$num
+	add	$car2,$car1,$car1
+	st	$car1,[$tp+8]
+
+	srlx	$car1,32,$car2
+	bl,a	%icc,.Louter
+	ld	[$bp+$i],$mul0		! bp[i]
+!.Louter
+
+	add	$tp,12,$tp
+
+.Ltail:
+	add	$np,$num,$np
+	add	$rp,$num,$rp
+	mov	$tp,$ap
+	sub	%g0,$num,%o7		! k=-num
+	ba	.Lsub
+	subcc	%g0,%g0,%g0		! clear %icc.c
+.align	16
+.Lsub:
+	ld	[$tp+%o7],%o0
+	ld	[$np+%o7],%o1
+	subccc	%o0,%o1,%o1		! tp[j]-np[j]
+	add	$rp,%o7,$i
+	add	%o7,4,%o7
+	brnz	%o7,.Lsub
+	st	%o1,[$i]
+	subc	$car2,0,$car2		! handle upmost overflow bit
+	and	$tp,$car2,$ap
+	andn	$rp,$car2,$np
+	or	$ap,$np,$ap
+	sub	%g0,$num,%o7
+
+.Lcopy:
+	ld	[$ap+%o7],%o0		! copy or in-place refresh
+	st	%g0,[$tp+%o7]		! zap tp
+	st	%o0,[$rp+%o7]
+	add	%o7,4,%o7
+	brnz	%o7,.Lcopy
+	nop
+	mov	1,%i0
+	ret
+	restore
+___
+
+########
+######## .Lbn_sqr_mont gives up to 20% *overall* improvement over
+######## code without following dedicated squaring procedure.
+########
+$sbit="%i2";		# re-use $bp!
+
+$code.=<<___;
+.align	32
+.Lbn_sqr_mont:
+	mulx	$mul0,$mul0,$car0		! ap[0]*ap[0]
+	mulx	$apj,$mul0,$tmp0		!prologue!
+	and	$car0,$mask,$acc0
+	add	%sp,$bias+$frame,$tp
+	ld	[$ap+8],$apj			!prologue!
+
+	mulx	$n0,$acc0,$mul1			! "t[0]"*n0
+	srlx	$car0,32,$car0
+	and	$mul1,$mask,$mul1
+
+	mulx	$car1,$mul1,$car1		! np[0]*"t[0]"*n0
+	mulx	$npj,$mul1,$acc1		!prologue!
+	and	$car0,1,$sbit
+	ld	[$np+8],$npj			!prologue!
+	srlx	$car0,1,$car0
+	add	$acc0,$car1,$car1
+	srlx	$car1,32,$car1
+	mov	$tmp0,$acc0			!prologue!
+
+.Lsqr_1st:
+	mulx	$apj,$mul0,$tmp0
+	mulx	$npj,$mul1,$tmp1
+	add	$acc0,$car0,$car0		! ap[j]*a0+c0
+	add	$acc1,$car1,$car1
+	ld	[$ap+$j],$apj			! ap[j]
+	and	$car0,$mask,$acc0
+	ld	[$np+$j],$npj			! np[j]
+	srlx	$car0,32,$car0
+	add	$acc0,$acc0,$acc0
+	or	$sbit,$acc0,$acc0
+	mov	$tmp1,$acc1
+	srlx	$acc0,32,$sbit
+	add	$j,4,$j				! j++
+	and	$acc0,$mask,$acc0
+	cmp	$j,$num
+	add	$acc0,$car1,$car1
+	st	$car1,[$tp]
+	mov	$tmp0,$acc0
+	srlx	$car1,32,$car1
+	bl	%icc,.Lsqr_1st
+	add	$tp,4,$tp			! tp++
+!.Lsqr_1st
+
+	mulx	$apj,$mul0,$tmp0		! epilogue
+	mulx	$npj,$mul1,$tmp1
+	add	$acc0,$car0,$car0		! ap[j]*a0+c0
+	add	$acc1,$car1,$car1
+	and	$car0,$mask,$acc0
+	srlx	$car0,32,$car0
+	add	$acc0,$acc0,$acc0
+	or	$sbit,$acc0,$acc0
+	srlx	$acc0,32,$sbit
+	and	$acc0,$mask,$acc0
+	add	$acc0,$car1,$car1
+	st	$car1,[$tp]
+	srlx	$car1,32,$car1
+
+	add	$tmp0,$car0,$car0		! ap[j]*a0+c0
+	add	$tmp1,$car1,$car1
+	and	$car0,$mask,$acc0
+	srlx	$car0,32,$car0
+	add	$acc0,$acc0,$acc0
+	or	$sbit,$acc0,$acc0
+	srlx	$acc0,32,$sbit
+	and	$acc0,$mask,$acc0
+	add	$acc0,$car1,$car1
+	st	$car1,[$tp+4]
+	srlx	$car1,32,$car1
+
+	add	$car0,$car0,$car0
+	or	$sbit,$car0,$car0
+	add	$car0,$car1,$car1
+	st	$car1,[$tp+8]
+	srlx	$car1,32,$car2
+
+	ld	[%sp+$bias+$frame],$tmp0	! tp[0]
+	ld	[%sp+$bias+$frame+4],$tmp1	! tp[1]
+	ld	[%sp+$bias+$frame+8],$tpj	! tp[2]
+	ld	[$ap+4],$mul0			! ap[1]
+	ld	[$ap+8],$apj			! ap[2]
+	ld	[$np],$car1			! np[0]
+	ld	[$np+4],$npj			! np[1]
+	mulx	$n0,$tmp0,$mul1
+
+	mulx	$mul0,$mul0,$car0
+	and	$mul1,$mask,$mul1
+
+	mulx	$car1,$mul1,$car1
+	mulx	$npj,$mul1,$acc1
+	add	$tmp0,$car1,$car1
+	and	$car0,$mask,$acc0
+	ld	[$np+8],$npj			! np[2]
+	srlx	$car1,32,$car1
+	add	$tmp1,$car1,$car1
+	srlx	$car0,32,$car0
+	add	$acc0,$car1,$car1
+	and	$car0,1,$sbit
+	add	$acc1,$car1,$car1
+	srlx	$car0,1,$car0
+	mov	12,$j
+	st	$car1,[%sp+$bias+$frame]	! tp[0]=
+	srlx	$car1,32,$car1
+	add	%sp,$bias+$frame+4,$tp
+
+.Lsqr_2nd:
+	mulx	$apj,$mul0,$acc0
+	mulx	$npj,$mul1,$acc1
+	add	$acc0,$car0,$car0
+	add	$tpj,$car1,$car1
+	ld	[$ap+$j],$apj			! ap[j]
+	and	$car0,$mask,$acc0
+	ld	[$np+$j],$npj			! np[j]
+	srlx	$car0,32,$car0
+	add	$acc1,$car1,$car1
+	ld	[$tp+8],$tpj			! tp[j]
+	add	$acc0,$acc0,$acc0
+	add	$j,4,$j				! j++
+	or	$sbit,$acc0,$acc0
+	srlx	$acc0,32,$sbit
+	and	$acc0,$mask,$acc0
+	cmp	$j,$num
+	add	$acc0,$car1,$car1
+	st	$car1,[$tp]			! tp[j-1]
+	srlx	$car1,32,$car1
+	bl	%icc,.Lsqr_2nd
+	add	$tp,4,$tp			! tp++
+!.Lsqr_2nd
+
+	mulx	$apj,$mul0,$acc0
+	mulx	$npj,$mul1,$acc1
+	add	$acc0,$car0,$car0
+	add	$tpj,$car1,$car1
+	and	$car0,$mask,$acc0
+	srlx	$car0,32,$car0
+	add	$acc1,$car1,$car1
+	add	$acc0,$acc0,$acc0
+	or	$sbit,$acc0,$acc0
+	srlx	$acc0,32,$sbit
+	and	$acc0,$mask,$acc0
+	add	$acc0,$car1,$car1
+	st	$car1,[$tp]			! tp[j-1]
+	srlx	$car1,32,$car1
+
+	add	$car0,$car0,$car0
+	or	$sbit,$car0,$car0
+	add	$car0,$car1,$car1
+	add	$car2,$car1,$car1
+	st	$car1,[$tp+4]
+	srlx	$car1,32,$car2
+
+	ld	[%sp+$bias+$frame],$tmp1	! tp[0]
+	ld	[%sp+$bias+$frame+4],$tpj	! tp[1]
+	ld	[$ap+8],$mul0			! ap[2]
+	ld	[$np],$car1			! np[0]
+	ld	[$np+4],$npj			! np[1]
+	mulx	$n0,$tmp1,$mul1
+	and	$mul1,$mask,$mul1
+	mov	8,$i
+
+	mulx	$mul0,$mul0,$car0
+	mulx	$car1,$mul1,$car1
+	and	$car0,$mask,$acc0
+	add	$tmp1,$car1,$car1
+	srlx	$car0,32,$car0
+	add	%sp,$bias+$frame,$tp
+	srlx	$car1,32,$car1
+	and	$car0,1,$sbit
+	srlx	$car0,1,$car0
+	mov	4,$j
+
+.Lsqr_outer:
+.Lsqr_inner1:
+	mulx	$npj,$mul1,$acc1
+	add	$tpj,$car1,$car1
+	add	$j,4,$j
+	ld	[$tp+8],$tpj
+	cmp	$j,$i
+	add	$acc1,$car1,$car1
+	ld	[$np+$j],$npj
+	st	$car1,[$tp]
+	srlx	$car1,32,$car1
+	bl	%icc,.Lsqr_inner1
+	add	$tp,4,$tp
+!.Lsqr_inner1
+
+	add	$j,4,$j
+	ld	[$ap+$j],$apj			! ap[j]
+	mulx	$npj,$mul1,$acc1
+	add	$tpj,$car1,$car1
+	ld	[$np+$j],$npj			! np[j]
+	add	$acc0,$car1,$car1
+	ld	[$tp+8],$tpj			! tp[j]
+	add	$acc1,$car1,$car1
+	st	$car1,[$tp]
+	srlx	$car1,32,$car1
+
+	add	$j,4,$j
+	cmp	$j,$num
+	be,pn	%icc,.Lsqr_no_inner2
+	add	$tp,4,$tp
+
+.Lsqr_inner2:
+	mulx	$apj,$mul0,$acc0
+	mulx	$npj,$mul1,$acc1
+	add	$tpj,$car1,$car1
+	add	$acc0,$car0,$car0
+	ld	[$ap+$j],$apj			! ap[j]
+	and	$car0,$mask,$acc0
+	ld	[$np+$j],$npj			! np[j]
+	srlx	$car0,32,$car0
+	add	$acc0,$acc0,$acc0
+	ld	[$tp+8],$tpj			! tp[j]
+	or	$sbit,$acc0,$acc0
+	add	$j,4,$j				! j++
+	srlx	$acc0,32,$sbit
+	and	$acc0,$mask,$acc0
+	cmp	$j,$num
+	add	$acc0,$car1,$car1
+	add	$acc1,$car1,$car1
+	st	$car1,[$tp]			! tp[j-1]
+	srlx	$car1,32,$car1
+	bl	%icc,.Lsqr_inner2
+	add	$tp,4,$tp			! tp++
+
+.Lsqr_no_inner2:
+	mulx	$apj,$mul0,$acc0
+	mulx	$npj,$mul1,$acc1
+	add	$tpj,$car1,$car1
+	add	$acc0,$car0,$car0
+	and	$car0,$mask,$acc0
+	srlx	$car0,32,$car0
+	add	$acc0,$acc0,$acc0
+	or	$sbit,$acc0,$acc0
+	srlx	$acc0,32,$sbit
+	and	$acc0,$mask,$acc0
+	add	$acc0,$car1,$car1
+	add	$acc1,$car1,$car1
+	st	$car1,[$tp]			! tp[j-1]
+	srlx	$car1,32,$car1
+
+	add	$car0,$car0,$car0
+	or	$sbit,$car0,$car0
+	add	$car0,$car1,$car1
+	add	$car2,$car1,$car1
+	st	$car1,[$tp+4]
+	srlx	$car1,32,$car2
+
+	add	$i,4,$i				! i++
+	ld	[%sp+$bias+$frame],$tmp1	! tp[0]
+	ld	[%sp+$bias+$frame+4],$tpj	! tp[1]
+	ld	[$ap+$i],$mul0			! ap[j]
+	ld	[$np],$car1			! np[0]
+	ld	[$np+4],$npj			! np[1]
+	mulx	$n0,$tmp1,$mul1
+	and	$mul1,$mask,$mul1
+	add	$i,4,$tmp0
+
+	mulx	$mul0,$mul0,$car0
+	mulx	$car1,$mul1,$car1
+	and	$car0,$mask,$acc0
+	add	$tmp1,$car1,$car1
+	srlx	$car0,32,$car0
+	add	%sp,$bias+$frame,$tp
+	srlx	$car1,32,$car1
+	and	$car0,1,$sbit
+	srlx	$car0,1,$car0
+
+	cmp	$tmp0,$num			! i<num-1
+	bl	%icc,.Lsqr_outer
+	mov	4,$j
+
+.Lsqr_last:
+	mulx	$npj,$mul1,$acc1
+	add	$tpj,$car1,$car1
+	add	$j,4,$j
+	ld	[$tp+8],$tpj
+	cmp	$j,$i
+	add	$acc1,$car1,$car1
+	ld	[$np+$j],$npj
+	st	$car1,[$tp]
+	srlx	$car1,32,$car1
+	bl	%icc,.Lsqr_last
+	add	$tp,4,$tp
+!.Lsqr_last
+
+	mulx	$npj,$mul1,$acc1
+	add	$tpj,$car1,$car1
+	add	$acc0,$car1,$car1
+	add	$acc1,$car1,$car1
+	st	$car1,[$tp]
+	srlx	$car1,32,$car1
+
+	add	$car0,$car0,$car0		! recover $car0
+	or	$sbit,$car0,$car0
+	add	$car0,$car1,$car1
+	add	$car2,$car1,$car1
+	st	$car1,[$tp+4]
+	srlx	$car1,32,$car2
+
+	ba	.Ltail
+	add	$tp,8,$tp
+.type	$fname,#function
+.size	$fname,(.-$fname)
+.asciz	"Montgomery Multipltication for SPARCv9, CRYPTOGAMS by <appro\@openssl.org>"
+.align	32
+___
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/sparcv9a-mont.pl b/crypto/bn/asm/sparcv9a-mont.pl
new file mode 100755
index 000000000000..a14205f2f006
--- /dev/null
+++ b/crypto/bn/asm/sparcv9a-mont.pl
@@ -0,0 +1,882 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# October 2005
+#
+# "Teaser" Montgomery multiplication module for UltraSPARC. Why FPU?
+# Because unlike integer multiplier, which simply stalls whole CPU,
+# FPU is fully pipelined and can effectively emit 48 bit partial
+# product every cycle. Why not blended SPARC v9? One can argue that
+# making this module dependent on UltraSPARC VIS extension limits its
+# binary compatibility. Well yes, it does exclude SPARC64 prior-V(!)
+# implementations from compatibility matrix. But the rest, whole Sun
+# UltraSPARC family and brand new Fujitsu's SPARC64 V, all support
+# VIS extension instructions used in this module. This is considered
+# good enough to not care about HAL SPARC64 users [if any] who have
+# integer-only pure SPARCv9 module to "fall down" to.
+
+# USI&II cores currently exhibit uniform 2x improvement [over pre-
+# bn_mul_mont codebase] for all key lengths and benchmarks. On USIII
+# performance improves few percents for shorter keys and worsens few
+# percents for longer keys. This is because USIII integer multiplier
+# is >3x faster than USI&II one, which is harder to match [but see
+# TODO list below]. It should also be noted that SPARC64 V features
+# out-of-order execution, which *might* mean that integer multiplier
+# is pipelined, which in turn *might* be impossible to match... On
+# additional note, SPARC64 V implements FP Multiply-Add instruction,
+# which is perfectly usable in this context... In other words, as far
+# as Fujitsu SPARC64 V goes, talk to the author:-)
+
+# The implementation implies following "non-natural" limitations on
+# input arguments:
+# - num may not be less than 4;
+# - num has to be even;
+# Failure to meet either condition has no fatal effects, simply
+# doesn't give any performance gain.
+
+# TODO:
+# - modulo-schedule inner loop for better performance (on in-order
+#   execution core such as UltraSPARC this shall result in further
+#   noticeable(!) improvement);
+# - dedicated squaring procedure[?];
+
+######################################################################
+# November 2006
+#
+# Modulo-scheduled inner loops allow to interleave floating point and
+# integer instructions and minimize Read-After-Write penalties. This
+# results in *further* 20-50% perfromance improvement [depending on
+# key length, more for longer keys] on USI&II cores and 30-80% - on
+# USIII&IV.
+
+$fname="bn_mul_mont_fpu";
+$bits=32;
+for (@ARGV) { $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+
+if ($bits==64) {
+	$bias=2047;
+	$frame=192;
+} else {
+	$bias=0;
+	$frame=128;	# 96 rounded up to largest known cache-line
+}
+$locals=64;
+
+# In order to provide for 32-/64-bit ABI duality, I keep integers wider
+# than 32 bit in %g1-%g4 and %o0-%o5. %l0-%l7 and %i0-%i5 are used
+# exclusively for pointers, indexes and other small values...
+# int bn_mul_mont(
+$rp="%i0";	# BN_ULONG *rp,
+$ap="%i1";	# const BN_ULONG *ap,
+$bp="%i2";	# const BN_ULONG *bp,
+$np="%i3";	# const BN_ULONG *np,
+$n0="%i4";	# const BN_ULONG *n0,
+$num="%i5";	# int num);
+
+$tp="%l0";	# t[num]
+$ap_l="%l1";	# a[num],n[num] are smashed to 32-bit words and saved
+$ap_h="%l2";	# to these four vectors as double-precision FP values.
+$np_l="%l3";	# This way a bunch of fxtods are eliminated in second
+$np_h="%l4";	# loop and L1-cache aliasing is minimized...
+$i="%l5";
+$j="%l6";
+$mask="%l7";	# 16-bit mask, 0xffff
+
+$n0="%g4";	# reassigned(!) to "64-bit" register
+$carry="%i4";	# %i4 reused(!) for a carry bit
+
+# FP register naming chart
+#
+#     ..HILO
+#       dcba
+#   --------
+#        LOa
+#       LOb
+#      LOc
+#     LOd
+#      HIa
+#     HIb
+#    HIc
+#   HId
+#    ..a
+#   ..b
+$ba="%f0";    $bb="%f2";    $bc="%f4";    $bd="%f6";
+$na="%f8";    $nb="%f10";   $nc="%f12";   $nd="%f14";
+$alo="%f16";  $alo_="%f17"; $ahi="%f18";  $ahi_="%f19";
+$nlo="%f20";  $nlo_="%f21"; $nhi="%f22";  $nhi_="%f23";
+
+$dota="%f24"; $dotb="%f26";
+
+$aloa="%f32"; $alob="%f34"; $aloc="%f36"; $alod="%f38";
+$ahia="%f40"; $ahib="%f42"; $ahic="%f44"; $ahid="%f46";
+$nloa="%f48"; $nlob="%f50"; $nloc="%f52"; $nlod="%f54";
+$nhia="%f56"; $nhib="%f58"; $nhic="%f60"; $nhid="%f62";
+
+$ASI_FL16_P=0xD2;	# magic ASI value to engage 16-bit FP load
+
+$code=<<___;
+.section	".text",#alloc,#execinstr
+
+.global $fname
+.align  32
+$fname:
+	save	%sp,-$frame-$locals,%sp
+
+	cmp	$num,4
+	bl,a,pn %icc,.Lret
+	clr	%i0
+	andcc	$num,1,%g0		! $num has to be even...
+	bnz,a,pn %icc,.Lret
+	clr	%i0			! signal "unsupported input value"
+
+	srl	$num,1,$num
+	sethi	%hi(0xffff),$mask
+	ld	[%i4+0],$n0		! $n0 reassigned, remember?
+	or	$mask,%lo(0xffff),$mask
+	ld	[%i4+4],%o0
+	sllx	%o0,32,%o0
+	or	%o0,$n0,$n0		! $n0=n0[1].n0[0]
+
+	sll	$num,3,$num		! num*=8
+
+	add	%sp,$bias,%o0		! real top of stack
+	sll	$num,2,%o1
+	add	%o1,$num,%o1		! %o1=num*5
+	sub	%o0,%o1,%o0
+	and	%o0,-2048,%o0		! optimize TLB utilization
+	sub	%o0,$bias,%sp		! alloca(5*num*8)
+
+	rd	%asi,%o7		! save %asi
+	add	%sp,$bias+$frame+$locals,$tp
+	add	$tp,$num,$ap_l
+	add	$ap_l,$num,$ap_l	! [an]p_[lh] point at the vectors' ends !
+	add	$ap_l,$num,$ap_h
+	add	$ap_h,$num,$np_l
+	add	$np_l,$num,$np_h
+
+	wr	%g0,$ASI_FL16_P,%asi	! setup %asi for 16-bit FP loads
+
+	add	$rp,$num,$rp		! readjust input pointers to point
+	add	$ap,$num,$ap		! at the ends too...
+	add	$bp,$num,$bp
+	add	$np,$num,$np
+
+	stx	%o7,[%sp+$bias+$frame+48]	! save %asi
+
+	sub	%g0,$num,$i		! i=-num
+	sub	%g0,$num,$j		! j=-num
+
+	add	$ap,$j,%o3
+	add	$bp,$i,%o4
+
+	ld	[%o3+4],%g1		! bp[0]
+	ld	[%o3+0],%o0
+	ld	[%o4+4],%g5		! ap[0]
+	sllx	%g1,32,%g1
+	ld	[%o4+0],%o1
+	sllx	%g5,32,%g5
+	or	%g1,%o0,%o0
+	or	%g5,%o1,%o1
+
+	add	$np,$j,%o5
+
+	mulx	%o1,%o0,%o0		! ap[0]*bp[0]
+	mulx	$n0,%o0,%o0		! ap[0]*bp[0]*n0
+	stx	%o0,[%sp+$bias+$frame+0]
+
+	ld	[%o3+0],$alo_	! load a[j] as pair of 32-bit words
+	fzeros	$alo
+	ld	[%o3+4],$ahi_
+	fzeros	$ahi
+	ld	[%o5+0],$nlo_	! load n[j] as pair of 32-bit words
+	fzeros	$nlo
+	ld	[%o5+4],$nhi_
+	fzeros	$nhi
+
+	! transfer b[i] to FPU as 4x16-bit values
+	ldda	[%o4+2]%asi,$ba
+	fxtod	$alo,$alo
+	ldda	[%o4+0]%asi,$bb
+	fxtod	$ahi,$ahi
+	ldda	[%o4+6]%asi,$bc
+	fxtod	$nlo,$nlo
+	ldda	[%o4+4]%asi,$bd
+	fxtod	$nhi,$nhi
+
+	! transfer ap[0]*b[0]*n0 to FPU as 4x16-bit values
+	ldda	[%sp+$bias+$frame+6]%asi,$na
+	fxtod	$ba,$ba
+	ldda	[%sp+$bias+$frame+4]%asi,$nb
+	fxtod	$bb,$bb
+	ldda	[%sp+$bias+$frame+2]%asi,$nc
+	fxtod	$bc,$bc
+	ldda	[%sp+$bias+$frame+0]%asi,$nd
+	fxtod	$bd,$bd
+
+	std	$alo,[$ap_l+$j]		! save smashed ap[j] in double format
+	fxtod	$na,$na
+	std	$ahi,[$ap_h+$j]
+	fxtod	$nb,$nb
+	std	$nlo,[$np_l+$j]		! save smashed np[j] in double format
+	fxtod	$nc,$nc
+	std	$nhi,[$np_h+$j]
+	fxtod	$nd,$nd
+
+		fmuld	$alo,$ba,$aloa
+		fmuld	$nlo,$na,$nloa
+		fmuld	$alo,$bb,$alob
+		fmuld	$nlo,$nb,$nlob
+		fmuld	$alo,$bc,$aloc
+	faddd	$aloa,$nloa,$nloa
+		fmuld	$nlo,$nc,$nloc
+		fmuld	$alo,$bd,$alod
+	faddd	$alob,$nlob,$nlob
+		fmuld	$nlo,$nd,$nlod
+		fmuld	$ahi,$ba,$ahia
+	faddd	$aloc,$nloc,$nloc
+		fmuld	$nhi,$na,$nhia
+		fmuld	$ahi,$bb,$ahib
+	faddd	$alod,$nlod,$nlod
+		fmuld	$nhi,$nb,$nhib
+		fmuld	$ahi,$bc,$ahic
+	faddd	$ahia,$nhia,$nhia
+		fmuld	$nhi,$nc,$nhic
+		fmuld	$ahi,$bd,$ahid
+	faddd	$ahib,$nhib,$nhib
+		fmuld	$nhi,$nd,$nhid
+
+	faddd	$ahic,$nhic,$dota	! $nhic
+	faddd	$ahid,$nhid,$dotb	! $nhid
+
+	faddd	$nloc,$nhia,$nloc
+	faddd	$nlod,$nhib,$nlod
+
+	fdtox	$nloa,$nloa
+	fdtox	$nlob,$nlob
+	fdtox	$nloc,$nloc
+	fdtox	$nlod,$nlod
+
+	std	$nloa,[%sp+$bias+$frame+0]
+	add	$j,8,$j
+	std	$nlob,[%sp+$bias+$frame+8]
+	add	$ap,$j,%o4
+	std	$nloc,[%sp+$bias+$frame+16]
+	add	$np,$j,%o5
+	std	$nlod,[%sp+$bias+$frame+24]
+
+	ld	[%o4+0],$alo_	! load a[j] as pair of 32-bit words
+	fzeros	$alo
+	ld	[%o4+4],$ahi_
+	fzeros	$ahi
+	ld	[%o5+0],$nlo_	! load n[j] as pair of 32-bit words
+	fzeros	$nlo
+	ld	[%o5+4],$nhi_
+	fzeros	$nhi
+
+	fxtod	$alo,$alo
+	fxtod	$ahi,$ahi
+	fxtod	$nlo,$nlo
+	fxtod	$nhi,$nhi
+
+	ldx	[%sp+$bias+$frame+0],%o0
+		fmuld	$alo,$ba,$aloa
+	ldx	[%sp+$bias+$frame+8],%o1
+		fmuld	$nlo,$na,$nloa
+	ldx	[%sp+$bias+$frame+16],%o2
+		fmuld	$alo,$bb,$alob
+	ldx	[%sp+$bias+$frame+24],%o3
+		fmuld	$nlo,$nb,$nlob
+
+	srlx	%o0,16,%o7
+	std	$alo,[$ap_l+$j]		! save smashed ap[j] in double format
+		fmuld	$alo,$bc,$aloc
+	add	%o7,%o1,%o1
+	std	$ahi,[$ap_h+$j]
+		faddd	$aloa,$nloa,$nloa
+		fmuld	$nlo,$nc,$nloc
+	srlx	%o1,16,%o7
+	std	$nlo,[$np_l+$j]		! save smashed np[j] in double format
+		fmuld	$alo,$bd,$alod
+	add	%o7,%o2,%o2
+	std	$nhi,[$np_h+$j]
+		faddd	$alob,$nlob,$nlob
+		fmuld	$nlo,$nd,$nlod
+	srlx	%o2,16,%o7
+		fmuld	$ahi,$ba,$ahia
+	add	%o7,%o3,%o3		! %o3.%o2[0..15].%o1[0..15].%o0[0..15]
+		faddd	$aloc,$nloc,$nloc
+		fmuld	$nhi,$na,$nhia
+	!and	%o0,$mask,%o0
+	!and	%o1,$mask,%o1
+	!and	%o2,$mask,%o2
+	!sllx	%o1,16,%o1
+	!sllx	%o2,32,%o2
+	!sllx	%o3,48,%o7
+	!or	%o1,%o0,%o0
+	!or	%o2,%o0,%o0
+	!or	%o7,%o0,%o0		! 64-bit result
+	srlx	%o3,16,%g1		! 34-bit carry
+		fmuld	$ahi,$bb,$ahib
+
+	faddd	$alod,$nlod,$nlod
+		fmuld	$nhi,$nb,$nhib
+		fmuld	$ahi,$bc,$ahic
+	faddd	$ahia,$nhia,$nhia
+		fmuld	$nhi,$nc,$nhic
+		fmuld	$ahi,$bd,$ahid
+	faddd	$ahib,$nhib,$nhib
+		fmuld	$nhi,$nd,$nhid
+
+	faddd	$dota,$nloa,$nloa
+	faddd	$dotb,$nlob,$nlob
+	faddd	$ahic,$nhic,$dota	! $nhic
+	faddd	$ahid,$nhid,$dotb	! $nhid
+
+	faddd	$nloc,$nhia,$nloc
+	faddd	$nlod,$nhib,$nlod
+
+	fdtox	$nloa,$nloa
+	fdtox	$nlob,$nlob
+	fdtox	$nloc,$nloc
+	fdtox	$nlod,$nlod
+
+	std	$nloa,[%sp+$bias+$frame+0]
+	std	$nlob,[%sp+$bias+$frame+8]
+	addcc	$j,8,$j
+	std	$nloc,[%sp+$bias+$frame+16]
+	bz,pn	%icc,.L1stskip
+	std	$nlod,[%sp+$bias+$frame+24]
+
+.align	32			! incidentally already aligned !
+.L1st:
+	add	$ap,$j,%o4
+	add	$np,$j,%o5
+	ld	[%o4+0],$alo_	! load a[j] as pair of 32-bit words
+	fzeros	$alo
+	ld	[%o4+4],$ahi_
+	fzeros	$ahi
+	ld	[%o5+0],$nlo_	! load n[j] as pair of 32-bit words
+	fzeros	$nlo
+	ld	[%o5+4],$nhi_
+	fzeros	$nhi
+
+	fxtod	$alo,$alo
+	fxtod	$ahi,$ahi
+	fxtod	$nlo,$nlo
+	fxtod	$nhi,$nhi
+
+	ldx	[%sp+$bias+$frame+0],%o0
+		fmuld	$alo,$ba,$aloa
+	ldx	[%sp+$bias+$frame+8],%o1
+		fmuld	$nlo,$na,$nloa
+	ldx	[%sp+$bias+$frame+16],%o2
+		fmuld	$alo,$bb,$alob
+	ldx	[%sp+$bias+$frame+24],%o3
+		fmuld	$nlo,$nb,$nlob
+
+	srlx	%o0,16,%o7
+	std	$alo,[$ap_l+$j]		! save smashed ap[j] in double format
+		fmuld	$alo,$bc,$aloc
+	add	%o7,%o1,%o1
+	std	$ahi,[$ap_h+$j]
+		faddd	$aloa,$nloa,$nloa
+		fmuld	$nlo,$nc,$nloc
+	srlx	%o1,16,%o7
+	std	$nlo,[$np_l+$j]		! save smashed np[j] in double format
+		fmuld	$alo,$bd,$alod
+	add	%o7,%o2,%o2
+	std	$nhi,[$np_h+$j]
+		faddd	$alob,$nlob,$nlob
+		fmuld	$nlo,$nd,$nlod
+	srlx	%o2,16,%o7
+		fmuld	$ahi,$ba,$ahia
+	add	%o7,%o3,%o3		! %o3.%o2[0..15].%o1[0..15].%o0[0..15]
+	and	%o0,$mask,%o0
+		faddd	$aloc,$nloc,$nloc
+		fmuld	$nhi,$na,$nhia
+	and	%o1,$mask,%o1
+	and	%o2,$mask,%o2
+		fmuld	$ahi,$bb,$ahib
+	sllx	%o1,16,%o1
+		faddd	$alod,$nlod,$nlod
+		fmuld	$nhi,$nb,$nhib
+	sllx	%o2,32,%o2
+		fmuld	$ahi,$bc,$ahic
+	sllx	%o3,48,%o7
+	or	%o1,%o0,%o0
+		faddd	$ahia,$nhia,$nhia
+		fmuld	$nhi,$nc,$nhic
+	or	%o2,%o0,%o0
+		fmuld	$ahi,$bd,$ahid
+	or	%o7,%o0,%o0		! 64-bit result
+		faddd	$ahib,$nhib,$nhib
+		fmuld	$nhi,$nd,$nhid
+	addcc	%g1,%o0,%o0
+		faddd	$dota,$nloa,$nloa
+	srlx	%o3,16,%g1		! 34-bit carry
+		faddd	$dotb,$nlob,$nlob
+	bcs,a	%xcc,.+8
+	add	%g1,1,%g1
+
+	stx	%o0,[$tp]		! tp[j-1]=
+
+	faddd	$ahic,$nhic,$dota	! $nhic
+	faddd	$ahid,$nhid,$dotb	! $nhid
+
+	faddd	$nloc,$nhia,$nloc
+	faddd	$nlod,$nhib,$nlod
+
+	fdtox	$nloa,$nloa
+	fdtox	$nlob,$nlob
+	fdtox	$nloc,$nloc
+	fdtox	$nlod,$nlod
+
+	std	$nloa,[%sp+$bias+$frame+0]
+	std	$nlob,[%sp+$bias+$frame+8]
+	std	$nloc,[%sp+$bias+$frame+16]
+	std	$nlod,[%sp+$bias+$frame+24]
+
+	addcc	$j,8,$j
+	bnz,pt	%icc,.L1st
+	add	$tp,8,$tp
+
+.L1stskip:
+	fdtox	$dota,$dota
+	fdtox	$dotb,$dotb
+
+	ldx	[%sp+$bias+$frame+0],%o0
+	ldx	[%sp+$bias+$frame+8],%o1
+	ldx	[%sp+$bias+$frame+16],%o2
+	ldx	[%sp+$bias+$frame+24],%o3
+
+	srlx	%o0,16,%o7
+	std	$dota,[%sp+$bias+$frame+32]
+	add	%o7,%o1,%o1
+	std	$dotb,[%sp+$bias+$frame+40]
+	srlx	%o1,16,%o7
+	add	%o7,%o2,%o2
+	srlx	%o2,16,%o7
+	add	%o7,%o3,%o3		! %o3.%o2[0..15].%o1[0..15].%o0[0..15]
+	and	%o0,$mask,%o0
+	and	%o1,$mask,%o1
+	and	%o2,$mask,%o2
+	sllx	%o1,16,%o1
+	sllx	%o2,32,%o2
+	sllx	%o3,48,%o7
+	or	%o1,%o0,%o0
+	or	%o2,%o0,%o0
+	or	%o7,%o0,%o0		! 64-bit result
+	ldx	[%sp+$bias+$frame+32],%o4
+	addcc	%g1,%o0,%o0
+	ldx	[%sp+$bias+$frame+40],%o5
+	srlx	%o3,16,%g1		! 34-bit carry
+	bcs,a	%xcc,.+8
+	add	%g1,1,%g1
+
+	stx	%o0,[$tp]		! tp[j-1]=
+	add	$tp,8,$tp
+
+	srlx	%o4,16,%o7
+	add	%o7,%o5,%o5
+	and	%o4,$mask,%o4
+	sllx	%o5,16,%o7
+	or	%o7,%o4,%o4
+	addcc	%g1,%o4,%o4
+	srlx	%o5,48,%g1
+	bcs,a	%xcc,.+8
+	add	%g1,1,%g1
+
+	mov	%g1,$carry
+	stx	%o4,[$tp]		! tp[num-1]=
+
+	ba	.Louter
+	add	$i,8,$i
+.align	32
+.Louter:
+	sub	%g0,$num,$j		! j=-num
+	add	%sp,$bias+$frame+$locals,$tp
+
+	add	$ap,$j,%o3
+	add	$bp,$i,%o4
+
+	ld	[%o3+4],%g1		! bp[i]
+	ld	[%o3+0],%o0
+	ld	[%o4+4],%g5		! ap[0]
+	sllx	%g1,32,%g1
+	ld	[%o4+0],%o1
+	sllx	%g5,32,%g5
+	or	%g1,%o0,%o0
+	or	%g5,%o1,%o1
+
+	ldx	[$tp],%o2		! tp[0]
+	mulx	%o1,%o0,%o0
+	addcc	%o2,%o0,%o0
+	mulx	$n0,%o0,%o0		! (ap[0]*bp[i]+t[0])*n0
+	stx	%o0,[%sp+$bias+$frame+0]
+
+	! transfer b[i] to FPU as 4x16-bit values
+	ldda	[%o4+2]%asi,$ba
+	ldda	[%o4+0]%asi,$bb
+	ldda	[%o4+6]%asi,$bc
+	ldda	[%o4+4]%asi,$bd
+
+	! transfer (ap[0]*b[i]+t[0])*n0 to FPU as 4x16-bit values
+	ldda	[%sp+$bias+$frame+6]%asi,$na
+	fxtod	$ba,$ba
+	ldda	[%sp+$bias+$frame+4]%asi,$nb
+	fxtod	$bb,$bb
+	ldda	[%sp+$bias+$frame+2]%asi,$nc
+	fxtod	$bc,$bc
+	ldda	[%sp+$bias+$frame+0]%asi,$nd
+	fxtod	$bd,$bd
+	ldd	[$ap_l+$j],$alo		! load a[j] in double format
+	fxtod	$na,$na
+	ldd	[$ap_h+$j],$ahi
+	fxtod	$nb,$nb
+	ldd	[$np_l+$j],$nlo		! load n[j] in double format
+	fxtod	$nc,$nc
+	ldd	[$np_h+$j],$nhi
+	fxtod	$nd,$nd
+
+		fmuld	$alo,$ba,$aloa
+		fmuld	$nlo,$na,$nloa
+		fmuld	$alo,$bb,$alob
+		fmuld	$nlo,$nb,$nlob
+		fmuld	$alo,$bc,$aloc
+	faddd	$aloa,$nloa,$nloa
+		fmuld	$nlo,$nc,$nloc
+		fmuld	$alo,$bd,$alod
+	faddd	$alob,$nlob,$nlob
+		fmuld	$nlo,$nd,$nlod
+		fmuld	$ahi,$ba,$ahia
+	faddd	$aloc,$nloc,$nloc
+		fmuld	$nhi,$na,$nhia
+		fmuld	$ahi,$bb,$ahib
+	faddd	$alod,$nlod,$nlod
+		fmuld	$nhi,$nb,$nhib
+		fmuld	$ahi,$bc,$ahic
+	faddd	$ahia,$nhia,$nhia
+		fmuld	$nhi,$nc,$nhic
+		fmuld	$ahi,$bd,$ahid
+	faddd	$ahib,$nhib,$nhib
+		fmuld	$nhi,$nd,$nhid
+
+	faddd	$ahic,$nhic,$dota	! $nhic
+	faddd	$ahid,$nhid,$dotb	! $nhid
+
+	faddd	$nloc,$nhia,$nloc
+	faddd	$nlod,$nhib,$nlod
+
+	fdtox	$nloa,$nloa
+	fdtox	$nlob,$nlob
+	fdtox	$nloc,$nloc
+	fdtox	$nlod,$nlod
+
+	std	$nloa,[%sp+$bias+$frame+0]
+	std	$nlob,[%sp+$bias+$frame+8]
+	std	$nloc,[%sp+$bias+$frame+16]
+	add	$j,8,$j
+	std	$nlod,[%sp+$bias+$frame+24]
+
+	ldd	[$ap_l+$j],$alo		! load a[j] in double format
+	ldd	[$ap_h+$j],$ahi
+	ldd	[$np_l+$j],$nlo		! load n[j] in double format
+	ldd	[$np_h+$j],$nhi
+
+		fmuld	$alo,$ba,$aloa
+		fmuld	$nlo,$na,$nloa
+		fmuld	$alo,$bb,$alob
+		fmuld	$nlo,$nb,$nlob
+		fmuld	$alo,$bc,$aloc
+	ldx	[%sp+$bias+$frame+0],%o0
+		faddd	$aloa,$nloa,$nloa
+		fmuld	$nlo,$nc,$nloc
+	ldx	[%sp+$bias+$frame+8],%o1
+		fmuld	$alo,$bd,$alod
+	ldx	[%sp+$bias+$frame+16],%o2
+		faddd	$alob,$nlob,$nlob
+		fmuld	$nlo,$nd,$nlod
+	ldx	[%sp+$bias+$frame+24],%o3
+		fmuld	$ahi,$ba,$ahia
+
+	srlx	%o0,16,%o7
+		faddd	$aloc,$nloc,$nloc
+		fmuld	$nhi,$na,$nhia
+	add	%o7,%o1,%o1
+		fmuld	$ahi,$bb,$ahib
+	srlx	%o1,16,%o7
+		faddd	$alod,$nlod,$nlod
+		fmuld	$nhi,$nb,$nhib
+	add	%o7,%o2,%o2
+		fmuld	$ahi,$bc,$ahic
+	srlx	%o2,16,%o7
+		faddd	$ahia,$nhia,$nhia
+		fmuld	$nhi,$nc,$nhic
+	add	%o7,%o3,%o3		! %o3.%o2[0..15].%o1[0..15].%o0[0..15]
+	! why?
+	and	%o0,$mask,%o0
+		fmuld	$ahi,$bd,$ahid
+	and	%o1,$mask,%o1
+	and	%o2,$mask,%o2
+		faddd	$ahib,$nhib,$nhib
+		fmuld	$nhi,$nd,$nhid
+	sllx	%o1,16,%o1
+		faddd	$dota,$nloa,$nloa
+	sllx	%o2,32,%o2
+		faddd	$dotb,$nlob,$nlob
+	sllx	%o3,48,%o7
+	or	%o1,%o0,%o0
+		faddd	$ahic,$nhic,$dota	! $nhic
+	or	%o2,%o0,%o0
+		faddd	$ahid,$nhid,$dotb	! $nhid
+	or	%o7,%o0,%o0		! 64-bit result
+	ldx	[$tp],%o7
+		faddd	$nloc,$nhia,$nloc
+	addcc	%o7,%o0,%o0
+	! end-of-why?
+		faddd	$nlod,$nhib,$nlod
+	srlx	%o3,16,%g1		! 34-bit carry
+		fdtox	$nloa,$nloa
+	bcs,a	%xcc,.+8
+	add	%g1,1,%g1
+
+	fdtox	$nlob,$nlob
+	fdtox	$nloc,$nloc
+	fdtox	$nlod,$nlod
+
+	std	$nloa,[%sp+$bias+$frame+0]
+	std	$nlob,[%sp+$bias+$frame+8]
+	addcc	$j,8,$j
+	std	$nloc,[%sp+$bias+$frame+16]
+	bz,pn	%icc,.Linnerskip
+	std	$nlod,[%sp+$bias+$frame+24]
+
+	ba	.Linner
+	nop
+.align	32
+.Linner:
+	ldd	[$ap_l+$j],$alo		! load a[j] in double format
+	ldd	[$ap_h+$j],$ahi
+	ldd	[$np_l+$j],$nlo		! load n[j] in double format
+	ldd	[$np_h+$j],$nhi
+
+		fmuld	$alo,$ba,$aloa
+		fmuld	$nlo,$na,$nloa
+		fmuld	$alo,$bb,$alob
+		fmuld	$nlo,$nb,$nlob
+		fmuld	$alo,$bc,$aloc
+	ldx	[%sp+$bias+$frame+0],%o0
+		faddd	$aloa,$nloa,$nloa
+		fmuld	$nlo,$nc,$nloc
+	ldx	[%sp+$bias+$frame+8],%o1
+		fmuld	$alo,$bd,$alod
+	ldx	[%sp+$bias+$frame+16],%o2
+		faddd	$alob,$nlob,$nlob
+		fmuld	$nlo,$nd,$nlod
+	ldx	[%sp+$bias+$frame+24],%o3
+		fmuld	$ahi,$ba,$ahia
+
+	srlx	%o0,16,%o7
+		faddd	$aloc,$nloc,$nloc
+		fmuld	$nhi,$na,$nhia
+	add	%o7,%o1,%o1
+		fmuld	$ahi,$bb,$ahib
+	srlx	%o1,16,%o7
+		faddd	$alod,$nlod,$nlod
+		fmuld	$nhi,$nb,$nhib
+	add	%o7,%o2,%o2
+		fmuld	$ahi,$bc,$ahic
+	srlx	%o2,16,%o7
+		faddd	$ahia,$nhia,$nhia
+		fmuld	$nhi,$nc,$nhic
+	add	%o7,%o3,%o3		! %o3.%o2[0..15].%o1[0..15].%o0[0..15]
+	and	%o0,$mask,%o0
+		fmuld	$ahi,$bd,$ahid
+	and	%o1,$mask,%o1
+	and	%o2,$mask,%o2
+		faddd	$ahib,$nhib,$nhib
+		fmuld	$nhi,$nd,$nhid
+	sllx	%o1,16,%o1
+		faddd	$dota,$nloa,$nloa
+	sllx	%o2,32,%o2
+		faddd	$dotb,$nlob,$nlob
+	sllx	%o3,48,%o7
+	or	%o1,%o0,%o0
+		faddd	$ahic,$nhic,$dota	! $nhic
+	or	%o2,%o0,%o0
+		faddd	$ahid,$nhid,$dotb	! $nhid
+	or	%o7,%o0,%o0		! 64-bit result
+		faddd	$nloc,$nhia,$nloc
+	addcc	%g1,%o0,%o0
+	ldx	[$tp+8],%o7		! tp[j]
+		faddd	$nlod,$nhib,$nlod
+	srlx	%o3,16,%g1		! 34-bit carry
+		fdtox	$nloa,$nloa
+	bcs,a	%xcc,.+8
+	add	%g1,1,%g1
+		fdtox	$nlob,$nlob
+	addcc	%o7,%o0,%o0
+		fdtox	$nloc,$nloc
+	bcs,a	%xcc,.+8
+	add	%g1,1,%g1
+
+	stx	%o0,[$tp]		! tp[j-1]
+		fdtox	$nlod,$nlod
+
+	std	$nloa,[%sp+$bias+$frame+0]
+	std	$nlob,[%sp+$bias+$frame+8]
+	std	$nloc,[%sp+$bias+$frame+16]
+	addcc	$j,8,$j
+	std	$nlod,[%sp+$bias+$frame+24]
+	bnz,pt	%icc,.Linner
+	add	$tp,8,$tp
+
+.Linnerskip:
+	fdtox	$dota,$dota
+	fdtox	$dotb,$dotb
+
+	ldx	[%sp+$bias+$frame+0],%o0
+	ldx	[%sp+$bias+$frame+8],%o1
+	ldx	[%sp+$bias+$frame+16],%o2
+	ldx	[%sp+$bias+$frame+24],%o3
+
+	srlx	%o0,16,%o7
+	std	$dota,[%sp+$bias+$frame+32]
+	add	%o7,%o1,%o1
+	std	$dotb,[%sp+$bias+$frame+40]
+	srlx	%o1,16,%o7
+	add	%o7,%o2,%o2
+	srlx	%o2,16,%o7
+	add	%o7,%o3,%o3		! %o3.%o2[0..15].%o1[0..15].%o0[0..15]
+	and	%o0,$mask,%o0
+	and	%o1,$mask,%o1
+	and	%o2,$mask,%o2
+	sllx	%o1,16,%o1
+	sllx	%o2,32,%o2
+	sllx	%o3,48,%o7
+	or	%o1,%o0,%o0
+	or	%o2,%o0,%o0
+	ldx	[%sp+$bias+$frame+32],%o4
+	or	%o7,%o0,%o0		! 64-bit result
+	ldx	[%sp+$bias+$frame+40],%o5
+	addcc	%g1,%o0,%o0
+	ldx	[$tp+8],%o7		! tp[j]
+	srlx	%o3,16,%g1		! 34-bit carry
+	bcs,a	%xcc,.+8
+	add	%g1,1,%g1
+
+	addcc	%o7,%o0,%o0
+	bcs,a	%xcc,.+8
+	add	%g1,1,%g1
+
+	stx	%o0,[$tp]		! tp[j-1]
+	add	$tp,8,$tp
+
+	srlx	%o4,16,%o7
+	add	%o7,%o5,%o5
+	and	%o4,$mask,%o4
+	sllx	%o5,16,%o7
+	or	%o7,%o4,%o4
+	addcc	%g1,%o4,%o4
+	srlx	%o5,48,%g1
+	bcs,a	%xcc,.+8
+	add	%g1,1,%g1
+
+	addcc	$carry,%o4,%o4
+	stx	%o4,[$tp]		! tp[num-1]
+	mov	%g1,$carry
+	bcs,a	%xcc,.+8
+	add	$carry,1,$carry
+
+	addcc	$i,8,$i
+	bnz	%icc,.Louter
+	nop
+
+	add	$tp,8,$tp		! adjust tp to point at the end
+	orn	%g0,%g0,%g4
+	sub	%g0,$num,%o7		! n=-num
+	ba	.Lsub
+	subcc	%g0,%g0,%g0		! clear %icc.c
+
+.align	32
+.Lsub:
+	ldx	[$tp+%o7],%o0
+	add	$np,%o7,%g1
+	ld	[%g1+0],%o2
+	ld	[%g1+4],%o3
+	srlx	%o0,32,%o1
+	subccc	%o0,%o2,%o2
+	add	$rp,%o7,%g1
+	subccc	%o1,%o3,%o3
+	st	%o2,[%g1+0]
+	add	%o7,8,%o7
+	brnz,pt	%o7,.Lsub
+	st	%o3,[%g1+4]
+	subc	$carry,0,%g4
+	sub	%g0,$num,%o7		! n=-num
+	ba	.Lcopy
+	nop
+
+.align	32
+.Lcopy:
+	ldx	[$tp+%o7],%o0
+	add	$rp,%o7,%g1
+	ld	[%g1+0],%o2
+	ld	[%g1+4],%o3
+	stx	%g0,[$tp+%o7]
+	and	%o0,%g4,%o0
+	srlx	%o0,32,%o1
+	andn	%o2,%g4,%o2
+	andn	%o3,%g4,%o3
+	or	%o2,%o0,%o0
+	or	%o3,%o1,%o1
+	st	%o0,[%g1+0]
+	add	%o7,8,%o7
+	brnz,pt	%o7,.Lcopy
+	st	%o1,[%g1+4]
+	sub	%g0,$num,%o7		! n=-num
+
+.Lzap:
+	stx	%g0,[$ap_l+%o7]
+	stx	%g0,[$ap_h+%o7]
+	stx	%g0,[$np_l+%o7]
+	stx	%g0,[$np_h+%o7]
+	add	%o7,8,%o7
+	brnz,pt	%o7,.Lzap
+	nop
+
+	ldx	[%sp+$bias+$frame+48],%o7
+	wr	%g0,%o7,%asi		! restore %asi
+
+	mov	1,%i0
+.Lret:
+	ret
+	restore
+.type   $fname,#function
+.size	$fname,(.-$fname)
+.asciz	"Montgomery Multipltication for UltraSPARC, CRYPTOGAMS by <appro\@openssl.org>"
+.align	32
+___
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+
+# Below substitution makes it possible to compile without demanding
+# VIS extentions on command line, e.g. -xarch=v9 vs. -xarch=v9a. I
+# dare to do this, because VIS capability is detected at run-time now
+# and this routine is not called on CPU not capable to execute it. Do
+# note that fzeros is not the only VIS dependency! Another dependency
+# is implicit and is just _a_ numerical value loaded to %asi register,
+# which assembler can't recognize as VIS specific...
+$code =~ s/fzeros\s+%f([0-9]+)/
+	   sprintf(".word\t0x%x\t! fzeros %%f%d",0x81b00c20|($1<<25),$1)
+	  /gem;
+
+print $code;
+# flush
+close STDOUT;
diff --git a/crypto/bn/asm/via-mont.pl b/crypto/bn/asm/via-mont.pl
new file mode 100755
index 000000000000..c046a514c873
--- /dev/null
+++ b/crypto/bn/asm/via-mont.pl
@@ -0,0 +1,242 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# Wrapper around 'rep montmul', VIA-specific instruction accessing
+# PadLock Montgomery Multiplier. The wrapper is designed as drop-in
+# replacement for OpenSSL bn_mul_mont [first implemented in 0.9.9].
+#
+# Below are interleaved outputs from 'openssl speed rsa dsa' for 4
+# different software configurations on 1.5GHz VIA Esther processor.
+# Lines marked with "software integer" denote performance of hand-
+# coded integer-only assembler found in OpenSSL 0.9.7. "Software SSE2"
+# refers to hand-coded SSE2 Montgomery multiplication procedure found
+# OpenSSL 0.9.9. "Hardware VIA SDK" refers to padlock_pmm routine from
+# Padlock SDK 2.0.1 available for download from VIA, which naturally
+# utilizes the magic 'repz montmul' instruction. And finally "hardware
+# this" refers to *this* implementation which also uses 'repz montmul'
+#
+#                   sign    verify    sign/s verify/s
+# rsa  512 bits 0.001720s 0.000140s    581.4   7149.7	software integer
+# rsa  512 bits 0.000690s 0.000086s   1450.3  11606.0	software SSE2
+# rsa  512 bits 0.006136s 0.000201s    163.0   4974.5	hardware VIA SDK
+# rsa  512 bits 0.000712s 0.000050s   1404.9  19858.5	hardware this
+#
+# rsa 1024 bits 0.008518s 0.000413s    117.4   2420.8	software integer
+# rsa 1024 bits 0.004275s 0.000277s    233.9   3609.7	software SSE2
+# rsa 1024 bits 0.012136s 0.000260s     82.4   3844.5	hardware VIA SDK
+# rsa 1024 bits 0.002522s 0.000116s    396.5   8650.9	hardware this
+#
+# rsa 2048 bits 0.050101s 0.001371s     20.0    729.6	software integer
+# rsa 2048 bits 0.030273s 0.001008s     33.0    991.9	software SSE2
+# rsa 2048 bits 0.030833s 0.000976s     32.4   1025.1	hardware VIA SDK
+# rsa 2048 bits 0.011879s 0.000342s     84.2   2921.7	hardware this
+#
+# rsa 4096 bits 0.327097s 0.004859s      3.1    205.8	software integer
+# rsa 4096 bits 0.229318s 0.003859s      4.4    259.2	software SSE2
+# rsa 4096 bits 0.233953s 0.003274s      4.3    305.4	hardware VIA SDK
+# rsa 4096 bits 0.070493s 0.001166s     14.2    857.6	hardware this
+#
+# dsa  512 bits 0.001342s 0.001651s    745.2    605.7	software integer
+# dsa  512 bits 0.000844s 0.000987s   1185.3   1013.1	software SSE2
+# dsa  512 bits 0.001902s 0.002247s    525.6    444.9	hardware VIA SDK
+# dsa  512 bits 0.000458s 0.000524s   2182.2   1909.1	hardware this
+#
+# dsa 1024 bits 0.003964s 0.004926s    252.3    203.0	software integer
+# dsa 1024 bits 0.002686s 0.003166s    372.3    315.8	software SSE2
+# dsa 1024 bits 0.002397s 0.002823s    417.1    354.3	hardware VIA SDK
+# dsa 1024 bits 0.000978s 0.001170s   1022.2    855.0	hardware this
+#
+# dsa 2048 bits 0.013280s 0.016518s     75.3     60.5	software integer
+# dsa 2048 bits 0.009911s 0.011522s    100.9     86.8	software SSE2
+# dsa 2048 bits 0.009542s 0.011763s    104.8     85.0	hardware VIA SDK
+# dsa 2048 bits 0.002884s 0.003352s    346.8    298.3	hardware this
+#
+# To give you some other reference point here is output for 2.4GHz P4
+# running hand-coded SSE2 bn_mul_mont found in 0.9.9, i.e. "software
+# SSE2" in above terms.
+#
+# rsa  512 bits 0.000407s 0.000047s   2454.2  21137.0
+# rsa 1024 bits 0.002426s 0.000141s    412.1   7100.0
+# rsa 2048 bits 0.015046s 0.000491s     66.5   2034.9
+# rsa 4096 bits 0.109770s 0.002379s      9.1    420.3
+# dsa  512 bits 0.000438s 0.000525s   2281.1   1904.1
+# dsa 1024 bits 0.001346s 0.001595s    742.7    627.0
+# dsa 2048 bits 0.004745s 0.005582s    210.7    179.1
+#
+# Conclusions: 
+# - VIA SDK leaves a *lot* of room for improvement (which this
+#   implementation successfully fills:-);
+# - 'rep montmul' gives up to >3x performance improvement depending on
+#   key length;
+# - in terms of absolute performance it delivers approximately as much
+#   as modern out-of-order 32-bit cores [again, for longer keys].
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"via-mont.pl");
+
+# int bn_mul_mont(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np,const BN_ULONG *n0, int num);
+$func="bn_mul_mont_padlock";
+
+$pad=16*1;	# amount of reserved bytes on top of every vector
+
+# stack layout
+$mZeroPrime=&DWP(0,"esp");		# these are specified by VIA
+$A=&DWP(4,"esp");
+$B=&DWP(8,"esp");
+$T=&DWP(12,"esp");
+$M=&DWP(16,"esp");
+$scratch=&DWP(20,"esp");
+$rp=&DWP(24,"esp");			# these are mine
+$sp=&DWP(28,"esp");
+# &DWP(32,"esp")			# 32 byte scratch area
+# &DWP(64+(4*$num+$pad)*0,"esp")	# padded tp[num]
+# &DWP(64+(4*$num+$pad)*1,"esp")	# padded copy of ap[num]
+# &DWP(64+(4*$num+$pad)*2,"esp")	# padded copy of bp[num]
+# &DWP(64+(4*$num+$pad)*3,"esp")	# padded copy of np[num]
+# Note that SDK suggests to unconditionally allocate 2K per vector. This
+# has quite an impact on performance. It naturally depends on key length,
+# but to give an example 1024 bit private RSA key operations suffer >30%
+# penalty. I allocate only as much as actually required...
+
+&function_begin($func);
+	&xor	("eax","eax");
+	&mov	("ecx",&wparam(5));	# num
+	# meet VIA's limitations for num [note that the specification
+	# expresses them in bits, while we work with amount of 32-bit words]
+	&test	("ecx",3);
+	&jnz	(&label("leave"));	# num % 4 != 0
+	&cmp	("ecx",8);
+	&jb	(&label("leave"));	# num < 8
+	&cmp	("ecx",1024);
+	&ja	(&label("leave"));	# num > 1024
+
+	&pushf	();
+	&cld	();
+
+	&mov	("edi",&wparam(0));	# rp
+	&mov	("eax",&wparam(1));	# ap
+	&mov	("ebx",&wparam(2));	# bp
+	&mov	("edx",&wparam(3));	# np
+	&mov	("esi",&wparam(4));	# n0
+	&mov	("esi",&DWP(0,"esi"));	# *n0
+
+	&lea	("ecx",&DWP($pad,"","ecx",4));	# ecx becomes vector size in bytes
+	&lea	("ebp",&DWP(64,"","ecx",4));	# allocate 4 vectors + 64 bytes
+	&neg	("ebp");
+	&add	("ebp","esp");
+	&and	("ebp",-64);		# align to cache-line
+	&xchg	("ebp","esp");		# alloca
+
+	&mov	($rp,"edi");		# save rp
+	&mov	($sp,"ebp");		# save esp
+
+	&mov	($mZeroPrime,"esi");
+	&lea	("esi",&DWP(64,"esp"));	# tp
+	&mov	($T,"esi");
+	&lea	("edi",&DWP(32,"esp"));	# scratch area
+	&mov	($scratch,"edi");
+	&mov	("esi","eax");
+
+	&lea	("ebp",&DWP(-$pad,"ecx"));
+	&shr	("ebp",2);		# restore original num value in ebp
+
+	&xor	("eax","eax");
+
+	&mov	("ecx","ebp");
+	&lea	("ecx",&DWP((32+$pad)/4,"ecx"));# padded tp + scratch
+	&data_byte(0xf3,0xab);		# rep stosl, bzero
+
+	&mov	("ecx","ebp");
+	&lea	("edi",&DWP(64+$pad,"esp","ecx",4));# pointer to ap copy
+	&mov	($A,"edi");
+	&data_byte(0xf3,0xa5);		# rep movsl, memcpy
+	&mov	("ecx",$pad/4);
+	&data_byte(0xf3,0xab);		# rep stosl, bzero pad
+	# edi points at the end of padded ap copy...
+
+	&mov	("ecx","ebp");
+	&mov	("esi","ebx");
+	&mov	($B,"edi");
+	&data_byte(0xf3,0xa5);		# rep movsl, memcpy
+	&mov	("ecx",$pad/4);
+	&data_byte(0xf3,0xab);		# rep stosl, bzero pad
+	# edi points at the end of padded bp copy...
+
+	&mov	("ecx","ebp");
+	&mov	("esi","edx");
+	&mov	($M,"edi");
+	&data_byte(0xf3,0xa5);		# rep movsl, memcpy
+	&mov	("ecx",$pad/4);
+	&data_byte(0xf3,0xab);		# rep stosl, bzero pad
+	# edi points at the end of padded np copy...
+
+	# let magic happen...
+	&mov	("ecx","ebp");
+	&mov	("esi","esp");
+	&shl	("ecx",5);		# convert word counter to bit counter
+	&align	(4);
+	&data_byte(0xf3,0x0f,0xa6,0xc0);# rep montmul
+
+	&mov	("ecx","ebp");
+	&lea	("esi",&DWP(64,"esp"));		# tp
+	# edi still points at the end of padded np copy...
+	&neg	("ebp");
+	&lea	("ebp",&DWP(-$pad,"edi","ebp",4));	# so just "rewind"
+	&mov	("edi",$rp);			# restore rp
+	&xor	("edx","edx");			# i=0 and clear CF
+
+&set_label("sub",8);
+	&mov	("eax",&DWP(0,"esi","edx",4));
+	&sbb	("eax",&DWP(0,"ebp","edx",4));
+	&mov	(&DWP(0,"edi","edx",4),"eax");	# rp[i]=tp[i]-np[i]
+	&lea	("edx",&DWP(1,"edx"));		# i++
+	&loop	(&label("sub"));		# doesn't affect CF!
+
+	&mov	("eax",&DWP(0,"esi","edx",4));	# upmost overflow bit
+	&sbb	("eax",0);
+	&and	("esi","eax");
+	&not	("eax");
+	&mov	("ebp","edi");
+	&and	("ebp","eax");
+	&or	("esi","ebp");			# tp=carry?tp:rp
+
+	&mov	("ecx","edx");			# num
+	&xor	("edx","edx");			# i=0
+
+&set_label("copy",8);
+	&mov	("eax",&DWP(0,"esi","edx",4));
+	&mov	(&DWP(64,"esp","edx",4),"ecx");	# zap tp
+	&mov	(&DWP(0,"edi","edx",4),"eax");
+	&lea	("edx",&DWP(1,"edx"));		# i++
+	&loop	(&label("copy"));
+
+	&mov	("ebp",$sp);
+	&xor	("eax","eax");
+
+	&mov	("ecx",64/4);
+	&mov	("edi","esp");		# zap frame including scratch area
+	&data_byte(0xf3,0xab);		# rep stosl, bzero
+
+	# zap copies of ap, bp and np
+	&lea	("edi",&DWP(64+$pad,"esp","edx",4));# pointer to ap
+	&lea	("ecx",&DWP(3*$pad/4,"edx","edx",2));
+	&data_byte(0xf3,0xab);		# rep stosl, bzero
+
+	&mov	("esp","ebp");
+	&inc	("eax");		# signal "done"
+	&popf	();
+&set_label("leave");
+&function_end($func);
+
+&asciz("Padlock Montgomery Multiplication, CRYPTOGAMS by <appro\@openssl.org>");
+
+&asm_finish();
diff --git a/crypto/bn/asm/x86-gf2m.pl b/crypto/bn/asm/x86-gf2m.pl
new file mode 100755
index 000000000000..808a1e59691d
--- /dev/null
+++ b/crypto/bn/asm/x86-gf2m.pl
@@ -0,0 +1,313 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# May 2011
+#
+# The module implements bn_GF2m_mul_2x2 polynomial multiplication used
+# in bn_gf2m.c. It's kind of low-hanging mechanical port from C for
+# the time being... Except that it has three code paths: pure integer
+# code suitable for any x86 CPU, MMX code suitable for PIII and later
+# and PCLMULQDQ suitable for Westmere and later. Improvement varies
+# from one benchmark and µ-arch to another. Below are interval values
+# for 163- and 571-bit ECDH benchmarks relative to compiler-generated
+# code:
+#
+# PIII		16%-30%
+# P4		12%-12%
+# Opteron	18%-40%
+# Core2		19%-44%
+# Atom		38%-64%
+# Westmere	53%-121%(PCLMULQDQ)/20%-32%(MMX)
+# Sandy Bridge	72%-127%(PCLMULQDQ)/27%-23%(MMX)
+#
+# Note that above improvement coefficients are not coefficients for
+# bn_GF2m_mul_2x2 itself. For example 120% ECDH improvement is result
+# of bn_GF2m_mul_2x2 being >4x faster. As it gets faster, benchmark
+# is more and more dominated by other subroutines, most notably by
+# BN_GF2m_mod[_mul]_arr...
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],$0,$x86only = $ARGV[$#ARGV] eq "386");
+
+$sse2=0;
+for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); }
+
+&external_label("OPENSSL_ia32cap_P") if ($sse2);
+
+$a="eax";
+$b="ebx";
+($a1,$a2,$a4)=("ecx","edx","ebp");
+
+$R="mm0";
+@T=("mm1","mm2");
+($A,$B,$B30,$B31)=("mm2","mm3","mm4","mm5");
+@i=("esi","edi");
+
+					if (!$x86only) {
+&function_begin_B("_mul_1x1_mmx");
+	&sub	("esp",32+4);
+	 &mov	($a1,$a);
+	 &lea	($a2,&DWP(0,$a,$a));
+	 &and	($a1,0x3fffffff);
+	 &lea	($a4,&DWP(0,$a2,$a2));
+	 &mov	(&DWP(0*4,"esp"),0);
+	 &and	($a2,0x7fffffff);
+	&movd	($A,$a);
+	&movd	($B,$b);
+	 &mov	(&DWP(1*4,"esp"),$a1);	# a1
+	 &xor	($a1,$a2);		# a1^a2
+	&pxor	($B31,$B31);
+	&pxor	($B30,$B30);
+	 &mov	(&DWP(2*4,"esp"),$a2);	# a2
+	 &xor	($a2,$a4);		# a2^a4
+	 &mov	(&DWP(3*4,"esp"),$a1);	# a1^a2
+	&pcmpgtd($B31,$A);		# broadcast 31st bit
+	&paddd	($A,$A);		# $A<<=1
+	 &xor	($a1,$a2);		# a1^a4=a1^a2^a2^a4
+	 &mov	(&DWP(4*4,"esp"),$a4);	# a4
+	 &xor	($a4,$a2);		# a2=a4^a2^a4
+	&pand	($B31,$B);
+	&pcmpgtd($B30,$A);		# broadcast 30th bit
+	 &mov	(&DWP(5*4,"esp"),$a1);	# a1^a4
+	 &xor	($a4,$a1);		# a1^a2^a4
+	&psllq	($B31,31);
+	&pand	($B30,$B);
+	 &mov	(&DWP(6*4,"esp"),$a2);	# a2^a4
+	&mov	(@i[0],0x7);
+	 &mov	(&DWP(7*4,"esp"),$a4);	# a1^a2^a4
+	 &mov	($a4,@i[0]);
+	&and	(@i[0],$b);
+	&shr	($b,3);
+	&mov	(@i[1],$a4);
+	&psllq	($B30,30);
+	&and	(@i[1],$b);
+	&shr	($b,3);
+	&movd	($R,&DWP(0,"esp",@i[0],4));
+	&mov	(@i[0],$a4);
+	&and	(@i[0],$b);
+	&shr	($b,3);
+	for($n=1;$n<9;$n++) {
+		&movd	(@T[1],&DWP(0,"esp",@i[1],4));
+		&mov	(@i[1],$a4);
+		&psllq	(@T[1],3*$n);
+		&and	(@i[1],$b);
+		&shr	($b,3);
+		&pxor	($R,@T[1]);
+
+		push(@i,shift(@i)); push(@T,shift(@T));
+	}
+	&movd	(@T[1],&DWP(0,"esp",@i[1],4));
+	&pxor	($R,$B30);
+	&psllq	(@T[1],3*$n++);
+	&pxor	($R,@T[1]);
+
+	&movd	(@T[0],&DWP(0,"esp",@i[0],4));
+	&pxor	($R,$B31);
+	&psllq	(@T[0],3*$n);
+	&add	("esp",32+4);
+	&pxor	($R,@T[0]);
+	&ret	();
+&function_end_B("_mul_1x1_mmx");
+					}
+
+($lo,$hi)=("eax","edx");
+@T=("ecx","ebp");
+
+&function_begin_B("_mul_1x1_ialu");
+	&sub	("esp",32+4);
+	 &mov	($a1,$a);
+	 &lea	($a2,&DWP(0,$a,$a));
+	 &lea	($a4,&DWP(0,"",$a,4));
+	 &and	($a1,0x3fffffff);
+	&lea	(@i[1],&DWP(0,$lo,$lo));
+	&sar	($lo,31);		# broadcast 31st bit
+	 &mov	(&DWP(0*4,"esp"),0);
+	 &and	($a2,0x7fffffff);
+	 &mov	(&DWP(1*4,"esp"),$a1);	# a1
+	 &xor	($a1,$a2);		# a1^a2
+	 &mov	(&DWP(2*4,"esp"),$a2);	# a2
+	 &xor	($a2,$a4);		# a2^a4
+	 &mov	(&DWP(3*4,"esp"),$a1);	# a1^a2
+	 &xor	($a1,$a2);		# a1^a4=a1^a2^a2^a4
+	 &mov	(&DWP(4*4,"esp"),$a4);	# a4
+	 &xor	($a4,$a2);		# a2=a4^a2^a4
+	 &mov	(&DWP(5*4,"esp"),$a1);	# a1^a4
+	 &xor	($a4,$a1);		# a1^a2^a4
+	&sar	(@i[1],31);		# broardcast 30th bit
+	&and	($lo,$b);
+	 &mov	(&DWP(6*4,"esp"),$a2);	# a2^a4
+	&and	(@i[1],$b);
+	 &mov	(&DWP(7*4,"esp"),$a4);	# a1^a2^a4
+	&mov	($hi,$lo);
+	&shl	($lo,31);
+	&mov	(@T[0],@i[1]);
+	&shr	($hi,1);
+
+	 &mov	(@i[0],0x7);
+	&shl	(@i[1],30);
+	 &and	(@i[0],$b);
+	&shr	(@T[0],2);
+	&xor	($lo,@i[1]);
+
+	&shr	($b,3);
+	&mov	(@i[1],0x7);		# 5-byte instruction!?
+	&and	(@i[1],$b);
+	&shr	($b,3);
+	 &xor	($hi,@T[0]);
+	&xor	($lo,&DWP(0,"esp",@i[0],4));
+	&mov	(@i[0],0x7);
+	&and	(@i[0],$b);
+	&shr	($b,3);
+	for($n=1;$n<9;$n++) {
+		&mov	(@T[1],&DWP(0,"esp",@i[1],4));
+		&mov	(@i[1],0x7);
+		&mov	(@T[0],@T[1]);
+		&shl	(@T[1],3*$n);
+		&and	(@i[1],$b);
+		&shr	(@T[0],32-3*$n);
+		&xor	($lo,@T[1]);
+		&shr	($b,3);
+		&xor	($hi,@T[0]);
+
+		push(@i,shift(@i)); push(@T,shift(@T));
+	}
+	&mov	(@T[1],&DWP(0,"esp",@i[1],4));
+	&mov	(@T[0],@T[1]);
+	&shl	(@T[1],3*$n);
+	&mov	(@i[1],&DWP(0,"esp",@i[0],4));
+	&shr	(@T[0],32-3*$n);	$n++;
+	&mov	(@i[0],@i[1]);
+	&xor	($lo,@T[1]);
+	&shl	(@i[1],3*$n);
+	&xor	($hi,@T[0]);
+	&shr	(@i[0],32-3*$n);
+	&xor	($lo,@i[1]);
+	&xor	($hi,@i[0]);
+
+	&add	("esp",32+4);
+	&ret	();
+&function_end_B("_mul_1x1_ialu");
+
+# void bn_GF2m_mul_2x2(BN_ULONG *r, BN_ULONG a1, BN_ULONG a0, BN_ULONG b1, BN_ULONG b0);
+&function_begin_B("bn_GF2m_mul_2x2");
+if (!$x86only) {
+	&picmeup("edx","OPENSSL_ia32cap_P");
+	&mov	("eax",&DWP(0,"edx"));
+	&mov	("edx",&DWP(4,"edx"));
+	&test	("eax",1<<23);		# check MMX bit
+	&jz	(&label("ialu"));
+if ($sse2) {
+	&test	("eax",1<<24);		# check FXSR bit
+	&jz	(&label("mmx"));
+	&test	("edx",1<<1);		# check PCLMULQDQ bit
+	&jz	(&label("mmx"));
+
+	&movups		("xmm0",&QWP(8,"esp"));
+	&shufps		("xmm0","xmm0",0b10110001);
+	&pclmulqdq	("xmm0","xmm0",1);
+	&mov		("eax",&DWP(4,"esp"));
+	&movups		(&QWP(0,"eax"),"xmm0");
+	&ret	();
+
+&set_label("mmx",16);
+}
+	&push	("ebp");
+	&push	("ebx");
+	&push	("esi");
+	&push	("edi");
+	&mov	($a,&wparam(1));
+	&mov	($b,&wparam(3));
+	&call	("_mul_1x1_mmx");	# a1·b1
+	&movq	("mm7",$R);
+
+	&mov	($a,&wparam(2));
+	&mov	($b,&wparam(4));
+	&call	("_mul_1x1_mmx");	# a0·b0
+	&movq	("mm6",$R);
+
+	&mov	($a,&wparam(1));
+	&mov	($b,&wparam(3));
+	&xor	($a,&wparam(2));
+	&xor	($b,&wparam(4));
+	&call	("_mul_1x1_mmx");	# (a0+a1)·(b0+b1)
+	&pxor	($R,"mm7");
+	&mov	($a,&wparam(0));
+	&pxor	($R,"mm6");		# (a0+a1)·(b0+b1)-a1·b1-a0·b0
+
+	&movq	($A,$R);
+	&psllq	($R,32);
+	&pop	("edi");
+	&psrlq	($A,32);
+	&pop	("esi");
+	&pxor	($R,"mm6");
+	&pop	("ebx");
+	&pxor	($A,"mm7");
+	&movq	(&QWP(0,$a),$R);
+	&pop	("ebp");
+	&movq	(&QWP(8,$a),$A);
+	&emms	();
+	&ret	();
+&set_label("ialu",16);
+}
+	&push	("ebp");
+	&push	("ebx");
+	&push	("esi");
+	&push	("edi");
+	&stack_push(4+1);
+
+	&mov	($a,&wparam(1));
+	&mov	($b,&wparam(3));
+	&call	("_mul_1x1_ialu");	# a1·b1
+	&mov	(&DWP(8,"esp"),$lo);
+	&mov	(&DWP(12,"esp"),$hi);
+
+	&mov	($a,&wparam(2));
+	&mov	($b,&wparam(4));
+	&call	("_mul_1x1_ialu");	# a0·b0
+	&mov	(&DWP(0,"esp"),$lo);
+	&mov	(&DWP(4,"esp"),$hi);
+
+	&mov	($a,&wparam(1));
+	&mov	($b,&wparam(3));
+	&xor	($a,&wparam(2));
+	&xor	($b,&wparam(4));
+	&call	("_mul_1x1_ialu");	# (a0+a1)·(b0+b1)
+
+	&mov	("ebp",&wparam(0));
+		 @r=("ebx","ecx","edi","esi");
+	&mov	(@r[0],&DWP(0,"esp"));
+	&mov	(@r[1],&DWP(4,"esp"));
+	&mov	(@r[2],&DWP(8,"esp"));
+	&mov	(@r[3],&DWP(12,"esp"));
+
+	&xor	($lo,$hi);
+	&xor	($hi,@r[1]);
+	&xor	($lo,@r[0]);
+	&mov	(&DWP(0,"ebp"),@r[0]);
+	&xor	($hi,@r[2]);
+	&mov	(&DWP(12,"ebp"),@r[3]);
+	&xor	($lo,@r[3]);
+	&stack_pop(4+1);
+	&xor	($hi,@r[3]);
+	&pop	("edi");
+	&xor	($lo,$hi);
+	&pop	("esi");
+	&mov	(&DWP(8,"ebp"),$hi);
+	&pop	("ebx");
+	&mov	(&DWP(4,"ebp"),$lo);
+	&pop	("ebp");
+	&ret	();
+&function_end_B("bn_GF2m_mul_2x2");
+
+&asciz	("GF(2^m) Multiplication for x86, CRYPTOGAMS by <appro\@openssl.org>");
+
+&asm_finish();
diff --git a/crypto/bn/asm/mo-586.pl b/crypto/bn/asm/x86-mont.pl
index 061127e0b11d..e8f6b050842e 100755
--- a/crypto/bn/asm/mo-586.pl
+++ b/crypto/bn/asm/x86-mont.pl
@@ -1,18 +1,5 @@
 #!/usr/bin/env perl
 
-# This is crypto/bn/asm/x86-mont.pl (with asciz from crypto/perlasm/x86asm.pl)
-# from OpenSSL 0.9.9-dev 
-
-sub ::asciz
-{ my @str=unpack("C*",shift);
-    push @str,0;
-    while ($#str>15) {
-	&data_byte(@str[0..15]);
-	foreach (0..15) { shift @str; }
-    }
-    &data_byte(@str) if (@str);
-}
-
 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
 # project. The module is, however, dual licensed under OpenSSL and
@@ -39,7 +26,8 @@ sub ::asciz
 # Integer-only code [being equipped with dedicated squaring procedure]
 # gives ~40% on rsa512 sign benchmark...
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],$0);
diff --git a/crypto/bn/asm/x86_64-gcc.c b/crypto/bn/asm/x86_64-gcc.c
index b1b8a1109bf7..acb0b401181e 100644
--- a/crypto/bn/asm/x86_64-gcc.c
+++ b/crypto/bn/asm/x86_64-gcc.c
@@ -1,5 +1,5 @@
 #include "../bn_lcl.h"
-#ifdef __SUNPRO_C
+#if !(defined(__GNUC__) && __GNUC__>=2)
 # include "../bn_asm.c"	/* kind of dirty hack for Sun Studio */
 #else
 /*
@@ -55,7 +55,11 @@
  *    machine.
  */
 
+#ifdef _WIN64
+#define BN_ULONG unsigned long long
+#else
 #define BN_ULONG unsigned long
+#endif
 
 #undef mul
 #undef mul_add
@@ -187,7 +191,7 @@ BN_ULONG bn_add_words (BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,int
 
 	asm (
 	"	subq	%2,%2		\n"
-	".align 16			\n"
+	".p2align 4			\n"
 	"1:	movq	(%4,%2,8),%0	\n"
 	"	adcq	(%5,%2,8),%0	\n"
 	"	movq	%0,(%3,%2,8)	\n"
@@ -210,7 +214,7 @@ BN_ULONG bn_sub_words (BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,int
 
 	asm (
 	"	subq	%2,%2		\n"
-	".align 16			\n"
+	".p2align 4			\n"
 	"1:	movq	(%4,%2,8),%0	\n"
 	"	sbbq	(%5,%2,8),%0	\n"
 	"	movq	%0,(%3,%2,8)	\n"
diff --git a/crypto/bn/asm/x86_64-gf2m.pl b/crypto/bn/asm/x86_64-gf2m.pl
new file mode 100755
index 000000000000..1658acbbddd6
--- /dev/null
+++ b/crypto/bn/asm/x86_64-gf2m.pl
@@ -0,0 +1,389 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# May 2011
+#
+# The module implements bn_GF2m_mul_2x2 polynomial multiplication used
+# in bn_gf2m.c. It's kind of low-hanging mechanical port from C for
+# the time being... Except that it has two code paths: code suitable
+# for any x86_64 CPU and PCLMULQDQ one suitable for Westmere and
+# later. Improvement varies from one benchmark and µ-arch to another.
+# Vanilla code path is at most 20% faster than compiler-generated code
+# [not very impressive], while PCLMULQDQ - whole 85%-160% better on
+# 163- and 571-bit ECDH benchmarks on Intel CPUs. Keep in mind that
+# these coefficients are not ones for bn_GF2m_mul_2x2 itself, as not
+# all CPU time is burnt in it...
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+($lo,$hi)=("%rax","%rdx");	$a=$lo;
+($i0,$i1)=("%rsi","%rdi");
+($t0,$t1)=("%rbx","%rcx");
+($b,$mask)=("%rbp","%r8");
+($a1,$a2,$a4,$a8,$a12,$a48)=map("%r$_",(9..15));
+($R,$Tx)=("%xmm0","%xmm1");
+
+$code.=<<___;
+.text
+
+.type	_mul_1x1,\@abi-omnipotent
+.align	16
+_mul_1x1:
+	sub	\$128+8,%rsp
+	mov	\$-1,$a1
+	lea	($a,$a),$i0
+	shr	\$3,$a1
+	lea	(,$a,4),$i1
+	and	$a,$a1			# a1=a&0x1fffffffffffffff
+	lea	(,$a,8),$a8
+	sar	\$63,$a			# broadcast 63rd bit
+	lea	($a1,$a1),$a2
+	sar	\$63,$i0		# broadcast 62nd bit
+	lea	(,$a1,4),$a4
+	and	$b,$a
+	sar	\$63,$i1		# boardcast 61st bit
+	mov	$a,$hi			# $a is $lo
+	shl	\$63,$lo
+	and	$b,$i0
+	shr	\$1,$hi
+	mov	$i0,$t1
+	shl	\$62,$i0
+	and	$b,$i1
+	shr	\$2,$t1
+	xor	$i0,$lo
+	mov	$i1,$t0
+	shl	\$61,$i1
+	xor	$t1,$hi
+	shr	\$3,$t0
+	xor	$i1,$lo
+	xor	$t0,$hi
+
+	mov	$a1,$a12
+	movq	\$0,0(%rsp)		# tab[0]=0
+	xor	$a2,$a12		# a1^a2
+	mov	$a1,8(%rsp)		# tab[1]=a1
+	 mov	$a4,$a48
+	mov	$a2,16(%rsp)		# tab[2]=a2
+	 xor	$a8,$a48		# a4^a8
+	mov	$a12,24(%rsp)		# tab[3]=a1^a2
+
+	xor	$a4,$a1
+	mov	$a4,32(%rsp)		# tab[4]=a4
+	xor	$a4,$a2
+	mov	$a1,40(%rsp)		# tab[5]=a1^a4
+	xor	$a4,$a12
+	mov	$a2,48(%rsp)		# tab[6]=a2^a4
+	 xor	$a48,$a1		# a1^a4^a4^a8=a1^a8
+	mov	$a12,56(%rsp)		# tab[7]=a1^a2^a4
+	 xor	$a48,$a2		# a2^a4^a4^a8=a1^a8
+
+	mov	$a8,64(%rsp)		# tab[8]=a8
+	xor	$a48,$a12		# a1^a2^a4^a4^a8=a1^a2^a8
+	mov	$a1,72(%rsp)		# tab[9]=a1^a8
+	 xor	$a4,$a1			# a1^a8^a4
+	mov	$a2,80(%rsp)		# tab[10]=a2^a8
+	 xor	$a4,$a2			# a2^a8^a4
+	mov	$a12,88(%rsp)		# tab[11]=a1^a2^a8
+
+	xor	$a4,$a12		# a1^a2^a8^a4
+	mov	$a48,96(%rsp)		# tab[12]=a4^a8
+	 mov	$mask,$i0
+	mov	$a1,104(%rsp)		# tab[13]=a1^a4^a8
+	 and	$b,$i0
+	mov	$a2,112(%rsp)		# tab[14]=a2^a4^a8
+	 shr	\$4,$b
+	mov	$a12,120(%rsp)		# tab[15]=a1^a2^a4^a8
+	 mov	$mask,$i1
+	 and	$b,$i1
+	 shr	\$4,$b
+
+	movq	(%rsp,$i0,8),$R		# half of calculations is done in SSE2
+	mov	$mask,$i0
+	and	$b,$i0
+	shr	\$4,$b
+___
+    for ($n=1;$n<8;$n++) {
+	$code.=<<___;
+	mov	(%rsp,$i1,8),$t1
+	mov	$mask,$i1
+	mov	$t1,$t0
+	shl	\$`8*$n-4`,$t1
+	and	$b,$i1
+	 movq	(%rsp,$i0,8),$Tx
+	shr	\$`64-(8*$n-4)`,$t0
+	xor	$t1,$lo
+	 pslldq	\$$n,$Tx
+	 mov	$mask,$i0
+	shr	\$4,$b
+	xor	$t0,$hi
+	 and	$b,$i0
+	 shr	\$4,$b
+	 pxor	$Tx,$R
+___
+    }
+$code.=<<___;
+	mov	(%rsp,$i1,8),$t1
+	mov	$t1,$t0
+	shl	\$`8*$n-4`,$t1
+	movq	$R,$i0
+	shr	\$`64-(8*$n-4)`,$t0
+	xor	$t1,$lo
+	psrldq	\$8,$R
+	xor	$t0,$hi
+	movq	$R,$i1
+	xor	$i0,$lo
+	xor	$i1,$hi
+
+	add	\$128+8,%rsp
+	ret
+.Lend_mul_1x1:
+.size	_mul_1x1,.-_mul_1x1
+___
+
+($rp,$a1,$a0,$b1,$b0) = $win64?	("%rcx","%rdx","%r8", "%r9","%r10") :	# Win64 order
+				("%rdi","%rsi","%rdx","%rcx","%r8");	# Unix order
+
+$code.=<<___;
+.extern	OPENSSL_ia32cap_P
+.globl	bn_GF2m_mul_2x2
+.type	bn_GF2m_mul_2x2,\@abi-omnipotent
+.align	16
+bn_GF2m_mul_2x2:
+	mov	OPENSSL_ia32cap_P(%rip),%rax
+	bt	\$33,%rax
+	jnc	.Lvanilla_mul_2x2
+
+	movq		$a1,%xmm0
+	movq		$b1,%xmm1
+	movq		$a0,%xmm2
+___
+$code.=<<___ if ($win64);
+	movq		40(%rsp),%xmm3
+___
+$code.=<<___ if (!$win64);
+	movq		$b0,%xmm3
+___
+$code.=<<___;
+	movdqa		%xmm0,%xmm4
+	movdqa		%xmm1,%xmm5
+	pclmulqdq	\$0,%xmm1,%xmm0	# a1·b1
+	pxor		%xmm2,%xmm4
+	pxor		%xmm3,%xmm5
+	pclmulqdq	\$0,%xmm3,%xmm2	# a0·b0
+	pclmulqdq	\$0,%xmm5,%xmm4	# (a0+a1)·(b0+b1)
+	xorps		%xmm0,%xmm4
+	xorps		%xmm2,%xmm4	# (a0+a1)·(b0+b1)-a0·b0-a1·b1
+	movdqa		%xmm4,%xmm5
+	pslldq		\$8,%xmm4
+	psrldq		\$8,%xmm5
+	pxor		%xmm4,%xmm2
+	pxor		%xmm5,%xmm0
+	movdqu		%xmm2,0($rp)
+	movdqu		%xmm0,16($rp)
+	ret
+
+.align	16
+.Lvanilla_mul_2x2:
+	lea	-8*17(%rsp),%rsp
+___
+$code.=<<___ if ($win64);
+	mov	`8*17+40`(%rsp),$b0
+	mov	%rdi,8*15(%rsp)
+	mov	%rsi,8*16(%rsp)
+___
+$code.=<<___;
+	mov	%r14,8*10(%rsp)
+	mov	%r13,8*11(%rsp)
+	mov	%r12,8*12(%rsp)
+	mov	%rbp,8*13(%rsp)
+	mov	%rbx,8*14(%rsp)
+.Lbody_mul_2x2:
+	mov	$rp,32(%rsp)		# save the arguments
+	mov	$a1,40(%rsp)
+	mov	$a0,48(%rsp)
+	mov	$b1,56(%rsp)
+	mov	$b0,64(%rsp)
+
+	mov	\$0xf,$mask
+	mov	$a1,$a
+	mov	$b1,$b
+	call	_mul_1x1		# a1·b1
+	mov	$lo,16(%rsp)
+	mov	$hi,24(%rsp)
+
+	mov	48(%rsp),$a
+	mov	64(%rsp),$b
+	call	_mul_1x1		# a0·b0
+	mov	$lo,0(%rsp)
+	mov	$hi,8(%rsp)
+
+	mov	40(%rsp),$a
+	mov	56(%rsp),$b
+	xor	48(%rsp),$a
+	xor	64(%rsp),$b
+	call	_mul_1x1		# (a0+a1)·(b0+b1)
+___
+	@r=("%rbx","%rcx","%rdi","%rsi");
+$code.=<<___;
+	mov	0(%rsp),@r[0]
+	mov	8(%rsp),@r[1]
+	mov	16(%rsp),@r[2]
+	mov	24(%rsp),@r[3]
+	mov	32(%rsp),%rbp
+
+	xor	$hi,$lo
+	xor	@r[1],$hi
+	xor	@r[0],$lo
+	mov	@r[0],0(%rbp)
+	xor	@r[2],$hi
+	mov	@r[3],24(%rbp)
+	xor	@r[3],$lo
+	xor	@r[3],$hi
+	xor	$hi,$lo
+	mov	$hi,16(%rbp)
+	mov	$lo,8(%rbp)
+
+	mov	8*10(%rsp),%r14
+	mov	8*11(%rsp),%r13
+	mov	8*12(%rsp),%r12
+	mov	8*13(%rsp),%rbp
+	mov	8*14(%rsp),%rbx
+___
+$code.=<<___ if ($win64);
+	mov	8*15(%rsp),%rdi
+	mov	8*16(%rsp),%rsi
+___
+$code.=<<___;
+	lea	8*17(%rsp),%rsp
+	ret
+.Lend_mul_2x2:
+.size	bn_GF2m_mul_2x2,.-bn_GF2m_mul_2x2
+.asciz	"GF(2^m) Multiplication for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+.align	16
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#               CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern __imp_RtlVirtualUnwind
+
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	152($context),%rax	# pull context->Rsp
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lbody_mul_2x2(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<"prologue" label
+	jb	.Lin_prologue
+
+	mov	8*10(%rax),%r14		# mimic epilogue
+	mov	8*11(%rax),%r13
+	mov	8*12(%rax),%r12
+	mov	8*13(%rax),%rbp
+	mov	8*14(%rax),%rbx
+	mov	8*15(%rax),%rdi
+	mov	8*16(%rax),%rsi
+
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+
+.Lin_prologue:
+	lea	8*17(%rax),%rax
+	mov	%rax,152($context)	# restore context->Rsp
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	se_handler,.-se_handler
+
+.section	.pdata
+.align	4
+	.rva	_mul_1x1
+	.rva	.Lend_mul_1x1
+	.rva	.LSEH_info_1x1
+
+	.rva	.Lvanilla_mul_2x2
+	.rva	.Lend_mul_2x2
+	.rva	.LSEH_info_2x2
+.section	.xdata
+.align	8
+.LSEH_info_1x1:
+	.byte	0x01,0x07,0x02,0x00
+	.byte	0x07,0x01,0x11,0x00	# sub rsp,128+8
+.LSEH_info_2x2:
+	.byte	9,0,0,0
+	.rva	se_handler
+___
+}
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/bn/asm/x86_64-mont.pl b/crypto/bn/asm/x86_64-mont.pl
index c43b69592a5c..5d79b35e1cf2 100755
--- a/crypto/bn/asm/x86_64-mont.pl
+++ b/crypto/bn/asm/x86_64-mont.pl
@@ -1,7 +1,7 @@
 #!/usr/bin/env perl
 
 # ====================================================================
-# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
 # project. The module is, however, dual licensed under OpenSSL and
 # CRYPTOGAMS licenses depending on where you obtain it. For further
 # details see http://www.openssl.org/~appro/cryptogams/.
@@ -15,14 +15,32 @@
 # respectful 50%. It remains to be seen if loop unrolling and
 # dedicated squaring routine can provide further improvement...
 
-$output=shift;
+# July 2011.
+#
+# Add dedicated squaring procedure. Performance improvement varies
+# from platform to platform, but in average it's ~5%/15%/25%/33%
+# for 512-/1024-/2048-/4096-bit RSA *sign* benchmarks respectively.
+
+# August 2011.
+#
+# Unroll and modulo-schedule inner loops in such manner that they
+# are "fallen through" for input lengths of 8, which is critical for
+# 1024-bit RSA *sign*. Average performance improvement in comparison
+# to *initial* version of this module from 2005 is ~0%/30%/40%/45%
+# for 512-/1024-/2048-/4096-bit RSA *sign* benchmarks respectively.
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
 
 $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
 ( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
 ( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
 die "can't locate x86_64-xlate.pl";
 
-open STDOUT,"| $^X $xlate $output";
+open STDOUT,"| $^X $xlate $flavour $output";
 
 # int bn_mul_mont(
 $rp="%rdi";	# BN_ULONG *rp,
@@ -33,7 +51,6 @@ $n0="%r8";	# const BN_ULONG *n0,
 $num="%r9";	# int num);
 $lo0="%r10";
 $hi0="%r11";
-$bp="%r12";	# reassign $bp
 $hi1="%r13";
 $i="%r14";
 $j="%r15";
@@ -47,6 +64,16 @@ $code=<<___;
 .type	bn_mul_mont,\@function,6
 .align	16
 bn_mul_mont:
+	test	\$3,${num}d
+	jnz	.Lmul_enter
+	cmp	\$8,${num}d
+	jb	.Lmul_enter
+	cmp	$ap,$bp
+	jne	.Lmul4x_enter
+	jmp	.Lsqr4x_enter
+
+.align	16
+.Lmul_enter:
 	push	%rbx
 	push	%rbp
 	push	%r12
@@ -55,54 +82,73 @@ bn_mul_mont:
 	push	%r15
 
 	mov	${num}d,${num}d
-	lea	2($num),%rax
-	mov	%rsp,%rbp
-	neg	%rax
-	lea	(%rsp,%rax,8),%rsp	# tp=alloca(8*(num+2))
+	lea	2($num),%r10
+	mov	%rsp,%r11
+	neg	%r10
+	lea	(%rsp,%r10,8),%rsp	# tp=alloca(8*(num+2))
 	and	\$-1024,%rsp		# minimize TLB usage
 
-	mov	%rbp,8(%rsp,$num,8)	# tp[num+1]=%rsp
-	mov	%rdx,$bp		# $bp reassigned, remember?
-
+	mov	%r11,8(%rsp,$num,8)	# tp[num+1]=%rsp
+.Lmul_body:
+	mov	$bp,%r12		# reassign $bp
+___
+		$bp="%r12";
+$code.=<<___;
 	mov	($n0),$n0		# pull n0[0] value
+	mov	($bp),$m0		# m0=bp[0]
+	mov	($ap),%rax
 
 	xor	$i,$i			# i=0
 	xor	$j,$j			# j=0
 
-	mov	($bp),$m0		# m0=bp[0]
-	mov	($ap),%rax
+	mov	$n0,$m1
 	mulq	$m0			# ap[0]*bp[0]
 	mov	%rax,$lo0
-	mov	%rdx,$hi0
+	mov	($np),%rax
 
-	imulq	$n0,%rax		# "tp[0]"*n0
-	mov	%rax,$m1
+	imulq	$lo0,$m1		# "tp[0]"*n0
+	mov	%rdx,$hi0
 
-	mulq	($np)			# np[0]*m1
-	add	$lo0,%rax		# discarded
+	mulq	$m1			# np[0]*m1
+	add	%rax,$lo0		# discarded
+	mov	8($ap),%rax
 	adc	\$0,%rdx
 	mov	%rdx,$hi1
 
 	lea	1($j),$j		# j++
+	jmp	.L1st_enter
+
+.align	16
 .L1st:
+	add	%rax,$hi1
 	mov	($ap,$j,8),%rax
-	mulq	$m0			# ap[j]*bp[0]
-	add	$hi0,%rax
 	adc	\$0,%rdx
-	mov	%rax,$lo0
+	add	$hi0,$hi1		# np[j]*m1+ap[j]*bp[0]
+	mov	$lo0,$hi0
+	adc	\$0,%rdx
+	mov	$hi1,-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$hi1
+
+.L1st_enter:
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$hi0
 	mov	($np,$j,8),%rax
-	mov	%rdx,$hi0
+	adc	\$0,%rdx
+	lea	1($j),$j		# j++
+	mov	%rdx,$lo0
 
 	mulq	$m1			# np[j]*m1
-	add	$hi1,%rax
-	lea	1($j),$j		# j++
+	cmp	$num,$j
+	jne	.L1st
+
+	add	%rax,$hi1
+	mov	($ap),%rax		# ap[0]
 	adc	\$0,%rdx
-	add	$lo0,%rax		# np[j]*m1+ap[j]*bp[0]
+	add	$hi0,$hi1		# np[j]*m1+ap[j]*bp[0]
 	adc	\$0,%rdx
-	mov	%rax,-16(%rsp,$j,8)	# tp[j-1]
-	cmp	$num,$j
+	mov	$hi1,-16(%rsp,$j,8)	# tp[j-1]
 	mov	%rdx,$hi1
-	jl	.L1st
+	mov	$lo0,$hi0
 
 	xor	%rdx,%rdx
 	add	$hi0,$hi1
@@ -111,50 +157,64 @@ bn_mul_mont:
 	mov	%rdx,(%rsp,$num,8)	# store upmost overflow bit
 
 	lea	1($i),$i		# i++
-.align	4
+	jmp	.Louter
+.align	16
 .Louter:
-	xor	$j,$j			# j=0
-
 	mov	($bp,$i,8),$m0		# m0=bp[i]
-	mov	($ap),%rax		# ap[0]
+	xor	$j,$j			# j=0
+	mov	$n0,$m1
+	mov	(%rsp),$lo0
 	mulq	$m0			# ap[0]*bp[i]
-	add	(%rsp),%rax		# ap[0]*bp[i]+tp[0]
+	add	%rax,$lo0		# ap[0]*bp[i]+tp[0]
+	mov	($np),%rax
 	adc	\$0,%rdx
-	mov	%rax,$lo0
-	mov	%rdx,$hi0
 
-	imulq	$n0,%rax		# tp[0]*n0
-	mov	%rax,$m1
+	imulq	$lo0,$m1		# tp[0]*n0
+	mov	%rdx,$hi0
 
-	mulq	($np,$j,8)		# np[0]*m1
-	add	$lo0,%rax		# discarded
-	mov	8(%rsp),$lo0		# tp[1]
+	mulq	$m1			# np[0]*m1
+	add	%rax,$lo0		# discarded
+	mov	8($ap),%rax
 	adc	\$0,%rdx
+	mov	8(%rsp),$lo0		# tp[1]
 	mov	%rdx,$hi1
 
 	lea	1($j),$j		# j++
-.align	4
+	jmp	.Linner_enter
+
+.align	16
 .Linner:
+	add	%rax,$hi1
 	mov	($ap,$j,8),%rax
-	mulq	$m0			# ap[j]*bp[i]
-	add	$hi0,%rax
 	adc	\$0,%rdx
-	add	%rax,$lo0		# ap[j]*bp[i]+tp[j]
+	add	$lo0,$hi1		# np[j]*m1+ap[j]*bp[i]+tp[j]
+	mov	(%rsp,$j,8),$lo0
+	adc	\$0,%rdx
+	mov	$hi1,-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$hi1
+
+.Linner_enter:
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$hi0
 	mov	($np,$j,8),%rax
 	adc	\$0,%rdx
+	add	$hi0,$lo0		# ap[j]*bp[i]+tp[j]
 	mov	%rdx,$hi0
+	adc	\$0,$hi0
+	lea	1($j),$j		# j++
 
 	mulq	$m1			# np[j]*m1
-	add	$hi1,%rax
-	lea	1($j),$j		# j++
-	adc	\$0,%rdx
-	add	$lo0,%rax		# np[j]*m1+ap[j]*bp[i]+tp[j]
+	cmp	$num,$j
+	jne	.Linner
+
+	add	%rax,$hi1
+	mov	($ap),%rax		# ap[0]
 	adc	\$0,%rdx
+	add	$lo0,$hi1		# np[j]*m1+ap[j]*bp[i]+tp[j]
 	mov	(%rsp,$j,8),$lo0
-	cmp	$num,$j
-	mov	%rax,-16(%rsp,$j,8)	# tp[j-1]
+	adc	\$0,%rdx
+	mov	$hi1,-16(%rsp,$j,8)	# tp[j-1]
 	mov	%rdx,$hi1
-	jl	.Linner
 
 	xor	%rdx,%rdx
 	add	$hi0,$hi1
@@ -168,47 +228,1453 @@ bn_mul_mont:
 	cmp	$num,$i
 	jl	.Louter
 
-	lea	(%rsp),$ap		# borrow ap for tp
-	lea	-1($num),$j		# j=num-1
-
-	mov	($ap),%rax		# tp[0]
 	xor	$i,$i			# i=0 and clear CF!
+	mov	(%rsp),%rax		# tp[0]
+	lea	(%rsp),$ap		# borrow ap for tp
+	mov	$num,$j			# j=num
 	jmp	.Lsub
 .align	16
 .Lsub:	sbb	($np,$i,8),%rax
 	mov	%rax,($rp,$i,8)		# rp[i]=tp[i]-np[i]
-	dec	$j			# doesn't affect CF!
 	mov	8($ap,$i,8),%rax	# tp[i+1]
 	lea	1($i),$i		# i++
-	jge	.Lsub
+	dec	$j			# doesnn't affect CF!
+	jnz	.Lsub
 
 	sbb	\$0,%rax		# handle upmost overflow bit
+	xor	$i,$i
 	and	%rax,$ap
 	not	%rax
 	mov	$rp,$np
 	and	%rax,$np
-	lea	-1($num),$j
+	mov	$num,$j			# j=num
 	or	$np,$ap			# ap=borrow?tp:rp
 .align	16
 .Lcopy:					# copy or in-place refresh
+	mov	($ap,$i,8),%rax
+	mov	$i,(%rsp,$i,8)		# zap temporary vector
+	mov	%rax,($rp,$i,8)		# rp[i]=tp[i]
+	lea	1($i),$i
+	sub	\$1,$j
+	jnz	.Lcopy
+
+	mov	8(%rsp,$num,8),%rsi	# restore %rsp
+	mov	\$1,%rax
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lmul_epilogue:
+	ret
+.size	bn_mul_mont,.-bn_mul_mont
+___
+{{{
+my @A=("%r10","%r11");
+my @N=("%r13","%rdi");
+$code.=<<___;
+.type	bn_mul4x_mont,\@function,6
+.align	16
+bn_mul4x_mont:
+.Lmul4x_enter:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+
+	mov	${num}d,${num}d
+	lea	4($num),%r10
+	mov	%rsp,%r11
+	neg	%r10
+	lea	(%rsp,%r10,8),%rsp	# tp=alloca(8*(num+4))
+	and	\$-1024,%rsp		# minimize TLB usage
+
+	mov	%r11,8(%rsp,$num,8)	# tp[num+1]=%rsp
+.Lmul4x_body:
+	mov	$rp,16(%rsp,$num,8)	# tp[num+2]=$rp
+	mov	%rdx,%r12		# reassign $bp
+___
+		$bp="%r12";
+$code.=<<___;
+	mov	($n0),$n0		# pull n0[0] value
+	mov	($bp),$m0		# m0=bp[0]
+	mov	($ap),%rax
+
+	xor	$i,$i			# i=0
+	xor	$j,$j			# j=0
+
+	mov	$n0,$m1
+	mulq	$m0			# ap[0]*bp[0]
+	mov	%rax,$A[0]
+	mov	($np),%rax
+
+	imulq	$A[0],$m1		# "tp[0]"*n0
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[0]*m1
+	add	%rax,$A[0]		# discarded
+	mov	8($ap),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$N[1]
+
+	mulq	$m0
+	add	%rax,$A[1]
+	mov	8($np),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1
+	add	%rax,$N[1]
+	mov	16($ap),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]
+	lea	4($j),$j		# j++
+	adc	\$0,%rdx
+	mov	$N[1],(%rsp)
+	mov	%rdx,$N[0]
+	jmp	.L1st4x
+.align	16
+.L1st4x:
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[0]
+	mov	-16($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	-8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[0],-24(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[1]
+	mov	-8($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
 	mov	($ap,$j,8),%rax
-	mov	%rax,($rp,$j,8)		# rp[i]=tp[i]
-	mov	$i,(%rsp,$j,8)		# zap temporary vector
+	adc	\$0,%rdx
+	add	$A[1],$N[1]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[1],-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[0]
+	mov	($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[0],-8(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[1]
+	mov	8($np,$j,8),%rax
+	adc	\$0,%rdx
+	lea	4($j),$j		# j++
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	-16($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[1],-32(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+	cmp	$num,$j
+	jl	.L1st4x
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[0]
+	mov	-16($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	-8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[0],-24(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[1]
+	mov	-8($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	($ap),%rax		# ap[0]
+	adc	\$0,%rdx
+	add	$A[1],$N[1]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[1],-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+
+	xor	$N[1],$N[1]
+	add	$A[0],$N[0]
+	adc	\$0,$N[1]
+	mov	$N[0],-8(%rsp,$j,8)
+	mov	$N[1],(%rsp,$j,8)	# store upmost overflow bit
+
+	lea	1($i),$i		# i++
+.align	4
+.Louter4x:
+	mov	($bp,$i,8),$m0		# m0=bp[i]
+	xor	$j,$j			# j=0
+	mov	(%rsp),$A[0]
+	mov	$n0,$m1
+	mulq	$m0			# ap[0]*bp[i]
+	add	%rax,$A[0]		# ap[0]*bp[i]+tp[0]
+	mov	($np),%rax
+	adc	\$0,%rdx
+
+	imulq	$A[0],$m1		# tp[0]*n0
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[0]*m1
+	add	%rax,$A[0]		# "$N[0]", discarded
+	mov	8($ap),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[1]
+	mov	8($np),%rax
+	adc	\$0,%rdx
+	add	8(%rsp),$A[1]		# +tp[1]
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	16($ap),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]		# np[j]*m1+ap[j]*bp[i]+tp[j]
+	lea	4($j),$j		# j+=2
+	adc	\$0,%rdx
+	mov	$N[1],(%rsp)		# tp[j-1]
+	mov	%rdx,$N[0]
+	jmp	.Linner4x
+.align	16
+.Linner4x:
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[0]
+	mov	-16($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	-16(%rsp,$j,8),$A[0]	# ap[j]*bp[i]+tp[j]
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	-8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]
+	adc	\$0,%rdx
+	mov	$N[0],-24(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[1]
+	mov	-8($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	-8(%rsp,$j,8),$A[1]
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]
+	adc	\$0,%rdx
+	mov	$N[1],-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[0]
+	mov	($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	(%rsp,$j,8),$A[0]	# ap[j]*bp[i]+tp[j]
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]
+	adc	\$0,%rdx
+	mov	$N[0],-8(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[1]
+	mov	8($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	8(%rsp,$j,8),$A[1]
+	adc	\$0,%rdx
+	lea	4($j),$j		# j++
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	-16($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]
+	adc	\$0,%rdx
+	mov	$N[1],-32(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+	cmp	$num,$j
+	jl	.Linner4x
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[0]
+	mov	-16($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	-16(%rsp,$j,8),$A[0]	# ap[j]*bp[i]+tp[j]
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	-8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]
+	adc	\$0,%rdx
+	mov	$N[0],-24(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[1]
+	mov	-8($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	-8(%rsp,$j,8),$A[1]
+	adc	\$0,%rdx
+	lea	1($i),$i		# i++
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	($ap),%rax		# ap[0]
+	adc	\$0,%rdx
+	add	$A[1],$N[1]
+	adc	\$0,%rdx
+	mov	$N[1],-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+
+	xor	$N[1],$N[1]
+	add	$A[0],$N[0]
+	adc	\$0,$N[1]
+	add	(%rsp,$num,8),$N[0]	# pull upmost overflow bit
+	adc	\$0,$N[1]
+	mov	$N[0],-8(%rsp,$j,8)
+	mov	$N[1],(%rsp,$j,8)	# store upmost overflow bit
+
+	cmp	$num,$i
+	jl	.Louter4x
+___
+{
+my @ri=("%rax","%rdx",$m0,$m1);
+$code.=<<___;
+	mov	16(%rsp,$num,8),$rp	# restore $rp
+	mov	0(%rsp),@ri[0]		# tp[0]
+	pxor	%xmm0,%xmm0
+	mov	8(%rsp),@ri[1]		# tp[1]
+	shr	\$2,$num		# num/=4
+	lea	(%rsp),$ap		# borrow ap for tp
+	xor	$i,$i			# i=0 and clear CF!
+
+	sub	0($np),@ri[0]
+	mov	16($ap),@ri[2]		# tp[2]
+	mov	24($ap),@ri[3]		# tp[3]
+	sbb	8($np),@ri[1]
+	lea	-1($num),$j		# j=num/4-1
+	jmp	.Lsub4x
+.align	16
+.Lsub4x:
+	mov	@ri[0],0($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	mov	@ri[1],8($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	sbb	16($np,$i,8),@ri[2]
+	mov	32($ap,$i,8),@ri[0]	# tp[i+1]
+	mov	40($ap,$i,8),@ri[1]
+	sbb	24($np,$i,8),@ri[3]
+	mov	@ri[2],16($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	mov	@ri[3],24($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	sbb	32($np,$i,8),@ri[0]
+	mov	48($ap,$i,8),@ri[2]
+	mov	56($ap,$i,8),@ri[3]
+	sbb	40($np,$i,8),@ri[1]
+	lea	4($i),$i		# i++
+	dec	$j			# doesnn't affect CF!
+	jnz	.Lsub4x
+
+	mov	@ri[0],0($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	mov	32($ap,$i,8),@ri[0]	# load overflow bit
+	sbb	16($np,$i,8),@ri[2]
+	mov	@ri[1],8($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	sbb	24($np,$i,8),@ri[3]
+	mov	@ri[2],16($rp,$i,8)	# rp[i]=tp[i]-np[i]
+
+	sbb	\$0,@ri[0]		# handle upmost overflow bit
+	mov	@ri[3],24($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	xor	$i,$i			# i=0
+	and	@ri[0],$ap
+	not	@ri[0]
+	mov	$rp,$np
+	and	@ri[0],$np
+	lea	-1($num),$j
+	or	$np,$ap			# ap=borrow?tp:rp
+
+	movdqu	($ap),%xmm1
+	movdqa	%xmm0,(%rsp)
+	movdqu	%xmm1,($rp)
+	jmp	.Lcopy4x
+.align	16
+.Lcopy4x:					# copy or in-place refresh
+	movdqu	16($ap,$i),%xmm2
+	movdqu	32($ap,$i),%xmm1
+	movdqa	%xmm0,16(%rsp,$i)
+	movdqu	%xmm2,16($rp,$i)
+	movdqa	%xmm0,32(%rsp,$i)
+	movdqu	%xmm1,32($rp,$i)
+	lea	32($i),$i
+	dec	$j
+	jnz	.Lcopy4x
+
+	shl	\$2,$num
+	movdqu	16($ap,$i),%xmm2
+	movdqa	%xmm0,16(%rsp,$i)
+	movdqu	%xmm2,16($rp,$i)
+___
+}
+$code.=<<___;
+	mov	8(%rsp,$num,8),%rsi	# restore %rsp
+	mov	\$1,%rax
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lmul4x_epilogue:
+	ret
+.size	bn_mul4x_mont,.-bn_mul4x_mont
+___
+}}}
+{{{
+######################################################################
+# void bn_sqr4x_mont(
+my $rptr="%rdi";	# const BN_ULONG *rptr,
+my $aptr="%rsi";	# const BN_ULONG *aptr,
+my $bptr="%rdx";	# not used
+my $nptr="%rcx";	# const BN_ULONG *nptr,
+my $n0  ="%r8";		# const BN_ULONG *n0);
+my $num ="%r9";		# int num, has to be divisible by 4 and
+			# not less than 8
+
+my ($i,$j,$tptr)=("%rbp","%rcx",$rptr);
+my @A0=("%r10","%r11");
+my @A1=("%r12","%r13");
+my ($a0,$a1,$ai)=("%r14","%r15","%rbx");
+
+$code.=<<___;
+.type	bn_sqr4x_mont,\@function,6
+.align	16
+bn_sqr4x_mont:
+.Lsqr4x_enter:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+
+	shl	\$3,${num}d		# convert $num to bytes
+	xor	%r10,%r10
+	mov	%rsp,%r11		# put aside %rsp
+	sub	$num,%r10		# -$num
+	mov	($n0),$n0		# *n0
+	lea	-72(%rsp,%r10,2),%rsp	# alloca(frame+2*$num)
+	and	\$-1024,%rsp		# minimize TLB usage
+	##############################################################
+	# Stack layout
+	#
+	# +0	saved $num, used in reduction section
+	# +8	&t[2*$num], used in reduction section
+	# +32	saved $rptr
+	# +40	saved $nptr
+	# +48	saved *n0
+	# +56	saved %rsp
+	# +64	t[2*$num]
+	#
+	mov	$rptr,32(%rsp)		# save $rptr
+	mov	$nptr,40(%rsp)
+	mov	$n0,  48(%rsp)
+	mov	%r11, 56(%rsp)		# save original %rsp
+.Lsqr4x_body:
+	##############################################################
+	# Squaring part:
+	#
+	# a) multiply-n-add everything but a[i]*a[i];
+	# b) shift result of a) by 1 to the left and accumulate
+	#    a[i]*a[i] products;
+	#
+	lea	32(%r10),$i		# $i=-($num-32)
+	lea	($aptr,$num),$aptr	# end of a[] buffer, ($aptr,$i)=&ap[2]
+
+	mov	$num,$j			# $j=$num
+
+					# comments apply to $num==8 case
+	mov	-32($aptr,$i),$a0	# a[0]
+	lea	64(%rsp,$num,2),$tptr	# end of tp[] buffer, &tp[2*$num]
+	mov	-24($aptr,$i),%rax	# a[1]
+	lea	-32($tptr,$i),$tptr	# end of tp[] window, &tp[2*$num-"$i"]
+	mov	-16($aptr,$i),$ai	# a[2]
+	mov	%rax,$a1
+
+	mul	$a0			# a[1]*a[0]
+	mov	%rax,$A0[0]		# a[1]*a[0]
+	 mov	$ai,%rax		# a[2]
+	mov	%rdx,$A0[1]
+	mov	$A0[0],-24($tptr,$i)	# t[1]
+
+	xor	$A0[0],$A0[0]
+	mul	$a0			# a[2]*a[0]
+	add	%rax,$A0[1]
+	 mov	$ai,%rax
+	adc	%rdx,$A0[0]
+	mov	$A0[1],-16($tptr,$i)	# t[2]
+
+	lea	-16($i),$j		# j=-16
+
+
+	 mov	8($aptr,$j),$ai		# a[3]
+	mul	$a1			# a[2]*a[1]
+	mov	%rax,$A1[0]		# a[2]*a[1]+t[3]
+	 mov	$ai,%rax
+	mov	%rdx,$A1[1]
+
+	xor	$A0[1],$A0[1]
+	add	$A1[0],$A0[0]
+	 lea	16($j),$j
+	adc	\$0,$A0[1]
+	mul	$a0			# a[3]*a[0]
+	add	%rax,$A0[0]		# a[3]*a[0]+a[2]*a[1]+t[3]
+	 mov	$ai,%rax
+	adc	%rdx,$A0[1]
+	mov	$A0[0],-8($tptr,$j)	# t[3]
+	jmp	.Lsqr4x_1st
+
+.align	16
+.Lsqr4x_1st:
+	 mov	($aptr,$j),$ai		# a[4]
+	xor	$A1[0],$A1[0]
+	mul	$a1			# a[3]*a[1]
+	add	%rax,$A1[1]		# a[3]*a[1]+t[4]
+	 mov	$ai,%rax
+	adc	%rdx,$A1[0]
+
+	xor	$A0[0],$A0[0]
+	add	$A1[1],$A0[1]
+	adc	\$0,$A0[0]
+	mul	$a0			# a[4]*a[0]
+	add	%rax,$A0[1]		# a[4]*a[0]+a[3]*a[1]+t[4]
+	 mov	$ai,%rax		# a[3]
+	adc	%rdx,$A0[0]
+	mov	$A0[1],($tptr,$j)	# t[4]
+
+
+	 mov	8($aptr,$j),$ai		# a[5]
+	xor	$A1[1],$A1[1]
+	mul	$a1			# a[4]*a[3]
+	add	%rax,$A1[0]		# a[4]*a[3]+t[5]
+	 mov	$ai,%rax
+	adc	%rdx,$A1[1]
+
+	xor	$A0[1],$A0[1]
+	add	$A1[0],$A0[0]
+	adc	\$0,$A0[1]
+	mul	$a0			# a[5]*a[2]
+	add	%rax,$A0[0]		# a[5]*a[2]+a[4]*a[3]+t[5]
+	 mov	$ai,%rax
+	adc	%rdx,$A0[1]
+	mov	$A0[0],8($tptr,$j)	# t[5]
+
+	 mov	16($aptr,$j),$ai	# a[6]
+	xor	$A1[0],$A1[0]
+	mul	$a1			# a[5]*a[3]
+	add	%rax,$A1[1]		# a[5]*a[3]+t[6]
+	 mov	$ai,%rax
+	adc	%rdx,$A1[0]
+
+	xor	$A0[0],$A0[0]
+	add	$A1[1],$A0[1]
+	adc	\$0,$A0[0]
+	mul	$a0			# a[6]*a[2]
+	add	%rax,$A0[1]		# a[6]*a[2]+a[5]*a[3]+t[6]
+	 mov	$ai,%rax		# a[3]
+	adc	%rdx,$A0[0]
+	mov	$A0[1],16($tptr,$j)	# t[6]
+
+
+	 mov	24($aptr,$j),$ai	# a[7]
+	xor	$A1[1],$A1[1]
+	mul	$a1			# a[6]*a[5]
+	add	%rax,$A1[0]		# a[6]*a[5]+t[7]
+	 mov	$ai,%rax
+	adc	%rdx,$A1[1]
+
+	xor	$A0[1],$A0[1]
+	add	$A1[0],$A0[0]
+	 lea	32($j),$j
+	adc	\$0,$A0[1]
+	mul	$a0			# a[7]*a[4]
+	add	%rax,$A0[0]		# a[7]*a[4]+a[6]*a[5]+t[6]
+	 mov	$ai,%rax
+	adc	%rdx,$A0[1]
+	mov	$A0[0],-8($tptr,$j)	# t[7]
+
+	cmp	\$0,$j
+	jne	.Lsqr4x_1st
+
+	xor	$A1[0],$A1[0]
+	add	$A0[1],$A1[1]
+	adc	\$0,$A1[0]
+	mul	$a1			# a[7]*a[5]
+	add	%rax,$A1[1]
+	adc	%rdx,$A1[0]
+
+	mov	$A1[1],($tptr)		# t[8]
+	lea	16($i),$i
+	mov	$A1[0],8($tptr)		# t[9]
+	jmp	.Lsqr4x_outer
+
+.align	16
+.Lsqr4x_outer:				# comments apply to $num==6 case
+	mov	-32($aptr,$i),$a0	# a[0]
+	lea	64(%rsp,$num,2),$tptr	# end of tp[] buffer, &tp[2*$num]
+	mov	-24($aptr,$i),%rax	# a[1]
+	lea	-32($tptr,$i),$tptr	# end of tp[] window, &tp[2*$num-"$i"]
+	mov	-16($aptr,$i),$ai	# a[2]
+	mov	%rax,$a1
+
+	mov	-24($tptr,$i),$A0[0]	# t[1]
+	xor	$A0[1],$A0[1]
+	mul	$a0			# a[1]*a[0]
+	add	%rax,$A0[0]		# a[1]*a[0]+t[1]
+	 mov	$ai,%rax		# a[2]
+	adc	%rdx,$A0[1]
+	mov	$A0[0],-24($tptr,$i)	# t[1]
+
+	xor	$A0[0],$A0[0]
+	add	-16($tptr,$i),$A0[1]	# a[2]*a[0]+t[2]
+	adc	\$0,$A0[0]
+	mul	$a0			# a[2]*a[0]
+	add	%rax,$A0[1]
+	 mov	$ai,%rax
+	adc	%rdx,$A0[0]
+	mov	$A0[1],-16($tptr,$i)	# t[2]
+
+	lea	-16($i),$j		# j=-16
+	xor	$A1[0],$A1[0]
+
+
+	 mov	8($aptr,$j),$ai		# a[3]
+	xor	$A1[1],$A1[1]
+	add	8($tptr,$j),$A1[0]
+	adc	\$0,$A1[1]
+	mul	$a1			# a[2]*a[1]
+	add	%rax,$A1[0]		# a[2]*a[1]+t[3]
+	 mov	$ai,%rax
+	adc	%rdx,$A1[1]
+
+	xor	$A0[1],$A0[1]
+	add	$A1[0],$A0[0]
+	adc	\$0,$A0[1]
+	mul	$a0			# a[3]*a[0]
+	add	%rax,$A0[0]		# a[3]*a[0]+a[2]*a[1]+t[3]
+	 mov	$ai,%rax
+	adc	%rdx,$A0[1]
+	mov	$A0[0],8($tptr,$j)	# t[3]
+
+	lea	16($j),$j
+	jmp	.Lsqr4x_inner
+
+.align	16
+.Lsqr4x_inner:
+	 mov	($aptr,$j),$ai		# a[4]
+	xor	$A1[0],$A1[0]
+	add	($tptr,$j),$A1[1]
+	adc	\$0,$A1[0]
+	mul	$a1			# a[3]*a[1]
+	add	%rax,$A1[1]		# a[3]*a[1]+t[4]
+	 mov	$ai,%rax
+	adc	%rdx,$A1[0]
+
+	xor	$A0[0],$A0[0]
+	add	$A1[1],$A0[1]
+	adc	\$0,$A0[0]
+	mul	$a0			# a[4]*a[0]
+	add	%rax,$A0[1]		# a[4]*a[0]+a[3]*a[1]+t[4]
+	 mov	$ai,%rax		# a[3]
+	adc	%rdx,$A0[0]
+	mov	$A0[1],($tptr,$j)	# t[4]
+
+	 mov	8($aptr,$j),$ai		# a[5]
+	xor	$A1[1],$A1[1]
+	add	8($tptr,$j),$A1[0]
+	adc	\$0,$A1[1]
+	mul	$a1			# a[4]*a[3]
+	add	%rax,$A1[0]		# a[4]*a[3]+t[5]
+	 mov	$ai,%rax
+	adc	%rdx,$A1[1]
+
+	xor	$A0[1],$A0[1]
+	add	$A1[0],$A0[0]
+	lea	16($j),$j		# j++
+	adc	\$0,$A0[1]
+	mul	$a0			# a[5]*a[2]
+	add	%rax,$A0[0]		# a[5]*a[2]+a[4]*a[3]+t[5]
+	 mov	$ai,%rax
+	adc	%rdx,$A0[1]
+	mov	$A0[0],-8($tptr,$j)	# t[5], "preloaded t[1]" below
+
+	cmp	\$0,$j
+	jne	.Lsqr4x_inner
+
+	xor	$A1[0],$A1[0]
+	add	$A0[1],$A1[1]
+	adc	\$0,$A1[0]
+	mul	$a1			# a[5]*a[3]
+	add	%rax,$A1[1]
+	adc	%rdx,$A1[0]
+
+	mov	$A1[1],($tptr)		# t[6], "preloaded t[2]" below
+	mov	$A1[0],8($tptr)		# t[7], "preloaded t[3]" below
+
+	add	\$16,$i
+	jnz	.Lsqr4x_outer
+
+					# comments apply to $num==4 case
+	mov	-32($aptr),$a0		# a[0]
+	lea	64(%rsp,$num,2),$tptr	# end of tp[] buffer, &tp[2*$num]
+	mov	-24($aptr),%rax		# a[1]
+	lea	-32($tptr,$i),$tptr	# end of tp[] window, &tp[2*$num-"$i"]
+	mov	-16($aptr),$ai		# a[2]
+	mov	%rax,$a1
+
+	xor	$A0[1],$A0[1]
+	mul	$a0			# a[1]*a[0]
+	add	%rax,$A0[0]		# a[1]*a[0]+t[1], preloaded t[1]
+	 mov	$ai,%rax		# a[2]
+	adc	%rdx,$A0[1]
+	mov	$A0[0],-24($tptr)	# t[1]
+
+	xor	$A0[0],$A0[0]
+	add	$A1[1],$A0[1]		# a[2]*a[0]+t[2], preloaded t[2]
+	adc	\$0,$A0[0]
+	mul	$a0			# a[2]*a[0]
+	add	%rax,$A0[1]
+	 mov	$ai,%rax
+	adc	%rdx,$A0[0]
+	mov	$A0[1],-16($tptr)	# t[2]
+
+	 mov	-8($aptr),$ai		# a[3]
+	mul	$a1			# a[2]*a[1]
+	add	%rax,$A1[0]		# a[2]*a[1]+t[3], preloaded t[3]
+	 mov	$ai,%rax
+	adc	\$0,%rdx
+
+	xor	$A0[1],$A0[1]
+	add	$A1[0],$A0[0]
+	 mov	%rdx,$A1[1]
+	adc	\$0,$A0[1]
+	mul	$a0			# a[3]*a[0]
+	add	%rax,$A0[0]		# a[3]*a[0]+a[2]*a[1]+t[3]
+	 mov	$ai,%rax
+	adc	%rdx,$A0[1]
+	mov	$A0[0],-8($tptr)	# t[3]
+
+	xor	$A1[0],$A1[0]
+	add	$A0[1],$A1[1]
+	adc	\$0,$A1[0]
+	mul	$a1			# a[3]*a[1]
+	add	%rax,$A1[1]
+	 mov	-16($aptr),%rax		# a[2]
+	adc	%rdx,$A1[0]
+
+	mov	$A1[1],($tptr)		# t[4]
+	mov	$A1[0],8($tptr)		# t[5]
+
+	mul	$ai			# a[2]*a[3]
+___
+{
+my ($shift,$carry)=($a0,$a1);
+my @S=(@A1,$ai,$n0);
+$code.=<<___;
+	 add	\$16,$i
+	 xor	$shift,$shift
+	 sub	$num,$i			# $i=16-$num
+	 xor	$carry,$carry
+
+	add	$A1[0],%rax		# t[5]
+	adc	\$0,%rdx
+	mov	%rax,8($tptr)		# t[5]
+	mov	%rdx,16($tptr)		# t[6]
+	mov	$carry,24($tptr)	# t[7]
+
+	 mov	-16($aptr,$i),%rax	# a[0]
+	lea	64(%rsp,$num,2),$tptr
+	 xor	$A0[0],$A0[0]		# t[0]
+	 mov	-24($tptr,$i,2),$A0[1]	# t[1]
+
+	lea	($shift,$A0[0],2),$S[0]	# t[2*i]<<1 | shift
+	shr	\$63,$A0[0]
+	lea	($j,$A0[1],2),$S[1]	# t[2*i+1]<<1 |
+	shr	\$63,$A0[1]
+	or	$A0[0],$S[1]		# | t[2*i]>>63
+	 mov	-16($tptr,$i,2),$A0[0]	# t[2*i+2]	# prefetch
+	mov	$A0[1],$shift		# shift=t[2*i+1]>>63
+	mul	%rax			# a[i]*a[i]
+	neg	$carry			# mov $carry,cf
+	 mov	-8($tptr,$i,2),$A0[1]	# t[2*i+2+1]	# prefetch
+	adc	%rax,$S[0]
+	 mov	-8($aptr,$i),%rax	# a[i+1]	# prefetch
+	mov	$S[0],-32($tptr,$i,2)
+	adc	%rdx,$S[1]
+
+	lea	($shift,$A0[0],2),$S[2]	# t[2*i]<<1 | shift
+	 mov	$S[1],-24($tptr,$i,2)
+	 sbb	$carry,$carry		# mov cf,$carry
+	shr	\$63,$A0[0]
+	lea	($j,$A0[1],2),$S[3]	# t[2*i+1]<<1 |
+	shr	\$63,$A0[1]
+	or	$A0[0],$S[3]		# | t[2*i]>>63
+	 mov	0($tptr,$i,2),$A0[0]	# t[2*i+2]	# prefetch
+	mov	$A0[1],$shift		# shift=t[2*i+1]>>63
+	mul	%rax			# a[i]*a[i]
+	neg	$carry			# mov $carry,cf
+	 mov	8($tptr,$i,2),$A0[1]	# t[2*i+2+1]	# prefetch
+	adc	%rax,$S[2]
+	 mov	0($aptr,$i),%rax	# a[i+1]	# prefetch
+	mov	$S[2],-16($tptr,$i,2)
+	adc	%rdx,$S[3]
+	lea	16($i),$i
+	mov	$S[3],-40($tptr,$i,2)
+	sbb	$carry,$carry		# mov cf,$carry
+	jmp	.Lsqr4x_shift_n_add
+
+.align	16
+.Lsqr4x_shift_n_add:
+	lea	($shift,$A0[0],2),$S[0]	# t[2*i]<<1 | shift
+	shr	\$63,$A0[0]
+	lea	($j,$A0[1],2),$S[1]	# t[2*i+1]<<1 |
+	shr	\$63,$A0[1]
+	or	$A0[0],$S[1]		# | t[2*i]>>63
+	 mov	-16($tptr,$i,2),$A0[0]	# t[2*i+2]	# prefetch
+	mov	$A0[1],$shift		# shift=t[2*i+1]>>63
+	mul	%rax			# a[i]*a[i]
+	neg	$carry			# mov $carry,cf
+	 mov	-8($tptr,$i,2),$A0[1]	# t[2*i+2+1]	# prefetch
+	adc	%rax,$S[0]
+	 mov	-8($aptr,$i),%rax	# a[i+1]	# prefetch
+	mov	$S[0],-32($tptr,$i,2)
+	adc	%rdx,$S[1]
+
+	lea	($shift,$A0[0],2),$S[2]	# t[2*i]<<1 | shift
+	 mov	$S[1],-24($tptr,$i,2)
+	 sbb	$carry,$carry		# mov cf,$carry
+	shr	\$63,$A0[0]
+	lea	($j,$A0[1],2),$S[3]	# t[2*i+1]<<1 |
+	shr	\$63,$A0[1]
+	or	$A0[0],$S[3]		# | t[2*i]>>63
+	 mov	0($tptr,$i,2),$A0[0]	# t[2*i+2]	# prefetch
+	mov	$A0[1],$shift		# shift=t[2*i+1]>>63
+	mul	%rax			# a[i]*a[i]
+	neg	$carry			# mov $carry,cf
+	 mov	8($tptr,$i,2),$A0[1]	# t[2*i+2+1]	# prefetch
+	adc	%rax,$S[2]
+	 mov	0($aptr,$i),%rax	# a[i+1]	# prefetch
+	mov	$S[2],-16($tptr,$i,2)
+	adc	%rdx,$S[3]
+
+	lea	($shift,$A0[0],2),$S[0]	# t[2*i]<<1 | shift
+	 mov	$S[3],-8($tptr,$i,2)
+	 sbb	$carry,$carry		# mov cf,$carry
+	shr	\$63,$A0[0]
+	lea	($j,$A0[1],2),$S[1]	# t[2*i+1]<<1 |
+	shr	\$63,$A0[1]
+	or	$A0[0],$S[1]		# | t[2*i]>>63
+	 mov	16($tptr,$i,2),$A0[0]	# t[2*i+2]	# prefetch
+	mov	$A0[1],$shift		# shift=t[2*i+1]>>63
+	mul	%rax			# a[i]*a[i]
+	neg	$carry			# mov $carry,cf
+	 mov	24($tptr,$i,2),$A0[1]	# t[2*i+2+1]	# prefetch
+	adc	%rax,$S[0]
+	 mov	8($aptr,$i),%rax	# a[i+1]	# prefetch
+	mov	$S[0],0($tptr,$i,2)
+	adc	%rdx,$S[1]
+
+	lea	($shift,$A0[0],2),$S[2]	# t[2*i]<<1 | shift
+	 mov	$S[1],8($tptr,$i,2)
+	 sbb	$carry,$carry		# mov cf,$carry
+	shr	\$63,$A0[0]
+	lea	($j,$A0[1],2),$S[3]	# t[2*i+1]<<1 |
+	shr	\$63,$A0[1]
+	or	$A0[0],$S[3]		# | t[2*i]>>63
+	 mov	32($tptr,$i,2),$A0[0]	# t[2*i+2]	# prefetch
+	mov	$A0[1],$shift		# shift=t[2*i+1]>>63
+	mul	%rax			# a[i]*a[i]
+	neg	$carry			# mov $carry,cf
+	 mov	40($tptr,$i,2),$A0[1]	# t[2*i+2+1]	# prefetch
+	adc	%rax,$S[2]
+	 mov	16($aptr,$i),%rax	# a[i+1]	# prefetch
+	mov	$S[2],16($tptr,$i,2)
+	adc	%rdx,$S[3]
+	mov	$S[3],24($tptr,$i,2)
+	sbb	$carry,$carry		# mov cf,$carry
+	add	\$32,$i
+	jnz	.Lsqr4x_shift_n_add
+
+	lea	($shift,$A0[0],2),$S[0]	# t[2*i]<<1 | shift
+	shr	\$63,$A0[0]
+	lea	($j,$A0[1],2),$S[1]	# t[2*i+1]<<1 |
+	shr	\$63,$A0[1]
+	or	$A0[0],$S[1]		# | t[2*i]>>63
+	 mov	-16($tptr),$A0[0]	# t[2*i+2]	# prefetch
+	mov	$A0[1],$shift		# shift=t[2*i+1]>>63
+	mul	%rax			# a[i]*a[i]
+	neg	$carry			# mov $carry,cf
+	 mov	-8($tptr),$A0[1]	# t[2*i+2+1]	# prefetch
+	adc	%rax,$S[0]
+	 mov	-8($aptr),%rax		# a[i+1]	# prefetch
+	mov	$S[0],-32($tptr)
+	adc	%rdx,$S[1]
+
+	lea	($shift,$A0[0],2),$S[2]	# t[2*i]<<1|shift
+	 mov	$S[1],-24($tptr)
+	 sbb	$carry,$carry		# mov cf,$carry
+	shr	\$63,$A0[0]
+	lea	($j,$A0[1],2),$S[3]	# t[2*i+1]<<1 |
+	shr	\$63,$A0[1]
+	or	$A0[0],$S[3]		# | t[2*i]>>63
+	mul	%rax			# a[i]*a[i]
+	neg	$carry			# mov $carry,cf
+	adc	%rax,$S[2]
+	adc	%rdx,$S[3]
+	mov	$S[2],-16($tptr)
+	mov	$S[3],-8($tptr)
+___
+}
+##############################################################
+# Montgomery reduction part, "word-by-word" algorithm.
+#
+{
+my ($topbit,$nptr)=("%rbp",$aptr);
+my ($m0,$m1)=($a0,$a1);
+my @Ni=("%rbx","%r9");
+$code.=<<___;
+	mov	40(%rsp),$nptr		# restore $nptr
+	mov	48(%rsp),$n0		# restore *n0
+	xor	$j,$j
+	mov	$num,0(%rsp)		# save $num
+	sub	$num,$j			# $j=-$num
+	 mov	64(%rsp),$A0[0]		# t[0]		# modsched #
+	 mov	$n0,$m0			#		# modsched #
+	lea	64(%rsp,$num,2),%rax	# end of t[] buffer
+	lea	64(%rsp,$num),$tptr	# end of t[] window
+	mov	%rax,8(%rsp)		# save end of t[] buffer
+	lea	($nptr,$num),$nptr	# end of n[] buffer
+	xor	$topbit,$topbit		# $topbit=0
+
+	mov	0($nptr,$j),%rax	# n[0]		# modsched #
+	mov	8($nptr,$j),$Ni[1]	# n[1]		# modsched #
+	 imulq	$A0[0],$m0		# m0=t[0]*n0	# modsched #
+	 mov	%rax,$Ni[0]		#		# modsched #
+	jmp	.Lsqr4x_mont_outer
+
+.align	16
+.Lsqr4x_mont_outer:
+	xor	$A0[1],$A0[1]
+	mul	$m0			# n[0]*m0
+	add	%rax,$A0[0]		# n[0]*m0+t[0]
+	 mov	$Ni[1],%rax
+	adc	%rdx,$A0[1]
+	mov	$n0,$m1
+
+	xor	$A0[0],$A0[0]
+	add	8($tptr,$j),$A0[1]
+	adc	\$0,$A0[0]
+	mul	$m0			# n[1]*m0
+	add	%rax,$A0[1]		# n[1]*m0+t[1]
+	 mov	$Ni[0],%rax
+	adc	%rdx,$A0[0]
+
+	imulq	$A0[1],$m1
+
+	mov	16($nptr,$j),$Ni[0]	# n[2]
+	xor	$A1[1],$A1[1]
+	add	$A0[1],$A1[0]
+	adc	\$0,$A1[1]
+	mul	$m1			# n[0]*m1
+	add	%rax,$A1[0]		# n[0]*m1+"t[1]"
+	 mov	$Ni[0],%rax
+	adc	%rdx,$A1[1]
+	mov	$A1[0],8($tptr,$j)	# "t[1]"
+
+	xor	$A0[1],$A0[1]
+	add	16($tptr,$j),$A0[0]
+	adc	\$0,$A0[1]
+	mul	$m0			# n[2]*m0
+	add	%rax,$A0[0]		# n[2]*m0+t[2]
+	 mov	$Ni[1],%rax
+	adc	%rdx,$A0[1]
+
+	mov	24($nptr,$j),$Ni[1]	# n[3]
+	xor	$A1[0],$A1[0]
+	add	$A0[0],$A1[1]
+	adc	\$0,$A1[0]
+	mul	$m1			# n[1]*m1
+	add	%rax,$A1[1]		# n[1]*m1+"t[2]"
+	 mov	$Ni[1],%rax
+	adc	%rdx,$A1[0]
+	mov	$A1[1],16($tptr,$j)	# "t[2]"
+
+	xor	$A0[0],$A0[0]
+	add	24($tptr,$j),$A0[1]
+	lea	32($j),$j
+	adc	\$0,$A0[0]
+	mul	$m0			# n[3]*m0
+	add	%rax,$A0[1]		# n[3]*m0+t[3]
+	 mov	$Ni[0],%rax
+	adc	%rdx,$A0[0]
+	jmp	.Lsqr4x_mont_inner
+
+.align	16
+.Lsqr4x_mont_inner:
+	mov	($nptr,$j),$Ni[0]	# n[4]
+	xor	$A1[1],$A1[1]
+	add	$A0[1],$A1[0]
+	adc	\$0,$A1[1]
+	mul	$m1			# n[2]*m1
+	add	%rax,$A1[0]		# n[2]*m1+"t[3]"
+	 mov	$Ni[0],%rax
+	adc	%rdx,$A1[1]
+	mov	$A1[0],-8($tptr,$j)	# "t[3]"
+
+	xor	$A0[1],$A0[1]
+	add	($tptr,$j),$A0[0]
+	adc	\$0,$A0[1]
+	mul	$m0			# n[4]*m0
+	add	%rax,$A0[0]		# n[4]*m0+t[4]
+	 mov	$Ni[1],%rax
+	adc	%rdx,$A0[1]
+
+	mov	8($nptr,$j),$Ni[1]	# n[5]
+	xor	$A1[0],$A1[0]
+	add	$A0[0],$A1[1]
+	adc	\$0,$A1[0]
+	mul	$m1			# n[3]*m1
+	add	%rax,$A1[1]		# n[3]*m1+"t[4]"
+	 mov	$Ni[1],%rax
+	adc	%rdx,$A1[0]
+	mov	$A1[1],($tptr,$j)	# "t[4]"
+
+	xor	$A0[0],$A0[0]
+	add	8($tptr,$j),$A0[1]
+	adc	\$0,$A0[0]
+	mul	$m0			# n[5]*m0
+	add	%rax,$A0[1]		# n[5]*m0+t[5]
+	 mov	$Ni[0],%rax
+	adc	%rdx,$A0[0]
+
+
+	mov	16($nptr,$j),$Ni[0]	# n[6]
+	xor	$A1[1],$A1[1]
+	add	$A0[1],$A1[0]
+	adc	\$0,$A1[1]
+	mul	$m1			# n[4]*m1
+	add	%rax,$A1[0]		# n[4]*m1+"t[5]"
+	 mov	$Ni[0],%rax
+	adc	%rdx,$A1[1]
+	mov	$A1[0],8($tptr,$j)	# "t[5]"
+
+	xor	$A0[1],$A0[1]
+	add	16($tptr,$j),$A0[0]
+	adc	\$0,$A0[1]
+	mul	$m0			# n[6]*m0
+	add	%rax,$A0[0]		# n[6]*m0+t[6]
+	 mov	$Ni[1],%rax
+	adc	%rdx,$A0[1]
+
+	mov	24($nptr,$j),$Ni[1]	# n[7]
+	xor	$A1[0],$A1[0]
+	add	$A0[0],$A1[1]
+	adc	\$0,$A1[0]
+	mul	$m1			# n[5]*m1
+	add	%rax,$A1[1]		# n[5]*m1+"t[6]"
+	 mov	$Ni[1],%rax
+	adc	%rdx,$A1[0]
+	mov	$A1[1],16($tptr,$j)	# "t[6]"
+
+	xor	$A0[0],$A0[0]
+	add	24($tptr,$j),$A0[1]
+	lea	32($j),$j
+	adc	\$0,$A0[0]
+	mul	$m0			# n[7]*m0
+	add	%rax,$A0[1]		# n[7]*m0+t[7]
+	 mov	$Ni[0],%rax
+	adc	%rdx,$A0[0]
+	cmp	\$0,$j
+	jne	.Lsqr4x_mont_inner
+
+	 sub	0(%rsp),$j		# $j=-$num	# modsched #
+	 mov	$n0,$m0			#		# modsched #
+
+	xor	$A1[1],$A1[1]
+	add	$A0[1],$A1[0]
+	adc	\$0,$A1[1]
+	mul	$m1			# n[6]*m1
+	add	%rax,$A1[0]		# n[6]*m1+"t[7]"
+	mov	$Ni[1],%rax
+	adc	%rdx,$A1[1]
+	mov	$A1[0],-8($tptr)	# "t[7]"
+
+	xor	$A0[1],$A0[1]
+	add	($tptr),$A0[0]		# +t[8]
+	adc	\$0,$A0[1]
+	 mov	0($nptr,$j),$Ni[0]	# n[0]		# modsched #
+	add	$topbit,$A0[0]
+	adc	\$0,$A0[1]
+
+	 imulq	16($tptr,$j),$m0	# m0=t[0]*n0	# modsched #
+	xor	$A1[0],$A1[0]
+	 mov	8($nptr,$j),$Ni[1]	# n[1]		# modsched #
+	add	$A0[0],$A1[1]
+	 mov	16($tptr,$j),$A0[0]	# t[0]		# modsched #
+	adc	\$0,$A1[0]
+	mul	$m1			# n[7]*m1
+	add	%rax,$A1[1]		# n[7]*m1+"t[8]"
+	 mov	$Ni[0],%rax		#		# modsched #
+	adc	%rdx,$A1[0]
+	mov	$A1[1],($tptr)		# "t[8]"
+
+	xor	$topbit,$topbit
+	add	8($tptr),$A1[0]		# +t[9]
+	adc	$topbit,$topbit
+	add	$A0[1],$A1[0]
+	lea	16($tptr),$tptr		# "t[$num]>>128"
+	adc	\$0,$topbit
+	mov	$A1[0],-8($tptr)	# "t[9]"
+	cmp	8(%rsp),$tptr		# are we done?
+	jb	.Lsqr4x_mont_outer
+
+	mov	0(%rsp),$num		# restore $num
+	mov	$topbit,($tptr)		# save $topbit
+___
+}
+##############################################################
+# Post-condition, 4x unrolled copy from bn_mul_mont
+#
+{
+my ($tptr,$nptr)=("%rbx",$aptr);
+my @ri=("%rax","%rdx","%r10","%r11");
+$code.=<<___;
+	mov	64(%rsp,$num),@ri[0]	# tp[0]
+	lea	64(%rsp,$num),$tptr	# upper half of t[2*$num] holds result
+	mov	40(%rsp),$nptr		# restore $nptr
+	shr	\$5,$num		# num/4
+	mov	8($tptr),@ri[1]		# t[1]
+	xor	$i,$i			# i=0 and clear CF!
+
+	mov	32(%rsp),$rptr		# restore $rptr
+	sub	0($nptr),@ri[0]
+	mov	16($tptr),@ri[2]	# t[2]
+	mov	24($tptr),@ri[3]	# t[3]
+	sbb	8($nptr),@ri[1]
+	lea	-1($num),$j		# j=num/4-1
+	jmp	.Lsqr4x_sub
+.align	16
+.Lsqr4x_sub:
+	mov	@ri[0],0($rptr,$i,8)	# rp[i]=tp[i]-np[i]
+	mov	@ri[1],8($rptr,$i,8)	# rp[i]=tp[i]-np[i]
+	sbb	16($nptr,$i,8),@ri[2]
+	mov	32($tptr,$i,8),@ri[0]	# tp[i+1]
+	mov	40($tptr,$i,8),@ri[1]
+	sbb	24($nptr,$i,8),@ri[3]
+	mov	@ri[2],16($rptr,$i,8)	# rp[i]=tp[i]-np[i]
+	mov	@ri[3],24($rptr,$i,8)	# rp[i]=tp[i]-np[i]
+	sbb	32($nptr,$i,8),@ri[0]
+	mov	48($tptr,$i,8),@ri[2]
+	mov	56($tptr,$i,8),@ri[3]
+	sbb	40($nptr,$i,8),@ri[1]
+	lea	4($i),$i		# i++
+	dec	$j			# doesn't affect CF!
+	jnz	.Lsqr4x_sub
+
+	mov	@ri[0],0($rptr,$i,8)	# rp[i]=tp[i]-np[i]
+	mov	32($tptr,$i,8),@ri[0]	# load overflow bit
+	sbb	16($nptr,$i,8),@ri[2]
+	mov	@ri[1],8($rptr,$i,8)	# rp[i]=tp[i]-np[i]
+	sbb	24($nptr,$i,8),@ri[3]
+	mov	@ri[2],16($rptr,$i,8)	# rp[i]=tp[i]-np[i]
+
+	sbb	\$0,@ri[0]		# handle upmost overflow bit
+	mov	@ri[3],24($rptr,$i,8)	# rp[i]=tp[i]-np[i]
+	xor	$i,$i			# i=0
+	and	@ri[0],$tptr
+	not	@ri[0]
+	mov	$rptr,$nptr
+	and	@ri[0],$nptr
+	lea	-1($num),$j
+	or	$nptr,$tptr		# tp=borrow?tp:rp
+
+	pxor	%xmm0,%xmm0
+	lea	64(%rsp,$num,8),$nptr
+	movdqu	($tptr),%xmm1
+	lea	($nptr,$num,8),$nptr
+	movdqa	%xmm0,64(%rsp)		# zap lower half of temporary vector
+	movdqa	%xmm0,($nptr)		# zap upper half of temporary vector
+	movdqu	%xmm1,($rptr)
+	jmp	.Lsqr4x_copy
+.align	16
+.Lsqr4x_copy:				# copy or in-place refresh
+	movdqu	16($tptr,$i),%xmm2
+	movdqu	32($tptr,$i),%xmm1
+	movdqa	%xmm0,80(%rsp,$i)	# zap lower half of temporary vector
+	movdqa	%xmm0,96(%rsp,$i)	# zap lower half of temporary vector
+	movdqa	%xmm0,16($nptr,$i)	# zap upper half of temporary vector
+	movdqa	%xmm0,32($nptr,$i)	# zap upper half of temporary vector
+	movdqu	%xmm2,16($rptr,$i)
+	movdqu	%xmm1,32($rptr,$i)
+	lea	32($i),$i
 	dec	$j
-	jge	.Lcopy
+	jnz	.Lsqr4x_copy
 
-	mov	8(%rsp,$num,8),%rsp	# restore %rsp
+	movdqu	16($tptr,$i),%xmm2
+	movdqa	%xmm0,80(%rsp,$i)	# zap lower half of temporary vector
+	movdqa	%xmm0,16($nptr,$i)	# zap upper half of temporary vector
+	movdqu	%xmm2,16($rptr,$i)
+___
+}
+$code.=<<___;
+	mov	56(%rsp),%rsi		# restore %rsp
 	mov	\$1,%rax
+	mov	0(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lsqr4x_epilogue:
+	ret
+.size	bn_sqr4x_mont,.-bn_sqr4x_mont
+___
+}}}
+$code.=<<___;
+.asciz	"Montgomery Multiplication for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+.align	16
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	mul_handler,\@abi-omnipotent
+.align	16
+mul_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# end of prologue label
+	cmp	%r10,%rbx		# context->Rip<end of prologue label
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lcommon_seh_tail
+
+	mov	192($context),%r10	# pull $num
+	mov	8(%rax,%r10,8),%rax	# pull saved stack pointer
+	lea	48(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	-40(%rax),%r14
+	mov	-48(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+	jmp	.Lcommon_seh_tail
+.size	mul_handler,.-mul_handler
+
+.type	sqr_handler,\@abi-omnipotent
+.align	16
+sqr_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lsqr4x_body(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lsqr_body
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lsqr4x_epilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lsqr_epilogue
+	jae	.Lcommon_seh_tail
+
+	mov	56(%rax),%rax		# pull saved stack pointer
+	lea	48(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	-40(%rax),%r14
+	mov	-48(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lcommon_seh_tail:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
 	pop	%r15
 	pop	%r14
 	pop	%r13
 	pop	%r12
 	pop	%rbp
 	pop	%rbx
+	pop	%rdi
+	pop	%rsi
 	ret
-.size	bn_mul_mont,.-bn_mul_mont
-.asciz	"Montgomery Multiplication for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+.size	sqr_handler,.-sqr_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_bn_mul_mont
+	.rva	.LSEH_end_bn_mul_mont
+	.rva	.LSEH_info_bn_mul_mont
+
+	.rva	.LSEH_begin_bn_mul4x_mont
+	.rva	.LSEH_end_bn_mul4x_mont
+	.rva	.LSEH_info_bn_mul4x_mont
+
+	.rva	.LSEH_begin_bn_sqr4x_mont
+	.rva	.LSEH_end_bn_sqr4x_mont
+	.rva	.LSEH_info_bn_sqr4x_mont
+
+.section	.xdata
+.align	8
+.LSEH_info_bn_mul_mont:
+	.byte	9,0,0,0
+	.rva	mul_handler
+	.rva	.Lmul_body,.Lmul_epilogue	# HandlerData[]
+.LSEH_info_bn_mul4x_mont:
+	.byte	9,0,0,0
+	.rva	mul_handler
+	.rva	.Lmul4x_body,.Lmul4x_epilogue	# HandlerData[]
+.LSEH_info_bn_sqr4x_mont:
+	.byte	9,0,0,0
+	.rva	sqr_handler
 ___
+}
 
 print $code;
 close STDOUT;
diff --git a/crypto/bn/asm/x86_64-mont5.pl b/crypto/bn/asm/x86_64-mont5.pl
new file mode 100755
index 000000000000..057cda28aaed
--- /dev/null
+++ b/crypto/bn/asm/x86_64-mont5.pl
@@ -0,0 +1,1070 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# August 2011.
+#
+# Companion to x86_64-mont.pl that optimizes cache-timing attack
+# countermeasures. The subroutines are produced by replacing bp[i]
+# references in their x86_64-mont.pl counterparts with cache-neutral
+# references to powers table computed in BN_mod_exp_mont_consttime.
+# In addition subroutine that scatters elements of the powers table
+# is implemented, so that scatter-/gathering can be tuned without
+# bn_exp.c modifications.
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+# int bn_mul_mont_gather5(
+$rp="%rdi";	# BN_ULONG *rp,
+$ap="%rsi";	# const BN_ULONG *ap,
+$bp="%rdx";	# const BN_ULONG *bp,
+$np="%rcx";	# const BN_ULONG *np,
+$n0="%r8";	# const BN_ULONG *n0,
+$num="%r9";	# int num,
+		# int idx);	# 0 to 2^5-1, "index" in $bp holding
+				# pre-computed powers of a', interlaced
+				# in such manner that b[0] is $bp[idx],
+				# b[1] is [2^5+idx], etc.
+$lo0="%r10";
+$hi0="%r11";
+$hi1="%r13";
+$i="%r14";
+$j="%r15";
+$m0="%rbx";
+$m1="%rbp";
+
+$code=<<___;
+.text
+
+.globl	bn_mul_mont_gather5
+.type	bn_mul_mont_gather5,\@function,6
+.align	64
+bn_mul_mont_gather5:
+	test	\$3,${num}d
+	jnz	.Lmul_enter
+	cmp	\$8,${num}d
+	jb	.Lmul_enter
+	jmp	.Lmul4x_enter
+
+.align	16
+.Lmul_enter:
+	mov	${num}d,${num}d
+	mov	`($win64?56:8)`(%rsp),%r10d	# load 7th argument
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+___
+$code.=<<___ if ($win64);
+	lea	-0x28(%rsp),%rsp
+	movaps	%xmm6,(%rsp)
+	movaps	%xmm7,0x10(%rsp)
+.Lmul_alloca:
+___
+$code.=<<___;
+	mov	%rsp,%rax
+	lea	2($num),%r11
+	neg	%r11
+	lea	(%rsp,%r11,8),%rsp	# tp=alloca(8*(num+2))
+	and	\$-1024,%rsp		# minimize TLB usage
+
+	mov	%rax,8(%rsp,$num,8)	# tp[num+1]=%rsp
+.Lmul_body:
+	mov	$bp,%r12		# reassign $bp
+___
+		$bp="%r12";
+		$STRIDE=2**5*8;		# 5 is "window size"
+		$N=$STRIDE/4;		# should match cache line size
+$code.=<<___;
+	mov	%r10,%r11
+	shr	\$`log($N/8)/log(2)`,%r10
+	and	\$`$N/8-1`,%r11
+	not	%r10
+	lea	.Lmagic_masks(%rip),%rax
+	and	\$`2**5/($N/8)-1`,%r10	# 5 is "window size"
+	lea	96($bp,%r11,8),$bp	# pointer within 1st cache line
+	movq	0(%rax,%r10,8),%xmm4	# set of masks denoting which
+	movq	8(%rax,%r10,8),%xmm5	# cache line contains element
+	movq	16(%rax,%r10,8),%xmm6	# denoted by 7th argument
+	movq	24(%rax,%r10,8),%xmm7
+
+	movq	`0*$STRIDE/4-96`($bp),%xmm0
+	movq	`1*$STRIDE/4-96`($bp),%xmm1
+	pand	%xmm4,%xmm0
+	movq	`2*$STRIDE/4-96`($bp),%xmm2
+	pand	%xmm5,%xmm1
+	movq	`3*$STRIDE/4-96`($bp),%xmm3
+	pand	%xmm6,%xmm2
+	por	%xmm1,%xmm0
+	pand	%xmm7,%xmm3
+	por	%xmm2,%xmm0
+	lea	$STRIDE($bp),$bp
+	por	%xmm3,%xmm0
+
+	movq	%xmm0,$m0		# m0=bp[0]
+
+	mov	($n0),$n0		# pull n0[0] value
+	mov	($ap),%rax
+
+	xor	$i,$i			# i=0
+	xor	$j,$j			# j=0
+
+	movq	`0*$STRIDE/4-96`($bp),%xmm0
+	movq	`1*$STRIDE/4-96`($bp),%xmm1
+	pand	%xmm4,%xmm0
+	movq	`2*$STRIDE/4-96`($bp),%xmm2
+	pand	%xmm5,%xmm1
+
+	mov	$n0,$m1
+	mulq	$m0			# ap[0]*bp[0]
+	mov	%rax,$lo0
+	mov	($np),%rax
+
+	movq	`3*$STRIDE/4-96`($bp),%xmm3
+	pand	%xmm6,%xmm2
+	por	%xmm1,%xmm0
+	pand	%xmm7,%xmm3
+
+	imulq	$lo0,$m1		# "tp[0]"*n0
+	mov	%rdx,$hi0
+
+	por	%xmm2,%xmm0
+	lea	$STRIDE($bp),$bp
+	por	%xmm3,%xmm0
+
+	mulq	$m1			# np[0]*m1
+	add	%rax,$lo0		# discarded
+	mov	8($ap),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$hi1
+
+	lea	1($j),$j		# j++
+	jmp	.L1st_enter
+
+.align	16
+.L1st:
+	add	%rax,$hi1
+	mov	($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$hi0,$hi1		# np[j]*m1+ap[j]*bp[0]
+	mov	$lo0,$hi0
+	adc	\$0,%rdx
+	mov	$hi1,-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$hi1
+
+.L1st_enter:
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$hi0
+	mov	($np,$j,8),%rax
+	adc	\$0,%rdx
+	lea	1($j),$j		# j++
+	mov	%rdx,$lo0
+
+	mulq	$m1			# np[j]*m1
+	cmp	$num,$j
+	jne	.L1st
+
+	movq	%xmm0,$m0		# bp[1]
+
+	add	%rax,$hi1
+	mov	($ap),%rax		# ap[0]
+	adc	\$0,%rdx
+	add	$hi0,$hi1		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$hi1,-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$hi1
+	mov	$lo0,$hi0
+
+	xor	%rdx,%rdx
+	add	$hi0,$hi1
+	adc	\$0,%rdx
+	mov	$hi1,-8(%rsp,$num,8)
+	mov	%rdx,(%rsp,$num,8)	# store upmost overflow bit
+
+	lea	1($i),$i		# i++
+	jmp	.Louter
+.align	16
+.Louter:
+	xor	$j,$j			# j=0
+	mov	$n0,$m1
+	mov	(%rsp),$lo0
+
+	movq	`0*$STRIDE/4-96`($bp),%xmm0
+	movq	`1*$STRIDE/4-96`($bp),%xmm1
+	pand	%xmm4,%xmm0
+	movq	`2*$STRIDE/4-96`($bp),%xmm2
+	pand	%xmm5,%xmm1
+
+	mulq	$m0			# ap[0]*bp[i]
+	add	%rax,$lo0		# ap[0]*bp[i]+tp[0]
+	mov	($np),%rax
+	adc	\$0,%rdx
+
+	movq	`3*$STRIDE/4-96`($bp),%xmm3
+	pand	%xmm6,%xmm2
+	por	%xmm1,%xmm0
+	pand	%xmm7,%xmm3
+
+	imulq	$lo0,$m1		# tp[0]*n0
+	mov	%rdx,$hi0
+
+	por	%xmm2,%xmm0
+	lea	$STRIDE($bp),$bp
+	por	%xmm3,%xmm0
+
+	mulq	$m1			# np[0]*m1
+	add	%rax,$lo0		# discarded
+	mov	8($ap),%rax
+	adc	\$0,%rdx
+	mov	8(%rsp),$lo0		# tp[1]
+	mov	%rdx,$hi1
+
+	lea	1($j),$j		# j++
+	jmp	.Linner_enter
+
+.align	16
+.Linner:
+	add	%rax,$hi1
+	mov	($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$lo0,$hi1		# np[j]*m1+ap[j]*bp[i]+tp[j]
+	mov	(%rsp,$j,8),$lo0
+	adc	\$0,%rdx
+	mov	$hi1,-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$hi1
+
+.Linner_enter:
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$hi0
+	mov	($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	$hi0,$lo0		# ap[j]*bp[i]+tp[j]
+	mov	%rdx,$hi0
+	adc	\$0,$hi0
+	lea	1($j),$j		# j++
+
+	mulq	$m1			# np[j]*m1
+	cmp	$num,$j
+	jne	.Linner
+
+	movq	%xmm0,$m0		# bp[i+1]
+
+	add	%rax,$hi1
+	mov	($ap),%rax		# ap[0]
+	adc	\$0,%rdx
+	add	$lo0,$hi1		# np[j]*m1+ap[j]*bp[i]+tp[j]
+	mov	(%rsp,$j,8),$lo0
+	adc	\$0,%rdx
+	mov	$hi1,-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$hi1
+
+	xor	%rdx,%rdx
+	add	$hi0,$hi1
+	adc	\$0,%rdx
+	add	$lo0,$hi1		# pull upmost overflow bit
+	adc	\$0,%rdx
+	mov	$hi1,-8(%rsp,$num,8)
+	mov	%rdx,(%rsp,$num,8)	# store upmost overflow bit
+
+	lea	1($i),$i		# i++
+	cmp	$num,$i
+	jl	.Louter
+
+	xor	$i,$i			# i=0 and clear CF!
+	mov	(%rsp),%rax		# tp[0]
+	lea	(%rsp),$ap		# borrow ap for tp
+	mov	$num,$j			# j=num
+	jmp	.Lsub
+.align	16
+.Lsub:	sbb	($np,$i,8),%rax
+	mov	%rax,($rp,$i,8)		# rp[i]=tp[i]-np[i]
+	mov	8($ap,$i,8),%rax	# tp[i+1]
+	lea	1($i),$i		# i++
+	dec	$j			# doesnn't affect CF!
+	jnz	.Lsub
+
+	sbb	\$0,%rax		# handle upmost overflow bit
+	xor	$i,$i
+	and	%rax,$ap
+	not	%rax
+	mov	$rp,$np
+	and	%rax,$np
+	mov	$num,$j			# j=num
+	or	$np,$ap			# ap=borrow?tp:rp
+.align	16
+.Lcopy:					# copy or in-place refresh
+	mov	($ap,$i,8),%rax
+	mov	$i,(%rsp,$i,8)		# zap temporary vector
+	mov	%rax,($rp,$i,8)		# rp[i]=tp[i]
+	lea	1($i),$i
+	sub	\$1,$j
+	jnz	.Lcopy
+
+	mov	8(%rsp,$num,8),%rsi	# restore %rsp
+	mov	\$1,%rax
+___
+$code.=<<___ if ($win64);
+	movaps	(%rsi),%xmm6
+	movaps	0x10(%rsi),%xmm7
+	lea	0x28(%rsi),%rsi
+___
+$code.=<<___;
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lmul_epilogue:
+	ret
+.size	bn_mul_mont_gather5,.-bn_mul_mont_gather5
+___
+{{{
+my @A=("%r10","%r11");
+my @N=("%r13","%rdi");
+$code.=<<___;
+.type	bn_mul4x_mont_gather5,\@function,6
+.align	16
+bn_mul4x_mont_gather5:
+.Lmul4x_enter:
+	mov	${num}d,${num}d
+	mov	`($win64?56:8)`(%rsp),%r10d	# load 7th argument
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+___
+$code.=<<___ if ($win64);
+	lea	-0x28(%rsp),%rsp
+	movaps	%xmm6,(%rsp)
+	movaps	%xmm7,0x10(%rsp)
+.Lmul4x_alloca:
+___
+$code.=<<___;
+	mov	%rsp,%rax
+	lea	4($num),%r11
+	neg	%r11
+	lea	(%rsp,%r11,8),%rsp	# tp=alloca(8*(num+4))
+	and	\$-1024,%rsp		# minimize TLB usage
+
+	mov	%rax,8(%rsp,$num,8)	# tp[num+1]=%rsp
+.Lmul4x_body:
+	mov	$rp,16(%rsp,$num,8)	# tp[num+2]=$rp
+	mov	%rdx,%r12		# reassign $bp
+___
+		$bp="%r12";
+		$STRIDE=2**5*8;		# 5 is "window size"
+		$N=$STRIDE/4;		# should match cache line size
+$code.=<<___;
+	mov	%r10,%r11
+	shr	\$`log($N/8)/log(2)`,%r10
+	and	\$`$N/8-1`,%r11
+	not	%r10
+	lea	.Lmagic_masks(%rip),%rax
+	and	\$`2**5/($N/8)-1`,%r10	# 5 is "window size"
+	lea	96($bp,%r11,8),$bp	# pointer within 1st cache line
+	movq	0(%rax,%r10,8),%xmm4	# set of masks denoting which
+	movq	8(%rax,%r10,8),%xmm5	# cache line contains element
+	movq	16(%rax,%r10,8),%xmm6	# denoted by 7th argument
+	movq	24(%rax,%r10,8),%xmm7
+
+	movq	`0*$STRIDE/4-96`($bp),%xmm0
+	movq	`1*$STRIDE/4-96`($bp),%xmm1
+	pand	%xmm4,%xmm0
+	movq	`2*$STRIDE/4-96`($bp),%xmm2
+	pand	%xmm5,%xmm1
+	movq	`3*$STRIDE/4-96`($bp),%xmm3
+	pand	%xmm6,%xmm2
+	por	%xmm1,%xmm0
+	pand	%xmm7,%xmm3
+	por	%xmm2,%xmm0
+	lea	$STRIDE($bp),$bp
+	por	%xmm3,%xmm0
+
+	movq	%xmm0,$m0		# m0=bp[0]
+	mov	($n0),$n0		# pull n0[0] value
+	mov	($ap),%rax
+
+	xor	$i,$i			# i=0
+	xor	$j,$j			# j=0
+
+	movq	`0*$STRIDE/4-96`($bp),%xmm0
+	movq	`1*$STRIDE/4-96`($bp),%xmm1
+	pand	%xmm4,%xmm0
+	movq	`2*$STRIDE/4-96`($bp),%xmm2
+	pand	%xmm5,%xmm1
+
+	mov	$n0,$m1
+	mulq	$m0			# ap[0]*bp[0]
+	mov	%rax,$A[0]
+	mov	($np),%rax
+
+	movq	`3*$STRIDE/4-96`($bp),%xmm3
+	pand	%xmm6,%xmm2
+	por	%xmm1,%xmm0
+	pand	%xmm7,%xmm3
+
+	imulq	$A[0],$m1		# "tp[0]"*n0
+	mov	%rdx,$A[1]
+
+	por	%xmm2,%xmm0
+	lea	$STRIDE($bp),$bp
+	por	%xmm3,%xmm0
+
+	mulq	$m1			# np[0]*m1
+	add	%rax,$A[0]		# discarded
+	mov	8($ap),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$N[1]
+
+	mulq	$m0
+	add	%rax,$A[1]
+	mov	8($np),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1
+	add	%rax,$N[1]
+	mov	16($ap),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]
+	lea	4($j),$j		# j++
+	adc	\$0,%rdx
+	mov	$N[1],(%rsp)
+	mov	%rdx,$N[0]
+	jmp	.L1st4x
+.align	16
+.L1st4x:
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[0]
+	mov	-16($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	-8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[0],-24(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[1]
+	mov	-8($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[1],-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[0]
+	mov	($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[0],-8(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[1]
+	mov	8($np,$j,8),%rax
+	adc	\$0,%rdx
+	lea	4($j),$j		# j++
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	-16($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[1],-32(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+	cmp	$num,$j
+	jl	.L1st4x
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[0]
+	mov	-16($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	-8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[0],-24(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[0]
+	add	%rax,$A[1]
+	mov	-8($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	($ap),%rax		# ap[0]
+	adc	\$0,%rdx
+	add	$A[1],$N[1]		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	$N[1],-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+
+	movq	%xmm0,$m0		# bp[1]
+
+	xor	$N[1],$N[1]
+	add	$A[0],$N[0]
+	adc	\$0,$N[1]
+	mov	$N[0],-8(%rsp,$j,8)
+	mov	$N[1],(%rsp,$j,8)	# store upmost overflow bit
+
+	lea	1($i),$i		# i++
+.align	4
+.Louter4x:
+	xor	$j,$j			# j=0
+	movq	`0*$STRIDE/4-96`($bp),%xmm0
+	movq	`1*$STRIDE/4-96`($bp),%xmm1
+	pand	%xmm4,%xmm0
+	movq	`2*$STRIDE/4-96`($bp),%xmm2
+	pand	%xmm5,%xmm1
+
+	mov	(%rsp),$A[0]
+	mov	$n0,$m1
+	mulq	$m0			# ap[0]*bp[i]
+	add	%rax,$A[0]		# ap[0]*bp[i]+tp[0]
+	mov	($np),%rax
+	adc	\$0,%rdx
+
+	movq	`3*$STRIDE/4-96`($bp),%xmm3
+	pand	%xmm6,%xmm2
+	por	%xmm1,%xmm0
+	pand	%xmm7,%xmm3
+
+	imulq	$A[0],$m1		# tp[0]*n0
+	mov	%rdx,$A[1]
+
+	por	%xmm2,%xmm0
+	lea	$STRIDE($bp),$bp
+	por	%xmm3,%xmm0
+
+	mulq	$m1			# np[0]*m1
+	add	%rax,$A[0]		# "$N[0]", discarded
+	mov	8($ap),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[1]
+	mov	8($np),%rax
+	adc	\$0,%rdx
+	add	8(%rsp),$A[1]		# +tp[1]
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	16($ap),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]		# np[j]*m1+ap[j]*bp[i]+tp[j]
+	lea	4($j),$j		# j+=2
+	adc	\$0,%rdx
+	mov	%rdx,$N[0]
+	jmp	.Linner4x
+.align	16
+.Linner4x:
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[0]
+	mov	-16($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	-16(%rsp,$j,8),$A[0]	# ap[j]*bp[i]+tp[j]
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	-8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]
+	adc	\$0,%rdx
+	mov	$N[1],-32(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[1]
+	mov	-8($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	-8(%rsp,$j,8),$A[1]
+	adc	\$0,%rdx
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]
+	adc	\$0,%rdx
+	mov	$N[0],-24(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[0]
+	mov	($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	(%rsp,$j,8),$A[0]	# ap[j]*bp[i]+tp[j]
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]
+	adc	\$0,%rdx
+	mov	$N[1],-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[1]
+	mov	8($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	8(%rsp,$j,8),$A[1]
+	adc	\$0,%rdx
+	lea	4($j),$j		# j++
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	-16($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[1],$N[1]
+	adc	\$0,%rdx
+	mov	$N[0],-40(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+	cmp	$num,$j
+	jl	.Linner4x
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[0]
+	mov	-16($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	-16(%rsp,$j,8),$A[0]	# ap[j]*bp[i]+tp[j]
+	adc	\$0,%rdx
+	mov	%rdx,$A[1]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[0]
+	mov	-8($ap,$j,8),%rax
+	adc	\$0,%rdx
+	add	$A[0],$N[0]
+	adc	\$0,%rdx
+	mov	$N[1],-32(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[1]
+
+	mulq	$m0			# ap[j]*bp[i]
+	add	%rax,$A[1]
+	mov	-8($np,$j,8),%rax
+	adc	\$0,%rdx
+	add	-8(%rsp,$j,8),$A[1]
+	adc	\$0,%rdx
+	lea	1($i),$i		# i++
+	mov	%rdx,$A[0]
+
+	mulq	$m1			# np[j]*m1
+	add	%rax,$N[1]
+	mov	($ap),%rax		# ap[0]
+	adc	\$0,%rdx
+	add	$A[1],$N[1]
+	adc	\$0,%rdx
+	mov	$N[0],-24(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$N[0]
+
+	movq	%xmm0,$m0		# bp[i+1]
+	mov	$N[1],-16(%rsp,$j,8)	# tp[j-1]
+
+	xor	$N[1],$N[1]
+	add	$A[0],$N[0]
+	adc	\$0,$N[1]
+	add	(%rsp,$num,8),$N[0]	# pull upmost overflow bit
+	adc	\$0,$N[1]
+	mov	$N[0],-8(%rsp,$j,8)
+	mov	$N[1],(%rsp,$j,8)	# store upmost overflow bit
+
+	cmp	$num,$i
+	jl	.Louter4x
+___
+{
+my @ri=("%rax","%rdx",$m0,$m1);
+$code.=<<___;
+	mov	16(%rsp,$num,8),$rp	# restore $rp
+	mov	0(%rsp),@ri[0]		# tp[0]
+	pxor	%xmm0,%xmm0
+	mov	8(%rsp),@ri[1]		# tp[1]
+	shr	\$2,$num		# num/=4
+	lea	(%rsp),$ap		# borrow ap for tp
+	xor	$i,$i			# i=0 and clear CF!
+
+	sub	0($np),@ri[0]
+	mov	16($ap),@ri[2]		# tp[2]
+	mov	24($ap),@ri[3]		# tp[3]
+	sbb	8($np),@ri[1]
+	lea	-1($num),$j		# j=num/4-1
+	jmp	.Lsub4x
+.align	16
+.Lsub4x:
+	mov	@ri[0],0($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	mov	@ri[1],8($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	sbb	16($np,$i,8),@ri[2]
+	mov	32($ap,$i,8),@ri[0]	# tp[i+1]
+	mov	40($ap,$i,8),@ri[1]
+	sbb	24($np,$i,8),@ri[3]
+	mov	@ri[2],16($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	mov	@ri[3],24($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	sbb	32($np,$i,8),@ri[0]
+	mov	48($ap,$i,8),@ri[2]
+	mov	56($ap,$i,8),@ri[3]
+	sbb	40($np,$i,8),@ri[1]
+	lea	4($i),$i		# i++
+	dec	$j			# doesnn't affect CF!
+	jnz	.Lsub4x
+
+	mov	@ri[0],0($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	mov	32($ap,$i,8),@ri[0]	# load overflow bit
+	sbb	16($np,$i,8),@ri[2]
+	mov	@ri[1],8($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	sbb	24($np,$i,8),@ri[3]
+	mov	@ri[2],16($rp,$i,8)	# rp[i]=tp[i]-np[i]
+
+	sbb	\$0,@ri[0]		# handle upmost overflow bit
+	mov	@ri[3],24($rp,$i,8)	# rp[i]=tp[i]-np[i]
+	xor	$i,$i			# i=0
+	and	@ri[0],$ap
+	not	@ri[0]
+	mov	$rp,$np
+	and	@ri[0],$np
+	lea	-1($num),$j
+	or	$np,$ap			# ap=borrow?tp:rp
+
+	movdqu	($ap),%xmm1
+	movdqa	%xmm0,(%rsp)
+	movdqu	%xmm1,($rp)
+	jmp	.Lcopy4x
+.align	16
+.Lcopy4x:					# copy or in-place refresh
+	movdqu	16($ap,$i),%xmm2
+	movdqu	32($ap,$i),%xmm1
+	movdqa	%xmm0,16(%rsp,$i)
+	movdqu	%xmm2,16($rp,$i)
+	movdqa	%xmm0,32(%rsp,$i)
+	movdqu	%xmm1,32($rp,$i)
+	lea	32($i),$i
+	dec	$j
+	jnz	.Lcopy4x
+
+	shl	\$2,$num
+	movdqu	16($ap,$i),%xmm2
+	movdqa	%xmm0,16(%rsp,$i)
+	movdqu	%xmm2,16($rp,$i)
+___
+}
+$code.=<<___;
+	mov	8(%rsp,$num,8),%rsi	# restore %rsp
+	mov	\$1,%rax
+___
+$code.=<<___ if ($win64);
+	movaps	(%rsi),%xmm6
+	movaps	0x10(%rsi),%xmm7
+	lea	0x28(%rsi),%rsi
+___
+$code.=<<___;
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lmul4x_epilogue:
+	ret
+.size	bn_mul4x_mont_gather5,.-bn_mul4x_mont_gather5
+___
+}}}
+
+{
+my ($inp,$num,$tbl,$idx)=$win64?("%rcx","%rdx","%r8", "%r9") : # Win64 order
+				("%rdi","%rsi","%rdx","%rcx"); # Unix order
+my $out=$inp;
+my $STRIDE=2**5*8;
+my $N=$STRIDE/4;
+
+$code.=<<___;
+.globl	bn_scatter5
+.type	bn_scatter5,\@abi-omnipotent
+.align	16
+bn_scatter5:
+	cmp	\$0, $num
+	jz	.Lscatter_epilogue
+	lea	($tbl,$idx,8),$tbl
+.Lscatter:
+	mov	($inp),%rax
+	lea	8($inp),$inp
+	mov	%rax,($tbl)
+	lea	32*8($tbl),$tbl
+	sub	\$1,$num
+	jnz	.Lscatter
+.Lscatter_epilogue:
+	ret
+.size	bn_scatter5,.-bn_scatter5
+
+.globl	bn_gather5
+.type	bn_gather5,\@abi-omnipotent
+.align	16
+bn_gather5:
+___
+$code.=<<___ if ($win64);
+.LSEH_begin_bn_gather5:
+	# I can't trust assembler to use specific encoding:-(
+	.byte	0x48,0x83,0xec,0x28		#sub	\$0x28,%rsp
+	.byte	0x0f,0x29,0x34,0x24		#movaps	%xmm6,(%rsp)
+	.byte	0x0f,0x29,0x7c,0x24,0x10	#movdqa	%xmm7,0x10(%rsp)
+___
+$code.=<<___;
+	mov	$idx,%r11
+	shr	\$`log($N/8)/log(2)`,$idx
+	and	\$`$N/8-1`,%r11
+	not	$idx
+	lea	.Lmagic_masks(%rip),%rax
+	and	\$`2**5/($N/8)-1`,$idx	# 5 is "window size"
+	lea	96($tbl,%r11,8),$tbl	# pointer within 1st cache line
+	movq	0(%rax,$idx,8),%xmm4	# set of masks denoting which
+	movq	8(%rax,$idx,8),%xmm5	# cache line contains element
+	movq	16(%rax,$idx,8),%xmm6	# denoted by 7th argument
+	movq	24(%rax,$idx,8),%xmm7
+	jmp	.Lgather
+.align	16
+.Lgather:
+	movq	`0*$STRIDE/4-96`($tbl),%xmm0
+	movq	`1*$STRIDE/4-96`($tbl),%xmm1
+	pand	%xmm4,%xmm0
+	movq	`2*$STRIDE/4-96`($tbl),%xmm2
+	pand	%xmm5,%xmm1
+	movq	`3*$STRIDE/4-96`($tbl),%xmm3
+	pand	%xmm6,%xmm2
+	por	%xmm1,%xmm0
+	pand	%xmm7,%xmm3
+	por	%xmm2,%xmm0
+	lea	$STRIDE($tbl),$tbl
+	por	%xmm3,%xmm0
+
+	movq	%xmm0,($out)		# m0=bp[0]
+	lea	8($out),$out
+	sub	\$1,$num
+	jnz	.Lgather
+___
+$code.=<<___ if ($win64);
+	movaps	%xmm6,(%rsp)
+	movaps	%xmm7,0x10(%rsp)
+	lea	0x28(%rsp),%rsp
+___
+$code.=<<___;
+	ret
+.LSEH_end_bn_gather5:
+.size	bn_gather5,.-bn_gather5
+___
+}
+$code.=<<___;
+.align	64
+.Lmagic_masks:
+	.long	0,0, 0,0, 0,0, -1,-1
+	.long	0,0, 0,0, 0,0,  0,0
+.asciz	"Montgomery Multiplication with scatter/gather for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	mul_handler,\@abi-omnipotent
+.align	16
+mul_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# end of prologue label
+	cmp	%r10,%rbx		# context->Rip<end of prologue label
+	jb	.Lcommon_seh_tail
+
+	lea	`40+48`(%rax),%rax
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# end of alloca label
+	cmp	%r10,%rbx		# context->Rip<end of alloca label
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	8(%r11),%r10d		# HandlerData[2]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lcommon_seh_tail
+
+	mov	192($context),%r10	# pull $num
+	mov	8(%rax,%r10,8),%rax	# pull saved stack pointer
+
+	movaps	(%rax),%xmm0
+	movaps	16(%rax),%xmm1
+	lea	`40+48`(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	-40(%rax),%r14
+	mov	-48(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+	movups	%xmm0,512($context)	# restore context->Xmm6
+	movups	%xmm1,528($context)	# restore context->Xmm7
+
+.Lcommon_seh_tail:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	mul_handler,.-mul_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_bn_mul_mont_gather5
+	.rva	.LSEH_end_bn_mul_mont_gather5
+	.rva	.LSEH_info_bn_mul_mont_gather5
+
+	.rva	.LSEH_begin_bn_mul4x_mont_gather5
+	.rva	.LSEH_end_bn_mul4x_mont_gather5
+	.rva	.LSEH_info_bn_mul4x_mont_gather5
+
+	.rva	.LSEH_begin_bn_gather5
+	.rva	.LSEH_end_bn_gather5
+	.rva	.LSEH_info_bn_gather5
+
+.section	.xdata
+.align	8
+.LSEH_info_bn_mul_mont_gather5:
+	.byte	9,0,0,0
+	.rva	mul_handler
+	.rva	.Lmul_alloca,.Lmul_body,.Lmul_epilogue		# HandlerData[]
+.align	8
+.LSEH_info_bn_mul4x_mont_gather5:
+	.byte	9,0,0,0
+	.rva	mul_handler
+	.rva	.Lmul4x_alloca,.Lmul4x_body,.Lmul4x_epilogue	# HandlerData[]
+.align	8
+.LSEH_info_bn_gather5:
+        .byte   0x01,0x0d,0x05,0x00
+        .byte   0x0d,0x78,0x01,0x00	#movaps	0x10(rsp),xmm7
+        .byte   0x08,0x68,0x00,0x00	#movaps	(rsp),xmm6
+        .byte   0x04,0x42,0x00,0x00	#sub	rsp,0x28
+.align	8
+___
+}
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+
+print $code;
+close STDOUT;
diff --git a/crypto/bn/bn.h b/crypto/bn/bn.h
index f1719a5877f7..f34248ec4f87 100644
--- a/crypto/bn/bn.h
+++ b/crypto/bn/bn.h
@@ -56,6 +56,59 @@
  * [including the GNU Public Licence.]
  */
 /* ====================================================================
+ * Copyright (c) 1998-2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+/* ====================================================================
  * Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
  *
  * Portions of the attached software ("Contribution") are developed by 
@@ -77,6 +130,7 @@
 #include <stdio.h> /* FILE */
 #endif
 #include <openssl/ossl_typ.h>
+#include <openssl/crypto.h>
 
 #ifdef  __cplusplus
 extern "C" {
@@ -94,9 +148,11 @@ extern "C" {
 /* #define BN_DEBUG */
 /* #define BN_DEBUG_RAND */
 
+#ifndef OPENSSL_SMALL_FOOTPRINT
 #define BN_MUL_COMBA
 #define BN_SQR_COMBA
 #define BN_RECURSION
+#endif
 
 /* This next option uses the C libraries (2 word)/(1 word) function.
  * If it is not defined, I use my C version (which is slower).
@@ -137,6 +193,8 @@ extern "C" {
 #define BN_DEC_FMT1	"%lu"
 #define BN_DEC_FMT2	"%019lu"
 #define BN_DEC_NUM	19
+#define BN_HEX_FMT1	"%lX"
+#define BN_HEX_FMT2	"%016lX"
 #endif
 
 /* This is where the long long data type is 64 bits, but long is 32.
@@ -162,84 +220,56 @@ extern "C" {
 #define BN_DEC_FMT1	"%llu"
 #define BN_DEC_FMT2	"%019llu"
 #define BN_DEC_NUM	19
+#define BN_HEX_FMT1	"%llX"
+#define BN_HEX_FMT2	"%016llX"
 #endif
 
 #ifdef THIRTY_TWO_BIT
 #ifdef BN_LLONG
-# if defined(OPENSSL_SYS_WIN32) && !defined(__GNUC__)
+# if defined(_WIN32) && !defined(__GNUC__)
 #  define BN_ULLONG	unsigned __int64
+#  define BN_MASK	(0xffffffffffffffffI64)
 # else
 #  define BN_ULLONG	unsigned long long
+#  define BN_MASK	(0xffffffffffffffffLL)
 # endif
 #endif
-#define BN_ULONG	unsigned long
-#define BN_LONG		long
+#define BN_ULONG	unsigned int
+#define BN_LONG		int
 #define BN_BITS		64
 #define BN_BYTES	4
 #define BN_BITS2	32
 #define BN_BITS4	16
-#ifdef OPENSSL_SYS_WIN32
-/* VC++ doesn't like the LL suffix */
-#define BN_MASK		(0xffffffffffffffffL)
-#else
-#define BN_MASK		(0xffffffffffffffffLL)
-#endif
 #define BN_MASK2	(0xffffffffL)
 #define BN_MASK2l	(0xffff)
 #define BN_MASK2h1	(0xffff8000L)
 #define BN_MASK2h	(0xffff0000L)
 #define BN_TBIT		(0x80000000L)
 #define BN_DEC_CONV	(1000000000L)
-#define BN_DEC_FMT1	"%lu"
-#define BN_DEC_FMT2	"%09lu"
-#define BN_DEC_NUM	9
-#endif
-
-#ifdef SIXTEEN_BIT
-#ifndef BN_DIV2W
-#define BN_DIV2W
-#endif
-#define BN_ULLONG	unsigned long
-#define BN_ULONG	unsigned short
-#define BN_LONG		short
-#define BN_BITS		32
-#define BN_BYTES	2
-#define BN_BITS2	16
-#define BN_BITS4	8
-#define BN_MASK		(0xffffffff)
-#define BN_MASK2	(0xffff)
-#define BN_MASK2l	(0xff)
-#define BN_MASK2h1	(0xff80)
-#define BN_MASK2h	(0xff00)
-#define BN_TBIT		(0x8000)
-#define BN_DEC_CONV	(100000)
 #define BN_DEC_FMT1	"%u"
-#define BN_DEC_FMT2	"%05u"
-#define BN_DEC_NUM	5
+#define BN_DEC_FMT2	"%09u"
+#define BN_DEC_NUM	9
+#define BN_HEX_FMT1	"%X"
+#define BN_HEX_FMT2	"%08X"
 #endif
 
-#ifdef EIGHT_BIT
-#ifndef BN_DIV2W
-#define BN_DIV2W
-#endif
-#define BN_ULLONG	unsigned short
-#define BN_ULONG	unsigned char
-#define BN_LONG		char
-#define BN_BITS		16
-#define BN_BYTES	1
-#define BN_BITS2	8
-#define BN_BITS4	4
-#define BN_MASK		(0xffff)
-#define BN_MASK2	(0xff)
-#define BN_MASK2l	(0xf)
-#define BN_MASK2h1	(0xf8)
-#define BN_MASK2h	(0xf0)
-#define BN_TBIT		(0x80)
-#define BN_DEC_CONV	(100)
-#define BN_DEC_FMT1	"%u"
-#define BN_DEC_FMT2	"%02u"
-#define BN_DEC_NUM	2
-#endif
+/* 2011-02-22 SMS.
+ * In various places, a size_t variable or a type cast to size_t was
+ * used to perform integer-only operations on pointers.  This failed on
+ * VMS with 64-bit pointers (CC /POINTER_SIZE = 64) because size_t is
+ * still only 32 bits.  What's needed in these cases is an integer type
+ * with the same size as a pointer, which size_t is not certain to be. 
+ * The only fix here is VMS-specific.
+ */
+#if defined(OPENSSL_SYS_VMS)
+# if __INITIAL_POINTER_SIZE == 64
+#  define PTR_SIZE_INT long long
+# else /* __INITIAL_POINTER_SIZE == 64 */
+#  define PTR_SIZE_INT int
+# endif /* __INITIAL_POINTER_SIZE == 64 [else] */
+#else /* defined(OPENSSL_SYS_VMS) */
+# define PTR_SIZE_INT size_t
+#endif /* defined(OPENSSL_SYS_VMS) [else] */
 
 #define BN_DEFAULT_BITS	1280
 
@@ -303,12 +333,8 @@ struct bn_mont_ctx_st
 	BIGNUM N;      /* The modulus */
 	BIGNUM Ni;     /* R*(1/R mod N) - N*Ni = 1
 	                * (Ni is only stored for bignum algorithm) */
-#if 0
-	/* OpenSSL 0.9.9 preview: */
-	BN_ULONG n0[2];/* least significant word(s) of Ni */
-#else
-	BN_ULONG n0;   /* least significant word of Ni */
-#endif
+	BN_ULONG n0[2];/* least significant word(s) of Ni;
+	                  (type changed with 0.9.9, was "BN_ULONG n0;" before) */
 	int flags;
 	};
 
@@ -504,6 +530,7 @@ char *	BN_bn2hex(const BIGNUM *a);
 char *	BN_bn2dec(const BIGNUM *a);
 int 	BN_hex2bn(BIGNUM **a, const char *str);
 int 	BN_dec2bn(BIGNUM **a, const char *str);
+int	BN_asc2bn(BIGNUM **a, const char *str);
 int	BN_gcd(BIGNUM *r,const BIGNUM *a,const BIGNUM *b,BN_CTX *ctx);
 int	BN_kronecker(const BIGNUM *a,const BIGNUM *b,BN_CTX *ctx); /* returns -2 for error */
 BIGNUM *BN_mod_inverse(BIGNUM *ret,
@@ -560,19 +587,22 @@ BN_MONT_CTX *BN_MONT_CTX_set_locked(BN_MONT_CTX **pmont, int lock,
 #define	BN_BLINDING_NO_UPDATE	0x00000001
 #define	BN_BLINDING_NO_RECREATE	0x00000002
 
-BN_BLINDING *BN_BLINDING_new(const BIGNUM *A, const BIGNUM *Ai, /* const */ BIGNUM *mod);
+BN_BLINDING *BN_BLINDING_new(const BIGNUM *A, const BIGNUM *Ai, BIGNUM *mod);
 void BN_BLINDING_free(BN_BLINDING *b);
 int BN_BLINDING_update(BN_BLINDING *b,BN_CTX *ctx);
 int BN_BLINDING_convert(BIGNUM *n, BN_BLINDING *b, BN_CTX *ctx);
 int BN_BLINDING_invert(BIGNUM *n, BN_BLINDING *b, BN_CTX *ctx);
 int BN_BLINDING_convert_ex(BIGNUM *n, BIGNUM *r, BN_BLINDING *b, BN_CTX *);
 int BN_BLINDING_invert_ex(BIGNUM *n, const BIGNUM *r, BN_BLINDING *b, BN_CTX *);
+#ifndef OPENSSL_NO_DEPRECATED
 unsigned long BN_BLINDING_get_thread_id(const BN_BLINDING *);
 void BN_BLINDING_set_thread_id(BN_BLINDING *, unsigned long);
+#endif
+CRYPTO_THREADID *BN_BLINDING_thread_id(BN_BLINDING *);
 unsigned long BN_BLINDING_get_flags(const BN_BLINDING *);
 void BN_BLINDING_set_flags(BN_BLINDING *, unsigned long);
 BN_BLINDING *BN_BLINDING_create_param(BN_BLINDING *b,
-	const BIGNUM *e, /* const */ BIGNUM *m, BN_CTX *ctx,
+	const BIGNUM *e, BIGNUM *m, BN_CTX *ctx,
 	int (*bn_mod_exp)(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
 			  const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *m_ctx),
 	BN_MONT_CTX *m_ctx);
@@ -593,6 +623,8 @@ int	BN_mod_exp_recp(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
 int	BN_div_recp(BIGNUM *dv, BIGNUM *rem, const BIGNUM *m,
 	BN_RECP_CTX *recp, BN_CTX *ctx);
 
+#ifndef OPENSSL_NO_EC2M
+
 /* Functions for arithmetic over binary polynomials represented by BIGNUMs. 
  *
  * The BIGNUM::neg property of BIGNUMs representing binary polynomials is
@@ -625,24 +657,26 @@ int	BN_GF2m_mod_solve_quad(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
  *     t^p[0] + t^p[1] + ... + t^p[k]
  * where m = p[0] > p[1] > ... > p[k] = 0.
  */
-int	BN_GF2m_mod_arr(BIGNUM *r, const BIGNUM *a, const unsigned int p[]);
+int	BN_GF2m_mod_arr(BIGNUM *r, const BIGNUM *a, const int p[]);
 	/* r = a mod p */
 int	BN_GF2m_mod_mul_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
-	const unsigned int p[], BN_CTX *ctx); /* r = (a * b) mod p */
-int	BN_GF2m_mod_sqr_arr(BIGNUM *r, const BIGNUM *a, const unsigned int p[],
+	const int p[], BN_CTX *ctx); /* r = (a * b) mod p */
+int	BN_GF2m_mod_sqr_arr(BIGNUM *r, const BIGNUM *a, const int p[],
 	BN_CTX *ctx); /* r = (a * a) mod p */
-int	BN_GF2m_mod_inv_arr(BIGNUM *r, const BIGNUM *b, const unsigned int p[],
+int	BN_GF2m_mod_inv_arr(BIGNUM *r, const BIGNUM *b, const int p[],
 	BN_CTX *ctx); /* r = (1 / b) mod p */
 int	BN_GF2m_mod_div_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
-	const unsigned int p[], BN_CTX *ctx); /* r = (a / b) mod p */
+	const int p[], BN_CTX *ctx); /* r = (a / b) mod p */
 int	BN_GF2m_mod_exp_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
-	const unsigned int p[], BN_CTX *ctx); /* r = (a ^ b) mod p */
+	const int p[], BN_CTX *ctx); /* r = (a ^ b) mod p */
 int	BN_GF2m_mod_sqrt_arr(BIGNUM *r, const BIGNUM *a,
-	const unsigned int p[], BN_CTX *ctx); /* r = sqrt(a) mod p */
+	const int p[], BN_CTX *ctx); /* r = sqrt(a) mod p */
 int	BN_GF2m_mod_solve_quad_arr(BIGNUM *r, const BIGNUM *a,
-	const unsigned int p[], BN_CTX *ctx); /* r^2 + r = a mod p */
-int	BN_GF2m_poly2arr(const BIGNUM *a, unsigned int p[], int max);
-int	BN_GF2m_arr2poly(const unsigned int p[], BIGNUM *a);
+	const int p[], BN_CTX *ctx); /* r^2 + r = a mod p */
+int	BN_GF2m_poly2arr(const BIGNUM *a, int p[], int max);
+int	BN_GF2m_arr2poly(const int p[], BIGNUM *a);
+
+#endif
 
 /* faster mod functions for the 'NIST primes' 
  * 0 <= a < p^2 */
@@ -751,10 +785,12 @@ int RAND_pseudo_bytes(unsigned char *buf,int num);
 #define bn_correct_top(a) \
         { \
         BN_ULONG *ftl; \
-	if ((a)->top > 0) \
+	int tmp_top = (a)->top; \
+	if (tmp_top > 0) \
 		{ \
-		for (ftl= &((a)->d[(a)->top-1]); (a)->top > 0; (a)->top--) \
-		if (*(ftl--)) break; \
+		for (ftl= &((a)->d[tmp_top-1]); tmp_top > 0; tmp_top--) \
+			if (*(ftl--)) break; \
+		(a)->top = tmp_top; \
 		} \
 	bn_pollute(a); \
 	}
diff --git a/crypto/bn/bn_asm.c b/crypto/bn/bn_asm.c
index 99bc2de4913e..c43c91cc09f4 100644
--- a/crypto/bn/bn_asm.c
+++ b/crypto/bn/bn_asm.c
@@ -75,6 +75,7 @@ BN_ULONG bn_mul_add_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w)
 	assert(num >= 0);
 	if (num <= 0) return(c1);
 
+#ifndef OPENSSL_SMALL_FOOTPRINT
 	while (num&~3)
 		{
 		mul_add(rp[0],ap[0],w,c1);
@@ -83,11 +84,11 @@ BN_ULONG bn_mul_add_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w)
 		mul_add(rp[3],ap[3],w,c1);
 		ap+=4; rp+=4; num-=4;
 		}
-	if (num)
+#endif
+	while (num)
 		{
-		mul_add(rp[0],ap[0],w,c1); if (--num==0) return c1;
-		mul_add(rp[1],ap[1],w,c1); if (--num==0) return c1;
-		mul_add(rp[2],ap[2],w,c1); return c1;
+		mul_add(rp[0],ap[0],w,c1);
+		ap++; rp++; num--;
 		}
 	
 	return(c1);
@@ -100,6 +101,7 @@ BN_ULONG bn_mul_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w)
 	assert(num >= 0);
 	if (num <= 0) return(c1);
 
+#ifndef OPENSSL_SMALL_FOOTPRINT
 	while (num&~3)
 		{
 		mul(rp[0],ap[0],w,c1);
@@ -108,11 +110,11 @@ BN_ULONG bn_mul_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w)
 		mul(rp[3],ap[3],w,c1);
 		ap+=4; rp+=4; num-=4;
 		}
-	if (num)
+#endif
+	while (num)
 		{
-		mul(rp[0],ap[0],w,c1); if (--num == 0) return c1;
-		mul(rp[1],ap[1],w,c1); if (--num == 0) return c1;
-		mul(rp[2],ap[2],w,c1);
+		mul(rp[0],ap[0],w,c1);
+		ap++; rp++; num--;
 		}
 	return(c1);
 	} 
@@ -121,6 +123,8 @@ void bn_sqr_words(BN_ULONG *r, const BN_ULONG *a, int n)
         {
 	assert(n >= 0);
 	if (n <= 0) return;
+
+#ifndef OPENSSL_SMALL_FOOTPRINT
 	while (n&~3)
 		{
 		sqr(r[0],r[1],a[0]);
@@ -129,11 +133,11 @@ void bn_sqr_words(BN_ULONG *r, const BN_ULONG *a, int n)
 		sqr(r[6],r[7],a[3]);
 		a+=4; r+=8; n-=4;
 		}
-	if (n)
+#endif
+	while (n)
 		{
-		sqr(r[0],r[1],a[0]); if (--n == 0) return;
-		sqr(r[2],r[3],a[1]); if (--n == 0) return;
-		sqr(r[4],r[5],a[2]);
+		sqr(r[0],r[1],a[0]);
+		a++; r+=2; n--;
 		}
 	}
 
@@ -150,18 +154,20 @@ BN_ULONG bn_mul_add_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w)
 	bl=LBITS(w);
 	bh=HBITS(w);
 
-	for (;;)
+#ifndef OPENSSL_SMALL_FOOTPRINT
+	while (num&~3)
 		{
 		mul_add(rp[0],ap[0],bl,bh,c);
-		if (--num == 0) break;
 		mul_add(rp[1],ap[1],bl,bh,c);
-		if (--num == 0) break;
 		mul_add(rp[2],ap[2],bl,bh,c);
-		if (--num == 0) break;
 		mul_add(rp[3],ap[3],bl,bh,c);
-		if (--num == 0) break;
-		ap+=4;
-		rp+=4;
+		ap+=4; rp+=4; num-=4;
+		}
+#endif
+	while (num)
+		{
+		mul_add(rp[0],ap[0],bl,bh,c);
+		ap++; rp++; num--;
 		}
 	return(c);
 	} 
@@ -177,18 +183,20 @@ BN_ULONG bn_mul_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w)
 	bl=LBITS(w);
 	bh=HBITS(w);
 
-	for (;;)
+#ifndef OPENSSL_SMALL_FOOTPRINT
+	while (num&~3)
 		{
 		mul(rp[0],ap[0],bl,bh,carry);
-		if (--num == 0) break;
 		mul(rp[1],ap[1],bl,bh,carry);
-		if (--num == 0) break;
 		mul(rp[2],ap[2],bl,bh,carry);
-		if (--num == 0) break;
 		mul(rp[3],ap[3],bl,bh,carry);
-		if (--num == 0) break;
-		ap+=4;
-		rp+=4;
+		ap+=4; rp+=4; num-=4;
+		}
+#endif
+	while (num)
+		{
+		mul(rp[0],ap[0],bl,bh,carry);
+		ap++; rp++; num--;
 		}
 	return(carry);
 	} 
@@ -197,22 +205,21 @@ void bn_sqr_words(BN_ULONG *r, const BN_ULONG *a, int n)
         {
 	assert(n >= 0);
 	if (n <= 0) return;
-	for (;;)
+
+#ifndef OPENSSL_SMALL_FOOTPRINT
+	while (n&~3)
 		{
 		sqr64(r[0],r[1],a[0]);
-		if (--n == 0) break;
-
 		sqr64(r[2],r[3],a[1]);
-		if (--n == 0) break;
-
 		sqr64(r[4],r[5],a[2]);
-		if (--n == 0) break;
-
 		sqr64(r[6],r[7],a[3]);
-		if (--n == 0) break;
-
-		a+=4;
-		r+=8;
+		a+=4; r+=8; n-=4;
+		}
+#endif
+	while (n)
+		{
+		sqr64(r[0],r[1],a[0]);
+		a++; r+=2; n--;
 		}
 	}
 
@@ -303,31 +310,30 @@ BN_ULONG bn_add_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b, int n)
 	assert(n >= 0);
 	if (n <= 0) return((BN_ULONG)0);
 
-	for (;;)
+#ifndef OPENSSL_SMALL_FOOTPRINT
+	while (n&~3)
 		{
 		ll+=(BN_ULLONG)a[0]+b[0];
 		r[0]=(BN_ULONG)ll&BN_MASK2;
 		ll>>=BN_BITS2;
-		if (--n <= 0) break;
-
 		ll+=(BN_ULLONG)a[1]+b[1];
 		r[1]=(BN_ULONG)ll&BN_MASK2;
 		ll>>=BN_BITS2;
-		if (--n <= 0) break;
-
 		ll+=(BN_ULLONG)a[2]+b[2];
 		r[2]=(BN_ULONG)ll&BN_MASK2;
 		ll>>=BN_BITS2;
-		if (--n <= 0) break;
-
 		ll+=(BN_ULLONG)a[3]+b[3];
 		r[3]=(BN_ULONG)ll&BN_MASK2;
 		ll>>=BN_BITS2;
-		if (--n <= 0) break;
-
-		a+=4;
-		b+=4;
-		r+=4;
+		a+=4; b+=4; r+=4; n-=4;
+		}
+#endif
+	while (n)
+		{
+		ll+=(BN_ULLONG)a[0]+b[0];
+		r[0]=(BN_ULONG)ll&BN_MASK2;
+		ll>>=BN_BITS2;
+		a++; b++; r++; n--;
 		}
 	return((BN_ULONG)ll);
 	}
@@ -340,7 +346,8 @@ BN_ULONG bn_add_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b, int n)
 	if (n <= 0) return((BN_ULONG)0);
 
 	c=0;
-	for (;;)
+#ifndef OPENSSL_SMALL_FOOTPRINT
+	while (n&~3)
 		{
 		t=a[0];
 		t=(t+c)&BN_MASK2;
@@ -348,35 +355,36 @@ BN_ULONG bn_add_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b, int n)
 		l=(t+b[0])&BN_MASK2;
 		c+=(l < t);
 		r[0]=l;
-		if (--n <= 0) break;
-
 		t=a[1];
 		t=(t+c)&BN_MASK2;
 		c=(t < c);
 		l=(t+b[1])&BN_MASK2;
 		c+=(l < t);
 		r[1]=l;
-		if (--n <= 0) break;
-
 		t=a[2];
 		t=(t+c)&BN_MASK2;
 		c=(t < c);
 		l=(t+b[2])&BN_MASK2;
 		c+=(l < t);
 		r[2]=l;
-		if (--n <= 0) break;
-
 		t=a[3];
 		t=(t+c)&BN_MASK2;
 		c=(t < c);
 		l=(t+b[3])&BN_MASK2;
 		c+=(l < t);
 		r[3]=l;
-		if (--n <= 0) break;
-
-		a+=4;
-		b+=4;
-		r+=4;
+		a+=4; b+=4; r+=4; n-=4;
+		}
+#endif
+	while(n)
+		{
+		t=a[0];
+		t=(t+c)&BN_MASK2;
+		c=(t < c);
+		l=(t+b[0])&BN_MASK2;
+		c+=(l < t);
+		r[0]=l;
+		a++; b++; r++; n--;
 		}
 	return((BN_ULONG)c);
 	}
@@ -390,36 +398,35 @@ BN_ULONG bn_sub_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b, int n)
 	assert(n >= 0);
 	if (n <= 0) return((BN_ULONG)0);
 
-	for (;;)
+#ifndef OPENSSL_SMALL_FOOTPRINT
+	while (n&~3)
 		{
 		t1=a[0]; t2=b[0];
 		r[0]=(t1-t2-c)&BN_MASK2;
 		if (t1 != t2) c=(t1 < t2);
-		if (--n <= 0) break;
-
 		t1=a[1]; t2=b[1];
 		r[1]=(t1-t2-c)&BN_MASK2;
 		if (t1 != t2) c=(t1 < t2);
-		if (--n <= 0) break;
-
 		t1=a[2]; t2=b[2];
 		r[2]=(t1-t2-c)&BN_MASK2;
 		if (t1 != t2) c=(t1 < t2);
-		if (--n <= 0) break;
-
 		t1=a[3]; t2=b[3];
 		r[3]=(t1-t2-c)&BN_MASK2;
 		if (t1 != t2) c=(t1 < t2);
-		if (--n <= 0) break;
-
-		a+=4;
-		b+=4;
-		r+=4;
+		a+=4; b+=4; r+=4; n-=4;
+		}
+#endif
+	while (n)
+		{
+		t1=a[0]; t2=b[0];
+		r[0]=(t1-t2-c)&BN_MASK2;
+		if (t1 != t2) c=(t1 < t2);
+		a++; b++; r++; n--;
 		}
 	return(c);
 	}
 
-#ifdef BN_MUL_COMBA
+#if defined(BN_MUL_COMBA) && !defined(OPENSSL_SMALL_FOOTPRINT)
 
 #undef bn_mul_comba8
 #undef bn_mul_comba4
@@ -820,18 +827,134 @@ void bn_sqr_comba4(BN_ULONG *r, const BN_ULONG *a)
 	r[6]=c1;
 	r[7]=c2;
 	}
+
+#ifdef OPENSSL_NO_ASM
+#ifdef OPENSSL_BN_ASM_MONT
+#include <alloca.h>
+/*
+ * This is essentially reference implementation, which may or may not
+ * result in performance improvement. E.g. on IA-32 this routine was
+ * observed to give 40% faster rsa1024 private key operations and 10%
+ * faster rsa4096 ones, while on AMD64 it improves rsa1024 sign only
+ * by 10% and *worsens* rsa4096 sign by 15%. Once again, it's a
+ * reference implementation, one to be used as starting point for
+ * platform-specific assembler. Mentioned numbers apply to compiler
+ * generated code compiled with and without -DOPENSSL_BN_ASM_MONT and
+ * can vary not only from platform to platform, but even for compiler
+ * versions. Assembler vs. assembler improvement coefficients can
+ * [and are known to] differ and are to be documented elsewhere.
+ */
+int bn_mul_mont(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np,const BN_ULONG *n0p, int num)
+	{
+	BN_ULONG c0,c1,ml,*tp,n0;
+#ifdef mul64
+	BN_ULONG mh;
+#endif
+	volatile BN_ULONG *vp;
+	int i=0,j;
+
+#if 0	/* template for platform-specific implementation */
+	if (ap==bp)	return bn_sqr_mont(rp,ap,np,n0p,num);
+#endif
+	vp = tp = alloca((num+2)*sizeof(BN_ULONG));
+
+	n0 = *n0p;
+
+	c0 = 0;
+	ml = bp[0];
+#ifdef mul64
+	mh = HBITS(ml);
+	ml = LBITS(ml);
+	for (j=0;j<num;++j)
+		mul(tp[j],ap[j],ml,mh,c0);
+#else
+	for (j=0;j<num;++j)
+		mul(tp[j],ap[j],ml,c0);
+#endif
+
+	tp[num]   = c0;
+	tp[num+1] = 0;
+	goto enter;
+
+	for(i=0;i<num;i++)
+		{
+		c0 = 0;
+		ml = bp[i];
+#ifdef mul64
+		mh = HBITS(ml);
+		ml = LBITS(ml);
+		for (j=0;j<num;++j)
+			mul_add(tp[j],ap[j],ml,mh,c0);
+#else
+		for (j=0;j<num;++j)
+			mul_add(tp[j],ap[j],ml,c0);
+#endif
+		c1 = (tp[num] + c0)&BN_MASK2;
+		tp[num]   = c1;
+		tp[num+1] = (c1<c0?1:0);
+	enter:
+		c1  = tp[0];
+		ml = (c1*n0)&BN_MASK2;
+		c0 = 0;
+#ifdef mul64
+		mh = HBITS(ml);
+		ml = LBITS(ml);
+		mul_add(c1,np[0],ml,mh,c0);
+#else
+		mul_add(c1,ml,np[0],c0);
+#endif
+		for(j=1;j<num;j++)
+			{
+			c1 = tp[j];
+#ifdef mul64
+			mul_add(c1,np[j],ml,mh,c0);
+#else
+			mul_add(c1,ml,np[j],c0);
+#endif
+			tp[j-1] = c1&BN_MASK2;
+			}
+		c1        = (tp[num] + c0)&BN_MASK2;
+		tp[num-1] = c1;
+		tp[num]   = tp[num+1] + (c1<c0?1:0);
+		}
+
+	if (tp[num]!=0 || tp[num-1]>=np[num-1])
+		{
+		c0 = bn_sub_words(rp,tp,np,num);
+		if (tp[num]!=0 || c0==0)
+			{
+			for(i=0;i<num+2;i++)	vp[i] = 0;
+			return 1;
+			}
+		}
+	for(i=0;i<num;i++)	rp[i] = tp[i],	vp[i] = 0;
+	vp[num]   = 0;
+	vp[num+1] = 0;
+	return 1;
+	}
+#else
+/*
+ * Return value of 0 indicates that multiplication/convolution was not
+ * performed to signal the caller to fall down to alternative/original
+ * code-path.
+ */
+int bn_mul_mont(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np,const BN_ULONG *n0, int num)
+{	return 0;	}
+#endif /* OPENSSL_BN_ASM_MONT */
+#endif
+
 #else /* !BN_MUL_COMBA */
 
 /* hmm... is it faster just to do a multiply? */
 #undef bn_sqr_comba4
-void bn_sqr_comba4(BN_ULONG *r, BN_ULONG *a)
+void bn_sqr_comba4(BN_ULONG *r, const BN_ULONG *a)
 	{
 	BN_ULONG t[8];
 	bn_sqr_normal(r,a,4,t);
 	}
 
 #undef bn_sqr_comba8
-void bn_sqr_comba8(BN_ULONG *r, BN_ULONG *a)
+void bn_sqr_comba8(BN_ULONG *r, const BN_ULONG *a)
 	{
 	BN_ULONG t[16];
 	bn_sqr_normal(r,a,8,t);
@@ -857,4 +980,51 @@ void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
 	r[15]=bn_mul_add_words(&(r[7]),a,8,b[7]);
 	}
 
+#ifdef OPENSSL_NO_ASM
+#ifdef OPENSSL_BN_ASM_MONT
+#include <alloca.h>
+int bn_mul_mont(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np,const BN_ULONG *n0p, int num)
+	{
+	BN_ULONG c0,c1,*tp,n0=*n0p;
+	volatile BN_ULONG *vp;
+	int i=0,j;
+
+	vp = tp = alloca((num+2)*sizeof(BN_ULONG));
+
+	for(i=0;i<=num;i++)	tp[i]=0;
+
+	for(i=0;i<num;i++)
+		{
+		c0         = bn_mul_add_words(tp,ap,num,bp[i]);
+		c1         = (tp[num] + c0)&BN_MASK2;
+		tp[num]    = c1;
+		tp[num+1]  = (c1<c0?1:0);
+
+		c0         = bn_mul_add_words(tp,np,num,tp[0]*n0);
+		c1         = (tp[num] + c0)&BN_MASK2;
+		tp[num]    = c1;
+		tp[num+1] += (c1<c0?1:0);
+		for(j=0;j<=num;j++)	tp[j]=tp[j+1];
+		}
+
+	if (tp[num]!=0 || tp[num-1]>=np[num-1])
+		{
+		c0 = bn_sub_words(rp,tp,np,num);
+		if (tp[num]!=0 || c0==0)
+			{
+			for(i=0;i<num+2;i++)	vp[i] = 0;
+			return 1;
+			}
+		}
+	for(i=0;i<num;i++)	rp[i] = tp[i],	vp[i] = 0;
+	vp[num]   = 0;
+	vp[num+1] = 0;
+	return 1;
+	}
+#else
+int bn_mul_mont(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np,const BN_ULONG *n0, int num)
+{	return 0;	}
+#endif /* OPENSSL_BN_ASM_MONT */
+#endif
+
 #endif /* !BN_MUL_COMBA */
diff --git a/crypto/bn/bn_blind.c b/crypto/bn/bn_blind.c
index ca7f996bc8a1..9ed8bc2b40b8 100644
--- a/crypto/bn/bn_blind.c
+++ b/crypto/bn/bn_blind.c
@@ -1,6 +1,6 @@
 /* crypto/bn/bn_blind.c */
 /* ====================================================================
- * Copyright (c) 1998-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1998-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -121,8 +121,11 @@ struct bn_blinding_st
 	BIGNUM *Ai;
 	BIGNUM *e;
 	BIGNUM *mod; /* just a reference */
+#ifndef OPENSSL_NO_DEPRECATED
 	unsigned long thread_id; /* added in OpenSSL 0.9.6j and 0.9.7b;
 				  * used only by crypto/rsa/rsa_eay.c, rsa_lib.c */
+#endif
+	CRYPTO_THREADID tid;
 	int counter;
 	unsigned long flags;
 	BN_MONT_CTX *m_ctx;
@@ -131,7 +134,7 @@ struct bn_blinding_st
 			  BN_MONT_CTX *m_ctx);
 	};
 
-BN_BLINDING *BN_BLINDING_new(const BIGNUM *A, const BIGNUM *Ai, /* const */ BIGNUM *mod)
+BN_BLINDING *BN_BLINDING_new(const BIGNUM *A, const BIGNUM *Ai, BIGNUM *mod)
 	{
 	BN_BLINDING *ret=NULL;
 
@@ -161,6 +164,7 @@ BN_BLINDING *BN_BLINDING_new(const BIGNUM *A, const BIGNUM *Ai, /* const */ BIGN
 	 * to indicate that this is never-used fresh blinding
 	 * that does not need updating before first use. */
 	ret->counter = -1;
+	CRYPTO_THREADID_current(&ret->tid);
 	return(ret);
 err:
 	if (ret != NULL) BN_BLINDING_free(ret);
@@ -272,6 +276,7 @@ int BN_BLINDING_invert_ex(BIGNUM *n, const BIGNUM *r, BN_BLINDING *b, BN_CTX *ct
 	return(ret);
 	}
 
+#ifndef OPENSSL_NO_DEPRECATED
 unsigned long BN_BLINDING_get_thread_id(const BN_BLINDING *b)
 	{
 	return b->thread_id;
@@ -281,6 +286,12 @@ void BN_BLINDING_set_thread_id(BN_BLINDING *b, unsigned long n)
 	{
 	b->thread_id = n;
 	}
+#endif
+
+CRYPTO_THREADID *BN_BLINDING_thread_id(BN_BLINDING *b)
+	{
+	return &b->tid;
+	}
 
 unsigned long BN_BLINDING_get_flags(const BN_BLINDING *b)
 	{
@@ -293,7 +304,7 @@ void BN_BLINDING_set_flags(BN_BLINDING *b, unsigned long flags)
 	}
 
 BN_BLINDING *BN_BLINDING_create_param(BN_BLINDING *b,
-	const BIGNUM *e, /* const */ BIGNUM *m, BN_CTX *ctx,
+	const BIGNUM *e, BIGNUM *m, BN_CTX *ctx,
 	int (*bn_mod_exp)(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
 			  const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *m_ctx),
 	BN_MONT_CTX *m_ctx)
diff --git a/crypto/bn/bn_ctx.c b/crypto/bn/bn_ctx.c
index b3452f1a91e3..3f2256f67575 100644
--- a/crypto/bn/bn_ctx.c
+++ b/crypto/bn/bn_ctx.c
@@ -161,7 +161,7 @@ static void ctxdbg(BN_CTX *ctx)
 	fprintf(stderr,"(%08x): ", (unsigned int)ctx);
 	while(bnidx < ctx->used)
 		{
-		fprintf(stderr,"%02x ", item->vals[bnidx++ % BN_CTX_POOL_SIZE].dmax);
+		fprintf(stderr,"%03x ", item->vals[bnidx++ % BN_CTX_POOL_SIZE].dmax);
 		if(!(bnidx % BN_CTX_POOL_SIZE))
 			item = item->next;
 		}
@@ -171,8 +171,8 @@ static void ctxdbg(BN_CTX *ctx)
 	while(fpidx < stack->depth)
 		{
 		while(bnidx++ < stack->indexes[fpidx])
-			fprintf(stderr,"   ");
-		fprintf(stderr,"^^ ");
+			fprintf(stderr,"    ");
+		fprintf(stderr,"^^^ ");
 		bnidx++;
 		fpidx++;
 		}
diff --git a/crypto/bn/bn_div.c b/crypto/bn/bn_div.c
index 78c6507113bf..52b3304293a5 100644
--- a/crypto/bn/bn_div.c
+++ b/crypto/bn/bn_div.c
@@ -169,15 +169,13 @@ int BN_div(BIGNUM *dv, BIGNUM *rem, const BIGNUM *m, const BIGNUM *d,
 #endif /* OPENSSL_NO_ASM */
 
 
-/* BN_div[_no_branch] computes  dv := num / divisor,  rounding towards
+/* BN_div computes  dv := num / divisor,  rounding towards
  * zero, and sets up rm  such that  dv*divisor + rm = num  holds.
  * Thus:
  *     dv->neg == num->neg ^ divisor->neg  (unless the result is zero)
  *     rm->neg == num->neg                 (unless the remainder is zero)
  * If 'dv' or 'rm' is NULL, the respective value is not returned.
  */
-static int BN_div_no_branch(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num,
-        const BIGNUM *divisor, BN_CTX *ctx);
 int BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor,
 	   BN_CTX *ctx)
 	{
@@ -186,6 +184,7 @@ int BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor,
 	BN_ULONG *resp,*wnump;
 	BN_ULONG d0,d1;
 	int num_n,div_n;
+	int no_branch=0;
 
 	/* Invalid zero-padding would have particularly bad consequences
 	 * in the case of 'num', so don't just rely on bn_check_top() for this one
@@ -200,7 +199,7 @@ int BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor,
 
 	if ((BN_get_flags(num, BN_FLG_CONSTTIME) != 0) || (BN_get_flags(divisor, BN_FLG_CONSTTIME) != 0))
 		{
-		return BN_div_no_branch(dv, rm, num, divisor, ctx);
+		no_branch=1;
 		}
 
 	bn_check_top(dv);
@@ -214,7 +213,7 @@ int BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor,
 		return(0);
 		}
 
-	if (BN_ucmp(num,divisor) < 0)
+	if (!no_branch && BN_ucmp(num,divisor) < 0)
 		{
 		if (rm != NULL)
 			{ if (BN_copy(rm,num) == NULL) return(0); }
@@ -239,242 +238,25 @@ int BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor,
 	norm_shift+=BN_BITS2;
 	if (!(BN_lshift(snum,num,norm_shift))) goto err;
 	snum->neg=0;
-	div_n=sdiv->top;
-	num_n=snum->top;
-	loop=num_n-div_n;
-	/* Lets setup a 'window' into snum
-	 * This is the part that corresponds to the current
-	 * 'area' being divided */
-	wnum.neg   = 0;
-	wnum.d     = &(snum->d[loop]);
-	wnum.top   = div_n;
-	/* only needed when BN_ucmp messes up the values between top and max */
-	wnum.dmax  = snum->dmax - loop; /* so we don't step out of bounds */
-
-	/* Get the top 2 words of sdiv */
-	/* div_n=sdiv->top; */
-	d0=sdiv->d[div_n-1];
-	d1=(div_n == 1)?0:sdiv->d[div_n-2];
-
-	/* pointer to the 'top' of snum */
-	wnump= &(snum->d[num_n-1]);
-
-	/* Setup to 'res' */
-	res->neg= (num->neg^divisor->neg);
-	if (!bn_wexpand(res,(loop+1))) goto err;
-	res->top=loop;
-	resp= &(res->d[loop-1]);
-
-	/* space for temp */
-	if (!bn_wexpand(tmp,(div_n+1))) goto err;
 
-	if (BN_ucmp(&wnum,sdiv) >= 0)
+	if (no_branch)
 		{
-		/* If BN_DEBUG_RAND is defined BN_ucmp changes (via
-		 * bn_pollute) the const bignum arguments =>
-		 * clean the values between top and max again */
-		bn_clear_top2max(&wnum);
-		bn_sub_words(wnum.d, wnum.d, sdiv->d, div_n);
-		*resp=1;
-		}
-	else
-		res->top--;
-	/* if res->top == 0 then clear the neg value otherwise decrease
-	 * the resp pointer */
-	if (res->top == 0)
-		res->neg = 0;
-	else
-		resp--;
-
-	for (i=0; i<loop-1; i++, wnump--, resp--)
-		{
-		BN_ULONG q,l0;
-		/* the first part of the loop uses the top two words of
-		 * snum and sdiv to calculate a BN_ULONG q such that
-		 * | wnum - sdiv * q | < sdiv */
-#if defined(BN_DIV3W) && !defined(OPENSSL_NO_ASM)
-		BN_ULONG bn_div_3_words(BN_ULONG*,BN_ULONG,BN_ULONG);
-		q=bn_div_3_words(wnump,d1,d0);
-#else
-		BN_ULONG n0,n1,rem=0;
-
-		n0=wnump[0];
-		n1=wnump[-1];
-		if (n0 == d0)
-			q=BN_MASK2;
-		else 			/* n0 < d0 */
+		/* Since we don't know whether snum is larger than sdiv,
+		 * we pad snum with enough zeroes without changing its
+		 * value. 
+		 */
+		if (snum->top <= sdiv->top+1) 
 			{
-#ifdef BN_LLONG
-			BN_ULLONG t2;
-
-#if defined(BN_LLONG) && defined(BN_DIV2W) && !defined(bn_div_words)
-			q=(BN_ULONG)(((((BN_ULLONG)n0)<<BN_BITS2)|n1)/d0);
-#else
-			q=bn_div_words(n0,n1,d0);
-#ifdef BN_DEBUG_LEVITTE
-			fprintf(stderr,"DEBUG: bn_div_words(0x%08X,0x%08X,0x%08\
-X) -> 0x%08X\n",
-				n0, n1, d0, q);
-#endif
-#endif
-
-#ifndef REMAINDER_IS_ALREADY_CALCULATED
-			/*
-			 * rem doesn't have to be BN_ULLONG. The least we
-			 * know it's less that d0, isn't it?
-			 */
-			rem=(n1-q*d0)&BN_MASK2;
-#endif
-			t2=(BN_ULLONG)d1*q;
-
-			for (;;)
-				{
-				if (t2 <= ((((BN_ULLONG)rem)<<BN_BITS2)|wnump[-2]))
-					break;
-				q--;
-				rem += d0;
-				if (rem < d0) break; /* don't let rem overflow */
-				t2 -= d1;
-				}
-#else /* !BN_LLONG */
-			BN_ULONG t2l,t2h;
-#if !defined(BN_UMULT_LOHI) && !defined(BN_UMULT_HIGH)
-			BN_ULONG ql,qh;
-#endif
-
-			q=bn_div_words(n0,n1,d0);
-#ifdef BN_DEBUG_LEVITTE
-			fprintf(stderr,"DEBUG: bn_div_words(0x%08X,0x%08X,0x%08\
-X) -> 0x%08X\n",
-				n0, n1, d0, q);
-#endif
-#ifndef REMAINDER_IS_ALREADY_CALCULATED
-			rem=(n1-q*d0)&BN_MASK2;
-#endif
-
-#if defined(BN_UMULT_LOHI)
-			BN_UMULT_LOHI(t2l,t2h,d1,q);
-#elif defined(BN_UMULT_HIGH)
-			t2l = d1 * q;
-			t2h = BN_UMULT_HIGH(d1,q);
-#else
-			t2l=LBITS(d1); t2h=HBITS(d1);
-			ql =LBITS(q);  qh =HBITS(q);
-			mul64(t2l,t2h,ql,qh); /* t2=(BN_ULLONG)d1*q; */
-#endif
-
-			for (;;)
-				{
-				if ((t2h < rem) ||
-					((t2h == rem) && (t2l <= wnump[-2])))
-					break;
-				q--;
-				rem += d0;
-				if (rem < d0) break; /* don't let rem overflow */
-				if (t2l < d1) t2h--; t2l -= d1;
-				}
-#endif /* !BN_LLONG */
+			if (bn_wexpand(snum, sdiv->top + 2) == NULL) goto err;
+			for (i = snum->top; i < sdiv->top + 2; i++) snum->d[i] = 0;
+			snum->top = sdiv->top + 2;
 			}
-#endif /* !BN_DIV3W */
-
-		l0=bn_mul_words(tmp->d,sdiv->d,div_n,q);
-		tmp->d[div_n]=l0;
-		wnum.d--;
-		/* ingore top values of the bignums just sub the two 
-		 * BN_ULONG arrays with bn_sub_words */
-		if (bn_sub_words(wnum.d, wnum.d, tmp->d, div_n+1))
+		else
 			{
-			/* Note: As we have considered only the leading
-			 * two BN_ULONGs in the calculation of q, sdiv * q
-			 * might be greater than wnum (but then (q-1) * sdiv
-			 * is less or equal than wnum)
-			 */
-			q--;
-			if (bn_add_words(wnum.d, wnum.d, sdiv->d, div_n))
-				/* we can't have an overflow here (assuming
-				 * that q != 0, but if q == 0 then tmp is
-				 * zero anyway) */
-				(*wnump)++;
+			if (bn_wexpand(snum, snum->top + 1) == NULL) goto err;
+			snum->d[snum->top] = 0;
+			snum->top ++;
 			}
-		/* store part of the result */
-		*resp = q;
-		}
-	bn_correct_top(snum);
-	if (rm != NULL)
-		{
-		/* Keep a copy of the neg flag in num because if rm==num
-		 * BN_rshift() will overwrite it.
-		 */
-		int neg = num->neg;
-		BN_rshift(rm,snum,norm_shift);
-		if (!BN_is_zero(rm))
-			rm->neg = neg;
-		bn_check_top(rm);
-		}
-	BN_CTX_end(ctx);
-	return(1);
-err:
-	bn_check_top(rm);
-	BN_CTX_end(ctx);
-	return(0);
-	}
-
-
-/* BN_div_no_branch is a special version of BN_div. It does not contain
- * branches that may leak sensitive information.
- */
-static int BN_div_no_branch(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, 
-	const BIGNUM *divisor, BN_CTX *ctx)
-	{
-	int norm_shift,i,loop;
-	BIGNUM *tmp,wnum,*snum,*sdiv,*res;
-	BN_ULONG *resp,*wnump;
-	BN_ULONG d0,d1;
-	int num_n,div_n;
-
-	bn_check_top(dv);
-	bn_check_top(rm);
-	/* bn_check_top(num); */ /* 'num' has been checked in BN_div() */
-	bn_check_top(divisor);
-
-	if (BN_is_zero(divisor))
-		{
-		BNerr(BN_F_BN_DIV_NO_BRANCH,BN_R_DIV_BY_ZERO);
-		return(0);
-		}
-
-	BN_CTX_start(ctx);
-	tmp=BN_CTX_get(ctx);
-	snum=BN_CTX_get(ctx);
-	sdiv=BN_CTX_get(ctx);
-	if (dv == NULL)
-		res=BN_CTX_get(ctx);
-	else	res=dv;
-	if (sdiv == NULL || res == NULL) goto err;
-
-	/* First we normalise the numbers */
-	norm_shift=BN_BITS2-((BN_num_bits(divisor))%BN_BITS2);
-	if (!(BN_lshift(sdiv,divisor,norm_shift))) goto err;
-	sdiv->neg=0;
-	norm_shift+=BN_BITS2;
-	if (!(BN_lshift(snum,num,norm_shift))) goto err;
-	snum->neg=0;
-
-	/* Since we don't know whether snum is larger than sdiv,
-	 * we pad snum with enough zeroes without changing its
-	 * value. 
-	 */
-	if (snum->top <= sdiv->top+1) 
-		{
-		if (bn_wexpand(snum, sdiv->top + 2) == NULL) goto err;
-		for (i = snum->top; i < sdiv->top + 2; i++) snum->d[i] = 0;
-		snum->top = sdiv->top + 2;
-		}
-	else
-		{
-		if (bn_wexpand(snum, snum->top + 1) == NULL) goto err;
-		snum->d[snum->top] = 0;
-		snum->top ++;
 		}
 
 	div_n=sdiv->top;
@@ -500,12 +282,27 @@ static int BN_div_no_branch(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num,
 	/* Setup to 'res' */
 	res->neg= (num->neg^divisor->neg);
 	if (!bn_wexpand(res,(loop+1))) goto err;
-	res->top=loop-1;
+	res->top=loop-no_branch;
 	resp= &(res->d[loop-1]);
 
 	/* space for temp */
 	if (!bn_wexpand(tmp,(div_n+1))) goto err;
 
+	if (!no_branch)
+		{
+		if (BN_ucmp(&wnum,sdiv) >= 0)
+			{
+			/* If BN_DEBUG_RAND is defined BN_ucmp changes (via
+			 * bn_pollute) the const bignum arguments =>
+			 * clean the values between top and max again */
+			bn_clear_top2max(&wnum);
+			bn_sub_words(wnum.d, wnum.d, sdiv->d, div_n);
+			*resp=1;
+			}
+		else
+			res->top--;
+		}
+
 	/* if res->top == 0 then clear the neg value otherwise decrease
 	 * the resp pointer */
 	if (res->top == 0)
@@ -565,9 +362,6 @@ X) -> 0x%08X\n",
 				}
 #else /* !BN_LLONG */
 			BN_ULONG t2l,t2h;
-#if !defined(BN_UMULT_LOHI) && !defined(BN_UMULT_HIGH)
-			BN_ULONG ql,qh;
-#endif
 
 			q=bn_div_words(n0,n1,d0);
 #ifdef BN_DEBUG_LEVITTE
@@ -585,9 +379,12 @@ X) -> 0x%08X\n",
 			t2l = d1 * q;
 			t2h = BN_UMULT_HIGH(d1,q);
 #else
+			{
+			BN_ULONG ql, qh;
 			t2l=LBITS(d1); t2h=HBITS(d1);
 			ql =LBITS(q);  qh =HBITS(q);
 			mul64(t2l,t2h,ql,qh); /* t2=(BN_ULLONG)d1*q; */
+			}
 #endif
 
 			for (;;)
@@ -638,7 +435,7 @@ X) -> 0x%08X\n",
 			rm->neg = neg;
 		bn_check_top(rm);
 		}
-	bn_correct_top(res);
+	if (no_branch)	bn_correct_top(res);
 	BN_CTX_end(ctx);
 	return(1);
 err:
@@ -646,5 +443,4 @@ err:
 	BN_CTX_end(ctx);
 	return(0);
 	}
-
 #endif
diff --git a/crypto/bn/bn_exp.c b/crypto/bn/bn_exp.c
index d9b6c737fc82..2abf6fd67871 100644
--- a/crypto/bn/bn_exp.c
+++ b/crypto/bn/bn_exp.c
@@ -113,6 +113,18 @@
 #include "cryptlib.h"
 #include "bn_lcl.h"
 
+#include <stdlib.h>
+#ifdef _WIN32
+# include <malloc.h>
+# ifndef alloca
+#  define alloca _alloca
+# endif
+#elif defined(__GNUC__)
+# ifndef alloca
+#  define alloca(s) __builtin_alloca((s))
+# endif
+#endif
+
 /* maximum precomputation table size for *variable* sliding windows */
 #define TABLE_SIZE	32
 
@@ -522,23 +534,17 @@ err:
  * as cache lines are concerned.  The following functions are used to transfer a BIGNUM
  * from/to that table. */
 
-static int MOD_EXP_CTIME_COPY_TO_PREBUF(BIGNUM *b, int top, unsigned char *buf, int idx, int width)
+static int MOD_EXP_CTIME_COPY_TO_PREBUF(const BIGNUM *b, int top, unsigned char *buf, int idx, int width)
 	{
 	size_t i, j;
 
-	if (bn_wexpand(b, top) == NULL)
-		return 0;
-	while (b->top < top)
-		{
-		b->d[b->top++] = 0;
-		}
-	
+	if (top > b->top)
+		top = b->top; /* this works because 'buf' is explicitly zeroed */
 	for (i = 0, j=idx; i < top * sizeof b->d[0]; i++, j+=width)
 		{
 		buf[j] = ((unsigned char*)b->d)[i];
 		}
 
-	bn_correct_top(b);
 	return 1;
 	}
 
@@ -561,7 +567,7 @@ static int MOD_EXP_CTIME_COPY_FROM_PREBUF(BIGNUM *b, int top, unsigned char *buf
 
 /* Given a pointer value, compute the next address that is a cache line multiple. */
 #define MOD_EXP_CTIME_ALIGN(x_) \
-	((unsigned char*)(x_) + (MOD_EXP_CTIME_MIN_CACHE_LINE_WIDTH - (((BN_ULONG)(x_)) & (MOD_EXP_CTIME_MIN_CACHE_LINE_MASK))))
+	((unsigned char*)(x_) + (MOD_EXP_CTIME_MIN_CACHE_LINE_WIDTH - (((size_t)(x_)) & (MOD_EXP_CTIME_MIN_CACHE_LINE_MASK))))
 
 /* This variant of BN_mod_exp_mont() uses fixed windows and the special
  * precomputation memory layout to limit data-dependency to a minimum
@@ -572,17 +578,15 @@ static int MOD_EXP_CTIME_COPY_FROM_PREBUF(BIGNUM *b, int top, unsigned char *buf
 int BN_mod_exp_mont_consttime(BIGNUM *rr, const BIGNUM *a, const BIGNUM *p,
 		    const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *in_mont)
 	{
-	int i,bits,ret=0,idx,window,wvalue;
+	int i,bits,ret=0,window,wvalue;
 	int top;
- 	BIGNUM *r;
-	const BIGNUM *aa;
 	BN_MONT_CTX *mont=NULL;
 
 	int numPowers;
 	unsigned char *powerbufFree=NULL;
 	int powerbufLen = 0;
 	unsigned char *powerbuf=NULL;
-	BIGNUM *computeTemp=NULL, *am=NULL;
+	BIGNUM tmp, am;
 
 	bn_check_top(a);
 	bn_check_top(p);
@@ -602,10 +606,7 @@ int BN_mod_exp_mont_consttime(BIGNUM *rr, const BIGNUM *a, const BIGNUM *p,
 		return ret;
 		}
 
- 	/* Initialize BIGNUM context and allocate intermediate result */
 	BN_CTX_start(ctx);
-	r = BN_CTX_get(ctx);
-	if (r == NULL) goto err;
 
 	/* Allocate a montgomery context if it was not supplied by the caller.
 	 * If this is not done, things will break in the montgomery part.
@@ -620,40 +621,154 @@ int BN_mod_exp_mont_consttime(BIGNUM *rr, const BIGNUM *a, const BIGNUM *p,
 
 	/* Get the window size to use with size of p. */
 	window = BN_window_bits_for_ctime_exponent_size(bits);
+#if defined(OPENSSL_BN_ASM_MONT5)
+	if (window==6 && bits<=1024) window=5;	/* ~5% improvement of 2048-bit RSA sign */
+#endif
 
 	/* Allocate a buffer large enough to hold all of the pre-computed
-	 * powers of a.
+	 * powers of am, am itself and tmp.
 	 */
 	numPowers = 1 << window;
-	powerbufLen = sizeof(m->d[0])*top*numPowers;
+	powerbufLen = sizeof(m->d[0])*(top*numPowers +
+				((2*top)>numPowers?(2*top):numPowers));
+#ifdef alloca
+	if (powerbufLen < 3072)
+		powerbufFree = alloca(powerbufLen+MOD_EXP_CTIME_MIN_CACHE_LINE_WIDTH);
+	else
+#endif
 	if ((powerbufFree=(unsigned char*)OPENSSL_malloc(powerbufLen+MOD_EXP_CTIME_MIN_CACHE_LINE_WIDTH)) == NULL)
 		goto err;
 		
 	powerbuf = MOD_EXP_CTIME_ALIGN(powerbufFree);
 	memset(powerbuf, 0, powerbufLen);
 
- 	/* Initialize the intermediate result. Do this early to save double conversion,
-	 * once each for a^0 and intermediate result.
-	 */
- 	if (!BN_to_montgomery(r,BN_value_one(),mont,ctx)) goto err;
-	if (!MOD_EXP_CTIME_COPY_TO_PREBUF(r, top, powerbuf, 0, numPowers)) goto err;
+#ifdef alloca
+	if (powerbufLen < 3072)
+		powerbufFree = NULL;
+#endif
 
-	/* Initialize computeTemp as a^1 with montgomery precalcs */
-	computeTemp = BN_CTX_get(ctx);
-	am = BN_CTX_get(ctx);
-	if (computeTemp==NULL || am==NULL) goto err;
+	/* lay down tmp and am right after powers table */
+	tmp.d     = (BN_ULONG *)(powerbuf + sizeof(m->d[0])*top*numPowers);
+	am.d      = tmp.d + top;
+	tmp.top   = am.top  = 0;
+	tmp.dmax  = am.dmax = top;
+	tmp.neg   = am.neg  = 0;
+	tmp.flags = am.flags = BN_FLG_STATIC_DATA;
+
+	/* prepare a^0 in Montgomery domain */
+#if 1
+ 	if (!BN_to_montgomery(&tmp,BN_value_one(),mont,ctx))	goto err;
+#else
+	tmp.d[0] = (0-m->d[0])&BN_MASK2;	/* 2^(top*BN_BITS2) - m */
+	for (i=1;i<top;i++)
+		tmp.d[i] = (~m->d[i])&BN_MASK2;
+	tmp.top = top;
+#endif
 
+	/* prepare a^1 in Montgomery domain */
 	if (a->neg || BN_ucmp(a,m) >= 0)
 		{
-		if (!BN_mod(am,a,m,ctx))
-			goto err;
-		aa= am;
+		if (!BN_mod(&am,a,m,ctx))			goto err;
+		if (!BN_to_montgomery(&am,&am,mont,ctx))	goto err;
 		}
-	else
-		aa=a;
-	if (!BN_to_montgomery(am,aa,mont,ctx)) goto err;
-	if (!BN_copy(computeTemp, am)) goto err;
-	if (!MOD_EXP_CTIME_COPY_TO_PREBUF(am, top, powerbuf, 1, numPowers)) goto err;
+	else	if (!BN_to_montgomery(&am,a,mont,ctx))		goto err;
+
+#if defined(OPENSSL_BN_ASM_MONT5)
+    /* This optimization uses ideas from http://eprint.iacr.org/2011/239,
+     * specifically optimization of cache-timing attack countermeasures
+     * and pre-computation optimization. */
+
+    /* Dedicated window==4 case improves 512-bit RSA sign by ~15%, but as
+     * 512-bit RSA is hardly relevant, we omit it to spare size... */ 
+    if (window==5)
+	{
+	void bn_mul_mont_gather5(BN_ULONG *rp,const BN_ULONG *ap,
+			const void *table,const BN_ULONG *np,
+			const BN_ULONG *n0,int num,int power);
+	void bn_scatter5(const BN_ULONG *inp,size_t num,
+			void *table,size_t power);
+	void bn_gather5(BN_ULONG *out,size_t num,
+			void *table,size_t power);
+
+	BN_ULONG *np=mont->N.d, *n0=mont->n0;
+
+	/* BN_to_montgomery can contaminate words above .top
+	 * [in BN_DEBUG[_DEBUG] build]... */
+	for (i=am.top; i<top; i++)	am.d[i]=0;
+	for (i=tmp.top; i<top; i++)	tmp.d[i]=0;
+
+	bn_scatter5(tmp.d,top,powerbuf,0);
+	bn_scatter5(am.d,am.top,powerbuf,1);
+	bn_mul_mont(tmp.d,am.d,am.d,np,n0,top);
+	bn_scatter5(tmp.d,top,powerbuf,2);
+
+#if 0
+	for (i=3; i<32; i++)
+		{
+		/* Calculate a^i = a^(i-1) * a */
+		bn_mul_mont_gather5(tmp.d,am.d,powerbuf,np,n0,top,i-1);
+		bn_scatter5(tmp.d,top,powerbuf,i);
+		}
+#else
+	/* same as above, but uses squaring for 1/2 of operations */
+	for (i=4; i<32; i*=2)
+		{
+		bn_mul_mont(tmp.d,tmp.d,tmp.d,np,n0,top);
+		bn_scatter5(tmp.d,top,powerbuf,i);
+		}
+	for (i=3; i<8; i+=2)
+		{
+		int j;
+		bn_mul_mont_gather5(tmp.d,am.d,powerbuf,np,n0,top,i-1);
+		bn_scatter5(tmp.d,top,powerbuf,i);
+		for (j=2*i; j<32; j*=2)
+			{
+			bn_mul_mont(tmp.d,tmp.d,tmp.d,np,n0,top);
+			bn_scatter5(tmp.d,top,powerbuf,j);
+			}
+		}
+	for (; i<16; i+=2)
+		{
+		bn_mul_mont_gather5(tmp.d,am.d,powerbuf,np,n0,top,i-1);
+		bn_scatter5(tmp.d,top,powerbuf,i);
+		bn_mul_mont(tmp.d,tmp.d,tmp.d,np,n0,top);
+		bn_scatter5(tmp.d,top,powerbuf,2*i);
+		}
+	for (; i<32; i+=2)
+		{
+		bn_mul_mont_gather5(tmp.d,am.d,powerbuf,np,n0,top,i-1);
+		bn_scatter5(tmp.d,top,powerbuf,i);
+		}
+#endif
+	bits--;
+	for (wvalue=0, i=bits%5; i>=0; i--,bits--)
+		wvalue = (wvalue<<1)+BN_is_bit_set(p,bits);
+	bn_gather5(tmp.d,top,powerbuf,wvalue);
+
+	/* Scan the exponent one window at a time starting from the most
+	 * significant bits.
+	 */
+	while (bits >= 0)
+		{
+		for (wvalue=0, i=0; i<5; i++,bits--)
+			wvalue = (wvalue<<1)+BN_is_bit_set(p,bits);
+
+		bn_mul_mont(tmp.d,tmp.d,tmp.d,np,n0,top);
+		bn_mul_mont(tmp.d,tmp.d,tmp.d,np,n0,top);
+		bn_mul_mont(tmp.d,tmp.d,tmp.d,np,n0,top);
+		bn_mul_mont(tmp.d,tmp.d,tmp.d,np,n0,top);
+		bn_mul_mont(tmp.d,tmp.d,tmp.d,np,n0,top);
+		bn_mul_mont_gather5(tmp.d,tmp.d,powerbuf,np,n0,top,wvalue);
+		}
+
+	tmp.top=top;
+	bn_correct_top(&tmp);
+	}
+    else
+#endif
+	{
+	if (!MOD_EXP_CTIME_COPY_TO_PREBUF(&tmp, top, powerbuf, 0, numPowers)) goto err;
+	if (!MOD_EXP_CTIME_COPY_TO_PREBUF(&am,  top, powerbuf, 1, numPowers)) goto err;
 
 	/* If the window size is greater than 1, then calculate
 	 * val[i=2..2^winsize-1]. Powers are computed as a*a^(i-1)
@@ -662,62 +777,54 @@ int BN_mod_exp_mont_consttime(BIGNUM *rr, const BIGNUM *a, const BIGNUM *p,
 	 */
 	if (window > 1)
 		{
-		for (i=2; i<numPowers; i++)
+		if (!BN_mod_mul_montgomery(&tmp,&am,&am,mont,ctx))	goto err;
+		if (!MOD_EXP_CTIME_COPY_TO_PREBUF(&tmp, top, powerbuf, 2, numPowers)) goto err;
+		for (i=3; i<numPowers; i++)
 			{
 			/* Calculate a^i = a^(i-1) * a */
-			if (!BN_mod_mul_montgomery(computeTemp,am,computeTemp,mont,ctx))
+			if (!BN_mod_mul_montgomery(&tmp,&am,&tmp,mont,ctx))
 				goto err;
-			if (!MOD_EXP_CTIME_COPY_TO_PREBUF(computeTemp, top, powerbuf, i, numPowers)) goto err;
+			if (!MOD_EXP_CTIME_COPY_TO_PREBUF(&tmp, top, powerbuf, i, numPowers)) goto err;
 			}
 		}
 
- 	/* Adjust the number of bits up to a multiple of the window size.
- 	 * If the exponent length is not a multiple of the window size, then
- 	 * this pads the most significant bits with zeros to normalize the
- 	 * scanning loop to there's no special cases.
- 	 *
- 	 * * NOTE: Making the window size a power of two less than the native
-	 * * word size ensures that the padded bits won't go past the last
- 	 * * word in the internal BIGNUM structure. Going past the end will
- 	 * * still produce the correct result, but causes a different branch
- 	 * * to be taken in the BN_is_bit_set function.
- 	 */
- 	bits = ((bits+window-1)/window)*window;
- 	idx=bits-1;	/* The top bit of the window */
-
- 	/* Scan the exponent one window at a time starting from the most
- 	 * significant bits.
- 	 */
- 	while (idx >= 0)
+	bits--;
+	for (wvalue=0, i=bits%window; i>=0; i--,bits--)
+		wvalue = (wvalue<<1)+BN_is_bit_set(p,bits);
+	if (!MOD_EXP_CTIME_COPY_FROM_PREBUF(&tmp,top,powerbuf,wvalue,numPowers)) goto err;
+ 
+	/* Scan the exponent one window at a time starting from the most
+	 * significant bits.
+	 */
+ 	while (bits >= 0)
   		{
  		wvalue=0; /* The 'value' of the window */
  		
  		/* Scan the window, squaring the result as we go */
- 		for (i=0; i<window; i++,idx--)
+ 		for (i=0; i<window; i++,bits--)
  			{
-			if (!BN_mod_mul_montgomery(r,r,r,mont,ctx))	goto err;
-			wvalue = (wvalue<<1)+BN_is_bit_set(p,idx);
+			if (!BN_mod_mul_montgomery(&tmp,&tmp,&tmp,mont,ctx))	goto err;
+			wvalue = (wvalue<<1)+BN_is_bit_set(p,bits);
   			}
  		
 		/* Fetch the appropriate pre-computed value from the pre-buf */
-		if (!MOD_EXP_CTIME_COPY_FROM_PREBUF(computeTemp, top, powerbuf, wvalue, numPowers)) goto err;
+		if (!MOD_EXP_CTIME_COPY_FROM_PREBUF(&am, top, powerbuf, wvalue, numPowers)) goto err;
 
  		/* Multiply the result into the intermediate result */
- 		if (!BN_mod_mul_montgomery(r,r,computeTemp,mont,ctx)) goto err;
+ 		if (!BN_mod_mul_montgomery(&tmp,&tmp,&am,mont,ctx)) goto err;
   		}
+	}
 
  	/* Convert the final result from montgomery to standard format */
-	if (!BN_from_montgomery(rr,r,mont,ctx)) goto err;
+	if (!BN_from_montgomery(rr,&tmp,mont,ctx)) goto err;
 	ret=1;
 err:
 	if ((in_mont == NULL) && (mont != NULL)) BN_MONT_CTX_free(mont);
 	if (powerbuf!=NULL)
 		{
 		OPENSSL_cleanse(powerbuf,powerbufLen);
-		OPENSSL_free(powerbufFree);
+		if (powerbufFree) OPENSSL_free(powerbufFree);
 		}
- 	if (am!=NULL) BN_clear(am);
- 	if (computeTemp!=NULL) BN_clear(computeTemp);
 	BN_CTX_end(ctx);
 	return(ret);
 	}
@@ -988,4 +1095,3 @@ err:
 	bn_check_top(r);
 	return(ret);
 	}
-
diff --git a/crypto/bn/bn_gf2m.c b/crypto/bn/bn_gf2m.c
index 5d90f1e88be2..8a4dc20ad980 100644
--- a/crypto/bn/bn_gf2m.c
+++ b/crypto/bn/bn_gf2m.c
@@ -94,6 +94,8 @@
 #include "cryptlib.h"
 #include "bn_lcl.h"
 
+#ifndef OPENSSL_NO_EC2M
+
 /* Maximum number of iterations before BN_GF2m_mod_solve_quad_arr should fail. */
 #define MAX_ITERATIONS 50
 
@@ -121,74 +123,13 @@ static const BN_ULONG SQR_tb[16] =
     SQR_tb[(w) >> 12 & 0xF] << 24 | SQR_tb[(w) >>  8 & 0xF] << 16 | \
     SQR_tb[(w) >>  4 & 0xF] <<  8 | SQR_tb[(w)       & 0xF]
 #endif
-#ifdef SIXTEEN_BIT
-#define SQR1(w) \
-    SQR_tb[(w) >> 12 & 0xF] <<  8 | SQR_tb[(w) >>  8 & 0xF]
-#define SQR0(w) \
-    SQR_tb[(w) >>  4 & 0xF] <<  8 | SQR_tb[(w)       & 0xF]
-#endif
-#ifdef EIGHT_BIT
-#define SQR1(w) \
-    SQR_tb[(w) >>  4 & 0xF]
-#define SQR0(w) \
-    SQR_tb[(w)       & 15]
-#endif
 
+#if !defined(OPENSSL_BN_ASM_GF2m)
 /* Product of two polynomials a, b each with degree < BN_BITS2 - 1,
  * result is a polynomial r with degree < 2 * BN_BITS - 1
  * The caller MUST ensure that the variables have the right amount
  * of space allocated.
  */
-#ifdef EIGHT_BIT
-static void bn_GF2m_mul_1x1(BN_ULONG *r1, BN_ULONG *r0, const BN_ULONG a, const BN_ULONG b)
-	{
-	register BN_ULONG h, l, s;
-	BN_ULONG tab[4], top1b = a >> 7;
-	register BN_ULONG a1, a2;
-
-	a1 = a & (0x7F); a2 = a1 << 1;
-
-	tab[0] = 0; tab[1] = a1; tab[2] = a2; tab[3] = a1^a2;
-
-	s = tab[b      & 0x3]; l  = s;
-	s = tab[b >> 2 & 0x3]; l ^= s << 2; h  = s >> 6;
-	s = tab[b >> 4 & 0x3]; l ^= s << 4; h ^= s >> 4;
-	s = tab[b >> 6      ]; l ^= s << 6; h ^= s >> 2;
-	
-	/* compensate for the top bit of a */
-
-	if (top1b & 01) { l ^= b << 7; h ^= b >> 1; } 
-
-	*r1 = h; *r0 = l;
-	} 
-#endif
-#ifdef SIXTEEN_BIT
-static void bn_GF2m_mul_1x1(BN_ULONG *r1, BN_ULONG *r0, const BN_ULONG a, const BN_ULONG b)
-	{
-	register BN_ULONG h, l, s;
-	BN_ULONG tab[4], top1b = a >> 15; 
-	register BN_ULONG a1, a2;
-
-	a1 = a & (0x7FFF); a2 = a1 << 1;
-
-	tab[0] = 0; tab[1] = a1; tab[2] = a2; tab[3] = a1^a2;
-
-	s = tab[b      & 0x3]; l  = s;
-	s = tab[b >> 2 & 0x3]; l ^= s <<  2; h  = s >> 14;
-	s = tab[b >> 4 & 0x3]; l ^= s <<  4; h ^= s >> 12;
-	s = tab[b >> 6 & 0x3]; l ^= s <<  6; h ^= s >> 10;
-	s = tab[b >> 8 & 0x3]; l ^= s <<  8; h ^= s >>  8;
-	s = tab[b >>10 & 0x3]; l ^= s << 10; h ^= s >>  6;
-	s = tab[b >>12 & 0x3]; l ^= s << 12; h ^= s >>  4;
-	s = tab[b >>14      ]; l ^= s << 14; h ^= s >>  2;
-
-	/* compensate for the top bit of a */
-
-	if (top1b & 01) { l ^= b << 15; h ^= b >> 1; } 
-
-	*r1 = h; *r0 = l;
-	} 
-#endif
 #ifdef THIRTY_TWO_BIT
 static void bn_GF2m_mul_1x1(BN_ULONG *r1, BN_ULONG *r0, const BN_ULONG a, const BN_ULONG b)
 	{
@@ -278,7 +219,9 @@ static void bn_GF2m_mul_2x2(BN_ULONG *r, const BN_ULONG a1, const BN_ULONG a0, c
 	r[2] ^= m1 ^ r[1] ^ r[3];  /* h0 ^= m1 ^ l1 ^ h1; */
 	r[1] = r[3] ^ r[2] ^ r[0] ^ m1 ^ m0;  /* l1 ^= l0 ^ h0 ^ m0; */
 	}
-
+#else
+void bn_GF2m_mul_2x2(BN_ULONG *r, BN_ULONG a1, BN_ULONG a0, BN_ULONG b1, BN_ULONG b0);
+#endif 
 
 /* Add polynomials a and b and store result in r; r could be a or b, a and b 
  * could be equal; r is the bitwise XOR of a and b.
@@ -321,7 +264,7 @@ int	BN_GF2m_add(BIGNUM *r, const BIGNUM *a, const BIGNUM *b)
 
 
 /* Performs modular reduction of a and store result in r.  r could be a. */
-int BN_GF2m_mod_arr(BIGNUM *r, const BIGNUM *a, const unsigned int p[])
+int BN_GF2m_mod_arr(BIGNUM *r, const BIGNUM *a, const int p[])
 	{
 	int j, k;
 	int n, dN, d0, d1;
@@ -422,21 +365,17 @@ int BN_GF2m_mod_arr(BIGNUM *r, const BIGNUM *a, const unsigned int p[])
 int	BN_GF2m_mod(BIGNUM *r, const BIGNUM *a, const BIGNUM *p)
 	{
 	int ret = 0;
-	const int max = BN_num_bits(p);
-	unsigned int *arr=NULL;
+	int arr[6];
 	bn_check_top(a);
 	bn_check_top(p);
-	if ((arr = (unsigned int *)OPENSSL_malloc(sizeof(unsigned int) * max)) == NULL) goto err;
-	ret = BN_GF2m_poly2arr(p, arr, max);
-	if (!ret || ret > max)
+	ret = BN_GF2m_poly2arr(p, arr, sizeof(arr)/sizeof(arr[0]));
+	if (!ret || ret > (int)(sizeof(arr)/sizeof(arr[0])))
 		{
 		BNerr(BN_F_BN_GF2M_MOD,BN_R_INVALID_LENGTH);
-		goto err;
+		return 0;
 		}
 	ret = BN_GF2m_mod_arr(r, a, arr);
 	bn_check_top(r);
-err:
-	if (arr) OPENSSL_free(arr);
 	return ret;
 	}
 
@@ -444,7 +383,7 @@ err:
 /* Compute the product of two polynomials a and b, reduce modulo p, and store
  * the result in r.  r could be a or b; a could be b.
  */
-int	BN_GF2m_mod_mul_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const unsigned int p[], BN_CTX *ctx)
+int	BN_GF2m_mod_mul_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const int p[], BN_CTX *ctx)
 	{
 	int zlen, i, j, k, ret = 0;
 	BIGNUM *s;
@@ -500,12 +439,12 @@ err:
 int	BN_GF2m_mod_mul(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const BIGNUM *p, BN_CTX *ctx)
 	{
 	int ret = 0;
-	const int max = BN_num_bits(p);
-	unsigned int *arr=NULL;
+	const int max = BN_num_bits(p) + 1;
+	int *arr=NULL;
 	bn_check_top(a);
 	bn_check_top(b);
 	bn_check_top(p);
-	if ((arr = (unsigned int *)OPENSSL_malloc(sizeof(unsigned int) * max)) == NULL) goto err;
+	if ((arr = (int *)OPENSSL_malloc(sizeof(int) * max)) == NULL) goto err;
 	ret = BN_GF2m_poly2arr(p, arr, max);
 	if (!ret || ret > max)
 		{
@@ -521,7 +460,7 @@ err:
 
 
 /* Square a, reduce the result mod p, and store it in a.  r could be a. */
-int	BN_GF2m_mod_sqr_arr(BIGNUM *r, const BIGNUM *a, const unsigned int p[], BN_CTX *ctx)
+int	BN_GF2m_mod_sqr_arr(BIGNUM *r, const BIGNUM *a, const int p[], BN_CTX *ctx)
 	{
 	int i, ret = 0;
 	BIGNUM *s;
@@ -556,12 +495,12 @@ err:
 int	BN_GF2m_mod_sqr(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx)
 	{
 	int ret = 0;
-	const int max = BN_num_bits(p);
-	unsigned int *arr=NULL;
+	const int max = BN_num_bits(p) + 1;
+	int *arr=NULL;
 
 	bn_check_top(a);
 	bn_check_top(p);
-	if ((arr = (unsigned int *)OPENSSL_malloc(sizeof(unsigned int) * max)) == NULL) goto err;
+	if ((arr = (int *)OPENSSL_malloc(sizeof(int) * max)) == NULL) goto err;
 	ret = BN_GF2m_poly2arr(p, arr, max);
 	if (!ret || ret > max)
 		{
@@ -583,7 +522,7 @@ err:
  */
 int BN_GF2m_mod_inv(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx)
 	{
-	BIGNUM *b, *c, *u, *v, *tmp;
+	BIGNUM *b, *c = NULL, *u = NULL, *v = NULL, *tmp;
 	int ret = 0;
 
 	bn_check_top(a);
@@ -591,18 +530,18 @@ int BN_GF2m_mod_inv(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx)
 
 	BN_CTX_start(ctx);
 	
-	b = BN_CTX_get(ctx);
-	c = BN_CTX_get(ctx);
-	u = BN_CTX_get(ctx);
-	v = BN_CTX_get(ctx);
-	if (v == NULL) goto err;
+	if ((b = BN_CTX_get(ctx))==NULL) goto err;
+	if ((c = BN_CTX_get(ctx))==NULL) goto err;
+	if ((u = BN_CTX_get(ctx))==NULL) goto err;
+	if ((v = BN_CTX_get(ctx))==NULL) goto err;
 
-	if (!BN_one(b)) goto err;
 	if (!BN_GF2m_mod(u, a, p)) goto err;
-	if (!BN_copy(v, p)) goto err;
-
 	if (BN_is_zero(u)) goto err;
 
+	if (!BN_copy(v, p)) goto err;
+#if 0
+	if (!BN_one(b)) goto err;
+
 	while (1)
 		{
 		while (!BN_is_odd(u))
@@ -627,13 +566,89 @@ int BN_GF2m_mod_inv(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx)
 		if (!BN_GF2m_add(u, u, v)) goto err;
 		if (!BN_GF2m_add(b, b, c)) goto err;
 		}
+#else
+	{
+	int i,	ubits = BN_num_bits(u),
+		vbits = BN_num_bits(v),	/* v is copy of p */
+		top = p->top;
+	BN_ULONG *udp,*bdp,*vdp,*cdp;
+
+	bn_wexpand(u,top);	udp = u->d;
+				for (i=u->top;i<top;i++) udp[i] = 0;
+				u->top = top;
+	bn_wexpand(b,top);	bdp = b->d;
+				bdp[0] = 1;
+				for (i=1;i<top;i++) bdp[i] = 0;
+				b->top = top;
+	bn_wexpand(c,top);	cdp = c->d;
+				for (i=0;i<top;i++) cdp[i] = 0;
+				c->top = top;
+	vdp = v->d;	/* It pays off to "cache" *->d pointers, because
+			 * it allows optimizer to be more aggressive.
+			 * But we don't have to "cache" p->d, because *p
+			 * is declared 'const'... */
+	while (1)
+		{
+		while (ubits && !(udp[0]&1))
+			{
+			BN_ULONG u0,u1,b0,b1,mask;
+
+			u0   = udp[0];
+			b0   = bdp[0];
+			mask = (BN_ULONG)0-(b0&1);
+			b0  ^= p->d[0]&mask;
+			for (i=0;i<top-1;i++)
+				{
+				u1 = udp[i+1];
+				udp[i] = ((u0>>1)|(u1<<(BN_BITS2-1)))&BN_MASK2;
+				u0 = u1;
+				b1 = bdp[i+1]^(p->d[i+1]&mask);
+				bdp[i] = ((b0>>1)|(b1<<(BN_BITS2-1)))&BN_MASK2;
+				b0 = b1;
+				}
+			udp[i] = u0>>1;
+			bdp[i] = b0>>1;
+			ubits--;
+			}
+
+		if (ubits<=BN_BITS2 && udp[0]==1) break;
 
+		if (ubits<vbits)
+			{
+			i = ubits; ubits = vbits; vbits = i;
+			tmp = u; u = v; v = tmp;
+			tmp = b; b = c; c = tmp;
+			udp = vdp; vdp = v->d;
+			bdp = cdp; cdp = c->d;
+			}
+		for(i=0;i<top;i++)
+			{
+			udp[i] ^= vdp[i];
+			bdp[i] ^= cdp[i];
+			}
+		if (ubits==vbits)
+			{
+			BN_ULONG ul;
+			int utop = (ubits-1)/BN_BITS2;
+
+			while ((ul=udp[utop])==0 && utop) utop--;
+			ubits = utop*BN_BITS2 + BN_num_bits_word(ul);
+			}
+		}
+	bn_correct_top(b);
+	}
+#endif
 
 	if (!BN_copy(r, b)) goto err;
 	bn_check_top(r);
 	ret = 1;
 
 err:
+#ifdef BN_DEBUG /* BN_CTX_end would complain about the expanded form */
+        bn_correct_top(c);
+        bn_correct_top(u);
+        bn_correct_top(v);
+#endif
   	BN_CTX_end(ctx);
 	return ret;
 	}
@@ -644,7 +659,7 @@ err:
  * function is only provided for convenience; for best performance, use the 
  * BN_GF2m_mod_inv function.
  */
-int BN_GF2m_mod_inv_arr(BIGNUM *r, const BIGNUM *xx, const unsigned int p[], BN_CTX *ctx)
+int BN_GF2m_mod_inv_arr(BIGNUM *r, const BIGNUM *xx, const int p[], BN_CTX *ctx)
 	{
 	BIGNUM *field;
 	int ret = 0;
@@ -770,7 +785,7 @@ err:
  * function is only provided for convenience; for best performance, use the 
  * BN_GF2m_mod_div function.
  */
-int BN_GF2m_mod_div_arr(BIGNUM *r, const BIGNUM *yy, const BIGNUM *xx, const unsigned int p[], BN_CTX *ctx)
+int BN_GF2m_mod_div_arr(BIGNUM *r, const BIGNUM *yy, const BIGNUM *xx, const int p[], BN_CTX *ctx)
 	{
 	BIGNUM *field;
 	int ret = 0;
@@ -795,7 +810,7 @@ err:
  * the result in r.  r could be a.
  * Uses simple square-and-multiply algorithm A.5.1 from IEEE P1363.
  */
-int	BN_GF2m_mod_exp_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const unsigned int p[], BN_CTX *ctx)
+int	BN_GF2m_mod_exp_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const int p[], BN_CTX *ctx)
 	{
 	int ret = 0, i, n;
 	BIGNUM *u;
@@ -841,12 +856,12 @@ err:
 int BN_GF2m_mod_exp(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const BIGNUM *p, BN_CTX *ctx)
 	{
 	int ret = 0;
-	const int max = BN_num_bits(p);
-	unsigned int *arr=NULL;
+	const int max = BN_num_bits(p) + 1;
+	int *arr=NULL;
 	bn_check_top(a);
 	bn_check_top(b);
 	bn_check_top(p);
-	if ((arr = (unsigned int *)OPENSSL_malloc(sizeof(unsigned int) * max)) == NULL) goto err;
+	if ((arr = (int *)OPENSSL_malloc(sizeof(int) * max)) == NULL) goto err;
 	ret = BN_GF2m_poly2arr(p, arr, max);
 	if (!ret || ret > max)
 		{
@@ -864,7 +879,7 @@ err:
  * the result in r.  r could be a.
  * Uses exponentiation as in algorithm A.4.1 from IEEE P1363.
  */
-int	BN_GF2m_mod_sqrt_arr(BIGNUM *r, const BIGNUM *a, const unsigned int p[], BN_CTX *ctx)
+int	BN_GF2m_mod_sqrt_arr(BIGNUM *r, const BIGNUM *a, const int p[], BN_CTX *ctx)
 	{
 	int ret = 0;
 	BIGNUM *u;
@@ -900,11 +915,11 @@ err:
 int BN_GF2m_mod_sqrt(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx)
 	{
 	int ret = 0;
-	const int max = BN_num_bits(p);
-	unsigned int *arr=NULL;
+	const int max = BN_num_bits(p) + 1;
+	int *arr=NULL;
 	bn_check_top(a);
 	bn_check_top(p);
-	if ((arr = (unsigned int *)OPENSSL_malloc(sizeof(unsigned int) * max)) == NULL) goto err;
+	if ((arr = (int *)OPENSSL_malloc(sizeof(int) * max)) == NULL) goto err;
 	ret = BN_GF2m_poly2arr(p, arr, max);
 	if (!ret || ret > max)
 		{
@@ -921,10 +936,9 @@ err:
 /* Find r such that r^2 + r = a mod p.  r could be a. If no r exists returns 0.
  * Uses algorithms A.4.7 and A.4.6 from IEEE P1363.
  */
-int BN_GF2m_mod_solve_quad_arr(BIGNUM *r, const BIGNUM *a_, const unsigned int p[], BN_CTX *ctx)
+int BN_GF2m_mod_solve_quad_arr(BIGNUM *r, const BIGNUM *a_, const int p[], BN_CTX *ctx)
 	{
-	int ret = 0, count = 0;
-	unsigned int j;
+	int ret = 0, count = 0, j;
 	BIGNUM *a, *z, *rho, *w, *w2, *tmp;
 
 	bn_check_top(a_);
@@ -1019,11 +1033,11 @@ err:
 int BN_GF2m_mod_solve_quad(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx)
 	{
 	int ret = 0;
-	const int max = BN_num_bits(p);
-	unsigned int *arr=NULL;
+	const int max = BN_num_bits(p) + 1;
+	int *arr=NULL;
 	bn_check_top(a);
 	bn_check_top(p);
-	if ((arr = (unsigned int *)OPENSSL_malloc(sizeof(unsigned int) *
+	if ((arr = (int *)OPENSSL_malloc(sizeof(int) *
 						max)) == NULL) goto err;
 	ret = BN_GF2m_poly2arr(p, arr, max);
 	if (!ret || ret > max)
@@ -1039,20 +1053,17 @@ err:
 	}
 
 /* Convert the bit-string representation of a polynomial
- * ( \sum_{i=0}^n a_i * x^i , where a_0 is *not* zero) into an array
- * of integers corresponding to the bits with non-zero coefficient.
+ * ( \sum_{i=0}^n a_i * x^i) into an array of integers corresponding 
+ * to the bits with non-zero coefficient.  Array is terminated with -1.
  * Up to max elements of the array will be filled.  Return value is total
- * number of coefficients that would be extracted if array was large enough.
+ * number of array elements that would be filled if array was large enough.
  */
-int BN_GF2m_poly2arr(const BIGNUM *a, unsigned int p[], int max)
+int BN_GF2m_poly2arr(const BIGNUM *a, int p[], int max)
 	{
 	int i, j, k = 0;
 	BN_ULONG mask;
 
-	if (BN_is_zero(a) || !BN_is_bit_set(a, 0))
-		/* a_0 == 0 => return error (the unsigned int array
-		 * must be terminated by 0)
-		 */
+	if (BN_is_zero(a))
 		return 0;
 
 	for (i = a->top - 1; i >= 0; i--)
@@ -1072,26 +1083,31 @@ int BN_GF2m_poly2arr(const BIGNUM *a, unsigned int p[], int max)
 			}
 		}
 
+	if (k < max) {
+		p[k] = -1;
+		k++;
+	}
+
 	return k;
 	}
 
 /* Convert the coefficient array representation of a polynomial to a 
- * bit-string.  The array must be terminated by 0.
+ * bit-string.  The array must be terminated by -1.
  */
-int BN_GF2m_arr2poly(const unsigned int p[], BIGNUM *a)
+int BN_GF2m_arr2poly(const int p[], BIGNUM *a)
 	{
 	int i;
 
 	bn_check_top(a);
 	BN_zero(a);
-	for (i = 0; p[i] != 0; i++)
+	for (i = 0; p[i] != -1; i++)
 		{
 		if (BN_set_bit(a, p[i]) == 0)
 			return 0;
 		}
-	BN_set_bit(a, 0);
 	bn_check_top(a);
 
 	return 1;
 	}
 
+#endif
diff --git a/crypto/bn/bn_lcl.h b/crypto/bn/bn_lcl.h
index 27ac4397a151..eecfd8cc99e7 100644
--- a/crypto/bn/bn_lcl.h
+++ b/crypto/bn/bn_lcl.h
@@ -238,7 +238,7 @@ extern "C" {
 #  if defined(__DECC)
 #   include <c_asm.h>
 #   define BN_UMULT_HIGH(a,b)	(BN_ULONG)asm("umulh %a0,%a1,%v0",(a),(b))
-#  elif defined(__GNUC__)
+#  elif defined(__GNUC__) && __GNUC__>=2
 #   define BN_UMULT_HIGH(a,b)	({	\
 	register BN_ULONG ret;		\
 	asm ("umulh	%1,%2,%0"	\
@@ -247,7 +247,7 @@ extern "C" {
 	ret;			})
 #  endif	/* compiler */
 # elif defined(_ARCH_PPC) && defined(__64BIT__) && defined(SIXTY_FOUR_BIT_LONG)
-#  if defined(__GNUC__)
+#  if defined(__GNUC__) && __GNUC__>=2
 #   define BN_UMULT_HIGH(a,b)	({	\
 	register BN_ULONG ret;		\
 	asm ("mulhdu	%0,%1,%2"	\
@@ -255,8 +255,9 @@ extern "C" {
 	     : "r"(a), "r"(b));		\
 	ret;			})
 #  endif	/* compiler */
-# elif defined(__x86_64) && defined(SIXTY_FOUR_BIT_LONG)
-#  if defined(__GNUC__)
+# elif (defined(__x86_64) || defined(__x86_64__)) && \
+       (defined(SIXTY_FOUR_BIT_LONG) || defined(SIXTY_FOUR_BIT))
+#  if defined(__GNUC__) && __GNUC__>=2
 #   define BN_UMULT_HIGH(a,b)	({	\
 	register BN_ULONG ret,discard;	\
 	asm ("mulq	%3"		\
@@ -279,6 +280,19 @@ extern "C" {
 #   define BN_UMULT_HIGH(a,b)		__umulh((a),(b))
 #   define BN_UMULT_LOHI(low,high,a,b)	((low)=_umul128((a),(b),&(high)))
 #  endif
+# elif defined(__mips) && (defined(SIXTY_FOUR_BIT) || defined(SIXTY_FOUR_BIT_LONG))
+#  if defined(__GNUC__) && __GNUC__>=2
+#   define BN_UMULT_HIGH(a,b)	({	\
+	register BN_ULONG ret;		\
+	asm ("dmultu	%1,%2"		\
+	     : "=h"(ret)		\
+	     : "r"(a), "r"(b) : "l");	\
+	ret;			})
+#   define BN_UMULT_LOHI(low,high,a,b)	\
+	asm ("dmultu	%2,%3"		\
+	     : "=l"(low),"=h"(high)	\
+	     : "r"(a), "r"(b));
+#  endif
 # endif		/* cpu */
 #endif		/* OPENSSL_NO_ASM */
 
@@ -458,6 +472,10 @@ extern "C" {
 	}
 #endif /* !BN_LLONG */
 
+#if defined(OPENSSL_DOING_MAKEDEPEND) && defined(OPENSSL_FIPS)
+#undef bn_div_words
+#endif
+
 void bn_mul_normal(BN_ULONG *r,BN_ULONG *a,int na,BN_ULONG *b,int nb);
 void bn_mul_comba8(BN_ULONG *r,BN_ULONG *a,BN_ULONG *b);
 void bn_mul_comba4(BN_ULONG *r,BN_ULONG *a,BN_ULONG *b);
diff --git a/crypto/bn/bn_lib.c b/crypto/bn/bn_lib.c
index 32a8fbaf51ee..7a5676de6927 100644
--- a/crypto/bn/bn_lib.c
+++ b/crypto/bn/bn_lib.c
@@ -133,15 +133,15 @@ int BN_get_params(int which)
 
 const BIGNUM *BN_value_one(void)
 	{
-	static BN_ULONG data_one=1L;
-	static BIGNUM const_one={&data_one,1,1,0,BN_FLG_STATIC_DATA};
+	static const BN_ULONG data_one=1L;
+	static const BIGNUM const_one={(BN_ULONG *)&data_one,1,1,0,BN_FLG_STATIC_DATA};
 
 	return(&const_one);
 	}
 
 int BN_num_bits_word(BN_ULONG l)
 	{
-	static const char bits[256]={
+	static const unsigned char bits[256]={
 		0,1,2,2,3,3,3,3,4,4,4,4,4,4,4,4,
 		5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,
 		6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,
@@ -216,7 +216,7 @@ int BN_num_bits_word(BN_ULONG l)
 		else
 #endif
 			{
-#if defined(SIXTEEN_BIT) || defined(THIRTY_TWO_BIT) || defined(SIXTY_FOUR_BIT) || defined(SIXTY_FOUR_BIT_LONG)
+#if defined(THIRTY_TWO_BIT) || defined(SIXTY_FOUR_BIT) || defined(SIXTY_FOUR_BIT_LONG)
 			if (l & 0xff00L)
 				return(bits[(int)(l>>8)]+8);
 			else	
@@ -744,7 +744,7 @@ int BN_is_bit_set(const BIGNUM *a, int n)
 	i=n/BN_BITS2;
 	j=n%BN_BITS2;
 	if (a->top <= i) return 0;
-	return(((a->d[i])>>j)&((BN_ULONG)1));
+	return (int)(((a->d[i])>>j)&((BN_ULONG)1));
 	}
 
 int BN_mask_bits(BIGNUM *a, int n)
diff --git a/crypto/bn/bn_mont.c b/crypto/bn/bn_mont.c
index 4799b152ddcb..427b5cf4df95 100644
--- a/crypto/bn/bn_mont.c
+++ b/crypto/bn/bn_mont.c
@@ -122,26 +122,10 @@
 
 #define MONT_WORD /* use the faster word-based algorithm */
 
-#if defined(MONT_WORD) && defined(OPENSSL_BN_ASM_MONT) && (BN_BITS2<=32)
-/* This condition means we have a specific non-default build:
- * In the 0.9.8 branch, OPENSSL_BN_ASM_MONT is normally not set for any
- * BN_BITS2<=32 platform; an explicit "enable-montasm" is required.
- * I.e., if we are here, the user intentionally deviates from the
- * normal stable build to get better Montgomery performance from
- * the 0.9.9-dev backport.
- *
- * In this case only, we also enable BN_from_montgomery_word()
- * (another non-stable feature from 0.9.9-dev).
- */
-#define MONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD
-#endif
-
-#ifdef MONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD
+#ifdef MONT_WORD
 static int BN_from_montgomery_word(BIGNUM *ret, BIGNUM *r, BN_MONT_CTX *mont);
 #endif
 
-
-
 int BN_mod_mul_montgomery(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
 			  BN_MONT_CTX *mont, BN_CTX *ctx)
 	{
@@ -153,11 +137,7 @@ int BN_mod_mul_montgomery(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
 	if (num>1 && a->top==num && b->top==num)
 		{
 		if (bn_wexpand(r,num) == NULL) return(0);
-#if 0 /* for OpenSSL 0.9.9 mont->n0 */
 		if (bn_mul_mont(r->d,a->d,b->d,mont->N.d,mont->n0,num))
-#else
-		if (bn_mul_mont(r->d,a->d,b->d,mont->N.d,&mont->n0,num))
-#endif
 			{
 			r->neg = a->neg^b->neg;
 			r->top = num;
@@ -181,7 +161,7 @@ int BN_mod_mul_montgomery(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
 		if (!BN_mul(tmp,a,b,ctx)) goto err;
 		}
 	/* reduce from aRR to aR */
-#ifdef MONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD
+#ifdef MONT_WORD
 	if (!BN_from_montgomery_word(r,tmp,mont)) goto err;
 #else
 	if (!BN_from_montgomery(r,tmp,mont,ctx)) goto err;
@@ -193,44 +173,39 @@ err:
 	return(ret);
 	}
 
-#ifdef MONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD
+#ifdef MONT_WORD
 static int BN_from_montgomery_word(BIGNUM *ret, BIGNUM *r, BN_MONT_CTX *mont)
 	{
 	BIGNUM *n;
-	BN_ULONG *ap,*np,*rp,n0,v,*nrp;
-	int al,nl,max,i,x,ri;
+	BN_ULONG *ap,*np,*rp,n0,v,carry;
+	int nl,max,i;
 
 	n= &(mont->N);
-	/* mont->ri is the size of mont->N in bits (rounded up
-	   to the word size) */
-	al=ri=mont->ri/BN_BITS2;
-
 	nl=n->top;
-	if ((al == 0) || (nl == 0)) { ret->top=0; return(1); }
+	if (nl == 0) { ret->top=0; return(1); }
 
-	max=(nl+al+1); /* allow for overflow (no?) XXX */
+	max=(2*nl); /* carry is stored separately */
 	if (bn_wexpand(r,max) == NULL) return(0);
 
 	r->neg^=n->neg;
 	np=n->d;
 	rp=r->d;
-	nrp= &(r->d[nl]);
 
 	/* clear the top words of T */
+#if 1
 	for (i=r->top; i<max; i++) /* memset? XXX */
-		r->d[i]=0;
+		rp[i]=0;
+#else
+	memset(&(rp[r->top]),0,(max-r->top)*sizeof(BN_ULONG)); 
+#endif
 
 	r->top=max;
-#if 0 /* for OpenSSL 0.9.9 mont->n0 */
 	n0=mont->n0[0];
-#else
-	n0=mont->n0;
-#endif
 
 #ifdef BN_COUNT
 	fprintf(stderr,"word BN_from_montgomery_word %d * %d\n",nl,nl);
 #endif
-	for (i=0; i<nl; i++)
+	for (carry=0, i=0; i<nl; i++, rp++)
 		{
 #ifdef __TANDEM
                 {
@@ -248,59 +223,33 @@ static int BN_from_montgomery_word(BIGNUM *ret, BIGNUM *r, BN_MONT_CTX *mont)
 #else
 		v=bn_mul_add_words(rp,np,nl,(rp[0]*n0)&BN_MASK2);
 #endif
-		nrp++;
-		rp++;
-		if (((nrp[-1]+=v)&BN_MASK2) >= v)
-			continue;
-		else
-			{
-			if (((++nrp[0])&BN_MASK2) != 0) continue;
-			if (((++nrp[1])&BN_MASK2) != 0) continue;
-			for (x=2; (((++nrp[x])&BN_MASK2) == 0); x++) ;
-			}
+		v = (v+carry+rp[nl])&BN_MASK2;
+		carry |= (v != rp[nl]);
+		carry &= (v <= rp[nl]);
+		rp[nl]=v;
 		}
-	bn_correct_top(r);
 
-	/* mont->ri will be a multiple of the word size and below code
-	 * is kind of BN_rshift(ret,r,mont->ri) equivalent */
-	if (r->top <= ri)
-		{
-		ret->top=0;
-		return(1);
-		}
-	al=r->top-ri;
-
-	if (bn_wexpand(ret,ri) == NULL) return(0);
-	x=0-(((al-ri)>>(sizeof(al)*8-1))&1);
-	ret->top=x=(ri&~x)|(al&x);	/* min(ri,al) */
+	if (bn_wexpand(ret,nl) == NULL) return(0);
+	ret->top=nl;
 	ret->neg=r->neg;
 
 	rp=ret->d;
-	ap=&(r->d[ri]);
+	ap=&(r->d[nl]);
 
+#define BRANCH_FREE 1
+#if BRANCH_FREE
 	{
-	size_t m1,m2;
-
-	v=bn_sub_words(rp,ap,np,ri);
-	/* this ----------------^^ works even in al<ri case
-	 * thanks to zealous zeroing of top of the vector in the
-	 * beginning. */
+	BN_ULONG *nrp;
+	size_t m;
 
-	/* if (al==ri && !v) || al>ri) nrp=rp; else nrp=ap; */
-	/* in other words if subtraction result is real, then
+	v=bn_sub_words(rp,ap,np,nl)-carry;
+	/* if subtraction result is real, then
 	 * trick unconditional memcpy below to perform in-place
 	 * "refresh" instead of actual copy. */
-	m1=0-(size_t)(((al-ri)>>(sizeof(al)*8-1))&1);	/* al<ri */
-	m2=0-(size_t)(((ri-al)>>(sizeof(al)*8-1))&1);	/* al>ri */
-	m1|=m2;			/* (al!=ri) */
-	m1|=(0-(size_t)v);	/* (al!=ri || v) */
-	m1&=~m2;		/* (al!=ri || v) && !al>ri */
-	nrp=(BN_ULONG *)(((size_t)rp&~m1)|((size_t)ap&m1));
-	}
+	m=(0-(size_t)v);
+	nrp=(BN_ULONG *)(((PTR_SIZE_INT)rp&~m)|((PTR_SIZE_INT)ap&m));
 
-	/* 'i<ri' is chosen to eliminate dependency on input data, even
-	 * though it results in redundant copy in al<ri case. */
-	for (i=0,ri-=4; i<ri; i+=4)
+	for (i=0,nl-=4; i<nl; i+=4)
 		{
 		BN_ULONG t1,t2,t3,t4;
 		
@@ -313,192 +262,33 @@ static int BN_from_montgomery_word(BIGNUM *ret, BIGNUM *r, BN_MONT_CTX *mont)
 		rp[i+2]=t3;
 		rp[i+3]=t4;
 		}
-	for (ri+=4; i<ri; i++)
+	for (nl+=4; i<nl; i++)
 		rp[i]=nrp[i], ap[i]=0;
+	}
+#else
+	if (bn_sub_words (rp,ap,np,nl)-carry)
+		memcpy(rp,ap,nl*sizeof(BN_ULONG));
+#endif
 	bn_correct_top(r);
 	bn_correct_top(ret);
 	bn_check_top(ret);
 
 	return(1);
 	}
+#endif	/* MONT_WORD */
 
 int BN_from_montgomery(BIGNUM *ret, const BIGNUM *a, BN_MONT_CTX *mont,
 	     BN_CTX *ctx)
 	{
 	int retn=0;
+#ifdef MONT_WORD
 	BIGNUM *t;
 
 	BN_CTX_start(ctx);
 	if ((t = BN_CTX_get(ctx)) && BN_copy(t,a))
 		retn = BN_from_montgomery_word(ret,t,mont);
 	BN_CTX_end(ctx);
-	return retn;
-	}
-
-#else /* !MONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD */
-
-int BN_from_montgomery(BIGNUM *ret, const BIGNUM *a, BN_MONT_CTX *mont,
-	     BN_CTX *ctx)
-	{
-	int retn=0;
-
-#ifdef MONT_WORD
-	BIGNUM *n,*r;
-	BN_ULONG *ap,*np,*rp,n0,v,*nrp;
-	int al,nl,max,i,x,ri;
-
-	BN_CTX_start(ctx);
-	if ((r = BN_CTX_get(ctx)) == NULL) goto err;
-
-	if (!BN_copy(r,a)) goto err;
-	n= &(mont->N);
-
-	ap=a->d;
-	/* mont->ri is the size of mont->N in bits (rounded up
-	   to the word size) */
-	al=ri=mont->ri/BN_BITS2;
-	
-	nl=n->top;
-	if ((al == 0) || (nl == 0)) { r->top=0; return(1); }
-
-	max=(nl+al+1); /* allow for overflow (no?) XXX */
-	if (bn_wexpand(r,max) == NULL) goto err;
-
-	r->neg=a->neg^n->neg;
-	np=n->d;
-	rp=r->d;
-	nrp= &(r->d[nl]);
-
-	/* clear the top words of T */
-#if 1
-	for (i=r->top; i<max; i++) /* memset? XXX */
-		r->d[i]=0;
-#else
-	memset(&(r->d[r->top]),0,(max-r->top)*sizeof(BN_ULONG)); 
-#endif
-
-	r->top=max;
-	n0=mont->n0;
-
-#ifdef BN_COUNT
-	fprintf(stderr,"word BN_from_montgomery %d * %d\n",nl,nl);
-#endif
-	for (i=0; i<nl; i++)
-		{
-#ifdef __TANDEM
-                {
-                   long long t1;
-                   long long t2;
-                   long long t3;
-                   t1 = rp[0] * (n0 & 0177777);
-                   t2 = 037777600000l;
-                   t2 = n0 & t2;
-                   t3 = rp[0] & 0177777;
-                   t2 = (t3 * t2) & BN_MASK2;
-                   t1 = t1 + t2;
-                   v=bn_mul_add_words(rp,np,nl,(BN_ULONG) t1);
-                }
-#else
-		v=bn_mul_add_words(rp,np,nl,(rp[0]*n0)&BN_MASK2);
-#endif
-		nrp++;
-		rp++;
-		if (((nrp[-1]+=v)&BN_MASK2) >= v)
-			continue;
-		else
-			{
-			if (((++nrp[0])&BN_MASK2) != 0) continue;
-			if (((++nrp[1])&BN_MASK2) != 0) continue;
-			for (x=2; (((++nrp[x])&BN_MASK2) == 0); x++) ;
-			}
-		}
-	bn_correct_top(r);
-	
-	/* mont->ri will be a multiple of the word size and below code
-	 * is kind of BN_rshift(ret,r,mont->ri) equivalent */
-	if (r->top <= ri)
-		{
-		ret->top=0;
-		retn=1;
-		goto err;
-		}
-	al=r->top-ri;
-
-# define BRANCH_FREE 1
-# if BRANCH_FREE
-	if (bn_wexpand(ret,ri) == NULL) goto err;
-	x=0-(((al-ri)>>(sizeof(al)*8-1))&1);
-	ret->top=x=(ri&~x)|(al&x);	/* min(ri,al) */
-	ret->neg=r->neg;
-
-	rp=ret->d;
-	ap=&(r->d[ri]);
-
-	{
-	size_t m1,m2;
-
-	v=bn_sub_words(rp,ap,np,ri);
-	/* this ----------------^^ works even in al<ri case
-	 * thanks to zealous zeroing of top of the vector in the
-	 * beginning. */
-
-	/* if (al==ri && !v) || al>ri) nrp=rp; else nrp=ap; */
-	/* in other words if subtraction result is real, then
-	 * trick unconditional memcpy below to perform in-place
-	 * "refresh" instead of actual copy. */
-	m1=0-(size_t)(((al-ri)>>(sizeof(al)*8-1))&1);	/* al<ri */
-	m2=0-(size_t)(((ri-al)>>(sizeof(al)*8-1))&1);	/* al>ri */
-	m1|=m2;			/* (al!=ri) */
-	m1|=(0-(size_t)v);	/* (al!=ri || v) */
-	m1&=~m2;		/* (al!=ri || v) && !al>ri */
-	nrp=(BN_ULONG *)(((size_t)rp&~m1)|((size_t)ap&m1));
-	}
-
-	/* 'i<ri' is chosen to eliminate dependency on input data, even
-	 * though it results in redundant copy in al<ri case. */
-	for (i=0,ri-=4; i<ri; i+=4)
-		{
-		BN_ULONG t1,t2,t3,t4;
-		
-		t1=nrp[i+0];
-		t2=nrp[i+1];
-		t3=nrp[i+2];	ap[i+0]=0;
-		t4=nrp[i+3];	ap[i+1]=0;
-		rp[i+0]=t1;	ap[i+2]=0;
-		rp[i+1]=t2;	ap[i+3]=0;
-		rp[i+2]=t3;
-		rp[i+3]=t4;
-		}
-	for (ri+=4; i<ri; i++)
-		rp[i]=nrp[i], ap[i]=0;
-	bn_correct_top(r);
-	bn_correct_top(ret);
-# else
-	if (bn_wexpand(ret,al) == NULL) goto err;
-	ret->top=al;
-	ret->neg=r->neg;
-
-	rp=ret->d;
-	ap=&(r->d[ri]);
-	al-=4;
-	for (i=0; i<al; i+=4)
-		{
-		BN_ULONG t1,t2,t3,t4;
-		
-		t1=ap[i+0];
-		t2=ap[i+1];
-		t3=ap[i+2];
-		t4=ap[i+3];
-		rp[i+0]=t1;
-		rp[i+1]=t2;
-		rp[i+2]=t3;
-		rp[i+3]=t4;
-		}
-	al+=4;
-	for (; i<al; i++)
-		rp[i]=ap[i];
-# endif
-#else /* !MONT_WORD */ 
+#else /* !MONT_WORD */
 	BIGNUM *t1,*t2;
 
 	BN_CTX_start(ctx);
@@ -515,21 +305,18 @@ int BN_from_montgomery(BIGNUM *ret, const BIGNUM *a, BN_MONT_CTX *mont,
 	if (!BN_mul(t1,t2,&mont->N,ctx)) goto err;
 	if (!BN_add(t2,a,t1)) goto err;
 	if (!BN_rshift(ret,t2,mont->ri)) goto err;
-#endif /* MONT_WORD */
 
-#if !defined(BRANCH_FREE) || BRANCH_FREE==0
 	if (BN_ucmp(ret, &(mont->N)) >= 0)
 		{
 		if (!BN_usub(ret,ret,&(mont->N))) goto err;
 		}
-#endif
 	retn=1;
 	bn_check_top(ret);
  err:
 	BN_CTX_end(ctx);
+#endif /* MONT_WORD */
 	return(retn);
 	}
-#endif /* MONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD */
 
 BN_MONT_CTX *BN_MONT_CTX_new(void)
 	{
@@ -549,11 +336,7 @@ void BN_MONT_CTX_init(BN_MONT_CTX *ctx)
 	BN_init(&(ctx->RR));
 	BN_init(&(ctx->N));
 	BN_init(&(ctx->Ni));
-#if 0 /* for OpenSSL 0.9.9 mont->n0 */
 	ctx->n0[0] = ctx->n0[1] = 0;
-#else
-	ctx->n0 = 0;
-#endif
 	ctx->flags=0;
 	}
 
@@ -585,26 +368,22 @@ int BN_MONT_CTX_set(BN_MONT_CTX *mont, const BIGNUM *mod, BN_CTX *ctx)
 		BIGNUM tmod;
 		BN_ULONG buf[2];
 
-		mont->ri=(BN_num_bits(mod)+(BN_BITS2-1))/BN_BITS2*BN_BITS2;
-		BN_zero(R);
-#if 0 /* for OpenSSL 0.9.9 mont->n0, would be "#if defined(OPENSSL_BN_ASM_MONT) && (BN_BITS2<=32)",
-         only certain BN_BITS2<=32 platforms actually need this */
-		if (!(BN_set_bit(R,2*BN_BITS2))) goto err;	/* R */
-#else
-		if (!(BN_set_bit(R,BN_BITS2))) goto err;	/* R */
-#endif
-
-		buf[0]=mod->d[0]; /* tmod = N mod word size */
-		buf[1]=0;
-
 		BN_init(&tmod);
 		tmod.d=buf;
-		tmod.top = buf[0] != 0 ? 1 : 0;
 		tmod.dmax=2;
 		tmod.neg=0;
 
-#if 0 /* for OpenSSL 0.9.9 mont->n0, would be "#if defined(OPENSSL_BN_ASM_MONT) && (BN_BITS2<=32)";
-         only certain BN_BITS2<=32 platforms actually need this */
+		mont->ri=(BN_num_bits(mod)+(BN_BITS2-1))/BN_BITS2*BN_BITS2;
+
+#if defined(OPENSSL_BN_ASM_MONT) && (BN_BITS2<=32)
+		/* Only certain BN_BITS2<=32 platforms actually make use of
+		 * n0[1], and we could use the #else case (with a shorter R
+		 * value) for the others.  However, currently only the assembler
+		 * files do know which is which. */
+
+		BN_zero(R);
+		if (!(BN_set_bit(R,2*BN_BITS2))) goto err;
+
 								tmod.top=0;
 		if ((buf[0] = mod->d[0]))			tmod.top=1;
 		if ((buf[1] = mod->top>1 ? mod->d[1] : 0))	tmod.top=2;
@@ -632,6 +411,12 @@ int BN_MONT_CTX_set(BN_MONT_CTX *mont, const BIGNUM *mod, BN_CTX *ctx)
 		mont->n0[0] = (Ri->top > 0) ? Ri->d[0] : 0;
 		mont->n0[1] = (Ri->top > 1) ? Ri->d[1] : 0;
 #else
+		BN_zero(R);
+		if (!(BN_set_bit(R,BN_BITS2))) goto err;	/* R */
+
+		buf[0]=mod->d[0]; /* tmod = N mod word size */
+		buf[1]=0;
+		tmod.top = buf[0] != 0 ? 1 : 0;
 							/* Ri = R^-1 mod N*/
 		if ((BN_mod_inverse(Ri,R,&tmod,ctx)) == NULL)
 			goto err;
@@ -647,12 +432,8 @@ int BN_MONT_CTX_set(BN_MONT_CTX *mont, const BIGNUM *mod, BN_CTX *ctx)
 		if (!BN_div(Ri,NULL,Ri,&tmod,ctx)) goto err;
 		/* Ni = (R*Ri-1)/N,
 		 * keep only least significant word: */
-# if 0 /* for OpenSSL 0.9.9 mont->n0 */
 		mont->n0[0] = (Ri->top > 0) ? Ri->d[0] : 0;
 		mont->n0[1] = 0;
-# else
-		mont->n0 = (Ri->top > 0) ? Ri->d[0] : 0;
-# endif
 #endif
 		}
 #else /* !MONT_WORD */
@@ -689,12 +470,8 @@ BN_MONT_CTX *BN_MONT_CTX_copy(BN_MONT_CTX *to, BN_MONT_CTX *from)
 	if (!BN_copy(&(to->N),&(from->N))) return NULL;
 	if (!BN_copy(&(to->Ni),&(from->Ni))) return NULL;
 	to->ri=from->ri;
-#if 0 /* for OpenSSL 0.9.9 mont->n0 */
 	to->n0[0]=from->n0[0];
 	to->n0[1]=from->n0[1];
-#else
-	to->n0=from->n0;
-#endif
 	return(to);
 	}
 
diff --git a/crypto/bn/bn_nist.c b/crypto/bn/bn_nist.c
index 2ca5b0139111..43caee477031 100644
--- a/crypto/bn/bn_nist.c
+++ b/crypto/bn/bn_nist.c
@@ -319,6 +319,13 @@ static void nist_cp_bn(BN_ULONG *buf, BN_ULONG *a, int top)
 						:(to[(n)/2] =((m)&1)?(from[(m)/2]>>32):(from[(m)/2]&BN_MASK2l)))
 #define bn_32_set_0(to, n)		(((n)&1)?(to[(n)/2]&=BN_MASK2l):(to[(n)/2]=0));
 #define bn_cp_32(to,n,from,m)		((m)>=0)?bn_cp_32_naked(to,n,from,m):bn_32_set_0(to,n)
+# if defined(L_ENDIAN)
+#  if defined(__arch64__)
+#   define NIST_INT64 long
+#  else
+#   define NIST_INT64 long long
+#  endif
+# endif
 #else
 #define bn_cp_64(to, n, from, m) \
 	{ \
@@ -330,13 +337,15 @@ static void nist_cp_bn(BN_ULONG *buf, BN_ULONG *a, int top)
 	bn_32_set_0(to, (n)*2); \
 	bn_32_set_0(to, (n)*2+1); \
 	}
-#if BN_BITS2 == 32
 #define bn_cp_32(to, n, from, m)	(to)[n] = (m>=0)?((from)[m]):0;
 #define bn_32_set_0(to, n)		(to)[n] = (BN_ULONG)0;
-#endif
+# if defined(_WIN32) && !defined(__GNUC__)
+#  define NIST_INT64 __int64
+# elif defined(BN_LLONG)
+#  define NIST_INT64 long long
+# endif
 #endif /* BN_BITS2 != 64 */
 
-
 #define nist_set_192(to, from, a1, a2, a3) \
 	{ \
 	bn_cp_64(to, 0, from, (a3) - 3) \
@@ -350,11 +359,13 @@ int BN_nist_mod_192(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	int      top = a->top, i;
 	int      carry;
 	register BN_ULONG *r_d, *a_d = a->d;
-	BN_ULONG t_d[BN_NIST_192_TOP],
-	         buf[BN_NIST_192_TOP],
-		 c_d[BN_NIST_192_TOP],
+	union	{
+		BN_ULONG	bn[BN_NIST_192_TOP];
+		unsigned int	ui[BN_NIST_192_TOP*sizeof(BN_ULONG)/sizeof(unsigned int)];
+		} buf;
+	BN_ULONG c_d[BN_NIST_192_TOP],
 		*res;
-	size_t   mask;
+	PTR_SIZE_INT mask;
 	static const BIGNUM _bignum_nist_p_192_sqr = {
 		(BN_ULONG *)_nist_p_192_sqr,
 		sizeof(_nist_p_192_sqr)/sizeof(_nist_p_192_sqr[0]),
@@ -385,15 +396,48 @@ int BN_nist_mod_192(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	else
 		r_d = a_d;
 
-	nist_cp_bn_0(buf, a_d + BN_NIST_192_TOP, top - BN_NIST_192_TOP, BN_NIST_192_TOP);
+	nist_cp_bn_0(buf.bn, a_d + BN_NIST_192_TOP, top - BN_NIST_192_TOP, BN_NIST_192_TOP);
+
+#if defined(NIST_INT64)
+	{
+	NIST_INT64		acc;	/* accumulator */
+	unsigned int		*rp=(unsigned int *)r_d;
+	const unsigned int	*bp=(const unsigned int *)buf.ui;
+
+	acc  = rp[0];	acc += bp[3*2-6];
+			acc += bp[5*2-6]; rp[0] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[1];	acc += bp[3*2-5];
+			acc += bp[5*2-5]; rp[1] = (unsigned int)acc; acc >>= 32;
 
-	nist_set_192(t_d, buf, 0, 3, 3);
+	acc += rp[2];	acc += bp[3*2-6];
+			acc += bp[4*2-6];
+			acc += bp[5*2-6]; rp[2] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[3];	acc += bp[3*2-5];
+			acc += bp[4*2-5];
+			acc += bp[5*2-5]; rp[3] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[4];	acc += bp[4*2-6];
+			acc += bp[5*2-6]; rp[4] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[5];	acc += bp[4*2-5];
+			acc += bp[5*2-5]; rp[5] = (unsigned int)acc;
+
+	carry = (int)(acc>>32);
+	}
+#else
+	{
+	BN_ULONG t_d[BN_NIST_192_TOP];
+
+	nist_set_192(t_d, buf.bn, 0, 3, 3);
 	carry = (int)bn_add_words(r_d, r_d, t_d, BN_NIST_192_TOP);
-	nist_set_192(t_d, buf, 4, 4, 0);
+	nist_set_192(t_d, buf.bn, 4, 4, 0);
 	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_192_TOP);
-	nist_set_192(t_d, buf, 5, 5, 5)
+	nist_set_192(t_d, buf.bn, 5, 5, 5)
 	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_192_TOP);
-
+	}
+#endif
 	if (carry > 0)
 		carry = (int)bn_sub_words(r_d,r_d,_nist_p_192[carry-1],BN_NIST_192_TOP);
 	else
@@ -405,9 +449,10 @@ int BN_nist_mod_192(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	 * 'tmp=result-modulus; if (!carry || !borrow) result=tmp;'
 	 * this is what happens below, but without explicit if:-) a.
 	 */
-	mask  = 0-(size_t)bn_sub_words(c_d,r_d,_nist_p_192[0],BN_NIST_192_TOP);
-	mask &= 0-(size_t)carry;
-	res   = (BN_ULONG *)(((size_t)c_d&~mask) | ((size_t)r_d&mask));
+	mask  = 0-(PTR_SIZE_INT)bn_sub_words(c_d,r_d,_nist_p_192[0],BN_NIST_192_TOP);
+	mask &= 0-(PTR_SIZE_INT)carry;
+	res   = (BN_ULONG *)
+	 (((PTR_SIZE_INT)c_d&~mask) | ((PTR_SIZE_INT)r_d&mask));
 	nist_cp_bn(r_d, res, BN_NIST_192_TOP);
 	r->top = BN_NIST_192_TOP;
 	bn_correct_top(r);
@@ -434,12 +479,11 @@ int BN_nist_mod_224(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	int	top = a->top, i;
 	int	carry;
 	BN_ULONG *r_d, *a_d = a->d;
-	BN_ULONG t_d[BN_NIST_224_TOP],
-	         buf[BN_NIST_224_TOP],
+	BN_ULONG buf[BN_NIST_224_TOP],
 		 c_d[BN_NIST_224_TOP],
 		*res;
-	size_t   mask;
-	union { bn_addsub_f f; size_t p; } u;
+	PTR_SIZE_INT mask;
+	union { bn_addsub_f f; PTR_SIZE_INT p; } u;
 	static const BIGNUM _bignum_nist_p_224_sqr = {
 		(BN_ULONG *)_nist_p_224_sqr,
 		sizeof(_nist_p_224_sqr)/sizeof(_nist_p_224_sqr[0]),
@@ -473,14 +517,54 @@ int BN_nist_mod_224(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 
 #if BN_BITS2==64
 	/* copy upper 256 bits of 448 bit number ... */
-	nist_cp_bn_0(t_d, a_d + (BN_NIST_224_TOP-1), top - (BN_NIST_224_TOP-1), BN_NIST_224_TOP);
+	nist_cp_bn_0(c_d, a_d + (BN_NIST_224_TOP-1), top - (BN_NIST_224_TOP-1), BN_NIST_224_TOP);
 	/* ... and right shift by 32 to obtain upper 224 bits */
-	nist_set_224(buf, t_d, 14, 13, 12, 11, 10, 9, 8);
+	nist_set_224(buf, c_d, 14, 13, 12, 11, 10, 9, 8);
 	/* truncate lower part to 224 bits too */
 	r_d[BN_NIST_224_TOP-1] &= BN_MASK2l;
 #else
 	nist_cp_bn_0(buf, a_d + BN_NIST_224_TOP, top - BN_NIST_224_TOP, BN_NIST_224_TOP);
 #endif
+
+#if defined(NIST_INT64) && BN_BITS2!=64
+	{
+	NIST_INT64		acc;	/* accumulator */
+	unsigned int		*rp=(unsigned int *)r_d;
+	const unsigned int	*bp=(const unsigned int *)buf;
+
+	acc  = rp[0];	acc -= bp[7-7];
+			acc -= bp[11-7]; rp[0] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[1];	acc -= bp[8-7];
+			acc -= bp[12-7]; rp[1] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[2];	acc -= bp[9-7];
+			acc -= bp[13-7]; rp[2] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[3];	acc += bp[7-7];
+			acc += bp[11-7];
+			acc -= bp[10-7]; rp[3] = (unsigned int)acc; acc>>= 32;
+
+	acc += rp[4];	acc += bp[8-7];
+			acc += bp[12-7];
+			acc -= bp[11-7]; rp[4] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[5];	acc += bp[9-7];
+			acc += bp[13-7];
+			acc -= bp[12-7]; rp[5] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[6];	acc += bp[10-7];
+			acc -= bp[13-7]; rp[6] = (unsigned int)acc;
+
+	carry = (int)(acc>>32);
+# if BN_BITS2==64
+	rp[7] = carry;
+# endif
+	}	
+#else
+	{
+	BN_ULONG t_d[BN_NIST_224_TOP];
+
 	nist_set_224(t_d, buf, 10, 9, 8, 7, 0, 0, 0);
 	carry = (int)bn_add_words(r_d, r_d, t_d, BN_NIST_224_TOP);
 	nist_set_224(t_d, buf, 0, 13, 12, 11, 0, 0, 0);
@@ -493,6 +577,8 @@ int BN_nist_mod_224(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 #if BN_BITS2==64
 	carry = (int)(r_d[BN_NIST_224_TOP-1]>>32);
 #endif
+	}
+#endif
 	u.f = bn_sub_words;
 	if (carry > 0)
 		{
@@ -510,16 +596,18 @@ int BN_nist_mod_224(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 		 * to be compared to the modulus and conditionally
 		 * adjusted by *subtracting* the latter. */
 		carry = (int)bn_add_words(r_d,r_d,_nist_p_224[-carry-1],BN_NIST_224_TOP);
-		mask = 0-(size_t)carry;
-		u.p = ((size_t)bn_sub_words&mask) | ((size_t)bn_add_words&~mask);
+		mask = 0-(PTR_SIZE_INT)carry;
+		u.p = ((PTR_SIZE_INT)bn_sub_words&mask) |
+		 ((PTR_SIZE_INT)bn_add_words&~mask);
 		}
 	else
 		carry = 1;
 
 	/* otherwise it's effectively same as in BN_nist_mod_192... */
-	mask  = 0-(size_t)(*u.f)(c_d,r_d,_nist_p_224[0],BN_NIST_224_TOP);
-	mask &= 0-(size_t)carry;
-	res   = (BN_ULONG *)(((size_t)c_d&~mask) | ((size_t)r_d&mask));
+	mask  = 0-(PTR_SIZE_INT)(*u.f)(c_d,r_d,_nist_p_224[0],BN_NIST_224_TOP);
+	mask &= 0-(PTR_SIZE_INT)carry;
+	res   = (BN_ULONG *)(((PTR_SIZE_INT)c_d&~mask) |
+	 ((PTR_SIZE_INT)r_d&mask));
 	nist_cp_bn(r_d, res, BN_NIST_224_TOP);
 	r->top = BN_NIST_224_TOP;
 	bn_correct_top(r);
@@ -545,12 +633,14 @@ int BN_nist_mod_256(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	int	i, top = a->top;
 	int	carry = 0;
 	register BN_ULONG *a_d = a->d, *r_d;
-	BN_ULONG t_d[BN_NIST_256_TOP],
-	         buf[BN_NIST_256_TOP],
-		 c_d[BN_NIST_256_TOP],
+	union	{
+		BN_ULONG bn[BN_NIST_256_TOP];
+		unsigned int ui[BN_NIST_256_TOP*sizeof(BN_ULONG)/sizeof(unsigned int)];
+		} buf;
+	BN_ULONG c_d[BN_NIST_256_TOP],
 		*res;
-	size_t   mask;
-	union { bn_addsub_f f; size_t p; } u;
+	PTR_SIZE_INT mask;
+	union { bn_addsub_f f; PTR_SIZE_INT p; } u;
 	static const BIGNUM _bignum_nist_p_256_sqr = {
 		(BN_ULONG *)_nist_p_256_sqr,
 		sizeof(_nist_p_256_sqr)/sizeof(_nist_p_256_sqr[0]),
@@ -581,12 +671,87 @@ int BN_nist_mod_256(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	else
 		r_d = a_d;
 
-	nist_cp_bn_0(buf, a_d + BN_NIST_256_TOP, top - BN_NIST_256_TOP, BN_NIST_256_TOP);
+	nist_cp_bn_0(buf.bn, a_d + BN_NIST_256_TOP, top - BN_NIST_256_TOP, BN_NIST_256_TOP);
+
+#if defined(NIST_INT64)
+	{
+	NIST_INT64		acc;	/* accumulator */
+	unsigned int		*rp=(unsigned int *)r_d;
+	const unsigned int	*bp=(const unsigned int *)buf.ui;
+
+	acc = rp[0];	acc += bp[8-8];
+			acc += bp[9-8];
+			acc -= bp[11-8];
+			acc -= bp[12-8];
+			acc -= bp[13-8];
+			acc -= bp[14-8]; rp[0] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[1];	acc += bp[9-8];
+			acc += bp[10-8];
+			acc -= bp[12-8];
+			acc -= bp[13-8];
+			acc -= bp[14-8];
+			acc -= bp[15-8]; rp[1] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[2];	acc += bp[10-8];
+			acc += bp[11-8];
+			acc -= bp[13-8];
+			acc -= bp[14-8];
+			acc -= bp[15-8]; rp[2] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[3];	acc += bp[11-8];
+			acc += bp[11-8];
+			acc += bp[12-8];
+			acc += bp[12-8];
+			acc += bp[13-8];
+			acc -= bp[15-8];
+			acc -= bp[8-8];
+			acc -= bp[9-8];  rp[3] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[4];	acc += bp[12-8];
+			acc += bp[12-8];
+			acc += bp[13-8];
+			acc += bp[13-8];
+			acc += bp[14-8];
+			acc -= bp[9-8];
+			acc -= bp[10-8]; rp[4] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[5];	acc += bp[13-8];
+			acc += bp[13-8];
+			acc += bp[14-8];
+			acc += bp[14-8];
+			acc += bp[15-8];
+			acc -= bp[10-8];
+			acc -= bp[11-8]; rp[5] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[6];	acc += bp[14-8];
+			acc += bp[14-8];
+			acc += bp[15-8];
+			acc += bp[15-8];
+			acc += bp[14-8];
+			acc += bp[13-8];
+			acc -= bp[8-8];
+			acc -= bp[9-8];  rp[6] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[7];	acc += bp[15-8];
+			acc += bp[15-8];
+			acc += bp[15-8];
+			acc += bp[8 -8];
+			acc -= bp[10-8];
+			acc -= bp[11-8];
+			acc -= bp[12-8];
+			acc -= bp[13-8]; rp[7] = (unsigned int)acc;
+
+	carry = (int)(acc>>32);
+	}
+#else
+	{
+	BN_ULONG t_d[BN_NIST_256_TOP];
 
 	/*S1*/
-	nist_set_256(t_d, buf, 15, 14, 13, 12, 11, 0, 0, 0);
+	nist_set_256(t_d, buf.bn, 15, 14, 13, 12, 11, 0, 0, 0);
 	/*S2*/
-	nist_set_256(c_d, buf, 0, 15, 14, 13, 12, 0, 0, 0);
+	nist_set_256(c_d, buf.bn, 0, 15, 14, 13, 12, 0, 0, 0);
 	carry = (int)bn_add_words(t_d, t_d, c_d, BN_NIST_256_TOP);
 	/* left shift */
 		{
@@ -604,24 +769,26 @@ int BN_nist_mod_256(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 		}
 	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_256_TOP);
 	/*S3*/
-	nist_set_256(t_d, buf, 15, 14, 0, 0, 0, 10, 9, 8);
+	nist_set_256(t_d, buf.bn, 15, 14, 0, 0, 0, 10, 9, 8);
 	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_256_TOP);
 	/*S4*/
-	nist_set_256(t_d, buf, 8, 13, 15, 14, 13, 11, 10, 9);
+	nist_set_256(t_d, buf.bn, 8, 13, 15, 14, 13, 11, 10, 9);
 	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_256_TOP);
 	/*D1*/
-	nist_set_256(t_d, buf, 10, 8, 0, 0, 0, 13, 12, 11);
+	nist_set_256(t_d, buf.bn, 10, 8, 0, 0, 0, 13, 12, 11);
 	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_256_TOP);
 	/*D2*/
-	nist_set_256(t_d, buf, 11, 9, 0, 0, 15, 14, 13, 12);
+	nist_set_256(t_d, buf.bn, 11, 9, 0, 0, 15, 14, 13, 12);
 	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_256_TOP);
 	/*D3*/
-	nist_set_256(t_d, buf, 12, 0, 10, 9, 8, 15, 14, 13);
+	nist_set_256(t_d, buf.bn, 12, 0, 10, 9, 8, 15, 14, 13);
 	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_256_TOP);
 	/*D4*/
-	nist_set_256(t_d, buf, 13, 0, 11, 10, 9, 0, 15, 14);
+	nist_set_256(t_d, buf.bn, 13, 0, 11, 10, 9, 0, 15, 14);
 	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_256_TOP);
 
+	}
+#endif
 	/* see BN_nist_mod_224 for explanation */
 	u.f = bn_sub_words;
 	if (carry > 0)
@@ -629,15 +796,17 @@ int BN_nist_mod_256(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	else if (carry < 0)
 		{
 		carry = (int)bn_add_words(r_d,r_d,_nist_p_256[-carry-1],BN_NIST_256_TOP);
-		mask = 0-(size_t)carry;
-		u.p = ((size_t)bn_sub_words&mask) | ((size_t)bn_add_words&~mask);
+		mask = 0-(PTR_SIZE_INT)carry;
+		u.p = ((PTR_SIZE_INT)bn_sub_words&mask) |
+		 ((PTR_SIZE_INT)bn_add_words&~mask);
 		}
 	else
 		carry = 1;
 
-	mask  = 0-(size_t)(*u.f)(c_d,r_d,_nist_p_256[0],BN_NIST_256_TOP);
-	mask &= 0-(size_t)carry;
-	res   = (BN_ULONG *)(((size_t)c_d&~mask) | ((size_t)r_d&mask));
+	mask  = 0-(PTR_SIZE_INT)(*u.f)(c_d,r_d,_nist_p_256[0],BN_NIST_256_TOP);
+	mask &= 0-(PTR_SIZE_INT)carry;
+	res   = (BN_ULONG *)(((PTR_SIZE_INT)c_d&~mask) |
+	 ((PTR_SIZE_INT)r_d&mask));
 	nist_cp_bn(r_d, res, BN_NIST_256_TOP);
 	r->top = BN_NIST_256_TOP;
 	bn_correct_top(r);
@@ -667,12 +836,14 @@ int BN_nist_mod_384(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	int	i, top = a->top;
 	int	carry = 0;
 	register BN_ULONG *r_d, *a_d = a->d;
-	BN_ULONG t_d[BN_NIST_384_TOP],
-	         buf[BN_NIST_384_TOP],
-		 c_d[BN_NIST_384_TOP],
+	union	{
+		BN_ULONG bn[BN_NIST_384_TOP];
+		unsigned int ui[BN_NIST_384_TOP*sizeof(BN_ULONG)/sizeof(unsigned int)];
+		} buf;
+	BN_ULONG c_d[BN_NIST_384_TOP],
 		*res;
-	size_t	 mask;
-	union { bn_addsub_f f; size_t p; } u;
+	PTR_SIZE_INT mask;
+	union { bn_addsub_f f; PTR_SIZE_INT p; } u;
 	static const BIGNUM _bignum_nist_p_384_sqr = {
 		(BN_ULONG *)_nist_p_384_sqr,
 		sizeof(_nist_p_384_sqr)/sizeof(_nist_p_384_sqr[0]),
@@ -704,10 +875,100 @@ int BN_nist_mod_384(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	else
 		r_d = a_d;
 
-	nist_cp_bn_0(buf, a_d + BN_NIST_384_TOP, top - BN_NIST_384_TOP, BN_NIST_384_TOP);
+	nist_cp_bn_0(buf.bn, a_d + BN_NIST_384_TOP, top - BN_NIST_384_TOP, BN_NIST_384_TOP);
+
+#if defined(NIST_INT64)
+	{
+	NIST_INT64		acc;	/* accumulator */
+	unsigned int		*rp=(unsigned int *)r_d;
+	const unsigned int	*bp=(const unsigned int *)buf.ui;
+
+	acc = rp[0];	acc += bp[12-12];
+			acc += bp[21-12];
+			acc += bp[20-12];
+			acc -= bp[23-12]; rp[0] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[1];	acc += bp[13-12];
+			acc += bp[22-12];
+			acc += bp[23-12];
+			acc -= bp[12-12];
+			acc -= bp[20-12]; rp[1] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[2];	acc += bp[14-12];
+			acc += bp[23-12];
+			acc -= bp[13-12];
+			acc -= bp[21-12]; rp[2] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[3];	acc += bp[15-12];
+			acc += bp[12-12];
+			acc += bp[20-12];
+			acc += bp[21-12];
+			acc -= bp[14-12];
+			acc -= bp[22-12];
+			acc -= bp[23-12]; rp[3] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[4];	acc += bp[21-12];
+			acc += bp[21-12];
+			acc += bp[16-12];
+			acc += bp[13-12];
+			acc += bp[12-12];
+			acc += bp[20-12];
+			acc += bp[22-12];
+			acc -= bp[15-12];
+			acc -= bp[23-12];
+			acc -= bp[23-12]; rp[4] = (unsigned int)acc; acc >>= 32;
+
+	acc += rp[5];	acc += bp[22-12];
+			acc += bp[22-12];
+			acc += bp[17-12];
+			acc += bp[14-12];
+			acc += bp[13-12];
+			acc += bp[21-12];
+			acc += bp[23-12];
+			acc -= bp[16-12]; rp[5] = (unsigned int)acc; acc >>= 32;
+			
+	acc += rp[6];	acc += bp[23-12];
+			acc += bp[23-12];
+			acc += bp[18-12];
+			acc += bp[15-12];
+			acc += bp[14-12];
+			acc += bp[22-12];
+			acc -= bp[17-12]; rp[6] = (unsigned int)acc; acc >>= 32;
+			
+	acc += rp[7];	acc += bp[19-12];
+			acc += bp[16-12];
+			acc += bp[15-12];
+			acc += bp[23-12];
+			acc -= bp[18-12]; rp[7] = (unsigned int)acc; acc >>= 32;
+			
+	acc += rp[8];	acc += bp[20-12];
+			acc += bp[17-12];
+			acc += bp[16-12];
+			acc -= bp[19-12]; rp[8] = (unsigned int)acc; acc >>= 32;
+			
+	acc += rp[9];	acc += bp[21-12];
+			acc += bp[18-12];
+			acc += bp[17-12];
+			acc -= bp[20-12]; rp[9] = (unsigned int)acc; acc >>= 32;
+			
+	acc += rp[10];	acc += bp[22-12];
+			acc += bp[19-12];
+			acc += bp[18-12];
+			acc -= bp[21-12]; rp[10] = (unsigned int)acc; acc >>= 32;
+			
+	acc += rp[11];	acc += bp[23-12];
+			acc += bp[20-12];
+			acc += bp[19-12];
+			acc -= bp[22-12]; rp[11] = (unsigned int)acc;
+
+	carry = (int)(acc>>32);
+	}
+#else
+	{
+	BN_ULONG t_d[BN_NIST_384_TOP];
 
 	/*S1*/
-	nist_set_256(t_d, buf, 0, 0, 0, 0, 0, 23-4, 22-4, 21-4);
+	nist_set_256(t_d, buf.bn, 0, 0, 0, 0, 0, 23-4, 22-4, 21-4);
 		/* left shift */
 		{
 		register BN_ULONG *ap,t,c;
@@ -724,29 +985,31 @@ int BN_nist_mod_384(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	carry = (int)bn_add_words(r_d+(128/BN_BITS2), r_d+(128/BN_BITS2), 
 		t_d, BN_NIST_256_TOP);
 	/*S2 */
-	carry += (int)bn_add_words(r_d, r_d, buf, BN_NIST_384_TOP);
+	carry += (int)bn_add_words(r_d, r_d, buf.bn, BN_NIST_384_TOP);
 	/*S3*/
-	nist_set_384(t_d,buf,20,19,18,17,16,15,14,13,12,23,22,21);
+	nist_set_384(t_d,buf.bn,20,19,18,17,16,15,14,13,12,23,22,21);
 	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_384_TOP);
 	/*S4*/
-	nist_set_384(t_d,buf,19,18,17,16,15,14,13,12,20,0,23,0);
+	nist_set_384(t_d,buf.bn,19,18,17,16,15,14,13,12,20,0,23,0);
 	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_384_TOP);
 	/*S5*/
-	nist_set_384(t_d, buf,0,0,0,0,23,22,21,20,0,0,0,0);
+	nist_set_384(t_d, buf.bn,0,0,0,0,23,22,21,20,0,0,0,0);
 	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_384_TOP);
 	/*S6*/
-	nist_set_384(t_d,buf,0,0,0,0,0,0,23,22,21,0,0,20);
+	nist_set_384(t_d,buf.bn,0,0,0,0,0,0,23,22,21,0,0,20);
 	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_384_TOP);
 	/*D1*/
-	nist_set_384(t_d,buf,22,21,20,19,18,17,16,15,14,13,12,23);
+	nist_set_384(t_d,buf.bn,22,21,20,19,18,17,16,15,14,13,12,23);
 	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_384_TOP);
 	/*D2*/
-	nist_set_384(t_d,buf,0,0,0,0,0,0,0,23,22,21,20,0);
+	nist_set_384(t_d,buf.bn,0,0,0,0,0,0,0,23,22,21,20,0);
 	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_384_TOP);
 	/*D3*/
-	nist_set_384(t_d,buf,0,0,0,0,0,0,0,23,23,0,0,0);
+	nist_set_384(t_d,buf.bn,0,0,0,0,0,0,0,23,23,0,0,0);
 	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_384_TOP);
 
+	}
+#endif
 	/* see BN_nist_mod_224 for explanation */
 	u.f = bn_sub_words;
 	if (carry > 0)
@@ -754,15 +1017,17 @@ int BN_nist_mod_384(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	else if (carry < 0)
 		{
 		carry = (int)bn_add_words(r_d,r_d,_nist_p_384[-carry-1],BN_NIST_384_TOP);
-		mask = 0-(size_t)carry;
-		u.p = ((size_t)bn_sub_words&mask) | ((size_t)bn_add_words&~mask);
+		mask = 0-(PTR_SIZE_INT)carry;
+		u.p = ((PTR_SIZE_INT)bn_sub_words&mask) |
+		 ((PTR_SIZE_INT)bn_add_words&~mask);
 		}
 	else
 		carry = 1;
 
-	mask  = 0-(size_t)(*u.f)(c_d,r_d,_nist_p_384[0],BN_NIST_384_TOP);
-	mask &= 0-(size_t)carry;
-	res   = (BN_ULONG *)(((size_t)c_d&~mask) | ((size_t)r_d&mask));
+	mask  = 0-(PTR_SIZE_INT)(*u.f)(c_d,r_d,_nist_p_384[0],BN_NIST_384_TOP);
+	mask &= 0-(PTR_SIZE_INT)carry;
+	res   = (BN_ULONG *)(((PTR_SIZE_INT)c_d&~mask) |
+	 ((PTR_SIZE_INT)r_d&mask));
 	nist_cp_bn(r_d, res, BN_NIST_384_TOP);
 	r->top = BN_NIST_384_TOP;
 	bn_correct_top(r);
@@ -781,7 +1046,7 @@ int BN_nist_mod_521(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	BN_ULONG *r_d, *a_d = a->d,
 		 t_d[BN_NIST_521_TOP],
 		 val,tmp,*res;
-	size_t	mask;
+	PTR_SIZE_INT mask;
 	static const BIGNUM _bignum_nist_p_521_sqr = {
 		(BN_ULONG *)_nist_p_521_sqr,
 		sizeof(_nist_p_521_sqr)/sizeof(_nist_p_521_sqr[0]),
@@ -826,8 +1091,9 @@ int BN_nist_mod_521(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
 	r_d[i] &= BN_NIST_521_TOP_MASK;
 
 	bn_add_words(r_d,r_d,t_d,BN_NIST_521_TOP);
-	mask = 0-(size_t)bn_sub_words(t_d,r_d,_nist_p_521,BN_NIST_521_TOP);
-	res  = (BN_ULONG *)(((size_t)t_d&~mask) | ((size_t)r_d&mask));
+	mask = 0-(PTR_SIZE_INT)bn_sub_words(t_d,r_d,_nist_p_521,BN_NIST_521_TOP);
+	res  = (BN_ULONG *)(((PTR_SIZE_INT)t_d&~mask) |
+	 ((PTR_SIZE_INT)r_d&mask));
 	nist_cp_bn(r_d,res,BN_NIST_521_TOP);
 	r->top = BN_NIST_521_TOP;
 	bn_correct_top(r);
diff --git a/crypto/bn/bn_opt.c b/crypto/bn/bn_opt.c
deleted file mode 100644
index 21cbb38f622b..000000000000
--- a/crypto/bn/bn_opt.c
+++ /dev/null
@@ -1,87 +0,0 @@
-/* crypto/bn/bn_opt.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#ifndef BN_DEBUG
-# undef NDEBUG /* avoid conflicting definitions */
-# define NDEBUG
-#endif
-
-#include <assert.h>
-#include <limits.h>
-#include <stdio.h>
-#include "cryptlib.h"
-#include "bn_lcl.h"
-
-char *BN_options(void)
-	{
-	static int init=0;
-	static char data[16];
-
-	if (!init)
-		{
-		init++;
-#ifdef BN_LLONG
-		BIO_snprintf(data,sizeof data,"bn(%d,%d)",
-			     (int)sizeof(BN_ULLONG)*8,(int)sizeof(BN_ULONG)*8);
-#else
-		BIO_snprintf(data,sizeof data,"bn(%d,%d)",
-			     (int)sizeof(BN_ULONG)*8,(int)sizeof(BN_ULONG)*8);
-#endif
-		}
-	return(data);
-	}
diff --git a/crypto/bn/bn_print.c b/crypto/bn/bn_print.c
index 810dde34e15c..1743b6a7e212 100644
--- a/crypto/bn/bn_print.c
+++ b/crypto/bn/bn_print.c
@@ -294,6 +294,27 @@ err:
 	return(0);
 	}
 
+int BN_asc2bn(BIGNUM **bn, const char *a)
+	{
+	const char *p = a;
+	if (*p == '-')
+		p++;
+
+	if (p[0] == '0' && (p[1] == 'X' || p[1] == 'x'))
+		{		
+		if (!BN_hex2bn(bn, p + 2))
+			return 0;
+		}
+	else
+		{
+		if (!BN_dec2bn(bn, p))
+			return 0;
+		}
+	if (*a == '-')
+		(*bn)->neg = 1;
+	return 1;
+	}
+
 #ifndef OPENSSL_NO_BIO
 #ifndef OPENSSL_NO_FP_API
 int BN_print_fp(FILE *fp, const BIGNUM *a)
@@ -336,3 +357,22 @@ end:
 	return(ret);
 	}
 #endif
+
+char *BN_options(void)
+	{
+	static int init=0;
+	static char data[16];
+
+	if (!init)
+		{
+		init++;
+#ifdef BN_LLONG
+		BIO_snprintf(data,sizeof data,"bn(%d,%d)",
+			     (int)sizeof(BN_ULLONG)*8,(int)sizeof(BN_ULONG)*8);
+#else
+		BIO_snprintf(data,sizeof data,"bn(%d,%d)",
+			     (int)sizeof(BN_ULONG)*8,(int)sizeof(BN_ULONG)*8);
+#endif
+		}
+	return(data);
+	}
diff --git a/crypto/bn/bn_shift.c b/crypto/bn/bn_shift.c
index c4d301afc467..a6fca2c424f0 100644
--- a/crypto/bn/bn_shift.c
+++ b/crypto/bn/bn_shift.c
@@ -99,7 +99,7 @@ int BN_lshift1(BIGNUM *r, const BIGNUM *a)
 int BN_rshift1(BIGNUM *r, const BIGNUM *a)
 	{
 	BN_ULONG *ap,*rp,t,c;
-	int i;
+	int i,j;
 
 	bn_check_top(r);
 	bn_check_top(a);
@@ -109,22 +109,25 @@ int BN_rshift1(BIGNUM *r, const BIGNUM *a)
 		BN_zero(r);
 		return(1);
 		}
+	i = a->top;
+	ap= a->d;
+	j = i-(ap[i-1]==1);
 	if (a != r)
 		{
-		if (bn_wexpand(r,a->top) == NULL) return(0);
-		r->top=a->top;
+		if (bn_wexpand(r,j) == NULL) return(0);
 		r->neg=a->neg;
 		}
-	ap=a->d;
 	rp=r->d;
-	c=0;
-	for (i=a->top-1; i>=0; i--)
+	t=ap[--i];
+	c=(t&1)?BN_TBIT:0;
+	if (t>>=1) rp[i]=t;
+	while (i>0)
 		{
-		t=ap[i];
+		t=ap[--i];
 		rp[i]=((t>>1)&BN_MASK2)|c;
 		c=(t&1)?BN_TBIT:0;
 		}
-	bn_correct_top(r);
+	r->top=j;
 	bn_check_top(r);
 	return(1);
 	}
@@ -182,10 +185,11 @@ int BN_rshift(BIGNUM *r, const BIGNUM *a, int n)
 		BN_zero(r);
 		return(1);
 		}
+	i = (BN_num_bits(a)-n+(BN_BITS2-1))/BN_BITS2;
 	if (r != a)
 		{
 		r->neg=a->neg;
-		if (bn_wexpand(r,a->top-nw+1) == NULL) return(0);
+		if (bn_wexpand(r,i) == NULL) return(0);
 		}
 	else
 		{
@@ -196,7 +200,7 @@ int BN_rshift(BIGNUM *r, const BIGNUM *a, int n)
 	f= &(a->d[nw]);
 	t=r->d;
 	j=a->top-nw;
-	r->top=j;
+	r->top=i;
 
 	if (rb == 0)
 		{
@@ -212,9 +216,8 @@ int BN_rshift(BIGNUM *r, const BIGNUM *a, int n)
 			l= *(f++);
 			*(t++) =(tmp|(l<<lb))&BN_MASK2;
 			}
-		*(t++) =(l>>rb)&BN_MASK2;
+		if ((l = (l>>rb)&BN_MASK2)) *(t) = l;
 		}
-	bn_correct_top(r);
 	bn_check_top(r);
 	return(1);
 	}
diff --git a/crypto/bn/bntest.c b/crypto/bn/bntest.c
index d41daac5fedb..06f5954acc39 100644
--- a/crypto/bn/bntest.c
+++ b/crypto/bn/bntest.c
@@ -262,7 +262,7 @@ int main(int argc, char *argv[])
 	message(out,"BN_mod_sqrt");
 	if (!test_sqrt(out,ctx)) goto err;
 	(void)BIO_flush(out);
-
+#ifndef OPENSSL_NO_EC2M
 	message(out,"BN_GF2m_add");
 	if (!test_gf2m_add(out)) goto err;
 	(void)BIO_flush(out);
@@ -298,7 +298,7 @@ int main(int argc, char *argv[])
 	message(out,"BN_GF2m_mod_solve_quad");
 	if (!test_gf2m_mod_solve_quad(out,ctx)) goto err;
 	(void)BIO_flush(out);
-
+#endif
 	BN_CTX_free(ctx);
 	BIO_free(out);
 
@@ -486,7 +486,7 @@ static void print_word(BIO *bp,BN_ULONG w)
 		return;
 		}
 #endif
-	BIO_printf(bp,"%lX",w);
+	BIO_printf(bp,BN_HEX_FMT1,w);
 	}
 
 int test_div_word(BIO *bp)
@@ -732,6 +732,8 @@ int test_mont(BIO *bp, BN_CTX *ctx)
 	BN_init(&n);
 
 	mont=BN_MONT_CTX_new();
+	if (mont == NULL)
+		return 0;
 
 	BN_bntest_rand(&a,100,0,0); /**/
 	BN_bntest_rand(&b,100,0,0); /**/
@@ -1059,7 +1061,7 @@ int test_exp(BIO *bp, BN_CTX *ctx)
 	BN_free(one);
 	return(1);
 	}
-
+#ifndef OPENSSL_NO_EC2M
 int test_gf2m_add(BIO *bp)
 	{
 	BIGNUM a,b,c;
@@ -1116,8 +1118,8 @@ int test_gf2m_mod(BIO *bp)
 	{
 	BIGNUM *a,*b[2],*c,*d,*e;
 	int i, j, ret = 0;
-	unsigned int p0[] = {163,7,6,3,0};
-	unsigned int p1[] = {193,15,0};
+	int p0[] = {163,7,6,3,0,-1};
+	int p1[] = {193,15,0,-1};
 
 	a=BN_new();
 	b[0]=BN_new();
@@ -1174,8 +1176,8 @@ int test_gf2m_mod_mul(BIO *bp,BN_CTX *ctx)
 	{
 	BIGNUM *a,*b[2],*c,*d,*e,*f,*g,*h;
 	int i, j, ret = 0;
-	unsigned int p0[] = {163,7,6,3,0};
-	unsigned int p1[] = {193,15,0};
+	int p0[] = {163,7,6,3,0,-1};
+	int p1[] = {193,15,0,-1};
 
 	a=BN_new();
 	b[0]=BN_new();
@@ -1245,8 +1247,8 @@ int test_gf2m_mod_sqr(BIO *bp,BN_CTX *ctx)
 	{
 	BIGNUM *a,*b[2],*c,*d;
 	int i, j, ret = 0;
-	unsigned int p0[] = {163,7,6,3,0};
-	unsigned int p1[] = {193,15,0};
+	int p0[] = {163,7,6,3,0,-1};
+	int p1[] = {193,15,0,-1};
 
 	a=BN_new();
 	b[0]=BN_new();
@@ -1304,8 +1306,8 @@ int test_gf2m_mod_inv(BIO *bp,BN_CTX *ctx)
 	{
 	BIGNUM *a,*b[2],*c,*d;
 	int i, j, ret = 0;
-	unsigned int p0[] = {163,7,6,3,0};
-	unsigned int p1[] = {193,15,0};
+	int p0[] = {163,7,6,3,0,-1};
+	int p1[] = {193,15,0,-1};
 
 	a=BN_new();
 	b[0]=BN_new();
@@ -1359,8 +1361,8 @@ int test_gf2m_mod_div(BIO *bp,BN_CTX *ctx)
 	{
 	BIGNUM *a,*b[2],*c,*d,*e,*f;
 	int i, j, ret = 0;
-	unsigned int p0[] = {163,7,6,3,0};
-	unsigned int p1[] = {193,15,0};
+	int p0[] = {163,7,6,3,0,-1};
+	int p1[] = {193,15,0,-1};
 
 	a=BN_new();
 	b[0]=BN_new();
@@ -1422,8 +1424,8 @@ int test_gf2m_mod_exp(BIO *bp,BN_CTX *ctx)
 	{
 	BIGNUM *a,*b[2],*c,*d,*e,*f;
 	int i, j, ret = 0;
-	unsigned int p0[] = {163,7,6,3,0};
-	unsigned int p1[] = {193,15,0};
+	int p0[] = {163,7,6,3,0,-1};
+	int p1[] = {193,15,0,-1};
 
 	a=BN_new();
 	b[0]=BN_new();
@@ -1493,8 +1495,8 @@ int test_gf2m_mod_sqrt(BIO *bp,BN_CTX *ctx)
 	{
 	BIGNUM *a,*b[2],*c,*d,*e,*f;
 	int i, j, ret = 0;
-	unsigned int p0[] = {163,7,6,3,0};
-	unsigned int p1[] = {193,15,0};
+	int p0[] = {163,7,6,3,0,-1};
+	int p1[] = {193,15,0,-1};
 
 	a=BN_new();
 	b[0]=BN_new();
@@ -1552,8 +1554,8 @@ int test_gf2m_mod_solve_quad(BIO *bp,BN_CTX *ctx)
 	{
 	BIGNUM *a,*b[2],*c,*d,*e;
 	int i, j, s = 0, t, ret = 0;
-	unsigned int p0[] = {163,7,6,3,0};
-	unsigned int p1[] = {193,15,0};
+	int p0[] = {163,7,6,3,0,-1};
+	int p1[] = {193,15,0,-1};
 
 	a=BN_new();
 	b[0]=BN_new();
@@ -1634,7 +1636,7 @@ int test_gf2m_mod_solve_quad(BIO *bp,BN_CTX *ctx)
 	BN_free(e);
 	return ret;
 	}
-
+#endif
 static int genprime_cb(int p, int n, BN_GENCB *arg)
 	{
 	char c='*';
diff --git a/crypto/bn/exptest.c b/crypto/bn/exptest.c
index f598a07cf5c9..074a8e882a8e 100644
--- a/crypto/bn/exptest.c
+++ b/crypto/bn/exptest.c
@@ -163,7 +163,7 @@ int main(int argc, char *argv[])
 		  	{
 			if (BN_cmp(r_simple,r_mont) != 0)
 				printf("\nsimple and mont results differ\n");
-			if (BN_cmp(r_simple,r_mont) != 0)
+			if (BN_cmp(r_simple,r_mont_const) != 0)
 				printf("\nsimple and mont const time results differ\n");
 			if (BN_cmp(r_simple,r_recp) != 0)
 				printf("\nsimple and recp results differ\n");
@@ -187,7 +187,7 @@ int main(int argc, char *argv[])
 	BN_free(b);
 	BN_free(m);
 	BN_CTX_free(ctx);
-	ERR_remove_state(0);
+	ERR_remove_thread_state(NULL);
 	CRYPTO_mem_leaks(out);
 	BIO_free(out);
 	printf(" done\n");
diff --git a/crypto/buffer/Makefile b/crypto/buffer/Makefile
index 9e0f46e19ab0..2efba47f0700 100644
--- a/crypto/buffer/Makefile
+++ b/crypto/buffer/Makefile
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
diff --git a/crypto/buffer/buf_err.c b/crypto/buffer/buf_err.c
index 3e25bbe8798a..8f1de6192be4 100644
--- a/crypto/buffer/buf_err.c
+++ b/crypto/buffer/buf_err.c
@@ -1,6 +1,6 @@
 /* crypto/buffer/buf_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
diff --git a/crypto/buffer/buf_str.c b/crypto/buffer/buf_str.c
index 28dd1e401e2a..151f5ea971aa 100644
--- a/crypto/buffer/buf_str.c
+++ b/crypto/buffer/buf_str.c
@@ -1,56 +1,59 @@
-/* crypto/buffer/buf_str.c */
-/* ====================================================================
- * Copyright (c) 2007 The OpenSSL Project.  All rights reserved.
+/* crypto/buffer/buffer.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
  *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
  * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    licensing@OpenSSL.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
  */
 
 #include <stdio.h>
diff --git a/crypto/buffer/buffer.c b/crypto/buffer/buffer.c
index 3b4c79f7048c..d7aa79ad7f0a 100644
--- a/crypto/buffer/buffer.c
+++ b/crypto/buffer/buffer.c
@@ -94,16 +94,11 @@ void BUF_MEM_free(BUF_MEM *a)
 	OPENSSL_free(a);
 	}
 
-int BUF_MEM_grow(BUF_MEM *str, int len)
+int BUF_MEM_grow(BUF_MEM *str, size_t len)
 	{
 	char *ret;
-	unsigned int n;
+	size_t n;
 
-	if (len < 0)
-		{
-		BUFerr(BUF_F_BUF_MEM_GROW,ERR_R_MALLOC_FAILURE);
-		return 0;
-		}
 	if (str->length >= len)
 		{
 		str->length=len;
@@ -141,16 +136,11 @@ int BUF_MEM_grow(BUF_MEM *str, int len)
 	return(len);
 	}
 
-int BUF_MEM_grow_clean(BUF_MEM *str, int len)
+int BUF_MEM_grow_clean(BUF_MEM *str, size_t len)
 	{
 	char *ret;
-	unsigned int n;
+	size_t n;
 
-	if (len < 0)
-		{
-		BUFerr(BUF_F_BUF_MEM_GROW_CLEAN,ERR_R_MALLOC_FAILURE);
-		return 0;
-		}
 	if (str->length >= len)
 		{
 		memset(&str->data[len],0,str->length-len);
@@ -188,3 +178,26 @@ int BUF_MEM_grow_clean(BUF_MEM *str, int len)
 		}
 	return(len);
 	}
+
+void BUF_reverse(unsigned char *out, unsigned char *in, size_t size)
+	{
+	size_t i;
+	if (in)
+		{
+		out += size - 1;
+		for (i = 0; i < size; i++)
+			*in++ = *out--;
+		}
+	else
+		{
+		unsigned char *q;
+		char c;
+		q = out + size - 1;
+		for (i = 0; i < size/2; i++)
+			{
+			c = *q;
+			*q-- = *out;
+			*out++ = c;
+			}
+		}
+	}
diff --git a/crypto/buffer/buffer.h b/crypto/buffer/buffer.h
index 1db960745037..178e418282b2 100644
--- a/crypto/buffer/buffer.h
+++ b/crypto/buffer/buffer.h
@@ -76,18 +76,19 @@ extern "C" {
 
 struct buf_mem_st
 	{
-	int length;	/* current number of bytes */
+	size_t length;	/* current number of bytes */
 	char *data;
-	int max;	/* size of buffer */
+	size_t max;	/* size of buffer */
 	};
 
 BUF_MEM *BUF_MEM_new(void);
 void	BUF_MEM_free(BUF_MEM *a);
-int	BUF_MEM_grow(BUF_MEM *str, int len);
-int	BUF_MEM_grow_clean(BUF_MEM *str, int len);
+int	BUF_MEM_grow(BUF_MEM *str, size_t len);
+int	BUF_MEM_grow_clean(BUF_MEM *str, size_t len);
 char *	BUF_strdup(const char *str);
 char *	BUF_strndup(const char *str, size_t siz);
 void *	BUF_memdup(const void *data, size_t siz);
+void	BUF_reverse(unsigned char *out, unsigned char *in, size_t siz);
 
 /* safe string functions */
 size_t BUF_strlcpy(char *dst,const char *src,size_t siz);
diff --git a/crypto/camellia/Makefile b/crypto/camellia/Makefile
index dfd1a75a939e..6ce6fc99cd20 100644
--- a/crypto/camellia/Makefile
+++ b/crypto/camellia/Makefile
@@ -11,7 +11,7 @@ CFLAG=-g
 MAKEFILE=	Makefile
 AR=		ar r
 
-CAMELLIA_ASM_OBJ=
+CMLL_ENC= camellia.o cmll_misc.o cmll_cbc.o
 
 CFLAGS= $(INCLUDES) $(CFLAG)
 ASFLAGS= $(INCLUDES) $(ASFLAG)
@@ -23,10 +23,9 @@ APPS=
 
 LIB=$(TOP)/libcrypto.a
 LIBSRC=camellia.c cmll_misc.c cmll_ecb.c cmll_cbc.c cmll_ofb.c \
-	   cmll_cfb.c cmll_ctr.c 
+	   cmll_cfb.c cmll_ctr.c cmll_utl.c
 
-LIBOBJ= camellia.o cmll_misc.o cmll_ecb.o cmll_cbc.o cmll_ofb.o \
-		cmll_cfb.o cmll_ctr.o $(CAMELLIA_ASM_OBJ)
+LIBOBJ= cmll_ecb.o cmll_ofb.o cmll_cfb.o cmll_ctr.o cmll_utl.o $(CMLL_ENC)
 
 SRC= $(LIBSRC)
 
@@ -41,12 +40,14 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-$(LIBOBJ): $(LIBSRC)
-
+cmll-x86.s:	asm/cmll-x86.pl ../perlasm/x86asm.pl
+	$(PERL) asm/cmll-x86.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+cmll-x86_64.s:  asm/cmll-x86_64.pl
+	$(PERL) asm/cmll-x86_64.pl $(PERLASM_SCHEME) > $@
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
@@ -85,22 +86,25 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
-camellia.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-camellia.o: camellia.c camellia.h cmll_locl.h
-cmll_cbc.o: ../../include/openssl/camellia.h ../../include/openssl/e_os2.h
-cmll_cbc.o: ../../include/openssl/opensslconf.h cmll_cbc.c cmll_locl.h
-cmll_cfb.o: ../../e_os.h ../../include/openssl/camellia.h
-cmll_cfb.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-cmll_cfb.o: cmll_cfb.c cmll_locl.h
-cmll_ctr.o: ../../include/openssl/camellia.h ../../include/openssl/e_os2.h
-cmll_ctr.o: ../../include/openssl/opensslconf.h cmll_ctr.c cmll_locl.h
-cmll_ecb.o: ../../include/openssl/camellia.h ../../include/openssl/e_os2.h
+camellia.o: ../../include/openssl/opensslconf.h camellia.c camellia.h
+camellia.o: cmll_locl.h
+cmll_cbc.o: ../../include/openssl/camellia.h ../../include/openssl/modes.h
+cmll_cbc.o: ../../include/openssl/opensslconf.h cmll_cbc.c
+cmll_cfb.o: ../../include/openssl/camellia.h ../../include/openssl/modes.h
+cmll_cfb.o: ../../include/openssl/opensslconf.h cmll_cfb.c
+cmll_ctr.o: ../../include/openssl/camellia.h ../../include/openssl/modes.h
+cmll_ctr.o: ../../include/openssl/opensslconf.h cmll_ctr.c
+cmll_ecb.o: ../../include/openssl/camellia.h
 cmll_ecb.o: ../../include/openssl/opensslconf.h cmll_ecb.c cmll_locl.h
 cmll_misc.o: ../../include/openssl/camellia.h ../../include/openssl/crypto.h
-cmll_misc.o: ../../include/openssl/e_os2.h ../../include/openssl/fips.h
-cmll_misc.o: ../../include/openssl/opensslconf.h
+cmll_misc.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
 cmll_misc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 cmll_misc.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 cmll_misc.o: ../../include/openssl/symhacks.h cmll_locl.h cmll_misc.c
-cmll_ofb.o: ../../include/openssl/camellia.h ../../include/openssl/e_os2.h
-cmll_ofb.o: ../../include/openssl/opensslconf.h cmll_locl.h cmll_ofb.c
+cmll_ofb.o: ../../include/openssl/camellia.h ../../include/openssl/modes.h
+cmll_ofb.o: ../../include/openssl/opensslconf.h cmll_ofb.c
+cmll_utl.o: ../../include/openssl/camellia.h ../../include/openssl/crypto.h
+cmll_utl.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
+cmll_utl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cmll_utl.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+cmll_utl.o: ../../include/openssl/symhacks.h cmll_locl.h cmll_utl.c
diff --git a/crypto/camellia/asm/cmll-x86.pl b/crypto/camellia/asm/cmll-x86.pl
new file mode 100755
index 000000000000..c314d62312f0
--- /dev/null
+++ b/crypto/camellia/asm/cmll-x86.pl
@@ -0,0 +1,1138 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Copyright (c) 2008 Andy Polyakov <appro@openssl.org>
+#
+# This module may be used under the terms of either the GNU General
+# Public License version 2 or later, the GNU Lesser General Public
+# License version 2.1 or later, the Mozilla Public License version
+# 1.1 or the BSD License. The exact terms of either license are
+# distributed along with this module. For further details see
+# http://www.openssl.org/~appro/camellia/.
+# ====================================================================
+
+# Performance in cycles per processed byte (less is better) in
+# 'openssl speed ...' benchmark:
+#
+#			AMD K8	Core2	PIII	P4
+# -evp camellia-128-ecb	21.5	22.8	27.0	28.9
+# + over gcc 3.4.6	+90/11% +70/10%	+53/4%	+160/64%
+# + over icc 8.0	+48/19% +21/15%	+21/17%	+55/37%
+#
+# camellia-128-cbc	17.3	21.1	23.9	25.9
+#
+# 128-bit key setup	196	280	256	240	cycles/key
+# + over gcc 3.4.6	+30/0%	+17/11%	+11/0%	+63/40%
+# + over icc 8.0	+18/3%	+10/0%	+10/3%	+21/10%
+#
+# Pairs of numbers in "+" rows represent performance improvement over
+# compiler generated position-independent code, PIC, and non-PIC
+# respectively. PIC results are of greater relevance, as this module
+# is position-independent, i.e. suitable for a shared library or PIE.
+# Position independence "costs" one register, which is why compilers
+# are so close with non-PIC results, they have an extra register to
+# spare. CBC results are better than ECB ones thanks to "zero-copy"
+# private _x86_* interface, and are ~30-40% better than with compiler
+# generated cmll_cbc.o, and reach ~80-90% of x86_64 performance on
+# same CPU (where applicable).
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+$OPENSSL=1;
+
+&asm_init($ARGV[0],"cmll-586.pl",$ARGV[$#ARGV] eq "386");
+
+@T=("eax","ebx","ecx","edx");
+$idx="esi";
+$key="edi";
+$Tbl="ebp";
+
+# stack frame layout in _x86_Camellia_* routines, frame is allocated
+# by caller
+$__ra=&DWP(0,"esp");	# return address
+$__s0=&DWP(4,"esp");	# s0 backing store
+$__s1=&DWP(8,"esp");	# s1 backing store
+$__s2=&DWP(12,"esp");	# s2 backing store
+$__s3=&DWP(16,"esp");	# s3 backing store
+$__end=&DWP(20,"esp");	# pointer to end/start of key schedule
+
+# stack frame layout in Camellia_[en|crypt] routines, which differs from
+# above by 4 and overlaps by pointer to end/start of key schedule
+$_end=&DWP(16,"esp");
+$_esp=&DWP(20,"esp");
+
+# const unsigned int Camellia_SBOX[4][256];
+# Well, sort of... Camellia_SBOX[0][] is interleaved with [1][],
+# and [2][] - with [3][]. This is done to optimize code size.
+$SBOX1_1110=0;		# Camellia_SBOX[0]
+$SBOX4_4404=4;		# Camellia_SBOX[1]
+$SBOX2_0222=2048;	# Camellia_SBOX[2]
+$SBOX3_3033=2052;	# Camellia_SBOX[3]
+&static_label("Camellia_SIGMA");
+&static_label("Camellia_SBOX");
+
+sub Camellia_Feistel {
+my $i=@_[0];
+my $seed=defined(@_[1])?@_[1]:0;
+my $scale=$seed<0?-8:8;
+my $frame=defined(@_[2])?@_[2]:0;
+my $j=($i&1)*2;
+my $t0=@T[($j)%4],$t1=@T[($j+1)%4],$t2=@T[($j+2)%4],$t3=@T[($j+3)%4];
+
+	&xor	($t0,$idx);				# t0^=key[0]
+	&xor	($t1,&DWP($seed+$i*$scale+4,$key));	# t1^=key[1]
+	&movz	($idx,&HB($t0));			# (t0>>8)&0xff
+	&mov	($t3,&DWP($SBOX3_3033,$Tbl,$idx,8));	# t3=SBOX3_3033[0]
+	&movz	($idx,&LB($t0));			# (t0>>0)&0xff
+	&xor	($t3,&DWP($SBOX4_4404,$Tbl,$idx,8));	# t3^=SBOX4_4404[0]
+	&shr	($t0,16);
+	&movz	($idx,&LB($t1));			# (t1>>0)&0xff
+	&mov	($t2,&DWP($SBOX1_1110,$Tbl,$idx,8));	# t2=SBOX1_1110[1]
+	&movz	($idx,&HB($t0));			# (t0>>24)&0xff
+	&xor	($t3,&DWP($SBOX1_1110,$Tbl,$idx,8));	# t3^=SBOX1_1110[0]
+	&movz	($idx,&HB($t1));			# (t1>>8)&0xff
+	&xor	($t2,&DWP($SBOX4_4404,$Tbl,$idx,8));	# t2^=SBOX4_4404[1]
+	&shr	($t1,16);
+	&movz	($t0,&LB($t0));				# (t0>>16)&0xff
+	&xor	($t3,&DWP($SBOX2_0222,$Tbl,$t0,8));	# t3^=SBOX2_0222[0]
+	&movz	($idx,&HB($t1));			# (t1>>24)&0xff
+	&mov	($t0,&DWP($frame+4*(($j+3)%4),"esp"));	# prefetch "s3"
+	&xor	($t2,$t3);				# t2^=t3
+	&rotr	($t3,8);				# t3=RightRotate(t3,8)
+	&xor	($t2,&DWP($SBOX2_0222,$Tbl,$idx,8));	# t2^=SBOX2_0222[1]
+	&movz	($idx,&LB($t1));			# (t1>>16)&0xff
+	&mov	($t1,&DWP($frame+4*(($j+2)%4),"esp"));	# prefetch "s2"
+	&xor	($t3,$t0);				# t3^=s3
+	&xor	($t2,&DWP($SBOX3_3033,$Tbl,$idx,8));	# t2^=SBOX3_3033[1]
+	&mov	($idx,&DWP($seed+($i+1)*$scale,$key));	# prefetch key[i+1]
+	&xor	($t3,$t2);				# t3^=t2
+	&mov	(&DWP($frame+4*(($j+3)%4),"esp"),$t3);	# s3=t3
+	&xor	($t2,$t1);				# t2^=s2
+	&mov	(&DWP($frame+4*(($j+2)%4),"esp"),$t2);	# s2=t2
+}
+
+# void Camellia_EncryptBlock_Rounds(
+#		int grandRounds,
+#		const Byte plaintext[],
+#		const KEY_TABLE_TYPE keyTable,
+#		Byte ciphertext[])
+&function_begin("Camellia_EncryptBlock_Rounds");
+	&mov	("eax",&wparam(0));	# load grandRounds
+	&mov	($idx,&wparam(1));	# load plaintext pointer
+	&mov	($key,&wparam(2));	# load key schedule pointer
+
+	&mov	("ebx","esp");
+	&sub	("esp",7*4);		# place for s[0-3],keyEnd,esp and ra
+	&and	("esp",-64);
+
+	# place stack frame just "above mod 1024" the key schedule
+	# this ensures that cache associativity of 2 suffices
+	&lea	("ecx",&DWP(-64-63,$key));
+	&sub	("ecx","esp");
+	&neg	("ecx");
+	&and	("ecx",0x3C0);	# modulo 1024, but aligned to cache-line
+	&sub	("esp","ecx");
+	&add	("esp",4);	# 4 is reserved for callee's return address
+
+	&shl	("eax",6);
+	&lea	("eax",&DWP(0,$key,"eax"));
+	&mov	($_esp,"ebx");	# save %esp
+	&mov	($_end,"eax");	# save keyEnd
+
+	&call	(&label("pic_point"));
+	&set_label("pic_point");
+	&blindpop($Tbl);
+	&lea	($Tbl,&DWP(&label("Camellia_SBOX")."-".&label("pic_point"),$Tbl));
+
+	&mov	(@T[0],&DWP(0,$idx));	# load plaintext
+	&mov	(@T[1],&DWP(4,$idx));
+	&mov	(@T[2],&DWP(8,$idx));
+	&bswap	(@T[0]);
+	&mov	(@T[3],&DWP(12,$idx));
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+
+	&call	("_x86_Camellia_encrypt");
+
+	&mov	("esp",$_esp);
+	&bswap	(@T[0]);
+	&mov	($idx,&wparam(3));	# load ciphertext pointer
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+	&mov	(&DWP(0,$idx),@T[0]);	# write ciphertext
+	&mov	(&DWP(4,$idx),@T[1]);
+	&mov	(&DWP(8,$idx),@T[2]);
+	&mov	(&DWP(12,$idx),@T[3]);
+&function_end("Camellia_EncryptBlock_Rounds");
+# V1.x API
+&function_begin_B("Camellia_EncryptBlock");
+	&mov	("eax",128);
+	&sub	("eax",&wparam(0));	# load keyBitLength
+	&mov	("eax",3);
+	&adc	("eax",0);		# keyBitLength==128?3:4
+	&mov	(&wparam(0),"eax");
+	&jmp	(&label("Camellia_EncryptBlock_Rounds"));
+&function_end_B("Camellia_EncryptBlock");
+
+if ($OPENSSL) {
+# void Camellia_encrypt(
+#		const unsigned char *in,
+#		unsigned char *out,
+#		const CAMELLIA_KEY *key)
+&function_begin("Camellia_encrypt");
+	&mov	($idx,&wparam(0));	# load plaintext pointer
+	&mov	($key,&wparam(2));	# load key schedule pointer
+
+	&mov	("ebx","esp");
+	&sub	("esp",7*4);		# place for s[0-3],keyEnd,esp and ra
+	&and	("esp",-64);
+	&mov	("eax",&DWP(272,$key));	# load grandRounds counter
+
+	# place stack frame just "above mod 1024" the key schedule
+	# this ensures that cache associativity of 2 suffices
+	&lea	("ecx",&DWP(-64-63,$key));
+	&sub	("ecx","esp");
+	&neg	("ecx");
+	&and	("ecx",0x3C0);	# modulo 1024, but aligned to cache-line
+	&sub	("esp","ecx");
+	&add	("esp",4);	# 4 is reserved for callee's return address
+
+	&shl	("eax",6);
+	&lea	("eax",&DWP(0,$key,"eax"));
+	&mov	($_esp,"ebx");	# save %esp
+	&mov	($_end,"eax");	# save keyEnd
+
+	&call	(&label("pic_point"));
+	&set_label("pic_point");
+	&blindpop($Tbl);
+	&lea	($Tbl,&DWP(&label("Camellia_SBOX")."-".&label("pic_point"),$Tbl));
+
+	&mov	(@T[0],&DWP(0,$idx));	# load plaintext
+	&mov	(@T[1],&DWP(4,$idx));
+	&mov	(@T[2],&DWP(8,$idx));
+	&bswap	(@T[0]);
+	&mov	(@T[3],&DWP(12,$idx));
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+
+	&call	("_x86_Camellia_encrypt");
+
+	&mov	("esp",$_esp);
+	&bswap	(@T[0]);
+	&mov	($idx,&wparam(1));	# load ciphertext pointer
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+	&mov	(&DWP(0,$idx),@T[0]);	# write ciphertext
+	&mov	(&DWP(4,$idx),@T[1]);
+	&mov	(&DWP(8,$idx),@T[2]);
+	&mov	(&DWP(12,$idx),@T[3]);
+&function_end("Camellia_encrypt");
+}
+
+&function_begin_B("_x86_Camellia_encrypt");
+	&xor	(@T[0],&DWP(0,$key));	# ^=key[0-3]
+	&xor	(@T[1],&DWP(4,$key));
+	&xor	(@T[2],&DWP(8,$key));
+	&xor	(@T[3],&DWP(12,$key));
+	&mov	($idx,&DWP(16,$key));	# prefetch key[4]
+
+	&mov	($__s0,@T[0]);		# save s[0-3]
+	&mov	($__s1,@T[1]);
+	&mov	($__s2,@T[2]);
+	&mov	($__s3,@T[3]);
+
+&set_label("loop",16);
+	for ($i=0;$i<6;$i++) { Camellia_Feistel($i,16,4); }
+
+	&add	($key,16*4);
+	&cmp	($key,$__end);
+	&je	(&label("done"));
+
+	# @T[0-1] are preloaded, $idx is preloaded with key[0]
+	&and	($idx,@T[0]);
+	 &mov	 (@T[3],$__s3);
+	&rotl	($idx,1);
+	 &mov	 (@T[2],@T[3]);
+	&xor	(@T[1],$idx);
+	 &or	 (@T[2],&DWP(12,$key));
+	&mov	($__s1,@T[1]);		# s1^=LeftRotate(s0&key[0],1);
+	 &xor	 (@T[2],$__s2);
+
+	&mov	($idx,&DWP(4,$key));
+	 &mov	 ($__s2,@T[2]);		# s2^=s3|key[3];
+	&or	($idx,@T[1]);
+	 &and	 (@T[2],&DWP(8,$key));
+	&xor	(@T[0],$idx);
+	 &rotl	 (@T[2],1);
+	&mov	($__s0,@T[0]);		# s0^=s1|key[1];
+	 &xor	 (@T[3],@T[2]);
+	&mov	($idx,&DWP(16,$key));		# prefetch key[4]
+	 &mov	 ($__s3,@T[3]);		# s3^=LeftRotate(s2&key[2],1);
+	&jmp	(&label("loop"));
+
+&set_label("done",8);
+	&mov	(@T[2],@T[0]);		# SwapHalf
+	&mov	(@T[3],@T[1]);
+	&mov	(@T[0],$__s2);
+	&mov	(@T[1],$__s3);
+	&xor	(@T[0],$idx);		# $idx is preloaded with key[0]
+	&xor	(@T[1],&DWP(4,$key));
+	&xor	(@T[2],&DWP(8,$key));
+	&xor	(@T[3],&DWP(12,$key));
+	&ret	();
+&function_end_B("_x86_Camellia_encrypt");
+
+# void Camellia_DecryptBlock_Rounds(
+#		int grandRounds,
+#		const Byte ciphertext[],
+#		const KEY_TABLE_TYPE keyTable,
+#		Byte plaintext[])
+&function_begin("Camellia_DecryptBlock_Rounds");
+	&mov	("eax",&wparam(0));	# load grandRounds
+	&mov	($idx,&wparam(1));	# load ciphertext pointer
+	&mov	($key,&wparam(2));	# load key schedule pointer
+
+	&mov	("ebx","esp");
+	&sub	("esp",7*4);		# place for s[0-3],keyEnd,esp and ra
+	&and	("esp",-64);
+
+	# place stack frame just "above mod 1024" the key schedule
+	# this ensures that cache associativity of 2 suffices
+	&lea	("ecx",&DWP(-64-63,$key));
+	&sub	("ecx","esp");
+	&neg	("ecx");
+	&and	("ecx",0x3C0);	# modulo 1024, but aligned to cache-line
+	&sub	("esp","ecx");
+	&add	("esp",4);	# 4 is reserved for callee's return address
+
+	&shl	("eax",6);
+	&mov	(&DWP(4*4,"esp"),$key);	# save keyStart
+	&lea	($key,&DWP(0,$key,"eax"));
+	&mov	(&DWP(5*4,"esp"),"ebx");# save %esp
+
+	&call	(&label("pic_point"));
+	&set_label("pic_point");
+	&blindpop($Tbl);
+	&lea	($Tbl,&DWP(&label("Camellia_SBOX")."-".&label("pic_point"),$Tbl));
+
+	&mov	(@T[0],&DWP(0,$idx));	# load ciphertext
+	&mov	(@T[1],&DWP(4,$idx));
+	&mov	(@T[2],&DWP(8,$idx));
+	&bswap	(@T[0]);
+	&mov	(@T[3],&DWP(12,$idx));
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+
+	&call	("_x86_Camellia_decrypt");
+
+	&mov	("esp",&DWP(5*4,"esp"));
+	&bswap	(@T[0]);
+	&mov	($idx,&wparam(3));	# load plaintext pointer
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+	&mov	(&DWP(0,$idx),@T[0]);	# write plaintext
+	&mov	(&DWP(4,$idx),@T[1]);
+	&mov	(&DWP(8,$idx),@T[2]);
+	&mov	(&DWP(12,$idx),@T[3]);
+&function_end("Camellia_DecryptBlock_Rounds");
+# V1.x API
+&function_begin_B("Camellia_DecryptBlock");
+	&mov	("eax",128);
+	&sub	("eax",&wparam(0));	# load keyBitLength
+	&mov	("eax",3);
+	&adc	("eax",0);		# keyBitLength==128?3:4
+	&mov	(&wparam(0),"eax");
+	&jmp	(&label("Camellia_DecryptBlock_Rounds"));
+&function_end_B("Camellia_DecryptBlock");
+
+if ($OPENSSL) {
+# void Camellia_decrypt(
+#		const unsigned char *in,
+#		unsigned char *out,
+#		const CAMELLIA_KEY *key)
+&function_begin("Camellia_decrypt");
+	&mov	($idx,&wparam(0));	# load ciphertext pointer
+	&mov	($key,&wparam(2));	# load key schedule pointer
+
+	&mov	("ebx","esp");
+	&sub	("esp",7*4);		# place for s[0-3],keyEnd,esp and ra
+	&and	("esp",-64);
+	&mov	("eax",&DWP(272,$key));	# load grandRounds counter
+
+	# place stack frame just "above mod 1024" the key schedule
+	# this ensures that cache associativity of 2 suffices
+	&lea	("ecx",&DWP(-64-63,$key));
+	&sub	("ecx","esp");
+	&neg	("ecx");
+	&and	("ecx",0x3C0);	# modulo 1024, but aligned to cache-line
+	&sub	("esp","ecx");
+	&add	("esp",4);	# 4 is reserved for callee's return address
+
+	&shl	("eax",6);
+	&mov	(&DWP(4*4,"esp"),$key);	# save keyStart
+	&lea	($key,&DWP(0,$key,"eax"));
+	&mov	(&DWP(5*4,"esp"),"ebx");# save %esp
+
+	&call	(&label("pic_point"));
+	&set_label("pic_point");
+	&blindpop($Tbl);
+	&lea	($Tbl,&DWP(&label("Camellia_SBOX")."-".&label("pic_point"),$Tbl));
+
+	&mov	(@T[0],&DWP(0,$idx));	# load ciphertext
+	&mov	(@T[1],&DWP(4,$idx));
+	&mov	(@T[2],&DWP(8,$idx));
+	&bswap	(@T[0]);
+	&mov	(@T[3],&DWP(12,$idx));
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+
+	&call	("_x86_Camellia_decrypt");
+
+	&mov	("esp",&DWP(5*4,"esp"));
+	&bswap	(@T[0]);
+	&mov	($idx,&wparam(1));	# load plaintext pointer
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+	&mov	(&DWP(0,$idx),@T[0]);	# write plaintext
+	&mov	(&DWP(4,$idx),@T[1]);
+	&mov	(&DWP(8,$idx),@T[2]);
+	&mov	(&DWP(12,$idx),@T[3]);
+&function_end("Camellia_decrypt");
+}
+
+&function_begin_B("_x86_Camellia_decrypt");
+	&xor	(@T[0],&DWP(0,$key));	# ^=key[0-3]
+	&xor	(@T[1],&DWP(4,$key));
+	&xor	(@T[2],&DWP(8,$key));
+	&xor	(@T[3],&DWP(12,$key));
+	&mov	($idx,&DWP(-8,$key));	# prefetch key[-2]
+
+	&mov	($__s0,@T[0]);		# save s[0-3]
+	&mov	($__s1,@T[1]);
+	&mov	($__s2,@T[2]);
+	&mov	($__s3,@T[3]);
+
+&set_label("loop",16);
+	for ($i=0;$i<6;$i++) { Camellia_Feistel($i,-8,4); }
+
+	&sub	($key,16*4);
+	&cmp	($key,$__end);
+	&je	(&label("done"));
+
+	# @T[0-1] are preloaded, $idx is preloaded with key[2]
+	&and	($idx,@T[0]);
+	 &mov	 (@T[3],$__s3);
+	&rotl	($idx,1);
+	 &mov	 (@T[2],@T[3]);
+	&xor	(@T[1],$idx);
+	 &or	 (@T[2],&DWP(4,$key));
+	&mov	($__s1,@T[1]);		# s1^=LeftRotate(s0&key[0],1);
+	 &xor	 (@T[2],$__s2);
+
+	&mov	($idx,&DWP(12,$key));
+	 &mov	 ($__s2,@T[2]);		# s2^=s3|key[3];
+	&or	($idx,@T[1]);
+	 &and	 (@T[2],&DWP(0,$key));
+	&xor	(@T[0],$idx);
+	 &rotl	 (@T[2],1);
+	&mov	($__s0,@T[0]);		# s0^=s1|key[1];
+	 &xor	 (@T[3],@T[2]);
+	&mov	($idx,&DWP(-8,$key));	# prefetch key[4]
+	 &mov	 ($__s3,@T[3]);		# s3^=LeftRotate(s2&key[2],1);
+	&jmp	(&label("loop"));
+
+&set_label("done",8);
+	&mov	(@T[2],@T[0]);		# SwapHalf
+	&mov	(@T[3],@T[1]);
+	&mov	(@T[0],$__s2);
+	&mov	(@T[1],$__s3);
+	&xor	(@T[2],$idx);		# $idx is preloaded with key[2]
+	&xor	(@T[3],&DWP(12,$key));
+	&xor	(@T[0],&DWP(0,$key));
+	&xor	(@T[1],&DWP(4,$key));
+	&ret	();
+&function_end_B("_x86_Camellia_decrypt");
+
+# shld is very slow on Intel P4 family. Even on AMD it limits
+# instruction decode rate [because it's VectorPath] and consequently
+# performance. PIII, PM and Core[2] seem to be the only ones which
+# execute this code ~7% faster...
+sub __rotl128 {
+  my ($i0,$i1,$i2,$i3,$rot,$rnd,@T)=@_;
+
+    $rnd *= 2;
+    if ($rot) {
+	&mov	($idx,$i0);
+	&shld	($i0,$i1,$rot);
+	&shld	($i1,$i2,$rot);
+	&shld	($i2,$i3,$rot);
+	&shld	($i3,$idx,$rot);
+    }
+    &mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i0 eq @T[0]);
+    &mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i1 eq @T[0]);
+    &mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i2 eq @T[0]);
+    &mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i3 eq @T[0]);
+}
+
+# ... Implementing 128-bit rotate without shld gives >3x performance
+# improvement on P4, only ~7% degradation on other Intel CPUs and
+# not worse performance on AMD. This is therefore preferred.
+sub _rotl128 {
+  my ($i0,$i1,$i2,$i3,$rot,$rnd,@T)=@_;
+
+    $rnd *= 2;
+    if ($rot) {
+	&mov	($Tbl,$i0);
+	&shl	($i0,$rot);
+	&mov	($idx,$i1);
+	&shr	($idx,32-$rot);
+	&shl	($i1,$rot);
+	&or	($i0,$idx);
+	&mov	($idx,$i2);
+	&shl	($i2,$rot);
+	&mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i0 eq @T[0]);
+	&shr	($idx,32-$rot);
+	&or	($i1,$idx);
+	&shr	($Tbl,32-$rot);
+	&mov	($idx,$i3);
+	&shr	($idx,32-$rot);
+	&mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i1 eq @T[0]);
+	&shl	($i3,$rot);
+	&or	($i2,$idx);
+	&or	($i3,$Tbl);
+	&mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i2 eq @T[0]);
+	&mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i3 eq @T[0]);
+    } else {
+	&mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i0 eq @T[0]);
+	&mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i1 eq @T[0]);
+	&mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i2 eq @T[0]);
+	&mov	(&DWP(-128+4*$rnd++,$key),shift(@T))	if ($i3 eq @T[0]);
+    }
+}
+
+sub _saveround {
+my ($rnd,$key,@T)=@_;
+my $bias=int(@T[0])?shift(@T):0;
+
+	&mov	(&DWP($bias+$rnd*8+0,$key),@T[0]);
+	&mov	(&DWP($bias+$rnd*8+4,$key),@T[1])	if ($#T>=1);
+	&mov	(&DWP($bias+$rnd*8+8,$key),@T[2])	if ($#T>=2);
+	&mov	(&DWP($bias+$rnd*8+12,$key),@T[3])	if ($#T>=3);
+}
+
+sub _loadround {
+my ($rnd,$key,@T)=@_;
+my $bias=int(@T[0])?shift(@T):0;
+
+	&mov	(@T[0],&DWP($bias+$rnd*8+0,$key));
+	&mov	(@T[1],&DWP($bias+$rnd*8+4,$key))	if ($#T>=1);
+	&mov	(@T[2],&DWP($bias+$rnd*8+8,$key))	if ($#T>=2);
+	&mov	(@T[3],&DWP($bias+$rnd*8+12,$key))	if ($#T>=3);
+}
+
+# void Camellia_Ekeygen(
+#		const int keyBitLength,
+#		const Byte *rawKey,
+#		KEY_TABLE_TYPE keyTable)
+&function_begin("Camellia_Ekeygen");
+{ my $step=0;
+
+	&stack_push(4);				# place for s[0-3]
+
+	&mov	($Tbl,&wparam(0));		# load arguments
+	&mov	($idx,&wparam(1));
+	&mov	($key,&wparam(2));
+
+	&mov	(@T[0],&DWP(0,$idx));		# load 0-127 bits
+	&mov	(@T[1],&DWP(4,$idx));
+	&mov	(@T[2],&DWP(8,$idx));
+	&mov	(@T[3],&DWP(12,$idx));
+
+	&bswap	(@T[0]);
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+
+	&_saveround	(0,$key,@T);		# KL<<<0
+
+	&cmp	($Tbl,128);
+	&je	(&label("1st128"));
+
+	&mov	(@T[0],&DWP(16,$idx));		# load 128-191 bits
+	&mov	(@T[1],&DWP(20,$idx));
+	&cmp	($Tbl,192);
+	&je	(&label("1st192"));
+	&mov	(@T[2],&DWP(24,$idx));		# load 192-255 bits
+	&mov	(@T[3],&DWP(28,$idx));
+	&jmp	(&label("1st256"));
+&set_label("1st192",4);
+	&mov	(@T[2],@T[0]);
+	&mov	(@T[3],@T[1]);
+	&not	(@T[2]);
+	&not	(@T[3]);
+&set_label("1st256",4);
+	&bswap	(@T[0]);
+	&bswap	(@T[1]);
+	&bswap	(@T[2]);
+	&bswap	(@T[3]);
+
+	&_saveround	(4,$key,@T);		# temporary storage for KR!
+
+	&xor	(@T[0],&DWP(0*8+0,$key));	# KR^KL
+	&xor	(@T[1],&DWP(0*8+4,$key));
+	&xor	(@T[2],&DWP(1*8+0,$key));
+	&xor	(@T[3],&DWP(1*8+4,$key));
+
+&set_label("1st128",4);
+	&call	(&label("pic_point"));
+	&set_label("pic_point");
+	&blindpop($Tbl);
+	&lea	($Tbl,&DWP(&label("Camellia_SBOX")."-".&label("pic_point"),$Tbl));
+	&lea	($key,&DWP(&label("Camellia_SIGMA")."-".&label("Camellia_SBOX"),$Tbl));
+
+	&mov	($idx,&DWP($step*8,$key));	# prefetch SIGMA[0]
+	&mov	(&swtmp(0),@T[0]);		# save s[0-3]
+	&mov	(&swtmp(1),@T[1]);
+	&mov	(&swtmp(2),@T[2]);
+	&mov	(&swtmp(3),@T[3]);
+	&Camellia_Feistel($step++);
+	&Camellia_Feistel($step++);
+	&mov	(@T[2],&swtmp(2));
+	&mov	(@T[3],&swtmp(3));
+
+	&mov	($idx,&wparam(2));
+	&xor	(@T[0],&DWP(0*8+0,$idx));	# ^KL
+	&xor	(@T[1],&DWP(0*8+4,$idx));
+	&xor	(@T[2],&DWP(1*8+0,$idx));
+	&xor	(@T[3],&DWP(1*8+4,$idx));
+
+	&mov	($idx,&DWP($step*8,$key));	# prefetch SIGMA[4]
+	&mov	(&swtmp(0),@T[0]);		# save s[0-3]
+	&mov	(&swtmp(1),@T[1]);
+	&mov	(&swtmp(2),@T[2]);
+	&mov	(&swtmp(3),@T[3]);
+	&Camellia_Feistel($step++);
+	&Camellia_Feistel($step++);
+	&mov	(@T[2],&swtmp(2));
+	&mov	(@T[3],&swtmp(3));
+
+	&mov	($idx,&wparam(0));
+	&cmp	($idx,128);
+	&jne	(&label("2nd256"));
+
+	&mov	($key,&wparam(2));
+	&lea	($key,&DWP(128,$key));		# size optimization
+
+	####### process KA
+	&_saveround	(2,$key,-128,@T);	# KA<<<0
+	&_rotl128	(@T,15,6,@T);		# KA<<<15
+	&_rotl128	(@T,15,8,@T);		# KA<<<(15+15=30)
+	&_rotl128	(@T,15,12,@T[0],@T[1]);	# KA<<<(30+15=45)
+	&_rotl128	(@T,15,14,@T);		# KA<<<(45+15=60)
+	push		(@T,shift(@T));		# rotl128(@T,32);
+	&_rotl128	(@T,2,20,@T);		# KA<<<(60+32+2=94)
+	&_rotl128	(@T,17,24,@T);		# KA<<<(94+17=111)
+
+	####### process KL
+	&_loadround	(0,$key,-128,@T);	# load KL
+	&_rotl128	(@T,15,4,@T);		# KL<<<15
+	&_rotl128	(@T,30,10,@T);		# KL<<<(15+30=45)
+	&_rotl128	(@T,15,13,@T[2],@T[3]);	# KL<<<(45+15=60)
+	&_rotl128	(@T,17,16,@T);		# KL<<<(60+17=77)
+	&_rotl128	(@T,17,18,@T);		# KL<<<(77+17=94)
+	&_rotl128	(@T,17,22,@T);		# KL<<<(94+17=111)
+
+	while (@T[0] ne "eax")			# restore order
+	{   unshift	(@T,pop(@T));   }
+
+	&mov	("eax",3);			# 3 grandRounds
+	&jmp	(&label("done"));
+
+&set_label("2nd256",16);
+	&mov	($idx,&wparam(2));
+	&_saveround	(6,$idx,@T);		# temporary storage for KA!
+
+	&xor	(@T[0],&DWP(4*8+0,$idx));	# KA^KR
+	&xor	(@T[1],&DWP(4*8+4,$idx));
+	&xor	(@T[2],&DWP(5*8+0,$idx));
+	&xor	(@T[3],&DWP(5*8+4,$idx));
+
+	&mov	($idx,&DWP($step*8,$key));	# prefetch SIGMA[8]
+	&mov	(&swtmp(0),@T[0]);		# save s[0-3]
+	&mov	(&swtmp(1),@T[1]);
+	&mov	(&swtmp(2),@T[2]);
+	&mov	(&swtmp(3),@T[3]);
+	&Camellia_Feistel($step++);
+	&Camellia_Feistel($step++);
+	&mov	(@T[2],&swtmp(2));
+	&mov	(@T[3],&swtmp(3));
+
+	&mov	($key,&wparam(2));
+	&lea	($key,&DWP(128,$key));		# size optimization
+
+	####### process KB
+	&_saveround	(2,$key,-128,@T);	# KB<<<0
+	&_rotl128	(@T,30,10,@T);		# KB<<<30
+	&_rotl128	(@T,30,20,@T);		# KB<<<(30+30=60)
+	push		(@T,shift(@T));		# rotl128(@T,32);
+	&_rotl128	(@T,19,32,@T);		# KB<<<(60+32+19=111)
+
+	####### process KR
+	&_loadround	(4,$key,-128,@T);	# load KR
+	&_rotl128	(@T,15,4,@T);		# KR<<<15
+	&_rotl128	(@T,15,8,@T);		# KR<<<(15+15=30)
+	&_rotl128	(@T,30,18,@T);		# KR<<<(30+30=60)
+	push		(@T,shift(@T));		# rotl128(@T,32);
+	&_rotl128	(@T,2,26,@T);		# KR<<<(60+32+2=94)
+
+	####### process KA
+	&_loadround	(6,$key,-128,@T);	# load KA
+	&_rotl128	(@T,15,6,@T);		# KA<<<15
+	&_rotl128	(@T,30,14,@T);		# KA<<<(15+30=45)
+	push		(@T,shift(@T));		# rotl128(@T,32);
+	&_rotl128	(@T,0,24,@T);		# KA<<<(45+32+0=77)
+	&_rotl128	(@T,17,28,@T);		# KA<<<(77+17=94)
+
+	####### process KL
+	&_loadround	(0,$key,-128,@T);	# load KL
+	push		(@T,shift(@T));		# rotl128(@T,32);
+	&_rotl128	(@T,13,12,@T);		# KL<<<(32+13=45)
+	&_rotl128	(@T,15,16,@T);		# KL<<<(45+15=60)
+	&_rotl128	(@T,17,22,@T);		# KL<<<(60+17=77)
+	push		(@T,shift(@T));		# rotl128(@T,32);
+	&_rotl128	(@T,2,30,@T);		# KL<<<(77+32+2=111)
+
+	while (@T[0] ne "eax")			# restore order
+	{   unshift	(@T,pop(@T));   }
+
+	&mov	("eax",4);			# 4 grandRounds
+&set_label("done");
+	&lea	("edx",&DWP(272-128,$key));	# end of key schedule
+	&stack_pop(4);
+}
+&function_end("Camellia_Ekeygen");
+
+if ($OPENSSL) {
+# int private_Camellia_set_key (
+#		const unsigned char *userKey,
+#		int bits,
+#		CAMELLIA_KEY *key)
+&function_begin_B("private_Camellia_set_key");
+	&push	("ebx");
+	&mov	("ecx",&wparam(0));	# pull arguments
+	&mov	("ebx",&wparam(1));
+	&mov	("edx",&wparam(2));
+
+	&mov	("eax",-1);
+	&test	("ecx","ecx");
+	&jz	(&label("done"));	# userKey==NULL?
+	&test	("edx","edx");
+	&jz	(&label("done"));	# key==NULL?
+
+	&mov	("eax",-2);
+	&cmp	("ebx",256);
+	&je	(&label("arg_ok"));	# bits==256?
+	&cmp	("ebx",192);
+	&je	(&label("arg_ok"));	# bits==192?
+	&cmp	("ebx",128);
+	&jne	(&label("done"));	# bits!=128?
+&set_label("arg_ok",4);
+
+	&push	("edx");		# push arguments
+	&push	("ecx");
+	&push	("ebx");
+	&call	("Camellia_Ekeygen");
+	&stack_pop(3);
+
+	# eax holds grandRounds and edx points at where to put it
+	&mov	(&DWP(0,"edx"),"eax");
+	&xor	("eax","eax");
+&set_label("done",4);
+	&pop	("ebx");
+	&ret	();
+&function_end_B("private_Camellia_set_key");
+}
+
+@SBOX=(
+112,130, 44,236,179, 39,192,229,228,133, 87, 53,234, 12,174, 65,
+ 35,239,107,147, 69, 25,165, 33,237, 14, 79, 78, 29,101,146,189,
+134,184,175,143,124,235, 31,206, 62, 48,220, 95, 94,197, 11, 26,
+166,225, 57,202,213, 71, 93, 61,217,  1, 90,214, 81, 86,108, 77,
+139, 13,154,102,251,204,176, 45,116, 18, 43, 32,240,177,132,153,
+223, 76,203,194, 52,126,118,  5,109,183,169, 49,209, 23,  4,215,
+ 20, 88, 58, 97,222, 27, 17, 28, 50, 15,156, 22, 83, 24,242, 34,
+254, 68,207,178,195,181,122,145, 36,  8,232,168, 96,252,105, 80,
+170,208,160,125,161,137, 98,151, 84, 91, 30,149,224,255,100,210,
+ 16,196,  0, 72,163,247,117,219,138,  3,230,218,  9, 63,221,148,
+135, 92,131,  2,205, 74,144, 51,115,103,246,243,157,127,191,226,
+ 82,155,216, 38,200, 55,198, 59,129,150,111, 75, 19,190, 99, 46,
+233,121,167,140,159,110,188,142, 41,245,249,182, 47,253,180, 89,
+120,152,  6,106,231, 70,113,186,212, 37,171, 66,136,162,141,250,
+114,  7,185, 85,248,238,172, 10, 54, 73, 42,104, 60, 56,241,164,
+ 64, 40,211,123,187,201, 67,193, 21,227,173,244,119,199,128,158);
+
+sub S1110 { my $i=shift; $i=@SBOX[$i]; return $i<<24|$i<<16|$i<<8; }
+sub S4404 { my $i=shift; $i=($i<<1|$i>>7)&0xff; $i=@SBOX[$i]; return $i<<24|$i<<16|$i; }	
+sub S0222 { my $i=shift; $i=@SBOX[$i]; $i=($i<<1|$i>>7)&0xff; return $i<<16|$i<<8|$i; }	
+sub S3033 { my $i=shift; $i=@SBOX[$i]; $i=($i>>1|$i<<7)&0xff; return $i<<24|$i<<8|$i; }	
+
+&set_label("Camellia_SIGMA",64);
+&data_word(
+    0xa09e667f, 0x3bcc908b, 0xb67ae858, 0x4caa73b2,
+    0xc6ef372f, 0xe94f82be, 0x54ff53a5, 0xf1d36f1c,
+    0x10e527fa, 0xde682d1d, 0xb05688c2, 0xb3e6c1fd,
+    0,          0,          0,          0);
+&set_label("Camellia_SBOX",64);
+# tables are interleaved, remember?
+for ($i=0;$i<256;$i++) { &data_word(&S1110($i),&S4404($i)); }
+for ($i=0;$i<256;$i++) { &data_word(&S0222($i),&S3033($i)); }
+
+# void Camellia_cbc_encrypt (const void char *inp, unsigned char *out,
+#			size_t length, const CAMELLIA_KEY *key,
+#			unsigned char *ivp,const int enc);
+{
+# stack frame layout
+#             -4(%esp)		# return address	 0(%esp)
+#              0(%esp)		# s0			 4(%esp)
+#              4(%esp)		# s1			 8(%esp)
+#              8(%esp)		# s2			12(%esp)
+#             12(%esp)		# s3			16(%esp)
+#             16(%esp)		# end of key schedule	20(%esp)
+#             20(%esp)		# %esp backup
+my $_inp=&DWP(24,"esp");	#copy of wparam(0)
+my $_out=&DWP(28,"esp");	#copy of wparam(1)
+my $_len=&DWP(32,"esp");	#copy of wparam(2)
+my $_key=&DWP(36,"esp");	#copy of wparam(3)
+my $_ivp=&DWP(40,"esp");	#copy of wparam(4)
+my $ivec=&DWP(44,"esp");	#ivec[16]
+my $_tmp=&DWP(44,"esp");	#volatile variable [yes, aliases with ivec]
+my ($s0,$s1,$s2,$s3) = @T;
+
+&function_begin("Camellia_cbc_encrypt");
+	&mov	($s2 eq "ecx"? $s2 : "",&wparam(2));	# load len
+	&cmp	($s2,0);
+	&je	(&label("enc_out"));
+
+	&pushf	();
+	&cld	();
+
+	&mov	($s0,&wparam(0));	# load inp
+	&mov	($s1,&wparam(1));	# load out
+	#&mov	($s2,&wparam(2));	# load len
+	&mov	($s3,&wparam(3));	# load key
+	&mov	($Tbl,&wparam(4));	# load ivp
+
+	# allocate aligned stack frame...
+	&lea	($idx,&DWP(-64,"esp"));
+	&and	($idx,-64);
+
+	# place stack frame just "above mod 1024" the key schedule
+	# this ensures that cache associativity of 2 suffices
+	&lea	($key,&DWP(-64-63,$s3));
+	&sub	($key,$idx);
+	&neg	($key);
+	&and	($key,0x3C0);	# modulo 1024, but aligned to cache-line
+	&sub	($idx,$key);
+
+	&mov	($key,&wparam(5));	# load enc
+
+	&exch	("esp",$idx);
+	&add	("esp",4);		# reserve for return address!
+	&mov	($_esp,$idx);		# save %esp
+
+	&mov	($_inp,$s0);		# save copy of inp
+	&mov	($_out,$s1);		# save copy of out
+	&mov	($_len,$s2);		# save copy of len
+	&mov	($_key,$s3);		# save copy of key
+	&mov	($_ivp,$Tbl);		# save copy of ivp
+
+	&call   (&label("pic_point"));	# make it PIC!
+	&set_label("pic_point");
+	&blindpop($Tbl);
+	&lea    ($Tbl,&DWP(&label("Camellia_SBOX")."-".&label("pic_point"),$Tbl));
+
+	&mov	($idx,32);
+	&set_label("prefetch_sbox",4);
+		&mov	($s0,&DWP(0,$Tbl));
+		&mov	($s1,&DWP(32,$Tbl));
+		&mov	($s2,&DWP(64,$Tbl));
+		&mov	($s3,&DWP(96,$Tbl));
+		&lea	($Tbl,&DWP(128,$Tbl));
+		&dec	($idx);
+	&jnz	(&label("prefetch_sbox"));
+	&mov	($s0,$_key);
+	&sub	($Tbl,4096);
+	&mov	($idx,$_inp);
+	&mov	($s3,&DWP(272,$s0));		# load grandRounds
+
+	&cmp	($key,0);
+	&je	(&label("DECRYPT"));
+
+	&mov	($s2,$_len);
+	&mov	($key,$_ivp);
+	&shl	($s3,6);
+	&lea	($s3,&DWP(0,$s0,$s3));
+	&mov	($_end,$s3);
+
+	&test	($s2,0xFFFFFFF0);
+	&jz	(&label("enc_tail"));		# short input...
+
+	&mov	($s0,&DWP(0,$key));		# load iv
+	&mov	($s1,&DWP(4,$key));
+
+	&set_label("enc_loop",4);
+		&mov	($s2,&DWP(8,$key));
+		&mov	($s3,&DWP(12,$key));
+
+		&xor	($s0,&DWP(0,$idx));	# xor input data
+		&xor	($s1,&DWP(4,$idx));
+		&xor	($s2,&DWP(8,$idx));
+		&bswap	($s0);
+		&xor	($s3,&DWP(12,$idx));
+		&bswap	($s1);
+		&mov	($key,$_key);		# load key
+		&bswap	($s2);
+		&bswap	($s3);
+
+		&call	("_x86_Camellia_encrypt");
+
+		&mov	($idx,$_inp);		# load inp
+		&mov	($key,$_out);		# load out
+
+		&bswap	($s0);
+		&bswap	($s1);
+		&bswap	($s2);
+		&mov	(&DWP(0,$key),$s0);	# save output data
+		&bswap	($s3);
+		&mov	(&DWP(4,$key),$s1);
+		&mov	(&DWP(8,$key),$s2);
+		&mov	(&DWP(12,$key),$s3);
+
+		&mov	($s2,$_len);		# load len
+
+		&lea	($idx,&DWP(16,$idx));
+		&mov	($_inp,$idx);		# save inp
+
+		&lea	($s3,&DWP(16,$key));
+		&mov	($_out,$s3);		# save out
+
+		&sub	($s2,16);
+		&test	($s2,0xFFFFFFF0);
+		&mov	($_len,$s2);		# save len
+	&jnz	(&label("enc_loop"));
+	&test	($s2,15);
+	&jnz	(&label("enc_tail"));
+	&mov	($idx,$_ivp);		# load ivp
+	&mov	($s2,&DWP(8,$key));	# restore last dwords
+	&mov	($s3,&DWP(12,$key));
+	&mov	(&DWP(0,$idx),$s0);	# save ivec
+	&mov	(&DWP(4,$idx),$s1);
+	&mov	(&DWP(8,$idx),$s2);
+	&mov	(&DWP(12,$idx),$s3);
+
+	&mov	("esp",$_esp);
+	&popf	();
+    &set_label("enc_out");
+	&function_end_A();
+	&pushf	();			# kludge, never executed
+
+    &set_label("enc_tail",4);
+	&mov	($s0,$key eq "edi" ? $key : "");
+	&mov	($key,$_out);			# load out
+	&push	($s0);				# push ivp
+	&mov	($s1,16);
+	&sub	($s1,$s2);
+	&cmp	($key,$idx);			# compare with inp
+	&je	(&label("enc_in_place"));
+	&align	(4);
+	&data_word(0xA4F3F689);	# rep movsb	# copy input
+	&jmp	(&label("enc_skip_in_place"));
+    &set_label("enc_in_place");
+	&lea	($key,&DWP(0,$key,$s2));
+    &set_label("enc_skip_in_place");
+	&mov	($s2,$s1);
+	&xor	($s0,$s0);
+	&align	(4);
+	&data_word(0xAAF3F689);	# rep stosb	# zero tail
+	&pop	($key);				# pop ivp
+
+	&mov	($idx,$_out);			# output as input
+	&mov	($s0,&DWP(0,$key));
+	&mov	($s1,&DWP(4,$key));
+	&mov	($_len,16);			# len=16
+	&jmp	(&label("enc_loop"));		# one more spin...
+
+#----------------------------- DECRYPT -----------------------------#
+&set_label("DECRYPT",16);
+	&shl	($s3,6);
+	&lea	($s3,&DWP(0,$s0,$s3));
+	&mov	($_end,$s0);
+	&mov	($_key,$s3);
+
+	&cmp	($idx,$_out);
+	&je	(&label("dec_in_place"));	# in-place processing...
+
+	&mov	($key,$_ivp);			# load ivp
+	&mov	($_tmp,$key);
+
+	&set_label("dec_loop",4);
+		&mov	($s0,&DWP(0,$idx));	# read input
+		&mov	($s1,&DWP(4,$idx));
+		&mov	($s2,&DWP(8,$idx));
+		&bswap	($s0);
+		&mov	($s3,&DWP(12,$idx));
+		&bswap	($s1);
+		&mov	($key,$_key);		# load key
+		&bswap	($s2);
+		&bswap	($s3);
+
+		&call	("_x86_Camellia_decrypt");
+
+		&mov	($key,$_tmp);		# load ivp
+		&mov	($idx,$_len);		# load len
+
+		&bswap	($s0);
+		&bswap	($s1);
+		&bswap	($s2);
+		&xor	($s0,&DWP(0,$key));	# xor iv
+		&bswap	($s3);
+		&xor	($s1,&DWP(4,$key));
+		&xor	($s2,&DWP(8,$key));
+		&xor	($s3,&DWP(12,$key));
+
+		&sub	($idx,16);
+		&jc	(&label("dec_partial"));
+		&mov	($_len,$idx);		# save len
+		&mov	($idx,$_inp);		# load inp
+		&mov	($key,$_out);		# load out
+
+		&mov	(&DWP(0,$key),$s0);	# write output
+		&mov	(&DWP(4,$key),$s1);
+		&mov	(&DWP(8,$key),$s2);
+		&mov	(&DWP(12,$key),$s3);
+
+		&mov	($_tmp,$idx);		# save ivp
+		&lea	($idx,&DWP(16,$idx));
+		&mov	($_inp,$idx);		# save inp
+
+		&lea	($key,&DWP(16,$key));
+		&mov	($_out,$key);		# save out
+
+	&jnz	(&label("dec_loop"));
+	&mov	($key,$_tmp);		# load temp ivp
+    &set_label("dec_end");
+	&mov	($idx,$_ivp);		# load user ivp
+	&mov	($s0,&DWP(0,$key));	# load iv
+	&mov	($s1,&DWP(4,$key));
+	&mov	($s2,&DWP(8,$key));
+	&mov	($s3,&DWP(12,$key));
+	&mov	(&DWP(0,$idx),$s0);	# copy back to user
+	&mov	(&DWP(4,$idx),$s1);
+	&mov	(&DWP(8,$idx),$s2);
+	&mov	(&DWP(12,$idx),$s3);
+	&jmp	(&label("dec_out"));
+
+    &set_label("dec_partial",4);
+	&lea	($key,$ivec);
+	&mov	(&DWP(0,$key),$s0);	# dump output to stack
+	&mov	(&DWP(4,$key),$s1);
+	&mov	(&DWP(8,$key),$s2);
+	&mov	(&DWP(12,$key),$s3);
+	&lea	($s2 eq "ecx" ? $s2 : "",&DWP(16,$idx));
+	&mov	($idx eq "esi" ? $idx : "",$key);
+	&mov	($key eq "edi" ? $key : "",$_out);	# load out
+	&data_word(0xA4F3F689);	# rep movsb		# copy output
+	&mov	($key,$_inp);				# use inp as temp ivp
+	&jmp	(&label("dec_end"));
+
+    &set_label("dec_in_place",4);
+	&set_label("dec_in_place_loop");
+		&lea	($key,$ivec);
+		&mov	($s0,&DWP(0,$idx));	# read input
+		&mov	($s1,&DWP(4,$idx));
+		&mov	($s2,&DWP(8,$idx));
+		&mov	($s3,&DWP(12,$idx));
+
+		&mov	(&DWP(0,$key),$s0);	# copy to temp
+		&mov	(&DWP(4,$key),$s1);
+		&mov	(&DWP(8,$key),$s2);
+		&bswap	($s0);
+		&mov	(&DWP(12,$key),$s3);
+		&bswap	($s1);
+		&mov	($key,$_key);		# load key
+		&bswap	($s2);
+		&bswap	($s3);
+
+		&call	("_x86_Camellia_decrypt");
+
+		&mov	($key,$_ivp);		# load ivp
+		&mov	($idx,$_out);		# load out
+
+		&bswap	($s0);
+		&bswap	($s1);
+		&bswap	($s2);
+		&xor	($s0,&DWP(0,$key));	# xor iv
+		&bswap	($s3);
+		&xor	($s1,&DWP(4,$key));
+		&xor	($s2,&DWP(8,$key));
+		&xor	($s3,&DWP(12,$key));
+
+		&mov	(&DWP(0,$idx),$s0);	# write output
+		&mov	(&DWP(4,$idx),$s1);
+		&mov	(&DWP(8,$idx),$s2);
+		&mov	(&DWP(12,$idx),$s3);
+
+		&lea	($idx,&DWP(16,$idx));
+		&mov	($_out,$idx);		# save out
+
+		&lea	($idx,$ivec);
+		&mov	($s0,&DWP(0,$idx));	# read temp
+		&mov	($s1,&DWP(4,$idx));
+		&mov	($s2,&DWP(8,$idx));
+		&mov	($s3,&DWP(12,$idx));
+
+		&mov	(&DWP(0,$key),$s0);	# copy iv
+		&mov	(&DWP(4,$key),$s1);
+		&mov	(&DWP(8,$key),$s2);
+		&mov	(&DWP(12,$key),$s3);
+
+		&mov	($idx,$_inp);		# load inp
+
+		&lea	($idx,&DWP(16,$idx));
+		&mov	($_inp,$idx);		# save inp
+
+		&mov	($s2,$_len);		# load len
+		&sub	($s2,16);
+		&jc	(&label("dec_in_place_partial"));
+		&mov	($_len,$s2);		# save len
+	&jnz	(&label("dec_in_place_loop"));
+	&jmp	(&label("dec_out"));
+
+    &set_label("dec_in_place_partial",4);
+	# one can argue if this is actually required...
+	&mov	($key eq "edi" ? $key : "",$_out);
+	&lea	($idx eq "esi" ? $idx : "",$ivec);
+	&lea	($key,&DWP(0,$key,$s2));
+	&lea	($idx,&DWP(16,$idx,$s2));
+	&neg	($s2 eq "ecx" ? $s2 : "");
+	&data_word(0xA4F3F689);	# rep movsb	# restore tail
+
+    &set_label("dec_out",4);
+    &mov	("esp",$_esp);
+    &popf	();
+&function_end("Camellia_cbc_encrypt");
+}
+
+&asciz("Camellia for x86 by <appro\@openssl.org>");
+
+&asm_finish();
diff --git a/crypto/camellia/asm/cmll-x86_64.pl b/crypto/camellia/asm/cmll-x86_64.pl
new file mode 100755
index 000000000000..76955e47265c
--- /dev/null
+++ b/crypto/camellia/asm/cmll-x86_64.pl
@@ -0,0 +1,1080 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Copyright (c) 2008 Andy Polyakov <appro@openssl.org>
+#
+# This module may be used under the terms of either the GNU General
+# Public License version 2 or later, the GNU Lesser General Public
+# License version 2.1 or later, the Mozilla Public License version
+# 1.1 or the BSD License. The exact terms of either license are
+# distributed along with this module. For further details see
+# http://www.openssl.org/~appro/camellia/.
+# ====================================================================
+
+# Performance in cycles per processed byte (less is better) in
+# 'openssl speed ...' benchmark:
+#
+#			AMD64	Core2	EM64T
+# -evp camellia-128-ecb	16.7	21.0	22.7
+# + over gcc 3.4.6	+25%	+5%	0%
+#
+# camellia-128-cbc	15.7	20.4	21.1
+#
+# 128-bit key setup	128	216	205	cycles/key
+# + over gcc 3.4.6	+54%	+39%	+15%
+#
+# Numbers in "+" rows represent performance improvement over compiler
+# generated code. Key setup timings are impressive on AMD and Core2
+# thanks to 64-bit operations being covertly deployed. Improvement on
+# EM64T, pre-Core2 Intel x86_64 CPU, is not as impressive, because it
+# apparently emulates some of 64-bit operations in [32-bit] microcode.
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+sub hi() { my $r=shift; $r =~ s/%[er]([a-d])x/%\1h/;    $r; }
+sub lo() { my $r=shift; $r =~ s/%[er]([a-d])x/%\1l/;
+                        $r =~ s/%[er]([sd]i)/%\1l/;
+                        $r =~ s/%(r[0-9]+)[d]?/%\1b/;   $r; }
+
+$t0="%eax";$t1="%ebx";$t2="%ecx";$t3="%edx";
+@S=("%r8d","%r9d","%r10d","%r11d");
+$i0="%esi";
+$i1="%edi";
+$Tbl="%rbp";	# size optimization
+$inp="%r12";
+$out="%r13";
+$key="%r14";
+$keyend="%r15";
+$arg0d=$win64?"%ecx":"%edi";
+
+# const unsigned int Camellia_SBOX[4][256];
+# Well, sort of... Camellia_SBOX[0][] is interleaved with [1][],
+# and [2][] - with [3][]. This is done to minimize code size.
+$SBOX1_1110=0;		# Camellia_SBOX[0]
+$SBOX4_4404=4;		# Camellia_SBOX[1]
+$SBOX2_0222=2048;	# Camellia_SBOX[2]
+$SBOX3_3033=2052;	# Camellia_SBOX[3]
+
+sub Camellia_Feistel {
+my $i=@_[0];
+my $seed=defined(@_[1])?@_[1]:0;
+my $scale=$seed<0?-8:8;
+my $j=($i&1)*2;
+my $s0=@S[($j)%4],$s1=@S[($j+1)%4],$s2=@S[($j+2)%4],$s3=@S[($j+3)%4];
+
+$code.=<<___;
+	xor	$s0,$t0				# t0^=key[0]
+	xor	$s1,$t1				# t1^=key[1]
+	movz	`&hi("$t0")`,$i0		# (t0>>8)&0xff
+	movz	`&lo("$t1")`,$i1		# (t1>>0)&0xff
+	mov	$SBOX3_3033($Tbl,$i0,8),$t3	# t3=SBOX3_3033[0]
+	mov	$SBOX1_1110($Tbl,$i1,8),$t2	# t2=SBOX1_1110[1]
+	movz	`&lo("$t0")`,$i0		# (t0>>0)&0xff
+	shr	\$16,$t0
+	movz	`&hi("$t1")`,$i1		# (t1>>8)&0xff
+	xor	$SBOX4_4404($Tbl,$i0,8),$t3	# t3^=SBOX4_4404[0]
+	shr	\$16,$t1
+	xor	$SBOX4_4404($Tbl,$i1,8),$t2	# t2^=SBOX4_4404[1]
+	movz	`&hi("$t0")`,$i0		# (t0>>24)&0xff
+	movz	`&lo("$t1")`,$i1		# (t1>>16)&0xff
+	xor	$SBOX1_1110($Tbl,$i0,8),$t3	# t3^=SBOX1_1110[0]
+	xor	$SBOX3_3033($Tbl,$i1,8),$t2	# t2^=SBOX3_3033[1]
+	movz	`&lo("$t0")`,$i0		# (t0>>16)&0xff
+	movz	`&hi("$t1")`,$i1		# (t1>>24)&0xff
+	xor	$SBOX2_0222($Tbl,$i0,8),$t3	# t3^=SBOX2_0222[0]
+	xor	$SBOX2_0222($Tbl,$i1,8),$t2	# t2^=SBOX2_0222[1]
+	mov	`$seed+($i+1)*$scale`($key),$t1	# prefetch key[i+1]
+	mov	`$seed+($i+1)*$scale+4`($key),$t0
+	xor	$t3,$t2				# t2^=t3
+	ror	\$8,$t3				# t3=RightRotate(t3,8)
+	xor	$t2,$s2
+	xor	$t2,$s3
+	xor	$t3,$s3
+___
+}
+
+# void Camellia_EncryptBlock_Rounds(
+#		int grandRounds,
+#		const Byte plaintext[],
+#		const KEY_TABLE_TYPE keyTable,
+#		Byte ciphertext[])
+$code=<<___;
+.text
+
+# V1.x API
+.globl	Camellia_EncryptBlock
+.type	Camellia_EncryptBlock,\@abi-omnipotent
+.align	16
+Camellia_EncryptBlock:
+	movl	\$128,%eax
+	subl	$arg0d,%eax
+	movl	\$3,$arg0d
+	adcl	\$0,$arg0d	# keyBitLength==128?3:4
+	jmp	.Lenc_rounds
+.size	Camellia_EncryptBlock,.-Camellia_EncryptBlock
+# V2
+.globl	Camellia_EncryptBlock_Rounds
+.type	Camellia_EncryptBlock_Rounds,\@function,4
+.align	16
+.Lenc_rounds:
+Camellia_EncryptBlock_Rounds:
+	push	%rbx
+	push	%rbp
+	push	%r13
+	push	%r14
+	push	%r15
+.Lenc_prologue:
+
+	#mov	%rsi,$inp		# put away arguments
+	mov	%rcx,$out
+	mov	%rdx,$key
+
+	shl	\$6,%edi		# process grandRounds
+	lea	.LCamellia_SBOX(%rip),$Tbl
+	lea	($key,%rdi),$keyend
+
+	mov	0(%rsi),@S[0]		# load plaintext
+	mov	4(%rsi),@S[1]
+	mov	8(%rsi),@S[2]
+	bswap	@S[0]
+	mov	12(%rsi),@S[3]
+	bswap	@S[1]
+	bswap	@S[2]
+	bswap	@S[3]
+
+	call	_x86_64_Camellia_encrypt
+
+	bswap	@S[0]
+	bswap	@S[1]
+	bswap	@S[2]
+	mov	@S[0],0($out)
+	bswap	@S[3]
+	mov	@S[1],4($out)
+	mov	@S[2],8($out)
+	mov	@S[3],12($out)
+
+	mov	0(%rsp),%r15
+	mov	8(%rsp),%r14
+	mov	16(%rsp),%r13
+	mov	24(%rsp),%rbp
+	mov	32(%rsp),%rbx
+	lea	40(%rsp),%rsp
+.Lenc_epilogue:
+	ret
+.size	Camellia_EncryptBlock_Rounds,.-Camellia_EncryptBlock_Rounds
+
+.type	_x86_64_Camellia_encrypt,\@abi-omnipotent
+.align	16
+_x86_64_Camellia_encrypt:
+	xor	0($key),@S[1]
+	xor	4($key),@S[0]		# ^=key[0-3]
+	xor	8($key),@S[3]
+	xor	12($key),@S[2]
+.align	16
+.Leloop:
+	mov	16($key),$t1		# prefetch key[4-5]
+	mov	20($key),$t0
+
+___
+	for ($i=0;$i<6;$i++) { Camellia_Feistel($i,16); }
+$code.=<<___;
+	lea	16*4($key),$key
+	cmp	$keyend,$key
+	mov	8($key),$t3		# prefetch key[2-3]
+	mov	12($key),$t2
+	je	.Ledone
+
+	and	@S[0],$t0
+	or	@S[3],$t3
+	rol	\$1,$t0
+	xor	$t3,@S[2]		# s2^=s3|key[3];
+	xor	$t0,@S[1]		# s1^=LeftRotate(s0&key[0],1);
+	and	@S[2],$t2
+	or	@S[1],$t1
+	rol	\$1,$t2
+	xor	$t1,@S[0]		# s0^=s1|key[1];
+	xor	$t2,@S[3]		# s3^=LeftRotate(s2&key[2],1);
+	jmp	.Leloop
+
+.align	16
+.Ledone:
+	xor	@S[2],$t0		# SwapHalf
+	xor	@S[3],$t1
+	xor	@S[0],$t2
+	xor	@S[1],$t3
+
+	mov	$t0,@S[0]
+	mov	$t1,@S[1]
+	mov	$t2,@S[2]
+	mov	$t3,@S[3]
+
+	.byte	0xf3,0xc3		# rep ret
+.size	_x86_64_Camellia_encrypt,.-_x86_64_Camellia_encrypt
+
+# V1.x API
+.globl	Camellia_DecryptBlock
+.type	Camellia_DecryptBlock,\@abi-omnipotent
+.align	16
+Camellia_DecryptBlock:
+	movl	\$128,%eax
+	subl	$arg0d,%eax
+	movl	\$3,$arg0d
+	adcl	\$0,$arg0d	# keyBitLength==128?3:4
+	jmp	.Ldec_rounds
+.size	Camellia_DecryptBlock,.-Camellia_DecryptBlock
+# V2
+.globl	Camellia_DecryptBlock_Rounds
+.type	Camellia_DecryptBlock_Rounds,\@function,4
+.align	16
+.Ldec_rounds:
+Camellia_DecryptBlock_Rounds:
+	push	%rbx
+	push	%rbp
+	push	%r13
+	push	%r14
+	push	%r15
+.Ldec_prologue:
+
+	#mov	%rsi,$inp		# put away arguments
+	mov	%rcx,$out
+	mov	%rdx,$keyend
+
+	shl	\$6,%edi		# process grandRounds
+	lea	.LCamellia_SBOX(%rip),$Tbl
+	lea	($keyend,%rdi),$key
+
+	mov	0(%rsi),@S[0]		# load plaintext
+	mov	4(%rsi),@S[1]
+	mov	8(%rsi),@S[2]
+	bswap	@S[0]
+	mov	12(%rsi),@S[3]
+	bswap	@S[1]
+	bswap	@S[2]
+	bswap	@S[3]
+
+	call	_x86_64_Camellia_decrypt
+
+	bswap	@S[0]
+	bswap	@S[1]
+	bswap	@S[2]
+	mov	@S[0],0($out)
+	bswap	@S[3]
+	mov	@S[1],4($out)
+	mov	@S[2],8($out)
+	mov	@S[3],12($out)
+
+	mov	0(%rsp),%r15
+	mov	8(%rsp),%r14
+	mov	16(%rsp),%r13
+	mov	24(%rsp),%rbp
+	mov	32(%rsp),%rbx
+	lea	40(%rsp),%rsp
+.Ldec_epilogue:
+	ret
+.size	Camellia_DecryptBlock_Rounds,.-Camellia_DecryptBlock_Rounds
+
+.type	_x86_64_Camellia_decrypt,\@abi-omnipotent
+.align	16
+_x86_64_Camellia_decrypt:
+	xor	0($key),@S[1]
+	xor	4($key),@S[0]		# ^=key[0-3]
+	xor	8($key),@S[3]
+	xor	12($key),@S[2]
+.align	16
+.Ldloop:
+	mov	-8($key),$t1		# prefetch key[4-5]
+	mov	-4($key),$t0
+
+___
+	for ($i=0;$i<6;$i++) { Camellia_Feistel($i,-8); }
+$code.=<<___;
+	lea	-16*4($key),$key
+	cmp	$keyend,$key
+	mov	0($key),$t3		# prefetch key[2-3]
+	mov	4($key),$t2
+	je	.Lddone
+
+	and	@S[0],$t0
+	or	@S[3],$t3
+	rol	\$1,$t0
+	xor	$t3,@S[2]		# s2^=s3|key[3];
+	xor	$t0,@S[1]		# s1^=LeftRotate(s0&key[0],1);
+	and	@S[2],$t2
+	or	@S[1],$t1
+	rol	\$1,$t2
+	xor	$t1,@S[0]		# s0^=s1|key[1];
+	xor	$t2,@S[3]		# s3^=LeftRotate(s2&key[2],1);
+
+	jmp	.Ldloop
+
+.align	16
+.Lddone:
+	xor	@S[2],$t2
+	xor	@S[3],$t3
+	xor	@S[0],$t0
+	xor	@S[1],$t1
+
+	mov	$t2,@S[0]		# SwapHalf
+	mov	$t3,@S[1]
+	mov	$t0,@S[2]
+	mov	$t1,@S[3]
+
+	.byte	0xf3,0xc3		# rep ret
+.size	_x86_64_Camellia_decrypt,.-_x86_64_Camellia_decrypt
+___
+
+sub _saveround {
+my ($rnd,$key,@T)=@_;
+my $bias=int(@T[0])?shift(@T):0;
+
+    if ($#T==3) {
+	$code.=<<___;
+	mov	@T[1],`$bias+$rnd*8+0`($key)
+	mov	@T[0],`$bias+$rnd*8+4`($key)
+	mov	@T[3],`$bias+$rnd*8+8`($key)
+	mov	@T[2],`$bias+$rnd*8+12`($key)
+___
+    } else {
+	$code.="	mov	@T[0],`$bias+$rnd*8+0`($key)\n";
+	$code.="	mov	@T[1],`$bias+$rnd*8+8`($key)\n"	if ($#T>=1);
+    }
+}
+
+sub _loadround {
+my ($rnd,$key,@T)=@_;
+my $bias=int(@T[0])?shift(@T):0;
+
+$code.="	mov	`$bias+$rnd*8+0`($key),@T[0]\n";
+$code.="	mov	`$bias+$rnd*8+8`($key),@T[1]\n"	if ($#T>=1);
+}
+
+# shld is very slow on Intel EM64T family. Even on AMD it limits
+# instruction decode rate [because it's VectorPath] and consequently
+# performance...
+sub __rotl128 {
+my ($i0,$i1,$rot)=@_;
+
+    if ($rot) {
+	$code.=<<___;
+	mov	$i0,%r11
+	shld	\$$rot,$i1,$i0
+	shld	\$$rot,%r11,$i1
+___
+    }
+}
+
+# ... Implementing 128-bit rotate without shld gives 80% better
+# performance EM64T, +15% on AMD64 and only ~7% degradation on
+# Core2. This is therefore preferred.
+sub _rotl128 {
+my ($i0,$i1,$rot)=@_;
+
+    if ($rot) {
+	$code.=<<___;
+	mov	$i0,%r11
+	shl	\$$rot,$i0
+	mov	$i1,%r9
+	shr	\$`64-$rot`,%r9
+	shr	\$`64-$rot`,%r11
+	or	%r9,$i0
+	shl	\$$rot,$i1
+	or	%r11,$i1
+___
+    }
+}
+
+{ my $step=0;
+
+$code.=<<___;
+.globl	Camellia_Ekeygen
+.type	Camellia_Ekeygen,\@function,3
+.align	16
+Camellia_Ekeygen:
+	push	%rbx
+	push	%rbp
+	push	%r13
+	push	%r14
+	push	%r15
+.Lkey_prologue:
+
+	mov	%rdi,$keyend		# put away arguments, keyBitLength
+	mov	%rdx,$out		# keyTable
+
+	mov	0(%rsi),@S[0]		# load 0-127 bits
+	mov	4(%rsi),@S[1]
+	mov	8(%rsi),@S[2]
+	mov	12(%rsi),@S[3]
+
+	bswap	@S[0]
+	bswap	@S[1]
+	bswap	@S[2]
+	bswap	@S[3]
+___
+	&_saveround	(0,$out,@S);	# KL<<<0
+$code.=<<___;
+	cmp	\$128,$keyend		# check keyBitLength
+	je	.L1st128
+
+	mov	16(%rsi),@S[0]		# load 128-191 bits
+	mov	20(%rsi),@S[1]
+	cmp	\$192,$keyend
+	je	.L1st192
+	mov	24(%rsi),@S[2]		# load 192-255 bits
+	mov	28(%rsi),@S[3]
+	jmp	.L1st256
+.L1st192:
+	mov	@S[0],@S[2]
+	mov	@S[1],@S[3]
+	not	@S[2]
+	not	@S[3]
+.L1st256:
+	bswap	@S[0]
+	bswap	@S[1]
+	bswap	@S[2]
+	bswap	@S[3]
+___
+	&_saveround	(4,$out,@S);	# temp storage for KR!
+$code.=<<___;
+	xor	0($out),@S[1]		# KR^KL
+	xor	4($out),@S[0]
+	xor	8($out),@S[3]
+	xor	12($out),@S[2]
+
+.L1st128:
+	lea	.LCamellia_SIGMA(%rip),$key
+	lea	.LCamellia_SBOX(%rip),$Tbl
+
+	mov	0($key),$t1
+	mov	4($key),$t0
+___
+	&Camellia_Feistel($step++);
+	&Camellia_Feistel($step++);
+$code.=<<___;
+	xor	0($out),@S[1]		# ^KL
+	xor	4($out),@S[0]
+	xor	8($out),@S[3]
+	xor	12($out),@S[2]
+___
+	&Camellia_Feistel($step++);
+	&Camellia_Feistel($step++);
+$code.=<<___;
+	cmp	\$128,$keyend
+	jne	.L2nd256
+
+	lea	128($out),$out		# size optimization
+	shl	\$32,%r8		# @S[0]||
+	shl	\$32,%r10		# @S[2]||
+	or	%r9,%r8			# ||@S[1]
+	or	%r11,%r10		# ||@S[3]
+___
+	&_loadround	(0,$out,-128,"%rax","%rbx");	# KL
+	&_saveround	(2,$out,-128,"%r8","%r10");	# KA<<<0
+	&_rotl128	("%rax","%rbx",15);
+	&_saveround	(4,$out,-128,"%rax","%rbx");	# KL<<<15
+	&_rotl128	("%r8","%r10",15);
+	&_saveround	(6,$out,-128,"%r8","%r10");	# KA<<<15
+	&_rotl128	("%r8","%r10",15);		# 15+15=30
+	&_saveround	(8,$out,-128,"%r8","%r10");	# KA<<<30
+	&_rotl128	("%rax","%rbx",30);		# 15+30=45
+	&_saveround	(10,$out,-128,"%rax","%rbx");	# KL<<<45
+	&_rotl128	("%r8","%r10",15);		# 30+15=45
+	&_saveround	(12,$out,-128,"%r8");		# KA<<<45
+	&_rotl128	("%rax","%rbx",15);		# 45+15=60
+	&_saveround	(13,$out,-128,"%rbx");		# KL<<<60
+	&_rotl128	("%r8","%r10",15);		# 45+15=60
+	&_saveround	(14,$out,-128,"%r8","%r10");	# KA<<<60
+	&_rotl128	("%rax","%rbx",17);		# 60+17=77
+	&_saveround	(16,$out,-128,"%rax","%rbx");	# KL<<<77
+	&_rotl128	("%rax","%rbx",17);		# 77+17=94
+	&_saveround	(18,$out,-128,"%rax","%rbx");	# KL<<<94
+	&_rotl128	("%r8","%r10",34);		# 60+34=94
+	&_saveround	(20,$out,-128,"%r8","%r10");	# KA<<<94
+	&_rotl128	("%rax","%rbx",17);		# 94+17=111
+	&_saveround	(22,$out,-128,"%rax","%rbx");	# KL<<<111
+	&_rotl128	("%r8","%r10",17);		# 94+17=111
+	&_saveround	(24,$out,-128,"%r8","%r10");	# KA<<<111
+$code.=<<___;
+	mov	\$3,%eax
+	jmp	.Ldone
+.align	16
+.L2nd256:
+___
+	&_saveround	(6,$out,@S);	# temp storage for KA!
+$code.=<<___;
+	xor	`4*8+0`($out),@S[1]	# KA^KR
+	xor	`4*8+4`($out),@S[0]
+	xor	`5*8+0`($out),@S[3]
+	xor	`5*8+4`($out),@S[2]
+___
+	&Camellia_Feistel($step++);
+	&Camellia_Feistel($step++);
+
+	&_loadround	(0,$out,"%rax","%rbx");	# KL
+	&_loadround	(4,$out,"%rcx","%rdx");	# KR
+	&_loadround	(6,$out,"%r14","%r15");	# KA
+$code.=<<___;
+	lea	128($out),$out		# size optimization
+	shl	\$32,%r8		# @S[0]||
+	shl	\$32,%r10		# @S[2]||
+	or	%r9,%r8			# ||@S[1]
+	or	%r11,%r10		# ||@S[3]
+___
+	&_saveround	(2,$out,-128,"%r8","%r10");	# KB<<<0
+	&_rotl128	("%rcx","%rdx",15);
+	&_saveround	(4,$out,-128,"%rcx","%rdx");	# KR<<<15
+	&_rotl128	("%r14","%r15",15);
+	&_saveround	(6,$out,-128,"%r14","%r15");	# KA<<<15
+	&_rotl128	("%rcx","%rdx",15);		# 15+15=30
+	&_saveround	(8,$out,-128,"%rcx","%rdx");	# KR<<<30
+	&_rotl128	("%r8","%r10",30);
+	&_saveround	(10,$out,-128,"%r8","%r10");	# KB<<<30
+	&_rotl128	("%rax","%rbx",45);
+	&_saveround	(12,$out,-128,"%rax","%rbx");	# KL<<<45
+	&_rotl128	("%r14","%r15",30);		# 15+30=45
+	&_saveround	(14,$out,-128,"%r14","%r15");	# KA<<<45
+	&_rotl128	("%rax","%rbx",15);		# 45+15=60
+	&_saveround	(16,$out,-128,"%rax","%rbx");	# KL<<<60
+	&_rotl128	("%rcx","%rdx",30);		# 30+30=60
+	&_saveround	(18,$out,-128,"%rcx","%rdx");	# KR<<<60
+	&_rotl128	("%r8","%r10",30);		# 30+30=60
+	&_saveround	(20,$out,-128,"%r8","%r10");	# KB<<<60
+	&_rotl128	("%rax","%rbx",17);		# 60+17=77
+	&_saveround	(22,$out,-128,"%rax","%rbx");	# KL<<<77
+	&_rotl128	("%r14","%r15",32);		# 45+32=77
+	&_saveround	(24,$out,-128,"%r14","%r15");	# KA<<<77
+	&_rotl128	("%rcx","%rdx",34);		# 60+34=94
+	&_saveround	(26,$out,-128,"%rcx","%rdx");	# KR<<<94
+	&_rotl128	("%r14","%r15",17);		# 77+17=94
+	&_saveround	(28,$out,-128,"%r14","%r15");	# KA<<<77
+	&_rotl128	("%rax","%rbx",34);		# 77+34=111
+	&_saveround	(30,$out,-128,"%rax","%rbx");	# KL<<<111
+	&_rotl128	("%r8","%r10",51);		# 60+51=111
+	&_saveround	(32,$out,-128,"%r8","%r10");	# KB<<<111
+$code.=<<___;
+	mov	\$4,%eax
+.Ldone:
+	mov	0(%rsp),%r15
+	mov	8(%rsp),%r14
+	mov	16(%rsp),%r13
+	mov	24(%rsp),%rbp
+	mov	32(%rsp),%rbx
+	lea	40(%rsp),%rsp
+.Lkey_epilogue:
+	ret
+.size	Camellia_Ekeygen,.-Camellia_Ekeygen
+___
+}
+
+@SBOX=(
+112,130, 44,236,179, 39,192,229,228,133, 87, 53,234, 12,174, 65,
+ 35,239,107,147, 69, 25,165, 33,237, 14, 79, 78, 29,101,146,189,
+134,184,175,143,124,235, 31,206, 62, 48,220, 95, 94,197, 11, 26,
+166,225, 57,202,213, 71, 93, 61,217,  1, 90,214, 81, 86,108, 77,
+139, 13,154,102,251,204,176, 45,116, 18, 43, 32,240,177,132,153,
+223, 76,203,194, 52,126,118,  5,109,183,169, 49,209, 23,  4,215,
+ 20, 88, 58, 97,222, 27, 17, 28, 50, 15,156, 22, 83, 24,242, 34,
+254, 68,207,178,195,181,122,145, 36,  8,232,168, 96,252,105, 80,
+170,208,160,125,161,137, 98,151, 84, 91, 30,149,224,255,100,210,
+ 16,196,  0, 72,163,247,117,219,138,  3,230,218,  9, 63,221,148,
+135, 92,131,  2,205, 74,144, 51,115,103,246,243,157,127,191,226,
+ 82,155,216, 38,200, 55,198, 59,129,150,111, 75, 19,190, 99, 46,
+233,121,167,140,159,110,188,142, 41,245,249,182, 47,253,180, 89,
+120,152,  6,106,231, 70,113,186,212, 37,171, 66,136,162,141,250,
+114,  7,185, 85,248,238,172, 10, 54, 73, 42,104, 60, 56,241,164,
+ 64, 40,211,123,187,201, 67,193, 21,227,173,244,119,199,128,158);
+
+sub S1110 { my $i=shift; $i=@SBOX[$i]; $i=$i<<24|$i<<16|$i<<8; sprintf("0x%08x",$i); }
+sub S4404 { my $i=shift; $i=($i<<1|$i>>7)&0xff; $i=@SBOX[$i]; $i=$i<<24|$i<<16|$i; sprintf("0x%08x",$i); }
+sub S0222 { my $i=shift; $i=@SBOX[$i]; $i=($i<<1|$i>>7)&0xff; $i=$i<<16|$i<<8|$i; sprintf("0x%08x",$i); }
+sub S3033 { my $i=shift; $i=@SBOX[$i]; $i=($i>>1|$i<<7)&0xff; $i=$i<<24|$i<<8|$i; sprintf("0x%08x",$i); }
+
+$code.=<<___;
+.align	64
+.LCamellia_SIGMA:
+.long	0x3bcc908b, 0xa09e667f, 0x4caa73b2, 0xb67ae858
+.long	0xe94f82be, 0xc6ef372f, 0xf1d36f1c, 0x54ff53a5
+.long	0xde682d1d, 0x10e527fa, 0xb3e6c1fd, 0xb05688c2
+.long	0,          0,          0,          0
+.LCamellia_SBOX:
+___
+# tables are interleaved, remember?
+sub data_word { $code.=".long\t".join(',',@_)."\n"; }
+for ($i=0;$i<256;$i++) { &data_word(&S1110($i),&S4404($i)); }
+for ($i=0;$i<256;$i++) { &data_word(&S0222($i),&S3033($i)); }
+
+# void Camellia_cbc_encrypt (const void char *inp, unsigned char *out,
+#			size_t length, const CAMELLIA_KEY *key,
+#			unsigned char *ivp,const int enc);
+{
+$_key="0(%rsp)";
+$_end="8(%rsp)";	# inp+len&~15
+$_res="16(%rsp)";	# len&15
+$ivec="24(%rsp)";
+$_ivp="40(%rsp)";
+$_rsp="48(%rsp)";
+
+$code.=<<___;
+.globl	Camellia_cbc_encrypt
+.type	Camellia_cbc_encrypt,\@function,6
+.align	16
+Camellia_cbc_encrypt:
+	cmp	\$0,%rdx
+	je	.Lcbc_abort
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+.Lcbc_prologue:
+
+	mov	%rsp,%rbp
+	sub	\$64,%rsp
+	and	\$-64,%rsp
+
+	# place stack frame just "above mod 1024" the key schedule,
+	# this ensures that cache associativity suffices
+	lea	-64-63(%rcx),%r10
+	sub	%rsp,%r10
+	neg	%r10
+	and	\$0x3C0,%r10
+	sub	%r10,%rsp
+	#add	\$8,%rsp		# 8 is reserved for callee's ra
+
+	mov	%rdi,$inp		# inp argument
+	mov	%rsi,$out		# out argument
+	mov	%r8,%rbx		# ivp argument
+	mov	%rcx,$key		# key argument
+	mov	272(%rcx),${keyend}d	# grandRounds
+
+	mov	%r8,$_ivp
+	mov	%rbp,$_rsp
+
+.Lcbc_body:
+	lea	.LCamellia_SBOX(%rip),$Tbl
+
+	mov	\$32,%ecx
+.align	4
+.Lcbc_prefetch_sbox:
+	mov	0($Tbl),%rax
+	mov	32($Tbl),%rsi
+	mov	64($Tbl),%rdi
+	mov	96($Tbl),%r11
+	lea	128($Tbl),$Tbl
+	loop	.Lcbc_prefetch_sbox
+	sub	\$4096,$Tbl
+	shl	\$6,$keyend
+	mov	%rdx,%rcx		# len argument
+	lea	($key,$keyend),$keyend
+
+	cmp	\$0,%r9d		# enc argument
+	je	.LCBC_DECRYPT
+
+	and	\$-16,%rdx
+	and	\$15,%rcx		# length residue
+	lea	($inp,%rdx),%rdx
+	mov	$key,$_key
+	mov	%rdx,$_end
+	mov	%rcx,$_res
+
+	cmp	$inp,%rdx
+	mov	0(%rbx),@S[0]		# load IV
+	mov	4(%rbx),@S[1]
+	mov	8(%rbx),@S[2]
+	mov	12(%rbx),@S[3]
+	je	.Lcbc_enc_tail
+	jmp	.Lcbc_eloop
+
+.align	16
+.Lcbc_eloop:
+	xor	0($inp),@S[0]
+	xor	4($inp),@S[1]
+	xor	8($inp),@S[2]
+	bswap	@S[0]
+	xor	12($inp),@S[3]
+	bswap	@S[1]
+	bswap	@S[2]
+	bswap	@S[3]
+
+	call	_x86_64_Camellia_encrypt
+
+	mov	$_key,$key		# "rewind" the key
+	bswap	@S[0]
+	mov	$_end,%rdx
+	bswap	@S[1]
+	mov	$_res,%rcx
+	bswap	@S[2]
+	mov	@S[0],0($out)
+	bswap	@S[3]
+	mov	@S[1],4($out)
+	mov	@S[2],8($out)
+	lea	16($inp),$inp
+	mov	@S[3],12($out)
+	cmp	%rdx,$inp
+	lea	16($out),$out
+	jne	.Lcbc_eloop
+
+	cmp	\$0,%rcx
+	jne	.Lcbc_enc_tail
+
+	mov	$_ivp,$out
+	mov	@S[0],0($out)		# write out IV residue
+	mov	@S[1],4($out)
+	mov	@S[2],8($out)
+	mov	@S[3],12($out)
+	jmp	.Lcbc_done
+
+.align	16
+.Lcbc_enc_tail:
+	xor	%rax,%rax
+	mov	%rax,0+$ivec
+	mov	%rax,8+$ivec
+	mov	%rax,$_res
+
+.Lcbc_enc_pushf:
+	pushfq
+	cld
+	mov	$inp,%rsi
+	lea	8+$ivec,%rdi
+	.long	0x9066A4F3		# rep movsb
+	popfq
+.Lcbc_enc_popf:
+
+	lea	$ivec,$inp
+	lea	16+$ivec,%rax
+	mov	%rax,$_end
+	jmp	.Lcbc_eloop		# one more time
+
+.align	16
+.LCBC_DECRYPT:
+	xchg	$key,$keyend
+	add	\$15,%rdx
+	and	\$15,%rcx		# length residue
+	and	\$-16,%rdx
+	mov	$key,$_key
+	lea	($inp,%rdx),%rdx
+	mov	%rdx,$_end
+	mov	%rcx,$_res
+
+	mov	(%rbx),%rax		# load IV
+	mov	8(%rbx),%rbx
+	jmp	.Lcbc_dloop
+.align	16
+.Lcbc_dloop:
+	mov	0($inp),@S[0]
+	mov	4($inp),@S[1]
+	mov	8($inp),@S[2]
+	bswap	@S[0]
+	mov	12($inp),@S[3]
+	bswap	@S[1]
+	mov	%rax,0+$ivec		# save IV to temporary storage
+	bswap	@S[2]
+	mov	%rbx,8+$ivec
+	bswap	@S[3]
+
+	call	_x86_64_Camellia_decrypt
+
+	mov	$_key,$key		# "rewind" the key
+	mov	$_end,%rdx
+	mov	$_res,%rcx
+
+	bswap	@S[0]
+	mov	($inp),%rax		# load IV for next iteration
+	bswap	@S[1]
+	mov	8($inp),%rbx
+	bswap	@S[2]
+	xor	0+$ivec,@S[0]
+	bswap	@S[3]
+	xor	4+$ivec,@S[1]
+	xor	8+$ivec,@S[2]
+	lea	16($inp),$inp
+	xor	12+$ivec,@S[3]
+	cmp	%rdx,$inp
+	je	.Lcbc_ddone
+
+	mov	@S[0],0($out)
+	mov	@S[1],4($out)
+	mov	@S[2],8($out)
+	mov	@S[3],12($out)
+
+	lea	16($out),$out
+	jmp	.Lcbc_dloop
+
+.align	16
+.Lcbc_ddone:
+	mov	$_ivp,%rdx
+	cmp	\$0,%rcx
+	jne	.Lcbc_dec_tail
+
+	mov	@S[0],0($out)
+	mov	@S[1],4($out)
+	mov	@S[2],8($out)
+	mov	@S[3],12($out)
+
+	mov	%rax,(%rdx)		# write out IV residue
+	mov	%rbx,8(%rdx)
+	jmp	.Lcbc_done
+.align	16
+.Lcbc_dec_tail:
+	mov	@S[0],0+$ivec
+	mov	@S[1],4+$ivec
+	mov	@S[2],8+$ivec
+	mov	@S[3],12+$ivec
+
+.Lcbc_dec_pushf:
+	pushfq
+	cld
+	lea	8+$ivec,%rsi
+	lea	($out),%rdi
+	.long	0x9066A4F3		# rep movsb
+	popfq
+.Lcbc_dec_popf:
+
+	mov	%rax,(%rdx)		# write out IV residue
+	mov	%rbx,8(%rdx)
+	jmp	.Lcbc_done
+
+.align	16
+.Lcbc_done:
+	mov	$_rsp,%rcx
+	mov	0(%rcx),%r15
+	mov	8(%rcx),%r14
+	mov	16(%rcx),%r13
+	mov	24(%rcx),%r12
+	mov	32(%rcx),%rbp
+	mov	40(%rcx),%rbx
+	lea	48(%rcx),%rsp
+.Lcbc_abort:
+	ret
+.size	Camellia_cbc_encrypt,.-Camellia_cbc_encrypt
+
+.asciz	"Camellia for x86_64 by <appro\@openssl.org>"
+___
+}
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	common_se_handler,\@abi-omnipotent
+.align	16
+common_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	lea	-64(%rsp),%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lin_prologue
+
+	lea	40(%rax),%rax
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r13
+	mov	-32(%rax),%r14
+	mov	-40(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	jmp	.Lcommon_seh_exit
+.size	common_se_handler,.-common_se_handler
+
+.type	cbc_se_handler,\@abi-omnipotent
+.align	16
+cbc_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	lea	-64(%rsp),%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lcbc_prologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lcbc_prologue
+	jb	.Lin_cbc_prologue
+
+	lea	.Lcbc_body(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lcbc_body
+	jb	.Lin_cbc_frame_setup
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lcbc_abort(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lcbc_abort
+	jae	.Lin_cbc_prologue
+
+	# handle pushf/popf in Camellia_cbc_encrypt
+	lea	.Lcbc_enc_pushf(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<=.Lcbc_enc_pushf
+	jbe	.Lin_cbc_no_flag
+	lea	8(%rax),%rax
+	lea	.Lcbc_enc_popf(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lcbc_enc_popf
+	jb	.Lin_cbc_no_flag
+	lea	-8(%rax),%rax
+	lea	.Lcbc_dec_pushf(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<=.Lcbc_dec_pushf
+	jbe	.Lin_cbc_no_flag
+	lea	8(%rax),%rax
+	lea	.Lcbc_dec_popf(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lcbc_dec_popf
+	jb	.Lin_cbc_no_flag
+	lea	-8(%rax),%rax
+
+.Lin_cbc_no_flag:
+	mov	48(%rax),%rax		# $_rsp
+	lea	48(%rax),%rax
+
+.Lin_cbc_frame_setup:
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	-40(%rax),%r14
+	mov	-48(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_cbc_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+.align	4
+.Lcommon_seh_exit:
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$`1232/8`,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	lea	64(%rsp),%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	cbc_se_handler,.-cbc_se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_Camellia_EncryptBlock_Rounds
+	.rva	.LSEH_end_Camellia_EncryptBlock_Rounds
+	.rva	.LSEH_info_Camellia_EncryptBlock_Rounds
+
+	.rva	.LSEH_begin_Camellia_DecryptBlock_Rounds
+	.rva	.LSEH_end_Camellia_DecryptBlock_Rounds
+	.rva	.LSEH_info_Camellia_DecryptBlock_Rounds
+
+	.rva	.LSEH_begin_Camellia_Ekeygen
+	.rva	.LSEH_end_Camellia_Ekeygen
+	.rva	.LSEH_info_Camellia_Ekeygen
+
+	.rva	.LSEH_begin_Camellia_cbc_encrypt
+	.rva	.LSEH_end_Camellia_cbc_encrypt
+	.rva	.LSEH_info_Camellia_cbc_encrypt
+
+.section	.xdata
+.align	8
+.LSEH_info_Camellia_EncryptBlock_Rounds:
+	.byte	9,0,0,0
+	.rva	common_se_handler
+	.rva	.Lenc_prologue,.Lenc_epilogue	# HandlerData[]
+.LSEH_info_Camellia_DecryptBlock_Rounds:
+	.byte	9,0,0,0
+	.rva	common_se_handler
+	.rva	.Ldec_prologue,.Ldec_epilogue	# HandlerData[]
+.LSEH_info_Camellia_Ekeygen:
+	.byte	9,0,0,0
+	.rva	common_se_handler
+	.rva	.Lkey_prologue,.Lkey_epilogue	# HandlerData[]
+.LSEH_info_Camellia_cbc_encrypt:
+	.byte	9,0,0,0
+	.rva	cbc_se_handler
+___
+}
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/camellia/camellia.c b/crypto/camellia/camellia.c
index 491c26b39e96..75fc8991c087 100644
--- a/crypto/camellia/camellia.c
+++ b/crypto/camellia/camellia.c
@@ -68,1557 +68,515 @@
 /* Algorithm Specification 
    http://info.isl.ntt.co.jp/crypt/eng/camellia/specifications.html
 */
-
-
-#include <string.h>
-#include <stdlib.h>
+ 
+/*
+ * This release balances code size and performance. In particular key
+ * schedule setup is fully unrolled, because doing so *significantly*
+ * reduces amount of instructions per setup round and code increase is
+ * justifiable. In block functions on the other hand only inner loops
+ * are unrolled, as full unroll gives only nominal performance boost,
+ * while code size grows 4 or 7 times. Also, unlike previous versions
+ * this one "encourages" compiler to keep intermediate variables in
+ * registers, which should give better "all round" results, in other
+ * words reasonable performance even with not so modern compilers.
+ */
 
 #include "camellia.h"
 #include "cmll_locl.h"
+#include <string.h>
+#include <stdlib.h>
 
-/* key constants */
-#define CAMELLIA_SIGMA1L (0xA09E667FL)
-#define CAMELLIA_SIGMA1R (0x3BCC908BL)
-#define CAMELLIA_SIGMA2L (0xB67AE858L)
-#define CAMELLIA_SIGMA2R (0x4CAA73B2L)
-#define CAMELLIA_SIGMA3L (0xC6EF372FL)
-#define CAMELLIA_SIGMA3R (0xE94F82BEL)
-#define CAMELLIA_SIGMA4L (0x54FF53A5L)
-#define CAMELLIA_SIGMA4R (0xF1D36F1CL)
-#define CAMELLIA_SIGMA5L (0x10E527FAL)
-#define CAMELLIA_SIGMA5R (0xDE682D1DL)
-#define CAMELLIA_SIGMA6L (0xB05688C2L)
-#define CAMELLIA_SIGMA6R (0xB3E6C1FDL)
-
+/* 32-bit rotations */
+#if !defined(PEDANTIC) && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
+# if defined(_MSC_VER) && (defined(_M_IX86) || defined(_M_AMD64) || defined(_M_X64))
+#  define RightRotate(x, s) _lrotr(x, s)
+#  define LeftRotate(x, s)  _lrotl(x, s)
+#  if _MSC_VER >= 1400
+#   define SWAP(x) _byteswap_ulong(x)
+#  else
+#   define SWAP(x) (_lrotl(x, 8) & 0x00ff00ff | _lrotr(x, 8) & 0xff00ff00)
+#  endif
+#  define GETU32(p)   SWAP(*((u32 *)(p)))
+#  define PUTU32(p,v) (*((u32 *)(p)) = SWAP((v)))
+# elif defined(__GNUC__) && __GNUC__>=2
+#  if defined(__i386) || defined(__x86_64)
+#   define RightRotate(x,s) ({u32 ret; asm ("rorl %1,%0":"=r"(ret):"I"(s),"0"(x):"cc"); ret; })
+#   define LeftRotate(x,s)  ({u32 ret; asm ("roll %1,%0":"=r"(ret):"I"(s),"0"(x):"cc"); ret; })
+#   if defined(B_ENDIAN) /* stratus.com does it */
+#    define GETU32(p)   (*(u32 *)(p))
+#    define PUTU32(p,v) (*(u32 *)(p)=(v))
+#   else
+#    define GETU32(p)   ({u32 r=*(const u32 *)(p); asm("bswapl %0":"=r"(r):"0"(r)); r; })
+#    define PUTU32(p,v) ({u32 r=(v); asm("bswapl %0":"=r"(r):"0"(r)); *(u32 *)(p)=r; })
+#   endif
+#  elif defined(_ARCH_PPC) || defined(_ARCH_PPC64) || \
+        defined(__powerpc) || defined(__ppc__) || defined(__powerpc64__)
+#   define LeftRotate(x,s)  ({u32 ret; asm ("rlwinm %0,%1,%2,0,31":"=r"(ret):"r"(x),"I"(s)); ret; })
+#   define RightRotate(x,s) LeftRotate(x,(32-s))
+#  elif defined(__s390x__)
+#   define LeftRotate(x,s)  ({u32 ret; asm ("rll %0,%1,%2":"=r"(ret):"r"(x),"I"(s)); ret; })
+#   define RightRotate(x,s) LeftRotate(x,(32-s))
+#   define GETU32(p)   (*(u32 *)(p))
+#   define PUTU32(p,v) (*(u32 *)(p)=(v))
+#  endif
+# endif
+#endif
+
+#if !defined(RightRotate) && !defined(LeftRotate)
+# define RightRotate(x, s) ( ((x) >> (s)) + ((x) << (32 - s)) )
+# define LeftRotate(x, s)  ( ((x) << (s)) + ((x) >> (32 - s)) )
+#endif
+
+#if !defined(GETU32) && !defined(PUTU32)
+# define GETU32(p)   (((u32)(p)[0] << 24) ^ ((u32)(p)[1] << 16) ^ ((u32)(p)[2] <<  8) ^ ((u32)(p)[3]))
+# define PUTU32(p,v) ((p)[0] = (u8)((v) >> 24), (p)[1] = (u8)((v) >> 16), (p)[2] = (u8)((v) >>  8), (p)[3] = (u8)(v))
+#endif
+
+/* S-box data */
+#define SBOX1_1110 Camellia_SBOX[0]
+#define SBOX4_4404 Camellia_SBOX[1]
+#define SBOX2_0222 Camellia_SBOX[2]
+#define SBOX3_3033 Camellia_SBOX[3]
+static const u32 Camellia_SBOX[][256] = {
+{   0x70707000, 0x82828200, 0x2c2c2c00, 0xececec00, 0xb3b3b300, 0x27272700, 
+    0xc0c0c000, 0xe5e5e500, 0xe4e4e400, 0x85858500, 0x57575700, 0x35353500, 
+    0xeaeaea00, 0x0c0c0c00, 0xaeaeae00, 0x41414100, 0x23232300, 0xefefef00, 
+    0x6b6b6b00, 0x93939300, 0x45454500, 0x19191900, 0xa5a5a500, 0x21212100, 
+    0xededed00, 0x0e0e0e00, 0x4f4f4f00, 0x4e4e4e00, 0x1d1d1d00, 0x65656500, 
+    0x92929200, 0xbdbdbd00, 0x86868600, 0xb8b8b800, 0xafafaf00, 0x8f8f8f00, 
+    0x7c7c7c00, 0xebebeb00, 0x1f1f1f00, 0xcecece00, 0x3e3e3e00, 0x30303000, 
+    0xdcdcdc00, 0x5f5f5f00, 0x5e5e5e00, 0xc5c5c500, 0x0b0b0b00, 0x1a1a1a00, 
+    0xa6a6a600, 0xe1e1e100, 0x39393900, 0xcacaca00, 0xd5d5d500, 0x47474700, 
+    0x5d5d5d00, 0x3d3d3d00, 0xd9d9d900, 0x01010100, 0x5a5a5a00, 0xd6d6d600, 
+    0x51515100, 0x56565600, 0x6c6c6c00, 0x4d4d4d00, 0x8b8b8b00, 0x0d0d0d00, 
+    0x9a9a9a00, 0x66666600, 0xfbfbfb00, 0xcccccc00, 0xb0b0b000, 0x2d2d2d00, 
+    0x74747400, 0x12121200, 0x2b2b2b00, 0x20202000, 0xf0f0f000, 0xb1b1b100, 
+    0x84848400, 0x99999900, 0xdfdfdf00, 0x4c4c4c00, 0xcbcbcb00, 0xc2c2c200, 
+    0x34343400, 0x7e7e7e00, 0x76767600, 0x05050500, 0x6d6d6d00, 0xb7b7b700, 
+    0xa9a9a900, 0x31313100, 0xd1d1d100, 0x17171700, 0x04040400, 0xd7d7d700, 
+    0x14141400, 0x58585800, 0x3a3a3a00, 0x61616100, 0xdedede00, 0x1b1b1b00, 
+    0x11111100, 0x1c1c1c00, 0x32323200, 0x0f0f0f00, 0x9c9c9c00, 0x16161600, 
+    0x53535300, 0x18181800, 0xf2f2f200, 0x22222200, 0xfefefe00, 0x44444400, 
+    0xcfcfcf00, 0xb2b2b200, 0xc3c3c300, 0xb5b5b500, 0x7a7a7a00, 0x91919100, 
+    0x24242400, 0x08080800, 0xe8e8e800, 0xa8a8a800, 0x60606000, 0xfcfcfc00, 
+    0x69696900, 0x50505000, 0xaaaaaa00, 0xd0d0d000, 0xa0a0a000, 0x7d7d7d00, 
+    0xa1a1a100, 0x89898900, 0x62626200, 0x97979700, 0x54545400, 0x5b5b5b00, 
+    0x1e1e1e00, 0x95959500, 0xe0e0e000, 0xffffff00, 0x64646400, 0xd2d2d200, 
+    0x10101000, 0xc4c4c400, 0x00000000, 0x48484800, 0xa3a3a300, 0xf7f7f700, 
+    0x75757500, 0xdbdbdb00, 0x8a8a8a00, 0x03030300, 0xe6e6e600, 0xdadada00, 
+    0x09090900, 0x3f3f3f00, 0xdddddd00, 0x94949400, 0x87878700, 0x5c5c5c00, 
+    0x83838300, 0x02020200, 0xcdcdcd00, 0x4a4a4a00, 0x90909000, 0x33333300, 
+    0x73737300, 0x67676700, 0xf6f6f600, 0xf3f3f300, 0x9d9d9d00, 0x7f7f7f00, 
+    0xbfbfbf00, 0xe2e2e200, 0x52525200, 0x9b9b9b00, 0xd8d8d800, 0x26262600, 
+    0xc8c8c800, 0x37373700, 0xc6c6c600, 0x3b3b3b00, 0x81818100, 0x96969600, 
+    0x6f6f6f00, 0x4b4b4b00, 0x13131300, 0xbebebe00, 0x63636300, 0x2e2e2e00, 
+    0xe9e9e900, 0x79797900, 0xa7a7a700, 0x8c8c8c00, 0x9f9f9f00, 0x6e6e6e00, 
+    0xbcbcbc00, 0x8e8e8e00, 0x29292900, 0xf5f5f500, 0xf9f9f900, 0xb6b6b600, 
+    0x2f2f2f00, 0xfdfdfd00, 0xb4b4b400, 0x59595900, 0x78787800, 0x98989800, 
+    0x06060600, 0x6a6a6a00, 0xe7e7e700, 0x46464600, 0x71717100, 0xbababa00, 
+    0xd4d4d400, 0x25252500, 0xababab00, 0x42424200, 0x88888800, 0xa2a2a200, 
+    0x8d8d8d00, 0xfafafa00, 0x72727200, 0x07070700, 0xb9b9b900, 0x55555500, 
+    0xf8f8f800, 0xeeeeee00, 0xacacac00, 0x0a0a0a00, 0x36363600, 0x49494900, 
+    0x2a2a2a00, 0x68686800, 0x3c3c3c00, 0x38383800, 0xf1f1f100, 0xa4a4a400, 
+    0x40404000, 0x28282800, 0xd3d3d300, 0x7b7b7b00, 0xbbbbbb00, 0xc9c9c900, 
+    0x43434300, 0xc1c1c100, 0x15151500, 0xe3e3e300, 0xadadad00, 0xf4f4f400, 
+    0x77777700, 0xc7c7c700, 0x80808000, 0x9e9e9e00 },
+{   0x70700070, 0x2c2c002c, 0xb3b300b3, 0xc0c000c0, 0xe4e400e4, 0x57570057, 
+    0xeaea00ea, 0xaeae00ae, 0x23230023, 0x6b6b006b, 0x45450045, 0xa5a500a5, 
+    0xeded00ed, 0x4f4f004f, 0x1d1d001d, 0x92920092, 0x86860086, 0xafaf00af, 
+    0x7c7c007c, 0x1f1f001f, 0x3e3e003e, 0xdcdc00dc, 0x5e5e005e, 0x0b0b000b, 
+    0xa6a600a6, 0x39390039, 0xd5d500d5, 0x5d5d005d, 0xd9d900d9, 0x5a5a005a, 
+    0x51510051, 0x6c6c006c, 0x8b8b008b, 0x9a9a009a, 0xfbfb00fb, 0xb0b000b0, 
+    0x74740074, 0x2b2b002b, 0xf0f000f0, 0x84840084, 0xdfdf00df, 0xcbcb00cb, 
+    0x34340034, 0x76760076, 0x6d6d006d, 0xa9a900a9, 0xd1d100d1, 0x04040004, 
+    0x14140014, 0x3a3a003a, 0xdede00de, 0x11110011, 0x32320032, 0x9c9c009c, 
+    0x53530053, 0xf2f200f2, 0xfefe00fe, 0xcfcf00cf, 0xc3c300c3, 0x7a7a007a, 
+    0x24240024, 0xe8e800e8, 0x60600060, 0x69690069, 0xaaaa00aa, 0xa0a000a0, 
+    0xa1a100a1, 0x62620062, 0x54540054, 0x1e1e001e, 0xe0e000e0, 0x64640064, 
+    0x10100010, 0x00000000, 0xa3a300a3, 0x75750075, 0x8a8a008a, 0xe6e600e6, 
+    0x09090009, 0xdddd00dd, 0x87870087, 0x83830083, 0xcdcd00cd, 0x90900090, 
+    0x73730073, 0xf6f600f6, 0x9d9d009d, 0xbfbf00bf, 0x52520052, 0xd8d800d8, 
+    0xc8c800c8, 0xc6c600c6, 0x81810081, 0x6f6f006f, 0x13130013, 0x63630063, 
+    0xe9e900e9, 0xa7a700a7, 0x9f9f009f, 0xbcbc00bc, 0x29290029, 0xf9f900f9, 
+    0x2f2f002f, 0xb4b400b4, 0x78780078, 0x06060006, 0xe7e700e7, 0x71710071, 
+    0xd4d400d4, 0xabab00ab, 0x88880088, 0x8d8d008d, 0x72720072, 0xb9b900b9, 
+    0xf8f800f8, 0xacac00ac, 0x36360036, 0x2a2a002a, 0x3c3c003c, 0xf1f100f1, 
+    0x40400040, 0xd3d300d3, 0xbbbb00bb, 0x43430043, 0x15150015, 0xadad00ad, 
+    0x77770077, 0x80800080, 0x82820082, 0xecec00ec, 0x27270027, 0xe5e500e5, 
+    0x85850085, 0x35350035, 0x0c0c000c, 0x41410041, 0xefef00ef, 0x93930093, 
+    0x19190019, 0x21210021, 0x0e0e000e, 0x4e4e004e, 0x65650065, 0xbdbd00bd, 
+    0xb8b800b8, 0x8f8f008f, 0xebeb00eb, 0xcece00ce, 0x30300030, 0x5f5f005f, 
+    0xc5c500c5, 0x1a1a001a, 0xe1e100e1, 0xcaca00ca, 0x47470047, 0x3d3d003d, 
+    0x01010001, 0xd6d600d6, 0x56560056, 0x4d4d004d, 0x0d0d000d, 0x66660066, 
+    0xcccc00cc, 0x2d2d002d, 0x12120012, 0x20200020, 0xb1b100b1, 0x99990099, 
+    0x4c4c004c, 0xc2c200c2, 0x7e7e007e, 0x05050005, 0xb7b700b7, 0x31310031, 
+    0x17170017, 0xd7d700d7, 0x58580058, 0x61610061, 0x1b1b001b, 0x1c1c001c, 
+    0x0f0f000f, 0x16160016, 0x18180018, 0x22220022, 0x44440044, 0xb2b200b2, 
+    0xb5b500b5, 0x91910091, 0x08080008, 0xa8a800a8, 0xfcfc00fc, 0x50500050, 
+    0xd0d000d0, 0x7d7d007d, 0x89890089, 0x97970097, 0x5b5b005b, 0x95950095, 
+    0xffff00ff, 0xd2d200d2, 0xc4c400c4, 0x48480048, 0xf7f700f7, 0xdbdb00db, 
+    0x03030003, 0xdada00da, 0x3f3f003f, 0x94940094, 0x5c5c005c, 0x02020002, 
+    0x4a4a004a, 0x33330033, 0x67670067, 0xf3f300f3, 0x7f7f007f, 0xe2e200e2, 
+    0x9b9b009b, 0x26260026, 0x37370037, 0x3b3b003b, 0x96960096, 0x4b4b004b, 
+    0xbebe00be, 0x2e2e002e, 0x79790079, 0x8c8c008c, 0x6e6e006e, 0x8e8e008e, 
+    0xf5f500f5, 0xb6b600b6, 0xfdfd00fd, 0x59590059, 0x98980098, 0x6a6a006a, 
+    0x46460046, 0xbaba00ba, 0x25250025, 0x42420042, 0xa2a200a2, 0xfafa00fa, 
+    0x07070007, 0x55550055, 0xeeee00ee, 0x0a0a000a, 0x49490049, 0x68680068, 
+    0x38380038, 0xa4a400a4, 0x28280028, 0x7b7b007b, 0xc9c900c9, 0xc1c100c1, 
+    0xe3e300e3, 0xf4f400f4, 0xc7c700c7, 0x9e9e009e },
+{   0x00e0e0e0, 0x00050505, 0x00585858, 0x00d9d9d9, 0x00676767, 0x004e4e4e, 
+    0x00818181, 0x00cbcbcb, 0x00c9c9c9, 0x000b0b0b, 0x00aeaeae, 0x006a6a6a, 
+    0x00d5d5d5, 0x00181818, 0x005d5d5d, 0x00828282, 0x00464646, 0x00dfdfdf, 
+    0x00d6d6d6, 0x00272727, 0x008a8a8a, 0x00323232, 0x004b4b4b, 0x00424242, 
+    0x00dbdbdb, 0x001c1c1c, 0x009e9e9e, 0x009c9c9c, 0x003a3a3a, 0x00cacaca, 
+    0x00252525, 0x007b7b7b, 0x000d0d0d, 0x00717171, 0x005f5f5f, 0x001f1f1f, 
+    0x00f8f8f8, 0x00d7d7d7, 0x003e3e3e, 0x009d9d9d, 0x007c7c7c, 0x00606060, 
+    0x00b9b9b9, 0x00bebebe, 0x00bcbcbc, 0x008b8b8b, 0x00161616, 0x00343434, 
+    0x004d4d4d, 0x00c3c3c3, 0x00727272, 0x00959595, 0x00ababab, 0x008e8e8e, 
+    0x00bababa, 0x007a7a7a, 0x00b3b3b3, 0x00020202, 0x00b4b4b4, 0x00adadad, 
+    0x00a2a2a2, 0x00acacac, 0x00d8d8d8, 0x009a9a9a, 0x00171717, 0x001a1a1a, 
+    0x00353535, 0x00cccccc, 0x00f7f7f7, 0x00999999, 0x00616161, 0x005a5a5a, 
+    0x00e8e8e8, 0x00242424, 0x00565656, 0x00404040, 0x00e1e1e1, 0x00636363, 
+    0x00090909, 0x00333333, 0x00bfbfbf, 0x00989898, 0x00979797, 0x00858585, 
+    0x00686868, 0x00fcfcfc, 0x00ececec, 0x000a0a0a, 0x00dadada, 0x006f6f6f, 
+    0x00535353, 0x00626262, 0x00a3a3a3, 0x002e2e2e, 0x00080808, 0x00afafaf, 
+    0x00282828, 0x00b0b0b0, 0x00747474, 0x00c2c2c2, 0x00bdbdbd, 0x00363636, 
+    0x00222222, 0x00383838, 0x00646464, 0x001e1e1e, 0x00393939, 0x002c2c2c, 
+    0x00a6a6a6, 0x00303030, 0x00e5e5e5, 0x00444444, 0x00fdfdfd, 0x00888888, 
+    0x009f9f9f, 0x00656565, 0x00878787, 0x006b6b6b, 0x00f4f4f4, 0x00232323, 
+    0x00484848, 0x00101010, 0x00d1d1d1, 0x00515151, 0x00c0c0c0, 0x00f9f9f9, 
+    0x00d2d2d2, 0x00a0a0a0, 0x00555555, 0x00a1a1a1, 0x00414141, 0x00fafafa, 
+    0x00434343, 0x00131313, 0x00c4c4c4, 0x002f2f2f, 0x00a8a8a8, 0x00b6b6b6, 
+    0x003c3c3c, 0x002b2b2b, 0x00c1c1c1, 0x00ffffff, 0x00c8c8c8, 0x00a5a5a5, 
+    0x00202020, 0x00898989, 0x00000000, 0x00909090, 0x00474747, 0x00efefef, 
+    0x00eaeaea, 0x00b7b7b7, 0x00151515, 0x00060606, 0x00cdcdcd, 0x00b5b5b5, 
+    0x00121212, 0x007e7e7e, 0x00bbbbbb, 0x00292929, 0x000f0f0f, 0x00b8b8b8, 
+    0x00070707, 0x00040404, 0x009b9b9b, 0x00949494, 0x00212121, 0x00666666, 
+    0x00e6e6e6, 0x00cecece, 0x00ededed, 0x00e7e7e7, 0x003b3b3b, 0x00fefefe, 
+    0x007f7f7f, 0x00c5c5c5, 0x00a4a4a4, 0x00373737, 0x00b1b1b1, 0x004c4c4c, 
+    0x00919191, 0x006e6e6e, 0x008d8d8d, 0x00767676, 0x00030303, 0x002d2d2d, 
+    0x00dedede, 0x00969696, 0x00262626, 0x007d7d7d, 0x00c6c6c6, 0x005c5c5c, 
+    0x00d3d3d3, 0x00f2f2f2, 0x004f4f4f, 0x00191919, 0x003f3f3f, 0x00dcdcdc, 
+    0x00797979, 0x001d1d1d, 0x00525252, 0x00ebebeb, 0x00f3f3f3, 0x006d6d6d, 
+    0x005e5e5e, 0x00fbfbfb, 0x00696969, 0x00b2b2b2, 0x00f0f0f0, 0x00313131, 
+    0x000c0c0c, 0x00d4d4d4, 0x00cfcfcf, 0x008c8c8c, 0x00e2e2e2, 0x00757575, 
+    0x00a9a9a9, 0x004a4a4a, 0x00575757, 0x00848484, 0x00111111, 0x00454545, 
+    0x001b1b1b, 0x00f5f5f5, 0x00e4e4e4, 0x000e0e0e, 0x00737373, 0x00aaaaaa, 
+    0x00f1f1f1, 0x00dddddd, 0x00595959, 0x00141414, 0x006c6c6c, 0x00929292, 
+    0x00545454, 0x00d0d0d0, 0x00787878, 0x00707070, 0x00e3e3e3, 0x00494949, 
+    0x00808080, 0x00505050, 0x00a7a7a7, 0x00f6f6f6, 0x00777777, 0x00939393, 
+    0x00868686, 0x00838383, 0x002a2a2a, 0x00c7c7c7, 0x005b5b5b, 0x00e9e9e9, 
+    0x00eeeeee, 0x008f8f8f, 0x00010101, 0x003d3d3d },
+{   0x38003838, 0x41004141, 0x16001616, 0x76007676, 0xd900d9d9, 0x93009393, 
+    0x60006060, 0xf200f2f2, 0x72007272, 0xc200c2c2, 0xab00abab, 0x9a009a9a, 
+    0x75007575, 0x06000606, 0x57005757, 0xa000a0a0, 0x91009191, 0xf700f7f7, 
+    0xb500b5b5, 0xc900c9c9, 0xa200a2a2, 0x8c008c8c, 0xd200d2d2, 0x90009090, 
+    0xf600f6f6, 0x07000707, 0xa700a7a7, 0x27002727, 0x8e008e8e, 0xb200b2b2, 
+    0x49004949, 0xde00dede, 0x43004343, 0x5c005c5c, 0xd700d7d7, 0xc700c7c7, 
+    0x3e003e3e, 0xf500f5f5, 0x8f008f8f, 0x67006767, 0x1f001f1f, 0x18001818, 
+    0x6e006e6e, 0xaf00afaf, 0x2f002f2f, 0xe200e2e2, 0x85008585, 0x0d000d0d, 
+    0x53005353, 0xf000f0f0, 0x9c009c9c, 0x65006565, 0xea00eaea, 0xa300a3a3, 
+    0xae00aeae, 0x9e009e9e, 0xec00ecec, 0x80008080, 0x2d002d2d, 0x6b006b6b, 
+    0xa800a8a8, 0x2b002b2b, 0x36003636, 0xa600a6a6, 0xc500c5c5, 0x86008686, 
+    0x4d004d4d, 0x33003333, 0xfd00fdfd, 0x66006666, 0x58005858, 0x96009696, 
+    0x3a003a3a, 0x09000909, 0x95009595, 0x10001010, 0x78007878, 0xd800d8d8, 
+    0x42004242, 0xcc00cccc, 0xef00efef, 0x26002626, 0xe500e5e5, 0x61006161, 
+    0x1a001a1a, 0x3f003f3f, 0x3b003b3b, 0x82008282, 0xb600b6b6, 0xdb00dbdb, 
+    0xd400d4d4, 0x98009898, 0xe800e8e8, 0x8b008b8b, 0x02000202, 0xeb00ebeb, 
+    0x0a000a0a, 0x2c002c2c, 0x1d001d1d, 0xb000b0b0, 0x6f006f6f, 0x8d008d8d, 
+    0x88008888, 0x0e000e0e, 0x19001919, 0x87008787, 0x4e004e4e, 0x0b000b0b, 
+    0xa900a9a9, 0x0c000c0c, 0x79007979, 0x11001111, 0x7f007f7f, 0x22002222, 
+    0xe700e7e7, 0x59005959, 0xe100e1e1, 0xda00dada, 0x3d003d3d, 0xc800c8c8, 
+    0x12001212, 0x04000404, 0x74007474, 0x54005454, 0x30003030, 0x7e007e7e, 
+    0xb400b4b4, 0x28002828, 0x55005555, 0x68006868, 0x50005050, 0xbe00bebe, 
+    0xd000d0d0, 0xc400c4c4, 0x31003131, 0xcb00cbcb, 0x2a002a2a, 0xad00adad, 
+    0x0f000f0f, 0xca00caca, 0x70007070, 0xff00ffff, 0x32003232, 0x69006969, 
+    0x08000808, 0x62006262, 0x00000000, 0x24002424, 0xd100d1d1, 0xfb00fbfb, 
+    0xba00baba, 0xed00eded, 0x45004545, 0x81008181, 0x73007373, 0x6d006d6d, 
+    0x84008484, 0x9f009f9f, 0xee00eeee, 0x4a004a4a, 0xc300c3c3, 0x2e002e2e, 
+    0xc100c1c1, 0x01000101, 0xe600e6e6, 0x25002525, 0x48004848, 0x99009999, 
+    0xb900b9b9, 0xb300b3b3, 0x7b007b7b, 0xf900f9f9, 0xce00cece, 0xbf00bfbf, 
+    0xdf00dfdf, 0x71007171, 0x29002929, 0xcd00cdcd, 0x6c006c6c, 0x13001313, 
+    0x64006464, 0x9b009b9b, 0x63006363, 0x9d009d9d, 0xc000c0c0, 0x4b004b4b, 
+    0xb700b7b7, 0xa500a5a5, 0x89008989, 0x5f005f5f, 0xb100b1b1, 0x17001717, 
+    0xf400f4f4, 0xbc00bcbc, 0xd300d3d3, 0x46004646, 0xcf00cfcf, 0x37003737, 
+    0x5e005e5e, 0x47004747, 0x94009494, 0xfa00fafa, 0xfc00fcfc, 0x5b005b5b, 
+    0x97009797, 0xfe00fefe, 0x5a005a5a, 0xac00acac, 0x3c003c3c, 0x4c004c4c, 
+    0x03000303, 0x35003535, 0xf300f3f3, 0x23002323, 0xb800b8b8, 0x5d005d5d, 
+    0x6a006a6a, 0x92009292, 0xd500d5d5, 0x21002121, 0x44004444, 0x51005151, 
+    0xc600c6c6, 0x7d007d7d, 0x39003939, 0x83008383, 0xdc00dcdc, 0xaa00aaaa, 
+    0x7c007c7c, 0x77007777, 0x56005656, 0x05000505, 0x1b001b1b, 0xa400a4a4, 
+    0x15001515, 0x34003434, 0x1e001e1e, 0x1c001c1c, 0xf800f8f8, 0x52005252, 
+    0x20002020, 0x14001414, 0xe900e9e9, 0xbd00bdbd, 0xdd00dddd, 0xe400e4e4, 
+    0xa100a1a1, 0xe000e0e0, 0x8a008a8a, 0xf100f1f1, 0xd600d6d6, 0x7a007a7a, 
+    0xbb00bbbb, 0xe300e3e3, 0x40004040, 0x4f004f4f }
+};
+
+/* Key generation constants */
+static const u32 SIGMA[] = {
+    0xa09e667f, 0x3bcc908b, 0xb67ae858, 0x4caa73b2, 0xc6ef372f, 0xe94f82be,
+    0x54ff53a5, 0xf1d36f1c, 0x10e527fa, 0xde682d1d, 0xb05688c2, 0xb3e6c1fd
+};
+
+/* The phi algorithm given in C.2.7 of the Camellia spec document. */
 /*
- *  macros
+ * This version does not attempt to minimize amount of temporary
+ * variables, but instead explicitly exposes algorithm's parallelism.
+ * It is therefore most appropriate for platforms with not less than
+ * ~16 registers. For platforms with less registers [well, x86 to be
+ * specific] assembler version should be/is provided anyway...
  */
-
-/* e is pointer of subkey */
-#define CamelliaSubkeyL(INDEX) (subkey[(INDEX)*2])
-#define CamelliaSubkeyR(INDEX) (subkey[(INDEX)*2 + 1])
-
-/* rotation right shift 1byte */
-#define CAMELLIA_RR8(x) (((x) >> 8) + ((x) << 24))
-/* rotation left shift 1bit */
-#define CAMELLIA_RL1(x) (((x) << 1) + ((x) >> 31))
-/* rotation left shift 1byte */
-#define CAMELLIA_RL8(x) (((x) << 8) + ((x) >> 24))
-
-#define CAMELLIA_ROLDQ(ll, lr, rl, rr, w0, w1, bits)	\
-do							\
-	{						\
-	w0 = ll;					\
-	ll = (ll << bits) + (lr >> (32 - bits));	\
-	lr = (lr << bits) + (rl >> (32 - bits));	\
-	rl = (rl << bits) + (rr >> (32 - bits));	\
-	rr = (rr << bits) + (w0 >> (32 - bits));	\
-	} while(0)
-
-#define CAMELLIA_ROLDQo32(ll, lr, rl, rr, w0, w1, bits)	\
-do							\
-	{						\
-	w0 = ll;					\
-	w1 = lr;					\
-	ll = (lr << (bits - 32)) + (rl >> (64 - bits));	\
-	lr = (rl << (bits - 32)) + (rr >> (64 - bits));	\
-	rl = (rr << (bits - 32)) + (w0 >> (64 - bits));	\
-	rr = (w0 << (bits - 32)) + (w1 >> (64 - bits));	\
-	} while(0)
-
-#define CAMELLIA_SP1110(INDEX) (camellia_sp1110[(INDEX)])
-#define CAMELLIA_SP0222(INDEX) (camellia_sp0222[(INDEX)])
-#define CAMELLIA_SP3033(INDEX) (camellia_sp3033[(INDEX)])
-#define CAMELLIA_SP4404(INDEX) (camellia_sp4404[(INDEX)])
-
-#define CAMELLIA_F(xl, xr, kl, kr, yl, yr, il, ir, t0, t1)		\
-do									\
-	{								\
-	il = xl ^ kl;							\
-	ir = xr ^ kr;							\
-	t0 = il >> 16;							\
-	t1 = ir >> 16;							\
-	yl = CAMELLIA_SP1110(ir & 0xff)					\
-		^ CAMELLIA_SP0222((t1 >> 8) & 0xff)			\
-		^ CAMELLIA_SP3033(t1 & 0xff)				\
-		^ CAMELLIA_SP4404((ir >> 8) & 0xff);			\
-	yr = CAMELLIA_SP1110((t0 >> 8) & 0xff)				\
-		^ CAMELLIA_SP0222(t0 & 0xff)				\
-		^ CAMELLIA_SP3033((il >> 8) & 0xff)			\
-		^ CAMELLIA_SP4404(il & 0xff);				\
-	yl ^= yr;							\
-	yr = CAMELLIA_RR8(yr);						\
-	yr ^= yl;							\
-	} while(0)
-
+#define Camellia_Feistel(_s0,_s1,_s2,_s3,_key) do {\
+	register u32 _t0,_t1,_t2,_t3;\
+\
+	_t0  = _s0 ^ (_key)[0];\
+	_t3  = SBOX4_4404[_t0&0xff];\
+	_t1  = _s1 ^ (_key)[1];\
+	_t3 ^= SBOX3_3033[(_t0 >> 8)&0xff];\
+	_t2  = SBOX1_1110[_t1&0xff];\
+	_t3 ^= SBOX2_0222[(_t0 >> 16)&0xff];\
+	_t2 ^= SBOX4_4404[(_t1 >> 8)&0xff];\
+	_t3 ^= SBOX1_1110[(_t0 >> 24)];\
+	_t2 ^= _t3;\
+	_t3  = RightRotate(_t3,8);\
+	_t2 ^= SBOX3_3033[(_t1 >> 16)&0xff];\
+	_s3 ^= _t3;\
+	_t2 ^= SBOX2_0222[(_t1 >> 24)];\
+	_s2 ^= _t2; \
+	_s3 ^= _t2;\
+} while(0)
 
 /*
- * for speed up
- *
+ * Note that n has to be less than 32. Rotations for larger amount
+ * of bits are achieved by "rotating" order of s-elements and
+ * adjusting n accordingly, e.g. RotLeft128(s1,s2,s3,s0,n-32).
  */
-#define CAMELLIA_FLS(ll, lr, rl, rr, kll, klr, krl, krr, t0, t1, t2, t3) \
-do									\
-	{								\
-	t0 = kll;							\
-	t0 &= ll;							\
-	lr ^= CAMELLIA_RL1(t0);						\
-	t1 = klr;							\
-	t1 |= lr;							\
-	ll ^= t1;							\
-									\
-	t2 = krr;							\
-	t2 |= rr;							\
-	rl ^= t2;							\
-	t3 = krl;							\
-	t3 &= rl;							\
-	rr ^= CAMELLIA_RL1(t3);						\
-	} while(0)
-
-#define CAMELLIA_ROUNDSM(xl, xr, kl, kr, yl, yr, il, ir, t0, t1)	\
-do									\
-	{								\
-	il = xl;							\
-	ir = xr;							\
-	t0 = il >> 16;							\
-	t1 = ir >> 16;							\
-	ir = CAMELLIA_SP1110(ir & 0xff)					\
-		^ CAMELLIA_SP0222((t1 >> 8) & 0xff)			\
-		^ CAMELLIA_SP3033(t1 & 0xff)				\
-		^ CAMELLIA_SP4404((ir >> 8) & 0xff);			\
-	il = CAMELLIA_SP1110((t0 >> 8) & 0xff)				\
-		^ CAMELLIA_SP0222(t0 & 0xff)				\
-		^ CAMELLIA_SP3033((il >> 8) & 0xff)			\
-		^ CAMELLIA_SP4404(il & 0xff);				\
-	il ^= kl;							\
-	ir ^= kr;							\
-	ir ^= il;							\
-	il = CAMELLIA_RR8(il);						\
-	il ^= ir;							\
-	yl ^= ir;							\
-	yr ^= il;							\
-	} while(0)
-
-static const u32 camellia_sp1110[256] =
-	{
-	0x70707000,0x82828200,0x2c2c2c00,0xececec00,
-	0xb3b3b300,0x27272700,0xc0c0c000,0xe5e5e500,
-	0xe4e4e400,0x85858500,0x57575700,0x35353500,
-	0xeaeaea00,0x0c0c0c00,0xaeaeae00,0x41414100,
-	0x23232300,0xefefef00,0x6b6b6b00,0x93939300,
-	0x45454500,0x19191900,0xa5a5a500,0x21212100,
-	0xededed00,0x0e0e0e00,0x4f4f4f00,0x4e4e4e00,
-	0x1d1d1d00,0x65656500,0x92929200,0xbdbdbd00,
-	0x86868600,0xb8b8b800,0xafafaf00,0x8f8f8f00,
-	0x7c7c7c00,0xebebeb00,0x1f1f1f00,0xcecece00,
-	0x3e3e3e00,0x30303000,0xdcdcdc00,0x5f5f5f00,
-	0x5e5e5e00,0xc5c5c500,0x0b0b0b00,0x1a1a1a00,
-	0xa6a6a600,0xe1e1e100,0x39393900,0xcacaca00,
-	0xd5d5d500,0x47474700,0x5d5d5d00,0x3d3d3d00,
-	0xd9d9d900,0x01010100,0x5a5a5a00,0xd6d6d600,
-	0x51515100,0x56565600,0x6c6c6c00,0x4d4d4d00,
-	0x8b8b8b00,0x0d0d0d00,0x9a9a9a00,0x66666600,
-	0xfbfbfb00,0xcccccc00,0xb0b0b000,0x2d2d2d00,
-	0x74747400,0x12121200,0x2b2b2b00,0x20202000,
-	0xf0f0f000,0xb1b1b100,0x84848400,0x99999900,
-	0xdfdfdf00,0x4c4c4c00,0xcbcbcb00,0xc2c2c200,
-	0x34343400,0x7e7e7e00,0x76767600,0x05050500,
-	0x6d6d6d00,0xb7b7b700,0xa9a9a900,0x31313100,
-	0xd1d1d100,0x17171700,0x04040400,0xd7d7d700,
-	0x14141400,0x58585800,0x3a3a3a00,0x61616100,
-	0xdedede00,0x1b1b1b00,0x11111100,0x1c1c1c00,
-	0x32323200,0x0f0f0f00,0x9c9c9c00,0x16161600,
-	0x53535300,0x18181800,0xf2f2f200,0x22222200,
-	0xfefefe00,0x44444400,0xcfcfcf00,0xb2b2b200,
-	0xc3c3c300,0xb5b5b500,0x7a7a7a00,0x91919100,
-	0x24242400,0x08080800,0xe8e8e800,0xa8a8a800,
-	0x60606000,0xfcfcfc00,0x69696900,0x50505000,
-	0xaaaaaa00,0xd0d0d000,0xa0a0a000,0x7d7d7d00,
-	0xa1a1a100,0x89898900,0x62626200,0x97979700,
-	0x54545400,0x5b5b5b00,0x1e1e1e00,0x95959500,
-	0xe0e0e000,0xffffff00,0x64646400,0xd2d2d200,
-	0x10101000,0xc4c4c400,0x00000000,0x48484800,
-	0xa3a3a300,0xf7f7f700,0x75757500,0xdbdbdb00,
-	0x8a8a8a00,0x03030300,0xe6e6e600,0xdadada00,
-	0x09090900,0x3f3f3f00,0xdddddd00,0x94949400,
-	0x87878700,0x5c5c5c00,0x83838300,0x02020200,
-	0xcdcdcd00,0x4a4a4a00,0x90909000,0x33333300,
-	0x73737300,0x67676700,0xf6f6f600,0xf3f3f300,
-	0x9d9d9d00,0x7f7f7f00,0xbfbfbf00,0xe2e2e200,
-	0x52525200,0x9b9b9b00,0xd8d8d800,0x26262600,
-	0xc8c8c800,0x37373700,0xc6c6c600,0x3b3b3b00,
-	0x81818100,0x96969600,0x6f6f6f00,0x4b4b4b00,
-	0x13131300,0xbebebe00,0x63636300,0x2e2e2e00,
-	0xe9e9e900,0x79797900,0xa7a7a700,0x8c8c8c00,
-	0x9f9f9f00,0x6e6e6e00,0xbcbcbc00,0x8e8e8e00,
-	0x29292900,0xf5f5f500,0xf9f9f900,0xb6b6b600,
-	0x2f2f2f00,0xfdfdfd00,0xb4b4b400,0x59595900,
-	0x78787800,0x98989800,0x06060600,0x6a6a6a00,
-	0xe7e7e700,0x46464600,0x71717100,0xbababa00,
-	0xd4d4d400,0x25252500,0xababab00,0x42424200,
-	0x88888800,0xa2a2a200,0x8d8d8d00,0xfafafa00,
-	0x72727200,0x07070700,0xb9b9b900,0x55555500,
-	0xf8f8f800,0xeeeeee00,0xacacac00,0x0a0a0a00,
-	0x36363600,0x49494900,0x2a2a2a00,0x68686800,
-	0x3c3c3c00,0x38383800,0xf1f1f100,0xa4a4a400,
-	0x40404000,0x28282800,0xd3d3d300,0x7b7b7b00,
-	0xbbbbbb00,0xc9c9c900,0x43434300,0xc1c1c100,
-	0x15151500,0xe3e3e300,0xadadad00,0xf4f4f400,
-	0x77777700,0xc7c7c700,0x80808000,0x9e9e9e00,
-	};
-
-static const u32 camellia_sp0222[256] =
+#define RotLeft128(_s0,_s1,_s2,_s3,_n) do {\
+	u32 _t0=_s0>>(32-_n);\
+	_s0 = (_s0<<_n) | (_s1>>(32-_n));\
+	_s1 = (_s1<<_n) | (_s2>>(32-_n));\
+	_s2 = (_s2<<_n) | (_s3>>(32-_n));\
+	_s3 = (_s3<<_n) | _t0;\
+} while (0)
+
+int Camellia_Ekeygen(int keyBitLength, const u8 *rawKey, KEY_TABLE_TYPE k)
 	{
-	0x00e0e0e0,0x00050505,0x00585858,0x00d9d9d9,
-	0x00676767,0x004e4e4e,0x00818181,0x00cbcbcb,
-	0x00c9c9c9,0x000b0b0b,0x00aeaeae,0x006a6a6a,
-	0x00d5d5d5,0x00181818,0x005d5d5d,0x00828282,
-	0x00464646,0x00dfdfdf,0x00d6d6d6,0x00272727,
-	0x008a8a8a,0x00323232,0x004b4b4b,0x00424242,
-	0x00dbdbdb,0x001c1c1c,0x009e9e9e,0x009c9c9c,
-	0x003a3a3a,0x00cacaca,0x00252525,0x007b7b7b,
-	0x000d0d0d,0x00717171,0x005f5f5f,0x001f1f1f,
-	0x00f8f8f8,0x00d7d7d7,0x003e3e3e,0x009d9d9d,
-	0x007c7c7c,0x00606060,0x00b9b9b9,0x00bebebe,
-	0x00bcbcbc,0x008b8b8b,0x00161616,0x00343434,
-	0x004d4d4d,0x00c3c3c3,0x00727272,0x00959595,
-	0x00ababab,0x008e8e8e,0x00bababa,0x007a7a7a,
-	0x00b3b3b3,0x00020202,0x00b4b4b4,0x00adadad,
-	0x00a2a2a2,0x00acacac,0x00d8d8d8,0x009a9a9a,
-	0x00171717,0x001a1a1a,0x00353535,0x00cccccc,
-	0x00f7f7f7,0x00999999,0x00616161,0x005a5a5a,
-	0x00e8e8e8,0x00242424,0x00565656,0x00404040,
-	0x00e1e1e1,0x00636363,0x00090909,0x00333333,
-	0x00bfbfbf,0x00989898,0x00979797,0x00858585,
-	0x00686868,0x00fcfcfc,0x00ececec,0x000a0a0a,
-	0x00dadada,0x006f6f6f,0x00535353,0x00626262,
-	0x00a3a3a3,0x002e2e2e,0x00080808,0x00afafaf,
-	0x00282828,0x00b0b0b0,0x00747474,0x00c2c2c2,
-	0x00bdbdbd,0x00363636,0x00222222,0x00383838,
-	0x00646464,0x001e1e1e,0x00393939,0x002c2c2c,
-	0x00a6a6a6,0x00303030,0x00e5e5e5,0x00444444,
-	0x00fdfdfd,0x00888888,0x009f9f9f,0x00656565,
-	0x00878787,0x006b6b6b,0x00f4f4f4,0x00232323,
-	0x00484848,0x00101010,0x00d1d1d1,0x00515151,
-	0x00c0c0c0,0x00f9f9f9,0x00d2d2d2,0x00a0a0a0,
-	0x00555555,0x00a1a1a1,0x00414141,0x00fafafa,
-	0x00434343,0x00131313,0x00c4c4c4,0x002f2f2f,
-	0x00a8a8a8,0x00b6b6b6,0x003c3c3c,0x002b2b2b,
-	0x00c1c1c1,0x00ffffff,0x00c8c8c8,0x00a5a5a5,
-	0x00202020,0x00898989,0x00000000,0x00909090,
-	0x00474747,0x00efefef,0x00eaeaea,0x00b7b7b7,
-	0x00151515,0x00060606,0x00cdcdcd,0x00b5b5b5,
-	0x00121212,0x007e7e7e,0x00bbbbbb,0x00292929,
-	0x000f0f0f,0x00b8b8b8,0x00070707,0x00040404,
-	0x009b9b9b,0x00949494,0x00212121,0x00666666,
-	0x00e6e6e6,0x00cecece,0x00ededed,0x00e7e7e7,
-	0x003b3b3b,0x00fefefe,0x007f7f7f,0x00c5c5c5,
-	0x00a4a4a4,0x00373737,0x00b1b1b1,0x004c4c4c,
-	0x00919191,0x006e6e6e,0x008d8d8d,0x00767676,
-	0x00030303,0x002d2d2d,0x00dedede,0x00969696,
-	0x00262626,0x007d7d7d,0x00c6c6c6,0x005c5c5c,
-	0x00d3d3d3,0x00f2f2f2,0x004f4f4f,0x00191919,
-	0x003f3f3f,0x00dcdcdc,0x00797979,0x001d1d1d,
-	0x00525252,0x00ebebeb,0x00f3f3f3,0x006d6d6d,
-	0x005e5e5e,0x00fbfbfb,0x00696969,0x00b2b2b2,
-	0x00f0f0f0,0x00313131,0x000c0c0c,0x00d4d4d4,
-	0x00cfcfcf,0x008c8c8c,0x00e2e2e2,0x00757575,
-	0x00a9a9a9,0x004a4a4a,0x00575757,0x00848484,
-	0x00111111,0x00454545,0x001b1b1b,0x00f5f5f5,
-	0x00e4e4e4,0x000e0e0e,0x00737373,0x00aaaaaa,
-	0x00f1f1f1,0x00dddddd,0x00595959,0x00141414,
-	0x006c6c6c,0x00929292,0x00545454,0x00d0d0d0,
-	0x00787878,0x00707070,0x00e3e3e3,0x00494949,
-	0x00808080,0x00505050,0x00a7a7a7,0x00f6f6f6,
-	0x00777777,0x00939393,0x00868686,0x00838383,
-	0x002a2a2a,0x00c7c7c7,0x005b5b5b,0x00e9e9e9,
-	0x00eeeeee,0x008f8f8f,0x00010101,0x003d3d3d,
-	};
-
-static const u32 camellia_sp3033[256] =
-	{
-	0x38003838,0x41004141,0x16001616,0x76007676,
-	0xd900d9d9,0x93009393,0x60006060,0xf200f2f2,
-	0x72007272,0xc200c2c2,0xab00abab,0x9a009a9a,
-	0x75007575,0x06000606,0x57005757,0xa000a0a0,
-	0x91009191,0xf700f7f7,0xb500b5b5,0xc900c9c9,
-	0xa200a2a2,0x8c008c8c,0xd200d2d2,0x90009090,
-	0xf600f6f6,0x07000707,0xa700a7a7,0x27002727,
-	0x8e008e8e,0xb200b2b2,0x49004949,0xde00dede,
-	0x43004343,0x5c005c5c,0xd700d7d7,0xc700c7c7,
-	0x3e003e3e,0xf500f5f5,0x8f008f8f,0x67006767,
-	0x1f001f1f,0x18001818,0x6e006e6e,0xaf00afaf,
-	0x2f002f2f,0xe200e2e2,0x85008585,0x0d000d0d,
-	0x53005353,0xf000f0f0,0x9c009c9c,0x65006565,
-	0xea00eaea,0xa300a3a3,0xae00aeae,0x9e009e9e,
-	0xec00ecec,0x80008080,0x2d002d2d,0x6b006b6b,
-	0xa800a8a8,0x2b002b2b,0x36003636,0xa600a6a6,
-	0xc500c5c5,0x86008686,0x4d004d4d,0x33003333,
-	0xfd00fdfd,0x66006666,0x58005858,0x96009696,
-	0x3a003a3a,0x09000909,0x95009595,0x10001010,
-	0x78007878,0xd800d8d8,0x42004242,0xcc00cccc,
-	0xef00efef,0x26002626,0xe500e5e5,0x61006161,
-	0x1a001a1a,0x3f003f3f,0x3b003b3b,0x82008282,
-	0xb600b6b6,0xdb00dbdb,0xd400d4d4,0x98009898,
-	0xe800e8e8,0x8b008b8b,0x02000202,0xeb00ebeb,
-	0x0a000a0a,0x2c002c2c,0x1d001d1d,0xb000b0b0,
-	0x6f006f6f,0x8d008d8d,0x88008888,0x0e000e0e,
-	0x19001919,0x87008787,0x4e004e4e,0x0b000b0b,
-	0xa900a9a9,0x0c000c0c,0x79007979,0x11001111,
-	0x7f007f7f,0x22002222,0xe700e7e7,0x59005959,
-	0xe100e1e1,0xda00dada,0x3d003d3d,0xc800c8c8,
-	0x12001212,0x04000404,0x74007474,0x54005454,
-	0x30003030,0x7e007e7e,0xb400b4b4,0x28002828,
-	0x55005555,0x68006868,0x50005050,0xbe00bebe,
-	0xd000d0d0,0xc400c4c4,0x31003131,0xcb00cbcb,
-	0x2a002a2a,0xad00adad,0x0f000f0f,0xca00caca,
-	0x70007070,0xff00ffff,0x32003232,0x69006969,
-	0x08000808,0x62006262,0x00000000,0x24002424,
-	0xd100d1d1,0xfb00fbfb,0xba00baba,0xed00eded,
-	0x45004545,0x81008181,0x73007373,0x6d006d6d,
-	0x84008484,0x9f009f9f,0xee00eeee,0x4a004a4a,
-	0xc300c3c3,0x2e002e2e,0xc100c1c1,0x01000101,
-	0xe600e6e6,0x25002525,0x48004848,0x99009999,
-	0xb900b9b9,0xb300b3b3,0x7b007b7b,0xf900f9f9,
-	0xce00cece,0xbf00bfbf,0xdf00dfdf,0x71007171,
-	0x29002929,0xcd00cdcd,0x6c006c6c,0x13001313,
-	0x64006464,0x9b009b9b,0x63006363,0x9d009d9d,
-	0xc000c0c0,0x4b004b4b,0xb700b7b7,0xa500a5a5,
-	0x89008989,0x5f005f5f,0xb100b1b1,0x17001717,
-	0xf400f4f4,0xbc00bcbc,0xd300d3d3,0x46004646,
-	0xcf00cfcf,0x37003737,0x5e005e5e,0x47004747,
-	0x94009494,0xfa00fafa,0xfc00fcfc,0x5b005b5b,
-	0x97009797,0xfe00fefe,0x5a005a5a,0xac00acac,
-	0x3c003c3c,0x4c004c4c,0x03000303,0x35003535,
-	0xf300f3f3,0x23002323,0xb800b8b8,0x5d005d5d,
-	0x6a006a6a,0x92009292,0xd500d5d5,0x21002121,
-	0x44004444,0x51005151,0xc600c6c6,0x7d007d7d,
-	0x39003939,0x83008383,0xdc00dcdc,0xaa00aaaa,
-	0x7c007c7c,0x77007777,0x56005656,0x05000505,
-	0x1b001b1b,0xa400a4a4,0x15001515,0x34003434,
-	0x1e001e1e,0x1c001c1c,0xf800f8f8,0x52005252,
-	0x20002020,0x14001414,0xe900e9e9,0xbd00bdbd,
-	0xdd00dddd,0xe400e4e4,0xa100a1a1,0xe000e0e0,
-	0x8a008a8a,0xf100f1f1,0xd600d6d6,0x7a007a7a,
-	0xbb00bbbb,0xe300e3e3,0x40004040,0x4f004f4f,
-	};
-
-static const u32 camellia_sp4404[256] =
-	{
-	0x70700070,0x2c2c002c,0xb3b300b3,0xc0c000c0,
-	0xe4e400e4,0x57570057,0xeaea00ea,0xaeae00ae,
-	0x23230023,0x6b6b006b,0x45450045,0xa5a500a5,
-	0xeded00ed,0x4f4f004f,0x1d1d001d,0x92920092,
-	0x86860086,0xafaf00af,0x7c7c007c,0x1f1f001f,
-	0x3e3e003e,0xdcdc00dc,0x5e5e005e,0x0b0b000b,
-	0xa6a600a6,0x39390039,0xd5d500d5,0x5d5d005d,
-	0xd9d900d9,0x5a5a005a,0x51510051,0x6c6c006c,
-	0x8b8b008b,0x9a9a009a,0xfbfb00fb,0xb0b000b0,
-	0x74740074,0x2b2b002b,0xf0f000f0,0x84840084,
-	0xdfdf00df,0xcbcb00cb,0x34340034,0x76760076,
-	0x6d6d006d,0xa9a900a9,0xd1d100d1,0x04040004,
-	0x14140014,0x3a3a003a,0xdede00de,0x11110011,
-	0x32320032,0x9c9c009c,0x53530053,0xf2f200f2,
-	0xfefe00fe,0xcfcf00cf,0xc3c300c3,0x7a7a007a,
-	0x24240024,0xe8e800e8,0x60600060,0x69690069,
-	0xaaaa00aa,0xa0a000a0,0xa1a100a1,0x62620062,
-	0x54540054,0x1e1e001e,0xe0e000e0,0x64640064,
-	0x10100010,0x00000000,0xa3a300a3,0x75750075,
-	0x8a8a008a,0xe6e600e6,0x09090009,0xdddd00dd,
-	0x87870087,0x83830083,0xcdcd00cd,0x90900090,
-	0x73730073,0xf6f600f6,0x9d9d009d,0xbfbf00bf,
-	0x52520052,0xd8d800d8,0xc8c800c8,0xc6c600c6,
-	0x81810081,0x6f6f006f,0x13130013,0x63630063,
-	0xe9e900e9,0xa7a700a7,0x9f9f009f,0xbcbc00bc,
-	0x29290029,0xf9f900f9,0x2f2f002f,0xb4b400b4,
-	0x78780078,0x06060006,0xe7e700e7,0x71710071,
-	0xd4d400d4,0xabab00ab,0x88880088,0x8d8d008d,
-	0x72720072,0xb9b900b9,0xf8f800f8,0xacac00ac,
-	0x36360036,0x2a2a002a,0x3c3c003c,0xf1f100f1,
-	0x40400040,0xd3d300d3,0xbbbb00bb,0x43430043,
-	0x15150015,0xadad00ad,0x77770077,0x80800080,
-	0x82820082,0xecec00ec,0x27270027,0xe5e500e5,
-	0x85850085,0x35350035,0x0c0c000c,0x41410041,
-	0xefef00ef,0x93930093,0x19190019,0x21210021,
-	0x0e0e000e,0x4e4e004e,0x65650065,0xbdbd00bd,
-	0xb8b800b8,0x8f8f008f,0xebeb00eb,0xcece00ce,
-	0x30300030,0x5f5f005f,0xc5c500c5,0x1a1a001a,
-	0xe1e100e1,0xcaca00ca,0x47470047,0x3d3d003d,
-	0x01010001,0xd6d600d6,0x56560056,0x4d4d004d,
-	0x0d0d000d,0x66660066,0xcccc00cc,0x2d2d002d,
-	0x12120012,0x20200020,0xb1b100b1,0x99990099,
-	0x4c4c004c,0xc2c200c2,0x7e7e007e,0x05050005,
-	0xb7b700b7,0x31310031,0x17170017,0xd7d700d7,
-	0x58580058,0x61610061,0x1b1b001b,0x1c1c001c,
-	0x0f0f000f,0x16160016,0x18180018,0x22220022,
-	0x44440044,0xb2b200b2,0xb5b500b5,0x91910091,
-	0x08080008,0xa8a800a8,0xfcfc00fc,0x50500050,
-	0xd0d000d0,0x7d7d007d,0x89890089,0x97970097,
-	0x5b5b005b,0x95950095,0xffff00ff,0xd2d200d2,
-	0xc4c400c4,0x48480048,0xf7f700f7,0xdbdb00db,
-	0x03030003,0xdada00da,0x3f3f003f,0x94940094,
-	0x5c5c005c,0x02020002,0x4a4a004a,0x33330033,
-	0x67670067,0xf3f300f3,0x7f7f007f,0xe2e200e2,
-	0x9b9b009b,0x26260026,0x37370037,0x3b3b003b,
-	0x96960096,0x4b4b004b,0xbebe00be,0x2e2e002e,
-	0x79790079,0x8c8c008c,0x6e6e006e,0x8e8e008e,
-	0xf5f500f5,0xb6b600b6,0xfdfd00fd,0x59590059,
-	0x98980098,0x6a6a006a,0x46460046,0xbaba00ba,
-	0x25250025,0x42420042,0xa2a200a2,0xfafa00fa,
-	0x07070007,0x55550055,0xeeee00ee,0x0a0a000a,
-	0x49490049,0x68680068,0x38380038,0xa4a400a4,
-	0x28280028,0x7b7b007b,0xc9c900c9,0xc1c100c1,
-	0xe3e300e3,0xf4f400f4,0xc7c700c7,0x9e9e009e,
-	};
-
-/**
- * Stuff related to the Camellia key schedule
- */
-#define subl(x) subL[(x)]
-#define subr(x) subR[(x)]
-
-void camellia_setup128(const u8 *key, u32 *subkey)
-	{
-	u32 kll, klr, krl, krr;
-	u32 il, ir, t0, t1, w0, w1;
-	u32 kw4l, kw4r, dw, tl, tr;
-	u32 subL[26];
-	u32 subR[26];
-
-	/**
-	 *  k == kll || klr || krl || krr (|| is concatination)
-	 */
-	kll = GETU32(key     );
-	klr = GETU32(key +  4);
-	krl = GETU32(key +  8);
-	krr = GETU32(key + 12);
-	/**
-	 * generate KL dependent subkeys
-	 */
-	/* kw1 */
-	subl(0) = kll; subr(0) = klr;
-	/* kw2 */
-	subl(1) = krl; subr(1) = krr;
-	/* rotation left shift 15bit */
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 15);
-	/* k3 */
-	subl(4) = kll; subr(4) = klr;
-	/* k4 */
-	subl(5) = krl; subr(5) = krr;
-	/* rotation left shift 15+30bit */
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 30);
-	/* k7 */
-	subl(10) = kll; subr(10) = klr;
-	/* k8 */
-	subl(11) = krl; subr(11) = krr;
-	/* rotation left shift 15+30+15bit */
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 15);
-	/* k10 */
-	subl(13) = krl; subr(13) = krr;
-	/* rotation left shift 15+30+15+17 bit */
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 17);
-	/* kl3 */
-	subl(16) = kll; subr(16) = klr;
-	/* kl4 */
-	subl(17) = krl; subr(17) = krr;
-	/* rotation left shift 15+30+15+17+17 bit */
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 17);
-	/* k13 */
-	subl(18) = kll; subr(18) = klr;
-	/* k14 */
-	subl(19) = krl; subr(19) = krr;
-	/* rotation left shift 15+30+15+17+17+17 bit */
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 17);
-	/* k17 */
-	subl(22) = kll; subr(22) = klr;
-	/* k18 */
-	subl(23) = krl; subr(23) = krr;
-
-	/* generate KA */
-	kll = subl(0); klr = subr(0);
-	krl = subl(1); krr = subr(1);
-	CAMELLIA_F(kll, klr,
-		CAMELLIA_SIGMA1L, CAMELLIA_SIGMA1R,
-		w0, w1, il, ir, t0, t1);
-	krl ^= w0; krr ^= w1;
-	CAMELLIA_F(krl, krr,
-		CAMELLIA_SIGMA2L, CAMELLIA_SIGMA2R,
-		kll, klr, il, ir, t0, t1);
-	/* current status == (kll, klr, w0, w1) */
-	CAMELLIA_F(kll, klr,
-		CAMELLIA_SIGMA3L, CAMELLIA_SIGMA3R,
-		krl, krr, il, ir, t0, t1);
-	krl ^= w0; krr ^= w1;
-	CAMELLIA_F(krl, krr,
-		CAMELLIA_SIGMA4L, CAMELLIA_SIGMA4R,
-		w0, w1, il, ir, t0, t1);
-	kll ^= w0; klr ^= w1;
-
-	/* generate KA dependent subkeys */
-	/* k1, k2 */
-	subl(2) = kll; subr(2) = klr;
-	subl(3) = krl; subr(3) = krr;
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 15);
-	/* k5,k6 */
-	subl(6) = kll; subr(6) = klr;
-	subl(7) = krl; subr(7) = krr;
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 15);
-	/* kl1, kl2 */
-	subl(8) = kll; subr(8) = klr;
-	subl(9) = krl; subr(9) = krr;
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 15);
-	/* k9 */
-	subl(12) = kll; subr(12) = klr;
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 15);
-	/* k11, k12 */
-	subl(14) = kll; subr(14) = klr;
-	subl(15) = krl; subr(15) = krr;
-	CAMELLIA_ROLDQo32(kll, klr, krl, krr, w0, w1, 34);
-	/* k15, k16 */
-	subl(20) = kll; subr(20) = klr;
-	subl(21) = krl; subr(21) = krr;
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 17);
-	/* kw3, kw4 */
-	subl(24) = kll; subr(24) = klr;
-	subl(25) = krl; subr(25) = krr;
-
-
-	/* absorb kw2 to other subkeys */
-/* round 2 */
-	subl(3) ^= subl(1); subr(3) ^= subr(1);
-/* round 4 */
-	subl(5) ^= subl(1); subr(5) ^= subr(1);
-/* round 6 */
-	subl(7) ^= subl(1); subr(7) ^= subr(1);
-	subl(1) ^= subr(1) & ~subr(9);
-	dw = subl(1) & subl(9),
-		subr(1) ^= CAMELLIA_RL1(dw); /* modified for FLinv(kl2) */
-/* round 8 */
-	subl(11) ^= subl(1); subr(11) ^= subr(1);
-/* round 10 */
-	subl(13) ^= subl(1); subr(13) ^= subr(1);
-/* round 12 */
-	subl(15) ^= subl(1); subr(15) ^= subr(1);
-	subl(1) ^= subr(1) & ~subr(17);
-	dw = subl(1) & subl(17),
-		subr(1) ^= CAMELLIA_RL1(dw); /* modified for FLinv(kl4) */
-/* round 14 */
-	subl(19) ^= subl(1); subr(19) ^= subr(1);
-/* round 16 */
-	subl(21) ^= subl(1); subr(21) ^= subr(1);
-/* round 18 */
-	subl(23) ^= subl(1); subr(23) ^= subr(1);
-/* kw3 */
-	subl(24) ^= subl(1); subr(24) ^= subr(1);
-
-	/* absorb kw4 to other subkeys */
-	kw4l = subl(25); kw4r = subr(25);
-/* round 17 */
-	subl(22) ^= kw4l; subr(22) ^= kw4r;
-/* round 15 */
-	subl(20) ^= kw4l; subr(20) ^= kw4r;
-/* round 13 */
-	subl(18) ^= kw4l; subr(18) ^= kw4r;
-	kw4l ^= kw4r & ~subr(16);
-	dw = kw4l & subl(16),
-		kw4r ^= CAMELLIA_RL1(dw); /* modified for FL(kl3) */
-/* round 11 */
-	subl(14) ^= kw4l; subr(14) ^= kw4r;
-/* round 9 */
-	subl(12) ^= kw4l; subr(12) ^= kw4r;
-/* round 7 */
-	subl(10) ^= kw4l; subr(10) ^= kw4r;
-	kw4l ^= kw4r & ~subr(8);
-	dw = kw4l & subl(8),
-		kw4r ^= CAMELLIA_RL1(dw); /* modified for FL(kl1) */
-/* round 5 */
-	subl(6) ^= kw4l; subr(6) ^= kw4r;
-/* round 3 */
-	subl(4) ^= kw4l; subr(4) ^= kw4r;
-/* round 1 */
-	subl(2) ^= kw4l; subr(2) ^= kw4r;
-/* kw1 */
-	subl(0) ^= kw4l; subr(0) ^= kw4r;
-
-
-	/* key XOR is end of F-function */
-	CamelliaSubkeyL(0) = subl(0) ^ subl(2);/* kw1 */
-	CamelliaSubkeyR(0) = subr(0) ^ subr(2);
-	CamelliaSubkeyL(2) = subl(3);       /* round 1 */
-	CamelliaSubkeyR(2) = subr(3);
-	CamelliaSubkeyL(3) = subl(2) ^ subl(4); /* round 2 */
-	CamelliaSubkeyR(3) = subr(2) ^ subr(4);
-	CamelliaSubkeyL(4) = subl(3) ^ subl(5); /* round 3 */
-	CamelliaSubkeyR(4) = subr(3) ^ subr(5);
-	CamelliaSubkeyL(5) = subl(4) ^ subl(6); /* round 4 */
-	CamelliaSubkeyR(5) = subr(4) ^ subr(6);
-	CamelliaSubkeyL(6) = subl(5) ^ subl(7); /* round 5 */
-	CamelliaSubkeyR(6) = subr(5) ^ subr(7);
-	tl = subl(10) ^ (subr(10) & ~subr(8));
-	dw = tl & subl(8),  /* FL(kl1) */
-		tr = subr(10) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(7) = subl(6) ^ tl; /* round 6 */
-	CamelliaSubkeyR(7) = subr(6) ^ tr;
-	CamelliaSubkeyL(8) = subl(8);       /* FL(kl1) */
-	CamelliaSubkeyR(8) = subr(8);
-	CamelliaSubkeyL(9) = subl(9);       /* FLinv(kl2) */
-	CamelliaSubkeyR(9) = subr(9);
-	tl = subl(7) ^ (subr(7) & ~subr(9));
-	dw = tl & subl(9),  /* FLinv(kl2) */
-		tr = subr(7) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(10) = tl ^ subl(11); /* round 7 */
-	CamelliaSubkeyR(10) = tr ^ subr(11);
-	CamelliaSubkeyL(11) = subl(10) ^ subl(12); /* round 8 */
-	CamelliaSubkeyR(11) = subr(10) ^ subr(12);
-	CamelliaSubkeyL(12) = subl(11) ^ subl(13); /* round 9 */
-	CamelliaSubkeyR(12) = subr(11) ^ subr(13);
-	CamelliaSubkeyL(13) = subl(12) ^ subl(14); /* round 10 */
-	CamelliaSubkeyR(13) = subr(12) ^ subr(14);
-	CamelliaSubkeyL(14) = subl(13) ^ subl(15); /* round 11 */
-	CamelliaSubkeyR(14) = subr(13) ^ subr(15);
-	tl = subl(18) ^ (subr(18) & ~subr(16));
-	dw = tl & subl(16), /* FL(kl3) */
-		tr = subr(18) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(15) = subl(14) ^ tl; /* round 12 */
-	CamelliaSubkeyR(15) = subr(14) ^ tr;
-	CamelliaSubkeyL(16) = subl(16);     /* FL(kl3) */
-	CamelliaSubkeyR(16) = subr(16);
-	CamelliaSubkeyL(17) = subl(17);     /* FLinv(kl4) */
-	CamelliaSubkeyR(17) = subr(17);
-	tl = subl(15) ^ (subr(15) & ~subr(17));
-	dw = tl & subl(17), /* FLinv(kl4) */
-		tr = subr(15) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(18) = tl ^ subl(19); /* round 13 */
-	CamelliaSubkeyR(18) = tr ^ subr(19);
-	CamelliaSubkeyL(19) = subl(18) ^ subl(20); /* round 14 */
-	CamelliaSubkeyR(19) = subr(18) ^ subr(20);
-	CamelliaSubkeyL(20) = subl(19) ^ subl(21); /* round 15 */
-	CamelliaSubkeyR(20) = subr(19) ^ subr(21);
-	CamelliaSubkeyL(21) = subl(20) ^ subl(22); /* round 16 */
-	CamelliaSubkeyR(21) = subr(20) ^ subr(22);
-	CamelliaSubkeyL(22) = subl(21) ^ subl(23); /* round 17 */
-	CamelliaSubkeyR(22) = subr(21) ^ subr(23);
-	CamelliaSubkeyL(23) = subl(22);     /* round 18 */
-	CamelliaSubkeyR(23) = subr(22);
-	CamelliaSubkeyL(24) = subl(24) ^ subl(23); /* kw3 */
-	CamelliaSubkeyR(24) = subr(24) ^ subr(23);
-
-	/* apply the inverse of the last half of P-function */
-	dw = CamelliaSubkeyL(2) ^ CamelliaSubkeyR(2),
-		dw = CAMELLIA_RL8(dw);/* round 1 */
-	CamelliaSubkeyR(2) = CamelliaSubkeyL(2) ^ dw,
-		CamelliaSubkeyL(2) = dw;
-	dw = CamelliaSubkeyL(3) ^ CamelliaSubkeyR(3),
-		dw = CAMELLIA_RL8(dw);/* round 2 */
-	CamelliaSubkeyR(3) = CamelliaSubkeyL(3) ^ dw,
-		CamelliaSubkeyL(3) = dw;
-	dw = CamelliaSubkeyL(4) ^ CamelliaSubkeyR(4),
-		dw = CAMELLIA_RL8(dw);/* round 3 */
-	CamelliaSubkeyR(4) = CamelliaSubkeyL(4) ^ dw,
-		CamelliaSubkeyL(4) = dw;
-	dw = CamelliaSubkeyL(5) ^ CamelliaSubkeyR(5),
-		dw = CAMELLIA_RL8(dw);/* round 4 */
-	CamelliaSubkeyR(5) = CamelliaSubkeyL(5) ^ dw,
-		CamelliaSubkeyL(5) = dw;
-	dw = CamelliaSubkeyL(6) ^ CamelliaSubkeyR(6),
-		dw = CAMELLIA_RL8(dw);/* round 5 */
-	CamelliaSubkeyR(6) = CamelliaSubkeyL(6) ^ dw,
-		CamelliaSubkeyL(6) = dw;
-	dw = CamelliaSubkeyL(7) ^ CamelliaSubkeyR(7),
-		dw = CAMELLIA_RL8(dw);/* round 6 */
-	CamelliaSubkeyR(7) = CamelliaSubkeyL(7) ^ dw,
-		CamelliaSubkeyL(7) = dw;
-	dw = CamelliaSubkeyL(10) ^ CamelliaSubkeyR(10),
-		dw = CAMELLIA_RL8(dw);/* round 7 */
-	CamelliaSubkeyR(10) = CamelliaSubkeyL(10) ^ dw,
-		CamelliaSubkeyL(10) = dw;
-	dw = CamelliaSubkeyL(11) ^ CamelliaSubkeyR(11),
-		dw = CAMELLIA_RL8(dw);/* round 8 */
-	CamelliaSubkeyR(11) = CamelliaSubkeyL(11) ^ dw,
-		CamelliaSubkeyL(11) = dw;
-	dw = CamelliaSubkeyL(12) ^ CamelliaSubkeyR(12),
-		dw = CAMELLIA_RL8(dw);/* round 9 */
-	CamelliaSubkeyR(12) = CamelliaSubkeyL(12) ^ dw,
-		CamelliaSubkeyL(12) = dw;
-	dw = CamelliaSubkeyL(13) ^ CamelliaSubkeyR(13),
-		dw = CAMELLIA_RL8(dw);/* round 10 */
-	CamelliaSubkeyR(13) = CamelliaSubkeyL(13) ^ dw,
-		CamelliaSubkeyL(13) = dw;
-	dw = CamelliaSubkeyL(14) ^ CamelliaSubkeyR(14),
-		dw = CAMELLIA_RL8(dw);/* round 11 */
-	CamelliaSubkeyR(14) = CamelliaSubkeyL(14) ^ dw,
-		CamelliaSubkeyL(14) = dw;
-	dw = CamelliaSubkeyL(15) ^ CamelliaSubkeyR(15),
-		dw = CAMELLIA_RL8(dw);/* round 12 */
-	CamelliaSubkeyR(15) = CamelliaSubkeyL(15) ^ dw,
-		CamelliaSubkeyL(15) = dw;
-	dw = CamelliaSubkeyL(18) ^ CamelliaSubkeyR(18),
-		dw = CAMELLIA_RL8(dw);/* round 13 */
-	CamelliaSubkeyR(18) = CamelliaSubkeyL(18) ^ dw,
-		CamelliaSubkeyL(18) = dw;
-	dw = CamelliaSubkeyL(19) ^ CamelliaSubkeyR(19),
-		dw = CAMELLIA_RL8(dw);/* round 14 */
-	CamelliaSubkeyR(19) = CamelliaSubkeyL(19) ^ dw,
-		CamelliaSubkeyL(19) = dw;
-	dw = CamelliaSubkeyL(20) ^ CamelliaSubkeyR(20),
-		dw = CAMELLIA_RL8(dw);/* round 15 */
-	CamelliaSubkeyR(20) = CamelliaSubkeyL(20) ^ dw,
-		CamelliaSubkeyL(20) = dw;
-	dw = CamelliaSubkeyL(21) ^ CamelliaSubkeyR(21),
-		dw = CAMELLIA_RL8(dw);/* round 16 */
-	CamelliaSubkeyR(21) = CamelliaSubkeyL(21) ^ dw,
-		CamelliaSubkeyL(21) = dw;
-	dw = CamelliaSubkeyL(22) ^ CamelliaSubkeyR(22),
-		dw = CAMELLIA_RL8(dw);/* round 17 */
-	CamelliaSubkeyR(22) = CamelliaSubkeyL(22) ^ dw,
-		CamelliaSubkeyL(22) = dw;
-	dw = CamelliaSubkeyL(23) ^ CamelliaSubkeyR(23),
-		dw = CAMELLIA_RL8(dw);/* round 18 */
-	CamelliaSubkeyR(23) = CamelliaSubkeyL(23) ^ dw,
-		CamelliaSubkeyL(23) = dw;
-
-	return;
+	register u32 s0,s1,s2,s3;
+
+	k[0] = s0 = GETU32(rawKey);
+	k[1] = s1 = GETU32(rawKey+4);
+	k[2] = s2 = GETU32(rawKey+8);
+	k[3] = s3 = GETU32(rawKey+12);
+
+	if (keyBitLength != 128)
+		{
+		k[8] = s0 = GETU32(rawKey+16);
+		k[9] = s1 = GETU32(rawKey+20);
+		if (keyBitLength == 192)
+			{
+			k[10] = s2 = ~s0;
+			k[11] = s3 = ~s1;
+			}
+		else
+			{
+			k[10] = s2 = GETU32(rawKey+24);
+			k[11] = s3 = GETU32(rawKey+28);
+			}
+		s0 ^= k[0], s1 ^= k[1], s2 ^= k[2], s3 ^= k[3];
+		}
+
+	/* Use the Feistel routine to scramble the key material */
+	Camellia_Feistel(s0,s1,s2,s3,SIGMA+0);
+	Camellia_Feistel(s2,s3,s0,s1,SIGMA+2);
+
+	s0 ^= k[0], s1 ^= k[1], s2 ^= k[2], s3 ^= k[3];
+	Camellia_Feistel(s0,s1,s2,s3,SIGMA+4);
+	Camellia_Feistel(s2,s3,s0,s1,SIGMA+6);
+
+	/* Fill the keyTable. Requires many block rotations. */
+	if (keyBitLength == 128)
+		{
+		k[ 4] = s0, k[ 5] = s1, k[ 6] = s2, k[ 7] = s3;
+		RotLeft128(s0,s1,s2,s3,15);	/* KA <<< 15 */
+		k[12] = s0, k[13] = s1, k[14] = s2, k[15] = s3;
+		RotLeft128(s0,s1,s2,s3,15);	/* KA <<< 30 */
+		k[16] = s0, k[17] = s1, k[18] = s2, k[19] = s3;
+		RotLeft128(s0,s1,s2,s3,15);	/* KA <<< 45 */
+		k[24] = s0, k[25] = s1;
+		RotLeft128(s0,s1,s2,s3,15);	/* KA <<< 60 */
+		k[28] = s0, k[29] = s1, k[30] = s2, k[31] = s3;
+		RotLeft128(s1,s2,s3,s0,2);	/* KA <<< 94 */
+		k[40] = s1, k[41] = s2, k[42] = s3, k[43] = s0;
+		RotLeft128(s1,s2,s3,s0,17);	/* KA <<<111 */
+		k[48] = s1, k[49] = s2, k[50] = s3, k[51] = s0;
+
+		s0 = k[ 0], s1 = k[ 1], s2 = k[ 2], s3 = k[ 3];
+		RotLeft128(s0,s1,s2,s3,15);	/* KL <<< 15 */
+		k[ 8] = s0, k[ 9] = s1, k[10] = s2, k[11] = s3;
+		RotLeft128(s0,s1,s2,s3,30);	/* KL <<< 45 */
+		k[20] = s0, k[21] = s1, k[22] = s2, k[23] = s3;
+		RotLeft128(s0,s1,s2,s3,15);	/* KL <<< 60 */
+		k[26] = s2, k[27] = s3;
+		RotLeft128(s0,s1,s2,s3,17);	/* KL <<< 77 */
+		k[32] = s0, k[33] = s1, k[34] = s2, k[35] = s3;
+		RotLeft128(s0,s1,s2,s3,17);	/* KL <<< 94 */
+		k[36] = s0, k[37] = s1, k[38] = s2, k[39] = s3;
+		RotLeft128(s0,s1,s2,s3,17);	/* KL <<<111 */
+		k[44] = s0, k[45] = s1, k[46] = s2, k[47] = s3;
+
+		return 3;	/* grand rounds */
+		}
+	else
+		{
+		k[12] = s0, k[13] = s1, k[14] = s2, k[15] = s3;
+		s0 ^= k[8], s1 ^= k[9], s2 ^=k[10], s3 ^=k[11];
+		Camellia_Feistel(s0,s1,s2,s3,(SIGMA+8));
+		Camellia_Feistel(s2,s3,s0,s1,(SIGMA+10));
+
+		k[ 4] = s0, k[ 5] = s1, k[ 6] = s2, k[ 7] = s3;
+		RotLeft128(s0,s1,s2,s3,30);	/* KB <<< 30 */
+		k[20] = s0, k[21] = s1, k[22] = s2, k[23] = s3;
+		RotLeft128(s0,s1,s2,s3,30);	/* KB <<< 60 */
+		k[40] = s0, k[41] = s1, k[42] = s2, k[43] = s3;
+		RotLeft128(s1,s2,s3,s0,19);	/* KB <<<111 */
+		k[64] = s1, k[65] = s2, k[66] = s3, k[67] = s0;
+
+		s0 = k[ 8], s1 = k[ 9], s2 = k[10], s3 = k[11];
+		RotLeft128(s0,s1,s2,s3,15);	/* KR <<< 15 */
+		k[ 8] = s0, k[ 9] = s1, k[10] = s2, k[11] = s3;
+		RotLeft128(s0,s1,s2,s3,15);	/* KR <<< 30 */
+		k[16] = s0, k[17] = s1, k[18] = s2, k[19] = s3;
+		RotLeft128(s0,s1,s2,s3,30);	/* KR <<< 60 */
+		k[36] = s0, k[37] = s1, k[38] = s2, k[39] = s3;
+		RotLeft128(s1,s2,s3,s0,2);	/* KR <<< 94 */
+		k[52] = s1, k[53] = s2, k[54] = s3, k[55] = s0;
+
+		s0 = k[12], s1 = k[13], s2 = k[14], s3 = k[15];
+		RotLeft128(s0,s1,s2,s3,15);	/* KA <<< 15 */
+		k[12] = s0, k[13] = s1, k[14] = s2, k[15] = s3;
+		RotLeft128(s0,s1,s2,s3,30);	/* KA <<< 45 */
+		k[28] = s0, k[29] = s1, k[30] = s2, k[31] = s3;
+						/* KA <<< 77 */
+		k[48] = s1, k[49] = s2, k[50] = s3, k[51] = s0;
+		RotLeft128(s1,s2,s3,s0,17);	/* KA <<< 94 */
+		k[56] = s1, k[57] = s2, k[58] = s3, k[59] = s0;
+
+		s0 = k[ 0], s1 = k[ 1], s2 = k[ 2], s3 = k[ 3];
+		RotLeft128(s1,s2,s3,s0,13);	/* KL <<< 45 */
+		k[24] = s1, k[25] = s2, k[26] = s3, k[27] = s0;
+		RotLeft128(s1,s2,s3,s0,15);	/* KL <<< 60 */
+		k[32] = s1, k[33] = s2, k[34] = s3, k[35] = s0;
+		RotLeft128(s1,s2,s3,s0,17);	/* KL <<< 77 */
+		k[44] = s1, k[45] = s2, k[46] = s3, k[47] = s0;
+		RotLeft128(s2,s3,s0,s1,2);	/* KL <<<111 */
+		k[60] = s2, k[61] = s3, k[62] = s0, k[63] = s1;
+
+		return 4;	/* grand rounds */
+		}
+	/*
+	 * It is possible to perform certain precalculations, which
+	 * would spare few cycles in block procedure. It's not done,
+	 * because it upsets the performance balance between key
+	 * setup and block procedures, negatively affecting overall
+	 * throughput in applications operating on short messages
+	 * and volatile keys.
+	 */ 
 	}
 
-void camellia_setup256(const u8 *key, u32 *subkey)
+void Camellia_EncryptBlock_Rounds(int grandRounds, const u8 plaintext[], 
+		const KEY_TABLE_TYPE keyTable, u8 ciphertext[])
 	{
-	u32 kll,klr,krl,krr;           /* left half of key */
-	u32 krll,krlr,krrl,krrr;       /* right half of key */
-	u32 il, ir, t0, t1, w0, w1;    /* temporary variables */
-	u32 kw4l, kw4r, dw, tl, tr;
-	u32 subL[34];
-	u32 subR[34];
-
-	/**
-	 *  key = (kll || klr || krl || krr || krll || krlr || krrl || krrr)
-	 *  (|| is concatination)
-	 */
-
-	kll  = GETU32(key     );
-	klr  = GETU32(key +  4);
-	krl  = GETU32(key +  8);
-	krr  = GETU32(key + 12);
-	krll = GETU32(key + 16);
-	krlr = GETU32(key + 20);
-	krrl = GETU32(key + 24);
-	krrr = GETU32(key + 28);
-
-	/* generate KL dependent subkeys */
-	/* kw1 */
-	subl(0) = kll; subr(0) = klr;
-	/* kw2 */
-	subl(1) = krl; subr(1) = krr;
-	CAMELLIA_ROLDQo32(kll, klr, krl, krr, w0, w1, 45);
-	/* k9 */
-	subl(12) = kll; subr(12) = klr;
-	/* k10 */
-	subl(13) = krl; subr(13) = krr;
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 15);
-	/* kl3 */
-	subl(16) = kll; subr(16) = klr;
-	/* kl4 */
-	subl(17) = krl; subr(17) = krr;
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 17);
-	/* k17 */
-	subl(22) = kll; subr(22) = klr;
-	/* k18 */
-	subl(23) = krl; subr(23) = krr;
-	CAMELLIA_ROLDQo32(kll, klr, krl, krr, w0, w1, 34);
-	/* k23 */
-	subl(30) = kll; subr(30) = klr;
-	/* k24 */
-	subl(31) = krl; subr(31) = krr;
-
-	/* generate KR dependent subkeys */
-	CAMELLIA_ROLDQ(krll, krlr, krrl, krrr, w0, w1, 15);
-	/* k3 */
-	subl(4) = krll; subr(4) = krlr;
-	/* k4 */
-	subl(5) = krrl; subr(5) = krrr;
-	CAMELLIA_ROLDQ(krll, krlr, krrl, krrr, w0, w1, 15);
-	/* kl1 */
-	subl(8) = krll; subr(8) = krlr;
-	/* kl2 */
-	subl(9) = krrl; subr(9) = krrr;
-	CAMELLIA_ROLDQ(krll, krlr, krrl, krrr, w0, w1, 30);
-	/* k13 */
-	subl(18) = krll; subr(18) = krlr;
-	/* k14 */
-	subl(19) = krrl; subr(19) = krrr;
-	CAMELLIA_ROLDQo32(krll, krlr, krrl, krrr, w0, w1, 34);
-	/* k19 */
-	subl(26) = krll; subr(26) = krlr;
-	/* k20 */
-	subl(27) = krrl; subr(27) = krrr;
-	CAMELLIA_ROLDQo32(krll, krlr, krrl, krrr, w0, w1, 34);
-
-	/* generate KA */
-	kll = subl(0) ^ krll; klr = subr(0) ^ krlr;
-	krl = subl(1) ^ krrl; krr = subr(1) ^ krrr;
-	CAMELLIA_F(kll, klr,
-		CAMELLIA_SIGMA1L, CAMELLIA_SIGMA1R,
-		w0, w1, il, ir, t0, t1);
-	krl ^= w0; krr ^= w1;
-	CAMELLIA_F(krl, krr,
-		CAMELLIA_SIGMA2L, CAMELLIA_SIGMA2R,
-		kll, klr, il, ir, t0, t1);
-	kll ^= krll; klr ^= krlr;
-	CAMELLIA_F(kll, klr,
-		CAMELLIA_SIGMA3L, CAMELLIA_SIGMA3R,
-		krl, krr, il, ir, t0, t1);
-	krl ^= w0 ^ krrl; krr ^= w1 ^ krrr;
-	CAMELLIA_F(krl, krr,
-		CAMELLIA_SIGMA4L, CAMELLIA_SIGMA4R,
-		w0, w1, il, ir, t0, t1);
-	kll ^= w0; klr ^= w1;
-
-	/* generate KB */
-	krll ^= kll; krlr ^= klr;
-	krrl ^= krl; krrr ^= krr;
-	CAMELLIA_F(krll, krlr,
-		CAMELLIA_SIGMA5L, CAMELLIA_SIGMA5R,
-		w0, w1, il, ir, t0, t1);
-	krrl ^= w0; krrr ^= w1;
-	CAMELLIA_F(krrl, krrr,
-		CAMELLIA_SIGMA6L, CAMELLIA_SIGMA6R,
-		w0, w1, il, ir, t0, t1);
-	krll ^= w0; krlr ^= w1;
-
-	/* generate KA dependent subkeys */
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 15);
-	/* k5 */
-	subl(6) = kll; subr(6) = klr;
-	/* k6 */
-	subl(7) = krl; subr(7) = krr;
-	CAMELLIA_ROLDQ(kll, klr, krl, krr, w0, w1, 30);
-	/* k11 */
-	subl(14) = kll; subr(14) = klr;
-	/* k12 */
-	subl(15) = krl; subr(15) = krr;
-	/* rotation left shift 32bit */
-	/* kl5 */
-	subl(24) = klr; subr(24) = krl;
-	/* kl6 */
-	subl(25) = krr; subr(25) = kll;
-	/* rotation left shift 49 from k11,k12 -> k21,k22 */
-	CAMELLIA_ROLDQo32(kll, klr, krl, krr, w0, w1, 49);
-	/* k21 */
-	subl(28) = kll; subr(28) = klr;
-	/* k22 */
-	subl(29) = krl; subr(29) = krr;
-
-	/* generate KB dependent subkeys */
-	/* k1 */
-	subl(2) = krll; subr(2) = krlr;
-	/* k2 */
-	subl(3) = krrl; subr(3) = krrr;
-	CAMELLIA_ROLDQ(krll, krlr, krrl, krrr, w0, w1, 30);
-	/* k7 */
-	subl(10) = krll; subr(10) = krlr;
-	/* k8 */
-	subl(11) = krrl; subr(11) = krrr;
-	CAMELLIA_ROLDQ(krll, krlr, krrl, krrr, w0, w1, 30);
-	/* k15 */
-	subl(20) = krll; subr(20) = krlr;
-	/* k16 */
-	subl(21) = krrl; subr(21) = krrr;
-	CAMELLIA_ROLDQo32(krll, krlr, krrl, krrr, w0, w1, 51);
-	/* kw3 */
-	subl(32) = krll; subr(32) = krlr;
-	/* kw4 */
-	subl(33) = krrl; subr(33) = krrr;
-
-	/* absorb kw2 to other subkeys */
-/* round 2 */
-	subl(3) ^= subl(1); subr(3) ^= subr(1);
-/* round 4 */
-	subl(5) ^= subl(1); subr(5) ^= subr(1);
-/* round 6 */
-	subl(7) ^= subl(1); subr(7) ^= subr(1);
-	subl(1) ^= subr(1) & ~subr(9);
-	dw = subl(1) & subl(9),
-		subr(1) ^= CAMELLIA_RL1(dw); /* modified for FLinv(kl2) */
-/* round 8 */
-	subl(11) ^= subl(1); subr(11) ^= subr(1);
-/* round 10 */
-	subl(13) ^= subl(1); subr(13) ^= subr(1);
-/* round 12 */
-	subl(15) ^= subl(1); subr(15) ^= subr(1);
-	subl(1) ^= subr(1) & ~subr(17);
-	dw = subl(1) & subl(17),
-		subr(1) ^= CAMELLIA_RL1(dw); /* modified for FLinv(kl4) */
-/* round 14 */
-	subl(19) ^= subl(1); subr(19) ^= subr(1);
-/* round 16 */
-	subl(21) ^= subl(1); subr(21) ^= subr(1);
-/* round 18 */
-	subl(23) ^= subl(1); subr(23) ^= subr(1);
-	subl(1) ^= subr(1) & ~subr(25);
-	dw = subl(1) & subl(25),
-		subr(1) ^= CAMELLIA_RL1(dw); /* modified for FLinv(kl6) */
-/* round 20 */
-	subl(27) ^= subl(1); subr(27) ^= subr(1);
-/* round 22 */
-	subl(29) ^= subl(1); subr(29) ^= subr(1);
-/* round 24 */
-	subl(31) ^= subl(1); subr(31) ^= subr(1);
-/* kw3 */
-	subl(32) ^= subl(1); subr(32) ^= subr(1);
-
-
-	/* absorb kw4 to other subkeys */
-	kw4l = subl(33); kw4r = subr(33);
-/* round 23 */
-	subl(30) ^= kw4l; subr(30) ^= kw4r;
-/* round 21 */
-	subl(28) ^= kw4l; subr(28) ^= kw4r;
-/* round 19 */
-	subl(26) ^= kw4l; subr(26) ^= kw4r;
-	kw4l ^= kw4r & ~subr(24);
-	dw = kw4l & subl(24),
-		kw4r ^= CAMELLIA_RL1(dw); /* modified for FL(kl5) */
-/* round 17 */
-	subl(22) ^= kw4l; subr(22) ^= kw4r;
-/* round 15 */
-	subl(20) ^= kw4l; subr(20) ^= kw4r;
-/* round 13 */
-	subl(18) ^= kw4l; subr(18) ^= kw4r;
-	kw4l ^= kw4r & ~subr(16);
-	dw = kw4l & subl(16),
-		kw4r ^= CAMELLIA_RL1(dw); /* modified for FL(kl3) */
-/* round 11 */
-	subl(14) ^= kw4l; subr(14) ^= kw4r;
-/* round 9 */
-	subl(12) ^= kw4l; subr(12) ^= kw4r;
-/* round 7 */
-	subl(10) ^= kw4l; subr(10) ^= kw4r;
-	kw4l ^= kw4r & ~subr(8);
-	dw = kw4l & subl(8),
-		kw4r ^= CAMELLIA_RL1(dw); /* modified for FL(kl1) */
-/* round 5 */
-	subl(6) ^= kw4l; subr(6) ^= kw4r;
-/* round 3 */
-	subl(4) ^= kw4l; subr(4) ^= kw4r;
-/* round 1 */
-	subl(2) ^= kw4l; subr(2) ^= kw4r;
-/* kw1 */
-	subl(0) ^= kw4l; subr(0) ^= kw4r;
-
-	/* key XOR is end of F-function */
-	CamelliaSubkeyL(0) = subl(0) ^ subl(2);/* kw1 */
-	CamelliaSubkeyR(0) = subr(0) ^ subr(2);
-	CamelliaSubkeyL(2) = subl(3);       /* round 1 */
-	CamelliaSubkeyR(2) = subr(3);
-	CamelliaSubkeyL(3) = subl(2) ^ subl(4); /* round 2 */
-	CamelliaSubkeyR(3) = subr(2) ^ subr(4);
-	CamelliaSubkeyL(4) = subl(3) ^ subl(5); /* round 3 */
-	CamelliaSubkeyR(4) = subr(3) ^ subr(5);
-	CamelliaSubkeyL(5) = subl(4) ^ subl(6); /* round 4 */
-	CamelliaSubkeyR(5) = subr(4) ^ subr(6);
-	CamelliaSubkeyL(6) = subl(5) ^ subl(7); /* round 5 */
-	CamelliaSubkeyR(6) = subr(5) ^ subr(7);
-	tl = subl(10) ^ (subr(10) & ~subr(8));
-	dw = tl & subl(8),  /* FL(kl1) */
-		tr = subr(10) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(7) = subl(6) ^ tl; /* round 6 */
-	CamelliaSubkeyR(7) = subr(6) ^ tr;
-	CamelliaSubkeyL(8) = subl(8);       /* FL(kl1) */
-	CamelliaSubkeyR(8) = subr(8);
-	CamelliaSubkeyL(9) = subl(9);       /* FLinv(kl2) */
-	CamelliaSubkeyR(9) = subr(9);
-	tl = subl(7) ^ (subr(7) & ~subr(9));
-	dw = tl & subl(9),  /* FLinv(kl2) */
-		tr = subr(7) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(10) = tl ^ subl(11); /* round 7 */
-	CamelliaSubkeyR(10) = tr ^ subr(11);
-	CamelliaSubkeyL(11) = subl(10) ^ subl(12); /* round 8 */
-	CamelliaSubkeyR(11) = subr(10) ^ subr(12);
-	CamelliaSubkeyL(12) = subl(11) ^ subl(13); /* round 9 */
-	CamelliaSubkeyR(12) = subr(11) ^ subr(13);
-	CamelliaSubkeyL(13) = subl(12) ^ subl(14); /* round 10 */
-	CamelliaSubkeyR(13) = subr(12) ^ subr(14);
-	CamelliaSubkeyL(14) = subl(13) ^ subl(15); /* round 11 */
-	CamelliaSubkeyR(14) = subr(13) ^ subr(15);
-	tl = subl(18) ^ (subr(18) & ~subr(16));
-	dw = tl & subl(16), /* FL(kl3) */
-		tr = subr(18) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(15) = subl(14) ^ tl; /* round 12 */
-	CamelliaSubkeyR(15) = subr(14) ^ tr;
-	CamelliaSubkeyL(16) = subl(16);     /* FL(kl3) */
-	CamelliaSubkeyR(16) = subr(16);
-	CamelliaSubkeyL(17) = subl(17);     /* FLinv(kl4) */
-	CamelliaSubkeyR(17) = subr(17);
-	tl = subl(15) ^ (subr(15) & ~subr(17));
-	dw = tl & subl(17), /* FLinv(kl4) */
-		tr = subr(15) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(18) = tl ^ subl(19); /* round 13 */
-	CamelliaSubkeyR(18) = tr ^ subr(19);
-	CamelliaSubkeyL(19) = subl(18) ^ subl(20); /* round 14 */
-	CamelliaSubkeyR(19) = subr(18) ^ subr(20);
-	CamelliaSubkeyL(20) = subl(19) ^ subl(21); /* round 15 */
-	CamelliaSubkeyR(20) = subr(19) ^ subr(21);
-	CamelliaSubkeyL(21) = subl(20) ^ subl(22); /* round 16 */
-	CamelliaSubkeyR(21) = subr(20) ^ subr(22);
-	CamelliaSubkeyL(22) = subl(21) ^ subl(23); /* round 17 */
-	CamelliaSubkeyR(22) = subr(21) ^ subr(23);
-	tl = subl(26) ^ (subr(26)
-		& ~subr(24));
-	dw = tl & subl(24), /* FL(kl5) */
-		tr = subr(26) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(23) = subl(22) ^ tl; /* round 18 */
-	CamelliaSubkeyR(23) = subr(22) ^ tr;
-	CamelliaSubkeyL(24) = subl(24);     /* FL(kl5) */
-	CamelliaSubkeyR(24) = subr(24);
-	CamelliaSubkeyL(25) = subl(25);     /* FLinv(kl6) */
-	CamelliaSubkeyR(25) = subr(25);
-	tl = subl(23) ^ (subr(23) &
-		~subr(25));
-	dw = tl & subl(25), /* FLinv(kl6) */
-		tr = subr(23) ^ CAMELLIA_RL1(dw);
-	CamelliaSubkeyL(26) = tl ^ subl(27); /* round 19 */
-	CamelliaSubkeyR(26) = tr ^ subr(27);
-	CamelliaSubkeyL(27) = subl(26) ^ subl(28); /* round 20 */
-	CamelliaSubkeyR(27) = subr(26) ^ subr(28);
-	CamelliaSubkeyL(28) = subl(27) ^ subl(29); /* round 21 */
-	CamelliaSubkeyR(28) = subr(27) ^ subr(29);
-	CamelliaSubkeyL(29) = subl(28) ^ subl(30); /* round 22 */
-	CamelliaSubkeyR(29) = subr(28) ^ subr(30);
-	CamelliaSubkeyL(30) = subl(29) ^ subl(31); /* round 23 */
-	CamelliaSubkeyR(30) = subr(29) ^ subr(31);
-	CamelliaSubkeyL(31) = subl(30);     /* round 24 */
-	CamelliaSubkeyR(31) = subr(30);
-	CamelliaSubkeyL(32) = subl(32) ^ subl(31); /* kw3 */
-	CamelliaSubkeyR(32) = subr(32) ^ subr(31);
-
-	/* apply the inverse of the last half of P-function */
-	dw = CamelliaSubkeyL(2) ^ CamelliaSubkeyR(2),
-		dw = CAMELLIA_RL8(dw);/* round 1 */
-	CamelliaSubkeyR(2) = CamelliaSubkeyL(2) ^ dw,
-		CamelliaSubkeyL(2) = dw;
-	dw = CamelliaSubkeyL(3) ^ CamelliaSubkeyR(3),
-		dw = CAMELLIA_RL8(dw);/* round 2 */
-	CamelliaSubkeyR(3) = CamelliaSubkeyL(3) ^ dw,
-		CamelliaSubkeyL(3) = dw;
-	dw = CamelliaSubkeyL(4) ^ CamelliaSubkeyR(4),
-		dw = CAMELLIA_RL8(dw);/* round 3 */
-	CamelliaSubkeyR(4) = CamelliaSubkeyL(4) ^ dw,
-		CamelliaSubkeyL(4) = dw;
-	dw = CamelliaSubkeyL(5) ^ CamelliaSubkeyR(5),
-		dw = CAMELLIA_RL8(dw);/* round 4 */
-	CamelliaSubkeyR(5) = CamelliaSubkeyL(5) ^ dw,
-		CamelliaSubkeyL(5) = dw;
-	dw = CamelliaSubkeyL(6) ^ CamelliaSubkeyR(6),
-		dw = CAMELLIA_RL8(dw);/* round 5 */
-	CamelliaSubkeyR(6) = CamelliaSubkeyL(6) ^ dw,
-		CamelliaSubkeyL(6) = dw;
-	dw = CamelliaSubkeyL(7) ^ CamelliaSubkeyR(7),
-		dw = CAMELLIA_RL8(dw);/* round 6 */
-	CamelliaSubkeyR(7) = CamelliaSubkeyL(7) ^ dw,
-		CamelliaSubkeyL(7) = dw;
-	dw = CamelliaSubkeyL(10) ^ CamelliaSubkeyR(10),
-		dw = CAMELLIA_RL8(dw);/* round 7 */
-	CamelliaSubkeyR(10) = CamelliaSubkeyL(10) ^ dw,
-		CamelliaSubkeyL(10) = dw;
-	dw = CamelliaSubkeyL(11) ^ CamelliaSubkeyR(11),
-		dw = CAMELLIA_RL8(dw);/* round 8 */
-	CamelliaSubkeyR(11) = CamelliaSubkeyL(11) ^ dw,
-		CamelliaSubkeyL(11) = dw;
-	dw = CamelliaSubkeyL(12) ^ CamelliaSubkeyR(12),
-		dw = CAMELLIA_RL8(dw);/* round 9 */
-	CamelliaSubkeyR(12) = CamelliaSubkeyL(12) ^ dw,
-		CamelliaSubkeyL(12) = dw;
-	dw = CamelliaSubkeyL(13) ^ CamelliaSubkeyR(13),
-		dw = CAMELLIA_RL8(dw);/* round 10 */
-	CamelliaSubkeyR(13) = CamelliaSubkeyL(13) ^ dw,
-		CamelliaSubkeyL(13) = dw;
-	dw = CamelliaSubkeyL(14) ^ CamelliaSubkeyR(14),
-		dw = CAMELLIA_RL8(dw);/* round 11 */
-	CamelliaSubkeyR(14) = CamelliaSubkeyL(14) ^ dw,
-		CamelliaSubkeyL(14) = dw;
-	dw = CamelliaSubkeyL(15) ^ CamelliaSubkeyR(15),
-		dw = CAMELLIA_RL8(dw);/* round 12 */
-	CamelliaSubkeyR(15) = CamelliaSubkeyL(15) ^ dw,
-		CamelliaSubkeyL(15) = dw;
-	dw = CamelliaSubkeyL(18) ^ CamelliaSubkeyR(18),
-		dw = CAMELLIA_RL8(dw);/* round 13 */
-	CamelliaSubkeyR(18) = CamelliaSubkeyL(18) ^ dw,
-		CamelliaSubkeyL(18) = dw;
-	dw = CamelliaSubkeyL(19) ^ CamelliaSubkeyR(19),
-		dw = CAMELLIA_RL8(dw);/* round 14 */
-	CamelliaSubkeyR(19) = CamelliaSubkeyL(19) ^ dw,
-		CamelliaSubkeyL(19) = dw;
-	dw = CamelliaSubkeyL(20) ^ CamelliaSubkeyR(20),
-		dw = CAMELLIA_RL8(dw);/* round 15 */
-	CamelliaSubkeyR(20) = CamelliaSubkeyL(20) ^ dw,
-		CamelliaSubkeyL(20) = dw;
-	dw = CamelliaSubkeyL(21) ^ CamelliaSubkeyR(21),
-		dw = CAMELLIA_RL8(dw);/* round 16 */
-	CamelliaSubkeyR(21) = CamelliaSubkeyL(21) ^ dw,
-		CamelliaSubkeyL(21) = dw;
-	dw = CamelliaSubkeyL(22) ^ CamelliaSubkeyR(22),
-		dw = CAMELLIA_RL8(dw);/* round 17 */
-	CamelliaSubkeyR(22) = CamelliaSubkeyL(22) ^ dw,
-		CamelliaSubkeyL(22) = dw;
-	dw = CamelliaSubkeyL(23) ^ CamelliaSubkeyR(23),
-		dw = CAMELLIA_RL8(dw);/* round 18 */
-	CamelliaSubkeyR(23) = CamelliaSubkeyL(23) ^ dw,
-		CamelliaSubkeyL(23) = dw;
-	dw = CamelliaSubkeyL(26) ^ CamelliaSubkeyR(26),
-		dw = CAMELLIA_RL8(dw);/* round 19 */
-	CamelliaSubkeyR(26) = CamelliaSubkeyL(26) ^ dw,
-		CamelliaSubkeyL(26) = dw;
-	dw = CamelliaSubkeyL(27) ^ CamelliaSubkeyR(27),
-		dw = CAMELLIA_RL8(dw);/* round 20 */
-	CamelliaSubkeyR(27) = CamelliaSubkeyL(27) ^ dw,
-		CamelliaSubkeyL(27) = dw;
-	dw = CamelliaSubkeyL(28) ^ CamelliaSubkeyR(28),
-		dw = CAMELLIA_RL8(dw);/* round 21 */
-	CamelliaSubkeyR(28) = CamelliaSubkeyL(28) ^ dw,
-		CamelliaSubkeyL(28) = dw;
-	dw = CamelliaSubkeyL(29) ^ CamelliaSubkeyR(29),
-		dw = CAMELLIA_RL8(dw);/* round 22 */
-	CamelliaSubkeyR(29) = CamelliaSubkeyL(29) ^ dw,
-		CamelliaSubkeyL(29) = dw;
-	dw = CamelliaSubkeyL(30) ^ CamelliaSubkeyR(30),
-		dw = CAMELLIA_RL8(dw);/* round 23 */
-	CamelliaSubkeyR(30) = CamelliaSubkeyL(30) ^ dw,
-		CamelliaSubkeyL(30) = dw;
-	dw = CamelliaSubkeyL(31) ^ CamelliaSubkeyR(31),
-		dw = CAMELLIA_RL8(dw);/* round 24 */
-	CamelliaSubkeyR(31) = CamelliaSubkeyL(31) ^ dw,
-		CamelliaSubkeyL(31) = dw;
-
-    
-	return;
+	register u32 s0,s1,s2,s3; 
+	const u32 *k = keyTable,*kend = keyTable+grandRounds*16; 
+
+	s0 = GETU32(plaintext)    ^ k[0];
+	s1 = GETU32(plaintext+4)  ^ k[1];
+	s2 = GETU32(plaintext+8)  ^ k[2];
+	s3 = GETU32(plaintext+12) ^ k[3];
+	k += 4;
+
+	while (1)
+		{
+		/* Camellia makes 6 Feistel rounds */
+		Camellia_Feistel(s0,s1,s2,s3,k+0);
+		Camellia_Feistel(s2,s3,s0,s1,k+2);
+		Camellia_Feistel(s0,s1,s2,s3,k+4);
+		Camellia_Feistel(s2,s3,s0,s1,k+6);
+		Camellia_Feistel(s0,s1,s2,s3,k+8);
+		Camellia_Feistel(s2,s3,s0,s1,k+10);
+		k += 12;
+
+		if (k == kend) break;
+
+		/* This is the same function as the diffusion function D
+		 * of the accompanying documentation. See section 3.2
+		 * for properties of the FLlayer function. */
+		s1 ^= LeftRotate(s0 & k[0], 1);
+		s2 ^= s3 | k[3];
+		s0 ^= s1 | k[1];
+		s3 ^= LeftRotate(s2 & k[2], 1);
+		k += 4;
+		}
+
+	s2 ^= k[0], s3 ^= k[1], s0 ^= k[2], s1 ^= k[3];
+
+	PUTU32(ciphertext,   s2);
+	PUTU32(ciphertext+4, s3);
+	PUTU32(ciphertext+8, s0);
+	PUTU32(ciphertext+12,s1);
 	}
-
-void camellia_setup192(const u8 *key, u32 *subkey)
-	{
-	u8 kk[32];
-	u32 krll, krlr, krrl,krrr;
-
-	memcpy(kk, key, 24);
-	memcpy((u8 *)&krll, key+16,4);
-	memcpy((u8 *)&krlr, key+20,4);
-	krrl = ~krll;
-	krrr = ~krlr;
-	memcpy(kk+24, (u8 *)&krrl, 4);
-	memcpy(kk+28, (u8 *)&krrr, 4);
-	camellia_setup256(kk, subkey);
-	return;
-	}
-
-
-/**
- * Stuff related to camellia encryption/decryption
- */
-void camellia_encrypt128(const u32 *subkey, u32 *io)
-	{
-	u32 il, ir, t0, t1;
-
-	/* pre whitening but absorb kw2*/
-	io[0] ^= CamelliaSubkeyL(0);
-	io[1] ^= CamelliaSubkeyR(0);
-	/* main iteration */
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(2),CamelliaSubkeyR(2),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(3),CamelliaSubkeyR(3),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(4),CamelliaSubkeyR(4),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(5),CamelliaSubkeyR(5),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(6),CamelliaSubkeyR(6),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(7),CamelliaSubkeyR(7),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(8),CamelliaSubkeyR(8),
-		CamelliaSubkeyL(9),CamelliaSubkeyR(9),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(10),CamelliaSubkeyR(10),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(11),CamelliaSubkeyR(11),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(12),CamelliaSubkeyR(12),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(13),CamelliaSubkeyR(13),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(14),CamelliaSubkeyR(14),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(15),CamelliaSubkeyR(15),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(16),CamelliaSubkeyR(16),
-		CamelliaSubkeyL(17),CamelliaSubkeyR(17),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(18),CamelliaSubkeyR(18),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(19),CamelliaSubkeyR(19),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(20),CamelliaSubkeyR(20),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(21),CamelliaSubkeyR(21),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(22),CamelliaSubkeyR(22),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(23),CamelliaSubkeyR(23),
-		io[0],io[1],il,ir,t0,t1);
-
-	/* post whitening but kw4 */
-	io[2] ^= CamelliaSubkeyL(24);
-	io[3] ^= CamelliaSubkeyR(24);
-
-	t0 = io[0];
-	t1 = io[1];
-	io[0] = io[2];
-	io[1] = io[3];
-	io[2] = t0;
-	io[3] = t1;
-
-	return;
-	}
-
-void camellia_decrypt128(const u32 *subkey, u32 *io)
+void Camellia_EncryptBlock(int keyBitLength, const u8 plaintext[], 
+		const KEY_TABLE_TYPE keyTable, u8 ciphertext[])
 	{
-	u32 il,ir,t0,t1;               /* temporary valiables */
-
-	/* pre whitening but absorb kw2*/
-	io[0] ^= CamelliaSubkeyL(24);
-	io[1] ^= CamelliaSubkeyR(24);
-
-	/* main iteration */
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(23),CamelliaSubkeyR(23),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(22),CamelliaSubkeyR(22),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(21),CamelliaSubkeyR(21),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(20),CamelliaSubkeyR(20),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(19),CamelliaSubkeyR(19),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(18),CamelliaSubkeyR(18),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(17),CamelliaSubkeyR(17),
-		CamelliaSubkeyL(16),CamelliaSubkeyR(16),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(15),CamelliaSubkeyR(15),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(14),CamelliaSubkeyR(14),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(13),CamelliaSubkeyR(13),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(12),CamelliaSubkeyR(12),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(11),CamelliaSubkeyR(11),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(10),CamelliaSubkeyR(10),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(9),CamelliaSubkeyR(9),
-		CamelliaSubkeyL(8),CamelliaSubkeyR(8),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(7),CamelliaSubkeyR(7),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(6),CamelliaSubkeyR(6),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(5),CamelliaSubkeyR(5),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(4),CamelliaSubkeyR(4),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(3),CamelliaSubkeyR(3),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(2),CamelliaSubkeyR(2),
-		io[0],io[1],il,ir,t0,t1);
-
-	/* post whitening but kw4 */
-	io[2] ^= CamelliaSubkeyL(0);
-	io[3] ^= CamelliaSubkeyR(0);
-
-	t0 = io[0];
-	t1 = io[1];
-	io[0] = io[2];
-	io[1] = io[3];
-	io[2] = t0;
-	io[3] = t1;
-
-	return;
+	Camellia_EncryptBlock_Rounds(keyBitLength==128?3:4,
+			plaintext,keyTable,ciphertext);
 	}
 
-/**
- * stuff for 192 and 256bit encryption/decryption
- */
-void camellia_encrypt256(const u32 *subkey, u32 *io)
+void Camellia_DecryptBlock_Rounds(int grandRounds, const u8 ciphertext[], 
+		const KEY_TABLE_TYPE keyTable, u8 plaintext[])
 	{
-	u32 il,ir,t0,t1;           /* temporary valiables */
-
-	/* pre whitening but absorb kw2*/
-	io[0] ^= CamelliaSubkeyL(0);
-	io[1] ^= CamelliaSubkeyR(0);
-
-	/* main iteration */
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(2),CamelliaSubkeyR(2),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(3),CamelliaSubkeyR(3),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(4),CamelliaSubkeyR(4),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(5),CamelliaSubkeyR(5),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(6),CamelliaSubkeyR(6),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(7),CamelliaSubkeyR(7),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(8),CamelliaSubkeyR(8),
-		CamelliaSubkeyL(9),CamelliaSubkeyR(9),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(10),CamelliaSubkeyR(10),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(11),CamelliaSubkeyR(11),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(12),CamelliaSubkeyR(12),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(13),CamelliaSubkeyR(13),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(14),CamelliaSubkeyR(14),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(15),CamelliaSubkeyR(15),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(16),CamelliaSubkeyR(16),
-		CamelliaSubkeyL(17),CamelliaSubkeyR(17),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(18),CamelliaSubkeyR(18),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(19),CamelliaSubkeyR(19),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(20),CamelliaSubkeyR(20),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(21),CamelliaSubkeyR(21),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(22),CamelliaSubkeyR(22),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(23),CamelliaSubkeyR(23),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(24),CamelliaSubkeyR(24),
-		CamelliaSubkeyL(25),CamelliaSubkeyR(25),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(26),CamelliaSubkeyR(26),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(27),CamelliaSubkeyR(27),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(28),CamelliaSubkeyR(28),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(29),CamelliaSubkeyR(29),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(30),CamelliaSubkeyR(30),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(31),CamelliaSubkeyR(31),
-		io[0],io[1],il,ir,t0,t1);
-
-	/* post whitening but kw4 */
-	io[2] ^= CamelliaSubkeyL(32);
-	io[3] ^= CamelliaSubkeyR(32);
-
-	t0 = io[0];
-	t1 = io[1];
-	io[0] = io[2];
-	io[1] = io[3];
-	io[2] = t0;
-	io[3] = t1;
-
-	return;
+	u32 s0,s1,s2,s3; 
+	const u32 *k = keyTable+grandRounds*16,*kend = keyTable+4; 
+
+	s0 = GETU32(ciphertext)    ^ k[0];
+	s1 = GETU32(ciphertext+4)  ^ k[1];
+	s2 = GETU32(ciphertext+8)  ^ k[2];
+	s3 = GETU32(ciphertext+12) ^ k[3];
+
+	while (1)
+		{
+		/* Camellia makes 6 Feistel rounds */
+		k -= 12;
+		Camellia_Feistel(s0,s1,s2,s3,k+10);
+		Camellia_Feistel(s2,s3,s0,s1,k+8);
+		Camellia_Feistel(s0,s1,s2,s3,k+6);
+		Camellia_Feistel(s2,s3,s0,s1,k+4);
+		Camellia_Feistel(s0,s1,s2,s3,k+2);
+		Camellia_Feistel(s2,s3,s0,s1,k+0);
+
+		if (k == kend) break;
+
+		/* This is the same function as the diffusion function D
+		 * of the accompanying documentation. See section 3.2
+		 * for properties of the FLlayer function. */
+		k -= 4;
+		s1 ^= LeftRotate(s0 & k[2], 1);
+		s2 ^= s3 | k[1];
+		s0 ^= s1 | k[3];
+		s3 ^= LeftRotate(s2 & k[0], 1);
+		}
+
+	k -= 4;
+	s2 ^= k[0], s3 ^= k[1], s0 ^= k[2], s1 ^= k[3];
+
+	PUTU32(plaintext,   s2);
+	PUTU32(plaintext+4, s3);
+	PUTU32(plaintext+8, s0);
+	PUTU32(plaintext+12,s1);
 	}
-
-void camellia_decrypt256(const u32 *subkey, u32 *io)
+void Camellia_DecryptBlock(int keyBitLength, const u8 plaintext[], 
+		const KEY_TABLE_TYPE keyTable, u8 ciphertext[])
 	{
-	u32 il,ir,t0,t1;           /* temporary valiables */
-
-	/* pre whitening but absorb kw2*/
-	io[0] ^= CamelliaSubkeyL(32);
-	io[1] ^= CamelliaSubkeyR(32);
-	
-	/* main iteration */
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(31),CamelliaSubkeyR(31),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(30),CamelliaSubkeyR(30),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(29),CamelliaSubkeyR(29),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(28),CamelliaSubkeyR(28),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(27),CamelliaSubkeyR(27),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(26),CamelliaSubkeyR(26),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(25),CamelliaSubkeyR(25),
-		CamelliaSubkeyL(24),CamelliaSubkeyR(24),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(23),CamelliaSubkeyR(23),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(22),CamelliaSubkeyR(22),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(21),CamelliaSubkeyR(21),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(20),CamelliaSubkeyR(20),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(19),CamelliaSubkeyR(19),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(18),CamelliaSubkeyR(18),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(17),CamelliaSubkeyR(17),
-		CamelliaSubkeyL(16),CamelliaSubkeyR(16),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(15),CamelliaSubkeyR(15),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(14),CamelliaSubkeyR(14),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(13),CamelliaSubkeyR(13),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(12),CamelliaSubkeyR(12),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(11),CamelliaSubkeyR(11),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(10),CamelliaSubkeyR(10),
-		io[0],io[1],il,ir,t0,t1);
-
-	CAMELLIA_FLS(io[0],io[1],io[2],io[3],
-		CamelliaSubkeyL(9),CamelliaSubkeyR(9),
-		CamelliaSubkeyL(8),CamelliaSubkeyR(8),
-		t0,t1,il,ir);
-
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(7),CamelliaSubkeyR(7),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(6),CamelliaSubkeyR(6),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(5),CamelliaSubkeyR(5),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(4),CamelliaSubkeyR(4),
-		io[0],io[1],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[0],io[1],
-		CamelliaSubkeyL(3),CamelliaSubkeyR(3),
-		io[2],io[3],il,ir,t0,t1);
-	CAMELLIA_ROUNDSM(io[2],io[3],
-		CamelliaSubkeyL(2),CamelliaSubkeyR(2),
-		io[0],io[1],il,ir,t0,t1);
-
-	/* post whitening but kw4 */
-	io[2] ^= CamelliaSubkeyL(0);
-	io[3] ^= CamelliaSubkeyR(0);
-
-	t0 = io[0];
-	t1 = io[1];
-	io[0] = io[2];
-	io[1] = io[3];
-	io[2] = t0;
-	io[3] = t1;
-
-	return;
+	Camellia_DecryptBlock_Rounds(keyBitLength==128?3:4,
+			plaintext,keyTable,ciphertext);
 	}
-
diff --git a/crypto/camellia/camellia.h b/crypto/camellia/camellia.h
index b8a8b6e10b9e..67911e0adf83 100644
--- a/crypto/camellia/camellia.h
+++ b/crypto/camellia/camellia.h
@@ -58,6 +58,8 @@
 #error CAMELLIA is disabled.
 #endif
 
+#include <stddef.h>
+
 #define CAMELLIA_ENCRYPT	1
 #define CAMELLIA_DECRYPT	0
 
@@ -74,24 +76,22 @@ extern "C" {
 #define CAMELLIA_TABLE_BYTE_LEN 272
 #define CAMELLIA_TABLE_WORD_LEN (CAMELLIA_TABLE_BYTE_LEN / 4)
 
- /* to match with WORD */
-typedef unsigned int KEY_TABLE_TYPE[CAMELLIA_TABLE_WORD_LEN];
+typedef unsigned int KEY_TABLE_TYPE[CAMELLIA_TABLE_WORD_LEN]; /* to match with WORD */
 
 struct camellia_key_st 
 	{
-	KEY_TABLE_TYPE rd_key;
-	int bitLength;
-	void (*enc)(const unsigned int *subkey, unsigned int *io);
-	void (*dec)(const unsigned int *subkey, unsigned int *io);
+	union	{
+		double d;	/* ensures 64-bit align */
+		KEY_TABLE_TYPE rd_key;
+		} u;
+	int grand_rounds;
 	};
-
 typedef struct camellia_key_st CAMELLIA_KEY;
 
 #ifdef OPENSSL_FIPS
 int private_Camellia_set_key(const unsigned char *userKey, const int bits,
 	CAMELLIA_KEY *key);
 #endif
-
 int Camellia_set_key(const unsigned char *userKey, const int bits,
 	CAMELLIA_KEY *key);
 
@@ -103,25 +103,22 @@ void Camellia_decrypt(const unsigned char *in, unsigned char *out,
 void Camellia_ecb_encrypt(const unsigned char *in, unsigned char *out,
 	const CAMELLIA_KEY *key, const int enc);
 void Camellia_cbc_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char *ivec, const int enc);
 void Camellia_cfb128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char *ivec, int *num, const int enc);
 void Camellia_cfb1_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char *ivec, int *num, const int enc);
 void Camellia_cfb8_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char *ivec, int *num, const int enc);
-void Camellia_cfbr_encrypt_block(const unsigned char *in,unsigned char *out,
-	const int nbits,const CAMELLIA_KEY *key,
-	unsigned char *ivec,const int enc);
 void Camellia_ofb128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char *ivec, int *num);
 void Camellia_ctr128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char ivec[CAMELLIA_BLOCK_SIZE],
 	unsigned char ecount_buf[CAMELLIA_BLOCK_SIZE],
 	unsigned int *num);
@@ -131,4 +128,3 @@ void Camellia_ctr128_encrypt(const unsigned char *in, unsigned char *out,
 #endif
 
 #endif /* !HEADER_Camellia_H */
-
diff --git a/crypto/camellia/cmll_cbc.c b/crypto/camellia/cmll_cbc.c
index 4141a7b59bb7..4c8d455adef3 100644
--- a/crypto/camellia/cmll_cbc.c
+++ b/crypto/camellia/cmll_cbc.c
@@ -49,225 +49,16 @@
  *
  */
 
-#ifndef CAMELLIA_DEBUG
-# ifndef NDEBUG
-#  define NDEBUG
-# endif
-#endif
-#include <assert.h>
-#include <stdio.h>
-#include <string.h>
-
 #include <openssl/camellia.h>
-#include "cmll_locl.h"
+#include <openssl/modes.h>
 
 void Camellia_cbc_encrypt(const unsigned char *in, unsigned char *out,
-		     const unsigned long length, const CAMELLIA_KEY *key,
-		     unsigned char *ivec, const int enc) {
-
-	unsigned long n;
-	unsigned long len = length;
-	const unsigned char *iv = ivec;
-	union {	u32 t32[CAMELLIA_BLOCK_SIZE/sizeof(u32)];
-		u8  t8 [CAMELLIA_BLOCK_SIZE]; } tmp;
-	const union { long one; char little; } camellia_endian = {1};
-
-
-	assert(in && out && key && ivec);
-	assert((CAMELLIA_ENCRYPT == enc)||(CAMELLIA_DECRYPT == enc));
+	size_t len, const CAMELLIA_KEY *key,
+	unsigned char *ivec, const int enc) 
+	{
 
-	if(((size_t)in|(size_t)out|(size_t)ivec) % sizeof(u32) == 0)
-		{
-		if (CAMELLIA_ENCRYPT == enc)
-			{
-			while (len >= CAMELLIA_BLOCK_SIZE)
-				{
-				XOR4WORD2((u32 *)out,
-					(u32 *)in, (u32 *)iv);
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				key->enc(key->rd_key, (u32 *)out);
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				iv = out;
-				len -= CAMELLIA_BLOCK_SIZE;
-				in += CAMELLIA_BLOCK_SIZE;
-				out += CAMELLIA_BLOCK_SIZE;
-				}
-			if (len)
-				{
-				for(n=0; n < len; ++n)
-					out[n] = in[n] ^ iv[n];
-				for(n=len; n < CAMELLIA_BLOCK_SIZE; ++n)
-					out[n] = iv[n];
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				key->enc(key->rd_key, (u32 *)out);
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				iv = out;
-				}
-			memcpy(ivec,iv,CAMELLIA_BLOCK_SIZE);
-			}
-		else if (in != out)
-			{
-			while (len >= CAMELLIA_BLOCK_SIZE)
-				{
-				memcpy(out,in,CAMELLIA_BLOCK_SIZE);
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				key->dec(key->rd_key,(u32 *)out);
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				XOR4WORD((u32 *)out, (u32 *)iv);
-				iv = in;
-				len -= CAMELLIA_BLOCK_SIZE;
-				in  += CAMELLIA_BLOCK_SIZE;
-				out += CAMELLIA_BLOCK_SIZE;
-				}
-			if (len)
-				{
-				memcpy(tmp.t8, in, CAMELLIA_BLOCK_SIZE);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				key->dec(key->rd_key, tmp.t32);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				for(n=0; n < len; ++n)
-					out[n] = tmp.t8[n] ^ iv[n];
-				iv = in;
-				}
-			memcpy(ivec,iv,CAMELLIA_BLOCK_SIZE);
-			}
-		else /* in == out */
-			{
-			while (len >= CAMELLIA_BLOCK_SIZE)
-				{
-				memcpy(tmp.t8, in, CAMELLIA_BLOCK_SIZE);
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				key->dec(key->rd_key, (u32 *)out);
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				XOR4WORD((u32 *)out, (u32 *)ivec);
-				memcpy(ivec, tmp.t8, CAMELLIA_BLOCK_SIZE);
-				len -= CAMELLIA_BLOCK_SIZE;
-				in += CAMELLIA_BLOCK_SIZE;
-				out += CAMELLIA_BLOCK_SIZE;
-				}
-			if (len)
-				{
-				memcpy(tmp.t8, in, CAMELLIA_BLOCK_SIZE);
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				key->dec(key->rd_key,(u32 *)out);
-				if (camellia_endian.little)
-					SWAP4WORD((u32 *)out);
-				for(n=0; n < len; ++n)
-					out[n] ^= ivec[n];
-				for(n=len; n < CAMELLIA_BLOCK_SIZE; ++n)
-					out[n] = tmp.t8[n];
-				memcpy(ivec, tmp.t8, CAMELLIA_BLOCK_SIZE);
-				}
-			}
-		}
-	else /* no aligned */
-		{
-		if (CAMELLIA_ENCRYPT == enc)
-			{
-			while (len >= CAMELLIA_BLOCK_SIZE)
-				{
-				for(n=0; n < CAMELLIA_BLOCK_SIZE; ++n)
-					tmp.t8[n] = in[n] ^ iv[n];
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				key->enc(key->rd_key, tmp.t32);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				memcpy(out, tmp.t8, CAMELLIA_BLOCK_SIZE);
-				iv = out;
-				len -= CAMELLIA_BLOCK_SIZE;
-				in += CAMELLIA_BLOCK_SIZE;
-				out += CAMELLIA_BLOCK_SIZE;
-				}
-			if (len)
-				{
-				for(n=0; n < len; ++n)
-					tmp.t8[n] = in[n] ^ iv[n];
-				for(n=len; n < CAMELLIA_BLOCK_SIZE; ++n)
-					tmp.t8[n] = iv[n];
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				key->enc(key->rd_key, tmp.t32);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				memcpy(out, tmp.t8, CAMELLIA_BLOCK_SIZE);
-				iv = out;
-				}
-			memcpy(ivec,iv,CAMELLIA_BLOCK_SIZE);
-			}
-		else if (in != out)
-			{
-			while (len >= CAMELLIA_BLOCK_SIZE)
-				{
-				memcpy(tmp.t8,in,CAMELLIA_BLOCK_SIZE);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				key->dec(key->rd_key,tmp.t32);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				for(n=0; n < CAMELLIA_BLOCK_SIZE; ++n)
-					out[n] = tmp.t8[n] ^ iv[n];
-				iv = in;
-				len -= CAMELLIA_BLOCK_SIZE;
-				in  += CAMELLIA_BLOCK_SIZE;
-				out += CAMELLIA_BLOCK_SIZE;
-				}
-			if (len)
-				{
-				memcpy(tmp.t8, in, CAMELLIA_BLOCK_SIZE);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				key->dec(key->rd_key, tmp.t32);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				for(n=0; n < len; ++n)
-					out[n] = tmp.t8[n] ^ iv[n];
-				iv = in;
-				}
-			memcpy(ivec,iv,CAMELLIA_BLOCK_SIZE);
-			}
-		else
-			{
-			while (len >= CAMELLIA_BLOCK_SIZE)
-				{
-				memcpy(tmp.t8, in, CAMELLIA_BLOCK_SIZE);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				key->dec(key->rd_key, tmp.t32);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				for(n=0; n < CAMELLIA_BLOCK_SIZE; ++n)
-					tmp.t8[n] ^= ivec[n];
-				memcpy(ivec, in, CAMELLIA_BLOCK_SIZE);
-				memcpy(out, tmp.t8, CAMELLIA_BLOCK_SIZE);
-				len -= CAMELLIA_BLOCK_SIZE;
-				in += CAMELLIA_BLOCK_SIZE;
-				out += CAMELLIA_BLOCK_SIZE;
-				}
-			if (len)
-				{
-				memcpy(tmp.t8, in, CAMELLIA_BLOCK_SIZE);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				key->dec(key->rd_key,tmp.t32);
-				if (camellia_endian.little)
-					SWAP4WORD(tmp.t32);
-				for(n=0; n < len; ++n)
-					tmp.t8[n] ^= ivec[n];
-				memcpy(ivec, in, CAMELLIA_BLOCK_SIZE);
-				memcpy(out,tmp.t8,len);
-				}
-			}
-		}
-}
+	if (enc)
+		CRYPTO_cbc128_encrypt(in,out,len,key,ivec,(block128_f)Camellia_encrypt);
+	else
+		CRYPTO_cbc128_decrypt(in,out,len,key,ivec,(block128_f)Camellia_decrypt);
+	}
diff --git a/crypto/camellia/cmll_cfb.c b/crypto/camellia/cmll_cfb.c
index af0f9f49ad90..3d81b51d3f42 100644
--- a/crypto/camellia/cmll_cfb.c
+++ b/crypto/camellia/cmll_cfb.c
@@ -105,17 +105,8 @@
  * [including the GNU Public Licence.]
  */
 
-#ifndef CAMELLIA_DEBUG
-# ifndef NDEBUG
-#  define NDEBUG
-# endif
-#endif
-#include <assert.h>
-#include <string.h>
-
 #include <openssl/camellia.h>
-#include "cmll_locl.h"
-#include "e_os.h"
+#include <openssl/modes.h>
 
 
 /* The input and output encrypted as though 128bit cfb mode is being
@@ -124,112 +115,25 @@
  */
 
 void Camellia_cfb128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char *ivec, int *num, const int enc)
 	{
 
-	unsigned int n;
-	unsigned long l = length;
-	unsigned char c;
-
-	assert(in && out && key && ivec && num);
-
-	n = *num;
-
-	if (enc) 
-		{
-		while (l--) 
-			{
-			if (n == 0) 
-				{
-				Camellia_encrypt(ivec, ivec, key);
-				}
-			ivec[n] = *(out++) = *(in++) ^ ivec[n];
-			n = (n+1) % CAMELLIA_BLOCK_SIZE;
-			}
-		} 
-	else 
-		{
-		while (l--) 
-			{
-			if (n == 0) 
-				{
-				Camellia_encrypt(ivec, ivec, key);
-				}
-			c = *(in);
-			*(out++) = *(in++) ^ ivec[n];
-			ivec[n] = c;
-			n = (n+1) % CAMELLIA_BLOCK_SIZE;
-			}
-		}
-
-	*num=n;
-	}
-
-/* This expects a single block of size nbits for both in and out. Note that
-   it corrupts any extra bits in the last byte of out */
-void Camellia_cfbr_encrypt_block(const unsigned char *in,unsigned char *out,
-	const int nbits,const CAMELLIA_KEY *key,
-	unsigned char *ivec,const int enc)
-	{
-	int n,rem,num;
-	unsigned char ovec[CAMELLIA_BLOCK_SIZE*2];
-
-	if (nbits<=0 || nbits>128) return;
-
-	/* fill in the first half of the new IV with the current IV */
-	memcpy(ovec,ivec,CAMELLIA_BLOCK_SIZE);
-	/* construct the new IV */
-	Camellia_encrypt(ivec,ivec,key);
-	num = (nbits+7)/8;
-	if (enc)	/* encrypt the input */
-		for(n=0 ; n < num ; ++n)
-			out[n] = (ovec[CAMELLIA_BLOCK_SIZE+n] = in[n] ^ ivec[n]);
-	else		/* decrypt the input */
-		for(n=0 ; n < num ; ++n)
-			out[n] = (ovec[CAMELLIA_BLOCK_SIZE+n] = in[n]) ^ ivec[n];
-	/* shift ovec left... */
-	rem = nbits%8;
-	num = nbits/8;
-	if(rem==0)
-		memcpy(ivec,ovec+num,CAMELLIA_BLOCK_SIZE);
-	else
-		for(n=0 ; n < CAMELLIA_BLOCK_SIZE ; ++n)
-			ivec[n] = ovec[n+num]<<rem | ovec[n+num+1]>>(8-rem);
-
-	/* it is not necessary to cleanse ovec, since the IV is not secret */
+	CRYPTO_cfb128_encrypt(in,out,length,key,ivec,num,enc,(block128_f)Camellia_encrypt);
 	}
 
 /* N.B. This expects the input to be packed, MS bit first */
 void Camellia_cfb1_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char *ivec, int *num, const int enc)
 	{
-	unsigned int n;
-	unsigned char c[1],d[1];
-
-	assert(in && out && key && ivec && num);
-	assert(*num == 0);
-
-	memset(out,0,(length+7)/8);
-	for(n=0 ; n < length ; ++n)
-		{
-		c[0]=(in[n/8]&(1 << (7-n%8))) ? 0x80 : 0;
-		Camellia_cfbr_encrypt_block(c,d,1,key,ivec,enc);
-		out[n/8]=(out[n/8]&~(1 << (7-n%8)))|((d[0]&0x80) >> (n%8));
-		}
+	CRYPTO_cfb128_1_encrypt(in,out,length,key,ivec,num,enc,(block128_f)Camellia_encrypt);
 	}
 
 void Camellia_cfb8_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char *ivec, int *num, const int enc)
 	{
-	unsigned int n;
-
-	assert(in && out && key && ivec && num);
-	assert(*num == 0);
-
-	for(n=0 ; n < length ; ++n)
-		Camellia_cfbr_encrypt_block(&in[n],&out[n],8,key,ivec,enc);
+	CRYPTO_cfb128_8_encrypt(in,out,length,key,ivec,num,enc,(block128_f)Camellia_encrypt);
 	}
 
diff --git a/crypto/camellia/cmll_ctr.c b/crypto/camellia/cmll_ctr.c
index cc21b70890db..014e621a34bc 100644
--- a/crypto/camellia/cmll_ctr.c
+++ b/crypto/camellia/cmll_ctr.c
@@ -49,95 +49,16 @@
  *
  */
 
-#ifndef CAMELLIA_DEBUG
-# ifndef NDEBUG
-#  define NDEBUG
-# endif
-#endif
-#include <assert.h>
-
 #include <openssl/camellia.h>
-#include "cmll_locl.h"
-
-/* NOTE: the IV/counter CTR mode is big-endian.  The rest of the Camellia code
- * is endian-neutral. */
-/* increment counter (128-bit int) by 1 */
-static void Camellia_ctr128_inc(unsigned char *counter) 
-	{
-	unsigned long c;
-
-	/* Grab bottom dword of counter and increment */
-	c = GETU32(counter + 12);
-	c++;	c &= 0xFFFFFFFF;
-	PUTU32(counter + 12, c);
-
-	/* if no overflow, we're done */
-	if (c)
-		return;
-
-	/* Grab 1st dword of counter and increment */
-	c = GETU32(counter +  8);
-	c++;	c &= 0xFFFFFFFF;
-	PUTU32(counter +  8, c);
-
-	/* if no overflow, we're done */
-	if (c)
-		return;
-
-	/* Grab 2nd dword of counter and increment */
-	c = GETU32(counter +  4);
-	c++;	c &= 0xFFFFFFFF;
-	PUTU32(counter +  4, c);
-
-	/* if no overflow, we're done */
-	if (c)
-		return;
+#include <openssl/modes.h>
 
-	/* Grab top dword of counter and increment */
-	c = GETU32(counter +  0);
-	c++;	c &= 0xFFFFFFFF;
-	PUTU32(counter +  0, c);
-	}
-
-/* The input encrypted as though 128bit counter mode is being
- * used.  The extra state information to record how much of the
- * 128bit block we have used is contained in *num, and the
- * encrypted counter is kept in ecount_buf.  Both *num and
- * ecount_buf must be initialised with zeros before the first
- * call to Camellia_ctr128_encrypt().
- *
- * This algorithm assumes that the counter is in the x lower bits
- * of the IV (ivec), and that the application has full control over
- * overflow and the rest of the IV.  This implementation takes NO
- * responsability for checking that the counter doesn't overflow
- * into the rest of the IV when incremented.
- */
 void Camellia_ctr128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char ivec[CAMELLIA_BLOCK_SIZE],
 	unsigned char ecount_buf[CAMELLIA_BLOCK_SIZE],
 	unsigned int *num) 
 	{
 
-	unsigned int n;
-	unsigned long l=length;
-
-	assert(in && out && key && counter && num);
-	assert(*num < CAMELLIA_BLOCK_SIZE);
-
-	n = *num;
-
-	while (l--) 
-		{
-		if (n == 0) 
-			{
-			Camellia_encrypt(ivec, ecount_buf, key);
-			Camellia_ctr128_inc(ivec);
-			}
-		*(out++) = *(in++) ^ ecount_buf[n];
-		n = (n+1) % CAMELLIA_BLOCK_SIZE;
-		}
-
-	*num=n;
+	CRYPTO_ctr128_encrypt(in,out,length,key,ivec,ecount_buf,num,(block128_f)Camellia_encrypt);
 	}
 
diff --git a/crypto/camellia/cmll_locl.h b/crypto/camellia/cmll_locl.h
index 2ac2e9543571..246b6ce1d8cf 100644
--- a/crypto/camellia/cmll_locl.h
+++ b/crypto/camellia/cmll_locl.h
@@ -68,98 +68,19 @@
 #ifndef HEADER_CAMELLIA_LOCL_H
 #define HEADER_CAMELLIA_LOCL_H
 
-#include "openssl/e_os2.h"
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-
+typedef unsigned int  u32;
 typedef unsigned char u8;
-typedef unsigned int u32;
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-#if defined(_MSC_VER) && (defined(_M_IX86) || defined(_M_AMD64) || defined(_M_X64))
-# define SWAP(x) ( _lrotl(x, 8) & 0x00ff00ff | _lrotr(x, 8) & 0xff00ff00 )
-# define GETU32(p) SWAP(*((u32 *)(p)))
-# define PUTU32(ct, st) { *((u32 *)(ct)) = SWAP((st)); }
-# define CAMELLIA_SWAP4(x) (x = ( _lrotl(x, 8) & 0x00ff00ff | _lrotr(x, 8) & 0xff00ff00) )
-
-#else /* not windows */
-# define GETU32(pt) (((u32)(pt)[0] << 24) \
-	^ ((u32)(pt)[1] << 16) \
-	^ ((u32)(pt)[2] <<  8) \
-	^ ((u32)(pt)[3]))
-
-# define PUTU32(ct, st) { (ct)[0] = (u8)((st) >> 24); \
-	(ct)[1] = (u8)((st) >> 16); \
-	(ct)[2] = (u8)((st) >>  8); \
-	(ct)[3] = (u8)(st); }
-
-#if (defined (__GNUC__) && (defined(__x86_64__) || defined(__x86_64)))
-#define CAMELLIA_SWAP4(x) \
-  do{\
-    asm("bswap %1" : "+r" (x));\
-  }while(0)
-#else
-#define CAMELLIA_SWAP4(x) \
-   do{\
-     x = ((u32)x << 16) + ((u32)x >> 16);\
-     x = (((u32)x & 0xff00ff) << 8) + (((u32)x >> 8) & 0xff00ff);\
-   } while(0)
-#endif
-#endif
-
-#define COPY4WORD(dst, src)	 \
-	     do			 \
-		     {		 \
-		     (dst)[0]=(src)[0];		\
-		     (dst)[1]=(src)[1];		\
-		     (dst)[2]=(src)[2];		\
-		     (dst)[3]=(src)[3];		\
-		     }while(0)
-
-#define SWAP4WORD(word)				\
-   do						\
-	   {					\
-	   CAMELLIA_SWAP4((word)[0]);			\
-	   CAMELLIA_SWAP4((word)[1]);			\
-	   CAMELLIA_SWAP4((word)[2]);			\
-	   CAMELLIA_SWAP4((word)[3]);			\
-	   }while(0)
-
-#define XOR4WORD(a, b)/* a = a ^ b */		\
-   do						\
-	{					\
-	(a)[0]^=(b)[0];				\
-	(a)[1]^=(b)[1];				\
-	(a)[2]^=(b)[2];				\
-	(a)[3]^=(b)[3];				\
-	}while(0)
-
-#define XOR4WORD2(a, b, c)/* a = b ^ c */	\
-   do						\
-	{					\
-	(a)[0]=(b)[0]^(c)[0];			\
-	(a)[1]=(b)[1]^(c)[1];				\
-	(a)[2]=(b)[2]^(c)[2];				\
-	(a)[3]=(b)[3]^(c)[3];				\
-	}while(0)
-
-
-void camellia_setup128(const u8 *key, u32 *subkey);
-void camellia_setup192(const u8 *key, u32 *subkey);
-void camellia_setup256(const u8 *key, u32 *subkey);
-
-void camellia_encrypt128(const u32 *subkey, u32 *io);
-void camellia_decrypt128(const u32 *subkey, u32 *io);
-void camellia_encrypt256(const u32 *subkey, u32 *io);
-void camellia_decrypt256(const u32 *subkey, u32 *io);
-
-#ifdef __cplusplus
-}
-#endif
 
+int Camellia_Ekeygen(int keyBitLength, const u8 *rawKey,
+		     KEY_TABLE_TYPE keyTable);
+void Camellia_EncryptBlock_Rounds(int grandRounds, const u8 plaintext[], 
+		const KEY_TABLE_TYPE keyTable, u8 ciphertext[]);
+void Camellia_DecryptBlock_Rounds(int grandRounds, const u8 ciphertext[], 
+		const KEY_TABLE_TYPE keyTable, u8 plaintext[]);
+void Camellia_EncryptBlock(int keyBitLength, const u8 plaintext[], 
+		const KEY_TABLE_TYPE keyTable, u8 ciphertext[]);
+void Camellia_DecryptBlock(int keyBitLength, const u8 ciphertext[], 
+		const KEY_TABLE_TYPE keyTable, u8 plaintext[]);
+int private_Camellia_set_key(const unsigned char *userKey, const int bits,
+			     CAMELLIA_KEY *key);
 #endif /* #ifndef HEADER_CAMELLIA_LOCL_H */
-
diff --git a/crypto/camellia/cmll_misc.c b/crypto/camellia/cmll_misc.c
index 2cd7aba9bbd4..f44d48564c2a 100644
--- a/crypto/camellia/cmll_misc.c
+++ b/crypto/camellia/cmll_misc.c
@@ -50,80 +50,31 @@
  */
  
 #include <openssl/opensslv.h>
+#include <openssl/crypto.h>
 #include <openssl/camellia.h>
 #include "cmll_locl.h"
-#include <openssl/crypto.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
 
 const char CAMELLIA_version[]="CAMELLIA" OPENSSL_VERSION_PTEXT;
 
-int Camellia_set_key(const unsigned char *userKey, const int bits,
-	CAMELLIA_KEY *key)
-#ifdef OPENSSL_FIPS
-	{
-	if (FIPS_mode())
-		FIPS_BAD_ABORT(CAMELLIA)
-	return private_Camellia_set_key(userKey, bits, key);
-	}
 int private_Camellia_set_key(const unsigned char *userKey, const int bits,
 	CAMELLIA_KEY *key)
-#endif
 	{
-	if (!userKey || !key)
-		{
+	if(!userKey || !key)
 		return -1;
-		}
-	
-	switch(bits)
-		{
-	case 128:
-		camellia_setup128(userKey, (unsigned int *)key->rd_key);
-		key->enc = camellia_encrypt128;
-		key->dec = camellia_decrypt128;
-		break;
-	case 192:
-		camellia_setup192(userKey, (unsigned int *)key->rd_key);
-		key->enc = camellia_encrypt256;
-		key->dec = camellia_decrypt256;
-		break;
-	case 256:
-		camellia_setup256(userKey, (unsigned int *)key->rd_key);
-		key->enc = camellia_encrypt256;
-		key->dec = camellia_decrypt256;
-		break;
-	default:
+	if(bits != 128 && bits != 192 && bits != 256)
 		return -2;
-		}
-	
-	key->bitLength = bits;
+	key->grand_rounds = Camellia_Ekeygen(bits , userKey, key->u.rd_key);
 	return 0;
 	}
 
 void Camellia_encrypt(const unsigned char *in, unsigned char *out,
 	const CAMELLIA_KEY *key)
 	{
-	u32 tmp[CAMELLIA_BLOCK_SIZE/sizeof(u32)];
-	const union { long one; char little; } camellia_endian = {1};
-
-	memcpy(tmp, in, CAMELLIA_BLOCK_SIZE);
-	if (camellia_endian.little) SWAP4WORD(tmp);
-	key->enc(key->rd_key, tmp);
-	if (camellia_endian.little) SWAP4WORD(tmp);
-	memcpy(out, tmp, CAMELLIA_BLOCK_SIZE);
+	Camellia_EncryptBlock_Rounds(key->grand_rounds, in , key->u.rd_key , out);
 	}
 
 void Camellia_decrypt(const unsigned char *in, unsigned char *out,
 	const CAMELLIA_KEY *key)
 	{
-	u32 tmp[CAMELLIA_BLOCK_SIZE/sizeof(u32)];
-	const union { long one; char little; } camellia_endian = {1};
-
-	memcpy(tmp, in, CAMELLIA_BLOCK_SIZE);
-	if (camellia_endian.little) SWAP4WORD(tmp);
-	key->dec(key->rd_key, tmp);
-	if (camellia_endian.little) SWAP4WORD(tmp);
-	memcpy(out, tmp, CAMELLIA_BLOCK_SIZE);
+	Camellia_DecryptBlock_Rounds(key->grand_rounds, in , key->u.rd_key , out);
 	}
-
diff --git a/crypto/camellia/cmll_ofb.c b/crypto/camellia/cmll_ofb.c
index d89cf9f3b324..a482befc7470 100644
--- a/crypto/camellia/cmll_ofb.c
+++ b/crypto/camellia/cmll_ofb.c
@@ -105,37 +105,15 @@
  * [including the GNU Public Licence.]
  */
 
-#ifndef CAMELLIA_DEBUG
-# ifndef NDEBUG
-#  define NDEBUG
-# endif
-#endif
-#include <assert.h>
 #include <openssl/camellia.h>
-#include "cmll_locl.h"
+#include <openssl/modes.h>
 
 /* The input and output encrypted as though 128bit ofb mode is being
  * used.  The extra state information to record how much of the
  * 128bit block we have used is contained in *num;
  */
 void Camellia_ofb128_encrypt(const unsigned char *in, unsigned char *out,
-	const unsigned long length, const CAMELLIA_KEY *key,
+	size_t length, const CAMELLIA_KEY *key,
 	unsigned char *ivec, int *num) {
-
-	unsigned int n;
-	unsigned long l=length;
-
-	assert(in && out && key && ivec && num);
-
-	n = *num;
-
-	while (l--) {
-		if (n == 0) {
-			Camellia_encrypt(ivec, ivec, key);
-		}
-		*(out++) = *(in++) ^ ivec[n];
-		n = (n+1) % CAMELLIA_BLOCK_SIZE;
-	}
-
-	*num=n;
+	CRYPTO_ofb128_encrypt(in,out,length,key,ivec,num,(block128_f)Camellia_encrypt);
 }
diff --git a/crypto/camellia/cmll_utl.c b/crypto/camellia/cmll_utl.c
new file mode 100644
index 000000000000..7a35711ec1c2
--- /dev/null
+++ b/crypto/camellia/cmll_utl.c
@@ -0,0 +1,64 @@
+/* crypto/camellia/cmll_utl.c -*- mode:C; c-file-style: "eay" -*- */
+/* ====================================================================
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ */
+ 
+#include <openssl/opensslv.h>
+#include <openssl/crypto.h>
+#include <openssl/camellia.h>
+#include "cmll_locl.h"
+
+int Camellia_set_key(const unsigned char *userKey, const int bits,
+	CAMELLIA_KEY *key)
+	{
+#ifdef OPENSSL_FIPS
+	fips_cipher_abort(Camellia);
+#endif
+	return private_Camellia_set_key(userKey, bits, key);
+	}
diff --git a/crypto/cast/Makefile b/crypto/cast/Makefile
index 2e026dbe0da9..f3f4859886a6 100644
--- a/crypto/cast/Makefile
+++ b/crypto/cast/Makefile
@@ -38,19 +38,12 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-# ELF
-cx86-elf.s: asm/cast-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) cast-586.pl elf $(CLAGS) $(PROCESSOR) > ../$@)
-# COFF
-cx86-cof.s: asm/cast-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) cast-586.pl coff $(CLAGS) $(PROCESSOR) > ../$@)
-# a.out
-cx86-out.s: asm/cast-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) cast-586.pl a.out $(CLAGS) $(PROCESSOR) > ../$@)
+cast-586.s:	asm/cast-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
+	$(PERL)	asm/cast-586.pl $(PERLASM_SCHEME) $(CLAGS) $(PROCESSOR) > $@
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
@@ -103,7 +96,7 @@ c_ofb64.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
 c_ofb64.o: c_ofb64.c cast_lcl.h
 c_skey.o: ../../e_os.h ../../include/openssl/cast.h
 c_skey.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
-c_skey.o: ../../include/openssl/fips.h ../../include/openssl/opensslconf.h
-c_skey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-c_skey.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-c_skey.o: ../../include/openssl/symhacks.h c_skey.c cast_lcl.h cast_s.h
+c_skey.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+c_skey.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+c_skey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+c_skey.o: c_skey.c cast_lcl.h cast_s.h
diff --git a/crypto/cast/asm/cast-586.pl b/crypto/cast/asm/cast-586.pl
index 6be0bfe57245..bf6810d335f2 100644
--- a/crypto/cast/asm/cast-586.pl
+++ b/crypto/cast/asm/cast-586.pl
@@ -3,7 +3,8 @@
 # define for pentium pro friendly version
 $ppro=1;
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 require "cbc.pl";
 
diff --git a/crypto/cast/c_skey.c b/crypto/cast/c_skey.c
index 68e690a60c2b..cb6bf9fee37f 100644
--- a/crypto/cast/c_skey.c
+++ b/crypto/cast/c_skey.c
@@ -56,12 +56,8 @@
  * [including the GNU Public Licence.]
  */
 
-#include <openssl/cast.h>
 #include <openssl/crypto.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/cast.h>
 #include "cast_lcl.h"
 #include "cast_s.h"
 
@@ -76,8 +72,14 @@
 #define S5 CAST_S_table5
 #define S6 CAST_S_table6
 #define S7 CAST_S_table7
-
-FIPS_NON_FIPS_VCIPHER_Init(CAST)
+void CAST_set_key(CAST_KEY *key, int len, const unsigned char *data)
+#ifdef OPENSSL_FIPS
+	{
+	fips_cipher_abort(CAST);
+	private_CAST_set_key(key, len, data);
+	}
+void private_CAST_set_key(CAST_KEY *key, int len, const unsigned char *data)
+#endif
 	{
 	CAST_LONG x[16];
 	CAST_LONG z[16];
diff --git a/crypto/cast/cast.h b/crypto/cast/cast.h
index 6e0cd31ae447..203922ea2b4c 100644
--- a/crypto/cast/cast.h
+++ b/crypto/cast/cast.h
@@ -72,7 +72,7 @@ extern "C" {
 #define CAST_ENCRYPT	1
 #define CAST_DECRYPT	0
 
-#define CAST_LONG unsigned long
+#define CAST_LONG unsigned int
 
 #define CAST_BLOCK	8
 #define CAST_KEY_LENGTH	16
diff --git a/crypto/cast/casts.cpp b/crypto/cast/casts.cpp
deleted file mode 100644
index 8d7bd468d229..000000000000
--- a/crypto/cast/casts.cpp
+++ /dev/null
@@ -1,70 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/cast.h>
-
-void main(int argc,char *argv[])
-	{
-	CAST_KEY key;
-	unsigned long s1,s2,e1,e2;
-	unsigned long data[2];
-	int i,j;
-	static unsigned char d[16]={0x01,0x23,0x45,0x67,0x89,0xAB,0xCD,0xEF};
-
-	CAST_set_key(&key, 16,d);
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<1000; i++) /**/
-			{
-			CAST_encrypt(&data[0],&key);
-			GetTSC(s1);
-			CAST_encrypt(&data[0],&key);
-			CAST_encrypt(&data[0],&key);
-			CAST_encrypt(&data[0],&key);
-			GetTSC(e1);
-			GetTSC(s2);
-			CAST_encrypt(&data[0],&key);
-			CAST_encrypt(&data[0],&key);
-			CAST_encrypt(&data[0],&key);
-			CAST_encrypt(&data[0],&key);
-			GetTSC(e2);
-			CAST_encrypt(&data[0],&key);
-			}
-
-		printf("cast %d %d (%d)\n",
-			e1-s1,e2-s2,((e2-s2)-(e1-s1)));
-		}
-	}
-
diff --git a/crypto/cmac/Makefile b/crypto/cmac/Makefile
new file mode 100644
index 000000000000..54e7cc39d52b
--- /dev/null
+++ b/crypto/cmac/Makefile
@@ -0,0 +1,111 @@
+#
+# OpenSSL/crypto/cmac/Makefile
+#
+
+DIR=	cmac
+TOP=	../..
+CC=	cc
+INCLUDES=
+CFLAG=-g
+MAKEFILE=	Makefile
+AR=		ar r
+
+CFLAGS= $(INCLUDES) $(CFLAG)
+
+GENERAL=Makefile
+TEST=
+APPS=
+
+LIB=$(TOP)/libcrypto.a
+LIBSRC=cmac.c cm_ameth.c cm_pmeth.c
+LIBOBJ=cmac.o cm_ameth.o cm_pmeth.o
+
+SRC= $(LIBSRC)
+
+EXHEADER= cmac.h
+HEADER=	$(EXHEADER)
+
+ALL=    $(GENERAL) $(SRC) $(HEADER)
+
+top:
+	(cd ../..; $(MAKE) DIRS=crypto SDIRS=$(DIR) sub_all)
+
+all:	lib
+
+lib:	$(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
+	$(RANLIB) $(LIB) || echo Never mind.
+	@touch lib
+
+files:
+	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
+
+links:
+	@$(PERL) $(TOP)/util/mklink.pl ../../include/openssl $(EXHEADER)
+	@$(PERL) $(TOP)/util/mklink.pl ../../test $(TEST)
+	@$(PERL) $(TOP)/util/mklink.pl ../../apps $(APPS)
+
+install:
+	@[ -n "$(INSTALLTOP)" ] # should be set by top Makefile...
+	@headerlist="$(EXHEADER)"; for i in $$headerlist ; \
+	do  \
+	(cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i; \
+	chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i ); \
+	done;
+
+tags:
+	ctags $(SRC)
+
+tests:
+
+lint:
+	lint -DLINT $(INCLUDES) $(SRC)>fluff
+
+depend:
+	@[ -n "$(MAKEDEPEND)" ] # should be set by upper Makefile...
+	$(MAKEDEPEND) -- $(CFLAG) $(INCLUDES) $(DEPFLAG) -- $(PROGS) $(LIBSRC)
+
+dclean:
+	$(PERL) -pe 'if (/^# DO NOT DELETE THIS LINE/) {print; exit(0);}' $(MAKEFILE) >Makefile.new
+	mv -f Makefile.new $(MAKEFILE)
+
+clean:
+	rm -f *.o *.obj lib tags core .pure .nfs* *.old *.bak fluff
+
+# DO NOT DELETE THIS LINE -- make depend depends on it.
+
+cm_ameth.o: ../../e_os.h ../../include/openssl/asn1.h
+cm_ameth.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+cm_ameth.o: ../../include/openssl/cmac.h ../../include/openssl/crypto.h
+cm_ameth.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
+cm_ameth.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+cm_ameth.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+cm_ameth.o: ../../include/openssl/opensslconf.h
+cm_ameth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cm_ameth.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+cm_ameth.o: ../../include/openssl/symhacks.h ../asn1/asn1_locl.h ../cryptlib.h
+cm_ameth.o: cm_ameth.c
+cm_pmeth.o: ../../e_os.h ../../include/openssl/asn1.h
+cm_pmeth.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+cm_pmeth.o: ../../include/openssl/cmac.h ../../include/openssl/conf.h
+cm_pmeth.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+cm_pmeth.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+cm_pmeth.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+cm_pmeth.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+cm_pmeth.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+cm_pmeth.o: ../../include/openssl/opensslconf.h
+cm_pmeth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cm_pmeth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+cm_pmeth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+cm_pmeth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+cm_pmeth.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+cm_pmeth.o: ../cryptlib.h ../evp/evp_locl.h cm_pmeth.c
+cmac.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
+cmac.o: ../../include/openssl/buffer.h ../../include/openssl/cmac.h
+cmac.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+cmac.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+cmac.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+cmac.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+cmac.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cmac.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+cmac.o: ../../include/openssl/symhacks.h ../cryptlib.h cmac.c
diff --git a/crypto/cmac/cm_ameth.c b/crypto/cmac/cm_ameth.c
new file mode 100644
index 000000000000..0b8e5670b0ec
--- /dev/null
+++ b/crypto/cmac/cm_ameth.c
@@ -0,0 +1,97 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2010.
+ */
+/* ====================================================================
+ * Copyright (c) 2010 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/evp.h>
+#include <openssl/cmac.h>
+#include "asn1_locl.h"
+
+/* CMAC "ASN1" method. This is just here to indicate the
+ * maximum CMAC output length and to free up a CMAC
+ * key.
+ */
+
+static int cmac_size(const EVP_PKEY *pkey)
+	{
+	return EVP_MAX_BLOCK_LENGTH;
+	}
+
+static void cmac_key_free(EVP_PKEY *pkey)
+	{
+	CMAC_CTX *cmctx = (CMAC_CTX *)pkey->pkey.ptr;
+	if (cmctx)
+		CMAC_CTX_free(cmctx);
+	}
+
+const EVP_PKEY_ASN1_METHOD cmac_asn1_meth = 
+	{
+	EVP_PKEY_CMAC,
+	EVP_PKEY_CMAC,
+	0,
+
+	"CMAC",
+	"OpenSSL CMAC method",
+
+	0,0,0,0,
+
+	0,0,0,
+
+	cmac_size,
+	0,
+	0,0,0,0,0,0,0,
+
+	cmac_key_free,
+	0,
+	0,0
+	};
+
diff --git a/crypto/cmac/cm_pmeth.c b/crypto/cmac/cm_pmeth.c
new file mode 100644
index 000000000000..072228ec7fa4
--- /dev/null
+++ b/crypto/cmac/cm_pmeth.c
@@ -0,0 +1,224 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2010.
+ */
+/* ====================================================================
+ * Copyright (c) 2010 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/x509.h>
+#include <openssl/x509v3.h>
+#include <openssl/evp.h>
+#include <openssl/cmac.h>
+#include "evp_locl.h"
+
+/* The context structure and "key" is simply a CMAC_CTX */
+
+static int pkey_cmac_init(EVP_PKEY_CTX *ctx)
+	{
+	ctx->data = CMAC_CTX_new();
+	if (!ctx->data)
+		return 0;
+	ctx->keygen_info_count = 0;
+	return 1;
+	}
+
+static int pkey_cmac_copy(EVP_PKEY_CTX *dst, EVP_PKEY_CTX *src)
+	{
+	if (!pkey_cmac_init(dst))
+		return 0;
+	if (!CMAC_CTX_copy(dst->data, src->data))
+		return 0;
+	return 1;
+	}
+
+static void pkey_cmac_cleanup(EVP_PKEY_CTX *ctx)
+	{
+	CMAC_CTX_free(ctx->data);
+	}
+
+static int pkey_cmac_keygen(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey)
+	{
+	CMAC_CTX *cmkey = CMAC_CTX_new();
+	CMAC_CTX *cmctx = ctx->data;
+	if (!cmkey)
+		return 0;
+	if (!CMAC_CTX_copy(cmkey, cmctx))
+		{
+		CMAC_CTX_free(cmkey);
+		return 0;
+		}
+	EVP_PKEY_assign(pkey, EVP_PKEY_CMAC, cmkey);
+	
+	return 1;
+	}
+
+static int int_update(EVP_MD_CTX *ctx,const void *data,size_t count)
+	{
+	if (!CMAC_Update(ctx->pctx->data, data, count))
+		return 0;
+	return 1;
+	}
+
+static int cmac_signctx_init(EVP_PKEY_CTX *ctx, EVP_MD_CTX *mctx)
+	{
+	EVP_MD_CTX_set_flags(mctx, EVP_MD_CTX_FLAG_NO_INIT);
+	mctx->update = int_update;
+	return 1;
+	}
+
+static int cmac_signctx(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					EVP_MD_CTX *mctx)
+	{
+	return CMAC_Final(ctx->data, sig, siglen);
+	}
+
+static int pkey_cmac_ctrl(EVP_PKEY_CTX *ctx, int type, int p1, void *p2)
+	{
+	CMAC_CTX *cmctx = ctx->data;
+	switch (type)
+		{
+
+		case EVP_PKEY_CTRL_SET_MAC_KEY:
+		if (!p2 || p1 < 0)
+			return 0;
+		if (!CMAC_Init(cmctx, p2, p1, NULL, NULL))
+			return 0;
+		break;
+
+		case EVP_PKEY_CTRL_CIPHER:
+		if (!CMAC_Init(cmctx, NULL, 0, p2, ctx->engine))
+			return 0;
+		break;
+
+		case EVP_PKEY_CTRL_MD:
+		if (ctx->pkey && !CMAC_CTX_copy(ctx->data,
+					(CMAC_CTX *)ctx->pkey->pkey.ptr))
+			return 0;
+		if (!CMAC_Init(cmctx, NULL, 0, NULL, NULL))
+			return 0;
+		break;
+
+		default:
+		return -2;
+
+		}
+	return 1;
+	}
+
+static int pkey_cmac_ctrl_str(EVP_PKEY_CTX *ctx,
+			const char *type, const char *value)
+	{
+	if (!value)
+		{
+		return 0;
+		}
+	if (!strcmp(type, "key"))
+		{
+		void *p = (void *)value;
+		return pkey_cmac_ctrl(ctx, EVP_PKEY_CTRL_SET_MAC_KEY,
+								strlen(p), p);
+		}
+	if (!strcmp(type, "cipher"))
+		{
+		const EVP_CIPHER *c;
+		c = EVP_get_cipherbyname(value);
+		if (!c)
+			return 0;
+		return pkey_cmac_ctrl(ctx, EVP_PKEY_CTRL_CIPHER, -1, (void *)c);
+		}
+	if (!strcmp(type, "hexkey"))
+		{
+		unsigned char *key;
+		int r;
+		long keylen;
+		key = string_to_hex(value, &keylen);
+		if (!key)
+			return 0;
+		r = pkey_cmac_ctrl(ctx, EVP_PKEY_CTRL_SET_MAC_KEY, keylen, key);
+		OPENSSL_free(key);
+		return r;
+		}
+	return -2;
+	}
+
+const EVP_PKEY_METHOD cmac_pkey_meth = 
+	{
+	EVP_PKEY_CMAC,
+	EVP_PKEY_FLAG_SIGCTX_CUSTOM,
+	pkey_cmac_init,
+	pkey_cmac_copy,
+	pkey_cmac_cleanup,
+
+	0, 0,
+
+	0,
+	pkey_cmac_keygen,
+
+	0, 0,
+
+	0, 0,
+
+	0,0,
+
+	cmac_signctx_init,
+	cmac_signctx,
+
+	0,0,
+
+	0,0,
+
+	0,0,
+
+	0,0,
+
+	pkey_cmac_ctrl,
+	pkey_cmac_ctrl_str
+
+	};
diff --git a/crypto/cmac/cmac.c b/crypto/cmac/cmac.c
new file mode 100644
index 000000000000..8b72b096813c
--- /dev/null
+++ b/crypto/cmac/cmac.c
@@ -0,0 +1,308 @@
+/* crypto/cmac/cmac.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project.
+ */
+/* ====================================================================
+ * Copyright (c) 2010 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include "cryptlib.h"
+#include <openssl/cmac.h>
+
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
+
+struct CMAC_CTX_st
+	{
+	/* Cipher context to use */
+	EVP_CIPHER_CTX cctx;
+	/* Keys k1 and k2 */
+	unsigned char k1[EVP_MAX_BLOCK_LENGTH];
+	unsigned char k2[EVP_MAX_BLOCK_LENGTH];
+	/* Temporary block */
+	unsigned char tbl[EVP_MAX_BLOCK_LENGTH];
+	/* Last (possibly partial) block */
+	unsigned char last_block[EVP_MAX_BLOCK_LENGTH];
+	/* Number of bytes in last block: -1 means context not initialised */
+	int nlast_block;
+	};
+
+
+/* Make temporary keys K1 and K2 */
+
+static void make_kn(unsigned char *k1, unsigned char *l, int bl)
+	{
+	int i;
+	/* Shift block to left, including carry */
+	for (i = 0; i < bl; i++)
+		{
+		k1[i] = l[i] << 1;
+		if (i < bl - 1 && l[i + 1] & 0x80)
+			k1[i] |= 1;
+		}
+	/* If MSB set fixup with R */
+	if (l[0] & 0x80)
+		k1[bl - 1] ^= bl == 16 ? 0x87 : 0x1b;
+	}
+
+CMAC_CTX *CMAC_CTX_new(void)
+	{
+	CMAC_CTX *ctx;
+	ctx = OPENSSL_malloc(sizeof(CMAC_CTX));
+	if (!ctx)
+		return NULL;
+	EVP_CIPHER_CTX_init(&ctx->cctx);
+	ctx->nlast_block = -1;
+	return ctx;
+	}
+
+void CMAC_CTX_cleanup(CMAC_CTX *ctx)
+	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !ctx->cctx.engine)
+		{
+		FIPS_cmac_ctx_cleanup(ctx);
+		return;
+		}
+#endif
+	EVP_CIPHER_CTX_cleanup(&ctx->cctx);
+	OPENSSL_cleanse(ctx->tbl, EVP_MAX_BLOCK_LENGTH);
+	OPENSSL_cleanse(ctx->k1, EVP_MAX_BLOCK_LENGTH);
+	OPENSSL_cleanse(ctx->k2, EVP_MAX_BLOCK_LENGTH);
+	OPENSSL_cleanse(ctx->last_block, EVP_MAX_BLOCK_LENGTH);
+	ctx->nlast_block = -1;
+	}
+
+EVP_CIPHER_CTX *CMAC_CTX_get0_cipher_ctx(CMAC_CTX *ctx)
+	{
+	return &ctx->cctx;
+	}
+
+void CMAC_CTX_free(CMAC_CTX *ctx)
+	{
+	CMAC_CTX_cleanup(ctx);
+	OPENSSL_free(ctx);
+	}
+
+int CMAC_CTX_copy(CMAC_CTX *out, const CMAC_CTX *in)
+	{
+	int bl;
+	if (in->nlast_block == -1)
+		return 0;
+	if (!EVP_CIPHER_CTX_copy(&out->cctx, &in->cctx))
+		return 0;
+	bl = EVP_CIPHER_CTX_block_size(&in->cctx);
+	memcpy(out->k1, in->k1, bl);
+	memcpy(out->k2, in->k2, bl);
+	memcpy(out->tbl, in->tbl, bl);
+	memcpy(out->last_block, in->last_block, bl);
+	out->nlast_block = in->nlast_block;
+	return 1;
+	}
+
+int CMAC_Init(CMAC_CTX *ctx, const void *key, size_t keylen, 
+			const EVP_CIPHER *cipher, ENGINE *impl)
+	{
+	static unsigned char zero_iv[EVP_MAX_BLOCK_LENGTH];
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode())
+		{
+		/* If we have an ENGINE need to allow non FIPS */
+		if ((impl || ctx->cctx.engine)
+			&& !(ctx->cctx.flags & EVP_CIPH_FLAG_NON_FIPS_ALLOW))
+
+			{
+			EVPerr(EVP_F_CMAC_INIT, EVP_R_DISABLED_FOR_FIPS);
+			return 0;
+			}
+		/* Other algorithm blocking will be done in FIPS_cmac_init,
+		 * via FIPS_cipherinit().
+		 */
+		if (!impl && !ctx->cctx.engine)
+			return FIPS_cmac_init(ctx, key, keylen, cipher, NULL);
+		}
+#endif
+	/* All zeros means restart */
+	if (!key && !cipher && !impl && keylen == 0)
+		{
+		/* Not initialised */
+		if (ctx->nlast_block == -1)
+			return 0;
+		if (!EVP_EncryptInit_ex(&ctx->cctx, NULL, NULL, NULL, zero_iv))
+			return 0;
+		memset(ctx->tbl, 0, EVP_CIPHER_CTX_block_size(&ctx->cctx));
+		ctx->nlast_block = 0;
+		return 1;
+		}
+	/* Initialiase context */
+	if (cipher && !EVP_EncryptInit_ex(&ctx->cctx, cipher, impl, NULL, NULL))
+		return 0;
+	/* Non-NULL key means initialisation complete */
+	if (key)
+		{
+		int bl;
+		if (!EVP_CIPHER_CTX_cipher(&ctx->cctx))
+			return 0;
+		if (!EVP_CIPHER_CTX_set_key_length(&ctx->cctx, keylen))
+			return 0;
+		if (!EVP_EncryptInit_ex(&ctx->cctx, NULL, NULL, key, zero_iv))
+			return 0;
+		bl = EVP_CIPHER_CTX_block_size(&ctx->cctx);
+		if (!EVP_Cipher(&ctx->cctx, ctx->tbl, zero_iv, bl))
+			return 0;
+		make_kn(ctx->k1, ctx->tbl, bl);
+		make_kn(ctx->k2, ctx->k1, bl);
+		OPENSSL_cleanse(ctx->tbl, bl);
+		/* Reset context again ready for first data block */
+		if (!EVP_EncryptInit_ex(&ctx->cctx, NULL, NULL, NULL, zero_iv))
+			return 0;
+		/* Zero tbl so resume works */
+		memset(ctx->tbl, 0, bl);
+		ctx->nlast_block = 0;
+		}
+	return 1;
+	}
+
+int CMAC_Update(CMAC_CTX *ctx, const void *in, size_t dlen)
+	{
+	const unsigned char *data = in;
+	size_t bl;
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !ctx->cctx.engine)
+		return FIPS_cmac_update(ctx, in, dlen);
+#endif
+	if (ctx->nlast_block == -1)
+		return 0;
+	if (dlen == 0)
+		return 1;
+	bl = EVP_CIPHER_CTX_block_size(&ctx->cctx);
+	/* Copy into partial block if we need to */
+	if (ctx->nlast_block > 0)
+		{
+		size_t nleft;
+		nleft = bl - ctx->nlast_block;
+		if (dlen < nleft)
+			nleft = dlen;
+		memcpy(ctx->last_block + ctx->nlast_block, data, nleft);
+		dlen -= nleft;
+		ctx->nlast_block += nleft;
+		/* If no more to process return */
+		if (dlen == 0)
+			return 1;
+		data += nleft;
+		/* Else not final block so encrypt it */
+		if (!EVP_Cipher(&ctx->cctx, ctx->tbl, ctx->last_block,bl))
+			return 0;
+		}
+	/* Encrypt all but one of the complete blocks left */
+	while(dlen > bl)
+		{
+		if (!EVP_Cipher(&ctx->cctx, ctx->tbl, data, bl))
+			return 0;
+		dlen -= bl;
+		data += bl;
+		}
+	/* Copy any data left to last block buffer */
+	memcpy(ctx->last_block, data, dlen);
+	ctx->nlast_block = dlen;
+	return 1;
+
+	}
+
+int CMAC_Final(CMAC_CTX *ctx, unsigned char *out, size_t *poutlen)
+	{
+	int i, bl, lb;
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !ctx->cctx.engine)
+		return FIPS_cmac_final(ctx, out, poutlen);
+#endif
+	if (ctx->nlast_block == -1)
+		return 0;
+	bl = EVP_CIPHER_CTX_block_size(&ctx->cctx);
+	*poutlen = (size_t)bl;
+	if (!out)
+		return 1;
+	lb = ctx->nlast_block;
+	/* Is last block complete? */
+	if (lb == bl)
+		{
+		for (i = 0; i < bl; i++)
+			out[i] = ctx->last_block[i] ^ ctx->k1[i];
+		}
+	else
+		{
+		ctx->last_block[lb] = 0x80;
+		if (bl - lb > 1)
+			memset(ctx->last_block + lb + 1, 0, bl - lb - 1);
+		for (i = 0; i < bl; i++)
+			out[i] = ctx->last_block[i] ^ ctx->k2[i];
+		}
+	if (!EVP_Cipher(&ctx->cctx, out, out, bl))
+		{
+		OPENSSL_cleanse(out, bl);	
+		return 0;
+		}
+	return 1;
+	}
+
+int CMAC_resume(CMAC_CTX *ctx)
+	{
+	if (ctx->nlast_block == -1)
+		return 0;
+	/* The buffer "tbl" containes the last fully encrypted block
+	 * which is the last IV (or all zeroes if no last encrypted block).
+	 * The last block has not been modified since CMAC_final().
+	 * So reinitliasing using the last decrypted block will allow
+	 * CMAC to continue after calling CMAC_Final(). 
+	 */
+	return EVP_EncryptInit_ex(&ctx->cctx, NULL, NULL, NULL, ctx->tbl);
+	}
diff --git a/crypto/cmac/cmac.h b/crypto/cmac/cmac.h
new file mode 100644
index 000000000000..712e92dced2c
--- /dev/null
+++ b/crypto/cmac/cmac.h
@@ -0,0 +1,82 @@
+/* crypto/cmac/cmac.h */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project.
+ */
+/* ====================================================================
+ * Copyright (c) 2010 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+
+#ifndef HEADER_CMAC_H
+#define HEADER_CMAC_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <openssl/evp.h>
+
+/* Opaque */
+typedef struct CMAC_CTX_st CMAC_CTX;
+
+CMAC_CTX *CMAC_CTX_new(void);
+void CMAC_CTX_cleanup(CMAC_CTX *ctx);
+void CMAC_CTX_free(CMAC_CTX *ctx);
+EVP_CIPHER_CTX *CMAC_CTX_get0_cipher_ctx(CMAC_CTX *ctx);
+int CMAC_CTX_copy(CMAC_CTX *out, const CMAC_CTX *in);
+
+int CMAC_Init(CMAC_CTX *ctx, const void *key, size_t keylen, 
+			const EVP_CIPHER *cipher, ENGINE *impl);
+int CMAC_Update(CMAC_CTX *ctx, const void *data, size_t dlen);
+int CMAC_Final(CMAC_CTX *ctx, unsigned char *out, size_t *poutlen);
+int CMAC_resume(CMAC_CTX *ctx);
+
+#ifdef  __cplusplus
+}
+#endif
+#endif
diff --git a/crypto/cms/Makefile b/crypto/cms/Makefile
index 1c137e0cff6a..9820adb21278 100644
--- a/crypto/cms/Makefile
+++ b/crypto/cms/Makefile
@@ -18,9 +18,11 @@ APPS=
 
 LIB=$(TOP)/libcrypto.a
 LIBSRC= cms_lib.c cms_asn1.c cms_att.c cms_io.c cms_smime.c cms_err.c \
-	cms_sd.c cms_dd.c cms_cd.c cms_env.c cms_enc.c cms_ess.c
+	cms_sd.c cms_dd.c cms_cd.c cms_env.c cms_enc.c cms_ess.c \
+	cms_pwri.c
 LIBOBJ= cms_lib.o cms_asn1.o cms_att.o cms_io.o cms_smime.o cms_err.o \
-	cms_sd.o cms_dd.o cms_cd.o cms_env.o cms_enc.o cms_ess.o
+	cms_sd.o cms_dd.o cms_cd.o cms_env.o cms_enc.o cms_ess.o \
+	cms_pwri.o
 
 SRC= $(LIBSRC)
 
@@ -37,7 +39,7 @@ test:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -108,6 +110,71 @@ cms_att.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 cms_att.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 cms_att.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
 cms_att.o: cms.h cms_att.c cms_lcl.h
+cms_cd.o: ../../e_os.h ../../include/openssl/asn1.h
+cms_cd.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+cms_cd.o: ../../include/openssl/buffer.h ../../include/openssl/cms.h
+cms_cd.o: ../../include/openssl/comp.h ../../include/openssl/conf.h
+cms_cd.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+cms_cd.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+cms_cd.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+cms_cd.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+cms_cd.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+cms_cd.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+cms_cd.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
+cms_cd.o: ../../include/openssl/pem2.h ../../include/openssl/pkcs7.h
+cms_cd.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+cms_cd.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+cms_cd.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+cms_cd.o: ../../include/openssl/x509v3.h ../cryptlib.h cms_cd.c cms_lcl.h
+cms_dd.o: ../../e_os.h ../../include/openssl/asn1.h
+cms_dd.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+cms_dd.o: ../../include/openssl/buffer.h ../../include/openssl/cms.h
+cms_dd.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+cms_dd.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+cms_dd.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+cms_dd.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+cms_dd.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+cms_dd.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+cms_dd.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cms_dd.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
+cms_dd.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+cms_dd.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+cms_dd.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+cms_dd.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+cms_dd.o: ../cryptlib.h cms_dd.c cms_lcl.h
+cms_enc.o: ../../e_os.h ../../include/openssl/asn1.h
+cms_enc.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+cms_enc.o: ../../include/openssl/buffer.h ../../include/openssl/cms.h
+cms_enc.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+cms_enc.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+cms_enc.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+cms_enc.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+cms_enc.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+cms_enc.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+cms_enc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cms_enc.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
+cms_enc.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+cms_enc.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+cms_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+cms_enc.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+cms_enc.o: ../../include/openssl/x509v3.h ../cryptlib.h cms_enc.c cms_lcl.h
+cms_env.o: ../../e_os.h ../../include/openssl/aes.h
+cms_env.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
+cms_env.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+cms_env.o: ../../include/openssl/cms.h ../../include/openssl/conf.h
+cms_env.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+cms_env.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+cms_env.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+cms_env.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+cms_env.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+cms_env.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+cms_env.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
+cms_env.o: ../../include/openssl/pem2.h ../../include/openssl/pkcs7.h
+cms_env.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
+cms_env.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+cms_env.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+cms_env.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+cms_env.o: ../asn1/asn1_locl.h ../cryptlib.h cms_env.c cms_lcl.h
 cms_err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 cms_err.o: ../../include/openssl/buffer.h ../../include/openssl/cms.h
 cms_err.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
@@ -121,6 +188,22 @@ cms_err.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 cms_err.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 cms_err.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
 cms_err.o: cms_err.c
+cms_ess.o: ../../e_os.h ../../include/openssl/asn1.h
+cms_ess.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+cms_ess.o: ../../include/openssl/buffer.h ../../include/openssl/cms.h
+cms_ess.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+cms_ess.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+cms_ess.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+cms_ess.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+cms_ess.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+cms_ess.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+cms_ess.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cms_ess.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
+cms_ess.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+cms_ess.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+cms_ess.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+cms_ess.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+cms_ess.o: ../../include/openssl/x509v3.h ../cryptlib.h cms_ess.c cms_lcl.h
 cms_io.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
 cms_io.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 cms_io.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
@@ -149,6 +232,24 @@ cms_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 cms_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 cms_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h cms.h
 cms_lib.o: cms_lcl.h cms_lib.c
+cms_pwri.o: ../../e_os.h ../../include/openssl/aes.h
+cms_pwri.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
+cms_pwri.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+cms_pwri.o: ../../include/openssl/cms.h ../../include/openssl/conf.h
+cms_pwri.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+cms_pwri.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+cms_pwri.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+cms_pwri.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+cms_pwri.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+cms_pwri.o: ../../include/openssl/opensslconf.h
+cms_pwri.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cms_pwri.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
+cms_pwri.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+cms_pwri.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+cms_pwri.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+cms_pwri.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+cms_pwri.o: ../../include/openssl/x509v3.h ../asn1/asn1_locl.h ../cryptlib.h
+cms_pwri.o: cms_lcl.h cms_pwri.c
 cms_sd.o: ../../e_os.h ../../include/openssl/asn1.h
 cms_sd.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 cms_sd.o: ../../include/openssl/buffer.h ../../include/openssl/cms.h
@@ -164,7 +265,7 @@ cms_sd.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 cms_sd.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 cms_sd.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 cms_sd.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-cms_sd.o: ../cryptlib.h cms_lcl.h cms_sd.c
+cms_sd.o: ../asn1/asn1_locl.h ../cryptlib.h cms_lcl.h cms_sd.c
 cms_smime.o: ../../e_os.h ../../include/openssl/asn1.h
 cms_smime.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 cms_smime.o: ../../include/openssl/buffer.h ../../include/openssl/cms.h
diff --git a/crypto/cms/cms.h b/crypto/cms/cms.h
index 75e3be0e4bc5..36994fa6a270 100644
--- a/crypto/cms/cms.h
+++ b/crypto/cms/cms.h
@@ -76,8 +76,9 @@ typedef struct CMS_Receipt_st CMS_Receipt;
 
 DECLARE_STACK_OF(CMS_SignerInfo)
 DECLARE_STACK_OF(GENERAL_NAMES)
-DECLARE_ASN1_FUNCTIONS_const(CMS_ContentInfo)
-DECLARE_ASN1_FUNCTIONS_const(CMS_ReceiptRequest)
+DECLARE_ASN1_FUNCTIONS(CMS_ContentInfo)
+DECLARE_ASN1_FUNCTIONS(CMS_ReceiptRequest)
+DECLARE_ASN1_PRINT_FUNCTION(CMS_ContentInfo)
 
 #define CMS_SIGNERINFO_ISSUER_SERIAL	0
 #define CMS_SIGNERINFO_KEYIDENTIFIER	1
@@ -125,9 +126,13 @@ int CMS_set_detached(CMS_ContentInfo *cms, int detached);
 DECLARE_PEM_rw_const(CMS, CMS_ContentInfo)
 #endif
 
+int CMS_stream(unsigned char ***boundary, CMS_ContentInfo *cms);
 CMS_ContentInfo *d2i_CMS_bio(BIO *bp, CMS_ContentInfo **cms);
 int i2d_CMS_bio(BIO *bp, CMS_ContentInfo *cms);
 
+BIO *BIO_new_CMS(BIO *out, CMS_ContentInfo *cms);
+int i2d_CMS_bio_stream(BIO *out, CMS_ContentInfo *cms, BIO *in, int flags);
+int PEM_write_bio_CMS_stream(BIO *out, CMS_ContentInfo *cms, BIO *in, int flags);
 CMS_ContentInfo *SMIME_read_CMS(BIO *bio, BIO **bcont);
 int SMIME_write_CMS(BIO *bio, CMS_ContentInfo *cms, BIO *data, int flags);
 
@@ -180,6 +185,8 @@ int CMS_decrypt_set1_pkey(CMS_ContentInfo *cms, EVP_PKEY *pk, X509 *cert);
 int CMS_decrypt_set1_key(CMS_ContentInfo *cms, 
 				unsigned char *key, size_t keylen,
 				unsigned char *id, size_t idlen);
+int CMS_decrypt_set1_password(CMS_ContentInfo *cms, 
+				unsigned char *pass, ossl_ssize_t passlen);
 
 STACK_OF(CMS_RecipientInfo) *CMS_get0_RecipientInfos(CMS_ContentInfo *cms);
 int CMS_RecipientInfo_type(CMS_RecipientInfo *ri);
@@ -215,6 +222,16 @@ int CMS_RecipientInfo_set0_key(CMS_RecipientInfo *ri,
 int CMS_RecipientInfo_kekri_id_cmp(CMS_RecipientInfo *ri, 
 					const unsigned char *id, size_t idlen);
 
+int CMS_RecipientInfo_set0_password(CMS_RecipientInfo *ri, 
+					unsigned char *pass,
+					ossl_ssize_t passlen);
+
+CMS_RecipientInfo *CMS_add0_recipient_password(CMS_ContentInfo *cms,
+					int iter, int wrap_nid, int pbe_nid,
+					unsigned char *pass,
+					ossl_ssize_t passlen,
+					const EVP_CIPHER *kekciph);
+
 int CMS_RecipientInfo_decrypt(CMS_ContentInfo *cms, CMS_RecipientInfo *ri);
 	
 int CMS_uncompress(CMS_ContentInfo *cms, BIO *dcont, BIO *out,
@@ -231,6 +248,7 @@ STACK_OF(X509) *CMS_get1_certs(CMS_ContentInfo *cms);
 
 CMS_RevocationInfoChoice *CMS_add0_RevocationInfoChoice(CMS_ContentInfo *cms);
 int CMS_add0_crl(CMS_ContentInfo *cms, X509_CRL *crl);
+int CMS_add1_crl(CMS_ContentInfo *cms, X509_CRL *crl);
 STACK_OF(X509_CRL) *CMS_get1_crls(CMS_ContentInfo *cms);
 
 int CMS_SignedData_init(CMS_ContentInfo *cms);
@@ -325,6 +343,7 @@ void ERR_load_CMS_strings(void);
 #define CMS_F_CHECK_CONTENT				 99
 #define CMS_F_CMS_ADD0_CERT				 164
 #define CMS_F_CMS_ADD0_RECIPIENT_KEY			 100
+#define CMS_F_CMS_ADD0_RECIPIENT_PASSWORD		 165
 #define CMS_F_CMS_ADD1_RECEIPTREQUEST			 158
 #define CMS_F_CMS_ADD1_RECIPIENT_CERT			 101
 #define CMS_F_CMS_ADD1_SIGNER				 102
@@ -339,6 +358,7 @@ void ERR_load_CMS_strings(void);
 #define CMS_F_CMS_DATAINIT				 111
 #define CMS_F_CMS_DECRYPT				 112
 #define CMS_F_CMS_DECRYPT_SET1_KEY			 113
+#define CMS_F_CMS_DECRYPT_SET1_PASSWORD			 166
 #define CMS_F_CMS_DECRYPT_SET1_PKEY			 114
 #define CMS_F_CMS_DIGESTALGORITHM_FIND_CTX		 115
 #define CMS_F_CMS_DIGESTALGORITHM_INIT_BIO		 116
@@ -373,7 +393,9 @@ void ERR_load_CMS_strings(void);
 #define CMS_F_CMS_RECIPIENTINFO_KTRI_ENCRYPT		 141
 #define CMS_F_CMS_RECIPIENTINFO_KTRI_GET0_ALGS		 142
 #define CMS_F_CMS_RECIPIENTINFO_KTRI_GET0_SIGNER_ID	 143
+#define CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT		 167
 #define CMS_F_CMS_RECIPIENTINFO_SET0_KEY		 144
+#define CMS_F_CMS_RECIPIENTINFO_SET0_PASSWORD		 168
 #define CMS_F_CMS_RECIPIENTINFO_SET0_PKEY		 145
 #define CMS_F_CMS_SET1_SIGNERIDENTIFIER			 146
 #define CMS_F_CMS_SET_DETACHED				 147
@@ -414,6 +436,7 @@ void ERR_load_CMS_strings(void);
 #define CMS_R_ERROR_SETTING_KEY				 115
 #define CMS_R_ERROR_SETTING_RECIPIENTINFO		 116
 #define CMS_R_INVALID_ENCRYPTED_KEY_LENGTH		 117
+#define CMS_R_INVALID_KEY_ENCRYPTION_PARAMETER		 176
 #define CMS_R_INVALID_KEY_LENGTH			 118
 #define CMS_R_MD_BIO_INIT_ERROR				 119
 #define CMS_R_MESSAGEDIGEST_ATTRIBUTE_WRONG_LENGTH	 120
@@ -426,6 +449,7 @@ void ERR_load_CMS_strings(void);
 #define CMS_R_NOT_ENCRYPTED_DATA			 122
 #define CMS_R_NOT_KEK					 123
 #define CMS_R_NOT_KEY_TRANSPORT				 124
+#define CMS_R_NOT_PWRI					 177
 #define CMS_R_NOT_SUPPORTED_FOR_THIS_KEY_TYPE		 125
 #define CMS_R_NO_CIPHER					 126
 #define CMS_R_NO_CONTENT				 127
@@ -438,6 +462,7 @@ void ERR_load_CMS_strings(void);
 #define CMS_R_NO_MATCHING_RECIPIENT			 132
 #define CMS_R_NO_MATCHING_SIGNATURE			 166
 #define CMS_R_NO_MSGSIGDIGEST				 167
+#define CMS_R_NO_PASSWORD				 178
 #define CMS_R_NO_PRIVATE_KEY				 133
 #define CMS_R_NO_PUBLIC_KEY				 134
 #define CMS_R_NO_RECEIPT_REQUEST			 168
@@ -461,10 +486,12 @@ void ERR_load_CMS_strings(void);
 #define CMS_R_UNSUPPORTED_COMPRESSION_ALGORITHM		 151
 #define CMS_R_UNSUPPORTED_CONTENT_TYPE			 152
 #define CMS_R_UNSUPPORTED_KEK_ALGORITHM			 153
+#define CMS_R_UNSUPPORTED_KEY_ENCRYPTION_ALGORITHM	 179
 #define CMS_R_UNSUPPORTED_RECIPIENT_TYPE		 154
 #define CMS_R_UNSUPPORTED_RECPIENTINFO_TYPE		 155
 #define CMS_R_UNSUPPORTED_TYPE				 156
 #define CMS_R_UNWRAP_ERROR				 157
+#define CMS_R_UNWRAP_FAILURE				 180
 #define CMS_R_VERIFICATION_FAILURE			 158
 #define CMS_R_WRAP_ERROR				 159
 
diff --git a/crypto/cms/cms_asn1.c b/crypto/cms/cms_asn1.c
index b253d54b5701..cfe67fb6c183 100644
--- a/crypto/cms/cms_asn1.c
+++ b/crypto/cms/cms_asn1.c
@@ -87,7 +87,8 @@ ASN1_NDEF_SEQUENCE(CMS_EncapsulatedContentInfo) = {
 } ASN1_NDEF_SEQUENCE_END(CMS_EncapsulatedContentInfo)
 
 /* Minor tweak to operation: free up signer key, cert */
-static int cms_si_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int cms_si_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+							void *exarg)
 	{
 	if(operation == ASN1_OP_FREE_POST)
 		{
@@ -213,7 +214,8 @@ ASN1_SEQUENCE(CMS_OtherRecipientInfo) = {
 } ASN1_SEQUENCE_END(CMS_OtherRecipientInfo)
 
 /* Free up RecipientInfo additional data */
-static int cms_ri_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int cms_ri_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+							void *exarg)
 	{
 	if(operation == ASN1_OP_FREE_PRE)
 		{
@@ -235,6 +237,15 @@ static int cms_ri_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
 				OPENSSL_free(kekri->key);
 				}
 			}
+		else if (ri->type == CMS_RECIPINFO_PASS)
+			{
+			CMS_PasswordRecipientInfo *pwri = ri->d.pwri;
+			if (pwri->pass)
+				{
+				OPENSSL_cleanse(pwri->pass, pwri->passlen);
+				OPENSSL_free(pwri->pass);
+				}
+			}
 		}
 	return 1;
 	}
@@ -300,10 +311,42 @@ ASN1_ADB(CMS_ContentInfo) = {
 	ADB_ENTRY(NID_id_smime_ct_compressedData, ASN1_NDEF_EXP(CMS_ContentInfo, d.compressedData, CMS_CompressedData, 0)),
 } ASN1_ADB_END(CMS_ContentInfo, 0, contentType, 0, &cms_default_tt, NULL);
 
-ASN1_NDEF_SEQUENCE(CMS_ContentInfo) = {
+/* CMS streaming support */
+static int cms_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+							void *exarg)
+	{
+	ASN1_STREAM_ARG *sarg = exarg;
+	CMS_ContentInfo *cms = NULL;
+	if (pval)
+		cms = (CMS_ContentInfo *)*pval;
+	else
+		return 1;
+	switch(operation)
+		{
+
+		case ASN1_OP_STREAM_PRE:
+		if (CMS_stream(&sarg->boundary, cms) <= 0)
+			return 0;
+		case ASN1_OP_DETACHED_PRE:
+		sarg->ndef_bio = CMS_dataInit(cms, sarg->out);
+		if (!sarg->ndef_bio)
+			return 0;
+		break;
+
+		case ASN1_OP_STREAM_POST:
+		case ASN1_OP_DETACHED_POST:
+		if (CMS_dataFinal(cms, sarg->ndef_bio) <= 0)
+			return 0;
+		break;
+
+		}
+	return 1;
+	}
+
+ASN1_NDEF_SEQUENCE_cb(CMS_ContentInfo, cms_cb) = {
 	ASN1_SIMPLE(CMS_ContentInfo, contentType, ASN1_OBJECT),
 	ASN1_ADB_OBJECT(CMS_ContentInfo)
-} ASN1_NDEF_SEQUENCE_END(CMS_ContentInfo)
+} ASN1_NDEF_SEQUENCE_END_cb(CMS_ContentInfo, CMS_ContentInfo)
 
 /* Specials for signed attributes */
 
diff --git a/crypto/cms/cms_env.c b/crypto/cms/cms_env.c
index b8685fa17590..be20b1c024cf 100644
--- a/crypto/cms/cms_env.c
+++ b/crypto/cms/cms_env.c
@@ -60,18 +60,18 @@
 #include <openssl/rand.h>
 #include <openssl/aes.h>
 #include "cms_lcl.h"
+#include "asn1_locl.h"
 
 /* CMS EnvelopedData Utilities */
 
 DECLARE_ASN1_ITEM(CMS_EnvelopedData)
-DECLARE_ASN1_ITEM(CMS_RecipientInfo)
 DECLARE_ASN1_ITEM(CMS_KeyTransRecipientInfo)
 DECLARE_ASN1_ITEM(CMS_KEKRecipientInfo)
 DECLARE_ASN1_ITEM(CMS_OtherKeyAttribute)
 
 DECLARE_STACK_OF(CMS_RecipientInfo)
 
-static CMS_EnvelopedData *cms_get0_enveloped(CMS_ContentInfo *cms)
+CMS_EnvelopedData *cms_get0_enveloped(CMS_ContentInfo *cms)
 	{
 	if (OBJ_obj2nid(cms->contentType) != NID_pkcs7_enveloped)
 		{
@@ -151,7 +151,7 @@ CMS_RecipientInfo *CMS_add1_recipient_cert(CMS_ContentInfo *cms,
 	CMS_KeyTransRecipientInfo *ktri;
 	CMS_EnvelopedData *env;
 	EVP_PKEY *pk = NULL;
-	int type;
+	int i, type;
 	env = cms_get0_enveloped(cms);
 	if (!env)
 		goto err;
@@ -200,21 +200,22 @@ CMS_RecipientInfo *CMS_add1_recipient_cert(CMS_ContentInfo *cms,
 	if (!cms_set1_SignerIdentifier(ktri->rid, recip, type))
 		goto err;
 
-	/* Since we have no EVP_PKEY_ASN1_METHOD in OpenSSL 0.9.8,
-	 * hard code algorithm parameters.
-	 */
-
-	if (pk->type == EVP_PKEY_RSA)
-		{
-		X509_ALGOR_set0(ktri->keyEncryptionAlgorithm,
-					OBJ_nid2obj(NID_rsaEncryption), 
-					V_ASN1_NULL, 0);
-		}
-	else
+	if (pk->ameth && pk->ameth->pkey_ctrl)
 		{
-		CMSerr(CMS_F_CMS_ADD1_RECIPIENT_CERT,
+		i = pk->ameth->pkey_ctrl(pk, ASN1_PKEY_CTRL_CMS_ENVELOPE,
+						0, ri);
+		if (i == -2)
+			{
+			CMSerr(CMS_F_CMS_ADD1_RECIPIENT_CERT,
 				CMS_R_NOT_SUPPORTED_FOR_THIS_KEY_TYPE);
-		goto err;
+			goto err;
+			}
+		if (i <= 0)
+			{
+			CMSerr(CMS_F_CMS_ADD1_RECIPIENT_CERT,
+				CMS_R_CTRL_FAILURE);
+			goto err;
+			}
 		}
 
 	if (!sk_CMS_RecipientInfo_push(env->recipientInfos, ri))
@@ -301,8 +302,9 @@ static int cms_RecipientInfo_ktri_encrypt(CMS_ContentInfo *cms,
 	{
 	CMS_KeyTransRecipientInfo *ktri;
 	CMS_EncryptedContentInfo *ec;
+	EVP_PKEY_CTX *pctx = NULL;
 	unsigned char *ek = NULL;
-	int eklen;
+	size_t eklen;
 
 	int ret = 0;
 
@@ -315,7 +317,22 @@ static int cms_RecipientInfo_ktri_encrypt(CMS_ContentInfo *cms,
 	ktri = ri->d.ktri;
 	ec = cms->d.envelopedData->encryptedContentInfo;
 
-	eklen = EVP_PKEY_size(ktri->pkey);
+	pctx = EVP_PKEY_CTX_new(ktri->pkey, NULL);
+	if (!pctx)
+		return 0;
+
+	if (EVP_PKEY_encrypt_init(pctx) <= 0)
+		goto err;
+
+	if (EVP_PKEY_CTX_ctrl(pctx, -1, EVP_PKEY_OP_ENCRYPT,
+				EVP_PKEY_CTRL_CMS_ENCRYPT, 0, ri) <= 0)
+		{
+		CMSerr(CMS_F_CMS_RECIPIENTINFO_KTRI_ENCRYPT, CMS_R_CTRL_ERROR);
+		goto err;
+		}
+
+	if (EVP_PKEY_encrypt(pctx, NULL, &eklen, ec->key, ec->keylen) <= 0)
+		goto err;
 
 	ek = OPENSSL_malloc(eklen);
 
@@ -326,9 +343,7 @@ static int cms_RecipientInfo_ktri_encrypt(CMS_ContentInfo *cms,
 		goto err;
 		}
 
-	eklen = EVP_PKEY_encrypt(ek, ec->key, ec->keylen, ktri->pkey);
-
-	if (eklen <= 0)
+	if (EVP_PKEY_encrypt(pctx, ek, &eklen, ec->key, ec->keylen) <= 0)
 		goto err;
 
 	ASN1_STRING_set0(ktri->encryptedKey, ek, eklen);
@@ -337,6 +352,8 @@ static int cms_RecipientInfo_ktri_encrypt(CMS_ContentInfo *cms,
 	ret = 1;
 
 	err:
+	if (pctx)
+		EVP_PKEY_CTX_free(pctx);
 	if (ek)
 		OPENSSL_free(ek);
 	return ret;
@@ -349,8 +366,9 @@ static int cms_RecipientInfo_ktri_decrypt(CMS_ContentInfo *cms,
 							CMS_RecipientInfo *ri)
 	{
 	CMS_KeyTransRecipientInfo *ktri = ri->d.ktri;
+	EVP_PKEY_CTX *pctx = NULL;
 	unsigned char *ek = NULL;
-	int eklen;
+	size_t eklen;
 	int ret = 0;
 	CMS_EncryptedContentInfo *ec;
 	ec = cms->d.envelopedData->encryptedContentInfo;
@@ -362,7 +380,24 @@ static int cms_RecipientInfo_ktri_decrypt(CMS_ContentInfo *cms,
 		return 0;
 		}
 
-	eklen = EVP_PKEY_size(ktri->pkey);
+	pctx = EVP_PKEY_CTX_new(ktri->pkey, NULL);
+	if (!pctx)
+		return 0;
+
+	if (EVP_PKEY_decrypt_init(pctx) <= 0)
+		goto err;
+
+	if (EVP_PKEY_CTX_ctrl(pctx, -1, EVP_PKEY_OP_DECRYPT,
+				EVP_PKEY_CTRL_CMS_DECRYPT, 0, ri) <= 0)
+		{
+		CMSerr(CMS_F_CMS_RECIPIENTINFO_KTRI_DECRYPT, CMS_R_CTRL_ERROR);
+		goto err;
+		}
+
+	if (EVP_PKEY_decrypt(pctx, NULL, &eklen,
+				ktri->encryptedKey->data,
+				ktri->encryptedKey->length) <= 0)
+		goto err;
 
 	ek = OPENSSL_malloc(eklen);
 
@@ -373,10 +408,9 @@ static int cms_RecipientInfo_ktri_decrypt(CMS_ContentInfo *cms,
 		goto err;
 		}
 
-	eklen = EVP_PKEY_decrypt(ek, 
+	if (EVP_PKEY_decrypt(pctx, ek, &eklen,
 				ktri->encryptedKey->data,
-				ktri->encryptedKey->length, ktri->pkey);
-	if (eklen <= 0)
+				ktri->encryptedKey->length) <= 0)
 		{
 		CMSerr(CMS_F_CMS_RECIPIENTINFO_KTRI_DECRYPT, CMS_R_CMS_LIB);
 		goto err;
@@ -394,6 +428,8 @@ static int cms_RecipientInfo_ktri_decrypt(CMS_ContentInfo *cms,
 	ec->keylen = eklen;
 
 	err:
+	if (pctx)
+		EVP_PKEY_CTX_free(pctx);
 	if (!ret && ek)
 		OPENSSL_free(ek);
 
@@ -757,6 +793,9 @@ int CMS_RecipientInfo_decrypt(CMS_ContentInfo *cms, CMS_RecipientInfo *ri)
 		case CMS_RECIPINFO_KEK:
 		return cms_RecipientInfo_kekri_decrypt(cms, ri);
 
+		case CMS_RECIPINFO_PASS:
+		return cms_RecipientInfo_pwri_crypt(cms, ri, 0);
+
 		default:
 		CMSerr(CMS_F_CMS_RECIPIENTINFO_DECRYPT,
 			CMS_R_UNSUPPORTED_RECPIENTINFO_TYPE);
@@ -800,6 +839,10 @@ BIO *cms_EnvelopedData_init_bio(CMS_ContentInfo *cms)
 			r = cms_RecipientInfo_kekri_encrypt(cms, ri);
 			break;
 
+			case CMS_RECIPINFO_PASS:
+			r = cms_RecipientInfo_pwri_crypt(cms, ri, 1);
+			break;
+
 			default:
 			CMSerr(CMS_F_CMS_ENVELOPEDDATA_INIT_BIO,
 				CMS_R_UNSUPPORTED_RECIPIENT_TYPE);
diff --git a/crypto/cms/cms_err.c b/crypto/cms/cms_err.c
index 52fa53954f32..8330ead7eda3 100644
--- a/crypto/cms/cms_err.c
+++ b/crypto/cms/cms_err.c
@@ -1,6 +1,6 @@
 /* crypto/cms/cms_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2008 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2009 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -73,6 +73,7 @@ static ERR_STRING_DATA CMS_str_functs[]=
 {ERR_FUNC(CMS_F_CHECK_CONTENT),	"CHECK_CONTENT"},
 {ERR_FUNC(CMS_F_CMS_ADD0_CERT),	"CMS_add0_cert"},
 {ERR_FUNC(CMS_F_CMS_ADD0_RECIPIENT_KEY),	"CMS_add0_recipient_key"},
+{ERR_FUNC(CMS_F_CMS_ADD0_RECIPIENT_PASSWORD),	"CMS_add0_recipient_password"},
 {ERR_FUNC(CMS_F_CMS_ADD1_RECEIPTREQUEST),	"CMS_add1_ReceiptRequest"},
 {ERR_FUNC(CMS_F_CMS_ADD1_RECIPIENT_CERT),	"CMS_add1_recipient_cert"},
 {ERR_FUNC(CMS_F_CMS_ADD1_SIGNER),	"CMS_add1_signer"},
@@ -87,6 +88,7 @@ static ERR_STRING_DATA CMS_str_functs[]=
 {ERR_FUNC(CMS_F_CMS_DATAINIT),	"CMS_dataInit"},
 {ERR_FUNC(CMS_F_CMS_DECRYPT),	"CMS_decrypt"},
 {ERR_FUNC(CMS_F_CMS_DECRYPT_SET1_KEY),	"CMS_decrypt_set1_key"},
+{ERR_FUNC(CMS_F_CMS_DECRYPT_SET1_PASSWORD),	"CMS_decrypt_set1_password"},
 {ERR_FUNC(CMS_F_CMS_DECRYPT_SET1_PKEY),	"CMS_decrypt_set1_pkey"},
 {ERR_FUNC(CMS_F_CMS_DIGESTALGORITHM_FIND_CTX),	"cms_DigestAlgorithm_find_ctx"},
 {ERR_FUNC(CMS_F_CMS_DIGESTALGORITHM_INIT_BIO),	"cms_DigestAlgorithm_init_bio"},
@@ -105,7 +107,7 @@ static ERR_STRING_DATA CMS_str_functs[]=
 {ERR_FUNC(CMS_F_CMS_GET0_CERTIFICATE_CHOICES),	"CMS_GET0_CERTIFICATE_CHOICES"},
 {ERR_FUNC(CMS_F_CMS_GET0_CONTENT),	"CMS_get0_content"},
 {ERR_FUNC(CMS_F_CMS_GET0_ECONTENT_TYPE),	"CMS_GET0_ECONTENT_TYPE"},
-{ERR_FUNC(CMS_F_CMS_GET0_ENVELOPED),	"CMS_GET0_ENVELOPED"},
+{ERR_FUNC(CMS_F_CMS_GET0_ENVELOPED),	"cms_get0_enveloped"},
 {ERR_FUNC(CMS_F_CMS_GET0_REVOCATION_CHOICES),	"CMS_GET0_REVOCATION_CHOICES"},
 {ERR_FUNC(CMS_F_CMS_GET0_SIGNED),	"CMS_GET0_SIGNED"},
 {ERR_FUNC(CMS_F_CMS_MSGSIGDIGEST_ADD1),	"cms_msgSigDigest_add1"},
@@ -121,7 +123,9 @@ static ERR_STRING_DATA CMS_str_functs[]=
 {ERR_FUNC(CMS_F_CMS_RECIPIENTINFO_KTRI_ENCRYPT),	"CMS_RECIPIENTINFO_KTRI_ENCRYPT"},
 {ERR_FUNC(CMS_F_CMS_RECIPIENTINFO_KTRI_GET0_ALGS),	"CMS_RecipientInfo_ktri_get0_algs"},
 {ERR_FUNC(CMS_F_CMS_RECIPIENTINFO_KTRI_GET0_SIGNER_ID),	"CMS_RecipientInfo_ktri_get0_signer_id"},
+{ERR_FUNC(CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT),	"cms_RecipientInfo_pwri_crypt"},
 {ERR_FUNC(CMS_F_CMS_RECIPIENTINFO_SET0_KEY),	"CMS_RecipientInfo_set0_key"},
+{ERR_FUNC(CMS_F_CMS_RECIPIENTINFO_SET0_PASSWORD),	"CMS_RecipientInfo_set0_password"},
 {ERR_FUNC(CMS_F_CMS_RECIPIENTINFO_SET0_PKEY),	"CMS_RecipientInfo_set0_pkey"},
 {ERR_FUNC(CMS_F_CMS_SET1_SIGNERIDENTIFIER),	"cms_set1_SignerIdentifier"},
 {ERR_FUNC(CMS_F_CMS_SET_DETACHED),	"CMS_set_detached"},
@@ -133,7 +137,7 @@ static ERR_STRING_DATA CMS_str_functs[]=
 {ERR_FUNC(CMS_F_CMS_SIGNERINFO_VERIFY_CERT),	"CMS_SIGNERINFO_VERIFY_CERT"},
 {ERR_FUNC(CMS_F_CMS_SIGNERINFO_VERIFY_CONTENT),	"CMS_SignerInfo_verify_content"},
 {ERR_FUNC(CMS_F_CMS_SIGN_RECEIPT),	"CMS_sign_receipt"},
-{ERR_FUNC(CMS_F_CMS_STREAM),	"CMS_STREAM"},
+{ERR_FUNC(CMS_F_CMS_STREAM),	"CMS_stream"},
 {ERR_FUNC(CMS_F_CMS_UNCOMPRESS),	"CMS_uncompress"},
 {ERR_FUNC(CMS_F_CMS_VERIFY),	"CMS_verify"},
 {0,NULL}
@@ -165,6 +169,7 @@ static ERR_STRING_DATA CMS_str_reasons[]=
 {ERR_REASON(CMS_R_ERROR_SETTING_KEY)     ,"error setting key"},
 {ERR_REASON(CMS_R_ERROR_SETTING_RECIPIENTINFO),"error setting recipientinfo"},
 {ERR_REASON(CMS_R_INVALID_ENCRYPTED_KEY_LENGTH),"invalid encrypted key length"},
+{ERR_REASON(CMS_R_INVALID_KEY_ENCRYPTION_PARAMETER),"invalid key encryption parameter"},
 {ERR_REASON(CMS_R_INVALID_KEY_LENGTH)    ,"invalid key length"},
 {ERR_REASON(CMS_R_MD_BIO_INIT_ERROR)     ,"md bio init error"},
 {ERR_REASON(CMS_R_MESSAGEDIGEST_ATTRIBUTE_WRONG_LENGTH),"messagedigest attribute wrong length"},
@@ -177,6 +182,7 @@ static ERR_STRING_DATA CMS_str_reasons[]=
 {ERR_REASON(CMS_R_NOT_ENCRYPTED_DATA)    ,"not encrypted data"},
 {ERR_REASON(CMS_R_NOT_KEK)               ,"not kek"},
 {ERR_REASON(CMS_R_NOT_KEY_TRANSPORT)     ,"not key transport"},
+{ERR_REASON(CMS_R_NOT_PWRI)              ,"not pwri"},
 {ERR_REASON(CMS_R_NOT_SUPPORTED_FOR_THIS_KEY_TYPE),"not supported for this key type"},
 {ERR_REASON(CMS_R_NO_CIPHER)             ,"no cipher"},
 {ERR_REASON(CMS_R_NO_CONTENT)            ,"no content"},
@@ -189,6 +195,7 @@ static ERR_STRING_DATA CMS_str_reasons[]=
 {ERR_REASON(CMS_R_NO_MATCHING_RECIPIENT) ,"no matching recipient"},
 {ERR_REASON(CMS_R_NO_MATCHING_SIGNATURE) ,"no matching signature"},
 {ERR_REASON(CMS_R_NO_MSGSIGDIGEST)       ,"no msgsigdigest"},
+{ERR_REASON(CMS_R_NO_PASSWORD)           ,"no password"},
 {ERR_REASON(CMS_R_NO_PRIVATE_KEY)        ,"no private key"},
 {ERR_REASON(CMS_R_NO_PUBLIC_KEY)         ,"no public key"},
 {ERR_REASON(CMS_R_NO_RECEIPT_REQUEST)    ,"no receipt request"},
@@ -212,10 +219,12 @@ static ERR_STRING_DATA CMS_str_reasons[]=
 {ERR_REASON(CMS_R_UNSUPPORTED_COMPRESSION_ALGORITHM),"unsupported compression algorithm"},
 {ERR_REASON(CMS_R_UNSUPPORTED_CONTENT_TYPE),"unsupported content type"},
 {ERR_REASON(CMS_R_UNSUPPORTED_KEK_ALGORITHM),"unsupported kek algorithm"},
+{ERR_REASON(CMS_R_UNSUPPORTED_KEY_ENCRYPTION_ALGORITHM),"unsupported key encryption algorithm"},
 {ERR_REASON(CMS_R_UNSUPPORTED_RECIPIENT_TYPE),"unsupported recipient type"},
 {ERR_REASON(CMS_R_UNSUPPORTED_RECPIENTINFO_TYPE),"unsupported recpientinfo type"},
 {ERR_REASON(CMS_R_UNSUPPORTED_TYPE)      ,"unsupported type"},
 {ERR_REASON(CMS_R_UNWRAP_ERROR)          ,"unwrap error"},
+{ERR_REASON(CMS_R_UNWRAP_FAILURE)        ,"unwrap failure"},
 {ERR_REASON(CMS_R_VERIFICATION_FAILURE)  ,"verification failure"},
 {ERR_REASON(CMS_R_WRAP_ERROR)            ,"wrap error"},
 {0,NULL}
diff --git a/crypto/cms/cms_ess.c b/crypto/cms/cms_ess.c
index 65613aab6529..90c0b82fb568 100644
--- a/crypto/cms/cms_ess.c
+++ b/crypto/cms/cms_ess.c
@@ -63,7 +63,7 @@
 DECLARE_ASN1_ITEM(CMS_ReceiptRequest)
 DECLARE_ASN1_ITEM(CMS_Receipt)
 
-IMPLEMENT_ASN1_FUNCTIONS_const(CMS_ReceiptRequest)
+IMPLEMENT_ASN1_FUNCTIONS(CMS_ReceiptRequest)
 
 /* ESS services: for now just Signed Receipt related */
 
diff --git a/crypto/cms/cms_io.c b/crypto/cms/cms_io.c
index 6d3edba5f728..1cb0264cc531 100644
--- a/crypto/cms/cms_io.c
+++ b/crypto/cms/cms_io.c
@@ -58,6 +58,25 @@
 #include "cms.h"
 #include "cms_lcl.h"
 
+int CMS_stream(unsigned char ***boundary, CMS_ContentInfo *cms)
+	{
+	ASN1_OCTET_STRING **pos;
+	pos = CMS_get0_content(cms);
+	if (!pos)
+		return 0;
+	if (!*pos)
+		*pos = ASN1_OCTET_STRING_new();
+	if (*pos)
+		{
+		(*pos)->flags |= ASN1_STRING_FLAG_NDEF;
+		(*pos)->flags &= ~ASN1_STRING_FLAG_CONT;
+		*boundary = &(*pos)->data;
+		return 1;
+		}
+	CMSerr(CMS_F_CMS_STREAM, ERR_R_MALLOC_FAILURE);
+	return 0;
+	}
+
 CMS_ContentInfo *d2i_CMS_bio(BIO *bp, CMS_ContentInfo **cms)
 	{
 	return ASN1_item_d2i_bio(ASN1_ITEM_rptr(CMS_ContentInfo), bp, cms);
@@ -70,52 +89,26 @@ int i2d_CMS_bio(BIO *bp, CMS_ContentInfo *cms)
 
 IMPLEMENT_PEM_rw_const(CMS, CMS_ContentInfo, PEM_STRING_CMS, CMS_ContentInfo)
 
-/* Callback for int_smime_write_ASN1 */
-
-static int cms_output_data(BIO *out, BIO *data, ASN1_VALUE *val, int flags,
-					const ASN1_ITEM *it)
+BIO *BIO_new_CMS(BIO *out, CMS_ContentInfo *cms) 
 	{
-	CMS_ContentInfo *cms = (CMS_ContentInfo *)val;
-	BIO *tmpbio, *cmsbio;
-	int r = 0;
-
-	if (!(flags & SMIME_DETACHED))
-		{
-		SMIME_crlf_copy(data, out, flags);
-		return 1;
-		}
-
-	/* Let CMS code prepend any needed BIOs */
-
-	cmsbio = CMS_dataInit(cms, out);
-
-	if (!cmsbio)
-		return 0;
-
-	/* Copy data across, passing through filter BIOs for processing */
-	SMIME_crlf_copy(data, cmsbio, flags);
-
-	/* Finalize structure */
-	if (CMS_dataFinal(cms, cmsbio) <= 0)
-		goto err;
-
-	r = 1;
-
-	err:
-
-	/* Now remove any digests prepended to the BIO */
-
-	while (cmsbio != out)
-		{
-		tmpbio = BIO_pop(cmsbio);
-		BIO_free(cmsbio);
-		cmsbio = tmpbio;
-		}
+	return BIO_new_NDEF(out, (ASN1_VALUE *)cms,
+				ASN1_ITEM_rptr(CMS_ContentInfo));
+	}
 
-	return r;
+/* CMS wrappers round generalised stream and MIME routines */
 
+int i2d_CMS_bio_stream(BIO *out, CMS_ContentInfo *cms, BIO *in, int flags)
+	{
+	return i2d_ASN1_bio_stream(out, (ASN1_VALUE *)cms, in, flags,
+					ASN1_ITEM_rptr(CMS_ContentInfo));
 	}
 
+int PEM_write_bio_CMS_stream(BIO *out, CMS_ContentInfo *cms, BIO *in, int flags)
+	{
+	return PEM_write_bio_ASN1_stream(out, (ASN1_VALUE *) cms, in, flags,
+					"CMS",
+					ASN1_ITEM_rptr(CMS_ContentInfo));
+	}
 
 int SMIME_write_CMS(BIO *bio, CMS_ContentInfo *cms, BIO *data, int flags)
 	{
@@ -127,9 +120,8 @@ int SMIME_write_CMS(BIO *bio, CMS_ContentInfo *cms, BIO *data, int flags)
 	else
 		mdalgs = NULL;
 
-	return int_smime_write_ASN1(bio, (ASN1_VALUE *)cms, data, flags,
+	return SMIME_write_ASN1(bio, (ASN1_VALUE *)cms, data, flags,
 					ctype_nid, econt_nid, mdalgs,
-					cms_output_data,
 					ASN1_ITEM_rptr(CMS_ContentInfo));	
 	}
 
@@ -138,3 +130,4 @@ CMS_ContentInfo *SMIME_read_CMS(BIO *bio, BIO **bcont)
 	return (CMS_ContentInfo *)SMIME_read_ASN1(bio, bcont,
 					ASN1_ITEM_rptr(CMS_ContentInfo));
 	}
+
diff --git a/crypto/cms/cms_lcl.h b/crypto/cms/cms_lcl.h
index ce65d6ef665c..a9f9730157b3 100644
--- a/crypto/cms/cms_lcl.h
+++ b/crypto/cms/cms_lcl.h
@@ -275,6 +275,9 @@ struct CMS_PasswordRecipientInfo_st
  	X509_ALGOR *keyDerivationAlgorithm;
  	X509_ALGOR *keyEncryptionAlgorithm;
  	ASN1_OCTET_STRING *encryptedKey;
+	/* Extra info: password to use */
+	unsigned char *pass;
+	size_t passlen;
 	};
 
 struct CMS_OtherRecipientInfo_st
@@ -408,10 +411,13 @@ struct CMS_Receipt_st
 	ASN1_OCTET_STRING *originatorSignatureValue;
 	};
 
+DECLARE_ASN1_FUNCTIONS(CMS_ContentInfo)
 DECLARE_ASN1_ITEM(CMS_SignerInfo)
 DECLARE_ASN1_ITEM(CMS_IssuerAndSerialNumber)
 DECLARE_ASN1_ITEM(CMS_Attributes_Sign)
 DECLARE_ASN1_ITEM(CMS_Attributes_Verify)
+DECLARE_ASN1_ITEM(CMS_RecipientInfo)
+DECLARE_ASN1_ITEM(CMS_PasswordRecipientInfo)
 DECLARE_ASN1_ALLOC_FUNCTIONS(CMS_IssuerAndSerialNumber)
 
 #define CMS_SIGNERINFO_ISSUER_SERIAL	0
@@ -455,6 +461,11 @@ int cms_msgSigDigest_add1(CMS_SignerInfo *dest, CMS_SignerInfo *src);
 ASN1_OCTET_STRING *cms_encode_Receipt(CMS_SignerInfo *si);
 
 BIO *cms_EnvelopedData_init_bio(CMS_ContentInfo *cms);
+CMS_EnvelopedData *cms_get0_enveloped(CMS_ContentInfo *cms);
+
+/* PWRI routines */
+int cms_RecipientInfo_pwri_crypt(CMS_ContentInfo *cms, CMS_RecipientInfo *ri,
+							int en_de);
 	
 #ifdef  __cplusplus
 }
diff --git a/crypto/cms/cms_lib.c b/crypto/cms/cms_lib.c
index cc00526d3e07..f88e8f3b5257 100644
--- a/crypto/cms/cms_lib.c
+++ b/crypto/cms/cms_lib.c
@@ -60,7 +60,8 @@
 #include "cms.h"
 #include "cms_lcl.h"
 
-IMPLEMENT_ASN1_FUNCTIONS_const(CMS_ContentInfo)
+IMPLEMENT_ASN1_FUNCTIONS(CMS_ContentInfo)
+IMPLEMENT_ASN1_PRINT_FUNCTION(CMS_ContentInfo)
 
 DECLARE_ASN1_ITEM(CMS_CertificateChoices)
 DECLARE_ASN1_ITEM(CMS_RevocationInfoChoice)
@@ -346,20 +347,10 @@ void cms_DigestAlgorithm_set(X509_ALGOR *alg, const EVP_MD *md)
 	{
 	int param_type;
 
-	switch (EVP_MD_type(md))
-		{
-		case NID_sha1:
-		case NID_sha224:
-		case NID_sha256:
-		case NID_sha384:
-		case NID_sha512:
+	if (md->flags & EVP_MD_FLAG_DIGALGID_ABSENT)
 		param_type = V_ASN1_UNDEF;
-		break;
-	
-		default:
+	else
 		param_type = V_ASN1_NULL;
-		break;
-		}
 
 	X509_ALGOR_set0(alg, OBJ_nid2obj(EVP_MD_type(md)), param_type, NULL);
 
@@ -421,8 +412,7 @@ int cms_DigestAlgorithm_find_ctx(EVP_MD_CTX *mctx, BIO *chain,
 		 */
 			|| EVP_MD_pkey_type(EVP_MD_CTX_md(mtmp)) == nid)
 			{
-			EVP_MD_CTX_copy_ex(mctx, mtmp);
-			return 1;
+			return EVP_MD_CTX_copy_ex(mctx, mtmp);
 			}
 		chain = BIO_next(chain);
 		}
@@ -561,6 +551,15 @@ int CMS_add0_crl(CMS_ContentInfo *cms, X509_CRL *crl)
 	return 1;
 	}
 
+int CMS_add1_crl(CMS_ContentInfo *cms, X509_CRL *crl)
+	{
+	int r;
+	r = CMS_add0_crl(cms, crl);
+	if (r > 0)
+		CRYPTO_add(&crl->references, 1, CRYPTO_LOCK_X509_CRL);
+	return r;
+	}
+
 STACK_OF(X509) *CMS_get1_certs(CMS_ContentInfo *cms)
 	{
 	STACK_OF(X509) *certs = NULL;
diff --git a/crypto/cms/cms_pwri.c b/crypto/cms/cms_pwri.c
new file mode 100644
index 000000000000..b79612a12df4
--- /dev/null
+++ b/crypto/cms/cms_pwri.c
@@ -0,0 +1,454 @@
+/* crypto/cms/cms_pwri.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project.
+ */
+/* ====================================================================
+ * Copyright (c) 2009 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#include "cryptlib.h"
+#include <openssl/asn1t.h>
+#include <openssl/pem.h>
+#include <openssl/x509v3.h>
+#include <openssl/err.h>
+#include <openssl/cms.h>
+#include <openssl/rand.h>
+#include <openssl/aes.h>
+#include "cms_lcl.h"
+#include "asn1_locl.h"
+
+int CMS_RecipientInfo_set0_password(CMS_RecipientInfo *ri, 
+				unsigned char *pass, ossl_ssize_t passlen)
+	{
+	CMS_PasswordRecipientInfo *pwri;
+	if (ri->type != CMS_RECIPINFO_PASS)
+		{
+		CMSerr(CMS_F_CMS_RECIPIENTINFO_SET0_PASSWORD, CMS_R_NOT_PWRI);
+		return 0;
+		}
+
+	pwri = ri->d.pwri;
+	pwri->pass = pass;
+	if (pass && passlen < 0)
+		passlen = strlen((char *)pass);
+	pwri->passlen = passlen;
+	return 1;
+	}
+
+CMS_RecipientInfo *CMS_add0_recipient_password(CMS_ContentInfo *cms,
+					int iter, int wrap_nid, int pbe_nid,
+					unsigned char *pass,
+					ossl_ssize_t passlen,
+					const EVP_CIPHER *kekciph)
+	{
+	CMS_RecipientInfo *ri = NULL;
+	CMS_EnvelopedData *env;
+	CMS_PasswordRecipientInfo *pwri;
+	EVP_CIPHER_CTX ctx;
+	X509_ALGOR *encalg = NULL;
+	unsigned char iv[EVP_MAX_IV_LENGTH];
+	int ivlen;
+	env = cms_get0_enveloped(cms);
+	if (!env)
+		goto err;
+
+	if (wrap_nid <= 0)
+		wrap_nid = NID_id_alg_PWRI_KEK;
+
+	if (pbe_nid <= 0)
+		pbe_nid = NID_id_pbkdf2;
+
+	/* Get from enveloped data */
+	if (kekciph == NULL)
+		kekciph = env->encryptedContentInfo->cipher;
+
+	if (kekciph == NULL)
+		{
+		CMSerr(CMS_F_CMS_ADD0_RECIPIENT_PASSWORD, CMS_R_NO_CIPHER);
+		return NULL;
+		}
+	if (wrap_nid != NID_id_alg_PWRI_KEK)
+		{
+		CMSerr(CMS_F_CMS_ADD0_RECIPIENT_PASSWORD,
+				CMS_R_UNSUPPORTED_KEY_ENCRYPTION_ALGORITHM);
+		return NULL;
+		}
+
+	/* Setup algorithm identifier for cipher */
+	encalg = X509_ALGOR_new();
+	EVP_CIPHER_CTX_init(&ctx);
+
+	if (EVP_EncryptInit_ex(&ctx, kekciph, NULL, NULL, NULL) <= 0)
+		{
+		CMSerr(CMS_F_CMS_ADD0_RECIPIENT_PASSWORD, ERR_R_EVP_LIB);
+		goto err;
+		}
+
+	ivlen = EVP_CIPHER_CTX_iv_length(&ctx);
+
+	if (ivlen > 0)
+		{
+		if (RAND_pseudo_bytes(iv, ivlen) <= 0)
+			goto err;
+		if (EVP_EncryptInit_ex(&ctx, NULL, NULL, NULL, iv) <= 0)
+			{
+			CMSerr(CMS_F_CMS_ADD0_RECIPIENT_PASSWORD,
+							ERR_R_EVP_LIB);
+			goto err;
+			}
+		encalg->parameter = ASN1_TYPE_new();
+		if (!encalg->parameter)
+			{
+			CMSerr(CMS_F_CMS_ADD0_RECIPIENT_PASSWORD,
+							ERR_R_MALLOC_FAILURE);
+			goto err;
+			}
+		if (EVP_CIPHER_param_to_asn1(&ctx, encalg->parameter) <= 0)
+			{
+			CMSerr(CMS_F_CMS_ADD0_RECIPIENT_PASSWORD,
+				CMS_R_CIPHER_PARAMETER_INITIALISATION_ERROR);
+			goto err;
+			}
+		}
+
+
+	encalg->algorithm = OBJ_nid2obj(EVP_CIPHER_CTX_type(&ctx));
+
+	EVP_CIPHER_CTX_cleanup(&ctx);
+
+	/* Initialize recipient info */
+	ri = M_ASN1_new_of(CMS_RecipientInfo);
+	if (!ri)
+		goto merr;
+
+	ri->d.pwri = M_ASN1_new_of(CMS_PasswordRecipientInfo);
+	if (!ri->d.pwri)
+		goto merr;
+	ri->type = CMS_RECIPINFO_PASS;
+
+	pwri = ri->d.pwri;
+	/* Since this is overwritten, free up empty structure already there */
+	X509_ALGOR_free(pwri->keyEncryptionAlgorithm);
+	pwri->keyEncryptionAlgorithm = X509_ALGOR_new();
+	if (!pwri->keyEncryptionAlgorithm)
+		goto merr;
+	pwri->keyEncryptionAlgorithm->algorithm = OBJ_nid2obj(wrap_nid);
+	pwri->keyEncryptionAlgorithm->parameter = ASN1_TYPE_new();
+	if (!pwri->keyEncryptionAlgorithm->parameter)
+		goto merr;
+
+        if(!ASN1_item_pack(encalg, ASN1_ITEM_rptr(X509_ALGOR),
+	    &pwri->keyEncryptionAlgorithm->parameter->value.sequence))
+		goto merr;
+        pwri->keyEncryptionAlgorithm->parameter->type = V_ASN1_SEQUENCE;
+
+	X509_ALGOR_free(encalg);
+	encalg = NULL;
+
+	/* Setup PBE algorithm */
+
+	pwri->keyDerivationAlgorithm = PKCS5_pbkdf2_set(iter, NULL, 0, -1, -1);
+
+	if (!pwri->keyDerivationAlgorithm)
+		goto err;
+
+	CMS_RecipientInfo_set0_password(ri, pass, passlen);
+	pwri->version = 0;
+
+	if (!sk_CMS_RecipientInfo_push(env->recipientInfos, ri))
+		goto merr;
+
+	return ri;
+
+	merr:
+	CMSerr(CMS_F_CMS_ADD0_RECIPIENT_PASSWORD, ERR_R_MALLOC_FAILURE);
+	err:
+	EVP_CIPHER_CTX_cleanup(&ctx);
+	if (ri)
+		M_ASN1_free_of(ri, CMS_RecipientInfo);
+	if (encalg)
+		X509_ALGOR_free(encalg);
+	return NULL;
+
+	}
+
+/* This is an implementation of the key wrapping mechanism in RFC3211,
+ * at some point this should go into EVP.
+ */
+
+static int kek_unwrap_key(unsigned char *out, size_t *outlen,
+		const unsigned char *in, size_t inlen, EVP_CIPHER_CTX *ctx)
+	{
+	size_t blocklen = EVP_CIPHER_CTX_block_size(ctx);
+	unsigned char *tmp;
+	int outl, rv = 0;
+	if (inlen < 2 * blocklen)
+		{
+		/* too small */
+		return 0;
+		}
+	if (inlen % blocklen)
+		{
+		/* Invalid size */
+		return 0;
+		}
+	tmp = OPENSSL_malloc(inlen);
+	/* setup IV by decrypting last two blocks */
+	EVP_DecryptUpdate(ctx, tmp + inlen - 2 * blocklen, &outl,
+				in  + inlen - 2 * blocklen, blocklen * 2);
+	/* Do a decrypt of last decrypted block to set IV to correct value
+	 * output it to start of buffer so we don't corrupt decrypted block
+	 * this works because buffer is at least two block lengths long.
+	 */
+	EVP_DecryptUpdate(ctx, tmp, &outl,
+				tmp  + inlen - blocklen, blocklen);
+	/* Can now decrypt first n - 1 blocks */
+	EVP_DecryptUpdate(ctx, tmp, &outl, in, inlen - blocklen);
+
+	/* Reset IV to original value */
+	EVP_DecryptInit_ex(ctx, NULL, NULL, NULL, NULL);
+	/* Decrypt again */
+	EVP_DecryptUpdate(ctx, tmp, &outl, tmp, inlen);
+	/* Check check bytes */
+	if (((tmp[1] ^ tmp[4]) & (tmp[2] ^ tmp[5]) & (tmp[3] ^ tmp[6])) != 0xff)
+		{
+		/* Check byte failure */
+		goto err;
+		}
+	if (inlen < (size_t)(tmp[0] - 4 ))
+		{
+		/* Invalid length value */
+		goto err;
+		}
+	*outlen = (size_t)tmp[0];
+	memcpy(out, tmp + 4, *outlen);
+	rv = 1;
+	err:
+	OPENSSL_cleanse(tmp, inlen);
+	OPENSSL_free(tmp);
+	return rv;
+
+	}
+
+static int kek_wrap_key(unsigned char *out, size_t *outlen,
+		const unsigned char *in, size_t inlen, EVP_CIPHER_CTX *ctx)
+	{
+	size_t blocklen = EVP_CIPHER_CTX_block_size(ctx);
+	size_t olen;
+	int dummy;
+	/* First decide length of output buffer: need header and round up to
+	 * multiple of block length.
+	 */
+	olen = (inlen + 4 + blocklen - 1)/blocklen;
+	olen *= blocklen;
+	if (olen < 2 * blocklen)
+		{
+		/* Key too small */
+		return 0;
+		}
+	if (inlen > 0xFF)
+		{
+		/* Key too large */
+		return 0;
+		}
+	if (out)
+		{
+		/* Set header */
+		out[0] = (unsigned char)inlen;
+		out[1] = in[0] ^ 0xFF;
+		out[2] = in[1] ^ 0xFF;
+		out[3] = in[2] ^ 0xFF;
+		memcpy(out + 4, in, inlen);
+		/* Add random padding to end */
+		if (olen > inlen + 4)
+			RAND_pseudo_bytes(out + 4 + inlen, olen - 4 - inlen);
+		/* Encrypt twice */
+		EVP_EncryptUpdate(ctx, out, &dummy, out, olen);
+		EVP_EncryptUpdate(ctx, out, &dummy, out, olen);
+		}
+
+	*outlen = olen;
+
+	return 1;
+	}
+
+/* Encrypt/Decrypt content key in PWRI recipient info */
+
+int cms_RecipientInfo_pwri_crypt(CMS_ContentInfo *cms, CMS_RecipientInfo *ri,
+							int en_de)
+	{
+	CMS_EncryptedContentInfo *ec;
+	CMS_PasswordRecipientInfo *pwri;
+	const unsigned char *p = NULL;
+	int plen;
+	int r = 0;
+	X509_ALGOR *algtmp, *kekalg = NULL;
+	EVP_CIPHER_CTX kekctx;
+	const EVP_CIPHER *kekcipher;
+	unsigned char *key = NULL;
+	size_t keylen;
+
+	ec = cms->d.envelopedData->encryptedContentInfo;
+
+	pwri = ri->d.pwri;
+	EVP_CIPHER_CTX_init(&kekctx);
+
+	if (!pwri->pass)
+		{
+		CMSerr(CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT, CMS_R_NO_PASSWORD);
+		return 0;
+		}
+	algtmp = pwri->keyEncryptionAlgorithm;
+
+	if (!algtmp || OBJ_obj2nid(algtmp->algorithm) != NID_id_alg_PWRI_KEK)
+		{
+		CMSerr(CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT,
+				CMS_R_UNSUPPORTED_KEY_ENCRYPTION_ALGORITHM);
+		return 0;
+		}
+
+	if (algtmp->parameter->type == V_ASN1_SEQUENCE)
+		{
+		p = algtmp->parameter->value.sequence->data;
+		plen = algtmp->parameter->value.sequence->length;
+		kekalg = d2i_X509_ALGOR(NULL, &p, plen);
+		}
+	if (kekalg == NULL)
+		{
+		CMSerr(CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT,
+				CMS_R_INVALID_KEY_ENCRYPTION_PARAMETER);
+		return 0;
+		}
+
+	kekcipher = EVP_get_cipherbyobj(kekalg->algorithm);
+		
+	if(!kekcipher)
+		{
+		CMSerr(CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT,
+				CMS_R_UNKNOWN_CIPHER);
+		goto err;
+		}
+
+	/* Fixup cipher based on AlgorithmIdentifier to set IV etc */
+	if (!EVP_CipherInit_ex(&kekctx, kekcipher, NULL, NULL, NULL, en_de))
+		goto err;
+	EVP_CIPHER_CTX_set_padding(&kekctx, 0);
+	if(EVP_CIPHER_asn1_to_param(&kekctx, kekalg->parameter) < 0)
+		{
+		CMSerr(CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT,
+				CMS_R_CIPHER_PARAMETER_INITIALISATION_ERROR);
+		goto err;
+		}
+
+	algtmp = pwri->keyDerivationAlgorithm;
+
+	/* Finish password based key derivation to setup key in "ctx" */
+
+	if (EVP_PBE_CipherInit(algtmp->algorithm,
+				(char *)pwri->pass, pwri->passlen,
+				algtmp->parameter, &kekctx, en_de) < 0)
+		{
+		CMSerr(CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT, ERR_R_EVP_LIB);
+		goto err;
+		}
+
+	/* Finally wrap/unwrap the key */
+
+	if (en_de)
+		{
+
+		if (!kek_wrap_key(NULL, &keylen, ec->key, ec->keylen, &kekctx))
+			goto err;
+
+		key = OPENSSL_malloc(keylen);
+
+		if (!key)
+			goto err;
+
+		if (!kek_wrap_key(key, &keylen, ec->key, ec->keylen, &kekctx))
+			goto err;
+		pwri->encryptedKey->data = key;
+		pwri->encryptedKey->length = keylen;
+		}
+	else
+		{
+		key = OPENSSL_malloc(pwri->encryptedKey->length);
+
+		if (!key)
+			{
+			CMSerr(CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT,
+							ERR_R_MALLOC_FAILURE);
+			goto err;
+			}
+		if (!kek_unwrap_key(key, &keylen,
+					pwri->encryptedKey->data,
+					pwri->encryptedKey->length, &kekctx))
+			{
+			CMSerr(CMS_F_CMS_RECIPIENTINFO_PWRI_CRYPT,
+							CMS_R_UNWRAP_FAILURE);
+			goto err;
+			}
+
+		ec->key = key;
+		ec->keylen = keylen;
+
+		}
+
+	r = 1;
+
+	err:
+
+	EVP_CIPHER_CTX_cleanup(&kekctx);
+
+	if (!r && key)
+		OPENSSL_free(key);
+	X509_ALGOR_free(kekalg);
+
+	return r;
+
+	}
diff --git a/crypto/cms/cms_sd.c b/crypto/cms/cms_sd.c
index cdac3b870dd9..77fbd1359679 100644
--- a/crypto/cms/cms_sd.c
+++ b/crypto/cms/cms_sd.c
@@ -58,6 +58,7 @@
 #include <openssl/err.h>
 #include <openssl/cms.h>
 #include "cms_lcl.h"
+#include "asn1_locl.h"
 
 /* CMS SignedData Utilities */
 
@@ -218,10 +219,9 @@ int cms_set1_SignerIdentifier(CMS_SignerIdentifier *sid, X509 *cert, int type)
 		if (!X509_NAME_set(&sid->d.issuerAndSerialNumber->issuer,
 					X509_get_issuer_name(cert)))
 			goto merr;
-		ASN1_STRING_free(sid->d.issuerAndSerialNumber->serialNumber);
-		sid->d.issuerAndSerialNumber->serialNumber =
-				ASN1_STRING_dup(X509_get_serialNumber(cert));
-		if(!sid->d.issuerAndSerialNumber->serialNumber)
+		if (!ASN1_STRING_copy(
+			sid->d.issuerAndSerialNumber->serialNumber,
+				X509_get_serialNumber(cert)))
 			goto merr;
 		break;
 
@@ -341,16 +341,22 @@ CMS_SignerInfo *CMS_add1_signer(CMS_ContentInfo *cms,
 	if (!cms_set1_SignerIdentifier(si->sid, signer, type))
 		goto err;
 
-	/* Since no EVP_PKEY_METHOD in 0.9.8 hard code SHA1 as default */
 	if (md == NULL)
-		md = EVP_sha1();
-
-	/* OpenSSL 0.9.8 only supports SHA1 with non-RSA keys */
+		{
+		int def_nid;
+		if (EVP_PKEY_get_default_digest_nid(pk, &def_nid) <= 0)
+			goto err;
+		md = EVP_get_digestbynid(def_nid);
+		if (md == NULL)
+			{
+			CMSerr(CMS_F_CMS_ADD1_SIGNER, CMS_R_NO_DEFAULT_DIGEST);
+			goto err;
+			}
+		}
 
-	if ((pk->type != EVP_PKEY_RSA) && (EVP_MD_type(md) != NID_sha1))
+	if (!md)
 		{
-		CMSerr(CMS_F_CMS_ADD1_SIGNER,
-				CMS_R_NOT_SUPPORTED_FOR_THIS_KEY_TYPE);
+		CMSerr(CMS_F_CMS_ADD1_SIGNER, CMS_R_NO_DIGEST_SET);
 		goto err;
 		}
 
@@ -379,37 +385,21 @@ CMS_SignerInfo *CMS_add1_signer(CMS_ContentInfo *cms,
 			}
 		}
 
-	/* Since we have no EVP_PKEY_ASN1_METHOD in OpenSSL 0.9.8,
-	 * hard code algorithm parameters.
-	 */
-
-	switch (pk->type)
+	if (pk->ameth && pk->ameth->pkey_ctrl)
 		{
-
-		case EVP_PKEY_RSA:
-		X509_ALGOR_set0(si->signatureAlgorithm,
-					OBJ_nid2obj(NID_rsaEncryption),
-					V_ASN1_NULL, 0);
-		break;
-
-		case EVP_PKEY_DSA:
-		X509_ALGOR_set0(si->signatureAlgorithm,
-					OBJ_nid2obj(NID_dsaWithSHA1),
-					V_ASN1_UNDEF, 0);
-		break;
-
-
-		case EVP_PKEY_EC:
-		X509_ALGOR_set0(si->signatureAlgorithm,
-					OBJ_nid2obj(NID_ecdsa_with_SHA1),
-					V_ASN1_UNDEF, 0);
-		break;
-
-		default:
-		CMSerr(CMS_F_CMS_ADD1_SIGNER,
+		i = pk->ameth->pkey_ctrl(pk, ASN1_PKEY_CTRL_CMS_SIGN,
+						0, si);
+		if (i == -2)
+			{
+			CMSerr(CMS_F_CMS_ADD1_SIGNER,
 				CMS_R_NOT_SUPPORTED_FOR_THIS_KEY_TYPE);
-		goto err;
-
+			goto err;
+			}
+		if (i <= 0)
+			{
+			CMSerr(CMS_F_CMS_ADD1_SIGNER, CMS_R_CTRL_FAILURE);
+			goto err;
+			}
 		}
 
 	if (!(flags & CMS_NOATTR))
@@ -626,25 +616,6 @@ void CMS_SignerInfo_get0_algs(CMS_SignerInfo *si, EVP_PKEY **pk, X509 **signer,
 		*psig = si->signatureAlgorithm;
 	}
 
-/* In OpenSSL 0.9.8 we have the link between digest types and public
- * key types so we need to fixup the digest type if the public key
- * type is not appropriate.
- */
-
-static void cms_fixup_mctx(EVP_MD_CTX *mctx, EVP_PKEY *pkey)
-	{
-	if (EVP_MD_CTX_type(mctx) != NID_sha1)
-		return;
-#ifndef OPENSSL_NO_DSA
-	if (pkey->type == EVP_PKEY_DSA)
-		mctx->digest = EVP_dss1();	
-#endif
-#ifndef OPENSSL_NO_ECDSA
-	if (pkey->type == EVP_PKEY_EC)
-		mctx->digest = EVP_ecdsa();	
-#endif
-	}
-
 static int cms_SignerInfo_content_sign(CMS_ContentInfo *cms,
 					CMS_SignerInfo *si, BIO *chain)
 	{
@@ -670,7 +641,8 @@ static int cms_SignerInfo_content_sign(CMS_ContentInfo *cms,
 			cms->d.signedData->encapContentInfo->eContentType; 
 		unsigned char md[EVP_MAX_MD_SIZE];
 		unsigned int mdlen;
-		EVP_DigestFinal_ex(&mctx, md, &mdlen);
+		if (!EVP_DigestFinal_ex(&mctx, md, &mdlen))
+			goto err;
 		if (!CMS_signed_add1_attr_by_NID(si, NID_pkcs9_messageDigest,
 						V_ASN1_OCTET_STRING,
 						md, mdlen))
@@ -693,7 +665,6 @@ static int cms_SignerInfo_content_sign(CMS_ContentInfo *cms,
 					ERR_R_MALLOC_FAILURE);
 			goto err;
 			}
-		cms_fixup_mctx(&mctx, si->pkey);
 		if (!EVP_SignFinal(&mctx, sig, &siglen, si->pkey))
 			{
 			CMSerr(CMS_F_CMS_SIGNERINFO_CONTENT_SIGN,
@@ -731,9 +702,10 @@ int cms_SignedData_final(CMS_ContentInfo *cms, BIO *chain)
 int CMS_SignerInfo_sign(CMS_SignerInfo *si)
 	{
 	EVP_MD_CTX mctx;
+	EVP_PKEY_CTX *pctx;
 	unsigned char *abuf = NULL;
 	int alen;
-	unsigned int siglen;
+	size_t siglen;
 	const EVP_MD *md = NULL;
 
 	md = EVP_get_digestbyobj(si->digestAlgorithm->algorithm);
@@ -748,40 +720,38 @@ int CMS_SignerInfo_sign(CMS_SignerInfo *si)
 			goto err;
 		}
 
-	if (EVP_SignInit_ex(&mctx, md, NULL) <= 0)
+	if (EVP_DigestSignInit(&mctx, &pctx, md, NULL, si->pkey) <= 0)
 		goto err;
 
-#if 0
 	if (EVP_PKEY_CTX_ctrl(pctx, -1, EVP_PKEY_OP_SIGN,
 				EVP_PKEY_CTRL_CMS_SIGN, 0, si) <= 0)
 		{
 		CMSerr(CMS_F_CMS_SIGNERINFO_SIGN, CMS_R_CTRL_ERROR);
 		goto err;
 		}
-#endif
 
 	alen = ASN1_item_i2d((ASN1_VALUE *)si->signedAttrs,&abuf,
 				ASN1_ITEM_rptr(CMS_Attributes_Sign));
 	if(!abuf)
 		goto err;
-	if (EVP_SignUpdate(&mctx, abuf, alen) <= 0)
+	if (EVP_DigestSignUpdate(&mctx, abuf, alen) <= 0)
+		goto err;
+	if (EVP_DigestSignFinal(&mctx, NULL, &siglen) <= 0)
 		goto err;
-	siglen = EVP_PKEY_size(si->pkey);
 	OPENSSL_free(abuf);
 	abuf = OPENSSL_malloc(siglen);
 	if(!abuf)
 		goto err;
-	cms_fixup_mctx(&mctx, si->pkey);
-	if (EVP_SignFinal(&mctx, abuf, &siglen, si->pkey) <= 0)
+	if (EVP_DigestSignFinal(&mctx, abuf, &siglen) <= 0)
 		goto err;
-#if 0
+
 	if (EVP_PKEY_CTX_ctrl(pctx, -1, EVP_PKEY_OP_SIGN,
 				EVP_PKEY_CTRL_CMS_SIGN, 1, si) <= 0)
 		{
 		CMSerr(CMS_F_CMS_SIGNERINFO_SIGN, CMS_R_CTRL_ERROR);
 		goto err;
 		}
-#endif
+
 	EVP_MD_CTX_cleanup(&mctx);
 
 	ASN1_STRING_set0(si->signature, abuf, siglen);
@@ -799,6 +769,7 @@ int CMS_SignerInfo_sign(CMS_SignerInfo *si)
 int CMS_SignerInfo_verify(CMS_SignerInfo *si)
 	{
 	EVP_MD_CTX mctx;
+	EVP_PKEY_CTX *pctx;
 	unsigned char *abuf = NULL;
 	int alen, r = -1;
 	const EVP_MD *md = NULL;
@@ -813,23 +784,22 @@ int CMS_SignerInfo_verify(CMS_SignerInfo *si)
 	if (md == NULL)
 		return -1;
 	EVP_MD_CTX_init(&mctx);
-	if (EVP_VerifyInit_ex(&mctx, md, NULL) <= 0)
+	if (EVP_DigestVerifyInit(&mctx, &pctx, md, NULL, si->pkey) <= 0)
 		goto err;
 
 	alen = ASN1_item_i2d((ASN1_VALUE *)si->signedAttrs,&abuf,
 				ASN1_ITEM_rptr(CMS_Attributes_Verify));
 	if(!abuf)
 		goto err;
-	r = EVP_VerifyUpdate(&mctx, abuf, alen);
+	r = EVP_DigestVerifyUpdate(&mctx, abuf, alen);
 	OPENSSL_free(abuf);
 	if (r <= 0)
 		{
 		r = -1;
 		goto err;
 		}
-	cms_fixup_mctx(&mctx, si->pkey);
-	r = EVP_VerifyFinal(&mctx,
-			si->signature->data, si->signature->length, si->pkey);
+	r = EVP_DigestVerifyFinal(&mctx,
+			si->signature->data, si->signature->length);
 	if (r <= 0)
 		CMSerr(CMS_F_CMS_SIGNERINFO_VERIFY, CMS_R_VERIFICATION_FAILURE);
 	err:
@@ -922,7 +892,6 @@ int CMS_SignerInfo_verify_content(CMS_SignerInfo *si, BIO *chain)
 		}
 	else
 		{
-		cms_fixup_mctx(&mctx, si->pkey);
 		r = EVP_VerifyFinal(&mctx, si->signature->data,
 					si->signature->length, si->pkey);
 		if (r <= 0)
@@ -991,17 +960,19 @@ static int cms_add_cipher_smcap(STACK_OF(X509_ALGOR) **sk, int nid, int arg)
 		return CMS_add_simple_smimecap(sk, nid, arg);
 	return 1;
 	}
-#if 0
+
 static int cms_add_digest_smcap(STACK_OF(X509_ALGOR) **sk, int nid, int arg)
 	{
 	if (EVP_get_digestbynid(nid))
 		return CMS_add_simple_smimecap(sk, nid, arg);
 	return 1;
 	}
-#endif
+
 int CMS_add_standard_smimecap(STACK_OF(X509_ALGOR) **smcap)
 	{
 	if (!cms_add_cipher_smcap(smcap, NID_aes_256_cbc, -1)
+		|| !cms_add_digest_smcap(smcap, NID_id_GostR3411_94, -1)
+		|| !cms_add_cipher_smcap(smcap, NID_id_Gost28147_89, -1)
 		|| !cms_add_cipher_smcap(smcap, NID_aes_192_cbc, -1)
 		|| !cms_add_cipher_smcap(smcap, NID_aes_128_cbc, -1)
 		|| !cms_add_cipher_smcap(smcap, NID_des_ede3_cbc, -1)
diff --git a/crypto/cms/cms_smime.c b/crypto/cms/cms_smime.c
index 2be07c2099af..8c56e3a8520d 100644
--- a/crypto/cms/cms_smime.c
+++ b/crypto/cms/cms_smime.c
@@ -171,7 +171,7 @@ CMS_ContentInfo *CMS_data_create(BIO *in, unsigned int flags)
 	if (!cms)
 		return NULL;
 
-	if (CMS_final(cms, in, NULL, flags))
+	if ((flags & CMS_STREAM) || CMS_final(cms, in, NULL, flags))
 		return cms;
 
 	CMS_ContentInfo_free(cms);
@@ -214,10 +214,7 @@ CMS_ContentInfo *CMS_digest_create(BIO *in, const EVP_MD *md,
 		return NULL;
 
 	if(!(flags & CMS_DETACHED))
-		{
-		flags &= ~CMS_STREAM;
 		CMS_set_detached(cms, 0);
-		}
 
 	if ((flags & CMS_STREAM) || CMS_final(cms, in, NULL, flags))
 		return cms;
@@ -269,10 +266,7 @@ CMS_ContentInfo *CMS_EncryptedData_encrypt(BIO *in, const EVP_CIPHER *cipher,
 		return NULL;
 
 	if(!(flags & CMS_DETACHED))
-		{
-		flags &= ~CMS_STREAM;
 		CMS_set_detached(cms, 0);
-		}
 
 	if ((flags & (CMS_STREAM|CMS_PARTIAL))
 		|| CMS_final(cms, in, NULL, flags))
@@ -456,6 +450,7 @@ int CMS_verify_receipt(CMS_ContentInfo *rcms, CMS_ContentInfo *ocms,
 			X509_STORE *store, unsigned int flags)
 	{
 	int r;
+	flags &= ~(CMS_DETACHED|CMS_TEXT);
 	r = CMS_verify(rcms, certs, store, NULL, NULL, flags);
 	if (r <= 0)
 		return r;
@@ -486,10 +481,7 @@ CMS_ContentInfo *CMS_sign(X509 *signcert, EVP_PKEY *pkey, STACK_OF(X509) *certs,
 		}
 
 	if(!(flags & CMS_DETACHED))
-		{
-		flags &= ~CMS_STREAM;
 		CMS_set_detached(cms, 0);
-		}
 
 	if ((flags & (CMS_STREAM|CMS_PARTIAL))
 		|| CMS_final(cms, data, NULL, flags))
@@ -517,7 +509,7 @@ CMS_ContentInfo *CMS_sign_receipt(CMS_SignerInfo *si,
 	BIO *rct_cont = NULL;
 	int r = 0;
 
-	flags &= ~CMS_STREAM;
+	flags &= ~(CMS_STREAM|CMS_TEXT);
 	/* Not really detached but avoids content being allocated */
 	flags |= CMS_PARTIAL|CMS_BINARY|CMS_DETACHED;
 	if (!pkey || !signcert)
@@ -598,10 +590,7 @@ CMS_ContentInfo *CMS_encrypt(STACK_OF(X509) *certs, BIO *data,
 		}
 
 	if(!(flags & CMS_DETACHED))
-		{
-		flags &= ~CMS_STREAM;
 		CMS_set_detached(cms, 0);
-		}
 
 	if ((flags & (CMS_STREAM|CMS_PARTIAL))
 		|| CMS_final(cms, data, NULL, flags))
@@ -715,6 +704,30 @@ int CMS_decrypt_set1_key(CMS_ContentInfo *cms,
 	return 0;
 
 	}
+
+int CMS_decrypt_set1_password(CMS_ContentInfo *cms, 
+				unsigned char *pass, ossl_ssize_t passlen)
+	{
+	STACK_OF(CMS_RecipientInfo) *ris;
+	CMS_RecipientInfo *ri;
+	int i, r;
+	ris = CMS_get0_RecipientInfos(cms);
+	for (i = 0; i < sk_CMS_RecipientInfo_num(ris); i++)
+		{
+		ri = sk_CMS_RecipientInfo_value(ris, i);
+		if (CMS_RecipientInfo_type(ri) != CMS_RECIPINFO_PASS)
+				continue;
+		CMS_RecipientInfo_set0_password(ri, pass, passlen);
+		r = CMS_RecipientInfo_decrypt(cms, ri);
+		CMS_RecipientInfo_set0_password(ri, NULL, 0);
+		if (r > 0)
+			return 1;
+		}
+
+	CMSerr(CMS_F_CMS_DECRYPT_SET1_PASSWORD, CMS_R_NO_MATCHING_RECIPIENT);
+	return 0;
+
+	}
 	
 int CMS_decrypt(CMS_ContentInfo *cms, EVP_PKEY *pk, X509 *cert,
 				BIO *dcont, BIO *out,
@@ -810,12 +823,9 @@ CMS_ContentInfo *CMS_compress(BIO *in, int comp_nid, unsigned int flags)
 		return NULL;
 
 	if(!(flags & CMS_DETACHED))
-		{
-		flags &= ~CMS_STREAM;
 		CMS_set_detached(cms, 0);
-		}
 
-	if (CMS_final(cms, in, NULL, flags))
+	if ((flags & CMS_STREAM) || CMS_final(cms, in, NULL, flags))
 		return cms;
 
 	CMS_ContentInfo_free(cms);
diff --git a/crypto/comp/Makefile b/crypto/comp/Makefile
index 5d364b85133b..efda832dce47 100644
--- a/crypto/comp/Makefile
+++ b/crypto/comp/Makefile
@@ -36,7 +36,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
diff --git a/crypto/comp/c_rle.c b/crypto/comp/c_rle.c
index 18bceae51e76..47dfb67fbd52 100644
--- a/crypto/comp/c_rle.c
+++ b/crypto/comp/c_rle.c
@@ -30,7 +30,7 @@ static int rle_compress_block(COMP_CTX *ctx, unsigned char *out,
 	{
 	/* int i; */
 
-	if (olen < (ilen+1))
+	if (ilen == 0 || olen < (ilen-1))
 		{
 		/* ZZZZZZZZZZZZZZZZZZZZZZ */
 		return(-1);
@@ -46,7 +46,7 @@ static int rle_expand_block(COMP_CTX *ctx, unsigned char *out,
 	{
 	int i;
 
-	if (ilen == 0 || olen < (ilen-1))
+	if (olen < (ilen-1))
 		{
 		/* ZZZZZZZZZZZZZZZZZZZZZZ */
 		return(-1);
diff --git a/crypto/comp/c_zlib.c b/crypto/comp/c_zlib.c
index 8df7792c50b7..8adf35f3fc9c 100644
--- a/crypto/comp/c_zlib.c
+++ b/crypto/comp/c_zlib.c
@@ -781,6 +781,7 @@ static long bio_zlib_ctrl(BIO *b, int cmd, long num, void *ptr)
 	default:
 		ret = BIO_ctrl(b->next_bio, cmd, num, ptr);
 		break;
+
 		}
 
 	return ret;
diff --git a/crypto/comp/comp_err.c b/crypto/comp/comp_err.c
index 187d68b725ca..661c94c3a43b 100644
--- a/crypto/comp/comp_err.c
+++ b/crypto/comp/comp_err.c
@@ -1,6 +1,6 @@
 /* crypto/comp/comp_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2008 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
diff --git a/crypto/conf/Makefile b/crypto/conf/Makefile
index ccd07213326e..78bb3241065d 100644
--- a/crypto/conf/Makefile
+++ b/crypto/conf/Makefile
@@ -36,7 +36,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -114,8 +114,8 @@ conf_mall.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
 conf_mall.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 conf_mall.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 conf_mall.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-conf_mall.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-conf_mall.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+conf_mall.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+conf_mall.o: ../../include/openssl/objects.h
 conf_mall.o: ../../include/openssl/opensslconf.h
 conf_mall.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 conf_mall.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
@@ -128,9 +128,9 @@ conf_mod.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 conf_mod.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
 conf_mod.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 conf_mod.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-conf_mod.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-conf_mod.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-conf_mod.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+conf_mod.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+conf_mod.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+conf_mod.o: ../../include/openssl/opensslconf.h
 conf_mod.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 conf_mod.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 conf_mod.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -143,9 +143,8 @@ conf_sap.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
 conf_sap.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 conf_sap.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 conf_sap.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-conf_sap.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-conf_sap.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-conf_sap.o: ../../include/openssl/opensslconf.h
+conf_sap.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+conf_sap.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 conf_sap.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 conf_sap.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 conf_sap.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
diff --git a/crypto/conf/README b/crypto/conf/README
index ca58d0240f57..96e53b34ed85 100644
--- a/crypto/conf/README
+++ b/crypto/conf/README
@@ -1,8 +1,3 @@
-WARNING WARNING WARNING!!!
-
-This stuff is experimental, may change radically or be deleted altogether
-before OpenSSL 0.9.7 release. You have been warned!
-
 Configuration modules. These are a set of modules which can perform
 various configuration functions.
 
@@ -13,7 +8,7 @@ The routines read a configuration file set up like this:
 
 -----
 #default section
-openssl_init=init_section
+openssl_conf=init_section
 
 [init_section]
 
@@ -30,29 +25,27 @@ path=/some/path/to/some/dso.so
 other_stuff=other_value
 ----
 
-When this file is loaded a configuration module with the specified
-string (module* in the above example) is looked up and its init
-function called as:
+When this file is loaded a configuration module with the specified string
+(module* in the above example) is looked up and its init function called as:
 
 int conf_init_func(CONF_IMODULE *md, CONF *cnf);
 
-The function can then take whatever action is appropriate, for example
-further lookups based on the value. Multiple instances of the same 
-config module can be loaded.
+The function can then take whatever action is appropriate, for example further
+lookups based on the value. Multiple instances of the same config module can be
+loaded.
 
-When the application closes down the modules are cleaned up by calling
-an optional finish function:
+When the application closes down the modules are cleaned up by calling an
+optional finish function:
 
 void conf_finish_func(CONF_IMODULE *md);
 
 The finish functions are called in reverse order: that is the last module
 loaded is the first one cleaned up.
 
-If no module exists with a given name then an attempt is made to load
-a DSO with the supplied name. This might mean that "module3" attempts
-to load a DSO called libmodule3.so or module3.dll for example. An explicit
-DSO name can be given by including a separate section as in the module4 example
-above.
+If no module exists with a given name then an attempt is made to load a DSO
+with the supplied name. This might mean that "module3" attempts to load a DSO
+called libmodule3.so or module3.dll for example. An explicit DSO name can be
+given by including a separate section as in the module4 example above.
 
 The DSO is expected to at least contain an initialization function:
 
@@ -64,15 +57,17 @@ void OPENSSL_finish(CONF_IMODULE *md);
 
 Static modules can also be added using,
 
-int CONF_module_add(char *name, dso_mod_init_func *ifunc, dso_mod_finish_func *ffunc);
+int CONF_module_add(char *name, dso_mod_init_func *ifunc, dso_mod_finish_func
+*ffunc);
 
-where "name" is the name in the configuration file this function corresponds to.
+where "name" is the name in the configuration file this function corresponds
+to.
 
-A set of builtin modules (currently only an ASN1 non functional test module) can be 
-added by calling OPENSSL_load_builtin_modules(). 
+A set of builtin modules (currently only an ASN1 non functional test module)
+can be added by calling OPENSSL_load_builtin_modules(). 
 
-The function OPENSSL_config() is intended as a simple configuration function that
-any application can call to perform various default configuration tasks. It uses the
-file openssl.cnf in the usual locations.
+The function OPENSSL_config() is intended as a simple configuration function
+that any application can call to perform various default configuration tasks.
+It uses the file openssl.cnf in the usual locations.
 
 
diff --git a/crypto/conf/conf.h b/crypto/conf/conf.h
index 8aa06bc5ecbd..c2199978a34c 100644
--- a/crypto/conf/conf.h
+++ b/crypto/conf/conf.h
@@ -79,8 +79,7 @@ typedef struct
 	} CONF_VALUE;
 
 DECLARE_STACK_OF(CONF_VALUE)
-DECLARE_STACK_OF(CONF_MODULE)
-DECLARE_STACK_OF(CONF_IMODULE)
+DECLARE_LHASH_OF(CONF_VALUE);
 
 struct conf_st;
 struct conf_method_st;
@@ -105,6 +104,9 @@ struct conf_method_st
 typedef struct conf_imodule_st CONF_IMODULE;
 typedef struct conf_module_st CONF_MODULE;
 
+DECLARE_STACK_OF(CONF_MODULE)
+DECLARE_STACK_OF(CONF_IMODULE)
+
 /* DSO module function typedefs */
 typedef int conf_init_func(CONF_IMODULE *md, const CONF *cnf);
 typedef void conf_finish_func(CONF_IMODULE *md);
@@ -117,18 +119,23 @@ typedef void conf_finish_func(CONF_IMODULE *md);
 #define CONF_MFLAGS_DEFAULT_SECTION	0x20
 
 int CONF_set_default_method(CONF_METHOD *meth);
-void CONF_set_nconf(CONF *conf,LHASH *hash);
-LHASH *CONF_load(LHASH *conf,const char *file,long *eline);
+void CONF_set_nconf(CONF *conf,LHASH_OF(CONF_VALUE) *hash);
+LHASH_OF(CONF_VALUE) *CONF_load(LHASH_OF(CONF_VALUE) *conf,const char *file,
+				long *eline);
 #ifndef OPENSSL_NO_FP_API
-LHASH *CONF_load_fp(LHASH *conf, FILE *fp,long *eline);
+LHASH_OF(CONF_VALUE) *CONF_load_fp(LHASH_OF(CONF_VALUE) *conf, FILE *fp,
+				   long *eline);
 #endif
-LHASH *CONF_load_bio(LHASH *conf, BIO *bp,long *eline);
-STACK_OF(CONF_VALUE) *CONF_get_section(LHASH *conf,const char *section);
-char *CONF_get_string(LHASH *conf,const char *group,const char *name);
-long CONF_get_number(LHASH *conf,const char *group,const char *name);
-void CONF_free(LHASH *conf);
-int CONF_dump_fp(LHASH *conf, FILE *out);
-int CONF_dump_bio(LHASH *conf, BIO *out);
+LHASH_OF(CONF_VALUE) *CONF_load_bio(LHASH_OF(CONF_VALUE) *conf, BIO *bp,long *eline);
+STACK_OF(CONF_VALUE) *CONF_get_section(LHASH_OF(CONF_VALUE) *conf,
+				       const char *section);
+char *CONF_get_string(LHASH_OF(CONF_VALUE) *conf,const char *group,
+		      const char *name);
+long CONF_get_number(LHASH_OF(CONF_VALUE) *conf,const char *group,
+		     const char *name);
+void CONF_free(LHASH_OF(CONF_VALUE) *conf);
+int CONF_dump_fp(LHASH_OF(CONF_VALUE) *conf, FILE *out);
+int CONF_dump_bio(LHASH_OF(CONF_VALUE) *conf, BIO *out);
 
 void OPENSSL_config(const char *config_name);
 void OPENSSL_no_config(void);
@@ -140,7 +147,7 @@ struct conf_st
 	{
 	CONF_METHOD *meth;
 	void *meth_data;
-	LHASH *data;
+	LHASH_OF(CONF_VALUE) *data;
 	};
 
 CONF *NCONF_new(CONF_METHOD *meth);
@@ -214,6 +221,7 @@ void ERR_load_CONF_strings(void);
 #define CONF_F_CONF_LOAD_BIO				 102
 #define CONF_F_CONF_LOAD_FP				 103
 #define CONF_F_CONF_MODULES_LOAD			 116
+#define CONF_F_CONF_PARSE_LIST				 119
 #define CONF_F_DEF_LOAD					 120
 #define CONF_F_DEF_LOAD_BIO				 121
 #define CONF_F_MODULE_INIT				 115
@@ -233,6 +241,7 @@ void ERR_load_CONF_strings(void);
 
 /* Reason codes. */
 #define CONF_R_ERROR_LOADING_DSO			 110
+#define CONF_R_LIST_CANNOT_BE_NULL			 115
 #define CONF_R_MISSING_CLOSE_SQUARE_BRACKET		 100
 #define CONF_R_MISSING_EQUAL_SIGN			 101
 #define CONF_R_MISSING_FINISH_FUNCTION			 111
diff --git a/crypto/conf/conf_api.c b/crypto/conf/conf_api.c
index 17bae8359f0c..f5fcbb9f6b15 100644
--- a/crypto/conf/conf_api.c
+++ b/crypto/conf/conf_api.c
@@ -70,16 +70,12 @@
 #include <openssl/conf_api.h>
 #include "e_os.h"
 
-static void value_free_hash(CONF_VALUE *a, LHASH *conf);
-static void value_free_stack(CONF_VALUE *a,LHASH *conf);
-static IMPLEMENT_LHASH_DOALL_ARG_FN(value_free_hash, CONF_VALUE *, LHASH *)
-static IMPLEMENT_LHASH_DOALL_ARG_FN(value_free_stack, CONF_VALUE *, LHASH *)
-/* We don't use function pointer casting or wrapper functions - but cast each
- * callback parameter inside the callback functions. */
-/* static unsigned long hash(CONF_VALUE *v); */
-static unsigned long hash(const void *v_void);
-/* static int cmp_conf(CONF_VALUE *a,CONF_VALUE *b); */
-static int cmp_conf(const void *a_void,const void *b_void);
+static void value_free_hash_doall_arg(CONF_VALUE *a,
+				      LHASH_OF(CONF_VALUE) *conf);
+static void value_free_stack_doall(CONF_VALUE *a);
+static IMPLEMENT_LHASH_DOALL_ARG_FN(value_free_hash, CONF_VALUE,
+				    LHASH_OF(CONF_VALUE))
+static IMPLEMENT_LHASH_DOALL_FN(value_free_stack, CONF_VALUE)
 
 /* Up until OpenSSL 0.9.5a, this was get_section */
 CONF_VALUE *_CONF_get_section(const CONF *conf, const char *section)
@@ -89,7 +85,7 @@ CONF_VALUE *_CONF_get_section(const CONF *conf, const char *section)
 	if ((conf == NULL) || (section == NULL)) return(NULL);
 	vv.name=NULL;
 	vv.section=(char *)section;
-	v=(CONF_VALUE *)lh_retrieve(conf->data,&vv);
+	v=lh_CONF_VALUE_retrieve(conf->data,&vv);
 	return(v);
 	}
 
@@ -119,7 +115,7 @@ int _CONF_add_string(CONF *conf, CONF_VALUE *section, CONF_VALUE *value)
 		return 0;
 		}
 
-	v = (CONF_VALUE *)lh_insert(conf->data, value);
+	v = lh_CONF_VALUE_insert(conf->data, value);
 	if (v != NULL)
 		{
 		(void)sk_CONF_VALUE_delete_ptr(ts,v);
@@ -142,24 +138,24 @@ char *_CONF_get_string(const CONF *conf, const char *section, const char *name)
 			{
 			vv.name=(char *)name;
 			vv.section=(char *)section;
-			v=(CONF_VALUE *)lh_retrieve(conf->data,&vv);
+			v=lh_CONF_VALUE_retrieve(conf->data,&vv);
 			if (v != NULL) return(v->value);
 			if (strcmp(section,"ENV") == 0)
 				{
-				p=Getenv(name);
+				p=getenv(name);
 				if (p != NULL) return(p);
 				}
 			}
 		vv.section="default";
 		vv.name=(char *)name;
-		v=(CONF_VALUE *)lh_retrieve(conf->data,&vv);
+		v=lh_CONF_VALUE_retrieve(conf->data,&vv);
 		if (v != NULL)
 			return(v->value);
 		else
 			return(NULL);
 		}
 	else
-		return(Getenv(name));
+		return(getenv(name));
 	}
 
 #if 0 /* There's no way to provide error checking with this function, so
@@ -183,6 +179,34 @@ long _CONF_get_number(CONF *conf, char *section, char *name)
 	}
 #endif
 
+static unsigned long conf_value_hash(const CONF_VALUE *v)
+	{
+	return (lh_strhash(v->section)<<2)^lh_strhash(v->name);
+	}
+static IMPLEMENT_LHASH_HASH_FN(conf_value, CONF_VALUE)
+
+static int conf_value_cmp(const CONF_VALUE *a, const CONF_VALUE *b)
+	{
+	int i;
+
+	if (a->section != b->section)
+		{
+		i=strcmp(a->section,b->section);
+		if (i) return(i);
+		}
+
+	if ((a->name != NULL) && (b->name != NULL))
+		{
+		i=strcmp(a->name,b->name);
+		return(i);
+		}
+	else if (a->name == b->name)
+		return(0);
+	else
+		return((a->name == NULL)?-1:1);
+	}
+static IMPLEMENT_LHASH_COMP_FN(conf_value, CONF_VALUE)
+
 int _CONF_new_data(CONF *conf)
 	{
 	if (conf == NULL)
@@ -190,7 +214,7 @@ int _CONF_new_data(CONF *conf)
 		return 0;
 		}
 	if (conf->data == NULL)
-		if ((conf->data = lh_new(hash, cmp_conf)) == NULL)
+		if ((conf->data = lh_CONF_VALUE_new()) == NULL)
 			{
 			return 0;
 			}
@@ -201,105 +225,73 @@ void _CONF_free_data(CONF *conf)
 	{
 	if (conf == NULL || conf->data == NULL) return;
 
-	conf->data->down_load=0; /* evil thing to make sure the 'OPENSSL_free()'
-				  * works as expected */
-	lh_doall_arg(conf->data, LHASH_DOALL_ARG_FN(value_free_hash),
-			conf->data);
+	lh_CONF_VALUE_down_load(conf->data)=0; /* evil thing to make
+				  * sure the 'OPENSSL_free()' works as
+				  * expected */
+	lh_CONF_VALUE_doall_arg(conf->data,
+				LHASH_DOALL_ARG_FN(value_free_hash),
+				LHASH_OF(CONF_VALUE), conf->data);
 
 	/* We now have only 'section' entries in the hash table.
 	 * Due to problems with */
 
-	lh_doall_arg(conf->data, LHASH_DOALL_ARG_FN(value_free_stack),
-			conf->data);
-	lh_free(conf->data);
+	lh_CONF_VALUE_doall(conf->data, LHASH_DOALL_FN(value_free_stack));
+	lh_CONF_VALUE_free(conf->data);
 	}
 
-static void value_free_hash(CONF_VALUE *a, LHASH *conf)
+static void value_free_hash_doall_arg(CONF_VALUE *a, LHASH_OF(CONF_VALUE) *conf)
 	{
 	if (a->name != NULL)
-		{
-		a=(CONF_VALUE *)lh_delete(conf,a);
-		}
+		(void)lh_CONF_VALUE_delete(conf,a);
 	}
 
-static void value_free_stack(CONF_VALUE *a, LHASH *conf)
+static void value_free_stack_doall(CONF_VALUE *a)
 	{
 	CONF_VALUE *vv;
-	STACK *sk;
+	STACK_OF(CONF_VALUE) *sk;
 	int i;
 
 	if (a->name != NULL) return;
 
-	sk=(STACK *)a->value;
-	for (i=sk_num(sk)-1; i>=0; i--)
+	sk=(STACK_OF(CONF_VALUE) *)a->value;
+	for (i=sk_CONF_VALUE_num(sk)-1; i>=0; i--)
 		{
-		vv=(CONF_VALUE *)sk_value(sk,i);
+		vv=sk_CONF_VALUE_value(sk,i);
 		OPENSSL_free(vv->value);
 		OPENSSL_free(vv->name);
 		OPENSSL_free(vv);
 		}
-	if (sk != NULL) sk_free(sk);
+	if (sk != NULL) sk_CONF_VALUE_free(sk);
 	OPENSSL_free(a->section);
 	OPENSSL_free(a);
 	}
 
-/* static unsigned long hash(CONF_VALUE *v) */
-static unsigned long hash(const void *v_void)
-	{
-	CONF_VALUE *v = (CONF_VALUE *)v_void;
-	return((lh_strhash(v->section)<<2)^lh_strhash(v->name));
-	}
-
-/* static int cmp_conf(CONF_VALUE *a, CONF_VALUE *b) */
-static int cmp_conf(const void *a_void,const  void *b_void)
-	{
-	int i;
-	CONF_VALUE *a = (CONF_VALUE *)a_void;
-	CONF_VALUE *b = (CONF_VALUE *)b_void;
-
-	if (a->section != b->section)
-		{
-		i=strcmp(a->section,b->section);
-		if (i) return(i);
-		}
-
-	if ((a->name != NULL) && (b->name != NULL))
-		{
-		i=strcmp(a->name,b->name);
-		return(i);
-		}
-	else if (a->name == b->name)
-		return(0);
-	else
-		return((a->name == NULL)?-1:1);
-	}
-
 /* Up until OpenSSL 0.9.5a, this was new_section */
 CONF_VALUE *_CONF_new_section(CONF *conf, const char *section)
 	{
-	STACK *sk=NULL;
+	STACK_OF(CONF_VALUE) *sk=NULL;
 	int ok=0,i;
 	CONF_VALUE *v=NULL,*vv;
 
-	if ((sk=sk_new_null()) == NULL)
+	if ((sk=sk_CONF_VALUE_new_null()) == NULL)
 		goto err;
-	if ((v=(CONF_VALUE *)OPENSSL_malloc(sizeof(CONF_VALUE))) == NULL)
+	if ((v=OPENSSL_malloc(sizeof(CONF_VALUE))) == NULL)
 		goto err;
 	i=strlen(section)+1;
-	if ((v->section=(char *)OPENSSL_malloc(i)) == NULL)
+	if ((v->section=OPENSSL_malloc(i)) == NULL)
 		goto err;
 
 	memcpy(v->section,section,i);
 	v->name=NULL;
 	v->value=(char *)sk;
 	
-	vv=(CONF_VALUE *)lh_insert(conf->data,v);
-	assert(vv == NULL);
+	vv=lh_CONF_VALUE_insert(conf->data,v);
+	OPENSSL_assert(vv == NULL);
 	ok=1;
 err:
 	if (!ok)
 		{
-		if (sk != NULL) sk_free(sk);
+		if (sk != NULL) sk_CONF_VALUE_free(sk);
 		if (v != NULL) OPENSSL_free(v);
 		v=NULL;
 		}
diff --git a/crypto/conf/conf_def.c b/crypto/conf/conf_def.c
index 3c58936d9928..cf951320af8e 100644
--- a/crypto/conf/conf_def.c
+++ b/crypto/conf/conf_def.c
@@ -129,7 +129,7 @@ static CONF *def_create(CONF_METHOD *meth)
 	{
 	CONF *ret;
 
-	ret = (CONF *)OPENSSL_malloc(sizeof(CONF) + sizeof(unsigned short *));
+	ret = OPENSSL_malloc(sizeof(CONF) + sizeof(unsigned short *));
 	if (ret)
 		if (meth->init(ret) == 0)
 			{
@@ -145,7 +145,7 @@ static int def_init_default(CONF *conf)
 		return 0;
 
 	conf->meth = &default_method;
-	conf->meth_data = (void *)CONF_type_default;
+	conf->meth_data = CONF_type_default;
 	conf->data = NULL;
 
 	return 1;
@@ -219,8 +219,6 @@ static int def_load_bio(CONF *conf, BIO *in, long *line)
 	CONF_VALUE *v=NULL,*tv;
 	CONF_VALUE *sv=NULL;
 	char *section=NULL,*buf;
-/*	STACK_OF(CONF_VALUE) *section_sk=NULL;*/
-/*	STACK_OF(CONF_VALUE) *ts=NULL;*/
 	char *start,*psection,*pname;
 	void *h = (void *)(conf->data);
 
@@ -251,7 +249,6 @@ static int def_load_bio(CONF *conf, BIO *in, long *line)
 					CONF_R_UNABLE_TO_CREATE_NEW_SECTION);
 		goto err;
 		}
-/*	section_sk=(STACK_OF(CONF_VALUE) *)sv->value;*/
 
 	bufnum=0;
 	again=0;
@@ -343,7 +340,6 @@ again:
 					CONF_R_UNABLE_TO_CREATE_NEW_SECTION);
 				goto err;
 				}
-/*			section_sk=(STACK_OF(CONF_VALUE) *)sv->value;*/
 			continue;
 			}
 		else
@@ -406,13 +402,9 @@ again:
 					   CONF_R_UNABLE_TO_CREATE_NEW_SECTION);
 					goto err;
 					}
-/*				ts=(STACK_OF(CONF_VALUE) *)tv->value;*/
 				}
 			else
-				{
 				tv=sv;
-/*				ts=section_sk;*/
-				}
 #if 1
 			if (_CONF_add_string(conf, tv, v) == 0)
 				{
@@ -719,7 +711,7 @@ static char *scan_dquote(CONF *conf, char *p)
 	return(p);
 	}
 
-static void dump_value(CONF_VALUE *a, BIO *out)
+static void dump_value_doall_arg(CONF_VALUE *a, BIO *out)
 	{
 	if (a->name)
 		BIO_printf(out, "[%s] %s=%s\n", a->section, a->name, a->value);
@@ -727,11 +719,12 @@ static void dump_value(CONF_VALUE *a, BIO *out)
 		BIO_printf(out, "[[%s]]\n", a->section);
 	}
 
-static IMPLEMENT_LHASH_DOALL_ARG_FN(dump_value, CONF_VALUE *, BIO *)
+static IMPLEMENT_LHASH_DOALL_ARG_FN(dump_value, CONF_VALUE, BIO)
 
 static int def_dump(const CONF *conf, BIO *out)
 	{
-	lh_doall_arg(conf->data, LHASH_DOALL_ARG_FN(dump_value), out);
+	lh_CONF_VALUE_doall_arg(conf->data, LHASH_DOALL_ARG_FN(dump_value),
+				BIO, out);
 	return 1;
 	}
 
diff --git a/crypto/conf/conf_err.c b/crypto/conf/conf_err.c
index a16a5e0bd44b..25bb5dc9aa36 100644
--- a/crypto/conf/conf_err.c
+++ b/crypto/conf/conf_err.c
@@ -1,6 +1,6 @@
 /* crypto/conf/conf_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -75,6 +75,7 @@ static ERR_STRING_DATA CONF_str_functs[]=
 {ERR_FUNC(CONF_F_CONF_LOAD_BIO),	"CONF_load_bio"},
 {ERR_FUNC(CONF_F_CONF_LOAD_FP),	"CONF_load_fp"},
 {ERR_FUNC(CONF_F_CONF_MODULES_LOAD),	"CONF_modules_load"},
+{ERR_FUNC(CONF_F_CONF_PARSE_LIST),	"CONF_parse_list"},
 {ERR_FUNC(CONF_F_DEF_LOAD),	"DEF_LOAD"},
 {ERR_FUNC(CONF_F_DEF_LOAD_BIO),	"DEF_LOAD_BIO"},
 {ERR_FUNC(CONF_F_MODULE_INIT),	"MODULE_INIT"},
@@ -97,6 +98,7 @@ static ERR_STRING_DATA CONF_str_functs[]=
 static ERR_STRING_DATA CONF_str_reasons[]=
 	{
 {ERR_REASON(CONF_R_ERROR_LOADING_DSO)    ,"error loading dso"},
+{ERR_REASON(CONF_R_LIST_CANNOT_BE_NULL)  ,"list cannot be null"},
 {ERR_REASON(CONF_R_MISSING_CLOSE_SQUARE_BRACKET),"missing close square bracket"},
 {ERR_REASON(CONF_R_MISSING_EQUAL_SIGN)   ,"missing equal sign"},
 {ERR_REASON(CONF_R_MISSING_FINISH_FUNCTION),"missing finish function"},
diff --git a/crypto/conf/conf_lib.c b/crypto/conf/conf_lib.c
index 2a3399d269e4..54046defca8b 100644
--- a/crypto/conf/conf_lib.c
+++ b/crypto/conf/conf_lib.c
@@ -69,7 +69,7 @@ static CONF_METHOD *default_CONF_method=NULL;
 
 /* Init a 'CONF' structure from an old LHASH */
 
-void CONF_set_nconf(CONF *conf, LHASH *hash)
+void CONF_set_nconf(CONF *conf, LHASH_OF(CONF_VALUE) *hash)
 	{
 	if (default_CONF_method == NULL)
 		default_CONF_method = NCONF_default();
@@ -87,9 +87,10 @@ int CONF_set_default_method(CONF_METHOD *meth)
 	return 1;
 	}
 
-LHASH *CONF_load(LHASH *conf, const char *file, long *eline)
+LHASH_OF(CONF_VALUE) *CONF_load(LHASH_OF(CONF_VALUE) *conf, const char *file,
+				long *eline)
 	{
-	LHASH *ltmp;
+	LHASH_OF(CONF_VALUE) *ltmp;
 	BIO *in=NULL;
 
 #ifdef OPENSSL_SYS_VMS
@@ -110,10 +111,11 @@ LHASH *CONF_load(LHASH *conf, const char *file, long *eline)
 	}
 
 #ifndef OPENSSL_NO_FP_API
-LHASH *CONF_load_fp(LHASH *conf, FILE *fp,long *eline)
+LHASH_OF(CONF_VALUE) *CONF_load_fp(LHASH_OF(CONF_VALUE) *conf, FILE *fp,
+				   long *eline)
 	{
 	BIO *btmp;
-	LHASH *ltmp;
+	LHASH_OF(CONF_VALUE) *ltmp;
 	if(!(btmp = BIO_new_fp(fp, BIO_NOCLOSE))) {
 		CONFerr(CONF_F_CONF_LOAD_FP,ERR_R_BUF_LIB);
 		return NULL;
@@ -124,7 +126,8 @@ LHASH *CONF_load_fp(LHASH *conf, FILE *fp,long *eline)
 	}
 #endif
 
-LHASH *CONF_load_bio(LHASH *conf, BIO *bp,long *eline)
+LHASH_OF(CONF_VALUE) *CONF_load_bio(LHASH_OF(CONF_VALUE) *conf, BIO *bp,
+				    long *eline)
 	{
 	CONF ctmp;
 	int ret;
@@ -137,7 +140,8 @@ LHASH *CONF_load_bio(LHASH *conf, BIO *bp,long *eline)
 	return NULL;
 	}
 
-STACK_OF(CONF_VALUE) *CONF_get_section(LHASH *conf,const char *section)
+STACK_OF(CONF_VALUE) *CONF_get_section(LHASH_OF(CONF_VALUE) *conf,
+				       const char *section)
 	{
 	if (conf == NULL)
 		{
@@ -151,7 +155,8 @@ STACK_OF(CONF_VALUE) *CONF_get_section(LHASH *conf,const char *section)
 		}
 	}
 
-char *CONF_get_string(LHASH *conf,const char *group,const char *name)
+char *CONF_get_string(LHASH_OF(CONF_VALUE) *conf,const char *group,
+		      const char *name)
 	{
 	if (conf == NULL)
 		{
@@ -165,7 +170,8 @@ char *CONF_get_string(LHASH *conf,const char *group,const char *name)
 		}
 	}
 
-long CONF_get_number(LHASH *conf,const char *group,const char *name)
+long CONF_get_number(LHASH_OF(CONF_VALUE) *conf,const char *group,
+		     const char *name)
 	{
 	int status;
 	long result = 0;
@@ -189,7 +195,7 @@ long CONF_get_number(LHASH *conf,const char *group,const char *name)
 	return result;
 	}
 
-void CONF_free(LHASH *conf)
+void CONF_free(LHASH_OF(CONF_VALUE) *conf)
 	{
 	CONF ctmp;
 	CONF_set_nconf(&ctmp, conf);
@@ -197,7 +203,7 @@ void CONF_free(LHASH *conf)
 	}
 
 #ifndef OPENSSL_NO_FP_API
-int CONF_dump_fp(LHASH *conf, FILE *out)
+int CONF_dump_fp(LHASH_OF(CONF_VALUE) *conf, FILE *out)
 	{
 	BIO *btmp;
 	int ret;
@@ -212,7 +218,7 @@ int CONF_dump_fp(LHASH *conf, FILE *out)
 	}
 #endif
 
-int CONF_dump_bio(LHASH *conf, BIO *out)
+int CONF_dump_bio(LHASH_OF(CONF_VALUE) *conf, BIO *out)
 	{
 	CONF ctmp;
 	CONF_set_nconf(&ctmp, conf);
diff --git a/crypto/conf/conf_mall.c b/crypto/conf/conf_mall.c
index 1cc1fd55345a..c6f4cb2d55be 100644
--- a/crypto/conf/conf_mall.c
+++ b/crypto/conf/conf_mall.c
@@ -63,7 +63,6 @@
 #include <openssl/dso.h>
 #include <openssl/x509.h>
 #include <openssl/asn1.h>
-#include <openssl/evp.h>
 #ifndef OPENSSL_NO_ENGINE
 #include <openssl/engine.h>
 #endif
@@ -77,6 +76,5 @@ void OPENSSL_load_builtin_modules(void)
 #ifndef OPENSSL_NO_ENGINE
 	ENGINE_add_conf_module();
 #endif
-	EVP_add_alg_module();
 	}
 
diff --git a/crypto/conf/conf_mod.c b/crypto/conf/conf_mod.c
index ee9c677d9bda..df1642a0a562 100644
--- a/crypto/conf/conf_mod.c
+++ b/crypto/conf/conf_mod.c
@@ -582,8 +582,14 @@ int CONF_parse_list(const char *list_, int sep, int nospc,
 	{
 	int ret;
 	const char *lstart, *tmpend, *p;
-	lstart = list_;
 
+	if(list_ == NULL)
+		{
+		CONFerr(CONF_F_CONF_PARSE_LIST, CONF_R_LIST_CANNOT_BE_NULL);
+		return 0;
+		}
+
+	lstart = list_;
 	for(;;)
 		{
 		if (nospc)
diff --git a/crypto/cpt_err.c b/crypto/cpt_err.c
index 9fd41fff8cbd..289005f662f3 100644
--- a/crypto/cpt_err.c
+++ b/crypto/cpt_err.c
@@ -1,6 +1,6 @@
 /* crypto/cpt_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -76,6 +76,7 @@ static ERR_STRING_DATA CRYPTO_str_functs[]=
 {ERR_FUNC(CRYPTO_F_CRYPTO_SET_EX_DATA),	"CRYPTO_set_ex_data"},
 {ERR_FUNC(CRYPTO_F_DEF_ADD_INDEX),	"DEF_ADD_INDEX"},
 {ERR_FUNC(CRYPTO_F_DEF_GET_CLASS),	"DEF_GET_CLASS"},
+{ERR_FUNC(CRYPTO_F_FIPS_MODE_SET),	"FIPS_mode_set"},
 {ERR_FUNC(CRYPTO_F_INT_DUP_EX_DATA),	"INT_DUP_EX_DATA"},
 {ERR_FUNC(CRYPTO_F_INT_FREE_EX_DATA),	"INT_FREE_EX_DATA"},
 {ERR_FUNC(CRYPTO_F_INT_NEW_EX_DATA),	"INT_NEW_EX_DATA"},
@@ -84,6 +85,7 @@ static ERR_STRING_DATA CRYPTO_str_functs[]=
 
 static ERR_STRING_DATA CRYPTO_str_reasons[]=
 	{
+{ERR_REASON(CRYPTO_R_FIPS_MODE_NOT_SUPPORTED),"fips mode not supported"},
 {ERR_REASON(CRYPTO_R_NO_DYNLOCK_CREATE_CALLBACK),"no dynlock create callback"},
 {0,NULL}
 	};
diff --git a/crypto/cryptlib.c b/crypto/cryptlib.c
index dd74ea8855f4..766ea8cac737 100644
--- a/crypto/cryptlib.c
+++ b/crypto/cryptlib.c
@@ -1,6 +1,6 @@
 /* crypto/cryptlib.c */
 /* ====================================================================
- * Copyright (c) 1998-2003 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1998-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -121,17 +121,279 @@
 static double SSLeay_MSVC5_hack=0.0; /* and for VC1.5 */
 #endif
 
+DECLARE_STACK_OF(CRYPTO_dynlock)
+
+/* real #defines in crypto.h, keep these upto date */
+static const char* const lock_names[CRYPTO_NUM_LOCKS] =
+	{
+	"<<ERROR>>",
+	"err",
+	"ex_data",
+	"x509",
+	"x509_info",
+	"x509_pkey",
+	"x509_crl",
+	"x509_req",
+	"dsa",
+	"rsa",
+	"evp_pkey",
+	"x509_store",
+	"ssl_ctx",
+	"ssl_cert",
+	"ssl_session",
+	"ssl_sess_cert",
+	"ssl",
+	"ssl_method",
+	"rand",
+	"rand2",
+	"debug_malloc",
+	"BIO",
+	"gethostbyname",
+	"getservbyname",
+	"readdir",
+	"RSA_blinding",
+	"dh",
+	"debug_malloc2",
+	"dso",
+	"dynlock",
+	"engine",
+	"ui",
+	"ecdsa",
+	"ec",
+	"ecdh",
+	"bn",
+	"ec_pre_comp",
+	"store",
+	"comp",
+	"fips",
+	"fips2",
+#if CRYPTO_NUM_LOCKS != 41
+# error "Inconsistency between crypto.h and cryptlib.c"
+#endif
+	};
+
+/* This is for applications to allocate new type names in the non-dynamic
+   array of lock names.  These are numbered with positive numbers.  */
+static STACK_OF(OPENSSL_STRING) *app_locks=NULL;
+
+/* For applications that want a more dynamic way of handling threads, the
+   following stack is used.  These are externally numbered with negative
+   numbers.  */
+static STACK_OF(CRYPTO_dynlock) *dyn_locks=NULL;
+
+
 static void (MS_FAR *locking_callback)(int mode,int type,
-	const char *file,int line)=NULL;
+	const char *file,int line)=0;
 static int (MS_FAR *add_lock_callback)(int *pointer,int amount,
-	int type,const char *file,int line)=NULL;
-static unsigned long (MS_FAR *id_callback)(void)=NULL;
+	int type,const char *file,int line)=0;
+#ifndef OPENSSL_NO_DEPRECATED
+static unsigned long (MS_FAR *id_callback)(void)=0;
+#endif
+static void (MS_FAR *threadid_callback)(CRYPTO_THREADID *)=0;
+static struct CRYPTO_dynlock_value *(MS_FAR *dynlock_create_callback)
+	(const char *file,int line)=0;
+static void (MS_FAR *dynlock_lock_callback)(int mode,
+	struct CRYPTO_dynlock_value *l, const char *file,int line)=0;
+static void (MS_FAR *dynlock_destroy_callback)(struct CRYPTO_dynlock_value *l,
+	const char *file,int line)=0;
+
+int CRYPTO_get_new_lockid(char *name)
+	{
+	char *str;
+	int i;
+
+#if defined(OPENSSL_SYS_WIN32) || defined(OPENSSL_SYS_WIN16)
+	/* A hack to make Visual C++ 5.0 work correctly when linking as
+	 * a DLL using /MT. Without this, the application cannot use
+	 * any floating point printf's.
+	 * It also seems to be needed for Visual C 1.5 (win16) */
+	SSLeay_MSVC5_hack=(double)name[0]*(double)name[1];
+#endif
+
+	if ((app_locks == NULL) && ((app_locks=sk_OPENSSL_STRING_new_null()) == NULL))
+		{
+		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_LOCKID,ERR_R_MALLOC_FAILURE);
+		return(0);
+		}
+	if ((str=BUF_strdup(name)) == NULL)
+		{
+		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_LOCKID,ERR_R_MALLOC_FAILURE);
+		return(0);
+		}
+	i=sk_OPENSSL_STRING_push(app_locks,str);
+	if (!i)
+		OPENSSL_free(str);
+	else
+		i+=CRYPTO_NUM_LOCKS; /* gap of one :-) */
+	return(i);
+	}
 
 int CRYPTO_num_locks(void)
 	{
 	return CRYPTO_NUM_LOCKS;
 	}
 
+int CRYPTO_get_new_dynlockid(void)
+	{
+	int i = 0;
+	CRYPTO_dynlock *pointer = NULL;
+
+	if (dynlock_create_callback == NULL)
+		{
+		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_DYNLOCKID,CRYPTO_R_NO_DYNLOCK_CREATE_CALLBACK);
+		return(0);
+		}
+	CRYPTO_w_lock(CRYPTO_LOCK_DYNLOCK);
+	if ((dyn_locks == NULL)
+		&& ((dyn_locks=sk_CRYPTO_dynlock_new_null()) == NULL))
+		{
+		CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
+		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_DYNLOCKID,ERR_R_MALLOC_FAILURE);
+		return(0);
+		}
+	CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
+
+	pointer = (CRYPTO_dynlock *)OPENSSL_malloc(sizeof(CRYPTO_dynlock));
+	if (pointer == NULL)
+		{
+		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_DYNLOCKID,ERR_R_MALLOC_FAILURE);
+		return(0);
+		}
+	pointer->references = 1;
+	pointer->data = dynlock_create_callback(__FILE__,__LINE__);
+	if (pointer->data == NULL)
+		{
+		OPENSSL_free(pointer);
+		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_DYNLOCKID,ERR_R_MALLOC_FAILURE);
+		return(0);
+		}
+
+	CRYPTO_w_lock(CRYPTO_LOCK_DYNLOCK);
+	/* First, try to find an existing empty slot */
+	i=sk_CRYPTO_dynlock_find(dyn_locks,NULL);
+	/* If there was none, push, thereby creating a new one */
+	if (i == -1)
+		/* Since sk_push() returns the number of items on the
+		   stack, not the location of the pushed item, we need
+		   to transform the returned number into a position,
+		   by decreasing it.  */
+		i=sk_CRYPTO_dynlock_push(dyn_locks,pointer) - 1;
+	else
+		/* If we found a place with a NULL pointer, put our pointer
+		   in it.  */
+		(void)sk_CRYPTO_dynlock_set(dyn_locks,i,pointer);
+	CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
+
+	if (i == -1)
+		{
+		dynlock_destroy_callback(pointer->data,__FILE__,__LINE__);
+		OPENSSL_free(pointer);
+		}
+	else
+		i += 1; /* to avoid 0 */
+	return -i;
+	}
+
+void CRYPTO_destroy_dynlockid(int i)
+	{
+	CRYPTO_dynlock *pointer = NULL;
+	if (i)
+		i = -i-1;
+	if (dynlock_destroy_callback == NULL)
+		return;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_DYNLOCK);
+
+	if (dyn_locks == NULL || i >= sk_CRYPTO_dynlock_num(dyn_locks))
+		{
+		CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
+		return;
+		}
+	pointer = sk_CRYPTO_dynlock_value(dyn_locks, i);
+	if (pointer != NULL)
+		{
+		--pointer->references;
+#ifdef REF_CHECK
+		if (pointer->references < 0)
+			{
+			fprintf(stderr,"CRYPTO_destroy_dynlockid, bad reference count\n");
+			abort();
+			}
+		else
+#endif
+			if (pointer->references <= 0)
+				{
+				(void)sk_CRYPTO_dynlock_set(dyn_locks, i, NULL);
+				}
+			else
+				pointer = NULL;
+		}
+	CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
+
+	if (pointer)
+		{
+		dynlock_destroy_callback(pointer->data,__FILE__,__LINE__);
+		OPENSSL_free(pointer);
+		}
+	}
+
+struct CRYPTO_dynlock_value *CRYPTO_get_dynlock_value(int i)
+	{
+	CRYPTO_dynlock *pointer = NULL;
+	if (i)
+		i = -i-1;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_DYNLOCK);
+
+	if (dyn_locks != NULL && i < sk_CRYPTO_dynlock_num(dyn_locks))
+		pointer = sk_CRYPTO_dynlock_value(dyn_locks, i);
+	if (pointer)
+		pointer->references++;
+
+	CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
+
+	if (pointer)
+		return pointer->data;
+	return NULL;
+	}
+
+struct CRYPTO_dynlock_value *(*CRYPTO_get_dynlock_create_callback(void))
+	(const char *file,int line)
+	{
+	return(dynlock_create_callback);
+	}
+
+void (*CRYPTO_get_dynlock_lock_callback(void))(int mode,
+	struct CRYPTO_dynlock_value *l, const char *file,int line)
+	{
+	return(dynlock_lock_callback);
+	}
+
+void (*CRYPTO_get_dynlock_destroy_callback(void))
+	(struct CRYPTO_dynlock_value *l, const char *file,int line)
+	{
+	return(dynlock_destroy_callback);
+	}
+
+void CRYPTO_set_dynlock_create_callback(struct CRYPTO_dynlock_value *(*func)
+	(const char *file, int line))
+	{
+	dynlock_create_callback=func;
+	}
+
+void CRYPTO_set_dynlock_lock_callback(void (*func)(int mode,
+	struct CRYPTO_dynlock_value *l, const char *file, int line))
+	{
+	dynlock_lock_callback=func;
+	}
+
+void CRYPTO_set_dynlock_destroy_callback(void (*func)
+	(struct CRYPTO_dynlock_value *l, const char *file, int line))
+	{
+	dynlock_destroy_callback=func;
+	}
+
+
 void (*CRYPTO_get_locking_callback(void))(int mode,int type,const char *file,
 		int line)
 	{
@@ -147,6 +409,10 @@ int (*CRYPTO_get_add_lock_callback(void))(int *num,int mount,int type,
 void CRYPTO_set_locking_callback(void (*func)(int mode,int type,
 					      const char *file,int line))
 	{
+	/* Calling this here ensures initialisation before any threads
+	 * are started.
+	 */
+	OPENSSL_init();
 	locking_callback=func;
 	}
 
@@ -156,6 +422,108 @@ void CRYPTO_set_add_lock_callback(int (*func)(int *num,int mount,int type,
 	add_lock_callback=func;
 	}
 
+/* the memset() here and in set_pointer() seem overkill, but for the sake of
+ * CRYPTO_THREADID_cmp() this avoids any platform silliness that might cause two
+ * "equal" THREADID structs to not be memcmp()-identical. */
+void CRYPTO_THREADID_set_numeric(CRYPTO_THREADID *id, unsigned long val)
+	{
+	memset(id, 0, sizeof(*id));
+	id->val = val;
+	}
+
+static const unsigned char hash_coeffs[] = { 3, 5, 7, 11, 13, 17, 19, 23 };
+void CRYPTO_THREADID_set_pointer(CRYPTO_THREADID *id, void *ptr)
+	{
+	unsigned char *dest = (void *)&id->val;
+	unsigned int accum = 0;
+	unsigned char dnum = sizeof(id->val);
+
+	memset(id, 0, sizeof(*id));
+	id->ptr = ptr;
+	if (sizeof(id->val) >= sizeof(id->ptr))
+		{
+		/* 'ptr' can be embedded in 'val' without loss of uniqueness */
+		id->val = (unsigned long)id->ptr;
+		return;
+		}
+	/* hash ptr ==> val. Each byte of 'val' gets the mod-256 total of a
+	 * linear function over the bytes in 'ptr', the co-efficients of which
+	 * are a sequence of low-primes (hash_coeffs is an 8-element cycle) -
+	 * the starting prime for the sequence varies for each byte of 'val'
+	 * (unique polynomials unless pointers are >64-bit). For added spice,
+	 * the totals accumulate rather than restarting from zero, and the index
+	 * of the 'val' byte is added each time (position dependence). If I was
+	 * a black-belt, I'd scan big-endian pointers in reverse to give
+	 * low-order bits more play, but this isn't crypto and I'd prefer nobody
+	 * mistake it as such. Plus I'm lazy. */
+	while (dnum--)
+		{
+		const unsigned char *src = (void *)&id->ptr;
+		unsigned char snum = sizeof(id->ptr);
+		while (snum--)
+			accum += *(src++) * hash_coeffs[(snum + dnum) & 7];
+		accum += dnum;
+		*(dest++) = accum & 255;
+		}
+	}
+
+int CRYPTO_THREADID_set_callback(void (*func)(CRYPTO_THREADID *))
+	{
+	if (threadid_callback)
+		return 0;
+	threadid_callback = func;
+	return 1;
+	}
+
+void (*CRYPTO_THREADID_get_callback(void))(CRYPTO_THREADID *)
+	{
+	return threadid_callback;
+	}
+
+void CRYPTO_THREADID_current(CRYPTO_THREADID *id)
+	{
+	if (threadid_callback)
+		{
+		threadid_callback(id);
+		return;
+		}
+#ifndef OPENSSL_NO_DEPRECATED
+	/* If the deprecated callback was set, fall back to that */
+	if (id_callback)
+		{
+		CRYPTO_THREADID_set_numeric(id, id_callback());
+		return;
+		}
+#endif
+	/* Else pick a backup */
+#ifdef OPENSSL_SYS_WIN16
+	CRYPTO_THREADID_set_numeric(id, (unsigned long)GetCurrentTask());
+#elif defined(OPENSSL_SYS_WIN32)
+	CRYPTO_THREADID_set_numeric(id, (unsigned long)GetCurrentThreadId());
+#elif defined(OPENSSL_SYS_BEOS)
+	CRYPTO_THREADID_set_numeric(id, (unsigned long)find_thread(NULL));
+#else
+	/* For everything else, default to using the address of 'errno' */
+	CRYPTO_THREADID_set_pointer(id, &errno);
+#endif
+	}
+
+int CRYPTO_THREADID_cmp(const CRYPTO_THREADID *a, const CRYPTO_THREADID *b)
+	{
+	return memcmp(a, b, sizeof(*a));
+	}
+
+void CRYPTO_THREADID_cpy(CRYPTO_THREADID *dest, const CRYPTO_THREADID *src)
+	{
+	memcpy(dest, src, sizeof(*src));
+	}
+
+unsigned long CRYPTO_THREADID_hash(const CRYPTO_THREADID *id)
+	{
+	return id->val;
+	}
+
+#ifndef OPENSSL_NO_DEPRECATED
 unsigned long (*CRYPTO_get_id_callback(void))(void)
 	{
 	return(id_callback);
@@ -178,6 +546,8 @@ unsigned long CRYPTO_thread_id(void)
 		ret=(unsigned long)GetCurrentThreadId();
 #elif defined(GETPID_IS_MEANINGLESS)
 		ret=1L;
+#elif defined(OPENSSL_SYS_BEOS)
+		ret=(unsigned long)find_thread(NULL);
 #else
 		ret=(unsigned long)getpid();
 #endif
@@ -186,19 +556,13 @@ unsigned long CRYPTO_thread_id(void)
 		ret=id_callback();
 	return(ret);
 	}
-
-static void (*do_dynlock_cb)(int mode, int type, const char *file, int line);
-
-void int_CRYPTO_set_do_dynlock_callback(
-	void (*dyn_cb)(int mode, int type, const char *file, int line))
-	{
-	do_dynlock_cb = dyn_cb;
-	}
+#endif
 
 void CRYPTO_lock(int mode, int type, const char *file, int line)
 	{
 #ifdef LOCK_DEBUG
 		{
+		CRYPTO_THREADID id;
 		char *rw_text,*operation_text;
 
 		if (mode & CRYPTO_LOCK)
@@ -215,15 +579,25 @@ void CRYPTO_lock(int mode, int type, const char *file, int line)
 		else
 			rw_text="ERROR";
 
+		CRYPTO_THREADID_current(&id);
 		fprintf(stderr,"lock:%08lx:(%s)%s %-18s %s:%d\n",
-			CRYPTO_thread_id(), rw_text, operation_text,
+			CRYPTO_THREADID_hash(&id), rw_text, operation_text,
 			CRYPTO_get_lock_name(type), file, line);
 		}
 #endif
 	if (type < 0)
 		{
-		if (do_dynlock_cb)
-			do_dynlock_cb(mode, type, file, line);
+		if (dynlock_lock_callback != NULL)
+			{
+			struct CRYPTO_dynlock_value *pointer
+				= CRYPTO_get_dynlock_value(type);
+
+			OPENSSL_assert(pointer != NULL);
+
+			dynlock_lock_callback(mode, pointer, file, line);
+
+			CRYPTO_destroy_dynlockid(type);
+			}
 		}
 	else
 		if (locking_callback != NULL)
@@ -243,11 +617,14 @@ int CRYPTO_add_lock(int *pointer, int amount, int type, const char *file,
 
 		ret=add_lock_callback(pointer,amount,type,file,line);
 #ifdef LOCK_DEBUG
+		{
+		CRYPTO_THREADID id;
+		CRYPTO_THREADID_current(&id);
 		fprintf(stderr,"ladd:%08lx:%2d+%2d->%2d %-18s %s:%d\n",
-			CRYPTO_thread_id(),
-			before,amount,ret,
+			CRYPTO_THREADID_hash(&id), before,amount,ret,
 			CRYPTO_get_lock_name(type),
 			file,line);
+		}
 #endif
 		}
 	else
@@ -256,11 +633,15 @@ int CRYPTO_add_lock(int *pointer, int amount, int type, const char *file,
 
 		ret= *pointer+amount;
 #ifdef LOCK_DEBUG
+		{
+		CRYPTO_THREADID id;
+		CRYPTO_THREADID_current(&id);
 		fprintf(stderr,"ladd:%08lx:%2d+%2d->%2d %-18s %s:%d\n",
-			CRYPTO_thread_id(),
+			CRYPTO_THREADID_hash(&id),
 			*pointer,amount,ret,
 			CRYPTO_get_lock_name(type),
 			file,line);
+		}
 #endif
 		*pointer=ret;
 		CRYPTO_lock(CRYPTO_UNLOCK|CRYPTO_WRITE,type,file,line);
@@ -268,32 +649,68 @@ int CRYPTO_add_lock(int *pointer, int amount, int type, const char *file,
 	return(ret);
 	}
 
+const char *CRYPTO_get_lock_name(int type)
+	{
+	if (type < 0)
+		return("dynamic");
+	else if (type < CRYPTO_NUM_LOCKS)
+		return(lock_names[type]);
+	else if (type-CRYPTO_NUM_LOCKS > sk_OPENSSL_STRING_num(app_locks))
+		return("ERROR");
+	else
+		return(sk_OPENSSL_STRING_value(app_locks,type-CRYPTO_NUM_LOCKS));
+	}
+
 #if	defined(__i386)   || defined(__i386__)   || defined(_M_IX86) || \
 	defined(__INTEL__) || \
 	defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64) || defined(_M_X64)
 
-unsigned long  OPENSSL_ia32cap_P=0;
-unsigned long *OPENSSL_ia32cap_loc(void) { return &OPENSSL_ia32cap_P; }
+unsigned int  OPENSSL_ia32cap_P[2];
+unsigned long *OPENSSL_ia32cap_loc(void)
+{   if (sizeof(long)==4)
+	/*
+	 * If 32-bit application pulls address of OPENSSL_ia32cap_P[0]
+	 * clear second element to maintain the illusion that vector
+	 * is 32-bit.
+	 */
+	OPENSSL_ia32cap_P[1]=0;
+    return (unsigned long *)OPENSSL_ia32cap_P;
+}
 
 #if defined(OPENSSL_CPUID_OBJ) && !defined(OPENSSL_NO_ASM) && !defined(I386_ONLY)
 #define OPENSSL_CPUID_SETUP
+#if defined(_WIN32)
+typedef unsigned __int64 IA32CAP;
+#else
+typedef unsigned long long IA32CAP;
+#endif
 void OPENSSL_cpuid_setup(void)
 { static int trigger=0;
-  unsigned long OPENSSL_ia32_cpuid(void);
+  IA32CAP OPENSSL_ia32_cpuid(void);
+  IA32CAP vec;
   char *env;
 
     if (trigger)	return;
 
     trigger=1;
-    if ((env=getenv("OPENSSL_ia32cap")))
-	OPENSSL_ia32cap_P = strtoul(env,NULL,0)|(1<<10);
+    if ((env=getenv("OPENSSL_ia32cap"))) {
+	int off = (env[0]=='~')?1:0;
+#if defined(_WIN32)
+	if (!sscanf(env+off,"%I64i",&vec)) vec = strtoul(env+off,NULL,0);
+#else
+	if (!sscanf(env+off,"%lli",(long long *)&vec)) vec = strtoul(env+off,NULL,0);
+#endif
+	if (off) vec = OPENSSL_ia32_cpuid()&~vec;
+    }
     else
-	OPENSSL_ia32cap_P = OPENSSL_ia32_cpuid()|(1<<10);
+	vec = OPENSSL_ia32_cpuid();
     /*
      * |(1<<10) sets a reserved bit to signal that variable
      * was initialized already... This is to avoid interference
      * with cpuid snippets in ELF .init segment.
      */
+    OPENSSL_ia32cap_P[0] = (unsigned int)vec|(1<<10);
+    OPENSSL_ia32cap_P[1] = (unsigned int)(vec>>32);
 }
 #endif
 
@@ -301,70 +718,16 @@ void OPENSSL_cpuid_setup(void)
 unsigned long *OPENSSL_ia32cap_loc(void) { return NULL; }
 #endif
 int OPENSSL_NONPIC_relocated = 0;
-#if !defined(OPENSSL_CPUID_SETUP)
+#if !defined(OPENSSL_CPUID_SETUP) && !defined(OPENSSL_CPUID_OBJ)
 void OPENSSL_cpuid_setup(void) {}
 #endif
 
 #if (defined(_WIN32) || defined(__CYGWIN__)) && defined(_WINDLL)
-
-#ifdef OPENSSL_FIPS
-
-#include <tlhelp32.h>
-#if defined(__GNUC__) && __GNUC__>=2
-static int DllInit(void) __attribute__((constructor));
-#elif defined(_MSC_VER)
-static int DllInit(void);
-# ifdef _WIN64
-# pragma section(".CRT$XCU",read)
-  __declspec(allocate(".CRT$XCU"))
-# else
-# pragma data_seg(".CRT$XCU")
-# endif
-  static int (*p)(void) = DllInit;
-# pragma data_seg()
-#endif
-
-static int DllInit(void)
-{
-#if defined(_WIN32_WINNT)
-	union	{ int(*f)(void); BYTE *p; } t = { DllInit };
-        HANDLE	hModuleSnap = INVALID_HANDLE_VALUE;
-	IMAGE_DOS_HEADER *dos_header;
-	IMAGE_NT_HEADERS *nt_headers;
-	MODULEENTRY32 me32 = {sizeof(me32)};
-
-	hModuleSnap = CreateToolhelp32Snapshot(TH32CS_SNAPMODULE,0);
-	if (hModuleSnap != INVALID_HANDLE_VALUE &&
-	    Module32First(hModuleSnap,&me32)) do
-		{
-		if (t.p >= me32.modBaseAddr &&
-		    t.p <  me32.modBaseAddr+me32.modBaseSize)
-			{
-			dos_header=(IMAGE_DOS_HEADER *)me32.modBaseAddr;
-			if (dos_header->e_magic==IMAGE_DOS_SIGNATURE)
-				{
-				nt_headers=(IMAGE_NT_HEADERS *)
-					((BYTE *)dos_header+dos_header->e_lfanew);
-				if (nt_headers->Signature==IMAGE_NT_SIGNATURE &&
-				    me32.modBaseAddr!=(BYTE*)nt_headers->OptionalHeader.ImageBase)
-					OPENSSL_NONPIC_relocated=1;
-				}
-			break;
-			}
-		} while (Module32Next(hModuleSnap,&me32));
-
-	if (hModuleSnap != INVALID_HANDLE_VALUE)
-		CloseHandle(hModuleSnap);
-#endif
-	OPENSSL_cpuid_setup();
-	return 0;
-}
-
-#else
-
 #ifdef __CYGWIN__
 /* pick DLL_[PROCESS|THREAD]_[ATTACH|DETACH] definitions */
 #include <windows.h>
+/* this has side-effect of _WIN32 getting defined, which otherwise
+ * is mutually exclusive with __CYGWIN__... */
 #endif
 
 /* All we really need to do is remove the 'error' state when a thread
@@ -404,16 +767,37 @@ BOOL WINAPI DllMain(HINSTANCE hinstDLL, DWORD fdwReason,
 	}
 #endif
 
-#endif
-
 #if defined(_WIN32) && !defined(__CYGWIN__)
 #include <tchar.h>
+#include <signal.h>
+#ifdef __WATCOMC__
+#if defined(_UNICODE) || defined(__UNICODE__)
+#define _vsntprintf _vsnwprintf
+#else
+#define _vsntprintf _vsnprintf
+#endif
+#endif
+#ifdef _MSC_VER
+#define alloca _alloca
+#endif
 
 #if defined(_WIN32_WINNT) && _WIN32_WINNT>=0x0333
 int OPENSSL_isservice(void)
 { HWINSTA h;
   DWORD len;
   WCHAR *name;
+  static union { void *p; int (*f)(void); } _OPENSSL_isservice = { NULL };
+
+    if (_OPENSSL_isservice.p == NULL) {
+	HANDLE h = GetModuleHandle(NULL);
+	if (h != NULL)
+	    _OPENSSL_isservice.p = GetProcAddress(h,"_OPENSSL_isservice");
+	if (_OPENSSL_isservice.p == NULL)
+	    _OPENSSL_isservice.p = (void *)-1;
+    }
+
+    if (_OPENSSL_isservice.p != (void *)-1)
+	return (*_OPENSSL_isservice.f)();
 
     (void)GetDesktopWindow(); /* return value is ignored */
 
@@ -426,11 +810,7 @@ int OPENSSL_isservice(void)
 
     if (len>512) return -1;		/* paranoia */
     len++,len&=~1;			/* paranoia */
-#ifdef _MSC_VER
-    name=(WCHAR *)_alloca(len+sizeof(WCHAR));
-#else
     name=(WCHAR *)alloca(len+sizeof(WCHAR));
-#endif
     if (!GetUserObjectInformationW (h,UOI_NAME,name,len,&len))
 	return -1;
 
@@ -475,11 +855,7 @@ void OPENSSL_showfatal (const char *fmta,...)
       size_t len_0=strlen(fmta)+1,i;
       WCHAR *fmtw;
 
-#ifdef _MSC_VER
-	fmtw = (WCHAR *)_alloca (len_0*sizeof(WCHAR));
-#else
-	fmtw = (WCHAR *)alloca (len_0*sizeof(WCHAR));
-#endif
+	fmtw = (WCHAR *)alloca(len_0*sizeof(WCHAR));
 	if (fmtw == NULL) { fmt=(const TCHAR *)L"no stack?"; break; }
 
 #ifndef OPENSSL_NO_MULTIBYTE
@@ -538,7 +914,13 @@ void OpenSSLDie(const char *file,int line,const char *assertion)
 	OPENSSL_showfatal(
 		"%s(%d): OpenSSL internal error, assertion failed: %s\n",
 		file,line,assertion);
+#if !defined(_WIN32) || defined(__CYGWIN__)
 	abort();
+#else
+	/* Win32 abort() customarily shows a dialog, but we just did that... */
+	raise(SIGABRT);
+	_exit(3);
+#endif
 	}
 
 void *OPENSSL_stderr(void)	{ return stderr; }
diff --git a/crypto/cryptlib.h b/crypto/cryptlib.h
index fc249c57f379..1761f6b66879 100644
--- a/crypto/cryptlib.h
+++ b/crypto/cryptlib.h
@@ -99,7 +99,7 @@ extern "C" {
 #define HEX_SIZE(type)		(sizeof(type)*2)
 
 void OPENSSL_cpuid_setup(void);
-extern unsigned long OPENSSL_ia32cap_P;
+extern unsigned int OPENSSL_ia32cap_P[];
 void OPENSSL_showfatal(const char *,...);
 void *OPENSSL_stderr(void);
 extern int OPENSSL_NONPIC_relocated;
diff --git a/crypto/crypto.h b/crypto/crypto.h
index fc1374fad5f7..6aeda0a9ace3 100644
--- a/crypto/crypto.h
+++ b/crypto/crypto.h
@@ -1,6 +1,6 @@
 /* crypto/crypto.h */
 /* ====================================================================
- * Copyright (c) 1998-2003 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1998-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -219,13 +219,9 @@ typedef struct openssl_item_st
 #define CRYPTO_LOCK_EC_PRE_COMP		36
 #define CRYPTO_LOCK_STORE		37
 #define CRYPTO_LOCK_COMP		38
-#ifndef OPENSSL_FIPS
-#define CRYPTO_NUM_LOCKS		39
-#else
 #define CRYPTO_LOCK_FIPS		39
 #define CRYPTO_LOCK_FIPS2		40
 #define CRYPTO_NUM_LOCKS		41
-#endif
 
 #define CRYPTO_LOCK		1
 #define CRYPTO_UNLOCK		2
@@ -288,9 +284,10 @@ typedef struct bio_st BIO_dummy;
 
 struct crypto_ex_data_st
 	{
-	STACK *sk;
+	STACK_OF(void) *sk;
 	int dummy; /* gcc is screwing up this data structure :-( */
 	};
+DECLARE_STACK_OF(void)
 
 /* This stuff is basically class callback functions
  * The current classes are SSL_CTX, SSL, SSL_SESSION, and a few more */
@@ -347,7 +344,14 @@ DECLARE_STACK_OF(CRYPTO_EX_DATA_FUNCS)
 
 /* Set standard debugging functions (not done by default
  * unless CRYPTO_MDEBUG is defined) */
-void CRYPTO_malloc_debug_init(void);
+#define CRYPTO_malloc_debug_init()	do {\
+	CRYPTO_set_mem_debug_functions(\
+		CRYPTO_dbg_malloc,\
+		CRYPTO_dbg_realloc,\
+		CRYPTO_dbg_free,\
+		CRYPTO_dbg_set_options,\
+		CRYPTO_dbg_get_options);\
+	} while(0)
 
 int CRYPTO_mem_ctrl(int mode);
 int CRYPTO_is_mem_check_on(void);
@@ -420,16 +424,32 @@ void CRYPTO_set_add_lock_callback(int (*func)(int *num,int mount,int type,
 					      const char *file, int line));
 int (*CRYPTO_get_add_lock_callback(void))(int *num,int mount,int type,
 					  const char *file,int line);
+
+/* Don't use this structure directly. */
+typedef struct crypto_threadid_st
+	{
+	void *ptr;
+	unsigned long val;
+	} CRYPTO_THREADID;
+/* Only use CRYPTO_THREADID_set_[numeric|pointer]() within callbacks */
+void CRYPTO_THREADID_set_numeric(CRYPTO_THREADID *id, unsigned long val);
+void CRYPTO_THREADID_set_pointer(CRYPTO_THREADID *id, void *ptr);
+int CRYPTO_THREADID_set_callback(void (*threadid_func)(CRYPTO_THREADID *));
+void (*CRYPTO_THREADID_get_callback(void))(CRYPTO_THREADID *);
+void CRYPTO_THREADID_current(CRYPTO_THREADID *id);
+int CRYPTO_THREADID_cmp(const CRYPTO_THREADID *a, const CRYPTO_THREADID *b);
+void CRYPTO_THREADID_cpy(CRYPTO_THREADID *dest, const CRYPTO_THREADID *src);
+unsigned long CRYPTO_THREADID_hash(const CRYPTO_THREADID *id);
+#ifndef OPENSSL_NO_DEPRECATED
 void CRYPTO_set_id_callback(unsigned long (*func)(void));
 unsigned long (*CRYPTO_get_id_callback(void))(void);
 unsigned long CRYPTO_thread_id(void);
+#endif
+
 const char *CRYPTO_get_lock_name(int type);
 int CRYPTO_add_lock(int *pointer,int amount,int type, const char *file,
 		    int line);
 
-void int_CRYPTO_set_do_dynlock_callback(
-	void (*do_dynlock_cb)(int mode, int type, const char *file, int line));
-
 int CRYPTO_get_new_dynlockid(void);
 void CRYPTO_destroy_dynlockid(int i);
 struct CRYPTO_dynlock_value *CRYPTO_get_dynlock_value(int i);
@@ -454,10 +474,6 @@ int CRYPTO_set_mem_debug_functions(void (*m)(void *,int,const char *,int,int),
 				   void (*f)(void *,int),
 				   void (*so)(long),
 				   long (*go)(void));
-void CRYPTO_set_mem_info_functions(
-	int  (*push_info_fn)(const char *info, const char *file, int line),
-	int  (*pop_info_fn)(void),
-	int (*remove_all_info_fn)(void));
 void CRYPTO_get_mem_functions(void *(**m)(size_t),void *(**r)(void *, size_t), void (**f)(void *));
 void CRYPTO_get_locked_mem_functions(void *(**m)(size_t), void (**f)(void *));
 void CRYPTO_get_mem_ex_functions(void *(**m)(size_t,const char *,int),
@@ -514,9 +530,6 @@ void CRYPTO_dbg_free(void *addr,int before_p);
 void CRYPTO_dbg_set_options(long bits);
 long CRYPTO_dbg_get_options(void);
 
-int CRYPTO_dbg_push_info(const char *info, const char *file, int line);
-int CRYPTO_dbg_pop_info(void);
-int CRYPTO_dbg_remove_all_info(void);
 
 #ifndef OPENSSL_NO_FP_API
 void CRYPTO_mem_leaks_fp(FILE *);
@@ -534,62 +547,32 @@ unsigned long *OPENSSL_ia32cap_loc(void);
 #define OPENSSL_ia32cap (*(OPENSSL_ia32cap_loc()))
 int OPENSSL_isservice(void);
 
-#ifdef OPENSSL_FIPS
-#define FIPS_ERROR_IGNORED(alg) OpenSSLDie(__FILE__, __LINE__, \
-		alg " previous FIPS forbidden algorithm error ignored");
-
-#define FIPS_BAD_ABORT(alg) OpenSSLDie(__FILE__, __LINE__, \
-		#alg " Algorithm forbidden in FIPS mode");
+int FIPS_mode(void);
+int FIPS_mode_set(int r);
 
-#ifdef OPENSSL_FIPS_STRICT
-#define FIPS_BAD_ALGORITHM(alg) FIPS_BAD_ABORT(alg)
-#else
-#define FIPS_BAD_ALGORITHM(alg) \
-	{ \
-	FIPSerr(FIPS_F_HASH_FINAL,FIPS_R_NON_FIPS_METHOD); \
-	ERR_add_error_data(2, "Algorithm=", #alg); \
-	return 0; \
-	}
-#endif
-
-/* Low level digest API blocking macro */
+void OPENSSL_init(void);
 
-#define FIPS_NON_FIPS_MD_Init(alg) \
-	int alg##_Init(alg##_CTX *c) \
-		{ \
-		if (FIPS_mode()) \
-			FIPS_BAD_ALGORITHM(alg) \
-		return private_##alg##_Init(c); \
-		} \
-	int private_##alg##_Init(alg##_CTX *c)
+#define fips_md_init(alg) fips_md_init_ctx(alg, alg)
 
-/* For ciphers the API often varies from cipher to cipher and each needs to
- * be treated as a special case. Variable key length ciphers (Blowfish, RC4,
- * CAST) however are very similar and can use a blocking macro.
- */
+#ifdef OPENSSL_FIPS
+#define fips_md_init_ctx(alg, cx) \
+	int alg##_Init(cx##_CTX *c) \
+	{ \
+	if (FIPS_mode()) OpenSSLDie(__FILE__, __LINE__, \
+		"Low level API call to digest " #alg " forbidden in FIPS mode!"); \
+	return private_##alg##_Init(c); \
+	} \
+	int private_##alg##_Init(cx##_CTX *c)
 
-#define FIPS_NON_FIPS_VCIPHER_Init(alg) \
-	void alg##_set_key(alg##_KEY *key, int len, const unsigned char *data) \
-		{ \
-		if (FIPS_mode()) \
-			FIPS_BAD_ABORT(alg) \
-		private_##alg##_set_key(key, len, data); \
-		} \
-	void private_##alg##_set_key(alg##_KEY *key, int len, \
-					const unsigned char *data)
+#define fips_cipher_abort(alg) \
+	if (FIPS_mode()) OpenSSLDie(__FILE__, __LINE__, \
+		"Low level API call to cipher " #alg " forbidden in FIPS mode!")
 
 #else
-
-#define FIPS_NON_FIPS_VCIPHER_Init(alg) \
-	void alg##_set_key(alg##_KEY *key, int len, const unsigned char *data)
-
-#define FIPS_NON_FIPS_MD_Init(alg) \
-	int alg##_Init(alg##_CTX *c) 
-
-#endif /* def OPENSSL_FIPS */
-
-#define OPENSSL_HAVE_INIT	1
-void OPENSSL_init(void);
+#define fips_md_init_ctx(alg, cx) \
+	int alg##_Init(cx##_CTX *c)
+#define fips_cipher_abort(alg) while(0)
+#endif
 
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
@@ -606,11 +589,13 @@ void ERR_load_CRYPTO_strings(void);
 #define CRYPTO_F_CRYPTO_SET_EX_DATA			 102
 #define CRYPTO_F_DEF_ADD_INDEX				 104
 #define CRYPTO_F_DEF_GET_CLASS				 105
+#define CRYPTO_F_FIPS_MODE_SET				 109
 #define CRYPTO_F_INT_DUP_EX_DATA			 106
 #define CRYPTO_F_INT_FREE_EX_DATA			 107
 #define CRYPTO_F_INT_NEW_EX_DATA			 108
 
 /* Reason codes. */
+#define CRYPTO_R_FIPS_MODE_NOT_SUPPORTED		 101
 #define CRYPTO_R_NO_DYNLOCK_CREATE_CALLBACK		 100
 
 #ifdef  __cplusplus
diff --git a/crypto/des/Makefile b/crypto/des/Makefile
index 786e68802ea6..a6e1001329a5 100644
--- a/crypto/des/Makefile
+++ b/crypto/des/Makefile
@@ -12,8 +12,6 @@ MAKEFILE=	Makefile
 AR=		ar r
 RANLIB=		ranlib
 DES_ENC=	des_enc.o fcrypt_b.o
-# or use
-#DES_ENC=	dx86-elf.o yx86-elf.o
 
 CFLAGS= $(INCLUDES) $(CFLAG)
 ASFLAGS= $(INCLUDES) $(ASFLAG)
@@ -24,7 +22,7 @@ TEST=destest.c
 APPS=
 
 LIB=$(TOP)/libcrypto.a
-LIBSRC=	des_lib.c cbc_cksm.c cbc_enc.c  cfb64enc.c cfb_enc.c  \
+LIBSRC=	cbc_cksm.c cbc_enc.c  cfb64enc.c cfb_enc.c  \
 	ecb3_enc.c ecb_enc.c  enc_read.c enc_writ.c \
 	fcrypt.c ofb64enc.c ofb_enc.c  pcbc_enc.c \
 	qud_cksm.c rand_key.c rpc_enc.c  set_key.c  \
@@ -33,7 +31,7 @@ LIBSRC=	des_lib.c cbc_cksm.c cbc_enc.c  cfb64enc.c cfb_enc.c  \
 	str2key.c  cfb64ede.c ofb64ede.c ede_cbcm_enc.c des_old.c des_old2.c \
 	read2pwd.c
 
-LIBOBJ= des_lib.o set_key.o  ecb_enc.o  cbc_enc.o \
+LIBOBJ= set_key.o  ecb_enc.o  cbc_enc.o \
 	ecb3_enc.o cfb64enc.o cfb64ede.o cfb_enc.o  ofb64ede.o \
 	enc_read.o enc_writ.o ofb64enc.o \
 	ofb_enc.o  str2key.o  pcbc_enc.o qud_cksm.o rand_key.o \
@@ -54,7 +52,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -64,21 +62,10 @@ des: des.o cbc3_enc.o lib
 des_enc-sparc.S:	asm/des_enc.m4
 	m4 -B 8192 asm/des_enc.m4 > des_enc-sparc.S
 
-# ELF
-dx86-elf.s:	asm/des-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) des-586.pl elf $(CFLAGS) > ../$@)
-yx86-elf.s:	asm/crypt586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) crypt586.pl elf $(CFLAGS) > ../$@)
-# COFF
-dx86-cof.s: asm/des-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) des-586.pl coff $(CFLAGS) > ../$@)
-yx86-cof.s: asm/crypt586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) crypt586.pl coff $(CFLAGS) > ../$@)
-# a.out
-dx86-out.s: asm/des-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) des-586.pl a.out $(CFLAGS) > ../$@)
-yx86-out.s: asm/crypt586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) crypt586.pl a.out $(CFLAGS) > ../$@)
+des-586.s:	asm/des-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
+	$(PERL) asm/des-586.pl $(PERLASM_SCHEME) $(CFLAGS) > $@
+crypt586.s:	asm/crypt586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
+	$(PERL) asm/crypt586.pl $(PERLASM_SCHEME) $(CFLAGS) > $@
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
@@ -156,14 +143,7 @@ des_enc.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
 des_enc.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
 des_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 des_enc.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
-des_enc.o: des_enc.c des_locl.h ncbc_enc.c
-des_lib.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
-des_lib.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
-des_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-des_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-des_lib.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-des_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
-des_lib.o: ../../include/openssl/ui_compat.h des_lib.c des_locl.h des_ver.h
+des_enc.o: des_enc.c des_locl.h ncbc_enc.c spr.h
 des_old.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
 des_old.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
 des_old.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
@@ -182,12 +162,13 @@ ecb3_enc.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
 ecb3_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 ecb3_enc.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
 ecb3_enc.o: des_locl.h ecb3_enc.c
+ecb_enc.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
 ecb_enc.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
 ecb_enc.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-ecb_enc.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-ecb_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-ecb_enc.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
-ecb_enc.o: des_locl.h ecb_enc.c spr.h
+ecb_enc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ecb_enc.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+ecb_enc.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
+ecb_enc.o: ../../include/openssl/ui_compat.h des_locl.h des_ver.h ecb_enc.c
 ede_cbcm_enc.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
 ede_cbcm_enc.o: ../../include/openssl/e_os2.h
 ede_cbcm_enc.o: ../../include/openssl/opensslconf.h
@@ -276,12 +257,13 @@ rpc_enc.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
 rpc_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 rpc_enc.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
 rpc_enc.o: des_locl.h des_ver.h rpc_des.h rpc_enc.c
-set_key.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
-set_key.o: ../../include/openssl/e_os2.h ../../include/openssl/fips.h
-set_key.o: ../../include/openssl/opensslconf.h ../../include/openssl/ossl_typ.h
-set_key.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-set_key.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
-set_key.o: ../../include/openssl/ui_compat.h des_locl.h set_key.c
+set_key.o: ../../include/openssl/crypto.h ../../include/openssl/des.h
+set_key.o: ../../include/openssl/des_old.h ../../include/openssl/e_os2.h
+set_key.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+set_key.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+set_key.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+set_key.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
+set_key.o: des_locl.h set_key.c
 str2key.o: ../../include/openssl/crypto.h ../../include/openssl/des.h
 str2key.o: ../../include/openssl/des_old.h ../../include/openssl/e_os2.h
 str2key.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
diff --git a/crypto/des/asm/crypt586.pl b/crypto/des/asm/crypt586.pl
index 1d04ed6def10..e36f7d44bd7d 100644
--- a/crypto/des/asm/crypt586.pl
+++ b/crypto/des/asm/crypt586.pl
@@ -6,7 +6,8 @@
 # things perfect.
 #
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],"crypt586.pl");
@@ -22,7 +23,7 @@ sub fcrypt_body
 	{
 	local($name,$do_ip)=@_;
 
-	&function_begin($name,"EXTRN   _DES_SPtrans:DWORD");
+	&function_begin($name);
 
 	&comment("");
 	&comment("Load the 2 words");
diff --git a/crypto/des/asm/des-586.pl b/crypto/des/asm/des-586.pl
index b75d3c6b3a44..5b5f39cebd13 100644
--- a/crypto/des/asm/des-586.pl
+++ b/crypto/des/asm/des-586.pl
@@ -4,7 +4,8 @@
 # Svend Olaf Mikkelsen <svolaf@inet.uni-c.dk>
 #
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 require "cbc.pl";
 require "desboth.pl";
@@ -18,29 +19,110 @@ require "desboth.pl";
 
 $L="edi";
 $R="esi";
+$trans="ebp";
+$small_footprint=1 if (grep(/\-DOPENSSL_SMALL_FOOTPRINT/,@ARGV));
+# one can discuss setting this variable to 1 unconditionally, as
+# the folded loop is only 3% slower than unrolled, but >7 times smaller
 
-&external_label("DES_SPtrans");
+&public_label("DES_SPtrans");
+
+&DES_encrypt_internal();
+&DES_decrypt_internal();
 &DES_encrypt("DES_encrypt1",1);
 &DES_encrypt("DES_encrypt2",0);
 &DES_encrypt3("DES_encrypt3",1);
 &DES_encrypt3("DES_decrypt3",0);
 &cbc("DES_ncbc_encrypt","DES_encrypt1","DES_encrypt1",0,4,5,3,5,-1);
 &cbc("DES_ede3_cbc_encrypt","DES_encrypt3","DES_decrypt3",0,6,7,3,4,5);
+&DES_SPtrans();
 
 &asm_finish();
 
+sub DES_encrypt_internal()
+	{
+	&function_begin_B("_x86_DES_encrypt");
+
+	if ($small_footprint)
+	    {
+	    &lea("edx",&DWP(128,"ecx"));
+	    &push("edx");
+	    &push("ecx");
+	    &set_label("eloop");
+		&D_ENCRYPT(0,$L,$R,0,$trans,"eax","ebx","ecx","edx",&swtmp(0));
+		&comment("");
+		&D_ENCRYPT(1,$R,$L,2,$trans,"eax","ebx","ecx","edx",&swtmp(0));
+		&comment("");
+		&add("ecx",16);
+		&cmp("ecx",&swtmp(1));
+		&mov(&swtmp(0),"ecx");
+		&jb(&label("eloop"));
+	    &add("esp",8);
+	    }
+	else
+	    {
+	    &push("ecx");
+	    for ($i=0; $i<16; $i+=2)
+		{
+		&comment("Round $i");
+		&D_ENCRYPT($i,$L,$R,$i*2,$trans,"eax","ebx","ecx","edx",&swtmp(0));
+		&comment("Round ".sprintf("%d",$i+1));
+		&D_ENCRYPT($i+1,$R,$L,($i+1)*2,$trans,"eax","ebx","ecx","edx",&swtmp(0));
+		}
+	    &add("esp",4);
+	}
+	&ret();
+
+	&function_end_B("_x86_DES_encrypt");
+	}
+	
+sub DES_decrypt_internal()
+	{
+	&function_begin_B("_x86_DES_decrypt");
+
+	if ($small_footprint)
+	    {
+	    &push("ecx");
+	    &lea("ecx",&DWP(128,"ecx"));
+	    &push("ecx");
+	    &set_label("dloop");
+		&D_ENCRYPT(0,$L,$R,-2,$trans,"eax","ebx","ecx","edx",&swtmp(0));
+		&comment("");
+		&D_ENCRYPT(1,$R,$L,-4,$trans,"eax","ebx","ecx","edx",&swtmp(0));
+		&comment("");
+		&sub("ecx",16);
+		&cmp("ecx",&swtmp(1));
+		&mov(&swtmp(0),"ecx");
+		&ja(&label("dloop"));
+	    &add("esp",8);
+	    }
+	else
+	    {
+	    &push("ecx");
+	    for ($i=15; $i>0; $i-=2)
+		{
+		&comment("Round $i");
+		&D_ENCRYPT(15-$i,$L,$R,$i*2,$trans,"eax","ebx","ecx","edx",&swtmp(0));
+		&comment("Round ".sprintf("%d",$i-1));
+		&D_ENCRYPT(15-$i+1,$R,$L,($i-1)*2,$trans,"eax","ebx","ecx","edx",&swtmp(0));
+		}
+	    &add("esp",4);
+	    }
+	&ret();
+
+	&function_end_B("_x86_DES_decrypt");
+	}
+	
 sub DES_encrypt
 	{
 	local($name,$do_ip)=@_;
 
-	&function_begin_B($name,"EXTRN   _DES_SPtrans:DWORD");
+	&function_begin_B($name);
 
 	&push("esi");
 	&push("edi");
 
 	&comment("");
 	&comment("Load the 2 words");
-	$trans="ebp";
 
 	if ($do_ip)
 		{
@@ -73,39 +155,20 @@ sub DES_encrypt
 		}
 
 	# PIC-ification:-)
-	&picmeup($trans,"DES_SPtrans");
-	#if ($cpp)	{ &picmeup($trans,"DES_SPtrans");   }
-	#else		{ &lea($trans,&DWP("DES_SPtrans")); }
+	&call	(&label("pic_point"));
+	&set_label("pic_point");
+	&blindpop($trans);
+	&lea	($trans,&DWP(&label("DES_SPtrans")."-".&label("pic_point"),$trans));
 
 	&mov(	"ecx",	&wparam(1)	);
-	&cmp("ebx","0");
-	&je(&label("start_decrypt"));
-
-	for ($i=0; $i<16; $i+=2)
-		{
-		&comment("");
-		&comment("Round $i");
-		&D_ENCRYPT($i,$L,$R,$i*2,$trans,"eax","ebx","ecx","edx");
-
-		&comment("");
-		&comment("Round ".sprintf("%d",$i+1));
-		&D_ENCRYPT($i+1,$R,$L,($i+1)*2,$trans,"eax","ebx","ecx","edx");
-		}
-	&jmp(&label("end"));
 
-	&set_label("start_decrypt");
-
-	for ($i=15; $i>0; $i-=2)
-		{
-		&comment("");
-		&comment("Round $i");
-		&D_ENCRYPT(15-$i,$L,$R,$i*2,$trans,"eax","ebx","ecx","edx");
-		&comment("");
-		&comment("Round ".sprintf("%d",$i-1));
-		&D_ENCRYPT(15-$i+1,$R,$L,($i-1)*2,$trans,"eax","ebx","ecx","edx");
-		}
-
-	&set_label("end");
+	&cmp("ebx","0");
+	&je(&label("decrypt"));
+	&call("_x86_DES_encrypt");
+	&jmp(&label("done"));
+	&set_label("decrypt");
+	&call("_x86_DES_decrypt");
+	&set_label("done");
 
 	if ($do_ip)
 		{
@@ -139,7 +202,7 @@ sub DES_encrypt
 
 sub D_ENCRYPT
 	{
-	local($r,$L,$R,$S,$trans,$u,$tmp1,$tmp2,$t)=@_;
+	local($r,$L,$R,$S,$trans,$u,$tmp1,$tmp2,$t,$wp1)=@_;
 
 	 &mov(	$u,		&DWP(&n2a($S*4),$tmp2,"",0));
 	&xor(	$tmp1,		$tmp1);
@@ -166,7 +229,7 @@ sub D_ENCRYPT
 	&and(	$t,		"0xff"	);
 	 &xor(	$L,		&DWP("0x600",$trans,$tmp1,0));
 	 &xor(	$L,		&DWP("0x700",$trans,$tmp2,0));
-	&mov(	$tmp2,		&wparam(1)	);
+	&mov(	$tmp2,		$wp1	);
 	 &xor(	$L,		&DWP("0x400",$trans,$u,0));
 	 &xor(	$L,		&DWP("0x500",$trans,$t,0));
 	}
@@ -249,3 +312,142 @@ sub FP_new
 	&rotr($tt	, 4);
 	}
 
+sub DES_SPtrans
+	{
+	&set_label("DES_SPtrans",64);
+	&data_word(0x02080800, 0x00080000, 0x02000002, 0x02080802);
+	&data_word(0x02000000, 0x00080802, 0x00080002, 0x02000002);
+	&data_word(0x00080802, 0x02080800, 0x02080000, 0x00000802);
+	&data_word(0x02000802, 0x02000000, 0x00000000, 0x00080002);
+	&data_word(0x00080000, 0x00000002, 0x02000800, 0x00080800);
+	&data_word(0x02080802, 0x02080000, 0x00000802, 0x02000800);
+	&data_word(0x00000002, 0x00000800, 0x00080800, 0x02080002);
+	&data_word(0x00000800, 0x02000802, 0x02080002, 0x00000000);
+	&data_word(0x00000000, 0x02080802, 0x02000800, 0x00080002);
+	&data_word(0x02080800, 0x00080000, 0x00000802, 0x02000800);
+	&data_word(0x02080002, 0x00000800, 0x00080800, 0x02000002);
+	&data_word(0x00080802, 0x00000002, 0x02000002, 0x02080000);
+	&data_word(0x02080802, 0x00080800, 0x02080000, 0x02000802);
+	&data_word(0x02000000, 0x00000802, 0x00080002, 0x00000000);
+	&data_word(0x00080000, 0x02000000, 0x02000802, 0x02080800);
+	&data_word(0x00000002, 0x02080002, 0x00000800, 0x00080802);
+	# nibble 1
+	&data_word(0x40108010, 0x00000000, 0x00108000, 0x40100000);
+	&data_word(0x40000010, 0x00008010, 0x40008000, 0x00108000);
+	&data_word(0x00008000, 0x40100010, 0x00000010, 0x40008000);
+	&data_word(0x00100010, 0x40108000, 0x40100000, 0x00000010);
+	&data_word(0x00100000, 0x40008010, 0x40100010, 0x00008000);
+	&data_word(0x00108010, 0x40000000, 0x00000000, 0x00100010);
+	&data_word(0x40008010, 0x00108010, 0x40108000, 0x40000010);
+	&data_word(0x40000000, 0x00100000, 0x00008010, 0x40108010);
+	&data_word(0x00100010, 0x40108000, 0x40008000, 0x00108010);
+	&data_word(0x40108010, 0x00100010, 0x40000010, 0x00000000);
+	&data_word(0x40000000, 0x00008010, 0x00100000, 0x40100010);
+	&data_word(0x00008000, 0x40000000, 0x00108010, 0x40008010);
+	&data_word(0x40108000, 0x00008000, 0x00000000, 0x40000010);
+	&data_word(0x00000010, 0x40108010, 0x00108000, 0x40100000);
+	&data_word(0x40100010, 0x00100000, 0x00008010, 0x40008000);
+	&data_word(0x40008010, 0x00000010, 0x40100000, 0x00108000);
+	# nibble 2
+	&data_word(0x04000001, 0x04040100, 0x00000100, 0x04000101);
+	&data_word(0x00040001, 0x04000000, 0x04000101, 0x00040100);
+	&data_word(0x04000100, 0x00040000, 0x04040000, 0x00000001);
+	&data_word(0x04040101, 0x00000101, 0x00000001, 0x04040001);
+	&data_word(0x00000000, 0x00040001, 0x04040100, 0x00000100);
+	&data_word(0x00000101, 0x04040101, 0x00040000, 0x04000001);
+	&data_word(0x04040001, 0x04000100, 0x00040101, 0x04040000);
+	&data_word(0x00040100, 0x00000000, 0x04000000, 0x00040101);
+	&data_word(0x04040100, 0x00000100, 0x00000001, 0x00040000);
+	&data_word(0x00000101, 0x00040001, 0x04040000, 0x04000101);
+	&data_word(0x00000000, 0x04040100, 0x00040100, 0x04040001);
+	&data_word(0x00040001, 0x04000000, 0x04040101, 0x00000001);
+	&data_word(0x00040101, 0x04000001, 0x04000000, 0x04040101);
+	&data_word(0x00040000, 0x04000100, 0x04000101, 0x00040100);
+	&data_word(0x04000100, 0x00000000, 0x04040001, 0x00000101);
+	&data_word(0x04000001, 0x00040101, 0x00000100, 0x04040000);
+	# nibble 3
+	&data_word(0x00401008, 0x10001000, 0x00000008, 0x10401008);
+	&data_word(0x00000000, 0x10400000, 0x10001008, 0x00400008);
+	&data_word(0x10401000, 0x10000008, 0x10000000, 0x00001008);
+	&data_word(0x10000008, 0x00401008, 0x00400000, 0x10000000);
+	&data_word(0x10400008, 0x00401000, 0x00001000, 0x00000008);
+	&data_word(0x00401000, 0x10001008, 0x10400000, 0x00001000);
+	&data_word(0x00001008, 0x00000000, 0x00400008, 0x10401000);
+	&data_word(0x10001000, 0x10400008, 0x10401008, 0x00400000);
+	&data_word(0x10400008, 0x00001008, 0x00400000, 0x10000008);
+	&data_word(0x00401000, 0x10001000, 0x00000008, 0x10400000);
+	&data_word(0x10001008, 0x00000000, 0x00001000, 0x00400008);
+	&data_word(0x00000000, 0x10400008, 0x10401000, 0x00001000);
+	&data_word(0x10000000, 0x10401008, 0x00401008, 0x00400000);
+	&data_word(0x10401008, 0x00000008, 0x10001000, 0x00401008);
+	&data_word(0x00400008, 0x00401000, 0x10400000, 0x10001008);
+	&data_word(0x00001008, 0x10000000, 0x10000008, 0x10401000);
+	# nibble 4
+	&data_word(0x08000000, 0x00010000, 0x00000400, 0x08010420);
+	&data_word(0x08010020, 0x08000400, 0x00010420, 0x08010000);
+	&data_word(0x00010000, 0x00000020, 0x08000020, 0x00010400);
+	&data_word(0x08000420, 0x08010020, 0x08010400, 0x00000000);
+	&data_word(0x00010400, 0x08000000, 0x00010020, 0x00000420);
+	&data_word(0x08000400, 0x00010420, 0x00000000, 0x08000020);
+	&data_word(0x00000020, 0x08000420, 0x08010420, 0x00010020);
+	&data_word(0x08010000, 0x00000400, 0x00000420, 0x08010400);
+	&data_word(0x08010400, 0x08000420, 0x00010020, 0x08010000);
+	&data_word(0x00010000, 0x00000020, 0x08000020, 0x08000400);
+	&data_word(0x08000000, 0x00010400, 0x08010420, 0x00000000);
+	&data_word(0x00010420, 0x08000000, 0x00000400, 0x00010020);
+	&data_word(0x08000420, 0x00000400, 0x00000000, 0x08010420);
+	&data_word(0x08010020, 0x08010400, 0x00000420, 0x00010000);
+	&data_word(0x00010400, 0x08010020, 0x08000400, 0x00000420);
+	&data_word(0x00000020, 0x00010420, 0x08010000, 0x08000020);
+	# nibble 5
+	&data_word(0x80000040, 0x00200040, 0x00000000, 0x80202000);
+	&data_word(0x00200040, 0x00002000, 0x80002040, 0x00200000);
+	&data_word(0x00002040, 0x80202040, 0x00202000, 0x80000000);
+	&data_word(0x80002000, 0x80000040, 0x80200000, 0x00202040);
+	&data_word(0x00200000, 0x80002040, 0x80200040, 0x00000000);
+	&data_word(0x00002000, 0x00000040, 0x80202000, 0x80200040);
+	&data_word(0x80202040, 0x80200000, 0x80000000, 0x00002040);
+	&data_word(0x00000040, 0x00202000, 0x00202040, 0x80002000);
+	&data_word(0x00002040, 0x80000000, 0x80002000, 0x00202040);
+	&data_word(0x80202000, 0x00200040, 0x00000000, 0x80002000);
+	&data_word(0x80000000, 0x00002000, 0x80200040, 0x00200000);
+	&data_word(0x00200040, 0x80202040, 0x00202000, 0x00000040);
+	&data_word(0x80202040, 0x00202000, 0x00200000, 0x80002040);
+	&data_word(0x80000040, 0x80200000, 0x00202040, 0x00000000);
+	&data_word(0x00002000, 0x80000040, 0x80002040, 0x80202000);
+	&data_word(0x80200000, 0x00002040, 0x00000040, 0x80200040);
+	# nibble 6
+	&data_word(0x00004000, 0x00000200, 0x01000200, 0x01000004);
+	&data_word(0x01004204, 0x00004004, 0x00004200, 0x00000000);
+	&data_word(0x01000000, 0x01000204, 0x00000204, 0x01004000);
+	&data_word(0x00000004, 0x01004200, 0x01004000, 0x00000204);
+	&data_word(0x01000204, 0x00004000, 0x00004004, 0x01004204);
+	&data_word(0x00000000, 0x01000200, 0x01000004, 0x00004200);
+	&data_word(0x01004004, 0x00004204, 0x01004200, 0x00000004);
+	&data_word(0x00004204, 0x01004004, 0x00000200, 0x01000000);
+	&data_word(0x00004204, 0x01004000, 0x01004004, 0x00000204);
+	&data_word(0x00004000, 0x00000200, 0x01000000, 0x01004004);
+	&data_word(0x01000204, 0x00004204, 0x00004200, 0x00000000);
+	&data_word(0x00000200, 0x01000004, 0x00000004, 0x01000200);
+	&data_word(0x00000000, 0x01000204, 0x01000200, 0x00004200);
+	&data_word(0x00000204, 0x00004000, 0x01004204, 0x01000000);
+	&data_word(0x01004200, 0x00000004, 0x00004004, 0x01004204);
+	&data_word(0x01000004, 0x01004200, 0x01004000, 0x00004004);
+	# nibble 7
+	&data_word(0x20800080, 0x20820000, 0x00020080, 0x00000000);
+	&data_word(0x20020000, 0x00800080, 0x20800000, 0x20820080);
+	&data_word(0x00000080, 0x20000000, 0x00820000, 0x00020080);
+	&data_word(0x00820080, 0x20020080, 0x20000080, 0x20800000);
+	&data_word(0x00020000, 0x00820080, 0x00800080, 0x20020000);
+	&data_word(0x20820080, 0x20000080, 0x00000000, 0x00820000);
+	&data_word(0x20000000, 0x00800000, 0x20020080, 0x20800080);
+	&data_word(0x00800000, 0x00020000, 0x20820000, 0x00000080);
+	&data_word(0x00800000, 0x00020000, 0x20000080, 0x20820080);
+	&data_word(0x00020080, 0x20000000, 0x00000000, 0x00820000);
+	&data_word(0x20800080, 0x20020080, 0x20020000, 0x00800080);
+	&data_word(0x20820000, 0x00000080, 0x00800080, 0x20020000);
+	&data_word(0x20820080, 0x00800000, 0x20800000, 0x20000080);
+	&data_word(0x00820000, 0x00020080, 0x20020080, 0x20800000);
+	&data_word(0x00000080, 0x20820000, 0x00820080, 0x00000000);
+	&data_word(0x20000000, 0x20800080, 0x00020000, 0x00820080);
+	}
diff --git a/crypto/des/asm/des686.pl b/crypto/des/asm/des686.pl
deleted file mode 100644
index d3ad5d5edd5f..000000000000
--- a/crypto/des/asm/des686.pl
+++ /dev/null
@@ -1,230 +0,0 @@
-#!/usr/local/bin/perl
-
-$prog="des686.pl";
-
-# base code is in microsft
-# op dest, source
-# format.
-#
-
-# WILL NOT WORK ANYMORE WITH desboth.pl
-require "desboth.pl";
-
-if (	($ARGV[0] eq "elf"))
-	{ require "x86unix.pl"; }
-elsif (	($ARGV[0] eq "a.out"))
-	{ $aout=1; require "x86unix.pl"; }
-elsif (	($ARGV[0] eq "sol"))
-	{ $sol=1; require "x86unix.pl"; }
-elsif ( ($ARGV[0] eq "cpp"))
-	{ $cpp=1; require "x86unix.pl"; }
-elsif (	($ARGV[0] eq "win32"))
-	{ require "x86ms.pl"; }
-else
-	{
-	print STDERR <<"EOF";
-Pick one target type from
-	elf	- linux, FreeBSD etc
-	a.out	- old linux
-	sol	- x86 solaris
-	cpp	- format so x86unix.cpp can be used
-	win32	- Windows 95/Windows NT
-EOF
-	exit(1);
-	}
-
-&comment("Don't even think of reading this code");
-&comment("It was automatically generated by $prog");
-&comment("Which is a perl program used to generate the x86 assember for");
-&comment("any of elf, a.out, Win32, or Solaris");
-&comment("It can be found in SSLeay 0.6.5+ or in libdes 3.26+");
-&comment("eric <eay\@cryptsoft.com>");
-&comment("");
-
-&file("dx86xxxx");
-
-$L="edi";
-$R="esi";
-
-&DES_encrypt("DES_encrypt1",1);
-&DES_encrypt("DES_encrypt2",0);
-
-&DES_encrypt3("DES_encrypt3",1);
-&DES_encrypt3("DES_decrypt3",0);
-
-&file_end();
-
-sub DES_encrypt
-	{
-	local($name,$do_ip)=@_;
-
-	&function_begin($name,"EXTRN   _DES_SPtrans:DWORD");
-
-	&comment("");
-	&comment("Load the 2 words");
-	&mov("eax",&wparam(0));
-	&mov($L,&DWP(0,"eax","",0));
-	&mov($R,&DWP(4,"eax","",0));
-
-	$ksp=&wparam(1);
-
-	if ($do_ip)
-		{
-		&comment("");
-		&comment("IP");
-		&IP_new($L,$R,"eax");
-		}
-
-	&comment("");
-	&comment("fixup rotate");
-	&rotl($R,3);
-	&rotl($L,3);
-	&exch($L,$R);
-
-	&comment("");
-	&comment("load counter, key_schedule and enc flag");
-	&mov("eax",&wparam(2));	# get encrypt flag
-	&mov("ebp",&wparam(1));	# get ks
-	&cmp("eax","0");
-	&je(&label("start_decrypt"));
-
-	# encrypting part
-
-	for ($i=0; $i<16; $i+=2)
-		{
-		&comment("");
-		&comment("Round $i");
-		&D_ENCRYPT($L,$R,$i*2,"ebp","DES_SPtrans","ecx","edx","eax","ebx");
-
-		&comment("");
-		&comment("Round ".sprintf("%d",$i+1));
-		&D_ENCRYPT($R,$L,($i+1)*2,"ebp","DES_SPtrans","ecx","edx","eax","ebx");
-		}
-	&jmp(&label("end"));
-
-	&set_label("start_decrypt");
-
-	for ($i=15; $i>0; $i-=2)
-		{
-		&comment("");
-		&comment("Round $i");
-		&D_ENCRYPT($L,$R,$i*2,"ebp","DES_SPtrans","ecx","edx","eax","ebx");
-		&comment("");
-		&comment("Round ".sprintf("%d",$i-1));
-		&D_ENCRYPT($R,$L,($i-1)*2,"ebp","DES_SPtrans","ecx","edx","eax","ebx");
-		}
-
-	&set_label("end");
-
-	&comment("");
-	&comment("Fixup");
-	&rotr($L,3);		# r
-	&rotr($R,3);		# l
-
-	if ($do_ip)
-		{
-		&comment("");
-		&comment("FP");
-		&FP_new($R,$L,"eax");
-		}
-
-	&mov("eax",&wparam(0));
-	&mov(&DWP(0,"eax","",0),$L);
-	&mov(&DWP(4,"eax","",0),$R);
-
-	&function_end($name);
-	}
-
-
-# The logic is to load R into 2 registers and operate on both at the same time.
-# We also load the 2 R's into 2 more registers so we can do the 'move word down a byte'
-# while also masking the other copy and doing a lookup.  We then also accumulate the
-# L value in 2 registers then combine them at the end.
-sub D_ENCRYPT
-	{
-	local($L,$R,$S,$ks,$desSP,$u,$t,$tmp1,$tmp2,$tmp3)=@_;
-
-	&mov(	$u,		&DWP(&n2a($S*4),$ks,"",0));
-	&mov(	$t,		&DWP(&n2a(($S+1)*4),$ks,"",0));
-	&xor(	$u,		$R		);
-	&xor(	$t,		$R		);
-	&rotr(	$t,		4		);
-
-	# the numbers at the end of the line are origional instruction order
-	&mov(	$tmp2,		$u		);			# 1 2
-	&mov(	$tmp1,		$t		);			# 1 1
-	&and(	$tmp2,		"0xfc"		);			# 1 4
-	&and(	$tmp1,		"0xfc"		);			# 1 3
-	&shr(	$t,		8		);			# 1 5
-	&xor(	$L,		&DWP("0x100+$desSP",$tmp1,"",0));	# 1 7
-	&shr(	$u,		8		);			# 1 6
-	&mov(	$tmp1,		&DWP("      $desSP",$tmp2,"",0));	# 1 8
-
-	&mov(	$tmp2,		$u		);			# 2 2
-	&xor(	$L,		$tmp1		);			# 1 9
-	&and(	$tmp2,		"0xfc"		);			# 2 4
-	&mov(	$tmp1,		$t		);			# 2 1
-	&and(	$tmp1,		"0xfc"		);			# 2 3
-	&shr(	$t,		8		);			# 2 5
-	&xor(	$L,		&DWP("0x300+$desSP",$tmp1,"",0));	# 2 7
-	&shr(	$u,		8		);			# 2 6
-	&mov(	$tmp1,		&DWP("0x200+$desSP",$tmp2,"",0));	# 2 8
-	&mov(	$tmp2,		$u		);			# 3 2
-
-	&xor(	$L,		$tmp1		);			# 2 9
-	&and(	$tmp2,		"0xfc"		);			# 3 4
-
-	&mov(	$tmp1,		$t		);			# 3 1 
-	&shr(	$u,		8		);			# 3 6
-	&and(	$tmp1,		"0xfc"		);			# 3 3
-	&shr(	$t,		8		);			# 3 5
-	&xor(	$L,		&DWP("0x500+$desSP",$tmp1,"",0));	# 3 7
-	&mov(	$tmp1,		&DWP("0x400+$desSP",$tmp2,"",0));	# 3 8
-
-	&and(	$t,		"0xfc"		);			# 4 1
-	&xor(	$L,		$tmp1		);			# 3 9
-
-	&and(	$u,		"0xfc"		);			# 4 2
-	&xor(	$L,		&DWP("0x700+$desSP",$t,"",0));		# 4 3
-	&xor(	$L,		&DWP("0x600+$desSP",$u,"",0));		# 4 4
-	}
-
-sub PERM_OP
-	{
-	local($a,$b,$tt,$shift,$mask)=@_;
-
-	&mov(	$tt,		$a		);
-	&shr(	$tt,		$shift		);
-	&xor(	$tt,		$b		);
-	&and(	$tt,		$mask		);
-	&xor(	$b,		$tt		);
-	&shl(	$tt,		$shift		);
-	&xor(	$a,		$tt		);
-	}
-
-sub IP_new
-	{
-	local($l,$r,$tt)=@_;
-
-	&PERM_OP($r,$l,$tt, 4,"0x0f0f0f0f");
-	&PERM_OP($l,$r,$tt,16,"0x0000ffff");
-	&PERM_OP($r,$l,$tt, 2,"0x33333333");
-	&PERM_OP($l,$r,$tt, 8,"0x00ff00ff");
-	&PERM_OP($r,$l,$tt, 1,"0x55555555");
-	}
-
-sub FP_new
-	{
-	local($l,$r,$tt)=@_;
-
-	&PERM_OP($l,$r,$tt, 1,"0x55555555");
-        &PERM_OP($r,$l,$tt, 8,"0x00ff00ff");
-        &PERM_OP($l,$r,$tt, 2,"0x33333333");
-        &PERM_OP($r,$l,$tt,16,"0x0000ffff");
-        &PERM_OP($l,$r,$tt, 4,"0x0f0f0f0f");
-	}
-
-sub n2a
-	{
-	sprintf("%d",$_[0]);
-	}
diff --git a/crypto/des/asm/des_enc.m4 b/crypto/des/asm/des_enc.m4
index f59333a03085..328059547820 100644
--- a/crypto/des/asm/des_enc.m4
+++ b/crypto/des/asm/des_enc.m4
@@ -1954,9 +1954,11 @@ DES_ede3_cbc_encrypt:
 	.word	LOOPS                     ! 280
 	.word	0x0000FC00                ! 284
 
-	.type	.PIC.DES_SPtrans,#object
-	.size	.PIC.DES_SPtrans,2048
+	.global	DES_SPtrans
+	.type	DES_SPtrans,#object
+	.size	DES_SPtrans,2048
 .align	64
+DES_SPtrans:
 .PIC.DES_SPtrans:
 	! nibble 0
 	.word	0x02080800, 0x00080000, 0x02000002, 0x02080802
diff --git a/crypto/des/des.h b/crypto/des/des.h
index 92b666359984..1eaedcbd24b7 100644
--- a/crypto/des/des.h
+++ b/crypto/des/des.h
@@ -224,6 +224,9 @@ int DES_set_key(const_DES_cblock *key,DES_key_schedule *schedule);
 int DES_key_sched(const_DES_cblock *key,DES_key_schedule *schedule);
 int DES_set_key_checked(const_DES_cblock *key,DES_key_schedule *schedule);
 void DES_set_key_unchecked(const_DES_cblock *key,DES_key_schedule *schedule);
+#ifdef OPENSSL_FIPS
+void private_DES_set_key_unchecked(const_DES_cblock *key,DES_key_schedule *schedule);
+#endif
 void DES_string_to_key(const char *str,DES_cblock *key);
 void DES_string_to_2keys(const char *str,DES_cblock *key1,DES_cblock *key2);
 void DES_cfb64_encrypt(const unsigned char *in,unsigned char *out,long length,
diff --git a/crypto/des/des3s.cpp b/crypto/des/des3s.cpp
deleted file mode 100644
index 02d527c057c2..000000000000
--- a/crypto/des/des3s.cpp
+++ /dev/null
@@ -1,67 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/des.h>
-
-void main(int argc,char *argv[])
-	{
-	des_key_schedule key1,key2,key3;
-	unsigned long s1,s2,e1,e2;
-	unsigned long data[2];
-	int i,j;
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<1000; i++) /**/
-			{
-			des_encrypt3(&data[0],key1,key2,key3);
-			GetTSC(s1);
-			des_encrypt3(&data[0],key1,key2,key3);
-			des_encrypt3(&data[0],key1,key2,key3);
-			des_encrypt3(&data[0],key1,key2,key3);
-			GetTSC(e1);
-			GetTSC(s2);
-			des_encrypt3(&data[0],key1,key2,key3);
-			des_encrypt3(&data[0],key1,key2,key3);
-			des_encrypt3(&data[0],key1,key2,key3);
-			des_encrypt3(&data[0],key1,key2,key3);
-			GetTSC(e2);
-			des_encrypt3(&data[0],key1,key2,key3);
-			}
-
-		printf("des %d %d (%d)\n",
-			e1-s1,e2-s2,((e2-s2)-(e1-s1)));
-		}
-	}
-
diff --git a/crypto/des/des_enc.c b/crypto/des/des_enc.c
index cf71965acae0..828feba208af 100644
--- a/crypto/des/des_enc.c
+++ b/crypto/des/des_enc.c
@@ -57,6 +57,7 @@
  */
 
 #include "des_locl.h"
+#include "spr.h"
 
 void DES_encrypt1(DES_LONG *data, DES_key_schedule *ks, int enc)
 	{
@@ -107,12 +108,10 @@ void DES_encrypt1(DES_LONG *data, DES_key_schedule *ks, int enc)
 		D_ENCRYPT(l,r,28); /*  15 */
 		D_ENCRYPT(r,l,30); /*  16 */
 #else
-		for (i=0; i<32; i+=8)
+		for (i=0; i<32; i+=4)
 			{
 			D_ENCRYPT(l,r,i+0); /*  1 */
 			D_ENCRYPT(r,l,i+2); /*  2 */
-			D_ENCRYPT(l,r,i+4); /*  3 */
-			D_ENCRYPT(r,l,i+6); /*  4 */
 			}
 #endif
 		}
@@ -136,12 +135,10 @@ void DES_encrypt1(DES_LONG *data, DES_key_schedule *ks, int enc)
 		D_ENCRYPT(l,r, 2); /*  2 */
 		D_ENCRYPT(r,l, 0); /*  1 */
 #else
-		for (i=30; i>0; i-=8)
+		for (i=30; i>0; i-=4)
 			{
 			D_ENCRYPT(l,r,i-0); /* 16 */
 			D_ENCRYPT(r,l,i-2); /* 15 */
-			D_ENCRYPT(l,r,i-4); /* 14 */
-			D_ENCRYPT(r,l,i-6); /* 13 */
 			}
 #endif
 		}
@@ -203,12 +200,10 @@ void DES_encrypt2(DES_LONG *data, DES_key_schedule *ks, int enc)
 		D_ENCRYPT(l,r,28); /*  15 */
 		D_ENCRYPT(r,l,30); /*  16 */
 #else
-		for (i=0; i<32; i+=8)
+		for (i=0; i<32; i+=4)
 			{
 			D_ENCRYPT(l,r,i+0); /*  1 */
 			D_ENCRYPT(r,l,i+2); /*  2 */
-			D_ENCRYPT(l,r,i+4); /*  3 */
-			D_ENCRYPT(r,l,i+6); /*  4 */
 			}
 #endif
 		}
@@ -232,12 +227,10 @@ void DES_encrypt2(DES_LONG *data, DES_key_schedule *ks, int enc)
 		D_ENCRYPT(l,r, 2); /*  2 */
 		D_ENCRYPT(r,l, 0); /*  1 */
 #else
-		for (i=30; i>0; i-=8)
+		for (i=30; i>0; i-=4)
 			{
 			D_ENCRYPT(l,r,i-0); /* 16 */
 			D_ENCRYPT(r,l,i-2); /* 15 */
-			D_ENCRYPT(l,r,i-4); /* 14 */
-			D_ENCRYPT(r,l,i-6); /* 13 */
 			}
 #endif
 		}
@@ -289,8 +282,6 @@ void DES_decrypt3(DES_LONG *data, DES_key_schedule *ks1,
 
 #ifndef DES_DEFAULT_OPTIONS
 
-#if !defined(OPENSSL_FIPS_DES_ASM)
-
 #undef CBC_ENC_C__DONT_UPDATE_IV
 #include "ncbc_enc.c" /* DES_ncbc_encrypt */
 
@@ -406,6 +397,4 @@ void DES_ede3_cbc_encrypt(const unsigned char *input, unsigned char *output,
 	tin[0]=tin[1]=0;
 	}
 
-#endif
-
 #endif /* DES_DEFAULT_OPTIONS */
diff --git a/crypto/des/des_lib.c b/crypto/des/des_lib.c
deleted file mode 100644
index d4b304793272..000000000000
--- a/crypto/des/des_lib.c
+++ /dev/null
@@ -1,106 +0,0 @@
-/* crypto/des/ecb_enc.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#include "des_locl.h"
-#include "des_ver.h"
-#include <openssl/opensslv.h>
-#include <openssl/bio.h>
-
-OPENSSL_GLOBAL const char libdes_version[]="libdes" OPENSSL_VERSION_PTEXT;
-OPENSSL_GLOBAL const char DES_version[]="DES" OPENSSL_VERSION_PTEXT;
-
-const char *DES_options(void)
-	{
-	static int init=1;
-	static char buf[32];
-
-	if (init)
-		{
-		const char *ptr,*unroll,*risc,*size;
-
-#ifdef DES_PTR
-		ptr="ptr";
-#else
-		ptr="idx";
-#endif
-#if defined(DES_RISC1) || defined(DES_RISC2)
-#ifdef DES_RISC1
-		risc="risc1";
-#endif
-#ifdef DES_RISC2
-		risc="risc2";
-#endif
-#else
-		risc="cisc";
-#endif
-#ifdef DES_UNROLL
-		unroll="16";
-#else
-		unroll="4";
-#endif
-		if (sizeof(DES_LONG) != sizeof(long))
-			size="int";
-		else
-			size="long";
-		BIO_snprintf(buf,sizeof buf,"des(%s,%s,%s,%s)",ptr,risc,unroll,
-			     size);
-		init=0;
-		}
-	return(buf);
-	}
-
diff --git a/crypto/des/des_locl.h b/crypto/des/des_locl.h
index 4b9ecff23391..a3b512e9b040 100644
--- a/crypto/des/des_locl.h
+++ b/crypto/des/des_locl.h
@@ -61,7 +61,7 @@
 
 #include <openssl/e_os2.h>
 
-#if defined(OPENSSL_SYS_WIN32) || defined(OPENSSL_SYS_WIN16)
+#if defined(OPENSSL_SYS_WIN32)
 #ifndef OPENSSL_SYS_MSDOS
 #define OPENSSL_SYS_MSDOS
 #endif
@@ -425,4 +425,8 @@ extern const DES_LONG DES_SPtrans[8][64];
 
 void fcrypt_body(DES_LONG *out,DES_key_schedule *ks,
 		 DES_LONG Eswap0, DES_LONG Eswap1);
+
+#ifdef OPENSSL_SMALL_FOOTPRINT
+#undef DES_UNROLL
+#endif
 #endif
diff --git a/crypto/des/dess.cpp b/crypto/des/dess.cpp
deleted file mode 100644
index 5549bab90afb..000000000000
--- a/crypto/des/dess.cpp
+++ /dev/null
@@ -1,67 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/des.h>
-
-void main(int argc,char *argv[])
-	{
-	des_key_schedule key;
-	unsigned long s1,s2,e1,e2;
-	unsigned long data[2];
-	int i,j;
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<1000; i++) /**/
-			{
-			des_encrypt1(&data[0],key,1);
-			GetTSC(s1);
-			des_encrypt1(&data[0],key,1);
-			des_encrypt1(&data[0],key,1);
-			des_encrypt1(&data[0],key,1);
-			GetTSC(e1);
-			GetTSC(s2);
-			des_encrypt1(&data[0],key,1);
-			des_encrypt1(&data[0],key,1);
-			des_encrypt1(&data[0],key,1);
-			des_encrypt1(&data[0],key,1);
-			GetTSC(e2);
-			des_encrypt1(&data[0],key,1);
-			}
-
-		printf("des %d %d (%d)\n",
-			e1-s1,e2-s2,((e2-s2)-(e1-s1)));
-		}
-	}
-
diff --git a/crypto/des/ecb_enc.c b/crypto/des/ecb_enc.c
index 75ae6cf8bb6c..0684e769b3e0 100644
--- a/crypto/des/ecb_enc.c
+++ b/crypto/des/ecb_enc.c
@@ -57,7 +57,53 @@
  */
 
 #include "des_locl.h"
-#include "spr.h"
+#include "des_ver.h"
+#include <openssl/opensslv.h>
+#include <openssl/bio.h>
+
+OPENSSL_GLOBAL const char libdes_version[]="libdes" OPENSSL_VERSION_PTEXT;
+OPENSSL_GLOBAL const char DES_version[]="DES" OPENSSL_VERSION_PTEXT;
+
+const char *DES_options(void)
+	{
+	static int init=1;
+	static char buf[32];
+
+	if (init)
+		{
+		const char *ptr,*unroll,*risc,*size;
+
+#ifdef DES_PTR
+		ptr="ptr";
+#else
+		ptr="idx";
+#endif
+#if defined(DES_RISC1) || defined(DES_RISC2)
+#ifdef DES_RISC1
+		risc="risc1";
+#endif
+#ifdef DES_RISC2
+		risc="risc2";
+#endif
+#else
+		risc="cisc";
+#endif
+#ifdef DES_UNROLL
+		unroll="16";
+#else
+		unroll="2";
+#endif
+		if (sizeof(DES_LONG) != sizeof(long))
+			size="int";
+		else
+			size="long";
+		BIO_snprintf(buf,sizeof buf,"des(%s,%s,%s,%s)",ptr,risc,unroll,
+			     size);
+		init=0;
+		}
+	return(buf);
+	}
+		
 
 void DES_ecb_encrypt(const_DES_cblock *input, DES_cblock *output,
 		     DES_key_schedule *ks, int enc)
diff --git a/crypto/des/enc_read.c b/crypto/des/enc_read.c
index e7da2ec66b08..edb6620d0856 100644
--- a/crypto/des/enc_read.c
+++ b/crypto/des/enc_read.c
@@ -63,7 +63,7 @@
 
 /* This has some uglies in it but it works - even over sockets. */
 /*extern int errno;*/
-OPENSSL_IMPLEMENT_GLOBAL(int,DES_rw_mode)=DES_PCBC_MODE;
+OPENSSL_IMPLEMENT_GLOBAL(int,DES_rw_mode,DES_PCBC_MODE)
 
 
 /*
@@ -87,6 +87,9 @@ OPENSSL_IMPLEMENT_GLOBAL(int,DES_rw_mode)=DES_PCBC_MODE;
 int DES_enc_read(int fd, void *buf, int len, DES_key_schedule *sched,
 		 DES_cblock *iv)
 	{
+#if defined(OPENSSL_NO_POSIX_IO)
+	return(0);
+#else
 	/* data to be unencrypted */
 	int net_num=0;
 	static unsigned char *net=NULL;
@@ -147,7 +150,7 @@ int DES_enc_read(int fd, void *buf, int len, DES_key_schedule *sched,
 	/* first - get the length */
 	while (net_num < HDRSIZE) 
 		{
-#ifndef _WIN32
+#ifndef OPENSSL_SYS_WIN32
 		i=read(fd,(void *)&(net[net_num]),HDRSIZE-net_num);
 #else
 		i=_read(fd,(void *)&(net[net_num]),HDRSIZE-net_num);
@@ -173,7 +176,11 @@ int DES_enc_read(int fd, void *buf, int len, DES_key_schedule *sched,
 	net_num=0;
 	while (net_num < rnum)
 		{
+#ifndef OPENSSL_SYS_WIN32
 		i=read(fd,(void *)&(net[net_num]),rnum-net_num);
+#else
+		i=_read(fd,(void *)&(net[net_num]),rnum-net_num);
+#endif
 #ifdef EINTR
 		if ((i == -1) && (errno == EINTR)) continue;
 #endif
@@ -228,5 +235,6 @@ int DES_enc_read(int fd, void *buf, int len, DES_key_schedule *sched,
 			}
 		}
 	return num;
+#endif /* OPENSSL_NO_POSIX_IO */
 	}
 
diff --git a/crypto/des/enc_writ.c b/crypto/des/enc_writ.c
index c2f032c9a6af..2353ac1e892f 100644
--- a/crypto/des/enc_writ.c
+++ b/crypto/des/enc_writ.c
@@ -80,6 +80,9 @@
 int DES_enc_write(int fd, const void *_buf, int len,
 		  DES_key_schedule *sched, DES_cblock *iv)
 	{
+#if defined(OPENSSL_NO_POSIX_IO)
+	return (-1);
+#else
 #ifdef _LIBC
 	extern unsigned long time();
 	extern int write();
@@ -172,4 +175,5 @@ int DES_enc_write(int fd, const void *_buf, int len,
 		}
 
 	return(len);
+#endif /* OPENSSL_NO_POSIX_IO */
 	}
diff --git a/crypto/des/fcrypt_b.c b/crypto/des/fcrypt_b.c
index 1390138787ff..8822816938dd 100644
--- a/crypto/des/fcrypt_b.c
+++ b/crypto/des/fcrypt_b.c
@@ -100,12 +100,10 @@ void fcrypt_body(DES_LONG *out, DES_key_schedule *ks, DES_LONG Eswap0,
 #ifndef DES_UNROLL
 		register int i;
 
-		for (i=0; i<32; i+=8)
+		for (i=0; i<32; i+=4)
 			{
 			D_ENCRYPT(l,r,i+0); /*  1 */
 			D_ENCRYPT(r,l,i+2); /*  2 */
-			D_ENCRYPT(l,r,i+4); /*  1 */
-			D_ENCRYPT(r,l,i+6); /*  2 */
 			}
 #else
 		D_ENCRYPT(l,r, 0); /*  1 */
diff --git a/crypto/des/set_key.c b/crypto/des/set_key.c
index c0806d593c1f..d3e69ca8b586 100644
--- a/crypto/des/set_key.c
+++ b/crypto/des/set_key.c
@@ -64,12 +64,10 @@
  * 1.0 First working version
  */
 #include "des_locl.h"
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
 
+#include <openssl/crypto.h>
 
-OPENSSL_IMPLEMENT_GLOBAL(int,DES_check_key);	/* defaults to false */
+OPENSSL_IMPLEMENT_GLOBAL(int,DES_check_key,0)	/* defaults to false */
 
 static const unsigned char odd_parity[256]={
   1,  1,  2,  2,  4,  4,  7,  7,  8,  8, 11, 11, 13, 13, 14, 14,
@@ -339,8 +337,15 @@ int DES_set_key_checked(const_DES_cblock *key, DES_key_schedule *schedule)
 	}
 
 void DES_set_key_unchecked(const_DES_cblock *key, DES_key_schedule *schedule)
+#ifdef OPENSSL_FIPS
+	{
+	fips_cipher_abort(DES);
+	private_DES_set_key_unchecked(key, schedule);
+	}
+void private_DES_set_key_unchecked(const_DES_cblock *key, DES_key_schedule *schedule)
+#endif
 	{
-	static int shifts2[16]={0,0,1,1,1,1,1,1,0,1,1,1,1,1,1,0};
+	static const int shifts2[16]={0,0,1,1,1,1,1,1,0,1,1,1,1,1,1,0};
 	register DES_LONG c,d,t,s,t2;
 	register const unsigned char *in;
 	register DES_LONG *k;
@@ -353,10 +358,6 @@ void DES_set_key_unchecked(const_DES_cblock *key, DES_key_schedule *schedule)
 	k = &schedule->ks->deslong[0];
 	in = &(*key)[0];
 
-#ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
-
 	c2l(in,c);
 	c2l(in,d);
 
@@ -413,4 +414,3 @@ void des_fixup_key_parity(des_cblock *key)
 	des_set_odd_parity(key);
 	}
 */
-
diff --git a/crypto/des/t/test b/crypto/des/t/test
deleted file mode 100644
index 97acd0552e43..000000000000
--- a/crypto/des/t/test
+++ /dev/null
@@ -1,27 +0,0 @@
-#!./perl
-
-BEGIN { push(@INC, qw(../../../lib ../../lib ../lib lib)); }
-
-use DES;
-
-$key='00000000';
-$ks=DES::set_key($key);
-@a=split(//,$ks);
-foreach (@a) { printf "%02x-",ord($_); }
-print "\n";
-
-
-$key=DES::random_key();
-print "($_)\n";
-@a=split(//,$key);
-foreach (@a) { printf "%02x-",ord($_); }
-print "\n";
-$str="this is and again into the breach";
-($k1,$k2)=DES::string_to_2keys($str);
-@a=split(//,$k1);
-foreach (@a) { printf "%02x-",ord($_); }
-print "\n";
-@a=split(//,$k2);
-foreach (@a) { printf "%02x-",ord($_); }
-print "\n";
-
diff --git a/crypto/des/times/486-50.sol b/crypto/des/times/486-50.sol
deleted file mode 100644
index 0de62d6db31e..000000000000
--- a/crypto/des/times/486-50.sol
+++ /dev/null
@@ -1,16 +0,0 @@
-Solaris 2.4, 486 50mhz, gcc 2.6.3
-options    des ecb/s
-16 r2 i     43552.51 100.0%
-16 r1 i     43487.45  99.9%
-16  c p     43003.23  98.7%
-16 r2 p     42339.00  97.2%
-16  c i     41900.91  96.2%
-16 r1 p     41360.64  95.0%
- 4  c i     38728.48  88.9%
- 4  c p     38225.63  87.8%
- 4 r1 i     38085.79  87.4%
- 4 r2 i     37825.64  86.9%
- 4 r2 p     34611.00  79.5%
- 4 r1 p     31802.00  73.0%
--DDES_UNROLL -DDES_RISC2
-
diff --git a/crypto/des/times/586-100.lnx b/crypto/des/times/586-100.lnx
deleted file mode 100644
index 4323914a11bb..000000000000
--- a/crypto/des/times/586-100.lnx
+++ /dev/null
@@ -1,20 +0,0 @@
-Pentium 100
-Linux 2 kernel
-gcc 2.7.0 -O3 -fomit-frame-pointer
-No X server running, just a console, it makes the top speed jump from 151,000
-to 158,000 :-).
-options    des ecb/s
-assember   281000.00 177.1%
-16 r1 p    158667.40 100.0%
-16 r1 i    148471.70  93.6%
-16 r2 p    143961.80  90.7%
-16 r2 i    141689.20  89.3%
- 4 r1 i    140100.00  88.3%
- 4 r2 i    134049.40  84.5%
-16  c i    124145.20  78.2%
-16  c p    121584.20  76.6%
- 4  c i    118116.00  74.4%
- 4 r2 p    117977.90  74.4%
- 4  c p    114971.40  72.5%
- 4 r1 p    114578.40  72.2%
--DDES_UNROLL -DDES_RISC1 -DDES_PTR
diff --git a/crypto/des/times/686-200.fre b/crypto/des/times/686-200.fre
deleted file mode 100644
index 7d83f6adee15..000000000000
--- a/crypto/des/times/686-200.fre
+++ /dev/null
@@ -1,18 +0,0 @@
-Pentium 100
-Free BSD 2.1.5 kernel
-gcc 2.7.2.2 -O3 -fomit-frame-pointer
-options    des ecb/s
-assember   578000.00 133.1%
-16 r2 i    434454.80 100.0%
-16 r1 i    433621.43  99.8%
-16 r2 p    431375.69  99.3%
- 4 r1 i    423722.30  97.5%
- 4 r2 i    422399.40  97.2%
-16 r1 p    421739.40  97.1%
-16  c i    399027.94  91.8%
-16  c p    372251.70  85.7%
- 4  c i    365118.35  84.0%
- 4  c p    352880.51  81.2%
- 4 r2 p    255104.90  58.7%
- 4 r1 p    251289.18  57.8%
--DDES_UNROLL -DDES_RISC2
diff --git a/crypto/des/times/aix.cc b/crypto/des/times/aix.cc
deleted file mode 100644
index d96b74e2cedd..000000000000
--- a/crypto/des/times/aix.cc
+++ /dev/null
@@ -1,26 +0,0 @@
-From: Paco Garcia <pgarcia@cam.es>
-
-This machine is a Bull Estrella  Minitower Model MT604-100
-Processor        : PPC604 
-P.Speed          : 100Mhz 
-Data/Instr Cache :    16 K
-L2 Cache         :   256 K
-PCI BUS Speed    :    33 Mhz
-TransfRate PCI   :   132 MB/s
-Memory           :    96 MB
-
-options    des ecb/s       
- 4  c p    275118.61 100.0%
- 4  c i    273545.07  99.4%
- 4 r2 p    270441.02  98.3%
- 4 r1 p    253052.15  92.0%
- 4 r2 i    240842.97  87.5%
- 4 r1 i    240556.66  87.4%
-16  c i    224603.99  81.6%
-16  c p    224483.98  81.6%
-16 r2 p    215691.19  78.4%
-16 r1 p    208332.83  75.7%
-16 r1 i    199206.50  72.4%
-16 r2 i    198963.70  72.3%
--DDES_PTR
-
diff --git a/crypto/des/times/alpha.cc b/crypto/des/times/alpha.cc
deleted file mode 100644
index 95c17efae7e5..000000000000
--- a/crypto/des/times/alpha.cc
+++ /dev/null
@@ -1,18 +0,0 @@
-cc -O2
-DES_LONG is 'unsigned int'
-
-options    des ecb/s
- 4 r2 p    181146.14 100.0%
-16 r2 p    172102.94  95.0%
- 4 r2 i    165424.11  91.3%
-16  c p    160468.64  88.6%
- 4  c p    156653.59  86.5%
- 4  c i    155245.18  85.7%
- 4 r1 p    154729.68  85.4%
-16 r2 i    154137.69  85.1%
-16 r1 p    152357.96  84.1%
-16  c i    148743.91  82.1%
- 4 r1 i    146695.59  81.0%
-16 r1 i    144961.00  80.0%
--DDES_RISC2 -DDES_PTR
-
diff --git a/crypto/des/times/hpux.cc b/crypto/des/times/hpux.cc
deleted file mode 100644
index 3de856ddac5a..000000000000
--- a/crypto/des/times/hpux.cc
+++ /dev/null
@@ -1,17 +0,0 @@
-HPUX 10 - 9000/887 - cc -D_HPUX_SOURCE -Aa +ESlit +O2 -Wl,-a,archive
-
-options    des ecb/s
-16  c i    149448.90 100.0%
- 4  c i    145861.79  97.6%
-16 r2 i    141710.96  94.8%
-16 r1 i    139455.33  93.3%
- 4 r2 i    138800.00  92.9%
- 4 r1 i    136692.65  91.5%
-16 r2 p    110228.17  73.8%
-16 r1 p    109397.07  73.2%
-16  c p    109209.89  73.1%
- 4  c p    108014.71  72.3%
- 4 r2 p    107873.88  72.2%
- 4 r1 p    107685.83  72.1%
--DDES_UNROLL
-
diff --git a/crypto/des/times/sparc.gcc b/crypto/des/times/sparc.gcc
deleted file mode 100644
index 8eaa04210406..000000000000
--- a/crypto/des/times/sparc.gcc
+++ /dev/null
@@ -1,17 +0,0 @@
-solaris 2.5.1 - sparc 10 50mhz - gcc 2.7.2
-
-options    des ecb/s
-16  c i    124382.70 100.0%
- 4  c i    118884.68  95.6%
-16  c p    112261.20  90.3%
-16 r2 i    111777.10  89.9%
-16 r2 p    108896.30  87.5%
-16 r1 p    108791.59  87.5%
- 4  c p    107290.10  86.3%
- 4 r1 p    104583.80  84.1%
-16 r1 i    104206.20  83.8%
- 4 r2 p    103709.80  83.4%
- 4 r2 i     98306.43  79.0%
- 4 r1 i     91525.80  73.6%
--DDES_UNROLL
-      
diff --git a/crypto/des/times/usparc.cc b/crypto/des/times/usparc.cc
deleted file mode 100644
index 0864285ef6ac..000000000000
--- a/crypto/des/times/usparc.cc
+++ /dev/null
@@ -1,31 +0,0 @@
-solaris 2.5.1 usparc 167mhz?? - SC4.0 cc -fast -Xa -xO5
-
-For the ultra sparc, SunC 4.0 cc -fast -Xa -xO5, running 'des_opts'
-gives a speed of 475,000 des/s while 'speed' gives 417,000 des/s.
-I believe the difference is tied up in optimisation that the compiler
-is able to perform when the code is 'inlined'.  For 'speed', the DES
-routines are being linked from a library.  I'll record the higher
-speed since if performance is everything, you can always inline
-'des_enc.c'.
-
-[ 16-Jan-06 - I've been playing with the
-  '-xtarget=ultra -xarch=v8plus -Xa -xO5 -Xa'
-  and while it makes the des_opts numbers much slower, it makes the
-  actual 'speed' numbers look better which is a realistic version of
-  using the libraries. ]
-
-options    des ecb/s
-16 r1 p    475516.90 100.0%
-16 r2 p    439388.10  92.4%
-16  c i    427001.40  89.8%
-16  c p    419516.50  88.2%
- 4 r2 p    409491.70  86.1%
- 4 r1 p    404266.90  85.0%
- 4  c p    398121.00  83.7%
- 4  c i    370588.40  77.9%
- 4 r1 i    362742.20  76.3%
-16 r2 i    331275.50  69.7%
-16 r1 i    324730.60  68.3%
- 4 r2 i     63535.10  13.4%	<-- very very weird, must be cache problems.
--DDES_UNROLL -DDES_RISC1 -DDES_PTR
-
diff --git a/crypto/des/xcbc_enc.c b/crypto/des/xcbc_enc.c
index dc0c761b71f7..058cab6bce0c 100644
--- a/crypto/des/xcbc_enc.c
+++ b/crypto/des/xcbc_enc.c
@@ -61,7 +61,7 @@
 /* RSA's DESX */
 
 #if 0 /* broken code, preserved just in case anyone specifically looks for this */
-static unsigned char desx_white_in2out[256]={
+static const unsigned char desx_white_in2out[256]={
 0xBD,0x56,0xEA,0xF2,0xA2,0xF1,0xAC,0x2A,0xB0,0x93,0xD1,0x9C,0x1B,0x33,0xFD,0xD0,
 0x30,0x04,0xB6,0xDC,0x7D,0xDF,0x32,0x4B,0xF7,0xCB,0x45,0x9B,0x31,0xBB,0x21,0x5A,
 0x41,0x9F,0xE1,0xD9,0x4A,0x4D,0x9E,0xDA,0xA0,0x68,0x2C,0xC3,0x27,0x5F,0x80,0x36,
diff --git a/crypto/dh/Makefile b/crypto/dh/Makefile
index d01fa960ebc7..f23b4f7fde84 100644
--- a/crypto/dh/Makefile
+++ b/crypto/dh/Makefile
@@ -17,8 +17,10 @@ TEST= dhtest.c
 APPS=
 
 LIB=$(TOP)/libcrypto.a
-LIBSRC= dh_asn1.c dh_gen.c dh_key.c dh_lib.c dh_check.c dh_err.c dh_depr.c
-LIBOBJ= dh_asn1.o dh_gen.o dh_key.o dh_lib.o dh_check.o dh_err.o dh_depr.o
+LIBSRC= dh_asn1.c dh_gen.c dh_key.c dh_lib.c dh_check.c dh_err.c dh_depr.c \
+	dh_ameth.c dh_pmeth.c dh_prn.c
+LIBOBJ= dh_asn1.o dh_gen.o dh_key.o dh_lib.o dh_check.o dh_err.o dh_depr.o \
+	dh_ameth.o dh_pmeth.o dh_prn.o
 
 SRC= $(LIBSRC)
 
@@ -33,7 +35,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -74,6 +76,21 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
+dh_ameth.o: ../../e_os.h ../../include/openssl/asn1.h
+dh_ameth.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+dh_ameth.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+dh_ameth.o: ../../include/openssl/dh.h ../../include/openssl/e_os2.h
+dh_ameth.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+dh_ameth.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+dh_ameth.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+dh_ameth.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+dh_ameth.o: ../../include/openssl/opensslconf.h
+dh_ameth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+dh_ameth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+dh_ameth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+dh_ameth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+dh_ameth.o: ../../include/openssl/x509_vfy.h ../asn1/asn1_locl.h ../cryptlib.h
+dh_ameth.o: dh_ameth.c
 dh_asn1.o: ../../e_os.h ../../include/openssl/asn1.h
 dh_asn1.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 dh_asn1.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
@@ -129,11 +146,35 @@ dh_lib.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
 dh_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 dh_lib.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 dh_lib.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-dh_lib.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-dh_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-dh_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-dh_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-dh_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-dh_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-dh_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-dh_lib.o: ../../include/openssl/x509_vfy.h ../cryptlib.h dh_lib.c
+dh_lib.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+dh_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+dh_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+dh_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+dh_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+dh_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+dh_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+dh_lib.o: ../cryptlib.h dh_lib.c
+dh_pmeth.o: ../../e_os.h ../../include/openssl/asn1.h
+dh_pmeth.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+dh_pmeth.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
+dh_pmeth.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
+dh_pmeth.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+dh_pmeth.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+dh_pmeth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+dh_pmeth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+dh_pmeth.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+dh_pmeth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+dh_pmeth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+dh_pmeth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+dh_pmeth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+dh_pmeth.o: ../../include/openssl/x509_vfy.h ../cryptlib.h ../evp/evp_locl.h
+dh_pmeth.o: dh_pmeth.c
+dh_prn.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
+dh_prn.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+dh_prn.o: ../../include/openssl/dh.h ../../include/openssl/e_os2.h
+dh_prn.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+dh_prn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+dh_prn.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+dh_prn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+dh_prn.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+dh_prn.o: ../../include/openssl/symhacks.h ../cryptlib.h dh_prn.c
diff --git a/crypto/dh/dh.h b/crypto/dh/dh.h
index 10475ac4b3c3..ea59e610ef05 100644
--- a/crypto/dh/dh.h
+++ b/crypto/dh/dh.h
@@ -77,8 +77,6 @@
 # define OPENSSL_DH_MAX_MODULUS_BITS	10000
 #endif
 
-#define OPENSSL_DH_FIPS_MIN_MODULUS_BITS 1024
-
 #define DH_FLAG_CACHE_MONT_P     0x01
 #define DH_FLAG_NO_EXP_CONSTTIME 0x02 /* new with 0.9.7h; the built-in DH
                                        * implementation now uses constant time
@@ -88,6 +86,21 @@
                                        * be used for all exponents.
                                        */
 
+/* If this flag is set the DH method is FIPS compliant and can be used
+ * in FIPS mode. This is set in the validated module method. If an
+ * application sets this flag in its own methods it is its reposibility
+ * to ensure the result is compliant.
+ */
+
+#define DH_FLAG_FIPS_METHOD			0x0400
+
+/* If this flag is set the operations normally disabled in FIPS mode are
+ * permitted it is then the applications responsibility to ensure that the
+ * usage is compliant.
+ */
+
+#define DH_FLAG_NON_FIPS_ALLOW			0x0400
+
 #ifdef  __cplusplus
 extern "C" {
 #endif
@@ -159,7 +172,6 @@ struct dh_st
    this for backward compatibility: */
 #define DH_CHECK_P_NOT_STRONG_PRIME	DH_CHECK_P_NOT_SAFE_PRIME
 
-#define DHparams_dup(x) ASN1_dup_of_const(DH,i2d_DHparams,d2i_DHparams,x)
 #define d2i_DHparams_fp(fp,x) (DH *)ASN1_d2i_fp((char *(*)())DH_new, \
 		(char *(*)())d2i_DHparams,(fp),(unsigned char **)(x))
 #define i2d_DHparams_fp(fp,x) ASN1_i2d_fp(i2d_DHparams,(fp), \
@@ -167,12 +179,9 @@ struct dh_st
 #define d2i_DHparams_bio(bp,x) ASN1_d2i_bio_of(DH,DH_new,d2i_DHparams,bp,x)
 #define i2d_DHparams_bio(bp,x) ASN1_i2d_bio_of_const(DH,i2d_DHparams,bp,x)
 
-const DH_METHOD *DH_OpenSSL(void);
+DH *DHparams_dup(DH *);
 
-#ifdef OPENSSL_FIPS
-DH *	FIPS_dh_new(void);
-void	FIPS_dh_free(DH *dh);
-#endif
+const DH_METHOD *DH_OpenSSL(void);
 
 void DH_set_default_method(const DH_METHOD *meth);
 const DH_METHOD *DH_get_default_method(void);
@@ -212,6 +221,18 @@ int	DHparams_print(BIO *bp, const DH *x);
 int	DHparams_print(char *bp, const DH *x);
 #endif
 
+#define EVP_PKEY_CTX_set_dh_paramgen_prime_len(ctx, len) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_DH, EVP_PKEY_OP_PARAMGEN, \
+			EVP_PKEY_CTRL_DH_PARAMGEN_PRIME_LEN, len, NULL)
+
+#define EVP_PKEY_CTX_set_dh_paramgen_generator(ctx, gen) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_DH, EVP_PKEY_OP_PARAMGEN, \
+			EVP_PKEY_CTRL_DH_PARAMGEN_GENERATOR, gen, NULL)
+
+#define	EVP_PKEY_CTRL_DH_PARAMGEN_PRIME_LEN	(EVP_PKEY_ALG_CTRL + 1)
+#define	EVP_PKEY_CTRL_DH_PARAMGEN_GENERATOR	(EVP_PKEY_ALG_CTRL + 2)
+		
+
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
  * made after this point may be overwritten when the script is next run.
@@ -222,22 +243,36 @@ void ERR_load_DH_strings(void);
 
 /* Function codes. */
 #define DH_F_COMPUTE_KEY				 102
-#define DH_F_DHPARAMS_PRINT				 100
 #define DH_F_DHPARAMS_PRINT_FP				 101
 #define DH_F_DH_BUILTIN_GENPARAMS			 106
-#define DH_F_DH_COMPUTE_KEY				 107
-#define DH_F_DH_GENERATE_KEY				 108
-#define DH_F_DH_GENERATE_PARAMETERS			 109
+#define DH_F_DH_COMPUTE_KEY				 114
+#define DH_F_DH_GENERATE_KEY				 115
+#define DH_F_DH_GENERATE_PARAMETERS_EX			 116
 #define DH_F_DH_NEW_METHOD				 105
+#define DH_F_DH_PARAM_DECODE				 107
+#define DH_F_DH_PRIV_DECODE				 110
+#define DH_F_DH_PRIV_ENCODE				 111
+#define DH_F_DH_PUB_DECODE				 108
+#define DH_F_DH_PUB_ENCODE				 109
+#define DH_F_DO_DH_PRINT				 100
 #define DH_F_GENERATE_KEY				 103
 #define DH_F_GENERATE_PARAMETERS			 104
+#define DH_F_PKEY_DH_DERIVE				 112
+#define DH_F_PKEY_DH_KEYGEN				 113
 
 /* Reason codes. */
 #define DH_R_BAD_GENERATOR				 101
+#define DH_R_BN_DECODE_ERROR				 109
+#define DH_R_BN_ERROR					 106
+#define DH_R_DECODE_ERROR				 104
 #define DH_R_INVALID_PUBKEY				 102
-#define DH_R_KEY_SIZE_TOO_SMALL				 104
+#define DH_R_KEYS_NOT_SET				 108
+#define DH_R_KEY_SIZE_TOO_SMALL				 110
 #define DH_R_MODULUS_TOO_LARGE				 103
+#define DH_R_NON_FIPS_METHOD				 111
+#define DH_R_NO_PARAMETERS_SET				 107
 #define DH_R_NO_PRIVATE_VALUE				 100
+#define DH_R_PARAMETER_ENCODING_ERROR			 105
 
 #ifdef  __cplusplus
 }
diff --git a/crypto/dh/dh_ameth.c b/crypto/dh/dh_ameth.c
new file mode 100644
index 000000000000..02ec2d47b4bd
--- /dev/null
+++ b/crypto/dh/dh_ameth.c
@@ -0,0 +1,501 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/x509.h>
+#include <openssl/asn1.h>
+#include <openssl/dh.h>
+#include <openssl/bn.h>
+#include "asn1_locl.h"
+
+static void int_dh_free(EVP_PKEY *pkey)
+	{
+	DH_free(pkey->pkey.dh);
+	}
+
+static int dh_pub_decode(EVP_PKEY *pkey, X509_PUBKEY *pubkey)
+	{
+	const unsigned char *p, *pm;
+	int pklen, pmlen;
+	int ptype;
+	void *pval;
+	ASN1_STRING *pstr;
+	X509_ALGOR *palg;
+	ASN1_INTEGER *public_key = NULL;
+
+	DH *dh = NULL;
+
+	if (!X509_PUBKEY_get0_param(NULL, &p, &pklen, &palg, pubkey))
+		return 0;
+	X509_ALGOR_get0(NULL, &ptype, &pval, palg);
+
+	if (ptype != V_ASN1_SEQUENCE)
+		{
+		DHerr(DH_F_DH_PUB_DECODE, DH_R_PARAMETER_ENCODING_ERROR);
+		goto err;
+		}
+
+	pstr = pval;	
+	pm = pstr->data;
+	pmlen = pstr->length;
+
+	if (!(dh = d2i_DHparams(NULL, &pm, pmlen)))
+		{
+		DHerr(DH_F_DH_PUB_DECODE, DH_R_DECODE_ERROR);
+		goto err;
+		}
+
+	if (!(public_key=d2i_ASN1_INTEGER(NULL, &p, pklen)))
+		{
+		DHerr(DH_F_DH_PUB_DECODE, DH_R_DECODE_ERROR);
+		goto err;
+		}
+
+	/* We have parameters now set public key */
+	if (!(dh->pub_key = ASN1_INTEGER_to_BN(public_key, NULL)))
+		{
+		DHerr(DH_F_DH_PUB_DECODE, DH_R_BN_DECODE_ERROR);
+		goto err;
+		}
+
+	ASN1_INTEGER_free(public_key);
+	EVP_PKEY_assign_DH(pkey, dh);
+	return 1;
+
+	err:
+	if (public_key)
+		ASN1_INTEGER_free(public_key);
+	if (dh)
+		DH_free(dh);
+	return 0;
+
+	}
+
+static int dh_pub_encode(X509_PUBKEY *pk, const EVP_PKEY *pkey)
+	{
+	DH *dh;
+	void *pval = NULL;
+	int ptype;
+	unsigned char *penc = NULL;
+	int penclen;
+	ASN1_STRING *str;
+	ASN1_INTEGER *pub_key = NULL;
+
+	dh=pkey->pkey.dh;
+
+	str = ASN1_STRING_new();
+	str->length = i2d_DHparams(dh, &str->data);
+	if (str->length <= 0)
+		{
+		DHerr(DH_F_DH_PUB_ENCODE, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+	pval = str;
+	ptype = V_ASN1_SEQUENCE;
+
+	pub_key = BN_to_ASN1_INTEGER(dh->pub_key, NULL);
+	if (!pub_key)
+		goto err;
+
+	penclen = i2d_ASN1_INTEGER(pub_key, &penc);
+
+	ASN1_INTEGER_free(pub_key);
+
+	if (penclen <= 0)
+		{
+		DHerr(DH_F_DH_PUB_ENCODE, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	if (X509_PUBKEY_set0_param(pk, OBJ_nid2obj(EVP_PKEY_DH),
+				ptype, pval, penc, penclen))
+		return 1;
+
+	err:
+	if (penc)
+		OPENSSL_free(penc);
+	if (pval)
+		ASN1_STRING_free(pval);
+
+	return 0;
+	}
+
+
+/* PKCS#8 DH is defined in PKCS#11 of all places. It is similar to DH in
+ * that the AlgorithmIdentifier contains the paramaters, the private key
+ * is explcitly included and the pubkey must be recalculated.
+ */
+	
+static int dh_priv_decode(EVP_PKEY *pkey, PKCS8_PRIV_KEY_INFO *p8)
+	{
+	const unsigned char *p, *pm;
+	int pklen, pmlen;
+	int ptype;
+	void *pval;
+	ASN1_STRING *pstr;
+	X509_ALGOR *palg;
+	ASN1_INTEGER *privkey = NULL;
+
+	DH *dh = NULL;
+
+	if (!PKCS8_pkey_get0(NULL, &p, &pklen, &palg, p8))
+		return 0;
+
+	X509_ALGOR_get0(NULL, &ptype, &pval, palg);
+
+	if (ptype != V_ASN1_SEQUENCE)
+			goto decerr;
+
+	if (!(privkey=d2i_ASN1_INTEGER(NULL, &p, pklen)))
+		goto decerr;
+
+
+	pstr = pval;	
+	pm = pstr->data;
+	pmlen = pstr->length;
+	if (!(dh = d2i_DHparams(NULL, &pm, pmlen)))
+		goto decerr;
+	/* We have parameters now set private key */
+	if (!(dh->priv_key = ASN1_INTEGER_to_BN(privkey, NULL)))
+		{
+		DHerr(DH_F_DH_PRIV_DECODE,DH_R_BN_ERROR);
+		goto dherr;
+		}
+	/* Calculate public key */
+	if (!DH_generate_key(dh))
+		goto dherr;
+
+	EVP_PKEY_assign_DH(pkey, dh);
+
+	ASN1_INTEGER_free(privkey);
+
+	return 1;
+
+	decerr:
+	DHerr(DH_F_DH_PRIV_DECODE, EVP_R_DECODE_ERROR);
+	dherr:
+	DH_free(dh);
+	return 0;
+	}
+
+static int dh_priv_encode(PKCS8_PRIV_KEY_INFO *p8, const EVP_PKEY *pkey)
+{
+	ASN1_STRING *params = NULL;
+	ASN1_INTEGER *prkey = NULL;
+	unsigned char *dp = NULL;
+	int dplen;
+
+	params = ASN1_STRING_new();
+
+	if (!params)
+		{
+		DHerr(DH_F_DH_PRIV_ENCODE,ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	params->length = i2d_DHparams(pkey->pkey.dh, &params->data);
+	if (params->length <= 0)
+		{
+		DHerr(DH_F_DH_PRIV_ENCODE,ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+	params->type = V_ASN1_SEQUENCE;
+
+	/* Get private key into integer */
+	prkey = BN_to_ASN1_INTEGER(pkey->pkey.dh->priv_key, NULL);
+
+	if (!prkey)
+		{
+		DHerr(DH_F_DH_PRIV_ENCODE,DH_R_BN_ERROR);
+		goto err;
+		}
+
+	dplen = i2d_ASN1_INTEGER(prkey, &dp);
+
+	ASN1_INTEGER_free(prkey);
+
+	if (!PKCS8_pkey_set0(p8, OBJ_nid2obj(NID_dhKeyAgreement), 0,
+				V_ASN1_SEQUENCE, params, dp, dplen))
+		goto err;
+
+	return 1;
+
+err:
+	if (dp != NULL)
+		OPENSSL_free(dp);
+	if (params != NULL)
+		ASN1_STRING_free(params);
+	if (prkey != NULL)
+		ASN1_INTEGER_free(prkey);
+	return 0;
+}
+
+
+static void update_buflen(const BIGNUM *b, size_t *pbuflen)
+	{
+	size_t i;
+	if (!b)
+		return;
+	if (*pbuflen < (i = (size_t)BN_num_bytes(b)))
+			*pbuflen = i;
+	}
+
+static int dh_param_decode(EVP_PKEY *pkey,
+					const unsigned char **pder, int derlen)
+	{
+	DH *dh;
+	if (!(dh = d2i_DHparams(NULL, pder, derlen)))
+		{
+		DHerr(DH_F_DH_PARAM_DECODE, ERR_R_DH_LIB);
+		return 0;
+		}
+	EVP_PKEY_assign_DH(pkey, dh);
+	return 1;
+	}
+
+static int dh_param_encode(const EVP_PKEY *pkey, unsigned char **pder)
+	{
+	return i2d_DHparams(pkey->pkey.dh, pder);
+	}
+
+static int do_dh_print(BIO *bp, const DH *x, int indent,
+						ASN1_PCTX *ctx, int ptype)
+	{
+	unsigned char *m=NULL;
+	int reason=ERR_R_BUF_LIB,ret=0;
+	size_t buf_len=0;
+
+	const char *ktype = NULL;
+
+	BIGNUM *priv_key, *pub_key;
+
+	if (ptype == 2)
+		priv_key = x->priv_key;
+	else
+		priv_key = NULL;
+
+	if (ptype > 0)
+		pub_key = x->pub_key;
+	else
+		pub_key = NULL;
+
+	update_buflen(x->p, &buf_len);
+
+	if (buf_len == 0)
+		{
+		reason = ERR_R_PASSED_NULL_PARAMETER;
+		goto err;
+		}
+
+	update_buflen(x->g, &buf_len);
+	update_buflen(pub_key, &buf_len);
+	update_buflen(priv_key, &buf_len);
+
+	if (ptype == 2)
+		ktype = "PKCS#3 DH Private-Key";
+	else if (ptype == 1)
+		ktype = "PKCS#3 DH Public-Key";
+	else
+		ktype = "PKCS#3 DH Parameters";
+
+	m= OPENSSL_malloc(buf_len+10);
+	if (m == NULL)
+		{
+		reason=ERR_R_MALLOC_FAILURE;
+		goto err;
+		}
+
+	BIO_indent(bp, indent, 128);
+	if (BIO_printf(bp,"%s: (%d bit)\n", ktype, BN_num_bits(x->p)) <= 0)
+		goto err;
+	indent += 4;
+
+	if (!ASN1_bn_print(bp,"private-key:",priv_key,m,indent)) goto err;
+	if (!ASN1_bn_print(bp,"public-key:",pub_key,m,indent)) goto err;
+
+	if (!ASN1_bn_print(bp,"prime:",x->p,m,indent)) goto err;
+	if (!ASN1_bn_print(bp,"generator:",x->g,m,indent)) goto err;
+	if (x->length != 0)
+		{
+		BIO_indent(bp, indent, 128);
+		if (BIO_printf(bp,"recommended-private-length: %d bits\n",
+			(int)x->length) <= 0) goto err;
+		}
+
+
+	ret=1;
+	if (0)
+		{
+err:
+		DHerr(DH_F_DO_DH_PRINT,reason);
+		}
+	if (m != NULL) OPENSSL_free(m);
+	return(ret);
+	}
+
+static int int_dh_size(const EVP_PKEY *pkey)
+	{
+	return(DH_size(pkey->pkey.dh));
+	}
+
+static int dh_bits(const EVP_PKEY *pkey)
+	{
+	return BN_num_bits(pkey->pkey.dh->p);
+	}
+
+static int dh_cmp_parameters(const EVP_PKEY *a, const EVP_PKEY *b)
+	{
+	if (	BN_cmp(a->pkey.dh->p,b->pkey.dh->p) ||
+		BN_cmp(a->pkey.dh->g,b->pkey.dh->g))
+		return 0;
+	else
+		return 1;
+	}
+
+static int dh_copy_parameters(EVP_PKEY *to, const EVP_PKEY *from)
+	{
+	BIGNUM *a;
+
+	if ((a=BN_dup(from->pkey.dh->p)) == NULL)
+		return 0;
+	if (to->pkey.dh->p != NULL)
+		BN_free(to->pkey.dh->p);
+	to->pkey.dh->p=a;
+
+	if ((a=BN_dup(from->pkey.dh->g)) == NULL)
+		return 0;
+	if (to->pkey.dh->g != NULL)
+		BN_free(to->pkey.dh->g);
+	to->pkey.dh->g=a;
+
+	return 1;
+	}
+
+static int dh_missing_parameters(const EVP_PKEY *a)
+	{
+	if (!a->pkey.dh->p || !a->pkey.dh->g)
+		return 1;
+	return 0;
+	}
+
+static int dh_pub_cmp(const EVP_PKEY *a, const EVP_PKEY *b)
+	{
+	if (dh_cmp_parameters(a, b) == 0)
+		return 0;
+	if (BN_cmp(b->pkey.dh->pub_key,a->pkey.dh->pub_key) != 0)
+		return 0;
+	else
+		return 1;
+	}
+
+static int dh_param_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_dh_print(bp, pkey->pkey.dh, indent, ctx, 0);
+	}
+
+static int dh_public_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_dh_print(bp, pkey->pkey.dh, indent, ctx, 1);
+	}
+
+static int dh_private_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_dh_print(bp, pkey->pkey.dh, indent, ctx, 2);
+	}
+
+int DHparams_print(BIO *bp, const DH *x)
+	{
+	return do_dh_print(bp, x, 4, NULL, 0);
+	}
+
+const EVP_PKEY_ASN1_METHOD dh_asn1_meth = 
+	{
+	EVP_PKEY_DH,
+	EVP_PKEY_DH,
+	0,
+
+	"DH",
+	"OpenSSL PKCS#3 DH method",
+
+	dh_pub_decode,
+	dh_pub_encode,
+	dh_pub_cmp,
+	dh_public_print,
+
+	dh_priv_decode,
+	dh_priv_encode,
+	dh_private_print,
+
+	int_dh_size,
+	dh_bits,
+
+	dh_param_decode,
+	dh_param_encode,
+	dh_missing_parameters,
+	dh_copy_parameters,
+	dh_cmp_parameters,
+	dh_param_print,
+	0,
+
+	int_dh_free,
+	0
+	};
+
diff --git a/crypto/dh/dh_asn1.c b/crypto/dh/dh_asn1.c
index 76740af2bd16..0b4357d60530 100644
--- a/crypto/dh/dh_asn1.c
+++ b/crypto/dh/dh_asn1.c
@@ -3,7 +3,7 @@
  * project 2000.
  */
 /* ====================================================================
- * Copyright (c) 2000 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2000-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -64,7 +64,8 @@
 #include <openssl/asn1t.h>
 
 /* Override the default free and new methods */
-static int dh_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int dh_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+						void *exarg)
 {
 	if(operation == ASN1_OP_NEW_PRE) {
 		*pval = (ASN1_VALUE *)DH_new();
@@ -85,3 +86,8 @@ ASN1_SEQUENCE_cb(DHparams, dh_cb) = {
 } ASN1_SEQUENCE_END_cb(DH, DHparams)
 
 IMPLEMENT_ASN1_ENCODE_FUNCTIONS_const_fname(DH, DHparams, DHparams)
+
+DH *DHparams_dup(DH *dh)
+	{
+	return ASN1_item_dup(ASN1_ITEM_rptr(DHparams), dh);
+	}
diff --git a/crypto/dh/dh_check.c b/crypto/dh/dh_check.c
index 316cb9221dff..066898174e39 100644
--- a/crypto/dh/dh_check.c
+++ b/crypto/dh/dh_check.c
@@ -70,8 +70,6 @@
  * should hold.
  */
 
-#ifndef OPENSSL_FIPS
-
 int DH_check(const DH *dh, int *ret)
 	{
 	int ok=0;
@@ -130,11 +128,11 @@ int DH_check_pub_key(const DH *dh, const BIGNUM *pub_key, int *ret)
 	q=BN_new();
 	if (q == NULL) goto err;
 	BN_set_word(q,1);
-	if (BN_cmp(pub_key,q) <= 0)
+	if (BN_cmp(pub_key,q)<=0)
 		*ret|=DH_CHECK_PUBKEY_TOO_SMALL;
 	BN_copy(q,dh->p);
 	BN_sub_word(q,1);
-	if (BN_cmp(pub_key,q) >= 0)
+	if (BN_cmp(pub_key,q)>=0)
 		*ret|=DH_CHECK_PUBKEY_TOO_LARGE;
 
 	ok = 1;
@@ -142,5 +140,3 @@ err:
 	if (q != NULL) BN_free(q);
 	return(ok);
 	}
-
-#endif
diff --git a/crypto/dh/dh_err.c b/crypto/dh/dh_err.c
index 13263c81c1b0..56d3df73563d 100644
--- a/crypto/dh/dh_err.c
+++ b/crypto/dh/dh_err.c
@@ -1,6 +1,6 @@
 /* crypto/dh/dh_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -71,25 +71,39 @@
 static ERR_STRING_DATA DH_str_functs[]=
 	{
 {ERR_FUNC(DH_F_COMPUTE_KEY),	"COMPUTE_KEY"},
-{ERR_FUNC(DH_F_DHPARAMS_PRINT),	"DHparams_print"},
 {ERR_FUNC(DH_F_DHPARAMS_PRINT_FP),	"DHparams_print_fp"},
 {ERR_FUNC(DH_F_DH_BUILTIN_GENPARAMS),	"DH_BUILTIN_GENPARAMS"},
 {ERR_FUNC(DH_F_DH_COMPUTE_KEY),	"DH_compute_key"},
 {ERR_FUNC(DH_F_DH_GENERATE_KEY),	"DH_generate_key"},
-{ERR_FUNC(DH_F_DH_GENERATE_PARAMETERS),	"DH_generate_parameters"},
+{ERR_FUNC(DH_F_DH_GENERATE_PARAMETERS_EX),	"DH_generate_parameters_ex"},
 {ERR_FUNC(DH_F_DH_NEW_METHOD),	"DH_new_method"},
+{ERR_FUNC(DH_F_DH_PARAM_DECODE),	"DH_PARAM_DECODE"},
+{ERR_FUNC(DH_F_DH_PRIV_DECODE),	"DH_PRIV_DECODE"},
+{ERR_FUNC(DH_F_DH_PRIV_ENCODE),	"DH_PRIV_ENCODE"},
+{ERR_FUNC(DH_F_DH_PUB_DECODE),	"DH_PUB_DECODE"},
+{ERR_FUNC(DH_F_DH_PUB_ENCODE),	"DH_PUB_ENCODE"},
+{ERR_FUNC(DH_F_DO_DH_PRINT),	"DO_DH_PRINT"},
 {ERR_FUNC(DH_F_GENERATE_KEY),	"GENERATE_KEY"},
 {ERR_FUNC(DH_F_GENERATE_PARAMETERS),	"GENERATE_PARAMETERS"},
+{ERR_FUNC(DH_F_PKEY_DH_DERIVE),	"PKEY_DH_DERIVE"},
+{ERR_FUNC(DH_F_PKEY_DH_KEYGEN),	"PKEY_DH_KEYGEN"},
 {0,NULL}
 	};
 
 static ERR_STRING_DATA DH_str_reasons[]=
 	{
 {ERR_REASON(DH_R_BAD_GENERATOR)          ,"bad generator"},
+{ERR_REASON(DH_R_BN_DECODE_ERROR)        ,"bn decode error"},
+{ERR_REASON(DH_R_BN_ERROR)               ,"bn error"},
+{ERR_REASON(DH_R_DECODE_ERROR)           ,"decode error"},
 {ERR_REASON(DH_R_INVALID_PUBKEY)         ,"invalid public key"},
+{ERR_REASON(DH_R_KEYS_NOT_SET)           ,"keys not set"},
 {ERR_REASON(DH_R_KEY_SIZE_TOO_SMALL)     ,"key size too small"},
 {ERR_REASON(DH_R_MODULUS_TOO_LARGE)      ,"modulus too large"},
+{ERR_REASON(DH_R_NON_FIPS_METHOD)        ,"non fips method"},
+{ERR_REASON(DH_R_NO_PARAMETERS_SET)      ,"no parameters set"},
 {ERR_REASON(DH_R_NO_PRIVATE_VALUE)       ,"no private value"},
+{ERR_REASON(DH_R_PARAMETER_ENCODING_ERROR),"parameter encoding error"},
 {0,NULL}
 	};
 
diff --git a/crypto/dh/dh_gen.c b/crypto/dh/dh_gen.c
index 999e1deb4092..7b1fe9c9cbba 100644
--- a/crypto/dh/dh_gen.c
+++ b/crypto/dh/dh_gen.c
@@ -66,14 +66,29 @@
 #include <openssl/bn.h>
 #include <openssl/dh.h>
 
-#ifndef OPENSSL_FIPS
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 
 static int dh_builtin_genparams(DH *ret, int prime_len, int generator, BN_GENCB *cb);
 
 int DH_generate_parameters_ex(DH *ret, int prime_len, int generator, BN_GENCB *cb)
 	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(ret->meth->flags & DH_FLAG_FIPS_METHOD)
+			&& !(ret->flags & DH_FLAG_NON_FIPS_ALLOW))
+		{
+		DHerr(DH_F_DH_GENERATE_PARAMETERS_EX, DH_R_NON_FIPS_METHOD);
+		return 0;
+		}
+#endif
 	if(ret->meth->generate_params)
 		return ret->meth->generate_params(ret, prime_len, generator, cb);
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode())
+		return FIPS_dh_generate_parameters_ex(ret, prime_len,
+							generator, cb);
+#endif
 	return dh_builtin_genparams(ret, prime_len, generator, cb);
 	}
 
@@ -175,5 +190,3 @@ err:
 		}
 	return ok;
 	}
-
-#endif
diff --git a/crypto/dh/dh_key.c b/crypto/dh/dh_key.c
index 79dd33186341..89a74db4e691 100644
--- a/crypto/dh/dh_key.c
+++ b/crypto/dh/dh_key.c
@@ -62,8 +62,6 @@
 #include <openssl/rand.h>
 #include <openssl/dh.h>
 
-#ifndef OPENSSL_FIPS
-
 static int generate_key(DH *dh);
 static int compute_key(unsigned char *key, const BIGNUM *pub_key, DH *dh);
 static int dh_bn_mod_exp(const DH *dh, BIGNUM *r,
@@ -75,11 +73,27 @@ static int dh_finish(DH *dh);
 
 int DH_generate_key(DH *dh)
 	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(dh->meth->flags & DH_FLAG_FIPS_METHOD)
+			&& !(dh->flags & DH_FLAG_NON_FIPS_ALLOW))
+		{
+		DHerr(DH_F_DH_GENERATE_KEY, DH_R_NON_FIPS_METHOD);
+		return 0;
+		}
+#endif
 	return dh->meth->generate_key(dh);
 	}
 
 int DH_compute_key(unsigned char *key, const BIGNUM *pub_key, DH *dh)
 	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(dh->meth->flags & DH_FLAG_FIPS_METHOD)
+			&& !(dh->flags & DH_FLAG_NON_FIPS_ALLOW))
+		{
+		DHerr(DH_F_DH_COMPUTE_KEY, DH_R_NON_FIPS_METHOD);
+		return 0;
+		}
+#endif
 	return dh->meth->compute_key(key, pub_key, dh);
 	}
 
@@ -140,8 +154,21 @@ static int generate_key(DH *dh)
 
 	if (generate_new_key)
 		{
-		l = dh->length ? dh->length : BN_num_bits(dh->p)-1; /* secret exponent length */
-		if (!BN_rand(priv_key, l, 0, 0)) goto err;
+		if (dh->q)
+			{
+			do
+				{
+				if (!BN_rand_range(priv_key, dh->q))
+					goto err;
+				}
+			while (BN_is_zero(priv_key) || BN_is_one(priv_key));
+			}
+		else
+			{
+			/* secret exponent length */
+			l = dh->length ? dh->length : BN_num_bits(dh->p)-1;
+			if (!BN_rand(priv_key, l, 0, 0)) goto err;
+			}
 		}
 
 	{
@@ -263,5 +290,3 @@ static int dh_finish(DH *dh)
 		BN_MONT_CTX_free(dh->method_mont_p);
 	return(1);
 	}
-
-#endif
diff --git a/crypto/dh/dh_lib.c b/crypto/dh/dh_lib.c
index 7aef080e7abb..00218f2b92bb 100644
--- a/crypto/dh/dh_lib.c
+++ b/crypto/dh/dh_lib.c
@@ -64,6 +64,10 @@
 #include <openssl/engine.h>
 #endif
 
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
+
 const char DH_version[]="Diffie-Hellman" OPENSSL_VERSION_PTEXT;
 
 static const DH_METHOD *default_DH_method = NULL;
@@ -76,7 +80,16 @@ void DH_set_default_method(const DH_METHOD *meth)
 const DH_METHOD *DH_get_default_method(void)
 	{
 	if(!default_DH_method)
+		{
+#ifdef OPENSSL_FIPS
+		if (FIPS_mode())
+			return FIPS_dh_openssl();
+		else
+			return DH_OpenSSL();
+#else
 		default_DH_method = DH_OpenSSL();
+#endif
+		}
 	return default_DH_method;
 	}
 
@@ -156,7 +169,7 @@ DH *DH_new_method(ENGINE *engine)
 	ret->counter = NULL;
 	ret->method_mont_p=NULL;
 	ret->references = 1;
-	ret->flags=ret->meth->flags;
+	ret->flags=ret->meth->flags & ~DH_FLAG_NON_FIPS_ALLOW;
 	CRYPTO_new_ex_data(CRYPTO_EX_INDEX_DH, ret, &ret->ex_data);
 	if ((ret->meth->init != NULL) && !ret->meth->init(ret))
 		{
diff --git a/crypto/dh/dh_pmeth.c b/crypto/dh/dh_pmeth.c
new file mode 100644
index 000000000000..5ae72b7d4cc2
--- /dev/null
+++ b/crypto/dh/dh_pmeth.c
@@ -0,0 +1,254 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/asn1t.h>
+#include <openssl/x509.h>
+#include <openssl/evp.h>
+#include <openssl/dh.h>
+#include <openssl/bn.h>
+#include "evp_locl.h"
+
+/* DH pkey context structure */
+
+typedef struct
+	{
+	/* Parameter gen parameters */
+	int prime_len;
+	int generator;
+	int use_dsa;
+	/* Keygen callback info */
+	int gentmp[2];
+	/* message digest */
+	} DH_PKEY_CTX;
+
+static int pkey_dh_init(EVP_PKEY_CTX *ctx)
+	{
+	DH_PKEY_CTX *dctx;
+	dctx = OPENSSL_malloc(sizeof(DH_PKEY_CTX));
+	if (!dctx)
+		return 0;
+	dctx->prime_len = 1024;
+	dctx->generator = 2;
+	dctx->use_dsa = 0;
+
+	ctx->data = dctx;
+	ctx->keygen_info = dctx->gentmp;
+	ctx->keygen_info_count = 2;
+	
+	return 1;
+	}
+
+static int pkey_dh_copy(EVP_PKEY_CTX *dst, EVP_PKEY_CTX *src)
+	{
+	DH_PKEY_CTX *dctx, *sctx;
+	if (!pkey_dh_init(dst))
+		return 0;
+       	sctx = src->data;
+	dctx = dst->data;
+	dctx->prime_len = sctx->prime_len;
+	dctx->generator = sctx->generator;
+	dctx->use_dsa = sctx->use_dsa;
+	return 1;
+	}
+
+static void pkey_dh_cleanup(EVP_PKEY_CTX *ctx)
+	{
+	DH_PKEY_CTX *dctx = ctx->data;
+	if (dctx)
+		OPENSSL_free(dctx);
+	}
+
+static int pkey_dh_ctrl(EVP_PKEY_CTX *ctx, int type, int p1, void *p2)
+	{
+	DH_PKEY_CTX *dctx = ctx->data;
+	switch (type)
+		{
+		case EVP_PKEY_CTRL_DH_PARAMGEN_PRIME_LEN:
+		if (p1 < 256)
+			return -2;
+		dctx->prime_len = p1;
+		return 1;
+
+		case EVP_PKEY_CTRL_DH_PARAMGEN_GENERATOR:
+		dctx->generator = p1;
+		return 1;
+
+		case EVP_PKEY_CTRL_PEER_KEY:
+		/* Default behaviour is OK */
+		return 1;
+
+		default:
+		return -2;
+
+		}
+	}
+
+			
+static int pkey_dh_ctrl_str(EVP_PKEY_CTX *ctx,
+			const char *type, const char *value)
+	{
+	if (!strcmp(type, "dh_paramgen_prime_len"))
+		{
+		int len;
+		len = atoi(value);
+		return EVP_PKEY_CTX_set_dh_paramgen_prime_len(ctx, len);
+		}
+	if (!strcmp(type, "dh_paramgen_generator"))
+		{
+		int len;
+		len = atoi(value);
+		return EVP_PKEY_CTX_set_dh_paramgen_generator(ctx, len);
+		}
+	return -2;
+	}
+
+static int pkey_dh_paramgen(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey)
+	{
+	DH *dh = NULL;
+	DH_PKEY_CTX *dctx = ctx->data;
+	BN_GENCB *pcb, cb;
+	int ret;
+	if (ctx->pkey_gencb)
+		{
+		pcb = &cb;
+		evp_pkey_set_cb_translate(pcb, ctx);
+		}
+	else
+		pcb = NULL;
+	dh = DH_new();
+	if (!dh)
+		return 0;
+	ret = DH_generate_parameters_ex(dh,
+					dctx->prime_len, dctx->generator, pcb);
+	if (ret)
+		EVP_PKEY_assign_DH(pkey, dh);
+	else
+		DH_free(dh);
+	return ret;
+	}
+
+static int pkey_dh_keygen(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey)
+	{
+	DH *dh = NULL;
+	if (ctx->pkey == NULL)
+		{
+		DHerr(DH_F_PKEY_DH_KEYGEN, DH_R_NO_PARAMETERS_SET);
+		return 0;
+		}
+	dh = DH_new();
+	if (!dh)
+		return 0;
+	EVP_PKEY_assign_DH(pkey, dh);
+	/* Note: if error return, pkey is freed by parent routine */
+	if (!EVP_PKEY_copy_parameters(pkey, ctx->pkey))
+		return 0;
+	return DH_generate_key(pkey->pkey.dh);
+	}
+
+static int pkey_dh_derive(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen)
+	{
+	int ret;
+	if (!ctx->pkey || !ctx->peerkey)
+		{
+		DHerr(DH_F_PKEY_DH_DERIVE, DH_R_KEYS_NOT_SET);
+		return 0;
+		}
+	ret = DH_compute_key(key, ctx->peerkey->pkey.dh->pub_key,
+							ctx->pkey->pkey.dh);
+	if (ret < 0)
+		return ret;
+	*keylen = ret;
+	return 1;
+	}
+
+const EVP_PKEY_METHOD dh_pkey_meth = 
+	{
+	EVP_PKEY_DH,
+	EVP_PKEY_FLAG_AUTOARGLEN,
+	pkey_dh_init,
+	pkey_dh_copy,
+	pkey_dh_cleanup,
+
+	0,
+	pkey_dh_paramgen,
+
+	0,
+	pkey_dh_keygen,
+
+	0,
+	0,
+
+	0,
+	0,
+
+	0,0,
+
+	0,0,0,0,
+
+	0,0,
+
+	0,0,
+
+	0,
+	pkey_dh_derive,
+
+	pkey_dh_ctrl,
+	pkey_dh_ctrl_str
+
+	};
diff --git a/crypto/err/err_bio.c b/crypto/dh/dh_prn.c
index a42f804840b1..ae58c2ac873b 100644
--- a/crypto/err/err_bio.c
+++ b/crypto/dh/dh_prn.c
@@ -1,4 +1,4 @@
-/* crypto/err/err_prn.c */
+/* crypto/asn1/t_pkey.c */
 /* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
  * All rights reserved.
  *
@@ -58,18 +58,23 @@
 
 #include <stdio.h>
 #include "cryptlib.h"
-#include <openssl/lhash.h>
-#include <openssl/crypto.h>
-#include <openssl/buffer.h>
-#include <openssl/err.h>
+#include <openssl/evp.h>
+#include <openssl/dh.h>
 
-static int print_bio(const char *str, size_t len, void *bp)
+#ifndef OPENSSL_NO_FP_API
+int DHparams_print_fp(FILE *fp, const DH *x)
 	{
-	return BIO_write((BIO *)bp, str, len);
-	}
-void ERR_print_errors(BIO *bp)
-	{
-	ERR_print_errors_cb(print_bio, bp);
-	}
+	BIO *b;
+	int ret;
 
-	
+	if ((b=BIO_new(BIO_s_file())) == NULL)
+		{
+		DHerr(DH_F_DHPARAMS_PRINT_FP,ERR_R_BUF_LIB);
+		return(0);
+		}
+	BIO_set_fp(b,fp,BIO_NOCLOSE);
+	ret=DHparams_print(b, x);
+	BIO_free(b);
+	return(ret);
+	}
+#endif
diff --git a/crypto/dsa/Makefile b/crypto/dsa/Makefile
index 6c9578c04ecf..5fef4ca5adfc 100644
--- a/crypto/dsa/Makefile
+++ b/crypto/dsa/Makefile
@@ -18,14 +18,14 @@ APPS=
 
 LIB=$(TOP)/libcrypto.a
 LIBSRC= dsa_gen.c dsa_key.c dsa_lib.c dsa_asn1.c dsa_vrf.c dsa_sign.c \
-	dsa_err.c dsa_ossl.c dsa_depr.c dsa_utl.c
+	dsa_err.c dsa_ossl.c dsa_depr.c dsa_ameth.c dsa_pmeth.c dsa_prn.c
 LIBOBJ= dsa_gen.o dsa_key.o dsa_lib.o dsa_asn1.o dsa_vrf.o dsa_sign.o \
-	dsa_err.o dsa_ossl.o dsa_depr.o dsa_utl.o
+	dsa_err.o dsa_ossl.o dsa_depr.o dsa_ameth.o dsa_pmeth.o dsa_prn.o
 
 SRC= $(LIBSRC)
 
 EXHEADER= dsa.h
-HEADER=	$(EXHEADER)
+HEADER=	dsa_locl.h $(EXHEADER)
 
 ALL=    $(GENERAL) $(SRC) $(HEADER)
 
@@ -35,7 +35,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -76,12 +76,27 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
+dsa_ameth.o: ../../e_os.h ../../include/openssl/asn1.h
+dsa_ameth.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+dsa_ameth.o: ../../include/openssl/buffer.h ../../include/openssl/cms.h
+dsa_ameth.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
+dsa_ameth.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+dsa_ameth.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+dsa_ameth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+dsa_ameth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+dsa_ameth.o: ../../include/openssl/objects.h
+dsa_ameth.o: ../../include/openssl/opensslconf.h
+dsa_ameth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+dsa_ameth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+dsa_ameth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+dsa_ameth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+dsa_ameth.o: ../../include/openssl/x509_vfy.h ../asn1/asn1_locl.h ../cryptlib.h
+dsa_ameth.o: dsa_ameth.c
 dsa_asn1.o: ../../e_os.h ../../include/openssl/asn1.h
 dsa_asn1.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
-dsa_asn1.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
-dsa_asn1.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
-dsa_asn1.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-dsa_asn1.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+dsa_asn1.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+dsa_asn1.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
+dsa_asn1.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
 dsa_asn1.o: ../../include/openssl/opensslconf.h
 dsa_asn1.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 dsa_asn1.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
@@ -92,9 +107,8 @@ dsa_depr.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 dsa_depr.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 dsa_depr.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 dsa_depr.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-dsa_depr.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-dsa_depr.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-dsa_depr.o: ../../include/openssl/opensslconf.h
+dsa_depr.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+dsa_depr.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 dsa_depr.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 dsa_depr.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
 dsa_depr.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -111,13 +125,12 @@ dsa_gen.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 dsa_gen.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 dsa_gen.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 dsa_gen.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-dsa_gen.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-dsa_gen.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-dsa_gen.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-dsa_gen.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
-dsa_gen.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-dsa_gen.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-dsa_gen.o: ../cryptlib.h dsa_gen.c
+dsa_gen.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+dsa_gen.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+dsa_gen.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+dsa_gen.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
+dsa_gen.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+dsa_gen.o: ../../include/openssl/symhacks.h ../cryptlib.h dsa_gen.c dsa_locl.h
 dsa_key.o: ../../e_os.h ../../include/openssl/bio.h ../../include/openssl/bn.h
 dsa_key.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 dsa_key.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
@@ -133,14 +146,14 @@ dsa_lib.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
 dsa_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 dsa_lib.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 dsa_lib.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-dsa_lib.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-dsa_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-dsa_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-dsa_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-dsa_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-dsa_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-dsa_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-dsa_lib.o: ../../include/openssl/x509_vfy.h ../cryptlib.h dsa_lib.c
+dsa_lib.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+dsa_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+dsa_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+dsa_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+dsa_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+dsa_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+dsa_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+dsa_lib.o: ../cryptlib.h dsa_lib.c
 dsa_ossl.o: ../../e_os.h ../../include/openssl/asn1.h
 dsa_ossl.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 dsa_ossl.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
@@ -149,40 +162,48 @@ dsa_ossl.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
 dsa_ossl.o: ../../include/openssl/opensslconf.h
 dsa_ossl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 dsa_ossl.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-dsa_ossl.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-dsa_ossl.o: ../cryptlib.h dsa_ossl.c
-dsa_sign.o: ../../e_os.h ../../include/openssl/asn1.h
-dsa_sign.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+dsa_ossl.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+dsa_ossl.o: ../../include/openssl/symhacks.h ../cryptlib.h dsa_ossl.c
+dsa_pmeth.o: ../../e_os.h ../../include/openssl/asn1.h
+dsa_pmeth.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+dsa_pmeth.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
+dsa_pmeth.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
+dsa_pmeth.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+dsa_pmeth.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+dsa_pmeth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+dsa_pmeth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+dsa_pmeth.o: ../../include/openssl/objects.h
+dsa_pmeth.o: ../../include/openssl/opensslconf.h
+dsa_pmeth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+dsa_pmeth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+dsa_pmeth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+dsa_pmeth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+dsa_pmeth.o: ../../include/openssl/x509_vfy.h ../cryptlib.h ../evp/evp_locl.h
+dsa_pmeth.o: dsa_locl.h dsa_pmeth.c
+dsa_prn.o: ../../e_os.h ../../include/openssl/asn1.h
+dsa_prn.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+dsa_prn.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
+dsa_prn.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
+dsa_prn.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+dsa_prn.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+dsa_prn.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+dsa_prn.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+dsa_prn.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+dsa_prn.o: ../cryptlib.h dsa_prn.c
+dsa_sign.o: ../../e_os.h ../../include/openssl/bio.h ../../include/openssl/bn.h
 dsa_sign.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 dsa_sign.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
-dsa_sign.o: ../../include/openssl/err.h ../../include/openssl/fips.h
-dsa_sign.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
+dsa_sign.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
+dsa_sign.o: ../../include/openssl/opensslconf.h
 dsa_sign.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 dsa_sign.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
 dsa_sign.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 dsa_sign.o: ../cryptlib.h dsa_sign.c
-dsa_utl.o: ../../e_os.h ../../include/openssl/asn1.h
-dsa_utl.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
-dsa_utl.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-dsa_utl.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
-dsa_utl.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
-dsa_utl.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-dsa_utl.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-dsa_utl.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-dsa_utl.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-dsa_utl.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-dsa_utl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-dsa_utl.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-dsa_utl.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-dsa_utl.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-dsa_utl.o: ../../include/openssl/x509_vfy.h ../cryptlib.h dsa_utl.c
-dsa_vrf.o: ../../e_os.h ../../include/openssl/asn1.h
-dsa_vrf.o: ../../include/openssl/asn1_mac.h ../../include/openssl/bio.h
-dsa_vrf.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
-dsa_vrf.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
-dsa_vrf.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-dsa_vrf.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+dsa_vrf.o: ../../e_os.h ../../include/openssl/bio.h
+dsa_vrf.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+dsa_vrf.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
+dsa_vrf.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
 dsa_vrf.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-dsa_vrf.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
-dsa_vrf.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-dsa_vrf.o: ../../include/openssl/symhacks.h ../cryptlib.h dsa_vrf.c
+dsa_vrf.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+dsa_vrf.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+dsa_vrf.o: ../cryptlib.h dsa_vrf.c
diff --git a/crypto/dsa/dsa.h b/crypto/dsa/dsa.h
index 702c50d6dc87..a6f6d0b0b243 100644
--- a/crypto/dsa/dsa.h
+++ b/crypto/dsa/dsa.h
@@ -88,8 +88,6 @@
 # define OPENSSL_DSA_MAX_MODULUS_BITS	10000
 #endif
 
-#define OPENSSL_DSA_FIPS_MIN_MODULUS_BITS 1024
-
 #define DSA_FLAG_CACHE_MONT_P	0x01
 #define DSA_FLAG_NO_EXP_CONSTTIME       0x02 /* new with 0.9.7h; the built-in DSA
                                               * implementation now uses constant time
@@ -114,10 +112,6 @@
 
 #define DSA_FLAG_NON_FIPS_ALLOW			0x0400
 
-#ifdef OPENSSL_FIPS
-#define FIPS_DSA_SIZE_T	int
-#endif
-
 #ifdef  __cplusplus
 extern "C" {
 #endif
@@ -139,7 +133,7 @@ struct dsa_method
 	int (*dsa_sign_setup)(DSA *dsa, BN_CTX *ctx_in, BIGNUM **kinvp,
 								BIGNUM **rp);
 	int (*dsa_do_verify)(const unsigned char *dgst, int dgst_len,
-							DSA_SIG *sig, DSA *dsa);
+			     DSA_SIG *sig, DSA *dsa);
 	int (*dsa_mod_exp)(DSA *dsa, BIGNUM *rr, BIGNUM *a1, BIGNUM *p1,
 			BIGNUM *a2, BIGNUM *p2, BIGNUM *m, BN_CTX *ctx,
 			BN_MONT_CTX *in_mont);
@@ -152,7 +146,7 @@ struct dsa_method
 	char *app_data;
 	/* If this is non-NULL, it is used to generate DSA parameters */
 	int (*dsa_paramgen)(DSA *dsa, int bits,
-			unsigned char *seed, int seed_len,
+			const unsigned char *seed, int seed_len,
 			int *counter_ret, unsigned long *h_ret,
 			BN_GENCB *cb);
 	/* If this is non-NULL, it is used to generate DSA keys */
@@ -186,7 +180,6 @@ struct dsa_st
 	ENGINE *engine;
 	};
 
-#define DSAparams_dup(x) ASN1_dup_of_const(DSA,i2d_DSAparams,d2i_DSAparams,x)
 #define d2i_DSAparams_fp(fp,x) (DSA *)ASN1_d2i_fp((char *(*)())DSA_new, \
 		(char *(*)())d2i_DSAparams,(fp),(unsigned char **)(x))
 #define i2d_DSAparams_fp(fp,x) ASN1_i2d_fp(i2d_DSAparams,(fp), \
@@ -195,6 +188,7 @@ struct dsa_st
 #define i2d_DSAparams_bio(bp,x) ASN1_i2d_bio_of_const(DSA,i2d_DSAparams,bp,x)
 
 
+DSA *DSAparams_dup(DSA *x);
 DSA_SIG * DSA_SIG_new(void);
 void	DSA_SIG_free(DSA_SIG *a);
 int	i2d_DSA_SIG(const DSA_SIG *a, unsigned char **pp);
@@ -210,11 +204,6 @@ void	DSA_set_default_method(const DSA_METHOD *);
 const DSA_METHOD *DSA_get_default_method(void);
 int	DSA_set_method(DSA *dsa, const DSA_METHOD *);
 
-#ifdef OPENSSL_FIPS
-DSA *	FIPS_dsa_new(void);
-void	FIPS_dsa_free (DSA *r);
-#endif
-
 DSA *	DSA_new(void);
 DSA *	DSA_new_method(ENGINE *engine);
 void	DSA_free (DSA *r);
@@ -246,7 +235,7 @@ DSA *	DSA_generate_parameters(int bits,
 
 /* New version */
 int	DSA_generate_parameters_ex(DSA *dsa, int bits,
-		unsigned char *seed,int seed_len,
+		const unsigned char *seed,int seed_len,
 		int *counter_ret, unsigned long *h_ret, BN_GENCB *cb);
 
 int	DSA_generate_key(DSA *a);
@@ -275,10 +264,13 @@ int	DSA_print_fp(FILE *bp, const DSA *x, int off);
 DH *DSA_dup_DH(const DSA *r);
 #endif
 
-#ifdef OPENSSL_FIPS
-int FIPS_dsa_sig_encode(unsigned char *out, DSA_SIG *sig);
-int FIPS_dsa_sig_decode(DSA_SIG *sig, const unsigned char *in, int inlen);
-#endif
+#define EVP_PKEY_CTX_set_dsa_paramgen_bits(ctx, nbits) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_DSA, EVP_PKEY_OP_PARAMGEN, \
+				EVP_PKEY_CTRL_DSA_PARAMGEN_BITS, nbits, NULL)
+
+#define	EVP_PKEY_CTRL_DSA_PARAMGEN_BITS		(EVP_PKEY_ALG_CTRL + 1)
+#define	EVP_PKEY_CTRL_DSA_PARAMGEN_Q_BITS	(EVP_PKEY_ALG_CTRL + 2)
+#define	EVP_PKEY_CTRL_DSA_PARAMGEN_MD		(EVP_PKEY_ALG_CTRL + 3)
 
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
@@ -290,33 +282,44 @@ void ERR_load_DSA_strings(void);
 
 /* Function codes. */
 #define DSA_F_D2I_DSA_SIG				 110
+#define DSA_F_DO_DSA_PRINT				 104
 #define DSA_F_DSAPARAMS_PRINT				 100
 #define DSA_F_DSAPARAMS_PRINT_FP			 101
-#define DSA_F_DSA_BUILTIN_KEYGEN			 119
-#define DSA_F_DSA_BUILTIN_PARAMGEN			 118
 #define DSA_F_DSA_DO_SIGN				 112
 #define DSA_F_DSA_DO_VERIFY				 113
-#define DSA_F_DSA_GENERATE_PARAMETERS			 117
+#define DSA_F_DSA_GENERATE_KEY				 124
+#define DSA_F_DSA_GENERATE_PARAMETERS_EX		 123
 #define DSA_F_DSA_NEW_METHOD				 103
-#define DSA_F_DSA_PRINT					 104
+#define DSA_F_DSA_PARAM_DECODE				 119
 #define DSA_F_DSA_PRINT_FP				 105
-#define DSA_F_DSA_SET_DEFAULT_METHOD			 115
-#define DSA_F_DSA_SET_METHOD				 116
+#define DSA_F_DSA_PRIV_DECODE				 115
+#define DSA_F_DSA_PRIV_ENCODE				 116
+#define DSA_F_DSA_PUB_DECODE				 117
+#define DSA_F_DSA_PUB_ENCODE				 118
 #define DSA_F_DSA_SIGN					 106
 #define DSA_F_DSA_SIGN_SETUP				 107
 #define DSA_F_DSA_SIG_NEW				 109
+#define DSA_F_DSA_SIG_PRINT				 125
 #define DSA_F_DSA_VERIFY				 108
 #define DSA_F_I2D_DSA_SIG				 111
+#define DSA_F_OLD_DSA_PRIV_DECODE			 122
+#define DSA_F_PKEY_DSA_CTRL				 120
+#define DSA_F_PKEY_DSA_KEYGEN				 121
 #define DSA_F_SIG_CB					 114
 
 /* Reason codes. */
 #define DSA_R_BAD_Q_VALUE				 102
+#define DSA_R_BN_DECODE_ERROR				 108
+#define DSA_R_BN_ERROR					 109
 #define DSA_R_DATA_TOO_LARGE_FOR_KEY_SIZE		 100
-#define DSA_R_KEY_SIZE_TOO_SMALL			 106
+#define DSA_R_DECODE_ERROR				 104
+#define DSA_R_INVALID_DIGEST_TYPE			 106
 #define DSA_R_MISSING_PARAMETERS			 101
 #define DSA_R_MODULUS_TOO_LARGE				 103
-#define DSA_R_NON_FIPS_METHOD				 104
-#define DSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE	 105
+#define DSA_R_NEED_NEW_SETUP_VALUES			 110
+#define DSA_R_NON_FIPS_DSA_METHOD			 111
+#define DSA_R_NO_PARAMETERS_SET				 107
+#define DSA_R_PARAMETER_ENCODING_ERROR			 105
 
 #ifdef  __cplusplus
 }
diff --git a/crypto/dsa/dsa_ameth.c b/crypto/dsa/dsa_ameth.c
new file mode 100644
index 000000000000..376156ec5ef3
--- /dev/null
+++ b/crypto/dsa/dsa_ameth.c
@@ -0,0 +1,704 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/x509.h>
+#include <openssl/asn1.h>
+#include <openssl/dsa.h>
+#include <openssl/bn.h>
+#ifndef OPENSSL_NO_CMS
+#include <openssl/cms.h>
+#endif
+#include "asn1_locl.h"
+
+static int dsa_pub_decode(EVP_PKEY *pkey, X509_PUBKEY *pubkey)
+	{
+	const unsigned char *p, *pm;
+	int pklen, pmlen;
+	int ptype;
+	void *pval;
+	ASN1_STRING *pstr;
+	X509_ALGOR *palg;
+	ASN1_INTEGER *public_key = NULL;
+
+	DSA *dsa = NULL;
+
+	if (!X509_PUBKEY_get0_param(NULL, &p, &pklen, &palg, pubkey))
+		return 0;
+	X509_ALGOR_get0(NULL, &ptype, &pval, palg);
+
+
+	if (ptype == V_ASN1_SEQUENCE)
+		{
+		pstr = pval;	
+		pm = pstr->data;
+		pmlen = pstr->length;
+
+		if (!(dsa = d2i_DSAparams(NULL, &pm, pmlen)))
+			{
+			DSAerr(DSA_F_DSA_PUB_DECODE, DSA_R_DECODE_ERROR);
+			goto err;
+			}
+
+		}
+	else if ((ptype == V_ASN1_NULL) || (ptype == V_ASN1_UNDEF))
+		{
+		if (!(dsa = DSA_new()))
+			{
+			DSAerr(DSA_F_DSA_PUB_DECODE, ERR_R_MALLOC_FAILURE);
+			goto err;
+			}
+		}
+	else
+		{
+		DSAerr(DSA_F_DSA_PUB_DECODE, DSA_R_PARAMETER_ENCODING_ERROR);
+		goto err;
+		}
+
+	if (!(public_key=d2i_ASN1_INTEGER(NULL, &p, pklen)))
+		{
+		DSAerr(DSA_F_DSA_PUB_DECODE, DSA_R_DECODE_ERROR);
+		goto err;
+		}
+
+	if (!(dsa->pub_key = ASN1_INTEGER_to_BN(public_key, NULL)))
+		{
+		DSAerr(DSA_F_DSA_PUB_DECODE, DSA_R_BN_DECODE_ERROR);
+		goto err;
+		}
+
+	ASN1_INTEGER_free(public_key);
+	EVP_PKEY_assign_DSA(pkey, dsa);
+	return 1;
+
+	err:
+	if (public_key)
+		ASN1_INTEGER_free(public_key);
+	if (dsa)
+		DSA_free(dsa);
+	return 0;
+
+	}
+
+static int dsa_pub_encode(X509_PUBKEY *pk, const EVP_PKEY *pkey)
+	{
+	DSA *dsa;
+	void *pval = NULL;
+	int ptype;
+	unsigned char *penc = NULL;
+	int penclen;
+
+	dsa=pkey->pkey.dsa;
+	if (pkey->save_parameters && dsa->p && dsa->q && dsa->g)
+		{
+		ASN1_STRING *str;
+		str = ASN1_STRING_new();
+		str->length = i2d_DSAparams(dsa, &str->data);
+		if (str->length <= 0)
+			{
+			DSAerr(DSA_F_DSA_PUB_ENCODE, ERR_R_MALLOC_FAILURE);
+			goto err;
+			}
+		pval = str;
+		ptype = V_ASN1_SEQUENCE;
+		}
+	else
+		ptype = V_ASN1_UNDEF;
+
+	dsa->write_params=0;
+
+	penclen = i2d_DSAPublicKey(dsa, &penc);
+
+	if (penclen <= 0)
+		{
+		DSAerr(DSA_F_DSA_PUB_ENCODE, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	if (X509_PUBKEY_set0_param(pk, OBJ_nid2obj(EVP_PKEY_DSA),
+				ptype, pval, penc, penclen))
+		return 1;
+
+	err:
+	if (penc)
+		OPENSSL_free(penc);
+	if (pval)
+		ASN1_STRING_free(pval);
+
+	return 0;
+	}
+
+/* In PKCS#8 DSA: you just get a private key integer and parameters in the
+ * AlgorithmIdentifier the pubkey must be recalculated.
+ */
+	
+static int dsa_priv_decode(EVP_PKEY *pkey, PKCS8_PRIV_KEY_INFO *p8)
+	{
+	const unsigned char *p, *pm;
+	int pklen, pmlen;
+	int ptype;
+	void *pval;
+	ASN1_STRING *pstr;
+	X509_ALGOR *palg;
+	ASN1_INTEGER *privkey = NULL;
+	BN_CTX *ctx = NULL;
+
+	STACK_OF(ASN1_TYPE) *ndsa = NULL;
+	DSA *dsa = NULL;
+
+	if (!PKCS8_pkey_get0(NULL, &p, &pklen, &palg, p8))
+		return 0;
+	X509_ALGOR_get0(NULL, &ptype, &pval, palg);
+
+	/* Check for broken DSA PKCS#8, UGH! */
+	if (*p == (V_ASN1_SEQUENCE|V_ASN1_CONSTRUCTED))
+		{
+		ASN1_TYPE *t1, *t2;
+	    	if(!(ndsa = d2i_ASN1_SEQUENCE_ANY(NULL, &p, pklen)))
+			goto decerr;
+		if (sk_ASN1_TYPE_num(ndsa) != 2)
+			goto decerr;
+		/* Handle Two broken types:
+	    	 * SEQUENCE {parameters, priv_key}
+		 * SEQUENCE {pub_key, priv_key}
+		 */
+
+		t1 = sk_ASN1_TYPE_value(ndsa, 0);
+		t2 = sk_ASN1_TYPE_value(ndsa, 1);
+		if (t1->type == V_ASN1_SEQUENCE)
+			{
+			p8->broken = PKCS8_EMBEDDED_PARAM;
+			pval = t1->value.ptr;
+			}
+		else if (ptype == V_ASN1_SEQUENCE)
+			p8->broken = PKCS8_NS_DB;
+		else
+			goto decerr;
+
+		if (t2->type != V_ASN1_INTEGER)
+			goto decerr;
+
+		privkey = t2->value.integer;
+		}
+	else
+		{
+		const unsigned char *q = p;
+		if (!(privkey=d2i_ASN1_INTEGER(NULL, &p, pklen)))
+			goto decerr;
+		if (privkey->type == V_ASN1_NEG_INTEGER)
+			{
+			p8->broken = PKCS8_NEG_PRIVKEY;
+			ASN1_INTEGER_free(privkey);
+			if (!(privkey=d2i_ASN1_UINTEGER(NULL, &q, pklen)))
+				goto decerr;
+			}
+		if (ptype != V_ASN1_SEQUENCE)
+			goto decerr;
+		}
+
+	pstr = pval;	
+	pm = pstr->data;
+	pmlen = pstr->length;
+	if (!(dsa = d2i_DSAparams(NULL, &pm, pmlen)))
+		goto decerr;
+	/* We have parameters now set private key */
+	if (!(dsa->priv_key = ASN1_INTEGER_to_BN(privkey, NULL)))
+		{
+		DSAerr(DSA_F_DSA_PRIV_DECODE,DSA_R_BN_ERROR);
+		goto dsaerr;
+		}
+	/* Calculate public key */
+	if (!(dsa->pub_key = BN_new()))
+		{
+		DSAerr(DSA_F_DSA_PRIV_DECODE, ERR_R_MALLOC_FAILURE);
+		goto dsaerr;
+		}
+	if (!(ctx = BN_CTX_new()))
+		{
+		DSAerr(DSA_F_DSA_PRIV_DECODE, ERR_R_MALLOC_FAILURE);
+		goto dsaerr;
+		}
+			
+	if (!BN_mod_exp(dsa->pub_key, dsa->g, dsa->priv_key, dsa->p, ctx))
+		{
+		DSAerr(DSA_F_DSA_PRIV_DECODE,DSA_R_BN_ERROR);
+		goto dsaerr;
+		}
+
+	EVP_PKEY_assign_DSA(pkey, dsa);
+	BN_CTX_free (ctx);
+	if(ndsa)
+		sk_ASN1_TYPE_pop_free(ndsa, ASN1_TYPE_free);
+	else
+		ASN1_INTEGER_free(privkey);
+
+	return 1;
+
+	decerr:
+	DSAerr(DSA_F_DSA_PRIV_DECODE, EVP_R_DECODE_ERROR);
+	dsaerr:
+	BN_CTX_free (ctx);
+	if (privkey)
+		ASN1_INTEGER_free(privkey);
+	sk_ASN1_TYPE_pop_free(ndsa, ASN1_TYPE_free);
+	DSA_free(dsa);
+	return 0;
+	}
+
+static int dsa_priv_encode(PKCS8_PRIV_KEY_INFO *p8, const EVP_PKEY *pkey)
+{
+	ASN1_STRING *params = NULL;
+	ASN1_INTEGER *prkey = NULL;
+	unsigned char *dp = NULL;
+	int dplen;
+
+	params = ASN1_STRING_new();
+
+	if (!params)
+		{
+		DSAerr(DSA_F_DSA_PRIV_ENCODE,ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	params->length = i2d_DSAparams(pkey->pkey.dsa, &params->data);
+	if (params->length <= 0)
+		{
+		DSAerr(DSA_F_DSA_PRIV_ENCODE,ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+	params->type = V_ASN1_SEQUENCE;
+
+	/* Get private key into integer */
+	prkey = BN_to_ASN1_INTEGER(pkey->pkey.dsa->priv_key, NULL);
+
+	if (!prkey)
+		{
+		DSAerr(DSA_F_DSA_PRIV_ENCODE,DSA_R_BN_ERROR);
+		goto err;
+		}
+
+	dplen = i2d_ASN1_INTEGER(prkey, &dp);
+
+	ASN1_INTEGER_free(prkey);
+
+	if (!PKCS8_pkey_set0(p8, OBJ_nid2obj(NID_dsa), 0,
+				V_ASN1_SEQUENCE, params, dp, dplen))
+		goto err;
+
+	return 1;
+
+err:
+	if (dp != NULL)
+		OPENSSL_free(dp);
+	if (params != NULL)
+		ASN1_STRING_free(params);
+	if (prkey != NULL)
+		ASN1_INTEGER_free(prkey);
+	return 0;
+}
+
+static int int_dsa_size(const EVP_PKEY *pkey)
+	{
+	return(DSA_size(pkey->pkey.dsa));
+	}
+
+static int dsa_bits(const EVP_PKEY *pkey)
+	{
+	return BN_num_bits(pkey->pkey.dsa->p);
+	}
+
+static int dsa_missing_parameters(const EVP_PKEY *pkey)
+	{
+	DSA *dsa;
+	dsa=pkey->pkey.dsa;
+	if ((dsa->p == NULL) || (dsa->q == NULL) || (dsa->g == NULL))
+			return 1;
+	return 0;
+	}
+
+static int dsa_copy_parameters(EVP_PKEY *to, const EVP_PKEY *from)
+	{
+	BIGNUM *a;
+
+	if ((a=BN_dup(from->pkey.dsa->p)) == NULL)
+		return 0;
+	if (to->pkey.dsa->p != NULL)
+		BN_free(to->pkey.dsa->p);
+	to->pkey.dsa->p=a;
+
+	if ((a=BN_dup(from->pkey.dsa->q)) == NULL)
+		return 0;
+	if (to->pkey.dsa->q != NULL)
+		BN_free(to->pkey.dsa->q);
+	to->pkey.dsa->q=a;
+
+	if ((a=BN_dup(from->pkey.dsa->g)) == NULL)
+		return 0;
+	if (to->pkey.dsa->g != NULL)
+		BN_free(to->pkey.dsa->g);
+	to->pkey.dsa->g=a;
+	return 1;
+	}
+
+static int dsa_cmp_parameters(const EVP_PKEY *a, const EVP_PKEY *b)
+	{
+	if (	BN_cmp(a->pkey.dsa->p,b->pkey.dsa->p) ||
+		BN_cmp(a->pkey.dsa->q,b->pkey.dsa->q) ||
+		BN_cmp(a->pkey.dsa->g,b->pkey.dsa->g))
+		return 0;
+	else
+		return 1;
+	}
+
+static int dsa_pub_cmp(const EVP_PKEY *a, const EVP_PKEY *b)
+	{
+	if (BN_cmp(b->pkey.dsa->pub_key,a->pkey.dsa->pub_key) != 0)
+		return 0;
+	else
+		return 1;
+	}
+
+static void int_dsa_free(EVP_PKEY *pkey)
+	{
+	DSA_free(pkey->pkey.dsa);
+	}
+
+static void update_buflen(const BIGNUM *b, size_t *pbuflen)
+	{
+	size_t i;
+	if (!b)
+		return;
+	if (*pbuflen < (i = (size_t)BN_num_bytes(b)))
+			*pbuflen = i;
+	}
+
+static int do_dsa_print(BIO *bp, const DSA *x, int off, int ptype)
+	{
+	unsigned char *m=NULL;
+	int ret=0;
+	size_t buf_len=0;
+	const char *ktype = NULL;
+
+	const BIGNUM *priv_key, *pub_key;
+
+	if (ptype == 2)
+		priv_key = x->priv_key;
+	else
+		priv_key = NULL;
+
+	if (ptype > 0)
+		pub_key = x->pub_key;
+	else
+		pub_key = NULL;
+
+	if (ptype == 2)
+		ktype = "Private-Key";
+	else if (ptype == 1)
+		ktype = "Public-Key";
+	else
+		ktype = "DSA-Parameters";
+
+	update_buflen(x->p, &buf_len);
+	update_buflen(x->q, &buf_len);
+	update_buflen(x->g, &buf_len);
+	update_buflen(priv_key, &buf_len);
+	update_buflen(pub_key, &buf_len);
+
+	m=(unsigned char *)OPENSSL_malloc(buf_len+10);
+	if (m == NULL)
+		{
+		DSAerr(DSA_F_DO_DSA_PRINT,ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	if (priv_key)
+		{
+		if(!BIO_indent(bp,off,128))
+		   goto err;
+		if (BIO_printf(bp,"%s: (%d bit)\n",ktype, BN_num_bits(x->p))
+			<= 0) goto err;
+		}
+
+	if (!ASN1_bn_print(bp,"priv:",priv_key,m,off))
+		goto err;
+	if (!ASN1_bn_print(bp,"pub: ",pub_key,m,off))
+		goto err;
+	if (!ASN1_bn_print(bp,"P:   ",x->p,m,off)) goto err;
+	if (!ASN1_bn_print(bp,"Q:   ",x->q,m,off)) goto err;
+	if (!ASN1_bn_print(bp,"G:   ",x->g,m,off)) goto err;
+	ret=1;
+err:
+	if (m != NULL) OPENSSL_free(m);
+	return(ret);
+	}
+
+static int dsa_param_decode(EVP_PKEY *pkey,
+					const unsigned char **pder, int derlen)
+	{
+	DSA *dsa;
+	if (!(dsa = d2i_DSAparams(NULL, pder, derlen)))
+		{
+		DSAerr(DSA_F_DSA_PARAM_DECODE, ERR_R_DSA_LIB);
+		return 0;
+		}
+	EVP_PKEY_assign_DSA(pkey, dsa);
+	return 1;
+	}
+
+static int dsa_param_encode(const EVP_PKEY *pkey, unsigned char **pder)
+	{
+	return i2d_DSAparams(pkey->pkey.dsa, pder);
+	}
+
+static int dsa_param_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_dsa_print(bp, pkey->pkey.dsa, indent, 0);
+	}
+
+static int dsa_pub_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_dsa_print(bp, pkey->pkey.dsa, indent, 1);
+	}
+
+
+static int dsa_priv_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_dsa_print(bp, pkey->pkey.dsa, indent, 2);
+	}
+
+static int old_dsa_priv_decode(EVP_PKEY *pkey,
+					const unsigned char **pder, int derlen)
+	{
+	DSA *dsa;
+	if (!(dsa = d2i_DSAPrivateKey (NULL, pder, derlen)))
+		{
+		DSAerr(DSA_F_OLD_DSA_PRIV_DECODE, ERR_R_DSA_LIB);
+		return 0;
+		}
+	EVP_PKEY_assign_DSA(pkey, dsa);
+	return 1;
+	}
+
+static int old_dsa_priv_encode(const EVP_PKEY *pkey, unsigned char **pder)
+	{
+	return i2d_DSAPrivateKey(pkey->pkey.dsa, pder);
+	}
+
+static int dsa_sig_print(BIO *bp, const X509_ALGOR *sigalg,
+					const ASN1_STRING *sig,
+					int indent, ASN1_PCTX *pctx)
+	{
+	DSA_SIG *dsa_sig;
+	const unsigned char *p;
+	if (!sig)
+		{
+		if (BIO_puts(bp, "\n") <= 0)
+			return 0;
+		else
+			return 1;
+		}
+	p = sig->data;
+	dsa_sig = d2i_DSA_SIG(NULL, &p, sig->length);
+	if (dsa_sig)
+		{
+		int rv = 0;
+		size_t buf_len = 0;
+		unsigned char *m=NULL;
+		update_buflen(dsa_sig->r, &buf_len);
+		update_buflen(dsa_sig->s, &buf_len);
+		m = OPENSSL_malloc(buf_len+10);
+		if (m == NULL)
+			{
+			DSAerr(DSA_F_DSA_SIG_PRINT,ERR_R_MALLOC_FAILURE);
+			goto err;
+			}
+
+		if (BIO_write(bp, "\n", 1) != 1)
+			goto err;
+
+		if (!ASN1_bn_print(bp,"r:   ",dsa_sig->r,m,indent))
+			goto err;
+		if (!ASN1_bn_print(bp,"s:   ",dsa_sig->s,m,indent))
+			goto err;
+		rv = 1;
+		err:
+		if (m)
+			OPENSSL_free(m);
+		DSA_SIG_free(dsa_sig);
+		return rv;
+		}
+	return X509_signature_dump(bp, sig, indent);
+	}
+
+static int dsa_pkey_ctrl(EVP_PKEY *pkey, int op, long arg1, void *arg2)
+	{
+	switch (op)
+		{
+		case ASN1_PKEY_CTRL_PKCS7_SIGN:
+		if (arg1 == 0)
+			{
+			int snid, hnid;
+			X509_ALGOR *alg1, *alg2;
+			PKCS7_SIGNER_INFO_get0_algs(arg2, NULL, &alg1, &alg2);
+			if (alg1 == NULL || alg1->algorithm == NULL)
+				return -1;
+			hnid = OBJ_obj2nid(alg1->algorithm);
+			if (hnid == NID_undef)
+				return -1;
+			if (!OBJ_find_sigid_by_algs(&snid, hnid, EVP_PKEY_id(pkey)))
+				return -1; 
+			X509_ALGOR_set0(alg2, OBJ_nid2obj(snid), V_ASN1_UNDEF, 0);
+			}
+		return 1;
+#ifndef OPENSSL_NO_CMS
+		case ASN1_PKEY_CTRL_CMS_SIGN:
+		if (arg1 == 0)
+			{
+			int snid, hnid;
+			X509_ALGOR *alg1, *alg2;
+			CMS_SignerInfo_get0_algs(arg2, NULL, NULL, &alg1, &alg2);
+			if (alg1 == NULL || alg1->algorithm == NULL)
+				return -1;
+			hnid = OBJ_obj2nid(alg1->algorithm);
+			if (hnid == NID_undef)
+				return -1;
+			if (!OBJ_find_sigid_by_algs(&snid, hnid, EVP_PKEY_id(pkey)))
+				return -1; 
+			X509_ALGOR_set0(alg2, OBJ_nid2obj(snid), V_ASN1_UNDEF, 0);
+			}
+		return 1;
+#endif
+
+		case ASN1_PKEY_CTRL_DEFAULT_MD_NID:
+		*(int *)arg2 = NID_sha1;
+		return 2;
+
+		default:
+		return -2;
+
+		}
+
+	}
+
+/* NB these are sorted in pkey_id order, lowest first */
+
+const EVP_PKEY_ASN1_METHOD dsa_asn1_meths[] = 
+	{
+
+		{
+		EVP_PKEY_DSA2,
+		EVP_PKEY_DSA,
+		ASN1_PKEY_ALIAS
+		},
+
+		{
+		EVP_PKEY_DSA1,
+		EVP_PKEY_DSA,
+		ASN1_PKEY_ALIAS
+		},
+
+		{
+		EVP_PKEY_DSA4,
+		EVP_PKEY_DSA,
+		ASN1_PKEY_ALIAS
+		},
+
+		{
+		EVP_PKEY_DSA3,
+		EVP_PKEY_DSA,
+		ASN1_PKEY_ALIAS
+		},
+
+		{
+		EVP_PKEY_DSA,
+		EVP_PKEY_DSA,
+		0,
+
+		"DSA",
+		"OpenSSL DSA method",
+
+		dsa_pub_decode,
+		dsa_pub_encode,
+		dsa_pub_cmp,
+		dsa_pub_print,
+
+		dsa_priv_decode,
+		dsa_priv_encode,
+		dsa_priv_print,
+
+		int_dsa_size,
+		dsa_bits,
+
+		dsa_param_decode,
+		dsa_param_encode,
+		dsa_missing_parameters,
+		dsa_copy_parameters,
+		dsa_cmp_parameters,
+		dsa_param_print,
+		dsa_sig_print,
+
+		int_dsa_free,
+		dsa_pkey_ctrl,
+		old_dsa_priv_decode,
+		old_dsa_priv_encode
+		}
+	};
+
diff --git a/crypto/dsa/dsa_asn1.c b/crypto/dsa/dsa_asn1.c
index bc7d7a02a347..60585343746d 100644
--- a/crypto/dsa/dsa_asn1.c
+++ b/crypto/dsa/dsa_asn1.c
@@ -3,7 +3,7 @@
  * project 2000.
  */
 /* ====================================================================
- * Copyright (c) 2000 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2000-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -61,25 +61,24 @@
 #include <openssl/dsa.h>
 #include <openssl/asn1.h>
 #include <openssl/asn1t.h>
-#include <openssl/bn.h>
 #include <openssl/rand.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 
 /* Override the default new methods */
-static int sig_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int sig_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+								void *exarg)
 {
 	if(operation == ASN1_OP_NEW_PRE) {
 		DSA_SIG *sig;
 		sig = OPENSSL_malloc(sizeof(DSA_SIG));
+		if (!sig)
+			{
+			DSAerr(DSA_F_SIG_CB, ERR_R_MALLOC_FAILURE);
+			return 0;
+			}
 		sig->r = NULL;
 		sig->s = NULL;
 		*pval = (ASN1_VALUE *)sig;
-		if(sig) return 2;
-		DSAerr(DSA_F_SIG_CB, ERR_R_MALLOC_FAILURE);
-		return 0;
+		return 2;
 	}
 	return 1;
 }
@@ -89,10 +88,11 @@ ASN1_SEQUENCE_cb(DSA_SIG, sig_cb) = {
 	ASN1_SIMPLE(DSA_SIG, s, CBIGNUM)
 } ASN1_SEQUENCE_END_cb(DSA_SIG, DSA_SIG)
 
-IMPLEMENT_ASN1_ENCODE_FUNCTIONS_const_fname(DSA_SIG,DSA_SIG,DSA_SIG)
+IMPLEMENT_ASN1_ENCODE_FUNCTIONS_const_fname(DSA_SIG, DSA_SIG, DSA_SIG)
 
 /* Override the default free and new methods */
-static int dsa_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int dsa_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+							void *exarg)
 {
 	if(operation == ASN1_OP_NEW_PRE) {
 		*pval = (ASN1_VALUE *)DSA_new();
@@ -145,17 +145,15 @@ ASN1_CHOICE_cb(DSAPublicKey, dsa_cb) = {
 
 IMPLEMENT_ASN1_ENCODE_FUNCTIONS_const_fname(DSA, DSAPublicKey, DSAPublicKey)
 
+DSA *DSAparams_dup(DSA *dsa)
+	{
+	return ASN1_item_dup(ASN1_ITEM_rptr(DSAparams), dsa);
+	}
+
 int DSA_sign(int type, const unsigned char *dgst, int dlen, unsigned char *sig,
 	     unsigned int *siglen, DSA *dsa)
 	{
 	DSA_SIG *s;
-#ifdef OPENSSL_FIPS
-	if(FIPS_mode() && !(dsa->flags & DSA_FLAG_NON_FIPS_ALLOW))
-		{
-		DSAerr(DSA_F_DSA_SIGN, DSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
-		return 0;
-		}
-#endif
 	RAND_seed(dgst, dlen);
 	s=DSA_do_sign(dgst,dlen,dsa);
 	if (s == NULL)
@@ -168,28 +166,6 @@ int DSA_sign(int type, const unsigned char *dgst, int dlen, unsigned char *sig,
 	return(1);
 	}
 
-int DSA_size(const DSA *r)
-	{
-	int ret,i;
-	ASN1_INTEGER bs;
-	unsigned char buf[4];	/* 4 bytes looks really small.
-				   However, i2d_ASN1_INTEGER() will not look
-				   beyond the first byte, as long as the second
-				   parameter is NULL. */
-
-	i=BN_num_bits(r->q);
-	bs.length=(i+7)/8;
-	bs.data=buf;
-	bs.type=V_ASN1_INTEGER;
-	/* If the top bit is set the asn1 encoding is 1 larger. */
-	buf[0]=0xff;	
-
-	i=i2d_ASN1_INTEGER(&bs,NULL);
-	i+=i; /* r and s */
-	ret=ASN1_object_size(1,i,V_ASN1_SEQUENCE);
-	return(ret);
-	}
-
 /* data has already been hashed (probably with SHA or SHA-1). */
 /* returns
  *      1: correct signature
@@ -201,13 +177,6 @@ int DSA_verify(int type, const unsigned char *dgst, int dgst_len,
 	{
 	DSA_SIG *s;
 	int ret=-1;
-#ifdef OPENSSL_FIPS
-	if(FIPS_mode() && !(dsa->flags & DSA_FLAG_NON_FIPS_ALLOW))
-		{
-		DSAerr(DSA_F_DSA_VERIFY, DSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
-		return 0;
-		}
-#endif
 
 	s = DSA_SIG_new();
 	if (s == NULL) return(ret);
@@ -217,4 +186,3 @@ err:
 	DSA_SIG_free(s);
 	return(ret);
 	}
-
diff --git a/crypto/dsa/dsa_err.c b/crypto/dsa/dsa_err.c
index 872839af944b..00545b7b9fa9 100644
--- a/crypto/dsa/dsa_err.c
+++ b/crypto/dsa/dsa_err.c
@@ -1,6 +1,6 @@
 /* crypto/dsa/dsa_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -71,23 +71,29 @@
 static ERR_STRING_DATA DSA_str_functs[]=
 	{
 {ERR_FUNC(DSA_F_D2I_DSA_SIG),	"d2i_DSA_SIG"},
+{ERR_FUNC(DSA_F_DO_DSA_PRINT),	"DO_DSA_PRINT"},
 {ERR_FUNC(DSA_F_DSAPARAMS_PRINT),	"DSAparams_print"},
 {ERR_FUNC(DSA_F_DSAPARAMS_PRINT_FP),	"DSAparams_print_fp"},
-{ERR_FUNC(DSA_F_DSA_BUILTIN_KEYGEN),	"DSA_BUILTIN_KEYGEN"},
-{ERR_FUNC(DSA_F_DSA_BUILTIN_PARAMGEN),	"DSA_BUILTIN_PARAMGEN"},
 {ERR_FUNC(DSA_F_DSA_DO_SIGN),	"DSA_do_sign"},
 {ERR_FUNC(DSA_F_DSA_DO_VERIFY),	"DSA_do_verify"},
-{ERR_FUNC(DSA_F_DSA_GENERATE_PARAMETERS),	"DSA_generate_parameters"},
+{ERR_FUNC(DSA_F_DSA_GENERATE_KEY),	"DSA_generate_key"},
+{ERR_FUNC(DSA_F_DSA_GENERATE_PARAMETERS_EX),	"DSA_generate_parameters_ex"},
 {ERR_FUNC(DSA_F_DSA_NEW_METHOD),	"DSA_new_method"},
-{ERR_FUNC(DSA_F_DSA_PRINT),	"DSA_print"},
+{ERR_FUNC(DSA_F_DSA_PARAM_DECODE),	"DSA_PARAM_DECODE"},
 {ERR_FUNC(DSA_F_DSA_PRINT_FP),	"DSA_print_fp"},
-{ERR_FUNC(DSA_F_DSA_SET_DEFAULT_METHOD),	"DSA_set_default_method"},
-{ERR_FUNC(DSA_F_DSA_SET_METHOD),	"DSA_set_method"},
+{ERR_FUNC(DSA_F_DSA_PRIV_DECODE),	"DSA_PRIV_DECODE"},
+{ERR_FUNC(DSA_F_DSA_PRIV_ENCODE),	"DSA_PRIV_ENCODE"},
+{ERR_FUNC(DSA_F_DSA_PUB_DECODE),	"DSA_PUB_DECODE"},
+{ERR_FUNC(DSA_F_DSA_PUB_ENCODE),	"DSA_PUB_ENCODE"},
 {ERR_FUNC(DSA_F_DSA_SIGN),	"DSA_sign"},
 {ERR_FUNC(DSA_F_DSA_SIGN_SETUP),	"DSA_sign_setup"},
 {ERR_FUNC(DSA_F_DSA_SIG_NEW),	"DSA_SIG_new"},
+{ERR_FUNC(DSA_F_DSA_SIG_PRINT),	"DSA_SIG_PRINT"},
 {ERR_FUNC(DSA_F_DSA_VERIFY),	"DSA_verify"},
 {ERR_FUNC(DSA_F_I2D_DSA_SIG),	"i2d_DSA_SIG"},
+{ERR_FUNC(DSA_F_OLD_DSA_PRIV_DECODE),	"OLD_DSA_PRIV_DECODE"},
+{ERR_FUNC(DSA_F_PKEY_DSA_CTRL),	"PKEY_DSA_CTRL"},
+{ERR_FUNC(DSA_F_PKEY_DSA_KEYGEN),	"PKEY_DSA_KEYGEN"},
 {ERR_FUNC(DSA_F_SIG_CB),	"SIG_CB"},
 {0,NULL}
 	};
@@ -95,12 +101,17 @@ static ERR_STRING_DATA DSA_str_functs[]=
 static ERR_STRING_DATA DSA_str_reasons[]=
 	{
 {ERR_REASON(DSA_R_BAD_Q_VALUE)           ,"bad q value"},
+{ERR_REASON(DSA_R_BN_DECODE_ERROR)       ,"bn decode error"},
+{ERR_REASON(DSA_R_BN_ERROR)              ,"bn error"},
 {ERR_REASON(DSA_R_DATA_TOO_LARGE_FOR_KEY_SIZE),"data too large for key size"},
-{ERR_REASON(DSA_R_KEY_SIZE_TOO_SMALL)    ,"key size too small"},
+{ERR_REASON(DSA_R_DECODE_ERROR)          ,"decode error"},
+{ERR_REASON(DSA_R_INVALID_DIGEST_TYPE)   ,"invalid digest type"},
 {ERR_REASON(DSA_R_MISSING_PARAMETERS)    ,"missing parameters"},
 {ERR_REASON(DSA_R_MODULUS_TOO_LARGE)     ,"modulus too large"},
-{ERR_REASON(DSA_R_NON_FIPS_METHOD)       ,"non fips method"},
-{ERR_REASON(DSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE),"operation not allowed in fips mode"},
+{ERR_REASON(DSA_R_NEED_NEW_SETUP_VALUES) ,"need new setup values"},
+{ERR_REASON(DSA_R_NON_FIPS_DSA_METHOD)   ,"non fips dsa method"},
+{ERR_REASON(DSA_R_NO_PARAMETERS_SET)     ,"no parameters set"},
+{ERR_REASON(DSA_R_PARAMETER_ENCODING_ERROR),"parameter encoding error"},
 {0,NULL}
 	};
 
diff --git a/crypto/dsa/dsa_gen.c b/crypto/dsa/dsa_gen.c
index 7a9d1881cc48..c398761d0ddd 100644
--- a/crypto/dsa/dsa_gen.c
+++ b/crypto/dsa/dsa_gen.c
@@ -74,69 +74,109 @@
 #ifndef OPENSSL_NO_SHA
 
 #include <stdio.h>
-#include <time.h>
 #include "cryptlib.h"
 #include <openssl/evp.h>
 #include <openssl/bn.h>
-#include <openssl/dsa.h>
 #include <openssl/rand.h>
 #include <openssl/sha.h>
+#include "dsa_locl.h"
 
-#ifndef OPENSSL_FIPS
-
-static int dsa_builtin_paramgen(DSA *ret, int bits,
-		unsigned char *seed_in, int seed_len,
-		int *counter_ret, unsigned long *h_ret, BN_GENCB *cb);
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 
 int DSA_generate_parameters_ex(DSA *ret, int bits,
-		unsigned char *seed_in, int seed_len,
+		const unsigned char *seed_in, int seed_len,
 		int *counter_ret, unsigned long *h_ret, BN_GENCB *cb)
 	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(ret->meth->flags & DSA_FLAG_FIPS_METHOD)
+			&& !(ret->flags & DSA_FLAG_NON_FIPS_ALLOW))
+		{
+		DSAerr(DSA_F_DSA_GENERATE_PARAMETERS_EX, DSA_R_NON_FIPS_DSA_METHOD);
+		return 0;
+		}
+#endif
 	if(ret->meth->dsa_paramgen)
 		return ret->meth->dsa_paramgen(ret, bits, seed_in, seed_len,
 				counter_ret, h_ret, cb);
-	return dsa_builtin_paramgen(ret, bits, seed_in, seed_len,
-			counter_ret, h_ret, cb);
+#ifdef OPENSSL_FIPS
+	else if (FIPS_mode())
+		{
+		return FIPS_dsa_generate_parameters_ex(ret, bits, 
+							seed_in, seed_len,
+							counter_ret, h_ret, cb);
+		}
+#endif
+	else
+		{
+		const EVP_MD *evpmd;
+		size_t qbits = bits >= 2048 ? 256 : 160;
+
+		if (bits >= 2048)
+			{
+			qbits = 256;
+			evpmd = EVP_sha256();
+			}
+		else
+			{
+			qbits = 160;
+			evpmd = EVP_sha1();
+			}
+
+		return dsa_builtin_paramgen(ret, bits, qbits, evpmd,
+			seed_in, seed_len, NULL, counter_ret, h_ret, cb);
+		}
 	}
 
-static int dsa_builtin_paramgen(DSA *ret, int bits,
-		unsigned char *seed_in, int seed_len,
-		int *counter_ret, unsigned long *h_ret, BN_GENCB *cb)
+int dsa_builtin_paramgen(DSA *ret, size_t bits, size_t qbits,
+	const EVP_MD *evpmd, const unsigned char *seed_in, size_t seed_len,
+	unsigned char *seed_out,
+	int *counter_ret, unsigned long *h_ret, BN_GENCB *cb)
 	{
 	int ok=0;
-	unsigned char seed[SHA_DIGEST_LENGTH];
-	unsigned char md[SHA_DIGEST_LENGTH];
-	unsigned char buf[SHA_DIGEST_LENGTH],buf2[SHA_DIGEST_LENGTH];
+	unsigned char seed[SHA256_DIGEST_LENGTH];
+	unsigned char md[SHA256_DIGEST_LENGTH];
+	unsigned char buf[SHA256_DIGEST_LENGTH],buf2[SHA256_DIGEST_LENGTH];
 	BIGNUM *r0,*W,*X,*c,*test;
 	BIGNUM *g=NULL,*q=NULL,*p=NULL;
 	BN_MONT_CTX *mont=NULL;
-	int k,n=0,i,m=0;
+	int i, k, n=0, m=0, qsize = qbits >> 3;
 	int counter=0;
 	int r=0;
 	BN_CTX *ctx=NULL;
 	unsigned int h=2;
 
-	if (bits < 512) bits=512;
-	bits=(bits+63)/64*64;
+	if (qsize != SHA_DIGEST_LENGTH && qsize != SHA224_DIGEST_LENGTH &&
+	    qsize != SHA256_DIGEST_LENGTH)
+		/* invalid q size */
+		return 0;
+
+	if (evpmd == NULL)
+		/* use SHA1 as default */
+		evpmd = EVP_sha1();
+
+	if (bits < 512)
+		bits = 512;
+
+	bits = (bits+63)/64*64;
 
 	/* NB: seed_len == 0 is special case: copy generated seed to
  	 * seed_in if it is not NULL.
  	 */
-	if (seed_len && (seed_len < 20))
-		seed_in = NULL; /* seed buffer too small -- ignore */
-	if (seed_len > 20) 
-		seed_len = 20; /* App. 2.2 of FIPS PUB 186 allows larger SEED,
-		                * but our internal buffers are restricted to 160 bits*/
-	if ((seed_in != NULL) && (seed_len == 20))
-		{
-		memcpy(seed,seed_in,seed_len);
-		/* set seed_in to NULL to avoid it being copied back */
-		seed_in = NULL;
-		}
-
-	if ((ctx=BN_CTX_new()) == NULL) goto err;
+	if (seed_len && (seed_len < (size_t)qsize))
+		seed_in = NULL;		/* seed buffer too small -- ignore */
+	if (seed_len > (size_t)qsize) 
+		seed_len = qsize;	/* App. 2.2 of FIPS PUB 186 allows larger SEED,
+					 * but our internal buffers are restricted to 160 bits*/
+	if (seed_in != NULL)
+		memcpy(seed, seed_in, seed_len);
+
+	if ((ctx=BN_CTX_new()) == NULL)
+		goto err;
 
-	if ((mont=BN_MONT_CTX_new()) == NULL) goto err;
+	if ((mont=BN_MONT_CTX_new()) == NULL)
+		goto err;
 
 	BN_CTX_start(ctx);
 	r0 = BN_CTX_get(ctx);
@@ -163,7 +203,7 @@ static int dsa_builtin_paramgen(DSA *ret, int bits,
 
 			if (!seed_len)
 				{
-				RAND_pseudo_bytes(seed,SHA_DIGEST_LENGTH);
+				RAND_pseudo_bytes(seed, qsize);
 				seed_is_random = 1;
 				}
 			else
@@ -171,25 +211,29 @@ static int dsa_builtin_paramgen(DSA *ret, int bits,
 				seed_is_random = 0;
 				seed_len=0; /* use random seed if 'seed_in' turns out to be bad*/
 				}
-			memcpy(buf,seed,SHA_DIGEST_LENGTH);
-			memcpy(buf2,seed,SHA_DIGEST_LENGTH);
+			memcpy(buf , seed, qsize);
+			memcpy(buf2, seed, qsize);
 			/* precompute "SEED + 1" for step 7: */
-			for (i=SHA_DIGEST_LENGTH-1; i >= 0; i--)
+			for (i = qsize-1; i >= 0; i--)
 				{
 				buf[i]++;
-				if (buf[i] != 0) break;
+				if (buf[i] != 0)
+					break;
 				}
 
 			/* step 2 */
-			EVP_Digest(seed,SHA_DIGEST_LENGTH,md,NULL,HASH, NULL);
-			EVP_Digest(buf,SHA_DIGEST_LENGTH,buf2,NULL,HASH, NULL);
-			for (i=0; i<SHA_DIGEST_LENGTH; i++)
+			if (!EVP_Digest(seed, qsize, md,   NULL, evpmd, NULL))
+				goto err;
+			if (!EVP_Digest(buf,  qsize, buf2, NULL, evpmd, NULL))
+				goto err;
+			for (i = 0; i < qsize; i++)
 				md[i]^=buf2[i];
 
 			/* step 3 */
-			md[0]|=0x80;
-			md[SHA_DIGEST_LENGTH-1]|=0x01;
-			if (!BN_bin2bn(md,SHA_DIGEST_LENGTH,q)) goto err;
+			md[0] |= 0x80;
+			md[qsize-1] |= 0x01;
+			if (!BN_bin2bn(md, qsize, q))
+				goto err;
 
 			/* step 4 */
 			r = BN_is_prime_fasttest_ex(q, DSS_prime_checks, ctx,
@@ -223,18 +267,21 @@ static int dsa_builtin_paramgen(DSA *ret, int bits,
 			for (k=0; k<=n; k++)
 				{
 				/* obtain "SEED + offset + k" by incrementing: */
-				for (i=SHA_DIGEST_LENGTH-1; i >= 0; i--)
+				for (i = qsize-1; i >= 0; i--)
 					{
 					buf[i]++;
-					if (buf[i] != 0) break;
+					if (buf[i] != 0)
+						break;
 					}
 
-				EVP_Digest(buf,SHA_DIGEST_LENGTH,md,NULL,HASH, NULL);
+				if (!EVP_Digest(buf, qsize, md ,NULL, evpmd,
+									NULL))
+					goto err;
 
 				/* step 8 */
-				if (!BN_bin2bn(md,SHA_DIGEST_LENGTH,r0))
+				if (!BN_bin2bn(md, qsize, r0))
 					goto err;
-				if (!BN_lshift(r0,r0,160*k)) goto err;
+				if (!BN_lshift(r0,r0,(qsize << 3)*k)) goto err;
 				if (!BN_add(W,W,r0)) goto err;
 				}
 
@@ -308,9 +355,10 @@ err:
 			ok=0;
 			goto err;
 			}
-		if (seed_in != NULL) memcpy(seed_in,seed,20);
 		if (counter_ret != NULL) *counter_ret=counter;
 		if (h_ret != NULL) *h_ret=h;
+		if (seed_out)
+			memcpy(seed_out, seed, qsize);
 		}
 	if(ctx)
 		{
@@ -321,4 +369,3 @@ err:
 	return ok;
 	}
 #endif
-#endif
diff --git a/crypto/dsa/dsa_key.c b/crypto/dsa/dsa_key.c
index 5e3912423014..9cf669b921ac 100644
--- a/crypto/dsa/dsa_key.c
+++ b/crypto/dsa/dsa_key.c
@@ -64,14 +64,28 @@
 #include <openssl/dsa.h>
 #include <openssl/rand.h>
 
-#ifndef OPENSSL_FIPS
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 
 static int dsa_builtin_keygen(DSA *dsa);
 
 int DSA_generate_key(DSA *dsa)
 	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(dsa->meth->flags & DSA_FLAG_FIPS_METHOD)
+			&& !(dsa->flags & DSA_FLAG_NON_FIPS_ALLOW))
+		{
+		DSAerr(DSA_F_DSA_GENERATE_KEY, DSA_R_NON_FIPS_DSA_METHOD);
+		return 0;
+		}
+#endif
 	if(dsa->meth->dsa_keygen)
 		return dsa->meth->dsa_keygen(dsa);
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode())
+		return FIPS_dsa_generate_key(dsa);
+#endif
 	return dsa_builtin_keygen(dsa);
 	}
 
@@ -128,5 +142,3 @@ err:
 	return(ok);
 	}
 #endif
-
-#endif
diff --git a/crypto/dsa/dsa_lib.c b/crypto/dsa/dsa_lib.c
index 85556d12d681..96d8d0c4b453 100644
--- a/crypto/dsa/dsa_lib.c
+++ b/crypto/dsa/dsa_lib.c
@@ -70,27 +70,32 @@
 #include <openssl/dh.h>
 #endif
 
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
+
 const char DSA_version[]="DSA" OPENSSL_VERSION_PTEXT;
 
 static const DSA_METHOD *default_DSA_method = NULL;
 
 void DSA_set_default_method(const DSA_METHOD *meth)
 	{
-#ifdef OPENSSL_FIPS
-	if (FIPS_mode() && !(meth->flags & DSA_FLAG_FIPS_METHOD))
-		{
-		DSAerr(DSA_F_DSA_SET_DEFAULT_METHOD, DSA_R_NON_FIPS_METHOD);
-		return;
-		}
-#endif
-		
 	default_DSA_method = meth;
 	}
 
 const DSA_METHOD *DSA_get_default_method(void)
 	{
 	if(!default_DSA_method)
+		{
+#ifdef OPENSSL_FIPS
+		if (FIPS_mode())
+			return FIPS_dsa_openssl();
+		else
+			return DSA_OpenSSL();
+#else
 		default_DSA_method = DSA_OpenSSL();
+#endif
+		}
 	return default_DSA_method;
 	}
 
@@ -104,13 +109,6 @@ int DSA_set_method(DSA *dsa, const DSA_METHOD *meth)
 	/* NB: The caller is specifically setting a method, so it's not up to us
 	 * to deal with which ENGINE it comes from. */
         const DSA_METHOD *mtmp;
-#ifdef OPENSSL_FIPS
-	if (FIPS_mode() && !(meth->flags & DSA_FLAG_FIPS_METHOD))
-		{
-		DSAerr(DSA_F_DSA_SET_METHOD, DSA_R_NON_FIPS_METHOD);
-		return 0;
-		}
-#endif
         mtmp = dsa->meth;
         if (mtmp->finish) mtmp->finish(dsa);
 #ifndef OPENSSL_NO_ENGINE
@@ -162,18 +160,6 @@ DSA *DSA_new_method(ENGINE *engine)
 			}
 		}
 #endif
-#ifdef OPENSSL_FIPS
-	if (FIPS_mode() && !(ret->meth->flags & DSA_FLAG_FIPS_METHOD))
-		{
-		DSAerr(DSA_F_DSA_NEW_METHOD, DSA_R_NON_FIPS_METHOD);
-#ifndef OPENSSL_NO_ENGINE
-		if (ret->engine)
-			ENGINE_finish(ret->engine);
-#endif
-		OPENSSL_free(ret);
-		return NULL;
-		}
-#endif
 
 	ret->pad=0;
 	ret->version=0;
@@ -260,6 +246,28 @@ int DSA_up_ref(DSA *r)
 	return ((i > 1) ? 1 : 0);
 	}
 
+int DSA_size(const DSA *r)
+	{
+	int ret,i;
+	ASN1_INTEGER bs;
+	unsigned char buf[4];	/* 4 bytes looks really small.
+				   However, i2d_ASN1_INTEGER() will not look
+				   beyond the first byte, as long as the second
+				   parameter is NULL. */
+
+	i=BN_num_bits(r->q);
+	bs.length=(i+7)/8;
+	bs.data=buf;
+	bs.type=V_ASN1_INTEGER;
+	/* If the top bit is set the asn1 encoding is 1 larger. */
+	buf[0]=0xff;	
+
+	i=i2d_ASN1_INTEGER(&bs,NULL);
+	i+=i; /* r and s */
+	ret=ASN1_object_size(1,i,V_ASN1_SEQUENCE);
+	return(ret);
+	}
+
 int DSA_get_ex_new_index(long argl, void *argp, CRYPTO_EX_new *new_func,
 	     CRYPTO_EX_dup *dup_func, CRYPTO_EX_free *free_func)
         {
@@ -281,7 +289,8 @@ void *DSA_get_ex_data(DSA *d, int idx)
 DH *DSA_dup_DH(const DSA *r)
 	{
 	/* DSA has p, q, g, optional pub_key, optional priv_key.
-	 * DH has p, optional length, g, optional pub_key, optional priv_key.
+	 * DH has p, optional length, g, optional pub_key, optional priv_key,
+	 * optional q.
 	 */ 
 
 	DH *ret = NULL;
@@ -295,7 +304,11 @@ DH *DSA_dup_DH(const DSA *r)
 		if ((ret->p = BN_dup(r->p)) == NULL)
 			goto err;
 	if (r->q != NULL)
+		{
 		ret->length = BN_num_bits(r->q);
+		if ((ret->q = BN_dup(r->q)) == NULL)
+			goto err;
+		}
 	if (r->g != NULL)
 		if ((ret->g = BN_dup(r->g)) == NULL)
 			goto err;
diff --git a/crypto/ec/ec2_smpt.c b/crypto/dsa/dsa_locl.h
index 72a8d570517f..21e2e4524224 100644
--- a/crypto/ec/ec2_smpt.c
+++ b/crypto/dsa/dsa_locl.h
@@ -1,9 +1,5 @@
-/* crypto/ec/ec2_smpt.c */
-/* This code was originally written by Douglas Stebila 
- * <dstebila@student.math.uwaterloo.ca> for the OpenSSL project.
- */
 /* ====================================================================
- * Copyright (c) 1998-2002 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2007 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -56,86 +52,9 @@
  *
  */
 
+#include <openssl/dsa.h>
 
-/* Calaculates and sets the affine coordinates of an EC_POINT from the given
- * compressed coordinates.  Uses algorithm 2.3.4 of SEC 1. 
- * Note that the simple implementation only uses affine coordinates.
- *
- * This algorithm is patented by Certicom Corp. under US Patent 6,141,420
- * (for licensing information, contact licensing@certicom.com).
- * This function is disabled by default and can be enabled by defining the 
- * preprocessor macro OPENSSL_EC_BIN_PT_COMP at Configure-time.
- */
-int ec_GF2m_simple_set_compressed_coordinates(const EC_GROUP *group, EC_POINT *point,
-	const BIGNUM *x_, int y_bit, BN_CTX *ctx)
-	{
-#ifndef OPENSSL_EC_BIN_PT_COMP	
-	ECerr(EC_F_EC_GF2M_SIMPLE_SET_COMPRESSED_COORDINATES, ERR_R_DISABLED);
-	return 0;
-#else
-	BN_CTX *new_ctx = NULL;
-	BIGNUM *tmp, *x, *y, *z;
-	int ret = 0, z0;
-
-	/* clear error queue */
-	ERR_clear_error();
-
-	if (ctx == NULL)
-		{
-		ctx = new_ctx = BN_CTX_new();
-		if (ctx == NULL)
-			return 0;
-		}
-
-	y_bit = (y_bit != 0) ? 1 : 0;
-
-	BN_CTX_start(ctx);
-	tmp = BN_CTX_get(ctx);
-	x = BN_CTX_get(ctx);
-	y = BN_CTX_get(ctx);
-	z = BN_CTX_get(ctx);
-	if (z == NULL) goto err;
-
-	if (!BN_GF2m_mod_arr(x, x_, group->poly)) goto err;
-	if (BN_is_zero(x))
-		{
-		if (!BN_GF2m_mod_sqrt_arr(y, &group->b, group->poly, ctx)) goto err;
-		}
-	else
-		{
-		if (!group->meth->field_sqr(group, tmp, x, ctx)) goto err;
-		if (!group->meth->field_div(group, tmp, &group->b, tmp, ctx)) goto err;
-		if (!BN_GF2m_add(tmp, &group->a, tmp)) goto err;
-		if (!BN_GF2m_add(tmp, x, tmp)) goto err;
-		if (!BN_GF2m_mod_solve_quad_arr(z, tmp, group->poly, ctx))
-			{
-			unsigned long err = ERR_peek_last_error();
-			
-			if (ERR_GET_LIB(err) == ERR_LIB_BN && ERR_GET_REASON(err) == BN_R_NO_SOLUTION)
-				{
-				ERR_clear_error();
-				ECerr(EC_F_EC_GF2M_SIMPLE_SET_COMPRESSED_COORDINATES, EC_R_INVALID_COMPRESSED_POINT);
-				}
-			else
-				ECerr(EC_F_EC_GF2M_SIMPLE_SET_COMPRESSED_COORDINATES, ERR_R_BN_LIB);
-			goto err;
-			}
-		z0 = (BN_is_odd(z)) ? 1 : 0;
-		if (!group->meth->field_mul(group, y, x, z, ctx)) goto err;
-		if (z0 != y_bit)
-			{
-			if (!BN_GF2m_add(y, y, x)) goto err;
-			}
-		}
-
-	if (!EC_POINT_set_affine_coordinates_GF2m(group, point, x, y, ctx)) goto err;
-
-	ret = 1;
-
- err:
-	BN_CTX_end(ctx);
-	if (new_ctx != NULL)
-		BN_CTX_free(new_ctx);
-	return ret;
-#endif
-	}
+int dsa_builtin_paramgen(DSA *ret, size_t bits, size_t qbits,
+	const EVP_MD *evpmd, const unsigned char *seed_in, size_t seed_len,
+	unsigned char *seed_out,
+	int *counter_ret, unsigned long *h_ret, BN_GENCB *cb);
diff --git a/crypto/dsa/dsa_ossl.c b/crypto/dsa/dsa_ossl.c
index 172776080669..b3d78e524cf4 100644
--- a/crypto/dsa/dsa_ossl.c
+++ b/crypto/dsa/dsa_ossl.c
@@ -61,16 +61,15 @@
 #include <stdio.h>
 #include "cryptlib.h"
 #include <openssl/bn.h>
+#include <openssl/sha.h>
 #include <openssl/dsa.h>
 #include <openssl/rand.h>
 #include <openssl/asn1.h>
 
-#ifndef OPENSSL_FIPS
-
 static DSA_SIG *dsa_do_sign(const unsigned char *dgst, int dlen, DSA *dsa);
 static int dsa_sign_setup(DSA *dsa, BN_CTX *ctx_in, BIGNUM **kinvp, BIGNUM **rp);
 static int dsa_do_verify(const unsigned char *dgst, int dgst_len, DSA_SIG *sig,
-		  DSA *dsa);
+			 DSA *dsa);
 static int dsa_init(DSA *dsa);
 static int dsa_finish(DSA *dsa);
 
@@ -135,8 +134,9 @@ static DSA_SIG *dsa_do_sign(const unsigned char *dgst, int dlen, DSA *dsa)
 	BIGNUM m;
 	BIGNUM xr;
 	BN_CTX *ctx=NULL;
-	int i,reason=ERR_R_BN_LIB;
+	int reason=ERR_R_BN_LIB;
 	DSA_SIG *ret=NULL;
+	int noredo = 0;
 
 	BN_init(&m);
 	BN_init(&xr);
@@ -149,17 +149,9 @@ static DSA_SIG *dsa_do_sign(const unsigned char *dgst, int dlen, DSA *dsa)
 
 	s=BN_new();
 	if (s == NULL) goto err;
-
-	i=BN_num_bytes(dsa->q); /* should be 20 */
-	if ((dlen > i) || (dlen > 50))
-		{
-		reason=DSA_R_DATA_TOO_LARGE_FOR_KEY_SIZE;
-		goto err;
-		}
-
 	ctx=BN_CTX_new();
 	if (ctx == NULL) goto err;
-
+redo:
 	if ((dsa->kinv == NULL) || (dsa->r == NULL))
 		{
 		if (!DSA_sign_setup(dsa,ctx,&kinv,&r)) goto err;
@@ -170,20 +162,39 @@ static DSA_SIG *dsa_do_sign(const unsigned char *dgst, int dlen, DSA *dsa)
 		dsa->kinv=NULL;
 		r=dsa->r;
 		dsa->r=NULL;
+		noredo = 1;
 		}
 
-	if (BN_bin2bn(dgst,dlen,&m) == NULL) goto err;
+	
+	if (dlen > BN_num_bytes(dsa->q))
+		/* if the digest length is greater than the size of q use the
+		 * BN_num_bits(dsa->q) leftmost bits of the digest, see
+		 * fips 186-3, 4.2 */
+		dlen = BN_num_bytes(dsa->q);
+	if (BN_bin2bn(dgst,dlen,&m) == NULL)
+		goto err;
 
 	/* Compute  s = inv(k) (m + xr) mod q */
 	if (!BN_mod_mul(&xr,dsa->priv_key,r,dsa->q,ctx)) goto err;/* s = xr */
 	if (!BN_add(s, &xr, &m)) goto err;		/* s = m + xr */
 	if (BN_cmp(s,dsa->q) > 0)
-		if (!BN_sub(s,s,dsa->q))
-			goto err;
+		if (!BN_sub(s,s,dsa->q)) goto err;
 	if (!BN_mod_mul(s,s,kinv,dsa->q,ctx)) goto err;
 
 	ret=DSA_SIG_new();
 	if (ret == NULL) goto err;
+	/* Redo if r or s is zero as required by FIPS 186-3: this is
+	 * very unlikely.
+	 */
+	if (BN_is_zero(r) || BN_is_zero(s))
+		{
+		if (noredo)
+			{
+			reason = DSA_R_NEED_NEW_SETUP_VALUES;
+			goto err;
+			}
+		goto redo;
+		}
 	ret->r = r;
 	ret->s = s;
 	
@@ -284,30 +295,31 @@ err:
 	if (!ret)
 		{
 		DSAerr(DSA_F_DSA_SIGN_SETUP,ERR_R_BN_LIB);
-		if (kinv != NULL) BN_clear_free(kinv);
-		if (r != NULL) BN_clear_free(r);
+		if (r != NULL)
+			BN_clear_free(r);
 		}
 	if (ctx_in == NULL) BN_CTX_free(ctx);
-	if (kinv != NULL) BN_clear_free(kinv);
 	BN_clear_free(&k);
 	BN_clear_free(&kq);
 	return(ret);
 	}
 
 static int dsa_do_verify(const unsigned char *dgst, int dgst_len, DSA_SIG *sig,
-		  DSA *dsa)
+			 DSA *dsa)
 	{
 	BN_CTX *ctx;
 	BIGNUM u1,u2,t1;
 	BN_MONT_CTX *mont=NULL;
-	int ret = -1;
+	int ret = -1, i;
 	if (!dsa->p || !dsa->q || !dsa->g)
 		{
 		DSAerr(DSA_F_DSA_DO_VERIFY,DSA_R_MISSING_PARAMETERS);
 		return -1;
 		}
 
-	if (BN_num_bits(dsa->q) != 160)
+	i = BN_num_bits(dsa->q);
+	/* fips 186-3 allows only different sizes for q */
+	if (i != 160 && i != 224 && i != 256)
 		{
 		DSAerr(DSA_F_DSA_DO_VERIFY,DSA_R_BAD_Q_VALUE);
 		return -1;
@@ -318,7 +330,6 @@ static int dsa_do_verify(const unsigned char *dgst, int dgst_len, DSA_SIG *sig,
 		DSAerr(DSA_F_DSA_DO_VERIFY,DSA_R_MODULUS_TOO_LARGE);
 		return -1;
 		}
-
 	BN_init(&u1);
 	BN_init(&u2);
 	BN_init(&t1);
@@ -343,6 +354,11 @@ static int dsa_do_verify(const unsigned char *dgst, int dgst_len, DSA_SIG *sig,
 	if ((BN_mod_inverse(&u2,sig->s,dsa->q,ctx)) == NULL) goto err;
 
 	/* save M in u1 */
+	if (dgst_len > (i >> 3))
+		/* if the digest length is greater than the size of q use the
+		 * BN_num_bits(dsa->q) leftmost bits of the digest, see
+		 * fips 186-3, 4.2 */
+		dgst_len = (i >> 3);
 	if (BN_bin2bn(dgst,dgst_len,&u1) == NULL) goto err;
 
 	/* u1 = M * w mod q */
@@ -394,4 +410,3 @@ static int dsa_finish(DSA *dsa)
 	return(1);
 }
 
-#endif
diff --git a/crypto/dsa/dsa_pmeth.c b/crypto/dsa/dsa_pmeth.c
new file mode 100644
index 000000000000..715d8d675bb7
--- /dev/null
+++ b/crypto/dsa/dsa_pmeth.c
@@ -0,0 +1,318 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/asn1t.h>
+#include <openssl/x509.h>
+#include <openssl/evp.h>
+#include <openssl/bn.h>
+#include "evp_locl.h"
+#include "dsa_locl.h"
+
+/* DSA pkey context structure */
+
+typedef struct
+	{
+	/* Parameter gen parameters */
+	int nbits;		/* size of p in bits (default: 1024) */
+	int qbits;		/* size of q in bits (default: 160)  */
+	const EVP_MD *pmd;	/* MD for parameter generation */
+	/* Keygen callback info */
+	int gentmp[2];
+	/* message digest */
+	const EVP_MD *md;	/* MD for the signature */
+	} DSA_PKEY_CTX;
+
+static int pkey_dsa_init(EVP_PKEY_CTX *ctx)
+	{
+	DSA_PKEY_CTX *dctx;
+	dctx = OPENSSL_malloc(sizeof(DSA_PKEY_CTX));
+	if (!dctx)
+		return 0;
+	dctx->nbits = 1024;
+	dctx->qbits = 160;
+	dctx->pmd = NULL;
+	dctx->md = NULL;
+
+	ctx->data = dctx;
+	ctx->keygen_info = dctx->gentmp;
+	ctx->keygen_info_count = 2;
+	
+	return 1;
+	}
+
+static int pkey_dsa_copy(EVP_PKEY_CTX *dst, EVP_PKEY_CTX *src)
+	{
+	DSA_PKEY_CTX *dctx, *sctx;
+	if (!pkey_dsa_init(dst))
+		return 0;
+       	sctx = src->data;
+	dctx = dst->data;
+	dctx->nbits = sctx->nbits;
+	dctx->qbits = sctx->qbits;
+	dctx->pmd = sctx->pmd;
+	dctx->md  = sctx->md;
+	return 1;
+	}
+
+static void pkey_dsa_cleanup(EVP_PKEY_CTX *ctx)
+	{
+	DSA_PKEY_CTX *dctx = ctx->data;
+	if (dctx)
+		OPENSSL_free(dctx);
+	}
+
+static int pkey_dsa_sign(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					const unsigned char *tbs, size_t tbslen)
+	{
+	int ret, type;
+	unsigned int sltmp;
+	DSA_PKEY_CTX *dctx = ctx->data;
+	DSA *dsa = ctx->pkey->pkey.dsa;
+
+	if (dctx->md)
+		type = EVP_MD_type(dctx->md);
+	else
+		type = NID_sha1;
+
+	ret = DSA_sign(type, tbs, tbslen, sig, &sltmp, dsa);
+
+	if (ret <= 0)
+		return ret;
+	*siglen = sltmp;
+	return 1;
+	}
+
+static int pkey_dsa_verify(EVP_PKEY_CTX *ctx,
+					const unsigned char *sig, size_t siglen,
+					const unsigned char *tbs, size_t tbslen)
+	{
+	int ret, type;
+	DSA_PKEY_CTX *dctx = ctx->data;
+	DSA *dsa = ctx->pkey->pkey.dsa;
+
+	if (dctx->md)
+		type = EVP_MD_type(dctx->md);
+	else
+		type = NID_sha1;
+
+	ret = DSA_verify(type, tbs, tbslen, sig, siglen, dsa);
+
+	return ret;
+	}
+
+static int pkey_dsa_ctrl(EVP_PKEY_CTX *ctx, int type, int p1, void *p2)
+	{
+	DSA_PKEY_CTX *dctx = ctx->data;
+	switch (type)
+		{
+		case EVP_PKEY_CTRL_DSA_PARAMGEN_BITS:
+		if (p1 < 256)
+			return -2;
+		dctx->nbits = p1;
+		return 1;
+
+		case EVP_PKEY_CTRL_DSA_PARAMGEN_Q_BITS:
+		if (p1 != 160 && p1 != 224 && p1 && p1 != 256)
+			return -2;
+		dctx->qbits = p1;
+		return 1;
+
+		case EVP_PKEY_CTRL_DSA_PARAMGEN_MD:
+		if (EVP_MD_type((const EVP_MD *)p2) != NID_sha1   &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha224 &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha256)
+			{
+			DSAerr(DSA_F_PKEY_DSA_CTRL, DSA_R_INVALID_DIGEST_TYPE);
+			return 0;
+			}
+		dctx->md = p2;
+		return 1;
+
+		case EVP_PKEY_CTRL_MD:
+		if (EVP_MD_type((const EVP_MD *)p2) != NID_sha1   &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_dsa    &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_dsaWithSHA    &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha224 &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha256 &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha384 &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha512)
+			{
+			DSAerr(DSA_F_PKEY_DSA_CTRL, DSA_R_INVALID_DIGEST_TYPE);
+			return 0;
+			}
+		dctx->md = p2;
+		return 1;
+
+		case EVP_PKEY_CTRL_DIGESTINIT:
+		case EVP_PKEY_CTRL_PKCS7_SIGN:
+		case EVP_PKEY_CTRL_CMS_SIGN:
+		return 1;
+		
+		case EVP_PKEY_CTRL_PEER_KEY:
+			DSAerr(DSA_F_PKEY_DSA_CTRL,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+			return -2;	
+		default:
+		return -2;
+
+		}
+	}
+			
+static int pkey_dsa_ctrl_str(EVP_PKEY_CTX *ctx,
+			const char *type, const char *value)
+	{
+	if (!strcmp(type, "dsa_paramgen_bits"))
+		{
+		int nbits;
+		nbits = atoi(value);
+		return EVP_PKEY_CTX_set_dsa_paramgen_bits(ctx, nbits);
+		}
+	if (!strcmp(type, "dsa_paramgen_q_bits"))
+		{
+		int qbits = atoi(value);
+		return EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_DSA, EVP_PKEY_OP_PARAMGEN,
+		                         EVP_PKEY_CTRL_DSA_PARAMGEN_Q_BITS, qbits, NULL);
+		}
+	if (!strcmp(type, "dsa_paramgen_md"))
+		{
+		return EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_DSA, EVP_PKEY_OP_PARAMGEN,
+		                         EVP_PKEY_CTRL_DSA_PARAMGEN_MD, 0, 
+		                         (void *)EVP_get_digestbyname(value));
+		}
+	return -2;
+	}
+
+static int pkey_dsa_paramgen(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey)
+	{
+	DSA *dsa = NULL;
+	DSA_PKEY_CTX *dctx = ctx->data;
+	BN_GENCB *pcb, cb;
+	int ret;
+	if (ctx->pkey_gencb)
+		{
+		pcb = &cb;
+		evp_pkey_set_cb_translate(pcb, ctx);
+		}
+	else
+		pcb = NULL;
+	dsa = DSA_new();
+	if (!dsa)
+		return 0;
+	ret = dsa_builtin_paramgen(dsa, dctx->nbits, dctx->qbits, dctx->pmd,
+	                           NULL, 0, NULL, NULL, NULL, pcb);
+	if (ret)
+		EVP_PKEY_assign_DSA(pkey, dsa);
+	else
+		DSA_free(dsa);
+	return ret;
+	}
+
+static int pkey_dsa_keygen(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey)
+	{
+	DSA *dsa = NULL;
+	if (ctx->pkey == NULL)
+		{
+		DSAerr(DSA_F_PKEY_DSA_KEYGEN, DSA_R_NO_PARAMETERS_SET);
+		return 0;
+		}
+	dsa = DSA_new();
+	if (!dsa)
+		return 0;
+	EVP_PKEY_assign_DSA(pkey, dsa);
+	/* Note: if error return, pkey is freed by parent routine */
+	if (!EVP_PKEY_copy_parameters(pkey, ctx->pkey))
+		return 0;
+	return DSA_generate_key(pkey->pkey.dsa);
+	}
+
+const EVP_PKEY_METHOD dsa_pkey_meth = 
+	{
+	EVP_PKEY_DSA,
+	EVP_PKEY_FLAG_AUTOARGLEN,
+	pkey_dsa_init,
+	pkey_dsa_copy,
+	pkey_dsa_cleanup,
+
+	0,
+	pkey_dsa_paramgen,
+
+	0,
+	pkey_dsa_keygen,
+
+	0,
+	pkey_dsa_sign,
+
+	0,
+	pkey_dsa_verify,
+
+	0,0,
+
+	0,0,0,0,
+
+	0,0,
+
+	0,0,
+
+	0,0,
+
+	pkey_dsa_ctrl,
+	pkey_dsa_ctrl_str
+
+
+	};
diff --git a/crypto/evp/evp_cnf.c b/crypto/dsa/dsa_prn.c
index 2e4db3023593..6f29f5e2409b 100644
--- a/crypto/evp/evp_cnf.c
+++ b/crypto/dsa/dsa_prn.c
@@ -1,9 +1,9 @@
-/* evp_cnf.c */
-/* Written by Stephen Henson (steve@openssl.org) for the OpenSSL
- * project 2007.
+/* crypto/dsa/dsa_prn.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
  */
 /* ====================================================================
- * Copyright (c) 2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -57,69 +57,65 @@
  */
 
 #include <stdio.h>
-#include <ctype.h>
-#include <openssl/crypto.h>
 #include "cryptlib.h"
-#include <openssl/conf.h>
-#include <openssl/dso.h>
-#include <openssl/x509.h>
-#include <openssl/x509v3.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
+#include <openssl/evp.h>
+#include <openssl/dsa.h>
 
+#ifndef OPENSSL_NO_FP_API
+int DSA_print_fp(FILE *fp, const DSA *x, int off)
+	{
+	BIO *b;
+	int ret;
 
-/* Algorithm configuration module. */
+	if ((b=BIO_new(BIO_s_file())) == NULL)
+		{
+		DSAerr(DSA_F_DSA_PRINT_FP,ERR_R_BUF_LIB);
+		return(0);
+		}
+	BIO_set_fp(b,fp,BIO_NOCLOSE);
+	ret=DSA_print(b,x,off);
+	BIO_free(b);
+	return(ret);
+	}
 
-static int alg_module_init(CONF_IMODULE *md, const CONF *cnf)
+int DSAparams_print_fp(FILE *fp, const DSA *x)
 	{
-	int i;
-	const char *oid_section;
-	STACK_OF(CONF_VALUE) *sktmp;
-	CONF_VALUE *oval;
-	oid_section = CONF_imodule_get_value(md);
-	if(!(sktmp = NCONF_get_section(cnf, oid_section)))
+	BIO *b;
+	int ret;
+
+	if ((b=BIO_new(BIO_s_file())) == NULL)
 		{
-		EVPerr(EVP_F_ALG_MODULE_INIT, EVP_R_ERROR_LOADING_SECTION);
-		return 0;
+		DSAerr(DSA_F_DSAPARAMS_PRINT_FP,ERR_R_BUF_LIB);
+		return(0);
 		}
-	for(i = 0; i < sk_CONF_VALUE_num(sktmp); i++)
-		{
-		oval = sk_CONF_VALUE_value(sktmp, i);
-		if (!strcmp(oval->name, "fips_mode"))
-			{
-			int m;
-			if (!X509V3_get_value_bool(oval, &m))
-				{
-				EVPerr(EVP_F_ALG_MODULE_INIT, EVP_R_INVALID_FIPS_MODE);
-				return 0;
-				}
-			if (m > 0)
-				{
-#ifdef OPENSSL_FIPS
-				if (!FIPS_mode() && !FIPS_mode_set(1))
-					{
-					EVPerr(EVP_F_ALG_MODULE_INIT, EVP_R_ERROR_SETTING_FIPS_MODE);
-					return 0;
-					}
-#else
-				EVPerr(EVP_F_ALG_MODULE_INIT, EVP_R_FIPS_MODE_NOT_SUPPORTED);
-				return 0;
+	BIO_set_fp(b,fp,BIO_NOCLOSE);
+	ret=DSAparams_print(b, x);
+	BIO_free(b);
+	return(ret);
+	}
 #endif
-				}
-			}
-		else
-			{
-			EVPerr(EVP_F_ALG_MODULE_INIT, EVP_R_UNKNOWN_OPTION);
-			ERR_add_error_data(4, "name=", oval->name,
-						", value=", oval->value);
-			}
-				
-		}
-	return 1;
+
+int DSA_print(BIO *bp, const DSA *x, int off)
+	{
+	EVP_PKEY *pk;
+	int ret;
+	pk = EVP_PKEY_new();
+	if (!pk || !EVP_PKEY_set1_DSA(pk, (DSA *)x))
+		return 0;
+	ret = EVP_PKEY_print_private(bp, pk, off, NULL);
+	EVP_PKEY_free(pk);
+	return ret;
 	}
 
-void EVP_add_alg_module(void)
+int DSAparams_print(BIO *bp, const DSA *x)
 	{
-	CONF_module_add("alg_section", alg_module_init, 0);
+	EVP_PKEY *pk;
+	int ret;
+	pk = EVP_PKEY_new();
+	if (!pk || !EVP_PKEY_set1_DSA(pk, (DSA *)x))
+		return 0;
+	ret = EVP_PKEY_print_params(bp, pk, 4, NULL);
+	EVP_PKEY_free(pk);
+	return ret;
 	}
+
diff --git a/crypto/dsa/dsa_sign.c b/crypto/dsa/dsa_sign.c
index 4cfbbe57a802..c3cc3642cea8 100644
--- a/crypto/dsa/dsa_sign.c
+++ b/crypto/dsa/dsa_sign.c
@@ -58,23 +58,18 @@
 
 /* Original version from Steven Schoch <schoch@sheba.arc.nasa.gov> */
 
-#include <stdio.h>
 #include "cryptlib.h"
-#include <openssl/bn.h>
 #include <openssl/dsa.h>
 #include <openssl/rand.h>
-#include <openssl/asn1.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/bn.h>
 
 DSA_SIG * DSA_do_sign(const unsigned char *dgst, int dlen, DSA *dsa)
 	{
 #ifdef OPENSSL_FIPS
-	if(FIPS_mode() && !(dsa->flags & DSA_FLAG_NON_FIPS_ALLOW))
+	if (FIPS_mode() && !(dsa->meth->flags & DSA_FLAG_FIPS_METHOD)
+			&& !(dsa->flags & DSA_FLAG_NON_FIPS_ALLOW))
 		{
-		DSAerr(DSA_F_DSA_DO_SIGN, DSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
+		DSAerr(DSA_F_DSA_DO_SIGN, DSA_R_NON_FIPS_DSA_METHOD);
 		return NULL;
 		}
 #endif
@@ -84,12 +79,36 @@ DSA_SIG * DSA_do_sign(const unsigned char *dgst, int dlen, DSA *dsa)
 int DSA_sign_setup(DSA *dsa, BN_CTX *ctx_in, BIGNUM **kinvp, BIGNUM **rp)
 	{
 #ifdef OPENSSL_FIPS
-	if(FIPS_mode() && !(dsa->flags & DSA_FLAG_NON_FIPS_ALLOW))
+	if (FIPS_mode() && !(dsa->meth->flags & DSA_FLAG_FIPS_METHOD)
+			&& !(dsa->flags & DSA_FLAG_NON_FIPS_ALLOW))
 		{
-		DSAerr(DSA_F_DSA_SIGN_SETUP, DSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
+		DSAerr(DSA_F_DSA_SIGN_SETUP, DSA_R_NON_FIPS_DSA_METHOD);
 		return 0;
 		}
 #endif
 	return dsa->meth->dsa_sign_setup(dsa, ctx_in, kinvp, rp);
 	}
 
+DSA_SIG *DSA_SIG_new(void)
+	{
+	DSA_SIG *sig;
+	sig = OPENSSL_malloc(sizeof(DSA_SIG));
+	if (!sig)
+		return NULL;
+	sig->r = NULL;
+	sig->s = NULL;
+	return sig;
+	}
+
+void DSA_SIG_free(DSA_SIG *sig)
+	{
+	if (sig)
+		{
+		if (sig->r)
+			BN_free(sig->r);
+		if (sig->s)
+			BN_free(sig->s);
+		OPENSSL_free(sig);
+		}
+	}
+
diff --git a/crypto/dsa/dsa_utl.c b/crypto/dsa/dsa_utl.c
deleted file mode 100644
index 24c021d12017..000000000000
--- a/crypto/dsa/dsa_utl.c
+++ /dev/null
@@ -1,95 +0,0 @@
-/* crypto/dsa/dsa_lib.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-/* Original version from Steven Schoch <schoch@sheba.arc.nasa.gov> */
-
-#include <stdio.h>
-#include "cryptlib.h"
-#include <openssl/bn.h>
-#include <openssl/dsa.h>
-#include <openssl/asn1.h>
-#ifndef OPENSSL_NO_ENGINE
-#include <openssl/engine.h>
-#endif
-#ifndef OPENSSL_NO_DH
-#include <openssl/dh.h>
-#endif
-
-DSA_SIG *DSA_SIG_new(void)
-	{
-	DSA_SIG *sig;
-	sig = OPENSSL_malloc(sizeof(DSA_SIG));
-	if (!sig)
-		return NULL;
-	sig->r = NULL;
-	sig->s = NULL;
-	return sig;
-	}
-
-void DSA_SIG_free(DSA_SIG *sig)
-	{
-	if (sig)
-		{
-		if (sig->r)
-			BN_free(sig->r);
-		if (sig->s)
-			BN_free(sig->s);
-		OPENSSL_free(sig);
-		}
-	}
-
diff --git a/crypto/dsa/dsa_vrf.c b/crypto/dsa/dsa_vrf.c
index c75e423048ac..674cb5fa5f5b 100644
--- a/crypto/dsa/dsa_vrf.c
+++ b/crypto/dsa/dsa_vrf.c
@@ -58,26 +58,18 @@
 
 /* Original version from Steven Schoch <schoch@sheba.arc.nasa.gov> */
 
-#include <stdio.h>
 #include "cryptlib.h"
-#include <openssl/bn.h>
 #include <openssl/dsa.h>
-#include <openssl/rand.h>
-#include <openssl/asn1.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
-#include <openssl/asn1_mac.h>
 
 int DSA_do_verify(const unsigned char *dgst, int dgst_len, DSA_SIG *sig,
 		  DSA *dsa)
 	{
 #ifdef OPENSSL_FIPS
-	if(FIPS_mode() && !(dsa->flags & DSA_FLAG_NON_FIPS_ALLOW))
+	if (FIPS_mode() && !(dsa->meth->flags & DSA_FLAG_FIPS_METHOD)
+			&& !(dsa->flags & DSA_FLAG_NON_FIPS_ALLOW))
 		{
-		DSAerr(DSA_F_DSA_DO_VERIFY, DSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
-		return 0;
+		DSAerr(DSA_F_DSA_DO_VERIFY, DSA_R_NON_FIPS_DSA_METHOD);
+		return -1;
 		}
 #endif
 	return dsa->meth->dsa_do_verify(dgst, dgst_len, sig, dsa);
diff --git a/crypto/dsa/dsatest.c b/crypto/dsa/dsatest.c
index 912317bb443c..edffd24e6ba2 100644
--- a/crypto/dsa/dsatest.c
+++ b/crypto/dsa/dsatest.c
@@ -169,7 +169,6 @@ int main(int argc, char **argv)
 		}
 	BIO_printf(bio_err,"\ncounter=%d h=%ld\n",counter,h);
 		
-	if (dsa == NULL) goto end;
 	DSA_print(bio_err,dsa,0);
 	if (counter != 105) 
 		{
@@ -223,7 +222,7 @@ end:
 		ERR_print_errors(bio_err);
 	if (dsa != NULL) DSA_free(dsa);
 	CRYPTO_cleanup_all_ex_data();
-	ERR_remove_state(0);
+	ERR_remove_thread_state(NULL);
 	ERR_free_strings();
 	CRYPTO_mem_leaks(bio_err);
 	if (bio_err != NULL)
diff --git a/crypto/dso/Makefile b/crypto/dso/Makefile
index 52f152888c4b..fb2709ed63a7 100644
--- a/crypto/dso/Makefile
+++ b/crypto/dso/Makefile
@@ -18,9 +18,9 @@ APPS=
 
 LIB=$(TOP)/libcrypto.a
 LIBSRC= dso_dl.c dso_dlfcn.c dso_err.c dso_lib.c dso_null.c \
-	dso_openssl.c dso_win32.c dso_vms.c
+	dso_openssl.c dso_win32.c dso_vms.c dso_beos.c
 LIBOBJ= dso_dl.o dso_dlfcn.o dso_err.o dso_lib.o dso_null.o \
-	dso_openssl.o dso_win32.o dso_vms.o
+	dso_openssl.o dso_win32.o dso_vms.o dso_beos.o
 
 SRC= $(LIBSRC)
 
@@ -35,7 +35,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -76,6 +76,14 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
+dso_beos.o: ../../e_os.h ../../include/openssl/bio.h
+dso_beos.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+dso_beos.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
+dso_beos.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
+dso_beos.o: ../../include/openssl/opensslconf.h
+dso_beos.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+dso_beos.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+dso_beos.o: ../../include/openssl/symhacks.h ../cryptlib.h dso_beos.c
 dso_dl.o: ../../e_os.h ../../include/openssl/bio.h
 dso_dl.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 dso_dl.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
diff --git a/crypto/dso/dso.h b/crypto/dso/dso.h
index 3e51913a725b..839f2e061708 100644
--- a/crypto/dso/dso.h
+++ b/crypto/dso/dso.h
@@ -170,6 +170,11 @@ typedef struct dso_meth_st
 	/* [De]Initialisation handlers. */
 	int (*init)(DSO *dso);
 	int (*finish)(DSO *dso);
+
+	/* Return pathname of the module containing location */
+	int (*pathbyaddr)(void *addr,char *path,int sz);
+	/* Perform global symbol lookup, i.e. among *all* modules */
+	void *(*globallookup)(const char *symname);
 	} DSO_METHOD;
 
 /**********************************************************************/
@@ -183,7 +188,7 @@ struct dso_st
 	 * for use in the dso_bind handler. All in all, let each
 	 * method control its own destiny. "Handles" and such go in
 	 * a STACK. */
-	STACK *meth_data;
+	STACK_OF(void) *meth_data;
 	int references;
 	int flags;
 	/* For use by applications etc ... use this for your bits'n'pieces,
@@ -296,6 +301,30 @@ DSO_METHOD *DSO_METHOD_win32(void);
 /* If VMS is defined, use shared images. If not, return NULL. */
 DSO_METHOD *DSO_METHOD_vms(void);
 
+/* This function writes null-terminated pathname of DSO module
+ * containing 'addr' into 'sz' large caller-provided 'path' and
+ * returns the number of characters [including trailing zero]
+ * written to it. If 'sz' is 0 or negative, 'path' is ignored and
+ * required amount of charachers [including trailing zero] to
+ * accomodate pathname is returned. If 'addr' is NULL, then
+ * pathname of cryptolib itself is returned. Negative or zero
+ * return value denotes error.
+ */
+int DSO_pathbyaddr(void *addr,char *path,int sz);
+
+/* This function should be used with caution! It looks up symbols in
+ * *all* loaded modules and if module gets unloaded by somebody else
+ * attempt to dereference the pointer is doomed to have fatal
+ * consequences. Primary usage for this function is to probe *core*
+ * system functionality, e.g. check if getnameinfo(3) is available
+ * at run-time without bothering about OS-specific details such as
+ * libc.so.versioning or where does it actually reside: in libc
+ * itself or libsocket. */
+void *DSO_global_lookup(const char *name);
+
+/* If BeOS is defined, use shared images. If not, return NULL. */
+DSO_METHOD *DSO_METHOD_beos(void);
+
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
  * made after this point may be overwritten when the script is next run.
@@ -305,6 +334,11 @@ void ERR_load_DSO_strings(void);
 /* Error codes for the DSO functions. */
 
 /* Function codes. */
+#define DSO_F_BEOS_BIND_FUNC				 144
+#define DSO_F_BEOS_BIND_VAR				 145
+#define DSO_F_BEOS_LOAD					 146
+#define DSO_F_BEOS_NAME_CONVERTER			 147
+#define DSO_F_BEOS_UNLOAD				 148
 #define DSO_F_DLFCN_BIND_FUNC				 100
 #define DSO_F_DLFCN_BIND_VAR				 101
 #define DSO_F_DLFCN_LOAD				 102
@@ -324,22 +358,29 @@ void ERR_load_DSO_strings(void);
 #define DSO_F_DSO_FREE					 111
 #define DSO_F_DSO_GET_FILENAME				 127
 #define DSO_F_DSO_GET_LOADED_FILENAME			 128
+#define DSO_F_DSO_GLOBAL_LOOKUP				 139
 #define DSO_F_DSO_LOAD					 112
 #define DSO_F_DSO_MERGE					 132
 #define DSO_F_DSO_NEW_METHOD				 113
+#define DSO_F_DSO_PATHBYADDR				 140
 #define DSO_F_DSO_SET_FILENAME				 129
 #define DSO_F_DSO_SET_NAME_CONVERTER			 122
 #define DSO_F_DSO_UP_REF				 114
+#define DSO_F_GLOBAL_LOOKUP_FUNC			 138
+#define DSO_F_PATHBYADDR				 137
 #define DSO_F_VMS_BIND_SYM				 115
 #define DSO_F_VMS_LOAD					 116
 #define DSO_F_VMS_MERGER				 133
 #define DSO_F_VMS_UNLOAD				 117
 #define DSO_F_WIN32_BIND_FUNC				 118
 #define DSO_F_WIN32_BIND_VAR				 119
+#define DSO_F_WIN32_GLOBALLOOKUP			 142
+#define DSO_F_WIN32_GLOBALLOOKUP_FUNC			 143
 #define DSO_F_WIN32_JOINER				 135
 #define DSO_F_WIN32_LOAD				 120
 #define DSO_F_WIN32_MERGER				 134
 #define DSO_F_WIN32_NAME_CONVERTER			 125
+#define DSO_F_WIN32_PATHBYADDR				 141
 #define DSO_F_WIN32_SPLITTER				 136
 #define DSO_F_WIN32_UNLOAD				 121
 
diff --git a/crypto/dso/dso_beos.c b/crypto/dso/dso_beos.c
new file mode 100644
index 000000000000..553966e6993e
--- /dev/null
+++ b/crypto/dso/dso_beos.c
@@ -0,0 +1,270 @@
+/* dso_beos.c */
+/* Written by Marcin Konicki (ahwayakchih@neoni.net) for the OpenSSL
+ * project 2000.
+ */
+/* ====================================================================
+ * Copyright (c) 2000 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include "cryptlib.h"
+#include <openssl/dso.h>
+
+#if !defined(OPENSSL_SYS_BEOS)
+DSO_METHOD *DSO_METHOD_beos(void)
+	{
+	return NULL;
+	}
+#else
+
+#include <kernel/image.h>
+
+static int beos_load(DSO *dso);
+static int beos_unload(DSO *dso);
+static void *beos_bind_var(DSO *dso, const char *symname);
+static DSO_FUNC_TYPE beos_bind_func(DSO *dso, const char *symname);
+#if 0
+static int beos_unbind_var(DSO *dso, char *symname, void *symptr);
+static int beos_unbind_func(DSO *dso, char *symname, DSO_FUNC_TYPE symptr);
+static int beos_init(DSO *dso);
+static int beos_finish(DSO *dso);
+static long beos_ctrl(DSO *dso, int cmd, long larg, void *parg);
+#endif
+static char *beos_name_converter(DSO *dso, const char *filename);
+
+static DSO_METHOD dso_meth_beos = {
+	"OpenSSL 'beos' shared library method",
+	beos_load,
+	beos_unload,
+	beos_bind_var,
+	beos_bind_func,
+/* For now, "unbind" doesn't exist */
+#if 0
+	NULL, /* unbind_var */
+	NULL, /* unbind_func */
+#endif
+	NULL, /* ctrl */
+	beos_name_converter,
+	NULL, /* init */
+	NULL  /* finish */
+	};
+
+DSO_METHOD *DSO_METHOD_beos(void)
+	{
+	return(&dso_meth_beos);
+	}
+
+/* For this DSO_METHOD, our meth_data STACK will contain;
+ * (i) a pointer to the handle (image_id) returned from
+ *     load_add_on().
+ */
+
+static int beos_load(DSO *dso)
+	{
+	image_id id;
+	/* See applicable comments from dso_dl.c */
+	char *filename = DSO_convert_filename(dso, NULL);
+
+	if(filename == NULL)
+		{
+		DSOerr(DSO_F_BEOS_LOAD,DSO_R_NO_FILENAME);
+		goto err;
+		}
+	id = load_add_on(filename);
+	if(id < 1)
+		{
+		DSOerr(DSO_F_BEOS_LOAD,DSO_R_LOAD_FAILED);
+		ERR_add_error_data(3, "filename(", filename, ")");
+		goto err;
+		}
+	if(!sk_push(dso->meth_data, (char *)id))
+		{
+		DSOerr(DSO_F_BEOS_LOAD,DSO_R_STACK_ERROR);
+		goto err;
+		}
+	/* Success */
+	dso->loaded_filename = filename;
+	return(1);
+err:
+	/* Cleanup !*/
+	if(filename != NULL)
+		OPENSSL_free(filename);
+	if(id > 0)
+		unload_add_on(id);
+	return(0);
+	}
+
+static int beos_unload(DSO *dso)
+	{
+	image_id id;
+	if(dso == NULL)
+		{
+		DSOerr(DSO_F_BEOS_UNLOAD,ERR_R_PASSED_NULL_PARAMETER);
+		return(0);
+		}
+	if(sk_num(dso->meth_data) < 1)
+		return(1);
+	id = (image_id)sk_pop(dso->meth_data);
+	if(id < 1)
+		{
+		DSOerr(DSO_F_BEOS_UNLOAD,DSO_R_NULL_HANDLE);
+		return(0);
+		}
+	if(unload_add_on(id) != B_OK)
+		{
+		DSOerr(DSO_F_BEOS_UNLOAD,DSO_R_UNLOAD_FAILED);
+		/* We should push the value back onto the stack in
+		 * case of a retry. */
+		sk_push(dso->meth_data, (char *)id);
+		return(0);
+		}
+	return(1);
+	}
+
+static void *beos_bind_var(DSO *dso, const char *symname)
+	{
+	image_id id;
+	void *sym;
+
+	if((dso == NULL) || (symname == NULL))
+		{
+		DSOerr(DSO_F_BEOS_BIND_VAR,ERR_R_PASSED_NULL_PARAMETER);
+		return(NULL);
+		}
+	if(sk_num(dso->meth_data) < 1)
+		{
+		DSOerr(DSO_F_BEOS_BIND_VAR,DSO_R_STACK_ERROR);
+		return(NULL);
+		}
+	id = (image_id)sk_value(dso->meth_data, sk_num(dso->meth_data) - 1);
+	if(id < 1)
+		{
+		DSOerr(DSO_F_BEOS_BIND_VAR,DSO_R_NULL_HANDLE);
+		return(NULL);
+		}
+	if(get_image_symbol(id, symname, B_SYMBOL_TYPE_DATA, &sym) != B_OK)
+		{
+		DSOerr(DSO_F_BEOS_BIND_VAR,DSO_R_SYM_FAILURE);
+		ERR_add_error_data(3, "symname(", symname, ")");
+		return(NULL);
+		}
+	return(sym);
+	}
+
+static DSO_FUNC_TYPE beos_bind_func(DSO *dso, const char *symname)
+	{
+	image_id id;
+	void *sym;
+
+	if((dso == NULL) || (symname == NULL))
+		{
+		DSOerr(DSO_F_BEOS_BIND_FUNC,ERR_R_PASSED_NULL_PARAMETER);
+		return(NULL);
+		}
+	if(sk_num(dso->meth_data) < 1)
+		{
+		DSOerr(DSO_F_BEOS_BIND_FUNC,DSO_R_STACK_ERROR);
+		return(NULL);
+		}
+	id = (image_id)sk_value(dso->meth_data, sk_num(dso->meth_data) - 1);
+	if(id < 1)
+		{
+		DSOerr(DSO_F_BEOS_BIND_FUNC,DSO_R_NULL_HANDLE);
+		return(NULL);
+		}
+	if(get_image_symbol(id, symname, B_SYMBOL_TYPE_TEXT, &sym) != B_OK)
+		{
+		DSOerr(DSO_F_BEOS_BIND_FUNC,DSO_R_SYM_FAILURE);
+		ERR_add_error_data(3, "symname(", symname, ")");
+		return(NULL);
+		}
+	return((DSO_FUNC_TYPE)sym);
+	}
+
+/* This one is the same as the one in dlfcn */
+static char *beos_name_converter(DSO *dso, const char *filename)
+	{
+	char *translated;
+	int len, rsize, transform;
+
+	len = strlen(filename);
+	rsize = len + 1;
+	transform = (strstr(filename, "/") == NULL);
+	if(transform)
+		{
+		/* We will convert this to "%s.so" or "lib%s.so" */
+		rsize += 3;	/* The length of ".so" */
+		if ((DSO_flags(dso) & DSO_FLAG_NAME_TRANSLATION_EXT_ONLY) == 0)
+			rsize += 3; /* The length of "lib" */
+		}
+	translated = OPENSSL_malloc(rsize);
+	if(translated == NULL)
+		{
+		DSOerr(DSO_F_BEOS_NAME_CONVERTER,
+				DSO_R_NAME_TRANSLATION_FAILED);
+		return(NULL);
+		}
+	if(transform)
+		{
+		if ((DSO_flags(dso) & DSO_FLAG_NAME_TRANSLATION_EXT_ONLY) == 0)
+			sprintf(translated, "lib%s.so", filename);
+		else
+			sprintf(translated, "%s.so", filename);
+		}
+	else
+		sprintf(translated, "%s", filename);
+	return(translated);
+	}
+
+#endif
diff --git a/crypto/dso/dso_dl.c b/crypto/dso/dso_dl.c
index 417abb6ea95f..fc4236bd9ab0 100644
--- a/crypto/dso/dso_dl.c
+++ b/crypto/dso/dso_dl.c
@@ -85,6 +85,8 @@ static int dl_ctrl(DSO *dso, int cmd, long larg, void *parg);
 #endif
 static char *dl_name_converter(DSO *dso, const char *filename);
 static char *dl_merger(DSO *dso, const char *filespec1, const char *filespec2);
+static int dl_pathbyaddr(void *addr,char *path,int sz);
+static void *dl_globallookup(const char *name);
 
 static DSO_METHOD dso_meth_dl = {
 	"OpenSSL 'dl' shared library method",
@@ -101,7 +103,9 @@ static DSO_METHOD dso_meth_dl = {
 	dl_name_converter,
 	dl_merger,
 	NULL, /* init */
-	NULL  /* finish */
+	NULL, /* finish */
+	dl_pathbyaddr,
+	dl_globallookup
 	};
 
 DSO_METHOD *DSO_METHOD_dl(void)
@@ -350,4 +354,40 @@ static char *dl_name_converter(DSO *dso, const char *filename)
 	return(translated);
 	}
 
+static int dl_pathbyaddr(void *addr,char *path,int sz)
+	{
+	struct shl_descriptor inf;
+	int i,len;
+
+	if (addr == NULL)
+		{
+		union	{ int(*f)(void*,char*,int); void *p; } t =
+			{ dl_pathbyaddr };
+		addr = t.p;
+		}
+
+	for (i=-1;shl_get_r(i,&inf)==0;i++)
+		{
+		if (((size_t)addr >= inf.tstart && (size_t)addr < inf.tend) ||
+		    ((size_t)addr >= inf.dstart && (size_t)addr < inf.dend))
+			{
+			len = (int)strlen(inf.filename);
+			if (sz <= 0) return len+1;
+			if (len >= sz) len=sz-1;
+			memcpy(path,inf.filename,len);
+			path[len++] = 0;
+			return len;
+			}
+		}
+
+	return -1;
+	}
+
+static void *dl_globallookup(const char *name)
+	{
+	void *ret;
+	shl_t h = NULL;
+
+	return shl_findsym(&h,name,TYPE_UNDEFINED,&ret) ? NULL : ret;
+	}
 #endif /* DSO_DL */
diff --git a/crypto/dso/dso_dlfcn.c b/crypto/dso/dso_dlfcn.c
index d91e821af675..5f2254806ca9 100644
--- a/crypto/dso/dso_dlfcn.c
+++ b/crypto/dso/dso_dlfcn.c
@@ -56,6 +56,16 @@
  *
  */
 
+/* We need to do this early, because stdio.h includes the header files
+   that handle _GNU_SOURCE and other similar macros.  Defining it later
+   is simply too late, because those headers are protected from re-
+   inclusion.  */
+#ifdef __linux
+# ifndef _GNU_SOURCE
+#  define _GNU_SOURCE	/* make sure dladdr is declared */
+# endif
+#endif
+
 #include <stdio.h>
 #include "cryptlib.h"
 #include <openssl/dso.h>
@@ -68,7 +78,18 @@ DSO_METHOD *DSO_METHOD_dlfcn(void)
 #else
 
 #ifdef HAVE_DLFCN_H
-#include <dlfcn.h>
+# ifdef __osf__
+#  define __EXTENSIONS__
+# endif
+# include <dlfcn.h>
+# define HAVE_DLINFO 1
+# if defined(_AIX) || defined(__CYGWIN__) || \
+     defined(__SCO_VERSION__) || defined(_SCO_ELF) || \
+     (defined(__osf__) && !defined(RTLD_NEXT))     || \
+     (defined(__OpenBSD__) && !defined(RTLD_SELF)) || \
+	defined(__ANDROID__)
+#  undef HAVE_DLINFO
+# endif
 #endif
 
 /* Part of the hack in "dlfcn_load" ... */
@@ -87,6 +108,8 @@ static long dlfcn_ctrl(DSO *dso, int cmd, long larg, void *parg);
 static char *dlfcn_name_converter(DSO *dso, const char *filename);
 static char *dlfcn_merger(DSO *dso, const char *filespec1,
 	const char *filespec2);
+static int dlfcn_pathbyaddr(void *addr,char *path,int sz);
+static void *dlfcn_globallookup(const char *name);
 
 static DSO_METHOD dso_meth_dlfcn = {
 	"OpenSSL 'dlfcn' shared library method",
@@ -103,7 +126,9 @@ static DSO_METHOD dso_meth_dlfcn = {
 	dlfcn_name_converter,
 	dlfcn_merger,
 	NULL, /* init */
-	NULL  /* finish */
+	NULL, /* finish */
+	dlfcn_pathbyaddr,
+	dlfcn_globallookup
 	};
 
 DSO_METHOD *DSO_METHOD_dlfcn(void)
@@ -163,7 +188,7 @@ static int dlfcn_load(DSO *dso)
 		ERR_add_error_data(4, "filename(", filename, "): ", dlerror());
 		goto err;
 		}
-	if(!sk_push(dso->meth_data, (char *)ptr))
+	if(!sk_void_push(dso->meth_data, (char *)ptr))
 		{
 		DSOerr(DSO_F_DLFCN_LOAD,DSO_R_STACK_ERROR);
 		goto err;
@@ -188,15 +213,15 @@ static int dlfcn_unload(DSO *dso)
 		DSOerr(DSO_F_DLFCN_UNLOAD,ERR_R_PASSED_NULL_PARAMETER);
 		return(0);
 		}
-	if(sk_num(dso->meth_data) < 1)
+	if(sk_void_num(dso->meth_data) < 1)
 		return(1);
-	ptr = (void *)sk_pop(dso->meth_data);
+	ptr = sk_void_pop(dso->meth_data);
 	if(ptr == NULL)
 		{
 		DSOerr(DSO_F_DLFCN_UNLOAD,DSO_R_NULL_HANDLE);
 		/* Should push the value back onto the stack in
 		 * case of a retry. */
-		sk_push(dso->meth_data, (char *)ptr);
+		sk_void_push(dso->meth_data, ptr);
 		return(0);
 		}
 	/* For now I'm not aware of any errors associated with dlclose() */
@@ -213,12 +238,12 @@ static void *dlfcn_bind_var(DSO *dso, const char *symname)
 		DSOerr(DSO_F_DLFCN_BIND_VAR,ERR_R_PASSED_NULL_PARAMETER);
 		return(NULL);
 		}
-	if(sk_num(dso->meth_data) < 1)
+	if(sk_void_num(dso->meth_data) < 1)
 		{
 		DSOerr(DSO_F_DLFCN_BIND_VAR,DSO_R_STACK_ERROR);
 		return(NULL);
 		}
-	ptr = (void *)sk_value(dso->meth_data, sk_num(dso->meth_data) - 1);
+	ptr = sk_void_value(dso->meth_data, sk_void_num(dso->meth_data) - 1);
 	if(ptr == NULL)
 		{
 		DSOerr(DSO_F_DLFCN_BIND_VAR,DSO_R_NULL_HANDLE);
@@ -247,12 +272,12 @@ static DSO_FUNC_TYPE dlfcn_bind_func(DSO *dso, const char *symname)
 		DSOerr(DSO_F_DLFCN_BIND_FUNC,ERR_R_PASSED_NULL_PARAMETER);
 		return(NULL);
 		}
-	if(sk_num(dso->meth_data) < 1)
+	if(sk_void_num(dso->meth_data) < 1)
 		{
 		DSOerr(DSO_F_DLFCN_BIND_FUNC,DSO_R_STACK_ERROR);
 		return(NULL);
 		}
-	ptr = (void *)sk_value(dso->meth_data, sk_num(dso->meth_data) - 1);
+	ptr = sk_void_value(dso->meth_data, sk_void_num(dso->meth_data) - 1);
 	if(ptr == NULL)
 		{
 		DSOerr(DSO_F_DLFCN_BIND_FUNC,DSO_R_NULL_HANDLE);
@@ -281,13 +306,12 @@ static char *dlfcn_merger(DSO *dso, const char *filespec1,
 		}
 	/* If the first file specification is a rooted path, it rules.
 	   same goes if the second file specification is missing. */
-	if (!filespec2 || filespec1[0] == '/')
+	if (!filespec2 || (filespec1 != NULL && filespec1[0] == '/'))
 		{
 		merged = OPENSSL_malloc(strlen(filespec1) + 1);
 		if(!merged)
 			{
-			DSOerr(DSO_F_DLFCN_MERGER,
-				ERR_R_MALLOC_FAILURE);
+			DSOerr(DSO_F_DLFCN_MERGER, ERR_R_MALLOC_FAILURE);
 			return(NULL);
 			}
 		strcpy(merged, filespec1);
@@ -313,7 +337,7 @@ static char *dlfcn_merger(DSO *dso, const char *filespec1,
 		{
 		int spec2len, len;
 
-		spec2len = (filespec2 ? strlen(filespec2) : 0);
+		spec2len = strlen(filespec2);
 		len = spec2len + (filespec1 ? strlen(filespec1) : 0);
 
 		if(filespec2 && filespec2[spec2len - 1] == '/')
@@ -378,4 +402,83 @@ static char *dlfcn_name_converter(DSO *dso, const char *filename)
 	return(translated);
 	}
 
+#ifdef __sgi
+/*
+This is a quote from IRIX manual for dladdr(3c):
+
+     <dlfcn.h> does not contain a prototype for dladdr or definition of
+     Dl_info.  The #include <dlfcn.h>  in the SYNOPSIS line is traditional,
+     but contains no dladdr prototype and no IRIX library contains an
+     implementation.  Write your own declaration based on the code below.
+
+     The following code is dependent on internal interfaces that are not
+     part of the IRIX compatibility guarantee; however, there is no future
+     intention to change this interface, so on a practical level, the code
+     below is safe to use on IRIX.
+*/
+#include <rld_interface.h>
+#ifndef _RLD_INTERFACE_DLFCN_H_DLADDR
+#define _RLD_INTERFACE_DLFCN_H_DLADDR
+typedef struct Dl_info {
+    const char * dli_fname;
+    void       * dli_fbase;
+    const char * dli_sname;
+    void       * dli_saddr;
+    int          dli_version;
+    int          dli_reserved1;
+    long         dli_reserved[4];
+} Dl_info;
+#else
+typedef struct Dl_info Dl_info;
+#endif
+#define _RLD_DLADDR             14
+
+static int dladdr(void *address, Dl_info *dl)
+{
+	void *v;
+	v = _rld_new_interface(_RLD_DLADDR,address,dl);
+	return (int)v;
+}
+#endif /* __sgi */
+
+static int dlfcn_pathbyaddr(void *addr,char *path,int sz)
+	{
+#ifdef HAVE_DLINFO
+	Dl_info dli;
+	int len;
+
+	if (addr == NULL)
+		{
+		union	{ int(*f)(void*,char*,int); void *p; } t =
+			{ dlfcn_pathbyaddr };
+		addr = t.p;
+		}
+
+	if (dladdr(addr,&dli))
+		{
+		len = (int)strlen(dli.dli_fname);
+		if (sz <= 0) return len+1;
+		if (len >= sz) len=sz-1;
+		memcpy(path,dli.dli_fname,len);
+		path[len++]=0;
+		return len;
+		}
+
+	ERR_add_error_data(4, "dlfcn_pathbyaddr(): ", dlerror());
+#endif
+	return -1;
+	}
+
+static void *dlfcn_globallookup(const char *name)
+	{
+	void *ret = NULL,*handle = dlopen(NULL,RTLD_LAZY);
+	
+	if (handle)
+		{
+		ret = dlsym(handle,name);
+		dlclose(handle);
+		}
+
+	return ret;
+	}
 #endif /* DSO_DLFCN */
diff --git a/crypto/dso/dso_err.c b/crypto/dso/dso_err.c
index a8b0a210de43..2bb07c25143a 100644
--- a/crypto/dso/dso_err.c
+++ b/crypto/dso/dso_err.c
@@ -1,6 +1,6 @@
 /* crypto/dso/dso_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,6 +70,11 @@
 
 static ERR_STRING_DATA DSO_str_functs[]=
 	{
+{ERR_FUNC(DSO_F_BEOS_BIND_FUNC),	"BEOS_BIND_FUNC"},
+{ERR_FUNC(DSO_F_BEOS_BIND_VAR),	"BEOS_BIND_VAR"},
+{ERR_FUNC(DSO_F_BEOS_LOAD),	"BEOS_LOAD"},
+{ERR_FUNC(DSO_F_BEOS_NAME_CONVERTER),	"BEOS_NAME_CONVERTER"},
+{ERR_FUNC(DSO_F_BEOS_UNLOAD),	"BEOS_UNLOAD"},
 {ERR_FUNC(DSO_F_DLFCN_BIND_FUNC),	"DLFCN_BIND_FUNC"},
 {ERR_FUNC(DSO_F_DLFCN_BIND_VAR),	"DLFCN_BIND_VAR"},
 {ERR_FUNC(DSO_F_DLFCN_LOAD),	"DLFCN_LOAD"},
@@ -89,22 +94,29 @@ static ERR_STRING_DATA DSO_str_functs[]=
 {ERR_FUNC(DSO_F_DSO_FREE),	"DSO_free"},
 {ERR_FUNC(DSO_F_DSO_GET_FILENAME),	"DSO_get_filename"},
 {ERR_FUNC(DSO_F_DSO_GET_LOADED_FILENAME),	"DSO_get_loaded_filename"},
+{ERR_FUNC(DSO_F_DSO_GLOBAL_LOOKUP),	"DSO_global_lookup"},
 {ERR_FUNC(DSO_F_DSO_LOAD),	"DSO_load"},
 {ERR_FUNC(DSO_F_DSO_MERGE),	"DSO_merge"},
 {ERR_FUNC(DSO_F_DSO_NEW_METHOD),	"DSO_new_method"},
+{ERR_FUNC(DSO_F_DSO_PATHBYADDR),	"DSO_pathbyaddr"},
 {ERR_FUNC(DSO_F_DSO_SET_FILENAME),	"DSO_set_filename"},
 {ERR_FUNC(DSO_F_DSO_SET_NAME_CONVERTER),	"DSO_set_name_converter"},
 {ERR_FUNC(DSO_F_DSO_UP_REF),	"DSO_up_ref"},
+{ERR_FUNC(DSO_F_GLOBAL_LOOKUP_FUNC),	"GLOBAL_LOOKUP_FUNC"},
+{ERR_FUNC(DSO_F_PATHBYADDR),	"PATHBYADDR"},
 {ERR_FUNC(DSO_F_VMS_BIND_SYM),	"VMS_BIND_SYM"},
 {ERR_FUNC(DSO_F_VMS_LOAD),	"VMS_LOAD"},
 {ERR_FUNC(DSO_F_VMS_MERGER),	"VMS_MERGER"},
 {ERR_FUNC(DSO_F_VMS_UNLOAD),	"VMS_UNLOAD"},
 {ERR_FUNC(DSO_F_WIN32_BIND_FUNC),	"WIN32_BIND_FUNC"},
 {ERR_FUNC(DSO_F_WIN32_BIND_VAR),	"WIN32_BIND_VAR"},
+{ERR_FUNC(DSO_F_WIN32_GLOBALLOOKUP),	"WIN32_GLOBALLOOKUP"},
+{ERR_FUNC(DSO_F_WIN32_GLOBALLOOKUP_FUNC),	"WIN32_GLOBALLOOKUP_FUNC"},
 {ERR_FUNC(DSO_F_WIN32_JOINER),	"WIN32_JOINER"},
 {ERR_FUNC(DSO_F_WIN32_LOAD),	"WIN32_LOAD"},
 {ERR_FUNC(DSO_F_WIN32_MERGER),	"WIN32_MERGER"},
 {ERR_FUNC(DSO_F_WIN32_NAME_CONVERTER),	"WIN32_NAME_CONVERTER"},
+{ERR_FUNC(DSO_F_WIN32_PATHBYADDR),	"WIN32_PATHBYADDR"},
 {ERR_FUNC(DSO_F_WIN32_SPLITTER),	"WIN32_SPLITTER"},
 {ERR_FUNC(DSO_F_WIN32_UNLOAD),	"WIN32_UNLOAD"},
 {0,NULL}
diff --git a/crypto/dso/dso_lib.c b/crypto/dso/dso_lib.c
index 49bdd7130976..8a15b794abef 100644
--- a/crypto/dso/dso_lib.c
+++ b/crypto/dso/dso_lib.c
@@ -107,7 +107,7 @@ DSO *DSO_new_method(DSO_METHOD *meth)
 		return(NULL);
 		}
 	memset(ret, 0, sizeof(DSO));
-	ret->meth_data = sk_new_null();
+	ret->meth_data = sk_void_new_null();
 	if(ret->meth_data == NULL)
 		{
 		/* sk_new doesn't generate any errors so we do */
@@ -163,7 +163,7 @@ int DSO_free(DSO *dso)
 		return(0);
 		}
 	
-	sk_free(dso->meth_data);
+	sk_void_free(dso->meth_data);
 	if(dso->filename != NULL)
 		OPENSSL_free(dso->filename);
 	if(dso->loaded_filename != NULL)
@@ -399,13 +399,6 @@ char *DSO_merge(DSO *dso, const char *filespec1, const char *filespec2)
 		DSOerr(DSO_F_DSO_MERGE,ERR_R_PASSED_NULL_PARAMETER);
 		return(NULL);
 		}
-	if(filespec1 == NULL)
-		filespec1 = dso->filename;
-	if(filespec1 == NULL)
-		{
-		DSOerr(DSO_F_DSO_MERGE,DSO_R_NO_FILE_SPECIFICATION);
-		return(NULL);
-		}
 	if((dso->flags & DSO_FLAG_NO_NAME_TRANSLATION) == 0)
 		{
 		if(dso->merger != NULL)
@@ -464,3 +457,27 @@ const char *DSO_get_loaded_filename(DSO *dso)
 		}
 	return(dso->loaded_filename);
 	}
+
+int DSO_pathbyaddr(void *addr,char *path,int sz)
+	{
+	DSO_METHOD *meth = default_DSO_meth;
+	if (meth == NULL) meth = DSO_METHOD_openssl();
+	if (meth->pathbyaddr == NULL)
+		{
+		DSOerr(DSO_F_DSO_PATHBYADDR,DSO_R_UNSUPPORTED);
+		return -1;
+		}
+	return (*meth->pathbyaddr)(addr,path,sz);
+	}
+
+void *DSO_global_lookup(const char *name)
+	{
+	DSO_METHOD *meth = default_DSO_meth;
+	if (meth == NULL) meth = DSO_METHOD_openssl();
+	if (meth->globallookup == NULL)
+		{
+		DSOerr(DSO_F_DSO_GLOBAL_LOOKUP,DSO_R_UNSUPPORTED);
+		return NULL;
+		}
+	return (*meth->globallookup)(name);
+	}
diff --git a/crypto/dso/dso_null.c b/crypto/dso/dso_null.c
index 497298465127..49d842d1f567 100644
--- a/crypto/dso/dso_null.c
+++ b/crypto/dso/dso_null.c
@@ -78,7 +78,9 @@ static DSO_METHOD dso_meth_null = {
 	NULL, /* dso_name_converter */
 	NULL, /* dso_merger */
 	NULL, /* init */
-	NULL  /* finish */
+	NULL, /* finish */
+	NULL, /* pathbyaddr */
+	NULL  /* globallookup */
 	};
 
 DSO_METHOD *DSO_METHOD_null(void)
diff --git a/crypto/dso/dso_openssl.c b/crypto/dso/dso_openssl.c
index a4395ebffeca..b17e8e8e9e64 100644
--- a/crypto/dso/dso_openssl.c
+++ b/crypto/dso/dso_openssl.c
@@ -74,6 +74,8 @@ DSO_METHOD *DSO_METHOD_openssl(void)
 	return(DSO_METHOD_win32());
 #elif defined(DSO_VMS)
 	return(DSO_METHOD_vms());
+#elif defined(DSO_BEOS)
+	return(DSO_METHOD_beos());
 #else
 	return(DSO_METHOD_null());
 #endif
diff --git a/crypto/dyn_lck.c b/crypto/dyn_lck.c
deleted file mode 100644
index 7f82c4126493..000000000000
--- a/crypto/dyn_lck.c
+++ /dev/null
@@ -1,428 +0,0 @@
-/* crypto/cryptlib.c */
-/* ====================================================================
- * Copyright (c) 1998-2003 The OpenSSL Project.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    openssl-core@openssl.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
- */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-/* ====================================================================
- * Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
- * ECDH support in OpenSSL originally developed by 
- * SUN MICROSYSTEMS, INC., and contributed to the OpenSSL project.
- */
-
-#include "cryptlib.h"
-#include <openssl/safestack.h>
-
-#if defined(OPENSSL_SYS_WIN32) || defined(OPENSSL_SYS_WIN16)
-static double SSLeay_MSVC5_hack=0.0; /* and for VC1.5 */
-#endif
-
-DECLARE_STACK_OF(CRYPTO_dynlock)
-IMPLEMENT_STACK_OF(CRYPTO_dynlock)
-
-/* real #defines in crypto.h, keep these upto date */
-static const char* const lock_names[CRYPTO_NUM_LOCKS] =
-	{
-	"<<ERROR>>",
-	"err",
-	"ex_data",
-	"x509",
-	"x509_info",
-	"x509_pkey",
-	"x509_crl",
-	"x509_req",
-	"dsa",
-	"rsa",
-	"evp_pkey",
-	"x509_store",
-	"ssl_ctx",
-	"ssl_cert",
-	"ssl_session",
-	"ssl_sess_cert",
-	"ssl",
-	"ssl_method",
-	"rand",
-	"rand2",
-	"debug_malloc",
-	"BIO",
-	"gethostbyname",
-	"getservbyname",
-	"readdir",
-	"RSA_blinding",
-	"dh",
-	"debug_malloc2",
-	"dso",
-	"dynlock",
-	"engine",
-	"ui",
-	"ecdsa",
-	"ec",
-	"ecdh",
-	"bn",
-	"ec_pre_comp",
-	"store",
-	"comp",
-#ifndef OPENSSL_FIPS
-# if CRYPTO_NUM_LOCKS != 39
-#  error "Inconsistency between crypto.h and cryptlib.c"
-# endif
-#else
-	"fips",
-	"fips2",
-# if CRYPTO_NUM_LOCKS != 41
-#  error "Inconsistency between crypto.h and cryptlib.c"
-# endif
-#endif
-	};
-
-/* This is for applications to allocate new type names in the non-dynamic
-   array of lock names.  These are numbered with positive numbers.  */
-static STACK *app_locks=NULL;
-
-/* For applications that want a more dynamic way of handling threads, the
-   following stack is used.  These are externally numbered with negative
-   numbers.  */
-static STACK_OF(CRYPTO_dynlock) *dyn_locks=NULL;
-
-
-static struct CRYPTO_dynlock_value *(MS_FAR *dynlock_create_callback)
-	(const char *file,int line)=NULL;
-static void (MS_FAR *dynlock_lock_callback)(int mode,
-	struct CRYPTO_dynlock_value *l, const char *file,int line)=NULL;
-static void (MS_FAR *dynlock_destroy_callback)(struct CRYPTO_dynlock_value *l,
-	const char *file,int line)=NULL;
-
-int CRYPTO_get_new_lockid(char *name)
-	{
-	char *str;
-	int i;
-
-#if defined(OPENSSL_SYS_WIN32) || defined(OPENSSL_SYS_WIN16)
-	/* A hack to make Visual C++ 5.0 work correctly when linking as
-	 * a DLL using /MT. Without this, the application cannot use
-	 * and floating point printf's.
-	 * It also seems to be needed for Visual C 1.5 (win16) */
-	SSLeay_MSVC5_hack=(double)name[0]*(double)name[1];
-#endif
-
-	if ((app_locks == NULL) && ((app_locks=sk_new_null()) == NULL))
-		{
-		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_LOCKID,ERR_R_MALLOC_FAILURE);
-		return(0);
-		}
-	if ((str=BUF_strdup(name)) == NULL)
-		{
-		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_LOCKID,ERR_R_MALLOC_FAILURE);
-		return(0);
-		}
-	i=sk_push(app_locks,str);
-	if (!i)
-		OPENSSL_free(str);
-	else
-		i+=CRYPTO_NUM_LOCKS; /* gap of one :-) */
-	return(i);
-	}
-
-int CRYPTO_get_new_dynlockid(void)
-	{
-	int i = 0;
-	CRYPTO_dynlock *pointer = NULL;
-
-	if (dynlock_create_callback == NULL)
-		{
-		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_DYNLOCKID,CRYPTO_R_NO_DYNLOCK_CREATE_CALLBACK);
-		return(0);
-		}
-	CRYPTO_w_lock(CRYPTO_LOCK_DYNLOCK);
-	if ((dyn_locks == NULL)
-		&& ((dyn_locks=sk_CRYPTO_dynlock_new_null()) == NULL))
-		{
-		CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
-		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_DYNLOCKID,ERR_R_MALLOC_FAILURE);
-		return(0);
-		}
-	CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
-
-	pointer = (CRYPTO_dynlock *)OPENSSL_malloc(sizeof(CRYPTO_dynlock));
-	if (pointer == NULL)
-		{
-		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_DYNLOCKID,ERR_R_MALLOC_FAILURE);
-		return(0);
-		}
-	pointer->references = 1;
-	pointer->data = dynlock_create_callback(__FILE__,__LINE__);
-	if (pointer->data == NULL)
-		{
-		OPENSSL_free(pointer);
-		CRYPTOerr(CRYPTO_F_CRYPTO_GET_NEW_DYNLOCKID,ERR_R_MALLOC_FAILURE);
-		return(0);
-		}
-
-	CRYPTO_w_lock(CRYPTO_LOCK_DYNLOCK);
-	/* First, try to find an existing empty slot */
-	i=sk_CRYPTO_dynlock_find(dyn_locks,NULL);
-	/* If there was none, push, thereby creating a new one */
-	if (i == -1)
-		/* Since sk_push() returns the number of items on the
-		   stack, not the location of the pushed item, we need
-		   to transform the returned number into a position,
-		   by decreasing it.  */
-		i=sk_CRYPTO_dynlock_push(dyn_locks,pointer) - 1;
-	else
-		/* If we found a place with a NULL pointer, put our pointer
-		   in it.  */
-		(void)sk_CRYPTO_dynlock_set(dyn_locks,i,pointer);
-	CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
-
-	if (i == -1)
-		{
-		dynlock_destroy_callback(pointer->data,__FILE__,__LINE__);
-		OPENSSL_free(pointer);
-		}
-	else
-		i += 1; /* to avoid 0 */
-	return -i;
-	}
-
-void CRYPTO_destroy_dynlockid(int i)
-	{
-	CRYPTO_dynlock *pointer = NULL;
-	if (i)
-		i = -i-1;
-	if (dynlock_destroy_callback == NULL)
-		return;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_DYNLOCK);
-
-	if (dyn_locks == NULL || i >= sk_CRYPTO_dynlock_num(dyn_locks))
-		{
-		CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
-		return;
-		}
-	pointer = sk_CRYPTO_dynlock_value(dyn_locks, i);
-	if (pointer != NULL)
-		{
-		--pointer->references;
-#ifdef REF_CHECK
-		if (pointer->references < 0)
-			{
-			fprintf(stderr,"CRYPTO_destroy_dynlockid, bad reference count\n");
-			abort();
-			}
-		else
-#endif
-			if (pointer->references <= 0)
-				{
-				(void)sk_CRYPTO_dynlock_set(dyn_locks, i, NULL);
-				}
-			else
-				pointer = NULL;
-		}
-	CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
-
-	if (pointer)
-		{
-		dynlock_destroy_callback(pointer->data,__FILE__,__LINE__);
-		OPENSSL_free(pointer);
-		}
-	}
-
-struct CRYPTO_dynlock_value *CRYPTO_get_dynlock_value(int i)
-	{
-	CRYPTO_dynlock *pointer = NULL;
-	if (i)
-		i = -i-1;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_DYNLOCK);
-
-	if (dyn_locks != NULL && i < sk_CRYPTO_dynlock_num(dyn_locks))
-		pointer = sk_CRYPTO_dynlock_value(dyn_locks, i);
-	if (pointer)
-		pointer->references++;
-
-	CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
-
-	if (pointer)
-		return pointer->data;
-	return NULL;
-	}
-
-struct CRYPTO_dynlock_value *(*CRYPTO_get_dynlock_create_callback(void))
-	(const char *file,int line)
-	{
-	return(dynlock_create_callback);
-	}
-
-void (*CRYPTO_get_dynlock_lock_callback(void))(int mode,
-	struct CRYPTO_dynlock_value *l, const char *file,int line)
-	{
-	return(dynlock_lock_callback);
-	}
-
-void (*CRYPTO_get_dynlock_destroy_callback(void))
-	(struct CRYPTO_dynlock_value *l, const char *file,int line)
-	{
-	return(dynlock_destroy_callback);
-	}
-
-void CRYPTO_set_dynlock_create_callback(struct CRYPTO_dynlock_value *(*func)
-	(const char *file, int line))
-	{
-	dynlock_create_callback=func;
-	}
-
-static void do_dynlock(int mode, int type, const char *file, int line)
-	{
-	if (dynlock_lock_callback != NULL)
-		{
-		struct CRYPTO_dynlock_value *pointer
-				= CRYPTO_get_dynlock_value(type);
-
-		OPENSSL_assert(pointer != NULL);
-
-		dynlock_lock_callback(mode, pointer, file, line);
-
-		CRYPTO_destroy_dynlockid(type);
-		}
-	}
-
-void CRYPTO_set_dynlock_lock_callback(void (*func)(int mode,
-	struct CRYPTO_dynlock_value *l, const char *file, int line))
-	{
-	/* Set callback so CRYPTO_lock() can now handle dynamic locks.
-	 * This is OK because at this point and application shouldn't be using
-	 * OpenSSL from multiple threads because it is setting up the locking
-	 * callbacks.
-	 */
-	static int done = 0;
-	if (!done)
-		{
-		int_CRYPTO_set_do_dynlock_callback(do_dynlock);
-		done = 1;
-		}
-		
-	dynlock_lock_callback=func;
-	}
-
-void CRYPTO_set_dynlock_destroy_callback(void (*func)
-	(struct CRYPTO_dynlock_value *l, const char *file, int line))
-	{
-	dynlock_destroy_callback=func;
-	}
-
-const char *CRYPTO_get_lock_name(int type)
-	{
-	if (type < 0)
-		return("dynamic");
-	else if (type < CRYPTO_NUM_LOCKS)
-		return(lock_names[type]);
-	else if (type-CRYPTO_NUM_LOCKS > sk_num(app_locks))
-		return("ERROR");
-	else
-		return(sk_value(app_locks,type-CRYPTO_NUM_LOCKS));
-	}
-
diff --git a/crypto/ec/Makefile b/crypto/ec/Makefile
index b5bbc9faa1a2..f85fc845ca2b 100644
--- a/crypto/ec/Makefile
+++ b/crypto/ec/Makefile
@@ -19,11 +19,15 @@ APPS=
 LIB=$(TOP)/libcrypto.a
 LIBSRC=	ec_lib.c ecp_smpl.c ecp_mont.c ecp_nist.c ec_cvt.c ec_mult.c\
 	ec_err.c ec_curve.c ec_check.c ec_print.c ec_asn1.c ec_key.c\
-	ec2_smpl.c ec2_smpt.c ec2_mult.c
+	ec2_smpl.c ec2_mult.c ec_ameth.c ec_pmeth.c eck_prn.c \
+	ecp_nistp224.c ecp_nistp256.c ecp_nistp521.c ecp_nistputil.c \
+	ecp_oct.c ec2_oct.c ec_oct.c
 
 LIBOBJ=	ec_lib.o ecp_smpl.o ecp_mont.o ecp_nist.o ec_cvt.o ec_mult.o\
 	ec_err.o ec_curve.o ec_check.o ec_print.o ec_asn1.o ec_key.o\
-	ec2_smpl.o ec2_mult.o
+	ec2_smpl.o ec2_mult.o ec_ameth.o ec_pmeth.o eck_prn.o \
+	ecp_nistp224.o ecp_nistp256.o ecp_nistp521.o ecp_nistputil.o \
+	ecp_oct.o ec2_oct.o ec_oct.o
 
 SRC= $(LIBSRC)
 
@@ -38,7 +42,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -87,6 +91,14 @@ ec2_mult.o: ../../include/openssl/obj_mac.h ../../include/openssl/opensslconf.h
 ec2_mult.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ec2_mult.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 ec2_mult.o: ../../include/openssl/symhacks.h ec2_mult.c ec_lcl.h
+ec2_oct.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+ec2_oct.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
+ec2_oct.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ec2_oct.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
+ec2_oct.o: ../../include/openssl/obj_mac.h ../../include/openssl/opensslconf.h
+ec2_oct.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ec2_oct.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+ec2_oct.o: ../../include/openssl/symhacks.h ec2_oct.c ec_lcl.h
 ec2_smpl.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 ec2_smpl.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
 ec2_smpl.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
@@ -94,8 +106,22 @@ ec2_smpl.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
 ec2_smpl.o: ../../include/openssl/obj_mac.h ../../include/openssl/opensslconf.h
 ec2_smpl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ec2_smpl.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-ec2_smpl.o: ../../include/openssl/symhacks.h ec2_smpl.c ec2_smpt.c ec_lcl.h
-ec2_smpt.o: ec2_smpt.c
+ec2_smpl.o: ../../include/openssl/symhacks.h ec2_smpl.c ec_lcl.h
+ec_ameth.o: ../../e_os.h ../../include/openssl/asn1.h
+ec_ameth.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+ec_ameth.o: ../../include/openssl/buffer.h ../../include/openssl/cms.h
+ec_ameth.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+ec_ameth.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+ec_ameth.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+ec_ameth.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ec_ameth.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ec_ameth.o: ../../include/openssl/opensslconf.h
+ec_ameth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ec_ameth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+ec_ameth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ec_ameth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+ec_ameth.o: ../../include/openssl/x509_vfy.h ../asn1/asn1_locl.h ../cryptlib.h
+ec_ameth.o: ec_ameth.c
 ec_asn1.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
 ec_asn1.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 ec_asn1.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
@@ -160,6 +186,28 @@ ec_mult.o: ../../include/openssl/obj_mac.h ../../include/openssl/opensslconf.h
 ec_mult.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ec_mult.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 ec_mult.o: ../../include/openssl/symhacks.h ec_lcl.h ec_mult.c
+ec_oct.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+ec_oct.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
+ec_oct.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ec_oct.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
+ec_oct.o: ../../include/openssl/obj_mac.h ../../include/openssl/opensslconf.h
+ec_oct.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ec_oct.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+ec_oct.o: ../../include/openssl/symhacks.h ec_lcl.h ec_oct.c
+ec_pmeth.o: ../../e_os.h ../../include/openssl/asn1.h
+ec_pmeth.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+ec_pmeth.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+ec_pmeth.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ec_pmeth.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ec_pmeth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+ec_pmeth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ec_pmeth.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+ec_pmeth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ec_pmeth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+ec_pmeth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ec_pmeth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+ec_pmeth.o: ../../include/openssl/x509_vfy.h ../cryptlib.h ../evp/evp_locl.h
+ec_pmeth.o: ec_pmeth.c
 ec_print.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 ec_print.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
 ec_print.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
@@ -167,6 +215,16 @@ ec_print.o: ../../include/openssl/obj_mac.h ../../include/openssl/opensslconf.h
 ec_print.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ec_print.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 ec_print.o: ../../include/openssl/symhacks.h ec_lcl.h ec_print.c
+eck_prn.o: ../../e_os.h ../../include/openssl/asn1.h
+eck_prn.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+eck_prn.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+eck_prn.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+eck_prn.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+eck_prn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+eck_prn.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+eck_prn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+eck_prn.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+eck_prn.o: ../../include/openssl/symhacks.h ../cryptlib.h eck_prn.c
 ecp_mont.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 ecp_mont.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
 ecp_mont.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
@@ -183,6 +241,18 @@ ecp_nist.o: ../../include/openssl/obj_mac.h ../../include/openssl/opensslconf.h
 ecp_nist.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ecp_nist.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 ecp_nist.o: ../../include/openssl/symhacks.h ec_lcl.h ecp_nist.c
+ecp_nistp224.o: ../../include/openssl/opensslconf.h ecp_nistp224.c
+ecp_nistp256.o: ../../include/openssl/opensslconf.h ecp_nistp256.c
+ecp_nistp521.o: ../../include/openssl/opensslconf.h ecp_nistp521.c
+ecp_nistputil.o: ../../include/openssl/opensslconf.h ecp_nistputil.c
+ecp_oct.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+ecp_oct.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
+ecp_oct.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ecp_oct.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
+ecp_oct.o: ../../include/openssl/obj_mac.h ../../include/openssl/opensslconf.h
+ecp_oct.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ecp_oct.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+ecp_oct.o: ../../include/openssl/symhacks.h ec_lcl.h ecp_oct.c
 ecp_smpl.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 ecp_smpl.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
 ecp_smpl.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
diff --git a/crypto/ec/ec.h b/crypto/ec/ec.h
index 8bc2a235b1a7..9d01325af33b 100644
--- a/crypto/ec/ec.h
+++ b/crypto/ec/ec.h
@@ -2,8 +2,12 @@
 /*
  * Originally written by Bodo Moeller for the OpenSSL project.
  */
+/**
+ * \file crypto/ec/ec.h Include file for the OpenSSL EC functions
+ * \author Originally written by Bodo Moeller for the OpenSSL project
+ */
 /* ====================================================================
- * Copyright (c) 1998-2003 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1998-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -92,15 +96,21 @@ extern "C" {
 # endif
 #endif
 
-
+  
 #ifndef OPENSSL_ECC_MAX_FIELD_BITS
 # define OPENSSL_ECC_MAX_FIELD_BITS 661
 #endif
 
+/** Enum for the point conversion form as defined in X9.62 (ECDSA)
+ *  for the encoding of a elliptic curve point (x,y) */
 typedef enum {
-	/* values as defined in X9.62 (ECDSA) and elsewhere */
+	/** the point is encoded as z||x, where the octet z specifies 
+	 *  which solution of the quadratic equation y is  */
 	POINT_CONVERSION_COMPRESSED = 2,
+	/** the point is encoded as z||x||y, where z is the octet 0x02  */
 	POINT_CONVERSION_UNCOMPRESSED = 4,
+	/** the point is encoded as z||x||y, where the octet z specifies
+         *  which solution of the quadratic equation y is  */
 	POINT_CONVERSION_HYBRID = 6
 } point_conversion_form_t;
 
@@ -121,37 +131,148 @@ typedef struct ec_group_st
 typedef struct ec_point_st EC_POINT;
 
 
-/* EC_METHODs for curves over GF(p).
- * EC_GFp_simple_method provides the basis for the optimized methods.
+/********************************************************************/
+/*               EC_METHODs for curves over GF(p)                   */       
+/********************************************************************/
+
+/** Returns the basic GFp ec methods which provides the basis for the
+ *  optimized methods. 
+ *  \return  EC_METHOD object
  */
 const EC_METHOD *EC_GFp_simple_method(void);
+
+/** Returns GFp methods using montgomery multiplication.
+ *  \return  EC_METHOD object
+ */
 const EC_METHOD *EC_GFp_mont_method(void);
+
+/** Returns GFp methods using optimized methods for NIST recommended curves
+ *  \return  EC_METHOD object
+ */
 const EC_METHOD *EC_GFp_nist_method(void);
 
-/* EC_METHOD for curves over GF(2^m).
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+/** Returns 64-bit optimized methods for nistp224
+ *  \return  EC_METHOD object
+ */
+const EC_METHOD *EC_GFp_nistp224_method(void);
+
+/** Returns 64-bit optimized methods for nistp256
+ *  \return  EC_METHOD object
+ */
+const EC_METHOD *EC_GFp_nistp256_method(void);
+
+/** Returns 64-bit optimized methods for nistp521
+ *  \return  EC_METHOD object
+ */
+const EC_METHOD *EC_GFp_nistp521_method(void);
+#endif
+
+#ifndef OPENSSL_NO_EC2M
+/********************************************************************/ 
+/*           EC_METHOD for curves over GF(2^m)                      */
+/********************************************************************/
+
+/** Returns the basic GF2m ec method 
+ *  \return  EC_METHOD object
  */
 const EC_METHOD *EC_GF2m_simple_method(void);
 
+#endif
+
 
-EC_GROUP *EC_GROUP_new(const EC_METHOD *);
-void EC_GROUP_free(EC_GROUP *);
-void EC_GROUP_clear_free(EC_GROUP *);
-int EC_GROUP_copy(EC_GROUP *, const EC_GROUP *);
-EC_GROUP *EC_GROUP_dup(const EC_GROUP *);
+/********************************************************************/
+/*                   EC_GROUP functions                             */
+/********************************************************************/
 
-const EC_METHOD *EC_GROUP_method_of(const EC_GROUP *);
-int EC_METHOD_get_field_type(const EC_METHOD *);
+/** Creates a new EC_GROUP object
+ *  \param   meth  EC_METHOD to use
+ *  \return  newly created EC_GROUP object or NULL in case of an error.
+ */
+EC_GROUP *EC_GROUP_new(const EC_METHOD *meth);
 
-int EC_GROUP_set_generator(EC_GROUP *, const EC_POINT *generator, const BIGNUM *order, const BIGNUM *cofactor);
-const EC_POINT *EC_GROUP_get0_generator(const EC_GROUP *);
-int EC_GROUP_get_order(const EC_GROUP *, BIGNUM *order, BN_CTX *);
-int EC_GROUP_get_cofactor(const EC_GROUP *, BIGNUM *cofactor, BN_CTX *);
+/** Frees a EC_GROUP object
+ *  \param  group  EC_GROUP object to be freed.
+ */
+void EC_GROUP_free(EC_GROUP *group);
 
-void EC_GROUP_set_curve_name(EC_GROUP *, int nid);
-int EC_GROUP_get_curve_name(const EC_GROUP *);
+/** Clears and frees a EC_GROUP object
+ *  \param  group  EC_GROUP object to be cleared and freed.
+ */
+void EC_GROUP_clear_free(EC_GROUP *group);
 
-void EC_GROUP_set_asn1_flag(EC_GROUP *, int flag);
-int EC_GROUP_get_asn1_flag(const EC_GROUP *);
+/** Copies EC_GROUP objects. Note: both EC_GROUPs must use the same EC_METHOD.
+ *  \param  dst  destination EC_GROUP object
+ *  \param  src  source EC_GROUP object
+ *  \return 1 on success and 0 if an error occurred.
+ */
+int EC_GROUP_copy(EC_GROUP *dst, const EC_GROUP *src);
+
+/** Creates a new EC_GROUP object and copies the copies the content
+ *  form src to the newly created EC_KEY object
+ *  \param  src  source EC_GROUP object
+ *  \return newly created EC_GROUP object or NULL in case of an error.
+ */
+EC_GROUP *EC_GROUP_dup(const EC_GROUP *src);
+
+/** Returns the EC_METHOD of the EC_GROUP object.
+ *  \param  group  EC_GROUP object 
+ *  \return EC_METHOD used in this EC_GROUP object.
+ */
+const EC_METHOD *EC_GROUP_method_of(const EC_GROUP *group);
+
+/** Returns the field type of the EC_METHOD.
+ *  \param  meth  EC_METHOD object
+ *  \return NID of the underlying field type OID.
+ */
+int EC_METHOD_get_field_type(const EC_METHOD *meth);
+
+/** Sets the generator and it's order/cofactor of a EC_GROUP object.
+ *  \param  group      EC_GROUP object 
+ *  \param  generator  EC_POINT object with the generator.
+ *  \param  order      the order of the group generated by the generator.
+ *  \param  cofactor   the index of the sub-group generated by the generator
+ *                     in the group of all points on the elliptic curve.
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_GROUP_set_generator(EC_GROUP *group, const EC_POINT *generator, const BIGNUM *order, const BIGNUM *cofactor);
+
+/** Returns the generator of a EC_GROUP object.
+ *  \param  group  EC_GROUP object
+ *  \return the currently used generator (possibly NULL).
+ */
+const EC_POINT *EC_GROUP_get0_generator(const EC_GROUP *group);
+
+/** Gets the order of a EC_GROUP
+ *  \param  group  EC_GROUP object
+ *  \param  order  BIGNUM to which the order is copied
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_GROUP_get_order(const EC_GROUP *group, BIGNUM *order, BN_CTX *ctx);
+
+/** Gets the cofactor of a EC_GROUP
+ *  \param  group     EC_GROUP object
+ *  \param  cofactor  BIGNUM to which the cofactor is copied
+ *  \param  ctx       BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_GROUP_get_cofactor(const EC_GROUP *group, BIGNUM *cofactor, BN_CTX *ctx);
+
+/** Sets the name of a EC_GROUP object
+ *  \param  group  EC_GROUP object
+ *  \param  nid    NID of the curve name OID
+ */
+void EC_GROUP_set_curve_name(EC_GROUP *group, int nid);
+
+/** Returns the curve name of a EC_GROUP object
+ *  \param  group  EC_GROUP object
+ *  \return NID of the curve name OID or 0 if not set.
+ */
+int EC_GROUP_get_curve_name(const EC_GROUP *group);
+
+void EC_GROUP_set_asn1_flag(EC_GROUP *group, int flag);
+int EC_GROUP_get_asn1_flag(const EC_GROUP *group);
 
 void EC_GROUP_set_point_conversion_form(EC_GROUP *, point_conversion_form_t);
 point_conversion_form_t EC_GROUP_get_point_conversion_form(const EC_GROUP *);
@@ -160,36 +281,115 @@ unsigned char *EC_GROUP_get0_seed(const EC_GROUP *);
 size_t EC_GROUP_get_seed_len(const EC_GROUP *);
 size_t EC_GROUP_set_seed(EC_GROUP *, const unsigned char *, size_t len);
 
-int EC_GROUP_set_curve_GFp(EC_GROUP *, const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *);
-int EC_GROUP_get_curve_GFp(const EC_GROUP *, BIGNUM *p, BIGNUM *a, BIGNUM *b, BN_CTX *);
-int EC_GROUP_set_curve_GF2m(EC_GROUP *, const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *);
-int EC_GROUP_get_curve_GF2m(const EC_GROUP *, BIGNUM *p, BIGNUM *a, BIGNUM *b, BN_CTX *);
-
-/* returns the number of bits needed to represent a field element */
-int EC_GROUP_get_degree(const EC_GROUP *);
+/** Sets the parameter of a ec over GFp defined by y^2 = x^3 + a*x + b
+ *  \param  group  EC_GROUP object
+ *  \param  p      BIGNUM with the prime number
+ *  \param  a      BIGNUM with parameter a of the equation
+ *  \param  b      BIGNUM with parameter b of the equation
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_GROUP_set_curve_GFp(EC_GROUP *group, const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx);
+
+/** Gets the parameter of the ec over GFp defined by y^2 = x^3 + a*x + b
+ *  \param  group  EC_GROUP object
+ *  \param  p      BIGNUM for the prime number
+ *  \param  a      BIGNUM for parameter a of the equation
+ *  \param  b      BIGNUM for parameter b of the equation
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_GROUP_get_curve_GFp(const EC_GROUP *group, BIGNUM *p, BIGNUM *a, BIGNUM *b, BN_CTX *ctx);
+
+#ifndef OPENSSL_NO_EC2M
+/** Sets the parameter of a ec over GF2m defined by y^2 + x*y = x^3 + a*x^2 + b
+ *  \param  group  EC_GROUP object
+ *  \param  p      BIGNUM with the polynomial defining the underlying field
+ *  \param  a      BIGNUM with parameter a of the equation
+ *  \param  b      BIGNUM with parameter b of the equation
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_GROUP_set_curve_GF2m(EC_GROUP *group, const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx);
+
+/** Gets the parameter of the ec over GF2m defined by y^2 + x*y = x^3 + a*x^2 + b
+ *  \param  group  EC_GROUP object
+ *  \param  p      BIGNUM for the polynomial defining the underlying field
+ *  \param  a      BIGNUM for parameter a of the equation
+ *  \param  b      BIGNUM for parameter b of the equation
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_GROUP_get_curve_GF2m(const EC_GROUP *group, BIGNUM *p, BIGNUM *a, BIGNUM *b, BN_CTX *ctx);
+#endif
+/** Returns the number of bits needed to represent a field element 
+ *  \param  group  EC_GROUP object
+ *  \return number of bits needed to represent a field element
+ */
+int EC_GROUP_get_degree(const EC_GROUP *group);
 
-/* EC_GROUP_check() returns 1 if 'group' defines a valid group, 0 otherwise */
+/** Checks whether the parameter in the EC_GROUP define a valid ec group
+ *  \param  group  EC_GROUP object
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 if group is a valid ec group and 0 otherwise
+ */
 int EC_GROUP_check(const EC_GROUP *group, BN_CTX *ctx);
-/* EC_GROUP_check_discriminant() returns 1 if the discriminant of the
- * elliptic curve is not zero, 0 otherwise */
-int EC_GROUP_check_discriminant(const EC_GROUP *, BN_CTX *);
 
-/* EC_GROUP_cmp() returns 0 if both groups are equal and 1 otherwise */
-int EC_GROUP_cmp(const EC_GROUP *, const EC_GROUP *, BN_CTX *);
+/** Checks whether the discriminant of the elliptic curve is zero or not
+ *  \param  group  EC_GROUP object
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 if the discriminant is not zero and 0 otherwise
+ */
+int EC_GROUP_check_discriminant(const EC_GROUP *group, BN_CTX *ctx);
+
+/** Compares two EC_GROUP objects
+ *  \param  a    first EC_GROUP object
+ *  \param  b    second EC_GROUP object
+ *  \param  ctx  BN_CTX object (optional)
+ *  \return 0 if both groups are equal and 1 otherwise
+ */
+int EC_GROUP_cmp(const EC_GROUP *a, const EC_GROUP *b, BN_CTX *ctx);
 
 /* EC_GROUP_new_GF*() calls EC_GROUP_new() and EC_GROUP_set_GF*()
  * after choosing an appropriate EC_METHOD */
-EC_GROUP *EC_GROUP_new_curve_GFp(const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *);
-EC_GROUP *EC_GROUP_new_curve_GF2m(const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *);
 
-/* EC_GROUP_new_by_curve_name() creates a EC_GROUP structure
- * specified by a curve name (in form of a NID) */
+/** Creates a new EC_GROUP object with the specified parameters defined
+ *  over GFp (defined by the equation y^2 = x^3 + a*x + b)
+ *  \param  p    BIGNUM with the prime number
+ *  \param  a    BIGNUM with the parameter a of the equation
+ *  \param  b    BIGNUM with the parameter b of the equation
+ *  \param  ctx  BN_CTX object (optional)
+ *  \return newly created EC_GROUP object with the specified parameters
+ */
+EC_GROUP *EC_GROUP_new_curve_GFp(const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx);
+#ifndef OPENSSL_NO_EC2M
+/** Creates a new EC_GROUP object with the specified parameters defined
+ *  over GF2m (defined by the equation y^2 + x*y = x^3 + a*x^2 + b)
+ *  \param  p    BIGNUM with the polynomial defining the underlying field
+ *  \param  a    BIGNUM with the parameter a of the equation
+ *  \param  b    BIGNUM with the parameter b of the equation
+ *  \param  ctx  BN_CTX object (optional)
+ *  \return newly created EC_GROUP object with the specified parameters
+ */
+EC_GROUP *EC_GROUP_new_curve_GF2m(const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx);
+#endif
+/** Creates a EC_GROUP object with a curve specified by a NID
+ *  \param  nid  NID of the OID of the curve name
+ *  \return newly created EC_GROUP object with specified curve or NULL
+ *          if an error occurred
+ */
 EC_GROUP *EC_GROUP_new_by_curve_name(int nid);
-/* handling of internal curves */
+
+
+/********************************************************************/
+/*               handling of internal curves                        */
+/********************************************************************/
+
 typedef struct { 
 	int nid;
 	const char *comment;
 	} EC_builtin_curve;
+
 /* EC_builtin_curves(EC_builtin_curve *r, size_t size) returns number 
  * of all available curves or zero if a error occurred. 
  * In case r ist not zero nitems EC_builtin_curve structures 
@@ -197,39 +397,168 @@ typedef struct {
 size_t EC_get_builtin_curves(EC_builtin_curve *r, size_t nitems);
 
 
-/* EC_POINT functions */
+/********************************************************************/
+/*                    EC_POINT functions                            */
+/********************************************************************/
+
+/** Creates a new EC_POINT object for the specified EC_GROUP
+ *  \param  group  EC_GROUP the underlying EC_GROUP object
+ *  \return newly created EC_POINT object or NULL if an error occurred
+ */
+EC_POINT *EC_POINT_new(const EC_GROUP *group);
+
+/** Frees a EC_POINT object
+ *  \param  point  EC_POINT object to be freed
+ */
+void EC_POINT_free(EC_POINT *point);
 
-EC_POINT *EC_POINT_new(const EC_GROUP *);
-void EC_POINT_free(EC_POINT *);
-void EC_POINT_clear_free(EC_POINT *);
-int EC_POINT_copy(EC_POINT *, const EC_POINT *);
-EC_POINT *EC_POINT_dup(const EC_POINT *, const EC_GROUP *);
+/** Clears and frees a EC_POINT object
+ *  \param  point  EC_POINT object to be cleared and freed
+ */
+void EC_POINT_clear_free(EC_POINT *point);
+
+/** Copies EC_POINT object
+ *  \param  dst  destination EC_POINT object
+ *  \param  src  source EC_POINT object
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_copy(EC_POINT *dst, const EC_POINT *src);
+
+/** Creates a new EC_POINT object and copies the content of the supplied
+ *  EC_POINT
+ *  \param  src    source EC_POINT object
+ *  \param  group  underlying the EC_GROUP object
+ *  \return newly created EC_POINT object or NULL if an error occurred 
+ */
+EC_POINT *EC_POINT_dup(const EC_POINT *src, const EC_GROUP *group);
  
-const EC_METHOD *EC_POINT_method_of(const EC_POINT *);
-
-int EC_POINT_set_to_infinity(const EC_GROUP *, EC_POINT *);
-int EC_POINT_set_Jprojective_coordinates_GFp(const EC_GROUP *, EC_POINT *,
-	const BIGNUM *x, const BIGNUM *y, const BIGNUM *z, BN_CTX *);
-int EC_POINT_get_Jprojective_coordinates_GFp(const EC_GROUP *, const EC_POINT *,
-	BIGNUM *x, BIGNUM *y, BIGNUM *z, BN_CTX *);
-int EC_POINT_set_affine_coordinates_GFp(const EC_GROUP *, EC_POINT *,
-	const BIGNUM *x, const BIGNUM *y, BN_CTX *);
-int EC_POINT_get_affine_coordinates_GFp(const EC_GROUP *, const EC_POINT *,
-	BIGNUM *x, BIGNUM *y, BN_CTX *);
-int EC_POINT_set_compressed_coordinates_GFp(const EC_GROUP *, EC_POINT *,
-	const BIGNUM *x, int y_bit, BN_CTX *);
-
-int EC_POINT_set_affine_coordinates_GF2m(const EC_GROUP *, EC_POINT *,
-	const BIGNUM *x, const BIGNUM *y, BN_CTX *);
-int EC_POINT_get_affine_coordinates_GF2m(const EC_GROUP *, const EC_POINT *,
-	BIGNUM *x, BIGNUM *y, BN_CTX *);
-int EC_POINT_set_compressed_coordinates_GF2m(const EC_GROUP *, EC_POINT *,
-	const BIGNUM *x, int y_bit, BN_CTX *);
-
-size_t EC_POINT_point2oct(const EC_GROUP *, const EC_POINT *, point_conversion_form_t form,
-        unsigned char *buf, size_t len, BN_CTX *);
-int EC_POINT_oct2point(const EC_GROUP *, EC_POINT *,
-        const unsigned char *buf, size_t len, BN_CTX *);
+/** Returns the EC_METHOD used in EC_POINT object 
+ *  \param  point  EC_POINT object
+ *  \return the EC_METHOD used
+ */
+const EC_METHOD *EC_POINT_method_of(const EC_POINT *point);
+
+/** Sets a point to infinity (neutral element)
+ *  \param  group  underlying EC_GROUP object
+ *  \param  point  EC_POINT to set to infinity
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_set_to_infinity(const EC_GROUP *group, EC_POINT *point);
+
+/** Sets the jacobian projective coordinates of a EC_POINT over GFp
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  x      BIGNUM with the x-coordinate
+ *  \param  y      BIGNUM with the y-coordinate
+ *  \param  z      BIGNUM with the z-coordinate
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_set_Jprojective_coordinates_GFp(const EC_GROUP *group, EC_POINT *p,
+	const BIGNUM *x, const BIGNUM *y, const BIGNUM *z, BN_CTX *ctx);
+
+/** Gets the jacobian projective coordinates of a EC_POINT over GFp
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  x      BIGNUM for the x-coordinate
+ *  \param  y      BIGNUM for the y-coordinate
+ *  \param  z      BIGNUM for the z-coordinate
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_get_Jprojective_coordinates_GFp(const EC_GROUP *group,
+	const EC_POINT *p, BIGNUM *x, BIGNUM *y, BIGNUM *z, BN_CTX *ctx);
+
+/** Sets the affine coordinates of a EC_POINT over GFp
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  x      BIGNUM with the x-coordinate
+ *  \param  y      BIGNUM with the y-coordinate
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_set_affine_coordinates_GFp(const EC_GROUP *group, EC_POINT *p,
+	const BIGNUM *x, const BIGNUM *y, BN_CTX *ctx);
+
+/** Gets the affine coordinates of a EC_POINT over GFp
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  x      BIGNUM for the x-coordinate
+ *  \param  y      BIGNUM for the y-coordinate
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_get_affine_coordinates_GFp(const EC_GROUP *group,
+	const EC_POINT *p, BIGNUM *x, BIGNUM *y, BN_CTX *ctx);
+
+/** Sets the x9.62 compressed coordinates of a EC_POINT over GFp
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  x      BIGNUM with x-coordinate
+ *  \param  y_bit  integer with the y-Bit (either 0 or 1)
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_set_compressed_coordinates_GFp(const EC_GROUP *group, EC_POINT *p,
+	const BIGNUM *x, int y_bit, BN_CTX *ctx);
+#ifndef OPENSSL_NO_EC2M
+/** Sets the affine coordinates of a EC_POINT over GF2m
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  x      BIGNUM with the x-coordinate
+ *  \param  y      BIGNUM with the y-coordinate
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_set_affine_coordinates_GF2m(const EC_GROUP *group, EC_POINT *p,
+	const BIGNUM *x, const BIGNUM *y, BN_CTX *ctx);
+
+/** Gets the affine coordinates of a EC_POINT over GF2m
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  x      BIGNUM for the x-coordinate
+ *  \param  y      BIGNUM for the y-coordinate
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_get_affine_coordinates_GF2m(const EC_GROUP *group,
+	const EC_POINT *p, BIGNUM *x, BIGNUM *y, BN_CTX *ctx);
+
+/** Sets the x9.62 compressed coordinates of a EC_POINT over GF2m
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  x      BIGNUM with x-coordinate
+ *  \param  y_bit  integer with the y-Bit (either 0 or 1)
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_set_compressed_coordinates_GF2m(const EC_GROUP *group, EC_POINT *p,
+	const BIGNUM *x, int y_bit, BN_CTX *ctx);
+#endif
+/** Encodes a EC_POINT object to a octet string
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  form   point conversion form
+ *  \param  buf    memory buffer for the result. If NULL the function returns
+ *                 required buffer size.
+ *  \param  len    length of the memory buffer
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return the length of the encoded octet string or 0 if an error occurred
+ */
+size_t EC_POINT_point2oct(const EC_GROUP *group, const EC_POINT *p,
+	point_conversion_form_t form,
+        unsigned char *buf, size_t len, BN_CTX *ctx);
+
+/** Decodes a EC_POINT from a octet string
+ *  \param  group  underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \param  buf    memory buffer with the encoded ec point
+ *  \param  len    length of the encoded ec point
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_oct2point(const EC_GROUP *group, EC_POINT *p,
+        const unsigned char *buf, size_t len, BN_CTX *ctx);
 
 /* other interfaces to point2oct/oct2point: */
 BIGNUM *EC_POINT_point2bn(const EC_GROUP *, const EC_POINT *,
@@ -241,36 +570,114 @@ char *EC_POINT_point2hex(const EC_GROUP *, const EC_POINT *,
 EC_POINT *EC_POINT_hex2point(const EC_GROUP *, const char *,
 	EC_POINT *, BN_CTX *);
 
-int EC_POINT_add(const EC_GROUP *, EC_POINT *r, const EC_POINT *a, const EC_POINT *b, BN_CTX *);
-int EC_POINT_dbl(const EC_GROUP *, EC_POINT *r, const EC_POINT *a, BN_CTX *);
-int EC_POINT_invert(const EC_GROUP *, EC_POINT *, BN_CTX *);
 
-int EC_POINT_is_at_infinity(const EC_GROUP *, const EC_POINT *);
-int EC_POINT_is_on_curve(const EC_GROUP *, const EC_POINT *, BN_CTX *);
-int EC_POINT_cmp(const EC_GROUP *, const EC_POINT *a, const EC_POINT *b, BN_CTX *);
+/********************************************************************/
+/*         functions for doing EC_POINT arithmetic                  */
+/********************************************************************/
+
+/** Computes the sum of two EC_POINT 
+ *  \param  group  underlying EC_GROUP object
+ *  \param  r      EC_POINT object for the result (r = a + b)
+ *  \param  a      EC_POINT object with the first summand
+ *  \param  b      EC_POINT object with the second summand
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_add(const EC_GROUP *group, EC_POINT *r, const EC_POINT *a, const EC_POINT *b, BN_CTX *ctx);
+
+/** Computes the double of a EC_POINT
+ *  \param  group  underlying EC_GROUP object
+ *  \param  r      EC_POINT object for the result (r = 2 * a)
+ *  \param  a      EC_POINT object 
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_dbl(const EC_GROUP *group, EC_POINT *r, const EC_POINT *a, BN_CTX *ctx);
+
+/** Computes the inverse of a EC_POINT
+ *  \param  group  underlying EC_GROUP object
+ *  \param  a      EC_POINT object to be inverted (it's used for the result as well)
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_invert(const EC_GROUP *group, EC_POINT *a, BN_CTX *ctx);
+
+/** Checks whether the point is the neutral element of the group
+ *  \param  group  the underlying EC_GROUP object
+ *  \param  p      EC_POINT object
+ *  \return 1 if the point is the neutral element and 0 otherwise
+ */
+int EC_POINT_is_at_infinity(const EC_GROUP *group, const EC_POINT *p);
+
+/** Checks whether the point is on the curve 
+ *  \param  group  underlying EC_GROUP object
+ *  \param  point  EC_POINT object to check
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 if point if on the curve and 0 otherwise
+ */
+int EC_POINT_is_on_curve(const EC_GROUP *group, const EC_POINT *point, BN_CTX *ctx);
+
+/** Compares two EC_POINTs 
+ *  \param  group  underlying EC_GROUP object
+ *  \param  a      first EC_POINT object
+ *  \param  b      second EC_POINT object
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 0 if both points are equal and a value != 0 otherwise
+ */
+int EC_POINT_cmp(const EC_GROUP *group, const EC_POINT *a, const EC_POINT *b, BN_CTX *ctx);
 
 int EC_POINT_make_affine(const EC_GROUP *, EC_POINT *, BN_CTX *);
 int EC_POINTs_make_affine(const EC_GROUP *, size_t num, EC_POINT *[], BN_CTX *);
 
+/** Computes r = generator * n sum_{i=0}^num p[i] * m[i]
+ *  \param  group  underlying EC_GROUP object
+ *  \param  r      EC_POINT object for the result
+ *  \param  n      BIGNUM with the multiplier for the group generator (optional)
+ *  \param  num    number futher summands
+ *  \param  p      array of size num of EC_POINT objects
+ *  \param  m      array of size num of BIGNUM objects
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINTs_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *n, size_t num, const EC_POINT *p[], const BIGNUM *m[], BN_CTX *ctx);
+
+/** Computes r = generator * n + q * m
+ *  \param  group  underlying EC_GROUP object
+ *  \param  r      EC_POINT object for the result
+ *  \param  n      BIGNUM with the multiplier for the group generator (optional)
+ *  \param  q      EC_POINT object with the first factor of the second summand
+ *  \param  m      BIGNUM with the second factor of the second summand
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_POINT_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *n, const EC_POINT *q, const BIGNUM *m, BN_CTX *ctx);
 
-int EC_POINTs_mul(const EC_GROUP *, EC_POINT *r, const BIGNUM *, size_t num, const EC_POINT *[], const BIGNUM *[], BN_CTX *);
-int EC_POINT_mul(const EC_GROUP *, EC_POINT *r, const BIGNUM *, const EC_POINT *, const BIGNUM *, BN_CTX *);
-
-/* EC_GROUP_precompute_mult() stores multiples of generator for faster point multiplication */
-int EC_GROUP_precompute_mult(EC_GROUP *, BN_CTX *);
-/* EC_GROUP_have_precompute_mult() reports whether such precomputation has been done */
-int EC_GROUP_have_precompute_mult(const EC_GROUP *);
+/** Stores multiples of generator for faster point multiplication
+ *  \param  group  EC_GROUP object
+ *  \param  ctx    BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occured
+ */
+int EC_GROUP_precompute_mult(EC_GROUP *group, BN_CTX *ctx);
 
+/** Reports whether a precomputation has been done
+ *  \param  group  EC_GROUP object
+ *  \return 1 if a pre-computation has been done and 0 otherwise
+ */
+int EC_GROUP_have_precompute_mult(const EC_GROUP *group);
 
 
-/* ASN1 stuff */
+/********************************************************************/
+/*                       ASN1 stuff                                 */
+/********************************************************************/
 
 /* EC_GROUP_get_basis_type() returns the NID of the basis type
  * used to represent the field elements */
 int EC_GROUP_get_basis_type(const EC_GROUP *);
+#ifndef OPENSSL_NO_EC2M
 int EC_GROUP_get_trinomial_basis(const EC_GROUP *, unsigned int *k);
 int EC_GROUP_get_pentanomial_basis(const EC_GROUP *, unsigned int *k1, 
 	unsigned int *k2, unsigned int *k3);
+#endif
 
 #define OPENSSL_EC_NAMED_CURVE	0x001
 
@@ -293,28 +700,106 @@ int     ECPKParameters_print(BIO *bp, const EC_GROUP *x, int off);
 int     ECPKParameters_print_fp(FILE *fp, const EC_GROUP *x, int off);
 #endif
 
-/* the EC_KEY stuff */
+
+/********************************************************************/
+/*                      EC_KEY functions                            */
+/********************************************************************/
+
 typedef struct ec_key_st EC_KEY;
 
 /* some values for the encoding_flag */
 #define EC_PKEY_NO_PARAMETERS	0x001
 #define EC_PKEY_NO_PUBKEY	0x002
 
+/* some values for the flags field */
+#define EC_FLAG_NON_FIPS_ALLOW	0x1
+#define EC_FLAG_FIPS_CHECKED	0x2
+
+/** Creates a new EC_KEY object.
+ *  \return EC_KEY object or NULL if an error occurred.
+ */
 EC_KEY *EC_KEY_new(void);
+
+int EC_KEY_get_flags(const EC_KEY *key);
+
+void EC_KEY_set_flags(EC_KEY *key, int flags);
+
+void EC_KEY_clear_flags(EC_KEY *key, int flags);
+
+/** Creates a new EC_KEY object using a named curve as underlying
+ *  EC_GROUP object.
+ *  \param  nid  NID of the named curve.
+ *  \return EC_KEY object or NULL if an error occurred. 
+ */
 EC_KEY *EC_KEY_new_by_curve_name(int nid);
-void EC_KEY_free(EC_KEY *);
-EC_KEY *EC_KEY_copy(EC_KEY *, const EC_KEY *);
-EC_KEY *EC_KEY_dup(const EC_KEY *);
-
-int EC_KEY_up_ref(EC_KEY *);
-
-const EC_GROUP *EC_KEY_get0_group(const EC_KEY *);
-int EC_KEY_set_group(EC_KEY *, const EC_GROUP *);
-const BIGNUM *EC_KEY_get0_private_key(const EC_KEY *);
-int EC_KEY_set_private_key(EC_KEY *, const BIGNUM *);
-const EC_POINT *EC_KEY_get0_public_key(const EC_KEY *);
-int EC_KEY_set_public_key(EC_KEY *, const EC_POINT *);
-unsigned EC_KEY_get_enc_flags(const EC_KEY *);
+
+/** Frees a EC_KEY object.
+ *  \param  key  EC_KEY object to be freed.
+ */
+void EC_KEY_free(EC_KEY *key);
+
+/** Copies a EC_KEY object.
+ *  \param  dst  destination EC_KEY object
+ *  \param  src  src EC_KEY object
+ *  \return dst or NULL if an error occurred.
+ */
+EC_KEY *EC_KEY_copy(EC_KEY *dst, const EC_KEY *src);
+
+/** Creates a new EC_KEY object and copies the content from src to it.
+ *  \param  src  the source EC_KEY object
+ *  \return newly created EC_KEY object or NULL if an error occurred.
+ */
+EC_KEY *EC_KEY_dup(const EC_KEY *src);
+
+/** Increases the internal reference count of a EC_KEY object.
+ *  \param  key  EC_KEY object
+ *  \return 1 on success and 0 if an error occurred.
+ */
+int EC_KEY_up_ref(EC_KEY *key);
+
+/** Returns the EC_GROUP object of a EC_KEY object
+ *  \param  key  EC_KEY object
+ *  \return the EC_GROUP object (possibly NULL).
+ */
+const EC_GROUP *EC_KEY_get0_group(const EC_KEY *key);
+
+/** Sets the EC_GROUP of a EC_KEY object.
+ *  \param  key    EC_KEY object
+ *  \param  group  EC_GROUP to use in the EC_KEY object (note: the EC_KEY
+ *                 object will use an own copy of the EC_GROUP).
+ *  \return 1 on success and 0 if an error occurred.
+ */
+int EC_KEY_set_group(EC_KEY *key, const EC_GROUP *group);
+
+/** Returns the private key of a EC_KEY object.
+ *  \param  key  EC_KEY object
+ *  \return a BIGNUM with the private key (possibly NULL).
+ */
+const BIGNUM *EC_KEY_get0_private_key(const EC_KEY *key);
+
+/** Sets the private key of a EC_KEY object.
+ *  \param  key  EC_KEY object
+ *  \param  prv  BIGNUM with the private key (note: the EC_KEY object
+ *               will use an own copy of the BIGNUM).
+ *  \return 1 on success and 0 if an error occurred.
+ */
+int EC_KEY_set_private_key(EC_KEY *key, const BIGNUM *prv);
+
+/** Returns the public key of a EC_KEY object.
+ *  \param  key  the EC_KEY object
+ *  \return a EC_POINT object with the public key (possibly NULL)
+ */
+const EC_POINT *EC_KEY_get0_public_key(const EC_KEY *key);
+
+/** Sets the public key of a EC_KEY object.
+ *  \param  key  EC_KEY object
+ *  \param  pub  EC_POINT object with the public key (note: the EC_KEY object
+ *               will use an own copy of the EC_POINT object).
+ *  \return 1 on success and 0 if an error occurred.
+ */
+int EC_KEY_set_public_key(EC_KEY *key, const EC_POINT *pub);
+
+unsigned EC_KEY_get_enc_flags(const EC_KEY *key);
 void EC_KEY_set_enc_flags(EC_KEY *, unsigned int);
 point_conversion_form_t EC_KEY_get_conv_form(const EC_KEY *);
 void EC_KEY_set_conv_form(EC_KEY *, point_conversion_form_t);
@@ -325,31 +810,135 @@ void EC_KEY_insert_key_method_data(EC_KEY *, void *data,
 	void *(*dup_func)(void *), void (*free_func)(void *), void (*clear_free_func)(void *));
 /* wrapper functions for the underlying EC_GROUP object */
 void EC_KEY_set_asn1_flag(EC_KEY *, int);
-int EC_KEY_precompute_mult(EC_KEY *, BN_CTX *ctx);
-
-/* EC_KEY_generate_key() creates a ec private (public) key */
-int EC_KEY_generate_key(EC_KEY *);
-/* EC_KEY_check_key() */
-int EC_KEY_check_key(const EC_KEY *);
-
-/* de- and encoding functions for SEC1 ECPrivateKey */
-EC_KEY *d2i_ECPrivateKey(EC_KEY **a, const unsigned char **in, long len);
-int i2d_ECPrivateKey(EC_KEY *a, unsigned char **out);
-/* de- and encoding functions for EC parameters */
-EC_KEY *d2i_ECParameters(EC_KEY **a, const unsigned char **in, long len);
-int i2d_ECParameters(EC_KEY *a, unsigned char **out);
-/* de- and encoding functions for EC public key
- * (octet string, not DER -- hence 'o2i' and 'i2o') */
-EC_KEY *o2i_ECPublicKey(EC_KEY **a, const unsigned char **in, long len);
-int i2o_ECPublicKey(EC_KEY *a, unsigned char **out);
+
+/** Creates a table of pre-computed multiples of the generator to 
+ *  accelerate further EC_KEY operations.
+ *  \param  key  EC_KEY object
+ *  \param  ctx  BN_CTX object (optional)
+ *  \return 1 on success and 0 if an error occurred.
+ */
+int EC_KEY_precompute_mult(EC_KEY *key, BN_CTX *ctx);
+
+/** Creates a new ec private (and optional a new public) key.
+ *  \param  key  EC_KEY object
+ *  \return 1 on success and 0 if an error occurred.
+ */
+int EC_KEY_generate_key(EC_KEY *key);
+
+/** Verifies that a private and/or public key is valid.
+ *  \param  key  the EC_KEY object
+ *  \return 1 on success and 0 otherwise.
+ */
+int EC_KEY_check_key(const EC_KEY *key);
+
+/** Sets a public key from affine coordindates performing
+ *  neccessary NIST PKV tests.
+ *  \param  key  the EC_KEY object
+ *  \param  x    public key x coordinate
+ *  \param  y    public key y coordinate
+ *  \return 1 on success and 0 otherwise.
+ */
+int EC_KEY_set_public_key_affine_coordinates(EC_KEY *key, BIGNUM *x, BIGNUM *y);
+
+
+/********************************************************************/
+/*        de- and encoding functions for SEC1 ECPrivateKey          */
+/********************************************************************/
+
+/** Decodes a private key from a memory buffer.
+ *  \param  key  a pointer to a EC_KEY object which should be used (or NULL)
+ *  \param  in   pointer to memory with the DER encoded private key
+ *  \param  len  length of the DER encoded private key
+ *  \return the decoded private key or NULL if an error occurred.
+ */
+EC_KEY *d2i_ECPrivateKey(EC_KEY **key, const unsigned char **in, long len);
+
+/** Encodes a private key object and stores the result in a buffer.
+ *  \param  key  the EC_KEY object to encode
+ *  \param  out  the buffer for the result (if NULL the function returns number
+ *               of bytes needed).
+ *  \return 1 on success and 0 if an error occurred.
+ */
+int i2d_ECPrivateKey(EC_KEY *key, unsigned char **out);
+
+
+/********************************************************************/
+/*        de- and encoding functions for EC parameters              */
+/********************************************************************/
+
+/** Decodes ec parameter from a memory buffer.
+ *  \param  key  a pointer to a EC_KEY object which should be used (or NULL)
+ *  \param  in   pointer to memory with the DER encoded ec parameters
+ *  \param  len  length of the DER encoded ec parameters
+ *  \return a EC_KEY object with the decoded parameters or NULL if an error
+ *          occurred.
+ */
+EC_KEY *d2i_ECParameters(EC_KEY **key, const unsigned char **in, long len);
+
+/** Encodes ec parameter and stores the result in a buffer.
+ *  \param  key  the EC_KEY object with ec paramters to encode
+ *  \param  out  the buffer for the result (if NULL the function returns number
+ *               of bytes needed).
+ *  \return 1 on success and 0 if an error occurred.
+ */
+int i2d_ECParameters(EC_KEY *key, unsigned char **out);
+
+
+/********************************************************************/
+/*         de- and encoding functions for EC public key             */
+/*         (octet string, not DER -- hence 'o2i' and 'i2o')         */
+/********************************************************************/
+
+/** Decodes a ec public key from a octet string.
+ *  \param  key  a pointer to a EC_KEY object which should be used
+ *  \param  in   memory buffer with the encoded public key
+ *  \param  len  length of the encoded public key
+ *  \return EC_KEY object with decoded public key or NULL if an error
+ *          occurred.
+ */
+EC_KEY *o2i_ECPublicKey(EC_KEY **key, const unsigned char **in, long len);
+
+/** Encodes a ec public key in an octet string.
+ *  \param  key  the EC_KEY object with the public key
+ *  \param  out  the buffer for the result (if NULL the function returns number
+ *               of bytes needed).
+ *  \return 1 on success and 0 if an error occurred
+ */
+int i2o_ECPublicKey(EC_KEY *key, unsigned char **out);
 
 #ifndef OPENSSL_NO_BIO
-int	ECParameters_print(BIO *bp, const EC_KEY *x);
-int	EC_KEY_print(BIO *bp, const EC_KEY *x, int off);
+/** Prints out the ec parameters on human readable form.
+ *  \param  bp   BIO object to which the information is printed
+ *  \param  key  EC_KEY object
+ *  \return 1 on success and 0 if an error occurred
+ */
+int	ECParameters_print(BIO *bp, const EC_KEY *key);
+
+/** Prints out the contents of a EC_KEY object
+ *  \param  bp   BIO object to which the information is printed
+ *  \param  key  EC_KEY object
+ *  \param  off  line offset 
+ *  \return 1 on success and 0 if an error occurred
+ */
+int	EC_KEY_print(BIO *bp, const EC_KEY *key, int off);
+
 #endif
 #ifndef OPENSSL_NO_FP_API
-int	ECParameters_print_fp(FILE *fp, const EC_KEY *x);
-int	EC_KEY_print_fp(FILE *fp, const EC_KEY *x, int off);
+/** Prints out the ec parameters on human readable form.
+ *  \param  fp   file descriptor to which the information is printed
+ *  \param  key  EC_KEY object
+ *  \return 1 on success and 0 if an error occurred
+ */
+int	ECParameters_print_fp(FILE *fp, const EC_KEY *key);
+
+/** Prints out the contents of a EC_KEY object
+ *  \param  fp   file descriptor to which the information is printed
+ *  \param  key  EC_KEY object
+ *  \param  off  line offset 
+ *  \return 1 on success and 0 if an error occurred
+ */
+int	EC_KEY_print_fp(FILE *fp, const EC_KEY *key, int off);
+
 #endif
 
 #define ECParameters_dup(x) ASN1_dup_of(EC_KEY,i2d_ECParameters,d2i_ECParameters,x)
@@ -362,6 +951,13 @@ int	EC_KEY_print_fp(FILE *fp, const EC_KEY *x, int off);
 # endif
 #endif
 
+#define EVP_PKEY_CTX_set_ec_paramgen_curve_nid(ctx, nid) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_EC, EVP_PKEY_OP_PARAMGEN, \
+				EVP_PKEY_CTRL_EC_PARAMGEN_CURVE_NID, nid, NULL)
+
+
+#define EVP_PKEY_CTRL_EC_PARAMGEN_CURVE_NID		(EVP_PKEY_ALG_CTRL + 1)
+
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
  * made after this point may be overwritten when the script is next run.
@@ -371,10 +967,19 @@ void ERR_load_EC_strings(void);
 /* Error codes for the EC functions. */
 
 /* Function codes. */
+#define EC_F_BN_TO_FELEM				 224
 #define EC_F_COMPUTE_WNAF				 143
 #define EC_F_D2I_ECPARAMETERS				 144
 #define EC_F_D2I_ECPKPARAMETERS				 145
 #define EC_F_D2I_ECPRIVATEKEY				 146
+#define EC_F_DO_EC_KEY_PRINT				 221
+#define EC_F_ECKEY_PARAM2TYPE				 223
+#define EC_F_ECKEY_PARAM_DECODE				 212
+#define EC_F_ECKEY_PRIV_DECODE				 213
+#define EC_F_ECKEY_PRIV_ENCODE				 214
+#define EC_F_ECKEY_PUB_DECODE				 215
+#define EC_F_ECKEY_PUB_ENCODE				 216
+#define EC_F_ECKEY_TYPE2PARAM				 220
 #define EC_F_ECPARAMETERS_PRINT				 147
 #define EC_F_ECPARAMETERS_PRINT_FP			 148
 #define EC_F_ECPKPARAMETERS_PRINT			 149
@@ -405,6 +1010,15 @@ void ERR_load_EC_strings(void);
 #define EC_F_EC_GFP_MONT_FIELD_SQR			 132
 #define EC_F_EC_GFP_MONT_GROUP_SET_CURVE		 189
 #define EC_F_EC_GFP_MONT_GROUP_SET_CURVE_GFP		 135
+#define EC_F_EC_GFP_NISTP224_GROUP_SET_CURVE		 225
+#define EC_F_EC_GFP_NISTP224_POINTS_MUL			 228
+#define EC_F_EC_GFP_NISTP224_POINT_GET_AFFINE_COORDINATES 226
+#define EC_F_EC_GFP_NISTP256_GROUP_SET_CURVE		 230
+#define EC_F_EC_GFP_NISTP256_POINTS_MUL			 231
+#define EC_F_EC_GFP_NISTP256_POINT_GET_AFFINE_COORDINATES 232
+#define EC_F_EC_GFP_NISTP521_GROUP_SET_CURVE		 233
+#define EC_F_EC_GFP_NISTP521_POINTS_MUL			 234
+#define EC_F_EC_GFP_NISTP521_POINT_GET_AFFINE_COORDINATES 235
 #define EC_F_EC_GFP_NIST_FIELD_MUL			 200
 #define EC_F_EC_GFP_NIST_FIELD_SQR			 201
 #define EC_F_EC_GFP_NIST_GROUP_SET_CURVE		 202
@@ -447,8 +1061,8 @@ void ERR_load_EC_strings(void);
 #define EC_F_EC_KEY_NEW					 182
 #define EC_F_EC_KEY_PRINT				 180
 #define EC_F_EC_KEY_PRINT_FP				 181
+#define EC_F_EC_KEY_SET_PUBLIC_KEY_AFFINE_COORDINATES	 229
 #define EC_F_EC_POINTS_MAKE_AFFINE			 136
-#define EC_F_EC_POINTS_MUL				 138
 #define EC_F_EC_POINT_ADD				 112
 #define EC_F_EC_POINT_CMP				 113
 #define EC_F_EC_POINT_COPY				 114
@@ -478,22 +1092,38 @@ void ERR_load_EC_strings(void);
 #define EC_F_I2D_ECPKPARAMETERS				 191
 #define EC_F_I2D_ECPRIVATEKEY				 192
 #define EC_F_I2O_ECPUBLICKEY				 151
+#define EC_F_NISTP224_PRE_COMP_NEW			 227
+#define EC_F_NISTP256_PRE_COMP_NEW			 236
+#define EC_F_NISTP521_PRE_COMP_NEW			 237
 #define EC_F_O2I_ECPUBLICKEY				 152
+#define EC_F_OLD_EC_PRIV_DECODE				 222
+#define EC_F_PKEY_EC_CTRL				 197
+#define EC_F_PKEY_EC_CTRL_STR				 198
+#define EC_F_PKEY_EC_DERIVE				 217
+#define EC_F_PKEY_EC_KEYGEN				 199
+#define EC_F_PKEY_EC_PARAMGEN				 219
+#define EC_F_PKEY_EC_SIGN				 218
 
 /* Reason codes. */
 #define EC_R_ASN1_ERROR					 115
 #define EC_R_ASN1_UNKNOWN_FIELD				 116
+#define EC_R_BIGNUM_OUT_OF_RANGE			 144
 #define EC_R_BUFFER_TOO_SMALL				 100
+#define EC_R_COORDINATES_OUT_OF_RANGE			 146
 #define EC_R_D2I_ECPKPARAMETERS_FAILURE			 117
+#define EC_R_DECODE_ERROR				 142
 #define EC_R_DISCRIMINANT_IS_ZERO			 118
 #define EC_R_EC_GROUP_NEW_BY_NAME_FAILURE		 119
-#define EC_R_FIELD_TOO_LARGE				 138
+#define EC_R_FIELD_TOO_LARGE				 143
+#define EC_R_GF2M_NOT_SUPPORTED				 147
 #define EC_R_GROUP2PKPARAMETERS_FAILURE			 120
 #define EC_R_I2D_ECPKPARAMETERS_FAILURE			 121
 #define EC_R_INCOMPATIBLE_OBJECTS			 101
 #define EC_R_INVALID_ARGUMENT				 112
 #define EC_R_INVALID_COMPRESSED_POINT			 110
 #define EC_R_INVALID_COMPRESSION_BIT			 109
+#define EC_R_INVALID_CURVE				 141
+#define EC_R_INVALID_DIGEST_TYPE			 138
 #define EC_R_INVALID_ENCODING				 102
 #define EC_R_INVALID_FIELD				 103
 #define EC_R_INVALID_FORM				 104
@@ -501,6 +1131,7 @@ void ERR_load_EC_strings(void);
 #define EC_R_INVALID_PENTANOMIAL_BASIS			 132
 #define EC_R_INVALID_PRIVATE_KEY			 123
 #define EC_R_INVALID_TRINOMIAL_BASIS			 137
+#define EC_R_KEYS_NOT_SET				 140
 #define EC_R_MISSING_PARAMETERS				 124
 #define EC_R_MISSING_PRIVATE_KEY			 125
 #define EC_R_NOT_A_NIST_PRIME				 135
@@ -508,6 +1139,7 @@ void ERR_load_EC_strings(void);
 #define EC_R_NOT_IMPLEMENTED				 126
 #define EC_R_NOT_INITIALIZED				 111
 #define EC_R_NO_FIELD_MOD				 133
+#define EC_R_NO_PARAMETERS_SET				 139
 #define EC_R_PASSED_NULL_PARAMETER			 134
 #define EC_R_PKPARAMETERS2GROUP_FAILURE			 127
 #define EC_R_POINT_AT_INFINITY				 106
@@ -518,6 +1150,7 @@ void ERR_load_EC_strings(void);
 #define EC_R_UNKNOWN_GROUP				 129
 #define EC_R_UNKNOWN_ORDER				 114
 #define EC_R_UNSUPPORTED_FIELD				 131
+#define EC_R_WRONG_CURVE_PARAMETERS			 145
 #define EC_R_WRONG_ORDER				 130
 
 #ifdef  __cplusplus
diff --git a/crypto/ec/ec2_mult.c b/crypto/ec/ec2_mult.c
index 7dca5e4bcd43..26f4a783fcc1 100644
--- a/crypto/ec/ec2_mult.c
+++ b/crypto/ec/ec2_mult.c
@@ -71,12 +71,14 @@
 
 #include "ec_lcl.h"
 
+#ifndef OPENSSL_NO_EC2M
+
 
 /* Compute the x-coordinate x/z for the point 2*(x/z) in Montgomery projective 
  * coordinates.
  * Uses algorithm Mdouble in appendix of 
  *     Lopez, J. and Dahab, R.  "Fast multiplication on elliptic curves over 
- *     GF(2^m) without precomputation".
+ *     GF(2^m) without precomputation" (CHES '99, LNCS 1717).
  * modified to not require precomputation of c=b^{2^{m-1}}.
  */
 static int gf2m_Mdouble(const EC_GROUP *group, BIGNUM *x, BIGNUM *z, BN_CTX *ctx)
@@ -107,8 +109,8 @@ static int gf2m_Mdouble(const EC_GROUP *group, BIGNUM *x, BIGNUM *z, BN_CTX *ctx
 /* Compute the x-coordinate x1/z1 for the point (x1/z1)+(x2/x2) in Montgomery 
  * projective coordinates.
  * Uses algorithm Madd in appendix of 
- *     Lopex, J. and Dahab, R.  "Fast multiplication on elliptic curves over 
- *     GF(2^m) without precomputation".
+ *     Lopez, J. and Dahab, R.  "Fast multiplication on elliptic curves over 
+ *     GF(2^m) without precomputation" (CHES '99, LNCS 1717).
  */
 static int gf2m_Madd(const EC_GROUP *group, const BIGNUM *x, BIGNUM *x1, BIGNUM *z1, 
 	const BIGNUM *x2, const BIGNUM *z2, BN_CTX *ctx)
@@ -140,8 +142,8 @@ static int gf2m_Madd(const EC_GROUP *group, const BIGNUM *x, BIGNUM *x1, BIGNUM
 
 /* Compute the x, y affine coordinates from the point (x1, z1) (x2, z2) 
  * using Montgomery point multiplication algorithm Mxy() in appendix of 
- *     Lopex, J. and Dahab, R.  "Fast multiplication on elliptic curves over 
- *     GF(2^m) without precomputation".
+ *     Lopez, J. and Dahab, R.  "Fast multiplication on elliptic curves over 
+ *     GF(2^m) without precomputation" (CHES '99, LNCS 1717).
  * Returns:
  *     0 on error
  *     1 if return value should be the point at infinity
@@ -209,15 +211,15 @@ static int gf2m_Mxy(const EC_GROUP *group, const BIGNUM *x, const BIGNUM *y, BIG
 /* Computes scalar*point and stores the result in r.
  * point can not equal r.
  * Uses algorithm 2P of
- *     Lopex, J. and Dahab, R.  "Fast multiplication on elliptic curves over 
- *     GF(2^m) without precomputation".
+ *     Lopez, J. and Dahab, R.  "Fast multiplication on elliptic curves over 
+ *     GF(2^m) without precomputation" (CHES '99, LNCS 1717).
  */
 static int ec_GF2m_montgomery_point_multiply(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar,
 	const EC_POINT *point, BN_CTX *ctx)
 	{
 	BIGNUM *x1, *x2, *z1, *z2;
-	int ret = 0, i, j;
-	BN_ULONG mask;
+	int ret = 0, i;
+	BN_ULONG mask,word;
 
 	if (r == point)
 		{
@@ -251,22 +253,24 @@ static int ec_GF2m_montgomery_point_multiply(const EC_GROUP *group, EC_POINT *r,
 	if (!BN_GF2m_add(x2, x2, &group->b)) goto err; /* x2 = x^4 + b */
 
 	/* find top most bit and go one past it */
-	i = scalar->top - 1; j = BN_BITS2 - 1;
+	i = scalar->top - 1;
 	mask = BN_TBIT;
-	while (!(scalar->d[i] & mask)) { mask >>= 1; j--; }
-	mask >>= 1; j--;
+	word = scalar->d[i];
+	while (!(word & mask)) mask >>= 1;
+	mask >>= 1;
 	/* if top most bit was at word break, go to next word */
 	if (!mask) 
 		{
-		i--; j = BN_BITS2 - 1;
+		i--;
 		mask = BN_TBIT;
 		}
 
 	for (; i >= 0; i--)
 		{
-		for (; j >= 0; j--)
+		word = scalar->d[i];
+		while (mask)
 			{
-			if (scalar->d[i] & mask)
+			if (word & mask)
 				{
 				if (!gf2m_Madd(group, &point->X, x1, z1, x2, z2, ctx)) goto err;
 				if (!gf2m_Mdouble(group, x2, z2, ctx)) goto err;
@@ -278,7 +282,6 @@ static int ec_GF2m_montgomery_point_multiply(const EC_GROUP *group, EC_POINT *r,
 				}
 			mask >>= 1;
 			}
-		j = BN_BITS2 - 1;
 		mask = BN_TBIT;
 		}
 
@@ -383,3 +386,5 @@ int ec_GF2m_have_precompute_mult(const EC_GROUP *group)
 	{
 	return ec_wNAF_have_precompute_mult(group);
  	}
+
+#endif
diff --git a/crypto/ec/ec2_oct.c b/crypto/ec/ec2_oct.c
new file mode 100644
index 000000000000..f1d75e5ddf66
--- /dev/null
+++ b/crypto/ec/ec2_oct.c
@@ -0,0 +1,407 @@
+/* crypto/ec/ec2_oct.c */
+/* ====================================================================
+ * Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
+ *
+ * The Elliptic Curve Public-Key Crypto Library (ECC Code) included
+ * herein is developed by SUN MICROSYSTEMS, INC., and is contributed
+ * to the OpenSSL project.
+ *
+ * The ECC Code is licensed pursuant to the OpenSSL open source
+ * license provided below.
+ *
+ * The software is originally written by Sheueling Chang Shantz and
+ * Douglas Stebila of Sun Microsystems Laboratories.
+ *
+ */
+/* ====================================================================
+ * Copyright (c) 1998-2005 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <openssl/err.h>
+
+#include "ec_lcl.h"
+
+#ifndef OPENSSL_NO_EC2M
+
+/* Calculates and sets the affine coordinates of an EC_POINT from the given
+ * compressed coordinates.  Uses algorithm 2.3.4 of SEC 1. 
+ * Note that the simple implementation only uses affine coordinates.
+ *
+ * The method is from the following publication:
+ * 
+ *     Harper, Menezes, Vanstone:
+ *     "Public-Key Cryptosystems with Very Small Key Lengths",
+ *     EUROCRYPT '92, Springer-Verlag LNCS 658,
+ *     published February 1993
+ *
+ * US Patents 6,141,420 and 6,618,483 (Vanstone, Mullin, Agnew) describe
+ * the same method, but claim no priority date earlier than July 29, 1994
+ * (and additionally fail to cite the EUROCRYPT '92 publication as prior art).
+ */
+int ec_GF2m_simple_set_compressed_coordinates(const EC_GROUP *group, EC_POINT *point,
+	const BIGNUM *x_, int y_bit, BN_CTX *ctx)
+	{
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *tmp, *x, *y, *z;
+	int ret = 0, z0;
+
+	/* clear error queue */
+	ERR_clear_error();
+
+	if (ctx == NULL)
+		{
+		ctx = new_ctx = BN_CTX_new();
+		if (ctx == NULL)
+			return 0;
+		}
+
+	y_bit = (y_bit != 0) ? 1 : 0;
+
+	BN_CTX_start(ctx);
+	tmp = BN_CTX_get(ctx);
+	x = BN_CTX_get(ctx);
+	y = BN_CTX_get(ctx);
+	z = BN_CTX_get(ctx);
+	if (z == NULL) goto err;
+
+	if (!BN_GF2m_mod_arr(x, x_, group->poly)) goto err;
+	if (BN_is_zero(x))
+		{
+		if (!BN_GF2m_mod_sqrt_arr(y, &group->b, group->poly, ctx)) goto err;
+		}
+	else
+		{
+		if (!group->meth->field_sqr(group, tmp, x, ctx)) goto err;
+		if (!group->meth->field_div(group, tmp, &group->b, tmp, ctx)) goto err;
+		if (!BN_GF2m_add(tmp, &group->a, tmp)) goto err;
+		if (!BN_GF2m_add(tmp, x, tmp)) goto err;
+		if (!BN_GF2m_mod_solve_quad_arr(z, tmp, group->poly, ctx))
+			{
+			unsigned long err = ERR_peek_last_error();
+			
+			if (ERR_GET_LIB(err) == ERR_LIB_BN && ERR_GET_REASON(err) == BN_R_NO_SOLUTION)
+				{
+				ERR_clear_error();
+				ECerr(EC_F_EC_GF2M_SIMPLE_SET_COMPRESSED_COORDINATES, EC_R_INVALID_COMPRESSED_POINT);
+				}
+			else
+				ECerr(EC_F_EC_GF2M_SIMPLE_SET_COMPRESSED_COORDINATES, ERR_R_BN_LIB);
+			goto err;
+			}
+		z0 = (BN_is_odd(z)) ? 1 : 0;
+		if (!group->meth->field_mul(group, y, x, z, ctx)) goto err;
+		if (z0 != y_bit)
+			{
+			if (!BN_GF2m_add(y, y, x)) goto err;
+			}
+		}
+
+	if (!EC_POINT_set_affine_coordinates_GF2m(group, point, x, y, ctx)) goto err;
+
+	ret = 1;
+
+ err:
+	BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return ret;
+	}
+
+
+/* Converts an EC_POINT to an octet string.  
+ * If buf is NULL, the encoded length will be returned.
+ * If the length len of buf is smaller than required an error will be returned.
+ */
+size_t ec_GF2m_simple_point2oct(const EC_GROUP *group, const EC_POINT *point, point_conversion_form_t form,
+	unsigned char *buf, size_t len, BN_CTX *ctx)
+	{
+	size_t ret;
+	BN_CTX *new_ctx = NULL;
+	int used_ctx = 0;
+	BIGNUM *x, *y, *yxi;
+	size_t field_len, i, skip;
+
+	if ((form != POINT_CONVERSION_COMPRESSED)
+		&& (form != POINT_CONVERSION_UNCOMPRESSED)
+		&& (form != POINT_CONVERSION_HYBRID))
+		{
+		ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, EC_R_INVALID_FORM);
+		goto err;
+		}
+
+	if (EC_POINT_is_at_infinity(group, point))
+		{
+		/* encodes to a single 0 octet */
+		if (buf != NULL)
+			{
+			if (len < 1)
+				{
+				ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, EC_R_BUFFER_TOO_SMALL);
+				return 0;
+				}
+			buf[0] = 0;
+			}
+		return 1;
+		}
+
+
+	/* ret := required output buffer length */
+	field_len = (EC_GROUP_get_degree(group) + 7) / 8;
+	ret = (form == POINT_CONVERSION_COMPRESSED) ? 1 + field_len : 1 + 2*field_len;
+
+	/* if 'buf' is NULL, just return required length */
+	if (buf != NULL)
+		{
+		if (len < ret)
+			{
+			ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, EC_R_BUFFER_TOO_SMALL);
+			goto err;
+			}
+
+		if (ctx == NULL)
+			{
+			ctx = new_ctx = BN_CTX_new();
+			if (ctx == NULL)
+				return 0;
+			}
+
+		BN_CTX_start(ctx);
+		used_ctx = 1;
+		x = BN_CTX_get(ctx);
+		y = BN_CTX_get(ctx);
+		yxi = BN_CTX_get(ctx);
+		if (yxi == NULL) goto err;
+
+		if (!EC_POINT_get_affine_coordinates_GF2m(group, point, x, y, ctx)) goto err;
+
+		buf[0] = form;
+		if ((form != POINT_CONVERSION_UNCOMPRESSED) && !BN_is_zero(x))
+			{
+			if (!group->meth->field_div(group, yxi, y, x, ctx)) goto err;
+			if (BN_is_odd(yxi)) buf[0]++;
+			}
+
+		i = 1;
+		
+		skip = field_len - BN_num_bytes(x);
+		if (skip > field_len)
+			{
+			ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
+			goto err;
+			}
+		while (skip > 0)
+			{
+			buf[i++] = 0;
+			skip--;
+			}
+		skip = BN_bn2bin(x, buf + i);
+		i += skip;
+		if (i != 1 + field_len)
+			{
+			ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
+			goto err;
+			}
+
+		if (form == POINT_CONVERSION_UNCOMPRESSED || form == POINT_CONVERSION_HYBRID)
+			{
+			skip = field_len - BN_num_bytes(y);
+			if (skip > field_len)
+				{
+				ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
+				goto err;
+				}
+			while (skip > 0)
+				{
+				buf[i++] = 0;
+				skip--;
+				}
+			skip = BN_bn2bin(y, buf + i);
+			i += skip;
+			}
+
+		if (i != ret)
+			{
+			ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
+			goto err;
+			}
+		}
+	
+	if (used_ctx)
+		BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return ret;
+
+ err:
+	if (used_ctx)
+		BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return 0;
+	}
+
+
+/* Converts an octet string representation to an EC_POINT. 
+ * Note that the simple implementation only uses affine coordinates.
+ */
+int ec_GF2m_simple_oct2point(const EC_GROUP *group, EC_POINT *point,
+	const unsigned char *buf, size_t len, BN_CTX *ctx)
+	{
+	point_conversion_form_t form;
+	int y_bit;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *x, *y, *yxi;
+	size_t field_len, enc_len;
+	int ret = 0;
+
+	if (len == 0)
+		{
+		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_BUFFER_TOO_SMALL);
+		return 0;
+		}
+	form = buf[0];
+	y_bit = form & 1;
+	form = form & ~1U;
+	if ((form != 0)	&& (form != POINT_CONVERSION_COMPRESSED)
+		&& (form != POINT_CONVERSION_UNCOMPRESSED)
+		&& (form != POINT_CONVERSION_HYBRID))
+		{
+		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+		return 0;
+		}
+	if ((form == 0 || form == POINT_CONVERSION_UNCOMPRESSED) && y_bit)
+		{
+		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+		return 0;
+		}
+
+	if (form == 0)
+		{
+		if (len != 1)
+			{
+			ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+			return 0;
+			}
+
+		return EC_POINT_set_to_infinity(group, point);
+		}
+	
+	field_len = (EC_GROUP_get_degree(group) + 7) / 8;
+	enc_len = (form == POINT_CONVERSION_COMPRESSED) ? 1 + field_len : 1 + 2*field_len;
+
+	if (len != enc_len)
+		{
+		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+		return 0;
+		}
+
+	if (ctx == NULL)
+		{
+		ctx = new_ctx = BN_CTX_new();
+		if (ctx == NULL)
+			return 0;
+		}
+
+	BN_CTX_start(ctx);
+	x = BN_CTX_get(ctx);
+	y = BN_CTX_get(ctx);
+	yxi = BN_CTX_get(ctx);
+	if (yxi == NULL) goto err;
+
+	if (!BN_bin2bn(buf + 1, field_len, x)) goto err;
+	if (BN_ucmp(x, &group->field) >= 0)
+		{
+		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+		goto err;
+		}
+
+	if (form == POINT_CONVERSION_COMPRESSED)
+		{
+		if (!EC_POINT_set_compressed_coordinates_GF2m(group, point, x, y_bit, ctx)) goto err;
+		}
+	else
+		{
+		if (!BN_bin2bn(buf + 1 + field_len, field_len, y)) goto err;
+		if (BN_ucmp(y, &group->field) >= 0)
+			{
+			ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+			goto err;
+			}
+		if (form == POINT_CONVERSION_HYBRID)
+			{
+			if (!group->meth->field_div(group, yxi, y, x, ctx)) goto err;
+			if (y_bit != BN_is_odd(yxi))
+				{
+				ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+				goto err;
+				}
+			}
+
+		if (!EC_POINT_set_affine_coordinates_GF2m(group, point, x, y, ctx)) goto err;
+		}
+	
+	if (!EC_POINT_is_on_curve(group, point, ctx)) /* test required by X9.62 */
+		{
+		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_POINT_IS_NOT_ON_CURVE);
+		goto err;
+		}
+
+	ret = 1;
+	
+ err:
+	BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return ret;
+	}
+#endif
diff --git a/crypto/ec/ec2_smpl.c b/crypto/ec/ec2_smpl.c
index c06b3b667f0b..e0e59c7d8299 100644
--- a/crypto/ec/ec2_smpl.c
+++ b/crypto/ec/ec2_smpl.c
@@ -14,7 +14,7 @@
  *
  */
 /* ====================================================================
- * Copyright (c) 1998-2003 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1998-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -71,10 +71,20 @@
 
 #include "ec_lcl.h"
 
+#ifndef OPENSSL_NO_EC2M
+
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
+
 
 const EC_METHOD *EC_GF2m_simple_method(void)
 	{
+#ifdef OPENSSL_FIPS
+	return fips_ec_gf2m_simple_method();
+#else
 	static const EC_METHOD ret = {
+		EC_FLAGS_DEFAULT_OCT,
 		NID_X9_62_characteristic_two_field,
 		ec_GF2m_simple_group_init,
 		ec_GF2m_simple_group_finish,
@@ -93,9 +103,7 @@ const EC_METHOD *EC_GF2m_simple_method(void)
 		0 /* get_Jprojective_coordinates_GFp */,
 		ec_GF2m_simple_point_set_affine_coordinates,
 		ec_GF2m_simple_point_get_affine_coordinates,
-		ec_GF2m_simple_set_compressed_coordinates,
-		ec_GF2m_simple_point2oct,
-		ec_GF2m_simple_oct2point,
+		0,0,0,
 		ec_GF2m_simple_add,
 		ec_GF2m_simple_dbl,
 		ec_GF2m_simple_invert,
@@ -118,6 +126,7 @@ const EC_METHOD *EC_GF2m_simple_method(void)
 		0 /* field_set_to_one */ };
 
 	return &ret;
+#endif
 	}
 
 
@@ -157,6 +166,7 @@ void ec_GF2m_simple_group_clear_finish(EC_GROUP *group)
 	group->poly[2] = 0;
 	group->poly[3] = 0;
 	group->poly[4] = 0;
+	group->poly[5] = -1;
 	}
 
 
@@ -174,10 +184,9 @@ int ec_GF2m_simple_group_copy(EC_GROUP *dest, const EC_GROUP *src)
 	dest->poly[2] = src->poly[2];
 	dest->poly[3] = src->poly[3];
 	dest->poly[4] = src->poly[4];
-	if(bn_wexpand(&dest->a, (int)(dest->poly[0] + BN_BITS2 - 1) / BN_BITS2) == NULL)
-		return 0;
-	if(bn_wexpand(&dest->b, (int)(dest->poly[0] + BN_BITS2 - 1) / BN_BITS2) == NULL)
-		return 0;
+	dest->poly[5] = src->poly[5];
+	if (bn_wexpand(&dest->a, (int)(dest->poly[0] + BN_BITS2 - 1) / BN_BITS2) == NULL) return 0;
+	if (bn_wexpand(&dest->b, (int)(dest->poly[0] + BN_BITS2 - 1) / BN_BITS2) == NULL) return 0;
 	for (i = dest->a.top; i < dest->a.dmax; i++) dest->a.d[i] = 0;
 	for (i = dest->b.top; i < dest->b.dmax; i++) dest->b.d[i] = 0;
 	return 1;
@@ -192,7 +201,7 @@ int ec_GF2m_simple_group_set_curve(EC_GROUP *group,
 
 	/* group->field */
 	if (!BN_copy(&group->field, p)) goto err;
-	i = BN_GF2m_poly2arr(&group->field, group->poly, 5);
+	i = BN_GF2m_poly2arr(&group->field, group->poly, 6) - 1;
 	if ((i != 5) && (i != 3))
 		{
 		ECerr(EC_F_EC_GF2M_SIMPLE_GROUP_SET_CURVE, EC_R_UNSUPPORTED_FIELD);
@@ -405,274 +414,6 @@ int ec_GF2m_simple_point_get_affine_coordinates(const EC_GROUP *group, const EC_
 	return ret;
 	}
 
-
-/* Include patented algorithms. */
-#include "ec2_smpt.c"
-
-
-/* Converts an EC_POINT to an octet string.  
- * If buf is NULL, the encoded length will be returned.
- * If the length len of buf is smaller than required an error will be returned.
- *
- * The point compression section of this function is patented by Certicom Corp. 
- * under US Patent 6,141,420.  Point compression is disabled by default and can 
- * be enabled by defining the preprocessor macro OPENSSL_EC_BIN_PT_COMP at 
- * Configure-time.
- */
-size_t ec_GF2m_simple_point2oct(const EC_GROUP *group, const EC_POINT *point, point_conversion_form_t form,
-	unsigned char *buf, size_t len, BN_CTX *ctx)
-	{
-	size_t ret;
-	BN_CTX *new_ctx = NULL;
-	int used_ctx = 0;
-	BIGNUM *x, *y, *yxi;
-	size_t field_len, i, skip;
-
-#ifndef OPENSSL_EC_BIN_PT_COMP
-	if ((form == POINT_CONVERSION_COMPRESSED) || (form == POINT_CONVERSION_HYBRID)) 
-		{
-		ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, ERR_R_DISABLED);
-		goto err;
-		}
-#endif
-
-	if ((form != POINT_CONVERSION_COMPRESSED)
-		&& (form != POINT_CONVERSION_UNCOMPRESSED)
-		&& (form != POINT_CONVERSION_HYBRID))
-		{
-		ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, EC_R_INVALID_FORM);
-		goto err;
-		}
-
-	if (EC_POINT_is_at_infinity(group, point))
-		{
-		/* encodes to a single 0 octet */
-		if (buf != NULL)
-			{
-			if (len < 1)
-				{
-				ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, EC_R_BUFFER_TOO_SMALL);
-				return 0;
-				}
-			buf[0] = 0;
-			}
-		return 1;
-		}
-
-
-	/* ret := required output buffer length */
-	field_len = (EC_GROUP_get_degree(group) + 7) / 8;
-	ret = (form == POINT_CONVERSION_COMPRESSED) ? 1 + field_len : 1 + 2*field_len;
-
-	/* if 'buf' is NULL, just return required length */
-	if (buf != NULL)
-		{
-		if (len < ret)
-			{
-			ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, EC_R_BUFFER_TOO_SMALL);
-			goto err;
-			}
-
-		if (ctx == NULL)
-			{
-			ctx = new_ctx = BN_CTX_new();
-			if (ctx == NULL)
-				return 0;
-			}
-
-		BN_CTX_start(ctx);
-		used_ctx = 1;
-		x = BN_CTX_get(ctx);
-		y = BN_CTX_get(ctx);
-		yxi = BN_CTX_get(ctx);
-		if (yxi == NULL) goto err;
-
-		if (!EC_POINT_get_affine_coordinates_GF2m(group, point, x, y, ctx)) goto err;
-
-		buf[0] = form;
-#ifdef OPENSSL_EC_BIN_PT_COMP
-		if ((form != POINT_CONVERSION_UNCOMPRESSED) && !BN_is_zero(x))
-			{
-			if (!group->meth->field_div(group, yxi, y, x, ctx)) goto err;
-			if (BN_is_odd(yxi)) buf[0]++;
-			}
-#endif
-
-		i = 1;
-		
-		skip = field_len - BN_num_bytes(x);
-		if (skip > field_len)
-			{
-			ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
-			goto err;
-			}
-		while (skip > 0)
-			{
-			buf[i++] = 0;
-			skip--;
-			}
-		skip = BN_bn2bin(x, buf + i);
-		i += skip;
-		if (i != 1 + field_len)
-			{
-			ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
-			goto err;
-			}
-
-		if (form == POINT_CONVERSION_UNCOMPRESSED || form == POINT_CONVERSION_HYBRID)
-			{
-			skip = field_len - BN_num_bytes(y);
-			if (skip > field_len)
-				{
-				ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
-				goto err;
-				}
-			while (skip > 0)
-				{
-				buf[i++] = 0;
-				skip--;
-				}
-			skip = BN_bn2bin(y, buf + i);
-			i += skip;
-			}
-
-		if (i != ret)
-			{
-			ECerr(EC_F_EC_GF2M_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
-			goto err;
-			}
-		}
-	
-	if (used_ctx)
-		BN_CTX_end(ctx);
-	if (new_ctx != NULL)
-		BN_CTX_free(new_ctx);
-	return ret;
-
- err:
-	if (used_ctx)
-		BN_CTX_end(ctx);
-	if (new_ctx != NULL)
-		BN_CTX_free(new_ctx);
-	return 0;
-	}
-
-
-/* Converts an octet string representation to an EC_POINT. 
- * Note that the simple implementation only uses affine coordinates.
- */
-int ec_GF2m_simple_oct2point(const EC_GROUP *group, EC_POINT *point,
-	const unsigned char *buf, size_t len, BN_CTX *ctx)
-	{
-	point_conversion_form_t form;
-	int y_bit;
-	BN_CTX *new_ctx = NULL;
-	BIGNUM *x, *y, *yxi;
-	size_t field_len, enc_len;
-	int ret = 0;
-
-	if (len == 0)
-		{
-		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_BUFFER_TOO_SMALL);
-		return 0;
-		}
-	form = buf[0];
-	y_bit = form & 1;
-	form = form & ~1U;
-	if ((form != 0)	&& (form != POINT_CONVERSION_COMPRESSED)
-		&& (form != POINT_CONVERSION_UNCOMPRESSED)
-		&& (form != POINT_CONVERSION_HYBRID))
-		{
-		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-		return 0;
-		}
-	if ((form == 0 || form == POINT_CONVERSION_UNCOMPRESSED) && y_bit)
-		{
-		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-		return 0;
-		}
-
-	if (form == 0)
-		{
-		if (len != 1)
-			{
-			ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-			return 0;
-			}
-
-		return EC_POINT_set_to_infinity(group, point);
-		}
-	
-	field_len = (EC_GROUP_get_degree(group) + 7) / 8;
-	enc_len = (form == POINT_CONVERSION_COMPRESSED) ? 1 + field_len : 1 + 2*field_len;
-
-	if (len != enc_len)
-		{
-		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-		return 0;
-		}
-
-	if (ctx == NULL)
-		{
-		ctx = new_ctx = BN_CTX_new();
-		if (ctx == NULL)
-			return 0;
-		}
-
-	BN_CTX_start(ctx);
-	x = BN_CTX_get(ctx);
-	y = BN_CTX_get(ctx);
-	yxi = BN_CTX_get(ctx);
-	if (yxi == NULL) goto err;
-
-	if (!BN_bin2bn(buf + 1, field_len, x)) goto err;
-	if (BN_ucmp(x, &group->field) >= 0)
-		{
-		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-		goto err;
-		}
-
-	if (form == POINT_CONVERSION_COMPRESSED)
-		{
-		if (!EC_POINT_set_compressed_coordinates_GF2m(group, point, x, y_bit, ctx)) goto err;
-		}
-	else
-		{
-		if (!BN_bin2bn(buf + 1 + field_len, field_len, y)) goto err;
-		if (BN_ucmp(y, &group->field) >= 0)
-			{
-			ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-			goto err;
-			}
-		if (form == POINT_CONVERSION_HYBRID)
-			{
-			if (!group->meth->field_div(group, yxi, y, x, ctx)) goto err;
-			if (y_bit != BN_is_odd(yxi))
-				{
-				ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-				goto err;
-				}
-			}
-
-		if (!EC_POINT_set_affine_coordinates_GF2m(group, point, x, y, ctx)) goto err;
-		}
-	
-	if (!EC_POINT_is_on_curve(group, point, ctx)) /* test required by X9.62 */
-		{
-		ECerr(EC_F_EC_GF2M_SIMPLE_OCT2POINT, EC_R_POINT_IS_NOT_ON_CURVE);
-		goto err;
-		}
-
-	ret = 1;
-	
- err:
-	BN_CTX_end(ctx);
-	if (new_ctx != NULL)
-		BN_CTX_free(new_ctx);
-	return ret;
-	}
-
-
 /* Computes a + b and stores the result in r.  r could be a or b, a could be b.
  * Uses algorithm A.10.2 of IEEE P1363.
  */
@@ -974,3 +715,5 @@ int ec_GF2m_simple_field_div(const EC_GROUP *group, BIGNUM *r, const BIGNUM *a,
 	{
 	return BN_GF2m_mod_div(r, a, b, &group->field, ctx);
 	}
+
+#endif
diff --git a/crypto/ec/ec_ameth.c b/crypto/ec/ec_ameth.c
new file mode 100644
index 000000000000..83909c185359
--- /dev/null
+++ b/crypto/ec/ec_ameth.c
@@ -0,0 +1,660 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/x509.h>
+#include <openssl/ec.h>
+#include <openssl/bn.h>
+#ifndef OPENSSL_NO_CMS
+#include <openssl/cms.h>
+#endif
+#include "asn1_locl.h"
+
+static int eckey_param2type(int *pptype, void **ppval, EC_KEY *ec_key)
+	{
+	const EC_GROUP  *group;
+	int nid;
+	if (ec_key == NULL || (group = EC_KEY_get0_group(ec_key)) == NULL) 
+	{
+		ECerr(EC_F_ECKEY_PARAM2TYPE, EC_R_MISSING_PARAMETERS);
+		return 0;
+	}
+	if (EC_GROUP_get_asn1_flag(group)
+                     && (nid = EC_GROUP_get_curve_name(group)))
+		/* we have a 'named curve' => just set the OID */
+		{
+		*ppval = OBJ_nid2obj(nid);
+		*pptype = V_ASN1_OBJECT;
+		}
+	else	/* explicit parameters */
+		{
+		ASN1_STRING *pstr = NULL;
+		pstr = ASN1_STRING_new();
+		if (!pstr)
+			return 0;
+		pstr->length = i2d_ECParameters(ec_key, &pstr->data);
+		if (pstr->length < 0)
+			{
+			ASN1_STRING_free(pstr);
+			ECerr(EC_F_ECKEY_PARAM2TYPE, ERR_R_EC_LIB);
+			return 0;
+			}
+		*ppval = pstr;
+		*pptype = V_ASN1_SEQUENCE;
+		}
+	return 1;
+	}
+
+static int eckey_pub_encode(X509_PUBKEY *pk, const EVP_PKEY *pkey)
+	{
+	EC_KEY *ec_key = pkey->pkey.ec;
+	void *pval = NULL;
+	int ptype;
+	unsigned char *penc = NULL, *p;
+	int penclen;
+
+	if (!eckey_param2type(&ptype, &pval, ec_key))
+		{
+		ECerr(EC_F_ECKEY_PUB_ENCODE, ERR_R_EC_LIB);
+		return 0;
+		}
+	penclen = i2o_ECPublicKey(ec_key, NULL);
+	if (penclen <= 0)
+		goto err;
+	penc = OPENSSL_malloc(penclen);
+	if (!penc)
+		goto err;
+	p = penc;
+	penclen = i2o_ECPublicKey(ec_key, &p);
+	if (penclen <= 0)
+		goto err;
+	if (X509_PUBKEY_set0_param(pk, OBJ_nid2obj(EVP_PKEY_EC),
+				ptype, pval, penc, penclen))
+		return 1;
+	err:
+	if (ptype == V_ASN1_OBJECT)
+		ASN1_OBJECT_free(pval);
+	else
+		ASN1_STRING_free(pval);
+	if (penc)
+		OPENSSL_free(penc);
+	return 0;
+	}
+
+static EC_KEY *eckey_type2param(int ptype, void *pval)
+	{
+	EC_KEY *eckey = NULL;
+	if (ptype == V_ASN1_SEQUENCE)
+		{
+		ASN1_STRING *pstr = pval;
+		const unsigned char *pm = NULL;
+		int pmlen;
+		pm = pstr->data;
+		pmlen = pstr->length;
+		if (!(eckey = d2i_ECParameters(NULL, &pm, pmlen)))
+			{
+			ECerr(EC_F_ECKEY_TYPE2PARAM, EC_R_DECODE_ERROR);
+			goto ecerr;
+			}
+		}
+	else if (ptype == V_ASN1_OBJECT)
+		{
+		ASN1_OBJECT *poid = pval;
+		EC_GROUP *group;
+
+		/* type == V_ASN1_OBJECT => the parameters are given
+		 * by an asn1 OID
+		 */
+		if ((eckey = EC_KEY_new()) == NULL)
+			{
+			ECerr(EC_F_ECKEY_TYPE2PARAM, ERR_R_MALLOC_FAILURE);
+			goto ecerr;
+			}
+		group = EC_GROUP_new_by_curve_name(OBJ_obj2nid(poid));
+		if (group == NULL)
+			goto ecerr;
+		EC_GROUP_set_asn1_flag(group, OPENSSL_EC_NAMED_CURVE);
+		if (EC_KEY_set_group(eckey, group) == 0)
+			goto ecerr;
+		EC_GROUP_free(group);
+		}
+	else
+		{
+		ECerr(EC_F_ECKEY_TYPE2PARAM, EC_R_DECODE_ERROR);
+		goto ecerr;
+		}
+
+	return eckey;
+
+	ecerr:
+	if (eckey)
+		EC_KEY_free(eckey);
+	return NULL;
+	}
+
+static int eckey_pub_decode(EVP_PKEY *pkey, X509_PUBKEY *pubkey)
+	{
+	const unsigned char *p = NULL;
+	void *pval;
+	int ptype, pklen;
+	EC_KEY *eckey = NULL;
+	X509_ALGOR *palg;
+
+	if (!X509_PUBKEY_get0_param(NULL, &p, &pklen, &palg, pubkey))
+		return 0;
+	X509_ALGOR_get0(NULL, &ptype, &pval, palg);
+
+	eckey = eckey_type2param(ptype, pval);
+
+	if (!eckey)
+		{
+		ECerr(EC_F_ECKEY_PUB_DECODE, ERR_R_EC_LIB);
+		return 0;
+		}
+
+	/* We have parameters now set public key */
+	if (!o2i_ECPublicKey(&eckey, &p, pklen))
+		{
+		ECerr(EC_F_ECKEY_PUB_DECODE, EC_R_DECODE_ERROR);
+		goto ecerr;
+		}
+
+	EVP_PKEY_assign_EC_KEY(pkey, eckey);
+	return 1;
+
+	ecerr:
+	if (eckey)
+		EC_KEY_free(eckey);
+	return 0;
+	}
+
+static int eckey_pub_cmp(const EVP_PKEY *a, const EVP_PKEY *b)
+	{
+	int  r;
+	const EC_GROUP *group = EC_KEY_get0_group(b->pkey.ec);
+	const EC_POINT *pa = EC_KEY_get0_public_key(a->pkey.ec),
+	               *pb = EC_KEY_get0_public_key(b->pkey.ec);
+	r = EC_POINT_cmp(group, pa, pb, NULL);
+	if (r == 0)
+		return 1;
+	if (r == 1)
+		return 0;
+	return -2;
+	}
+
+static int eckey_priv_decode(EVP_PKEY *pkey, PKCS8_PRIV_KEY_INFO *p8)
+	{
+	const unsigned char *p = NULL;
+	void *pval;
+	int ptype, pklen;
+	EC_KEY *eckey = NULL;
+	X509_ALGOR *palg;
+
+	if (!PKCS8_pkey_get0(NULL, &p, &pklen, &palg, p8))
+		return 0;
+	X509_ALGOR_get0(NULL, &ptype, &pval, palg);
+
+	eckey = eckey_type2param(ptype, pval);
+
+	if (!eckey)
+		goto ecliberr;
+
+	/* We have parameters now set private key */
+	if (!d2i_ECPrivateKey(&eckey, &p, pklen))
+		{
+		ECerr(EC_F_ECKEY_PRIV_DECODE, EC_R_DECODE_ERROR);
+		goto ecerr;
+		}
+
+	/* calculate public key (if necessary) */
+	if (EC_KEY_get0_public_key(eckey) == NULL)
+		{
+		const BIGNUM *priv_key;
+		const EC_GROUP *group;
+		EC_POINT *pub_key;
+		/* the public key was not included in the SEC1 private
+		 * key => calculate the public key */
+		group   = EC_KEY_get0_group(eckey);
+		pub_key = EC_POINT_new(group);
+		if (pub_key == NULL)
+			{
+			ECerr(EC_F_ECKEY_PRIV_DECODE, ERR_R_EC_LIB);
+			goto ecliberr;
+			}
+		if (!EC_POINT_copy(pub_key, EC_GROUP_get0_generator(group)))
+			{
+			EC_POINT_free(pub_key);
+			ECerr(EC_F_ECKEY_PRIV_DECODE, ERR_R_EC_LIB);
+			goto ecliberr;
+			}
+		priv_key = EC_KEY_get0_private_key(eckey);
+		if (!EC_POINT_mul(group, pub_key, priv_key, NULL, NULL, NULL))
+			{
+			EC_POINT_free(pub_key);
+			ECerr(EC_F_ECKEY_PRIV_DECODE, ERR_R_EC_LIB);
+			goto ecliberr;
+			}
+		if (EC_KEY_set_public_key(eckey, pub_key) == 0)
+			{
+			EC_POINT_free(pub_key);
+			ECerr(EC_F_ECKEY_PRIV_DECODE, ERR_R_EC_LIB);
+			goto ecliberr;
+			}
+		EC_POINT_free(pub_key);
+		}
+
+	EVP_PKEY_assign_EC_KEY(pkey, eckey);
+	return 1;
+
+	ecliberr:
+	ECerr(EC_F_ECKEY_PRIV_DECODE, ERR_R_EC_LIB);
+	ecerr:
+	if (eckey)
+		EC_KEY_free(eckey);
+	return 0;
+	}
+
+static int eckey_priv_encode(PKCS8_PRIV_KEY_INFO *p8, const EVP_PKEY *pkey)
+{
+	EC_KEY		*ec_key;
+	unsigned char	*ep, *p;
+	int 		eplen, ptype;
+	void		*pval;
+	unsigned int    tmp_flags, old_flags;
+
+	ec_key = pkey->pkey.ec;
+
+	if (!eckey_param2type(&ptype, &pval, ec_key))
+		{
+		ECerr(EC_F_ECKEY_PRIV_ENCODE, EC_R_DECODE_ERROR);
+		return 0;
+		}
+
+	/* set the private key */
+
+	/* do not include the parameters in the SEC1 private key
+	 * see PKCS#11 12.11 */
+	old_flags = EC_KEY_get_enc_flags(ec_key);
+	tmp_flags = old_flags | EC_PKEY_NO_PARAMETERS;
+	EC_KEY_set_enc_flags(ec_key, tmp_flags);
+	eplen = i2d_ECPrivateKey(ec_key, NULL);
+	if (!eplen)
+	{
+		EC_KEY_set_enc_flags(ec_key, old_flags);
+		ECerr(EC_F_ECKEY_PRIV_ENCODE, ERR_R_EC_LIB);
+		return 0;
+	}
+	ep = (unsigned char *) OPENSSL_malloc(eplen);
+	if (!ep)
+	{
+		EC_KEY_set_enc_flags(ec_key, old_flags);
+		ECerr(EC_F_ECKEY_PRIV_ENCODE, ERR_R_MALLOC_FAILURE);
+		return 0;
+	}
+	p = ep;
+	if (!i2d_ECPrivateKey(ec_key, &p))
+	{
+		EC_KEY_set_enc_flags(ec_key, old_flags);
+		OPENSSL_free(ep);
+		ECerr(EC_F_ECKEY_PRIV_ENCODE, ERR_R_EC_LIB);
+	}
+	/* restore old encoding flags */
+	EC_KEY_set_enc_flags(ec_key, old_flags);
+
+	if (!PKCS8_pkey_set0(p8, OBJ_nid2obj(NID_X9_62_id_ecPublicKey), 0,
+				ptype, pval, ep, eplen))
+		return 0;
+
+	return 1;
+}
+
+static int int_ec_size(const EVP_PKEY *pkey)
+	{
+	return ECDSA_size(pkey->pkey.ec);
+	}
+
+static int ec_bits(const EVP_PKEY *pkey)
+	{
+	BIGNUM *order = BN_new();
+	const EC_GROUP *group;
+	int ret;
+
+	if (!order)
+		{
+		ERR_clear_error();
+		return 0;
+		}
+	group = EC_KEY_get0_group(pkey->pkey.ec);
+	if (!EC_GROUP_get_order(group, order, NULL))
+		{
+		ERR_clear_error();
+		return 0;
+		}
+
+	ret = BN_num_bits(order);
+	BN_free(order);
+	return ret;
+	}
+
+static int ec_missing_parameters(const EVP_PKEY *pkey)
+	{
+	if (EC_KEY_get0_group(pkey->pkey.ec) == NULL)
+		return 1;
+	return 0;
+	}
+
+static int ec_copy_parameters(EVP_PKEY *to, const EVP_PKEY *from)
+	{
+	EC_GROUP *group = EC_GROUP_dup(EC_KEY_get0_group(from->pkey.ec));
+	if (group == NULL)
+		return 0;
+	if (EC_KEY_set_group(to->pkey.ec, group) == 0)
+		return 0;
+	EC_GROUP_free(group);
+	return 1;
+	}
+
+static int ec_cmp_parameters(const EVP_PKEY *a, const EVP_PKEY *b)
+	{
+	const EC_GROUP *group_a = EC_KEY_get0_group(a->pkey.ec),
+	               *group_b = EC_KEY_get0_group(b->pkey.ec);
+	if (EC_GROUP_cmp(group_a, group_b, NULL))
+		return 0;
+	else
+		return 1;
+	}
+
+static void int_ec_free(EVP_PKEY *pkey)
+	{
+	EC_KEY_free(pkey->pkey.ec);
+	}
+
+static int do_EC_KEY_print(BIO *bp, const EC_KEY *x, int off, int ktype)
+	{
+	unsigned char *buffer=NULL;
+	const char *ecstr;
+	size_t	buf_len=0, i;
+	int     ret=0, reason=ERR_R_BIO_LIB;
+	BIGNUM  *pub_key=NULL, *order=NULL;
+	BN_CTX  *ctx=NULL;
+	const EC_GROUP *group;
+	const EC_POINT *public_key;
+	const BIGNUM *priv_key;
+ 
+	if (x == NULL || (group = EC_KEY_get0_group(x)) == NULL)
+		{
+		reason = ERR_R_PASSED_NULL_PARAMETER;
+		goto err;
+		}
+
+	ctx = BN_CTX_new();
+	if (ctx == NULL)
+		{
+		reason = ERR_R_MALLOC_FAILURE;
+		goto err;
+		}
+
+	if (ktype > 0)
+		{
+		public_key = EC_KEY_get0_public_key(x);
+		if ((pub_key = EC_POINT_point2bn(group, public_key,
+			EC_KEY_get_conv_form(x), NULL, ctx)) == NULL)
+			{
+			reason = ERR_R_EC_LIB;
+			goto err;
+			}
+		if (pub_key)
+			buf_len = (size_t)BN_num_bytes(pub_key);
+		}
+
+	if (ktype == 2)
+		{
+		priv_key = EC_KEY_get0_private_key(x);
+		if (priv_key && (i = (size_t)BN_num_bytes(priv_key)) > buf_len)
+			buf_len = i;
+		}
+	else
+		priv_key = NULL;
+
+	if (ktype > 0)
+		{
+		buf_len += 10;
+		if ((buffer = OPENSSL_malloc(buf_len)) == NULL)
+			{
+			reason = ERR_R_MALLOC_FAILURE;
+			goto err;
+			}
+		}
+	if (ktype == 2)
+		ecstr = "Private-Key";
+	else if (ktype == 1)
+		ecstr = "Public-Key";
+	else
+		ecstr = "ECDSA-Parameters";
+
+	if (!BIO_indent(bp, off, 128))
+		goto err;
+	if ((order = BN_new()) == NULL)
+		goto err;
+	if (!EC_GROUP_get_order(group, order, NULL))
+		goto err;
+	if (BIO_printf(bp, "%s: (%d bit)\n", ecstr,
+		BN_num_bits(order)) <= 0) goto err;
+  
+	if ((priv_key != NULL) && !ASN1_bn_print(bp, "priv:", priv_key, 
+		buffer, off))
+		goto err;
+	if ((pub_key != NULL) && !ASN1_bn_print(bp, "pub: ", pub_key,
+		buffer, off))
+		goto err;
+	if (!ECPKParameters_print(bp, group, off))
+		goto err;
+	ret=1;
+err:
+	if (!ret)
+ 		ECerr(EC_F_DO_EC_KEY_PRINT, reason);
+	if (pub_key) 
+		BN_free(pub_key);
+	if (order)
+		BN_free(order);
+	if (ctx)
+		BN_CTX_free(ctx);
+	if (buffer != NULL)
+		OPENSSL_free(buffer);
+	return(ret);
+	}
+
+static int eckey_param_decode(EVP_PKEY *pkey,
+					const unsigned char **pder, int derlen)
+	{
+	EC_KEY *eckey;
+	if (!(eckey = d2i_ECParameters(NULL, pder, derlen)))
+		{
+		ECerr(EC_F_ECKEY_PARAM_DECODE, ERR_R_EC_LIB);
+		return 0;
+		}
+	EVP_PKEY_assign_EC_KEY(pkey, eckey);
+	return 1;
+	}
+
+static int eckey_param_encode(const EVP_PKEY *pkey, unsigned char **pder)
+	{
+	return i2d_ECParameters(pkey->pkey.ec, pder);
+	}
+
+static int eckey_param_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_EC_KEY_print(bp, pkey->pkey.ec, indent, 0);
+	}
+
+static int eckey_pub_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_EC_KEY_print(bp, pkey->pkey.ec, indent, 1);
+	}
+
+
+static int eckey_priv_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_EC_KEY_print(bp, pkey->pkey.ec, indent, 2);
+	}
+
+static int old_ec_priv_decode(EVP_PKEY *pkey,
+					const unsigned char **pder, int derlen)
+	{
+	EC_KEY *ec;
+	if (!(ec = d2i_ECPrivateKey (NULL, pder, derlen)))
+		{
+		ECerr(EC_F_OLD_EC_PRIV_DECODE, EC_R_DECODE_ERROR);
+		return 0;
+		}
+	EVP_PKEY_assign_EC_KEY(pkey, ec);
+	return 1;
+	}
+
+static int old_ec_priv_encode(const EVP_PKEY *pkey, unsigned char **pder)
+	{
+	return i2d_ECPrivateKey(pkey->pkey.ec, pder);
+	}
+
+static int ec_pkey_ctrl(EVP_PKEY *pkey, int op, long arg1, void *arg2)
+	{
+	switch (op)
+		{
+		case ASN1_PKEY_CTRL_PKCS7_SIGN:
+		if (arg1 == 0)
+			{
+			int snid, hnid;
+			X509_ALGOR *alg1, *alg2;
+			PKCS7_SIGNER_INFO_get0_algs(arg2, NULL, &alg1, &alg2);
+			if (alg1 == NULL || alg1->algorithm == NULL)
+				return -1;
+			hnid = OBJ_obj2nid(alg1->algorithm);
+			if (hnid == NID_undef)
+				return -1;
+			if (!OBJ_find_sigid_by_algs(&snid, hnid, EVP_PKEY_id(pkey)))
+				return -1; 
+			X509_ALGOR_set0(alg2, OBJ_nid2obj(snid), V_ASN1_UNDEF, 0);
+			}
+		return 1;
+#ifndef OPENSSL_NO_CMS
+		case ASN1_PKEY_CTRL_CMS_SIGN:
+		if (arg1 == 0)
+			{
+			int snid, hnid;
+			X509_ALGOR *alg1, *alg2;
+			CMS_SignerInfo_get0_algs(arg2, NULL, NULL,
+								&alg1, &alg2);
+			if (alg1 == NULL || alg1->algorithm == NULL)
+				return -1;
+			hnid = OBJ_obj2nid(alg1->algorithm);
+			if (hnid == NID_undef)
+				return -1;
+			if (!OBJ_find_sigid_by_algs(&snid, hnid, EVP_PKEY_id(pkey)))
+				return -1; 
+			X509_ALGOR_set0(alg2, OBJ_nid2obj(snid), V_ASN1_UNDEF, 0);
+			}
+		return 1;
+#endif
+
+		case ASN1_PKEY_CTRL_DEFAULT_MD_NID:
+		*(int *)arg2 = NID_sha1;
+		return 2;
+
+		default:
+		return -2;
+
+		}
+
+	}
+
+const EVP_PKEY_ASN1_METHOD eckey_asn1_meth = 
+	{
+	EVP_PKEY_EC,
+	EVP_PKEY_EC,
+	0,
+	"EC",
+	"OpenSSL EC algorithm",
+
+	eckey_pub_decode,
+	eckey_pub_encode,
+	eckey_pub_cmp,
+	eckey_pub_print,
+
+	eckey_priv_decode,
+	eckey_priv_encode,
+	eckey_priv_print,
+
+	int_ec_size,
+	ec_bits,
+
+	eckey_param_decode,
+	eckey_param_encode,
+	ec_missing_parameters,
+	ec_copy_parameters,
+	ec_cmp_parameters,
+	eckey_param_print,
+	0,
+
+	int_ec_free,
+	ec_pkey_ctrl,
+	old_ec_priv_decode,
+	old_ec_priv_encode
+	};
diff --git a/crypto/ec/ec_asn1.c b/crypto/ec/ec_asn1.c
index ae555398594b..175eec53428b 100644
--- a/crypto/ec/ec_asn1.c
+++ b/crypto/ec/ec_asn1.c
@@ -83,7 +83,7 @@ int EC_GROUP_get_basis_type(const EC_GROUP *group)
 		/* everything else is currently not supported */
 		return 0;
 	}
-
+#ifndef OPENSSL_NO_EC2M
 int EC_GROUP_get_trinomial_basis(const EC_GROUP *group, unsigned int *k)
 	{
 	if (group == NULL)
@@ -101,7 +101,6 @@ int EC_GROUP_get_trinomial_basis(const EC_GROUP *group, unsigned int *k)
 
 	return 1;
 	}
-
 int EC_GROUP_get_pentanomial_basis(const EC_GROUP *group, unsigned int *k1,
 	unsigned int *k2, unsigned int *k3)
 	{
@@ -124,7 +123,7 @@ int EC_GROUP_get_pentanomial_basis(const EC_GROUP *group, unsigned int *k1,
 
 	return 1;
 	}
-
+#endif
 
 
 /* some structures needed for the asn1 encoding */
@@ -340,6 +339,12 @@ static int ec_asn1_group2fieldid(const EC_GROUP *group, X9_62_FIELDID *field)
 			}
 		}
 	else	/* nid == NID_X9_62_characteristic_two_field */
+#ifdef OPENSSL_NO_EC2M
+		{
+		ECerr(EC_F_EC_ASN1_GROUP2FIELDID, EC_R_GF2M_NOT_SUPPORTED);
+		goto err;
+		}
+#else
 		{
 		int		field_type;
 		X9_62_CHARACTERISTIC_TWO *char_two;
@@ -419,6 +424,7 @@ static int ec_asn1_group2fieldid(const EC_GROUP *group, X9_62_FIELDID *field)
 				}
 			}
 		}
+#endif
 
 	ok = 1;
 
@@ -456,6 +462,7 @@ static int ec_asn1_group2curve(const EC_GROUP *group, X9_62_CURVE *curve)
 			goto err;
 			}
 		}
+#ifndef OPENSSL_NO_EC2M
 	else	/* nid == NID_X9_62_characteristic_two_field */
 		{
 		if (!EC_GROUP_get_curve_GF2m(group, NULL, tmp_1, tmp_2, NULL))
@@ -464,7 +471,7 @@ static int ec_asn1_group2curve(const EC_GROUP *group, X9_62_CURVE *curve)
 			goto err;
 			}
 		}
-
+#endif
 	len_1 = (size_t)BN_num_bytes(tmp_1);
 	len_2 = (size_t)BN_num_bytes(tmp_2);
 
@@ -775,8 +782,13 @@ static EC_GROUP *ec_asn1_parameters2group(const ECPARAMETERS *params)
 
 	/* get the field parameters */
 	tmp = OBJ_obj2nid(params->fieldID->fieldType);
-
 	if (tmp == NID_X9_62_characteristic_two_field)
+#ifdef OPENSSL_NO_EC2M
+		{
+		ECerr(EC_F_EC_ASN1_PARAMETERS2GROUP, EC_R_GF2M_NOT_SUPPORTED);
+		goto err;
+		}
+#else
 		{
 		X9_62_CHARACTERISTIC_TWO *char_two;
 
@@ -862,6 +874,7 @@ static EC_GROUP *ec_asn1_parameters2group(const ECPARAMETERS *params)
 		/* create the EC_GROUP structure */
 		ret = EC_GROUP_new_curve_GF2m(p, a, b, NULL);
 		}
+#endif
 	else if (tmp == NID_X9_62_prime_field)
 		{
 		/* we have a curve over a prime field */
@@ -1065,6 +1078,7 @@ EC_GROUP *d2i_ECPKParameters(EC_GROUP **a, const unsigned char **in, long len)
 	if ((group = ec_asn1_pkparameters2group(params)) == NULL)
 		{
 		ECerr(EC_F_D2I_ECPKPARAMETERS, EC_R_PKPARAMETERS2GROUP_FAILURE);
+		ECPKPARAMETERS_free(params);
 		return NULL; 
 		}
 
diff --git a/crypto/ec/ec_curve.c b/crypto/ec/ec_curve.c
index beac20969b75..c72fb2697ca2 100644
--- a/crypto/ec/ec_curve.c
+++ b/crypto/ec/ec_curve.c
@@ -3,7 +3,7 @@
  * Written by Nils Larsch for the OpenSSL project.
  */
 /* ====================================================================
- * Copyright (c) 1998-2004 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1998-2010 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -72,927 +72,1694 @@
 #include "ec_lcl.h"
 #include <openssl/err.h>
 #include <openssl/obj_mac.h>
+#include <openssl/opensslconf.h>
 
-typedef struct ec_curve_data_st {
-	int	field_type;	/* either NID_X9_62_prime_field or
+typedef struct {
+	int	field_type,	/* either NID_X9_62_prime_field or
 				 * NID_X9_62_characteristic_two_field */
-	const char *p;		/* either a prime number or a polynomial */
-	const char *a;
-	const char *b;
-	const char *x;		/* the x coordinate of the generator */
-	const char *y;		/* the y coordinate of the generator */
-	const char *order;	/* the order of the group generated by the
-				 * generator */
-	const BN_ULONG cofactor;/* the cofactor */
-	const unsigned char *seed;/* the seed (optional) */
-	size_t	seed_len;
-	const char *comment;	/* a short description of the curve */
+		seed_len,
+		param_len;
+	unsigned int cofactor;	/* promoted to BN_ULONG */
 } EC_CURVE_DATA;
 
 /* the nist prime curves */
-static const unsigned char _EC_NIST_PRIME_192_SEED[] = {
-	0x30,0x45,0xAE,0x6F,0xC8,0x42,0x2F,0x64,0xED,0x57,
-	0x95,0x28,0xD3,0x81,0x20,0xEA,0xE1,0x21,0x96,0xD5};
-static const EC_CURVE_DATA _EC_NIST_PRIME_192 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFFFFFFFFFFFF",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFFFFFFFFFFFC",
-	"64210519E59C80E70FA7E9AB72243049FEB8DEECC146B9B1",
-	"188DA80EB03090F67CBF20EB43A18800F4FF0AFD82FF1012",
-	"07192b95ffc8da78631011ed6b24cdd573f977a11e794811",
-	"FFFFFFFFFFFFFFFFFFFFFFFF99DEF836146BC9B1B4D22831",1,
-	_EC_NIST_PRIME_192_SEED, 20,
-	"NIST/X9.62/SECG curve over a 192 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+24*6]; }
+	_EC_NIST_PRIME_192 = {
+	{ NID_X9_62_prime_field,20,24,1 },
+	{ 0x30,0x45,0xAE,0x6F,0xC8,0x42,0x2F,0x64,0xED,0x57,	/* seed */
+	  0x95,0x28,0xD3,0x81,0x20,0xEA,0xE1,0x21,0x96,0xD5,
+
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFC,
+	  0x64,0x21,0x05,0x19,0xE5,0x9C,0x80,0xE7,0x0F,0xA7,	/* b */
+	  0xE9,0xAB,0x72,0x24,0x30,0x49,0xFE,0xB8,0xDE,0xEC,
+	  0xC1,0x46,0xB9,0xB1,
+	  0x18,0x8D,0xA8,0x0E,0xB0,0x30,0x90,0xF6,0x7C,0xBF,	/* x */
+	  0x20,0xEB,0x43,0xA1,0x88,0x00,0xF4,0xFF,0x0A,0xFD,
+	  0x82,0xFF,0x10,0x12,
+	  0x07,0x19,0x2b,0x95,0xff,0xc8,0xda,0x78,0x63,0x10,	/* y */
+	  0x11,0xed,0x6b,0x24,0xcd,0xd5,0x73,0xf9,0x77,0xa1,
+	  0x1e,0x79,0x48,0x11,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0x99,0xDE,0xF8,0x36,0x14,0x6B,0xC9,0xB1,
+	  0xB4,0xD2,0x28,0x31 }
 	};
 
-static const unsigned char _EC_NIST_PRIME_224_SEED[] = {
-	0xBD,0x71,0x34,0x47,0x99,0xD5,0xC7,0xFC,0xDC,0x45,
-	0xB5,0x9F,0xA3,0xB9,0xAB,0x8F,0x6A,0x94,0x8B,0xC5};
-static const EC_CURVE_DATA _EC_NIST_PRIME_224 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF000000000000000000000001",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFFFFFFFFFFFFFFFFFFFE",
-	"B4050A850C04B3ABF54132565044B0B7D7BFD8BA270B39432355FFB4",
-	"B70E0CBD6BB4BF7F321390B94A03C1D356C21122343280D6115C1D21",
-	"bd376388b5f723fb4c22dfe6cd4375a05a07476444d5819985007e34",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFF16A2E0B8F03E13DD29455C5C2A3D",1,
-	_EC_NIST_PRIME_224_SEED, 20,
-	"NIST/SECG curve over a 224 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+28*6]; }
+	_EC_NIST_PRIME_224 = {
+	{ NID_X9_62_prime_field,20,28,1 },
+	{ 0xBD,0x71,0x34,0x47,0x99,0xD5,0xC7,0xFC,0xDC,0x45,	/* seed */
+	  0xB5,0x9F,0xA3,0xB9,0xAB,0x8F,0x6A,0x94,0x8B,0xC5,
+
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,
+	  0xB4,0x05,0x0A,0x85,0x0C,0x04,0xB3,0xAB,0xF5,0x41,	/* b */
+	  0x32,0x56,0x50,0x44,0xB0,0xB7,0xD7,0xBF,0xD8,0xBA,
+	  0x27,0x0B,0x39,0x43,0x23,0x55,0xFF,0xB4,
+	  0xB7,0x0E,0x0C,0xBD,0x6B,0xB4,0xBF,0x7F,0x32,0x13,	/* x */
+	  0x90,0xB9,0x4A,0x03,0xC1,0xD3,0x56,0xC2,0x11,0x22,
+	  0x34,0x32,0x80,0xD6,0x11,0x5C,0x1D,0x21,
+	  0xbd,0x37,0x63,0x88,0xb5,0xf7,0x23,0xfb,0x4c,0x22,	/* y */
+	  0xdf,0xe6,0xcd,0x43,0x75,0xa0,0x5a,0x07,0x47,0x64,
+	  0x44,0xd5,0x81,0x99,0x85,0x00,0x7e,0x34,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0x16,0xA2,0xE0,0xB8,0xF0,0x3E,
+	  0x13,0xDD,0x29,0x45,0x5C,0x5C,0x2A,0x3D }
 	};
 
-static const unsigned char _EC_NIST_PRIME_384_SEED[] = {
-	0xA3,0x35,0x92,0x6A,0xA3,0x19,0xA2,0x7A,0x1D,0x00,
-	0x89,0x6A,0x67,0x73,0xA4,0x82,0x7A,0xCD,0xAC,0x73};
-static const EC_CURVE_DATA _EC_NIST_PRIME_384 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFF"
-	"FFF0000000000000000FFFFFFFF",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFF"
-	"FFF0000000000000000FFFFFFFC",
-	"B3312FA7E23EE7E4988E056BE3F82D19181D9C6EFE8141120314088F5013875AC6563"
-	"98D8A2ED19D2A85C8EDD3EC2AEF",
-	"AA87CA22BE8B05378EB1C71EF320AD746E1D3B628BA79B9859F741E082542A385502F"
-	"25DBF55296C3A545E3872760AB7",
-	"3617de4a96262c6f5d9e98bf9292dc29f8f41dbd289a147ce9da3113b5f0b8c00a60b"
-	"1ce1d7e819d7a431d7c90ea0e5f",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFC7634D81F4372DDF581A0"
-	"DB248B0A77AECEC196ACCC52973",1,
-	_EC_NIST_PRIME_384_SEED, 20,
-	"NIST/SECG curve over a 384 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+48*6]; }
+	_EC_NIST_PRIME_384 = {
+	{ NID_X9_62_prime_field,20,48,1 },
+	{ 0xA3,0x35,0x92,0x6A,0xA3,0x19,0xA2,0x7A,0x1D,0x00,	/* seed */
+	  0x89,0x6A,0x67,0x73,0xA4,0x82,0x7A,0xCD,0xAC,0x73,
+
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0xFF,0xFF,0xFF,0xFC,
+	  0xB3,0x31,0x2F,0xA7,0xE2,0x3E,0xE7,0xE4,0x98,0x8E,	/* b */
+	  0x05,0x6B,0xE3,0xF8,0x2D,0x19,0x18,0x1D,0x9C,0x6E,
+	  0xFE,0x81,0x41,0x12,0x03,0x14,0x08,0x8F,0x50,0x13,
+	  0x87,0x5A,0xC6,0x56,0x39,0x8D,0x8A,0x2E,0xD1,0x9D,
+	  0x2A,0x85,0xC8,0xED,0xD3,0xEC,0x2A,0xEF,
+	  0xAA,0x87,0xCA,0x22,0xBE,0x8B,0x05,0x37,0x8E,0xB1,	/* x */
+	  0xC7,0x1E,0xF3,0x20,0xAD,0x74,0x6E,0x1D,0x3B,0x62,
+	  0x8B,0xA7,0x9B,0x98,0x59,0xF7,0x41,0xE0,0x82,0x54,
+	  0x2A,0x38,0x55,0x02,0xF2,0x5D,0xBF,0x55,0x29,0x6C,
+	  0x3A,0x54,0x5E,0x38,0x72,0x76,0x0A,0xB7,
+	  0x36,0x17,0xde,0x4a,0x96,0x26,0x2c,0x6f,0x5d,0x9e,	/* y */
+	  0x98,0xbf,0x92,0x92,0xdc,0x29,0xf8,0xf4,0x1d,0xbd,
+	  0x28,0x9a,0x14,0x7c,0xe9,0xda,0x31,0x13,0xb5,0xf0,
+	  0xb8,0xc0,0x0a,0x60,0xb1,0xce,0x1d,0x7e,0x81,0x9d,
+	  0x7a,0x43,0x1d,0x7c,0x90,0xea,0x0e,0x5f,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xC7,0x63,0x4D,0x81,0xF4,0x37,
+	  0x2D,0xDF,0x58,0x1A,0x0D,0xB2,0x48,0xB0,0xA7,0x7A,
+	  0xEC,0xEC,0x19,0x6A,0xCC,0xC5,0x29,0x73 }
 	};
 
-static const unsigned char _EC_NIST_PRIME_521_SEED[] = {
-	0xD0,0x9E,0x88,0x00,0x29,0x1C,0xB8,0x53,0x96,0xCC,
-	0x67,0x17,0x39,0x32,0x84,0xAA,0xA0,0xDA,0x64,0xBA};
-static const EC_CURVE_DATA _EC_NIST_PRIME_521 = {
-	NID_X9_62_prime_field,
-	"1FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF"
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF",
-	"1FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF"
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFC",
-	"051953EB9618E1C9A1F929A21A0B68540EEA2DA725B99B315F3B8B489918EF109E156"
-	"193951EC7E937B1652C0BD3BB1BF073573DF883D2C34F1EF451FD46B503F00",
-	"C6858E06B70404E9CD9E3ECB662395B4429C648139053FB521F828AF606B4D3DBAA14"
-	"B5E77EFE75928FE1DC127A2FFA8DE3348B3C1856A429BF97E7E31C2E5BD66",
-	"011839296a789a3bc0045c8a5fb42c7d1bd998f54449579b446817afbd17273e662c9"
-	"7ee72995ef42640c550b9013fad0761353c7086a272c24088be94769fd16650",
-	"1FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFA51"
-	"868783BF2F966B7FCC0148F709A5D03BB5C9B8899C47AEBB6FB71E91386409",1,
-	_EC_NIST_PRIME_521_SEED, 20,
-	"NIST/SECG curve over a 521 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+66*6]; }
+	_EC_NIST_PRIME_521 = {
+	{ NID_X9_62_prime_field,20,66,1 },
+	{ 0xD0,0x9E,0x88,0x00,0x29,0x1C,0xB8,0x53,0x96,0xCC,	/* seed */
+	  0x67,0x17,0x39,0x32,0x84,0xAA,0xA0,0xDA,0x64,0xBA,
+
+	  0x01,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0x01,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFC,
+	  0x00,0x51,0x95,0x3E,0xB9,0x61,0x8E,0x1C,0x9A,0x1F,	/* b */
+	  0x92,0x9A,0x21,0xA0,0xB6,0x85,0x40,0xEE,0xA2,0xDA,
+	  0x72,0x5B,0x99,0xB3,0x15,0xF3,0xB8,0xB4,0x89,0x91,
+	  0x8E,0xF1,0x09,0xE1,0x56,0x19,0x39,0x51,0xEC,0x7E,
+	  0x93,0x7B,0x16,0x52,0xC0,0xBD,0x3B,0xB1,0xBF,0x07,
+	  0x35,0x73,0xDF,0x88,0x3D,0x2C,0x34,0xF1,0xEF,0x45,
+	  0x1F,0xD4,0x6B,0x50,0x3F,0x00,
+	  0x00,0xC6,0x85,0x8E,0x06,0xB7,0x04,0x04,0xE9,0xCD,	/* x */
+	  0x9E,0x3E,0xCB,0x66,0x23,0x95,0xB4,0x42,0x9C,0x64,
+	  0x81,0x39,0x05,0x3F,0xB5,0x21,0xF8,0x28,0xAF,0x60,
+	  0x6B,0x4D,0x3D,0xBA,0xA1,0x4B,0x5E,0x77,0xEF,0xE7,
+	  0x59,0x28,0xFE,0x1D,0xC1,0x27,0xA2,0xFF,0xA8,0xDE,
+	  0x33,0x48,0xB3,0xC1,0x85,0x6A,0x42,0x9B,0xF9,0x7E,
+	  0x7E,0x31,0xC2,0xE5,0xBD,0x66,
+	  0x01,0x18,0x39,0x29,0x6a,0x78,0x9a,0x3b,0xc0,0x04,	/* y */
+	  0x5c,0x8a,0x5f,0xb4,0x2c,0x7d,0x1b,0xd9,0x98,0xf5,
+	  0x44,0x49,0x57,0x9b,0x44,0x68,0x17,0xaf,0xbd,0x17,
+	  0x27,0x3e,0x66,0x2c,0x97,0xee,0x72,0x99,0x5e,0xf4,
+	  0x26,0x40,0xc5,0x50,0xb9,0x01,0x3f,0xad,0x07,0x61,
+	  0x35,0x3c,0x70,0x86,0xa2,0x72,0xc2,0x40,0x88,0xbe,
+	  0x94,0x76,0x9f,0xd1,0x66,0x50,
+	  0x01,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFA,0x51,0x86,0x87,0x83,0xBF,0x2F,
+	  0x96,0x6B,0x7F,0xCC,0x01,0x48,0xF7,0x09,0xA5,0xD0,
+	  0x3B,0xB5,0xC9,0xB8,0x89,0x9C,0x47,0xAE,0xBB,0x6F,
+	  0xB7,0x1E,0x91,0x38,0x64,0x09 }
 	};
+
 /* the x9.62 prime curves (minus the nist prime curves) */
-static const unsigned char _EC_X9_62_PRIME_192V2_SEED[] = {
-	0x31,0xA9,0x2E,0xE2,0x02,0x9F,0xD1,0x0D,0x90,0x1B,
-	0x11,0x3E,0x99,0x07,0x10,0xF0,0xD2,0x1A,0xC6,0xB6};
-static const EC_CURVE_DATA _EC_X9_62_PRIME_192V2 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFFFFFFFFFFFF",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFFFFFFFFFFFC",
-	"CC22D6DFB95C6B25E49C0D6364A4E5980C393AA21668D953",
-	"EEA2BAE7E1497842F2DE7769CFE9C989C072AD696F48034A",
-	"6574d11d69b6ec7a672bb82a083df2f2b0847de970b2de15",
-	"FFFFFFFFFFFFFFFFFFFFFFFE5FB1A724DC80418648D8DD31",1,
-	_EC_X9_62_PRIME_192V2_SEED, 20,
-	"X9.62 curve over a 192 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+24*6]; }
+	_EC_X9_62_PRIME_192V2 = {
+	{ NID_X9_62_prime_field,20,24,1 },
+	{ 0x31,0xA9,0x2E,0xE2,0x02,0x9F,0xD1,0x0D,0x90,0x1B,	/* seed */
+	  0x11,0x3E,0x99,0x07,0x10,0xF0,0xD2,0x1A,0xC6,0xB6,
+
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFC,
+	  0xCC,0x22,0xD6,0xDF,0xB9,0x5C,0x6B,0x25,0xE4,0x9C,	/* b */
+	  0x0D,0x63,0x64,0xA4,0xE5,0x98,0x0C,0x39,0x3A,0xA2,
+	  0x16,0x68,0xD9,0x53,
+	  0xEE,0xA2,0xBA,0xE7,0xE1,0x49,0x78,0x42,0xF2,0xDE,	/* x */
+	  0x77,0x69,0xCF,0xE9,0xC9,0x89,0xC0,0x72,0xAD,0x69,
+	  0x6F,0x48,0x03,0x4A,
+	  0x65,0x74,0xd1,0x1d,0x69,0xb6,0xec,0x7a,0x67,0x2b,	/* y */
+	  0xb8,0x2a,0x08,0x3d,0xf2,0xf2,0xb0,0x84,0x7d,0xe9,
+	  0x70,0xb2,0xde,0x15,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFE,0x5F,0xB1,0xA7,0x24,0xDC,0x80,0x41,0x86,
+	  0x48,0xD8,0xDD,0x31 }
 	};
 
-static const unsigned char _EC_X9_62_PRIME_192V3_SEED[] = {
-	0xC4,0x69,0x68,0x44,0x35,0xDE,0xB3,0x78,0xC4,0xB6,
-	0x5C,0xA9,0x59,0x1E,0x2A,0x57,0x63,0x05,0x9A,0x2E};
-static const EC_CURVE_DATA _EC_X9_62_PRIME_192V3 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFFFFFFFFFFFF",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFFFFFFFFFFFC",
-	"22123DC2395A05CAA7423DAECCC94760A7D462256BD56916",
-	"7D29778100C65A1DA1783716588DCE2B8B4AEE8E228F1896",
-	"38a90f22637337334b49dcb66a6dc8f9978aca7648a943b0",
-	"FFFFFFFFFFFFFFFFFFFFFFFF7A62D031C83F4294F640EC13",1,
-	_EC_X9_62_PRIME_192V3_SEED, 20,
-	"X9.62 curve over a 192 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+24*6]; }
+	_EC_X9_62_PRIME_192V3 = {
+	{ NID_X9_62_prime_field,20,24,1 },
+	{ 0xC4,0x69,0x68,0x44,0x35,0xDE,0xB3,0x78,0xC4,0xB6,	/* seed */
+	  0x5C,0xA9,0x59,0x1E,0x2A,0x57,0x63,0x05,0x9A,0x2E,
+
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFC,
+	  0x22,0x12,0x3D,0xC2,0x39,0x5A,0x05,0xCA,0xA7,0x42,	/* b */
+	  0x3D,0xAE,0xCC,0xC9,0x47,0x60,0xA7,0xD4,0x62,0x25,
+	  0x6B,0xD5,0x69,0x16,
+	  0x7D,0x29,0x77,0x81,0x00,0xC6,0x5A,0x1D,0xA1,0x78,	/* x */
+	  0x37,0x16,0x58,0x8D,0xCE,0x2B,0x8B,0x4A,0xEE,0x8E,
+	  0x22,0x8F,0x18,0x96,
+	  0x38,0xa9,0x0f,0x22,0x63,0x73,0x37,0x33,0x4b,0x49,	/* y */
+	  0xdc,0xb6,0x6a,0x6d,0xc8,0xf9,0x97,0x8a,0xca,0x76,
+	  0x48,0xa9,0x43,0xb0,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0x7A,0x62,0xD0,0x31,0xC8,0x3F,0x42,0x94,
+	  0xF6,0x40,0xEC,0x13 }
 	};
 
-static const unsigned char _EC_X9_62_PRIME_239V1_SEED[] = {
-	0xE4,0x3B,0xB4,0x60,0xF0,0xB8,0x0C,0xC0,0xC0,0xB0,
-	0x75,0x79,0x8E,0x94,0x80,0x60,0xF8,0x32,0x1B,0x7D};
-static const EC_CURVE_DATA _EC_X9_62_PRIME_239V1 = {
-	NID_X9_62_prime_field,
-	"7FFFFFFFFFFFFFFFFFFFFFFF7FFFFFFFFFFF8000000000007FFFFFFFFFFF",
-	"7FFFFFFFFFFFFFFFFFFFFFFF7FFFFFFFFFFF8000000000007FFFFFFFFFFC",
-	"6B016C3BDCF18941D0D654921475CA71A9DB2FB27D1D37796185C2942C0A",
-	"0FFA963CDCA8816CCC33B8642BEDF905C3D358573D3F27FBBD3B3CB9AAAF",
-	"7debe8e4e90a5dae6e4054ca530ba04654b36818ce226b39fccb7b02f1ae",
-	"7FFFFFFFFFFFFFFFFFFFFFFF7FFFFF9E5E9A9F5D9071FBD1522688909D0B",1,
-	_EC_X9_62_PRIME_239V1_SEED, 20,
-	"X9.62 curve over a 239 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+30*6]; }
+	_EC_X9_62_PRIME_239V1 = {
+	{ NID_X9_62_prime_field,20,30,1 },
+	{ 0xE4,0x3B,0xB4,0x60,0xF0,0xB8,0x0C,0xC0,0xC0,0xB0,	/* seed */
+	  0x75,0x79,0x8E,0x94,0x80,0x60,0xF8,0x32,0x1B,0x7D,
+
+	  0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0x80,0x00,
+	  0x00,0x00,0x00,0x00,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,
+
+	  0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0x80,0x00,
+	  0x00,0x00,0x00,0x00,0x7F,0xFF,0xFF,0xFF,0xFF,0xFC,
+
+	  0x6B,0x01,0x6C,0x3B,0xDC,0xF1,0x89,0x41,0xD0,0xD6,	/* b */
+	  0x54,0x92,0x14,0x75,0xCA,0x71,0xA9,0xDB,0x2F,0xB2,
+	  0x7D,0x1D,0x37,0x79,0x61,0x85,0xC2,0x94,0x2C,0x0A,
+
+	  0x0F,0xFA,0x96,0x3C,0xDC,0xA8,0x81,0x6C,0xCC,0x33,	/* x */
+	  0xB8,0x64,0x2B,0xED,0xF9,0x05,0xC3,0xD3,0x58,0x57,
+	  0x3D,0x3F,0x27,0xFB,0xBD,0x3B,0x3C,0xB9,0xAA,0xAF,
+
+	  0x7d,0xeb,0xe8,0xe4,0xe9,0x0a,0x5d,0xae,0x6e,0x40,	/* y */
+	  0x54,0xca,0x53,0x0b,0xa0,0x46,0x54,0xb3,0x68,0x18,
+	  0xce,0x22,0x6b,0x39,0xfc,0xcb,0x7b,0x02,0xf1,0xae,
+
+	  0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0x7F,0xFF,0xFF,0x9E,0x5E,0x9A,0x9F,0x5D,
+	  0x90,0x71,0xFB,0xD1,0x52,0x26,0x88,0x90,0x9D,0x0B }
 	};
 
-static const unsigned char _EC_X9_62_PRIME_239V2_SEED[] = {
-	0xE8,0xB4,0x01,0x16,0x04,0x09,0x53,0x03,0xCA,0x3B,
-	0x80,0x99,0x98,0x2B,0xE0,0x9F,0xCB,0x9A,0xE6,0x16};
-static const EC_CURVE_DATA _EC_X9_62_PRIME_239V2 = {
-	NID_X9_62_prime_field,
-	"7FFFFFFFFFFFFFFFFFFFFFFF7FFFFFFFFFFF8000000000007FFFFFFFFFFF",
-	"7FFFFFFFFFFFFFFFFFFFFFFF7FFFFFFFFFFF8000000000007FFFFFFFFFFC",
-	"617FAB6832576CBBFED50D99F0249C3FEE58B94BA0038C7AE84C8C832F2C",
-	"38AF09D98727705120C921BB5E9E26296A3CDCF2F35757A0EAFD87B830E7",
-	"5b0125e4dbea0ec7206da0fc01d9b081329fb555de6ef460237dff8be4ba",
-	"7FFFFFFFFFFFFFFFFFFFFFFF800000CFA7E8594377D414C03821BC582063",1,
-	_EC_X9_62_PRIME_239V2_SEED, 20,
-	"X9.62 curve over a 239 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+30*6]; }
+	_EC_X9_62_PRIME_239V2 = {
+	{ NID_X9_62_prime_field,20,30,1 },
+	{ 0xE8,0xB4,0x01,0x16,0x04,0x09,0x53,0x03,0xCA,0x3B,	/* seed */
+	  0x80,0x99,0x98,0x2B,0xE0,0x9F,0xCB,0x9A,0xE6,0x16,
+
+	  0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0x80,0x00,
+	  0x00,0x00,0x00,0x00,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,
+
+	  0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0x80,0x00,
+	  0x00,0x00,0x00,0x00,0x7F,0xFF,0xFF,0xFF,0xFF,0xFC,
+
+	  0x61,0x7F,0xAB,0x68,0x32,0x57,0x6C,0xBB,0xFE,0xD5,	/* b */
+	  0x0D,0x99,0xF0,0x24,0x9C,0x3F,0xEE,0x58,0xB9,0x4B,
+	  0xA0,0x03,0x8C,0x7A,0xE8,0x4C,0x8C,0x83,0x2F,0x2C,
+
+	  0x38,0xAF,0x09,0xD9,0x87,0x27,0x70,0x51,0x20,0xC9,	/* x */
+	  0x21,0xBB,0x5E,0x9E,0x26,0x29,0x6A,0x3C,0xDC,0xF2,
+	  0xF3,0x57,0x57,0xA0,0xEA,0xFD,0x87,0xB8,0x30,0xE7,
+
+	  0x5b,0x01,0x25,0xe4,0xdb,0xea,0x0e,0xc7,0x20,0x6d,	/* y */
+	  0xa0,0xfc,0x01,0xd9,0xb0,0x81,0x32,0x9f,0xb5,0x55,
+	  0xde,0x6e,0xf4,0x60,0x23,0x7d,0xff,0x8b,0xe4,0xba,
+
+	  0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0x80,0x00,0x00,0xCF,0xA7,0xE8,0x59,0x43,
+	  0x77,0xD4,0x14,0xC0,0x38,0x21,0xBC,0x58,0x20,0x63 }
 	};
 
-static const unsigned char _EC_X9_62_PRIME_239V3_SEED[] = {
-	0x7D,0x73,0x74,0x16,0x8F,0xFE,0x34,0x71,0xB6,0x0A,
-	0x85,0x76,0x86,0xA1,0x94,0x75,0xD3,0xBF,0xA2,0xFF};
-static const EC_CURVE_DATA _EC_X9_62_PRIME_239V3 = {
-	NID_X9_62_prime_field,
-	"7FFFFFFFFFFFFFFFFFFFFFFF7FFFFFFFFFFF8000000000007FFFFFFFFFFF",
-	"7FFFFFFFFFFFFFFFFFFFFFFF7FFFFFFFFFFF8000000000007FFFFFFFFFFC",
-	"255705FA2A306654B1F4CB03D6A750A30C250102D4988717D9BA15AB6D3E",
-	"6768AE8E18BB92CFCF005C949AA2C6D94853D0E660BBF854B1C9505FE95A",
-	"1607e6898f390c06bc1d552bad226f3b6fcfe48b6e818499af18e3ed6cf3",
-	"7FFFFFFFFFFFFFFFFFFFFFFF7FFFFF975DEB41B3A6057C3C432146526551",1,
-	_EC_X9_62_PRIME_239V3_SEED, 20,
-	"X9.62 curve over a 239 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+30*6]; }
+	_EC_X9_62_PRIME_239V3 = {
+	{ NID_X9_62_prime_field,20,30,1 },
+	{ 0x7D,0x73,0x74,0x16,0x8F,0xFE,0x34,0x71,0xB6,0x0A,	/* seed */
+	  0x85,0x76,0x86,0xA1,0x94,0x75,0xD3,0xBF,0xA2,0xFF,
+
+	  0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0x80,0x00,
+	  0x00,0x00,0x00,0x00,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,
+
+	  0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0x80,0x00,
+	  0x00,0x00,0x00,0x00,0x7F,0xFF,0xFF,0xFF,0xFF,0xFC,
+
+	  0x25,0x57,0x05,0xFA,0x2A,0x30,0x66,0x54,0xB1,0xF4,	/* b */
+	  0xCB,0x03,0xD6,0xA7,0x50,0xA3,0x0C,0x25,0x01,0x02,
+	  0xD4,0x98,0x87,0x17,0xD9,0xBA,0x15,0xAB,0x6D,0x3E,
+
+	  0x67,0x68,0xAE,0x8E,0x18,0xBB,0x92,0xCF,0xCF,0x00,	/* x */
+	  0x5C,0x94,0x9A,0xA2,0xC6,0xD9,0x48,0x53,0xD0,0xE6,
+	  0x60,0xBB,0xF8,0x54,0xB1,0xC9,0x50,0x5F,0xE9,0x5A,
+
+	  0x16,0x07,0xe6,0x89,0x8f,0x39,0x0c,0x06,0xbc,0x1d,	/* y */
+	  0x55,0x2b,0xad,0x22,0x6f,0x3b,0x6f,0xcf,0xe4,0x8b,
+	  0x6e,0x81,0x84,0x99,0xaf,0x18,0xe3,0xed,0x6c,0xf3,
+
+	  0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0x7F,0xFF,0xFF,0x97,0x5D,0xEB,0x41,0xB3,
+	  0xA6,0x05,0x7C,0x3C,0x43,0x21,0x46,0x52,0x65,0x51 }
 	};
 
-static const unsigned char _EC_X9_62_PRIME_256V1_SEED[] = {
-	0xC4,0x9D,0x36,0x08,0x86,0xE7,0x04,0x93,0x6A,0x66,
-	0x78,0xE1,0x13,0x9D,0x26,0xB7,0x81,0x9F,0x7E,0x90};
-static const EC_CURVE_DATA _EC_X9_62_PRIME_256V1 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFF00000001000000000000000000000000FFFFFFFFFFFFFFFFFFFFFFFF",
-	"FFFFFFFF00000001000000000000000000000000FFFFFFFFFFFFFFFFFFFFFFFC",
-	"5AC635D8AA3A93E7B3EBBD55769886BC651D06B0CC53B0F63BCE3C3E27D2604B",
-	"6B17D1F2E12C4247F8BCE6E563A440F277037D812DEB33A0F4A13945D898C296",
-	"4fe342e2fe1a7f9b8ee7eb4a7c0f9e162bce33576b315ececbb6406837bf51f5",
-	"FFFFFFFF00000000FFFFFFFFFFFFFFFFBCE6FAADA7179E84F3B9CAC2FC632551",1,
-	_EC_X9_62_PRIME_256V1_SEED, 20,
-	"X9.62/SECG curve over a 256 bit prime field"
+
+static const struct { EC_CURVE_DATA h; unsigned char data[20+32*6]; }
+	_EC_X9_62_PRIME_256V1 = {
+	{ NID_X9_62_prime_field,20,32,1 },
+	{ 0xC4,0x9D,0x36,0x08,0x86,0xE7,0x04,0x93,0x6A,0x66,	/* seed */
+	  0x78,0xE1,0x13,0x9D,0x26,0xB7,0x81,0x9F,0x7E,0x90,
+
+	  0xFF,0xFF,0xFF,0xFF,0x00,0x00,0x00,0x01,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0x00,0x00,0x00,0x01,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFC,
+	  0x5A,0xC6,0x35,0xD8,0xAA,0x3A,0x93,0xE7,0xB3,0xEB,	/* b */
+	  0xBD,0x55,0x76,0x98,0x86,0xBC,0x65,0x1D,0x06,0xB0,
+	  0xCC,0x53,0xB0,0xF6,0x3B,0xCE,0x3C,0x3E,0x27,0xD2,
+	  0x60,0x4B,
+	  0x6B,0x17,0xD1,0xF2,0xE1,0x2C,0x42,0x47,0xF8,0xBC,	/* x */
+	  0xE6,0xE5,0x63,0xA4,0x40,0xF2,0x77,0x03,0x7D,0x81,
+	  0x2D,0xEB,0x33,0xA0,0xF4,0xA1,0x39,0x45,0xD8,0x98,
+	  0xC2,0x96,
+	  0x4f,0xe3,0x42,0xe2,0xfe,0x1a,0x7f,0x9b,0x8e,0xe7,	/* y */
+	  0xeb,0x4a,0x7c,0x0f,0x9e,0x16,0x2b,0xce,0x33,0x57,
+	  0x6b,0x31,0x5e,0xce,0xcb,0xb6,0x40,0x68,0x37,0xbf,
+	  0x51,0xf5,
+	  0xFF,0xFF,0xFF,0xFF,0x00,0x00,0x00,0x00,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xBC,0xE6,0xFA,0xAD,
+	  0xA7,0x17,0x9E,0x84,0xF3,0xB9,0xCA,0xC2,0xFC,0x63,
+	  0x25,0x51 }
 	};
+
 /* the secg prime curves (minus the nist and x9.62 prime curves) */
-static const unsigned char _EC_SECG_PRIME_112R1_SEED[] = {
-	0x00,0xF5,0x0B,0x02,0x8E,0x4D,0x69,0x6E,0x67,0x68,
-	0x75,0x61,0x51,0x75,0x29,0x04,0x72,0x78,0x3F,0xB1};
-static const EC_CURVE_DATA _EC_SECG_PRIME_112R1 = {
-	NID_X9_62_prime_field,
-	"DB7C2ABF62E35E668076BEAD208B",
-	"DB7C2ABF62E35E668076BEAD2088",
-	"659EF8BA043916EEDE8911702B22",
-	"09487239995A5EE76B55F9C2F098",
-	"a89ce5af8724c0a23e0e0ff77500",
-	"DB7C2ABF62E35E7628DFAC6561C5",1,
-	_EC_SECG_PRIME_112R1_SEED, 20,
-	"SECG/WTLS curve over a 112 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+14*6]; }
+	_EC_SECG_PRIME_112R1 = {
+	{ NID_X9_62_prime_field,20,14,1 },
+	{ 0x00,0xF5,0x0B,0x02,0x8E,0x4D,0x69,0x6E,0x67,0x68,	/* seed */
+	  0x75,0x61,0x51,0x75,0x29,0x04,0x72,0x78,0x3F,0xB1,
+
+	  0xDB,0x7C,0x2A,0xBF,0x62,0xE3,0x5E,0x66,0x80,0x76,	/* p */
+	  0xBE,0xAD,0x20,0x8B,
+	  0xDB,0x7C,0x2A,0xBF,0x62,0xE3,0x5E,0x66,0x80,0x76,	/* a */
+	  0xBE,0xAD,0x20,0x88,
+	  0x65,0x9E,0xF8,0xBA,0x04,0x39,0x16,0xEE,0xDE,0x89,	/* b */
+	  0x11,0x70,0x2B,0x22,
+	  0x09,0x48,0x72,0x39,0x99,0x5A,0x5E,0xE7,0x6B,0x55,	/* x */
+	  0xF9,0xC2,0xF0,0x98,
+	  0xa8,0x9c,0xe5,0xaf,0x87,0x24,0xc0,0xa2,0x3e,0x0e,	/* y */
+	  0x0f,0xf7,0x75,0x00,
+	  0xDB,0x7C,0x2A,0xBF,0x62,0xE3,0x5E,0x76,0x28,0xDF,	/* order */
+	  0xAC,0x65,0x61,0xC5 }
 	};
 
-static const unsigned char _EC_SECG_PRIME_112R2_SEED[] = {
-	0x00,0x27,0x57,0xA1,0x11,0x4D,0x69,0x6E,0x67,0x68,
-	0x75,0x61,0x51,0x75,0x53,0x16,0xC0,0x5E,0x0B,0xD4};
-static const EC_CURVE_DATA _EC_SECG_PRIME_112R2 = {
-	NID_X9_62_prime_field,
-	"DB7C2ABF62E35E668076BEAD208B",
-	"6127C24C05F38A0AAAF65C0EF02C",
-	"51DEF1815DB5ED74FCC34C85D709",
-	"4BA30AB5E892B4E1649DD0928643",
-	"adcd46f5882e3747def36e956e97",
-	"36DF0AAFD8B8D7597CA10520D04B",4, 
-	_EC_SECG_PRIME_112R2_SEED, 20,
-	"SECG curve over a 112 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+14*6]; }
+	_EC_SECG_PRIME_112R2 = {
+	{ NID_X9_62_prime_field,20,14,4 },
+	{ 0x00,0x27,0x57,0xA1,0x11,0x4D,0x69,0x6E,0x67,0x68,	/* seed */
+	  0x75,0x61,0x51,0x75,0x53,0x16,0xC0,0x5E,0x0B,0xD4,
+
+	  0xDB,0x7C,0x2A,0xBF,0x62,0xE3,0x5E,0x66,0x80,0x76,	/* p */
+	  0xBE,0xAD,0x20,0x8B,
+	  0x61,0x27,0xC2,0x4C,0x05,0xF3,0x8A,0x0A,0xAA,0xF6,	/* a */
+	  0x5C,0x0E,0xF0,0x2C,
+	  0x51,0xDE,0xF1,0x81,0x5D,0xB5,0xED,0x74,0xFC,0xC3,	/* b */
+	  0x4C,0x85,0xD7,0x09,
+	  0x4B,0xA3,0x0A,0xB5,0xE8,0x92,0xB4,0xE1,0x64,0x9D,	/* x */
+	  0xD0,0x92,0x86,0x43,
+	  0xad,0xcd,0x46,0xf5,0x88,0x2e,0x37,0x47,0xde,0xf3,	/* y */
+	  0x6e,0x95,0x6e,0x97,
+	  0x36,0xDF,0x0A,0xAF,0xD8,0xB8,0xD7,0x59,0x7C,0xA1,	/* order */
+	  0x05,0x20,0xD0,0x4B }
 	};
 
-static const unsigned char _EC_SECG_PRIME_128R1_SEED[] = {
-	0x00,0x0E,0x0D,0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,
-	0x51,0x75,0x0C,0xC0,0x3A,0x44,0x73,0xD0,0x36,0x79};
-static const EC_CURVE_DATA _EC_SECG_PRIME_128R1 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFDFFFFFFFFFFFFFFFFFFFFFFFF",
-	"FFFFFFFDFFFFFFFFFFFFFFFFFFFFFFFC",
-	"E87579C11079F43DD824993C2CEE5ED3",
-	"161FF7528B899B2D0C28607CA52C5B86",
-	"cf5ac8395bafeb13c02da292dded7a83",
-	"FFFFFFFE0000000075A30D1B9038A115",1,
-	_EC_SECG_PRIME_128R1_SEED, 20,
-	"SECG curve over a 128 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+16*6]; }
+	_EC_SECG_PRIME_128R1 = {
+	{ NID_X9_62_prime_field,20,16,1 },
+	{ 0x00,0x0E,0x0D,0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,	/* seed */
+	  0x51,0x75,0x0C,0xC0,0x3A,0x44,0x73,0xD0,0x36,0x79,
+
+	  0xFF,0xFF,0xFF,0xFD,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFD,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFC,
+	  0xE8,0x75,0x79,0xC1,0x10,0x79,0xF4,0x3D,0xD8,0x24,	/* b */
+	  0x99,0x3C,0x2C,0xEE,0x5E,0xD3,
+	  0x16,0x1F,0xF7,0x52,0x8B,0x89,0x9B,0x2D,0x0C,0x28,	/* x */
+	  0x60,0x7C,0xA5,0x2C,0x5B,0x86,
+	  0xcf,0x5a,0xc8,0x39,0x5b,0xaf,0xeb,0x13,0xc0,0x2d,	/* y */
+	  0xa2,0x92,0xdd,0xed,0x7a,0x83,
+	  0xFF,0xFF,0xFF,0xFE,0x00,0x00,0x00,0x00,0x75,0xA3,	/* order */
+	  0x0D,0x1B,0x90,0x38,0xA1,0x15 }
 	};
 
-static const unsigned char _EC_SECG_PRIME_128R2_SEED[] = {
-	0x00,0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,0x51,0x75,
-	0x12,0xD8,0xF0,0x34,0x31,0xFC,0xE6,0x3B,0x88,0xF4};
-static const EC_CURVE_DATA _EC_SECG_PRIME_128R2 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFDFFFFFFFFFFFFFFFFFFFFFFFF",
-	"D6031998D1B3BBFEBF59CC9BBFF9AEE1",
-	"5EEEFCA380D02919DC2C6558BB6D8A5D",
-	"7B6AA5D85E572983E6FB32A7CDEBC140",
-	"27b6916a894d3aee7106fe805fc34b44",
-	"3FFFFFFF7FFFFFFFBE0024720613B5A3",4,
-	_EC_SECG_PRIME_128R2_SEED, 20,
-	"SECG curve over a 128 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+16*6]; }
+	_EC_SECG_PRIME_128R2 = {
+	{ NID_X9_62_prime_field,20,16,4 },
+	{ 0x00,0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,0x51,0x75,	/* seed */
+	  0x12,0xD8,0xF0,0x34,0x31,0xFC,0xE6,0x3B,0x88,0xF4,
+
+	  0xFF,0xFF,0xFF,0xFD,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xD6,0x03,0x19,0x98,0xD1,0xB3,0xBB,0xFE,0xBF,0x59,	/* a */
+	  0xCC,0x9B,0xBF,0xF9,0xAE,0xE1,
+	  0x5E,0xEE,0xFC,0xA3,0x80,0xD0,0x29,0x19,0xDC,0x2C,	/* b */
+	  0x65,0x58,0xBB,0x6D,0x8A,0x5D,
+	  0x7B,0x6A,0xA5,0xD8,0x5E,0x57,0x29,0x83,0xE6,0xFB,	/* x */
+	  0x32,0xA7,0xCD,0xEB,0xC1,0x40,
+	  0x27,0xb6,0x91,0x6a,0x89,0x4d,0x3a,0xee,0x71,0x06,	/* y */
+	  0xfe,0x80,0x5f,0xc3,0x4b,0x44,
+	  0x3F,0xFF,0xFF,0xFF,0x7F,0xFF,0xFF,0xFF,0xBE,0x00,	/* order */
+	  0x24,0x72,0x06,0x13,0xB5,0xA3 }
 	};
 
-static const EC_CURVE_DATA _EC_SECG_PRIME_160K1 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFAC73",
-	"0",
-	"7",
-	"3B4C382CE37AA192A4019E763036F4F5DD4D7EBB",
-	"938cf935318fdced6bc28286531733c3f03c4fee",
-	"0100000000000000000001B8FA16DFAB9ACA16B6B3",1,
-	NULL, 0,
-	"SECG curve over a 160 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+21*6]; }
+	_EC_SECG_PRIME_160K1 = {
+	{ NID_X9_62_prime_field,0,21,1 },
+	{							/* no seed */
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xAC,
+	  0x73,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x07,
+	  0x00,0x3B,0x4C,0x38,0x2C,0xE3,0x7A,0xA1,0x92,0xA4,	/* x */
+	  0x01,0x9E,0x76,0x30,0x36,0xF4,0xF5,0xDD,0x4D,0x7E,
+	  0xBB,
+	  0x00,0x93,0x8c,0xf9,0x35,0x31,0x8f,0xdc,0xed,0x6b,	/* y */
+	  0xc2,0x82,0x86,0x53,0x17,0x33,0xc3,0xf0,0x3c,0x4f,
+	  0xee,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x01,0xB8,0xFA,0x16,0xDF,0xAB,0x9A,0xCA,0x16,0xB6,
+	  0xB3 }
 	};
 
-static const unsigned char _EC_SECG_PRIME_160R1_SEED[] = {
-	0x10,0x53,0xCD,0xE4,0x2C,0x14,0xD6,0x96,0xE6,0x76,
-	0x87,0x56,0x15,0x17,0x53,0x3B,0xF3,0xF8,0x33,0x45};
-static const EC_CURVE_DATA _EC_SECG_PRIME_160R1 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF7FFFFFFF",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF7FFFFFFC",
-	"1C97BEFC54BD7A8B65ACF89F81D4D4ADC565FA45",
-	"4A96B5688EF573284664698968C38BB913CBFC82",
-	"23a628553168947d59dcc912042351377ac5fb32",
-	"0100000000000000000001F4C8F927AED3CA752257",1,
-	_EC_SECG_PRIME_160R1_SEED, 20,
-	"SECG curve over a 160 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+21*6]; }
+	_EC_SECG_PRIME_160R1 = {
+	{ NID_X9_62_prime_field,20,21,1 },
+	{ 0x10,0x53,0xCD,0xE4,0x2C,0x14,0xD6,0x96,0xE6,0x76,	/* seed */
+	  0x87,0x56,0x15,0x17,0x53,0x3B,0xF3,0xF8,0x33,0x45,
+
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0x7F,0xFF,0xFF,
+	  0xFF,
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0x7F,0xFF,0xFF,
+	  0xFC,
+	  0x00,0x1C,0x97,0xBE,0xFC,0x54,0xBD,0x7A,0x8B,0x65,	/* b */
+	  0xAC,0xF8,0x9F,0x81,0xD4,0xD4,0xAD,0xC5,0x65,0xFA,
+	  0x45,
+	  0x00,0x4A,0x96,0xB5,0x68,0x8E,0xF5,0x73,0x28,0x46,	/* x */
+	  0x64,0x69,0x89,0x68,0xC3,0x8B,0xB9,0x13,0xCB,0xFC,
+	  0x82,
+	  0x00,0x23,0xa6,0x28,0x55,0x31,0x68,0x94,0x7d,0x59,	/* y */
+	  0xdc,0xc9,0x12,0x04,0x23,0x51,0x37,0x7a,0xc5,0xfb,
+	  0x32,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x01,0xF4,0xC8,0xF9,0x27,0xAE,0xD3,0xCA,0x75,0x22,
+	  0x57 }
 	};
 
-static const unsigned char _EC_SECG_PRIME_160R2_SEED[] = {
-	0xB9,0x9B,0x99,0xB0,0x99,0xB3,0x23,0xE0,0x27,0x09,
-	0xA4,0xD6,0x96,0xE6,0x76,0x87,0x56,0x15,0x17,0x51};
-static const EC_CURVE_DATA _EC_SECG_PRIME_160R2 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFAC73",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFAC70",
-	"B4E134D3FB59EB8BAB57274904664D5AF50388BA",
-	"52DCB034293A117E1F4FF11B30F7199D3144CE6D",
-	"feaffef2e331f296e071fa0df9982cfea7d43f2e",
-	"0100000000000000000000351EE786A818F3A1A16B",1,
-	_EC_SECG_PRIME_160R2_SEED, 20,
-	"SECG/WTLS curve over a 160 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+21*6]; }
+	_EC_SECG_PRIME_160R2 = {
+	{ NID_X9_62_prime_field,20,21,1 },
+	{ 0xB9,0x9B,0x99,0xB0,0x99,0xB3,0x23,0xE0,0x27,0x09,	/* seed */
+	  0xA4,0xD6,0x96,0xE6,0x76,0x87,0x56,0x15,0x17,0x51,
+
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xAC,
+	  0x73,
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xAC,
+	  0x70,
+	  0x00,0xB4,0xE1,0x34,0xD3,0xFB,0x59,0xEB,0x8B,0xAB,	/* b */
+	  0x57,0x27,0x49,0x04,0x66,0x4D,0x5A,0xF5,0x03,0x88,
+	  0xBA,
+	  0x00,0x52,0xDC,0xB0,0x34,0x29,0x3A,0x11,0x7E,0x1F,	/* x */
+	  0x4F,0xF1,0x1B,0x30,0xF7,0x19,0x9D,0x31,0x44,0xCE,
+	  0x6D,
+	  0x00,0xfe,0xaf,0xfe,0xf2,0xe3,0x31,0xf2,0x96,0xe0,	/* y */
+	  0x71,0xfa,0x0d,0xf9,0x98,0x2c,0xfe,0xa7,0xd4,0x3f,
+	  0x2e,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x35,0x1E,0xE7,0x86,0xA8,0x18,0xF3,0xA1,0xA1,
+	  0x6B }
 	};
 
-static const EC_CURVE_DATA _EC_SECG_PRIME_192K1 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFEE37",
-	"0",
-	"3",
-	"DB4FF10EC057E9AE26B07D0280B7F4341DA5D1B1EAE06C7D",
-	"9b2f2f6d9c5628a7844163d015be86344082aa88d95e2f9d",
-	"FFFFFFFFFFFFFFFFFFFFFFFE26F2FC170F69466A74DEFD8D",1,
-	NULL, 20,
-	"SECG curve over a 192 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+24*6]; }
+	_EC_SECG_PRIME_192K1 = {
+	{ NID_X9_62_prime_field,0,24,1 },
+	{							/* no seed */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,
+	  0xFF,0xFF,0xEE,0x37,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x03,
+	  0xDB,0x4F,0xF1,0x0E,0xC0,0x57,0xE9,0xAE,0x26,0xB0,	/* x */
+	  0x7D,0x02,0x80,0xB7,0xF4,0x34,0x1D,0xA5,0xD1,0xB1,
+	  0xEA,0xE0,0x6C,0x7D,
+	  0x9b,0x2f,0x2f,0x6d,0x9c,0x56,0x28,0xa7,0x84,0x41,	/* y */
+	  0x63,0xd0,0x15,0xbe,0x86,0x34,0x40,0x82,0xaa,0x88,
+	  0xd9,0x5e,0x2f,0x9d,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFE,0x26,0xF2,0xFC,0x17,0x0F,0x69,0x46,0x6A,
+	  0x74,0xDE,0xFD,0x8D }
 	};
 
-static const EC_CURVE_DATA _EC_SECG_PRIME_224K1 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFE56D",
-	"0",
-	"5",
-	"A1455B334DF099DF30FC28A169A467E9E47075A90F7E650EB6B7A45C",
-	"7e089fed7fba344282cafbd6f7e319f7c0b0bd59e2ca4bdb556d61a5",
-	"010000000000000000000000000001DCE8D2EC6184CAF0A971769FB1F7",1,
-	NULL, 20,
-	"SECG curve over a 224 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+29*6]; }
+	_EC_SECG_PRIME_224K1 = {
+	{ NID_X9_62_prime_field,0,29,1 },
+	{							/* no seed */
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xE5,0x6D,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x05,
+	  0x00,0xA1,0x45,0x5B,0x33,0x4D,0xF0,0x99,0xDF,0x30,	/* x */
+	  0xFC,0x28,0xA1,0x69,0xA4,0x67,0xE9,0xE4,0x70,0x75,
+	  0xA9,0x0F,0x7E,0x65,0x0E,0xB6,0xB7,0xA4,0x5C,
+	  0x00,0x7e,0x08,0x9f,0xed,0x7f,0xba,0x34,0x42,0x82,	/* y */
+	  0xca,0xfb,0xd6,0xf7,0xe3,0x19,0xf7,0xc0,0xb0,0xbd,
+	  0x59,0xe2,0xca,0x4b,0xdb,0x55,0x6d,0x61,0xa5,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x00,0x00,0x01,0xDC,0xE8,0xD2,0xEC,0x61,
+	  0x84,0xCA,0xF0,0xA9,0x71,0x76,0x9F,0xB1,0xF7 }
 	};
 
-static const EC_CURVE_DATA _EC_SECG_PRIME_256K1 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFC2F",
-	"0",
-	"7",
-	"79BE667EF9DCBBAC55A06295CE870B07029BFCDB2DCE28D959F2815B16F81798",
-	"483ada7726a3c4655da4fbfc0e1108a8fd17b448a68554199c47d08ffb10d4b8",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEBAAEDCE6AF48A03BBFD25E8CD0364141",1,
-	NULL, 20,
-	"SECG curve over a 256 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+32*6]; }
+	_EC_SECG_PRIME_256K1 = {
+	{ NID_X9_62_prime_field,0,32,1 },
+	{							/* no seed */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,
+	  0xFC,0x2F,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x07,
+	  0x79,0xBE,0x66,0x7E,0xF9,0xDC,0xBB,0xAC,0x55,0xA0,	/* x */
+	  0x62,0x95,0xCE,0x87,0x0B,0x07,0x02,0x9B,0xFC,0xDB,
+	  0x2D,0xCE,0x28,0xD9,0x59,0xF2,0x81,0x5B,0x16,0xF8,
+	  0x17,0x98,
+	  0x48,0x3a,0xda,0x77,0x26,0xa3,0xc4,0x65,0x5d,0xa4,	/* y */
+	  0xfb,0xfc,0x0e,0x11,0x08,0xa8,0xfd,0x17,0xb4,0x48,
+	  0xa6,0x85,0x54,0x19,0x9c,0x47,0xd0,0x8f,0xfb,0x10,
+	  0xd4,0xb8,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xBA,0xAE,0xDC,0xE6,
+	  0xAF,0x48,0xA0,0x3B,0xBF,0xD2,0x5E,0x8C,0xD0,0x36,
+	  0x41,0x41 }
 	};
 
 /* some wap/wtls curves */
-static const EC_CURVE_DATA _EC_WTLS_8 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFDE7",
-	"0",
-	"3",
-	"1",
-	"2",
-	"0100000000000001ECEA551AD837E9",1,
-	NULL, 20,
-	"WTLS curve over a 112 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+15*6]; }
+	_EC_WTLS_8 = {
+	{ NID_X9_62_prime_field,0,15,1 },
+	{							/* no seed */
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFD,0xE7,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x03,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* x */
+	  0x00,0x00,0x00,0x00,0x01,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* y */
+	  0x00,0x00,0x00,0x00,0x02,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x01,0xEC,0xEA,	/* order */
+	  0x55,0x1A,0xD8,0x37,0xE9 }
 	};
 
-static const EC_CURVE_DATA _EC_WTLS_9 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFC808F",
-	"0",
-	"3",
-	"1",
-	"2",
-	"0100000000000000000001CDC98AE0E2DE574ABF33",1,
-	NULL, 20,
-	"WTLS curve over a 160 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+21*6]; }
+	_EC_WTLS_9 = {
+	{ NID_X9_62_prime_field,0,21,1 },
+	{							/* no seed */
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFC,0x80,
+	  0x8F,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x03,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* x */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x01,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* y */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x02,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x01,0xCD,0xC9,0x8A,0xE0,0xE2,0xDE,0x57,0x4A,0xBF,
+	  0x33 }
 	};
 
-static const EC_CURVE_DATA _EC_WTLS_12 = {
-	NID_X9_62_prime_field,
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF000000000000000000000001",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFFFFFFFFFFFFFFFFFFFE",
-	"B4050A850C04B3ABF54132565044B0B7D7BFD8BA270B39432355FFB4",
-	"B70E0CBD6BB4BF7F321390B94A03C1D356C21122343280D6115C1D21",
-	"bd376388b5f723fb4c22dfe6cd4375a05a07476444d5819985007e34",
-	"FFFFFFFFFFFFFFFFFFFFFFFFFFFF16A2E0B8F03E13DD29455C5C2A3D", 1,
-	NULL, 0,
-	"WTLS curvs over a 224 bit prime field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+28*6]; }
+	_EC_WTLS_12 = {
+	{ NID_X9_62_prime_field,0,28,1 },
+	{							/* no seed */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* p */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* a */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,
+	  0xB4,0x05,0x0A,0x85,0x0C,0x04,0xB3,0xAB,0xF5,0x41,	/* b */
+	  0x32,0x56,0x50,0x44,0xB0,0xB7,0xD7,0xBF,0xD8,0xBA,
+	  0x27,0x0B,0x39,0x43,0x23,0x55,0xFF,0xB4,
+	  0xB7,0x0E,0x0C,0xBD,0x6B,0xB4,0xBF,0x7F,0x32,0x13,	/* x */
+	  0x90,0xB9,0x4A,0x03,0xC1,0xD3,0x56,0xC2,0x11,0x22,
+	  0x34,0x32,0x80,0xD6,0x11,0x5C,0x1D,0x21,
+	  0xbd,0x37,0x63,0x88,0xb5,0xf7,0x23,0xfb,0x4c,0x22,	/* y */
+	  0xdf,0xe6,0xcd,0x43,0x75,0xa0,0x5a,0x07,0x47,0x64,
+	  0x44,0xd5,0x81,0x99,0x85,0x00,0x7e,0x34,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0x16,0xA2,0xE0,0xB8,0xF0,0x3E,
+	  0x13,0xDD,0x29,0x45,0x5C,0x5C,0x2A,0x3D }
 	};
 
+#ifndef OPENSSL_NO_EC2M
+
 /* characteristic two curves */
-static const unsigned char _EC_SECG_CHAR2_113R1_SEED[] = {
-	0x10,0xE7,0x23,0xAB,0x14,0xD6,0x96,0xE6,0x76,0x87,
-	0x56,0x15,0x17,0x56,0xFE,0xBF,0x8F,0xCB,0x49,0xA9};
-static const EC_CURVE_DATA _EC_SECG_CHAR2_113R1 = {
-	NID_X9_62_characteristic_two_field,
-	"020000000000000000000000000201",
-	"003088250CA6E7C7FE649CE85820F7",
-	"00E8BEE4D3E2260744188BE0E9C723",
-	"009D73616F35F4AB1407D73562C10F",
-	"00A52830277958EE84D1315ED31886",
-	"0100000000000000D9CCEC8A39E56F", 2,
-	_EC_SECG_CHAR2_113R1_SEED, 20,
-	"SECG curve over a 113 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+15*6]; }
+	_EC_SECG_CHAR2_113R1 = {
+	{ NID_X9_62_characteristic_two_field,20,15,2 },
+	{ 0x10,0xE7,0x23,0xAB,0x14,0xD6,0x96,0xE6,0x76,0x87,	/* seed */
+	  0x56,0x15,0x17,0x56,0xFE,0xBF,0x8F,0xCB,0x49,0xA9,
+
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x02,0x01,
+	  0x00,0x30,0x88,0x25,0x0C,0xA6,0xE7,0xC7,0xFE,0x64,	/* a */
+	  0x9C,0xE8,0x58,0x20,0xF7,
+	  0x00,0xE8,0xBE,0xE4,0xD3,0xE2,0x26,0x07,0x44,0x18,	/* b */
+	  0x8B,0xE0,0xE9,0xC7,0x23,
+	  0x00,0x9D,0x73,0x61,0x6F,0x35,0xF4,0xAB,0x14,0x07,	/* x */
+	  0xD7,0x35,0x62,0xC1,0x0F,
+	  0x00,0xA5,0x28,0x30,0x27,0x79,0x58,0xEE,0x84,0xD1,	/* y */
+	  0x31,0x5E,0xD3,0x18,0x86,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xD9,0xCC,	/* order */
+	  0xEC,0x8A,0x39,0xE5,0x6F }
 	};
 
-static const unsigned char _EC_SECG_CHAR2_113R2_SEED[] = {
-	0x10,0xC0,0xFB,0x15,0x76,0x08,0x60,0xDE,0xF1,0xEE,
-	0xF4,0xD6,0x96,0xE6,0x76,0x87,0x56,0x15,0x17,0x5D};
-static const EC_CURVE_DATA _EC_SECG_CHAR2_113R2 = {
-	NID_X9_62_characteristic_two_field,
-	"020000000000000000000000000201",
-	"00689918DBEC7E5A0DD6DFC0AA55C7",
-	"0095E9A9EC9B297BD4BF36E059184F",
-	"01A57A6A7B26CA5EF52FCDB8164797",
-	"00B3ADC94ED1FE674C06E695BABA1D",
-	"010000000000000108789B2496AF93", 2,
-	_EC_SECG_CHAR2_113R2_SEED, 20,
-	"SECG curve over a 113 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+15*6]; }
+	_EC_SECG_CHAR2_113R2 = {
+	{ NID_X9_62_characteristic_two_field,20,15,2 },
+	{ 0x10,0xC0,0xFB,0x15,0x76,0x08,0x60,0xDE,0xF1,0xEE,	/* seed */
+	  0xF4,0xD6,0x96,0xE6,0x76,0x87,0x56,0x15,0x17,0x5D,
+
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x02,0x01,
+	  0x00,0x68,0x99,0x18,0xDB,0xEC,0x7E,0x5A,0x0D,0xD6,	/* a */
+	  0xDF,0xC0,0xAA,0x55,0xC7,
+	  0x00,0x95,0xE9,0xA9,0xEC,0x9B,0x29,0x7B,0xD4,0xBF,	/* b */
+	  0x36,0xE0,0x59,0x18,0x4F,
+	  0x01,0xA5,0x7A,0x6A,0x7B,0x26,0xCA,0x5E,0xF5,0x2F,	/* x */
+	  0xCD,0xB8,0x16,0x47,0x97,
+	  0x00,0xB3,0xAD,0xC9,0x4E,0xD1,0xFE,0x67,0x4C,0x06,	/* y */
+	  0xE6,0x95,0xBA,0xBA,0x1D,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x01,0x08,0x78,	/* order */
+	  0x9B,0x24,0x96,0xAF,0x93 }
 	};
 
-static const unsigned char _EC_SECG_CHAR2_131R1_SEED[] = {
-	0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,0x51,0x75,0x98,
-	0x5B,0xD3,0xAD,0xBA,0xDA,0x21,0xB4,0x3A,0x97,0xE2};
-static const EC_CURVE_DATA _EC_SECG_CHAR2_131R1 = {
-	NID_X9_62_characteristic_two_field,
-	"080000000000000000000000000000010D",
-	"07A11B09A76B562144418FF3FF8C2570B8",
-	"0217C05610884B63B9C6C7291678F9D341",
-	"0081BAF91FDF9833C40F9C181343638399",
-	"078C6E7EA38C001F73C8134B1B4EF9E150",
-	"0400000000000000023123953A9464B54D", 2,
-	_EC_SECG_CHAR2_131R1_SEED, 20,
-	"SECG/WTLS curve over a 131 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+17*6]; }
+	_EC_SECG_CHAR2_131R1 = {
+	{ NID_X9_62_characteristic_two_field,20,17,2 },
+	{ 0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,0x51,0x75,0x98,	/* seed */
+	  0x5B,0xD3,0xAD,0xBA,0xDA,0x21,0xB4,0x3A,0x97,0xE2,
+
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x01,0x0D,
+	  0x07,0xA1,0x1B,0x09,0xA7,0x6B,0x56,0x21,0x44,0x41,	/* a */
+	  0x8F,0xF3,0xFF,0x8C,0x25,0x70,0xB8,
+	  0x02,0x17,0xC0,0x56,0x10,0x88,0x4B,0x63,0xB9,0xC6,	/* b */
+	  0xC7,0x29,0x16,0x78,0xF9,0xD3,0x41,
+	  0x00,0x81,0xBA,0xF9,0x1F,0xDF,0x98,0x33,0xC4,0x0F,	/* x */
+	  0x9C,0x18,0x13,0x43,0x63,0x83,0x99,
+	  0x07,0x8C,0x6E,0x7E,0xA3,0x8C,0x00,0x1F,0x73,0xC8,	/* y */
+	  0x13,0x4B,0x1B,0x4E,0xF9,0xE1,0x50,
+	  0x04,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x02,0x31,	/* order */
+	  0x23,0x95,0x3A,0x94,0x64,0xB5,0x4D }
 	};
 
-static const unsigned char _EC_SECG_CHAR2_131R2_SEED[] = {
-	0x98,0x5B,0xD3,0xAD,0xBA,0xD4,0xD6,0x96,0xE6,0x76,
-	0x87,0x56,0x15,0x17,0x5A,0x21,0xB4,0x3A,0x97,0xE3};
-static const EC_CURVE_DATA _EC_SECG_CHAR2_131R2 = {
-	NID_X9_62_characteristic_two_field,
-	"080000000000000000000000000000010D",
-	"03E5A88919D7CAFCBF415F07C2176573B2",
-	"04B8266A46C55657AC734CE38F018F2192",
-	"0356DCD8F2F95031AD652D23951BB366A8",
-	"0648F06D867940A5366D9E265DE9EB240F",
-	"0400000000000000016954A233049BA98F", 2,
-	_EC_SECG_CHAR2_131R2_SEED, 20,
-	"SECG curve over a 131 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+17*6]; }
+	_EC_SECG_CHAR2_131R2 = {
+	{ NID_X9_62_characteristic_two_field,20,17,2 },
+	{ 0x98,0x5B,0xD3,0xAD,0xBA,0xD4,0xD6,0x96,0xE6,0x76,	/* seed */
+	  0x87,0x56,0x15,0x17,0x5A,0x21,0xB4,0x3A,0x97,0xE3,
+
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x01,0x0D,
+	  0x03,0xE5,0xA8,0x89,0x19,0xD7,0xCA,0xFC,0xBF,0x41,	/* a */
+	  0x5F,0x07,0xC2,0x17,0x65,0x73,0xB2,
+	  0x04,0xB8,0x26,0x6A,0x46,0xC5,0x56,0x57,0xAC,0x73,	/* b */
+	  0x4C,0xE3,0x8F,0x01,0x8F,0x21,0x92,
+	  0x03,0x56,0xDC,0xD8,0xF2,0xF9,0x50,0x31,0xAD,0x65,	/* x */
+	  0x2D,0x23,0x95,0x1B,0xB3,0x66,0xA8,
+	  0x06,0x48,0xF0,0x6D,0x86,0x79,0x40,0xA5,0x36,0x6D,	/* y */
+	  0x9E,0x26,0x5D,0xE9,0xEB,0x24,0x0F,
+	  0x04,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,0x69,	/* order */
+	  0x54,0xA2,0x33,0x04,0x9B,0xA9,0x8F }
 	};
 
-static const EC_CURVE_DATA _EC_NIST_CHAR2_163K = {
-	NID_X9_62_characteristic_two_field,
-	"0800000000000000000000000000000000000000C9",
-	"1",
-	"1",
-	"02FE13C0537BBC11ACAA07D793DE4E6D5E5C94EEE8",
-	"0289070FB05D38FF58321F2E800536D538CCDAA3D9",
-	"04000000000000000000020108A2E0CC0D99F8A5EF", 2,
-	NULL, 0,
-	"NIST/SECG/WTLS curve over a 163 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+21*6]; }
+	_EC_NIST_CHAR2_163K = {
+	{ NID_X9_62_characteristic_two_field,0,21,2 },
+	{							/* no seed */
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0xC9,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x01,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x01,
+	  0x02,0xFE,0x13,0xC0,0x53,0x7B,0xBC,0x11,0xAC,0xAA,	/* x */
+	  0x07,0xD7,0x93,0xDE,0x4E,0x6D,0x5E,0x5C,0x94,0xEE,
+	  0xE8,
+	  0x02,0x89,0x07,0x0F,0xB0,0x5D,0x38,0xFF,0x58,0x32,	/* y */
+	  0x1F,0x2E,0x80,0x05,0x36,0xD5,0x38,0xCC,0xDA,0xA3,
+	  0xD9,
+	  0x04,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x02,0x01,0x08,0xA2,0xE0,0xCC,0x0D,0x99,0xF8,0xA5,
+	  0xEF }
 	};
 
-static const unsigned char _EC_SECG_CHAR2_163R1_SEED[] = {
-	0x24,0xB7,0xB1,0x37,0xC8,0xA1,0x4D,0x69,0x6E,0x67,
-	0x68,0x75,0x61,0x51,0x75,0x6F,0xD0,0xDA,0x2E,0x5C};
-static const EC_CURVE_DATA _EC_SECG_CHAR2_163R1 = {
-	NID_X9_62_characteristic_two_field,
-	"0800000000000000000000000000000000000000C9",
-	"07B6882CAAEFA84F9554FF8428BD88E246D2782AE2",
-	"0713612DCDDCB40AAB946BDA29CA91F73AF958AFD9",
-	"0369979697AB43897789566789567F787A7876A654",
-	"00435EDB42EFAFB2989D51FEFCE3C80988F41FF883",
-	"03FFFFFFFFFFFFFFFFFFFF48AAB689C29CA710279B", 2,
+static const struct { EC_CURVE_DATA h; unsigned char data[0+21*6]; }
+	_EC_SECG_CHAR2_163R1 = {
+	{ NID_X9_62_characteristic_two_field,0,21,2 },
+	{							/* no seed */
+#if 0
 /* The algorithm used to derive the curve parameters from
  * the seed used here is slightly different than the
- * algorithm described in X9.62 .
- */
-#if 0
-	_EC_SECG_CHAR2_163R1_SEED, 20,
-#else
-	NULL, 0,
+ * algorithm described in X9.62 . */
+	  0x24,0xB7,0xB1,0x37,0xC8,0xA1,0x4D,0x69,0x6E,0x67,
+	  0x68,0x75,0x61,0x51,0x75,0x6F,0xD0,0xDA,0x2E,0x5C,
 #endif
-	"SECG curve over a 163 bit binary field"
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0xC9,
+	  0x07,0xB6,0x88,0x2C,0xAA,0xEF,0xA8,0x4F,0x95,0x54,	/* a */
+	  0xFF,0x84,0x28,0xBD,0x88,0xE2,0x46,0xD2,0x78,0x2A,
+	  0xE2,
+	  0x07,0x13,0x61,0x2D,0xCD,0xDC,0xB4,0x0A,0xAB,0x94,	/* b */
+	  0x6B,0xDA,0x29,0xCA,0x91,0xF7,0x3A,0xF9,0x58,0xAF,
+	  0xD9,
+	  0x03,0x69,0x97,0x96,0x97,0xAB,0x43,0x89,0x77,0x89,	/* x */
+	  0x56,0x67,0x89,0x56,0x7F,0x78,0x7A,0x78,0x76,0xA6,
+	  0x54,
+	  0x00,0x43,0x5E,0xDB,0x42,0xEF,0xAF,0xB2,0x98,0x9D,	/* y */
+	  0x51,0xFE,0xFC,0xE3,0xC8,0x09,0x88,0xF4,0x1F,0xF8,
+	  0x83,
+	  0x03,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0x48,0xAA,0xB6,0x89,0xC2,0x9C,0xA7,0x10,0x27,
+	  0x9B }
 	};
 
-static const unsigned char _EC_NIST_CHAR2_163B_SEED[] = {
-	0x85,0xE2,0x5B,0xFE,0x5C,0x86,0x22,0x6C,0xDB,0x12,
-	0x01,0x6F,0x75,0x53,0xF9,0xD0,0xE6,0x93,0xA2,0x68};
-static const EC_CURVE_DATA _EC_NIST_CHAR2_163B ={
-	NID_X9_62_characteristic_two_field,
-	"0800000000000000000000000000000000000000C9",
-	"1",
-	"020A601907B8C953CA1481EB10512F78744A3205FD",
-	"03F0EBA16286A2D57EA0991168D4994637E8343E36",
-	"00D51FBC6C71A0094FA2CDD545B11C5C0C797324F1",
-	"040000000000000000000292FE77E70C12A4234C33", 2,
-/* The seed here was used to created the curve parameters in normal
- * basis representation (and not the polynomial representation used here) 
- */
+static const struct { EC_CURVE_DATA h; unsigned char data[0+21*6]; }
+	_EC_NIST_CHAR2_163B = {
+	{ NID_X9_62_characteristic_two_field,0,21,2 },
+	{							/* no seed */
 #if 0
-	_EC_NIST_CHAR2_163B_SEED, 20,
-#else
-	NULL, 0,
+/* The seed here was used to created the curve parameters in normal
+ * basis representation (and not the polynomial representation used here) */
+	  0x85,0xE2,0x5B,0xFE,0x5C,0x86,0x22,0x6C,0xDB,0x12,
+	  0x01,0x6F,0x75,0x53,0xF9,0xD0,0xE6,0x93,0xA2,0x68,
 #endif
-	"NIST/SECG curve over a 163 bit binary field"
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0xC9,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x01,
+	  0x02,0x0A,0x60,0x19,0x07,0xB8,0xC9,0x53,0xCA,0x14,	/* b */
+	  0x81,0xEB,0x10,0x51,0x2F,0x78,0x74,0x4A,0x32,0x05,
+	  0xFD,
+	  0x03,0xF0,0xEB,0xA1,0x62,0x86,0xA2,0xD5,0x7E,0xA0,	/* x */
+	  0x99,0x11,0x68,0xD4,0x99,0x46,0x37,0xE8,0x34,0x3E,
+	  0x36,
+	  0x00,0xD5,0x1F,0xBC,0x6C,0x71,0xA0,0x09,0x4F,0xA2,	/* y */
+	  0xCD,0xD5,0x45,0xB1,0x1C,0x5C,0x0C,0x79,0x73,0x24,
+	  0xF1,
+	  0x04,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x02,0x92,0xFE,0x77,0xE7,0x0C,0x12,0xA4,0x23,0x4C,
+	  0x33 }
 	};
 
-static const unsigned char _EC_SECG_CHAR2_193R1_SEED[] = {
-	0x10,0x3F,0xAE,0xC7,0x4D,0x69,0x6E,0x67,0x68,0x75,
-	0x61,0x51,0x75,0x77,0x7F,0xC5,0xB1,0x91,0xEF,0x30};
-static const EC_CURVE_DATA _EC_SECG_CHAR2_193R1 = {
-	NID_X9_62_characteristic_two_field,
-	"02000000000000000000000000000000000000000000008001",
-	"0017858FEB7A98975169E171F77B4087DE098AC8A911DF7B01",
-	"00FDFB49BFE6C3A89FACADAA7A1E5BBC7CC1C2E5D831478814",
-	"01F481BC5F0FF84A74AD6CDF6FDEF4BF6179625372D8C0C5E1",
-	"0025E399F2903712CCF3EA9E3A1AD17FB0B3201B6AF7CE1B05",
-	"01000000000000000000000000C7F34A778F443ACC920EBA49", 2,
-	_EC_SECG_CHAR2_193R1_SEED, 20,
-	"SECG curve over a 193 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+25*6]; }
+	_EC_SECG_CHAR2_193R1 = {
+	{ NID_X9_62_characteristic_two_field,20,25,2 },
+	{ 0x10,0x3F,0xAE,0xC7,0x4D,0x69,0x6E,0x67,0x68,0x75,	/* seed */
+	  0x61,0x51,0x75,0x77,0x7F,0xC5,0xB1,0x91,0xEF,0x30,
+
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x80,0x01,
+	  0x00,0x17,0x85,0x8F,0xEB,0x7A,0x98,0x97,0x51,0x69,	/* a */
+	  0xE1,0x71,0xF7,0x7B,0x40,0x87,0xDE,0x09,0x8A,0xC8,
+	  0xA9,0x11,0xDF,0x7B,0x01,
+	  0x00,0xFD,0xFB,0x49,0xBF,0xE6,0xC3,0xA8,0x9F,0xAC,	/* b */
+	  0xAD,0xAA,0x7A,0x1E,0x5B,0xBC,0x7C,0xC1,0xC2,0xE5,
+	  0xD8,0x31,0x47,0x88,0x14,
+	  0x01,0xF4,0x81,0xBC,0x5F,0x0F,0xF8,0x4A,0x74,0xAD,	/* x */
+	  0x6C,0xDF,0x6F,0xDE,0xF4,0xBF,0x61,0x79,0x62,0x53,
+	  0x72,0xD8,0xC0,0xC5,0xE1,
+	  0x00,0x25,0xE3,0x99,0xF2,0x90,0x37,0x12,0xCC,0xF3,	/* y */
+	  0xEA,0x9E,0x3A,0x1A,0xD1,0x7F,0xB0,0xB3,0x20,0x1B,
+	  0x6A,0xF7,0xCE,0x1B,0x05,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x00,0xC7,0xF3,0x4A,0x77,0x8F,0x44,0x3A,
+	  0xCC,0x92,0x0E,0xBA,0x49 }
 	};
 
-static const unsigned char _EC_SECG_CHAR2_193R2_SEED[] = {
-	0x10,0xB7,0xB4,0xD6,0x96,0xE6,0x76,0x87,0x56,0x15,
-	0x17,0x51,0x37,0xC8,0xA1,0x6F,0xD0,0xDA,0x22,0x11};
-static const EC_CURVE_DATA _EC_SECG_CHAR2_193R2 = {
-	NID_X9_62_characteristic_two_field,
-	"02000000000000000000000000000000000000000000008001",
-	"0163F35A5137C2CE3EA6ED8667190B0BC43ECD69977702709B",
-	"00C9BB9E8927D4D64C377E2AB2856A5B16E3EFB7F61D4316AE",
-	"00D9B67D192E0367C803F39E1A7E82CA14A651350AAE617E8F",
-	"01CE94335607C304AC29E7DEFBD9CA01F596F927224CDECF6C",
-	"010000000000000000000000015AAB561B005413CCD4EE99D5", 2,
-	_EC_SECG_CHAR2_193R2_SEED, 20,
-	"SECG curve over a 193 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+25*6]; }
+	_EC_SECG_CHAR2_193R2 = {
+	{ NID_X9_62_characteristic_two_field,20,25,2 },
+	{ 0x10,0xB7,0xB4,0xD6,0x96,0xE6,0x76,0x87,0x56,0x15,	/* seed */
+	  0x17,0x51,0x37,0xC8,0xA1,0x6F,0xD0,0xDA,0x22,0x11,
+
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x80,0x01,
+	  0x01,0x63,0xF3,0x5A,0x51,0x37,0xC2,0xCE,0x3E,0xA6,	/* a */
+	  0xED,0x86,0x67,0x19,0x0B,0x0B,0xC4,0x3E,0xCD,0x69,
+	  0x97,0x77,0x02,0x70,0x9B,
+	  0x00,0xC9,0xBB,0x9E,0x89,0x27,0xD4,0xD6,0x4C,0x37,	/* b */
+	  0x7E,0x2A,0xB2,0x85,0x6A,0x5B,0x16,0xE3,0xEF,0xB7,
+	  0xF6,0x1D,0x43,0x16,0xAE,
+	  0x00,0xD9,0xB6,0x7D,0x19,0x2E,0x03,0x67,0xC8,0x03,	/* x */
+	  0xF3,0x9E,0x1A,0x7E,0x82,0xCA,0x14,0xA6,0x51,0x35,
+	  0x0A,0xAE,0x61,0x7E,0x8F,
+	  0x01,0xCE,0x94,0x33,0x56,0x07,0xC3,0x04,0xAC,0x29,	/* y */
+	  0xE7,0xDE,0xFB,0xD9,0xCA,0x01,0xF5,0x96,0xF9,0x27,
+	  0x22,0x4C,0xDE,0xCF,0x6C,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x01,0x5A,0xAB,0x56,0x1B,0x00,0x54,0x13,
+	  0xCC,0xD4,0xEE,0x99,0xD5 }
 	};
 
-static const EC_CURVE_DATA _EC_NIST_CHAR2_233K = {
-	NID_X9_62_characteristic_two_field,
-	"020000000000000000000000000000000000000004000000000000000001",
-	"0",
-	"1",
-	"017232BA853A7E731AF129F22FF4149563A419C26BF50A4C9D6EEFAD6126",
-	"01DB537DECE819B7F70F555A67C427A8CD9BF18AEB9B56E0C11056FAE6A3",
-	"008000000000000000000000000000069D5BB915BCD46EFB1AD5F173ABDF", 4,
-	NULL, 0,
-	"NIST/SECG/WTLS curve over a 233 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+30*6]; }
+	_EC_NIST_CHAR2_233K = {
+	{ NID_X9_62_characteristic_two_field,0,30,4 },
+	{							/* no seed */
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x04,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+
+	  0x01,0x72,0x32,0xBA,0x85,0x3A,0x7E,0x73,0x1A,0xF1,	/* x */
+	  0x29,0xF2,0x2F,0xF4,0x14,0x95,0x63,0xA4,0x19,0xC2,
+	  0x6B,0xF5,0x0A,0x4C,0x9D,0x6E,0xEF,0xAD,0x61,0x26,
+
+	  0x01,0xDB,0x53,0x7D,0xEC,0xE8,0x19,0xB7,0xF7,0x0F,	/* y */
+	  0x55,0x5A,0x67,0xC4,0x27,0xA8,0xCD,0x9B,0xF1,0x8A,
+	  0xEB,0x9B,0x56,0xE0,0xC1,0x10,0x56,0xFA,0xE6,0xA3,
+
+	  0x00,0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x00,0x00,0x00,0x06,0x9D,0x5B,0xB9,0x15,
+	  0xBC,0xD4,0x6E,0xFB,0x1A,0xD5,0xF1,0x73,0xAB,0xDF }
 	};
 
-static const unsigned char _EC_NIST_CHAR2_233B_SEED[] = {
-	0x74,0xD5,0x9F,0xF0,0x7F,0x6B,0x41,0x3D,0x0E,0xA1,
-	0x4B,0x34,0x4B,0x20,0xA2,0xDB,0x04,0x9B,0x50,0xC3};
-static const EC_CURVE_DATA _EC_NIST_CHAR2_233B = {
-	NID_X9_62_characteristic_two_field,
-	"020000000000000000000000000000000000000004000000000000000001",
-	"000000000000000000000000000000000000000000000000000000000001",
-	"0066647EDE6C332C7F8C0923BB58213B333B20E9CE4281FE115F7D8F90AD",
-	"00FAC9DFCBAC8313BB2139F1BB755FEF65BC391F8B36F8F8EB7371FD558B",
-	"01006A08A41903350678E58528BEBF8A0BEFF867A7CA36716F7E01F81052",
-	"01000000000000000000000000000013E974E72F8A6922031D2603CFE0D7", 2,
-	_EC_NIST_CHAR2_233B_SEED, 20,
-	"NIST/SECG/WTLS curve over a 233 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+30*6]; }
+	_EC_NIST_CHAR2_233B = {
+	{ NID_X9_62_characteristic_two_field,20,30,2 },
+	{ 0x74,0xD5,0x9F,0xF0,0x7F,0x6B,0x41,0x3D,0x0E,0xA1,	/* seed */
+	  0x4B,0x34,0x4B,0x20,0xA2,0xDB,0x04,0x9B,0x50,0xC3,
+
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x04,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+
+	  0x00,0x66,0x64,0x7E,0xDE,0x6C,0x33,0x2C,0x7F,0x8C,	/* b */
+	  0x09,0x23,0xBB,0x58,0x21,0x3B,0x33,0x3B,0x20,0xE9,
+	  0xCE,0x42,0x81,0xFE,0x11,0x5F,0x7D,0x8F,0x90,0xAD,
+
+	  0x00,0xFA,0xC9,0xDF,0xCB,0xAC,0x83,0x13,0xBB,0x21,	/* x */
+	  0x39,0xF1,0xBB,0x75,0x5F,0xEF,0x65,0xBC,0x39,0x1F,
+	  0x8B,0x36,0xF8,0xF8,0xEB,0x73,0x71,0xFD,0x55,0x8B,
+
+	  0x01,0x00,0x6A,0x08,0xA4,0x19,0x03,0x35,0x06,0x78,	/* y */
+	  0xE5,0x85,0x28,0xBE,0xBF,0x8A,0x0B,0xEF,0xF8,0x67,
+	  0xA7,0xCA,0x36,0x71,0x6F,0x7E,0x01,0xF8,0x10,0x52,
+
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x00,0x00,0x00,0x13,0xE9,0x74,0xE7,0x2F,
+	  0x8A,0x69,0x22,0x03,0x1D,0x26,0x03,0xCF,0xE0,0xD7 }
 	};
 
-static const EC_CURVE_DATA _EC_SECG_CHAR2_239K1 = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000004000000000000000000000000000000000000001",
-	"0",
-	"1",
-	"29A0B6A887A983E9730988A68727A8B2D126C44CC2CC7B2A6555193035DC",
-	"76310804F12E549BDB011C103089E73510ACB275FC312A5DC6B76553F0CA",
-	"2000000000000000000000000000005A79FEC67CB6E91F1C1DA800E478A5", 4,
-	NULL, 0,
-	"SECG curve over a 239 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+30*6]; }
+	_EC_SECG_CHAR2_239K1 = {
+	{ NID_X9_62_characteristic_two_field,0,30,4 },
+	{							/* no seed */
+	  0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x40,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+
+	  0x29,0xA0,0xB6,0xA8,0x87,0xA9,0x83,0xE9,0x73,0x09,	/* x */
+	  0x88,0xA6,0x87,0x27,0xA8,0xB2,0xD1,0x26,0xC4,0x4C,
+	  0xC2,0xCC,0x7B,0x2A,0x65,0x55,0x19,0x30,0x35,0xDC,
+
+	  0x76,0x31,0x08,0x04,0xF1,0x2E,0x54,0x9B,0xDB,0x01,	/* y */
+	  0x1C,0x10,0x30,0x89,0xE7,0x35,0x10,0xAC,0xB2,0x75,
+	  0xFC,0x31,0x2A,0x5D,0xC6,0xB7,0x65,0x53,0xF0,0xCA,
+
+	  0x20,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x00,0x00,0x00,0x5A,0x79,0xFE,0xC6,0x7C,
+	  0xB6,0xE9,0x1F,0x1C,0x1D,0xA8,0x00,0xE4,0x78,0xA5 }
 	};
 
-static const EC_CURVE_DATA _EC_NIST_CHAR2_283K = {
-	NID_X9_62_characteristic_two_field,
-	"080000000000000000000000000000000000000000000000000000000000000000001"
-	"0A1",
-	"0",
-	"1",
-	"0503213F78CA44883F1A3B8162F188E553CD265F23C1567A16876913B0C2AC2458492"
-	"836",
-	"01CCDA380F1C9E318D90F95D07E5426FE87E45C0E8184698E45962364E34116177DD2"
-	"259",
-	"01FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFE9AE2ED07577265DFF7F94451E061E163"
-	"C61", 4,
-	NULL, 20,
-	"NIST/SECG curve over a 283 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+36*6]; }
+	_EC_NIST_CHAR2_283K = {
+	{ NID_X9_62_characteristic_two_field,0,36,4 },
+	{							/* no seed */
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x10,0xA1,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x01,
+	  0x05,0x03,0x21,0x3F,0x78,0xCA,0x44,0x88,0x3F,0x1A,	/* x */
+	  0x3B,0x81,0x62,0xF1,0x88,0xE5,0x53,0xCD,0x26,0x5F,
+	  0x23,0xC1,0x56,0x7A,0x16,0x87,0x69,0x13,0xB0,0xC2,
+	  0xAC,0x24,0x58,0x49,0x28,0x36,
+	  0x01,0xCC,0xDA,0x38,0x0F,0x1C,0x9E,0x31,0x8D,0x90,	/* y */
+	  0xF9,0x5D,0x07,0xE5,0x42,0x6F,0xE8,0x7E,0x45,0xC0,
+	  0xE8,0x18,0x46,0x98,0xE4,0x59,0x62,0x36,0x4E,0x34,
+	  0x11,0x61,0x77,0xDD,0x22,0x59,
+	  0x01,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xE9,0xAE,
+	  0x2E,0xD0,0x75,0x77,0x26,0x5D,0xFF,0x7F,0x94,0x45,
+	  0x1E,0x06,0x1E,0x16,0x3C,0x61 }
 	};
 
-static const unsigned char _EC_NIST_CHAR2_283B_SEED[] = {
-	0x77,0xE2,0xB0,0x73,0x70,0xEB,0x0F,0x83,0x2A,0x6D,
-	0xD5,0xB6,0x2D,0xFC,0x88,0xCD,0x06,0xBB,0x84,0xBE};
-static const EC_CURVE_DATA _EC_NIST_CHAR2_283B = {
-	NID_X9_62_characteristic_two_field,
-	"080000000000000000000000000000000000000000000000000000000000000000001"
-	"0A1",
-	"000000000000000000000000000000000000000000000000000000000000000000000"
-	"001",
-	"027B680AC8B8596DA5A4AF8A19A0303FCA97FD7645309FA2A581485AF6263E313B79A"
-	"2F5",
-	"05F939258DB7DD90E1934F8C70B0DFEC2EED25B8557EAC9C80E2E198F8CDBECD86B12"
-	"053",
-	"03676854FE24141CB98FE6D4B20D02B4516FF702350EDDB0826779C813F0DF45BE811"
-	"2F4",
-	"03FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEF90399660FC938A90165B042A7CEFADB"
-	"307", 2,
-	_EC_NIST_CHAR2_283B_SEED, 20,
-	"NIST/SECG curve over a 283 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+36*6]; }
+	_EC_NIST_CHAR2_283B = {
+	{ NID_X9_62_characteristic_two_field,20,36,2 },
+	{ 0x77,0xE2,0xB0,0x73,0x70,0xEB,0x0F,0x83,0x2A,0x6D,	/* no seed */
+	  0xD5,0xB6,0x2D,0xFC,0x88,0xCD,0x06,0xBB,0x84,0xBE,
+
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x10,0xA1,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x01,
+	  0x02,0x7B,0x68,0x0A,0xC8,0xB8,0x59,0x6D,0xA5,0xA4,	/* b */
+	  0xAF,0x8A,0x19,0xA0,0x30,0x3F,0xCA,0x97,0xFD,0x76,
+	  0x45,0x30,0x9F,0xA2,0xA5,0x81,0x48,0x5A,0xF6,0x26,
+	  0x3E,0x31,0x3B,0x79,0xA2,0xF5,
+	  0x05,0xF9,0x39,0x25,0x8D,0xB7,0xDD,0x90,0xE1,0x93,	/* x */
+	  0x4F,0x8C,0x70,0xB0,0xDF,0xEC,0x2E,0xED,0x25,0xB8,
+	  0x55,0x7E,0xAC,0x9C,0x80,0xE2,0xE1,0x98,0xF8,0xCD,
+	  0xBE,0xCD,0x86,0xB1,0x20,0x53,
+	  0x03,0x67,0x68,0x54,0xFE,0x24,0x14,0x1C,0xB9,0x8F,	/* y */
+	  0xE6,0xD4,0xB2,0x0D,0x02,0xB4,0x51,0x6F,0xF7,0x02,
+	  0x35,0x0E,0xDD,0xB0,0x82,0x67,0x79,0xC8,0x13,0xF0,
+	  0xDF,0x45,0xBE,0x81,0x12,0xF4,
+	  0x03,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xEF,0x90,
+	  0x39,0x96,0x60,0xFC,0x93,0x8A,0x90,0x16,0x5B,0x04,
+	  0x2A,0x7C,0xEF,0xAD,0xB3,0x07 }
 	};
 
-static const EC_CURVE_DATA _EC_NIST_CHAR2_409K = {
-	NID_X9_62_characteristic_two_field,
-	"020000000000000000000000000000000000000000000000000000000000000000000"
-	"00000000000008000000000000000000001",
-	"0",
-	"1",
-	"0060F05F658F49C1AD3AB1890F7184210EFD0987E307C84C27ACCFB8F9F67CC2C4601"
-	"89EB5AAAA62EE222EB1B35540CFE9023746",
-	"01E369050B7C4E42ACBA1DACBF04299C3460782F918EA427E6325165E9EA10E3DA5F6"
-	"C42E9C55215AA9CA27A5863EC48D8E0286B",
-	"007FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFE5F83B2D4EA20400"
-	"EC4557D5ED3E3E7CA5B4B5C83B8E01E5FCF", 4,
-	NULL, 0,
-	"NIST/SECG curve over a 409 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+52*6]; }
+	_EC_NIST_CHAR2_409K = {
+	{ NID_X9_62_characteristic_two_field,0,52,4 },
+	{							/* no seed */
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x01,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x01,
+	  0x00,0x60,0xF0,0x5F,0x65,0x8F,0x49,0xC1,0xAD,0x3A,	/* x */
+	  0xB1,0x89,0x0F,0x71,0x84,0x21,0x0E,0xFD,0x09,0x87,
+	  0xE3,0x07,0xC8,0x4C,0x27,0xAC,0xCF,0xB8,0xF9,0xF6,
+	  0x7C,0xC2,0xC4,0x60,0x18,0x9E,0xB5,0xAA,0xAA,0x62,
+	  0xEE,0x22,0x2E,0xB1,0xB3,0x55,0x40,0xCF,0xE9,0x02,
+	  0x37,0x46,
+	  0x01,0xE3,0x69,0x05,0x0B,0x7C,0x4E,0x42,0xAC,0xBA,	/* y */
+	  0x1D,0xAC,0xBF,0x04,0x29,0x9C,0x34,0x60,0x78,0x2F,
+	  0x91,0x8E,0xA4,0x27,0xE6,0x32,0x51,0x65,0xE9,0xEA,
+	  0x10,0xE3,0xDA,0x5F,0x6C,0x42,0xE9,0xC5,0x52,0x15,
+	  0xAA,0x9C,0xA2,0x7A,0x58,0x63,0xEC,0x48,0xD8,0xE0,
+	  0x28,0x6B,
+	  0x00,0x7F,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0x5F,0x83,0xB2,
+	  0xD4,0xEA,0x20,0x40,0x0E,0xC4,0x55,0x7D,0x5E,0xD3,
+	  0xE3,0xE7,0xCA,0x5B,0x4B,0x5C,0x83,0xB8,0xE0,0x1E,
+	  0x5F,0xCF }
 	};
 
-static const unsigned char _EC_NIST_CHAR2_409B_SEED[] = {
-	0x40,0x99,0xB5,0xA4,0x57,0xF9,0xD6,0x9F,0x79,0x21,
-	0x3D,0x09,0x4C,0x4B,0xCD,0x4D,0x42,0x62,0x21,0x0B};
-static const EC_CURVE_DATA _EC_NIST_CHAR2_409B = {
-	NID_X9_62_characteristic_two_field,
-	"020000000000000000000000000000000000000000000000000000000000000000000"
-	"00000000000008000000000000000000001",
-	"000000000000000000000000000000000000000000000000000000000000000000000"
-	"00000000000000000000000000000000001",
-	"0021A5C2C8EE9FEB5C4B9A753B7B476B7FD6422EF1F3DD674761FA99D6AC27C8A9A19"
-	"7B272822F6CD57A55AA4F50AE317B13545F",
-	"015D4860D088DDB3496B0C6064756260441CDE4AF1771D4DB01FFE5B34E59703DC255"
-	"A868A1180515603AEAB60794E54BB7996A7",
-	"0061B1CFAB6BE5F32BBFA78324ED106A7636B9C5A7BD198D0158AA4F5488D08F38514"
-	"F1FDF4B4F40D2181B3681C364BA0273C706",
-	"010000000000000000000000000000000000000000000000000001E2AAD6A612F3330"
-	"7BE5FA47C3C9E052F838164CD37D9A21173", 2,
-	_EC_NIST_CHAR2_409B_SEED, 20,
-	"NIST/SECG curve over a 409 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+52*6]; }
+	_EC_NIST_CHAR2_409B = {
+	{ NID_X9_62_characteristic_two_field,20,52,2 },
+	{ 0x40,0x99,0xB5,0xA4,0x57,0xF9,0xD6,0x9F,0x79,0x21,	/* seed */
+	  0x3D,0x09,0x4C,0x4B,0xCD,0x4D,0x42,0x62,0x21,0x0B,
+
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x01,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x01,
+	  0x00,0x21,0xA5,0xC2,0xC8,0xEE,0x9F,0xEB,0x5C,0x4B,	/* b */
+	  0x9A,0x75,0x3B,0x7B,0x47,0x6B,0x7F,0xD6,0x42,0x2E,
+	  0xF1,0xF3,0xDD,0x67,0x47,0x61,0xFA,0x99,0xD6,0xAC,
+	  0x27,0xC8,0xA9,0xA1,0x97,0xB2,0x72,0x82,0x2F,0x6C,
+	  0xD5,0x7A,0x55,0xAA,0x4F,0x50,0xAE,0x31,0x7B,0x13,
+	  0x54,0x5F,
+	  0x01,0x5D,0x48,0x60,0xD0,0x88,0xDD,0xB3,0x49,0x6B,	/* x */
+	  0x0C,0x60,0x64,0x75,0x62,0x60,0x44,0x1C,0xDE,0x4A,
+	  0xF1,0x77,0x1D,0x4D,0xB0,0x1F,0xFE,0x5B,0x34,0xE5,
+	  0x97,0x03,0xDC,0x25,0x5A,0x86,0x8A,0x11,0x80,0x51,
+	  0x56,0x03,0xAE,0xAB,0x60,0x79,0x4E,0x54,0xBB,0x79,
+	  0x96,0xA7,
+	  0x00,0x61,0xB1,0xCF,0xAB,0x6B,0xE5,0xF3,0x2B,0xBF,	/* y */
+	  0xA7,0x83,0x24,0xED,0x10,0x6A,0x76,0x36,0xB9,0xC5,
+	  0xA7,0xBD,0x19,0x8D,0x01,0x58,0xAA,0x4F,0x54,0x88,
+	  0xD0,0x8F,0x38,0x51,0x4F,0x1F,0xDF,0x4B,0x4F,0x40,
+	  0xD2,0x18,0x1B,0x36,0x81,0xC3,0x64,0xBA,0x02,0x73,
+	  0xC7,0x06,
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x01,0xE2,0xAA,0xD6,
+	  0xA6,0x12,0xF3,0x33,0x07,0xBE,0x5F,0xA4,0x7C,0x3C,
+	  0x9E,0x05,0x2F,0x83,0x81,0x64,0xCD,0x37,0xD9,0xA2,
+	  0x11,0x73 }
 	};
 
-static const EC_CURVE_DATA _EC_NIST_CHAR2_571K = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000000000000000000000000000"
-	"000000000000000000000000000000000000000000000000000000000000000000000"
-	"00425",
-	"0",
-	"1",
-	"026EB7A859923FBC82189631F8103FE4AC9CA2970012D5D46024804801841CA443709"
-	"58493B205E647DA304DB4CEB08CBBD1BA39494776FB988B47174DCA88C7E2945283A0"
-	"1C8972",
-	"0349DC807F4FBF374F4AEADE3BCA95314DD58CEC9F307A54FFC61EFC006D8A2C9D497"
-	"9C0AC44AEA74FBEBBB9F772AEDCB620B01A7BA7AF1B320430C8591984F601CD4C143E"
-	"F1C7A3",
-	"020000000000000000000000000000000000000000000000000000000000000000000"
-	"000131850E1F19A63E4B391A8DB917F4138B630D84BE5D639381E91DEB45CFE778F63"
-	"7C1001", 4,
-	NULL, 0,
-	"NIST/SECG curve over a 571 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+72*6]; }
+	_EC_NIST_CHAR2_571K = {
+	{ NID_X9_62_characteristic_two_field,0,72,4 },
+	{							/* no seed */
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x04,0x25,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x01,
+	  0x02,0x6E,0xB7,0xA8,0x59,0x92,0x3F,0xBC,0x82,0x18,	/* x */
+	  0x96,0x31,0xF8,0x10,0x3F,0xE4,0xAC,0x9C,0xA2,0x97,
+	  0x00,0x12,0xD5,0xD4,0x60,0x24,0x80,0x48,0x01,0x84,
+	  0x1C,0xA4,0x43,0x70,0x95,0x84,0x93,0xB2,0x05,0xE6,
+	  0x47,0xDA,0x30,0x4D,0xB4,0xCE,0xB0,0x8C,0xBB,0xD1,
+	  0xBA,0x39,0x49,0x47,0x76,0xFB,0x98,0x8B,0x47,0x17,
+	  0x4D,0xCA,0x88,0xC7,0xE2,0x94,0x52,0x83,0xA0,0x1C,
+	  0x89,0x72,
+	  0x03,0x49,0xDC,0x80,0x7F,0x4F,0xBF,0x37,0x4F,0x4A,	/* y */
+	  0xEA,0xDE,0x3B,0xCA,0x95,0x31,0x4D,0xD5,0x8C,0xEC,
+	  0x9F,0x30,0x7A,0x54,0xFF,0xC6,0x1E,0xFC,0x00,0x6D,
+	  0x8A,0x2C,0x9D,0x49,0x79,0xC0,0xAC,0x44,0xAE,0xA7,
+	  0x4F,0xBE,0xBB,0xB9,0xF7,0x72,0xAE,0xDC,0xB6,0x20,
+	  0xB0,0x1A,0x7B,0xA7,0xAF,0x1B,0x32,0x04,0x30,0xC8,
+	  0x59,0x19,0x84,0xF6,0x01,0xCD,0x4C,0x14,0x3E,0xF1,
+	  0xC7,0xA3,
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x13,0x18,0x50,0xE1,
+	  0xF1,0x9A,0x63,0xE4,0xB3,0x91,0xA8,0xDB,0x91,0x7F,
+	  0x41,0x38,0xB6,0x30,0xD8,0x4B,0xE5,0xD6,0x39,0x38,
+	  0x1E,0x91,0xDE,0xB4,0x5C,0xFE,0x77,0x8F,0x63,0x7C,
+	  0x10,0x01 }
 	};
 
-static const unsigned char _EC_NIST_CHAR2_571B_SEED[] = {
-	0x2A,0xA0,0x58,0xF7,0x3A,0x0E,0x33,0xAB,0x48,0x6B,
-	0x0F,0x61,0x04,0x10,0xC5,0x3A,0x7F,0x13,0x23,0x10};
-static const EC_CURVE_DATA _EC_NIST_CHAR2_571B = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000000000000000000000000000"
-	"000000000000000000000000000000000000000000000000000000000000000000000"
-	"00425",
-	"000000000000000000000000000000000000000000000000000000000000000000000"
-	"000000000000000000000000000000000000000000000000000000000000000000000"
-	"000001",
-	"02F40E7E2221F295DE297117B7F3D62F5C6A97FFCB8CEFF1CD6BA8CE4A9A18AD84FFA"
-	"BBD8EFA59332BE7AD6756A66E294AFD185A78FF12AA520E4DE739BACA0C7FFEFF7F29"
-	"55727A",
-	"0303001D34B856296C16C0D40D3CD7750A93D1D2955FA80AA5F40FC8DB7B2ABDBDE53"
-	"950F4C0D293CDD711A35B67FB1499AE60038614F1394ABFA3B4C850D927E1E7769C8E"
-	"EC2D19",
-	"037BF27342DA639B6DCCFFFEB73D69D78C6C27A6009CBBCA1980F8533921E8A684423"
-	"E43BAB08A576291AF8F461BB2A8B3531D2F0485C19B16E2F1516E23DD3C1A4827AF1B"
-	"8AC15B",
-	"03FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF"
-	"FFFE661CE18FF55987308059B186823851EC7DD9CA1161DE93D5174D66E8382E9BB2F"
-	"E84E47", 2,
-	_EC_NIST_CHAR2_571B_SEED, 20,
-	"NIST/SECG curve over a 571 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+72*6]; }
+	_EC_NIST_CHAR2_571B = {
+	{ NID_X9_62_characteristic_two_field,20,72,2 },
+	{ 0x2A,0xA0,0x58,0xF7,0x3A,0x0E,0x33,0xAB,0x48,0x6B,	/* seed */
+	  0x0F,0x61,0x04,0x10,0xC5,0x3A,0x7F,0x13,0x23,0x10,
+
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x04,0x25,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x01,
+	  0x02,0xF4,0x0E,0x7E,0x22,0x21,0xF2,0x95,0xDE,0x29,	/* b */
+	  0x71,0x17,0xB7,0xF3,0xD6,0x2F,0x5C,0x6A,0x97,0xFF,
+	  0xCB,0x8C,0xEF,0xF1,0xCD,0x6B,0xA8,0xCE,0x4A,0x9A,
+	  0x18,0xAD,0x84,0xFF,0xAB,0xBD,0x8E,0xFA,0x59,0x33,
+	  0x2B,0xE7,0xAD,0x67,0x56,0xA6,0x6E,0x29,0x4A,0xFD,
+	  0x18,0x5A,0x78,0xFF,0x12,0xAA,0x52,0x0E,0x4D,0xE7,
+	  0x39,0xBA,0xCA,0x0C,0x7F,0xFE,0xFF,0x7F,0x29,0x55,
+	  0x72,0x7A,
+	  0x03,0x03,0x00,0x1D,0x34,0xB8,0x56,0x29,0x6C,0x16,	/* x */
+	  0xC0,0xD4,0x0D,0x3C,0xD7,0x75,0x0A,0x93,0xD1,0xD2,
+	  0x95,0x5F,0xA8,0x0A,0xA5,0xF4,0x0F,0xC8,0xDB,0x7B,
+	  0x2A,0xBD,0xBD,0xE5,0x39,0x50,0xF4,0xC0,0xD2,0x93,
+	  0xCD,0xD7,0x11,0xA3,0x5B,0x67,0xFB,0x14,0x99,0xAE,
+	  0x60,0x03,0x86,0x14,0xF1,0x39,0x4A,0xBF,0xA3,0xB4,
+	  0xC8,0x50,0xD9,0x27,0xE1,0xE7,0x76,0x9C,0x8E,0xEC,
+	  0x2D,0x19,
+	  0x03,0x7B,0xF2,0x73,0x42,0xDA,0x63,0x9B,0x6D,0xCC,	/* y */
+	  0xFF,0xFE,0xB7,0x3D,0x69,0xD7,0x8C,0x6C,0x27,0xA6,
+	  0x00,0x9C,0xBB,0xCA,0x19,0x80,0xF8,0x53,0x39,0x21,
+	  0xE8,0xA6,0x84,0x42,0x3E,0x43,0xBA,0xB0,0x8A,0x57,
+	  0x62,0x91,0xAF,0x8F,0x46,0x1B,0xB2,0xA8,0xB3,0x53,
+	  0x1D,0x2F,0x04,0x85,0xC1,0x9B,0x16,0xE2,0xF1,0x51,
+	  0x6E,0x23,0xDD,0x3C,0x1A,0x48,0x27,0xAF,0x1B,0x8A,
+	  0xC1,0x5B,
+	  0x03,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
+	  0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xE6,0x61,0xCE,0x18,
+	  0xFF,0x55,0x98,0x73,0x08,0x05,0x9B,0x18,0x68,0x23,
+	  0x85,0x1E,0xC7,0xDD,0x9C,0xA1,0x16,0x1D,0xE9,0x3D,
+	  0x51,0x74,0xD6,0x6E,0x83,0x82,0xE9,0xBB,0x2F,0xE8,
+	  0x4E,0x47 }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_163V1_SEED[] = {
-	0xD2,0xC0,0xFB,0x15,0x76,0x08,0x60,0xDE,0xF1,0xEE,
-	0xF4,0xD6,0x96,0xE6,0x76,0x87,0x56,0x15,0x17,0x54};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_163V1 = {
-	NID_X9_62_characteristic_two_field,
-	"080000000000000000000000000000000000000107",
-	"072546B5435234A422E0789675F432C89435DE5242",
-	"00C9517D06D5240D3CFF38C74B20B6CD4D6F9DD4D9",
-	"07AF69989546103D79329FCC3D74880F33BBE803CB",
-	"01EC23211B5966ADEA1D3F87F7EA5848AEF0B7CA9F",
-	"0400000000000000000001E60FC8821CC74DAEAFC1", 2,
-	_EC_X9_62_CHAR2_163V1_SEED, 20,
-	"X9.62 curve over a 163 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+21*6]; }
+	_EC_X9_62_CHAR2_163V1 = {
+	{ NID_X9_62_characteristic_two_field,20,21,2 },
+	{ 0xD2,0xC0,0xFB,0x15,0x76,0x08,0x60,0xDE,0xF1,0xEE,
+	  0xF4,0xD6,0x96,0xE6,0x76,0x87,0x56,0x15,0x17,0x54,	/* seed */
+
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+	  0x07,
+	  0x07,0x25,0x46,0xB5,0x43,0x52,0x34,0xA4,0x22,0xE0,	/* a */
+	  0x78,0x96,0x75,0xF4,0x32,0xC8,0x94,0x35,0xDE,0x52,
+	  0x42,
+	  0x00,0xC9,0x51,0x7D,0x06,0xD5,0x24,0x0D,0x3C,0xFF,	/* b */
+	  0x38,0xC7,0x4B,0x20,0xB6,0xCD,0x4D,0x6F,0x9D,0xD4,
+	  0xD9,
+	  0x07,0xAF,0x69,0x98,0x95,0x46,0x10,0x3D,0x79,0x32,	/* x */
+	  0x9F,0xCC,0x3D,0x74,0x88,0x0F,0x33,0xBB,0xE8,0x03,
+	  0xCB,
+	  0x01,0xEC,0x23,0x21,0x1B,0x59,0x66,0xAD,0xEA,0x1D,	/* y */
+	  0x3F,0x87,0xF7,0xEA,0x58,0x48,0xAE,0xF0,0xB7,0xCA,
+	  0x9F,
+	  0x04,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x01,0xE6,0x0F,0xC8,0x82,0x1C,0xC7,0x4D,0xAE,0xAF,
+	  0xC1 }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_163V2_SEED[] = {
-	0x53,0x81,0x4C,0x05,0x0D,0x44,0xD6,0x96,0xE6,0x76,
-	0x87,0x56,0x15,0x17,0x58,0x0C,0xA4,0xE2,0x9F,0xFD};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_163V2 = {
-	NID_X9_62_characteristic_two_field,
- 	"080000000000000000000000000000000000000107",
-	"0108B39E77C4B108BED981ED0E890E117C511CF072",
-	"0667ACEB38AF4E488C407433FFAE4F1C811638DF20",
-	"0024266E4EB5106D0A964D92C4860E2671DB9B6CC5",
-	"079F684DDF6684C5CD258B3890021B2386DFD19FC5",
-	"03FFFFFFFFFFFFFFFFFFFDF64DE1151ADBB78F10A7", 2,
-	_EC_X9_62_CHAR2_163V2_SEED, 20,
-	"X9.62 curve over a 163 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+21*6]; }
+	_EC_X9_62_CHAR2_163V2 = {
+	{ NID_X9_62_characteristic_two_field,20,21,2 },
+	{ 0x53,0x81,0x4C,0x05,0x0D,0x44,0xD6,0x96,0xE6,0x76,	/* seed */
+	  0x87,0x56,0x15,0x17,0x58,0x0C,0xA4,0xE2,0x9F,0xFD,
+
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+	  0x07,
+	  0x01,0x08,0xB3,0x9E,0x77,0xC4,0xB1,0x08,0xBE,0xD9,	/* a */
+	  0x81,0xED,0x0E,0x89,0x0E,0x11,0x7C,0x51,0x1C,0xF0,
+	  0x72,
+	  0x06,0x67,0xAC,0xEB,0x38,0xAF,0x4E,0x48,0x8C,0x40,	/* b */
+	  0x74,0x33,0xFF,0xAE,0x4F,0x1C,0x81,0x16,0x38,0xDF,
+	  0x20,
+	  0x00,0x24,0x26,0x6E,0x4E,0xB5,0x10,0x6D,0x0A,0x96,	/* x */
+	  0x4D,0x92,0xC4,0x86,0x0E,0x26,0x71,0xDB,0x9B,0x6C,
+	  0xC5,
+	  0x07,0x9F,0x68,0x4D,0xDF,0x66,0x84,0xC5,0xCD,0x25,	/* y */
+	  0x8B,0x38,0x90,0x02,0x1B,0x23,0x86,0xDF,0xD1,0x9F,
+	  0xC5,
+	  0x03,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFD,0xF6,0x4D,0xE1,0x15,0x1A,0xDB,0xB7,0x8F,0x10,
+	  0xA7 }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_163V3_SEED[] = {
-	0x50,0xCB,0xF1,0xD9,0x5C,0xA9,0x4D,0x69,0x6E,0x67,
-	0x68,0x75,0x61,0x51,0x75,0xF1,0x6A,0x36,0xA3,0xB8};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_163V3 = {
-	NID_X9_62_characteristic_two_field,
-	"080000000000000000000000000000000000000107",
-	"07A526C63D3E25A256A007699F5447E32AE456B50E",
-	"03F7061798EB99E238FD6F1BF95B48FEEB4854252B",
-	"02F9F87B7C574D0BDECF8A22E6524775F98CDEBDCB",
-	"05B935590C155E17EA48EB3FF3718B893DF59A05D0",
-	"03FFFFFFFFFFFFFFFFFFFE1AEE140F110AFF961309", 2,
-	_EC_X9_62_CHAR2_163V3_SEED, 20,
-	"X9.62 curve over a 163 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+21*6]; }
+	_EC_X9_62_CHAR2_163V3 = {
+	{ NID_X9_62_characteristic_two_field,20,21,2 },
+	{ 0x50,0xCB,0xF1,0xD9,0x5C,0xA9,0x4D,0x69,0x6E,0x67,	/* seed */
+	  0x68,0x75,0x61,0x51,0x75,0xF1,0x6A,0x36,0xA3,0xB8,
+
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+	  0x07,
+	  0x07,0xA5,0x26,0xC6,0x3D,0x3E,0x25,0xA2,0x56,0xA0,	/* a */
+	  0x07,0x69,0x9F,0x54,0x47,0xE3,0x2A,0xE4,0x56,0xB5,
+	  0x0E,
+	  0x03,0xF7,0x06,0x17,0x98,0xEB,0x99,0xE2,0x38,0xFD,	/* b */
+	  0x6F,0x1B,0xF9,0x5B,0x48,0xFE,0xEB,0x48,0x54,0x25,
+	  0x2B,
+	  0x02,0xF9,0xF8,0x7B,0x7C,0x57,0x4D,0x0B,0xDE,0xCF,	/* x */
+	  0x8A,0x22,0xE6,0x52,0x47,0x75,0xF9,0x8C,0xDE,0xBD,
+	  0xCB,
+	  0x05,0xB9,0x35,0x59,0x0C,0x15,0x5E,0x17,0xEA,0x48,	/* y */
+	  0xEB,0x3F,0xF3,0x71,0x8B,0x89,0x3D,0xF5,0x9A,0x05,
+	  0xD0,
+	  0x03,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFE,0x1A,0xEE,0x14,0x0F,0x11,0x0A,0xFF,0x96,0x13,
+	  0x09 }
 	};
 
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_176V1 = {
-	NID_X9_62_characteristic_two_field,
-	"0100000000000000000000000000000000080000000007",
-	"E4E6DB2995065C407D9D39B8D0967B96704BA8E9C90B",
-	"5DDA470ABE6414DE8EC133AE28E9BBD7FCEC0AE0FFF2",
-	"8D16C2866798B600F9F08BB4A8E860F3298CE04A5798",
-	"6FA4539C2DADDDD6BAB5167D61B436E1D92BB16A562C",
-	"00010092537397ECA4F6145799D62B0A19CE06FE26AD", 0xFF6E,
-	NULL, 0,
-	"X9.62 curve over a 176 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+23*6]; }
+	_EC_X9_62_CHAR2_176V1 = {
+	{ NID_X9_62_characteristic_two_field,0,23,0xFF6E },
+	{							/* no seed */
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x08,0x00,0x00,
+	  0x00,0x00,0x07,
+	  0x00,0xE4,0xE6,0xDB,0x29,0x95,0x06,0x5C,0x40,0x7D,	/* a */
+	  0x9D,0x39,0xB8,0xD0,0x96,0x7B,0x96,0x70,0x4B,0xA8,
+	  0xE9,0xC9,0x0B,
+	  0x00,0x5D,0xDA,0x47,0x0A,0xBE,0x64,0x14,0xDE,0x8E,	/* b */
+	  0xC1,0x33,0xAE,0x28,0xE9,0xBB,0xD7,0xFC,0xEC,0x0A,
+	  0xE0,0xFF,0xF2,
+	  0x00,0x8D,0x16,0xC2,0x86,0x67,0x98,0xB6,0x00,0xF9,	/* x */
+	  0xF0,0x8B,0xB4,0xA8,0xE8,0x60,0xF3,0x29,0x8C,0xE0,
+	  0x4A,0x57,0x98,
+	  0x00,0x6F,0xA4,0x53,0x9C,0x2D,0xAD,0xDD,0xD6,0xBA,	/* y */
+	  0xB5,0x16,0x7D,0x61,0xB4,0x36,0xE1,0xD9,0x2B,0xB1,
+	  0x6A,0x56,0x2C,
+	  0x00,0x00,0x01,0x00,0x92,0x53,0x73,0x97,0xEC,0xA4,	/* order */
+	  0xF6,0x14,0x57,0x99,0xD6,0x2B,0x0A,0x19,0xCE,0x06,
+	  0xFE,0x26,0xAD }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_191V1_SEED[] = {
-	0x4E,0x13,0xCA,0x54,0x27,0x44,0xD6,0x96,0xE6,0x76,
-	0x87,0x56,0x15,0x17,0x55,0x2F,0x27,0x9A,0x8C,0x84};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_191V1 = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000201",
-	"2866537B676752636A68F56554E12640276B649EF7526267",
-	"2E45EF571F00786F67B0081B9495A3D95462F5DE0AA185EC",
-	"36B3DAF8A23206F9C4F299D7B21A9C369137F2C84AE1AA0D",
-	"765BE73433B3F95E332932E70EA245CA2418EA0EF98018FB",
-	"40000000000000000000000004A20E90C39067C893BBB9A5", 2,
-	_EC_X9_62_CHAR2_191V1_SEED, 20,
-	"X9.62 curve over a 191 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+24*6]; }
+	_EC_X9_62_CHAR2_191V1 = {
+	{ NID_X9_62_characteristic_two_field,20,24,2 },
+	{ 0x4E,0x13,0xCA,0x54,0x27,0x44,0xD6,0x96,0xE6,0x76,	/* seed */
+	  0x87,0x56,0x15,0x17,0x55,0x2F,0x27,0x9A,0x8C,0x84,
+
+	  0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x02,0x01,
+	  0x28,0x66,0x53,0x7B,0x67,0x67,0x52,0x63,0x6A,0x68,	/* a */
+	  0xF5,0x65,0x54,0xE1,0x26,0x40,0x27,0x6B,0x64,0x9E,
+	  0xF7,0x52,0x62,0x67,
+	  0x2E,0x45,0xEF,0x57,0x1F,0x00,0x78,0x6F,0x67,0xB0,	/* b */
+	  0x08,0x1B,0x94,0x95,0xA3,0xD9,0x54,0x62,0xF5,0xDE,
+	  0x0A,0xA1,0x85,0xEC,
+	  0x36,0xB3,0xDA,0xF8,0xA2,0x32,0x06,0xF9,0xC4,0xF2,	/* x */
+	  0x99,0xD7,0xB2,0x1A,0x9C,0x36,0x91,0x37,0xF2,0xC8,
+	  0x4A,0xE1,0xAA,0x0D,
+	  0x76,0x5B,0xE7,0x34,0x33,0xB3,0xF9,0x5E,0x33,0x29,	/* y */
+	  0x32,0xE7,0x0E,0xA2,0x45,0xCA,0x24,0x18,0xEA,0x0E,
+	  0xF9,0x80,0x18,0xFB,
+	  0x40,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x04,0xA2,0x0E,0x90,0xC3,0x90,0x67,0xC8,
+	  0x93,0xBB,0xB9,0xA5 }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_191V2_SEED[] = {
-	0x08,0x71,0xEF,0x2F,0xEF,0x24,0xD6,0x96,0xE6,0x76,
-	0x87,0x56,0x15,0x17,0x58,0xBE,0xE0,0xD9,0x5C,0x15};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_191V2 = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000201",
-	"401028774D7777C7B7666D1366EA432071274F89FF01E718",
-	"0620048D28BCBD03B6249C99182B7C8CD19700C362C46A01",
-	"3809B2B7CC1B28CC5A87926AAD83FD28789E81E2C9E3BF10",
-	"17434386626D14F3DBF01760D9213A3E1CF37AEC437D668A",
-	"20000000000000000000000050508CB89F652824E06B8173", 4,
-	_EC_X9_62_CHAR2_191V2_SEED, 20,
-	"X9.62 curve over a 191 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+24*6]; }
+	_EC_X9_62_CHAR2_191V2 = {
+	{ NID_X9_62_characteristic_two_field,20,24,4 },
+	{ 0x08,0x71,0xEF,0x2F,0xEF,0x24,0xD6,0x96,0xE6,0x76,	/* seed */
+	  0x87,0x56,0x15,0x17,0x58,0xBE,0xE0,0xD9,0x5C,0x15,
+
+	  0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x02,0x01,
+	  0x40,0x10,0x28,0x77,0x4D,0x77,0x77,0xC7,0xB7,0x66,	/* a */
+	  0x6D,0x13,0x66,0xEA,0x43,0x20,0x71,0x27,0x4F,0x89,
+	  0xFF,0x01,0xE7,0x18,
+	  0x06,0x20,0x04,0x8D,0x28,0xBC,0xBD,0x03,0xB6,0x24,	/* b */
+	  0x9C,0x99,0x18,0x2B,0x7C,0x8C,0xD1,0x97,0x00,0xC3,
+	  0x62,0xC4,0x6A,0x01,
+	  0x38,0x09,0xB2,0xB7,0xCC,0x1B,0x28,0xCC,0x5A,0x87,	/* x */
+	  0x92,0x6A,0xAD,0x83,0xFD,0x28,0x78,0x9E,0x81,0xE2,
+	  0xC9,0xE3,0xBF,0x10,
+	  0x17,0x43,0x43,0x86,0x62,0x6D,0x14,0xF3,0xDB,0xF0,	/* y */
+	  0x17,0x60,0xD9,0x21,0x3A,0x3E,0x1C,0xF3,0x7A,0xEC,
+	  0x43,0x7D,0x66,0x8A,
+	  0x20,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x50,0x50,0x8C,0xB8,0x9F,0x65,0x28,0x24,
+	  0xE0,0x6B,0x81,0x73 }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_191V3_SEED[] = {
-	0xE0,0x53,0x51,0x2D,0xC6,0x84,0xD6,0x96,0xE6,0x76,
-	0x87,0x56,0x15,0x17,0x50,0x67,0xAE,0x78,0x6D,0x1F};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_191V3 = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000201",
-	"6C01074756099122221056911C77D77E77A777E7E7E77FCB",
-	"71FE1AF926CF847989EFEF8DB459F66394D90F32AD3F15E8",
-	"375D4CE24FDE434489DE8746E71786015009E66E38A926DD",
-	"545A39176196575D985999366E6AD34CE0A77CD7127B06BE",
-	"155555555555555555555555610C0B196812BFB6288A3EA3", 6,
-	_EC_X9_62_CHAR2_191V3_SEED, 20,
-	"X9.62 curve over a 191 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+24*6]; }
+	_EC_X9_62_CHAR2_191V3 = {
+	{ NID_X9_62_characteristic_two_field,20,24,6 },
+	{ 0xE0,0x53,0x51,0x2D,0xC6,0x84,0xD6,0x96,0xE6,0x76,	/* seed */
+	  0x87,0x56,0x15,0x17,0x50,0x67,0xAE,0x78,0x6D,0x1F,
+
+	  0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x02,0x01,
+	  0x6C,0x01,0x07,0x47,0x56,0x09,0x91,0x22,0x22,0x10,	/* a */
+	  0x56,0x91,0x1C,0x77,0xD7,0x7E,0x77,0xA7,0x77,0xE7,
+	  0xE7,0xE7,0x7F,0xCB,
+	  0x71,0xFE,0x1A,0xF9,0x26,0xCF,0x84,0x79,0x89,0xEF,	/* b */
+	  0xEF,0x8D,0xB4,0x59,0xF6,0x63,0x94,0xD9,0x0F,0x32,
+	  0xAD,0x3F,0x15,0xE8,
+	  0x37,0x5D,0x4C,0xE2,0x4F,0xDE,0x43,0x44,0x89,0xDE,	/* x */
+	  0x87,0x46,0xE7,0x17,0x86,0x01,0x50,0x09,0xE6,0x6E,
+	  0x38,0xA9,0x26,0xDD,
+	  0x54,0x5A,0x39,0x17,0x61,0x96,0x57,0x5D,0x98,0x59,	/* y */
+	  0x99,0x36,0x6E,0x6A,0xD3,0x4C,0xE0,0xA7,0x7C,0xD7,
+	  0x12,0x7B,0x06,0xBE,
+	  0x15,0x55,0x55,0x55,0x55,0x55,0x55,0x55,0x55,0x55,	/* order */
+	  0x55,0x55,0x61,0x0C,0x0B,0x19,0x68,0x12,0xBF,0xB6,
+	  0x28,0x8A,0x3E,0xA3 }
 	};
 
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_208W1 = {
-	NID_X9_62_characteristic_two_field,
-	"010000000000000000000000000000000800000000000000000007",
-	"0000000000000000000000000000000000000000000000000000",
-	"C8619ED45A62E6212E1160349E2BFA844439FAFC2A3FD1638F9E",
-	"89FDFBE4ABE193DF9559ECF07AC0CE78554E2784EB8C1ED1A57A",
-	"0F55B51A06E78E9AC38A035FF520D8B01781BEB1A6BB08617DE3",
-	"000101BAF95C9723C57B6C21DA2EFF2D5ED588BDD5717E212F9D", 0xFE48,
-	NULL, 0,
-	"X9.62 curve over a 208 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+27*6]; }
+	_EC_X9_62_CHAR2_208W1 = {
+	{ NID_X9_62_characteristic_two_field,0,27,0xFE48 },
+	{							/* no seed */
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x08,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x07,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0xC8,0x61,0x9E,0xD4,0x5A,0x62,0xE6,0x21,0x2E,	/* b */
+	  0x11,0x60,0x34,0x9E,0x2B,0xFA,0x84,0x44,0x39,0xFA,
+	  0xFC,0x2A,0x3F,0xD1,0x63,0x8F,0x9E,
+	  0x00,0x89,0xFD,0xFB,0xE4,0xAB,0xE1,0x93,0xDF,0x95,	/* x */
+	  0x59,0xEC,0xF0,0x7A,0xC0,0xCE,0x78,0x55,0x4E,0x27,
+	  0x84,0xEB,0x8C,0x1E,0xD1,0xA5,0x7A,
+	  0x00,0x0F,0x55,0xB5,0x1A,0x06,0xE7,0x8E,0x9A,0xC3,	/* y */
+	  0x8A,0x03,0x5F,0xF5,0x20,0xD8,0xB0,0x17,0x81,0xBE,
+	  0xB1,0xA6,0xBB,0x08,0x61,0x7D,0xE3,
+	  0x00,0x00,0x01,0x01,0xBA,0xF9,0x5C,0x97,0x23,0xC5,	/* order */
+	  0x7B,0x6C,0x21,0xDA,0x2E,0xFF,0x2D,0x5E,0xD5,0x88,
+	  0xBD,0xD5,0x71,0x7E,0x21,0x2F,0x9D }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_239V1_SEED[] = {
-	0xD3,0x4B,0x9A,0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,
-	0x51,0x75,0xCA,0x71,0xB9,0x20,0xBF,0xEF,0xB0,0x5D};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_239V1 = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000000001000000001",
-	"32010857077C5431123A46B808906756F543423E8D27877578125778AC76",
-	"790408F2EEDAF392B012EDEFB3392F30F4327C0CA3F31FC383C422AA8C16",
-	"57927098FA932E7C0A96D3FD5B706EF7E5F5C156E16B7E7C86038552E91D",
-	"61D8EE5077C33FECF6F1A16B268DE469C3C7744EA9A971649FC7A9616305",
-	"2000000000000000000000000000000F4D42FFE1492A4993F1CAD666E447", 4,
-	_EC_X9_62_CHAR2_239V1_SEED, 20,
-	"X9.62 curve over a 239 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+30*6]; }
+	_EC_X9_62_CHAR2_239V1 = {
+	{ NID_X9_62_characteristic_two_field,20,30,4 },
+	{ 0xD3,0x4B,0x9A,0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,	/* seed */
+	  0x51,0x75,0xCA,0x71,0xB9,0x20,0xBF,0xEF,0xB0,0x5D,
+
+	  0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x00,0x00,0x01,
+
+	  0x32,0x01,0x08,0x57,0x07,0x7C,0x54,0x31,0x12,0x3A,	/* a */
+	  0x46,0xB8,0x08,0x90,0x67,0x56,0xF5,0x43,0x42,0x3E,
+	  0x8D,0x27,0x87,0x75,0x78,0x12,0x57,0x78,0xAC,0x76,
+
+	  0x79,0x04,0x08,0xF2,0xEE,0xDA,0xF3,0x92,0xB0,0x12,	/* b */
+	  0xED,0xEF,0xB3,0x39,0x2F,0x30,0xF4,0x32,0x7C,0x0C,
+	  0xA3,0xF3,0x1F,0xC3,0x83,0xC4,0x22,0xAA,0x8C,0x16,
+
+	  0x57,0x92,0x70,0x98,0xFA,0x93,0x2E,0x7C,0x0A,0x96,	/* x */
+	  0xD3,0xFD,0x5B,0x70,0x6E,0xF7,0xE5,0xF5,0xC1,0x56,
+	  0xE1,0x6B,0x7E,0x7C,0x86,0x03,0x85,0x52,0xE9,0x1D,
+
+	  0x61,0xD8,0xEE,0x50,0x77,0xC3,0x3F,0xEC,0xF6,0xF1,	/* y */
+	  0xA1,0x6B,0x26,0x8D,0xE4,0x69,0xC3,0xC7,0x74,0x4E,
+	  0xA9,0xA9,0x71,0x64,0x9F,0xC7,0xA9,0x61,0x63,0x05,
+
+	  0x20,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* order */
+	  0x00,0x00,0x00,0x00,0x00,0x0F,0x4D,0x42,0xFF,0xE1,
+	  0x49,0x2A,0x49,0x93,0xF1,0xCA,0xD6,0x66,0xE4,0x47 }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_239V2_SEED[] = {
-	0x2A,0xA6,0x98,0x2F,0xDF,0xA4,0xD6,0x96,0xE6,0x76,
-	0x87,0x56,0x15,0x17,0x5D,0x26,0x67,0x27,0x27,0x7D};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_239V2 = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000000001000000001",
-	"4230017757A767FAE42398569B746325D45313AF0766266479B75654E65F",
-	"5037EA654196CFF0CD82B2C14A2FCF2E3FF8775285B545722F03EACDB74B",
-	"28F9D04E900069C8DC47A08534FE76D2B900B7D7EF31F5709F200C4CA205",
-	"5667334C45AFF3B5A03BAD9DD75E2C71A99362567D5453F7FA6E227EC833",
-	"1555555555555555555555555555553C6F2885259C31E3FCDF154624522D", 6,
-	_EC_X9_62_CHAR2_239V2_SEED, 20,
-	"X9.62 curve over a 239 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+30*6]; }
+	_EC_X9_62_CHAR2_239V2 = {
+	{ NID_X9_62_characteristic_two_field,20,30,6 },
+	{ 0x2A,0xA6,0x98,0x2F,0xDF,0xA4,0xD6,0x96,0xE6,0x76,	/* seed */
+	  0x87,0x56,0x15,0x17,0x5D,0x26,0x67,0x27,0x27,0x7D,
+
+	  0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x00,0x00,0x01,
+
+	  0x42,0x30,0x01,0x77,0x57,0xA7,0x67,0xFA,0xE4,0x23,	/* a */
+	  0x98,0x56,0x9B,0x74,0x63,0x25,0xD4,0x53,0x13,0xAF,
+	  0x07,0x66,0x26,0x64,0x79,0xB7,0x56,0x54,0xE6,0x5F,
+
+	  0x50,0x37,0xEA,0x65,0x41,0x96,0xCF,0xF0,0xCD,0x82,	/* b */
+	  0xB2,0xC1,0x4A,0x2F,0xCF,0x2E,0x3F,0xF8,0x77,0x52,
+	  0x85,0xB5,0x45,0x72,0x2F,0x03,0xEA,0xCD,0xB7,0x4B,
+
+	  0x28,0xF9,0xD0,0x4E,0x90,0x00,0x69,0xC8,0xDC,0x47,	/* x */
+	  0xA0,0x85,0x34,0xFE,0x76,0xD2,0xB9,0x00,0xB7,0xD7,
+	  0xEF,0x31,0xF5,0x70,0x9F,0x20,0x0C,0x4C,0xA2,0x05,
+
+	  0x56,0x67,0x33,0x4C,0x45,0xAF,0xF3,0xB5,0xA0,0x3B,	/* y */
+	  0xAD,0x9D,0xD7,0x5E,0x2C,0x71,0xA9,0x93,0x62,0x56,
+	  0x7D,0x54,0x53,0xF7,0xFA,0x6E,0x22,0x7E,0xC8,0x33,
+
+	  0x15,0x55,0x55,0x55,0x55,0x55,0x55,0x55,0x55,0x55,	/* order */
+	  0x55,0x55,0x55,0x55,0x55,0x3C,0x6F,0x28,0x85,0x25,
+	  0x9C,0x31,0xE3,0xFC,0xDF,0x15,0x46,0x24,0x52,0x2D }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_239V3_SEED[] = {
-	0x9E,0x07,0x6F,0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,
-	0x51,0x75,0xE1,0x1E,0x9F,0xDD,0x77,0xF9,0x20,0x41};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_239V3 = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000000001000000001",
-	"01238774666A67766D6676F778E676B66999176666E687666D8766C66A9F",
-	"6A941977BA9F6A435199ACFC51067ED587F519C5ECB541B8E44111DE1D40",
-	"70F6E9D04D289C4E89913CE3530BFDE903977D42B146D539BF1BDE4E9C92",
-	"2E5A0EAF6E5E1305B9004DCE5C0ED7FE59A35608F33837C816D80B79F461",
-	"0CCCCCCCCCCCCCCCCCCCCCCCCCCCCCAC4912D2D9DF903EF9888B8A0E4CFF", 0xA,
-	_EC_X9_62_CHAR2_239V3_SEED, 20,
-	"X9.62 curve over a 239 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+30*6]; }
+	_EC_X9_62_CHAR2_239V3 = {
+	{ NID_X9_62_characteristic_two_field,20,30,0xA },
+	{ 0x9E,0x07,0x6F,0x4D,0x69,0x6E,0x67,0x68,0x75,0x61,	/* seed */
+	  0x51,0x75,0xE1,0x1E,0x9F,0xDD,0x77,0xF9,0x20,0x41,
+
+	  0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x00,0x00,0x01,
+
+	  0x01,0x23,0x87,0x74,0x66,0x6A,0x67,0x76,0x6D,0x66,	/* a */
+	  0x76,0xF7,0x78,0xE6,0x76,0xB6,0x69,0x99,0x17,0x66,
+	  0x66,0xE6,0x87,0x66,0x6D,0x87,0x66,0xC6,0x6A,0x9F,
+
+	  0x6A,0x94,0x19,0x77,0xBA,0x9F,0x6A,0x43,0x51,0x99,	/* b */
+	  0xAC,0xFC,0x51,0x06,0x7E,0xD5,0x87,0xF5,0x19,0xC5,
+	  0xEC,0xB5,0x41,0xB8,0xE4,0x41,0x11,0xDE,0x1D,0x40,
+
+	  0x70,0xF6,0xE9,0xD0,0x4D,0x28,0x9C,0x4E,0x89,0x91,	/* x */
+	  0x3C,0xE3,0x53,0x0B,0xFD,0xE9,0x03,0x97,0x7D,0x42,
+	  0xB1,0x46,0xD5,0x39,0xBF,0x1B,0xDE,0x4E,0x9C,0x92,
+
+	  0x2E,0x5A,0x0E,0xAF,0x6E,0x5E,0x13,0x05,0xB9,0x00,	/* y */
+	  0x4D,0xCE,0x5C,0x0E,0xD7,0xFE,0x59,0xA3,0x56,0x08,
+	  0xF3,0x38,0x37,0xC8,0x16,0xD8,0x0B,0x79,0xF4,0x61,
+
+	  0x0C,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,	/* order */
+	  0xCC,0xCC,0xCC,0xCC,0xCC,0xAC,0x49,0x12,0xD2,0xD9,
+	  0xDF,0x90,0x3E,0xF9,0x88,0x8B,0x8A,0x0E,0x4C,0xFF }
 	};
 
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_272W1 = {
-	NID_X9_62_characteristic_two_field,
-	"010000000000000000000000000000000000000000000000000000010000000000000"
-	"B",
-	"91A091F03B5FBA4AB2CCF49C4EDD220FB028712D42BE752B2C40094DBACDB586FB20",
-	"7167EFC92BB2E3CE7C8AAAFF34E12A9C557003D7C73A6FAF003F99F6CC8482E540F7",
-	"6108BABB2CEEBCF787058A056CBE0CFE622D7723A289E08A07AE13EF0D10D171DD8D",
-	"10C7695716851EEF6BA7F6872E6142FBD241B830FF5EFCACECCAB05E02005DDE9D23",
-	"000100FAF51354E0E39E4892DF6E319C72C8161603FA45AA7B998A167B8F1E629521",
-	0xFF06,
-	NULL, 0,
-	"X9.62 curve over a 272 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+35*6]; }
+	_EC_X9_62_CHAR2_272W1 = {
+	{ NID_X9_62_characteristic_two_field,0,35,0xFF06 },
+	{							/* no seed */
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x0B,
+	  0x00,0x91,0xA0,0x91,0xF0,0x3B,0x5F,0xBA,0x4A,0xB2,	/* a */
+	  0xCC,0xF4,0x9C,0x4E,0xDD,0x22,0x0F,0xB0,0x28,0x71,
+	  0x2D,0x42,0xBE,0x75,0x2B,0x2C,0x40,0x09,0x4D,0xBA,
+	  0xCD,0xB5,0x86,0xFB,0x20,
+	  0x00,0x71,0x67,0xEF,0xC9,0x2B,0xB2,0xE3,0xCE,0x7C,	/* b */
+	  0x8A,0xAA,0xFF,0x34,0xE1,0x2A,0x9C,0x55,0x70,0x03,
+	  0xD7,0xC7,0x3A,0x6F,0xAF,0x00,0x3F,0x99,0xF6,0xCC,
+	  0x84,0x82,0xE5,0x40,0xF7,
+	  0x00,0x61,0x08,0xBA,0xBB,0x2C,0xEE,0xBC,0xF7,0x87,	/* x */
+	  0x05,0x8A,0x05,0x6C,0xBE,0x0C,0xFE,0x62,0x2D,0x77,
+	  0x23,0xA2,0x89,0xE0,0x8A,0x07,0xAE,0x13,0xEF,0x0D,
+	  0x10,0xD1,0x71,0xDD,0x8D,
+	  0x00,0x10,0xC7,0x69,0x57,0x16,0x85,0x1E,0xEF,0x6B,	/* y */
+	  0xA7,0xF6,0x87,0x2E,0x61,0x42,0xFB,0xD2,0x41,0xB8,
+	  0x30,0xFF,0x5E,0xFC,0xAC,0xEC,0xCA,0xB0,0x5E,0x02,
+	  0x00,0x5D,0xDE,0x9D,0x23,
+	  0x00,0x00,0x01,0x00,0xFA,0xF5,0x13,0x54,0xE0,0xE3,	/* order */
+	  0x9E,0x48,0x92,0xDF,0x6E,0x31,0x9C,0x72,0xC8,0x16,
+	  0x16,0x03,0xFA,0x45,0xAA,0x7B,0x99,0x8A,0x16,0x7B,
+	  0x8F,0x1E,0x62,0x95,0x21 }
 	};
 
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_304W1 = {
-	NID_X9_62_characteristic_two_field,
-	"010000000000000000000000000000000000000000000000000000000000000000000"
-	"000000807",
-	"FD0D693149A118F651E6DCE6802085377E5F882D1B510B44160074C1288078365A039"
-	"6C8E681",
-	"BDDB97E555A50A908E43B01C798EA5DAA6788F1EA2794EFCF57166B8C14039601E558"
-	"27340BE",
-	"197B07845E9BE2D96ADB0F5F3C7F2CFFBD7A3EB8B6FEC35C7FD67F26DDF6285A644F7"
-	"40A2614",
-	"E19FBEB76E0DA171517ECF401B50289BF014103288527A9B416A105E80260B549FDC1"
-	"B92C03B",
-	"000101D556572AABAC800101D556572AABAC8001022D5C91DD173F8FB561DA6899164"
-	"443051D", 0xFE2E,
-	NULL, 0,
-	"X9.62 curve over a 304 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+39*6]; }
+	_EC_X9_62_CHAR2_304W1 = {
+	{ NID_X9_62_characteristic_two_field,0,39,0xFE2E },
+	{							/* no seed */
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x08,0x07,
+	  0x00,0xFD,0x0D,0x69,0x31,0x49,0xA1,0x18,0xF6,0x51,	/* a */
+	  0xE6,0xDC,0xE6,0x80,0x20,0x85,0x37,0x7E,0x5F,0x88,
+	  0x2D,0x1B,0x51,0x0B,0x44,0x16,0x00,0x74,0xC1,0x28,
+	  0x80,0x78,0x36,0x5A,0x03,0x96,0xC8,0xE6,0x81,
+	  0x00,0xBD,0xDB,0x97,0xE5,0x55,0xA5,0x0A,0x90,0x8E,	/* b */
+	  0x43,0xB0,0x1C,0x79,0x8E,0xA5,0xDA,0xA6,0x78,0x8F,
+	  0x1E,0xA2,0x79,0x4E,0xFC,0xF5,0x71,0x66,0xB8,0xC1,
+	  0x40,0x39,0x60,0x1E,0x55,0x82,0x73,0x40,0xBE,
+	  0x00,0x19,0x7B,0x07,0x84,0x5E,0x9B,0xE2,0xD9,0x6A,	/* x */
+	  0xDB,0x0F,0x5F,0x3C,0x7F,0x2C,0xFF,0xBD,0x7A,0x3E,
+	  0xB8,0xB6,0xFE,0xC3,0x5C,0x7F,0xD6,0x7F,0x26,0xDD,
+	  0xF6,0x28,0x5A,0x64,0x4F,0x74,0x0A,0x26,0x14,
+	  0x00,0xE1,0x9F,0xBE,0xB7,0x6E,0x0D,0xA1,0x71,0x51,	/* y */
+	  0x7E,0xCF,0x40,0x1B,0x50,0x28,0x9B,0xF0,0x14,0x10,
+	  0x32,0x88,0x52,0x7A,0x9B,0x41,0x6A,0x10,0x5E,0x80,
+	  0x26,0x0B,0x54,0x9F,0xDC,0x1B,0x92,0xC0,0x3B,
+	  0x00,0x00,0x01,0x01,0xD5,0x56,0x57,0x2A,0xAB,0xAC,	/* order */
+	  0x80,0x01,0x01,0xD5,0x56,0x57,0x2A,0xAB,0xAC,0x80,
+	  0x01,0x02,0x2D,0x5C,0x91,0xDD,0x17,0x3F,0x8F,0xB5,
+	  0x61,0xDA,0x68,0x99,0x16,0x44,0x43,0x05,0x1D }
 	};
 
-static const unsigned char _EC_X9_62_CHAR2_359V1_SEED[] = {
-	0x2B,0x35,0x49,0x20,0xB7,0x24,0xD6,0x96,0xE6,0x76,
-	0x87,0x56,0x15,0x17,0x58,0x5B,0xA1,0x33,0x2D,0xC6};
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_359V1 = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000000000000000000000000000"
-	"000100000000000000001",
-	"5667676A654B20754F356EA92017D946567C46675556F19556A04616B567D223A5E05"
-	"656FB549016A96656A557",
-	"2472E2D0197C49363F1FE7F5B6DB075D52B6947D135D8CA445805D39BC34562608968"
-	"7742B6329E70680231988",
-	"3C258EF3047767E7EDE0F1FDAA79DAEE3841366A132E163ACED4ED2401DF9C6BDCDE9"
-	"8E8E707C07A2239B1B097",
-	"53D7E08529547048121E9C95F3791DD804963948F34FAE7BF44EA82365DC7868FE57E"
-	"4AE2DE211305A407104BD",
-	"01AF286BCA1AF286BCA1AF286BCA1AF286BCA1AF286BC9FB8F6B85C556892C20A7EB9"
-	"64FE7719E74F490758D3B", 0x4C,
-	_EC_X9_62_CHAR2_359V1_SEED, 20,
-	"X9.62 curve over a 359 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[20+45*6]; }
+	_EC_X9_62_CHAR2_359V1 = {
+	{ NID_X9_62_characteristic_two_field,20,45,0x4C },
+	{ 0x2B,0x35,0x49,0x20,0xB7,0x24,0xD6,0x96,0xE6,0x76,	/* seed */
+	  0x87,0x56,0x15,0x17,0x58,0x5B,0xA1,0x33,0x2D,0xC6,
+
+	  0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x01,
+	  0x56,0x67,0x67,0x6A,0x65,0x4B,0x20,0x75,0x4F,0x35,	/* a */
+	  0x6E,0xA9,0x20,0x17,0xD9,0x46,0x56,0x7C,0x46,0x67,
+	  0x55,0x56,0xF1,0x95,0x56,0xA0,0x46,0x16,0xB5,0x67,
+	  0xD2,0x23,0xA5,0xE0,0x56,0x56,0xFB,0x54,0x90,0x16,
+	  0xA9,0x66,0x56,0xA5,0x57,
+	  0x24,0x72,0xE2,0xD0,0x19,0x7C,0x49,0x36,0x3F,0x1F,	/* b */
+	  0xE7,0xF5,0xB6,0xDB,0x07,0x5D,0x52,0xB6,0x94,0x7D,
+	  0x13,0x5D,0x8C,0xA4,0x45,0x80,0x5D,0x39,0xBC,0x34,
+	  0x56,0x26,0x08,0x96,0x87,0x74,0x2B,0x63,0x29,0xE7,
+	  0x06,0x80,0x23,0x19,0x88,
+	  0x3C,0x25,0x8E,0xF3,0x04,0x77,0x67,0xE7,0xED,0xE0,	/* x */
+	  0xF1,0xFD,0xAA,0x79,0xDA,0xEE,0x38,0x41,0x36,0x6A,
+	  0x13,0x2E,0x16,0x3A,0xCE,0xD4,0xED,0x24,0x01,0xDF,
+	  0x9C,0x6B,0xDC,0xDE,0x98,0xE8,0xE7,0x07,0xC0,0x7A,
+	  0x22,0x39,0xB1,0xB0,0x97,
+	  0x53,0xD7,0xE0,0x85,0x29,0x54,0x70,0x48,0x12,0x1E,	/* y */
+	  0x9C,0x95,0xF3,0x79,0x1D,0xD8,0x04,0x96,0x39,0x48,
+	  0xF3,0x4F,0xAE,0x7B,0xF4,0x4E,0xA8,0x23,0x65,0xDC,
+	  0x78,0x68,0xFE,0x57,0xE4,0xAE,0x2D,0xE2,0x11,0x30,
+	  0x5A,0x40,0x71,0x04,0xBD,
+	  0x01,0xAF,0x28,0x6B,0xCA,0x1A,0xF2,0x86,0xBC,0xA1,	/* order */
+	  0xAF,0x28,0x6B,0xCA,0x1A,0xF2,0x86,0xBC,0xA1,0xAF,
+	  0x28,0x6B,0xC9,0xFB,0x8F,0x6B,0x85,0xC5,0x56,0x89,
+	  0x2C,0x20,0xA7,0xEB,0x96,0x4F,0xE7,0x71,0x9E,0x74,
+	  0xF4,0x90,0x75,0x8D,0x3B }
 	};
 
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_368W1 = {
-	NID_X9_62_characteristic_two_field,
-	"010000000000000000000000000000000000000000000000000000000000000000000"
-	"0002000000000000000000007",
-	"E0D2EE25095206F5E2A4F9ED229F1F256E79A0E2B455970D8D0D865BD94778C576D62"
-	"F0AB7519CCD2A1A906AE30D",
-	"FC1217D4320A90452C760A58EDCD30C8DD069B3C34453837A34ED50CB54917E1C2112"
-	"D84D164F444F8F74786046A",
-	"1085E2755381DCCCE3C1557AFA10C2F0C0C2825646C5B34A394CBCFA8BC16B22E7E78"
-	"9E927BE216F02E1FB136A5F",
-	"7B3EB1BDDCBA62D5D8B2059B525797FC73822C59059C623A45FF3843CEE8F87CD1855"
-	"ADAA81E2A0750B80FDA2310",
-	"00010090512DA9AF72B08349D98A5DD4C7B0532ECA51CE03E2D10F3B7AC579BD87E90"
-	"9AE40A6F131E9CFCE5BD967", 0xFF70,
-	NULL, 0,
-	"X9.62 curve over a 368 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+47*6]; }
+	_EC_X9_62_CHAR2_368W1 = {
+	{ NID_X9_62_characteristic_two_field,0,47,0xFF70 },
+	{							/* no seed */
+	  0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x20,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x07,
+	  0x00,0xE0,0xD2,0xEE,0x25,0x09,0x52,0x06,0xF5,0xE2,	/* a */
+	  0xA4,0xF9,0xED,0x22,0x9F,0x1F,0x25,0x6E,0x79,0xA0,
+	  0xE2,0xB4,0x55,0x97,0x0D,0x8D,0x0D,0x86,0x5B,0xD9,
+	  0x47,0x78,0xC5,0x76,0xD6,0x2F,0x0A,0xB7,0x51,0x9C,
+	  0xCD,0x2A,0x1A,0x90,0x6A,0xE3,0x0D,
+	  0x00,0xFC,0x12,0x17,0xD4,0x32,0x0A,0x90,0x45,0x2C,	/* b */
+	  0x76,0x0A,0x58,0xED,0xCD,0x30,0xC8,0xDD,0x06,0x9B,
+	  0x3C,0x34,0x45,0x38,0x37,0xA3,0x4E,0xD5,0x0C,0xB5,
+	  0x49,0x17,0xE1,0xC2,0x11,0x2D,0x84,0xD1,0x64,0xF4,
+	  0x44,0xF8,0xF7,0x47,0x86,0x04,0x6A,
+	  0x00,0x10,0x85,0xE2,0x75,0x53,0x81,0xDC,0xCC,0xE3,	/* x */
+	  0xC1,0x55,0x7A,0xFA,0x10,0xC2,0xF0,0xC0,0xC2,0x82,
+	  0x56,0x46,0xC5,0xB3,0x4A,0x39,0x4C,0xBC,0xFA,0x8B,
+	  0xC1,0x6B,0x22,0xE7,0xE7,0x89,0xE9,0x27,0xBE,0x21,
+	  0x6F,0x02,0xE1,0xFB,0x13,0x6A,0x5F,
+	  0x00,0x7B,0x3E,0xB1,0xBD,0xDC,0xBA,0x62,0xD5,0xD8,	/* y */
+	  0xB2,0x05,0x9B,0x52,0x57,0x97,0xFC,0x73,0x82,0x2C,
+	  0x59,0x05,0x9C,0x62,0x3A,0x45,0xFF,0x38,0x43,0xCE,
+	  0xE8,0xF8,0x7C,0xD1,0x85,0x5A,0xDA,0xA8,0x1E,0x2A,
+	  0x07,0x50,0xB8,0x0F,0xDA,0x23,0x10,
+	  0x00,0x00,0x01,0x00,0x90,0x51,0x2D,0xA9,0xAF,0x72,	/* order */
+	  0xB0,0x83,0x49,0xD9,0x8A,0x5D,0xD4,0xC7,0xB0,0x53,
+	  0x2E,0xCA,0x51,0xCE,0x03,0xE2,0xD1,0x0F,0x3B,0x7A,
+	  0xC5,0x79,0xBD,0x87,0xE9,0x09,0xAE,0x40,0xA6,0xF1,
+	  0x31,0xE9,0xCF,0xCE,0x5B,0xD9,0x67 }
 	};
 
-static const EC_CURVE_DATA _EC_X9_62_CHAR2_431R1 = {
-	NID_X9_62_characteristic_two_field,
-	"800000000000000000000000000000000000000000000000000000000000000000000"
-	"000000001000000000000000000000000000001",
-	"1A827EF00DD6FC0E234CAF046C6A5D8A85395B236CC4AD2CF32A0CADBDC9DDF620B0E"
-	"B9906D0957F6C6FEACD615468DF104DE296CD8F",
-	"10D9B4A3D9047D8B154359ABFB1B7F5485B04CEB868237DDC9DEDA982A679A5A919B6"
-	"26D4E50A8DD731B107A9962381FB5D807BF2618",
-	"120FC05D3C67A99DE161D2F4092622FECA701BE4F50F4758714E8A87BBF2A658EF8C2"
-	"1E7C5EFE965361F6C2999C0C247B0DBD70CE6B7",
-	"20D0AF8903A96F8D5FA2C255745D3C451B302C9346D9B7E485E7BCE41F6B591F3E8F6"
-	"ADDCBB0BC4C2F947A7DE1A89B625D6A598B3760",
-	"0340340340340340340340340340340340340340340340340340340323C313FAB5058"
-	"9703B5EC68D3587FEC60D161CC149C1AD4A91", 0x2760,
-	NULL, 0,
-	"X9.62 curve over a 431 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+54*6]; }
+	_EC_X9_62_CHAR2_431R1 = {
+	{ NID_X9_62_characteristic_two_field,0,54,0x2760 },
+	{							/* no seed */
+	  0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x01,
+	  0x1A,0x82,0x7E,0xF0,0x0D,0xD6,0xFC,0x0E,0x23,0x4C,	/* a */
+	  0xAF,0x04,0x6C,0x6A,0x5D,0x8A,0x85,0x39,0x5B,0x23,
+	  0x6C,0xC4,0xAD,0x2C,0xF3,0x2A,0x0C,0xAD,0xBD,0xC9,
+	  0xDD,0xF6,0x20,0xB0,0xEB,0x99,0x06,0xD0,0x95,0x7F,
+	  0x6C,0x6F,0xEA,0xCD,0x61,0x54,0x68,0xDF,0x10,0x4D,
+	  0xE2,0x96,0xCD,0x8F,
+	  0x10,0xD9,0xB4,0xA3,0xD9,0x04,0x7D,0x8B,0x15,0x43,	/* b */
+	  0x59,0xAB,0xFB,0x1B,0x7F,0x54,0x85,0xB0,0x4C,0xEB,
+	  0x86,0x82,0x37,0xDD,0xC9,0xDE,0xDA,0x98,0x2A,0x67,
+	  0x9A,0x5A,0x91,0x9B,0x62,0x6D,0x4E,0x50,0xA8,0xDD,
+	  0x73,0x1B,0x10,0x7A,0x99,0x62,0x38,0x1F,0xB5,0xD8,
+	  0x07,0xBF,0x26,0x18,
+	  0x12,0x0F,0xC0,0x5D,0x3C,0x67,0xA9,0x9D,0xE1,0x61,	/* x */
+	  0xD2,0xF4,0x09,0x26,0x22,0xFE,0xCA,0x70,0x1B,0xE4,
+	  0xF5,0x0F,0x47,0x58,0x71,0x4E,0x8A,0x87,0xBB,0xF2,
+	  0xA6,0x58,0xEF,0x8C,0x21,0xE7,0xC5,0xEF,0xE9,0x65,
+	  0x36,0x1F,0x6C,0x29,0x99,0xC0,0xC2,0x47,0xB0,0xDB,
+	  0xD7,0x0C,0xE6,0xB7,
+	  0x20,0xD0,0xAF,0x89,0x03,0xA9,0x6F,0x8D,0x5F,0xA2,	/* y */
+	  0xC2,0x55,0x74,0x5D,0x3C,0x45,0x1B,0x30,0x2C,0x93,
+	  0x46,0xD9,0xB7,0xE4,0x85,0xE7,0xBC,0xE4,0x1F,0x6B,
+	  0x59,0x1F,0x3E,0x8F,0x6A,0xDD,0xCB,0xB0,0xBC,0x4C,
+	  0x2F,0x94,0x7A,0x7D,0xE1,0xA8,0x9B,0x62,0x5D,0x6A,
+	  0x59,0x8B,0x37,0x60,
+	  0x00,0x03,0x40,0x34,0x03,0x40,0x34,0x03,0x40,0x34,	/* order */
+	  0x03,0x40,0x34,0x03,0x40,0x34,0x03,0x40,0x34,0x03,
+	  0x40,0x34,0x03,0x40,0x34,0x03,0x40,0x34,0x03,0x23,
+	  0xC3,0x13,0xFA,0xB5,0x05,0x89,0x70,0x3B,0x5E,0xC6,
+	  0x8D,0x35,0x87,0xFE,0xC6,0x0D,0x16,0x1C,0xC1,0x49,
+	  0xC1,0xAD,0x4A,0x91 }
 	};
 
-static const EC_CURVE_DATA _EC_WTLS_1 = {
-	NID_X9_62_characteristic_two_field,
-	"020000000000000000000000000201",
-	"1",
-	"1",
-	"01667979A40BA497E5D5C270780617",
-	"00F44B4AF1ECC2630E08785CEBCC15",
-	"00FFFFFFFFFFFFFFFDBF91AF6DEA73", 2,
-	NULL, 0,
-	"WTLS curve over a 113 bit binary field"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+15*6]; }
+	_EC_WTLS_1 = {
+	{ NID_X9_62_characteristic_two_field,0,15,2 },
+	{							/* no seed */
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x02,0x01,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x01,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x01,
+	  0x01,0x66,0x79,0x79,0xA4,0x0B,0xA4,0x97,0xE5,0xD5,	/* x */
+	  0xC2,0x70,0x78,0x06,0x17,
+	  0x00,0xF4,0x4B,0x4A,0xF1,0xEC,0xC2,0x63,0x0E,0x08,	/* y */
+	  0x78,0x5C,0xEB,0xCC,0x15,
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFD,0xBF,	/* order */
+	  0x91,0xAF,0x6D,0xEA,0x73 }
 	};
 
 /* IPSec curves */
@@ -1001,17 +1768,27 @@ static const EC_CURVE_DATA _EC_WTLS_1 = {
  * As the group order is not a prime this curve is not suitable
  * for ECDSA.
  */
-static const EC_CURVE_DATA _EC_IPSEC_155_ID3 = {
-	NID_X9_62_characteristic_two_field,
-	"0800000000000000000000004000000000000001",
-	"0",
-	"07338f",
-	"7b",
-	"1c8",
-	"2AAAAAAAAAAAAAAAAAAC7F3C7881BD0868FA86C",3,
-	NULL, 0,
-	"\n\tIPSec/IKE/Oakley curve #3 over a 155 bit binary field.\n"
-	"\tNot suitable for ECDSA.\n\tQuestionable extension field!"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+20*6]; }
+	_EC_IPSEC_155_ID3 = {
+	{ NID_X9_62_characteristic_two_field,0,20,3 },
+	{							/* no seed */
+	  0x08,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x40,0x00,0x00,0x00,0x00,0x00,0x00,0x01,
+
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x07,0x33,0x8f,
+
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* x */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x7b,
+
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* y */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,0xc8,
+
+	  0x02,0xAA,0xAA,0xAA,0xAA,0xAA,0xAA,0xAA,0xAA,0xAA,	/* order */
+	  0xC7,0xF3,0xC7,0x88,0x1B,0xD0,0x86,0x8F,0xA8,0x6C }
 	};
 
 /* NOTE: The of curves over a extension field of non prime degree
@@ -1019,136 +1796,185 @@ static const EC_CURVE_DATA _EC_IPSEC_155_ID3 = {
  * As the group order is not a prime this curve is not suitable
  * for ECDSA.
  */
-static const EC_CURVE_DATA _EC_IPSEC_185_ID4 = {
-	NID_X9_62_characteristic_two_field,
-	"020000000000000000000000000000200000000000000001",
-	"0",
-	"1ee9",
-	"18",
-	"0d",
-	"FFFFFFFFFFFFFFFFFFFFFFEDF97C44DB9F2420BAFCA75E",2,
-	NULL, 0,
-	"\n\tIPSec/IKE/Oakley curve #4 over a 185 bit binary field.\n"
-	"\tNot suitable for ECDSA.\n\tQuestionable extension field!"
+static const struct { EC_CURVE_DATA h; unsigned char data[0+24*6]; }
+	_EC_IPSEC_185_ID4 = {
+	{ NID_X9_62_characteristic_two_field,0,24,2 },
+	{							/* no seed */
+	  0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* p */
+	  0x00,0x00,0x00,0x00,0x00,0x20,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x01,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* a */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* b */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x1e,0xe9,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* x */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x18,
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,	/* y */
+	  0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+	  0x00,0x00,0x00,0x0d,
+	  0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,	/* order */
+	  0xFF,0xFF,0xED,0xF9,0x7C,0x44,0xDB,0x9F,0x24,0x20,
+	  0xBA,0xFC,0xA7,0x5E }
 	};
 
+#endif
+
 typedef struct _ec_list_element_st {
 	int	nid;
 	const EC_CURVE_DATA *data;
+	const EC_METHOD *(*meth)(void);
+	const char *comment;
 	} ec_list_element;
 
 static const ec_list_element curve_list[] = {
-	/* prime field curves */	
+	/* prime field curves */
 	/* secg curves */
-	{ NID_secp112r1, &_EC_SECG_PRIME_112R1},
-	{ NID_secp112r2, &_EC_SECG_PRIME_112R2},
-	{ NID_secp128r1, &_EC_SECG_PRIME_128R1},
-	{ NID_secp128r2, &_EC_SECG_PRIME_128R2},
-	{ NID_secp160k1, &_EC_SECG_PRIME_160K1},
-	{ NID_secp160r1, &_EC_SECG_PRIME_160R1},
-	{ NID_secp160r2, &_EC_SECG_PRIME_160R2},
+	{ NID_secp112r1, &_EC_SECG_PRIME_112R1.h, 0, "SECG/WTLS curve over a 112 bit prime field" },
+	{ NID_secp112r2, &_EC_SECG_PRIME_112R2.h, 0, "SECG curve over a 112 bit prime field" },
+	{ NID_secp128r1, &_EC_SECG_PRIME_128R1.h, 0, "SECG curve over a 128 bit prime field" },
+	{ NID_secp128r2, &_EC_SECG_PRIME_128R2.h, 0, "SECG curve over a 128 bit prime field" },
+	{ NID_secp160k1, &_EC_SECG_PRIME_160K1.h, 0, "SECG curve over a 160 bit prime field" },
+	{ NID_secp160r1, &_EC_SECG_PRIME_160R1.h, 0, "SECG curve over a 160 bit prime field" },
+	{ NID_secp160r2, &_EC_SECG_PRIME_160R2.h, 0, "SECG/WTLS curve over a 160 bit prime field" },
 	/* SECG secp192r1 is the same as X9.62 prime192v1 and hence omitted */
-	{ NID_secp192k1, &_EC_SECG_PRIME_192K1},
-	{ NID_secp224k1, &_EC_SECG_PRIME_224K1},
-	{ NID_secp224r1, &_EC_NIST_PRIME_224},
-	{ NID_secp256k1, &_EC_SECG_PRIME_256K1},
+	{ NID_secp192k1, &_EC_SECG_PRIME_192K1.h, 0, "SECG curve over a 192 bit prime field" },
+	{ NID_secp224k1, &_EC_SECG_PRIME_224K1.h, 0, "SECG curve over a 224 bit prime field" },
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+	{ NID_secp224r1, &_EC_NIST_PRIME_224.h, EC_GFp_nistp224_method, "NIST/SECG curve over a 224 bit prime field" },
+#else
+	{ NID_secp224r1, &_EC_NIST_PRIME_224.h, 0, "NIST/SECG curve over a 224 bit prime field" },
+#endif
+	{ NID_secp256k1, &_EC_SECG_PRIME_256K1.h, 0, "SECG curve over a 256 bit prime field" },
 	/* SECG secp256r1 is the same as X9.62 prime256v1 and hence omitted */
-	{ NID_secp384r1, &_EC_NIST_PRIME_384},
-	{ NID_secp521r1, &_EC_NIST_PRIME_521},
+	{ NID_secp384r1, &_EC_NIST_PRIME_384.h, 0, "NIST/SECG curve over a 384 bit prime field" },
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+	{ NID_secp521r1, &_EC_NIST_PRIME_521.h, EC_GFp_nistp521_method, "NIST/SECG curve over a 521 bit prime field" },
+#else
+	{ NID_secp521r1, &_EC_NIST_PRIME_521.h, 0, "NIST/SECG curve over a 521 bit prime field" },
+#endif
 	/* X9.62 curves */
-	{ NID_X9_62_prime192v1, &_EC_NIST_PRIME_192},
-	{ NID_X9_62_prime192v2, &_EC_X9_62_PRIME_192V2},
-	{ NID_X9_62_prime192v3, &_EC_X9_62_PRIME_192V3},
-	{ NID_X9_62_prime239v1, &_EC_X9_62_PRIME_239V1},
-	{ NID_X9_62_prime239v2, &_EC_X9_62_PRIME_239V2},
-	{ NID_X9_62_prime239v3, &_EC_X9_62_PRIME_239V3},
-	{ NID_X9_62_prime256v1, &_EC_X9_62_PRIME_256V1},
+	{ NID_X9_62_prime192v1, &_EC_NIST_PRIME_192.h, 0, "NIST/X9.62/SECG curve over a 192 bit prime field" },
+	{ NID_X9_62_prime192v2, &_EC_X9_62_PRIME_192V2.h, 0, "X9.62 curve over a 192 bit prime field" },
+	{ NID_X9_62_prime192v3, &_EC_X9_62_PRIME_192V3.h, 0, "X9.62 curve over a 192 bit prime field" },
+	{ NID_X9_62_prime239v1, &_EC_X9_62_PRIME_239V1.h, 0, "X9.62 curve over a 239 bit prime field" },
+	{ NID_X9_62_prime239v2, &_EC_X9_62_PRIME_239V2.h, 0, "X9.62 curve over a 239 bit prime field" },
+	{ NID_X9_62_prime239v3, &_EC_X9_62_PRIME_239V3.h, 0, "X9.62 curve over a 239 bit prime field" },
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+	{ NID_X9_62_prime256v1, &_EC_X9_62_PRIME_256V1.h, EC_GFp_nistp256_method, "X9.62/SECG curve over a 256 bit prime field" },
+#else
+	{ NID_X9_62_prime256v1, &_EC_X9_62_PRIME_256V1.h, 0, "X9.62/SECG curve over a 256 bit prime field" },
+#endif
+#ifndef OPENSSL_NO_EC2M
 	/* characteristic two field curves */
 	/* NIST/SECG curves */
-	{ NID_sect113r1, &_EC_SECG_CHAR2_113R1},
-	{ NID_sect113r2, &_EC_SECG_CHAR2_113R2},
-	{ NID_sect131r1, &_EC_SECG_CHAR2_131R1},
-	{ NID_sect131r2, &_EC_SECG_CHAR2_131R2},
-	{ NID_sect163k1, &_EC_NIST_CHAR2_163K },
-	{ NID_sect163r1, &_EC_SECG_CHAR2_163R1},
-	{ NID_sect163r2, &_EC_NIST_CHAR2_163B },
-	{ NID_sect193r1, &_EC_SECG_CHAR2_193R1},
-	{ NID_sect193r2, &_EC_SECG_CHAR2_193R2},
-	{ NID_sect233k1, &_EC_NIST_CHAR2_233K },
-	{ NID_sect233r1, &_EC_NIST_CHAR2_233B },
-	{ NID_sect239k1, &_EC_SECG_CHAR2_239K1},
-	{ NID_sect283k1, &_EC_NIST_CHAR2_283K },
-	{ NID_sect283r1, &_EC_NIST_CHAR2_283B },
-	{ NID_sect409k1, &_EC_NIST_CHAR2_409K },
-	{ NID_sect409r1, &_EC_NIST_CHAR2_409B },
-	{ NID_sect571k1, &_EC_NIST_CHAR2_571K },
-	{ NID_sect571r1, &_EC_NIST_CHAR2_571B },
+	{ NID_sect113r1, &_EC_SECG_CHAR2_113R1.h, 0, "SECG curve over a 113 bit binary field" },
+	{ NID_sect113r2, &_EC_SECG_CHAR2_113R2.h, 0, "SECG curve over a 113 bit binary field" },
+	{ NID_sect131r1, &_EC_SECG_CHAR2_131R1.h, 0, "SECG/WTLS curve over a 131 bit binary field" },
+	{ NID_sect131r2, &_EC_SECG_CHAR2_131R2.h, 0, "SECG curve over a 131 bit binary field" },
+	{ NID_sect163k1, &_EC_NIST_CHAR2_163K.h, 0, "NIST/SECG/WTLS curve over a 163 bit binary field" },
+	{ NID_sect163r1, &_EC_SECG_CHAR2_163R1.h, 0, "SECG curve over a 163 bit binary field" },
+	{ NID_sect163r2, &_EC_NIST_CHAR2_163B.h, 0, "NIST/SECG curve over a 163 bit binary field" },
+	{ NID_sect193r1, &_EC_SECG_CHAR2_193R1.h, 0, "SECG curve over a 193 bit binary field" },
+	{ NID_sect193r2, &_EC_SECG_CHAR2_193R2.h, 0, "SECG curve over a 193 bit binary field" },
+	{ NID_sect233k1, &_EC_NIST_CHAR2_233K.h, 0, "NIST/SECG/WTLS curve over a 233 bit binary field" },
+	{ NID_sect233r1, &_EC_NIST_CHAR2_233B.h, 0, "NIST/SECG/WTLS curve over a 233 bit binary field" },
+	{ NID_sect239k1, &_EC_SECG_CHAR2_239K1.h, 0, "SECG curve over a 239 bit binary field" },
+	{ NID_sect283k1, &_EC_NIST_CHAR2_283K.h, 0, "NIST/SECG curve over a 283 bit binary field" },
+	{ NID_sect283r1, &_EC_NIST_CHAR2_283B.h, 0, "NIST/SECG curve over a 283 bit binary field" },
+	{ NID_sect409k1, &_EC_NIST_CHAR2_409K.h, 0, "NIST/SECG curve over a 409 bit binary field" },
+	{ NID_sect409r1, &_EC_NIST_CHAR2_409B.h, 0, "NIST/SECG curve over a 409 bit binary field" },
+	{ NID_sect571k1, &_EC_NIST_CHAR2_571K.h, 0, "NIST/SECG curve over a 571 bit binary field" },
+	{ NID_sect571r1, &_EC_NIST_CHAR2_571B.h, 0, "NIST/SECG curve over a 571 bit binary field" },
 	/* X9.62 curves */
-	{ NID_X9_62_c2pnb163v1, &_EC_X9_62_CHAR2_163V1},
-	{ NID_X9_62_c2pnb163v2, &_EC_X9_62_CHAR2_163V2},
-	{ NID_X9_62_c2pnb163v3, &_EC_X9_62_CHAR2_163V3},
-	{ NID_X9_62_c2pnb176v1, &_EC_X9_62_CHAR2_176V1},
-	{ NID_X9_62_c2tnb191v1, &_EC_X9_62_CHAR2_191V1},
-	{ NID_X9_62_c2tnb191v2, &_EC_X9_62_CHAR2_191V2},
-	{ NID_X9_62_c2tnb191v3, &_EC_X9_62_CHAR2_191V3},
-	{ NID_X9_62_c2pnb208w1, &_EC_X9_62_CHAR2_208W1},
-	{ NID_X9_62_c2tnb239v1, &_EC_X9_62_CHAR2_239V1},
-	{ NID_X9_62_c2tnb239v2, &_EC_X9_62_CHAR2_239V2},
-	{ NID_X9_62_c2tnb239v3, &_EC_X9_62_CHAR2_239V3},
-	{ NID_X9_62_c2pnb272w1, &_EC_X9_62_CHAR2_272W1},
-	{ NID_X9_62_c2pnb304w1, &_EC_X9_62_CHAR2_304W1},
-	{ NID_X9_62_c2tnb359v1, &_EC_X9_62_CHAR2_359V1},
-	{ NID_X9_62_c2pnb368w1, &_EC_X9_62_CHAR2_368W1},
-	{ NID_X9_62_c2tnb431r1, &_EC_X9_62_CHAR2_431R1},
+	{ NID_X9_62_c2pnb163v1, &_EC_X9_62_CHAR2_163V1.h, 0, "X9.62 curve over a 163 bit binary field" },
+	{ NID_X9_62_c2pnb163v2, &_EC_X9_62_CHAR2_163V2.h, 0, "X9.62 curve over a 163 bit binary field" },
+	{ NID_X9_62_c2pnb163v3, &_EC_X9_62_CHAR2_163V3.h, 0, "X9.62 curve over a 163 bit binary field" },
+	{ NID_X9_62_c2pnb176v1, &_EC_X9_62_CHAR2_176V1.h, 0, "X9.62 curve over a 176 bit binary field" },
+	{ NID_X9_62_c2tnb191v1, &_EC_X9_62_CHAR2_191V1.h, 0, "X9.62 curve over a 191 bit binary field" },
+	{ NID_X9_62_c2tnb191v2, &_EC_X9_62_CHAR2_191V2.h, 0, "X9.62 curve over a 191 bit binary field" },
+	{ NID_X9_62_c2tnb191v3, &_EC_X9_62_CHAR2_191V3.h, 0, "X9.62 curve over a 191 bit binary field" },
+	{ NID_X9_62_c2pnb208w1, &_EC_X9_62_CHAR2_208W1.h, 0, "X9.62 curve over a 208 bit binary field" },
+	{ NID_X9_62_c2tnb239v1, &_EC_X9_62_CHAR2_239V1.h, 0, "X9.62 curve over a 239 bit binary field" },
+	{ NID_X9_62_c2tnb239v2, &_EC_X9_62_CHAR2_239V2.h, 0, "X9.62 curve over a 239 bit binary field" },
+	{ NID_X9_62_c2tnb239v3, &_EC_X9_62_CHAR2_239V3.h, 0, "X9.62 curve over a 239 bit binary field" },
+	{ NID_X9_62_c2pnb272w1, &_EC_X9_62_CHAR2_272W1.h, 0, "X9.62 curve over a 272 bit binary field" },
+	{ NID_X9_62_c2pnb304w1, &_EC_X9_62_CHAR2_304W1.h, 0, "X9.62 curve over a 304 bit binary field" },
+	{ NID_X9_62_c2tnb359v1, &_EC_X9_62_CHAR2_359V1.h, 0, "X9.62 curve over a 359 bit binary field" },
+	{ NID_X9_62_c2pnb368w1, &_EC_X9_62_CHAR2_368W1.h, 0, "X9.62 curve over a 368 bit binary field" },
+	{ NID_X9_62_c2tnb431r1, &_EC_X9_62_CHAR2_431R1.h, 0, "X9.62 curve over a 431 bit binary field" },
 	/* the WAP/WTLS curves
 	 * [unlike SECG, spec has its own OIDs for curves from X9.62] */
-	{ NID_wap_wsg_idm_ecid_wtls1, &_EC_WTLS_1},
-	{ NID_wap_wsg_idm_ecid_wtls3, &_EC_NIST_CHAR2_163K},
-	{ NID_wap_wsg_idm_ecid_wtls4, &_EC_SECG_CHAR2_113R1},
-	{ NID_wap_wsg_idm_ecid_wtls5, &_EC_X9_62_CHAR2_163V1},
-	{ NID_wap_wsg_idm_ecid_wtls6, &_EC_SECG_PRIME_112R1},
-	{ NID_wap_wsg_idm_ecid_wtls7, &_EC_SECG_PRIME_160R2},
-	{ NID_wap_wsg_idm_ecid_wtls8, &_EC_WTLS_8},
-	{ NID_wap_wsg_idm_ecid_wtls9, &_EC_WTLS_9 },
-	{ NID_wap_wsg_idm_ecid_wtls10, &_EC_NIST_CHAR2_233K},
-	{ NID_wap_wsg_idm_ecid_wtls11, &_EC_NIST_CHAR2_233B},
-	{ NID_wap_wsg_idm_ecid_wtls12, &_EC_WTLS_12},
+	{ NID_wap_wsg_idm_ecid_wtls1, &_EC_WTLS_1.h, 0, "WTLS curve over a 113 bit binary field" },
+	{ NID_wap_wsg_idm_ecid_wtls3, &_EC_NIST_CHAR2_163K.h, 0, "NIST/SECG/WTLS curve over a 163 bit binary field" },
+	{ NID_wap_wsg_idm_ecid_wtls4, &_EC_SECG_CHAR2_113R1.h, 0, "SECG curve over a 113 bit binary field" },
+	{ NID_wap_wsg_idm_ecid_wtls5, &_EC_X9_62_CHAR2_163V1.h, 0, "X9.62 curve over a 163 bit binary field" },
+#endif
+	{ NID_wap_wsg_idm_ecid_wtls6, &_EC_SECG_PRIME_112R1.h, 0, "SECG/WTLS curve over a 112 bit prime field" },
+	{ NID_wap_wsg_idm_ecid_wtls7, &_EC_SECG_PRIME_160R2.h, 0, "SECG/WTLS curve over a 160 bit prime field" },
+	{ NID_wap_wsg_idm_ecid_wtls8, &_EC_WTLS_8.h, 0, "WTLS curve over a 112 bit prime field" },
+	{ NID_wap_wsg_idm_ecid_wtls9, &_EC_WTLS_9.h, 0, "WTLS curve over a 160 bit prime field" },
+#ifndef OPENSSL_NO_EC2M
+	{ NID_wap_wsg_idm_ecid_wtls10, &_EC_NIST_CHAR2_233K.h, 0, "NIST/SECG/WTLS curve over a 233 bit binary field" },
+	{ NID_wap_wsg_idm_ecid_wtls11, &_EC_NIST_CHAR2_233B.h, 0, "NIST/SECG/WTLS curve over a 233 bit binary field" },
+#endif
+	{ NID_wap_wsg_idm_ecid_wtls12, &_EC_WTLS_12.h, 0, "WTLS curvs over a 224 bit prime field" },
+#ifndef OPENSSL_NO_EC2M
 	/* IPSec curves */
-	{ NID_ipsec3, &_EC_IPSEC_155_ID3},
-	{ NID_ipsec4, &_EC_IPSEC_185_ID4},
+	{ NID_ipsec3, &_EC_IPSEC_155_ID3.h, 0, "\n\tIPSec/IKE/Oakley curve #3 over a 155 bit binary field.\n"
+	  "\tNot suitable for ECDSA.\n\tQuestionable extension field!" },
+	{ NID_ipsec4, &_EC_IPSEC_185_ID4.h, 0, "\n\tIPSec/IKE/Oakley curve #4 over a 185 bit binary field.\n"
+	  "\tNot suitable for ECDSA.\n\tQuestionable extension field!" },
+#endif
 };
 
-static size_t curve_list_length = sizeof(curve_list)/sizeof(ec_list_element);
+#define curve_list_length (sizeof(curve_list)/sizeof(ec_list_element))
 
-static EC_GROUP *ec_group_new_from_data(const EC_CURVE_DATA *data)
+static EC_GROUP *ec_group_new_from_data(const ec_list_element curve)
 	{
 	EC_GROUP *group=NULL;
 	EC_POINT *P=NULL;
 	BN_CTX	 *ctx=NULL;
-	BIGNUM 	 *p=NULL, *a=NULL, *b=NULL, *x=NULL, *y=NULL, *order=NULL;
+	BIGNUM	 *p=NULL, *a=NULL, *b=NULL, *x=NULL, *y=NULL, *order=NULL;
 	int	 ok=0;
+	int	 seed_len,param_len;
+	const EC_METHOD *meth;
+	const EC_CURVE_DATA *data;
+	const unsigned char *params;
 
 	if ((ctx = BN_CTX_new()) == NULL)
 		{
 		ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_MALLOC_FAILURE);
 		goto err;
 		}
-	if ((p = BN_new()) == NULL || (a = BN_new()) == NULL || 
-		(b = BN_new()) == NULL || (x = BN_new()) == NULL ||
-		(y = BN_new()) == NULL || (order = BN_new()) == NULL)
-		{
-		ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_MALLOC_FAILURE);
-		goto err;
-		}
-	
-	if (!BN_hex2bn(&p, data->p) || !BN_hex2bn(&a, data->a)
-		|| !BN_hex2bn(&b, data->b))
+
+	data = curve.data;
+	seed_len  = data->seed_len;
+	param_len = data->param_len;
+	params	  = (const unsigned char *)(data+1);	/* skip header */
+	params	 += seed_len;				/* skip seed   */
+
+	if (!(p = BN_bin2bn(params+0*param_len, param_len, NULL))
+		|| !(a = BN_bin2bn(params+1*param_len, param_len, NULL))
+		|| !(b = BN_bin2bn(params+2*param_len, param_len, NULL)))
 		{
 		ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_BN_LIB);
 		goto err;
 		}
 
-	if (data->field_type == NID_X9_62_prime_field)
+	if (curve.meth != 0)
+		{
+		meth = curve.meth();
+		if (((group = EC_GROUP_new(meth)) == NULL) ||
+			(!(group->meth->group_set_curve(group, p, a, b, ctx))))
+			{
+			ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_EC_LIB);
+			goto err;
+			}
+		}
+	else if (data->field_type == NID_X9_62_prime_field)
 		{
 		if ((group = EC_GROUP_new_curve_GFp(p, a, b, ctx)) == NULL)
 			{
@@ -1156,32 +1982,36 @@ static EC_GROUP *ec_group_new_from_data(const EC_CURVE_DATA *data)
 			goto err;
 			}
 		}
-		else
-		{ /* field_type == NID_X9_62_characteristic_two_field */
+#ifndef OPENSSL_NO_EC2M
+	else	/* field_type == NID_X9_62_characteristic_two_field */
+		{
 		if ((group = EC_GROUP_new_curve_GF2m(p, a, b, ctx)) == NULL)
 			{
 			ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_EC_LIB);
 			goto err;
 			}
 		}
+#endif
 
 	if ((P = EC_POINT_new(group)) == NULL)
 		{
 		ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_EC_LIB);
 		goto err;
 		}
-	
-	if (!BN_hex2bn(&x, data->x) || !BN_hex2bn(&y, data->y))
+
+	if (!(x = BN_bin2bn(params+3*param_len, param_len, NULL))
+		|| !(y = BN_bin2bn(params+4*param_len, param_len, NULL)))
 		{
 		ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_BN_LIB);
 		goto err;
 		}
-	if (!EC_POINT_set_affine_coordinates_GF2m(group, P, x, y, ctx))
+	if (!EC_POINT_set_affine_coordinates_GFp(group, P, x, y, ctx))
 		{
 		ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_EC_LIB);
 		goto err;
 		}
-	if (!BN_hex2bn(&order, data->order) || !BN_set_word(x, data->cofactor))
+	if (!(order = BN_bin2bn(params+5*param_len, param_len, NULL))
+		|| !BN_set_word(x, (BN_ULONG)data->cofactor))
 		{
 		ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_BN_LIB);
 		goto err;
@@ -1191,9 +2021,9 @@ static EC_GROUP *ec_group_new_from_data(const EC_CURVE_DATA *data)
 		ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_EC_LIB);
 		goto err;
 		}
-	if (data->seed)
+	if (seed_len)
 		{
-		if (!EC_GROUP_set_seed(group, data->seed, data->seed_len))
+		if (!EC_GROUP_set_seed(group, params-seed_len, seed_len))
 			{
 			ECerr(EC_F_EC_GROUP_NEW_FROM_DATA, ERR_R_EC_LIB);
 			goto err;
@@ -1236,7 +2066,7 @@ EC_GROUP *EC_GROUP_new_by_curve_name(int nid)
 	for (i=0; i<curve_list_length; i++)
 		if (curve_list[i].nid == nid)
 			{
-			ret = ec_group_new_from_data(curve_list[i].data);
+			ret = ec_group_new_from_data(curve_list[i]);
 			break;
 			}
 
@@ -1263,7 +2093,7 @@ size_t EC_get_builtin_curves(EC_builtin_curve *r, size_t nitems)
 	for (i = 0; i < min; i++)
 		{
 		r[i].nid = curve_list[i].nid;
-		r[i].comment = curve_list[i].data->comment;
+		r[i].comment = curve_list[i].comment;
 		}
 
 	return curve_list_length;
diff --git a/crypto/ec/ec_cvt.c b/crypto/ec/ec_cvt.c
index d45640bab902..bfcbab35fe69 100644
--- a/crypto/ec/ec_cvt.c
+++ b/crypto/ec/ec_cvt.c
@@ -78,7 +78,32 @@ EC_GROUP *EC_GROUP_new_curve_GFp(const BIGNUM *p, const BIGNUM *a, const BIGNUM
 	const EC_METHOD *meth;
 	EC_GROUP *ret;
 
+#if defined(OPENSSL_BN_ASM_MONT)
+	/*
+	 * This might appear controversial, but the fact is that generic
+	 * prime method was observed to deliver better performance even
+	 * for NIST primes on a range of platforms, e.g.: 60%-15%
+	 * improvement on IA-64, ~25% on ARM, 30%-90% on P4, 20%-25%
+	 * in 32-bit build and 35%--12% in 64-bit build on Core2...
+	 * Coefficients are relative to optimized bn_nist.c for most
+	 * intensive ECDSA verify and ECDH operations for 192- and 521-
+	 * bit keys respectively. Choice of these boundary values is
+	 * arguable, because the dependency of improvement coefficient
+	 * from key length is not a "monotone" curve. For example while
+	 * 571-bit result is 23% on ARM, 384-bit one is -1%. But it's
+	 * generally faster, sometimes "respectfully" faster, sometimes
+	 * "tolerably" slower... What effectively happens is that loop
+	 * with bn_mul_add_words is put against bn_mul_mont, and the
+	 * latter "wins" on short vectors. Correct solution should be
+	 * implementing dedicated NxN multiplication subroutines for
+	 * small N. But till it materializes, let's stick to generic
+	 * prime method...
+	 *						<appro>
+	 */
+	meth = EC_GFp_mont_method();
+#else
 	meth = EC_GFp_nist_method();
+#endif
 	
 	ret = EC_GROUP_new(meth);
 	if (ret == NULL)
@@ -122,7 +147,7 @@ EC_GROUP *EC_GROUP_new_curve_GFp(const BIGNUM *p, const BIGNUM *a, const BIGNUM
 	return ret;
 	}
 
-
+#ifndef OPENSSL_NO_EC2M
 EC_GROUP *EC_GROUP_new_curve_GF2m(const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx)
 	{
 	const EC_METHOD *meth;
@@ -142,3 +167,4 @@ EC_GROUP *EC_GROUP_new_curve_GF2m(const BIGNUM *p, const BIGNUM *a, const BIGNUM
 
 	return ret;
 	}
+#endif
diff --git a/crypto/ec/ec_err.c b/crypto/ec/ec_err.c
index d04c8955604e..0d19398731ac 100644
--- a/crypto/ec/ec_err.c
+++ b/crypto/ec/ec_err.c
@@ -1,6 +1,6 @@
 /* crypto/ec/ec_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,10 +70,19 @@
 
 static ERR_STRING_DATA EC_str_functs[]=
 	{
+{ERR_FUNC(EC_F_BN_TO_FELEM),	"BN_TO_FELEM"},
 {ERR_FUNC(EC_F_COMPUTE_WNAF),	"COMPUTE_WNAF"},
 {ERR_FUNC(EC_F_D2I_ECPARAMETERS),	"d2i_ECParameters"},
 {ERR_FUNC(EC_F_D2I_ECPKPARAMETERS),	"d2i_ECPKParameters"},
 {ERR_FUNC(EC_F_D2I_ECPRIVATEKEY),	"d2i_ECPrivateKey"},
+{ERR_FUNC(EC_F_DO_EC_KEY_PRINT),	"DO_EC_KEY_PRINT"},
+{ERR_FUNC(EC_F_ECKEY_PARAM2TYPE),	"ECKEY_PARAM2TYPE"},
+{ERR_FUNC(EC_F_ECKEY_PARAM_DECODE),	"ECKEY_PARAM_DECODE"},
+{ERR_FUNC(EC_F_ECKEY_PRIV_DECODE),	"ECKEY_PRIV_DECODE"},
+{ERR_FUNC(EC_F_ECKEY_PRIV_ENCODE),	"ECKEY_PRIV_ENCODE"},
+{ERR_FUNC(EC_F_ECKEY_PUB_DECODE),	"ECKEY_PUB_DECODE"},
+{ERR_FUNC(EC_F_ECKEY_PUB_ENCODE),	"ECKEY_PUB_ENCODE"},
+{ERR_FUNC(EC_F_ECKEY_TYPE2PARAM),	"ECKEY_TYPE2PARAM"},
 {ERR_FUNC(EC_F_ECPARAMETERS_PRINT),	"ECParameters_print"},
 {ERR_FUNC(EC_F_ECPARAMETERS_PRINT_FP),	"ECParameters_print_fp"},
 {ERR_FUNC(EC_F_ECPKPARAMETERS_PRINT),	"ECPKParameters_print"},
@@ -104,6 +113,15 @@ static ERR_STRING_DATA EC_str_functs[]=
 {ERR_FUNC(EC_F_EC_GFP_MONT_FIELD_SQR),	"ec_GFp_mont_field_sqr"},
 {ERR_FUNC(EC_F_EC_GFP_MONT_GROUP_SET_CURVE),	"ec_GFp_mont_group_set_curve"},
 {ERR_FUNC(EC_F_EC_GFP_MONT_GROUP_SET_CURVE_GFP),	"EC_GFP_MONT_GROUP_SET_CURVE_GFP"},
+{ERR_FUNC(EC_F_EC_GFP_NISTP224_GROUP_SET_CURVE),	"ec_GFp_nistp224_group_set_curve"},
+{ERR_FUNC(EC_F_EC_GFP_NISTP224_POINTS_MUL),	"ec_GFp_nistp224_points_mul"},
+{ERR_FUNC(EC_F_EC_GFP_NISTP224_POINT_GET_AFFINE_COORDINATES),	"ec_GFp_nistp224_point_get_affine_coordinates"},
+{ERR_FUNC(EC_F_EC_GFP_NISTP256_GROUP_SET_CURVE),	"ec_GFp_nistp256_group_set_curve"},
+{ERR_FUNC(EC_F_EC_GFP_NISTP256_POINTS_MUL),	"ec_GFp_nistp256_points_mul"},
+{ERR_FUNC(EC_F_EC_GFP_NISTP256_POINT_GET_AFFINE_COORDINATES),	"ec_GFp_nistp256_point_get_affine_coordinates"},
+{ERR_FUNC(EC_F_EC_GFP_NISTP521_GROUP_SET_CURVE),	"ec_GFp_nistp521_group_set_curve"},
+{ERR_FUNC(EC_F_EC_GFP_NISTP521_POINTS_MUL),	"ec_GFp_nistp521_points_mul"},
+{ERR_FUNC(EC_F_EC_GFP_NISTP521_POINT_GET_AFFINE_COORDINATES),	"ec_GFp_nistp521_point_get_affine_coordinates"},
 {ERR_FUNC(EC_F_EC_GFP_NIST_FIELD_MUL),	"ec_GFp_nist_field_mul"},
 {ERR_FUNC(EC_F_EC_GFP_NIST_FIELD_SQR),	"ec_GFp_nist_field_sqr"},
 {ERR_FUNC(EC_F_EC_GFP_NIST_GROUP_SET_CURVE),	"ec_GFp_nist_group_set_curve"},
@@ -146,8 +164,8 @@ static ERR_STRING_DATA EC_str_functs[]=
 {ERR_FUNC(EC_F_EC_KEY_NEW),	"EC_KEY_new"},
 {ERR_FUNC(EC_F_EC_KEY_PRINT),	"EC_KEY_print"},
 {ERR_FUNC(EC_F_EC_KEY_PRINT_FP),	"EC_KEY_print_fp"},
+{ERR_FUNC(EC_F_EC_KEY_SET_PUBLIC_KEY_AFFINE_COORDINATES),	"EC_KEY_set_public_key_affine_coordinates"},
 {ERR_FUNC(EC_F_EC_POINTS_MAKE_AFFINE),	"EC_POINTs_make_affine"},
-{ERR_FUNC(EC_F_EC_POINTS_MUL),	"EC_POINTs_mul"},
 {ERR_FUNC(EC_F_EC_POINT_ADD),	"EC_POINT_add"},
 {ERR_FUNC(EC_F_EC_POINT_CMP),	"EC_POINT_cmp"},
 {ERR_FUNC(EC_F_EC_POINT_COPY),	"EC_POINT_copy"},
@@ -177,7 +195,17 @@ static ERR_STRING_DATA EC_str_functs[]=
 {ERR_FUNC(EC_F_I2D_ECPKPARAMETERS),	"i2d_ECPKParameters"},
 {ERR_FUNC(EC_F_I2D_ECPRIVATEKEY),	"i2d_ECPrivateKey"},
 {ERR_FUNC(EC_F_I2O_ECPUBLICKEY),	"i2o_ECPublicKey"},
+{ERR_FUNC(EC_F_NISTP224_PRE_COMP_NEW),	"NISTP224_PRE_COMP_NEW"},
+{ERR_FUNC(EC_F_NISTP256_PRE_COMP_NEW),	"NISTP256_PRE_COMP_NEW"},
+{ERR_FUNC(EC_F_NISTP521_PRE_COMP_NEW),	"NISTP521_PRE_COMP_NEW"},
 {ERR_FUNC(EC_F_O2I_ECPUBLICKEY),	"o2i_ECPublicKey"},
+{ERR_FUNC(EC_F_OLD_EC_PRIV_DECODE),	"OLD_EC_PRIV_DECODE"},
+{ERR_FUNC(EC_F_PKEY_EC_CTRL),	"PKEY_EC_CTRL"},
+{ERR_FUNC(EC_F_PKEY_EC_CTRL_STR),	"PKEY_EC_CTRL_STR"},
+{ERR_FUNC(EC_F_PKEY_EC_DERIVE),	"PKEY_EC_DERIVE"},
+{ERR_FUNC(EC_F_PKEY_EC_KEYGEN),	"PKEY_EC_KEYGEN"},
+{ERR_FUNC(EC_F_PKEY_EC_PARAMGEN),	"PKEY_EC_PARAMGEN"},
+{ERR_FUNC(EC_F_PKEY_EC_SIGN),	"PKEY_EC_SIGN"},
 {0,NULL}
 	};
 
@@ -185,17 +213,23 @@ static ERR_STRING_DATA EC_str_reasons[]=
 	{
 {ERR_REASON(EC_R_ASN1_ERROR)             ,"asn1 error"},
 {ERR_REASON(EC_R_ASN1_UNKNOWN_FIELD)     ,"asn1 unknown field"},
+{ERR_REASON(EC_R_BIGNUM_OUT_OF_RANGE)    ,"bignum out of range"},
 {ERR_REASON(EC_R_BUFFER_TOO_SMALL)       ,"buffer too small"},
+{ERR_REASON(EC_R_COORDINATES_OUT_OF_RANGE),"coordinates out of range"},
 {ERR_REASON(EC_R_D2I_ECPKPARAMETERS_FAILURE),"d2i ecpkparameters failure"},
+{ERR_REASON(EC_R_DECODE_ERROR)           ,"decode error"},
 {ERR_REASON(EC_R_DISCRIMINANT_IS_ZERO)   ,"discriminant is zero"},
 {ERR_REASON(EC_R_EC_GROUP_NEW_BY_NAME_FAILURE),"ec group new by name failure"},
 {ERR_REASON(EC_R_FIELD_TOO_LARGE)        ,"field too large"},
+{ERR_REASON(EC_R_GF2M_NOT_SUPPORTED)     ,"gf2m not supported"},
 {ERR_REASON(EC_R_GROUP2PKPARAMETERS_FAILURE),"group2pkparameters failure"},
 {ERR_REASON(EC_R_I2D_ECPKPARAMETERS_FAILURE),"i2d ecpkparameters failure"},
 {ERR_REASON(EC_R_INCOMPATIBLE_OBJECTS)   ,"incompatible objects"},
 {ERR_REASON(EC_R_INVALID_ARGUMENT)       ,"invalid argument"},
 {ERR_REASON(EC_R_INVALID_COMPRESSED_POINT),"invalid compressed point"},
 {ERR_REASON(EC_R_INVALID_COMPRESSION_BIT),"invalid compression bit"},
+{ERR_REASON(EC_R_INVALID_CURVE)          ,"invalid curve"},
+{ERR_REASON(EC_R_INVALID_DIGEST_TYPE)    ,"invalid digest type"},
 {ERR_REASON(EC_R_INVALID_ENCODING)       ,"invalid encoding"},
 {ERR_REASON(EC_R_INVALID_FIELD)          ,"invalid field"},
 {ERR_REASON(EC_R_INVALID_FORM)           ,"invalid form"},
@@ -203,6 +237,7 @@ static ERR_STRING_DATA EC_str_reasons[]=
 {ERR_REASON(EC_R_INVALID_PENTANOMIAL_BASIS),"invalid pentanomial basis"},
 {ERR_REASON(EC_R_INVALID_PRIVATE_KEY)    ,"invalid private key"},
 {ERR_REASON(EC_R_INVALID_TRINOMIAL_BASIS),"invalid trinomial basis"},
+{ERR_REASON(EC_R_KEYS_NOT_SET)           ,"keys not set"},
 {ERR_REASON(EC_R_MISSING_PARAMETERS)     ,"missing parameters"},
 {ERR_REASON(EC_R_MISSING_PRIVATE_KEY)    ,"missing private key"},
 {ERR_REASON(EC_R_NOT_A_NIST_PRIME)       ,"not a NIST prime"},
@@ -210,6 +245,7 @@ static ERR_STRING_DATA EC_str_reasons[]=
 {ERR_REASON(EC_R_NOT_IMPLEMENTED)        ,"not implemented"},
 {ERR_REASON(EC_R_NOT_INITIALIZED)        ,"not initialized"},
 {ERR_REASON(EC_R_NO_FIELD_MOD)           ,"no field mod"},
+{ERR_REASON(EC_R_NO_PARAMETERS_SET)      ,"no parameters set"},
 {ERR_REASON(EC_R_PASSED_NULL_PARAMETER)  ,"passed null parameter"},
 {ERR_REASON(EC_R_PKPARAMETERS2GROUP_FAILURE),"pkparameters2group failure"},
 {ERR_REASON(EC_R_POINT_AT_INFINITY)      ,"point at infinity"},
@@ -220,6 +256,7 @@ static ERR_STRING_DATA EC_str_reasons[]=
 {ERR_REASON(EC_R_UNKNOWN_GROUP)          ,"unknown group"},
 {ERR_REASON(EC_R_UNKNOWN_ORDER)          ,"unknown order"},
 {ERR_REASON(EC_R_UNSUPPORTED_FIELD)      ,"unsupported field"},
+{ERR_REASON(EC_R_WRONG_CURVE_PARAMETERS) ,"wrong curve parameters"},
 {ERR_REASON(EC_R_WRONG_ORDER)            ,"wrong order"},
 {0,NULL}
 	};
diff --git a/crypto/ec/ec_key.c b/crypto/ec/ec_key.c
index 522802c07ae1..bf9fd2dc2c43 100644
--- a/crypto/ec/ec_key.c
+++ b/crypto/ec/ec_key.c
@@ -64,7 +64,9 @@
 #include <string.h>
 #include "ec_lcl.h"
 #include <openssl/err.h>
-#include <string.h>
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 
 EC_KEY *EC_KEY_new(void)
 	{
@@ -78,6 +80,7 @@ EC_KEY *EC_KEY_new(void)
 		}
 
 	ret->version = 1;	
+	ret->flags = 0;
 	ret->group   = NULL;
 	ret->pub_key = NULL;
 	ret->priv_key= NULL;
@@ -197,6 +200,7 @@ EC_KEY *EC_KEY_copy(EC_KEY *dest, const EC_KEY *src)
 	dest->enc_flag  = src->enc_flag;
 	dest->conv_form = src->conv_form;
 	dest->version   = src->version;
+	dest->flags = src->flags;
 
 	return dest;
 	}
@@ -237,6 +241,11 @@ int EC_KEY_generate_key(EC_KEY *eckey)
 	BIGNUM	*priv_key = NULL, *order = NULL;
 	EC_POINT *pub_key = NULL;
 
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode())
+		return FIPS_ec_key_generate_key(eckey);
+#endif
+
 	if (!eckey || !eckey->group)
 		{
 		ECerr(EC_F_EC_KEY_GENERATE_KEY, ERR_R_PASSED_NULL_PARAMETER);
@@ -371,6 +380,82 @@ err:
 	return(ok);
 	}
 
+int EC_KEY_set_public_key_affine_coordinates(EC_KEY *key, BIGNUM *x, BIGNUM *y)
+	{
+	BN_CTX *ctx = NULL;
+	BIGNUM *tx, *ty;
+	EC_POINT *point = NULL;
+	int ok = 0, tmp_nid, is_char_two = 0;
+
+	if (!key || !key->group || !x || !y)
+		{
+		ECerr(EC_F_EC_KEY_SET_PUBLIC_KEY_AFFINE_COORDINATES,
+						ERR_R_PASSED_NULL_PARAMETER);
+		return 0;
+		}
+	ctx = BN_CTX_new();
+	if (!ctx)
+		goto err;
+
+	point = EC_POINT_new(key->group);
+
+	if (!point)
+		goto err;
+
+	tmp_nid = EC_METHOD_get_field_type(EC_GROUP_method_of(key->group));
+
+        if (tmp_nid == NID_X9_62_characteristic_two_field)
+		is_char_two = 1;
+
+	tx = BN_CTX_get(ctx);
+	ty = BN_CTX_get(ctx);
+#ifndef OPENSSL_NO_EC2M
+	if (is_char_two)
+		{
+		if (!EC_POINT_set_affine_coordinates_GF2m(key->group, point,
+								x, y, ctx))
+			goto err;
+		if (!EC_POINT_get_affine_coordinates_GF2m(key->group, point,
+								tx, ty, ctx))
+			goto err;
+		}
+	else
+#endif
+		{
+		if (!EC_POINT_set_affine_coordinates_GFp(key->group, point,
+								x, y, ctx))
+			goto err;
+		if (!EC_POINT_get_affine_coordinates_GFp(key->group, point,
+								tx, ty, ctx))
+			goto err;
+		}
+	/* Check if retrieved coordinates match originals: if not values
+	 * are out of range.
+	 */
+	if (BN_cmp(x, tx) || BN_cmp(y, ty))
+		{
+		ECerr(EC_F_EC_KEY_SET_PUBLIC_KEY_AFFINE_COORDINATES,
+			EC_R_COORDINATES_OUT_OF_RANGE);
+		goto err;
+		}
+
+	if (!EC_KEY_set_public_key(key, point))
+		goto err;
+
+	if (EC_KEY_check_key(key) == 0)
+		goto err;
+
+	ok = 1;
+
+	err:
+	if (ctx)
+		BN_CTX_free(ctx);
+	if (point)
+		EC_POINT_free(point);
+	return ok;
+
+	}
+
 const EC_GROUP *EC_KEY_get0_group(const EC_KEY *key)
 	{
 	return key->group;
@@ -461,3 +546,18 @@ int EC_KEY_precompute_mult(EC_KEY *key, BN_CTX *ctx)
 		return 0;
 	return EC_GROUP_precompute_mult(key->group, ctx);
 	}
+
+int EC_KEY_get_flags(const EC_KEY *key)
+	{
+	return key->flags;
+	}
+
+void EC_KEY_set_flags(EC_KEY *key, int flags)
+	{
+	key->flags |= flags;
+	}
+
+void EC_KEY_clear_flags(EC_KEY *key, int flags)
+	{
+	key->flags &= ~flags;
+	}
diff --git a/crypto/ec/ec_lcl.h b/crypto/ec/ec_lcl.h
index fdd7aa275563..da7967df38a3 100644
--- a/crypto/ec/ec_lcl.h
+++ b/crypto/ec/ec_lcl.h
@@ -3,7 +3,7 @@
  * Originally written by Bodo Moeller for the OpenSSL project.
  */
 /* ====================================================================
- * Copyright (c) 1998-2003 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1998-2010 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -82,10 +82,15 @@
 # endif
 #endif
 
+/* Use default functions for poin2oct, oct2point and compressed coordinates */
+#define EC_FLAGS_DEFAULT_OCT	0x1
+
 /* Structure details are not part of the exported interface,
  * so all this may change in future versions. */
 
 struct ec_method_st {
+	/* Various method flags */
+	int flags;
 	/* used by EC_METHOD_get_field_type: */
 	int field_type; /* a NID */
 
@@ -205,11 +210,14 @@ struct ec_group_st {
 	               * irreducible polynomial defining the field.
 	               */
 
-	unsigned int poly[5]; /* Field specification for curves over GF(2^m).
-	                       * The irreducible f(t) is then of the form:
-	                       *     t^poly[0] + t^poly[1] + ... + t^poly[k]
-	                       * where m = poly[0] > poly[1] > ... > poly[k] = 0.
-	                       */
+	int poly[6]; /* Field specification for curves over GF(2^m).
+	              * The irreducible f(t) is then of the form:
+	              *     t^poly[0] + t^poly[1] + ... + t^poly[k]
+	              * where m = poly[0] > poly[1] > ... > poly[k] = 0.
+	              * The array is terminated with poly[k+1]=-1.
+	              * All elliptic curve irreducibles have at most 5
+	              * non-zero terms.
+	              */
 
 	BIGNUM a, b; /* Curve coefficients.
 	              * (Here the assumption is that BIGNUMs can be used
@@ -241,6 +249,7 @@ struct ec_key_st {
 	point_conversion_form_t conv_form;
 
 	int 	references;
+	int	flags;
 
 	EC_EXTRA_DATA *method_data;
 } /* EC_KEY */;
@@ -388,3 +397,50 @@ int ec_GF2m_simple_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar,
 	size_t num, const EC_POINT *points[], const BIGNUM *scalars[], BN_CTX *);
 int ec_GF2m_precompute_mult(EC_GROUP *group, BN_CTX *ctx);
 int ec_GF2m_have_precompute_mult(const EC_GROUP *group);
+
+/* method functions in ec2_mult.c */
+int ec_GF2m_simple_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar,
+	size_t num, const EC_POINT *points[], const BIGNUM *scalars[], BN_CTX *);
+int ec_GF2m_precompute_mult(EC_GROUP *group, BN_CTX *ctx);
+int ec_GF2m_have_precompute_mult(const EC_GROUP *group);
+
+#ifndef OPENSSL_EC_NISTP_64_GCC_128
+/* method functions in ecp_nistp224.c */
+int ec_GFp_nistp224_group_init(EC_GROUP *group);
+int ec_GFp_nistp224_group_set_curve(EC_GROUP *group, const BIGNUM *p, const BIGNUM *a, const BIGNUM *n, BN_CTX *);
+int ec_GFp_nistp224_point_get_affine_coordinates(const EC_GROUP *group, const EC_POINT *point, BIGNUM *x, BIGNUM *y, BN_CTX *ctx);
+int ec_GFp_nistp224_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar, size_t num, const EC_POINT *points[], const BIGNUM *scalars[], BN_CTX *);
+int ec_GFp_nistp224_points_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar, size_t num, const EC_POINT *points[], const BIGNUM *scalars[], BN_CTX *ctx);
+int ec_GFp_nistp224_precompute_mult(EC_GROUP *group, BN_CTX *ctx);
+int ec_GFp_nistp224_have_precompute_mult(const EC_GROUP *group);
+
+/* method functions in ecp_nistp256.c */
+int ec_GFp_nistp256_group_init(EC_GROUP *group);
+int ec_GFp_nistp256_group_set_curve(EC_GROUP *group, const BIGNUM *p, const BIGNUM *a, const BIGNUM *n, BN_CTX *);
+int ec_GFp_nistp256_point_get_affine_coordinates(const EC_GROUP *group, const EC_POINT *point, BIGNUM *x, BIGNUM *y, BN_CTX *ctx);
+int ec_GFp_nistp256_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar, size_t num, const EC_POINT *points[], const BIGNUM *scalars[], BN_CTX *);
+int ec_GFp_nistp256_points_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar, size_t num, const EC_POINT *points[], const BIGNUM *scalars[], BN_CTX *ctx);
+int ec_GFp_nistp256_precompute_mult(EC_GROUP *group, BN_CTX *ctx);
+int ec_GFp_nistp256_have_precompute_mult(const EC_GROUP *group);
+
+/* method functions in ecp_nistp521.c */
+int ec_GFp_nistp521_group_init(EC_GROUP *group);
+int ec_GFp_nistp521_group_set_curve(EC_GROUP *group, const BIGNUM *p, const BIGNUM *a, const BIGNUM *n, BN_CTX *);
+int ec_GFp_nistp521_point_get_affine_coordinates(const EC_GROUP *group, const EC_POINT *point, BIGNUM *x, BIGNUM *y, BN_CTX *ctx);
+int ec_GFp_nistp521_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar, size_t num, const EC_POINT *points[], const BIGNUM *scalars[], BN_CTX *);
+int ec_GFp_nistp521_points_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar, size_t num, const EC_POINT *points[], const BIGNUM *scalars[], BN_CTX *ctx);
+int ec_GFp_nistp521_precompute_mult(EC_GROUP *group, BN_CTX *ctx);
+int ec_GFp_nistp521_have_precompute_mult(const EC_GROUP *group);
+
+/* utility functions in ecp_nistputil.c */
+void ec_GFp_nistp_points_make_affine_internal(size_t num, void *point_array,
+	size_t felem_size, void *tmp_felems,
+	void (*felem_one)(void *out),
+	int (*felem_is_zero)(const void *in),
+	void (*felem_assign)(void *out, const void *in),
+	void (*felem_square)(void *out, const void *in),
+	void (*felem_mul)(void *out, const void *in1, const void *in2),
+	void (*felem_inv)(void *out, const void *in),
+	void (*felem_contract)(void *out, const void *in));
+void ec_GFp_nistp_recode_scalar_bits(unsigned char *sign, unsigned char *digit, unsigned char in);
+#endif
diff --git a/crypto/ec/ec_lib.c b/crypto/ec/ec_lib.c
index 5af84376c602..25247b580332 100644
--- a/crypto/ec/ec_lib.c
+++ b/crypto/ec/ec_lib.c
@@ -79,7 +79,7 @@ EC_GROUP *EC_GROUP_new(const EC_METHOD *meth)
 
 	if (meth == NULL)
 		{
-		ECerr(EC_F_EC_GROUP_NEW, ERR_R_PASSED_NULL_PARAMETER);
+		ECerr(EC_F_EC_GROUP_NEW, EC_R_SLOT_FULL);
 		return NULL;
 		}
 	if (meth->group_init == 0)
@@ -425,7 +425,7 @@ int EC_GROUP_get_curve_GFp(const EC_GROUP *group, BIGNUM *p, BIGNUM *a, BIGNUM *
 	return group->meth->group_get_curve(group, p, a, b, ctx);
 	}
 
-
+#ifndef OPENSSL_NO_EC2M
 int EC_GROUP_set_curve_GF2m(EC_GROUP *group, const BIGNUM *p, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx)
 	{
 	if (group->meth->group_set_curve == 0)
@@ -446,7 +446,7 @@ int EC_GROUP_get_curve_GF2m(const EC_GROUP *group, BIGNUM *p, BIGNUM *a, BIGNUM
 		}
 	return group->meth->group_get_curve(group, p, a, b, ctx);
 	}
-
+#endif
 
 int EC_GROUP_get_degree(const EC_GROUP *group)
 	{
@@ -740,7 +740,7 @@ void EC_POINT_clear_free(EC_POINT *point)
 
 	if (point->meth->point_clear_finish != 0)
 		point->meth->point_clear_finish(point);
-	else if (point->meth != NULL && point->meth->point_finish != 0)
+	else if (point->meth->point_finish != 0)
 		point->meth->point_finish(point);
 	OPENSSL_cleanse(point, sizeof *point);
 	OPENSSL_free(point);
@@ -856,7 +856,7 @@ int EC_POINT_set_affine_coordinates_GFp(const EC_GROUP *group, EC_POINT *point,
 	return group->meth->point_set_affine_coordinates(group, point, x, y, ctx);
 	}
 
-
+#ifndef OPENSSL_NO_EC2M
 int EC_POINT_set_affine_coordinates_GF2m(const EC_GROUP *group, EC_POINT *point,
 	const BIGNUM *x, const BIGNUM *y, BN_CTX *ctx)
 	{
@@ -872,7 +872,7 @@ int EC_POINT_set_affine_coordinates_GF2m(const EC_GROUP *group, EC_POINT *point,
 		}
 	return group->meth->point_set_affine_coordinates(group, point, x, y, ctx);
 	}
-
+#endif
 
 int EC_POINT_get_affine_coordinates_GFp(const EC_GROUP *group, const EC_POINT *point,
 	BIGNUM *x, BIGNUM *y, BN_CTX *ctx)
@@ -890,7 +890,7 @@ int EC_POINT_get_affine_coordinates_GFp(const EC_GROUP *group, const EC_POINT *p
 	return group->meth->point_get_affine_coordinates(group, point, x, y, ctx);
 	}
 
-
+#ifndef OPENSSL_NO_EC2M
 int EC_POINT_get_affine_coordinates_GF2m(const EC_GROUP *group, const EC_POINT *point,
 	BIGNUM *x, BIGNUM *y, BN_CTX *ctx)
 	{
@@ -906,75 +906,7 @@ int EC_POINT_get_affine_coordinates_GF2m(const EC_GROUP *group, const EC_POINT *
 		}
 	return group->meth->point_get_affine_coordinates(group, point, x, y, ctx);
 	}
-
-
-int EC_POINT_set_compressed_coordinates_GFp(const EC_GROUP *group, EC_POINT *point,
-	const BIGNUM *x, int y_bit, BN_CTX *ctx)
-	{
-	if (group->meth->point_set_compressed_coordinates == 0)
-		{
-		ECerr(EC_F_EC_POINT_SET_COMPRESSED_COORDINATES_GFP, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED);
-		return 0;
-		}
-	if (group->meth != point->meth)
-		{
-		ECerr(EC_F_EC_POINT_SET_COMPRESSED_COORDINATES_GFP, EC_R_INCOMPATIBLE_OBJECTS);
-		return 0;
-		}
-	return group->meth->point_set_compressed_coordinates(group, point, x, y_bit, ctx);
-	}
-
-
-int EC_POINT_set_compressed_coordinates_GF2m(const EC_GROUP *group, EC_POINT *point,
-	const BIGNUM *x, int y_bit, BN_CTX *ctx)
-	{
-	if (group->meth->point_set_compressed_coordinates == 0)
-		{
-		ECerr(EC_F_EC_POINT_SET_COMPRESSED_COORDINATES_GF2M, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED);
-		return 0;
-		}
-	if (group->meth != point->meth)
-		{
-		ECerr(EC_F_EC_POINT_SET_COMPRESSED_COORDINATES_GF2M, EC_R_INCOMPATIBLE_OBJECTS);
-		return 0;
-		}
-	return group->meth->point_set_compressed_coordinates(group, point, x, y_bit, ctx);
-	}
-
-
-size_t EC_POINT_point2oct(const EC_GROUP *group, const EC_POINT *point, point_conversion_form_t form,
-        unsigned char *buf, size_t len, BN_CTX *ctx)
-	{
-	if (group->meth->point2oct == 0)
-		{
-		ECerr(EC_F_EC_POINT_POINT2OCT, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED);
-		return 0;
-		}
-	if (group->meth != point->meth)
-		{
-		ECerr(EC_F_EC_POINT_POINT2OCT, EC_R_INCOMPATIBLE_OBJECTS);
-		return 0;
-		}
-	return group->meth->point2oct(group, point, form, buf, len, ctx);
-	}
-
-
-int EC_POINT_oct2point(const EC_GROUP *group, EC_POINT *point,
-        const unsigned char *buf, size_t len, BN_CTX *ctx)
-	{
-	if (group->meth->oct2point == 0)
-		{
-		ECerr(EC_F_EC_POINT_OCT2POINT, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED);
-		return 0;
-		}
-	if (group->meth != point->meth)
-		{
-		ECerr(EC_F_EC_POINT_OCT2POINT, EC_R_INCOMPATIBLE_OBJECTS);
-		return 0;
-		}
-	return group->meth->oct2point(group, point, buf, len, ctx);
-	}
-
+#endif
 
 int EC_POINT_add(const EC_GROUP *group, EC_POINT *r, const EC_POINT *a, const EC_POINT *b, BN_CTX *ctx)
 	{
diff --git a/crypto/ec/ec_mult.c b/crypto/ec/ec_mult.c
index ee422697267e..19f21675fbd8 100644
--- a/crypto/ec/ec_mult.c
+++ b/crypto/ec/ec_mult.c
@@ -226,6 +226,12 @@ static signed char *compute_wNAF(const BIGNUM *scalar, int w, size_t *ret_len)
 		sign = -1;
 		}
 
+	if (scalar->d == NULL || scalar->top == 0)
+		{
+		ECerr(EC_F_COMPUTE_WNAF, ERR_R_INTERNAL_ERROR);
+		goto err;
+		}
+
 	len = BN_num_bits(scalar);
 	r = OPENSSL_malloc(len + 1); /* modified wNAF may be one digit longer than binary representation
 	                              * (*ret_len will be set to the actual length, i.e. at most
@@ -235,12 +241,6 @@ static signed char *compute_wNAF(const BIGNUM *scalar, int w, size_t *ret_len)
 		ECerr(EC_F_COMPUTE_WNAF, ERR_R_MALLOC_FAILURE);
 		goto err;
 		}
-
-	if (scalar->d == NULL || scalar->top == 0)
-		{
-		ECerr(EC_F_COMPUTE_WNAF, ERR_R_INTERNAL_ERROR);
-		goto err;
-		}
 	window_val = scalar->d[0] & mask;
 	j = 0;
 	while ((window_val != 0) || (j + w + 1 < len)) /* if j+w+1 >= len, window_val will not increase */
@@ -421,7 +421,7 @@ int ec_wNAF_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar,
 			if (numblocks > pre_comp->numblocks)
 				numblocks = pre_comp->numblocks;
 
-			pre_points_per_block = 1u << (pre_comp->w - 1);
+			pre_points_per_block = (size_t)1 << (pre_comp->w - 1);
 
 			/* check that pre_comp looks sane */
 			if (pre_comp->num != (pre_comp->numblocks * pre_points_per_block))
@@ -463,7 +463,7 @@ int ec_wNAF_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar,
 
 		bits = i < num ? BN_num_bits(scalars[i]) : BN_num_bits(scalar);
 		wsize[i] = EC_window_bits_for_scalar_size(bits);
-		num_val += 1u << (wsize[i] - 1);
+		num_val += (size_t)1 << (wsize[i] - 1);
 		wNAF[i + 1] = NULL; /* make sure we always have a pivot */
 		wNAF[i] = compute_wNAF((i < num ? scalars[i] : scalar), wsize[i], &wNAF_len[i]);
 		if (wNAF[i] == NULL)
@@ -602,7 +602,7 @@ int ec_wNAF_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar,
 	for (i = 0; i < num + num_scalar; i++)
 		{
 		val_sub[i] = v;
-		for (j = 0; j < (1u << (wsize[i] - 1)); j++)
+		for (j = 0; j < ((size_t)1 << (wsize[i] - 1)); j++)
 			{
 			*v = EC_POINT_new(group);
 			if (*v == NULL) goto err;
@@ -638,7 +638,7 @@ int ec_wNAF_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar,
 		if (wsize[i] > 1)
 			{
 			if (!EC_POINT_dbl(group, tmp, val_sub[i][0], ctx)) goto err;
-			for (j = 1; j < (1u << (wsize[i] - 1)); j++)
+			for (j = 1; j < ((size_t)1 << (wsize[i] - 1)); j++)
 				{
 				if (!EC_POINT_add(group, val_sub[i][j], val_sub[i][j - 1], tmp, ctx)) goto err;
 				}
@@ -822,7 +822,7 @@ int ec_wNAF_precompute_mult(EC_GROUP *group, BN_CTX *ctx)
 
 	numblocks = (bits + blocksize - 1) / blocksize; /* max. number of blocks to use for wNAF splitting */
 	
-	pre_points_per_block = 1u << (w - 1);
+	pre_points_per_block = (size_t)1 << (w - 1);
 	num = pre_points_per_block * numblocks; /* number of points to compute and store */
 
 	points = OPENSSL_malloc(sizeof (EC_POINT*)*(num + 1));
diff --git a/crypto/ec/ec_oct.c b/crypto/ec/ec_oct.c
new file mode 100644
index 000000000000..fd9db0798d3d
--- /dev/null
+++ b/crypto/ec/ec_oct.c
@@ -0,0 +1,199 @@
+/* crypto/ec/ec_lib.c */
+/*
+ * Originally written by Bodo Moeller for the OpenSSL project.
+ */
+/* ====================================================================
+ * Copyright (c) 1998-2003 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+/* ====================================================================
+ * Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
+ * Binary polynomial ECC support in OpenSSL originally developed by 
+ * SUN MICROSYSTEMS, INC., and contributed to the OpenSSL project.
+ */
+
+#include <string.h>
+
+#include <openssl/err.h>
+#include <openssl/opensslv.h>
+
+#include "ec_lcl.h"
+
+int EC_POINT_set_compressed_coordinates_GFp(const EC_GROUP *group, EC_POINT *point,
+	const BIGNUM *x, int y_bit, BN_CTX *ctx)
+	{
+	if (group->meth->point_set_compressed_coordinates == 0
+		&& !(group->meth->flags & EC_FLAGS_DEFAULT_OCT))
+		{
+		ECerr(EC_F_EC_POINT_SET_COMPRESSED_COORDINATES_GFP, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED);
+		return 0;
+		}
+	if (group->meth != point->meth)
+		{
+		ECerr(EC_F_EC_POINT_SET_COMPRESSED_COORDINATES_GFP, EC_R_INCOMPATIBLE_OBJECTS);
+		return 0;
+		}
+	if(group->meth->flags & EC_FLAGS_DEFAULT_OCT)
+		{
+		if (group->meth->field_type == NID_X9_62_prime_field)
+			return ec_GFp_simple_set_compressed_coordinates(
+					group, point, x, y_bit, ctx);
+		else
+#ifdef OPENSSL_NO_EC2M
+			{
+			ECerr(EC_F_EC_POINT_SET_COMPRESSED_COORDINATES_GFP, EC_R_GF2M_NOT_SUPPORTED);
+			return 0;
+			}
+#else
+			return ec_GF2m_simple_set_compressed_coordinates(
+					group, point, x, y_bit, ctx);
+#endif
+		}
+	return group->meth->point_set_compressed_coordinates(group, point, x, y_bit, ctx);
+	}
+
+#ifndef OPENSSL_NO_EC2M
+int EC_POINT_set_compressed_coordinates_GF2m(const EC_GROUP *group, EC_POINT *point,
+	const BIGNUM *x, int y_bit, BN_CTX *ctx)
+	{
+	if (group->meth->point_set_compressed_coordinates == 0
+		&& !(group->meth->flags & EC_FLAGS_DEFAULT_OCT))
+		{
+		ECerr(EC_F_EC_POINT_SET_COMPRESSED_COORDINATES_GF2M, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED);
+		return 0;
+		}
+	if (group->meth != point->meth)
+		{
+		ECerr(EC_F_EC_POINT_SET_COMPRESSED_COORDINATES_GF2M, EC_R_INCOMPATIBLE_OBJECTS);
+		return 0;
+		}
+	if(group->meth->flags & EC_FLAGS_DEFAULT_OCT)
+		{
+		if (group->meth->field_type == NID_X9_62_prime_field)
+			return ec_GFp_simple_set_compressed_coordinates(
+					group, point, x, y_bit, ctx);
+		else
+			return ec_GF2m_simple_set_compressed_coordinates(
+					group, point, x, y_bit, ctx);
+		}
+	return group->meth->point_set_compressed_coordinates(group, point, x, y_bit, ctx);
+	}
+#endif
+
+size_t EC_POINT_point2oct(const EC_GROUP *group, const EC_POINT *point, point_conversion_form_t form,
+        unsigned char *buf, size_t len, BN_CTX *ctx)
+	{
+	if (group->meth->point2oct == 0
+		&& !(group->meth->flags & EC_FLAGS_DEFAULT_OCT))
+		{
+		ECerr(EC_F_EC_POINT_POINT2OCT, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED);
+		return 0;
+		}
+	if (group->meth != point->meth)
+		{
+		ECerr(EC_F_EC_POINT_POINT2OCT, EC_R_INCOMPATIBLE_OBJECTS);
+		return 0;
+		}
+	if(group->meth->flags & EC_FLAGS_DEFAULT_OCT)
+		{
+		if (group->meth->field_type == NID_X9_62_prime_field)
+			return ec_GFp_simple_point2oct(group, point,
+							form, buf, len, ctx);
+		else
+#ifdef OPENSSL_NO_EC2M
+			{
+			ECerr(EC_F_EC_POINT_POINT2OCT, EC_R_GF2M_NOT_SUPPORTED);
+			return 0;
+			}
+#else
+			return ec_GF2m_simple_point2oct(group, point,
+							form, buf, len, ctx);
+#endif
+		}
+			
+	return group->meth->point2oct(group, point, form, buf, len, ctx);
+	}
+
+
+int EC_POINT_oct2point(const EC_GROUP *group, EC_POINT *point,
+        const unsigned char *buf, size_t len, BN_CTX *ctx)
+	{
+	if (group->meth->oct2point == 0
+		&& !(group->meth->flags & EC_FLAGS_DEFAULT_OCT))
+		{
+		ECerr(EC_F_EC_POINT_OCT2POINT, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED);
+		return 0;
+		}
+	if (group->meth != point->meth)
+		{
+		ECerr(EC_F_EC_POINT_OCT2POINT, EC_R_INCOMPATIBLE_OBJECTS);
+		return 0;
+		}
+	if(group->meth->flags & EC_FLAGS_DEFAULT_OCT)
+		{
+		if (group->meth->field_type == NID_X9_62_prime_field)
+			return ec_GFp_simple_oct2point(group, point,
+							buf, len, ctx);
+		else
+#ifdef OPENSSL_NO_EC2M
+			{
+			ECerr(EC_F_EC_POINT_OCT2POINT, EC_R_GF2M_NOT_SUPPORTED);
+			return 0;
+			}
+#else
+			return ec_GF2m_simple_oct2point(group, point,
+							buf, len, ctx);
+#endif
+		}
+	return group->meth->oct2point(group, point, buf, len, ctx);
+	}
+
diff --git a/crypto/ec/ec_pmeth.c b/crypto/ec/ec_pmeth.c
new file mode 100644
index 000000000000..d1ed66c37e76
--- /dev/null
+++ b/crypto/ec/ec_pmeth.c
@@ -0,0 +1,341 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/asn1t.h>
+#include <openssl/x509.h>
+#include <openssl/ec.h>
+#include <openssl/ecdsa.h>
+#include <openssl/evp.h>
+#include "evp_locl.h"
+
+/* EC pkey context structure */
+
+typedef struct
+	{
+	/* Key and paramgen group */
+	EC_GROUP *gen_group;
+	/* message digest */
+	const EVP_MD *md;
+	} EC_PKEY_CTX;
+
+static int pkey_ec_init(EVP_PKEY_CTX *ctx)
+	{
+	EC_PKEY_CTX *dctx;
+	dctx = OPENSSL_malloc(sizeof(EC_PKEY_CTX));
+	if (!dctx)
+		return 0;
+	dctx->gen_group = NULL;
+	dctx->md = NULL;
+
+	ctx->data = dctx;
+
+	return 1;
+	}
+
+static int pkey_ec_copy(EVP_PKEY_CTX *dst, EVP_PKEY_CTX *src)
+	{
+	EC_PKEY_CTX *dctx, *sctx;
+	if (!pkey_ec_init(dst))
+		return 0;
+       	sctx = src->data;
+	dctx = dst->data;
+	if (sctx->gen_group)
+		{
+		dctx->gen_group = EC_GROUP_dup(sctx->gen_group);
+		if (!dctx->gen_group)
+			return 0;
+		}
+	dctx->md = sctx->md;
+	return 1;
+	}
+
+static void pkey_ec_cleanup(EVP_PKEY_CTX *ctx)
+	{
+	EC_PKEY_CTX *dctx = ctx->data;
+	if (dctx)
+		{
+		if (dctx->gen_group)
+			EC_GROUP_free(dctx->gen_group);
+		OPENSSL_free(dctx);
+		}
+	}
+
+static int pkey_ec_sign(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					const unsigned char *tbs, size_t tbslen)
+	{
+	int ret, type;
+	unsigned int sltmp;
+	EC_PKEY_CTX *dctx = ctx->data;
+	EC_KEY *ec = ctx->pkey->pkey.ec;
+
+	if (!sig)
+		{
+		*siglen = ECDSA_size(ec);
+		return 1;
+		}
+	else if(*siglen < (size_t)ECDSA_size(ec))
+		{
+		ECerr(EC_F_PKEY_EC_SIGN, EC_R_BUFFER_TOO_SMALL);
+		return 0;
+		}
+
+	if (dctx->md)
+		type = EVP_MD_type(dctx->md);
+	else
+		type = NID_sha1;
+
+
+	ret = ECDSA_sign(type, tbs, tbslen, sig, &sltmp, ec);
+
+	if (ret <= 0)
+		return ret;
+	*siglen = (size_t)sltmp;
+	return 1;
+	}
+
+static int pkey_ec_verify(EVP_PKEY_CTX *ctx,
+					const unsigned char *sig, size_t siglen,
+					const unsigned char *tbs, size_t tbslen)
+	{
+	int ret, type;
+	EC_PKEY_CTX *dctx = ctx->data;
+	EC_KEY *ec = ctx->pkey->pkey.ec;
+
+	if (dctx->md)
+		type = EVP_MD_type(dctx->md);
+	else
+		type = NID_sha1;
+
+	ret = ECDSA_verify(type, tbs, tbslen, sig, siglen, ec);
+
+	return ret;
+	}
+
+static int pkey_ec_derive(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen)
+	{
+	int ret;
+	size_t outlen;
+	const EC_POINT *pubkey = NULL;
+	if (!ctx->pkey || !ctx->peerkey)
+		{
+		ECerr(EC_F_PKEY_EC_DERIVE, EC_R_KEYS_NOT_SET);
+		return 0;
+		}
+
+	if (!key)
+		{
+		const EC_GROUP *group;
+		group = EC_KEY_get0_group(ctx->pkey->pkey.ec);
+		*keylen = (EC_GROUP_get_degree(group) + 7)/8;
+		return 1;
+		}
+
+	pubkey = EC_KEY_get0_public_key(ctx->peerkey->pkey.ec);
+
+	/* NB: unlike PKS#3 DH, if *outlen is less than maximum size this is
+	 * not an error, the result is truncated.
+	 */
+
+	outlen = *keylen;
+		
+	ret = ECDH_compute_key(key, outlen, pubkey, ctx->pkey->pkey.ec, 0);
+	if (ret < 0)
+		return ret;
+	*keylen = ret;
+	return 1;
+	}
+
+static int pkey_ec_ctrl(EVP_PKEY_CTX *ctx, int type, int p1, void *p2)
+	{
+	EC_PKEY_CTX *dctx = ctx->data;
+	EC_GROUP *group;
+	switch (type)
+		{
+		case EVP_PKEY_CTRL_EC_PARAMGEN_CURVE_NID:
+		group = EC_GROUP_new_by_curve_name(p1);
+		if (group == NULL)
+			{
+			ECerr(EC_F_PKEY_EC_CTRL, EC_R_INVALID_CURVE);
+			return 0;
+			}
+		if (dctx->gen_group)
+			EC_GROUP_free(dctx->gen_group);
+		dctx->gen_group = group;
+		return 1;
+
+		case EVP_PKEY_CTRL_MD:
+		if (EVP_MD_type((const EVP_MD *)p2) != NID_sha1 &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_ecdsa_with_SHA1 &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha224 &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha256 &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha384 &&
+		    EVP_MD_type((const EVP_MD *)p2) != NID_sha512)
+			{
+			ECerr(EC_F_PKEY_EC_CTRL, EC_R_INVALID_DIGEST_TYPE);
+			return 0;
+			}
+		dctx->md = p2;
+		return 1;
+
+		case EVP_PKEY_CTRL_PEER_KEY:
+		/* Default behaviour is OK */
+		case EVP_PKEY_CTRL_DIGESTINIT:
+		case EVP_PKEY_CTRL_PKCS7_SIGN:
+		case EVP_PKEY_CTRL_CMS_SIGN:
+		return 1;
+
+		default:
+		return -2;
+
+		}
+	}
+			
+static int pkey_ec_ctrl_str(EVP_PKEY_CTX *ctx,
+			const char *type, const char *value)
+	{
+	if (!strcmp(type, "ec_paramgen_curve"))
+		{
+		int nid;
+		nid = OBJ_sn2nid(value);
+		if (nid == NID_undef)
+			nid = OBJ_ln2nid(value);
+		if (nid == NID_undef)
+			{
+			ECerr(EC_F_PKEY_EC_CTRL_STR, EC_R_INVALID_CURVE);
+			return 0;
+			}
+		return EVP_PKEY_CTX_set_ec_paramgen_curve_nid(ctx, nid);
+		}
+	return -2;
+	}
+
+static int pkey_ec_paramgen(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey)
+	{
+	EC_KEY *ec = NULL;
+	EC_PKEY_CTX *dctx = ctx->data;
+	int ret = 0;
+	if (dctx->gen_group == NULL)
+		{
+		ECerr(EC_F_PKEY_EC_PARAMGEN, EC_R_NO_PARAMETERS_SET);
+		return 0;
+		}
+	ec = EC_KEY_new();
+	if (!ec)
+		return 0;
+	ret = EC_KEY_set_group(ec, dctx->gen_group);
+	if (ret)
+		EVP_PKEY_assign_EC_KEY(pkey, ec);
+	else
+		EC_KEY_free(ec);
+	return ret;
+	}
+
+static int pkey_ec_keygen(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey)
+	{
+	EC_KEY *ec = NULL;
+	if (ctx->pkey == NULL)
+		{
+		ECerr(EC_F_PKEY_EC_KEYGEN, EC_R_NO_PARAMETERS_SET);
+		return 0;
+		}
+	ec = EC_KEY_new();
+	if (!ec)
+		return 0;
+	EVP_PKEY_assign_EC_KEY(pkey, ec);
+	/* Note: if error return, pkey is freed by parent routine */
+	if (!EVP_PKEY_copy_parameters(pkey, ctx->pkey))
+		return 0;
+	return EC_KEY_generate_key(pkey->pkey.ec);
+	}
+
+const EVP_PKEY_METHOD ec_pkey_meth = 
+	{
+	EVP_PKEY_EC,
+	0,
+	pkey_ec_init,
+	pkey_ec_copy,
+	pkey_ec_cleanup,
+
+	0,
+	pkey_ec_paramgen,
+
+	0,
+	pkey_ec_keygen,
+
+	0,
+	pkey_ec_sign,
+
+	0,
+	pkey_ec_verify,
+
+	0,0,
+
+	0,0,0,0,
+
+	0,0,
+
+	0,0,
+
+	0,
+	pkey_ec_derive,
+
+	pkey_ec_ctrl,
+	pkey_ec_ctrl_str
+
+	};
diff --git a/crypto/ec/eck_prn.c b/crypto/ec/eck_prn.c
new file mode 100644
index 000000000000..06de8f3959d5
--- /dev/null
+++ b/crypto/ec/eck_prn.c
@@ -0,0 +1,392 @@
+/* crypto/ec/eck_prn.c */
+/*
+ * Written by Nils Larsch for the OpenSSL project.
+ */
+/* ====================================================================
+ * Copyright (c) 1998-2005 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+/* ====================================================================
+ * Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
+ * Portions originally developed by SUN MICROSYSTEMS, INC., and 
+ * contributed to the OpenSSL project.
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/evp.h>
+#include <openssl/ec.h>
+#include <openssl/bn.h>
+
+#ifndef OPENSSL_NO_FP_API
+int ECPKParameters_print_fp(FILE *fp, const EC_GROUP *x, int off)
+	{
+	BIO *b;
+	int ret;
+
+	if ((b=BIO_new(BIO_s_file())) == NULL)
+		{
+		ECerr(EC_F_ECPKPARAMETERS_PRINT_FP,ERR_R_BUF_LIB);
+		return(0);
+		}
+	BIO_set_fp(b, fp, BIO_NOCLOSE);
+	ret = ECPKParameters_print(b, x, off);
+	BIO_free(b);
+	return(ret);
+	}
+
+int EC_KEY_print_fp(FILE *fp, const EC_KEY *x, int off)
+	{
+	BIO *b;
+	int ret;
+ 
+	if ((b=BIO_new(BIO_s_file())) == NULL)
+		{
+		ECerr(EC_F_EC_KEY_PRINT_FP, ERR_R_BIO_LIB);
+		return(0);
+		}
+	BIO_set_fp(b, fp, BIO_NOCLOSE);
+	ret = EC_KEY_print(b, x, off);
+	BIO_free(b);
+	return(ret);
+	}
+
+int ECParameters_print_fp(FILE *fp, const EC_KEY *x)
+	{
+	BIO *b;
+	int ret;
+ 
+	if ((b=BIO_new(BIO_s_file())) == NULL)
+		{
+		ECerr(EC_F_ECPARAMETERS_PRINT_FP, ERR_R_BIO_LIB);
+		return(0);
+		}
+	BIO_set_fp(b, fp, BIO_NOCLOSE);
+	ret = ECParameters_print(b, x);
+	BIO_free(b);
+	return(ret);
+	}
+#endif
+
+int EC_KEY_print(BIO *bp, const EC_KEY *x, int off)
+	{
+	EVP_PKEY *pk;
+	int ret;
+	pk = EVP_PKEY_new();
+	if (!pk || !EVP_PKEY_set1_EC_KEY(pk, (EC_KEY *)x))
+		return 0;
+	ret = EVP_PKEY_print_private(bp, pk, off, NULL);
+	EVP_PKEY_free(pk);
+	return ret;
+	}
+
+int ECParameters_print(BIO *bp, const EC_KEY *x)
+	{
+	EVP_PKEY *pk;
+	int ret;
+	pk = EVP_PKEY_new();
+	if (!pk || !EVP_PKEY_set1_EC_KEY(pk, (EC_KEY *)x))
+		return 0;
+	ret = EVP_PKEY_print_params(bp, pk, 4, NULL);
+	EVP_PKEY_free(pk);
+	return ret;
+	}
+
+static int print_bin(BIO *fp, const char *str, const unsigned char *num,
+		size_t len, int off);
+
+int ECPKParameters_print(BIO *bp, const EC_GROUP *x, int off)
+	{
+	unsigned char *buffer=NULL;
+	size_t	buf_len=0, i;
+	int     ret=0, reason=ERR_R_BIO_LIB;
+	BN_CTX  *ctx=NULL;
+	const EC_POINT *point=NULL;
+	BIGNUM	*p=NULL, *a=NULL, *b=NULL, *gen=NULL,
+		*order=NULL, *cofactor=NULL;
+	const unsigned char *seed;
+	size_t	seed_len=0;
+	
+	static const char *gen_compressed = "Generator (compressed):";
+	static const char *gen_uncompressed = "Generator (uncompressed):";
+	static const char *gen_hybrid = "Generator (hybrid):";
+ 
+	if (!x)
+		{
+		reason = ERR_R_PASSED_NULL_PARAMETER;
+		goto err;
+		}
+
+	ctx = BN_CTX_new();
+	if (ctx == NULL)
+		{
+		reason = ERR_R_MALLOC_FAILURE;
+		goto err;
+		}
+
+	if (EC_GROUP_get_asn1_flag(x))
+		{
+		/* the curve parameter are given by an asn1 OID */
+		int nid;
+
+		if (!BIO_indent(bp, off, 128))
+			goto err;
+
+		nid = EC_GROUP_get_curve_name(x);
+		if (nid == 0)
+			goto err;
+
+		if (BIO_printf(bp, "ASN1 OID: %s", OBJ_nid2sn(nid)) <= 0)
+			goto err;
+		if (BIO_printf(bp, "\n") <= 0)
+			goto err;
+		}
+	else
+		{
+		/* explicit parameters */
+		int is_char_two = 0;
+		point_conversion_form_t form;
+		int tmp_nid = EC_METHOD_get_field_type(EC_GROUP_method_of(x));
+
+		if (tmp_nid == NID_X9_62_characteristic_two_field)
+			is_char_two = 1;
+
+		if ((p = BN_new()) == NULL || (a = BN_new()) == NULL ||
+			(b = BN_new()) == NULL || (order = BN_new()) == NULL ||
+			(cofactor = BN_new()) == NULL)
+			{
+			reason = ERR_R_MALLOC_FAILURE;
+			goto err;
+			}
+#ifndef OPENSSL_NO_EC2M
+		if (is_char_two)
+			{
+			if (!EC_GROUP_get_curve_GF2m(x, p, a, b, ctx))
+				{
+				reason = ERR_R_EC_LIB;
+				goto err;
+				}
+			}
+		else /* prime field */
+#endif
+			{
+			if (!EC_GROUP_get_curve_GFp(x, p, a, b, ctx))
+				{
+				reason = ERR_R_EC_LIB;
+				goto err;
+				}
+			}
+
+		if ((point = EC_GROUP_get0_generator(x)) == NULL)
+			{
+			reason = ERR_R_EC_LIB;
+			goto err;
+			}
+		if (!EC_GROUP_get_order(x, order, NULL) || 
+            		!EC_GROUP_get_cofactor(x, cofactor, NULL))
+			{
+			reason = ERR_R_EC_LIB;
+			goto err;
+			}
+		
+		form = EC_GROUP_get_point_conversion_form(x);
+
+		if ((gen = EC_POINT_point2bn(x, point, 
+				form, NULL, ctx)) == NULL)
+			{
+			reason = ERR_R_EC_LIB;
+			goto err;
+			}
+
+		buf_len = (size_t)BN_num_bytes(p);
+		if (buf_len < (i = (size_t)BN_num_bytes(a)))
+			buf_len = i;
+		if (buf_len < (i = (size_t)BN_num_bytes(b)))
+			buf_len = i;
+		if (buf_len < (i = (size_t)BN_num_bytes(gen)))
+			buf_len = i;
+		if (buf_len < (i = (size_t)BN_num_bytes(order)))
+			buf_len = i;
+		if (buf_len < (i = (size_t)BN_num_bytes(cofactor))) 
+			buf_len = i;
+
+		if ((seed = EC_GROUP_get0_seed(x)) != NULL)
+			seed_len = EC_GROUP_get_seed_len(x);
+
+		buf_len += 10;
+		if ((buffer = OPENSSL_malloc(buf_len)) == NULL)
+			{
+			reason = ERR_R_MALLOC_FAILURE;
+			goto err;
+			}
+
+		if (!BIO_indent(bp, off, 128))
+			goto err;
+
+		/* print the 'short name' of the field type */
+		if (BIO_printf(bp, "Field Type: %s\n", OBJ_nid2sn(tmp_nid))
+			<= 0)
+			goto err;  
+
+		if (is_char_two)
+			{
+			/* print the 'short name' of the base type OID */
+			int basis_type = EC_GROUP_get_basis_type(x);
+			if (basis_type == 0)
+				goto err;
+
+			if (!BIO_indent(bp, off, 128))
+				goto err;
+
+			if (BIO_printf(bp, "Basis Type: %s\n", 
+				OBJ_nid2sn(basis_type)) <= 0)
+				goto err;
+
+			/* print the polynomial */
+			if ((p != NULL) && !ASN1_bn_print(bp, "Polynomial:", p, buffer,
+				off))
+				goto err;
+			}
+		else
+			{
+			if ((p != NULL) && !ASN1_bn_print(bp, "Prime:", p, buffer,off))
+				goto err;
+			}
+		if ((a != NULL) && !ASN1_bn_print(bp, "A:   ", a, buffer, off)) 
+			goto err;
+		if ((b != NULL) && !ASN1_bn_print(bp, "B:   ", b, buffer, off))
+			goto err;
+		if (form == POINT_CONVERSION_COMPRESSED)
+			{
+			if ((gen != NULL) && !ASN1_bn_print(bp, gen_compressed, gen,
+				buffer, off))
+				goto err;
+			}
+		else if (form == POINT_CONVERSION_UNCOMPRESSED)
+			{
+			if ((gen != NULL) && !ASN1_bn_print(bp, gen_uncompressed, gen,
+				buffer, off))
+				goto err;
+			}
+		else /* form == POINT_CONVERSION_HYBRID */
+			{
+			if ((gen != NULL) && !ASN1_bn_print(bp, gen_hybrid, gen,
+				buffer, off))
+				goto err;
+			}
+		if ((order != NULL) && !ASN1_bn_print(bp, "Order: ", order, 
+			buffer, off)) goto err;
+		if ((cofactor != NULL) && !ASN1_bn_print(bp, "Cofactor: ", cofactor, 
+			buffer, off)) goto err;
+		if (seed && !print_bin(bp, "Seed:", seed, seed_len, off))
+			goto err;
+		}
+	ret=1;
+err:
+	if (!ret)
+ 		ECerr(EC_F_ECPKPARAMETERS_PRINT, reason);
+	if (p) 
+		BN_free(p);
+	if (a) 
+		BN_free(a);
+	if (b)
+		BN_free(b);
+	if (gen)
+		BN_free(gen);
+	if (order)
+		BN_free(order);
+	if (cofactor)
+		BN_free(cofactor);
+	if (ctx)
+		BN_CTX_free(ctx);
+	if (buffer != NULL) 
+		OPENSSL_free(buffer);
+	return(ret);	
+	}
+
+static int print_bin(BIO *fp, const char *name, const unsigned char *buf,
+		size_t len, int off)
+	{
+	size_t i;
+	char str[128];
+
+	if (buf == NULL)
+		return 1;
+	if (off)
+		{
+		if (off > 128)
+			off=128;
+		memset(str,' ',off);
+		if (BIO_write(fp, str, off) <= 0)
+			return 0;
+		}
+
+	if (BIO_printf(fp,"%s", name) <= 0)
+		return 0;
+
+	for (i=0; i<len; i++)
+		{
+		if ((i%15) == 0)
+			{
+			str[0]='\n';
+			memset(&(str[1]),' ',off+4);
+			if (BIO_write(fp, str, off+1+4) <= 0)
+				return 0;
+			}
+		if (BIO_printf(fp,"%02x%s",buf[i],((i+1) == len)?"":":") <= 0)
+			return 0;
+		}
+	if (BIO_write(fp,"\n",1) <= 0)
+		return 0;
+
+	return 1;
+	}
diff --git a/crypto/ec/ecp_mont.c b/crypto/ec/ecp_mont.c
index 9fc4a466a59f..079e47431b4f 100644
--- a/crypto/ec/ecp_mont.c
+++ b/crypto/ec/ecp_mont.c
@@ -63,12 +63,20 @@
 
 #include <openssl/err.h>
 
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
+
 #include "ec_lcl.h"
 
 
 const EC_METHOD *EC_GFp_mont_method(void)
 	{
+#ifdef OPENSSL_FIPS
+	return fips_ec_gfp_mont_method();
+#else
 	static const EC_METHOD ret = {
+		EC_FLAGS_DEFAULT_OCT,
 		NID_X9_62_prime_field,
 		ec_GFp_mont_group_init,
 		ec_GFp_mont_group_finish,
@@ -87,9 +95,7 @@ const EC_METHOD *EC_GFp_mont_method(void)
 		ec_GFp_simple_get_Jprojective_coordinates_GFp,
 		ec_GFp_simple_point_set_affine_coordinates,
 		ec_GFp_simple_point_get_affine_coordinates,
-		ec_GFp_simple_set_compressed_coordinates,
-		ec_GFp_simple_point2oct,
-		ec_GFp_simple_oct2point,
+		0,0,0,
 		ec_GFp_simple_add,
 		ec_GFp_simple_dbl,
 		ec_GFp_simple_invert,
@@ -108,7 +114,9 @@ const EC_METHOD *EC_GFp_mont_method(void)
 		ec_GFp_mont_field_decode,
 		ec_GFp_mont_field_set_to_one };
 
+
 	return &ret;
+#endif
 	}
 
 
diff --git a/crypto/ec/ecp_nist.c b/crypto/ec/ecp_nist.c
index 71893d5eaba1..aad2d5f44389 100644
--- a/crypto/ec/ecp_nist.c
+++ b/crypto/ec/ecp_nist.c
@@ -67,9 +67,17 @@
 #include <openssl/obj_mac.h>
 #include "ec_lcl.h"
 
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
+
 const EC_METHOD *EC_GFp_nist_method(void)
 	{
+#ifdef OPENSSL_FIPS
+	return fips_ec_gfp_nist_method();
+#else
 	static const EC_METHOD ret = {
+		EC_FLAGS_DEFAULT_OCT,
 		NID_X9_62_prime_field,
 		ec_GFp_simple_group_init,
 		ec_GFp_simple_group_finish,
@@ -88,9 +96,7 @@ const EC_METHOD *EC_GFp_nist_method(void)
 		ec_GFp_simple_get_Jprojective_coordinates_GFp,
 		ec_GFp_simple_point_set_affine_coordinates,
 		ec_GFp_simple_point_get_affine_coordinates,
-		ec_GFp_simple_set_compressed_coordinates,
-		ec_GFp_simple_point2oct,
-		ec_GFp_simple_oct2point,
+		0,0,0,
 		ec_GFp_simple_add,
 		ec_GFp_simple_dbl,
 		ec_GFp_simple_invert,
@@ -110,11 +116,8 @@ const EC_METHOD *EC_GFp_nist_method(void)
 		0 /* field_set_to_one */ };
 
 	return &ret;
-	}
-
-#if BN_BITS2 == 64
-#define	NO_32_BIT_TYPE
 #endif
+	}
 
 int ec_GFp_nist_group_copy(EC_GROUP *dest, const EC_GROUP *src)
 	{
@@ -139,34 +142,12 @@ int ec_GFp_nist_group_set_curve(EC_GROUP *group, const BIGNUM *p,
 	if (BN_ucmp(BN_get0_nist_prime_192(), p) == 0)
 		group->field_mod_func = BN_nist_mod_192;
 	else if (BN_ucmp(BN_get0_nist_prime_224(), p) == 0)
-		{
-#ifndef NO_32_BIT_TYPE
 		group->field_mod_func = BN_nist_mod_224;
-#else
-		ECerr(EC_F_EC_GFP_NIST_GROUP_SET_CURVE, EC_R_NOT_A_SUPPORTED_NIST_PRIME);
-		goto err;
-#endif
-		}
 	else if (BN_ucmp(BN_get0_nist_prime_256(), p) == 0)
-		{
-#ifndef NO_32_BIT_TYPE
 		group->field_mod_func = BN_nist_mod_256;
-#else
-		ECerr(EC_F_EC_GFP_NIST_GROUP_SET_CURVE, EC_R_NOT_A_SUPPORTED_NIST_PRIME);
-		goto err;
-#endif
-		}
 	else if (BN_ucmp(BN_get0_nist_prime_384(), p) == 0)
-		{
-#ifndef NO_32_BIT_TYPE
 		group->field_mod_func = BN_nist_mod_384;
-#else
-		ECerr(EC_F_EC_GFP_NIST_GROUP_SET_CURVE, EC_R_NOT_A_SUPPORTED_NIST_PRIME);
-		goto err;
-#endif
-		}
 	else if (BN_ucmp(BN_get0_nist_prime_521(), p) == 0)
-		/* this one works in the NO_32_BIT_TYPE case */
 		group->field_mod_func = BN_nist_mod_521;
 	else
 		{
diff --git a/crypto/ec/ecp_nistp224.c b/crypto/ec/ecp_nistp224.c
new file mode 100644
index 000000000000..b5ff56c25277
--- /dev/null
+++ b/crypto/ec/ecp_nistp224.c
@@ -0,0 +1,1658 @@
+/* crypto/ec/ecp_nistp224.c */
+/*
+ * Written by Emilia Kasper (Google) for the OpenSSL project.
+ */
+/* Copyright 2011 Google Inc.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ *
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ *  Unless required by applicable law or agreed to in writing, software
+ *  distributed under the License is distributed on an "AS IS" BASIS,
+ *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ *  See the License for the specific language governing permissions and
+ *  limitations under the License.
+ */
+
+/*
+ * A 64-bit implementation of the NIST P-224 elliptic curve point multiplication
+ *
+ * Inspired by Daniel J. Bernstein's public domain nistp224 implementation
+ * and Adam Langley's public domain 64-bit C implementation of curve25519
+ */
+
+#include <openssl/opensslconf.h>
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+
+#ifndef OPENSSL_SYS_VMS
+#include <stdint.h>
+#else
+#include <inttypes.h>
+#endif
+
+#include <string.h>
+#include <openssl/err.h>
+#include "ec_lcl.h"
+
+#if defined(__GNUC__) && (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1))
+  /* even with gcc, the typedef won't work for 32-bit platforms */
+  typedef __uint128_t uint128_t; /* nonstandard; implemented by gcc on 64-bit platforms */
+#else
+  #error "Need GCC 3.1 or later to define type uint128_t"
+#endif
+
+typedef uint8_t u8;
+typedef uint64_t u64;
+typedef int64_t s64;
+
+
+/******************************************************************************/
+/*		    INTERNAL REPRESENTATION OF FIELD ELEMENTS
+ *
+ * Field elements are represented as a_0 + 2^56*a_1 + 2^112*a_2 + 2^168*a_3
+ * using 64-bit coefficients called 'limbs',
+ * and sometimes (for multiplication results) as
+ * b_0 + 2^56*b_1 + 2^112*b_2 + 2^168*b_3 + 2^224*b_4 + 2^280*b_5 + 2^336*b_6
+ * using 128-bit coefficients called 'widelimbs'.
+ * A 4-limb representation is an 'felem';
+ * a 7-widelimb representation is a 'widefelem'.
+ * Even within felems, bits of adjacent limbs overlap, and we don't always
+ * reduce the representations: we ensure that inputs to each felem
+ * multiplication satisfy a_i < 2^60, so outputs satisfy b_i < 4*2^60*2^60,
+ * and fit into a 128-bit word without overflow. The coefficients are then
+ * again partially reduced to obtain an felem satisfying a_i < 2^57.
+ * We only reduce to the unique minimal representation at the end of the
+ * computation.
+ */
+
+typedef uint64_t limb;
+typedef uint128_t widelimb;
+
+typedef limb felem[4];
+typedef widelimb widefelem[7];
+
+/* Field element represented as a byte arrary.
+ * 28*8 = 224 bits is also the group order size for the elliptic curve,
+ * and we also use this type for scalars for point multiplication.
+  */
+typedef u8 felem_bytearray[28];
+
+static const felem_bytearray nistp224_curve_params[5] = {
+	{0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,    /* p */
+	 0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0x00,0x00,0x00,0x00,
+	 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01},
+	{0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,    /* a */
+	 0xFF,0xFF,0xFF,0xFF,0xFF,0xFE,0xFF,0xFF,0xFF,0xFF,
+	 0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFE},
+	{0xB4,0x05,0x0A,0x85,0x0C,0x04,0xB3,0xAB,0xF5,0x41,    /* b */
+	 0x32,0x56,0x50,0x44,0xB0,0xB7,0xD7,0xBF,0xD8,0xBA,
+	 0x27,0x0B,0x39,0x43,0x23,0x55,0xFF,0xB4},
+	{0xB7,0x0E,0x0C,0xBD,0x6B,0xB4,0xBF,0x7F,0x32,0x13,    /* x */
+	 0x90,0xB9,0x4A,0x03,0xC1,0xD3,0x56,0xC2,0x11,0x22,
+	 0x34,0x32,0x80,0xD6,0x11,0x5C,0x1D,0x21},
+	{0xbd,0x37,0x63,0x88,0xb5,0xf7,0x23,0xfb,0x4c,0x22,    /* y */
+	 0xdf,0xe6,0xcd,0x43,0x75,0xa0,0x5a,0x07,0x47,0x64,
+	 0x44,0xd5,0x81,0x99,0x85,0x00,0x7e,0x34}
+};
+
+/* Precomputed multiples of the standard generator
+ * Points are given in coordinates (X, Y, Z) where Z normally is 1
+ * (0 for the point at infinity).
+ * For each field element, slice a_0 is word 0, etc.
+ *
+ * The table has 2 * 16 elements, starting with the following:
+ * index | bits    | point
+ * ------+---------+------------------------------
+ *     0 | 0 0 0 0 | 0G
+ *     1 | 0 0 0 1 | 1G
+ *     2 | 0 0 1 0 | 2^56G
+ *     3 | 0 0 1 1 | (2^56 + 1)G
+ *     4 | 0 1 0 0 | 2^112G
+ *     5 | 0 1 0 1 | (2^112 + 1)G
+ *     6 | 0 1 1 0 | (2^112 + 2^56)G
+ *     7 | 0 1 1 1 | (2^112 + 2^56 + 1)G
+ *     8 | 1 0 0 0 | 2^168G
+ *     9 | 1 0 0 1 | (2^168 + 1)G
+ *    10 | 1 0 1 0 | (2^168 + 2^56)G
+ *    11 | 1 0 1 1 | (2^168 + 2^56 + 1)G
+ *    12 | 1 1 0 0 | (2^168 + 2^112)G
+ *    13 | 1 1 0 1 | (2^168 + 2^112 + 1)G
+ *    14 | 1 1 1 0 | (2^168 + 2^112 + 2^56)G
+ *    15 | 1 1 1 1 | (2^168 + 2^112 + 2^56 + 1)G
+ * followed by a copy of this with each element multiplied by 2^28.
+ *
+ * The reason for this is so that we can clock bits into four different
+ * locations when doing simple scalar multiplies against the base point,
+ * and then another four locations using the second 16 elements.
+ */
+static const felem gmul[2][16][3] =
+{{{{0, 0, 0, 0},
+   {0, 0, 0, 0},
+   {0, 0, 0, 0}},
+  {{0x3280d6115c1d21, 0xc1d356c2112234, 0x7f321390b94a03, 0xb70e0cbd6bb4bf},
+   {0xd5819985007e34, 0x75a05a07476444, 0xfb4c22dfe6cd43, 0xbd376388b5f723},
+   {1, 0, 0, 0}},
+  {{0xfd9675666ebbe9, 0xbca7664d40ce5e, 0x2242df8d8a2a43, 0x1f49bbb0f99bc5},
+   {0x29e0b892dc9c43, 0xece8608436e662, 0xdc858f185310d0, 0x9812dd4eb8d321},
+   {1, 0, 0, 0}},
+  {{0x6d3e678d5d8eb8, 0x559eed1cb362f1, 0x16e9a3bbce8a3f, 0xeedcccd8c2a748},
+   {0xf19f90ed50266d, 0xabf2b4bf65f9df, 0x313865468fafec, 0x5cb379ba910a17},
+   {1, 0, 0, 0}},
+  {{0x0641966cab26e3, 0x91fb2991fab0a0, 0xefec27a4e13a0b, 0x0499aa8a5f8ebe},
+   {0x7510407766af5d, 0x84d929610d5450, 0x81d77aae82f706, 0x6916f6d4338c5b},
+   {1, 0, 0, 0}},
+  {{0xea95ac3b1f15c6, 0x086000905e82d4, 0xdd323ae4d1c8b1, 0x932b56be7685a3},
+   {0x9ef93dea25dbbf, 0x41665960f390f0, 0xfdec76dbe2a8a7, 0x523e80f019062a},
+   {1, 0, 0, 0}},
+  {{0x822fdd26732c73, 0xa01c83531b5d0f, 0x363f37347c1ba4, 0xc391b45c84725c},
+   {0xbbd5e1b2d6ad24, 0xddfbcde19dfaec, 0xc393da7e222a7f, 0x1efb7890ede244},
+   {1, 0, 0, 0}},
+  {{0x4c9e90ca217da1, 0xd11beca79159bb, 0xff8d33c2c98b7c, 0x2610b39409f849},
+   {0x44d1352ac64da0, 0xcdbb7b2c46b4fb, 0x966c079b753c89, 0xfe67e4e820b112},
+   {1, 0, 0, 0}},
+  {{0xe28cae2df5312d, 0xc71b61d16f5c6e, 0x79b7619a3e7c4c, 0x05c73240899b47},
+   {0x9f7f6382c73e3a, 0x18615165c56bda, 0x641fab2116fd56, 0x72855882b08394},
+   {1, 0, 0, 0}},
+  {{0x0469182f161c09, 0x74a98ca8d00fb5, 0xb89da93489a3e0, 0x41c98768fb0c1d},
+   {0xe5ea05fb32da81, 0x3dce9ffbca6855, 0x1cfe2d3fbf59e6, 0x0e5e03408738a7},
+   {1, 0, 0, 0}},
+  {{0xdab22b2333e87f, 0x4430137a5dd2f6, 0xe03ab9f738beb8, 0xcb0c5d0dc34f24},
+   {0x764a7df0c8fda5, 0x185ba5c3fa2044, 0x9281d688bcbe50, 0xc40331df893881},
+   {1, 0, 0, 0}},
+  {{0xb89530796f0f60, 0xade92bd26909a3, 0x1a0c83fb4884da, 0x1765bf22a5a984},
+   {0x772a9ee75db09e, 0x23bc6c67cec16f, 0x4c1edba8b14e2f, 0xe2a215d9611369},
+   {1, 0, 0, 0}},
+  {{0x571e509fb5efb3, 0xade88696410552, 0xc8ae85fada74fe, 0x6c7e4be83bbde3},
+   {0xff9f51160f4652, 0xb47ce2495a6539, 0xa2946c53b582f4, 0x286d2db3ee9a60},
+   {1, 0, 0, 0}},
+  {{0x40bbd5081a44af, 0x0995183b13926c, 0xbcefba6f47f6d0, 0x215619e9cc0057},
+   {0x8bc94d3b0df45e, 0xf11c54a3694f6f, 0x8631b93cdfe8b5, 0xe7e3f4b0982db9},
+   {1, 0, 0, 0}},
+  {{0xb17048ab3e1c7b, 0xac38f36ff8a1d8, 0x1c29819435d2c6, 0xc813132f4c07e9},
+   {0x2891425503b11f, 0x08781030579fea, 0xf5426ba5cc9674, 0x1e28ebf18562bc},
+   {1, 0, 0, 0}},
+  {{0x9f31997cc864eb, 0x06cd91d28b5e4c, 0xff17036691a973, 0xf1aef351497c58},
+   {0xdd1f2d600564ff, 0xdead073b1402db, 0x74a684435bd693, 0xeea7471f962558},
+   {1, 0, 0, 0}}},
+ {{{0, 0, 0, 0},
+   {0, 0, 0, 0},
+   {0, 0, 0, 0}},
+  {{0x9665266dddf554, 0x9613d78b60ef2d, 0xce27a34cdba417, 0xd35ab74d6afc31},
+   {0x85ccdd22deb15e, 0x2137e5783a6aab, 0xa141cffd8c93c6, 0x355a1830e90f2d},
+   {1, 0, 0, 0}},
+  {{0x1a494eadaade65, 0xd6da4da77fe53c, 0xe7992996abec86, 0x65c3553c6090e3},
+   {0xfa610b1fb09346, 0xf1c6540b8a4aaf, 0xc51a13ccd3cbab, 0x02995b1b18c28a},
+   {1, 0, 0, 0}},
+  {{0x7874568e7295ef, 0x86b419fbe38d04, 0xdc0690a7550d9a, 0xd3966a44beac33},
+   {0x2b7280ec29132f, 0xbeaa3b6a032df3, 0xdc7dd88ae41200, 0xd25e2513e3a100},
+   {1, 0, 0, 0}},
+  {{0x924857eb2efafd, 0xac2bce41223190, 0x8edaa1445553fc, 0x825800fd3562d5},
+   {0x8d79148ea96621, 0x23a01c3dd9ed8d, 0xaf8b219f9416b5, 0xd8db0cc277daea},
+   {1, 0, 0, 0}},
+  {{0x76a9c3b1a700f0, 0xe9acd29bc7e691, 0x69212d1a6b0327, 0x6322e97fe154be},
+   {0x469fc5465d62aa, 0x8d41ed18883b05, 0x1f8eae66c52b88, 0xe4fcbe9325be51},
+   {1, 0, 0, 0}},
+  {{0x825fdf583cac16, 0x020b857c7b023a, 0x683c17744b0165, 0x14ffd0a2daf2f1},
+   {0x323b36184218f9, 0x4944ec4e3b47d4, 0xc15b3080841acf, 0x0bced4b01a28bb},
+   {1, 0, 0, 0}},
+  {{0x92ac22230df5c4, 0x52f33b4063eda8, 0xcb3f19870c0c93, 0x40064f2ba65233},
+   {0xfe16f0924f8992, 0x012da25af5b517, 0x1a57bb24f723a6, 0x06f8bc76760def},
+   {1, 0, 0, 0}},
+  {{0x4a7084f7817cb9, 0xbcab0738ee9a78, 0x3ec11e11d9c326, 0xdc0fe90e0f1aae},
+   {0xcf639ea5f98390, 0x5c350aa22ffb74, 0x9afae98a4047b7, 0x956ec2d617fc45},
+   {1, 0, 0, 0}},
+  {{0x4306d648c1be6a, 0x9247cd8bc9a462, 0xf5595e377d2f2e, 0xbd1c3caff1a52e},
+   {0x045e14472409d0, 0x29f3e17078f773, 0x745a602b2d4f7d, 0x191837685cdfbb},
+   {1, 0, 0, 0}},
+  {{0x5b6ee254a8cb79, 0x4953433f5e7026, 0xe21faeb1d1def4, 0xc4c225785c09de},
+   {0x307ce7bba1e518, 0x31b125b1036db8, 0x47e91868839e8f, 0xc765866e33b9f3},
+   {1, 0, 0, 0}},
+  {{0x3bfece24f96906, 0x4794da641e5093, 0xde5df64f95db26, 0x297ecd89714b05},
+   {0x701bd3ebb2c3aa, 0x7073b4f53cb1d5, 0x13c5665658af16, 0x9895089d66fe58},
+   {1, 0, 0, 0}},
+  {{0x0fef05f78c4790, 0x2d773633b05d2e, 0x94229c3a951c94, 0xbbbd70df4911bb},
+   {0xb2c6963d2c1168, 0x105f47a72b0d73, 0x9fdf6111614080, 0x7b7e94b39e67b0},
+   {1, 0, 0, 0}},
+  {{0xad1a7d6efbe2b3, 0xf012482c0da69d, 0x6b3bdf12438345, 0x40d7558d7aa4d9},
+   {0x8a09fffb5c6d3d, 0x9a356e5d9ffd38, 0x5973f15f4f9b1c, 0xdcd5f59f63c3ea},
+   {1, 0, 0, 0}},
+  {{0xacf39f4c5ca7ab, 0x4c8071cc5fd737, 0xc64e3602cd1184, 0x0acd4644c9abba},
+   {0x6c011a36d8bf6e, 0xfecd87ba24e32a, 0x19f6f56574fad8, 0x050b204ced9405},
+   {1, 0, 0, 0}},
+  {{0xed4f1cae7d9a96, 0x5ceef7ad94c40a, 0x778e4a3bf3ef9b, 0x7405783dc3b55e},
+   {0x32477c61b6e8c6, 0xb46a97570f018b, 0x91176d0a7e95d1, 0x3df90fbc4c7d0e},
+   {1, 0, 0, 0}}}};
+
+/* Precomputation for the group generator. */
+typedef struct {
+	felem g_pre_comp[2][16][3];
+	int references;
+} NISTP224_PRE_COMP;
+
+const EC_METHOD *EC_GFp_nistp224_method(void)
+	{
+	static const EC_METHOD ret = {
+		EC_FLAGS_DEFAULT_OCT,
+		NID_X9_62_prime_field,
+		ec_GFp_nistp224_group_init,
+		ec_GFp_simple_group_finish,
+		ec_GFp_simple_group_clear_finish,
+		ec_GFp_nist_group_copy,
+		ec_GFp_nistp224_group_set_curve,
+		ec_GFp_simple_group_get_curve,
+		ec_GFp_simple_group_get_degree,
+		ec_GFp_simple_group_check_discriminant,
+		ec_GFp_simple_point_init,
+		ec_GFp_simple_point_finish,
+		ec_GFp_simple_point_clear_finish,
+		ec_GFp_simple_point_copy,
+		ec_GFp_simple_point_set_to_infinity,
+		ec_GFp_simple_set_Jprojective_coordinates_GFp,
+		ec_GFp_simple_get_Jprojective_coordinates_GFp,
+		ec_GFp_simple_point_set_affine_coordinates,
+		ec_GFp_nistp224_point_get_affine_coordinates,
+		0 /* point_set_compressed_coordinates */,
+		0 /* point2oct */,
+		0 /* oct2point */,
+		ec_GFp_simple_add,
+		ec_GFp_simple_dbl,
+		ec_GFp_simple_invert,
+		ec_GFp_simple_is_at_infinity,
+		ec_GFp_simple_is_on_curve,
+		ec_GFp_simple_cmp,
+		ec_GFp_simple_make_affine,
+		ec_GFp_simple_points_make_affine,
+		ec_GFp_nistp224_points_mul,
+		ec_GFp_nistp224_precompute_mult,
+		ec_GFp_nistp224_have_precompute_mult,
+		ec_GFp_nist_field_mul,
+		ec_GFp_nist_field_sqr,
+		0 /* field_div */,
+		0 /* field_encode */,
+		0 /* field_decode */,
+		0 /* field_set_to_one */ };
+
+	return &ret;
+	}
+
+/* Helper functions to convert field elements to/from internal representation */
+static void bin28_to_felem(felem out, const u8 in[28])
+	{
+	out[0] = *((const uint64_t *)(in)) & 0x00ffffffffffffff;
+	out[1] = (*((const uint64_t *)(in+7))) & 0x00ffffffffffffff;
+	out[2] = (*((const uint64_t *)(in+14))) & 0x00ffffffffffffff;
+	out[3] = (*((const uint64_t *)(in+21))) & 0x00ffffffffffffff;
+	}
+
+static void felem_to_bin28(u8 out[28], const felem in)
+	{
+	unsigned i;
+	for (i = 0; i < 7; ++i)
+		{
+		out[i]	  = in[0]>>(8*i);
+		out[i+7]  = in[1]>>(8*i);
+		out[i+14] = in[2]>>(8*i);
+		out[i+21] = in[3]>>(8*i);
+		}
+	}
+
+/* To preserve endianness when using BN_bn2bin and BN_bin2bn */
+static void flip_endian(u8 *out, const u8 *in, unsigned len)
+	{
+	unsigned i;
+	for (i = 0; i < len; ++i)
+		out[i] = in[len-1-i];
+	}
+
+/* From OpenSSL BIGNUM to internal representation */
+static int BN_to_felem(felem out, const BIGNUM *bn)
+	{
+	felem_bytearray b_in;
+	felem_bytearray b_out;
+	unsigned num_bytes;
+
+	/* BN_bn2bin eats leading zeroes */
+	memset(b_out, 0, sizeof b_out);
+	num_bytes = BN_num_bytes(bn);
+	if (num_bytes > sizeof b_out)
+		{
+		ECerr(EC_F_BN_TO_FELEM, EC_R_BIGNUM_OUT_OF_RANGE);
+		return 0;
+		}
+	if (BN_is_negative(bn))
+		{
+		ECerr(EC_F_BN_TO_FELEM, EC_R_BIGNUM_OUT_OF_RANGE);
+		return 0;
+		}
+	num_bytes = BN_bn2bin(bn, b_in);
+	flip_endian(b_out, b_in, num_bytes);
+	bin28_to_felem(out, b_out);
+	return 1;
+	}
+
+/* From internal representation to OpenSSL BIGNUM */
+static BIGNUM *felem_to_BN(BIGNUM *out, const felem in)
+	{
+	felem_bytearray b_in, b_out;
+	felem_to_bin28(b_in, in);
+	flip_endian(b_out, b_in, sizeof b_out);
+	return BN_bin2bn(b_out, sizeof b_out, out);
+	}
+
+/******************************************************************************/
+/*				FIELD OPERATIONS
+ *
+ * Field operations, using the internal representation of field elements.
+ * NB! These operations are specific to our point multiplication and cannot be
+ * expected to be correct in general - e.g., multiplication with a large scalar
+ * will cause an overflow.
+ *
+ */
+
+static void felem_one(felem out)
+	{
+	out[0] = 1;
+	out[1] = 0;
+	out[2] = 0;
+	out[3] = 0;
+	}
+
+static void felem_assign(felem out, const felem in)
+	{
+	out[0] = in[0];
+	out[1] = in[1];
+	out[2] = in[2];
+	out[3] = in[3];
+	}
+
+/* Sum two field elements: out += in */
+static void felem_sum(felem out, const felem in)
+	{
+	out[0] += in[0];
+	out[1] += in[1];
+	out[2] += in[2];
+	out[3] += in[3];
+	}
+
+/* Get negative value: out = -in */
+/* Assumes in[i] < 2^57 */
+static void felem_neg(felem out, const felem in)
+	{
+	static const limb two58p2 = (((limb) 1) << 58) + (((limb) 1) << 2);
+	static const limb two58m2 = (((limb) 1) << 58) - (((limb) 1) << 2);
+	static const limb two58m42m2 = (((limb) 1) << 58) -
+	    (((limb) 1) << 42) - (((limb) 1) << 2);
+
+	/* Set to 0 mod 2^224-2^96+1 to ensure out > in */
+	out[0] = two58p2 - in[0];
+	out[1] = two58m42m2 - in[1];
+	out[2] = two58m2 - in[2];
+	out[3] = two58m2 - in[3];
+	}
+
+/* Subtract field elements: out -= in */
+/* Assumes in[i] < 2^57 */
+static void felem_diff(felem out, const felem in)
+	{
+	static const limb two58p2 = (((limb) 1) << 58) + (((limb) 1) << 2);
+	static const limb two58m2 = (((limb) 1) << 58) - (((limb) 1) << 2);
+	static const limb two58m42m2 = (((limb) 1) << 58) -
+	    (((limb) 1) << 42) - (((limb) 1) << 2);
+
+	/* Add 0 mod 2^224-2^96+1 to ensure out > in */
+	out[0] += two58p2;
+	out[1] += two58m42m2;
+	out[2] += two58m2;
+	out[3] += two58m2;
+
+	out[0] -= in[0];
+	out[1] -= in[1];
+	out[2] -= in[2];
+	out[3] -= in[3];
+	}
+
+/* Subtract in unreduced 128-bit mode: out -= in */
+/* Assumes in[i] < 2^119 */
+static void widefelem_diff(widefelem out, const widefelem in)
+	{
+	static const widelimb two120 = ((widelimb) 1) << 120;
+	static const widelimb two120m64 = (((widelimb) 1) << 120) -
+		(((widelimb) 1) << 64);
+	static const widelimb two120m104m64 = (((widelimb) 1) << 120) -
+		(((widelimb) 1) << 104) - (((widelimb) 1) << 64);
+
+	/* Add 0 mod 2^224-2^96+1 to ensure out > in */
+	out[0] += two120;
+	out[1] += two120m64;
+	out[2] += two120m64;
+	out[3] += two120;
+	out[4] += two120m104m64;
+	out[5] += two120m64;
+	out[6] += two120m64;
+
+	out[0] -= in[0];
+	out[1] -= in[1];
+	out[2] -= in[2];
+	out[3] -= in[3];
+	out[4] -= in[4];
+	out[5] -= in[5];
+	out[6] -= in[6];
+	}
+
+/* Subtract in mixed mode: out128 -= in64 */
+/* in[i] < 2^63 */
+static void felem_diff_128_64(widefelem out, const felem in)
+	{
+	static const widelimb two64p8 = (((widelimb) 1) << 64) +
+		(((widelimb) 1) << 8);
+	static const widelimb two64m8 = (((widelimb) 1) << 64) -
+		(((widelimb) 1) << 8);
+	static const widelimb two64m48m8 = (((widelimb) 1) << 64) -
+		(((widelimb) 1) << 48) - (((widelimb) 1) << 8);
+
+	/* Add 0 mod 2^224-2^96+1 to ensure out > in */
+	out[0] += two64p8;
+	out[1] += two64m48m8;
+	out[2] += two64m8;
+	out[3] += two64m8;
+
+	out[0] -= in[0];
+	out[1] -= in[1];
+	out[2] -= in[2];
+	out[3] -= in[3];
+	}
+
+/* Multiply a field element by a scalar: out = out * scalar
+ * The scalars we actually use are small, so results fit without overflow */
+static void felem_scalar(felem out, const limb scalar)
+	{
+	out[0] *= scalar;
+	out[1] *= scalar;
+	out[2] *= scalar;
+	out[3] *= scalar;
+	}
+
+/* Multiply an unreduced field element by a scalar: out = out * scalar
+ * The scalars we actually use are small, so results fit without overflow */
+static void widefelem_scalar(widefelem out, const widelimb scalar)
+	{
+	out[0] *= scalar;
+	out[1] *= scalar;
+	out[2] *= scalar;
+	out[3] *= scalar;
+	out[4] *= scalar;
+	out[5] *= scalar;
+	out[6] *= scalar;
+	}
+
+/* Square a field element: out = in^2 */
+static void felem_square(widefelem out, const felem in)
+	{
+	limb tmp0, tmp1, tmp2;
+	tmp0 = 2 * in[0]; tmp1 = 2 * in[1]; tmp2 = 2 * in[2];
+	out[0] = ((widelimb) in[0]) * in[0];
+	out[1] = ((widelimb) in[0]) * tmp1;
+	out[2] = ((widelimb) in[0]) * tmp2 + ((widelimb) in[1]) * in[1];
+	out[3] = ((widelimb) in[3]) * tmp0 +
+		((widelimb) in[1]) * tmp2;
+	out[4] = ((widelimb) in[3]) * tmp1 + ((widelimb) in[2]) * in[2];
+	out[5] = ((widelimb) in[3]) * tmp2;
+	out[6] = ((widelimb) in[3]) * in[3];
+	}
+
+/* Multiply two field elements: out = in1 * in2 */
+static void felem_mul(widefelem out, const felem in1, const felem in2)
+	{
+	out[0] = ((widelimb) in1[0]) * in2[0];
+	out[1] = ((widelimb) in1[0]) * in2[1] + ((widelimb) in1[1]) * in2[0];
+	out[2] = ((widelimb) in1[0]) * in2[2] + ((widelimb) in1[1]) * in2[1] +
+		((widelimb) in1[2]) * in2[0];
+	out[3] = ((widelimb) in1[0]) * in2[3] + ((widelimb) in1[1]) * in2[2] +
+		((widelimb) in1[2]) * in2[1] + ((widelimb) in1[3]) * in2[0];
+	out[4] = ((widelimb) in1[1]) * in2[3] + ((widelimb) in1[2]) * in2[2] +
+		((widelimb) in1[3]) * in2[1];
+	out[5] = ((widelimb) in1[2]) * in2[3] + ((widelimb) in1[3]) * in2[2];
+	out[6] = ((widelimb) in1[3]) * in2[3];
+	}
+
+/* Reduce seven 128-bit coefficients to four 64-bit coefficients.
+ * Requires in[i] < 2^126,
+ * ensures out[0] < 2^56, out[1] < 2^56, out[2] < 2^56, out[3] <= 2^56 + 2^16 */
+static void felem_reduce(felem out, const widefelem in)
+	{
+	static const widelimb two127p15 = (((widelimb) 1) << 127) +
+		(((widelimb) 1) << 15);
+	static const widelimb two127m71 = (((widelimb) 1) << 127) -
+		(((widelimb) 1) << 71);
+	static const widelimb two127m71m55 = (((widelimb) 1) << 127) -
+		(((widelimb) 1) << 71) - (((widelimb) 1) << 55);
+	widelimb output[5];
+
+	/* Add 0 mod 2^224-2^96+1 to ensure all differences are positive */
+	output[0] = in[0] + two127p15;
+	output[1] = in[1] + two127m71m55;
+	output[2] = in[2] + two127m71;
+	output[3] = in[3];
+	output[4] = in[4];
+
+	/* Eliminate in[4], in[5], in[6] */
+	output[4] += in[6] >> 16;
+	output[3] += (in[6] & 0xffff) << 40;
+	output[2] -= in[6];
+
+	output[3] += in[5] >> 16;
+	output[2] += (in[5] & 0xffff) << 40;
+	output[1] -= in[5];
+
+	output[2] += output[4] >> 16;
+	output[1] += (output[4] & 0xffff) << 40;
+	output[0] -= output[4];
+
+	/* Carry 2 -> 3 -> 4 */
+	output[3] += output[2] >> 56;
+	output[2] &= 0x00ffffffffffffff;
+
+	output[4] = output[3] >> 56;
+	output[3] &= 0x00ffffffffffffff;
+
+	/* Now output[2] < 2^56, output[3] < 2^56, output[4] < 2^72 */
+
+	/* Eliminate output[4] */
+	output[2] += output[4] >> 16;
+	/* output[2] < 2^56 + 2^56 = 2^57 */
+	output[1] += (output[4] & 0xffff) << 40;
+	output[0] -= output[4];
+
+	/* Carry 0 -> 1 -> 2 -> 3 */
+	output[1] += output[0] >> 56;
+	out[0] = output[0] & 0x00ffffffffffffff;
+
+	output[2] += output[1] >> 56;
+	/* output[2] < 2^57 + 2^72 */
+	out[1] = output[1] & 0x00ffffffffffffff;
+	output[3] += output[2] >> 56;
+	/* output[3] <= 2^56 + 2^16 */
+	out[2] = output[2] & 0x00ffffffffffffff;
+
+	/* out[0] < 2^56, out[1] < 2^56, out[2] < 2^56,
+	 * out[3] <= 2^56 + 2^16 (due to final carry),
+	 * so out < 2*p */
+	out[3] = output[3];
+	}
+
+static void felem_square_reduce(felem out, const felem in)
+	{
+	widefelem tmp;
+	felem_square(tmp, in);
+	felem_reduce(out, tmp);
+	}
+
+static void felem_mul_reduce(felem out, const felem in1, const felem in2)
+	{
+	widefelem tmp;
+	felem_mul(tmp, in1, in2);
+	felem_reduce(out, tmp);
+	}
+
+/* Reduce to unique minimal representation.
+ * Requires 0 <= in < 2*p (always call felem_reduce first) */
+static void felem_contract(felem out, const felem in)
+	{
+	static const int64_t two56 = ((limb) 1) << 56;
+	/* 0 <= in < 2*p, p = 2^224 - 2^96 + 1 */
+	/* if in > p , reduce in = in - 2^224 + 2^96 - 1 */
+	int64_t tmp[4], a;
+	tmp[0] = in[0];
+	tmp[1] = in[1];
+	tmp[2] = in[2];
+	tmp[3] = in[3];
+	/* Case 1: a = 1 iff in >= 2^224 */
+	a = (in[3] >> 56);
+	tmp[0] -= a;
+	tmp[1] += a << 40;
+	tmp[3] &= 0x00ffffffffffffff;
+	/* Case 2: a = 0 iff p <= in < 2^224, i.e.,
+	 * the high 128 bits are all 1 and the lower part is non-zero */
+	a = ((in[3] & in[2] & (in[1] | 0x000000ffffffffff)) + 1) |
+		(((int64_t)(in[0] + (in[1] & 0x000000ffffffffff)) - 1) >> 63);
+	a &= 0x00ffffffffffffff;
+	/* turn a into an all-one mask (if a = 0) or an all-zero mask */
+	a = (a - 1) >> 63;
+	/* subtract 2^224 - 2^96 + 1 if a is all-one*/
+	tmp[3] &= a ^ 0xffffffffffffffff;
+	tmp[2] &= a ^ 0xffffffffffffffff;
+	tmp[1] &= (a ^ 0xffffffffffffffff) | 0x000000ffffffffff;
+	tmp[0] -= 1 & a;
+
+	/* eliminate negative coefficients: if tmp[0] is negative, tmp[1] must
+	 * be non-zero, so we only need one step */
+	a = tmp[0] >> 63;
+	tmp[0] += two56 & a;
+	tmp[1] -= 1 & a;
+
+	/* carry 1 -> 2 -> 3 */
+	tmp[2] += tmp[1] >> 56;
+	tmp[1] &= 0x00ffffffffffffff;
+
+	tmp[3] += tmp[2] >> 56;
+	tmp[2] &= 0x00ffffffffffffff;
+
+	/* Now 0 <= out < p */
+	out[0] = tmp[0];
+	out[1] = tmp[1];
+	out[2] = tmp[2];
+	out[3] = tmp[3];
+	}
+
+/* Zero-check: returns 1 if input is 0, and 0 otherwise.
+ * We know that field elements are reduced to in < 2^225,
+ * so we only need to check three cases: 0, 2^224 - 2^96 + 1,
+ * and 2^225 - 2^97 + 2 */
+static limb felem_is_zero(const felem in)
+	{
+	limb zero, two224m96p1, two225m97p2;
+
+	zero = in[0] | in[1] | in[2] | in[3];
+	zero = (((int64_t)(zero) - 1) >> 63) & 1;
+	two224m96p1 = (in[0] ^ 1) | (in[1] ^ 0x00ffff0000000000)
+		| (in[2] ^ 0x00ffffffffffffff) | (in[3] ^ 0x00ffffffffffffff);
+	two224m96p1 = (((int64_t)(two224m96p1) - 1) >> 63) & 1;
+	two225m97p2 = (in[0] ^ 2) | (in[1] ^ 0x00fffe0000000000)
+		| (in[2] ^ 0x00ffffffffffffff) | (in[3] ^ 0x01ffffffffffffff);
+	two225m97p2 = (((int64_t)(two225m97p2) - 1) >> 63) & 1;
+	return (zero | two224m96p1 | two225m97p2);
+	}
+
+static limb felem_is_zero_int(const felem in)
+	{
+	return (int) (felem_is_zero(in) & ((limb)1));
+	}
+
+/* Invert a field element */
+/* Computation chain copied from djb's code */
+static void felem_inv(felem out, const felem in)
+	{
+	felem ftmp, ftmp2, ftmp3, ftmp4;
+	widefelem tmp;
+	unsigned i;
+
+	felem_square(tmp, in); felem_reduce(ftmp, tmp);		/* 2 */
+	felem_mul(tmp, in, ftmp); felem_reduce(ftmp, tmp);	/* 2^2 - 1 */
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);	/* 2^3 - 2 */
+	felem_mul(tmp, in, ftmp); felem_reduce(ftmp, tmp);	/* 2^3 - 1 */
+	felem_square(tmp, ftmp); felem_reduce(ftmp2, tmp);	/* 2^4 - 2 */
+	felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);	/* 2^5 - 4 */
+	felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);	/* 2^6 - 8 */
+	felem_mul(tmp, ftmp2, ftmp); felem_reduce(ftmp, tmp);	/* 2^6 - 1 */
+	felem_square(tmp, ftmp); felem_reduce(ftmp2, tmp);	/* 2^7 - 2 */
+	for (i = 0; i < 5; ++i)					/* 2^12 - 2^6 */
+		{
+		felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);
+		}
+	felem_mul(tmp, ftmp2, ftmp); felem_reduce(ftmp2, tmp);	/* 2^12 - 1 */
+	felem_square(tmp, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^13 - 2 */
+	for (i = 0; i < 11; ++i)				/* 2^24 - 2^12 */
+		{
+		felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);
+		}
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp2, tmp); /* 2^24 - 1 */
+	felem_square(tmp, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^25 - 2 */
+	for (i = 0; i < 23; ++i)				/* 2^48 - 2^24 */
+		{
+		felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);
+		}
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp3, tmp); /* 2^48 - 1 */
+	felem_square(tmp, ftmp3); felem_reduce(ftmp4, tmp);	/* 2^49 - 2 */
+	for (i = 0; i < 47; ++i)				/* 2^96 - 2^48 */
+		{
+		felem_square(tmp, ftmp4); felem_reduce(ftmp4, tmp);
+		}
+	felem_mul(tmp, ftmp3, ftmp4); felem_reduce(ftmp3, tmp); /* 2^96 - 1 */
+	felem_square(tmp, ftmp3); felem_reduce(ftmp4, tmp);	/* 2^97 - 2 */
+	for (i = 0; i < 23; ++i)				/* 2^120 - 2^24 */
+		{
+		felem_square(tmp, ftmp4); felem_reduce(ftmp4, tmp);
+		}
+	felem_mul(tmp, ftmp2, ftmp4); felem_reduce(ftmp2, tmp); /* 2^120 - 1 */
+	for (i = 0; i < 6; ++i)					/* 2^126 - 2^6 */
+		{
+		felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);
+		}
+	felem_mul(tmp, ftmp2, ftmp); felem_reduce(ftmp, tmp);	/* 2^126 - 1 */
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);	/* 2^127 - 2 */
+	felem_mul(tmp, ftmp, in); felem_reduce(ftmp, tmp);	/* 2^127 - 1 */
+	for (i = 0; i < 97; ++i)				/* 2^224 - 2^97 */
+		{
+		felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);
+		}
+	felem_mul(tmp, ftmp, ftmp3); felem_reduce(out, tmp);	/* 2^224 - 2^96 - 1 */
+	}
+
+/* Copy in constant time:
+ * if icopy == 1, copy in to out,
+ * if icopy == 0, copy out to itself. */
+static void
+copy_conditional(felem out, const felem in, limb icopy)
+	{
+	unsigned i;
+	/* icopy is a (64-bit) 0 or 1, so copy is either all-zero or all-one */
+	const limb copy = -icopy;
+	for (i = 0; i < 4; ++i)
+		{
+		const limb tmp = copy & (in[i] ^ out[i]);
+		out[i] ^= tmp;
+		}
+	}
+
+/******************************************************************************/
+/*			 ELLIPTIC CURVE POINT OPERATIONS
+ *
+ * Points are represented in Jacobian projective coordinates:
+ * (X, Y, Z) corresponds to the affine point (X/Z^2, Y/Z^3),
+ * or to the point at infinity if Z == 0.
+ *
+ */
+
+/* Double an elliptic curve point:
+ * (X', Y', Z') = 2 * (X, Y, Z), where
+ * X' = (3 * (X - Z^2) * (X + Z^2))^2 - 8 * X * Y^2
+ * Y' = 3 * (X - Z^2) * (X + Z^2) * (4 * X * Y^2 - X') - 8 * Y^2
+ * Z' = (Y + Z)^2 - Y^2 - Z^2 = 2 * Y * Z
+ * Outputs can equal corresponding inputs, i.e., x_out == x_in is allowed,
+ * while x_out == y_in is not (maybe this works, but it's not tested). */
+static void
+point_double(felem x_out, felem y_out, felem z_out,
+             const felem x_in, const felem y_in, const felem z_in)
+	{
+	widefelem tmp, tmp2;
+	felem delta, gamma, beta, alpha, ftmp, ftmp2;
+
+	felem_assign(ftmp, x_in);
+	felem_assign(ftmp2, x_in);
+
+	/* delta = z^2 */
+	felem_square(tmp, z_in);
+	felem_reduce(delta, tmp);
+
+	/* gamma = y^2 */
+	felem_square(tmp, y_in);
+	felem_reduce(gamma, tmp);
+
+	/* beta = x*gamma */
+	felem_mul(tmp, x_in, gamma);
+	felem_reduce(beta, tmp);
+
+	/* alpha = 3*(x-delta)*(x+delta) */
+	felem_diff(ftmp, delta);
+	/* ftmp[i] < 2^57 + 2^58 + 2 < 2^59 */
+	felem_sum(ftmp2, delta);
+	/* ftmp2[i] < 2^57 + 2^57 = 2^58 */
+	felem_scalar(ftmp2, 3);
+	/* ftmp2[i] < 3 * 2^58 < 2^60 */
+	felem_mul(tmp, ftmp, ftmp2);
+	/* tmp[i] < 2^60 * 2^59 * 4 = 2^121 */
+	felem_reduce(alpha, tmp);
+
+	/* x' = alpha^2 - 8*beta */
+	felem_square(tmp, alpha);
+	/* tmp[i] < 4 * 2^57 * 2^57 = 2^116 */
+	felem_assign(ftmp, beta);
+	felem_scalar(ftmp, 8);
+	/* ftmp[i] < 8 * 2^57 = 2^60 */
+	felem_diff_128_64(tmp, ftmp);
+	/* tmp[i] < 2^116 + 2^64 + 8 < 2^117 */
+	felem_reduce(x_out, tmp);
+
+	/* z' = (y + z)^2 - gamma - delta */
+	felem_sum(delta, gamma);
+	/* delta[i] < 2^57 + 2^57 = 2^58 */
+	felem_assign(ftmp, y_in);
+	felem_sum(ftmp, z_in);
+	/* ftmp[i] < 2^57 + 2^57 = 2^58 */
+	felem_square(tmp, ftmp);
+	/* tmp[i] < 4 * 2^58 * 2^58 = 2^118 */
+	felem_diff_128_64(tmp, delta);
+	/* tmp[i] < 2^118 + 2^64 + 8 < 2^119 */
+	felem_reduce(z_out, tmp);
+
+	/* y' = alpha*(4*beta - x') - 8*gamma^2 */
+	felem_scalar(beta, 4);
+	/* beta[i] < 4 * 2^57 = 2^59 */
+	felem_diff(beta, x_out);
+	/* beta[i] < 2^59 + 2^58 + 2 < 2^60 */
+	felem_mul(tmp, alpha, beta);
+	/* tmp[i] < 4 * 2^57 * 2^60 = 2^119 */
+	felem_square(tmp2, gamma);
+	/* tmp2[i] < 4 * 2^57 * 2^57 = 2^116 */
+	widefelem_scalar(tmp2, 8);
+	/* tmp2[i] < 8 * 2^116 = 2^119 */
+	widefelem_diff(tmp, tmp2);
+	/* tmp[i] < 2^119 + 2^120 < 2^121 */
+	felem_reduce(y_out, tmp);
+	}
+
+/* Add two elliptic curve points:
+ * (X_1, Y_1, Z_1) + (X_2, Y_2, Z_2) = (X_3, Y_3, Z_3), where
+ * X_3 = (Z_1^3 * Y_2 - Z_2^3 * Y_1)^2 - (Z_1^2 * X_2 - Z_2^2 * X_1)^3 -
+ * 2 * Z_2^2 * X_1 * (Z_1^2 * X_2 - Z_2^2 * X_1)^2
+ * Y_3 = (Z_1^3 * Y_2 - Z_2^3 * Y_1) * (Z_2^2 * X_1 * (Z_1^2 * X_2 - Z_2^2 * X_1)^2 - X_3) -
+ *        Z_2^3 * Y_1 * (Z_1^2 * X_2 - Z_2^2 * X_1)^3
+ * Z_3 = (Z_1^2 * X_2 - Z_2^2 * X_1) * (Z_1 * Z_2)
+ *
+ * This runs faster if 'mixed' is set, which requires Z_2 = 1 or Z_2 = 0.
+ */
+
+/* This function is not entirely constant-time:
+ * it includes a branch for checking whether the two input points are equal,
+ * (while not equal to the point at infinity).
+ * This case never happens during single point multiplication,
+ * so there is no timing leak for ECDH or ECDSA signing. */
+static void point_add(felem x3, felem y3, felem z3,
+	const felem x1, const felem y1, const felem z1,
+	const int mixed, const felem x2, const felem y2, const felem z2)
+	{
+	felem ftmp, ftmp2, ftmp3, ftmp4, ftmp5, x_out, y_out, z_out;
+	widefelem tmp, tmp2;
+	limb z1_is_zero, z2_is_zero, x_equal, y_equal;
+
+	if (!mixed)
+		{
+		/* ftmp2 = z2^2 */
+		felem_square(tmp, z2);
+		felem_reduce(ftmp2, tmp);
+
+		/* ftmp4 = z2^3 */
+		felem_mul(tmp, ftmp2, z2);
+		felem_reduce(ftmp4, tmp);
+
+		/* ftmp4 = z2^3*y1 */
+		felem_mul(tmp2, ftmp4, y1);
+		felem_reduce(ftmp4, tmp2);
+
+		/* ftmp2 = z2^2*x1 */
+		felem_mul(tmp2, ftmp2, x1);
+		felem_reduce(ftmp2, tmp2);
+		}
+	else
+		{
+		/* We'll assume z2 = 1 (special case z2 = 0 is handled later) */
+
+		/* ftmp4 = z2^3*y1 */
+		felem_assign(ftmp4, y1);
+
+		/* ftmp2 = z2^2*x1 */
+		felem_assign(ftmp2, x1);
+		}
+
+	/* ftmp = z1^2 */
+	felem_square(tmp, z1);
+	felem_reduce(ftmp, tmp);
+
+	/* ftmp3 = z1^3 */
+	felem_mul(tmp, ftmp, z1);
+	felem_reduce(ftmp3, tmp);
+
+	/* tmp = z1^3*y2 */
+	felem_mul(tmp, ftmp3, y2);
+	/* tmp[i] < 4 * 2^57 * 2^57 = 2^116 */
+
+	/* ftmp3 = z1^3*y2 - z2^3*y1 */
+	felem_diff_128_64(tmp, ftmp4);
+	/* tmp[i] < 2^116 + 2^64 + 8 < 2^117 */
+	felem_reduce(ftmp3, tmp);
+
+	/* tmp = z1^2*x2 */
+	felem_mul(tmp, ftmp, x2);
+	/* tmp[i] < 4 * 2^57 * 2^57 = 2^116 */
+
+	/* ftmp = z1^2*x2 - z2^2*x1 */
+	felem_diff_128_64(tmp, ftmp2);
+	/* tmp[i] < 2^116 + 2^64 + 8 < 2^117 */
+	felem_reduce(ftmp, tmp);
+
+	/* the formulae are incorrect if the points are equal
+	 * so we check for this and do doubling if this happens */
+	x_equal = felem_is_zero(ftmp);
+	y_equal = felem_is_zero(ftmp3);
+	z1_is_zero = felem_is_zero(z1);
+	z2_is_zero = felem_is_zero(z2);
+	/* In affine coordinates, (X_1, Y_1) == (X_2, Y_2) */
+	if (x_equal && y_equal && !z1_is_zero && !z2_is_zero)
+		{
+		point_double(x3, y3, z3, x1, y1, z1);
+		return;
+		}
+
+	/* ftmp5 = z1*z2 */
+	if (!mixed)
+		{
+		felem_mul(tmp, z1, z2);
+		felem_reduce(ftmp5, tmp);
+		}
+	else
+		{
+		/* special case z2 = 0 is handled later */
+		felem_assign(ftmp5, z1);
+		}
+
+	/* z_out = (z1^2*x2 - z2^2*x1)*(z1*z2) */
+	felem_mul(tmp, ftmp, ftmp5);
+	felem_reduce(z_out, tmp);
+
+	/* ftmp = (z1^2*x2 - z2^2*x1)^2 */
+	felem_assign(ftmp5, ftmp);
+	felem_square(tmp, ftmp);
+	felem_reduce(ftmp, tmp);
+
+	/* ftmp5 = (z1^2*x2 - z2^2*x1)^3 */
+	felem_mul(tmp, ftmp, ftmp5);
+	felem_reduce(ftmp5, tmp);
+
+	/* ftmp2 = z2^2*x1*(z1^2*x2 - z2^2*x1)^2 */
+	felem_mul(tmp, ftmp2, ftmp);
+	felem_reduce(ftmp2, tmp);
+
+	/* tmp = z2^3*y1*(z1^2*x2 - z2^2*x1)^3 */
+	felem_mul(tmp, ftmp4, ftmp5);
+	/* tmp[i] < 4 * 2^57 * 2^57 = 2^116 */
+
+	/* tmp2 = (z1^3*y2 - z2^3*y1)^2 */
+	felem_square(tmp2, ftmp3);
+	/* tmp2[i] < 4 * 2^57 * 2^57 < 2^116 */
+
+	/* tmp2 = (z1^3*y2 - z2^3*y1)^2 - (z1^2*x2 - z2^2*x1)^3 */
+	felem_diff_128_64(tmp2, ftmp5);
+	/* tmp2[i] < 2^116 + 2^64 + 8 < 2^117 */
+
+	/* ftmp5 = 2*z2^2*x1*(z1^2*x2 - z2^2*x1)^2 */
+	felem_assign(ftmp5, ftmp2);
+	felem_scalar(ftmp5, 2);
+	/* ftmp5[i] < 2 * 2^57 = 2^58 */
+
+	/* x_out = (z1^3*y2 - z2^3*y1)^2 - (z1^2*x2 - z2^2*x1)^3 -
+	   2*z2^2*x1*(z1^2*x2 - z2^2*x1)^2 */
+	felem_diff_128_64(tmp2, ftmp5);
+	/* tmp2[i] < 2^117 + 2^64 + 8 < 2^118 */
+	felem_reduce(x_out, tmp2);
+
+	/* ftmp2 = z2^2*x1*(z1^2*x2 - z2^2*x1)^2 - x_out */
+	felem_diff(ftmp2, x_out);
+	/* ftmp2[i] < 2^57 + 2^58 + 2 < 2^59 */
+
+	/* tmp2 = (z1^3*y2 - z2^3*y1)*(z2^2*x1*(z1^2*x2 - z2^2*x1)^2 - x_out) */
+	felem_mul(tmp2, ftmp3, ftmp2);
+	/* tmp2[i] < 4 * 2^57 * 2^59 = 2^118 */
+
+	/* y_out = (z1^3*y2 - z2^3*y1)*(z2^2*x1*(z1^2*x2 - z2^2*x1)^2 - x_out) -
+	   z2^3*y1*(z1^2*x2 - z2^2*x1)^3 */
+	widefelem_diff(tmp2, tmp);
+	/* tmp2[i] < 2^118 + 2^120 < 2^121 */
+	felem_reduce(y_out, tmp2);
+
+	/* the result (x_out, y_out, z_out) is incorrect if one of the inputs is
+	 * the point at infinity, so we need to check for this separately */
+
+	/* if point 1 is at infinity, copy point 2 to output, and vice versa */
+	copy_conditional(x_out, x2, z1_is_zero);
+	copy_conditional(x_out, x1, z2_is_zero);
+	copy_conditional(y_out, y2, z1_is_zero);
+	copy_conditional(y_out, y1, z2_is_zero);
+	copy_conditional(z_out, z2, z1_is_zero);
+	copy_conditional(z_out, z1, z2_is_zero);
+	felem_assign(x3, x_out);
+	felem_assign(y3, y_out);
+	felem_assign(z3, z_out);
+	}
+
+/* select_point selects the |idx|th point from a precomputation table and
+ * copies it to out. */
+static void select_point(const u64 idx, unsigned int size, const felem pre_comp[/*size*/][3], felem out[3])
+	{
+	unsigned i, j;
+	limb *outlimbs = &out[0][0];
+	memset(outlimbs, 0, 3 * sizeof(felem));
+
+	for (i = 0; i < size; i++)
+		{
+		const limb *inlimbs = &pre_comp[i][0][0];
+		u64 mask = i ^ idx;
+		mask |= mask >> 4;
+		mask |= mask >> 2;
+		mask |= mask >> 1;
+		mask &= 1;
+		mask--;
+		for (j = 0; j < 4 * 3; j++)
+			outlimbs[j] |= inlimbs[j] & mask;
+		}
+	}
+
+/* get_bit returns the |i|th bit in |in| */
+static char get_bit(const felem_bytearray in, unsigned i)
+	{
+	if (i >= 224)
+		return 0;
+	return (in[i >> 3] >> (i & 7)) & 1;
+	}
+
+/* Interleaved point multiplication using precomputed point multiples:
+ * The small point multiples 0*P, 1*P, ..., 16*P are in pre_comp[],
+ * the scalars in scalars[]. If g_scalar is non-NULL, we also add this multiple
+ * of the generator, using certain (large) precomputed multiples in g_pre_comp.
+ * Output point (X, Y, Z) is stored in x_out, y_out, z_out */
+static void batch_mul(felem x_out, felem y_out, felem z_out,
+	const felem_bytearray scalars[], const unsigned num_points, const u8 *g_scalar,
+	const int mixed, const felem pre_comp[][17][3], const felem g_pre_comp[2][16][3])
+	{
+	int i, skip;
+	unsigned num;
+	unsigned gen_mul = (g_scalar != NULL);
+	felem nq[3], tmp[4];
+	u64 bits;
+	u8 sign, digit;
+
+	/* set nq to the point at infinity */
+	memset(nq, 0, 3 * sizeof(felem));
+
+	/* Loop over all scalars msb-to-lsb, interleaving additions
+	 * of multiples of the generator (two in each of the last 28 rounds)
+	 * and additions of other points multiples (every 5th round).
+	 */
+	skip = 1; /* save two point operations in the first round */
+	for (i = (num_points ? 220 : 27); i >= 0; --i)
+		{
+		/* double */
+		if (!skip)
+			point_double(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2]);
+
+		/* add multiples of the generator */
+		if (gen_mul && (i <= 27))
+			{
+			/* first, look 28 bits upwards */
+			bits = get_bit(g_scalar, i + 196) << 3;
+			bits |= get_bit(g_scalar, i + 140) << 2;
+			bits |= get_bit(g_scalar, i + 84) << 1;
+			bits |= get_bit(g_scalar, i + 28);
+			/* select the point to add, in constant time */
+			select_point(bits, 16, g_pre_comp[1], tmp);
+
+			if (!skip)
+				{
+				point_add(nq[0], nq[1], nq[2],
+					nq[0], nq[1], nq[2],
+					1 /* mixed */, tmp[0], tmp[1], tmp[2]);
+				}
+			else
+				{
+				memcpy(nq, tmp, 3 * sizeof(felem));
+				skip = 0;
+				}
+
+			/* second, look at the current position */
+			bits = get_bit(g_scalar, i + 168) << 3;
+			bits |= get_bit(g_scalar, i + 112) << 2;
+			bits |= get_bit(g_scalar, i + 56) << 1;
+			bits |= get_bit(g_scalar, i);
+			/* select the point to add, in constant time */
+			select_point(bits, 16, g_pre_comp[0], tmp);
+			point_add(nq[0], nq[1], nq[2],
+				nq[0], nq[1], nq[2],
+				1 /* mixed */, tmp[0], tmp[1], tmp[2]);
+			}
+
+		/* do other additions every 5 doublings */
+		if (num_points && (i % 5 == 0))
+			{
+			/* loop over all scalars */
+			for (num = 0; num < num_points; ++num)
+				{
+				bits = get_bit(scalars[num], i + 4) << 5;
+				bits |= get_bit(scalars[num], i + 3) << 4;
+				bits |= get_bit(scalars[num], i + 2) << 3;
+				bits |= get_bit(scalars[num], i + 1) << 2;
+				bits |= get_bit(scalars[num], i) << 1;
+				bits |= get_bit(scalars[num], i - 1);
+				ec_GFp_nistp_recode_scalar_bits(&sign, &digit, bits);
+
+				/* select the point to add or subtract */
+				select_point(digit, 17, pre_comp[num], tmp);
+				felem_neg(tmp[3], tmp[1]); /* (X, -Y, Z) is the negative point */
+				copy_conditional(tmp[1], tmp[3], sign);
+
+				if (!skip)
+					{
+					point_add(nq[0], nq[1], nq[2],
+						nq[0], nq[1], nq[2],
+						mixed, tmp[0], tmp[1], tmp[2]);
+					}
+				else
+					{
+					memcpy(nq, tmp, 3 * sizeof(felem));
+					skip = 0;
+					}
+				}
+			}
+		}
+	felem_assign(x_out, nq[0]);
+	felem_assign(y_out, nq[1]);
+	felem_assign(z_out, nq[2]);
+	}
+
+/******************************************************************************/
+/*		       FUNCTIONS TO MANAGE PRECOMPUTATION
+ */
+
+static NISTP224_PRE_COMP *nistp224_pre_comp_new()
+	{
+	NISTP224_PRE_COMP *ret = NULL;
+	ret = (NISTP224_PRE_COMP *) OPENSSL_malloc(sizeof *ret);
+	if (!ret)
+		{
+		ECerr(EC_F_NISTP224_PRE_COMP_NEW, ERR_R_MALLOC_FAILURE);
+		return ret;
+		}
+	memset(ret->g_pre_comp, 0, sizeof(ret->g_pre_comp));
+	ret->references = 1;
+	return ret;
+	}
+
+static void *nistp224_pre_comp_dup(void *src_)
+	{
+	NISTP224_PRE_COMP *src = src_;
+
+	/* no need to actually copy, these objects never change! */
+	CRYPTO_add(&src->references, 1, CRYPTO_LOCK_EC_PRE_COMP);
+
+	return src_;
+	}
+
+static void nistp224_pre_comp_free(void *pre_)
+	{
+	int i;
+	NISTP224_PRE_COMP *pre = pre_;
+
+	if (!pre)
+		return;
+
+	i = CRYPTO_add(&pre->references, -1, CRYPTO_LOCK_EC_PRE_COMP);
+	if (i > 0)
+		return;
+
+	OPENSSL_free(pre);
+	}
+
+static void nistp224_pre_comp_clear_free(void *pre_)
+	{
+	int i;
+	NISTP224_PRE_COMP *pre = pre_;
+
+	if (!pre)
+		return;
+
+	i = CRYPTO_add(&pre->references, -1, CRYPTO_LOCK_EC_PRE_COMP);
+	if (i > 0)
+		return;
+
+	OPENSSL_cleanse(pre, sizeof *pre);
+	OPENSSL_free(pre);
+	}
+
+/******************************************************************************/
+/*			   OPENSSL EC_METHOD FUNCTIONS
+ */
+
+int ec_GFp_nistp224_group_init(EC_GROUP *group)
+	{
+	int ret;
+	ret = ec_GFp_simple_group_init(group);
+	group->a_is_minus3 = 1;
+	return ret;
+	}
+
+int ec_GFp_nistp224_group_set_curve(EC_GROUP *group, const BIGNUM *p,
+	const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx)
+	{
+	int ret = 0;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *curve_p, *curve_a, *curve_b;
+
+	if (ctx == NULL)
+		if ((ctx = new_ctx = BN_CTX_new()) == NULL) return 0;
+	BN_CTX_start(ctx);
+	if (((curve_p = BN_CTX_get(ctx)) == NULL) ||
+		((curve_a = BN_CTX_get(ctx)) == NULL) ||
+		((curve_b = BN_CTX_get(ctx)) == NULL)) goto err;
+	BN_bin2bn(nistp224_curve_params[0], sizeof(felem_bytearray), curve_p);
+	BN_bin2bn(nistp224_curve_params[1], sizeof(felem_bytearray), curve_a);
+	BN_bin2bn(nistp224_curve_params[2], sizeof(felem_bytearray), curve_b);
+	if ((BN_cmp(curve_p, p)) || (BN_cmp(curve_a, a)) ||
+		(BN_cmp(curve_b, b)))
+		{
+		ECerr(EC_F_EC_GFP_NISTP224_GROUP_SET_CURVE,
+			EC_R_WRONG_CURVE_PARAMETERS);
+		goto err;
+		}
+	group->field_mod_func = BN_nist_mod_224;
+	ret = ec_GFp_simple_group_set_curve(group, p, a, b, ctx);
+err:
+	BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return ret;
+	}
+
+/* Takes the Jacobian coordinates (X, Y, Z) of a point and returns
+ * (X', Y') = (X/Z^2, Y/Z^3) */
+int ec_GFp_nistp224_point_get_affine_coordinates(const EC_GROUP *group,
+	const EC_POINT *point, BIGNUM *x, BIGNUM *y, BN_CTX *ctx)
+	{
+	felem z1, z2, x_in, y_in, x_out, y_out;
+	widefelem tmp;
+
+	if (EC_POINT_is_at_infinity(group, point))
+		{
+		ECerr(EC_F_EC_GFP_NISTP224_POINT_GET_AFFINE_COORDINATES,
+			EC_R_POINT_AT_INFINITY);
+		return 0;
+		}
+	if ((!BN_to_felem(x_in, &point->X)) || (!BN_to_felem(y_in, &point->Y)) ||
+		(!BN_to_felem(z1, &point->Z))) return 0;
+	felem_inv(z2, z1);
+	felem_square(tmp, z2); felem_reduce(z1, tmp);
+	felem_mul(tmp, x_in, z1); felem_reduce(x_in, tmp);
+	felem_contract(x_out, x_in);
+	if (x != NULL)
+		{
+		if (!felem_to_BN(x, x_out)) {
+		ECerr(EC_F_EC_GFP_NISTP224_POINT_GET_AFFINE_COORDINATES,
+			ERR_R_BN_LIB);
+		return 0;
+		}
+		}
+	felem_mul(tmp, z1, z2); felem_reduce(z1, tmp);
+	felem_mul(tmp, y_in, z1); felem_reduce(y_in, tmp);
+	felem_contract(y_out, y_in);
+	if (y != NULL)
+		{
+		if (!felem_to_BN(y, y_out)) {
+		ECerr(EC_F_EC_GFP_NISTP224_POINT_GET_AFFINE_COORDINATES,
+			ERR_R_BN_LIB);
+		return 0;
+		}
+		}
+	return 1;
+	}
+
+static void make_points_affine(size_t num, felem points[/*num*/][3], felem tmp_felems[/*num+1*/])
+	{
+	/* Runs in constant time, unless an input is the point at infinity
+	 * (which normally shouldn't happen). */
+	ec_GFp_nistp_points_make_affine_internal(
+		num,
+		points,
+		sizeof(felem),
+		tmp_felems,
+		(void (*)(void *)) felem_one,
+		(int (*)(const void *)) felem_is_zero_int,
+		(void (*)(void *, const void *)) felem_assign,
+		(void (*)(void *, const void *)) felem_square_reduce,
+		(void (*)(void *, const void *, const void *)) felem_mul_reduce,
+		(void (*)(void *, const void *)) felem_inv,
+		(void (*)(void *, const void *)) felem_contract);
+	}
+
+/* Computes scalar*generator + \sum scalars[i]*points[i], ignoring NULL values
+ * Result is stored in r (r can equal one of the inputs). */
+int ec_GFp_nistp224_points_mul(const EC_GROUP *group, EC_POINT *r,
+	const BIGNUM *scalar, size_t num, const EC_POINT *points[],
+	const BIGNUM *scalars[], BN_CTX *ctx)
+	{
+	int ret = 0;
+	int j;
+	unsigned i;
+	int mixed = 0;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *x, *y, *z, *tmp_scalar;
+	felem_bytearray g_secret;
+	felem_bytearray *secrets = NULL;
+	felem (*pre_comp)[17][3] = NULL;
+	felem *tmp_felems = NULL;
+	felem_bytearray tmp;
+	unsigned num_bytes;
+	int have_pre_comp = 0;
+	size_t num_points = num;
+	felem x_in, y_in, z_in, x_out, y_out, z_out;
+	NISTP224_PRE_COMP *pre = NULL;
+	const felem (*g_pre_comp)[16][3] = NULL;
+	EC_POINT *generator = NULL;
+	const EC_POINT *p = NULL;
+	const BIGNUM *p_scalar = NULL;
+
+	if (ctx == NULL)
+		if ((ctx = new_ctx = BN_CTX_new()) == NULL) return 0;
+	BN_CTX_start(ctx);
+	if (((x = BN_CTX_get(ctx)) == NULL) ||
+		((y = BN_CTX_get(ctx)) == NULL) ||
+		((z = BN_CTX_get(ctx)) == NULL) ||
+		((tmp_scalar = BN_CTX_get(ctx)) == NULL))
+		goto err;
+
+	if (scalar != NULL)
+		{
+		pre = EC_EX_DATA_get_data(group->extra_data,
+			nistp224_pre_comp_dup, nistp224_pre_comp_free,
+			nistp224_pre_comp_clear_free);
+		if (pre)
+			/* we have precomputation, try to use it */
+			g_pre_comp = (const felem (*)[16][3]) pre->g_pre_comp;
+		else
+			/* try to use the standard precomputation */
+			g_pre_comp = &gmul[0];
+		generator = EC_POINT_new(group);
+		if (generator == NULL)
+			goto err;
+		/* get the generator from precomputation */
+		if (!felem_to_BN(x, g_pre_comp[0][1][0]) ||
+			!felem_to_BN(y, g_pre_comp[0][1][1]) ||
+			!felem_to_BN(z, g_pre_comp[0][1][2]))
+			{
+			ECerr(EC_F_EC_GFP_NISTP224_POINTS_MUL, ERR_R_BN_LIB);
+			goto err;
+			}
+		if (!EC_POINT_set_Jprojective_coordinates_GFp(group,
+				generator, x, y, z, ctx))
+			goto err;
+		if (0 == EC_POINT_cmp(group, generator, group->generator, ctx))
+			/* precomputation matches generator */
+			have_pre_comp = 1;
+		else
+			/* we don't have valid precomputation:
+			 * treat the generator as a random point */
+			num_points = num_points + 1;
+		}
+
+	if (num_points > 0)
+		{
+		if (num_points >= 3)
+			{
+			/* unless we precompute multiples for just one or two points,
+			 * converting those into affine form is time well spent  */
+			mixed = 1;
+			}
+		secrets = OPENSSL_malloc(num_points * sizeof(felem_bytearray));
+		pre_comp = OPENSSL_malloc(num_points * 17 * 3 * sizeof(felem));
+		if (mixed)
+			tmp_felems = OPENSSL_malloc((num_points * 17 + 1) * sizeof(felem));
+		if ((secrets == NULL) || (pre_comp == NULL) || (mixed && (tmp_felems == NULL)))
+			{
+			ECerr(EC_F_EC_GFP_NISTP224_POINTS_MUL, ERR_R_MALLOC_FAILURE);
+			goto err;
+			}
+
+		/* we treat NULL scalars as 0, and NULL points as points at infinity,
+		 * i.e., they contribute nothing to the linear combination */
+		memset(secrets, 0, num_points * sizeof(felem_bytearray));
+		memset(pre_comp, 0, num_points * 17 * 3 * sizeof(felem));
+		for (i = 0; i < num_points; ++i)
+			{
+			if (i == num)
+				/* the generator */
+				{
+				p = EC_GROUP_get0_generator(group);
+				p_scalar = scalar;
+				}
+			else
+				/* the i^th point */
+				{
+				p = points[i];
+				p_scalar = scalars[i];
+				}
+			if ((p_scalar != NULL) && (p != NULL))
+				{
+				/* reduce scalar to 0 <= scalar < 2^224 */
+				if ((BN_num_bits(p_scalar) > 224) || (BN_is_negative(p_scalar)))
+					{
+					/* this is an unusual input, and we don't guarantee
+					 * constant-timeness */
+					if (!BN_nnmod(tmp_scalar, p_scalar, &group->order, ctx))
+						{
+						ECerr(EC_F_EC_GFP_NISTP224_POINTS_MUL, ERR_R_BN_LIB);
+						goto err;
+						}
+					num_bytes = BN_bn2bin(tmp_scalar, tmp);
+					}
+				else
+					num_bytes = BN_bn2bin(p_scalar, tmp);
+				flip_endian(secrets[i], tmp, num_bytes);
+				/* precompute multiples */
+				if ((!BN_to_felem(x_out, &p->X)) ||
+					(!BN_to_felem(y_out, &p->Y)) ||
+					(!BN_to_felem(z_out, &p->Z))) goto err;
+				felem_assign(pre_comp[i][1][0], x_out);
+				felem_assign(pre_comp[i][1][1], y_out);
+				felem_assign(pre_comp[i][1][2], z_out);
+				for (j = 2; j <= 16; ++j)
+					{
+					if (j & 1)
+						{
+						point_add(
+							pre_comp[i][j][0], pre_comp[i][j][1], pre_comp[i][j][2],
+							pre_comp[i][1][0], pre_comp[i][1][1], pre_comp[i][1][2],
+							0, pre_comp[i][j-1][0], pre_comp[i][j-1][1], pre_comp[i][j-1][2]);
+						}
+					else
+						{
+						point_double(
+							pre_comp[i][j][0], pre_comp[i][j][1], pre_comp[i][j][2],
+							pre_comp[i][j/2][0], pre_comp[i][j/2][1], pre_comp[i][j/2][2]);
+						}
+					}
+				}
+			}
+		if (mixed)
+			make_points_affine(num_points * 17, pre_comp[0], tmp_felems);
+		}
+
+	/* the scalar for the generator */
+	if ((scalar != NULL) && (have_pre_comp))
+		{
+		memset(g_secret, 0, sizeof g_secret);
+		/* reduce scalar to 0 <= scalar < 2^224 */
+		if ((BN_num_bits(scalar) > 224) || (BN_is_negative(scalar)))
+			{
+			/* this is an unusual input, and we don't guarantee
+			 * constant-timeness */
+			if (!BN_nnmod(tmp_scalar, scalar, &group->order, ctx))
+				{
+				ECerr(EC_F_EC_GFP_NISTP224_POINTS_MUL, ERR_R_BN_LIB);
+				goto err;
+				}
+			num_bytes = BN_bn2bin(tmp_scalar, tmp);
+			}
+		else
+			num_bytes = BN_bn2bin(scalar, tmp);
+		flip_endian(g_secret, tmp, num_bytes);
+		/* do the multiplication with generator precomputation*/
+		batch_mul(x_out, y_out, z_out,
+			(const felem_bytearray (*)) secrets, num_points,
+			g_secret,
+			mixed, (const felem (*)[17][3]) pre_comp,
+			g_pre_comp);
+		}
+	else
+		/* do the multiplication without generator precomputation */
+		batch_mul(x_out, y_out, z_out,
+			(const felem_bytearray (*)) secrets, num_points,
+			NULL, mixed, (const felem (*)[17][3]) pre_comp, NULL);
+	/* reduce the output to its unique minimal representation */
+	felem_contract(x_in, x_out);
+	felem_contract(y_in, y_out);
+	felem_contract(z_in, z_out);
+	if ((!felem_to_BN(x, x_in)) || (!felem_to_BN(y, y_in)) ||
+		(!felem_to_BN(z, z_in)))
+		{
+		ECerr(EC_F_EC_GFP_NISTP224_POINTS_MUL, ERR_R_BN_LIB);
+		goto err;
+		}
+	ret = EC_POINT_set_Jprojective_coordinates_GFp(group, r, x, y, z, ctx);
+
+err:
+	BN_CTX_end(ctx);
+	if (generator != NULL)
+		EC_POINT_free(generator);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	if (secrets != NULL)
+		OPENSSL_free(secrets);
+	if (pre_comp != NULL)
+		OPENSSL_free(pre_comp);
+	if (tmp_felems != NULL)
+		OPENSSL_free(tmp_felems);
+	return ret;
+	}
+
+int ec_GFp_nistp224_precompute_mult(EC_GROUP *group, BN_CTX *ctx)
+	{
+	int ret = 0;
+	NISTP224_PRE_COMP *pre = NULL;
+	int i, j;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *x, *y;
+	EC_POINT *generator = NULL;
+	felem tmp_felems[32];
+
+	/* throw away old precomputation */
+	EC_EX_DATA_free_data(&group->extra_data, nistp224_pre_comp_dup,
+		nistp224_pre_comp_free, nistp224_pre_comp_clear_free);
+	if (ctx == NULL)
+		if ((ctx = new_ctx = BN_CTX_new()) == NULL) return 0;
+	BN_CTX_start(ctx);
+	if (((x = BN_CTX_get(ctx)) == NULL) ||
+		((y = BN_CTX_get(ctx)) == NULL))
+		goto err;
+	/* get the generator */
+	if (group->generator == NULL) goto err;
+	generator = EC_POINT_new(group);
+	if (generator == NULL)
+		goto err;
+	BN_bin2bn(nistp224_curve_params[3], sizeof (felem_bytearray), x);
+	BN_bin2bn(nistp224_curve_params[4], sizeof (felem_bytearray), y);
+	if (!EC_POINT_set_affine_coordinates_GFp(group, generator, x, y, ctx))
+		goto err;
+	if ((pre = nistp224_pre_comp_new()) == NULL)
+		goto err;
+	/* if the generator is the standard one, use built-in precomputation */
+	if (0 == EC_POINT_cmp(group, generator, group->generator, ctx))
+		{
+		memcpy(pre->g_pre_comp, gmul, sizeof(pre->g_pre_comp));
+		ret = 1;
+		goto err;
+		}
+	if ((!BN_to_felem(pre->g_pre_comp[0][1][0], &group->generator->X)) ||
+		(!BN_to_felem(pre->g_pre_comp[0][1][1], &group->generator->Y)) ||
+		(!BN_to_felem(pre->g_pre_comp[0][1][2], &group->generator->Z)))
+		goto err;
+	/* compute 2^56*G, 2^112*G, 2^168*G for the first table,
+	 * 2^28*G, 2^84*G, 2^140*G, 2^196*G for the second one
+	 */
+	for (i = 1; i <= 8; i <<= 1)
+		{
+		point_double(
+			pre->g_pre_comp[1][i][0], pre->g_pre_comp[1][i][1], pre->g_pre_comp[1][i][2],
+			pre->g_pre_comp[0][i][0], pre->g_pre_comp[0][i][1], pre->g_pre_comp[0][i][2]);
+		for (j = 0; j < 27; ++j)
+			{
+			point_double(
+				pre->g_pre_comp[1][i][0], pre->g_pre_comp[1][i][1], pre->g_pre_comp[1][i][2],
+				pre->g_pre_comp[1][i][0], pre->g_pre_comp[1][i][1], pre->g_pre_comp[1][i][2]);
+			}
+		if (i == 8)
+			break;
+		point_double(
+			pre->g_pre_comp[0][2*i][0], pre->g_pre_comp[0][2*i][1], pre->g_pre_comp[0][2*i][2],
+			pre->g_pre_comp[1][i][0], pre->g_pre_comp[1][i][1], pre->g_pre_comp[1][i][2]);
+		for (j = 0; j < 27; ++j)
+			{
+			point_double(
+				pre->g_pre_comp[0][2*i][0], pre->g_pre_comp[0][2*i][1], pre->g_pre_comp[0][2*i][2],
+				pre->g_pre_comp[0][2*i][0], pre->g_pre_comp[0][2*i][1], pre->g_pre_comp[0][2*i][2]);
+			}
+		}
+	for (i = 0; i < 2; i++)
+		{
+		/* g_pre_comp[i][0] is the point at infinity */
+		memset(pre->g_pre_comp[i][0], 0, sizeof(pre->g_pre_comp[i][0]));
+		/* the remaining multiples */
+		/* 2^56*G + 2^112*G resp. 2^84*G + 2^140*G */
+		point_add(
+			pre->g_pre_comp[i][6][0], pre->g_pre_comp[i][6][1],
+			pre->g_pre_comp[i][6][2], pre->g_pre_comp[i][4][0],
+			pre->g_pre_comp[i][4][1], pre->g_pre_comp[i][4][2],
+			0, pre->g_pre_comp[i][2][0], pre->g_pre_comp[i][2][1],
+			pre->g_pre_comp[i][2][2]);
+		/* 2^56*G + 2^168*G resp. 2^84*G + 2^196*G */
+		point_add(
+			pre->g_pre_comp[i][10][0], pre->g_pre_comp[i][10][1],
+			pre->g_pre_comp[i][10][2], pre->g_pre_comp[i][8][0],
+			pre->g_pre_comp[i][8][1], pre->g_pre_comp[i][8][2],
+			0, pre->g_pre_comp[i][2][0], pre->g_pre_comp[i][2][1],
+			pre->g_pre_comp[i][2][2]);
+		/* 2^112*G + 2^168*G resp. 2^140*G + 2^196*G */
+		point_add(
+			pre->g_pre_comp[i][12][0], pre->g_pre_comp[i][12][1],
+			pre->g_pre_comp[i][12][2], pre->g_pre_comp[i][8][0],
+			pre->g_pre_comp[i][8][1], pre->g_pre_comp[i][8][2],
+			0, pre->g_pre_comp[i][4][0], pre->g_pre_comp[i][4][1],
+			pre->g_pre_comp[i][4][2]);
+		/* 2^56*G + 2^112*G + 2^168*G resp. 2^84*G + 2^140*G + 2^196*G */
+		point_add(
+			pre->g_pre_comp[i][14][0], pre->g_pre_comp[i][14][1],
+			pre->g_pre_comp[i][14][2], pre->g_pre_comp[i][12][0],
+			pre->g_pre_comp[i][12][1], pre->g_pre_comp[i][12][2],
+			0, pre->g_pre_comp[i][2][0], pre->g_pre_comp[i][2][1],
+			pre->g_pre_comp[i][2][2]);
+		for (j = 1; j < 8; ++j)
+			{
+			/* odd multiples: add G resp. 2^28*G */
+			point_add(
+				pre->g_pre_comp[i][2*j+1][0], pre->g_pre_comp[i][2*j+1][1],
+				pre->g_pre_comp[i][2*j+1][2], pre->g_pre_comp[i][2*j][0],
+				pre->g_pre_comp[i][2*j][1], pre->g_pre_comp[i][2*j][2],
+				0, pre->g_pre_comp[i][1][0], pre->g_pre_comp[i][1][1],
+				pre->g_pre_comp[i][1][2]);
+			}
+		}
+	make_points_affine(31, &(pre->g_pre_comp[0][1]), tmp_felems);
+
+	if (!EC_EX_DATA_set_data(&group->extra_data, pre, nistp224_pre_comp_dup,
+			nistp224_pre_comp_free, nistp224_pre_comp_clear_free))
+		goto err;
+	ret = 1;
+	pre = NULL;
+ err:
+	BN_CTX_end(ctx);
+	if (generator != NULL)
+		EC_POINT_free(generator);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	if (pre)
+		nistp224_pre_comp_free(pre);
+	return ret;
+	}
+
+int ec_GFp_nistp224_have_precompute_mult(const EC_GROUP *group)
+	{
+	if (EC_EX_DATA_get_data(group->extra_data, nistp224_pre_comp_dup,
+			nistp224_pre_comp_free, nistp224_pre_comp_clear_free)
+		!= NULL)
+		return 1;
+	else
+		return 0;
+	}
+
+#else
+static void *dummy=&dummy;
+#endif
diff --git a/crypto/ec/ecp_nistp256.c b/crypto/ec/ecp_nistp256.c
new file mode 100644
index 000000000000..4bc0f5dce022
--- /dev/null
+++ b/crypto/ec/ecp_nistp256.c
@@ -0,0 +1,2171 @@
+/* crypto/ec/ecp_nistp256.c */
+/*
+ * Written by Adam Langley (Google) for the OpenSSL project
+ */
+/* Copyright 2011 Google Inc.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ *
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ *  Unless required by applicable law or agreed to in writing, software
+ *  distributed under the License is distributed on an "AS IS" BASIS,
+ *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ *  See the License for the specific language governing permissions and
+ *  limitations under the License.
+ */
+
+/*
+ * A 64-bit implementation of the NIST P-256 elliptic curve point multiplication
+ *
+ * OpenSSL integration was taken from Emilia Kasper's work in ecp_nistp224.c.
+ * Otherwise based on Emilia's P224 work, which was inspired by my curve25519
+ * work which got its smarts from Daniel J. Bernstein's work on the same.
+ */
+
+#include <openssl/opensslconf.h>
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+
+#ifndef OPENSSL_SYS_VMS
+#include <stdint.h>
+#else
+#include <inttypes.h>
+#endif
+
+#include <string.h>
+#include <openssl/err.h>
+#include "ec_lcl.h"
+
+#if defined(__GNUC__) && (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1))
+  /* even with gcc, the typedef won't work for 32-bit platforms */
+  typedef __uint128_t uint128_t; /* nonstandard; implemented by gcc on 64-bit platforms */
+  typedef __int128_t int128_t;
+#else
+  #error "Need GCC 3.1 or later to define type uint128_t"
+#endif
+
+typedef uint8_t u8;
+typedef uint32_t u32;
+typedef uint64_t u64;
+typedef int64_t s64;
+
+/* The underlying field.
+ *
+ * P256 operates over GF(2^256-2^224+2^192+2^96-1). We can serialise an element
+ * of this field into 32 bytes. We call this an felem_bytearray. */
+
+typedef u8 felem_bytearray[32];
+
+/* These are the parameters of P256, taken from FIPS 186-3, page 86. These
+ * values are big-endian. */
+static const felem_bytearray nistp256_curve_params[5] = {
+	{0xff, 0xff, 0xff, 0xff, 0x00, 0x00, 0x00, 0x01,       /* p */
+	 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	 0x00, 0x00, 0x00, 0x00, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff},
+	{0xff, 0xff, 0xff, 0xff, 0x00, 0x00, 0x00, 0x01,       /* a = -3 */
+	 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	 0x00, 0x00, 0x00, 0x00, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xfc},      /* b */
+	{0x5a, 0xc6, 0x35, 0xd8, 0xaa, 0x3a, 0x93, 0xe7,
+	 0xb3, 0xeb, 0xbd, 0x55, 0x76, 0x98, 0x86, 0xbc,
+	 0x65, 0x1d, 0x06, 0xb0, 0xcc, 0x53, 0xb0, 0xf6,
+	 0x3b, 0xce, 0x3c, 0x3e, 0x27, 0xd2, 0x60, 0x4b},
+	{0x6b, 0x17, 0xd1, 0xf2, 0xe1, 0x2c, 0x42, 0x47,       /* x */
+	 0xf8, 0xbc, 0xe6, 0xe5, 0x63, 0xa4, 0x40, 0xf2,
+	 0x77, 0x03, 0x7d, 0x81, 0x2d, 0xeb, 0x33, 0xa0,
+	 0xf4, 0xa1, 0x39, 0x45, 0xd8, 0x98, 0xc2, 0x96},
+	{0x4f, 0xe3, 0x42, 0xe2, 0xfe, 0x1a, 0x7f, 0x9b,       /* y */
+	 0x8e, 0xe7, 0xeb, 0x4a, 0x7c, 0x0f, 0x9e, 0x16,
+	 0x2b, 0xce, 0x33, 0x57, 0x6b, 0x31, 0x5e, 0xce,
+	 0xcb, 0xb6, 0x40, 0x68, 0x37, 0xbf, 0x51, 0xf5}
+};
+
+/* The representation of field elements.
+ * ------------------------------------
+ *
+ * We represent field elements with either four 128-bit values, eight 128-bit
+ * values, or four 64-bit values. The field element represented is:
+ *   v[0]*2^0 + v[1]*2^64 + v[2]*2^128 + v[3]*2^192  (mod p)
+ * or:
+ *   v[0]*2^0 + v[1]*2^64 + v[2]*2^128 + ... + v[8]*2^512  (mod p)
+ *
+ * 128-bit values are called 'limbs'. Since the limbs are spaced only 64 bits
+ * apart, but are 128-bits wide, the most significant bits of each limb overlap
+ * with the least significant bits of the next.
+ *
+ * A field element with four limbs is an 'felem'. One with eight limbs is a
+ * 'longfelem'
+ *
+ * A field element with four, 64-bit values is called a 'smallfelem'. Small
+ * values are used as intermediate values before multiplication.
+ */
+
+#define NLIMBS 4
+
+typedef uint128_t limb;
+typedef limb felem[NLIMBS];
+typedef limb longfelem[NLIMBS * 2];
+typedef u64 smallfelem[NLIMBS];
+
+/* This is the value of the prime as four 64-bit words, little-endian. */
+static const u64 kPrime[4] = { 0xfffffffffffffffful, 0xffffffff, 0, 0xffffffff00000001ul };
+static const limb bottom32bits = 0xffffffff;
+static const u64 bottom63bits = 0x7ffffffffffffffful;
+
+/* bin32_to_felem takes a little-endian byte array and converts it into felem
+ * form. This assumes that the CPU is little-endian. */
+static void bin32_to_felem(felem out, const u8 in[32])
+	{
+	out[0] = *((u64*) &in[0]);
+	out[1] = *((u64*) &in[8]);
+	out[2] = *((u64*) &in[16]);
+	out[3] = *((u64*) &in[24]);
+	}
+
+/* smallfelem_to_bin32 takes a smallfelem and serialises into a little endian,
+ * 32 byte array. This assumes that the CPU is little-endian. */
+static void smallfelem_to_bin32(u8 out[32], const smallfelem in)
+	{
+	*((u64*) &out[0]) = in[0];
+	*((u64*) &out[8]) = in[1];
+	*((u64*) &out[16]) = in[2];
+	*((u64*) &out[24]) = in[3];
+	}
+
+/* To preserve endianness when using BN_bn2bin and BN_bin2bn */
+static void flip_endian(u8 *out, const u8 *in, unsigned len)
+	{
+	unsigned i;
+	for (i = 0; i < len; ++i)
+		out[i] = in[len-1-i];
+	}
+
+/* BN_to_felem converts an OpenSSL BIGNUM into an felem */
+static int BN_to_felem(felem out, const BIGNUM *bn)
+	{
+	felem_bytearray b_in;
+	felem_bytearray b_out;
+	unsigned num_bytes;
+
+	/* BN_bn2bin eats leading zeroes */
+	memset(b_out, 0, sizeof b_out);
+	num_bytes = BN_num_bytes(bn);
+	if (num_bytes > sizeof b_out)
+		{
+		ECerr(EC_F_BN_TO_FELEM, EC_R_BIGNUM_OUT_OF_RANGE);
+		return 0;
+		}
+	if (BN_is_negative(bn))
+		{
+		ECerr(EC_F_BN_TO_FELEM, EC_R_BIGNUM_OUT_OF_RANGE);
+		return 0;
+		}
+	num_bytes = BN_bn2bin(bn, b_in);
+	flip_endian(b_out, b_in, num_bytes);
+	bin32_to_felem(out, b_out);
+	return 1;
+	}
+
+/* felem_to_BN converts an felem into an OpenSSL BIGNUM */
+static BIGNUM *smallfelem_to_BN(BIGNUM *out, const smallfelem in)
+	{
+	felem_bytearray b_in, b_out;
+	smallfelem_to_bin32(b_in, in);
+	flip_endian(b_out, b_in, sizeof b_out);
+	return BN_bin2bn(b_out, sizeof b_out, out);
+	}
+
+
+/* Field operations
+ * ---------------- */
+
+static void smallfelem_one(smallfelem out)
+	{
+	out[0] = 1;
+	out[1] = 0;
+	out[2] = 0;
+	out[3] = 0;
+	}
+
+static void smallfelem_assign(smallfelem out, const smallfelem in)
+	{
+	out[0] = in[0];
+	out[1] = in[1];
+	out[2] = in[2];
+	out[3] = in[3];
+	}
+
+static void felem_assign(felem out, const felem in)
+	{
+	out[0] = in[0];
+	out[1] = in[1];
+	out[2] = in[2];
+	out[3] = in[3];
+	}
+
+/* felem_sum sets out = out + in. */
+static void felem_sum(felem out, const felem in)
+	{
+	out[0] += in[0];
+	out[1] += in[1];
+	out[2] += in[2];
+	out[3] += in[3];
+	}
+
+/* felem_small_sum sets out = out + in. */
+static void felem_small_sum(felem out, const smallfelem in)
+	{
+	out[0] += in[0];
+	out[1] += in[1];
+	out[2] += in[2];
+	out[3] += in[3];
+	}
+
+/* felem_scalar sets out = out * scalar */
+static void felem_scalar(felem out, const u64 scalar)
+	{
+	out[0] *= scalar;
+	out[1] *= scalar;
+	out[2] *= scalar;
+	out[3] *= scalar;
+	}
+
+/* longfelem_scalar sets out = out * scalar */
+static void longfelem_scalar(longfelem out, const u64 scalar)
+	{
+	out[0] *= scalar;
+	out[1] *= scalar;
+	out[2] *= scalar;
+	out[3] *= scalar;
+	out[4] *= scalar;
+	out[5] *= scalar;
+	out[6] *= scalar;
+	out[7] *= scalar;
+	}
+
+#define two105m41m9 (((limb)1) << 105) - (((limb)1) << 41) - (((limb)1) << 9)
+#define two105 (((limb)1) << 105)
+#define two105m41p9 (((limb)1) << 105) - (((limb)1) << 41) + (((limb)1) << 9)
+
+/* zero105 is 0 mod p */
+static const felem zero105 = { two105m41m9, two105, two105m41p9, two105m41p9 };
+
+/* smallfelem_neg sets |out| to |-small|
+ * On exit:
+ *   out[i] < out[i] + 2^105
+ */
+static void smallfelem_neg(felem out, const smallfelem small)
+	{
+	/* In order to prevent underflow, we subtract from 0 mod p. */
+	out[0] = zero105[0] - small[0];
+	out[1] = zero105[1] - small[1];
+	out[2] = zero105[2] - small[2];
+	out[3] = zero105[3] - small[3];
+	}
+
+/* felem_diff subtracts |in| from |out|
+ * On entry:
+ *   in[i] < 2^104
+ * On exit:
+ *   out[i] < out[i] + 2^105
+ */
+static void felem_diff(felem out, const felem in)
+	{
+	/* In order to prevent underflow, we add 0 mod p before subtracting. */
+	out[0] += zero105[0];
+	out[1] += zero105[1];
+	out[2] += zero105[2];
+	out[3] += zero105[3];
+
+	out[0] -= in[0];
+	out[1] -= in[1];
+	out[2] -= in[2];
+	out[3] -= in[3];
+	}
+
+#define two107m43m11 (((limb)1) << 107) - (((limb)1) << 43) - (((limb)1) << 11)
+#define two107 (((limb)1) << 107)
+#define two107m43p11 (((limb)1) << 107) - (((limb)1) << 43) + (((limb)1) << 11)
+
+/* zero107 is 0 mod p */
+static const felem zero107 = { two107m43m11, two107, two107m43p11, two107m43p11 };
+
+/* An alternative felem_diff for larger inputs |in|
+ * felem_diff_zero107 subtracts |in| from |out|
+ * On entry:
+ *   in[i] < 2^106
+ * On exit:
+ *   out[i] < out[i] + 2^107
+ */
+static void felem_diff_zero107(felem out, const felem in)
+	{
+	/* In order to prevent underflow, we add 0 mod p before subtracting. */
+	out[0] += zero107[0];
+	out[1] += zero107[1];
+	out[2] += zero107[2];
+	out[3] += zero107[3];
+
+	out[0] -= in[0];
+	out[1] -= in[1];
+	out[2] -= in[2];
+	out[3] -= in[3];
+	}
+
+/* longfelem_diff subtracts |in| from |out|
+ * On entry:
+ *   in[i] < 7*2^67
+ * On exit:
+ *   out[i] < out[i] + 2^70 + 2^40
+ */
+static void longfelem_diff(longfelem out, const longfelem in)
+	{
+	static const limb two70m8p6 = (((limb)1) << 70) - (((limb)1) << 8) + (((limb)1) << 6);
+	static const limb two70p40 = (((limb)1) << 70) + (((limb)1) << 40);
+	static const limb two70 = (((limb)1) << 70);
+	static const limb two70m40m38p6 = (((limb)1) << 70) - (((limb)1) << 40) - (((limb)1) << 38) + (((limb)1) << 6);
+	static const limb two70m6 = (((limb)1) << 70) - (((limb)1) << 6);
+
+	/* add 0 mod p to avoid underflow */
+	out[0] += two70m8p6;
+	out[1] += two70p40;
+	out[2] += two70;
+	out[3] += two70m40m38p6;
+	out[4] += two70m6;
+	out[5] += two70m6;
+	out[6] += two70m6;
+	out[7] += two70m6;
+
+	/* in[i] < 7*2^67 < 2^70 - 2^40 - 2^38 + 2^6 */
+	out[0] -= in[0];
+	out[1] -= in[1];
+	out[2] -= in[2];
+	out[3] -= in[3];
+	out[4] -= in[4];
+	out[5] -= in[5];
+	out[6] -= in[6];
+	out[7] -= in[7];
+	}
+
+#define two64m0 (((limb)1) << 64) - 1
+#define two110p32m0 (((limb)1) << 110) + (((limb)1) << 32) - 1
+#define two64m46 (((limb)1) << 64) - (((limb)1) << 46)
+#define two64m32 (((limb)1) << 64) - (((limb)1) << 32)
+
+/* zero110 is 0 mod p */
+static const felem zero110 = { two64m0, two110p32m0, two64m46, two64m32 };
+
+/* felem_shrink converts an felem into a smallfelem. The result isn't quite
+ * minimal as the value may be greater than p.
+ *
+ * On entry:
+ *   in[i] < 2^109
+ * On exit:
+ *   out[i] < 2^64
+ */
+static void felem_shrink(smallfelem out, const felem in)
+	{
+	felem tmp;
+	u64 a, b, mask;
+	s64 high, low;
+	static const u64 kPrime3Test = 0x7fffffff00000001ul; /* 2^63 - 2^32 + 1 */
+
+	/* Carry 2->3 */
+	tmp[3] = zero110[3] + in[3] + ((u64) (in[2] >> 64));
+	/* tmp[3] < 2^110 */
+
+	tmp[2] = zero110[2] + (u64) in[2];
+	tmp[0] = zero110[0] + in[0];
+	tmp[1] = zero110[1] + in[1];
+	/* tmp[0] < 2**110, tmp[1] < 2^111, tmp[2] < 2**65 */
+
+	/* We perform two partial reductions where we eliminate the
+	 * high-word of tmp[3]. We don't update the other words till the end.
+	 */
+	a = tmp[3] >> 64; /* a < 2^46 */
+	tmp[3] = (u64) tmp[3];
+	tmp[3] -= a;
+	tmp[3] += ((limb)a) << 32;
+	/* tmp[3] < 2^79 */
+
+	b = a;
+	a = tmp[3] >> 64; /* a < 2^15 */
+	b += a; /* b < 2^46 + 2^15 < 2^47 */
+	tmp[3] = (u64) tmp[3];
+	tmp[3] -= a;
+	tmp[3] += ((limb)a) << 32;
+	/* tmp[3] < 2^64 + 2^47 */
+
+	/* This adjusts the other two words to complete the two partial
+	 * reductions. */
+	tmp[0] += b;
+	tmp[1] -= (((limb)b) << 32);
+
+	/* In order to make space in tmp[3] for the carry from 2 -> 3, we
+	 * conditionally subtract kPrime if tmp[3] is large enough. */
+	high = tmp[3] >> 64;
+	/* As tmp[3] < 2^65, high is either 1 or 0 */
+	high <<= 63;
+	high >>= 63;
+	/* high is:
+	 *   all ones   if the high word of tmp[3] is 1
+	 *   all zeros  if the high word of tmp[3] if 0 */
+	low = tmp[3];
+	mask = low >> 63;
+	/* mask is:
+	 *   all ones   if the MSB of low is 1
+	 *   all zeros  if the MSB of low if 0 */
+	low &= bottom63bits;
+	low -= kPrime3Test;
+	/* if low was greater than kPrime3Test then the MSB is zero */
+	low = ~low;
+	low >>= 63;
+	/* low is:
+	 *   all ones   if low was > kPrime3Test
+	 *   all zeros  if low was <= kPrime3Test */
+	mask = (mask & low) | high;
+	tmp[0] -= mask & kPrime[0];
+	tmp[1] -= mask & kPrime[1];
+	/* kPrime[2] is zero, so omitted */
+	tmp[3] -= mask & kPrime[3];
+	/* tmp[3] < 2**64 - 2**32 + 1 */
+
+	tmp[1] += ((u64) (tmp[0] >> 64)); tmp[0] = (u64) tmp[0];
+	tmp[2] += ((u64) (tmp[1] >> 64)); tmp[1] = (u64) tmp[1];
+	tmp[3] += ((u64) (tmp[2] >> 64)); tmp[2] = (u64) tmp[2];
+	/* tmp[i] < 2^64 */
+
+	out[0] = tmp[0];
+	out[1] = tmp[1];
+	out[2] = tmp[2];
+	out[3] = tmp[3];
+	}
+
+/* smallfelem_expand converts a smallfelem to an felem */
+static void smallfelem_expand(felem out, const smallfelem in)
+	{
+	out[0] = in[0];
+	out[1] = in[1];
+	out[2] = in[2];
+	out[3] = in[3];
+	}
+
+/* smallfelem_square sets |out| = |small|^2
+ * On entry:
+ *   small[i] < 2^64
+ * On exit:
+ *   out[i] < 7 * 2^64 < 2^67
+ */
+static void smallfelem_square(longfelem out, const smallfelem small)
+	{
+	limb a;
+	u64 high, low;
+
+	a = ((uint128_t) small[0]) * small[0];
+	low = a;
+	high = a >> 64;
+	out[0] = low;
+	out[1] = high;
+
+	a = ((uint128_t) small[0]) * small[1];
+	low = a;
+	high = a >> 64;
+	out[1] += low;
+	out[1] += low;
+	out[2] = high;
+
+	a = ((uint128_t) small[0]) * small[2];
+	low = a;
+	high = a >> 64;
+	out[2] += low;
+	out[2] *= 2;
+	out[3] = high;
+
+	a = ((uint128_t) small[0]) * small[3];
+	low = a;
+	high = a >> 64;
+	out[3] += low;
+	out[4] = high;
+
+	a = ((uint128_t) small[1]) * small[2];
+	low = a;
+	high = a >> 64;
+	out[3] += low;
+	out[3] *= 2;
+	out[4] += high;
+
+	a = ((uint128_t) small[1]) * small[1];
+	low = a;
+	high = a >> 64;
+	out[2] += low;
+	out[3] += high;
+
+	a = ((uint128_t) small[1]) * small[3];
+	low = a;
+	high = a >> 64;
+	out[4] += low;
+	out[4] *= 2;
+	out[5] = high;
+
+	a = ((uint128_t) small[2]) * small[3];
+	low = a;
+	high = a >> 64;
+	out[5] += low;
+	out[5] *= 2;
+	out[6] = high;
+	out[6] += high;
+
+	a = ((uint128_t) small[2]) * small[2];
+	low = a;
+	high = a >> 64;
+	out[4] += low;
+	out[5] += high;
+
+	a = ((uint128_t) small[3]) * small[3];
+	low = a;
+	high = a >> 64;
+	out[6] += low;
+	out[7] = high;
+	}
+
+/* felem_square sets |out| = |in|^2
+ * On entry:
+ *   in[i] < 2^109
+ * On exit:
+ *   out[i] < 7 * 2^64 < 2^67
+ */
+static void felem_square(longfelem out, const felem in)
+	{
+	u64 small[4];
+	felem_shrink(small, in);
+	smallfelem_square(out, small);
+	}
+
+/* smallfelem_mul sets |out| = |small1| * |small2|
+ * On entry:
+ *   small1[i] < 2^64
+ *   small2[i] < 2^64
+ * On exit:
+ *   out[i] < 7 * 2^64 < 2^67
+ */
+static void smallfelem_mul(longfelem out, const smallfelem small1, const smallfelem small2)
+	{
+	limb a;
+	u64 high, low;
+
+	a = ((uint128_t) small1[0]) * small2[0];
+	low = a;
+	high = a >> 64;
+	out[0] = low;
+	out[1] = high;
+
+
+	a = ((uint128_t) small1[0]) * small2[1];
+	low = a;
+	high = a >> 64;
+	out[1] += low;
+	out[2] = high;
+
+	a = ((uint128_t) small1[1]) * small2[0];
+	low = a;
+	high = a >> 64;
+	out[1] += low;
+	out[2] += high;
+
+
+	a = ((uint128_t) small1[0]) * small2[2];
+	low = a;
+	high = a >> 64;
+	out[2] += low;
+	out[3] = high;
+
+	a = ((uint128_t) small1[1]) * small2[1];
+	low = a;
+	high = a >> 64;
+	out[2] += low;
+	out[3] += high;
+
+	a = ((uint128_t) small1[2]) * small2[0];
+	low = a;
+	high = a >> 64;
+	out[2] += low;
+	out[3] += high;
+
+
+	a = ((uint128_t) small1[0]) * small2[3];
+	low = a;
+	high = a >> 64;
+	out[3] += low;
+	out[4] = high;
+
+	a = ((uint128_t) small1[1]) * small2[2];
+	low = a;
+	high = a >> 64;
+	out[3] += low;
+	out[4] += high;
+
+	a = ((uint128_t) small1[2]) * small2[1];
+	low = a;
+	high = a >> 64;
+	out[3] += low;
+	out[4] += high;
+
+	a = ((uint128_t) small1[3]) * small2[0];
+	low = a;
+	high = a >> 64;
+	out[3] += low;
+	out[4] += high;
+
+
+	a = ((uint128_t) small1[1]) * small2[3];
+	low = a;
+	high = a >> 64;
+	out[4] += low;
+	out[5] = high;
+
+	a = ((uint128_t) small1[2]) * small2[2];
+	low = a;
+	high = a >> 64;
+	out[4] += low;
+	out[5] += high;
+
+	a = ((uint128_t) small1[3]) * small2[1];
+	low = a;
+	high = a >> 64;
+	out[4] += low;
+	out[5] += high;
+
+
+	a = ((uint128_t) small1[2]) * small2[3];
+	low = a;
+	high = a >> 64;
+	out[5] += low;
+	out[6] = high;
+
+	a = ((uint128_t) small1[3]) * small2[2];
+	low = a;
+	high = a >> 64;
+	out[5] += low;
+	out[6] += high;
+
+
+	a = ((uint128_t) small1[3]) * small2[3];
+	low = a;
+	high = a >> 64;
+	out[6] += low;
+	out[7] = high;
+	}
+
+/* felem_mul sets |out| = |in1| * |in2|
+ * On entry:
+ *   in1[i] < 2^109
+ *   in2[i] < 2^109
+ * On exit:
+ *   out[i] < 7 * 2^64 < 2^67
+ */
+static void felem_mul(longfelem out, const felem in1, const felem in2)
+	{
+	smallfelem small1, small2;
+	felem_shrink(small1, in1);
+	felem_shrink(small2, in2);
+	smallfelem_mul(out, small1, small2);
+	}
+
+/* felem_small_mul sets |out| = |small1| * |in2|
+ * On entry:
+ *   small1[i] < 2^64
+ *   in2[i] < 2^109
+ * On exit:
+ *   out[i] < 7 * 2^64 < 2^67
+ */
+static void felem_small_mul(longfelem out, const smallfelem small1, const felem in2)
+	{
+	smallfelem small2;
+	felem_shrink(small2, in2);
+	smallfelem_mul(out, small1, small2);
+	}
+
+#define two100m36m4 (((limb)1) << 100) - (((limb)1) << 36) - (((limb)1) << 4)
+#define two100 (((limb)1) << 100)
+#define two100m36p4 (((limb)1) << 100) - (((limb)1) << 36) + (((limb)1) << 4)
+/* zero100 is 0 mod p */
+static const felem zero100 = { two100m36m4, two100, two100m36p4, two100m36p4 };
+
+/* Internal function for the different flavours of felem_reduce.
+ * felem_reduce_ reduces the higher coefficients in[4]-in[7].
+ * On entry:
+ *   out[0] >= in[6] + 2^32*in[6] + in[7] + 2^32*in[7] 
+ *   out[1] >= in[7] + 2^32*in[4]
+ *   out[2] >= in[5] + 2^32*in[5]
+ *   out[3] >= in[4] + 2^32*in[5] + 2^32*in[6]
+ * On exit:
+ *   out[0] <= out[0] + in[4] + 2^32*in[5]
+ *   out[1] <= out[1] + in[5] + 2^33*in[6]
+ *   out[2] <= out[2] + in[7] + 2*in[6] + 2^33*in[7]
+ *   out[3] <= out[3] + 2^32*in[4] + 3*in[7]
+ */
+static void felem_reduce_(felem out, const longfelem in)
+	{
+	int128_t c;
+	/* combine common terms from below */
+	c = in[4] + (in[5] << 32);
+	out[0] += c;
+	out[3] -= c;
+
+	c = in[5] - in[7];
+	out[1] += c;
+	out[2] -= c;
+
+	/* the remaining terms */
+	/* 256: [(0,1),(96,-1),(192,-1),(224,1)] */
+	out[1] -= (in[4] << 32);
+	out[3] += (in[4] << 32);
+
+	/* 320: [(32,1),(64,1),(128,-1),(160,-1),(224,-1)] */
+	out[2] -= (in[5] << 32);
+
+	/* 384: [(0,-1),(32,-1),(96,2),(128,2),(224,-1)] */
+	out[0] -= in[6];
+	out[0] -= (in[6] << 32);
+	out[1] += (in[6] << 33);
+	out[2] += (in[6] * 2);
+	out[3] -= (in[6] << 32);
+
+	/* 448: [(0,-1),(32,-1),(64,-1),(128,1),(160,2),(192,3)] */
+	out[0] -= in[7];
+	out[0] -= (in[7] << 32);
+	out[2] += (in[7] << 33);
+	out[3] += (in[7] * 3);
+	}
+
+/* felem_reduce converts a longfelem into an felem.
+ * To be called directly after felem_square or felem_mul.
+ * On entry:
+ *   in[0] < 2^64, in[1] < 3*2^64, in[2] < 5*2^64, in[3] < 7*2^64
+ *   in[4] < 7*2^64, in[5] < 5*2^64, in[6] < 3*2^64, in[7] < 2*64
+ * On exit:
+ *   out[i] < 2^101
+ */
+static void felem_reduce(felem out, const longfelem in)
+	{
+	out[0] = zero100[0] + in[0];
+	out[1] = zero100[1] + in[1];
+	out[2] = zero100[2] + in[2];
+	out[3] = zero100[3] + in[3];
+
+	felem_reduce_(out, in);
+
+	/* out[0] > 2^100 - 2^36 - 2^4 - 3*2^64 - 3*2^96 - 2^64 - 2^96 > 0
+	 * out[1] > 2^100 - 2^64 - 7*2^96 > 0
+	 * out[2] > 2^100 - 2^36 + 2^4 - 5*2^64 - 5*2^96 > 0
+	 * out[3] > 2^100 - 2^36 + 2^4 - 7*2^64 - 5*2^96 - 3*2^96 > 0
+	 *
+	 * out[0] < 2^100 + 2^64 + 7*2^64 + 5*2^96 < 2^101
+	 * out[1] < 2^100 + 3*2^64 + 5*2^64 + 3*2^97 < 2^101
+	 * out[2] < 2^100 + 5*2^64 + 2^64 + 3*2^65 + 2^97 < 2^101
+	 * out[3] < 2^100 + 7*2^64 + 7*2^96 + 3*2^64 < 2^101
+	 */
+	}
+
+/* felem_reduce_zero105 converts a larger longfelem into an felem.
+ * On entry:
+ *   in[0] < 2^71
+ * On exit:
+ *   out[i] < 2^106
+ */
+static void felem_reduce_zero105(felem out, const longfelem in)
+	{
+	out[0] = zero105[0] + in[0];
+	out[1] = zero105[1] + in[1];
+	out[2] = zero105[2] + in[2];
+	out[3] = zero105[3] + in[3];
+
+	felem_reduce_(out, in);
+
+	/* out[0] > 2^105 - 2^41 - 2^9 - 2^71 - 2^103 - 2^71 - 2^103 > 0
+	 * out[1] > 2^105 - 2^71 - 2^103 > 0
+	 * out[2] > 2^105 - 2^41 + 2^9 - 2^71 - 2^103 > 0
+	 * out[3] > 2^105 - 2^41 + 2^9 - 2^71 - 2^103 - 2^103 > 0
+	 *
+	 * out[0] < 2^105 + 2^71 + 2^71 + 2^103 < 2^106
+	 * out[1] < 2^105 + 2^71 + 2^71 + 2^103 < 2^106
+	 * out[2] < 2^105 + 2^71 + 2^71 + 2^71 + 2^103 < 2^106
+	 * out[3] < 2^105 + 2^71 + 2^103 + 2^71 < 2^106
+	 */
+	}
+
+/* subtract_u64 sets *result = *result - v and *carry to one if the subtraction
+ * underflowed. */
+static void subtract_u64(u64* result, u64* carry, u64 v)
+	{
+	uint128_t r = *result;
+	r -= v;
+	*carry = (r >> 64) & 1;
+	*result = (u64) r;
+	}
+
+/* felem_contract converts |in| to its unique, minimal representation.
+ * On entry:
+ *   in[i] < 2^109
+ */
+static void felem_contract(smallfelem out, const felem in)
+	{
+	unsigned i;
+	u64 all_equal_so_far = 0, result = 0, carry;
+
+	felem_shrink(out, in);
+	/* small is minimal except that the value might be > p */
+
+	all_equal_so_far--;
+	/* We are doing a constant time test if out >= kPrime. We need to
+	 * compare each u64, from most-significant to least significant. For
+	 * each one, if all words so far have been equal (m is all ones) then a
+	 * non-equal result is the answer. Otherwise we continue. */
+	for (i = 3; i < 4; i--)
+		{
+		u64 equal;
+		uint128_t a = ((uint128_t) kPrime[i]) - out[i];
+		/* if out[i] > kPrime[i] then a will underflow and the high
+		 * 64-bits will all be set. */
+		result |= all_equal_so_far & ((u64) (a >> 64));
+
+		/* if kPrime[i] == out[i] then |equal| will be all zeros and
+		 * the decrement will make it all ones. */
+		equal = kPrime[i] ^ out[i];
+		equal--;
+		equal &= equal << 32;
+		equal &= equal << 16;
+		equal &= equal << 8;
+		equal &= equal << 4;
+		equal &= equal << 2;
+		equal &= equal << 1;
+		equal = ((s64) equal) >> 63;
+
+		all_equal_so_far &= equal;
+		}
+
+	/* if all_equal_so_far is still all ones then the two values are equal
+	 * and so out >= kPrime is true. */
+	result |= all_equal_so_far;
+
+	/* if out >= kPrime then we subtract kPrime. */
+	subtract_u64(&out[0], &carry, result & kPrime[0]);
+	subtract_u64(&out[1], &carry, carry);
+	subtract_u64(&out[2], &carry, carry);
+	subtract_u64(&out[3], &carry, carry);
+
+	subtract_u64(&out[1], &carry, result & kPrime[1]);
+	subtract_u64(&out[2], &carry, carry);
+	subtract_u64(&out[3], &carry, carry);
+
+	subtract_u64(&out[2], &carry, result & kPrime[2]);
+	subtract_u64(&out[3], &carry, carry);
+
+	subtract_u64(&out[3], &carry, result & kPrime[3]);
+	}
+
+static void smallfelem_square_contract(smallfelem out, const smallfelem in)
+	{
+	longfelem longtmp;
+	felem tmp;
+
+	smallfelem_square(longtmp, in);
+	felem_reduce(tmp, longtmp);
+	felem_contract(out, tmp);
+	}
+
+static void smallfelem_mul_contract(smallfelem out, const smallfelem in1, const smallfelem in2)
+	{
+	longfelem longtmp;
+	felem tmp;
+
+	smallfelem_mul(longtmp, in1, in2);
+	felem_reduce(tmp, longtmp);
+	felem_contract(out, tmp);
+	}
+
+/* felem_is_zero returns a limb with all bits set if |in| == 0 (mod p) and 0
+ * otherwise.
+ * On entry:
+ *   small[i] < 2^64
+ */
+static limb smallfelem_is_zero(const smallfelem small)
+	{
+	limb result;
+	u64 is_p;
+
+	u64 is_zero = small[0] | small[1] | small[2] | small[3];
+	is_zero--;
+	is_zero &= is_zero << 32;
+	is_zero &= is_zero << 16;
+	is_zero &= is_zero << 8;
+	is_zero &= is_zero << 4;
+	is_zero &= is_zero << 2;
+	is_zero &= is_zero << 1;
+	is_zero = ((s64) is_zero) >> 63;
+
+	is_p = (small[0] ^ kPrime[0]) |
+	       (small[1] ^ kPrime[1]) |
+	       (small[2] ^ kPrime[2]) |
+	       (small[3] ^ kPrime[3]);
+	is_p--;
+	is_p &= is_p << 32;
+	is_p &= is_p << 16;
+	is_p &= is_p << 8;
+	is_p &= is_p << 4;
+	is_p &= is_p << 2;
+	is_p &= is_p << 1;
+	is_p = ((s64) is_p) >> 63;
+
+	is_zero |= is_p;
+
+	result = is_zero;
+	result |= ((limb) is_zero) << 64;
+	return result;
+	}
+
+static int smallfelem_is_zero_int(const smallfelem small)
+	{
+	return (int) (smallfelem_is_zero(small) & ((limb)1));
+	}
+
+/* felem_inv calculates |out| = |in|^{-1}
+ *
+ * Based on Fermat's Little Theorem:
+ *   a^p = a (mod p)
+ *   a^{p-1} = 1 (mod p)
+ *   a^{p-2} = a^{-1} (mod p)
+ */
+static void felem_inv(felem out, const felem in)
+	{
+	felem ftmp, ftmp2;
+	/* each e_I will hold |in|^{2^I - 1} */
+	felem e2, e4, e8, e16, e32, e64;
+	longfelem tmp;
+	unsigned i;
+
+	felem_square(tmp, in); felem_reduce(ftmp, tmp);			/* 2^1 */
+	felem_mul(tmp, in, ftmp); felem_reduce(ftmp, tmp);		/* 2^2 - 2^0 */
+	felem_assign(e2, ftmp);
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);		/* 2^3 - 2^1 */
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);		/* 2^4 - 2^2 */
+	felem_mul(tmp, ftmp, e2); felem_reduce(ftmp, tmp);		/* 2^4 - 2^0 */
+	felem_assign(e4, ftmp);
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);		/* 2^5 - 2^1 */
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);		/* 2^6 - 2^2 */
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);		/* 2^7 - 2^3 */
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);		/* 2^8 - 2^4 */
+	felem_mul(tmp, ftmp, e4); felem_reduce(ftmp, tmp);		/* 2^8 - 2^0 */
+	felem_assign(e8, ftmp);
+	for (i = 0; i < 8; i++) {
+		felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);
+	}								/* 2^16 - 2^8 */
+	felem_mul(tmp, ftmp, e8); felem_reduce(ftmp, tmp);		/* 2^16 - 2^0 */
+	felem_assign(e16, ftmp);
+	for (i = 0; i < 16; i++) {
+		felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);
+	}								/* 2^32 - 2^16 */
+	felem_mul(tmp, ftmp, e16); felem_reduce(ftmp, tmp);		/* 2^32 - 2^0 */
+	felem_assign(e32, ftmp);
+	for (i = 0; i < 32; i++) {
+		felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);
+	}								/* 2^64 - 2^32 */
+	felem_assign(e64, ftmp);
+	felem_mul(tmp, ftmp, in); felem_reduce(ftmp, tmp);		/* 2^64 - 2^32 + 2^0 */
+	for (i = 0; i < 192; i++) {
+		felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);
+	}								/* 2^256 - 2^224 + 2^192 */
+
+	felem_mul(tmp, e64, e32); felem_reduce(ftmp2, tmp);		/* 2^64 - 2^0 */
+	for (i = 0; i < 16; i++) {
+		felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);
+	}								/* 2^80 - 2^16 */
+	felem_mul(tmp, ftmp2, e16); felem_reduce(ftmp2, tmp);		/* 2^80 - 2^0 */
+	for (i = 0; i < 8; i++) {
+		felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);
+	}								/* 2^88 - 2^8 */
+	felem_mul(tmp, ftmp2, e8); felem_reduce(ftmp2, tmp);		/* 2^88 - 2^0 */
+	for (i = 0; i < 4; i++) {
+		felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);
+	}								/* 2^92 - 2^4 */
+	felem_mul(tmp, ftmp2, e4); felem_reduce(ftmp2, tmp);		/* 2^92 - 2^0 */
+	felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);		/* 2^93 - 2^1 */
+	felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);		/* 2^94 - 2^2 */
+	felem_mul(tmp, ftmp2, e2); felem_reduce(ftmp2, tmp);		/* 2^94 - 2^0 */
+	felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);		/* 2^95 - 2^1 */
+	felem_square(tmp, ftmp2); felem_reduce(ftmp2, tmp);		/* 2^96 - 2^2 */
+	felem_mul(tmp, ftmp2, in); felem_reduce(ftmp2, tmp);		/* 2^96 - 3 */
+
+	felem_mul(tmp, ftmp2, ftmp); felem_reduce(out, tmp); /* 2^256 - 2^224 + 2^192 + 2^96 - 3 */
+	}
+
+static void smallfelem_inv_contract(smallfelem out, const smallfelem in)
+	{
+	felem tmp;
+
+	smallfelem_expand(tmp, in);
+	felem_inv(tmp, tmp);
+	felem_contract(out, tmp);
+	}
+
+/* Group operations
+ * ----------------
+ *
+ * Building on top of the field operations we have the operations on the
+ * elliptic curve group itself. Points on the curve are represented in Jacobian
+ * coordinates */
+
+/* point_double calculates 2*(x_in, y_in, z_in)
+ *
+ * The method is taken from:
+ *   http://hyperelliptic.org/EFD/g1p/auto-shortw-jacobian-3.html#doubling-dbl-2001-b
+ *
+ * Outputs can equal corresponding inputs, i.e., x_out == x_in is allowed.
+ * while x_out == y_in is not (maybe this works, but it's not tested). */
+static void
+point_double(felem x_out, felem y_out, felem z_out,
+	     const felem x_in, const felem y_in, const felem z_in)
+	{
+	longfelem tmp, tmp2;
+	felem delta, gamma, beta, alpha, ftmp, ftmp2;
+	smallfelem small1, small2;
+
+	felem_assign(ftmp, x_in);
+	/* ftmp[i] < 2^106 */
+	felem_assign(ftmp2, x_in);
+	/* ftmp2[i] < 2^106 */
+
+	/* delta = z^2 */
+	felem_square(tmp, z_in);
+	felem_reduce(delta, tmp);
+	/* delta[i] < 2^101 */
+
+	/* gamma = y^2 */
+	felem_square(tmp, y_in);
+	felem_reduce(gamma, tmp);
+	/* gamma[i] < 2^101 */
+	felem_shrink(small1, gamma);
+
+	/* beta = x*gamma */
+	felem_small_mul(tmp, small1, x_in);
+	felem_reduce(beta, tmp);
+	/* beta[i] < 2^101 */
+
+	/* alpha = 3*(x-delta)*(x+delta) */
+	felem_diff(ftmp, delta);
+	/* ftmp[i] < 2^105 + 2^106 < 2^107 */
+	felem_sum(ftmp2, delta);
+	/* ftmp2[i] < 2^105 + 2^106 < 2^107 */
+	felem_scalar(ftmp2, 3);
+	/* ftmp2[i] < 3 * 2^107 < 2^109 */
+	felem_mul(tmp, ftmp, ftmp2);
+	felem_reduce(alpha, tmp);
+	/* alpha[i] < 2^101 */
+	felem_shrink(small2, alpha);
+
+	/* x' = alpha^2 - 8*beta */
+	smallfelem_square(tmp, small2);
+	felem_reduce(x_out, tmp);
+	felem_assign(ftmp, beta);
+	felem_scalar(ftmp, 8);
+	/* ftmp[i] < 8 * 2^101 = 2^104 */
+	felem_diff(x_out, ftmp);
+	/* x_out[i] < 2^105 + 2^101 < 2^106 */
+
+	/* z' = (y + z)^2 - gamma - delta */
+	felem_sum(delta, gamma);
+	/* delta[i] < 2^101 + 2^101 = 2^102 */
+	felem_assign(ftmp, y_in);
+	felem_sum(ftmp, z_in);
+	/* ftmp[i] < 2^106 + 2^106 = 2^107 */
+	felem_square(tmp, ftmp);
+	felem_reduce(z_out, tmp);
+	felem_diff(z_out, delta);
+	/* z_out[i] < 2^105 + 2^101 < 2^106 */
+
+	/* y' = alpha*(4*beta - x') - 8*gamma^2 */
+	felem_scalar(beta, 4);
+	/* beta[i] < 4 * 2^101 = 2^103 */
+	felem_diff_zero107(beta, x_out);
+	/* beta[i] < 2^107 + 2^103 < 2^108 */
+	felem_small_mul(tmp, small2, beta);
+	/* tmp[i] < 7 * 2^64 < 2^67 */
+	smallfelem_square(tmp2, small1);
+	/* tmp2[i] < 7 * 2^64 */
+	longfelem_scalar(tmp2, 8);
+	/* tmp2[i] < 8 * 7 * 2^64 = 7 * 2^67 */
+	longfelem_diff(tmp, tmp2);
+	/* tmp[i] < 2^67 + 2^70 + 2^40 < 2^71 */
+	felem_reduce_zero105(y_out, tmp);
+	/* y_out[i] < 2^106 */
+	}
+
+/* point_double_small is the same as point_double, except that it operates on
+ * smallfelems */
+static void
+point_double_small(smallfelem x_out, smallfelem y_out, smallfelem z_out,
+		   const smallfelem x_in, const smallfelem y_in, const smallfelem z_in)
+	{
+	felem felem_x_out, felem_y_out, felem_z_out;
+	felem felem_x_in, felem_y_in, felem_z_in;
+
+	smallfelem_expand(felem_x_in, x_in);
+	smallfelem_expand(felem_y_in, y_in);
+	smallfelem_expand(felem_z_in, z_in);
+	point_double(felem_x_out, felem_y_out, felem_z_out,
+		     felem_x_in, felem_y_in, felem_z_in);
+	felem_shrink(x_out, felem_x_out);
+	felem_shrink(y_out, felem_y_out);
+	felem_shrink(z_out, felem_z_out);
+	}
+
+/* copy_conditional copies in to out iff mask is all ones. */
+static void
+copy_conditional(felem out, const felem in, limb mask)
+	{
+	unsigned i;
+	for (i = 0; i < NLIMBS; ++i)
+		{
+		const limb tmp = mask & (in[i] ^ out[i]);
+		out[i] ^= tmp;
+		}
+	}
+
+/* copy_small_conditional copies in to out iff mask is all ones. */
+static void
+copy_small_conditional(felem out, const smallfelem in, limb mask)
+	{
+	unsigned i;
+	const u64 mask64 = mask;
+	for (i = 0; i < NLIMBS; ++i)
+		{
+		out[i] = ((limb) (in[i] & mask64)) | (out[i] & ~mask);
+		}
+	}
+
+/* point_add calcuates (x1, y1, z1) + (x2, y2, z2)
+ *
+ * The method is taken from:
+ *   http://hyperelliptic.org/EFD/g1p/auto-shortw-jacobian-3.html#addition-add-2007-bl,
+ * adapted for mixed addition (z2 = 1, or z2 = 0 for the point at infinity).
+ *
+ * This function includes a branch for checking whether the two input points
+ * are equal, (while not equal to the point at infinity). This case never
+ * happens during single point multiplication, so there is no timing leak for
+ * ECDH or ECDSA signing. */
+static void point_add(felem x3, felem y3, felem z3,
+	const felem x1, const felem y1, const felem z1,
+	const int mixed, const smallfelem x2, const smallfelem y2, const smallfelem z2)
+	{
+	felem ftmp, ftmp2, ftmp3, ftmp4, ftmp5, ftmp6, x_out, y_out, z_out;
+	longfelem tmp, tmp2;
+	smallfelem small1, small2, small3, small4, small5;
+	limb x_equal, y_equal, z1_is_zero, z2_is_zero;
+
+	felem_shrink(small3, z1);
+
+	z1_is_zero = smallfelem_is_zero(small3);
+	z2_is_zero = smallfelem_is_zero(z2);
+
+	/* ftmp = z1z1 = z1**2 */
+	smallfelem_square(tmp, small3);
+	felem_reduce(ftmp, tmp);
+	/* ftmp[i] < 2^101 */
+	felem_shrink(small1, ftmp);
+
+	if(!mixed)
+		{
+		/* ftmp2 = z2z2 = z2**2 */
+		smallfelem_square(tmp, z2);
+		felem_reduce(ftmp2, tmp);
+		/* ftmp2[i] < 2^101 */
+		felem_shrink(small2, ftmp2);
+
+		felem_shrink(small5, x1);
+
+		/* u1 = ftmp3 = x1*z2z2 */
+		smallfelem_mul(tmp, small5, small2);
+		felem_reduce(ftmp3, tmp);
+		/* ftmp3[i] < 2^101 */
+
+		/* ftmp5 = z1 + z2 */
+		felem_assign(ftmp5, z1);
+		felem_small_sum(ftmp5, z2);
+		/* ftmp5[i] < 2^107 */
+
+		/* ftmp5 = (z1 + z2)**2 - (z1z1 + z2z2) = 2z1z2 */
+		felem_square(tmp, ftmp5);
+		felem_reduce(ftmp5, tmp);
+		/* ftmp2 = z2z2 + z1z1 */
+		felem_sum(ftmp2, ftmp);
+		/* ftmp2[i] < 2^101 + 2^101 = 2^102 */
+		felem_diff(ftmp5, ftmp2);
+		/* ftmp5[i] < 2^105 + 2^101 < 2^106 */
+
+		/* ftmp2 = z2 * z2z2 */
+		smallfelem_mul(tmp, small2, z2);
+		felem_reduce(ftmp2, tmp);
+
+		/* s1 = ftmp2 = y1 * z2**3 */
+		felem_mul(tmp, y1, ftmp2);
+		felem_reduce(ftmp6, tmp);
+		/* ftmp6[i] < 2^101 */
+		}
+	else
+		{
+		/* We'll assume z2 = 1 (special case z2 = 0 is handled later) */
+
+		/* u1 = ftmp3 = x1*z2z2 */
+		felem_assign(ftmp3, x1);
+		/* ftmp3[i] < 2^106 */
+
+		/* ftmp5 = 2z1z2 */
+		felem_assign(ftmp5, z1);
+		felem_scalar(ftmp5, 2);
+		/* ftmp5[i] < 2*2^106 = 2^107 */
+
+		/* s1 = ftmp2 = y1 * z2**3 */
+		felem_assign(ftmp6, y1);
+		/* ftmp6[i] < 2^106 */
+		}
+
+	/* u2 = x2*z1z1 */
+	smallfelem_mul(tmp, x2, small1);
+	felem_reduce(ftmp4, tmp);
+
+	/* h = ftmp4 = u2 - u1 */
+	felem_diff_zero107(ftmp4, ftmp3);
+	/* ftmp4[i] < 2^107 + 2^101 < 2^108 */
+	felem_shrink(small4, ftmp4);
+
+	x_equal = smallfelem_is_zero(small4);
+
+	/* z_out = ftmp5 * h */
+	felem_small_mul(tmp, small4, ftmp5);
+	felem_reduce(z_out, tmp);
+	/* z_out[i] < 2^101 */
+
+	/* ftmp = z1 * z1z1 */
+	smallfelem_mul(tmp, small1, small3);
+	felem_reduce(ftmp, tmp);
+
+	/* s2 = tmp = y2 * z1**3 */
+	felem_small_mul(tmp, y2, ftmp);
+	felem_reduce(ftmp5, tmp);
+
+	/* r = ftmp5 = (s2 - s1)*2 */
+	felem_diff_zero107(ftmp5, ftmp6);
+	/* ftmp5[i] < 2^107 + 2^107 = 2^108*/
+	felem_scalar(ftmp5, 2);
+	/* ftmp5[i] < 2^109 */
+	felem_shrink(small1, ftmp5);
+	y_equal = smallfelem_is_zero(small1);
+
+	if (x_equal && y_equal && !z1_is_zero && !z2_is_zero)
+		{
+		point_double(x3, y3, z3, x1, y1, z1);
+		return;
+		}
+
+	/* I = ftmp = (2h)**2 */
+	felem_assign(ftmp, ftmp4);
+	felem_scalar(ftmp, 2);
+	/* ftmp[i] < 2*2^108 = 2^109 */
+	felem_square(tmp, ftmp);
+	felem_reduce(ftmp, tmp);
+
+	/* J = ftmp2 = h * I */
+	felem_mul(tmp, ftmp4, ftmp);
+	felem_reduce(ftmp2, tmp);
+
+	/* V = ftmp4 = U1 * I */
+	felem_mul(tmp, ftmp3, ftmp);
+	felem_reduce(ftmp4, tmp);
+
+	/* x_out = r**2 - J - 2V */
+	smallfelem_square(tmp, small1);
+	felem_reduce(x_out, tmp);
+	felem_assign(ftmp3, ftmp4);
+	felem_scalar(ftmp4, 2);
+	felem_sum(ftmp4, ftmp2);
+	/* ftmp4[i] < 2*2^101 + 2^101 < 2^103 */
+	felem_diff(x_out, ftmp4);
+	/* x_out[i] < 2^105 + 2^101 */
+
+	/* y_out = r(V-x_out) - 2 * s1 * J */
+	felem_diff_zero107(ftmp3, x_out);
+	/* ftmp3[i] < 2^107 + 2^101 < 2^108 */
+	felem_small_mul(tmp, small1, ftmp3);
+	felem_mul(tmp2, ftmp6, ftmp2);
+	longfelem_scalar(tmp2, 2);
+	/* tmp2[i] < 2*2^67 = 2^68 */
+	longfelem_diff(tmp, tmp2);
+	/* tmp[i] < 2^67 + 2^70 + 2^40 < 2^71 */
+	felem_reduce_zero105(y_out, tmp);
+	/* y_out[i] < 2^106 */
+
+	copy_small_conditional(x_out, x2, z1_is_zero);
+	copy_conditional(x_out, x1, z2_is_zero);
+	copy_small_conditional(y_out, y2, z1_is_zero);
+	copy_conditional(y_out, y1, z2_is_zero);
+	copy_small_conditional(z_out, z2, z1_is_zero);
+	copy_conditional(z_out, z1, z2_is_zero);
+	felem_assign(x3, x_out);
+	felem_assign(y3, y_out);
+	felem_assign(z3, z_out);
+	}
+
+/* point_add_small is the same as point_add, except that it operates on
+ * smallfelems */
+static void point_add_small(smallfelem x3, smallfelem y3, smallfelem z3,
+			    smallfelem x1, smallfelem y1, smallfelem z1,
+			    smallfelem x2, smallfelem y2, smallfelem z2)
+	{
+	felem felem_x3, felem_y3, felem_z3;
+	felem felem_x1, felem_y1, felem_z1;
+	smallfelem_expand(felem_x1, x1);
+	smallfelem_expand(felem_y1, y1);
+	smallfelem_expand(felem_z1, z1);
+	point_add(felem_x3, felem_y3, felem_z3, felem_x1, felem_y1, felem_z1, 0, x2, y2, z2);
+	felem_shrink(x3, felem_x3);
+	felem_shrink(y3, felem_y3);
+	felem_shrink(z3, felem_z3);
+	}
+
+/* Base point pre computation
+ * --------------------------
+ *
+ * Two different sorts of precomputed tables are used in the following code.
+ * Each contain various points on the curve, where each point is three field
+ * elements (x, y, z).
+ *
+ * For the base point table, z is usually 1 (0 for the point at infinity).
+ * This table has 2 * 16 elements, starting with the following:
+ * index | bits    | point
+ * ------+---------+------------------------------
+ *     0 | 0 0 0 0 | 0G
+ *     1 | 0 0 0 1 | 1G
+ *     2 | 0 0 1 0 | 2^64G
+ *     3 | 0 0 1 1 | (2^64 + 1)G
+ *     4 | 0 1 0 0 | 2^128G
+ *     5 | 0 1 0 1 | (2^128 + 1)G
+ *     6 | 0 1 1 0 | (2^128 + 2^64)G
+ *     7 | 0 1 1 1 | (2^128 + 2^64 + 1)G
+ *     8 | 1 0 0 0 | 2^192G
+ *     9 | 1 0 0 1 | (2^192 + 1)G
+ *    10 | 1 0 1 0 | (2^192 + 2^64)G
+ *    11 | 1 0 1 1 | (2^192 + 2^64 + 1)G
+ *    12 | 1 1 0 0 | (2^192 + 2^128)G
+ *    13 | 1 1 0 1 | (2^192 + 2^128 + 1)G
+ *    14 | 1 1 1 0 | (2^192 + 2^128 + 2^64)G
+ *    15 | 1 1 1 1 | (2^192 + 2^128 + 2^64 + 1)G
+ * followed by a copy of this with each element multiplied by 2^32.
+ *
+ * The reason for this is so that we can clock bits into four different
+ * locations when doing simple scalar multiplies against the base point,
+ * and then another four locations using the second 16 elements.
+ *
+ * Tables for other points have table[i] = iG for i in 0 .. 16. */
+
+/* gmul is the table of precomputed base points */
+static const smallfelem gmul[2][16][3] =
+{{{{0, 0, 0, 0},
+   {0, 0, 0, 0},
+   {0, 0, 0, 0}},
+  {{0xf4a13945d898c296, 0x77037d812deb33a0, 0xf8bce6e563a440f2, 0x6b17d1f2e12c4247},
+   {0xcbb6406837bf51f5, 0x2bce33576b315ece, 0x8ee7eb4a7c0f9e16, 0x4fe342e2fe1a7f9b},
+   {1, 0, 0, 0}},
+  {{0x90e75cb48e14db63, 0x29493baaad651f7e, 0x8492592e326e25de, 0x0fa822bc2811aaa5},
+   {0xe41124545f462ee7, 0x34b1a65050fe82f5, 0x6f4ad4bcb3df188b, 0xbff44ae8f5dba80d},
+   {1, 0, 0, 0}},
+  {{0x93391ce2097992af, 0xe96c98fd0d35f1fa, 0xb257c0de95e02789, 0x300a4bbc89d6726f},
+   {0xaa54a291c08127a0, 0x5bb1eeada9d806a5, 0x7f1ddb25ff1e3c6f, 0x72aac7e0d09b4644},
+   {1, 0, 0, 0}},
+  {{0x57c84fc9d789bd85, 0xfc35ff7dc297eac3, 0xfb982fd588c6766e, 0x447d739beedb5e67},
+   {0x0c7e33c972e25b32, 0x3d349b95a7fae500, 0xe12e9d953a4aaff7, 0x2d4825ab834131ee},
+   {1, 0, 0, 0}},
+  {{0x13949c932a1d367f, 0xef7fbd2b1a0a11b7, 0xddc6068bb91dfc60, 0xef9519328a9c72ff},
+   {0x196035a77376d8a8, 0x23183b0895ca1740, 0xc1ee9807022c219c, 0x611e9fc37dbb2c9b},
+   {1, 0, 0, 0}},
+  {{0xcae2b1920b57f4bc, 0x2936df5ec6c9bc36, 0x7dea6482e11238bf, 0x550663797b51f5d8},
+   {0x44ffe216348a964c, 0x9fb3d576dbdefbe1, 0x0afa40018d9d50e5, 0x157164848aecb851},
+   {1, 0, 0, 0}},
+  {{0xe48ecafffc5cde01, 0x7ccd84e70d715f26, 0xa2e8f483f43e4391, 0xeb5d7745b21141ea},
+   {0xcac917e2731a3479, 0x85f22cfe2844b645, 0x0990e6a158006cee, 0xeafd72ebdbecc17b},
+   {1, 0, 0, 0}},
+  {{0x6cf20ffb313728be, 0x96439591a3c6b94a, 0x2736ff8344315fc5, 0xa6d39677a7849276},
+   {0xf2bab833c357f5f4, 0x824a920c2284059b, 0x66b8babd2d27ecdf, 0x674f84749b0b8816},
+   {1, 0, 0, 0}},
+  {{0x2df48c04677c8a3e, 0x74e02f080203a56b, 0x31855f7db8c7fedb, 0x4e769e7672c9ddad},
+   {0xa4c36165b824bbb0, 0xfb9ae16f3b9122a5, 0x1ec0057206947281, 0x42b99082de830663},
+   {1, 0, 0, 0}},
+  {{0x6ef95150dda868b9, 0xd1f89e799c0ce131, 0x7fdc1ca008a1c478, 0x78878ef61c6ce04d},
+   {0x9c62b9121fe0d976, 0x6ace570ebde08d4f, 0xde53142c12309def, 0xb6cb3f5d7b72c321},
+   {1, 0, 0, 0}},
+  {{0x7f991ed2c31a3573, 0x5b82dd5bd54fb496, 0x595c5220812ffcae, 0x0c88bc4d716b1287},
+   {0x3a57bf635f48aca8, 0x7c8181f4df2564f3, 0x18d1b5b39c04e6aa, 0xdd5ddea3f3901dc6},
+   {1, 0, 0, 0}},
+  {{0xe96a79fb3e72ad0c, 0x43a0a28c42ba792f, 0xefe0a423083e49f3, 0x68f344af6b317466},
+   {0xcdfe17db3fb24d4a, 0x668bfc2271f5c626, 0x604ed93c24d67ff3, 0x31b9c405f8540a20},
+   {1, 0, 0, 0}},
+  {{0xd36b4789a2582e7f, 0x0d1a10144ec39c28, 0x663c62c3edbad7a0, 0x4052bf4b6f461db9},
+   {0x235a27c3188d25eb, 0xe724f33999bfcc5b, 0x862be6bd71d70cc8, 0xfecf4d5190b0fc61},
+   {1, 0, 0, 0}},
+  {{0x74346c10a1d4cfac, 0xafdf5cc08526a7a4, 0x123202a8f62bff7a, 0x1eddbae2c802e41a},
+   {0x8fa0af2dd603f844, 0x36e06b7e4c701917, 0x0c45f45273db33a0, 0x43104d86560ebcfc},
+   {1, 0, 0, 0}},
+  {{0x9615b5110d1d78e5, 0x66b0de3225c4744b, 0x0a4a46fb6aaf363a, 0xb48e26b484f7a21c},
+   {0x06ebb0f621a01b2d, 0xc004e4048b7b0f98, 0x64131bcdfed6f668, 0xfac015404d4d3dab},
+   {1, 0, 0, 0}}},
+ {{{0, 0, 0, 0},
+   {0, 0, 0, 0},
+   {0, 0, 0, 0}},
+  {{0x3a5a9e22185a5943, 0x1ab919365c65dfb6, 0x21656b32262c71da, 0x7fe36b40af22af89},
+   {0xd50d152c699ca101, 0x74b3d5867b8af212, 0x9f09f40407dca6f1, 0xe697d45825b63624},
+   {1, 0, 0, 0}},
+  {{0xa84aa9397512218e, 0xe9a521b074ca0141, 0x57880b3a18a2e902, 0x4a5b506612a677a6},
+   {0x0beada7a4c4f3840, 0x626db15419e26d9d, 0xc42604fbe1627d40, 0xeb13461ceac089f1},
+   {1, 0, 0, 0}},
+  {{0xf9faed0927a43281, 0x5e52c4144103ecbc, 0xc342967aa815c857, 0x0781b8291c6a220a},
+   {0x5a8343ceeac55f80, 0x88f80eeee54a05e3, 0x97b2a14f12916434, 0x690cde8df0151593},
+   {1, 0, 0, 0}},
+  {{0xaee9c75df7f82f2a, 0x9e4c35874afdf43a, 0xf5622df437371326, 0x8a535f566ec73617},
+   {0xc5f9a0ac223094b7, 0xcde533864c8c7669, 0x37e02819085a92bf, 0x0455c08468b08bd7},
+   {1, 0, 0, 0}},
+  {{0x0c0a6e2c9477b5d9, 0xf9a4bf62876dc444, 0x5050a949b6cdc279, 0x06bada7ab77f8276},
+   {0xc8b4aed1ea48dac9, 0xdebd8a4b7ea1070f, 0x427d49101366eb70, 0x5b476dfd0e6cb18a},
+   {1, 0, 0, 0}},
+  {{0x7c5c3e44278c340a, 0x4d54606812d66f3b, 0x29a751b1ae23c5d8, 0x3e29864e8a2ec908},
+   {0x142d2a6626dbb850, 0xad1744c4765bd780, 0x1f150e68e322d1ed, 0x239b90ea3dc31e7e},
+   {1, 0, 0, 0}},
+  {{0x78c416527a53322a, 0x305dde6709776f8e, 0xdbcab759f8862ed4, 0x820f4dd949f72ff7},
+   {0x6cc544a62b5debd4, 0x75be5d937b4e8cc4, 0x1b481b1b215c14d3, 0x140406ec783a05ec},
+   {1, 0, 0, 0}},
+  {{0x6a703f10e895df07, 0xfd75f3fa01876bd8, 0xeb5b06e70ce08ffe, 0x68f6b8542783dfee},
+   {0x90c76f8a78712655, 0xcf5293d2f310bf7f, 0xfbc8044dfda45028, 0xcbe1feba92e40ce6},
+   {1, 0, 0, 0}},
+  {{0xe998ceea4396e4c1, 0xfc82ef0b6acea274, 0x230f729f2250e927, 0xd0b2f94d2f420109},
+   {0x4305adddb38d4966, 0x10b838f8624c3b45, 0x7db2636658954e7a, 0x971459828b0719e5},
+   {1, 0, 0, 0}},
+  {{0x4bd6b72623369fc9, 0x57f2929e53d0b876, 0xc2d5cba4f2340687, 0x961610004a866aba},
+   {0x49997bcd2e407a5e, 0x69ab197d92ddcb24, 0x2cf1f2438fe5131c, 0x7acb9fadcee75e44},
+   {1, 0, 0, 0}},
+  {{0x254e839423d2d4c0, 0xf57f0c917aea685b, 0xa60d880f6f75aaea, 0x24eb9acca333bf5b},
+   {0xe3de4ccb1cda5dea, 0xfeef9341c51a6b4f, 0x743125f88bac4c4d, 0x69f891c5acd079cc},
+   {1, 0, 0, 0}},
+  {{0xeee44b35702476b5, 0x7ed031a0e45c2258, 0xb422d1e7bd6f8514, 0xe51f547c5972a107},
+   {0xa25bcd6fc9cf343d, 0x8ca922ee097c184e, 0xa62f98b3a9fe9a06, 0x1c309a2b25bb1387},
+   {1, 0, 0, 0}},
+  {{0x9295dbeb1967c459, 0xb00148833472c98e, 0xc504977708011828, 0x20b87b8aa2c4e503},
+   {0x3063175de057c277, 0x1bd539338fe582dd, 0x0d11adef5f69a044, 0xf5c6fa49919776be},
+   {1, 0, 0, 0}},
+  {{0x8c944e760fd59e11, 0x3876cba1102fad5f, 0xa454c3fad83faa56, 0x1ed7d1b9332010b9},
+   {0xa1011a270024b889, 0x05e4d0dcac0cd344, 0x52b520f0eb6a2a24, 0x3a2b03f03217257a},
+   {1, 0, 0, 0}},
+  {{0xf20fc2afdf1d043d, 0xf330240db58d5a62, 0xfc7d229ca0058c3b, 0x15fee545c78dd9f6},
+   {0x501e82885bc98cda, 0x41ef80e5d046ac04, 0x557d9f49461210fb, 0x4ab5b6b2b8753f81},
+   {1, 0, 0, 0}}}};
+
+/* select_point selects the |idx|th point from a precomputation table and
+ * copies it to out. */
+static void select_point(const u64 idx, unsigned int size, const smallfelem pre_comp[16][3], smallfelem out[3])
+	{
+	unsigned i, j;
+	u64 *outlimbs = &out[0][0];
+	memset(outlimbs, 0, 3 * sizeof(smallfelem));
+
+	for (i = 0; i < size; i++)
+		{
+		const u64 *inlimbs = (u64*) &pre_comp[i][0][0];
+		u64 mask = i ^ idx;
+		mask |= mask >> 4;
+		mask |= mask >> 2;
+		mask |= mask >> 1;
+		mask &= 1;
+		mask--;
+		for (j = 0; j < NLIMBS * 3; j++)
+			outlimbs[j] |= inlimbs[j] & mask;
+		}
+	}
+
+/* get_bit returns the |i|th bit in |in| */
+static char get_bit(const felem_bytearray in, int i)
+	{
+	if ((i < 0) || (i >= 256))
+		return 0;
+	return (in[i >> 3] >> (i & 7)) & 1;
+	}
+
+/* Interleaved point multiplication using precomputed point multiples:
+ * The small point multiples 0*P, 1*P, ..., 17*P are in pre_comp[],
+ * the scalars in scalars[]. If g_scalar is non-NULL, we also add this multiple
+ * of the generator, using certain (large) precomputed multiples in g_pre_comp.
+ * Output point (X, Y, Z) is stored in x_out, y_out, z_out */
+static void batch_mul(felem x_out, felem y_out, felem z_out,
+	const felem_bytearray scalars[], const unsigned num_points, const u8 *g_scalar,
+	const int mixed, const smallfelem pre_comp[][17][3], const smallfelem g_pre_comp[2][16][3])
+	{
+	int i, skip;
+	unsigned num, gen_mul = (g_scalar != NULL);
+	felem nq[3], ftmp;
+	smallfelem tmp[3];
+	u64 bits;
+	u8 sign, digit;
+
+	/* set nq to the point at infinity */
+	memset(nq, 0, 3 * sizeof(felem));
+
+	/* Loop over all scalars msb-to-lsb, interleaving additions
+	 * of multiples of the generator (two in each of the last 32 rounds)
+	 * and additions of other points multiples (every 5th round).
+	 */
+	skip = 1; /* save two point operations in the first round */
+	for (i = (num_points ? 255 : 31); i >= 0; --i)
+		{
+		/* double */
+		if (!skip)
+			point_double(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2]);
+
+		/* add multiples of the generator */
+		if (gen_mul && (i <= 31))
+			{
+			/* first, look 32 bits upwards */
+			bits = get_bit(g_scalar, i + 224) << 3;
+			bits |= get_bit(g_scalar, i + 160) << 2;
+			bits |= get_bit(g_scalar, i + 96) << 1;
+			bits |= get_bit(g_scalar, i + 32);
+			/* select the point to add, in constant time */
+			select_point(bits, 16, g_pre_comp[1], tmp);
+
+			if (!skip)
+				{
+				point_add(nq[0], nq[1], nq[2],
+					nq[0], nq[1], nq[2],
+					1 /* mixed */, tmp[0], tmp[1], tmp[2]);
+				}
+			else
+				{
+				smallfelem_expand(nq[0], tmp[0]);
+				smallfelem_expand(nq[1], tmp[1]);
+				smallfelem_expand(nq[2], tmp[2]);
+				skip = 0;
+				}
+
+			/* second, look at the current position */
+			bits = get_bit(g_scalar, i + 192) << 3;
+			bits |= get_bit(g_scalar, i + 128) << 2;
+			bits |= get_bit(g_scalar, i + 64) << 1;
+			bits |= get_bit(g_scalar, i);
+			/* select the point to add, in constant time */
+			select_point(bits, 16, g_pre_comp[0], tmp);
+			point_add(nq[0], nq[1], nq[2],
+				nq[0], nq[1], nq[2],
+				1 /* mixed */, tmp[0], tmp[1], tmp[2]);
+			}
+
+		/* do other additions every 5 doublings */
+		if (num_points && (i % 5 == 0))
+			{
+			/* loop over all scalars */
+			for (num = 0; num < num_points; ++num)
+				{
+				bits = get_bit(scalars[num], i + 4) << 5;
+				bits |= get_bit(scalars[num], i + 3) << 4;
+				bits |= get_bit(scalars[num], i + 2) << 3;
+				bits |= get_bit(scalars[num], i + 1) << 2;
+				bits |= get_bit(scalars[num], i) << 1;
+				bits |= get_bit(scalars[num], i - 1);
+				ec_GFp_nistp_recode_scalar_bits(&sign, &digit, bits);
+
+				/* select the point to add or subtract, in constant time */
+				select_point(digit, 17, pre_comp[num], tmp);
+				smallfelem_neg(ftmp, tmp[1]); /* (X, -Y, Z) is the negative point */
+				copy_small_conditional(ftmp, tmp[1], (((limb) sign) - 1));
+				felem_contract(tmp[1], ftmp);
+
+				if (!skip)
+					{
+					point_add(nq[0], nq[1], nq[2],
+						nq[0], nq[1], nq[2],
+						mixed, tmp[0], tmp[1], tmp[2]);
+					}
+				else
+					{
+					smallfelem_expand(nq[0], tmp[0]);
+					smallfelem_expand(nq[1], tmp[1]);
+					smallfelem_expand(nq[2], tmp[2]);
+					skip = 0;
+					}
+				}
+			}
+		}
+	felem_assign(x_out, nq[0]);
+	felem_assign(y_out, nq[1]);
+	felem_assign(z_out, nq[2]);
+	}
+
+/* Precomputation for the group generator. */
+typedef struct {
+	smallfelem g_pre_comp[2][16][3];
+	int references;
+} NISTP256_PRE_COMP;
+
+const EC_METHOD *EC_GFp_nistp256_method(void)
+	{
+	static const EC_METHOD ret = {
+		EC_FLAGS_DEFAULT_OCT,
+		NID_X9_62_prime_field,
+		ec_GFp_nistp256_group_init,
+		ec_GFp_simple_group_finish,
+		ec_GFp_simple_group_clear_finish,
+		ec_GFp_nist_group_copy,
+		ec_GFp_nistp256_group_set_curve,
+		ec_GFp_simple_group_get_curve,
+		ec_GFp_simple_group_get_degree,
+		ec_GFp_simple_group_check_discriminant,
+		ec_GFp_simple_point_init,
+		ec_GFp_simple_point_finish,
+		ec_GFp_simple_point_clear_finish,
+		ec_GFp_simple_point_copy,
+		ec_GFp_simple_point_set_to_infinity,
+		ec_GFp_simple_set_Jprojective_coordinates_GFp,
+		ec_GFp_simple_get_Jprojective_coordinates_GFp,
+		ec_GFp_simple_point_set_affine_coordinates,
+		ec_GFp_nistp256_point_get_affine_coordinates,
+		0 /* point_set_compressed_coordinates */,
+		0 /* point2oct */,
+		0 /* oct2point */,
+		ec_GFp_simple_add,
+		ec_GFp_simple_dbl,
+		ec_GFp_simple_invert,
+		ec_GFp_simple_is_at_infinity,
+		ec_GFp_simple_is_on_curve,
+		ec_GFp_simple_cmp,
+		ec_GFp_simple_make_affine,
+		ec_GFp_simple_points_make_affine,
+		ec_GFp_nistp256_points_mul,
+		ec_GFp_nistp256_precompute_mult,
+		ec_GFp_nistp256_have_precompute_mult,
+		ec_GFp_nist_field_mul,
+		ec_GFp_nist_field_sqr,
+		0 /* field_div */,
+		0 /* field_encode */,
+		0 /* field_decode */,
+		0 /* field_set_to_one */ };
+
+	return &ret;
+	}
+
+/******************************************************************************/
+/*		       FUNCTIONS TO MANAGE PRECOMPUTATION
+ */
+
+static NISTP256_PRE_COMP *nistp256_pre_comp_new()
+	{
+	NISTP256_PRE_COMP *ret = NULL;
+	ret = (NISTP256_PRE_COMP *) OPENSSL_malloc(sizeof *ret);
+	if (!ret)
+		{
+		ECerr(EC_F_NISTP256_PRE_COMP_NEW, ERR_R_MALLOC_FAILURE);
+		return ret;
+		}
+	memset(ret->g_pre_comp, 0, sizeof(ret->g_pre_comp));
+	ret->references = 1;
+	return ret;
+	}
+
+static void *nistp256_pre_comp_dup(void *src_)
+	{
+	NISTP256_PRE_COMP *src = src_;
+
+	/* no need to actually copy, these objects never change! */
+	CRYPTO_add(&src->references, 1, CRYPTO_LOCK_EC_PRE_COMP);
+
+	return src_;
+	}
+
+static void nistp256_pre_comp_free(void *pre_)
+	{
+	int i;
+	NISTP256_PRE_COMP *pre = pre_;
+
+	if (!pre)
+		return;
+
+	i = CRYPTO_add(&pre->references, -1, CRYPTO_LOCK_EC_PRE_COMP);
+	if (i > 0)
+		return;
+
+	OPENSSL_free(pre);
+	}
+
+static void nistp256_pre_comp_clear_free(void *pre_)
+	{
+	int i;
+	NISTP256_PRE_COMP *pre = pre_;
+
+	if (!pre)
+		return;
+
+	i = CRYPTO_add(&pre->references, -1, CRYPTO_LOCK_EC_PRE_COMP);
+	if (i > 0)
+		return;
+
+	OPENSSL_cleanse(pre, sizeof *pre);
+	OPENSSL_free(pre);
+	}
+
+/******************************************************************************/
+/*			   OPENSSL EC_METHOD FUNCTIONS
+ */
+
+int ec_GFp_nistp256_group_init(EC_GROUP *group)
+	{
+	int ret;
+	ret = ec_GFp_simple_group_init(group);
+	group->a_is_minus3 = 1;
+	return ret;
+	}
+
+int ec_GFp_nistp256_group_set_curve(EC_GROUP *group, const BIGNUM *p,
+	const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx)
+	{
+	int ret = 0;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *curve_p, *curve_a, *curve_b;
+
+	if (ctx == NULL)
+		if ((ctx = new_ctx = BN_CTX_new()) == NULL) return 0;
+	BN_CTX_start(ctx);
+	if (((curve_p = BN_CTX_get(ctx)) == NULL) ||
+		((curve_a = BN_CTX_get(ctx)) == NULL) ||
+		((curve_b = BN_CTX_get(ctx)) == NULL)) goto err;
+	BN_bin2bn(nistp256_curve_params[0], sizeof(felem_bytearray), curve_p);
+	BN_bin2bn(nistp256_curve_params[1], sizeof(felem_bytearray), curve_a);
+	BN_bin2bn(nistp256_curve_params[2], sizeof(felem_bytearray), curve_b);
+	if ((BN_cmp(curve_p, p)) || (BN_cmp(curve_a, a)) ||
+		(BN_cmp(curve_b, b)))
+		{
+		ECerr(EC_F_EC_GFP_NISTP256_GROUP_SET_CURVE,
+			EC_R_WRONG_CURVE_PARAMETERS);
+		goto err;
+		}
+	group->field_mod_func = BN_nist_mod_256;
+	ret = ec_GFp_simple_group_set_curve(group, p, a, b, ctx);
+err:
+	BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return ret;
+	}
+
+/* Takes the Jacobian coordinates (X, Y, Z) of a point and returns
+ * (X', Y') = (X/Z^2, Y/Z^3) */
+int ec_GFp_nistp256_point_get_affine_coordinates(const EC_GROUP *group,
+	const EC_POINT *point, BIGNUM *x, BIGNUM *y, BN_CTX *ctx)
+	{
+	felem z1, z2, x_in, y_in;
+	smallfelem x_out, y_out;
+	longfelem tmp;
+
+	if (EC_POINT_is_at_infinity(group, point))
+		{
+		ECerr(EC_F_EC_GFP_NISTP256_POINT_GET_AFFINE_COORDINATES,
+			EC_R_POINT_AT_INFINITY);
+		return 0;
+		}
+	if ((!BN_to_felem(x_in, &point->X)) || (!BN_to_felem(y_in, &point->Y)) ||
+		(!BN_to_felem(z1, &point->Z))) return 0;
+	felem_inv(z2, z1);
+	felem_square(tmp, z2); felem_reduce(z1, tmp);
+	felem_mul(tmp, x_in, z1); felem_reduce(x_in, tmp);
+	felem_contract(x_out, x_in);
+	if (x != NULL)
+		{
+		if (!smallfelem_to_BN(x, x_out)) {
+		ECerr(EC_F_EC_GFP_NISTP256_POINT_GET_AFFINE_COORDINATES,
+			ERR_R_BN_LIB);
+		return 0;
+		}
+		}
+	felem_mul(tmp, z1, z2); felem_reduce(z1, tmp);
+	felem_mul(tmp, y_in, z1); felem_reduce(y_in, tmp);
+	felem_contract(y_out, y_in);
+	if (y != NULL)
+		{
+		if (!smallfelem_to_BN(y, y_out))
+			{
+			ECerr(EC_F_EC_GFP_NISTP256_POINT_GET_AFFINE_COORDINATES,
+				ERR_R_BN_LIB);
+			return 0;
+			}
+		}
+	return 1;
+	}
+
+static void make_points_affine(size_t num, smallfelem points[/* num */][3], smallfelem tmp_smallfelems[/* num+1 */])
+	{
+	/* Runs in constant time, unless an input is the point at infinity
+	 * (which normally shouldn't happen). */
+	ec_GFp_nistp_points_make_affine_internal(
+		num,
+		points,
+		sizeof(smallfelem),
+		tmp_smallfelems,
+		(void (*)(void *)) smallfelem_one,
+		(int (*)(const void *)) smallfelem_is_zero_int,
+		(void (*)(void *, const void *)) smallfelem_assign,
+		(void (*)(void *, const void *)) smallfelem_square_contract,
+		(void (*)(void *, const void *, const void *)) smallfelem_mul_contract,
+		(void (*)(void *, const void *)) smallfelem_inv_contract,
+		(void (*)(void *, const void *)) smallfelem_assign /* nothing to contract */);
+	}
+
+/* Computes scalar*generator + \sum scalars[i]*points[i], ignoring NULL values
+ * Result is stored in r (r can equal one of the inputs). */
+int ec_GFp_nistp256_points_mul(const EC_GROUP *group, EC_POINT *r,
+	const BIGNUM *scalar, size_t num, const EC_POINT *points[],
+	const BIGNUM *scalars[], BN_CTX *ctx)
+	{
+	int ret = 0;
+	int j;
+	int mixed = 0;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *x, *y, *z, *tmp_scalar;
+	felem_bytearray g_secret;
+	felem_bytearray *secrets = NULL;
+	smallfelem (*pre_comp)[17][3] = NULL;
+	smallfelem *tmp_smallfelems = NULL;
+	felem_bytearray tmp;
+	unsigned i, num_bytes;
+	int have_pre_comp = 0;
+	size_t num_points = num;
+	smallfelem x_in, y_in, z_in;
+	felem x_out, y_out, z_out;
+	NISTP256_PRE_COMP *pre = NULL;
+	const smallfelem (*g_pre_comp)[16][3] = NULL;
+	EC_POINT *generator = NULL;
+	const EC_POINT *p = NULL;
+	const BIGNUM *p_scalar = NULL;
+
+	if (ctx == NULL)
+		if ((ctx = new_ctx = BN_CTX_new()) == NULL) return 0;
+	BN_CTX_start(ctx);
+	if (((x = BN_CTX_get(ctx)) == NULL) ||
+		((y = BN_CTX_get(ctx)) == NULL) ||
+		((z = BN_CTX_get(ctx)) == NULL) ||
+		((tmp_scalar = BN_CTX_get(ctx)) == NULL))
+		goto err;
+
+	if (scalar != NULL)
+		{
+		pre = EC_EX_DATA_get_data(group->extra_data,
+			nistp256_pre_comp_dup, nistp256_pre_comp_free,
+			nistp256_pre_comp_clear_free);
+		if (pre)
+			/* we have precomputation, try to use it */
+			g_pre_comp = (const smallfelem (*)[16][3]) pre->g_pre_comp;
+		else
+			/* try to use the standard precomputation */
+			g_pre_comp = &gmul[0];
+		generator = EC_POINT_new(group);
+		if (generator == NULL)
+			goto err;
+		/* get the generator from precomputation */
+		if (!smallfelem_to_BN(x, g_pre_comp[0][1][0]) ||
+			!smallfelem_to_BN(y, g_pre_comp[0][1][1]) ||
+			!smallfelem_to_BN(z, g_pre_comp[0][1][2]))
+			{
+			ECerr(EC_F_EC_GFP_NISTP256_POINTS_MUL, ERR_R_BN_LIB);
+			goto err;
+			}
+		if (!EC_POINT_set_Jprojective_coordinates_GFp(group,
+				generator, x, y, z, ctx))
+			goto err;
+		if (0 == EC_POINT_cmp(group, generator, group->generator, ctx))
+			/* precomputation matches generator */
+			have_pre_comp = 1;
+		else
+			/* we don't have valid precomputation:
+			 * treat the generator as a random point */
+			num_points++;
+		}
+	if (num_points > 0)
+		{
+		if (num_points >= 3)
+			{
+			/* unless we precompute multiples for just one or two points,
+			 * converting those into affine form is time well spent  */
+			mixed = 1;
+			}
+		secrets = OPENSSL_malloc(num_points * sizeof(felem_bytearray));
+		pre_comp = OPENSSL_malloc(num_points * 17 * 3 * sizeof(smallfelem));
+		if (mixed)
+			tmp_smallfelems = OPENSSL_malloc((num_points * 17 + 1) * sizeof(smallfelem));
+		if ((secrets == NULL) || (pre_comp == NULL) || (mixed && (tmp_smallfelems == NULL)))
+			{
+			ECerr(EC_F_EC_GFP_NISTP256_POINTS_MUL, ERR_R_MALLOC_FAILURE);
+			goto err;
+			}
+
+		/* we treat NULL scalars as 0, and NULL points as points at infinity,
+		 * i.e., they contribute nothing to the linear combination */
+		memset(secrets, 0, num_points * sizeof(felem_bytearray));
+		memset(pre_comp, 0, num_points * 17 * 3 * sizeof(smallfelem));
+		for (i = 0; i < num_points; ++i)
+			{
+			if (i == num)
+				/* we didn't have a valid precomputation, so we pick
+				 * the generator */
+				{
+				p = EC_GROUP_get0_generator(group);
+				p_scalar = scalar;
+				}
+			else
+				/* the i^th point */
+				{
+				p = points[i];
+				p_scalar = scalars[i];
+				}
+			if ((p_scalar != NULL) && (p != NULL))
+				{
+				/* reduce scalar to 0 <= scalar < 2^256 */
+				if ((BN_num_bits(p_scalar) > 256) || (BN_is_negative(p_scalar)))
+					{
+					/* this is an unusual input, and we don't guarantee
+					 * constant-timeness */
+					if (!BN_nnmod(tmp_scalar, p_scalar, &group->order, ctx))
+						{
+						ECerr(EC_F_EC_GFP_NISTP256_POINTS_MUL, ERR_R_BN_LIB);
+						goto err;
+						}
+					num_bytes = BN_bn2bin(tmp_scalar, tmp);
+					}
+				else
+					num_bytes = BN_bn2bin(p_scalar, tmp);
+				flip_endian(secrets[i], tmp, num_bytes);
+				/* precompute multiples */
+				if ((!BN_to_felem(x_out, &p->X)) ||
+					(!BN_to_felem(y_out, &p->Y)) ||
+					(!BN_to_felem(z_out, &p->Z))) goto err;
+				felem_shrink(pre_comp[i][1][0], x_out);
+				felem_shrink(pre_comp[i][1][1], y_out);
+				felem_shrink(pre_comp[i][1][2], z_out);
+				for (j = 2; j <= 16; ++j)
+					{
+					if (j & 1)
+						{
+						point_add_small(
+							pre_comp[i][j][0], pre_comp[i][j][1], pre_comp[i][j][2],
+							pre_comp[i][1][0], pre_comp[i][1][1], pre_comp[i][1][2],
+							pre_comp[i][j-1][0], pre_comp[i][j-1][1], pre_comp[i][j-1][2]);
+						}
+					else
+						{
+						point_double_small(
+							pre_comp[i][j][0], pre_comp[i][j][1], pre_comp[i][j][2],
+							pre_comp[i][j/2][0], pre_comp[i][j/2][1], pre_comp[i][j/2][2]);
+						}
+					}
+				}
+			}
+		if (mixed)
+			make_points_affine(num_points * 17, pre_comp[0], tmp_smallfelems);
+		}
+
+	/* the scalar for the generator */
+	if ((scalar != NULL) && (have_pre_comp))
+		{
+		memset(g_secret, 0, sizeof(g_secret));
+		/* reduce scalar to 0 <= scalar < 2^256 */
+		if ((BN_num_bits(scalar) > 256) || (BN_is_negative(scalar)))
+			{
+			/* this is an unusual input, and we don't guarantee
+			 * constant-timeness */
+			if (!BN_nnmod(tmp_scalar, scalar, &group->order, ctx))
+				{
+				ECerr(EC_F_EC_GFP_NISTP256_POINTS_MUL, ERR_R_BN_LIB);
+				goto err;
+				}
+			num_bytes = BN_bn2bin(tmp_scalar, tmp);
+			}
+		else
+			num_bytes = BN_bn2bin(scalar, tmp);
+		flip_endian(g_secret, tmp, num_bytes);
+		/* do the multiplication with generator precomputation*/
+		batch_mul(x_out, y_out, z_out,
+			(const felem_bytearray (*)) secrets, num_points,
+			g_secret,
+			mixed, (const smallfelem (*)[17][3]) pre_comp,
+			g_pre_comp);
+		}
+	else
+		/* do the multiplication without generator precomputation */
+		batch_mul(x_out, y_out, z_out,
+			(const felem_bytearray (*)) secrets, num_points,
+			NULL, mixed, (const smallfelem (*)[17][3]) pre_comp, NULL);
+	/* reduce the output to its unique minimal representation */
+	felem_contract(x_in, x_out);
+	felem_contract(y_in, y_out);
+	felem_contract(z_in, z_out);
+	if ((!smallfelem_to_BN(x, x_in)) || (!smallfelem_to_BN(y, y_in)) ||
+		(!smallfelem_to_BN(z, z_in)))
+		{
+		ECerr(EC_F_EC_GFP_NISTP256_POINTS_MUL, ERR_R_BN_LIB);
+		goto err;
+		}
+	ret = EC_POINT_set_Jprojective_coordinates_GFp(group, r, x, y, z, ctx);
+
+err:
+	BN_CTX_end(ctx);
+	if (generator != NULL)
+		EC_POINT_free(generator);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	if (secrets != NULL)
+		OPENSSL_free(secrets);
+	if (pre_comp != NULL)
+		OPENSSL_free(pre_comp);
+	if (tmp_smallfelems != NULL)
+		OPENSSL_free(tmp_smallfelems);
+	return ret;
+	}
+
+int ec_GFp_nistp256_precompute_mult(EC_GROUP *group, BN_CTX *ctx)
+	{
+	int ret = 0;
+	NISTP256_PRE_COMP *pre = NULL;
+	int i, j;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *x, *y;
+	EC_POINT *generator = NULL;
+	smallfelem tmp_smallfelems[32];
+	felem x_tmp, y_tmp, z_tmp;
+
+	/* throw away old precomputation */
+	EC_EX_DATA_free_data(&group->extra_data, nistp256_pre_comp_dup,
+		nistp256_pre_comp_free, nistp256_pre_comp_clear_free);
+	if (ctx == NULL)
+		if ((ctx = new_ctx = BN_CTX_new()) == NULL) return 0;
+	BN_CTX_start(ctx);
+	if (((x = BN_CTX_get(ctx)) == NULL) ||
+		((y = BN_CTX_get(ctx)) == NULL))
+		goto err;
+	/* get the generator */
+	if (group->generator == NULL) goto err;
+	generator = EC_POINT_new(group);
+	if (generator == NULL)
+		goto err;
+	BN_bin2bn(nistp256_curve_params[3], sizeof (felem_bytearray), x);
+	BN_bin2bn(nistp256_curve_params[4], sizeof (felem_bytearray), y);
+	if (!EC_POINT_set_affine_coordinates_GFp(group, generator, x, y, ctx))
+		goto err;
+	if ((pre = nistp256_pre_comp_new()) == NULL)
+		goto err;
+	/* if the generator is the standard one, use built-in precomputation */
+	if (0 == EC_POINT_cmp(group, generator, group->generator, ctx))
+		{
+		memcpy(pre->g_pre_comp, gmul, sizeof(pre->g_pre_comp));
+		ret = 1;
+		goto err;
+		}
+	if ((!BN_to_felem(x_tmp, &group->generator->X)) ||
+		(!BN_to_felem(y_tmp, &group->generator->Y)) ||
+		(!BN_to_felem(z_tmp, &group->generator->Z)))
+		goto err;
+	felem_shrink(pre->g_pre_comp[0][1][0], x_tmp);
+	felem_shrink(pre->g_pre_comp[0][1][1], y_tmp);
+	felem_shrink(pre->g_pre_comp[0][1][2], z_tmp);
+	/* compute 2^64*G, 2^128*G, 2^192*G for the first table,
+	 * 2^32*G, 2^96*G, 2^160*G, 2^224*G for the second one
+	 */
+	for (i = 1; i <= 8; i <<= 1)
+		{
+		point_double_small(
+			pre->g_pre_comp[1][i][0], pre->g_pre_comp[1][i][1], pre->g_pre_comp[1][i][2],
+			pre->g_pre_comp[0][i][0], pre->g_pre_comp[0][i][1], pre->g_pre_comp[0][i][2]);
+		for (j = 0; j < 31; ++j)
+			{
+			point_double_small(
+				pre->g_pre_comp[1][i][0], pre->g_pre_comp[1][i][1], pre->g_pre_comp[1][i][2],
+				pre->g_pre_comp[1][i][0], pre->g_pre_comp[1][i][1], pre->g_pre_comp[1][i][2]);
+			}
+		if (i == 8)
+			break;
+		point_double_small(
+			pre->g_pre_comp[0][2*i][0], pre->g_pre_comp[0][2*i][1], pre->g_pre_comp[0][2*i][2],
+			pre->g_pre_comp[1][i][0], pre->g_pre_comp[1][i][1], pre->g_pre_comp[1][i][2]);
+		for (j = 0; j < 31; ++j)
+			{
+			point_double_small(
+				pre->g_pre_comp[0][2*i][0], pre->g_pre_comp[0][2*i][1], pre->g_pre_comp[0][2*i][2],
+				pre->g_pre_comp[0][2*i][0], pre->g_pre_comp[0][2*i][1], pre->g_pre_comp[0][2*i][2]);
+			}
+		}
+	for (i = 0; i < 2; i++)
+		{
+		/* g_pre_comp[i][0] is the point at infinity */
+		memset(pre->g_pre_comp[i][0], 0, sizeof(pre->g_pre_comp[i][0]));
+		/* the remaining multiples */
+		/* 2^64*G + 2^128*G resp. 2^96*G + 2^160*G */
+		point_add_small(
+			pre->g_pre_comp[i][6][0], pre->g_pre_comp[i][6][1], pre->g_pre_comp[i][6][2],
+			pre->g_pre_comp[i][4][0], pre->g_pre_comp[i][4][1], pre->g_pre_comp[i][4][2],
+			pre->g_pre_comp[i][2][0], pre->g_pre_comp[i][2][1], pre->g_pre_comp[i][2][2]);
+		/* 2^64*G + 2^192*G resp. 2^96*G + 2^224*G */
+		point_add_small(
+			pre->g_pre_comp[i][10][0], pre->g_pre_comp[i][10][1], pre->g_pre_comp[i][10][2],
+			pre->g_pre_comp[i][8][0], pre->g_pre_comp[i][8][1], pre->g_pre_comp[i][8][2],
+			pre->g_pre_comp[i][2][0], pre->g_pre_comp[i][2][1], pre->g_pre_comp[i][2][2]);
+		/* 2^128*G + 2^192*G resp. 2^160*G + 2^224*G */
+		point_add_small(
+			pre->g_pre_comp[i][12][0], pre->g_pre_comp[i][12][1], pre->g_pre_comp[i][12][2],
+			pre->g_pre_comp[i][8][0], pre->g_pre_comp[i][8][1], pre->g_pre_comp[i][8][2],
+			pre->g_pre_comp[i][4][0], pre->g_pre_comp[i][4][1], pre->g_pre_comp[i][4][2]);
+		/* 2^64*G + 2^128*G + 2^192*G resp. 2^96*G + 2^160*G + 2^224*G */
+		point_add_small(
+			pre->g_pre_comp[i][14][0], pre->g_pre_comp[i][14][1], pre->g_pre_comp[i][14][2],
+			pre->g_pre_comp[i][12][0], pre->g_pre_comp[i][12][1], pre->g_pre_comp[i][12][2],
+			pre->g_pre_comp[i][2][0], pre->g_pre_comp[i][2][1], pre->g_pre_comp[i][2][2]);
+		for (j = 1; j < 8; ++j)
+			{
+			/* odd multiples: add G resp. 2^32*G */
+			point_add_small(
+				pre->g_pre_comp[i][2*j+1][0], pre->g_pre_comp[i][2*j+1][1], pre->g_pre_comp[i][2*j+1][2],
+				pre->g_pre_comp[i][2*j][0], pre->g_pre_comp[i][2*j][1], pre->g_pre_comp[i][2*j][2],
+				pre->g_pre_comp[i][1][0], pre->g_pre_comp[i][1][1], pre->g_pre_comp[i][1][2]);
+			}
+		}
+	make_points_affine(31, &(pre->g_pre_comp[0][1]), tmp_smallfelems);
+
+	if (!EC_EX_DATA_set_data(&group->extra_data, pre, nistp256_pre_comp_dup,
+			nistp256_pre_comp_free, nistp256_pre_comp_clear_free))
+		goto err;
+	ret = 1;
+	pre = NULL;
+ err:
+	BN_CTX_end(ctx);
+	if (generator != NULL)
+		EC_POINT_free(generator);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	if (pre)
+		nistp256_pre_comp_free(pre);
+	return ret;
+	}
+
+int ec_GFp_nistp256_have_precompute_mult(const EC_GROUP *group)
+	{
+	if (EC_EX_DATA_get_data(group->extra_data, nistp256_pre_comp_dup,
+			nistp256_pre_comp_free, nistp256_pre_comp_clear_free)
+		!= NULL)
+		return 1;
+	else
+		return 0;
+	}
+#else
+static void *dummy=&dummy;
+#endif
diff --git a/crypto/ec/ecp_nistp521.c b/crypto/ec/ecp_nistp521.c
new file mode 100644
index 000000000000..178b655f7f17
--- /dev/null
+++ b/crypto/ec/ecp_nistp521.c
@@ -0,0 +1,2025 @@
+/* crypto/ec/ecp_nistp521.c */
+/*
+ * Written by Adam Langley (Google) for the OpenSSL project
+ */
+/* Copyright 2011 Google Inc.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ *
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ *  Unless required by applicable law or agreed to in writing, software
+ *  distributed under the License is distributed on an "AS IS" BASIS,
+ *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ *  See the License for the specific language governing permissions and
+ *  limitations under the License.
+ */
+
+/*
+ * A 64-bit implementation of the NIST P-521 elliptic curve point multiplication
+ *
+ * OpenSSL integration was taken from Emilia Kasper's work in ecp_nistp224.c.
+ * Otherwise based on Emilia's P224 work, which was inspired by my curve25519
+ * work which got its smarts from Daniel J. Bernstein's work on the same.
+ */
+
+#include <openssl/opensslconf.h>
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+
+#ifndef OPENSSL_SYS_VMS
+#include <stdint.h>
+#else
+#include <inttypes.h>
+#endif
+
+#include <string.h>
+#include <openssl/err.h>
+#include "ec_lcl.h"
+
+#if defined(__GNUC__) && (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1))
+  /* even with gcc, the typedef won't work for 32-bit platforms */
+  typedef __uint128_t uint128_t; /* nonstandard; implemented by gcc on 64-bit platforms */
+#else
+  #error "Need GCC 3.1 or later to define type uint128_t"
+#endif
+
+typedef uint8_t u8;
+typedef uint64_t u64;
+typedef int64_t s64;
+
+/* The underlying field.
+ *
+ * P521 operates over GF(2^521-1). We can serialise an element of this field
+ * into 66 bytes where the most significant byte contains only a single bit. We
+ * call this an felem_bytearray. */
+
+typedef u8 felem_bytearray[66];
+
+/* These are the parameters of P521, taken from FIPS 186-3, section D.1.2.5.
+ * These values are big-endian. */
+static const felem_bytearray nistp521_curve_params[5] =
+	{
+	{0x01, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,  /* p */
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff},
+	{0x01, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,  /* a = -3 */
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	 0xff, 0xfc},
+	{0x00, 0x51, 0x95, 0x3e, 0xb9, 0x61, 0x8e, 0x1c,  /* b */
+	 0x9a, 0x1f, 0x92, 0x9a, 0x21, 0xa0, 0xb6, 0x85,
+	 0x40, 0xee, 0xa2, 0xda, 0x72, 0x5b, 0x99, 0xb3,
+	 0x15, 0xf3, 0xb8, 0xb4, 0x89, 0x91, 0x8e, 0xf1,
+	 0x09, 0xe1, 0x56, 0x19, 0x39, 0x51, 0xec, 0x7e,
+	 0x93, 0x7b, 0x16, 0x52, 0xc0, 0xbd, 0x3b, 0xb1,
+	 0xbf, 0x07, 0x35, 0x73, 0xdf, 0x88, 0x3d, 0x2c,
+	 0x34, 0xf1, 0xef, 0x45, 0x1f, 0xd4, 0x6b, 0x50,
+	 0x3f, 0x00},
+	{0x00, 0xc6, 0x85, 0x8e, 0x06, 0xb7, 0x04, 0x04,  /* x */
+	 0xe9, 0xcd, 0x9e, 0x3e, 0xcb, 0x66, 0x23, 0x95,
+	 0xb4, 0x42, 0x9c, 0x64, 0x81, 0x39, 0x05, 0x3f,
+	 0xb5, 0x21, 0xf8, 0x28, 0xaf, 0x60, 0x6b, 0x4d,
+	 0x3d, 0xba, 0xa1, 0x4b, 0x5e, 0x77, 0xef, 0xe7,
+	 0x59, 0x28, 0xfe, 0x1d, 0xc1, 0x27, 0xa2, 0xff,
+	 0xa8, 0xde, 0x33, 0x48, 0xb3, 0xc1, 0x85, 0x6a,
+	 0x42, 0x9b, 0xf9, 0x7e, 0x7e, 0x31, 0xc2, 0xe5,
+	 0xbd, 0x66},
+	{0x01, 0x18, 0x39, 0x29, 0x6a, 0x78, 0x9a, 0x3b,  /* y */
+	 0xc0, 0x04, 0x5c, 0x8a, 0x5f, 0xb4, 0x2c, 0x7d,
+	 0x1b, 0xd9, 0x98, 0xf5, 0x44, 0x49, 0x57, 0x9b,
+	 0x44, 0x68, 0x17, 0xaf, 0xbd, 0x17, 0x27, 0x3e,
+	 0x66, 0x2c, 0x97, 0xee, 0x72, 0x99, 0x5e, 0xf4,
+	 0x26, 0x40, 0xc5, 0x50, 0xb9, 0x01, 0x3f, 0xad,
+	 0x07, 0x61, 0x35, 0x3c, 0x70, 0x86, 0xa2, 0x72,
+	 0xc2, 0x40, 0x88, 0xbe, 0x94, 0x76, 0x9f, 0xd1,
+	 0x66, 0x50}
+	};
+
+/* The representation of field elements.
+ * ------------------------------------
+ *
+ * We represent field elements with nine values. These values are either 64 or
+ * 128 bits and the field element represented is:
+ *   v[0]*2^0 + v[1]*2^58 + v[2]*2^116 + ... + v[8]*2^464  (mod p)
+ * Each of the nine values is called a 'limb'. Since the limbs are spaced only
+ * 58 bits apart, but are greater than 58 bits in length, the most significant
+ * bits of each limb overlap with the least significant bits of the next.
+ *
+ * A field element with 64-bit limbs is an 'felem'. One with 128-bit limbs is a
+ * 'largefelem' */
+
+#define NLIMBS 9
+
+typedef uint64_t limb;
+typedef limb felem[NLIMBS];
+typedef uint128_t largefelem[NLIMBS];
+
+static const limb bottom57bits = 0x1ffffffffffffff;
+static const limb bottom58bits = 0x3ffffffffffffff;
+
+/* bin66_to_felem takes a little-endian byte array and converts it into felem
+ * form. This assumes that the CPU is little-endian. */
+static void bin66_to_felem(felem out, const u8 in[66])
+	{
+	out[0] = (*((limb*) &in[0])) & bottom58bits;
+	out[1] = (*((limb*) &in[7]) >> 2) & bottom58bits;
+	out[2] = (*((limb*) &in[14]) >> 4) & bottom58bits;
+	out[3] = (*((limb*) &in[21]) >> 6) & bottom58bits;
+	out[4] = (*((limb*) &in[29])) & bottom58bits;
+	out[5] = (*((limb*) &in[36]) >> 2) & bottom58bits;
+	out[6] = (*((limb*) &in[43]) >> 4) & bottom58bits;
+	out[7] = (*((limb*) &in[50]) >> 6) & bottom58bits;
+	out[8] = (*((limb*) &in[58])) & bottom57bits;
+	}
+
+/* felem_to_bin66 takes an felem and serialises into a little endian, 66 byte
+ * array. This assumes that the CPU is little-endian. */
+static void felem_to_bin66(u8 out[66], const felem in)
+	{
+	memset(out, 0, 66);
+	(*((limb*) &out[0])) = in[0];
+	(*((limb*) &out[7])) |= in[1] << 2;
+	(*((limb*) &out[14])) |= in[2] << 4;
+	(*((limb*) &out[21])) |= in[3] << 6;
+	(*((limb*) &out[29])) = in[4];
+	(*((limb*) &out[36])) |= in[5] << 2;
+	(*((limb*) &out[43])) |= in[6] << 4;
+	(*((limb*) &out[50])) |= in[7] << 6;
+	(*((limb*) &out[58])) = in[8];
+	}
+
+/* To preserve endianness when using BN_bn2bin and BN_bin2bn */
+static void flip_endian(u8 *out, const u8 *in, unsigned len)
+	{
+	unsigned i;
+	for (i = 0; i < len; ++i)
+		out[i] = in[len-1-i];
+	}
+
+/* BN_to_felem converts an OpenSSL BIGNUM into an felem */
+static int BN_to_felem(felem out, const BIGNUM *bn)
+	{
+	felem_bytearray b_in;
+	felem_bytearray b_out;
+	unsigned num_bytes;
+
+	/* BN_bn2bin eats leading zeroes */
+	memset(b_out, 0, sizeof b_out);
+	num_bytes = BN_num_bytes(bn);
+	if (num_bytes > sizeof b_out)
+		{
+		ECerr(EC_F_BN_TO_FELEM, EC_R_BIGNUM_OUT_OF_RANGE);
+		return 0;
+		}
+	if (BN_is_negative(bn))
+		{
+		ECerr(EC_F_BN_TO_FELEM, EC_R_BIGNUM_OUT_OF_RANGE);
+		return 0;
+		}
+	num_bytes = BN_bn2bin(bn, b_in);
+	flip_endian(b_out, b_in, num_bytes);
+	bin66_to_felem(out, b_out);
+	return 1;
+	}
+
+/* felem_to_BN converts an felem into an OpenSSL BIGNUM */
+static BIGNUM *felem_to_BN(BIGNUM *out, const felem in)
+	{
+	felem_bytearray b_in, b_out;
+	felem_to_bin66(b_in, in);
+	flip_endian(b_out, b_in, sizeof b_out);
+	return BN_bin2bn(b_out, sizeof b_out, out);
+	}
+
+
+/* Field operations
+ * ---------------- */
+
+static void felem_one(felem out)
+	{
+	out[0] = 1;
+	out[1] = 0;
+	out[2] = 0;
+	out[3] = 0;
+	out[4] = 0;
+	out[5] = 0;
+	out[6] = 0;
+	out[7] = 0;
+	out[8] = 0;
+	}
+
+static void felem_assign(felem out, const felem in)
+	{
+	out[0] = in[0];
+	out[1] = in[1];
+	out[2] = in[2];
+	out[3] = in[3];
+	out[4] = in[4];
+	out[5] = in[5];
+	out[6] = in[6];
+	out[7] = in[7];
+	out[8] = in[8];
+	}
+
+/* felem_sum64 sets out = out + in. */
+static void felem_sum64(felem out, const felem in)
+	{
+	out[0] += in[0];
+	out[1] += in[1];
+	out[2] += in[2];
+	out[3] += in[3];
+	out[4] += in[4];
+	out[5] += in[5];
+	out[6] += in[6];
+	out[7] += in[7];
+	out[8] += in[8];
+	}
+
+/* felem_scalar sets out = in * scalar */
+static void felem_scalar(felem out, const felem in, limb scalar)
+	{
+	out[0] = in[0] * scalar;
+	out[1] = in[1] * scalar;
+	out[2] = in[2] * scalar;
+	out[3] = in[3] * scalar;
+	out[4] = in[4] * scalar;
+	out[5] = in[5] * scalar;
+	out[6] = in[6] * scalar;
+	out[7] = in[7] * scalar;
+	out[8] = in[8] * scalar;
+	}
+
+/* felem_scalar64 sets out = out * scalar */
+static void felem_scalar64(felem out, limb scalar)
+	{
+	out[0] *= scalar;
+	out[1] *= scalar;
+	out[2] *= scalar;
+	out[3] *= scalar;
+	out[4] *= scalar;
+	out[5] *= scalar;
+	out[6] *= scalar;
+	out[7] *= scalar;
+	out[8] *= scalar;
+	}
+
+/* felem_scalar128 sets out = out * scalar */
+static void felem_scalar128(largefelem out, limb scalar)
+	{
+	out[0] *= scalar;
+	out[1] *= scalar;
+	out[2] *= scalar;
+	out[3] *= scalar;
+	out[4] *= scalar;
+	out[5] *= scalar;
+	out[6] *= scalar;
+	out[7] *= scalar;
+	out[8] *= scalar;
+	}
+
+/* felem_neg sets |out| to |-in|
+ * On entry:
+ *   in[i] < 2^59 + 2^14
+ * On exit:
+ *   out[i] < 2^62
+ */
+static void felem_neg(felem out, const felem in)
+	{
+	/* In order to prevent underflow, we subtract from 0 mod p. */
+	static const limb two62m3 = (((limb)1) << 62) - (((limb)1) << 5);
+	static const limb two62m2 = (((limb)1) << 62) - (((limb)1) << 4);
+
+	out[0] = two62m3 - in[0];
+	out[1] = two62m2 - in[1];
+	out[2] = two62m2 - in[2];
+	out[3] = two62m2 - in[3];
+	out[4] = two62m2 - in[4];
+	out[5] = two62m2 - in[5];
+	out[6] = two62m2 - in[6];
+	out[7] = two62m2 - in[7];
+	out[8] = two62m2 - in[8];
+	}
+
+/* felem_diff64 subtracts |in| from |out|
+ * On entry:
+ *   in[i] < 2^59 + 2^14
+ * On exit:
+ *   out[i] < out[i] + 2^62
+ */
+static void felem_diff64(felem out, const felem in)
+	{
+	/* In order to prevent underflow, we add 0 mod p before subtracting. */
+	static const limb two62m3 = (((limb)1) << 62) - (((limb)1) << 5);
+	static const limb two62m2 = (((limb)1) << 62) - (((limb)1) << 4);
+
+	out[0] += two62m3 - in[0];
+	out[1] += two62m2 - in[1];
+	out[2] += two62m2 - in[2];
+	out[3] += two62m2 - in[3];
+	out[4] += two62m2 - in[4];
+	out[5] += two62m2 - in[5];
+	out[6] += two62m2 - in[6];
+	out[7] += two62m2 - in[7];
+	out[8] += two62m2 - in[8];
+	}
+
+/* felem_diff_128_64 subtracts |in| from |out|
+ * On entry:
+ *   in[i] < 2^62 + 2^17
+ * On exit:
+ *   out[i] < out[i] + 2^63
+ */
+static void felem_diff_128_64(largefelem out, const felem in)
+	{
+	/* In order to prevent underflow, we add 0 mod p before subtracting. */
+	static const limb two63m6 = (((limb)1) << 62) - (((limb)1) << 5);
+	static const limb two63m5 = (((limb)1) << 62) - (((limb)1) << 4);
+
+	out[0] += two63m6 - in[0];
+	out[1] += two63m5 - in[1];
+	out[2] += two63m5 - in[2];
+	out[3] += two63m5 - in[3];
+	out[4] += two63m5 - in[4];
+	out[5] += two63m5 - in[5];
+	out[6] += two63m5 - in[6];
+	out[7] += two63m5 - in[7];
+	out[8] += two63m5 - in[8];
+	}
+
+/* felem_diff_128_64 subtracts |in| from |out|
+ * On entry:
+ *   in[i] < 2^126
+ * On exit:
+ *   out[i] < out[i] + 2^127 - 2^69
+ */
+static void felem_diff128(largefelem out, const largefelem in)
+	{
+	/* In order to prevent underflow, we add 0 mod p before subtracting. */
+	static const uint128_t two127m70 = (((uint128_t)1) << 127) - (((uint128_t)1) << 70);
+	static const uint128_t two127m69 = (((uint128_t)1) << 127) - (((uint128_t)1) << 69);
+
+	out[0] += (two127m70 - in[0]);
+	out[1] += (two127m69 - in[1]);
+	out[2] += (two127m69 - in[2]);
+	out[3] += (two127m69 - in[3]);
+	out[4] += (two127m69 - in[4]);
+	out[5] += (two127m69 - in[5]);
+	out[6] += (two127m69 - in[6]);
+	out[7] += (two127m69 - in[7]);
+	out[8] += (two127m69 - in[8]);
+	}
+
+/* felem_square sets |out| = |in|^2
+ * On entry:
+ *   in[i] < 2^62
+ * On exit:
+ *   out[i] < 17 * max(in[i]) * max(in[i])
+ */
+static void felem_square(largefelem out, const felem in)
+	{
+	felem inx2, inx4;
+	felem_scalar(inx2, in, 2);
+	felem_scalar(inx4, in, 4);
+
+	/* We have many cases were we want to do
+	 *   in[x] * in[y] +
+	 *   in[y] * in[x]
+	 * This is obviously just
+	 *   2 * in[x] * in[y]
+	 * However, rather than do the doubling on the 128 bit result, we
+	 * double one of the inputs to the multiplication by reading from
+	 * |inx2| */
+
+	out[0] = ((uint128_t) in[0]) * in[0];
+	out[1] = ((uint128_t) in[0]) * inx2[1];
+	out[2] = ((uint128_t) in[0]) * inx2[2] +
+		 ((uint128_t) in[1]) * in[1];
+	out[3] = ((uint128_t) in[0]) * inx2[3] +
+		 ((uint128_t) in[1]) * inx2[2];
+	out[4] = ((uint128_t) in[0]) * inx2[4] +
+		 ((uint128_t) in[1]) * inx2[3] +
+		 ((uint128_t) in[2]) * in[2];
+	out[5] = ((uint128_t) in[0]) * inx2[5] +
+		 ((uint128_t) in[1]) * inx2[4] +
+		 ((uint128_t) in[2]) * inx2[3];
+	out[6] = ((uint128_t) in[0]) * inx2[6] +
+		 ((uint128_t) in[1]) * inx2[5] +
+		 ((uint128_t) in[2]) * inx2[4] +
+		 ((uint128_t) in[3]) * in[3];
+	out[7] = ((uint128_t) in[0]) * inx2[7] +
+		 ((uint128_t) in[1]) * inx2[6] +
+		 ((uint128_t) in[2]) * inx2[5] +
+		 ((uint128_t) in[3]) * inx2[4];
+	out[8] = ((uint128_t) in[0]) * inx2[8] +
+		 ((uint128_t) in[1]) * inx2[7] +
+		 ((uint128_t) in[2]) * inx2[6] +
+		 ((uint128_t) in[3]) * inx2[5] +
+		 ((uint128_t) in[4]) * in[4];
+
+	/* The remaining limbs fall above 2^521, with the first falling at
+	 * 2^522. They correspond to locations one bit up from the limbs
+	 * produced above so we would have to multiply by two to align them.
+	 * Again, rather than operate on the 128-bit result, we double one of
+	 * the inputs to the multiplication. If we want to double for both this
+	 * reason, and the reason above, then we end up multiplying by four. */
+
+	/* 9 */
+	out[0] += ((uint128_t) in[1]) * inx4[8] +
+		  ((uint128_t) in[2]) * inx4[7] +
+		  ((uint128_t) in[3]) * inx4[6] +
+		  ((uint128_t) in[4]) * inx4[5];
+
+	/* 10 */
+	out[1] += ((uint128_t) in[2]) * inx4[8] +
+		  ((uint128_t) in[3]) * inx4[7] +
+		  ((uint128_t) in[4]) * inx4[6] +
+		  ((uint128_t) in[5]) * inx2[5];
+
+	/* 11 */
+	out[2] += ((uint128_t) in[3]) * inx4[8] +
+		  ((uint128_t) in[4]) * inx4[7] +
+		  ((uint128_t) in[5]) * inx4[6];
+
+	/* 12 */
+	out[3] += ((uint128_t) in[4]) * inx4[8] +
+		  ((uint128_t) in[5]) * inx4[7] +
+		  ((uint128_t) in[6]) * inx2[6];
+
+	/* 13 */
+	out[4] += ((uint128_t) in[5]) * inx4[8] +
+		  ((uint128_t) in[6]) * inx4[7];
+
+	/* 14 */
+	out[5] += ((uint128_t) in[6]) * inx4[8] +
+		  ((uint128_t) in[7]) * inx2[7];
+
+	/* 15 */
+	out[6] += ((uint128_t) in[7]) * inx4[8];
+
+	/* 16 */
+	out[7] += ((uint128_t) in[8]) * inx2[8];
+	}
+
+/* felem_mul sets |out| = |in1| * |in2|
+ * On entry:
+ *   in1[i] < 2^64
+ *   in2[i] < 2^63
+ * On exit:
+ *   out[i] < 17 * max(in1[i]) * max(in2[i])
+ */
+static void felem_mul(largefelem out, const felem in1, const felem in2)
+	{
+	felem in2x2;
+	felem_scalar(in2x2, in2, 2);
+
+	out[0] = ((uint128_t) in1[0]) * in2[0];
+
+	out[1] = ((uint128_t) in1[0]) * in2[1] +
+	         ((uint128_t) in1[1]) * in2[0];
+
+	out[2] = ((uint128_t) in1[0]) * in2[2] +
+		 ((uint128_t) in1[1]) * in2[1] +
+	         ((uint128_t) in1[2]) * in2[0];
+
+	out[3] = ((uint128_t) in1[0]) * in2[3] +
+		 ((uint128_t) in1[1]) * in2[2] +
+		 ((uint128_t) in1[2]) * in2[1] +
+		 ((uint128_t) in1[3]) * in2[0];
+
+	out[4] = ((uint128_t) in1[0]) * in2[4] +
+		 ((uint128_t) in1[1]) * in2[3] +
+		 ((uint128_t) in1[2]) * in2[2] +
+		 ((uint128_t) in1[3]) * in2[1] +
+		 ((uint128_t) in1[4]) * in2[0];
+
+	out[5] = ((uint128_t) in1[0]) * in2[5] +
+		 ((uint128_t) in1[1]) * in2[4] +
+		 ((uint128_t) in1[2]) * in2[3] +
+		 ((uint128_t) in1[3]) * in2[2] +
+		 ((uint128_t) in1[4]) * in2[1] +
+		 ((uint128_t) in1[5]) * in2[0];
+
+	out[6] = ((uint128_t) in1[0]) * in2[6] +
+		 ((uint128_t) in1[1]) * in2[5] +
+		 ((uint128_t) in1[2]) * in2[4] +
+		 ((uint128_t) in1[3]) * in2[3] +
+		 ((uint128_t) in1[4]) * in2[2] +
+		 ((uint128_t) in1[5]) * in2[1] +
+		 ((uint128_t) in1[6]) * in2[0];
+
+	out[7] = ((uint128_t) in1[0]) * in2[7] +
+		 ((uint128_t) in1[1]) * in2[6] +
+		 ((uint128_t) in1[2]) * in2[5] +
+		 ((uint128_t) in1[3]) * in2[4] +
+		 ((uint128_t) in1[4]) * in2[3] +
+		 ((uint128_t) in1[5]) * in2[2] +
+		 ((uint128_t) in1[6]) * in2[1] +
+		 ((uint128_t) in1[7]) * in2[0];
+
+	out[8] = ((uint128_t) in1[0]) * in2[8] +
+		 ((uint128_t) in1[1]) * in2[7] +
+		 ((uint128_t) in1[2]) * in2[6] +
+		 ((uint128_t) in1[3]) * in2[5] +
+		 ((uint128_t) in1[4]) * in2[4] +
+		 ((uint128_t) in1[5]) * in2[3] +
+		 ((uint128_t) in1[6]) * in2[2] +
+		 ((uint128_t) in1[7]) * in2[1] +
+		 ((uint128_t) in1[8]) * in2[0];
+
+	/* See comment in felem_square about the use of in2x2 here */
+
+	out[0] += ((uint128_t) in1[1]) * in2x2[8] +
+		  ((uint128_t) in1[2]) * in2x2[7] +
+		  ((uint128_t) in1[3]) * in2x2[6] +
+		  ((uint128_t) in1[4]) * in2x2[5] +
+		  ((uint128_t) in1[5]) * in2x2[4] +
+		  ((uint128_t) in1[6]) * in2x2[3] +
+		  ((uint128_t) in1[7]) * in2x2[2] +
+		  ((uint128_t) in1[8]) * in2x2[1];
+
+	out[1] += ((uint128_t) in1[2]) * in2x2[8] +
+		  ((uint128_t) in1[3]) * in2x2[7] +
+		  ((uint128_t) in1[4]) * in2x2[6] +
+		  ((uint128_t) in1[5]) * in2x2[5] +
+		  ((uint128_t) in1[6]) * in2x2[4] +
+		  ((uint128_t) in1[7]) * in2x2[3] +
+		  ((uint128_t) in1[8]) * in2x2[2];
+
+	out[2] += ((uint128_t) in1[3]) * in2x2[8] +
+		  ((uint128_t) in1[4]) * in2x2[7] +
+		  ((uint128_t) in1[5]) * in2x2[6] +
+		  ((uint128_t) in1[6]) * in2x2[5] +
+		  ((uint128_t) in1[7]) * in2x2[4] +
+		  ((uint128_t) in1[8]) * in2x2[3];
+
+	out[3] += ((uint128_t) in1[4]) * in2x2[8] +
+		  ((uint128_t) in1[5]) * in2x2[7] +
+		  ((uint128_t) in1[6]) * in2x2[6] +
+		  ((uint128_t) in1[7]) * in2x2[5] +
+		  ((uint128_t) in1[8]) * in2x2[4];
+
+	out[4] += ((uint128_t) in1[5]) * in2x2[8] +
+		  ((uint128_t) in1[6]) * in2x2[7] +
+		  ((uint128_t) in1[7]) * in2x2[6] +
+		  ((uint128_t) in1[8]) * in2x2[5];
+
+	out[5] += ((uint128_t) in1[6]) * in2x2[8] +
+		  ((uint128_t) in1[7]) * in2x2[7] +
+		  ((uint128_t) in1[8]) * in2x2[6];
+
+	out[6] += ((uint128_t) in1[7]) * in2x2[8] +
+		  ((uint128_t) in1[8]) * in2x2[7];
+
+	out[7] += ((uint128_t) in1[8]) * in2x2[8];
+	}
+
+static const limb bottom52bits = 0xfffffffffffff;
+
+/* felem_reduce converts a largefelem to an felem.
+ * On entry:
+ *   in[i] < 2^128
+ * On exit:
+ *   out[i] < 2^59 + 2^14
+ */
+static void felem_reduce(felem out, const largefelem in)
+	{
+	u64 overflow1, overflow2;
+
+	out[0] = ((limb) in[0]) & bottom58bits;
+	out[1] = ((limb) in[1]) & bottom58bits;
+	out[2] = ((limb) in[2]) & bottom58bits;
+	out[3] = ((limb) in[3]) & bottom58bits;
+	out[4] = ((limb) in[4]) & bottom58bits;
+	out[5] = ((limb) in[5]) & bottom58bits;
+	out[6] = ((limb) in[6]) & bottom58bits;
+	out[7] = ((limb) in[7]) & bottom58bits;
+	out[8] = ((limb) in[8]) & bottom58bits;
+
+	/* out[i] < 2^58 */
+
+	out[1] += ((limb) in[0]) >> 58;
+	out[1] += (((limb) (in[0] >> 64)) & bottom52bits) << 6;
+	/* out[1] < 2^58 + 2^6 + 2^58
+	 *        = 2^59 + 2^6 */
+	out[2] += ((limb) (in[0] >> 64)) >> 52;
+
+	out[2] += ((limb) in[1]) >> 58;
+	out[2] += (((limb) (in[1] >> 64)) & bottom52bits) << 6;
+	out[3] += ((limb) (in[1] >> 64)) >> 52;
+
+	out[3] += ((limb) in[2]) >> 58;
+	out[3] += (((limb) (in[2] >> 64)) & bottom52bits) << 6;
+	out[4] += ((limb) (in[2] >> 64)) >> 52;
+
+	out[4] += ((limb) in[3]) >> 58;
+	out[4] += (((limb) (in[3] >> 64)) & bottom52bits) << 6;
+	out[5] += ((limb) (in[3] >> 64)) >> 52;
+
+	out[5] += ((limb) in[4]) >> 58;
+	out[5] += (((limb) (in[4] >> 64)) & bottom52bits) << 6;
+	out[6] += ((limb) (in[4] >> 64)) >> 52;
+
+	out[6] += ((limb) in[5]) >> 58;
+	out[6] += (((limb) (in[5] >> 64)) & bottom52bits) << 6;
+	out[7] += ((limb) (in[5] >> 64)) >> 52;
+
+	out[7] += ((limb) in[6]) >> 58;
+	out[7] += (((limb) (in[6] >> 64)) & bottom52bits) << 6;
+	out[8] += ((limb) (in[6] >> 64)) >> 52;
+
+	out[8] += ((limb) in[7]) >> 58;
+	out[8] += (((limb) (in[7] >> 64)) & bottom52bits) << 6;
+	/* out[x > 1] < 2^58 + 2^6 + 2^58 + 2^12
+	 *            < 2^59 + 2^13 */
+	overflow1 = ((limb) (in[7] >> 64)) >> 52;
+
+	overflow1 += ((limb) in[8]) >> 58;
+	overflow1 += (((limb) (in[8] >> 64)) & bottom52bits) << 6;
+	overflow2 = ((limb) (in[8] >> 64)) >> 52;
+
+	overflow1 <<= 1;  /* overflow1 < 2^13 + 2^7 + 2^59 */
+	overflow2 <<= 1;  /* overflow2 < 2^13 */
+
+	out[0] += overflow1;  /* out[0] < 2^60 */
+	out[1] += overflow2;  /* out[1] < 2^59 + 2^6 + 2^13 */
+
+	out[1] += out[0] >> 58; out[0] &= bottom58bits;
+	/* out[0] < 2^58
+	 * out[1] < 2^59 + 2^6 + 2^13 + 2^2
+	 *        < 2^59 + 2^14 */
+	}
+
+static void felem_square_reduce(felem out, const felem in)
+	{
+	largefelem tmp;
+	felem_square(tmp, in);
+	felem_reduce(out, tmp);
+	}
+
+static void felem_mul_reduce(felem out, const felem in1, const felem in2)
+	{
+	largefelem tmp;
+	felem_mul(tmp, in1, in2);
+	felem_reduce(out, tmp);
+	}
+
+/* felem_inv calculates |out| = |in|^{-1}
+ *
+ * Based on Fermat's Little Theorem:
+ *   a^p = a (mod p)
+ *   a^{p-1} = 1 (mod p)
+ *   a^{p-2} = a^{-1} (mod p)
+ */
+static void felem_inv(felem out, const felem in)
+	{
+	felem ftmp, ftmp2, ftmp3, ftmp4;
+	largefelem tmp;
+	unsigned i;
+
+	felem_square(tmp, in); felem_reduce(ftmp, tmp);		/* 2^1 */
+	felem_mul(tmp, in, ftmp); felem_reduce(ftmp, tmp);	/* 2^2 - 2^0 */
+	felem_assign(ftmp2, ftmp);
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);	/* 2^3 - 2^1 */
+	felem_mul(tmp, in, ftmp); felem_reduce(ftmp, tmp);	/* 2^3 - 2^0 */
+	felem_square(tmp, ftmp); felem_reduce(ftmp, tmp);	/* 2^4 - 2^1 */
+
+	felem_square(tmp, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^3 - 2^1 */
+	felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^4 - 2^2 */
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^4 - 2^0 */
+
+	felem_assign(ftmp2, ftmp3);
+	felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^5 - 2^1 */
+	felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^6 - 2^2 */
+	felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^7 - 2^3 */
+	felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^8 - 2^4 */
+	felem_assign(ftmp4, ftmp3);
+	felem_mul(tmp, ftmp3, ftmp); felem_reduce(ftmp4, tmp);	/* 2^8 - 2^1 */
+	felem_square(tmp, ftmp4); felem_reduce(ftmp4, tmp);	/* 2^9 - 2^2 */
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^8 - 2^0 */
+	felem_assign(ftmp2, ftmp3);
+
+	for (i = 0; i < 8; i++)
+		{
+		felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^16 - 2^8 */
+		}
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^16 - 2^0 */
+	felem_assign(ftmp2, ftmp3);
+
+	for (i = 0; i < 16; i++)
+		{
+		felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^32 - 2^16 */
+		}
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^32 - 2^0 */
+	felem_assign(ftmp2, ftmp3);
+
+	for (i = 0; i < 32; i++)
+		{
+		felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^64 - 2^32 */
+		}
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^64 - 2^0 */
+	felem_assign(ftmp2, ftmp3);
+
+	for (i = 0; i < 64; i++)
+		{
+		felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^128 - 2^64 */
+		}
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^128 - 2^0 */
+	felem_assign(ftmp2, ftmp3);
+
+	for (i = 0; i < 128; i++)
+		{
+		felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^256 - 2^128 */
+		}
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^256 - 2^0 */
+	felem_assign(ftmp2, ftmp3);
+
+	for (i = 0; i < 256; i++)
+		{
+		felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^512 - 2^256 */
+		}
+	felem_mul(tmp, ftmp3, ftmp2); felem_reduce(ftmp3, tmp);	/* 2^512 - 2^0 */
+
+	for (i = 0; i < 9; i++)
+		{
+		felem_square(tmp, ftmp3); felem_reduce(ftmp3, tmp);	/* 2^521 - 2^9 */
+		}
+	felem_mul(tmp, ftmp3, ftmp4); felem_reduce(ftmp3, tmp);	/* 2^512 - 2^2 */
+	felem_mul(tmp, ftmp3, in); felem_reduce(out, tmp);	/* 2^512 - 3 */
+}
+
+/* This is 2^521-1, expressed as an felem */
+static const felem kPrime =
+	{
+	0x03ffffffffffffff, 0x03ffffffffffffff, 0x03ffffffffffffff,
+	0x03ffffffffffffff, 0x03ffffffffffffff, 0x03ffffffffffffff,
+	0x03ffffffffffffff, 0x03ffffffffffffff, 0x01ffffffffffffff
+	};
+
+/* felem_is_zero returns a limb with all bits set if |in| == 0 (mod p) and 0
+ * otherwise.
+ * On entry:
+ *   in[i] < 2^59 + 2^14
+ */
+static limb felem_is_zero(const felem in)
+	{
+	felem ftmp;
+	limb is_zero, is_p;
+	felem_assign(ftmp, in);
+
+	ftmp[0] += ftmp[8] >> 57; ftmp[8] &= bottom57bits;
+	/* ftmp[8] < 2^57 */
+	ftmp[1] += ftmp[0] >> 58; ftmp[0] &= bottom58bits;
+	ftmp[2] += ftmp[1] >> 58; ftmp[1] &= bottom58bits;
+	ftmp[3] += ftmp[2] >> 58; ftmp[2] &= bottom58bits;
+	ftmp[4] += ftmp[3] >> 58; ftmp[3] &= bottom58bits;
+	ftmp[5] += ftmp[4] >> 58; ftmp[4] &= bottom58bits;
+	ftmp[6] += ftmp[5] >> 58; ftmp[5] &= bottom58bits;
+	ftmp[7] += ftmp[6] >> 58; ftmp[6] &= bottom58bits;
+	ftmp[8] += ftmp[7] >> 58; ftmp[7] &= bottom58bits;
+	/* ftmp[8] < 2^57 + 4 */
+
+	/* The ninth limb of 2*(2^521-1) is 0x03ffffffffffffff, which is
+	 * greater than our bound for ftmp[8]. Therefore we only have to check
+	 * if the zero is zero or 2^521-1. */
+
+	is_zero = 0;
+	is_zero |= ftmp[0];
+	is_zero |= ftmp[1];
+	is_zero |= ftmp[2];
+	is_zero |= ftmp[3];
+	is_zero |= ftmp[4];
+	is_zero |= ftmp[5];
+	is_zero |= ftmp[6];
+	is_zero |= ftmp[7];
+	is_zero |= ftmp[8];
+
+	is_zero--;
+	/* We know that ftmp[i] < 2^63, therefore the only way that the top bit
+	 * can be set is if is_zero was 0 before the decrement. */
+	is_zero = ((s64) is_zero) >> 63;
+
+	is_p = ftmp[0] ^ kPrime[0];
+	is_p |= ftmp[1] ^ kPrime[1];
+	is_p |= ftmp[2] ^ kPrime[2];
+	is_p |= ftmp[3] ^ kPrime[3];
+	is_p |= ftmp[4] ^ kPrime[4];
+	is_p |= ftmp[5] ^ kPrime[5];
+	is_p |= ftmp[6] ^ kPrime[6];
+	is_p |= ftmp[7] ^ kPrime[7];
+	is_p |= ftmp[8] ^ kPrime[8];
+
+	is_p--;
+	is_p = ((s64) is_p) >> 63;
+
+	is_zero |= is_p;
+	return is_zero;
+	}
+
+static int felem_is_zero_int(const felem in)
+	{
+	return (int) (felem_is_zero(in) & ((limb)1));
+	}
+
+/* felem_contract converts |in| to its unique, minimal representation.
+ * On entry:
+ *   in[i] < 2^59 + 2^14
+ */
+static void felem_contract(felem out, const felem in)
+	{
+	limb is_p, is_greater, sign;
+	static const limb two58 = ((limb)1) << 58;
+
+	felem_assign(out, in);
+
+	out[0] += out[8] >> 57; out[8] &= bottom57bits;
+	/* out[8] < 2^57 */
+	out[1] += out[0] >> 58; out[0] &= bottom58bits;
+	out[2] += out[1] >> 58; out[1] &= bottom58bits;
+	out[3] += out[2] >> 58; out[2] &= bottom58bits;
+	out[4] += out[3] >> 58; out[3] &= bottom58bits;
+	out[5] += out[4] >> 58; out[4] &= bottom58bits;
+	out[6] += out[5] >> 58; out[5] &= bottom58bits;
+	out[7] += out[6] >> 58; out[6] &= bottom58bits;
+	out[8] += out[7] >> 58; out[7] &= bottom58bits;
+	/* out[8] < 2^57 + 4 */
+
+	/* If the value is greater than 2^521-1 then we have to subtract
+	 * 2^521-1 out. See the comments in felem_is_zero regarding why we
+	 * don't test for other multiples of the prime. */
+
+	/* First, if |out| is equal to 2^521-1, we subtract it out to get zero. */
+
+	is_p = out[0] ^ kPrime[0];
+	is_p |= out[1] ^ kPrime[1];
+	is_p |= out[2] ^ kPrime[2];
+	is_p |= out[3] ^ kPrime[3];
+	is_p |= out[4] ^ kPrime[4];
+	is_p |= out[5] ^ kPrime[5];
+	is_p |= out[6] ^ kPrime[6];
+	is_p |= out[7] ^ kPrime[7];
+	is_p |= out[8] ^ kPrime[8];
+
+	is_p--;
+	is_p &= is_p << 32;
+	is_p &= is_p << 16;
+	is_p &= is_p << 8;
+	is_p &= is_p << 4;
+	is_p &= is_p << 2;
+	is_p &= is_p << 1;
+	is_p = ((s64) is_p) >> 63;
+	is_p = ~is_p;
+
+	/* is_p is 0 iff |out| == 2^521-1 and all ones otherwise */
+
+	out[0] &= is_p;
+	out[1] &= is_p;
+	out[2] &= is_p;
+	out[3] &= is_p;
+	out[4] &= is_p;
+	out[5] &= is_p;
+	out[6] &= is_p;
+	out[7] &= is_p;
+	out[8] &= is_p;
+
+	/* In order to test that |out| >= 2^521-1 we need only test if out[8]
+	 * >> 57 is greater than zero as (2^521-1) + x >= 2^522 */
+	is_greater = out[8] >> 57;
+	is_greater |= is_greater << 32;
+	is_greater |= is_greater << 16;
+	is_greater |= is_greater << 8;
+	is_greater |= is_greater << 4;
+	is_greater |= is_greater << 2;
+	is_greater |= is_greater << 1;
+	is_greater = ((s64) is_greater) >> 63;
+
+	out[0] -= kPrime[0] & is_greater;
+	out[1] -= kPrime[1] & is_greater;
+	out[2] -= kPrime[2] & is_greater;
+	out[3] -= kPrime[3] & is_greater;
+	out[4] -= kPrime[4] & is_greater;
+	out[5] -= kPrime[5] & is_greater;
+	out[6] -= kPrime[6] & is_greater;
+	out[7] -= kPrime[7] & is_greater;
+	out[8] -= kPrime[8] & is_greater;
+
+	/* Eliminate negative coefficients */
+	sign = -(out[0] >> 63); out[0] += (two58 & sign); out[1] -= (1 & sign);
+	sign = -(out[1] >> 63); out[1] += (two58 & sign); out[2] -= (1 & sign);
+	sign = -(out[2] >> 63); out[2] += (two58 & sign); out[3] -= (1 & sign);
+	sign = -(out[3] >> 63); out[3] += (two58 & sign); out[4] -= (1 & sign);
+	sign = -(out[4] >> 63); out[4] += (two58 & sign); out[5] -= (1 & sign);
+	sign = -(out[0] >> 63); out[5] += (two58 & sign); out[6] -= (1 & sign);
+	sign = -(out[6] >> 63); out[6] += (two58 & sign); out[7] -= (1 & sign);
+	sign = -(out[7] >> 63); out[7] += (two58 & sign); out[8] -= (1 & sign);
+	sign = -(out[5] >> 63); out[5] += (two58 & sign); out[6] -= (1 & sign);
+	sign = -(out[6] >> 63); out[6] += (two58 & sign); out[7] -= (1 & sign);
+	sign = -(out[7] >> 63); out[7] += (two58 & sign); out[8] -= (1 & sign);
+	}
+
+/* Group operations
+ * ----------------
+ *
+ * Building on top of the field operations we have the operations on the
+ * elliptic curve group itself. Points on the curve are represented in Jacobian
+ * coordinates */
+
+/* point_double calcuates 2*(x_in, y_in, z_in)
+ *
+ * The method is taken from:
+ *   http://hyperelliptic.org/EFD/g1p/auto-shortw-jacobian-3.html#doubling-dbl-2001-b
+ *
+ * Outputs can equal corresponding inputs, i.e., x_out == x_in is allowed.
+ * while x_out == y_in is not (maybe this works, but it's not tested). */
+static void
+point_double(felem x_out, felem y_out, felem z_out,
+	     const felem x_in, const felem y_in, const felem z_in)
+	{
+	largefelem tmp, tmp2;
+	felem delta, gamma, beta, alpha, ftmp, ftmp2;
+
+	felem_assign(ftmp, x_in);
+	felem_assign(ftmp2, x_in);
+
+	/* delta = z^2 */
+	felem_square(tmp, z_in);
+	felem_reduce(delta, tmp);  /* delta[i] < 2^59 + 2^14 */
+
+	/* gamma = y^2 */
+	felem_square(tmp, y_in);
+	felem_reduce(gamma, tmp);  /* gamma[i] < 2^59 + 2^14 */
+
+	/* beta = x*gamma */
+	felem_mul(tmp, x_in, gamma);
+	felem_reduce(beta, tmp);  /* beta[i] < 2^59 + 2^14 */
+
+	/* alpha = 3*(x-delta)*(x+delta) */
+	felem_diff64(ftmp, delta);
+	/* ftmp[i] < 2^61 */
+	felem_sum64(ftmp2, delta);
+	/* ftmp2[i] < 2^60 + 2^15 */
+	felem_scalar64(ftmp2, 3);
+	/* ftmp2[i] < 3*2^60 + 3*2^15 */
+	felem_mul(tmp, ftmp, ftmp2);
+	/* tmp[i] < 17(3*2^121 + 3*2^76)
+	 *        = 61*2^121 + 61*2^76
+	 *        < 64*2^121 + 64*2^76
+	 *        = 2^127 + 2^82
+	 *        < 2^128 */
+	felem_reduce(alpha, tmp);
+
+	/* x' = alpha^2 - 8*beta */
+	felem_square(tmp, alpha);
+	/* tmp[i] < 17*2^120
+	 *        < 2^125 */
+	felem_assign(ftmp, beta);
+	felem_scalar64(ftmp, 8);
+	/* ftmp[i] < 2^62 + 2^17 */
+	felem_diff_128_64(tmp, ftmp);
+	/* tmp[i] < 2^125 + 2^63 + 2^62 + 2^17 */
+	felem_reduce(x_out, tmp);
+
+	/* z' = (y + z)^2 - gamma - delta */
+	felem_sum64(delta, gamma);
+	/* delta[i] < 2^60 + 2^15 */
+	felem_assign(ftmp, y_in);
+	felem_sum64(ftmp, z_in);
+	/* ftmp[i] < 2^60 + 2^15 */
+	felem_square(tmp, ftmp);
+	/* tmp[i] < 17(2^122)
+	 *        < 2^127 */
+	felem_diff_128_64(tmp, delta);
+	/* tmp[i] < 2^127 + 2^63 */
+	felem_reduce(z_out, tmp);
+
+	/* y' = alpha*(4*beta - x') - 8*gamma^2 */
+	felem_scalar64(beta, 4);
+	/* beta[i] < 2^61 + 2^16 */
+	felem_diff64(beta, x_out);
+	/* beta[i] < 2^61 + 2^60 + 2^16 */
+	felem_mul(tmp, alpha, beta);
+	/* tmp[i] < 17*((2^59 + 2^14)(2^61 + 2^60 + 2^16))
+	 *        = 17*(2^120 + 2^75 + 2^119 + 2^74 + 2^75 + 2^30) 
+	 *        = 17*(2^120 + 2^119 + 2^76 + 2^74 + 2^30)
+	 *        < 2^128 */
+	felem_square(tmp2, gamma);
+	/* tmp2[i] < 17*(2^59 + 2^14)^2
+	 *         = 17*(2^118 + 2^74 + 2^28) */
+	felem_scalar128(tmp2, 8);
+	/* tmp2[i] < 8*17*(2^118 + 2^74 + 2^28)
+	 *         = 2^125 + 2^121 + 2^81 + 2^77 + 2^35 + 2^31
+	 *         < 2^126 */
+	felem_diff128(tmp, tmp2);
+	/* tmp[i] < 2^127 - 2^69 + 17(2^120 + 2^119 + 2^76 + 2^74 + 2^30)
+	 *        = 2^127 + 2^124 + 2^122 + 2^120 + 2^118 + 2^80 + 2^78 + 2^76 +
+	 *          2^74 + 2^69 + 2^34 + 2^30
+	 *        < 2^128 */
+	felem_reduce(y_out, tmp);
+	}
+
+/* copy_conditional copies in to out iff mask is all ones. */
+static void
+copy_conditional(felem out, const felem in, limb mask)
+	{
+	unsigned i;
+	for (i = 0; i < NLIMBS; ++i)
+		{
+		const limb tmp = mask & (in[i] ^ out[i]);
+		out[i] ^= tmp;
+		}
+	}
+
+/* point_add calcuates (x1, y1, z1) + (x2, y2, z2)
+ *
+ * The method is taken from
+ *   http://hyperelliptic.org/EFD/g1p/auto-shortw-jacobian-3.html#addition-add-2007-bl,
+ * adapted for mixed addition (z2 = 1, or z2 = 0 for the point at infinity).
+ *
+ * This function includes a branch for checking whether the two input points
+ * are equal (while not equal to the point at infinity). This case never
+ * happens during single point multiplication, so there is no timing leak for
+ * ECDH or ECDSA signing. */
+static void point_add(felem x3, felem y3, felem z3,
+	const felem x1, const felem y1, const felem z1,
+	const int mixed, const felem x2, const felem y2, const felem z2)
+	{
+	felem ftmp, ftmp2, ftmp3, ftmp4, ftmp5, ftmp6, x_out, y_out, z_out;
+	largefelem tmp, tmp2;
+	limb x_equal, y_equal, z1_is_zero, z2_is_zero;
+
+	z1_is_zero = felem_is_zero(z1);
+	z2_is_zero = felem_is_zero(z2);
+
+	/* ftmp = z1z1 = z1**2 */
+	felem_square(tmp, z1);
+	felem_reduce(ftmp, tmp);
+
+	if (!mixed)
+		{
+		/* ftmp2 = z2z2 = z2**2 */
+		felem_square(tmp, z2);
+		felem_reduce(ftmp2, tmp);
+
+		/* u1 = ftmp3 = x1*z2z2 */
+		felem_mul(tmp, x1, ftmp2);
+		felem_reduce(ftmp3, tmp);
+
+		/* ftmp5 = z1 + z2 */
+		felem_assign(ftmp5, z1);
+		felem_sum64(ftmp5, z2);
+		/* ftmp5[i] < 2^61 */
+
+		/* ftmp5 = (z1 + z2)**2 - z1z1 - z2z2 = 2*z1z2 */
+		felem_square(tmp, ftmp5);
+		/* tmp[i] < 17*2^122 */
+		felem_diff_128_64(tmp, ftmp);
+		/* tmp[i] < 17*2^122 + 2^63 */
+		felem_diff_128_64(tmp, ftmp2);
+		/* tmp[i] < 17*2^122 + 2^64 */
+		felem_reduce(ftmp5, tmp);
+
+		/* ftmp2 = z2 * z2z2 */
+		felem_mul(tmp, ftmp2, z2);
+		felem_reduce(ftmp2, tmp);
+
+		/* s1 = ftmp6 = y1 * z2**3 */
+		felem_mul(tmp, y1, ftmp2);
+		felem_reduce(ftmp6, tmp);
+		}
+	else
+		{
+		/* We'll assume z2 = 1 (special case z2 = 0 is handled later) */
+
+		/* u1 = ftmp3 = x1*z2z2 */
+		felem_assign(ftmp3, x1);
+
+		/* ftmp5 = 2*z1z2 */
+		felem_scalar(ftmp5, z1, 2);
+
+		/* s1 = ftmp6 = y1 * z2**3 */
+		felem_assign(ftmp6, y1);
+		}
+
+	/* u2 = x2*z1z1 */
+	felem_mul(tmp, x2, ftmp);
+	/* tmp[i] < 17*2^120 */
+
+	/* h = ftmp4 = u2 - u1 */
+	felem_diff_128_64(tmp, ftmp3);
+	/* tmp[i] < 17*2^120 + 2^63 */
+	felem_reduce(ftmp4, tmp);
+
+	x_equal = felem_is_zero(ftmp4);
+
+	/* z_out = ftmp5 * h */
+	felem_mul(tmp, ftmp5, ftmp4);
+	felem_reduce(z_out, tmp);
+
+	/* ftmp = z1 * z1z1 */
+	felem_mul(tmp, ftmp, z1);
+	felem_reduce(ftmp, tmp);
+
+	/* s2 = tmp = y2 * z1**3 */
+	felem_mul(tmp, y2, ftmp);
+	/* tmp[i] < 17*2^120 */
+
+	/* r = ftmp5 = (s2 - s1)*2 */
+	felem_diff_128_64(tmp, ftmp6);
+	/* tmp[i] < 17*2^120 + 2^63 */
+	felem_reduce(ftmp5, tmp);
+	y_equal = felem_is_zero(ftmp5);
+	felem_scalar64(ftmp5, 2);
+	/* ftmp5[i] < 2^61 */
+
+	if (x_equal && y_equal && !z1_is_zero && !z2_is_zero)
+		{
+		point_double(x3, y3, z3, x1, y1, z1);
+		return;
+		}
+
+	/* I = ftmp = (2h)**2 */
+	felem_assign(ftmp, ftmp4);
+	felem_scalar64(ftmp, 2);
+	/* ftmp[i] < 2^61 */
+	felem_square(tmp, ftmp);
+	/* tmp[i] < 17*2^122 */
+	felem_reduce(ftmp, tmp);
+
+	/* J = ftmp2 = h * I */
+	felem_mul(tmp, ftmp4, ftmp);
+	felem_reduce(ftmp2, tmp);
+
+	/* V = ftmp4 = U1 * I */
+	felem_mul(tmp, ftmp3, ftmp);
+	felem_reduce(ftmp4, tmp);
+
+	/* x_out = r**2 - J - 2V */
+	felem_square(tmp, ftmp5);
+	/* tmp[i] < 17*2^122 */
+	felem_diff_128_64(tmp, ftmp2);
+	/* tmp[i] < 17*2^122 + 2^63 */
+	felem_assign(ftmp3, ftmp4);
+	felem_scalar64(ftmp4, 2);
+	/* ftmp4[i] < 2^61 */
+	felem_diff_128_64(tmp, ftmp4);
+	/* tmp[i] < 17*2^122 + 2^64 */
+	felem_reduce(x_out, tmp);
+
+	/* y_out = r(V-x_out) - 2 * s1 * J */
+	felem_diff64(ftmp3, x_out);
+	/* ftmp3[i] < 2^60 + 2^60
+	 *          = 2^61 */
+	felem_mul(tmp, ftmp5, ftmp3);
+	/* tmp[i] < 17*2^122 */
+	felem_mul(tmp2, ftmp6, ftmp2);
+	/* tmp2[i] < 17*2^120 */
+	felem_scalar128(tmp2, 2);
+	/* tmp2[i] < 17*2^121 */
+	felem_diff128(tmp, tmp2);
+	/* tmp[i] < 2^127 - 2^69 + 17*2^122
+	 *        = 2^126 - 2^122 - 2^6 - 2^2 - 1
+	 *        < 2^127 */
+	felem_reduce(y_out, tmp);
+
+	copy_conditional(x_out, x2, z1_is_zero);
+	copy_conditional(x_out, x1, z2_is_zero);
+	copy_conditional(y_out, y2, z1_is_zero);
+	copy_conditional(y_out, y1, z2_is_zero);
+	copy_conditional(z_out, z2, z1_is_zero);
+	copy_conditional(z_out, z1, z2_is_zero);
+	felem_assign(x3, x_out);
+	felem_assign(y3, y_out);
+	felem_assign(z3, z_out);
+	}
+
+/* Base point pre computation
+ * --------------------------
+ *
+ * Two different sorts of precomputed tables are used in the following code.
+ * Each contain various points on the curve, where each point is three field
+ * elements (x, y, z).
+ *
+ * For the base point table, z is usually 1 (0 for the point at infinity).
+ * This table has 16 elements:
+ * index | bits    | point
+ * ------+---------+------------------------------
+ *     0 | 0 0 0 0 | 0G
+ *     1 | 0 0 0 1 | 1G
+ *     2 | 0 0 1 0 | 2^130G
+ *     3 | 0 0 1 1 | (2^130 + 1)G
+ *     4 | 0 1 0 0 | 2^260G
+ *     5 | 0 1 0 1 | (2^260 + 1)G
+ *     6 | 0 1 1 0 | (2^260 + 2^130)G
+ *     7 | 0 1 1 1 | (2^260 + 2^130 + 1)G
+ *     8 | 1 0 0 0 | 2^390G
+ *     9 | 1 0 0 1 | (2^390 + 1)G
+ *    10 | 1 0 1 0 | (2^390 + 2^130)G
+ *    11 | 1 0 1 1 | (2^390 + 2^130 + 1)G
+ *    12 | 1 1 0 0 | (2^390 + 2^260)G
+ *    13 | 1 1 0 1 | (2^390 + 2^260 + 1)G
+ *    14 | 1 1 1 0 | (2^390 + 2^260 + 2^130)G
+ *    15 | 1 1 1 1 | (2^390 + 2^260 + 2^130 + 1)G
+ *
+ * The reason for this is so that we can clock bits into four different
+ * locations when doing simple scalar multiplies against the base point.
+ *
+ * Tables for other points have table[i] = iG for i in 0 .. 16. */
+
+/* gmul is the table of precomputed base points */
+static const felem gmul[16][3] =
+	{{{0, 0, 0, 0, 0, 0, 0, 0, 0},
+	  {0, 0, 0, 0, 0, 0, 0, 0, 0},
+	  {0, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x017e7e31c2e5bd66, 0x022cf0615a90a6fe, 0x00127a2ffa8de334,
+	   0x01dfbf9d64a3f877, 0x006b4d3dbaa14b5e, 0x014fed487e0a2bd8,
+	   0x015b4429c6481390, 0x03a73678fb2d988e, 0x00c6858e06b70404},
+	  {0x00be94769fd16650, 0x031c21a89cb09022, 0x039013fad0761353,
+	   0x02657bd099031542, 0x03273e662c97ee72, 0x01e6d11a05ebef45,
+	   0x03d1bd998f544495, 0x03001172297ed0b1, 0x011839296a789a3b},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x0373faacbc875bae, 0x00f325023721c671, 0x00f666fd3dbde5ad,
+	   0x01a6932363f88ea7, 0x01fc6d9e13f9c47b, 0x03bcbffc2bbf734e,
+	   0x013ee3c3647f3a92, 0x029409fefe75d07d, 0x00ef9199963d85e5},
+	  {0x011173743ad5b178, 0x02499c7c21bf7d46, 0x035beaeabb8b1a58,
+	   0x00f989c4752ea0a3, 0x0101e1de48a9c1a3, 0x01a20076be28ba6c,
+	   0x02f8052e5eb2de95, 0x01bfe8f82dea117c, 0x0160074d3c36ddb7},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x012f3fc373393b3b, 0x03d3d6172f1419fa, 0x02adc943c0b86873,
+	   0x00d475584177952b, 0x012a4d1673750ee2, 0x00512517a0f13b0c,
+	   0x02b184671a7b1734, 0x0315b84236f1a50a, 0x00a4afc472edbdb9},
+	  {0x00152a7077f385c4, 0x03044007d8d1c2ee, 0x0065829d61d52b52,
+	   0x00494ff6b6631d0d, 0x00a11d94d5f06bcf, 0x02d2f89474d9282e,
+	   0x0241c5727c06eeb9, 0x0386928710fbdb9d, 0x01f883f727b0dfbe},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x019b0c3c9185544d, 0x006243a37c9d97db, 0x02ee3cbe030a2ad2,
+	   0x00cfdd946bb51e0d, 0x0271c00932606b91, 0x03f817d1ec68c561,
+	   0x03f37009806a369c, 0x03c1f30baf184fd5, 0x01091022d6d2f065},
+	  {0x0292c583514c45ed, 0x0316fca51f9a286c, 0x00300af507c1489a,
+	   0x0295f69008298cf1, 0x02c0ed8274943d7b, 0x016509b9b47a431e,
+	   0x02bc9de9634868ce, 0x005b34929bffcb09, 0x000c1a0121681524},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x0286abc0292fb9f2, 0x02665eee9805b3f7, 0x01ed7455f17f26d6,
+	   0x0346355b83175d13, 0x006284944cd0a097, 0x0191895bcdec5e51,
+	   0x02e288370afda7d9, 0x03b22312bfefa67a, 0x01d104d3fc0613fe},
+	  {0x0092421a12f7e47f, 0x0077a83fa373c501, 0x03bd25c5f696bd0d,
+	   0x035c41e4d5459761, 0x01ca0d1742b24f53, 0x00aaab27863a509c,
+	   0x018b6de47df73917, 0x025c0b771705cd01, 0x01fd51d566d760a7},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x01dd92ff6b0d1dbd, 0x039c5e2e8f8afa69, 0x0261ed13242c3b27,
+	   0x0382c6e67026e6a0, 0x01d60b10be2089f9, 0x03c15f3dce86723f,
+	   0x03c764a32d2a062d, 0x017307eac0fad056, 0x018207c0b96c5256},
+	  {0x0196a16d60e13154, 0x03e6ce74c0267030, 0x00ddbf2b4e52a5aa,
+	   0x012738241bbf31c8, 0x00ebe8dc04685a28, 0x024c2ad6d380d4a2,
+	   0x035ee062a6e62d0e, 0x0029ed74af7d3a0f, 0x00eef32aec142ebd},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x00c31ec398993b39, 0x03a9f45bcda68253, 0x00ac733c24c70890,
+	   0x00872b111401ff01, 0x01d178c23195eafb, 0x03bca2c816b87f74,
+	   0x0261a9af46fbad7a, 0x0324b2a8dd3d28f9, 0x00918121d8f24e23},
+	  {0x032bc8c1ca983cd7, 0x00d869dfb08fc8c6, 0x01693cb61fce1516,
+	   0x012a5ea68f4e88a8, 0x010869cab88d7ae3, 0x009081ad277ceee1,
+	   0x033a77166d064cdc, 0x03955235a1fb3a95, 0x01251a4a9b25b65e},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x00148a3a1b27f40b, 0x0123186df1b31fdc, 0x00026e7beaad34ce,
+	   0x01db446ac1d3dbba, 0x0299c1a33437eaec, 0x024540610183cbb7,
+	   0x0173bb0e9ce92e46, 0x02b937e43921214b, 0x01ab0436a9bf01b5},
+	  {0x0383381640d46948, 0x008dacbf0e7f330f, 0x03602122bcc3f318,
+	   0x01ee596b200620d6, 0x03bd0585fda430b3, 0x014aed77fd123a83,
+	   0x005ace749e52f742, 0x0390fe041da2b842, 0x0189a8ceb3299242},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x012a19d6b3282473, 0x00c0915918b423ce, 0x023a954eb94405ae,
+	   0x00529f692be26158, 0x0289fa1b6fa4b2aa, 0x0198ae4ceea346ef,
+	   0x0047d8cdfbdedd49, 0x00cc8c8953f0f6b8, 0x001424abbff49203},
+	  {0x0256732a1115a03a, 0x0351bc38665c6733, 0x03f7b950fb4a6447,
+	   0x000afffa94c22155, 0x025763d0a4dab540, 0x000511e92d4fc283,
+	   0x030a7e9eda0ee96c, 0x004c3cd93a28bf0a, 0x017edb3a8719217f},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x011de5675a88e673, 0x031d7d0f5e567fbe, 0x0016b2062c970ae5,
+	   0x03f4a2be49d90aa7, 0x03cef0bd13822866, 0x03f0923dcf774a6c,
+	   0x0284bebc4f322f72, 0x016ab2645302bb2c, 0x01793f95dace0e2a},
+	  {0x010646e13527a28f, 0x01ca1babd59dc5e7, 0x01afedfd9a5595df,
+	   0x01f15785212ea6b1, 0x0324e5d64f6ae3f4, 0x02d680f526d00645,
+	   0x0127920fadf627a7, 0x03b383f75df4f684, 0x0089e0057e783b0a},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x00f334b9eb3c26c6, 0x0298fdaa98568dce, 0x01c2d24843a82292,
+	   0x020bcb24fa1b0711, 0x02cbdb3d2b1875e6, 0x0014907598f89422,
+	   0x03abe3aa43b26664, 0x02cbf47f720bc168, 0x0133b5e73014b79b},
+	  {0x034aab5dab05779d, 0x00cdc5d71fee9abb, 0x0399f16bd4bd9d30,
+	   0x03582fa592d82647, 0x02be1cdfb775b0e9, 0x0034f7cea32e94cb,
+	   0x0335a7f08f56f286, 0x03b707e9565d1c8b, 0x0015c946ea5b614f},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x024676f6cff72255, 0x00d14625cac96378, 0x00532b6008bc3767,
+	   0x01fc16721b985322, 0x023355ea1b091668, 0x029de7afdc0317c3,
+	   0x02fc8a7ca2da037c, 0x02de1217d74a6f30, 0x013f7173175b73bf},
+	  {0x0344913f441490b5, 0x0200f9e272b61eca, 0x0258a246b1dd55d2,
+	   0x03753db9ea496f36, 0x025e02937a09c5ef, 0x030cbd3d14012692,
+	   0x01793a67e70dc72a, 0x03ec1d37048a662e, 0x006550f700c32a8d},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x00d3f48a347eba27, 0x008e636649b61bd8, 0x00d3b93716778fb3,
+	   0x004d1915757bd209, 0x019d5311a3da44e0, 0x016d1afcbbe6aade,
+	   0x0241bf5f73265616, 0x0384672e5d50d39b, 0x005009fee522b684},
+	  {0x029b4fab064435fe, 0x018868ee095bbb07, 0x01ea3d6936cc92b8,
+	   0x000608b00f78a2f3, 0x02db911073d1c20f, 0x018205938470100a,
+	   0x01f1e4964cbe6ff2, 0x021a19a29eed4663, 0x01414485f42afa81},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x01612b3a17f63e34, 0x03813992885428e6, 0x022b3c215b5a9608,
+	   0x029b4057e19f2fcb, 0x0384059a587af7e6, 0x02d6400ace6fe610,
+	   0x029354d896e8e331, 0x00c047ee6dfba65e, 0x0037720542e9d49d},
+	  {0x02ce9eed7c5e9278, 0x0374ed703e79643b, 0x01316c54c4072006,
+	   0x005aaa09054b2ee8, 0x002824000c840d57, 0x03d4eba24771ed86,
+	   0x0189c50aabc3bdae, 0x0338c01541e15510, 0x00466d56e38eed42},
+	  {1, 0, 0, 0, 0, 0, 0, 0, 0}},
+	 {{0x007efd8330ad8bd6, 0x02465ed48047710b, 0x0034c6606b215e0c,
+	   0x016ae30c53cbf839, 0x01fa17bd37161216, 0x018ead4e61ce8ab9,
+	   0x005482ed5f5dee46, 0x037543755bba1d7f, 0x005e5ac7e70a9d0f},
+	  {0x0117e1bb2fdcb2a2, 0x03deea36249f40c4, 0x028d09b4a6246cb7,
+	   0x03524b8855bcf756, 0x023d7d109d5ceb58, 0x0178e43e3223ef9c,
+	   0x0154536a0c6e966a, 0x037964d1286ee9fe, 0x0199bcd90e125055},
+	 {1, 0, 0, 0, 0, 0, 0, 0, 0}}};
+
+/* select_point selects the |idx|th point from a precomputation table and
+ * copies it to out. */
+static void select_point(const limb idx, unsigned int size, const felem pre_comp[/* size */][3],
+			 felem out[3])
+	{
+	unsigned i, j;
+	limb *outlimbs = &out[0][0];
+	memset(outlimbs, 0, 3 * sizeof(felem));
+
+	for (i = 0; i < size; i++)
+		{
+		const limb *inlimbs = &pre_comp[i][0][0];
+		limb mask = i ^ idx;
+		mask |= mask >> 4;
+		mask |= mask >> 2;
+		mask |= mask >> 1;
+		mask &= 1;
+		mask--;
+		for (j = 0; j < NLIMBS * 3; j++)
+			outlimbs[j] |= inlimbs[j] & mask;
+		}
+	}
+
+/* get_bit returns the |i|th bit in |in| */
+static char get_bit(const felem_bytearray in, int i)
+	{
+	if (i < 0)
+		return 0;
+	return (in[i >> 3] >> (i & 7)) & 1;
+	}
+
+/* Interleaved point multiplication using precomputed point multiples:
+ * The small point multiples 0*P, 1*P, ..., 16*P are in pre_comp[],
+ * the scalars in scalars[]. If g_scalar is non-NULL, we also add this multiple
+ * of the generator, using certain (large) precomputed multiples in g_pre_comp.
+ * Output point (X, Y, Z) is stored in x_out, y_out, z_out */
+static void batch_mul(felem x_out, felem y_out, felem z_out,
+	const felem_bytearray scalars[], const unsigned num_points, const u8 *g_scalar,
+	const int mixed, const felem pre_comp[][17][3], const felem g_pre_comp[16][3])
+	{
+	int i, skip;
+	unsigned num, gen_mul = (g_scalar != NULL);
+	felem nq[3], tmp[4];
+	limb bits;
+	u8 sign, digit;
+
+	/* set nq to the point at infinity */
+	memset(nq, 0, 3 * sizeof(felem));
+
+	/* Loop over all scalars msb-to-lsb, interleaving additions
+	 * of multiples of the generator (last quarter of rounds)
+	 * and additions of other points multiples (every 5th round).
+	 */
+	skip = 1; /* save two point operations in the first round */
+	for (i = (num_points ? 520 : 130); i >= 0; --i)
+		{
+		/* double */
+		if (!skip)
+			point_double(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2]);
+
+		/* add multiples of the generator */
+		if (gen_mul && (i <= 130))
+			{
+			bits = get_bit(g_scalar, i + 390) << 3;
+			if (i < 130)
+				{
+				bits |= get_bit(g_scalar, i + 260) << 2;
+				bits |= get_bit(g_scalar, i + 130) << 1;
+				bits |= get_bit(g_scalar, i);
+				}
+			/* select the point to add, in constant time */
+			select_point(bits, 16, g_pre_comp, tmp);
+			if (!skip)
+				{
+				point_add(nq[0], nq[1], nq[2],
+					nq[0], nq[1], nq[2],
+					1 /* mixed */, tmp[0], tmp[1], tmp[2]);
+				}
+			else
+				{
+				memcpy(nq, tmp, 3 * sizeof(felem));
+				skip = 0;
+				}
+			}
+
+		/* do other additions every 5 doublings */
+		if (num_points && (i % 5 == 0))
+			{
+			/* loop over all scalars */
+			for (num = 0; num < num_points; ++num)
+				{
+				bits = get_bit(scalars[num], i + 4) << 5;
+				bits |= get_bit(scalars[num], i + 3) << 4;
+				bits |= get_bit(scalars[num], i + 2) << 3;
+				bits |= get_bit(scalars[num], i + 1) << 2;
+				bits |= get_bit(scalars[num], i) << 1;
+				bits |= get_bit(scalars[num], i - 1);
+				ec_GFp_nistp_recode_scalar_bits(&sign, &digit, bits);
+
+				/* select the point to add or subtract, in constant time */
+				select_point(digit, 17, pre_comp[num], tmp);
+				felem_neg(tmp[3], tmp[1]); /* (X, -Y, Z) is the negative point */
+				copy_conditional(tmp[1], tmp[3], (-(limb) sign));
+
+				if (!skip)
+					{
+					point_add(nq[0], nq[1], nq[2],
+						nq[0], nq[1], nq[2],
+						mixed, tmp[0], tmp[1], tmp[2]);
+					}
+				else
+					{
+					memcpy(nq, tmp, 3 * sizeof(felem));
+					skip = 0;
+					}
+				}
+			}
+		}
+	felem_assign(x_out, nq[0]);
+	felem_assign(y_out, nq[1]);
+	felem_assign(z_out, nq[2]);
+	}
+
+
+/* Precomputation for the group generator. */
+typedef struct {
+	felem g_pre_comp[16][3];
+	int references;
+} NISTP521_PRE_COMP;
+
+const EC_METHOD *EC_GFp_nistp521_method(void)
+	{
+	static const EC_METHOD ret = {
+		EC_FLAGS_DEFAULT_OCT,
+		NID_X9_62_prime_field,
+		ec_GFp_nistp521_group_init,
+		ec_GFp_simple_group_finish,
+		ec_GFp_simple_group_clear_finish,
+		ec_GFp_nist_group_copy,
+		ec_GFp_nistp521_group_set_curve,
+		ec_GFp_simple_group_get_curve,
+		ec_GFp_simple_group_get_degree,
+		ec_GFp_simple_group_check_discriminant,
+		ec_GFp_simple_point_init,
+		ec_GFp_simple_point_finish,
+		ec_GFp_simple_point_clear_finish,
+		ec_GFp_simple_point_copy,
+		ec_GFp_simple_point_set_to_infinity,
+		ec_GFp_simple_set_Jprojective_coordinates_GFp,
+		ec_GFp_simple_get_Jprojective_coordinates_GFp,
+		ec_GFp_simple_point_set_affine_coordinates,
+		ec_GFp_nistp521_point_get_affine_coordinates,
+		0 /* point_set_compressed_coordinates */,
+		0 /* point2oct */,
+		0 /* oct2point */,
+		ec_GFp_simple_add,
+		ec_GFp_simple_dbl,
+		ec_GFp_simple_invert,
+		ec_GFp_simple_is_at_infinity,
+		ec_GFp_simple_is_on_curve,
+		ec_GFp_simple_cmp,
+		ec_GFp_simple_make_affine,
+		ec_GFp_simple_points_make_affine,
+		ec_GFp_nistp521_points_mul,
+		ec_GFp_nistp521_precompute_mult,
+		ec_GFp_nistp521_have_precompute_mult,
+		ec_GFp_nist_field_mul,
+		ec_GFp_nist_field_sqr,
+		0 /* field_div */,
+		0 /* field_encode */,
+		0 /* field_decode */,
+		0 /* field_set_to_one */ };
+
+	return &ret;
+	}
+
+
+/******************************************************************************/
+/*		       FUNCTIONS TO MANAGE PRECOMPUTATION
+ */
+
+static NISTP521_PRE_COMP *nistp521_pre_comp_new()
+	{
+	NISTP521_PRE_COMP *ret = NULL;
+	ret = (NISTP521_PRE_COMP *)OPENSSL_malloc(sizeof(NISTP521_PRE_COMP));
+	if (!ret)
+		{
+		ECerr(EC_F_NISTP521_PRE_COMP_NEW, ERR_R_MALLOC_FAILURE);
+		return ret;
+		}
+	memset(ret->g_pre_comp, 0, sizeof(ret->g_pre_comp));
+	ret->references = 1;
+	return ret;
+	}
+
+static void *nistp521_pre_comp_dup(void *src_)
+	{
+	NISTP521_PRE_COMP *src = src_;
+
+	/* no need to actually copy, these objects never change! */
+	CRYPTO_add(&src->references, 1, CRYPTO_LOCK_EC_PRE_COMP);
+
+	return src_;
+	}
+
+static void nistp521_pre_comp_free(void *pre_)
+	{
+	int i;
+	NISTP521_PRE_COMP *pre = pre_;
+
+	if (!pre)
+		return;
+
+	i = CRYPTO_add(&pre->references, -1, CRYPTO_LOCK_EC_PRE_COMP);
+	if (i > 0)
+		return;
+
+	OPENSSL_free(pre);
+	}
+
+static void nistp521_pre_comp_clear_free(void *pre_)
+	{
+	int i;
+	NISTP521_PRE_COMP *pre = pre_;
+
+	if (!pre)
+		return;
+
+	i = CRYPTO_add(&pre->references, -1, CRYPTO_LOCK_EC_PRE_COMP);
+	if (i > 0)
+		return;
+
+	OPENSSL_cleanse(pre, sizeof(*pre));
+	OPENSSL_free(pre);
+	}
+
+/******************************************************************************/
+/*			   OPENSSL EC_METHOD FUNCTIONS
+ */
+
+int ec_GFp_nistp521_group_init(EC_GROUP *group)
+	{
+	int ret;
+	ret = ec_GFp_simple_group_init(group);
+	group->a_is_minus3 = 1;
+	return ret;
+	}
+
+int ec_GFp_nistp521_group_set_curve(EC_GROUP *group, const BIGNUM *p,
+	const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx)
+	{
+	int ret = 0;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *curve_p, *curve_a, *curve_b;
+
+	if (ctx == NULL)
+		if ((ctx = new_ctx = BN_CTX_new()) == NULL) return 0;
+	BN_CTX_start(ctx);
+	if (((curve_p = BN_CTX_get(ctx)) == NULL) ||
+		((curve_a = BN_CTX_get(ctx)) == NULL) ||
+		((curve_b = BN_CTX_get(ctx)) == NULL)) goto err;
+	BN_bin2bn(nistp521_curve_params[0], sizeof(felem_bytearray), curve_p);
+	BN_bin2bn(nistp521_curve_params[1], sizeof(felem_bytearray), curve_a);
+	BN_bin2bn(nistp521_curve_params[2], sizeof(felem_bytearray), curve_b);
+	if ((BN_cmp(curve_p, p)) || (BN_cmp(curve_a, a)) ||
+		(BN_cmp(curve_b, b)))
+		{
+		ECerr(EC_F_EC_GFP_NISTP521_GROUP_SET_CURVE,
+			EC_R_WRONG_CURVE_PARAMETERS);
+		goto err;
+		}
+	group->field_mod_func = BN_nist_mod_521;
+	ret = ec_GFp_simple_group_set_curve(group, p, a, b, ctx);
+err:
+	BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return ret;
+	}
+
+/* Takes the Jacobian coordinates (X, Y, Z) of a point and returns
+ * (X', Y') = (X/Z^2, Y/Z^3) */
+int ec_GFp_nistp521_point_get_affine_coordinates(const EC_GROUP *group,
+	const EC_POINT *point, BIGNUM *x, BIGNUM *y, BN_CTX *ctx)
+	{
+	felem z1, z2, x_in, y_in, x_out, y_out;
+	largefelem tmp;
+
+	if (EC_POINT_is_at_infinity(group, point))
+		{
+		ECerr(EC_F_EC_GFP_NISTP521_POINT_GET_AFFINE_COORDINATES,
+			EC_R_POINT_AT_INFINITY);
+		return 0;
+		}
+	if ((!BN_to_felem(x_in, &point->X)) || (!BN_to_felem(y_in, &point->Y)) ||
+		(!BN_to_felem(z1, &point->Z))) return 0;
+	felem_inv(z2, z1);
+	felem_square(tmp, z2); felem_reduce(z1, tmp);
+	felem_mul(tmp, x_in, z1); felem_reduce(x_in, tmp);
+	felem_contract(x_out, x_in);
+	if (x != NULL)
+		{
+		if (!felem_to_BN(x, x_out))
+			{
+			ECerr(EC_F_EC_GFP_NISTP521_POINT_GET_AFFINE_COORDINATES, ERR_R_BN_LIB);
+			return 0;
+			}
+		}
+	felem_mul(tmp, z1, z2); felem_reduce(z1, tmp);
+	felem_mul(tmp, y_in, z1); felem_reduce(y_in, tmp);
+	felem_contract(y_out, y_in);
+	if (y != NULL)
+		{
+		if (!felem_to_BN(y, y_out))
+			{
+			ECerr(EC_F_EC_GFP_NISTP521_POINT_GET_AFFINE_COORDINATES, ERR_R_BN_LIB);
+			return 0;
+			}
+		}
+	return 1;
+	}
+
+static void make_points_affine(size_t num, felem points[/* num */][3], felem tmp_felems[/* num+1 */])
+	{
+	/* Runs in constant time, unless an input is the point at infinity
+	 * (which normally shouldn't happen). */
+	ec_GFp_nistp_points_make_affine_internal(
+		num,
+		points,
+		sizeof(felem),
+		tmp_felems,
+		(void (*)(void *)) felem_one,
+		(int (*)(const void *)) felem_is_zero_int,
+		(void (*)(void *, const void *)) felem_assign,
+		(void (*)(void *, const void *)) felem_square_reduce,
+		(void (*)(void *, const void *, const void *)) felem_mul_reduce,
+		(void (*)(void *, const void *)) felem_inv,
+		(void (*)(void *, const void *)) felem_contract);
+	}
+
+/* Computes scalar*generator + \sum scalars[i]*points[i], ignoring NULL values
+ * Result is stored in r (r can equal one of the inputs). */
+int ec_GFp_nistp521_points_mul(const EC_GROUP *group, EC_POINT *r,
+	const BIGNUM *scalar, size_t num, const EC_POINT *points[],
+	const BIGNUM *scalars[], BN_CTX *ctx)
+	{
+	int ret = 0;
+	int j;
+	int mixed = 0;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *x, *y, *z, *tmp_scalar;
+	felem_bytearray g_secret;
+	felem_bytearray *secrets = NULL;
+	felem (*pre_comp)[17][3] = NULL;
+	felem *tmp_felems = NULL;
+	felem_bytearray tmp;
+	unsigned i, num_bytes;
+	int have_pre_comp = 0;
+	size_t num_points = num;
+	felem x_in, y_in, z_in, x_out, y_out, z_out;
+	NISTP521_PRE_COMP *pre = NULL;
+	felem (*g_pre_comp)[3] = NULL;
+	EC_POINT *generator = NULL;
+	const EC_POINT *p = NULL;
+	const BIGNUM *p_scalar = NULL;
+
+	if (ctx == NULL)
+		if ((ctx = new_ctx = BN_CTX_new()) == NULL) return 0;
+	BN_CTX_start(ctx);
+	if (((x = BN_CTX_get(ctx)) == NULL) ||
+		((y = BN_CTX_get(ctx)) == NULL) ||
+		((z = BN_CTX_get(ctx)) == NULL) ||
+		((tmp_scalar = BN_CTX_get(ctx)) == NULL))
+		goto err;
+
+	if (scalar != NULL)
+		{
+		pre = EC_EX_DATA_get_data(group->extra_data,
+			nistp521_pre_comp_dup, nistp521_pre_comp_free,
+			nistp521_pre_comp_clear_free);
+		if (pre)
+			/* we have precomputation, try to use it */
+			g_pre_comp = &pre->g_pre_comp[0];
+		else
+			/* try to use the standard precomputation */
+			g_pre_comp = (felem (*)[3]) gmul;
+		generator = EC_POINT_new(group);
+		if (generator == NULL)
+			goto err;
+		/* get the generator from precomputation */
+		if (!felem_to_BN(x, g_pre_comp[1][0]) ||
+			!felem_to_BN(y, g_pre_comp[1][1]) ||
+			!felem_to_BN(z, g_pre_comp[1][2]))
+			{
+			ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_BN_LIB);
+			goto err;
+			}
+		if (!EC_POINT_set_Jprojective_coordinates_GFp(group,
+				generator, x, y, z, ctx))
+			goto err;
+		if (0 == EC_POINT_cmp(group, generator, group->generator, ctx))
+			/* precomputation matches generator */
+			have_pre_comp = 1;
+		else
+			/* we don't have valid precomputation:
+			 * treat the generator as a random point */
+			num_points++;
+		}
+
+	if (num_points > 0)
+		{
+		if (num_points >= 2)
+			{
+			/* unless we precompute multiples for just one point,
+			 * converting those into affine form is time well spent  */
+			mixed = 1;
+			}
+		secrets = OPENSSL_malloc(num_points * sizeof(felem_bytearray));
+		pre_comp = OPENSSL_malloc(num_points * 17 * 3 * sizeof(felem));
+		if (mixed)
+			tmp_felems = OPENSSL_malloc((num_points * 17 + 1) * sizeof(felem));
+		if ((secrets == NULL) || (pre_comp == NULL) || (mixed && (tmp_felems == NULL)))
+			{
+			ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_MALLOC_FAILURE);
+			goto err;
+			}
+
+		/* we treat NULL scalars as 0, and NULL points as points at infinity,
+		 * i.e., they contribute nothing to the linear combination */
+		memset(secrets, 0, num_points * sizeof(felem_bytearray));
+		memset(pre_comp, 0, num_points * 17 * 3 * sizeof(felem));
+		for (i = 0; i < num_points; ++i)
+			{
+			if (i == num)
+				/* we didn't have a valid precomputation, so we pick
+				 * the generator */
+				{
+				p = EC_GROUP_get0_generator(group);
+				p_scalar = scalar;
+				}
+			else
+				/* the i^th point */
+				{
+				p = points[i];
+				p_scalar = scalars[i];
+				}
+			if ((p_scalar != NULL) && (p != NULL))
+				{
+				/* reduce scalar to 0 <= scalar < 2^521 */
+				if ((BN_num_bits(p_scalar) > 521) || (BN_is_negative(p_scalar)))
+					{
+					/* this is an unusual input, and we don't guarantee
+					 * constant-timeness */
+					if (!BN_nnmod(tmp_scalar, p_scalar, &group->order, ctx))
+						{
+						ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_BN_LIB);
+						goto err;
+						}
+					num_bytes = BN_bn2bin(tmp_scalar, tmp);
+					}
+				else
+					num_bytes = BN_bn2bin(p_scalar, tmp);
+				flip_endian(secrets[i], tmp, num_bytes);
+				/* precompute multiples */
+				if ((!BN_to_felem(x_out, &p->X)) ||
+					(!BN_to_felem(y_out, &p->Y)) ||
+					(!BN_to_felem(z_out, &p->Z))) goto err;
+				memcpy(pre_comp[i][1][0], x_out, sizeof(felem));
+				memcpy(pre_comp[i][1][1], y_out, sizeof(felem));
+				memcpy(pre_comp[i][1][2], z_out, sizeof(felem));
+				for (j = 2; j <= 16; ++j)
+					{
+					if (j & 1)
+						{
+						point_add(
+							pre_comp[i][j][0], pre_comp[i][j][1], pre_comp[i][j][2],
+							pre_comp[i][1][0], pre_comp[i][1][1], pre_comp[i][1][2],
+							0, pre_comp[i][j-1][0], pre_comp[i][j-1][1], pre_comp[i][j-1][2]);
+						}
+					else
+						{
+						point_double(
+							pre_comp[i][j][0], pre_comp[i][j][1], pre_comp[i][j][2],
+							pre_comp[i][j/2][0], pre_comp[i][j/2][1], pre_comp[i][j/2][2]);
+						}
+					}
+				}
+			}
+		if (mixed)
+			make_points_affine(num_points * 17, pre_comp[0], tmp_felems);
+		}
+
+	/* the scalar for the generator */
+	if ((scalar != NULL) && (have_pre_comp))
+		{
+		memset(g_secret, 0, sizeof(g_secret));
+		/* reduce scalar to 0 <= scalar < 2^521 */
+		if ((BN_num_bits(scalar) > 521) || (BN_is_negative(scalar)))
+			{
+			/* this is an unusual input, and we don't guarantee
+			 * constant-timeness */
+			if (!BN_nnmod(tmp_scalar, scalar, &group->order, ctx))
+				{
+				ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_BN_LIB);
+				goto err;
+				}
+			num_bytes = BN_bn2bin(tmp_scalar, tmp);
+			}
+		else
+			num_bytes = BN_bn2bin(scalar, tmp);
+		flip_endian(g_secret, tmp, num_bytes);
+		/* do the multiplication with generator precomputation*/
+		batch_mul(x_out, y_out, z_out,
+			(const felem_bytearray (*)) secrets, num_points,
+			g_secret,
+			mixed, (const felem (*)[17][3]) pre_comp,
+			(const felem (*)[3]) g_pre_comp);
+		}
+	else
+		/* do the multiplication without generator precomputation */
+		batch_mul(x_out, y_out, z_out,
+			(const felem_bytearray (*)) secrets, num_points,
+			NULL, mixed, (const felem (*)[17][3]) pre_comp, NULL);
+	/* reduce the output to its unique minimal representation */
+	felem_contract(x_in, x_out);
+	felem_contract(y_in, y_out);
+	felem_contract(z_in, z_out);
+	if ((!felem_to_BN(x, x_in)) || (!felem_to_BN(y, y_in)) ||
+		(!felem_to_BN(z, z_in)))
+		{
+		ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_BN_LIB);
+		goto err;
+		}
+	ret = EC_POINT_set_Jprojective_coordinates_GFp(group, r, x, y, z, ctx);
+
+err:
+	BN_CTX_end(ctx);
+	if (generator != NULL)
+		EC_POINT_free(generator);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	if (secrets != NULL)
+		OPENSSL_free(secrets);
+	if (pre_comp != NULL)
+		OPENSSL_free(pre_comp);
+	if (tmp_felems != NULL)
+		OPENSSL_free(tmp_felems);
+	return ret;
+	}
+
+int ec_GFp_nistp521_precompute_mult(EC_GROUP *group, BN_CTX *ctx)
+	{
+	int ret = 0;
+	NISTP521_PRE_COMP *pre = NULL;
+	int i, j;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *x, *y;
+	EC_POINT *generator = NULL;
+	felem tmp_felems[16];
+
+	/* throw away old precomputation */
+	EC_EX_DATA_free_data(&group->extra_data, nistp521_pre_comp_dup,
+		nistp521_pre_comp_free, nistp521_pre_comp_clear_free);
+	if (ctx == NULL)
+		if ((ctx = new_ctx = BN_CTX_new()) == NULL) return 0;
+	BN_CTX_start(ctx);
+	if (((x = BN_CTX_get(ctx)) == NULL) ||
+		((y = BN_CTX_get(ctx)) == NULL))
+		goto err;
+	/* get the generator */
+	if (group->generator == NULL) goto err;
+	generator = EC_POINT_new(group);
+	if (generator == NULL)
+		goto err;
+	BN_bin2bn(nistp521_curve_params[3], sizeof (felem_bytearray), x);
+	BN_bin2bn(nistp521_curve_params[4], sizeof (felem_bytearray), y);
+	if (!EC_POINT_set_affine_coordinates_GFp(group, generator, x, y, ctx))
+		goto err;
+	if ((pre = nistp521_pre_comp_new()) == NULL)
+		goto err;
+	/* if the generator is the standard one, use built-in precomputation */
+	if (0 == EC_POINT_cmp(group, generator, group->generator, ctx))
+		{
+		memcpy(pre->g_pre_comp, gmul, sizeof(pre->g_pre_comp));
+		ret = 1;
+		goto err;
+		}
+	if ((!BN_to_felem(pre->g_pre_comp[1][0], &group->generator->X)) ||
+		(!BN_to_felem(pre->g_pre_comp[1][1], &group->generator->Y)) ||
+		(!BN_to_felem(pre->g_pre_comp[1][2], &group->generator->Z)))
+		goto err;
+	/* compute 2^130*G, 2^260*G, 2^390*G */
+	for (i = 1; i <= 4; i <<= 1)
+		{
+		point_double(pre->g_pre_comp[2*i][0], pre->g_pre_comp[2*i][1],
+			pre->g_pre_comp[2*i][2], pre->g_pre_comp[i][0],
+			pre->g_pre_comp[i][1], pre->g_pre_comp[i][2]);
+		for (j = 0; j < 129; ++j)
+			{
+			point_double(pre->g_pre_comp[2*i][0],
+				pre->g_pre_comp[2*i][1],
+				pre->g_pre_comp[2*i][2],
+				pre->g_pre_comp[2*i][0],
+				pre->g_pre_comp[2*i][1],
+				pre->g_pre_comp[2*i][2]);
+			}
+		}
+	/* g_pre_comp[0] is the point at infinity */
+	memset(pre->g_pre_comp[0], 0, sizeof(pre->g_pre_comp[0]));
+	/* the remaining multiples */
+	/* 2^130*G + 2^260*G */
+	point_add(pre->g_pre_comp[6][0], pre->g_pre_comp[6][1],
+		pre->g_pre_comp[6][2], pre->g_pre_comp[4][0],
+		pre->g_pre_comp[4][1], pre->g_pre_comp[4][2],
+		0, pre->g_pre_comp[2][0], pre->g_pre_comp[2][1],
+		pre->g_pre_comp[2][2]);
+	/* 2^130*G + 2^390*G */
+	point_add(pre->g_pre_comp[10][0], pre->g_pre_comp[10][1],
+		pre->g_pre_comp[10][2], pre->g_pre_comp[8][0],
+		pre->g_pre_comp[8][1], pre->g_pre_comp[8][2],
+		0, pre->g_pre_comp[2][0], pre->g_pre_comp[2][1],
+		pre->g_pre_comp[2][2]);
+	/* 2^260*G + 2^390*G */
+	point_add(pre->g_pre_comp[12][0], pre->g_pre_comp[12][1],
+		pre->g_pre_comp[12][2], pre->g_pre_comp[8][0],
+		pre->g_pre_comp[8][1], pre->g_pre_comp[8][2],
+		0, pre->g_pre_comp[4][0], pre->g_pre_comp[4][1],
+		pre->g_pre_comp[4][2]);
+	/* 2^130*G + 2^260*G + 2^390*G */
+	point_add(pre->g_pre_comp[14][0], pre->g_pre_comp[14][1],
+		pre->g_pre_comp[14][2], pre->g_pre_comp[12][0],
+		pre->g_pre_comp[12][1], pre->g_pre_comp[12][2],
+		0, pre->g_pre_comp[2][0], pre->g_pre_comp[2][1],
+		pre->g_pre_comp[2][2]);
+	for (i = 1; i < 8; ++i)
+		{
+		/* odd multiples: add G */
+		point_add(pre->g_pre_comp[2*i+1][0], pre->g_pre_comp[2*i+1][1],
+			pre->g_pre_comp[2*i+1][2], pre->g_pre_comp[2*i][0],
+			pre->g_pre_comp[2*i][1], pre->g_pre_comp[2*i][2],
+			0, pre->g_pre_comp[1][0], pre->g_pre_comp[1][1],
+			pre->g_pre_comp[1][2]);
+		}
+	make_points_affine(15, &(pre->g_pre_comp[1]), tmp_felems);
+
+	if (!EC_EX_DATA_set_data(&group->extra_data, pre, nistp521_pre_comp_dup,
+			nistp521_pre_comp_free, nistp521_pre_comp_clear_free))
+		goto err;
+	ret = 1;
+	pre = NULL;
+ err:
+	BN_CTX_end(ctx);
+	if (generator != NULL)
+		EC_POINT_free(generator);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	if (pre)
+		nistp521_pre_comp_free(pre);
+	return ret;
+	}
+
+int ec_GFp_nistp521_have_precompute_mult(const EC_GROUP *group)
+	{
+	if (EC_EX_DATA_get_data(group->extra_data, nistp521_pre_comp_dup,
+			nistp521_pre_comp_free, nistp521_pre_comp_clear_free)
+		!= NULL)
+		return 1;
+	else
+		return 0;
+	}
+
+#else
+static void *dummy=&dummy;
+#endif
diff --git a/crypto/ec/ecp_nistputil.c b/crypto/ec/ecp_nistputil.c
new file mode 100644
index 000000000000..c8140c807fb0
--- /dev/null
+++ b/crypto/ec/ecp_nistputil.c
@@ -0,0 +1,197 @@
+/* crypto/ec/ecp_nistputil.c */
+/*
+ * Written by Bodo Moeller for the OpenSSL project.
+ */
+/* Copyright 2011 Google Inc.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ *
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ *  Unless required by applicable law or agreed to in writing, software
+ *  distributed under the License is distributed on an "AS IS" BASIS,
+ *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ *  See the License for the specific language governing permissions and
+ *  limitations under the License.
+ */
+
+#include <openssl/opensslconf.h>
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+
+/*
+ * Common utility functions for ecp_nistp224.c, ecp_nistp256.c, ecp_nistp521.c.
+ */
+
+#include <stddef.h>
+#include "ec_lcl.h"
+
+/* Convert an array of points into affine coordinates.
+ * (If the point at infinity is found (Z = 0), it remains unchanged.)
+ * This function is essentially an equivalent to EC_POINTs_make_affine(), but
+ * works with the internal representation of points as used by ecp_nistp###.c
+ * rather than with (BIGNUM-based) EC_POINT data structures.
+ *
+ * point_array is the input/output buffer ('num' points in projective form,
+ * i.e. three coordinates each), based on an internal representation of
+ * field elements of size 'felem_size'.
+ *
+ * tmp_felems needs to point to a temporary array of 'num'+1 field elements
+ * for storage of intermediate values.
+ */
+void ec_GFp_nistp_points_make_affine_internal(size_t num, void *point_array,
+	size_t felem_size, void *tmp_felems,
+	void (*felem_one)(void *out),
+	int (*felem_is_zero)(const void *in),
+	void (*felem_assign)(void *out, const void *in),
+	void (*felem_square)(void *out, const void *in),
+	void (*felem_mul)(void *out, const void *in1, const void *in2),
+	void (*felem_inv)(void *out, const void *in),
+	void (*felem_contract)(void *out, const void *in))
+	{
+	int i = 0;
+
+#define tmp_felem(I) (&((char *)tmp_felems)[(I) * felem_size])
+#define X(I) (&((char *)point_array)[3*(I) * felem_size])
+#define Y(I) (&((char *)point_array)[(3*(I) + 1) * felem_size])
+#define Z(I) (&((char *)point_array)[(3*(I) + 2) * felem_size])
+
+	if (!felem_is_zero(Z(0)))
+		felem_assign(tmp_felem(0), Z(0));
+	else
+		felem_one(tmp_felem(0));
+	for (i = 1; i < (int)num; i++)
+		{
+		if (!felem_is_zero(Z(i)))
+			felem_mul(tmp_felem(i), tmp_felem(i-1), Z(i));
+		else
+			felem_assign(tmp_felem(i), tmp_felem(i-1));
+		}
+	/* Now each tmp_felem(i) is the product of Z(0) .. Z(i), skipping any zero-valued factors:
+	 * if Z(i) = 0, we essentially pretend that Z(i) = 1 */
+
+	felem_inv(tmp_felem(num-1), tmp_felem(num-1));
+	for (i = num - 1; i >= 0; i--)
+		{
+		if (i > 0)
+			/* tmp_felem(i-1) is the product of Z(0) .. Z(i-1),
+			 * tmp_felem(i) is the inverse of the product of Z(0) .. Z(i)
+			 */
+			felem_mul(tmp_felem(num), tmp_felem(i-1), tmp_felem(i)); /* 1/Z(i) */
+		else
+			felem_assign(tmp_felem(num), tmp_felem(0)); /* 1/Z(0) */
+
+		if (!felem_is_zero(Z(i)))
+			{
+			if (i > 0)
+				/* For next iteration, replace tmp_felem(i-1) by its inverse */
+				felem_mul(tmp_felem(i-1), tmp_felem(i), Z(i));
+
+			/* Convert point (X, Y, Z) into affine form (X/(Z^2), Y/(Z^3), 1) */
+			felem_square(Z(i), tmp_felem(num)); /* 1/(Z^2) */
+			felem_mul(X(i), X(i), Z(i)); /* X/(Z^2) */
+			felem_mul(Z(i), Z(i), tmp_felem(num)); /* 1/(Z^3) */
+			felem_mul(Y(i), Y(i), Z(i)); /* Y/(Z^3) */
+			felem_contract(X(i), X(i));
+			felem_contract(Y(i), Y(i));
+			felem_one(Z(i));
+			}
+		else
+			{
+			if (i > 0)
+				/* For next iteration, replace tmp_felem(i-1) by its inverse */
+				felem_assign(tmp_felem(i-1), tmp_felem(i));
+			}
+		}
+	}
+
+/*
+ * This function looks at 5+1 scalar bits (5 current, 1 adjacent less
+ * significant bit), and recodes them into a signed digit for use in fast point
+ * multiplication: the use of signed rather than unsigned digits means that
+ * fewer points need to be precomputed, given that point inversion is easy
+ * (a precomputed point dP makes -dP available as well).
+ *
+ * BACKGROUND:
+ *
+ * Signed digits for multiplication were introduced by Booth ("A signed binary
+ * multiplication technique", Quart. Journ. Mech. and Applied Math., vol. IV,
+ * pt. 2 (1951), pp. 236-240), in that case for multiplication of integers.
+ * Booth's original encoding did not generally improve the density of nonzero
+ * digits over the binary representation, and was merely meant to simplify the
+ * handling of signed factors given in two's complement; but it has since been
+ * shown to be the basis of various signed-digit representations that do have
+ * further advantages, including the wNAF, using the following general approach:
+ *
+ * (1) Given a binary representation
+ *
+ *       b_k  ...  b_2  b_1  b_0,
+ *
+ *     of a nonnegative integer (b_k in {0, 1}), rewrite it in digits 0, 1, -1
+ *     by using bit-wise subtraction as follows:
+ *
+ *        b_k b_(k-1)  ...  b_2  b_1  b_0
+ *      -     b_k      ...  b_3  b_2  b_1  b_0
+ *       -------------------------------------
+ *        s_k b_(k-1)  ...  s_3  s_2  s_1  s_0
+ *
+ *     A left-shift followed by subtraction of the original value yields a new
+ *     representation of the same value, using signed bits s_i = b_(i+1) - b_i.
+ *     This representation from Booth's paper has since appeared in the
+ *     literature under a variety of different names including "reversed binary
+ *     form", "alternating greedy expansion", "mutual opposite form", and
+ *     "sign-alternating {+-1}-representation".
+ *
+ *     An interesting property is that among the nonzero bits, values 1 and -1
+ *     strictly alternate.
+ *
+ * (2) Various window schemes can be applied to the Booth representation of
+ *     integers: for example, right-to-left sliding windows yield the wNAF
+ *     (a signed-digit encoding independently discovered by various researchers
+ *     in the 1990s), and left-to-right sliding windows yield a left-to-right
+ *     equivalent of the wNAF (independently discovered by various researchers
+ *     around 2004).
+ *
+ * To prevent leaking information through side channels in point multiplication,
+ * we need to recode the given integer into a regular pattern: sliding windows
+ * as in wNAFs won't do, we need their fixed-window equivalent -- which is a few
+ * decades older: we'll be using the so-called "modified Booth encoding" due to
+ * MacSorley ("High-speed arithmetic in binary computers", Proc. IRE, vol. 49
+ * (1961), pp. 67-91), in a radix-2^5 setting.  That is, we always combine five
+ * signed bits into a signed digit:
+ *
+ *       s_(4j + 4) s_(4j + 3) s_(4j + 2) s_(4j + 1) s_(4j)
+ *
+ * The sign-alternating property implies that the resulting digit values are
+ * integers from -16 to 16.
+ *
+ * Of course, we don't actually need to compute the signed digits s_i as an
+ * intermediate step (that's just a nice way to see how this scheme relates
+ * to the wNAF): a direct computation obtains the recoded digit from the
+ * six bits b_(4j + 4) ... b_(4j - 1).
+ *
+ * This function takes those five bits as an integer (0 .. 63), writing the
+ * recoded digit to *sign (0 for positive, 1 for negative) and *digit (absolute
+ * value, in the range 0 .. 8).  Note that this integer essentially provides the
+ * input bits "shifted to the left" by one position: for example, the input to
+ * compute the least significant recoded digit, given that there's no bit b_-1,
+ * has to be b_4 b_3 b_2 b_1 b_0 0.
+ *
+ */
+void ec_GFp_nistp_recode_scalar_bits(unsigned char *sign, unsigned char *digit, unsigned char in)
+	{
+	unsigned char s, d;
+
+	s = ~((in >> 5) - 1); /* sets all bits to MSB(in), 'in' seen as 6-bit value */
+	d = (1 << 6) - in - 1;
+	d = (d & s) | (in & ~s);
+	d = (d >> 1) + (d & 1);
+
+	*sign = s & 1;
+	*digit = d;
+	}
+#else
+static void *dummy=&dummy;
+#endif
diff --git a/crypto/ec/ecp_oct.c b/crypto/ec/ecp_oct.c
new file mode 100644
index 000000000000..374a0ee731ae
--- /dev/null
+++ b/crypto/ec/ecp_oct.c
@@ -0,0 +1,433 @@
+/* crypto/ec/ecp_oct.c */
+/* Includes code written by Lenka Fibikova <fibikova@exp-math.uni-essen.de>
+ * for the OpenSSL project. 
+ * Includes code written by Bodo Moeller for the OpenSSL project.
+*/
+/* ====================================================================
+ * Copyright (c) 1998-2002 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+/* ====================================================================
+ * Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
+ * Portions of this software developed by SUN MICROSYSTEMS, INC.,
+ * and contributed to the OpenSSL project.
+ */
+
+#include <openssl/err.h>
+#include <openssl/symhacks.h>
+
+#include "ec_lcl.h"
+
+int ec_GFp_simple_set_compressed_coordinates(const EC_GROUP *group, EC_POINT *point,
+	const BIGNUM *x_, int y_bit, BN_CTX *ctx)
+	{
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *tmp1, *tmp2, *x, *y;
+	int ret = 0;
+
+	/* clear error queue*/
+	ERR_clear_error();
+
+	if (ctx == NULL)
+		{
+		ctx = new_ctx = BN_CTX_new();
+		if (ctx == NULL)
+			return 0;
+		}
+
+	y_bit = (y_bit != 0);
+
+	BN_CTX_start(ctx);
+	tmp1 = BN_CTX_get(ctx);
+	tmp2 = BN_CTX_get(ctx);
+	x = BN_CTX_get(ctx);
+	y = BN_CTX_get(ctx);
+	if (y == NULL) goto err;
+
+	/* Recover y.  We have a Weierstrass equation
+	 *     y^2 = x^3 + a*x + b,
+	 * so  y  is one of the square roots of  x^3 + a*x + b.
+	 */
+
+	/* tmp1 := x^3 */
+	if (!BN_nnmod(x, x_, &group->field,ctx)) goto err;
+	if (group->meth->field_decode == 0)
+		{
+		/* field_{sqr,mul} work on standard representation */
+		if (!group->meth->field_sqr(group, tmp2, x_, ctx)) goto err;
+		if (!group->meth->field_mul(group, tmp1, tmp2, x_, ctx)) goto err;
+		}
+	else
+		{
+		if (!BN_mod_sqr(tmp2, x_, &group->field, ctx)) goto err;
+		if (!BN_mod_mul(tmp1, tmp2, x_, &group->field, ctx)) goto err;
+		}
+	
+	/* tmp1 := tmp1 + a*x */
+	if (group->a_is_minus3)
+		{
+		if (!BN_mod_lshift1_quick(tmp2, x, &group->field)) goto err;
+		if (!BN_mod_add_quick(tmp2, tmp2, x, &group->field)) goto err;
+		if (!BN_mod_sub_quick(tmp1, tmp1, tmp2, &group->field)) goto err;
+		}
+	else
+		{
+		if (group->meth->field_decode)
+			{
+			if (!group->meth->field_decode(group, tmp2, &group->a, ctx)) goto err;
+			if (!BN_mod_mul(tmp2, tmp2, x, &group->field, ctx)) goto err;
+			}
+		else
+			{
+			/* field_mul works on standard representation */
+			if (!group->meth->field_mul(group, tmp2, &group->a, x, ctx)) goto err;
+			}
+		
+		if (!BN_mod_add_quick(tmp1, tmp1, tmp2, &group->field)) goto err;
+		}
+	
+	/* tmp1 := tmp1 + b */
+	if (group->meth->field_decode)
+		{
+		if (!group->meth->field_decode(group, tmp2, &group->b, ctx)) goto err;
+		if (!BN_mod_add_quick(tmp1, tmp1, tmp2, &group->field)) goto err;
+		}
+	else
+		{
+		if (!BN_mod_add_quick(tmp1, tmp1, &group->b, &group->field)) goto err;
+		}
+	
+	if (!BN_mod_sqrt(y, tmp1, &group->field, ctx))
+		{
+		unsigned long err = ERR_peek_last_error();
+		
+		if (ERR_GET_LIB(err) == ERR_LIB_BN && ERR_GET_REASON(err) == BN_R_NOT_A_SQUARE)
+			{
+			ERR_clear_error();
+			ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, EC_R_INVALID_COMPRESSED_POINT);
+			}
+		else
+			ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, ERR_R_BN_LIB);
+		goto err;
+		}
+
+	if (y_bit != BN_is_odd(y))
+		{
+		if (BN_is_zero(y))
+			{
+			int kron;
+
+			kron = BN_kronecker(x, &group->field, ctx);
+			if (kron == -2) goto err;
+
+			if (kron == 1)
+				ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, EC_R_INVALID_COMPRESSION_BIT);
+			else
+				/* BN_mod_sqrt() should have cought this error (not a square) */
+				ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, EC_R_INVALID_COMPRESSED_POINT);
+			goto err;
+			}
+		if (!BN_usub(y, &group->field, y)) goto err;
+		}
+	if (y_bit != BN_is_odd(y))
+		{
+		ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, ERR_R_INTERNAL_ERROR);
+		goto err;
+		}
+
+	if (!EC_POINT_set_affine_coordinates_GFp(group, point, x, y, ctx)) goto err;
+
+	ret = 1;
+
+ err:
+	BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return ret;
+	}
+
+
+size_t ec_GFp_simple_point2oct(const EC_GROUP *group, const EC_POINT *point, point_conversion_form_t form,
+	unsigned char *buf, size_t len, BN_CTX *ctx)
+	{
+	size_t ret;
+	BN_CTX *new_ctx = NULL;
+	int used_ctx = 0;
+	BIGNUM *x, *y;
+	size_t field_len, i, skip;
+
+	if ((form != POINT_CONVERSION_COMPRESSED)
+		&& (form != POINT_CONVERSION_UNCOMPRESSED)
+		&& (form != POINT_CONVERSION_HYBRID))
+		{
+		ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, EC_R_INVALID_FORM);
+		goto err;
+		}
+
+	if (EC_POINT_is_at_infinity(group, point))
+		{
+		/* encodes to a single 0 octet */
+		if (buf != NULL)
+			{
+			if (len < 1)
+				{
+				ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, EC_R_BUFFER_TOO_SMALL);
+				return 0;
+				}
+			buf[0] = 0;
+			}
+		return 1;
+		}
+
+
+	/* ret := required output buffer length */
+	field_len = BN_num_bytes(&group->field);
+	ret = (form == POINT_CONVERSION_COMPRESSED) ? 1 + field_len : 1 + 2*field_len;
+
+	/* if 'buf' is NULL, just return required length */
+	if (buf != NULL)
+		{
+		if (len < ret)
+			{
+			ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, EC_R_BUFFER_TOO_SMALL);
+			goto err;
+			}
+
+		if (ctx == NULL)
+			{
+			ctx = new_ctx = BN_CTX_new();
+			if (ctx == NULL)
+				return 0;
+			}
+
+		BN_CTX_start(ctx);
+		used_ctx = 1;
+		x = BN_CTX_get(ctx);
+		y = BN_CTX_get(ctx);
+		if (y == NULL) goto err;
+
+		if (!EC_POINT_get_affine_coordinates_GFp(group, point, x, y, ctx)) goto err;
+
+		if ((form == POINT_CONVERSION_COMPRESSED || form == POINT_CONVERSION_HYBRID) && BN_is_odd(y))
+			buf[0] = form + 1;
+		else
+			buf[0] = form;
+	
+		i = 1;
+		
+		skip = field_len - BN_num_bytes(x);
+		if (skip > field_len)
+			{
+			ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
+			goto err;
+			}
+		while (skip > 0)
+			{
+			buf[i++] = 0;
+			skip--;
+			}
+		skip = BN_bn2bin(x, buf + i);
+		i += skip;
+		if (i != 1 + field_len)
+			{
+			ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
+			goto err;
+			}
+
+		if (form == POINT_CONVERSION_UNCOMPRESSED || form == POINT_CONVERSION_HYBRID)
+			{
+			skip = field_len - BN_num_bytes(y);
+			if (skip > field_len)
+				{
+				ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
+				goto err;
+				}
+			while (skip > 0)
+				{
+				buf[i++] = 0;
+				skip--;
+				}
+			skip = BN_bn2bin(y, buf + i);
+			i += skip;
+			}
+
+		if (i != ret)
+			{
+			ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
+			goto err;
+			}
+		}
+	
+	if (used_ctx)
+		BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return ret;
+
+ err:
+	if (used_ctx)
+		BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return 0;
+	}
+
+
+int ec_GFp_simple_oct2point(const EC_GROUP *group, EC_POINT *point,
+	const unsigned char *buf, size_t len, BN_CTX *ctx)
+	{
+	point_conversion_form_t form;
+	int y_bit;
+	BN_CTX *new_ctx = NULL;
+	BIGNUM *x, *y;
+	size_t field_len, enc_len;
+	int ret = 0;
+
+	if (len == 0)
+		{
+		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_BUFFER_TOO_SMALL);
+		return 0;
+		}
+	form = buf[0];
+	y_bit = form & 1;
+	form = form & ~1U;
+	if ((form != 0)	&& (form != POINT_CONVERSION_COMPRESSED)
+		&& (form != POINT_CONVERSION_UNCOMPRESSED)
+		&& (form != POINT_CONVERSION_HYBRID))
+		{
+		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+		return 0;
+		}
+	if ((form == 0 || form == POINT_CONVERSION_UNCOMPRESSED) && y_bit)
+		{
+		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+		return 0;
+		}
+
+	if (form == 0)
+		{
+		if (len != 1)
+			{
+			ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+			return 0;
+			}
+
+		return EC_POINT_set_to_infinity(group, point);
+		}
+	
+	field_len = BN_num_bytes(&group->field);
+	enc_len = (form == POINT_CONVERSION_COMPRESSED) ? 1 + field_len : 1 + 2*field_len;
+
+	if (len != enc_len)
+		{
+		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+		return 0;
+		}
+
+	if (ctx == NULL)
+		{
+		ctx = new_ctx = BN_CTX_new();
+		if (ctx == NULL)
+			return 0;
+		}
+
+	BN_CTX_start(ctx);
+	x = BN_CTX_get(ctx);
+	y = BN_CTX_get(ctx);
+	if (y == NULL) goto err;
+
+	if (!BN_bin2bn(buf + 1, field_len, x)) goto err;
+	if (BN_ucmp(x, &group->field) >= 0)
+		{
+		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+		goto err;
+		}
+
+	if (form == POINT_CONVERSION_COMPRESSED)
+		{
+		if (!EC_POINT_set_compressed_coordinates_GFp(group, point, x, y_bit, ctx)) goto err;
+		}
+	else
+		{
+		if (!BN_bin2bn(buf + 1 + field_len, field_len, y)) goto err;
+		if (BN_ucmp(y, &group->field) >= 0)
+			{
+			ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+			goto err;
+			}
+		if (form == POINT_CONVERSION_HYBRID)
+			{
+			if (y_bit != BN_is_odd(y))
+				{
+				ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
+				goto err;
+				}
+			}
+
+		if (!EC_POINT_set_affine_coordinates_GFp(group, point, x, y, ctx)) goto err;
+		}
+	
+	if (!EC_POINT_is_on_curve(group, point, ctx)) /* test required by X9.62 */
+		{
+		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_POINT_IS_NOT_ON_CURVE);
+		goto err;
+		}
+
+	ret = 1;
+	
+ err:
+	BN_CTX_end(ctx);
+	if (new_ctx != NULL)
+		BN_CTX_free(new_ctx);
+	return ret;
+	}
+
diff --git a/crypto/ec/ecp_smpl.c b/crypto/ec/ecp_smpl.c
index 66a92e2a9005..7cbb321f9aae 100644
--- a/crypto/ec/ecp_smpl.c
+++ b/crypto/ec/ecp_smpl.c
@@ -65,11 +65,19 @@
 #include <openssl/err.h>
 #include <openssl/symhacks.h>
 
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
+
 #include "ec_lcl.h"
 
 const EC_METHOD *EC_GFp_simple_method(void)
 	{
+#ifdef OPENSSL_FIPS
+	return fips_ec_gfp_simple_method();
+#else
 	static const EC_METHOD ret = {
+		EC_FLAGS_DEFAULT_OCT,
 		NID_X9_62_prime_field,
 		ec_GFp_simple_group_init,
 		ec_GFp_simple_group_finish,
@@ -88,9 +96,7 @@ const EC_METHOD *EC_GFp_simple_method(void)
 		ec_GFp_simple_get_Jprojective_coordinates_GFp,
 		ec_GFp_simple_point_set_affine_coordinates,
 		ec_GFp_simple_point_get_affine_coordinates,
-		ec_GFp_simple_set_compressed_coordinates,
-		ec_GFp_simple_point2oct,
-		ec_GFp_simple_oct2point,
+		0,0,0,
 		ec_GFp_simple_add,
 		ec_GFp_simple_dbl,
 		ec_GFp_simple_invert,
@@ -110,6 +116,7 @@ const EC_METHOD *EC_GFp_simple_method(void)
 		0 /* field_set_to_one */ };
 
 	return &ret;
+#endif
 	}
 
 
@@ -633,372 +640,6 @@ int ec_GFp_simple_point_get_affine_coordinates(const EC_GROUP *group, const EC_P
 	return ret;
 	}
 
-
-int ec_GFp_simple_set_compressed_coordinates(const EC_GROUP *group, EC_POINT *point,
-	const BIGNUM *x_, int y_bit, BN_CTX *ctx)
-	{
-	BN_CTX *new_ctx = NULL;
-	BIGNUM *tmp1, *tmp2, *x, *y;
-	int ret = 0;
-
-	/* clear error queue*/
-	ERR_clear_error();
-
-	if (ctx == NULL)
-		{
-		ctx = new_ctx = BN_CTX_new();
-		if (ctx == NULL)
-			return 0;
-		}
-
-	y_bit = (y_bit != 0);
-
-	BN_CTX_start(ctx);
-	tmp1 = BN_CTX_get(ctx);
-	tmp2 = BN_CTX_get(ctx);
-	x = BN_CTX_get(ctx);
-	y = BN_CTX_get(ctx);
-	if (y == NULL) goto err;
-
-	/* Recover y.  We have a Weierstrass equation
-	 *     y^2 = x^3 + a*x + b,
-	 * so  y  is one of the square roots of  x^3 + a*x + b.
-	 */
-
-	/* tmp1 := x^3 */
-	if (!BN_nnmod(x, x_, &group->field,ctx)) goto err;
-	if (group->meth->field_decode == 0)
-		{
-		/* field_{sqr,mul} work on standard representation */
-		if (!group->meth->field_sqr(group, tmp2, x_, ctx)) goto err;
-		if (!group->meth->field_mul(group, tmp1, tmp2, x_, ctx)) goto err;
-		}
-	else
-		{
-		if (!BN_mod_sqr(tmp2, x_, &group->field, ctx)) goto err;
-		if (!BN_mod_mul(tmp1, tmp2, x_, &group->field, ctx)) goto err;
-		}
-	
-	/* tmp1 := tmp1 + a*x */
-	if (group->a_is_minus3)
-		{
-		if (!BN_mod_lshift1_quick(tmp2, x, &group->field)) goto err;
-		if (!BN_mod_add_quick(tmp2, tmp2, x, &group->field)) goto err;
-		if (!BN_mod_sub_quick(tmp1, tmp1, tmp2, &group->field)) goto err;
-		}
-	else
-		{
-		if (group->meth->field_decode)
-			{
-			if (!group->meth->field_decode(group, tmp2, &group->a, ctx)) goto err;
-			if (!BN_mod_mul(tmp2, tmp2, x, &group->field, ctx)) goto err;
-			}
-		else
-			{
-			/* field_mul works on standard representation */
-			if (!group->meth->field_mul(group, tmp2, &group->a, x, ctx)) goto err;
-			}
-		
-		if (!BN_mod_add_quick(tmp1, tmp1, tmp2, &group->field)) goto err;
-		}
-	
-	/* tmp1 := tmp1 + b */
-	if (group->meth->field_decode)
-		{
-		if (!group->meth->field_decode(group, tmp2, &group->b, ctx)) goto err;
-		if (!BN_mod_add_quick(tmp1, tmp1, tmp2, &group->field)) goto err;
-		}
-	else
-		{
-		if (!BN_mod_add_quick(tmp1, tmp1, &group->b, &group->field)) goto err;
-		}
-	
-	if (!BN_mod_sqrt(y, tmp1, &group->field, ctx))
-		{
-		unsigned long err = ERR_peek_last_error();
-		
-		if (ERR_GET_LIB(err) == ERR_LIB_BN && ERR_GET_REASON(err) == BN_R_NOT_A_SQUARE)
-			{
-			ERR_clear_error();
-			ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, EC_R_INVALID_COMPRESSED_POINT);
-			}
-		else
-			ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, ERR_R_BN_LIB);
-		goto err;
-		}
-
-	if (y_bit != BN_is_odd(y))
-		{
-		if (BN_is_zero(y))
-			{
-			int kron;
-
-			kron = BN_kronecker(x, &group->field, ctx);
-			if (kron == -2) goto err;
-
-			if (kron == 1)
-				ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, EC_R_INVALID_COMPRESSION_BIT);
-			else
-				/* BN_mod_sqrt() should have cought this error (not a square) */
-				ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, EC_R_INVALID_COMPRESSED_POINT);
-			goto err;
-			}
-		if (!BN_usub(y, &group->field, y)) goto err;
-		}
-	if (y_bit != BN_is_odd(y))
-		{
-		ECerr(EC_F_EC_GFP_SIMPLE_SET_COMPRESSED_COORDINATES, ERR_R_INTERNAL_ERROR);
-		goto err;
-		}
-
-	if (!EC_POINT_set_affine_coordinates_GFp(group, point, x, y, ctx)) goto err;
-
-	ret = 1;
-
- err:
-	BN_CTX_end(ctx);
-	if (new_ctx != NULL)
-		BN_CTX_free(new_ctx);
-	return ret;
-	}
-
-
-size_t ec_GFp_simple_point2oct(const EC_GROUP *group, const EC_POINT *point, point_conversion_form_t form,
-	unsigned char *buf, size_t len, BN_CTX *ctx)
-	{
-	size_t ret;
-	BN_CTX *new_ctx = NULL;
-	int used_ctx = 0;
-	BIGNUM *x, *y;
-	size_t field_len, i, skip;
-
-	if ((form != POINT_CONVERSION_COMPRESSED)
-		&& (form != POINT_CONVERSION_UNCOMPRESSED)
-		&& (form != POINT_CONVERSION_HYBRID))
-		{
-		ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, EC_R_INVALID_FORM);
-		goto err;
-		}
-
-	if (EC_POINT_is_at_infinity(group, point))
-		{
-		/* encodes to a single 0 octet */
-		if (buf != NULL)
-			{
-			if (len < 1)
-				{
-				ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, EC_R_BUFFER_TOO_SMALL);
-				return 0;
-				}
-			buf[0] = 0;
-			}
-		return 1;
-		}
-
-
-	/* ret := required output buffer length */
-	field_len = BN_num_bytes(&group->field);
-	ret = (form == POINT_CONVERSION_COMPRESSED) ? 1 + field_len : 1 + 2*field_len;
-
-	/* if 'buf' is NULL, just return required length */
-	if (buf != NULL)
-		{
-		if (len < ret)
-			{
-			ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, EC_R_BUFFER_TOO_SMALL);
-			goto err;
-			}
-
-		if (ctx == NULL)
-			{
-			ctx = new_ctx = BN_CTX_new();
-			if (ctx == NULL)
-				return 0;
-			}
-
-		BN_CTX_start(ctx);
-		used_ctx = 1;
-		x = BN_CTX_get(ctx);
-		y = BN_CTX_get(ctx);
-		if (y == NULL) goto err;
-
-		if (!EC_POINT_get_affine_coordinates_GFp(group, point, x, y, ctx)) goto err;
-
-		if ((form == POINT_CONVERSION_COMPRESSED || form == POINT_CONVERSION_HYBRID) && BN_is_odd(y))
-			buf[0] = form + 1;
-		else
-			buf[0] = form;
-	
-		i = 1;
-		
-		skip = field_len - BN_num_bytes(x);
-		if (skip > field_len)
-			{
-			ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
-			goto err;
-			}
-		while (skip > 0)
-			{
-			buf[i++] = 0;
-			skip--;
-			}
-		skip = BN_bn2bin(x, buf + i);
-		i += skip;
-		if (i != 1 + field_len)
-			{
-			ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
-			goto err;
-			}
-
-		if (form == POINT_CONVERSION_UNCOMPRESSED || form == POINT_CONVERSION_HYBRID)
-			{
-			skip = field_len - BN_num_bytes(y);
-			if (skip > field_len)
-				{
-				ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
-				goto err;
-				}
-			while (skip > 0)
-				{
-				buf[i++] = 0;
-				skip--;
-				}
-			skip = BN_bn2bin(y, buf + i);
-			i += skip;
-			}
-
-		if (i != ret)
-			{
-			ECerr(EC_F_EC_GFP_SIMPLE_POINT2OCT, ERR_R_INTERNAL_ERROR);
-			goto err;
-			}
-		}
-	
-	if (used_ctx)
-		BN_CTX_end(ctx);
-	if (new_ctx != NULL)
-		BN_CTX_free(new_ctx);
-	return ret;
-
- err:
-	if (used_ctx)
-		BN_CTX_end(ctx);
-	if (new_ctx != NULL)
-		BN_CTX_free(new_ctx);
-	return 0;
-	}
-
-
-int ec_GFp_simple_oct2point(const EC_GROUP *group, EC_POINT *point,
-	const unsigned char *buf, size_t len, BN_CTX *ctx)
-	{
-	point_conversion_form_t form;
-	int y_bit;
-	BN_CTX *new_ctx = NULL;
-	BIGNUM *x, *y;
-	size_t field_len, enc_len;
-	int ret = 0;
-
-	if (len == 0)
-		{
-		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_BUFFER_TOO_SMALL);
-		return 0;
-		}
-	form = buf[0];
-	y_bit = form & 1;
-	form = form & ~1U;
-	if ((form != 0)	&& (form != POINT_CONVERSION_COMPRESSED)
-		&& (form != POINT_CONVERSION_UNCOMPRESSED)
-		&& (form != POINT_CONVERSION_HYBRID))
-		{
-		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-		return 0;
-		}
-	if ((form == 0 || form == POINT_CONVERSION_UNCOMPRESSED) && y_bit)
-		{
-		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-		return 0;
-		}
-
-	if (form == 0)
-		{
-		if (len != 1)
-			{
-			ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-			return 0;
-			}
-
-		return EC_POINT_set_to_infinity(group, point);
-		}
-	
-	field_len = BN_num_bytes(&group->field);
-	enc_len = (form == POINT_CONVERSION_COMPRESSED) ? 1 + field_len : 1 + 2*field_len;
-
-	if (len != enc_len)
-		{
-		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-		return 0;
-		}
-
-	if (ctx == NULL)
-		{
-		ctx = new_ctx = BN_CTX_new();
-		if (ctx == NULL)
-			return 0;
-		}
-
-	BN_CTX_start(ctx);
-	x = BN_CTX_get(ctx);
-	y = BN_CTX_get(ctx);
-	if (y == NULL) goto err;
-
-	if (!BN_bin2bn(buf + 1, field_len, x)) goto err;
-	if (BN_ucmp(x, &group->field) >= 0)
-		{
-		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-		goto err;
-		}
-
-	if (form == POINT_CONVERSION_COMPRESSED)
-		{
-		if (!EC_POINT_set_compressed_coordinates_GFp(group, point, x, y_bit, ctx)) goto err;
-		}
-	else
-		{
-		if (!BN_bin2bn(buf + 1 + field_len, field_len, y)) goto err;
-		if (BN_ucmp(y, &group->field) >= 0)
-			{
-			ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-			goto err;
-			}
-		if (form == POINT_CONVERSION_HYBRID)
-			{
-			if (y_bit != BN_is_odd(y))
-				{
-				ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_INVALID_ENCODING);
-				goto err;
-				}
-			}
-
-		if (!EC_POINT_set_affine_coordinates_GFp(group, point, x, y, ctx)) goto err;
-		}
-	
-	if (!EC_POINT_is_on_curve(group, point, ctx)) /* test required by X9.62 */
-		{
-		ECerr(EC_F_EC_GFP_SIMPLE_OCT2POINT, EC_R_POINT_IS_NOT_ON_CURVE);
-		goto err;
-		}
-
-	ret = 1;
-	
- err:
-	BN_CTX_end(ctx);
-	if (new_ctx != NULL)
-		BN_CTX_free(new_ctx);
-	return ret;
-	}
-
-
 int ec_GFp_simple_add(const EC_GROUP *group, EC_POINT *r, const EC_POINT *a, const EC_POINT *b, BN_CTX *ctx)
 	{
 	int (*field_mul)(const EC_GROUP *, BIGNUM *, const BIGNUM *, const BIGNUM *, BN_CTX *);
diff --git a/crypto/ec/ectest.c b/crypto/ec/ectest.c
index 6148d553f9d5..f107782de006 100644
--- a/crypto/ec/ectest.c
+++ b/crypto/ec/ectest.c
@@ -94,6 +94,7 @@ int main(int argc, char * argv[]) { puts("Elliptic curves are disabled."); retur
 #include <openssl/objects.h>
 #include <openssl/rand.h>
 #include <openssl/bn.h>
+#include <openssl/opensslconf.h>
 
 #if defined(_MSC_VER) && defined(_MIPS_) && (_MSC_VER/100==12)
 /* suppress "too big too optimize" warning */
@@ -107,10 +108,6 @@ int main(int argc, char * argv[]) { puts("Elliptic curves are disabled."); retur
 	EXIT(1); \
 } while (0)
 
-void prime_field_tests(void);
-void char2_field_tests(void);
-void internal_curve_test(void);
-
 #define TIMING_BASE_PT 0
 #define TIMING_RAND_PT 1
 #define TIMING_SIMUL 2
@@ -195,7 +192,50 @@ static void timings(EC_GROUP *group, int type, BN_CTX *ctx)
 	}
 #endif
 
-void prime_field_tests()
+/* test multiplication with group order, long and negative scalars */
+static void group_order_tests(EC_GROUP *group)
+	{
+	BIGNUM *n1, *n2, *order;
+	EC_POINT *P = EC_POINT_new(group);
+	EC_POINT *Q = EC_POINT_new(group);
+	BN_CTX *ctx = BN_CTX_new();
+
+	n1 = BN_new(); n2 = BN_new(); order = BN_new();
+	fprintf(stdout, "verify group order ...");
+	fflush(stdout);
+	if (!EC_GROUP_get_order(group, order, ctx)) ABORT;
+	if (!EC_POINT_mul(group, Q, order, NULL, NULL, ctx)) ABORT;
+	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
+	fprintf(stdout, ".");
+	fflush(stdout);
+	if (!EC_GROUP_precompute_mult(group, ctx)) ABORT;
+	if (!EC_POINT_mul(group, Q, order, NULL, NULL, ctx)) ABORT;
+	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
+	fprintf(stdout, " ok\n");
+	fprintf(stdout, "long/negative scalar tests ... ");
+	if (!BN_one(n1)) ABORT;
+	/* n1 = 1 - order */
+	if (!BN_sub(n1, n1, order)) ABORT;
+	if(!EC_POINT_mul(group, Q, NULL, P, n1, ctx)) ABORT;
+	if (0 != EC_POINT_cmp(group, Q, P, ctx)) ABORT;
+	/* n2 = 1 + order */
+	if (!BN_add(n2, order, BN_value_one())) ABORT;
+	if(!EC_POINT_mul(group, Q, NULL, P, n2, ctx)) ABORT;
+	if (0 != EC_POINT_cmp(group, Q, P, ctx)) ABORT;
+	/* n2 = (1 - order) * (1 + order) */
+	if (!BN_mul(n2, n1, n2, ctx)) ABORT;
+	if(!EC_POINT_mul(group, Q, NULL, P, n2, ctx)) ABORT;
+	if (0 != EC_POINT_cmp(group, Q, P, ctx)) ABORT;
+	fprintf(stdout, "ok\n");
+	EC_POINT_free(P);
+	EC_POINT_free(Q);
+	BN_free(n1);
+	BN_free(n2);
+	BN_free(order);
+	BN_CTX_free(ctx);
+	}
+
+static void prime_field_tests(void)
 	{	
 	BN_CTX *ctx = NULL;
 	BIGNUM *p, *a, *b;
@@ -321,21 +361,21 @@ void prime_field_tests()
 	if (len == 0) ABORT;
 	if (!EC_POINT_oct2point(group, P, buf, len, ctx)) ABORT;
 	if (0 != EC_POINT_cmp(group, P, Q, ctx)) ABORT;
-	fprintf(stdout, "Generator as octect string, compressed form:\n     ");
+	fprintf(stdout, "Generator as octet string, compressed form:\n     ");
 	for (i = 0; i < len; i++) fprintf(stdout, "%02X", buf[i]);
 	
 	len = EC_POINT_point2oct(group, Q, POINT_CONVERSION_UNCOMPRESSED, buf, sizeof buf, ctx);
 	if (len == 0) ABORT;
 	if (!EC_POINT_oct2point(group, P, buf, len, ctx)) ABORT;
 	if (0 != EC_POINT_cmp(group, P, Q, ctx)) ABORT;
-	fprintf(stdout, "\nGenerator as octect string, uncompressed form:\n     ");
+	fprintf(stdout, "\nGenerator as octet string, uncompressed form:\n     ");
 	for (i = 0; i < len; i++) fprintf(stdout, "%02X", buf[i]);
 	
 	len = EC_POINT_point2oct(group, Q, POINT_CONVERSION_HYBRID, buf, sizeof buf, ctx);
 	if (len == 0) ABORT;
 	if (!EC_POINT_oct2point(group, P, buf, len, ctx)) ABORT;
 	if (0 != EC_POINT_cmp(group, P, Q, ctx)) ABORT;
-	fprintf(stdout, "\nGenerator as octect string, hybrid form:\n     ");
+	fprintf(stdout, "\nGenerator as octet string, hybrid form:\n     ");
 	for (i = 0; i < len; i++) fprintf(stdout, "%02X", buf[i]);
 	
 	if (!EC_POINT_get_Jprojective_coordinates_GFp(group, R, x, y, z, ctx)) ABORT;
@@ -381,17 +421,7 @@ void prime_field_tests()
 	if (EC_GROUP_get_degree(group) != 160) ABORT;
 	fprintf(stdout, " ok\n");
 	
-	fprintf(stdout, "verify group order ...");
-	fflush(stdout);
-	if (!EC_GROUP_get_order(group, z, ctx)) ABORT;
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, ".");
-	fflush(stdout);
-	if (!EC_GROUP_precompute_mult(group, ctx)) ABORT;
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, " ok\n");
+	group_order_tests(group);
 
 	if (!(P_160 = EC_GROUP_new(EC_GROUP_method_of(group)))) ABORT;
 	if (!EC_GROUP_copy(P_160, group)) ABORT;
@@ -425,19 +455,7 @@ void prime_field_tests()
 	if (EC_GROUP_get_degree(group) != 192) ABORT;
 	fprintf(stdout, " ok\n");
 	
-	fprintf(stdout, "verify group order ...");
-	fflush(stdout);
-	if (!EC_GROUP_get_order(group, z, ctx)) ABORT;
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, ".");
-	fflush(stdout);
-#if 0
-	if (!EC_GROUP_precompute_mult(group, ctx)) ABORT;
-#endif
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, " ok\n");
+	group_order_tests(group);
 
 	if (!(P_192 = EC_GROUP_new(EC_GROUP_method_of(group)))) ABORT;
 	if (!EC_GROUP_copy(P_192, group)) ABORT;
@@ -471,19 +489,7 @@ void prime_field_tests()
 	if (EC_GROUP_get_degree(group) != 224) ABORT;
 	fprintf(stdout, " ok\n");
 	
-	fprintf(stdout, "verify group order ...");
-	fflush(stdout);
-	if (!EC_GROUP_get_order(group, z, ctx)) ABORT;
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, ".");
-	fflush(stdout);
-#if 0
-	if (!EC_GROUP_precompute_mult(group, ctx)) ABORT;
-#endif
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, " ok\n");
+	group_order_tests(group);
 
 	if (!(P_224 = EC_GROUP_new(EC_GROUP_method_of(group)))) ABORT;
 	if (!EC_GROUP_copy(P_224, group)) ABORT;
@@ -518,19 +524,7 @@ void prime_field_tests()
 	if (EC_GROUP_get_degree(group) != 256) ABORT;
 	fprintf(stdout, " ok\n");
 	
-	fprintf(stdout, "verify group order ...");
-	fflush(stdout);
-	if (!EC_GROUP_get_order(group, z, ctx)) ABORT;
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, ".");
-	fflush(stdout);
-#if 0
-	if (!EC_GROUP_precompute_mult(group, ctx)) ABORT;
-#endif
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, " ok\n");
+	group_order_tests(group);
 
 	if (!(P_256 = EC_GROUP_new(EC_GROUP_method_of(group)))) ABORT;
 	if (!EC_GROUP_copy(P_256, group)) ABORT;
@@ -569,20 +563,8 @@ void prime_field_tests()
 	fprintf(stdout, "verify degree ...");
 	if (EC_GROUP_get_degree(group) != 384) ABORT;
 	fprintf(stdout, " ok\n");
-	
-	fprintf(stdout, "verify group order ...");
-	fflush(stdout);
-	if (!EC_GROUP_get_order(group, z, ctx)) ABORT;
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, ".");
-	fflush(stdout);
-#if 0
-	if (!EC_GROUP_precompute_mult(group, ctx)) ABORT;
-#endif
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, " ok\n");
+
+	group_order_tests(group);
 
 	if (!(P_384 = EC_GROUP_new(EC_GROUP_method_of(group)))) ABORT;
 	if (!EC_GROUP_copy(P_384, group)) ABORT;
@@ -627,20 +609,8 @@ void prime_field_tests()
 	fprintf(stdout, "verify degree ...");
 	if (EC_GROUP_get_degree(group) != 521) ABORT;
 	fprintf(stdout, " ok\n");
-	
-	fprintf(stdout, "verify group order ...");
-	fflush(stdout);
-	if (!EC_GROUP_get_order(group, z, ctx)) ABORT;
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, ".");
-	fflush(stdout);
-#if 0
-	if (!EC_GROUP_precompute_mult(group, ctx)) ABORT;
-#endif
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT;
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT;
-	fprintf(stdout, " ok\n");
+
+ 	group_order_tests(group);
 
 	if (!(P_521 = EC_GROUP_new(EC_GROUP_method_of(group)))) ABORT;
 	if (!EC_GROUP_copy(P_521, group)) ABORT;
@@ -669,6 +639,7 @@ void prime_field_tests()
 		points[2] = Q;
 		points[3] = Q;
 
+		if (!EC_GROUP_get_order(group, z, ctx)) ABORT;
 		if (!BN_add(y, z, BN_value_one())) ABORT;
 		if (BN_is_odd(y)) ABORT;
 		if (!BN_rshift1(y, y)) ABORT;
@@ -802,22 +773,14 @@ void prime_field_tests()
 	fprintf(stdout, "verify degree ..."); \
 	if (EC_GROUP_get_degree(group) != _degree) ABORT; \
 	fprintf(stdout, " ok\n"); \
-	fprintf(stdout, "verify group order ..."); \
-	fflush(stdout); \
-	if (!EC_GROUP_get_order(group, z, ctx)) ABORT; \
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT; \
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT; \
-	fprintf(stdout, "."); \
-	fflush(stdout); \
-	/* if (!EC_GROUP_precompute_mult(group, ctx)) ABORT; */ \
-	if (!EC_POINT_mul(group, Q, z, NULL, NULL, ctx)) ABORT; \
-	if (!EC_POINT_is_at_infinity(group, Q)) ABORT; \
-	fprintf(stdout, " ok\n"); \
+	group_order_tests(group); \
 	if (!(_variable = EC_GROUP_new(EC_GROUP_method_of(group)))) ABORT; \
-	if (!EC_GROUP_copy(_variable, group)) ABORT;
+	if (!EC_GROUP_copy(_variable, group)) ABORT; \
 
-void char2_field_tests()
-	{	
+#ifndef OPENSSL_NO_EC2M
+
+static void char2_field_tests(void)
+	{
 	BN_CTX *ctx = NULL;
 	BIGNUM *p, *a, *b;
 	EC_GROUP *group;
@@ -1249,8 +1212,9 @@ void char2_field_tests()
 	if (C2_B571) EC_GROUP_free(C2_B571);
 
 	}
+#endif
 
-void internal_curve_test(void)
+static void internal_curve_test(void)
 	{
 	EC_builtin_curve *curves = NULL;
 	size_t crv_len = 0, n = 0;
@@ -1297,13 +1261,189 @@ void internal_curve_test(void)
 		EC_GROUP_free(group);
 		}
 	if (ok)
-		fprintf(stdout, " ok\n");
+		fprintf(stdout, " ok\n\n");
 	else
-		fprintf(stdout, " failed\n");
+		{
+		fprintf(stdout, " failed\n\n");
+		ABORT;
+		}
 	OPENSSL_free(curves);
 	return;
 	}
 
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+/* nistp_test_params contains magic numbers for testing our optimized
+ * implementations of several NIST curves with characteristic > 3. */
+struct nistp_test_params
+	{
+	const EC_METHOD* (*meth) ();
+	int degree;
+	/* Qx, Qy and D are taken from
+	 * http://csrc.nist.gov/groups/ST/toolkit/documents/Examples/ECDSA_Prime.pdf
+	 * Otherwise, values are standard curve parameters from FIPS 180-3 */
+	const char *p, *a, *b, *Qx, *Qy, *Gx, *Gy, *order, *d;
+	};
+
+static const struct nistp_test_params nistp_tests_params[] =
+	{
+		{
+		/* P-224 */
+		EC_GFp_nistp224_method,
+		224,
+		"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF000000000000000000000001", /* p */
+		"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFEFFFFFFFFFFFFFFFFFFFFFFFE", /* a */
+		"B4050A850C04B3ABF54132565044B0B7D7BFD8BA270B39432355FFB4", /* b */
+		"E84FB0B8E7000CB657D7973CF6B42ED78B301674276DF744AF130B3E", /* Qx */
+		"4376675C6FC5612C21A0FF2D2A89D2987DF7A2BC52183B5982298555", /* Qy */
+		"B70E0CBD6BB4BF7F321390B94A03C1D356C21122343280D6115C1D21", /* Gx */
+		"BD376388B5F723FB4C22DFE6CD4375A05A07476444D5819985007E34", /* Gy */
+		"FFFFFFFFFFFFFFFFFFFFFFFFFFFF16A2E0B8F03E13DD29455C5C2A3D", /* order */
+		"3F0C488E987C80BE0FEE521F8D90BE6034EC69AE11CA72AA777481E8", /* d */
+		},
+		{
+		/* P-256 */
+		EC_GFp_nistp256_method,
+		256,
+		"ffffffff00000001000000000000000000000000ffffffffffffffffffffffff", /* p */
+		"ffffffff00000001000000000000000000000000fffffffffffffffffffffffc", /* a */
+		"5ac635d8aa3a93e7b3ebbd55769886bc651d06b0cc53b0f63bce3c3e27d2604b", /* b */
+		"b7e08afdfe94bad3f1dc8c734798ba1c62b3a0ad1e9ea2a38201cd0889bc7a19", /* Qx */
+		"3603f747959dbf7a4bb226e41928729063adc7ae43529e61b563bbc606cc5e09", /* Qy */
+		"6b17d1f2e12c4247f8bce6e563a440f277037d812deb33a0f4a13945d898c296", /* Gx */
+		"4fe342e2fe1a7f9b8ee7eb4a7c0f9e162bce33576b315ececbb6406837bf51f5", /* Gy */
+		"ffffffff00000000ffffffffffffffffbce6faada7179e84f3b9cac2fc632551", /* order */
+		"c477f9f65c22cce20657faa5b2d1d8122336f851a508a1ed04e479c34985bf96", /* d */
+		},
+		{
+		/* P-521 */
+		EC_GFp_nistp521_method,
+		521,
+		"1ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", /* p */
+		"1fffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffc", /* a */
+		"051953eb9618e1c9a1f929a21a0b68540eea2da725b99b315f3b8b489918ef109e156193951ec7e937b1652c0bd3bb1bf073573df883d2c34f1ef451fd46b503f00", /* b */
+		"0098e91eef9a68452822309c52fab453f5f117c1da8ed796b255e9ab8f6410cca16e59df403a6bdc6ca467a37056b1e54b3005d8ac030decfeb68df18b171885d5c4", /* Qx */
+		"0164350c321aecfc1cca1ba4364c9b15656150b4b78d6a48d7d28e7f31985ef17be8554376b72900712c4b83ad668327231526e313f5f092999a4632fd50d946bc2e", /* Qy */
+		"c6858e06b70404e9cd9e3ecb662395b4429c648139053fb521f828af606b4d3dbaa14b5e77efe75928fe1dc127a2ffa8de3348b3c1856a429bf97e7e31c2e5bd66", /* Gx */
+		"11839296a789a3bc0045c8a5fb42c7d1bd998f54449579b446817afbd17273e662c97ee72995ef42640c550b9013fad0761353c7086a272c24088be94769fd16650", /* Gy */
+		"1fffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffa51868783bf2f966b7fcc0148f709a5d03bb5c9b8899c47aebb6fb71e91386409", /* order */
+		"0100085f47b8e1b8b11b7eb33028c0b2888e304bfc98501955b45bba1478dc184eeedf09b86a5f7c21994406072787205e69a63709fe35aa93ba333514b24f961722", /* d */
+		},
+	};
+
+void nistp_single_test(const struct nistp_test_params *test)
+	{
+	BN_CTX *ctx;
+	BIGNUM *p, *a, *b, *x, *y, *n, *m, *order;
+	EC_GROUP *NISTP;
+	EC_POINT *G, *P, *Q, *Q_CHECK;
+
+	fprintf(stdout, "\nNIST curve P-%d (optimised implementation):\n", test->degree);
+	ctx = BN_CTX_new();
+	p = BN_new();
+	a = BN_new();
+	b = BN_new();
+	x = BN_new(); y = BN_new();
+	m = BN_new(); n = BN_new(); order = BN_new();
+
+	NISTP = EC_GROUP_new(test->meth());
+	if(!NISTP) ABORT;
+	if (!BN_hex2bn(&p, test->p)) ABORT;
+	if (1 != BN_is_prime_ex(p, BN_prime_checks, ctx, NULL)) ABORT;
+	if (!BN_hex2bn(&a, test->a)) ABORT;
+	if (!BN_hex2bn(&b, test->b)) ABORT;
+	if (!EC_GROUP_set_curve_GFp(NISTP, p, a, b, ctx)) ABORT;
+	G = EC_POINT_new(NISTP);
+	P = EC_POINT_new(NISTP);
+	Q = EC_POINT_new(NISTP);
+	Q_CHECK = EC_POINT_new(NISTP);
+	if(!BN_hex2bn(&x, test->Qx)) ABORT;
+	if(!BN_hex2bn(&y, test->Qy)) ABORT;
+	if(!EC_POINT_set_affine_coordinates_GFp(NISTP, Q_CHECK, x, y, ctx)) ABORT;
+	if (!BN_hex2bn(&x, test->Gx)) ABORT;
+	if (!BN_hex2bn(&y, test->Gy)) ABORT;
+	if (!EC_POINT_set_affine_coordinates_GFp(NISTP, G, x, y, ctx)) ABORT;
+	if (!BN_hex2bn(&order, test->order)) ABORT;
+	if (!EC_GROUP_set_generator(NISTP, G, order, BN_value_one())) ABORT;
+
+	fprintf(stdout, "verify degree ... ");
+	if (EC_GROUP_get_degree(NISTP) != test->degree) ABORT;
+	fprintf(stdout, "ok\n");
+
+	fprintf(stdout, "NIST test vectors ... ");
+	if (!BN_hex2bn(&n, test->d)) ABORT;
+	/* fixed point multiplication */
+	EC_POINT_mul(NISTP, Q, n, NULL, NULL, ctx);
+	if (0 != EC_POINT_cmp(NISTP, Q, Q_CHECK, ctx)) ABORT;
+	/* random point multiplication */
+	EC_POINT_mul(NISTP, Q, NULL, G, n, ctx);
+	if (0 != EC_POINT_cmp(NISTP, Q, Q_CHECK, ctx)) ABORT;
+
+	/* set generator to P = 2*G, where G is the standard generator */
+	if (!EC_POINT_dbl(NISTP, P, G, ctx)) ABORT;
+	if (!EC_GROUP_set_generator(NISTP, P, order, BN_value_one())) ABORT;
+	/* set the scalar to m=n/2, where n is the NIST test scalar */
+	if (!BN_rshift(m, n, 1)) ABORT;
+
+	/* test the non-standard generator */
+	/* fixed point multiplication */
+	EC_POINT_mul(NISTP, Q, m, NULL, NULL, ctx);
+	if (0 != EC_POINT_cmp(NISTP, Q, Q_CHECK, ctx)) ABORT;
+	/* random point multiplication */
+	EC_POINT_mul(NISTP, Q, NULL, P, m, ctx);
+	if (0 != EC_POINT_cmp(NISTP, Q, Q_CHECK, ctx)) ABORT;
+
+	/* now repeat all tests with precomputation */
+	if (!EC_GROUP_precompute_mult(NISTP, ctx)) ABORT;
+
+	/* fixed point multiplication */
+	EC_POINT_mul(NISTP, Q, m, NULL, NULL, ctx);
+	if (0 != EC_POINT_cmp(NISTP, Q, Q_CHECK, ctx)) ABORT;
+	/* random point multiplication */
+	EC_POINT_mul(NISTP, Q, NULL, P, m, ctx);
+	if (0 != EC_POINT_cmp(NISTP, Q, Q_CHECK, ctx)) ABORT;
+
+	/* reset generator */
+	if (!EC_GROUP_set_generator(NISTP, G, order, BN_value_one())) ABORT;
+	/* fixed point multiplication */
+	EC_POINT_mul(NISTP, Q, n, NULL, NULL, ctx);
+	if (0 != EC_POINT_cmp(NISTP, Q, Q_CHECK, ctx)) ABORT;
+	/* random point multiplication */
+	EC_POINT_mul(NISTP, Q, NULL, G, n, ctx);
+	if (0 != EC_POINT_cmp(NISTP, Q, Q_CHECK, ctx)) ABORT;
+
+	fprintf(stdout, "ok\n");
+	group_order_tests(NISTP);
+#if 0
+	timings(NISTP, TIMING_BASE_PT, ctx);
+	timings(NISTP, TIMING_RAND_PT, ctx);
+#endif
+	EC_GROUP_free(NISTP);
+	EC_POINT_free(G);
+	EC_POINT_free(P);
+	EC_POINT_free(Q);
+	EC_POINT_free(Q_CHECK);
+	BN_free(n);
+	BN_free(m);
+	BN_free(p);
+	BN_free(a);
+	BN_free(b);
+	BN_free(x);
+	BN_free(y);
+	BN_free(order);
+	BN_CTX_free(ctx);
+	}
+
+void nistp_tests()
+	{
+	unsigned i;
+
+	for (i = 0; i < sizeof(nistp_tests_params) / sizeof(struct nistp_test_params); i++)
+		{
+		nistp_single_test(&nistp_tests_params[i]);
+		}
+	}
+#endif
+
 static const char rnd_seed[] = "string to make the random number generator think it has entropy";
 
 int main(int argc, char *argv[])
@@ -1327,7 +1467,12 @@ int main(int argc, char *argv[])
 
 	prime_field_tests();
 	puts("");
+#ifndef OPENSSL_NO_EC2M
 	char2_field_tests();
+#endif
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+	nistp_tests();
+#endif
 	/* test the internal curves */
 	internal_curve_test();
 
@@ -1336,7 +1481,7 @@ int main(int argc, char *argv[])
 #endif
 	CRYPTO_cleanup_all_ex_data();
 	ERR_free_strings();
-	ERR_remove_state(0);
+	ERR_remove_thread_state(NULL);
 	CRYPTO_mem_leaks_fp(stderr);
 	
 	return 0;
diff --git a/crypto/ecdh/Makefile b/crypto/ecdh/Makefile
index 7a7b618eeb8e..65d8904ee8a6 100644
--- a/crypto/ecdh/Makefile
+++ b/crypto/ecdh/Makefile
@@ -34,7 +34,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -88,27 +88,26 @@ ech_key.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 ech_key.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 ech_key.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 ech_key.o: ../../include/openssl/engine.h ../../include/openssl/evp.h
-ech_key.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-ech_key.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-ech_key.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-ech_key.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-ech_key.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-ech_key.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-ech_key.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-ech_key.o: ech_key.c ech_locl.h
+ech_key.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ech_key.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+ech_key.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ech_key.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+ech_key.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ech_key.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+ech_key.o: ../../include/openssl/x509_vfy.h ech_key.c ech_locl.h
 ech_lib.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 ech_lib.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 ech_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 ech_lib.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 ech_lib.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-ech_lib.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-ech_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-ech_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-ech_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-ech_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-ech_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-ech_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-ech_lib.o: ../../include/openssl/x509_vfy.h ech_lib.c ech_locl.h
+ech_lib.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ech_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ech_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+ech_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ech_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+ech_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+ech_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ech_lib.o: ech_lib.c ech_locl.h
 ech_ossl.o: ../../e_os.h ../../include/openssl/asn1.h
 ech_ossl.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 ech_ossl.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
diff --git a/crypto/ecdh/ecdh.h b/crypto/ecdh/ecdh.h
index b4b58ee65ba2..8887102c0b56 100644
--- a/crypto/ecdh/ecdh.h
+++ b/crypto/ecdh/ecdh.h
@@ -109,11 +109,13 @@ void ERR_load_ECDH_strings(void);
 /* Error codes for the ECDH functions. */
 
 /* Function codes. */
+#define ECDH_F_ECDH_CHECK				 102
 #define ECDH_F_ECDH_COMPUTE_KEY				 100
 #define ECDH_F_ECDH_DATA_NEW_METHOD			 101
 
 /* Reason codes. */
 #define ECDH_R_KDF_FAILED				 102
+#define ECDH_R_NON_FIPS_METHOD				 103
 #define ECDH_R_NO_PRIVATE_VALUE				 100
 #define ECDH_R_POINT_ARITHMETIC_FAILURE			 101
 
diff --git a/crypto/ecdh/ecdhtest.c b/crypto/ecdh/ecdhtest.c
index 1575006b5161..823d7baa657c 100644
--- a/crypto/ecdh/ecdhtest.c
+++ b/crypto/ecdh/ecdhtest.c
@@ -158,11 +158,13 @@ static int test_ecdh_curve(int nid, const char *text, BN_CTX *ctx, BIO *out)
 		if (!EC_POINT_get_affine_coordinates_GFp(group,
 			EC_KEY_get0_public_key(a), x_a, y_a, ctx)) goto err;
 		}
+#ifndef OPENSSL_NO_EC2M
 	else
 		{
 		if (!EC_POINT_get_affine_coordinates_GF2m(group,
 			EC_KEY_get0_public_key(a), x_a, y_a, ctx)) goto err;
 		}
+#endif
 #ifdef NOISY
 	BIO_puts(out,"  pri 1=");
 	BN_print(out,a->priv_key);
@@ -183,11 +185,13 @@ static int test_ecdh_curve(int nid, const char *text, BN_CTX *ctx, BIO *out)
 		if (!EC_POINT_get_affine_coordinates_GFp(group, 
 			EC_KEY_get0_public_key(b), x_b, y_b, ctx)) goto err;
 		}
+#ifndef OPENSSL_NO_EC2M
 	else
 		{
 		if (!EC_POINT_get_affine_coordinates_GF2m(group, 
 			EC_KEY_get0_public_key(b), x_b, y_b, ctx)) goto err;
 		}
+#endif
 
 #ifdef NOISY
 	BIO_puts(out,"  pri 2=");
@@ -324,6 +328,7 @@ int main(int argc, char *argv[])
 	if (!test_ecdh_curve(NID_X9_62_prime256v1, "NIST Prime-Curve P-256", ctx, out)) goto err;
 	if (!test_ecdh_curve(NID_secp384r1, "NIST Prime-Curve P-384", ctx, out)) goto err;
 	if (!test_ecdh_curve(NID_secp521r1, "NIST Prime-Curve P-521", ctx, out)) goto err;
+#ifndef OPENSSL_NO_EC2M
 	/* NIST BINARY CURVES TESTS */
 	if (!test_ecdh_curve(NID_sect163k1, "NIST Binary-Curve K-163", ctx, out)) goto err;
 	if (!test_ecdh_curve(NID_sect163r2, "NIST Binary-Curve B-163", ctx, out)) goto err;
@@ -335,6 +340,7 @@ int main(int argc, char *argv[])
 	if (!test_ecdh_curve(NID_sect409r1, "NIST Binary-Curve B-409", ctx, out)) goto err;
 	if (!test_ecdh_curve(NID_sect571k1, "NIST Binary-Curve K-571", ctx, out)) goto err;
 	if (!test_ecdh_curve(NID_sect571r1, "NIST Binary-Curve B-571", ctx, out)) goto err;
+#endif
 
 	ret = 0;
 
@@ -343,7 +349,7 @@ err:
 	if (ctx) BN_CTX_free(ctx);
 	BIO_free(out);
 	CRYPTO_cleanup_all_ex_data();
-	ERR_remove_state(0);
+	ERR_remove_thread_state(NULL);
 	CRYPTO_mem_leaks_fp(stderr);
 	EXIT(ret);
 	return(ret);
diff --git a/crypto/ecdh/ech_err.c b/crypto/ecdh/ech_err.c
index 4d2ede75bd37..3bd247398db5 100644
--- a/crypto/ecdh/ech_err.c
+++ b/crypto/ecdh/ech_err.c
@@ -1,6 +1,6 @@
 /* crypto/ecdh/ech_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,14 +70,16 @@
 
 static ERR_STRING_DATA ECDH_str_functs[]=
 	{
+{ERR_FUNC(ECDH_F_ECDH_CHECK),	"ECDH_CHECK"},
 {ERR_FUNC(ECDH_F_ECDH_COMPUTE_KEY),	"ECDH_compute_key"},
-{ERR_FUNC(ECDH_F_ECDH_DATA_NEW_METHOD),	"ECDH_DATA_NEW_METHOD"},
+{ERR_FUNC(ECDH_F_ECDH_DATA_NEW_METHOD),	"ECDH_DATA_new_method"},
 {0,NULL}
 	};
 
 static ERR_STRING_DATA ECDH_str_reasons[]=
 	{
 {ERR_REASON(ECDH_R_KDF_FAILED)           ,"KDF failed"},
+{ERR_REASON(ECDH_R_NON_FIPS_METHOD)      ,"non fips method"},
 {ERR_REASON(ECDH_R_NO_PRIVATE_VALUE)     ,"no private value"},
 {ERR_REASON(ECDH_R_POINT_ARITHMETIC_FAILURE),"point arithmetic failure"},
 {0,NULL}
diff --git a/crypto/ecdh/ech_lib.c b/crypto/ecdh/ech_lib.c
index bf2223477852..dadbfd3c49f5 100644
--- a/crypto/ecdh/ech_lib.c
+++ b/crypto/ecdh/ech_lib.c
@@ -73,6 +73,9 @@
 #include <openssl/engine.h>
 #endif
 #include <openssl/err.h>
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 
 const char ECDH_version[]="ECDH" OPENSSL_VERSION_PTEXT;
 
@@ -90,7 +93,16 @@ void ECDH_set_default_method(const ECDH_METHOD *meth)
 const ECDH_METHOD *ECDH_get_default_method(void)
 	{
 	if(!default_ECDH_method) 
+		{
+#ifdef OPENSSL_FIPS
+		if (FIPS_mode())
+			return FIPS_ecdh_openssl();
+		else
+			return ECDH_OpenSSL();
+#else
 		default_ECDH_method = ECDH_OpenSSL();
+#endif
+		}
 	return default_ECDH_method;
 	}
 
@@ -103,6 +115,11 @@ int ECDH_set_method(EC_KEY *eckey, const ECDH_METHOD *meth)
 	if (ecdh == NULL)
 		return 0;
 
+#if 0
+        mtmp = ecdh->meth;
+        if (mtmp->finish)
+		mtmp->finish(eckey);
+#endif
 #ifndef OPENSSL_NO_ENGINE
 	if (ecdh->engine)
 		{
@@ -210,6 +227,14 @@ ECDH_DATA *ecdh_check(EC_KEY *key)
 	}
 	else
 		ecdh_data = (ECDH_DATA *)data;
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(ecdh_data->flags & ECDH_FLAG_FIPS_METHOD)
+			&& !(EC_KEY_get_flags(key) & EC_FLAG_NON_FIPS_ALLOW))
+		{
+		ECDHerr(ECDH_F_ECDH_CHECK, ECDH_R_NON_FIPS_METHOD);
+		return NULL;
+		}
+#endif
 	
 
 	return ecdh_data;
diff --git a/crypto/ecdh/ech_locl.h b/crypto/ecdh/ech_locl.h
index f658526a7e3a..f6cad6a894bc 100644
--- a/crypto/ecdh/ech_locl.h
+++ b/crypto/ecdh/ech_locl.h
@@ -75,6 +75,14 @@ struct ecdh_method
 	char *app_data;
 	};
 
+/* If this flag is set the ECDH method is FIPS compliant and can be used
+ * in FIPS mode. This is set in the validated module method. If an
+ * application sets this flag in its own methods it is its responsibility
+ * to ensure the result is compliant.
+ */
+
+#define ECDH_FLAG_FIPS_METHOD	0x1
+
 typedef struct ecdh_data_st {
 	/* EC_KEY_METH_DATA part */
 	int (*init)(EC_KEY *);
diff --git a/crypto/ecdh/ech_ossl.c b/crypto/ecdh/ech_ossl.c
index 2a40ff12dfa8..4a30628fbcc9 100644
--- a/crypto/ecdh/ech_ossl.c
+++ b/crypto/ecdh/ech_ossl.c
@@ -157,6 +157,7 @@ static int ecdh_compute_key(void *out, size_t outlen, const EC_POINT *pub_key,
 			goto err;
 			}
 		}
+#ifndef OPENSSL_NO_EC2M
 	else
 		{
 		if (!EC_POINT_get_affine_coordinates_GF2m(group, tmp, x, y, ctx)) 
@@ -165,6 +166,7 @@ static int ecdh_compute_key(void *out, size_t outlen, const EC_POINT *pub_key,
 			goto err;
 			}
 		}
+#endif
 
 	buflen = (EC_GROUP_get_degree(group) + 7)/8;
 	len = BN_num_bytes(x);
diff --git a/crypto/ecdsa/Makefile b/crypto/ecdsa/Makefile
index 49e2681575a1..e89e0c010c6b 100644
--- a/crypto/ecdsa/Makefile
+++ b/crypto/ecdsa/Makefile
@@ -34,7 +34,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -97,14 +97,13 @@ ecs_lib.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 ecs_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 ecs_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 ecs_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-ecs_lib.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-ecs_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-ecs_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-ecs_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-ecs_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-ecs_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-ecs_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-ecs_lib.o: ecs_lib.c ecs_locl.h
+ecs_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ecs_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+ecs_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ecs_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+ecs_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ecs_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+ecs_lib.o: ../../include/openssl/x509_vfy.h ecs_lib.c ecs_locl.h
 ecs_ossl.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 ecs_ossl.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
 ecs_ossl.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
@@ -119,9 +118,8 @@ ecs_sign.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 ecs_sign.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 ecs_sign.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 ecs_sign.o: ../../include/openssl/engine.h ../../include/openssl/evp.h
-ecs_sign.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-ecs_sign.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-ecs_sign.o: ../../include/openssl/opensslconf.h
+ecs_sign.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ecs_sign.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 ecs_sign.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ecs_sign.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
 ecs_sign.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -133,11 +131,10 @@ ecs_vrf.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 ecs_vrf.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 ecs_vrf.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 ecs_vrf.o: ../../include/openssl/engine.h ../../include/openssl/evp.h
-ecs_vrf.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-ecs_vrf.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-ecs_vrf.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-ecs_vrf.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-ecs_vrf.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-ecs_vrf.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-ecs_vrf.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-ecs_vrf.o: ecs_locl.h ecs_vrf.c
+ecs_vrf.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ecs_vrf.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+ecs_vrf.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ecs_vrf.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+ecs_vrf.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ecs_vrf.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+ecs_vrf.o: ../../include/openssl/x509_vfy.h ecs_locl.h ecs_vrf.c
diff --git a/crypto/ecdsa/ecdsa.h b/crypto/ecdsa/ecdsa.h
index f20c8ee7381e..7fb5254b62e6 100644
--- a/crypto/ecdsa/ecdsa.h
+++ b/crypto/ecdsa/ecdsa.h
@@ -4,7 +4,7 @@
  * \author Written by Nils Larsch for the OpenSSL project
  */
 /* ====================================================================
- * Copyright (c) 2000-2003 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2000-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -81,156 +81,143 @@ typedef struct ECDSA_SIG_st
 	BIGNUM *s;
 	} ECDSA_SIG;
 
-/** ECDSA_SIG *ECDSA_SIG_new(void)
- * allocates and initialize a ECDSA_SIG structure
- * \return pointer to a ECDSA_SIG structure or NULL if an error occurred
+/** Allocates and initialize a ECDSA_SIG structure
+ *  \return pointer to a ECDSA_SIG structure or NULL if an error occurred
  */
 ECDSA_SIG *ECDSA_SIG_new(void);
 
-/** ECDSA_SIG_free
- * frees a ECDSA_SIG structure
- * \param a pointer to the ECDSA_SIG structure
+/** frees a ECDSA_SIG structure
+ *  \param  sig  pointer to the ECDSA_SIG structure
  */
-void	  ECDSA_SIG_free(ECDSA_SIG *a);
+void	  ECDSA_SIG_free(ECDSA_SIG *sig);
 
-/** i2d_ECDSA_SIG
- * DER encode content of ECDSA_SIG object (note: this function modifies *pp
- * (*pp += length of the DER encoded signature)).
- * \param a  pointer to the ECDSA_SIG object
- * \param pp pointer to a unsigned char pointer for the output or NULL
- * \return the length of the DER encoded ECDSA_SIG object or 0 
+/** DER encode content of ECDSA_SIG object (note: this function modifies *pp
+ *  (*pp += length of the DER encoded signature)).
+ *  \param  sig  pointer to the ECDSA_SIG object
+ *  \param  pp   pointer to a unsigned char pointer for the output or NULL
+ *  \return the length of the DER encoded ECDSA_SIG object or 0 
  */
-int	  i2d_ECDSA_SIG(const ECDSA_SIG *a, unsigned char **pp);
+int	  i2d_ECDSA_SIG(const ECDSA_SIG *sig, unsigned char **pp);
 
-/** d2i_ECDSA_SIG
- * decodes a DER encoded ECDSA signature (note: this function changes *pp
- * (*pp += len)). 
- * \param v pointer to ECDSA_SIG pointer (may be NULL)
- * \param pp buffer with the DER encoded signature
- * \param len bufferlength
- * \return pointer to the decoded ECDSA_SIG structure (or NULL)
+/** Decodes a DER encoded ECDSA signature (note: this function changes *pp
+ *  (*pp += len)). 
+ *  \param  sig  pointer to ECDSA_SIG pointer (may be NULL)
+ *  \param  pp   memory buffer with the DER encoded signature
+ *  \param  len  length of the buffer
+ *  \return pointer to the decoded ECDSA_SIG structure (or NULL)
  */
-ECDSA_SIG *d2i_ECDSA_SIG(ECDSA_SIG **v, const unsigned char **pp, long len);
+ECDSA_SIG *d2i_ECDSA_SIG(ECDSA_SIG **sig, const unsigned char **pp, long len);
 
-/** ECDSA_do_sign
- * computes the ECDSA signature of the given hash value using
- * the supplied private key and returns the created signature.
- * \param dgst pointer to the hash value
- * \param dgst_len length of the hash value
- * \param eckey pointer to the EC_KEY object containing a private EC key
- * \return pointer to a ECDSA_SIG structure or NULL
+/** Computes the ECDSA signature of the given hash value using
+ *  the supplied private key and returns the created signature.
+ *  \param  dgst      pointer to the hash value
+ *  \param  dgst_len  length of the hash value
+ *  \param  eckey     EC_KEY object containing a private EC key
+ *  \return pointer to a ECDSA_SIG structure or NULL if an error occurred
  */
 ECDSA_SIG *ECDSA_do_sign(const unsigned char *dgst,int dgst_len,EC_KEY *eckey);
 
-/** ECDSA_do_sign_ex
- * computes ECDSA signature of a given hash value using the supplied
- * private key (note: sig must point to ECDSA_size(eckey) bytes of memory).
- * \param dgst pointer to the hash value to sign
- * \param dgstlen length of the hash value
- * \param kinv optional pointer to a pre-computed inverse k
- * \param rp optional pointer to the pre-computed rp value (see 
- *        ECDSA_sign_setup
- * \param eckey pointer to the EC_KEY object containing a private EC key
- * \return pointer to a ECDSA_SIG structure or NULL
+/** Computes ECDSA signature of a given hash value using the supplied
+ *  private key (note: sig must point to ECDSA_size(eckey) bytes of memory).
+ *  \param  dgst     pointer to the hash value to sign
+ *  \param  dgstlen  length of the hash value
+ *  \param  kinv     BIGNUM with a pre-computed inverse k (optional)
+ *  \param  rp       BIGNUM with a pre-computed rp value (optioanl), 
+ *                   see ECDSA_sign_setup
+ *  \param  eckey    EC_KEY object containing a private EC key
+ *  \return pointer to a ECDSA_SIG structure or NULL if an error occurred
  */
 ECDSA_SIG *ECDSA_do_sign_ex(const unsigned char *dgst, int dgstlen, 
 		const BIGNUM *kinv, const BIGNUM *rp, EC_KEY *eckey);
 
-/** ECDSA_do_verify
- * verifies that the supplied signature is a valid ECDSA
- * signature of the supplied hash value using the supplied public key.
- * \param dgst pointer to the hash value
- * \param dgst_len length of the hash value
- * \param sig  pointer to the ECDSA_SIG structure
- * \param eckey pointer to the EC_KEY object containing a public EC key
- * \return 1 if the signature is valid, 0 if the signature is invalid and -1 on error
+/** Verifies that the supplied signature is a valid ECDSA
+ *  signature of the supplied hash value using the supplied public key.
+ *  \param  dgst      pointer to the hash value
+ *  \param  dgst_len  length of the hash value
+ *  \param  sig       ECDSA_SIG structure
+ *  \param  eckey     EC_KEY object containing a public EC key
+ *  \return 1 if the signature is valid, 0 if the signature is invalid
+ *          and -1 on error
  */
 int	  ECDSA_do_verify(const unsigned char *dgst, int dgst_len,
 		const ECDSA_SIG *sig, EC_KEY* eckey);
 
 const ECDSA_METHOD *ECDSA_OpenSSL(void);
 
-/** ECDSA_set_default_method
- * sets the default ECDSA method
- * \param meth the new default ECDSA_METHOD
+/** Sets the default ECDSA method
+ *  \param  meth  new default ECDSA_METHOD
  */
 void	  ECDSA_set_default_method(const ECDSA_METHOD *meth);
 
-/** ECDSA_get_default_method
- * returns the default ECDSA method
- * \return pointer to ECDSA_METHOD structure containing the default method
+/** Returns the default ECDSA method
+ *  \return pointer to ECDSA_METHOD structure containing the default method
  */
 const ECDSA_METHOD *ECDSA_get_default_method(void);
 
-/** ECDSA_set_method
- * sets method to be used for the ECDSA operations
- * \param eckey pointer to the EC_KEY object
- * \param meth  pointer to the new method
- * \return 1 on success and 0 otherwise 
+/** Sets method to be used for the ECDSA operations
+ *  \param  eckey  EC_KEY object
+ *  \param  meth   new method
+ *  \return 1 on success and 0 otherwise 
  */
 int 	  ECDSA_set_method(EC_KEY *eckey, const ECDSA_METHOD *meth);
 
-/** ECDSA_size
- * returns the maximum length of the DER encoded signature
- * \param  eckey pointer to a EC_KEY object
- * \return numbers of bytes required for the DER encoded signature
+/** Returns the maximum length of the DER encoded signature
+ *  \param  eckey  EC_KEY object
+ *  \return numbers of bytes required for the DER encoded signature
  */
 int	  ECDSA_size(const EC_KEY *eckey);
 
-/** ECDSA_sign_setup
- * precompute parts of the signing operation. 
- * \param eckey pointer to the EC_KEY object containing a private EC key
- * \param ctx  pointer to a BN_CTX object (may be NULL)
- * \param kinv pointer to a BIGNUM pointer for the inverse of k
- * \param rp   pointer to a BIGNUM pointer for x coordinate of k * generator
- * \return 1 on success and 0 otherwise
+/** Precompute parts of the signing operation
+ *  \param  eckey  EC_KEY object containing a private EC key
+ *  \param  ctx    BN_CTX object (optional)
+ *  \param  kinv   BIGNUM pointer for the inverse of k
+ *  \param  rp     BIGNUM pointer for x coordinate of k * generator
+ *  \return 1 on success and 0 otherwise
  */
 int 	  ECDSA_sign_setup(EC_KEY *eckey, BN_CTX *ctx, BIGNUM **kinv, 
 		BIGNUM **rp);
 
-/** ECDSA_sign
- * computes ECDSA signature of a given hash value using the supplied
- * private key (note: sig must point to ECDSA_size(eckey) bytes of memory).
- * \param type this parameter is ignored
- * \param dgst pointer to the hash value to sign
- * \param dgstlen length of the hash value
- * \param sig buffer to hold the DER encoded signature
- * \param siglen pointer to the length of the returned signature
- * \param eckey pointer to the EC_KEY object containing a private EC key
- * \return 1 on success and 0 otherwise
+/** Computes ECDSA signature of a given hash value using the supplied
+ *  private key (note: sig must point to ECDSA_size(eckey) bytes of memory).
+ *  \param  type     this parameter is ignored
+ *  \param  dgst     pointer to the hash value to sign
+ *  \param  dgstlen  length of the hash value
+ *  \param  sig      memory for the DER encoded created signature
+ *  \param  siglen   pointer to the length of the returned signature
+ *  \param  eckey    EC_KEY object containing a private EC key
+ *  \return 1 on success and 0 otherwise
  */
 int	  ECDSA_sign(int type, const unsigned char *dgst, int dgstlen, 
 		unsigned char *sig, unsigned int *siglen, EC_KEY *eckey);
 
 
-/** ECDSA_sign_ex
- * computes ECDSA signature of a given hash value using the supplied
- * private key (note: sig must point to ECDSA_size(eckey) bytes of memory).
- * \param type this parameter is ignored
- * \param dgst pointer to the hash value to sign
- * \param dgstlen length of the hash value
- * \param sig buffer to hold the DER encoded signature
- * \param siglen pointer to the length of the returned signature
- * \param kinv optional pointer to a pre-computed inverse k
- * \param rp optional pointer to the pre-computed rp value (see 
- *        ECDSA_sign_setup
- * \param eckey pointer to the EC_KEY object containing a private EC key
- * \return 1 on success and 0 otherwise
+/** Computes ECDSA signature of a given hash value using the supplied
+ *  private key (note: sig must point to ECDSA_size(eckey) bytes of memory).
+ *  \param  type     this parameter is ignored
+ *  \param  dgst     pointer to the hash value to sign
+ *  \param  dgstlen  length of the hash value
+ *  \param  sig      buffer to hold the DER encoded signature
+ *  \param  siglen   pointer to the length of the returned signature
+ *  \param  kinv     BIGNUM with a pre-computed inverse k (optional)
+ *  \param  rp       BIGNUM with a pre-computed rp value (optioanl), 
+ *                   see ECDSA_sign_setup
+ *  \param  eckey    EC_KEY object containing a private EC key
+ *  \return 1 on success and 0 otherwise
  */
 int	  ECDSA_sign_ex(int type, const unsigned char *dgst, int dgstlen, 
 		unsigned char *sig, unsigned int *siglen, const BIGNUM *kinv,
 		const BIGNUM *rp, EC_KEY *eckey);
 
-/** ECDSA_verify
- * verifies that the given signature is valid ECDSA signature
- * of the supplied hash value using the specified public key.
- * \param type this parameter is ignored
- * \param dgst pointer to the hash value 
- * \param dgstlen length of the hash value
- * \param sig  pointer to the DER encoded signature
- * \param siglen length of the DER encoded signature
- * \param eckey pointer to the EC_KEY object containing a public EC key
- * \return 1 if the signature is valid, 0 if the signature is invalid and -1 on error
+/** Verifies that the given signature is valid ECDSA signature
+ *  of the supplied hash value using the specified public key.
+ *  \param  type     this parameter is ignored
+ *  \param  dgst     pointer to the hash value 
+ *  \param  dgstlen  length of the hash value
+ *  \param  sig      pointer to the DER encoded signature
+ *  \param  siglen   length of the DER encoded signature
+ *  \param  eckey    EC_KEY object containing a public EC key
+ *  \return 1 if the signature is valid, 0 if the signature is invalid
+ *          and -1 on error
  */
 int 	  ECDSA_verify(int type, const unsigned char *dgst, int dgstlen, 
 		const unsigned char *sig, int siglen, EC_KEY *eckey);
@@ -251,6 +238,7 @@ void ERR_load_ECDSA_strings(void);
 /* Error codes for the ECDSA functions. */
 
 /* Function codes. */
+#define ECDSA_F_ECDSA_CHECK				 104
 #define ECDSA_F_ECDSA_DATA_NEW_METHOD			 100
 #define ECDSA_F_ECDSA_DO_SIGN				 101
 #define ECDSA_F_ECDSA_DO_VERIFY				 102
@@ -262,6 +250,7 @@ void ERR_load_ECDSA_strings(void);
 #define ECDSA_R_ERR_EC_LIB				 102
 #define ECDSA_R_MISSING_PARAMETERS			 103
 #define ECDSA_R_NEED_NEW_SETUP_VALUES			 106
+#define ECDSA_R_NON_FIPS_METHOD				 107
 #define ECDSA_R_RANDOM_NUMBER_GENERATION_FAILED		 104
 #define ECDSA_R_SIGNATURE_MALLOC_FAILED			 105
 
diff --git a/crypto/ecdsa/ecdsatest.c b/crypto/ecdsa/ecdsatest.c
index 1ad507d47172..537bb30362cf 100644
--- a/crypto/ecdsa/ecdsatest.c
+++ b/crypto/ecdsa/ecdsatest.c
@@ -262,6 +262,7 @@ int x9_62_tests(BIO *out)
 		"3238135532097973577080787768312505059318910517550078427819"
 		"78505179448783"))
 		goto x962_err;
+#ifndef OPENSSL_NO_EC2M
 	if (!x9_62_test_internal(out, NID_X9_62_c2tnb191v1,
 		"87194383164871543355722284926904419997237591535066528048",
 		"308992691965804947361541664549085895292153777025772063598"))
@@ -272,7 +273,7 @@ int x9_62_tests(BIO *out)
 		"1970303740007316867383349976549972270528498040721988191026"
 		"49413465737174"))
 		goto x962_err;
-
+#endif
 	ret = 1;
 x962_err:
 	if (!restore_rand())
@@ -561,7 +562,7 @@ err:
 	if (ret)
 		ERR_print_errors(out);
 	CRYPTO_cleanup_all_ex_data();
-	ERR_remove_state(0);
+	ERR_remove_thread_state(NULL);
 	ERR_free_strings();
 	CRYPTO_mem_leaks(out);
 	if (out != NULL)
diff --git a/crypto/ecdsa/ecs_err.c b/crypto/ecdsa/ecs_err.c
index d2a53730ea5f..81542e6d1539 100644
--- a/crypto/ecdsa/ecs_err.c
+++ b/crypto/ecdsa/ecs_err.c
@@ -1,6 +1,6 @@
 /* crypto/ecdsa/ecs_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,6 +70,7 @@
 
 static ERR_STRING_DATA ECDSA_str_functs[]=
 	{
+{ERR_FUNC(ECDSA_F_ECDSA_CHECK),	"ECDSA_CHECK"},
 {ERR_FUNC(ECDSA_F_ECDSA_DATA_NEW_METHOD),	"ECDSA_DATA_NEW_METHOD"},
 {ERR_FUNC(ECDSA_F_ECDSA_DO_SIGN),	"ECDSA_do_sign"},
 {ERR_FUNC(ECDSA_F_ECDSA_DO_VERIFY),	"ECDSA_do_verify"},
@@ -84,6 +85,7 @@ static ERR_STRING_DATA ECDSA_str_reasons[]=
 {ERR_REASON(ECDSA_R_ERR_EC_LIB)          ,"err ec lib"},
 {ERR_REASON(ECDSA_R_MISSING_PARAMETERS)  ,"missing parameters"},
 {ERR_REASON(ECDSA_R_NEED_NEW_SETUP_VALUES),"need new setup values"},
+{ERR_REASON(ECDSA_R_NON_FIPS_METHOD)     ,"non fips method"},
 {ERR_REASON(ECDSA_R_RANDOM_NUMBER_GENERATION_FAILED),"random number generation failed"},
 {ERR_REASON(ECDSA_R_SIGNATURE_MALLOC_FAILED),"signature malloc failed"},
 {0,NULL}
diff --git a/crypto/ecdsa/ecs_lib.c b/crypto/ecdsa/ecs_lib.c
index 2ebae3aa27d7..e477da430bab 100644
--- a/crypto/ecdsa/ecs_lib.c
+++ b/crypto/ecdsa/ecs_lib.c
@@ -60,6 +60,9 @@
 #endif
 #include <openssl/err.h>
 #include <openssl/bn.h>
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 
 const char ECDSA_version[]="ECDSA" OPENSSL_VERSION_PTEXT;
 
@@ -77,7 +80,16 @@ void ECDSA_set_default_method(const ECDSA_METHOD *meth)
 const ECDSA_METHOD *ECDSA_get_default_method(void)
 {
 	if(!default_ECDSA_method) 
+		{
+#ifdef OPENSSL_FIPS
+		if (FIPS_mode())
+			return FIPS_ecdsa_openssl();
+		else
+			return ECDSA_OpenSSL();
+#else
 		default_ECDSA_method = ECDSA_OpenSSL();
+#endif
+		}
 	return default_ECDSA_method;
 }
 
@@ -193,7 +205,14 @@ ECDSA_DATA *ecdsa_check(EC_KEY *key)
 	}
 	else
 		ecdsa_data = (ECDSA_DATA *)data;
-	
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(ecdsa_data->flags & ECDSA_FLAG_FIPS_METHOD)
+			&& !(EC_KEY_get_flags(key) & EC_FLAG_NON_FIPS_ALLOW))
+		{
+		ECDSAerr(ECDSA_F_ECDSA_CHECK, ECDSA_R_NON_FIPS_METHOD);
+		return NULL;
+		}
+#endif
 
 	return ecdsa_data;
 }
diff --git a/crypto/ecdsa/ecs_locl.h b/crypto/ecdsa/ecs_locl.h
index 3a69a840e211..cb3be13cfc30 100644
--- a/crypto/ecdsa/ecs_locl.h
+++ b/crypto/ecdsa/ecs_locl.h
@@ -82,6 +82,14 @@ struct ecdsa_method
 	char *app_data;
 	};
 
+/* If this flag is set the ECDSA method is FIPS compliant and can be used
+ * in FIPS mode. This is set in the validated module method. If an
+ * application sets this flag in its own methods it is its responsibility
+ * to ensure the result is compliant.
+ */
+
+#define ECDSA_FLAG_FIPS_METHOD	0x1
+
 typedef struct ecdsa_data_st {
 	/* EC_KEY_METH_DATA part */
 	int (*init)(EC_KEY *);
diff --git a/crypto/ecdsa/ecs_ossl.c b/crypto/ecdsa/ecs_ossl.c
index 1bbf328de54f..7725935610ed 100644
--- a/crypto/ecdsa/ecs_ossl.c
+++ b/crypto/ecdsa/ecs_ossl.c
@@ -167,6 +167,7 @@ static int ecdsa_sign_setup(EC_KEY *eckey, BN_CTX *ctx_in, BIGNUM **kinvp,
 				goto err;
 			}
 		}
+#ifndef OPENSSL_NO_EC2M
 		else /* NID_X9_62_characteristic_two_field */
 		{
 			if (!EC_POINT_get_affine_coordinates_GF2m(group,
@@ -176,6 +177,7 @@ static int ecdsa_sign_setup(EC_KEY *eckey, BN_CTX *ctx_in, BIGNUM **kinvp,
 				goto err;
 			}
 		}
+#endif
 		if (!BN_nnmod(r, X, order, ctx))
 		{
 			ECDSAerr(ECDSA_F_ECDSA_SIGN_SETUP, ERR_R_BN_LIB);
@@ -454,6 +456,7 @@ static int ecdsa_do_verify(const unsigned char *dgst, int dgst_len,
 			goto err;
 		}
 	}
+#ifndef OPENSSL_NO_EC2M
 	else /* NID_X9_62_characteristic_two_field */
 	{
 		if (!EC_POINT_get_affine_coordinates_GF2m(group,
@@ -463,7 +466,7 @@ static int ecdsa_do_verify(const unsigned char *dgst, int dgst_len,
 			goto err;
 		}
 	}
-	
+#endif	
 	if (!BN_nnmod(u1, X, order, ctx))
 	{
 		ECDSAerr(ECDSA_F_ECDSA_DO_VERIFY, ERR_R_BN_LIB);
diff --git a/crypto/engine/Makefile b/crypto/engine/Makefile
index b52fa48065b6..d29bdd09a096 100644
--- a/crypto/engine/Makefile
+++ b/crypto/engine/Makefile
@@ -20,13 +20,15 @@ LIB=$(TOP)/libcrypto.a
 LIBSRC= eng_err.c eng_lib.c eng_list.c eng_init.c eng_ctrl.c \
 	eng_table.c eng_pkey.c eng_fat.c eng_all.c \
 	tb_rsa.c tb_dsa.c tb_ecdsa.c tb_dh.c tb_ecdh.c tb_rand.c tb_store.c \
-	tb_cipher.c tb_digest.c \
-	eng_openssl.c eng_cnf.c eng_dyn.c eng_cryptodev.c eng_padlock.c
+	tb_cipher.c tb_digest.c tb_pkmeth.c tb_asnmth.c \
+	eng_openssl.c eng_cnf.c eng_dyn.c eng_cryptodev.c \
+	eng_rsax.c eng_rdrand.c
 LIBOBJ= eng_err.o eng_lib.o eng_list.o eng_init.o eng_ctrl.o \
 	eng_table.o eng_pkey.o eng_fat.o eng_all.o \
 	tb_rsa.o tb_dsa.o tb_ecdsa.o tb_dh.o tb_ecdh.o tb_rand.o tb_store.o \
-	tb_cipher.o tb_digest.o \
-	eng_openssl.o eng_cnf.o eng_dyn.o eng_cryptodev.o eng_padlock.o
+	tb_cipher.o tb_digest.o tb_pkmeth.o tb_asnmth.o \
+	eng_openssl.o eng_cnf.o eng_dyn.o eng_cryptodev.o \
+	eng_rsax.o eng_rdrand.o
 
 SRC= $(LIBSRC)
 
@@ -41,7 +43,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -88,54 +90,50 @@ eng_all.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 eng_all.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 eng_all.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 eng_all.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-eng_all.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-eng_all.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-eng_all.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-eng_all.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-eng_all.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-eng_all.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-eng_all.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-eng_all.o: ../cryptlib.h eng_all.c eng_int.h
+eng_all.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+eng_all.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+eng_all.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+eng_all.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+eng_all.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+eng_all.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+eng_all.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_all.c eng_int.h
 eng_cnf.o: ../../e_os.h ../../include/openssl/asn1.h
 eng_cnf.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 eng_cnf.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 eng_cnf.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 eng_cnf.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 eng_cnf.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-eng_cnf.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-eng_cnf.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-eng_cnf.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-eng_cnf.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-eng_cnf.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-eng_cnf.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-eng_cnf.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-eng_cnf.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_cnf.c eng_int.h
+eng_cnf.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+eng_cnf.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+eng_cnf.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+eng_cnf.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+eng_cnf.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+eng_cnf.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+eng_cnf.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+eng_cnf.o: ../cryptlib.h eng_cnf.c eng_int.h
 eng_cryptodev.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 eng_cryptodev.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
-eng_cryptodev.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
-eng_cryptodev.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
+eng_cryptodev.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 eng_cryptodev.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 eng_cryptodev.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
-eng_cryptodev.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-eng_cryptodev.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+eng_cryptodev.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
 eng_cryptodev.o: ../../include/openssl/obj_mac.h
 eng_cryptodev.o: ../../include/openssl/objects.h
 eng_cryptodev.o: ../../include/openssl/opensslconf.h
 eng_cryptodev.o: ../../include/openssl/opensslv.h
 eng_cryptodev.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-eng_cryptodev.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
-eng_cryptodev.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-eng_cryptodev.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-eng_cryptodev.o: ../../include/openssl/x509_vfy.h eng_cryptodev.c
+eng_cryptodev.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+eng_cryptodev.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+eng_cryptodev.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+eng_cryptodev.o: eng_cryptodev.c
 eng_ctrl.o: ../../e_os.h ../../include/openssl/asn1.h
 eng_ctrl.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 eng_ctrl.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 eng_ctrl.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 eng_ctrl.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 eng_ctrl.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-eng_ctrl.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-eng_ctrl.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-eng_ctrl.o: ../../include/openssl/opensslconf.h
+eng_ctrl.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+eng_ctrl.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 eng_ctrl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 eng_ctrl.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 eng_ctrl.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -147,50 +145,49 @@ eng_dyn.o: ../../include/openssl/crypto.h ../../include/openssl/dso.h
 eng_dyn.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 eng_dyn.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 eng_dyn.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-eng_dyn.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-eng_dyn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-eng_dyn.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-eng_dyn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-eng_dyn.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-eng_dyn.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-eng_dyn.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-eng_dyn.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_dyn.c eng_int.h
+eng_dyn.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+eng_dyn.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+eng_dyn.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+eng_dyn.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+eng_dyn.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+eng_dyn.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+eng_dyn.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+eng_dyn.o: ../cryptlib.h eng_dyn.c eng_int.h
 eng_err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 eng_err.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 eng_err.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 eng_err.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 eng_err.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-eng_err.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-eng_err.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-eng_err.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-eng_err.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-eng_err.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-eng_err.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-eng_err.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-eng_err.o: ../../include/openssl/x509_vfy.h eng_err.c
+eng_err.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+eng_err.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+eng_err.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+eng_err.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+eng_err.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+eng_err.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+eng_err.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+eng_err.o: eng_err.c
 eng_fat.o: ../../e_os.h ../../include/openssl/asn1.h
 eng_fat.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 eng_fat.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 eng_fat.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 eng_fat.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 eng_fat.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-eng_fat.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-eng_fat.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-eng_fat.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-eng_fat.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-eng_fat.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-eng_fat.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-eng_fat.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-eng_fat.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_fat.c eng_int.h
+eng_fat.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+eng_fat.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+eng_fat.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+eng_fat.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+eng_fat.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+eng_fat.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+eng_fat.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+eng_fat.o: ../cryptlib.h eng_fat.c eng_int.h
 eng_init.o: ../../e_os.h ../../include/openssl/asn1.h
 eng_init.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 eng_init.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 eng_init.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 eng_init.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 eng_init.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-eng_init.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-eng_init.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-eng_init.o: ../../include/openssl/opensslconf.h
+eng_init.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+eng_init.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 eng_init.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 eng_init.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 eng_init.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -202,23 +199,22 @@ eng_lib.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 eng_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 eng_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 eng_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-eng_lib.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-eng_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-eng_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-eng_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-eng_lib.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-eng_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-eng_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-eng_lib.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h eng_lib.c
+eng_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+eng_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+eng_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+eng_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+eng_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+eng_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+eng_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+eng_lib.o: ../cryptlib.h eng_int.h eng_lib.c
 eng_list.o: ../../e_os.h ../../include/openssl/asn1.h
 eng_list.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 eng_list.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 eng_list.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 eng_list.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 eng_list.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-eng_list.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-eng_list.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-eng_list.o: ../../include/openssl/opensslconf.h
+eng_list.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+eng_list.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 eng_list.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 eng_list.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 eng_list.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -231,9 +227,8 @@ eng_openssl.o: ../../include/openssl/dsa.h ../../include/openssl/dso.h
 eng_openssl.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 eng_openssl.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 eng_openssl.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-eng_openssl.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-eng_openssl.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-eng_openssl.o: ../../include/openssl/objects.h
+eng_openssl.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+eng_openssl.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
 eng_openssl.o: ../../include/openssl/opensslconf.h
 eng_openssl.o: ../../include/openssl/opensslv.h
 eng_openssl.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
@@ -243,44 +238,55 @@ eng_openssl.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
 eng_openssl.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 eng_openssl.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 eng_openssl.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_openssl.c
-eng_padlock.o: ../../include/openssl/aes.h ../../include/openssl/asn1.h
-eng_padlock.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
-eng_padlock.o: ../../include/openssl/crypto.h ../../include/openssl/dso.h
-eng_padlock.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
-eng_padlock.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-eng_padlock.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-eng_padlock.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-eng_padlock.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-eng_padlock.o: ../../include/openssl/objects.h
-eng_padlock.o: ../../include/openssl/opensslconf.h
-eng_padlock.o: ../../include/openssl/opensslv.h
-eng_padlock.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-eng_padlock.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-eng_padlock.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-eng_padlock.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-eng_padlock.o: ../../include/openssl/x509_vfy.h eng_padlock.c
 eng_pkey.o: ../../e_os.h ../../include/openssl/asn1.h
 eng_pkey.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 eng_pkey.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 eng_pkey.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 eng_pkey.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 eng_pkey.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-eng_pkey.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-eng_pkey.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-eng_pkey.o: ../../include/openssl/opensslconf.h
+eng_pkey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+eng_pkey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 eng_pkey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 eng_pkey.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 eng_pkey.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 eng_pkey.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 eng_pkey.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h eng_pkey.c
+eng_rdrand.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+eng_rdrand.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+eng_rdrand.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+eng_rdrand.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+eng_rdrand.o: ../../include/openssl/engine.h ../../include/openssl/err.h
+eng_rdrand.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+eng_rdrand.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+eng_rdrand.o: ../../include/openssl/opensslconf.h
+eng_rdrand.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+eng_rdrand.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+eng_rdrand.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+eng_rdrand.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+eng_rdrand.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+eng_rdrand.o: eng_rdrand.c
+eng_rsax.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+eng_rsax.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
+eng_rsax.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+eng_rsax.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+eng_rsax.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
+eng_rsax.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+eng_rsax.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+eng_rsax.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+eng_rsax.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+eng_rsax.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+eng_rsax.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+eng_rsax.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+eng_rsax.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+eng_rsax.o: eng_rsax.c
 eng_table.o: ../../e_os.h ../../include/openssl/asn1.h
 eng_table.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 eng_table.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 eng_table.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 eng_table.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 eng_table.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-eng_table.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-eng_table.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+eng_table.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+eng_table.o: ../../include/openssl/objects.h
 eng_table.o: ../../include/openssl/opensslconf.h
 eng_table.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 eng_table.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
@@ -288,14 +294,29 @@ eng_table.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 eng_table.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 eng_table.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h
 eng_table.o: eng_table.c
+tb_asnmth.o: ../../e_os.h ../../include/openssl/asn1.h
+tb_asnmth.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+tb_asnmth.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+tb_asnmth.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+tb_asnmth.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
+tb_asnmth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+tb_asnmth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+tb_asnmth.o: ../../include/openssl/objects.h
+tb_asnmth.o: ../../include/openssl/opensslconf.h
+tb_asnmth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+tb_asnmth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+tb_asnmth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+tb_asnmth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+tb_asnmth.o: ../../include/openssl/x509_vfy.h ../asn1/asn1_locl.h ../cryptlib.h
+tb_asnmth.o: eng_int.h tb_asnmth.c
 tb_cipher.o: ../../e_os.h ../../include/openssl/asn1.h
 tb_cipher.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 tb_cipher.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 tb_cipher.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 tb_cipher.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 tb_cipher.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-tb_cipher.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-tb_cipher.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+tb_cipher.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+tb_cipher.o: ../../include/openssl/objects.h
 tb_cipher.o: ../../include/openssl/opensslconf.h
 tb_cipher.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 tb_cipher.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
@@ -308,22 +329,22 @@ tb_dh.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 tb_dh.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 tb_dh.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 tb_dh.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-tb_dh.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-tb_dh.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-tb_dh.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-tb_dh.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-tb_dh.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-tb_dh.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-tb_dh.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-tb_dh.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h tb_dh.c
+tb_dh.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+tb_dh.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+tb_dh.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+tb_dh.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+tb_dh.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+tb_dh.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+tb_dh.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+tb_dh.o: ../cryptlib.h eng_int.h tb_dh.c
 tb_digest.o: ../../e_os.h ../../include/openssl/asn1.h
 tb_digest.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 tb_digest.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 tb_digest.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 tb_digest.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 tb_digest.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-tb_digest.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-tb_digest.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+tb_digest.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+tb_digest.o: ../../include/openssl/objects.h
 tb_digest.o: ../../include/openssl/opensslconf.h
 tb_digest.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 tb_digest.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
@@ -336,78 +357,89 @@ tb_dsa.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 tb_dsa.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 tb_dsa.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 tb_dsa.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-tb_dsa.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-tb_dsa.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-tb_dsa.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-tb_dsa.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-tb_dsa.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-tb_dsa.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-tb_dsa.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-tb_dsa.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h tb_dsa.c
+tb_dsa.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+tb_dsa.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+tb_dsa.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+tb_dsa.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+tb_dsa.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+tb_dsa.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+tb_dsa.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+tb_dsa.o: ../cryptlib.h eng_int.h tb_dsa.c
 tb_ecdh.o: ../../e_os.h ../../include/openssl/asn1.h
 tb_ecdh.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 tb_ecdh.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 tb_ecdh.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 tb_ecdh.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 tb_ecdh.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-tb_ecdh.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-tb_ecdh.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-tb_ecdh.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-tb_ecdh.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-tb_ecdh.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-tb_ecdh.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-tb_ecdh.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-tb_ecdh.o: ../cryptlib.h eng_int.h tb_ecdh.c
+tb_ecdh.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+tb_ecdh.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+tb_ecdh.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+tb_ecdh.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+tb_ecdh.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+tb_ecdh.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+tb_ecdh.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h tb_ecdh.c
 tb_ecdsa.o: ../../e_os.h ../../include/openssl/asn1.h
 tb_ecdsa.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 tb_ecdsa.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 tb_ecdsa.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 tb_ecdsa.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 tb_ecdsa.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-tb_ecdsa.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-tb_ecdsa.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-tb_ecdsa.o: ../../include/openssl/opensslconf.h
+tb_ecdsa.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+tb_ecdsa.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 tb_ecdsa.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 tb_ecdsa.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 tb_ecdsa.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 tb_ecdsa.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 tb_ecdsa.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h tb_ecdsa.c
+tb_pkmeth.o: ../../e_os.h ../../include/openssl/asn1.h
+tb_pkmeth.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+tb_pkmeth.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+tb_pkmeth.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+tb_pkmeth.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
+tb_pkmeth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+tb_pkmeth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+tb_pkmeth.o: ../../include/openssl/objects.h
+tb_pkmeth.o: ../../include/openssl/opensslconf.h
+tb_pkmeth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+tb_pkmeth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+tb_pkmeth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+tb_pkmeth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+tb_pkmeth.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h
+tb_pkmeth.o: tb_pkmeth.c
 tb_rand.o: ../../e_os.h ../../include/openssl/asn1.h
 tb_rand.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 tb_rand.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 tb_rand.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 tb_rand.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 tb_rand.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-tb_rand.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-tb_rand.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-tb_rand.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-tb_rand.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-tb_rand.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-tb_rand.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-tb_rand.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-tb_rand.o: ../cryptlib.h eng_int.h tb_rand.c
+tb_rand.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+tb_rand.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+tb_rand.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+tb_rand.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+tb_rand.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+tb_rand.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+tb_rand.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h tb_rand.c
 tb_rsa.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 tb_rsa.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 tb_rsa.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 tb_rsa.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 tb_rsa.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-tb_rsa.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-tb_rsa.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-tb_rsa.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-tb_rsa.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-tb_rsa.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-tb_rsa.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-tb_rsa.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-tb_rsa.o: ../../include/openssl/x509_vfy.h ../cryptlib.h eng_int.h tb_rsa.c
+tb_rsa.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+tb_rsa.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+tb_rsa.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+tb_rsa.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+tb_rsa.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+tb_rsa.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+tb_rsa.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+tb_rsa.o: ../cryptlib.h eng_int.h tb_rsa.c
 tb_store.o: ../../e_os.h ../../include/openssl/asn1.h
 tb_store.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 tb_store.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 tb_store.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 tb_store.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 tb_store.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-tb_store.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-tb_store.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-tb_store.o: ../../include/openssl/opensslconf.h
+tb_store.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+tb_store.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 tb_store.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 tb_store.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 tb_store.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
diff --git a/crypto/engine/eng_all.c b/crypto/engine/eng_all.c
index f29c167c0698..6093376df41a 100644
--- a/crypto/engine/eng_all.c
+++ b/crypto/engine/eng_all.c
@@ -61,15 +61,23 @@
 
 void ENGINE_load_builtin_engines(void)
 	{
+	/* Some ENGINEs need this */
+	OPENSSL_cpuid_setup();
+#if 0
 	/* There's no longer any need for an "openssl" ENGINE unless, one day,
 	 * it is the *only* way for standard builtin implementations to be be
 	 * accessed (ie. it would be possible to statically link binaries with
 	 * *no* builtin implementations). */
-#if 0
 	ENGINE_load_openssl();
 #endif
-#if !defined(OPENSSL_NO_HW) && !defined(OPENSSL_NO_HW_PADLOCK)
-	ENGINE_load_padlock();
+#if !defined(OPENSSL_NO_HW) && (defined(__OpenBSD__) || defined(__FreeBSD__) || defined(HAVE_CRYPTODEV))
+	ENGINE_load_cryptodev();
+#endif
+#ifndef OPENSSL_NO_RSAX
+	ENGINE_load_rsax();
+#endif
+#ifndef OPENSSL_NO_RDRAND
+	ENGINE_load_rdrand();
 #endif
 	ENGINE_load_dynamic();
 #ifndef OPENSSL_NO_STATIC_ENGINE
@@ -98,22 +106,24 @@ void ENGINE_load_builtin_engines(void)
 #ifndef OPENSSL_NO_HW_UBSEC
 	ENGINE_load_ubsec();
 #endif
+#ifndef OPENSSL_NO_HW_PADLOCK
+	ENGINE_load_padlock();
 #endif
-#if !defined(OPENSSL_NO_GMP) && !defined(OPENSSL_NO_HW_GMP)
-	ENGINE_load_gmp();
 #endif
+#ifndef OPENSSL_NO_GOST
+	ENGINE_load_gost();
 #endif
-#ifndef OPENSSL_NO_HW
-#if defined(__OpenBSD__) || defined(__FreeBSD__) || defined(HAVE_CRYPTODEV)
-	ENGINE_load_cryptodev();
+#ifndef OPENSSL_NO_GMP
+	ENGINE_load_gmp();
 #endif
 #if defined(OPENSSL_SYS_WIN32) && !defined(OPENSSL_NO_CAPIENG)
 	ENGINE_load_capi();
 #endif
 #endif
+	ENGINE_register_all_complete();
 	}
 
-#if defined(__OpenBSD__) || defined(__FreeBSD__)
+#if defined(__OpenBSD__) || defined(__FreeBSD__) || defined(HAVE_CRYPTODEV)
 void ENGINE_setup_bsd_cryptodev(void) {
 	static int bsd_cryptodev_default_loaded = 0;
 	if (!bsd_cryptodev_default_loaded) {
diff --git a/crypto/engine/eng_cryptodev.c b/crypto/engine/eng_cryptodev.c
index eef1e2d81424..5a715aca4f26 100644
--- a/crypto/engine/eng_cryptodev.c
+++ b/crypto/engine/eng_cryptodev.c
@@ -30,10 +30,6 @@
 #include <openssl/engine.h>
 #include <openssl/evp.h>
 #include <openssl/bn.h>
-#include <openssl/dsa.h>
-#include <openssl/rsa.h>
-#include <openssl/dh.h>
-#include <openssl/err.h>
 
 #if (defined(__unix__) || defined(unix)) && !defined(USG) && \
 	(defined(OpenBSD) || defined(__FreeBSD__))
@@ -59,6 +55,10 @@ ENGINE_load_cryptodev(void)
  
 #include <sys/types.h>
 #include <crypto/cryptodev.h>
+#include <crypto/dh/dh.h>
+#include <crypto/dsa/dsa.h>
+#include <crypto/err/err.h>
+#include <crypto/rsa/rsa.h>
 #include <sys/ioctl.h>
 #include <errno.h>
 #include <stdio.h>
@@ -72,6 +72,14 @@ ENGINE_load_cryptodev(void)
 struct dev_crypto_state {
 	struct session_op d_sess;
 	int d_fd;
+
+#ifdef USE_CRYPTODEV_DIGESTS
+	char dummy_mac_key[HASH_MAX_LEN];
+
+	unsigned char digest_res[HASH_MAX_LEN];
+	char *mac_data;
+	int mac_len;
+#endif
 };
 
 static u_int32_t cryptodev_asymfeat = 0;
@@ -79,15 +87,14 @@ static u_int32_t cryptodev_asymfeat = 0;
 static int get_asym_dev_crypto(void);
 static int open_dev_crypto(void);
 static int get_dev_crypto(void);
-static int cryptodev_max_iv(int cipher);
-static int cryptodev_key_length_valid(int cipher, int len);
-static int cipher_nid_to_cryptodev(int nid);
 static int get_cryptodev_ciphers(const int **cnids);
-/*static int get_cryptodev_digests(const int **cnids);*/
+#ifdef USE_CRYPTODEV_DIGESTS
+static int get_cryptodev_digests(const int **cnids);
+#endif
 static int cryptodev_usable_ciphers(const int **nids);
 static int cryptodev_usable_digests(const int **nids);
 static int cryptodev_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-    const unsigned char *in, unsigned int inl);
+    const unsigned char *in, size_t inl);
 static int cryptodev_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
     const unsigned char *iv, int enc);
 static int cryptodev_cleanup(EVP_CIPHER_CTX *ctx);
@@ -121,7 +128,7 @@ static int cryptodev_mod_exp_dh(const DH *dh, BIGNUM *r, const BIGNUM *a,
 static int cryptodev_dh_compute_key(unsigned char *key,
     const BIGNUM *pub_key, DH *dh);
 static int cryptodev_ctrl(ENGINE *e, int cmd, long i, void *p,
-    void (*f)());
+    void (*f)(void));
 void ENGINE_load_cryptodev(void);
 
 static const ENGINE_CMD_DEFN cryptodev_defns[] = {
@@ -134,27 +141,32 @@ static struct {
 	int	ivmax;
 	int	keylen;
 } ciphers[] = {
+	{ CRYPTO_ARC4,			NID_rc4,		0,	16, },
 	{ CRYPTO_DES_CBC,		NID_des_cbc,		8,	 8, },
 	{ CRYPTO_3DES_CBC,		NID_des_ede3_cbc,	8,	24, },
 	{ CRYPTO_AES_CBC,		NID_aes_128_cbc,	16,	16, },
+	{ CRYPTO_AES_CBC,		NID_aes_192_cbc,	16,	24, },
+	{ CRYPTO_AES_CBC,		NID_aes_256_cbc,	16,	32, },
 	{ CRYPTO_BLF_CBC,		NID_bf_cbc,		8,	16, },
 	{ CRYPTO_CAST_CBC,		NID_cast5_cbc,		8,	16, },
 	{ CRYPTO_SKIPJACK_CBC,		NID_undef,		0,	 0, },
 	{ 0,				NID_undef,		0,	 0, },
 };
 
-#if 0
+#ifdef USE_CRYPTODEV_DIGESTS
 static struct {
 	int	id;
 	int	nid;
+	int 	keylen;
 } digests[] = {
-	{ CRYPTO_SHA1_HMAC,		NID_hmacWithSHA1,	},
-	{ CRYPTO_RIPEMD160_HMAC,	NID_ripemd160,		},
-	{ CRYPTO_MD5_KPDK,		NID_undef,		},
-	{ CRYPTO_SHA1_KPDK,		NID_undef,		},
-	{ CRYPTO_MD5,			NID_md5,		},
-	{ CRYPTO_SHA1,			NID_undef,		},
-	{ 0,				NID_undef,		},
+	{ CRYPTO_MD5_HMAC,		NID_hmacWithMD5,	16},
+	{ CRYPTO_SHA1_HMAC,		NID_hmacWithSHA1,	20},
+	{ CRYPTO_RIPEMD160_HMAC,	NID_ripemd160,		16/*?*/},
+	{ CRYPTO_MD5_KPDK,		NID_undef,		0},
+	{ CRYPTO_SHA1_KPDK,		NID_undef,		0},
+	{ CRYPTO_MD5,			NID_md5,		16},
+	{ CRYPTO_SHA1,			NID_sha1,		20},
+	{ 0,				NID_undef,		0},
 };
 #endif
 
@@ -186,6 +198,7 @@ get_dev_crypto(void)
 
 	if ((fd = open_dev_crypto()) == -1)
 		return (-1);
+#ifndef CRIOGET_NOT_NEEDED
 	if (ioctl(fd, CRIOGET, &retfd) == -1)
 		return (-1);
 
@@ -194,9 +207,19 @@ get_dev_crypto(void)
 		close(retfd);
 		return (-1);
 	}
+#else
+        retfd = fd;
+#endif
 	return (retfd);
 }
 
+static void put_dev_crypto(int fd)
+{
+#ifndef CRIOGET_NOT_NEEDED
+	close(fd);
+#endif
+}
+
 /* Caching version for asym operations */
 static int
 get_asym_dev_crypto(void)
@@ -209,50 +232,6 @@ get_asym_dev_crypto(void)
 }
 
 /*
- * XXXX this needs to be set for each alg - and determined from
- * a running card.
- */
-static int
-cryptodev_max_iv(int cipher)
-{
-	int i;
-
-	for (i = 0; ciphers[i].id; i++)
-		if (ciphers[i].id == cipher)
-			return (ciphers[i].ivmax);
-	return (0);
-}
-
-/*
- * XXXX this needs to be set for each alg - and determined from
- * a running card. For now, fake it out - but most of these
- * for real devices should return 1 for the supported key
- * sizes the device can handle.
- */
-static int
-cryptodev_key_length_valid(int cipher, int len)
-{
-	int i;
-
-	for (i = 0; ciphers[i].id; i++)
-		if (ciphers[i].id == cipher)
-			return (ciphers[i].keylen == len);
-	return (0);
-}
-
-/* convert libcrypto nids to cryptodev */
-static int
-cipher_nid_to_cryptodev(int nid)
-{
-	int i;
-
-	for (i = 0; ciphers[i].id; i++)
-		if (ciphers[i].nid == nid)
-			return (ciphers[i].id);
-	return (0);
-}
-
-/*
  * Find out what ciphers /dev/crypto will let us have a session for.
  * XXX note, that some of these openssl doesn't deal with yet!
  * returning them here is harmless, as long as we return NULL
@@ -270,7 +249,7 @@ get_cryptodev_ciphers(const int **cnids)
 		return (0);
 	}
 	memset(&sess, 0, sizeof(sess));
-	sess.key = (caddr_t)"123456781234567812345678";
+	sess.key = (caddr_t)"123456789abcdefghijklmno";
 
 	for (i = 0; ciphers[i].id && count < CRYPTO_ALGORITHM_MAX; i++) {
 		if (ciphers[i].nid == NID_undef)
@@ -282,7 +261,7 @@ get_cryptodev_ciphers(const int **cnids)
 		    ioctl(fd, CIOCFSESSION, &sess.ses) != -1)
 			nids[count++] = ciphers[i].nid;
 	}
-	close(fd);
+	put_dev_crypto(fd);
 
 	if (count > 0)
 		*cnids = nids;
@@ -291,7 +270,7 @@ get_cryptodev_ciphers(const int **cnids)
 	return (count);
 }
 
-#if 0  /* unused */
+#ifdef USE_CRYPTODEV_DIGESTS
 /*
  * Find out what digests /dev/crypto will let us have a session for.
  * XXX note, that some of these openssl doesn't deal with yet!
@@ -310,16 +289,18 @@ get_cryptodev_digests(const int **cnids)
 		return (0);
 	}
 	memset(&sess, 0, sizeof(sess));
+	sess.mackey = (caddr_t)"123456789abcdefghijklmno";
 	for (i = 0; digests[i].id && count < CRYPTO_ALGORITHM_MAX; i++) {
 		if (digests[i].nid == NID_undef)
 			continue;
 		sess.mac = digests[i].id;
+		sess.mackeylen = digests[i].keylen;
 		sess.cipher = 0;
 		if (ioctl(fd, CIOCGSESSION, &sess) != -1 &&
 		    ioctl(fd, CIOCFSESSION, &sess.ses) != -1)
 			nids[count++] = digests[i].nid;
 	}
-	close(fd);
+	put_dev_crypto(fd);
 
 	if (count > 0)
 		*cnids = nids;
@@ -327,8 +308,7 @@ get_cryptodev_digests(const int **cnids)
 		*cnids = NULL;
 	return (count);
 }
-
-#endif
+#endif  /* 0 */
 
 /*
  * Find the useable ciphers|digests from dev/crypto - this is the first
@@ -360,6 +340,9 @@ cryptodev_usable_ciphers(const int **nids)
 static int
 cryptodev_usable_digests(const int **nids)
 {
+#ifdef USE_CRYPTODEV_DIGESTS
+	return (get_cryptodev_digests(nids));
+#else
 	/*
 	 * XXXX just disable all digests for now, because it sucks.
 	 * we need a better way to decide this - i.e. I may not
@@ -374,11 +357,12 @@ cryptodev_usable_digests(const int **nids)
 	 */
 	*nids = NULL;
 	return (0);
+#endif
 }
 
 static int
 cryptodev_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-    const unsigned char *in, unsigned int inl)
+    const unsigned char *in, size_t inl)
 {
 	struct crypt_op cryp;
 	struct dev_crypto_state *state = ctx->cipher_data;
@@ -436,28 +420,32 @@ cryptodev_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 {
 	struct dev_crypto_state *state = ctx->cipher_data;
 	struct session_op *sess = &state->d_sess;
-	int cipher;
+	int cipher = -1, i;
 
-	if ((cipher = cipher_nid_to_cryptodev(ctx->cipher->nid)) == NID_undef)
-		return (0);
-
-	if (ctx->cipher->iv_len > cryptodev_max_iv(cipher))
-		return (0);
+	for (i = 0; ciphers[i].id; i++)
+		if (ctx->cipher->nid == ciphers[i].nid &&
+		    ctx->cipher->iv_len <= ciphers[i].ivmax &&
+		    ctx->key_len == ciphers[i].keylen) {
+			cipher = ciphers[i].id;
+			break;
+		}
 
-	if (!cryptodev_key_length_valid(cipher, ctx->key_len))
+	if (!ciphers[i].id) {
+		state->d_fd = -1;
 		return (0);
+	}
 
 	memset(sess, 0, sizeof(struct session_op));
 
 	if ((state->d_fd = get_dev_crypto()) < 0)
 		return (0);
 
-	sess->key = (char *)key;
+	sess->key = (caddr_t)key;
 	sess->keylen = ctx->key_len;
 	sess->cipher = cipher;
 
 	if (ioctl(state->d_fd, CIOCGSESSION, sess) == -1) {
-		close(state->d_fd);
+		put_dev_crypto(state->d_fd);
 		state->d_fd = -1;
 		return (0);
 	}
@@ -494,7 +482,7 @@ cryptodev_cleanup(EVP_CIPHER_CTX *ctx)
 	} else {
 		ret = 1;
 	}
-	close(state->d_fd);
+	put_dev_crypto(state->d_fd);
 	state->d_fd = -1;
 
 	return (ret);
@@ -505,6 +493,20 @@ cryptodev_cleanup(EVP_CIPHER_CTX *ctx)
  * gets called when libcrypto requests a cipher NID.
  */
 
+/* RC4 */
+const EVP_CIPHER cryptodev_rc4 = {
+	NID_rc4,
+	1, 16, 0,
+	EVP_CIPH_VARIABLE_LENGTH,
+	cryptodev_init_key,
+	cryptodev_cipher,
+	cryptodev_cleanup,
+	sizeof(struct dev_crypto_state),
+	NULL,
+	NULL,
+	NULL
+};
+
 /* DES CBC EVP */
 const EVP_CIPHER cryptodev_des_cbc = {
 	NID_des_cbc,
@@ -572,6 +574,32 @@ const EVP_CIPHER cryptodev_aes_cbc = {
 	NULL
 };
 
+const EVP_CIPHER cryptodev_aes_192_cbc = {
+	NID_aes_192_cbc,
+	16, 24, 16,
+	EVP_CIPH_CBC_MODE,
+	cryptodev_init_key,
+	cryptodev_cipher,
+	cryptodev_cleanup,
+	sizeof(struct dev_crypto_state),
+	EVP_CIPHER_set_asn1_iv,
+	EVP_CIPHER_get_asn1_iv,
+	NULL
+};
+
+const EVP_CIPHER cryptodev_aes_256_cbc = {
+	NID_aes_256_cbc,
+	16, 32, 16,
+	EVP_CIPH_CBC_MODE,
+	cryptodev_init_key,
+	cryptodev_cipher,
+	cryptodev_cleanup,
+	sizeof(struct dev_crypto_state),
+	EVP_CIPHER_set_asn1_iv,
+	EVP_CIPHER_get_asn1_iv,
+	NULL
+};
+
 /*
  * Registered by the ENGINE when used to find out how to deal with
  * a particular NID in the ENGINE. this says what we'll do at the
@@ -585,6 +613,9 @@ cryptodev_engine_ciphers(ENGINE *e, const EVP_CIPHER **cipher,
 		return (cryptodev_usable_ciphers(nids));
 
 	switch (nid) {
+	case NID_rc4:
+		*cipher = &cryptodev_rc4;
+		break;
 	case NID_des_ede3_cbc:
 		*cipher = &cryptodev_3des_cbc;
 		break;
@@ -600,6 +631,12 @@ cryptodev_engine_ciphers(ENGINE *e, const EVP_CIPHER **cipher,
 	case NID_aes_128_cbc:
 		*cipher = &cryptodev_aes_cbc;
 		break;
+	case NID_aes_192_cbc:
+		*cipher = &cryptodev_aes_192_cbc;
+		break;
+	case NID_aes_256_cbc:
+		*cipher = &cryptodev_aes_256_cbc;
+		break;
 	default:
 		*cipher = NULL;
 		break;
@@ -607,6 +644,256 @@ cryptodev_engine_ciphers(ENGINE *e, const EVP_CIPHER **cipher,
 	return (*cipher != NULL);
 }
 
+
+#ifdef USE_CRYPTODEV_DIGESTS
+
+/* convert digest type to cryptodev */
+static int
+digest_nid_to_cryptodev(int nid)
+{
+	int i;
+
+	for (i = 0; digests[i].id; i++)
+		if (digests[i].nid == nid)
+			return (digests[i].id);
+	return (0);
+}
+
+
+static int
+digest_key_length(int nid)
+{
+	int i;
+
+	for (i = 0; digests[i].id; i++)
+		if (digests[i].nid == nid)
+			return digests[i].keylen;
+	return (0);
+}
+
+
+static int cryptodev_digest_init(EVP_MD_CTX *ctx)
+{
+	struct dev_crypto_state *state = ctx->md_data;
+	struct session_op *sess = &state->d_sess;
+	int digest;
+
+	if ((digest = digest_nid_to_cryptodev(ctx->digest->type)) == NID_undef){
+		printf("cryptodev_digest_init: Can't get digest \n");
+		return (0);
+	}
+
+	memset(state, 0, sizeof(struct dev_crypto_state));
+
+	if ((state->d_fd = get_dev_crypto()) < 0) {
+		printf("cryptodev_digest_init: Can't get Dev \n");
+		return (0);
+	}
+
+	sess->mackey = state->dummy_mac_key;
+	sess->mackeylen = digest_key_length(ctx->digest->type);
+	sess->mac = digest;
+
+	if (ioctl(state->d_fd, CIOCGSESSION, sess) < 0) {
+		put_dev_crypto(state->d_fd);
+		state->d_fd = -1;
+		printf("cryptodev_digest_init: Open session failed\n");
+		return (0);
+	}
+
+	return (1);
+}
+
+static int cryptodev_digest_update(EVP_MD_CTX *ctx, const void *data,
+		size_t count)
+{
+	struct crypt_op cryp;
+	struct dev_crypto_state *state = ctx->md_data;
+	struct session_op *sess = &state->d_sess;
+
+	if (!data || state->d_fd < 0) {
+		printf("cryptodev_digest_update: illegal inputs \n");
+		return (0);
+	}
+
+	if (!count) {
+		return (0);
+	}
+
+	if (!(ctx->flags & EVP_MD_CTX_FLAG_ONESHOT)) {
+		/* if application doesn't support one buffer */
+		state->mac_data = OPENSSL_realloc(state->mac_data, state->mac_len + count);
+
+		if (!state->mac_data) {
+			printf("cryptodev_digest_update: realloc failed\n");
+			return (0);
+		}
+
+		memcpy(state->mac_data + state->mac_len, data, count);
+   		state->mac_len += count;
+	
+		return (1);
+	}
+
+	memset(&cryp, 0, sizeof(cryp));
+
+	cryp.ses = sess->ses;
+	cryp.flags = 0;
+	cryp.len = count;
+	cryp.src = (caddr_t) data;
+	cryp.dst = NULL;
+	cryp.mac = (caddr_t) state->digest_res;
+	if (ioctl(state->d_fd, CIOCCRYPT, &cryp) < 0) {
+		printf("cryptodev_digest_update: digest failed\n");
+		return (0);
+	}
+	return (1);
+}
+
+
+static int cryptodev_digest_final(EVP_MD_CTX *ctx, unsigned char *md)
+{
+	struct crypt_op cryp;
+	struct dev_crypto_state *state = ctx->md_data;
+	struct session_op *sess = &state->d_sess;
+
+	int ret = 1;
+
+	if (!md || state->d_fd < 0) {
+		printf("cryptodev_digest_final: illegal input\n");
+		return(0);
+	}
+
+	if (! (ctx->flags & EVP_MD_CTX_FLAG_ONESHOT) ) {
+		/* if application doesn't support one buffer */
+		memset(&cryp, 0, sizeof(cryp));
+		cryp.ses = sess->ses;
+		cryp.flags = 0;
+		cryp.len = state->mac_len;
+		cryp.src = state->mac_data;
+		cryp.dst = NULL;
+		cryp.mac = (caddr_t)md;
+		if (ioctl(state->d_fd, CIOCCRYPT, &cryp) < 0) {
+			printf("cryptodev_digest_final: digest failed\n");
+			return (0);
+		}
+
+		return 1;
+	}
+
+	memcpy(md, state->digest_res, ctx->digest->md_size);
+
+	return (ret);
+}
+
+
+static int cryptodev_digest_cleanup(EVP_MD_CTX *ctx)
+{
+	int ret = 1;
+	struct dev_crypto_state *state = ctx->md_data;
+	struct session_op *sess = &state->d_sess;
+
+	if (state == NULL)
+	  return 0;
+
+	if (state->d_fd < 0) {
+		printf("cryptodev_digest_cleanup: illegal input\n");
+		return (0);
+	}
+
+	if (state->mac_data) {
+		OPENSSL_free(state->mac_data);
+		state->mac_data = NULL;
+		state->mac_len = 0;
+	}
+
+	if (ioctl(state->d_fd, CIOCFSESSION, &sess->ses) < 0) {
+		printf("cryptodev_digest_cleanup: failed to close session\n");
+		ret = 0;
+	} else {
+		ret = 1;
+	}
+	put_dev_crypto(state->d_fd);	
+	state->d_fd = -1;
+
+	return (ret);
+}
+
+static int cryptodev_digest_copy(EVP_MD_CTX *to,const EVP_MD_CTX *from)
+{
+	struct dev_crypto_state *fstate = from->md_data;
+	struct dev_crypto_state *dstate = to->md_data;
+	struct session_op *sess;
+	int digest;
+
+	if (dstate == NULL || fstate == NULL)
+	  return 1;
+
+       	memcpy(dstate, fstate, sizeof(struct dev_crypto_state));
+
+	sess = &dstate->d_sess;
+
+	digest = digest_nid_to_cryptodev(to->digest->type);
+
+	sess->mackey = dstate->dummy_mac_key;
+	sess->mackeylen = digest_key_length(to->digest->type);
+	sess->mac = digest;
+
+	dstate->d_fd = get_dev_crypto();
+
+	if (ioctl(dstate->d_fd, CIOCGSESSION, sess) < 0) {
+		put_dev_crypto(dstate->d_fd);
+		dstate->d_fd = -1;
+		printf("cryptodev_digest_init: Open session failed\n");
+		return (0);
+	}
+
+	if (fstate->mac_len != 0) {
+	        if (fstate->mac_data != NULL)
+	                {
+        		dstate->mac_data = OPENSSL_malloc(fstate->mac_len);
+	        	memcpy(dstate->mac_data, fstate->mac_data, fstate->mac_len);
+           		dstate->mac_len = fstate->mac_len;
+	        	}
+	}
+
+	return 1;
+}
+
+
+const EVP_MD cryptodev_sha1 = {
+	NID_sha1,
+	NID_undef, 
+	SHA_DIGEST_LENGTH, 
+	EVP_MD_FLAG_ONESHOT,
+	cryptodev_digest_init,
+	cryptodev_digest_update,
+	cryptodev_digest_final,
+	cryptodev_digest_copy,
+	cryptodev_digest_cleanup,
+	EVP_PKEY_NULL_method,
+	SHA_CBLOCK,
+	sizeof(struct dev_crypto_state),
+};
+
+const EVP_MD cryptodev_md5 = {
+	NID_md5,
+	NID_undef, 
+	16 /* MD5_DIGEST_LENGTH */, 
+	EVP_MD_FLAG_ONESHOT,
+	cryptodev_digest_init,
+	cryptodev_digest_update,
+	cryptodev_digest_final,
+	cryptodev_digest_copy,
+	cryptodev_digest_cleanup,
+	EVP_PKEY_NULL_method,
+	64 /* MD5_CBLOCK */,
+	sizeof(struct dev_crypto_state),
+};
+
+#endif /* USE_CRYPTODEV_DIGESTS */
+
+
 static int
 cryptodev_engine_digests(ENGINE *e, const EVP_MD **digest,
     const int **nids, int nid)
@@ -615,10 +902,15 @@ cryptodev_engine_digests(ENGINE *e, const EVP_MD **digest,
 		return (cryptodev_usable_digests(nids));
 
 	switch (nid) {
+#ifdef USE_CRYPTODEV_DIGESTS
 	case NID_md5:
-		*digest = NULL; /* need to make a clean md5 critter */
+		*digest = &cryptodev_md5; 
 		break;
+	case NID_sha1:
+		*digest = &cryptodev_sha1;
+ 		break;
 	default:
+#endif /* USE_CRYPTODEV_DIGESTS */
 		*digest = NULL;
 		break;
 	}
@@ -646,8 +938,9 @@ bn2crparam(const BIGNUM *a, struct crparam *crp)
 	b = malloc(bytes);
 	if (b == NULL)
 		return (1);
+	memset(b, 0, bytes);
 
-	crp->crp_p = (char *)b;
+	crp->crp_p = (caddr_t) b;
 	crp->crp_nbits = bits;
 
 	for (i = 0, j = 0; i < a->top; i++) {
@@ -690,7 +983,7 @@ zapparams(struct crypt_kop *kop)
 {
 	int i;
 
-	for (i = 0; i <= kop->crk_iparams + kop->crk_oparams; i++) {
+	for (i = 0; i < kop->crk_iparams + kop->crk_oparams; i++) {
 		if (kop->crk_param[i].crp_p)
 			free(kop->crk_param[i].crp_p);
 		kop->crk_param[i].crp_p = NULL;
@@ -776,8 +1069,9 @@ static int
 cryptodev_rsa_nocrt_mod_exp(BIGNUM *r0, const BIGNUM *I, RSA *rsa, BN_CTX *ctx)
 {
 	int r;
-
+	ctx = BN_CTX_new();
 	r = cryptodev_bn_mod_exp(r0, I, rsa->d, rsa->n, ctx, NULL);
+	BN_CTX_free(ctx);
 	return (r);
 }
 
@@ -1017,7 +1311,7 @@ cryptodev_dh_compute_key(unsigned char *key, const BIGNUM *pub_key, DH *dh)
 		goto err;
 	kop.crk_iparams = 3;
 
-	kop.crk_param[3].crp_p = (char *)key;
+	kop.crk_param[3].crp_p = (caddr_t) key;
 	kop.crk_param[3].crp_nbits = keylen * 8;
 	kop.crk_oparams = 1;
 
@@ -1048,7 +1342,7 @@ static DH_METHOD cryptodev_dh = {
  * but I expect we'll want some options soon.
  */
 static int
-cryptodev_ctrl(ENGINE *e, int cmd, long i, void *p, void (*f)())
+cryptodev_ctrl(ENGINE *e, int cmd, long i, void *p, void (*f)(void))
 {
 #ifdef HAVE_SYSLOG_R
 	struct syslog_data sd = SYSLOG_DATA_INIT;
@@ -1084,11 +1378,11 @@ ENGINE_load_cryptodev(void)
 	 * find out what asymmetric crypto algorithms we support
 	 */
 	if (ioctl(fd, CIOCASYMFEAT, &cryptodev_asymfeat) == -1) {
-		close(fd);
+		put_dev_crypto(fd);
 		ENGINE_free(engine);
 		return;
 	}
-	close(fd);
+	put_dev_crypto(fd);
 
 	if (!ENGINE_set_id(engine, "cryptodev") ||
 	    !ENGINE_set_name(engine, "BSD cryptodev engine") ||
diff --git a/crypto/engine/eng_dyn.c b/crypto/engine/eng_dyn.c
index acb30c34d899..807da7a5eb84 100644
--- a/crypto/engine/eng_dyn.c
+++ b/crypto/engine/eng_dyn.c
@@ -146,14 +146,14 @@ struct st_dynamic_data_ctx
 	 * 'dirs' for loading. Default is to use 'dirs' as a fallback. */
 	int dir_load;
 	/* A stack of directories from which ENGINEs could be loaded */
-	STACK *dirs;
+	STACK_OF(OPENSSL_STRING) *dirs;
 	};
 
 /* This is the "ex_data" index we obtain and reserve for use with our context
  * structure. */
 static int dynamic_ex_data_idx = -1;
 
-static void int_free_str(void *s) { OPENSSL_free(s); }
+static void int_free_str(char *s) { OPENSSL_free(s); }
 /* Because our ex_data element may or may not get allocated depending on whether
  * a "first-use" occurs before the ENGINE is freed, we have a memory leak
  * problem to solve. We can't declare a "new" handler for the ex_data as we
@@ -174,7 +174,7 @@ static void dynamic_data_ctx_free_func(void *parent, void *ptr,
 		if(ctx->engine_id)
 			OPENSSL_free((void*)ctx->engine_id);
 		if(ctx->dirs)
-			sk_pop_free(ctx->dirs, int_free_str);
+			sk_OPENSSL_STRING_pop_free(ctx->dirs, int_free_str);
 		OPENSSL_free(ctx);
 		}
 	}
@@ -203,7 +203,7 @@ static int dynamic_set_data_ctx(ENGINE *e, dynamic_data_ctx **ctx)
 	c->DYNAMIC_F1 = "v_check";
 	c->DYNAMIC_F2 = "bind_engine";
 	c->dir_load = 1;
-	c->dirs = sk_new_null();
+	c->dirs = sk_OPENSSL_STRING_new_null();
 	if(!c->dirs)
 		{
 		ENGINEerr(ENGINE_F_DYNAMIC_SET_DATA_CTX,ERR_R_MALLOC_FAILURE);
@@ -393,7 +393,7 @@ static int dynamic_ctrl(ENGINE *e, int cmd, long i, void *p, void (*f)(void))
 				ERR_R_MALLOC_FAILURE);
 			return 0;
 			}
-		sk_insert(ctx->dirs, tmp_str, -1);
+		sk_OPENSSL_STRING_insert(ctx->dirs, tmp_str, -1);
 		}
 		return 1;
 	default:
@@ -411,11 +411,11 @@ static int int_load(dynamic_data_ctx *ctx)
 				ctx->DYNAMIC_LIBNAME, NULL, 0)) != NULL)
 		return 1;
 	/* If we're not allowed to use 'dirs' or we have none, fail */
-	if(!ctx->dir_load || ((num = sk_num(ctx->dirs)) < 1))
+	if(!ctx->dir_load || (num = sk_OPENSSL_STRING_num(ctx->dirs)) < 1)
 		return 0;
 	for(loop = 0; loop < num; loop++)
 		{
-		const char *s = sk_value(ctx->dirs, loop);
+		const char *s = sk_OPENSSL_STRING_value(ctx->dirs, loop);
 		char *merge = DSO_merge(ctx->dynamic_dso, ctx->DYNAMIC_LIBNAME, s);
 		if(!merge)
 			return 0;
diff --git a/crypto/engine/eng_err.c b/crypto/engine/eng_err.c
index ac74dd127a19..81c70acfa820 100644
--- a/crypto/engine/eng_err.c
+++ b/crypto/engine/eng_err.c
@@ -86,6 +86,8 @@ static ERR_STRING_DATA ENGINE_str_functs[]=
 {ERR_FUNC(ENGINE_F_ENGINE_GET_DEFAULT_TYPE),	"ENGINE_GET_DEFAULT_TYPE"},
 {ERR_FUNC(ENGINE_F_ENGINE_GET_DIGEST),	"ENGINE_get_digest"},
 {ERR_FUNC(ENGINE_F_ENGINE_GET_NEXT),	"ENGINE_get_next"},
+{ERR_FUNC(ENGINE_F_ENGINE_GET_PKEY_ASN1_METH),	"ENGINE_get_pkey_asn1_meth"},
+{ERR_FUNC(ENGINE_F_ENGINE_GET_PKEY_METH),	"ENGINE_get_pkey_meth"},
 {ERR_FUNC(ENGINE_F_ENGINE_GET_PREV),	"ENGINE_get_prev"},
 {ERR_FUNC(ENGINE_F_ENGINE_INIT),	"ENGINE_init"},
 {ERR_FUNC(ENGINE_F_ENGINE_LIST_ADD),	"ENGINE_LIST_ADD"},
@@ -151,6 +153,7 @@ static ERR_STRING_DATA ENGINE_str_reasons[]=
 {ERR_REASON(ENGINE_R_RSA_NOT_IMPLEMENTED),"rsa not implemented"},
 {ERR_REASON(ENGINE_R_UNIMPLEMENTED_CIPHER),"unimplemented cipher"},
 {ERR_REASON(ENGINE_R_UNIMPLEMENTED_DIGEST),"unimplemented digest"},
+{ERR_REASON(ENGINE_R_UNIMPLEMENTED_PUBLIC_KEY_METHOD),"unimplemented public key method"},
 {ERR_REASON(ENGINE_R_VERSION_INCOMPATIBILITY),"version incompatibility"},
 {0,NULL}
 	};
diff --git a/crypto/engine/eng_fat.c b/crypto/engine/eng_fat.c
index 27c1662f6254..789b8d57e58b 100644
--- a/crypto/engine/eng_fat.c
+++ b/crypto/engine/eng_fat.c
@@ -89,6 +89,12 @@ int ENGINE_set_default(ENGINE *e, unsigned int flags)
 #endif
 	if((flags & ENGINE_METHOD_RAND) && !ENGINE_set_default_RAND(e))
 		return 0;
+	if((flags & ENGINE_METHOD_PKEY_METHS)
+				&& !ENGINE_set_default_pkey_meths(e))
+		return 0;
+	if((flags & ENGINE_METHOD_PKEY_ASN1_METHS)
+				&& !ENGINE_set_default_pkey_asn1_meths(e))
+		return 0;
 	return 1;
 	}
 
@@ -115,6 +121,13 @@ static int int_def_cb(const char *alg, int len, void *arg)
 		*pflags |= ENGINE_METHOD_CIPHERS;
 	else if (!strncmp(alg, "DIGESTS", len))
 		*pflags |= ENGINE_METHOD_DIGESTS;
+	else if (!strncmp(alg, "PKEY", len))
+		*pflags |=
+			ENGINE_METHOD_PKEY_METHS|ENGINE_METHOD_PKEY_ASN1_METHS;
+	else if (!strncmp(alg, "PKEY_CRYPTO", len))
+		*pflags |= ENGINE_METHOD_PKEY_METHS;
+	else if (!strncmp(alg, "PKEY_ASN1", len))
+		*pflags |= ENGINE_METHOD_PKEY_ASN1_METHS;
 	else
 		return 0;
 	return 1;
@@ -154,6 +167,7 @@ int ENGINE_register_complete(ENGINE *e)
 	ENGINE_register_ECDSA(e);
 #endif
 	ENGINE_register_RAND(e);
+	ENGINE_register_pkey_meths(e);
 	return 1;
 	}
 
@@ -162,6 +176,7 @@ int ENGINE_register_all_complete(void)
 	ENGINE *e;
 
 	for(e=ENGINE_get_first() ; e ; e=ENGINE_get_next(e))
-		ENGINE_register_complete(e);
+		if (!(e->flags & ENGINE_FLAGS_NO_REGISTER_ALL))
+			ENGINE_register_complete(e);
 	return 1;
 	}
diff --git a/crypto/engine/eng_int.h b/crypto/engine/eng_int.h
index a66f107a44d0..451ef8feb820 100644
--- a/crypto/engine/eng_int.h
+++ b/crypto/engine/eng_int.h
@@ -127,6 +127,8 @@ ENGINE *engine_table_select(ENGINE_TABLE **table, int nid);
 ENGINE *engine_table_select_tmp(ENGINE_TABLE **table, int nid, const char *f, int l);
 #define engine_table_select(t,n) engine_table_select_tmp(t,n,__FILE__,__LINE__)
 #endif
+typedef void (engine_table_doall_cb)(int nid, STACK_OF(ENGINE) *sk, ENGINE *def, void *arg);
+void engine_table_doall(ENGINE_TABLE *table, engine_table_doall_cb *cb, void *arg);
 
 /* Internal versions of API functions that have control over locking. These are
  * used between C files when functionality needs to be shared but the caller may
@@ -143,6 +145,11 @@ void engine_set_all_null(ENGINE *e);
 /* NB: Bitwise OR-able values for the "flags" variable in ENGINE are now exposed
  * in engine.h. */
 
+/* Free up dynamically allocated public key methods associated with ENGINE */
+
+void engine_pkey_meths_free(ENGINE *e);
+void engine_pkey_asn1_meths_free(ENGINE *e);
+
 /* This is a structure for storing implementations of various crypto
  * algorithms and functions. */
 struct engine_st
@@ -160,7 +167,10 @@ struct engine_st
 	ENGINE_CIPHERS_PTR ciphers;
 	/* Digest handling is via this callback */
 	ENGINE_DIGESTS_PTR digests;
-
+	/* Public key handling via this callback */
+	ENGINE_PKEY_METHS_PTR pkey_meths;
+	/* ASN1 public key handling via this callback */
+	ENGINE_PKEY_ASN1_METHS_PTR pkey_asn1_meths;
 
 	ENGINE_GEN_INT_FUNC_PTR	destroy;
 
diff --git a/crypto/engine/eng_lib.c b/crypto/engine/eng_lib.c
index 5815b867f493..18a66646458a 100644
--- a/crypto/engine/eng_lib.c
+++ b/crypto/engine/eng_lib.c
@@ -125,6 +125,9 @@ int engine_free_util(ENGINE *e, int locked)
 		abort();
 		}
 #endif
+	/* Free up any dynamically allocated public key methods */
+	engine_pkey_meths_free(e);
+	engine_pkey_asn1_meths_free(e);
 	/* Give the ENGINE a chance to do any structural cleanup corresponding
 	 * to allocation it did in its constructor (eg. unload error strings) */
 	if(e->destroy)
diff --git a/crypto/engine/eng_list.c b/crypto/engine/eng_list.c
index fa2ab9732ae6..27846edb1e9a 100644
--- a/crypto/engine/eng_list.c
+++ b/crypto/engine/eng_list.c
@@ -336,6 +336,7 @@ static void engine_cpy(ENGINE *dest, const ENGINE *src)
 	dest->store_meth = src->store_meth;
 	dest->ciphers = src->ciphers;
 	dest->digests = src->digests;
+	dest->pkey_meths = src->pkey_meths;
 	dest->destroy = src->destroy;
 	dest->init = src->init;
 	dest->finish = src->finish;
diff --git a/crypto/engine/eng_openssl.c b/crypto/engine/eng_openssl.c
index 7c139ae2efcb..9abb95cc22bf 100644
--- a/crypto/engine/eng_openssl.c
+++ b/crypto/engine/eng_openssl.c
@@ -238,7 +238,7 @@ static int test_rc4_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 	return 1;
 	}
 static int test_rc4_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-		      const unsigned char *in, unsigned int inl)
+		      const unsigned char *in, size_t inl)
 	{
 #ifdef TEST_ENG_OPENSSL_RC4_P_CIPHER
 	fprintf(stderr, "(TEST_ENG_OPENSSL_RC4) test_cipher() called\n");
diff --git a/crypto/engine/eng_padlock.c b/crypto/engine/eng_padlock.c
deleted file mode 100644
index 743558ab3360..000000000000
--- a/crypto/engine/eng_padlock.c
+++ /dev/null
@@ -1,1219 +0,0 @@
-/* 
- * Support for VIA PadLock Advanced Cryptography Engine (ACE)
- * Written by Michal Ludvig <michal@logix.cz>
- *            http://www.logix.cz/michal
- *
- * Big thanks to Andy Polyakov for a help with optimization, 
- * assembler fixes, port to MS Windows and a lot of other 
- * valuable work on this engine!
- */
-
-/* ====================================================================
- * Copyright (c) 1999-2001 The OpenSSL Project.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    licensing@OpenSSL.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
- */
-
-
-#include <stdio.h>
-#include <string.h>
-
-#include <openssl/opensslconf.h>
-#include <openssl/crypto.h>
-#include <openssl/dso.h>
-#include <openssl/engine.h>
-#include <openssl/evp.h>
-#ifndef OPENSSL_NO_AES
-#include <openssl/aes.h>
-#endif
-#include <openssl/rand.h>
-#include <openssl/err.h>
-
-#ifndef OPENSSL_NO_HW
-#ifndef OPENSSL_NO_HW_PADLOCK
-
-/* Attempt to have a single source for both 0.9.7 and 0.9.8 :-) */
-#if (OPENSSL_VERSION_NUMBER >= 0x00908000L)
-#  ifndef OPENSSL_NO_DYNAMIC_ENGINE
-#    define DYNAMIC_ENGINE
-#  endif
-#elif (OPENSSL_VERSION_NUMBER >= 0x00907000L)
-#  ifdef ENGINE_DYNAMIC_SUPPORT
-#    define DYNAMIC_ENGINE
-#  endif
-#else
-#  error "Only OpenSSL >= 0.9.7 is supported"
-#endif
-
-/* VIA PadLock AES is available *ONLY* on some x86 CPUs.
-   Not only that it doesn't exist elsewhere, but it
-   even can't be compiled on other platforms!
- 
-   In addition, because of the heavy use of inline assembler,
-   compiler choice is limited to GCC and Microsoft C. */
-#undef COMPILE_HW_PADLOCK
-#if !defined(I386_ONLY) && !defined(OPENSSL_NO_INLINE_ASM)
-# if (defined(__GNUC__) && (defined(__i386__) || defined(__i386))) || \
-     (defined(_MSC_VER) && defined(_M_IX86))
-#  define COMPILE_HW_PADLOCK
-static ENGINE *ENGINE_padlock (void);
-# endif
-#endif
-
-void ENGINE_load_padlock (void)
-{
-/* On non-x86 CPUs it just returns. */
-#ifdef COMPILE_HW_PADLOCK
-	ENGINE *toadd = ENGINE_padlock ();
-	if (!toadd) return;
-	ENGINE_add (toadd);
-	ENGINE_free (toadd);
-	ERR_clear_error ();
-#endif
-}
-
-#ifdef COMPILE_HW_PADLOCK
-/* We do these includes here to avoid header problems on platforms that
-   do not have the VIA padlock anyway... */
-#ifdef _MSC_VER
-# include <malloc.h>
-# define alloca _alloca
-#elif defined(NETWARE_CLIB) && defined(__GNUC__)
-  void *alloca(size_t);
-# define alloca(s) __builtin_alloca(s)
-#else
-# include <stdlib.h>
-#endif
-
-/* Function for ENGINE detection and control */
-static int padlock_available(void);
-static int padlock_init(ENGINE *e);
-
-/* RNG Stuff */
-static RAND_METHOD padlock_rand;
-
-/* Cipher Stuff */
-#ifndef OPENSSL_NO_AES
-static int padlock_ciphers(ENGINE *e, const EVP_CIPHER **cipher, const int **nids, int nid);
-#endif
-
-/* Engine names */
-static const char *padlock_id = "padlock";
-static char padlock_name[100];
-
-/* Available features */
-static int padlock_use_ace = 0;	/* Advanced Cryptography Engine */
-static int padlock_use_rng = 0;	/* Random Number Generator */
-#ifndef OPENSSL_NO_AES
-static int padlock_aes_align_required = 1;
-#endif
-
-/* ===== Engine "management" functions ===== */
-
-/* Prepare the ENGINE structure for registration */
-static int
-padlock_bind_helper(ENGINE *e)
-{
-	/* Check available features */
-	padlock_available();
-
-#if 1	/* disable RNG for now, see commentary in vicinity of RNG code */
-	padlock_use_rng=0;
-#endif
-
-	/* Generate a nice engine name with available features */
-	BIO_snprintf(padlock_name, sizeof(padlock_name),
-		"VIA PadLock (%s, %s)", 
-		 padlock_use_rng ? "RNG" : "no-RNG",
-		 padlock_use_ace ? "ACE" : "no-ACE");
-
-	/* Register everything or return with an error */ 
-	if (!ENGINE_set_id(e, padlock_id) ||
-	    !ENGINE_set_name(e, padlock_name) ||
-
-	    !ENGINE_set_init_function(e, padlock_init) ||
-#ifndef OPENSSL_NO_AES
-	    (padlock_use_ace && !ENGINE_set_ciphers (e, padlock_ciphers)) ||
-#endif
-	    (padlock_use_rng && !ENGINE_set_RAND (e, &padlock_rand))) {
-		return 0;
-	}
-
-	/* Everything looks good */
-	return 1;
-}
-
-/* Constructor */
-static ENGINE *
-ENGINE_padlock(void)
-{
-	ENGINE *eng = ENGINE_new();
-
-	if (!eng) {
-		return NULL;
-	}
-
-	if (!padlock_bind_helper(eng)) {
-		ENGINE_free(eng);
-		return NULL;
-	}
-
-	return eng;
-}
-
-/* Check availability of the engine */
-static int
-padlock_init(ENGINE *e)
-{
-	return (padlock_use_rng || padlock_use_ace);
-}
-
-/* This stuff is needed if this ENGINE is being compiled into a self-contained
- * shared-library.
- */
-#ifdef DYNAMIC_ENGINE
-static int
-padlock_bind_fn(ENGINE *e, const char *id)
-{
-	if (id && (strcmp(id, padlock_id) != 0)) {
-		return 0;
-	}
-
-	if (!padlock_bind_helper(e))  {
-		return 0;
-	}
-
-	return 1;
-}
-
-IMPLEMENT_DYNAMIC_CHECK_FN ()
-IMPLEMENT_DYNAMIC_BIND_FN (padlock_bind_fn)
-#endif /* DYNAMIC_ENGINE */
-
-/* ===== Here comes the "real" engine ===== */
-
-#ifndef OPENSSL_NO_AES
-/* Some AES-related constants */
-#define AES_BLOCK_SIZE		16
-#define AES_KEY_SIZE_128	16
-#define AES_KEY_SIZE_192	24
-#define AES_KEY_SIZE_256	32
-
-/* Here we store the status information relevant to the 
-   current context. */
-/* BIG FAT WARNING:
- * 	Inline assembler in PADLOCK_XCRYPT_ASM()
- * 	depends on the order of items in this structure.
- * 	Don't blindly modify, reorder, etc!
- */
-struct padlock_cipher_data
-{
-	unsigned char iv[AES_BLOCK_SIZE];	/* Initialization vector */
-	union {	unsigned int pad[4];
-		struct {
-			int rounds:4;
-			int dgst:1;	/* n/a in C3 */
-			int align:1;	/* n/a in C3 */
-			int ciphr:1;	/* n/a in C3 */
-			unsigned int keygen:1;
-			int interm:1;
-			unsigned int encdec:1;
-			int ksize:2;
-		} b;
-	} cword;		/* Control word */
-	AES_KEY ks;		/* Encryption key */
-};
-
-/*
- * Essentially this variable belongs in thread local storage.
- * Having this variable global on the other hand can only cause
- * few bogus key reloads [if any at all on single-CPU system],
- * so we accept the penatly...
- */
-static volatile struct padlock_cipher_data *padlock_saved_context;
-#endif
-
-/*
- * =======================================================
- * Inline assembler section(s).
- * =======================================================
- * Order of arguments is chosen to facilitate Windows port
- * using __fastcall calling convention. If you wish to add
- * more routines, keep in mind that first __fastcall
- * argument is passed in %ecx and second - in %edx.
- * =======================================================
- */
-#if defined(__GNUC__) && __GNUC__>=2
-/*
- * As for excessive "push %ebx"/"pop %ebx" found all over.
- * When generating position-independent code GCC won't let
- * us use "b" in assembler templates nor even respect "ebx"
- * in "clobber description." Therefore the trouble...
- */
-
-/* Helper function - check if a CPUID instruction
-   is available on this CPU */
-static int
-padlock_insn_cpuid_available(void)
-{
-	int result = -1;
-
-	/* We're checking if the bit #21 of EFLAGS 
-	   can be toggled. If yes = CPUID is available. */
-	asm volatile (
-		"pushf\n"
-		"popl %%eax\n"
-		"xorl $0x200000, %%eax\n"
-		"movl %%eax, %%ecx\n"
-		"andl $0x200000, %%ecx\n"
-		"pushl %%eax\n"
-		"popf\n"
-		"pushf\n"
-		"popl %%eax\n"
-		"andl $0x200000, %%eax\n"
-		"xorl %%eax, %%ecx\n"
-		"movl %%ecx, %0\n"
-		: "=r" (result) : : "eax", "ecx");
-	
-	return (result == 0);
-}
-
-/* Load supported features of the CPU to see if
-   the PadLock is available. */
-static int
-padlock_available(void)
-{
-	char vendor_string[16];
-	unsigned int eax, edx;
-
-	/* First check if the CPUID instruction is available at all... */
-	if (! padlock_insn_cpuid_available())
-		return 0;
-
-	/* Are we running on the Centaur (VIA) CPU? */
-	eax = 0x00000000;
-	vendor_string[12] = 0;
-	asm volatile (
-		"pushl	%%ebx\n"
-		"cpuid\n"
-		"movl	%%ebx,(%%edi)\n"
-		"movl	%%edx,4(%%edi)\n"
-		"movl	%%ecx,8(%%edi)\n"
-		"popl	%%ebx"
-		: "+a"(eax) : "D"(vendor_string) : "ecx", "edx");
-	if (strcmp(vendor_string, "CentaurHauls") != 0)
-		return 0;
-
-	/* Check for Centaur Extended Feature Flags presence */
-	eax = 0xC0000000;
-	asm volatile ("pushl %%ebx; cpuid; popl	%%ebx"
-		: "+a"(eax) : : "ecx", "edx");
-	if (eax < 0xC0000001)
-		return 0;
-
-	/* Read the Centaur Extended Feature Flags */
-	eax = 0xC0000001;
-	asm volatile ("pushl %%ebx; cpuid; popl %%ebx"
-		: "+a"(eax), "=d"(edx) : : "ecx");
-
-	/* Fill up some flags */
-	padlock_use_ace = ((edx & (0x3<<6)) == (0x3<<6));
-	padlock_use_rng = ((edx & (0x3<<2)) == (0x3<<2));
-
-	return padlock_use_ace + padlock_use_rng;
-}
-
-#ifndef OPENSSL_NO_AES
-/* Our own htonl()/ntohl() */
-static inline void
-padlock_bswapl(AES_KEY *ks)
-{
-	size_t i = sizeof(ks->rd_key)/sizeof(ks->rd_key[0]);
-	unsigned int *key = ks->rd_key;
-
-	while (i--) {
-		asm volatile ("bswapl %0" : "+r"(*key));
-		key++;
-	}
-}
-#endif
-
-/* Force key reload from memory to the CPU microcode.
-   Loading EFLAGS from the stack clears EFLAGS[30] 
-   which does the trick. */
-static inline void
-padlock_reload_key(void)
-{
-	asm volatile ("pushfl; popfl");
-}
-
-#ifndef OPENSSL_NO_AES
-/*
- * This is heuristic key context tracing. At first one
- * believes that one should use atomic swap instructions,
- * but it's not actually necessary. Point is that if
- * padlock_saved_context was changed by another thread
- * after we've read it and before we compare it with cdata,
- * our key *shall* be reloaded upon thread context switch
- * and we are therefore set in either case...
- */
-static inline void
-padlock_verify_context(struct padlock_cipher_data *cdata)
-{
-	asm volatile (
-	"pushfl\n"
-"	btl	$30,(%%esp)\n"
-"	jnc	1f\n"
-"	cmpl	%2,%1\n"
-"	je	1f\n"
-"	popfl\n"
-"	subl	$4,%%esp\n"
-"1:	addl	$4,%%esp\n"
-"	movl	%2,%0"
-	:"+m"(padlock_saved_context)
-	: "r"(padlock_saved_context), "r"(cdata) : "cc");
-}
-
-/* Template for padlock_xcrypt_* modes */
-/* BIG FAT WARNING: 
- * 	The offsets used with 'leal' instructions
- * 	describe items of the 'padlock_cipher_data'
- * 	structure.
- */
-#define PADLOCK_XCRYPT_ASM(name,rep_xcrypt)	\
-static inline void *name(size_t cnt,		\
-	struct padlock_cipher_data *cdata,	\
-	void *out, const void *inp) 		\
-{	void *iv; 				\
-	asm volatile ( "pushl	%%ebx\n"	\
-		"	leal	16(%0),%%edx\n"	\
-		"	leal	32(%0),%%ebx\n"	\
-			rep_xcrypt "\n"		\
-		"	popl	%%ebx"		\
-		: "=a"(iv), "=c"(cnt), "=D"(out), "=S"(inp) \
-		: "0"(cdata), "1"(cnt), "2"(out), "3"(inp)  \
-		: "edx", "cc", "memory");	\
-	return iv;				\
-}
-
-/* Generate all functions with appropriate opcodes */
-PADLOCK_XCRYPT_ASM(padlock_xcrypt_ecb, ".byte 0xf3,0x0f,0xa7,0xc8")	/* rep xcryptecb */
-PADLOCK_XCRYPT_ASM(padlock_xcrypt_cbc, ".byte 0xf3,0x0f,0xa7,0xd0")	/* rep xcryptcbc */
-PADLOCK_XCRYPT_ASM(padlock_xcrypt_cfb, ".byte 0xf3,0x0f,0xa7,0xe0")	/* rep xcryptcfb */
-PADLOCK_XCRYPT_ASM(padlock_xcrypt_ofb, ".byte 0xf3,0x0f,0xa7,0xe8")	/* rep xcryptofb */
-#endif
-
-/* The RNG call itself */
-static inline unsigned int
-padlock_xstore(void *addr, unsigned int edx_in)
-{
-	unsigned int eax_out;
-
-	asm volatile (".byte 0x0f,0xa7,0xc0"	/* xstore */
-	    : "=a"(eax_out),"=m"(*(unsigned *)addr)
-	    : "D"(addr), "d" (edx_in)
-	    );
-
-	return eax_out;
-}
-
-/* Why not inline 'rep movsd'? I failed to find information on what
- * value in Direction Flag one can expect and consequently have to
- * apply "better-safe-than-sorry" approach and assume "undefined."
- * I could explicitly clear it and restore the original value upon
- * return from padlock_aes_cipher, but it's presumably too much
- * trouble for too little gain...
- *
- * In case you wonder 'rep xcrypt*' instructions above are *not*
- * affected by the Direction Flag and pointers advance toward
- * larger addresses unconditionally.
- */ 
-static inline unsigned char *
-padlock_memcpy(void *dst,const void *src,size_t n)
-{
-	long       *d=dst;
-	const long *s=src;
-
-	n /= sizeof(*d);
-	do { *d++ = *s++; } while (--n);
-
-	return dst;
-}
-
-#elif defined(_MSC_VER)
-/*
- * Unlike GCC these are real functions. In order to minimize impact
- * on performance we adhere to __fastcall calling convention in
- * order to get two first arguments passed through %ecx and %edx.
- * Which kind of suits very well, as instructions in question use
- * both %ecx and %edx as input:-)
- */
-#define REP_XCRYPT(code)		\
-	_asm _emit 0xf3			\
-	_asm _emit 0x0f _asm _emit 0xa7	\
-	_asm _emit code
-
-/* BIG FAT WARNING: 
- * 	The offsets used with 'lea' instructions
- * 	describe items of the 'padlock_cipher_data'
- * 	structure.
- */
-#define PADLOCK_XCRYPT_ASM(name,code)	\
-static void * __fastcall 		\
-	name (size_t cnt, void *cdata,	\
-	void *outp, const void *inp)	\
-{	_asm	mov	eax,edx		\
-	_asm	lea	edx,[eax+16]	\
-	_asm	lea	ebx,[eax+32]	\
-	_asm	mov	edi,outp	\
-	_asm	mov	esi,inp		\
-	REP_XCRYPT(code)		\
-}
-
-PADLOCK_XCRYPT_ASM(padlock_xcrypt_ecb,0xc8)
-PADLOCK_XCRYPT_ASM(padlock_xcrypt_cbc,0xd0)
-PADLOCK_XCRYPT_ASM(padlock_xcrypt_cfb,0xe0)
-PADLOCK_XCRYPT_ASM(padlock_xcrypt_ofb,0xe8)
-
-static int __fastcall
-padlock_xstore(void *outp,unsigned int code)
-{	_asm	mov	edi,ecx
-	_asm _emit 0x0f _asm _emit 0xa7 _asm _emit 0xc0
-}
-
-static void __fastcall
-padlock_reload_key(void)
-{	_asm pushfd _asm popfd		}
-
-static void __fastcall
-padlock_verify_context(void *cdata)
-{	_asm	{
-		pushfd
-		bt	DWORD PTR[esp],30
-		jnc	skip
-		cmp	ecx,padlock_saved_context
-		je	skip
-		popfd
-		sub	esp,4
-	skip:	add	esp,4
-		mov	padlock_saved_context,ecx
-		}
-}
-
-static int
-padlock_available(void)
-{	_asm	{
-		pushfd
-		pop	eax
-		mov	ecx,eax
-		xor	eax,1<<21
-		push	eax
-		popfd
-		pushfd
-		pop	eax
-		xor	eax,ecx
-		bt	eax,21
-		jnc	noluck
-		mov	eax,0
-		cpuid
-		xor	eax,eax
-		cmp	ebx,'tneC'
-		jne	noluck
-		cmp	edx,'Hrua'
-		jne	noluck
-		cmp	ecx,'slua'
-		jne	noluck
-		mov	eax,0xC0000000
-		cpuid
-		mov	edx,eax
-		xor	eax,eax
-		cmp	edx,0xC0000001
-		jb	noluck
-		mov	eax,0xC0000001
-		cpuid
-		xor	eax,eax
-		bt	edx,6
-		jnc	skip_a
-		bt	edx,7
-		jnc	skip_a
-		mov	padlock_use_ace,1
-		inc	eax
-	skip_a:	bt	edx,2
-		jnc	skip_r
-		bt	edx,3
-		jnc	skip_r
-		mov	padlock_use_rng,1
-		inc	eax
-	skip_r:
-	noluck:
-		}
-}
-
-static void __fastcall
-padlock_bswapl(void *key)
-{	_asm	{
-		pushfd
-		cld
-		mov	esi,ecx
-		mov	edi,ecx
-		mov	ecx,60
-	up:	lodsd
-		bswap	eax
-		stosd
-		loop	up
-		popfd
-		}
-}
-
-/* MS actually specifies status of Direction Flag and compiler even
- * manages to compile following as 'rep movsd' all by itself...
- */
-#define padlock_memcpy(o,i,n) ((unsigned char *)memcpy((o),(i),(n)&~3U))
-#endif
-
-/* ===== AES encryption/decryption ===== */
-#ifndef OPENSSL_NO_AES
-
-#if defined(NID_aes_128_cfb128) && ! defined (NID_aes_128_cfb)
-#define NID_aes_128_cfb	NID_aes_128_cfb128
-#endif
-
-#if defined(NID_aes_128_ofb128) && ! defined (NID_aes_128_ofb)
-#define NID_aes_128_ofb	NID_aes_128_ofb128
-#endif
-
-#if defined(NID_aes_192_cfb128) && ! defined (NID_aes_192_cfb)
-#define NID_aes_192_cfb	NID_aes_192_cfb128
-#endif
-
-#if defined(NID_aes_192_ofb128) && ! defined (NID_aes_192_ofb)
-#define NID_aes_192_ofb	NID_aes_192_ofb128
-#endif
-
-#if defined(NID_aes_256_cfb128) && ! defined (NID_aes_256_cfb)
-#define NID_aes_256_cfb	NID_aes_256_cfb128
-#endif
-
-#if defined(NID_aes_256_ofb128) && ! defined (NID_aes_256_ofb)
-#define NID_aes_256_ofb	NID_aes_256_ofb128
-#endif
-
-/* List of supported ciphers. */
-static int padlock_cipher_nids[] = {
-	NID_aes_128_ecb,
-	NID_aes_128_cbc,
-	NID_aes_128_cfb,
-	NID_aes_128_ofb,
-
-	NID_aes_192_ecb,
-	NID_aes_192_cbc,
-	NID_aes_192_cfb,
-	NID_aes_192_ofb,
-
-	NID_aes_256_ecb,
-	NID_aes_256_cbc,
-	NID_aes_256_cfb,
-	NID_aes_256_ofb,
-};
-static int padlock_cipher_nids_num = (sizeof(padlock_cipher_nids)/
-				      sizeof(padlock_cipher_nids[0]));
-
-/* Function prototypes ... */
-static int padlock_aes_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
-				const unsigned char *iv, int enc);
-static int padlock_aes_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			      const unsigned char *in, size_t nbytes);
-
-#define NEAREST_ALIGNED(ptr) ( (unsigned char *)(ptr) +		\
-	( (0x10 - ((size_t)(ptr) & 0x0F)) & 0x0F )	)
-#define ALIGNED_CIPHER_DATA(ctx) ((struct padlock_cipher_data *)\
-	NEAREST_ALIGNED(ctx->cipher_data))
-
-#define EVP_CIPHER_block_size_ECB	AES_BLOCK_SIZE
-#define EVP_CIPHER_block_size_CBC	AES_BLOCK_SIZE
-#define EVP_CIPHER_block_size_OFB	1
-#define EVP_CIPHER_block_size_CFB	1
-
-/* Declaring so many ciphers by hand would be a pain.
-   Instead introduce a bit of preprocessor magic :-) */
-#define	DECLARE_AES_EVP(ksize,lmode,umode)	\
-static const EVP_CIPHER padlock_aes_##ksize##_##lmode = {	\
-	NID_aes_##ksize##_##lmode,		\
-	EVP_CIPHER_block_size_##umode,	\
-	AES_KEY_SIZE_##ksize,		\
-	AES_BLOCK_SIZE,			\
-	0 | EVP_CIPH_##umode##_MODE,	\
-	padlock_aes_init_key,		\
-	padlock_aes_cipher,		\
-	NULL,				\
-	sizeof(struct padlock_cipher_data) + 16,	\
-	EVP_CIPHER_set_asn1_iv,		\
-	EVP_CIPHER_get_asn1_iv,		\
-	NULL,				\
-	NULL				\
-}
-
-DECLARE_AES_EVP(128,ecb,ECB);
-DECLARE_AES_EVP(128,cbc,CBC);
-DECLARE_AES_EVP(128,cfb,CFB);
-DECLARE_AES_EVP(128,ofb,OFB);
-
-DECLARE_AES_EVP(192,ecb,ECB);
-DECLARE_AES_EVP(192,cbc,CBC);
-DECLARE_AES_EVP(192,cfb,CFB);
-DECLARE_AES_EVP(192,ofb,OFB);
-
-DECLARE_AES_EVP(256,ecb,ECB);
-DECLARE_AES_EVP(256,cbc,CBC);
-DECLARE_AES_EVP(256,cfb,CFB);
-DECLARE_AES_EVP(256,ofb,OFB);
-
-static int
-padlock_ciphers (ENGINE *e, const EVP_CIPHER **cipher, const int **nids, int nid)
-{
-	/* No specific cipher => return a list of supported nids ... */
-	if (!cipher) {
-		*nids = padlock_cipher_nids;
-		return padlock_cipher_nids_num;
-	}
-
-	/* ... or the requested "cipher" otherwise */
-	switch (nid) {
-	  case NID_aes_128_ecb:
-	    *cipher = &padlock_aes_128_ecb;
-	    break;
-	  case NID_aes_128_cbc:
-	    *cipher = &padlock_aes_128_cbc;
-	    break;
-	  case NID_aes_128_cfb:
-	    *cipher = &padlock_aes_128_cfb;
-	    break;
-	  case NID_aes_128_ofb:
-	    *cipher = &padlock_aes_128_ofb;
-	    break;
-
-	  case NID_aes_192_ecb:
-	    *cipher = &padlock_aes_192_ecb;
-	    break;
-	  case NID_aes_192_cbc:
-	    *cipher = &padlock_aes_192_cbc;
-	    break;
-	  case NID_aes_192_cfb:
-	    *cipher = &padlock_aes_192_cfb;
-	    break;
-	  case NID_aes_192_ofb:
-	    *cipher = &padlock_aes_192_ofb;
-	    break;
-
-	  case NID_aes_256_ecb:
-	    *cipher = &padlock_aes_256_ecb;
-	    break;
-	  case NID_aes_256_cbc:
-	    *cipher = &padlock_aes_256_cbc;
-	    break;
-	  case NID_aes_256_cfb:
-	    *cipher = &padlock_aes_256_cfb;
-	    break;
-	  case NID_aes_256_ofb:
-	    *cipher = &padlock_aes_256_ofb;
-	    break;
-
-	  default:
-	    /* Sorry, we don't support this NID */
-	    *cipher = NULL;
-	    return 0;
-	}
-
-	return 1;
-}
-
-/* Prepare the encryption key for PadLock usage */
-static int
-padlock_aes_init_key (EVP_CIPHER_CTX *ctx, const unsigned char *key,
-		      const unsigned char *iv, int enc)
-{
-	struct padlock_cipher_data *cdata;
-	int key_len = EVP_CIPHER_CTX_key_length(ctx) * 8;
-
-	if (key==NULL) return 0;	/* ERROR */
-
-	cdata = ALIGNED_CIPHER_DATA(ctx);
-	memset(cdata, 0, sizeof(struct padlock_cipher_data));
-
-	/* Prepare Control word. */
-	if (EVP_CIPHER_CTX_mode(ctx) == EVP_CIPH_OFB_MODE)
-		cdata->cword.b.encdec = 0;
-	else
-		cdata->cword.b.encdec = (ctx->encrypt == 0);
-	cdata->cword.b.rounds = 10 + (key_len - 128) / 32;
-	cdata->cword.b.ksize = (key_len - 128) / 64;
-
-	switch(key_len) {
-		case 128:
-			/* PadLock can generate an extended key for
-			   AES128 in hardware */
-			memcpy(cdata->ks.rd_key, key, AES_KEY_SIZE_128);
-			cdata->cword.b.keygen = 0;
-			break;
-
-		case 192:
-		case 256:
-			/* Generate an extended AES key in software.
-			   Needed for AES192/AES256 */
-			/* Well, the above applies to Stepping 8 CPUs
-			   and is listed as hardware errata. They most
-			   likely will fix it at some point and then
-			   a check for stepping would be due here. */
-			if (EVP_CIPHER_CTX_mode(ctx) == EVP_CIPH_CFB_MODE ||
-			    EVP_CIPHER_CTX_mode(ctx) == EVP_CIPH_OFB_MODE ||
-			    enc)
-				AES_set_encrypt_key(key, key_len, &cdata->ks);
-			else
-				AES_set_decrypt_key(key, key_len, &cdata->ks);
-#ifndef AES_ASM
-			/* OpenSSL C functions use byte-swapped extended key. */
-			padlock_bswapl(&cdata->ks);
-#endif
-			cdata->cword.b.keygen = 1;
-			break;
-
-		default:
-			/* ERROR */
-			return 0;
-	}
-
-	/*
-	 * This is done to cover for cases when user reuses the
-	 * context for new key. The catch is that if we don't do
-	 * this, padlock_eas_cipher might proceed with old key...
-	 */
-	padlock_reload_key ();
-
-	return 1;
-}
-
-/* 
- * Simplified version of padlock_aes_cipher() used when
- * 1) both input and output buffers are at aligned addresses.
- * or when
- * 2) running on a newer CPU that doesn't require aligned buffers.
- */
-static int
-padlock_aes_cipher_omnivorous(EVP_CIPHER_CTX *ctx, unsigned char *out_arg,
-		const unsigned char *in_arg, size_t nbytes)
-{
-	struct padlock_cipher_data *cdata;
-	void  *iv;
-
-	cdata = ALIGNED_CIPHER_DATA(ctx);
-	padlock_verify_context(cdata);
-
-	switch (EVP_CIPHER_CTX_mode(ctx)) {
-	case EVP_CIPH_ECB_MODE:
-		padlock_xcrypt_ecb(nbytes/AES_BLOCK_SIZE, cdata, out_arg, in_arg);
-		break;
-
-	case EVP_CIPH_CBC_MODE:
-		memcpy(cdata->iv, ctx->iv, AES_BLOCK_SIZE);
-		iv = padlock_xcrypt_cbc(nbytes/AES_BLOCK_SIZE, cdata, out_arg, in_arg);
-		memcpy(ctx->iv, iv, AES_BLOCK_SIZE);
-		break;
-
-	case EVP_CIPH_CFB_MODE:
-		memcpy(cdata->iv, ctx->iv, AES_BLOCK_SIZE);
-		iv = padlock_xcrypt_cfb(nbytes/AES_BLOCK_SIZE, cdata, out_arg, in_arg);
-		memcpy(ctx->iv, iv, AES_BLOCK_SIZE);
-		break;
-
-	case EVP_CIPH_OFB_MODE:
-		memcpy(cdata->iv, ctx->iv, AES_BLOCK_SIZE);
-		padlock_xcrypt_ofb(nbytes/AES_BLOCK_SIZE, cdata, out_arg, in_arg);
-		memcpy(ctx->iv, cdata->iv, AES_BLOCK_SIZE);
-		break;
-
-	default:
-		return 0;
-	}
-
-	memset(cdata->iv, 0, AES_BLOCK_SIZE);
-
-	return 1;
-}
-
-#ifndef  PADLOCK_CHUNK
-# define PADLOCK_CHUNK	512	/* Must be a power of 2 larger than 16 */
-#endif
-#if PADLOCK_CHUNK<16 || PADLOCK_CHUNK&(PADLOCK_CHUNK-1)
-# error "insane PADLOCK_CHUNK..."
-#endif
-
-/* Re-align the arguments to 16-Bytes boundaries and run the 
-   encryption function itself. This function is not AES-specific. */
-static int
-padlock_aes_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out_arg,
-		   const unsigned char *in_arg, size_t nbytes)
-{
-	struct padlock_cipher_data *cdata;
-	const  void *inp;
-	unsigned char  *out;
-	void  *iv;
-	int    inp_misaligned, out_misaligned, realign_in_loop;
-	size_t chunk, allocated=0;
-
-	/* ctx->num is maintained in byte-oriented modes,
-	   such as CFB and OFB... */
-	if ((chunk = ctx->num)) { /* borrow chunk variable */
-		unsigned char *ivp=ctx->iv;
-
-		switch (EVP_CIPHER_CTX_mode(ctx)) {
-		case EVP_CIPH_CFB_MODE:
-			if (chunk >= AES_BLOCK_SIZE)
-				return 0; /* bogus value */
-
-			if (ctx->encrypt)
-				while (chunk<AES_BLOCK_SIZE && nbytes!=0) {
-					ivp[chunk] = *(out_arg++) = *(in_arg++) ^ ivp[chunk];
-					chunk++, nbytes--;
-				}
-			else	while (chunk<AES_BLOCK_SIZE && nbytes!=0) {
-					unsigned char c = *(in_arg++);
-					*(out_arg++) = c ^ ivp[chunk];
-					ivp[chunk++] = c, nbytes--;
-				}
-
-			ctx->num = chunk%AES_BLOCK_SIZE;
-			break;
-		case EVP_CIPH_OFB_MODE:
-			if (chunk >= AES_BLOCK_SIZE)
-				return 0; /* bogus value */
-
-			while (chunk<AES_BLOCK_SIZE && nbytes!=0) {
-				*(out_arg++) = *(in_arg++) ^ ivp[chunk];
-				chunk++, nbytes--;
-			}
-
-			ctx->num = chunk%AES_BLOCK_SIZE;
-			break;
-		}
-	}
-
-	if (nbytes == 0)
-		return 1;
-#if 0
-	if (nbytes % AES_BLOCK_SIZE)
-		return 0; /* are we expected to do tail processing? */
-#else
-	/* nbytes is always multiple of AES_BLOCK_SIZE in ECB and CBC
-	   modes and arbitrary value in byte-oriented modes, such as
-	   CFB and OFB... */
-#endif
-
-	/* VIA promises CPUs that won't require alignment in the future.
-	   For now padlock_aes_align_required is initialized to 1 and
-	   the condition is never met... */
-	/* C7 core is capable to manage unaligned input in non-ECB[!]
-	   mode, but performance penalties appear to be approximately
-	   same as for software alignment below or ~3x. They promise to
-	   improve it in the future, but for now we can just as well
-	   pretend that it can only handle aligned input... */
-	if (!padlock_aes_align_required && (nbytes%AES_BLOCK_SIZE)==0)
-		return padlock_aes_cipher_omnivorous(ctx, out_arg, in_arg, nbytes);
-
-	inp_misaligned = (((size_t)in_arg) & 0x0F);
-	out_misaligned = (((size_t)out_arg) & 0x0F);
-
-	/* Note that even if output is aligned and input not,
-	 * I still prefer to loop instead of copy the whole
-	 * input and then encrypt in one stroke. This is done
-	 * in order to improve L1 cache utilization... */
-	realign_in_loop = out_misaligned|inp_misaligned;
-
-	if (!realign_in_loop && (nbytes%AES_BLOCK_SIZE)==0)
-		return padlock_aes_cipher_omnivorous(ctx, out_arg, in_arg, nbytes);
-
-	/* this takes one "if" out of the loops */
-	chunk  = nbytes;
-	chunk %= PADLOCK_CHUNK;
-	if (chunk==0) chunk = PADLOCK_CHUNK;
-
-	if (out_misaligned) {
-		/* optmize for small input */
-		allocated = (chunk<nbytes?PADLOCK_CHUNK:nbytes);
-		out = alloca(0x10 + allocated);
-		out = NEAREST_ALIGNED(out);
-	}
-	else
-		out = out_arg;
-
-	cdata = ALIGNED_CIPHER_DATA(ctx);
-	padlock_verify_context(cdata);
-
-	switch (EVP_CIPHER_CTX_mode(ctx)) {
-	case EVP_CIPH_ECB_MODE:
-		do	{
-			if (inp_misaligned)
-				inp = padlock_memcpy(out, in_arg, chunk);
-			else
-				inp = in_arg;
-			in_arg += chunk;
-
-			padlock_xcrypt_ecb(chunk/AES_BLOCK_SIZE, cdata, out, inp);
-
-			if (out_misaligned)
-				out_arg = padlock_memcpy(out_arg, out, chunk) + chunk;
-			else
-				out     = out_arg+=chunk;
-
-			nbytes -= chunk;
-			chunk   = PADLOCK_CHUNK;
-		} while (nbytes);
-		break;
-
-	case EVP_CIPH_CBC_MODE:
-		memcpy(cdata->iv, ctx->iv, AES_BLOCK_SIZE);
-		goto cbc_shortcut;
-		do	{
-			if (iv != cdata->iv)
-				memcpy(cdata->iv, iv, AES_BLOCK_SIZE);
-			chunk = PADLOCK_CHUNK;
-		cbc_shortcut: /* optimize for small input */
-			if (inp_misaligned)
-				inp = padlock_memcpy(out, in_arg, chunk);
-			else
-				inp = in_arg;
-			in_arg += chunk;
-
-			iv = padlock_xcrypt_cbc(chunk/AES_BLOCK_SIZE, cdata, out, inp);
-
-			if (out_misaligned)
-				out_arg = padlock_memcpy(out_arg, out, chunk) + chunk;
-			else
-				out     = out_arg+=chunk;
-
-		} while (nbytes -= chunk);
-		memcpy(ctx->iv, iv, AES_BLOCK_SIZE);
-		break;
-
-	case EVP_CIPH_CFB_MODE:
-		memcpy (iv = cdata->iv, ctx->iv, AES_BLOCK_SIZE);
-		chunk &= ~(AES_BLOCK_SIZE-1);
-		if (chunk)	goto cfb_shortcut;
-		else		goto cfb_skiploop;
-		do	{
-			if (iv != cdata->iv)
-				memcpy(cdata->iv, iv, AES_BLOCK_SIZE);
-			chunk = PADLOCK_CHUNK;
-		cfb_shortcut: /* optimize for small input */
-			if (inp_misaligned)
-				inp = padlock_memcpy(out, in_arg, chunk);
-			else
-				inp = in_arg;
-			in_arg += chunk;
-
-			iv = padlock_xcrypt_cfb(chunk/AES_BLOCK_SIZE, cdata, out, inp);
-
-			if (out_misaligned)
-				out_arg = padlock_memcpy(out_arg, out, chunk) + chunk;
-			else
-				out     = out_arg+=chunk;
-
-			nbytes -= chunk;
-		} while (nbytes >= AES_BLOCK_SIZE);
-
-		cfb_skiploop:
-		if (nbytes) {
-			unsigned char *ivp = cdata->iv;
-
-			if (iv != ivp) {
-				memcpy(ivp, iv, AES_BLOCK_SIZE);
-				iv = ivp;
-			}
-			ctx->num = nbytes;
-			if (cdata->cword.b.encdec) {
-				cdata->cword.b.encdec=0;
-				padlock_reload_key();
-				padlock_xcrypt_ecb(1,cdata,ivp,ivp);
-				cdata->cword.b.encdec=1;
-				padlock_reload_key();
-				while(nbytes) {
-					unsigned char c = *(in_arg++);
-					*(out_arg++) = c ^ *ivp;
-					*(ivp++) = c, nbytes--;
-				}
-			}
-			else {	padlock_reload_key();
-				padlock_xcrypt_ecb(1,cdata,ivp,ivp);
-				padlock_reload_key();
-				while (nbytes) {
-					*ivp = *(out_arg++) = *(in_arg++) ^ *ivp;
-					ivp++, nbytes--;
-				}
-			}
-		}
-
-		memcpy(ctx->iv, iv, AES_BLOCK_SIZE);
-		break;
-
-	case EVP_CIPH_OFB_MODE:
-		memcpy(cdata->iv, ctx->iv, AES_BLOCK_SIZE);
-		chunk &= ~(AES_BLOCK_SIZE-1);
-		if (chunk) do	{
-			if (inp_misaligned)
-				inp = padlock_memcpy(out, in_arg, chunk);
-			else
-				inp = in_arg;
-			in_arg += chunk;
-
-			padlock_xcrypt_ofb(chunk/AES_BLOCK_SIZE, cdata, out, inp);
-
-			if (out_misaligned)
-				out_arg = padlock_memcpy(out_arg, out, chunk) + chunk;
-			else
-				out     = out_arg+=chunk;
-
-			nbytes -= chunk;
-			chunk   = PADLOCK_CHUNK;
-		} while (nbytes >= AES_BLOCK_SIZE);
-
-		if (nbytes) {
-			unsigned char *ivp = cdata->iv;
-
-			ctx->num = nbytes;
-			padlock_reload_key();	/* empirically found */
-			padlock_xcrypt_ecb(1,cdata,ivp,ivp);
-			padlock_reload_key();	/* empirically found */
-			while (nbytes) {
-				*(out_arg++) = *(in_arg++) ^ *ivp;
-				ivp++, nbytes--;
-			}
-		}
-
-		memcpy(ctx->iv, cdata->iv, AES_BLOCK_SIZE);
-		break;
-
-	default:
-		return 0;
-	}
-
-	/* Clean the realign buffer if it was used */
-	if (out_misaligned) {
-		volatile unsigned long *p=(void *)out;
-		size_t   n = allocated/sizeof(*p);
-		while (n--) *p++=0;
-	}
-
-	memset(cdata->iv, 0, AES_BLOCK_SIZE);
-
-	return 1;
-}
-
-#endif /* OPENSSL_NO_AES */
-
-/* ===== Random Number Generator ===== */
-/*
- * This code is not engaged. The reason is that it does not comply
- * with recommendations for VIA RNG usage for secure applications
- * (posted at http://www.via.com.tw/en/viac3/c3.jsp) nor does it
- * provide meaningful error control...
- */
-/* Wrapper that provides an interface between the API and 
-   the raw PadLock RNG */
-static int
-padlock_rand_bytes(unsigned char *output, int count)
-{
-	unsigned int eax, buf;
-
-	while (count >= 8) {
-		eax = padlock_xstore(output, 0);
-		if (!(eax&(1<<6)))	return 0; /* RNG disabled */
-		/* this ---vv--- covers DC bias, Raw Bits and String Filter */
-		if (eax&(0x1F<<10))	return 0;
-		if ((eax&0x1F)==0)	continue; /* no data, retry... */
-		if ((eax&0x1F)!=8)	return 0; /* fatal failure...  */
-		output += 8;
-		count  -= 8;
-	}
-	while (count > 0) {
-		eax = padlock_xstore(&buf, 3);
-		if (!(eax&(1<<6)))	return 0; /* RNG disabled */
-		/* this ---vv--- covers DC bias, Raw Bits and String Filter */
-		if (eax&(0x1F<<10))	return 0;
-		if ((eax&0x1F)==0)	continue; /* no data, retry... */
-		if ((eax&0x1F)!=1)	return 0; /* fatal failure...  */
-		*output++ = (unsigned char)buf;
-		count--;
-	}
-	*(volatile unsigned int *)&buf=0;
-
-	return 1;
-}
-
-/* Dummy but necessary function */
-static int
-padlock_rand_status(void)
-{
-	return 1;
-}
-
-/* Prepare structure for registration */
-static RAND_METHOD padlock_rand = {
-	NULL,			/* seed */
-	padlock_rand_bytes,	/* bytes */
-	NULL,			/* cleanup */
-	NULL,			/* add */
-	padlock_rand_bytes,	/* pseudorand */
-	padlock_rand_status,	/* rand status */
-};
-
-#endif /* COMPILE_HW_PADLOCK */
-
-#endif /* !OPENSSL_NO_HW_PADLOCK */
-#endif /* !OPENSSL_NO_HW */
diff --git a/crypto/engine/eng_rdrand.c b/crypto/engine/eng_rdrand.c
new file mode 100644
index 000000000000..a9ba5ae6f9ff
--- /dev/null
+++ b/crypto/engine/eng_rdrand.c
@@ -0,0 +1,142 @@
+/* ====================================================================
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#include <openssl/opensslconf.h>
+
+#include <stdio.h>
+#include <string.h>
+#include <openssl/engine.h>
+#include <openssl/rand.h>
+#include <openssl/err.h>
+
+#if (defined(__i386)   || defined(__i386__)   || defined(_M_IX86) || \
+     defined(__x86_64) || defined(__x86_64__) || \
+     defined(_M_AMD64) || defined (_M_X64)) && defined(OPENSSL_CPUID_OBJ)
+
+size_t OPENSSL_ia32_rdrand(void);
+
+static int get_random_bytes (unsigned char *buf, int num)
+	{
+	size_t rnd;
+
+	while (num>=(int)sizeof(size_t)) {
+		if ((rnd = OPENSSL_ia32_rdrand()) == 0) return 0;
+
+		*((size_t *)buf) = rnd;
+		buf += sizeof(size_t);
+		num -= sizeof(size_t);
+	}
+	if (num) {
+		if ((rnd = OPENSSL_ia32_rdrand()) == 0) return 0;
+
+		memcpy (buf,&rnd,num);
+	}
+
+	return 1;
+	}
+
+static int random_status (void)
+{	return 1;	}
+
+static RAND_METHOD rdrand_meth =
+	{
+	NULL,	/* seed */
+	get_random_bytes,
+	NULL,	/* cleanup */
+	NULL,	/* add */
+	get_random_bytes,
+	random_status,
+	};
+
+static int rdrand_init(ENGINE *e)
+{	return 1;	}
+
+static const char *engine_e_rdrand_id = "rdrand";
+static const char *engine_e_rdrand_name = "Intel RDRAND engine";
+
+static int bind_helper(ENGINE *e)
+	{
+	if (!ENGINE_set_id(e, engine_e_rdrand_id) ||
+	    !ENGINE_set_name(e, engine_e_rdrand_name) ||
+	    !ENGINE_set_init_function(e, rdrand_init) ||
+	    !ENGINE_set_RAND(e, &rdrand_meth) )
+		return 0;
+
+	return 1;
+	}
+
+static ENGINE *ENGINE_rdrand(void)
+	{
+	ENGINE *ret = ENGINE_new();
+	if(!ret)
+		return NULL;
+	if(!bind_helper(ret))
+		{
+		ENGINE_free(ret);
+		return NULL;
+		}
+	return ret;
+	}
+
+void ENGINE_load_rdrand (void)
+	{
+	extern unsigned int OPENSSL_ia32cap_P[];
+
+	if (OPENSSL_ia32cap_P[1] & (1<<(62-32)))
+		{
+		ENGINE *toadd = ENGINE_rdrand();
+		if(!toadd) return;
+		ENGINE_add(toadd);
+		ENGINE_free(toadd);
+		ERR_clear_error();
+		}
+	}
+#else
+void ENGINE_load_rdrand (void) {}
+#endif
diff --git a/crypto/engine/eng_rsax.c b/crypto/engine/eng_rsax.c
new file mode 100644
index 000000000000..96e63477eedb
--- /dev/null
+++ b/crypto/engine/eng_rsax.c
@@ -0,0 +1,668 @@
+/* crypto/engine/eng_rsax.c */
+/* Copyright (c) 2010-2010 Intel Corp.
+ *   Author: Vinodh.Gopal@intel.com
+ *           Jim Guilford
+ *           Erdinc.Ozturk@intel.com
+ *           Maxim.Perminov@intel.com
+ *           Ying.Huang@intel.com
+ *
+ * More information about algorithm used can be found at:
+ *   http://www.cse.buffalo.edu/srds2009/escs2009_submission_Gopal.pdf
+ */
+/* ====================================================================
+ * Copyright (c) 1999-2001 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ */
+
+#include <openssl/opensslconf.h>
+
+#include <stdio.h>
+#include <string.h>
+#include <openssl/crypto.h>
+#include <openssl/buffer.h>
+#include <openssl/engine.h>
+#ifndef OPENSSL_NO_RSA
+#include <openssl/rsa.h>
+#endif
+#include <openssl/bn.h>
+#include <openssl/err.h>
+
+/* RSAX is available **ONLY* on x86_64 CPUs */
+#undef COMPILE_RSAX
+
+#if (defined(__x86_64) || defined(__x86_64__) || \
+     defined(_M_AMD64) || defined (_M_X64)) && !defined(OPENSSL_NO_ASM)
+#define COMPILE_RSAX
+static ENGINE *ENGINE_rsax (void);
+#endif
+
+void ENGINE_load_rsax (void)
+	{
+/* On non-x86 CPUs it just returns. */
+#ifdef COMPILE_RSAX
+	ENGINE *toadd = ENGINE_rsax();
+	if(!toadd) return;
+	ENGINE_add(toadd);
+	ENGINE_free(toadd);
+	ERR_clear_error();
+#endif
+	}
+
+#ifdef COMPILE_RSAX
+#define E_RSAX_LIB_NAME "rsax engine"
+
+static int e_rsax_destroy(ENGINE *e);
+static int e_rsax_init(ENGINE *e);
+static int e_rsax_finish(ENGINE *e);
+static int e_rsax_ctrl(ENGINE *e, int cmd, long i, void *p, void (*f)(void));
+
+#ifndef OPENSSL_NO_RSA
+/* RSA stuff */
+static int e_rsax_rsa_mod_exp(BIGNUM *r, const BIGNUM *I, RSA *rsa, BN_CTX *ctx);
+static int e_rsax_rsa_finish(RSA *r);
+#endif
+
+static const ENGINE_CMD_DEFN e_rsax_cmd_defns[] = {
+	{0, NULL, NULL, 0}
+	};
+
+#ifndef OPENSSL_NO_RSA
+/* Our internal RSA_METHOD that we provide pointers to */
+static RSA_METHOD e_rsax_rsa =
+	{
+	"Intel RSA-X method",
+	NULL,
+	NULL,
+	NULL,
+	NULL,
+	e_rsax_rsa_mod_exp,
+	NULL,
+	NULL,
+	e_rsax_rsa_finish,
+	RSA_FLAG_CACHE_PUBLIC|RSA_FLAG_CACHE_PRIVATE,
+	NULL,
+	NULL,
+	NULL
+	};
+#endif
+
+/* Constants used when creating the ENGINE */
+static const char *engine_e_rsax_id = "rsax";
+static const char *engine_e_rsax_name = "RSAX engine support";
+
+/* This internal function is used by ENGINE_rsax() */
+static int bind_helper(ENGINE *e)
+	{
+#ifndef OPENSSL_NO_RSA
+	const RSA_METHOD *meth1;
+#endif
+	if(!ENGINE_set_id(e, engine_e_rsax_id) ||
+			!ENGINE_set_name(e, engine_e_rsax_name) ||
+#ifndef OPENSSL_NO_RSA
+			!ENGINE_set_RSA(e, &e_rsax_rsa) ||
+#endif
+			!ENGINE_set_destroy_function(e, e_rsax_destroy) ||
+			!ENGINE_set_init_function(e, e_rsax_init) ||
+			!ENGINE_set_finish_function(e, e_rsax_finish) ||
+			!ENGINE_set_ctrl_function(e, e_rsax_ctrl) ||
+			!ENGINE_set_cmd_defns(e, e_rsax_cmd_defns))
+		return 0;
+
+#ifndef OPENSSL_NO_RSA
+	meth1 = RSA_PKCS1_SSLeay();
+	e_rsax_rsa.rsa_pub_enc = meth1->rsa_pub_enc;
+	e_rsax_rsa.rsa_pub_dec = meth1->rsa_pub_dec;
+	e_rsax_rsa.rsa_priv_enc = meth1->rsa_priv_enc;
+	e_rsax_rsa.rsa_priv_dec = meth1->rsa_priv_dec;
+	e_rsax_rsa.bn_mod_exp = meth1->bn_mod_exp;
+#endif
+	return 1;
+	}
+
+static ENGINE *ENGINE_rsax(void)
+	{
+	ENGINE *ret = ENGINE_new();
+	if(!ret)
+		return NULL;
+	if(!bind_helper(ret))
+		{
+		ENGINE_free(ret);
+		return NULL;
+		}
+	return ret;
+	}
+
+#ifndef OPENSSL_NO_RSA
+/* Used to attach our own key-data to an RSA structure */
+static int rsax_ex_data_idx = -1;
+#endif
+
+static int e_rsax_destroy(ENGINE *e)
+	{
+	return 1;
+	}
+
+/* (de)initialisation functions. */
+static int e_rsax_init(ENGINE *e)
+	{
+#ifndef OPENSSL_NO_RSA
+	if (rsax_ex_data_idx == -1)
+		rsax_ex_data_idx = RSA_get_ex_new_index(0,
+			NULL,
+			NULL, NULL, NULL);
+#endif
+	if (rsax_ex_data_idx  == -1)
+		return 0;
+	return 1;
+	}
+
+static int e_rsax_finish(ENGINE *e)
+	{
+	return 1;
+	}
+
+static int e_rsax_ctrl(ENGINE *e, int cmd, long i, void *p, void (*f)(void))
+	{
+	int to_return = 1;
+
+	switch(cmd)
+		{
+	/* The command isn't understood by this engine */
+	default:
+		to_return = 0;
+		break;
+		}
+
+	return to_return;
+	}
+
+
+#ifndef OPENSSL_NO_RSA
+
+#ifdef _WIN32
+typedef unsigned __int64 UINT64;
+#else
+typedef unsigned long long UINT64;
+#endif
+typedef unsigned short UINT16;
+
+/* Table t is interleaved in the following manner:
+ * The order in memory is t[0][0], t[0][1], ..., t[0][7], t[1][0], ...
+ * A particular 512-bit value is stored in t[][index] rather than the more
+ * normal t[index][]; i.e. the qwords of a particular entry in t are not
+ * adjacent in memory
+ */
+
+/* Init BIGNUM b from the interleaved UINT64 array */
+static int interleaved_array_to_bn_512(BIGNUM* b, UINT64 *array);
+
+/* Extract array elements from BIGNUM b
+ * To set the whole array from b, call with n=8
+ */
+static int bn_extract_to_array_512(const BIGNUM* b, unsigned int n, UINT64 *array);
+
+struct mod_ctx_512 {
+    UINT64 t[8][8];
+    UINT64 m[8];
+    UINT64 m1[8]; /* 2^278 % m */
+    UINT64 m2[8]; /* 2^640 % m */
+    UINT64 k1[2]; /* (- 1/m) % 2^128 */
+};
+
+static int mod_exp_pre_compute_data_512(UINT64 *m, struct mod_ctx_512 *data);
+
+void mod_exp_512(UINT64 *result, /* 512 bits, 8 qwords */
+		 UINT64 *g,      /* 512 bits, 8 qwords */
+		 UINT64 *exp,    /* 512 bits, 8 qwords */
+		 struct mod_ctx_512 *data);
+
+typedef struct st_e_rsax_mod_ctx
+{
+  UINT64 type;
+  union {
+    struct mod_ctx_512 b512;
+  } ctx;
+
+} E_RSAX_MOD_CTX;
+
+static E_RSAX_MOD_CTX *e_rsax_get_ctx(RSA *rsa, int idx, BIGNUM* m)
+{
+	E_RSAX_MOD_CTX *hptr;
+
+        if (idx < 0 || idx > 2)
+           return NULL;
+
+	hptr = RSA_get_ex_data(rsa, rsax_ex_data_idx);
+	if (!hptr) {
+	    hptr = OPENSSL_malloc(3*sizeof(E_RSAX_MOD_CTX));
+	    if (!hptr) return NULL;
+            hptr[2].type = hptr[1].type= hptr[0].type = 0;
+	    RSA_set_ex_data(rsa, rsax_ex_data_idx, hptr);
+        }
+
+        if (hptr[idx].type == (UINT64)BN_num_bits(m))
+            return hptr+idx;
+
+        if (BN_num_bits(m) == 512) {
+  	    UINT64 _m[8];
+	    bn_extract_to_array_512(m, 8, _m);
+	    memset( &hptr[idx].ctx.b512, 0, sizeof(struct mod_ctx_512));
+	    mod_exp_pre_compute_data_512(_m, &hptr[idx].ctx.b512);
+	}
+
+        hptr[idx].type = BN_num_bits(m);
+	return hptr+idx;
+}
+
+static int e_rsax_rsa_finish(RSA *rsa)
+	{
+	E_RSAX_MOD_CTX *hptr = RSA_get_ex_data(rsa, rsax_ex_data_idx);
+	if(hptr)
+		{
+		OPENSSL_free(hptr);
+		RSA_set_ex_data(rsa, rsax_ex_data_idx, NULL);
+		}
+	if (rsa->_method_mod_n)
+		BN_MONT_CTX_free(rsa->_method_mod_n);
+	if (rsa->_method_mod_p)
+		BN_MONT_CTX_free(rsa->_method_mod_p);
+	if (rsa->_method_mod_q)
+		BN_MONT_CTX_free(rsa->_method_mod_q);
+	return 1;
+	}
+
+
+static int e_rsax_bn_mod_exp(BIGNUM *r, const BIGNUM *g, const BIGNUM *e,
+                    const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *in_mont, E_RSAX_MOD_CTX* rsax_mod_ctx )
+{
+	if (rsax_mod_ctx && BN_get_flags(e, BN_FLG_CONSTTIME) != 0) {
+	   if (BN_num_bits(m) == 512) {
+  		UINT64 _r[8];
+   	  	UINT64 _g[8];
+		UINT64 _e[8];
+
+		/* Init the arrays from the BIGNUMs */
+		bn_extract_to_array_512(g, 8, _g);
+		bn_extract_to_array_512(e, 8, _e);
+
+		mod_exp_512(_r, _g, _e, &rsax_mod_ctx->ctx.b512);
+		/* Return the result in the BIGNUM */
+		interleaved_array_to_bn_512(r, _r);
+                return 1;
+           }
+        }
+
+	return BN_mod_exp_mont(r, g, e, m, ctx, in_mont);
+}
+
+/* Declares for the Intel CIAP 512-bit / CRT / 1024 bit RSA modular
+ * exponentiation routine precalculations and a structure to hold the
+ * necessary values.  These files are meant to live in crypto/rsa/ in
+ * the target openssl.
+ */
+
+/*
+ * Local method: extracts a piece from a BIGNUM, to fit it into
+ * an array. Call with n=8 to extract an entire 512-bit BIGNUM
+ */
+static int bn_extract_to_array_512(const BIGNUM* b, unsigned int n, UINT64 *array)
+{
+	int i;
+	UINT64 tmp;
+	unsigned char bn_buff[64];
+	memset(bn_buff, 0, 64);
+	if (BN_num_bytes(b) > 64) {
+		printf ("Can't support this byte size\n");
+		return 0; }
+	if (BN_num_bytes(b)!=0) {
+		if (!BN_bn2bin(b, bn_buff+(64-BN_num_bytes(b)))) {
+			printf ("Error's in bn2bin\n");
+			/* We have to error, here */
+			return 0; } }
+	while (n-- > 0) {
+		array[n] = 0;
+		for (i=7; i>=0; i--) {
+			tmp = bn_buff[63-(n*8+i)];
+			array[n] |= tmp << (8*i); } }
+	return 1;
+}
+
+/* Init a 512-bit BIGNUM from the UINT64*_ (8 * 64) interleaved array */
+static int interleaved_array_to_bn_512(BIGNUM* b, UINT64 *array)
+{
+	unsigned char tmp[64];
+	int n=8;
+	int i;
+	while (n-- > 0) {
+		for (i = 7; i>=0; i--) {
+			tmp[63-(n*8+i)] = (unsigned char)(array[n]>>(8*i)); } }
+	BN_bin2bn(tmp, 64, b);
+        return 0;
+}
+
+
+/* The main 512bit precompute call */
+static int mod_exp_pre_compute_data_512(UINT64 *m, struct mod_ctx_512 *data)
+ {
+    BIGNUM two_768, two_640, two_128, two_512, tmp, _m, tmp2;
+
+    /* We need a BN_CTX for the modulo functions */
+    BN_CTX* ctx;
+    /* Some tmps */
+    UINT64 _t[8];
+    int i, j, ret = 0;
+
+    /* Init _m with m */
+    BN_init(&_m);
+    interleaved_array_to_bn_512(&_m, m);
+    memset(_t, 0, 64);
+
+    /* Inits */
+    BN_init(&two_768);
+    BN_init(&two_640);
+    BN_init(&two_128);
+    BN_init(&two_512);
+    BN_init(&tmp);
+    BN_init(&tmp2);
+
+    /* Create our context */
+    if ((ctx=BN_CTX_new()) == NULL) { goto err; }
+	BN_CTX_start(ctx);
+
+    /*
+     * For production, if you care, these only need to be set once,
+     * and may be made constants.
+     */
+    BN_lshift(&two_768, BN_value_one(), 768);
+    BN_lshift(&two_640, BN_value_one(), 640);
+    BN_lshift(&two_128, BN_value_one(), 128);
+    BN_lshift(&two_512, BN_value_one(), 512);
+
+    if (0 == (m[7] & 0x8000000000000000)) {
+        exit(1);
+    }
+    if (0 == (m[0] & 0x1)) { /* Odd modulus required for Mont */
+        exit(1);
+    }
+
+    /* Precompute m1 */
+    BN_mod(&tmp, &two_768, &_m, ctx);
+    if (!bn_extract_to_array_512(&tmp, 8, &data->m1[0])) {
+	    goto err; }
+
+    /* Precompute m2 */
+    BN_mod(&tmp, &two_640, &_m, ctx);
+    if (!bn_extract_to_array_512(&tmp, 8, &data->m2[0])) {
+	    goto err;
+    }
+
+    /*
+     * Precompute k1, a 128b number = ((-1)* m-1 ) mod 2128; k1 should
+     * be non-negative.
+     */
+    BN_mod_inverse(&tmp, &_m, &two_128, ctx);
+    if (!BN_is_zero(&tmp)) { BN_sub(&tmp, &two_128, &tmp); }
+    if (!bn_extract_to_array_512(&tmp, 2, &data->k1[0])) {
+	    goto err; }
+
+    /* Precompute t */
+    for (i=0; i<8; i++) {
+        BN_zero(&tmp);
+        if (i & 1) { BN_add(&tmp, &two_512, &tmp); }
+        if (i & 2) { BN_add(&tmp, &two_512, &tmp); }
+        if (i & 4) { BN_add(&tmp, &two_640, &tmp); }
+
+        BN_nnmod(&tmp2, &tmp, &_m, ctx);
+        if (!bn_extract_to_array_512(&tmp2, 8, _t)) {
+	        goto err; }
+        for (j=0; j<8; j++) data->t[j][i] = _t[j]; }
+
+    /* Precompute m */
+    for (i=0; i<8; i++) {
+        data->m[i] = m[i]; }
+
+    ret = 1;
+
+err:
+    /* Cleanup */
+	if (ctx != NULL) {
+		BN_CTX_end(ctx); BN_CTX_free(ctx); }
+    BN_free(&two_768);
+    BN_free(&two_640);
+    BN_free(&two_128);
+    BN_free(&two_512);
+    BN_free(&tmp);
+    BN_free(&tmp2);
+    BN_free(&_m);
+
+    return ret;
+}
+
+
+static int e_rsax_rsa_mod_exp(BIGNUM *r0, const BIGNUM *I, RSA *rsa, BN_CTX *ctx)
+	{
+	BIGNUM *r1,*m1,*vrfy;
+	BIGNUM local_dmp1,local_dmq1,local_c,local_r1;
+	BIGNUM *dmp1,*dmq1,*c,*pr1;
+	int ret=0;
+
+	BN_CTX_start(ctx);
+	r1 = BN_CTX_get(ctx);
+	m1 = BN_CTX_get(ctx);
+	vrfy = BN_CTX_get(ctx);
+
+	{
+		BIGNUM local_p, local_q;
+		BIGNUM *p = NULL, *q = NULL;
+		int error = 0;
+
+		/* Make sure BN_mod_inverse in Montgomery
+		 * intialization uses the BN_FLG_CONSTTIME flag
+		 * (unless RSA_FLAG_NO_CONSTTIME is set)
+		 */
+		if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+			{
+			BN_init(&local_p);
+			p = &local_p;
+			BN_with_flags(p, rsa->p, BN_FLG_CONSTTIME);
+
+			BN_init(&local_q);
+			q = &local_q;
+			BN_with_flags(q, rsa->q, BN_FLG_CONSTTIME);
+			}
+		else
+			{
+			p = rsa->p;
+			q = rsa->q;
+			}
+
+		if (rsa->flags & RSA_FLAG_CACHE_PRIVATE)
+			{
+			if (!BN_MONT_CTX_set_locked(&rsa->_method_mod_p, CRYPTO_LOCK_RSA, p, ctx))
+				error = 1;
+			if (!BN_MONT_CTX_set_locked(&rsa->_method_mod_q, CRYPTO_LOCK_RSA, q, ctx))
+				error = 1;
+			}
+
+		/* clean up */
+		if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+			{
+			BN_free(&local_p);
+			BN_free(&local_q);
+			}
+		if ( error )
+			goto err;
+	}
+
+	if (rsa->flags & RSA_FLAG_CACHE_PUBLIC)
+		if (!BN_MONT_CTX_set_locked(&rsa->_method_mod_n, CRYPTO_LOCK_RSA, rsa->n, ctx))
+			goto err;
+
+	/* compute I mod q */
+	if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+		{
+		c = &local_c;
+		BN_with_flags(c, I, BN_FLG_CONSTTIME);
+		if (!BN_mod(r1,c,rsa->q,ctx)) goto err;
+		}
+	else
+		{
+		if (!BN_mod(r1,I,rsa->q,ctx)) goto err;
+		}
+
+	/* compute r1^dmq1 mod q */
+	if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+		{
+		dmq1 = &local_dmq1;
+		BN_with_flags(dmq1, rsa->dmq1, BN_FLG_CONSTTIME);
+		}
+	else
+		dmq1 = rsa->dmq1;
+
+	if (!e_rsax_bn_mod_exp(m1,r1,dmq1,rsa->q,ctx,
+		rsa->_method_mod_q, e_rsax_get_ctx(rsa, 0, rsa->q) )) goto err;
+
+	/* compute I mod p */
+	if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+		{
+		c = &local_c;
+		BN_with_flags(c, I, BN_FLG_CONSTTIME);
+		if (!BN_mod(r1,c,rsa->p,ctx)) goto err;
+		}
+	else
+		{
+		if (!BN_mod(r1,I,rsa->p,ctx)) goto err;
+		}
+
+	/* compute r1^dmp1 mod p */
+	if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+		{
+		dmp1 = &local_dmp1;
+		BN_with_flags(dmp1, rsa->dmp1, BN_FLG_CONSTTIME);
+		}
+	else
+		dmp1 = rsa->dmp1;
+
+	if (!e_rsax_bn_mod_exp(r0,r1,dmp1,rsa->p,ctx,
+		rsa->_method_mod_p, e_rsax_get_ctx(rsa, 1, rsa->p) )) goto err;
+
+	if (!BN_sub(r0,r0,m1)) goto err;
+	/* This will help stop the size of r0 increasing, which does
+	 * affect the multiply if it optimised for a power of 2 size */
+	if (BN_is_negative(r0))
+		if (!BN_add(r0,r0,rsa->p)) goto err;
+
+	if (!BN_mul(r1,r0,rsa->iqmp,ctx)) goto err;
+
+	/* Turn BN_FLG_CONSTTIME flag on before division operation */
+	if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+		{
+		pr1 = &local_r1;
+		BN_with_flags(pr1, r1, BN_FLG_CONSTTIME);
+		}
+	else
+		pr1 = r1;
+	if (!BN_mod(r0,pr1,rsa->p,ctx)) goto err;
+
+	/* If p < q it is occasionally possible for the correction of
+         * adding 'p' if r0 is negative above to leave the result still
+	 * negative. This can break the private key operations: the following
+	 * second correction should *always* correct this rare occurrence.
+	 * This will *never* happen with OpenSSL generated keys because
+         * they ensure p > q [steve]
+         */
+	if (BN_is_negative(r0))
+		if (!BN_add(r0,r0,rsa->p)) goto err;
+	if (!BN_mul(r1,r0,rsa->q,ctx)) goto err;
+	if (!BN_add(r0,r1,m1)) goto err;
+
+	if (rsa->e && rsa->n)
+		{
+		if (!e_rsax_bn_mod_exp(vrfy,r0,rsa->e,rsa->n,ctx,rsa->_method_mod_n, e_rsax_get_ctx(rsa, 2, rsa->n) ))
+                    goto err;
+
+		/* If 'I' was greater than (or equal to) rsa->n, the operation
+		 * will be equivalent to using 'I mod n'. However, the result of
+		 * the verify will *always* be less than 'n' so we don't check
+		 * for absolute equality, just congruency. */
+		if (!BN_sub(vrfy, vrfy, I)) goto err;
+		if (!BN_mod(vrfy, vrfy, rsa->n, ctx)) goto err;
+		if (BN_is_negative(vrfy))
+			if (!BN_add(vrfy, vrfy, rsa->n)) goto err;
+		if (!BN_is_zero(vrfy))
+			{
+			/* 'I' and 'vrfy' aren't congruent mod n. Don't leak
+			 * miscalculated CRT output, just do a raw (slower)
+			 * mod_exp and return that instead. */
+
+			BIGNUM local_d;
+			BIGNUM *d = NULL;
+
+			if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+				{
+				d = &local_d;
+				BN_with_flags(d, rsa->d, BN_FLG_CONSTTIME);
+				}
+			else
+				d = rsa->d;
+			if (!e_rsax_bn_mod_exp(r0,I,d,rsa->n,ctx,
+						   rsa->_method_mod_n, e_rsax_get_ctx(rsa, 2, rsa->n) )) goto err;
+			}
+		}
+	ret=1;
+
+err:
+	BN_CTX_end(ctx);
+
+	return ret;
+	}
+#endif /* !OPENSSL_NO_RSA */
+#endif /* !COMPILE_RSAX */
diff --git a/crypto/engine/eng_table.c b/crypto/engine/eng_table.c
index 8fc47b3399c8..4fde9481852c 100644
--- a/crypto/engine/eng_table.c
+++ b/crypto/engine/eng_table.c
@@ -70,12 +70,22 @@ typedef struct st_engine_pile
 	int uptodate;
 	} ENGINE_PILE;
 
+DECLARE_LHASH_OF(ENGINE_PILE);
+
 /* The type exposed in eng_int.h */
 struct st_engine_table
 	{
-	LHASH piles;
+	LHASH_OF(ENGINE_PILE) piles;
 	}; /* ENGINE_TABLE */
 
+
+typedef struct st_engine_pile_doall
+	{
+	engine_table_doall_cb *cb;
+	void *arg;
+	} ENGINE_PILE_DOALL;
+	
+
 /* Global flags (ENGINE_TABLE_FLAG_***). */
 static unsigned int table_flags = 0;
 
@@ -84,6 +94,7 @@ unsigned int ENGINE_get_table_flags(void)
 	{
 	return table_flags;
 	}
+
 void ENGINE_set_table_flags(unsigned int flags)
 	{
 	table_flags = flags;
@@ -94,19 +105,21 @@ static unsigned long engine_pile_hash(const ENGINE_PILE *c)
 	{
 	return c->nid;
 	}
+
 static int engine_pile_cmp(const ENGINE_PILE *a, const ENGINE_PILE *b)
 	{
 	return a->nid - b->nid;
 	}
-static IMPLEMENT_LHASH_HASH_FN(engine_pile_hash, const ENGINE_PILE *)
-static IMPLEMENT_LHASH_COMP_FN(engine_pile_cmp, const ENGINE_PILE *)
+static IMPLEMENT_LHASH_HASH_FN(engine_pile, ENGINE_PILE)
+static IMPLEMENT_LHASH_COMP_FN(engine_pile, ENGINE_PILE)
+
 static int int_table_check(ENGINE_TABLE **t, int create)
 	{
-	LHASH *lh;
+	LHASH_OF(ENGINE_PILE) *lh;
+
 	if(*t) return 1;
 	if(!create) return 0;
-	if((lh = lh_new(LHASH_HASH_FN(engine_pile_hash),
-			LHASH_COMP_FN(engine_pile_cmp))) == NULL)
+	if((lh = lh_ENGINE_PILE_new()) == NULL)
 		return 0;
 	*t = (ENGINE_TABLE *)lh;
 	return 1;
@@ -130,7 +143,7 @@ int engine_table_register(ENGINE_TABLE **table, ENGINE_CLEANUP_CB *cleanup,
 	while(num_nids--)
 		{
 		tmplate.nid = *nids;
-		fnd = lh_retrieve(&(*table)->piles, &tmplate);
+		fnd = lh_ENGINE_PILE_retrieve(&(*table)->piles, &tmplate);
 		if(!fnd)
 			{
 			fnd = OPENSSL_malloc(sizeof(ENGINE_PILE));
@@ -144,7 +157,7 @@ int engine_table_register(ENGINE_TABLE **table, ENGINE_CLEANUP_CB *cleanup,
 				goto end;
 				}
 			fnd->funct = NULL;
-			lh_insert(&(*table)->piles, fnd);
+			(void)lh_ENGINE_PILE_insert(&(*table)->piles, fnd);
 			}
 		/* A registration shouldn't add duplciate entries */
 		(void)sk_ENGINE_delete_ptr(fnd->sk, e);
@@ -173,7 +186,7 @@ end:
 	CRYPTO_w_unlock(CRYPTO_LOCK_ENGINE);
 	return ret;
 	}
-static void int_unregister_cb(ENGINE_PILE *pile, ENGINE *e)
+static void int_unregister_cb_doall_arg(ENGINE_PILE *pile, ENGINE *e)
 	{
 	int n;
 	/* Iterate the 'c->sk' stack removing any occurance of 'e' */
@@ -188,31 +201,35 @@ static void int_unregister_cb(ENGINE_PILE *pile, ENGINE *e)
 		pile->funct = NULL;
 		}
 	}
-static IMPLEMENT_LHASH_DOALL_ARG_FN(int_unregister_cb,ENGINE_PILE *,ENGINE *)
+static IMPLEMENT_LHASH_DOALL_ARG_FN(int_unregister_cb, ENGINE_PILE, ENGINE)
+
 void engine_table_unregister(ENGINE_TABLE **table, ENGINE *e)
 	{
 	CRYPTO_w_lock(CRYPTO_LOCK_ENGINE);
 	if(int_table_check(table, 0))
-		lh_doall_arg(&(*table)->piles,
-			LHASH_DOALL_ARG_FN(int_unregister_cb), e);
+		lh_ENGINE_PILE_doall_arg(&(*table)->piles,
+					 LHASH_DOALL_ARG_FN(int_unregister_cb),
+					 ENGINE, e);
 	CRYPTO_w_unlock(CRYPTO_LOCK_ENGINE);
 	}
 
-static void int_cleanup_cb(ENGINE_PILE *p)
+static void int_cleanup_cb_doall(ENGINE_PILE *p)
 	{
 	sk_ENGINE_free(p->sk);
 	if(p->funct)
 		engine_unlocked_finish(p->funct, 0);
 	OPENSSL_free(p);
 	}
-static IMPLEMENT_LHASH_DOALL_FN(int_cleanup_cb,ENGINE_PILE *)
+static IMPLEMENT_LHASH_DOALL_FN(int_cleanup_cb, ENGINE_PILE)
+
 void engine_table_cleanup(ENGINE_TABLE **table)
 	{
 	CRYPTO_w_lock(CRYPTO_LOCK_ENGINE);
 	if(*table)
 		{
-		lh_doall(&(*table)->piles, LHASH_DOALL_FN(int_cleanup_cb));
-		lh_free(&(*table)->piles);
+		lh_ENGINE_PILE_doall(&(*table)->piles,
+				     LHASH_DOALL_FN(int_cleanup_cb));
+		lh_ENGINE_PILE_free(&(*table)->piles);
 		*table = NULL;
 		}
 	CRYPTO_w_unlock(CRYPTO_LOCK_ENGINE);
@@ -243,7 +260,7 @@ ENGINE *engine_table_select_tmp(ENGINE_TABLE **table, int nid, const char *f, in
 	 * operations. But don't worry about a fprintf(stderr). */
 	if(!int_table_check(table, 0)) goto end;
 	tmplate.nid = nid;
-	fnd = lh_retrieve(&(*table)->piles, &tmplate);
+	fnd = lh_ENGINE_PILE_retrieve(&(*table)->piles, &tmplate);
 	if(!fnd) goto end;
 	if(fnd->funct && engine_unlocked_init(fnd->funct))
 		{
@@ -314,3 +331,21 @@ end:
 	ERR_pop_to_mark();
 	return ret;
 	}
+
+/* Table enumeration */
+
+static void int_cb_doall_arg(ENGINE_PILE *pile, ENGINE_PILE_DOALL *dall)
+	{
+	dall->cb(pile->nid, pile->sk, pile->funct, dall->arg);
+	}
+static IMPLEMENT_LHASH_DOALL_ARG_FN(int_cb, ENGINE_PILE,ENGINE_PILE_DOALL)
+
+void engine_table_doall(ENGINE_TABLE *table, engine_table_doall_cb *cb,
+								void *arg)
+	{
+	ENGINE_PILE_DOALL dall;
+	dall.cb = cb;
+	dall.arg = arg;
+	lh_ENGINE_PILE_doall_arg(&table->piles, LHASH_DOALL_ARG_FN(int_cb),
+				 ENGINE_PILE_DOALL, &dall);
+	}
diff --git a/crypto/engine/engine.h b/crypto/engine/engine.h
index d4bc1efc794c..f8be49772446 100644
--- a/crypto/engine/engine.h
+++ b/crypto/engine/engine.h
@@ -88,16 +88,15 @@
 #include <openssl/ecdsa.h>
 #endif
 #include <openssl/rand.h>
-#include <openssl/store.h>
 #include <openssl/ui.h>
 #include <openssl/err.h>
 #endif
 
-#include <openssl/x509.h>
-
 #include <openssl/ossl_typ.h>
 #include <openssl/symhacks.h>
 
+#include <openssl/x509.h>
+
 #ifdef  __cplusplus
 extern "C" {
 #endif
@@ -113,6 +112,8 @@ extern "C" {
 #define ENGINE_METHOD_CIPHERS		(unsigned int)0x0040
 #define ENGINE_METHOD_DIGESTS		(unsigned int)0x0080
 #define ENGINE_METHOD_STORE		(unsigned int)0x0100
+#define ENGINE_METHOD_PKEY_METHS	(unsigned int)0x0200
+#define ENGINE_METHOD_PKEY_ASN1_METHS	(unsigned int)0x0400
 /* Obvious all-or-nothing cases. */
 #define ENGINE_METHOD_ALL		(unsigned int)0xFFFF
 #define ENGINE_METHOD_NONE		(unsigned int)0x0000
@@ -140,6 +141,13 @@ extern "C" {
  * the existing ENGINE's structural reference count. */
 #define ENGINE_FLAGS_BY_ID_COPY		(int)0x0004
 
+/* This flag if for an ENGINE that does not want its methods registered as 
+ * part of ENGINE_register_all_complete() for example if the methods are
+ * not usable as default methods.
+ */
+
+#define ENGINE_FLAGS_NO_REGISTER_ALL	(int)0x0008
+
 /* ENGINEs can support their own command types, and these flags are used in
  * ENGINE_CTRL_GET_CMD_FLAGS to indicate to the caller what kind of input each
  * command expects. Currently only numeric and string input is supported. If a
@@ -297,7 +305,8 @@ typedef int (*ENGINE_SSL_CLIENT_CERT_PTR)(ENGINE *, SSL *ssl,
  * parameter is non-NULL it is set to the size of the returned array. */
 typedef int (*ENGINE_CIPHERS_PTR)(ENGINE *, const EVP_CIPHER **, const int **, int);
 typedef int (*ENGINE_DIGESTS_PTR)(ENGINE *, const EVP_MD **, const int **, int);
-
+typedef int (*ENGINE_PKEY_METHS_PTR)(ENGINE *, EVP_PKEY_METHOD **, const int **, int);
+typedef int (*ENGINE_PKEY_ASN1_METHS_PTR)(ENGINE *, EVP_PKEY_ASN1_METHOD **, const int **, int);
 /* STRUCTURE functions ... all of these functions deal with pointers to ENGINE
  * structures where the pointers have a "structural reference". This means that
  * their reference is to allowed access to the structure but it does not imply
@@ -329,21 +338,22 @@ void ENGINE_load_aep(void);
 void ENGINE_load_atalla(void);
 void ENGINE_load_chil(void);
 void ENGINE_load_cswift(void);
-#ifndef OPENSSL_NO_GMP
-void ENGINE_load_gmp(void);
-#endif
 void ENGINE_load_nuron(void);
 void ENGINE_load_sureware(void);
 void ENGINE_load_ubsec(void);
-#endif
-void ENGINE_load_cryptodev(void);
 void ENGINE_load_padlock(void);
-void ENGINE_load_builtin_engines(void);
-#ifdef OPENSSL_SYS_WIN32
-#ifndef OPENSSL_NO_CAPIENG
 void ENGINE_load_capi(void);
+#ifndef OPENSSL_NO_GMP
+void ENGINE_load_gmp(void);
+#endif
+#ifndef OPENSSL_NO_GOST
+void ENGINE_load_gost(void);
 #endif
 #endif
+void ENGINE_load_cryptodev(void);
+void ENGINE_load_rsax(void);
+void ENGINE_load_rdrand(void);
+void ENGINE_load_builtin_engines(void);
 
 /* Get and set global flags (ENGINE_TABLE_FLAG_***) for the implementation
  * "registry" handling. */
@@ -394,6 +404,14 @@ int ENGINE_register_digests(ENGINE *e);
 void ENGINE_unregister_digests(ENGINE *e);
 void ENGINE_register_all_digests(void);
 
+int ENGINE_register_pkey_meths(ENGINE *e);
+void ENGINE_unregister_pkey_meths(ENGINE *e);
+void ENGINE_register_all_pkey_meths(void);
+
+int ENGINE_register_pkey_asn1_meths(ENGINE *e);
+void ENGINE_unregister_pkey_asn1_meths(ENGINE *e);
+void ENGINE_register_all_pkey_asn1_meths(void);
+
 /* These functions register all support from the above categories. Note, use of
  * these functions can result in static linkage of code your application may not
  * need. If you only need a subset of functionality, consider using more
@@ -473,6 +491,8 @@ int ENGINE_set_load_ssl_client_cert_function(ENGINE *e,
 				ENGINE_SSL_CLIENT_CERT_PTR loadssl_f);
 int ENGINE_set_ciphers(ENGINE *e, ENGINE_CIPHERS_PTR f);
 int ENGINE_set_digests(ENGINE *e, ENGINE_DIGESTS_PTR f);
+int ENGINE_set_pkey_meths(ENGINE *e, ENGINE_PKEY_METHS_PTR f);
+int ENGINE_set_pkey_asn1_meths(ENGINE *e, ENGINE_PKEY_ASN1_METHS_PTR f);
 int ENGINE_set_flags(ENGINE *e, int flags);
 int ENGINE_set_cmd_defns(ENGINE *e, const ENGINE_CMD_DEFN *defns);
 /* These functions allow control over any per-structure ENGINE data. */
@@ -509,8 +529,16 @@ ENGINE_LOAD_KEY_PTR ENGINE_get_load_pubkey_function(const ENGINE *e);
 ENGINE_SSL_CLIENT_CERT_PTR ENGINE_get_ssl_client_cert_function(const ENGINE *e);
 ENGINE_CIPHERS_PTR ENGINE_get_ciphers(const ENGINE *e);
 ENGINE_DIGESTS_PTR ENGINE_get_digests(const ENGINE *e);
+ENGINE_PKEY_METHS_PTR ENGINE_get_pkey_meths(const ENGINE *e);
+ENGINE_PKEY_ASN1_METHS_PTR ENGINE_get_pkey_asn1_meths(const ENGINE *e);
 const EVP_CIPHER *ENGINE_get_cipher(ENGINE *e, int nid);
 const EVP_MD *ENGINE_get_digest(ENGINE *e, int nid);
+const EVP_PKEY_METHOD *ENGINE_get_pkey_meth(ENGINE *e, int nid);
+const EVP_PKEY_ASN1_METHOD *ENGINE_get_pkey_asn1_meth(ENGINE *e, int nid);
+const EVP_PKEY_ASN1_METHOD *ENGINE_get_pkey_asn1_meth_str(ENGINE *e,
+					const char *str, int len);
+const EVP_PKEY_ASN1_METHOD *ENGINE_pkey_asn1_find_str(ENGINE **pe,
+					const char *str, int len);
 const ENGINE_CMD_DEFN *ENGINE_get_cmd_defns(const ENGINE *e);
 int ENGINE_get_flags(const ENGINE *e);
 
@@ -562,6 +590,8 @@ ENGINE *ENGINE_get_default_RAND(void);
  * ciphering or digesting corresponding to "nid". */
 ENGINE *ENGINE_get_cipher_engine(int nid);
 ENGINE *ENGINE_get_digest_engine(int nid);
+ENGINE *ENGINE_get_pkey_meth_engine(int nid);
+ENGINE *ENGINE_get_pkey_asn1_meth_engine(int nid);
 
 /* This sets a new default ENGINE structure for performing RSA
  * operations. If the result is non-zero (success) then the ENGINE
@@ -577,6 +607,8 @@ int ENGINE_set_default_DH(ENGINE *e);
 int ENGINE_set_default_RAND(ENGINE *e);
 int ENGINE_set_default_ciphers(ENGINE *e);
 int ENGINE_set_default_digests(ENGINE *e);
+int ENGINE_set_default_pkey_meths(ENGINE *e);
+int ENGINE_set_default_pkey_asn1_meths(ENGINE *e);
 
 /* The combination "set" - the flags are bitwise "OR"d from the
  * ENGINE_METHOD_*** defines above. As with the "ENGINE_register_complete()"
@@ -654,6 +686,7 @@ typedef struct st_dynamic_fns {
  * can be fully instantiated with IMPLEMENT_DYNAMIC_CHECK_FN(). */
 typedef unsigned long (*dynamic_v_check_fn)(unsigned long ossl_version);
 #define IMPLEMENT_DYNAMIC_CHECK_FN() \
+	OPENSSL_EXPORT unsigned long v_check(unsigned long v); \
 	OPENSSL_EXPORT unsigned long v_check(unsigned long v) { \
 		if(v >= OSSL_DYNAMIC_OLDEST) return OSSL_DYNAMIC_VERSION; \
 		return 0; }
@@ -677,6 +710,8 @@ typedef int (*dynamic_bind_engine)(ENGINE *e, const char *id,
 				const dynamic_fns *fns);
 #define IMPLEMENT_DYNAMIC_BIND_FN(fn) \
 	OPENSSL_EXPORT \
+	int bind_engine(ENGINE *e, const char *id, const dynamic_fns *fns); \
+	OPENSSL_EXPORT \
 	int bind_engine(ENGINE *e, const char *id, const dynamic_fns *fns) { \
 		if(ENGINE_get_static_state() == fns->static_state) goto skip_cbs; \
 		if(!CRYPTO_set_mem_functions(fns->mem_fns.malloc_cb, \
@@ -705,7 +740,7 @@ typedef int (*dynamic_bind_engine)(ENGINE *e, const char *id,
  * values. */
 void *ENGINE_get_static_state(void);
 
-#if defined(__OpenBSD__) || defined(__FreeBSD__)
+#if defined(__OpenBSD__) || defined(__FreeBSD__) || defined(HAVE_CRYPTODEV)
 void ENGINE_setup_bsd_cryptodev(void);
 #endif
 
@@ -734,13 +769,15 @@ void ERR_load_ENGINE_strings(void);
 #define ENGINE_F_ENGINE_GET_DEFAULT_TYPE		 177
 #define ENGINE_F_ENGINE_GET_DIGEST			 186
 #define ENGINE_F_ENGINE_GET_NEXT			 115
+#define ENGINE_F_ENGINE_GET_PKEY_ASN1_METH		 193
+#define ENGINE_F_ENGINE_GET_PKEY_METH			 192
 #define ENGINE_F_ENGINE_GET_PREV			 116
 #define ENGINE_F_ENGINE_INIT				 119
 #define ENGINE_F_ENGINE_LIST_ADD			 120
 #define ENGINE_F_ENGINE_LIST_REMOVE			 121
 #define ENGINE_F_ENGINE_LOAD_PRIVATE_KEY		 150
 #define ENGINE_F_ENGINE_LOAD_PUBLIC_KEY			 151
-#define ENGINE_F_ENGINE_LOAD_SSL_CLIENT_CERT		 192
+#define ENGINE_F_ENGINE_LOAD_SSL_CLIENT_CERT		 194
 #define ENGINE_F_ENGINE_NEW				 122
 #define ENGINE_F_ENGINE_REMOVE				 123
 #define ENGINE_F_ENGINE_SET_DEFAULT_STRING		 189
@@ -769,7 +806,7 @@ void ERR_load_ENGINE_strings(void);
 #define ENGINE_R_DSO_FAILURE				 104
 #define ENGINE_R_DSO_NOT_FOUND				 132
 #define ENGINE_R_ENGINES_SECTION_ERROR			 148
-#define ENGINE_R_ENGINE_CONFIGURATION_ERROR		 101
+#define ENGINE_R_ENGINE_CONFIGURATION_ERROR		 102
 #define ENGINE_R_ENGINE_IS_NOT_IN_LIST			 105
 #define ENGINE_R_ENGINE_SECTION_ERROR			 149
 #define ENGINE_R_FAILED_LOADING_PRIVATE_KEY		 128
@@ -796,6 +833,7 @@ void ERR_load_ENGINE_strings(void);
 #define ENGINE_R_RSA_NOT_IMPLEMENTED			 141
 #define ENGINE_R_UNIMPLEMENTED_CIPHER			 146
 #define ENGINE_R_UNIMPLEMENTED_DIGEST			 147
+#define ENGINE_R_UNIMPLEMENTED_PUBLIC_KEY_METHOD	 101
 #define ENGINE_R_VERSION_INCOMPATIBILITY		 145
 
 #ifdef  __cplusplus
diff --git a/crypto/engine/enginetest.c b/crypto/engine/enginetest.c
index e3834611dbdb..f4d70e7e0a60 100644
--- a/crypto/engine/enginetest.c
+++ b/crypto/engine/enginetest.c
@@ -276,7 +276,7 @@ end:
 	ENGINE_cleanup();
 	CRYPTO_cleanup_all_ex_data();
 	ERR_free_strings();
-	ERR_remove_state(0);
+	ERR_remove_thread_state(NULL);
 	CRYPTO_mem_leaks_fp(stderr);
 	return to_return;
 	}
diff --git a/crypto/engine/tb_asnmth.c b/crypto/engine/tb_asnmth.c
new file mode 100644
index 000000000000..75090339f7a2
--- /dev/null
+++ b/crypto/engine/tb_asnmth.c
@@ -0,0 +1,246 @@
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include "eng_int.h"
+#include "asn1_locl.h"
+#include <openssl/evp.h>
+
+/* If this symbol is defined then ENGINE_get_pkey_asn1_meth_engine(), the
+ * function that is used by EVP to hook in pkey_asn1_meth code and cache
+ * defaults (etc), will display brief debugging summaries to stderr with the
+ * 'nid'. */
+/* #define ENGINE_PKEY_ASN1_METH_DEBUG */
+
+static ENGINE_TABLE *pkey_asn1_meth_table = NULL;
+
+void ENGINE_unregister_pkey_asn1_meths(ENGINE *e)
+	{
+	engine_table_unregister(&pkey_asn1_meth_table, e);
+	}
+
+static void engine_unregister_all_pkey_asn1_meths(void)
+	{
+	engine_table_cleanup(&pkey_asn1_meth_table);
+	}
+
+int ENGINE_register_pkey_asn1_meths(ENGINE *e)
+	{
+	if(e->pkey_asn1_meths)
+		{
+		const int *nids;
+		int num_nids = e->pkey_asn1_meths(e, NULL, &nids, 0);
+		if(num_nids > 0)
+			return engine_table_register(&pkey_asn1_meth_table,
+				engine_unregister_all_pkey_asn1_meths, e, nids,
+					num_nids, 0);
+		}
+	return 1;
+	}
+
+void ENGINE_register_all_pkey_asn1_meths(void)
+	{
+	ENGINE *e;
+
+	for(e=ENGINE_get_first() ; e ; e=ENGINE_get_next(e))
+		ENGINE_register_pkey_asn1_meths(e);
+	}
+
+int ENGINE_set_default_pkey_asn1_meths(ENGINE *e)
+	{
+	if(e->pkey_asn1_meths)
+		{
+		const int *nids;
+		int num_nids = e->pkey_asn1_meths(e, NULL, &nids, 0);
+		if(num_nids > 0)
+			return engine_table_register(&pkey_asn1_meth_table,
+				engine_unregister_all_pkey_asn1_meths, e, nids,
+					num_nids, 1);
+		}
+	return 1;
+	}
+
+/* Exposed API function to get a functional reference from the implementation
+ * table (ie. try to get a functional reference from the tabled structural
+ * references) for a given pkey_asn1_meth 'nid' */
+ENGINE *ENGINE_get_pkey_asn1_meth_engine(int nid)
+	{
+	return engine_table_select(&pkey_asn1_meth_table, nid);
+	}
+
+/* Obtains a pkey_asn1_meth implementation from an ENGINE functional reference */
+const EVP_PKEY_ASN1_METHOD *ENGINE_get_pkey_asn1_meth(ENGINE *e, int nid)
+	{
+	EVP_PKEY_ASN1_METHOD *ret;
+	ENGINE_PKEY_ASN1_METHS_PTR fn = ENGINE_get_pkey_asn1_meths(e);
+	if(!fn || !fn(e, &ret, NULL, nid))
+		{
+		ENGINEerr(ENGINE_F_ENGINE_GET_PKEY_ASN1_METH,
+				ENGINE_R_UNIMPLEMENTED_PUBLIC_KEY_METHOD);
+		return NULL;
+		}
+	return ret;
+	}
+
+/* Gets the pkey_asn1_meth callback from an ENGINE structure */
+ENGINE_PKEY_ASN1_METHS_PTR ENGINE_get_pkey_asn1_meths(const ENGINE *e)
+	{
+	return e->pkey_asn1_meths;
+	}
+
+/* Sets the pkey_asn1_meth callback in an ENGINE structure */
+int ENGINE_set_pkey_asn1_meths(ENGINE *e, ENGINE_PKEY_ASN1_METHS_PTR f)
+	{
+	e->pkey_asn1_meths = f;
+	return 1;
+	}
+
+/* Internal function to free up EVP_PKEY_ASN1_METHOD structures before an
+ * ENGINE is destroyed
+ */
+
+void engine_pkey_asn1_meths_free(ENGINE *e)
+	{
+	int i;
+	EVP_PKEY_ASN1_METHOD *pkm;
+	if (e->pkey_asn1_meths)
+		{
+		const int *pknids;
+		int npknids;
+		npknids = e->pkey_asn1_meths(e, NULL, &pknids, 0);
+		for (i = 0; i < npknids; i++)
+			{
+			if (e->pkey_asn1_meths(e, &pkm, NULL, pknids[i]))
+				{
+				EVP_PKEY_asn1_free(pkm);
+				}
+			}
+		}
+	}
+
+/* Find a method based on a string. This does a linear search through
+ * all implemented algorithms. This is OK in practice because only
+ * a small number of algorithms are likely to be implemented in an engine
+ * and it is not used for speed critical operations.
+ */
+
+const EVP_PKEY_ASN1_METHOD *ENGINE_get_pkey_asn1_meth_str(ENGINE *e,
+					const char *str, int len)
+	{
+	int i, nidcount;
+	const int *nids;
+	EVP_PKEY_ASN1_METHOD *ameth;
+	if (!e->pkey_asn1_meths)
+		return NULL;
+	if (len == -1)
+		len = strlen(str);
+	nidcount = e->pkey_asn1_meths(e, NULL, &nids, 0);
+	for (i = 0; i < nidcount; i++)
+		{
+		e->pkey_asn1_meths(e, &ameth, NULL, nids[i]);
+		if (((int)strlen(ameth->pem_str) == len) && 
+					!strncasecmp(ameth->pem_str, str, len))
+			return ameth;
+		}
+	return NULL;
+	}
+
+typedef struct 
+	{
+	ENGINE *e;
+	const EVP_PKEY_ASN1_METHOD *ameth;
+	const char *str;
+	int len;
+	} ENGINE_FIND_STR;
+
+static void look_str_cb(int nid, STACK_OF(ENGINE) *sk, ENGINE *def, void *arg)
+	{
+	ENGINE_FIND_STR *lk = arg;
+	int i;
+	if (lk->ameth)
+		return;
+	for (i = 0; i < sk_ENGINE_num(sk); i++)
+		{
+		ENGINE *e = sk_ENGINE_value(sk, i);
+		EVP_PKEY_ASN1_METHOD *ameth;
+		e->pkey_asn1_meths(e, &ameth, NULL, nid);
+		if (((int)strlen(ameth->pem_str) == lk->len) && 
+				!strncasecmp(ameth->pem_str, lk->str, lk->len))
+			{
+			lk->e = e;
+			lk->ameth = ameth;
+			return;
+			}
+		}
+	}
+
+const EVP_PKEY_ASN1_METHOD *ENGINE_pkey_asn1_find_str(ENGINE **pe,
+					const char *str, int len)
+	{
+	ENGINE_FIND_STR fstr;
+	fstr.e = NULL;
+	fstr.ameth = NULL;
+	fstr.str = str;
+	fstr.len = len;
+	CRYPTO_w_lock(CRYPTO_LOCK_ENGINE);
+	engine_table_doall(pkey_asn1_meth_table, look_str_cb, &fstr);
+	/* If found obtain a structural reference to engine */
+	if (fstr.e)
+		{
+		fstr.e->struct_ref++;
+		engine_ref_debug(fstr.e, 0, 1)
+		}
+	*pe = fstr.e;
+	CRYPTO_w_unlock(CRYPTO_LOCK_ENGINE);
+	return fstr.ameth;
+	}
diff --git a/crypto/engine/tb_pkmeth.c b/crypto/engine/tb_pkmeth.c
new file mode 100644
index 000000000000..1cdb967f2536
--- /dev/null
+++ b/crypto/engine/tb_pkmeth.c
@@ -0,0 +1,167 @@
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include "eng_int.h"
+#include <openssl/evp.h>
+
+/* If this symbol is defined then ENGINE_get_pkey_meth_engine(), the function
+ * that is used by EVP to hook in pkey_meth code and cache defaults (etc), will
+ * display brief debugging summaries to stderr with the 'nid'. */
+/* #define ENGINE_PKEY_METH_DEBUG */
+
+static ENGINE_TABLE *pkey_meth_table = NULL;
+
+void ENGINE_unregister_pkey_meths(ENGINE *e)
+	{
+	engine_table_unregister(&pkey_meth_table, e);
+	}
+
+static void engine_unregister_all_pkey_meths(void)
+	{
+	engine_table_cleanup(&pkey_meth_table);
+	}
+
+int ENGINE_register_pkey_meths(ENGINE *e)
+	{
+	if(e->pkey_meths)
+		{
+		const int *nids;
+		int num_nids = e->pkey_meths(e, NULL, &nids, 0);
+		if(num_nids > 0)
+			return engine_table_register(&pkey_meth_table,
+				engine_unregister_all_pkey_meths, e, nids,
+					num_nids, 0);
+		}
+	return 1;
+	}
+
+void ENGINE_register_all_pkey_meths()
+	{
+	ENGINE *e;
+
+	for(e=ENGINE_get_first() ; e ; e=ENGINE_get_next(e))
+		ENGINE_register_pkey_meths(e);
+	}
+
+int ENGINE_set_default_pkey_meths(ENGINE *e)
+	{
+	if(e->pkey_meths)
+		{
+		const int *nids;
+		int num_nids = e->pkey_meths(e, NULL, &nids, 0);
+		if(num_nids > 0)
+			return engine_table_register(&pkey_meth_table,
+				engine_unregister_all_pkey_meths, e, nids,
+					num_nids, 1);
+		}
+	return 1;
+	}
+
+/* Exposed API function to get a functional reference from the implementation
+ * table (ie. try to get a functional reference from the tabled structural
+ * references) for a given pkey_meth 'nid' */
+ENGINE *ENGINE_get_pkey_meth_engine(int nid)
+	{
+	return engine_table_select(&pkey_meth_table, nid);
+	}
+
+/* Obtains a pkey_meth implementation from an ENGINE functional reference */
+const EVP_PKEY_METHOD *ENGINE_get_pkey_meth(ENGINE *e, int nid)
+	{
+	EVP_PKEY_METHOD *ret;
+	ENGINE_PKEY_METHS_PTR fn = ENGINE_get_pkey_meths(e);
+	if(!fn || !fn(e, &ret, NULL, nid))
+		{
+		ENGINEerr(ENGINE_F_ENGINE_GET_PKEY_METH,
+				ENGINE_R_UNIMPLEMENTED_PUBLIC_KEY_METHOD);
+		return NULL;
+		}
+	return ret;
+	}
+
+/* Gets the pkey_meth callback from an ENGINE structure */
+ENGINE_PKEY_METHS_PTR ENGINE_get_pkey_meths(const ENGINE *e)
+	{
+	return e->pkey_meths;
+	}
+
+/* Sets the pkey_meth callback in an ENGINE structure */
+int ENGINE_set_pkey_meths(ENGINE *e, ENGINE_PKEY_METHS_PTR f)
+	{
+	e->pkey_meths = f;
+	return 1;
+	}
+
+/* Internal function to free up EVP_PKEY_METHOD structures before an
+ * ENGINE is destroyed
+ */
+
+void engine_pkey_meths_free(ENGINE *e)
+	{
+	int i;
+	EVP_PKEY_METHOD *pkm;
+	if (e->pkey_meths)
+		{
+		const int *pknids;
+		int npknids;
+		npknids = e->pkey_meths(e, NULL, &pknids, 0);
+		for (i = 0; i < npknids; i++)
+			{
+			if (e->pkey_meths(e, &pkm, NULL, pknids[i]))
+				{
+				EVP_PKEY_meth_free(pkm);
+				}
+			}
+		}
+	}
diff --git a/crypto/err/Makefile b/crypto/err/Makefile
index 96d8a1a45eae..862b23ba1768 100644
--- a/crypto/err/Makefile
+++ b/crypto/err/Makefile
@@ -17,8 +17,8 @@ TEST=
 APPS=
 
 LIB=$(TOP)/libcrypto.a
-LIBSRC=err.c err_def.c err_all.c err_prn.c err_str.c err_bio.c
-LIBOBJ=err.o err_def.o err_all.o err_prn.o err_str.o err_bio.o
+LIBSRC=err.c err_all.c err_prn.c
+LIBOBJ=err.o err_all.o err_prn.o
 
 SRC= $(LIBSRC)
 
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -83,13 +83,13 @@ err.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 err.o: ../cryptlib.h err.c
 err_all.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 err_all.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
-err_all.o: ../../include/openssl/comp.h ../../include/openssl/conf.h
-err_all.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
-err_all.o: ../../include/openssl/dsa.h ../../include/openssl/dso.h
-err_all.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
-err_all.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-err_all.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-err_all.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
+err_all.o: ../../include/openssl/cms.h ../../include/openssl/comp.h
+err_all.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+err_all.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
+err_all.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
+err_all.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+err_all.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
+err_all.o: ../../include/openssl/err.h ../../include/openssl/evp.h
 err_all.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
 err_all.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
 err_all.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
@@ -98,23 +98,9 @@ err_all.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
 err_all.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
 err_all.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 err_all.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-err_all.o: ../../include/openssl/ui.h ../../include/openssl/x509.h
-err_all.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-err_all.o: err_all.c
-err_bio.o: ../../e_os.h ../../include/openssl/bio.h
-err_bio.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-err_bio.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-err_bio.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
-err_bio.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-err_bio.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-err_bio.o: ../../include/openssl/symhacks.h ../cryptlib.h err_bio.c
-err_def.o: ../../e_os.h ../../include/openssl/bio.h
-err_def.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-err_def.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-err_def.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
-err_def.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-err_def.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-err_def.o: ../../include/openssl/symhacks.h ../cryptlib.h err_def.c
+err_all.o: ../../include/openssl/ts.h ../../include/openssl/ui.h
+err_all.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+err_all.o: ../../include/openssl/x509v3.h err_all.c
 err_prn.o: ../../e_os.h ../../include/openssl/bio.h
 err_prn.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 err_prn.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
@@ -122,10 +108,3 @@ err_prn.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 err_prn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 err_prn.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 err_prn.o: ../../include/openssl/symhacks.h ../cryptlib.h err_prn.c
-err_str.o: ../../e_os.h ../../include/openssl/bio.h
-err_str.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-err_str.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-err_str.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
-err_str.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-err_str.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-err_str.o: ../../include/openssl/symhacks.h ../cryptlib.h err_str.c
diff --git a/crypto/err/err.c b/crypto/err/err.c
index 292404a2fb99..fcdb244008f6 100644
--- a/crypto/err/err.c
+++ b/crypto/err/err.c
@@ -56,7 +56,7 @@
  * [including the GNU Public Licence.]
  */
 /* ====================================================================
- * Copyright (c) 1998-2001 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1998-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -119,9 +119,507 @@
 #include <openssl/bio.h>
 #include <openssl/err.h>
 
-static unsigned long get_error_values(int inc,int top,
-					const char **file,int *line,
-					const char **data,int *flags);
+DECLARE_LHASH_OF(ERR_STRING_DATA);
+DECLARE_LHASH_OF(ERR_STATE);
+
+static void err_load_strings(int lib, ERR_STRING_DATA *str);
+
+static void ERR_STATE_free(ERR_STATE *s);
+#ifndef OPENSSL_NO_ERR
+static ERR_STRING_DATA ERR_str_libraries[]=
+	{
+{ERR_PACK(ERR_LIB_NONE,0,0)		,"unknown library"},
+{ERR_PACK(ERR_LIB_SYS,0,0)		,"system library"},
+{ERR_PACK(ERR_LIB_BN,0,0)		,"bignum routines"},
+{ERR_PACK(ERR_LIB_RSA,0,0)		,"rsa routines"},
+{ERR_PACK(ERR_LIB_DH,0,0)		,"Diffie-Hellman routines"},
+{ERR_PACK(ERR_LIB_EVP,0,0)		,"digital envelope routines"},
+{ERR_PACK(ERR_LIB_BUF,0,0)		,"memory buffer routines"},
+{ERR_PACK(ERR_LIB_OBJ,0,0)		,"object identifier routines"},
+{ERR_PACK(ERR_LIB_PEM,0,0)		,"PEM routines"},
+{ERR_PACK(ERR_LIB_DSA,0,0)		,"dsa routines"},
+{ERR_PACK(ERR_LIB_X509,0,0)		,"x509 certificate routines"},
+{ERR_PACK(ERR_LIB_ASN1,0,0)		,"asn1 encoding routines"},
+{ERR_PACK(ERR_LIB_CONF,0,0)		,"configuration file routines"},
+{ERR_PACK(ERR_LIB_CRYPTO,0,0)		,"common libcrypto routines"},
+{ERR_PACK(ERR_LIB_EC,0,0)		,"elliptic curve routines"},
+{ERR_PACK(ERR_LIB_SSL,0,0)		,"SSL routines"},
+{ERR_PACK(ERR_LIB_BIO,0,0)		,"BIO routines"},
+{ERR_PACK(ERR_LIB_PKCS7,0,0)		,"PKCS7 routines"},
+{ERR_PACK(ERR_LIB_X509V3,0,0)		,"X509 V3 routines"},
+{ERR_PACK(ERR_LIB_PKCS12,0,0)		,"PKCS12 routines"},
+{ERR_PACK(ERR_LIB_RAND,0,0)		,"random number generator"},
+{ERR_PACK(ERR_LIB_DSO,0,0)		,"DSO support routines"},
+{ERR_PACK(ERR_LIB_TS,0,0)		,"time stamp routines"},
+{ERR_PACK(ERR_LIB_ENGINE,0,0)		,"engine routines"},
+{ERR_PACK(ERR_LIB_OCSP,0,0)		,"OCSP routines"},
+{ERR_PACK(ERR_LIB_FIPS,0,0)		,"FIPS routines"},
+{ERR_PACK(ERR_LIB_CMS,0,0)		,"CMS routines"},
+{ERR_PACK(ERR_LIB_HMAC,0,0)		,"HMAC routines"},
+{0,NULL},
+	};
+
+static ERR_STRING_DATA ERR_str_functs[]=
+	{
+	{ERR_PACK(0,SYS_F_FOPEN,0),     	"fopen"},
+	{ERR_PACK(0,SYS_F_CONNECT,0),		"connect"},
+	{ERR_PACK(0,SYS_F_GETSERVBYNAME,0),	"getservbyname"},
+	{ERR_PACK(0,SYS_F_SOCKET,0),		"socket"}, 
+	{ERR_PACK(0,SYS_F_IOCTLSOCKET,0),	"ioctlsocket"},
+	{ERR_PACK(0,SYS_F_BIND,0),		"bind"},
+	{ERR_PACK(0,SYS_F_LISTEN,0),		"listen"},
+	{ERR_PACK(0,SYS_F_ACCEPT,0),		"accept"},
+#ifdef OPENSSL_SYS_WINDOWS
+	{ERR_PACK(0,SYS_F_WSASTARTUP,0),	"WSAstartup"},
+#endif
+	{ERR_PACK(0,SYS_F_OPENDIR,0),		"opendir"},
+	{ERR_PACK(0,SYS_F_FREAD,0),		"fread"},
+	{0,NULL},
+	};
+
+static ERR_STRING_DATA ERR_str_reasons[]=
+	{
+{ERR_R_SYS_LIB				,"system lib"},
+{ERR_R_BN_LIB				,"BN lib"},
+{ERR_R_RSA_LIB				,"RSA lib"},
+{ERR_R_DH_LIB				,"DH lib"},
+{ERR_R_EVP_LIB				,"EVP lib"},
+{ERR_R_BUF_LIB				,"BUF lib"},
+{ERR_R_OBJ_LIB				,"OBJ lib"},
+{ERR_R_PEM_LIB				,"PEM lib"},
+{ERR_R_DSA_LIB				,"DSA lib"},
+{ERR_R_X509_LIB				,"X509 lib"},
+{ERR_R_ASN1_LIB				,"ASN1 lib"},
+{ERR_R_CONF_LIB				,"CONF lib"},
+{ERR_R_CRYPTO_LIB			,"CRYPTO lib"},
+{ERR_R_EC_LIB				,"EC lib"},
+{ERR_R_SSL_LIB				,"SSL lib"},
+{ERR_R_BIO_LIB				,"BIO lib"},
+{ERR_R_PKCS7_LIB			,"PKCS7 lib"},
+{ERR_R_X509V3_LIB			,"X509V3 lib"},
+{ERR_R_PKCS12_LIB			,"PKCS12 lib"},
+{ERR_R_RAND_LIB				,"RAND lib"},
+{ERR_R_DSO_LIB				,"DSO lib"},
+{ERR_R_ENGINE_LIB			,"ENGINE lib"},
+{ERR_R_OCSP_LIB				,"OCSP lib"},
+{ERR_R_TS_LIB				,"TS lib"},
+
+{ERR_R_NESTED_ASN1_ERROR		,"nested asn1 error"},
+{ERR_R_BAD_ASN1_OBJECT_HEADER		,"bad asn1 object header"},
+{ERR_R_BAD_GET_ASN1_OBJECT_CALL		,"bad get asn1 object call"},
+{ERR_R_EXPECTING_AN_ASN1_SEQUENCE	,"expecting an asn1 sequence"},
+{ERR_R_ASN1_LENGTH_MISMATCH		,"asn1 length mismatch"},
+{ERR_R_MISSING_ASN1_EOS			,"missing asn1 eos"},
+
+{ERR_R_FATAL                            ,"fatal"},
+{ERR_R_MALLOC_FAILURE			,"malloc failure"},
+{ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED	,"called a function you should not call"},
+{ERR_R_PASSED_NULL_PARAMETER		,"passed a null parameter"},
+{ERR_R_INTERNAL_ERROR			,"internal error"},
+{ERR_R_DISABLED				,"called a function that was disabled at compile-time"},
+
+{0,NULL},
+	};
+#endif
+
+
+/* Define the predeclared (but externally opaque) "ERR_FNS" type */
+struct st_ERR_FNS
+	{
+	/* Works on the "error_hash" string table */
+	LHASH_OF(ERR_STRING_DATA) *(*cb_err_get)(int create);
+	void (*cb_err_del)(void);
+	ERR_STRING_DATA *(*cb_err_get_item)(const ERR_STRING_DATA *);
+	ERR_STRING_DATA *(*cb_err_set_item)(ERR_STRING_DATA *);
+	ERR_STRING_DATA *(*cb_err_del_item)(ERR_STRING_DATA *);
+	/* Works on the "thread_hash" error-state table */
+	LHASH_OF(ERR_STATE) *(*cb_thread_get)(int create);
+	void (*cb_thread_release)(LHASH_OF(ERR_STATE) **hash);
+	ERR_STATE *(*cb_thread_get_item)(const ERR_STATE *);
+	ERR_STATE *(*cb_thread_set_item)(ERR_STATE *);
+	void (*cb_thread_del_item)(const ERR_STATE *);
+	/* Returns the next available error "library" numbers */
+	int (*cb_get_next_lib)(void);
+	};
+
+/* Predeclarations of the "err_defaults" functions */
+static LHASH_OF(ERR_STRING_DATA) *int_err_get(int create);
+static void int_err_del(void);
+static ERR_STRING_DATA *int_err_get_item(const ERR_STRING_DATA *);
+static ERR_STRING_DATA *int_err_set_item(ERR_STRING_DATA *);
+static ERR_STRING_DATA *int_err_del_item(ERR_STRING_DATA *);
+static LHASH_OF(ERR_STATE) *int_thread_get(int create);
+static void int_thread_release(LHASH_OF(ERR_STATE) **hash);
+static ERR_STATE *int_thread_get_item(const ERR_STATE *);
+static ERR_STATE *int_thread_set_item(ERR_STATE *);
+static void int_thread_del_item(const ERR_STATE *);
+static int int_err_get_next_lib(void);
+/* The static ERR_FNS table using these defaults functions */
+static const ERR_FNS err_defaults =
+	{
+	int_err_get,
+	int_err_del,
+	int_err_get_item,
+	int_err_set_item,
+	int_err_del_item,
+	int_thread_get,
+	int_thread_release,
+	int_thread_get_item,
+	int_thread_set_item,
+	int_thread_del_item,
+	int_err_get_next_lib
+	};
+
+/* The replacable table of ERR_FNS functions we use at run-time */
+static const ERR_FNS *err_fns = NULL;
+
+/* Eg. rather than using "err_get()", use "ERRFN(err_get)()". */
+#define ERRFN(a) err_fns->cb_##a
+
+/* The internal state used by "err_defaults" - as such, the setting, reading,
+ * creating, and deleting of this data should only be permitted via the
+ * "err_defaults" functions. This way, a linked module can completely defer all
+ * ERR state operation (together with requisite locking) to the implementations
+ * and state in the loading application. */
+static LHASH_OF(ERR_STRING_DATA) *int_error_hash = NULL;
+static LHASH_OF(ERR_STATE) *int_thread_hash = NULL;
+static int int_thread_hash_references = 0;
+static int int_err_library_number= ERR_LIB_USER;
+
+/* Internal function that checks whether "err_fns" is set and if not, sets it to
+ * the defaults. */
+static void err_fns_check(void)
+	{
+	if (err_fns) return;
+	
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	if (!err_fns)
+		err_fns = &err_defaults;
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+	}
+
+/* API functions to get or set the underlying ERR functions. */
+
+const ERR_FNS *ERR_get_implementation(void)
+	{
+	err_fns_check();
+	return err_fns;
+	}
+
+int ERR_set_implementation(const ERR_FNS *fns)
+	{
+	int ret = 0;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	/* It's too late if 'err_fns' is non-NULL. BTW: not much point setting
+	 * an error is there?! */
+	if (!err_fns)
+		{
+		err_fns = fns;
+		ret = 1;
+		}
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+	return ret;
+	}
+
+/* These are the callbacks provided to "lh_new()" when creating the LHASH tables
+ * internal to the "err_defaults" implementation. */
+
+static unsigned long get_error_values(int inc,int top,const char **file,int *line,
+				      const char **data,int *flags);
+
+/* The internal functions used in the "err_defaults" implementation */
+
+static unsigned long err_string_data_hash(const ERR_STRING_DATA *a)
+	{
+	unsigned long ret,l;
+
+	l=a->error;
+	ret=l^ERR_GET_LIB(l)^ERR_GET_FUNC(l);
+	return(ret^ret%19*13);
+	}
+static IMPLEMENT_LHASH_HASH_FN(err_string_data, ERR_STRING_DATA)
+
+static int err_string_data_cmp(const ERR_STRING_DATA *a,
+			       const ERR_STRING_DATA *b)
+	{
+	return (int)(a->error - b->error);
+	}
+static IMPLEMENT_LHASH_COMP_FN(err_string_data, ERR_STRING_DATA)
+
+static LHASH_OF(ERR_STRING_DATA) *int_err_get(int create)
+	{
+	LHASH_OF(ERR_STRING_DATA) *ret = NULL;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	if (!int_error_hash && create)
+		{
+		CRYPTO_push_info("int_err_get (err.c)");
+		int_error_hash = lh_ERR_STRING_DATA_new();
+		CRYPTO_pop_info();
+		}
+	if (int_error_hash)
+		ret = int_error_hash;
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+
+	return ret;
+	}
+
+static void int_err_del(void)
+	{
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	if (int_error_hash)
+		{
+		lh_ERR_STRING_DATA_free(int_error_hash);
+		int_error_hash = NULL;
+		}
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+	}
+
+static ERR_STRING_DATA *int_err_get_item(const ERR_STRING_DATA *d)
+	{
+	ERR_STRING_DATA *p;
+	LHASH_OF(ERR_STRING_DATA) *hash;
+
+	err_fns_check();
+	hash = ERRFN(err_get)(0);
+	if (!hash)
+		return NULL;
+
+	CRYPTO_r_lock(CRYPTO_LOCK_ERR);
+	p = lh_ERR_STRING_DATA_retrieve(hash, d);
+	CRYPTO_r_unlock(CRYPTO_LOCK_ERR);
+
+	return p;
+	}
+
+static ERR_STRING_DATA *int_err_set_item(ERR_STRING_DATA *d)
+	{
+	ERR_STRING_DATA *p;
+	LHASH_OF(ERR_STRING_DATA) *hash;
+
+	err_fns_check();
+	hash = ERRFN(err_get)(1);
+	if (!hash)
+		return NULL;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	p = lh_ERR_STRING_DATA_insert(hash, d);
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+
+	return p;
+	}
+
+static ERR_STRING_DATA *int_err_del_item(ERR_STRING_DATA *d)
+	{
+	ERR_STRING_DATA *p;
+	LHASH_OF(ERR_STRING_DATA) *hash;
+
+	err_fns_check();
+	hash = ERRFN(err_get)(0);
+	if (!hash)
+		return NULL;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	p = lh_ERR_STRING_DATA_delete(hash, d);
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+
+	return p;
+	}
+
+static unsigned long err_state_hash(const ERR_STATE *a)
+	{
+	return CRYPTO_THREADID_hash(&a->tid) * 13;
+	}
+static IMPLEMENT_LHASH_HASH_FN(err_state, ERR_STATE)
+
+static int err_state_cmp(const ERR_STATE *a, const ERR_STATE *b)
+	{
+	return CRYPTO_THREADID_cmp(&a->tid, &b->tid);
+	}
+static IMPLEMENT_LHASH_COMP_FN(err_state, ERR_STATE)
+
+static LHASH_OF(ERR_STATE) *int_thread_get(int create)
+	{
+	LHASH_OF(ERR_STATE) *ret = NULL;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	if (!int_thread_hash && create)
+		{
+		CRYPTO_push_info("int_thread_get (err.c)");
+		int_thread_hash = lh_ERR_STATE_new();
+		CRYPTO_pop_info();
+		}
+	if (int_thread_hash)
+		{
+		int_thread_hash_references++;
+		ret = int_thread_hash;
+		}
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+	return ret;
+	}
+
+static void int_thread_release(LHASH_OF(ERR_STATE) **hash)
+	{
+	int i;
+
+	if (hash == NULL || *hash == NULL)
+		return;
+
+	i = CRYPTO_add(&int_thread_hash_references, -1, CRYPTO_LOCK_ERR);
+
+#ifdef REF_PRINT
+	fprintf(stderr,"%4d:%s\n",int_thread_hash_references,"ERR");
+#endif
+	if (i > 0) return;
+#ifdef REF_CHECK
+	if (i < 0)
+		{
+		fprintf(stderr,"int_thread_release, bad reference count\n");
+		abort(); /* ok */
+		}
+#endif
+	*hash = NULL;
+	}
+
+static ERR_STATE *int_thread_get_item(const ERR_STATE *d)
+	{
+	ERR_STATE *p;
+	LHASH_OF(ERR_STATE) *hash;
+
+	err_fns_check();
+	hash = ERRFN(thread_get)(0);
+	if (!hash)
+		return NULL;
+
+	CRYPTO_r_lock(CRYPTO_LOCK_ERR);
+	p = lh_ERR_STATE_retrieve(hash, d);
+	CRYPTO_r_unlock(CRYPTO_LOCK_ERR);
+
+	ERRFN(thread_release)(&hash);
+	return p;
+	}
+
+static ERR_STATE *int_thread_set_item(ERR_STATE *d)
+	{
+	ERR_STATE *p;
+	LHASH_OF(ERR_STATE) *hash;
+
+	err_fns_check();
+	hash = ERRFN(thread_get)(1);
+	if (!hash)
+		return NULL;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	p = lh_ERR_STATE_insert(hash, d);
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+
+	ERRFN(thread_release)(&hash);
+	return p;
+	}
+
+static void int_thread_del_item(const ERR_STATE *d)
+	{
+	ERR_STATE *p;
+	LHASH_OF(ERR_STATE) *hash;
+
+	err_fns_check();
+	hash = ERRFN(thread_get)(0);
+	if (!hash)
+		return;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	p = lh_ERR_STATE_delete(hash, d);
+	/* make sure we don't leak memory */
+	if (int_thread_hash_references == 1
+	    && int_thread_hash && lh_ERR_STATE_num_items(int_thread_hash) == 0)
+		{
+		lh_ERR_STATE_free(int_thread_hash);
+		int_thread_hash = NULL;
+		}
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+
+	ERRFN(thread_release)(&hash);
+	if (p)
+		ERR_STATE_free(p);
+	}
+
+static int int_err_get_next_lib(void)
+	{
+	int ret;
+
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	ret = int_err_library_number++;
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+
+	return ret;
+	}
+
+
+#ifndef OPENSSL_NO_ERR
+#define NUM_SYS_STR_REASONS 127
+#define LEN_SYS_STR_REASON 32
+
+static ERR_STRING_DATA SYS_str_reasons[NUM_SYS_STR_REASONS + 1];
+/* SYS_str_reasons is filled with copies of strerror() results at
+ * initialization.
+ * 'errno' values up to 127 should cover all usual errors,
+ * others will be displayed numerically by ERR_error_string.
+ * It is crucial that we have something for each reason code
+ * that occurs in ERR_str_reasons, or bogus reason strings
+ * will be returned for SYSerr(), which always gets an errno
+ * value and never one of those 'standard' reason codes. */
+
+static void build_SYS_str_reasons(void)
+	{
+	/* OPENSSL_malloc cannot be used here, use static storage instead */
+	static char strerror_tab[NUM_SYS_STR_REASONS][LEN_SYS_STR_REASON];
+	int i;
+	static int init = 1;
+
+	CRYPTO_r_lock(CRYPTO_LOCK_ERR);
+	if (!init)
+		{
+		CRYPTO_r_unlock(CRYPTO_LOCK_ERR);
+		return;
+		}
+	
+	CRYPTO_r_unlock(CRYPTO_LOCK_ERR);
+	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
+	if (!init)
+		{
+		CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+		return;
+		}
+
+	for (i = 1; i <= NUM_SYS_STR_REASONS; i++)
+		{
+		ERR_STRING_DATA *str = &SYS_str_reasons[i - 1];
+
+		str->error = (unsigned long)i;
+		if (str->string == NULL)
+			{
+			char (*dest)[LEN_SYS_STR_REASON] = &(strerror_tab[i - 1]);
+			char *src = strerror(i);
+			if (src != NULL)
+				{
+				strncpy(*dest, src, sizeof *dest);
+				(*dest)[sizeof *dest - 1] = '\0';
+				str->string = *dest;
+				}
+			}
+		if (str->string == NULL)
+			str->string = "unknown";
+		}
+
+	/* Now we still have SYS_str_reasons[NUM_SYS_STR_REASONS] = {0, NULL},
+	 * as required by ERR_load_strings. */
+
+	init = 0;
+	
+	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
+	}
+#endif
 
 #define err_clear_data(p,i) \
 	do { \
@@ -143,6 +641,68 @@ static unsigned long get_error_values(int inc,int top,
 	(p)->err_line[i]= -1; \
 	} while(0)
 
+static void ERR_STATE_free(ERR_STATE *s)
+	{
+	int i;
+
+	if (s == NULL)
+	    return;
+
+	for (i=0; i<ERR_NUM_ERRORS; i++)
+		{
+		err_clear_data(s,i);
+		}
+	OPENSSL_free(s);
+	}
+
+void ERR_load_ERR_strings(void)
+	{
+	err_fns_check();
+#ifndef OPENSSL_NO_ERR
+	err_load_strings(0,ERR_str_libraries);
+	err_load_strings(0,ERR_str_reasons);
+	err_load_strings(ERR_LIB_SYS,ERR_str_functs);
+	build_SYS_str_reasons();
+	err_load_strings(ERR_LIB_SYS,SYS_str_reasons);
+#endif
+	}
+
+static void err_load_strings(int lib, ERR_STRING_DATA *str)
+	{
+	while (str->error)
+		{
+		if (lib)
+			str->error|=ERR_PACK(lib,0,0);
+		ERRFN(err_set_item)(str);
+		str++;
+		}
+	}
+
+void ERR_load_strings(int lib, ERR_STRING_DATA *str)
+	{
+	ERR_load_ERR_strings();
+	err_load_strings(lib, str);
+	}
+
+void ERR_unload_strings(int lib, ERR_STRING_DATA *str)
+	{
+	while (str->error)
+		{
+		if (lib)
+			str->error|=ERR_PACK(lib,0,0);
+		ERRFN(err_del_item)(str);
+		str++;
+		}
+	}
+
+void ERR_free_strings(void)
+	{
+	err_fns_check();
+	ERRFN(err_del)();
+	}
+
+/********************************************************/
+
 void ERR_put_error(int lib, int func, int reason, const char *file,
 	     int line)
 	{
@@ -297,6 +857,196 @@ static unsigned long get_error_values(int inc, int top, const char **file, int *
 	return ret;
 	}
 
+void ERR_error_string_n(unsigned long e, char *buf, size_t len)
+	{
+	char lsbuf[64], fsbuf[64], rsbuf[64];
+	const char *ls,*fs,*rs;
+	unsigned long l,f,r;
+
+	l=ERR_GET_LIB(e);
+	f=ERR_GET_FUNC(e);
+	r=ERR_GET_REASON(e);
+
+	ls=ERR_lib_error_string(e);
+	fs=ERR_func_error_string(e);
+	rs=ERR_reason_error_string(e);
+
+	if (ls == NULL) 
+		BIO_snprintf(lsbuf, sizeof(lsbuf), "lib(%lu)", l);
+	if (fs == NULL)
+		BIO_snprintf(fsbuf, sizeof(fsbuf), "func(%lu)", f);
+	if (rs == NULL)
+		BIO_snprintf(rsbuf, sizeof(rsbuf), "reason(%lu)", r);
+
+	BIO_snprintf(buf, len,"error:%08lX:%s:%s:%s", e, ls?ls:lsbuf, 
+		fs?fs:fsbuf, rs?rs:rsbuf);
+	if (strlen(buf) == len-1)
+		{
+		/* output may be truncated; make sure we always have 5 
+		 * colon-separated fields, i.e. 4 colons ... */
+#define NUM_COLONS 4
+		if (len > NUM_COLONS) /* ... if possible */
+			{
+			int i;
+			char *s = buf;
+			
+			for (i = 0; i < NUM_COLONS; i++)
+				{
+				char *colon = strchr(s, ':');
+				if (colon == NULL || colon > &buf[len-1] - NUM_COLONS + i)
+					{
+					/* set colon no. i at last possible position
+					 * (buf[len-1] is the terminating 0)*/
+					colon = &buf[len-1] - NUM_COLONS + i;
+					*colon = ':';
+					}
+				s = colon + 1;
+				}
+			}
+		}
+	}
+
+/* BAD for multi-threading: uses a local buffer if ret == NULL */
+/* ERR_error_string_n should be used instead for ret != NULL
+ * as ERR_error_string cannot know how large the buffer is */
+char *ERR_error_string(unsigned long e, char *ret)
+	{
+	static char buf[256];
+
+	if (ret == NULL) ret=buf;
+	ERR_error_string_n(e, ret, 256);
+
+	return ret;
+	}
+
+LHASH_OF(ERR_STRING_DATA) *ERR_get_string_table(void)
+	{
+	err_fns_check();
+	return ERRFN(err_get)(0);
+	}
+
+LHASH_OF(ERR_STATE) *ERR_get_err_state_table(void)
+	{
+	err_fns_check();
+	return ERRFN(thread_get)(0);
+	}
+
+void ERR_release_err_state_table(LHASH_OF(ERR_STATE) **hash)
+	{
+	err_fns_check();
+	ERRFN(thread_release)(hash);
+	}
+
+const char *ERR_lib_error_string(unsigned long e)
+	{
+	ERR_STRING_DATA d,*p;
+	unsigned long l;
+
+	err_fns_check();
+	l=ERR_GET_LIB(e);
+	d.error=ERR_PACK(l,0,0);
+	p=ERRFN(err_get_item)(&d);
+	return((p == NULL)?NULL:p->string);
+	}
+
+const char *ERR_func_error_string(unsigned long e)
+	{
+	ERR_STRING_DATA d,*p;
+	unsigned long l,f;
+
+	err_fns_check();
+	l=ERR_GET_LIB(e);
+	f=ERR_GET_FUNC(e);
+	d.error=ERR_PACK(l,f,0);
+	p=ERRFN(err_get_item)(&d);
+	return((p == NULL)?NULL:p->string);
+	}
+
+const char *ERR_reason_error_string(unsigned long e)
+	{
+	ERR_STRING_DATA d,*p=NULL;
+	unsigned long l,r;
+
+	err_fns_check();
+	l=ERR_GET_LIB(e);
+	r=ERR_GET_REASON(e);
+	d.error=ERR_PACK(l,0,r);
+	p=ERRFN(err_get_item)(&d);
+	if (!p)
+		{
+		d.error=ERR_PACK(0,0,r);
+		p=ERRFN(err_get_item)(&d);
+		}
+	return((p == NULL)?NULL:p->string);
+	}
+
+void ERR_remove_thread_state(const CRYPTO_THREADID *id)
+	{
+	ERR_STATE tmp;
+
+	if (id)
+		CRYPTO_THREADID_cpy(&tmp.tid, id);
+	else
+		CRYPTO_THREADID_current(&tmp.tid);
+	err_fns_check();
+	/* thread_del_item automatically destroys the LHASH if the number of
+	 * items reaches zero. */
+	ERRFN(thread_del_item)(&tmp);
+	}
+
+#ifndef OPENSSL_NO_DEPRECATED
+void ERR_remove_state(unsigned long pid)
+	{
+	ERR_remove_thread_state(NULL);
+	}
+#endif
+
+ERR_STATE *ERR_get_state(void)
+	{
+	static ERR_STATE fallback;
+	ERR_STATE *ret,tmp,*tmpp=NULL;
+	int i;
+	CRYPTO_THREADID tid;
+
+	err_fns_check();
+	CRYPTO_THREADID_current(&tid);
+	CRYPTO_THREADID_cpy(&tmp.tid, &tid);
+	ret=ERRFN(thread_get_item)(&tmp);
+
+	/* ret == the error state, if NULL, make a new one */
+	if (ret == NULL)
+		{
+		ret=(ERR_STATE *)OPENSSL_malloc(sizeof(ERR_STATE));
+		if (ret == NULL) return(&fallback);
+		CRYPTO_THREADID_cpy(&ret->tid, &tid);
+		ret->top=0;
+		ret->bottom=0;
+		for (i=0; i<ERR_NUM_ERRORS; i++)
+			{
+			ret->err_data[i]=NULL;
+			ret->err_data_flags[i]=0;
+			}
+		tmpp = ERRFN(thread_set_item)(ret);
+		/* To check if insertion failed, do a get. */
+		if (ERRFN(thread_get_item)(ret) != ret)
+			{
+			ERR_STATE_free(ret); /* could not insert it */
+			return(&fallback);
+			}
+		/* If a race occured in this function and we came second, tmpp
+		 * is the first one that we just replaced. */
+		if (tmpp)
+			ERR_STATE_free(tmpp);
+		}
+	return ret;
+	}
+
+int ERR_get_next_error_library(void)
+	{
+	err_fns_check();
+	return ERRFN(get_next_lib)();
+	}
+
 void ERR_set_error_data(char *data, int flags)
 	{
 	ERR_STATE *es;
@@ -316,6 +1066,13 @@ void ERR_set_error_data(char *data, int flags)
 void ERR_add_error_data(int num, ...)
 	{
 	va_list args;
+	va_start(args, num);
+	ERR_add_error_vdata(num, args);
+	va_end(args);
+	}
+
+void ERR_add_error_vdata(int num, va_list args)
+	{
 	int i,n,s;
 	char *str,*p,*a;
 
@@ -324,7 +1081,6 @@ void ERR_add_error_data(int num, ...)
 	if (str == NULL) return;
 	str[0]='\0';
 
-	va_start(args, num);
 	n=0;
 	for (i=0; i<num; i++)
 		{
@@ -340,7 +1096,7 @@ void ERR_add_error_data(int num, ...)
 				if (p == NULL)
 					{
 					OPENSSL_free(str);
-					goto err;
+					return;
 					}
 				else
 					str=p;
@@ -349,9 +1105,6 @@ void ERR_add_error_data(int num, ...)
 			}
 		}
 	ERR_set_error_data(str,ERR_TXT_MALLOCED|ERR_TXT_STRING);
-
-err:
-	va_end(args);
 	}
 
 int ERR_set_mark(void)
@@ -383,34 +1136,3 @@ int ERR_pop_to_mark(void)
 	es->err_flags[es->top]&=~ERR_FLAG_MARK;
 	return 1;
 	}
-
-#ifdef OPENSSL_FIPS
-
-static ERR_STATE *fget_state(void)
-	{
-	static ERR_STATE fstate;
-	return &fstate;
-	}
-
-ERR_STATE *(*get_state_func)(void) = fget_state;
-void (*remove_state_func)(unsigned long pid);
-
-ERR_STATE *ERR_get_state(void)
-	{
-	return get_state_func();
-	}
-
-void int_ERR_set_state_func(ERR_STATE *(*get_func)(void),
-				void (*remove_func)(unsigned long pid))
-	{
-	get_state_func = get_func;
-	remove_state_func = remove_func;
-	}
-
-void ERR_remove_state(unsigned long pid)
-	{
-	if (remove_state_func)
-		remove_state_func(pid);
-	}
-
-#endif
diff --git a/crypto/err/err.h b/crypto/err/err.h
index dcac41523101..974cc9cc6f90 100644
--- a/crypto/err/err.h
+++ b/crypto/err/err.h
@@ -55,6 +55,59 @@
  * copied and put under another distribution licence
  * [including the GNU Public Licence.]
  */
+/* ====================================================================
+ * Copyright (c) 1998-2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
 
 #ifndef HEADER_ERR_H
 #define HEADER_ERR_H
@@ -94,7 +147,7 @@ extern "C" {
 #define ERR_NUM_ERRORS	16
 typedef struct err_state_st
 	{
-	unsigned long pid;
+	CRYPTO_THREADID tid;
 	int err_flags[ERR_NUM_ERRORS];
 	unsigned long err_buffer[ERR_NUM_ERRORS];
 	char *err_data[ERR_NUM_ERRORS];
@@ -142,7 +195,9 @@ typedef struct err_state_st
 #define ERR_LIB_STORE           44
 #define ERR_LIB_FIPS		45
 #define ERR_LIB_CMS		46
-#define ERR_LIB_JPAKE		47
+#define ERR_LIB_TS		47
+#define ERR_LIB_HMAC		48
+#define ERR_LIB_JPAKE		49
 
 #define ERR_LIB_USER		128
 
@@ -176,6 +231,8 @@ typedef struct err_state_st
 #define STOREerr(f,r) ERR_PUT_error(ERR_LIB_STORE,(f),(r),__FILE__,__LINE__)
 #define FIPSerr(f,r) ERR_PUT_error(ERR_LIB_FIPS,(f),(r),__FILE__,__LINE__)
 #define CMSerr(f,r) ERR_PUT_error(ERR_LIB_CMS,(f),(r),__FILE__,__LINE__)
+#define TSerr(f,r) ERR_PUT_error(ERR_LIB_TS,(f),(r),__FILE__,__LINE__)
+#define HMACerr(f,r) ERR_PUT_error(ERR_LIB_HMAC,(f),(r),__FILE__,__LINE__)
 #define JPAKEerr(f,r) ERR_PUT_error(ERR_LIB_JPAKE,(f),(r),__FILE__,__LINE__)
 
 /* Borland C seems too stupid to be able to shift and do longs in
@@ -232,6 +289,7 @@ typedef struct err_state_st
 #define ERR_R_ECDSA_LIB ERR_LIB_ECDSA	 /* 42 */
 #define ERR_R_ECDH_LIB  ERR_LIB_ECDH	 /* 43 */
 #define ERR_R_STORE_LIB ERR_LIB_STORE    /* 44 */
+#define ERR_R_TS_LIB	ERR_LIB_TS       /* 45 */
 
 #define ERR_R_NESTED_ASN1_ERROR			58
 #define ERR_R_BAD_ASN1_OBJECT_HEADER		59
@@ -286,21 +344,25 @@ void ERR_print_errors_fp(FILE *fp);
 #endif
 #ifndef OPENSSL_NO_BIO
 void ERR_print_errors(BIO *bp);
-void ERR_add_error_data(int num, ...);
 #endif
+void ERR_add_error_data(int num, ...);
+void ERR_add_error_vdata(int num, va_list args);
 void ERR_load_strings(int lib,ERR_STRING_DATA str[]);
 void ERR_unload_strings(int lib,ERR_STRING_DATA str[]);
 void ERR_load_ERR_strings(void);
 void ERR_load_crypto_strings(void);
 void ERR_free_strings(void);
 
+void ERR_remove_thread_state(const CRYPTO_THREADID *tid);
+#ifndef OPENSSL_NO_DEPRECATED
 void ERR_remove_state(unsigned long pid); /* if zero we look it up */
+#endif
 ERR_STATE *ERR_get_state(void);
 
 #ifndef OPENSSL_NO_LHASH
-LHASH *ERR_get_string_table(void);
-LHASH *ERR_get_err_state_table(void);
-void ERR_release_err_state_table(LHASH **hash);
+LHASH_OF(ERR_STRING_DATA) *ERR_get_string_table(void);
+LHASH_OF(ERR_STATE) *ERR_get_err_state_table(void);
+void ERR_release_err_state_table(LHASH_OF(ERR_STATE) **hash);
 #endif
 
 int ERR_get_next_error_library(void);
@@ -308,12 +370,6 @@ int ERR_get_next_error_library(void);
 int ERR_set_mark(void);
 int ERR_pop_to_mark(void);
 
-#ifdef OPENSSL_FIPS
-void int_ERR_set_state_func(ERR_STATE *(*get_func)(void),
-				void (*remove_func)(unsigned long pid));
-void int_ERR_lib_init(void);
-#endif
-
 /* Already defined in ossl_typ.h */
 /* typedef struct st_ERR_FNS ERR_FNS; */
 /* An application can use this function and provide the return value to loaded
diff --git a/crypto/err/err_all.c b/crypto/err/err_all.c
index 39796f7cc8cd..bd8946d8ba0f 100644
--- a/crypto/err/err_all.c
+++ b/crypto/err/err_all.c
@@ -64,6 +64,7 @@
 #endif
 #include <openssl/buffer.h>
 #include <openssl/bio.h>
+#include <openssl/comp.h>
 #ifndef OPENSSL_NO_RSA
 #include <openssl/rsa.h>
 #endif
@@ -94,10 +95,7 @@
 #include <openssl/ui.h>
 #include <openssl/ocsp.h>
 #include <openssl/err.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/ts.h>
 #ifndef OPENSSL_NO_CMS
 #include <openssl/cms.h>
 #endif
@@ -106,6 +104,10 @@
 #endif
 #include <openssl/comp.h>
 
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
+
 void ERR_load_crypto_strings(void)
 	{
 #ifndef OPENSSL_NO_ERR
@@ -128,6 +130,7 @@ void ERR_load_crypto_strings(void)
 	ERR_load_ASN1_strings();
 	ERR_load_CONF_strings();
 	ERR_load_CRYPTO_strings();
+	ERR_load_COMP_strings();
 #ifndef OPENSSL_NO_EC
 	ERR_load_EC_strings();
 #endif
@@ -144,14 +147,12 @@ void ERR_load_crypto_strings(void)
 	ERR_load_PKCS12_strings();
 	ERR_load_RAND_strings();
 	ERR_load_DSO_strings();
+	ERR_load_TS_strings();
 #ifndef OPENSSL_NO_ENGINE
 	ERR_load_ENGINE_strings();
 #endif
 	ERR_load_OCSP_strings();
 	ERR_load_UI_strings();
-#ifdef OPENSSL_FIPS
-	ERR_load_FIPS_strings();
-#endif
 #ifndef OPENSSL_NO_CMS
 	ERR_load_CMS_strings();
 #endif
@@ -160,4 +161,7 @@ void ERR_load_crypto_strings(void)
 #endif
 	ERR_load_COMP_strings();
 #endif
+#ifdef OPENSSL_FIPS
+	ERR_load_FIPS_strings();
+#endif
 	}
diff --git a/crypto/err/err_def.c b/crypto/err/err_def.c
deleted file mode 100644
index 7ed3d8495538..000000000000
--- a/crypto/err/err_def.c
+++ /dev/null
@@ -1,665 +0,0 @@
-/* crypto/err/err_def.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-/* ====================================================================
- * Copyright (c) 1998-2001 The OpenSSL Project.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    openssl-core@openssl.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
- */
-
-#include <stdio.h>
-#include <stdarg.h>
-#include <string.h>
-#include "cryptlib.h"
-#include <openssl/lhash.h>
-#include <openssl/crypto.h>
-#include <openssl/buffer.h>
-#include <openssl/bio.h>
-#include <openssl/err.h>
-
-#define err_clear_data(p,i) \
-	do { \
-	if (((p)->err_data[i] != NULL) && \
-		(p)->err_data_flags[i] & ERR_TXT_MALLOCED) \
-		{  \
-		OPENSSL_free((p)->err_data[i]); \
-		(p)->err_data[i]=NULL; \
-		} \
-	(p)->err_data_flags[i]=0; \
-	} while(0)
-
-#define err_clear(p,i) \
-	do { \
-	(p)->err_flags[i]=0; \
-	(p)->err_buffer[i]=0; \
-	err_clear_data(p,i); \
-	(p)->err_file[i]=NULL; \
-	(p)->err_line[i]= -1; \
-	} while(0)
-
-static void err_load_strings(int lib, ERR_STRING_DATA *str);
-
-static void ERR_STATE_free(ERR_STATE *s);
-
-/* Define the predeclared (but externally opaque) "ERR_FNS" type */
-struct st_ERR_FNS
-	{
-	/* Works on the "error_hash" string table */
-	LHASH *(*cb_err_get)(int create);
-	void (*cb_err_del)(void);
-	ERR_STRING_DATA *(*cb_err_get_item)(const ERR_STRING_DATA *);
-	ERR_STRING_DATA *(*cb_err_set_item)(ERR_STRING_DATA *);
-	ERR_STRING_DATA *(*cb_err_del_item)(ERR_STRING_DATA *);
-	/* Works on the "thread_hash" error-state table */
-	LHASH *(*cb_thread_get)(int create);
-	void (*cb_thread_release)(LHASH **hash);
-	ERR_STATE *(*cb_thread_get_item)(const ERR_STATE *);
-	ERR_STATE *(*cb_thread_set_item)(ERR_STATE *);
-	void (*cb_thread_del_item)(const ERR_STATE *);
-	/* Returns the next available error "library" numbers */
-	int (*cb_get_next_lib)(void);
-	};
-
-/* Predeclarations of the "err_defaults" functions */
-static LHASH *int_err_get(int create);
-static void int_err_del(void);
-static ERR_STRING_DATA *int_err_get_item(const ERR_STRING_DATA *);
-static ERR_STRING_DATA *int_err_set_item(ERR_STRING_DATA *);
-static ERR_STRING_DATA *int_err_del_item(ERR_STRING_DATA *);
-static LHASH *int_thread_get(int create);
-static void int_thread_release(LHASH **hash);
-static ERR_STATE *int_thread_get_item(const ERR_STATE *);
-static ERR_STATE *int_thread_set_item(ERR_STATE *);
-static void int_thread_del_item(const ERR_STATE *);
-static int int_err_get_next_lib(void);
-/* The static ERR_FNS table using these defaults functions */
-static const ERR_FNS err_defaults =
-	{
-	int_err_get,
-	int_err_del,
-	int_err_get_item,
-	int_err_set_item,
-	int_err_del_item,
-	int_thread_get,
-	int_thread_release,
-	int_thread_get_item,
-	int_thread_set_item,
-	int_thread_del_item,
-	int_err_get_next_lib
-	};
-
-/* The replacable table of ERR_FNS functions we use at run-time */
-static const ERR_FNS *err_fns = NULL;
-
-/* Eg. rather than using "err_get()", use "ERRFN(err_get)()". */
-#define ERRFN(a) err_fns->cb_##a
-
-/* The internal state used by "err_defaults" - as such, the setting, reading,
- * creating, and deleting of this data should only be permitted via the
- * "err_defaults" functions. This way, a linked module can completely defer all
- * ERR state operation (together with requisite locking) to the implementations
- * and state in the loading application. */
-static LHASH *int_error_hash = NULL;
-static LHASH *int_thread_hash = NULL;
-static int int_thread_hash_references = 0;
-static int int_err_library_number= ERR_LIB_USER;
-
-/* Internal function that checks whether "err_fns" is set and if not, sets it to
- * the defaults. */
-static void err_fns_check(void)
-	{
-	if (err_fns) return;
-	
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	if (!err_fns)
-		err_fns = &err_defaults;
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-	}
-
-/* API functions to get or set the underlying ERR functions. */
-
-const ERR_FNS *ERR_get_implementation(void)
-	{
-	err_fns_check();
-	return err_fns;
-	}
-
-int ERR_set_implementation(const ERR_FNS *fns)
-	{
-	int ret = 0;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	/* It's too late if 'err_fns' is non-NULL. BTW: not much point setting
-	 * an error is there?! */
-	if (!err_fns)
-		{
-		err_fns = fns;
-		ret = 1;
-		}
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-	return ret;
-	}
-
-/* These are the callbacks provided to "lh_new()" when creating the LHASH tables
- * internal to the "err_defaults" implementation. */
-
-/* static unsigned long err_hash(ERR_STRING_DATA *a); */
-static unsigned long err_hash(const void *a_void);
-/* static int err_cmp(ERR_STRING_DATA *a, ERR_STRING_DATA *b); */
-static int err_cmp(const void *a_void, const void *b_void);
-/* static unsigned long pid_hash(ERR_STATE *pid); */
-static unsigned long pid_hash(const void *pid_void);
-/* static int pid_cmp(ERR_STATE *a,ERR_STATE *pid); */
-static int pid_cmp(const void *a_void,const void *pid_void);
-
-/* The internal functions used in the "err_defaults" implementation */
-
-static LHASH *int_err_get(int create)
-	{
-	LHASH *ret = NULL;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	if (!int_error_hash && create)
-		{
-		CRYPTO_push_info("int_err_get (err.c)");
-		int_error_hash = lh_new(err_hash, err_cmp);
-		CRYPTO_pop_info();
-		}
-	if (int_error_hash)
-		ret = int_error_hash;
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-
-	return ret;
-	}
-
-static void int_err_del(void)
-	{
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	if (int_error_hash)
-		{
-		lh_free(int_error_hash);
-		int_error_hash = NULL;
-		}
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-	}
-
-static ERR_STRING_DATA *int_err_get_item(const ERR_STRING_DATA *d)
-	{
-	ERR_STRING_DATA *p;
-	LHASH *hash;
-
-	err_fns_check();
-	hash = ERRFN(err_get)(0);
-	if (!hash)
-		return NULL;
-
-	CRYPTO_r_lock(CRYPTO_LOCK_ERR);
-	p = (ERR_STRING_DATA *)lh_retrieve(hash, d);
-	CRYPTO_r_unlock(CRYPTO_LOCK_ERR);
-
-	return p;
-	}
-
-static ERR_STRING_DATA *int_err_set_item(ERR_STRING_DATA *d)
-	{
-	ERR_STRING_DATA *p;
-	LHASH *hash;
-
-	err_fns_check();
-	hash = ERRFN(err_get)(1);
-	if (!hash)
-		return NULL;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	p = (ERR_STRING_DATA *)lh_insert(hash, d);
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-
-	return p;
-	}
-
-static ERR_STRING_DATA *int_err_del_item(ERR_STRING_DATA *d)
-	{
-	ERR_STRING_DATA *p;
-	LHASH *hash;
-
-	err_fns_check();
-	hash = ERRFN(err_get)(0);
-	if (!hash)
-		return NULL;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	p = (ERR_STRING_DATA *)lh_delete(hash, d);
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-
-	return p;
-	}
-
-static LHASH *int_thread_get(int create)
-	{
-	LHASH *ret = NULL;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	if (!int_thread_hash && create)
-		{
-		CRYPTO_push_info("int_thread_get (err.c)");
-		int_thread_hash = lh_new(pid_hash, pid_cmp);
-		CRYPTO_pop_info();
-		}
-	if (int_thread_hash)
-		{
-		int_thread_hash_references++;
-		ret = int_thread_hash;
-		}
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-	return ret;
-	}
-
-static void int_thread_release(LHASH **hash)
-	{
-	int i;
-
-	if (hash == NULL || *hash == NULL)
-		return;
-
-	i = CRYPTO_add(&int_thread_hash_references, -1, CRYPTO_LOCK_ERR);
-
-#ifdef REF_PRINT
-	fprintf(stderr,"%4d:%s\n",int_thread_hash_references,"ERR");
-#endif
-	if (i > 0) return;
-#ifdef REF_CHECK
-	if (i < 0)
-		{
-		fprintf(stderr,"int_thread_release, bad reference count\n");
-		abort(); /* ok */
-		}
-#endif
-	*hash = NULL;
-	}
-
-static ERR_STATE *int_thread_get_item(const ERR_STATE *d)
-	{
-	ERR_STATE *p;
-	LHASH *hash;
-
-	err_fns_check();
-	hash = ERRFN(thread_get)(0);
-	if (!hash)
-		return NULL;
-
-	CRYPTO_r_lock(CRYPTO_LOCK_ERR);
-	p = (ERR_STATE *)lh_retrieve(hash, d);
-	CRYPTO_r_unlock(CRYPTO_LOCK_ERR);
-
-	ERRFN(thread_release)(&hash);
-	return p;
-	}
-
-static ERR_STATE *int_thread_set_item(ERR_STATE *d)
-	{
-	ERR_STATE *p;
-	LHASH *hash;
-
-	err_fns_check();
-	hash = ERRFN(thread_get)(1);
-	if (!hash)
-		return NULL;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	p = (ERR_STATE *)lh_insert(hash, d);
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-
-	ERRFN(thread_release)(&hash);
-	return p;
-	}
-
-static void int_thread_del_item(const ERR_STATE *d)
-	{
-	ERR_STATE *p;
-	LHASH *hash;
-
-	err_fns_check();
-	hash = ERRFN(thread_get)(0);
-	if (!hash)
-		return;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	p = (ERR_STATE *)lh_delete(hash, d);
-	/* make sure we don't leak memory */
-	if (int_thread_hash_references == 1
-		&& int_thread_hash && (lh_num_items(int_thread_hash) == 0))
-		{
-		lh_free(int_thread_hash);
-		int_thread_hash = NULL;
-		}
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-
-	ERRFN(thread_release)(&hash);
-	if (p)
-		ERR_STATE_free(p);
-	}
-
-static int int_err_get_next_lib(void)
-	{
-	int ret;
-
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	ret = int_err_library_number++;
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-
-	return ret;
-	}
-
-static void ERR_STATE_free(ERR_STATE *s)
-	{
-	int i;
-
-	if (s == NULL)
-	    return;
-
-	for (i=0; i<ERR_NUM_ERRORS; i++)
-		{
-		err_clear_data(s,i);
-		}
-	OPENSSL_free(s);
-	}
-
-static void err_load_strings(int lib, ERR_STRING_DATA *str)
-	{
-	while (str->error)
-		{
-		if (lib)
-			str->error|=ERR_PACK(lib,0,0);
-		ERRFN(err_set_item)(str);
-		str++;
-		}
-	}
-
-void ERR_load_strings(int lib, ERR_STRING_DATA *str)
-	{
-	err_fns_check();
-	err_load_strings(lib, str);
-	}
-
-void ERR_unload_strings(int lib, ERR_STRING_DATA *str)
-	{
-	while (str->error)
-		{
-		if (lib)
-			str->error|=ERR_PACK(lib,0,0);
-		ERRFN(err_del_item)(str);
-		str++;
-		}
-	}
-
-void ERR_free_strings(void)
-	{
-	err_fns_check();
-	ERRFN(err_del)();
-	}
-
-LHASH *ERR_get_string_table(void)
-	{
-	err_fns_check();
-	return ERRFN(err_get)(0);
-	}
-
-LHASH *ERR_get_err_state_table(void)
-	{
-	err_fns_check();
-	return ERRFN(thread_get)(0);
-	}
-
-void ERR_release_err_state_table(LHASH **hash)
-	{
-	err_fns_check();
-	ERRFN(thread_release)(hash);
-	}
-
-const char *ERR_lib_error_string(unsigned long e)
-	{
-	ERR_STRING_DATA d,*p;
-	unsigned long l;
-
-	err_fns_check();
-	l=ERR_GET_LIB(e);
-	d.error=ERR_PACK(l,0,0);
-	p=ERRFN(err_get_item)(&d);
-	return((p == NULL)?NULL:p->string);
-	}
-
-const char *ERR_func_error_string(unsigned long e)
-	{
-	ERR_STRING_DATA d,*p;
-	unsigned long l,f;
-
-	err_fns_check();
-	l=ERR_GET_LIB(e);
-	f=ERR_GET_FUNC(e);
-	d.error=ERR_PACK(l,f,0);
-	p=ERRFN(err_get_item)(&d);
-	return((p == NULL)?NULL:p->string);
-	}
-
-const char *ERR_reason_error_string(unsigned long e)
-	{
-	ERR_STRING_DATA d,*p=NULL;
-	unsigned long l,r;
-
-	err_fns_check();
-	l=ERR_GET_LIB(e);
-	r=ERR_GET_REASON(e);
-	d.error=ERR_PACK(l,0,r);
-	p=ERRFN(err_get_item)(&d);
-	if (!p)
-		{
-		d.error=ERR_PACK(0,0,r);
-		p=ERRFN(err_get_item)(&d);
-		}
-	return((p == NULL)?NULL:p->string);
-	}
-
-/* static unsigned long err_hash(ERR_STRING_DATA *a) */
-static unsigned long err_hash(const void *a_void)
-	{
-	unsigned long ret,l;
-
-	l=((const ERR_STRING_DATA *)a_void)->error;
-	ret=l^ERR_GET_LIB(l)^ERR_GET_FUNC(l);
-	return(ret^ret%19*13);
-	}
-
-/* static int err_cmp(ERR_STRING_DATA *a, ERR_STRING_DATA *b) */
-static int err_cmp(const void *a_void, const void *b_void)
-	{
-	return((int)(((const ERR_STRING_DATA *)a_void)->error -
-			((const ERR_STRING_DATA *)b_void)->error));
-	}
-
-/* static unsigned long pid_hash(ERR_STATE *a) */
-static unsigned long pid_hash(const void *a_void)
-	{
-	return(((const ERR_STATE *)a_void)->pid*13);
-	}
-
-/* static int pid_cmp(ERR_STATE *a, ERR_STATE *b) */
-static int pid_cmp(const void *a_void, const void *b_void)
-	{
-	return((int)((long)((const ERR_STATE *)a_void)->pid -
-			(long)((const ERR_STATE *)b_void)->pid));
-	}
-#ifdef OPENSSL_FIPS
-static void int_err_remove_state(unsigned long pid)
-#else
-void ERR_remove_state(unsigned long pid)
-#endif
-	{
-	ERR_STATE tmp;
-
-	err_fns_check();
-	if (pid == 0)
-		pid=(unsigned long)CRYPTO_thread_id();
-	tmp.pid=pid;
-	/* thread_del_item automatically destroys the LHASH if the number of
-	 * items reaches zero. */
-	ERRFN(thread_del_item)(&tmp);
-	}
-
-#ifdef OPENSSL_FIPS
-	static ERR_STATE *int_err_get_state(void)
-#else
-ERR_STATE *ERR_get_state(void)
-#endif
-	{
-	static ERR_STATE fallback;
-	ERR_STATE *ret,tmp,*tmpp=NULL;
-	int i;
-	unsigned long pid;
-
-	err_fns_check();
-	pid=(unsigned long)CRYPTO_thread_id();
-	tmp.pid=pid;
-	ret=ERRFN(thread_get_item)(&tmp);
-
-	/* ret == the error state, if NULL, make a new one */
-	if (ret == NULL)
-		{
-		ret=(ERR_STATE *)OPENSSL_malloc(sizeof(ERR_STATE));
-		if (ret == NULL) return(&fallback);
-		ret->pid=pid;
-		ret->top=0;
-		ret->bottom=0;
-		for (i=0; i<ERR_NUM_ERRORS; i++)
-			{
-			ret->err_data[i]=NULL;
-			ret->err_data_flags[i]=0;
-			}
-		tmpp = ERRFN(thread_set_item)(ret);
-		/* To check if insertion failed, do a get. */
-		if (ERRFN(thread_get_item)(ret) != ret)
-			{
-			ERR_STATE_free(ret); /* could not insert it */
-			return(&fallback);
-			}
-		/* If a race occured in this function and we came second, tmpp
-		 * is the first one that we just replaced. */
-		if (tmpp)
-			ERR_STATE_free(tmpp);
-		}
-	return ret;
-	}
-
-#ifdef OPENSSL_FIPS
-void int_ERR_lib_init(void)
-	{
-	int_ERR_set_state_func(int_err_get_state, int_err_remove_state);
-	}
-#endif
-
-int ERR_get_next_error_library(void)
-	{
-	err_fns_check();
-	return ERRFN(get_next_lib)();
-	}
diff --git a/crypto/err/err_prn.c b/crypto/err/err_prn.c
index 1e46f93ecf81..a0168ac8ed62 100644
--- a/crypto/err/err_prn.c
+++ b/crypto/err/err_prn.c
@@ -72,8 +72,10 @@ void ERR_print_errors_cb(int (*cb)(const char *str, size_t len, void *u),
 	const char *file,*data;
 	int line,flags;
 	unsigned long es;
+	CRYPTO_THREADID cur;
 
-	es=CRYPTO_thread_id();
+	CRYPTO_THREADID_current(&cur);
+	es=CRYPTO_THREADID_hash(&cur);
 	while ((l=ERR_get_error_line_data(&file,&line,&data,&flags)) != 0)
 		{
 		ERR_error_string_n(l, buf, sizeof buf);
@@ -100,64 +102,13 @@ void ERR_print_errors_fp(FILE *fp)
 	}
 #endif
 
-void ERR_error_string_n(unsigned long e, char *buf, size_t len)
+static int print_bio(const char *str, size_t len, void *bp)
 	{
-	char lsbuf[64], fsbuf[64], rsbuf[64];
-	const char *ls,*fs,*rs;
-	unsigned long l,f,r;
-
-	l=ERR_GET_LIB(e);
-	f=ERR_GET_FUNC(e);
-	r=ERR_GET_REASON(e);
-
-	ls=ERR_lib_error_string(e);
-	fs=ERR_func_error_string(e);
-	rs=ERR_reason_error_string(e);
-
-	if (ls == NULL) 
-		BIO_snprintf(lsbuf, sizeof(lsbuf), "lib(%lu)", l);
-	if (fs == NULL)
-		BIO_snprintf(fsbuf, sizeof(fsbuf), "func(%lu)", f);
-	if (rs == NULL)
-		BIO_snprintf(rsbuf, sizeof(rsbuf), "reason(%lu)", r);
-
-	BIO_snprintf(buf, len,"error:%08lX:%s:%s:%s", e, ls?ls:lsbuf, 
-		fs?fs:fsbuf, rs?rs:rsbuf);
-	if (strlen(buf) == len-1)
-		{
-		/* output may be truncated; make sure we always have 5 
-		 * colon-separated fields, i.e. 4 colons ... */
-#define NUM_COLONS 4
-		if (len > NUM_COLONS) /* ... if possible */
-			{
-			int i;
-			char *s = buf;
-			
-			for (i = 0; i < NUM_COLONS; i++)
-				{
-				char *colon = strchr(s, ':');
-				if (colon == NULL || colon > &buf[len-1] - NUM_COLONS + i)
-					{
-					/* set colon no. i at last possible position
-					 * (buf[len-1] is the terminating 0)*/
-					colon = &buf[len-1] - NUM_COLONS + i;
-					*colon = ':';
-					}
-				s = colon + 1;
-				}
-			}
-		}
+	return BIO_write((BIO *)bp, str, len);
 	}
-
-/* BAD for multi-threading: uses a local buffer if ret == NULL */
-/* ERR_error_string_n should be used instead for ret != NULL
- * as ERR_error_string cannot know how large the buffer is */
-char *ERR_error_string(unsigned long e, char *ret)
+void ERR_print_errors(BIO *bp)
 	{
-	static char buf[256];
-
-	if (ret == NULL) ret=buf;
-	ERR_error_string_n(e, ret, 256);
-
-	return ret;
+	ERR_print_errors_cb(print_bio, bp);
 	}
+
+	
diff --git a/crypto/err/err_str.c b/crypto/err/err_str.c
deleted file mode 100644
index d39040888d07..000000000000
--- a/crypto/err/err_str.c
+++ /dev/null
@@ -1,295 +0,0 @@
-/* crypto/err/err_str.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-/* ====================================================================
- * Copyright (c) 1998-2001 The OpenSSL Project.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    openssl-core@openssl.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
- */
-
-#include <stdio.h>
-#include <stdarg.h>
-#include <string.h>
-#include "cryptlib.h"
-#include <openssl/lhash.h>
-#include <openssl/crypto.h>
-#include <openssl/buffer.h>
-#include <openssl/bio.h>
-#include <openssl/err.h>
-
-#ifndef OPENSSL_NO_ERR
-static ERR_STRING_DATA ERR_str_libraries[]=
-	{
-{ERR_PACK(ERR_LIB_NONE,0,0)		,"unknown library"},
-{ERR_PACK(ERR_LIB_SYS,0,0)		,"system library"},
-{ERR_PACK(ERR_LIB_BN,0,0)		,"bignum routines"},
-{ERR_PACK(ERR_LIB_RSA,0,0)		,"rsa routines"},
-{ERR_PACK(ERR_LIB_DH,0,0)		,"Diffie-Hellman routines"},
-{ERR_PACK(ERR_LIB_EVP,0,0)		,"digital envelope routines"},
-{ERR_PACK(ERR_LIB_BUF,0,0)		,"memory buffer routines"},
-{ERR_PACK(ERR_LIB_OBJ,0,0)		,"object identifier routines"},
-{ERR_PACK(ERR_LIB_PEM,0,0)		,"PEM routines"},
-{ERR_PACK(ERR_LIB_DSA,0,0)		,"dsa routines"},
-{ERR_PACK(ERR_LIB_X509,0,0)		,"x509 certificate routines"},
-{ERR_PACK(ERR_LIB_ASN1,0,0)		,"asn1 encoding routines"},
-{ERR_PACK(ERR_LIB_CONF,0,0)		,"configuration file routines"},
-{ERR_PACK(ERR_LIB_CRYPTO,0,0)		,"common libcrypto routines"},
-{ERR_PACK(ERR_LIB_EC,0,0)		,"elliptic curve routines"},
-{ERR_PACK(ERR_LIB_SSL,0,0)		,"SSL routines"},
-{ERR_PACK(ERR_LIB_BIO,0,0)		,"BIO routines"},
-{ERR_PACK(ERR_LIB_PKCS7,0,0)		,"PKCS7 routines"},
-{ERR_PACK(ERR_LIB_X509V3,0,0)		,"X509 V3 routines"},
-{ERR_PACK(ERR_LIB_PKCS12,0,0)		,"PKCS12 routines"},
-{ERR_PACK(ERR_LIB_RAND,0,0)		,"random number generator"},
-{ERR_PACK(ERR_LIB_DSO,0,0)		,"DSO support routines"},
-{ERR_PACK(ERR_LIB_ENGINE,0,0)		,"engine routines"},
-{ERR_PACK(ERR_LIB_OCSP,0,0)		,"OCSP routines"},
-{ERR_PACK(ERR_LIB_FIPS,0,0)		,"FIPS routines"},
-{ERR_PACK(ERR_LIB_CMS,0,0)		,"CMS routines"},
-{ERR_PACK(ERR_LIB_JPAKE,0,0)		,"JPAKE routines"},
-{0,NULL},
-	};
-
-static ERR_STRING_DATA ERR_str_functs[]=
-	{
-	{ERR_PACK(0,SYS_F_FOPEN,0),     	"fopen"},
-	{ERR_PACK(0,SYS_F_CONNECT,0),		"connect"},
-	{ERR_PACK(0,SYS_F_GETSERVBYNAME,0),	"getservbyname"},
-	{ERR_PACK(0,SYS_F_SOCKET,0),		"socket"}, 
-	{ERR_PACK(0,SYS_F_IOCTLSOCKET,0),	"ioctlsocket"},
-	{ERR_PACK(0,SYS_F_BIND,0),		"bind"},
-	{ERR_PACK(0,SYS_F_LISTEN,0),		"listen"},
-	{ERR_PACK(0,SYS_F_ACCEPT,0),		"accept"},
-#ifdef OPENSSL_SYS_WINDOWS
-	{ERR_PACK(0,SYS_F_WSASTARTUP,0),	"WSAstartup"},
-#endif
-	{ERR_PACK(0,SYS_F_OPENDIR,0),		"opendir"},
-	{ERR_PACK(0,SYS_F_FREAD,0),		"fread"},
-	{0,NULL},
-	};
-
-static ERR_STRING_DATA ERR_str_reasons[]=
-	{
-{ERR_R_SYS_LIB				,"system lib"},
-{ERR_R_BN_LIB				,"BN lib"},
-{ERR_R_RSA_LIB				,"RSA lib"},
-{ERR_R_DH_LIB				,"DH lib"},
-{ERR_R_EVP_LIB				,"EVP lib"},
-{ERR_R_BUF_LIB				,"BUF lib"},
-{ERR_R_OBJ_LIB				,"OBJ lib"},
-{ERR_R_PEM_LIB				,"PEM lib"},
-{ERR_R_DSA_LIB				,"DSA lib"},
-{ERR_R_X509_LIB				,"X509 lib"},
-{ERR_R_ASN1_LIB				,"ASN1 lib"},
-{ERR_R_CONF_LIB				,"CONF lib"},
-{ERR_R_CRYPTO_LIB			,"CRYPTO lib"},
-{ERR_R_EC_LIB				,"EC lib"},
-{ERR_R_SSL_LIB				,"SSL lib"},
-{ERR_R_BIO_LIB				,"BIO lib"},
-{ERR_R_PKCS7_LIB			,"PKCS7 lib"},
-{ERR_R_X509V3_LIB			,"X509V3 lib"},
-{ERR_R_PKCS12_LIB			,"PKCS12 lib"},
-{ERR_R_RAND_LIB				,"RAND lib"},
-{ERR_R_DSO_LIB				,"DSO lib"},
-{ERR_R_ENGINE_LIB			,"ENGINE lib"},
-{ERR_R_OCSP_LIB				,"OCSP lib"},
-
-{ERR_R_NESTED_ASN1_ERROR		,"nested asn1 error"},
-{ERR_R_BAD_ASN1_OBJECT_HEADER		,"bad asn1 object header"},
-{ERR_R_BAD_GET_ASN1_OBJECT_CALL		,"bad get asn1 object call"},
-{ERR_R_EXPECTING_AN_ASN1_SEQUENCE	,"expecting an asn1 sequence"},
-{ERR_R_ASN1_LENGTH_MISMATCH		,"asn1 length mismatch"},
-{ERR_R_MISSING_ASN1_EOS			,"missing asn1 eos"},
-
-{ERR_R_FATAL                            ,"fatal"},
-{ERR_R_MALLOC_FAILURE			,"malloc failure"},
-{ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED	,"called a function you should not call"},
-{ERR_R_PASSED_NULL_PARAMETER		,"passed a null parameter"},
-{ERR_R_INTERNAL_ERROR			,"internal error"},
-{ERR_R_DISABLED				,"called a function that was disabled at compile-time"},
-
-{0,NULL},
-	};
-#endif
-
-#ifndef OPENSSL_NO_ERR
-#define NUM_SYS_STR_REASONS 127
-#define LEN_SYS_STR_REASON 32
-
-static ERR_STRING_DATA SYS_str_reasons[NUM_SYS_STR_REASONS + 1];
-/* SYS_str_reasons is filled with copies of strerror() results at
- * initialization.
- * 'errno' values up to 127 should cover all usual errors,
- * others will be displayed numerically by ERR_error_string.
- * It is crucial that we have something for each reason code
- * that occurs in ERR_str_reasons, or bogus reason strings
- * will be returned for SYSerr, which always gets an errno
- * value and never one of those 'standard' reason codes. */
-
-static void build_SYS_str_reasons(void)
-	{
-	/* OPENSSL_malloc cannot be used here, use static storage instead */
-	static char strerror_tab[NUM_SYS_STR_REASONS][LEN_SYS_STR_REASON];
-	int i;
-	static int init = 1;
-
-	CRYPTO_r_lock(CRYPTO_LOCK_ERR);
-	if (!init)
-		{
-		CRYPTO_r_unlock(CRYPTO_LOCK_ERR);
-		return;
-		}
-	
-	CRYPTO_r_unlock(CRYPTO_LOCK_ERR);
-	CRYPTO_w_lock(CRYPTO_LOCK_ERR);
-	if (!init)
-		{
-		CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-		return;
-		}
-
-	for (i = 1; i <= NUM_SYS_STR_REASONS; i++)
-		{
-		ERR_STRING_DATA *str = &SYS_str_reasons[i - 1];
-
-		str->error = (unsigned long)i;
-		if (str->string == NULL)
-			{
-			char (*dest)[LEN_SYS_STR_REASON] = &(strerror_tab[i - 1]);
-			char *src = strerror(i);
-			if (src != NULL)
-				{
-				strncpy(*dest, src, sizeof *dest);
-				(*dest)[sizeof *dest - 1] = '\0';
-				str->string = *dest;
-				}
-			}
-		if (str->string == NULL)
-			str->string = "unknown";
-		}
-
-	/* Now we still have SYS_str_reasons[NUM_SYS_STR_REASONS] = {0, NULL},
-	 * as required by ERR_load_strings. */
-
-	init = 0;
-	
-	CRYPTO_w_unlock(CRYPTO_LOCK_ERR);
-	}
-#endif
-
-void ERR_load_ERR_strings(void)
-	{
-#ifndef OPENSSL_NO_ERR
-	if (ERR_func_error_string(ERR_str_functs[0].error) == NULL)
-		{
-		ERR_load_strings(0,ERR_str_libraries);
-		ERR_load_strings(0,ERR_str_reasons);
-		ERR_load_strings(ERR_LIB_SYS,ERR_str_functs);
-		build_SYS_str_reasons();
-		ERR_load_strings(ERR_LIB_SYS,SYS_str_reasons);
-		}
-#endif
-	}
-
diff --git a/crypto/err/openssl.ec b/crypto/err/openssl.ec
deleted file mode 100644
index 868826624dbc..000000000000
--- a/crypto/err/openssl.ec
+++ /dev/null
@@ -1,89 +0,0 @@
-# crypto/err/openssl.ec
-
-# configuration file for util/mkerr.pl
-
-# files that may have to be rewritten by util/mkerr.pl
-L ERR		NONE				NONE
-L BN		crypto/bn/bn.h			crypto/bn/bn_err.c
-L RSA		crypto/rsa/rsa.h		crypto/rsa/rsa_err.c
-L DH		crypto/dh/dh.h			crypto/dh/dh_err.c
-L EVP		crypto/evp/evp.h		crypto/evp/evp_err.c
-L BUF		crypto/buffer/buffer.h		crypto/buffer/buf_err.c
-L OBJ		crypto/objects/objects.h	crypto/objects/obj_err.c
-L PEM		crypto/pem/pem.h		crypto/pem/pem_err.c
-L DSA		crypto/dsa/dsa.h		crypto/dsa/dsa_err.c
-L X509		crypto/x509/x509.h		crypto/x509/x509_err.c
-L ASN1		crypto/asn1/asn1.h		crypto/asn1/asn1_err.c
-L CONF		crypto/conf/conf.h		crypto/conf/conf_err.c
-L CRYPTO	crypto/crypto.h			crypto/cpt_err.c
-L EC		crypto/ec/ec.h			crypto/ec/ec_err.c
-L SSL		ssl/ssl.h			ssl/ssl_err.c
-L BIO		crypto/bio/bio.h		crypto/bio/bio_err.c
-L PKCS7		crypto/pkcs7/pkcs7.h		crypto/pkcs7/pkcs7err.c
-L X509V3	crypto/x509v3/x509v3.h		crypto/x509v3/v3err.c
-L PKCS12	crypto/pkcs12/pkcs12.h		crypto/pkcs12/pk12err.c
-L RAND		crypto/rand/rand.h		crypto/rand/rand_err.c
-L DSO		crypto/dso/dso.h		crypto/dso/dso_err.c
-L ENGINE	crypto/engine/engine.h		crypto/engine/eng_err.c
-L OCSP		crypto/ocsp/ocsp.h		crypto/ocsp/ocsp_err.c
-L UI		crypto/ui/ui.h			crypto/ui/ui_err.c
-L COMP		crypto/comp/comp.h		crypto/comp/comp_err.c
-L ECDSA		crypto/ecdsa/ecdsa.h		crypto/ecdsa/ecs_err.c
-L ECDH		crypto/ecdh/ecdh.h		crypto/ecdh/ech_err.c
-L STORE		crypto/store/store.h		crypto/store/str_err.c
-L FIPS		fips/fips.h			crypto/fips_err.h
-L CMS		crypto/cms/cms.h		crypto/cms/cms_err.c
-L JPAKE		crypto/jpake/jpake.h		crypto/jpake/jpake_err.c
-
-# additional header files to be scanned for function names
-L NONE		crypto/x509/x509_vfy.h		NONE
-L NONE		crypto/ec/ec_lcl.h		NONE
-L NONE		crypto/cms/cms_lcl.h		NONE
-
-
-F RSAREF_F_RSA_BN2BIN
-F RSAREF_F_RSA_PRIVATE_DECRYPT
-F RSAREF_F_RSA_PRIVATE_ENCRYPT
-F RSAREF_F_RSA_PUBLIC_DECRYPT
-F RSAREF_F_RSA_PUBLIC_ENCRYPT
-#F SSL_F_CLIENT_CERTIFICATE
-
-R SSL_R_SSLV3_ALERT_UNEXPECTED_MESSAGE		1010
-R SSL_R_SSLV3_ALERT_BAD_RECORD_MAC		1020
-R SSL_R_TLSV1_ALERT_DECRYPTION_FAILED		1021
-R SSL_R_TLSV1_ALERT_RECORD_OVERFLOW		1022
-R SSL_R_SSLV3_ALERT_DECOMPRESSION_FAILURE	1030
-R SSL_R_SSLV3_ALERT_HANDSHAKE_FAILURE		1040
-R SSL_R_SSLV3_ALERT_NO_CERTIFICATE		1041
-R SSL_R_SSLV3_ALERT_BAD_CERTIFICATE		1042
-R SSL_R_SSLV3_ALERT_UNSUPPORTED_CERTIFICATE	1043
-R SSL_R_SSLV3_ALERT_CERTIFICATE_REVOKED		1044
-R SSL_R_SSLV3_ALERT_CERTIFICATE_EXPIRED		1045
-R SSL_R_SSLV3_ALERT_CERTIFICATE_UNKNOWN		1046
-R SSL_R_SSLV3_ALERT_ILLEGAL_PARAMETER		1047
-R SSL_R_TLSV1_ALERT_UNKNOWN_CA			1048
-R SSL_R_TLSV1_ALERT_ACCESS_DENIED		1049
-R SSL_R_TLSV1_ALERT_DECODE_ERROR		1050
-R SSL_R_TLSV1_ALERT_DECRYPT_ERROR		1051
-R SSL_R_TLSV1_ALERT_EXPORT_RESTRICTION		1060
-R SSL_R_TLSV1_ALERT_PROTOCOL_VERSION		1070
-R SSL_R_TLSV1_ALERT_INSUFFICIENT_SECURITY	1071
-R SSL_R_TLSV1_ALERT_INTERNAL_ERROR		1080
-R SSL_R_TLSV1_ALERT_USER_CANCELLED		1090
-R SSL_R_TLSV1_ALERT_NO_RENEGOTIATION		1100
-
-R RSAREF_R_CONTENT_ENCODING			0x0400
-R RSAREF_R_DATA					0x0401
-R RSAREF_R_DIGEST_ALGORITHM			0x0402
-R RSAREF_R_ENCODING				0x0403
-R RSAREF_R_KEY					0x0404
-R RSAREF_R_KEY_ENCODING				0x0405
-R RSAREF_R_LEN					0x0406
-R RSAREF_R_MODULUS_LEN				0x0407
-R RSAREF_R_NEED_RANDOM				0x0408
-R RSAREF_R_PRIVATE_KEY				0x0409
-R RSAREF_R_PUBLIC_KEY				0x040a
-R RSAREF_R_SIGNATURE				0x040b
-R RSAREF_R_SIGNATURE_ENCODING			0x040c
-R RSAREF_R_ENCRYPTION_ALGORITHM			0x040d
-
diff --git a/crypto/evp/Makefile b/crypto/evp/Makefile
index c204f84c1d66..0fe1b96bff45 100644
--- a/crypto/evp/Makefile
+++ b/crypto/evp/Makefile
@@ -18,34 +18,36 @@ TESTDATA=evptests.txt
 APPS=
 
 LIB=$(TOP)/libcrypto.a
-LIBSRC= encode.c digest.c dig_eng.c evp_enc.c evp_key.c evp_acnf.c evp_cnf.c \
+LIBSRC= encode.c digest.c evp_enc.c evp_key.c evp_acnf.c \
 	e_des.c e_bf.c e_idea.c e_des3.c e_camellia.c\
 	e_rc4.c e_aes.c names.c e_seed.c \
-	e_xcbc_d.c e_rc2.c e_cast.c e_rc5.c enc_min.c \
-	m_null.c m_md2.c m_md4.c m_md5.c m_sha.c m_sha1.c \
+	e_xcbc_d.c e_rc2.c e_cast.c e_rc5.c \
+	m_null.c m_md2.c m_md4.c m_md5.c m_sha.c m_sha1.c m_wp.c \
 	m_dss.c m_dss1.c m_mdc2.c m_ripemd.c m_ecdsa.c\
 	p_open.c p_seal.c p_sign.c p_verify.c p_lib.c p_enc.c p_dec.c \
 	bio_md.c bio_b64.c bio_enc.c evp_err.c e_null.c \
 	c_all.c c_allc.c c_alld.c evp_lib.c bio_ok.c \
 	evp_pkey.c evp_pbe.c p5_crpt.c p5_crpt2.c \
-	e_old.c
+	e_old.c pmeth_lib.c pmeth_fn.c pmeth_gn.c m_sigver.c evp_fips.c	\
+	e_aes_cbc_hmac_sha1.c e_rc4_hmac_md5.c
 
-LIBOBJ=	encode.o digest.o dig_eng.o evp_enc.o evp_key.o evp_acnf.o evp_cnf.o \
+LIBOBJ=	encode.o digest.o evp_enc.o evp_key.o evp_acnf.o \
 	e_des.o e_bf.o e_idea.o e_des3.o e_camellia.o\
 	e_rc4.o e_aes.o names.o e_seed.o \
-	e_xcbc_d.o e_rc2.o e_cast.o e_rc5.o enc_min.o \
-	m_null.o m_md2.o m_md4.o m_md5.o m_sha.o m_sha1.o \
+	e_xcbc_d.o e_rc2.o e_cast.o e_rc5.o \
+	m_null.o m_md2.o m_md4.o m_md5.o m_sha.o m_sha1.o m_wp.o \
 	m_dss.o m_dss1.o m_mdc2.o m_ripemd.o m_ecdsa.o\
 	p_open.o p_seal.o p_sign.o p_verify.o p_lib.o p_enc.o p_dec.o \
 	bio_md.o bio_b64.o bio_enc.o evp_err.o e_null.o \
 	c_all.o c_allc.o c_alld.o evp_lib.o bio_ok.o \
 	evp_pkey.o evp_pbe.o p5_crpt.o p5_crpt2.o \
-	e_old.o
+	e_old.o pmeth_lib.o pmeth_fn.o pmeth_gn.o m_sigver.o evp_fips.o \
+	e_aes_cbc_hmac_sha1.o e_rc4_hmac_md5.o
 
 SRC= $(LIBSRC)
 
 EXHEADER= evp.h
-HEADER=	$(EXHEADER)
+HEADER=	evp_locl.h $(EXHEADER)
 
 ALL=    $(GENERAL) $(SRC) $(HEADER)
 
@@ -55,7 +57,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -101,201 +103,212 @@ bio_b64.o: ../../e_os.h ../../include/openssl/asn1.h
 bio_b64.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 bio_b64.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 bio_b64.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-bio_b64.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-bio_b64.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-bio_b64.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-bio_b64.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-bio_b64.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-bio_b64.o: ../cryptlib.h bio_b64.c
+bio_b64.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+bio_b64.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+bio_b64.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+bio_b64.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+bio_b64.o: ../../include/openssl/symhacks.h ../cryptlib.h bio_b64.c
 bio_enc.o: ../../e_os.h ../../include/openssl/asn1.h
 bio_enc.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 bio_enc.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 bio_enc.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-bio_enc.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-bio_enc.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-bio_enc.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-bio_enc.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-bio_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-bio_enc.o: ../cryptlib.h bio_enc.c
+bio_enc.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+bio_enc.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+bio_enc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+bio_enc.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+bio_enc.o: ../../include/openssl/symhacks.h ../cryptlib.h bio_enc.c
 bio_md.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 bio_md.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 bio_md.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-bio_md.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-bio_md.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-bio_md.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-bio_md.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-bio_md.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-bio_md.o: ../../include/openssl/symhacks.h ../cryptlib.h bio_md.c
+bio_md.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+bio_md.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+bio_md.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+bio_md.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+bio_md.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+bio_md.o: ../cryptlib.h bio_md.c
 bio_ok.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 bio_ok.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 bio_ok.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-bio_ok.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-bio_ok.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-bio_ok.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-bio_ok.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-bio_ok.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-bio_ok.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-bio_ok.o: ../cryptlib.h bio_ok.c
+bio_ok.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+bio_ok.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+bio_ok.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+bio_ok.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
+bio_ok.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+bio_ok.o: ../../include/openssl/symhacks.h ../cryptlib.h bio_ok.c
 c_all.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 c_all.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 c_all.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 c_all.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 c_all.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-c_all.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-c_all.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-c_all.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-c_all.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-c_all.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-c_all.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-c_all.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-c_all.o: ../../include/openssl/x509_vfy.h ../cryptlib.h c_all.c
+c_all.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+c_all.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+c_all.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+c_all.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+c_all.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+c_all.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+c_all.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+c_all.o: ../cryptlib.h c_all.c
 c_allc.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 c_allc.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 c_allc.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 c_allc.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 c_allc.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-c_allc.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-c_allc.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-c_allc.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-c_allc.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
-c_allc.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-c_allc.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-c_allc.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-c_allc.o: ../../include/openssl/x509_vfy.h ../cryptlib.h c_allc.c
+c_allc.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+c_allc.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+c_allc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+c_allc.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
+c_allc.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+c_allc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+c_allc.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+c_allc.o: ../cryptlib.h c_allc.c
 c_alld.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 c_alld.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 c_alld.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 c_alld.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 c_alld.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-c_alld.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-c_alld.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-c_alld.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-c_alld.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
-c_alld.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-c_alld.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-c_alld.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-c_alld.o: ../../include/openssl/x509_vfy.h ../cryptlib.h c_alld.c
-dig_eng.o: ../../e_os.h ../../include/openssl/asn1.h
-dig_eng.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
-dig_eng.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
-dig_eng.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
-dig_eng.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
-dig_eng.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-dig_eng.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-dig_eng.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-dig_eng.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-dig_eng.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-dig_eng.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-dig_eng.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-dig_eng.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-dig_eng.o: ../cryptlib.h dig_eng.c evp_locl.h
+c_alld.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+c_alld.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+c_alld.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+c_alld.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
+c_alld.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+c_alld.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+c_alld.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+c_alld.o: ../cryptlib.h c_alld.c
 digest.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 digest.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 digest.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 digest.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 digest.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-digest.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-digest.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-digest.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-digest.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-digest.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-digest.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-digest.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-digest.o: ../../include/openssl/x509_vfy.h ../cryptlib.h digest.c evp_locl.h
+digest.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+digest.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+digest.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+digest.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+digest.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+digest.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+digest.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+digest.o: ../cryptlib.h digest.c
 e_aes.o: ../../include/openssl/aes.h ../../include/openssl/asn1.h
 e_aes.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
 e_aes.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-e_aes.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-e_aes.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+e_aes.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+e_aes.o: ../../include/openssl/modes.h ../../include/openssl/obj_mac.h
 e_aes.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 e_aes.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-e_aes.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-e_aes.o: ../../include/openssl/symhacks.h e_aes.c evp_locl.h
+e_aes.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
+e_aes.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+e_aes.o: ../modes/modes_lcl.h e_aes.c evp_locl.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/aes.h ../../include/openssl/asn1.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/bio.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/crypto.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/e_os2.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/evp.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/obj_mac.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/objects.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/opensslconf.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/opensslv.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/ossl_typ.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/safestack.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/sha.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/stack.h
+e_aes_cbc_hmac_sha1.o: ../../include/openssl/symhacks.h e_aes_cbc_hmac_sha1.c
+e_aes_cbc_hmac_sha1.o: evp_locl.h
 e_bf.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 e_bf.o: ../../include/openssl/blowfish.h ../../include/openssl/buffer.h
 e_bf.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 e_bf.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-e_bf.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-e_bf.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-e_bf.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-e_bf.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-e_bf.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-e_bf.o: ../cryptlib.h e_bf.c evp_locl.h
-e_camellia.o: ../../include/openssl/opensslconf.h e_camellia.c
+e_bf.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+e_bf.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+e_bf.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+e_bf.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+e_bf.o: ../../include/openssl/symhacks.h ../cryptlib.h e_bf.c evp_locl.h
+e_camellia.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+e_camellia.o: ../../include/openssl/camellia.h ../../include/openssl/crypto.h
+e_camellia.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
+e_camellia.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+e_camellia.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+e_camellia.o: ../../include/openssl/opensslconf.h
+e_camellia.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+e_camellia.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+e_camellia.o: ../../include/openssl/symhacks.h e_camellia.c evp_locl.h
 e_cast.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 e_cast.o: ../../include/openssl/buffer.h ../../include/openssl/cast.h
 e_cast.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 e_cast.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-e_cast.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-e_cast.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-e_cast.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-e_cast.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-e_cast.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-e_cast.o: ../cryptlib.h e_cast.c evp_locl.h
+e_cast.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+e_cast.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+e_cast.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+e_cast.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+e_cast.o: ../../include/openssl/symhacks.h ../cryptlib.h e_cast.c evp_locl.h
 e_des.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 e_des.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 e_des.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
 e_des.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-e_des.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-e_des.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-e_des.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-e_des.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-e_des.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-e_des.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-e_des.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
-e_des.o: ../cryptlib.h e_des.c evp_locl.h
+e_des.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+e_des.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+e_des.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+e_des.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
+e_des.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+e_des.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
+e_des.o: ../../include/openssl/ui_compat.h ../cryptlib.h e_des.c evp_locl.h
 e_des3.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 e_des3.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 e_des3.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
 e_des3.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-e_des3.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-e_des3.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-e_des3.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-e_des3.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-e_des3.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-e_des3.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-e_des3.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
-e_des3.o: ../cryptlib.h e_des3.c evp_locl.h
+e_des3.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+e_des3.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+e_des3.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+e_des3.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
+e_des3.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+e_des3.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
+e_des3.o: ../../include/openssl/ui_compat.h ../cryptlib.h e_des3.c evp_locl.h
 e_idea.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 e_idea.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 e_idea.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-e_idea.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-e_idea.o: ../../include/openssl/idea.h ../../include/openssl/lhash.h
-e_idea.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-e_idea.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-e_idea.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-e_idea.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-e_idea.o: ../cryptlib.h e_idea.c evp_locl.h
+e_idea.o: ../../include/openssl/evp.h ../../include/openssl/idea.h
+e_idea.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+e_idea.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+e_idea.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+e_idea.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+e_idea.o: ../../include/openssl/symhacks.h ../cryptlib.h e_idea.c evp_locl.h
 e_null.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 e_null.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 e_null.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-e_null.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-e_null.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-e_null.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-e_null.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-e_null.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-e_null.o: ../../include/openssl/symhacks.h ../cryptlib.h e_null.c
+e_null.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+e_null.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+e_null.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+e_null.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+e_null.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+e_null.o: ../cryptlib.h e_null.c
 e_old.o: e_old.c
 e_rc2.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 e_rc2.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 e_rc2.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-e_rc2.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-e_rc2.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-e_rc2.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-e_rc2.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-e_rc2.o: ../../include/openssl/rc2.h ../../include/openssl/safestack.h
-e_rc2.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-e_rc2.o: ../cryptlib.h e_rc2.c evp_locl.h
+e_rc2.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+e_rc2.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+e_rc2.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+e_rc2.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rc2.h
+e_rc2.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+e_rc2.o: ../../include/openssl/symhacks.h ../cryptlib.h e_rc2.c evp_locl.h
 e_rc4.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 e_rc4.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 e_rc4.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-e_rc4.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-e_rc4.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-e_rc4.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-e_rc4.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-e_rc4.o: ../../include/openssl/rc4.h ../../include/openssl/safestack.h
-e_rc4.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-e_rc4.o: ../cryptlib.h e_rc4.c evp_locl.h
+e_rc4.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+e_rc4.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+e_rc4.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+e_rc4.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rc4.h
+e_rc4.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+e_rc4.o: ../../include/openssl/symhacks.h ../cryptlib.h e_rc4.c evp_locl.h
+e_rc4_hmac_md5.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+e_rc4_hmac_md5.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+e_rc4_hmac_md5.o: ../../include/openssl/evp.h ../../include/openssl/md5.h
+e_rc4_hmac_md5.o: ../../include/openssl/obj_mac.h
+e_rc4_hmac_md5.o: ../../include/openssl/objects.h
+e_rc4_hmac_md5.o: ../../include/openssl/opensslconf.h
+e_rc4_hmac_md5.o: ../../include/openssl/opensslv.h
+e_rc4_hmac_md5.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rc4.h
+e_rc4_hmac_md5.o: ../../include/openssl/safestack.h
+e_rc4_hmac_md5.o: ../../include/openssl/stack.h
+e_rc4_hmac_md5.o: ../../include/openssl/symhacks.h e_rc4_hmac_md5.c
 e_rc5.o: ../../e_os.h ../../include/openssl/bio.h
 e_rc5.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 e_rc5.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
@@ -306,364 +319,361 @@ e_rc5.o: ../../include/openssl/symhacks.h ../cryptlib.h e_rc5.c
 e_seed.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 e_seed.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 e_seed.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-e_seed.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-e_seed.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-e_seed.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-e_seed.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+e_seed.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+e_seed.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+e_seed.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+e_seed.o: ../../include/openssl/safestack.h ../../include/openssl/seed.h
 e_seed.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-e_seed.o: e_seed.c
+e_seed.o: e_seed.c evp_locl.h
 e_xcbc_d.o: ../../e_os.h ../../include/openssl/asn1.h
 e_xcbc_d.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 e_xcbc_d.o: ../../include/openssl/crypto.h ../../include/openssl/des.h
 e_xcbc_d.o: ../../include/openssl/des_old.h ../../include/openssl/e_os2.h
 e_xcbc_d.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-e_xcbc_d.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-e_xcbc_d.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-e_xcbc_d.o: ../../include/openssl/opensslconf.h
+e_xcbc_d.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+e_xcbc_d.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 e_xcbc_d.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 e_xcbc_d.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 e_xcbc_d.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
 e_xcbc_d.o: ../../include/openssl/ui_compat.h ../cryptlib.h e_xcbc_d.c
-enc_min.o: ../../e_os.h ../../include/openssl/asn1.h
-enc_min.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
-enc_min.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
-enc_min.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
-enc_min.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
-enc_min.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-enc_min.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-enc_min.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-enc_min.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-enc_min.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-enc_min.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-enc_min.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-enc_min.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-enc_min.o: ../../include/openssl/x509_vfy.h ../cryptlib.h enc_min.c evp_locl.h
+e_xcbc_d.o: evp_locl.h
 encode.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 encode.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 encode.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-encode.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-encode.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-encode.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-encode.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-encode.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-encode.o: ../../include/openssl/symhacks.h ../cryptlib.h encode.c
+encode.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+encode.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+encode.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+encode.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+encode.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+encode.o: ../cryptlib.h encode.c
 evp_acnf.o: ../../e_os.h ../../include/openssl/asn1.h
 evp_acnf.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 evp_acnf.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 evp_acnf.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-evp_acnf.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-evp_acnf.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-evp_acnf.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+evp_acnf.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+evp_acnf.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+evp_acnf.o: ../../include/openssl/opensslconf.h
 evp_acnf.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 evp_acnf.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 evp_acnf.o: ../../include/openssl/symhacks.h ../cryptlib.h evp_acnf.c
-evp_cnf.o: ../../e_os.h ../../include/openssl/asn1.h
-evp_cnf.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
-evp_cnf.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
-evp_cnf.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
-evp_cnf.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
-evp_cnf.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-evp_cnf.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-evp_cnf.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-evp_cnf.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-evp_cnf.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-evp_cnf.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-evp_cnf.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-evp_cnf.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-evp_cnf.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-evp_cnf.o: ../cryptlib.h evp_cnf.c
 evp_enc.o: ../../e_os.h ../../include/openssl/asn1.h
 evp_enc.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 evp_enc.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 evp_enc.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 evp_enc.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 evp_enc.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-evp_enc.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-evp_enc.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-evp_enc.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-evp_enc.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-evp_enc.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-evp_enc.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-evp_enc.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-evp_enc.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_enc.c evp_locl.h
+evp_enc.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+evp_enc.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+evp_enc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+evp_enc.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+evp_enc.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+evp_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+evp_enc.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+evp_enc.o: ../cryptlib.h evp_enc.c evp_locl.h
 evp_err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 evp_err.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 evp_err.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-evp_err.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-evp_err.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-evp_err.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-evp_err.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-evp_err.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-evp_err.o: evp_err.c
+evp_err.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+evp_err.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+evp_err.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+evp_err.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+evp_err.o: ../../include/openssl/symhacks.h evp_err.c
+evp_fips.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+evp_fips.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+evp_fips.o: ../../include/openssl/evp.h ../../include/openssl/obj_mac.h
+evp_fips.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+evp_fips.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+evp_fips.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+evp_fips.o: ../../include/openssl/symhacks.h evp_fips.c
 evp_key.o: ../../e_os.h ../../include/openssl/asn1.h
 evp_key.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 evp_key.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 evp_key.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 evp_key.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-evp_key.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-evp_key.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-evp_key.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-evp_key.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-evp_key.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-evp_key.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-evp_key.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
-evp_key.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-evp_key.o: ../cryptlib.h evp_key.c
+evp_key.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+evp_key.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+evp_key.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+evp_key.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+evp_key.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+evp_key.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+evp_key.o: ../../include/openssl/ui.h ../../include/openssl/x509.h
+evp_key.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_key.c
 evp_lib.o: ../../e_os.h ../../include/openssl/asn1.h
 evp_lib.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 evp_lib.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 evp_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-evp_lib.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-evp_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-evp_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-evp_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-evp_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-evp_lib.o: ../cryptlib.h evp_lib.c
+evp_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+evp_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+evp_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+evp_lib.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+evp_lib.o: ../../include/openssl/symhacks.h ../cryptlib.h evp_lib.c
 evp_pbe.o: ../../e_os.h ../../include/openssl/asn1.h
 evp_pbe.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 evp_pbe.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 evp_pbe.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 evp_pbe.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-evp_pbe.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-evp_pbe.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-evp_pbe.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-evp_pbe.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+evp_pbe.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+evp_pbe.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+evp_pbe.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+evp_pbe.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
 evp_pbe.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 evp_pbe.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 evp_pbe.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-evp_pbe.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_pbe.c
+evp_pbe.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_locl.h evp_pbe.c
 evp_pkey.o: ../../e_os.h ../../include/openssl/asn1.h
-evp_pkey.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
-evp_pkey.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-evp_pkey.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
+evp_pkey.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+evp_pkey.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 evp_pkey.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 evp_pkey.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-evp_pkey.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-evp_pkey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-evp_pkey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+evp_pkey.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+evp_pkey.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+evp_pkey.o: ../../include/openssl/opensslconf.h
 evp_pkey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 evp_pkey.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
-evp_pkey.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
-evp_pkey.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-evp_pkey.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-evp_pkey.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_pkey.c
+evp_pkey.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+evp_pkey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+evp_pkey.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+evp_pkey.o: ../asn1/asn1_locl.h ../cryptlib.h evp_pkey.c
 m_dss.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 m_dss.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 m_dss.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 m_dss.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 m_dss.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-m_dss.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-m_dss.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-m_dss.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-m_dss.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-m_dss.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-m_dss.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-m_dss.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-m_dss.o: ../../include/openssl/x509_vfy.h ../cryptlib.h m_dss.c
+m_dss.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+m_dss.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+m_dss.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+m_dss.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+m_dss.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+m_dss.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+m_dss.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+m_dss.o: ../cryptlib.h m_dss.c
 m_dss1.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 m_dss1.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 m_dss1.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 m_dss1.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 m_dss1.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-m_dss1.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-m_dss1.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-m_dss1.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-m_dss1.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-m_dss1.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-m_dss1.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-m_dss1.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-m_dss1.o: ../../include/openssl/x509_vfy.h ../cryptlib.h m_dss1.c
+m_dss1.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+m_dss1.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+m_dss1.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+m_dss1.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+m_dss1.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+m_dss1.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+m_dss1.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+m_dss1.o: ../cryptlib.h m_dss1.c
 m_ecdsa.o: ../../e_os.h ../../include/openssl/asn1.h
 m_ecdsa.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 m_ecdsa.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 m_ecdsa.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 m_ecdsa.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-m_ecdsa.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-m_ecdsa.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-m_ecdsa.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-m_ecdsa.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-m_ecdsa.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-m_ecdsa.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-m_ecdsa.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-m_ecdsa.o: ../../include/openssl/x509_vfy.h ../cryptlib.h m_ecdsa.c
-m_md2.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
+m_ecdsa.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+m_ecdsa.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+m_ecdsa.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+m_ecdsa.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+m_ecdsa.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+m_ecdsa.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+m_ecdsa.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+m_ecdsa.o: ../cryptlib.h m_ecdsa.c
+m_md2.o: ../../e_os.h ../../include/openssl/bio.h
 m_md2.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-m_md2.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
-m_md2.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-m_md2.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-m_md2.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-m_md2.o: ../../include/openssl/md2.h ../../include/openssl/obj_mac.h
-m_md2.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+m_md2.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
+m_md2.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 m_md2.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-m_md2.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
-m_md2.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-m_md2.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-m_md2.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-m_md2.o: ../cryptlib.h evp_locl.h m_md2.c
+m_md2.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+m_md2.o: ../../include/openssl/symhacks.h ../cryptlib.h m_md2.c
 m_md4.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 m_md4.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 m_md4.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 m_md4.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 m_md4.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-m_md4.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-m_md4.o: ../../include/openssl/md4.h ../../include/openssl/obj_mac.h
-m_md4.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-m_md4.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-m_md4.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
-m_md4.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-m_md4.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-m_md4.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-m_md4.o: ../cryptlib.h evp_locl.h m_md4.c
+m_md4.o: ../../include/openssl/lhash.h ../../include/openssl/md4.h
+m_md4.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+m_md4.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+m_md4.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+m_md4.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+m_md4.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+m_md4.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+m_md4.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_locl.h m_md4.c
 m_md5.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 m_md5.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 m_md5.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 m_md5.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 m_md5.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-m_md5.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-m_md5.o: ../../include/openssl/md5.h ../../include/openssl/obj_mac.h
-m_md5.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-m_md5.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-m_md5.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
-m_md5.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-m_md5.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-m_md5.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-m_md5.o: ../cryptlib.h evp_locl.h m_md5.c
-m_mdc2.o: ../../e_os.h ../../include/openssl/bio.h
+m_md5.o: ../../include/openssl/lhash.h ../../include/openssl/md5.h
+m_md5.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+m_md5.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+m_md5.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+m_md5.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+m_md5.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+m_md5.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+m_md5.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_locl.h m_md5.c
+m_mdc2.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 m_mdc2.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-m_mdc2.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-m_mdc2.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
-m_mdc2.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-m_mdc2.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
-m_mdc2.o: ../../include/openssl/symhacks.h ../cryptlib.h evp_locl.h m_mdc2.c
+m_mdc2.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
+m_mdc2.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+m_mdc2.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+m_mdc2.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+m_mdc2.o: ../../include/openssl/lhash.h ../../include/openssl/mdc2.h
+m_mdc2.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+m_mdc2.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+m_mdc2.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+m_mdc2.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+m_mdc2.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+m_mdc2.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
+m_mdc2.o: ../../include/openssl/ui_compat.h ../../include/openssl/x509.h
+m_mdc2.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_locl.h m_mdc2.c
 m_null.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 m_null.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 m_null.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 m_null.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 m_null.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-m_null.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-m_null.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-m_null.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-m_null.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-m_null.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-m_null.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-m_null.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-m_null.o: ../cryptlib.h m_null.c
+m_null.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+m_null.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+m_null.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+m_null.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+m_null.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+m_null.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+m_null.o: ../../include/openssl/x509_vfy.h ../cryptlib.h m_null.c
 m_ripemd.o: ../../e_os.h ../../include/openssl/asn1.h
 m_ripemd.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 m_ripemd.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 m_ripemd.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 m_ripemd.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-m_ripemd.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-m_ripemd.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-m_ripemd.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+m_ripemd.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+m_ripemd.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+m_ripemd.o: ../../include/openssl/opensslconf.h
 m_ripemd.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 m_ripemd.o: ../../include/openssl/pkcs7.h ../../include/openssl/ripemd.h
 m_ripemd.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
 m_ripemd.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 m_ripemd.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-m_ripemd.o: ../../include/openssl/x509_vfy.h ../cryptlib.h m_ripemd.c
+m_ripemd.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_locl.h
+m_ripemd.o: m_ripemd.c
 m_sha.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 m_sha.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 m_sha.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 m_sha.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 m_sha.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-m_sha.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-m_sha.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-m_sha.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-m_sha.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-m_sha.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
-m_sha.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-m_sha.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-m_sha.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_locl.h m_sha.c
+m_sha.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+m_sha.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+m_sha.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+m_sha.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+m_sha.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+m_sha.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+m_sha.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+m_sha.o: ../cryptlib.h evp_locl.h m_sha.c
 m_sha1.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 m_sha1.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 m_sha1.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 m_sha1.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 m_sha1.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-m_sha1.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-m_sha1.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-m_sha1.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-m_sha1.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-m_sha1.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
-m_sha1.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-m_sha1.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-m_sha1.o: ../../include/openssl/x509_vfy.h ../cryptlib.h m_sha1.c
+m_sha1.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+m_sha1.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+m_sha1.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+m_sha1.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+m_sha1.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+m_sha1.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+m_sha1.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+m_sha1.o: ../cryptlib.h m_sha1.c
+m_sigver.o: ../../e_os.h ../../include/openssl/asn1.h
+m_sigver.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+m_sigver.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+m_sigver.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+m_sigver.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+m_sigver.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+m_sigver.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+m_sigver.o: ../../include/openssl/opensslconf.h
+m_sigver.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+m_sigver.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+m_sigver.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+m_sigver.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+m_sigver.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_locl.h
+m_sigver.o: m_sigver.c
+m_wp.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
+m_wp.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+m_wp.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+m_wp.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+m_wp.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+m_wp.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+m_wp.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+m_wp.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+m_wp.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+m_wp.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+m_wp.o: ../../include/openssl/symhacks.h ../../include/openssl/whrlpool.h
+m_wp.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+m_wp.o: ../cryptlib.h evp_locl.h m_wp.c
 names.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 names.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 names.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 names.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 names.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-names.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-names.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-names.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-names.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-names.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-names.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-names.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-names.o: ../cryptlib.h names.c
+names.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+names.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+names.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+names.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+names.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+names.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+names.o: ../../include/openssl/x509_vfy.h ../cryptlib.h names.c
 p5_crpt.o: ../../e_os.h ../../include/openssl/asn1.h
 p5_crpt.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p5_crpt.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p5_crpt.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p5_crpt.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p5_crpt.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p5_crpt.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p5_crpt.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-p5_crpt.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-p5_crpt.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-p5_crpt.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-p5_crpt.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-p5_crpt.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p5_crpt.c
+p5_crpt.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p5_crpt.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p5_crpt.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+p5_crpt.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+p5_crpt.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+p5_crpt.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+p5_crpt.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+p5_crpt.o: ../cryptlib.h p5_crpt.c
 p5_crpt2.o: ../../e_os.h ../../include/openssl/asn1.h
 p5_crpt2.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p5_crpt2.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p5_crpt2.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p5_crpt2.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p5_crpt2.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p5_crpt2.o: ../../include/openssl/hmac.h ../../include/openssl/lhash.h
-p5_crpt2.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p5_crpt2.o: ../../include/openssl/opensslconf.h
+p5_crpt2.o: ../../include/openssl/evp.h ../../include/openssl/hmac.h
+p5_crpt2.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p5_crpt2.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 p5_crpt2.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p5_crpt2.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 p5_crpt2.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 p5_crpt2.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-p5_crpt2.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p5_crpt2.c
+p5_crpt2.o: ../../include/openssl/x509_vfy.h ../cryptlib.h evp_locl.h
+p5_crpt2.o: p5_crpt2.c
 p_dec.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 p_dec.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p_dec.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p_dec.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p_dec.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p_dec.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p_dec.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p_dec.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-p_dec.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-p_dec.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
-p_dec.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p_dec.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p_dec.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p_dec.o: ../cryptlib.h p_dec.c
+p_dec.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p_dec.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p_dec.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+p_dec.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+p_dec.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+p_dec.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p_dec.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p_dec.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p_dec.c
 p_enc.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 p_enc.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p_enc.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p_enc.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p_enc.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p_enc.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p_enc.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p_enc.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-p_enc.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-p_enc.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
-p_enc.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p_enc.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p_enc.o: ../cryptlib.h p_enc.c
+p_enc.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p_enc.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p_enc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+p_enc.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+p_enc.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+p_enc.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p_enc.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p_enc.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p_enc.c
 p_lib.o: ../../e_os.h ../../include/openssl/asn1.h
 p_lib.o: ../../include/openssl/asn1_mac.h ../../include/openssl/bio.h
 p_lib.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
 p_lib.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
 p_lib.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 p_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
-p_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p_lib.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
+p_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
+p_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
 p_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
 p_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 p_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
@@ -671,57 +681,91 @@ p_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
 p_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 p_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 p_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p_lib.o: ../cryptlib.h p_lib.c
+p_lib.o: ../asn1/asn1_locl.h ../cryptlib.h p_lib.c
 p_open.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 p_open.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p_open.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p_open.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p_open.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p_open.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p_open.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p_open.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-p_open.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-p_open.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
-p_open.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-p_open.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-p_open.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p_open.c
+p_open.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p_open.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p_open.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+p_open.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+p_open.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+p_open.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+p_open.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+p_open.o: ../cryptlib.h p_open.c
 p_seal.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 p_seal.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p_seal.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p_seal.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p_seal.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p_seal.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p_seal.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p_seal.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-p_seal.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-p_seal.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
-p_seal.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p_seal.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p_seal.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p_seal.o: ../cryptlib.h p_seal.c
+p_seal.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p_seal.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p_seal.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+p_seal.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+p_seal.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+p_seal.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p_seal.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p_seal.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p_seal.c
 p_sign.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 p_sign.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p_sign.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p_sign.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p_sign.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p_sign.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p_sign.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p_sign.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-p_sign.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-p_sign.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p_sign.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p_sign.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p_sign.o: ../cryptlib.h p_sign.c
+p_sign.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p_sign.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p_sign.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+p_sign.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+p_sign.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p_sign.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p_sign.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p_sign.c
 p_verify.o: ../../e_os.h ../../include/openssl/asn1.h
 p_verify.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p_verify.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p_verify.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p_verify.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p_verify.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p_verify.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p_verify.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p_verify.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p_verify.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p_verify.o: ../../include/openssl/opensslconf.h
 p_verify.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p_verify.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 p_verify.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 p_verify.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 p_verify.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p_verify.c
+pmeth_fn.o: ../../e_os.h ../../include/openssl/asn1.h
+pmeth_fn.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+pmeth_fn.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+pmeth_fn.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+pmeth_fn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pmeth_fn.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pmeth_fn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+pmeth_fn.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+pmeth_fn.o: ../../include/openssl/symhacks.h ../cryptlib.h evp_locl.h
+pmeth_fn.o: pmeth_fn.c
+pmeth_gn.o: ../../e_os.h ../../include/openssl/asn1.h
+pmeth_gn.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+pmeth_gn.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+pmeth_gn.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
+pmeth_gn.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pmeth_gn.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pmeth_gn.o: ../../include/openssl/opensslconf.h
+pmeth_gn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+pmeth_gn.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+pmeth_gn.o: ../../include/openssl/symhacks.h ../cryptlib.h evp_locl.h
+pmeth_gn.o: pmeth_gn.c
+pmeth_lib.o: ../../e_os.h ../../include/openssl/asn1.h
+pmeth_lib.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+pmeth_lib.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+pmeth_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+pmeth_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
+pmeth_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+pmeth_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pmeth_lib.o: ../../include/openssl/objects.h
+pmeth_lib.o: ../../include/openssl/opensslconf.h
+pmeth_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+pmeth_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+pmeth_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+pmeth_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+pmeth_lib.o: ../../include/openssl/x509_vfy.h ../asn1/asn1_locl.h ../cryptlib.h
+pmeth_lib.o: evp_locl.h pmeth_lib.c
diff --git a/crypto/evp/bio_enc.c b/crypto/evp/bio_enc.c
index f6ac94c6e1bb..b6efb5fbc4b1 100644
--- a/crypto/evp/bio_enc.c
+++ b/crypto/evp/bio_enc.c
@@ -361,8 +361,10 @@ again:
 	case BIO_CTRL_DUP:
 		dbio=(BIO *)ptr;
 		dctx=(BIO_ENC_CTX *)dbio->ptr;
-		memcpy(&(dctx->cipher),&(ctx->cipher),sizeof(ctx->cipher));
-		dbio->init=1;
+		EVP_CIPHER_CTX_init(&dctx->cipher);
+		ret = EVP_CIPHER_CTX_copy(&dctx->cipher,&ctx->cipher);
+		if (ret)
+			dbio->init=1;
 		break;
 	default:
 		ret=BIO_ctrl(b->next_bio,cmd,num,ptr);
diff --git a/crypto/evp/bio_md.c b/crypto/evp/bio_md.c
index ed5c1135fd45..144fdfd56a0e 100644
--- a/crypto/evp/bio_md.c
+++ b/crypto/evp/bio_md.c
@@ -130,8 +130,8 @@ static int md_read(BIO *b, char *out, int outl)
 		{
 		if (ret > 0)
 			{
-			EVP_DigestUpdate(ctx,(unsigned char *)out,
-				(unsigned int)ret);
+			if (EVP_DigestUpdate(ctx,(unsigned char *)out,
+				(unsigned int)ret)<=0) return (-1);
 			}
 		}
 	BIO_clear_retry_flags(b);
@@ -153,12 +153,19 @@ static int md_write(BIO *b, const char *in, int inl)
 		{
 		if (ret > 0)
 			{
-			EVP_DigestUpdate(ctx,(const unsigned char *)in,
-				(unsigned int)ret);
+			if (!EVP_DigestUpdate(ctx,(const unsigned char *)in,
+				(unsigned int)ret))
+				{
+				BIO_clear_retry_flags(b);
+				return 0;
+				}
 			}
 		}
-	BIO_clear_retry_flags(b);
-	BIO_copy_next_retry(b);
+	if(b->next_bio != NULL)
+		{
+		BIO_clear_retry_flags(b);
+		BIO_copy_next_retry(b);
+		}
 	return(ret);
 	}
 
@@ -194,6 +201,7 @@ static long md_ctrl(BIO *b, int cmd, long num, void *ptr)
 	case BIO_C_GET_MD_CTX:
 		pctx=ptr;
 		*pctx=ctx;
+		b->init = 1;
 		break;
 	case BIO_C_SET_MD_CTX:
 		if (b->init)
@@ -216,7 +224,8 @@ static long md_ctrl(BIO *b, int cmd, long num, void *ptr)
 	case BIO_CTRL_DUP:
 		dbio=ptr;
 		dctx=dbio->ptr;
-		EVP_MD_CTX_copy_ex(dctx,ctx);
+		if (!EVP_MD_CTX_copy_ex(dctx,ctx))
+			return 0;
 		b->init=1;
 		break;
 	default:
@@ -249,7 +258,9 @@ static int md_gets(BIO *bp, char *buf, int size)
 	ctx=bp->ptr;
 	if (size < ctx->digest->md_size)
 		return(0);
-	EVP_DigestFinal_ex(ctx,(unsigned char *)buf,&ret);
+	if (EVP_DigestFinal_ex(ctx,(unsigned char *)buf,&ret)<=0) 
+		return -1;
+		
 	return((int)ret);
 	}
 
diff --git a/crypto/evp/bio_ok.c b/crypto/evp/bio_ok.c
index 98bc1ab40963..e64335353fd5 100644
--- a/crypto/evp/bio_ok.c
+++ b/crypto/evp/bio_ok.c
@@ -133,10 +133,10 @@ static int ok_new(BIO *h);
 static int ok_free(BIO *data);
 static long ok_callback_ctrl(BIO *h, int cmd, bio_info_cb *fp);
 
-static void sig_out(BIO* b);
-static void sig_in(BIO* b);
-static void block_out(BIO* b);
-static void block_in(BIO* b);
+static int sig_out(BIO* b);
+static int sig_in(BIO* b);
+static int block_out(BIO* b);
+static int block_in(BIO* b);
 #define OK_BLOCK_SIZE	(1024*4)
 #define OK_BLOCK_BLOCK	4
 #define IOBS		(OK_BLOCK_SIZE+ OK_BLOCK_BLOCK+ 3*EVP_MAX_MD_SIZE)
@@ -266,10 +266,24 @@ static int ok_read(BIO *b, char *out, int outl)
 		ctx->buf_len+= i;
 
 		/* no signature yet -- check if we got one */
-		if (ctx->sigio == 1) sig_in(b);
+		if (ctx->sigio == 1)
+			{
+			if (!sig_in(b))
+				{
+				BIO_clear_retry_flags(b);
+				return 0;
+				}
+			}
 
 		/* signature ok -- check if we got block */
-		if (ctx->sigio == 0) block_in(b);
+		if (ctx->sigio == 0)
+			{
+			if (!block_in(b))
+				{
+				BIO_clear_retry_flags(b);
+				return 0;
+				}
+			}
 
 		/* invalid block -- cancel */
 		if (ctx->cont <= 0) break;
@@ -293,7 +307,8 @@ static int ok_write(BIO *b, const char *in, int inl)
 
 	if ((ctx == NULL) || (b->next_bio == NULL) || (b->init == 0)) return(0);
 
-	if(ctx->sigio) sig_out(b);
+	if(ctx->sigio && !sig_out(b))
+		return 0;
 
 	do{
 		BIO_clear_retry_flags(b);
@@ -332,7 +347,11 @@ static int ok_write(BIO *b, const char *in, int inl)
 
 		if(ctx->buf_len >= OK_BLOCK_SIZE+ OK_BLOCK_BLOCK)
 			{
-			block_out(b);
+			if (!block_out(b))
+				{
+				BIO_clear_retry_flags(b);
+				return 0;
+				}
 			}
 	}while(inl > 0);
 
@@ -379,7 +398,8 @@ static long ok_ctrl(BIO *b, int cmd, long num, void *ptr)
 	case BIO_CTRL_FLUSH:
 		/* do a final write */
 		if(ctx->blockout == 0)
-			block_out(b);
+			if (!block_out(b))
+				return 0;
 
 		while (ctx->blockout)
 			{
@@ -408,7 +428,8 @@ static long ok_ctrl(BIO *b, int cmd, long num, void *ptr)
 		break;
 	case BIO_C_SET_MD:
 		md=ptr;
-		EVP_DigestInit_ex(&ctx->md, md, NULL);
+		if (!EVP_DigestInit_ex(&ctx->md, md, NULL))
+			return 0;
 		b->init=1;
 		break;
 	case BIO_C_GET_MD:
@@ -455,7 +476,7 @@ static void longswap(void *_ptr, size_t len)
 	}
 }
 
-static void sig_out(BIO* b)
+static int sig_out(BIO* b)
 	{
 	BIO_OK_CTX *ctx;
 	EVP_MD_CTX *md;
@@ -463,9 +484,10 @@ static void sig_out(BIO* b)
 	ctx=b->ptr;
 	md=&ctx->md;
 
-	if(ctx->buf_len+ 2* md->digest->md_size > OK_BLOCK_SIZE) return;
+	if(ctx->buf_len+ 2* md->digest->md_size > OK_BLOCK_SIZE) return 1;
 
-	EVP_DigestInit_ex(md, md->digest, NULL);
+	if (!EVP_DigestInit_ex(md, md->digest, NULL))
+		goto berr;
 	/* FIXME: there's absolutely no guarantee this makes any sense at all,
 	 * particularly now EVP_MD_CTX has been restructured.
 	 */
@@ -474,14 +496,20 @@ static void sig_out(BIO* b)
 	longswap(&(ctx->buf[ctx->buf_len]), md->digest->md_size);
 	ctx->buf_len+= md->digest->md_size;
 
-	EVP_DigestUpdate(md, WELLKNOWN, strlen(WELLKNOWN));
-	EVP_DigestFinal_ex(md, &(ctx->buf[ctx->buf_len]), NULL);
+	if (!EVP_DigestUpdate(md, WELLKNOWN, strlen(WELLKNOWN)))
+		goto berr;
+	if (!EVP_DigestFinal_ex(md, &(ctx->buf[ctx->buf_len]), NULL))
+		goto berr;
 	ctx->buf_len+= md->digest->md_size;
 	ctx->blockout= 1;
 	ctx->sigio= 0;
+	return 1;
+	berr:
+	BIO_clear_retry_flags(b);
+	return 0;
 	}
 
-static void sig_in(BIO* b)
+static int sig_in(BIO* b)
 	{
 	BIO_OK_CTX *ctx;
 	EVP_MD_CTX *md;
@@ -491,15 +519,18 @@ static void sig_in(BIO* b)
 	ctx=b->ptr;
 	md=&ctx->md;
 
-	if((int)(ctx->buf_len-ctx->buf_off) < 2*md->digest->md_size) return;
+	if((int)(ctx->buf_len-ctx->buf_off) < 2*md->digest->md_size) return 1;
 
-	EVP_DigestInit_ex(md, md->digest, NULL);
+	if (!EVP_DigestInit_ex(md, md->digest, NULL))
+		goto berr;
 	memcpy(md->md_data, &(ctx->buf[ctx->buf_off]), md->digest->md_size);
 	longswap(md->md_data, md->digest->md_size);
 	ctx->buf_off+= md->digest->md_size;
 
-	EVP_DigestUpdate(md, WELLKNOWN, strlen(WELLKNOWN));
-	EVP_DigestFinal_ex(md, tmp, NULL);
+	if (!EVP_DigestUpdate(md, WELLKNOWN, strlen(WELLKNOWN)))
+		goto berr;
+	if (!EVP_DigestFinal_ex(md, tmp, NULL))
+		goto berr;
 	ret= memcmp(&(ctx->buf[ctx->buf_off]), tmp, md->digest->md_size) == 0;
 	ctx->buf_off+= md->digest->md_size;
 	if(ret == 1)
@@ -516,9 +547,13 @@ static void sig_in(BIO* b)
 		{
 		ctx->cont= 0;
 		}
+	return 1;
+	berr:
+	BIO_clear_retry_flags(b);
+	return 0;
 	}
 
-static void block_out(BIO* b)
+static int block_out(BIO* b)
 	{
 	BIO_OK_CTX *ctx;
 	EVP_MD_CTX *md;
@@ -532,13 +567,20 @@ static void block_out(BIO* b)
 	ctx->buf[1]=(unsigned char)(tl>>16);
 	ctx->buf[2]=(unsigned char)(tl>>8);
 	ctx->buf[3]=(unsigned char)(tl);
-	EVP_DigestUpdate(md, (unsigned char*) &(ctx->buf[OK_BLOCK_BLOCK]), tl);
-	EVP_DigestFinal_ex(md, &(ctx->buf[ctx->buf_len]), NULL);
+	if (!EVP_DigestUpdate(md,
+		(unsigned char*) &(ctx->buf[OK_BLOCK_BLOCK]), tl))
+		goto berr;
+	if (!EVP_DigestFinal_ex(md, &(ctx->buf[ctx->buf_len]), NULL))
+		goto berr;
 	ctx->buf_len+= md->digest->md_size;
 	ctx->blockout= 1;
+	return 1;
+	berr:
+	BIO_clear_retry_flags(b);
+	return 0;
 	}
 
-static void block_in(BIO* b)
+static int block_in(BIO* b)
 	{
 	BIO_OK_CTX *ctx;
 	EVP_MD_CTX *md;
@@ -554,10 +596,13 @@ static void block_in(BIO* b)
 	tl|=ctx->buf[2]; tl<<=8;
 	tl|=ctx->buf[3];
 
-	if (ctx->buf_len < tl+ OK_BLOCK_BLOCK+ md->digest->md_size) return;
+	if (ctx->buf_len < tl+ OK_BLOCK_BLOCK+ md->digest->md_size) return 1;
  
-	EVP_DigestUpdate(md, (unsigned char*) &(ctx->buf[OK_BLOCK_BLOCK]), tl);
-	EVP_DigestFinal_ex(md, tmp, NULL);
+	if (!EVP_DigestUpdate(md,
+			(unsigned char*) &(ctx->buf[OK_BLOCK_BLOCK]), tl))
+		goto berr;
+	if (!EVP_DigestFinal_ex(md, tmp, NULL))
+		goto berr;
 	if(memcmp(&(ctx->buf[tl+ OK_BLOCK_BLOCK]), tmp, md->digest->md_size) == 0)
 		{
 		/* there might be parts from next block lurking around ! */
@@ -571,5 +616,9 @@ static void block_in(BIO* b)
 		{
 		ctx->cont= 0;
 		}
+	return 1;
+	berr:
+	BIO_clear_retry_flags(b);
+	return 0;
 	}
 
diff --git a/crypto/evp/c_all.c b/crypto/evp/c_all.c
index a5da52e62d8b..766c4cecdf7e 100644
--- a/crypto/evp/c_all.c
+++ b/crypto/evp/c_all.c
@@ -83,7 +83,7 @@ void OPENSSL_add_all_algorithms_noconf(void)
 	OpenSSL_add_all_ciphers();
 	OpenSSL_add_all_digests();
 #ifndef OPENSSL_NO_ENGINE
-# if defined(__OpenBSD__) || defined(__FreeBSD__)
+# if defined(__OpenBSD__) || defined(__FreeBSD__) || defined(HAVE_CRYPTODEV)
 	ENGINE_setup_bsd_cryptodev();
 # endif
 #endif
diff --git a/crypto/evp/c_allc.c b/crypto/evp/c_allc.c
index e45cee8ab080..2a45d435e58e 100644
--- a/crypto/evp/c_allc.c
+++ b/crypto/evp/c_allc.c
@@ -98,6 +98,9 @@ void OpenSSL_add_all_ciphers(void)
 #ifndef OPENSSL_NO_RC4
 	EVP_add_cipher(EVP_rc4());
 	EVP_add_cipher(EVP_rc4_40());
+#ifndef OPENSSL_NO_MD5
+	EVP_add_cipher(EVP_rc4_hmac_md5());
+#endif
 #endif
 
 #ifndef OPENSSL_NO_IDEA
@@ -166,9 +169,9 @@ void OpenSSL_add_all_ciphers(void)
 	EVP_add_cipher(EVP_aes_128_cfb1());
 	EVP_add_cipher(EVP_aes_128_cfb8());
 	EVP_add_cipher(EVP_aes_128_ofb());
-#if 0
 	EVP_add_cipher(EVP_aes_128_ctr());
-#endif
+	EVP_add_cipher(EVP_aes_128_gcm());
+	EVP_add_cipher(EVP_aes_128_xts());
 	EVP_add_cipher_alias(SN_aes_128_cbc,"AES128");
 	EVP_add_cipher_alias(SN_aes_128_cbc,"aes128");
 	EVP_add_cipher(EVP_aes_192_ecb());
@@ -177,9 +180,8 @@ void OpenSSL_add_all_ciphers(void)
 	EVP_add_cipher(EVP_aes_192_cfb1());
 	EVP_add_cipher(EVP_aes_192_cfb8());
 	EVP_add_cipher(EVP_aes_192_ofb());
-#if 0
 	EVP_add_cipher(EVP_aes_192_ctr());
-#endif
+	EVP_add_cipher(EVP_aes_192_gcm());
 	EVP_add_cipher_alias(SN_aes_192_cbc,"AES192");
 	EVP_add_cipher_alias(SN_aes_192_cbc,"aes192");
 	EVP_add_cipher(EVP_aes_256_ecb());
@@ -188,11 +190,15 @@ void OpenSSL_add_all_ciphers(void)
 	EVP_add_cipher(EVP_aes_256_cfb1());
 	EVP_add_cipher(EVP_aes_256_cfb8());
 	EVP_add_cipher(EVP_aes_256_ofb());
-#if 0
 	EVP_add_cipher(EVP_aes_256_ctr());
-#endif
+	EVP_add_cipher(EVP_aes_256_gcm());
+	EVP_add_cipher(EVP_aes_256_xts());
 	EVP_add_cipher_alias(SN_aes_256_cbc,"AES256");
 	EVP_add_cipher_alias(SN_aes_256_cbc,"aes256");
+#if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA1)
+	EVP_add_cipher(EVP_aes_128_cbc_hmac_sha1());
+	EVP_add_cipher(EVP_aes_256_cbc_hmac_sha1());
+#endif
 #endif
 
 #ifndef OPENSSL_NO_CAMELLIA
@@ -221,7 +227,4 @@ void OpenSSL_add_all_ciphers(void)
 	EVP_add_cipher_alias(SN_camellia_256_cbc,"CAMELLIA256");
 	EVP_add_cipher_alias(SN_camellia_256_cbc,"camellia256");
 #endif
-
-	PKCS12_PBE_add();
-	PKCS5_PBE_add();
 	}
diff --git a/crypto/evp/c_alld.c b/crypto/evp/c_alld.c
index e0841d12f278..311e1fe2f851 100644
--- a/crypto/evp/c_alld.c
+++ b/crypto/evp/c_alld.c
@@ -78,7 +78,7 @@ void OpenSSL_add_all_digests(void)
 	EVP_add_digest(EVP_dss());
 #endif
 #endif
-#ifndef OPENSSL_NO_SHA
+#if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA1)
 	EVP_add_digest(EVP_sha1());
 	EVP_add_digest_alias(SN_sha1,"ssl3-sha1");
 	EVP_add_digest_alias(SN_sha1WithRSAEncryption,SN_sha1WithRSA);
@@ -108,4 +108,7 @@ void OpenSSL_add_all_digests(void)
 	EVP_add_digest(EVP_sha384());
 	EVP_add_digest(EVP_sha512());
 #endif
+#ifndef OPENSSL_NO_WHIRLPOOL
+	EVP_add_digest(EVP_whirlpool());
+#endif
 	}
diff --git a/crypto/evp/dig_eng.c b/crypto/evp/dig_eng.c
deleted file mode 100644
index 64cdf9366c9c..000000000000
--- a/crypto/evp/dig_eng.c
+++ /dev/null
@@ -1,180 +0,0 @@
-/* crypto/evp/digest.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-/* ====================================================================
- * Copyright (c) 1998-2001 The OpenSSL Project.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    openssl-core@openssl.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
- */
-
-#include <stdio.h>
-#include "cryptlib.h"
-#include <openssl/objects.h>
-#include <openssl/evp.h>
-#ifndef OPENSSL_NO_ENGINE
-#include <openssl/engine.h>
-#endif
-#include "evp_locl.h"
-
-#ifndef OPENSSL_NO_ENGINE
-
-#ifdef OPENSSL_FIPS
-
-static int do_evp_md_engine_full(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl)
-	{
-	if (*ptype)
-		{
-		/* Ensure an ENGINE left lying around from last time is cleared
-		 * (the previous check attempted to avoid this if the same
-		 * ENGINE and EVP_MD could be used). */
-		if(ctx->engine)
-			ENGINE_finish(ctx->engine);
-		if(impl)
-			{
-			if (!ENGINE_init(impl))
-				{
-				EVPerr(EVP_F_DO_EVP_MD_ENGINE_FULL,EVP_R_INITIALIZATION_ERROR);
-				return 0;
-				}
-			}
-		else
-			/* Ask if an ENGINE is reserved for this job */
-			impl = ENGINE_get_digest_engine((*ptype)->type);
-		if(impl)
-			{
-			/* There's an ENGINE for this job ... (apparently) */
-			const EVP_MD *d = ENGINE_get_digest(impl, (*ptype)->type);
-			if(!d)
-				{
-				/* Same comment from evp_enc.c */
-				EVPerr(EVP_F_DO_EVP_MD_ENGINE_FULL,EVP_R_INITIALIZATION_ERROR);
-				return 0;
-				}
-			/* We'll use the ENGINE's private digest definition */
-			*ptype = d;
-			/* Store the ENGINE functional reference so we know
-			 * 'type' came from an ENGINE and we need to release
-			 * it when done. */
-			ctx->engine = impl;
-			}
-		else
-			ctx->engine = NULL;
-		}
-	else
-	if(!ctx->digest)
-		{
-		EVPerr(EVP_F_DO_EVP_MD_ENGINE_FULL,EVP_R_NO_DIGEST_SET);
-		return 0;
-		}
-	return 1;
-	}
-
-void int_EVP_MD_init_engine_callbacks(void)
-	{
-	int_EVP_MD_set_engine_callbacks(
-		ENGINE_init, ENGINE_finish, do_evp_md_engine_full);
-	}
-#endif
-#endif
diff --git a/crypto/evp/digest.c b/crypto/evp/digest.c
index 10a36071f71b..467e6b5ae9cf 100644
--- a/crypto/evp/digest.c
+++ b/crypto/evp/digest.c
@@ -116,7 +116,10 @@
 #ifndef OPENSSL_NO_ENGINE
 #include <openssl/engine.h>
 #endif
-#include "evp_locl.h"
+
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 
 void EVP_MD_CTX_init(EVP_MD_CTX *ctx)
 	{
@@ -139,77 +142,18 @@ int EVP_DigestInit(EVP_MD_CTX *ctx, const EVP_MD *type)
 	return EVP_DigestInit_ex(ctx, type, NULL);
 	}
 
-#ifdef OPENSSL_FIPS
-
-/* The purpose of these is to trap programs that attempt to use non FIPS
- * algorithms in FIPS mode and ignore the errors.
- */
-
-static int bad_init(EVP_MD_CTX *ctx)
-	{ FIPS_ERROR_IGNORED("Digest init"); return 0;}
-
-static int bad_update(EVP_MD_CTX *ctx,const void *data,size_t count)
-	{ FIPS_ERROR_IGNORED("Digest update"); return 0;}
-
-static int bad_final(EVP_MD_CTX *ctx,unsigned char *md)
-	{ FIPS_ERROR_IGNORED("Digest Final"); return 0;}
-
-static const EVP_MD bad_md =
+int EVP_DigestInit_ex(EVP_MD_CTX *ctx, const EVP_MD *type, ENGINE *impl)
 	{
-	0,
-	0,
-	0,
-	0,
-	bad_init,
-	bad_update,
-	bad_final,
-	NULL,
-	NULL,
-	NULL,
-	0,
-	{0,0,0,0},
-	};
-
-#endif
-
+	EVP_MD_CTX_clear_flags(ctx,EVP_MD_CTX_FLAG_CLEANED);
 #ifndef OPENSSL_NO_ENGINE
-
-#ifdef OPENSSL_FIPS
-
-static int do_engine_null(ENGINE *impl) { return 0;}
-static int do_evp_md_engine_null(EVP_MD_CTX *ctx,
-				const EVP_MD **ptype, ENGINE *impl)
-	{ return 1; }
-
-static int (*do_engine_init)(ENGINE *impl)
-		= do_engine_null;
-
-static int (*do_engine_finish)(ENGINE *impl)
-		= do_engine_null;
-
-static int (*do_evp_md_engine)
-	(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl)
-		= do_evp_md_engine_null;
-
-void int_EVP_MD_set_engine_callbacks(
-	int (*eng_md_init)(ENGINE *impl),
-	int (*eng_md_fin)(ENGINE *impl),
-	int (*eng_md_evp)
-		(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl))
-	{
-	do_engine_init = eng_md_init;
-	do_engine_finish = eng_md_fin;
-	do_evp_md_engine = eng_md_evp;
-	}
-
-#else
-
-#define do_engine_init	ENGINE_init
-#define do_engine_finish ENGINE_finish
-
-static int do_evp_md_engine(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl)
-	{
-	if (*ptype)
+	/* Whether it's nice or not, "Inits" can be used on "Final"'d contexts
+	 * so this context may already have an ENGINE! Try to avoid releasing
+	 * the previous handle, re-querying for an ENGINE, and having a
+	 * reinitialisation, when it may all be unecessary. */
+	if (ctx->engine && ctx->digest && (!type ||
+			(type && (type->type == ctx->digest->type))))
+		goto skip_to_init;
+	if (type)
 		{
 		/* Ensure an ENGINE left lying around from last time is cleared
 		 * (the previous check attempted to avoid this if the same
@@ -220,26 +164,26 @@ static int do_evp_md_engine(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl)
 			{
 			if (!ENGINE_init(impl))
 				{
-				EVPerr(EVP_F_DO_EVP_MD_ENGINE,EVP_R_INITIALIZATION_ERROR);
+				EVPerr(EVP_F_EVP_DIGESTINIT_EX,EVP_R_INITIALIZATION_ERROR);
 				return 0;
 				}
 			}
 		else
 			/* Ask if an ENGINE is reserved for this job */
-			impl = ENGINE_get_digest_engine((*ptype)->type);
+			impl = ENGINE_get_digest_engine(type->type);
 		if(impl)
 			{
 			/* There's an ENGINE for this job ... (apparently) */
-			const EVP_MD *d = ENGINE_get_digest(impl, (*ptype)->type);
+			const EVP_MD *d = ENGINE_get_digest(impl, type->type);
 			if(!d)
 				{
 				/* Same comment from evp_enc.c */
-				EVPerr(EVP_F_DO_EVP_MD_ENGINE,EVP_R_INITIALIZATION_ERROR);
+				EVPerr(EVP_F_EVP_DIGESTINIT_EX,EVP_R_INITIALIZATION_ERROR);
 				ENGINE_finish(impl);
 				return 0;
 				}
 			/* We'll use the ENGINE's private digest definition */
-			*ptype = d;
+			type = d;
 			/* Store the ENGINE functional reference so we know
 			 * 'type' came from an ENGINE and we need to release
 			 * it when done. */
@@ -251,78 +195,60 @@ static int do_evp_md_engine(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl)
 	else
 	if(!ctx->digest)
 		{
-		EVPerr(EVP_F_DO_EVP_MD_ENGINE,EVP_R_NO_DIGEST_SET);
+		EVPerr(EVP_F_EVP_DIGESTINIT_EX,EVP_R_NO_DIGEST_SET);
 		return 0;
 		}
-	return 1;
-	}
-
-#endif
-
-#endif
-
-int EVP_DigestInit_ex(EVP_MD_CTX *ctx, const EVP_MD *type, ENGINE *impl)
-	{
-	M_EVP_MD_CTX_clear_flags(ctx,EVP_MD_CTX_FLAG_CLEANED);
-#ifdef OPENSSL_FIPS
-	if(FIPS_selftest_failed())
-		{
-		FIPSerr(FIPS_F_EVP_DIGESTINIT_EX,FIPS_R_FIPS_SELFTEST_FAILED);
-		ctx->digest = &bad_md;
-		return 0;
-		}
-#endif
-#ifndef OPENSSL_NO_ENGINE
-	/* Whether it's nice or not, "Inits" can be used on "Final"'d contexts
-	 * so this context may already have an ENGINE! Try to avoid releasing
-	 * the previous handle, re-querying for an ENGINE, and having a
-	 * reinitialisation, when it may all be unecessary. */
-	if (ctx->engine && ctx->digest && (!type ||
-			(type && (type->type == ctx->digest->type))))
-		goto skip_to_init;
-	if (!do_evp_md_engine(ctx, &type, impl))
-		return 0;
 #endif
 	if (ctx->digest != type)
 		{
-#ifdef OPENSSL_FIPS
-		if (FIPS_mode())
-			{
-			if (!(type->flags & EVP_MD_FLAG_FIPS) 
-			 && !(ctx->flags & EVP_MD_CTX_FLAG_NON_FIPS_ALLOW))
-				{
-				EVPerr(EVP_F_EVP_DIGESTINIT_EX, EVP_R_DISABLED_FOR_FIPS);
-				ctx->digest = &bad_md;
-				return 0;
-				}
-			}
-#endif
 		if (ctx->digest && ctx->digest->ctx_size)
 			OPENSSL_free(ctx->md_data);
 		ctx->digest=type;
-		if (type->ctx_size)
+		if (!(ctx->flags & EVP_MD_CTX_FLAG_NO_INIT) && type->ctx_size)
 			{
+			ctx->update = type->update;
 			ctx->md_data=OPENSSL_malloc(type->ctx_size);
-			if (!ctx->md_data)
+			if (ctx->md_data == NULL)
 				{
-				EVPerr(EVP_F_EVP_DIGESTINIT_EX, ERR_R_MALLOC_FAILURE);
+				EVPerr(EVP_F_EVP_DIGESTINIT_EX,
+							ERR_R_MALLOC_FAILURE);
 				return 0;
 				}
 			}
 		}
 #ifndef OPENSSL_NO_ENGINE
-	skip_to_init:
+skip_to_init:
+#endif
+	if (ctx->pctx)
+		{
+		int r;
+		r = EVP_PKEY_CTX_ctrl(ctx->pctx, -1, EVP_PKEY_OP_TYPE_SIG,
+					EVP_PKEY_CTRL_DIGESTINIT, 0, ctx);
+		if (r <= 0 && (r != -2))
+			return 0;
+		}
+	if (ctx->flags & EVP_MD_CTX_FLAG_NO_INIT)
+		return 1;
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode())
+		{
+		if (FIPS_digestinit(ctx, type))
+			return 1;
+		OPENSSL_free(ctx->md_data);
+		ctx->md_data = NULL;
+		return 0;
+		}
 #endif
 	return ctx->digest->init(ctx);
 	}
 
-int EVP_DigestUpdate(EVP_MD_CTX *ctx, const void *data,
-	     size_t count)
+int EVP_DigestUpdate(EVP_MD_CTX *ctx, const void *data, size_t count)
 	{
 #ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
+	return FIPS_digestupdate(ctx, data, count);
+#else
+	return ctx->update(ctx,data,count);
 #endif
-	return ctx->digest->update(ctx,data,count);
 	}
 
 /* The caller can assume that this removes any secret data from the context */
@@ -337,11 +263,10 @@ int EVP_DigestFinal(EVP_MD_CTX *ctx, unsigned char *md, unsigned int *size)
 /* The caller can assume that this removes any secret data from the context */
 int EVP_DigestFinal_ex(EVP_MD_CTX *ctx, unsigned char *md, unsigned int *size)
 	{
-	int ret;
 #ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
-
+	return FIPS_digestfinal(ctx, md, size);
+#else
+	int ret;
 	OPENSSL_assert(ctx->digest->md_size <= EVP_MAX_MD_SIZE);
 	ret=ctx->digest->final(ctx,md);
 	if (size != NULL)
@@ -349,10 +274,11 @@ int EVP_DigestFinal_ex(EVP_MD_CTX *ctx, unsigned char *md, unsigned int *size)
 	if (ctx->digest->cleanup)
 		{
 		ctx->digest->cleanup(ctx);
-		M_EVP_MD_CTX_set_flags(ctx,EVP_MD_CTX_FLAG_CLEANED);
+		EVP_MD_CTX_set_flags(ctx,EVP_MD_CTX_FLAG_CLEANED);
 		}
 	memset(ctx->md_data,0,ctx->digest->ctx_size);
 	return ret;
+#endif
 	}
 
 int EVP_MD_CTX_copy(EVP_MD_CTX *out, const EVP_MD_CTX *in)
@@ -371,7 +297,7 @@ int EVP_MD_CTX_copy_ex(EVP_MD_CTX *out, const EVP_MD_CTX *in)
 		}
 #ifndef OPENSSL_NO_ENGINE
 	/* Make sure it's safe to copy a digest context using an ENGINE */
-	if (in->engine && !do_engine_init(in->engine))
+	if (in->engine && !ENGINE_init(in->engine))
 		{
 		EVPerr(EVP_F_EVP_MD_CTX_COPY_EX,ERR_R_ENGINE_LIB);
 		return 0;
@@ -381,13 +307,13 @@ int EVP_MD_CTX_copy_ex(EVP_MD_CTX *out, const EVP_MD_CTX *in)
 	if (out->digest == in->digest)
 		{
 		tmp_buf = out->md_data;
-	    	M_EVP_MD_CTX_set_flags(out,EVP_MD_CTX_FLAG_REUSE);
+	    	EVP_MD_CTX_set_flags(out,EVP_MD_CTX_FLAG_REUSE);
 		}
 	else tmp_buf = NULL;
 	EVP_MD_CTX_cleanup(out);
 	memcpy(out,in,sizeof *out);
 
-	if (out->digest->ctx_size)
+	if (in->md_data && out->digest->ctx_size)
 		{
 		if (tmp_buf)
 			out->md_data = tmp_buf;
@@ -403,6 +329,18 @@ int EVP_MD_CTX_copy_ex(EVP_MD_CTX *out, const EVP_MD_CTX *in)
 		memcpy(out->md_data,in->md_data,out->digest->ctx_size);
 		}
 
+	out->update = in->update;
+
+	if (in->pctx)
+		{
+		out->pctx = EVP_PKEY_CTX_dup(in->pctx);
+		if (!out->pctx)
+			{
+			EVP_MD_CTX_cleanup(out);
+			return 0;
+			}
+		}
+
 	if (out->digest->copy)
 		return out->digest->copy(out,in);
 	
@@ -416,7 +354,7 @@ int EVP_Digest(const void *data, size_t count,
 	int ret;
 
 	EVP_MD_CTX_init(&ctx);
-	M_EVP_MD_CTX_set_flags(&ctx,EVP_MD_CTX_FLAG_ONESHOT);
+	EVP_MD_CTX_set_flags(&ctx,EVP_MD_CTX_FLAG_ONESHOT);
 	ret=EVP_DigestInit_ex(&ctx, type, impl)
 	  && EVP_DigestUpdate(&ctx, data, count)
 	  && EVP_DigestFinal_ex(&ctx, md, size);
@@ -434,23 +372,30 @@ void EVP_MD_CTX_destroy(EVP_MD_CTX *ctx)
 /* This call frees resources associated with the context */
 int EVP_MD_CTX_cleanup(EVP_MD_CTX *ctx)
 	{
+#ifndef OPENSSL_FIPS
 	/* Don't assume ctx->md_data was cleaned in EVP_Digest_Final,
 	 * because sometimes only copies of the context are ever finalised.
 	 */
 	if (ctx->digest && ctx->digest->cleanup
-	    && !M_EVP_MD_CTX_test_flags(ctx,EVP_MD_CTX_FLAG_CLEANED))
+	    && !EVP_MD_CTX_test_flags(ctx,EVP_MD_CTX_FLAG_CLEANED))
 		ctx->digest->cleanup(ctx);
 	if (ctx->digest && ctx->digest->ctx_size && ctx->md_data
-	    && !M_EVP_MD_CTX_test_flags(ctx, EVP_MD_CTX_FLAG_REUSE))
+	    && !EVP_MD_CTX_test_flags(ctx, EVP_MD_CTX_FLAG_REUSE))
 		{
 		OPENSSL_cleanse(ctx->md_data,ctx->digest->ctx_size);
 		OPENSSL_free(ctx->md_data);
 		}
+#endif
+	if (ctx->pctx)
+		EVP_PKEY_CTX_free(ctx->pctx);
 #ifndef OPENSSL_NO_ENGINE
 	if(ctx->engine)
 		/* The EVP_MD we used belongs to an ENGINE, release the
 		 * functional reference we held for this reason. */
-		do_engine_finish(ctx->engine);
+		ENGINE_finish(ctx->engine);
+#endif
+#ifdef OPENSSL_FIPS
+	FIPS_md_ctx_cleanup(ctx);
 #endif
 	memset(ctx,'\0',sizeof *ctx);
 
diff --git a/crypto/evp/e_aes.c b/crypto/evp/e_aes.c
index c9a5ee8d75a0..1e4af0cb7511 100644
--- a/crypto/evp/e_aes.c
+++ b/crypto/evp/e_aes.c
@@ -1,5 +1,5 @@
 /* ====================================================================
- * Copyright (c) 2001 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2001-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -56,54 +56,511 @@
 #include <assert.h>
 #include <openssl/aes.h>
 #include "evp_locl.h"
-
-static int aes_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
-					const unsigned char *iv, int enc);
+#ifndef OPENSSL_FIPS
+#include "modes_lcl.h"
+#include <openssl/rand.h>
 
 typedef struct
 	{
 	AES_KEY ks;
+	block128_f block;
+	union {
+		cbc128_f cbc;
+		ctr128_f ctr;
+	} stream;
 	} EVP_AES_KEY;
 
-#define data(ctx)	EVP_C_DATA(EVP_AES_KEY,ctx)
-
-IMPLEMENT_BLOCK_CIPHER(aes_128, ks, AES, EVP_AES_KEY,
-		       NID_aes_128, 16, 16, 16, 128,
-		       EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_DEFAULT_ASN1,
-		       aes_init_key,
-		       NULL, NULL, NULL, NULL)
-IMPLEMENT_BLOCK_CIPHER(aes_192, ks, AES, EVP_AES_KEY,
-		       NID_aes_192, 16, 24, 16, 128,
-		       EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_DEFAULT_ASN1,
-		       aes_init_key,
-		       NULL, NULL, NULL, NULL)
-IMPLEMENT_BLOCK_CIPHER(aes_256, ks, AES, EVP_AES_KEY,
-		       NID_aes_256, 16, 32, 16, 128,
-		       EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_DEFAULT_ASN1,
-		       aes_init_key,
-		       NULL, NULL, NULL, NULL)
-
-#define IMPLEMENT_AES_CFBR(ksize,cbits,flags)	IMPLEMENT_CFBR(aes,AES,EVP_AES_KEY,ks,ksize,cbits,16,flags)
-
-IMPLEMENT_AES_CFBR(128,1,EVP_CIPH_FLAG_FIPS)
-IMPLEMENT_AES_CFBR(192,1,EVP_CIPH_FLAG_FIPS)
-IMPLEMENT_AES_CFBR(256,1,EVP_CIPH_FLAG_FIPS)
-
-IMPLEMENT_AES_CFBR(128,8,EVP_CIPH_FLAG_FIPS)
-IMPLEMENT_AES_CFBR(192,8,EVP_CIPH_FLAG_FIPS)
-IMPLEMENT_AES_CFBR(256,8,EVP_CIPH_FLAG_FIPS)
+typedef struct
+	{
+	AES_KEY ks;		/* AES key schedule to use */
+	int key_set;		/* Set if key initialised */
+	int iv_set;		/* Set if an iv is set */
+	GCM128_CONTEXT gcm;
+	unsigned char *iv;	/* Temporary IV store */
+	int ivlen;		/* IV length */
+	int taglen;
+	int iv_gen;		/* It is OK to generate IVs */
+	int tls_aad_len;	/* TLS AAD length */
+	ctr128_f ctr;
+	} EVP_AES_GCM_CTX;
+
+typedef struct
+	{
+	AES_KEY ks1, ks2;	/* AES key schedules to use */
+	XTS128_CONTEXT xts;
+	void     (*stream)(const unsigned char *in,
+			unsigned char *out, size_t length,
+			const AES_KEY *key1, const AES_KEY *key2,
+			const unsigned char iv[16]);
+	} EVP_AES_XTS_CTX;
+
+typedef struct
+	{
+	AES_KEY ks;		/* AES key schedule to use */
+	int key_set;		/* Set if key initialised */
+	int iv_set;		/* Set if an iv is set */
+	int tag_set;		/* Set if tag is valid */
+	int len_set;		/* Set if message length set */
+	int L, M;		/* L and M parameters from RFC3610 */
+	CCM128_CONTEXT ccm;
+	ccm128_f str;
+	} EVP_AES_CCM_CTX;
+
+#define MAXBITCHUNK	((size_t)1<<(sizeof(size_t)*8-4))
+
+#ifdef VPAES_ASM
+int vpaes_set_encrypt_key(const unsigned char *userKey, int bits,
+			AES_KEY *key);
+int vpaes_set_decrypt_key(const unsigned char *userKey, int bits,
+			AES_KEY *key);
+
+void vpaes_encrypt(const unsigned char *in, unsigned char *out,
+			const AES_KEY *key);
+void vpaes_decrypt(const unsigned char *in, unsigned char *out,
+			const AES_KEY *key);
+
+void vpaes_cbc_encrypt(const unsigned char *in,
+			unsigned char *out,
+			size_t length,
+			const AES_KEY *key,
+			unsigned char *ivec, int enc);
+#endif
+#ifdef BSAES_ASM
+void bsaes_cbc_encrypt(const unsigned char *in, unsigned char *out,
+			size_t length, const AES_KEY *key,
+			unsigned char ivec[16], int enc);
+void bsaes_ctr32_encrypt_blocks(const unsigned char *in, unsigned char *out,
+			size_t len, const AES_KEY *key,
+			const unsigned char ivec[16]);
+void bsaes_xts_encrypt(const unsigned char *inp, unsigned char *out,
+			size_t len, const AES_KEY *key1,
+			const AES_KEY *key2, const unsigned char iv[16]);
+void bsaes_xts_decrypt(const unsigned char *inp, unsigned char *out,
+			size_t len, const AES_KEY *key1,
+			const AES_KEY *key2, const unsigned char iv[16]);
+#endif
+#ifdef AES_CTR_ASM
+void AES_ctr32_encrypt(const unsigned char *in, unsigned char *out,
+			size_t blocks, const AES_KEY *key,
+			const unsigned char ivec[AES_BLOCK_SIZE]);
+#endif
+#ifdef AES_XTS_ASM
+void AES_xts_encrypt(const char *inp,char *out,size_t len,
+			const AES_KEY *key1, const AES_KEY *key2,
+			const unsigned char iv[16]);
+void AES_xts_decrypt(const char *inp,char *out,size_t len,
+			const AES_KEY *key1, const AES_KEY *key2,
+			const unsigned char iv[16]);
+#endif
+
+#if	defined(AES_ASM) && !defined(I386_ONLY) &&	(  \
+	((defined(__i386)	|| defined(__i386__)	|| \
+	  defined(_M_IX86)) && defined(OPENSSL_IA32_SSE2))|| \
+	defined(__x86_64)	|| defined(__x86_64__)	|| \
+	defined(_M_AMD64)	|| defined(_M_X64)	|| \
+	defined(__INTEL__)				)
+
+extern unsigned int OPENSSL_ia32cap_P[2];
+
+#ifdef VPAES_ASM
+#define VPAES_CAPABLE	(OPENSSL_ia32cap_P[1]&(1<<(41-32)))
+#endif
+#ifdef BSAES_ASM
+#define BSAES_CAPABLE	VPAES_CAPABLE
+#endif
+/*
+ * AES-NI section
+ */
+#define	AESNI_CAPABLE	(OPENSSL_ia32cap_P[1]&(1<<(57-32)))
+
+int aesni_set_encrypt_key(const unsigned char *userKey, int bits,
+			AES_KEY *key);
+int aesni_set_decrypt_key(const unsigned char *userKey, int bits,
+			AES_KEY *key);
+
+void aesni_encrypt(const unsigned char *in, unsigned char *out,
+			const AES_KEY *key);
+void aesni_decrypt(const unsigned char *in, unsigned char *out,
+			const AES_KEY *key);
+
+void aesni_ecb_encrypt(const unsigned char *in,
+			unsigned char *out,
+			size_t length,
+			const AES_KEY *key,
+			int enc);
+void aesni_cbc_encrypt(const unsigned char *in,
+			unsigned char *out,
+			size_t length,
+			const AES_KEY *key,
+			unsigned char *ivec, int enc);
+
+void aesni_ctr32_encrypt_blocks(const unsigned char *in,
+			unsigned char *out,
+			size_t blocks,
+			const void *key,
+			const unsigned char *ivec);
+
+void aesni_xts_encrypt(const unsigned char *in,
+			unsigned char *out,
+			size_t length,
+			const AES_KEY *key1, const AES_KEY *key2,
+			const unsigned char iv[16]);
+
+void aesni_xts_decrypt(const unsigned char *in,
+			unsigned char *out,
+			size_t length,
+			const AES_KEY *key1, const AES_KEY *key2,
+			const unsigned char iv[16]);
+
+void aesni_ccm64_encrypt_blocks (const unsigned char *in,
+			unsigned char *out,
+			size_t blocks,
+			const void *key,
+			const unsigned char ivec[16],
+			unsigned char cmac[16]);
+
+void aesni_ccm64_decrypt_blocks (const unsigned char *in,
+			unsigned char *out,
+			size_t blocks,
+			const void *key,
+			const unsigned char ivec[16],
+			unsigned char cmac[16]);
+
+static int aesni_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
+		   const unsigned char *iv, int enc)
+	{
+	int ret, mode;
+	EVP_AES_KEY *dat = (EVP_AES_KEY *)ctx->cipher_data;
+
+	mode = ctx->cipher->flags & EVP_CIPH_MODE;
+	if ((mode == EVP_CIPH_ECB_MODE || mode == EVP_CIPH_CBC_MODE)
+	    && !enc)
+		{ 
+		ret = aesni_set_decrypt_key(key, ctx->key_len*8, ctx->cipher_data);
+		dat->block	= (block128_f)aesni_decrypt;
+		dat->stream.cbc	= mode==EVP_CIPH_CBC_MODE ?
+					(cbc128_f)aesni_cbc_encrypt :
+					NULL;
+		}
+	else	{
+		ret = aesni_set_encrypt_key(key, ctx->key_len*8, ctx->cipher_data);
+		dat->block	= (block128_f)aesni_encrypt;
+		if (mode==EVP_CIPH_CBC_MODE)
+			dat->stream.cbc	= (cbc128_f)aesni_cbc_encrypt;
+		else if (mode==EVP_CIPH_CTR_MODE)
+			dat->stream.ctr = (ctr128_f)aesni_ctr32_encrypt_blocks;
+		else
+			dat->stream.cbc = NULL;
+		}
+
+	if(ret < 0)
+		{
+		EVPerr(EVP_F_AESNI_INIT_KEY,EVP_R_AES_KEY_SETUP_FAILED);
+		return 0;
+		}
+
+	return 1;
+	}
+
+static int aesni_cbc_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in, size_t len)
+{
+	aesni_cbc_encrypt(in,out,len,ctx->cipher_data,ctx->iv,ctx->encrypt);
+
+	return 1;
+}
+
+static int aesni_ecb_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in, size_t len)
+{
+	size_t	bl = ctx->cipher->block_size;
+
+	if (len<bl)	return 1;
+
+	aesni_ecb_encrypt(in,out,len,ctx->cipher_data,ctx->encrypt);
+
+	return 1;
+}
+
+#define aesni_ofb_cipher aes_ofb_cipher
+static int aesni_ofb_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in,size_t len);
+
+#define aesni_cfb_cipher aes_cfb_cipher
+static int aesni_cfb_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in,size_t len);
+
+#define aesni_cfb8_cipher aes_cfb8_cipher
+static int aesni_cfb8_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in,size_t len);
+
+#define aesni_cfb1_cipher aes_cfb1_cipher
+static int aesni_cfb1_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in,size_t len);
+
+#define aesni_ctr_cipher aes_ctr_cipher
+static int aesni_ctr_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		const unsigned char *in, size_t len);
+
+static int aesni_gcm_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
+                        const unsigned char *iv, int enc)
+	{
+	EVP_AES_GCM_CTX *gctx = ctx->cipher_data;
+	if (!iv && !key)
+		return 1;
+	if (key)
+		{
+		aesni_set_encrypt_key(key, ctx->key_len * 8, &gctx->ks);
+		CRYPTO_gcm128_init(&gctx->gcm, &gctx->ks,
+				(block128_f)aesni_encrypt);
+		gctx->ctr = (ctr128_f)aesni_ctr32_encrypt_blocks;
+		/* If we have an iv can set it directly, otherwise use
+		 * saved IV.
+		 */
+		if (iv == NULL && gctx->iv_set)
+			iv = gctx->iv;
+		if (iv)
+			{
+			CRYPTO_gcm128_setiv(&gctx->gcm, iv, gctx->ivlen);
+			gctx->iv_set = 1;
+			}
+		gctx->key_set = 1;
+		}
+	else
+		{
+		/* If key set use IV, otherwise copy */
+		if (gctx->key_set)
+			CRYPTO_gcm128_setiv(&gctx->gcm, iv, gctx->ivlen);
+		else
+			memcpy(gctx->iv, iv, gctx->ivlen);
+		gctx->iv_set = 1;
+		gctx->iv_gen = 0;
+		}
+	return 1;
+	}
+
+#define aesni_gcm_cipher aes_gcm_cipher
+static int aesni_gcm_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		const unsigned char *in, size_t len);
+
+static int aesni_xts_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
+                        const unsigned char *iv, int enc)
+	{
+	EVP_AES_XTS_CTX *xctx = ctx->cipher_data;
+	if (!iv && !key)
+		return 1;
+
+	if (key)
+		{
+		/* key_len is two AES keys */
+		if (enc)
+			{
+			aesni_set_encrypt_key(key, ctx->key_len * 4, &xctx->ks1);
+			xctx->xts.block1 = (block128_f)aesni_encrypt;
+			xctx->stream = aesni_xts_encrypt;
+			}
+		else
+			{
+			aesni_set_decrypt_key(key, ctx->key_len * 4, &xctx->ks1);
+			xctx->xts.block1 = (block128_f)aesni_decrypt;
+			xctx->stream = aesni_xts_decrypt;
+			}
+
+		aesni_set_encrypt_key(key + ctx->key_len/2,
+						ctx->key_len * 4, &xctx->ks2);
+		xctx->xts.block2 = (block128_f)aesni_encrypt;
+
+		xctx->xts.key1 = &xctx->ks1;
+		}
+
+	if (iv)
+		{
+		xctx->xts.key2 = &xctx->ks2;
+		memcpy(ctx->iv, iv, 16);
+		}
+
+	return 1;
+	}
+
+#define aesni_xts_cipher aes_xts_cipher
+static int aesni_xts_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		const unsigned char *in, size_t len);
+
+static int aesni_ccm_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
+                        const unsigned char *iv, int enc)
+	{
+	EVP_AES_CCM_CTX *cctx = ctx->cipher_data;
+	if (!iv && !key)
+		return 1;
+	if (key)
+		{
+		aesni_set_encrypt_key(key, ctx->key_len * 8, &cctx->ks);
+		CRYPTO_ccm128_init(&cctx->ccm, cctx->M, cctx->L,
+					&cctx->ks, (block128_f)aesni_encrypt);
+		cctx->str = enc?(ccm128_f)aesni_ccm64_encrypt_blocks :
+				(ccm128_f)aesni_ccm64_decrypt_blocks;
+		cctx->key_set = 1;
+		}
+	if (iv)
+		{
+		memcpy(ctx->iv, iv, 15 - cctx->L);
+		cctx->iv_set = 1;
+		}
+	return 1;
+	}
+
+#define aesni_ccm_cipher aes_ccm_cipher
+static int aesni_ccm_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		const unsigned char *in, size_t len);
+
+#define BLOCK_CIPHER_generic(nid,keylen,blocksize,ivlen,nmode,mode,MODE,flags) \
+static const EVP_CIPHER aesni_##keylen##_##mode = { \
+	nid##_##keylen##_##nmode,blocksize,keylen/8,ivlen, \
+	flags|EVP_CIPH_##MODE##_MODE,	\
+	aesni_init_key,			\
+	aesni_##mode##_cipher,		\
+	NULL,				\
+	sizeof(EVP_AES_KEY),		\
+	NULL,NULL,NULL,NULL }; \
+static const EVP_CIPHER aes_##keylen##_##mode = { \
+	nid##_##keylen##_##nmode,blocksize,	\
+	keylen/8,ivlen, \
+	flags|EVP_CIPH_##MODE##_MODE,	\
+	aes_init_key,			\
+	aes_##mode##_cipher,		\
+	NULL,				\
+	sizeof(EVP_AES_KEY),		\
+	NULL,NULL,NULL,NULL }; \
+const EVP_CIPHER *EVP_aes_##keylen##_##mode(void) \
+{ return AESNI_CAPABLE?&aesni_##keylen##_##mode:&aes_##keylen##_##mode; }
+
+#define BLOCK_CIPHER_custom(nid,keylen,blocksize,ivlen,mode,MODE,flags) \
+static const EVP_CIPHER aesni_##keylen##_##mode = { \
+	nid##_##keylen##_##mode,blocksize, \
+	(EVP_CIPH_##MODE##_MODE==EVP_CIPH_XTS_MODE?2:1)*keylen/8, ivlen, \
+	flags|EVP_CIPH_##MODE##_MODE,	\
+	aesni_##mode##_init_key,	\
+	aesni_##mode##_cipher,		\
+	aes_##mode##_cleanup,		\
+	sizeof(EVP_AES_##MODE##_CTX),	\
+	NULL,NULL,aes_##mode##_ctrl,NULL }; \
+static const EVP_CIPHER aes_##keylen##_##mode = { \
+	nid##_##keylen##_##mode,blocksize, \
+	(EVP_CIPH_##MODE##_MODE==EVP_CIPH_XTS_MODE?2:1)*keylen/8, ivlen, \
+	flags|EVP_CIPH_##MODE##_MODE,	\
+	aes_##mode##_init_key,		\
+	aes_##mode##_cipher,		\
+	aes_##mode##_cleanup,		\
+	sizeof(EVP_AES_##MODE##_CTX),	\
+	NULL,NULL,aes_##mode##_ctrl,NULL }; \
+const EVP_CIPHER *EVP_aes_##keylen##_##mode(void) \
+{ return AESNI_CAPABLE?&aesni_##keylen##_##mode:&aes_##keylen##_##mode; }
+
+#else
+
+#define BLOCK_CIPHER_generic(nid,keylen,blocksize,ivlen,nmode,mode,MODE,flags) \
+static const EVP_CIPHER aes_##keylen##_##mode = { \
+	nid##_##keylen##_##nmode,blocksize,keylen/8,ivlen, \
+	flags|EVP_CIPH_##MODE##_MODE,	\
+	aes_init_key,			\
+	aes_##mode##_cipher,		\
+	NULL,				\
+	sizeof(EVP_AES_KEY),		\
+	NULL,NULL,NULL,NULL }; \
+const EVP_CIPHER *EVP_aes_##keylen##_##mode(void) \
+{ return &aes_##keylen##_##mode; }
+
+#define BLOCK_CIPHER_custom(nid,keylen,blocksize,ivlen,mode,MODE,flags) \
+static const EVP_CIPHER aes_##keylen##_##mode = { \
+	nid##_##keylen##_##mode,blocksize, \
+	(EVP_CIPH_##MODE##_MODE==EVP_CIPH_XTS_MODE?2:1)*keylen/8, ivlen, \
+	flags|EVP_CIPH_##MODE##_MODE,	\
+	aes_##mode##_init_key,		\
+	aes_##mode##_cipher,		\
+	aes_##mode##_cleanup,		\
+	sizeof(EVP_AES_##MODE##_CTX),	\
+	NULL,NULL,aes_##mode##_ctrl,NULL }; \
+const EVP_CIPHER *EVP_aes_##keylen##_##mode(void) \
+{ return &aes_##keylen##_##mode; }
+#endif
+
+#define BLOCK_CIPHER_generic_pack(nid,keylen,flags)		\
+	BLOCK_CIPHER_generic(nid,keylen,16,16,cbc,cbc,CBC,flags|EVP_CIPH_FLAG_DEFAULT_ASN1)	\
+	BLOCK_CIPHER_generic(nid,keylen,16,0,ecb,ecb,ECB,flags|EVP_CIPH_FLAG_DEFAULT_ASN1)	\
+	BLOCK_CIPHER_generic(nid,keylen,1,16,ofb128,ofb,OFB,flags|EVP_CIPH_FLAG_DEFAULT_ASN1)	\
+	BLOCK_CIPHER_generic(nid,keylen,1,16,cfb128,cfb,CFB,flags|EVP_CIPH_FLAG_DEFAULT_ASN1)	\
+	BLOCK_CIPHER_generic(nid,keylen,1,16,cfb1,cfb1,CFB,flags)	\
+	BLOCK_CIPHER_generic(nid,keylen,1,16,cfb8,cfb8,CFB,flags)	\
+	BLOCK_CIPHER_generic(nid,keylen,1,16,ctr,ctr,CTR,flags)
 
 static int aes_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 		   const unsigned char *iv, int enc)
 	{
-	int ret;
+	int ret, mode;
+	EVP_AES_KEY *dat = (EVP_AES_KEY *)ctx->cipher_data;
 
-	if ((ctx->cipher->flags & EVP_CIPH_MODE) == EVP_CIPH_CFB_MODE
-	    || (ctx->cipher->flags & EVP_CIPH_MODE) == EVP_CIPH_OFB_MODE
-	    || enc) 
-		ret=AES_set_encrypt_key(key, ctx->key_len * 8, ctx->cipher_data);
+	mode = ctx->cipher->flags & EVP_CIPH_MODE;
+	if ((mode == EVP_CIPH_ECB_MODE || mode == EVP_CIPH_CBC_MODE)
+	    && !enc)
+#ifdef BSAES_CAPABLE
+	    if (BSAES_CAPABLE && mode==EVP_CIPH_CBC_MODE)
+		{
+		ret = AES_set_decrypt_key(key,ctx->key_len*8,&dat->ks);
+		dat->block	= (block128_f)AES_decrypt;
+		dat->stream.cbc	= (cbc128_f)bsaes_cbc_encrypt;
+		}
+	    else
+#endif
+#ifdef VPAES_CAPABLE
+	    if (VPAES_CAPABLE)
+		{
+		ret = vpaes_set_decrypt_key(key,ctx->key_len*8,&dat->ks);
+		dat->block	= (block128_f)vpaes_decrypt;
+		dat->stream.cbc	= mode==EVP_CIPH_CBC_MODE ?
+					(cbc128_f)vpaes_cbc_encrypt :
+					NULL;
+		}
+	    else
+#endif
+		{
+		ret = AES_set_decrypt_key(key,ctx->key_len*8,&dat->ks);
+		dat->block	= (block128_f)AES_decrypt;
+		dat->stream.cbc	= mode==EVP_CIPH_CBC_MODE ?
+					(cbc128_f)AES_cbc_encrypt :
+					NULL;
+		}
 	else
-		ret=AES_set_decrypt_key(key, ctx->key_len * 8, ctx->cipher_data);
+#ifdef BSAES_CAPABLE
+	    if (BSAES_CAPABLE && mode==EVP_CIPH_CTR_MODE)
+		{
+		ret = AES_set_encrypt_key(key,ctx->key_len*8,&dat->ks);
+		dat->block	= (block128_f)AES_encrypt;
+		dat->stream.ctr	= (ctr128_f)bsaes_ctr32_encrypt_blocks;
+		}
+	    else
+#endif
+#ifdef VPAES_CAPABLE
+	    if (VPAES_CAPABLE)
+		{
+		ret = vpaes_set_encrypt_key(key,ctx->key_len*8,&dat->ks);
+		dat->block	= (block128_f)vpaes_encrypt;
+		dat->stream.cbc	= mode==EVP_CIPH_CBC_MODE ?
+					(cbc128_f)vpaes_cbc_encrypt :
+					NULL;
+		}
+	    else
+#endif
+		{
+		ret = AES_set_encrypt_key(key,ctx->key_len*8,&dat->ks);
+		dat->block	= (block128_f)AES_encrypt;
+		dat->stream.cbc	= mode==EVP_CIPH_CBC_MODE ?
+					(cbc128_f)AES_cbc_encrypt :
+					NULL;
+#ifdef AES_CTR_ASM
+		if (mode==EVP_CIPH_CTR_MODE)
+			dat->stream.ctr = (ctr128_f)AES_ctr32_encrypt;
+#endif
+		}
 
 	if(ret < 0)
 		{
@@ -114,4 +571,743 @@ static int aes_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 	return 1;
 	}
 
+static int aes_cbc_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in, size_t len)
+{
+	EVP_AES_KEY *dat = (EVP_AES_KEY *)ctx->cipher_data;
+
+	if (dat->stream.cbc)
+		(*dat->stream.cbc)(in,out,len,&dat->ks,ctx->iv,ctx->encrypt);
+	else if (ctx->encrypt)
+		CRYPTO_cbc128_encrypt(in,out,len,&dat->ks,ctx->iv,dat->block);
+	else
+		CRYPTO_cbc128_encrypt(in,out,len,&dat->ks,ctx->iv,dat->block);
+
+	return 1;
+}
+
+static int aes_ecb_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in, size_t len)
+{
+	size_t	bl = ctx->cipher->block_size;
+	size_t	i;
+	EVP_AES_KEY *dat = (EVP_AES_KEY *)ctx->cipher_data;
+
+	if (len<bl)	return 1;
+
+	for (i=0,len-=bl;i<=len;i+=bl)
+		(*dat->block)(in+i,out+i,&dat->ks);
+
+	return 1;
+}
+
+static int aes_ofb_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in,size_t len)
+{
+	EVP_AES_KEY *dat = (EVP_AES_KEY *)ctx->cipher_data;
+
+	CRYPTO_ofb128_encrypt(in,out,len,&dat->ks,
+			ctx->iv,&ctx->num,dat->block);
+	return 1;
+}
+
+static int aes_cfb_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in,size_t len)
+{
+	EVP_AES_KEY *dat = (EVP_AES_KEY *)ctx->cipher_data;
+
+	CRYPTO_cfb128_encrypt(in,out,len,&dat->ks,
+			ctx->iv,&ctx->num,ctx->encrypt,dat->block);
+	return 1;
+}
+
+static int aes_cfb8_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in,size_t len)
+{
+	EVP_AES_KEY *dat = (EVP_AES_KEY *)ctx->cipher_data;
+
+	CRYPTO_cfb128_8_encrypt(in,out,len,&dat->ks,
+			ctx->iv,&ctx->num,ctx->encrypt,dat->block);
+	return 1;
+}
+
+static int aes_cfb1_cipher(EVP_CIPHER_CTX *ctx,unsigned char *out,
+	const unsigned char *in,size_t len)
+{
+	EVP_AES_KEY *dat = (EVP_AES_KEY *)ctx->cipher_data;
+
+	if (ctx->flags&EVP_CIPH_FLAG_LENGTH_BITS) {
+		CRYPTO_cfb128_1_encrypt(in,out,len,&dat->ks,
+			ctx->iv,&ctx->num,ctx->encrypt,dat->block);
+		return 1;
+	}
+
+	while (len>=MAXBITCHUNK) {
+		CRYPTO_cfb128_1_encrypt(in,out,MAXBITCHUNK*8,&dat->ks,
+			ctx->iv,&ctx->num,ctx->encrypt,dat->block);
+		len-=MAXBITCHUNK;
+	}
+	if (len)
+		CRYPTO_cfb128_1_encrypt(in,out,len*8,&dat->ks,
+			ctx->iv,&ctx->num,ctx->encrypt,dat->block);
+	
+	return 1;
+}
+
+static int aes_ctr_cipher (EVP_CIPHER_CTX *ctx, unsigned char *out,
+		const unsigned char *in, size_t len)
+{
+	unsigned int num = ctx->num;
+	EVP_AES_KEY *dat = (EVP_AES_KEY *)ctx->cipher_data;
+
+	if (dat->stream.ctr)
+		CRYPTO_ctr128_encrypt_ctr32(in,out,len,&dat->ks,
+			ctx->iv,ctx->buf,&num,dat->stream.ctr);
+	else
+		CRYPTO_ctr128_encrypt(in,out,len,&dat->ks,
+			ctx->iv,ctx->buf,&num,dat->block);
+	ctx->num = (size_t)num;
+	return 1;
+}
+
+BLOCK_CIPHER_generic_pack(NID_aes,128,EVP_CIPH_FLAG_FIPS)
+BLOCK_CIPHER_generic_pack(NID_aes,192,EVP_CIPH_FLAG_FIPS)
+BLOCK_CIPHER_generic_pack(NID_aes,256,EVP_CIPH_FLAG_FIPS)
+
+static int aes_gcm_cleanup(EVP_CIPHER_CTX *c)
+	{
+	EVP_AES_GCM_CTX *gctx = c->cipher_data;
+	OPENSSL_cleanse(&gctx->gcm, sizeof(gctx->gcm));
+	if (gctx->iv != c->iv)
+		OPENSSL_free(gctx->iv);
+	return 1;
+	}
+
+/* increment counter (64-bit int) by 1 */
+static void ctr64_inc(unsigned char *counter) {
+	int n=8;
+	unsigned char  c;
+
+	do {
+		--n;
+		c = counter[n];
+		++c;
+		counter[n] = c;
+		if (c) return;
+	} while (n);
+}
+
+static int aes_gcm_ctrl(EVP_CIPHER_CTX *c, int type, int arg, void *ptr)
+	{
+	EVP_AES_GCM_CTX *gctx = c->cipher_data;
+	switch (type)
+		{
+	case EVP_CTRL_INIT:
+		gctx->key_set = 0;
+		gctx->iv_set = 0;
+		gctx->ivlen = c->cipher->iv_len;
+		gctx->iv = c->iv;
+		gctx->taglen = -1;
+		gctx->iv_gen = 0;
+		gctx->tls_aad_len = -1;
+		return 1;
+
+	case EVP_CTRL_GCM_SET_IVLEN:
+		if (arg <= 0)
+			return 0;
+#ifdef OPENSSL_FIPS
+		if (FIPS_module_mode() && !(c->flags & EVP_CIPH_FLAG_NON_FIPS_ALLOW)
+						 && arg < 12)
+			return 0;
+#endif
+		/* Allocate memory for IV if needed */
+		if ((arg > EVP_MAX_IV_LENGTH) && (arg > gctx->ivlen))
+			{
+			if (gctx->iv != c->iv)
+				OPENSSL_free(gctx->iv);
+			gctx->iv = OPENSSL_malloc(arg);
+			if (!gctx->iv)
+				return 0;
+			}
+		gctx->ivlen = arg;
+		return 1;
+
+	case EVP_CTRL_GCM_SET_TAG:
+		if (arg <= 0 || arg > 16 || c->encrypt)
+			return 0;
+		memcpy(c->buf, ptr, arg);
+		gctx->taglen = arg;
+		return 1;
+
+	case EVP_CTRL_GCM_GET_TAG:
+		if (arg <= 0 || arg > 16 || !c->encrypt || gctx->taglen < 0)
+			return 0;
+		memcpy(ptr, c->buf, arg);
+		return 1;
+
+	case EVP_CTRL_GCM_SET_IV_FIXED:
+		/* Special case: -1 length restores whole IV */
+		if (arg == -1)
+			{
+			memcpy(gctx->iv, ptr, gctx->ivlen);
+			gctx->iv_gen = 1;
+			return 1;
+			}
+		/* Fixed field must be at least 4 bytes and invocation field
+		 * at least 8.
+		 */
+		if ((arg < 4) || (gctx->ivlen - arg) < 8)
+			return 0;
+		if (arg)
+			memcpy(gctx->iv, ptr, arg);
+		if (c->encrypt &&
+			RAND_bytes(gctx->iv + arg, gctx->ivlen - arg) <= 0)
+			return 0;
+		gctx->iv_gen = 1;
+		return 1;
+
+	case EVP_CTRL_GCM_IV_GEN:
+		if (gctx->iv_gen == 0 || gctx->key_set == 0)
+			return 0;
+		CRYPTO_gcm128_setiv(&gctx->gcm, gctx->iv, gctx->ivlen);
+		if (arg <= 0 || arg > gctx->ivlen)
+			arg = gctx->ivlen;
+		memcpy(ptr, gctx->iv + gctx->ivlen - arg, arg);
+		/* Invocation field will be at least 8 bytes in size and
+		 * so no need to check wrap around or increment more than
+		 * last 8 bytes.
+		 */
+		ctr64_inc(gctx->iv + gctx->ivlen - 8);
+		gctx->iv_set = 1;
+		return 1;
+
+	case EVP_CTRL_GCM_SET_IV_INV:
+		if (gctx->iv_gen == 0 || gctx->key_set == 0 || c->encrypt)
+			return 0;
+		memcpy(gctx->iv + gctx->ivlen - arg, ptr, arg);
+		CRYPTO_gcm128_setiv(&gctx->gcm, gctx->iv, gctx->ivlen);
+		gctx->iv_set = 1;
+		return 1;
+
+	case EVP_CTRL_AEAD_TLS1_AAD:
+		/* Save the AAD for later use */
+		if (arg != 13)
+			return 0;
+		memcpy(c->buf, ptr, arg);
+		gctx->tls_aad_len = arg;
+			{
+			unsigned int len=c->buf[arg-2]<<8|c->buf[arg-1];
+			/* Correct length for explicit IV */
+			len -= EVP_GCM_TLS_EXPLICIT_IV_LEN;
+			/* If decrypting correct for tag too */
+			if (!c->encrypt)
+				len -= EVP_GCM_TLS_TAG_LEN;
+                        c->buf[arg-2] = len>>8;
+                        c->buf[arg-1] = len & 0xff;
+			}
+		/* Extra padding: tag appended to record */
+		return EVP_GCM_TLS_TAG_LEN;
+
+	default:
+		return -1;
+
+		}
+	}
+
+static int aes_gcm_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
+                        const unsigned char *iv, int enc)
+	{
+	EVP_AES_GCM_CTX *gctx = ctx->cipher_data;
+	if (!iv && !key)
+		return 1;
+	if (key)
+		{ do {
+#ifdef BSAES_CAPABLE
+		if (BSAES_CAPABLE)
+			{
+			AES_set_encrypt_key(key,ctx->key_len*8,&gctx->ks);
+			CRYPTO_gcm128_init(&gctx->gcm,&gctx->ks,
+					(block128_f)AES_encrypt);
+			gctx->ctr = (ctr128_f)bsaes_ctr32_encrypt_blocks;
+			break;
+			}
+		else
+#endif
+#ifdef VPAES_CAPABLE
+		if (VPAES_CAPABLE)
+			{
+			vpaes_set_encrypt_key(key,ctx->key_len*8,&gctx->ks);
+			CRYPTO_gcm128_init(&gctx->gcm,&gctx->ks,
+					(block128_f)vpaes_encrypt);
+			gctx->ctr = NULL;
+			break;
+			}
+#endif
+		AES_set_encrypt_key(key, ctx->key_len * 8, &gctx->ks);
+		CRYPTO_gcm128_init(&gctx->gcm, &gctx->ks, (block128_f)AES_encrypt);
+#ifdef AES_CTR_ASM
+		gctx->ctr = (ctr128_f)AES_ctr32_encrypt;
+#else
+		gctx->ctr = NULL;
+#endif
+		} while (0);
+
+		/* If we have an iv can set it directly, otherwise use
+		 * saved IV.
+		 */
+		if (iv == NULL && gctx->iv_set)
+			iv = gctx->iv;
+		if (iv)
+			{
+			CRYPTO_gcm128_setiv(&gctx->gcm, iv, gctx->ivlen);
+			gctx->iv_set = 1;
+			}
+		gctx->key_set = 1;
+		}
+	else
+		{
+		/* If key set use IV, otherwise copy */
+		if (gctx->key_set)
+			CRYPTO_gcm128_setiv(&gctx->gcm, iv, gctx->ivlen);
+		else
+			memcpy(gctx->iv, iv, gctx->ivlen);
+		gctx->iv_set = 1;
+		gctx->iv_gen = 0;
+		}
+	return 1;
+	}
+
+/* Handle TLS GCM packet format. This consists of the last portion of the IV
+ * followed by the payload and finally the tag. On encrypt generate IV,
+ * encrypt payload and write the tag. On verify retrieve IV, decrypt payload
+ * and verify tag.
+ */
+
+static int aes_gcm_tls_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		const unsigned char *in, size_t len)
+	{
+	EVP_AES_GCM_CTX *gctx = ctx->cipher_data;
+	int rv = -1;
+	/* Encrypt/decrypt must be performed in place */
+	if (out != in || len < (EVP_GCM_TLS_EXPLICIT_IV_LEN+EVP_GCM_TLS_TAG_LEN))
+		return -1;
+	/* Set IV from start of buffer or generate IV and write to start
+	 * of buffer.
+	 */
+	if (EVP_CIPHER_CTX_ctrl(ctx, ctx->encrypt ?
+				EVP_CTRL_GCM_IV_GEN : EVP_CTRL_GCM_SET_IV_INV,
+				EVP_GCM_TLS_EXPLICIT_IV_LEN, out) <= 0)
+		goto err;
+	/* Use saved AAD */
+	if (CRYPTO_gcm128_aad(&gctx->gcm, ctx->buf, gctx->tls_aad_len))
+		goto err;
+	/* Fix buffer and length to point to payload */
+	in += EVP_GCM_TLS_EXPLICIT_IV_LEN;
+	out += EVP_GCM_TLS_EXPLICIT_IV_LEN;
+	len -= EVP_GCM_TLS_EXPLICIT_IV_LEN + EVP_GCM_TLS_TAG_LEN;
+	if (ctx->encrypt)
+		{
+		/* Encrypt payload */
+		if (gctx->ctr)
+			{
+			if (CRYPTO_gcm128_encrypt_ctr32(&gctx->gcm,
+							in, out, len,
+							gctx->ctr))
+				goto err;
+			}
+		else	{
+			if (CRYPTO_gcm128_encrypt(&gctx->gcm, in, out, len))
+				goto err;
+			}
+		out += len;
+		/* Finally write tag */
+		CRYPTO_gcm128_tag(&gctx->gcm, out, EVP_GCM_TLS_TAG_LEN);
+		rv = len + EVP_GCM_TLS_EXPLICIT_IV_LEN + EVP_GCM_TLS_TAG_LEN;
+		}
+	else
+		{
+		/* Decrypt */
+		if (gctx->ctr)
+			{
+			if (CRYPTO_gcm128_decrypt_ctr32(&gctx->gcm,
+							in, out, len,
+							gctx->ctr))
+				goto err;
+			}
+		else	{
+			if (CRYPTO_gcm128_decrypt(&gctx->gcm, in, out, len))
+				goto err;
+			}
+		/* Retrieve tag */
+		CRYPTO_gcm128_tag(&gctx->gcm, ctx->buf,
+					EVP_GCM_TLS_TAG_LEN);
+		/* If tag mismatch wipe buffer */
+		if (memcmp(ctx->buf, in + len, EVP_GCM_TLS_TAG_LEN))
+			{
+			OPENSSL_cleanse(out, len);
+			goto err;
+			}
+		rv = len;
+		}
+
+	err:
+	gctx->iv_set = 0;
+	gctx->tls_aad_len = -1;
+	return rv;
+	}
+
+static int aes_gcm_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		const unsigned char *in, size_t len)
+	{
+	EVP_AES_GCM_CTX *gctx = ctx->cipher_data;
+	/* If not set up, return error */
+	if (!gctx->key_set)
+		return -1;
+
+	if (gctx->tls_aad_len >= 0)
+		return aes_gcm_tls_cipher(ctx, out, in, len);
+
+	if (!gctx->iv_set)
+		return -1;
+	if (!ctx->encrypt && gctx->taglen < 0)
+		return -1;
+	if (in)
+		{
+		if (out == NULL)
+			{
+			if (CRYPTO_gcm128_aad(&gctx->gcm, in, len))
+				return -1;
+			}
+		else if (ctx->encrypt)
+			{
+			if (gctx->ctr)
+				{
+				if (CRYPTO_gcm128_encrypt_ctr32(&gctx->gcm,
+							in, out, len,
+							gctx->ctr))
+					return -1;
+				}
+			else	{
+				if (CRYPTO_gcm128_encrypt(&gctx->gcm, in, out, len))
+					return -1;
+				}
+			}
+		else
+			{
+			if (gctx->ctr)
+				{
+				if (CRYPTO_gcm128_decrypt_ctr32(&gctx->gcm,
+							in, out, len,
+							gctx->ctr))
+					return -1;
+				}
+			else	{
+				if (CRYPTO_gcm128_decrypt(&gctx->gcm, in, out, len))
+					return -1;
+				}
+			}
+		return len;
+		}
+	else
+		{
+		if (!ctx->encrypt)
+			{
+			if (CRYPTO_gcm128_finish(&gctx->gcm,
+					ctx->buf, gctx->taglen) != 0)
+				return -1;
+			gctx->iv_set = 0;
+			return 0;
+			}
+		CRYPTO_gcm128_tag(&gctx->gcm, ctx->buf, 16);
+		gctx->taglen = 16;
+		/* Don't reuse the IV */
+		gctx->iv_set = 0;
+		return 0;
+		}
+
+	}
+
+#define CUSTOM_FLAGS	(EVP_CIPH_FLAG_DEFAULT_ASN1 \
+		| EVP_CIPH_CUSTOM_IV | EVP_CIPH_FLAG_CUSTOM_CIPHER \
+		| EVP_CIPH_ALWAYS_CALL_INIT | EVP_CIPH_CTRL_INIT)
+
+BLOCK_CIPHER_custom(NID_aes,128,1,12,gcm,GCM,
+		EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_AEAD_CIPHER|CUSTOM_FLAGS)
+BLOCK_CIPHER_custom(NID_aes,192,1,12,gcm,GCM,
+		EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_AEAD_CIPHER|CUSTOM_FLAGS)
+BLOCK_CIPHER_custom(NID_aes,256,1,12,gcm,GCM,
+		EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_AEAD_CIPHER|CUSTOM_FLAGS)
+
+static int aes_xts_ctrl(EVP_CIPHER_CTX *c, int type, int arg, void *ptr)
+	{
+	EVP_AES_XTS_CTX *xctx = c->cipher_data;
+	if (type != EVP_CTRL_INIT)
+		return -1;
+	/* key1 and key2 are used as an indicator both key and IV are set */
+	xctx->xts.key1 = NULL;
+	xctx->xts.key2 = NULL;
+	return 1;
+	}
+
+static int aes_xts_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
+                        const unsigned char *iv, int enc)
+	{
+	EVP_AES_XTS_CTX *xctx = ctx->cipher_data;
+	if (!iv && !key)
+		return 1;
+
+	if (key) do
+		{
+#ifdef AES_XTS_ASM
+		xctx->stream = enc ? AES_xts_encrypt : AES_xts_decrypt;
+#else
+		xctx->stream = NULL;
+#endif
+		/* key_len is two AES keys */
+#ifdef BSAES_CAPABLE
+		if (BSAES_CAPABLE)
+			xctx->stream = enc ? bsaes_xts_encrypt : bsaes_xts_decrypt;
+		else
+#endif
+#ifdef VPAES_CAPABLE
+		if (VPAES_CAPABLE)
+		    {
+		    if (enc)
+			{
+			vpaes_set_encrypt_key(key, ctx->key_len * 4, &xctx->ks1);
+			xctx->xts.block1 = (block128_f)vpaes_encrypt;
+			}
+		    else
+			{
+			vpaes_set_decrypt_key(key, ctx->key_len * 4, &xctx->ks1);
+			xctx->xts.block1 = (block128_f)vpaes_decrypt;
+			}
+
+		vpaes_set_encrypt_key(key + ctx->key_len/2,
+						ctx->key_len * 4, &xctx->ks2);
+		xctx->xts.block2 = (block128_f)vpaes_encrypt;
+
+		xctx->xts.key1 = &xctx->ks1;
+		break;
+		}
+#endif
+		if (enc)
+			{
+			AES_set_encrypt_key(key, ctx->key_len * 4, &xctx->ks1);
+			xctx->xts.block1 = (block128_f)AES_encrypt;
+			}
+		else
+			{
+			AES_set_decrypt_key(key, ctx->key_len * 4, &xctx->ks1);
+			xctx->xts.block1 = (block128_f)AES_decrypt;
+			}
+
+		AES_set_encrypt_key(key + ctx->key_len/2,
+						ctx->key_len * 4, &xctx->ks2);
+		xctx->xts.block2 = (block128_f)AES_encrypt;
+
+		xctx->xts.key1 = &xctx->ks1;
+		} while (0);
+
+	if (iv)
+		{
+		xctx->xts.key2 = &xctx->ks2;
+		memcpy(ctx->iv, iv, 16);
+		}
+
+	return 1;
+	}
+
+static int aes_xts_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		const unsigned char *in, size_t len)
+	{
+	EVP_AES_XTS_CTX *xctx = ctx->cipher_data;
+	if (!xctx->xts.key1 || !xctx->xts.key2)
+		return 0;
+	if (!out || !in || len<AES_BLOCK_SIZE)
+		return 0;
+#ifdef OPENSSL_FIPS
+	/* Requirement of SP800-38E */
+	if (FIPS_module_mode() && !(ctx->flags & EVP_CIPH_FLAG_NON_FIPS_ALLOW) &&
+			(len > (1UL<<20)*16))
+		{
+		EVPerr(EVP_F_AES_XTS_CIPHER, EVP_R_TOO_LARGE);
+		return 0;
+		}
+#endif
+	if (xctx->stream)
+		(*xctx->stream)(in, out, len,
+				xctx->xts.key1, xctx->xts.key2, ctx->iv);
+	else if (CRYPTO_xts128_encrypt(&xctx->xts, ctx->iv, in, out, len,
+								ctx->encrypt))
+		return 0;
+	return 1;
+	}
+
+#define aes_xts_cleanup NULL
+
+#define XTS_FLAGS	(EVP_CIPH_FLAG_DEFAULT_ASN1 | EVP_CIPH_CUSTOM_IV \
+			 | EVP_CIPH_ALWAYS_CALL_INIT | EVP_CIPH_CTRL_INIT)
+
+BLOCK_CIPHER_custom(NID_aes,128,1,16,xts,XTS,EVP_CIPH_FLAG_FIPS|XTS_FLAGS)
+BLOCK_CIPHER_custom(NID_aes,256,1,16,xts,XTS,EVP_CIPH_FLAG_FIPS|XTS_FLAGS)
+
+static int aes_ccm_ctrl(EVP_CIPHER_CTX *c, int type, int arg, void *ptr)
+	{
+	EVP_AES_CCM_CTX *cctx = c->cipher_data;
+	switch (type)
+		{
+	case EVP_CTRL_INIT:
+		cctx->key_set = 0;
+		cctx->iv_set = 0;
+		cctx->L = 8;
+		cctx->M = 12;
+		cctx->tag_set = 0;
+		cctx->len_set = 0;
+		return 1;
+
+	case EVP_CTRL_CCM_SET_IVLEN:
+		arg = 15 - arg;
+	case EVP_CTRL_CCM_SET_L:
+		if (arg < 2 || arg > 8)
+			return 0;
+		cctx->L = arg;
+		return 1;
+
+	case EVP_CTRL_CCM_SET_TAG:
+		if ((arg & 1) || arg < 4 || arg > 16)
+			return 0;
+		if ((c->encrypt && ptr) || (!c->encrypt && !ptr))
+			return 0;
+		if (ptr)
+			{
+			cctx->tag_set = 1;
+			memcpy(c->buf, ptr, arg);
+			}
+		cctx->M = arg;
+		return 1;
+
+	case EVP_CTRL_CCM_GET_TAG:
+		if (!c->encrypt || !cctx->tag_set)
+			return 0;
+		if(!CRYPTO_ccm128_tag(&cctx->ccm, ptr, (size_t)arg))
+			return 0;
+		cctx->tag_set = 0;
+		cctx->iv_set = 0;
+		cctx->len_set = 0;
+		return 1;
+
+	default:
+		return -1;
+
+		}
+	}
+
+static int aes_ccm_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
+                        const unsigned char *iv, int enc)
+	{
+	EVP_AES_CCM_CTX *cctx = ctx->cipher_data;
+	if (!iv && !key)
+		return 1;
+	if (key) do
+		{
+#ifdef VPAES_CAPABLE
+		if (VPAES_CAPABLE)
+			{
+			vpaes_set_encrypt_key(key, ctx->key_len*8, &cctx->ks);
+			CRYPTO_ccm128_init(&cctx->ccm, cctx->M, cctx->L,
+					&cctx->ks, (block128_f)vpaes_encrypt);
+			cctx->key_set = 1;
+			break;
+			}
+#endif
+		AES_set_encrypt_key(key, ctx->key_len * 8, &cctx->ks);
+		CRYPTO_ccm128_init(&cctx->ccm, cctx->M, cctx->L,
+					&cctx->ks, (block128_f)AES_encrypt);
+		cctx->str = NULL;
+		cctx->key_set = 1;
+		} while (0);
+	if (iv)
+		{
+		memcpy(ctx->iv, iv, 15 - cctx->L);
+		cctx->iv_set = 1;
+		}
+	return 1;
+	}
+
+static int aes_ccm_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		const unsigned char *in, size_t len)
+	{
+	EVP_AES_CCM_CTX *cctx = ctx->cipher_data;
+	CCM128_CONTEXT *ccm = &cctx->ccm;
+	/* If not set up, return error */
+	if (!cctx->iv_set && !cctx->key_set)
+		return -1;
+	if (!ctx->encrypt && !cctx->tag_set)
+		return -1;
+	if (!out)
+		{
+		if (!in)
+			{
+			if (CRYPTO_ccm128_setiv(ccm, ctx->iv, 15 - cctx->L,len))
+				return -1;
+			cctx->len_set = 1;
+			return len;
+			}
+		/* If have AAD need message length */
+		if (!cctx->len_set && len)
+			return -1;
+		CRYPTO_ccm128_aad(ccm, in, len);
+		return len;
+		}
+	/* EVP_*Final() doesn't return any data */
+	if (!in)
+		return 0;
+	/* If not set length yet do it */
+	if (!cctx->len_set)
+		{
+		if (CRYPTO_ccm128_setiv(ccm, ctx->iv, 15 - cctx->L, len))
+			return -1;
+		cctx->len_set = 1;
+		}
+	if (ctx->encrypt)
+		{
+		if (cctx->str ? CRYPTO_ccm128_encrypt_ccm64(ccm, in, out, len,
+						cctx->str) :
+				CRYPTO_ccm128_encrypt(ccm, in, out, len))
+			return -1;
+		cctx->tag_set = 1;
+		return len;
+		}
+	else
+		{
+		int rv = -1;
+		if (cctx->str ? !CRYPTO_ccm128_decrypt_ccm64(ccm, in, out, len,
+						cctx->str) :
+				!CRYPTO_ccm128_decrypt(ccm, in, out, len))
+			{
+			unsigned char tag[16];
+			if (CRYPTO_ccm128_tag(ccm, tag, cctx->M))
+				{
+				if (!memcmp(tag, ctx->buf, cctx->M))
+					rv = len;
+				}
+			}
+		if (rv == -1)
+			OPENSSL_cleanse(out, len);
+		cctx->iv_set = 0;
+		cctx->tag_set = 0;
+		cctx->len_set = 0;
+		return rv;
+		}
+
+	}
+
+#define aes_ccm_cleanup NULL
+
+BLOCK_CIPHER_custom(NID_aes,128,1,12,ccm,CCM,EVP_CIPH_FLAG_FIPS|CUSTOM_FLAGS)
+BLOCK_CIPHER_custom(NID_aes,192,1,12,ccm,CCM,EVP_CIPH_FLAG_FIPS|CUSTOM_FLAGS)
+BLOCK_CIPHER_custom(NID_aes,256,1,12,ccm,CCM,EVP_CIPH_FLAG_FIPS|CUSTOM_FLAGS)
+
+#endif
 #endif
diff --git a/crypto/evp/e_aes_cbc_hmac_sha1.c b/crypto/evp/e_aes_cbc_hmac_sha1.c
new file mode 100644
index 000000000000..710fb79baf45
--- /dev/null
+++ b/crypto/evp/e_aes_cbc_hmac_sha1.c
@@ -0,0 +1,406 @@
+/* ====================================================================
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#include <openssl/opensslconf.h>
+
+#include <stdio.h>
+#include <string.h>
+
+#if !defined(OPENSSL_NO_AES) && !defined(OPENSSL_NO_SHA1)
+
+#include <openssl/evp.h>
+#include <openssl/objects.h>
+#include <openssl/aes.h>
+#include <openssl/sha.h>
+#include "evp_locl.h"
+
+#ifndef EVP_CIPH_FLAG_AEAD_CIPHER
+#define EVP_CIPH_FLAG_AEAD_CIPHER	0x200000
+#define EVP_CTRL_AEAD_TLS1_AAD		0x16
+#define EVP_CTRL_AEAD_SET_MAC_KEY	0x17
+#endif
+
+#if !defined(EVP_CIPH_FLAG_DEFAULT_ASN1)
+#define EVP_CIPH_FLAG_DEFAULT_ASN1 0
+#endif
+
+#define TLS1_1_VERSION 0x0302
+
+typedef struct
+    {
+    AES_KEY		ks;
+    SHA_CTX		head,tail,md;
+    size_t		payload_length;	/* AAD length in decrypt case */
+    union {
+	unsigned int	tls_ver;
+    	unsigned char	tls_aad[16];	/* 13 used */
+    } aux;
+    } EVP_AES_HMAC_SHA1;
+
+#define NO_PAYLOAD_LENGTH	((size_t)-1)
+
+#if	defined(AES_ASM) &&	( \
+	defined(__x86_64)	|| defined(__x86_64__)	|| \
+	defined(_M_AMD64)	|| defined(_M_X64)	|| \
+	defined(__INTEL__)	)
+
+extern unsigned int OPENSSL_ia32cap_P[2];
+#define AESNI_CAPABLE   (1<<(57-32))
+
+int aesni_set_encrypt_key(const unsigned char *userKey, int bits,
+			      AES_KEY *key);
+int aesni_set_decrypt_key(const unsigned char *userKey, int bits,
+			      AES_KEY *key);
+
+void aesni_cbc_encrypt(const unsigned char *in,
+			   unsigned char *out,
+			   size_t length,
+			   const AES_KEY *key,
+			   unsigned char *ivec, int enc);
+
+void aesni_cbc_sha1_enc (const void *inp, void *out, size_t blocks,
+		const AES_KEY *key, unsigned char iv[16],
+		SHA_CTX *ctx,const void *in0);
+
+#define data(ctx) ((EVP_AES_HMAC_SHA1 *)(ctx)->cipher_data)
+
+static int aesni_cbc_hmac_sha1_init_key(EVP_CIPHER_CTX *ctx,
+			const unsigned char *inkey,
+			const unsigned char *iv, int enc)
+	{
+	EVP_AES_HMAC_SHA1 *key = data(ctx);
+	int ret;
+
+	if (enc)
+		ret=aesni_set_encrypt_key(inkey,ctx->key_len*8,&key->ks);
+	else
+		ret=aesni_set_decrypt_key(inkey,ctx->key_len*8,&key->ks);
+
+	SHA1_Init(&key->head);	/* handy when benchmarking */
+	key->tail = key->head;
+	key->md   = key->head;
+
+	key->payload_length = NO_PAYLOAD_LENGTH;
+
+	return ret<0?0:1;
+	}
+
+#define	STITCHED_CALL
+
+#if !defined(STITCHED_CALL)
+#define	aes_off 0
+#endif
+
+void sha1_block_data_order (void *c,const void *p,size_t len);
+
+static void sha1_update(SHA_CTX *c,const void *data,size_t len)
+{	const unsigned char *ptr = data;
+	size_t res;
+
+	if ((res = c->num)) {
+		res = SHA_CBLOCK-res;
+		if (len<res) res=len;
+		SHA1_Update (c,ptr,res);
+		ptr += res;
+		len -= res;
+	}
+
+	res = len % SHA_CBLOCK;
+	len -= res;
+
+	if (len) {
+		sha1_block_data_order(c,ptr,len/SHA_CBLOCK);
+
+		ptr += len;
+		c->Nh += len>>29;
+		c->Nl += len<<=3;
+		if (c->Nl<(unsigned int)len) c->Nh++;
+	}
+
+	if (res)
+		SHA1_Update(c,ptr,res);
+}
+
+#define SHA1_Update sha1_update
+
+static int aesni_cbc_hmac_sha1_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		      const unsigned char *in, size_t len)
+	{
+	EVP_AES_HMAC_SHA1 *key = data(ctx);
+	unsigned int l;
+	size_t	plen = key->payload_length,
+		iv = 0,		/* explicit IV in TLS 1.1 and later */
+		sha_off = 0;
+#if defined(STITCHED_CALL)
+	size_t	aes_off = 0,
+		blocks;
+
+	sha_off = SHA_CBLOCK-key->md.num;
+#endif
+
+	if (len%AES_BLOCK_SIZE) return 0;
+
+	if (ctx->encrypt) {
+		if (plen==NO_PAYLOAD_LENGTH)
+			plen = len;
+		else if (len!=((plen+SHA_DIGEST_LENGTH+AES_BLOCK_SIZE)&-AES_BLOCK_SIZE))
+			return 0;
+		else if (key->aux.tls_ver >= TLS1_1_VERSION)
+			iv = AES_BLOCK_SIZE;
+
+#if defined(STITCHED_CALL)
+		if (plen>(sha_off+iv) && (blocks=(plen-(sha_off+iv))/SHA_CBLOCK)) {
+			SHA1_Update(&key->md,in+iv,sha_off);
+
+			aesni_cbc_sha1_enc(in,out,blocks,&key->ks,
+				ctx->iv,&key->md,in+iv+sha_off);
+			blocks *= SHA_CBLOCK;
+			aes_off += blocks;
+			sha_off += blocks;
+			key->md.Nh += blocks>>29;
+			key->md.Nl += blocks<<=3;
+			if (key->md.Nl<(unsigned int)blocks) key->md.Nh++;
+		} else {
+			sha_off = 0;
+		}
+#endif
+		sha_off += iv;
+		SHA1_Update(&key->md,in+sha_off,plen-sha_off);
+
+		if (plen!=len)	{	/* "TLS" mode of operation */
+			if (in!=out)
+				memcpy(out+aes_off,in+aes_off,plen-aes_off);
+
+			/* calculate HMAC and append it to payload */
+			SHA1_Final(out+plen,&key->md);
+			key->md = key->tail;
+			SHA1_Update(&key->md,out+plen,SHA_DIGEST_LENGTH);
+			SHA1_Final(out+plen,&key->md);
+
+			/* pad the payload|hmac */
+			plen += SHA_DIGEST_LENGTH;
+			for (l=len-plen-1;plen<len;plen++) out[plen]=l;
+			/* encrypt HMAC|padding at once */
+			aesni_cbc_encrypt(out+aes_off,out+aes_off,len-aes_off,
+					&key->ks,ctx->iv,1);
+		} else {
+			aesni_cbc_encrypt(in+aes_off,out+aes_off,len-aes_off,
+					&key->ks,ctx->iv,1);
+		}
+	} else {
+		unsigned char mac[SHA_DIGEST_LENGTH];
+
+		/* decrypt HMAC|padding at once */
+		aesni_cbc_encrypt(in,out,len,
+				&key->ks,ctx->iv,0);
+
+		if (plen) {	/* "TLS" mode of operation */
+			/* figure out payload length */
+			if (len<(size_t)(out[len-1]+1+SHA_DIGEST_LENGTH))
+				return 0;
+
+			len -= (out[len-1]+1+SHA_DIGEST_LENGTH);
+
+			if ((key->aux.tls_aad[plen-4]<<8|key->aux.tls_aad[plen-3])
+			    >= TLS1_1_VERSION) {
+				len -= AES_BLOCK_SIZE;
+				iv = AES_BLOCK_SIZE;
+			}
+
+			key->aux.tls_aad[plen-2] = len>>8;
+			key->aux.tls_aad[plen-1] = len;
+
+			/* calculate HMAC and verify it */
+			key->md = key->head;
+			SHA1_Update(&key->md,key->aux.tls_aad,plen);
+			SHA1_Update(&key->md,out+iv,len);
+			SHA1_Final(mac,&key->md);
+
+			key->md = key->tail;
+			SHA1_Update(&key->md,mac,SHA_DIGEST_LENGTH);
+			SHA1_Final(mac,&key->md);
+
+			if (memcmp(out+iv+len,mac,SHA_DIGEST_LENGTH))
+				return 0;
+		} else {
+			SHA1_Update(&key->md,out,len);
+		}
+	}
+
+	key->payload_length = NO_PAYLOAD_LENGTH;
+
+	return 1;
+	}
+
+static int aesni_cbc_hmac_sha1_ctrl(EVP_CIPHER_CTX *ctx, int type, int arg, void *ptr)
+	{
+	EVP_AES_HMAC_SHA1 *key = data(ctx);
+
+	switch (type)
+		{
+	case EVP_CTRL_AEAD_SET_MAC_KEY:
+		{
+		unsigned int  i;
+		unsigned char hmac_key[64];
+
+		memset (hmac_key,0,sizeof(hmac_key));
+
+		if (arg > (int)sizeof(hmac_key)) {
+			SHA1_Init(&key->head);
+			SHA1_Update(&key->head,ptr,arg);
+			SHA1_Final(hmac_key,&key->head);
+		} else {
+			memcpy(hmac_key,ptr,arg);
+		}
+
+		for (i=0;i<sizeof(hmac_key);i++)
+			hmac_key[i] ^= 0x36;		/* ipad */
+		SHA1_Init(&key->head);
+		SHA1_Update(&key->head,hmac_key,sizeof(hmac_key));
+
+		for (i=0;i<sizeof(hmac_key);i++)
+			hmac_key[i] ^= 0x36^0x5c;	/* opad */
+		SHA1_Init(&key->tail);
+		SHA1_Update(&key->tail,hmac_key,sizeof(hmac_key));
+
+		return 1;
+		}
+	case EVP_CTRL_AEAD_TLS1_AAD:
+		{
+		unsigned char *p=ptr;
+		unsigned int   len=p[arg-2]<<8|p[arg-1];
+
+		if (ctx->encrypt)
+			{
+			key->payload_length = len;
+			if ((key->aux.tls_ver=p[arg-4]<<8|p[arg-3]) >= TLS1_1_VERSION) {
+				len -= AES_BLOCK_SIZE;
+				p[arg-2] = len>>8;
+				p[arg-1] = len;
+			}
+			key->md = key->head;
+			SHA1_Update(&key->md,p,arg);
+
+			return (int)(((len+SHA_DIGEST_LENGTH+AES_BLOCK_SIZE)&-AES_BLOCK_SIZE)
+				- len);
+			}
+		else
+			{
+			if (arg>13) arg = 13;
+			memcpy(key->aux.tls_aad,ptr,arg);
+			key->payload_length = arg;
+
+			return SHA_DIGEST_LENGTH;
+			}
+		}
+	default:
+		return -1;
+		}
+	}
+
+static EVP_CIPHER aesni_128_cbc_hmac_sha1_cipher =
+	{
+#ifdef NID_aes_128_cbc_hmac_sha1
+	NID_aes_128_cbc_hmac_sha1,
+#else
+	NID_undef,
+#endif
+	16,16,16,
+	EVP_CIPH_CBC_MODE|EVP_CIPH_FLAG_DEFAULT_ASN1|EVP_CIPH_FLAG_AEAD_CIPHER,
+	aesni_cbc_hmac_sha1_init_key,
+	aesni_cbc_hmac_sha1_cipher,
+	NULL,
+	sizeof(EVP_AES_HMAC_SHA1),
+	EVP_CIPH_FLAG_DEFAULT_ASN1?NULL:EVP_CIPHER_set_asn1_iv,
+	EVP_CIPH_FLAG_DEFAULT_ASN1?NULL:EVP_CIPHER_get_asn1_iv,
+	aesni_cbc_hmac_sha1_ctrl,
+	NULL
+	};
+
+static EVP_CIPHER aesni_256_cbc_hmac_sha1_cipher =
+	{
+#ifdef NID_aes_256_cbc_hmac_sha1
+	NID_aes_256_cbc_hmac_sha1,
+#else
+	NID_undef,
+#endif
+	16,32,16,
+	EVP_CIPH_CBC_MODE|EVP_CIPH_FLAG_DEFAULT_ASN1|EVP_CIPH_FLAG_AEAD_CIPHER,
+	aesni_cbc_hmac_sha1_init_key,
+	aesni_cbc_hmac_sha1_cipher,
+	NULL,
+	sizeof(EVP_AES_HMAC_SHA1),
+	EVP_CIPH_FLAG_DEFAULT_ASN1?NULL:EVP_CIPHER_set_asn1_iv,
+	EVP_CIPH_FLAG_DEFAULT_ASN1?NULL:EVP_CIPHER_get_asn1_iv,
+	aesni_cbc_hmac_sha1_ctrl,
+	NULL
+	};
+
+const EVP_CIPHER *EVP_aes_128_cbc_hmac_sha1(void)
+	{
+	return(OPENSSL_ia32cap_P[1]&AESNI_CAPABLE?
+		&aesni_128_cbc_hmac_sha1_cipher:NULL);
+	}
+
+const EVP_CIPHER *EVP_aes_256_cbc_hmac_sha1(void)
+	{
+	return(OPENSSL_ia32cap_P[1]&AESNI_CAPABLE?
+		&aesni_256_cbc_hmac_sha1_cipher:NULL);
+	}
+#else
+const EVP_CIPHER *EVP_aes_128_cbc_hmac_sha1(void)
+	{
+	return NULL;
+	}
+const EVP_CIPHER *EVP_aes_256_cbc_hmac_sha1(void)
+	{
+	return NULL;
+	}
+#endif
+#endif
diff --git a/crypto/evp/e_camellia.c b/crypto/evp/e_camellia.c
index 365d39716455..a7b40d1c6001 100644
--- a/crypto/evp/e_camellia.c
+++ b/crypto/evp/e_camellia.c
@@ -93,7 +93,7 @@ IMPLEMENT_BLOCK_CIPHER(camellia_256, ks, Camellia, EVP_CAMELLIA_KEY,
 	EVP_CIPHER_get_asn1_iv,
 	NULL)
 
-#define IMPLEMENT_CAMELLIA_CFBR(ksize,cbits)	IMPLEMENT_CFBR(camellia,Camellia,EVP_CAMELLIA_KEY,ks,ksize,cbits,16,0)
+#define IMPLEMENT_CAMELLIA_CFBR(ksize,cbits)	IMPLEMENT_CFBR(camellia,Camellia,EVP_CAMELLIA_KEY,ks,ksize,cbits,16)
 
 IMPLEMENT_CAMELLIA_CFBR(128,1)
 IMPLEMENT_CAMELLIA_CFBR(192,1)
diff --git a/crypto/evp/e_des.c b/crypto/evp/e_des.c
index 04376df23245..ca009f2c5250 100644
--- a/crypto/evp/e_des.c
+++ b/crypto/evp/e_des.c
@@ -72,7 +72,7 @@ static int des_ctrl(EVP_CIPHER_CTX *c, int type, int arg, void *ptr);
 /* Because of various casts and different names can't use IMPLEMENT_BLOCK_CIPHER */
 
 static int des_ecb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			  const unsigned char *in, unsigned int inl)
+			  const unsigned char *in, size_t inl)
 {
 	BLOCK_CIPHER_ecb_loop()
 		DES_ecb_encrypt((DES_cblock *)(in + i), (DES_cblock *)(out + i), ctx->cipher_data, ctx->encrypt);
@@ -80,24 +80,52 @@ static int des_ecb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
 }
 
 static int des_ofb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			  const unsigned char *in, unsigned int inl)
+			  const unsigned char *in, size_t inl)
 {
-	DES_ofb64_encrypt(in, out, (long)inl, ctx->cipher_data, (DES_cblock *)ctx->iv, &ctx->num);
+	while(inl>=EVP_MAXCHUNK)
+		{
+		DES_ofb64_encrypt(in, out, (long)EVP_MAXCHUNK, ctx->cipher_data,
+				(DES_cblock *)ctx->iv, &ctx->num);
+		inl-=EVP_MAXCHUNK;
+		in +=EVP_MAXCHUNK;
+		out+=EVP_MAXCHUNK;
+		}
+	if (inl)
+		DES_ofb64_encrypt(in, out, (long)inl, ctx->cipher_data,
+				(DES_cblock *)ctx->iv, &ctx->num);
 	return 1;
 }
 
 static int des_cbc_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			  const unsigned char *in, unsigned int inl)
+			  const unsigned char *in, size_t inl)
 {
-	DES_ncbc_encrypt(in, out, (long)inl, ctx->cipher_data,
-			 (DES_cblock *)ctx->iv, ctx->encrypt);
+	while(inl>=EVP_MAXCHUNK)
+		{
+		DES_ncbc_encrypt(in, out, (long)EVP_MAXCHUNK, ctx->cipher_data,
+				(DES_cblock *)ctx->iv, ctx->encrypt);
+		inl-=EVP_MAXCHUNK;
+		in +=EVP_MAXCHUNK;
+		out+=EVP_MAXCHUNK;
+		}
+	if (inl)
+		DES_ncbc_encrypt(in, out, (long)inl, ctx->cipher_data,
+				(DES_cblock *)ctx->iv, ctx->encrypt);
 	return 1;
 }
 
 static int des_cfb64_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			    const unsigned char *in, unsigned int inl)
+			    const unsigned char *in, size_t inl)
 {
-	DES_cfb64_encrypt(in, out, (long)inl, ctx->cipher_data,
+	while(inl>=EVP_MAXCHUNK)
+		{
+		DES_cfb64_encrypt(in,out, (long)EVP_MAXCHUNK, ctx->cipher_data,
+				(DES_cblock *)ctx->iv, &ctx->num, ctx->encrypt);
+		inl-=EVP_MAXCHUNK;
+		in +=EVP_MAXCHUNK;
+		out+=EVP_MAXCHUNK;
+		}
+	if (inl)
+		DES_cfb64_encrypt(in, out, (long)inl, ctx->cipher_data,
 			  (DES_cblock *)ctx->iv, &ctx->num, ctx->encrypt);
 	return 1;
 }
@@ -105,45 +133,62 @@ static int des_cfb64_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
 /* Although we have a CFB-r implementation for DES, it doesn't pack the right
    way, so wrap it here */
 static int des_cfb1_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			   const unsigned char *in, unsigned int inl)
+			   const unsigned char *in, size_t inl)
     {
-    unsigned int n;
+    size_t n,chunk=EVP_MAXCHUNK/8;
     unsigned char c[1],d[1];
 
-    for(n=0 ; n < inl ; ++n)
+    if (inl<chunk) chunk=inl;
+
+    while (inl && inl>=chunk)
 	{
-	c[0]=(in[n/8]&(1 << (7-n%8))) ? 0x80 : 0;
-	DES_cfb_encrypt(c,d,1,1,ctx->cipher_data,(DES_cblock *)ctx->iv,
+	for(n=0 ; n < chunk*8; ++n)
+	    {
+	    c[0]=(in[n/8]&(1 << (7-n%8))) ? 0x80 : 0;
+	    DES_cfb_encrypt(c,d,1,1,ctx->cipher_data,(DES_cblock *)ctx->iv,
 			ctx->encrypt);
-	out[n/8]=(out[n/8]&~(0x80 >> (n%8)))|((d[0]&0x80) >> (n%8));
+	    out[n/8]=(out[n/8]&~(0x80 >> (unsigned int)(n%8))) |
+		     ((d[0]&0x80) >> (unsigned int)(n%8));
+	    }
+	inl-=chunk;
+	in +=chunk;
+	out+=chunk;
+	if (inl<chunk) chunk=inl;
 	}
+
     return 1;
     }
 
 static int des_cfb8_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			   const unsigned char *in, unsigned int inl)
+			   const unsigned char *in, size_t inl)
     {
-    DES_cfb_encrypt(in,out,8,inl,ctx->cipher_data,(DES_cblock *)ctx->iv,
-		    ctx->encrypt);
+    while (inl>=EVP_MAXCHUNK)
+	{
+	DES_cfb_encrypt(in,out,8,(long)EVP_MAXCHUNK,ctx->cipher_data,
+			(DES_cblock *)ctx->iv,ctx->encrypt);
+	inl-=EVP_MAXCHUNK;
+	in +=EVP_MAXCHUNK;
+	out+=EVP_MAXCHUNK;
+	}
+    if (inl)
+	DES_cfb_encrypt(in,out,8,(long)inl,ctx->cipher_data,
+			(DES_cblock *)ctx->iv,ctx->encrypt);
     return 1;
     }
 
 BLOCK_CIPHER_defs(des, DES_key_schedule, NID_des, 8, 8, 8, 64,
-			EVP_CIPH_RAND_KEY,
-			des_init_key, NULL,
+			EVP_CIPH_RAND_KEY, des_init_key, NULL,
 			EVP_CIPHER_set_asn1_iv,
 			EVP_CIPHER_get_asn1_iv,
 			des_ctrl)
 
 BLOCK_CIPHER_def_cfb(des,DES_key_schedule,NID_des,8,8,1,
-		     EVP_CIPH_RAND_KEY,
-		     des_init_key, NULL,
+		     EVP_CIPH_RAND_KEY, des_init_key,NULL,
 		     EVP_CIPHER_set_asn1_iv,
 		     EVP_CIPHER_get_asn1_iv,des_ctrl)
 
 BLOCK_CIPHER_def_cfb(des,DES_key_schedule,NID_des,8,8,8,
-		     EVP_CIPH_RAND_KEY,
-		     des_init_key,NULL,
+		     EVP_CIPH_RAND_KEY,des_init_key,NULL,
 		     EVP_CIPHER_set_asn1_iv,
 		     EVP_CIPHER_get_asn1_iv,des_ctrl)
 
diff --git a/crypto/evp/e_des3.c b/crypto/evp/e_des3.c
index f910af19b110..1e699726627d 100644
--- a/crypto/evp/e_des3.c
+++ b/crypto/evp/e_des3.c
@@ -65,6 +65,8 @@
 #include <openssl/des.h>
 #include <openssl/rand.h>
 
+#ifndef OPENSSL_FIPS
+
 static int des_ede_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 			    const unsigned char *iv,int enc);
 
@@ -85,7 +87,7 @@ typedef struct
 /* Because of various casts and different args can't use IMPLEMENT_BLOCK_CIPHER */
 
 static int des_ede_ecb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			      const unsigned char *in, unsigned int inl)
+			      const unsigned char *in, size_t inl)
 {
 	BLOCK_CIPHER_ecb_loop()
 		DES_ecb3_encrypt((const_DES_cblock *)(in + i),
@@ -97,48 +99,80 @@ static int des_ede_ecb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
 }
 
 static int des_ede_ofb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			      const unsigned char *in, unsigned int inl)
+			      const unsigned char *in, size_t inl)
 {
-	DES_ede3_ofb64_encrypt(in, out, (long)inl,
+	if (inl>=EVP_MAXCHUNK)
+		{
+		DES_ede3_ofb64_encrypt(in, out, (long)EVP_MAXCHUNK,
 			       &data(ctx)->ks1, &data(ctx)->ks2, &data(ctx)->ks3,
 			       (DES_cblock *)ctx->iv, &ctx->num);
+		inl-=EVP_MAXCHUNK;
+		in +=EVP_MAXCHUNK;
+		out+=EVP_MAXCHUNK;
+		}
+	if (inl)
+		DES_ede3_ofb64_encrypt(in, out, (long)inl,
+				&data(ctx)->ks1, &data(ctx)->ks2, &data(ctx)->ks3,
+                               (DES_cblock *)ctx->iv, &ctx->num);
+
 	return 1;
 }
 
 static int des_ede_cbc_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			      const unsigned char *in, unsigned int inl)
+			      const unsigned char *in, size_t inl)
 {
 #ifdef KSSL_DEBUG
 	{
         int i;
-	printf("des_ede_cbc_cipher(ctx=%lx, buflen=%d)\n", (unsigned long)ctx, ctx->buf_len);
+        char *cp;
+	printf("des_ede_cbc_cipher(ctx=%lx, buflen=%d)\n", ctx, ctx->buf_len);
 	printf("\t iv= ");
         for(i=0;i<8;i++)
                 printf("%02X",ctx->iv[i]);
 	printf("\n");
 	}
 #endif    /* KSSL_DEBUG */
-	DES_ede3_cbc_encrypt(in, out, (long)inl,
+	if (inl>=EVP_MAXCHUNK)
+		{
+		DES_ede3_cbc_encrypt(in, out, (long)EVP_MAXCHUNK,
 			     &data(ctx)->ks1, &data(ctx)->ks2, &data(ctx)->ks3,
 			     (DES_cblock *)ctx->iv, ctx->encrypt);
+		inl-=EVP_MAXCHUNK;
+		in +=EVP_MAXCHUNK;
+		out+=EVP_MAXCHUNK;
+		}
+	if (inl)
+		DES_ede3_cbc_encrypt(in, out, (long)inl,
+			     &data(ctx)->ks1, &data(ctx)->ks2, &data(ctx)->ks3,
+                             (DES_cblock *)ctx->iv, ctx->encrypt);
 	return 1;
 }
 
 static int des_ede_cfb64_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			      const unsigned char *in, unsigned int inl)
+			      const unsigned char *in, size_t inl)
 {
-	DES_ede3_cfb64_encrypt(in, out, (long)inl, 
+	if (inl>=EVP_MAXCHUNK)
+		{
+		DES_ede3_cfb64_encrypt(in, out, (long)EVP_MAXCHUNK, 
 			       &data(ctx)->ks1, &data(ctx)->ks2, &data(ctx)->ks3,
 			       (DES_cblock *)ctx->iv, &ctx->num, ctx->encrypt);
+		inl-=EVP_MAXCHUNK;
+		in +=EVP_MAXCHUNK;
+		out+=EVP_MAXCHUNK;
+		}
+	if (inl)
+		DES_ede3_cfb64_encrypt(in, out, (long)inl,
+			       &data(ctx)->ks1, &data(ctx)->ks2, &data(ctx)->ks3,
+                               (DES_cblock *)ctx->iv, &ctx->num, ctx->encrypt);
 	return 1;
 }
 
 /* Although we have a CFB-r implementation for 3-DES, it doesn't pack the right
    way, so wrap it here */
 static int des_ede3_cfb1_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-				const unsigned char *in, unsigned int inl)
+				const unsigned char *in, size_t inl)
     {
-    unsigned int n;
+    size_t n;
     unsigned char c[1],d[1];
 
     for(n=0 ; n < inl ; ++n)
@@ -147,25 +181,36 @@ static int des_ede3_cfb1_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
 	DES_ede3_cfb_encrypt(c,d,1,1,
 			     &data(ctx)->ks1,&data(ctx)->ks2,&data(ctx)->ks3,
 			     (DES_cblock *)ctx->iv,ctx->encrypt);
-	out[n/8]=(out[n/8]&~(0x80 >> (n%8)))|((d[0]&0x80) >> (n%8));
+	out[n/8]=(out[n/8]&~(0x80 >> (unsigned int)(n%8))) |
+		 ((d[0]&0x80) >> (unsigned int)(n%8));
 	}
 
     return 1;
     }
 
 static int des_ede3_cfb8_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-				const unsigned char *in, unsigned int inl)
+				const unsigned char *in, size_t inl)
     {
-    DES_ede3_cfb_encrypt(in,out,8,inl,
+    while (inl>=EVP_MAXCHUNK)
+	{
+	DES_ede3_cfb_encrypt(in,out,8,(long)EVP_MAXCHUNK,
 			 &data(ctx)->ks1,&data(ctx)->ks2,&data(ctx)->ks3,
 			 (DES_cblock *)ctx->iv,ctx->encrypt);
+	inl-=EVP_MAXCHUNK;
+	in +=EVP_MAXCHUNK;
+	out+=EVP_MAXCHUNK;
+	}
+    if (inl)
+	DES_ede3_cfb_encrypt(in,out,8,(long)inl,
+			&data(ctx)->ks1,&data(ctx)->ks2,&data(ctx)->ks3,
+			(DES_cblock *)ctx->iv,ctx->encrypt);
     return 1;
     }
 
 BLOCK_CIPHER_defs(des_ede, DES_EDE_KEY, NID_des_ede, 8, 16, 8, 64,
-		EVP_CIPH_RAND_KEY|EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_DEFAULT_ASN1,
-			des_ede_init_key,
-			NULL, NULL, NULL,
+			EVP_CIPH_RAND_KEY, des_ede_init_key, NULL, 
+			EVP_CIPHER_set_asn1_iv,
+			EVP_CIPHER_get_asn1_iv,
 			des3_ctrl)
 
 #define des_ede3_cfb64_cipher des_ede_cfb64_cipher
@@ -174,21 +219,21 @@ BLOCK_CIPHER_defs(des_ede, DES_EDE_KEY, NID_des_ede, 8, 16, 8, 64,
 #define des_ede3_ecb_cipher des_ede_ecb_cipher
 
 BLOCK_CIPHER_defs(des_ede3, DES_EDE_KEY, NID_des_ede3, 8, 24, 8, 64,
-		EVP_CIPH_RAND_KEY|EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_DEFAULT_ASN1,
-			des_ede3_init_key,
-			NULL, NULL, NULL,
+			EVP_CIPH_RAND_KEY, des_ede3_init_key, NULL, 
+			EVP_CIPHER_set_asn1_iv,
+			EVP_CIPHER_get_asn1_iv,
 			des3_ctrl)
 
 BLOCK_CIPHER_def_cfb(des_ede3,DES_EDE_KEY,NID_des_ede3,24,8,1,
-		EVP_CIPH_RAND_KEY|EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_DEFAULT_ASN1,
-		     des_ede3_init_key,
-		     NULL, NULL, NULL,
+		     EVP_CIPH_RAND_KEY, des_ede3_init_key,NULL,
+		     EVP_CIPHER_set_asn1_iv,
+		     EVP_CIPHER_get_asn1_iv,
 		     des3_ctrl)
 
 BLOCK_CIPHER_def_cfb(des_ede3,DES_EDE_KEY,NID_des_ede3,24,8,8,
-		EVP_CIPH_RAND_KEY|EVP_CIPH_FLAG_FIPS|EVP_CIPH_FLAG_DEFAULT_ASN1,
-		     des_ede3_init_key,
-		     NULL, NULL, NULL,
+		     EVP_CIPH_RAND_KEY, des_ede3_init_key,NULL,
+		     EVP_CIPHER_set_asn1_iv,
+		     EVP_CIPHER_get_asn1_iv,
 		     des3_ctrl)
 
 static int des_ede_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
@@ -215,7 +260,7 @@ static int des_ede3_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 #ifdef KSSL_DEBUG
 	{
         int i;
-        printf("des_ede3_init_key(ctx=%lx)\n", (unsigned long)ctx);
+        printf("des_ede3_init_key(ctx=%lx)\n", ctx);
 	printf("\tKEY= ");
         for(i=0;i<24;i++) printf("%02X",key[i]); printf("\n");
 	printf("\t IV= ");
@@ -268,3 +313,4 @@ const EVP_CIPHER *EVP_des_ede3(void)
 	return &des_ede3_ecb;
 }
 #endif
+#endif
diff --git a/crypto/evp/e_idea.c b/crypto/evp/e_idea.c
index 48c33a774a5b..806b08036005 100644
--- a/crypto/evp/e_idea.c
+++ b/crypto/evp/e_idea.c
@@ -73,7 +73,7 @@ static int idea_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
  */
 
 static int idea_ecb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			   const unsigned char *in, unsigned int inl)
+			   const unsigned char *in, size_t inl)
 {
 	BLOCK_CIPHER_ecb_loop()
 		idea_ecb_encrypt(in + i, out + i, ctx->cipher_data);
diff --git a/crypto/evp/e_null.c b/crypto/evp/e_null.c
index 0872d733e475..f0c1f78b5fec 100644
--- a/crypto/evp/e_null.c
+++ b/crypto/evp/e_null.c
@@ -61,15 +61,17 @@
 #include <openssl/evp.h>
 #include <openssl/objects.h>
 
+#ifndef OPENSSL_FIPS
+
 static int null_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 	const unsigned char *iv,int enc);
 static int null_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-	const unsigned char *in, unsigned int inl);
+	const unsigned char *in, size_t inl);
 static const EVP_CIPHER n_cipher=
 	{
 	NID_undef,
 	1,0,0,
-	EVP_CIPH_FLAG_FIPS,
+	0,
 	null_init_key,
 	null_cipher,
 	NULL,
@@ -93,10 +95,10 @@ static int null_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 	}
 
 static int null_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-	     const unsigned char *in, unsigned int inl)
+	     const unsigned char *in, size_t inl)
 	{
 	if (in != out)
-		memcpy((char *)out,(const char *)in,(size_t)inl);
+		memcpy((char *)out,(const char *)in,inl);
 	return 1;
 	}
-
+#endif
diff --git a/crypto/evp/e_rc2.c b/crypto/evp/e_rc2.c
index d37726ffae4b..d4c33b58d4d5 100644
--- a/crypto/evp/e_rc2.c
+++ b/crypto/evp/e_rc2.c
@@ -183,7 +183,8 @@ static int rc2_get_asn1_type_and_iv(EVP_CIPHER_CTX *c, ASN1_TYPE *type)
 		key_bits =rc2_magic_to_meth((int)num);
 		if (!key_bits)
 			return(-1);
-		if(i > 0) EVP_CipherInit_ex(c, NULL, NULL, NULL, iv, -1);
+		if(i > 0 && !EVP_CipherInit_ex(c, NULL, NULL, NULL, iv, -1))
+			return -1;
 		EVP_CIPHER_CTX_ctrl(c, EVP_CTRL_SET_RC2_KEY_BITS, key_bits, NULL);
 		EVP_CIPHER_CTX_set_key_length(c, key_bits / 8);
 		}
@@ -223,6 +224,11 @@ static int rc2_ctrl(EVP_CIPHER_CTX *c, int type, int arg, void *ptr)
 			return 1;
 			}
 		return 0;
+#ifdef PBE_PRF_TEST
+	case EVP_CTRL_PBE_PRF_NID:
+		*(int *)ptr = NID_hmacWithMD5;
+		return 1;
+#endif
 
 	default:
 		return -1;
diff --git a/crypto/evp/e_rc4.c b/crypto/evp/e_rc4.c
index 55baad7446dc..b4f6bda82d40 100644
--- a/crypto/evp/e_rc4.c
+++ b/crypto/evp/e_rc4.c
@@ -62,9 +62,9 @@
 #ifndef OPENSSL_NO_RC4
 
 #include <openssl/evp.h>
+#include "evp_locl.h"
 #include <openssl/objects.h>
 #include <openssl/rc4.h>
-#include "evp_locl.h"
 
 /* FIXME: surely this is available elsewhere? */
 #define EVP_RC4_KEY_SIZE		16
@@ -79,7 +79,7 @@ typedef struct
 static int rc4_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 			const unsigned char *iv,int enc);
 static int rc4_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-		      const unsigned char *in, unsigned int inl);
+		      const unsigned char *in, size_t inl);
 static const EVP_CIPHER r4_cipher=
 	{
 	NID_rc4,
@@ -129,7 +129,7 @@ static int rc4_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 	}
 
 static int rc4_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-		      const unsigned char *in, unsigned int inl)
+		      const unsigned char *in, size_t inl)
 	{
 	RC4(&data(ctx)->ks,inl,in,out);
 	return 1;
diff --git a/crypto/evp/e_rc4_hmac_md5.c b/crypto/evp/e_rc4_hmac_md5.c
new file mode 100644
index 000000000000..56563191ba16
--- /dev/null
+++ b/crypto/evp/e_rc4_hmac_md5.c
@@ -0,0 +1,298 @@
+/* ====================================================================
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#include <openssl/opensslconf.h>
+
+#include <stdio.h>
+#include <string.h>
+
+#if !defined(OPENSSL_NO_RC4) && !defined(OPENSSL_NO_MD5)
+
+#include <openssl/evp.h>
+#include <openssl/objects.h>
+#include <openssl/rc4.h>
+#include <openssl/md5.h>
+
+#ifndef EVP_CIPH_FLAG_AEAD_CIPHER
+#define EVP_CIPH_FLAG_AEAD_CIPHER	0x200000
+#define EVP_CTRL_AEAD_TLS1_AAD		0x16
+#define EVP_CTRL_AEAD_SET_MAC_KEY	0x17
+#endif
+
+/* FIXME: surely this is available elsewhere? */
+#define EVP_RC4_KEY_SIZE		16
+
+typedef struct
+    {
+    RC4_KEY		ks;
+    MD5_CTX		head,tail,md;
+    size_t		payload_length;
+    } EVP_RC4_HMAC_MD5;
+
+#define NO_PAYLOAD_LENGTH	((size_t)-1)
+
+void rc4_md5_enc (RC4_KEY *key, const void *in0, void *out,
+		MD5_CTX *ctx,const void *inp,size_t blocks);
+
+#define data(ctx) ((EVP_RC4_HMAC_MD5 *)(ctx)->cipher_data)
+
+static int rc4_hmac_md5_init_key(EVP_CIPHER_CTX *ctx,
+			const unsigned char *inkey,
+			const unsigned char *iv, int enc)
+	{
+	EVP_RC4_HMAC_MD5 *key = data(ctx);
+
+	RC4_set_key(&key->ks,EVP_CIPHER_CTX_key_length(ctx),
+		    inkey);
+
+	MD5_Init(&key->head);	/* handy when benchmarking */
+	key->tail = key->head;
+	key->md   = key->head;
+
+	key->payload_length = NO_PAYLOAD_LENGTH;
+
+	return 1;
+	}
+
+#if	!defined(OPENSSL_NO_ASM) &&	( \
+	defined(__x86_64)	|| defined(__x86_64__)	|| \
+	defined(_M_AMD64)	|| defined(_M_X64)	|| \
+	defined(__INTEL__)		) && \
+	!(defined(__APPLE__) && defined(__MACH__))
+#define	STITCHED_CALL
+#endif
+
+#if !defined(STITCHED_CALL)
+#define	rc4_off 0
+#define	md5_off 0
+#endif
+
+static int rc4_hmac_md5_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
+		      const unsigned char *in, size_t len)
+	{
+	EVP_RC4_HMAC_MD5 *key = data(ctx);
+#if defined(STITCHED_CALL)
+	size_t	rc4_off = 32-1-(key->ks.x&(32-1)),	/* 32 is $MOD from rc4_md5-x86_64.pl */
+		md5_off = MD5_CBLOCK-key->md.num,
+		blocks;
+	unsigned int l;
+	extern unsigned int OPENSSL_ia32cap_P[];
+#endif
+	size_t	plen = key->payload_length;
+
+	if (plen!=NO_PAYLOAD_LENGTH && len!=(plen+MD5_DIGEST_LENGTH)) return 0;
+
+	if (ctx->encrypt) {
+		if (plen==NO_PAYLOAD_LENGTH) plen = len;
+#if defined(STITCHED_CALL)
+		/* cipher has to "fall behind" */
+		if (rc4_off>md5_off) md5_off+=MD5_CBLOCK;
+
+		if (plen>md5_off && (blocks=(plen-md5_off)/MD5_CBLOCK) &&
+		    (OPENSSL_ia32cap_P[0]&(1<<20))==0) {
+			MD5_Update(&key->md,in,md5_off);
+			RC4(&key->ks,rc4_off,in,out);
+
+			rc4_md5_enc(&key->ks,in+rc4_off,out+rc4_off,
+				&key->md,in+md5_off,blocks);
+			blocks *= MD5_CBLOCK;
+			rc4_off += blocks;
+			md5_off += blocks;
+			key->md.Nh += blocks>>29;
+			key->md.Nl += blocks<<=3;
+			if (key->md.Nl<(unsigned int)blocks) key->md.Nh++;
+		} else {
+			rc4_off = 0;
+			md5_off = 0;
+		}
+#endif
+		MD5_Update(&key->md,in+md5_off,plen-md5_off);
+
+		if (plen!=len) {	/* "TLS" mode of operation */
+			if (in!=out)
+				memcpy(out+rc4_off,in+rc4_off,plen-rc4_off);
+
+			/* calculate HMAC and append it to payload */
+			MD5_Final(out+plen,&key->md);
+			key->md = key->tail;
+			MD5_Update(&key->md,out+plen,MD5_DIGEST_LENGTH);
+			MD5_Final(out+plen,&key->md);
+			/* encrypt HMAC at once */
+			RC4(&key->ks,len-rc4_off,out+rc4_off,out+rc4_off);
+		} else {
+			RC4(&key->ks,len-rc4_off,in+rc4_off,out+rc4_off);
+		}
+	} else {
+		unsigned char mac[MD5_DIGEST_LENGTH];
+#if defined(STITCHED_CALL)
+		/* digest has to "fall behind" */
+		if (md5_off>rc4_off)	rc4_off += 2*MD5_CBLOCK;
+		else			rc4_off += MD5_CBLOCK;
+
+		if (len>rc4_off && (blocks=(len-rc4_off)/MD5_CBLOCK) &&
+		    (OPENSSL_ia32cap_P[0]&(1<<20))==0) {
+			RC4(&key->ks,rc4_off,in,out);
+			MD5_Update(&key->md,out,md5_off);
+
+			rc4_md5_enc(&key->ks,in+rc4_off,out+rc4_off,
+				&key->md,out+md5_off,blocks);
+			blocks *= MD5_CBLOCK;
+			rc4_off += blocks;
+			md5_off += blocks;
+			l = (key->md.Nl+(blocks<<3))&0xffffffffU;
+			if (l<key->md.Nl) key->md.Nh++;
+			key->md.Nl  = l;
+			key->md.Nh += blocks>>29;
+		} else {
+			md5_off=0;
+			rc4_off=0;
+		}
+#endif
+		/* decrypt HMAC at once */
+		RC4(&key->ks,len-rc4_off,in+rc4_off,out+rc4_off);
+		if (plen!=NO_PAYLOAD_LENGTH) {	/* "TLS" mode of operation */
+			MD5_Update(&key->md,out+md5_off,plen-md5_off);
+
+			/* calculate HMAC and verify it */
+			MD5_Final(mac,&key->md);
+			key->md = key->tail;
+			MD5_Update(&key->md,mac,MD5_DIGEST_LENGTH);
+			MD5_Final(mac,&key->md);
+
+			if (memcmp(out+plen,mac,MD5_DIGEST_LENGTH))
+				return 0;
+		} else {
+			MD5_Update(&key->md,out+md5_off,len-md5_off);
+		}
+	}
+
+	key->payload_length = NO_PAYLOAD_LENGTH;
+
+	return 1;
+	}
+
+static int rc4_hmac_md5_ctrl(EVP_CIPHER_CTX *ctx, int type, int arg, void *ptr)
+	{
+	EVP_RC4_HMAC_MD5 *key = data(ctx);
+
+	switch (type)
+		{
+	case EVP_CTRL_AEAD_SET_MAC_KEY:
+		{
+		unsigned int  i;
+		unsigned char hmac_key[64];
+
+		memset (hmac_key,0,sizeof(hmac_key));
+
+		if (arg > (int)sizeof(hmac_key)) {
+			MD5_Init(&key->head);
+			MD5_Update(&key->head,ptr,arg);
+			MD5_Final(hmac_key,&key->head);
+		} else {
+			memcpy(hmac_key,ptr,arg);
+		}
+
+		for (i=0;i<sizeof(hmac_key);i++)
+			hmac_key[i] ^= 0x36;		/* ipad */
+		MD5_Init(&key->head);
+		MD5_Update(&key->head,hmac_key,sizeof(hmac_key));
+
+		for (i=0;i<sizeof(hmac_key);i++)
+			hmac_key[i] ^= 0x36^0x5c;	/* opad */
+		MD5_Init(&key->tail);
+		MD5_Update(&key->tail,hmac_key,sizeof(hmac_key));
+
+		return 1;
+		}
+	case EVP_CTRL_AEAD_TLS1_AAD:
+		{
+		unsigned char *p=ptr;
+		unsigned int   len=p[arg-2]<<8|p[arg-1];
+
+		if (!ctx->encrypt)
+			{
+			len -= MD5_DIGEST_LENGTH;
+			p[arg-2] = len>>8;
+			p[arg-1] = len;
+			}
+		key->payload_length=len;
+		key->md = key->head;
+		MD5_Update(&key->md,p,arg);
+
+		return MD5_DIGEST_LENGTH;
+		}
+	default:
+		return -1;
+		}
+	}
+
+static EVP_CIPHER r4_hmac_md5_cipher=
+	{
+#ifdef NID_rc4_hmac_md5
+	NID_rc4_hmac_md5,
+#else
+	NID_undef,
+#endif
+	1,EVP_RC4_KEY_SIZE,0,
+	EVP_CIPH_STREAM_CIPHER|EVP_CIPH_VARIABLE_LENGTH|EVP_CIPH_FLAG_AEAD_CIPHER,
+	rc4_hmac_md5_init_key,
+	rc4_hmac_md5_cipher,
+	NULL,
+	sizeof(EVP_RC4_HMAC_MD5),
+	NULL,
+	NULL,
+	rc4_hmac_md5_ctrl,
+	NULL
+	};
+
+const EVP_CIPHER *EVP_rc4_hmac_md5(void)
+	{
+	return(&r4_hmac_md5_cipher);
+	}
+#endif
diff --git a/crypto/evp/e_seed.c b/crypto/evp/e_seed.c
index 8c1ec0d43a60..2d1759d27665 100644
--- a/crypto/evp/e_seed.c
+++ b/crypto/evp/e_seed.c
@@ -54,11 +54,11 @@
  */
 
 #include <openssl/opensslconf.h>
+#ifndef OPENSSL_NO_SEED
 #include <openssl/evp.h>
 #include <openssl/err.h>
 #include <string.h>
 #include <assert.h>
-#ifndef OPENSSL_NO_SEED
 #include <openssl/seed.h>
 #include "evp_locl.h"
 
diff --git a/crypto/evp/e_xcbc_d.c b/crypto/evp/e_xcbc_d.c
index 8832da24333c..250e88c8c5d7 100644
--- a/crypto/evp/e_xcbc_d.c
+++ b/crypto/evp/e_xcbc_d.c
@@ -63,12 +63,13 @@
 
 #include <openssl/evp.h>
 #include <openssl/objects.h>
+#include "evp_locl.h"
 #include <openssl/des.h>
 
 static int desx_cbc_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 			     const unsigned char *iv,int enc);
 static int desx_cbc_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			   const unsigned char *in, unsigned int inl);
+			   const unsigned char *in, size_t inl);
 
 
 typedef struct
@@ -113,13 +114,25 @@ static int desx_cbc_init_key(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 	}
 
 static int desx_cbc_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			   const unsigned char *in, unsigned int inl)
+			   const unsigned char *in, size_t inl)
 	{
-	DES_xcbc_encrypt(in,out,inl,&data(ctx)->ks,
+	while (inl>=EVP_MAXCHUNK)
+		{
+		DES_xcbc_encrypt(in,out,(long)EVP_MAXCHUNK,&data(ctx)->ks,
 			 (DES_cblock *)&(ctx->iv[0]),
 			 &data(ctx)->inw,
 			 &data(ctx)->outw,
 			 ctx->encrypt);
+		inl-=EVP_MAXCHUNK;
+		in +=EVP_MAXCHUNK;
+		out+=EVP_MAXCHUNK;
+		}
+	if (inl)
+		DES_xcbc_encrypt(in,out,(long)inl,&data(ctx)->ks,
+			(DES_cblock *)&(ctx->iv[0]),
+			&data(ctx)->inw,
+			&data(ctx)->outw,
+			ctx->encrypt);
 	return 1;
 	}
 #endif
diff --git a/crypto/evp/enc_min.c b/crypto/evp/enc_min.c
deleted file mode 100644
index a8c176fb3e98..000000000000
--- a/crypto/evp/enc_min.c
+++ /dev/null
@@ -1,391 +0,0 @@
-/* crypto/evp/enc_min.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#include <stdio.h>
-#include "cryptlib.h"
-#include <openssl/evp.h>
-#include <openssl/err.h>
-#include <openssl/rand.h>
-#ifndef OPENSSL_NO_ENGINE
-#include <openssl/engine.h>
-#endif
-#include "evp_locl.h"
-
-void EVP_CIPHER_CTX_init(EVP_CIPHER_CTX *ctx)
-	{
-#ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
-	memset(ctx,0,sizeof(EVP_CIPHER_CTX));
-	/* ctx->cipher=NULL; */
-	}
-
-#ifdef OPENSSL_FIPS
-
-/* The purpose of these is to trap programs that attempt to use non FIPS
- * algorithms in FIPS mode and ignore the errors.
- */
-
-static int bad_init(EVP_CIPHER_CTX *ctx, const unsigned char *key,
-		    const unsigned char *iv, int enc)
-	{ FIPS_ERROR_IGNORED("Cipher init"); return 0;}
-
-static int bad_do_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			 const unsigned char *in, unsigned int inl)
-	{ FIPS_ERROR_IGNORED("Cipher update"); return 0;}
-
-/* NB: no cleanup because it is allowed after failed init */
-
-static int bad_set_asn1(EVP_CIPHER_CTX *ctx, ASN1_TYPE *typ)
-	{ FIPS_ERROR_IGNORED("Cipher set_asn1"); return 0;}
-static int bad_get_asn1(EVP_CIPHER_CTX *ctx, ASN1_TYPE *typ)
-	{ FIPS_ERROR_IGNORED("Cipher get_asn1"); return 0;}
-static int bad_ctrl(EVP_CIPHER_CTX *ctx, int type, int arg, void *ptr)
-	{ FIPS_ERROR_IGNORED("Cipher ctrl"); return 0;}
-
-static const EVP_CIPHER bad_cipher =
-	{
-	0,
-	0,
-	0,
-	0,
-	0,
-	bad_init,
-	bad_do_cipher,
-	NULL,
-	0,
-	bad_set_asn1,
-	bad_get_asn1,
-	bad_ctrl,
-	NULL
-	};
-
-#endif
-
-#ifndef OPENSSL_NO_ENGINE
-
-#ifdef OPENSSL_FIPS
-
-static int do_engine_null(ENGINE *impl) { return 0;}
-static int do_evp_enc_engine_null(EVP_CIPHER_CTX *ctx,
-				const EVP_CIPHER **pciph, ENGINE *impl)
-	{ return 1; }
-
-static int (*do_engine_finish)(ENGINE *impl)
-		= do_engine_null;
-
-static int (*do_evp_enc_engine)
-	(EVP_CIPHER_CTX *ctx, const EVP_CIPHER **pciph, ENGINE *impl)
-		= do_evp_enc_engine_null;
-
-void int_EVP_CIPHER_set_engine_callbacks(
-	int (*eng_ciph_fin)(ENGINE *impl),
-	int (*eng_ciph_evp)
-		(EVP_CIPHER_CTX *ctx, const EVP_CIPHER **pciph, ENGINE *impl))
-	{
-	do_engine_finish = eng_ciph_fin;
-	do_evp_enc_engine = eng_ciph_evp;
-	}
-
-#else
-
-#define do_engine_finish ENGINE_finish
-
-static int do_evp_enc_engine(EVP_CIPHER_CTX *ctx, const EVP_CIPHER **pcipher, ENGINE *impl)
-	{
-	if(impl)
-		{
-		if (!ENGINE_init(impl))
-			{
-			EVPerr(EVP_F_DO_EVP_ENC_ENGINE, EVP_R_INITIALIZATION_ERROR);
-			return 0;
-			}
-		}
-	else
-		/* Ask if an ENGINE is reserved for this job */
-		impl = ENGINE_get_cipher_engine((*pcipher)->nid);
-	if(impl)
-		{
-		/* There's an ENGINE for this job ... (apparently) */
-		const EVP_CIPHER *c = ENGINE_get_cipher(impl, (*pcipher)->nid);
-		if(!c)
-			{
-			/* One positive side-effect of US's export
-			 * control history, is that we should at least
-			 * be able to avoid using US mispellings of
-			 * "initialisation"? */
-			EVPerr(EVP_F_DO_EVP_ENC_ENGINE, EVP_R_INITIALIZATION_ERROR);
-			return 0;
-			}
-		/* We'll use the ENGINE's private cipher definition */
-		*pcipher = c;
-		/* Store the ENGINE functional reference so we know
-		 * 'cipher' came from an ENGINE and we need to release
-		 * it when done. */
-		ctx->engine = impl;
-		}
-	else
-		ctx->engine = NULL;
-	return 1;
-	}
-
-#endif
-
-#endif
-
-int EVP_CipherInit_ex(EVP_CIPHER_CTX *ctx, const EVP_CIPHER *cipher, ENGINE *impl,
-	     const unsigned char *key, const unsigned char *iv, int enc)
-	{
-	if (enc == -1)
-		enc = ctx->encrypt;
-	else
-		{
-		if (enc)
-			enc = 1;
-		ctx->encrypt = enc;
-		}
-#ifdef OPENSSL_FIPS
-	if(FIPS_selftest_failed())
-		{
-		FIPSerr(FIPS_F_EVP_CIPHERINIT_EX,FIPS_R_FIPS_SELFTEST_FAILED);
-		ctx->cipher = &bad_cipher;
-		return 0;
-		}
-#endif
-#ifndef OPENSSL_NO_ENGINE
-	/* Whether it's nice or not, "Inits" can be used on "Final"'d contexts
-	 * so this context may already have an ENGINE! Try to avoid releasing
-	 * the previous handle, re-querying for an ENGINE, and having a
-	 * reinitialisation, when it may all be unecessary. */
-	if (ctx->engine && ctx->cipher && (!cipher ||
-			(cipher && (cipher->nid == ctx->cipher->nid))))
-		goto skip_to_init;
-#endif
-	if (cipher)
-		{
-		/* Ensure a context left lying around from last time is cleared
-		 * (the previous check attempted to avoid this if the same
-		 * ENGINE and EVP_CIPHER could be used). */
-		EVP_CIPHER_CTX_cleanup(ctx);
-
-		/* Restore encrypt field: it is zeroed by cleanup */
-		ctx->encrypt = enc;
-#ifndef OPENSSL_NO_ENGINE
-		if (!do_evp_enc_engine(ctx, &cipher, impl))
-			return 0;
-#endif
-
-		ctx->cipher=cipher;
-		if (ctx->cipher->ctx_size)
-			{
-			ctx->cipher_data=OPENSSL_malloc(ctx->cipher->ctx_size);
-			if (!ctx->cipher_data)
-				{
-				EVPerr(EVP_F_EVP_CIPHERINIT_EX, ERR_R_MALLOC_FAILURE);
-				return 0;
-				}
-			}
-		else
-			{
-			ctx->cipher_data = NULL;
-			}
-		ctx->key_len = cipher->key_len;
-		ctx->flags = 0;
-		if(ctx->cipher->flags & EVP_CIPH_CTRL_INIT)
-			{
-			if(!EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_INIT, 0, NULL))
-				{
-				EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_INITIALIZATION_ERROR);
-				return 0;
-				}
-			}
-		}
-	else if(!ctx->cipher)
-		{
-		EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_NO_CIPHER_SET);
-		return 0;
-		}
-#ifndef OPENSSL_NO_ENGINE
-skip_to_init:
-#endif
-	/* we assume block size is a power of 2 in *cryptUpdate */
-	OPENSSL_assert(ctx->cipher->block_size == 1
-	    || ctx->cipher->block_size == 8
-	    || ctx->cipher->block_size == 16);
-
-	if(!(EVP_CIPHER_CTX_flags(ctx) & EVP_CIPH_CUSTOM_IV)) {
-		switch(EVP_CIPHER_CTX_mode(ctx)) {
-
-			case EVP_CIPH_STREAM_CIPHER:
-			case EVP_CIPH_ECB_MODE:
-			break;
-
-			case EVP_CIPH_CFB_MODE:
-			case EVP_CIPH_OFB_MODE:
-
-			ctx->num = 0;
-			/* fall-through */
-
-			case EVP_CIPH_CBC_MODE:
-
-			OPENSSL_assert(EVP_CIPHER_CTX_iv_length(ctx) <=
-					(int)sizeof(ctx->iv));
-			if(iv) memcpy(ctx->oiv, iv, EVP_CIPHER_CTX_iv_length(ctx));
-			memcpy(ctx->iv, ctx->oiv, EVP_CIPHER_CTX_iv_length(ctx));
-			break;
-
-			default:
-			return 0;
-			break;
-		}
-	}
-
-#ifdef OPENSSL_FIPS
-	/* After 'key' is set no further parameters changes are permissible.
-	 * So only check for non FIPS enabling at this point.
-	 */
-	if (key && FIPS_mode())
-		{
-		if (!(ctx->cipher->flags & EVP_CIPH_FLAG_FIPS)
-			& !(ctx->flags & EVP_CIPH_FLAG_NON_FIPS_ALLOW))
-			{
-			EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_DISABLED_FOR_FIPS);
-#if 0
-			ERR_add_error_data(2, "cipher=",
-						EVP_CIPHER_name(ctx->cipher));
-#endif
-			ctx->cipher = &bad_cipher;
-			return 0;
-			}
-		}
-#endif
-
-	if(key || (ctx->cipher->flags & EVP_CIPH_ALWAYS_CALL_INIT)) {
-		if(!ctx->cipher->init(ctx,key,iv,enc)) return 0;
-	}
-	ctx->buf_len=0;
-	ctx->final_used=0;
-	ctx->block_mask=ctx->cipher->block_size-1;
-	return 1;
-	}
-
-int EVP_CIPHER_CTX_cleanup(EVP_CIPHER_CTX *c)
-	{
-	if (c->cipher != NULL)
-		{
-		if(c->cipher->cleanup && !c->cipher->cleanup(c))
-			return 0;
-		/* Cleanse cipher context data */
-		if (c->cipher_data)
-			OPENSSL_cleanse(c->cipher_data, c->cipher->ctx_size);
-		}
-	if (c->cipher_data)
-		OPENSSL_free(c->cipher_data);
-#ifndef OPENSSL_NO_ENGINE
-	if (c->engine)
-		/* The EVP_CIPHER we used belongs to an ENGINE, release the
-		 * functional reference we held for this reason. */
-		do_engine_finish(c->engine);
-#endif
-	memset(c,0,sizeof(EVP_CIPHER_CTX));
-	return 1;
-	}
-
-int EVP_Cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, unsigned int inl)
-	{
-#ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
-	return ctx->cipher->do_cipher(ctx,out,in,inl);
-	}
-
-int EVP_CIPHER_CTX_ctrl(EVP_CIPHER_CTX *ctx, int type, int arg, void *ptr)
-{
-	int ret;
-	if(!ctx->cipher) {
-		EVPerr(EVP_F_EVP_CIPHER_CTX_CTRL, EVP_R_NO_CIPHER_SET);
-		return 0;
-	}
-
-	if(!ctx->cipher->ctrl) {
-		EVPerr(EVP_F_EVP_CIPHER_CTX_CTRL, EVP_R_CTRL_NOT_IMPLEMENTED);
-		return 0;
-	}
-
-	ret = ctx->cipher->ctrl(ctx, type, arg, ptr);
-	if(ret == -1) {
-		EVPerr(EVP_F_EVP_CIPHER_CTX_CTRL, EVP_R_CTRL_OPERATION_NOT_IMPLEMENTED);
-		return 0;
-	}
-	return ret;
-}
-
-unsigned long EVP_CIPHER_CTX_flags(const EVP_CIPHER_CTX *ctx)
-	{
-	return ctx->cipher->flags;
-	}
-
-int EVP_CIPHER_CTX_iv_length(const EVP_CIPHER_CTX *ctx)
-	{
-	return ctx->cipher->iv_len;
-	}
-
-int EVP_CIPHER_nid(const EVP_CIPHER *cipher)
-	{
-	return cipher->nid;
-	}
diff --git a/crypto/evp/encode.c b/crypto/evp/encode.c
index e8a521814a5c..28546a84bc2b 100644
--- a/crypto/evp/encode.c
+++ b/crypto/evp/encode.c
@@ -85,7 +85,7 @@
 #define CHUNKS_PER_LINE (64/4)
 #define CHAR_PER_LINE   (64+1)
 
-static unsigned char data_bin2ascii[65]="ABCDEFGHIJKLMNOPQRSTUVWXYZ\
+static const unsigned char data_bin2ascii[65]="ABCDEFGHIJKLMNOPQRSTUVWXYZ\
 abcdefghijklmnopqrstuvwxyz0123456789+/";
 
 /* 0xF0 is a EOLN
@@ -102,7 +102,7 @@ abcdefghijklmnopqrstuvwxyz0123456789+/";
 #define B64_ERROR       	0xFF
 #define B64_NOT_BASE64(a)	(((a)|0x13) == 0xF3)
 
-static unsigned char data_ascii2bin[128]={
+static const unsigned char data_ascii2bin[128]={
 	0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
 	0xFF,0xE0,0xF0,0xFF,0xFF,0xF1,0xFF,0xFF,
 	0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,
diff --git a/crypto/evp/evp.h b/crypto/evp/evp.h
index 79c097181f48..0d1b20a7d38a 100644
--- a/crypto/evp/evp.h
+++ b/crypto/evp/evp.h
@@ -75,10 +75,6 @@
 #include <openssl/bio.h>
 #endif
 
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 /*
 #define EVP_RC2_KEY_SIZE		16
 #define EVP_RC4_KEY_SIZE		16
@@ -87,7 +83,7 @@
 #define EVP_RC5_32_12_16_KEY_SIZE	16
 */
 #define EVP_MAX_MD_SIZE			64	/* longest known is SHA512 */
-#define EVP_MAX_KEY_LENGTH		32
+#define EVP_MAX_KEY_LENGTH		64
 #define EVP_MAX_IV_LENGTH		16
 #define EVP_MAX_BLOCK_LENGTH		32
 
@@ -119,6 +115,8 @@
 #define EVP_PKEY_DSA4	NID_dsaWithSHA1_2
 #define EVP_PKEY_DH	NID_dhKeyAgreement
 #define EVP_PKEY_EC	NID_X9_62_id_ecPublicKey
+#define EVP_PKEY_HMAC	NID_hmac
+#define EVP_PKEY_CMAC	NID_cmac
 
 #ifdef	__cplusplus
 extern "C" {
@@ -132,6 +130,8 @@ struct evp_pkey_st
 	int type;
 	int save_type;
 	int references;
+	const EVP_PKEY_ASN1_METHOD *ameth;
+	ENGINE *engine;
 	union	{
 		char *ptr;
 #ifndef OPENSSL_NO_RSA
@@ -156,73 +156,6 @@ struct evp_pkey_st
 #define EVP_PKEY_MO_ENCRYPT	0x0004
 #define EVP_PKEY_MO_DECRYPT	0x0008
 
-#if 0
-/* This structure is required to tie the message digest and signing together.
- * The lookup can be done by md/pkey_method, oid, oid/pkey_method, or
- * oid, md and pkey.
- * This is required because for various smart-card perform the digest and
- * signing/verification on-board.  To handle this case, the specific
- * EVP_MD and EVP_PKEY_METHODs need to be closely associated.
- * When a PKEY is created, it will have a EVP_PKEY_METHOD associated with it.
- * This can either be software or a token to provide the required low level
- * routines.
- */
-typedef struct evp_pkey_md_st
-	{
-	int oid;
-	EVP_MD *md;
-	EVP_PKEY_METHOD *pkey;
-	} EVP_PKEY_MD;
-
-#define EVP_rsa_md2() \
-		EVP_PKEY_MD_add(NID_md2WithRSAEncryption,\
-			EVP_rsa_pkcs1(),EVP_md2())
-#define EVP_rsa_md5() \
-		EVP_PKEY_MD_add(NID_md5WithRSAEncryption,\
-			EVP_rsa_pkcs1(),EVP_md5())
-#define EVP_rsa_sha0() \
-		EVP_PKEY_MD_add(NID_shaWithRSAEncryption,\
-			EVP_rsa_pkcs1(),EVP_sha())
-#define EVP_rsa_sha1() \
-		EVP_PKEY_MD_add(NID_sha1WithRSAEncryption,\
-			EVP_rsa_pkcs1(),EVP_sha1())
-#define EVP_rsa_ripemd160() \
-		EVP_PKEY_MD_add(NID_ripemd160WithRSA,\
-			EVP_rsa_pkcs1(),EVP_ripemd160())
-#define EVP_rsa_mdc2() \
-		EVP_PKEY_MD_add(NID_mdc2WithRSA,\
-			EVP_rsa_octet_string(),EVP_mdc2())
-#define EVP_dsa_sha() \
-		EVP_PKEY_MD_add(NID_dsaWithSHA,\
-			EVP_dsa(),EVP_sha())
-#define EVP_dsa_sha1() \
-		EVP_PKEY_MD_add(NID_dsaWithSHA1,\
-			EVP_dsa(),EVP_sha1())
-
-typedef struct evp_pkey_method_st
-	{
-	char *name;
-	int flags;
-	int type;		/* RSA, DSA, an SSLeay specific constant */
-	int oid;		/* For the pub-key type */
-	int encrypt_oid;	/* pub/priv key encryption */
-
-	int (*sign)();
-	int (*verify)();
-	struct	{
-		int (*set)();	/* get and/or set the underlying type */
-		int (*get)();
-		int (*encrypt)();
-		int (*decrypt)();
-		int (*i2d)();
-		int (*d2i)();
-		int (*dup)();
-		} pub,priv;
-	int (*set_asn1_parameters)();
-	int (*get_asn1_parameters)();
-	} EVP_PKEY_METHOD;
-#endif
-
 #ifndef EVP_MD
 struct env_md_st
 	{
@@ -245,6 +178,8 @@ struct env_md_st
 	int required_pkey_type[5]; /*EVP_PKEY_xxx */
 	int block_size;
 	int ctx_size; /* how big does the ctx->md_data need to be */
+	/* control function */
+	int (*md_ctrl)(EVP_MD_CTX *ctx, int cmd, int p1, void *p2);
 	} /* EVP_MD */;
 
 typedef int evp_sign_method(int type,const unsigned char *m,
@@ -254,18 +189,44 @@ typedef int evp_verify_method(int type,const unsigned char *m,
 			    unsigned int m_length,const unsigned char *sigbuf,
 			    unsigned int siglen, void *key);
 
-typedef struct
-	{
-	EVP_MD_CTX *mctx;
-	void *key;
-	} EVP_MD_SVCTX;
-
 #define EVP_MD_FLAG_ONESHOT	0x0001 /* digest can only handle a single
 					* block */
 
+#define EVP_MD_FLAG_PKEY_DIGEST	0x0002 /* digest is a "clone" digest used
+					* which is a copy of an existing
+					* one for a specific public key type.
+					* EVP_dss1() etc */
+
+/* Digest uses EVP_PKEY_METHOD for signing instead of MD specific signing */
+
+#define EVP_MD_FLAG_PKEY_METHOD_SIGNATURE	0x0004
+
+/* DigestAlgorithmIdentifier flags... */
+
+#define EVP_MD_FLAG_DIGALGID_MASK		0x0018
+
+/* NULL or absent parameter accepted. Use NULL */
+
+#define EVP_MD_FLAG_DIGALGID_NULL		0x0000
+
+/* NULL or absent parameter accepted. Use NULL for PKCS#1 otherwise absent */
+
+#define EVP_MD_FLAG_DIGALGID_ABSENT		0x0008
+
+/* Custom handling via ctrl */
+
+#define EVP_MD_FLAG_DIGALGID_CUSTOM		0x0018
+
 #define EVP_MD_FLAG_FIPS	0x0400 /* Note if suitable for use in FIPS mode */
 
-#define EVP_MD_FLAG_SVCTX	0x0800 /* pass EVP_MD_SVCTX to sign/verify */
+/* Digest ctrls */
+
+#define	EVP_MD_CTRL_DIGALGID			0x1
+#define	EVP_MD_CTRL_MICALG			0x2
+
+/* Minimum Algorithm specific ctrl value */
+
+#define	EVP_MD_CTRL_ALG_CTRL			0x1000
 
 #define EVP_PKEY_NULL_method	NULL,NULL,{0,0,0,0}
 
@@ -307,6 +268,10 @@ struct env_md_ctx_st
 	ENGINE *engine; /* functional reference if 'digest' is ENGINE-provided */
 	unsigned long flags;
 	void *md_data;
+	/* Public key context for sign/verify */
+	EVP_PKEY_CTX *pctx;
+	/* Update function: usually copied from EVP_MD */
+	int (*update)(EVP_MD_CTX *ctx,const void *data,size_t count);
 	} /* EVP_MD_CTX */;
 
 /* values for EVP_MD_CTX flags */
@@ -317,17 +282,23 @@ struct env_md_ctx_st
 						* cleaned */
 #define EVP_MD_CTX_FLAG_REUSE		0x0004 /* Don't free up ctx->md_data
 						* in EVP_MD_CTX_cleanup */
+/* FIPS and pad options are ignored in 1.0.0, definitions are here
+ * so we don't accidentally reuse the values for other purposes.
+ */
+
 #define EVP_MD_CTX_FLAG_NON_FIPS_ALLOW	0x0008	/* Allow use of non FIPS digest
 						 * in FIPS mode */
 
+/* The following PAD options are also currently ignored in 1.0.0, digest
+ * parameters are handled through EVP_DigestSign*() and EVP_DigestVerify*()
+ * instead.
+ */
 #define EVP_MD_CTX_FLAG_PAD_MASK	0xF0	/* RSA mode to use */
 #define EVP_MD_CTX_FLAG_PAD_PKCS1	0x00	/* PKCS#1 v1.5 mode */
 #define EVP_MD_CTX_FLAG_PAD_X931	0x10	/* X9.31 mode */
 #define EVP_MD_CTX_FLAG_PAD_PSS		0x20	/* PSS mode */
-#define M_EVP_MD_CTX_FLAG_PSS_SALT(ctx) \
-		((ctx->flags>>16) &0xFFFF) /* seed length */
-#define EVP_MD_CTX_FLAG_PSS_MDLEN	0xFFFF	/* salt len same as digest */
-#define EVP_MD_CTX_FLAG_PSS_MREC	0xFFFE	/* salt max or auto recovered */
+
+#define EVP_MD_CTX_FLAG_NO_INIT		0x0100 /* Don't initialize md_data */
 
 struct evp_cipher_st
 	{
@@ -339,7 +310,7 @@ struct evp_cipher_st
 	int (*init)(EVP_CIPHER_CTX *ctx, const unsigned char *key,
 		    const unsigned char *iv, int enc);	/* init key */
 	int (*do_cipher)(EVP_CIPHER_CTX *ctx, unsigned char *out,
-			 const unsigned char *in, unsigned int inl);/* encrypt/decrypt data */
+			 const unsigned char *in, size_t inl);/* encrypt/decrypt data */
 	int (*cleanup)(EVP_CIPHER_CTX *); /* cleanup ctx */
 	int ctx_size;		/* how big ctx->cipher_data needs to be */
 	int (*set_asn1_parameters)(EVP_CIPHER_CTX *, ASN1_TYPE *); /* Populate a ASN1_TYPE with parameters */
@@ -357,7 +328,11 @@ struct evp_cipher_st
 #define		EVP_CIPH_CBC_MODE		0x2
 #define		EVP_CIPH_CFB_MODE		0x3
 #define		EVP_CIPH_OFB_MODE		0x4
-#define 	EVP_CIPH_MODE			0x7
+#define		EVP_CIPH_CTR_MODE		0x5
+#define		EVP_CIPH_GCM_MODE		0x6
+#define		EVP_CIPH_CCM_MODE		0x7
+#define		EVP_CIPH_XTS_MODE		0x10001
+#define 	EVP_CIPH_MODE			0xF0007
 /* Set if variable length cipher */
 #define 	EVP_CIPH_VARIABLE_LENGTH	0x8
 /* Set if the iv handling should be done by the cipher itself */
@@ -372,14 +347,21 @@ struct evp_cipher_st
 #define 	EVP_CIPH_NO_PADDING		0x100
 /* cipher handles random key generation */
 #define 	EVP_CIPH_RAND_KEY		0x200
-/* Note if suitable for use in FIPS mode */
-#define		EVP_CIPH_FLAG_FIPS		0x400
-/* Allow non FIPS cipher in FIPS mode */
-#define		EVP_CIPH_FLAG_NON_FIPS_ALLOW	0x800
+/* cipher has its own additional copying logic */
+#define 	EVP_CIPH_CUSTOM_COPY		0x400
 /* Allow use default ASN1 get/set iv */
 #define		EVP_CIPH_FLAG_DEFAULT_ASN1	0x1000
 /* Buffer length in bits not bytes: CFB1 mode only */
 #define		EVP_CIPH_FLAG_LENGTH_BITS	0x2000
+/* Note if suitable for use in FIPS mode */
+#define		EVP_CIPH_FLAG_FIPS		0x4000
+/* Allow non FIPS cipher in FIPS mode */
+#define		EVP_CIPH_FLAG_NON_FIPS_ALLOW	0x8000
+/* Cipher handles any and all padding logic as well
+ * as finalisation.
+ */
+#define 	EVP_CIPH_FLAG_CUSTOM_CIPHER	0x100000
+#define		EVP_CIPH_FLAG_AEAD_CIPHER	0x200000
 
 /* ctrl() values */
 
@@ -390,6 +372,36 @@ struct evp_cipher_st
 #define 	EVP_CTRL_GET_RC5_ROUNDS		0x4
 #define 	EVP_CTRL_SET_RC5_ROUNDS		0x5
 #define 	EVP_CTRL_RAND_KEY		0x6
+#define 	EVP_CTRL_PBE_PRF_NID		0x7
+#define 	EVP_CTRL_COPY			0x8
+#define 	EVP_CTRL_GCM_SET_IVLEN		0x9
+#define 	EVP_CTRL_GCM_GET_TAG		0x10
+#define 	EVP_CTRL_GCM_SET_TAG		0x11
+#define		EVP_CTRL_GCM_SET_IV_FIXED	0x12
+#define		EVP_CTRL_GCM_IV_GEN		0x13
+#define		EVP_CTRL_CCM_SET_IVLEN		EVP_CTRL_GCM_SET_IVLEN
+#define		EVP_CTRL_CCM_GET_TAG		EVP_CTRL_GCM_GET_TAG
+#define		EVP_CTRL_CCM_SET_TAG		EVP_CTRL_GCM_SET_TAG
+#define		EVP_CTRL_CCM_SET_L		0x14
+#define		EVP_CTRL_CCM_SET_MSGLEN		0x15
+/* AEAD cipher deduces payload length and returns number of bytes
+ * required to store MAC and eventual padding. Subsequent call to
+ * EVP_Cipher even appends/verifies MAC.
+ */
+#define		EVP_CTRL_AEAD_TLS1_AAD		0x16
+/* Used by composite AEAD ciphers, no-op in GCM, CCM... */
+#define		EVP_CTRL_AEAD_SET_MAC_KEY	0x17
+/* Set the GCM invocation field, decrypt only */
+#define		EVP_CTRL_GCM_SET_IV_INV		0x18
+
+/* GCM TLS constants */
+/* Length of fixed part of IV derived from PRF */
+#define EVP_GCM_TLS_FIXED_IV_LEN			4
+/* Length of explicit part of IV part of TLS records */
+#define EVP_GCM_TLS_EXPLICIT_IV_LEN			8
+/* Length of tag for TLS */
+#define EVP_GCM_TLS_TAG_LEN				16
+
 
 typedef struct evp_cipher_info_st
 	{
@@ -407,7 +419,7 @@ struct evp_cipher_ctx_st
 	unsigned char  oiv[EVP_MAX_IV_LENGTH];	/* original iv */
 	unsigned char  iv[EVP_MAX_IV_LENGTH];	/* working iv */
 	unsigned char buf[EVP_MAX_BLOCK_LENGTH];/* saved partial block */
-	int num;				/* used by cfb/ofb mode */
+	int num;				/* used by cfb/ofb/ctr mode */
 
 	void *app_data;		/* application stuff */
 	int key_len;		/* May change for variable length cipher */
@@ -462,26 +474,15 @@ typedef int (EVP_PBE_KEYGEN)(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
 #define EVP_get_cipherbynid(a) EVP_get_cipherbyname(OBJ_nid2sn(a))
 #define EVP_get_cipherbyobj(a) EVP_get_cipherbynid(OBJ_obj2nid(a))
 
-/* Macros to reduce FIPS dependencies: do NOT use in applications */
-#define M_EVP_MD_size(e)		((e)->md_size)
-#define M_EVP_MD_block_size(e)		((e)->block_size)
-#define M_EVP_MD_CTX_set_flags(ctx,flgs) ((ctx)->flags|=(flgs))
-#define M_EVP_MD_CTX_clear_flags(ctx,flgs) ((ctx)->flags&=~(flgs))
-#define M_EVP_MD_CTX_test_flags(ctx,flgs) ((ctx)->flags&(flgs))
-#define M_EVP_MD_type(e)			((e)->type)
-#define M_EVP_MD_CTX_type(e)		M_EVP_MD_type(M_EVP_MD_CTX_md(e))
-#define M_EVP_MD_CTX_md(e)			((e)->digest)
-
-#define M_EVP_CIPHER_CTX_set_flags(ctx,flgs) ((ctx)->flags|=(flgs))
-
 int EVP_MD_type(const EVP_MD *md);
 #define EVP_MD_nid(e)			EVP_MD_type(e)
 #define EVP_MD_name(e)			OBJ_nid2sn(EVP_MD_nid(e))
 int EVP_MD_pkey_type(const EVP_MD *md);	
 int EVP_MD_size(const EVP_MD *md);
 int EVP_MD_block_size(const EVP_MD *md);
+unsigned long EVP_MD_flags(const EVP_MD *md);
 
-const EVP_MD * EVP_MD_CTX_md(const EVP_MD_CTX *ctx);
+const EVP_MD *EVP_MD_CTX_md(const EVP_MD_CTX *ctx);
 #define EVP_MD_CTX_size(e)		EVP_MD_size(EVP_MD_CTX_md(e))
 #define EVP_MD_CTX_block_size(e)	EVP_MD_block_size(EVP_MD_CTX_md(e))
 #define EVP_MD_CTX_type(e)		EVP_MD_type(EVP_MD_CTX_md(e))
@@ -499,6 +500,7 @@ int EVP_CIPHER_CTX_nid(const EVP_CIPHER_CTX *ctx);
 int EVP_CIPHER_CTX_block_size(const EVP_CIPHER_CTX *ctx);
 int EVP_CIPHER_CTX_key_length(const EVP_CIPHER_CTX *ctx);
 int EVP_CIPHER_CTX_iv_length(const EVP_CIPHER_CTX *ctx);
+int EVP_CIPHER_CTX_copy(EVP_CIPHER_CTX *out, const EVP_CIPHER_CTX *in);
 void * EVP_CIPHER_CTX_get_app_data(const EVP_CIPHER_CTX *ctx);
 void EVP_CIPHER_CTX_set_app_data(EVP_CIPHER_CTX *ctx, void *data);
 #define EVP_CIPHER_CTX_type(c)         EVP_CIPHER_type(EVP_CIPHER_CTX_cipher(c))
@@ -516,6 +518,8 @@ unsigned long EVP_CIPHER_CTX_flags(const EVP_CIPHER_CTX *ctx);
 #define	EVP_VerifyUpdate(a,b,c)		EVP_DigestUpdate(a,b,c)
 #define EVP_OpenUpdate(a,b,c,d,e)	EVP_DecryptUpdate(a,b,c,d,e)
 #define EVP_SealUpdate(a,b,c,d,e)	EVP_EncryptUpdate(a,b,c,d,e)	
+#define EVP_DigestSignUpdate(a,b,c)	EVP_DigestUpdate(a,b,c)
+#define EVP_DigestVerifyUpdate(a,b,c)	EVP_DigestUpdate(a,b,c)
 
 #ifdef CONST_STRICT
 void BIO_set_md(BIO *,const EVP_MD *md);
@@ -562,6 +566,7 @@ int	EVP_DigestInit(EVP_MD_CTX *ctx, const EVP_MD *type);
 int	EVP_DigestFinal(EVP_MD_CTX *ctx,unsigned char *md,unsigned int *s);
 
 int	EVP_read_pw_string(char *buf,int length,const char *prompt,int verify);
+int	EVP_read_pw_string_min(char *buf,int minlen,int maxlen,const char *prompt,int verify);
 void	EVP_set_pw_prompt(const char *prompt);
 char *	EVP_get_pw_prompt(void);
 
@@ -608,6 +613,16 @@ int	EVP_SignFinal(EVP_MD_CTX *ctx,unsigned char *md,unsigned int *s,
 int	EVP_VerifyFinal(EVP_MD_CTX *ctx,const unsigned char *sigbuf,
 		unsigned int siglen,EVP_PKEY *pkey);
 
+int	EVP_DigestSignInit(EVP_MD_CTX *ctx, EVP_PKEY_CTX **pctx,
+			const EVP_MD *type, ENGINE *e, EVP_PKEY *pkey);
+int	EVP_DigestSignFinal(EVP_MD_CTX *ctx,
+			unsigned char *sigret, size_t *siglen);
+
+int	EVP_DigestVerifyInit(EVP_MD_CTX *ctx, EVP_PKEY_CTX **pctx,
+			const EVP_MD *type, ENGINE *e, EVP_PKEY *pkey);
+int	EVP_DigestVerifyFinal(EVP_MD_CTX *ctx,
+			unsigned char *sig, size_t siglen);
+
 int	EVP_OpenInit(EVP_CIPHER_CTX *ctx,const EVP_CIPHER *type,
 		const unsigned char *ek, int ekl, const unsigned char *iv,
 		EVP_PKEY *priv);
@@ -680,6 +695,9 @@ const EVP_MD *EVP_mdc2(void);
 #ifndef OPENSSL_NO_RIPEMD
 const EVP_MD *EVP_ripemd160(void);
 #endif
+#ifndef OPENSSL_NO_WHIRLPOOL
+const EVP_MD *EVP_whirlpool(void);
+#endif
 const EVP_CIPHER *EVP_enc_null(void);		/* does nothing :-) */
 #ifndef OPENSSL_NO_DES
 const EVP_CIPHER *EVP_des_ecb(void);
@@ -721,6 +739,9 @@ const EVP_MD *EVP_dev_crypto_md5(void);
 #ifndef OPENSSL_NO_RC4
 const EVP_CIPHER *EVP_rc4(void);
 const EVP_CIPHER *EVP_rc4_40(void);
+#ifndef OPENSSL_NO_MD5
+const EVP_CIPHER *EVP_rc4_hmac_md5(void);
+#endif
 #endif
 #ifndef OPENSSL_NO_IDEA
 const EVP_CIPHER *EVP_idea_ecb(void);
@@ -767,9 +788,10 @@ const EVP_CIPHER *EVP_aes_128_cfb8(void);
 const EVP_CIPHER *EVP_aes_128_cfb128(void);
 # define EVP_aes_128_cfb EVP_aes_128_cfb128
 const EVP_CIPHER *EVP_aes_128_ofb(void);
-#if 0
 const EVP_CIPHER *EVP_aes_128_ctr(void);
-#endif
+const EVP_CIPHER *EVP_aes_128_gcm(void);
+const EVP_CIPHER *EVP_aes_128_ccm(void);
+const EVP_CIPHER *EVP_aes_128_xts(void);
 const EVP_CIPHER *EVP_aes_192_ecb(void);
 const EVP_CIPHER *EVP_aes_192_cbc(void);
 const EVP_CIPHER *EVP_aes_192_cfb1(void);
@@ -777,9 +799,9 @@ const EVP_CIPHER *EVP_aes_192_cfb8(void);
 const EVP_CIPHER *EVP_aes_192_cfb128(void);
 # define EVP_aes_192_cfb EVP_aes_192_cfb128
 const EVP_CIPHER *EVP_aes_192_ofb(void);
-#if 0
 const EVP_CIPHER *EVP_aes_192_ctr(void);
-#endif
+const EVP_CIPHER *EVP_aes_192_gcm(void);
+const EVP_CIPHER *EVP_aes_192_ccm(void);
 const EVP_CIPHER *EVP_aes_256_ecb(void);
 const EVP_CIPHER *EVP_aes_256_cbc(void);
 const EVP_CIPHER *EVP_aes_256_cfb1(void);
@@ -787,8 +809,13 @@ const EVP_CIPHER *EVP_aes_256_cfb8(void);
 const EVP_CIPHER *EVP_aes_256_cfb128(void);
 # define EVP_aes_256_cfb EVP_aes_256_cfb128
 const EVP_CIPHER *EVP_aes_256_ofb(void);
-#if 0
 const EVP_CIPHER *EVP_aes_256_ctr(void);
+const EVP_CIPHER *EVP_aes_256_gcm(void);
+const EVP_CIPHER *EVP_aes_256_ccm(void);
+const EVP_CIPHER *EVP_aes_256_xts(void);
+#if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA1)
+const EVP_CIPHER *EVP_aes_128_cbc_hmac_sha1(void);
+const EVP_CIPHER *EVP_aes_256_cbc_hmac_sha1(void);
 #endif
 #endif
 #ifndef OPENSSL_NO_CAMELLIA
@@ -847,16 +874,31 @@ const EVP_CIPHER *EVP_get_cipherbyname(const char *name);
 const EVP_MD *EVP_get_digestbyname(const char *name);
 void EVP_cleanup(void);
 
-int		EVP_PKEY_decrypt(unsigned char *dec_key,
+void EVP_CIPHER_do_all(void (*fn)(const EVP_CIPHER *ciph,
+		const char *from, const char *to, void *x), void *arg);
+void EVP_CIPHER_do_all_sorted(void (*fn)(const EVP_CIPHER *ciph,
+		const char *from, const char *to, void *x), void *arg);
+
+void EVP_MD_do_all(void (*fn)(const EVP_MD *ciph,
+		const char *from, const char *to, void *x), void *arg);
+void EVP_MD_do_all_sorted(void (*fn)(const EVP_MD *ciph,
+		const char *from, const char *to, void *x), void *arg);
+
+int		EVP_PKEY_decrypt_old(unsigned char *dec_key,
 			const unsigned char *enc_key,int enc_key_len,
 			EVP_PKEY *private_key);
-int		EVP_PKEY_encrypt(unsigned char *enc_key,
+int		EVP_PKEY_encrypt_old(unsigned char *enc_key,
 			const unsigned char *key,int key_len,
 			EVP_PKEY *pub_key);
 int		EVP_PKEY_type(int type);
+int		EVP_PKEY_id(const EVP_PKEY *pkey);
+int		EVP_PKEY_base_id(const EVP_PKEY *pkey);
 int		EVP_PKEY_bits(EVP_PKEY *pkey);
 int		EVP_PKEY_size(EVP_PKEY *pkey);
-int 		EVP_PKEY_assign(EVP_PKEY *pkey,int type,char *key);
+int 		EVP_PKEY_set_type(EVP_PKEY *pkey,int type);
+int		EVP_PKEY_set_type_str(EVP_PKEY *pkey, const char *str, int len);
+int 		EVP_PKEY_assign(EVP_PKEY *pkey,int type,void *key);
+void *		EVP_PKEY_get0(EVP_PKEY *pkey);
 
 #ifndef OPENSSL_NO_RSA
 struct rsa_st;
@@ -899,6 +941,15 @@ int EVP_PKEY_cmp_parameters(const EVP_PKEY *a, const EVP_PKEY *b);
 
 int EVP_PKEY_cmp(const EVP_PKEY *a, const EVP_PKEY *b);
 
+int EVP_PKEY_print_public(BIO *out, const EVP_PKEY *pkey,
+				int indent, ASN1_PCTX *pctx);
+int EVP_PKEY_print_private(BIO *out, const EVP_PKEY *pkey,
+				int indent, ASN1_PCTX *pctx);
+int EVP_PKEY_print_params(BIO *out, const EVP_PKEY *pkey,
+				int indent, ASN1_PCTX *pctx);
+
+int EVP_PKEY_get_default_digest_nid(EVP_PKEY *pkey, int *pnid);
+
 int EVP_CIPHER_type(const EVP_CIPHER *ctx);
 
 /* calls methods */
@@ -916,6 +967,10 @@ int PKCS5_PBE_keyivgen(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
 int PKCS5_PBKDF2_HMAC_SHA1(const char *pass, int passlen,
 			   const unsigned char *salt, int saltlen, int iter,
 			   int keylen, unsigned char *out);
+int PKCS5_PBKDF2_HMAC(const char *pass, int passlen,
+			   const unsigned char *salt, int saltlen, int iter,
+			   const EVP_MD *digest,
+		      int keylen, unsigned char *out);
 int PKCS5_v2_PBE_keyivgen(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
 			 ASN1_TYPE *param, const EVP_CIPHER *cipher, const EVP_MD *md,
 			 int en_de);
@@ -924,27 +979,269 @@ void PKCS5_PBE_add(void);
 
 int EVP_PBE_CipherInit (ASN1_OBJECT *pbe_obj, const char *pass, int passlen,
 	     ASN1_TYPE *param, EVP_CIPHER_CTX *ctx, int en_de);
+
+/* PBE type */
+
+/* Can appear as the outermost AlgorithmIdentifier */
+#define EVP_PBE_TYPE_OUTER	0x0
+/* Is an PRF type OID */
+#define EVP_PBE_TYPE_PRF	0x1
+
+int EVP_PBE_alg_add_type(int pbe_type, int pbe_nid, int cipher_nid, int md_nid,
+	     EVP_PBE_KEYGEN *keygen);
 int EVP_PBE_alg_add(int nid, const EVP_CIPHER *cipher, const EVP_MD *md,
 		    EVP_PBE_KEYGEN *keygen);
+int EVP_PBE_find(int type, int pbe_nid,
+			int *pcnid, int *pmnid, EVP_PBE_KEYGEN **pkeygen);
 void EVP_PBE_cleanup(void);
 
-#ifdef OPENSSL_FIPS
-#ifndef OPENSSL_NO_ENGINE
-void int_EVP_MD_set_engine_callbacks(
-	int (*eng_md_init)(ENGINE *impl),
-	int (*eng_md_fin)(ENGINE *impl),
-	int (*eng_md_evp)
-		(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl));
-void int_EVP_MD_init_engine_callbacks(void);
-void int_EVP_CIPHER_set_engine_callbacks(
-	int (*eng_ciph_fin)(ENGINE *impl),
-	int (*eng_ciph_evp)
-		(EVP_CIPHER_CTX *ctx, const EVP_CIPHER **pciph, ENGINE *impl));
-void int_EVP_CIPHER_init_engine_callbacks(void);
-#endif
-#endif
-
-void EVP_add_alg_module(void);
+#define ASN1_PKEY_ALIAS		0x1
+#define ASN1_PKEY_DYNAMIC	0x2
+#define ASN1_PKEY_SIGPARAM_NULL	0x4
+
+#define ASN1_PKEY_CTRL_PKCS7_SIGN	0x1
+#define ASN1_PKEY_CTRL_PKCS7_ENCRYPT	0x2
+#define ASN1_PKEY_CTRL_DEFAULT_MD_NID	0x3
+#define ASN1_PKEY_CTRL_CMS_SIGN		0x5
+#define ASN1_PKEY_CTRL_CMS_ENVELOPE	0x7
+
+int EVP_PKEY_asn1_get_count(void);
+const EVP_PKEY_ASN1_METHOD *EVP_PKEY_asn1_get0(int idx);
+const EVP_PKEY_ASN1_METHOD *EVP_PKEY_asn1_find(ENGINE **pe, int type);
+const EVP_PKEY_ASN1_METHOD *EVP_PKEY_asn1_find_str(ENGINE **pe,
+					const char *str, int len);
+int EVP_PKEY_asn1_add0(const EVP_PKEY_ASN1_METHOD *ameth);
+int EVP_PKEY_asn1_add_alias(int to, int from);
+int EVP_PKEY_asn1_get0_info(int *ppkey_id, int *pkey_base_id, int *ppkey_flags,
+				const char **pinfo, const char **ppem_str,
+					const EVP_PKEY_ASN1_METHOD *ameth);
+
+const EVP_PKEY_ASN1_METHOD* EVP_PKEY_get0_asn1(EVP_PKEY *pkey);
+EVP_PKEY_ASN1_METHOD* EVP_PKEY_asn1_new(int id, int flags,
+					const char *pem_str, const char *info);
+void EVP_PKEY_asn1_copy(EVP_PKEY_ASN1_METHOD *dst, 
+			const EVP_PKEY_ASN1_METHOD *src);
+void EVP_PKEY_asn1_free(EVP_PKEY_ASN1_METHOD *ameth);
+void EVP_PKEY_asn1_set_public(EVP_PKEY_ASN1_METHOD *ameth,
+		int (*pub_decode)(EVP_PKEY *pk, X509_PUBKEY *pub),
+		int (*pub_encode)(X509_PUBKEY *pub, const EVP_PKEY *pk),
+		int (*pub_cmp)(const EVP_PKEY *a, const EVP_PKEY *b),
+		int (*pub_print)(BIO *out, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *pctx),
+		int (*pkey_size)(const EVP_PKEY *pk),
+		int (*pkey_bits)(const EVP_PKEY *pk));
+void EVP_PKEY_asn1_set_private(EVP_PKEY_ASN1_METHOD *ameth,
+		int (*priv_decode)(EVP_PKEY *pk, PKCS8_PRIV_KEY_INFO *p8inf),
+		int (*priv_encode)(PKCS8_PRIV_KEY_INFO *p8, const EVP_PKEY *pk),
+		int (*priv_print)(BIO *out, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *pctx));
+void EVP_PKEY_asn1_set_param(EVP_PKEY_ASN1_METHOD *ameth,
+		int (*param_decode)(EVP_PKEY *pkey,
+				const unsigned char **pder, int derlen),
+		int (*param_encode)(const EVP_PKEY *pkey, unsigned char **pder),
+		int (*param_missing)(const EVP_PKEY *pk),
+		int (*param_copy)(EVP_PKEY *to, const EVP_PKEY *from),
+		int (*param_cmp)(const EVP_PKEY *a, const EVP_PKEY *b),
+		int (*param_print)(BIO *out, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *pctx));
+
+void EVP_PKEY_asn1_set_free(EVP_PKEY_ASN1_METHOD *ameth,
+		void (*pkey_free)(EVP_PKEY *pkey));
+void EVP_PKEY_asn1_set_ctrl(EVP_PKEY_ASN1_METHOD *ameth,
+		int (*pkey_ctrl)(EVP_PKEY *pkey, int op,
+							long arg1, void *arg2));
+
+
+#define EVP_PKEY_OP_UNDEFINED		0
+#define EVP_PKEY_OP_PARAMGEN		(1<<1)
+#define EVP_PKEY_OP_KEYGEN		(1<<2)
+#define EVP_PKEY_OP_SIGN		(1<<3)
+#define EVP_PKEY_OP_VERIFY		(1<<4)
+#define EVP_PKEY_OP_VERIFYRECOVER	(1<<5)
+#define EVP_PKEY_OP_SIGNCTX		(1<<6)
+#define EVP_PKEY_OP_VERIFYCTX		(1<<7)
+#define EVP_PKEY_OP_ENCRYPT		(1<<8)
+#define EVP_PKEY_OP_DECRYPT		(1<<9)
+#define EVP_PKEY_OP_DERIVE		(1<<10)
+
+#define EVP_PKEY_OP_TYPE_SIG	\
+	(EVP_PKEY_OP_SIGN | EVP_PKEY_OP_VERIFY | EVP_PKEY_OP_VERIFYRECOVER \
+		| EVP_PKEY_OP_SIGNCTX | EVP_PKEY_OP_VERIFYCTX)
+
+#define EVP_PKEY_OP_TYPE_CRYPT \
+	(EVP_PKEY_OP_ENCRYPT | EVP_PKEY_OP_DECRYPT)
+
+#define EVP_PKEY_OP_TYPE_NOGEN \
+	(EVP_PKEY_OP_SIG | EVP_PKEY_OP_CRYPT | EVP_PKEY_OP_DERIVE)
+
+#define EVP_PKEY_OP_TYPE_GEN \
+		(EVP_PKEY_OP_PARAMGEN | EVP_PKEY_OP_KEYGEN)
+
+#define	 EVP_PKEY_CTX_set_signature_md(ctx, md)	\
+		EVP_PKEY_CTX_ctrl(ctx, -1, EVP_PKEY_OP_TYPE_SIG,  \
+					EVP_PKEY_CTRL_MD, 0, (void *)md)
+
+#define EVP_PKEY_CTRL_MD		1
+#define EVP_PKEY_CTRL_PEER_KEY		2
+
+#define EVP_PKEY_CTRL_PKCS7_ENCRYPT	3
+#define EVP_PKEY_CTRL_PKCS7_DECRYPT	4
+
+#define EVP_PKEY_CTRL_PKCS7_SIGN	5
+
+#define EVP_PKEY_CTRL_SET_MAC_KEY	6
+
+#define EVP_PKEY_CTRL_DIGESTINIT	7
+
+/* Used by GOST key encryption in TLS */
+#define EVP_PKEY_CTRL_SET_IV 		8
+
+#define EVP_PKEY_CTRL_CMS_ENCRYPT	9
+#define EVP_PKEY_CTRL_CMS_DECRYPT	10
+#define EVP_PKEY_CTRL_CMS_SIGN		11
+
+#define EVP_PKEY_CTRL_CIPHER		12
+
+#define EVP_PKEY_ALG_CTRL		0x1000
+
+
+#define EVP_PKEY_FLAG_AUTOARGLEN	2
+/* Method handles all operations: don't assume any digest related
+ * defaults.
+ */
+#define EVP_PKEY_FLAG_SIGCTX_CUSTOM	4
+
+const EVP_PKEY_METHOD *EVP_PKEY_meth_find(int type);
+EVP_PKEY_METHOD* EVP_PKEY_meth_new(int id, int flags);
+void EVP_PKEY_meth_get0_info(int *ppkey_id, int *pflags,
+				const EVP_PKEY_METHOD *meth);
+void EVP_PKEY_meth_copy(EVP_PKEY_METHOD *dst, const EVP_PKEY_METHOD *src);
+void EVP_PKEY_meth_free(EVP_PKEY_METHOD *pmeth);
+int EVP_PKEY_meth_add0(const EVP_PKEY_METHOD *pmeth);
+
+EVP_PKEY_CTX *EVP_PKEY_CTX_new(EVP_PKEY *pkey, ENGINE *e);
+EVP_PKEY_CTX *EVP_PKEY_CTX_new_id(int id, ENGINE *e);
+EVP_PKEY_CTX *EVP_PKEY_CTX_dup(EVP_PKEY_CTX *ctx);
+void EVP_PKEY_CTX_free(EVP_PKEY_CTX *ctx);
+
+int EVP_PKEY_CTX_ctrl(EVP_PKEY_CTX *ctx, int keytype, int optype,
+				int cmd, int p1, void *p2);
+int EVP_PKEY_CTX_ctrl_str(EVP_PKEY_CTX *ctx, const char *type,
+						const char *value);
+
+int EVP_PKEY_CTX_get_operation(EVP_PKEY_CTX *ctx);
+void EVP_PKEY_CTX_set0_keygen_info(EVP_PKEY_CTX *ctx, int *dat, int datlen);
+
+EVP_PKEY *EVP_PKEY_new_mac_key(int type, ENGINE *e,
+				const unsigned char *key, int keylen);
+
+void EVP_PKEY_CTX_set_data(EVP_PKEY_CTX *ctx, void *data);
+void *EVP_PKEY_CTX_get_data(EVP_PKEY_CTX *ctx);
+EVP_PKEY *EVP_PKEY_CTX_get0_pkey(EVP_PKEY_CTX *ctx);
+
+EVP_PKEY *EVP_PKEY_CTX_get0_peerkey(EVP_PKEY_CTX *ctx);
+
+void EVP_PKEY_CTX_set_app_data(EVP_PKEY_CTX *ctx, void *data);
+void *EVP_PKEY_CTX_get_app_data(EVP_PKEY_CTX *ctx);
+
+int EVP_PKEY_sign_init(EVP_PKEY_CTX *ctx);
+int EVP_PKEY_sign(EVP_PKEY_CTX *ctx,
+			unsigned char *sig, size_t *siglen,
+			const unsigned char *tbs, size_t tbslen);
+int EVP_PKEY_verify_init(EVP_PKEY_CTX *ctx);
+int EVP_PKEY_verify(EVP_PKEY_CTX *ctx,
+			const unsigned char *sig, size_t siglen,
+			const unsigned char *tbs, size_t tbslen);
+int EVP_PKEY_verify_recover_init(EVP_PKEY_CTX *ctx);
+int EVP_PKEY_verify_recover(EVP_PKEY_CTX *ctx,
+			unsigned char *rout, size_t *routlen,
+			const unsigned char *sig, size_t siglen);
+int EVP_PKEY_encrypt_init(EVP_PKEY_CTX *ctx);
+int EVP_PKEY_encrypt(EVP_PKEY_CTX *ctx,
+			unsigned char *out, size_t *outlen,
+			const unsigned char *in, size_t inlen);
+int EVP_PKEY_decrypt_init(EVP_PKEY_CTX *ctx);
+int EVP_PKEY_decrypt(EVP_PKEY_CTX *ctx,
+			unsigned char *out, size_t *outlen,
+			const unsigned char *in, size_t inlen);
+
+int EVP_PKEY_derive_init(EVP_PKEY_CTX *ctx);
+int EVP_PKEY_derive_set_peer(EVP_PKEY_CTX *ctx, EVP_PKEY *peer);
+int EVP_PKEY_derive(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen);
+
+typedef int EVP_PKEY_gen_cb(EVP_PKEY_CTX *ctx);
+
+int EVP_PKEY_paramgen_init(EVP_PKEY_CTX *ctx);
+int EVP_PKEY_paramgen(EVP_PKEY_CTX *ctx, EVP_PKEY **ppkey);
+int EVP_PKEY_keygen_init(EVP_PKEY_CTX *ctx);
+int EVP_PKEY_keygen(EVP_PKEY_CTX *ctx, EVP_PKEY **ppkey);
+
+void EVP_PKEY_CTX_set_cb(EVP_PKEY_CTX *ctx, EVP_PKEY_gen_cb *cb);
+EVP_PKEY_gen_cb *EVP_PKEY_CTX_get_cb(EVP_PKEY_CTX *ctx);
+
+int EVP_PKEY_CTX_get_keygen_info(EVP_PKEY_CTX *ctx, int idx);
+
+void EVP_PKEY_meth_set_init(EVP_PKEY_METHOD *pmeth,
+	int (*init)(EVP_PKEY_CTX *ctx));
+
+void EVP_PKEY_meth_set_copy(EVP_PKEY_METHOD *pmeth,
+	int (*copy)(EVP_PKEY_CTX *dst, EVP_PKEY_CTX *src));
+
+void EVP_PKEY_meth_set_cleanup(EVP_PKEY_METHOD *pmeth,
+	void (*cleanup)(EVP_PKEY_CTX *ctx));
+
+void EVP_PKEY_meth_set_paramgen(EVP_PKEY_METHOD *pmeth,
+	int (*paramgen_init)(EVP_PKEY_CTX *ctx),
+	int (*paramgen)(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey));
+
+void EVP_PKEY_meth_set_keygen(EVP_PKEY_METHOD *pmeth,
+	int (*keygen_init)(EVP_PKEY_CTX *ctx),
+	int (*keygen)(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey));
+
+void EVP_PKEY_meth_set_sign(EVP_PKEY_METHOD *pmeth,
+	int (*sign_init)(EVP_PKEY_CTX *ctx),
+	int (*sign)(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					const unsigned char *tbs, size_t tbslen));
+
+void EVP_PKEY_meth_set_verify(EVP_PKEY_METHOD *pmeth,
+	int (*verify_init)(EVP_PKEY_CTX *ctx),
+	int (*verify)(EVP_PKEY_CTX *ctx, const unsigned char *sig, size_t siglen,
+					const unsigned char *tbs, size_t tbslen));
+
+void EVP_PKEY_meth_set_verify_recover(EVP_PKEY_METHOD *pmeth,
+	int (*verify_recover_init)(EVP_PKEY_CTX *ctx),
+	int (*verify_recover)(EVP_PKEY_CTX *ctx,
+					unsigned char *sig, size_t *siglen,
+					const unsigned char *tbs, size_t tbslen));
+
+void EVP_PKEY_meth_set_signctx(EVP_PKEY_METHOD *pmeth,
+	int (*signctx_init)(EVP_PKEY_CTX *ctx, EVP_MD_CTX *mctx),
+	int (*signctx)(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					EVP_MD_CTX *mctx));
+
+void EVP_PKEY_meth_set_verifyctx(EVP_PKEY_METHOD *pmeth,
+	int (*verifyctx_init)(EVP_PKEY_CTX *ctx, EVP_MD_CTX *mctx),
+	int (*verifyctx)(EVP_PKEY_CTX *ctx, const unsigned char *sig,int siglen,
+					EVP_MD_CTX *mctx));
+
+void EVP_PKEY_meth_set_encrypt(EVP_PKEY_METHOD *pmeth,
+	int (*encrypt_init)(EVP_PKEY_CTX *ctx),
+	int (*encryptfn)(EVP_PKEY_CTX *ctx, unsigned char *out, size_t *outlen,
+					const unsigned char *in, size_t inlen));
+
+void EVP_PKEY_meth_set_decrypt(EVP_PKEY_METHOD *pmeth,
+	int (*decrypt_init)(EVP_PKEY_CTX *ctx),
+	int (*decrypt)(EVP_PKEY_CTX *ctx, unsigned char *out, size_t *outlen,
+					const unsigned char *in, size_t inlen));
+
+void EVP_PKEY_meth_set_derive(EVP_PKEY_METHOD *pmeth,
+	int (*derive_init)(EVP_PKEY_CTX *ctx),
+	int (*derive)(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen));
+
+void EVP_PKEY_meth_set_ctrl(EVP_PKEY_METHOD *pmeth,
+	int (*ctrl)(EVP_PKEY_CTX *ctx, int type, int p1, void *p2),
+	int (*ctrl_str)(EVP_PKEY_CTX *ctx,
+					const char *type, const char *value));
 
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
@@ -955,51 +1252,85 @@ void ERR_load_EVP_strings(void);
 /* Error codes for the EVP functions. */
 
 /* Function codes. */
+#define EVP_F_AESNI_INIT_KEY				 165
+#define EVP_F_AESNI_XTS_CIPHER				 176
 #define EVP_F_AES_INIT_KEY				 133
-#define EVP_F_ALG_MODULE_INIT				 138
+#define EVP_F_AES_XTS					 172
+#define EVP_F_AES_XTS_CIPHER				 175
 #define EVP_F_CAMELLIA_INIT_KEY				 159
+#define EVP_F_CMAC_INIT					 173
 #define EVP_F_D2I_PKEY					 100
-#define EVP_F_DO_EVP_ENC_ENGINE				 140
-#define EVP_F_DO_EVP_ENC_ENGINE_FULL			 141
-#define EVP_F_DO_EVP_MD_ENGINE				 139
-#define EVP_F_DO_EVP_MD_ENGINE_FULL			 142
+#define EVP_F_DO_SIGVER_INIT				 161
 #define EVP_F_DSAPKEY2PKCS8				 134
 #define EVP_F_DSA_PKEY2PKCS8				 135
 #define EVP_F_ECDSA_PKEY2PKCS8				 129
 #define EVP_F_ECKEY_PKEY2PKCS8				 132
-#define EVP_F_EVP_CIPHERINIT				 137
 #define EVP_F_EVP_CIPHERINIT_EX				 123
+#define EVP_F_EVP_CIPHER_CTX_COPY			 163
 #define EVP_F_EVP_CIPHER_CTX_CTRL			 124
 #define EVP_F_EVP_CIPHER_CTX_SET_KEY_LENGTH		 122
 #define EVP_F_EVP_DECRYPTFINAL_EX			 101
-#define EVP_F_EVP_DIGESTINIT				 136
 #define EVP_F_EVP_DIGESTINIT_EX				 128
 #define EVP_F_EVP_ENCRYPTFINAL_EX			 127
 #define EVP_F_EVP_MD_CTX_COPY_EX			 110
+#define EVP_F_EVP_MD_SIZE				 162
 #define EVP_F_EVP_OPENINIT				 102
 #define EVP_F_EVP_PBE_ALG_ADD				 115
+#define EVP_F_EVP_PBE_ALG_ADD_TYPE			 160
 #define EVP_F_EVP_PBE_CIPHERINIT			 116
 #define EVP_F_EVP_PKCS82PKEY				 111
+#define EVP_F_EVP_PKCS82PKEY_BROKEN			 136
 #define EVP_F_EVP_PKEY2PKCS8_BROKEN			 113
 #define EVP_F_EVP_PKEY_COPY_PARAMETERS			 103
+#define EVP_F_EVP_PKEY_CTX_CTRL				 137
+#define EVP_F_EVP_PKEY_CTX_CTRL_STR			 150
+#define EVP_F_EVP_PKEY_CTX_DUP				 156
 #define EVP_F_EVP_PKEY_DECRYPT				 104
+#define EVP_F_EVP_PKEY_DECRYPT_INIT			 138
+#define EVP_F_EVP_PKEY_DECRYPT_OLD			 151
+#define EVP_F_EVP_PKEY_DERIVE				 153
+#define EVP_F_EVP_PKEY_DERIVE_INIT			 154
+#define EVP_F_EVP_PKEY_DERIVE_SET_PEER			 155
 #define EVP_F_EVP_PKEY_ENCRYPT				 105
+#define EVP_F_EVP_PKEY_ENCRYPT_INIT			 139
+#define EVP_F_EVP_PKEY_ENCRYPT_OLD			 152
 #define EVP_F_EVP_PKEY_GET1_DH				 119
 #define EVP_F_EVP_PKEY_GET1_DSA				 120
 #define EVP_F_EVP_PKEY_GET1_ECDSA			 130
 #define EVP_F_EVP_PKEY_GET1_EC_KEY			 131
 #define EVP_F_EVP_PKEY_GET1_RSA				 121
+#define EVP_F_EVP_PKEY_KEYGEN				 146
+#define EVP_F_EVP_PKEY_KEYGEN_INIT			 147
 #define EVP_F_EVP_PKEY_NEW				 106
+#define EVP_F_EVP_PKEY_PARAMGEN				 148
+#define EVP_F_EVP_PKEY_PARAMGEN_INIT			 149
+#define EVP_F_EVP_PKEY_SIGN				 140
+#define EVP_F_EVP_PKEY_SIGN_INIT			 141
+#define EVP_F_EVP_PKEY_VERIFY				 142
+#define EVP_F_EVP_PKEY_VERIFY_INIT			 143
+#define EVP_F_EVP_PKEY_VERIFY_RECOVER			 144
+#define EVP_F_EVP_PKEY_VERIFY_RECOVER_INIT		 145
 #define EVP_F_EVP_RIJNDAEL				 126
 #define EVP_F_EVP_SIGNFINAL				 107
 #define EVP_F_EVP_VERIFYFINAL				 108
+#define EVP_F_FIPS_CIPHERINIT				 166
+#define EVP_F_FIPS_CIPHER_CTX_COPY			 170
+#define EVP_F_FIPS_CIPHER_CTX_CTRL			 167
+#define EVP_F_FIPS_CIPHER_CTX_SET_KEY_LENGTH		 171
+#define EVP_F_FIPS_DIGESTINIT				 168
+#define EVP_F_FIPS_MD_CTX_COPY				 169
+#define EVP_F_HMAC_INIT_EX				 174
+#define EVP_F_INT_CTX_NEW				 157
 #define EVP_F_PKCS5_PBE_KEYIVGEN			 117
 #define EVP_F_PKCS5_V2_PBE_KEYIVGEN			 118
+#define EVP_F_PKCS5_V2_PBKDF2_KEYIVGEN			 164
 #define EVP_F_PKCS8_SET_BROKEN				 112
+#define EVP_F_PKEY_SET_TYPE				 158
 #define EVP_F_RC2_MAGIC_TO_METH				 109
 #define EVP_F_RC5_CTRL					 125
 
 /* Reason codes. */
+#define EVP_R_AES_IV_SETUP_FAILED			 162
 #define EVP_R_AES_KEY_SETUP_FAILED			 143
 #define EVP_R_ASN1_LIB					 140
 #define EVP_R_BAD_BLOCK_LENGTH				 136
@@ -1007,41 +1338,54 @@ void ERR_load_EVP_strings(void);
 #define EVP_R_BAD_KEY_LENGTH				 137
 #define EVP_R_BN_DECODE_ERROR				 112
 #define EVP_R_BN_PUBKEY_ERROR				 113
+#define EVP_R_BUFFER_TOO_SMALL				 155
 #define EVP_R_CAMELLIA_KEY_SETUP_FAILED			 157
 #define EVP_R_CIPHER_PARAMETER_ERROR			 122
+#define EVP_R_COMMAND_NOT_SUPPORTED			 147
 #define EVP_R_CTRL_NOT_IMPLEMENTED			 132
 #define EVP_R_CTRL_OPERATION_NOT_IMPLEMENTED		 133
 #define EVP_R_DATA_NOT_MULTIPLE_OF_BLOCK_LENGTH		 138
 #define EVP_R_DECODE_ERROR				 114
 #define EVP_R_DIFFERENT_KEY_TYPES			 101
-#define EVP_R_DISABLED_FOR_FIPS				 144
+#define EVP_R_DIFFERENT_PARAMETERS			 153
+#define EVP_R_DISABLED_FOR_FIPS				 163
 #define EVP_R_ENCODE_ERROR				 115
-#define EVP_R_ERROR_LOADING_SECTION			 145
-#define EVP_R_ERROR_SETTING_FIPS_MODE			 146
 #define EVP_R_EVP_PBE_CIPHERINIT_ERROR			 119
 #define EVP_R_EXPECTING_AN_RSA_KEY			 127
 #define EVP_R_EXPECTING_A_DH_KEY			 128
 #define EVP_R_EXPECTING_A_DSA_KEY			 129
 #define EVP_R_EXPECTING_A_ECDSA_KEY			 141
 #define EVP_R_EXPECTING_A_EC_KEY			 142
-#define EVP_R_FIPS_MODE_NOT_SUPPORTED			 147
 #define EVP_R_INITIALIZATION_ERROR			 134
 #define EVP_R_INPUT_NOT_INITIALIZED			 111
-#define EVP_R_INVALID_FIPS_MODE				 148
+#define EVP_R_INVALID_DIGEST				 152
 #define EVP_R_INVALID_KEY_LENGTH			 130
+#define EVP_R_INVALID_OPERATION				 148
 #define EVP_R_IV_TOO_LARGE				 102
 #define EVP_R_KEYGEN_FAILURE				 120
+#define EVP_R_MESSAGE_DIGEST_IS_NULL			 159
+#define EVP_R_METHOD_NOT_SUPPORTED			 144
 #define EVP_R_MISSING_PARAMETERS			 103
 #define EVP_R_NO_CIPHER_SET				 131
+#define EVP_R_NO_DEFAULT_DIGEST				 158
 #define EVP_R_NO_DIGEST_SET				 139
 #define EVP_R_NO_DSA_PARAMETERS				 116
+#define EVP_R_NO_KEY_SET				 154
+#define EVP_R_NO_OPERATION_SET				 149
 #define EVP_R_NO_SIGN_FUNCTION_CONFIGURED		 104
 #define EVP_R_NO_VERIFY_FUNCTION_CONFIGURED		 105
+#define EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE	 150
+#define EVP_R_OPERATON_NOT_INITIALIZED			 151
 #define EVP_R_PKCS8_UNKNOWN_BROKEN_TYPE			 117
+#define EVP_R_PRIVATE_KEY_DECODE_ERROR			 145
+#define EVP_R_PRIVATE_KEY_ENCODE_ERROR			 146
 #define EVP_R_PUBLIC_KEY_NOT_RSA			 106
-#define EVP_R_UNKNOWN_OPTION				 149
+#define EVP_R_TOO_LARGE					 164
+#define EVP_R_UNKNOWN_CIPHER				 160
+#define EVP_R_UNKNOWN_DIGEST				 161
 #define EVP_R_UNKNOWN_PBE_ALGORITHM			 121
 #define EVP_R_UNSUPORTED_NUMBER_OF_ROUNDS		 135
+#define EVP_R_UNSUPPORTED_ALGORITHM			 156
 #define EVP_R_UNSUPPORTED_CIPHER			 107
 #define EVP_R_UNSUPPORTED_KEYLENGTH			 123
 #define EVP_R_UNSUPPORTED_KEY_DERIVATION_FUNCTION	 124
@@ -1051,7 +1395,6 @@ void ERR_load_EVP_strings(void);
 #define EVP_R_UNSUPPORTED_SALT_TYPE			 126
 #define EVP_R_WRONG_FINAL_BLOCK_LENGTH			 109
 #define EVP_R_WRONG_PUBLIC_KEY_TYPE			 110
-#define EVP_R_SEED_KEY_SETUP_FAILED			 162
 
 #ifdef  __cplusplus
 }
diff --git a/crypto/evp/evp_enc.c b/crypto/evp/evp_enc.c
index 30e0ca4d9faa..0c54f05e6ef5 100644
--- a/crypto/evp/evp_enc.c
+++ b/crypto/evp/evp_enc.c
@@ -64,18 +64,26 @@
 #ifndef OPENSSL_NO_ENGINE
 #include <openssl/engine.h>
 #endif
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 #include "evp_locl.h"
 
 #ifdef OPENSSL_FIPS
-	#define M_do_cipher(ctx, out, in, inl) \
-		EVP_Cipher(ctx,out,in,inl)
+#define M_do_cipher(ctx, out, in, inl) FIPS_cipher(ctx, out, in, inl)
 #else
-	#define M_do_cipher(ctx, out, in, inl) \
-		ctx->cipher->do_cipher(ctx,out,in,inl)
+#define M_do_cipher(ctx, out, in, inl) ctx->cipher->do_cipher(ctx, out, in, inl)
 #endif
 
+
 const char EVP_version[]="EVP" OPENSSL_VERSION_PTEXT;
 
+void EVP_CIPHER_CTX_init(EVP_CIPHER_CTX *ctx)
+	{
+	memset(ctx,0,sizeof(EVP_CIPHER_CTX));
+	/* ctx->cipher=NULL; */
+	}
+
 EVP_CIPHER_CTX *EVP_CIPHER_CTX_new(void)
 	{
 	EVP_CIPHER_CTX *ctx=OPENSSL_malloc(sizeof *ctx);
@@ -92,6 +100,164 @@ int EVP_CipherInit(EVP_CIPHER_CTX *ctx, const EVP_CIPHER *cipher,
 	return EVP_CipherInit_ex(ctx,cipher,NULL,key,iv,enc);
 	}
 
+int EVP_CipherInit_ex(EVP_CIPHER_CTX *ctx, const EVP_CIPHER *cipher, ENGINE *impl,
+	     const unsigned char *key, const unsigned char *iv, int enc)
+	{
+	if (enc == -1)
+		enc = ctx->encrypt;
+	else
+		{
+		if (enc)
+			enc = 1;
+		ctx->encrypt = enc;
+		}
+#ifndef OPENSSL_NO_ENGINE
+	/* Whether it's nice or not, "Inits" can be used on "Final"'d contexts
+	 * so this context may already have an ENGINE! Try to avoid releasing
+	 * the previous handle, re-querying for an ENGINE, and having a
+	 * reinitialisation, when it may all be unecessary. */
+	if (ctx->engine && ctx->cipher && (!cipher ||
+			(cipher && (cipher->nid == ctx->cipher->nid))))
+		goto skip_to_init;
+#endif
+	if (cipher)
+		{
+		/* Ensure a context left lying around from last time is cleared
+		 * (the previous check attempted to avoid this if the same
+		 * ENGINE and EVP_CIPHER could be used). */
+		if (ctx->cipher)
+			{
+			unsigned long flags = ctx->flags;
+			EVP_CIPHER_CTX_cleanup(ctx);
+			/* Restore encrypt and flags */
+			ctx->encrypt = enc;
+			ctx->flags = flags;
+			}
+#ifndef OPENSSL_NO_ENGINE
+		if(impl)
+			{
+			if (!ENGINE_init(impl))
+				{
+				EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_INITIALIZATION_ERROR);
+				return 0;
+				}
+			}
+		else
+			/* Ask if an ENGINE is reserved for this job */
+			impl = ENGINE_get_cipher_engine(cipher->nid);
+		if(impl)
+			{
+			/* There's an ENGINE for this job ... (apparently) */
+			const EVP_CIPHER *c = ENGINE_get_cipher(impl, cipher->nid);
+			if(!c)
+				{
+				/* One positive side-effect of US's export
+				 * control history, is that we should at least
+				 * be able to avoid using US mispellings of
+				 * "initialisation"? */
+				EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_INITIALIZATION_ERROR);
+				return 0;
+				}
+			/* We'll use the ENGINE's private cipher definition */
+			cipher = c;
+			/* Store the ENGINE functional reference so we know
+			 * 'cipher' came from an ENGINE and we need to release
+			 * it when done. */
+			ctx->engine = impl;
+			}
+		else
+			ctx->engine = NULL;
+#endif
+
+#ifdef OPENSSL_FIPS
+		if (FIPS_mode())
+			return FIPS_cipherinit(ctx, cipher, key, iv, enc);
+#endif
+		ctx->cipher=cipher;
+		if (ctx->cipher->ctx_size)
+			{
+			ctx->cipher_data=OPENSSL_malloc(ctx->cipher->ctx_size);
+			if (!ctx->cipher_data)
+				{
+				EVPerr(EVP_F_EVP_CIPHERINIT_EX, ERR_R_MALLOC_FAILURE);
+				return 0;
+				}
+			}
+		else
+			{
+			ctx->cipher_data = NULL;
+			}
+		ctx->key_len = cipher->key_len;
+		ctx->flags = 0;
+		if(ctx->cipher->flags & EVP_CIPH_CTRL_INIT)
+			{
+			if(!EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_INIT, 0, NULL))
+				{
+				EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_INITIALIZATION_ERROR);
+				return 0;
+				}
+			}
+		}
+	else if(!ctx->cipher)
+		{
+		EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_NO_CIPHER_SET);
+		return 0;
+		}
+#ifndef OPENSSL_NO_ENGINE
+skip_to_init:
+#endif
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode())
+		return FIPS_cipherinit(ctx, cipher, key, iv, enc);
+#endif
+	/* we assume block size is a power of 2 in *cryptUpdate */
+	OPENSSL_assert(ctx->cipher->block_size == 1
+	    || ctx->cipher->block_size == 8
+	    || ctx->cipher->block_size == 16);
+
+	if(!(EVP_CIPHER_CTX_flags(ctx) & EVP_CIPH_CUSTOM_IV)) {
+		switch(EVP_CIPHER_CTX_mode(ctx)) {
+
+			case EVP_CIPH_STREAM_CIPHER:
+			case EVP_CIPH_ECB_MODE:
+			break;
+
+			case EVP_CIPH_CFB_MODE:
+			case EVP_CIPH_OFB_MODE:
+
+			ctx->num = 0;
+			/* fall-through */
+
+			case EVP_CIPH_CBC_MODE:
+
+			OPENSSL_assert(EVP_CIPHER_CTX_iv_length(ctx) <=
+					(int)sizeof(ctx->iv));
+			if(iv) memcpy(ctx->oiv, iv, EVP_CIPHER_CTX_iv_length(ctx));
+			memcpy(ctx->iv, ctx->oiv, EVP_CIPHER_CTX_iv_length(ctx));
+			break;
+
+			case EVP_CIPH_CTR_MODE:
+			ctx->num = 0;
+			/* Don't reuse IV for CTR mode */
+			if(iv)
+				memcpy(ctx->iv, iv, EVP_CIPHER_CTX_iv_length(ctx));
+			break;
+
+			default:
+			return 0;
+			break;
+		}
+	}
+
+	if(key || (ctx->cipher->flags & EVP_CIPH_ALWAYS_CALL_INIT)) {
+		if(!ctx->cipher->init(ctx,key,iv,enc)) return 0;
+	}
+	ctx->buf_len=0;
+	ctx->final_used=0;
+	ctx->block_mask=ctx->cipher->block_size-1;
+	return 1;
+	}
+
 int EVP_CipherUpdate(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl,
 	     const unsigned char *in, int inl)
 	{
@@ -143,6 +309,16 @@ int EVP_EncryptUpdate(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl,
 	{
 	int i,j,bl;
 
+	if (ctx->cipher->flags & EVP_CIPH_FLAG_CUSTOM_CIPHER)
+		{
+		i = M_do_cipher(ctx, out, in, inl);
+		if (i < 0)
+			return 0;
+		else
+			*outl = i;
+		return 1;
+		}
+
 	if (inl <= 0)
 		{
 		*outl = 0;
@@ -213,6 +389,16 @@ int EVP_EncryptFinal_ex(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl)
 	int n,ret;
 	unsigned int i, b, bl;
 
+	if (ctx->cipher->flags & EVP_CIPH_FLAG_CUSTOM_CIPHER)
+		{
+		ret = M_do_cipher(ctx, out, NULL, 0);
+		if (ret < 0)
+			return 0;
+		else 
+			*outl = ret;
+		return 1;
+		}
+
 	b=ctx->cipher->block_size;
 	OPENSSL_assert(b <= sizeof ctx->buf);
 	if (b == 1)
@@ -250,6 +436,19 @@ int EVP_DecryptUpdate(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl,
 	int fix_len;
 	unsigned int b;
 
+	if (ctx->cipher->flags & EVP_CIPH_FLAG_CUSTOM_CIPHER)
+		{
+		fix_len = M_do_cipher(ctx, out, in, inl);
+		if (fix_len < 0)
+			{
+			*outl = 0;
+			return 0;
+			}
+		else
+			*outl = fix_len;
+		return 1;
+		}
+
 	if (inl <= 0)
 		{
 		*outl = 0;
@@ -303,8 +502,18 @@ int EVP_DecryptFinal_ex(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl)
 	{
 	int i,n;
 	unsigned int b;
-
 	*outl=0;
+
+	if (ctx->cipher->flags & EVP_CIPH_FLAG_CUSTOM_CIPHER)
+		{
+		i = M_do_cipher(ctx, out, NULL, 0);
+		if (i < 0)
+			return 0;
+		else
+			*outl = i;
+		return 1;
+		}
+
 	b=ctx->cipher->block_size;
 	if (ctx->flags & EVP_CIPH_NO_PADDING)
 		{
@@ -357,6 +566,33 @@ void EVP_CIPHER_CTX_free(EVP_CIPHER_CTX *ctx)
 		}
 	}
 
+int EVP_CIPHER_CTX_cleanup(EVP_CIPHER_CTX *c)
+	{
+#ifndef OPENSSL_FIPS
+	if (c->cipher != NULL)
+		{
+		if(c->cipher->cleanup && !c->cipher->cleanup(c))
+			return 0;
+		/* Cleanse cipher context data */
+		if (c->cipher_data)
+			OPENSSL_cleanse(c->cipher_data, c->cipher->ctx_size);
+		}
+	if (c->cipher_data)
+		OPENSSL_free(c->cipher_data);
+#endif
+#ifndef OPENSSL_NO_ENGINE
+	if (c->engine)
+		/* The EVP_CIPHER we used belongs to an ENGINE, release the
+		 * functional reference we held for this reason. */
+		ENGINE_finish(c->engine);
+#endif
+#ifdef OPENSSL_FIPS
+	FIPS_cipher_ctx_cleanup(c);
+#endif
+	memset(c,0,sizeof(EVP_CIPHER_CTX));
+	return 1;
+	}
+
 int EVP_CIPHER_CTX_set_key_length(EVP_CIPHER_CTX *c, int keylen)
 	{
 	if(c->cipher->flags & EVP_CIPH_CUSTOM_KEY_LENGTH) 
@@ -378,6 +614,27 @@ int EVP_CIPHER_CTX_set_padding(EVP_CIPHER_CTX *ctx, int pad)
 	return 1;
 	}
 
+int EVP_CIPHER_CTX_ctrl(EVP_CIPHER_CTX *ctx, int type, int arg, void *ptr)
+{
+	int ret;
+	if(!ctx->cipher) {
+		EVPerr(EVP_F_EVP_CIPHER_CTX_CTRL, EVP_R_NO_CIPHER_SET);
+		return 0;
+	}
+
+	if(!ctx->cipher->ctrl) {
+		EVPerr(EVP_F_EVP_CIPHER_CTX_CTRL, EVP_R_CTRL_NOT_IMPLEMENTED);
+		return 0;
+	}
+
+	ret = ctx->cipher->ctrl(ctx, type, arg, ptr);
+	if(ret == -1) {
+		EVPerr(EVP_F_EVP_CIPHER_CTX_CTRL, EVP_R_CTRL_OPERATION_NOT_IMPLEMENTED);
+		return 0;
+	}
+	return ret;
+}
+
 int EVP_CIPHER_CTX_rand_key(EVP_CIPHER_CTX *ctx, unsigned char *key)
 	{
 	if (ctx->cipher->flags & EVP_CIPH_RAND_KEY)
@@ -387,54 +644,38 @@ int EVP_CIPHER_CTX_rand_key(EVP_CIPHER_CTX *ctx, unsigned char *key)
 	return 1;
 	}
 
-#ifndef OPENSSL_NO_ENGINE
-
-#ifdef OPENSSL_FIPS
-
-static int do_evp_enc_engine_full(EVP_CIPHER_CTX *ctx, const EVP_CIPHER **pcipher, ENGINE *impl)
+int EVP_CIPHER_CTX_copy(EVP_CIPHER_CTX *out, const EVP_CIPHER_CTX *in)
 	{
-	if(impl)
+	if ((in == NULL) || (in->cipher == NULL))
 		{
-		if (!ENGINE_init(impl))
-			{
-			EVPerr(EVP_F_DO_EVP_ENC_ENGINE_FULL, EVP_R_INITIALIZATION_ERROR);
-			return 0;
-			}
+		EVPerr(EVP_F_EVP_CIPHER_CTX_COPY,EVP_R_INPUT_NOT_INITIALIZED);
+		return 0;
 		}
-	else
-		/* Ask if an ENGINE is reserved for this job */
-		impl = ENGINE_get_cipher_engine((*pcipher)->nid);
-	if(impl)
+#ifndef OPENSSL_NO_ENGINE
+	/* Make sure it's safe to copy a cipher context using an ENGINE */
+	if (in->engine && !ENGINE_init(in->engine))
+		{
+		EVPerr(EVP_F_EVP_CIPHER_CTX_COPY,ERR_R_ENGINE_LIB);
+		return 0;
+		}
+#endif
+
+	EVP_CIPHER_CTX_cleanup(out);
+	memcpy(out,in,sizeof *out);
+
+	if (in->cipher_data && in->cipher->ctx_size)
 		{
-		/* There's an ENGINE for this job ... (apparently) */
-		const EVP_CIPHER *c = ENGINE_get_cipher(impl, (*pcipher)->nid);
-		if(!c)
+		out->cipher_data=OPENSSL_malloc(in->cipher->ctx_size);
+		if (!out->cipher_data)
 			{
-			/* One positive side-effect of US's export
-			 * control history, is that we should at least
-			 * be able to avoid using US mispellings of
-			 * "initialisation"? */
-			EVPerr(EVP_F_DO_EVP_ENC_ENGINE_FULL, EVP_R_INITIALIZATION_ERROR);
+			EVPerr(EVP_F_EVP_CIPHER_CTX_COPY,ERR_R_MALLOC_FAILURE);
 			return 0;
 			}
-		/* We'll use the ENGINE's private cipher definition */
-		*pcipher = c;
-		/* Store the ENGINE functional reference so we know
-		 * 'cipher' came from an ENGINE and we need to release
-		 * it when done. */
-		ctx->engine = impl;
+		memcpy(out->cipher_data,in->cipher_data,in->cipher->ctx_size);
 		}
-	else
-		ctx->engine = NULL;
-	return 1;
-	}
 
-void int_EVP_CIPHER_init_engine_callbacks(void)
-	{
-	int_EVP_CIPHER_set_engine_callbacks(
-		ENGINE_finish, do_evp_enc_engine_full);
+	if (in->cipher->flags & EVP_CIPH_CUSTOM_COPY)
+		return in->cipher->ctrl((EVP_CIPHER_CTX *)in, EVP_CTRL_COPY, 0, out);
+	return 1;
 	}
 
-#endif
-
-#endif
diff --git a/crypto/evp/evp_err.c b/crypto/evp/evp_err.c
index b5b900d4fe4f..db0f76d59bcd 100644
--- a/crypto/evp/evp_err.c
+++ b/crypto/evp/evp_err.c
@@ -1,6 +1,6 @@
 /* crypto/evp/evp_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,47 +70,80 @@
 
 static ERR_STRING_DATA EVP_str_functs[]=
 	{
+{ERR_FUNC(EVP_F_AESNI_INIT_KEY),	"AESNI_INIT_KEY"},
+{ERR_FUNC(EVP_F_AESNI_XTS_CIPHER),	"AESNI_XTS_CIPHER"},
 {ERR_FUNC(EVP_F_AES_INIT_KEY),	"AES_INIT_KEY"},
-{ERR_FUNC(EVP_F_ALG_MODULE_INIT),	"ALG_MODULE_INIT"},
+{ERR_FUNC(EVP_F_AES_XTS),	"AES_XTS"},
+{ERR_FUNC(EVP_F_AES_XTS_CIPHER),	"AES_XTS_CIPHER"},
 {ERR_FUNC(EVP_F_CAMELLIA_INIT_KEY),	"CAMELLIA_INIT_KEY"},
+{ERR_FUNC(EVP_F_CMAC_INIT),	"CMAC_INIT"},
 {ERR_FUNC(EVP_F_D2I_PKEY),	"D2I_PKEY"},
-{ERR_FUNC(EVP_F_DO_EVP_ENC_ENGINE),	"DO_EVP_ENC_ENGINE"},
-{ERR_FUNC(EVP_F_DO_EVP_ENC_ENGINE_FULL),	"DO_EVP_ENC_ENGINE_FULL"},
-{ERR_FUNC(EVP_F_DO_EVP_MD_ENGINE),	"DO_EVP_MD_ENGINE"},
-{ERR_FUNC(EVP_F_DO_EVP_MD_ENGINE_FULL),	"DO_EVP_MD_ENGINE_FULL"},
+{ERR_FUNC(EVP_F_DO_SIGVER_INIT),	"DO_SIGVER_INIT"},
 {ERR_FUNC(EVP_F_DSAPKEY2PKCS8),	"DSAPKEY2PKCS8"},
 {ERR_FUNC(EVP_F_DSA_PKEY2PKCS8),	"DSA_PKEY2PKCS8"},
 {ERR_FUNC(EVP_F_ECDSA_PKEY2PKCS8),	"ECDSA_PKEY2PKCS8"},
 {ERR_FUNC(EVP_F_ECKEY_PKEY2PKCS8),	"ECKEY_PKEY2PKCS8"},
-{ERR_FUNC(EVP_F_EVP_CIPHERINIT),	"EVP_CipherInit"},
 {ERR_FUNC(EVP_F_EVP_CIPHERINIT_EX),	"EVP_CipherInit_ex"},
+{ERR_FUNC(EVP_F_EVP_CIPHER_CTX_COPY),	"EVP_CIPHER_CTX_copy"},
 {ERR_FUNC(EVP_F_EVP_CIPHER_CTX_CTRL),	"EVP_CIPHER_CTX_ctrl"},
 {ERR_FUNC(EVP_F_EVP_CIPHER_CTX_SET_KEY_LENGTH),	"EVP_CIPHER_CTX_set_key_length"},
 {ERR_FUNC(EVP_F_EVP_DECRYPTFINAL_EX),	"EVP_DecryptFinal_ex"},
-{ERR_FUNC(EVP_F_EVP_DIGESTINIT),	"EVP_DigestInit"},
 {ERR_FUNC(EVP_F_EVP_DIGESTINIT_EX),	"EVP_DigestInit_ex"},
 {ERR_FUNC(EVP_F_EVP_ENCRYPTFINAL_EX),	"EVP_EncryptFinal_ex"},
 {ERR_FUNC(EVP_F_EVP_MD_CTX_COPY_EX),	"EVP_MD_CTX_copy_ex"},
+{ERR_FUNC(EVP_F_EVP_MD_SIZE),	"EVP_MD_size"},
 {ERR_FUNC(EVP_F_EVP_OPENINIT),	"EVP_OpenInit"},
 {ERR_FUNC(EVP_F_EVP_PBE_ALG_ADD),	"EVP_PBE_alg_add"},
+{ERR_FUNC(EVP_F_EVP_PBE_ALG_ADD_TYPE),	"EVP_PBE_alg_add_type"},
 {ERR_FUNC(EVP_F_EVP_PBE_CIPHERINIT),	"EVP_PBE_CipherInit"},
 {ERR_FUNC(EVP_F_EVP_PKCS82PKEY),	"EVP_PKCS82PKEY"},
+{ERR_FUNC(EVP_F_EVP_PKCS82PKEY_BROKEN),	"EVP_PKCS82PKEY_BROKEN"},
 {ERR_FUNC(EVP_F_EVP_PKEY2PKCS8_BROKEN),	"EVP_PKEY2PKCS8_broken"},
 {ERR_FUNC(EVP_F_EVP_PKEY_COPY_PARAMETERS),	"EVP_PKEY_copy_parameters"},
+{ERR_FUNC(EVP_F_EVP_PKEY_CTX_CTRL),	"EVP_PKEY_CTX_ctrl"},
+{ERR_FUNC(EVP_F_EVP_PKEY_CTX_CTRL_STR),	"EVP_PKEY_CTX_ctrl_str"},
+{ERR_FUNC(EVP_F_EVP_PKEY_CTX_DUP),	"EVP_PKEY_CTX_dup"},
 {ERR_FUNC(EVP_F_EVP_PKEY_DECRYPT),	"EVP_PKEY_decrypt"},
+{ERR_FUNC(EVP_F_EVP_PKEY_DECRYPT_INIT),	"EVP_PKEY_decrypt_init"},
+{ERR_FUNC(EVP_F_EVP_PKEY_DECRYPT_OLD),	"EVP_PKEY_decrypt_old"},
+{ERR_FUNC(EVP_F_EVP_PKEY_DERIVE),	"EVP_PKEY_derive"},
+{ERR_FUNC(EVP_F_EVP_PKEY_DERIVE_INIT),	"EVP_PKEY_derive_init"},
+{ERR_FUNC(EVP_F_EVP_PKEY_DERIVE_SET_PEER),	"EVP_PKEY_derive_set_peer"},
 {ERR_FUNC(EVP_F_EVP_PKEY_ENCRYPT),	"EVP_PKEY_encrypt"},
+{ERR_FUNC(EVP_F_EVP_PKEY_ENCRYPT_INIT),	"EVP_PKEY_encrypt_init"},
+{ERR_FUNC(EVP_F_EVP_PKEY_ENCRYPT_OLD),	"EVP_PKEY_encrypt_old"},
 {ERR_FUNC(EVP_F_EVP_PKEY_GET1_DH),	"EVP_PKEY_get1_DH"},
 {ERR_FUNC(EVP_F_EVP_PKEY_GET1_DSA),	"EVP_PKEY_get1_DSA"},
 {ERR_FUNC(EVP_F_EVP_PKEY_GET1_ECDSA),	"EVP_PKEY_GET1_ECDSA"},
 {ERR_FUNC(EVP_F_EVP_PKEY_GET1_EC_KEY),	"EVP_PKEY_get1_EC_KEY"},
 {ERR_FUNC(EVP_F_EVP_PKEY_GET1_RSA),	"EVP_PKEY_get1_RSA"},
+{ERR_FUNC(EVP_F_EVP_PKEY_KEYGEN),	"EVP_PKEY_keygen"},
+{ERR_FUNC(EVP_F_EVP_PKEY_KEYGEN_INIT),	"EVP_PKEY_keygen_init"},
 {ERR_FUNC(EVP_F_EVP_PKEY_NEW),	"EVP_PKEY_new"},
+{ERR_FUNC(EVP_F_EVP_PKEY_PARAMGEN),	"EVP_PKEY_paramgen"},
+{ERR_FUNC(EVP_F_EVP_PKEY_PARAMGEN_INIT),	"EVP_PKEY_paramgen_init"},
+{ERR_FUNC(EVP_F_EVP_PKEY_SIGN),	"EVP_PKEY_sign"},
+{ERR_FUNC(EVP_F_EVP_PKEY_SIGN_INIT),	"EVP_PKEY_sign_init"},
+{ERR_FUNC(EVP_F_EVP_PKEY_VERIFY),	"EVP_PKEY_verify"},
+{ERR_FUNC(EVP_F_EVP_PKEY_VERIFY_INIT),	"EVP_PKEY_verify_init"},
+{ERR_FUNC(EVP_F_EVP_PKEY_VERIFY_RECOVER),	"EVP_PKEY_verify_recover"},
+{ERR_FUNC(EVP_F_EVP_PKEY_VERIFY_RECOVER_INIT),	"EVP_PKEY_verify_recover_init"},
 {ERR_FUNC(EVP_F_EVP_RIJNDAEL),	"EVP_RIJNDAEL"},
 {ERR_FUNC(EVP_F_EVP_SIGNFINAL),	"EVP_SignFinal"},
 {ERR_FUNC(EVP_F_EVP_VERIFYFINAL),	"EVP_VerifyFinal"},
+{ERR_FUNC(EVP_F_FIPS_CIPHERINIT),	"FIPS_CIPHERINIT"},
+{ERR_FUNC(EVP_F_FIPS_CIPHER_CTX_COPY),	"FIPS_CIPHER_CTX_COPY"},
+{ERR_FUNC(EVP_F_FIPS_CIPHER_CTX_CTRL),	"FIPS_CIPHER_CTX_CTRL"},
+{ERR_FUNC(EVP_F_FIPS_CIPHER_CTX_SET_KEY_LENGTH),	"FIPS_CIPHER_CTX_SET_KEY_LENGTH"},
+{ERR_FUNC(EVP_F_FIPS_DIGESTINIT),	"FIPS_DIGESTINIT"},
+{ERR_FUNC(EVP_F_FIPS_MD_CTX_COPY),	"FIPS_MD_CTX_COPY"},
+{ERR_FUNC(EVP_F_HMAC_INIT_EX),	"HMAC_Init_ex"},
+{ERR_FUNC(EVP_F_INT_CTX_NEW),	"INT_CTX_NEW"},
 {ERR_FUNC(EVP_F_PKCS5_PBE_KEYIVGEN),	"PKCS5_PBE_keyivgen"},
 {ERR_FUNC(EVP_F_PKCS5_V2_PBE_KEYIVGEN),	"PKCS5_v2_PBE_keyivgen"},
+{ERR_FUNC(EVP_F_PKCS5_V2_PBKDF2_KEYIVGEN),	"PKCS5_V2_PBKDF2_KEYIVGEN"},
 {ERR_FUNC(EVP_F_PKCS8_SET_BROKEN),	"PKCS8_set_broken"},
+{ERR_FUNC(EVP_F_PKEY_SET_TYPE),	"PKEY_SET_TYPE"},
 {ERR_FUNC(EVP_F_RC2_MAGIC_TO_METH),	"RC2_MAGIC_TO_METH"},
 {ERR_FUNC(EVP_F_RC5_CTRL),	"RC5_CTRL"},
 {0,NULL}
@@ -118,6 +151,7 @@ static ERR_STRING_DATA EVP_str_functs[]=
 
 static ERR_STRING_DATA EVP_str_reasons[]=
 	{
+{ERR_REASON(EVP_R_AES_IV_SETUP_FAILED)   ,"aes iv setup failed"},
 {ERR_REASON(EVP_R_AES_KEY_SETUP_FAILED)  ,"aes key setup failed"},
 {ERR_REASON(EVP_R_ASN1_LIB)              ,"asn1 lib"},
 {ERR_REASON(EVP_R_BAD_BLOCK_LENGTH)      ,"bad block length"},
@@ -125,42 +159,54 @@ static ERR_STRING_DATA EVP_str_reasons[]=
 {ERR_REASON(EVP_R_BAD_KEY_LENGTH)        ,"bad key length"},
 {ERR_REASON(EVP_R_BN_DECODE_ERROR)       ,"bn decode error"},
 {ERR_REASON(EVP_R_BN_PUBKEY_ERROR)       ,"bn pubkey error"},
+{ERR_REASON(EVP_R_BUFFER_TOO_SMALL)      ,"buffer too small"},
 {ERR_REASON(EVP_R_CAMELLIA_KEY_SETUP_FAILED),"camellia key setup failed"},
 {ERR_REASON(EVP_R_CIPHER_PARAMETER_ERROR),"cipher parameter error"},
+{ERR_REASON(EVP_R_COMMAND_NOT_SUPPORTED) ,"command not supported"},
 {ERR_REASON(EVP_R_CTRL_NOT_IMPLEMENTED)  ,"ctrl not implemented"},
 {ERR_REASON(EVP_R_CTRL_OPERATION_NOT_IMPLEMENTED),"ctrl operation not implemented"},
 {ERR_REASON(EVP_R_DATA_NOT_MULTIPLE_OF_BLOCK_LENGTH),"data not multiple of block length"},
 {ERR_REASON(EVP_R_DECODE_ERROR)          ,"decode error"},
 {ERR_REASON(EVP_R_DIFFERENT_KEY_TYPES)   ,"different key types"},
+{ERR_REASON(EVP_R_DIFFERENT_PARAMETERS)  ,"different parameters"},
 {ERR_REASON(EVP_R_DISABLED_FOR_FIPS)     ,"disabled for fips"},
 {ERR_REASON(EVP_R_ENCODE_ERROR)          ,"encode error"},
-{ERR_REASON(EVP_R_ERROR_LOADING_SECTION) ,"error loading section"},
-{ERR_REASON(EVP_R_ERROR_SETTING_FIPS_MODE),"error setting fips mode"},
 {ERR_REASON(EVP_R_EVP_PBE_CIPHERINIT_ERROR),"evp pbe cipherinit error"},
 {ERR_REASON(EVP_R_EXPECTING_AN_RSA_KEY)  ,"expecting an rsa key"},
 {ERR_REASON(EVP_R_EXPECTING_A_DH_KEY)    ,"expecting a dh key"},
 {ERR_REASON(EVP_R_EXPECTING_A_DSA_KEY)   ,"expecting a dsa key"},
 {ERR_REASON(EVP_R_EXPECTING_A_ECDSA_KEY) ,"expecting a ecdsa key"},
 {ERR_REASON(EVP_R_EXPECTING_A_EC_KEY)    ,"expecting a ec key"},
-{ERR_REASON(EVP_R_FIPS_MODE_NOT_SUPPORTED),"fips mode not supported"},
 {ERR_REASON(EVP_R_INITIALIZATION_ERROR)  ,"initialization error"},
 {ERR_REASON(EVP_R_INPUT_NOT_INITIALIZED) ,"input not initialized"},
-{ERR_REASON(EVP_R_INVALID_FIPS_MODE)     ,"invalid fips mode"},
+{ERR_REASON(EVP_R_INVALID_DIGEST)        ,"invalid digest"},
 {ERR_REASON(EVP_R_INVALID_KEY_LENGTH)    ,"invalid key length"},
+{ERR_REASON(EVP_R_INVALID_OPERATION)     ,"invalid operation"},
 {ERR_REASON(EVP_R_IV_TOO_LARGE)          ,"iv too large"},
 {ERR_REASON(EVP_R_KEYGEN_FAILURE)        ,"keygen failure"},
+{ERR_REASON(EVP_R_MESSAGE_DIGEST_IS_NULL),"message digest is null"},
+{ERR_REASON(EVP_R_METHOD_NOT_SUPPORTED)  ,"method not supported"},
 {ERR_REASON(EVP_R_MISSING_PARAMETERS)    ,"missing parameters"},
 {ERR_REASON(EVP_R_NO_CIPHER_SET)         ,"no cipher set"},
+{ERR_REASON(EVP_R_NO_DEFAULT_DIGEST)     ,"no default digest"},
 {ERR_REASON(EVP_R_NO_DIGEST_SET)         ,"no digest set"},
 {ERR_REASON(EVP_R_NO_DSA_PARAMETERS)     ,"no dsa parameters"},
+{ERR_REASON(EVP_R_NO_KEY_SET)            ,"no key set"},
+{ERR_REASON(EVP_R_NO_OPERATION_SET)      ,"no operation set"},
 {ERR_REASON(EVP_R_NO_SIGN_FUNCTION_CONFIGURED),"no sign function configured"},
 {ERR_REASON(EVP_R_NO_VERIFY_FUNCTION_CONFIGURED),"no verify function configured"},
+{ERR_REASON(EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE),"operation not supported for this keytype"},
+{ERR_REASON(EVP_R_OPERATON_NOT_INITIALIZED),"operaton not initialized"},
 {ERR_REASON(EVP_R_PKCS8_UNKNOWN_BROKEN_TYPE),"pkcs8 unknown broken type"},
+{ERR_REASON(EVP_R_PRIVATE_KEY_DECODE_ERROR),"private key decode error"},
+{ERR_REASON(EVP_R_PRIVATE_KEY_ENCODE_ERROR),"private key encode error"},
 {ERR_REASON(EVP_R_PUBLIC_KEY_NOT_RSA)    ,"public key not rsa"},
-{ERR_REASON(EVP_R_SEED_KEY_SETUP_FAILED) ,"seed key setup failed"},
-{ERR_REASON(EVP_R_UNKNOWN_OPTION)        ,"unknown option"},
+{ERR_REASON(EVP_R_TOO_LARGE)             ,"too large"},
+{ERR_REASON(EVP_R_UNKNOWN_CIPHER)        ,"unknown cipher"},
+{ERR_REASON(EVP_R_UNKNOWN_DIGEST)        ,"unknown digest"},
 {ERR_REASON(EVP_R_UNKNOWN_PBE_ALGORITHM) ,"unknown pbe algorithm"},
 {ERR_REASON(EVP_R_UNSUPORTED_NUMBER_OF_ROUNDS),"unsuported number of rounds"},
+{ERR_REASON(EVP_R_UNSUPPORTED_ALGORITHM) ,"unsupported algorithm"},
 {ERR_REASON(EVP_R_UNSUPPORTED_CIPHER)    ,"unsupported cipher"},
 {ERR_REASON(EVP_R_UNSUPPORTED_KEYLENGTH) ,"unsupported keylength"},
 {ERR_REASON(EVP_R_UNSUPPORTED_KEY_DERIVATION_FUNCTION),"unsupported key derivation function"},
diff --git a/crypto/evp/evp_fips.c b/crypto/evp/evp_fips.c
new file mode 100644
index 000000000000..cb7f4fc0faf4
--- /dev/null
+++ b/crypto/evp/evp_fips.c
@@ -0,0 +1,113 @@
+/* crypto/evp/evp_fips.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project.
+ */
+/* ====================================================================
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+
+#include <openssl/evp.h>
+
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+
+const EVP_CIPHER *EVP_aes_128_cbc(void)  { return FIPS_evp_aes_128_cbc(); }
+const EVP_CIPHER *EVP_aes_128_ccm(void)  { return FIPS_evp_aes_128_ccm(); }
+const EVP_CIPHER *EVP_aes_128_cfb1(void)  { return FIPS_evp_aes_128_cfb1(); }
+const EVP_CIPHER *EVP_aes_128_cfb128(void)  { return FIPS_evp_aes_128_cfb128(); }
+const EVP_CIPHER *EVP_aes_128_cfb8(void)  { return FIPS_evp_aes_128_cfb8(); }
+const EVP_CIPHER *EVP_aes_128_ctr(void)  { return FIPS_evp_aes_128_ctr(); }
+const EVP_CIPHER *EVP_aes_128_ecb(void)  { return FIPS_evp_aes_128_ecb(); }
+const EVP_CIPHER *EVP_aes_128_gcm(void)  { return FIPS_evp_aes_128_gcm(); }
+const EVP_CIPHER *EVP_aes_128_ofb(void)  { return FIPS_evp_aes_128_ofb(); }
+const EVP_CIPHER *EVP_aes_128_xts(void)  { return FIPS_evp_aes_128_xts(); }
+const EVP_CIPHER *EVP_aes_192_cbc(void)  { return FIPS_evp_aes_192_cbc(); }
+const EVP_CIPHER *EVP_aes_192_ccm(void)  { return FIPS_evp_aes_192_ccm(); }
+const EVP_CIPHER *EVP_aes_192_cfb1(void)  { return FIPS_evp_aes_192_cfb1(); }
+const EVP_CIPHER *EVP_aes_192_cfb128(void)  { return FIPS_evp_aes_192_cfb128(); }
+const EVP_CIPHER *EVP_aes_192_cfb8(void)  { return FIPS_evp_aes_192_cfb8(); }
+const EVP_CIPHER *EVP_aes_192_ctr(void)  { return FIPS_evp_aes_192_ctr(); }
+const EVP_CIPHER *EVP_aes_192_ecb(void)  { return FIPS_evp_aes_192_ecb(); }
+const EVP_CIPHER *EVP_aes_192_gcm(void)  { return FIPS_evp_aes_192_gcm(); }
+const EVP_CIPHER *EVP_aes_192_ofb(void)  { return FIPS_evp_aes_192_ofb(); }
+const EVP_CIPHER *EVP_aes_256_cbc(void)  { return FIPS_evp_aes_256_cbc(); }
+const EVP_CIPHER *EVP_aes_256_ccm(void)  { return FIPS_evp_aes_256_ccm(); }
+const EVP_CIPHER *EVP_aes_256_cfb1(void)  { return FIPS_evp_aes_256_cfb1(); }
+const EVP_CIPHER *EVP_aes_256_cfb128(void)  { return FIPS_evp_aes_256_cfb128(); }
+const EVP_CIPHER *EVP_aes_256_cfb8(void)  { return FIPS_evp_aes_256_cfb8(); }
+const EVP_CIPHER *EVP_aes_256_ctr(void)  { return FIPS_evp_aes_256_ctr(); }
+const EVP_CIPHER *EVP_aes_256_ecb(void)  { return FIPS_evp_aes_256_ecb(); }
+const EVP_CIPHER *EVP_aes_256_gcm(void)  { return FIPS_evp_aes_256_gcm(); }
+const EVP_CIPHER *EVP_aes_256_ofb(void)  { return FIPS_evp_aes_256_ofb(); }
+const EVP_CIPHER *EVP_aes_256_xts(void)  { return FIPS_evp_aes_256_xts(); }
+const EVP_CIPHER *EVP_des_ede(void)  { return FIPS_evp_des_ede(); }
+const EVP_CIPHER *EVP_des_ede3(void)  { return FIPS_evp_des_ede3(); }
+const EVP_CIPHER *EVP_des_ede3_cbc(void)  { return FIPS_evp_des_ede3_cbc(); }
+const EVP_CIPHER *EVP_des_ede3_cfb1(void)  { return FIPS_evp_des_ede3_cfb1(); }
+const EVP_CIPHER *EVP_des_ede3_cfb64(void)  { return FIPS_evp_des_ede3_cfb64(); }
+const EVP_CIPHER *EVP_des_ede3_cfb8(void)  { return FIPS_evp_des_ede3_cfb8(); }
+const EVP_CIPHER *EVP_des_ede3_ecb(void)  { return FIPS_evp_des_ede3_ecb(); }
+const EVP_CIPHER *EVP_des_ede3_ofb(void)  { return FIPS_evp_des_ede3_ofb(); }
+const EVP_CIPHER *EVP_des_ede_cbc(void)  { return FIPS_evp_des_ede_cbc(); }
+const EVP_CIPHER *EVP_des_ede_cfb64(void)  { return FIPS_evp_des_ede_cfb64(); }
+const EVP_CIPHER *EVP_des_ede_ecb(void)  { return FIPS_evp_des_ede_ecb(); }
+const EVP_CIPHER *EVP_des_ede_ofb(void)  { return FIPS_evp_des_ede_ofb(); }
+const EVP_CIPHER *EVP_enc_null(void)  { return FIPS_evp_enc_null(); }
+
+const EVP_MD *EVP_sha1(void)  { return FIPS_evp_sha1(); }
+const EVP_MD *EVP_sha224(void)  { return FIPS_evp_sha224(); }
+const EVP_MD *EVP_sha256(void)  { return FIPS_evp_sha256(); }
+const EVP_MD *EVP_sha384(void)  { return FIPS_evp_sha384(); }
+const EVP_MD *EVP_sha512(void)  { return FIPS_evp_sha512(); }
+
+const EVP_MD *EVP_dss(void)  { return FIPS_evp_dss(); }
+const EVP_MD *EVP_dss1(void)  { return FIPS_evp_dss1(); }
+const EVP_MD *EVP_ecdsa(void)  { return FIPS_evp_ecdsa(); }
+
+#endif
diff --git a/crypto/evp/evp_key.c b/crypto/evp/evp_key.c
index 361ea69ab6d5..7961fbebf2e7 100644
--- a/crypto/evp/evp_key.c
+++ b/crypto/evp/evp_key.c
@@ -90,6 +90,11 @@ char *EVP_get_pw_prompt(void)
  * this function will fail */
 int EVP_read_pw_string(char *buf, int len, const char *prompt, int verify)
 	{
+	return EVP_read_pw_string_min(buf, 0, len, prompt, verify);
+	}
+
+int EVP_read_pw_string_min(char *buf, int min, int len, const char *prompt, int verify)
+	{
 	int ret;
 	char buff[BUFSIZ];
 	UI *ui;
@@ -97,10 +102,10 @@ int EVP_read_pw_string(char *buf, int len, const char *prompt, int verify)
 	if ((prompt == NULL) && (prompt_string[0] != '\0'))
 		prompt=prompt_string;
 	ui = UI_new();
-	UI_add_input_string(ui,prompt,0,buf,0,(len>=BUFSIZ)?BUFSIZ-1:len);
+	UI_add_input_string(ui,prompt,0,buf,min,(len>=BUFSIZ)?BUFSIZ-1:len);
 	if (verify)
 		UI_add_verify_string(ui,prompt,0,
-			buff,0,(len>=BUFSIZ)?BUFSIZ-1:len,buf);
+			buff,min,(len>=BUFSIZ)?BUFSIZ-1:len,buf);
 	ret = UI_process(ui);
 	UI_free(ui);
 	OPENSSL_cleanse(buff,BUFSIZ);
@@ -115,7 +120,7 @@ int EVP_BytesToKey(const EVP_CIPHER *type, const EVP_MD *md,
 	unsigned char md_buf[EVP_MAX_MD_SIZE];
 	int niv,nkey,addmd=0;
 	unsigned int mds=0,i;
-
+	int rv = 0;
 	nkey=type->key_len;
 	niv=type->iv_len;
 	OPENSSL_assert(nkey <= EVP_MAX_KEY_LENGTH);
@@ -129,17 +134,24 @@ int EVP_BytesToKey(const EVP_CIPHER *type, const EVP_MD *md,
 		if (!EVP_DigestInit_ex(&c,md, NULL))
 			return 0;
 		if (addmd++)
-			EVP_DigestUpdate(&c,&(md_buf[0]),mds);
-		EVP_DigestUpdate(&c,data,datal);
+			if (!EVP_DigestUpdate(&c,&(md_buf[0]),mds))
+				goto err;
+		if (!EVP_DigestUpdate(&c,data,datal))
+			goto err;
 		if (salt != NULL)
-			EVP_DigestUpdate(&c,salt,PKCS5_SALT_LEN);
-		EVP_DigestFinal_ex(&c,&(md_buf[0]),&mds);
+			if (!EVP_DigestUpdate(&c,salt,PKCS5_SALT_LEN))
+				goto err;
+		if (!EVP_DigestFinal_ex(&c,&(md_buf[0]),&mds))
+			goto err;
 
 		for (i=1; i<(unsigned int)count; i++)
 			{
-			EVP_DigestInit_ex(&c,md, NULL);
-			EVP_DigestUpdate(&c,&(md_buf[0]),mds);
-			EVP_DigestFinal_ex(&c,&(md_buf[0]),&mds);
+			if (!EVP_DigestInit_ex(&c,md, NULL))
+				goto err;
+			if (!EVP_DigestUpdate(&c,&(md_buf[0]),mds))
+				goto err;
+			if (!EVP_DigestFinal_ex(&c,&(md_buf[0]),&mds))
+				goto err;
 			}
 		i=0;
 		if (nkey)
@@ -168,8 +180,10 @@ int EVP_BytesToKey(const EVP_CIPHER *type, const EVP_MD *md,
 			}
 		if ((nkey == 0) && (niv == 0)) break;
 		}
+	rv = type->key_len;
+	err:
 	EVP_MD_CTX_cleanup(&c);
 	OPENSSL_cleanse(&(md_buf[0]),EVP_MAX_MD_SIZE);
-	return(type->key_len);
+	return rv;
 	}
 
diff --git a/crypto/evp/evp_lib.c b/crypto/evp/evp_lib.c
index 9c200617a65f..b180e4828a8c 100644
--- a/crypto/evp/evp_lib.c
+++ b/crypto/evp/evp_lib.c
@@ -188,6 +188,11 @@ int EVP_CIPHER_CTX_block_size(const EVP_CIPHER_CTX *ctx)
 	return ctx->cipher->block_size;
 	}
 
+int EVP_Cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, unsigned int inl)
+	{
+	return ctx->cipher->do_cipher(ctx,out,in,inl);
+	}
+
 const EVP_CIPHER *EVP_CIPHER_CTX_cipher(const EVP_CIPHER_CTX *ctx)
 	{
 	return ctx->cipher;
@@ -198,6 +203,11 @@ unsigned long EVP_CIPHER_flags(const EVP_CIPHER *cipher)
 	return cipher->flags;
 	}
 
+unsigned long EVP_CIPHER_CTX_flags(const EVP_CIPHER_CTX *ctx)
+	{
+	return ctx->cipher->flags;
+	}
+
 void *EVP_CIPHER_CTX_get_app_data(const EVP_CIPHER_CTX *ctx)
 	{
 	return ctx->app_data;
@@ -213,6 +223,11 @@ int EVP_CIPHER_iv_length(const EVP_CIPHER *cipher)
 	return cipher->iv_len;
 	}
 
+int EVP_CIPHER_CTX_iv_length(const EVP_CIPHER_CTX *ctx)
+	{
+	return ctx->cipher->iv_len;
+	}
+
 int EVP_CIPHER_key_length(const EVP_CIPHER *cipher)
 	{
 	return cipher->key_len;
@@ -223,6 +238,11 @@ int EVP_CIPHER_CTX_key_length(const EVP_CIPHER_CTX *ctx)
 	return ctx->key_len;
 	}
 
+int EVP_CIPHER_nid(const EVP_CIPHER *cipher)
+	{
+	return cipher->nid;
+	}
+
 int EVP_CIPHER_CTX_nid(const EVP_CIPHER_CTX *ctx)
 	{
 	return ctx->cipher->nid;
@@ -245,11 +265,23 @@ int EVP_MD_pkey_type(const EVP_MD *md)
 
 int EVP_MD_size(const EVP_MD *md)
 	{
+	if (!md)
+		{
+		EVPerr(EVP_F_EVP_MD_SIZE, EVP_R_MESSAGE_DIGEST_IS_NULL);
+		return -1;
+		}
 	return md->md_size;
 	}
 
-const EVP_MD * EVP_MD_CTX_md(const EVP_MD_CTX *ctx)
+unsigned long EVP_MD_flags(const EVP_MD *md)
+	{
+	return md->flags;
+	}
+
+const EVP_MD *EVP_MD_CTX_md(const EVP_MD_CTX *ctx)
 	{
+	if (!ctx)
+		return NULL;
 	return ctx->digest;
 	}
 
diff --git a/crypto/evp/evp_locl.h b/crypto/evp/evp_locl.h
index ef6c432538aa..08c0a66d39c7 100644
--- a/crypto/evp/evp_locl.h
+++ b/crypto/evp/evp_locl.h
@@ -61,38 +61,66 @@
 /* Wrapper functions for each cipher mode */
 
 #define BLOCK_CIPHER_ecb_loop() \
-	unsigned int i, bl; \
+	size_t i, bl; \
 	bl = ctx->cipher->block_size;\
 	if(inl < bl) return 1;\
 	inl -= bl; \
 	for(i=0; i <= inl; i+=bl) 
 
 #define BLOCK_CIPHER_func_ecb(cname, cprefix, kstruct, ksched) \
-static int cname##_ecb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, unsigned int inl) \
+static int cname##_ecb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, size_t inl) \
 {\
 	BLOCK_CIPHER_ecb_loop() \
 		cprefix##_ecb_encrypt(in + i, out + i, &((kstruct *)ctx->cipher_data)->ksched, ctx->encrypt);\
 	return 1;\
 }
 
+#define EVP_MAXCHUNK ((size_t)1<<(sizeof(long)*8-2))
+
 #define BLOCK_CIPHER_func_ofb(cname, cprefix, cbits, kstruct, ksched) \
-static int cname##_ofb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, unsigned int inl) \
+static int cname##_ofb_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, size_t inl) \
 {\
-	cprefix##_ofb##cbits##_encrypt(in, out, (long)inl, &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, &ctx->num);\
+	while(inl>=EVP_MAXCHUNK)\
+	    {\
+	    cprefix##_ofb##cbits##_encrypt(in, out, (long)EVP_MAXCHUNK, &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, &ctx->num);\
+	    inl-=EVP_MAXCHUNK;\
+	    in +=EVP_MAXCHUNK;\
+	    out+=EVP_MAXCHUNK;\
+	    }\
+	if (inl)\
+	    cprefix##_ofb##cbits##_encrypt(in, out, (long)inl, &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, &ctx->num);\
 	return 1;\
 }
 
 #define BLOCK_CIPHER_func_cbc(cname, cprefix, kstruct, ksched) \
-static int cname##_cbc_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, unsigned int inl) \
+static int cname##_cbc_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, size_t inl) \
 {\
-	cprefix##_cbc_encrypt(in, out, (long)inl, &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, ctx->encrypt);\
+	while(inl>=EVP_MAXCHUNK) \
+	    {\
+	    cprefix##_cbc_encrypt(in, out, (long)EVP_MAXCHUNK, &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, ctx->encrypt);\
+	    inl-=EVP_MAXCHUNK;\
+	    in +=EVP_MAXCHUNK;\
+	    out+=EVP_MAXCHUNK;\
+	    }\
+	if (inl)\
+	    cprefix##_cbc_encrypt(in, out, (long)inl, &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, ctx->encrypt);\
 	return 1;\
 }
 
 #define BLOCK_CIPHER_func_cfb(cname, cprefix, cbits, kstruct, ksched) \
-static int cname##_cfb##cbits##_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, unsigned int inl) \
+static int cname##_cfb##cbits##_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, size_t inl) \
 {\
-	cprefix##_cfb##cbits##_encrypt(in, out, (long)((cbits==1) && !(ctx->flags & EVP_CIPH_FLAG_LENGTH_BITS) ?inl*8:inl), &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, &ctx->num, ctx->encrypt);\
+	size_t chunk=EVP_MAXCHUNK;\
+	if (cbits==1)  chunk>>=3;\
+	if (inl<chunk) chunk=inl;\
+	while(inl && inl>=chunk)\
+	    {\
+            cprefix##_cfb##cbits##_encrypt(in, out, (long)((cbits==1) && !(ctx->flags & EVP_CIPH_FLAG_LENGTH_BITS) ?inl*8:inl), &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, &ctx->num, ctx->encrypt);\
+	    inl-=chunk;\
+	    in +=chunk;\
+	    out+=chunk;\
+	    if(inl<chunk) chunk=inl;\
+	    }\
 	return 1;\
 }
 
@@ -226,27 +254,132 @@ const EVP_CIPHER *EVP_##cname##_ecb(void) { return &cname##_ecb; }
 
 #define EVP_C_DATA(kstruct, ctx)	((kstruct *)(ctx)->cipher_data)
 
-#define IMPLEMENT_CFBR(cipher,cprefix,kstruct,ksched,keysize,cbits,iv_len,fl) \
+#define IMPLEMENT_CFBR(cipher,cprefix,kstruct,ksched,keysize,cbits,iv_len) \
 	BLOCK_CIPHER_func_cfb(cipher##_##keysize,cprefix,cbits,kstruct,ksched) \
 	BLOCK_CIPHER_def_cfb(cipher##_##keysize,kstruct, \
 			     NID_##cipher##_##keysize, keysize/8, iv_len, cbits, \
-			     (fl)|EVP_CIPH_FLAG_DEFAULT_ASN1, \
-			     cipher##_init_key, NULL, NULL, NULL, NULL)
+			     0, cipher##_init_key, NULL, \
+			     EVP_CIPHER_set_asn1_iv, \
+			     EVP_CIPHER_get_asn1_iv, \
+			     NULL)
+
+struct evp_pkey_ctx_st
+	{
+	/* Method associated with this operation */
+	const EVP_PKEY_METHOD *pmeth;
+	/* Engine that implements this method or NULL if builtin */
+	ENGINE *engine;
+	/* Key: may be NULL */
+	EVP_PKEY *pkey;
+	/* Peer key for key agreement, may be NULL */
+	EVP_PKEY *peerkey;
+	/* Actual operation */
+	int operation;
+	/* Algorithm specific data */
+	void *data;
+	/* Application specific data */
+	void *app_data;
+	/* Keygen callback */
+	EVP_PKEY_gen_cb *pkey_gencb;
+	/* implementation specific keygen data */
+	int *keygen_info;
+	int keygen_info_count;
+	} /* EVP_PKEY_CTX */;
+
+#define EVP_PKEY_FLAG_DYNAMIC	1
+
+struct evp_pkey_method_st
+	{
+	int pkey_id;
+	int flags;
+
+	int (*init)(EVP_PKEY_CTX *ctx);
+	int (*copy)(EVP_PKEY_CTX *dst, EVP_PKEY_CTX *src);
+	void (*cleanup)(EVP_PKEY_CTX *ctx);
+
+	int (*paramgen_init)(EVP_PKEY_CTX *ctx);
+	int (*paramgen)(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey);
+
+	int (*keygen_init)(EVP_PKEY_CTX *ctx);
+	int (*keygen)(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey);
+
+	int (*sign_init)(EVP_PKEY_CTX *ctx);
+	int (*sign)(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+				const unsigned char *tbs, size_t tbslen);
+
+	int (*verify_init)(EVP_PKEY_CTX *ctx);
+	int (*verify)(EVP_PKEY_CTX *ctx,
+				const unsigned char *sig, size_t siglen,
+				const unsigned char *tbs, size_t tbslen);
+
+	int (*verify_recover_init)(EVP_PKEY_CTX *ctx);
+	int (*verify_recover)(EVP_PKEY_CTX *ctx,
+				unsigned char *rout, size_t *routlen,
+				const unsigned char *sig, size_t siglen);
+
+	int (*signctx_init)(EVP_PKEY_CTX *ctx, EVP_MD_CTX *mctx);
+	int (*signctx)(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					EVP_MD_CTX *mctx);
+
+	int (*verifyctx_init)(EVP_PKEY_CTX *ctx, EVP_MD_CTX *mctx);
+	int (*verifyctx)(EVP_PKEY_CTX *ctx, const unsigned char *sig,int siglen,
+					EVP_MD_CTX *mctx);
+
+	int (*encrypt_init)(EVP_PKEY_CTX *ctx);
+	int (*encrypt)(EVP_PKEY_CTX *ctx, unsigned char *out, size_t *outlen,
+					const unsigned char *in, size_t inlen);
+
+	int (*decrypt_init)(EVP_PKEY_CTX *ctx);
+	int (*decrypt)(EVP_PKEY_CTX *ctx, unsigned char *out, size_t *outlen,
+					const unsigned char *in, size_t inlen);
+
+	int (*derive_init)(EVP_PKEY_CTX *ctx);
+	int (*derive)(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen);
+
+	int (*ctrl)(EVP_PKEY_CTX *ctx, int type, int p1, void *p2);
+	int (*ctrl_str)(EVP_PKEY_CTX *ctx, const char *type, const char *value);
+
+
+	} /* EVP_PKEY_METHOD */;
+
+void evp_pkey_set_cb_translate(BN_GENCB *cb, EVP_PKEY_CTX *ctx);
+
+int PKCS5_v2_PBKDF2_keyivgen(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
+			     ASN1_TYPE *param,
+			     const EVP_CIPHER *c, const EVP_MD *md, int en_de);
 
 #ifdef OPENSSL_FIPS
-#define RC2_set_key	private_RC2_set_key
-#define RC4_set_key	private_RC4_set_key
-#define CAST_set_key	private_CAST_set_key
-#define RC5_32_set_key	private_RC5_32_set_key
-#define BF_set_key	private_BF_set_key
-#define Camellia_set_key private_Camellia_set_key
-#define idea_set_encrypt_key private_idea_set_encrypt_key
 
+#ifdef OPENSSL_DOING_MAKEDEPEND
+#undef SHA1_Init
+#undef SHA1_Update
+#undef SHA224_Init
+#undef SHA256_Init
+#undef SHA384_Init
+#undef SHA512_Init
+#undef DES_set_key_unchecked
+#endif
+
+#define RIPEMD160_Init	private_RIPEMD160_Init
+#define WHIRLPOOL_Init	private_WHIRLPOOL_Init
 #define MD5_Init	private_MD5_Init
 #define MD4_Init	private_MD4_Init
 #define MD2_Init	private_MD2_Init
 #define MDC2_Init	private_MDC2_Init
 #define SHA_Init	private_SHA_Init
+#define SHA1_Init	private_SHA1_Init
+#define SHA224_Init	private_SHA224_Init
+#define SHA256_Init	private_SHA256_Init
+#define SHA384_Init	private_SHA384_Init
+#define SHA512_Init	private_SHA512_Init
 
-#endif
+#define BF_set_key	private_BF_set_key
+#define CAST_set_key	private_CAST_set_key
+#define idea_set_encrypt_key	private_idea_set_encrypt_key
+#define SEED_set_key	private_SEED_set_key
+#define RC2_set_key	private_RC2_set_key
+#define RC4_set_key	private_RC4_set_key
+#define DES_set_key_unchecked	private_DES_set_key_unchecked
+#define Camellia_set_key	private_Camellia_set_key
 
+#endif
diff --git a/crypto/evp/evp_pbe.c b/crypto/evp/evp_pbe.c
index 766ea42b5aac..f8c32d825e79 100644
--- a/crypto/evp/evp_pbe.c
+++ b/crypto/evp/evp_pbe.c
@@ -3,7 +3,7 @@
  * project 1999.
  */
 /* ====================================================================
- * Copyright (c) 1999 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -59,112 +59,258 @@
 #include <stdio.h>
 #include "cryptlib.h"
 #include <openssl/evp.h>
+#include <openssl/pkcs12.h>
 #include <openssl/x509.h>
+#include "evp_locl.h"
 
 /* Password based encryption (PBE) functions */
 
-static STACK *pbe_algs;
+DECLARE_STACK_OF(EVP_PBE_CTL)
+static STACK_OF(EVP_PBE_CTL) *pbe_algs;
 
 /* Setup a cipher context from a PBE algorithm */
 
-typedef struct {
-int pbe_nid;
-const EVP_CIPHER *cipher;
-const EVP_MD *md;
-EVP_PBE_KEYGEN *keygen;
-} EVP_PBE_CTL;
+typedef struct
+	{
+	int pbe_type;
+	int pbe_nid;
+	int cipher_nid;
+	int md_nid;
+	EVP_PBE_KEYGEN *keygen;
+	} EVP_PBE_CTL;
 
-int EVP_PBE_CipherInit(ASN1_OBJECT *pbe_obj, const char *pass, int passlen,
-	     ASN1_TYPE *param, EVP_CIPHER_CTX *ctx, int en_de)
-{
+static const EVP_PBE_CTL builtin_pbe[] = 
+	{
+	{EVP_PBE_TYPE_OUTER, NID_pbeWithMD2AndDES_CBC,
+			NID_des_cbc, NID_md2, PKCS5_PBE_keyivgen},
+	{EVP_PBE_TYPE_OUTER, NID_pbeWithMD5AndDES_CBC,
+			NID_des_cbc, NID_md5, PKCS5_PBE_keyivgen},
+	{EVP_PBE_TYPE_OUTER, NID_pbeWithSHA1AndRC2_CBC,
+			NID_rc2_64_cbc, NID_sha1, PKCS5_PBE_keyivgen},
 
-	EVP_PBE_CTL *pbetmp, pbelu;
-	int i;
-	pbelu.pbe_nid = OBJ_obj2nid(pbe_obj);
-	if (pbelu.pbe_nid != NID_undef) i = sk_find(pbe_algs, (char *)&pbelu);
-	else i = -1;
+#ifndef OPENSSL_NO_HMAC
+	{EVP_PBE_TYPE_OUTER, NID_id_pbkdf2, -1, -1, PKCS5_v2_PBKDF2_keyivgen},
+#endif
+
+	{EVP_PBE_TYPE_OUTER, NID_pbe_WithSHA1And128BitRC4,
+			NID_rc4, NID_sha1, PKCS12_PBE_keyivgen},
+	{EVP_PBE_TYPE_OUTER, NID_pbe_WithSHA1And40BitRC4,
+			NID_rc4_40, NID_sha1, PKCS12_PBE_keyivgen},
+	{EVP_PBE_TYPE_OUTER, NID_pbe_WithSHA1And3_Key_TripleDES_CBC,
+		 	NID_des_ede3_cbc, NID_sha1, PKCS12_PBE_keyivgen},
+	{EVP_PBE_TYPE_OUTER, NID_pbe_WithSHA1And2_Key_TripleDES_CBC, 
+			NID_des_ede_cbc, NID_sha1, PKCS12_PBE_keyivgen},
+	{EVP_PBE_TYPE_OUTER, NID_pbe_WithSHA1And128BitRC2_CBC,
+			NID_rc2_cbc, NID_sha1, PKCS12_PBE_keyivgen},
+	{EVP_PBE_TYPE_OUTER, NID_pbe_WithSHA1And40BitRC2_CBC,
+			NID_rc2_40_cbc, NID_sha1, PKCS12_PBE_keyivgen},
+
+#ifndef OPENSSL_NO_HMAC
+	{EVP_PBE_TYPE_OUTER, NID_pbes2, -1, -1, PKCS5_v2_PBE_keyivgen},
+#endif
+	{EVP_PBE_TYPE_OUTER, NID_pbeWithMD2AndRC2_CBC,
+			NID_rc2_64_cbc, NID_md2, PKCS5_PBE_keyivgen},
+	{EVP_PBE_TYPE_OUTER, NID_pbeWithMD5AndRC2_CBC,
+			NID_rc2_64_cbc, NID_md5, PKCS5_PBE_keyivgen},
+	{EVP_PBE_TYPE_OUTER, NID_pbeWithSHA1AndDES_CBC,
+			NID_des_cbc, NID_sha1, PKCS5_PBE_keyivgen},
+
+
+	{EVP_PBE_TYPE_PRF, NID_hmacWithSHA1, -1, NID_sha1, 0},
+	{EVP_PBE_TYPE_PRF, NID_hmacWithMD5, -1, NID_md5, 0},
+	{EVP_PBE_TYPE_PRF, NID_hmacWithSHA224, -1, NID_sha224, 0},
+	{EVP_PBE_TYPE_PRF, NID_hmacWithSHA256, -1, NID_sha256, 0},
+	{EVP_PBE_TYPE_PRF, NID_hmacWithSHA384, -1, NID_sha384, 0},
+	{EVP_PBE_TYPE_PRF, NID_hmacWithSHA512, -1, NID_sha512, 0},
+	{EVP_PBE_TYPE_PRF, NID_id_HMACGostR3411_94, -1, NID_id_GostR3411_94, 0},
+	};
 
-	if (i == -1) {
+#ifdef TEST
+int main(int argc, char **argv)
+	{
+	int i, nid_md, nid_cipher;
+	EVP_PBE_CTL *tpbe, *tpbe2;
+	/*OpenSSL_add_all_algorithms();*/
+
+	for (i = 0; i < sizeof(builtin_pbe)/sizeof(EVP_PBE_CTL); i++)
+		{
+		tpbe = builtin_pbe + i;
+		fprintf(stderr, "%d %d %s ", tpbe->pbe_type, tpbe->pbe_nid,
+						OBJ_nid2sn(tpbe->pbe_nid));
+		if (EVP_PBE_find(tpbe->pbe_type, tpbe->pbe_nid,
+					&nid_cipher ,&nid_md,0))
+			fprintf(stderr, "Found %s %s\n",
+					OBJ_nid2sn(nid_cipher),
+					OBJ_nid2sn(nid_md));
+		else
+			fprintf(stderr, "Find ERROR!!\n");
+		}
+
+	return 0;
+	}
+#endif
+		
+
+
+int EVP_PBE_CipherInit(ASN1_OBJECT *pbe_obj, const char *pass, int passlen,
+		       ASN1_TYPE *param, EVP_CIPHER_CTX *ctx, int en_de)
+	{
+	const EVP_CIPHER *cipher;
+	const EVP_MD *md;
+	int cipher_nid, md_nid;
+	EVP_PBE_KEYGEN *keygen;
+
+	if (!EVP_PBE_find(EVP_PBE_TYPE_OUTER, OBJ_obj2nid(pbe_obj),
+					&cipher_nid, &md_nid, &keygen))
+		{
 		char obj_tmp[80];
 		EVPerr(EVP_F_EVP_PBE_CIPHERINIT,EVP_R_UNKNOWN_PBE_ALGORITHM);
 		if (!pbe_obj) BUF_strlcpy (obj_tmp, "NULL", sizeof obj_tmp);
 		else i2t_ASN1_OBJECT(obj_tmp, sizeof obj_tmp, pbe_obj);
 		ERR_add_error_data(2, "TYPE=", obj_tmp);
 		return 0;
-	}
-	if(!pass) passlen = 0;
-	else if (passlen == -1) passlen = strlen(pass);
-	pbetmp = (EVP_PBE_CTL *)sk_value (pbe_algs, i);
-	i = (*pbetmp->keygen)(ctx, pass, passlen, param, pbetmp->cipher,
-						 pbetmp->md, en_de);
-	if (!i) {
-		EVPerr(EVP_F_EVP_PBE_CIPHERINIT,EVP_R_KEYGEN_FAILURE);
-		return 0;
-	}
-	return 1;	
-}
-
-static int pbe_cmp(const char * const *a, const char * const *b)
-{
-	const EVP_PBE_CTL * const *pbe1 = (const EVP_PBE_CTL * const *) a,
-			* const *pbe2 = (const EVP_PBE_CTL * const *)b;
-	return ((*pbe1)->pbe_nid - (*pbe2)->pbe_nid);
-}
+		}
 
-/* Add a PBE algorithm */
+	if(!pass)
+		passlen = 0;
+	else if (passlen == -1)
+		passlen = strlen(pass);
 
-int EVP_PBE_alg_add(int nid, const EVP_CIPHER *cipher, const EVP_MD *md,
-	     EVP_PBE_KEYGEN *keygen)
-{
-	EVP_PBE_CTL *pbe_tmp = NULL, pbelu;
-	int i;
-	if (!pbe_algs)
+	if (cipher_nid == -1)
+		cipher = NULL;
+	else
 		{
-		pbe_algs = sk_new(pbe_cmp);
-		if (!pbe_algs)
+		cipher = EVP_get_cipherbynid(cipher_nid);
+		if (!cipher)
 			{
-			EVPerr(EVP_F_EVP_PBE_ALG_ADD,ERR_R_MALLOC_FAILURE);
+			EVPerr(EVP_F_EVP_PBE_CIPHERINIT,EVP_R_UNKNOWN_CIPHER);
 			return 0;
 			}
 		}
+
+	if (md_nid == -1)
+		md = NULL;
 	else
 		{
-		/* Check if already present */
-		pbelu.pbe_nid = nid;
-		i = sk_find(pbe_algs, (char *)&pbelu);
-		if (i >= 0)
+		md = EVP_get_digestbynid(md_nid);
+		if (!md)
 			{
-			pbe_tmp = (EVP_PBE_CTL *)sk_value(pbe_algs, i);
-			/* If everything identical leave alone */
-			if (pbe_tmp->cipher == cipher
-				&& pbe_tmp->md == md
-				&& pbe_tmp->keygen == keygen)
-				return 1;
+			EVPerr(EVP_F_EVP_PBE_CIPHERINIT,EVP_R_UNKNOWN_DIGEST);
+			return 0;
 			}
 		}
 
-	if (!pbe_tmp)
+	if (!keygen(ctx, pass, passlen, param, cipher, md, en_de))
 		{
-		pbe_tmp = OPENSSL_malloc (sizeof(EVP_PBE_CTL));
-		if (!pbe_tmp)
-			{
-			EVPerr(EVP_F_EVP_PBE_ALG_ADD,ERR_R_MALLOC_FAILURE);
-			return 0;
-			}
-		/* If adding a new PBE, set nid, append and sort */
-		pbe_tmp->pbe_nid = nid;
-		sk_push (pbe_algs, (char *)pbe_tmp);
-		sk_sort(pbe_algs);
+		EVPerr(EVP_F_EVP_PBE_CIPHERINIT,EVP_R_KEYGEN_FAILURE);
+		return 0;
 		}
-		
-	pbe_tmp->cipher = cipher;
-	pbe_tmp->md = md;
+	return 1;	
+}
+
+DECLARE_OBJ_BSEARCH_CMP_FN(EVP_PBE_CTL, EVP_PBE_CTL, pbe2);
+
+static int pbe2_cmp(const EVP_PBE_CTL *pbe1, const EVP_PBE_CTL *pbe2)
+	{
+	int ret = pbe1->pbe_type - pbe2->pbe_type;
+	if (ret)
+		return ret;
+	else
+		return pbe1->pbe_nid - pbe2->pbe_nid;
+	}
+
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(EVP_PBE_CTL, EVP_PBE_CTL, pbe2);
+
+static int pbe_cmp(const EVP_PBE_CTL * const *a, const EVP_PBE_CTL * const *b)
+	{
+	int ret = (*a)->pbe_type - (*b)->pbe_type;
+	if (ret)
+		return ret;
+	else
+		return (*a)->pbe_nid - (*b)->pbe_nid;
+	}
+
+/* Add a PBE algorithm */
+
+int EVP_PBE_alg_add_type(int pbe_type, int pbe_nid, int cipher_nid, int md_nid,
+			 EVP_PBE_KEYGEN *keygen)
+	{
+	EVP_PBE_CTL *pbe_tmp;
+	if (!pbe_algs)
+		pbe_algs = sk_EVP_PBE_CTL_new(pbe_cmp);
+	if (!(pbe_tmp = (EVP_PBE_CTL*) OPENSSL_malloc (sizeof(EVP_PBE_CTL))))
+		{
+		EVPerr(EVP_F_EVP_PBE_ALG_ADD_TYPE,ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	pbe_tmp->pbe_type = pbe_type;
+	pbe_tmp->pbe_nid = pbe_nid;
+	pbe_tmp->cipher_nid = cipher_nid;
+	pbe_tmp->md_nid = md_nid;
 	pbe_tmp->keygen = keygen;
+
+
+	sk_EVP_PBE_CTL_push (pbe_algs, pbe_tmp);
 	return 1;
-}
+	}
+
+int EVP_PBE_alg_add(int nid, const EVP_CIPHER *cipher, const EVP_MD *md,
+		    EVP_PBE_KEYGEN *keygen)
+	{
+	int cipher_nid, md_nid;
+	if (cipher)
+		cipher_nid = EVP_CIPHER_type(cipher);
+	else
+		cipher_nid = -1;
+	if (md)
+		md_nid = EVP_MD_type(md);
+	else
+		md_nid = -1;
+
+	return EVP_PBE_alg_add_type(EVP_PBE_TYPE_OUTER, nid,
+					cipher_nid, md_nid, keygen);
+	}
+
+int EVP_PBE_find(int type, int pbe_nid,
+		 int *pcnid, int *pmnid, EVP_PBE_KEYGEN **pkeygen)
+	{
+	EVP_PBE_CTL *pbetmp = NULL, pbelu;
+	int i;
+	if (pbe_nid == NID_undef)
+		return 0;
+
+	pbelu.pbe_type = type;
+	pbelu.pbe_nid = pbe_nid;
+
+	if (pbe_algs)
+		{
+		i = sk_EVP_PBE_CTL_find(pbe_algs, &pbelu);
+		if (i != -1)
+			pbetmp = sk_EVP_PBE_CTL_value (pbe_algs, i);
+		}
+	if (pbetmp == NULL)
+		{
+		pbetmp = OBJ_bsearch_pbe2(&pbelu, builtin_pbe,
+				     sizeof(builtin_pbe)/sizeof(EVP_PBE_CTL));
+		}
+	if (pbetmp == NULL)
+		return 0;
+	if (pcnid)
+		*pcnid = pbetmp->cipher_nid;
+	if (pmnid)
+		*pmnid = pbetmp->md_nid;
+	if (pkeygen)
+		*pkeygen = pbetmp->keygen;
+	return 1;
+	}
+
+static void free_evp_pbe_ctl(EVP_PBE_CTL *pbe)
+	 {
+	 OPENSSL_freeFunc(pbe);
+	 }
 
 void EVP_PBE_cleanup(void)
-{
-	sk_pop_free(pbe_algs, OPENSSL_freeFunc);
+	{
+	sk_EVP_PBE_CTL_pop_free(pbe_algs, free_evp_pbe_ctl);
 	pbe_algs = NULL;
-}
+	}
diff --git a/crypto/evp/evp_pkey.c b/crypto/evp/evp_pkey.c
index 10d9e9e772b8..ceebf6928488 100644
--- a/crypto/evp/evp_pkey.c
+++ b/crypto/evp/evp_pkey.c
@@ -3,7 +3,7 @@
  * project 1999.
  */
 /* ====================================================================
- * Copyright (c) 1999-2002 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -61,287 +61,52 @@
 #include "cryptlib.h"
 #include <openssl/x509.h>
 #include <openssl/rand.h>
-#ifndef OPENSSL_NO_RSA
-#include <openssl/rsa.h>
-#endif
-#ifndef OPENSSL_NO_DSA
-#include <openssl/dsa.h>
-#endif
-#include <openssl/bn.h>
-
-#ifndef OPENSSL_NO_DSA
-static int dsa_pkey2pkcs8(PKCS8_PRIV_KEY_INFO *p8inf, EVP_PKEY *pkey);
-#endif
-#ifndef OPENSSL_NO_EC
-static int eckey_pkey2pkcs8(PKCS8_PRIV_KEY_INFO *p8inf, EVP_PKEY *pkey);
-#endif
+#include "asn1_locl.h"
 
 /* Extract a private key from a PKCS8 structure */
 
 EVP_PKEY *EVP_PKCS82PKEY(PKCS8_PRIV_KEY_INFO *p8)
 {
 	EVP_PKEY *pkey = NULL;
-#ifndef OPENSSL_NO_RSA
-	RSA *rsa = NULL;
-#endif
-#ifndef OPENSSL_NO_DSA
-	DSA *dsa = NULL;
-	ASN1_TYPE *t1, *t2;
-	ASN1_INTEGER *privkey;
-	STACK_OF(ASN1_TYPE) *ndsa = NULL;
-#endif
-#ifndef OPENSSL_NO_EC
-	EC_KEY *eckey = NULL;
-	const unsigned char *p_tmp;
-#endif
-#if !defined(OPENSSL_NO_DSA) || !defined(OPENSSL_NO_EC)
-	ASN1_TYPE    *param = NULL;	
-	BN_CTX *ctx = NULL;
-	int plen;
-#endif
-	X509_ALGOR *a;
-	const unsigned char *p;
-	const unsigned char *cp;
-	int pkeylen;
-	int  nid;
+	ASN1_OBJECT *algoid;
 	char obj_tmp[80];
 
-	if(p8->pkey->type == V_ASN1_OCTET_STRING) {
-		p8->broken = PKCS8_OK;
-		p = p8->pkey->value.octet_string->data;
-		pkeylen = p8->pkey->value.octet_string->length;
-	} else {
-		p8->broken = PKCS8_NO_OCTET;
-		p = p8->pkey->value.sequence->data;
-		pkeylen = p8->pkey->value.sequence->length;
-	}
+	if (!PKCS8_pkey_get0(&algoid, NULL, NULL, NULL, p8))
+		return NULL;
+
 	if (!(pkey = EVP_PKEY_new())) {
 		EVPerr(EVP_F_EVP_PKCS82PKEY,ERR_R_MALLOC_FAILURE);
 		return NULL;
 	}
-	a = p8->pkeyalg;
-	nid = OBJ_obj2nid(a->algorithm);
-	switch(nid)
-	{
-#ifndef OPENSSL_NO_RSA
-		case NID_rsaEncryption:
-		cp = p;
-		if (!(rsa = d2i_RSAPrivateKey (NULL,&cp, pkeylen))) {
-			EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-			return NULL;
-		}
-		EVP_PKEY_assign_RSA (pkey, rsa);
-		break;
-#endif
-#ifndef OPENSSL_NO_DSA
-		case NID_dsa:
-		/* PKCS#8 DSA is weird: you just get a private key integer
-	         * and parameters in the AlgorithmIdentifier the pubkey must
-		 * be recalculated.
-		 */
-	
-		/* Check for broken DSA PKCS#8, UGH! */
-		if(*p == (V_ASN1_SEQUENCE|V_ASN1_CONSTRUCTED)) {
-		    if(!(ndsa = ASN1_seq_unpack_ASN1_TYPE(p, pkeylen, 
-							  d2i_ASN1_TYPE,
-							  ASN1_TYPE_free))) {
-			EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-			goto dsaerr;
-		    }
-		    if(sk_ASN1_TYPE_num(ndsa) != 2 ) {
-			EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-			goto dsaerr;
-		    }
-		    /* Handle Two broken types:
-		     * SEQUENCE {parameters, priv_key}
-		     * SEQUENCE {pub_key, priv_key}
-		     */
-
-		    t1 = sk_ASN1_TYPE_value(ndsa, 0);
-		    t2 = sk_ASN1_TYPE_value(ndsa, 1);
-		    if(t1->type == V_ASN1_SEQUENCE) {
-			p8->broken = PKCS8_EMBEDDED_PARAM;
-			param = t1;
-		    } else if(a->parameter->type == V_ASN1_SEQUENCE) {
-			p8->broken = PKCS8_NS_DB;
-			param = a->parameter;
-		    } else {
-			EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-			goto dsaerr;
-		    }
-
-		    if(t2->type != V_ASN1_INTEGER) {
-			EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-			goto dsaerr;
-		    }
-		    privkey = t2->value.integer;
-		} else {
-			if (!(privkey=d2i_ASN1_INTEGER (NULL, &p, pkeylen))) {
-				EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-				goto dsaerr;
-			}
-			param = p8->pkeyalg->parameter;
-		}
-		if (!param || (param->type != V_ASN1_SEQUENCE)) {
-			EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-			goto dsaerr;
-		}
-		cp = p = param->value.sequence->data;
-		plen = param->value.sequence->length;
-		if (!(dsa = d2i_DSAparams (NULL, &cp, plen))) {
-			EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-			goto dsaerr;
-		}
-		/* We have parameters now set private key */
-		if (!(dsa->priv_key = ASN1_INTEGER_to_BN(privkey, NULL))) {
-			EVPerr(EVP_F_EVP_PKCS82PKEY,EVP_R_BN_DECODE_ERROR);
-			goto dsaerr;
-		}
-		/* Calculate public key (ouch!) */
-		if (!(dsa->pub_key = BN_new())) {
-			EVPerr(EVP_F_EVP_PKCS82PKEY,ERR_R_MALLOC_FAILURE);
-			goto dsaerr;
-		}
-		if (!(ctx = BN_CTX_new())) {
-			EVPerr(EVP_F_EVP_PKCS82PKEY,ERR_R_MALLOC_FAILURE);
-			goto dsaerr;
-		}
-			
-		if (!BN_mod_exp(dsa->pub_key, dsa->g,
-						 dsa->priv_key, dsa->p, ctx)) {
-			
-			EVPerr(EVP_F_EVP_PKCS82PKEY,EVP_R_BN_PUBKEY_ERROR);
-			goto dsaerr;
-		}
 
-		EVP_PKEY_assign_DSA(pkey, dsa);
-		BN_CTX_free (ctx);
-		if(ndsa) sk_ASN1_TYPE_pop_free(ndsa, ASN1_TYPE_free);
-		else ASN1_INTEGER_free(privkey);
-		break;
-		dsaerr:
-		BN_CTX_free (ctx);
-		sk_ASN1_TYPE_pop_free(ndsa, ASN1_TYPE_free);
-		DSA_free(dsa);
-		EVP_PKEY_free(pkey);
-		return NULL;
-		break;
-#endif
-#ifndef OPENSSL_NO_EC
-		case NID_X9_62_id_ecPublicKey:
-		p_tmp = p;
-		/* extract the ec parameters */
-		param = p8->pkeyalg->parameter;
-
-		if (!param || ((param->type != V_ASN1_SEQUENCE) &&
-		    (param->type != V_ASN1_OBJECT)))
+	if (!EVP_PKEY_set_type(pkey, OBJ_obj2nid(algoid)))
 		{
-			EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-			goto ecerr;
+		EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_UNSUPPORTED_PRIVATE_KEY_ALGORITHM);
+		i2t_ASN1_OBJECT(obj_tmp, 80, algoid);
+		ERR_add_error_data(2, "TYPE=", obj_tmp);
+		goto error;
 		}
 
-		if (param->type == V_ASN1_SEQUENCE)
+	if (pkey->ameth->priv_decode)
 		{
-			cp = p = param->value.sequence->data;
-			plen = param->value.sequence->length;
-
-			if (!(eckey = d2i_ECParameters(NULL, &cp, plen)))
+		if (!pkey->ameth->priv_decode(pkey, p8))
 			{
-				EVPerr(EVP_F_EVP_PKCS82PKEY,
-					EVP_R_DECODE_ERROR);
-				goto ecerr;
+			EVPerr(EVP_F_EVP_PKCS82PKEY,
+					EVP_R_PRIVATE_KEY_DECODE_ERROR);
+			goto error;
 			}
 		}
-		else
+	else
 		{
-			EC_GROUP *group;
-			cp = p = param->value.object->data;
-			plen = param->value.object->length;
-
-			/* type == V_ASN1_OBJECT => the parameters are given
-			 * by an asn1 OID
-			 */
-			if ((eckey = EC_KEY_new()) == NULL)
-			{
-				EVPerr(EVP_F_EVP_PKCS82PKEY,
-					ERR_R_MALLOC_FAILURE);
-				goto ecerr;
-			}
-			group = EC_GROUP_new_by_curve_name(OBJ_obj2nid(a->parameter->value.object));
-			if (group == NULL)
-				goto ecerr;
-			EC_GROUP_set_asn1_flag(group, OPENSSL_EC_NAMED_CURVE);
-			if (EC_KEY_set_group(eckey, group) == 0)
-				goto ecerr;
-			EC_GROUP_free(group);
-		}
-
-		/* We have parameters now set private key */
-		if (!d2i_ECPrivateKey(&eckey, &p_tmp, pkeylen))
-		{
-			EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_DECODE_ERROR);
-			goto ecerr;
-		}
-
-		/* calculate public key (if necessary) */
-		if (EC_KEY_get0_public_key(eckey) == NULL)
-		{
-			const BIGNUM *priv_key;
-			const EC_GROUP *group;
-			EC_POINT *pub_key;
-			/* the public key was not included in the SEC1 private
-			 * key => calculate the public key */
-			group   = EC_KEY_get0_group(eckey);
-			pub_key = EC_POINT_new(group);
-			if (pub_key == NULL)
-			{
-				EVPerr(EVP_F_EVP_PKCS82PKEY, ERR_R_EC_LIB);
-				goto ecerr;
-			}
-			if (!EC_POINT_copy(pub_key, EC_GROUP_get0_generator(group)))
-			{
-				EC_POINT_free(pub_key);
-				EVPerr(EVP_F_EVP_PKCS82PKEY, ERR_R_EC_LIB);
-				goto ecerr;
-			}
-			priv_key = EC_KEY_get0_private_key(eckey);
-			if (!EC_POINT_mul(group, pub_key, priv_key, NULL, NULL, ctx))
-			{
-				EC_POINT_free(pub_key);
-				EVPerr(EVP_F_EVP_PKCS82PKEY, ERR_R_EC_LIB);
-				goto ecerr;
-			}
-			if (EC_KEY_set_public_key(eckey, pub_key) == 0)
-			{
-				EC_POINT_free(pub_key);
-				EVPerr(EVP_F_EVP_PKCS82PKEY, ERR_R_EC_LIB);
-				goto ecerr;
-			}
-			EC_POINT_free(pub_key);
+		EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_METHOD_NOT_SUPPORTED);
+		goto error;
 		}
 
-		EVP_PKEY_assign_EC_KEY(pkey, eckey);
-		if (ctx)
-			BN_CTX_free(ctx);
-		break;
-ecerr:
-		if (ctx)
-			BN_CTX_free(ctx);
-		if (eckey)
-			EC_KEY_free(eckey);
-		if (pkey)
-			EVP_PKEY_free(pkey);
-		return NULL;
-#endif
-		default:
-		EVPerr(EVP_F_EVP_PKCS82PKEY, EVP_R_UNSUPPORTED_PRIVATE_KEY_ALGORITHM);
-		if (!a->algorithm) BUF_strlcpy (obj_tmp, "NULL", sizeof obj_tmp);
-		else i2t_ASN1_OBJECT(obj_tmp, 80, a->algorithm);
-		ERR_add_error_data(2, "TYPE=", obj_tmp);
-		EVP_PKEY_free (pkey);
-		return NULL;
-	}
 	return pkey;
+
+	error:
+	EVP_PKEY_free (pkey);
+	return NULL;
 }
 
 PKCS8_PRIV_KEY_INFO *EVP_PKEY2PKCS8(EVP_PKEY *pkey)
@@ -360,59 +125,37 @@ PKCS8_PRIV_KEY_INFO *EVP_PKEY2PKCS8_broken(EVP_PKEY *pkey, int broken)
 		return NULL;
 	}
 	p8->broken = broken;
-	if (!ASN1_INTEGER_set(p8->version, 0)) {
-		EVPerr(EVP_F_EVP_PKEY2PKCS8_BROKEN,ERR_R_MALLOC_FAILURE);
-		PKCS8_PRIV_KEY_INFO_free (p8);
-		return NULL;
-	}
-	if (!(p8->pkeyalg->parameter = ASN1_TYPE_new ())) {
-		EVPerr(EVP_F_EVP_PKEY2PKCS8_BROKEN,ERR_R_MALLOC_FAILURE);
-		PKCS8_PRIV_KEY_INFO_free (p8);
-		return NULL;
-	}
-	p8->pkey->type = V_ASN1_OCTET_STRING;
-	switch (EVP_PKEY_type(pkey->type)) {
-#ifndef OPENSSL_NO_RSA
-		case EVP_PKEY_RSA:
 
-		if(p8->broken == PKCS8_NO_OCTET) p8->pkey->type = V_ASN1_SEQUENCE;
-
-		p8->pkeyalg->algorithm = OBJ_nid2obj(NID_rsaEncryption);
-		p8->pkeyalg->parameter->type = V_ASN1_NULL;
-		if (!ASN1_pack_string_of (EVP_PKEY,pkey, i2d_PrivateKey,
-					 &p8->pkey->value.octet_string)) {
-			EVPerr(EVP_F_EVP_PKEY2PKCS8_BROKEN,ERR_R_MALLOC_FAILURE);
-			PKCS8_PRIV_KEY_INFO_free (p8);
-			return NULL;
-		}
-		break;
-#endif
-#ifndef OPENSSL_NO_DSA
-		case EVP_PKEY_DSA:
-		if(!dsa_pkey2pkcs8(p8, pkey)) {
-			PKCS8_PRIV_KEY_INFO_free (p8);
-			return NULL;
+	if (pkey->ameth)
+		{
+		if (pkey->ameth->priv_encode)
+			{
+			if (!pkey->ameth->priv_encode(p8, pkey))
+				{
+				EVPerr(EVP_F_EVP_PKEY2PKCS8_BROKEN,
+					EVP_R_PRIVATE_KEY_ENCODE_ERROR);
+				goto error;
+				}
+			}
+		else
+			{
+			EVPerr(EVP_F_EVP_PKEY2PKCS8_BROKEN,
+					EVP_R_METHOD_NOT_SUPPORTED);
+			goto error;
+			}
 		}
-
-		break;
-#endif
-#ifndef OPENSSL_NO_EC
-		case EVP_PKEY_EC:
-		if (!eckey_pkey2pkcs8(p8, pkey))
+	else
 		{
-			PKCS8_PRIV_KEY_INFO_free(p8);
-			return(NULL);
+		EVPerr(EVP_F_EVP_PKEY2PKCS8_BROKEN,
+				EVP_R_UNSUPPORTED_PRIVATE_KEY_ALGORITHM);
+		goto error;
 		}
-		break;
-#endif
-		default:
-		EVPerr(EVP_F_EVP_PKEY2PKCS8_BROKEN, EVP_R_UNSUPPORTED_PRIVATE_KEY_ALGORITHM);
-		PKCS8_PRIV_KEY_INFO_free (p8);
-		return NULL;
-	}
 	RAND_add(p8->pkey->value.octet_string->data,
 		 p8->pkey->value.octet_string->length, 0.0);
 	return p8;
+	error:
+	PKCS8_PRIV_KEY_INFO_free(p8);
+	return NULL;
 }
 
 PKCS8_PRIV_KEY_INFO *PKCS8_set_broken(PKCS8_PRIV_KEY_INFO *p8, int broken)
@@ -436,301 +179,6 @@ PKCS8_PRIV_KEY_INFO *PKCS8_set_broken(PKCS8_PRIV_KEY_INFO *p8, int broken)
 	}
 }
 
-#ifndef OPENSSL_NO_DSA
-static int dsa_pkey2pkcs8(PKCS8_PRIV_KEY_INFO *p8, EVP_PKEY *pkey)
-{
-	ASN1_STRING *params = NULL;
-	ASN1_INTEGER *prkey = NULL;
-	ASN1_TYPE *ttmp = NULL;
-	STACK_OF(ASN1_TYPE) *ndsa = NULL;
-	unsigned char *p = NULL, *q;
-	int len;
-
-	p8->pkeyalg->algorithm = OBJ_nid2obj(NID_dsa);
-	len = i2d_DSAparams (pkey->pkey.dsa, NULL);
-	if (!(p = OPENSSL_malloc(len))) {
-		EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-		goto err;
-	}
-	q = p;
-	i2d_DSAparams (pkey->pkey.dsa, &q);
-	if (!(params = ASN1_STRING_new())) {
-		EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-		goto err;
-	}
-	if (!ASN1_STRING_set(params, p, len)) {
-		EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-		goto err;
-	}
-	OPENSSL_free(p);
-	p = NULL;
-	/* Get private key into integer */
-	if (!(prkey = BN_to_ASN1_INTEGER (pkey->pkey.dsa->priv_key, NULL))) {
-		EVPerr(EVP_F_DSA_PKEY2PKCS8,EVP_R_ENCODE_ERROR);
-		goto err;
-	}
-
-	switch(p8->broken) {
-
-		case PKCS8_OK:
-		case PKCS8_NO_OCTET:
-
-		if (!ASN1_pack_string_of(ASN1_INTEGER,prkey, i2d_ASN1_INTEGER,
-					 &p8->pkey->value.octet_string)) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-
-		M_ASN1_INTEGER_free (prkey);
-		prkey = NULL;
-		p8->pkeyalg->parameter->value.sequence = params;
-		params = NULL;
-		p8->pkeyalg->parameter->type = V_ASN1_SEQUENCE;
-
-		break;
-
-		case PKCS8_NS_DB:
-
-		p8->pkeyalg->parameter->value.sequence = params;
-		params = NULL;
-		p8->pkeyalg->parameter->type = V_ASN1_SEQUENCE;
-		if (!(ndsa = sk_ASN1_TYPE_new_null())) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		if (!(ttmp = ASN1_TYPE_new())) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		if (!(ttmp->value.integer =
-			BN_to_ASN1_INTEGER(pkey->pkey.dsa->pub_key, NULL))) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,EVP_R_ENCODE_ERROR);
-			goto err;
-		}
-		ttmp->type = V_ASN1_INTEGER;
-		if (!sk_ASN1_TYPE_push(ndsa, ttmp)) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-
-		if (!(ttmp = ASN1_TYPE_new())) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		ttmp->value.integer = prkey;
-		prkey = NULL;
-		ttmp->type = V_ASN1_INTEGER;
-		if (!sk_ASN1_TYPE_push(ndsa, ttmp)) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		ttmp = NULL;
-
-		if (!(p8->pkey->value.octet_string = ASN1_OCTET_STRING_new())) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-
-		if (!ASN1_seq_pack_ASN1_TYPE(ndsa, i2d_ASN1_TYPE,
-					 &p8->pkey->value.octet_string->data,
-					 &p8->pkey->value.octet_string->length)) {
-
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		sk_ASN1_TYPE_pop_free(ndsa, ASN1_TYPE_free);
-		break;
-
-		case PKCS8_EMBEDDED_PARAM:
-
-		p8->pkeyalg->parameter->type = V_ASN1_NULL;
-		if (!(ndsa = sk_ASN1_TYPE_new_null())) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		if (!(ttmp = ASN1_TYPE_new())) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		ttmp->value.sequence = params;
-		params = NULL;
-		ttmp->type = V_ASN1_SEQUENCE;
-		if (!sk_ASN1_TYPE_push(ndsa, ttmp)) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-
-		if (!(ttmp = ASN1_TYPE_new())) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		ttmp->value.integer = prkey;
-		prkey = NULL;
-		ttmp->type = V_ASN1_INTEGER;
-		if (!sk_ASN1_TYPE_push(ndsa, ttmp)) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		ttmp = NULL;
-
-		if (!(p8->pkey->value.octet_string = ASN1_OCTET_STRING_new())) {
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-
-		if (!ASN1_seq_pack_ASN1_TYPE(ndsa, i2d_ASN1_TYPE,
-					 &p8->pkey->value.octet_string->data,
-					 &p8->pkey->value.octet_string->length)) {
-
-			EVPerr(EVP_F_DSA_PKEY2PKCS8,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-		sk_ASN1_TYPE_pop_free(ndsa, ASN1_TYPE_free);
-		break;
-	}
-	return 1;
-err:
-	if (p != NULL) OPENSSL_free(p);
-	if (params != NULL) ASN1_STRING_free(params);
-	if (prkey != NULL) M_ASN1_INTEGER_free(prkey);
-	if (ttmp != NULL) ASN1_TYPE_free(ttmp);
-	if (ndsa != NULL) sk_ASN1_TYPE_pop_free(ndsa, ASN1_TYPE_free);
-	return 0;
-}
-#endif
-
-#ifndef OPENSSL_NO_EC
-static int eckey_pkey2pkcs8(PKCS8_PRIV_KEY_INFO *p8, EVP_PKEY *pkey)
-{
-	EC_KEY		*ec_key;
-	const EC_GROUP  *group;
-	unsigned char	*p, *pp;
-	int 		nid, i, ret = 0;
-	unsigned int    tmp_flags, old_flags;
-
-	ec_key = pkey->pkey.ec;
-	if (ec_key == NULL || (group = EC_KEY_get0_group(ec_key)) == NULL) 
-	{
-		EVPerr(EVP_F_ECKEY_PKEY2PKCS8, EVP_R_MISSING_PARAMETERS);
-		return 0;
-	}
-
-	/* set the ec parameters OID */
-	if (p8->pkeyalg->algorithm)
-		ASN1_OBJECT_free(p8->pkeyalg->algorithm);
-
-	p8->pkeyalg->algorithm = OBJ_nid2obj(NID_X9_62_id_ecPublicKey);
-
-	/* set the ec parameters */
-
-	if (p8->pkeyalg->parameter)
-	{
-		ASN1_TYPE_free(p8->pkeyalg->parameter);
-		p8->pkeyalg->parameter = NULL;
-	}
-
-	if ((p8->pkeyalg->parameter = ASN1_TYPE_new()) == NULL)
-	{
-		EVPerr(EVP_F_ECKEY_PKEY2PKCS8, ERR_R_MALLOC_FAILURE);
-		return 0;
-	}
-	
-	if (EC_GROUP_get_asn1_flag(group)
-                     && (nid = EC_GROUP_get_curve_name(group)))
-	{
-		/* we have a 'named curve' => just set the OID */
-		p8->pkeyalg->parameter->type = V_ASN1_OBJECT;
-		p8->pkeyalg->parameter->value.object = OBJ_nid2obj(nid);
-	}
-	else	/* explicit parameters */
-	{
-		if ((i = i2d_ECParameters(ec_key, NULL)) == 0)
-		{
-			EVPerr(EVP_F_ECKEY_PKEY2PKCS8, ERR_R_EC_LIB);
-			return 0;
-		}
-		if ((p = (unsigned char *) OPENSSL_malloc(i)) == NULL)
-		{
-			EVPerr(EVP_F_ECKEY_PKEY2PKCS8, ERR_R_MALLOC_FAILURE);
-			return 0;
-		}	
-		pp = p;
-		if (!i2d_ECParameters(ec_key, &pp))
-		{
-			EVPerr(EVP_F_ECKEY_PKEY2PKCS8, ERR_R_EC_LIB);
-			OPENSSL_free(p);
-			return 0;
-		}
-		p8->pkeyalg->parameter->type = V_ASN1_SEQUENCE;
-		if ((p8->pkeyalg->parameter->value.sequence 
-			= ASN1_STRING_new()) == NULL)
-		{
-			EVPerr(EVP_F_ECKEY_PKEY2PKCS8, ERR_R_ASN1_LIB);
-			OPENSSL_free(p);
-			return 0;
-		}
-		ASN1_STRING_set(p8->pkeyalg->parameter->value.sequence, p, i);
-		OPENSSL_free(p);
-	}
-
-	/* set the private key */
-
-	/* do not include the parameters in the SEC1 private key
-	 * see PKCS#11 12.11 */
-	old_flags = EC_KEY_get_enc_flags(pkey->pkey.ec);
-	tmp_flags = old_flags | EC_PKEY_NO_PARAMETERS;
-	EC_KEY_set_enc_flags(pkey->pkey.ec, tmp_flags);
-	i = i2d_ECPrivateKey(pkey->pkey.ec, NULL);
-	if (!i)
-	{
-		EC_KEY_set_enc_flags(pkey->pkey.ec, old_flags);
-		EVPerr(EVP_F_ECKEY_PKEY2PKCS8, ERR_R_EC_LIB);
-		return 0;
-	}
-	p = (unsigned char *) OPENSSL_malloc(i);
-	if (!p)
-	{
-		EC_KEY_set_enc_flags(pkey->pkey.ec, old_flags);
-		EVPerr(EVP_F_ECKEY_PKEY2PKCS8, ERR_R_MALLOC_FAILURE);
-		return 0;
-	}
-	pp = p;
-	if (!i2d_ECPrivateKey(pkey->pkey.ec, &pp))
-	{
-		EC_KEY_set_enc_flags(pkey->pkey.ec, old_flags);
-		EVPerr(EVP_F_ECKEY_PKEY2PKCS8, ERR_R_EC_LIB);
-		OPENSSL_free(p);
-		return 0;
-	}
-	/* restore old encoding flags */
-	EC_KEY_set_enc_flags(pkey->pkey.ec, old_flags);
-
-	switch(p8->broken) {
-
-		case PKCS8_OK:
-		p8->pkey->value.octet_string = ASN1_OCTET_STRING_new();
-		if (!p8->pkey->value.octet_string ||
-		    !M_ASN1_OCTET_STRING_set(p8->pkey->value.octet_string,
-		    (const void *)p, i))
-
-		{
-			EVPerr(EVP_F_ECKEY_PKEY2PKCS8, ERR_R_MALLOC_FAILURE);
-		}
-		else
-			ret = 1;
-		break;
-		case PKCS8_NO_OCTET:		/* RSA specific */
-		case PKCS8_NS_DB:		/* DSA specific */
-		case PKCS8_EMBEDDED_PARAM:	/* DSA specific */
-		default:
-			EVPerr(EVP_F_ECKEY_PKEY2PKCS8,EVP_R_ENCODE_ERROR);
-	}
-	OPENSSL_cleanse(p, (size_t)i);
-	OPENSSL_free(p);
-	return ret;
-}
-#endif
-
 /* EVP_PKEY attribute functions */
 
 int EVP_PKEY_get_attr_count(const EVP_PKEY *key)
diff --git a/crypto/evp/evp_test.c b/crypto/evp/evp_test.c
index a36ca7e4b376..55c7cdfdccf2 100644
--- a/crypto/evp/evp_test.c
+++ b/crypto/evp/evp_test.c
@@ -153,8 +153,8 @@ static void test1(const EVP_CIPHER *c,const unsigned char *key,int kn,
     
     if(kn != c->key_len)
 	{
-	fprintf(stderr,"Key length doesn't match, got %d expected %d\n",kn,
-		c->key_len);
+	fprintf(stderr,"Key length doesn't match, got %d expected %lu\n",kn,
+		(unsigned long)c->key_len);
 	test1_exit(5);
 	}
     EVP_CIPHER_CTX_init(&ctx);
@@ -442,7 +442,7 @@ int main(int argc,char **argv)
 #endif
     EVP_cleanup();
     CRYPTO_cleanup_all_ex_data();
-    ERR_remove_state(0);
+    ERR_remove_thread_state(NULL);
     ERR_free_strings();
     CRYPTO_mem_leaks_fp(stderr);
 
diff --git a/crypto/evp/evptests.txt b/crypto/evp/evptests.txt
index beb12144b6ae..c273707c1444 100644
--- a/crypto/evp/evptests.txt
+++ b/crypto/evp/evptests.txt
@@ -158,6 +158,19 @@ AES-256-OFB:603DEB1015CA71BE2B73AEF0857D77811F352C073B6108D72D9810A30914DFF4:B7B
 AES-256-OFB:603DEB1015CA71BE2B73AEF0857D77811F352C073B6108D72D9810A30914DFF4:E1C656305ED1A7A6563805746FE03EDC:30C81C46A35CE411E5FBC1191A0A52EF:71AB47A086E86EEDF39D1C5BBA97C408:0
 AES-256-OFB:603DEB1015CA71BE2B73AEF0857D77811F352C073B6108D72D9810A30914DFF4:41635BE625B48AFC1666DD42A09D96E7:F69F2445DF4F9B17AD2B417BE66C3710:0126141D67F37BE8538F5A8BE740E484:0
 
+# AES Counter test vectors from RFC3686
+aes-128-ctr:AE6852F8121067CC4BF7A5765577F39E:00000030000000000000000000000001:53696E676C6520626C6F636B206D7367:E4095D4FB7A7B3792D6175A3261311B8:1
+aes-128-ctr:7E24067817FAE0D743D6CE1F32539163:006CB6DBC0543B59DA48D90B00000001:000102030405060708090A0B0C0D0E0F101112131415161718191A1B1C1D1E1F:5104A106168A72D9790D41EE8EDAD388EB2E1EFC46DA57C8FCE630DF9141BE28:1
+aes-128-ctr:7691BE035E5020A8AC6E618529F9A0DC:00E0017B27777F3F4A1786F000000001:000102030405060708090A0B0C0D0E0F101112131415161718191A1B1C1D1E1F20212223:C1CF48A89F2FFDD9CF4652E9EFDB72D74540A42BDE6D7836D59A5CEAAEF3105325B2072F:1
+
+aes-192-ctr:16AF5B145FC9F579C175F93E3BFB0EED863D06CCFDB78515:0000004836733C147D6D93CB00000001:53696E676C6520626C6F636B206D7367:4B55384FE259C9C84E7935A003CBE928:1
+aes-192-ctr:7C5CB2401B3DC33C19E7340819E0F69C678C3DB8E6F6A91A:0096B03B020C6EADC2CB500D00000001:000102030405060708090A0B0C0D0E0F101112131415161718191A1B1C1D1E1F:453243FC609B23327EDFAAFA7131CD9F8490701C5AD4A79CFC1FE0FF42F4FB00:1
+aes-192-ctr:02BF391EE8ECB159B959617B0965279BF59B60A786D3E0FE:0007BDFD5CBD60278DCC091200000001:000102030405060708090A0B0C0D0E0F101112131415161718191A1B1C1D1E1F20212223:96893FC55E5C722F540B7DD1DDF7E758D288BC95C69165884536C811662F2188ABEE0935:1
+
+aes-256-ctr:776BEFF2851DB06F4C8A0542C8696F6C6A81AF1EEC96B4D37FC1D689E6C1C104:00000060DB5672C97AA8F0B200000001:53696E676C6520626C6F636B206D7367:145AD01DBF824EC7560863DC71E3E0C0:1
+aes-256-ctr:F6D66D6BD52D59BB0796365879EFF886C66DD51A5B6A99744B50590C87A23884:00FAAC24C1585EF15A43D87500000001:000102030405060708090A0B0C0D0E0F101112131415161718191A1B1C1D1E1F:F05E231B3894612C49EE000B804EB2A9B8306B508F839D6A5530831D9344AF1C:1
+aes-256-ctr:FF7A617CE69148E4F1726E2F43581DE2AA62D9F805532EDFF1EED687FB54153D:001CC5B751A51D70A1C1114800000001:000102030405060708090A0B0C0D0E0F101112131415161718191A1B1C1D1E1F20212223:EB6C52821D0BBBF7CE7594462ACA4FAAB407DF866569FD07F48CC0B583D6071F1EC0E6B8:1
+
 # DES ECB tests (from destest)
 
 DES-ECB:0000000000000000::0000000000000000:8CA64DE9C1B123A7
diff --git a/crypto/evp/m_dss.c b/crypto/evp/m_dss.c
index 6b0c0aa7a3f5..4ad63ada6fe2 100644
--- a/crypto/evp/m_dss.c
+++ b/crypto/evp/m_dss.c
@@ -66,6 +66,7 @@
 #endif
 
 #ifndef OPENSSL_NO_SHA
+#ifndef OPENSSL_FIPS
 
 static int init(EVP_MD_CTX *ctx)
 	{ return SHA1_Init(ctx->md_data); }
@@ -81,7 +82,7 @@ static const EVP_MD dsa_md=
 	NID_dsaWithSHA,
 	NID_dsaWithSHA,
 	SHA_DIGEST_LENGTH,
-	EVP_MD_FLAG_FIPS,
+	EVP_MD_FLAG_PKEY_DIGEST,
 	init,
 	update,
 	final,
@@ -97,3 +98,4 @@ const EVP_MD *EVP_dss(void)
 	return(&dsa_md);
 	}
 #endif
+#endif
diff --git a/crypto/evp/m_dss1.c b/crypto/evp/m_dss1.c
index da8babc147dd..f80170efeb98 100644
--- a/crypto/evp/m_dss1.c
+++ b/crypto/evp/m_dss1.c
@@ -68,7 +68,7 @@
 #include <openssl/dsa.h>
 #endif
 
-#ifndef OPENSSL_FIPS
+#ifndef OPENSSL_FIPS 
 
 static int init(EVP_MD_CTX *ctx)
 	{ return SHA1_Init(ctx->md_data); }
@@ -84,7 +84,7 @@ static const EVP_MD dss1_md=
 	NID_dsa,
 	NID_dsaWithSHA1,
 	SHA_DIGEST_LENGTH,
-	0,
+	EVP_MD_FLAG_PKEY_DIGEST,
 	init,
 	update,
 	final,
diff --git a/crypto/evp/m_ecdsa.c b/crypto/evp/m_ecdsa.c
index fad270faca2b..4b15fb0f6cec 100644
--- a/crypto/evp/m_ecdsa.c
+++ b/crypto/evp/m_ecdsa.c
@@ -116,6 +116,8 @@
 #include <openssl/x509.h>
 
 #ifndef OPENSSL_NO_SHA
+#ifndef OPENSSL_FIPS
+
 static int init(EVP_MD_CTX *ctx)
 	{ return SHA1_Init(ctx->md_data); }
 
@@ -130,7 +132,7 @@ static const EVP_MD ecdsa_md=
 	NID_ecdsa_with_SHA1,
 	NID_ecdsa_with_SHA1,
 	SHA_DIGEST_LENGTH,
-	0,
+	EVP_MD_FLAG_PKEY_DIGEST,
 	init,
 	update,
 	final,
@@ -146,3 +148,4 @@ const EVP_MD *EVP_ecdsa(void)
 	return(&ecdsa_md);
 	}
 #endif
+#endif
diff --git a/crypto/evp/m_md2.c b/crypto/evp/m_md2.c
index 8eee6236ba25..5ce849f161dd 100644
--- a/crypto/evp/m_md2.c
+++ b/crypto/evp/m_md2.c
@@ -58,7 +58,6 @@
 
 #include <stdio.h>
 #include "cryptlib.h"
-#include "evp_locl.h"
 
 #ifndef OPENSSL_NO_MD2
 
diff --git a/crypto/evp/m_md4.c b/crypto/evp/m_md4.c
index 5cd2ab5adeef..6d47f61b2745 100644
--- a/crypto/evp/m_md4.c
+++ b/crypto/evp/m_md4.c
@@ -58,7 +58,6 @@
 
 #include <stdio.h>
 #include "cryptlib.h"
-#include "evp_locl.h"
 
 #ifndef OPENSSL_NO_MD4
 
@@ -70,6 +69,8 @@
 #include <openssl/rsa.h>
 #endif
 
+#include "evp_locl.h"
+
 static int init(EVP_MD_CTX *ctx)
 	{ return MD4_Init(ctx->md_data); }
 
diff --git a/crypto/evp/m_md5.c b/crypto/evp/m_md5.c
index 6455829671ef..9a8bae025803 100644
--- a/crypto/evp/m_md5.c
+++ b/crypto/evp/m_md5.c
@@ -62,13 +62,13 @@
 #ifndef OPENSSL_NO_MD5
 
 #include <openssl/evp.h>
-#include "evp_locl.h"
 #include <openssl/objects.h>
 #include <openssl/x509.h>
 #include <openssl/md5.h>
 #ifndef OPENSSL_NO_RSA
 #include <openssl/rsa.h>
 #endif
+#include "evp_locl.h"
 
 static int init(EVP_MD_CTX *ctx)
 	{ return MD5_Init(ctx->md_data); }
diff --git a/crypto/evp/m_mdc2.c b/crypto/evp/m_mdc2.c
index 9f9bcf06ed2d..3602bed316a4 100644
--- a/crypto/evp/m_mdc2.c
+++ b/crypto/evp/m_mdc2.c
@@ -58,7 +58,6 @@
 
 #include <stdio.h>
 #include "cryptlib.h"
-#include "evp_locl.h"
 
 #ifndef OPENSSL_NO_MDC2
 
@@ -66,7 +65,11 @@
 #include <openssl/objects.h>
 #include <openssl/x509.h>
 #include <openssl/mdc2.h>
+#ifndef OPENSSL_NO_RSA
 #include <openssl/rsa.h>
+#endif
+
+#include "evp_locl.h"
 
 static int init(EVP_MD_CTX *ctx)
 	{ return MDC2_Init(ctx->md_data); }
diff --git a/crypto/evp/m_ripemd.c b/crypto/evp/m_ripemd.c
index a1d60ee78d28..7bf4804cf888 100644
--- a/crypto/evp/m_ripemd.c
+++ b/crypto/evp/m_ripemd.c
@@ -68,6 +68,7 @@
 #ifndef OPENSSL_NO_RSA
 #include <openssl/rsa.h>
 #endif
+#include "evp_locl.h"
 
 static int init(EVP_MD_CTX *ctx)
 	{ return RIPEMD160_Init(ctx->md_data); }
diff --git a/crypto/evp/m_sha.c b/crypto/evp/m_sha.c
index 3f30dfc579ca..8769cdd42fda 100644
--- a/crypto/evp/m_sha.c
+++ b/crypto/evp/m_sha.c
@@ -58,7 +58,6 @@
 
 #include <stdio.h>
 #include "cryptlib.h"
-#include "evp_locl.h"
 
 #if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA0)
 
@@ -68,6 +67,7 @@
 #ifndef OPENSSL_NO_RSA
 #include <openssl/rsa.h>
 #endif
+#include "evp_locl.h"
 
 static int init(EVP_MD_CTX *ctx)
 	{ return SHA_Init(ctx->md_data); }
diff --git a/crypto/evp/m_sha1.c b/crypto/evp/m_sha1.c
index 471ec30be013..3cb11f1ebb1f 100644
--- a/crypto/evp/m_sha1.c
+++ b/crypto/evp/m_sha1.c
@@ -59,6 +59,8 @@
 #include <stdio.h>
 #include "cryptlib.h"
 
+#ifndef OPENSSL_FIPS
+
 #ifndef OPENSSL_NO_SHA
 
 #include <openssl/evp.h>
@@ -68,7 +70,6 @@
 #include <openssl/rsa.h>
 #endif
 
-#ifndef OPENSSL_FIPS
 
 static int init(EVP_MD_CTX *ctx)
 	{ return SHA1_Init(ctx->md_data); }
@@ -84,7 +85,7 @@ static const EVP_MD sha1_md=
 	NID_sha1,
 	NID_sha1WithRSAEncryption,
 	SHA_DIGEST_LENGTH,
-	0,
+	EVP_MD_FLAG_PKEY_METHOD_SIGNATURE|EVP_MD_FLAG_DIGALGID_ABSENT,
 	init,
 	update,
 	final,
@@ -99,6 +100,7 @@ const EVP_MD *EVP_sha1(void)
 	{
 	return(&sha1_md);
 	}
+#endif
 
 #ifndef OPENSSL_NO_SHA256
 static int init224(EVP_MD_CTX *ctx)
@@ -120,7 +122,7 @@ static const EVP_MD sha224_md=
 	NID_sha224,
 	NID_sha224WithRSAEncryption,
 	SHA224_DIGEST_LENGTH,
-	0,
+	EVP_MD_FLAG_PKEY_METHOD_SIGNATURE|EVP_MD_FLAG_DIGALGID_ABSENT,
 	init224,
 	update256,
 	final256,
@@ -139,7 +141,7 @@ static const EVP_MD sha256_md=
 	NID_sha256,
 	NID_sha256WithRSAEncryption,
 	SHA256_DIGEST_LENGTH,
-	0,
+	EVP_MD_FLAG_PKEY_METHOD_SIGNATURE|EVP_MD_FLAG_DIGALGID_ABSENT,
 	init256,
 	update256,
 	final256,
@@ -170,7 +172,7 @@ static const EVP_MD sha384_md=
 	NID_sha384,
 	NID_sha384WithRSAEncryption,
 	SHA384_DIGEST_LENGTH,
-	0,
+	EVP_MD_FLAG_PKEY_METHOD_SIGNATURE|EVP_MD_FLAG_DIGALGID_ABSENT,
 	init384,
 	update512,
 	final512,
@@ -189,7 +191,7 @@ static const EVP_MD sha512_md=
 	NID_sha512,
 	NID_sha512WithRSAEncryption,
 	SHA512_DIGEST_LENGTH,
-	0,
+	EVP_MD_FLAG_PKEY_METHOD_SIGNATURE|EVP_MD_FLAG_DIGALGID_ABSENT,
 	init512,
 	update512,
 	final512,
@@ -205,5 +207,3 @@ const EVP_MD *EVP_sha512(void)
 #endif	/* ifndef OPENSSL_NO_SHA512 */
 
 #endif
-
-#endif
diff --git a/crypto/evp/m_sigver.c b/crypto/evp/m_sigver.c
new file mode 100644
index 000000000000..7e2731f4a4ea
--- /dev/null
+++ b/crypto/evp/m_sigver.c
@@ -0,0 +1,200 @@
+/* m_sigver.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006,2007 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/evp.h>
+#include <openssl/objects.h>
+#include <openssl/x509.h>
+#include "evp_locl.h"
+
+static int do_sigver_init(EVP_MD_CTX *ctx, EVP_PKEY_CTX **pctx,
+			  const EVP_MD *type, ENGINE *e, EVP_PKEY *pkey,
+			  int ver)
+	{
+	if (ctx->pctx == NULL)
+		ctx->pctx = EVP_PKEY_CTX_new(pkey, e);
+	if (ctx->pctx == NULL)
+		return 0;
+
+	if (type == NULL)
+		{
+		int def_nid;
+		if (EVP_PKEY_get_default_digest_nid(pkey, &def_nid) > 0)
+			type = EVP_get_digestbynid(def_nid);
+		}
+
+	if (type == NULL)
+		{
+		EVPerr(EVP_F_DO_SIGVER_INIT, EVP_R_NO_DEFAULT_DIGEST);
+		return 0;
+		}
+
+	if (ver)
+		{
+		if (ctx->pctx->pmeth->verifyctx_init)
+			{
+			if (ctx->pctx->pmeth->verifyctx_init(ctx->pctx, ctx) <=0)
+				return 0;
+			ctx->pctx->operation = EVP_PKEY_OP_VERIFYCTX;
+			}
+		else if (EVP_PKEY_verify_init(ctx->pctx) <= 0)
+			return 0;
+		}
+	else
+		{
+		if (ctx->pctx->pmeth->signctx_init)
+			{
+			if (ctx->pctx->pmeth->signctx_init(ctx->pctx, ctx) <= 0)
+				return 0;
+			ctx->pctx->operation = EVP_PKEY_OP_SIGNCTX;
+			}
+		else if (EVP_PKEY_sign_init(ctx->pctx) <= 0)
+			return 0;
+		}
+	if (EVP_PKEY_CTX_set_signature_md(ctx->pctx, type) <= 0)
+		return 0;
+	if (pctx)
+		*pctx = ctx->pctx;
+	if (!EVP_DigestInit_ex(ctx, type, e))
+		return 0;
+	return 1;
+	}
+
+int EVP_DigestSignInit(EVP_MD_CTX *ctx, EVP_PKEY_CTX **pctx,
+			const EVP_MD *type, ENGINE *e, EVP_PKEY *pkey)
+	{
+	return do_sigver_init(ctx, pctx, type, e, pkey, 0);
+	}
+
+int EVP_DigestVerifyInit(EVP_MD_CTX *ctx, EVP_PKEY_CTX **pctx,
+			const EVP_MD *type, ENGINE *e, EVP_PKEY *pkey)
+	{
+	return do_sigver_init(ctx, pctx, type, e, pkey, 1);
+	}
+
+int EVP_DigestSignFinal(EVP_MD_CTX *ctx, unsigned char *sigret, size_t *siglen)
+	{
+	int sctx, r = 0;
+	if (ctx->pctx->pmeth->signctx)
+		sctx = 1;
+	else
+		sctx = 0;
+	if (sigret)
+		{
+		EVP_MD_CTX tmp_ctx;
+		unsigned char md[EVP_MAX_MD_SIZE];
+		unsigned int mdlen;
+		EVP_MD_CTX_init(&tmp_ctx);
+		if (!EVP_MD_CTX_copy_ex(&tmp_ctx,ctx))
+		     	return 0;
+		if (sctx)
+			r = tmp_ctx.pctx->pmeth->signctx(tmp_ctx.pctx,
+					sigret, siglen, &tmp_ctx);
+		else
+			r = EVP_DigestFinal_ex(&tmp_ctx,md,&mdlen);
+		EVP_MD_CTX_cleanup(&tmp_ctx);
+		if (sctx || !r)
+			return r;
+		if (EVP_PKEY_sign(ctx->pctx, sigret, siglen, md, mdlen) <= 0)
+			return 0;
+		}
+	else
+		{
+		if (sctx)
+			{
+			if (ctx->pctx->pmeth->signctx(ctx->pctx, sigret, siglen, ctx) <= 0)
+				return 0;
+			}
+		else
+			{
+			int s = EVP_MD_size(ctx->digest);
+			if (s < 0 || EVP_PKEY_sign(ctx->pctx, sigret, siglen, NULL, s) <= 0)
+				return 0;
+			}
+		}
+	return 1;
+	}
+
+int EVP_DigestVerifyFinal(EVP_MD_CTX *ctx, unsigned char *sig, size_t siglen)
+	{
+	EVP_MD_CTX tmp_ctx;
+	unsigned char md[EVP_MAX_MD_SIZE];
+	int r;
+	unsigned int mdlen;
+	int vctx;
+
+	if (ctx->pctx->pmeth->verifyctx)
+		vctx = 1;
+	else
+		vctx = 0;
+	EVP_MD_CTX_init(&tmp_ctx);
+	if (!EVP_MD_CTX_copy_ex(&tmp_ctx,ctx))
+		return -1;	
+	if (vctx)
+		{
+		r = tmp_ctx.pctx->pmeth->verifyctx(tmp_ctx.pctx,
+					sig, siglen, &tmp_ctx);
+		}
+	else
+		r = EVP_DigestFinal_ex(&tmp_ctx,md,&mdlen);
+	EVP_MD_CTX_cleanup(&tmp_ctx);
+	if (vctx || !r)
+		return r;
+	return EVP_PKEY_verify(ctx->pctx, sig, siglen, md, mdlen);
+	}
diff --git a/crypto/evp/m_wp.c b/crypto/evp/m_wp.c
new file mode 100644
index 000000000000..c51bc2d5d1ee
--- /dev/null
+++ b/crypto/evp/m_wp.c
@@ -0,0 +1,43 @@
+/* crypto/evp/m_wp.c */
+
+#include <stdio.h>
+#include "cryptlib.h"
+
+#ifndef OPENSSL_NO_WHIRLPOOL
+
+#include <openssl/evp.h>
+#include <openssl/objects.h>
+#include <openssl/x509.h>
+#include <openssl/whrlpool.h>
+#include "evp_locl.h"
+
+static int init(EVP_MD_CTX *ctx)
+	{ return WHIRLPOOL_Init(ctx->md_data); }
+
+static int update(EVP_MD_CTX *ctx,const void *data,size_t count)
+	{ return WHIRLPOOL_Update(ctx->md_data,data,count); }
+
+static int final(EVP_MD_CTX *ctx,unsigned char *md)
+	{ return WHIRLPOOL_Final(md,ctx->md_data); }
+
+static const EVP_MD whirlpool_md=
+	{
+	NID_whirlpool,
+	0,
+	WHIRLPOOL_DIGEST_LENGTH,
+	0,
+	init,
+	update,
+	final,
+	NULL,
+	NULL,
+	EVP_PKEY_NULL_method,
+	WHIRLPOOL_BBLOCK/8,
+	sizeof(EVP_MD *)+sizeof(WHIRLPOOL_CTX),
+	};
+
+const EVP_MD *EVP_whirlpool(void)
+	{
+	return(&whirlpool_md);
+	}
+#endif
diff --git a/crypto/evp/names.c b/crypto/evp/names.c
index 945879dc39a6..6311ad7cfb77 100644
--- a/crypto/evp/names.c
+++ b/crypto/evp/names.c
@@ -66,27 +66,28 @@ int EVP_add_cipher(const EVP_CIPHER *c)
 	{
 	int r;
 
-#ifdef OPENSSL_FIPS
+	if (c == NULL) return 0;
+
 	OPENSSL_init();
-#endif
 
 	r=OBJ_NAME_add(OBJ_nid2sn(c->nid),OBJ_NAME_TYPE_CIPHER_METH,(const char *)c);
 	if (r == 0) return(0);
+	check_defer(c->nid);
 	r=OBJ_NAME_add(OBJ_nid2ln(c->nid),OBJ_NAME_TYPE_CIPHER_METH,(const char *)c);
 	return(r);
 	}
 
+
 int EVP_add_digest(const EVP_MD *md)
 	{
 	int r;
 	const char *name;
-
-#ifdef OPENSSL_FIPS
 	OPENSSL_init();
-#endif
+
 	name=OBJ_nid2sn(md->type);
 	r=OBJ_NAME_add(name,OBJ_NAME_TYPE_MD_METH,(const char *)md);
 	if (r == 0) return(0);
+	check_defer(md->type);
 	r=OBJ_NAME_add(OBJ_nid2ln(md->type),OBJ_NAME_TYPE_MD_METH,(const char *)md);
 	if (r == 0) return(0);
 
@@ -95,6 +96,7 @@ int EVP_add_digest(const EVP_MD *md)
 		r=OBJ_NAME_add(OBJ_nid2sn(md->pkey_type),
 			OBJ_NAME_TYPE_MD_METH|OBJ_NAME_ALIAS,name);
 		if (r == 0) return(0);
+		check_defer(md->pkey_type);
 		r=OBJ_NAME_add(OBJ_nid2ln(md->pkey_type),
 			OBJ_NAME_TYPE_MD_METH|OBJ_NAME_ALIAS,name);
 		}
@@ -127,4 +129,78 @@ void EVP_cleanup(void)
 	OBJ_NAME_cleanup(-1);
 
 	EVP_PBE_cleanup();
+	if (obj_cleanup_defer == 2)
+		{
+		obj_cleanup_defer = 0;
+		OBJ_cleanup();
+		}
+	OBJ_sigid_free();
+	}
+
+struct doall_cipher
+	{
+	void *arg;
+	void (*fn)(const EVP_CIPHER *ciph,
+			const char *from, const char *to, void *arg);
+	};
+
+static void do_all_cipher_fn(const OBJ_NAME *nm, void *arg)
+	{
+	struct doall_cipher *dc = arg;
+	if (nm->alias)
+		dc->fn(NULL, nm->name, nm->data, dc->arg);
+	else
+		dc->fn((const EVP_CIPHER *)nm->data, nm->name, NULL, dc->arg);
+	}
+
+void EVP_CIPHER_do_all(void (*fn)(const EVP_CIPHER *ciph,
+		const char *from, const char *to, void *x), void *arg)
+	{
+	struct doall_cipher dc;
+	dc.fn = fn;
+	dc.arg = arg;
+	OBJ_NAME_do_all(OBJ_NAME_TYPE_CIPHER_METH, do_all_cipher_fn, &dc);
+	}
+
+void EVP_CIPHER_do_all_sorted(void (*fn)(const EVP_CIPHER *ciph,
+		const char *from, const char *to, void *x), void *arg)
+	{
+	struct doall_cipher dc;
+	dc.fn = fn;
+	dc.arg = arg;
+	OBJ_NAME_do_all_sorted(OBJ_NAME_TYPE_CIPHER_METH, do_all_cipher_fn,&dc);
+	}
+
+struct doall_md
+	{
+	void *arg;
+	void (*fn)(const EVP_MD *ciph,
+			const char *from, const char *to, void *arg);
+	};
+
+static void do_all_md_fn(const OBJ_NAME *nm, void *arg)
+	{
+	struct doall_md *dc = arg;
+	if (nm->alias)
+		dc->fn(NULL, nm->name, nm->data, dc->arg);
+	else
+		dc->fn((const EVP_MD *)nm->data, nm->name, NULL, dc->arg);
+	}
+
+void EVP_MD_do_all(void (*fn)(const EVP_MD *md,
+		const char *from, const char *to, void *x), void *arg)
+	{
+	struct doall_md dc;
+	dc.fn = fn;
+	dc.arg = arg;
+	OBJ_NAME_do_all(OBJ_NAME_TYPE_MD_METH, do_all_md_fn, &dc);
+	}
+
+void EVP_MD_do_all_sorted(void (*fn)(const EVP_MD *md,
+		const char *from, const char *to, void *x), void *arg)
+	{
+	struct doall_md dc;
+	dc.fn = fn;
+	dc.arg = arg;
+	OBJ_NAME_do_all_sorted(OBJ_NAME_TYPE_MD_METH, do_all_md_fn, &dc);
 	}
diff --git a/crypto/evp/p5_crpt.c b/crypto/evp/p5_crpt.c
index 2a265fdee277..294cc90d878c 100644
--- a/crypto/evp/p5_crpt.c
+++ b/crypto/evp/p5_crpt.c
@@ -62,42 +62,11 @@
 #include <openssl/x509.h>
 #include <openssl/evp.h>
 
-/* PKCS#5 v1.5 compatible PBE functions: see PKCS#5 v2.0 for more info.
+/* Doesn't do anything now: Builtin PBE algorithms in static table.
  */
 
 void PKCS5_PBE_add(void)
 {
-#ifndef OPENSSL_NO_DES
-#  ifndef OPENSSL_NO_MD5
-EVP_PBE_alg_add(NID_pbeWithMD5AndDES_CBC, EVP_des_cbc(), EVP_md5(),
-							 PKCS5_PBE_keyivgen);
-#  endif
-#  ifndef OPENSSL_NO_MD2
-EVP_PBE_alg_add(NID_pbeWithMD2AndDES_CBC, EVP_des_cbc(), EVP_md2(),
-							 PKCS5_PBE_keyivgen);
-#  endif
-#  ifndef OPENSSL_NO_SHA
-EVP_PBE_alg_add(NID_pbeWithSHA1AndDES_CBC, EVP_des_cbc(), EVP_sha1(),
-							 PKCS5_PBE_keyivgen);
-#  endif
-#endif
-#ifndef OPENSSL_NO_RC2
-#  ifndef OPENSSL_NO_MD5
-EVP_PBE_alg_add(NID_pbeWithMD5AndRC2_CBC, EVP_rc2_64_cbc(), EVP_md5(),
-							 PKCS5_PBE_keyivgen);
-#  endif
-#  ifndef OPENSSL_NO_MD2
-EVP_PBE_alg_add(NID_pbeWithMD2AndRC2_CBC, EVP_rc2_64_cbc(), EVP_md2(),
-							 PKCS5_PBE_keyivgen);
-#  endif
-#  ifndef OPENSSL_NO_SHA
-EVP_PBE_alg_add(NID_pbeWithSHA1AndRC2_CBC, EVP_rc2_64_cbc(), EVP_sha1(),
-							 PKCS5_PBE_keyivgen);
-#  endif
-#endif
-#ifndef OPENSSL_NO_HMAC
-EVP_PBE_alg_add(NID_pbes2, NULL, NULL, PKCS5_v2_PBE_keyivgen);
-#endif
 }
 
 int PKCS5_PBE_keyivgen(EVP_CIPHER_CTX *cctx, const char *pass, int passlen,
@@ -112,6 +81,9 @@ int PKCS5_PBE_keyivgen(EVP_CIPHER_CTX *cctx, const char *pass, int passlen,
 	int saltlen, iter;
 	unsigned char *salt;
 	const unsigned char *pbuf;
+	int mdsize;
+	int rv = 0;
+	EVP_MD_CTX_init(&ctx);
 
 	/* Extract useful info from parameter */
 	if (param == NULL || param->type != V_ASN1_SEQUENCE ||
@@ -134,26 +106,38 @@ int PKCS5_PBE_keyivgen(EVP_CIPHER_CTX *cctx, const char *pass, int passlen,
 	if(!pass) passlen = 0;
 	else if(passlen == -1) passlen = strlen(pass);
 
-	EVP_MD_CTX_init(&ctx);
-	EVP_DigestInit_ex(&ctx, md, NULL);
-	EVP_DigestUpdate(&ctx, pass, passlen);
-	EVP_DigestUpdate(&ctx, salt, saltlen);
+	if (!EVP_DigestInit_ex(&ctx, md, NULL))
+		goto err;
+	if (!EVP_DigestUpdate(&ctx, pass, passlen))
+		goto err;
+	if (!EVP_DigestUpdate(&ctx, salt, saltlen))
+		goto err;
 	PBEPARAM_free(pbe);
-	EVP_DigestFinal_ex(&ctx, md_tmp, NULL);
+	if (!EVP_DigestFinal_ex(&ctx, md_tmp, NULL))
+		goto err;
+	mdsize = EVP_MD_size(md);
+	if (mdsize < 0)
+	    return 0;
 	for (i = 1; i < iter; i++) {
-		EVP_DigestInit_ex(&ctx, md, NULL);
-		EVP_DigestUpdate(&ctx, md_tmp, EVP_MD_size(md));
-		EVP_DigestFinal_ex (&ctx, md_tmp, NULL);
+		if (!EVP_DigestInit_ex(&ctx, md, NULL))
+			goto err;
+		if (!EVP_DigestUpdate(&ctx, md_tmp, mdsize))
+			goto err;
+		if (!EVP_DigestFinal_ex (&ctx, md_tmp, NULL))
+			goto err;
 	}
-	EVP_MD_CTX_cleanup(&ctx);
 	OPENSSL_assert(EVP_CIPHER_key_length(cipher) <= (int)sizeof(md_tmp));
 	memcpy(key, md_tmp, EVP_CIPHER_key_length(cipher));
 	OPENSSL_assert(EVP_CIPHER_iv_length(cipher) <= 16);
 	memcpy(iv, md_tmp + (16 - EVP_CIPHER_iv_length(cipher)),
 						 EVP_CIPHER_iv_length(cipher));
-	EVP_CipherInit_ex(cctx, cipher, NULL, key, iv, en_de);
+	if (!EVP_CipherInit_ex(cctx, cipher, NULL, key, iv, en_de))
+		goto err;
 	OPENSSL_cleanse(md_tmp, EVP_MAX_MD_SIZE);
 	OPENSSL_cleanse(key, EVP_MAX_KEY_LENGTH);
 	OPENSSL_cleanse(iv, EVP_MAX_IV_LENGTH);
-	return 1;
+	rv = 1;
+	err:
+	EVP_MD_CTX_cleanup(&ctx);
+	return rv;
 }
diff --git a/crypto/evp/p5_crpt2.c b/crypto/evp/p5_crpt2.c
index 6bec77baf964..975d004df473 100644
--- a/crypto/evp/p5_crpt2.c
+++ b/crypto/evp/p5_crpt2.c
@@ -3,7 +3,7 @@
  * project 1999.
  */
 /* ====================================================================
- * Copyright (c) 1999 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -62,6 +62,7 @@
 #include <openssl/x509.h>
 #include <openssl/evp.h>
 #include <openssl/hmac.h>
+#include "evp_locl.h"
 
 /* set this to print out info about the keygen algorithm */
 /* #define DEBUG_PKCS5V2 */
@@ -71,28 +72,38 @@
 #endif
 
 /* This is an implementation of PKCS#5 v2.0 password based encryption key
- * derivation function PBKDF2 using the only currently defined function HMAC
- * with SHA1. Verified against test vectors posted by Peter Gutmann
+ * derivation function PBKDF2.
+ * SHA1 version verified against test vectors posted by Peter Gutmann
  * <pgut001@cs.auckland.ac.nz> to the PKCS-TNG <pkcs-tng@rsa.com> mailing list.
  */
 
-int PKCS5_PBKDF2_HMAC_SHA1(const char *pass, int passlen,
+int PKCS5_PBKDF2_HMAC(const char *pass, int passlen,
 			   const unsigned char *salt, int saltlen, int iter,
+			   const EVP_MD *digest,
 			   int keylen, unsigned char *out)
-{
-	unsigned char digtmp[SHA_DIGEST_LENGTH], *p, itmp[4];
-	int cplen, j, k, tkeylen;
+	{
+	unsigned char digtmp[EVP_MAX_MD_SIZE], *p, itmp[4];
+	int cplen, j, k, tkeylen, mdlen;
 	unsigned long i = 1;
 	HMAC_CTX hctx;
 
+	mdlen = EVP_MD_size(digest);
+	if (mdlen < 0)
+		return 0;
+
 	HMAC_CTX_init(&hctx);
 	p = out;
 	tkeylen = keylen;
-	if(!pass) passlen = 0;
-	else if(passlen == -1) passlen = strlen(pass);
-	while(tkeylen) {
-		if(tkeylen > SHA_DIGEST_LENGTH) cplen = SHA_DIGEST_LENGTH;
-		else cplen = tkeylen;
+	if(!pass)
+		passlen = 0;
+	else if(passlen == -1)
+		passlen = strlen(pass);
+	while(tkeylen)
+		{
+		if(tkeylen > mdlen)
+			cplen = mdlen;
+		else
+			cplen = tkeylen;
 		/* We are unlikely to ever use more than 256 blocks (5120 bits!)
 		 * but just in case...
 		 */
@@ -100,20 +111,26 @@ int PKCS5_PBKDF2_HMAC_SHA1(const char *pass, int passlen,
 		itmp[1] = (unsigned char)((i >> 16) & 0xff);
 		itmp[2] = (unsigned char)((i >> 8) & 0xff);
 		itmp[3] = (unsigned char)(i & 0xff);
-		HMAC_Init_ex(&hctx, pass, passlen, EVP_sha1(), NULL);
-		HMAC_Update(&hctx, salt, saltlen);
-		HMAC_Update(&hctx, itmp, 4);
-		HMAC_Final(&hctx, digtmp, NULL);
+		if (!HMAC_Init_ex(&hctx, pass, passlen, digest, NULL)
+			|| !HMAC_Update(&hctx, salt, saltlen)
+			|| !HMAC_Update(&hctx, itmp, 4)
+			|| !HMAC_Final(&hctx, digtmp, NULL))
+			{
+			HMAC_CTX_cleanup(&hctx);
+			return 0;
+			}
 		memcpy(p, digtmp, cplen);
-		for(j = 1; j < iter; j++) {
-			HMAC(EVP_sha1(), pass, passlen,
-				 digtmp, SHA_DIGEST_LENGTH, digtmp, NULL);
-			for(k = 0; k < cplen; k++) p[k] ^= digtmp[k];
-		}
+		for(j = 1; j < iter; j++)
+			{
+			HMAC(digest, pass, passlen,
+				 digtmp, mdlen, digtmp, NULL);
+			for(k = 0; k < cplen; k++)
+				p[k] ^= digtmp[k];
+			}
 		tkeylen-= cplen;
 		i++;
 		p+= cplen;
-	}
+		}
 	HMAC_CTX_cleanup(&hctx);
 #ifdef DEBUG_PKCS5V2
 	fprintf(stderr, "Password:\n");
@@ -125,7 +142,15 @@ int PKCS5_PBKDF2_HMAC_SHA1(const char *pass, int passlen,
 	h__dump (out, keylen);
 #endif
 	return 1;
-}
+	}
+
+int PKCS5_PBKDF2_HMAC_SHA1(const char *pass, int passlen,
+			   const unsigned char *salt, int saltlen, int iter,
+			   int keylen, unsigned char *out)
+	{
+	return PKCS5_PBKDF2_HMAC(pass, passlen, salt, saltlen, iter, EVP_sha1(),
+					keylen, out);
+	}
 
 #ifdef DO_TEST
 main()
@@ -148,25 +173,24 @@ int PKCS5_v2_PBE_keyivgen(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
                          ASN1_TYPE *param, const EVP_CIPHER *c, const EVP_MD *md,
                          int en_de)
 {
-	unsigned char *salt, key[EVP_MAX_KEY_LENGTH];
 	const unsigned char *pbuf;
-	int saltlen, iter, plen;
-	unsigned int keylen;
+	int plen;
 	PBE2PARAM *pbe2 = NULL;
 	const EVP_CIPHER *cipher;
-	PBKDF2PARAM *kdf = NULL;
+
+	int rv = 0;
 
 	if (param == NULL || param->type != V_ASN1_SEQUENCE ||
 	    param->value.sequence == NULL) {
 		EVPerr(EVP_F_PKCS5_V2_PBE_KEYIVGEN,EVP_R_DECODE_ERROR);
-		return 0;
+		goto err;
 	}
 
 	pbuf = param->value.sequence->data;
 	plen = param->value.sequence->length;
 	if(!(pbe2 = d2i_PBE2PARAM(NULL, &pbuf, plen))) {
 		EVPerr(EVP_F_PKCS5_V2_PBE_KEYIVGEN,EVP_R_DECODE_ERROR);
-		return 0;
+		goto err;
 	}
 
 	/* See if we recognise the key derivation function */
@@ -180,8 +204,7 @@ int PKCS5_v2_PBE_keyivgen(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
 	/* lets see if we recognise the encryption algorithm.
 	 */
 
-	cipher = EVP_get_cipherbyname(
-			OBJ_nid2sn(OBJ_obj2nid(pbe2->encryption->algorithm)));
+	cipher = EVP_get_cipherbyobj(pbe2->encryption->algorithm);
 
 	if(!cipher) {
 		EVPerr(EVP_F_PKCS5_V2_PBE_KEYIVGEN,
@@ -190,49 +213,87 @@ int PKCS5_v2_PBE_keyivgen(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
 	}
 
 	/* Fixup cipher based on AlgorithmIdentifier */
-	EVP_CipherInit_ex(ctx, cipher, NULL, NULL, NULL, en_de);
+	if (!EVP_CipherInit_ex(ctx, cipher, NULL, NULL, NULL, en_de))
+		goto err;
 	if(EVP_CIPHER_asn1_to_param(ctx, pbe2->encryption->parameter) < 0) {
 		EVPerr(EVP_F_PKCS5_V2_PBE_KEYIVGEN,
 					EVP_R_CIPHER_PARAMETER_ERROR);
 		goto err;
 	}
+	rv = PKCS5_v2_PBKDF2_keyivgen(ctx, pass, passlen,
+					pbe2->keyfunc->parameter, c, md, en_de);
+	err:
+	PBE2PARAM_free(pbe2);
+	return rv;
+}
+
+int PKCS5_v2_PBKDF2_keyivgen(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
+                         ASN1_TYPE *param,
+			 const EVP_CIPHER *c, const EVP_MD *md, int en_de)
+{
+	unsigned char *salt, key[EVP_MAX_KEY_LENGTH];
+	const unsigned char *pbuf;
+	int saltlen, iter, plen;
+	int rv = 0;
+	unsigned int keylen = 0;
+	int prf_nid, hmac_md_nid;
+	PBKDF2PARAM *kdf = NULL;
+	const EVP_MD *prfmd;
+
+	if (EVP_CIPHER_CTX_cipher(ctx) == NULL)
+		{
+		EVPerr(EVP_F_PKCS5_V2_PBKDF2_KEYIVGEN,EVP_R_NO_CIPHER_SET);
+		goto err;
+		}
 	keylen = EVP_CIPHER_CTX_key_length(ctx);
 	OPENSSL_assert(keylen <= sizeof key);
 
-	/* Now decode key derivation function */
+	/* Decode parameter */
 
-	if(!pbe2->keyfunc->parameter ||
-		 (pbe2->keyfunc->parameter->type != V_ASN1_SEQUENCE))
+	if(!param || (param->type != V_ASN1_SEQUENCE))
 		{
-		EVPerr(EVP_F_PKCS5_V2_PBE_KEYIVGEN,EVP_R_DECODE_ERROR);
+		EVPerr(EVP_F_PKCS5_V2_PBKDF2_KEYIVGEN,EVP_R_DECODE_ERROR);
 		goto err;
 		}
 
-	pbuf = pbe2->keyfunc->parameter->value.sequence->data;
-	plen = pbe2->keyfunc->parameter->value.sequence->length;
+	pbuf = param->value.sequence->data;
+	plen = param->value.sequence->length;
+
 	if(!(kdf = d2i_PBKDF2PARAM(NULL, &pbuf, plen)) ) {
-		EVPerr(EVP_F_PKCS5_V2_PBE_KEYIVGEN,EVP_R_DECODE_ERROR);
+		EVPerr(EVP_F_PKCS5_V2_PBKDF2_KEYIVGEN,EVP_R_DECODE_ERROR);
 		goto err;
 	}
 
-	PBE2PARAM_free(pbe2);
-	pbe2 = NULL;
+	keylen = EVP_CIPHER_CTX_key_length(ctx);
 
 	/* Now check the parameters of the kdf */
 
 	if(kdf->keylength && (ASN1_INTEGER_get(kdf->keylength) != (int)keylen)){
-		EVPerr(EVP_F_PKCS5_V2_PBE_KEYIVGEN,
+		EVPerr(EVP_F_PKCS5_V2_PBKDF2_KEYIVGEN,
 						EVP_R_UNSUPPORTED_KEYLENGTH);
 		goto err;
 	}
 
-	if(kdf->prf && (OBJ_obj2nid(kdf->prf->algorithm) != NID_hmacWithSHA1)) {
-		EVPerr(EVP_F_PKCS5_V2_PBE_KEYIVGEN, EVP_R_UNSUPPORTED_PRF);
+	if (kdf->prf)
+		prf_nid = OBJ_obj2nid(kdf->prf->algorithm);
+	else
+		prf_nid = NID_hmacWithSHA1;
+
+	if (!EVP_PBE_find(EVP_PBE_TYPE_PRF, prf_nid, NULL, &hmac_md_nid, 0))
+		{
+		EVPerr(EVP_F_PKCS5_V2_PBKDF2_KEYIVGEN, EVP_R_UNSUPPORTED_PRF);
 		goto err;
-	}
+		}
+
+	prfmd = EVP_get_digestbynid(hmac_md_nid);
+	if (prfmd == NULL)
+		{
+		EVPerr(EVP_F_PKCS5_V2_PBKDF2_KEYIVGEN, EVP_R_UNSUPPORTED_PRF);
+		goto err;
+		}
 
 	if(kdf->salt->type != V_ASN1_OCTET_STRING) {
-		EVPerr(EVP_F_PKCS5_V2_PBE_KEYIVGEN,
+		EVPerr(EVP_F_PKCS5_V2_PBKDF2_KEYIVGEN,
 						EVP_R_UNSUPPORTED_SALT_TYPE);
 		goto err;
 	}
@@ -241,16 +302,14 @@ int PKCS5_v2_PBE_keyivgen(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
 	salt = kdf->salt->value.octet_string->data;
 	saltlen = kdf->salt->value.octet_string->length;
 	iter = ASN1_INTEGER_get(kdf->iter);
-	PKCS5_PBKDF2_HMAC_SHA1(pass, passlen, salt, saltlen, iter, keylen, key);
-	EVP_CipherInit_ex(ctx, NULL, NULL, key, NULL, en_de);
-	OPENSSL_cleanse(key, keylen);
-	PBKDF2PARAM_free(kdf);
-	return 1;
-
+	if(!PKCS5_PBKDF2_HMAC(pass, passlen, salt, saltlen, iter, prfmd,
+						   keylen, key))
+		goto err;
+	rv = EVP_CipherInit_ex(ctx, NULL, NULL, key, NULL, en_de);
 	err:
-	PBE2PARAM_free(pbe2);
+	OPENSSL_cleanse(key, keylen);
 	PBKDF2PARAM_free(kdf);
-	return 0;
+	return rv;
 }
 
 #ifdef DEBUG_PKCS5V2
diff --git a/crypto/evp/p_dec.c b/crypto/evp/p_dec.c
index f64901f65349..4201dcbad99e 100644
--- a/crypto/evp/p_dec.c
+++ b/crypto/evp/p_dec.c
@@ -66,7 +66,7 @@
 #include <openssl/objects.h>
 #include <openssl/x509.h>
 
-int EVP_PKEY_decrypt(unsigned char *key, const unsigned char *ek, int ekl,
+int EVP_PKEY_decrypt_old(unsigned char *key, const unsigned char *ek, int ekl,
 	     EVP_PKEY *priv)
 	{
 	int ret= -1;
@@ -75,7 +75,7 @@ int EVP_PKEY_decrypt(unsigned char *key, const unsigned char *ek, int ekl,
 	if (priv->type != EVP_PKEY_RSA)
 		{
 #endif
-		EVPerr(EVP_F_EVP_PKEY_DECRYPT,EVP_R_PUBLIC_KEY_NOT_RSA);
+		EVPerr(EVP_F_EVP_PKEY_DECRYPT_OLD,EVP_R_PUBLIC_KEY_NOT_RSA);
 #ifndef OPENSSL_NO_RSA
 		goto err;
                 }
diff --git a/crypto/evp/p_enc.c b/crypto/evp/p_enc.c
index c2dfdc52adc9..b5a3a84c41d3 100644
--- a/crypto/evp/p_enc.c
+++ b/crypto/evp/p_enc.c
@@ -66,7 +66,7 @@
 #include <openssl/objects.h>
 #include <openssl/x509.h>
 
-int EVP_PKEY_encrypt(unsigned char *ek, const unsigned char *key, int key_len,
+int EVP_PKEY_encrypt_old(unsigned char *ek, const unsigned char *key, int key_len,
 	     EVP_PKEY *pubk)
 	{
 	int ret=0;
@@ -75,7 +75,7 @@ int EVP_PKEY_encrypt(unsigned char *ek, const unsigned char *key, int key_len,
 	if (pubk->type != EVP_PKEY_RSA)
 		{
 #endif
-		EVPerr(EVP_F_EVP_PKEY_ENCRYPT,EVP_R_PUBLIC_KEY_NOT_RSA);
+		EVPerr(EVP_F_EVP_PKEY_ENCRYPT_OLD,EVP_R_PUBLIC_KEY_NOT_RSA);
 #ifndef OPENSSL_NO_RSA
 		goto err;
 		}
diff --git a/crypto/evp/p_lib.c b/crypto/evp/p_lib.c
index 22155ecf62c4..e26ccd0d086a 100644
--- a/crypto/evp/p_lib.c
+++ b/crypto/evp/p_lib.c
@@ -74,66 +74,26 @@
 #include <openssl/dh.h>
 #endif
 
+#ifndef OPENSSL_NO_ENGINE
+#include <openssl/engine.h>
+#endif
+
+#include "asn1_locl.h"
+
 static void EVP_PKEY_free_it(EVP_PKEY *x);
 
 int EVP_PKEY_bits(EVP_PKEY *pkey)
 	{
-	if (0)
-		return 0;
-#ifndef OPENSSL_NO_RSA
-	else if (pkey->type == EVP_PKEY_RSA)
-		return(BN_num_bits(pkey->pkey.rsa->n));
-#endif
-#ifndef OPENSSL_NO_DSA
-	else if (pkey->type == EVP_PKEY_DSA)
-		return(BN_num_bits(pkey->pkey.dsa->p));
-#endif
-#ifndef OPENSSL_NO_EC
-	else if (pkey->type == EVP_PKEY_EC)
-		{
-		BIGNUM *order = BN_new();
-		const EC_GROUP *group;
-		int ret;
-
-		if (!order)
-			{
-			ERR_clear_error();
-			return 0;
-			}
-		group = EC_KEY_get0_group(pkey->pkey.ec);
-		if (!EC_GROUP_get_order(group, order, NULL))
-			{
-			ERR_clear_error();
-			return 0;
-			}
-
-		ret = BN_num_bits(order);
-		BN_free(order);
-		return ret;
-		}
-#endif
-	return(0);
+	if (pkey && pkey->ameth && pkey->ameth->pkey_bits)
+		return pkey->ameth->pkey_bits(pkey);
+	return 0;
 	}
 
 int EVP_PKEY_size(EVP_PKEY *pkey)
 	{
-	if (pkey == NULL)
-		return(0);
-#ifndef OPENSSL_NO_RSA
-	if (pkey->type == EVP_PKEY_RSA)
-		return(RSA_size(pkey->pkey.rsa));
-	else
-#endif
-#ifndef OPENSSL_NO_DSA
-		if (pkey->type == EVP_PKEY_DSA)
-		return(DSA_size(pkey->pkey.dsa));
-#endif
-#ifndef OPENSSL_NO_ECDSA
-		if (pkey->type == EVP_PKEY_EC)
-		return(ECDSA_size(pkey->pkey.ec));
-#endif
-
-	return(0);
+	if (pkey && pkey->ameth && pkey->ameth->pkey_size)
+		return pkey->ameth->pkey_size(pkey);
+	return 0;
 	}
 
 int EVP_PKEY_save_parameters(EVP_PKEY *pkey, int mode)
@@ -174,88 +134,26 @@ int EVP_PKEY_copy_parameters(EVP_PKEY *to, const EVP_PKEY *from)
 		EVPerr(EVP_F_EVP_PKEY_COPY_PARAMETERS,EVP_R_MISSING_PARAMETERS);
 		goto err;
 		}
-#ifndef OPENSSL_NO_DSA
-	if (to->type == EVP_PKEY_DSA)
-		{
-		BIGNUM *a;
-
-		if ((a=BN_dup(from->pkey.dsa->p)) == NULL) goto err;
-		if (to->pkey.dsa->p != NULL) BN_free(to->pkey.dsa->p);
-		to->pkey.dsa->p=a;
-
-		if ((a=BN_dup(from->pkey.dsa->q)) == NULL) goto err;
-		if (to->pkey.dsa->q != NULL) BN_free(to->pkey.dsa->q);
-		to->pkey.dsa->q=a;
-
-		if ((a=BN_dup(from->pkey.dsa->g)) == NULL) goto err;
-		if (to->pkey.dsa->g != NULL) BN_free(to->pkey.dsa->g);
-		to->pkey.dsa->g=a;
-		}
-#endif
-#ifndef OPENSSL_NO_EC
-	if (to->type == EVP_PKEY_EC)
-		{
-		EC_GROUP *group = EC_GROUP_dup(EC_KEY_get0_group(from->pkey.ec));
-		if (group == NULL)
-			goto err;
-		if (EC_KEY_set_group(to->pkey.ec, group) == 0)
-			goto err;
-		EC_GROUP_free(group);
-		}
-#endif
-	return(1);
+	if (from->ameth && from->ameth->param_copy)
+		return from->ameth->param_copy(to, from);
 err:
-	return(0);
+	return 0;
 	}
 
 int EVP_PKEY_missing_parameters(const EVP_PKEY *pkey)
 	{
-#ifndef OPENSSL_NO_DSA
-	if (pkey->type == EVP_PKEY_DSA)
-		{
-		DSA *dsa;
-
-		dsa=pkey->pkey.dsa;
-		if ((dsa->p == NULL) || (dsa->q == NULL) || (dsa->g == NULL))
-			return(1);
-		}
-#endif
-#ifndef OPENSSL_NO_EC
-	if (pkey->type == EVP_PKEY_EC)
-		{
-		if (EC_KEY_get0_group(pkey->pkey.ec) == NULL)
-			return(1);
-		}
-#endif
-
-	return(0);
+	if (pkey->ameth && pkey->ameth->param_missing)
+		return pkey->ameth->param_missing(pkey);
+	return 0;
 	}
 
 int EVP_PKEY_cmp_parameters(const EVP_PKEY *a, const EVP_PKEY *b)
 	{
-#ifndef OPENSSL_NO_DSA
-	if ((a->type == EVP_PKEY_DSA) && (b->type == EVP_PKEY_DSA))
-		{
-		if (	BN_cmp(a->pkey.dsa->p,b->pkey.dsa->p) ||
-			BN_cmp(a->pkey.dsa->q,b->pkey.dsa->q) ||
-			BN_cmp(a->pkey.dsa->g,b->pkey.dsa->g))
-			return(0);
-		else
-			return(1);
-		}
-#endif
-#ifndef OPENSSL_NO_EC
-	if (a->type == EVP_PKEY_EC && b->type == EVP_PKEY_EC)
-		{
-		const EC_GROUP *group_a = EC_KEY_get0_group(a->pkey.ec),
-		               *group_b = EC_KEY_get0_group(b->pkey.ec);
-		if (EC_GROUP_cmp(group_a, group_b, NULL))
-			return 0;
-		else
-			return 1;
-		}
-#endif
-	return(-1);
+	if (a->type != b->type)
+		return -1;
+	if (a->ameth && a->ameth->param_cmp)
+		return a->ameth->param_cmp(a, b);
+	return -2;
 	}
 
 int EVP_PKEY_cmp(const EVP_PKEY *a, const EVP_PKEY *b)
@@ -263,51 +161,22 @@ int EVP_PKEY_cmp(const EVP_PKEY *a, const EVP_PKEY *b)
 	if (a->type != b->type)
 		return -1;
 
-	if (EVP_PKEY_cmp_parameters(a, b) == 0)
-		return 0;
-
-	switch (a->type)
+	if (a->ameth)
 		{
-#ifndef OPENSSL_NO_RSA
-	case EVP_PKEY_RSA:
-		if (BN_cmp(b->pkey.rsa->n,a->pkey.rsa->n) != 0
-			|| BN_cmp(b->pkey.rsa->e,a->pkey.rsa->e) != 0)
-			return 0;
-		break;
-#endif
-#ifndef OPENSSL_NO_DSA
-	case EVP_PKEY_DSA:
-		if (BN_cmp(b->pkey.dsa->pub_key,a->pkey.dsa->pub_key) != 0)
-			return 0;
-		break;
-#endif
-#ifndef OPENSSL_NO_EC
-	case EVP_PKEY_EC:
-		{
-		int  r;
-		const EC_GROUP *group = EC_KEY_get0_group(b->pkey.ec);
-		const EC_POINT *pa = EC_KEY_get0_public_key(a->pkey.ec),
-		               *pb = EC_KEY_get0_public_key(b->pkey.ec);
-		r = EC_POINT_cmp(group, pa, pb, NULL);
-		if (r != 0)
+		int ret;
+		/* Compare parameters if the algorithm has them */
+		if (a->ameth->param_cmp)
 			{
-			if (r == 1)
-				return 0;
-			else
-				return -2;
+			ret = a->ameth->param_cmp(a, b);
+			if (ret <= 0)
+				return ret;
 			}
-		}
- 		break;
-#endif
-#ifndef OPENSSL_NO_DH
-	case EVP_PKEY_DH:
-		return -2;
-#endif
-	default:
-		return -2;
+
+		if (a->ameth->pub_cmp)
+			return a->ameth->pub_cmp(a, b);
 		}
 
-	return 1;
+	return -2;
 	}
 
 EVP_PKEY *EVP_PKEY_new(void)
@@ -321,22 +190,87 @@ EVP_PKEY *EVP_PKEY_new(void)
 		return(NULL);
 		}
 	ret->type=EVP_PKEY_NONE;
+	ret->save_type=EVP_PKEY_NONE;
 	ret->references=1;
+	ret->ameth=NULL;
+	ret->engine=NULL;
 	ret->pkey.ptr=NULL;
 	ret->attributes=NULL;
 	ret->save_parameters=1;
 	return(ret);
 	}
 
-int EVP_PKEY_assign(EVP_PKEY *pkey, int type, char *key)
+/* Setup a public key ASN1 method and ENGINE from a NID or a string.
+ * If pkey is NULL just return 1 or 0 if the algorithm exists.
+ */
+
+static int pkey_set_type(EVP_PKEY *pkey, int type, const char *str, int len)
+	{
+	const EVP_PKEY_ASN1_METHOD *ameth;
+	ENGINE *e = NULL;
+	if (pkey)
+		{
+		if (pkey->pkey.ptr)
+			EVP_PKEY_free_it(pkey);
+		/* If key type matches and a method exists then this
+		 * lookup has succeeded once so just indicate success.
+		 */
+		if ((type == pkey->save_type) && pkey->ameth)
+			return 1;
+#ifndef OPENSSL_NO_ENGINE
+		/* If we have an ENGINE release it */
+		if (pkey->engine)
+			{
+			ENGINE_finish(pkey->engine);
+			pkey->engine = NULL;
+			}
+#endif
+		}
+	if (str)
+		ameth = EVP_PKEY_asn1_find_str(&e, str, len);
+	else
+		ameth = EVP_PKEY_asn1_find(&e, type);
+#ifndef OPENSSL_NO_ENGINE
+	if (!pkey && e)
+		ENGINE_finish(e);
+#endif
+	if (!ameth)
+		{
+		EVPerr(EVP_F_PKEY_SET_TYPE, EVP_R_UNSUPPORTED_ALGORITHM);
+		return 0;
+		}
+	if (pkey)
+		{
+		pkey->ameth = ameth;
+		pkey->engine = e;
+
+		pkey->type = pkey->ameth->pkey_id;
+		pkey->save_type=type;
+		}
+	return 1;
+	}
+
+int EVP_PKEY_set_type(EVP_PKEY *pkey, int type)
+	{
+	return pkey_set_type(pkey, type, NULL, -1);
+	}
+
+int EVP_PKEY_set_type_str(EVP_PKEY *pkey, const char *str, int len)
 	{
-	if (pkey == NULL) return(0);
-	if (pkey->pkey.ptr != NULL)
-		EVP_PKEY_free_it(pkey);
-	pkey->type=EVP_PKEY_type(type);
-	pkey->save_type=type;
+	return pkey_set_type(pkey, EVP_PKEY_NONE, str, len);
+	}
+
+int EVP_PKEY_assign(EVP_PKEY *pkey, int type, void *key)
+	{
+	if (!EVP_PKEY_set_type(pkey, type))
+		return 0;
 	pkey->pkey.ptr=key;
-	return(key != NULL);
+	return (key != NULL);
+	}
+
+void *EVP_PKEY_get0(EVP_PKEY *pkey)
+	{
+	return pkey->pkey.ptr;
 	}
 
 #ifndef OPENSSL_NO_RSA
@@ -425,24 +359,29 @@ DH *EVP_PKEY_get1_DH(EVP_PKEY *pkey)
 
 int EVP_PKEY_type(int type)
 	{
-	switch (type)
-		{
-	case EVP_PKEY_RSA:
-	case EVP_PKEY_RSA2:
-		return(EVP_PKEY_RSA);
-	case EVP_PKEY_DSA:
-	case EVP_PKEY_DSA1:
-	case EVP_PKEY_DSA2:
-	case EVP_PKEY_DSA3:
-	case EVP_PKEY_DSA4:
-		return(EVP_PKEY_DSA);
-	case EVP_PKEY_DH:
-		return(EVP_PKEY_DH);
-	case EVP_PKEY_EC:
-		return(EVP_PKEY_EC);
-	default:
-		return(NID_undef);
-		}
+	int ret;
+	const EVP_PKEY_ASN1_METHOD *ameth;
+	ENGINE *e;
+	ameth = EVP_PKEY_asn1_find(&e, type);
+	if (ameth)
+		ret = ameth->pkey_id;
+	else
+		ret = NID_undef;
+#ifndef OPENSSL_NO_ENGINE
+	if (e)
+		ENGINE_finish(e);
+#endif
+	return ret;
+	}
+
+int EVP_PKEY_id(const EVP_PKEY *pkey)
+	{
+	return pkey->type;
+	}
+
+int EVP_PKEY_base_id(const EVP_PKEY *pkey)
+	{
+	return EVP_PKEY_type(pkey->type);
 	}
 
 void EVP_PKEY_free(EVP_PKEY *x)
@@ -471,32 +410,60 @@ void EVP_PKEY_free(EVP_PKEY *x)
 
 static void EVP_PKEY_free_it(EVP_PKEY *x)
 	{
-	switch (x->type)
+	if (x->ameth && x->ameth->pkey_free)
 		{
-#ifndef OPENSSL_NO_RSA
-	case EVP_PKEY_RSA:
-	case EVP_PKEY_RSA2:
-		RSA_free(x->pkey.rsa);
-		break;
-#endif
-#ifndef OPENSSL_NO_DSA
-	case EVP_PKEY_DSA:
-	case EVP_PKEY_DSA2:
-	case EVP_PKEY_DSA3:
-	case EVP_PKEY_DSA4:
-		DSA_free(x->pkey.dsa);
-		break;
-#endif
-#ifndef OPENSSL_NO_EC
-	case EVP_PKEY_EC:
-		EC_KEY_free(x->pkey.ec);
-		break;
-#endif
-#ifndef OPENSSL_NO_DH
-	case EVP_PKEY_DH:
-		DH_free(x->pkey.dh);
-		break;
-#endif
+		x->ameth->pkey_free(x);
+		x->pkey.ptr = NULL;
+		}
+#ifndef OPENSSL_NO_ENGINE
+	if (x->engine)
+		{
+		ENGINE_finish(x->engine);
+		x->engine = NULL;
 		}
+#endif
+	}
+
+static int unsup_alg(BIO *out, const EVP_PKEY *pkey, int indent,
+				const char *kstr)
+	{
+	BIO_indent(out, indent, 128);
+	BIO_printf(out, "%s algorithm \"%s\" unsupported\n",
+						kstr, OBJ_nid2ln(pkey->type));
+	return 1;
+	}
+
+int EVP_PKEY_print_public(BIO *out, const EVP_PKEY *pkey,
+				int indent, ASN1_PCTX *pctx)
+	{
+	if (pkey->ameth && pkey->ameth->pub_print)
+		return pkey->ameth->pub_print(out, pkey, indent, pctx);
+	
+	return unsup_alg(out, pkey, indent, "Public Key");
+	}
+
+int EVP_PKEY_print_private(BIO *out, const EVP_PKEY *pkey,
+				int indent, ASN1_PCTX *pctx)
+	{
+	if (pkey->ameth && pkey->ameth->priv_print)
+		return pkey->ameth->priv_print(out, pkey, indent, pctx);
+	
+	return unsup_alg(out, pkey, indent, "Private Key");
+	}
+
+int EVP_PKEY_print_params(BIO *out, const EVP_PKEY *pkey,
+				int indent, ASN1_PCTX *pctx)
+	{
+	if (pkey->ameth && pkey->ameth->param_print)
+		return pkey->ameth->param_print(out, pkey, indent, pctx);
+	return unsup_alg(out, pkey, indent, "Parameters");
+	}
+
+int EVP_PKEY_get_default_digest_nid(EVP_PKEY *pkey, int *pnid)
+	{
+	if (!pkey->ameth || !pkey->ameth->pkey_ctrl)
+		return -2;
+	return pkey->ameth->pkey_ctrl(pkey, ASN1_PKEY_CTRL_DEFAULT_MD_NID,
+						0, pnid);
 	}
 
diff --git a/crypto/evp/p_open.c b/crypto/evp/p_open.c
index 9935206d0f3d..c748fbea8772 100644
--- a/crypto/evp/p_open.c
+++ b/crypto/evp/p_open.c
@@ -95,7 +95,7 @@ int EVP_OpenInit(EVP_CIPHER_CTX *ctx, const EVP_CIPHER *type,
 		goto err;
 		}
 
-	i=EVP_PKEY_decrypt(key,ek,ekl,priv);
+	i=EVP_PKEY_decrypt_old(key,ek,ekl,priv);
 	if ((i <= 0) || !EVP_CIPHER_CTX_set_key_length(ctx, i))
 		{
 		/* ERROR */
@@ -115,7 +115,8 @@ int EVP_OpenFinal(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl)
 	int i;
 
 	i=EVP_DecryptFinal_ex(ctx,out,outl);
-	EVP_DecryptInit_ex(ctx,NULL,NULL,NULL,NULL);
+	if (i)
+		i = EVP_DecryptInit_ex(ctx,NULL,NULL,NULL,NULL);
 	return(i);
 	}
 #else /* !OPENSSL_NO_RSA */
diff --git a/crypto/evp/p_seal.c b/crypto/evp/p_seal.c
index 8cc8fcb0bd05..e5919b0fbf98 100644
--- a/crypto/evp/p_seal.c
+++ b/crypto/evp/p_seal.c
@@ -87,7 +87,7 @@ int EVP_SealInit(EVP_CIPHER_CTX *ctx, const EVP_CIPHER *type, unsigned char **ek
 
 	for (i=0; i<npubk; i++)
 		{
-		ekl[i]=EVP_PKEY_encrypt(ek[i],key,EVP_CIPHER_CTX_key_length(ctx),
+		ekl[i]=EVP_PKEY_encrypt_old(ek[i],key,EVP_CIPHER_CTX_key_length(ctx),
 			pubk[i]);
 		if (ekl[i] <= 0) return(-1);
 		}
@@ -110,6 +110,7 @@ int EVP_SealFinal(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl)
 	{
 	int i;
 	i = EVP_EncryptFinal_ex(ctx,out,outl);
-	EVP_EncryptInit_ex(ctx,NULL,NULL,NULL,NULL);
+	if (i) 
+		i = EVP_EncryptInit_ex(ctx,NULL,NULL,NULL,NULL);
 	return i;
 	}
diff --git a/crypto/evp/p_sign.c b/crypto/evp/p_sign.c
index 782d14055a43..dfa48c157cf5 100644
--- a/crypto/evp/p_sign.c
+++ b/crypto/evp/p_sign.c
@@ -80,10 +80,38 @@ int EVP_SignFinal(EVP_MD_CTX *ctx, unsigned char *sigret, unsigned int *siglen,
 	{
 	unsigned char m[EVP_MAX_MD_SIZE];
 	unsigned int m_len;
-	int i,ok=0,v;
+	int i=0,ok=0,v;
 	EVP_MD_CTX tmp_ctx;
+	EVP_PKEY_CTX *pkctx = NULL;
 
 	*siglen=0;
+	EVP_MD_CTX_init(&tmp_ctx);
+	if (!EVP_MD_CTX_copy_ex(&tmp_ctx,ctx))
+		goto err;  
+	if (!EVP_DigestFinal_ex(&tmp_ctx,&(m[0]),&m_len))
+		goto err;
+	EVP_MD_CTX_cleanup(&tmp_ctx);
+
+	if (ctx->digest->flags & EVP_MD_FLAG_PKEY_METHOD_SIGNATURE)
+		{
+		size_t sltmp = (size_t)EVP_PKEY_size(pkey);
+		i = 0;
+		pkctx = EVP_PKEY_CTX_new(pkey, NULL);
+		if (!pkctx)
+			goto err;
+		if (EVP_PKEY_sign_init(pkctx) <= 0)
+			goto err;
+		if (EVP_PKEY_CTX_set_signature_md(pkctx, ctx->digest) <= 0)
+			goto err;
+		if (EVP_PKEY_sign(pkctx, sigret, &sltmp, m, m_len) <= 0)
+			goto err;
+		*siglen = sltmp;
+		i = 1;
+		err:
+		EVP_PKEY_CTX_free(pkctx);
+		return i;
+		}
+
 	for (i=0; i<4; i++)
 		{
 		v=ctx->digest->required_pkey_type[i];
@@ -99,28 +127,13 @@ int EVP_SignFinal(EVP_MD_CTX *ctx, unsigned char *sigret, unsigned int *siglen,
 		EVPerr(EVP_F_EVP_SIGNFINAL,EVP_R_WRONG_PUBLIC_KEY_TYPE);
 		return(0);
 		}
+
 	if (ctx->digest->sign == NULL)
 		{
 		EVPerr(EVP_F_EVP_SIGNFINAL,EVP_R_NO_SIGN_FUNCTION_CONFIGURED);
 		return(0);
 		}
-	EVP_MD_CTX_init(&tmp_ctx);
-	EVP_MD_CTX_copy_ex(&tmp_ctx,ctx);
-	if (ctx->digest->flags & EVP_MD_FLAG_SVCTX)
-		{
-		EVP_MD_SVCTX sctmp;
-		sctmp.mctx = &tmp_ctx;
-		sctmp.key = pkey->pkey.ptr;
-		i = ctx->digest->sign(ctx->digest->type,
-			NULL, -1, sigret, siglen, &sctmp);
-		}
-	else
-		{
-		EVP_DigestFinal_ex(&tmp_ctx,&(m[0]),&m_len);
-		i = ctx->digest->sign(ctx->digest->type,m,m_len,sigret,siglen,
-					pkey->pkey.ptr);
-		}
-	EVP_MD_CTX_cleanup(&tmp_ctx);
-	return i;
+	return(ctx->digest->sign(ctx->digest->type,m,m_len,sigret,siglen,
+		pkey->pkey.ptr));
 	}
 
diff --git a/crypto/evp/p_verify.c b/crypto/evp/p_verify.c
index 072c127951dc..5f5c409f45fc 100644
--- a/crypto/evp/p_verify.c
+++ b/crypto/evp/p_verify.c
@@ -67,8 +67,32 @@ int EVP_VerifyFinal(EVP_MD_CTX *ctx, const unsigned char *sigbuf,
 	{
 	unsigned char m[EVP_MAX_MD_SIZE];
 	unsigned int m_len;
-	int i,ok=0,v;
+	int i=-1,ok=0,v;
 	EVP_MD_CTX tmp_ctx;
+	EVP_PKEY_CTX *pkctx = NULL;
+
+	EVP_MD_CTX_init(&tmp_ctx);
+	if (!EVP_MD_CTX_copy_ex(&tmp_ctx,ctx))
+		goto err;    
+	if (!EVP_DigestFinal_ex(&tmp_ctx,&(m[0]),&m_len))
+		goto err;
+	EVP_MD_CTX_cleanup(&tmp_ctx);
+
+	if (ctx->digest->flags & EVP_MD_FLAG_PKEY_METHOD_SIGNATURE)
+		{
+		i = -1;
+		pkctx = EVP_PKEY_CTX_new(pkey, NULL);
+		if (!pkctx)
+			goto err;
+		if (EVP_PKEY_verify_init(pkctx) <= 0)
+			goto err;
+		if (EVP_PKEY_CTX_set_signature_md(pkctx, ctx->digest) <= 0)
+			goto err;
+		i = EVP_PKEY_verify(pkctx, sigbuf, siglen, m, m_len);
+		err:
+		EVP_PKEY_CTX_free(pkctx);
+		return i;
+		}
 
 	for (i=0; i<4; i++)
 		{
@@ -85,29 +109,13 @@ int EVP_VerifyFinal(EVP_MD_CTX *ctx, const unsigned char *sigbuf,
 		EVPerr(EVP_F_EVP_VERIFYFINAL,EVP_R_WRONG_PUBLIC_KEY_TYPE);
 		return(-1);
 		}
-	if (ctx->digest->verify == NULL)
+        if (ctx->digest->verify == NULL)
                 {
 		EVPerr(EVP_F_EVP_VERIFYFINAL,EVP_R_NO_VERIFY_FUNCTION_CONFIGURED);
 		return(0);
 		}
 
-	EVP_MD_CTX_init(&tmp_ctx);
-	EVP_MD_CTX_copy_ex(&tmp_ctx,ctx);     
-	if (ctx->digest->flags & EVP_MD_FLAG_SVCTX)
-		{
-		EVP_MD_SVCTX sctmp;
-		sctmp.mctx = &tmp_ctx;
-		sctmp.key = pkey->pkey.ptr;
-		i = ctx->digest->verify(ctx->digest->type,
-			NULL, -1, sigbuf, siglen, &sctmp);
-		}
-	else
-		{
-		EVP_DigestFinal_ex(&tmp_ctx,&(m[0]),&m_len);
-		i = ctx->digest->verify(ctx->digest->type,m,m_len,
-					sigbuf,siglen,pkey->pkey.ptr);
-		}
-	EVP_MD_CTX_cleanup(&tmp_ctx);
-	return i;
+	return(ctx->digest->verify(ctx->digest->type,m,m_len,
+		sigbuf,siglen,pkey->pkey.ptr));
 	}
 
diff --git a/crypto/evp/pmeth_fn.c b/crypto/evp/pmeth_fn.c
new file mode 100644
index 000000000000..c4676f2f8df5
--- /dev/null
+++ b/crypto/evp/pmeth_fn.c
@@ -0,0 +1,368 @@
+/* pmeth_fn.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/evp.h>
+#include "evp_locl.h"
+
+#define M_check_autoarg(ctx, arg, arglen, err) \
+	if (ctx->pmeth->flags & EVP_PKEY_FLAG_AUTOARGLEN) \
+		{ \
+		size_t pksize = (size_t)EVP_PKEY_size(ctx->pkey); \
+		if (!arg) \
+			{ \
+			*arglen = pksize; \
+			return 1; \
+			} \
+		else if (*arglen < pksize) \
+			{ \
+			EVPerr(err, EVP_R_BUFFER_TOO_SMALL); /*ckerr_ignore*/\
+			return 0; \
+			} \
+		}
+
+int EVP_PKEY_sign_init(EVP_PKEY_CTX *ctx)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->sign)
+		{
+		EVPerr(EVP_F_EVP_PKEY_SIGN_INIT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	ctx->operation = EVP_PKEY_OP_SIGN;
+	if (!ctx->pmeth->sign_init)
+		return 1;
+	ret = ctx->pmeth->sign_init(ctx);
+	if (ret <= 0)
+		ctx->operation = EVP_PKEY_OP_UNDEFINED;
+	return ret;
+	}
+
+int EVP_PKEY_sign(EVP_PKEY_CTX *ctx,
+			unsigned char *sig, size_t *siglen,
+			const unsigned char *tbs, size_t tbslen)
+	{
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->sign)
+		{
+		EVPerr(EVP_F_EVP_PKEY_SIGN,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	if (ctx->operation != EVP_PKEY_OP_SIGN)
+		{
+		EVPerr(EVP_F_EVP_PKEY_SIGN, EVP_R_OPERATON_NOT_INITIALIZED);
+		return -1;
+		}
+	M_check_autoarg(ctx, sig, siglen, EVP_F_EVP_PKEY_SIGN)
+	return ctx->pmeth->sign(ctx, sig, siglen, tbs, tbslen);
+	}
+
+int EVP_PKEY_verify_init(EVP_PKEY_CTX *ctx)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->verify)
+		{
+		EVPerr(EVP_F_EVP_PKEY_VERIFY_INIT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	ctx->operation = EVP_PKEY_OP_VERIFY;
+	if (!ctx->pmeth->verify_init)
+		return 1;
+	ret = ctx->pmeth->verify_init(ctx);
+	if (ret <= 0)
+		ctx->operation = EVP_PKEY_OP_UNDEFINED;
+	return ret;
+	}
+
+int EVP_PKEY_verify(EVP_PKEY_CTX *ctx,
+			const unsigned char *sig, size_t siglen,
+			const unsigned char *tbs, size_t tbslen)
+	{
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->verify)
+		{
+		EVPerr(EVP_F_EVP_PKEY_VERIFY,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	if (ctx->operation != EVP_PKEY_OP_VERIFY)
+		{
+		EVPerr(EVP_F_EVP_PKEY_VERIFY, EVP_R_OPERATON_NOT_INITIALIZED);
+		return -1;
+		}
+	return ctx->pmeth->verify(ctx, sig, siglen, tbs, tbslen);
+	}
+
+int EVP_PKEY_verify_recover_init(EVP_PKEY_CTX *ctx)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->verify_recover)
+		{
+		EVPerr(EVP_F_EVP_PKEY_VERIFY_RECOVER_INIT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	ctx->operation = EVP_PKEY_OP_VERIFYRECOVER;
+	if (!ctx->pmeth->verify_recover_init)
+		return 1;
+	ret = ctx->pmeth->verify_recover_init(ctx);
+	if (ret <= 0)
+		ctx->operation = EVP_PKEY_OP_UNDEFINED;
+	return ret;
+	}
+
+int EVP_PKEY_verify_recover(EVP_PKEY_CTX *ctx,
+			unsigned char *rout, size_t *routlen,
+			const unsigned char *sig, size_t siglen)
+	{
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->verify_recover)
+		{
+		EVPerr(EVP_F_EVP_PKEY_VERIFY_RECOVER,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	if (ctx->operation != EVP_PKEY_OP_VERIFYRECOVER)
+		{
+		EVPerr(EVP_F_EVP_PKEY_VERIFY_RECOVER, EVP_R_OPERATON_NOT_INITIALIZED);
+		return -1;
+		}
+	M_check_autoarg(ctx, rout, routlen, EVP_F_EVP_PKEY_VERIFY_RECOVER)
+	return ctx->pmeth->verify_recover(ctx, rout, routlen, sig, siglen);
+	}
+
+int EVP_PKEY_encrypt_init(EVP_PKEY_CTX *ctx)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->encrypt)
+		{
+		EVPerr(EVP_F_EVP_PKEY_ENCRYPT_INIT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	ctx->operation = EVP_PKEY_OP_ENCRYPT;
+	if (!ctx->pmeth->encrypt_init)
+		return 1;
+	ret = ctx->pmeth->encrypt_init(ctx);
+	if (ret <= 0)
+		ctx->operation = EVP_PKEY_OP_UNDEFINED;
+	return ret;
+	}
+
+int EVP_PKEY_encrypt(EVP_PKEY_CTX *ctx,
+			unsigned char *out, size_t *outlen,
+			const unsigned char *in, size_t inlen)
+	{
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->encrypt)
+		{
+		EVPerr(EVP_F_EVP_PKEY_ENCRYPT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	if (ctx->operation != EVP_PKEY_OP_ENCRYPT)
+		{
+		EVPerr(EVP_F_EVP_PKEY_ENCRYPT, EVP_R_OPERATON_NOT_INITIALIZED);
+		return -1;
+		}
+	M_check_autoarg(ctx, out, outlen, EVP_F_EVP_PKEY_ENCRYPT)
+	return ctx->pmeth->encrypt(ctx, out, outlen, in, inlen);
+	}
+
+int EVP_PKEY_decrypt_init(EVP_PKEY_CTX *ctx)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->decrypt)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DECRYPT_INIT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	ctx->operation = EVP_PKEY_OP_DECRYPT;
+	if (!ctx->pmeth->decrypt_init)
+		return 1;
+	ret = ctx->pmeth->decrypt_init(ctx);
+	if (ret <= 0)
+		ctx->operation = EVP_PKEY_OP_UNDEFINED;
+	return ret;
+	}
+
+int EVP_PKEY_decrypt(EVP_PKEY_CTX *ctx,
+			unsigned char *out, size_t *outlen,
+			const unsigned char *in, size_t inlen)
+	{
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->decrypt)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DECRYPT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	if (ctx->operation != EVP_PKEY_OP_DECRYPT)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DECRYPT, EVP_R_OPERATON_NOT_INITIALIZED);
+		return -1;
+		}
+	M_check_autoarg(ctx, out, outlen, EVP_F_EVP_PKEY_DECRYPT)
+	return ctx->pmeth->decrypt(ctx, out, outlen, in, inlen);
+	}
+
+
+int EVP_PKEY_derive_init(EVP_PKEY_CTX *ctx)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->derive)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DERIVE_INIT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	ctx->operation = EVP_PKEY_OP_DERIVE;
+	if (!ctx->pmeth->derive_init)
+		return 1;
+	ret = ctx->pmeth->derive_init(ctx);
+	if (ret <= 0)
+		ctx->operation = EVP_PKEY_OP_UNDEFINED;
+	return ret;
+	}
+
+int EVP_PKEY_derive_set_peer(EVP_PKEY_CTX *ctx, EVP_PKEY *peer)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !(ctx->pmeth->derive||ctx->pmeth->encrypt||ctx->pmeth->decrypt) || !ctx->pmeth->ctrl)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DERIVE_SET_PEER,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	if (ctx->operation != EVP_PKEY_OP_DERIVE && ctx->operation != EVP_PKEY_OP_ENCRYPT && ctx->operation != EVP_PKEY_OP_DECRYPT)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DERIVE_SET_PEER,
+					EVP_R_OPERATON_NOT_INITIALIZED);
+		return -1;
+		}
+
+	ret = ctx->pmeth->ctrl(ctx, EVP_PKEY_CTRL_PEER_KEY, 0, peer);
+
+	if (ret <= 0)
+		return ret;
+
+	if (ret == 2)
+		return 1;
+
+	if (!ctx->pkey)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DERIVE_SET_PEER, EVP_R_NO_KEY_SET);
+		return -1;
+		}
+
+	if (ctx->pkey->type != peer->type)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DERIVE_SET_PEER,
+						EVP_R_DIFFERENT_KEY_TYPES);
+		return -1;
+		}
+
+	/* ran@cryptocom.ru: For clarity.  The error is if parameters in peer are
+	 * present (!missing) but don't match.  EVP_PKEY_cmp_parameters may return
+	 * 1 (match), 0 (don't match) and -2 (comparison is not defined).  -1
+	 * (different key types) is impossible here because it is checked earlier.
+	 * -2 is OK for us here, as well as 1, so we can check for 0 only. */
+	if (!EVP_PKEY_missing_parameters(peer) &&
+		!EVP_PKEY_cmp_parameters(ctx->pkey, peer))
+		{
+		EVPerr(EVP_F_EVP_PKEY_DERIVE_SET_PEER,
+						EVP_R_DIFFERENT_PARAMETERS);
+		return -1;
+		}
+
+	if (ctx->peerkey)
+		EVP_PKEY_free(ctx->peerkey);
+	ctx->peerkey = peer;
+
+	ret = ctx->pmeth->ctrl(ctx, EVP_PKEY_CTRL_PEER_KEY, 1, peer);
+
+	if (ret <= 0)
+		{
+		ctx->peerkey = NULL;
+		return ret;
+		}
+
+	CRYPTO_add(&peer->references,1,CRYPTO_LOCK_EVP_PKEY);
+	return 1;
+	}
+
+
+int EVP_PKEY_derive(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *pkeylen)
+	{
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->derive)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DERIVE,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	if (ctx->operation != EVP_PKEY_OP_DERIVE)
+		{
+		EVPerr(EVP_F_EVP_PKEY_DERIVE, EVP_R_OPERATON_NOT_INITIALIZED);
+		return -1;
+		}
+	M_check_autoarg(ctx, key, pkeylen, EVP_F_EVP_PKEY_DERIVE)
+	return ctx->pmeth->derive(ctx, key, pkeylen);
+	}
+
diff --git a/crypto/evp/pmeth_gn.c b/crypto/evp/pmeth_gn.c
new file mode 100644
index 000000000000..4651c8137022
--- /dev/null
+++ b/crypto/evp/pmeth_gn.c
@@ -0,0 +1,221 @@
+/* pmeth_gn.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/evp.h>
+#include <openssl/bn.h>
+#include "evp_locl.h"
+
+int EVP_PKEY_paramgen_init(EVP_PKEY_CTX *ctx)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->paramgen)
+		{
+		EVPerr(EVP_F_EVP_PKEY_PARAMGEN_INIT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	ctx->operation = EVP_PKEY_OP_PARAMGEN;
+	if (!ctx->pmeth->paramgen_init)
+		return 1;
+	ret = ctx->pmeth->paramgen_init(ctx);
+	if (ret <= 0)
+		ctx->operation = EVP_PKEY_OP_UNDEFINED;
+	return ret;
+	}
+
+int EVP_PKEY_paramgen(EVP_PKEY_CTX *ctx, EVP_PKEY **ppkey)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->paramgen)
+		{
+		EVPerr(EVP_F_EVP_PKEY_PARAMGEN,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+
+	if (ctx->operation != EVP_PKEY_OP_PARAMGEN)
+		{
+		EVPerr(EVP_F_EVP_PKEY_PARAMGEN, EVP_R_OPERATON_NOT_INITIALIZED);
+		return -1;
+		}
+
+	if (!ppkey)
+		return -1;
+
+	if (!*ppkey)
+		*ppkey = EVP_PKEY_new();
+
+	ret = ctx->pmeth->paramgen(ctx, *ppkey);
+	if (ret <= 0)
+		{
+		EVP_PKEY_free(*ppkey);
+		*ppkey = NULL;
+		}
+	return ret;
+	}
+
+int EVP_PKEY_keygen_init(EVP_PKEY_CTX *ctx)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->keygen)
+		{
+		EVPerr(EVP_F_EVP_PKEY_KEYGEN_INIT,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	ctx->operation = EVP_PKEY_OP_KEYGEN;
+	if (!ctx->pmeth->keygen_init)
+		return 1;
+	ret = ctx->pmeth->keygen_init(ctx);
+	if (ret <= 0)
+		ctx->operation = EVP_PKEY_OP_UNDEFINED;
+	return ret;
+	}
+
+int EVP_PKEY_keygen(EVP_PKEY_CTX *ctx, EVP_PKEY **ppkey)
+	{
+	int ret;
+
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->keygen)
+		{
+		EVPerr(EVP_F_EVP_PKEY_KEYGEN,
+			EVP_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+		return -2;
+		}
+	if (ctx->operation != EVP_PKEY_OP_KEYGEN)
+		{
+		EVPerr(EVP_F_EVP_PKEY_KEYGEN, EVP_R_OPERATON_NOT_INITIALIZED);
+		return -1;
+		}
+
+	if (!ppkey)
+		return -1;
+
+	if (!*ppkey)
+		*ppkey = EVP_PKEY_new();
+
+	ret = ctx->pmeth->keygen(ctx, *ppkey);
+	if (ret <= 0)
+		{
+		EVP_PKEY_free(*ppkey);
+		*ppkey = NULL;
+		}
+	return ret;
+	}
+
+void EVP_PKEY_CTX_set_cb(EVP_PKEY_CTX *ctx, EVP_PKEY_gen_cb *cb)
+	{
+	ctx->pkey_gencb = cb;
+	}
+
+EVP_PKEY_gen_cb *EVP_PKEY_CTX_get_cb(EVP_PKEY_CTX *ctx)
+	{
+	return ctx->pkey_gencb;
+	}
+
+/* "translation callback" to call EVP_PKEY_CTX callbacks using BN_GENCB
+ * style callbacks.
+ */
+
+static int trans_cb(int a, int b, BN_GENCB *gcb)
+	{
+	EVP_PKEY_CTX *ctx = gcb->arg;
+	ctx->keygen_info[0] = a;
+	ctx->keygen_info[1] = b;
+	return ctx->pkey_gencb(ctx);
+	}	
+
+void evp_pkey_set_cb_translate(BN_GENCB *cb, EVP_PKEY_CTX *ctx)
+	{
+	BN_GENCB_set(cb, trans_cb, ctx)
+	}
+
+int EVP_PKEY_CTX_get_keygen_info(EVP_PKEY_CTX *ctx, int idx)
+	{
+	if (idx == -1)
+		return ctx->keygen_info_count; 
+	if (idx < 0 || idx > ctx->keygen_info_count)
+		return 0;
+	return ctx->keygen_info[idx];
+	}
+
+EVP_PKEY *EVP_PKEY_new_mac_key(int type, ENGINE *e,
+				const unsigned char *key, int keylen)
+	{
+	EVP_PKEY_CTX *mac_ctx = NULL;
+	EVP_PKEY *mac_key = NULL;
+	mac_ctx = EVP_PKEY_CTX_new_id(type, e);
+	if (!mac_ctx)
+		return NULL;
+	if (EVP_PKEY_keygen_init(mac_ctx) <= 0)
+		goto merr;
+	if (EVP_PKEY_CTX_ctrl(mac_ctx, -1, EVP_PKEY_OP_KEYGEN,
+				EVP_PKEY_CTRL_SET_MAC_KEY,
+				keylen, (void *)key) <= 0)
+		goto merr;
+	if (EVP_PKEY_keygen(mac_ctx, &mac_key) <= 0)
+		goto merr;
+	merr:
+	if (mac_ctx)
+		EVP_PKEY_CTX_free(mac_ctx);
+	return mac_key;
+	}
diff --git a/crypto/evp/pmeth_lib.c b/crypto/evp/pmeth_lib.c
new file mode 100644
index 000000000000..acfa7b6f873d
--- /dev/null
+++ b/crypto/evp/pmeth_lib.c
@@ -0,0 +1,593 @@
+/* pmeth_lib.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/evp.h>
+#ifndef OPENSSL_NO_ENGINE
+#include <openssl/engine.h>
+#endif
+#include "asn1_locl.h"
+#include "evp_locl.h"
+
+typedef int sk_cmp_fn_type(const char * const *a, const char * const *b);
+
+DECLARE_STACK_OF(EVP_PKEY_METHOD)
+STACK_OF(EVP_PKEY_METHOD) *app_pkey_methods = NULL;
+
+extern const EVP_PKEY_METHOD rsa_pkey_meth, dh_pkey_meth, dsa_pkey_meth;
+extern const EVP_PKEY_METHOD ec_pkey_meth, hmac_pkey_meth, cmac_pkey_meth;
+
+static const EVP_PKEY_METHOD *standard_methods[] =
+	{
+#ifndef OPENSSL_NO_RSA
+	&rsa_pkey_meth,
+#endif
+#ifndef OPENSSL_NO_DH
+	&dh_pkey_meth,
+#endif
+#ifndef OPENSSL_NO_DSA
+	&dsa_pkey_meth,
+#endif
+#ifndef OPENSSL_NO_EC
+	&ec_pkey_meth,
+#endif
+	&hmac_pkey_meth,
+	&cmac_pkey_meth
+	};
+
+DECLARE_OBJ_BSEARCH_CMP_FN(const EVP_PKEY_METHOD *, const EVP_PKEY_METHOD *,
+			   pmeth);
+
+static int pmeth_cmp(const EVP_PKEY_METHOD * const *a,
+		     const EVP_PKEY_METHOD * const *b)
+	{
+        return ((*a)->pkey_id - (*b)->pkey_id);
+	}
+
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(const EVP_PKEY_METHOD *, const EVP_PKEY_METHOD *,
+			     pmeth);
+
+const EVP_PKEY_METHOD *EVP_PKEY_meth_find(int type)
+	{
+	EVP_PKEY_METHOD tmp;
+	const EVP_PKEY_METHOD *t = &tmp, **ret;
+	tmp.pkey_id = type;
+	if (app_pkey_methods)
+		{
+		int idx;
+		idx = sk_EVP_PKEY_METHOD_find(app_pkey_methods, &tmp);
+		if (idx >= 0)
+			return sk_EVP_PKEY_METHOD_value(app_pkey_methods, idx);
+		}
+	ret = OBJ_bsearch_pmeth(&t, standard_methods,
+			  sizeof(standard_methods)/sizeof(EVP_PKEY_METHOD *));
+	if (!ret || !*ret)
+		return NULL;
+	return *ret;
+	}
+
+static EVP_PKEY_CTX *int_ctx_new(EVP_PKEY *pkey, ENGINE *e, int id)
+	{
+	EVP_PKEY_CTX *ret;
+	const EVP_PKEY_METHOD *pmeth;
+	if (id == -1)
+		{
+		if (!pkey || !pkey->ameth)
+			return NULL;
+		id = pkey->ameth->pkey_id;
+		}
+#ifndef OPENSSL_NO_ENGINE
+	if (pkey && pkey->engine)
+		e = pkey->engine;
+	/* Try to find an ENGINE which implements this method */
+	if (e)
+		{
+		if (!ENGINE_init(e))
+			{
+			EVPerr(EVP_F_INT_CTX_NEW,ERR_R_ENGINE_LIB);
+			return NULL;
+			}
+		}
+	else
+		e = ENGINE_get_pkey_meth_engine(id);
+
+	/* If an ENGINE handled this method look it up. Othewise
+	 * use internal tables.
+	 */
+
+	if (e)
+		pmeth = ENGINE_get_pkey_meth(e, id);
+	else
+#endif
+		pmeth = EVP_PKEY_meth_find(id);
+
+	if (pmeth == NULL)
+		{
+		EVPerr(EVP_F_INT_CTX_NEW,EVP_R_UNSUPPORTED_ALGORITHM);
+		return NULL;
+		}
+
+	ret = OPENSSL_malloc(sizeof(EVP_PKEY_CTX));
+	if (!ret)
+		{
+#ifndef OPENSSL_NO_ENGINE
+		if (e)
+			ENGINE_finish(e);
+#endif
+		EVPerr(EVP_F_INT_CTX_NEW,ERR_R_MALLOC_FAILURE);
+		return NULL;
+		}
+	ret->engine = e;
+	ret->pmeth = pmeth;
+	ret->operation = EVP_PKEY_OP_UNDEFINED;
+	ret->pkey = pkey;
+	ret->peerkey = NULL;
+	ret->pkey_gencb = 0;
+	if (pkey)
+		CRYPTO_add(&pkey->references,1,CRYPTO_LOCK_EVP_PKEY);
+	ret->data = NULL;
+
+	if (pmeth->init)
+		{
+		if (pmeth->init(ret) <= 0)
+			{
+			EVP_PKEY_CTX_free(ret);
+			return NULL;
+			}
+		}
+
+	return ret;
+	}
+
+EVP_PKEY_METHOD* EVP_PKEY_meth_new(int id, int flags)
+	{
+	EVP_PKEY_METHOD *pmeth;
+	pmeth = OPENSSL_malloc(sizeof(EVP_PKEY_METHOD));
+	if (!pmeth)
+		return NULL;
+
+	memset(pmeth, 0, sizeof(EVP_PKEY_METHOD));
+
+	pmeth->pkey_id = id;
+	pmeth->flags = flags | EVP_PKEY_FLAG_DYNAMIC;
+
+	pmeth->init = 0;
+	pmeth->copy = 0;
+	pmeth->cleanup = 0;
+	pmeth->paramgen_init = 0;
+	pmeth->paramgen = 0;
+	pmeth->keygen_init = 0;
+	pmeth->keygen = 0;
+	pmeth->sign_init = 0;
+	pmeth->sign = 0;
+	pmeth->verify_init = 0;
+	pmeth->verify = 0;
+	pmeth->verify_recover_init = 0;
+	pmeth->verify_recover = 0;
+	pmeth->signctx_init = 0;
+	pmeth->signctx = 0;
+	pmeth->verifyctx_init = 0;
+	pmeth->verifyctx = 0;
+	pmeth->encrypt_init = 0;
+	pmeth->encrypt = 0;
+	pmeth->decrypt_init = 0;
+	pmeth->decrypt = 0;
+	pmeth->derive_init = 0;
+	pmeth->derive = 0;
+	pmeth->ctrl = 0;
+	pmeth->ctrl_str = 0;
+
+	return pmeth;
+	}
+
+void EVP_PKEY_meth_get0_info(int *ppkey_id, int *pflags,
+				const EVP_PKEY_METHOD *meth)
+	{
+	if (ppkey_id)
+		*ppkey_id = meth->pkey_id;
+	if (pflags)
+		*pflags = meth->flags;
+	}
+
+void EVP_PKEY_meth_copy(EVP_PKEY_METHOD *dst, const EVP_PKEY_METHOD *src)
+	{
+
+	dst->init = src->init;
+	dst->copy = src->copy;
+	dst->cleanup = src->cleanup;
+
+	dst->paramgen_init = src->paramgen_init;
+	dst->paramgen = src->paramgen;
+
+	dst->keygen_init = src->keygen_init;
+	dst->keygen = src->keygen;
+
+	dst->sign_init = src->sign_init;
+	dst->sign = src->sign;
+
+	dst->verify_init = src->verify_init;
+	dst->verify = src->verify;
+
+	dst->verify_recover_init = src->verify_recover_init;
+	dst->verify_recover = src->verify_recover;
+
+	dst->signctx_init = src->signctx_init;
+	dst->signctx = src->signctx;
+
+	dst->verifyctx_init = src->verifyctx_init;
+	dst->verifyctx = src->verifyctx;
+
+	dst->encrypt_init = src->encrypt_init;
+	dst->encrypt = src->encrypt;
+
+	dst->decrypt_init = src->decrypt_init;
+	dst->decrypt = src->decrypt;
+
+	dst->derive_init = src->derive_init;
+	dst->derive = src->derive;
+
+	dst->ctrl = src->ctrl;
+	dst->ctrl_str = src->ctrl_str;
+	}
+
+void EVP_PKEY_meth_free(EVP_PKEY_METHOD *pmeth)
+	{
+	if (pmeth && (pmeth->flags & EVP_PKEY_FLAG_DYNAMIC))
+		OPENSSL_free(pmeth);
+	}
+
+EVP_PKEY_CTX *EVP_PKEY_CTX_new(EVP_PKEY *pkey, ENGINE *e)
+	{
+	return int_ctx_new(pkey, e, -1);
+	}
+
+EVP_PKEY_CTX *EVP_PKEY_CTX_new_id(int id, ENGINE *e)
+	{
+	return int_ctx_new(NULL, e, id);
+	}
+
+EVP_PKEY_CTX *EVP_PKEY_CTX_dup(EVP_PKEY_CTX *pctx)
+	{
+	EVP_PKEY_CTX *rctx;
+	if (!pctx->pmeth || !pctx->pmeth->copy)
+		return NULL;
+#ifndef OPENSSL_NO_ENGINE
+	/* Make sure it's safe to copy a pkey context using an ENGINE */
+	if (pctx->engine && !ENGINE_init(pctx->engine))
+		{
+		EVPerr(EVP_F_EVP_PKEY_CTX_DUP,ERR_R_ENGINE_LIB);
+		return 0;
+		}
+#endif
+	rctx = OPENSSL_malloc(sizeof(EVP_PKEY_CTX));
+	if (!rctx)
+		return NULL;
+
+	rctx->pmeth = pctx->pmeth;
+#ifndef OPENSSL_NO_ENGINE
+	rctx->engine = pctx->engine;
+#endif
+
+	if (pctx->pkey)
+		CRYPTO_add(&pctx->pkey->references,1,CRYPTO_LOCK_EVP_PKEY);
+
+	rctx->pkey = pctx->pkey;
+
+	if (pctx->peerkey)
+		CRYPTO_add(&pctx->peerkey->references,1,CRYPTO_LOCK_EVP_PKEY);
+
+	rctx->peerkey = pctx->peerkey;
+
+	rctx->data = NULL;
+	rctx->app_data = NULL;
+	rctx->operation = pctx->operation;
+
+	if (pctx->pmeth->copy(rctx, pctx) > 0)
+		return rctx;
+
+	EVP_PKEY_CTX_free(rctx);
+	return NULL;
+
+	}
+
+int EVP_PKEY_meth_add0(const EVP_PKEY_METHOD *pmeth)
+	{
+	if (app_pkey_methods == NULL)
+		{
+		app_pkey_methods = sk_EVP_PKEY_METHOD_new(pmeth_cmp);
+		if (!app_pkey_methods)
+			return 0;
+		}
+	if (!sk_EVP_PKEY_METHOD_push(app_pkey_methods, pmeth))
+		return 0;
+	sk_EVP_PKEY_METHOD_sort(app_pkey_methods);
+	return 1;
+	}
+
+void EVP_PKEY_CTX_free(EVP_PKEY_CTX *ctx)
+	{
+	if (ctx == NULL)
+		return;
+	if (ctx->pmeth && ctx->pmeth->cleanup)
+		ctx->pmeth->cleanup(ctx);
+	if (ctx->pkey)
+		EVP_PKEY_free(ctx->pkey);
+	if (ctx->peerkey)
+		EVP_PKEY_free(ctx->peerkey);
+#ifndef OPENSSL_NO_ENGINE
+	if(ctx->engine)
+		/* The EVP_PKEY_CTX we used belongs to an ENGINE, release the
+		 * functional reference we held for this reason. */
+		ENGINE_finish(ctx->engine);
+#endif
+	OPENSSL_free(ctx);
+	}
+
+int EVP_PKEY_CTX_ctrl(EVP_PKEY_CTX *ctx, int keytype, int optype,
+				int cmd, int p1, void *p2)
+	{
+	int ret;
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->ctrl)
+		{
+		EVPerr(EVP_F_EVP_PKEY_CTX_CTRL, EVP_R_COMMAND_NOT_SUPPORTED);
+		return -2;
+		}
+	if ((keytype != -1) && (ctx->pmeth->pkey_id != keytype))
+		return -1;
+
+	if (ctx->operation == EVP_PKEY_OP_UNDEFINED)
+		{
+		EVPerr(EVP_F_EVP_PKEY_CTX_CTRL, EVP_R_NO_OPERATION_SET);
+		return -1;
+		}
+
+	if ((optype != -1) && !(ctx->operation & optype))
+		{
+		EVPerr(EVP_F_EVP_PKEY_CTX_CTRL, EVP_R_INVALID_OPERATION);
+		return -1;
+		}
+
+	ret = ctx->pmeth->ctrl(ctx, cmd, p1, p2);
+
+	if (ret == -2)
+		EVPerr(EVP_F_EVP_PKEY_CTX_CTRL, EVP_R_COMMAND_NOT_SUPPORTED);
+
+	return ret;
+
+	}
+
+int EVP_PKEY_CTX_ctrl_str(EVP_PKEY_CTX *ctx,
+					const char *name, const char *value)
+	{
+	if (!ctx || !ctx->pmeth || !ctx->pmeth->ctrl_str)
+		{
+		EVPerr(EVP_F_EVP_PKEY_CTX_CTRL_STR,
+						EVP_R_COMMAND_NOT_SUPPORTED);
+		return -2;
+		}
+	if (!strcmp(name, "digest"))
+		{
+		const EVP_MD *md;
+		if (!value || !(md = EVP_get_digestbyname(value)))
+			{
+			EVPerr(EVP_F_EVP_PKEY_CTX_CTRL_STR,
+						EVP_R_INVALID_DIGEST);
+			return 0;
+			}
+		return EVP_PKEY_CTX_set_signature_md(ctx, md);
+		}
+	return ctx->pmeth->ctrl_str(ctx, name, value);
+	}
+
+int EVP_PKEY_CTX_get_operation(EVP_PKEY_CTX *ctx)
+	{
+	return ctx->operation;
+	}
+
+void EVP_PKEY_CTX_set0_keygen_info(EVP_PKEY_CTX *ctx, int *dat, int datlen)
+	{
+	ctx->keygen_info = dat;
+	ctx->keygen_info_count = datlen;
+	}
+
+void EVP_PKEY_CTX_set_data(EVP_PKEY_CTX *ctx, void *data)
+	{
+	ctx->data = data;
+	}
+
+void *EVP_PKEY_CTX_get_data(EVP_PKEY_CTX *ctx)
+	{
+	return ctx->data;
+	}
+
+EVP_PKEY *EVP_PKEY_CTX_get0_pkey(EVP_PKEY_CTX *ctx)
+	{
+	return ctx->pkey;
+	}
+
+EVP_PKEY *EVP_PKEY_CTX_get0_peerkey(EVP_PKEY_CTX *ctx)
+	{
+	return ctx->peerkey;
+	}
+	
+void EVP_PKEY_CTX_set_app_data(EVP_PKEY_CTX *ctx, void *data)
+	{
+	ctx->app_data = data;
+	}
+
+void *EVP_PKEY_CTX_get_app_data(EVP_PKEY_CTX *ctx)
+	{
+	return ctx->app_data;
+	}
+
+void EVP_PKEY_meth_set_init(EVP_PKEY_METHOD *pmeth,
+	int (*init)(EVP_PKEY_CTX *ctx))
+	{
+	pmeth->init = init;
+	}
+
+void EVP_PKEY_meth_set_copy(EVP_PKEY_METHOD *pmeth,
+	int (*copy)(EVP_PKEY_CTX *dst, EVP_PKEY_CTX *src))
+	{
+	pmeth->copy = copy;
+	}
+
+void EVP_PKEY_meth_set_cleanup(EVP_PKEY_METHOD *pmeth,
+	void (*cleanup)(EVP_PKEY_CTX *ctx))
+	{
+	pmeth->cleanup = cleanup;
+	}
+
+void EVP_PKEY_meth_set_paramgen(EVP_PKEY_METHOD *pmeth,
+	int (*paramgen_init)(EVP_PKEY_CTX *ctx),
+	int (*paramgen)(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey))
+	{
+	pmeth->paramgen_init = paramgen_init;
+	pmeth->paramgen = paramgen;
+	}
+
+void EVP_PKEY_meth_set_keygen(EVP_PKEY_METHOD *pmeth,
+	int (*keygen_init)(EVP_PKEY_CTX *ctx),
+	int (*keygen)(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey))
+	{
+	pmeth->keygen_init = keygen_init;
+	pmeth->keygen = keygen;
+	}
+
+void EVP_PKEY_meth_set_sign(EVP_PKEY_METHOD *pmeth,
+	int (*sign_init)(EVP_PKEY_CTX *ctx),
+	int (*sign)(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					const unsigned char *tbs, size_t tbslen))
+	{
+	pmeth->sign_init = sign_init;
+	pmeth->sign = sign;
+	}
+
+void EVP_PKEY_meth_set_verify(EVP_PKEY_METHOD *pmeth,
+	int (*verify_init)(EVP_PKEY_CTX *ctx),
+	int (*verify)(EVP_PKEY_CTX *ctx, const unsigned char *sig, size_t siglen,
+					const unsigned char *tbs, size_t tbslen))
+	{
+	pmeth->verify_init = verify_init;
+	pmeth->verify = verify;
+	}
+
+void EVP_PKEY_meth_set_verify_recover(EVP_PKEY_METHOD *pmeth,
+	int (*verify_recover_init)(EVP_PKEY_CTX *ctx),
+	int (*verify_recover)(EVP_PKEY_CTX *ctx,
+					unsigned char *sig, size_t *siglen,
+					const unsigned char *tbs, size_t tbslen))
+	{
+	pmeth->verify_recover_init = verify_recover_init;
+	pmeth->verify_recover = verify_recover;
+	}
+
+void EVP_PKEY_meth_set_signctx(EVP_PKEY_METHOD *pmeth,
+	int (*signctx_init)(EVP_PKEY_CTX *ctx, EVP_MD_CTX *mctx),
+	int (*signctx)(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					EVP_MD_CTX *mctx))
+	{
+	pmeth->signctx_init = signctx_init;
+	pmeth->signctx = signctx;
+	}
+
+void EVP_PKEY_meth_set_verifyctx(EVP_PKEY_METHOD *pmeth,
+	int (*verifyctx_init)(EVP_PKEY_CTX *ctx, EVP_MD_CTX *mctx),
+	int (*verifyctx)(EVP_PKEY_CTX *ctx, const unsigned char *sig,int siglen,
+					EVP_MD_CTX *mctx))
+	{
+	pmeth->verifyctx_init = verifyctx_init;
+	pmeth->verifyctx = verifyctx;
+	}
+
+void EVP_PKEY_meth_set_encrypt(EVP_PKEY_METHOD *pmeth,
+	int (*encrypt_init)(EVP_PKEY_CTX *ctx),
+	int (*encryptfn)(EVP_PKEY_CTX *ctx, unsigned char *out, size_t *outlen,
+					const unsigned char *in, size_t inlen))
+	{
+	pmeth->encrypt_init = encrypt_init;
+	pmeth->encrypt = encryptfn;
+	}
+
+void EVP_PKEY_meth_set_decrypt(EVP_PKEY_METHOD *pmeth,
+	int (*decrypt_init)(EVP_PKEY_CTX *ctx),
+	int (*decrypt)(EVP_PKEY_CTX *ctx, unsigned char *out, size_t *outlen,
+					const unsigned char *in, size_t inlen))
+	{
+	pmeth->decrypt_init = decrypt_init;
+	pmeth->decrypt = decrypt;
+	}
+
+void EVP_PKEY_meth_set_derive(EVP_PKEY_METHOD *pmeth,
+	int (*derive_init)(EVP_PKEY_CTX *ctx),
+	int (*derive)(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen))
+	{
+	pmeth->derive_init = derive_init;
+	pmeth->derive = derive;
+	}
+
+void EVP_PKEY_meth_set_ctrl(EVP_PKEY_METHOD *pmeth,
+	int (*ctrl)(EVP_PKEY_CTX *ctx, int type, int p1, void *p2),
+	int (*ctrl_str)(EVP_PKEY_CTX *ctx, const char *type, const char *value))
+	{
+	pmeth->ctrl = ctrl;
+	pmeth->ctrl_str = ctrl_str;
+	}
diff --git a/crypto/ex_data.c b/crypto/ex_data.c
index 3b11e7a55613..e2bc8298d0a7 100644
--- a/crypto/ex_data.c
+++ b/crypto/ex_data.c
@@ -245,18 +245,21 @@ typedef struct st_ex_class_item {
 static int ex_class = CRYPTO_EX_INDEX_USER;
 
 /* The global hash table of EX_CLASS_ITEM items */
-static LHASH *ex_data = NULL;
+DECLARE_LHASH_OF(EX_CLASS_ITEM);
+static LHASH_OF(EX_CLASS_ITEM) *ex_data = NULL;
 
 /* The callbacks required in the "ex_data" hash table */
-static unsigned long ex_hash_cb(const void *a_void)
+static unsigned long ex_class_item_hash(const EX_CLASS_ITEM *a)
 	{
-	return ((const EX_CLASS_ITEM *)a_void)->class_index;
+	return a->class_index;
 	}
-static int ex_cmp_cb(const void *a_void, const void *b_void)
+static IMPLEMENT_LHASH_HASH_FN(ex_class_item, EX_CLASS_ITEM)
+
+static int ex_class_item_cmp(const EX_CLASS_ITEM *a, const EX_CLASS_ITEM *b)
 	{
-	return (((const EX_CLASS_ITEM *)a_void)->class_index -
-		((const EX_CLASS_ITEM *)b_void)->class_index);
+	return a->class_index - b->class_index;
 	}
+static IMPLEMENT_LHASH_COMP_FN(ex_class_item, EX_CLASS_ITEM)
 
 /* Internal functions used by the "impl_default" implementation to access the
  * state */
@@ -265,7 +268,8 @@ static int ex_data_check(void)
 	{
 	int toret = 1;
 	CRYPTO_w_lock(CRYPTO_LOCK_EX_DATA);
-	if(!ex_data && ((ex_data = lh_new(ex_hash_cb, ex_cmp_cb)) == NULL))
+	if(!ex_data
+	   && (ex_data = lh_EX_CLASS_ITEM_new()) == NULL)
 		toret = 0;
 	CRYPTO_w_unlock(CRYPTO_LOCK_EX_DATA);
 	return toret;
@@ -298,7 +302,7 @@ static EX_CLASS_ITEM *def_get_class(int class_index)
 	EX_DATA_CHECK(return NULL;)
 	d.class_index = class_index;
 	CRYPTO_w_lock(CRYPTO_LOCK_EX_DATA);
-	p = lh_retrieve(ex_data, &d);
+	p = lh_EX_CLASS_ITEM_retrieve(ex_data, &d);
 	if(!p)
 		{
 		gen = OPENSSL_malloc(sizeof(EX_CLASS_ITEM));
@@ -313,7 +317,7 @@ static EX_CLASS_ITEM *def_get_class(int class_index)
 				{
 				/* Because we're inside the ex_data lock, the
 				 * return value from the insert will be NULL */
-				lh_insert(ex_data, gen);
+				(void)lh_EX_CLASS_ITEM_insert(ex_data, gen);
 				p = gen;
 				}
 			}
@@ -375,8 +379,8 @@ static int int_new_class(void)
 static void int_cleanup(void)
 	{
 	EX_DATA_CHECK(return;)
-	lh_doall(ex_data, def_cleanup_cb);
-	lh_free(ex_data);
+	lh_EX_CLASS_ITEM_doall(ex_data, def_cleanup_cb);
+	lh_EX_CLASS_ITEM_free(ex_data);
 	ex_data = NULL;
 	impl = NULL;
 	}
@@ -452,7 +456,7 @@ static int int_dup_ex_data(int class_index, CRYPTO_EX_DATA *to,
 		return 0;
 	CRYPTO_r_lock(CRYPTO_LOCK_EX_DATA);
 	mx = sk_CRYPTO_EX_DATA_FUNCS_num(item->meth);
-	j = sk_num(from->sk);
+	j = sk_void_num(from->sk);
 	if(j < mx)
 		mx = j;
 	if(mx > 0)
@@ -523,7 +527,7 @@ skip:
 		OPENSSL_free(storage);
 	if(ad->sk)
 		{
-		sk_free(ad->sk);
+		sk_void_free(ad->sk);
 		ad->sk=NULL;
 		}
 	}
@@ -596,24 +600,24 @@ int CRYPTO_set_ex_data(CRYPTO_EX_DATA *ad, int idx, void *val)
 
 	if (ad->sk == NULL)
 		{
-		if ((ad->sk=sk_new_null()) == NULL)
+		if ((ad->sk=sk_void_new_null()) == NULL)
 			{
 			CRYPTOerr(CRYPTO_F_CRYPTO_SET_EX_DATA,ERR_R_MALLOC_FAILURE);
 			return(0);
 			}
 		}
-	i=sk_num(ad->sk);
+	i=sk_void_num(ad->sk);
 
 	while (i <= idx)
 		{
-		if (!sk_push(ad->sk,NULL))
+		if (!sk_void_push(ad->sk,NULL))
 			{
 			CRYPTOerr(CRYPTO_F_CRYPTO_SET_EX_DATA,ERR_R_MALLOC_FAILURE);
 			return(0);
 			}
 		i++;
 		}
-	sk_set(ad->sk,idx,val);
+	sk_void_set(ad->sk,idx,val);
 	return(1);
 	}
 
@@ -623,10 +627,10 @@ void *CRYPTO_get_ex_data(const CRYPTO_EX_DATA *ad, int idx)
 	{
 	if (ad->sk == NULL)
 		return(0);
-	else if (idx >= sk_num(ad->sk))
+	else if (idx >= sk_void_num(ad->sk))
 		return(0);
 	else
-		return(sk_value(ad->sk,idx));
+		return(sk_void_value(ad->sk,idx));
 	}
 
 IMPLEMENT_STACK_OF(CRYPTO_EX_DATA_FUNCS)
diff --git a/crypto/fips_err.h b/crypto/fips_err.h
index b32861685819..c671691b4786 100644
--- a/crypto/fips_err.h
+++ b/crypto/fips_err.h
@@ -1,6 +1,6 @@
 /* crypto/fips_err.h */
 /* ====================================================================
- * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -71,53 +71,125 @@
 static ERR_STRING_DATA FIPS_str_functs[]=
 	{
 {ERR_FUNC(FIPS_F_DH_BUILTIN_GENPARAMS),	"DH_BUILTIN_GENPARAMS"},
+{ERR_FUNC(FIPS_F_DH_INIT),	"DH_INIT"},
+{ERR_FUNC(FIPS_F_DRBG_RESEED),	"DRBG_RESEED"},
 {ERR_FUNC(FIPS_F_DSA_BUILTIN_PARAMGEN),	"DSA_BUILTIN_PARAMGEN"},
+{ERR_FUNC(FIPS_F_DSA_BUILTIN_PARAMGEN2),	"DSA_BUILTIN_PARAMGEN2"},
 {ERR_FUNC(FIPS_F_DSA_DO_SIGN),	"DSA_do_sign"},
 {ERR_FUNC(FIPS_F_DSA_DO_VERIFY),	"DSA_do_verify"},
-{ERR_FUNC(FIPS_F_EVP_CIPHERINIT_EX),	"EVP_CipherInit_ex"},
-{ERR_FUNC(FIPS_F_EVP_DIGESTINIT_EX),	"EVP_DigestInit_ex"},
 {ERR_FUNC(FIPS_F_FIPS_CHECK_DSA),	"FIPS_CHECK_DSA"},
-{ERR_FUNC(FIPS_F_FIPS_CHECK_INCORE_FINGERPRINT),	"FIPS_CHECK_INCORE_FINGERPRINT"},
-{ERR_FUNC(FIPS_F_FIPS_CHECK_RSA),	"FIPS_CHECK_RSA"},
-{ERR_FUNC(FIPS_F_FIPS_DSA_CHECK),	"FIPS_DSA_CHECK"},
-{ERR_FUNC(FIPS_F_FIPS_MODE_SET),	"FIPS_mode_set"},
+{ERR_FUNC(FIPS_F_FIPS_CHECK_DSA_PRNG),	"fips_check_dsa_prng"},
+{ERR_FUNC(FIPS_F_FIPS_CHECK_EC),	"FIPS_CHECK_EC"},
+{ERR_FUNC(FIPS_F_FIPS_CHECK_EC_PRNG),	"fips_check_ec_prng"},
+{ERR_FUNC(FIPS_F_FIPS_CHECK_INCORE_FINGERPRINT),	"FIPS_check_incore_fingerprint"},
+{ERR_FUNC(FIPS_F_FIPS_CHECK_RSA),	"fips_check_rsa"},
+{ERR_FUNC(FIPS_F_FIPS_CHECK_RSA_PRNG),	"fips_check_rsa_prng"},
+{ERR_FUNC(FIPS_F_FIPS_CIPHER),	"FIPS_cipher"},
+{ERR_FUNC(FIPS_F_FIPS_CIPHERINIT),	"FIPS_cipherinit"},
+{ERR_FUNC(FIPS_F_FIPS_CIPHER_CTX_CTRL),	"FIPS_CIPHER_CTX_CTRL"},
+{ERR_FUNC(FIPS_F_FIPS_DIGESTFINAL),	"FIPS_digestfinal"},
+{ERR_FUNC(FIPS_F_FIPS_DIGESTINIT),	"FIPS_digestinit"},
+{ERR_FUNC(FIPS_F_FIPS_DIGESTUPDATE),	"FIPS_digestupdate"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_BYTES),	"FIPS_DRBG_BYTES"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_CHECK),	"FIPS_DRBG_CHECK"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_CPRNG_TEST),	"FIPS_DRBG_CPRNG_TEST"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_ERROR_CHECK),	"FIPS_DRBG_ERROR_CHECK"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_GENERATE),	"FIPS_drbg_generate"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_INIT),	"FIPS_drbg_init"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_INSTANTIATE),	"FIPS_drbg_instantiate"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_NEW),	"FIPS_drbg_new"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_RESEED),	"FIPS_drbg_reseed"},
+{ERR_FUNC(FIPS_F_FIPS_DRBG_SINGLE_KAT),	"FIPS_DRBG_SINGLE_KAT"},
+{ERR_FUNC(FIPS_F_FIPS_DSA_SIGN_DIGEST),	"FIPS_dsa_sign_digest"},
+{ERR_FUNC(FIPS_F_FIPS_DSA_VERIFY_DIGEST),	"FIPS_dsa_verify_digest"},
+{ERR_FUNC(FIPS_F_FIPS_GET_ENTROPY),	"FIPS_GET_ENTROPY"},
+{ERR_FUNC(FIPS_F_FIPS_MODULE_MODE_SET),	"FIPS_module_mode_set"},
 {ERR_FUNC(FIPS_F_FIPS_PKEY_SIGNATURE_TEST),	"fips_pkey_signature_test"},
+{ERR_FUNC(FIPS_F_FIPS_RAND_ADD),	"FIPS_rand_add"},
+{ERR_FUNC(FIPS_F_FIPS_RAND_BYTES),	"FIPS_rand_bytes"},
+{ERR_FUNC(FIPS_F_FIPS_RAND_PSEUDO_BYTES),	"FIPS_rand_pseudo_bytes"},
+{ERR_FUNC(FIPS_F_FIPS_RAND_SEED),	"FIPS_rand_seed"},
+{ERR_FUNC(FIPS_F_FIPS_RAND_SET_METHOD),	"FIPS_rand_set_method"},
+{ERR_FUNC(FIPS_F_FIPS_RAND_STATUS),	"FIPS_rand_status"},
+{ERR_FUNC(FIPS_F_FIPS_RSA_SIGN_DIGEST),	"FIPS_rsa_sign_digest"},
+{ERR_FUNC(FIPS_F_FIPS_RSA_VERIFY_DIGEST),	"FIPS_rsa_verify_digest"},
 {ERR_FUNC(FIPS_F_FIPS_SELFTEST_AES),	"FIPS_selftest_aes"},
+{ERR_FUNC(FIPS_F_FIPS_SELFTEST_AES_CCM),	"FIPS_selftest_aes_ccm"},
+{ERR_FUNC(FIPS_F_FIPS_SELFTEST_AES_GCM),	"FIPS_selftest_aes_gcm"},
+{ERR_FUNC(FIPS_F_FIPS_SELFTEST_AES_XTS),	"FIPS_selftest_aes_xts"},
+{ERR_FUNC(FIPS_F_FIPS_SELFTEST_CMAC),	"FIPS_selftest_cmac"},
 {ERR_FUNC(FIPS_F_FIPS_SELFTEST_DES),	"FIPS_selftest_des"},
 {ERR_FUNC(FIPS_F_FIPS_SELFTEST_DSA),	"FIPS_selftest_dsa"},
+{ERR_FUNC(FIPS_F_FIPS_SELFTEST_ECDSA),	"FIPS_selftest_ecdsa"},
 {ERR_FUNC(FIPS_F_FIPS_SELFTEST_HMAC),	"FIPS_selftest_hmac"},
-{ERR_FUNC(FIPS_F_FIPS_SELFTEST_RNG),	"FIPS_selftest_rng"},
 {ERR_FUNC(FIPS_F_FIPS_SELFTEST_SHA1),	"FIPS_selftest_sha1"},
+{ERR_FUNC(FIPS_F_FIPS_SELFTEST_X931),	"FIPS_selftest_x931"},
+{ERR_FUNC(FIPS_F_FIPS_SET_PRNG_KEY),	"FIPS_SET_PRNG_KEY"},
 {ERR_FUNC(FIPS_F_HASH_FINAL),	"HASH_FINAL"},
 {ERR_FUNC(FIPS_F_RSA_BUILTIN_KEYGEN),	"RSA_BUILTIN_KEYGEN"},
+{ERR_FUNC(FIPS_F_RSA_EAY_INIT),	"RSA_EAY_INIT"},
 {ERR_FUNC(FIPS_F_RSA_EAY_PRIVATE_DECRYPT),	"RSA_EAY_PRIVATE_DECRYPT"},
 {ERR_FUNC(FIPS_F_RSA_EAY_PRIVATE_ENCRYPT),	"RSA_EAY_PRIVATE_ENCRYPT"},
 {ERR_FUNC(FIPS_F_RSA_EAY_PUBLIC_DECRYPT),	"RSA_EAY_PUBLIC_DECRYPT"},
 {ERR_FUNC(FIPS_F_RSA_EAY_PUBLIC_ENCRYPT),	"RSA_EAY_PUBLIC_ENCRYPT"},
 {ERR_FUNC(FIPS_F_RSA_X931_GENERATE_KEY_EX),	"RSA_X931_generate_key_ex"},
-{ERR_FUNC(FIPS_F_SSLEAY_RAND_BYTES),	"SSLEAY_RAND_BYTES"},
 {0,NULL}
 	};
 
 static ERR_STRING_DATA FIPS_str_reasons[]=
 	{
-{ERR_REASON(FIPS_R_CANNOT_READ_EXE)      ,"cannot read exe"},
-{ERR_REASON(FIPS_R_CANNOT_READ_EXE_DIGEST),"cannot read exe digest"},
+{ERR_REASON(FIPS_R_ADDITIONAL_INPUT_ERROR_UNDETECTED),"additional input error undetected"},
+{ERR_REASON(FIPS_R_ADDITIONAL_INPUT_TOO_LONG),"additional input too long"},
+{ERR_REASON(FIPS_R_ALREADY_INSTANTIATED) ,"already instantiated"},
+{ERR_REASON(FIPS_R_AUTHENTICATION_FAILURE),"authentication failure"},
 {ERR_REASON(FIPS_R_CONTRADICTING_EVIDENCE),"contradicting evidence"},
-{ERR_REASON(FIPS_R_EXE_DIGEST_DOES_NOT_MATCH),"exe digest does not match"},
+{ERR_REASON(FIPS_R_DRBG_NOT_INITIALISED) ,"drbg not initialised"},
+{ERR_REASON(FIPS_R_DRBG_STUCK)           ,"drbg stuck"},
+{ERR_REASON(FIPS_R_ENTROPY_ERROR_UNDETECTED),"entropy error undetected"},
+{ERR_REASON(FIPS_R_ENTROPY_NOT_REQUESTED_FOR_RESEED),"entropy not requested for reseed"},
+{ERR_REASON(FIPS_R_ENTROPY_SOURCE_STUCK) ,"entropy source stuck"},
+{ERR_REASON(FIPS_R_ERROR_INITIALISING_DRBG),"error initialising drbg"},
+{ERR_REASON(FIPS_R_ERROR_INSTANTIATING_DRBG),"error instantiating drbg"},
+{ERR_REASON(FIPS_R_ERROR_RETRIEVING_ADDITIONAL_INPUT),"error retrieving additional input"},
+{ERR_REASON(FIPS_R_ERROR_RETRIEVING_ENTROPY),"error retrieving entropy"},
+{ERR_REASON(FIPS_R_ERROR_RETRIEVING_NONCE),"error retrieving nonce"},
 {ERR_REASON(FIPS_R_FINGERPRINT_DOES_NOT_MATCH),"fingerprint does not match"},
 {ERR_REASON(FIPS_R_FINGERPRINT_DOES_NOT_MATCH_NONPIC_RELOCATED),"fingerprint does not match nonpic relocated"},
 {ERR_REASON(FIPS_R_FINGERPRINT_DOES_NOT_MATCH_SEGMENT_ALIASING),"fingerprint does not match segment aliasing"},
 {ERR_REASON(FIPS_R_FIPS_MODE_ALREADY_SET),"fips mode already set"},
 {ERR_REASON(FIPS_R_FIPS_SELFTEST_FAILED) ,"fips selftest failed"},
+{ERR_REASON(FIPS_R_FUNCTION_ERROR)       ,"function error"},
+{ERR_REASON(FIPS_R_GENERATE_ERROR)       ,"generate error"},
+{ERR_REASON(FIPS_R_GENERATE_ERROR_UNDETECTED),"generate error undetected"},
+{ERR_REASON(FIPS_R_INSTANTIATE_ERROR)    ,"instantiate error"},
+{ERR_REASON(FIPS_R_INSUFFICIENT_SECURITY_STRENGTH),"insufficient security strength"},
+{ERR_REASON(FIPS_R_INTERNAL_ERROR)       ,"internal error"},
 {ERR_REASON(FIPS_R_INVALID_KEY_LENGTH)   ,"invalid key length"},
+{ERR_REASON(FIPS_R_INVALID_PARAMETERS)   ,"invalid parameters"},
+{ERR_REASON(FIPS_R_IN_ERROR_STATE)       ,"in error state"},
 {ERR_REASON(FIPS_R_KEY_TOO_SHORT)        ,"key too short"},
+{ERR_REASON(FIPS_R_NONCE_ERROR_UNDETECTED),"nonce error undetected"},
 {ERR_REASON(FIPS_R_NON_FIPS_METHOD)      ,"non fips method"},
+{ERR_REASON(FIPS_R_NOPR_TEST1_FAILURE)   ,"nopr test1 failure"},
+{ERR_REASON(FIPS_R_NOPR_TEST2_FAILURE)   ,"nopr test2 failure"},
+{ERR_REASON(FIPS_R_NOT_INSTANTIATED)     ,"not instantiated"},
 {ERR_REASON(FIPS_R_PAIRWISE_TEST_FAILED) ,"pairwise test failed"},
-{ERR_REASON(FIPS_R_RSA_DECRYPT_ERROR)    ,"rsa decrypt error"},
-{ERR_REASON(FIPS_R_RSA_ENCRYPT_ERROR)    ,"rsa encrypt error"},
+{ERR_REASON(FIPS_R_PERSONALISATION_ERROR_UNDETECTED),"personalisation error undetected"},
+{ERR_REASON(FIPS_R_PERSONALISATION_STRING_TOO_LONG),"personalisation string too long"},
+{ERR_REASON(FIPS_R_PRNG_STRENGTH_TOO_LOW),"prng strength too low"},
+{ERR_REASON(FIPS_R_PR_TEST1_FAILURE)     ,"pr test1 failure"},
+{ERR_REASON(FIPS_R_PR_TEST2_FAILURE)     ,"pr test2 failure"},
+{ERR_REASON(FIPS_R_REQUEST_LENGTH_ERROR_UNDETECTED),"request length error undetected"},
+{ERR_REASON(FIPS_R_REQUEST_TOO_LARGE_FOR_DRBG),"request too large for drbg"},
+{ERR_REASON(FIPS_R_RESEED_COUNTER_ERROR) ,"reseed counter error"},
+{ERR_REASON(FIPS_R_RESEED_ERROR)         ,"reseed error"},
 {ERR_REASON(FIPS_R_SELFTEST_FAILED)      ,"selftest failed"},
+{ERR_REASON(FIPS_R_SELFTEST_FAILURE)     ,"selftest failure"},
+{ERR_REASON(FIPS_R_STRENGTH_ERROR_UNDETECTED),"strength error undetected"},
 {ERR_REASON(FIPS_R_TEST_FAILURE)         ,"test failure"},
+{ERR_REASON(FIPS_R_UNINSTANTIATE_ERROR)  ,"uninstantiate error"},
+{ERR_REASON(FIPS_R_UNINSTANTIATE_ZEROISE_ERROR),"uninstantiate zeroise error"},
+{ERR_REASON(FIPS_R_UNSUPPORTED_DRBG_TYPE),"unsupported drbg type"},
 {ERR_REASON(FIPS_R_UNSUPPORTED_PLATFORM) ,"unsupported platform"},
 {0,NULL}
 	};
diff --git a/crypto/fips_err.c b/crypto/fips_ers.c
index 09f11748f609..09f11748f609 100644
--- a/crypto/fips_err.c
+++ b/crypto/fips_ers.c
diff --git a/crypto/hmac/Makefile b/crypto/hmac/Makefile
index 5cfa37d99c23..0e91709f64ca 100644
--- a/crypto/hmac/Makefile
+++ b/crypto/hmac/Makefile
@@ -17,8 +17,8 @@ TEST=hmactest.c
 APPS=
 
 LIB=$(TOP)/libcrypto.a
-LIBSRC=hmac.c
-LIBOBJ=hmac.o
+LIBSRC=hmac.c hm_ameth.c hm_pmeth.c
+LIBOBJ=hmac.o hm_ameth.o hm_pmeth.o
 
 SRC= $(LIBSRC)
 
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -74,13 +74,37 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
+hm_ameth.o: ../../e_os.h ../../include/openssl/asn1.h
+hm_ameth.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+hm_ameth.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+hm_ameth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+hm_ameth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+hm_ameth.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+hm_ameth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+hm_ameth.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+hm_ameth.o: ../../include/openssl/symhacks.h ../asn1/asn1_locl.h ../cryptlib.h
+hm_ameth.o: hm_ameth.c
+hm_pmeth.o: ../../e_os.h ../../include/openssl/asn1.h
+hm_pmeth.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+hm_pmeth.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+hm_pmeth.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+hm_pmeth.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+hm_pmeth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+hm_pmeth.o: ../../include/openssl/hmac.h ../../include/openssl/lhash.h
+hm_pmeth.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+hm_pmeth.o: ../../include/openssl/opensslconf.h
+hm_pmeth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+hm_pmeth.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+hm_pmeth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+hm_pmeth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+hm_pmeth.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+hm_pmeth.o: ../cryptlib.h ../evp/evp_locl.h hm_pmeth.c
 hmac.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 hmac.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 hmac.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-hmac.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-hmac.o: ../../include/openssl/hmac.h ../../include/openssl/lhash.h
-hmac.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-hmac.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-hmac.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-hmac.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-hmac.o: ../cryptlib.h hmac.c
+hmac.o: ../../include/openssl/evp.h ../../include/openssl/hmac.h
+hmac.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+hmac.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+hmac.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+hmac.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+hmac.o: ../../include/openssl/symhacks.h ../cryptlib.h hmac.c
diff --git a/crypto/hmac/hm_ameth.c b/crypto/hmac/hm_ameth.c
new file mode 100644
index 000000000000..e03f24aedab9
--- /dev/null
+++ b/crypto/hmac/hm_ameth.c
@@ -0,0 +1,167 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2007.
+ */
+/* ====================================================================
+ * Copyright (c) 2007 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/evp.h>
+#include "asn1_locl.h"
+
+#define HMAC_TEST_PRIVATE_KEY_FORMAT
+
+/* HMAC "ASN1" method. This is just here to indicate the
+ * maximum HMAC output length and to free up an HMAC
+ * key.
+ */
+
+static int hmac_size(const EVP_PKEY *pkey)
+	{
+	return EVP_MAX_MD_SIZE;
+	}
+
+static void hmac_key_free(EVP_PKEY *pkey)
+	{
+	ASN1_OCTET_STRING *os = (ASN1_OCTET_STRING *)pkey->pkey.ptr;
+	if (os)
+		{
+		if (os->data)
+			OPENSSL_cleanse(os->data, os->length);
+		ASN1_OCTET_STRING_free(os);
+		}
+	}
+
+
+static int hmac_pkey_ctrl(EVP_PKEY *pkey, int op, long arg1, void *arg2)
+	{
+	switch (op)
+		{
+		case ASN1_PKEY_CTRL_DEFAULT_MD_NID:
+		*(int *)arg2 = NID_sha1;
+		return 1;
+
+		default:
+		return -2;
+		}
+	}
+
+#ifdef HMAC_TEST_PRIVATE_KEY_FORMAT
+/* A bogus private key format for test purposes. This is simply the
+ * HMAC key with "HMAC PRIVATE KEY" in the headers. When enabled the
+ * genpkey utility can be used to "generate" HMAC keys.
+ */
+
+static int old_hmac_decode(EVP_PKEY *pkey,
+					const unsigned char **pder, int derlen)
+	{
+	ASN1_OCTET_STRING *os;
+	os = ASN1_OCTET_STRING_new();
+	if (!os || !ASN1_OCTET_STRING_set(os, *pder, derlen))
+		return 0;
+	EVP_PKEY_assign(pkey, EVP_PKEY_HMAC, os);
+	return 1;
+	}
+
+static int old_hmac_encode(const EVP_PKEY *pkey, unsigned char **pder)
+	{
+	int inc;
+	ASN1_OCTET_STRING *os = (ASN1_OCTET_STRING *)pkey->pkey.ptr;
+	if (pder)
+		{
+		if (!*pder)
+			{
+			*pder = OPENSSL_malloc(os->length);
+			inc = 0;
+			}
+		else inc = 1;
+
+		memcpy(*pder, os->data, os->length);
+
+		if (inc)
+			*pder += os->length;
+		}
+			
+	return os->length;
+	}
+
+#endif
+
+const EVP_PKEY_ASN1_METHOD hmac_asn1_meth = 
+	{
+	EVP_PKEY_HMAC,
+	EVP_PKEY_HMAC,
+	0,
+
+	"HMAC",
+	"OpenSSL HMAC method",
+
+	0,0,0,0,
+
+	0,0,0,
+
+	hmac_size,
+	0,
+	0,0,0,0,0,0,0,
+
+	hmac_key_free,
+	hmac_pkey_ctrl,
+#ifdef HMAC_TEST_PRIVATE_KEY_FORMAT
+	old_hmac_decode,
+	old_hmac_encode
+#else
+	0,0
+#endif
+	};
+
diff --git a/crypto/hmac/hm_pmeth.c b/crypto/hmac/hm_pmeth.c
new file mode 100644
index 000000000000..0daa44511d24
--- /dev/null
+++ b/crypto/hmac/hm_pmeth.c
@@ -0,0 +1,271 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2007.
+ */
+/* ====================================================================
+ * Copyright (c) 2007 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/x509.h>
+#include <openssl/x509v3.h>
+#include <openssl/evp.h>
+#include <openssl/hmac.h>
+#include "evp_locl.h"
+
+/* HMAC pkey context structure */
+
+typedef struct
+	{
+	const EVP_MD *md;	/* MD for HMAC use */
+	ASN1_OCTET_STRING ktmp; /* Temp storage for key */
+	HMAC_CTX ctx;
+	} HMAC_PKEY_CTX;
+
+static int pkey_hmac_init(EVP_PKEY_CTX *ctx)
+	{
+	HMAC_PKEY_CTX *hctx;
+	hctx = OPENSSL_malloc(sizeof(HMAC_PKEY_CTX));
+	if (!hctx)
+		return 0;
+	hctx->md = NULL;
+	hctx->ktmp.data = NULL;
+	hctx->ktmp.length = 0;
+	hctx->ktmp.flags = 0;
+	hctx->ktmp.type = V_ASN1_OCTET_STRING;
+	HMAC_CTX_init(&hctx->ctx);
+
+	ctx->data = hctx;
+	ctx->keygen_info_count = 0;
+
+	return 1;
+	}
+
+static int pkey_hmac_copy(EVP_PKEY_CTX *dst, EVP_PKEY_CTX *src)
+	{
+	HMAC_PKEY_CTX *sctx, *dctx;
+	if (!pkey_hmac_init(dst))
+		return 0;
+       	sctx = src->data;
+	dctx = dst->data;
+	dctx->md = sctx->md;
+	HMAC_CTX_init(&dctx->ctx);
+	if (!HMAC_CTX_copy(&dctx->ctx, &sctx->ctx))
+		return 0;
+	if (sctx->ktmp.data)
+		{
+		if (!ASN1_OCTET_STRING_set(&dctx->ktmp,
+					sctx->ktmp.data, sctx->ktmp.length))
+			return 0;
+		}
+	return 1;
+	}
+
+static void pkey_hmac_cleanup(EVP_PKEY_CTX *ctx)
+	{
+	HMAC_PKEY_CTX *hctx = ctx->data;
+	HMAC_CTX_cleanup(&hctx->ctx);
+	if (hctx->ktmp.data)
+		{
+		if (hctx->ktmp.length)
+			OPENSSL_cleanse(hctx->ktmp.data, hctx->ktmp.length);
+		OPENSSL_free(hctx->ktmp.data);
+		hctx->ktmp.data = NULL;
+		}
+	OPENSSL_free(hctx);
+	}
+
+static int pkey_hmac_keygen(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey)
+	{
+	ASN1_OCTET_STRING *hkey = NULL;
+	HMAC_PKEY_CTX *hctx = ctx->data;
+	if (!hctx->ktmp.data)
+		return 0;
+	hkey = ASN1_OCTET_STRING_dup(&hctx->ktmp);
+	if (!hkey)
+		return 0;
+	EVP_PKEY_assign(pkey, EVP_PKEY_HMAC, hkey);
+	
+	return 1;
+	}
+
+static int int_update(EVP_MD_CTX *ctx,const void *data,size_t count)
+	{
+	HMAC_PKEY_CTX *hctx = ctx->pctx->data;
+	if (!HMAC_Update(&hctx->ctx, data, count))
+		return 0;
+	return 1;
+	}
+
+static int hmac_signctx_init(EVP_PKEY_CTX *ctx, EVP_MD_CTX *mctx)
+	{
+	HMAC_PKEY_CTX *hctx = ctx->data;
+	HMAC_CTX_set_flags(&hctx->ctx, mctx->flags & ~EVP_MD_CTX_FLAG_NO_INIT);
+	EVP_MD_CTX_set_flags(mctx, EVP_MD_CTX_FLAG_NO_INIT);
+	mctx->update = int_update;
+	return 1;
+	}
+
+static int hmac_signctx(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					EVP_MD_CTX *mctx)
+	{
+	unsigned int hlen;
+	HMAC_PKEY_CTX *hctx = ctx->data;
+	int l = EVP_MD_CTX_size(mctx);
+
+	if (l < 0)
+		return 0;
+	*siglen = l;
+	if (!sig)
+		return 1;
+
+	if (!HMAC_Final(&hctx->ctx, sig, &hlen))
+		return 0;
+	*siglen = (size_t)hlen;
+	return 1;
+	}
+
+static int pkey_hmac_ctrl(EVP_PKEY_CTX *ctx, int type, int p1, void *p2)
+	{
+	HMAC_PKEY_CTX *hctx = ctx->data;
+	ASN1_OCTET_STRING *key;
+	switch (type)
+		{
+
+		case EVP_PKEY_CTRL_SET_MAC_KEY:
+		if ((!p2 && p1 > 0) || (p1 < -1))
+			return 0;
+		if (!ASN1_OCTET_STRING_set(&hctx->ktmp, p2, p1))
+			return 0;
+		break;
+
+		case EVP_PKEY_CTRL_MD:
+		hctx->md = p2;
+		break;
+
+		case EVP_PKEY_CTRL_DIGESTINIT:
+		key = (ASN1_OCTET_STRING *)ctx->pkey->pkey.ptr;
+		if (!HMAC_Init_ex(&hctx->ctx, key->data, key->length, hctx->md,
+				ctx->engine))
+			return 0;
+		break;
+
+		default:
+		return -2;
+
+		}
+	return 1;
+	}
+
+static int pkey_hmac_ctrl_str(EVP_PKEY_CTX *ctx,
+			const char *type, const char *value)
+	{
+	if (!value)
+		{
+		return 0;
+		}
+	if (!strcmp(type, "key"))
+		{
+		void *p = (void *)value;
+		return pkey_hmac_ctrl(ctx, EVP_PKEY_CTRL_SET_MAC_KEY,
+				-1, p);
+		}
+	if (!strcmp(type, "hexkey"))
+		{
+		unsigned char *key;
+		int r;
+		long keylen;
+		key = string_to_hex(value, &keylen);
+		if (!key)
+			return 0;
+		r = pkey_hmac_ctrl(ctx, EVP_PKEY_CTRL_SET_MAC_KEY, keylen, key);
+		OPENSSL_free(key);
+		return r;
+		}
+	return -2;
+	}
+
+const EVP_PKEY_METHOD hmac_pkey_meth = 
+	{
+	EVP_PKEY_HMAC,
+	0,
+	pkey_hmac_init,
+	pkey_hmac_copy,
+	pkey_hmac_cleanup,
+
+	0, 0,
+
+	0,
+	pkey_hmac_keygen,
+
+	0, 0,
+
+	0, 0,
+
+	0,0,
+
+	hmac_signctx_init,
+	hmac_signctx,
+
+	0,0,
+
+	0,0,
+
+	0,0,
+
+	0,0,
+
+	pkey_hmac_ctrl,
+	pkey_hmac_ctrl_str
+
+	};
diff --git a/crypto/hmac/hmac.c b/crypto/hmac/hmac.c
index 6899be63d4fb..ba27cbf56f2e 100644
--- a/crypto/hmac/hmac.c
+++ b/crypto/hmac/hmac.c
@@ -61,14 +61,34 @@
 #include "cryptlib.h"
 #include <openssl/hmac.h>
 
-#ifndef OPENSSL_FIPS
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 
-void HMAC_Init_ex(HMAC_CTX *ctx, const void *key, int len,
+int HMAC_Init_ex(HMAC_CTX *ctx, const void *key, int len,
 		  const EVP_MD *md, ENGINE *impl)
 	{
 	int i,j,reset=0;
 	unsigned char pad[HMAC_MAX_MD_CBLOCK];
 
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode())
+		{
+		/* If we have an ENGINE need to allow non FIPS */
+		if ((impl || ctx->i_ctx.engine)
+			&&  !(ctx->i_ctx.flags & EVP_CIPH_FLAG_NON_FIPS_ALLOW))
+			{
+			EVPerr(EVP_F_HMAC_INIT_EX, EVP_R_DISABLED_FOR_FIPS);
+			return 0;
+			}
+		/* Other algorithm blocking will be done in FIPS_cmac_init,
+		 * via FIPS_hmac_init_ex().
+		 */
+		if (!impl && !ctx->i_ctx.engine)
+			return FIPS_hmac_init_ex(ctx, key, len, md, NULL);
+		}
+#endif
+
 	if (md != NULL)
 		{
 		reset=1;
@@ -84,10 +104,13 @@ void HMAC_Init_ex(HMAC_CTX *ctx, const void *key, int len,
 		OPENSSL_assert(j <= (int)sizeof(ctx->key));
 		if (j < len)
 			{
-			EVP_DigestInit_ex(&ctx->md_ctx,md, impl);
-			EVP_DigestUpdate(&ctx->md_ctx,key,len);
-			EVP_DigestFinal_ex(&(ctx->md_ctx),ctx->key,
-				&ctx->key_length);
+			if (!EVP_DigestInit_ex(&ctx->md_ctx,md, impl))
+				goto err;
+			if (!EVP_DigestUpdate(&ctx->md_ctx,key,len))
+				goto err;
+			if (!EVP_DigestFinal_ex(&(ctx->md_ctx),ctx->key,
+				&ctx->key_length))
+				goto err;
 			}
 		else
 			{
@@ -104,39 +127,61 @@ void HMAC_Init_ex(HMAC_CTX *ctx, const void *key, int len,
 		{
 		for (i=0; i<HMAC_MAX_MD_CBLOCK; i++)
 			pad[i]=0x36^ctx->key[i];
-		EVP_DigestInit_ex(&ctx->i_ctx,md, impl);
-		EVP_DigestUpdate(&ctx->i_ctx,pad,EVP_MD_block_size(md));
+		if (!EVP_DigestInit_ex(&ctx->i_ctx,md, impl))
+			goto err;
+		if (!EVP_DigestUpdate(&ctx->i_ctx,pad,EVP_MD_block_size(md)))
+			goto err;
 
 		for (i=0; i<HMAC_MAX_MD_CBLOCK; i++)
 			pad[i]=0x5c^ctx->key[i];
-		EVP_DigestInit_ex(&ctx->o_ctx,md, impl);
-		EVP_DigestUpdate(&ctx->o_ctx,pad,EVP_MD_block_size(md));
+		if (!EVP_DigestInit_ex(&ctx->o_ctx,md, impl))
+			goto err;
+		if (!EVP_DigestUpdate(&ctx->o_ctx,pad,EVP_MD_block_size(md)))
+			goto err;
 		}
-	EVP_MD_CTX_copy_ex(&ctx->md_ctx,&ctx->i_ctx);
+	if (!EVP_MD_CTX_copy_ex(&ctx->md_ctx,&ctx->i_ctx))
+			goto err;
+	return 1;
+	err:
+	return 0;
 	}
 
-void HMAC_Init(HMAC_CTX *ctx, const void *key, int len,
-	       const EVP_MD *md)
+int HMAC_Init(HMAC_CTX *ctx, const void *key, int len, const EVP_MD *md)
 	{
 	if(key && md)
 	    HMAC_CTX_init(ctx);
-	HMAC_Init_ex(ctx,key,len,md, NULL);
+	return HMAC_Init_ex(ctx,key,len,md, NULL);
 	}
 
-void HMAC_Update(HMAC_CTX *ctx, const unsigned char *data, size_t len)
+int HMAC_Update(HMAC_CTX *ctx, const unsigned char *data, size_t len)
 	{
-	EVP_DigestUpdate(&ctx->md_ctx,data,len);
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !ctx->i_ctx.engine)
+		return FIPS_hmac_update(ctx, data, len);
+#endif
+	return EVP_DigestUpdate(&ctx->md_ctx,data,len);
 	}
 
-void HMAC_Final(HMAC_CTX *ctx, unsigned char *md, unsigned int *len)
+int HMAC_Final(HMAC_CTX *ctx, unsigned char *md, unsigned int *len)
 	{
 	unsigned int i;
 	unsigned char buf[EVP_MAX_MD_SIZE];
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !ctx->i_ctx.engine)
+		return FIPS_hmac_final(ctx, md, len);
+#endif
 
-	EVP_DigestFinal_ex(&ctx->md_ctx,buf,&i);
-	EVP_MD_CTX_copy_ex(&ctx->md_ctx,&ctx->o_ctx);
-	EVP_DigestUpdate(&ctx->md_ctx,buf,i);
-	EVP_DigestFinal_ex(&ctx->md_ctx,md,len);
+	if (!EVP_DigestFinal_ex(&ctx->md_ctx,buf,&i))
+		goto err;
+	if (!EVP_MD_CTX_copy_ex(&ctx->md_ctx,&ctx->o_ctx))
+		goto err;
+	if (!EVP_DigestUpdate(&ctx->md_ctx,buf,i))
+		goto err;
+	if (!EVP_DigestFinal_ex(&ctx->md_ctx,md,len))
+		goto err;
+	return 1;
+	err:
+	return 0;
 	}
 
 void HMAC_CTX_init(HMAC_CTX *ctx)
@@ -146,8 +191,31 @@ void HMAC_CTX_init(HMAC_CTX *ctx)
 	EVP_MD_CTX_init(&ctx->md_ctx);
 	}
 
+int HMAC_CTX_copy(HMAC_CTX *dctx, HMAC_CTX *sctx)
+	{
+	if (!EVP_MD_CTX_copy(&dctx->i_ctx, &sctx->i_ctx))
+		goto err;
+	if (!EVP_MD_CTX_copy(&dctx->o_ctx, &sctx->o_ctx))
+		goto err;
+	if (!EVP_MD_CTX_copy(&dctx->md_ctx, &sctx->md_ctx))
+		goto err;
+	memcpy(dctx->key, sctx->key, HMAC_MAX_MD_CBLOCK);
+	dctx->key_length = sctx->key_length;
+	dctx->md = sctx->md;
+	return 1;
+	err:
+	return 0;
+	}
+
 void HMAC_CTX_cleanup(HMAC_CTX *ctx)
 	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !ctx->i_ctx.engine)
+		{
+		FIPS_hmac_ctx_cleanup(ctx);
+		return;
+		}
+#endif
 	EVP_MD_CTX_cleanup(&ctx->i_ctx);
 	EVP_MD_CTX_cleanup(&ctx->o_ctx);
 	EVP_MD_CTX_cleanup(&ctx->md_ctx);
@@ -163,11 +231,16 @@ unsigned char *HMAC(const EVP_MD *evp_md, const void *key, int key_len,
 
 	if (md == NULL) md=m;
 	HMAC_CTX_init(&c);
-	HMAC_Init(&c,key,key_len,evp_md);
-	HMAC_Update(&c,d,n);
-	HMAC_Final(&c,md,md_len);
+	if (!HMAC_Init(&c,key,key_len,evp_md))
+		goto err;
+	if (!HMAC_Update(&c,d,n))
+		goto err;
+	if (!HMAC_Final(&c,md,md_len))
+		goto err;
 	HMAC_CTX_cleanup(&c);
-	return(md);
+	return md;
+	err:
+	return NULL;
 	}
 
 void HMAC_CTX_set_flags(HMAC_CTX *ctx, unsigned long flags)
@@ -176,5 +249,3 @@ void HMAC_CTX_set_flags(HMAC_CTX *ctx, unsigned long flags)
 	EVP_MD_CTX_set_flags(&ctx->o_ctx, flags);
 	EVP_MD_CTX_set_flags(&ctx->md_ctx, flags);
 	}
-
-#endif
diff --git a/crypto/hmac/hmac.h b/crypto/hmac/hmac.h
index fc38ffb52b3b..1be00221908d 100644
--- a/crypto/hmac/hmac.h
+++ b/crypto/hmac/hmac.h
@@ -90,15 +90,16 @@ void HMAC_CTX_cleanup(HMAC_CTX *ctx);
 
 #define HMAC_cleanup(ctx) HMAC_CTX_cleanup(ctx) /* deprecated */
 
-void HMAC_Init(HMAC_CTX *ctx, const void *key, int len,
+int HMAC_Init(HMAC_CTX *ctx, const void *key, int len,
 	       const EVP_MD *md); /* deprecated */
-void HMAC_Init_ex(HMAC_CTX *ctx, const void *key, int len,
+int HMAC_Init_ex(HMAC_CTX *ctx, const void *key, int len,
 		  const EVP_MD *md, ENGINE *impl);
-void HMAC_Update(HMAC_CTX *ctx, const unsigned char *data, size_t len);
-void HMAC_Final(HMAC_CTX *ctx, unsigned char *md, unsigned int *len);
+int HMAC_Update(HMAC_CTX *ctx, const unsigned char *data, size_t len);
+int HMAC_Final(HMAC_CTX *ctx, unsigned char *md, unsigned int *len);
 unsigned char *HMAC(const EVP_MD *evp_md, const void *key, int key_len,
 		    const unsigned char *d, size_t n, unsigned char *md,
 		    unsigned int *md_len);
+int HMAC_CTX_copy(HMAC_CTX *dctx, HMAC_CTX *sctx);
 
 void HMAC_CTX_set_flags(HMAC_CTX *ctx, unsigned long flags);
 
diff --git a/crypto/ia64cpuid.S b/crypto/ia64cpuid.S
index 04fbb3439eb7..7832b9b640ba 100644
--- a/crypto/ia64cpuid.S
+++ b/crypto/ia64cpuid.S
@@ -1,6 +1,13 @@
 // Works on all IA-64 platforms: Linux, HP-UX, Win64i...
 // On Win64i compile with ias.exe.
 .text
+
+.global	OPENSSL_cpuid_setup#
+.proc	OPENSSL_cpuid_setup#
+OPENSSL_cpuid_setup:
+{ .mib;	br.ret.sptk.many	b0		};;
+.endp	OPENSSL_cpuid_setup#
+
 .global	OPENSSL_rdtsc#
 .proc	OPENSSL_rdtsc#
 OPENSSL_rdtsc:
@@ -19,7 +26,7 @@ OPENSSL_atomic_add:
 { .mii;	mov		ar.ccv=r2
 	add		r8=r2,r33
 	mov		r3=r2		};;
-{ .mmi;	mf
+{ .mmi;	mf;;
 	cmpxchg4.acq	r2=[r32],r8,ar.ccv
 	nop.i		0		};;
 { .mib;	cmp.ne		p6,p0=r2,r3
@@ -119,3 +126,42 @@ OPENSSL_wipe_cpu:
 	mov		ar.lc=r3
 	br.ret.sptk	b0		};;
 .endp	OPENSSL_wipe_cpu#
+
+.global	OPENSSL_cleanse#
+.proc	OPENSSL_cleanse#
+OPENSSL_cleanse:
+{ .mib;	cmp.eq		p6,p0=0,r33	    // len==0
+#if defined(_HPUX_SOURCE) && !defined(_LP64)
+	addp4		r32=0,r32
+#endif
+(p6)	br.ret.spnt	b0		};;
+{ .mib;	and		r2=7,r32
+	cmp.leu		p6,p0=15,r33	    // len>=15
+(p6)	br.cond.dptk	.Lot		};;
+
+.Little:
+{ .mib;	st1		[r32]=r0,1
+	cmp.ltu		p6,p7=1,r33	}  // len>1
+{ .mbb;	add		r33=-1,r33	   // len--
+(p6)	br.cond.dptk	.Little
+(p7)	br.ret.sptk.many	b0	};;
+
+.Lot:
+{ .mib;	cmp.eq		p6,p0=0,r2
+(p6)	br.cond.dptk	.Laligned	};;
+{ .mmi;	st1		[r32]=r0,1;;
+	and		r2=7,r32	}
+{ .mib;	add		r33=-1,r33
+	br		.Lot		};;
+
+.Laligned:
+{ .mmi;	st8		[r32]=r0,8
+	and		r2=-8,r33	    // len&~7
+	add		r33=-8,r33	};; // len-=8
+{ .mib;	cmp.ltu		p6,p0=8,r2	    // ((len+8)&~7)>8
+(p6)	br.cond.dptk	.Laligned	};;
+
+{ .mbb;	cmp.eq		p6,p7=r0,r33
+(p7)	br.cond.dpnt	.Little
+(p6)	br.ret.sptk.many	b0	};;
+.endp	OPENSSL_cleanse#
diff --git a/crypto/idea/Makefile b/crypto/idea/Makefile
index 55c0d4dbff43..8af0acdad972 100644
--- a/crypto/idea/Makefile
+++ b/crypto/idea/Makefile
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -83,8 +83,7 @@ i_ecb.o: ../../include/openssl/opensslv.h i_ecb.c idea_lcl.h
 i_ofb64.o: ../../include/openssl/idea.h ../../include/openssl/opensslconf.h
 i_ofb64.o: i_ofb64.c idea_lcl.h
 i_skey.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
-i_skey.o: ../../include/openssl/fips.h ../../include/openssl/idea.h
-i_skey.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-i_skey.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-i_skey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-i_skey.o: i_skey.c idea_lcl.h
+i_skey.o: ../../include/openssl/idea.h ../../include/openssl/opensslconf.h
+i_skey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+i_skey.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+i_skey.o: ../../include/openssl/symhacks.h i_skey.c idea_lcl.h
diff --git a/crypto/idea/i_skey.c b/crypto/idea/i_skey.c
index fa75b1446613..afb830964df3 100644
--- a/crypto/idea/i_skey.c
+++ b/crypto/idea/i_skey.c
@@ -56,27 +56,18 @@
  * [including the GNU Public Licence.]
  */
 
-#include <openssl/idea.h>
 #include <openssl/crypto.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/idea.h>
 #include "idea_lcl.h"
 
 static IDEA_INT inverse(unsigned int xin);
-
-#ifdef OPENSSL_FIPS
 void idea_set_encrypt_key(const unsigned char *key, IDEA_KEY_SCHEDULE *ks)
+#ifdef OPENSSL_FIPS
 	{
-	if (FIPS_mode())
-		FIPS_BAD_ABORT(IDEA)
+	fips_cipher_abort(IDEA);
 	private_idea_set_encrypt_key(key, ks);
 	}
-void private_idea_set_encrypt_key(const unsigned char *key,
-						IDEA_KEY_SCHEDULE *ks)
-#else
-void idea_set_encrypt_key(const unsigned char *key, IDEA_KEY_SCHEDULE *ks)
+void private_idea_set_encrypt_key(const unsigned char *key, IDEA_KEY_SCHEDULE *ks)
 #endif
 	{
 	int i;
@@ -111,11 +102,10 @@ void idea_set_encrypt_key(const unsigned char *key, IDEA_KEY_SCHEDULE *ks)
 		}
 	}
 
-void idea_set_decrypt_key(const IDEA_KEY_SCHEDULE *ek, IDEA_KEY_SCHEDULE *dk)
+void idea_set_decrypt_key(IDEA_KEY_SCHEDULE *ek, IDEA_KEY_SCHEDULE *dk)
 	{
 	int r;
-	register IDEA_INT *tp,t;
-	const IDEA_INT *fp;
+	register IDEA_INT *fp,*tp,t;
 
 	tp= &(dk->data[0][0]);
 	fp= &(ek->data[8][0]);
diff --git a/crypto/idea/idea.h b/crypto/idea/idea.h
index a137d4cbce20..e9a1e7f1a5e6 100644
--- a/crypto/idea/idea.h
+++ b/crypto/idea/idea.h
@@ -87,7 +87,7 @@ void idea_ecb_encrypt(const unsigned char *in, unsigned char *out,
 void private_idea_set_encrypt_key(const unsigned char *key, IDEA_KEY_SCHEDULE *ks);
 #endif
 void idea_set_encrypt_key(const unsigned char *key, IDEA_KEY_SCHEDULE *ks);
-void idea_set_decrypt_key(const IDEA_KEY_SCHEDULE *ek, IDEA_KEY_SCHEDULE *dk);
+void idea_set_decrypt_key(IDEA_KEY_SCHEDULE *ek, IDEA_KEY_SCHEDULE *dk);
 void idea_cbc_encrypt(const unsigned char *in, unsigned char *out,
 	long length, IDEA_KEY_SCHEDULE *ks, unsigned char *iv,int enc);
 void idea_cfb64_encrypt(const unsigned char *in, unsigned char *out,
diff --git a/crypto/jpake/Makefile b/crypto/jpake/Makefile
index a4a1402f2ec8..110c49ce0b4a 100644
--- a/crypto/jpake/Makefile
+++ b/crypto/jpake/Makefile
@@ -16,7 +16,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
diff --git a/crypto/jpake/jpake.c b/crypto/jpake/jpake.c
index 9736f89854cc..8e4b633ccc53 100644
--- a/crypto/jpake/jpake.c
+++ b/crypto/jpake/jpake.c
@@ -4,7 +4,6 @@
 #include <openssl/sha.h>
 #include <openssl/err.h>
 #include <memory.h>
-#include <assert.h>
 
 /*
  * In the definition, (xa, xb, xc, xd) are Alice's (x1, x2, x3, x4) or
@@ -134,7 +133,7 @@ static void hashlength(SHA_CTX *sha, size_t l)
     {
     unsigned char b[2];
 
-    assert(l <= 0xffff);
+    OPENSSL_assert(l <= 0xffff);
     b[0] = l >> 8;
     b[1] = l&0xff;
     SHA1_Update(sha, b, 2);
@@ -172,7 +171,7 @@ static void zkp_hash(BIGNUM *h, const BIGNUM *zkpg, const JPAKE_STEP_PART *p,
     */
     SHA1_Init(&sha);
     hashbn(&sha, zkpg);
-    assert(!BN_is_zero(p->zkpx.gr));
+    OPENSSL_assert(!BN_is_zero(p->zkpx.gr));
     hashbn(&sha, p->zkpx.gr);
     hashbn(&sha, p->gx);
     hashstring(&sha, proof_name);
@@ -314,22 +313,21 @@ int JPAKE_STEP1_process(JPAKE_CTX *ctx, const JPAKE_STEP1 *received)
 	return 0;
 	}
 
-
-    /* verify their ZKP(xc) */
+   /* verify their ZKP(xc) */
     if(!verify_zkp(&received->p1, ctx->p.g, ctx))
 	{
 	JPAKEerr(JPAKE_F_JPAKE_STEP1_PROCESS, JPAKE_R_VERIFY_X3_FAILED);
 	return 0;
 	}
 
-    /* verify their ZKP(xd) */
+   /* verify their ZKP(xd) */
     if(!verify_zkp(&received->p2, ctx->p.g, ctx))
 	{
 	JPAKEerr(JPAKE_F_JPAKE_STEP1_PROCESS, JPAKE_R_VERIFY_X4_FAILED);
 	return 0;
 	}
 
-    /* g^xd != 1 */
+   /* g^xd != 1 */
     if(BN_is_one(received->p2.gx))
 	{
 	JPAKEerr(JPAKE_F_JPAKE_STEP1_PROCESS, JPAKE_R_G_TO_THE_X4_IS_ONE);
diff --git a/crypto/jpake/jpaketest.c b/crypto/jpake/jpaketest.c
index 792fc49eb466..eaba75ed8abe 100644
--- a/crypto/jpake/jpaketest.c
+++ b/crypto/jpake/jpaketest.c
@@ -182,7 +182,7 @@ int main(int argc, char **argv)
     BN_free(p);
 
     CRYPTO_cleanup_all_ex_data();
-    ERR_remove_state(0);
+    ERR_remove_thread_state(NULL);
     ERR_free_strings();
     CRYPTO_mem_leaks(bio_err);
 
diff --git a/crypto/krb5/Makefile b/crypto/krb5/Makefile
index 8efb9e8910aa..14077390d692 100644
--- a/crypto/krb5/Makefile
+++ b/crypto/krb5/Makefile
@@ -34,7 +34,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
diff --git a/crypto/lhash/Makefile b/crypto/lhash/Makefile
index 35f093297158..82bddac47450 100644
--- a/crypto/lhash/Makefile
+++ b/crypto/lhash/Makefile
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
diff --git a/crypto/lhash/lh_stats.c b/crypto/lhash/lh_stats.c
index 5aa7766aa61b..815615e338d9 100644
--- a/crypto/lhash/lh_stats.c
+++ b/crypto/lhash/lh_stats.c
@@ -139,7 +139,7 @@ void lh_node_usage_stats(LHASH *lh, FILE *out)
 #else
 
 #ifndef OPENSSL_NO_FP_API
-void lh_stats(const LHASH *lh, FILE *fp)
+void lh_stats(const _LHASH *lh, FILE *fp)
 	{
 	BIO *bp;
 
@@ -151,7 +151,7 @@ void lh_stats(const LHASH *lh, FILE *fp)
 end:;
 	}
 
-void lh_node_stats(const LHASH *lh, FILE *fp)
+void lh_node_stats(const _LHASH *lh, FILE *fp)
 	{
 	BIO *bp;
 
@@ -163,7 +163,7 @@ void lh_node_stats(const LHASH *lh, FILE *fp)
 end:;
 	}
 
-void lh_node_usage_stats(const LHASH *lh, FILE *fp)
+void lh_node_usage_stats(const _LHASH *lh, FILE *fp)
 	{
 	BIO *bp;
 
@@ -177,7 +177,7 @@ end:;
 
 #endif
 
-void lh_stats_bio(const LHASH *lh, BIO *out)
+void lh_stats_bio(const _LHASH *lh, BIO *out)
 	{
 	BIO_printf(out,"num_items             = %lu\n",lh->num_items);
 	BIO_printf(out,"num_nodes             = %u\n",lh->num_nodes);
@@ -205,7 +205,7 @@ void lh_stats_bio(const LHASH *lh, BIO *out)
 #endif
 	}
 
-void lh_node_stats_bio(const LHASH *lh, BIO *out)
+void lh_node_stats_bio(const _LHASH *lh, BIO *out)
 	{
 	LHASH_NODE *n;
 	unsigned int i,num;
@@ -218,7 +218,7 @@ void lh_node_stats_bio(const LHASH *lh, BIO *out)
 		}
 	}
 
-void lh_node_usage_stats_bio(const LHASH *lh, BIO *out)
+void lh_node_usage_stats_bio(const _LHASH *lh, BIO *out)
 	{
 	LHASH_NODE *n;
 	unsigned long num;
diff --git a/crypto/lhash/lhash.c b/crypto/lhash/lhash.c
index 0b41f8762159..47f748081bba 100644
--- a/crypto/lhash/lhash.c
+++ b/crypto/lhash/lhash.c
@@ -107,18 +107,18 @@ const char lh_version[]="lhash" OPENSSL_VERSION_PTEXT;
 #define UP_LOAD		(2*LH_LOAD_MULT) /* load times 256  (default 2) */
 #define DOWN_LOAD	(LH_LOAD_MULT)   /* load times 256  (default 1) */
 
-static void expand(LHASH *lh);
-static void contract(LHASH *lh);
-static LHASH_NODE **getrn(LHASH *lh, const void *data, unsigned long *rhash);
+static void expand(_LHASH *lh);
+static void contract(_LHASH *lh);
+static LHASH_NODE **getrn(_LHASH *lh, const void *data, unsigned long *rhash);
 
-LHASH *lh_new(LHASH_HASH_FN_TYPE h, LHASH_COMP_FN_TYPE c)
+_LHASH *lh_new(LHASH_HASH_FN_TYPE h, LHASH_COMP_FN_TYPE c)
 	{
-	LHASH *ret;
+	_LHASH *ret;
 	int i;
 
-	if ((ret=(LHASH *)OPENSSL_malloc(sizeof(LHASH))) == NULL)
+	if ((ret=OPENSSL_malloc(sizeof(_LHASH))) == NULL)
 		goto err0;
-	if ((ret->b=(LHASH_NODE **)OPENSSL_malloc(sizeof(LHASH_NODE *)*MIN_NODES)) == NULL)
+	if ((ret->b=OPENSSL_malloc(sizeof(LHASH_NODE *)*MIN_NODES)) == NULL)
 		goto err1;
 	for (i=0; i<MIN_NODES; i++)
 		ret->b[i]=NULL;
@@ -154,7 +154,7 @@ err0:
 	return(NULL);
 	}
 
-void lh_free(LHASH *lh)
+void lh_free(_LHASH *lh)
 	{
 	unsigned int i;
 	LHASH_NODE *n,*nn;
@@ -176,7 +176,7 @@ void lh_free(LHASH *lh)
 	OPENSSL_free(lh);
 	}
 
-void *lh_insert(LHASH *lh, void *data)
+void *lh_insert(_LHASH *lh, void *data)
 	{
 	unsigned long hash;
 	LHASH_NODE *nn,**rn;
@@ -214,7 +214,7 @@ void *lh_insert(LHASH *lh, void *data)
 	return(ret);
 	}
 
-void *lh_delete(LHASH *lh, const void *data)
+void *lh_delete(_LHASH *lh, const void *data)
 	{
 	unsigned long hash;
 	LHASH_NODE *nn,**rn;
@@ -245,7 +245,7 @@ void *lh_delete(LHASH *lh, const void *data)
 	return(ret);
 	}
 
-void *lh_retrieve(LHASH *lh, const void *data)
+void *lh_retrieve(_LHASH *lh, const void *data)
 	{
 	unsigned long hash;
 	LHASH_NODE **rn;
@@ -267,12 +267,15 @@ void *lh_retrieve(LHASH *lh, const void *data)
 	return(ret);
 	}
 
-static void doall_util_fn(LHASH *lh, int use_arg, LHASH_DOALL_FN_TYPE func,
+static void doall_util_fn(_LHASH *lh, int use_arg, LHASH_DOALL_FN_TYPE func,
 			  LHASH_DOALL_ARG_FN_TYPE func_arg, void *arg)
 	{
 	int i;
 	LHASH_NODE *a,*n;
 
+	if (lh == NULL)
+		return;
+
 	/* reverse the order so we search from 'top to bottom'
 	 * We were having memory leaks otherwise */
 	for (i=lh->num_nodes-1; i>=0; i--)
@@ -282,6 +285,8 @@ static void doall_util_fn(LHASH *lh, int use_arg, LHASH_DOALL_FN_TYPE func,
 			{
 			/* 28/05/91 - eay - n added so items can be deleted
 			 * via lh_doall */
+			/* 22/05/08 - ben - eh? since a is not passed,
+			 * this should not be needed */
 			n=a->next;
 			if(use_arg)
 				func_arg(a->data,arg);
@@ -292,53 +297,29 @@ static void doall_util_fn(LHASH *lh, int use_arg, LHASH_DOALL_FN_TYPE func,
 		}
 	}
 
-void lh_doall(LHASH *lh, LHASH_DOALL_FN_TYPE func)
+void lh_doall(_LHASH *lh, LHASH_DOALL_FN_TYPE func)
 	{
 	doall_util_fn(lh, 0, func, (LHASH_DOALL_ARG_FN_TYPE)0, NULL);
 	}
 
-void lh_doall_arg(LHASH *lh, LHASH_DOALL_ARG_FN_TYPE func, void *arg)
+void lh_doall_arg(_LHASH *lh, LHASH_DOALL_ARG_FN_TYPE func, void *arg)
 	{
 	doall_util_fn(lh, 1, (LHASH_DOALL_FN_TYPE)0, func, arg);
 	}
 
-static void expand(LHASH *lh)
+static void expand(_LHASH *lh)
 	{
 	LHASH_NODE **n,**n1,**n2,*np;
-	unsigned int p,i,j,pmax;
+	unsigned int p,i,j;
 	unsigned long hash,nni;
 
-	p=(int)lh->p++;
-	nni=lh->num_alloc_nodes;
-	pmax=lh->pmax;
-
-	if ((lh->p) >= lh->pmax)
-		{
-		j=(int)lh->num_alloc_nodes*2;
-		n=(LHASH_NODE **)OPENSSL_realloc(lh->b,
-			(int)sizeof(LHASH_NODE *)*j);
-		if (n == NULL)
-			{
-/*			fputs("realloc error in lhash",stderr); */
-			lh->error++;
-			lh->p=0;
-			return;
-			}
-		/* else */
-		for (i=(int)lh->num_alloc_nodes; i<j; i++)/* 26/02/92 eay */
-			n[i]=NULL;			  /* 02/03/92 eay */
-		lh->pmax=lh->num_alloc_nodes;
-		lh->num_alloc_nodes=j;
-		lh->num_expand_reallocs++;
-		lh->p=0;
-		lh->b=n;
-		}
-
 	lh->num_nodes++;
 	lh->num_expands++;
+	p=(int)lh->p++;
 	n1= &(lh->b[p]);
-	n2= &(lh->b[p+pmax]);
+	n2= &(lh->b[p+(int)lh->pmax]);
 	*n2=NULL;        /* 27/07/92 - eay - undefined pointer bug */
+	nni=lh->num_alloc_nodes;
 	
 	for (np= *n1; np != NULL; )
 		{
@@ -359,14 +340,35 @@ static void expand(LHASH *lh)
 		np= *n1;
 		}
 
+	if ((lh->p) >= lh->pmax)
+		{
+		j=(int)lh->num_alloc_nodes*2;
+		n=(LHASH_NODE **)OPENSSL_realloc(lh->b,
+			(int)(sizeof(LHASH_NODE *)*j));
+		if (n == NULL)
+			{
+/*			fputs("realloc error in lhash",stderr); */
+			lh->error++;
+			lh->p=0;
+			return;
+			}
+		/* else */
+		for (i=(int)lh->num_alloc_nodes; i<j; i++)/* 26/02/92 eay */
+			n[i]=NULL;			  /* 02/03/92 eay */
+		lh->pmax=lh->num_alloc_nodes;
+		lh->num_alloc_nodes=j;
+		lh->num_expand_reallocs++;
+		lh->p=0;
+		lh->b=n;
+		}
 	}
 
-static void contract(LHASH *lh)
+static void contract(_LHASH *lh)
 	{
 	LHASH_NODE **n,*n1,*np;
-	int idx = lh->p+lh->pmax-1;
 
-	np=lh->b[idx];
+	np=lh->b[lh->p+lh->pmax-1];
+	lh->b[lh->p+lh->pmax-1]=NULL; /* 24/07-92 - eay - weird but :-( */
 	if (lh->p == 0)
 		{
 		n=(LHASH_NODE **)OPENSSL_realloc(lh->b,
@@ -386,7 +388,6 @@ static void contract(LHASH *lh)
 	else
 		lh->p--;
 
-	lh->b[idx] = NULL;
 	lh->num_nodes--;
 	lh->num_contracts++;
 
@@ -401,7 +402,7 @@ static void contract(LHASH *lh)
 		}
 	}
 
-static LHASH_NODE **getrn(LHASH *lh, const void *data, unsigned long *rhash)
+static LHASH_NODE **getrn(_LHASH *lh, const void *data, unsigned long *rhash)
 	{
 	LHASH_NODE **ret,*n1;
 	unsigned long hash,nn;
@@ -468,7 +469,7 @@ unsigned long lh_strhash(const char *c)
 	return((ret>>16)^ret);
 	}
 
-unsigned long lh_num_items(const LHASH *lh)
+unsigned long lh_num_items(const _LHASH *lh)
 	{
 	return lh ? lh->num_items : 0;
 	}
diff --git a/crypto/lhash/lhash.h b/crypto/lhash/lhash.h
index d392d0cd80d7..e7d87635911a 100644
--- a/crypto/lhash/lhash.h
+++ b/crypto/lhash/lhash.h
@@ -98,42 +98,42 @@ typedef void (*LHASH_DOALL_ARG_FN_TYPE)(void *, void *);
  * macros if the functions are strictly internal. */
 
 /* First: "hash" functions */
-#define DECLARE_LHASH_HASH_FN(f_name,o_type) \
-	unsigned long f_name##_LHASH_HASH(const void *);
-#define IMPLEMENT_LHASH_HASH_FN(f_name,o_type) \
-	unsigned long f_name##_LHASH_HASH(const void *arg) { \
-		o_type a = (o_type)arg; \
-		return f_name(a); }
-#define LHASH_HASH_FN(f_name) f_name##_LHASH_HASH
+#define DECLARE_LHASH_HASH_FN(name, o_type) \
+	unsigned long name##_LHASH_HASH(const void *);
+#define IMPLEMENT_LHASH_HASH_FN(name, o_type) \
+	unsigned long name##_LHASH_HASH(const void *arg) { \
+		const o_type *a = arg; \
+		return name##_hash(a); }
+#define LHASH_HASH_FN(name) name##_LHASH_HASH
 
 /* Second: "compare" functions */
-#define DECLARE_LHASH_COMP_FN(f_name,o_type) \
-	int f_name##_LHASH_COMP(const void *, const void *);
-#define IMPLEMENT_LHASH_COMP_FN(f_name,o_type) \
-	int f_name##_LHASH_COMP(const void *arg1, const void *arg2) { \
-		o_type a = (o_type)arg1; \
-		o_type b = (o_type)arg2; \
-		return f_name(a,b); }
-#define LHASH_COMP_FN(f_name) f_name##_LHASH_COMP
+#define DECLARE_LHASH_COMP_FN(name, o_type) \
+	int name##_LHASH_COMP(const void *, const void *);
+#define IMPLEMENT_LHASH_COMP_FN(name, o_type) \
+	int name##_LHASH_COMP(const void *arg1, const void *arg2) { \
+		const o_type *a = arg1;		    \
+		const o_type *b = arg2; \
+		return name##_cmp(a,b); }
+#define LHASH_COMP_FN(name) name##_LHASH_COMP
 
 /* Third: "doall" functions */
-#define DECLARE_LHASH_DOALL_FN(f_name,o_type) \
-	void f_name##_LHASH_DOALL(void *);
-#define IMPLEMENT_LHASH_DOALL_FN(f_name,o_type) \
-	void f_name##_LHASH_DOALL(void *arg) { \
-		o_type a = (o_type)arg; \
-		f_name(a); }
-#define LHASH_DOALL_FN(f_name) f_name##_LHASH_DOALL
+#define DECLARE_LHASH_DOALL_FN(name, o_type) \
+	void name##_LHASH_DOALL(void *);
+#define IMPLEMENT_LHASH_DOALL_FN(name, o_type) \
+	void name##_LHASH_DOALL(void *arg) { \
+		o_type *a = arg; \
+		name##_doall(a); }
+#define LHASH_DOALL_FN(name) name##_LHASH_DOALL
 
 /* Fourth: "doall_arg" functions */
-#define DECLARE_LHASH_DOALL_ARG_FN(f_name,o_type,a_type) \
-	void f_name##_LHASH_DOALL_ARG(void *, void *);
-#define IMPLEMENT_LHASH_DOALL_ARG_FN(f_name,o_type,a_type) \
-	void f_name##_LHASH_DOALL_ARG(void *arg1, void *arg2) { \
-		o_type a = (o_type)arg1; \
-		a_type b = (a_type)arg2; \
-		f_name(a,b); }
-#define LHASH_DOALL_ARG_FN(f_name) f_name##_LHASH_DOALL_ARG
+#define DECLARE_LHASH_DOALL_ARG_FN(name, o_type, a_type) \
+	void name##_LHASH_DOALL_ARG(void *, void *);
+#define IMPLEMENT_LHASH_DOALL_ARG_FN(name, o_type, a_type) \
+	void name##_LHASH_DOALL_ARG(void *arg1, void *arg2) { \
+		o_type *a = arg1; \
+		a_type *b = arg2; \
+		name##_doall_arg(a, b); }
+#define LHASH_DOALL_ARG_FN(name) name##_LHASH_DOALL_ARG
 
 typedef struct lhash_st
 	{
@@ -163,7 +163,8 @@ typedef struct lhash_st
 	unsigned long num_hash_comps;
 
 	int error;
-	} LHASH;
+	} _LHASH;	/* Do not use _LHASH directly, use LHASH_OF
+			 * and friends */
 
 #define LH_LOAD_MULT	256
 
@@ -171,27 +172,67 @@ typedef struct lhash_st
  * in lh_insert(). */
 #define lh_error(lh)	((lh)->error)
 
-LHASH *lh_new(LHASH_HASH_FN_TYPE h, LHASH_COMP_FN_TYPE c);
-void lh_free(LHASH *lh);
-void *lh_insert(LHASH *lh, void *data);
-void *lh_delete(LHASH *lh, const void *data);
-void *lh_retrieve(LHASH *lh, const void *data);
-void lh_doall(LHASH *lh, LHASH_DOALL_FN_TYPE func);
-void lh_doall_arg(LHASH *lh, LHASH_DOALL_ARG_FN_TYPE func, void *arg);
+_LHASH *lh_new(LHASH_HASH_FN_TYPE h, LHASH_COMP_FN_TYPE c);
+void lh_free(_LHASH *lh);
+void *lh_insert(_LHASH *lh, void *data);
+void *lh_delete(_LHASH *lh, const void *data);
+void *lh_retrieve(_LHASH *lh, const void *data);
+void lh_doall(_LHASH *lh, LHASH_DOALL_FN_TYPE func);
+void lh_doall_arg(_LHASH *lh, LHASH_DOALL_ARG_FN_TYPE func, void *arg);
 unsigned long lh_strhash(const char *c);
-unsigned long lh_num_items(const LHASH *lh);
+unsigned long lh_num_items(const _LHASH *lh);
 
 #ifndef OPENSSL_NO_FP_API
-void lh_stats(const LHASH *lh, FILE *out);
-void lh_node_stats(const LHASH *lh, FILE *out);
-void lh_node_usage_stats(const LHASH *lh, FILE *out);
+void lh_stats(const _LHASH *lh, FILE *out);
+void lh_node_stats(const _LHASH *lh, FILE *out);
+void lh_node_usage_stats(const _LHASH *lh, FILE *out);
 #endif
 
 #ifndef OPENSSL_NO_BIO
-void lh_stats_bio(const LHASH *lh, BIO *out);
-void lh_node_stats_bio(const LHASH *lh, BIO *out);
-void lh_node_usage_stats_bio(const LHASH *lh, BIO *out);
+void lh_stats_bio(const _LHASH *lh, BIO *out);
+void lh_node_stats_bio(const _LHASH *lh, BIO *out);
+void lh_node_usage_stats_bio(const _LHASH *lh, BIO *out);
 #endif
+
+/* Type checking... */
+
+#define LHASH_OF(type) struct lhash_st_##type
+
+#define DECLARE_LHASH_OF(type) LHASH_OF(type) { int dummy; }
+
+#define CHECKED_LHASH_OF(type,lh) \
+  ((_LHASH *)CHECKED_PTR_OF(LHASH_OF(type),lh))
+
+/* Define wrapper functions. */
+#define LHM_lh_new(type, name) \
+  ((LHASH_OF(type) *)lh_new(LHASH_HASH_FN(name), LHASH_COMP_FN(name)))
+#define LHM_lh_error(type, lh) \
+  lh_error(CHECKED_LHASH_OF(type,lh))
+#define LHM_lh_insert(type, lh, inst) \
+  ((type *)lh_insert(CHECKED_LHASH_OF(type, lh), \
+		     CHECKED_PTR_OF(type, inst)))
+#define LHM_lh_retrieve(type, lh, inst) \
+  ((type *)lh_retrieve(CHECKED_LHASH_OF(type, lh), \
+		       CHECKED_PTR_OF(type, inst)))
+#define LHM_lh_delete(type, lh, inst) \
+  ((type *)lh_delete(CHECKED_LHASH_OF(type, lh),			\
+		     CHECKED_PTR_OF(type, inst)))
+#define LHM_lh_doall(type, lh,fn) lh_doall(CHECKED_LHASH_OF(type, lh), fn)
+#define LHM_lh_doall_arg(type, lh, fn, arg_type, arg) \
+  lh_doall_arg(CHECKED_LHASH_OF(type, lh), fn, CHECKED_PTR_OF(arg_type, arg))
+#define LHM_lh_num_items(type, lh) lh_num_items(CHECKED_LHASH_OF(type, lh))
+#define LHM_lh_down_load(type, lh) (CHECKED_LHASH_OF(type, lh)->down_load)
+#define LHM_lh_node_stats_bio(type, lh, out) \
+  lh_node_stats_bio(CHECKED_LHASH_OF(type, lh), out)
+#define LHM_lh_node_usage_stats_bio(type, lh, out) \
+  lh_node_usage_stats_bio(CHECKED_LHASH_OF(type, lh), out)
+#define LHM_lh_stats_bio(type, lh, out) \
+  lh_stats_bio(CHECKED_LHASH_OF(type, lh), out)
+#define LHM_lh_free(type, lh) lh_free(CHECKED_LHASH_OF(type, lh))
+
+DECLARE_LHASH_OF(OPENSSL_STRING);
+DECLARE_LHASH_OF(OPENSSL_CSTRING);
+
 #ifdef  __cplusplus
 }
 #endif
diff --git a/crypto/md2/Makefile b/crypto/md2/Makefile
index 7f43321ab2f6..17f878aeb7d4 100644
--- a/crypto/md2/Makefile
+++ b/crypto/md2/Makefile
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -74,9 +74,7 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
-md2_dgst.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
-md2_dgst.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-md2_dgst.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+md2_dgst.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 md2_dgst.o: ../../include/openssl/md2.h ../../include/openssl/opensslconf.h
 md2_dgst.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 md2_dgst.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
diff --git a/crypto/md2/md2_dgst.c b/crypto/md2/md2_dgst.c
index cc4eeaf7a7ad..bf89def73eb9 100644
--- a/crypto/md2/md2_dgst.c
+++ b/crypto/md2/md2_dgst.c
@@ -62,11 +62,6 @@
 #include <openssl/md2.h>
 #include <openssl/opensslv.h>
 #include <openssl/crypto.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
-#include <openssl/err.h>
 
 const char MD2_version[]="MD2" OPENSSL_VERSION_PTEXT;
 
@@ -78,7 +73,7 @@ const char MD2_version[]="MD2" OPENSSL_VERSION_PTEXT;
 static void md2_block(MD2_CTX *c, const unsigned char *d);
 /* The magic S table - I have converted it to hex since it is
  * basically just a random byte string. */
-static MD2_INT S[256]={
+static const MD2_INT S[256]={
 	0x29, 0x2E, 0x43, 0xC9, 0xA2, 0xD8, 0x7C, 0x01,
 	0x3D, 0x36, 0x54, 0xA1, 0xEC, 0xF0, 0x06, 0x13,
 	0x62, 0xA7, 0x05, 0xF3, 0xC0, 0xC7, 0x73, 0x8C,
@@ -121,7 +116,7 @@ const char *MD2_options(void)
 		return("md2(int)");
 	}
 
-FIPS_NON_FIPS_MD_Init(MD2)
+fips_md_init(MD2)
 	{
 	c->num=0;
 	memset(c->state,0,sizeof c->state);
diff --git a/crypto/md32_common.h b/crypto/md32_common.h
index e0deb78015a5..bb7381952a22 100644
--- a/crypto/md32_common.h
+++ b/crypto/md32_common.h
@@ -165,7 +165,7 @@
 				asm (			\
 				"roll %1,%0"		\
 				: "=r"(ret)		\
-				: "I"(n), "0"(a)	\
+				: "I"(n), "0"((unsigned int)(a))	\
 				: "cc");		\
 			   ret;				\
 			})
@@ -293,7 +293,7 @@ int HASH_UPDATE (HASH_CTX *c, const void *data_, size_t len)
 	 * Wei Dai <weidai@eskimo.com> for pointing it out. */
 	if (l < c->Nl) /* overflow */
 		c->Nh++;
-	c->Nh+=(len>>29);	/* might cause compiler warning on 16-bit */
+	c->Nh+=(HASH_LONG)(len>>29);	/* might cause compiler warning on 16-bit */
 	c->Nl=l;
 
 	n = c->num;
@@ -331,7 +331,7 @@ int HASH_UPDATE (HASH_CTX *c, const void *data_, size_t len)
 	if (len != 0)
 		{
 		p = (unsigned char *)c->data;
-		c->num = len;
+		c->num = (unsigned int)len;
 		memcpy (p,data,len);
 		}
 	return 1;
@@ -383,6 +383,7 @@ int HASH_FINAL (unsigned char *md, HASH_CTX *c)
 	}
 
 #ifndef MD32_REG_T
+#if defined(__alpha) || defined(__sparcv9) || defined(__mips)
 #define MD32_REG_T long
 /*
  * This comment was originaly written for MD5, which is why it
@@ -400,9 +401,15 @@ int HASH_FINAL (unsigned char *md, HASH_CTX *c)
  * Well, to be honest it should say that this *prevents* 
  * performance degradation.
  *				<appro@fy.chalmers.se>
- * Apparently there're LP64 compilers that generate better
- * code if A-D are declared int. Most notably GCC-x86_64
- * generates better code.
+ */
+#else
+/*
+ * Above is not absolute and there are LP64 compilers that
+ * generate better code if MD32_REG_T is defined int. The above
+ * pre-processor condition reflects the circumstances under which
+ * the conclusion was made and is subject to further extension.
  *				<appro@fy.chalmers.se>
  */
+#define MD32_REG_T int
+#endif
 #endif
diff --git a/crypto/md4/Makefile b/crypto/md4/Makefile
index 0bc489658514..e6f1e4478c04 100644
--- a/crypto/md4/Makefile
+++ b/crypto/md4/Makefile
@@ -34,7 +34,7 @@ top:
 all:    lib
 
 lib:    $(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -69,15 +69,14 @@ depend:
 dclean:
 	$(PERL) -pe 'if (/^# DO NOT DELETE THIS LINE/) {print; exit(0);}' $(MAKEFILE) >Makefile.new
 	mv -f Makefile.new $(MAKEFILE)
+	rm -f ../../include/openssl/$(EXHEADER) ../../test/$(TEST) ../../apps/$(APPS)
 
 clean:
 	rm -f asm/mx86unix.cpp *.o asm/*.o *.obj lib tags core .pure .nfs* *.old *.bak fluff
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
-md4_dgst.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
-md4_dgst.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-md4_dgst.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+md4_dgst.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 md4_dgst.o: ../../include/openssl/md4.h ../../include/openssl/opensslconf.h
 md4_dgst.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 md4_dgst.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
diff --git a/crypto/md4/md4.h b/crypto/md4/md4.h
index ba1fe4a6ee82..a55368a79095 100644
--- a/crypto/md4/md4.h
+++ b/crypto/md4/md4.h
@@ -77,7 +77,7 @@ extern "C" {
  * !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  */
 
-#if defined(OPENSSL_SYS_WIN16) || defined(__LP32__)
+#if defined(__LP32__)
 #define MD4_LONG unsigned long
 #elif defined(OPENSSL_SYS_CRAY) || defined(__ILP64__)
 #define MD4_LONG unsigned long
diff --git a/crypto/md4/md4_dgst.c b/crypto/md4/md4_dgst.c
index 0f5448601d8c..82c2cb2d98a9 100644
--- a/crypto/md4/md4_dgst.c
+++ b/crypto/md4/md4_dgst.c
@@ -57,13 +57,9 @@
  */
 
 #include <stdio.h>
-#include "md4_locl.h"
 #include <openssl/opensslv.h>
-#include <openssl/err.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/crypto.h>
+#include "md4_locl.h"
 
 const char MD4_version[]="MD4" OPENSSL_VERSION_PTEXT;
 
@@ -75,15 +71,13 @@ const char MD4_version[]="MD4" OPENSSL_VERSION_PTEXT;
 #define INIT_DATA_C (unsigned long)0x98badcfeL
 #define INIT_DATA_D (unsigned long)0x10325476L
 
-FIPS_NON_FIPS_MD_Init(MD4)
+fips_md_init(MD4)
 	{
+	memset (c,0,sizeof(*c));
 	c->A=INIT_DATA_A;
 	c->B=INIT_DATA_B;
 	c->C=INIT_DATA_C;
 	c->D=INIT_DATA_D;
-	c->Nl=0;
-	c->Nh=0;
-	c->num=0;
 	return 1;
 	}
 
diff --git a/crypto/md4/md4s.cpp b/crypto/md4/md4s.cpp
deleted file mode 100644
index c0ec97fc9f9b..000000000000
--- a/crypto/md4/md4s.cpp
+++ /dev/null
@@ -1,78 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/md4.h>
-
-extern "C" {
-void md4_block_x86(MD4_CTX *ctx, unsigned char *buffer,int num);
-}
-
-void main(int argc,char *argv[])
-	{
-	unsigned char buffer[64*256];
-	MD4_CTX ctx;
-	unsigned long s1,s2,e1,e2;
-	unsigned char k[16];
-	unsigned long data[2];
-	unsigned char iv[8];
-	int i,num=0,numm;
-	int j=0;
-
-	if (argc >= 2)
-		num=atoi(argv[1]);
-
-	if (num == 0) num=16;
-	if (num > 250) num=16;
-	numm=num+2;
-	num*=64;
-	numm*=64;
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<10; i++) /**/
-			{
-			md4_block_x86(&ctx,buffer,numm);
-			GetTSC(s1);
-			md4_block_x86(&ctx,buffer,numm);
-			GetTSC(e1);
-			GetTSC(s2);
-			md4_block_x86(&ctx,buffer,num);
-			GetTSC(e2);
-			md4_block_x86(&ctx,buffer,num);
-			}
-		printf("md4 (%d bytes) %d %d (%.2f)\n",num,
-			e1-s1,e2-s2,(double)((e1-s1)-(e2-s2))/2);
-		}
-	}
-
diff --git a/crypto/md5/Makefile b/crypto/md5/Makefile
index 3c450fcfc09f..b9e2ce9a386a 100644
--- a/crypto/md5/Makefile
+++ b/crypto/md5/Makefile
@@ -38,21 +38,19 @@ top:
 all:    lib
 
 lib:    $(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-# ELF
-mx86-elf.s: asm/md5-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) md5-586.pl elf $(CFLAGS) > ../$@)
-# COFF
-mx86-cof.s: asm/md5-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) md5-586.pl coff $(CFLAGS) > ../$@)
-# a.out
-mx86-out.s: asm/md5-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) md5-586.pl a.out $(CFLAGS) > ../$@)
+md5-586.s:	asm/md5-586.pl ../perlasm/x86asm.pl
+	$(PERL) asm/md5-586.pl $(PERLASM_SCHEME) $(CFLAGS) > $@
 
-md5-x86_64.s:	asm/md5-x86_64.pl;	$(PERL) asm/md5-x86_64.pl $@
+md5-x86_64.s:	asm/md5-x86_64.pl
+	$(PERL) asm/md5-x86_64.pl $(PERLASM_SCHEME) > $@
+
+md5-ia64.s: asm/md5-ia64.S
+	$(CC) $(CFLAGS) -E asm/md5-ia64.S | \
+	$(PERL) -ne 's/;\s+/;\n/g; print;' > $@
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
@@ -91,9 +89,7 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
-md5_dgst.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
-md5_dgst.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-md5_dgst.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+md5_dgst.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 md5_dgst.o: ../../include/openssl/md5.h ../../include/openssl/opensslconf.h
 md5_dgst.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 md5_dgst.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
diff --git a/crypto/md5/asm/md5-586.pl b/crypto/md5/asm/md5-586.pl
index 76ac235f7d03..6cb66bb49991 100644
--- a/crypto/md5/asm/md5-586.pl
+++ b/crypto/md5/asm/md5-586.pl
@@ -7,7 +7,8 @@
 
 $normal=0;
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],$0);
diff --git a/crypto/md5/asm/md5-ia64.S b/crypto/md5/asm/md5-ia64.S
new file mode 100644
index 000000000000..e7de08d46a2b
--- /dev/null
+++ b/crypto/md5/asm/md5-ia64.S
@@ -0,0 +1,992 @@
+/* Copyright (c) 2005 Hewlett-Packard Development Company, L.P.
+
+Permission is hereby granted, free of charge, to any person obtaining
+a copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice shall be
+included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
+LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.  */
+
+//	Common registers are assigned as follows:
+//
+//	COMMON
+//
+//	t0		Const Tbl Ptr	TPtr
+//	t1		Round Constant	TRound
+//	t4		Block residual	LenResid
+//	t5		Residual Data	DTmp
+//
+//	{in,out}0	Block 0 Cycle	RotateM0
+//	{in,out}1	Block Value 12	M12
+//	{in,out}2	Block Value 8	M8
+//	{in,out}3	Block Value 4	M4
+//	{in,out}4	Block Value 0	M0
+//	{in,out}5	Block 1 Cycle	RotateM1
+//	{in,out}6	Block Value 13	M13
+//	{in,out}7	Block Value 9	M9
+//	{in,out}8	Block Value 5	M5
+//	{in,out}9	Block Value 1	M1
+//	{in,out}10	Block 2 Cycle	RotateM2
+//	{in,out}11	Block Value 14	M14
+//	{in,out}12	Block Value 10	M10
+//	{in,out}13	Block Value 6	M6
+//	{in,out}14	Block Value 2	M2
+//	{in,out}15	Block 3 Cycle	RotateM3
+//	{in,out}16	Block Value 15	M15
+//	{in,out}17	Block Value 11	M11
+//	{in,out}18	Block Value 7	M7
+//	{in,out}19	Block Value 3	M3
+//	{in,out}20	Scratch			Z
+//	{in,out}21	Scratch			Y
+//	{in,out}22	Scratch			X
+//	{in,out}23	Scratch			W
+//	{in,out}24	Digest A		A
+//	{in,out}25	Digest B		B
+//	{in,out}26	Digest C		C
+//	{in,out}27	Digest D		D
+//	{in,out}28	Active Data Ptr	DPtr
+//	in28		Dummy Value		-
+//	out28		Dummy Value		-
+//	bt0			Coroutine Link	QUICK_RTN
+//
+///	These predicates are used for computing the padding block(s) and
+///	are shared between the driver and digest co-routines
+//
+//	pt0			Extra Pad Block	pExtra
+//	pt1			Load next word	pLoad
+//	pt2			Skip next word	pSkip
+//	pt3			Search for Pad	pNoPad
+//	pt4			Pad Word 0		pPad0
+//	pt5			Pad Word 1		pPad1
+//	pt6			Pad Word 2		pPad2
+//	pt7			Pad Word 3		pPad3
+
+#define	DTmp		r19
+#define	LenResid	r18
+#define	QUICK_RTN	b6
+#define	TPtr		r14
+#define	TRound		r15
+#define	pExtra		p6
+#define	pLoad		p7
+#define	pNoPad		p9
+#define	pPad0		p10
+#define	pPad1		p11
+#define	pPad2		p12
+#define	pPad3		p13
+#define	pSkip		p8
+
+#define	A_		out24
+#define	B_		out25
+#define	C_		out26
+#define	D_		out27
+#define	DPtr_		out28
+#define	M0_		out4
+#define	M1_		out9
+#define	M10_		out12
+#define	M11_		out17
+#define	M12_		out1
+#define	M13_		out6
+#define	M14_		out11
+#define	M15_		out16
+#define	M2_		out14
+#define	M3_		out19
+#define	M4_		out3
+#define	M5_		out8
+#define	M6_		out13
+#define	M7_		out18
+#define	M8_		out2
+#define	M9_		out7
+#define	RotateM0_	out0
+#define	RotateM1_	out5
+#define	RotateM2_	out10
+#define	RotateM3_	out15
+#define	W_		out23
+#define	X_		out22
+#define	Y_		out21
+#define	Z_		out20
+
+#define	A		in24
+#define	B		in25
+#define	C		in26
+#define	D		in27
+#define	DPtr		in28
+#define	M0		in4
+#define	M1		in9
+#define	M10		in12
+#define	M11		in17
+#define	M12		in1
+#define	M13		in6
+#define	M14		in11
+#define	M15		in16
+#define	M2		in14
+#define	M3		in19
+#define	M4		in3
+#define	M5		in8
+#define	M6		in13
+#define	M7		in18
+#define	M8		in2
+#define	M9		in7
+#define	RotateM0	in0
+#define	RotateM1	in5
+#define	RotateM2	in10
+#define	RotateM3	in15
+#define	W		in23
+#define	X		in22
+#define	Y		in21
+#define	Z		in20
+
+/* register stack configuration for md5_block_asm_data_order(): */
+#define	MD5_NINP	3
+#define	MD5_NLOC	0
+#define MD5_NOUT	29
+#define MD5_NROT	0
+
+/* register stack configuration for helpers: */
+#define	_NINPUTS	MD5_NOUT
+#define	_NLOCALS	0
+#define _NOUTPUT	0
+#define	_NROTATE	24	/* this must be <= _NINPUTS */
+
+#if defined(_HPUX_SOURCE) && !defined(_LP64)
+#define	ADDP	addp4
+#else
+#define	ADDP	add
+#endif
+
+#if defined(_HPUX_SOURCE) || defined(B_ENDIAN)
+#define HOST_IS_BIG_ENDIAN
+#endif
+
+//	Macros for getting the left and right portions of little-endian words
+
+#define	GETLW(dst, src, align)	dep.z dst = src, 32 - 8 * align, 8 * align
+#define	GETRW(dst, src, align)	extr.u dst = src, 8 * align, 32 - 8 * align
+
+//	MD5 driver
+//
+//		Reads an input block, then calls the digest block
+//		subroutine and adds the results to the accumulated
+//		digest.  It allocates 32 outs which the subroutine
+//		uses as it's inputs and rotating
+//		registers. Initializes the round constant pointer and
+//		takes care of saving/restoring ar.lc
+//
+///	INPUT
+//
+//	in0		Context Ptr		CtxPtr0
+//	in1		Input Data Ptr		DPtrIn
+//	in2		Integral Blocks		BlockCount
+//	rp		Return Address		-
+//
+///	CODE
+//
+//	v2		Input Align		InAlign
+//	t0		Shared w/digest		-
+//	t1		Shared w/digest		-
+//	t2		Shared w/digest		-
+//	t3		Shared w/digest		-
+//	t4		Shared w/digest		-
+//	t5		Shared w/digest		-
+//	t6		PFS Save		PFSSave
+//	t7		ar.lc Save		LCSave
+//	t8		Saved PR		PRSave
+//	t9		2nd CtxPtr		CtxPtr1
+//	t10		Table Base		CTable
+//	t11		Table[0]		CTable0
+//	t13		Accumulator A		AccumA
+//	t14		Accumulator B		AccumB
+//	t15		Accumulator C		AccumC
+//	t16		Accumulator D		AccumD
+//	pt0		Shared w/digest		-
+//	pt1		Shared w/digest		-
+//	pt2		Shared w/digest		-
+//	pt3		Shared w/digest		-
+//	pt4		Shared w/digest		-
+//	pt5		Shared w/digest		-
+//	pt6		Shared w/digest		-
+//	pt7		Shared w/digest		-
+//	pt8		Not Aligned		pOff
+//	pt8		Blocks Left		pAgain
+
+#define	AccumA		r27
+#define	AccumB		r28
+#define	AccumC		r29
+#define	AccumD		r30
+#define	CTable		r24
+#define	CTable0		r25
+#define	CtxPtr0		in0
+#define	CtxPtr1		r23
+#define	DPtrIn		in1
+#define	BlockCount	in2
+#define	InAlign		r10
+#define	LCSave		r21
+#define	PFSSave		r20
+#define	PRSave		r22
+#define	pAgain		p63
+#define	pOff		p63
+
+	.text
+
+/* md5_block_asm_data_order(MD5_CTX *c, const void *data, size_t num)
+
+     where:
+      c: a pointer to a structure of this type:
+
+	   typedef struct MD5state_st
+	     {
+	       MD5_LONG A,B,C,D;
+	       MD5_LONG Nl,Nh;
+	       MD5_LONG data[MD5_LBLOCK];
+	       unsigned int num;
+	     }
+	   MD5_CTX;
+
+      data: a pointer to the input data (may be misaligned)
+      num:  the number of 16-byte blocks to hash (i.e., the length
+            of DATA is 16*NUM.
+
+   */
+
+	.type	md5_block_asm_data_order, @function
+	.global	md5_block_asm_data_order
+	.align	32
+	.proc	md5_block_asm_data_order
+md5_block_asm_data_order:
+.md5_block:
+	.prologue
+{	.mmi
+	.save	ar.pfs, PFSSave
+	alloc	PFSSave = ar.pfs, MD5_NINP, MD5_NLOC, MD5_NOUT, MD5_NROT
+	ADDP	CtxPtr1 = 8, CtxPtr0
+	mov	CTable = ip
+}
+{	.mmi
+	ADDP	DPtrIn = 0, DPtrIn
+	ADDP	CtxPtr0 = 0, CtxPtr0
+	.save	ar.lc, LCSave
+	mov	LCSave = ar.lc
+}
+;;
+{	.mmi
+	add	CTable = .md5_tbl_data_order#-.md5_block#, CTable
+	and	InAlign = 0x3, DPtrIn
+}
+
+{	.mmi
+	ld4	AccumA = [CtxPtr0], 4
+	ld4	AccumC = [CtxPtr1], 4
+	.save pr, PRSave
+	mov	PRSave = pr
+	.body
+}
+;;
+{	.mmi
+	ld4	AccumB = [CtxPtr0]
+	ld4	AccumD = [CtxPtr1]
+	dep	DPtr_ = 0, DPtrIn, 0, 2
+} ;;
+#ifdef HOST_IS_BIG_ENDIAN
+	rum	psr.be;;	// switch to little-endian
+#endif
+{	.mmb
+	ld4	CTable0 = [CTable], 4
+	cmp.ne	pOff, p0 = 0, InAlign
+(pOff)	br.cond.spnt.many .md5_unaligned
+} ;;
+
+//	The FF load/compute loop rotates values three times, so that
+//	loading into M12 here produces the M0 value, M13 -> M1, etc.
+
+.md5_block_loop0:
+{	.mmi
+	ld4	M12_ = [DPtr_], 4
+	mov	TPtr = CTable
+	mov	TRound = CTable0
+} ;;
+{	.mmi
+	ld4	M13_ = [DPtr_], 4
+	mov	A_ = AccumA
+	mov	B_ = AccumB
+} ;;
+{	.mmi
+	ld4	M14_ = [DPtr_], 4
+	mov	C_ = AccumC
+	mov	D_ = AccumD
+} ;;
+{	.mmb
+	ld4	M15_ = [DPtr_], 4
+	add	BlockCount = -1, BlockCount
+	br.call.sptk.many QUICK_RTN = md5_digest_block0
+} ;;
+
+//	Now, we add the new digest values and do some clean-up
+//	before checking if there's another full block to process
+
+{	.mmi
+	add	AccumA = AccumA, A_
+	add	AccumB = AccumB, B_
+	cmp.ne	pAgain, p0 = 0, BlockCount
+}
+{	.mib
+	add	AccumC = AccumC, C_
+	add	AccumD = AccumD, D_
+(pAgain) br.cond.dptk.many .md5_block_loop0
+} ;;
+
+.md5_exit:
+#ifdef HOST_IS_BIG_ENDIAN
+	sum	psr.be;;	// switch back to big-endian mode
+#endif
+{	.mmi
+	st4	[CtxPtr0] = AccumB, -4
+	st4	[CtxPtr1] = AccumD, -4
+	mov	pr = PRSave, 0x1ffff ;;
+}
+{	.mmi
+	st4	[CtxPtr0] = AccumA
+	st4	[CtxPtr1] = AccumC
+	mov	ar.lc = LCSave
+} ;;
+{	.mib
+	mov	ar.pfs = PFSSave
+	br.ret.sptk.few	rp
+} ;;
+
+#define	MD5UNALIGNED(offset)						\
+.md5_process##offset:							\
+{	.mib ;								\
+	nop	0x0	;						\
+	GETRW(DTmp, DTmp, offset) ;					\
+} ;;									\
+.md5_block_loop##offset:						\
+{	.mmi ;								\
+	ld4	Y_ = [DPtr_], 4 ;					\
+	mov	TPtr = CTable ;						\
+	mov	TRound = CTable0 ;					\
+} ;;									\
+{	.mmi ;								\
+	ld4	M13_ = [DPtr_], 4 ;					\
+	mov	A_ = AccumA ;						\
+	mov	B_ = AccumB ;						\
+} ;;									\
+{	.mii ;								\
+	ld4	M14_ = [DPtr_], 4 ;					\
+	GETLW(W_, Y_, offset) ;						\
+	mov	C_ = AccumC ;						\
+}									\
+{	.mmi ;								\
+	mov	D_ = AccumD ;;						\
+	or	M12_ = W_, DTmp ;					\
+	GETRW(DTmp, Y_, offset) ;					\
+}									\
+{	.mib ;								\
+	ld4	M15_ = [DPtr_], 4 ;					\
+	add	BlockCount = -1, BlockCount ;				\
+	br.call.sptk.many QUICK_RTN = md5_digest_block##offset;		\
+} ;;									\
+{	.mmi ;								\
+	add	AccumA = AccumA, A_ ;					\
+	add	AccumB = AccumB, B_ ;					\
+	cmp.ne	pAgain, p0 = 0, BlockCount ;				\
+}									\
+{	.mib ;								\
+	add	AccumC = AccumC, C_ ;					\
+	add	AccumD = AccumD, D_ ;					\
+(pAgain) br.cond.dptk.many .md5_block_loop##offset ;			\
+} ;;									\
+{	.mib ;								\
+	nop	0x0 ;							\
+	nop	0x0 ;							\
+	br.cond.sptk.many .md5_exit ;					\
+} ;;
+
+	.align	32
+.md5_unaligned:
+//
+//	Because variable shifts are expensive, we special case each of
+//	the four alignements. In practice, this won't hurt too much
+//	since only one working set of code will be loaded.
+//
+{	.mib
+	ld4	DTmp = [DPtr_], 4
+	cmp.eq	pOff, p0 = 1, InAlign
+(pOff)	br.cond.dpnt.many .md5_process1
+} ;;
+{	.mib
+	cmp.eq	pOff, p0 = 2, InAlign
+	nop	0x0
+(pOff)	br.cond.dpnt.many .md5_process2
+} ;;
+	MD5UNALIGNED(3)
+	MD5UNALIGNED(1)
+	MD5UNALIGNED(2)
+
+	.endp md5_block_asm_data_order
+
+
+// MD5 Perform the F function and load
+//
+// Passed the first 4 words (M0 - M3) and initial (A, B, C, D) values,
+// computes the FF() round of functions, then branches to the common
+// digest code to finish up with GG(), HH, and II().
+//
+// INPUT
+//
+// rp Return Address -
+//
+// CODE
+//
+// v0 PFS bit bucket PFS
+// v1 Loop Trip Count LTrip
+// pt0 Load next word pMore
+
+/* For F round: */
+#define LTrip	r9
+#define PFS	r8
+#define pMore	p6
+
+/* For GHI rounds: */
+#define T	r9
+#define U	r10
+#define V	r11
+
+#define COMPUTE(a, b, s, M, R)			\
+{						\
+	.mii ;					\
+	ld4 TRound = [TPtr], 4 ;		\
+	dep.z Y = Z, 32, 32 ;;			\
+	shrp Z = Z, Y, 64 - s ;			\
+} ;;						\
+{						\
+	.mmi ;					\
+	add a = Z, b ;				\
+	mov R = M ;				\
+	nop 0x0 ;				\
+} ;;
+
+#define LOOP(a, b, s, M, R, label)		\
+{	.mii ;					\
+	ld4 TRound = [TPtr], 4 ;		\
+	dep.z Y = Z, 32, 32 ;;			\
+	shrp Z = Z, Y, 64 - s ;			\
+} ;;						\
+{	.mib ;					\
+	add a = Z, b ;				\
+	mov R = M ;				\
+	br.ctop.sptk.many label ;		\
+} ;;
+
+// G(B, C, D) = (B & D) | (C & ~D)
+
+#define G(a, b, c, d, M)			\
+{	.mmi ;					\
+	add Z = M, TRound ;			\
+	and Y = b, d ;				\
+	andcm X = c, d ;			\
+} ;;						\
+{	.mii ;					\
+	add Z = Z, a ;				\
+	or Y = Y, X ;;				\
+	add Z = Z, Y ;				\
+} ;;
+
+// H(B, C, D) = B ^ C ^ D
+
+#define H(a, b, c, d, M)			\
+{	.mmi ;					\
+	add Z = M, TRound ;			\
+	xor Y = b, c ;				\
+	nop 0x0 ;				\
+} ;;						\
+{	.mii ;					\
+	add Z = Z, a ;				\
+	xor Y = Y, d ;;				\
+	add Z = Z, Y ;				\
+} ;;
+
+// I(B, C, D) = C ^ (B | ~D)
+//
+// However, since we have an andcm operator, we use the fact that
+//
+// Y ^ Z == ~Y ^ ~Z
+//
+// to rewrite the expression as
+//
+// I(B, C, D) = ~C ^ (~B & D)
+
+#define I(a, b, c, d, M)			\
+{	.mmi ;					\
+	add Z = M, TRound ;			\
+	andcm Y = d, b ;			\
+	andcm X = -1, c ;			\
+} ;;						\
+{	.mii ;					\
+	add Z = Z, a ;				\
+	xor Y = Y, X ;;				\
+	add Z = Z, Y ;				\
+} ;;
+
+#define GG4(label)				\
+	G(A, B, C, D, M0)			\
+	COMPUTE(A, B, 5, M0, RotateM0)		\
+	G(D, A, B, C, M1)			\
+	COMPUTE(D, A, 9, M1, RotateM1)		\
+	G(C, D, A, B, M2)			\
+	COMPUTE(C, D, 14, M2, RotateM2)		\
+	G(B, C, D, A, M3)			\
+	LOOP(B, C, 20, M3, RotateM3, label)
+
+#define HH4(label)				\
+	H(A, B, C, D, M0)			\
+	COMPUTE(A, B, 4, M0, RotateM0)		\
+	H(D, A, B, C, M1)			\
+	COMPUTE(D, A, 11, M1, RotateM1)		\
+	H(C, D, A, B, M2)			\
+	COMPUTE(C, D, 16, M2, RotateM2)		\
+	H(B, C, D, A, M3)			\
+	LOOP(B, C, 23, M3, RotateM3, label)
+
+#define II4(label)				\
+	I(A, B, C, D, M0)			\
+	COMPUTE(A, B, 6, M0, RotateM0)		\
+	I(D, A, B, C, M1)			\
+	COMPUTE(D, A, 10, M1, RotateM1)		\
+	I(C, D, A, B, M2)			\
+	COMPUTE(C, D, 15, M2, RotateM2)		\
+	I(B, C, D, A, M3)			\
+	LOOP(B, C, 21, M3, RotateM3, label)
+
+#define FFLOAD(a, b, c, d, M, N, s)		\
+{	.mii ;					\
+(pMore) ld4 N = [DPtr], 4 ;			\
+	add Z = M, TRound ;			\
+	and Y = c, b ;				\
+}						\
+{	.mmi ;					\
+	andcm X = d, b ;;			\
+	add Z = Z, a ;				\
+	or Y = Y, X ;				\
+} ;;						\
+{	.mii ;					\
+	ld4 TRound = [TPtr], 4 ;		\
+	add Z = Z, Y ;;				\
+	dep.z Y = Z, 32, 32 ;			\
+} ;;						\
+{	.mii ;					\
+	nop 0x0 ;				\
+	shrp Z = Z, Y, 64 - s ;;		\
+	add a = Z, b ;				\
+} ;;
+
+#define FFLOOP(a, b, c, d, M, N, s, dest)	\
+{	.mii ;					\
+(pMore)	ld4 N = [DPtr], 4 ;			\
+	add Z = M, TRound ;			\
+	and Y = c, b ;				\
+}						\
+{	.mmi ;					\
+	andcm X = d, b ;;			\
+	add Z = Z, a ;				\
+	or Y = Y, X ;				\
+} ;;						\
+{	.mii ;					\
+	ld4 TRound = [TPtr], 4 ;		\
+	add Z = Z, Y ;;				\
+	dep.z Y = Z, 32, 32 ;			\
+} ;;						\
+{	.mii ;					\
+	nop 0x0 ;				\
+	shrp Z = Z, Y, 64 - s ;;		\
+	add a = Z, b ;				\
+}						\
+{	.mib ;					\
+	cmp.ne pMore, p0 = 0, LTrip ;		\
+	add LTrip = -1, LTrip ;			\
+	br.ctop.dptk.many dest ;		\
+} ;;
+
+	.type md5_digest_block0, @function
+	.align 32
+
+	.proc md5_digest_block0
+	.prologue
+md5_digest_block0:
+	.altrp QUICK_RTN
+	.body
+{	.mmi
+	alloc PFS = ar.pfs, _NINPUTS, _NLOCALS, _NOUTPUT, _NROTATE
+	mov LTrip = 2
+	mov ar.lc = 3
+} ;;
+{	.mii
+	cmp.eq pMore, p0 = r0, r0
+	mov ar.ec = 0
+	nop 0x0
+} ;;
+
+.md5_FF_round0:
+	FFLOAD(A, B, C, D, M12, RotateM0, 7)
+	FFLOAD(D, A, B, C, M13, RotateM1, 12)
+	FFLOAD(C, D, A, B, M14, RotateM2, 17)
+	FFLOOP(B, C, D, A, M15, RotateM3, 22, .md5_FF_round0)
+	//
+	// !!! Fall through to md5_digest_GHI
+	//
+	.endp md5_digest_block0
+
+	.type md5_digest_GHI, @function
+	.align 32
+
+	.proc md5_digest_GHI
+	.prologue
+	.regstk _NINPUTS, _NLOCALS, _NOUTPUT, _NROTATE
+md5_digest_GHI:
+	.altrp QUICK_RTN
+	.body
+//
+// The following sequence shuffles the block counstants round for the
+// next round:
+//
+// 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
+// 1 6 11 0 5 10 14 4 9 14 3 8 13 2 7 12
+//
+{	.mmi
+	mov Z = M0
+	mov Y = M15
+	mov ar.lc = 3
+}
+{	.mmi
+	mov X = M2
+	mov W = M9
+	mov V = M4
+} ;;
+
+{	.mmi
+	mov M0 = M1
+	mov M15 = M12
+	mov ar.ec = 1
+}
+{	.mmi
+	mov M2 = M11
+	mov M9 = M14
+	mov M4 = M5
+} ;;
+
+{	.mmi
+	mov M1 = M6
+	mov M12 = M13
+	mov U = M3
+}
+{	.mmi
+	mov M11 = M8
+	mov M14 = M7
+	mov M5 = M10
+} ;;
+
+{	.mmi
+	mov M6 = Y
+	mov M13 = X
+	mov M3 = Z
+}
+{	.mmi
+	mov M8 = W
+	mov M7 = V
+	mov M10 = U
+} ;;
+
+.md5_GG_round:
+	GG4(.md5_GG_round)
+
+// The following sequence shuffles the block constants round for the
+// next round:
+//
+// 1 6 11 0 5 10 14 4 9 14 3 8 13 2 7 12
+// 5 8 11 14 1 4 7 10 13 0 3 6 9 12 15 2
+
+{	.mmi
+	mov Z = M0
+	mov Y = M1
+	mov ar.lc = 3
+}
+{	.mmi
+	mov X = M3
+	mov W = M5
+	mov V = M6
+} ;;
+
+{	.mmi
+	mov M0 = M4
+	mov M1 = M11
+	mov ar.ec = 1
+}
+{	.mmi
+	mov M3 = M9
+	mov U = M8
+	mov T = M13
+} ;;
+
+{	.mmi
+	mov M4 = Z
+	mov M11 = Y
+	mov M5 = M7
+}
+{	.mmi
+	mov M6 = M14
+	mov M8 = M12
+	mov M13 = M15
+} ;;
+
+{	.mmi
+	mov M7 = W
+	mov M14 = V
+	nop 0x0
+}
+{	.mmi
+	mov M9 = X
+	mov M12 = U
+	mov M15 = T
+} ;;
+
+.md5_HH_round:
+	HH4(.md5_HH_round)
+
+// The following sequence shuffles the block constants round for the
+// next round:
+//
+// 5 8 11 14 1 4 7 10 13 0 3 6 9 12 15 2
+// 0 7 14 5 12 3 10 1 8 15 6 13 4 11 2 9
+
+{	.mmi
+	mov Z = M0
+	mov Y = M15
+	mov ar.lc = 3
+}
+{	.mmi
+	mov X = M10
+	mov W = M1
+	mov V = M4
+} ;;
+
+{	.mmi
+	mov M0 = M9
+	mov M15 = M12
+	mov ar.ec = 1
+}
+{	.mmi
+	mov M10 = M11
+	mov M1 = M6
+	mov M4 = M13
+} ;;
+
+{	.mmi
+	mov M9 = M14
+	mov M12 = M5
+	mov U = M3
+}
+{	.mmi
+	mov M11 = M8
+	mov M6 = M7
+	mov M13 = M2
+} ;;
+
+{	.mmi
+	mov M14 = Y
+	mov M5 = X
+	mov M3 = Z
+}
+{	.mmi
+	mov M8 = W
+	mov M7 = V
+	mov M2 = U
+} ;;
+
+.md5_II_round:
+	II4(.md5_II_round)
+
+{	.mib
+	nop 0x0
+	nop 0x0
+	br.ret.sptk.many QUICK_RTN
+} ;;
+
+	.endp md5_digest_GHI
+
+#define FFLOADU(a, b, c, d, M, P, N, s, offset)	\
+{	.mii ;					\
+(pMore) ld4 N = [DPtr], 4 ;			\
+	add Z = M, TRound ;			\
+	and Y = c, b ;				\
+}						\
+{	.mmi ;					\
+	andcm X = d, b ;;			\
+	add Z = Z, a ;				\
+	or Y = Y, X ;				\
+} ;;						\
+{	.mii ;					\
+	ld4 TRound = [TPtr], 4 ;		\
+	GETLW(W, P, offset) ;			\
+	add Z = Z, Y ;				\
+} ;;						\
+{	.mii ;					\
+	or W = W, DTmp ;			\
+	dep.z Y = Z, 32, 32 ;;			\
+	shrp Z = Z, Y, 64 - s ;			\
+} ;;						\
+{	.mii ;					\
+	add a = Z, b ;				\
+	GETRW(DTmp, P, offset) ;		\
+	mov P = W ;				\
+} ;;
+
+#define FFLOOPU(a, b, c, d, M, P, N, s, offset)		\
+{	.mii ;						\
+(pMore) ld4 N = [DPtr], 4 ;				\
+	add Z = M, TRound ;				\
+	and Y = c, b ;					\
+}							\
+{	.mmi ;						\
+	andcm X = d, b ;;				\
+	add Z = Z, a ;					\
+	or Y = Y, X ;					\
+} ;;							\
+{	.mii ;						\
+	ld4 TRound = [TPtr], 4 ;			\
+(pMore) GETLW(W, P, offset) 	;			\
+	add Z = Z, Y ;					\
+} ;;							\
+{	.mii ;						\
+(pMore) or W = W, DTmp ;				\
+	dep.z Y = Z, 32, 32 ;;				\
+	shrp Z = Z, Y, 64 - s ;				\
+} ;;							\
+{	.mii ;						\
+	add a = Z, b ;					\
+(pMore) GETRW(DTmp, P, offset) 	;			\
+(pMore) mov P = W ;					\
+}							\
+{	.mib ;						\
+	cmp.ne pMore, p0 = 0, LTrip ;			\
+	add LTrip = -1, LTrip ;				\
+	br.ctop.sptk.many .md5_FF_round##offset ;	\
+} ;;
+
+#define MD5FBLOCK(offset)						\
+	.type md5_digest_block##offset, @function ;			\
+									\
+	.align 32 ;							\
+	.proc md5_digest_block##offset ;				\
+	.prologue ;							\
+	.altrp QUICK_RTN ;						\
+	.body ;								\
+md5_digest_block##offset:						\
+{	.mmi ;								\
+	alloc PFS = ar.pfs, _NINPUTS, _NLOCALS, _NOUTPUT, _NROTATE ;	\
+	mov LTrip = 2 ;							\
+	mov ar.lc = 3 ;							\
+} ;;									\
+{	.mii ;								\
+	cmp.eq pMore, p0 = r0, r0 ;					\
+	mov ar.ec = 0 ;							\
+	nop 0x0 ;							\
+} ;;									\
+									\
+	.pred.rel "mutex", pLoad, pSkip ;				\
+.md5_FF_round##offset:							\
+	FFLOADU(A, B, C, D, M12, M13, RotateM0, 7, offset)		\
+	FFLOADU(D, A, B, C, M13, M14, RotateM1, 12, offset)		\
+	FFLOADU(C, D, A, B, M14, M15, RotateM2, 17, offset)		\
+	FFLOOPU(B, C, D, A, M15, RotateM0, RotateM3, 22, offset)	\
+									\
+{	.mib ;								\
+	nop 0x0 ;							\
+	nop 0x0 ;							\
+	br.cond.sptk.many md5_digest_GHI ;				\
+} ;;									\
+	.endp md5_digest_block##offset
+
+MD5FBLOCK(1)
+MD5FBLOCK(2)
+MD5FBLOCK(3)
+
+	.align 64
+	.type md5_constants, @object
+md5_constants:
+.md5_tbl_data_order:			// To ensure little-endian data
+					// order, code as bytes.
+	data1 0x78, 0xa4, 0x6a, 0xd7	//     0
+	data1 0x56, 0xb7, 0xc7, 0xe8	//     1
+	data1 0xdb, 0x70, 0x20, 0x24	//     2
+	data1 0xee, 0xce, 0xbd, 0xc1	//     3
+	data1 0xaf, 0x0f, 0x7c, 0xf5	//     4
+	data1 0x2a, 0xc6, 0x87, 0x47	//     5
+	data1 0x13, 0x46, 0x30, 0xa8	//     6
+	data1 0x01, 0x95, 0x46, 0xfd	//     7
+	data1 0xd8, 0x98, 0x80, 0x69	//     8
+	data1 0xaf, 0xf7, 0x44, 0x8b	//     9
+	data1 0xb1, 0x5b, 0xff, 0xff	//    10
+	data1 0xbe, 0xd7, 0x5c, 0x89	//    11
+	data1 0x22, 0x11, 0x90, 0x6b	//    12
+	data1 0x93, 0x71, 0x98, 0xfd	//    13
+	data1 0x8e, 0x43, 0x79, 0xa6	//    14
+	data1 0x21, 0x08, 0xb4, 0x49	//    15
+	data1 0x62, 0x25, 0x1e, 0xf6	//    16
+	data1 0x40, 0xb3, 0x40, 0xc0	//    17
+	data1 0x51, 0x5a, 0x5e, 0x26	//    18
+	data1 0xaa, 0xc7, 0xb6, 0xe9	//    19
+	data1 0x5d, 0x10, 0x2f, 0xd6	//    20
+	data1 0x53, 0x14, 0x44, 0x02	//    21
+	data1 0x81, 0xe6, 0xa1, 0xd8	//    22
+	data1 0xc8, 0xfb, 0xd3, 0xe7	//    23
+	data1 0xe6, 0xcd, 0xe1, 0x21	//    24
+	data1 0xd6, 0x07, 0x37, 0xc3	//    25
+	data1 0x87, 0x0d, 0xd5, 0xf4	//    26
+	data1 0xed, 0x14, 0x5a, 0x45	//    27
+	data1 0x05, 0xe9, 0xe3, 0xa9	//    28
+	data1 0xf8, 0xa3, 0xef, 0xfc	//    29
+	data1 0xd9, 0x02, 0x6f, 0x67	//    30
+	data1 0x8a, 0x4c, 0x2a, 0x8d	//    31
+	data1 0x42, 0x39, 0xfa, 0xff	//    32
+	data1 0x81, 0xf6, 0x71, 0x87	//    33
+	data1 0x22, 0x61, 0x9d, 0x6d	//    34
+	data1 0x0c, 0x38, 0xe5, 0xfd	//    35
+	data1 0x44, 0xea, 0xbe, 0xa4	//    36
+	data1 0xa9, 0xcf, 0xde, 0x4b	//    37
+	data1 0x60, 0x4b, 0xbb, 0xf6	//    38
+	data1 0x70, 0xbc, 0xbf, 0xbe	//    39
+	data1 0xc6, 0x7e, 0x9b, 0x28	//    40
+	data1 0xfa, 0x27, 0xa1, 0xea	//    41
+	data1 0x85, 0x30, 0xef, 0xd4	//    42
+	data1 0x05, 0x1d, 0x88, 0x04	//    43
+	data1 0x39, 0xd0, 0xd4, 0xd9	//    44
+	data1 0xe5, 0x99, 0xdb, 0xe6	//    45
+	data1 0xf8, 0x7c, 0xa2, 0x1f	//    46
+	data1 0x65, 0x56, 0xac, 0xc4	//    47
+	data1 0x44, 0x22, 0x29, 0xf4	//    48
+	data1 0x97, 0xff, 0x2a, 0x43	//    49
+	data1 0xa7, 0x23, 0x94, 0xab	//    50
+	data1 0x39, 0xa0, 0x93, 0xfc	//    51
+	data1 0xc3, 0x59, 0x5b, 0x65	//    52
+	data1 0x92, 0xcc, 0x0c, 0x8f	//    53
+	data1 0x7d, 0xf4, 0xef, 0xff	//    54
+	data1 0xd1, 0x5d, 0x84, 0x85	//    55
+	data1 0x4f, 0x7e, 0xa8, 0x6f	//    56
+	data1 0xe0, 0xe6, 0x2c, 0xfe	//    57
+	data1 0x14, 0x43, 0x01, 0xa3	//    58
+	data1 0xa1, 0x11, 0x08, 0x4e	//    59
+	data1 0x82, 0x7e, 0x53, 0xf7	//    60
+	data1 0x35, 0xf2, 0x3a, 0xbd	//    61
+	data1 0xbb, 0xd2, 0xd7, 0x2a	//    62
+	data1 0x91, 0xd3, 0x86, 0xeb	//    63
+.size	md5_constants#,64*4
diff --git a/crypto/md5/asm/md5-x86_64.pl b/crypto/md5/asm/md5-x86_64.pl
index 05d040f0b928..867885435e29 100755
--- a/crypto/md5/asm/md5-x86_64.pl
+++ b/crypto/md5/asm/md5-x86_64.pl
@@ -15,11 +15,10 @@ my $code;
 #   dst = x + ((dst + F(x,y,z) + X[k] + T_i) <<< s)
 #   %r10d = X[k_next]
 #   %r11d = z' (copy of z for the next step)
-# Each round1_step() takes about 5.71 clocks (9 instructions, 1.58 IPC)
+# Each round1_step() takes about 5.3 clocks (9 instructions, 1.7 IPC)
 sub round1_step
 {
     my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
-    $T_i = unpack("l",pack("l", hex($T_i))); # convert to 32-bit signed decimal
     $code .= " mov	0*4(%rsi),	%r10d		/* (NEXT STEP) X[0] */\n" if ($pos == -1);
     $code .= " mov	%edx,		%r11d		/* (NEXT STEP) z' = %edx */\n" if ($pos == -1);
     $code .= <<EOF;
@@ -38,23 +37,26 @@ EOF
 # round2_step() does:
 #   dst = x + ((dst + G(x,y,z) + X[k] + T_i) <<< s)
 #   %r10d = X[k_next]
-#   %r11d = y' (copy of y for the next step)
-# Each round2_step() takes about 6.22 clocks (9 instructions, 1.45 IPC)
+#   %r11d = z' (copy of z for the next step)
+#   %r12d = z' (copy of z for the next step)
+# Each round2_step() takes about 5.4 clocks (11 instructions, 2.0 IPC)
 sub round2_step
 {
     my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
-    $T_i = unpack("l",pack("l", hex($T_i))); # convert to 32-bit signed decimal
     $code .= " mov	1*4(%rsi),	%r10d		/* (NEXT STEP) X[1] */\n" if ($pos == -1);
-    $code .= " mov	%ecx,		%r11d		/* (NEXT STEP) y' = %ecx */\n" if ($pos == -1);
+    $code .= " mov	%edx,		%r11d		/* (NEXT STEP) z' = %edx */\n" if ($pos == -1);
+    $code .= " mov	%edx,		%r12d		/* (NEXT STEP) z' = %edx */\n" if ($pos == -1);
     $code .= <<EOF;
-	xor	$x,		%r11d		/* x ^ ... */
+	not	%r11d				/* not z */
 	lea	$T_i($dst,%r10d),$dst		/* Const + dst + ... */
-	and	$z,		%r11d		/* z & ... */
-	xor	$y,		%r11d		/* y ^ ... */
+	and	$x,		%r12d		/* x & z */
+	and	$y,		%r11d		/* y & (not z) */
 	mov	$k_next*4(%rsi),%r10d		/* (NEXT STEP) X[$k_next] */
-	add	%r11d,		$dst		/* dst += ... */
+	or	%r11d,		%r12d		/* (y & (not z)) | (x & z) */
+	mov	$y,		%r11d		/* (NEXT STEP) z' = $y */
+	add	%r12d,		$dst		/* dst += ... */
+	mov	$y,		%r12d		/* (NEXT STEP) z' = $y */
 	rol	\$$s,		$dst		/* dst <<< s */
-	mov	$x,		%r11d		/* (NEXT STEP) y' = $x */
 	add	$x,		$dst		/* dst += x */
 EOF
 }
@@ -63,11 +65,10 @@ EOF
 #   dst = x + ((dst + H(x,y,z) + X[k] + T_i) <<< s)
 #   %r10d = X[k_next]
 #   %r11d = y' (copy of y for the next step)
-# Each round3_step() takes about 4.26 clocks (8 instructions, 1.88 IPC)
+# Each round3_step() takes about 4.2 clocks (8 instructions, 1.9 IPC)
 sub round3_step
 {
     my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
-    $T_i = unpack("l",pack("l", hex($T_i))); # convert to 32-bit signed decimal
     $code .= " mov	5*4(%rsi),	%r10d		/* (NEXT STEP) X[5] */\n" if ($pos == -1);
     $code .= " mov	%ecx,		%r11d		/* (NEXT STEP) y' = %ecx */\n" if ($pos == -1);
     $code .= <<EOF;
@@ -86,11 +87,10 @@ EOF
 #   dst = x + ((dst + I(x,y,z) + X[k] + T_i) <<< s)
 #   %r10d = X[k_next]
 #   %r11d = not z' (copy of not z for the next step)
-# Each round4_step() takes about 5.27 clocks (9 instructions, 1.71 IPC)
+# Each round4_step() takes about 5.2 clocks (9 instructions, 1.7 IPC)
 sub round4_step
 {
     my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
-    $T_i = unpack("l",pack("l", hex($T_i))); # convert to 32-bit signed decimal
     $code .= " mov	0*4(%rsi),	%r10d		/* (NEXT STEP) X[0] */\n" if ($pos == -1);
     $code .= " mov	\$0xffffffff,	%r11d\n" if ($pos == -1);
     $code .= " xor	%edx,		%r11d		/* (NEXT STEP) not z' = not %edx*/\n"
@@ -108,8 +108,19 @@ sub round4_step
 EOF
 }
 
-my $output = shift;
-open STDOUT,"| $^X ../perlasm/x86_64-xlate.pl $output";
+my $flavour = shift;
+my $output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+my $win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; my $dir=$1; my $xlate;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+no warnings qw(uninitialized);
+open STDOUT,"| $^X $xlate $flavour $output";
 
 $code .= <<EOF;
 .text
@@ -120,8 +131,10 @@ $code .= <<EOF;
 md5_block_asm_data_order:
 	push	%rbp
 	push	%rbx
+	push	%r12
 	push	%r14
 	push	%r15
+.Lprologue:
 
 	# rdi = arg #1 (ctx, MD5_CTX pointer)
 	# rsi = arg #2 (ptr, data pointer)
@@ -236,13 +249,120 @@ $code .= <<EOF;
 	mov	%ecx,		2*4(%rbp)	# ctx->C = C
 	mov	%edx,		3*4(%rbp)	# ctx->D = D
 
+	mov	(%rsp),%r15
+	mov	8(%rsp),%r14
+	mov	16(%rsp),%r12
+	mov	24(%rsp),%rbx
+	mov	32(%rsp),%rbp
+	add	\$40,%rsp
+.Lepilogue:
+	ret
+.size md5_block_asm_data_order,.-md5_block_asm_data_order
+EOF
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+my $rec="%rcx";
+my $frame="%rdx";
+my $context="%r8";
+my $disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lprologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lprologue
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lepilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lepilogue
+	jae	.Lin_prologue
+
+	lea	40(%rax),%rax
+
+	mov	-8(%rax),%rbp
+	mov	-16(%rax),%rbx
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r14
+	mov	-40(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
 	pop	%r15
 	pop	%r14
-	pop	%rbx
+	pop	%r13
+	pop	%r12
 	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
 	ret
-.size md5_block_asm_data_order,.-md5_block_asm_data_order
-EOF
+.size	se_handler,.-se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_md5_block_asm_data_order
+	.rva	.LSEH_end_md5_block_asm_data_order
+	.rva	.LSEH_info_md5_block_asm_data_order
+
+.section	.xdata
+.align	8
+.LSEH_info_md5_block_asm_data_order:
+	.byte	9,0,0,0
+	.rva	se_handler
+___
+}
 
 print $code;
 
diff --git a/crypto/md5/md5.h b/crypto/md5/md5.h
index 0761f84a27bd..541cc925feb0 100644
--- a/crypto/md5/md5.h
+++ b/crypto/md5/md5.h
@@ -77,7 +77,7 @@ extern "C" {
  * !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  */
 
-#if defined(OPENSSL_SYS_WIN16) || defined(__LP32__)
+#if defined(__LP32__)
 #define MD5_LONG unsigned long
 #elif defined(OPENSSL_SYS_CRAY) || defined(__ILP64__)
 #define MD5_LONG unsigned long
diff --git a/crypto/md5/md5_dgst.c b/crypto/md5/md5_dgst.c
index 47bb9020ee1d..265890de52ba 100644
--- a/crypto/md5/md5_dgst.c
+++ b/crypto/md5/md5_dgst.c
@@ -59,11 +59,7 @@
 #include <stdio.h>
 #include "md5_locl.h"
 #include <openssl/opensslv.h>
-#include <openssl/err.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/crypto.h>
 
 const char MD5_version[]="MD5" OPENSSL_VERSION_PTEXT;
 
@@ -75,15 +71,13 @@ const char MD5_version[]="MD5" OPENSSL_VERSION_PTEXT;
 #define INIT_DATA_C (unsigned long)0x98badcfeL
 #define INIT_DATA_D (unsigned long)0x10325476L
 
-FIPS_NON_FIPS_MD_Init(MD5)
+fips_md_init(MD5)
 	{
+	memset (c,0,sizeof(*c));
 	c->A=INIT_DATA_A;
 	c->B=INIT_DATA_B;
 	c->C=INIT_DATA_C;
 	c->D=INIT_DATA_D;
-	c->Nl=0;
-	c->Nh=0;
-	c->num=0;
 	return 1;
 	}
 
diff --git a/crypto/md5/md5_locl.h b/crypto/md5/md5_locl.h
index 84e81b960de4..968d577995aa 100644
--- a/crypto/md5/md5_locl.h
+++ b/crypto/md5/md5_locl.h
@@ -69,6 +69,8 @@
 # if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__) || \
      defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64) || defined(_M_X64)
 #  define md5_block_data_order md5_block_asm_data_order
+# elif defined(__ia64) || defined(__ia64__) || defined(_M_IA64)
+#  define md5_block_data_order md5_block_asm_data_order
 # endif
 #endif
 
diff --git a/crypto/md5/md5s.cpp b/crypto/md5/md5s.cpp
deleted file mode 100644
index dd343fd4e6ed..000000000000
--- a/crypto/md5/md5s.cpp
+++ /dev/null
@@ -1,78 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/md5.h>
-
-extern "C" {
-void md5_block_x86(MD5_CTX *ctx, unsigned char *buffer,int num);
-}
-
-void main(int argc,char *argv[])
-	{
-	unsigned char buffer[64*256];
-	MD5_CTX ctx;
-	unsigned long s1,s2,e1,e2;
-	unsigned char k[16];
-	unsigned long data[2];
-	unsigned char iv[8];
-	int i,num=0,numm;
-	int j=0;
-
-	if (argc >= 2)
-		num=atoi(argv[1]);
-
-	if (num == 0) num=16;
-	if (num > 250) num=16;
-	numm=num+2;
-	num*=64;
-	numm*=64;
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<10; i++) /**/
-			{
-			md5_block_x86(&ctx,buffer,numm);
-			GetTSC(s1);
-			md5_block_x86(&ctx,buffer,numm);
-			GetTSC(e1);
-			GetTSC(s2);
-			md5_block_x86(&ctx,buffer,num);
-			GetTSC(e2);
-			md5_block_x86(&ctx,buffer,num);
-			}
-		printf("md5 (%d bytes) %d %d (%.2f)\n",num,
-			e1-s1,e2-s2,(double)((e1-s1)-(e2-s2))/2);
-		}
-	}
-
diff --git a/crypto/mdc2/Makefile b/crypto/mdc2/Makefile
index ea25688d888d..141553149d54 100644
--- a/crypto/mdc2/Makefile
+++ b/crypto/mdc2/Makefile
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -84,10 +84,10 @@ mdc2_one.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 mdc2_one.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 mdc2_one.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
 mdc2_one.o: ../../include/openssl/ui_compat.h ../cryptlib.h mdc2_one.c
-mdc2dgst.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
-mdc2dgst.o: ../../include/openssl/e_os2.h ../../include/openssl/mdc2.h
-mdc2dgst.o: ../../include/openssl/opensslconf.h
-mdc2dgst.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-mdc2dgst.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-mdc2dgst.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
-mdc2dgst.o: mdc2dgst.c
+mdc2dgst.o: ../../include/openssl/crypto.h ../../include/openssl/des.h
+mdc2dgst.o: ../../include/openssl/des_old.h ../../include/openssl/e_os2.h
+mdc2dgst.o: ../../include/openssl/mdc2.h ../../include/openssl/opensslconf.h
+mdc2dgst.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+mdc2dgst.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+mdc2dgst.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
+mdc2dgst.o: ../../include/openssl/ui_compat.h mdc2dgst.c
diff --git a/crypto/mdc2/mdc2.h b/crypto/mdc2/mdc2.h
index 7e1354116ad8..f3e8e579d237 100644
--- a/crypto/mdc2/mdc2.h
+++ b/crypto/mdc2/mdc2.h
@@ -80,6 +80,7 @@ typedef struct mdc2_ctx_st
 	int pad_type; /* either 1 or 2, default 1 */
 	} MDC2_CTX;
 
+
 #ifdef OPENSSL_FIPS
 int private_MDC2_Init(MDC2_CTX *c);
 #endif
diff --git a/crypto/mdc2/mdc2dgst.c b/crypto/mdc2/mdc2dgst.c
index a36b3f578ed1..b74bb1a75944 100644
--- a/crypto/mdc2/mdc2dgst.c
+++ b/crypto/mdc2/mdc2dgst.c
@@ -61,11 +61,7 @@
 #include <string.h>
 #include <openssl/des.h>
 #include <openssl/mdc2.h>
-#include <openssl/err.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/crypto.h>
 
 #undef c2l
 #define c2l(c,l)	(l =((DES_LONG)(*((c)++)))    , \
@@ -80,7 +76,7 @@
 			*((c)++)=(unsigned char)(((l)>>24L)&0xff))
 
 static void mdc2_body(MDC2_CTX *c, const unsigned char *in, size_t len);
-FIPS_NON_FIPS_MD_Init(MDC2)
+fips_md_init(MDC2)
 	{
 	c->num=0;
 	c->pad_type=1;
diff --git a/crypto/mem.c b/crypto/mem.c
index 05d7b9cd3d7a..21c001138083 100644
--- a/crypto/mem.c
+++ b/crypto/mem.c
@@ -101,7 +101,7 @@ static void (*free_locked_func)(void *)     = free;
 
 /* may be changed as long as 'allow_customize_debug' is set */
 /* XXX use correct function pointer types */
-#if defined(CRYPTO_MDEBUG) && !defined(OPENSSL_FIPS)
+#ifdef CRYPTO_MDEBUG
 /* use default functions from mem_dbg.c */
 static void (*malloc_debug_func)(void *,int,const char *,int,int)
 	= CRYPTO_dbg_malloc;
@@ -110,14 +110,6 @@ static void (*realloc_debug_func)(void *,void *,int,const char *,int,int)
 static void (*free_debug_func)(void *,int) = CRYPTO_dbg_free;
 static void (*set_debug_options_func)(long) = CRYPTO_dbg_set_options;
 static long (*get_debug_options_func)(void) = CRYPTO_dbg_get_options;
-
-static int  (*push_info_func)(const char *info, const char *file, int line)
-	= CRYPTO_dbg_push_info;
-static int  (*pop_info_func)(void)
-	= CRYPTO_dbg_pop_info;
-static int (*remove_all_info_func)(void)
-	= CRYPTO_dbg_remove_all_info;
-
 #else
 /* applications can use CRYPTO_malloc_debug_init() to select above case
  * at run-time */
@@ -127,19 +119,13 @@ static void (*realloc_debug_func)(void *,void *,int,const char *,int,int)
 static void (*free_debug_func)(void *,int) = NULL;
 static void (*set_debug_options_func)(long) = NULL;
 static long (*get_debug_options_func)(void) = NULL;
-
-
-static int  (*push_info_func)(const char *info, const char *file, int line)
-	= NULL;
-static int  (*pop_info_func)(void) = NULL;
-static int (*remove_all_info_func)(void) = NULL;
-
 #endif
 
 
 int CRYPTO_set_mem_functions(void *(*m)(size_t), void *(*r)(void *, size_t),
 	void (*f)(void *))
 	{
+	OPENSSL_init();
 	if (!allow_customize)
 		return 0;
 	if ((m == 0) || (r == 0) || (f == 0))
@@ -201,6 +187,7 @@ int CRYPTO_set_mem_debug_functions(void (*m)(void *,int,const char *,int,int),
 	{
 	if (!allow_customize_debug)
 		return 0;
+	OPENSSL_init();
 	malloc_debug_func=m;
 	realloc_debug_func=r;
 	free_debug_func=f;
@@ -209,15 +196,6 @@ int CRYPTO_set_mem_debug_functions(void (*m)(void *,int,const char *,int,int),
 	return 1;
 	}
 
-void CRYPTO_set_mem_info_functions(
-	int  (*push_info_fn)(const char *info, const char *file, int line),
-	int  (*pop_info_fn)(void),
-	int (*remove_all_info_fn)(void))
-	{
-	push_info_func = push_info_fn;
-	pop_info_func = pop_info_fn;
-	remove_all_info_func = remove_all_info_fn;
-	}
 
 void CRYPTO_get_mem_functions(void *(**m)(size_t), void *(**r)(void *, size_t),
 	void (**f)(void *))
@@ -274,7 +252,6 @@ void CRYPTO_get_mem_debug_functions(void (**m)(void *,int,const char *,int,int),
 void *CRYPTO_malloc_locked(int num, const char *file, int line)
 	{
 	void *ret = NULL;
-	extern unsigned char cleanse_ctr;
 
 	if (num <= 0) return NULL;
 
@@ -291,11 +268,15 @@ void *CRYPTO_malloc_locked(int num, const char *file, int line)
 	if (malloc_debug_func != NULL)
 		malloc_debug_func(ret, num, file, line, 1);
 
+#ifndef OPENSSL_CPUID_OBJ
         /* Create a dependency on the value of 'cleanse_ctr' so our memory
          * sanitisation function can't be optimised out. NB: We only do
          * this for >2Kb so the overhead doesn't bother us. */
         if(ret && (num > 2048))
+	{	extern unsigned char cleanse_ctr;
 		((unsigned char *)ret)[0] = cleanse_ctr;
+	}
+#endif
 
 	return ret;
 	}
@@ -315,7 +296,6 @@ void CRYPTO_free_locked(void *str)
 void *CRYPTO_malloc(int num, const char *file, int line)
 	{
 	void *ret = NULL;
-	extern unsigned char cleanse_ctr;
 
 	if (num <= 0) return NULL;
 
@@ -332,12 +312,23 @@ void *CRYPTO_malloc(int num, const char *file, int line)
 	if (malloc_debug_func != NULL)
 		malloc_debug_func(ret, num, file, line, 1);
 
+#ifndef OPENSSL_CPUID_OBJ
         /* Create a dependency on the value of 'cleanse_ctr' so our memory
          * sanitisation function can't be optimised out. NB: We only do
          * this for >2Kb so the overhead doesn't bother us. */
         if(ret && (num > 2048))
+	{	extern unsigned char cleanse_ctr;
                 ((unsigned char *)ret)[0] = cleanse_ctr;
+	}
+#endif
+
+	return ret;
+	}
+char *CRYPTO_strdup(const char *str, const char *file, int line)
+	{
+	char *ret = CRYPTO_malloc(strlen(str)+1, file, line);
 
+	strcpy(ret, str);
 	return ret;
 	}
 
@@ -427,24 +418,3 @@ long CRYPTO_get_mem_debug_options(void)
 		return get_debug_options_func();
 	return 0;
 	}
-
-int CRYPTO_push_info_(const char *info, const char *file, int line)
-	{
-	if (push_info_func)
-		return push_info_func(info, file, line);
-	return 1;
-	}
-
-int CRYPTO_pop_info(void)
-	{
-	if (pop_info_func)
-		return pop_info_func();
-	return 1;
-	}
-
-int CRYPTO_remove_all_info(void)
-	{
-	if (remove_all_info_func)
-		return remove_all_info_func();
-	return 1;
-	}
diff --git a/crypto/mem_dbg.c b/crypto/mem_dbg.c
index dfeb08479931..ac793397f10f 100644
--- a/crypto/mem_dbg.c
+++ b/crypto/mem_dbg.c
@@ -55,6 +55,59 @@
  * copied and put under another distribution licence
  * [including the GNU Public Licence.]
  */
+/* ====================================================================
+ * Copyright (c) 1998-2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
 
 #include <stdio.h>
 #include <stdlib.h>
@@ -81,8 +134,11 @@ static int mh_mode=CRYPTO_MEM_CHECK_OFF;
  */
 
 static unsigned long order = 0; /* number of memory requests */
-static LHASH *mh=NULL; /* hash-table of memory requests (address as key);
-                        * access requires MALLOC2 lock */
+
+DECLARE_LHASH_OF(MEM);
+static LHASH_OF(MEM) *mh=NULL; /* hash-table of memory requests
+				* (address as key); access requires
+				* MALLOC2 lock */
 
 
 typedef struct app_mem_info_st
@@ -93,8 +149,8 @@ typedef struct app_mem_info_st
  *   CRYPTO_pop_info()           to pop an entry,
  *   CRYPTO_remove_all_info()    to pop all entries.
  */
-	{	
-	unsigned long thread;
+	{
+	CRYPTO_THREADID threadid;
 	const char *file;
 	int line;
 	const char *info;
@@ -104,10 +160,13 @@ typedef struct app_mem_info_st
 
 static void app_info_free(APP_INFO *);
 
-static LHASH *amih=NULL; /* hash-table with those app_mem_info_st's
-                          * that are at the top of their thread's stack
-                          * (with `thread' as key);
-                          * access requires MALLOC2 lock */
+DECLARE_LHASH_OF(APP_INFO);
+static LHASH_OF(APP_INFO) *amih=NULL; /* hash-table with those
+				       * app_mem_info_st's that are at
+				       * the top of their thread's
+				       * stack (with `thread' as key);
+				       * access requires MALLOC2
+				       * lock */
 
 typedef struct mem_st
 /* memory-block description */
@@ -116,7 +175,7 @@ typedef struct mem_st
 	int num;
 	const char *file;
 	int line;
-	unsigned long thread;
+	CRYPTO_THREADID threadid;
 	unsigned long order;
 	time_t time;
 	APP_INFO *app_info;
@@ -136,11 +195,11 @@ static unsigned int num_disable = 0; /* num_disable > 0
                                       *     iff
                                       * mh_mode == CRYPTO_MEM_CHECK_ON (w/o ..._ENABLE)
                                       */
-static unsigned long disabling_thread = 0; /* Valid iff num_disable > 0.
-                                            * CRYPTO_LOCK_MALLOC2 is locked
-                                            * exactly in this case (by the
-                                            * thread named in disabling_thread).
-                                            */
+
+/* Valid iff num_disable > 0.  CRYPTO_LOCK_MALLOC2 is locked exactly in this
+ * case (by the thread named in disabling_thread).
+ */
+static CRYPTO_THREADID disabling_threadid;
 
 static void app_info_free(APP_INFO *inf)
 	{
@@ -177,7 +236,9 @@ int CRYPTO_mem_ctrl(int mode)
 	case CRYPTO_MEM_CHECK_DISABLE: /* aka MemCheck_off() */
 		if (mh_mode & CRYPTO_MEM_CHECK_ON)
 			{
-			if (!num_disable || (disabling_thread != CRYPTO_thread_id())) /* otherwise we already have the MALLOC2 lock */
+			CRYPTO_THREADID cur;
+			CRYPTO_THREADID_current(&cur);
+			if (!num_disable || CRYPTO_THREADID_cmp(&disabling_threadid, &cur)) /* otherwise we already have the MALLOC2 lock */
 				{
 				/* Long-time lock CRYPTO_LOCK_MALLOC2 must not be claimed while
 				 * we're holding CRYPTO_LOCK_MALLOC, or we'll deadlock if
@@ -195,7 +256,7 @@ int CRYPTO_mem_ctrl(int mode)
 				CRYPTO_w_lock(CRYPTO_LOCK_MALLOC2);
 				CRYPTO_w_lock(CRYPTO_LOCK_MALLOC);
 				mh_mode &= ~CRYPTO_MEM_CHECK_ENABLE;
-				disabling_thread=CRYPTO_thread_id();
+				CRYPTO_THREADID_cpy(&disabling_threadid, &cur);
 				}
 			num_disable++;
 			}
@@ -228,10 +289,12 @@ int CRYPTO_is_mem_check_on(void)
 
 	if (mh_mode & CRYPTO_MEM_CHECK_ON)
 		{
+		CRYPTO_THREADID cur;
+		CRYPTO_THREADID_current(&cur);
 		CRYPTO_r_lock(CRYPTO_LOCK_MALLOC);
 
 		ret = (mh_mode & CRYPTO_MEM_CHECK_ENABLE)
-			|| (disabling_thread != CRYPTO_thread_id());
+		        || CRYPTO_THREADID_cmp(&disabling_threadid, &cur);
 
 		CRYPTO_r_unlock(CRYPTO_LOCK_MALLOC);
 		}
@@ -249,49 +312,49 @@ long CRYPTO_dbg_get_options(void)
 	return options;
 	}
 
-/* static int mem_cmp(MEM *a, MEM *b) */
-static int mem_cmp(const void *a_void, const void *b_void)
+static int mem_cmp(const MEM *a, const MEM *b)
 	{
 #ifdef _WIN64
-	const char *a=(const char *)((const MEM *)a_void)->addr,
-		   *b=(const char *)((const MEM *)b_void)->addr;
-	if (a==b)	return 0;
-	else if (a>b)	return 1;
+	const char *ap=(const char *)a->addr,
+		   *bp=(const char *)b->addr;
+	if (ap==bp)	return 0;
+	else if (ap>bp)	return 1;
 	else		return -1;
 #else
-	return((const char *)((const MEM *)a_void)->addr
-		- (const char *)((const MEM *)b_void)->addr);
+	return (const char *)a->addr - (const char *)b->addr;
 #endif
 	}
+static IMPLEMENT_LHASH_COMP_FN(mem, MEM)
 
-/* static unsigned long mem_hash(MEM *a) */
-static unsigned long mem_hash(const void *a_void)
+static unsigned long mem_hash(const MEM *a)
 	{
 	unsigned long ret;
 
-	ret=(unsigned long)((const MEM *)a_void)->addr;
+	ret=(unsigned long)a->addr;
 
 	ret=ret*17851+(ret>>14)*7+(ret>>4)*251;
 	return(ret);
 	}
+static IMPLEMENT_LHASH_HASH_FN(mem, MEM)
 
 /* static int app_info_cmp(APP_INFO *a, APP_INFO *b) */
 static int app_info_cmp(const void *a_void, const void *b_void)
 	{
-	return(((const APP_INFO *)a_void)->thread
-		!= ((const APP_INFO *)b_void)->thread);
+	return CRYPTO_THREADID_cmp(&((const APP_INFO *)a_void)->threadid,
+				&((const APP_INFO *)b_void)->threadid);
 	}
+static IMPLEMENT_LHASH_COMP_FN(app_info, APP_INFO)
 
-/* static unsigned long app_info_hash(APP_INFO *a) */
-static unsigned long app_info_hash(const void *a_void)
+static unsigned long app_info_hash(const APP_INFO *a)
 	{
 	unsigned long ret;
 
-	ret=(unsigned long)((const APP_INFO *)a_void)->thread;
-
+	ret = CRYPTO_THREADID_hash(&a->threadid);
+	/* This is left in as a "who am I to question legacy?" measure */
 	ret=ret*17851+(ret>>14)*7+(ret>>4)*251;
 	return(ret);
 	}
+static IMPLEMENT_LHASH_HASH_FN(app_info, APP_INFO)
 
 static APP_INFO *pop_info(void)
 	{
@@ -300,21 +363,22 @@ static APP_INFO *pop_info(void)
 
 	if (amih != NULL)
 		{
-		tmp.thread=CRYPTO_thread_id();
-		if ((ret=(APP_INFO *)lh_delete(amih,&tmp)) != NULL)
+		CRYPTO_THREADID_current(&tmp.threadid);
+		if ((ret=lh_APP_INFO_delete(amih,&tmp)) != NULL)
 			{
 			APP_INFO *next=ret->next;
 
 			if (next != NULL)
 				{
 				next->references++;
-				lh_insert(amih,(char *)next);
+				(void)lh_APP_INFO_insert(amih,next);
 				}
 #ifdef LEVITTE_DEBUG_MEM
-			if (ret->thread != tmp.thread)
+			if (CRYPTO_THREADID_cmp(&ret->threadid, &tmp.threadid))
 				{
 				fprintf(stderr, "pop_info(): deleted info has other thread ID (%lu) than the current thread (%lu)!!!!\n",
-					ret->thread, tmp.thread);
+					CRYPTO_THREADID_hash(&ret->threadid),
+					CRYPTO_THREADID_hash(&tmp.threadid));
 				abort();
 				}
 #endif
@@ -330,7 +394,7 @@ static APP_INFO *pop_info(void)
 	return(ret);
 	}
 
-int CRYPTO_dbg_push_info(const char *info, const char *file, int line)
+int CRYPTO_push_info_(const char *info, const char *file, int line)
 	{
 	APP_INFO *ami, *amim;
 	int ret=0;
@@ -346,7 +410,7 @@ int CRYPTO_dbg_push_info(const char *info, const char *file, int line)
 			}
 		if (amih == NULL)
 			{
-			if ((amih=lh_new(app_info_hash, app_info_cmp)) == NULL)
+			if ((amih=lh_APP_INFO_new()) == NULL)
 				{
 				OPENSSL_free(ami);
 				ret=0;
@@ -354,20 +418,21 @@ int CRYPTO_dbg_push_info(const char *info, const char *file, int line)
 				}
 			}
 
-		ami->thread=CRYPTO_thread_id();
+		CRYPTO_THREADID_current(&ami->threadid);
 		ami->file=file;
 		ami->line=line;
 		ami->info=info;
 		ami->references=1;
 		ami->next=NULL;
 
-		if ((amim=(APP_INFO *)lh_insert(amih,(char *)ami)) != NULL)
+		if ((amim=lh_APP_INFO_insert(amih,ami)) != NULL)
 			{
 #ifdef LEVITTE_DEBUG_MEM
-			if (ami->thread != amim->thread)
+			if (CRYPTO_THREADID_cmp(&ami->threadid, &amim->threadid))
 				{
 				fprintf(stderr, "CRYPTO_push_info(): previous info has other thread ID (%lu) than the current thread (%lu)!!!!\n",
-					amim->thread, ami->thread);
+					CRYPTO_THREADID_hash(&amim->threadid),
+					CRYPTO_THREADID_hash(&ami->threadid));
 				abort();
 				}
 #endif
@@ -380,7 +445,7 @@ int CRYPTO_dbg_push_info(const char *info, const char *file, int line)
 	return(ret);
 	}
 
-int CRYPTO_dbg_pop_info(void)
+int CRYPTO_pop_info(void)
 	{
 	int ret=0;
 
@@ -395,7 +460,7 @@ int CRYPTO_dbg_pop_info(void)
 	return(ret);
 	}
 
-int CRYPTO_dbg_remove_all_info(void)
+int CRYPTO_remove_all_info(void)
 	{
 	int ret=0;
 
@@ -439,7 +504,7 @@ void CRYPTO_dbg_malloc(void *addr, int num, const char *file, int line,
 				}
 			if (mh == NULL)
 				{
-				if ((mh=lh_new(mem_hash, mem_cmp)) == NULL)
+				if ((mh=lh_MEM_new()) == NULL)
 					{
 					OPENSSL_free(addr);
 					OPENSSL_free(m);
@@ -453,9 +518,9 @@ void CRYPTO_dbg_malloc(void *addr, int num, const char *file, int line,
 			m->line=line;
 			m->num=num;
 			if (options & V_CRYPTO_MDEBUG_THREAD)
-				m->thread=CRYPTO_thread_id();
+				CRYPTO_THREADID_current(&m->threadid);
 			else
-				m->thread=0;
+				memset(&m->threadid, 0, sizeof(m->threadid));
 
 			if (order == break_order_num)
 				{
@@ -464,7 +529,7 @@ void CRYPTO_dbg_malloc(void *addr, int num, const char *file, int line,
 				}
 			m->order=order++;
 #ifdef LEVITTE_DEBUG_MEM
-			fprintf(stderr, "LEVITTE_DEBUG_MEM: [%5d] %c 0x%p (%d)\n",
+			fprintf(stderr, "LEVITTE_DEBUG_MEM: [%5ld] %c 0x%p (%d)\n",
 				m->order,
 				(before_p & 128) ? '*' : '+',
 				m->addr, m->num);
@@ -474,16 +539,16 @@ void CRYPTO_dbg_malloc(void *addr, int num, const char *file, int line,
 			else
 				m->time=0;
 
-			tmp.thread=CRYPTO_thread_id();
+			CRYPTO_THREADID_current(&tmp.threadid);
 			m->app_info=NULL;
 			if (amih != NULL
-				&& (amim=(APP_INFO *)lh_retrieve(amih,(char *)&tmp)) != NULL)
+			    && (amim=lh_APP_INFO_retrieve(amih,&tmp)) != NULL)
 				{
 				m->app_info = amim;
 				amim->references++;
 				}
 
-			if ((mm=(MEM *)lh_insert(mh,(char *)m)) != NULL)
+			if ((mm=lh_MEM_insert(mh, m)) != NULL)
 				{
 				/* Not good, but don't sweat it */
 				if (mm->app_info != NULL)
@@ -516,11 +581,11 @@ void CRYPTO_dbg_free(void *addr, int before_p)
 			MemCheck_off(); /* make sure we hold MALLOC2 lock */
 
 			m.addr=addr;
-			mp=(MEM *)lh_delete(mh,(char *)&m);
+			mp=lh_MEM_delete(mh,&m);
 			if (mp != NULL)
 				{
 #ifdef LEVITTE_DEBUG_MEM
-			fprintf(stderr, "LEVITTE_DEBUG_MEM: [%5d] - 0x%p (%d)\n",
+			fprintf(stderr, "LEVITTE_DEBUG_MEM: [%5ld] - 0x%p (%d)\n",
 				mp->order, mp->addr, mp->num);
 #endif
 				if (mp->app_info != NULL)
@@ -566,18 +631,18 @@ void CRYPTO_dbg_realloc(void *addr1, void *addr2, int num,
 			MemCheck_off(); /* make sure we hold MALLOC2 lock */
 
 			m.addr=addr1;
-			mp=(MEM *)lh_delete(mh,(char *)&m);
+			mp=lh_MEM_delete(mh,&m);
 			if (mp != NULL)
 				{
 #ifdef LEVITTE_DEBUG_MEM
-				fprintf(stderr, "LEVITTE_DEBUG_MEM: [%5d] * 0x%p (%d) -> 0x%p (%d)\n",
+				fprintf(stderr, "LEVITTE_DEBUG_MEM: [%5ld] * 0x%p (%d) -> 0x%p (%d)\n",
 					mp->order,
 					mp->addr, mp->num,
 					addr2, num);
 #endif
 				mp->addr=addr2;
 				mp->num=num;
-				lh_insert(mh,(char *)mp);
+				(void)lh_MEM_insert(mh,mp);
 				}
 
 			MemCheck_on(); /* release MALLOC2 lock
@@ -596,14 +661,14 @@ typedef struct mem_leak_st
 	long bytes;
 	} MEM_LEAK;
 
-static void print_leak(const MEM *m, MEM_LEAK *l)
+static void print_leak_doall_arg(const MEM *m, MEM_LEAK *l)
 	{
 	char buf[1024];
 	char *bufp = buf;
 	APP_INFO *amip;
 	int ami_cnt;
 	struct tm *lcl = NULL;
-	unsigned long ti;
+	CRYPTO_THREADID ti;
 
 #define BUF_REMAIN (sizeof buf - (size_t)(bufp - buf))
 
@@ -625,7 +690,8 @@ static void print_leak(const MEM *m, MEM_LEAK *l)
 
 	if (options & V_CRYPTO_MDEBUG_THREAD)
 		{
-		BIO_snprintf(bufp, BUF_REMAIN, "thread=%lu, ", m->thread);
+		BIO_snprintf(bufp, BUF_REMAIN, "thread=%lu, ",
+			CRYPTO_THREADID_hash(&m->threadid));
 		bufp += strlen(bufp);
 		}
 
@@ -642,8 +708,8 @@ static void print_leak(const MEM *m, MEM_LEAK *l)
 	ami_cnt=0;
 	if (!amip)
 		return;
-	ti=amip->thread;
-	
+	CRYPTO_THREADID_cpy(&ti, &amip->threadid);
+
 	do
 		{
 		int buf_len;
@@ -653,7 +719,8 @@ static void print_leak(const MEM *m, MEM_LEAK *l)
 		memset(buf,'>',ami_cnt);
 		BIO_snprintf(buf + ami_cnt, sizeof buf - ami_cnt,
 			" thread=%lu, file=%s, line=%d, info=\"",
-			amip->thread, amip->file, amip->line);
+			CRYPTO_THREADID_hash(&amip->threadid), amip->file,
+			amip->line);
 		buf_len=strlen(buf);
 		info_len=strlen(amip->info);
 		if (128 - buf_len - 3 < info_len)
@@ -673,8 +740,8 @@ static void print_leak(const MEM *m, MEM_LEAK *l)
 
 		amip = amip->next;
 		}
-	while(amip && amip->thread == ti);
-		
+	while(amip && !CRYPTO_THREADID_cmp(&amip->threadid, &ti));
+
 #ifdef LEVITTE_DEBUG_MEM
 	if (amip)
 		{
@@ -684,7 +751,7 @@ static void print_leak(const MEM *m, MEM_LEAK *l)
 #endif
 	}
 
-static IMPLEMENT_LHASH_DOALL_ARG_FN(print_leak, const MEM *, MEM_LEAK *)
+static IMPLEMENT_LHASH_DOALL_ARG_FN(print_leak, const MEM, MEM_LEAK)
 
 void CRYPTO_mem_leaks(BIO *b)
 	{
@@ -699,12 +766,15 @@ void CRYPTO_mem_leaks(BIO *b)
 	ml.bytes=0;
 	ml.chunks=0;
 	if (mh != NULL)
-		lh_doall_arg(mh, LHASH_DOALL_ARG_FN(print_leak),
-				(char *)&ml);
+		lh_MEM_doall_arg(mh, LHASH_DOALL_ARG_FN(print_leak), MEM_LEAK,
+				 &ml);
 	if (ml.chunks != 0)
 		{
 		BIO_printf(b,"%ld bytes leaked in %d chunks\n",
 			   ml.bytes,ml.chunks);
+#ifdef CRYPTO_MDEBUG_ABORT
+		abort();
+#endif
 		}
 	else
 		{
@@ -717,7 +787,7 @@ void CRYPTO_mem_leaks(BIO *b)
 		 * XXX    This should be in CRYPTO_mem_leaks_cb,
 		 * and CRYPTO_mem_leaks should be implemented by
 		 * using CRYPTO_mem_leaks_cb.
-		 * (Also their should be a variant of lh_doall_arg
+		 * (Also there should be a variant of lh_doall_arg
 		 * that takes a function pointer instead of a void *;
 		 * this would obviate the ugly and illegal
 		 * void_fn_to_char kludge in CRYPTO_mem_leaks_cb.
@@ -734,14 +804,14 @@ void CRYPTO_mem_leaks(BIO *b)
 
 		if (mh != NULL)
 			{
-			lh_free(mh);
+			lh_MEM_free(mh);
 			mh = NULL;
 			}
 		if (amih != NULL)
 			{
-			if (lh_num_items(amih) == 0) 
+			if (lh_APP_INFO_num_items(amih) == 0) 
 				{
-				lh_free(amih);
+				lh_APP_INFO_free(amih);
 				amih = NULL;
 				}
 			}
@@ -779,39 +849,26 @@ void CRYPTO_mem_leaks_fp(FILE *fp)
 /* NB: The prototypes have been typedef'd to CRYPTO_MEM_LEAK_CB inside crypto.h
  * If this code is restructured, remove the callback type if it is no longer
  * needed. -- Geoff Thorpe */
-static void cb_leak(const MEM *m, CRYPTO_MEM_LEAK_CB **cb)
+
+/* Can't pass CRYPTO_MEM_LEAK_CB directly to lh_MEM_doall_arg because it
+ * is a function pointer and conversion to void * is prohibited. Instead
+ * pass its address
+ */
+
+typedef CRYPTO_MEM_LEAK_CB *PCRYPTO_MEM_LEAK_CB;
+
+static void cb_leak_doall_arg(const MEM *m, PCRYPTO_MEM_LEAK_CB *cb)
 	{
-	(**cb)(m->order,m->file,m->line,m->num,m->addr);
+	(*cb)(m->order,m->file,m->line,m->num,m->addr);
 	}
 
-static IMPLEMENT_LHASH_DOALL_ARG_FN(cb_leak, const MEM *, CRYPTO_MEM_LEAK_CB **)
+static IMPLEMENT_LHASH_DOALL_ARG_FN(cb_leak, const MEM, PCRYPTO_MEM_LEAK_CB)
 
 void CRYPTO_mem_leaks_cb(CRYPTO_MEM_LEAK_CB *cb)
 	{
 	if (mh == NULL) return;
 	CRYPTO_w_lock(CRYPTO_LOCK_MALLOC2);
-	lh_doall_arg(mh, LHASH_DOALL_ARG_FN(cb_leak), &cb);
+	lh_MEM_doall_arg(mh, LHASH_DOALL_ARG_FN(cb_leak), PCRYPTO_MEM_LEAK_CB,
+			 &cb);
 	CRYPTO_w_unlock(CRYPTO_LOCK_MALLOC2);
 	}
-
-void CRYPTO_malloc_debug_init(void)
-	{
-	CRYPTO_set_mem_debug_functions(
-		CRYPTO_dbg_malloc,
-		CRYPTO_dbg_realloc,
-		CRYPTO_dbg_free,
-		CRYPTO_dbg_set_options,
-		CRYPTO_dbg_get_options);
-	CRYPTO_set_mem_info_functions(
-		CRYPTO_dbg_push_info,
-		CRYPTO_dbg_pop_info,
-		CRYPTO_dbg_remove_all_info);
-	}
-
-char *CRYPTO_strdup(const char *str, const char *file, int line)
-	{
-	char *ret = CRYPTO_malloc(strlen(str)+1, file, line);
-
-	strcpy(ret, str);
-	return ret;
-	}
diff --git a/crypto/modes/Makefile b/crypto/modes/Makefile
new file mode 100644
index 000000000000..c825b12f2585
--- /dev/null
+++ b/crypto/modes/Makefile
@@ -0,0 +1,141 @@
+#
+# OpenSSL/crypto/modes/Makefile
+#
+
+DIR=	modes
+TOP=	../..
+CC=	cc
+INCLUDES= -I.. -I$(TOP) -I../../include
+CFLAG=-g
+MAKEFILE=	Makefile
+AR=		ar r
+
+MODES_ASM_OBJ=
+
+CFLAGS= $(INCLUDES) $(CFLAG)
+ASFLAGS= $(INCLUDES) $(ASFLAG)
+AFLAGS= $(ASFLAGS)
+
+GENERAL=Makefile
+TEST=
+APPS=
+
+LIB=$(TOP)/libcrypto.a
+LIBSRC= cbc128.c ctr128.c cts128.c cfb128.c ofb128.c gcm128.c \
+	ccm128.c xts128.c
+LIBOBJ= cbc128.o ctr128.o cts128.o cfb128.o ofb128.o gcm128.o \
+	ccm128.o xts128.o $(MODES_ASM_OBJ)
+
+SRC= $(LIBSRC)
+
+#EXHEADER= store.h str_compat.h
+EXHEADER= modes.h
+HEADER=	modes_lcl.h $(EXHEADER)
+
+ALL=    $(GENERAL) $(SRC) $(HEADER)
+
+top:
+	(cd ../..; $(MAKE) DIRS=crypto SDIRS=$(DIR) sub_all)
+
+all:	lib
+
+lib:	$(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
+	$(RANLIB) $(LIB) || echo Never mind.
+	@touch lib
+
+ghash-ia64.s:	asm/ghash-ia64.pl
+	$(PERL) asm/ghash-ia64.pl $@ $(CFLAGS)
+ghash-x86.s:	asm/ghash-x86.pl
+	$(PERL) asm/ghash-x86.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+ghash-x86_64.s:	asm/ghash-x86_64.pl
+	$(PERL) asm/ghash-x86_64.pl $(PERLASM_SCHEME) > $@
+ghash-sparcv9.s:	asm/ghash-sparcv9.pl
+	$(PERL) asm/ghash-sparcv9.pl $@ $(CFLAGS)
+ghash-alpha.s:	asm/ghash-alpha.pl
+	$(PERL) $< | $(CC) -E - | tee $@ > /dev/null
+ghash-parisc.s:	asm/ghash-parisc.pl
+	$(PERL) asm/ghash-parisc.pl $(PERLASM_SCHEME) $@
+
+# GNU make "catch all"
+ghash-%.S:	asm/ghash-%.pl;	$(PERL) $< $(PERLASM_SCHEME) $@
+
+ghash-armv4.o:	ghash-armv4.S
+
+files:
+	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
+
+links:
+	@$(PERL) $(TOP)/util/mklink.pl ../../include/openssl $(EXHEADER)
+	@$(PERL) $(TOP)/util/mklink.pl ../../test $(TEST)
+	@$(PERL) $(TOP)/util/mklink.pl ../../apps $(APPS)
+
+install:
+	@[ -n "$(INSTALLTOP)" ] # should be set by top Makefile...
+	@headerlist="$(EXHEADER)"; for i in $$headerlist; \
+	do  \
+	(cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i; \
+	chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i ); \
+	done;
+
+tags:
+	ctags $(SRC)
+
+tests:
+
+lint:
+	lint -DLINT $(INCLUDES) $(SRC)>fluff
+
+depend:
+	@[ -n "$(MAKEDEPEND)" ] # should be set by upper Makefile...
+	$(MAKEDEPEND) -- $(CFLAG) $(INCLUDES) $(DEPFLAG) -- $(PROGS) $(LIBSRC)
+
+dclean:
+	$(PERL) -pe 'if (/^# DO NOT DELETE THIS LINE/) {print; exit(0);}' $(MAKEFILE) >Makefile.new
+	mv -f Makefile.new $(MAKEFILE)
+
+clean:
+	rm -f *.s *.o */*.o *.obj lib tags core .pure .nfs* *.old *.bak fluff
+
+# DO NOT DELETE THIS LINE -- make depend depends on it.
+
+cbc128.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+cbc128.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+cbc128.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cbc128.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+cbc128.o: ../../include/openssl/symhacks.h cbc128.c modes_lcl.h
+ccm128.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+ccm128.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+ccm128.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ccm128.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+ccm128.o: ../../include/openssl/symhacks.h ccm128.c modes_lcl.h
+cfb128.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+cfb128.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+cfb128.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cfb128.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+cfb128.o: ../../include/openssl/symhacks.h cfb128.c modes_lcl.h
+ctr128.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+ctr128.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+ctr128.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ctr128.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+ctr128.o: ../../include/openssl/symhacks.h ctr128.c modes_lcl.h
+cts128.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+cts128.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+cts128.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+cts128.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+cts128.o: ../../include/openssl/symhacks.h cts128.c modes_lcl.h
+gcm128.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+gcm128.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+gcm128.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+gcm128.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+gcm128.o: ../../include/openssl/symhacks.h gcm128.c modes_lcl.h
+ofb128.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+ofb128.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+ofb128.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ofb128.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+ofb128.o: ../../include/openssl/symhacks.h modes_lcl.h ofb128.c
+xts128.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+xts128.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+xts128.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+xts128.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+xts128.o: ../../include/openssl/symhacks.h modes_lcl.h xts128.c
diff --git a/crypto/modes/asm/ghash-armv4.pl b/crypto/modes/asm/ghash-armv4.pl
new file mode 100755
index 000000000000..d91586ee2925
--- /dev/null
+++ b/crypto/modes/asm/ghash-armv4.pl
@@ -0,0 +1,429 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# April 2010
+#
+# The module implements "4-bit" GCM GHASH function and underlying
+# single multiplication operation in GF(2^128). "4-bit" means that it
+# uses 256 bytes per-key table [+32 bytes shared table]. There is no
+# experimental performance data available yet. The only approximation
+# that can be made at this point is based on code size. Inner loop is
+# 32 instructions long and on single-issue core should execute in <40
+# cycles. Having verified that gcc 3.4 didn't unroll corresponding
+# loop, this assembler loop body was found to be ~3x smaller than
+# compiler-generated one...
+#
+# July 2010
+#
+# Rescheduling for dual-issue pipeline resulted in 8.5% improvement on
+# Cortex A8 core and ~25 cycles per processed byte (which was observed
+# to be ~3 times faster than gcc-generated code:-)
+#
+# February 2011
+#
+# Profiler-assisted and platform-specific optimization resulted in 7%
+# improvement on Cortex A8 core and ~23.5 cycles per byte.
+#
+# March 2011
+#
+# Add NEON implementation featuring polynomial multiplication, i.e. no
+# lookup tables involved. On Cortex A8 it was measured to process one
+# byte in 15 cycles or 55% faster than integer-only code.
+
+# ====================================================================
+# Note about "528B" variant. In ARM case it makes lesser sense to
+# implement it for following reasons:
+#
+# - performance improvement won't be anywhere near 50%, because 128-
+#   bit shift operation is neatly fused with 128-bit xor here, and
+#   "538B" variant would eliminate only 4-5 instructions out of 32
+#   in the inner loop (meaning that estimated improvement is ~15%);
+# - ARM-based systems are often embedded ones and extra memory
+#   consumption might be unappreciated (for so little improvement);
+#
+# Byte order [in]dependence. =========================================
+#
+# Caller is expected to maintain specific *dword* order in Htable,
+# namely with *least* significant dword of 128-bit value at *lower*
+# address. This differs completely from C code and has everything to
+# do with ldm instruction and order in which dwords are "consumed" by
+# algorithm. *Byte* order within these dwords in turn is whatever
+# *native* byte order on current platform. See gcm128.c for working
+# example...
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$Xi="r0";	# argument block
+$Htbl="r1";
+$inp="r2";
+$len="r3";
+
+$Zll="r4";	# variables
+$Zlh="r5";
+$Zhl="r6";
+$Zhh="r7";
+$Tll="r8";
+$Tlh="r9";
+$Thl="r10";
+$Thh="r11";
+$nlo="r12";
+################# r13 is stack pointer
+$nhi="r14";
+################# r15 is program counter
+
+$rem_4bit=$inp;	# used in gcm_gmult_4bit
+$cnt=$len;
+
+sub Zsmash() {
+  my $i=12;
+  my @args=@_;
+  for ($Zll,$Zlh,$Zhl,$Zhh) {
+    $code.=<<___;
+#if __ARM_ARCH__>=7 && defined(__ARMEL__)
+	rev	$_,$_
+	str	$_,[$Xi,#$i]
+#elif defined(__ARMEB__)
+	str	$_,[$Xi,#$i]
+#else
+	mov	$Tlh,$_,lsr#8
+	strb	$_,[$Xi,#$i+3]
+	mov	$Thl,$_,lsr#16
+	strb	$Tlh,[$Xi,#$i+2]
+	mov	$Thh,$_,lsr#24
+	strb	$Thl,[$Xi,#$i+1]
+	strb	$Thh,[$Xi,#$i]
+#endif
+___
+    $code.="\t".shift(@args)."\n";
+    $i-=4;
+  }
+}
+
+$code=<<___;
+#include "arm_arch.h"
+
+.text
+.code	32
+
+.type	rem_4bit,%object
+.align	5
+rem_4bit:
+.short	0x0000,0x1C20,0x3840,0x2460
+.short	0x7080,0x6CA0,0x48C0,0x54E0
+.short	0xE100,0xFD20,0xD940,0xC560
+.short	0x9180,0x8DA0,0xA9C0,0xB5E0
+.size	rem_4bit,.-rem_4bit
+
+.type	rem_4bit_get,%function
+rem_4bit_get:
+	sub	$rem_4bit,pc,#8
+	sub	$rem_4bit,$rem_4bit,#32	@ &rem_4bit
+	b	.Lrem_4bit_got
+	nop
+.size	rem_4bit_get,.-rem_4bit_get
+
+.global	gcm_ghash_4bit
+.type	gcm_ghash_4bit,%function
+gcm_ghash_4bit:
+	sub	r12,pc,#8
+	add	$len,$inp,$len		@ $len to point at the end
+	stmdb	sp!,{r3-r11,lr}		@ save $len/end too
+	sub	r12,r12,#48		@ &rem_4bit
+
+	ldmia	r12,{r4-r11}		@ copy rem_4bit ...
+	stmdb	sp!,{r4-r11}		@ ... to stack
+
+	ldrb	$nlo,[$inp,#15]
+	ldrb	$nhi,[$Xi,#15]
+.Louter:
+	eor	$nlo,$nlo,$nhi
+	and	$nhi,$nlo,#0xf0
+	and	$nlo,$nlo,#0x0f
+	mov	$cnt,#14
+
+	add	$Zhh,$Htbl,$nlo,lsl#4
+	ldmia	$Zhh,{$Zll-$Zhh}	@ load Htbl[nlo]
+	add	$Thh,$Htbl,$nhi
+	ldrb	$nlo,[$inp,#14]
+
+	and	$nhi,$Zll,#0xf		@ rem
+	ldmia	$Thh,{$Tll-$Thh}	@ load Htbl[nhi]
+	add	$nhi,$nhi,$nhi
+	eor	$Zll,$Tll,$Zll,lsr#4
+	ldrh	$Tll,[sp,$nhi]		@ rem_4bit[rem]
+	eor	$Zll,$Zll,$Zlh,lsl#28
+	ldrb	$nhi,[$Xi,#14]
+	eor	$Zlh,$Tlh,$Zlh,lsr#4
+	eor	$Zlh,$Zlh,$Zhl,lsl#28
+	eor	$Zhl,$Thl,$Zhl,lsr#4
+	eor	$Zhl,$Zhl,$Zhh,lsl#28
+	eor	$Zhh,$Thh,$Zhh,lsr#4
+	eor	$nlo,$nlo,$nhi
+	and	$nhi,$nlo,#0xf0
+	and	$nlo,$nlo,#0x0f
+	eor	$Zhh,$Zhh,$Tll,lsl#16
+
+.Linner:
+	add	$Thh,$Htbl,$nlo,lsl#4
+	and	$nlo,$Zll,#0xf		@ rem
+	subs	$cnt,$cnt,#1
+	add	$nlo,$nlo,$nlo
+	ldmia	$Thh,{$Tll-$Thh}	@ load Htbl[nlo]
+	eor	$Zll,$Tll,$Zll,lsr#4
+	eor	$Zll,$Zll,$Zlh,lsl#28
+	eor	$Zlh,$Tlh,$Zlh,lsr#4
+	eor	$Zlh,$Zlh,$Zhl,lsl#28
+	ldrh	$Tll,[sp,$nlo]		@ rem_4bit[rem]
+	eor	$Zhl,$Thl,$Zhl,lsr#4
+	ldrplb	$nlo,[$inp,$cnt]
+	eor	$Zhl,$Zhl,$Zhh,lsl#28
+	eor	$Zhh,$Thh,$Zhh,lsr#4
+
+	add	$Thh,$Htbl,$nhi
+	and	$nhi,$Zll,#0xf		@ rem
+	eor	$Zhh,$Zhh,$Tll,lsl#16	@ ^= rem_4bit[rem]
+	add	$nhi,$nhi,$nhi
+	ldmia	$Thh,{$Tll-$Thh}	@ load Htbl[nhi]
+	eor	$Zll,$Tll,$Zll,lsr#4
+	ldrplb	$Tll,[$Xi,$cnt]
+	eor	$Zll,$Zll,$Zlh,lsl#28
+	eor	$Zlh,$Tlh,$Zlh,lsr#4
+	ldrh	$Tlh,[sp,$nhi]
+	eor	$Zlh,$Zlh,$Zhl,lsl#28
+	eor	$Zhl,$Thl,$Zhl,lsr#4
+	eor	$Zhl,$Zhl,$Zhh,lsl#28
+	eorpl	$nlo,$nlo,$Tll
+	eor	$Zhh,$Thh,$Zhh,lsr#4
+	andpl	$nhi,$nlo,#0xf0
+	andpl	$nlo,$nlo,#0x0f
+	eor	$Zhh,$Zhh,$Tlh,lsl#16	@ ^= rem_4bit[rem]
+	bpl	.Linner
+
+	ldr	$len,[sp,#32]		@ re-load $len/end
+	add	$inp,$inp,#16
+	mov	$nhi,$Zll
+___
+	&Zsmash("cmp\t$inp,$len","ldrneb\t$nlo,[$inp,#15]");
+$code.=<<___;
+	bne	.Louter
+
+	add	sp,sp,#36
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r11,pc}
+#else
+	ldmia	sp!,{r4-r11,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.size	gcm_ghash_4bit,.-gcm_ghash_4bit
+
+.global	gcm_gmult_4bit
+.type	gcm_gmult_4bit,%function
+gcm_gmult_4bit:
+	stmdb	sp!,{r4-r11,lr}
+	ldrb	$nlo,[$Xi,#15]
+	b	rem_4bit_get
+.Lrem_4bit_got:
+	and	$nhi,$nlo,#0xf0
+	and	$nlo,$nlo,#0x0f
+	mov	$cnt,#14
+
+	add	$Zhh,$Htbl,$nlo,lsl#4
+	ldmia	$Zhh,{$Zll-$Zhh}	@ load Htbl[nlo]
+	ldrb	$nlo,[$Xi,#14]
+
+	add	$Thh,$Htbl,$nhi
+	and	$nhi,$Zll,#0xf		@ rem
+	ldmia	$Thh,{$Tll-$Thh}	@ load Htbl[nhi]
+	add	$nhi,$nhi,$nhi
+	eor	$Zll,$Tll,$Zll,lsr#4
+	ldrh	$Tll,[$rem_4bit,$nhi]	@ rem_4bit[rem]
+	eor	$Zll,$Zll,$Zlh,lsl#28
+	eor	$Zlh,$Tlh,$Zlh,lsr#4
+	eor	$Zlh,$Zlh,$Zhl,lsl#28
+	eor	$Zhl,$Thl,$Zhl,lsr#4
+	eor	$Zhl,$Zhl,$Zhh,lsl#28
+	eor	$Zhh,$Thh,$Zhh,lsr#4
+	and	$nhi,$nlo,#0xf0
+	eor	$Zhh,$Zhh,$Tll,lsl#16
+	and	$nlo,$nlo,#0x0f
+
+.Loop:
+	add	$Thh,$Htbl,$nlo,lsl#4
+	and	$nlo,$Zll,#0xf		@ rem
+	subs	$cnt,$cnt,#1
+	add	$nlo,$nlo,$nlo
+	ldmia	$Thh,{$Tll-$Thh}	@ load Htbl[nlo]
+	eor	$Zll,$Tll,$Zll,lsr#4
+	eor	$Zll,$Zll,$Zlh,lsl#28
+	eor	$Zlh,$Tlh,$Zlh,lsr#4
+	eor	$Zlh,$Zlh,$Zhl,lsl#28
+	ldrh	$Tll,[$rem_4bit,$nlo]	@ rem_4bit[rem]
+	eor	$Zhl,$Thl,$Zhl,lsr#4
+	ldrplb	$nlo,[$Xi,$cnt]
+	eor	$Zhl,$Zhl,$Zhh,lsl#28
+	eor	$Zhh,$Thh,$Zhh,lsr#4
+
+	add	$Thh,$Htbl,$nhi
+	and	$nhi,$Zll,#0xf		@ rem
+	eor	$Zhh,$Zhh,$Tll,lsl#16	@ ^= rem_4bit[rem]
+	add	$nhi,$nhi,$nhi
+	ldmia	$Thh,{$Tll-$Thh}	@ load Htbl[nhi]
+	eor	$Zll,$Tll,$Zll,lsr#4
+	eor	$Zll,$Zll,$Zlh,lsl#28
+	eor	$Zlh,$Tlh,$Zlh,lsr#4
+	ldrh	$Tll,[$rem_4bit,$nhi]	@ rem_4bit[rem]
+	eor	$Zlh,$Zlh,$Zhl,lsl#28
+	eor	$Zhl,$Thl,$Zhl,lsr#4
+	eor	$Zhl,$Zhl,$Zhh,lsl#28
+	eor	$Zhh,$Thh,$Zhh,lsr#4
+	andpl	$nhi,$nlo,#0xf0
+	andpl	$nlo,$nlo,#0x0f
+	eor	$Zhh,$Zhh,$Tll,lsl#16	@ ^= rem_4bit[rem]
+	bpl	.Loop
+___
+	&Zsmash();
+$code.=<<___;
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r11,pc}
+#else
+	ldmia	sp!,{r4-r11,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.size	gcm_gmult_4bit,.-gcm_gmult_4bit
+___
+{
+my $cnt=$Htbl;	# $Htbl is used once in the very beginning
+
+my ($Hhi, $Hlo, $Zo, $T, $xi, $mod) = map("d$_",(0..7));
+my ($Qhi, $Qlo, $Z,  $R, $zero, $Qpost, $IN) = map("q$_",(8..15));
+
+# Z:Zo keeps 128-bit result shifted by 1 to the right, with bottom bit
+# in Zo. Or should I say "top bit", because GHASH is specified in
+# reverse bit order? Otherwise straightforward 128-bt H by one input
+# byte multiplication and modulo-reduction, times 16.
+
+sub Dlo()   { shift=~m|q([1]?[0-9])|?"d".($1*2):"";     }
+sub Dhi()   { shift=~m|q([1]?[0-9])|?"d".($1*2+1):"";   }
+sub Q()     { shift=~m|d([1-3]?[02468])|?"q".($1/2):""; }
+
+$code.=<<___;
+#if __ARM_ARCH__>=7
+.fpu	neon
+
+.global	gcm_gmult_neon
+.type	gcm_gmult_neon,%function
+.align	4
+gcm_gmult_neon:
+	sub		$Htbl,#16		@ point at H in GCM128_CTX
+	vld1.64		`&Dhi("$IN")`,[$Xi,:64]!@ load Xi
+	vmov.i32	$mod,#0xe1		@ our irreducible polynomial
+	vld1.64		`&Dlo("$IN")`,[$Xi,:64]!
+	vshr.u64	$mod,#32
+	vldmia		$Htbl,{$Hhi-$Hlo}	@ load H
+	veor		$zero,$zero
+#ifdef __ARMEL__
+	vrev64.8	$IN,$IN
+#endif
+	veor		$Qpost,$Qpost
+	veor		$R,$R
+	mov		$cnt,#16
+	veor		$Z,$Z
+	mov		$len,#16
+	veor		$Zo,$Zo
+	vdup.8		$xi,`&Dlo("$IN")`[0]	@ broadcast lowest byte
+	b		.Linner_neon
+.size	gcm_gmult_neon,.-gcm_gmult_neon
+
+.global	gcm_ghash_neon
+.type	gcm_ghash_neon,%function
+.align	4
+gcm_ghash_neon:
+	vld1.64		`&Dhi("$Z")`,[$Xi,:64]!	@ load Xi
+	vmov.i32	$mod,#0xe1		@ our irreducible polynomial
+	vld1.64		`&Dlo("$Z")`,[$Xi,:64]!
+	vshr.u64	$mod,#32
+	vldmia		$Xi,{$Hhi-$Hlo}		@ load H
+	veor		$zero,$zero
+	nop
+#ifdef __ARMEL__
+	vrev64.8	$Z,$Z
+#endif
+.Louter_neon:
+	vld1.64		`&Dhi($IN)`,[$inp]!	@ load inp
+	veor		$Qpost,$Qpost
+	vld1.64		`&Dlo($IN)`,[$inp]!
+	veor		$R,$R
+	mov		$cnt,#16
+#ifdef __ARMEL__
+	vrev64.8	$IN,$IN
+#endif
+	veor		$Zo,$Zo
+	veor		$IN,$Z			@ inp^=Xi
+	veor		$Z,$Z
+	vdup.8		$xi,`&Dlo("$IN")`[0]	@ broadcast lowest byte
+.Linner_neon:
+	subs		$cnt,$cnt,#1
+	vmull.p8	$Qlo,$Hlo,$xi		@ H.lo·Xi[i]
+	vmull.p8	$Qhi,$Hhi,$xi		@ H.hi·Xi[i]
+	vext.8		$IN,$zero,#1		@ IN>>=8
+
+	veor		$Z,$Qpost		@ modulo-scheduled part
+	vshl.i64	`&Dlo("$R")`,#48
+	vdup.8		$xi,`&Dlo("$IN")`[0]	@ broadcast lowest byte
+	veor		$T,`&Dlo("$Qlo")`,`&Dlo("$Z")`
+
+	veor		`&Dhi("$Z")`,`&Dlo("$R")`
+	vuzp.8		$Qlo,$Qhi
+	vsli.8		$Zo,$T,#1		@ compose the "carry" byte
+	vext.8		$Z,$zero,#1		@ Z>>=8
+
+	vmull.p8	$R,$Zo,$mod		@ "carry"·0xe1
+	vshr.u8		$Zo,$T,#7		@ save Z's bottom bit
+	vext.8		$Qpost,$Qlo,$zero,#1	@ Qlo>>=8
+	veor		$Z,$Qhi
+	bne		.Linner_neon
+
+	veor		$Z,$Qpost		@ modulo-scheduled artefact
+	vshl.i64	`&Dlo("$R")`,#48
+	veor		`&Dhi("$Z")`,`&Dlo("$R")`
+
+	@ finalization, normalize Z:Zo
+	vand		$Zo,$mod		@ suffices to mask the bit
+	vshr.u64	`&Dhi(&Q("$Zo"))`,`&Dlo("$Z")`,#63
+	vshl.i64	$Z,#1
+	subs		$len,#16
+	vorr		$Z,`&Q("$Zo")`		@ Z=Z:Zo<<1
+	bne		.Louter_neon
+
+#ifdef __ARMEL__
+	vrev64.8	$Z,$Z
+#endif
+	sub		$Xi,#16	
+	vst1.64		`&Dhi("$Z")`,[$Xi,:64]!	@ write out Xi
+	vst1.64		`&Dlo("$Z")`,[$Xi,:64]
+
+	bx	lr
+.size	gcm_ghash_neon,.-gcm_ghash_neon
+#endif
+___
+}
+$code.=<<___;
+.asciz  "GHASH for ARMv4/NEON, CRYPTOGAMS by <appro\@openssl.org>"
+.align  2
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
+print $code;
+close STDOUT; # enforce flush
diff --git a/crypto/modes/asm/ghash-ia64.pl b/crypto/modes/asm/ghash-ia64.pl
new file mode 100755
index 000000000000..0354c9544485
--- /dev/null
+++ b/crypto/modes/asm/ghash-ia64.pl
@@ -0,0 +1,463 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# March 2010
+#
+# The module implements "4-bit" GCM GHASH function and underlying
+# single multiplication operation in GF(2^128). "4-bit" means that it
+# uses 256 bytes per-key table [+128 bytes shared table]. Streamed
+# GHASH performance was measured to be 6.67 cycles per processed byte
+# on Itanium 2, which is >90% better than Microsoft compiler generated
+# code. To anchor to something else sha1-ia64.pl module processes one
+# byte in 5.7 cycles. On Itanium GHASH should run at ~8.5 cycles per
+# byte.
+
+# September 2010
+#
+# It was originally thought that it makes lesser sense to implement
+# "528B" variant on Itanium 2 for following reason. Because number of
+# functional units is naturally limited, it appeared impossible to
+# implement "528B" loop in 4 cycles, only in 5. This would mean that
+# theoretically performance improvement couldn't be more than 20%.
+# But occasionally you prove yourself wrong:-) I figured out a way to
+# fold couple of instructions and having freed yet another instruction
+# slot by unrolling the loop... Resulting performance is 4.45 cycles
+# per processed byte and 50% better than "256B" version. On original
+# Itanium performance should remain the same as the "256B" version,
+# i.e. ~8.5 cycles.
+
+$output=shift and (open STDOUT,">$output" or die "can't open $output: $!");
+
+if ($^O eq "hpux") {
+    $ADDP="addp4";
+    for (@ARGV) { $ADDP="add" if (/[\+DD|\-mlp]64/); }
+} else { $ADDP="add"; }
+for (@ARGV)  {  $big_endian=1 if (/\-DB_ENDIAN/);
+                $big_endian=0 if (/\-DL_ENDIAN/);  }
+if (!defined($big_endian))
+             {  $big_endian=(unpack('L',pack('N',1))==1);  }
+
+sub loop() {
+my $label=shift;
+my ($p16,$p17)=(shift)?("p63","p63"):("p16","p17"); # mask references to inp
+
+# Loop is scheduled for 6 ticks on Itanium 2 and 8 on Itanium, i.e.
+# in scalable manner;-) Naturally assuming data in L1 cache...
+# Special note about 'dep' instruction, which is used to construct
+# &rem_4bit[Zlo&0xf]. It works, because rem_4bit is aligned at 128
+# bytes boundary and lower 7 bits of its address are guaranteed to
+# be zero.
+$code.=<<___;
+$label:
+{ .mfi;	(p18)	ld8	Hlo=[Hi[1]],-8
+	(p19)	dep	rem=Zlo,rem_4bitp,3,4	}
+{ .mfi;	(p19)	xor	Zhi=Zhi,Hhi
+	($p17)	xor	xi[1]=xi[1],in[1]	};;
+{ .mfi;	(p18)	ld8	Hhi=[Hi[1]]
+	(p19)	shrp	Zlo=Zhi,Zlo,4		}
+{ .mfi;	(p19)	ld8	rem=[rem]
+	(p18)	and	Hi[1]=mask0xf0,xi[2]	};;
+{ .mmi;	($p16)	ld1	in[0]=[inp],-1
+	(p18)	xor	Zlo=Zlo,Hlo
+	(p19)	shr.u	Zhi=Zhi,4		}
+{ .mib;	(p19)	xor	Hhi=Hhi,rem
+	(p18)	add	Hi[1]=Htbl,Hi[1]	};;
+
+{ .mfi;	(p18)	ld8	Hlo=[Hi[1]],-8
+	(p18)	dep	rem=Zlo,rem_4bitp,3,4	}
+{ .mfi;	(p17)	shladd	Hi[0]=xi[1],4,r0
+	(p18)	xor	Zhi=Zhi,Hhi		};;
+{ .mfi;	(p18)	ld8	Hhi=[Hi[1]]
+	(p18)	shrp	Zlo=Zhi,Zlo,4		}
+{ .mfi;	(p18)	ld8	rem=[rem]
+	(p17)	and	Hi[0]=mask0xf0,Hi[0]	};;
+{ .mmi;	(p16)	ld1	xi[0]=[Xi],-1
+	(p18)	xor	Zlo=Zlo,Hlo
+	(p18)	shr.u	Zhi=Zhi,4		}
+{ .mib;	(p18)	xor	Hhi=Hhi,rem
+	(p17)	add	Hi[0]=Htbl,Hi[0]
+	br.ctop.sptk	$label			};;
+___
+}
+
+$code=<<___;
+.explicit
+.text
+
+prevfs=r2;	prevlc=r3;	prevpr=r8;
+mask0xf0=r21;
+rem=r22;	rem_4bitp=r23;
+Xi=r24;		Htbl=r25;
+inp=r26;	end=r27;
+Hhi=r28;	Hlo=r29;
+Zhi=r30;	Zlo=r31;
+
+.align	128
+.skip	16					// aligns loop body
+.global	gcm_gmult_4bit#
+.proc	gcm_gmult_4bit#
+gcm_gmult_4bit:
+	.prologue
+{ .mmi;	.save	ar.pfs,prevfs
+	alloc	prevfs=ar.pfs,2,6,0,8
+	$ADDP	Xi=15,in0			// &Xi[15]
+	mov	rem_4bitp=ip		}
+{ .mii;	$ADDP	Htbl=8,in1			// &Htbl[0].lo
+	.save	ar.lc,prevlc
+	mov	prevlc=ar.lc
+	.save	pr,prevpr
+	mov	prevpr=pr		};;
+
+	.body
+	.rotr	in[3],xi[3],Hi[2]
+
+{ .mib;	ld1	xi[2]=[Xi],-1			// Xi[15]
+	mov	mask0xf0=0xf0
+	brp.loop.imp	.Loop1,.Lend1-16};;
+{ .mmi;	ld1	xi[1]=[Xi],-1			// Xi[14]
+					};;
+{ .mii;	shladd	Hi[1]=xi[2],4,r0
+	mov	pr.rot=0x7<<16
+	mov	ar.lc=13		};;
+{ .mii;	and	Hi[1]=mask0xf0,Hi[1]
+	mov	ar.ec=3
+	xor	Zlo=Zlo,Zlo		};;
+{ .mii;	add	Hi[1]=Htbl,Hi[1]		// &Htbl[nlo].lo
+	add	rem_4bitp=rem_4bit#-gcm_gmult_4bit#,rem_4bitp
+	xor	Zhi=Zhi,Zhi		};;
+___
+	&loop	(".Loop1",1);
+$code.=<<___;
+.Lend1:
+{ .mib;	xor	Zhi=Zhi,Hhi		};;	// modulo-scheduling artefact
+{ .mib;	mux1	Zlo=Zlo,\@rev		};;
+{ .mib;	mux1	Zhi=Zhi,\@rev		};;
+{ .mmi;	add	Hlo=9,Xi;;			// ;; is here to prevent
+	add	Hhi=1,Xi		};;	// pipeline flush on Itanium
+{ .mib;	st8	[Hlo]=Zlo
+	mov	pr=prevpr,0x1ffff	};;
+{ .mib;	st8	[Hhi]=Zhi
+	mov	ar.lc=prevlc
+	br.ret.sptk.many	b0	};;
+.endp	gcm_gmult_4bit#
+___
+
+######################################################################
+# "528B" (well, "512B" actualy) streamed GHASH
+#
+$Xip="in0";
+$Htbl="in1";
+$inp="in2";
+$len="in3";
+$rem_8bit="loc0";
+$mask0xff="loc1";
+($sum,$rum) = $big_endian ? ("nop.m","nop.m") : ("sum","rum");
+
+sub load_htable() {
+    for (my $i=0;$i<8;$i++) {
+	$code.=<<___;
+{ .mmi;	ld8	r`16+2*$i+1`=[r8],16		// Htable[$i].hi
+	ld8	r`16+2*$i`=[r9],16	}	// Htable[$i].lo
+{ .mmi;	ldf8	f`32+2*$i+1`=[r10],16		// Htable[`8+$i`].hi
+	ldf8	f`32+2*$i`=[r11],16		// Htable[`8+$i`].lo
+___
+	$code.=shift	if (($i+$#_)==7);
+	$code.="\t};;\n"
+    }
+}
+
+$code.=<<___;
+prevsp=r3;
+
+.align	32
+.skip	16					// aligns loop body
+.global	gcm_ghash_4bit#
+.proc	gcm_ghash_4bit#
+gcm_ghash_4bit:
+	.prologue
+{ .mmi;	.save	ar.pfs,prevfs
+	alloc	prevfs=ar.pfs,4,2,0,0
+	.vframe	prevsp
+	mov	prevsp=sp
+	mov	$rem_8bit=ip		};;
+	.body
+{ .mfi;	$ADDP	r8=0+0,$Htbl
+	$ADDP	r9=0+8,$Htbl		}
+{ .mfi;	$ADDP	r10=128+0,$Htbl
+	$ADDP	r11=128+8,$Htbl		};;
+___
+	&load_htable(
+	"	$ADDP	$Xip=15,$Xip",		# &Xi[15]
+	"	$ADDP	$len=$len,$inp",	# &inp[len]
+	"	$ADDP	$inp=15,$inp",		# &inp[15]
+	"	mov	$mask0xff=0xff",
+	"	add	sp=-512,sp",
+	"	andcm	sp=sp,$mask0xff",	# align stack frame
+	"	add	r14=0,sp",
+	"	add	r15=8,sp");
+$code.=<<___;
+{ .mmi;	$sum	1<<1				// go big-endian
+	add	r8=256+0,sp
+	add	r9=256+8,sp		}
+{ .mmi;	add	r10=256+128+0,sp
+	add	r11=256+128+8,sp
+	add	$len=-17,$len		};;
+___
+for($i=0;$i<8;$i++) {	# generate first half of Hshr4[]
+my ($rlo,$rhi)=("r".eval(16+2*$i),"r".eval(16+2*$i+1));
+$code.=<<___;
+{ .mmi;	st8	[r8]=$rlo,16			// Htable[$i].lo
+	st8	[r9]=$rhi,16			// Htable[$i].hi
+	shrp	$rlo=$rhi,$rlo,4	}//;;
+{ .mmi;	stf8	[r10]=f`32+2*$i`,16		// Htable[`8+$i`].lo
+	stf8	[r11]=f`32+2*$i+1`,16		// Htable[`8+$i`].hi
+	shr.u	$rhi=$rhi,4		};;
+{ .mmi;	st8	[r14]=$rlo,16			// Htable[$i].lo>>4
+	st8	[r15]=$rhi,16		}//;;	// Htable[$i].hi>>4
+___
+}
+$code.=<<___;
+{ .mmi;	ld8	r16=[r8],16			// Htable[8].lo
+	ld8	r17=[r9],16		};;	// Htable[8].hi
+{ .mmi;	ld8	r18=[r8],16			// Htable[9].lo
+	ld8	r19=[r9],16		}	// Htable[9].hi
+{ .mmi;	rum	1<<5				// clear um.mfh
+	shrp	r16=r17,r16,4		};;
+___
+for($i=0;$i<6;$i++) {	# generate second half of Hshr4[]
+$code.=<<___;
+{ .mmi;	ld8	r`20+2*$i`=[r8],16		// Htable[`10+$i`].lo
+	ld8	r`20+2*$i+1`=[r9],16		// Htable[`10+$i`].hi
+	shr.u	r`16+2*$i+1`=r`16+2*$i+1`,4	};;
+{ .mmi;	st8	[r14]=r`16+2*$i`,16		// Htable[`8+$i`].lo>>4
+	st8	[r15]=r`16+2*$i+1`,16		// Htable[`8+$i`].hi>>4
+	shrp	r`18+2*$i`=r`18+2*$i+1`,r`18+2*$i`,4	}
+___
+}
+$code.=<<___;
+{ .mmi;	shr.u	r`16+2*$i+1`=r`16+2*$i+1`,4	};;
+{ .mmi;	st8	[r14]=r`16+2*$i`,16		// Htable[`8+$i`].lo>>4
+	st8	[r15]=r`16+2*$i+1`,16		// Htable[`8+$i`].hi>>4
+	shrp	r`18+2*$i`=r`18+2*$i+1`,r`18+2*$i`,4	}
+{ .mmi;	add	$Htbl=256,sp			// &Htable[0]
+	add	$rem_8bit=rem_8bit#-gcm_ghash_4bit#,$rem_8bit
+	shr.u	r`18+2*$i+1`=r`18+2*$i+1`,4	};;
+{ .mmi;	st8	[r14]=r`18+2*$i`		// Htable[`8+$i`].lo>>4
+	st8	[r15]=r`18+2*$i+1`	}	// Htable[`8+$i`].hi>>4
+___
+
+$in="r15";
+@xi=("r16","r17");
+@rem=("r18","r19");
+($Alo,$Ahi,$Blo,$Bhi,$Zlo,$Zhi)=("r20","r21","r22","r23","r24","r25");
+($Atbl,$Btbl)=("r26","r27");
+
+$code.=<<___;	# (p16)
+{ .mmi;	ld1	$in=[$inp],-1			//(p16) *inp--
+	ld1	$xi[0]=[$Xip],-1		//(p16) *Xi--
+	cmp.eq	p0,p6=r0,r0		};;	//	clear p6
+___
+push (@xi,shift(@xi)); push (@rem,shift(@rem));	# "rotate" registers
+
+$code.=<<___;	# (p16),(p17)
+{ .mmi;	ld1	$xi[0]=[$Xip],-1		//(p16) *Xi--
+	xor	$xi[1]=$xi[1],$in	};;	//(p17) xi=$xi[i]^inp[i]
+{ .mii;	ld1	$in=[$inp],-1			//(p16) *inp--
+	dep	$Atbl=$xi[1],$Htbl,4,4		//(p17) &Htable[nlo].lo
+	and	$xi[1]=-16,$xi[1]	};;	//(p17) nhi=xi&0xf0
+.align	32
+.LOOP:
+{ .mmi;
+(p6)	st8	[$Xip]=$Zhi,13
+	xor	$Zlo=$Zlo,$Zlo
+	add	$Btbl=$xi[1],$Htbl	};;	//(p17) &Htable[nhi].lo
+___
+push (@xi,shift(@xi)); push (@rem,shift(@rem));	# "rotate" registers
+
+$code.=<<___;	# (p16),(p17),(p18)
+{ .mmi;	ld8	$Alo=[$Atbl],8			//(p18) Htable[nlo].lo,&Htable[nlo].hi
+	ld8	$rem[0]=[$Btbl],-256		//(p18) Htable[nhi].lo,&Hshr4[nhi].lo
+	xor	$xi[1]=$xi[1],$in	};;	//(p17) xi=$xi[i]^inp[i]
+{ .mfi;	ld8	$Ahi=[$Atbl]			//(p18) Htable[nlo].hi
+	dep	$Atbl=$xi[1],$Htbl,4,4	}	//(p17) &Htable[nlo].lo
+{ .mfi;	shladd	$rem[0]=$rem[0],4,r0		//(p18) Htable[nhi].lo<<4
+	xor	$Zlo=$Zlo,$Alo		};;	//(p18) Z.lo^=Htable[nlo].lo
+{ .mmi;	ld8	$Blo=[$Btbl],8			//(p18) Hshr4[nhi].lo,&Hshr4[nhi].hi
+	ld1	$in=[$inp],-1		}	//(p16) *inp--
+{ .mmi;	xor	$rem[0]=$rem[0],$Zlo		//(p18) Z.lo^(Htable[nhi].lo<<4)
+	mov	$Zhi=$Ahi			//(p18) Z.hi^=Htable[nlo].hi
+	and	$xi[1]=-16,$xi[1]	};;	//(p17) nhi=xi&0xf0
+{ .mmi;	ld8	$Bhi=[$Btbl]			//(p18) Hshr4[nhi].hi
+	ld1	$xi[0]=[$Xip],-1		//(p16) *Xi--
+	shrp	$Zlo=$Zhi,$Zlo,8	}	//(p18) Z.lo=(Z.hi<<56)|(Z.lo>>8)
+{ .mmi;	and	$rem[0]=$rem[0],$mask0xff	//(p18) rem=($Zlo^(Htable[nhi].lo<<4))&0xff
+	add	$Btbl=$xi[1],$Htbl	};;	//(p17) &Htable[nhi]
+___
+push (@xi,shift(@xi)); push (@rem,shift(@rem));	# "rotate" registers
+
+for ($i=1;$i<14;$i++) {
+# Above and below fragments are derived from this one by removing
+# unsuitable (p??) instructions.
+$code.=<<___;	# (p16),(p17),(p18),(p19)
+{ .mmi;	ld8	$Alo=[$Atbl],8			//(p18) Htable[nlo].lo,&Htable[nlo].hi
+	ld8	$rem[0]=[$Btbl],-256		//(p18) Htable[nhi].lo,&Hshr4[nhi].lo
+	shr.u	$Zhi=$Zhi,8		}	//(p19) Z.hi>>=8
+{ .mmi;	shladd	$rem[1]=$rem[1],1,$rem_8bit	//(p19) &rem_8bit[rem]
+	xor	$Zlo=$Zlo,$Blo			//(p19) Z.lo^=Hshr4[nhi].lo
+	xor	$xi[1]=$xi[1],$in	};;	//(p17) xi=$xi[i]^inp[i]
+{ .mmi;	ld8	$Ahi=[$Atbl]			//(p18) Htable[nlo].hi
+	ld2	$rem[1]=[$rem[1]]		//(p19) rem_8bit[rem]
+	dep	$Atbl=$xi[1],$Htbl,4,4	}	//(p17) &Htable[nlo].lo
+{ .mmi;	shladd	$rem[0]=$rem[0],4,r0		//(p18) Htable[nhi].lo<<4
+	xor	$Zlo=$Zlo,$Alo			//(p18) Z.lo^=Htable[nlo].lo
+	xor	$Zhi=$Zhi,$Bhi		};;	//(p19) Z.hi^=Hshr4[nhi].hi
+{ .mmi;	ld8	$Blo=[$Btbl],8			//(p18) Hshr4[nhi].lo,&Hshr4[nhi].hi
+	ld1	$in=[$inp],-1			//(p16) *inp--
+	shl	$rem[1]=$rem[1],48	}	//(p19) rem_8bit[rem]<<48
+{ .mmi;	xor	$rem[0]=$rem[0],$Zlo		//(p18) Z.lo^(Htable[nhi].lo<<4)
+	xor	$Zhi=$Zhi,$Ahi			//(p18) Z.hi^=Htable[nlo].hi
+	and	$xi[1]=-16,$xi[1]	};;	//(p17) nhi=xi&0xf0
+{ .mmi;	ld8	$Bhi=[$Btbl]			//(p18) Hshr4[nhi].hi
+	ld1	$xi[0]=[$Xip],-1		//(p16) *Xi--
+	shrp	$Zlo=$Zhi,$Zlo,8	}	//(p18) Z.lo=(Z.hi<<56)|(Z.lo>>8)
+{ .mmi;	and	$rem[0]=$rem[0],$mask0xff	//(p18) rem=($Zlo^(Htable[nhi].lo<<4))&0xff
+	xor	$Zhi=$Zhi,$rem[1]		//(p19) Z.hi^=rem_8bit[rem]<<48
+	add	$Btbl=$xi[1],$Htbl	};;	//(p17) &Htable[nhi]
+___
+push (@xi,shift(@xi)); push (@rem,shift(@rem));	# "rotate" registers
+}
+
+$code.=<<___;	# (p17),(p18),(p19)
+{ .mmi;	ld8	$Alo=[$Atbl],8			//(p18) Htable[nlo].lo,&Htable[nlo].hi
+	ld8	$rem[0]=[$Btbl],-256		//(p18) Htable[nhi].lo,&Hshr4[nhi].lo
+	shr.u	$Zhi=$Zhi,8		}	//(p19) Z.hi>>=8
+{ .mmi;	shladd	$rem[1]=$rem[1],1,$rem_8bit	//(p19) &rem_8bit[rem]
+	xor	$Zlo=$Zlo,$Blo			//(p19) Z.lo^=Hshr4[nhi].lo
+	xor	$xi[1]=$xi[1],$in	};;	//(p17) xi=$xi[i]^inp[i]
+{ .mmi;	ld8	$Ahi=[$Atbl]			//(p18) Htable[nlo].hi
+	ld2	$rem[1]=[$rem[1]]		//(p19) rem_8bit[rem]
+	dep	$Atbl=$xi[1],$Htbl,4,4	};;	//(p17) &Htable[nlo].lo
+{ .mmi;	shladd	$rem[0]=$rem[0],4,r0		//(p18) Htable[nhi].lo<<4
+	xor	$Zlo=$Zlo,$Alo			//(p18) Z.lo^=Htable[nlo].lo
+	xor	$Zhi=$Zhi,$Bhi		};;	//(p19) Z.hi^=Hshr4[nhi].hi
+{ .mmi;	ld8	$Blo=[$Btbl],8			//(p18) Hshr4[nhi].lo,&Hshr4[nhi].hi
+	shl	$rem[1]=$rem[1],48	}	//(p19) rem_8bit[rem]<<48
+{ .mmi;	xor	$rem[0]=$rem[0],$Zlo		//(p18) Z.lo^(Htable[nhi].lo<<4)
+	xor	$Zhi=$Zhi,$Ahi			//(p18) Z.hi^=Htable[nlo].hi
+	and	$xi[1]=-16,$xi[1]	};;	//(p17) nhi=xi&0xf0
+{ .mmi;	ld8	$Bhi=[$Btbl]			//(p18) Hshr4[nhi].hi
+	shrp	$Zlo=$Zhi,$Zlo,8	}	//(p18) Z.lo=(Z.hi<<56)|(Z.lo>>8)
+{ .mmi;	and	$rem[0]=$rem[0],$mask0xff	//(p18) rem=($Zlo^(Htable[nhi].lo<<4))&0xff
+	xor	$Zhi=$Zhi,$rem[1]		//(p19) Z.hi^=rem_8bit[rem]<<48
+	add	$Btbl=$xi[1],$Htbl	};;	//(p17) &Htable[nhi]
+___
+push (@xi,shift(@xi)); push (@rem,shift(@rem));	# "rotate" registers
+
+$code.=<<___;	# (p18),(p19)
+{ .mfi;	ld8	$Alo=[$Atbl],8			//(p18) Htable[nlo].lo,&Htable[nlo].hi
+	shr.u	$Zhi=$Zhi,8		}	//(p19) Z.hi>>=8
+{ .mfi;	shladd	$rem[1]=$rem[1],1,$rem_8bit	//(p19) &rem_8bit[rem]
+	xor	$Zlo=$Zlo,$Blo		};;	//(p19) Z.lo^=Hshr4[nhi].lo
+{ .mfi;	ld8	$Ahi=[$Atbl]			//(p18) Htable[nlo].hi
+	xor	$Zlo=$Zlo,$Alo		}	//(p18) Z.lo^=Htable[nlo].lo
+{ .mfi;	ld2	$rem[1]=[$rem[1]]		//(p19) rem_8bit[rem]
+	xor	$Zhi=$Zhi,$Bhi		};;	//(p19) Z.hi^=Hshr4[nhi].hi
+{ .mfi;	ld8	$Blo=[$Btbl],8			//(p18) Htable[nhi].lo,&Htable[nhi].hi
+	shl	$rem[1]=$rem[1],48	}	//(p19) rem_8bit[rem]<<48
+{ .mfi;	shladd	$rem[0]=$Zlo,4,r0		//(p18) Z.lo<<4
+	xor	$Zhi=$Zhi,$Ahi		};;	//(p18) Z.hi^=Htable[nlo].hi
+{ .mfi;	ld8	$Bhi=[$Btbl]			//(p18) Htable[nhi].hi
+	shrp	$Zlo=$Zhi,$Zlo,4	}	//(p18) Z.lo=(Z.hi<<60)|(Z.lo>>4)
+{ .mfi;	and	$rem[0]=$rem[0],$mask0xff	//(p18) rem=($Zlo^(Htable[nhi].lo<<4))&0xff
+	xor	$Zhi=$Zhi,$rem[1]	};;	//(p19) Z.hi^=rem_8bit[rem]<<48
+___
+push (@xi,shift(@xi)); push (@rem,shift(@rem));	# "rotate" registers
+
+$code.=<<___;	# (p19)
+{ .mmi;	cmp.ltu	p6,p0=$inp,$len
+	add	$inp=32,$inp
+	shr.u	$Zhi=$Zhi,4		}	//(p19) Z.hi>>=4
+{ .mmi;	shladd	$rem[1]=$rem[1],1,$rem_8bit	//(p19) &rem_8bit[rem]
+	xor	$Zlo=$Zlo,$Blo			//(p19) Z.lo^=Hshr4[nhi].lo
+	add	$Xip=9,$Xip		};;	//	&Xi.lo
+{ .mmi;	ld2	$rem[1]=[$rem[1]]		//(p19) rem_8bit[rem]
+(p6)	ld1	$in=[$inp],-1			//[p16] *inp--
+(p6)	extr.u	$xi[1]=$Zlo,8,8		}	//[p17] Xi[14]
+{ .mmi;	xor	$Zhi=$Zhi,$Bhi			//(p19) Z.hi^=Hshr4[nhi].hi
+(p6)	and	$xi[0]=$Zlo,$mask0xff	};;	//[p16] Xi[15]
+{ .mmi;	st8	[$Xip]=$Zlo,-8
+(p6)	xor	$xi[0]=$xi[0],$in		//[p17] xi=$xi[i]^inp[i]
+	shl	$rem[1]=$rem[1],48	};;	//(p19) rem_8bit[rem]<<48
+{ .mmi;
+(p6)	ld1	$in=[$inp],-1			//[p16] *inp--
+	xor	$Zhi=$Zhi,$rem[1]		//(p19) Z.hi^=rem_8bit[rem]<<48
+(p6)	dep	$Atbl=$xi[0],$Htbl,4,4	}	//[p17] &Htable[nlo].lo
+{ .mib;
+(p6)	and	$xi[0]=-16,$xi[0]		//[p17] nhi=xi&0xf0
+(p6)	br.cond.dptk.many	.LOOP	};;
+
+{ .mib;	st8	[$Xip]=$Zhi		};;
+{ .mib;	$rum	1<<1				// return to little-endian
+	.restore	sp
+	mov	sp=prevsp
+	br.ret.sptk.many	b0	};;
+.endp	gcm_ghash_4bit#
+___
+$code.=<<___;
+.align	128
+.type	rem_4bit#,\@object
+rem_4bit:
+        data8	0x0000<<48, 0x1C20<<48, 0x3840<<48, 0x2460<<48
+        data8	0x7080<<48, 0x6CA0<<48, 0x48C0<<48, 0x54E0<<48
+        data8	0xE100<<48, 0xFD20<<48, 0xD940<<48, 0xC560<<48
+        data8	0x9180<<48, 0x8DA0<<48, 0xA9C0<<48, 0xB5E0<<48
+.size	rem_4bit#,128
+.type	rem_8bit#,\@object
+rem_8bit:
+	data1	0x00,0x00, 0x01,0xC2, 0x03,0x84, 0x02,0x46, 0x07,0x08, 0x06,0xCA, 0x04,0x8C, 0x05,0x4E
+	data1	0x0E,0x10, 0x0F,0xD2, 0x0D,0x94, 0x0C,0x56, 0x09,0x18, 0x08,0xDA, 0x0A,0x9C, 0x0B,0x5E
+	data1	0x1C,0x20, 0x1D,0xE2, 0x1F,0xA4, 0x1E,0x66, 0x1B,0x28, 0x1A,0xEA, 0x18,0xAC, 0x19,0x6E
+	data1	0x12,0x30, 0x13,0xF2, 0x11,0xB4, 0x10,0x76, 0x15,0x38, 0x14,0xFA, 0x16,0xBC, 0x17,0x7E
+	data1	0x38,0x40, 0x39,0x82, 0x3B,0xC4, 0x3A,0x06, 0x3F,0x48, 0x3E,0x8A, 0x3C,0xCC, 0x3D,0x0E
+	data1	0x36,0x50, 0x37,0x92, 0x35,0xD4, 0x34,0x16, 0x31,0x58, 0x30,0x9A, 0x32,0xDC, 0x33,0x1E
+	data1	0x24,0x60, 0x25,0xA2, 0x27,0xE4, 0x26,0x26, 0x23,0x68, 0x22,0xAA, 0x20,0xEC, 0x21,0x2E
+	data1	0x2A,0x70, 0x2B,0xB2, 0x29,0xF4, 0x28,0x36, 0x2D,0x78, 0x2C,0xBA, 0x2E,0xFC, 0x2F,0x3E
+	data1	0x70,0x80, 0x71,0x42, 0x73,0x04, 0x72,0xC6, 0x77,0x88, 0x76,0x4A, 0x74,0x0C, 0x75,0xCE
+	data1	0x7E,0x90, 0x7F,0x52, 0x7D,0x14, 0x7C,0xD6, 0x79,0x98, 0x78,0x5A, 0x7A,0x1C, 0x7B,0xDE
+	data1	0x6C,0xA0, 0x6D,0x62, 0x6F,0x24, 0x6E,0xE6, 0x6B,0xA8, 0x6A,0x6A, 0x68,0x2C, 0x69,0xEE
+	data1	0x62,0xB0, 0x63,0x72, 0x61,0x34, 0x60,0xF6, 0x65,0xB8, 0x64,0x7A, 0x66,0x3C, 0x67,0xFE
+	data1	0x48,0xC0, 0x49,0x02, 0x4B,0x44, 0x4A,0x86, 0x4F,0xC8, 0x4E,0x0A, 0x4C,0x4C, 0x4D,0x8E
+	data1	0x46,0xD0, 0x47,0x12, 0x45,0x54, 0x44,0x96, 0x41,0xD8, 0x40,0x1A, 0x42,0x5C, 0x43,0x9E
+	data1	0x54,0xE0, 0x55,0x22, 0x57,0x64, 0x56,0xA6, 0x53,0xE8, 0x52,0x2A, 0x50,0x6C, 0x51,0xAE
+	data1	0x5A,0xF0, 0x5B,0x32, 0x59,0x74, 0x58,0xB6, 0x5D,0xF8, 0x5C,0x3A, 0x5E,0x7C, 0x5F,0xBE
+	data1	0xE1,0x00, 0xE0,0xC2, 0xE2,0x84, 0xE3,0x46, 0xE6,0x08, 0xE7,0xCA, 0xE5,0x8C, 0xE4,0x4E
+	data1	0xEF,0x10, 0xEE,0xD2, 0xEC,0x94, 0xED,0x56, 0xE8,0x18, 0xE9,0xDA, 0xEB,0x9C, 0xEA,0x5E
+	data1	0xFD,0x20, 0xFC,0xE2, 0xFE,0xA4, 0xFF,0x66, 0xFA,0x28, 0xFB,0xEA, 0xF9,0xAC, 0xF8,0x6E
+	data1	0xF3,0x30, 0xF2,0xF2, 0xF0,0xB4, 0xF1,0x76, 0xF4,0x38, 0xF5,0xFA, 0xF7,0xBC, 0xF6,0x7E
+	data1	0xD9,0x40, 0xD8,0x82, 0xDA,0xC4, 0xDB,0x06, 0xDE,0x48, 0xDF,0x8A, 0xDD,0xCC, 0xDC,0x0E
+	data1	0xD7,0x50, 0xD6,0x92, 0xD4,0xD4, 0xD5,0x16, 0xD0,0x58, 0xD1,0x9A, 0xD3,0xDC, 0xD2,0x1E
+	data1	0xC5,0x60, 0xC4,0xA2, 0xC6,0xE4, 0xC7,0x26, 0xC2,0x68, 0xC3,0xAA, 0xC1,0xEC, 0xC0,0x2E
+	data1	0xCB,0x70, 0xCA,0xB2, 0xC8,0xF4, 0xC9,0x36, 0xCC,0x78, 0xCD,0xBA, 0xCF,0xFC, 0xCE,0x3E
+	data1	0x91,0x80, 0x90,0x42, 0x92,0x04, 0x93,0xC6, 0x96,0x88, 0x97,0x4A, 0x95,0x0C, 0x94,0xCE
+	data1	0x9F,0x90, 0x9E,0x52, 0x9C,0x14, 0x9D,0xD6, 0x98,0x98, 0x99,0x5A, 0x9B,0x1C, 0x9A,0xDE
+	data1	0x8D,0xA0, 0x8C,0x62, 0x8E,0x24, 0x8F,0xE6, 0x8A,0xA8, 0x8B,0x6A, 0x89,0x2C, 0x88,0xEE
+	data1	0x83,0xB0, 0x82,0x72, 0x80,0x34, 0x81,0xF6, 0x84,0xB8, 0x85,0x7A, 0x87,0x3C, 0x86,0xFE
+	data1	0xA9,0xC0, 0xA8,0x02, 0xAA,0x44, 0xAB,0x86, 0xAE,0xC8, 0xAF,0x0A, 0xAD,0x4C, 0xAC,0x8E
+	data1	0xA7,0xD0, 0xA6,0x12, 0xA4,0x54, 0xA5,0x96, 0xA0,0xD8, 0xA1,0x1A, 0xA3,0x5C, 0xA2,0x9E
+	data1	0xB5,0xE0, 0xB4,0x22, 0xB6,0x64, 0xB7,0xA6, 0xB2,0xE8, 0xB3,0x2A, 0xB1,0x6C, 0xB0,0xAE
+	data1	0xBB,0xF0, 0xBA,0x32, 0xB8,0x74, 0xB9,0xB6, 0xBC,0xF8, 0xBD,0x3A, 0xBF,0x7C, 0xBE,0xBE
+.size	rem_8bit#,512
+stringz	"GHASH for IA64, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/mux1(\s+)\S+\@rev/nop.i$1 0x0/gm      if ($big_endian);
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+
+print $code;
+close STDOUT;
diff --git a/crypto/modes/asm/ghash-parisc.pl b/crypto/modes/asm/ghash-parisc.pl
new file mode 100755
index 000000000000..8c7454ee9340
--- /dev/null
+++ b/crypto/modes/asm/ghash-parisc.pl
@@ -0,0 +1,730 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# April 2010
+#
+# The module implements "4-bit" GCM GHASH function and underlying
+# single multiplication operation in GF(2^128). "4-bit" means that it
+# uses 256 bytes per-key table [+128 bytes shared table]. On PA-7100LC
+# it processes one byte in 19.6 cycles, which is more than twice as
+# fast as code generated by gcc 3.2. PA-RISC 2.0 loop is scheduled for
+# 8 cycles, but measured performance on PA-8600 system is ~9 cycles per
+# processed byte. This is ~2.2x faster than 64-bit code generated by
+# vendor compiler (which used to be very hard to beat:-).
+#
+# Special thanks to polarhome.com for providing HP-UX account.
+
+$flavour = shift;
+$output = shift;
+open STDOUT,">$output";
+
+if ($flavour =~ /64/) {
+	$LEVEL		="2.0W";
+	$SIZE_T		=8;
+	$FRAME_MARKER	=80;
+	$SAVED_RP	=16;
+	$PUSH		="std";
+	$PUSHMA		="std,ma";
+	$POP		="ldd";
+	$POPMB		="ldd,mb";
+	$NREGS		=6;
+} else {
+	$LEVEL		="1.0";	#"\n\t.ALLOW\t2.0";
+	$SIZE_T		=4;
+	$FRAME_MARKER	=48;
+	$SAVED_RP	=20;
+	$PUSH		="stw";
+	$PUSHMA		="stwm";
+	$POP		="ldw";
+	$POPMB		="ldwm";
+	$NREGS		=11;
+}
+
+$FRAME=10*$SIZE_T+$FRAME_MARKER;# NREGS saved regs + frame marker
+				#                 [+ argument transfer]
+
+################# volatile registers
+$Xi="%r26";	# argument block
+$Htbl="%r25";
+$inp="%r24";
+$len="%r23";
+$Hhh=$Htbl;	# variables
+$Hll="%r22";
+$Zhh="%r21";
+$Zll="%r20";
+$cnt="%r19";
+$rem_4bit="%r28";
+$rem="%r29";
+$mask0xf0="%r31";
+
+################# preserved registers
+$Thh="%r1";
+$Tll="%r2";
+$nlo="%r3";
+$nhi="%r4";
+$byte="%r5";
+if ($SIZE_T==4) {
+	$Zhl="%r6";
+	$Zlh="%r7";
+	$Hhl="%r8";
+	$Hlh="%r9";
+	$Thl="%r10";
+	$Tlh="%r11";
+}
+$rem2="%r6";	# used in PA-RISC 2.0 code
+
+$code.=<<___;
+	.LEVEL	$LEVEL
+	.SPACE	\$TEXT\$
+	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY
+
+	.EXPORT	gcm_gmult_4bit,ENTRY,ARGW0=GR,ARGW1=GR
+	.ALIGN	64
+gcm_gmult_4bit
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-10*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=$NREGS
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+___
+$code.=<<___ if ($SIZE_T==4);
+	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
+	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
+	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
+	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
+	$PUSH	%r11,`-$FRAME+8*$SIZE_T`(%sp)
+___
+$code.=<<___;
+	blr	%r0,$rem_4bit
+	ldi	3,$rem
+L\$pic_gmult
+	andcm	$rem_4bit,$rem,$rem_4bit
+	addl	$inp,$len,$len
+	ldo	L\$rem_4bit-L\$pic_gmult($rem_4bit),$rem_4bit
+	ldi	0xf0,$mask0xf0
+___
+$code.=<<___ if ($SIZE_T==4);
+	ldi	31,$rem
+	mtctl	$rem,%cr11
+	extrd,u,*= $rem,%sar,1,$rem	; executes on PA-RISC 1.0
+	b	L\$parisc1_gmult
+	nop
+___
+
+$code.=<<___;
+	ldb	15($Xi),$nlo
+	ldo	8($Htbl),$Hll
+
+	and	$mask0xf0,$nlo,$nhi
+	depd,z	$nlo,59,4,$nlo
+
+	ldd	$nlo($Hll),$Zll
+	ldd	$nlo($Hhh),$Zhh
+
+	depd,z	$Zll,60,4,$rem
+	shrpd	$Zhh,$Zll,4,$Zll
+	extrd,u	$Zhh,59,60,$Zhh
+	ldb	14($Xi),$nlo
+
+	ldd	$nhi($Hll),$Tll
+	ldd	$nhi($Hhh),$Thh
+	and	$mask0xf0,$nlo,$nhi
+	depd,z	$nlo,59,4,$nlo
+
+	xor	$Tll,$Zll,$Zll
+	xor	$Thh,$Zhh,$Zhh
+	ldd	$rem($rem_4bit),$rem
+	b	L\$oop_gmult_pa2
+	ldi	13,$cnt
+
+	.ALIGN	8
+L\$oop_gmult_pa2
+	xor	$rem,$Zhh,$Zhh		; moved here to work around gas bug
+	depd,z	$Zll,60,4,$rem
+
+	shrpd	$Zhh,$Zll,4,$Zll
+	extrd,u	$Zhh,59,60,$Zhh
+	ldd	$nlo($Hll),$Tll
+	ldd	$nlo($Hhh),$Thh
+
+	xor	$Tll,$Zll,$Zll
+	xor	$Thh,$Zhh,$Zhh
+	ldd	$rem($rem_4bit),$rem
+
+	xor	$rem,$Zhh,$Zhh
+	depd,z	$Zll,60,4,$rem
+	ldbx	$cnt($Xi),$nlo
+
+	shrpd	$Zhh,$Zll,4,$Zll
+	extrd,u	$Zhh,59,60,$Zhh
+	ldd	$nhi($Hll),$Tll
+	ldd	$nhi($Hhh),$Thh
+
+	and	$mask0xf0,$nlo,$nhi
+	depd,z	$nlo,59,4,$nlo
+	ldd	$rem($rem_4bit),$rem
+
+	xor	$Tll,$Zll,$Zll
+	addib,uv -1,$cnt,L\$oop_gmult_pa2
+	xor	$Thh,$Zhh,$Zhh
+
+	xor	$rem,$Zhh,$Zhh
+	depd,z	$Zll,60,4,$rem
+
+	shrpd	$Zhh,$Zll,4,$Zll
+	extrd,u	$Zhh,59,60,$Zhh
+	ldd	$nlo($Hll),$Tll
+	ldd	$nlo($Hhh),$Thh
+
+	xor	$Tll,$Zll,$Zll
+	xor	$Thh,$Zhh,$Zhh
+	ldd	$rem($rem_4bit),$rem
+
+	xor	$rem,$Zhh,$Zhh
+	depd,z	$Zll,60,4,$rem
+
+	shrpd	$Zhh,$Zll,4,$Zll
+	extrd,u	$Zhh,59,60,$Zhh
+	ldd	$nhi($Hll),$Tll
+	ldd	$nhi($Hhh),$Thh
+
+	xor	$Tll,$Zll,$Zll
+	xor	$Thh,$Zhh,$Zhh
+	ldd	$rem($rem_4bit),$rem
+
+	xor	$rem,$Zhh,$Zhh
+	std	$Zll,8($Xi)
+	std	$Zhh,0($Xi)
+___
+
+$code.=<<___ if ($SIZE_T==4);
+	b	L\$done_gmult
+	nop
+
+L\$parisc1_gmult
+	ldb	15($Xi),$nlo
+	ldo	12($Htbl),$Hll
+	ldo	8($Htbl),$Hlh
+	ldo	4($Htbl),$Hhl
+
+	and	$mask0xf0,$nlo,$nhi
+	zdep	$nlo,27,4,$nlo
+
+	ldwx	$nlo($Hll),$Zll
+	ldwx	$nlo($Hlh),$Zlh
+	ldwx	$nlo($Hhl),$Zhl
+	ldwx	$nlo($Hhh),$Zhh
+	zdep	$Zll,28,4,$rem
+	ldb	14($Xi),$nlo
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zlh,$Zll,4,$Zll
+	ldwx	$nhi($Hll),$Tll
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	ldwx	$nhi($Hlh),$Tlh
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	ldwx	$nhi($Hhl),$Thl
+	extru	$Zhh,27,28,$Zhh
+	ldwx	$nhi($Hhh),$Thh
+	xor	$rem,$Zhh,$Zhh
+	and	$mask0xf0,$nlo,$nhi
+	zdep	$nlo,27,4,$nlo
+
+	xor	$Tll,$Zll,$Zll
+	ldwx	$nlo($Hll),$Tll
+	xor	$Tlh,$Zlh,$Zlh
+	ldwx	$nlo($Hlh),$Tlh
+	xor	$Thl,$Zhl,$Zhl
+	b	L\$oop_gmult_pa1
+	ldi	13,$cnt
+
+	.ALIGN	8
+L\$oop_gmult_pa1
+	zdep	$Zll,28,4,$rem
+	ldwx	$nlo($Hhl),$Thl
+	xor	$Thh,$Zhh,$Zhh
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zlh,$Zll,4,$Zll
+	ldwx	$nlo($Hhh),$Thh
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	ldbx	$cnt($Xi),$nlo
+	xor	$Tll,$Zll,$Zll
+	ldwx	$nhi($Hll),$Tll
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	xor	$Tlh,$Zlh,$Zlh
+	ldwx	$nhi($Hlh),$Tlh
+	extru	$Zhh,27,28,$Zhh
+	xor	$Thl,$Zhl,$Zhl
+	ldwx	$nhi($Hhl),$Thl
+	xor	$rem,$Zhh,$Zhh
+	zdep	$Zll,28,4,$rem
+	xor	$Thh,$Zhh,$Zhh
+	ldwx	$nhi($Hhh),$Thh
+	shrpw	$Zlh,$Zll,4,$Zll
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	and	$mask0xf0,$nlo,$nhi
+	extru	$Zhh,27,28,$Zhh
+	zdep	$nlo,27,4,$nlo
+	xor	$Tll,$Zll,$Zll
+	ldwx	$nlo($Hll),$Tll
+	xor	$Tlh,$Zlh,$Zlh
+	ldwx	$nlo($Hlh),$Tlh
+	xor	$rem,$Zhh,$Zhh
+	addib,uv -1,$cnt,L\$oop_gmult_pa1
+	xor	$Thl,$Zhl,$Zhl
+
+	zdep	$Zll,28,4,$rem
+	ldwx	$nlo($Hhl),$Thl
+	xor	$Thh,$Zhh,$Zhh
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zlh,$Zll,4,$Zll
+	ldwx	$nlo($Hhh),$Thh
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	xor	$Tll,$Zll,$Zll
+	ldwx	$nhi($Hll),$Tll
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	xor	$Tlh,$Zlh,$Zlh
+	ldwx	$nhi($Hlh),$Tlh
+	extru	$Zhh,27,28,$Zhh
+	xor	$rem,$Zhh,$Zhh
+	xor	$Thl,$Zhl,$Zhl
+	ldwx	$nhi($Hhl),$Thl
+	xor	$Thh,$Zhh,$Zhh
+	ldwx	$nhi($Hhh),$Thh
+	zdep	$Zll,28,4,$rem
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zlh,$Zll,4,$Zll
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	extru	$Zhh,27,28,$Zhh
+	xor	$Tll,$Zll,$Zll
+	xor	$Tlh,$Zlh,$Zlh
+	xor	$rem,$Zhh,$Zhh
+	stw	$Zll,12($Xi)
+	xor	$Thl,$Zhl,$Zhl
+	stw	$Zlh,8($Xi)
+	xor	$Thh,$Zhh,$Zhh
+	stw	$Zhl,4($Xi)
+	stw	$Zhh,0($Xi)
+___
+$code.=<<___;
+L\$done_gmult
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2		; standard epilogue
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+___
+$code.=<<___ if ($SIZE_T==4);
+	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
+	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
+	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
+	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
+	$POP	`-$FRAME+8*$SIZE_T`(%sp),%r11
+___
+$code.=<<___;
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+
+	.EXPORT	gcm_ghash_4bit,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR,ARGW3=GR
+	.ALIGN	64
+gcm_ghash_4bit
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-10*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=11
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+___
+$code.=<<___ if ($SIZE_T==4);
+	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
+	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
+	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
+	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
+	$PUSH	%r11,`-$FRAME+8*$SIZE_T`(%sp)
+___
+$code.=<<___;
+	blr	%r0,$rem_4bit
+	ldi	3,$rem
+L\$pic_ghash
+	andcm	$rem_4bit,$rem,$rem_4bit
+	addl	$inp,$len,$len
+	ldo	L\$rem_4bit-L\$pic_ghash($rem_4bit),$rem_4bit
+	ldi	0xf0,$mask0xf0
+___
+$code.=<<___ if ($SIZE_T==4);
+	ldi	31,$rem
+	mtctl	$rem,%cr11
+	extrd,u,*= $rem,%sar,1,$rem	; executes on PA-RISC 1.0
+	b	L\$parisc1_ghash
+	nop
+___
+
+$code.=<<___;
+	ldb	15($Xi),$nlo
+	ldo	8($Htbl),$Hll
+
+L\$outer_ghash_pa2
+	ldb	15($inp),$nhi
+	xor	$nhi,$nlo,$nlo
+	and	$mask0xf0,$nlo,$nhi
+	depd,z	$nlo,59,4,$nlo
+
+	ldd	$nlo($Hll),$Zll
+	ldd	$nlo($Hhh),$Zhh
+
+	depd,z	$Zll,60,4,$rem
+	shrpd	$Zhh,$Zll,4,$Zll
+	extrd,u	$Zhh,59,60,$Zhh
+	ldb	14($Xi),$nlo
+	ldb	14($inp),$byte
+
+	ldd	$nhi($Hll),$Tll
+	ldd	$nhi($Hhh),$Thh
+	xor	$byte,$nlo,$nlo
+	and	$mask0xf0,$nlo,$nhi
+	depd,z	$nlo,59,4,$nlo
+
+	xor	$Tll,$Zll,$Zll
+	xor	$Thh,$Zhh,$Zhh
+	ldd	$rem($rem_4bit),$rem
+	b	L\$oop_ghash_pa2
+	ldi	13,$cnt
+
+	.ALIGN	8
+L\$oop_ghash_pa2
+	xor	$rem,$Zhh,$Zhh		; moved here to work around gas bug
+	depd,z	$Zll,60,4,$rem2
+
+	shrpd	$Zhh,$Zll,4,$Zll
+	extrd,u	$Zhh,59,60,$Zhh
+	ldd	$nlo($Hll),$Tll
+	ldd	$nlo($Hhh),$Thh
+
+	xor	$Tll,$Zll,$Zll
+	xor	$Thh,$Zhh,$Zhh
+	ldbx	$cnt($Xi),$nlo
+	ldbx	$cnt($inp),$byte
+
+	depd,z	$Zll,60,4,$rem
+	shrpd	$Zhh,$Zll,4,$Zll
+	ldd	$rem2($rem_4bit),$rem2
+
+	xor	$rem2,$Zhh,$Zhh
+	xor	$byte,$nlo,$nlo
+	ldd	$nhi($Hll),$Tll
+	ldd	$nhi($Hhh),$Thh
+
+	and	$mask0xf0,$nlo,$nhi
+	depd,z	$nlo,59,4,$nlo
+
+	extrd,u	$Zhh,59,60,$Zhh
+	xor	$Tll,$Zll,$Zll
+
+	ldd	$rem($rem_4bit),$rem
+	addib,uv -1,$cnt,L\$oop_ghash_pa2
+	xor	$Thh,$Zhh,$Zhh
+
+	xor	$rem,$Zhh,$Zhh
+	depd,z	$Zll,60,4,$rem2
+
+	shrpd	$Zhh,$Zll,4,$Zll
+	extrd,u	$Zhh,59,60,$Zhh
+	ldd	$nlo($Hll),$Tll
+	ldd	$nlo($Hhh),$Thh
+
+	xor	$Tll,$Zll,$Zll
+	xor	$Thh,$Zhh,$Zhh
+
+	depd,z	$Zll,60,4,$rem
+	shrpd	$Zhh,$Zll,4,$Zll
+	ldd	$rem2($rem_4bit),$rem2
+
+	xor	$rem2,$Zhh,$Zhh
+	ldd	$nhi($Hll),$Tll
+	ldd	$nhi($Hhh),$Thh
+
+	extrd,u	$Zhh,59,60,$Zhh
+	xor	$Tll,$Zll,$Zll
+	xor	$Thh,$Zhh,$Zhh
+	ldd	$rem($rem_4bit),$rem
+
+	xor	$rem,$Zhh,$Zhh
+	std	$Zll,8($Xi)
+	ldo	16($inp),$inp
+	std	$Zhh,0($Xi)
+	cmpb,*<> $inp,$len,L\$outer_ghash_pa2
+	copy	$Zll,$nlo
+___
+
+$code.=<<___ if ($SIZE_T==4);
+	b	L\$done_ghash
+	nop
+
+L\$parisc1_ghash
+	ldb	15($Xi),$nlo
+	ldo	12($Htbl),$Hll
+	ldo	8($Htbl),$Hlh
+	ldo	4($Htbl),$Hhl
+
+L\$outer_ghash_pa1
+	ldb	15($inp),$byte
+	xor	$byte,$nlo,$nlo
+	and	$mask0xf0,$nlo,$nhi
+	zdep	$nlo,27,4,$nlo
+
+	ldwx	$nlo($Hll),$Zll
+	ldwx	$nlo($Hlh),$Zlh
+	ldwx	$nlo($Hhl),$Zhl
+	ldwx	$nlo($Hhh),$Zhh
+	zdep	$Zll,28,4,$rem
+	ldb	14($Xi),$nlo
+	ldb	14($inp),$byte
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zlh,$Zll,4,$Zll
+	ldwx	$nhi($Hll),$Tll
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	ldwx	$nhi($Hlh),$Tlh
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	ldwx	$nhi($Hhl),$Thl
+	extru	$Zhh,27,28,$Zhh
+	ldwx	$nhi($Hhh),$Thh
+	xor	$byte,$nlo,$nlo
+	xor	$rem,$Zhh,$Zhh
+	and	$mask0xf0,$nlo,$nhi
+	zdep	$nlo,27,4,$nlo
+
+	xor	$Tll,$Zll,$Zll
+	ldwx	$nlo($Hll),$Tll
+	xor	$Tlh,$Zlh,$Zlh
+	ldwx	$nlo($Hlh),$Tlh
+	xor	$Thl,$Zhl,$Zhl
+	b	L\$oop_ghash_pa1
+	ldi	13,$cnt
+
+	.ALIGN	8
+L\$oop_ghash_pa1
+	zdep	$Zll,28,4,$rem
+	ldwx	$nlo($Hhl),$Thl
+	xor	$Thh,$Zhh,$Zhh
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zlh,$Zll,4,$Zll
+	ldwx	$nlo($Hhh),$Thh
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	ldbx	$cnt($Xi),$nlo
+	xor	$Tll,$Zll,$Zll
+	ldwx	$nhi($Hll),$Tll
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	ldbx	$cnt($inp),$byte
+	xor	$Tlh,$Zlh,$Zlh
+	ldwx	$nhi($Hlh),$Tlh
+	extru	$Zhh,27,28,$Zhh
+	xor	$Thl,$Zhl,$Zhl
+	ldwx	$nhi($Hhl),$Thl
+	xor	$rem,$Zhh,$Zhh
+	zdep	$Zll,28,4,$rem
+	xor	$Thh,$Zhh,$Zhh
+	ldwx	$nhi($Hhh),$Thh
+	shrpw	$Zlh,$Zll,4,$Zll
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	xor	$byte,$nlo,$nlo
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	and	$mask0xf0,$nlo,$nhi
+	extru	$Zhh,27,28,$Zhh
+	zdep	$nlo,27,4,$nlo
+	xor	$Tll,$Zll,$Zll
+	ldwx	$nlo($Hll),$Tll
+	xor	$Tlh,$Zlh,$Zlh
+	ldwx	$nlo($Hlh),$Tlh
+	xor	$rem,$Zhh,$Zhh
+	addib,uv -1,$cnt,L\$oop_ghash_pa1
+	xor	$Thl,$Zhl,$Zhl
+
+	zdep	$Zll,28,4,$rem
+	ldwx	$nlo($Hhl),$Thl
+	xor	$Thh,$Zhh,$Zhh
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zlh,$Zll,4,$Zll
+	ldwx	$nlo($Hhh),$Thh
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	xor	$Tll,$Zll,$Zll
+	ldwx	$nhi($Hll),$Tll
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	xor	$Tlh,$Zlh,$Zlh
+	ldwx	$nhi($Hlh),$Tlh
+	extru	$Zhh,27,28,$Zhh
+	xor	$rem,$Zhh,$Zhh
+	xor	$Thl,$Zhl,$Zhl
+	ldwx	$nhi($Hhl),$Thl
+	xor	$Thh,$Zhh,$Zhh
+	ldwx	$nhi($Hhh),$Thh
+	zdep	$Zll,28,4,$rem
+	ldwx	$rem($rem_4bit),$rem
+	shrpw	$Zlh,$Zll,4,$Zll
+	shrpw	$Zhl,$Zlh,4,$Zlh
+	shrpw	$Zhh,$Zhl,4,$Zhl
+	extru	$Zhh,27,28,$Zhh
+	xor	$Tll,$Zll,$Zll
+	xor	$Tlh,$Zlh,$Zlh
+	xor	$rem,$Zhh,$Zhh
+	stw	$Zll,12($Xi)
+	xor	$Thl,$Zhl,$Zhl
+	stw	$Zlh,8($Xi)
+	xor	$Thh,$Zhh,$Zhh
+	stw	$Zhl,4($Xi)
+	ldo	16($inp),$inp
+	stw	$Zhh,0($Xi)
+	comb,<>	$inp,$len,L\$outer_ghash_pa1
+	copy	$Zll,$nlo
+___
+$code.=<<___;
+L\$done_ghash
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2		; standard epilogue
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+___
+$code.=<<___ if ($SIZE_T==4);
+	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
+	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
+	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
+	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
+	$POP	`-$FRAME+8*$SIZE_T`(%sp),%r11
+___
+$code.=<<___;
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+
+	.ALIGN	64
+L\$rem_4bit
+	.WORD	`0x0000<<16`,0,`0x1C20<<16`,0,`0x3840<<16`,0,`0x2460<<16`,0
+	.WORD	`0x7080<<16`,0,`0x6CA0<<16`,0,`0x48C0<<16`,0,`0x54E0<<16`,0
+	.WORD	`0xE100<<16`,0,`0xFD20<<16`,0,`0xD940<<16`,0,`0xC560<<16`,0
+	.WORD	`0x9180<<16`,0,`0x8DA0<<16`,0,`0xA9C0<<16`,0,`0xB5E0<<16`,0
+	.STRINGZ "GHASH for PA-RISC, GRYPTOGAMS by <appro\@openssl.org>"
+	.ALIGN	64
+___
+
+# Explicitly encode PA-RISC 2.0 instructions used in this module, so
+# that it can be compiled with .LEVEL 1.0. It should be noted that I
+# wouldn't have to do this, if GNU assembler understood .ALLOW 2.0
+# directive...
+
+my $ldd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "ldd$mod\t$args";
+
+    if ($args =~ /%r([0-9]+)\(%r([0-9]+)\),%r([0-9]+)/)		# format 4
+    {	my $opcode=(0x03<<26)|($2<<21)|($1<<16)|(3<<6)|$3;
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    elsif ($args =~ /(\-?[0-9]+)\(%r([0-9]+)\),%r([0-9]+)/)	# format 5
+    {	my $opcode=(0x03<<26)|($2<<21)|(1<<12)|(3<<6)|$3;
+	$opcode|=(($1&0xF)<<17)|(($1&0x10)<<12);		# encode offset
+	$opcode|=(1<<5)  if ($mod =~ /^,m/);
+	$opcode|=(1<<13) if ($mod =~ /^,mb/);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $std = sub {
+  my ($mod,$args) = @_;
+  my $orig = "std$mod\t$args";
+
+    if ($args =~ /%r([0-9]+),(\-?[0-9]+)\(%r([0-9]+)\)/) # format 3 suffices
+    {	my $opcode=(0x1c<<26)|($3<<21)|($1<<16)|(($2&0x1FF8)<<1)|(($2>>13)&1);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $extrd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "extrd$mod\t$args";
+
+    # I only have ",u" completer, it's implicitly encoded...
+    if ($args =~ /%r([0-9]+),([0-9]+),([0-9]+),%r([0-9]+)/)	# format 15
+    {	my $opcode=(0x36<<26)|($1<<21)|($4<<16);
+	my $len=32-$3;
+	$opcode |= (($2&0x20)<<6)|(($2&0x1f)<<5);		# encode pos
+	$opcode |= (($len&0x20)<<7)|($len&0x1f);		# encode len
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    elsif ($args =~ /%r([0-9]+),%sar,([0-9]+),%r([0-9]+)/)	# format 12
+    {	my $opcode=(0x34<<26)|($1<<21)|($3<<16)|(2<<11)|(1<<9);
+	my $len=32-$2;
+	$opcode |= (($len&0x20)<<3)|($len&0x1f);		# encode len
+	$opcode |= (1<<13) if ($mod =~ /,\**=/);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $shrpd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "shrpd$mod\t$args";
+
+    if ($args =~ /%r([0-9]+),%r([0-9]+),([0-9]+),%r([0-9]+)/)	# format 14
+    {	my $opcode=(0x34<<26)|($2<<21)|($1<<16)|(1<<10)|$4;
+	my $cpos=63-$3;
+	$opcode |= (($cpos&0x20)<<6)|(($cpos&0x1f)<<5);		# encode sa
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    elsif ($args =~ /%r([0-9]+),%r([0-9]+),%sar,%r([0-9]+)/)	# format 11
+    {	sprintf "\t.WORD\t0x%08x\t; %s",
+		(0x34<<26)|($2<<21)|($1<<16)|(1<<9)|$3,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $depd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "depd$mod\t$args";
+
+    # I only have ",z" completer, it's impicitly encoded...
+    if ($args =~ /%r([0-9]+),([0-9]+),([0-9]+),%r([0-9]+)/)	# format 16
+    {	my $opcode=(0x3c<<26)|($4<<21)|($1<<16);
+    	my $cpos=63-$2;
+	my $len=32-$3;
+	$opcode |= (($cpos&0x20)<<6)|(($cpos&0x1f)<<5);		# encode pos
+	$opcode |= (($len&0x20)<<7)|($len&0x1f);		# encode len
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+sub assemble {
+  my ($mnemonic,$mod,$args)=@_;
+  my $opcode = eval("\$$mnemonic");
+
+    ref($opcode) eq 'CODE' ? &$opcode($mod,$args) : "\t$mnemonic$mod\t$args";
+}
+
+foreach (split("\n",$code)) {
+	s/\`([^\`]*)\`/eval $1/ge;
+	if ($SIZE_T==4) {
+		s/^\s+([a-z]+)([\S]*)\s+([\S]*)/&assemble($1,$2,$3)/e;
+		s/cmpb,\*/comb,/;
+		s/,\*/,/;
+	}
+	print $_,"\n";
+}
+
+close STDOUT;
diff --git a/crypto/modes/asm/ghash-s390x.pl b/crypto/modes/asm/ghash-s390x.pl
new file mode 100755
index 000000000000..6a40d5d89c0c
--- /dev/null
+++ b/crypto/modes/asm/ghash-s390x.pl
@@ -0,0 +1,262 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# September 2010.
+#
+# The module implements "4-bit" GCM GHASH function and underlying
+# single multiplication operation in GF(2^128). "4-bit" means that it
+# uses 256 bytes per-key table [+128 bytes shared table]. Performance
+# was measured to be ~18 cycles per processed byte on z10, which is
+# almost 40% better than gcc-generated code. It should be noted that
+# 18 cycles is worse result than expected: loop is scheduled for 12
+# and the result should be close to 12. In the lack of instruction-
+# level profiling data it's impossible to tell why...
+
+# November 2010.
+#
+# Adapt for -m31 build. If kernel supports what's called "highgprs"
+# feature on Linux [see /proc/cpuinfo], it's possible to use 64-bit
+# instructions and achieve "64-bit" performance even in 31-bit legacy
+# application context. The feature is not specific to any particular
+# processor, as long as it's "z-CPU". Latter implies that the code
+# remains z/Architecture specific. On z990 it was measured to perform
+# 2.8x better than 32-bit code generated by gcc 4.3.
+
+# March 2011.
+#
+# Support for hardware KIMD-GHASH is verified to produce correct
+# result and therefore is engaged. On z196 it was measured to process
+# 8KB buffer ~7 faster than software implementation. It's not as
+# impressive for smaller buffer sizes and for smallest 16-bytes buffer
+# it's actually almost 2 times slower. Which is the reason why
+# KIMD-GHASH is not used in gcm_gmult_4bit.
+
+$flavour = shift;
+
+if ($flavour =~ /3[12]/) {
+	$SIZE_T=4;
+	$g="";
+} else {
+	$SIZE_T=8;
+	$g="g";
+}
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$softonly=0;
+
+$Zhi="%r0";
+$Zlo="%r1";
+
+$Xi="%r2";	# argument block
+$Htbl="%r3";
+$inp="%r4";
+$len="%r5";
+
+$rem0="%r6";	# variables
+$rem1="%r7";
+$nlo="%r8";
+$nhi="%r9";
+$xi="%r10";
+$cnt="%r11";
+$tmp="%r12";
+$x78="%r13";
+$rem_4bit="%r14";
+
+$sp="%r15";
+
+$code.=<<___;
+.text
+
+.globl	gcm_gmult_4bit
+.align	32
+gcm_gmult_4bit:
+___
+$code.=<<___ if(!$softonly && 0);	# hardware is slow for single block...
+	larl	%r1,OPENSSL_s390xcap_P
+	lg	%r0,0(%r1)
+	tmhl	%r0,0x4000	# check for message-security-assist
+	jz	.Lsoft_gmult
+	lghi	%r0,0
+	la	%r1,16($sp)
+	.long	0xb93e0004	# kimd %r0,%r4
+	lg	%r1,24($sp)
+	tmhh	%r1,0x4000	# check for function 65
+	jz	.Lsoft_gmult
+	stg	%r0,16($sp)	# arrange 16 bytes of zero input
+	stg	%r0,24($sp)
+	lghi	%r0,65		# function 65
+	la	%r1,0($Xi)	# H lies right after Xi in gcm128_context
+	la	$inp,16($sp)
+	lghi	$len,16
+	.long	0xb93e0004	# kimd %r0,$inp
+	brc	1,.-4		# pay attention to "partial completion"
+	br	%r14
+.align	32
+.Lsoft_gmult:
+___
+$code.=<<___;
+	stm${g}	%r6,%r14,6*$SIZE_T($sp)
+
+	aghi	$Xi,-1
+	lghi	$len,1
+	lghi	$x78,`0xf<<3`
+	larl	$rem_4bit,rem_4bit
+
+	lg	$Zlo,8+1($Xi)		# Xi
+	j	.Lgmult_shortcut
+.type	gcm_gmult_4bit,\@function
+.size	gcm_gmult_4bit,(.-gcm_gmult_4bit)
+
+.globl	gcm_ghash_4bit
+.align	32
+gcm_ghash_4bit:
+___
+$code.=<<___ if(!$softonly);
+	larl	%r1,OPENSSL_s390xcap_P
+	lg	%r0,0(%r1)
+	tmhl	%r0,0x4000	# check for message-security-assist
+	jz	.Lsoft_ghash
+	lghi	%r0,0
+	la	%r1,16($sp)
+	.long	0xb93e0004	# kimd %r0,%r4
+	lg	%r1,24($sp)
+	tmhh	%r1,0x4000	# check for function 65
+	jz	.Lsoft_ghash
+	lghi	%r0,65		# function 65
+	la	%r1,0($Xi)	# H lies right after Xi in gcm128_context
+	.long	0xb93e0004	# kimd %r0,$inp
+	brc	1,.-4		# pay attention to "partial completion"
+	br	%r14
+.align	32
+.Lsoft_ghash:
+___
+$code.=<<___ if ($flavour =~ /3[12]/);
+	llgfr	$len,$len
+___
+$code.=<<___;
+	stm${g}	%r6,%r14,6*$SIZE_T($sp)
+
+	aghi	$Xi,-1
+	srlg	$len,$len,4
+	lghi	$x78,`0xf<<3`
+	larl	$rem_4bit,rem_4bit
+
+	lg	$Zlo,8+1($Xi)		# Xi
+	lg	$Zhi,0+1($Xi)
+	lghi	$tmp,0
+.Louter:
+	xg	$Zhi,0($inp)		# Xi ^= inp 
+	xg	$Zlo,8($inp)
+	xgr	$Zhi,$tmp
+	stg	$Zlo,8+1($Xi)
+	stg	$Zhi,0+1($Xi)
+
+.Lgmult_shortcut:
+	lghi	$tmp,0xf0
+	sllg	$nlo,$Zlo,4
+	srlg	$xi,$Zlo,8		# extract second byte
+	ngr	$nlo,$tmp
+	lgr	$nhi,$Zlo
+	lghi	$cnt,14
+	ngr	$nhi,$tmp
+
+	lg	$Zlo,8($nlo,$Htbl)
+	lg	$Zhi,0($nlo,$Htbl)
+
+	sllg	$nlo,$xi,4
+	sllg	$rem0,$Zlo,3
+	ngr	$nlo,$tmp
+	ngr	$rem0,$x78
+	ngr	$xi,$tmp
+
+	sllg	$tmp,$Zhi,60
+	srlg	$Zlo,$Zlo,4
+	srlg	$Zhi,$Zhi,4
+	xg	$Zlo,8($nhi,$Htbl)
+	xg	$Zhi,0($nhi,$Htbl)
+	lgr	$nhi,$xi
+	sllg	$rem1,$Zlo,3
+	xgr	$Zlo,$tmp
+	ngr	$rem1,$x78
+	j	.Lghash_inner
+.align	16
+.Lghash_inner:
+	srlg	$Zlo,$Zlo,4
+	sllg	$tmp,$Zhi,60
+	xg	$Zlo,8($nlo,$Htbl)
+	srlg	$Zhi,$Zhi,4
+	llgc	$xi,0($cnt,$Xi)
+	xg	$Zhi,0($nlo,$Htbl)
+	sllg	$nlo,$xi,4
+	xg	$Zhi,0($rem0,$rem_4bit)
+	nill	$nlo,0xf0
+	sllg	$rem0,$Zlo,3
+	xgr	$Zlo,$tmp
+	ngr	$rem0,$x78
+	nill	$xi,0xf0
+
+	sllg	$tmp,$Zhi,60
+	srlg	$Zlo,$Zlo,4
+	srlg	$Zhi,$Zhi,4
+	xg	$Zlo,8($nhi,$Htbl)
+	xg	$Zhi,0($nhi,$Htbl)
+	lgr	$nhi,$xi
+	xg	$Zhi,0($rem1,$rem_4bit)
+	sllg	$rem1,$Zlo,3
+	xgr	$Zlo,$tmp
+	ngr	$rem1,$x78
+	brct	$cnt,.Lghash_inner
+
+	sllg	$tmp,$Zhi,60
+	srlg	$Zlo,$Zlo,4
+	srlg	$Zhi,$Zhi,4
+	xg	$Zlo,8($nlo,$Htbl)
+	xg	$Zhi,0($nlo,$Htbl)
+	sllg	$xi,$Zlo,3
+	xg	$Zhi,0($rem0,$rem_4bit)
+	xgr	$Zlo,$tmp
+	ngr	$xi,$x78
+
+	sllg	$tmp,$Zhi,60
+	srlg	$Zlo,$Zlo,4
+	srlg	$Zhi,$Zhi,4
+	xg	$Zlo,8($nhi,$Htbl)
+	xg	$Zhi,0($nhi,$Htbl)
+	xgr	$Zlo,$tmp
+	xg	$Zhi,0($rem1,$rem_4bit)
+
+	lg	$tmp,0($xi,$rem_4bit)
+	la	$inp,16($inp)
+	sllg	$tmp,$tmp,4		# correct last rem_4bit[rem]
+	brctg	$len,.Louter
+
+	xgr	$Zhi,$tmp
+	stg	$Zlo,8+1($Xi)
+	stg	$Zhi,0+1($Xi)
+	lm${g}	%r6,%r14,6*$SIZE_T($sp)
+	br	%r14
+.type	gcm_ghash_4bit,\@function
+.size	gcm_ghash_4bit,(.-gcm_ghash_4bit)
+
+.align	64
+rem_4bit:
+	.long	`0x0000<<12`,0,`0x1C20<<12`,0,`0x3840<<12`,0,`0x2460<<12`,0
+	.long	`0x7080<<12`,0,`0x6CA0<<12`,0,`0x48C0<<12`,0,`0x54E0<<12`,0
+	.long	`0xE100<<12`,0,`0xFD20<<12`,0,`0xD940<<12`,0,`0xC560<<12`,0
+	.long	`0x9180<<12`,0,`0x8DA0<<12`,0,`0xA9C0<<12`,0,`0xB5E0<<12`,0
+.type	rem_4bit,\@object
+.size	rem_4bit,(.-rem_4bit)
+.string	"GHASH for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/modes/asm/ghash-sparcv9.pl b/crypto/modes/asm/ghash-sparcv9.pl
new file mode 100755
index 000000000000..70e7b044a3ec
--- /dev/null
+++ b/crypto/modes/asm/ghash-sparcv9.pl
@@ -0,0 +1,330 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# March 2010
+#
+# The module implements "4-bit" GCM GHASH function and underlying
+# single multiplication operation in GF(2^128). "4-bit" means that it
+# uses 256 bytes per-key table [+128 bytes shared table]. Performance
+# results are for streamed GHASH subroutine on UltraSPARC pre-Tx CPU
+# and are expressed in cycles per processed byte, less is better:
+#
+#		gcc 3.3.x	cc 5.2		this assembler
+#
+# 32-bit build	81.4		43.3		12.6	(+546%/+244%)
+# 64-bit build	20.2		21.2		12.6	(+60%/+68%)
+#
+# Here is data collected on UltraSPARC T1 system running Linux:
+#
+#		gcc 4.4.1			this assembler
+#
+# 32-bit build	566				50	(+1000%)
+# 64-bit build	56				50	(+12%)
+#
+# I don't quite understand why difference between 32-bit and 64-bit
+# compiler-generated code is so big. Compilers *were* instructed to
+# generate code for UltraSPARC and should have used 64-bit registers
+# for Z vector (see C code) even in 32-bit build... Oh well, it only
+# means more impressive improvement coefficients for this assembler
+# module;-) Loops are aggressively modulo-scheduled in respect to
+# references to input data and Z.hi updates to achieve 12 cycles
+# timing. To anchor to something else, sha1-sparcv9.pl spends 11.6
+# cycles to process one byte on UltraSPARC pre-Tx CPU and ~24 on T1.
+
+$bits=32;
+for (@ARGV)     { $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+if ($bits==64)  { $bias=2047; $frame=192; }
+else            { $bias=0;    $frame=112; }
+
+$output=shift;
+open STDOUT,">$output";
+
+$Zhi="%o0";	# 64-bit values
+$Zlo="%o1";
+$Thi="%o2";
+$Tlo="%o3";
+$rem="%o4";
+$tmp="%o5";
+
+$nhi="%l0";	# small values and pointers
+$nlo="%l1";
+$xi0="%l2";
+$xi1="%l3";
+$rem_4bit="%l4";
+$remi="%l5";
+$Htblo="%l6";
+$cnt="%l7";
+
+$Xi="%i0";	# input argument block
+$Htbl="%i1";
+$inp="%i2";
+$len="%i3";
+
+$code.=<<___;
+.section	".text",#alloc,#execinstr
+
+.align	64
+rem_4bit:
+	.long	`0x0000<<16`,0,`0x1C20<<16`,0,`0x3840<<16`,0,`0x2460<<16`,0
+	.long	`0x7080<<16`,0,`0x6CA0<<16`,0,`0x48C0<<16`,0,`0x54E0<<16`,0
+	.long	`0xE100<<16`,0,`0xFD20<<16`,0,`0xD940<<16`,0,`0xC560<<16`,0
+	.long	`0x9180<<16`,0,`0x8DA0<<16`,0,`0xA9C0<<16`,0,`0xB5E0<<16`,0
+.type	rem_4bit,#object
+.size	rem_4bit,(.-rem_4bit)
+
+.globl	gcm_ghash_4bit
+.align	32
+gcm_ghash_4bit:
+	save	%sp,-$frame,%sp
+	ldub	[$inp+15],$nlo
+	ldub	[$Xi+15],$xi0
+	ldub	[$Xi+14],$xi1
+	add	$len,$inp,$len
+	add	$Htbl,8,$Htblo
+
+1:	call	.+8
+	add	%o7,rem_4bit-1b,$rem_4bit
+
+.Louter:
+	xor	$xi0,$nlo,$nlo
+	and	$nlo,0xf0,$nhi
+	and	$nlo,0x0f,$nlo
+	sll	$nlo,4,$nlo
+	ldx	[$Htblo+$nlo],$Zlo
+	ldx	[$Htbl+$nlo],$Zhi
+
+	ldub	[$inp+14],$nlo
+
+	ldx	[$Htblo+$nhi],$Tlo
+	and	$Zlo,0xf,$remi
+	ldx	[$Htbl+$nhi],$Thi
+	sll	$remi,3,$remi
+	ldx	[$rem_4bit+$remi],$rem
+	srlx	$Zlo,4,$Zlo
+	mov	13,$cnt
+	sllx	$Zhi,60,$tmp
+	xor	$Tlo,$Zlo,$Zlo
+	srlx	$Zhi,4,$Zhi
+	xor	$Zlo,$tmp,$Zlo
+
+	xor	$xi1,$nlo,$nlo
+	and	$Zlo,0xf,$remi
+	and	$nlo,0xf0,$nhi
+	and	$nlo,0x0f,$nlo
+	ba	.Lghash_inner
+	sll	$nlo,4,$nlo
+.align	32
+.Lghash_inner:
+	ldx	[$Htblo+$nlo],$Tlo
+	sll	$remi,3,$remi
+	xor	$Thi,$Zhi,$Zhi
+	ldx	[$Htbl+$nlo],$Thi
+	srlx	$Zlo,4,$Zlo
+	xor	$rem,$Zhi,$Zhi
+	ldx	[$rem_4bit+$remi],$rem
+	sllx	$Zhi,60,$tmp
+	xor	$Tlo,$Zlo,$Zlo
+	ldub	[$inp+$cnt],$nlo
+	srlx	$Zhi,4,$Zhi
+	xor	$Zlo,$tmp,$Zlo
+	ldub	[$Xi+$cnt],$xi1
+	xor	$Thi,$Zhi,$Zhi
+	and	$Zlo,0xf,$remi
+
+	ldx	[$Htblo+$nhi],$Tlo
+	sll	$remi,3,$remi
+	xor	$rem,$Zhi,$Zhi
+	ldx	[$Htbl+$nhi],$Thi
+	srlx	$Zlo,4,$Zlo
+	ldx	[$rem_4bit+$remi],$rem
+	sllx	$Zhi,60,$tmp
+	xor	$xi1,$nlo,$nlo
+	srlx	$Zhi,4,$Zhi
+	and	$nlo,0xf0,$nhi
+	addcc	$cnt,-1,$cnt
+	xor	$Zlo,$tmp,$Zlo
+	and	$nlo,0x0f,$nlo
+	xor	$Tlo,$Zlo,$Zlo
+	sll	$nlo,4,$nlo
+	blu	.Lghash_inner
+	and	$Zlo,0xf,$remi
+
+	ldx	[$Htblo+$nlo],$Tlo
+	sll	$remi,3,$remi
+	xor	$Thi,$Zhi,$Zhi
+	ldx	[$Htbl+$nlo],$Thi
+	srlx	$Zlo,4,$Zlo
+	xor	$rem,$Zhi,$Zhi
+	ldx	[$rem_4bit+$remi],$rem
+	sllx	$Zhi,60,$tmp
+	xor	$Tlo,$Zlo,$Zlo
+	srlx	$Zhi,4,$Zhi
+	xor	$Zlo,$tmp,$Zlo
+	xor	$Thi,$Zhi,$Zhi
+
+	add	$inp,16,$inp
+	cmp	$inp,$len
+	be,pn	`$bits==64?"%xcc":"%icc"`,.Ldone
+	and	$Zlo,0xf,$remi
+
+	ldx	[$Htblo+$nhi],$Tlo
+	sll	$remi,3,$remi
+	xor	$rem,$Zhi,$Zhi
+	ldx	[$Htbl+$nhi],$Thi
+	srlx	$Zlo,4,$Zlo
+	ldx	[$rem_4bit+$remi],$rem
+	sllx	$Zhi,60,$tmp
+	xor	$Tlo,$Zlo,$Zlo
+	ldub	[$inp+15],$nlo
+	srlx	$Zhi,4,$Zhi
+	xor	$Zlo,$tmp,$Zlo
+	xor	$Thi,$Zhi,$Zhi
+	stx	$Zlo,[$Xi+8]
+	xor	$rem,$Zhi,$Zhi
+	stx	$Zhi,[$Xi]
+	srl	$Zlo,8,$xi1
+	and	$Zlo,0xff,$xi0
+	ba	.Louter
+	and	$xi1,0xff,$xi1
+.align	32
+.Ldone:
+	ldx	[$Htblo+$nhi],$Tlo
+	sll	$remi,3,$remi
+	xor	$rem,$Zhi,$Zhi
+	ldx	[$Htbl+$nhi],$Thi
+	srlx	$Zlo,4,$Zlo
+	ldx	[$rem_4bit+$remi],$rem
+	sllx	$Zhi,60,$tmp
+	xor	$Tlo,$Zlo,$Zlo
+	srlx	$Zhi,4,$Zhi
+	xor	$Zlo,$tmp,$Zlo
+	xor	$Thi,$Zhi,$Zhi
+	stx	$Zlo,[$Xi+8]
+	xor	$rem,$Zhi,$Zhi
+	stx	$Zhi,[$Xi]
+
+	ret
+	restore
+.type	gcm_ghash_4bit,#function
+.size	gcm_ghash_4bit,(.-gcm_ghash_4bit)
+___
+
+undef $inp;
+undef $len;
+
+$code.=<<___;
+.globl	gcm_gmult_4bit
+.align	32
+gcm_gmult_4bit:
+	save	%sp,-$frame,%sp
+	ldub	[$Xi+15],$nlo
+	add	$Htbl,8,$Htblo
+
+1:	call	.+8
+	add	%o7,rem_4bit-1b,$rem_4bit
+
+	and	$nlo,0xf0,$nhi
+	and	$nlo,0x0f,$nlo
+	sll	$nlo,4,$nlo
+	ldx	[$Htblo+$nlo],$Zlo
+	ldx	[$Htbl+$nlo],$Zhi
+
+	ldub	[$Xi+14],$nlo
+
+	ldx	[$Htblo+$nhi],$Tlo
+	and	$Zlo,0xf,$remi
+	ldx	[$Htbl+$nhi],$Thi
+	sll	$remi,3,$remi
+	ldx	[$rem_4bit+$remi],$rem
+	srlx	$Zlo,4,$Zlo
+	mov	13,$cnt
+	sllx	$Zhi,60,$tmp
+	xor	$Tlo,$Zlo,$Zlo
+	srlx	$Zhi,4,$Zhi
+	xor	$Zlo,$tmp,$Zlo
+
+	and	$Zlo,0xf,$remi
+	and	$nlo,0xf0,$nhi
+	and	$nlo,0x0f,$nlo
+	ba	.Lgmult_inner
+	sll	$nlo,4,$nlo
+.align	32
+.Lgmult_inner:
+	ldx	[$Htblo+$nlo],$Tlo
+	sll	$remi,3,$remi
+	xor	$Thi,$Zhi,$Zhi
+	ldx	[$Htbl+$nlo],$Thi
+	srlx	$Zlo,4,$Zlo
+	xor	$rem,$Zhi,$Zhi
+	ldx	[$rem_4bit+$remi],$rem
+	sllx	$Zhi,60,$tmp
+	xor	$Tlo,$Zlo,$Zlo
+	ldub	[$Xi+$cnt],$nlo
+	srlx	$Zhi,4,$Zhi
+	xor	$Zlo,$tmp,$Zlo
+	xor	$Thi,$Zhi,$Zhi
+	and	$Zlo,0xf,$remi
+
+	ldx	[$Htblo+$nhi],$Tlo
+	sll	$remi,3,$remi
+	xor	$rem,$Zhi,$Zhi
+	ldx	[$Htbl+$nhi],$Thi
+	srlx	$Zlo,4,$Zlo
+	ldx	[$rem_4bit+$remi],$rem
+	sllx	$Zhi,60,$tmp
+	srlx	$Zhi,4,$Zhi
+	and	$nlo,0xf0,$nhi
+	addcc	$cnt,-1,$cnt
+	xor	$Zlo,$tmp,$Zlo
+	and	$nlo,0x0f,$nlo
+	xor	$Tlo,$Zlo,$Zlo
+	sll	$nlo,4,$nlo
+	blu	.Lgmult_inner
+	and	$Zlo,0xf,$remi
+
+	ldx	[$Htblo+$nlo],$Tlo
+	sll	$remi,3,$remi
+	xor	$Thi,$Zhi,$Zhi
+	ldx	[$Htbl+$nlo],$Thi
+	srlx	$Zlo,4,$Zlo
+	xor	$rem,$Zhi,$Zhi
+	ldx	[$rem_4bit+$remi],$rem
+	sllx	$Zhi,60,$tmp
+	xor	$Tlo,$Zlo,$Zlo
+	srlx	$Zhi,4,$Zhi
+	xor	$Zlo,$tmp,$Zlo
+	xor	$Thi,$Zhi,$Zhi
+	and	$Zlo,0xf,$remi
+
+	ldx	[$Htblo+$nhi],$Tlo
+	sll	$remi,3,$remi
+	xor	$rem,$Zhi,$Zhi
+	ldx	[$Htbl+$nhi],$Thi
+	srlx	$Zlo,4,$Zlo
+	ldx	[$rem_4bit+$remi],$rem
+	sllx	$Zhi,60,$tmp
+	xor	$Tlo,$Zlo,$Zlo
+	srlx	$Zhi,4,$Zhi
+	xor	$Zlo,$tmp,$Zlo
+	xor	$Thi,$Zhi,$Zhi
+	stx	$Zlo,[$Xi+8]
+	xor	$rem,$Zhi,$Zhi
+	stx	$Zhi,[$Xi]
+
+	ret
+	restore
+.type	gcm_gmult_4bit,#function
+.size	gcm_gmult_4bit,(.-gcm_gmult_4bit)
+.asciz	"GHASH for SPARCv9, CRYPTOGAMS by <appro\@openssl.org>"
+.align	4
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/modes/asm/ghash-x86.pl b/crypto/modes/asm/ghash-x86.pl
new file mode 100755
index 000000000000..6b09669d474a
--- /dev/null
+++ b/crypto/modes/asm/ghash-x86.pl
@@ -0,0 +1,1342 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# March, May, June 2010
+#
+# The module implements "4-bit" GCM GHASH function and underlying
+# single multiplication operation in GF(2^128). "4-bit" means that it
+# uses 256 bytes per-key table [+64/128 bytes fixed table]. It has two
+# code paths: vanilla x86 and vanilla MMX. Former will be executed on
+# 486 and Pentium, latter on all others. MMX GHASH features so called
+# "528B" variant of "4-bit" method utilizing additional 256+16 bytes
+# of per-key storage [+512 bytes shared table]. Performance results
+# are for streamed GHASH subroutine and are expressed in cycles per
+# processed byte, less is better:
+#
+#		gcc 2.95.3(*)	MMX assembler	x86 assembler
+#
+# Pentium	105/111(**)	-		50
+# PIII		68 /75		12.2		24
+# P4		125/125		17.8		84(***)
+# Opteron	66 /70		10.1		30
+# Core2		54 /67		8.4		18
+#
+# (*)	gcc 3.4.x was observed to generate few percent slower code,
+#	which is one of reasons why 2.95.3 results were chosen,
+#	another reason is lack of 3.4.x results for older CPUs;
+#	comparison with MMX results is not completely fair, because C
+#	results are for vanilla "256B" implementation, while
+#	assembler results are for "528B";-)
+# (**)	second number is result for code compiled with -fPIC flag,
+#	which is actually more relevant, because assembler code is
+#	position-independent;
+# (***)	see comment in non-MMX routine for further details;
+#
+# To summarize, it's >2-5 times faster than gcc-generated code. To
+# anchor it to something else SHA1 assembler processes one byte in
+# 11-13 cycles on contemporary x86 cores. As for choice of MMX in
+# particular, see comment at the end of the file...
+
+# May 2010
+#
+# Add PCLMULQDQ version performing at 2.10 cycles per processed byte.
+# The question is how close is it to theoretical limit? The pclmulqdq
+# instruction latency appears to be 14 cycles and there can't be more
+# than 2 of them executing at any given time. This means that single
+# Karatsuba multiplication would take 28 cycles *plus* few cycles for
+# pre- and post-processing. Then multiplication has to be followed by
+# modulo-reduction. Given that aggregated reduction method [see
+# "Carry-less Multiplication and Its Usage for Computing the GCM Mode"
+# white paper by Intel] allows you to perform reduction only once in
+# a while we can assume that asymptotic performance can be estimated
+# as (28+Tmod/Naggr)/16, where Tmod is time to perform reduction
+# and Naggr is the aggregation factor.
+#
+# Before we proceed to this implementation let's have closer look at
+# the best-performing code suggested by Intel in their white paper.
+# By tracing inter-register dependencies Tmod is estimated as ~19
+# cycles and Naggr chosen by Intel is 4, resulting in 2.05 cycles per
+# processed byte. As implied, this is quite optimistic estimate,
+# because it does not account for Karatsuba pre- and post-processing,
+# which for a single multiplication is ~5 cycles. Unfortunately Intel
+# does not provide performance data for GHASH alone. But benchmarking
+# AES_GCM_encrypt ripped out of Fig. 15 of the white paper with aadt
+# alone resulted in 2.46 cycles per byte of out 16KB buffer. Note that
+# the result accounts even for pre-computing of degrees of the hash
+# key H, but its portion is negligible at 16KB buffer size.
+#
+# Moving on to the implementation in question. Tmod is estimated as
+# ~13 cycles and Naggr is 2, giving asymptotic performance of ...
+# 2.16. How is it possible that measured performance is better than
+# optimistic theoretical estimate? There is one thing Intel failed
+# to recognize. By serializing GHASH with CTR in same subroutine
+# former's performance is really limited to above (Tmul + Tmod/Naggr)
+# equation. But if GHASH procedure is detached, the modulo-reduction
+# can be interleaved with Naggr-1 multiplications at instruction level
+# and under ideal conditions even disappear from the equation. So that
+# optimistic theoretical estimate for this implementation is ...
+# 28/16=1.75, and not 2.16. Well, it's probably way too optimistic,
+# at least for such small Naggr. I'd argue that (28+Tproc/Naggr),
+# where Tproc is time required for Karatsuba pre- and post-processing,
+# is more realistic estimate. In this case it gives ... 1.91 cycles.
+# Or in other words, depending on how well we can interleave reduction
+# and one of the two multiplications the performance should be betwen
+# 1.91 and 2.16. As already mentioned, this implementation processes
+# one byte out of 8KB buffer in 2.10 cycles, while x86_64 counterpart
+# - in 2.02. x86_64 performance is better, because larger register
+# bank allows to interleave reduction and multiplication better.
+#
+# Does it make sense to increase Naggr? To start with it's virtually
+# impossible in 32-bit mode, because of limited register bank
+# capacity. Otherwise improvement has to be weighed agiainst slower
+# setup, as well as code size and complexity increase. As even
+# optimistic estimate doesn't promise 30% performance improvement,
+# there are currently no plans to increase Naggr.
+#
+# Special thanks to David Woodhouse <dwmw2@infradead.org> for
+# providing access to a Westmere-based system on behalf of Intel
+# Open Source Technology Centre.
+
+# January 2010
+#
+# Tweaked to optimize transitions between integer and FP operations
+# on same XMM register, PCLMULQDQ subroutine was measured to process
+# one byte in 2.07 cycles on Sandy Bridge, and in 2.12 - on Westmere.
+# The minor regression on Westmere is outweighed by ~15% improvement
+# on Sandy Bridge. Strangely enough attempt to modify 64-bit code in
+# similar manner resulted in almost 20% degradation on Sandy Bridge,
+# where original 64-bit code processes one byte in 1.95 cycles.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"ghash-x86.pl",$x86only = $ARGV[$#ARGV] eq "386");
+
+$sse2=0;
+for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); }
+
+($Zhh,$Zhl,$Zlh,$Zll) = ("ebp","edx","ecx","ebx");
+$inp  = "edi";
+$Htbl = "esi";
+
+$unroll = 0;	# Affects x86 loop. Folded loop performs ~7% worse
+		# than unrolled, which has to be weighted against
+		# 2.5x x86-specific code size reduction.
+
+sub x86_loop {
+    my $off = shift;
+    my $rem = "eax";
+
+	&mov	($Zhh,&DWP(4,$Htbl,$Zll));
+	&mov	($Zhl,&DWP(0,$Htbl,$Zll));
+	&mov	($Zlh,&DWP(12,$Htbl,$Zll));
+	&mov	($Zll,&DWP(8,$Htbl,$Zll));
+	&xor	($rem,$rem);	# avoid partial register stalls on PIII
+
+	# shrd practically kills P4, 2.5x deterioration, but P4 has
+	# MMX code-path to execute. shrd runs tad faster [than twice
+	# the shifts, move's and or's] on pre-MMX Pentium (as well as
+	# PIII and Core2), *but* minimizes code size, spares register
+	# and thus allows to fold the loop...
+	if (!$unroll) {
+	my $cnt = $inp;
+	&mov	($cnt,15);
+	&jmp	(&label("x86_loop"));
+	&set_label("x86_loop",16);
+	    for($i=1;$i<=2;$i++) {
+		&mov	(&LB($rem),&LB($Zll));
+		&shrd	($Zll,$Zlh,4);
+		&and	(&LB($rem),0xf);
+		&shrd	($Zlh,$Zhl,4);
+		&shrd	($Zhl,$Zhh,4);
+		&shr	($Zhh,4);
+		&xor	($Zhh,&DWP($off+16,"esp",$rem,4));
+
+		&mov	(&LB($rem),&BP($off,"esp",$cnt));
+		if ($i&1) {
+			&and	(&LB($rem),0xf0);
+		} else {
+			&shl	(&LB($rem),4);
+		}
+
+		&xor	($Zll,&DWP(8,$Htbl,$rem));
+		&xor	($Zlh,&DWP(12,$Htbl,$rem));
+		&xor	($Zhl,&DWP(0,$Htbl,$rem));
+		&xor	($Zhh,&DWP(4,$Htbl,$rem));
+
+		if ($i&1) {
+			&dec	($cnt);
+			&js	(&label("x86_break"));
+		} else {
+			&jmp	(&label("x86_loop"));
+		}
+	    }
+	&set_label("x86_break",16);
+	} else {
+	    for($i=1;$i<32;$i++) {
+		&comment($i);
+		&mov	(&LB($rem),&LB($Zll));
+		&shrd	($Zll,$Zlh,4);
+		&and	(&LB($rem),0xf);
+		&shrd	($Zlh,$Zhl,4);
+		&shrd	($Zhl,$Zhh,4);
+		&shr	($Zhh,4);
+		&xor	($Zhh,&DWP($off+16,"esp",$rem,4));
+
+		if ($i&1) {
+			&mov	(&LB($rem),&BP($off+15-($i>>1),"esp"));
+			&and	(&LB($rem),0xf0);
+		} else {
+			&mov	(&LB($rem),&BP($off+15-($i>>1),"esp"));
+			&shl	(&LB($rem),4);
+		}
+
+		&xor	($Zll,&DWP(8,$Htbl,$rem));
+		&xor	($Zlh,&DWP(12,$Htbl,$rem));
+		&xor	($Zhl,&DWP(0,$Htbl,$rem));
+		&xor	($Zhh,&DWP(4,$Htbl,$rem));
+	    }
+	}
+	&bswap	($Zll);
+	&bswap	($Zlh);
+	&bswap	($Zhl);
+	if (!$x86only) {
+		&bswap	($Zhh);
+	} else {
+		&mov	("eax",$Zhh);
+		&bswap	("eax");
+		&mov	($Zhh,"eax");
+	}
+}
+
+if ($unroll) {
+    &function_begin_B("_x86_gmult_4bit_inner");
+	&x86_loop(4);
+	&ret	();
+    &function_end_B("_x86_gmult_4bit_inner");
+}
+
+sub deposit_rem_4bit {
+    my $bias = shift;
+
+	&mov	(&DWP($bias+0, "esp"),0x0000<<16);
+	&mov	(&DWP($bias+4, "esp"),0x1C20<<16);
+	&mov	(&DWP($bias+8, "esp"),0x3840<<16);
+	&mov	(&DWP($bias+12,"esp"),0x2460<<16);
+	&mov	(&DWP($bias+16,"esp"),0x7080<<16);
+	&mov	(&DWP($bias+20,"esp"),0x6CA0<<16);
+	&mov	(&DWP($bias+24,"esp"),0x48C0<<16);
+	&mov	(&DWP($bias+28,"esp"),0x54E0<<16);
+	&mov	(&DWP($bias+32,"esp"),0xE100<<16);
+	&mov	(&DWP($bias+36,"esp"),0xFD20<<16);
+	&mov	(&DWP($bias+40,"esp"),0xD940<<16);
+	&mov	(&DWP($bias+44,"esp"),0xC560<<16);
+	&mov	(&DWP($bias+48,"esp"),0x9180<<16);
+	&mov	(&DWP($bias+52,"esp"),0x8DA0<<16);
+	&mov	(&DWP($bias+56,"esp"),0xA9C0<<16);
+	&mov	(&DWP($bias+60,"esp"),0xB5E0<<16);
+}
+
+$suffix = $x86only ? "" : "_x86";
+
+&function_begin("gcm_gmult_4bit".$suffix);
+	&stack_push(16+4+1);			# +1 for stack alignment
+	&mov	($inp,&wparam(0));		# load Xi
+	&mov	($Htbl,&wparam(1));		# load Htable
+
+	&mov	($Zhh,&DWP(0,$inp));		# load Xi[16]
+	&mov	($Zhl,&DWP(4,$inp));
+	&mov	($Zlh,&DWP(8,$inp));
+	&mov	($Zll,&DWP(12,$inp));
+
+	&deposit_rem_4bit(16);
+
+	&mov	(&DWP(0,"esp"),$Zhh);		# copy Xi[16] on stack
+	&mov	(&DWP(4,"esp"),$Zhl);
+	&mov	(&DWP(8,"esp"),$Zlh);
+	&mov	(&DWP(12,"esp"),$Zll);
+	&shr	($Zll,20);
+	&and	($Zll,0xf0);
+
+	if ($unroll) {
+		&call	("_x86_gmult_4bit_inner");
+	} else {
+		&x86_loop(0);
+		&mov	($inp,&wparam(0));
+	}
+
+	&mov	(&DWP(12,$inp),$Zll);
+	&mov	(&DWP(8,$inp),$Zlh);
+	&mov	(&DWP(4,$inp),$Zhl);
+	&mov	(&DWP(0,$inp),$Zhh);
+	&stack_pop(16+4+1);
+&function_end("gcm_gmult_4bit".$suffix);
+
+&function_begin("gcm_ghash_4bit".$suffix);
+	&stack_push(16+4+1);			# +1 for 64-bit alignment
+	&mov	($Zll,&wparam(0));		# load Xi
+	&mov	($Htbl,&wparam(1));		# load Htable
+	&mov	($inp,&wparam(2));		# load in
+	&mov	("ecx",&wparam(3));		# load len
+	&add	("ecx",$inp);
+	&mov	(&wparam(3),"ecx");
+
+	&mov	($Zhh,&DWP(0,$Zll));		# load Xi[16]
+	&mov	($Zhl,&DWP(4,$Zll));
+	&mov	($Zlh,&DWP(8,$Zll));
+	&mov	($Zll,&DWP(12,$Zll));
+
+	&deposit_rem_4bit(16);
+
+    &set_label("x86_outer_loop",16);
+	&xor	($Zll,&DWP(12,$inp));		# xor with input
+	&xor	($Zlh,&DWP(8,$inp));
+	&xor	($Zhl,&DWP(4,$inp));
+	&xor	($Zhh,&DWP(0,$inp));
+	&mov	(&DWP(12,"esp"),$Zll);		# dump it on stack
+	&mov	(&DWP(8,"esp"),$Zlh);
+	&mov	(&DWP(4,"esp"),$Zhl);
+	&mov	(&DWP(0,"esp"),$Zhh);
+
+	&shr	($Zll,20);
+	&and	($Zll,0xf0);
+
+	if ($unroll) {
+		&call	("_x86_gmult_4bit_inner");
+	} else {
+		&x86_loop(0);
+		&mov	($inp,&wparam(2));
+	}
+	&lea	($inp,&DWP(16,$inp));
+	&cmp	($inp,&wparam(3));
+	&mov	(&wparam(2),$inp)	if (!$unroll);
+	&jb	(&label("x86_outer_loop"));
+
+	&mov	($inp,&wparam(0));	# load Xi
+	&mov	(&DWP(12,$inp),$Zll);
+	&mov	(&DWP(8,$inp),$Zlh);
+	&mov	(&DWP(4,$inp),$Zhl);
+	&mov	(&DWP(0,$inp),$Zhh);
+	&stack_pop(16+4+1);
+&function_end("gcm_ghash_4bit".$suffix);
+
+if (!$x86only) {{{
+
+&static_label("rem_4bit");
+
+if (!$sse2) {{	# pure-MMX "May" version...
+
+$S=12;		# shift factor for rem_4bit
+
+&function_begin_B("_mmx_gmult_4bit_inner");
+# MMX version performs 3.5 times better on P4 (see comment in non-MMX
+# routine for further details), 100% better on Opteron, ~70% better
+# on Core2 and PIII... In other words effort is considered to be well
+# spent... Since initial release the loop was unrolled in order to
+# "liberate" register previously used as loop counter. Instead it's
+# used to optimize critical path in 'Z.hi ^= rem_4bit[Z.lo&0xf]'.
+# The path involves move of Z.lo from MMX to integer register,
+# effective address calculation and finally merge of value to Z.hi.
+# Reference to rem_4bit is scheduled so late that I had to >>4
+# rem_4bit elements. This resulted in 20-45% procent improvement
+# on contemporary µ-archs.
+{
+    my $cnt;
+    my $rem_4bit = "eax";
+    my @rem = ($Zhh,$Zll);
+    my $nhi = $Zhl;
+    my $nlo = $Zlh;
+
+    my ($Zlo,$Zhi) = ("mm0","mm1");
+    my $tmp = "mm2";
+
+	&xor	($nlo,$nlo);	# avoid partial register stalls on PIII
+	&mov	($nhi,$Zll);
+	&mov	(&LB($nlo),&LB($nhi));
+	&shl	(&LB($nlo),4);
+	&and	($nhi,0xf0);
+	&movq	($Zlo,&QWP(8,$Htbl,$nlo));
+	&movq	($Zhi,&QWP(0,$Htbl,$nlo));
+	&movd	($rem[0],$Zlo);
+
+	for ($cnt=28;$cnt>=-2;$cnt--) {
+	    my $odd = $cnt&1;
+	    my $nix = $odd ? $nlo : $nhi;
+
+		&shl	(&LB($nlo),4)			if ($odd);
+		&psrlq	($Zlo,4);
+		&movq	($tmp,$Zhi);
+		&psrlq	($Zhi,4);
+		&pxor	($Zlo,&QWP(8,$Htbl,$nix));
+		&mov	(&LB($nlo),&BP($cnt/2,$inp))	if (!$odd && $cnt>=0);
+		&psllq	($tmp,60);
+		&and	($nhi,0xf0)			if ($odd);
+		&pxor	($Zhi,&QWP(0,$rem_4bit,$rem[1],8)) if ($cnt<28);
+		&and	($rem[0],0xf);
+		&pxor	($Zhi,&QWP(0,$Htbl,$nix));
+		&mov	($nhi,$nlo)			if (!$odd && $cnt>=0);
+		&movd	($rem[1],$Zlo);
+		&pxor	($Zlo,$tmp);
+
+		push	(@rem,shift(@rem));		# "rotate" registers
+	}
+
+	&mov	($inp,&DWP(4,$rem_4bit,$rem[1],8));	# last rem_4bit[rem]
+
+	&psrlq	($Zlo,32);	# lower part of Zlo is already there
+	&movd	($Zhl,$Zhi);
+	&psrlq	($Zhi,32);
+	&movd	($Zlh,$Zlo);
+	&movd	($Zhh,$Zhi);
+	&shl	($inp,4);	# compensate for rem_4bit[i] being >>4
+
+	&bswap	($Zll);
+	&bswap	($Zhl);
+	&bswap	($Zlh);
+	&xor	($Zhh,$inp);
+	&bswap	($Zhh);
+
+	&ret	();
+}
+&function_end_B("_mmx_gmult_4bit_inner");
+
+&function_begin("gcm_gmult_4bit_mmx");
+	&mov	($inp,&wparam(0));	# load Xi
+	&mov	($Htbl,&wparam(1));	# load Htable
+
+	&call	(&label("pic_point"));
+	&set_label("pic_point");
+	&blindpop("eax");
+	&lea	("eax",&DWP(&label("rem_4bit")."-".&label("pic_point"),"eax"));
+
+	&movz	($Zll,&BP(15,$inp));
+
+	&call	("_mmx_gmult_4bit_inner");
+
+	&mov	($inp,&wparam(0));	# load Xi
+	&emms	();
+	&mov	(&DWP(12,$inp),$Zll);
+	&mov	(&DWP(4,$inp),$Zhl);
+	&mov	(&DWP(8,$inp),$Zlh);
+	&mov	(&DWP(0,$inp),$Zhh);
+&function_end("gcm_gmult_4bit_mmx");
+
+# Streamed version performs 20% better on P4, 7% on Opteron,
+# 10% on Core2 and PIII...
+&function_begin("gcm_ghash_4bit_mmx");
+	&mov	($Zhh,&wparam(0));	# load Xi
+	&mov	($Htbl,&wparam(1));	# load Htable
+	&mov	($inp,&wparam(2));	# load in
+	&mov	($Zlh,&wparam(3));	# load len
+
+	&call	(&label("pic_point"));
+	&set_label("pic_point");
+	&blindpop("eax");
+	&lea	("eax",&DWP(&label("rem_4bit")."-".&label("pic_point"),"eax"));
+
+	&add	($Zlh,$inp);
+	&mov	(&wparam(3),$Zlh);	# len to point at the end of input
+	&stack_push(4+1);		# +1 for stack alignment
+
+	&mov	($Zll,&DWP(12,$Zhh));	# load Xi[16]
+	&mov	($Zhl,&DWP(4,$Zhh));
+	&mov	($Zlh,&DWP(8,$Zhh));
+	&mov	($Zhh,&DWP(0,$Zhh));
+	&jmp	(&label("mmx_outer_loop"));
+
+    &set_label("mmx_outer_loop",16);
+	&xor	($Zll,&DWP(12,$inp));
+	&xor	($Zhl,&DWP(4,$inp));
+	&xor	($Zlh,&DWP(8,$inp));
+	&xor	($Zhh,&DWP(0,$inp));
+	&mov	(&wparam(2),$inp);
+	&mov	(&DWP(12,"esp"),$Zll);
+	&mov	(&DWP(4,"esp"),$Zhl);
+	&mov	(&DWP(8,"esp"),$Zlh);
+	&mov	(&DWP(0,"esp"),$Zhh);
+
+	&mov	($inp,"esp");
+	&shr	($Zll,24);
+
+	&call	("_mmx_gmult_4bit_inner");
+
+	&mov	($inp,&wparam(2));
+	&lea	($inp,&DWP(16,$inp));
+	&cmp	($inp,&wparam(3));
+	&jb	(&label("mmx_outer_loop"));
+
+	&mov	($inp,&wparam(0));	# load Xi
+	&emms	();
+	&mov	(&DWP(12,$inp),$Zll);
+	&mov	(&DWP(4,$inp),$Zhl);
+	&mov	(&DWP(8,$inp),$Zlh);
+	&mov	(&DWP(0,$inp),$Zhh);
+
+	&stack_pop(4+1);
+&function_end("gcm_ghash_4bit_mmx");
+
+}} else {{	# "June" MMX version...
+		# ... has slower "April" gcm_gmult_4bit_mmx with folded
+		# loop. This is done to conserve code size...
+$S=16;		# shift factor for rem_4bit
+
+sub mmx_loop() {
+# MMX version performs 2.8 times better on P4 (see comment in non-MMX
+# routine for further details), 40% better on Opteron and Core2, 50%
+# better on PIII... In other words effort is considered to be well
+# spent...
+    my $inp = shift;
+    my $rem_4bit = shift;
+    my $cnt = $Zhh;
+    my $nhi = $Zhl;
+    my $nlo = $Zlh;
+    my $rem = $Zll;
+
+    my ($Zlo,$Zhi) = ("mm0","mm1");
+    my $tmp = "mm2";
+
+	&xor	($nlo,$nlo);	# avoid partial register stalls on PIII
+	&mov	($nhi,$Zll);
+	&mov	(&LB($nlo),&LB($nhi));
+	&mov	($cnt,14);
+	&shl	(&LB($nlo),4);
+	&and	($nhi,0xf0);
+	&movq	($Zlo,&QWP(8,$Htbl,$nlo));
+	&movq	($Zhi,&QWP(0,$Htbl,$nlo));
+	&movd	($rem,$Zlo);
+	&jmp	(&label("mmx_loop"));
+
+    &set_label("mmx_loop",16);
+	&psrlq	($Zlo,4);
+	&and	($rem,0xf);
+	&movq	($tmp,$Zhi);
+	&psrlq	($Zhi,4);
+	&pxor	($Zlo,&QWP(8,$Htbl,$nhi));
+	&mov	(&LB($nlo),&BP(0,$inp,$cnt));
+	&psllq	($tmp,60);
+	&pxor	($Zhi,&QWP(0,$rem_4bit,$rem,8));
+	&dec	($cnt);
+	&movd	($rem,$Zlo);
+	&pxor	($Zhi,&QWP(0,$Htbl,$nhi));
+	&mov	($nhi,$nlo);
+	&pxor	($Zlo,$tmp);
+	&js	(&label("mmx_break"));
+
+	&shl	(&LB($nlo),4);
+	&and	($rem,0xf);
+	&psrlq	($Zlo,4);
+	&and	($nhi,0xf0);
+	&movq	($tmp,$Zhi);
+	&psrlq	($Zhi,4);
+	&pxor	($Zlo,&QWP(8,$Htbl,$nlo));
+	&psllq	($tmp,60);
+	&pxor	($Zhi,&QWP(0,$rem_4bit,$rem,8));
+	&movd	($rem,$Zlo);
+	&pxor	($Zhi,&QWP(0,$Htbl,$nlo));
+	&pxor	($Zlo,$tmp);
+	&jmp	(&label("mmx_loop"));
+
+    &set_label("mmx_break",16);
+	&shl	(&LB($nlo),4);
+	&and	($rem,0xf);
+	&psrlq	($Zlo,4);
+	&and	($nhi,0xf0);
+	&movq	($tmp,$Zhi);
+	&psrlq	($Zhi,4);
+	&pxor	($Zlo,&QWP(8,$Htbl,$nlo));
+	&psllq	($tmp,60);
+	&pxor	($Zhi,&QWP(0,$rem_4bit,$rem,8));
+	&movd	($rem,$Zlo);
+	&pxor	($Zhi,&QWP(0,$Htbl,$nlo));
+	&pxor	($Zlo,$tmp);
+
+	&psrlq	($Zlo,4);
+	&and	($rem,0xf);
+	&movq	($tmp,$Zhi);
+	&psrlq	($Zhi,4);
+	&pxor	($Zlo,&QWP(8,$Htbl,$nhi));
+	&psllq	($tmp,60);
+	&pxor	($Zhi,&QWP(0,$rem_4bit,$rem,8));
+	&movd	($rem,$Zlo);
+	&pxor	($Zhi,&QWP(0,$Htbl,$nhi));
+	&pxor	($Zlo,$tmp);
+
+	&psrlq	($Zlo,32);	# lower part of Zlo is already there
+	&movd	($Zhl,$Zhi);
+	&psrlq	($Zhi,32);
+	&movd	($Zlh,$Zlo);
+	&movd	($Zhh,$Zhi);
+
+	&bswap	($Zll);
+	&bswap	($Zhl);
+	&bswap	($Zlh);
+	&bswap	($Zhh);
+}
+
+&function_begin("gcm_gmult_4bit_mmx");
+	&mov	($inp,&wparam(0));	# load Xi
+	&mov	($Htbl,&wparam(1));	# load Htable
+
+	&call	(&label("pic_point"));
+	&set_label("pic_point");
+	&blindpop("eax");
+	&lea	("eax",&DWP(&label("rem_4bit")."-".&label("pic_point"),"eax"));
+
+	&movz	($Zll,&BP(15,$inp));
+
+	&mmx_loop($inp,"eax");
+
+	&emms	();
+	&mov	(&DWP(12,$inp),$Zll);
+	&mov	(&DWP(4,$inp),$Zhl);
+	&mov	(&DWP(8,$inp),$Zlh);
+	&mov	(&DWP(0,$inp),$Zhh);
+&function_end("gcm_gmult_4bit_mmx");
+
+######################################################################
+# Below subroutine is "528B" variant of "4-bit" GCM GHASH function
+# (see gcm128.c for details). It provides further 20-40% performance
+# improvement over above mentioned "May" version.
+
+&static_label("rem_8bit");
+
+&function_begin("gcm_ghash_4bit_mmx");
+{ my ($Zlo,$Zhi) = ("mm7","mm6");
+  my $rem_8bit = "esi";
+  my $Htbl = "ebx";
+
+    # parameter block
+    &mov	("eax",&wparam(0));		# Xi
+    &mov	("ebx",&wparam(1));		# Htable
+    &mov	("ecx",&wparam(2));		# inp
+    &mov	("edx",&wparam(3));		# len
+    &mov	("ebp","esp");			# original %esp
+    &call	(&label("pic_point"));
+    &set_label	("pic_point");
+    &blindpop	($rem_8bit);
+    &lea	($rem_8bit,&DWP(&label("rem_8bit")."-".&label("pic_point"),$rem_8bit));
+
+    &sub	("esp",512+16+16);		# allocate stack frame...
+    &and	("esp",-64);			# ...and align it
+    &sub	("esp",16);			# place for (u8)(H[]<<4)
+
+    &add	("edx","ecx");			# pointer to the end of input
+    &mov	(&DWP(528+16+0,"esp"),"eax");	# save Xi
+    &mov	(&DWP(528+16+8,"esp"),"edx");	# save inp+len
+    &mov	(&DWP(528+16+12,"esp"),"ebp");	# save original %esp
+
+    { my @lo  = ("mm0","mm1","mm2");
+      my @hi  = ("mm3","mm4","mm5");
+      my @tmp = ("mm6","mm7");
+      my $off1=0,$off2=0,$i;
+
+      &add	($Htbl,128);			# optimize for size
+      &lea	("edi",&DWP(16+128,"esp"));
+      &lea	("ebp",&DWP(16+256+128,"esp"));
+
+      # decompose Htable (low and high parts are kept separately),
+      # generate Htable[]>>4, (u8)(Htable[]<<4), save to stack...
+      for ($i=0;$i<18;$i++) {
+
+	&mov	("edx",&DWP(16*$i+8-128,$Htbl))		if ($i<16);
+	&movq	($lo[0],&QWP(16*$i+8-128,$Htbl))	if ($i<16);
+	&psllq	($tmp[1],60)				if ($i>1);
+	&movq	($hi[0],&QWP(16*$i+0-128,$Htbl))	if ($i<16);
+	&por	($lo[2],$tmp[1])			if ($i>1);
+	&movq	(&QWP($off1-128,"edi"),$lo[1])		if ($i>0 && $i<17);
+	&psrlq	($lo[1],4)				if ($i>0 && $i<17);
+	&movq	(&QWP($off1,"edi"),$hi[1])		if ($i>0 && $i<17);
+	&movq	($tmp[0],$hi[1])			if ($i>0 && $i<17);
+	&movq	(&QWP($off2-128,"ebp"),$lo[2])		if ($i>1);
+	&psrlq	($hi[1],4)				if ($i>0 && $i<17);
+	&movq	(&QWP($off2,"ebp"),$hi[2])		if ($i>1);
+	&shl	("edx",4)				if ($i<16);
+	&mov	(&BP($i,"esp"),&LB("edx"))		if ($i<16);
+
+	unshift	(@lo,pop(@lo));			# "rotate" registers
+	unshift	(@hi,pop(@hi));
+	unshift	(@tmp,pop(@tmp));
+	$off1 += 8	if ($i>0);
+	$off2 += 8	if ($i>1);
+      }
+    }
+
+    &movq	($Zhi,&QWP(0,"eax"));
+    &mov	("ebx",&DWP(8,"eax"));
+    &mov	("edx",&DWP(12,"eax"));		# load Xi
+
+&set_label("outer",16);
+  { my $nlo = "eax";
+    my $dat = "edx";
+    my @nhi = ("edi","ebp");
+    my @rem = ("ebx","ecx");
+    my @red = ("mm0","mm1","mm2");
+    my $tmp = "mm3";
+
+    &xor	($dat,&DWP(12,"ecx"));		# merge input data
+    &xor	("ebx",&DWP(8,"ecx"));
+    &pxor	($Zhi,&QWP(0,"ecx"));
+    &lea	("ecx",&DWP(16,"ecx"));		# inp+=16
+    #&mov	(&DWP(528+12,"esp"),$dat);	# save inp^Xi
+    &mov	(&DWP(528+8,"esp"),"ebx");
+    &movq	(&QWP(528+0,"esp"),$Zhi);
+    &mov	(&DWP(528+16+4,"esp"),"ecx");	# save inp
+
+    &xor	($nlo,$nlo);
+    &rol	($dat,8);
+    &mov	(&LB($nlo),&LB($dat));
+    &mov	($nhi[1],$nlo);
+    &and	(&LB($nlo),0x0f);
+    &shr	($nhi[1],4);
+    &pxor	($red[0],$red[0]);
+    &rol	($dat,8);			# next byte
+    &pxor	($red[1],$red[1]);
+    &pxor	($red[2],$red[2]);
+
+    # Just like in "May" verson modulo-schedule for critical path in
+    # 'Z.hi ^= rem_8bit[Z.lo&0xff^((u8)H[nhi]<<4)]<<48'. Final 'pxor'
+    # is scheduled so late that rem_8bit[] has to be shifted *right*
+    # by 16, which is why last argument to pinsrw is 2, which
+    # corresponds to <<32=<<48>>16...
+    for ($j=11,$i=0;$i<15;$i++) {
+
+      if ($i>0) {
+	&pxor	($Zlo,&QWP(16,"esp",$nlo,8));		# Z^=H[nlo]
+	&rol	($dat,8);				# next byte
+	&pxor	($Zhi,&QWP(16+128,"esp",$nlo,8));
+
+	&pxor	($Zlo,$tmp);
+	&pxor	($Zhi,&QWP(16+256+128,"esp",$nhi[0],8));
+	&xor	(&LB($rem[1]),&BP(0,"esp",$nhi[0]));	# rem^(H[nhi]<<4)
+      } else {
+	&movq	($Zlo,&QWP(16,"esp",$nlo,8));
+	&movq	($Zhi,&QWP(16+128,"esp",$nlo,8));
+      }
+
+	&mov	(&LB($nlo),&LB($dat));
+	&mov	($dat,&DWP(528+$j,"esp"))		if (--$j%4==0);
+
+	&movd	($rem[0],$Zlo);
+	&movz	($rem[1],&LB($rem[1]))			if ($i>0);
+	&psrlq	($Zlo,8);				# Z>>=8
+
+	&movq	($tmp,$Zhi);
+	&mov	($nhi[0],$nlo);
+	&psrlq	($Zhi,8);
+
+	&pxor	($Zlo,&QWP(16+256+0,"esp",$nhi[1],8));	# Z^=H[nhi]>>4
+	&and	(&LB($nlo),0x0f);
+	&psllq	($tmp,56);
+
+	&pxor	($Zhi,$red[1])				if ($i>1);
+	&shr	($nhi[0],4);
+	&pinsrw	($red[0],&WP(0,$rem_8bit,$rem[1],2),2)	if ($i>0);
+
+	unshift	(@red,pop(@red));			# "rotate" registers
+	unshift	(@rem,pop(@rem));
+	unshift	(@nhi,pop(@nhi));
+    }
+
+    &pxor	($Zlo,&QWP(16,"esp",$nlo,8));		# Z^=H[nlo]
+    &pxor	($Zhi,&QWP(16+128,"esp",$nlo,8));
+    &xor	(&LB($rem[1]),&BP(0,"esp",$nhi[0]));	# rem^(H[nhi]<<4)
+
+    &pxor	($Zlo,$tmp);
+    &pxor	($Zhi,&QWP(16+256+128,"esp",$nhi[0],8));
+    &movz	($rem[1],&LB($rem[1]));
+
+    &pxor	($red[2],$red[2]);			# clear 2nd word
+    &psllq	($red[1],4);
+
+    &movd	($rem[0],$Zlo);
+    &psrlq	($Zlo,4);				# Z>>=4
+
+    &movq	($tmp,$Zhi);
+    &psrlq	($Zhi,4);
+    &shl	($rem[0],4);				# rem<<4
+
+    &pxor	($Zlo,&QWP(16,"esp",$nhi[1],8));	# Z^=H[nhi]
+    &psllq	($tmp,60);
+    &movz	($rem[0],&LB($rem[0]));
+
+    &pxor	($Zlo,$tmp);
+    &pxor	($Zhi,&QWP(16+128,"esp",$nhi[1],8));
+
+    &pinsrw	($red[0],&WP(0,$rem_8bit,$rem[1],2),2);
+    &pxor	($Zhi,$red[1]);
+
+    &movd	($dat,$Zlo);
+    &pinsrw	($red[2],&WP(0,$rem_8bit,$rem[0],2),3);	# last is <<48
+
+    &psllq	($red[0],12);				# correct by <<16>>4
+    &pxor	($Zhi,$red[0]);
+    &psrlq	($Zlo,32);
+    &pxor	($Zhi,$red[2]);
+
+    &mov	("ecx",&DWP(528+16+4,"esp"));	# restore inp
+    &movd	("ebx",$Zlo);
+    &movq	($tmp,$Zhi);			# 01234567
+    &psllw	($Zhi,8);			# 1.3.5.7.
+    &psrlw	($tmp,8);			# .0.2.4.6
+    &por	($Zhi,$tmp);			# 10325476
+    &bswap	($dat);
+    &pshufw	($Zhi,$Zhi,0b00011011);		# 76543210
+    &bswap	("ebx");
+    
+    &cmp	("ecx",&DWP(528+16+8,"esp"));	# are we done?
+    &jne	(&label("outer"));
+  }
+
+    &mov	("eax",&DWP(528+16+0,"esp"));	# restore Xi
+    &mov	(&DWP(12,"eax"),"edx");
+    &mov	(&DWP(8,"eax"),"ebx");
+    &movq	(&QWP(0,"eax"),$Zhi);
+
+    &mov	("esp",&DWP(528+16+12,"esp"));	# restore original %esp
+    &emms	();
+}
+&function_end("gcm_ghash_4bit_mmx");
+}}
+
+if ($sse2) {{
+######################################################################
+# PCLMULQDQ version.
+
+$Xip="eax";
+$Htbl="edx";
+$const="ecx";
+$inp="esi";
+$len="ebx";
+
+($Xi,$Xhi)=("xmm0","xmm1");	$Hkey="xmm2";
+($T1,$T2,$T3)=("xmm3","xmm4","xmm5");
+($Xn,$Xhn)=("xmm6","xmm7");
+
+&static_label("bswap");
+
+sub clmul64x64_T2 {	# minimal "register" pressure
+my ($Xhi,$Xi,$Hkey)=@_;
+
+	&movdqa		($Xhi,$Xi);		#
+	&pshufd		($T1,$Xi,0b01001110);
+	&pshufd		($T2,$Hkey,0b01001110);
+	&pxor		($T1,$Xi);		#
+	&pxor		($T2,$Hkey);
+
+	&pclmulqdq	($Xi,$Hkey,0x00);	#######
+	&pclmulqdq	($Xhi,$Hkey,0x11);	#######
+	&pclmulqdq	($T1,$T2,0x00);		#######
+	&xorps		($T1,$Xi);		#
+	&xorps		($T1,$Xhi);		#
+
+	&movdqa		($T2,$T1);		#
+	&psrldq		($T1,8);
+	&pslldq		($T2,8);		#
+	&pxor		($Xhi,$T1);
+	&pxor		($Xi,$T2);		#
+}
+
+sub clmul64x64_T3 {
+# Even though this subroutine offers visually better ILP, it
+# was empirically found to be a tad slower than above version.
+# At least in gcm_ghash_clmul context. But it's just as well,
+# because loop modulo-scheduling is possible only thanks to
+# minimized "register" pressure...
+my ($Xhi,$Xi,$Hkey)=@_;
+
+	&movdqa		($T1,$Xi);		#
+	&movdqa		($Xhi,$Xi);
+	&pclmulqdq	($Xi,$Hkey,0x00);	#######
+	&pclmulqdq	($Xhi,$Hkey,0x11);	#######
+	&pshufd		($T2,$T1,0b01001110);	#
+	&pshufd		($T3,$Hkey,0b01001110);
+	&pxor		($T2,$T1);		#
+	&pxor		($T3,$Hkey);
+	&pclmulqdq	($T2,$T3,0x00);		#######
+	&pxor		($T2,$Xi);		#
+	&pxor		($T2,$Xhi);		#
+
+	&movdqa		($T3,$T2);		#
+	&psrldq		($T2,8);
+	&pslldq		($T3,8);		#
+	&pxor		($Xhi,$T2);
+	&pxor		($Xi,$T3);		#
+}
+
+if (1) {		# Algorithm 9 with <<1 twist.
+			# Reduction is shorter and uses only two
+			# temporary registers, which makes it better
+			# candidate for interleaving with 64x64
+			# multiplication. Pre-modulo-scheduled loop
+			# was found to be ~20% faster than Algorithm 5
+			# below. Algorithm 9 was therefore chosen for
+			# further optimization...
+
+sub reduction_alg9 {	# 17/13 times faster than Intel version
+my ($Xhi,$Xi) = @_;
+
+	# 1st phase
+	&movdqa		($T1,$Xi)		#
+	&psllq		($Xi,1);
+	&pxor		($Xi,$T1);		#
+	&psllq		($Xi,5);		#
+	&pxor		($Xi,$T1);		#
+	&psllq		($Xi,57);		#
+	&movdqa		($T2,$Xi);		#
+	&pslldq		($Xi,8);
+	&psrldq		($T2,8);		#
+	&pxor		($Xi,$T1);
+	&pxor		($Xhi,$T2);		#
+
+	# 2nd phase
+	&movdqa		($T2,$Xi);
+	&psrlq		($Xi,5);
+	&pxor		($Xi,$T2);		#
+	&psrlq		($Xi,1);		#
+	&pxor		($Xi,$T2);		#
+	&pxor		($T2,$Xhi);
+	&psrlq		($Xi,1);		#
+	&pxor		($Xi,$T2);		#
+}
+
+&function_begin_B("gcm_init_clmul");
+	&mov		($Htbl,&wparam(0));
+	&mov		($Xip,&wparam(1));
+
+	&call		(&label("pic"));
+&set_label("pic");
+	&blindpop	($const);
+	&lea		($const,&DWP(&label("bswap")."-".&label("pic"),$const));
+
+	&movdqu		($Hkey,&QWP(0,$Xip));
+	&pshufd		($Hkey,$Hkey,0b01001110);# dword swap
+
+	# <<1 twist
+	&pshufd		($T2,$Hkey,0b11111111);	# broadcast uppermost dword
+	&movdqa		($T1,$Hkey);
+	&psllq		($Hkey,1);
+	&pxor		($T3,$T3);		#
+	&psrlq		($T1,63);
+	&pcmpgtd	($T3,$T2);		# broadcast carry bit
+	&pslldq		($T1,8);
+	&por		($Hkey,$T1);		# H<<=1
+
+	# magic reduction
+	&pand		($T3,&QWP(16,$const));	# 0x1c2_polynomial
+	&pxor		($Hkey,$T3);		# if(carry) H^=0x1c2_polynomial
+
+	# calculate H^2
+	&movdqa		($Xi,$Hkey);
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey);
+	&reduction_alg9	($Xhi,$Xi);
+
+	&movdqu		(&QWP(0,$Htbl),$Hkey);	# save H
+	&movdqu		(&QWP(16,$Htbl),$Xi);	# save H^2
+
+	&ret		();
+&function_end_B("gcm_init_clmul");
+
+&function_begin_B("gcm_gmult_clmul");
+	&mov		($Xip,&wparam(0));
+	&mov		($Htbl,&wparam(1));
+
+	&call		(&label("pic"));
+&set_label("pic");
+	&blindpop	($const);
+	&lea		($const,&DWP(&label("bswap")."-".&label("pic"),$const));
+
+	&movdqu		($Xi,&QWP(0,$Xip));
+	&movdqa		($T3,&QWP(0,$const));
+	&movups		($Hkey,&QWP(0,$Htbl));
+	&pshufb		($Xi,$T3);
+
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey);
+	&reduction_alg9	($Xhi,$Xi);
+
+	&pshufb		($Xi,$T3);
+	&movdqu		(&QWP(0,$Xip),$Xi);
+
+	&ret	();
+&function_end_B("gcm_gmult_clmul");
+
+&function_begin("gcm_ghash_clmul");
+	&mov		($Xip,&wparam(0));
+	&mov		($Htbl,&wparam(1));
+	&mov		($inp,&wparam(2));
+	&mov		($len,&wparam(3));
+
+	&call		(&label("pic"));
+&set_label("pic");
+	&blindpop	($const);
+	&lea		($const,&DWP(&label("bswap")."-".&label("pic"),$const));
+
+	&movdqu		($Xi,&QWP(0,$Xip));
+	&movdqa		($T3,&QWP(0,$const));
+	&movdqu		($Hkey,&QWP(0,$Htbl));
+	&pshufb		($Xi,$T3);
+
+	&sub		($len,0x10);
+	&jz		(&label("odd_tail"));
+
+	#######
+	# Xi+2 =[H*(Ii+1 + Xi+1)] mod P =
+	#	[(H*Ii+1) + (H*Xi+1)] mod P =
+	#	[(H*Ii+1) + H^2*(Ii+Xi)] mod P
+	#
+	&movdqu		($T1,&QWP(0,$inp));	# Ii
+	&movdqu		($Xn,&QWP(16,$inp));	# Ii+1
+	&pshufb		($T1,$T3);
+	&pshufb		($Xn,$T3);
+	&pxor		($Xi,$T1);		# Ii+Xi
+
+	&clmul64x64_T2	($Xhn,$Xn,$Hkey);	# H*Ii+1
+	&movups		($Hkey,&QWP(16,$Htbl));	# load H^2
+
+	&lea		($inp,&DWP(32,$inp));	# i+=2
+	&sub		($len,0x20);
+	&jbe		(&label("even_tail"));
+
+&set_label("mod_loop");
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey);	# H^2*(Ii+Xi)
+	&movdqu		($T1,&QWP(0,$inp));	# Ii
+	&movups		($Hkey,&QWP(0,$Htbl));	# load H
+
+	&pxor		($Xi,$Xn);		# (H*Ii+1) + H^2*(Ii+Xi)
+	&pxor		($Xhi,$Xhn);
+
+	&movdqu		($Xn,&QWP(16,$inp));	# Ii+1
+	&pshufb		($T1,$T3);
+	&pshufb		($Xn,$T3);
+
+	&movdqa		($T3,$Xn);		#&clmul64x64_TX	($Xhn,$Xn,$Hkey); H*Ii+1
+	&movdqa		($Xhn,$Xn);
+	 &pxor		($Xhi,$T1);		# "Ii+Xi", consume early
+
+	  &movdqa	($T1,$Xi)		#&reduction_alg9($Xhi,$Xi); 1st phase
+	  &psllq	($Xi,1);
+	  &pxor		($Xi,$T1);		#
+	  &psllq	($Xi,5);		#
+	  &pxor		($Xi,$T1);		#
+	&pclmulqdq	($Xn,$Hkey,0x00);	#######
+	  &psllq	($Xi,57);		#
+	  &movdqa	($T2,$Xi);		#
+	  &pslldq	($Xi,8);
+	  &psrldq	($T2,8);		#	
+	  &pxor		($Xi,$T1);
+	&pshufd		($T1,$T3,0b01001110);
+	  &pxor		($Xhi,$T2);		#
+	&pxor		($T1,$T3);
+	&pshufd		($T3,$Hkey,0b01001110);
+	&pxor		($T3,$Hkey);		#
+
+	&pclmulqdq	($Xhn,$Hkey,0x11);	#######
+	  &movdqa	($T2,$Xi);		# 2nd phase
+	  &psrlq	($Xi,5);
+	  &pxor		($Xi,$T2);		#
+	  &psrlq	($Xi,1);		#
+	  &pxor		($Xi,$T2);		#
+	  &pxor		($T2,$Xhi);
+	  &psrlq	($Xi,1);		#
+	  &pxor		($Xi,$T2);		#
+
+	&pclmulqdq	($T1,$T3,0x00);		#######
+	&movups		($Hkey,&QWP(16,$Htbl));	# load H^2
+	&xorps		($T1,$Xn);		#
+	&xorps		($T1,$Xhn);		#
+
+	&movdqa		($T3,$T1);		#
+	&psrldq		($T1,8);
+	&pslldq		($T3,8);		#
+	&pxor		($Xhn,$T1);
+	&pxor		($Xn,$T3);		#
+	&movdqa		($T3,&QWP(0,$const));
+
+	&lea		($inp,&DWP(32,$inp));
+	&sub		($len,0x20);
+	&ja		(&label("mod_loop"));
+
+&set_label("even_tail");
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey);	# H^2*(Ii+Xi)
+
+	&pxor		($Xi,$Xn);		# (H*Ii+1) + H^2*(Ii+Xi)
+	&pxor		($Xhi,$Xhn);
+
+	&reduction_alg9	($Xhi,$Xi);
+
+	&test		($len,$len);
+	&jnz		(&label("done"));
+
+	&movups		($Hkey,&QWP(0,$Htbl));	# load H
+&set_label("odd_tail");
+	&movdqu		($T1,&QWP(0,$inp));	# Ii
+	&pshufb		($T1,$T3);
+	&pxor		($Xi,$T1);		# Ii+Xi
+
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey);	# H*(Ii+Xi)
+	&reduction_alg9	($Xhi,$Xi);
+
+&set_label("done");
+	&pshufb		($Xi,$T3);
+	&movdqu		(&QWP(0,$Xip),$Xi);
+&function_end("gcm_ghash_clmul");
+
+} else {		# Algorith 5. Kept for reference purposes.
+
+sub reduction_alg5 {	# 19/16 times faster than Intel version
+my ($Xhi,$Xi)=@_;
+
+	# <<1
+	&movdqa		($T1,$Xi);		#
+	&movdqa		($T2,$Xhi);
+	&pslld		($Xi,1);
+	&pslld		($Xhi,1);		#
+	&psrld		($T1,31);
+	&psrld		($T2,31);		#
+	&movdqa		($T3,$T1);
+	&pslldq		($T1,4);
+	&psrldq		($T3,12);		#
+	&pslldq		($T2,4);
+	&por		($Xhi,$T3);		#
+	&por		($Xi,$T1);
+	&por		($Xhi,$T2);		#
+
+	# 1st phase
+	&movdqa		($T1,$Xi);
+	&movdqa		($T2,$Xi);
+	&movdqa		($T3,$Xi);		#
+	&pslld		($T1,31);
+	&pslld		($T2,30);
+	&pslld		($Xi,25);		#
+	&pxor		($T1,$T2);
+	&pxor		($T1,$Xi);		#
+	&movdqa		($T2,$T1);		#
+	&pslldq		($T1,12);
+	&psrldq		($T2,4);		#
+	&pxor		($T3,$T1);
+
+	# 2nd phase
+	&pxor		($Xhi,$T3);		#
+	&movdqa		($Xi,$T3);
+	&movdqa		($T1,$T3);
+	&psrld		($Xi,1);		#
+	&psrld		($T1,2);
+	&psrld		($T3,7);		#
+	&pxor		($Xi,$T1);
+	&pxor		($Xhi,$T2);
+	&pxor		($Xi,$T3);		#
+	&pxor		($Xi,$Xhi);		#
+}
+
+&function_begin_B("gcm_init_clmul");
+	&mov		($Htbl,&wparam(0));
+	&mov		($Xip,&wparam(1));
+
+	&call		(&label("pic"));
+&set_label("pic");
+	&blindpop	($const);
+	&lea		($const,&DWP(&label("bswap")."-".&label("pic"),$const));
+
+	&movdqu		($Hkey,&QWP(0,$Xip));
+	&pshufd		($Hkey,$Hkey,0b01001110);# dword swap
+
+	# calculate H^2
+	&movdqa		($Xi,$Hkey);
+	&clmul64x64_T3	($Xhi,$Xi,$Hkey);
+	&reduction_alg5	($Xhi,$Xi);
+
+	&movdqu		(&QWP(0,$Htbl),$Hkey);	# save H
+	&movdqu		(&QWP(16,$Htbl),$Xi);	# save H^2
+
+	&ret		();
+&function_end_B("gcm_init_clmul");
+
+&function_begin_B("gcm_gmult_clmul");
+	&mov		($Xip,&wparam(0));
+	&mov		($Htbl,&wparam(1));
+
+	&call		(&label("pic"));
+&set_label("pic");
+	&blindpop	($const);
+	&lea		($const,&DWP(&label("bswap")."-".&label("pic"),$const));
+
+	&movdqu		($Xi,&QWP(0,$Xip));
+	&movdqa		($Xn,&QWP(0,$const));
+	&movdqu		($Hkey,&QWP(0,$Htbl));
+	&pshufb		($Xi,$Xn);
+
+	&clmul64x64_T3	($Xhi,$Xi,$Hkey);
+	&reduction_alg5	($Xhi,$Xi);
+
+	&pshufb		($Xi,$Xn);
+	&movdqu		(&QWP(0,$Xip),$Xi);
+
+	&ret	();
+&function_end_B("gcm_gmult_clmul");
+
+&function_begin("gcm_ghash_clmul");
+	&mov		($Xip,&wparam(0));
+	&mov		($Htbl,&wparam(1));
+	&mov		($inp,&wparam(2));
+	&mov		($len,&wparam(3));
+
+	&call		(&label("pic"));
+&set_label("pic");
+	&blindpop	($const);
+	&lea		($const,&DWP(&label("bswap")."-".&label("pic"),$const));
+
+	&movdqu		($Xi,&QWP(0,$Xip));
+	&movdqa		($T3,&QWP(0,$const));
+	&movdqu		($Hkey,&QWP(0,$Htbl));
+	&pshufb		($Xi,$T3);
+
+	&sub		($len,0x10);
+	&jz		(&label("odd_tail"));
+
+	#######
+	# Xi+2 =[H*(Ii+1 + Xi+1)] mod P =
+	#	[(H*Ii+1) + (H*Xi+1)] mod P =
+	#	[(H*Ii+1) + H^2*(Ii+Xi)] mod P
+	#
+	&movdqu		($T1,&QWP(0,$inp));	# Ii
+	&movdqu		($Xn,&QWP(16,$inp));	# Ii+1
+	&pshufb		($T1,$T3);
+	&pshufb		($Xn,$T3);
+	&pxor		($Xi,$T1);		# Ii+Xi
+
+	&clmul64x64_T3	($Xhn,$Xn,$Hkey);	# H*Ii+1
+	&movdqu		($Hkey,&QWP(16,$Htbl));	# load H^2
+
+	&sub		($len,0x20);
+	&lea		($inp,&DWP(32,$inp));	# i+=2
+	&jbe		(&label("even_tail"));
+
+&set_label("mod_loop");
+	&clmul64x64_T3	($Xhi,$Xi,$Hkey);	# H^2*(Ii+Xi)
+	&movdqu		($Hkey,&QWP(0,$Htbl));	# load H
+
+	&pxor		($Xi,$Xn);		# (H*Ii+1) + H^2*(Ii+Xi)
+	&pxor		($Xhi,$Xhn);
+
+	&reduction_alg5	($Xhi,$Xi);
+
+	#######
+	&movdqa		($T3,&QWP(0,$const));
+	&movdqu		($T1,&QWP(0,$inp));	# Ii
+	&movdqu		($Xn,&QWP(16,$inp));	# Ii+1
+	&pshufb		($T1,$T3);
+	&pshufb		($Xn,$T3);
+	&pxor		($Xi,$T1);		# Ii+Xi
+
+	&clmul64x64_T3	($Xhn,$Xn,$Hkey);	# H*Ii+1
+	&movdqu		($Hkey,&QWP(16,$Htbl));	# load H^2
+
+	&sub		($len,0x20);
+	&lea		($inp,&DWP(32,$inp));
+	&ja		(&label("mod_loop"));
+
+&set_label("even_tail");
+	&clmul64x64_T3	($Xhi,$Xi,$Hkey);	# H^2*(Ii+Xi)
+
+	&pxor		($Xi,$Xn);		# (H*Ii+1) + H^2*(Ii+Xi)
+	&pxor		($Xhi,$Xhn);
+
+	&reduction_alg5	($Xhi,$Xi);
+
+	&movdqa		($T3,&QWP(0,$const));
+	&test		($len,$len);
+	&jnz		(&label("done"));
+
+	&movdqu		($Hkey,&QWP(0,$Htbl));	# load H
+&set_label("odd_tail");
+	&movdqu		($T1,&QWP(0,$inp));	# Ii
+	&pshufb		($T1,$T3);
+	&pxor		($Xi,$T1);		# Ii+Xi
+
+	&clmul64x64_T3	($Xhi,$Xi,$Hkey);	# H*(Ii+Xi)
+	&reduction_alg5	($Xhi,$Xi);
+
+	&movdqa		($T3,&QWP(0,$const));
+&set_label("done");
+	&pshufb		($Xi,$T3);
+	&movdqu		(&QWP(0,$Xip),$Xi);
+&function_end("gcm_ghash_clmul");
+
+}
+
+&set_label("bswap",64);
+	&data_byte(15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0);
+	&data_byte(1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0xc2);	# 0x1c2_polynomial
+}}	# $sse2
+
+&set_label("rem_4bit",64);
+	&data_word(0,0x0000<<$S,0,0x1C20<<$S,0,0x3840<<$S,0,0x2460<<$S);
+	&data_word(0,0x7080<<$S,0,0x6CA0<<$S,0,0x48C0<<$S,0,0x54E0<<$S);
+	&data_word(0,0xE100<<$S,0,0xFD20<<$S,0,0xD940<<$S,0,0xC560<<$S);
+	&data_word(0,0x9180<<$S,0,0x8DA0<<$S,0,0xA9C0<<$S,0,0xB5E0<<$S);
+&set_label("rem_8bit",64);
+	&data_short(0x0000,0x01C2,0x0384,0x0246,0x0708,0x06CA,0x048C,0x054E);
+	&data_short(0x0E10,0x0FD2,0x0D94,0x0C56,0x0918,0x08DA,0x0A9C,0x0B5E);
+	&data_short(0x1C20,0x1DE2,0x1FA4,0x1E66,0x1B28,0x1AEA,0x18AC,0x196E);
+	&data_short(0x1230,0x13F2,0x11B4,0x1076,0x1538,0x14FA,0x16BC,0x177E);
+	&data_short(0x3840,0x3982,0x3BC4,0x3A06,0x3F48,0x3E8A,0x3CCC,0x3D0E);
+	&data_short(0x3650,0x3792,0x35D4,0x3416,0x3158,0x309A,0x32DC,0x331E);
+	&data_short(0x2460,0x25A2,0x27E4,0x2626,0x2368,0x22AA,0x20EC,0x212E);
+	&data_short(0x2A70,0x2BB2,0x29F4,0x2836,0x2D78,0x2CBA,0x2EFC,0x2F3E);
+	&data_short(0x7080,0x7142,0x7304,0x72C6,0x7788,0x764A,0x740C,0x75CE);
+	&data_short(0x7E90,0x7F52,0x7D14,0x7CD6,0x7998,0x785A,0x7A1C,0x7BDE);
+	&data_short(0x6CA0,0x6D62,0x6F24,0x6EE6,0x6BA8,0x6A6A,0x682C,0x69EE);
+	&data_short(0x62B0,0x6372,0x6134,0x60F6,0x65B8,0x647A,0x663C,0x67FE);
+	&data_short(0x48C0,0x4902,0x4B44,0x4A86,0x4FC8,0x4E0A,0x4C4C,0x4D8E);
+	&data_short(0x46D0,0x4712,0x4554,0x4496,0x41D8,0x401A,0x425C,0x439E);
+	&data_short(0x54E0,0x5522,0x5764,0x56A6,0x53E8,0x522A,0x506C,0x51AE);
+	&data_short(0x5AF0,0x5B32,0x5974,0x58B6,0x5DF8,0x5C3A,0x5E7C,0x5FBE);
+	&data_short(0xE100,0xE0C2,0xE284,0xE346,0xE608,0xE7CA,0xE58C,0xE44E);
+	&data_short(0xEF10,0xEED2,0xEC94,0xED56,0xE818,0xE9DA,0xEB9C,0xEA5E);
+	&data_short(0xFD20,0xFCE2,0xFEA4,0xFF66,0xFA28,0xFBEA,0xF9AC,0xF86E);
+	&data_short(0xF330,0xF2F2,0xF0B4,0xF176,0xF438,0xF5FA,0xF7BC,0xF67E);
+	&data_short(0xD940,0xD882,0xDAC4,0xDB06,0xDE48,0xDF8A,0xDDCC,0xDC0E);
+	&data_short(0xD750,0xD692,0xD4D4,0xD516,0xD058,0xD19A,0xD3DC,0xD21E);
+	&data_short(0xC560,0xC4A2,0xC6E4,0xC726,0xC268,0xC3AA,0xC1EC,0xC02E);
+	&data_short(0xCB70,0xCAB2,0xC8F4,0xC936,0xCC78,0xCDBA,0xCFFC,0xCE3E);
+	&data_short(0x9180,0x9042,0x9204,0x93C6,0x9688,0x974A,0x950C,0x94CE);
+	&data_short(0x9F90,0x9E52,0x9C14,0x9DD6,0x9898,0x995A,0x9B1C,0x9ADE);
+	&data_short(0x8DA0,0x8C62,0x8E24,0x8FE6,0x8AA8,0x8B6A,0x892C,0x88EE);
+	&data_short(0x83B0,0x8272,0x8034,0x81F6,0x84B8,0x857A,0x873C,0x86FE);
+	&data_short(0xA9C0,0xA802,0xAA44,0xAB86,0xAEC8,0xAF0A,0xAD4C,0xAC8E);
+	&data_short(0xA7D0,0xA612,0xA454,0xA596,0xA0D8,0xA11A,0xA35C,0xA29E);
+	&data_short(0xB5E0,0xB422,0xB664,0xB7A6,0xB2E8,0xB32A,0xB16C,0xB0AE);
+	&data_short(0xBBF0,0xBA32,0xB874,0xB9B6,0xBCF8,0xBD3A,0xBF7C,0xBEBE);
+}}}	# !$x86only
+
+&asciz("GHASH for x86, CRYPTOGAMS by <appro\@openssl.org>");
+&asm_finish();
+
+# A question was risen about choice of vanilla MMX. Or rather why wasn't
+# SSE2 chosen instead? In addition to the fact that MMX runs on legacy
+# CPUs such as PIII, "4-bit" MMX version was observed to provide better
+# performance than *corresponding* SSE2 one even on contemporary CPUs.
+# SSE2 results were provided by Peter-Michael Hager. He maintains SSE2
+# implementation featuring full range of lookup-table sizes, but with
+# per-invocation lookup table setup. Latter means that table size is
+# chosen depending on how much data is to be hashed in every given call,
+# more data - larger table. Best reported result for Core2 is ~4 cycles
+# per processed byte out of 64KB block. This number accounts even for
+# 64KB table setup overhead. As discussed in gcm128.c we choose to be
+# more conservative in respect to lookup table sizes, but how do the
+# results compare? Minimalistic "256B" MMX version delivers ~11 cycles
+# on same platform. As also discussed in gcm128.c, next in line "8-bit
+# Shoup's" or "4KB" method should deliver twice the performance of
+# "256B" one, in other words not worse than ~6 cycles per byte. It
+# should be also be noted that in SSE2 case improvement can be "super-
+# linear," i.e. more than twice, mostly because >>8 maps to single
+# instruction on SSE2 register. This is unlike "4-bit" case when >>4
+# maps to same amount of instructions in both MMX and SSE2 cases.
+# Bottom line is that switch to SSE2 is considered to be justifiable
+# only in case we choose to implement "8-bit" method...
diff --git a/crypto/modes/asm/ghash-x86_64.pl b/crypto/modes/asm/ghash-x86_64.pl
new file mode 100755
index 000000000000..a5ae180882db
--- /dev/null
+++ b/crypto/modes/asm/ghash-x86_64.pl
@@ -0,0 +1,805 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# March, June 2010
+#
+# The module implements "4-bit" GCM GHASH function and underlying
+# single multiplication operation in GF(2^128). "4-bit" means that
+# it uses 256 bytes per-key table [+128 bytes shared table]. GHASH
+# function features so called "528B" variant utilizing additional
+# 256+16 bytes of per-key storage [+512 bytes shared table].
+# Performance results are for this streamed GHASH subroutine and are
+# expressed in cycles per processed byte, less is better:
+#
+#		gcc 3.4.x(*)	assembler
+#
+# P4		28.6		14.0		+100%
+# Opteron	19.3		7.7		+150%
+# Core2		17.8		8.1(**)		+120%
+#
+# (*)	comparison is not completely fair, because C results are
+#	for vanilla "256B" implementation, while assembler results
+#	are for "528B";-)
+# (**)	it's mystery [to me] why Core2 result is not same as for
+#	Opteron;
+
+# May 2010
+#
+# Add PCLMULQDQ version performing at 2.02 cycles per processed byte.
+# See ghash-x86.pl for background information and details about coding
+# techniques.
+#
+# Special thanks to David Woodhouse <dwmw2@infradead.org> for
+# providing access to a Westmere-based system on behalf of Intel
+# Open Source Technology Centre.
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+# common register layout
+$nlo="%rax";
+$nhi="%rbx";
+$Zlo="%r8";
+$Zhi="%r9";
+$tmp="%r10";
+$rem_4bit = "%r11";
+
+$Xi="%rdi";
+$Htbl="%rsi";
+
+# per-function register layout
+$cnt="%rcx";
+$rem="%rdx";
+
+sub LB() { my $r=shift; $r =~ s/%[er]([a-d])x/%\1l/	or
+			$r =~ s/%[er]([sd]i)/%\1l/	or
+			$r =~ s/%[er](bp)/%\1l/		or
+			$r =~ s/%(r[0-9]+)[d]?/%\1b/;   $r; }
+
+sub AUTOLOAD()		# thunk [simplified] 32-bit style perlasm
+{ my $opcode = $AUTOLOAD; $opcode =~ s/.*:://;
+  my $arg = pop;
+    $arg = "\$$arg" if ($arg*1 eq $arg);
+    $code .= "\t$opcode\t".join(',',$arg,reverse @_)."\n";
+}
+
+{ my $N;
+  sub loop() {
+  my $inp = shift;
+
+	$N++;
+$code.=<<___;
+	xor	$nlo,$nlo
+	xor	$nhi,$nhi
+	mov	`&LB("$Zlo")`,`&LB("$nlo")`
+	mov	`&LB("$Zlo")`,`&LB("$nhi")`
+	shl	\$4,`&LB("$nlo")`
+	mov	\$14,$cnt
+	mov	8($Htbl,$nlo),$Zlo
+	mov	($Htbl,$nlo),$Zhi
+	and	\$0xf0,`&LB("$nhi")`
+	mov	$Zlo,$rem
+	jmp	.Loop$N
+
+.align	16
+.Loop$N:
+	shr	\$4,$Zlo
+	and	\$0xf,$rem
+	mov	$Zhi,$tmp
+	mov	($inp,$cnt),`&LB("$nlo")`
+	shr	\$4,$Zhi
+	xor	8($Htbl,$nhi),$Zlo
+	shl	\$60,$tmp
+	xor	($Htbl,$nhi),$Zhi
+	mov	`&LB("$nlo")`,`&LB("$nhi")`
+	xor	($rem_4bit,$rem,8),$Zhi
+	mov	$Zlo,$rem
+	shl	\$4,`&LB("$nlo")`
+	xor	$tmp,$Zlo
+	dec	$cnt
+	js	.Lbreak$N
+
+	shr	\$4,$Zlo
+	and	\$0xf,$rem
+	mov	$Zhi,$tmp
+	shr	\$4,$Zhi
+	xor	8($Htbl,$nlo),$Zlo
+	shl	\$60,$tmp
+	xor	($Htbl,$nlo),$Zhi
+	and	\$0xf0,`&LB("$nhi")`
+	xor	($rem_4bit,$rem,8),$Zhi
+	mov	$Zlo,$rem
+	xor	$tmp,$Zlo
+	jmp	.Loop$N
+
+.align	16
+.Lbreak$N:
+	shr	\$4,$Zlo
+	and	\$0xf,$rem
+	mov	$Zhi,$tmp
+	shr	\$4,$Zhi
+	xor	8($Htbl,$nlo),$Zlo
+	shl	\$60,$tmp
+	xor	($Htbl,$nlo),$Zhi
+	and	\$0xf0,`&LB("$nhi")`
+	xor	($rem_4bit,$rem,8),$Zhi
+	mov	$Zlo,$rem
+	xor	$tmp,$Zlo
+
+	shr	\$4,$Zlo
+	and	\$0xf,$rem
+	mov	$Zhi,$tmp
+	shr	\$4,$Zhi
+	xor	8($Htbl,$nhi),$Zlo
+	shl	\$60,$tmp
+	xor	($Htbl,$nhi),$Zhi
+	xor	$tmp,$Zlo
+	xor	($rem_4bit,$rem,8),$Zhi
+
+	bswap	$Zlo
+	bswap	$Zhi
+___
+}}
+
+$code=<<___;
+.text
+
+.globl	gcm_gmult_4bit
+.type	gcm_gmult_4bit,\@function,2
+.align	16
+gcm_gmult_4bit:
+	push	%rbx
+	push	%rbp		# %rbp and %r12 are pushed exclusively in
+	push	%r12		# order to reuse Win64 exception handler...
+.Lgmult_prologue:
+
+	movzb	15($Xi),$Zlo
+	lea	.Lrem_4bit(%rip),$rem_4bit
+___
+	&loop	($Xi);
+$code.=<<___;
+	mov	$Zlo,8($Xi)
+	mov	$Zhi,($Xi)
+
+	mov	16(%rsp),%rbx
+	lea	24(%rsp),%rsp
+.Lgmult_epilogue:
+	ret
+.size	gcm_gmult_4bit,.-gcm_gmult_4bit
+___
+
+# per-function register layout
+$inp="%rdx";
+$len="%rcx";
+$rem_8bit=$rem_4bit;
+
+$code.=<<___;
+.globl	gcm_ghash_4bit
+.type	gcm_ghash_4bit,\@function,4
+.align	16
+gcm_ghash_4bit:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	sub	\$280,%rsp
+.Lghash_prologue:
+	mov	$inp,%r14		# reassign couple of args
+	mov	$len,%r15
+___
+{ my $inp="%r14";
+  my $dat="%edx";
+  my $len="%r15";
+  my @nhi=("%ebx","%ecx");
+  my @rem=("%r12","%r13");
+  my $Hshr4="%rbp";
+
+	&sub	($Htbl,-128);		# size optimization
+	&lea	($Hshr4,"16+128(%rsp)");
+	{ my @lo =($nlo,$nhi);
+          my @hi =($Zlo,$Zhi);
+
+	  &xor	($dat,$dat);
+	  for ($i=0,$j=-2;$i<18;$i++,$j++) {
+	    &mov	("$j(%rsp)",&LB($dat))		if ($i>1);
+	    &or		($lo[0],$tmp)			if ($i>1);
+	    &mov	(&LB($dat),&LB($lo[1]))		if ($i>0 && $i<17);
+	    &shr	($lo[1],4)			if ($i>0 && $i<17);
+	    &mov	($tmp,$hi[1])			if ($i>0 && $i<17);
+	    &shr	($hi[1],4)			if ($i>0 && $i<17);
+	    &mov	("8*$j($Hshr4)",$hi[0])		if ($i>1);
+	    &mov	($hi[0],"16*$i+0-128($Htbl)")	if ($i<16);
+	    &shl	(&LB($dat),4)			if ($i>0 && $i<17);
+	    &mov	("8*$j-128($Hshr4)",$lo[0])	if ($i>1);
+	    &mov	($lo[0],"16*$i+8-128($Htbl)")	if ($i<16);
+	    &shl	($tmp,60)			if ($i>0 && $i<17);
+
+	    push	(@lo,shift(@lo));
+	    push	(@hi,shift(@hi));
+	  }
+	}
+	&add	($Htbl,-128);
+	&mov	($Zlo,"8($Xi)");
+	&mov	($Zhi,"0($Xi)");
+	&add	($len,$inp);		# pointer to the end of data
+	&lea	($rem_8bit,".Lrem_8bit(%rip)");
+	&jmp	(".Louter_loop");
+
+$code.=".align	16\n.Louter_loop:\n";
+	&xor	($Zhi,"($inp)");
+	&mov	("%rdx","8($inp)");
+	&lea	($inp,"16($inp)");
+	&xor	("%rdx",$Zlo);
+	&mov	("($Xi)",$Zhi);
+	&mov	("8($Xi)","%rdx");
+	&shr	("%rdx",32);
+
+	&xor	($nlo,$nlo);
+	&rol	($dat,8);
+	&mov	(&LB($nlo),&LB($dat));
+	&movz	($nhi[0],&LB($dat));
+	&shl	(&LB($nlo),4);
+	&shr	($nhi[0],4);
+
+	for ($j=11,$i=0;$i<15;$i++) {
+	    &rol	($dat,8);
+	    &xor	($Zlo,"8($Htbl,$nlo)")			if ($i>0);
+	    &xor	($Zhi,"($Htbl,$nlo)")			if ($i>0);
+	    &mov	($Zlo,"8($Htbl,$nlo)")			if ($i==0);
+	    &mov	($Zhi,"($Htbl,$nlo)")			if ($i==0);
+
+	    &mov	(&LB($nlo),&LB($dat));
+	    &xor	($Zlo,$tmp)				if ($i>0);
+	    &movzw	($rem[1],"($rem_8bit,$rem[1],2)")	if ($i>0);
+
+	    &movz	($nhi[1],&LB($dat));
+	    &shl	(&LB($nlo),4);
+	    &movzb	($rem[0],"(%rsp,$nhi[0])");
+
+	    &shr	($nhi[1],4)				if ($i<14);
+	    &and	($nhi[1],0xf0)				if ($i==14);
+	    &shl	($rem[1],48)				if ($i>0);
+	    &xor	($rem[0],$Zlo);
+
+	    &mov	($tmp,$Zhi);
+	    &xor	($Zhi,$rem[1])				if ($i>0);
+	    &shr	($Zlo,8);
+
+	    &movz	($rem[0],&LB($rem[0]));
+	    &mov	($dat,"$j($Xi)")			if (--$j%4==0);
+	    &shr	($Zhi,8);
+
+	    &xor	($Zlo,"-128($Hshr4,$nhi[0],8)");
+	    &shl	($tmp,56);
+	    &xor	($Zhi,"($Hshr4,$nhi[0],8)");
+
+	    unshift	(@nhi,pop(@nhi));		# "rotate" registers
+	    unshift	(@rem,pop(@rem));
+	}
+	&movzw	($rem[1],"($rem_8bit,$rem[1],2)");
+	&xor	($Zlo,"8($Htbl,$nlo)");
+	&xor	($Zhi,"($Htbl,$nlo)");
+
+	&shl	($rem[1],48);
+	&xor	($Zlo,$tmp);
+
+	&xor	($Zhi,$rem[1]);
+	&movz	($rem[0],&LB($Zlo));
+	&shr	($Zlo,4);
+
+	&mov	($tmp,$Zhi);
+	&shl	(&LB($rem[0]),4);
+	&shr	($Zhi,4);
+
+	&xor	($Zlo,"8($Htbl,$nhi[0])");
+	&movzw	($rem[0],"($rem_8bit,$rem[0],2)");
+	&shl	($tmp,60);
+
+	&xor	($Zhi,"($Htbl,$nhi[0])");
+	&xor	($Zlo,$tmp);
+	&shl	($rem[0],48);
+
+	&bswap	($Zlo);
+	&xor	($Zhi,$rem[0]);
+
+	&bswap	($Zhi);
+	&cmp	($inp,$len);
+	&jb	(".Louter_loop");
+}
+$code.=<<___;
+	mov	$Zlo,8($Xi)
+	mov	$Zhi,($Xi)
+
+	lea	280(%rsp),%rsi
+	mov	0(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lghash_epilogue:
+	ret
+.size	gcm_ghash_4bit,.-gcm_ghash_4bit
+___
+
+######################################################################
+# PCLMULQDQ version.
+
+@_4args=$win64?	("%rcx","%rdx","%r8", "%r9") :	# Win64 order
+		("%rdi","%rsi","%rdx","%rcx");	# Unix order
+
+($Xi,$Xhi)=("%xmm0","%xmm1");	$Hkey="%xmm2";
+($T1,$T2,$T3)=("%xmm3","%xmm4","%xmm5");
+
+sub clmul64x64_T2 {	# minimal register pressure
+my ($Xhi,$Xi,$Hkey,$modulo)=@_;
+
+$code.=<<___ if (!defined($modulo));
+	movdqa		$Xi,$Xhi		#
+	pshufd		\$0b01001110,$Xi,$T1
+	pshufd		\$0b01001110,$Hkey,$T2
+	pxor		$Xi,$T1			#
+	pxor		$Hkey,$T2
+___
+$code.=<<___;
+	pclmulqdq	\$0x00,$Hkey,$Xi	#######
+	pclmulqdq	\$0x11,$Hkey,$Xhi	#######
+	pclmulqdq	\$0x00,$T2,$T1		#######
+	pxor		$Xi,$T1			#
+	pxor		$Xhi,$T1		#
+
+	movdqa		$T1,$T2			#
+	psrldq		\$8,$T1
+	pslldq		\$8,$T2			#
+	pxor		$T1,$Xhi
+	pxor		$T2,$Xi			#
+___
+}
+
+sub reduction_alg9 {	# 17/13 times faster than Intel version
+my ($Xhi,$Xi) = @_;
+
+$code.=<<___;
+	# 1st phase
+	movdqa		$Xi,$T1			#
+	psllq		\$1,$Xi
+	pxor		$T1,$Xi			#
+	psllq		\$5,$Xi			#
+	pxor		$T1,$Xi			#
+	psllq		\$57,$Xi		#
+	movdqa		$Xi,$T2			#
+	pslldq		\$8,$Xi
+	psrldq		\$8,$T2			#	
+	pxor		$T1,$Xi
+	pxor		$T2,$Xhi		#
+
+	# 2nd phase
+	movdqa		$Xi,$T2
+	psrlq		\$5,$Xi
+	pxor		$T2,$Xi			#
+	psrlq		\$1,$Xi			#
+	pxor		$T2,$Xi			#
+	pxor		$Xhi,$T2
+	psrlq		\$1,$Xi			#
+	pxor		$T2,$Xi			#
+___
+}
+
+{ my ($Htbl,$Xip)=@_4args;
+
+$code.=<<___;
+.globl	gcm_init_clmul
+.type	gcm_init_clmul,\@abi-omnipotent
+.align	16
+gcm_init_clmul:
+	movdqu		($Xip),$Hkey
+	pshufd		\$0b01001110,$Hkey,$Hkey	# dword swap
+
+	# <<1 twist
+	pshufd		\$0b11111111,$Hkey,$T2	# broadcast uppermost dword
+	movdqa		$Hkey,$T1
+	psllq		\$1,$Hkey
+	pxor		$T3,$T3			#
+	psrlq		\$63,$T1
+	pcmpgtd		$T2,$T3			# broadcast carry bit
+	pslldq		\$8,$T1
+	por		$T1,$Hkey		# H<<=1
+
+	# magic reduction
+	pand		.L0x1c2_polynomial(%rip),$T3
+	pxor		$T3,$Hkey		# if(carry) H^=0x1c2_polynomial
+
+	# calculate H^2
+	movdqa		$Hkey,$Xi
+___
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey);
+	&reduction_alg9	($Xhi,$Xi);
+$code.=<<___;
+	movdqu		$Hkey,($Htbl)		# save H
+	movdqu		$Xi,16($Htbl)		# save H^2
+	ret
+.size	gcm_init_clmul,.-gcm_init_clmul
+___
+}
+
+{ my ($Xip,$Htbl)=@_4args;
+
+$code.=<<___;
+.globl	gcm_gmult_clmul
+.type	gcm_gmult_clmul,\@abi-omnipotent
+.align	16
+gcm_gmult_clmul:
+	movdqu		($Xip),$Xi
+	movdqa		.Lbswap_mask(%rip),$T3
+	movdqu		($Htbl),$Hkey
+	pshufb		$T3,$Xi
+___
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey);
+	&reduction_alg9	($Xhi,$Xi);
+$code.=<<___;
+	pshufb		$T3,$Xi
+	movdqu		$Xi,($Xip)
+	ret
+.size	gcm_gmult_clmul,.-gcm_gmult_clmul
+___
+}
+
+{ my ($Xip,$Htbl,$inp,$len)=@_4args;
+  my $Xn="%xmm6";
+  my $Xhn="%xmm7";
+  my $Hkey2="%xmm8";
+  my $T1n="%xmm9";
+  my $T2n="%xmm10";
+
+$code.=<<___;
+.globl	gcm_ghash_clmul
+.type	gcm_ghash_clmul,\@abi-omnipotent
+.align	16
+gcm_ghash_clmul:
+___
+$code.=<<___ if ($win64);
+.LSEH_begin_gcm_ghash_clmul:
+	# I can't trust assembler to use specific encoding:-(
+	.byte	0x48,0x83,0xec,0x58		#sub	\$0x58,%rsp
+	.byte	0x0f,0x29,0x34,0x24		#movaps	%xmm6,(%rsp)
+	.byte	0x0f,0x29,0x7c,0x24,0x10	#movdqa	%xmm7,0x10(%rsp)
+	.byte	0x44,0x0f,0x29,0x44,0x24,0x20	#movaps	%xmm8,0x20(%rsp)
+	.byte	0x44,0x0f,0x29,0x4c,0x24,0x30	#movaps	%xmm9,0x30(%rsp)
+	.byte	0x44,0x0f,0x29,0x54,0x24,0x40	#movaps	%xmm10,0x40(%rsp)
+___
+$code.=<<___;
+	movdqa		.Lbswap_mask(%rip),$T3
+
+	movdqu		($Xip),$Xi
+	movdqu		($Htbl),$Hkey
+	pshufb		$T3,$Xi
+
+	sub		\$0x10,$len
+	jz		.Lodd_tail
+
+	movdqu		16($Htbl),$Hkey2
+	#######
+	# Xi+2 =[H*(Ii+1 + Xi+1)] mod P =
+	#	[(H*Ii+1) + (H*Xi+1)] mod P =
+	#	[(H*Ii+1) + H^2*(Ii+Xi)] mod P
+	#
+	movdqu		($inp),$T1		# Ii
+	movdqu		16($inp),$Xn		# Ii+1
+	pshufb		$T3,$T1
+	pshufb		$T3,$Xn
+	pxor		$T1,$Xi			# Ii+Xi
+___
+	&clmul64x64_T2	($Xhn,$Xn,$Hkey);	# H*Ii+1
+$code.=<<___;
+	movdqa		$Xi,$Xhi		#
+	pshufd		\$0b01001110,$Xi,$T1
+	pshufd		\$0b01001110,$Hkey2,$T2
+	pxor		$Xi,$T1			#
+	pxor		$Hkey2,$T2
+
+	lea		32($inp),$inp		# i+=2
+	sub		\$0x20,$len
+	jbe		.Leven_tail
+
+.Lmod_loop:
+___
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey2,1);	# H^2*(Ii+Xi)
+$code.=<<___;
+	movdqu		($inp),$T1		# Ii
+	pxor		$Xn,$Xi			# (H*Ii+1) + H^2*(Ii+Xi)
+	pxor		$Xhn,$Xhi
+
+	movdqu		16($inp),$Xn		# Ii+1
+	pshufb		$T3,$T1
+	pshufb		$T3,$Xn
+
+	movdqa		$Xn,$Xhn		#
+	pshufd		\$0b01001110,$Xn,$T1n
+	pshufd		\$0b01001110,$Hkey,$T2n
+	pxor		$Xn,$T1n		#
+	pxor		$Hkey,$T2n
+	 pxor		$T1,$Xhi		# "Ii+Xi", consume early
+
+	  movdqa	$Xi,$T1			# 1st phase
+	  psllq		\$1,$Xi
+	  pxor		$T1,$Xi			#
+	  psllq		\$5,$Xi			#
+	  pxor		$T1,$Xi			#
+	pclmulqdq	\$0x00,$Hkey,$Xn	#######
+	  psllq		\$57,$Xi		#
+	  movdqa	$Xi,$T2			#
+	  pslldq	\$8,$Xi
+	  psrldq	\$8,$T2			#	
+	  pxor		$T1,$Xi
+	  pxor		$T2,$Xhi		#
+
+	pclmulqdq	\$0x11,$Hkey,$Xhn	#######
+	  movdqa	$Xi,$T2			# 2nd phase
+	  psrlq		\$5,$Xi
+	  pxor		$T2,$Xi			#
+	  psrlq		\$1,$Xi			#
+	  pxor		$T2,$Xi			#
+	  pxor		$Xhi,$T2
+	  psrlq		\$1,$Xi			#
+	  pxor		$T2,$Xi			#
+
+	pclmulqdq	\$0x00,$T2n,$T1n	#######
+	 movdqa		$Xi,$Xhi		#
+	 pshufd		\$0b01001110,$Xi,$T1
+	 pshufd		\$0b01001110,$Hkey2,$T2
+	 pxor		$Xi,$T1			#
+	 pxor		$Hkey2,$T2
+
+	pxor		$Xn,$T1n		#
+	pxor		$Xhn,$T1n		#
+	movdqa		$T1n,$T2n		#
+	psrldq		\$8,$T1n
+	pslldq		\$8,$T2n		#
+	pxor		$T1n,$Xhn
+	pxor		$T2n,$Xn		#
+
+	lea		32($inp),$inp
+	sub		\$0x20,$len
+	ja		.Lmod_loop
+
+.Leven_tail:
+___
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey2,1);	# H^2*(Ii+Xi)
+$code.=<<___;
+	pxor		$Xn,$Xi			# (H*Ii+1) + H^2*(Ii+Xi)
+	pxor		$Xhn,$Xhi
+___
+	&reduction_alg9	($Xhi,$Xi);
+$code.=<<___;
+	test		$len,$len
+	jnz		.Ldone
+
+.Lodd_tail:
+	movdqu		($inp),$T1		# Ii
+	pshufb		$T3,$T1
+	pxor		$T1,$Xi			# Ii+Xi
+___
+	&clmul64x64_T2	($Xhi,$Xi,$Hkey);	# H*(Ii+Xi)
+	&reduction_alg9	($Xhi,$Xi);
+$code.=<<___;
+.Ldone:
+	pshufb		$T3,$Xi
+	movdqu		$Xi,($Xip)
+___
+$code.=<<___ if ($win64);
+	movaps	(%rsp),%xmm6
+	movaps	0x10(%rsp),%xmm7
+	movaps	0x20(%rsp),%xmm8
+	movaps	0x30(%rsp),%xmm9
+	movaps	0x40(%rsp),%xmm10
+	add	\$0x58,%rsp
+___
+$code.=<<___;
+	ret
+.LSEH_end_gcm_ghash_clmul:
+.size	gcm_ghash_clmul,.-gcm_ghash_clmul
+___
+}
+
+$code.=<<___;
+.align	64
+.Lbswap_mask:
+	.byte	15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0
+.L0x1c2_polynomial:
+	.byte	1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0xc2
+.align	64
+.type	.Lrem_4bit,\@object
+.Lrem_4bit:
+	.long	0,`0x0000<<16`,0,`0x1C20<<16`,0,`0x3840<<16`,0,`0x2460<<16`
+	.long	0,`0x7080<<16`,0,`0x6CA0<<16`,0,`0x48C0<<16`,0,`0x54E0<<16`
+	.long	0,`0xE100<<16`,0,`0xFD20<<16`,0,`0xD940<<16`,0,`0xC560<<16`
+	.long	0,`0x9180<<16`,0,`0x8DA0<<16`,0,`0xA9C0<<16`,0,`0xB5E0<<16`
+.type	.Lrem_8bit,\@object
+.Lrem_8bit:
+	.value	0x0000,0x01C2,0x0384,0x0246,0x0708,0x06CA,0x048C,0x054E
+	.value	0x0E10,0x0FD2,0x0D94,0x0C56,0x0918,0x08DA,0x0A9C,0x0B5E
+	.value	0x1C20,0x1DE2,0x1FA4,0x1E66,0x1B28,0x1AEA,0x18AC,0x196E
+	.value	0x1230,0x13F2,0x11B4,0x1076,0x1538,0x14FA,0x16BC,0x177E
+	.value	0x3840,0x3982,0x3BC4,0x3A06,0x3F48,0x3E8A,0x3CCC,0x3D0E
+	.value	0x3650,0x3792,0x35D4,0x3416,0x3158,0x309A,0x32DC,0x331E
+	.value	0x2460,0x25A2,0x27E4,0x2626,0x2368,0x22AA,0x20EC,0x212E
+	.value	0x2A70,0x2BB2,0x29F4,0x2836,0x2D78,0x2CBA,0x2EFC,0x2F3E
+	.value	0x7080,0x7142,0x7304,0x72C6,0x7788,0x764A,0x740C,0x75CE
+	.value	0x7E90,0x7F52,0x7D14,0x7CD6,0x7998,0x785A,0x7A1C,0x7BDE
+	.value	0x6CA0,0x6D62,0x6F24,0x6EE6,0x6BA8,0x6A6A,0x682C,0x69EE
+	.value	0x62B0,0x6372,0x6134,0x60F6,0x65B8,0x647A,0x663C,0x67FE
+	.value	0x48C0,0x4902,0x4B44,0x4A86,0x4FC8,0x4E0A,0x4C4C,0x4D8E
+	.value	0x46D0,0x4712,0x4554,0x4496,0x41D8,0x401A,0x425C,0x439E
+	.value	0x54E0,0x5522,0x5764,0x56A6,0x53E8,0x522A,0x506C,0x51AE
+	.value	0x5AF0,0x5B32,0x5974,0x58B6,0x5DF8,0x5C3A,0x5E7C,0x5FBE
+	.value	0xE100,0xE0C2,0xE284,0xE346,0xE608,0xE7CA,0xE58C,0xE44E
+	.value	0xEF10,0xEED2,0xEC94,0xED56,0xE818,0xE9DA,0xEB9C,0xEA5E
+	.value	0xFD20,0xFCE2,0xFEA4,0xFF66,0xFA28,0xFBEA,0xF9AC,0xF86E
+	.value	0xF330,0xF2F2,0xF0B4,0xF176,0xF438,0xF5FA,0xF7BC,0xF67E
+	.value	0xD940,0xD882,0xDAC4,0xDB06,0xDE48,0xDF8A,0xDDCC,0xDC0E
+	.value	0xD750,0xD692,0xD4D4,0xD516,0xD058,0xD19A,0xD3DC,0xD21E
+	.value	0xC560,0xC4A2,0xC6E4,0xC726,0xC268,0xC3AA,0xC1EC,0xC02E
+	.value	0xCB70,0xCAB2,0xC8F4,0xC936,0xCC78,0xCDBA,0xCFFC,0xCE3E
+	.value	0x9180,0x9042,0x9204,0x93C6,0x9688,0x974A,0x950C,0x94CE
+	.value	0x9F90,0x9E52,0x9C14,0x9DD6,0x9898,0x995A,0x9B1C,0x9ADE
+	.value	0x8DA0,0x8C62,0x8E24,0x8FE6,0x8AA8,0x8B6A,0x892C,0x88EE
+	.value	0x83B0,0x8272,0x8034,0x81F6,0x84B8,0x857A,0x873C,0x86FE
+	.value	0xA9C0,0xA802,0xAA44,0xAB86,0xAEC8,0xAF0A,0xAD4C,0xAC8E
+	.value	0xA7D0,0xA612,0xA454,0xA596,0xA0D8,0xA11A,0xA35C,0xA29E
+	.value	0xB5E0,0xB422,0xB664,0xB7A6,0xB2E8,0xB32A,0xB16C,0xB0AE
+	.value	0xBBF0,0xBA32,0xB874,0xB9B6,0xBCF8,0xBD3A,0xBF7C,0xBEBE
+
+.asciz	"GHASH for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+.align	64
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lin_prologue
+
+	lea	24(%rax),%rax		# adjust "rsp"
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$`1232/8`,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	se_handler,.-se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_gcm_gmult_4bit
+	.rva	.LSEH_end_gcm_gmult_4bit
+	.rva	.LSEH_info_gcm_gmult_4bit
+
+	.rva	.LSEH_begin_gcm_ghash_4bit
+	.rva	.LSEH_end_gcm_ghash_4bit
+	.rva	.LSEH_info_gcm_ghash_4bit
+
+	.rva	.LSEH_begin_gcm_ghash_clmul
+	.rva	.LSEH_end_gcm_ghash_clmul
+	.rva	.LSEH_info_gcm_ghash_clmul
+
+.section	.xdata
+.align	8
+.LSEH_info_gcm_gmult_4bit:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lgmult_prologue,.Lgmult_epilogue	# HandlerData
+.LSEH_info_gcm_ghash_4bit:
+	.byte	9,0,0,0
+	.rva	se_handler
+	.rva	.Lghash_prologue,.Lghash_epilogue	# HandlerData
+.LSEH_info_gcm_ghash_clmul:
+	.byte	0x01,0x1f,0x0b,0x00
+	.byte	0x1f,0xa8,0x04,0x00	#movaps 0x40(rsp),xmm10
+	.byte	0x19,0x98,0x03,0x00	#movaps 0x30(rsp),xmm9
+	.byte	0x13,0x88,0x02,0x00	#movaps 0x20(rsp),xmm8
+	.byte	0x0d,0x78,0x01,0x00	#movaps 0x10(rsp),xmm7
+	.byte	0x08,0x68,0x00,0x00	#movaps (rsp),xmm6
+	.byte	0x04,0xa2,0x00,0x00	#sub	rsp,0x58
+___
+}
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+
+print $code;
+
+close STDOUT;
diff --git a/crypto/modes/cbc128.c b/crypto/modes/cbc128.c
new file mode 100644
index 000000000000..3d3782cbe118
--- /dev/null
+++ b/crypto/modes/cbc128.c
@@ -0,0 +1,202 @@
+/* ====================================================================
+ * Copyright (c) 2008 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ */
+
+#include <openssl/crypto.h>
+#include "modes_lcl.h"
+#include <string.h>
+
+#ifndef MODES_DEBUG
+# ifndef NDEBUG
+#  define NDEBUG
+# endif
+#endif
+#include <assert.h>
+
+#ifndef STRICT_ALIGNMENT
+#  define STRICT_ALIGNMENT 0
+#endif
+
+void CRYPTO_cbc128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block)
+{
+	size_t n;
+	const unsigned char *iv = ivec;
+
+	assert(in && out && key && ivec);
+
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+	if (STRICT_ALIGNMENT &&
+	    ((size_t)in|(size_t)out|(size_t)ivec)%sizeof(size_t) != 0) {
+		while (len>=16) {
+			for(n=0; n<16; ++n)
+				out[n] = in[n] ^ iv[n];
+			(*block)(out, out, key);
+			iv = out;
+			len -= 16;
+			in  += 16;
+			out += 16;
+		}
+	} else {
+		while (len>=16) {
+			for(n=0; n<16; n+=sizeof(size_t))
+				*(size_t*)(out+n) =
+				*(size_t*)(in+n) ^ *(size_t*)(iv+n);
+			(*block)(out, out, key);
+			iv = out;
+			len -= 16;
+			in  += 16;
+			out += 16;
+		}
+	}
+#endif
+	while (len) {
+		for(n=0; n<16 && n<len; ++n)
+			out[n] = in[n] ^ iv[n];
+		for(; n<16; ++n)
+			out[n] = iv[n];
+		(*block)(out, out, key);
+		iv = out;
+		if (len<=16) break;
+		len -= 16;
+		in  += 16;
+		out += 16;
+	}
+	memcpy(ivec,iv,16);
+}
+
+void CRYPTO_cbc128_decrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block)
+{
+	size_t n;
+	union { size_t align; unsigned char c[16]; } tmp;
+
+	assert(in && out && key && ivec);
+
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+	if (in != out) {
+		const unsigned char *iv = ivec;
+
+		if (STRICT_ALIGNMENT &&
+		    ((size_t)in|(size_t)out|(size_t)ivec)%sizeof(size_t) != 0) {
+			while (len>=16) {
+				(*block)(in, out, key);
+				for(n=0; n<16; ++n)
+					out[n] ^= iv[n];
+				iv = in;
+				len -= 16;
+				in  += 16;
+				out += 16;
+			}
+		}
+		else {
+			while (len>=16) {
+				(*block)(in, out, key);
+				for(n=0; n<16; n+=sizeof(size_t))
+					*(size_t *)(out+n) ^= *(size_t *)(iv+n);
+				iv = in;
+				len -= 16;
+				in  += 16;
+				out += 16;
+			}
+		}
+		memcpy(ivec,iv,16);
+	} else {
+		if (STRICT_ALIGNMENT &&
+		    ((size_t)in|(size_t)out|(size_t)ivec)%sizeof(size_t) != 0) {
+			unsigned char c;
+			while (len>=16) {
+				(*block)(in, tmp.c, key);
+				for(n=0; n<16; ++n) {
+					c = in[n];
+					out[n] = tmp.c[n] ^ ivec[n];
+					ivec[n] = c;
+				}
+				len -= 16;
+				in  += 16;
+				out += 16;
+			}
+		}
+		else {
+			size_t c;
+			while (len>=16) {
+				(*block)(in, tmp.c, key);
+				for(n=0; n<16; n+=sizeof(size_t)) {
+					c = *(size_t *)(in+n);
+					*(size_t *)(out+n) =
+					*(size_t *)(tmp.c+n) ^ *(size_t *)(ivec+n);
+					*(size_t *)(ivec+n) = c;
+				}
+				len -= 16;
+				in  += 16;
+				out += 16;
+			}
+		}
+	}
+#endif
+	while (len) {
+		unsigned char c;
+		(*block)(in, tmp.c, key);
+		for(n=0; n<16 && n<len; ++n) {
+			c = in[n];
+			out[n] = tmp.c[n] ^ ivec[n];
+			ivec[n] = c;
+		}
+		if (len<=16) {
+			for (; n<16; ++n)
+				ivec[n] = in[n];
+			break;
+		}
+		len -= 16;
+		in  += 16;
+		out += 16;
+	}
+}
diff --git a/crypto/modes/ccm128.c b/crypto/modes/ccm128.c
new file mode 100644
index 000000000000..c9b35e5b35e5
--- /dev/null
+++ b/crypto/modes/ccm128.c
@@ -0,0 +1,441 @@
+/* ====================================================================
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#include <openssl/crypto.h>
+#include "modes_lcl.h"
+#include <string.h>
+
+#ifndef MODES_DEBUG
+# ifndef NDEBUG
+#  define NDEBUG
+# endif
+#endif
+#include <assert.h>
+
+/* First you setup M and L parameters and pass the key schedule.
+ * This is called once per session setup... */
+void CRYPTO_ccm128_init(CCM128_CONTEXT *ctx,
+	unsigned int M,unsigned int L,void *key,block128_f block)
+{
+	memset(ctx->nonce.c,0,sizeof(ctx->nonce.c));
+	ctx->nonce.c[0] = ((u8)(L-1)&7) | (u8)(((M-2)/2)&7)<<3;
+	ctx->blocks = 0;
+	ctx->block = block;
+	ctx->key = key;
+}
+
+/* !!! Following interfaces are to be called *once* per packet !!! */
+
+/* Then you setup per-message nonce and pass the length of the message */
+int CRYPTO_ccm128_setiv(CCM128_CONTEXT *ctx,
+	const unsigned char *nonce,size_t nlen,size_t mlen)
+{
+	unsigned int L = ctx->nonce.c[0]&7;	/* the L parameter */
+
+	if (nlen<(14-L)) return -1;		/* nonce is too short */
+
+	if (sizeof(mlen)==8 && L>=3) {
+		ctx->nonce.c[8]  = (u8)(mlen>>(56%(sizeof(mlen)*8)));
+		ctx->nonce.c[9]  = (u8)(mlen>>(48%(sizeof(mlen)*8)));
+		ctx->nonce.c[10] = (u8)(mlen>>(40%(sizeof(mlen)*8)));
+		ctx->nonce.c[11] = (u8)(mlen>>(32%(sizeof(mlen)*8)));
+	}
+	else
+		*(u32*)(&ctx->nonce.c[8]) = 0;
+
+	ctx->nonce.c[12] = (u8)(mlen>>24);
+	ctx->nonce.c[13] = (u8)(mlen>>16);
+	ctx->nonce.c[14] = (u8)(mlen>>8);
+	ctx->nonce.c[15] = (u8)mlen;
+
+	ctx->nonce.c[0] &= ~0x40;	/* clear Adata flag */
+	memcpy(&ctx->nonce.c[1],nonce,14-L);
+
+	return 0;
+}
+
+/* Then you pass additional authentication data, this is optional */
+void CRYPTO_ccm128_aad(CCM128_CONTEXT *ctx,
+	const unsigned char *aad,size_t alen)
+{	unsigned int i;
+	block128_f block = ctx->block;
+
+	if (alen==0) return;
+
+	ctx->nonce.c[0] |= 0x40;	/* set Adata flag */
+	(*block)(ctx->nonce.c,ctx->cmac.c,ctx->key),
+	ctx->blocks++;
+
+	if (alen<(0x10000-0x100)) {
+		ctx->cmac.c[0] ^= (u8)(alen>>8);
+		ctx->cmac.c[1] ^= (u8)alen;
+		i=2;
+	}
+	else if (sizeof(alen)==8 && alen>=(size_t)1<<(32%(sizeof(alen)*8))) {
+		ctx->cmac.c[0] ^= 0xFF;
+		ctx->cmac.c[1] ^= 0xFF;
+		ctx->cmac.c[2] ^= (u8)(alen>>(56%(sizeof(alen)*8)));
+		ctx->cmac.c[3] ^= (u8)(alen>>(48%(sizeof(alen)*8)));
+		ctx->cmac.c[4] ^= (u8)(alen>>(40%(sizeof(alen)*8)));
+		ctx->cmac.c[5] ^= (u8)(alen>>(32%(sizeof(alen)*8)));
+		ctx->cmac.c[6] ^= (u8)(alen>>24);
+		ctx->cmac.c[7] ^= (u8)(alen>>16);
+		ctx->cmac.c[8] ^= (u8)(alen>>8);
+		ctx->cmac.c[9] ^= (u8)alen;
+		i=10;
+	}
+	else {
+		ctx->cmac.c[0] ^= 0xFF;
+		ctx->cmac.c[1] ^= 0xFE;
+		ctx->cmac.c[2] ^= (u8)(alen>>24);
+		ctx->cmac.c[3] ^= (u8)(alen>>16);
+		ctx->cmac.c[4] ^= (u8)(alen>>8);
+		ctx->cmac.c[5] ^= (u8)alen;
+		i=6;
+	}
+
+	do {
+		for(;i<16 && alen;++i,++aad,--alen)
+			ctx->cmac.c[i] ^= *aad;
+		(*block)(ctx->cmac.c,ctx->cmac.c,ctx->key),
+		ctx->blocks++;
+		i=0;
+	} while (alen);
+}
+
+/* Finally you encrypt or decrypt the message */
+
+/* counter part of nonce may not be larger than L*8 bits,
+ * L is not larger than 8, therefore 64-bit counter... */
+static void ctr64_inc(unsigned char *counter) {
+	unsigned int n=8;
+	u8  c;
+
+	counter += 8;
+	do {
+		--n;
+		c = counter[n];
+		++c;
+		counter[n] = c;
+		if (c) return;
+	} while (n);
+}
+
+int CRYPTO_ccm128_encrypt(CCM128_CONTEXT *ctx,
+	const unsigned char *inp, unsigned char *out,
+	size_t len)
+{
+	size_t		n;
+	unsigned int	i,L;
+	unsigned char	flags0	= ctx->nonce.c[0];
+	block128_f	block	= ctx->block;
+	void *		key	= ctx->key;
+	union { u64 u[2]; u8 c[16]; } scratch;
+
+	if (!(flags0&0x40))
+		(*block)(ctx->nonce.c,ctx->cmac.c,key),
+		ctx->blocks++;
+
+	ctx->nonce.c[0] = L = flags0&7;
+	for (n=0,i=15-L;i<15;++i) {
+		n |= ctx->nonce.c[i];
+		ctx->nonce.c[i]=0;
+		n <<= 8;
+	}
+	n |= ctx->nonce.c[15];	/* reconstructed length */
+	ctx->nonce.c[15]=1;
+
+	if (n!=len) return -1;	/* length mismatch */
+
+	ctx->blocks += ((len+15)>>3)|1;
+	if (ctx->blocks > (U64(1)<<61))	return -2; /* too much data */
+
+	while (len>=16) {
+#if defined(STRICT_ALIGNMENT)
+		union { u64 u[2]; u8 c[16]; } temp;
+
+		memcpy (temp.c,inp,16);
+		ctx->cmac.u[0] ^= temp.u[0];
+		ctx->cmac.u[1] ^= temp.u[1];
+#else
+		ctx->cmac.u[0] ^= ((u64*)inp)[0];
+		ctx->cmac.u[1] ^= ((u64*)inp)[1];
+#endif
+		(*block)(ctx->cmac.c,ctx->cmac.c,key);
+		(*block)(ctx->nonce.c,scratch.c,key);
+		ctr64_inc(ctx->nonce.c);
+#if defined(STRICT_ALIGNMENT)
+		temp.u[0] ^= scratch.u[0];
+		temp.u[1] ^= scratch.u[1];
+		memcpy(out,temp.c,16);
+#else
+		((u64*)out)[0] = scratch.u[0]^((u64*)inp)[0];
+		((u64*)out)[1] = scratch.u[1]^((u64*)inp)[1];
+#endif
+		inp += 16;
+		out += 16;
+		len -= 16;
+	}
+
+	if (len) {
+		for (i=0; i<len; ++i) ctx->cmac.c[i] ^= inp[i];
+		(*block)(ctx->cmac.c,ctx->cmac.c,key);
+		(*block)(ctx->nonce.c,scratch.c,key);
+		for (i=0; i<len; ++i) out[i] = scratch.c[i]^inp[i];
+	}
+
+	for (i=15-L;i<16;++i)
+		ctx->nonce.c[i]=0;
+
+	(*block)(ctx->nonce.c,scratch.c,key);
+	ctx->cmac.u[0] ^= scratch.u[0];
+	ctx->cmac.u[1] ^= scratch.u[1];
+
+	ctx->nonce.c[0] = flags0;
+
+	return 0;
+}
+
+int CRYPTO_ccm128_decrypt(CCM128_CONTEXT *ctx,
+	const unsigned char *inp, unsigned char *out,
+	size_t len)
+{
+	size_t		n;
+	unsigned int	i,L;
+	unsigned char	flags0	= ctx->nonce.c[0];
+	block128_f	block	= ctx->block;
+	void *		key	= ctx->key;
+	union { u64 u[2]; u8 c[16]; } scratch;
+
+	if (!(flags0&0x40))
+		(*block)(ctx->nonce.c,ctx->cmac.c,key);
+
+	ctx->nonce.c[0] = L = flags0&7;
+	for (n=0,i=15-L;i<15;++i) {
+		n |= ctx->nonce.c[i];
+		ctx->nonce.c[i]=0;
+		n <<= 8;
+	}
+	n |= ctx->nonce.c[15];	/* reconstructed length */
+	ctx->nonce.c[15]=1;
+
+	if (n!=len) return -1;
+
+	while (len>=16) {
+#if defined(STRICT_ALIGNMENT)
+		union { u64 u[2]; u8 c[16]; } temp;
+#endif
+		(*block)(ctx->nonce.c,scratch.c,key);
+		ctr64_inc(ctx->nonce.c);
+#if defined(STRICT_ALIGNMENT)
+		memcpy (temp.c,inp,16);
+		ctx->cmac.u[0] ^= (scratch.u[0] ^= temp.u[0]);
+		ctx->cmac.u[1] ^= (scratch.u[1] ^= temp.u[1]);
+		memcpy (out,scratch.c,16);
+#else
+		ctx->cmac.u[0] ^= (((u64*)out)[0] = scratch.u[0]^((u64*)inp)[0]);
+		ctx->cmac.u[1] ^= (((u64*)out)[1] = scratch.u[1]^((u64*)inp)[1]);
+#endif
+		(*block)(ctx->cmac.c,ctx->cmac.c,key);
+
+		inp += 16;
+		out += 16;
+		len -= 16;
+	}
+
+	if (len) {
+		(*block)(ctx->nonce.c,scratch.c,key);
+		for (i=0; i<len; ++i)
+			ctx->cmac.c[i] ^= (out[i] = scratch.c[i]^inp[i]);
+		(*block)(ctx->cmac.c,ctx->cmac.c,key);
+	}
+
+	for (i=15-L;i<16;++i)
+		ctx->nonce.c[i]=0;
+
+	(*block)(ctx->nonce.c,scratch.c,key);
+	ctx->cmac.u[0] ^= scratch.u[0];
+	ctx->cmac.u[1] ^= scratch.u[1];
+
+	ctx->nonce.c[0] = flags0;
+
+	return 0;
+}
+
+static void ctr64_add (unsigned char *counter,size_t inc)
+{	size_t n=8, val=0;
+
+	counter += 8;
+	do {
+		--n;
+		val += counter[n] + (inc&0xff);
+		counter[n] = (unsigned char)val;
+		val >>= 8;	/* carry bit */
+		inc >>= 8;
+	} while(n && (inc || val));
+}
+
+int CRYPTO_ccm128_encrypt_ccm64(CCM128_CONTEXT *ctx,
+	const unsigned char *inp, unsigned char *out,
+	size_t len,ccm128_f stream)
+{
+	size_t		n;
+	unsigned int	i,L;
+	unsigned char	flags0	= ctx->nonce.c[0];
+	block128_f	block	= ctx->block;
+	void *		key	= ctx->key;
+	union { u64 u[2]; u8 c[16]; } scratch;
+
+	if (!(flags0&0x40))
+		(*block)(ctx->nonce.c,ctx->cmac.c,key),
+		ctx->blocks++;
+
+	ctx->nonce.c[0] = L = flags0&7;
+	for (n=0,i=15-L;i<15;++i) {
+		n |= ctx->nonce.c[i];
+		ctx->nonce.c[i]=0;
+		n <<= 8;
+	}
+	n |= ctx->nonce.c[15];	/* reconstructed length */
+	ctx->nonce.c[15]=1;
+
+	if (n!=len) return -1;	/* length mismatch */
+
+	ctx->blocks += ((len+15)>>3)|1;
+	if (ctx->blocks > (U64(1)<<61))	return -2; /* too much data */
+
+	if ((n=len/16)) {
+		(*stream)(inp,out,n,key,ctx->nonce.c,ctx->cmac.c);
+		n   *= 16;
+		inp += n;
+		out += n;
+		len -= n;
+		if (len) ctr64_add(ctx->nonce.c,n/16);
+	}
+
+	if (len) {
+		for (i=0; i<len; ++i) ctx->cmac.c[i] ^= inp[i];
+		(*block)(ctx->cmac.c,ctx->cmac.c,key);
+		(*block)(ctx->nonce.c,scratch.c,key);
+		for (i=0; i<len; ++i) out[i] = scratch.c[i]^inp[i];
+	}
+
+	for (i=15-L;i<16;++i)
+		ctx->nonce.c[i]=0;
+
+	(*block)(ctx->nonce.c,scratch.c,key);
+	ctx->cmac.u[0] ^= scratch.u[0];
+	ctx->cmac.u[1] ^= scratch.u[1];
+
+	ctx->nonce.c[0] = flags0;
+
+	return 0;
+}
+
+int CRYPTO_ccm128_decrypt_ccm64(CCM128_CONTEXT *ctx,
+	const unsigned char *inp, unsigned char *out,
+	size_t len,ccm128_f stream)
+{
+	size_t		n;
+	unsigned int	i,L;
+	unsigned char	flags0	= ctx->nonce.c[0];
+	block128_f	block	= ctx->block;
+	void *		key	= ctx->key;
+	union { u64 u[2]; u8 c[16]; } scratch;
+
+	if (!(flags0&0x40))
+		(*block)(ctx->nonce.c,ctx->cmac.c,key);
+
+	ctx->nonce.c[0] = L = flags0&7;
+	for (n=0,i=15-L;i<15;++i) {
+		n |= ctx->nonce.c[i];
+		ctx->nonce.c[i]=0;
+		n <<= 8;
+	}
+	n |= ctx->nonce.c[15];	/* reconstructed length */
+	ctx->nonce.c[15]=1;
+
+	if (n!=len) return -1;
+
+	if ((n=len/16)) {
+		(*stream)(inp,out,n,key,ctx->nonce.c,ctx->cmac.c);
+		n   *= 16;
+		inp += n;
+		out += n;
+		len -= n;
+		if (len) ctr64_add(ctx->nonce.c,n/16);
+	}
+
+	if (len) {
+		(*block)(ctx->nonce.c,scratch.c,key);
+		for (i=0; i<len; ++i)
+			ctx->cmac.c[i] ^= (out[i] = scratch.c[i]^inp[i]);
+		(*block)(ctx->cmac.c,ctx->cmac.c,key);
+	}
+
+	for (i=15-L;i<16;++i)
+		ctx->nonce.c[i]=0;
+
+	(*block)(ctx->nonce.c,scratch.c,key);
+	ctx->cmac.u[0] ^= scratch.u[0];
+	ctx->cmac.u[1] ^= scratch.u[1];
+
+	ctx->nonce.c[0] = flags0;
+
+	return 0;
+}
+
+size_t CRYPTO_ccm128_tag(CCM128_CONTEXT *ctx,unsigned char *tag,size_t len)
+{	unsigned int M = (ctx->nonce.c[0]>>3)&7;	/* the M parameter */
+
+	M *= 2; M += 2;
+	if (len<M)	return 0;
+	memcpy(tag,ctx->cmac.c,M);
+	return M;
+}
diff --git a/crypto/modes/cfb128.c b/crypto/modes/cfb128.c
new file mode 100644
index 000000000000..4e6f5d35e13b
--- /dev/null
+++ b/crypto/modes/cfb128.c
@@ -0,0 +1,242 @@
+/* ====================================================================
+ * Copyright (c) 2008 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ */
+
+#include <openssl/crypto.h>
+#include "modes_lcl.h"
+#include <string.h>
+
+#ifndef MODES_DEBUG
+# ifndef NDEBUG
+#  define NDEBUG
+# endif
+#endif
+#include <assert.h>
+
+/* The input and output encrypted as though 128bit cfb mode is being
+ * used.  The extra state information to record how much of the
+ * 128bit block we have used is contained in *num;
+ */
+void CRYPTO_cfb128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], int *num,
+			int enc, block128_f block)
+{
+    unsigned int n;
+    size_t l = 0;
+
+    assert(in && out && key && ivec && num);
+
+    n = *num;
+
+    if (enc) {
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+	if (16%sizeof(size_t) == 0) do {	/* always true actually */
+		while (n && len) {
+			*(out++) = ivec[n] ^= *(in++);
+			--len;
+			n = (n+1) % 16;
+		}
+#if defined(STRICT_ALIGNMENT)
+		if (((size_t)in|(size_t)out|(size_t)ivec)%sizeof(size_t) != 0)
+			break;
+#endif
+		while (len>=16) {
+			(*block)(ivec, ivec, key);
+			for (; n<16; n+=sizeof(size_t)) {
+				*(size_t*)(out+n) =
+				*(size_t*)(ivec+n) ^= *(size_t*)(in+n);
+			}
+			len -= 16;
+			out += 16;
+			in  += 16;
+			n = 0;
+		}
+		if (len) {
+			(*block)(ivec, ivec, key);
+			while (len--) {
+				out[n] = ivec[n] ^= in[n];
+				++n;
+			}
+		}
+		*num = n;
+		return;
+	} while (0);
+	/* the rest would be commonly eliminated by x86* compiler */
+#endif
+	while (l<len) {
+		if (n == 0) {
+			(*block)(ivec, ivec, key);
+		}
+		out[l] = ivec[n] ^= in[l];
+		++l;
+		n = (n+1) % 16;
+	}
+	*num = n;
+    } else {
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+	if (16%sizeof(size_t) == 0) do {	/* always true actually */
+		while (n && len) {
+			unsigned char c;
+			*(out++) = ivec[n] ^ (c = *(in++)); ivec[n] = c;
+			--len;
+			n = (n+1) % 16;
+ 		}
+#if defined(STRICT_ALIGNMENT)
+		if (((size_t)in|(size_t)out|(size_t)ivec)%sizeof(size_t) != 0)
+			break;
+#endif
+		while (len>=16) {
+			(*block)(ivec, ivec, key);
+			for (; n<16; n+=sizeof(size_t)) {
+				size_t t = *(size_t*)(in+n);
+				*(size_t*)(out+n) = *(size_t*)(ivec+n) ^ t;
+				*(size_t*)(ivec+n) = t;
+			}
+			len -= 16;
+			out += 16;
+			in  += 16;
+			n = 0;
+		}
+		if (len) {
+			(*block)(ivec, ivec, key);
+			while (len--) {
+				unsigned char c;
+				out[n] = ivec[n] ^ (c = in[n]); ivec[n] = c;
+				++n;
+			}
+ 		}
+		*num = n;
+		return;
+	} while (0);
+	/* the rest would be commonly eliminated by x86* compiler */
+#endif
+	while (l<len) {
+		unsigned char c;
+		if (n == 0) {
+			(*block)(ivec, ivec, key);
+		}
+		out[l] = ivec[n] ^ (c = in[l]); ivec[n] = c;
+		++l;
+		n = (n+1) % 16;
+	}
+	*num=n;
+    }
+}
+
+/* This expects a single block of size nbits for both in and out. Note that
+   it corrupts any extra bits in the last byte of out */
+static void cfbr_encrypt_block(const unsigned char *in,unsigned char *out,
+			    int nbits,const void *key,
+			    unsigned char ivec[16],int enc,
+			    block128_f block)
+{
+    int n,rem,num;
+    unsigned char ovec[16*2 + 1];  /* +1 because we dererefence (but don't use) one byte off the end */
+
+    if (nbits<=0 || nbits>128) return;
+
+	/* fill in the first half of the new IV with the current IV */
+	memcpy(ovec,ivec,16);
+	/* construct the new IV */
+	(*block)(ivec,ivec,key);
+	num = (nbits+7)/8;
+	if (enc)	/* encrypt the input */
+	    for(n=0 ; n < num ; ++n)
+		out[n] = (ovec[16+n] = in[n] ^ ivec[n]);
+	else		/* decrypt the input */
+	    for(n=0 ; n < num ; ++n)
+		out[n] = (ovec[16+n] = in[n]) ^ ivec[n];
+	/* shift ovec left... */
+	rem = nbits%8;
+	num = nbits/8;
+	if(rem==0)
+	    memcpy(ivec,ovec+num,16);
+	else
+	    for(n=0 ; n < 16 ; ++n)
+		ivec[n] = ovec[n+num]<<rem | ovec[n+num+1]>>(8-rem);
+
+    /* it is not necessary to cleanse ovec, since the IV is not secret */
+}
+
+/* N.B. This expects the input to be packed, MS bit first */
+void CRYPTO_cfb128_1_encrypt(const unsigned char *in, unsigned char *out,
+		 	size_t bits, const void *key,
+			unsigned char ivec[16], int *num,
+			int enc, block128_f block)
+{
+    size_t n;
+    unsigned char c[1],d[1];
+
+    assert(in && out && key && ivec && num);
+    assert(*num == 0);
+
+    for(n=0 ; n<bits ; ++n)
+	{
+	c[0]=(in[n/8]&(1 << (7-n%8))) ? 0x80 : 0;
+	cfbr_encrypt_block(c,d,1,key,ivec,enc,block);
+	out[n/8]=(out[n/8]&~(1 << (unsigned int)(7-n%8))) |
+		 ((d[0]&0x80) >> (unsigned int)(n%8));
+	}
+}
+
+void CRYPTO_cfb128_8_encrypt(const unsigned char *in, unsigned char *out,
+			size_t length, const void *key,
+			unsigned char ivec[16], int *num,
+			int enc, block128_f block)
+{
+    size_t n;
+
+    assert(in && out && key && ivec && num);
+    assert(*num == 0);
+
+    for(n=0 ; n<length ; ++n)
+	cfbr_encrypt_block(&in[n],&out[n],8,key,ivec,enc,block);
+}
+
diff --git a/crypto/modes/ctr128.c b/crypto/modes/ctr128.c
new file mode 100644
index 000000000000..ee642c5863cd
--- /dev/null
+++ b/crypto/modes/ctr128.c
@@ -0,0 +1,252 @@
+/* ====================================================================
+ * Copyright (c) 2008 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ */
+
+#include <openssl/crypto.h>
+#include "modes_lcl.h"
+#include <string.h>
+
+#ifndef MODES_DEBUG
+# ifndef NDEBUG
+#  define NDEBUG
+# endif
+#endif
+#include <assert.h>
+
+/* NOTE: the IV/counter CTR mode is big-endian.  The code itself
+ * is endian-neutral. */
+
+/* increment counter (128-bit int) by 1 */
+static void ctr128_inc(unsigned char *counter) {
+	u32 n=16;
+	u8  c;
+
+	do {
+		--n;
+		c = counter[n];
+		++c;
+		counter[n] = c;
+		if (c) return;
+	} while (n);
+}
+
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+static void ctr128_inc_aligned(unsigned char *counter) {
+	size_t *data,c,n;
+	const union { long one; char little; } is_endian = {1};
+
+	if (is_endian.little) {
+		ctr128_inc(counter);
+		return;
+	}
+
+	data = (size_t *)counter;
+	n = 16/sizeof(size_t);
+	do {
+		--n;
+		c = data[n];
+		++c;
+		data[n] = c;
+		if (c) return;
+	} while (n);
+}
+#endif
+
+/* The input encrypted as though 128bit counter mode is being
+ * used.  The extra state information to record how much of the
+ * 128bit block we have used is contained in *num, and the
+ * encrypted counter is kept in ecount_buf.  Both *num and
+ * ecount_buf must be initialised with zeros before the first
+ * call to CRYPTO_ctr128_encrypt().
+ *
+ * This algorithm assumes that the counter is in the x lower bits
+ * of the IV (ivec), and that the application has full control over
+ * overflow and the rest of the IV.  This implementation takes NO
+ * responsability for checking that the counter doesn't overflow
+ * into the rest of the IV when incremented.
+ */
+void CRYPTO_ctr128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], unsigned char ecount_buf[16],
+			unsigned int *num, block128_f block)
+{
+	unsigned int n;
+	size_t l=0;
+
+	assert(in && out && key && ecount_buf && num);
+	assert(*num < 16);
+
+	n = *num;
+
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+	if (16%sizeof(size_t) == 0) do { /* always true actually */
+		while (n && len) {
+			*(out++) = *(in++) ^ ecount_buf[n];
+			--len;
+			n = (n+1) % 16;
+		}
+
+#if defined(STRICT_ALIGNMENT)
+		if (((size_t)in|(size_t)out|(size_t)ivec)%sizeof(size_t) != 0)
+			break;
+#endif
+		while (len>=16) {
+			(*block)(ivec, ecount_buf, key);
+			ctr128_inc_aligned(ivec);
+			for (; n<16; n+=sizeof(size_t))
+				*(size_t *)(out+n) =
+				*(size_t *)(in+n) ^ *(size_t *)(ecount_buf+n);
+			len -= 16;
+			out += 16;
+			in  += 16;
+			n = 0;
+		}
+		if (len) {
+			(*block)(ivec, ecount_buf, key);
+ 			ctr128_inc_aligned(ivec);
+			while (len--) {
+				out[n] = in[n] ^ ecount_buf[n];
+				++n;
+			}
+		}
+		*num = n;
+		return;
+	} while(0);
+	/* the rest would be commonly eliminated by x86* compiler */
+#endif
+	while (l<len) {
+		if (n==0) {
+			(*block)(ivec, ecount_buf, key);
+ 			ctr128_inc(ivec);
+		}
+		out[l] = in[l] ^ ecount_buf[n];
+		++l;
+		n = (n+1) % 16;
+	}
+
+	*num=n;
+}
+
+/* increment upper 96 bits of 128-bit counter by 1 */
+static void ctr96_inc(unsigned char *counter) {
+	u32 n=12;
+	u8  c;
+
+	do {
+		--n;
+		c = counter[n];
+		++c;
+		counter[n] = c;
+		if (c) return;
+	} while (n);
+}
+
+void CRYPTO_ctr128_encrypt_ctr32(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], unsigned char ecount_buf[16],
+			unsigned int *num, ctr128_f func)
+{
+	unsigned int n,ctr32;
+
+	assert(in && out && key && ecount_buf && num);
+	assert(*num < 16);
+
+	n = *num;
+
+	while (n && len) {
+		*(out++) = *(in++) ^ ecount_buf[n];
+		--len;
+		n = (n+1) % 16;
+	}
+
+	ctr32 = GETU32(ivec+12);
+	while (len>=16) {
+		size_t blocks = len/16;
+		/*
+		 * 1<<28 is just a not-so-small yet not-so-large number...
+		 * Below condition is practically never met, but it has to
+		 * be checked for code correctness.
+		 */
+		if (sizeof(size_t)>sizeof(unsigned int) && blocks>(1U<<28))
+			blocks = (1U<<28);
+		/*
+		 * As (*func) operates on 32-bit counter, caller
+		 * has to handle overflow. 'if' below detects the
+		 * overflow, which is then handled by limiting the
+		 * amount of blocks to the exact overflow point...
+		 */
+		ctr32 += (u32)blocks;
+		if (ctr32 < blocks) {
+			blocks -= ctr32;
+			ctr32   = 0;
+		}
+		(*func)(in,out,blocks,key,ivec);
+		/* (*ctr) does not update ivec, caller does: */
+		PUTU32(ivec+12,ctr32);
+		/* ... overflow was detected, propogate carry. */
+		if (ctr32 == 0)	ctr96_inc(ivec);
+		blocks *= 16;
+		len -= blocks;
+		out += blocks;
+		in  += blocks;
+	}
+	if (len) {
+		memset(ecount_buf,0,16);
+		(*func)(ecount_buf,ecount_buf,1,key,ivec);
+		++ctr32;
+		PUTU32(ivec+12,ctr32);
+		if (ctr32 == 0)	ctr96_inc(ivec);
+		while (len--) {
+			out[n] = in[n] ^ ecount_buf[n];
+			++n;
+		}
+	}
+
+	*num=n;
+}
diff --git a/crypto/modes/cts128.c b/crypto/modes/cts128.c
new file mode 100644
index 000000000000..c0e1f3696c6c
--- /dev/null
+++ b/crypto/modes/cts128.c
@@ -0,0 +1,465 @@
+/* ====================================================================
+ * Copyright (c) 2008 The OpenSSL Project. All rights reserved.
+ *
+ * Rights for redistribution and usage in source and binary
+ * forms are granted according to the OpenSSL license.
+ */
+
+#include <openssl/crypto.h>
+#include "modes_lcl.h"
+#include <string.h>
+
+#ifndef MODES_DEBUG
+# ifndef NDEBUG
+#  define NDEBUG
+# endif
+#endif
+#include <assert.h>
+
+/*
+ * Trouble with Ciphertext Stealing, CTS, mode is that there is no
+ * common official specification, but couple of cipher/application
+ * specific ones: RFC2040 and RFC3962. Then there is 'Proposal to
+ * Extend CBC Mode By "Ciphertext Stealing"' at NIST site, which
+ * deviates from mentioned RFCs. Most notably it allows input to be
+ * of block length and it doesn't flip the order of the last two
+ * blocks. CTS is being discussed even in ECB context, but it's not
+ * adopted for any known application. This implementation provides
+ * two interfaces: one compliant with above mentioned RFCs and one
+ * compliant with the NIST proposal, both extending CBC mode.
+ */
+
+size_t CRYPTO_cts128_encrypt_block(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block)
+{	size_t residue, n;
+
+	assert (in && out && key && ivec);
+
+	if (len <= 16) return 0;
+
+	if ((residue=len%16) == 0) residue = 16;
+
+	len -= residue;
+
+	CRYPTO_cbc128_encrypt(in,out,len,key,ivec,block);
+
+	in  += len;
+	out += len;
+
+	for (n=0; n<residue; ++n)
+		ivec[n] ^= in[n];
+	(*block)(ivec,ivec,key);
+	memcpy(out,out-16,residue);
+	memcpy(out-16,ivec,16); 
+
+	return len+residue;
+}
+
+size_t CRYPTO_nistcts128_encrypt_block(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block)
+{	size_t residue, n;
+
+	assert (in && out && key && ivec);
+
+	if (len < 16) return 0;
+
+	residue=len%16;
+
+	len -= residue;
+
+	CRYPTO_cbc128_encrypt(in,out,len,key,ivec,block);
+
+	if (residue==0)	return len;
+
+	in  += len;
+	out += len;
+
+	for (n=0; n<residue; ++n)
+		ivec[n] ^= in[n];
+	(*block)(ivec,ivec,key);
+	memcpy(out-16+residue,ivec,16);
+
+	return len+residue;
+}
+
+size_t CRYPTO_cts128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], cbc128_f cbc)
+{	size_t residue;
+	union { size_t align; unsigned char c[16]; } tmp;
+
+	assert (in && out && key && ivec);
+
+	if (len <= 16) return 0;
+
+	if ((residue=len%16) == 0) residue = 16;
+
+	len -= residue;
+
+	(*cbc)(in,out,len,key,ivec,1);
+
+	in  += len;
+	out += len;
+
+#if defined(CBC_HANDLES_TRUNCATED_IO)
+	memcpy(tmp.c,out-16,16);
+	(*cbc)(in,out-16,residue,key,ivec,1);
+	memcpy(out,tmp.c,residue);
+#else
+	{
+	size_t n;
+	for (n=0; n<16; n+=sizeof(size_t))
+		*(size_t *)(tmp.c+n) = 0;
+	memcpy(tmp.c,in,residue);
+	}
+	memcpy(out,out-16,residue);
+	(*cbc)(tmp.c,out-16,16,key,ivec,1);
+#endif
+	return len+residue;
+}
+
+size_t CRYPTO_nistcts128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], cbc128_f cbc)
+{	size_t residue;
+	union { size_t align; unsigned char c[16]; } tmp;
+
+	assert (in && out && key && ivec);
+
+	if (len < 16) return 0;
+
+	residue=len%16;
+
+	len -= residue;
+
+	(*cbc)(in,out,len,key,ivec,1);
+
+	if (residue==0) return len;
+
+	in  += len;
+	out += len;
+
+#if defined(CBC_HANDLES_TRUNCATED_IO)
+	(*cbc)(in,out-16+residue,residue,key,ivec,1);
+#else
+	{
+	size_t n;
+	for (n=0; n<16; n+=sizeof(size_t))
+		*(size_t *)(tmp.c+n) = 0;
+	memcpy(tmp.c,in,residue);
+	}
+	(*cbc)(tmp.c,out-16+residue,16,key,ivec,1);
+#endif
+	return len+residue;
+}
+
+size_t CRYPTO_cts128_decrypt_block(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block)
+{	size_t residue, n;
+	union { size_t align; unsigned char c[32]; } tmp;
+
+	assert (in && out && key && ivec);
+
+	if (len<=16) return 0;
+
+	if ((residue=len%16) == 0) residue = 16;
+
+	len -= 16+residue;
+
+	if (len) {
+		CRYPTO_cbc128_decrypt(in,out,len,key,ivec,block);
+		in  += len;
+		out += len;
+	}
+
+	(*block)(in,tmp.c+16,key);
+
+	for (n=0; n<16; n+=sizeof(size_t))
+		*(size_t *)(tmp.c+n) = *(size_t *)(tmp.c+16+n);
+	memcpy(tmp.c,in+16,residue);
+	(*block)(tmp.c,tmp.c,key);
+
+	for(n=0; n<16; ++n) {
+		unsigned char c = in[n];
+		out[n] = tmp.c[n] ^ ivec[n];
+		ivec[n] = c;
+	}
+	for(residue+=16; n<residue; ++n)
+		out[n] = tmp.c[n] ^ in[n];
+
+	return 16+len+residue;
+}
+
+size_t CRYPTO_nistcts128_decrypt_block(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block)
+{	size_t residue, n;
+	union { size_t align; unsigned char c[32]; } tmp;
+
+	assert (in && out && key && ivec);
+
+	if (len<16) return 0;
+
+	residue=len%16;
+
+	if (residue==0) {
+		CRYPTO_cbc128_decrypt(in,out,len,key,ivec,block);
+		return len;
+	}
+
+	len -= 16+residue;
+
+	if (len) {
+		CRYPTO_cbc128_decrypt(in,out,len,key,ivec,block);
+		in  += len;
+		out += len;
+	}
+
+	(*block)(in+residue,tmp.c+16,key);
+
+	for (n=0; n<16; n+=sizeof(size_t))
+		*(size_t *)(tmp.c+n) = *(size_t *)(tmp.c+16+n);
+	memcpy(tmp.c,in,residue);
+	(*block)(tmp.c,tmp.c,key);
+
+	for(n=0; n<16; ++n) {
+		unsigned char c = in[n];
+		out[n] = tmp.c[n] ^ ivec[n];
+		ivec[n] = in[n+residue];
+		tmp.c[n] = c;
+	}
+	for(residue+=16; n<residue; ++n)
+		out[n] = tmp.c[n] ^ tmp.c[n-16];
+
+	return 16+len+residue;
+}
+
+size_t CRYPTO_cts128_decrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], cbc128_f cbc)
+{	size_t residue, n;
+	union { size_t align; unsigned char c[32]; } tmp;
+
+	assert (in && out && key && ivec);
+
+	if (len<=16) return 0;
+
+	if ((residue=len%16) == 0) residue = 16;
+
+	len -= 16+residue;
+
+	if (len) {
+		(*cbc)(in,out,len,key,ivec,0);
+		in  += len;
+		out += len;
+	}
+
+	for (n=16; n<32; n+=sizeof(size_t))
+		*(size_t *)(tmp.c+n) = 0;
+	/* this places in[16] at &tmp.c[16] and decrypted block at &tmp.c[0] */
+	(*cbc)(in,tmp.c,16,key,tmp.c+16,0);
+
+	memcpy(tmp.c,in+16,residue);
+#if defined(CBC_HANDLES_TRUNCATED_IO)
+	(*cbc)(tmp.c,out,16+residue,key,ivec,0);
+#else
+	(*cbc)(tmp.c,tmp.c,32,key,ivec,0);
+	memcpy(out,tmp.c,16+residue);
+#endif
+	return 16+len+residue;
+}
+
+size_t CRYPTO_nistcts128_decrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], cbc128_f cbc)
+{	size_t residue, n;
+	union { size_t align; unsigned char c[32]; } tmp;
+
+	assert (in && out && key && ivec);
+
+	if (len<16) return 0;
+
+	residue=len%16;
+
+	if (residue==0) {
+		(*cbc)(in,out,len,key,ivec,0);
+		return len;
+	}
+
+	len -= 16+residue;
+
+	if (len) {
+		(*cbc)(in,out,len,key,ivec,0);
+		in  += len;
+		out += len;
+	}
+
+	for (n=16; n<32; n+=sizeof(size_t))
+		*(size_t *)(tmp.c+n) = 0;
+	/* this places in[16] at &tmp.c[16] and decrypted block at &tmp.c[0] */
+	(*cbc)(in+residue,tmp.c,16,key,tmp.c+16,0);
+
+	memcpy(tmp.c,in,residue);
+#if defined(CBC_HANDLES_TRUNCATED_IO)
+	(*cbc)(tmp.c,out,16+residue,key,ivec,0);
+#else
+	(*cbc)(tmp.c,tmp.c,32,key,ivec,0);
+	memcpy(out,tmp.c,16+residue);
+#endif
+	return 16+len+residue;
+}
+
+#if defined(SELFTEST)
+#include <stdio.h>
+#include <openssl/aes.h>
+
+/* test vectors from RFC 3962 */
+static const unsigned char test_key[16] = "chicken teriyaki";
+static const unsigned char test_input[64] =
+		"I would like the" " General Gau's C"
+		"hicken, please, " "and wonton soup.";
+static const unsigned char test_iv[16] = {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0};
+
+static const unsigned char vector_17[17] =
+{0xc6,0x35,0x35,0x68,0xf2,0xbf,0x8c,0xb4, 0xd8,0xa5,0x80,0x36,0x2d,0xa7,0xff,0x7f,
+ 0x97};
+static const unsigned char vector_31[31] =
+{0xfc,0x00,0x78,0x3e,0x0e,0xfd,0xb2,0xc1, 0xd4,0x45,0xd4,0xc8,0xef,0xf7,0xed,0x22,
+ 0x97,0x68,0x72,0x68,0xd6,0xec,0xcc,0xc0, 0xc0,0x7b,0x25,0xe2,0x5e,0xcf,0xe5};
+static const unsigned char vector_32[32] =
+{0x39,0x31,0x25,0x23,0xa7,0x86,0x62,0xd5, 0xbe,0x7f,0xcb,0xcc,0x98,0xeb,0xf5,0xa8,
+ 0x97,0x68,0x72,0x68,0xd6,0xec,0xcc,0xc0, 0xc0,0x7b,0x25,0xe2,0x5e,0xcf,0xe5,0x84};
+static const unsigned char vector_47[47] =
+{0x97,0x68,0x72,0x68,0xd6,0xec,0xcc,0xc0, 0xc0,0x7b,0x25,0xe2,0x5e,0xcf,0xe5,0x84,
+ 0xb3,0xff,0xfd,0x94,0x0c,0x16,0xa1,0x8c, 0x1b,0x55,0x49,0xd2,0xf8,0x38,0x02,0x9e,
+ 0x39,0x31,0x25,0x23,0xa7,0x86,0x62,0xd5, 0xbe,0x7f,0xcb,0xcc,0x98,0xeb,0xf5};
+static const unsigned char vector_48[48] =
+{0x97,0x68,0x72,0x68,0xd6,0xec,0xcc,0xc0, 0xc0,0x7b,0x25,0xe2,0x5e,0xcf,0xe5,0x84,
+ 0x9d,0xad,0x8b,0xbb,0x96,0xc4,0xcd,0xc0, 0x3b,0xc1,0x03,0xe1,0xa1,0x94,0xbb,0xd8,
+ 0x39,0x31,0x25,0x23,0xa7,0x86,0x62,0xd5, 0xbe,0x7f,0xcb,0xcc,0x98,0xeb,0xf5,0xa8};
+static const unsigned char vector_64[64] =
+{0x97,0x68,0x72,0x68,0xd6,0xec,0xcc,0xc0, 0xc0,0x7b,0x25,0xe2,0x5e,0xcf,0xe5,0x84,
+ 0x39,0x31,0x25,0x23,0xa7,0x86,0x62,0xd5, 0xbe,0x7f,0xcb,0xcc,0x98,0xeb,0xf5,0xa8,
+ 0x48,0x07,0xef,0xe8,0x36,0xee,0x89,0xa5, 0x26,0x73,0x0d,0xbc,0x2f,0x7b,0xc8,0x40,
+ 0x9d,0xad,0x8b,0xbb,0x96,0xc4,0xcd,0xc0, 0x3b,0xc1,0x03,0xe1,0xa1,0x94,0xbb,0xd8};
+
+static AES_KEY encks, decks;
+
+void test_vector(const unsigned char *vector,size_t len)
+{	unsigned char iv[sizeof(test_iv)];
+	unsigned char cleartext[64],ciphertext[64];
+	size_t tail;
+
+	printf("vector_%d\n",len); fflush(stdout);
+
+	if ((tail=len%16) == 0) tail = 16;
+	tail += 16;
+
+	/* test block-based encryption */
+	memcpy(iv,test_iv,sizeof(test_iv));
+	CRYPTO_cts128_encrypt_block(test_input,ciphertext,len,&encks,iv,(block128_f)AES_encrypt);
+	if (memcmp(ciphertext,vector,len))
+		fprintf(stderr,"output_%d mismatch\n",len), exit(1);
+	if (memcmp(iv,vector+len-tail,sizeof(iv)))
+		fprintf(stderr,"iv_%d mismatch\n",len), exit(1);
+
+	/* test block-based decryption */
+	memcpy(iv,test_iv,sizeof(test_iv));
+	CRYPTO_cts128_decrypt_block(ciphertext,cleartext,len,&decks,iv,(block128_f)AES_decrypt);
+	if (memcmp(cleartext,test_input,len))
+		fprintf(stderr,"input_%d mismatch\n",len), exit(2);
+	if (memcmp(iv,vector+len-tail,sizeof(iv)))
+		fprintf(stderr,"iv_%d mismatch\n",len), exit(2);
+
+	/* test streamed encryption */
+	memcpy(iv,test_iv,sizeof(test_iv));
+	CRYPTO_cts128_encrypt(test_input,ciphertext,len,&encks,iv,(cbc128_f)AES_cbc_encrypt);
+	if (memcmp(ciphertext,vector,len))
+		fprintf(stderr,"output_%d mismatch\n",len), exit(3);
+	if (memcmp(iv,vector+len-tail,sizeof(iv)))
+		fprintf(stderr,"iv_%d mismatch\n",len), exit(3);
+
+	/* test streamed decryption */
+	memcpy(iv,test_iv,sizeof(test_iv));
+	CRYPTO_cts128_decrypt(ciphertext,cleartext,len,&decks,iv,(cbc128_f)AES_cbc_encrypt);
+	if (memcmp(cleartext,test_input,len))
+		fprintf(stderr,"input_%d mismatch\n",len), exit(4);
+	if (memcmp(iv,vector+len-tail,sizeof(iv)))
+		fprintf(stderr,"iv_%d mismatch\n",len), exit(4);
+}
+
+void test_nistvector(const unsigned char *vector,size_t len)
+{	unsigned char iv[sizeof(test_iv)];
+	unsigned char cleartext[64],ciphertext[64],nistvector[64];
+	size_t tail;
+
+	printf("nistvector_%d\n",len); fflush(stdout);
+
+	if ((tail=len%16) == 0) tail = 16;
+
+	len -= 16 + tail;
+	memcpy(nistvector,vector,len);
+	/* flip two last blocks */
+	memcpy(nistvector+len,vector+len+16,tail);
+	memcpy(nistvector+len+tail,vector+len,16);
+	len += 16 + tail;
+	tail = 16;
+
+	/* test block-based encryption */
+	memcpy(iv,test_iv,sizeof(test_iv));
+	CRYPTO_nistcts128_encrypt_block(test_input,ciphertext,len,&encks,iv,(block128_f)AES_encrypt);
+	if (memcmp(ciphertext,nistvector,len))
+		fprintf(stderr,"output_%d mismatch\n",len), exit(1);
+	if (memcmp(iv,nistvector+len-tail,sizeof(iv)))
+		fprintf(stderr,"iv_%d mismatch\n",len), exit(1);
+
+	/* test block-based decryption */
+	memcpy(iv,test_iv,sizeof(test_iv));
+	CRYPTO_nistcts128_decrypt_block(ciphertext,cleartext,len,&decks,iv,(block128_f)AES_decrypt);
+	if (memcmp(cleartext,test_input,len))
+		fprintf(stderr,"input_%d mismatch\n",len), exit(2);
+	if (memcmp(iv,nistvector+len-tail,sizeof(iv)))
+		fprintf(stderr,"iv_%d mismatch\n",len), exit(2);
+
+	/* test streamed encryption */
+	memcpy(iv,test_iv,sizeof(test_iv));
+	CRYPTO_nistcts128_encrypt(test_input,ciphertext,len,&encks,iv,(cbc128_f)AES_cbc_encrypt);
+	if (memcmp(ciphertext,nistvector,len))
+		fprintf(stderr,"output_%d mismatch\n",len), exit(3);
+	if (memcmp(iv,nistvector+len-tail,sizeof(iv)))
+		fprintf(stderr,"iv_%d mismatch\n",len), exit(3);
+
+	/* test streamed decryption */
+	memcpy(iv,test_iv,sizeof(test_iv));
+	CRYPTO_nistcts128_decrypt(ciphertext,cleartext,len,&decks,iv,(cbc128_f)AES_cbc_encrypt);
+	if (memcmp(cleartext,test_input,len))
+		fprintf(stderr,"input_%d mismatch\n",len), exit(4);
+	if (memcmp(iv,nistvector+len-tail,sizeof(iv)))
+		fprintf(stderr,"iv_%d mismatch\n",len), exit(4);
+}
+
+int main()
+{
+	AES_set_encrypt_key(test_key,128,&encks);
+	AES_set_decrypt_key(test_key,128,&decks);
+
+	test_vector(vector_17,sizeof(vector_17));
+	test_vector(vector_31,sizeof(vector_31));
+	test_vector(vector_32,sizeof(vector_32));
+	test_vector(vector_47,sizeof(vector_47));
+	test_vector(vector_48,sizeof(vector_48));
+	test_vector(vector_64,sizeof(vector_64));
+
+	test_nistvector(vector_17,sizeof(vector_17));
+	test_nistvector(vector_31,sizeof(vector_31));
+	test_nistvector(vector_32,sizeof(vector_32));
+	test_nistvector(vector_47,sizeof(vector_47));
+	test_nistvector(vector_48,sizeof(vector_48));
+	test_nistvector(vector_64,sizeof(vector_64));
+
+	return 0;
+}
+#endif
diff --git a/crypto/modes/gcm128.c b/crypto/modes/gcm128.c
new file mode 100644
index 000000000000..7d6d03497027
--- /dev/null
+++ b/crypto/modes/gcm128.c
@@ -0,0 +1,1757 @@
+/* ====================================================================
+ * Copyright (c) 2010 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#define OPENSSL_FIPSAPI
+
+#include <openssl/crypto.h>
+#include "modes_lcl.h"
+#include <string.h>
+
+#ifndef MODES_DEBUG
+# ifndef NDEBUG
+#  define NDEBUG
+# endif
+#endif
+#include <assert.h>
+
+#if defined(BSWAP4) && defined(STRICT_ALIGNMENT)
+/* redefine, because alignment is ensured */
+#undef	GETU32
+#define	GETU32(p)	BSWAP4(*(const u32 *)(p))
+#undef	PUTU32
+#define	PUTU32(p,v)	*(u32 *)(p) = BSWAP4(v)
+#endif
+
+#define	PACK(s)		((size_t)(s)<<(sizeof(size_t)*8-16))
+#define REDUCE1BIT(V)	do { \
+	if (sizeof(size_t)==8) { \
+		u64 T = U64(0xe100000000000000) & (0-(V.lo&1)); \
+		V.lo  = (V.hi<<63)|(V.lo>>1); \
+		V.hi  = (V.hi>>1 )^T; \
+	} \
+	else { \
+		u32 T = 0xe1000000U & (0-(u32)(V.lo&1)); \
+		V.lo  = (V.hi<<63)|(V.lo>>1); \
+		V.hi  = (V.hi>>1 )^((u64)T<<32); \
+	} \
+} while(0)
+
+/*
+ * Even though permitted values for TABLE_BITS are 8, 4 and 1, it should
+ * never be set to 8. 8 is effectively reserved for testing purposes.
+ * TABLE_BITS>1 are lookup-table-driven implementations referred to as
+ * "Shoup's" in GCM specification. In other words OpenSSL does not cover
+ * whole spectrum of possible table driven implementations. Why? In
+ * non-"Shoup's" case memory access pattern is segmented in such manner,
+ * that it's trivial to see that cache timing information can reveal
+ * fair portion of intermediate hash value. Given that ciphertext is
+ * always available to attacker, it's possible for him to attempt to
+ * deduce secret parameter H and if successful, tamper with messages
+ * [which is nothing but trivial in CTR mode]. In "Shoup's" case it's
+ * not as trivial, but there is no reason to believe that it's resistant
+ * to cache-timing attack. And the thing about "8-bit" implementation is
+ * that it consumes 16 (sixteen) times more memory, 4KB per individual
+ * key + 1KB shared. Well, on pros side it should be twice as fast as
+ * "4-bit" version. And for gcc-generated x86[_64] code, "8-bit" version
+ * was observed to run ~75% faster, closer to 100% for commercial
+ * compilers... Yet "4-bit" procedure is preferred, because it's
+ * believed to provide better security-performance balance and adequate
+ * all-round performance. "All-round" refers to things like:
+ *
+ * - shorter setup time effectively improves overall timing for
+ *   handling short messages;
+ * - larger table allocation can become unbearable because of VM
+ *   subsystem penalties (for example on Windows large enough free
+ *   results in VM working set trimming, meaning that consequent
+ *   malloc would immediately incur working set expansion);
+ * - larger table has larger cache footprint, which can affect
+ *   performance of other code paths (not necessarily even from same
+ *   thread in Hyper-Threading world);
+ *
+ * Value of 1 is not appropriate for performance reasons.
+ */
+#if	TABLE_BITS==8
+
+static void gcm_init_8bit(u128 Htable[256], u64 H[2])
+{
+	int  i, j;
+	u128 V;
+
+	Htable[0].hi = 0;
+	Htable[0].lo = 0;
+	V.hi = H[0];
+	V.lo = H[1];
+
+	for (Htable[128]=V, i=64; i>0; i>>=1) {
+		REDUCE1BIT(V);
+		Htable[i] = V;
+	}
+
+	for (i=2; i<256; i<<=1) {
+		u128 *Hi = Htable+i, H0 = *Hi;
+		for (j=1; j<i; ++j) {
+			Hi[j].hi = H0.hi^Htable[j].hi;
+			Hi[j].lo = H0.lo^Htable[j].lo;
+		}
+	}
+}
+
+static void gcm_gmult_8bit(u64 Xi[2], const u128 Htable[256])
+{
+	u128 Z = { 0, 0};
+	const u8 *xi = (const u8 *)Xi+15;
+	size_t rem, n = *xi;
+	const union { long one; char little; } is_endian = {1};
+	static const size_t rem_8bit[256] = {
+		PACK(0x0000), PACK(0x01C2), PACK(0x0384), PACK(0x0246),
+		PACK(0x0708), PACK(0x06CA), PACK(0x048C), PACK(0x054E),
+		PACK(0x0E10), PACK(0x0FD2), PACK(0x0D94), PACK(0x0C56),
+		PACK(0x0918), PACK(0x08DA), PACK(0x0A9C), PACK(0x0B5E),
+		PACK(0x1C20), PACK(0x1DE2), PACK(0x1FA4), PACK(0x1E66),
+		PACK(0x1B28), PACK(0x1AEA), PACK(0x18AC), PACK(0x196E),
+		PACK(0x1230), PACK(0x13F2), PACK(0x11B4), PACK(0x1076),
+		PACK(0x1538), PACK(0x14FA), PACK(0x16BC), PACK(0x177E),
+		PACK(0x3840), PACK(0x3982), PACK(0x3BC4), PACK(0x3A06),
+		PACK(0x3F48), PACK(0x3E8A), PACK(0x3CCC), PACK(0x3D0E),
+		PACK(0x3650), PACK(0x3792), PACK(0x35D4), PACK(0x3416),
+		PACK(0x3158), PACK(0x309A), PACK(0x32DC), PACK(0x331E),
+		PACK(0x2460), PACK(0x25A2), PACK(0x27E4), PACK(0x2626),
+		PACK(0x2368), PACK(0x22AA), PACK(0x20EC), PACK(0x212E),
+		PACK(0x2A70), PACK(0x2BB2), PACK(0x29F4), PACK(0x2836),
+		PACK(0x2D78), PACK(0x2CBA), PACK(0x2EFC), PACK(0x2F3E),
+		PACK(0x7080), PACK(0x7142), PACK(0x7304), PACK(0x72C6),
+		PACK(0x7788), PACK(0x764A), PACK(0x740C), PACK(0x75CE),
+		PACK(0x7E90), PACK(0x7F52), PACK(0x7D14), PACK(0x7CD6),
+		PACK(0x7998), PACK(0x785A), PACK(0x7A1C), PACK(0x7BDE),
+		PACK(0x6CA0), PACK(0x6D62), PACK(0x6F24), PACK(0x6EE6),
+		PACK(0x6BA8), PACK(0x6A6A), PACK(0x682C), PACK(0x69EE),
+		PACK(0x62B0), PACK(0x6372), PACK(0x6134), PACK(0x60F6),
+		PACK(0x65B8), PACK(0x647A), PACK(0x663C), PACK(0x67FE),
+		PACK(0x48C0), PACK(0x4902), PACK(0x4B44), PACK(0x4A86),
+		PACK(0x4FC8), PACK(0x4E0A), PACK(0x4C4C), PACK(0x4D8E),
+		PACK(0x46D0), PACK(0x4712), PACK(0x4554), PACK(0x4496),
+		PACK(0x41D8), PACK(0x401A), PACK(0x425C), PACK(0x439E),
+		PACK(0x54E0), PACK(0x5522), PACK(0x5764), PACK(0x56A6),
+		PACK(0x53E8), PACK(0x522A), PACK(0x506C), PACK(0x51AE),
+		PACK(0x5AF0), PACK(0x5B32), PACK(0x5974), PACK(0x58B6),
+		PACK(0x5DF8), PACK(0x5C3A), PACK(0x5E7C), PACK(0x5FBE),
+		PACK(0xE100), PACK(0xE0C2), PACK(0xE284), PACK(0xE346),
+		PACK(0xE608), PACK(0xE7CA), PACK(0xE58C), PACK(0xE44E),
+		PACK(0xEF10), PACK(0xEED2), PACK(0xEC94), PACK(0xED56),
+		PACK(0xE818), PACK(0xE9DA), PACK(0xEB9C), PACK(0xEA5E),
+		PACK(0xFD20), PACK(0xFCE2), PACK(0xFEA4), PACK(0xFF66),
+		PACK(0xFA28), PACK(0xFBEA), PACK(0xF9AC), PACK(0xF86E),
+		PACK(0xF330), PACK(0xF2F2), PACK(0xF0B4), PACK(0xF176),
+		PACK(0xF438), PACK(0xF5FA), PACK(0xF7BC), PACK(0xF67E),
+		PACK(0xD940), PACK(0xD882), PACK(0xDAC4), PACK(0xDB06),
+		PACK(0xDE48), PACK(0xDF8A), PACK(0xDDCC), PACK(0xDC0E),
+		PACK(0xD750), PACK(0xD692), PACK(0xD4D4), PACK(0xD516),
+		PACK(0xD058), PACK(0xD19A), PACK(0xD3DC), PACK(0xD21E),
+		PACK(0xC560), PACK(0xC4A2), PACK(0xC6E4), PACK(0xC726),
+		PACK(0xC268), PACK(0xC3AA), PACK(0xC1EC), PACK(0xC02E),
+		PACK(0xCB70), PACK(0xCAB2), PACK(0xC8F4), PACK(0xC936),
+		PACK(0xCC78), PACK(0xCDBA), PACK(0xCFFC), PACK(0xCE3E),
+		PACK(0x9180), PACK(0x9042), PACK(0x9204), PACK(0x93C6),
+		PACK(0x9688), PACK(0x974A), PACK(0x950C), PACK(0x94CE),
+		PACK(0x9F90), PACK(0x9E52), PACK(0x9C14), PACK(0x9DD6),
+		PACK(0x9898), PACK(0x995A), PACK(0x9B1C), PACK(0x9ADE),
+		PACK(0x8DA0), PACK(0x8C62), PACK(0x8E24), PACK(0x8FE6),
+		PACK(0x8AA8), PACK(0x8B6A), PACK(0x892C), PACK(0x88EE),
+		PACK(0x83B0), PACK(0x8272), PACK(0x8034), PACK(0x81F6),
+		PACK(0x84B8), PACK(0x857A), PACK(0x873C), PACK(0x86FE),
+		PACK(0xA9C0), PACK(0xA802), PACK(0xAA44), PACK(0xAB86),
+		PACK(0xAEC8), PACK(0xAF0A), PACK(0xAD4C), PACK(0xAC8E),
+		PACK(0xA7D0), PACK(0xA612), PACK(0xA454), PACK(0xA596),
+		PACK(0xA0D8), PACK(0xA11A), PACK(0xA35C), PACK(0xA29E),
+		PACK(0xB5E0), PACK(0xB422), PACK(0xB664), PACK(0xB7A6),
+		PACK(0xB2E8), PACK(0xB32A), PACK(0xB16C), PACK(0xB0AE),
+		PACK(0xBBF0), PACK(0xBA32), PACK(0xB874), PACK(0xB9B6),
+		PACK(0xBCF8), PACK(0xBD3A), PACK(0xBF7C), PACK(0xBEBE) };
+
+	while (1) {
+		Z.hi ^= Htable[n].hi;
+		Z.lo ^= Htable[n].lo;
+
+		if ((u8 *)Xi==xi)	break;
+
+		n = *(--xi);
+
+		rem  = (size_t)Z.lo&0xff;
+		Z.lo = (Z.hi<<56)|(Z.lo>>8);
+		Z.hi = (Z.hi>>8);
+		if (sizeof(size_t)==8)
+			Z.hi ^= rem_8bit[rem];
+		else
+			Z.hi ^= (u64)rem_8bit[rem]<<32;
+	}
+
+	if (is_endian.little) {
+#ifdef BSWAP8
+		Xi[0] = BSWAP8(Z.hi);
+		Xi[1] = BSWAP8(Z.lo);
+#else
+		u8 *p = (u8 *)Xi;
+		u32 v;
+		v = (u32)(Z.hi>>32);	PUTU32(p,v);
+		v = (u32)(Z.hi);	PUTU32(p+4,v);
+		v = (u32)(Z.lo>>32);	PUTU32(p+8,v);
+		v = (u32)(Z.lo);	PUTU32(p+12,v);
+#endif
+	}
+	else {
+		Xi[0] = Z.hi;
+		Xi[1] = Z.lo;
+	}
+}
+#define GCM_MUL(ctx,Xi)   gcm_gmult_8bit(ctx->Xi.u,ctx->Htable)
+
+#elif	TABLE_BITS==4
+
+static void gcm_init_4bit(u128 Htable[16], u64 H[2])
+{
+	u128 V;
+#if defined(OPENSSL_SMALL_FOOTPRINT)
+	int  i;
+#endif
+
+	Htable[0].hi = 0;
+	Htable[0].lo = 0;
+	V.hi = H[0];
+	V.lo = H[1];
+
+#if defined(OPENSSL_SMALL_FOOTPRINT)
+	for (Htable[8]=V, i=4; i>0; i>>=1) {
+		REDUCE1BIT(V);
+		Htable[i] = V;
+	}
+
+	for (i=2; i<16; i<<=1) {
+		u128 *Hi = Htable+i;
+		int   j;
+		for (V=*Hi, j=1; j<i; ++j) {
+			Hi[j].hi = V.hi^Htable[j].hi;
+			Hi[j].lo = V.lo^Htable[j].lo;
+		}
+	}
+#else
+	Htable[8] = V;
+	REDUCE1BIT(V);
+	Htable[4] = V;
+	REDUCE1BIT(V);
+	Htable[2] = V;
+	REDUCE1BIT(V);
+	Htable[1] = V;
+	Htable[3].hi  = V.hi^Htable[2].hi, Htable[3].lo  = V.lo^Htable[2].lo;
+	V=Htable[4];
+	Htable[5].hi  = V.hi^Htable[1].hi, Htable[5].lo  = V.lo^Htable[1].lo;
+	Htable[6].hi  = V.hi^Htable[2].hi, Htable[6].lo  = V.lo^Htable[2].lo;
+	Htable[7].hi  = V.hi^Htable[3].hi, Htable[7].lo  = V.lo^Htable[3].lo;
+	V=Htable[8];
+	Htable[9].hi  = V.hi^Htable[1].hi, Htable[9].lo  = V.lo^Htable[1].lo;
+	Htable[10].hi = V.hi^Htable[2].hi, Htable[10].lo = V.lo^Htable[2].lo;
+	Htable[11].hi = V.hi^Htable[3].hi, Htable[11].lo = V.lo^Htable[3].lo;
+	Htable[12].hi = V.hi^Htable[4].hi, Htable[12].lo = V.lo^Htable[4].lo;
+	Htable[13].hi = V.hi^Htable[5].hi, Htable[13].lo = V.lo^Htable[5].lo;
+	Htable[14].hi = V.hi^Htable[6].hi, Htable[14].lo = V.lo^Htable[6].lo;
+	Htable[15].hi = V.hi^Htable[7].hi, Htable[15].lo = V.lo^Htable[7].lo;
+#endif
+#if defined(GHASH_ASM) && (defined(__arm__) || defined(__arm))
+	/*
+	 * ARM assembler expects specific dword order in Htable.
+	 */
+	{
+	int j;
+	const union { long one; char little; } is_endian = {1};
+
+	if (is_endian.little)
+		for (j=0;j<16;++j) {
+			V = Htable[j];
+			Htable[j].hi = V.lo;
+			Htable[j].lo = V.hi;
+		}
+	else
+		for (j=0;j<16;++j) {
+			V = Htable[j];
+			Htable[j].hi = V.lo<<32|V.lo>>32;
+			Htable[j].lo = V.hi<<32|V.hi>>32;
+		}
+	}
+#endif
+}
+
+#ifndef GHASH_ASM
+static const size_t rem_4bit[16] = {
+	PACK(0x0000), PACK(0x1C20), PACK(0x3840), PACK(0x2460),
+	PACK(0x7080), PACK(0x6CA0), PACK(0x48C0), PACK(0x54E0),
+	PACK(0xE100), PACK(0xFD20), PACK(0xD940), PACK(0xC560),
+	PACK(0x9180), PACK(0x8DA0), PACK(0xA9C0), PACK(0xB5E0) };
+
+static void gcm_gmult_4bit(u64 Xi[2], const u128 Htable[16])
+{
+	u128 Z;
+	int cnt = 15;
+	size_t rem, nlo, nhi;
+	const union { long one; char little; } is_endian = {1};
+
+	nlo  = ((const u8 *)Xi)[15];
+	nhi  = nlo>>4;
+	nlo &= 0xf;
+
+	Z.hi = Htable[nlo].hi;
+	Z.lo = Htable[nlo].lo;
+
+	while (1) {
+		rem  = (size_t)Z.lo&0xf;
+		Z.lo = (Z.hi<<60)|(Z.lo>>4);
+		Z.hi = (Z.hi>>4);
+		if (sizeof(size_t)==8)
+			Z.hi ^= rem_4bit[rem];
+		else
+			Z.hi ^= (u64)rem_4bit[rem]<<32;
+
+		Z.hi ^= Htable[nhi].hi;
+		Z.lo ^= Htable[nhi].lo;
+
+		if (--cnt<0)		break;
+
+		nlo  = ((const u8 *)Xi)[cnt];
+		nhi  = nlo>>4;
+		nlo &= 0xf;
+
+		rem  = (size_t)Z.lo&0xf;
+		Z.lo = (Z.hi<<60)|(Z.lo>>4);
+		Z.hi = (Z.hi>>4);
+		if (sizeof(size_t)==8)
+			Z.hi ^= rem_4bit[rem];
+		else
+			Z.hi ^= (u64)rem_4bit[rem]<<32;
+
+		Z.hi ^= Htable[nlo].hi;
+		Z.lo ^= Htable[nlo].lo;
+	}
+
+	if (is_endian.little) {
+#ifdef BSWAP8
+		Xi[0] = BSWAP8(Z.hi);
+		Xi[1] = BSWAP8(Z.lo);
+#else
+		u8 *p = (u8 *)Xi;
+		u32 v;
+		v = (u32)(Z.hi>>32);	PUTU32(p,v);
+		v = (u32)(Z.hi);	PUTU32(p+4,v);
+		v = (u32)(Z.lo>>32);	PUTU32(p+8,v);
+		v = (u32)(Z.lo);	PUTU32(p+12,v);
+#endif
+	}
+	else {
+		Xi[0] = Z.hi;
+		Xi[1] = Z.lo;
+	}
+}
+
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+/*
+ * Streamed gcm_mult_4bit, see CRYPTO_gcm128_[en|de]crypt for
+ * details... Compiler-generated code doesn't seem to give any
+ * performance improvement, at least not on x86[_64]. It's here
+ * mostly as reference and a placeholder for possible future
+ * non-trivial optimization[s]...
+ */
+static void gcm_ghash_4bit(u64 Xi[2],const u128 Htable[16],
+				const u8 *inp,size_t len)
+{
+    u128 Z;
+    int cnt;
+    size_t rem, nlo, nhi;
+    const union { long one; char little; } is_endian = {1};
+
+#if 1
+    do {
+	cnt  = 15;
+	nlo  = ((const u8 *)Xi)[15];
+	nlo ^= inp[15];
+	nhi  = nlo>>4;
+	nlo &= 0xf;
+
+	Z.hi = Htable[nlo].hi;
+	Z.lo = Htable[nlo].lo;
+
+	while (1) {
+		rem  = (size_t)Z.lo&0xf;
+		Z.lo = (Z.hi<<60)|(Z.lo>>4);
+		Z.hi = (Z.hi>>4);
+		if (sizeof(size_t)==8)
+			Z.hi ^= rem_4bit[rem];
+		else
+			Z.hi ^= (u64)rem_4bit[rem]<<32;
+
+		Z.hi ^= Htable[nhi].hi;
+		Z.lo ^= Htable[nhi].lo;
+
+		if (--cnt<0)		break;
+
+		nlo  = ((const u8 *)Xi)[cnt];
+		nlo ^= inp[cnt];
+		nhi  = nlo>>4;
+		nlo &= 0xf;
+
+		rem  = (size_t)Z.lo&0xf;
+		Z.lo = (Z.hi<<60)|(Z.lo>>4);
+		Z.hi = (Z.hi>>4);
+		if (sizeof(size_t)==8)
+			Z.hi ^= rem_4bit[rem];
+		else
+			Z.hi ^= (u64)rem_4bit[rem]<<32;
+
+		Z.hi ^= Htable[nlo].hi;
+		Z.lo ^= Htable[nlo].lo;
+	}
+#else
+    /*
+     * Extra 256+16 bytes per-key plus 512 bytes shared tables
+     * [should] give ~50% improvement... One could have PACK()-ed
+     * the rem_8bit even here, but the priority is to minimize
+     * cache footprint...
+     */ 
+    u128 Hshr4[16];	/* Htable shifted right by 4 bits */
+    u8   Hshl4[16];	/* Htable shifted left  by 4 bits */
+    static const unsigned short rem_8bit[256] = {
+	0x0000, 0x01C2, 0x0384, 0x0246, 0x0708, 0x06CA, 0x048C, 0x054E,
+	0x0E10, 0x0FD2, 0x0D94, 0x0C56, 0x0918, 0x08DA, 0x0A9C, 0x0B5E,
+	0x1C20, 0x1DE2, 0x1FA4, 0x1E66, 0x1B28, 0x1AEA, 0x18AC, 0x196E,
+	0x1230, 0x13F2, 0x11B4, 0x1076, 0x1538, 0x14FA, 0x16BC, 0x177E,
+	0x3840, 0x3982, 0x3BC4, 0x3A06, 0x3F48, 0x3E8A, 0x3CCC, 0x3D0E,
+	0x3650, 0x3792, 0x35D4, 0x3416, 0x3158, 0x309A, 0x32DC, 0x331E,
+	0x2460, 0x25A2, 0x27E4, 0x2626, 0x2368, 0x22AA, 0x20EC, 0x212E,
+	0x2A70, 0x2BB2, 0x29F4, 0x2836, 0x2D78, 0x2CBA, 0x2EFC, 0x2F3E,
+	0x7080, 0x7142, 0x7304, 0x72C6, 0x7788, 0x764A, 0x740C, 0x75CE,
+	0x7E90, 0x7F52, 0x7D14, 0x7CD6, 0x7998, 0x785A, 0x7A1C, 0x7BDE,
+	0x6CA0, 0x6D62, 0x6F24, 0x6EE6, 0x6BA8, 0x6A6A, 0x682C, 0x69EE,
+	0x62B0, 0x6372, 0x6134, 0x60F6, 0x65B8, 0x647A, 0x663C, 0x67FE,
+	0x48C0, 0x4902, 0x4B44, 0x4A86, 0x4FC8, 0x4E0A, 0x4C4C, 0x4D8E,
+	0x46D0, 0x4712, 0x4554, 0x4496, 0x41D8, 0x401A, 0x425C, 0x439E,
+	0x54E0, 0x5522, 0x5764, 0x56A6, 0x53E8, 0x522A, 0x506C, 0x51AE,
+	0x5AF0, 0x5B32, 0x5974, 0x58B6, 0x5DF8, 0x5C3A, 0x5E7C, 0x5FBE,
+	0xE100, 0xE0C2, 0xE284, 0xE346, 0xE608, 0xE7CA, 0xE58C, 0xE44E,
+	0xEF10, 0xEED2, 0xEC94, 0xED56, 0xE818, 0xE9DA, 0xEB9C, 0xEA5E,
+	0xFD20, 0xFCE2, 0xFEA4, 0xFF66, 0xFA28, 0xFBEA, 0xF9AC, 0xF86E,
+	0xF330, 0xF2F2, 0xF0B4, 0xF176, 0xF438, 0xF5FA, 0xF7BC, 0xF67E,
+	0xD940, 0xD882, 0xDAC4, 0xDB06, 0xDE48, 0xDF8A, 0xDDCC, 0xDC0E,
+	0xD750, 0xD692, 0xD4D4, 0xD516, 0xD058, 0xD19A, 0xD3DC, 0xD21E,
+	0xC560, 0xC4A2, 0xC6E4, 0xC726, 0xC268, 0xC3AA, 0xC1EC, 0xC02E,
+	0xCB70, 0xCAB2, 0xC8F4, 0xC936, 0xCC78, 0xCDBA, 0xCFFC, 0xCE3E,
+	0x9180, 0x9042, 0x9204, 0x93C6, 0x9688, 0x974A, 0x950C, 0x94CE,
+	0x9F90, 0x9E52, 0x9C14, 0x9DD6, 0x9898, 0x995A, 0x9B1C, 0x9ADE,
+	0x8DA0, 0x8C62, 0x8E24, 0x8FE6, 0x8AA8, 0x8B6A, 0x892C, 0x88EE,
+	0x83B0, 0x8272, 0x8034, 0x81F6, 0x84B8, 0x857A, 0x873C, 0x86FE,
+	0xA9C0, 0xA802, 0xAA44, 0xAB86, 0xAEC8, 0xAF0A, 0xAD4C, 0xAC8E,
+	0xA7D0, 0xA612, 0xA454, 0xA596, 0xA0D8, 0xA11A, 0xA35C, 0xA29E,
+	0xB5E0, 0xB422, 0xB664, 0xB7A6, 0xB2E8, 0xB32A, 0xB16C, 0xB0AE,
+	0xBBF0, 0xBA32, 0xB874, 0xB9B6, 0xBCF8, 0xBD3A, 0xBF7C, 0xBEBE };
+    /*
+     * This pre-processing phase slows down procedure by approximately
+     * same time as it makes each loop spin faster. In other words
+     * single block performance is approximately same as straightforward
+     * "4-bit" implementation, and then it goes only faster...
+     */
+    for (cnt=0; cnt<16; ++cnt) {
+	Z.hi = Htable[cnt].hi;
+	Z.lo = Htable[cnt].lo;
+	Hshr4[cnt].lo = (Z.hi<<60)|(Z.lo>>4);
+	Hshr4[cnt].hi = (Z.hi>>4);
+	Hshl4[cnt]    = (u8)(Z.lo<<4);
+    }
+
+    do {
+	for (Z.lo=0, Z.hi=0, cnt=15; cnt; --cnt) {
+		nlo  = ((const u8 *)Xi)[cnt];
+		nlo ^= inp[cnt];
+		nhi  = nlo>>4;
+		nlo &= 0xf;
+
+		Z.hi ^= Htable[nlo].hi;
+		Z.lo ^= Htable[nlo].lo;
+
+		rem = (size_t)Z.lo&0xff;
+
+		Z.lo = (Z.hi<<56)|(Z.lo>>8);
+		Z.hi = (Z.hi>>8);
+
+		Z.hi ^= Hshr4[nhi].hi;
+		Z.lo ^= Hshr4[nhi].lo;
+		Z.hi ^= (u64)rem_8bit[rem^Hshl4[nhi]]<<48;
+	}
+
+	nlo  = ((const u8 *)Xi)[0];
+	nlo ^= inp[0];
+	nhi  = nlo>>4;
+	nlo &= 0xf;
+
+	Z.hi ^= Htable[nlo].hi;
+	Z.lo ^= Htable[nlo].lo;
+
+	rem = (size_t)Z.lo&0xf;
+
+	Z.lo = (Z.hi<<60)|(Z.lo>>4);
+	Z.hi = (Z.hi>>4);
+
+	Z.hi ^= Htable[nhi].hi;
+	Z.lo ^= Htable[nhi].lo;
+	Z.hi ^= ((u64)rem_8bit[rem<<4])<<48;
+#endif
+
+	if (is_endian.little) {
+#ifdef BSWAP8
+		Xi[0] = BSWAP8(Z.hi);
+		Xi[1] = BSWAP8(Z.lo);
+#else
+		u8 *p = (u8 *)Xi;
+		u32 v;
+		v = (u32)(Z.hi>>32);	PUTU32(p,v);
+		v = (u32)(Z.hi);	PUTU32(p+4,v);
+		v = (u32)(Z.lo>>32);	PUTU32(p+8,v);
+		v = (u32)(Z.lo);	PUTU32(p+12,v);
+#endif
+	}
+	else {
+		Xi[0] = Z.hi;
+		Xi[1] = Z.lo;
+	}
+    } while (inp+=16, len-=16);
+}
+#endif
+#else
+void gcm_gmult_4bit(u64 Xi[2],const u128 Htable[16]);
+void gcm_ghash_4bit(u64 Xi[2],const u128 Htable[16],const u8 *inp,size_t len);
+#endif
+
+#define GCM_MUL(ctx,Xi)   gcm_gmult_4bit(ctx->Xi.u,ctx->Htable)
+#if defined(GHASH_ASM) || !defined(OPENSSL_SMALL_FOOTPRINT)
+#define GHASH(ctx,in,len) gcm_ghash_4bit((ctx)->Xi.u,(ctx)->Htable,in,len)
+/* GHASH_CHUNK is "stride parameter" missioned to mitigate cache
+ * trashing effect. In other words idea is to hash data while it's
+ * still in L1 cache after encryption pass... */
+#define GHASH_CHUNK       (3*1024)
+#endif
+
+#else	/* TABLE_BITS */
+
+static void gcm_gmult_1bit(u64 Xi[2],const u64 H[2])
+{
+	u128 V,Z = { 0,0 };
+	long X;
+	int  i,j;
+	const long *xi = (const long *)Xi;
+	const union { long one; char little; } is_endian = {1};
+
+	V.hi = H[0];	/* H is in host byte order, no byte swapping */
+	V.lo = H[1];
+
+	for (j=0; j<16/sizeof(long); ++j) {
+		if (is_endian.little) {
+			if (sizeof(long)==8) {
+#ifdef BSWAP8
+				X = (long)(BSWAP8(xi[j]));
+#else
+				const u8 *p = (const u8 *)(xi+j);
+				X = (long)((u64)GETU32(p)<<32|GETU32(p+4));
+#endif
+			}
+			else {
+				const u8 *p = (const u8 *)(xi+j);
+				X = (long)GETU32(p);
+			}
+		}
+		else
+			X = xi[j];
+
+		for (i=0; i<8*sizeof(long); ++i, X<<=1) {
+			u64 M = (u64)(X>>(8*sizeof(long)-1));
+			Z.hi ^= V.hi&M;
+			Z.lo ^= V.lo&M;
+
+			REDUCE1BIT(V);
+		}
+	}
+
+	if (is_endian.little) {
+#ifdef BSWAP8
+		Xi[0] = BSWAP8(Z.hi);
+		Xi[1] = BSWAP8(Z.lo);
+#else
+		u8 *p = (u8 *)Xi;
+		u32 v;
+		v = (u32)(Z.hi>>32);	PUTU32(p,v);
+		v = (u32)(Z.hi);	PUTU32(p+4,v);
+		v = (u32)(Z.lo>>32);	PUTU32(p+8,v);
+		v = (u32)(Z.lo);	PUTU32(p+12,v);
+#endif
+	}
+	else {
+		Xi[0] = Z.hi;
+		Xi[1] = Z.lo;
+	}
+}
+#define GCM_MUL(ctx,Xi)	  gcm_gmult_1bit(ctx->Xi.u,ctx->H.u)
+
+#endif
+
+#if	TABLE_BITS==4 && defined(GHASH_ASM)
+# if	!defined(I386_ONLY) && \
+	(defined(__i386)	|| defined(__i386__)	|| \
+	 defined(__x86_64)	|| defined(__x86_64__)	|| \
+	 defined(_M_IX86)	|| defined(_M_AMD64)	|| defined(_M_X64))
+#  define GHASH_ASM_X86_OR_64
+#  define GCM_FUNCREF_4BIT
+extern unsigned int OPENSSL_ia32cap_P[2];
+
+void gcm_init_clmul(u128 Htable[16],const u64 Xi[2]);
+void gcm_gmult_clmul(u64 Xi[2],const u128 Htable[16]);
+void gcm_ghash_clmul(u64 Xi[2],const u128 Htable[16],const u8 *inp,size_t len);
+
+#  if	defined(__i386) || defined(__i386__) || defined(_M_IX86)
+#   define GHASH_ASM_X86
+void gcm_gmult_4bit_mmx(u64 Xi[2],const u128 Htable[16]);
+void gcm_ghash_4bit_mmx(u64 Xi[2],const u128 Htable[16],const u8 *inp,size_t len);
+
+void gcm_gmult_4bit_x86(u64 Xi[2],const u128 Htable[16]);
+void gcm_ghash_4bit_x86(u64 Xi[2],const u128 Htable[16],const u8 *inp,size_t len);
+#  endif
+# elif defined(__arm__) || defined(__arm)
+#  include "arm_arch.h"
+#  if __ARM_ARCH__>=7
+#   define GHASH_ASM_ARM
+#   define GCM_FUNCREF_4BIT
+void gcm_gmult_neon(u64 Xi[2],const u128 Htable[16]);
+void gcm_ghash_neon(u64 Xi[2],const u128 Htable[16],const u8 *inp,size_t len);
+#  endif
+# endif
+#endif
+
+#ifdef GCM_FUNCREF_4BIT
+# undef  GCM_MUL
+# define GCM_MUL(ctx,Xi)	(*gcm_gmult_p)(ctx->Xi.u,ctx->Htable)
+# ifdef GHASH
+#  undef  GHASH
+#  define GHASH(ctx,in,len)	(*gcm_ghash_p)(ctx->Xi.u,ctx->Htable,in,len)
+# endif
+#endif
+
+void CRYPTO_gcm128_init(GCM128_CONTEXT *ctx,void *key,block128_f block)
+{
+	const union { long one; char little; } is_endian = {1};
+
+	memset(ctx,0,sizeof(*ctx));
+	ctx->block = block;
+	ctx->key   = key;
+
+	(*block)(ctx->H.c,ctx->H.c,key);
+
+	if (is_endian.little) {
+		/* H is stored in host byte order */
+#ifdef BSWAP8
+		ctx->H.u[0] = BSWAP8(ctx->H.u[0]);
+		ctx->H.u[1] = BSWAP8(ctx->H.u[1]);
+#else
+		u8 *p = ctx->H.c;
+		u64 hi,lo;
+		hi = (u64)GETU32(p)  <<32|GETU32(p+4);
+		lo = (u64)GETU32(p+8)<<32|GETU32(p+12);
+		ctx->H.u[0] = hi;
+		ctx->H.u[1] = lo;
+#endif
+	}
+
+#if	TABLE_BITS==8
+	gcm_init_8bit(ctx->Htable,ctx->H.u);
+#elif	TABLE_BITS==4
+# if	defined(GHASH_ASM_X86_OR_64)
+#  if	!defined(GHASH_ASM_X86) || defined(OPENSSL_IA32_SSE2)
+	if (OPENSSL_ia32cap_P[0]&(1<<24) &&	/* check FXSR bit */
+	    OPENSSL_ia32cap_P[1]&(1<<1) ) {	/* check PCLMULQDQ bit */
+		gcm_init_clmul(ctx->Htable,ctx->H.u);
+		ctx->gmult = gcm_gmult_clmul;
+		ctx->ghash = gcm_ghash_clmul;
+		return;
+	}
+#  endif
+	gcm_init_4bit(ctx->Htable,ctx->H.u);
+#  if	defined(GHASH_ASM_X86)			/* x86 only */
+#   if defined(OPENSSL_IA32_SSE2)
+	if (OPENSSL_ia32cap_P[0]&(1<<25)) {	/* check SSE bit */
+#   else
+	if (OPENSSL_ia32cap_P[0]&(1<<23)) {	/* check MMX bit */
+#   endif
+		ctx->gmult = gcm_gmult_4bit_mmx;
+		ctx->ghash = gcm_ghash_4bit_mmx;
+	} else {
+		ctx->gmult = gcm_gmult_4bit_x86;
+		ctx->ghash = gcm_ghash_4bit_x86;
+	}
+#  else
+	ctx->gmult = gcm_gmult_4bit;
+	ctx->ghash = gcm_ghash_4bit;
+#  endif
+# elif	defined(GHASH_ASM_ARM)
+	if (OPENSSL_armcap_P & ARMV7_NEON) {
+		ctx->gmult = gcm_gmult_neon;
+		ctx->ghash = gcm_ghash_neon;
+	} else {
+		gcm_init_4bit(ctx->Htable,ctx->H.u);
+		ctx->gmult = gcm_gmult_4bit;
+		ctx->ghash = gcm_ghash_4bit;
+	}
+# else
+	gcm_init_4bit(ctx->Htable,ctx->H.u);
+# endif
+#endif
+}
+
+void CRYPTO_gcm128_setiv(GCM128_CONTEXT *ctx,const unsigned char *iv,size_t len)
+{
+	const union { long one; char little; } is_endian = {1};
+	unsigned int ctr;
+#ifdef GCM_FUNCREF_4BIT
+	void (*gcm_gmult_p)(u64 Xi[2],const u128 Htable[16])	= ctx->gmult;
+#endif
+
+	ctx->Yi.u[0]  = 0;
+	ctx->Yi.u[1]  = 0;
+	ctx->Xi.u[0]  = 0;
+	ctx->Xi.u[1]  = 0;
+	ctx->len.u[0] = 0;	/* AAD length */
+	ctx->len.u[1] = 0;	/* message length */
+	ctx->ares = 0;
+	ctx->mres = 0;
+
+	if (len==12) {
+		memcpy(ctx->Yi.c,iv,12);
+		ctx->Yi.c[15]=1;
+		ctr=1;
+	}
+	else {
+		size_t i;
+		u64 len0 = len;
+
+		while (len>=16) {
+			for (i=0; i<16; ++i) ctx->Yi.c[i] ^= iv[i];
+			GCM_MUL(ctx,Yi);
+			iv += 16;
+			len -= 16;
+		}
+		if (len) {
+			for (i=0; i<len; ++i) ctx->Yi.c[i] ^= iv[i];
+			GCM_MUL(ctx,Yi);
+		}
+		len0 <<= 3;
+		if (is_endian.little) {
+#ifdef BSWAP8
+			ctx->Yi.u[1]  ^= BSWAP8(len0);
+#else
+			ctx->Yi.c[8]  ^= (u8)(len0>>56);
+			ctx->Yi.c[9]  ^= (u8)(len0>>48);
+			ctx->Yi.c[10] ^= (u8)(len0>>40);
+			ctx->Yi.c[11] ^= (u8)(len0>>32);
+			ctx->Yi.c[12] ^= (u8)(len0>>24);
+			ctx->Yi.c[13] ^= (u8)(len0>>16);
+			ctx->Yi.c[14] ^= (u8)(len0>>8);
+			ctx->Yi.c[15] ^= (u8)(len0);
+#endif
+		}
+		else
+			ctx->Yi.u[1]  ^= len0;
+
+		GCM_MUL(ctx,Yi);
+
+		if (is_endian.little)
+			ctr = GETU32(ctx->Yi.c+12);
+		else
+			ctr = ctx->Yi.d[3];
+	}
+
+	(*ctx->block)(ctx->Yi.c,ctx->EK0.c,ctx->key);
+	++ctr;
+	if (is_endian.little)
+		PUTU32(ctx->Yi.c+12,ctr);
+	else
+		ctx->Yi.d[3] = ctr;
+}
+
+int CRYPTO_gcm128_aad(GCM128_CONTEXT *ctx,const unsigned char *aad,size_t len)
+{
+	size_t i;
+	unsigned int n;
+	u64 alen = ctx->len.u[0];
+#ifdef GCM_FUNCREF_4BIT
+	void (*gcm_gmult_p)(u64 Xi[2],const u128 Htable[16])	= ctx->gmult;
+# ifdef GHASH
+	void (*gcm_ghash_p)(u64 Xi[2],const u128 Htable[16],
+				const u8 *inp,size_t len)	= ctx->ghash;
+# endif
+#endif
+
+	if (ctx->len.u[1]) return -2;
+
+	alen += len;
+	if (alen>(U64(1)<<61) || (sizeof(len)==8 && alen<len))
+		return -1;
+	ctx->len.u[0] = alen;
+
+	n = ctx->ares;
+	if (n) {
+		while (n && len) {
+			ctx->Xi.c[n] ^= *(aad++);
+			--len;
+			n = (n+1)%16;
+		}
+		if (n==0) GCM_MUL(ctx,Xi);
+		else {
+			ctx->ares = n;
+			return 0;
+		}
+	}
+
+#ifdef GHASH
+	if ((i = (len&(size_t)-16))) {
+		GHASH(ctx,aad,i);
+		aad += i;
+		len -= i;
+	}
+#else
+	while (len>=16) {
+		for (i=0; i<16; ++i) ctx->Xi.c[i] ^= aad[i];
+		GCM_MUL(ctx,Xi);
+		aad += 16;
+		len -= 16;
+	}
+#endif
+	if (len) {
+		n = (unsigned int)len;
+		for (i=0; i<len; ++i) ctx->Xi.c[i] ^= aad[i];
+	}
+
+	ctx->ares = n;
+	return 0;
+}
+
+int CRYPTO_gcm128_encrypt(GCM128_CONTEXT *ctx,
+		const unsigned char *in, unsigned char *out,
+		size_t len)
+{
+	const union { long one; char little; } is_endian = {1};
+	unsigned int n, ctr;
+	size_t i;
+	u64        mlen  = ctx->len.u[1];
+	block128_f block = ctx->block;
+	void      *key   = ctx->key;
+#ifdef GCM_FUNCREF_4BIT
+	void (*gcm_gmult_p)(u64 Xi[2],const u128 Htable[16])	= ctx->gmult;
+# ifdef GHASH
+	void (*gcm_ghash_p)(u64 Xi[2],const u128 Htable[16],
+				const u8 *inp,size_t len)	= ctx->ghash;
+# endif
+#endif
+
+#if 0
+	n = (unsigned int)mlen%16; /* alternative to ctx->mres */
+#endif
+	mlen += len;
+	if (mlen>((U64(1)<<36)-32) || (sizeof(len)==8 && mlen<len))
+		return -1;
+	ctx->len.u[1] = mlen;
+
+	if (ctx->ares) {
+		/* First call to encrypt finalizes GHASH(AAD) */
+		GCM_MUL(ctx,Xi);
+		ctx->ares = 0;
+	}
+
+	if (is_endian.little)
+		ctr = GETU32(ctx->Yi.c+12);
+	else
+		ctr = ctx->Yi.d[3];
+
+	n = ctx->mres;
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+	if (16%sizeof(size_t) == 0) do {	/* always true actually */
+		if (n) {
+			while (n && len) {
+				ctx->Xi.c[n] ^= *(out++) = *(in++)^ctx->EKi.c[n];
+				--len;
+				n = (n+1)%16;
+			}
+			if (n==0) GCM_MUL(ctx,Xi);
+			else {
+				ctx->mres = n;
+				return 0;
+			}
+		}
+#if defined(STRICT_ALIGNMENT)
+		if (((size_t)in|(size_t)out)%sizeof(size_t) != 0)
+			break;
+#endif
+#if defined(GHASH) && defined(GHASH_CHUNK)
+		while (len>=GHASH_CHUNK) {
+		    size_t j=GHASH_CHUNK;
+
+		    while (j) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+			for (i=0; i<16; i+=sizeof(size_t))
+				*(size_t *)(out+i) =
+				*(size_t *)(in+i)^*(size_t *)(ctx->EKi.c+i);
+			out += 16;
+			in  += 16;
+			j   -= 16;
+		    }
+		    GHASH(ctx,out-GHASH_CHUNK,GHASH_CHUNK);
+		    len -= GHASH_CHUNK;
+		}
+		if ((i = (len&(size_t)-16))) {
+		    size_t j=i;
+
+		    while (len>=16) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+			for (i=0; i<16; i+=sizeof(size_t))
+				*(size_t *)(out+i) =
+				*(size_t *)(in+i)^*(size_t *)(ctx->EKi.c+i);
+			out += 16;
+			in  += 16;
+			len -= 16;
+		    }
+		    GHASH(ctx,out-j,j);
+		}
+#else
+		while (len>=16) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+			for (i=0; i<16; i+=sizeof(size_t))
+				*(size_t *)(ctx->Xi.c+i) ^=
+				*(size_t *)(out+i) =
+				*(size_t *)(in+i)^*(size_t *)(ctx->EKi.c+i);
+			GCM_MUL(ctx,Xi);
+			out += 16;
+			in  += 16;
+			len -= 16;
+		}
+#endif
+		if (len) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+			while (len--) {
+				ctx->Xi.c[n] ^= out[n] = in[n]^ctx->EKi.c[n];
+				++n;
+			}
+		}
+
+		ctx->mres = n;
+		return 0;
+	} while(0);
+#endif
+	for (i=0;i<len;++i) {
+		if (n==0) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+		}
+		ctx->Xi.c[n] ^= out[i] = in[i]^ctx->EKi.c[n];
+		n = (n+1)%16;
+		if (n==0)
+			GCM_MUL(ctx,Xi);
+	}
+
+	ctx->mres = n;
+	return 0;
+}
+
+int CRYPTO_gcm128_decrypt(GCM128_CONTEXT *ctx,
+		const unsigned char *in, unsigned char *out,
+		size_t len)
+{
+	const union { long one; char little; } is_endian = {1};
+	unsigned int n, ctr;
+	size_t i;
+	u64        mlen  = ctx->len.u[1];
+	block128_f block = ctx->block;
+	void      *key   = ctx->key;
+#ifdef GCM_FUNCREF_4BIT
+	void (*gcm_gmult_p)(u64 Xi[2],const u128 Htable[16])	= ctx->gmult;
+# ifdef GHASH
+	void (*gcm_ghash_p)(u64 Xi[2],const u128 Htable[16],
+				const u8 *inp,size_t len)	= ctx->ghash;
+# endif
+#endif
+
+	mlen += len;
+	if (mlen>((U64(1)<<36)-32) || (sizeof(len)==8 && mlen<len))
+		return -1;
+	ctx->len.u[1] = mlen;
+
+	if (ctx->ares) {
+		/* First call to decrypt finalizes GHASH(AAD) */
+		GCM_MUL(ctx,Xi);
+		ctx->ares = 0;
+	}
+
+	if (is_endian.little)
+		ctr = GETU32(ctx->Yi.c+12);
+	else
+		ctr = ctx->Yi.d[3];
+
+	n = ctx->mres;
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+	if (16%sizeof(size_t) == 0) do {	/* always true actually */
+		if (n) {
+			while (n && len) {
+				u8 c = *(in++);
+				*(out++) = c^ctx->EKi.c[n];
+				ctx->Xi.c[n] ^= c;
+				--len;
+				n = (n+1)%16;
+			}
+			if (n==0) GCM_MUL (ctx,Xi);
+			else {
+				ctx->mres = n;
+				return 0;
+			}
+		}
+#if defined(STRICT_ALIGNMENT)
+		if (((size_t)in|(size_t)out)%sizeof(size_t) != 0)
+			break;
+#endif
+#if defined(GHASH) && defined(GHASH_CHUNK)
+		while (len>=GHASH_CHUNK) {
+		    size_t j=GHASH_CHUNK;
+
+		    GHASH(ctx,in,GHASH_CHUNK);
+		    while (j) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+			for (i=0; i<16; i+=sizeof(size_t))
+				*(size_t *)(out+i) =
+				*(size_t *)(in+i)^*(size_t *)(ctx->EKi.c+i);
+			out += 16;
+			in  += 16;
+			j   -= 16;
+		    }
+		    len -= GHASH_CHUNK;
+		}
+		if ((i = (len&(size_t)-16))) {
+		    GHASH(ctx,in,i);
+		    while (len>=16) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+			for (i=0; i<16; i+=sizeof(size_t))
+				*(size_t *)(out+i) =
+				*(size_t *)(in+i)^*(size_t *)(ctx->EKi.c+i);
+			out += 16;
+			in  += 16;
+			len -= 16;
+		    }
+		}
+#else
+		while (len>=16) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+			for (i=0; i<16; i+=sizeof(size_t)) {
+				size_t c = *(size_t *)(in+i);
+				*(size_t *)(out+i) = c^*(size_t *)(ctx->EKi.c+i);
+				*(size_t *)(ctx->Xi.c+i) ^= c;
+			}
+			GCM_MUL(ctx,Xi);
+			out += 16;
+			in  += 16;
+			len -= 16;
+		}
+#endif
+		if (len) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+			while (len--) {
+				u8 c = in[n];
+				ctx->Xi.c[n] ^= c;
+				out[n] = c^ctx->EKi.c[n];
+				++n;
+			}
+		}
+
+		ctx->mres = n;
+		return 0;
+	} while(0);
+#endif
+	for (i=0;i<len;++i) {
+		u8 c;
+		if (n==0) {
+			(*block)(ctx->Yi.c,ctx->EKi.c,key);
+			++ctr;
+			if (is_endian.little)
+				PUTU32(ctx->Yi.c+12,ctr);
+			else
+				ctx->Yi.d[3] = ctr;
+		}
+		c = in[i];
+		out[i] = c^ctx->EKi.c[n];
+		ctx->Xi.c[n] ^= c;
+		n = (n+1)%16;
+		if (n==0)
+			GCM_MUL(ctx,Xi);
+	}
+
+	ctx->mres = n;
+	return 0;
+}
+
+int CRYPTO_gcm128_encrypt_ctr32(GCM128_CONTEXT *ctx,
+		const unsigned char *in, unsigned char *out,
+		size_t len, ctr128_f stream)
+{
+	const union { long one; char little; } is_endian = {1};
+	unsigned int n, ctr;
+	size_t i;
+	u64   mlen = ctx->len.u[1];
+	void *key  = ctx->key;
+#ifdef GCM_FUNCREF_4BIT
+	void (*gcm_gmult_p)(u64 Xi[2],const u128 Htable[16])	= ctx->gmult;
+# ifdef GHASH
+	void (*gcm_ghash_p)(u64 Xi[2],const u128 Htable[16],
+				const u8 *inp,size_t len)	= ctx->ghash;
+# endif
+#endif
+
+	mlen += len;
+	if (mlen>((U64(1)<<36)-32) || (sizeof(len)==8 && mlen<len))
+		return -1;
+	ctx->len.u[1] = mlen;
+
+	if (ctx->ares) {
+		/* First call to encrypt finalizes GHASH(AAD) */
+		GCM_MUL(ctx,Xi);
+		ctx->ares = 0;
+	}
+
+	if (is_endian.little)
+		ctr = GETU32(ctx->Yi.c+12);
+	else
+		ctr = ctx->Yi.d[3];
+
+	n = ctx->mres;
+	if (n) {
+		while (n && len) {
+			ctx->Xi.c[n] ^= *(out++) = *(in++)^ctx->EKi.c[n];
+			--len;
+			n = (n+1)%16;
+		}
+		if (n==0) GCM_MUL(ctx,Xi);
+		else {
+			ctx->mres = n;
+			return 0;
+		}
+	}
+#if defined(GHASH) && !defined(OPENSSL_SMALL_FOOTPRINT)
+	while (len>=GHASH_CHUNK) {
+		(*stream)(in,out,GHASH_CHUNK/16,key,ctx->Yi.c);
+		ctr += GHASH_CHUNK/16;
+		if (is_endian.little)
+			PUTU32(ctx->Yi.c+12,ctr);
+		else
+			ctx->Yi.d[3] = ctr;
+		GHASH(ctx,out,GHASH_CHUNK);
+		out += GHASH_CHUNK;
+		in  += GHASH_CHUNK;
+		len -= GHASH_CHUNK;
+	}
+#endif
+	if ((i = (len&(size_t)-16))) {
+		size_t j=i/16;
+
+		(*stream)(in,out,j,key,ctx->Yi.c);
+		ctr += (unsigned int)j;
+		if (is_endian.little)
+			PUTU32(ctx->Yi.c+12,ctr);
+		else
+			ctx->Yi.d[3] = ctr;
+		in  += i;
+		len -= i;
+#if defined(GHASH)
+		GHASH(ctx,out,i);
+		out += i;
+#else
+		while (j--) {
+			for (i=0;i<16;++i) ctx->Xi.c[i] ^= out[i];
+			GCM_MUL(ctx,Xi);
+			out += 16;
+		}
+#endif
+	}
+	if (len) {
+		(*ctx->block)(ctx->Yi.c,ctx->EKi.c,key);
+		++ctr;
+		if (is_endian.little)
+			PUTU32(ctx->Yi.c+12,ctr);
+		else
+			ctx->Yi.d[3] = ctr;
+		while (len--) {
+			ctx->Xi.c[n] ^= out[n] = in[n]^ctx->EKi.c[n];
+			++n;
+		}
+	}
+
+	ctx->mres = n;
+	return 0;
+}
+
+int CRYPTO_gcm128_decrypt_ctr32(GCM128_CONTEXT *ctx,
+		const unsigned char *in, unsigned char *out,
+		size_t len,ctr128_f stream)
+{
+	const union { long one; char little; } is_endian = {1};
+	unsigned int n, ctr;
+	size_t i;
+	u64   mlen = ctx->len.u[1];
+	void *key  = ctx->key;
+#ifdef GCM_FUNCREF_4BIT
+	void (*gcm_gmult_p)(u64 Xi[2],const u128 Htable[16])	= ctx->gmult;
+# ifdef GHASH
+	void (*gcm_ghash_p)(u64 Xi[2],const u128 Htable[16],
+				const u8 *inp,size_t len)	= ctx->ghash;
+# endif
+#endif
+
+	mlen += len;
+	if (mlen>((U64(1)<<36)-32) || (sizeof(len)==8 && mlen<len))
+		return -1;
+	ctx->len.u[1] = mlen;
+
+	if (ctx->ares) {
+		/* First call to decrypt finalizes GHASH(AAD) */
+		GCM_MUL(ctx,Xi);
+		ctx->ares = 0;
+	}
+
+	if (is_endian.little)
+		ctr = GETU32(ctx->Yi.c+12);
+	else
+		ctr = ctx->Yi.d[3];
+
+	n = ctx->mres;
+	if (n) {
+		while (n && len) {
+			u8 c = *(in++);
+			*(out++) = c^ctx->EKi.c[n];
+			ctx->Xi.c[n] ^= c;
+			--len;
+			n = (n+1)%16;
+		}
+		if (n==0) GCM_MUL (ctx,Xi);
+		else {
+			ctx->mres = n;
+			return 0;
+		}
+	}
+#if defined(GHASH) && !defined(OPENSSL_SMALL_FOOTPRINT)
+	while (len>=GHASH_CHUNK) {
+		GHASH(ctx,in,GHASH_CHUNK);
+		(*stream)(in,out,GHASH_CHUNK/16,key,ctx->Yi.c);
+		ctr += GHASH_CHUNK/16;
+		if (is_endian.little)
+			PUTU32(ctx->Yi.c+12,ctr);
+		else
+			ctx->Yi.d[3] = ctr;
+		out += GHASH_CHUNK;
+		in  += GHASH_CHUNK;
+		len -= GHASH_CHUNK;
+	}
+#endif
+	if ((i = (len&(size_t)-16))) {
+		size_t j=i/16;
+
+#if defined(GHASH)
+		GHASH(ctx,in,i);
+#else
+		while (j--) {
+			size_t k;
+			for (k=0;k<16;++k) ctx->Xi.c[k] ^= in[k];
+			GCM_MUL(ctx,Xi);
+			in += 16;
+		}
+		j   = i/16;
+		in -= i;
+#endif
+		(*stream)(in,out,j,key,ctx->Yi.c);
+		ctr += (unsigned int)j;
+		if (is_endian.little)
+			PUTU32(ctx->Yi.c+12,ctr);
+		else
+			ctx->Yi.d[3] = ctr;
+		out += i;
+		in  += i;
+		len -= i;
+	}
+	if (len) {
+		(*ctx->block)(ctx->Yi.c,ctx->EKi.c,key);
+		++ctr;
+		if (is_endian.little)
+			PUTU32(ctx->Yi.c+12,ctr);
+		else
+			ctx->Yi.d[3] = ctr;
+		while (len--) {
+			u8 c = in[n];
+			ctx->Xi.c[n] ^= c;
+			out[n] = c^ctx->EKi.c[n];
+			++n;
+		}
+	}
+
+	ctx->mres = n;
+	return 0;
+}
+
+int CRYPTO_gcm128_finish(GCM128_CONTEXT *ctx,const unsigned char *tag,
+			size_t len)
+{
+	const union { long one; char little; } is_endian = {1};
+	u64 alen = ctx->len.u[0]<<3;
+	u64 clen = ctx->len.u[1]<<3;
+#ifdef GCM_FUNCREF_4BIT
+	void (*gcm_gmult_p)(u64 Xi[2],const u128 Htable[16])	= ctx->gmult;
+#endif
+
+	if (ctx->mres)
+		GCM_MUL(ctx,Xi);
+
+	if (is_endian.little) {
+#ifdef BSWAP8
+		alen = BSWAP8(alen);
+		clen = BSWAP8(clen);
+#else
+		u8 *p = ctx->len.c;
+
+		ctx->len.u[0] = alen;
+		ctx->len.u[1] = clen;
+
+		alen = (u64)GETU32(p)  <<32|GETU32(p+4);
+		clen = (u64)GETU32(p+8)<<32|GETU32(p+12);
+#endif
+	}
+
+	ctx->Xi.u[0] ^= alen;
+	ctx->Xi.u[1] ^= clen;
+	GCM_MUL(ctx,Xi);
+
+	ctx->Xi.u[0] ^= ctx->EK0.u[0];
+	ctx->Xi.u[1] ^= ctx->EK0.u[1];
+
+	if (tag && len<=sizeof(ctx->Xi))
+		return memcmp(ctx->Xi.c,tag,len);
+	else
+		return -1;
+}
+
+void CRYPTO_gcm128_tag(GCM128_CONTEXT *ctx, unsigned char *tag, size_t len)
+{
+	CRYPTO_gcm128_finish(ctx, NULL, 0);
+	memcpy(tag, ctx->Xi.c, len<=sizeof(ctx->Xi.c)?len:sizeof(ctx->Xi.c));
+}
+
+GCM128_CONTEXT *CRYPTO_gcm128_new(void *key, block128_f block)
+{
+	GCM128_CONTEXT *ret;
+
+	if ((ret = (GCM128_CONTEXT *)OPENSSL_malloc(sizeof(GCM128_CONTEXT))))
+		CRYPTO_gcm128_init(ret,key,block);
+
+	return ret;
+}
+
+void CRYPTO_gcm128_release(GCM128_CONTEXT *ctx)
+{
+	if (ctx) {
+		OPENSSL_cleanse(ctx,sizeof(*ctx));
+		OPENSSL_free(ctx);
+	}
+}
+
+#if defined(SELFTEST)
+#include <stdio.h>
+#include <openssl/aes.h>
+
+/* Test Case 1 */
+static const u8	K1[16],
+		*P1=NULL,
+		*A1=NULL,
+		IV1[12],
+		*C1=NULL,
+		T1[]=  {0x58,0xe2,0xfc,0xce,0xfa,0x7e,0x30,0x61,0x36,0x7f,0x1d,0x57,0xa4,0xe7,0x45,0x5a};
+
+/* Test Case 2 */
+#define K2 K1
+#define A2 A1
+#define IV2 IV1
+static const u8	P2[16],
+		C2[]=  {0x03,0x88,0xda,0xce,0x60,0xb6,0xa3,0x92,0xf3,0x28,0xc2,0xb9,0x71,0xb2,0xfe,0x78},
+		T2[]=  {0xab,0x6e,0x47,0xd4,0x2c,0xec,0x13,0xbd,0xf5,0x3a,0x67,0xb2,0x12,0x57,0xbd,0xdf};
+
+/* Test Case 3 */
+#define A3 A2
+static const u8	K3[]=  {0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c,0x6d,0x6a,0x8f,0x94,0x67,0x30,0x83,0x08},
+		P3[]=  {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
+			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
+			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
+			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39,0x1a,0xaf,0xd2,0x55},
+		IV3[]= {0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad,0xde,0xca,0xf8,0x88},
+		C3[]=  {0x42,0x83,0x1e,0xc2,0x21,0x77,0x74,0x24,0x4b,0x72,0x21,0xb7,0x84,0xd0,0xd4,0x9c,
+			0xe3,0xaa,0x21,0x2f,0x2c,0x02,0xa4,0xe0,0x35,0xc1,0x7e,0x23,0x29,0xac,0xa1,0x2e,
+			0x21,0xd5,0x14,0xb2,0x54,0x66,0x93,0x1c,0x7d,0x8f,0x6a,0x5a,0xac,0x84,0xaa,0x05,
+			0x1b,0xa3,0x0b,0x39,0x6a,0x0a,0xac,0x97,0x3d,0x58,0xe0,0x91,0x47,0x3f,0x59,0x85},
+		T3[]=  {0x4d,0x5c,0x2a,0xf3,0x27,0xcd,0x64,0xa6,0x2c,0xf3,0x5a,0xbd,0x2b,0xa6,0xfa,0xb4};
+
+/* Test Case 4 */
+#define K4 K3
+#define IV4 IV3
+static const u8	P4[]=  {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
+			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
+			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
+			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39},
+		A4[]=  {0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,
+			0xab,0xad,0xda,0xd2},
+		C4[]=  {0x42,0x83,0x1e,0xc2,0x21,0x77,0x74,0x24,0x4b,0x72,0x21,0xb7,0x84,0xd0,0xd4,0x9c,
+			0xe3,0xaa,0x21,0x2f,0x2c,0x02,0xa4,0xe0,0x35,0xc1,0x7e,0x23,0x29,0xac,0xa1,0x2e,
+			0x21,0xd5,0x14,0xb2,0x54,0x66,0x93,0x1c,0x7d,0x8f,0x6a,0x5a,0xac,0x84,0xaa,0x05,
+			0x1b,0xa3,0x0b,0x39,0x6a,0x0a,0xac,0x97,0x3d,0x58,0xe0,0x91},
+		T4[]=  {0x5b,0xc9,0x4f,0xbc,0x32,0x21,0xa5,0xdb,0x94,0xfa,0xe9,0x5a,0xe7,0x12,0x1a,0x47};
+
+/* Test Case 5 */
+#define K5 K4
+#define P5 P4
+#define A5 A4
+static const u8	IV5[]= {0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad},
+		C5[]=  {0x61,0x35,0x3b,0x4c,0x28,0x06,0x93,0x4a,0x77,0x7f,0xf5,0x1f,0xa2,0x2a,0x47,0x55,
+			0x69,0x9b,0x2a,0x71,0x4f,0xcd,0xc6,0xf8,0x37,0x66,0xe5,0xf9,0x7b,0x6c,0x74,0x23,
+			0x73,0x80,0x69,0x00,0xe4,0x9f,0x24,0xb2,0x2b,0x09,0x75,0x44,0xd4,0x89,0x6b,0x42,
+			0x49,0x89,0xb5,0xe1,0xeb,0xac,0x0f,0x07,0xc2,0x3f,0x45,0x98},
+		T5[]=  {0x36,0x12,0xd2,0xe7,0x9e,0x3b,0x07,0x85,0x56,0x1b,0xe1,0x4a,0xac,0xa2,0xfc,0xcb};
+
+/* Test Case 6 */
+#define K6 K5
+#define P6 P5
+#define A6 A5
+static const u8	IV6[]= {0x93,0x13,0x22,0x5d,0xf8,0x84,0x06,0xe5,0x55,0x90,0x9c,0x5a,0xff,0x52,0x69,0xaa,
+			0x6a,0x7a,0x95,0x38,0x53,0x4f,0x7d,0xa1,0xe4,0xc3,0x03,0xd2,0xa3,0x18,0xa7,0x28,
+			0xc3,0xc0,0xc9,0x51,0x56,0x80,0x95,0x39,0xfc,0xf0,0xe2,0x42,0x9a,0x6b,0x52,0x54,
+			0x16,0xae,0xdb,0xf5,0xa0,0xde,0x6a,0x57,0xa6,0x37,0xb3,0x9b},
+		C6[]=  {0x8c,0xe2,0x49,0x98,0x62,0x56,0x15,0xb6,0x03,0xa0,0x33,0xac,0xa1,0x3f,0xb8,0x94,
+			0xbe,0x91,0x12,0xa5,0xc3,0xa2,0x11,0xa8,0xba,0x26,0x2a,0x3c,0xca,0x7e,0x2c,0xa7,
+			0x01,0xe4,0xa9,0xa4,0xfb,0xa4,0x3c,0x90,0xcc,0xdc,0xb2,0x81,0xd4,0x8c,0x7c,0x6f,
+			0xd6,0x28,0x75,0xd2,0xac,0xa4,0x17,0x03,0x4c,0x34,0xae,0xe5},
+		T6[]=  {0x61,0x9c,0xc5,0xae,0xff,0xfe,0x0b,0xfa,0x46,0x2a,0xf4,0x3c,0x16,0x99,0xd0,0x50};
+
+/* Test Case 7 */
+static const u8 K7[24],
+		*P7=NULL,
+		*A7=NULL,
+		IV7[12],
+		*C7=NULL,
+		T7[]=  {0xcd,0x33,0xb2,0x8a,0xc7,0x73,0xf7,0x4b,0xa0,0x0e,0xd1,0xf3,0x12,0x57,0x24,0x35};
+
+/* Test Case 8 */
+#define K8 K7
+#define IV8 IV7
+#define A8 A7
+static const u8	P8[16],
+		C8[]=  {0x98,0xe7,0x24,0x7c,0x07,0xf0,0xfe,0x41,0x1c,0x26,0x7e,0x43,0x84,0xb0,0xf6,0x00},
+		T8[]=  {0x2f,0xf5,0x8d,0x80,0x03,0x39,0x27,0xab,0x8e,0xf4,0xd4,0x58,0x75,0x14,0xf0,0xfb};
+
+/* Test Case 9 */
+#define A9 A8
+static const u8	K9[]=  {0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c,0x6d,0x6a,0x8f,0x94,0x67,0x30,0x83,0x08,
+			0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c},
+		P9[]=  {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
+			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
+			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
+			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39,0x1a,0xaf,0xd2,0x55},
+		IV9[]= {0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad,0xde,0xca,0xf8,0x88},
+		C9[]=  {0x39,0x80,0xca,0x0b,0x3c,0x00,0xe8,0x41,0xeb,0x06,0xfa,0xc4,0x87,0x2a,0x27,0x57,
+			0x85,0x9e,0x1c,0xea,0xa6,0xef,0xd9,0x84,0x62,0x85,0x93,0xb4,0x0c,0xa1,0xe1,0x9c,
+			0x7d,0x77,0x3d,0x00,0xc1,0x44,0xc5,0x25,0xac,0x61,0x9d,0x18,0xc8,0x4a,0x3f,0x47,
+			0x18,0xe2,0x44,0x8b,0x2f,0xe3,0x24,0xd9,0xcc,0xda,0x27,0x10,0xac,0xad,0xe2,0x56},
+		T9[]=  {0x99,0x24,0xa7,0xc8,0x58,0x73,0x36,0xbf,0xb1,0x18,0x02,0x4d,0xb8,0x67,0x4a,0x14};
+
+/* Test Case 10 */
+#define K10 K9
+#define IV10 IV9
+static const u8	P10[]= {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
+			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
+			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
+			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39},
+		A10[]= {0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,
+			0xab,0xad,0xda,0xd2},
+		C10[]= {0x39,0x80,0xca,0x0b,0x3c,0x00,0xe8,0x41,0xeb,0x06,0xfa,0xc4,0x87,0x2a,0x27,0x57,
+			0x85,0x9e,0x1c,0xea,0xa6,0xef,0xd9,0x84,0x62,0x85,0x93,0xb4,0x0c,0xa1,0xe1,0x9c,
+			0x7d,0x77,0x3d,0x00,0xc1,0x44,0xc5,0x25,0xac,0x61,0x9d,0x18,0xc8,0x4a,0x3f,0x47,
+			0x18,0xe2,0x44,0x8b,0x2f,0xe3,0x24,0xd9,0xcc,0xda,0x27,0x10},
+		T10[]= {0x25,0x19,0x49,0x8e,0x80,0xf1,0x47,0x8f,0x37,0xba,0x55,0xbd,0x6d,0x27,0x61,0x8c};
+
+/* Test Case 11 */
+#define K11 K10
+#define P11 P10
+#define A11 A10
+static const u8	IV11[]={0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad},
+		C11[]= {0x0f,0x10,0xf5,0x99,0xae,0x14,0xa1,0x54,0xed,0x24,0xb3,0x6e,0x25,0x32,0x4d,0xb8,
+			0xc5,0x66,0x63,0x2e,0xf2,0xbb,0xb3,0x4f,0x83,0x47,0x28,0x0f,0xc4,0x50,0x70,0x57,
+			0xfd,0xdc,0x29,0xdf,0x9a,0x47,0x1f,0x75,0xc6,0x65,0x41,0xd4,0xd4,0xda,0xd1,0xc9,
+			0xe9,0x3a,0x19,0xa5,0x8e,0x8b,0x47,0x3f,0xa0,0xf0,0x62,0xf7},
+		T11[]= {0x65,0xdc,0xc5,0x7f,0xcf,0x62,0x3a,0x24,0x09,0x4f,0xcc,0xa4,0x0d,0x35,0x33,0xf8};
+
+/* Test Case 12 */
+#define K12 K11
+#define P12 P11
+#define A12 A11
+static const u8	IV12[]={0x93,0x13,0x22,0x5d,0xf8,0x84,0x06,0xe5,0x55,0x90,0x9c,0x5a,0xff,0x52,0x69,0xaa,
+			0x6a,0x7a,0x95,0x38,0x53,0x4f,0x7d,0xa1,0xe4,0xc3,0x03,0xd2,0xa3,0x18,0xa7,0x28,
+			0xc3,0xc0,0xc9,0x51,0x56,0x80,0x95,0x39,0xfc,0xf0,0xe2,0x42,0x9a,0x6b,0x52,0x54,
+			0x16,0xae,0xdb,0xf5,0xa0,0xde,0x6a,0x57,0xa6,0x37,0xb3,0x9b},
+		C12[]= {0xd2,0x7e,0x88,0x68,0x1c,0xe3,0x24,0x3c,0x48,0x30,0x16,0x5a,0x8f,0xdc,0xf9,0xff,
+			0x1d,0xe9,0xa1,0xd8,0xe6,0xb4,0x47,0xef,0x6e,0xf7,0xb7,0x98,0x28,0x66,0x6e,0x45,
+			0x81,0xe7,0x90,0x12,0xaf,0x34,0xdd,0xd9,0xe2,0xf0,0x37,0x58,0x9b,0x29,0x2d,0xb3,
+			0xe6,0x7c,0x03,0x67,0x45,0xfa,0x22,0xe7,0xe9,0xb7,0x37,0x3b},
+		T12[]= {0xdc,0xf5,0x66,0xff,0x29,0x1c,0x25,0xbb,0xb8,0x56,0x8f,0xc3,0xd3,0x76,0xa6,0xd9};
+
+/* Test Case 13 */
+static const u8	K13[32],
+		*P13=NULL,
+		*A13=NULL,
+		IV13[12],
+		*C13=NULL,
+		T13[]={0x53,0x0f,0x8a,0xfb,0xc7,0x45,0x36,0xb9,0xa9,0x63,0xb4,0xf1,0xc4,0xcb,0x73,0x8b};
+
+/* Test Case 14 */
+#define K14 K13
+#define A14 A13
+static const u8	P14[16],
+		IV14[12],
+		C14[]= {0xce,0xa7,0x40,0x3d,0x4d,0x60,0x6b,0x6e,0x07,0x4e,0xc5,0xd3,0xba,0xf3,0x9d,0x18},
+		T14[]= {0xd0,0xd1,0xc8,0xa7,0x99,0x99,0x6b,0xf0,0x26,0x5b,0x98,0xb5,0xd4,0x8a,0xb9,0x19};
+
+/* Test Case 15 */
+#define A15 A14
+static const u8	K15[]= {0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c,0x6d,0x6a,0x8f,0x94,0x67,0x30,0x83,0x08,
+			0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c,0x6d,0x6a,0x8f,0x94,0x67,0x30,0x83,0x08},
+		P15[]= {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
+			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
+			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
+			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39,0x1a,0xaf,0xd2,0x55},
+		IV15[]={0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad,0xde,0xca,0xf8,0x88},
+		C15[]= {0x52,0x2d,0xc1,0xf0,0x99,0x56,0x7d,0x07,0xf4,0x7f,0x37,0xa3,0x2a,0x84,0x42,0x7d,
+			0x64,0x3a,0x8c,0xdc,0xbf,0xe5,0xc0,0xc9,0x75,0x98,0xa2,0xbd,0x25,0x55,0xd1,0xaa,
+			0x8c,0xb0,0x8e,0x48,0x59,0x0d,0xbb,0x3d,0xa7,0xb0,0x8b,0x10,0x56,0x82,0x88,0x38,
+			0xc5,0xf6,0x1e,0x63,0x93,0xba,0x7a,0x0a,0xbc,0xc9,0xf6,0x62,0x89,0x80,0x15,0xad},
+		T15[]= {0xb0,0x94,0xda,0xc5,0xd9,0x34,0x71,0xbd,0xec,0x1a,0x50,0x22,0x70,0xe3,0xcc,0x6c};
+
+/* Test Case 16 */
+#define K16 K15
+#define IV16 IV15
+static const u8	P16[]= {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
+			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
+			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
+			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39},
+		A16[]= {0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,
+			0xab,0xad,0xda,0xd2},
+		C16[]= {0x52,0x2d,0xc1,0xf0,0x99,0x56,0x7d,0x07,0xf4,0x7f,0x37,0xa3,0x2a,0x84,0x42,0x7d,
+			0x64,0x3a,0x8c,0xdc,0xbf,0xe5,0xc0,0xc9,0x75,0x98,0xa2,0xbd,0x25,0x55,0xd1,0xaa,
+			0x8c,0xb0,0x8e,0x48,0x59,0x0d,0xbb,0x3d,0xa7,0xb0,0x8b,0x10,0x56,0x82,0x88,0x38,
+			0xc5,0xf6,0x1e,0x63,0x93,0xba,0x7a,0x0a,0xbc,0xc9,0xf6,0x62},
+		T16[]= {0x76,0xfc,0x6e,0xce,0x0f,0x4e,0x17,0x68,0xcd,0xdf,0x88,0x53,0xbb,0x2d,0x55,0x1b};
+
+/* Test Case 17 */
+#define K17 K16
+#define P17 P16
+#define A17 A16
+static const u8	IV17[]={0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad},
+		C17[]= {0xc3,0x76,0x2d,0xf1,0xca,0x78,0x7d,0x32,0xae,0x47,0xc1,0x3b,0xf1,0x98,0x44,0xcb,
+			0xaf,0x1a,0xe1,0x4d,0x0b,0x97,0x6a,0xfa,0xc5,0x2f,0xf7,0xd7,0x9b,0xba,0x9d,0xe0,
+			0xfe,0xb5,0x82,0xd3,0x39,0x34,0xa4,0xf0,0x95,0x4c,0xc2,0x36,0x3b,0xc7,0x3f,0x78,
+			0x62,0xac,0x43,0x0e,0x64,0xab,0xe4,0x99,0xf4,0x7c,0x9b,0x1f},
+		T17[]= {0x3a,0x33,0x7d,0xbf,0x46,0xa7,0x92,0xc4,0x5e,0x45,0x49,0x13,0xfe,0x2e,0xa8,0xf2};
+
+/* Test Case 18 */
+#define K18 K17
+#define P18 P17
+#define A18 A17
+static const u8	IV18[]={0x93,0x13,0x22,0x5d,0xf8,0x84,0x06,0xe5,0x55,0x90,0x9c,0x5a,0xff,0x52,0x69,0xaa,
+			0x6a,0x7a,0x95,0x38,0x53,0x4f,0x7d,0xa1,0xe4,0xc3,0x03,0xd2,0xa3,0x18,0xa7,0x28,
+			0xc3,0xc0,0xc9,0x51,0x56,0x80,0x95,0x39,0xfc,0xf0,0xe2,0x42,0x9a,0x6b,0x52,0x54,
+			0x16,0xae,0xdb,0xf5,0xa0,0xde,0x6a,0x57,0xa6,0x37,0xb3,0x9b},
+		C18[]= {0x5a,0x8d,0xef,0x2f,0x0c,0x9e,0x53,0xf1,0xf7,0x5d,0x78,0x53,0x65,0x9e,0x2a,0x20,
+			0xee,0xb2,0xb2,0x2a,0xaf,0xde,0x64,0x19,0xa0,0x58,0xab,0x4f,0x6f,0x74,0x6b,0xf4,
+			0x0f,0xc0,0xc3,0xb7,0x80,0xf2,0x44,0x45,0x2d,0xa3,0xeb,0xf1,0xc5,0xd8,0x2c,0xde,
+			0xa2,0x41,0x89,0x97,0x20,0x0e,0xf8,0x2e,0x44,0xae,0x7e,0x3f},
+		T18[]= {0xa4,0x4a,0x82,0x66,0xee,0x1c,0x8e,0xb0,0xc8,0xb5,0xd4,0xcf,0x5a,0xe9,0xf1,0x9a};
+
+#define TEST_CASE(n)	do {					\
+	u8 out[sizeof(P##n)];					\
+	AES_set_encrypt_key(K##n,sizeof(K##n)*8,&key);		\
+	CRYPTO_gcm128_init(&ctx,&key,(block128_f)AES_encrypt);	\
+	CRYPTO_gcm128_setiv(&ctx,IV##n,sizeof(IV##n));		\
+	memset(out,0,sizeof(out));				\
+	if (A##n) CRYPTO_gcm128_aad(&ctx,A##n,sizeof(A##n));	\
+	if (P##n) CRYPTO_gcm128_encrypt(&ctx,P##n,out,sizeof(out));	\
+	if (CRYPTO_gcm128_finish(&ctx,T##n,16) ||		\
+	    (C##n && memcmp(out,C##n,sizeof(out))))		\
+		ret++, printf ("encrypt test#%d failed.\n",n);	\
+	CRYPTO_gcm128_setiv(&ctx,IV##n,sizeof(IV##n));		\
+	memset(out,0,sizeof(out));				\
+	if (A##n) CRYPTO_gcm128_aad(&ctx,A##n,sizeof(A##n));	\
+	if (C##n) CRYPTO_gcm128_decrypt(&ctx,C##n,out,sizeof(out));	\
+	if (CRYPTO_gcm128_finish(&ctx,T##n,16) ||		\
+	    (P##n && memcmp(out,P##n,sizeof(out))))		\
+		ret++, printf ("decrypt test#%d failed.\n",n);	\
+	} while(0)
+
+int main()
+{
+	GCM128_CONTEXT ctx;
+	AES_KEY key;
+	int ret=0;
+
+	TEST_CASE(1);
+	TEST_CASE(2);
+	TEST_CASE(3);
+	TEST_CASE(4);
+	TEST_CASE(5);
+	TEST_CASE(6);
+	TEST_CASE(7);
+	TEST_CASE(8);
+	TEST_CASE(9);
+	TEST_CASE(10);
+	TEST_CASE(11);
+	TEST_CASE(12);
+	TEST_CASE(13);
+	TEST_CASE(14);
+	TEST_CASE(15);
+	TEST_CASE(16);
+	TEST_CASE(17);
+	TEST_CASE(18);
+
+#ifdef OPENSSL_CPUID_OBJ
+	{
+	size_t start,stop,gcm_t,ctr_t,OPENSSL_rdtsc();
+	union { u64 u; u8 c[1024]; } buf;
+	int i;
+
+	AES_set_encrypt_key(K1,sizeof(K1)*8,&key);
+	CRYPTO_gcm128_init(&ctx,&key,(block128_f)AES_encrypt);
+	CRYPTO_gcm128_setiv(&ctx,IV1,sizeof(IV1));
+
+	CRYPTO_gcm128_encrypt(&ctx,buf.c,buf.c,sizeof(buf));
+	start = OPENSSL_rdtsc();
+	CRYPTO_gcm128_encrypt(&ctx,buf.c,buf.c,sizeof(buf));
+	gcm_t = OPENSSL_rdtsc() - start;
+
+	CRYPTO_ctr128_encrypt(buf.c,buf.c,sizeof(buf),
+			&key,ctx.Yi.c,ctx.EKi.c,&ctx.mres,
+			(block128_f)AES_encrypt);
+	start = OPENSSL_rdtsc();
+	CRYPTO_ctr128_encrypt(buf.c,buf.c,sizeof(buf),
+			&key,ctx.Yi.c,ctx.EKi.c,&ctx.mres,
+			(block128_f)AES_encrypt);
+	ctr_t = OPENSSL_rdtsc() - start;
+
+	printf("%.2f-%.2f=%.2f\n",
+			gcm_t/(double)sizeof(buf),
+			ctr_t/(double)sizeof(buf),
+			(gcm_t-ctr_t)/(double)sizeof(buf));
+#ifdef GHASH
+	GHASH(&ctx,buf.c,sizeof(buf));
+	start = OPENSSL_rdtsc();
+	for (i=0;i<100;++i) GHASH(&ctx,buf.c,sizeof(buf));
+	gcm_t = OPENSSL_rdtsc() - start;
+	printf("%.2f\n",gcm_t/(double)sizeof(buf)/(double)i);
+#endif
+	}
+#endif
+
+	return ret;
+}
+#endif
diff --git a/crypto/modes/modes.h b/crypto/modes/modes.h
new file mode 100644
index 000000000000..f18215bb2b25
--- /dev/null
+++ b/crypto/modes/modes.h
@@ -0,0 +1,135 @@
+/* ====================================================================
+ * Copyright (c) 2008 The OpenSSL Project. All rights reserved.
+ *
+ * Rights for redistribution and usage in source and binary
+ * forms are granted according to the OpenSSL license.
+ */
+
+#include <stddef.h>
+
+typedef void (*block128_f)(const unsigned char in[16],
+			unsigned char out[16],
+			const void *key);
+
+typedef void (*cbc128_f)(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], int enc);
+
+typedef void (*ctr128_f)(const unsigned char *in, unsigned char *out,
+			size_t blocks, const void *key,
+			const unsigned char ivec[16]);
+
+typedef void (*ccm128_f)(const unsigned char *in, unsigned char *out,
+			size_t blocks, const void *key,
+			const unsigned char ivec[16],unsigned char cmac[16]);
+
+void CRYPTO_cbc128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block);
+void CRYPTO_cbc128_decrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block);
+
+void CRYPTO_ctr128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], unsigned char ecount_buf[16],
+			unsigned int *num, block128_f block);
+
+void CRYPTO_ctr128_encrypt_ctr32(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], unsigned char ecount_buf[16],
+			unsigned int *num, ctr128_f ctr);
+
+void CRYPTO_ofb128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], int *num,
+			block128_f block);
+
+void CRYPTO_cfb128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], int *num,
+			int enc, block128_f block);
+void CRYPTO_cfb128_8_encrypt(const unsigned char *in, unsigned char *out,
+			size_t length, const void *key,
+			unsigned char ivec[16], int *num,
+			int enc, block128_f block);
+void CRYPTO_cfb128_1_encrypt(const unsigned char *in, unsigned char *out,
+			size_t bits, const void *key,
+			unsigned char ivec[16], int *num,
+			int enc, block128_f block);
+
+size_t CRYPTO_cts128_encrypt_block(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block);
+size_t CRYPTO_cts128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], cbc128_f cbc);
+size_t CRYPTO_cts128_decrypt_block(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block);
+size_t CRYPTO_cts128_decrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], cbc128_f cbc);
+
+size_t CRYPTO_nistcts128_encrypt_block(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block);
+size_t CRYPTO_nistcts128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], cbc128_f cbc);
+size_t CRYPTO_nistcts128_decrypt_block(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], block128_f block);
+size_t CRYPTO_nistcts128_decrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], cbc128_f cbc);
+
+typedef struct gcm128_context GCM128_CONTEXT;
+
+GCM128_CONTEXT *CRYPTO_gcm128_new(void *key, block128_f block);
+void CRYPTO_gcm128_init(GCM128_CONTEXT *ctx,void *key,block128_f block);
+void CRYPTO_gcm128_setiv(GCM128_CONTEXT *ctx, const unsigned char *iv,
+			size_t len);
+int CRYPTO_gcm128_aad(GCM128_CONTEXT *ctx, const unsigned char *aad,
+			size_t len);
+int CRYPTO_gcm128_encrypt(GCM128_CONTEXT *ctx,
+			const unsigned char *in, unsigned char *out,
+			size_t len);
+int CRYPTO_gcm128_decrypt(GCM128_CONTEXT *ctx,
+			const unsigned char *in, unsigned char *out,
+			size_t len);
+int CRYPTO_gcm128_encrypt_ctr32(GCM128_CONTEXT *ctx,
+			const unsigned char *in, unsigned char *out,
+			size_t len, ctr128_f stream);
+int CRYPTO_gcm128_decrypt_ctr32(GCM128_CONTEXT *ctx,
+			const unsigned char *in, unsigned char *out,
+			size_t len, ctr128_f stream);
+int CRYPTO_gcm128_finish(GCM128_CONTEXT *ctx,const unsigned char *tag,
+			size_t len);
+void CRYPTO_gcm128_tag(GCM128_CONTEXT *ctx, unsigned char *tag, size_t len);
+void CRYPTO_gcm128_release(GCM128_CONTEXT *ctx);
+
+typedef struct ccm128_context CCM128_CONTEXT;
+
+void CRYPTO_ccm128_init(CCM128_CONTEXT *ctx,
+	unsigned int M, unsigned int L, void *key,block128_f block);
+int CRYPTO_ccm128_setiv(CCM128_CONTEXT *ctx,
+	const unsigned char *nonce, size_t nlen, size_t mlen);
+void CRYPTO_ccm128_aad(CCM128_CONTEXT *ctx,
+	const unsigned char *aad, size_t alen);
+int CRYPTO_ccm128_encrypt(CCM128_CONTEXT *ctx,
+	const unsigned char *inp, unsigned char *out, size_t len);
+int CRYPTO_ccm128_decrypt(CCM128_CONTEXT *ctx,
+	const unsigned char *inp, unsigned char *out, size_t len);
+int CRYPTO_ccm128_encrypt_ccm64(CCM128_CONTEXT *ctx,
+	const unsigned char *inp, unsigned char *out, size_t len,
+	ccm128_f stream);
+int CRYPTO_ccm128_decrypt_ccm64(CCM128_CONTEXT *ctx,
+	const unsigned char *inp, unsigned char *out, size_t len,
+	ccm128_f stream);
+size_t CRYPTO_ccm128_tag(CCM128_CONTEXT *ctx, unsigned char *tag, size_t len);
+
+typedef struct xts128_context XTS128_CONTEXT;
+
+int CRYPTO_xts128_encrypt(const XTS128_CONTEXT *ctx, const unsigned char iv[16],
+	const unsigned char *inp, unsigned char *out, size_t len, int enc);
diff --git a/crypto/modes/modes_lcl.h b/crypto/modes/modes_lcl.h
new file mode 100644
index 000000000000..b6dc3c336fe0
--- /dev/null
+++ b/crypto/modes/modes_lcl.h
@@ -0,0 +1,131 @@
+/* ====================================================================
+ * Copyright (c) 2010 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use is governed by OpenSSL license.
+ * ====================================================================
+ */
+
+#include <openssl/modes.h>
+
+
+#if (defined(_WIN32) || defined(_WIN64)) && !defined(__MINGW32__)
+typedef __int64 i64;
+typedef unsigned __int64 u64;
+#define U64(C) C##UI64
+#elif defined(__arch64__)
+typedef long i64;
+typedef unsigned long u64;
+#define U64(C) C##UL
+#else
+typedef long long i64;
+typedef unsigned long long u64;
+#define U64(C) C##ULL
+#endif
+
+typedef unsigned int u32;
+typedef unsigned char u8;
+
+#define STRICT_ALIGNMENT 1
+#if defined(__i386)	|| defined(__i386__)	|| \
+    defined(__x86_64)	|| defined(__x86_64__)	|| \
+    defined(_M_IX86)	|| defined(_M_AMD64)	|| defined(_M_X64) || \
+    defined(__s390__)	|| defined(__s390x__)	|| \
+    ( (defined(__arm__)	|| defined(__arm)) && \
+      (defined(__ARM_ARCH_7__)	|| defined(__ARM_ARCH_7A__) || \
+       defined(__ARM_ARCH_7R__)	|| defined(__ARM_ARCH_7M__)) )
+# undef STRICT_ALIGNMENT
+#endif
+
+#if !defined(PEDANTIC) && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
+#if defined(__GNUC__) && __GNUC__>=2
+# if defined(__x86_64) || defined(__x86_64__)
+#  define BSWAP8(x) ({	u64 ret=(x);			\
+			asm ("bswapq %0"		\
+			: "+r"(ret));	ret;		})
+#  define BSWAP4(x) ({	u32 ret=(x);			\
+			asm ("bswapl %0"		\
+			: "+r"(ret));	ret;		})
+# elif (defined(__i386) || defined(__i386__)) && !defined(I386_ONLY)
+#  define BSWAP8(x) ({	u32 lo=(u64)(x)>>32,hi=(x);	\
+			asm ("bswapl %0; bswapl %1"	\
+			: "+r"(hi),"+r"(lo));		\
+			(u64)hi<<32|lo;			})
+#  define BSWAP4(x) ({	u32 ret=(x);			\
+			asm ("bswapl %0"		\
+			: "+r"(ret));	ret;		})
+# elif (defined(__arm__) || defined(__arm)) && !defined(STRICT_ALIGNMENT)
+#  define BSWAP8(x) ({	u32 lo=(u64)(x)>>32,hi=(x);	\
+			asm ("rev %0,%0; rev %1,%1"	\
+			: "+r"(hi),"+r"(lo));		\
+			(u64)hi<<32|lo;			})
+#  define BSWAP4(x) ({	u32 ret;			\
+			asm ("rev %0,%1"		\
+			: "=r"(ret) : "r"((u32)(x)));	\
+			ret;				})
+# endif
+#elif defined(_MSC_VER)
+# if _MSC_VER>=1300
+#  pragma intrinsic(_byteswap_uint64,_byteswap_ulong)
+#  define BSWAP8(x)	_byteswap_uint64((u64)(x))
+#  define BSWAP4(x)	_byteswap_ulong((u32)(x))
+# elif defined(_M_IX86)
+   __inline u32 _bswap4(u32 val) {
+	_asm mov eax,val
+	_asm bswap eax
+   }
+#  define BSWAP4(x)	_bswap4(x)
+# endif
+#endif
+#endif
+
+#if defined(BSWAP4) && !defined(STRICT_ALIGNMENT)
+#define GETU32(p)	BSWAP4(*(const u32 *)(p))
+#define PUTU32(p,v)	*(u32 *)(p) = BSWAP4(v)
+#else
+#define GETU32(p)	((u32)(p)[0]<<24|(u32)(p)[1]<<16|(u32)(p)[2]<<8|(u32)(p)[3])
+#define PUTU32(p,v)	((p)[0]=(u8)((v)>>24),(p)[1]=(u8)((v)>>16),(p)[2]=(u8)((v)>>8),(p)[3]=(u8)(v))
+#endif
+
+/* GCM definitions */
+
+typedef struct { u64 hi,lo; } u128;
+
+#ifdef	TABLE_BITS
+#undef	TABLE_BITS
+#endif
+/*
+ * Even though permitted values for TABLE_BITS are 8, 4 and 1, it should
+ * never be set to 8 [or 1]. For further information see gcm128.c.
+ */
+#define	TABLE_BITS 4
+
+struct gcm128_context {
+	/* Following 6 names follow names in GCM specification */
+	union { u64 u[2]; u32 d[4]; u8 c[16]; }	Yi,EKi,EK0,len,
+						Xi,H;
+	/* Relative position of Xi, H and pre-computed Htable is used
+	 * in some assembler modules, i.e. don't change the order! */
+#if TABLE_BITS==8
+	u128 Htable[256];
+#else
+	u128 Htable[16];
+	void (*gmult)(u64 Xi[2],const u128 Htable[16]);
+	void (*ghash)(u64 Xi[2],const u128 Htable[16],const u8 *inp,size_t len);
+#endif
+	unsigned int mres, ares;
+	block128_f block;
+	void *key;
+};
+
+struct xts128_context {
+	void      *key1, *key2;
+	block128_f block1,block2;
+};
+
+struct ccm128_context {
+	union { u64 u[2]; u8 c[16]; } nonce, cmac;
+	u64 blocks;
+	block128_f block;
+	void *key;
+};
+
diff --git a/crypto/modes/ofb128.c b/crypto/modes/ofb128.c
new file mode 100644
index 000000000000..01c01702c4f9
--- /dev/null
+++ b/crypto/modes/ofb128.c
@@ -0,0 +1,121 @@
+/* ====================================================================
+ * Copyright (c) 2008 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ */
+
+#include <openssl/crypto.h>
+#include "modes_lcl.h"
+#include <string.h>
+
+#ifndef MODES_DEBUG
+# ifndef NDEBUG
+#  define NDEBUG
+# endif
+#endif
+#include <assert.h>
+
+/* The input and output encrypted as though 128bit ofb mode is being
+ * used.  The extra state information to record how much of the
+ * 128bit block we have used is contained in *num;
+ */
+void CRYPTO_ofb128_encrypt(const unsigned char *in, unsigned char *out,
+			size_t len, const void *key,
+			unsigned char ivec[16], int *num,
+			block128_f block)
+{
+	unsigned int n;
+	size_t l=0;
+
+	assert(in && out && key && ivec && num);
+
+	n = *num;
+
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
+	if (16%sizeof(size_t) == 0) do { /* always true actually */
+		while (n && len) {
+			*(out++) = *(in++) ^ ivec[n];
+			--len;
+			n = (n+1) % 16;
+		}
+#if defined(STRICT_ALIGNMENT)
+		if (((size_t)in|(size_t)out|(size_t)ivec)%sizeof(size_t) != 0)
+			break;
+#endif
+		while (len>=16) {
+			(*block)(ivec, ivec, key);
+			for (; n<16; n+=sizeof(size_t))
+				*(size_t*)(out+n) =
+				*(size_t*)(in+n) ^ *(size_t*)(ivec+n);
+			len -= 16;
+			out += 16;
+			in  += 16;
+			n = 0;
+		}
+		if (len) {
+			(*block)(ivec, ivec, key);
+			while (len--) {
+				out[n] = in[n] ^ ivec[n];
+				++n;
+			}
+		}
+		*num = n;
+		return;
+	} while(0);
+	/* the rest would be commonly eliminated by x86* compiler */
+#endif
+	while (l<len) {
+		if (n==0) {
+			(*block)(ivec, ivec, key);
+		}
+		out[l] = in[l] ^ ivec[n];
+		++l;
+		n = (n+1) % 16;
+	}
+
+	*num=n;
+}
diff --git a/crypto/modes/xts128.c b/crypto/modes/xts128.c
new file mode 100644
index 000000000000..9cf27a25e960
--- /dev/null
+++ b/crypto/modes/xts128.c
@@ -0,0 +1,187 @@
+/* ====================================================================
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ */
+
+#include <openssl/crypto.h>
+#include "modes_lcl.h"
+#include <string.h>
+
+#ifndef MODES_DEBUG
+# ifndef NDEBUG
+#  define NDEBUG
+# endif
+#endif
+#include <assert.h>
+
+int CRYPTO_xts128_encrypt(const XTS128_CONTEXT *ctx, const unsigned char iv[16],
+	const unsigned char *inp, unsigned char *out,
+	size_t len, int enc)
+{
+	const union { long one; char little; } is_endian = {1};
+	union { u64 u[2]; u32 d[4]; u8 c[16]; } tweak, scratch;
+	unsigned int i;
+
+	if (len<16) return -1;
+
+	memcpy(tweak.c, iv, 16);
+
+	(*ctx->block2)(tweak.c,tweak.c,ctx->key2);
+
+	if (!enc && (len%16)) len-=16;
+
+	while (len>=16) {
+#if defined(STRICT_ALIGNMENT)
+		memcpy(scratch.c,inp,16);
+		scratch.u[0] ^= tweak.u[0];
+		scratch.u[1] ^= tweak.u[1];
+#else
+		scratch.u[0] = ((u64*)inp)[0]^tweak.u[0];
+		scratch.u[1] = ((u64*)inp)[1]^tweak.u[1];
+#endif
+		(*ctx->block1)(scratch.c,scratch.c,ctx->key1);
+#if defined(STRICT_ALIGNMENT)
+		scratch.u[0] ^= tweak.u[0];
+		scratch.u[1] ^= tweak.u[1];
+		memcpy(out,scratch.c,16);
+#else
+		((u64*)out)[0] = scratch.u[0]^=tweak.u[0];
+		((u64*)out)[1] = scratch.u[1]^=tweak.u[1];
+#endif
+		inp += 16;
+		out += 16;
+		len -= 16;
+
+		if (len==0)	return 0;
+
+		if (is_endian.little) {
+			unsigned int carry,res;
+			
+			res = 0x87&(((int)tweak.d[3])>>31);
+			carry = (unsigned int)(tweak.u[0]>>63);
+			tweak.u[0] = (tweak.u[0]<<1)^res;
+			tweak.u[1] = (tweak.u[1]<<1)|carry;
+		}
+		else {
+			size_t c;
+
+			for (c=0,i=0;i<16;++i) {
+				/*+ substitutes for |, because c is 1 bit */ 
+				c += ((size_t)tweak.c[i])<<1;
+				tweak.c[i] = (u8)c;
+				c = c>>8;
+			}
+			tweak.c[0] ^= (u8)(0x87&(0-c));
+		}
+	}
+	if (enc) {
+		for (i=0;i<len;++i) {
+			u8 c = inp[i];
+			out[i] = scratch.c[i];
+			scratch.c[i] = c;
+		}
+		scratch.u[0] ^= tweak.u[0];
+		scratch.u[1] ^= tweak.u[1];
+		(*ctx->block1)(scratch.c,scratch.c,ctx->key1);
+		scratch.u[0] ^= tweak.u[0];
+		scratch.u[1] ^= tweak.u[1];
+		memcpy(out-16,scratch.c,16);
+	}
+	else {
+		union { u64 u[2]; u8 c[16]; } tweak1;
+
+		if (is_endian.little) {
+			unsigned int carry,res;
+
+			res = 0x87&(((int)tweak.d[3])>>31);
+			carry = (unsigned int)(tweak.u[0]>>63);
+			tweak1.u[0] = (tweak.u[0]<<1)^res;
+			tweak1.u[1] = (tweak.u[1]<<1)|carry;
+		}
+		else {
+			size_t c;
+
+			for (c=0,i=0;i<16;++i) {
+				/*+ substitutes for |, because c is 1 bit */ 
+				c += ((size_t)tweak.c[i])<<1;
+				tweak1.c[i] = (u8)c;
+				c = c>>8;
+			}
+			tweak1.c[0] ^= (u8)(0x87&(0-c));
+		}
+#if defined(STRICT_ALIGNMENT)
+		memcpy(scratch.c,inp,16);
+		scratch.u[0] ^= tweak1.u[0];
+		scratch.u[1] ^= tweak1.u[1];
+#else
+		scratch.u[0] = ((u64*)inp)[0]^tweak1.u[0];
+		scratch.u[1] = ((u64*)inp)[1]^tweak1.u[1];
+#endif
+		(*ctx->block1)(scratch.c,scratch.c,ctx->key1);
+		scratch.u[0] ^= tweak1.u[0];
+		scratch.u[1] ^= tweak1.u[1];
+
+		for (i=0;i<len;++i) {
+			u8 c = inp[16+i];
+			out[16+i] = scratch.c[i];
+			scratch.c[i] = c;
+		}
+		scratch.u[0] ^= tweak.u[0];
+		scratch.u[1] ^= tweak.u[1];
+		(*ctx->block1)(scratch.c,scratch.c,ctx->key1);
+#if defined(STRICT_ALIGNMENT)
+		scratch.u[0] ^= tweak.u[0];
+		scratch.u[1] ^= tweak.u[1];
+		memcpy (out,scratch.c,16);
+#else
+		((u64*)out)[0] = scratch.u[0]^tweak.u[0];
+		((u64*)out)[1] = scratch.u[1]^tweak.u[1];
+#endif
+	}
+
+	return 0;
+}
diff --git a/crypto/rand/rand_vms.c b/crypto/o_fips.c
index 1267a3acae7c..f6d1b21855ce 100644
--- a/crypto/rand/rand_vms.c
+++ b/crypto/o_fips.c
@@ -1,9 +1,8 @@
-/* crypto/rand/rand_vms.c -*- mode:C; c-file-style: "eay" -*- */
-/* Written by Richard Levitte <richard@levitte.org> for the OpenSSL
- * project 2000.
+/* Written by Stephen henson (steve@openssl.org) for the OpenSSL
+ * project 2011.
  */
 /* ====================================================================
- * Copyright (c) 1998-2000 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -56,81 +55,42 @@
  *
  */
 
+#include "cryptlib.h"
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#include <openssl/fips_rand.h>
 #include <openssl/rand.h>
-#include "rand_lcl.h"
-
-#if defined(OPENSSL_SYS_VMS)
-
-#include <descrip.h>
-#include <jpidef.h>
-#include <ssdef.h>
-#include <starlet.h>
-#ifdef __DECC
-# pragma message disable DOLLARID
 #endif
 
-static struct items_data_st
-	{
-	short length, code;	/* length is amount of bytes */
-	} items_data[] =
-		{ { 4, JPI$_BUFIO },
-		  { 4, JPI$_CPUTIM },
-		  { 4, JPI$_DIRIO },
-		  { 8, JPI$_LOGINTIM },
-		  { 4, JPI$_PAGEFLTS },
-		  { 4, JPI$_PID },
-		  { 4, JPI$_WSSIZE },
-		  { 0, 0 }
-		};
-		  
-int RAND_poll(void)
+int FIPS_mode(void)
 	{
-	long pid, iosb[2];
-	int status = 0;
-	struct
-		{
-		short length, code;
-		long *buffer;
-		int *retlen;
-		} item[32], *pitem;
-	unsigned char data_buffer[256];
-	short total_length = 0;
-	struct items_data_st *pitems_data;
-
-	pitems_data = items_data;
-	pitem = item;
-
-	/* Setup */
-	while (pitems_data->length
-		&& (total_length + pitems_data->length <= 256))
-		{
-		pitem->length = pitems_data->length;
-		pitem->code = pitems_data->code;
-		pitem->buffer = (long *)&data_buffer[total_length];
-		pitem->retlen = 0;
-		total_length += pitems_data->length;
-		pitems_data++;
-		pitem++;
-		}
-	pitem->length = pitem->code = 0;
+	OPENSSL_init();
+#ifdef OPENSSL_FIPS
+	return FIPS_module_mode();
+#else
+	return 0;
+#endif
+	}
 
-	/*
-	 * Scan through all the processes in the system and add entropy with
-	 * results from the processes that were possible to look at.
-	 * However, view the information as only half trustable.
-	 */
-	pid = -1;			/* search context */
-	while ((status = sys$getjpiw(0, &pid,  0, item, iosb, 0, 0))
-		!= SS$_NOMOREPROC)
-		{
-		if (status == SS$_NORMAL)
-			{
-			RAND_add(data_buffer, total_length, total_length/2);
-			}
-		}
-	sys$gettim(iosb);
-	RAND_add((unsigned char *)iosb, sizeof(iosb), sizeof(iosb)/2);
+int FIPS_mode_set(int r)
+	{
+	OPENSSL_init();
+#ifdef OPENSSL_FIPS
+#ifndef FIPS_AUTH_USER_PASS
+#define FIPS_AUTH_USER_PASS	"Default FIPS Crypto User Password"
+#endif
+	if (!FIPS_module_mode_set(r, FIPS_AUTH_USER_PASS))
+		return 0;
+	if (r)
+		RAND_set_rand_method(FIPS_rand_get_method());
+	else
+		RAND_set_rand_method(NULL);
 	return 1;
-}
-
+#else
+	if (r == 0)
+		return 1;
+	CRYPTOerr(CRYPTO_F_FIPS_MODE_SET, CRYPTO_R_FIPS_MODE_NOT_SUPPORTED);
+	return 0;
 #endif
+	}
+
diff --git a/crypto/o_init.c b/crypto/o_init.c
index d767a90a5f43..db4cdc443b8f 100644
--- a/crypto/o_init.c
+++ b/crypto/o_init.c
@@ -3,7 +3,7 @@
  * project.
  */
 /* ====================================================================
- * Copyright (c) 2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -50,23 +50,13 @@
  * OF THE POSSIBILITY OF SUCH DAMAGE.
  * ====================================================================
  *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
  */
 
 #include <e_os.h>
 #include <openssl/err.h>
-
-/* Internal only functions: only ever used here */
 #ifdef OPENSSL_FIPS
-extern	void int_ERR_lib_init(void);
-# ifndef OPENSSL_NO_ENGINE
-extern	void int_EVP_MD_init_engine_callbacks(void );
-extern	void int_EVP_CIPHER_init_engine_callbacks(void );
-extern	void int_RAND_init_engine_callbacks(void );
-# endif
+#include <openssl/fips.h>
+#include <openssl/rand.h>
 #endif
 
 /* Perform any essential OpenSSL initialization operations.
@@ -75,22 +65,18 @@ extern	void int_RAND_init_engine_callbacks(void );
 
 void OPENSSL_init(void)
 	{
-#ifdef OPENSSL_FIPS
 	static int done = 0;
-	if (!done)
-		{
-		int_ERR_lib_init();
-#ifdef CRYPTO_MDEBUG
-		CRYPTO_malloc_debug_init();
-#endif
-#ifndef OPENSSL_NO_ENGINE
-		int_EVP_MD_init_engine_callbacks();
-		int_EVP_CIPHER_init_engine_callbacks();
-		int_RAND_init_engine_callbacks();
+	if (done)
+		return;
+	done = 1;
+#ifdef OPENSSL_FIPS
+	FIPS_set_locking_callbacks(CRYPTO_lock, CRYPTO_add_lock);
+	FIPS_set_error_callbacks(ERR_put_error, ERR_add_error_vdata);
+	FIPS_set_malloc_callbacks(CRYPTO_malloc, CRYPTO_free);
+	RAND_init_fips();
 #endif
-		done = 1;
-		}
+#if 0
+	fprintf(stderr, "Called OPENSSL_init\n");
 #endif
 	}
-		
 
diff --git a/crypto/o_time.c b/crypto/o_time.c
index e29091d6504c..9030fdef7aed 100644
--- a/crypto/o_time.c
+++ b/crypto/o_time.c
@@ -2,6 +2,9 @@
 /* Written by Richard Levitte (richard@levitte.org) for the OpenSSL
  * project 2001.
  */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2008.
+ */
 /* ====================================================================
  * Copyright (c) 2001 The OpenSSL Project.  All rights reserved.
  *
@@ -61,24 +64,30 @@
 #include "o_time.h"
 
 #ifdef OPENSSL_SYS_VMS
-# include <libdtdef.h>
-# include <lib$routines.h>
-# include <lnmdef.h>
-# include <starlet.h>
-# include <descrip.h>
-# include <stdlib.h>
+# if __CRTL_VER >= 70000000 && \
+     (defined _POSIX_C_SOURCE || !defined _ANSI_C_SOURCE)
+#  define VMS_GMTIME_OK
+# endif
+# ifndef VMS_GMTIME_OK
+#  include <libdtdef.h>
+#  include <lib$routines.h>
+#  include <lnmdef.h>
+#  include <starlet.h>
+#  include <descrip.h>
+#  include <stdlib.h>
+# endif /* ndef VMS_GMTIME_OK */
 #endif
 
 struct tm *OPENSSL_gmtime(const time_t *timer, struct tm *result)
 	{
 	struct tm *ts = NULL;
 
-#if defined(OPENSSL_THREADS) && !defined(OPENSSL_SYS_WIN32) && !defined(OPENSSL_SYS_OS2) && !defined(__CYGWIN32__) && (!defined(OPENSSL_SYS_VMS) || defined(gmtime_r)) && !defined(OPENSSL_SYS_MACOSX) && !defined(OPENSSL_SYS_SUNOS)
+#if defined(OPENSSL_THREADS) && !defined(OPENSSL_SYS_WIN32) && !defined(OPENSSL_SYS_OS2) && (!defined(OPENSSL_SYS_VMS) || defined(gmtime_r)) && !defined(OPENSSL_SYS_MACOSX) && !defined(OPENSSL_SYS_SUNOS)
 	/* should return &data, but doesn't on some systems,
 	   so we don't even look at the return value */
 	gmtime_r(timer,result);
 	ts = result;
-#elif !defined(OPENSSL_SYS_VMS)
+#elif !defined(OPENSSL_SYS_VMS) || defined(VMS_GMTIME_OK)
 	ts = gmtime(timer);
 	if (ts == NULL)
 		return NULL;
@@ -86,7 +95,7 @@ struct tm *OPENSSL_gmtime(const time_t *timer, struct tm *result)
 	memcpy(result, ts, sizeof(struct tm));
 	ts = result;
 #endif
-#ifdef OPENSSL_SYS_VMS
+#if defined( OPENSSL_SYS_VMS) && !defined( VMS_GMTIME_OK)
 	if (ts == NULL)
 		{
 		static $DESCRIPTOR(tabnam,"LNM$DCL_LOGICAL");
@@ -214,4 +223,150 @@ struct tm *OPENSSL_gmtime(const time_t *timer, struct tm *result)
 		}
 #endif
 	return ts;
-	}	
+	}
+
+/* Take a tm structure and add an offset to it. This avoids any OS issues
+ * with restricted date types and overflows which cause the year 2038
+ * problem.
+ */
+
+#define SECS_PER_DAY (24 * 60 * 60)
+
+static long date_to_julian(int y, int m, int d);
+static void julian_to_date(long jd, int *y, int *m, int *d);
+
+int OPENSSL_gmtime_adj(struct tm *tm, int off_day, long offset_sec)
+	{
+	int offset_hms, offset_day;
+	long time_jd;
+	int time_year, time_month, time_day;
+	/* split offset into days and day seconds */
+	offset_day = offset_sec / SECS_PER_DAY;
+	/* Avoid sign issues with % operator */
+	offset_hms  = offset_sec - (offset_day * SECS_PER_DAY);
+	offset_day += off_day;
+	/* Add current time seconds to offset */
+	offset_hms += tm->tm_hour * 3600 + tm->tm_min * 60 + tm->tm_sec;
+	/* Adjust day seconds if overflow */
+	if (offset_hms >= SECS_PER_DAY)
+		{
+		offset_day++;
+		offset_hms -= SECS_PER_DAY;
+		}
+	else if (offset_hms < 0)
+		{
+		offset_day--;
+		offset_hms += SECS_PER_DAY;
+		}
+
+	/* Convert date of time structure into a Julian day number.
+	 */
+
+	time_year = tm->tm_year + 1900;
+	time_month = tm->tm_mon + 1;
+	time_day = tm->tm_mday;
+
+	time_jd = date_to_julian(time_year, time_month, time_day);
+
+	/* Work out Julian day of new date */
+	time_jd += offset_day;
+
+	if (time_jd < 0)
+		return 0;
+
+	/* Convert Julian day back to date */
+
+	julian_to_date(time_jd, &time_year, &time_month, &time_day);
+
+	if (time_year < 1900 || time_year > 9999)
+		return 0;
+
+	/* Update tm structure */
+
+	tm->tm_year = time_year - 1900;
+	tm->tm_mon = time_month - 1;
+	tm->tm_mday = time_day;
+
+	tm->tm_hour = offset_hms / 3600;
+	tm->tm_min = (offset_hms / 60) % 60;
+	tm->tm_sec = offset_hms % 60;
+
+	return 1;
+		
+}
+
+/* Convert date to and from julian day
+ * Uses Fliegel & Van Flandern algorithm
+ */
+static long date_to_julian(int y, int m, int d)
+{
+	return (1461 * (y + 4800 + (m - 14) / 12)) / 4 +
+		(367 * (m - 2 - 12 * ((m - 14) / 12))) / 12 -
+		(3 * ((y + 4900 + (m - 14) / 12) / 100)) / 4 +
+		d - 32075;
+}
+
+static void julian_to_date(long jd, int *y, int *m, int *d)
+	{
+	long  L = jd + 68569;
+	long  n = (4 * L) / 146097;
+	long  i, j;
+
+	L = L - (146097 * n + 3) / 4;
+	i = (4000 * (L + 1)) / 1461001;
+	L = L - (1461 * i) / 4 + 31;
+	j = (80 * L) / 2447;
+	*d = L - (2447 * j) / 80;
+	L = j / 11;
+	*m = j + 2 - (12 * L);
+	*y = 100 * (n - 49) + i + L;
+	}
+
+#ifdef OPENSSL_TIME_TEST
+
+#include <stdio.h>
+
+/* Time checking test code. Check times are identical for a wide range of
+ * offsets. This should be run on a machine with 64 bit time_t or it will
+ * trigger the very errors the routines fix.
+ */
+
+int main(int argc, char **argv)
+	{
+	long offset;
+	for (offset = 0; offset < 1000000; offset++)
+		{
+		check_time(offset);
+		check_time(-offset);
+		check_time(offset * 1000);
+		check_time(-offset * 1000);
+		}
+	}
+
+int check_time(long offset)
+	{
+	struct tm tm1, tm2;
+	time_t t1, t2;
+	time(&t1);
+	t2 = t1 + offset;
+	OPENSSL_gmtime(&t2, &tm2);
+	OPENSSL_gmtime(&t1, &tm1);
+	OPENSSL_gmtime_adj(&tm1, 0, offset);
+	if ((tm1.tm_year == tm2.tm_year) &&
+	    (tm1.tm_mon == tm2.tm_mon) &&
+	    (tm1.tm_mday == tm2.tm_mday) &&
+	    (tm1.tm_hour == tm2.tm_hour) &&
+	    (tm1.tm_min == tm2.tm_min) &&
+	    (tm1.tm_sec == tm2.tm_sec))
+		return 1;
+	fprintf(stderr, "TIME ERROR!!\n");
+	fprintf(stderr, "Time1: %d/%d/%d, %d:%02d:%02d\n",
+			tm2.tm_mday, tm2.tm_mon + 1, tm2.tm_year + 1900,
+			tm2.tm_hour, tm2.tm_min, tm2.tm_sec);
+	fprintf(stderr, "Time2: %d/%d/%d, %d:%02d:%02d\n",
+			tm1.tm_mday, tm1.tm_mon + 1, tm1.tm_year + 1900,
+			tm1.tm_hour, tm1.tm_min, tm1.tm_sec);
+	return 0;
+	}
+
+#endif
diff --git a/crypto/o_time.h b/crypto/o_time.h
index e66044626d1f..e391da75087e 100644
--- a/crypto/o_time.h
+++ b/crypto/o_time.h
@@ -62,5 +62,6 @@
 #include <time.h>
 
 struct tm *OPENSSL_gmtime(const time_t *timer, struct tm *result);
+int OPENSSL_gmtime_adj(struct tm *tm, int offset_day, long offset_sec);
 
 #endif
diff --git a/crypto/objects/Makefile b/crypto/objects/Makefile
index 25e8b23b5d05..a8aedbd42287 100644
--- a/crypto/objects/Makefile
+++ b/crypto/objects/Makefile
@@ -18,23 +18,23 @@ TEST=
 APPS=
 
 LIB=$(TOP)/libcrypto.a
-LIBSRC=	o_names.c obj_dat.c obj_lib.c obj_err.c
-LIBOBJ= o_names.o obj_dat.o obj_lib.o obj_err.o
+LIBSRC=	o_names.c obj_dat.c obj_lib.c obj_err.c obj_xref.c
+LIBOBJ= o_names.o obj_dat.o obj_lib.o obj_err.o obj_xref.o
 
 SRC= $(LIBSRC)
 
 EXHEADER= objects.h obj_mac.h
-HEADER=	$(EXHEADER) obj_dat.h
+HEADER=	$(EXHEADER) obj_dat.h obj_xref.h
 
 ALL=    $(GENERAL) $(SRC) $(HEADER)
 
 top:
 	(cd ../..; $(MAKE) DIRS=crypto SDIRS=$(DIR) sub_all)
 
-all:	obj_dat.h lib
+all:	obj_dat.h obj_xref.h lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -46,6 +46,10 @@ obj_mac.h: objects.pl objects.txt obj_mac.num
 	$(PERL) objects.pl objects.txt obj_mac.num obj_mac.h
 	@sleep 1; touch obj_mac.h; sleep 1
 
+obj_xref.h: objxref.pl obj_xref.txt obj_mac.num
+	$(PERL) objxref.pl obj_mac.num obj_xref.txt > obj_xref.h
+	@sleep 1; touch obj_xref.h; sleep 1
+
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
 
@@ -117,3 +121,10 @@ obj_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
 obj_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
 obj_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 obj_lib.o: ../cryptlib.h obj_lib.c
+obj_xref.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+obj_xref.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+obj_xref.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+obj_xref.o: ../../include/openssl/opensslconf.h
+obj_xref.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+obj_xref.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+obj_xref.o: ../../include/openssl/symhacks.h obj_xref.c obj_xref.h
diff --git a/crypto/objects/o_names.c b/crypto/objects/o_names.c
index adb5731f7659..84380a96a98c 100644
--- a/crypto/objects/o_names.c
+++ b/crypto/objects/o_names.c
@@ -22,7 +22,8 @@
 /* I use the ex_data stuff to manage the identifiers for the obj_name_types
  * that applications may define.  I only really use the free function field.
  */
-static LHASH *names_lh=NULL;
+DECLARE_LHASH_OF(OBJ_NAME);
+static LHASH_OF(OBJ_NAME) *names_lh=NULL;
 static int names_type_num=OBJ_NAME_TYPE_NUM;
 
 typedef struct name_funcs_st
@@ -46,11 +47,14 @@ static unsigned long obj_name_hash(const void *a_void);
 /* static int obj_name_cmp(OBJ_NAME *a,OBJ_NAME *b); */
 static int obj_name_cmp(const void *a_void,const void *b_void);
 
+static IMPLEMENT_LHASH_HASH_FN(obj_name, OBJ_NAME)
+static IMPLEMENT_LHASH_COMP_FN(obj_name, OBJ_NAME)
+
 int OBJ_NAME_init(void)
 	{
 	if (names_lh != NULL) return(1);
 	MemCheck_off();
-	names_lh=lh_new(obj_name_hash, obj_name_cmp);
+	names_lh=lh_OBJ_NAME_new();
 	MemCheck_on();
 	return(names_lh != NULL);
 	}
@@ -164,7 +168,7 @@ const char *OBJ_NAME_get(const char *name, int type)
 
 	for (;;)
 	{
-		ret=(OBJ_NAME *)lh_retrieve(names_lh,&on);
+		ret=lh_OBJ_NAME_retrieve(names_lh,&on);
 		if (ret == NULL) return(NULL);
 		if ((ret->alias) && !alias)
 			{
@@ -200,7 +204,7 @@ int OBJ_NAME_add(const char *name, int type, const char *data)
 	onp->type=type;
 	onp->data=data;
 
-	ret=(OBJ_NAME *)lh_insert(names_lh,onp);
+	ret=lh_OBJ_NAME_insert(names_lh,onp);
 	if (ret != NULL)
 		{
 		/* free things */
@@ -217,7 +221,7 @@ int OBJ_NAME_add(const char *name, int type, const char *data)
 		}
 	else
 		{
-		if (lh_error(names_lh))
+		if (lh_OBJ_NAME_error(names_lh))
 			{
 			/* ERROR */
 			return(0);
@@ -235,7 +239,7 @@ int OBJ_NAME_remove(const char *name, int type)
 	type&= ~OBJ_NAME_ALIAS;
 	on.name=name;
 	on.type=type;
-	ret=(OBJ_NAME *)lh_delete(names_lh,&on);
+	ret=lh_OBJ_NAME_delete(names_lh,&on);
 	if (ret != NULL)
 		{
 		/* free things */
@@ -262,13 +266,13 @@ struct doall
 	void *arg;
 	};
 
-static void do_all_fn(const OBJ_NAME *name,struct doall *d)
+static void do_all_fn_doall_arg(const OBJ_NAME *name,struct doall *d)
 	{
 	if(name->type == d->type)
 		d->fn(name,d->arg);
 	}
 
-static IMPLEMENT_LHASH_DOALL_ARG_FN(do_all_fn, const OBJ_NAME *, struct doall *)
+static IMPLEMENT_LHASH_DOALL_ARG_FN(do_all_fn, const OBJ_NAME, struct doall)
 
 void OBJ_NAME_do_all(int type,void (*fn)(const OBJ_NAME *,void *arg),void *arg)
 	{
@@ -278,7 +282,8 @@ void OBJ_NAME_do_all(int type,void (*fn)(const OBJ_NAME *,void *arg),void *arg)
 	d.fn=fn;
 	d.arg=arg;
 
-	lh_doall_arg(names_lh,LHASH_DOALL_ARG_FN(do_all_fn),&d);
+	lh_OBJ_NAME_doall_arg(names_lh, LHASH_DOALL_ARG_FN(do_all_fn),
+			      struct doall, &d);
 	}
 
 struct doall_sorted
@@ -313,7 +318,7 @@ void OBJ_NAME_do_all_sorted(int type,void (*fn)(const OBJ_NAME *,void *arg),
 	int n;
 
 	d.type=type;
-	d.names=OPENSSL_malloc(lh_num_items(names_lh)*sizeof *d.names);
+	d.names=OPENSSL_malloc(lh_OBJ_NAME_num_items(names_lh)*sizeof *d.names);
 	d.n=0;
 	OBJ_NAME_do_all(type,do_all_sorted_fn,&d);
 
@@ -327,18 +332,16 @@ void OBJ_NAME_do_all_sorted(int type,void (*fn)(const OBJ_NAME *,void *arg),
 
 static int free_type;
 
-static void names_lh_free(OBJ_NAME *onp)
-{
-	if(onp == NULL)
+static void names_lh_free_doall(OBJ_NAME *onp)
+	{
+	if (onp == NULL)
 		return;
 
-	if ((free_type < 0) || (free_type == onp->type))
-		{
+	if (free_type < 0 || free_type == onp->type)
 		OBJ_NAME_remove(onp->name,onp->type);
-		}
 	}
 
-static IMPLEMENT_LHASH_DOALL_FN(names_lh_free, OBJ_NAME *)
+static IMPLEMENT_LHASH_DOALL_FN(names_lh_free, OBJ_NAME)
 
 static void name_funcs_free(NAME_FUNCS *ptr)
 	{
@@ -352,18 +355,18 @@ void OBJ_NAME_cleanup(int type)
 	if (names_lh == NULL) return;
 
 	free_type=type;
-	down_load=names_lh->down_load;
-	names_lh->down_load=0;
+	down_load=lh_OBJ_NAME_down_load(names_lh);
+	lh_OBJ_NAME_down_load(names_lh)=0;
 
-	lh_doall(names_lh,LHASH_DOALL_FN(names_lh_free));
+	lh_OBJ_NAME_doall(names_lh,LHASH_DOALL_FN(names_lh_free));
 	if (type < 0)
 		{
-		lh_free(names_lh);
+		lh_OBJ_NAME_free(names_lh);
 		sk_NAME_FUNCS_pop_free(name_funcs_stack,name_funcs_free);
 		names_lh=NULL;
 		name_funcs_stack = NULL;
 		}
 	else
-		names_lh->down_load=down_load;
+		lh_OBJ_NAME_down_load(names_lh)=down_load;
 	}
 
diff --git a/crypto/objects/obj_dat.c b/crypto/objects/obj_dat.c
index 760af16dd9a7..8a342ba3eb3e 100644
--- a/crypto/objects/obj_dat.c
+++ b/crypto/objects/obj_dat.c
@@ -74,16 +74,17 @@
 #define NUM_SN 0
 #define NUM_LN 0
 #define NUM_OBJ 0
-static unsigned char lvalues[1];
-static ASN1_OBJECT nid_objs[1];
-static ASN1_OBJECT *sn_objs[1];
-static ASN1_OBJECT *ln_objs[1];
-static ASN1_OBJECT *obj_objs[1];
+static const unsigned char lvalues[1];
+static const ASN1_OBJECT nid_objs[1];
+static const unsigned int sn_objs[1];
+static const unsigned int ln_objs[1];
+static const unsigned int obj_objs[1];
 #endif
 
-static int sn_cmp(const void *a, const void *b);
-static int ln_cmp(const void *a, const void *b);
-static int obj_cmp(const void *a, const void *b);
+DECLARE_OBJ_BSEARCH_CMP_FN(const ASN1_OBJECT *, unsigned int, sn);
+DECLARE_OBJ_BSEARCH_CMP_FN(const ASN1_OBJECT *, unsigned int, ln);
+DECLARE_OBJ_BSEARCH_CMP_FN(const ASN1_OBJECT *, unsigned int, obj);
+
 #define ADDED_DATA	0
 #define ADDED_SNAME	1
 #define ADDED_LNAME	2
@@ -94,30 +95,27 @@ typedef struct added_obj_st
 	int type;
 	ASN1_OBJECT *obj;
 	} ADDED_OBJ;
+DECLARE_LHASH_OF(ADDED_OBJ);
 
 static int new_nid=NUM_NID;
-static LHASH *added=NULL;
+static LHASH_OF(ADDED_OBJ) *added=NULL;
 
-static int sn_cmp(const void *a, const void *b)
-	{
-	const ASN1_OBJECT * const *ap = a, * const *bp = b;
-	return(strcmp((*ap)->sn,(*bp)->sn));
-	}
+static int sn_cmp(const ASN1_OBJECT * const *a, const unsigned int *b)
+	{ return(strcmp((*a)->sn,nid_objs[*b].sn)); }
 
-static int ln_cmp(const void *a, const void *b)
-	{ 
-	const ASN1_OBJECT * const *ap = a, * const *bp = b;
-	return(strcmp((*ap)->ln,(*bp)->ln));
-	}
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(const ASN1_OBJECT *, unsigned int, sn);
 
-/* static unsigned long add_hash(ADDED_OBJ *ca) */
-static unsigned long add_hash(const void *ca_void)
+static int ln_cmp(const ASN1_OBJECT * const *a, const unsigned int *b)
+	{ return(strcmp((*a)->ln,nid_objs[*b].ln)); }
+
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(const ASN1_OBJECT *, unsigned int, ln);
+
+static unsigned long added_obj_hash(const ADDED_OBJ *ca)
 	{
 	const ASN1_OBJECT *a;
 	int i;
 	unsigned long ret=0;
 	unsigned char *p;
-	const ADDED_OBJ *ca = (const ADDED_OBJ *)ca_void;
 
 	a=ca->obj;
 	switch (ca->type)
@@ -145,14 +143,12 @@ static unsigned long add_hash(const void *ca_void)
 	ret|=ca->type<<30L;
 	return(ret);
 	}
+static IMPLEMENT_LHASH_HASH_FN(added_obj, ADDED_OBJ)
 
-/* static int add_cmp(ADDED_OBJ *ca, ADDED_OBJ *cb) */
-static int add_cmp(const void *ca_void, const void *cb_void)
+static int added_obj_cmp(const ADDED_OBJ *ca, const ADDED_OBJ *cb)
 	{
 	ASN1_OBJECT *a,*b;
 	int i;
-	const ADDED_OBJ *ca = (const ADDED_OBJ *)ca_void;
-	const ADDED_OBJ *cb = (const ADDED_OBJ *)cb_void;
 
 	i=ca->type-cb->type;
 	if (i) return(i);
@@ -179,15 +175,16 @@ static int add_cmp(const void *ca_void, const void *cb_void)
 		return 0;
 		}
 	}
+static IMPLEMENT_LHASH_COMP_FN(added_obj, ADDED_OBJ)
 
 static int init_added(void)
 	{
 	if (added != NULL) return(1);
-	added=lh_new(add_hash,add_cmp);
+	added=lh_ADDED_OBJ_new();
 	return(added != NULL);
 	}
 
-static void cleanup1(ADDED_OBJ *a)
+static void cleanup1_doall(ADDED_OBJ *a)
 	{
 	a->obj->nid=0;
 	a->obj->flags|=ASN1_OBJECT_FLAG_DYNAMIC|
@@ -195,28 +192,46 @@ static void cleanup1(ADDED_OBJ *a)
 			ASN1_OBJECT_FLAG_DYNAMIC_DATA;
 	}
 
-static void cleanup2(ADDED_OBJ *a)
+static void cleanup2_doall(ADDED_OBJ *a)
 	{ a->obj->nid++; }
 
-static void cleanup3(ADDED_OBJ *a)
+static void cleanup3_doall(ADDED_OBJ *a)
 	{
 	if (--a->obj->nid == 0)
 		ASN1_OBJECT_free(a->obj);
 	OPENSSL_free(a);
 	}
 
-static IMPLEMENT_LHASH_DOALL_FN(cleanup1, ADDED_OBJ *)
-static IMPLEMENT_LHASH_DOALL_FN(cleanup2, ADDED_OBJ *)
-static IMPLEMENT_LHASH_DOALL_FN(cleanup3, ADDED_OBJ *)
+static IMPLEMENT_LHASH_DOALL_FN(cleanup1, ADDED_OBJ)
+static IMPLEMENT_LHASH_DOALL_FN(cleanup2, ADDED_OBJ)
+static IMPLEMENT_LHASH_DOALL_FN(cleanup3, ADDED_OBJ)
+
+/* The purpose of obj_cleanup_defer is to avoid EVP_cleanup() attempting
+ * to use freed up OIDs. If neccessary the actual freeing up of OIDs is
+ * delayed.
+ */
+
+int obj_cleanup_defer = 0;
+
+void check_defer(int nid)
+	{
+	if (!obj_cleanup_defer && nid >= NUM_NID)
+			obj_cleanup_defer = 1;
+	}
 
 void OBJ_cleanup(void)
 	{
+	if (obj_cleanup_defer)
+		{
+		obj_cleanup_defer = 2;
+		return ;
+		}
 	if (added == NULL) return;
-	added->down_load=0;
-	lh_doall(added,LHASH_DOALL_FN(cleanup1)); /* zero counters */
-	lh_doall(added,LHASH_DOALL_FN(cleanup2)); /* set counters */
-	lh_doall(added,LHASH_DOALL_FN(cleanup3)); /* free objects */
-	lh_free(added);
+	lh_ADDED_OBJ_down_load(added) = 0;
+	lh_ADDED_OBJ_doall(added,LHASH_DOALL_FN(cleanup1)); /* zero counters */
+	lh_ADDED_OBJ_doall(added,LHASH_DOALL_FN(cleanup2)); /* set counters */
+	lh_ADDED_OBJ_doall(added,LHASH_DOALL_FN(cleanup3)); /* free objects */
+	lh_ADDED_OBJ_free(added);
 	added=NULL;
 	}
 
@@ -252,7 +267,7 @@ int OBJ_add_object(const ASN1_OBJECT *obj)
 			{
 			ao[i]->type=i;
 			ao[i]->obj=o;
-			aop=(ADDED_OBJ *)lh_insert(added,ao[i]);
+			aop=lh_ADDED_OBJ_insert(added,ao[i]);
 			/* memory leak, buit should not normally matter */
 			if (aop != NULL)
 				OPENSSL_free(aop);
@@ -292,7 +307,7 @@ ASN1_OBJECT *OBJ_nid2obj(int n)
 		ad.type=ADDED_NID;
 		ad.obj= &ob;
 		ob.nid=n;
-		adp=(ADDED_OBJ *)lh_retrieve(added,&ad);
+		adp=lh_ADDED_OBJ_retrieve(added,&ad);
 		if (adp != NULL)
 			return(adp->obj);
 		else
@@ -324,7 +339,7 @@ const char *OBJ_nid2sn(int n)
 		ad.type=ADDED_NID;
 		ad.obj= &ob;
 		ob.nid=n;
-		adp=(ADDED_OBJ *)lh_retrieve(added,&ad);
+		adp=lh_ADDED_OBJ_retrieve(added,&ad);
 		if (adp != NULL)
 			return(adp->obj->sn);
 		else
@@ -356,7 +371,7 @@ const char *OBJ_nid2ln(int n)
 		ad.type=ADDED_NID;
 		ad.obj= &ob;
 		ob.nid=n;
-		adp=(ADDED_OBJ *)lh_retrieve(added,&ad);
+		adp=lh_ADDED_OBJ_retrieve(added,&ad);
 		if (adp != NULL)
 			return(adp->obj->ln);
 		else
@@ -367,9 +382,22 @@ const char *OBJ_nid2ln(int n)
 		}
 	}
 
+static int obj_cmp(const ASN1_OBJECT * const *ap, const unsigned int *bp)
+	{
+	int j;
+	const ASN1_OBJECT *a= *ap;
+	const ASN1_OBJECT *b= &nid_objs[*bp];
+
+	j=(a->length - b->length);
+        if (j) return(j);
+	return(memcmp(a->data,b->data,a->length));
+	}
+
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(const ASN1_OBJECT *, unsigned int, obj);
+
 int OBJ_obj2nid(const ASN1_OBJECT *a)
 	{
-	ASN1_OBJECT **op;
+	const unsigned int *op;
 	ADDED_OBJ ad,*adp;
 
 	if (a == NULL)
@@ -381,14 +409,13 @@ int OBJ_obj2nid(const ASN1_OBJECT *a)
 		{
 		ad.type=ADDED_DATA;
 		ad.obj=(ASN1_OBJECT *)a; /* XXX: ugly but harmless */
-		adp=(ADDED_OBJ *)lh_retrieve(added,&ad);
+		adp=lh_ADDED_OBJ_retrieve(added,&ad);
 		if (adp != NULL) return (adp->obj->nid);
 		}
-	op=(ASN1_OBJECT **)OBJ_bsearch((const char *)&a,(const char *)obj_objs,
-		NUM_OBJ, sizeof(ASN1_OBJECT *),obj_cmp);
+	op=OBJ_bsearch_obj(&a, obj_objs, NUM_OBJ);
 	if (op == NULL)
 		return(NID_undef);
-	return((*op)->nid);
+	return(nid_objs[*op].nid);
 	}
 
 /* Convert an object name into an ASN1_OBJECT
@@ -441,7 +468,7 @@ int OBJ_obj2txt(char *buf, int buf_len, const ASN1_OBJECT *a, int no_name)
 	int i,n=0,len,nid, first, use_bn;
 	BIGNUM *bl;
 	unsigned long l;
-	unsigned char *p;
+	const unsigned char *p;
 	char tbuf[DECIMAL_SIZE(i)+DECIMAL_SIZE(l)+2];
 
 	if ((a == NULL) || (a->data == NULL)) {
@@ -610,62 +637,56 @@ int OBJ_txt2nid(const char *s)
 
 int OBJ_ln2nid(const char *s)
 	{
-	ASN1_OBJECT o,*oo= &o,**op;
+	ASN1_OBJECT o;
+	const ASN1_OBJECT *oo= &o;
 	ADDED_OBJ ad,*adp;
+	const unsigned int *op;
 
 	o.ln=s;
 	if (added != NULL)
 		{
 		ad.type=ADDED_LNAME;
 		ad.obj= &o;
-		adp=(ADDED_OBJ *)lh_retrieve(added,&ad);
+		adp=lh_ADDED_OBJ_retrieve(added,&ad);
 		if (adp != NULL) return (adp->obj->nid);
 		}
-	op=(ASN1_OBJECT **)OBJ_bsearch((char *)&oo,(char *)ln_objs, NUM_LN,
-		sizeof(ASN1_OBJECT *),ln_cmp);
+	op=OBJ_bsearch_ln(&oo, ln_objs, NUM_LN);
 	if (op == NULL) return(NID_undef);
-	return((*op)->nid);
+	return(nid_objs[*op].nid);
 	}
 
 int OBJ_sn2nid(const char *s)
 	{
-	ASN1_OBJECT o,*oo= &o,**op;
+	ASN1_OBJECT o;
+	const ASN1_OBJECT *oo= &o;
 	ADDED_OBJ ad,*adp;
+	const unsigned int *op;
 
 	o.sn=s;
 	if (added != NULL)
 		{
 		ad.type=ADDED_SNAME;
 		ad.obj= &o;
-		adp=(ADDED_OBJ *)lh_retrieve(added,&ad);
+		adp=lh_ADDED_OBJ_retrieve(added,&ad);
 		if (adp != NULL) return (adp->obj->nid);
 		}
-	op=(ASN1_OBJECT **)OBJ_bsearch((char *)&oo,(char *)sn_objs,NUM_SN,
-		sizeof(ASN1_OBJECT *),sn_cmp);
+	op=OBJ_bsearch_sn(&oo, sn_objs, NUM_SN);
 	if (op == NULL) return(NID_undef);
-	return((*op)->nid);
+	return(nid_objs[*op].nid);
 	}
 
-static int obj_cmp(const void *ap, const void *bp)
-	{
-	int j;
-	const ASN1_OBJECT *a= *(ASN1_OBJECT * const *)ap;
-	const ASN1_OBJECT *b= *(ASN1_OBJECT * const *)bp;
-
-	j=(a->length - b->length);
-        if (j) return(j);
-	return(memcmp(a->data,b->data,a->length));
-        }
-
-const char *OBJ_bsearch(const char *key, const char *base, int num, int size,
-	int (*cmp)(const void *, const void *))
+const void *OBJ_bsearch_(const void *key, const void *base, int num, int size,
+			 int (*cmp)(const void *, const void *))
 	{
-	return OBJ_bsearch_ex(key, base, num, size, cmp, 0);
+	return OBJ_bsearch_ex_(key, base, num, size, cmp, 0);
 	}
 
-const char *OBJ_bsearch_ex(const char *key, const char *base, int num,
-	int size, int (*cmp)(const void *, const void *), int flags)
+const void *OBJ_bsearch_ex_(const void *key, const void *base_, int num,
+			    int size,
+			    int (*cmp)(const void *, const void *),
+			    int flags)
 	{
+	const char *base=base_;
 	int l,h,i=0,c=0;
 	const char *p = NULL;
 
diff --git a/crypto/objects/obj_dat.h b/crypto/objects/obj_dat.h
index 23bdb4603a91..d404ad07c981 100644
--- a/crypto/objects/obj_dat.h
+++ b/crypto/objects/obj_dat.h
@@ -62,12 +62,12 @@
  * [including the GNU Public Licence.]
  */
 
-#define NUM_NID 893
-#define NUM_SN 886
-#define NUM_LN 886
-#define NUM_OBJ 840
+#define NUM_NID 920
+#define NUM_SN 913
+#define NUM_LN 913
+#define NUM_OBJ 857
 
-static unsigned char lvalues[5824]={
+static const unsigned char lvalues[5980]={
 0x00,                                        /* [  0] OBJ_undef */
 0x2A,0x86,0x48,0x86,0xF7,0x0D,               /* [  1] OBJ_rsadsi */
 0x2A,0x86,0x48,0x86,0xF7,0x0D,0x01,          /* [  7] OBJ_pkcs */
@@ -908,9 +908,26 @@ static unsigned char lvalues[5824]={
 0x55,0x04,0x34,                              /* [5814] OBJ_supportedAlgorithms */
 0x55,0x04,0x35,                              /* [5817] OBJ_deltaRevocationList */
 0x55,0x04,0x36,                              /* [5820] OBJ_dmdName */
+0x2A,0x86,0x48,0x86,0xF7,0x0D,0x01,0x09,0x10,0x03,0x09,/* [5823] OBJ_id_alg_PWRI_KEK */
+0x60,0x86,0x48,0x01,0x65,0x03,0x04,0x01,0x06,/* [5834] OBJ_aes_128_gcm */
+0x60,0x86,0x48,0x01,0x65,0x03,0x04,0x01,0x07,/* [5843] OBJ_aes_128_ccm */
+0x60,0x86,0x48,0x01,0x65,0x03,0x04,0x01,0x08,/* [5852] OBJ_id_aes128_wrap_pad */
+0x60,0x86,0x48,0x01,0x65,0x03,0x04,0x01,0x1A,/* [5861] OBJ_aes_192_gcm */
+0x60,0x86,0x48,0x01,0x65,0x03,0x04,0x01,0x1B,/* [5870] OBJ_aes_192_ccm */
+0x60,0x86,0x48,0x01,0x65,0x03,0x04,0x01,0x1C,/* [5879] OBJ_id_aes192_wrap_pad */
+0x60,0x86,0x48,0x01,0x65,0x03,0x04,0x01,0x2E,/* [5888] OBJ_aes_256_gcm */
+0x60,0x86,0x48,0x01,0x65,0x03,0x04,0x01,0x2F,/* [5897] OBJ_aes_256_ccm */
+0x60,0x86,0x48,0x01,0x65,0x03,0x04,0x01,0x30,/* [5906] OBJ_id_aes256_wrap_pad */
+0x2A,0x83,0x08,0x8C,0x9A,0x4B,0x3D,0x01,0x01,0x03,0x02,/* [5915] OBJ_id_camellia128_wrap */
+0x2A,0x83,0x08,0x8C,0x9A,0x4B,0x3D,0x01,0x01,0x03,0x03,/* [5926] OBJ_id_camellia192_wrap */
+0x2A,0x83,0x08,0x8C,0x9A,0x4B,0x3D,0x01,0x01,0x03,0x04,/* [5937] OBJ_id_camellia256_wrap */
+0x55,0x1D,0x25,0x00,                         /* [5948] OBJ_anyExtendedKeyUsage */
+0x2A,0x86,0x48,0x86,0xF7,0x0D,0x01,0x01,0x08,/* [5952] OBJ_mgf1 */
+0x2A,0x86,0x48,0x86,0xF7,0x0D,0x01,0x01,0x0A,/* [5961] OBJ_rsassaPss */
+0x2A,0x86,0x48,0x86,0xF7,0x0D,0x01,0x01,0x07,/* [5970] OBJ_rsaesOaep */
 };
 
-static ASN1_OBJECT nid_objs[NUM_NID]={
+static const ASN1_OBJECT nid_objs[NUM_NID]={
 {"UNDEF","undefined",NID_undef,1,&(lvalues[0]),0},
 {"rsadsi","RSA Data Security, Inc.",NID_rsadsi,6,&(lvalues[1]),0},
 {"pkcs","RSA Data Security, Inc. PKCS",NID_pkcs,7,&(lvalues[7]),0},
@@ -2351,2626 +2368,2735 @@ static ASN1_OBJECT nid_objs[NUM_NID]={
 {"deltaRevocationList","deltaRevocationList",NID_deltaRevocationList,
 	3,&(lvalues[5817]),0},
 {"dmdName","dmdName",NID_dmdName,3,&(lvalues[5820]),0},
+{"id-alg-PWRI-KEK","id-alg-PWRI-KEK",NID_id_alg_PWRI_KEK,11,
+	&(lvalues[5823]),0},
+{"CMAC","cmac",NID_cmac,0,NULL,0},
+{"id-aes128-GCM","aes-128-gcm",NID_aes_128_gcm,9,&(lvalues[5834]),0},
+{"id-aes128-CCM","aes-128-ccm",NID_aes_128_ccm,9,&(lvalues[5843]),0},
+{"id-aes128-wrap-pad","id-aes128-wrap-pad",NID_id_aes128_wrap_pad,9,
+	&(lvalues[5852]),0},
+{"id-aes192-GCM","aes-192-gcm",NID_aes_192_gcm,9,&(lvalues[5861]),0},
+{"id-aes192-CCM","aes-192-ccm",NID_aes_192_ccm,9,&(lvalues[5870]),0},
+{"id-aes192-wrap-pad","id-aes192-wrap-pad",NID_id_aes192_wrap_pad,9,
+	&(lvalues[5879]),0},
+{"id-aes256-GCM","aes-256-gcm",NID_aes_256_gcm,9,&(lvalues[5888]),0},
+{"id-aes256-CCM","aes-256-ccm",NID_aes_256_ccm,9,&(lvalues[5897]),0},
+{"id-aes256-wrap-pad","id-aes256-wrap-pad",NID_id_aes256_wrap_pad,9,
+	&(lvalues[5906]),0},
+{"AES-128-CTR","aes-128-ctr",NID_aes_128_ctr,0,NULL,0},
+{"AES-192-CTR","aes-192-ctr",NID_aes_192_ctr,0,NULL,0},
+{"AES-256-CTR","aes-256-ctr",NID_aes_256_ctr,0,NULL,0},
+{"id-camellia128-wrap","id-camellia128-wrap",NID_id_camellia128_wrap,
+	11,&(lvalues[5915]),0},
+{"id-camellia192-wrap","id-camellia192-wrap",NID_id_camellia192_wrap,
+	11,&(lvalues[5926]),0},
+{"id-camellia256-wrap","id-camellia256-wrap",NID_id_camellia256_wrap,
+	11,&(lvalues[5937]),0},
+{"anyExtendedKeyUsage","Any Extended Key Usage",
+	NID_anyExtendedKeyUsage,4,&(lvalues[5948]),0},
+{"MGF1","mgf1",NID_mgf1,9,&(lvalues[5952]),0},
+{"RSASSA-PSS","rsassaPss",NID_rsassaPss,9,&(lvalues[5961]),0},
+{"AES-128-XTS","aes-128-xts",NID_aes_128_xts,0,NULL,0},
+{"AES-256-XTS","aes-256-xts",NID_aes_256_xts,0,NULL,0},
+{"RC4-HMAC-MD5","rc4-hmac-md5",NID_rc4_hmac_md5,0,NULL,0},
+{"AES-128-CBC-HMAC-SHA1","aes-128-cbc-hmac-sha1",
+	NID_aes_128_cbc_hmac_sha1,0,NULL,0},
+{"AES-192-CBC-HMAC-SHA1","aes-192-cbc-hmac-sha1",
+	NID_aes_192_cbc_hmac_sha1,0,NULL,0},
+{"AES-256-CBC-HMAC-SHA1","aes-256-cbc-hmac-sha1",
+	NID_aes_256_cbc_hmac_sha1,0,NULL,0},
+{"RSAES-OAEP","rsaesOaep",NID_rsaesOaep,9,&(lvalues[5970]),0},
 };
 
-static ASN1_OBJECT *sn_objs[NUM_SN]={
-&(nid_objs[364]),/* "AD_DVCS" */
-&(nid_objs[419]),/* "AES-128-CBC" */
-&(nid_objs[421]),/* "AES-128-CFB" */
-&(nid_objs[650]),/* "AES-128-CFB1" */
-&(nid_objs[653]),/* "AES-128-CFB8" */
-&(nid_objs[418]),/* "AES-128-ECB" */
-&(nid_objs[420]),/* "AES-128-OFB" */
-&(nid_objs[423]),/* "AES-192-CBC" */
-&(nid_objs[425]),/* "AES-192-CFB" */
-&(nid_objs[651]),/* "AES-192-CFB1" */
-&(nid_objs[654]),/* "AES-192-CFB8" */
-&(nid_objs[422]),/* "AES-192-ECB" */
-&(nid_objs[424]),/* "AES-192-OFB" */
-&(nid_objs[427]),/* "AES-256-CBC" */
-&(nid_objs[429]),/* "AES-256-CFB" */
-&(nid_objs[652]),/* "AES-256-CFB1" */
-&(nid_objs[655]),/* "AES-256-CFB8" */
-&(nid_objs[426]),/* "AES-256-ECB" */
-&(nid_objs[428]),/* "AES-256-OFB" */
-&(nid_objs[91]),/* "BF-CBC" */
-&(nid_objs[93]),/* "BF-CFB" */
-&(nid_objs[92]),/* "BF-ECB" */
-&(nid_objs[94]),/* "BF-OFB" */
-&(nid_objs[14]),/* "C" */
-&(nid_objs[751]),/* "CAMELLIA-128-CBC" */
-&(nid_objs[757]),/* "CAMELLIA-128-CFB" */
-&(nid_objs[760]),/* "CAMELLIA-128-CFB1" */
-&(nid_objs[763]),/* "CAMELLIA-128-CFB8" */
-&(nid_objs[754]),/* "CAMELLIA-128-ECB" */
-&(nid_objs[766]),/* "CAMELLIA-128-OFB" */
-&(nid_objs[752]),/* "CAMELLIA-192-CBC" */
-&(nid_objs[758]),/* "CAMELLIA-192-CFB" */
-&(nid_objs[761]),/* "CAMELLIA-192-CFB1" */
-&(nid_objs[764]),/* "CAMELLIA-192-CFB8" */
-&(nid_objs[755]),/* "CAMELLIA-192-ECB" */
-&(nid_objs[767]),/* "CAMELLIA-192-OFB" */
-&(nid_objs[753]),/* "CAMELLIA-256-CBC" */
-&(nid_objs[759]),/* "CAMELLIA-256-CFB" */
-&(nid_objs[762]),/* "CAMELLIA-256-CFB1" */
-&(nid_objs[765]),/* "CAMELLIA-256-CFB8" */
-&(nid_objs[756]),/* "CAMELLIA-256-ECB" */
-&(nid_objs[768]),/* "CAMELLIA-256-OFB" */
-&(nid_objs[108]),/* "CAST5-CBC" */
-&(nid_objs[110]),/* "CAST5-CFB" */
-&(nid_objs[109]),/* "CAST5-ECB" */
-&(nid_objs[111]),/* "CAST5-OFB" */
-&(nid_objs[13]),/* "CN" */
-&(nid_objs[141]),/* "CRLReason" */
-&(nid_objs[417]),/* "CSPName" */
-&(nid_objs[367]),/* "CrlID" */
-&(nid_objs[391]),/* "DC" */
-&(nid_objs[31]),/* "DES-CBC" */
-&(nid_objs[643]),/* "DES-CDMF" */
-&(nid_objs[30]),/* "DES-CFB" */
-&(nid_objs[656]),/* "DES-CFB1" */
-&(nid_objs[657]),/* "DES-CFB8" */
-&(nid_objs[29]),/* "DES-ECB" */
-&(nid_objs[32]),/* "DES-EDE" */
-&(nid_objs[43]),/* "DES-EDE-CBC" */
-&(nid_objs[60]),/* "DES-EDE-CFB" */
-&(nid_objs[62]),/* "DES-EDE-OFB" */
-&(nid_objs[33]),/* "DES-EDE3" */
-&(nid_objs[44]),/* "DES-EDE3-CBC" */
-&(nid_objs[61]),/* "DES-EDE3-CFB" */
-&(nid_objs[658]),/* "DES-EDE3-CFB1" */
-&(nid_objs[659]),/* "DES-EDE3-CFB8" */
-&(nid_objs[63]),/* "DES-EDE3-OFB" */
-&(nid_objs[45]),/* "DES-OFB" */
-&(nid_objs[80]),/* "DESX-CBC" */
-&(nid_objs[380]),/* "DOD" */
-&(nid_objs[116]),/* "DSA" */
-&(nid_objs[66]),/* "DSA-SHA" */
-&(nid_objs[113]),/* "DSA-SHA1" */
-&(nid_objs[70]),/* "DSA-SHA1-old" */
-&(nid_objs[67]),/* "DSA-old" */
-&(nid_objs[297]),/* "DVCS" */
-&(nid_objs[99]),/* "GN" */
-&(nid_objs[855]),/* "HMAC" */
-&(nid_objs[780]),/* "HMAC-MD5" */
-&(nid_objs[781]),/* "HMAC-SHA1" */
-&(nid_objs[381]),/* "IANA" */
-&(nid_objs[34]),/* "IDEA-CBC" */
-&(nid_objs[35]),/* "IDEA-CFB" */
-&(nid_objs[36]),/* "IDEA-ECB" */
-&(nid_objs[46]),/* "IDEA-OFB" */
-&(nid_objs[181]),/* "ISO" */
-&(nid_objs[183]),/* "ISO-US" */
-&(nid_objs[645]),/* "ITU-T" */
-&(nid_objs[646]),/* "JOINT-ISO-ITU-T" */
-&(nid_objs[773]),/* "KISA" */
-&(nid_objs[15]),/* "L" */
-&(nid_objs[856]),/* "LocalKeySet" */
-&(nid_objs[ 3]),/* "MD2" */
-&(nid_objs[257]),/* "MD4" */
-&(nid_objs[ 4]),/* "MD5" */
-&(nid_objs[114]),/* "MD5-SHA1" */
-&(nid_objs[95]),/* "MDC2" */
-&(nid_objs[388]),/* "Mail" */
-&(nid_objs[393]),/* "NULL" */
-&(nid_objs[404]),/* "NULL" */
-&(nid_objs[57]),/* "Netscape" */
-&(nid_objs[366]),/* "Nonce" */
-&(nid_objs[17]),/* "O" */
-&(nid_objs[178]),/* "OCSP" */
-&(nid_objs[180]),/* "OCSPSigning" */
-&(nid_objs[379]),/* "ORG" */
-&(nid_objs[18]),/* "OU" */
-&(nid_objs[749]),/* "Oakley-EC2N-3" */
-&(nid_objs[750]),/* "Oakley-EC2N-4" */
-&(nid_objs[ 9]),/* "PBE-MD2-DES" */
-&(nid_objs[168]),/* "PBE-MD2-RC2-64" */
-&(nid_objs[10]),/* "PBE-MD5-DES" */
-&(nid_objs[169]),/* "PBE-MD5-RC2-64" */
-&(nid_objs[147]),/* "PBE-SHA1-2DES" */
-&(nid_objs[146]),/* "PBE-SHA1-3DES" */
-&(nid_objs[170]),/* "PBE-SHA1-DES" */
-&(nid_objs[148]),/* "PBE-SHA1-RC2-128" */
-&(nid_objs[149]),/* "PBE-SHA1-RC2-40" */
-&(nid_objs[68]),/* "PBE-SHA1-RC2-64" */
-&(nid_objs[144]),/* "PBE-SHA1-RC4-128" */
-&(nid_objs[145]),/* "PBE-SHA1-RC4-40" */
-&(nid_objs[161]),/* "PBES2" */
-&(nid_objs[69]),/* "PBKDF2" */
-&(nid_objs[162]),/* "PBMAC1" */
-&(nid_objs[127]),/* "PKIX" */
-&(nid_objs[98]),/* "RC2-40-CBC" */
-&(nid_objs[166]),/* "RC2-64-CBC" */
-&(nid_objs[37]),/* "RC2-CBC" */
-&(nid_objs[39]),/* "RC2-CFB" */
-&(nid_objs[38]),/* "RC2-ECB" */
-&(nid_objs[40]),/* "RC2-OFB" */
-&(nid_objs[ 5]),/* "RC4" */
-&(nid_objs[97]),/* "RC4-40" */
-&(nid_objs[120]),/* "RC5-CBC" */
-&(nid_objs[122]),/* "RC5-CFB" */
-&(nid_objs[121]),/* "RC5-ECB" */
-&(nid_objs[123]),/* "RC5-OFB" */
-&(nid_objs[117]),/* "RIPEMD160" */
-&(nid_objs[124]),/* "RLE" */
-&(nid_objs[19]),/* "RSA" */
-&(nid_objs[ 7]),/* "RSA-MD2" */
-&(nid_objs[396]),/* "RSA-MD4" */
-&(nid_objs[ 8]),/* "RSA-MD5" */
-&(nid_objs[96]),/* "RSA-MDC2" */
-&(nid_objs[104]),/* "RSA-NP-MD5" */
-&(nid_objs[119]),/* "RSA-RIPEMD160" */
-&(nid_objs[42]),/* "RSA-SHA" */
-&(nid_objs[65]),/* "RSA-SHA1" */
-&(nid_objs[115]),/* "RSA-SHA1-2" */
-&(nid_objs[671]),/* "RSA-SHA224" */
-&(nid_objs[668]),/* "RSA-SHA256" */
-&(nid_objs[669]),/* "RSA-SHA384" */
-&(nid_objs[670]),/* "RSA-SHA512" */
-&(nid_objs[777]),/* "SEED-CBC" */
-&(nid_objs[779]),/* "SEED-CFB" */
-&(nid_objs[776]),/* "SEED-ECB" */
-&(nid_objs[778]),/* "SEED-OFB" */
-&(nid_objs[41]),/* "SHA" */
-&(nid_objs[64]),/* "SHA1" */
-&(nid_objs[675]),/* "SHA224" */
-&(nid_objs[672]),/* "SHA256" */
-&(nid_objs[673]),/* "SHA384" */
-&(nid_objs[674]),/* "SHA512" */
-&(nid_objs[188]),/* "SMIME" */
-&(nid_objs[167]),/* "SMIME-CAPS" */
-&(nid_objs[100]),/* "SN" */
-&(nid_objs[16]),/* "ST" */
-&(nid_objs[143]),/* "SXNetID" */
-&(nid_objs[458]),/* "UID" */
-&(nid_objs[ 0]),/* "UNDEF" */
-&(nid_objs[11]),/* "X500" */
-&(nid_objs[378]),/* "X500algorithms" */
-&(nid_objs[12]),/* "X509" */
-&(nid_objs[184]),/* "X9-57" */
-&(nid_objs[185]),/* "X9cm" */
-&(nid_objs[125]),/* "ZLIB" */
-&(nid_objs[478]),/* "aRecord" */
-&(nid_objs[289]),/* "aaControls" */
-&(nid_objs[287]),/* "ac-auditEntity" */
-&(nid_objs[397]),/* "ac-proxying" */
-&(nid_objs[288]),/* "ac-targeting" */
-&(nid_objs[368]),/* "acceptableResponses" */
-&(nid_objs[446]),/* "account" */
-&(nid_objs[363]),/* "ad_timestamping" */
-&(nid_objs[376]),/* "algorithm" */
-&(nid_objs[405]),/* "ansi-X9-62" */
-&(nid_objs[746]),/* "anyPolicy" */
-&(nid_objs[370]),/* "archiveCutoff" */
-&(nid_objs[484]),/* "associatedDomain" */
-&(nid_objs[485]),/* "associatedName" */
-&(nid_objs[501]),/* "audio" */
-&(nid_objs[177]),/* "authorityInfoAccess" */
-&(nid_objs[90]),/* "authorityKeyIdentifier" */
-&(nid_objs[882]),/* "authorityRevocationList" */
-&(nid_objs[87]),/* "basicConstraints" */
-&(nid_objs[365]),/* "basicOCSPResponse" */
-&(nid_objs[285]),/* "biometricInfo" */
-&(nid_objs[494]),/* "buildingName" */
-&(nid_objs[860]),/* "businessCategory" */
-&(nid_objs[691]),/* "c2onb191v4" */
-&(nid_objs[692]),/* "c2onb191v5" */
-&(nid_objs[697]),/* "c2onb239v4" */
-&(nid_objs[698]),/* "c2onb239v5" */
-&(nid_objs[684]),/* "c2pnb163v1" */
-&(nid_objs[685]),/* "c2pnb163v2" */
-&(nid_objs[686]),/* "c2pnb163v3" */
-&(nid_objs[687]),/* "c2pnb176v1" */
-&(nid_objs[693]),/* "c2pnb208w1" */
-&(nid_objs[699]),/* "c2pnb272w1" */
-&(nid_objs[700]),/* "c2pnb304w1" */
-&(nid_objs[702]),/* "c2pnb368w1" */
-&(nid_objs[688]),/* "c2tnb191v1" */
-&(nid_objs[689]),/* "c2tnb191v2" */
-&(nid_objs[690]),/* "c2tnb191v3" */
-&(nid_objs[694]),/* "c2tnb239v1" */
-&(nid_objs[695]),/* "c2tnb239v2" */
-&(nid_objs[696]),/* "c2tnb239v3" */
-&(nid_objs[701]),/* "c2tnb359v1" */
-&(nid_objs[703]),/* "c2tnb431r1" */
-&(nid_objs[881]),/* "cACertificate" */
-&(nid_objs[483]),/* "cNAMERecord" */
-&(nid_objs[179]),/* "caIssuers" */
-&(nid_objs[785]),/* "caRepository" */
-&(nid_objs[443]),/* "caseIgnoreIA5StringSyntax" */
-&(nid_objs[152]),/* "certBag" */
-&(nid_objs[677]),/* "certicom-arc" */
-&(nid_objs[771]),/* "certificateIssuer" */
-&(nid_objs[89]),/* "certificatePolicies" */
-&(nid_objs[883]),/* "certificateRevocationList" */
-&(nid_objs[54]),/* "challengePassword" */
-&(nid_objs[407]),/* "characteristic-two-field" */
-&(nid_objs[395]),/* "clearance" */
-&(nid_objs[130]),/* "clientAuth" */
-&(nid_objs[131]),/* "codeSigning" */
-&(nid_objs[50]),/* "contentType" */
-&(nid_objs[53]),/* "countersignature" */
-&(nid_objs[153]),/* "crlBag" */
-&(nid_objs[103]),/* "crlDistributionPoints" */
-&(nid_objs[88]),/* "crlNumber" */
-&(nid_objs[884]),/* "crossCertificatePair" */
-&(nid_objs[806]),/* "cryptocom" */
-&(nid_objs[805]),/* "cryptopro" */
-&(nid_objs[500]),/* "dITRedirect" */
-&(nid_objs[451]),/* "dNSDomain" */
-&(nid_objs[495]),/* "dSAQuality" */
-&(nid_objs[434]),/* "data" */
-&(nid_objs[390]),/* "dcobject" */
-&(nid_objs[140]),/* "deltaCRL" */
-&(nid_objs[891]),/* "deltaRevocationList" */
-&(nid_objs[107]),/* "description" */
-&(nid_objs[871]),/* "destinationIndicator" */
-&(nid_objs[28]),/* "dhKeyAgreement" */
-&(nid_objs[382]),/* "directory" */
-&(nid_objs[887]),/* "distinguishedName" */
-&(nid_objs[892]),/* "dmdName" */
-&(nid_objs[174]),/* "dnQualifier" */
-&(nid_objs[447]),/* "document" */
-&(nid_objs[471]),/* "documentAuthor" */
-&(nid_objs[468]),/* "documentIdentifier" */
-&(nid_objs[472]),/* "documentLocation" */
-&(nid_objs[502]),/* "documentPublisher" */
-&(nid_objs[449]),/* "documentSeries" */
-&(nid_objs[469]),/* "documentTitle" */
-&(nid_objs[470]),/* "documentVersion" */
-&(nid_objs[392]),/* "domain" */
-&(nid_objs[452]),/* "domainRelatedObject" */
-&(nid_objs[802]),/* "dsa_with_SHA224" */
-&(nid_objs[803]),/* "dsa_with_SHA256" */
-&(nid_objs[791]),/* "ecdsa-with-Recommended" */
-&(nid_objs[416]),/* "ecdsa-with-SHA1" */
-&(nid_objs[793]),/* "ecdsa-with-SHA224" */
-&(nid_objs[794]),/* "ecdsa-with-SHA256" */
-&(nid_objs[795]),/* "ecdsa-with-SHA384" */
-&(nid_objs[796]),/* "ecdsa-with-SHA512" */
-&(nid_objs[792]),/* "ecdsa-with-Specified" */
-&(nid_objs[48]),/* "emailAddress" */
-&(nid_objs[132]),/* "emailProtection" */
-&(nid_objs[885]),/* "enhancedSearchGuide" */
-&(nid_objs[389]),/* "enterprises" */
-&(nid_objs[384]),/* "experimental" */
-&(nid_objs[172]),/* "extReq" */
-&(nid_objs[56]),/* "extendedCertificateAttributes" */
-&(nid_objs[126]),/* "extendedKeyUsage" */
-&(nid_objs[372]),/* "extendedStatus" */
-&(nid_objs[867]),/* "facsimileTelephoneNumber" */
-&(nid_objs[462]),/* "favouriteDrink" */
-&(nid_objs[857]),/* "freshestCRL" */
-&(nid_objs[453]),/* "friendlyCountry" */
-&(nid_objs[490]),/* "friendlyCountryName" */
-&(nid_objs[156]),/* "friendlyName" */
-&(nid_objs[509]),/* "generationQualifier" */
-&(nid_objs[815]),/* "gost-mac" */
-&(nid_objs[811]),/* "gost2001" */
-&(nid_objs[851]),/* "gost2001cc" */
-&(nid_objs[813]),/* "gost89" */
-&(nid_objs[814]),/* "gost89-cnt" */
-&(nid_objs[812]),/* "gost94" */
-&(nid_objs[850]),/* "gost94cc" */
-&(nid_objs[797]),/* "hmacWithMD5" */
-&(nid_objs[163]),/* "hmacWithSHA1" */
-&(nid_objs[798]),/* "hmacWithSHA224" */
-&(nid_objs[799]),/* "hmacWithSHA256" */
-&(nid_objs[800]),/* "hmacWithSHA384" */
-&(nid_objs[801]),/* "hmacWithSHA512" */
-&(nid_objs[432]),/* "holdInstructionCallIssuer" */
-&(nid_objs[430]),/* "holdInstructionCode" */
-&(nid_objs[431]),/* "holdInstructionNone" */
-&(nid_objs[433]),/* "holdInstructionReject" */
-&(nid_objs[486]),/* "homePostalAddress" */
-&(nid_objs[473]),/* "homeTelephoneNumber" */
-&(nid_objs[466]),/* "host" */
-&(nid_objs[889]),/* "houseIdentifier" */
-&(nid_objs[442]),/* "iA5StringSyntax" */
-&(nid_objs[783]),/* "id-DHBasedMac" */
-&(nid_objs[824]),/* "id-Gost28147-89-CryptoPro-A-ParamSet" */
-&(nid_objs[825]),/* "id-Gost28147-89-CryptoPro-B-ParamSet" */
-&(nid_objs[826]),/* "id-Gost28147-89-CryptoPro-C-ParamSet" */
-&(nid_objs[827]),/* "id-Gost28147-89-CryptoPro-D-ParamSet" */
-&(nid_objs[819]),/* "id-Gost28147-89-CryptoPro-KeyMeshing" */
-&(nid_objs[829]),/* "id-Gost28147-89-CryptoPro-Oscar-1-0-ParamSet" */
-&(nid_objs[828]),/* "id-Gost28147-89-CryptoPro-Oscar-1-1-ParamSet" */
-&(nid_objs[830]),/* "id-Gost28147-89-CryptoPro-RIC-1-ParamSet" */
-&(nid_objs[820]),/* "id-Gost28147-89-None-KeyMeshing" */
-&(nid_objs[823]),/* "id-Gost28147-89-TestParamSet" */
-&(nid_objs[849]),/* "id-Gost28147-89-cc" */
-&(nid_objs[840]),/* "id-GostR3410-2001-CryptoPro-A-ParamSet" */
-&(nid_objs[841]),/* "id-GostR3410-2001-CryptoPro-B-ParamSet" */
-&(nid_objs[842]),/* "id-GostR3410-2001-CryptoPro-C-ParamSet" */
-&(nid_objs[843]),/* "id-GostR3410-2001-CryptoPro-XchA-ParamSet" */
-&(nid_objs[844]),/* "id-GostR3410-2001-CryptoPro-XchB-ParamSet" */
-&(nid_objs[854]),/* "id-GostR3410-2001-ParamSet-cc" */
-&(nid_objs[839]),/* "id-GostR3410-2001-TestParamSet" */
-&(nid_objs[817]),/* "id-GostR3410-2001DH" */
-&(nid_objs[832]),/* "id-GostR3410-94-CryptoPro-A-ParamSet" */
-&(nid_objs[833]),/* "id-GostR3410-94-CryptoPro-B-ParamSet" */
-&(nid_objs[834]),/* "id-GostR3410-94-CryptoPro-C-ParamSet" */
-&(nid_objs[835]),/* "id-GostR3410-94-CryptoPro-D-ParamSet" */
-&(nid_objs[836]),/* "id-GostR3410-94-CryptoPro-XchA-ParamSet" */
-&(nid_objs[837]),/* "id-GostR3410-94-CryptoPro-XchB-ParamSet" */
-&(nid_objs[838]),/* "id-GostR3410-94-CryptoPro-XchC-ParamSet" */
-&(nid_objs[831]),/* "id-GostR3410-94-TestParamSet" */
-&(nid_objs[845]),/* "id-GostR3410-94-a" */
-&(nid_objs[846]),/* "id-GostR3410-94-aBis" */
-&(nid_objs[847]),/* "id-GostR3410-94-b" */
-&(nid_objs[848]),/* "id-GostR3410-94-bBis" */
-&(nid_objs[818]),/* "id-GostR3410-94DH" */
-&(nid_objs[822]),/* "id-GostR3411-94-CryptoProParamSet" */
-&(nid_objs[821]),/* "id-GostR3411-94-TestParamSet" */
-&(nid_objs[807]),/* "id-GostR3411-94-with-GostR3410-2001" */
-&(nid_objs[853]),/* "id-GostR3411-94-with-GostR3410-2001-cc" */
-&(nid_objs[808]),/* "id-GostR3411-94-with-GostR3410-94" */
-&(nid_objs[852]),/* "id-GostR3411-94-with-GostR3410-94-cc" */
-&(nid_objs[810]),/* "id-HMACGostR3411-94" */
-&(nid_objs[782]),/* "id-PasswordBasedMAC" */
-&(nid_objs[266]),/* "id-aca" */
-&(nid_objs[355]),/* "id-aca-accessIdentity" */
-&(nid_objs[354]),/* "id-aca-authenticationInfo" */
-&(nid_objs[356]),/* "id-aca-chargingIdentity" */
-&(nid_objs[399]),/* "id-aca-encAttrs" */
-&(nid_objs[357]),/* "id-aca-group" */
-&(nid_objs[358]),/* "id-aca-role" */
-&(nid_objs[176]),/* "id-ad" */
-&(nid_objs[788]),/* "id-aes128-wrap" */
-&(nid_objs[789]),/* "id-aes192-wrap" */
-&(nid_objs[790]),/* "id-aes256-wrap" */
-&(nid_objs[262]),/* "id-alg" */
-&(nid_objs[323]),/* "id-alg-des40" */
-&(nid_objs[326]),/* "id-alg-dh-pop" */
-&(nid_objs[325]),/* "id-alg-dh-sig-hmac-sha1" */
-&(nid_objs[324]),/* "id-alg-noSignature" */
-&(nid_objs[268]),/* "id-cct" */
-&(nid_objs[361]),/* "id-cct-PKIData" */
-&(nid_objs[362]),/* "id-cct-PKIResponse" */
-&(nid_objs[360]),/* "id-cct-crs" */
-&(nid_objs[81]),/* "id-ce" */
-&(nid_objs[680]),/* "id-characteristic-two-basis" */
-&(nid_objs[263]),/* "id-cmc" */
-&(nid_objs[334]),/* "id-cmc-addExtensions" */
-&(nid_objs[346]),/* "id-cmc-confirmCertAcceptance" */
-&(nid_objs[330]),/* "id-cmc-dataReturn" */
-&(nid_objs[336]),/* "id-cmc-decryptedPOP" */
-&(nid_objs[335]),/* "id-cmc-encryptedPOP" */
-&(nid_objs[339]),/* "id-cmc-getCRL" */
-&(nid_objs[338]),/* "id-cmc-getCert" */
-&(nid_objs[328]),/* "id-cmc-identification" */
-&(nid_objs[329]),/* "id-cmc-identityProof" */
-&(nid_objs[337]),/* "id-cmc-lraPOPWitness" */
-&(nid_objs[344]),/* "id-cmc-popLinkRandom" */
-&(nid_objs[345]),/* "id-cmc-popLinkWitness" */
-&(nid_objs[343]),/* "id-cmc-queryPending" */
-&(nid_objs[333]),/* "id-cmc-recipientNonce" */
-&(nid_objs[341]),/* "id-cmc-regInfo" */
-&(nid_objs[342]),/* "id-cmc-responseInfo" */
-&(nid_objs[340]),/* "id-cmc-revokeRequest" */
-&(nid_objs[332]),/* "id-cmc-senderNonce" */
-&(nid_objs[327]),/* "id-cmc-statusInfo" */
-&(nid_objs[331]),/* "id-cmc-transactionId" */
-&(nid_objs[787]),/* "id-ct-asciiTextWithCRLF" */
-&(nid_objs[408]),/* "id-ecPublicKey" */
-&(nid_objs[508]),/* "id-hex-multipart-message" */
-&(nid_objs[507]),/* "id-hex-partial-message" */
-&(nid_objs[260]),/* "id-it" */
-&(nid_objs[302]),/* "id-it-caKeyUpdateInfo" */
-&(nid_objs[298]),/* "id-it-caProtEncCert" */
-&(nid_objs[311]),/* "id-it-confirmWaitTime" */
-&(nid_objs[303]),/* "id-it-currentCRL" */
-&(nid_objs[300]),/* "id-it-encKeyPairTypes" */
-&(nid_objs[310]),/* "id-it-implicitConfirm" */
-&(nid_objs[308]),/* "id-it-keyPairParamRep" */
-&(nid_objs[307]),/* "id-it-keyPairParamReq" */
-&(nid_objs[312]),/* "id-it-origPKIMessage" */
-&(nid_objs[301]),/* "id-it-preferredSymmAlg" */
-&(nid_objs[309]),/* "id-it-revPassphrase" */
-&(nid_objs[299]),/* "id-it-signKeyPairTypes" */
-&(nid_objs[305]),/* "id-it-subscriptionRequest" */
-&(nid_objs[306]),/* "id-it-subscriptionResponse" */
-&(nid_objs[784]),/* "id-it-suppLangTags" */
-&(nid_objs[304]),/* "id-it-unsupportedOIDs" */
-&(nid_objs[128]),/* "id-kp" */
-&(nid_objs[280]),/* "id-mod-attribute-cert" */
-&(nid_objs[274]),/* "id-mod-cmc" */
-&(nid_objs[277]),/* "id-mod-cmp" */
-&(nid_objs[284]),/* "id-mod-cmp2000" */
-&(nid_objs[273]),/* "id-mod-crmf" */
-&(nid_objs[283]),/* "id-mod-dvcs" */
-&(nid_objs[275]),/* "id-mod-kea-profile-88" */
-&(nid_objs[276]),/* "id-mod-kea-profile-93" */
-&(nid_objs[282]),/* "id-mod-ocsp" */
-&(nid_objs[278]),/* "id-mod-qualified-cert-88" */
-&(nid_objs[279]),/* "id-mod-qualified-cert-93" */
-&(nid_objs[281]),/* "id-mod-timestamp-protocol" */
-&(nid_objs[264]),/* "id-on" */
-&(nid_objs[858]),/* "id-on-permanentIdentifier" */
-&(nid_objs[347]),/* "id-on-personalData" */
-&(nid_objs[265]),/* "id-pda" */
-&(nid_objs[352]),/* "id-pda-countryOfCitizenship" */
-&(nid_objs[353]),/* "id-pda-countryOfResidence" */
-&(nid_objs[348]),/* "id-pda-dateOfBirth" */
-&(nid_objs[351]),/* "id-pda-gender" */
-&(nid_objs[349]),/* "id-pda-placeOfBirth" */
-&(nid_objs[175]),/* "id-pe" */
-&(nid_objs[261]),/* "id-pkip" */
-&(nid_objs[258]),/* "id-pkix-mod" */
-&(nid_objs[269]),/* "id-pkix1-explicit-88" */
-&(nid_objs[271]),/* "id-pkix1-explicit-93" */
-&(nid_objs[270]),/* "id-pkix1-implicit-88" */
-&(nid_objs[272]),/* "id-pkix1-implicit-93" */
-&(nid_objs[662]),/* "id-ppl" */
-&(nid_objs[664]),/* "id-ppl-anyLanguage" */
-&(nid_objs[667]),/* "id-ppl-independent" */
-&(nid_objs[665]),/* "id-ppl-inheritAll" */
-&(nid_objs[267]),/* "id-qcs" */
-&(nid_objs[359]),/* "id-qcs-pkixQCSyntax-v1" */
-&(nid_objs[259]),/* "id-qt" */
-&(nid_objs[164]),/* "id-qt-cps" */
-&(nid_objs[165]),/* "id-qt-unotice" */
-&(nid_objs[313]),/* "id-regCtrl" */
-&(nid_objs[316]),/* "id-regCtrl-authenticator" */
-&(nid_objs[319]),/* "id-regCtrl-oldCertID" */
-&(nid_objs[318]),/* "id-regCtrl-pkiArchiveOptions" */
-&(nid_objs[317]),/* "id-regCtrl-pkiPublicationInfo" */
-&(nid_objs[320]),/* "id-regCtrl-protocolEncrKey" */
-&(nid_objs[315]),/* "id-regCtrl-regToken" */
-&(nid_objs[314]),/* "id-regInfo" */
-&(nid_objs[322]),/* "id-regInfo-certReq" */
-&(nid_objs[321]),/* "id-regInfo-utf8Pairs" */
-&(nid_objs[512]),/* "id-set" */
-&(nid_objs[191]),/* "id-smime-aa" */
-&(nid_objs[215]),/* "id-smime-aa-contentHint" */
-&(nid_objs[218]),/* "id-smime-aa-contentIdentifier" */
-&(nid_objs[221]),/* "id-smime-aa-contentReference" */
-&(nid_objs[240]),/* "id-smime-aa-dvcs-dvc" */
-&(nid_objs[217]),/* "id-smime-aa-encapContentType" */
-&(nid_objs[222]),/* "id-smime-aa-encrypKeyPref" */
-&(nid_objs[220]),/* "id-smime-aa-equivalentLabels" */
-&(nid_objs[232]),/* "id-smime-aa-ets-CertificateRefs" */
-&(nid_objs[233]),/* "id-smime-aa-ets-RevocationRefs" */
-&(nid_objs[238]),/* "id-smime-aa-ets-archiveTimeStamp" */
-&(nid_objs[237]),/* "id-smime-aa-ets-certCRLTimestamp" */
-&(nid_objs[234]),/* "id-smime-aa-ets-certValues" */
-&(nid_objs[227]),/* "id-smime-aa-ets-commitmentType" */
-&(nid_objs[231]),/* "id-smime-aa-ets-contentTimestamp" */
-&(nid_objs[236]),/* "id-smime-aa-ets-escTimeStamp" */
-&(nid_objs[230]),/* "id-smime-aa-ets-otherSigCert" */
-&(nid_objs[235]),/* "id-smime-aa-ets-revocationValues" */
-&(nid_objs[226]),/* "id-smime-aa-ets-sigPolicyId" */
-&(nid_objs[229]),/* "id-smime-aa-ets-signerAttr" */
-&(nid_objs[228]),/* "id-smime-aa-ets-signerLocation" */
-&(nid_objs[219]),/* "id-smime-aa-macValue" */
-&(nid_objs[214]),/* "id-smime-aa-mlExpandHistory" */
-&(nid_objs[216]),/* "id-smime-aa-msgSigDigest" */
-&(nid_objs[212]),/* "id-smime-aa-receiptRequest" */
-&(nid_objs[213]),/* "id-smime-aa-securityLabel" */
-&(nid_objs[239]),/* "id-smime-aa-signatureType" */
-&(nid_objs[223]),/* "id-smime-aa-signingCertificate" */
-&(nid_objs[224]),/* "id-smime-aa-smimeEncryptCerts" */
-&(nid_objs[225]),/* "id-smime-aa-timeStampToken" */
-&(nid_objs[192]),/* "id-smime-alg" */
-&(nid_objs[243]),/* "id-smime-alg-3DESwrap" */
-&(nid_objs[246]),/* "id-smime-alg-CMS3DESwrap" */
-&(nid_objs[247]),/* "id-smime-alg-CMSRC2wrap" */
-&(nid_objs[245]),/* "id-smime-alg-ESDH" */
-&(nid_objs[241]),/* "id-smime-alg-ESDHwith3DES" */
-&(nid_objs[242]),/* "id-smime-alg-ESDHwithRC2" */
-&(nid_objs[244]),/* "id-smime-alg-RC2wrap" */
-&(nid_objs[193]),/* "id-smime-cd" */
-&(nid_objs[248]),/* "id-smime-cd-ldap" */
-&(nid_objs[190]),/* "id-smime-ct" */
-&(nid_objs[210]),/* "id-smime-ct-DVCSRequestData" */
-&(nid_objs[211]),/* "id-smime-ct-DVCSResponseData" */
-&(nid_objs[208]),/* "id-smime-ct-TDTInfo" */
-&(nid_objs[207]),/* "id-smime-ct-TSTInfo" */
-&(nid_objs[205]),/* "id-smime-ct-authData" */
-&(nid_objs[786]),/* "id-smime-ct-compressedData" */
-&(nid_objs[209]),/* "id-smime-ct-contentInfo" */
-&(nid_objs[206]),/* "id-smime-ct-publishCert" */
-&(nid_objs[204]),/* "id-smime-ct-receipt" */
-&(nid_objs[195]),/* "id-smime-cti" */
-&(nid_objs[255]),/* "id-smime-cti-ets-proofOfApproval" */
-&(nid_objs[256]),/* "id-smime-cti-ets-proofOfCreation" */
-&(nid_objs[253]),/* "id-smime-cti-ets-proofOfDelivery" */
-&(nid_objs[251]),/* "id-smime-cti-ets-proofOfOrigin" */
-&(nid_objs[252]),/* "id-smime-cti-ets-proofOfReceipt" */
-&(nid_objs[254]),/* "id-smime-cti-ets-proofOfSender" */
-&(nid_objs[189]),/* "id-smime-mod" */
-&(nid_objs[196]),/* "id-smime-mod-cms" */
-&(nid_objs[197]),/* "id-smime-mod-ess" */
-&(nid_objs[202]),/* "id-smime-mod-ets-eSigPolicy-88" */
-&(nid_objs[203]),/* "id-smime-mod-ets-eSigPolicy-97" */
-&(nid_objs[200]),/* "id-smime-mod-ets-eSignature-88" */
-&(nid_objs[201]),/* "id-smime-mod-ets-eSignature-97" */
-&(nid_objs[199]),/* "id-smime-mod-msg-v3" */
-&(nid_objs[198]),/* "id-smime-mod-oid" */
-&(nid_objs[194]),/* "id-smime-spq" */
-&(nid_objs[250]),/* "id-smime-spq-ets-sqt-unotice" */
-&(nid_objs[249]),/* "id-smime-spq-ets-sqt-uri" */
-&(nid_objs[676]),/* "identified-organization" */
-&(nid_objs[461]),/* "info" */
-&(nid_objs[748]),/* "inhibitAnyPolicy" */
-&(nid_objs[101]),/* "initials" */
-&(nid_objs[647]),/* "international-organizations" */
-&(nid_objs[869]),/* "internationaliSDNNumber" */
-&(nid_objs[142]),/* "invalidityDate" */
-&(nid_objs[294]),/* "ipsecEndSystem" */
-&(nid_objs[295]),/* "ipsecTunnel" */
-&(nid_objs[296]),/* "ipsecUser" */
-&(nid_objs[86]),/* "issuerAltName" */
-&(nid_objs[770]),/* "issuingDistributionPoint" */
-&(nid_objs[492]),/* "janetMailbox" */
-&(nid_objs[150]),/* "keyBag" */
-&(nid_objs[83]),/* "keyUsage" */
-&(nid_objs[477]),/* "lastModifiedBy" */
-&(nid_objs[476]),/* "lastModifiedTime" */
-&(nid_objs[157]),/* "localKeyID" */
-&(nid_objs[480]),/* "mXRecord" */
-&(nid_objs[460]),/* "mail" */
-&(nid_objs[493]),/* "mailPreferenceOption" */
-&(nid_objs[467]),/* "manager" */
-&(nid_objs[809]),/* "md_gost94" */
-&(nid_objs[875]),/* "member" */
-&(nid_objs[182]),/* "member-body" */
-&(nid_objs[51]),/* "messageDigest" */
-&(nid_objs[383]),/* "mgmt" */
-&(nid_objs[504]),/* "mime-mhs" */
-&(nid_objs[506]),/* "mime-mhs-bodies" */
-&(nid_objs[505]),/* "mime-mhs-headings" */
-&(nid_objs[488]),/* "mobileTelephoneNumber" */
-&(nid_objs[136]),/* "msCTLSign" */
-&(nid_objs[135]),/* "msCodeCom" */
-&(nid_objs[134]),/* "msCodeInd" */
-&(nid_objs[138]),/* "msEFS" */
-&(nid_objs[171]),/* "msExtReq" */
-&(nid_objs[137]),/* "msSGC" */
-&(nid_objs[648]),/* "msSmartcardLogin" */
-&(nid_objs[649]),/* "msUPN" */
-&(nid_objs[481]),/* "nSRecord" */
-&(nid_objs[173]),/* "name" */
-&(nid_objs[666]),/* "nameConstraints" */
-&(nid_objs[369]),/* "noCheck" */
-&(nid_objs[403]),/* "noRevAvail" */
-&(nid_objs[72]),/* "nsBaseUrl" */
-&(nid_objs[76]),/* "nsCaPolicyUrl" */
-&(nid_objs[74]),/* "nsCaRevocationUrl" */
-&(nid_objs[58]),/* "nsCertExt" */
-&(nid_objs[79]),/* "nsCertSequence" */
-&(nid_objs[71]),/* "nsCertType" */
-&(nid_objs[78]),/* "nsComment" */
-&(nid_objs[59]),/* "nsDataType" */
-&(nid_objs[75]),/* "nsRenewalUrl" */
-&(nid_objs[73]),/* "nsRevocationUrl" */
-&(nid_objs[139]),/* "nsSGC" */
-&(nid_objs[77]),/* "nsSslServerName" */
-&(nid_objs[681]),/* "onBasis" */
-&(nid_objs[491]),/* "organizationalStatus" */
-&(nid_objs[475]),/* "otherMailbox" */
-&(nid_objs[876]),/* "owner" */
-&(nid_objs[489]),/* "pagerTelephoneNumber" */
-&(nid_objs[374]),/* "path" */
-&(nid_objs[112]),/* "pbeWithMD5AndCast5CBC" */
-&(nid_objs[499]),/* "personalSignature" */
-&(nid_objs[487]),/* "personalTitle" */
-&(nid_objs[464]),/* "photo" */
-&(nid_objs[863]),/* "physicalDeliveryOfficeName" */
-&(nid_objs[437]),/* "pilot" */
-&(nid_objs[439]),/* "pilotAttributeSyntax" */
-&(nid_objs[438]),/* "pilotAttributeType" */
-&(nid_objs[479]),/* "pilotAttributeType27" */
-&(nid_objs[456]),/* "pilotDSA" */
-&(nid_objs[441]),/* "pilotGroups" */
-&(nid_objs[444]),/* "pilotObject" */
-&(nid_objs[440]),/* "pilotObjectClass" */
-&(nid_objs[455]),/* "pilotOrganization" */
-&(nid_objs[445]),/* "pilotPerson" */
-&(nid_objs[ 2]),/* "pkcs" */
-&(nid_objs[186]),/* "pkcs1" */
-&(nid_objs[27]),/* "pkcs3" */
-&(nid_objs[187]),/* "pkcs5" */
-&(nid_objs[20]),/* "pkcs7" */
-&(nid_objs[21]),/* "pkcs7-data" */
-&(nid_objs[25]),/* "pkcs7-digestData" */
-&(nid_objs[26]),/* "pkcs7-encryptedData" */
-&(nid_objs[23]),/* "pkcs7-envelopedData" */
-&(nid_objs[24]),/* "pkcs7-signedAndEnvelopedData" */
-&(nid_objs[22]),/* "pkcs7-signedData" */
-&(nid_objs[151]),/* "pkcs8ShroudedKeyBag" */
-&(nid_objs[47]),/* "pkcs9" */
-&(nid_objs[401]),/* "policyConstraints" */
-&(nid_objs[747]),/* "policyMappings" */
-&(nid_objs[862]),/* "postOfficeBox" */
-&(nid_objs[861]),/* "postalAddress" */
-&(nid_objs[661]),/* "postalCode" */
-&(nid_objs[683]),/* "ppBasis" */
-&(nid_objs[872]),/* "preferredDeliveryMethod" */
-&(nid_objs[873]),/* "presentationAddress" */
-&(nid_objs[816]),/* "prf-gostr3411-94" */
-&(nid_objs[406]),/* "prime-field" */
-&(nid_objs[409]),/* "prime192v1" */
-&(nid_objs[410]),/* "prime192v2" */
-&(nid_objs[411]),/* "prime192v3" */
-&(nid_objs[412]),/* "prime239v1" */
-&(nid_objs[413]),/* "prime239v2" */
-&(nid_objs[414]),/* "prime239v3" */
-&(nid_objs[415]),/* "prime256v1" */
-&(nid_objs[385]),/* "private" */
-&(nid_objs[84]),/* "privateKeyUsagePeriod" */
-&(nid_objs[886]),/* "protocolInformation" */
-&(nid_objs[663]),/* "proxyCertInfo" */
-&(nid_objs[510]),/* "pseudonym" */
-&(nid_objs[435]),/* "pss" */
-&(nid_objs[286]),/* "qcStatements" */
-&(nid_objs[457]),/* "qualityLabelledData" */
-&(nid_objs[450]),/* "rFC822localPart" */
-&(nid_objs[870]),/* "registeredAddress" */
-&(nid_objs[400]),/* "role" */
-&(nid_objs[877]),/* "roleOccupant" */
-&(nid_objs[448]),/* "room" */
-&(nid_objs[463]),/* "roomNumber" */
-&(nid_objs[ 6]),/* "rsaEncryption" */
-&(nid_objs[644]),/* "rsaOAEPEncryptionSET" */
-&(nid_objs[377]),/* "rsaSignature" */
-&(nid_objs[ 1]),/* "rsadsi" */
-&(nid_objs[482]),/* "sOARecord" */
-&(nid_objs[155]),/* "safeContentsBag" */
-&(nid_objs[291]),/* "sbgp-autonomousSysNum" */
-&(nid_objs[290]),/* "sbgp-ipAddrBlock" */
-&(nid_objs[292]),/* "sbgp-routerIdentifier" */
-&(nid_objs[159]),/* "sdsiCertificate" */
-&(nid_objs[859]),/* "searchGuide" */
-&(nid_objs[704]),/* "secp112r1" */
-&(nid_objs[705]),/* "secp112r2" */
-&(nid_objs[706]),/* "secp128r1" */
-&(nid_objs[707]),/* "secp128r2" */
-&(nid_objs[708]),/* "secp160k1" */
-&(nid_objs[709]),/* "secp160r1" */
-&(nid_objs[710]),/* "secp160r2" */
-&(nid_objs[711]),/* "secp192k1" */
-&(nid_objs[712]),/* "secp224k1" */
-&(nid_objs[713]),/* "secp224r1" */
-&(nid_objs[714]),/* "secp256k1" */
-&(nid_objs[715]),/* "secp384r1" */
-&(nid_objs[716]),/* "secp521r1" */
-&(nid_objs[154]),/* "secretBag" */
-&(nid_objs[474]),/* "secretary" */
-&(nid_objs[717]),/* "sect113r1" */
-&(nid_objs[718]),/* "sect113r2" */
-&(nid_objs[719]),/* "sect131r1" */
-&(nid_objs[720]),/* "sect131r2" */
-&(nid_objs[721]),/* "sect163k1" */
-&(nid_objs[722]),/* "sect163r1" */
-&(nid_objs[723]),/* "sect163r2" */
-&(nid_objs[724]),/* "sect193r1" */
-&(nid_objs[725]),/* "sect193r2" */
-&(nid_objs[726]),/* "sect233k1" */
-&(nid_objs[727]),/* "sect233r1" */
-&(nid_objs[728]),/* "sect239k1" */
-&(nid_objs[729]),/* "sect283k1" */
-&(nid_objs[730]),/* "sect283r1" */
-&(nid_objs[731]),/* "sect409k1" */
-&(nid_objs[732]),/* "sect409r1" */
-&(nid_objs[733]),/* "sect571k1" */
-&(nid_objs[734]),/* "sect571r1" */
-&(nid_objs[386]),/* "security" */
-&(nid_objs[878]),/* "seeAlso" */
-&(nid_objs[394]),/* "selected-attribute-types" */
-&(nid_objs[105]),/* "serialNumber" */
-&(nid_objs[129]),/* "serverAuth" */
-&(nid_objs[371]),/* "serviceLocator" */
-&(nid_objs[625]),/* "set-addPolicy" */
-&(nid_objs[515]),/* "set-attr" */
-&(nid_objs[518]),/* "set-brand" */
-&(nid_objs[638]),/* "set-brand-AmericanExpress" */
-&(nid_objs[637]),/* "set-brand-Diners" */
-&(nid_objs[636]),/* "set-brand-IATA-ATA" */
-&(nid_objs[639]),/* "set-brand-JCB" */
-&(nid_objs[641]),/* "set-brand-MasterCard" */
-&(nid_objs[642]),/* "set-brand-Novus" */
-&(nid_objs[640]),/* "set-brand-Visa" */
-&(nid_objs[517]),/* "set-certExt" */
-&(nid_objs[513]),/* "set-ctype" */
-&(nid_objs[514]),/* "set-msgExt" */
-&(nid_objs[516]),/* "set-policy" */
-&(nid_objs[607]),/* "set-policy-root" */
-&(nid_objs[624]),/* "set-rootKeyThumb" */
-&(nid_objs[620]),/* "setAttr-Cert" */
-&(nid_objs[631]),/* "setAttr-GenCryptgrm" */
-&(nid_objs[623]),/* "setAttr-IssCap" */
-&(nid_objs[628]),/* "setAttr-IssCap-CVM" */
-&(nid_objs[630]),/* "setAttr-IssCap-Sig" */
-&(nid_objs[629]),/* "setAttr-IssCap-T2" */
-&(nid_objs[621]),/* "setAttr-PGWYcap" */
-&(nid_objs[635]),/* "setAttr-SecDevSig" */
-&(nid_objs[632]),/* "setAttr-T2Enc" */
-&(nid_objs[633]),/* "setAttr-T2cleartxt" */
-&(nid_objs[634]),/* "setAttr-TokICCsig" */
-&(nid_objs[627]),/* "setAttr-Token-B0Prime" */
-&(nid_objs[626]),/* "setAttr-Token-EMV" */
-&(nid_objs[622]),/* "setAttr-TokenType" */
-&(nid_objs[619]),/* "setCext-IssuerCapabilities" */
-&(nid_objs[615]),/* "setCext-PGWYcapabilities" */
-&(nid_objs[616]),/* "setCext-TokenIdentifier" */
-&(nid_objs[618]),/* "setCext-TokenType" */
-&(nid_objs[617]),/* "setCext-Track2Data" */
-&(nid_objs[611]),/* "setCext-cCertRequired" */
-&(nid_objs[609]),/* "setCext-certType" */
-&(nid_objs[608]),/* "setCext-hashedRoot" */
-&(nid_objs[610]),/* "setCext-merchData" */
-&(nid_objs[613]),/* "setCext-setExt" */
-&(nid_objs[614]),/* "setCext-setQualf" */
-&(nid_objs[612]),/* "setCext-tunneling" */
-&(nid_objs[540]),/* "setct-AcqCardCodeMsg" */
-&(nid_objs[576]),/* "setct-AcqCardCodeMsgTBE" */
-&(nid_objs[570]),/* "setct-AuthReqTBE" */
-&(nid_objs[534]),/* "setct-AuthReqTBS" */
-&(nid_objs[527]),/* "setct-AuthResBaggage" */
-&(nid_objs[571]),/* "setct-AuthResTBE" */
-&(nid_objs[572]),/* "setct-AuthResTBEX" */
-&(nid_objs[535]),/* "setct-AuthResTBS" */
-&(nid_objs[536]),/* "setct-AuthResTBSX" */
-&(nid_objs[528]),/* "setct-AuthRevReqBaggage" */
-&(nid_objs[577]),/* "setct-AuthRevReqTBE" */
-&(nid_objs[541]),/* "setct-AuthRevReqTBS" */
-&(nid_objs[529]),/* "setct-AuthRevResBaggage" */
-&(nid_objs[542]),/* "setct-AuthRevResData" */
-&(nid_objs[578]),/* "setct-AuthRevResTBE" */
-&(nid_objs[579]),/* "setct-AuthRevResTBEB" */
-&(nid_objs[543]),/* "setct-AuthRevResTBS" */
-&(nid_objs[573]),/* "setct-AuthTokenTBE" */
-&(nid_objs[537]),/* "setct-AuthTokenTBS" */
-&(nid_objs[600]),/* "setct-BCIDistributionTBS" */
-&(nid_objs[558]),/* "setct-BatchAdminReqData" */
-&(nid_objs[592]),/* "setct-BatchAdminReqTBE" */
-&(nid_objs[559]),/* "setct-BatchAdminResData" */
-&(nid_objs[593]),/* "setct-BatchAdminResTBE" */
-&(nid_objs[599]),/* "setct-CRLNotificationResTBS" */
-&(nid_objs[598]),/* "setct-CRLNotificationTBS" */
-&(nid_objs[580]),/* "setct-CapReqTBE" */
-&(nid_objs[581]),/* "setct-CapReqTBEX" */
-&(nid_objs[544]),/* "setct-CapReqTBS" */
-&(nid_objs[545]),/* "setct-CapReqTBSX" */
-&(nid_objs[546]),/* "setct-CapResData" */
-&(nid_objs[582]),/* "setct-CapResTBE" */
-&(nid_objs[583]),/* "setct-CapRevReqTBE" */
-&(nid_objs[584]),/* "setct-CapRevReqTBEX" */
-&(nid_objs[547]),/* "setct-CapRevReqTBS" */
-&(nid_objs[548]),/* "setct-CapRevReqTBSX" */
-&(nid_objs[549]),/* "setct-CapRevResData" */
-&(nid_objs[585]),/* "setct-CapRevResTBE" */
-&(nid_objs[538]),/* "setct-CapTokenData" */
-&(nid_objs[530]),/* "setct-CapTokenSeq" */
-&(nid_objs[574]),/* "setct-CapTokenTBE" */
-&(nid_objs[575]),/* "setct-CapTokenTBEX" */
-&(nid_objs[539]),/* "setct-CapTokenTBS" */
-&(nid_objs[560]),/* "setct-CardCInitResTBS" */
-&(nid_objs[566]),/* "setct-CertInqReqTBS" */
-&(nid_objs[563]),/* "setct-CertReqData" */
-&(nid_objs[595]),/* "setct-CertReqTBE" */
-&(nid_objs[596]),/* "setct-CertReqTBEX" */
-&(nid_objs[564]),/* "setct-CertReqTBS" */
-&(nid_objs[565]),/* "setct-CertResData" */
-&(nid_objs[597]),/* "setct-CertResTBE" */
-&(nid_objs[586]),/* "setct-CredReqTBE" */
-&(nid_objs[587]),/* "setct-CredReqTBEX" */
-&(nid_objs[550]),/* "setct-CredReqTBS" */
-&(nid_objs[551]),/* "setct-CredReqTBSX" */
-&(nid_objs[552]),/* "setct-CredResData" */
-&(nid_objs[588]),/* "setct-CredResTBE" */
-&(nid_objs[589]),/* "setct-CredRevReqTBE" */
-&(nid_objs[590]),/* "setct-CredRevReqTBEX" */
-&(nid_objs[553]),/* "setct-CredRevReqTBS" */
-&(nid_objs[554]),/* "setct-CredRevReqTBSX" */
-&(nid_objs[555]),/* "setct-CredRevResData" */
-&(nid_objs[591]),/* "setct-CredRevResTBE" */
-&(nid_objs[567]),/* "setct-ErrorTBS" */
-&(nid_objs[526]),/* "setct-HODInput" */
-&(nid_objs[561]),/* "setct-MeAqCInitResTBS" */
-&(nid_objs[522]),/* "setct-OIData" */
-&(nid_objs[519]),/* "setct-PANData" */
-&(nid_objs[521]),/* "setct-PANOnly" */
-&(nid_objs[520]),/* "setct-PANToken" */
-&(nid_objs[556]),/* "setct-PCertReqData" */
-&(nid_objs[557]),/* "setct-PCertResTBS" */
-&(nid_objs[523]),/* "setct-PI" */
-&(nid_objs[532]),/* "setct-PI-TBS" */
-&(nid_objs[524]),/* "setct-PIData" */
-&(nid_objs[525]),/* "setct-PIDataUnsigned" */
-&(nid_objs[568]),/* "setct-PIDualSignedTBE" */
-&(nid_objs[569]),/* "setct-PIUnsignedTBE" */
-&(nid_objs[531]),/* "setct-PInitResData" */
-&(nid_objs[533]),/* "setct-PResData" */
-&(nid_objs[594]),/* "setct-RegFormReqTBE" */
-&(nid_objs[562]),/* "setct-RegFormResTBS" */
-&(nid_objs[606]),/* "setext-cv" */
-&(nid_objs[601]),/* "setext-genCrypt" */
-&(nid_objs[602]),/* "setext-miAuth" */
-&(nid_objs[604]),/* "setext-pinAny" */
-&(nid_objs[603]),/* "setext-pinSecure" */
-&(nid_objs[605]),/* "setext-track2" */
-&(nid_objs[52]),/* "signingTime" */
-&(nid_objs[454]),/* "simpleSecurityObject" */
-&(nid_objs[496]),/* "singleLevelQuality" */
-&(nid_objs[387]),/* "snmpv2" */
-&(nid_objs[660]),/* "street" */
-&(nid_objs[85]),/* "subjectAltName" */
-&(nid_objs[769]),/* "subjectDirectoryAttributes" */
-&(nid_objs[398]),/* "subjectInfoAccess" */
-&(nid_objs[82]),/* "subjectKeyIdentifier" */
-&(nid_objs[498]),/* "subtreeMaximumQuality" */
-&(nid_objs[497]),/* "subtreeMinimumQuality" */
-&(nid_objs[890]),/* "supportedAlgorithms" */
-&(nid_objs[874]),/* "supportedApplicationContext" */
-&(nid_objs[402]),/* "targetInformation" */
-&(nid_objs[864]),/* "telephoneNumber" */
-&(nid_objs[866]),/* "teletexTerminalIdentifier" */
-&(nid_objs[865]),/* "telexNumber" */
-&(nid_objs[459]),/* "textEncodedORAddress" */
-&(nid_objs[293]),/* "textNotice" */
-&(nid_objs[133]),/* "timeStamping" */
-&(nid_objs[106]),/* "title" */
-&(nid_objs[682]),/* "tpBasis" */
-&(nid_objs[375]),/* "trustRoot" */
-&(nid_objs[436]),/* "ucl" */
-&(nid_objs[888]),/* "uniqueMember" */
-&(nid_objs[55]),/* "unstructuredAddress" */
-&(nid_objs[49]),/* "unstructuredName" */
-&(nid_objs[880]),/* "userCertificate" */
-&(nid_objs[465]),/* "userClass" */
-&(nid_objs[879]),/* "userPassword" */
-&(nid_objs[373]),/* "valid" */
-&(nid_objs[678]),/* "wap" */
-&(nid_objs[679]),/* "wap-wsg" */
-&(nid_objs[735]),/* "wap-wsg-idm-ecid-wtls1" */
-&(nid_objs[743]),/* "wap-wsg-idm-ecid-wtls10" */
-&(nid_objs[744]),/* "wap-wsg-idm-ecid-wtls11" */
-&(nid_objs[745]),/* "wap-wsg-idm-ecid-wtls12" */
-&(nid_objs[736]),/* "wap-wsg-idm-ecid-wtls3" */
-&(nid_objs[737]),/* "wap-wsg-idm-ecid-wtls4" */
-&(nid_objs[738]),/* "wap-wsg-idm-ecid-wtls5" */
-&(nid_objs[739]),/* "wap-wsg-idm-ecid-wtls6" */
-&(nid_objs[740]),/* "wap-wsg-idm-ecid-wtls7" */
-&(nid_objs[741]),/* "wap-wsg-idm-ecid-wtls8" */
-&(nid_objs[742]),/* "wap-wsg-idm-ecid-wtls9" */
-&(nid_objs[804]),/* "whirlpool" */
-&(nid_objs[868]),/* "x121Address" */
-&(nid_objs[503]),/* "x500UniqueIdentifier" */
-&(nid_objs[158]),/* "x509Certificate" */
-&(nid_objs[160]),/* "x509Crl" */
+static const unsigned int sn_objs[NUM_SN]={
+364,	/* "AD_DVCS" */
+419,	/* "AES-128-CBC" */
+916,	/* "AES-128-CBC-HMAC-SHA1" */
+421,	/* "AES-128-CFB" */
+650,	/* "AES-128-CFB1" */
+653,	/* "AES-128-CFB8" */
+904,	/* "AES-128-CTR" */
+418,	/* "AES-128-ECB" */
+420,	/* "AES-128-OFB" */
+913,	/* "AES-128-XTS" */
+423,	/* "AES-192-CBC" */
+917,	/* "AES-192-CBC-HMAC-SHA1" */
+425,	/* "AES-192-CFB" */
+651,	/* "AES-192-CFB1" */
+654,	/* "AES-192-CFB8" */
+905,	/* "AES-192-CTR" */
+422,	/* "AES-192-ECB" */
+424,	/* "AES-192-OFB" */
+427,	/* "AES-256-CBC" */
+918,	/* "AES-256-CBC-HMAC-SHA1" */
+429,	/* "AES-256-CFB" */
+652,	/* "AES-256-CFB1" */
+655,	/* "AES-256-CFB8" */
+906,	/* "AES-256-CTR" */
+426,	/* "AES-256-ECB" */
+428,	/* "AES-256-OFB" */
+914,	/* "AES-256-XTS" */
+91,	/* "BF-CBC" */
+93,	/* "BF-CFB" */
+92,	/* "BF-ECB" */
+94,	/* "BF-OFB" */
+14,	/* "C" */
+751,	/* "CAMELLIA-128-CBC" */
+757,	/* "CAMELLIA-128-CFB" */
+760,	/* "CAMELLIA-128-CFB1" */
+763,	/* "CAMELLIA-128-CFB8" */
+754,	/* "CAMELLIA-128-ECB" */
+766,	/* "CAMELLIA-128-OFB" */
+752,	/* "CAMELLIA-192-CBC" */
+758,	/* "CAMELLIA-192-CFB" */
+761,	/* "CAMELLIA-192-CFB1" */
+764,	/* "CAMELLIA-192-CFB8" */
+755,	/* "CAMELLIA-192-ECB" */
+767,	/* "CAMELLIA-192-OFB" */
+753,	/* "CAMELLIA-256-CBC" */
+759,	/* "CAMELLIA-256-CFB" */
+762,	/* "CAMELLIA-256-CFB1" */
+765,	/* "CAMELLIA-256-CFB8" */
+756,	/* "CAMELLIA-256-ECB" */
+768,	/* "CAMELLIA-256-OFB" */
+108,	/* "CAST5-CBC" */
+110,	/* "CAST5-CFB" */
+109,	/* "CAST5-ECB" */
+111,	/* "CAST5-OFB" */
+894,	/* "CMAC" */
+13,	/* "CN" */
+141,	/* "CRLReason" */
+417,	/* "CSPName" */
+367,	/* "CrlID" */
+391,	/* "DC" */
+31,	/* "DES-CBC" */
+643,	/* "DES-CDMF" */
+30,	/* "DES-CFB" */
+656,	/* "DES-CFB1" */
+657,	/* "DES-CFB8" */
+29,	/* "DES-ECB" */
+32,	/* "DES-EDE" */
+43,	/* "DES-EDE-CBC" */
+60,	/* "DES-EDE-CFB" */
+62,	/* "DES-EDE-OFB" */
+33,	/* "DES-EDE3" */
+44,	/* "DES-EDE3-CBC" */
+61,	/* "DES-EDE3-CFB" */
+658,	/* "DES-EDE3-CFB1" */
+659,	/* "DES-EDE3-CFB8" */
+63,	/* "DES-EDE3-OFB" */
+45,	/* "DES-OFB" */
+80,	/* "DESX-CBC" */
+380,	/* "DOD" */
+116,	/* "DSA" */
+66,	/* "DSA-SHA" */
+113,	/* "DSA-SHA1" */
+70,	/* "DSA-SHA1-old" */
+67,	/* "DSA-old" */
+297,	/* "DVCS" */
+99,	/* "GN" */
+855,	/* "HMAC" */
+780,	/* "HMAC-MD5" */
+781,	/* "HMAC-SHA1" */
+381,	/* "IANA" */
+34,	/* "IDEA-CBC" */
+35,	/* "IDEA-CFB" */
+36,	/* "IDEA-ECB" */
+46,	/* "IDEA-OFB" */
+181,	/* "ISO" */
+183,	/* "ISO-US" */
+645,	/* "ITU-T" */
+646,	/* "JOINT-ISO-ITU-T" */
+773,	/* "KISA" */
+15,	/* "L" */
+856,	/* "LocalKeySet" */
+ 3,	/* "MD2" */
+257,	/* "MD4" */
+ 4,	/* "MD5" */
+114,	/* "MD5-SHA1" */
+95,	/* "MDC2" */
+911,	/* "MGF1" */
+388,	/* "Mail" */
+393,	/* "NULL" */
+404,	/* "NULL" */
+57,	/* "Netscape" */
+366,	/* "Nonce" */
+17,	/* "O" */
+178,	/* "OCSP" */
+180,	/* "OCSPSigning" */
+379,	/* "ORG" */
+18,	/* "OU" */
+749,	/* "Oakley-EC2N-3" */
+750,	/* "Oakley-EC2N-4" */
+ 9,	/* "PBE-MD2-DES" */
+168,	/* "PBE-MD2-RC2-64" */
+10,	/* "PBE-MD5-DES" */
+169,	/* "PBE-MD5-RC2-64" */
+147,	/* "PBE-SHA1-2DES" */
+146,	/* "PBE-SHA1-3DES" */
+170,	/* "PBE-SHA1-DES" */
+148,	/* "PBE-SHA1-RC2-128" */
+149,	/* "PBE-SHA1-RC2-40" */
+68,	/* "PBE-SHA1-RC2-64" */
+144,	/* "PBE-SHA1-RC4-128" */
+145,	/* "PBE-SHA1-RC4-40" */
+161,	/* "PBES2" */
+69,	/* "PBKDF2" */
+162,	/* "PBMAC1" */
+127,	/* "PKIX" */
+98,	/* "RC2-40-CBC" */
+166,	/* "RC2-64-CBC" */
+37,	/* "RC2-CBC" */
+39,	/* "RC2-CFB" */
+38,	/* "RC2-ECB" */
+40,	/* "RC2-OFB" */
+ 5,	/* "RC4" */
+97,	/* "RC4-40" */
+915,	/* "RC4-HMAC-MD5" */
+120,	/* "RC5-CBC" */
+122,	/* "RC5-CFB" */
+121,	/* "RC5-ECB" */
+123,	/* "RC5-OFB" */
+117,	/* "RIPEMD160" */
+124,	/* "RLE" */
+19,	/* "RSA" */
+ 7,	/* "RSA-MD2" */
+396,	/* "RSA-MD4" */
+ 8,	/* "RSA-MD5" */
+96,	/* "RSA-MDC2" */
+104,	/* "RSA-NP-MD5" */
+119,	/* "RSA-RIPEMD160" */
+42,	/* "RSA-SHA" */
+65,	/* "RSA-SHA1" */
+115,	/* "RSA-SHA1-2" */
+671,	/* "RSA-SHA224" */
+668,	/* "RSA-SHA256" */
+669,	/* "RSA-SHA384" */
+670,	/* "RSA-SHA512" */
+919,	/* "RSAES-OAEP" */
+912,	/* "RSASSA-PSS" */
+777,	/* "SEED-CBC" */
+779,	/* "SEED-CFB" */
+776,	/* "SEED-ECB" */
+778,	/* "SEED-OFB" */
+41,	/* "SHA" */
+64,	/* "SHA1" */
+675,	/* "SHA224" */
+672,	/* "SHA256" */
+673,	/* "SHA384" */
+674,	/* "SHA512" */
+188,	/* "SMIME" */
+167,	/* "SMIME-CAPS" */
+100,	/* "SN" */
+16,	/* "ST" */
+143,	/* "SXNetID" */
+458,	/* "UID" */
+ 0,	/* "UNDEF" */
+11,	/* "X500" */
+378,	/* "X500algorithms" */
+12,	/* "X509" */
+184,	/* "X9-57" */
+185,	/* "X9cm" */
+125,	/* "ZLIB" */
+478,	/* "aRecord" */
+289,	/* "aaControls" */
+287,	/* "ac-auditEntity" */
+397,	/* "ac-proxying" */
+288,	/* "ac-targeting" */
+368,	/* "acceptableResponses" */
+446,	/* "account" */
+363,	/* "ad_timestamping" */
+376,	/* "algorithm" */
+405,	/* "ansi-X9-62" */
+910,	/* "anyExtendedKeyUsage" */
+746,	/* "anyPolicy" */
+370,	/* "archiveCutoff" */
+484,	/* "associatedDomain" */
+485,	/* "associatedName" */
+501,	/* "audio" */
+177,	/* "authorityInfoAccess" */
+90,	/* "authorityKeyIdentifier" */
+882,	/* "authorityRevocationList" */
+87,	/* "basicConstraints" */
+365,	/* "basicOCSPResponse" */
+285,	/* "biometricInfo" */
+494,	/* "buildingName" */
+860,	/* "businessCategory" */
+691,	/* "c2onb191v4" */
+692,	/* "c2onb191v5" */
+697,	/* "c2onb239v4" */
+698,	/* "c2onb239v5" */
+684,	/* "c2pnb163v1" */
+685,	/* "c2pnb163v2" */
+686,	/* "c2pnb163v3" */
+687,	/* "c2pnb176v1" */
+693,	/* "c2pnb208w1" */
+699,	/* "c2pnb272w1" */
+700,	/* "c2pnb304w1" */
+702,	/* "c2pnb368w1" */
+688,	/* "c2tnb191v1" */
+689,	/* "c2tnb191v2" */
+690,	/* "c2tnb191v3" */
+694,	/* "c2tnb239v1" */
+695,	/* "c2tnb239v2" */
+696,	/* "c2tnb239v3" */
+701,	/* "c2tnb359v1" */
+703,	/* "c2tnb431r1" */
+881,	/* "cACertificate" */
+483,	/* "cNAMERecord" */
+179,	/* "caIssuers" */
+785,	/* "caRepository" */
+443,	/* "caseIgnoreIA5StringSyntax" */
+152,	/* "certBag" */
+677,	/* "certicom-arc" */
+771,	/* "certificateIssuer" */
+89,	/* "certificatePolicies" */
+883,	/* "certificateRevocationList" */
+54,	/* "challengePassword" */
+407,	/* "characteristic-two-field" */
+395,	/* "clearance" */
+130,	/* "clientAuth" */
+131,	/* "codeSigning" */
+50,	/* "contentType" */
+53,	/* "countersignature" */
+153,	/* "crlBag" */
+103,	/* "crlDistributionPoints" */
+88,	/* "crlNumber" */
+884,	/* "crossCertificatePair" */
+806,	/* "cryptocom" */
+805,	/* "cryptopro" */
+500,	/* "dITRedirect" */
+451,	/* "dNSDomain" */
+495,	/* "dSAQuality" */
+434,	/* "data" */
+390,	/* "dcobject" */
+140,	/* "deltaCRL" */
+891,	/* "deltaRevocationList" */
+107,	/* "description" */
+871,	/* "destinationIndicator" */
+28,	/* "dhKeyAgreement" */
+382,	/* "directory" */
+887,	/* "distinguishedName" */
+892,	/* "dmdName" */
+174,	/* "dnQualifier" */
+447,	/* "document" */
+471,	/* "documentAuthor" */
+468,	/* "documentIdentifier" */
+472,	/* "documentLocation" */
+502,	/* "documentPublisher" */
+449,	/* "documentSeries" */
+469,	/* "documentTitle" */
+470,	/* "documentVersion" */
+392,	/* "domain" */
+452,	/* "domainRelatedObject" */
+802,	/* "dsa_with_SHA224" */
+803,	/* "dsa_with_SHA256" */
+791,	/* "ecdsa-with-Recommended" */
+416,	/* "ecdsa-with-SHA1" */
+793,	/* "ecdsa-with-SHA224" */
+794,	/* "ecdsa-with-SHA256" */
+795,	/* "ecdsa-with-SHA384" */
+796,	/* "ecdsa-with-SHA512" */
+792,	/* "ecdsa-with-Specified" */
+48,	/* "emailAddress" */
+132,	/* "emailProtection" */
+885,	/* "enhancedSearchGuide" */
+389,	/* "enterprises" */
+384,	/* "experimental" */
+172,	/* "extReq" */
+56,	/* "extendedCertificateAttributes" */
+126,	/* "extendedKeyUsage" */
+372,	/* "extendedStatus" */
+867,	/* "facsimileTelephoneNumber" */
+462,	/* "favouriteDrink" */
+857,	/* "freshestCRL" */
+453,	/* "friendlyCountry" */
+490,	/* "friendlyCountryName" */
+156,	/* "friendlyName" */
+509,	/* "generationQualifier" */
+815,	/* "gost-mac" */
+811,	/* "gost2001" */
+851,	/* "gost2001cc" */
+813,	/* "gost89" */
+814,	/* "gost89-cnt" */
+812,	/* "gost94" */
+850,	/* "gost94cc" */
+797,	/* "hmacWithMD5" */
+163,	/* "hmacWithSHA1" */
+798,	/* "hmacWithSHA224" */
+799,	/* "hmacWithSHA256" */
+800,	/* "hmacWithSHA384" */
+801,	/* "hmacWithSHA512" */
+432,	/* "holdInstructionCallIssuer" */
+430,	/* "holdInstructionCode" */
+431,	/* "holdInstructionNone" */
+433,	/* "holdInstructionReject" */
+486,	/* "homePostalAddress" */
+473,	/* "homeTelephoneNumber" */
+466,	/* "host" */
+889,	/* "houseIdentifier" */
+442,	/* "iA5StringSyntax" */
+783,	/* "id-DHBasedMac" */
+824,	/* "id-Gost28147-89-CryptoPro-A-ParamSet" */
+825,	/* "id-Gost28147-89-CryptoPro-B-ParamSet" */
+826,	/* "id-Gost28147-89-CryptoPro-C-ParamSet" */
+827,	/* "id-Gost28147-89-CryptoPro-D-ParamSet" */
+819,	/* "id-Gost28147-89-CryptoPro-KeyMeshing" */
+829,	/* "id-Gost28147-89-CryptoPro-Oscar-1-0-ParamSet" */
+828,	/* "id-Gost28147-89-CryptoPro-Oscar-1-1-ParamSet" */
+830,	/* "id-Gost28147-89-CryptoPro-RIC-1-ParamSet" */
+820,	/* "id-Gost28147-89-None-KeyMeshing" */
+823,	/* "id-Gost28147-89-TestParamSet" */
+849,	/* "id-Gost28147-89-cc" */
+840,	/* "id-GostR3410-2001-CryptoPro-A-ParamSet" */
+841,	/* "id-GostR3410-2001-CryptoPro-B-ParamSet" */
+842,	/* "id-GostR3410-2001-CryptoPro-C-ParamSet" */
+843,	/* "id-GostR3410-2001-CryptoPro-XchA-ParamSet" */
+844,	/* "id-GostR3410-2001-CryptoPro-XchB-ParamSet" */
+854,	/* "id-GostR3410-2001-ParamSet-cc" */
+839,	/* "id-GostR3410-2001-TestParamSet" */
+817,	/* "id-GostR3410-2001DH" */
+832,	/* "id-GostR3410-94-CryptoPro-A-ParamSet" */
+833,	/* "id-GostR3410-94-CryptoPro-B-ParamSet" */
+834,	/* "id-GostR3410-94-CryptoPro-C-ParamSet" */
+835,	/* "id-GostR3410-94-CryptoPro-D-ParamSet" */
+836,	/* "id-GostR3410-94-CryptoPro-XchA-ParamSet" */
+837,	/* "id-GostR3410-94-CryptoPro-XchB-ParamSet" */
+838,	/* "id-GostR3410-94-CryptoPro-XchC-ParamSet" */
+831,	/* "id-GostR3410-94-TestParamSet" */
+845,	/* "id-GostR3410-94-a" */
+846,	/* "id-GostR3410-94-aBis" */
+847,	/* "id-GostR3410-94-b" */
+848,	/* "id-GostR3410-94-bBis" */
+818,	/* "id-GostR3410-94DH" */
+822,	/* "id-GostR3411-94-CryptoProParamSet" */
+821,	/* "id-GostR3411-94-TestParamSet" */
+807,	/* "id-GostR3411-94-with-GostR3410-2001" */
+853,	/* "id-GostR3411-94-with-GostR3410-2001-cc" */
+808,	/* "id-GostR3411-94-with-GostR3410-94" */
+852,	/* "id-GostR3411-94-with-GostR3410-94-cc" */
+810,	/* "id-HMACGostR3411-94" */
+782,	/* "id-PasswordBasedMAC" */
+266,	/* "id-aca" */
+355,	/* "id-aca-accessIdentity" */
+354,	/* "id-aca-authenticationInfo" */
+356,	/* "id-aca-chargingIdentity" */
+399,	/* "id-aca-encAttrs" */
+357,	/* "id-aca-group" */
+358,	/* "id-aca-role" */
+176,	/* "id-ad" */
+896,	/* "id-aes128-CCM" */
+895,	/* "id-aes128-GCM" */
+788,	/* "id-aes128-wrap" */
+897,	/* "id-aes128-wrap-pad" */
+899,	/* "id-aes192-CCM" */
+898,	/* "id-aes192-GCM" */
+789,	/* "id-aes192-wrap" */
+900,	/* "id-aes192-wrap-pad" */
+902,	/* "id-aes256-CCM" */
+901,	/* "id-aes256-GCM" */
+790,	/* "id-aes256-wrap" */
+903,	/* "id-aes256-wrap-pad" */
+262,	/* "id-alg" */
+893,	/* "id-alg-PWRI-KEK" */
+323,	/* "id-alg-des40" */
+326,	/* "id-alg-dh-pop" */
+325,	/* "id-alg-dh-sig-hmac-sha1" */
+324,	/* "id-alg-noSignature" */
+907,	/* "id-camellia128-wrap" */
+908,	/* "id-camellia192-wrap" */
+909,	/* "id-camellia256-wrap" */
+268,	/* "id-cct" */
+361,	/* "id-cct-PKIData" */
+362,	/* "id-cct-PKIResponse" */
+360,	/* "id-cct-crs" */
+81,	/* "id-ce" */
+680,	/* "id-characteristic-two-basis" */
+263,	/* "id-cmc" */
+334,	/* "id-cmc-addExtensions" */
+346,	/* "id-cmc-confirmCertAcceptance" */
+330,	/* "id-cmc-dataReturn" */
+336,	/* "id-cmc-decryptedPOP" */
+335,	/* "id-cmc-encryptedPOP" */
+339,	/* "id-cmc-getCRL" */
+338,	/* "id-cmc-getCert" */
+328,	/* "id-cmc-identification" */
+329,	/* "id-cmc-identityProof" */
+337,	/* "id-cmc-lraPOPWitness" */
+344,	/* "id-cmc-popLinkRandom" */
+345,	/* "id-cmc-popLinkWitness" */
+343,	/* "id-cmc-queryPending" */
+333,	/* "id-cmc-recipientNonce" */
+341,	/* "id-cmc-regInfo" */
+342,	/* "id-cmc-responseInfo" */
+340,	/* "id-cmc-revokeRequest" */
+332,	/* "id-cmc-senderNonce" */
+327,	/* "id-cmc-statusInfo" */
+331,	/* "id-cmc-transactionId" */
+787,	/* "id-ct-asciiTextWithCRLF" */
+408,	/* "id-ecPublicKey" */
+508,	/* "id-hex-multipart-message" */
+507,	/* "id-hex-partial-message" */
+260,	/* "id-it" */
+302,	/* "id-it-caKeyUpdateInfo" */
+298,	/* "id-it-caProtEncCert" */
+311,	/* "id-it-confirmWaitTime" */
+303,	/* "id-it-currentCRL" */
+300,	/* "id-it-encKeyPairTypes" */
+310,	/* "id-it-implicitConfirm" */
+308,	/* "id-it-keyPairParamRep" */
+307,	/* "id-it-keyPairParamReq" */
+312,	/* "id-it-origPKIMessage" */
+301,	/* "id-it-preferredSymmAlg" */
+309,	/* "id-it-revPassphrase" */
+299,	/* "id-it-signKeyPairTypes" */
+305,	/* "id-it-subscriptionRequest" */
+306,	/* "id-it-subscriptionResponse" */
+784,	/* "id-it-suppLangTags" */
+304,	/* "id-it-unsupportedOIDs" */
+128,	/* "id-kp" */
+280,	/* "id-mod-attribute-cert" */
+274,	/* "id-mod-cmc" */
+277,	/* "id-mod-cmp" */
+284,	/* "id-mod-cmp2000" */
+273,	/* "id-mod-crmf" */
+283,	/* "id-mod-dvcs" */
+275,	/* "id-mod-kea-profile-88" */
+276,	/* "id-mod-kea-profile-93" */
+282,	/* "id-mod-ocsp" */
+278,	/* "id-mod-qualified-cert-88" */
+279,	/* "id-mod-qualified-cert-93" */
+281,	/* "id-mod-timestamp-protocol" */
+264,	/* "id-on" */
+858,	/* "id-on-permanentIdentifier" */
+347,	/* "id-on-personalData" */
+265,	/* "id-pda" */
+352,	/* "id-pda-countryOfCitizenship" */
+353,	/* "id-pda-countryOfResidence" */
+348,	/* "id-pda-dateOfBirth" */
+351,	/* "id-pda-gender" */
+349,	/* "id-pda-placeOfBirth" */
+175,	/* "id-pe" */
+261,	/* "id-pkip" */
+258,	/* "id-pkix-mod" */
+269,	/* "id-pkix1-explicit-88" */
+271,	/* "id-pkix1-explicit-93" */
+270,	/* "id-pkix1-implicit-88" */
+272,	/* "id-pkix1-implicit-93" */
+662,	/* "id-ppl" */
+664,	/* "id-ppl-anyLanguage" */
+667,	/* "id-ppl-independent" */
+665,	/* "id-ppl-inheritAll" */
+267,	/* "id-qcs" */
+359,	/* "id-qcs-pkixQCSyntax-v1" */
+259,	/* "id-qt" */
+164,	/* "id-qt-cps" */
+165,	/* "id-qt-unotice" */
+313,	/* "id-regCtrl" */
+316,	/* "id-regCtrl-authenticator" */
+319,	/* "id-regCtrl-oldCertID" */
+318,	/* "id-regCtrl-pkiArchiveOptions" */
+317,	/* "id-regCtrl-pkiPublicationInfo" */
+320,	/* "id-regCtrl-protocolEncrKey" */
+315,	/* "id-regCtrl-regToken" */
+314,	/* "id-regInfo" */
+322,	/* "id-regInfo-certReq" */
+321,	/* "id-regInfo-utf8Pairs" */
+512,	/* "id-set" */
+191,	/* "id-smime-aa" */
+215,	/* "id-smime-aa-contentHint" */
+218,	/* "id-smime-aa-contentIdentifier" */
+221,	/* "id-smime-aa-contentReference" */
+240,	/* "id-smime-aa-dvcs-dvc" */
+217,	/* "id-smime-aa-encapContentType" */
+222,	/* "id-smime-aa-encrypKeyPref" */
+220,	/* "id-smime-aa-equivalentLabels" */
+232,	/* "id-smime-aa-ets-CertificateRefs" */
+233,	/* "id-smime-aa-ets-RevocationRefs" */
+238,	/* "id-smime-aa-ets-archiveTimeStamp" */
+237,	/* "id-smime-aa-ets-certCRLTimestamp" */
+234,	/* "id-smime-aa-ets-certValues" */
+227,	/* "id-smime-aa-ets-commitmentType" */
+231,	/* "id-smime-aa-ets-contentTimestamp" */
+236,	/* "id-smime-aa-ets-escTimeStamp" */
+230,	/* "id-smime-aa-ets-otherSigCert" */
+235,	/* "id-smime-aa-ets-revocationValues" */
+226,	/* "id-smime-aa-ets-sigPolicyId" */
+229,	/* "id-smime-aa-ets-signerAttr" */
+228,	/* "id-smime-aa-ets-signerLocation" */
+219,	/* "id-smime-aa-macValue" */
+214,	/* "id-smime-aa-mlExpandHistory" */
+216,	/* "id-smime-aa-msgSigDigest" */
+212,	/* "id-smime-aa-receiptRequest" */
+213,	/* "id-smime-aa-securityLabel" */
+239,	/* "id-smime-aa-signatureType" */
+223,	/* "id-smime-aa-signingCertificate" */
+224,	/* "id-smime-aa-smimeEncryptCerts" */
+225,	/* "id-smime-aa-timeStampToken" */
+192,	/* "id-smime-alg" */
+243,	/* "id-smime-alg-3DESwrap" */
+246,	/* "id-smime-alg-CMS3DESwrap" */
+247,	/* "id-smime-alg-CMSRC2wrap" */
+245,	/* "id-smime-alg-ESDH" */
+241,	/* "id-smime-alg-ESDHwith3DES" */
+242,	/* "id-smime-alg-ESDHwithRC2" */
+244,	/* "id-smime-alg-RC2wrap" */
+193,	/* "id-smime-cd" */
+248,	/* "id-smime-cd-ldap" */
+190,	/* "id-smime-ct" */
+210,	/* "id-smime-ct-DVCSRequestData" */
+211,	/* "id-smime-ct-DVCSResponseData" */
+208,	/* "id-smime-ct-TDTInfo" */
+207,	/* "id-smime-ct-TSTInfo" */
+205,	/* "id-smime-ct-authData" */
+786,	/* "id-smime-ct-compressedData" */
+209,	/* "id-smime-ct-contentInfo" */
+206,	/* "id-smime-ct-publishCert" */
+204,	/* "id-smime-ct-receipt" */
+195,	/* "id-smime-cti" */
+255,	/* "id-smime-cti-ets-proofOfApproval" */
+256,	/* "id-smime-cti-ets-proofOfCreation" */
+253,	/* "id-smime-cti-ets-proofOfDelivery" */
+251,	/* "id-smime-cti-ets-proofOfOrigin" */
+252,	/* "id-smime-cti-ets-proofOfReceipt" */
+254,	/* "id-smime-cti-ets-proofOfSender" */
+189,	/* "id-smime-mod" */
+196,	/* "id-smime-mod-cms" */
+197,	/* "id-smime-mod-ess" */
+202,	/* "id-smime-mod-ets-eSigPolicy-88" */
+203,	/* "id-smime-mod-ets-eSigPolicy-97" */
+200,	/* "id-smime-mod-ets-eSignature-88" */
+201,	/* "id-smime-mod-ets-eSignature-97" */
+199,	/* "id-smime-mod-msg-v3" */
+198,	/* "id-smime-mod-oid" */
+194,	/* "id-smime-spq" */
+250,	/* "id-smime-spq-ets-sqt-unotice" */
+249,	/* "id-smime-spq-ets-sqt-uri" */
+676,	/* "identified-organization" */
+461,	/* "info" */
+748,	/* "inhibitAnyPolicy" */
+101,	/* "initials" */
+647,	/* "international-organizations" */
+869,	/* "internationaliSDNNumber" */
+142,	/* "invalidityDate" */
+294,	/* "ipsecEndSystem" */
+295,	/* "ipsecTunnel" */
+296,	/* "ipsecUser" */
+86,	/* "issuerAltName" */
+770,	/* "issuingDistributionPoint" */
+492,	/* "janetMailbox" */
+150,	/* "keyBag" */
+83,	/* "keyUsage" */
+477,	/* "lastModifiedBy" */
+476,	/* "lastModifiedTime" */
+157,	/* "localKeyID" */
+480,	/* "mXRecord" */
+460,	/* "mail" */
+493,	/* "mailPreferenceOption" */
+467,	/* "manager" */
+809,	/* "md_gost94" */
+875,	/* "member" */
+182,	/* "member-body" */
+51,	/* "messageDigest" */
+383,	/* "mgmt" */
+504,	/* "mime-mhs" */
+506,	/* "mime-mhs-bodies" */
+505,	/* "mime-mhs-headings" */
+488,	/* "mobileTelephoneNumber" */
+136,	/* "msCTLSign" */
+135,	/* "msCodeCom" */
+134,	/* "msCodeInd" */
+138,	/* "msEFS" */
+171,	/* "msExtReq" */
+137,	/* "msSGC" */
+648,	/* "msSmartcardLogin" */
+649,	/* "msUPN" */
+481,	/* "nSRecord" */
+173,	/* "name" */
+666,	/* "nameConstraints" */
+369,	/* "noCheck" */
+403,	/* "noRevAvail" */
+72,	/* "nsBaseUrl" */
+76,	/* "nsCaPolicyUrl" */
+74,	/* "nsCaRevocationUrl" */
+58,	/* "nsCertExt" */
+79,	/* "nsCertSequence" */
+71,	/* "nsCertType" */
+78,	/* "nsComment" */
+59,	/* "nsDataType" */
+75,	/* "nsRenewalUrl" */
+73,	/* "nsRevocationUrl" */
+139,	/* "nsSGC" */
+77,	/* "nsSslServerName" */
+681,	/* "onBasis" */
+491,	/* "organizationalStatus" */
+475,	/* "otherMailbox" */
+876,	/* "owner" */
+489,	/* "pagerTelephoneNumber" */
+374,	/* "path" */
+112,	/* "pbeWithMD5AndCast5CBC" */
+499,	/* "personalSignature" */
+487,	/* "personalTitle" */
+464,	/* "photo" */
+863,	/* "physicalDeliveryOfficeName" */
+437,	/* "pilot" */
+439,	/* "pilotAttributeSyntax" */
+438,	/* "pilotAttributeType" */
+479,	/* "pilotAttributeType27" */
+456,	/* "pilotDSA" */
+441,	/* "pilotGroups" */
+444,	/* "pilotObject" */
+440,	/* "pilotObjectClass" */
+455,	/* "pilotOrganization" */
+445,	/* "pilotPerson" */
+ 2,	/* "pkcs" */
+186,	/* "pkcs1" */
+27,	/* "pkcs3" */
+187,	/* "pkcs5" */
+20,	/* "pkcs7" */
+21,	/* "pkcs7-data" */
+25,	/* "pkcs7-digestData" */
+26,	/* "pkcs7-encryptedData" */
+23,	/* "pkcs7-envelopedData" */
+24,	/* "pkcs7-signedAndEnvelopedData" */
+22,	/* "pkcs7-signedData" */
+151,	/* "pkcs8ShroudedKeyBag" */
+47,	/* "pkcs9" */
+401,	/* "policyConstraints" */
+747,	/* "policyMappings" */
+862,	/* "postOfficeBox" */
+861,	/* "postalAddress" */
+661,	/* "postalCode" */
+683,	/* "ppBasis" */
+872,	/* "preferredDeliveryMethod" */
+873,	/* "presentationAddress" */
+816,	/* "prf-gostr3411-94" */
+406,	/* "prime-field" */
+409,	/* "prime192v1" */
+410,	/* "prime192v2" */
+411,	/* "prime192v3" */
+412,	/* "prime239v1" */
+413,	/* "prime239v2" */
+414,	/* "prime239v3" */
+415,	/* "prime256v1" */
+385,	/* "private" */
+84,	/* "privateKeyUsagePeriod" */
+886,	/* "protocolInformation" */
+663,	/* "proxyCertInfo" */
+510,	/* "pseudonym" */
+435,	/* "pss" */
+286,	/* "qcStatements" */
+457,	/* "qualityLabelledData" */
+450,	/* "rFC822localPart" */
+870,	/* "registeredAddress" */
+400,	/* "role" */
+877,	/* "roleOccupant" */
+448,	/* "room" */
+463,	/* "roomNumber" */
+ 6,	/* "rsaEncryption" */
+644,	/* "rsaOAEPEncryptionSET" */
+377,	/* "rsaSignature" */
+ 1,	/* "rsadsi" */
+482,	/* "sOARecord" */
+155,	/* "safeContentsBag" */
+291,	/* "sbgp-autonomousSysNum" */
+290,	/* "sbgp-ipAddrBlock" */
+292,	/* "sbgp-routerIdentifier" */
+159,	/* "sdsiCertificate" */
+859,	/* "searchGuide" */
+704,	/* "secp112r1" */
+705,	/* "secp112r2" */
+706,	/* "secp128r1" */
+707,	/* "secp128r2" */
+708,	/* "secp160k1" */
+709,	/* "secp160r1" */
+710,	/* "secp160r2" */
+711,	/* "secp192k1" */
+712,	/* "secp224k1" */
+713,	/* "secp224r1" */
+714,	/* "secp256k1" */
+715,	/* "secp384r1" */
+716,	/* "secp521r1" */
+154,	/* "secretBag" */
+474,	/* "secretary" */
+717,	/* "sect113r1" */
+718,	/* "sect113r2" */
+719,	/* "sect131r1" */
+720,	/* "sect131r2" */
+721,	/* "sect163k1" */
+722,	/* "sect163r1" */
+723,	/* "sect163r2" */
+724,	/* "sect193r1" */
+725,	/* "sect193r2" */
+726,	/* "sect233k1" */
+727,	/* "sect233r1" */
+728,	/* "sect239k1" */
+729,	/* "sect283k1" */
+730,	/* "sect283r1" */
+731,	/* "sect409k1" */
+732,	/* "sect409r1" */
+733,	/* "sect571k1" */
+734,	/* "sect571r1" */
+386,	/* "security" */
+878,	/* "seeAlso" */
+394,	/* "selected-attribute-types" */
+105,	/* "serialNumber" */
+129,	/* "serverAuth" */
+371,	/* "serviceLocator" */
+625,	/* "set-addPolicy" */
+515,	/* "set-attr" */
+518,	/* "set-brand" */
+638,	/* "set-brand-AmericanExpress" */
+637,	/* "set-brand-Diners" */
+636,	/* "set-brand-IATA-ATA" */
+639,	/* "set-brand-JCB" */
+641,	/* "set-brand-MasterCard" */
+642,	/* "set-brand-Novus" */
+640,	/* "set-brand-Visa" */
+517,	/* "set-certExt" */
+513,	/* "set-ctype" */
+514,	/* "set-msgExt" */
+516,	/* "set-policy" */
+607,	/* "set-policy-root" */
+624,	/* "set-rootKeyThumb" */
+620,	/* "setAttr-Cert" */
+631,	/* "setAttr-GenCryptgrm" */
+623,	/* "setAttr-IssCap" */
+628,	/* "setAttr-IssCap-CVM" */
+630,	/* "setAttr-IssCap-Sig" */
+629,	/* "setAttr-IssCap-T2" */
+621,	/* "setAttr-PGWYcap" */
+635,	/* "setAttr-SecDevSig" */
+632,	/* "setAttr-T2Enc" */
+633,	/* "setAttr-T2cleartxt" */
+634,	/* "setAttr-TokICCsig" */
+627,	/* "setAttr-Token-B0Prime" */
+626,	/* "setAttr-Token-EMV" */
+622,	/* "setAttr-TokenType" */
+619,	/* "setCext-IssuerCapabilities" */
+615,	/* "setCext-PGWYcapabilities" */
+616,	/* "setCext-TokenIdentifier" */
+618,	/* "setCext-TokenType" */
+617,	/* "setCext-Track2Data" */
+611,	/* "setCext-cCertRequired" */
+609,	/* "setCext-certType" */
+608,	/* "setCext-hashedRoot" */
+610,	/* "setCext-merchData" */
+613,	/* "setCext-setExt" */
+614,	/* "setCext-setQualf" */
+612,	/* "setCext-tunneling" */
+540,	/* "setct-AcqCardCodeMsg" */
+576,	/* "setct-AcqCardCodeMsgTBE" */
+570,	/* "setct-AuthReqTBE" */
+534,	/* "setct-AuthReqTBS" */
+527,	/* "setct-AuthResBaggage" */
+571,	/* "setct-AuthResTBE" */
+572,	/* "setct-AuthResTBEX" */
+535,	/* "setct-AuthResTBS" */
+536,	/* "setct-AuthResTBSX" */
+528,	/* "setct-AuthRevReqBaggage" */
+577,	/* "setct-AuthRevReqTBE" */
+541,	/* "setct-AuthRevReqTBS" */
+529,	/* "setct-AuthRevResBaggage" */
+542,	/* "setct-AuthRevResData" */
+578,	/* "setct-AuthRevResTBE" */
+579,	/* "setct-AuthRevResTBEB" */
+543,	/* "setct-AuthRevResTBS" */
+573,	/* "setct-AuthTokenTBE" */
+537,	/* "setct-AuthTokenTBS" */
+600,	/* "setct-BCIDistributionTBS" */
+558,	/* "setct-BatchAdminReqData" */
+592,	/* "setct-BatchAdminReqTBE" */
+559,	/* "setct-BatchAdminResData" */
+593,	/* "setct-BatchAdminResTBE" */
+599,	/* "setct-CRLNotificationResTBS" */
+598,	/* "setct-CRLNotificationTBS" */
+580,	/* "setct-CapReqTBE" */
+581,	/* "setct-CapReqTBEX" */
+544,	/* "setct-CapReqTBS" */
+545,	/* "setct-CapReqTBSX" */
+546,	/* "setct-CapResData" */
+582,	/* "setct-CapResTBE" */
+583,	/* "setct-CapRevReqTBE" */
+584,	/* "setct-CapRevReqTBEX" */
+547,	/* "setct-CapRevReqTBS" */
+548,	/* "setct-CapRevReqTBSX" */
+549,	/* "setct-CapRevResData" */
+585,	/* "setct-CapRevResTBE" */
+538,	/* "setct-CapTokenData" */
+530,	/* "setct-CapTokenSeq" */
+574,	/* "setct-CapTokenTBE" */
+575,	/* "setct-CapTokenTBEX" */
+539,	/* "setct-CapTokenTBS" */
+560,	/* "setct-CardCInitResTBS" */
+566,	/* "setct-CertInqReqTBS" */
+563,	/* "setct-CertReqData" */
+595,	/* "setct-CertReqTBE" */
+596,	/* "setct-CertReqTBEX" */
+564,	/* "setct-CertReqTBS" */
+565,	/* "setct-CertResData" */
+597,	/* "setct-CertResTBE" */
+586,	/* "setct-CredReqTBE" */
+587,	/* "setct-CredReqTBEX" */
+550,	/* "setct-CredReqTBS" */
+551,	/* "setct-CredReqTBSX" */
+552,	/* "setct-CredResData" */
+588,	/* "setct-CredResTBE" */
+589,	/* "setct-CredRevReqTBE" */
+590,	/* "setct-CredRevReqTBEX" */
+553,	/* "setct-CredRevReqTBS" */
+554,	/* "setct-CredRevReqTBSX" */
+555,	/* "setct-CredRevResData" */
+591,	/* "setct-CredRevResTBE" */
+567,	/* "setct-ErrorTBS" */
+526,	/* "setct-HODInput" */
+561,	/* "setct-MeAqCInitResTBS" */
+522,	/* "setct-OIData" */
+519,	/* "setct-PANData" */
+521,	/* "setct-PANOnly" */
+520,	/* "setct-PANToken" */
+556,	/* "setct-PCertReqData" */
+557,	/* "setct-PCertResTBS" */
+523,	/* "setct-PI" */
+532,	/* "setct-PI-TBS" */
+524,	/* "setct-PIData" */
+525,	/* "setct-PIDataUnsigned" */
+568,	/* "setct-PIDualSignedTBE" */
+569,	/* "setct-PIUnsignedTBE" */
+531,	/* "setct-PInitResData" */
+533,	/* "setct-PResData" */
+594,	/* "setct-RegFormReqTBE" */
+562,	/* "setct-RegFormResTBS" */
+606,	/* "setext-cv" */
+601,	/* "setext-genCrypt" */
+602,	/* "setext-miAuth" */
+604,	/* "setext-pinAny" */
+603,	/* "setext-pinSecure" */
+605,	/* "setext-track2" */
+52,	/* "signingTime" */
+454,	/* "simpleSecurityObject" */
+496,	/* "singleLevelQuality" */
+387,	/* "snmpv2" */
+660,	/* "street" */
+85,	/* "subjectAltName" */
+769,	/* "subjectDirectoryAttributes" */
+398,	/* "subjectInfoAccess" */
+82,	/* "subjectKeyIdentifier" */
+498,	/* "subtreeMaximumQuality" */
+497,	/* "subtreeMinimumQuality" */
+890,	/* "supportedAlgorithms" */
+874,	/* "supportedApplicationContext" */
+402,	/* "targetInformation" */
+864,	/* "telephoneNumber" */
+866,	/* "teletexTerminalIdentifier" */
+865,	/* "telexNumber" */
+459,	/* "textEncodedORAddress" */
+293,	/* "textNotice" */
+133,	/* "timeStamping" */
+106,	/* "title" */
+682,	/* "tpBasis" */
+375,	/* "trustRoot" */
+436,	/* "ucl" */
+888,	/* "uniqueMember" */
+55,	/* "unstructuredAddress" */
+49,	/* "unstructuredName" */
+880,	/* "userCertificate" */
+465,	/* "userClass" */
+879,	/* "userPassword" */
+373,	/* "valid" */
+678,	/* "wap" */
+679,	/* "wap-wsg" */
+735,	/* "wap-wsg-idm-ecid-wtls1" */
+743,	/* "wap-wsg-idm-ecid-wtls10" */
+744,	/* "wap-wsg-idm-ecid-wtls11" */
+745,	/* "wap-wsg-idm-ecid-wtls12" */
+736,	/* "wap-wsg-idm-ecid-wtls3" */
+737,	/* "wap-wsg-idm-ecid-wtls4" */
+738,	/* "wap-wsg-idm-ecid-wtls5" */
+739,	/* "wap-wsg-idm-ecid-wtls6" */
+740,	/* "wap-wsg-idm-ecid-wtls7" */
+741,	/* "wap-wsg-idm-ecid-wtls8" */
+742,	/* "wap-wsg-idm-ecid-wtls9" */
+804,	/* "whirlpool" */
+868,	/* "x121Address" */
+503,	/* "x500UniqueIdentifier" */
+158,	/* "x509Certificate" */
+160,	/* "x509Crl" */
 };
 
-static ASN1_OBJECT *ln_objs[NUM_LN]={
-&(nid_objs[363]),/* "AD Time Stamping" */
-&(nid_objs[405]),/* "ANSI X9.62" */
-&(nid_objs[368]),/* "Acceptable OCSP Responses" */
-&(nid_objs[664]),/* "Any language" */
-&(nid_objs[177]),/* "Authority Information Access" */
-&(nid_objs[365]),/* "Basic OCSP Response" */
-&(nid_objs[285]),/* "Biometric Info" */
-&(nid_objs[179]),/* "CA Issuers" */
-&(nid_objs[785]),/* "CA Repository" */
-&(nid_objs[131]),/* "Code Signing" */
-&(nid_objs[783]),/* "Diffie-Hellman based MAC" */
-&(nid_objs[382]),/* "Directory" */
-&(nid_objs[392]),/* "Domain" */
-&(nid_objs[132]),/* "E-mail Protection" */
-&(nid_objs[389]),/* "Enterprises" */
-&(nid_objs[384]),/* "Experimental" */
-&(nid_objs[372]),/* "Extended OCSP Status" */
-&(nid_objs[172]),/* "Extension Request" */
-&(nid_objs[813]),/* "GOST 28147-89" */
-&(nid_objs[849]),/* "GOST 28147-89 Cryptocom ParamSet" */
-&(nid_objs[815]),/* "GOST 28147-89 MAC" */
-&(nid_objs[851]),/* "GOST 34.10-2001 Cryptocom" */
-&(nid_objs[850]),/* "GOST 34.10-94 Cryptocom" */
-&(nid_objs[811]),/* "GOST R 34.10-2001" */
-&(nid_objs[817]),/* "GOST R 34.10-2001 DH" */
-&(nid_objs[812]),/* "GOST R 34.10-94" */
-&(nid_objs[818]),/* "GOST R 34.10-94 DH" */
-&(nid_objs[809]),/* "GOST R 34.11-94" */
-&(nid_objs[816]),/* "GOST R 34.11-94 PRF" */
-&(nid_objs[807]),/* "GOST R 34.11-94 with GOST R 34.10-2001" */
-&(nid_objs[853]),/* "GOST R 34.11-94 with GOST R 34.10-2001 Cryptocom" */
-&(nid_objs[808]),/* "GOST R 34.11-94 with GOST R 34.10-94" */
-&(nid_objs[852]),/* "GOST R 34.11-94 with GOST R 34.10-94 Cryptocom" */
-&(nid_objs[854]),/* "GOST R 3410-2001 Parameter Set Cryptocom" */
-&(nid_objs[810]),/* "HMAC GOST 34.11-94" */
-&(nid_objs[432]),/* "Hold Instruction Call Issuer" */
-&(nid_objs[430]),/* "Hold Instruction Code" */
-&(nid_objs[431]),/* "Hold Instruction None" */
-&(nid_objs[433]),/* "Hold Instruction Reject" */
-&(nid_objs[634]),/* "ICC or token signature" */
-&(nid_objs[294]),/* "IPSec End System" */
-&(nid_objs[295]),/* "IPSec Tunnel" */
-&(nid_objs[296]),/* "IPSec User" */
-&(nid_objs[182]),/* "ISO Member Body" */
-&(nid_objs[183]),/* "ISO US Member Body" */
-&(nid_objs[667]),/* "Independent" */
-&(nid_objs[665]),/* "Inherit all" */
-&(nid_objs[647]),/* "International Organizations" */
-&(nid_objs[142]),/* "Invalidity Date" */
-&(nid_objs[504]),/* "MIME MHS" */
-&(nid_objs[388]),/* "Mail" */
-&(nid_objs[383]),/* "Management" */
-&(nid_objs[417]),/* "Microsoft CSP Name" */
-&(nid_objs[135]),/* "Microsoft Commercial Code Signing" */
-&(nid_objs[138]),/* "Microsoft Encrypted File System" */
-&(nid_objs[171]),/* "Microsoft Extension Request" */
-&(nid_objs[134]),/* "Microsoft Individual Code Signing" */
-&(nid_objs[856]),/* "Microsoft Local Key set" */
-&(nid_objs[137]),/* "Microsoft Server Gated Crypto" */
-&(nid_objs[648]),/* "Microsoft Smartcardlogin" */
-&(nid_objs[136]),/* "Microsoft Trust List Signing" */
-&(nid_objs[649]),/* "Microsoft Universal Principal Name" */
-&(nid_objs[393]),/* "NULL" */
-&(nid_objs[404]),/* "NULL" */
-&(nid_objs[72]),/* "Netscape Base Url" */
-&(nid_objs[76]),/* "Netscape CA Policy Url" */
-&(nid_objs[74]),/* "Netscape CA Revocation Url" */
-&(nid_objs[71]),/* "Netscape Cert Type" */
-&(nid_objs[58]),/* "Netscape Certificate Extension" */
-&(nid_objs[79]),/* "Netscape Certificate Sequence" */
-&(nid_objs[78]),/* "Netscape Comment" */
-&(nid_objs[57]),/* "Netscape Communications Corp." */
-&(nid_objs[59]),/* "Netscape Data Type" */
-&(nid_objs[75]),/* "Netscape Renewal Url" */
-&(nid_objs[73]),/* "Netscape Revocation Url" */
-&(nid_objs[77]),/* "Netscape SSL Server Name" */
-&(nid_objs[139]),/* "Netscape Server Gated Crypto" */
-&(nid_objs[178]),/* "OCSP" */
-&(nid_objs[370]),/* "OCSP Archive Cutoff" */
-&(nid_objs[367]),/* "OCSP CRL ID" */
-&(nid_objs[369]),/* "OCSP No Check" */
-&(nid_objs[366]),/* "OCSP Nonce" */
-&(nid_objs[371]),/* "OCSP Service Locator" */
-&(nid_objs[180]),/* "OCSP Signing" */
-&(nid_objs[161]),/* "PBES2" */
-&(nid_objs[69]),/* "PBKDF2" */
-&(nid_objs[162]),/* "PBMAC1" */
-&(nid_objs[127]),/* "PKIX" */
-&(nid_objs[858]),/* "Permanent Identifier" */
-&(nid_objs[164]),/* "Policy Qualifier CPS" */
-&(nid_objs[165]),/* "Policy Qualifier User Notice" */
-&(nid_objs[385]),/* "Private" */
-&(nid_objs[663]),/* "Proxy Certificate Information" */
-&(nid_objs[ 1]),/* "RSA Data Security, Inc." */
-&(nid_objs[ 2]),/* "RSA Data Security, Inc. PKCS" */
-&(nid_objs[188]),/* "S/MIME" */
-&(nid_objs[167]),/* "S/MIME Capabilities" */
-&(nid_objs[387]),/* "SNMPv2" */
-&(nid_objs[512]),/* "Secure Electronic Transactions" */
-&(nid_objs[386]),/* "Security" */
-&(nid_objs[394]),/* "Selected Attribute Types" */
-&(nid_objs[143]),/* "Strong Extranet ID" */
-&(nid_objs[398]),/* "Subject Information Access" */
-&(nid_objs[130]),/* "TLS Web Client Authentication" */
-&(nid_objs[129]),/* "TLS Web Server Authentication" */
-&(nid_objs[133]),/* "Time Stamping" */
-&(nid_objs[375]),/* "Trust Root" */
-&(nid_objs[12]),/* "X509" */
-&(nid_objs[402]),/* "X509v3 AC Targeting" */
-&(nid_objs[746]),/* "X509v3 Any Policy" */
-&(nid_objs[90]),/* "X509v3 Authority Key Identifier" */
-&(nid_objs[87]),/* "X509v3 Basic Constraints" */
-&(nid_objs[103]),/* "X509v3 CRL Distribution Points" */
-&(nid_objs[88]),/* "X509v3 CRL Number" */
-&(nid_objs[141]),/* "X509v3 CRL Reason Code" */
-&(nid_objs[771]),/* "X509v3 Certificate Issuer" */
-&(nid_objs[89]),/* "X509v3 Certificate Policies" */
-&(nid_objs[140]),/* "X509v3 Delta CRL Indicator" */
-&(nid_objs[126]),/* "X509v3 Extended Key Usage" */
-&(nid_objs[857]),/* "X509v3 Freshest CRL" */
-&(nid_objs[748]),/* "X509v3 Inhibit Any Policy" */
-&(nid_objs[86]),/* "X509v3 Issuer Alternative Name" */
-&(nid_objs[770]),/* "X509v3 Issuing Distrubution Point" */
-&(nid_objs[83]),/* "X509v3 Key Usage" */
-&(nid_objs[666]),/* "X509v3 Name Constraints" */
-&(nid_objs[403]),/* "X509v3 No Revocation Available" */
-&(nid_objs[401]),/* "X509v3 Policy Constraints" */
-&(nid_objs[747]),/* "X509v3 Policy Mappings" */
-&(nid_objs[84]),/* "X509v3 Private Key Usage Period" */
-&(nid_objs[85]),/* "X509v3 Subject Alternative Name" */
-&(nid_objs[769]),/* "X509v3 Subject Directory Attributes" */
-&(nid_objs[82]),/* "X509v3 Subject Key Identifier" */
-&(nid_objs[184]),/* "X9.57" */
-&(nid_objs[185]),/* "X9.57 CM ?" */
-&(nid_objs[478]),/* "aRecord" */
-&(nid_objs[289]),/* "aaControls" */
-&(nid_objs[287]),/* "ac-auditEntity" */
-&(nid_objs[397]),/* "ac-proxying" */
-&(nid_objs[288]),/* "ac-targeting" */
-&(nid_objs[446]),/* "account" */
-&(nid_objs[364]),/* "ad dvcs" */
-&(nid_objs[606]),/* "additional verification" */
-&(nid_objs[419]),/* "aes-128-cbc" */
-&(nid_objs[421]),/* "aes-128-cfb" */
-&(nid_objs[650]),/* "aes-128-cfb1" */
-&(nid_objs[653]),/* "aes-128-cfb8" */
-&(nid_objs[418]),/* "aes-128-ecb" */
-&(nid_objs[420]),/* "aes-128-ofb" */
-&(nid_objs[423]),/* "aes-192-cbc" */
-&(nid_objs[425]),/* "aes-192-cfb" */
-&(nid_objs[651]),/* "aes-192-cfb1" */
-&(nid_objs[654]),/* "aes-192-cfb8" */
-&(nid_objs[422]),/* "aes-192-ecb" */
-&(nid_objs[424]),/* "aes-192-ofb" */
-&(nid_objs[427]),/* "aes-256-cbc" */
-&(nid_objs[429]),/* "aes-256-cfb" */
-&(nid_objs[652]),/* "aes-256-cfb1" */
-&(nid_objs[655]),/* "aes-256-cfb8" */
-&(nid_objs[426]),/* "aes-256-ecb" */
-&(nid_objs[428]),/* "aes-256-ofb" */
-&(nid_objs[376]),/* "algorithm" */
-&(nid_objs[484]),/* "associatedDomain" */
-&(nid_objs[485]),/* "associatedName" */
-&(nid_objs[501]),/* "audio" */
-&(nid_objs[882]),/* "authorityRevocationList" */
-&(nid_objs[91]),/* "bf-cbc" */
-&(nid_objs[93]),/* "bf-cfb" */
-&(nid_objs[92]),/* "bf-ecb" */
-&(nid_objs[94]),/* "bf-ofb" */
-&(nid_objs[494]),/* "buildingName" */
-&(nid_objs[860]),/* "businessCategory" */
-&(nid_objs[691]),/* "c2onb191v4" */
-&(nid_objs[692]),/* "c2onb191v5" */
-&(nid_objs[697]),/* "c2onb239v4" */
-&(nid_objs[698]),/* "c2onb239v5" */
-&(nid_objs[684]),/* "c2pnb163v1" */
-&(nid_objs[685]),/* "c2pnb163v2" */
-&(nid_objs[686]),/* "c2pnb163v3" */
-&(nid_objs[687]),/* "c2pnb176v1" */
-&(nid_objs[693]),/* "c2pnb208w1" */
-&(nid_objs[699]),/* "c2pnb272w1" */
-&(nid_objs[700]),/* "c2pnb304w1" */
-&(nid_objs[702]),/* "c2pnb368w1" */
-&(nid_objs[688]),/* "c2tnb191v1" */
-&(nid_objs[689]),/* "c2tnb191v2" */
-&(nid_objs[690]),/* "c2tnb191v3" */
-&(nid_objs[694]),/* "c2tnb239v1" */
-&(nid_objs[695]),/* "c2tnb239v2" */
-&(nid_objs[696]),/* "c2tnb239v3" */
-&(nid_objs[701]),/* "c2tnb359v1" */
-&(nid_objs[703]),/* "c2tnb431r1" */
-&(nid_objs[881]),/* "cACertificate" */
-&(nid_objs[483]),/* "cNAMERecord" */
-&(nid_objs[751]),/* "camellia-128-cbc" */
-&(nid_objs[757]),/* "camellia-128-cfb" */
-&(nid_objs[760]),/* "camellia-128-cfb1" */
-&(nid_objs[763]),/* "camellia-128-cfb8" */
-&(nid_objs[754]),/* "camellia-128-ecb" */
-&(nid_objs[766]),/* "camellia-128-ofb" */
-&(nid_objs[752]),/* "camellia-192-cbc" */
-&(nid_objs[758]),/* "camellia-192-cfb" */
-&(nid_objs[761]),/* "camellia-192-cfb1" */
-&(nid_objs[764]),/* "camellia-192-cfb8" */
-&(nid_objs[755]),/* "camellia-192-ecb" */
-&(nid_objs[767]),/* "camellia-192-ofb" */
-&(nid_objs[753]),/* "camellia-256-cbc" */
-&(nid_objs[759]),/* "camellia-256-cfb" */
-&(nid_objs[762]),/* "camellia-256-cfb1" */
-&(nid_objs[765]),/* "camellia-256-cfb8" */
-&(nid_objs[756]),/* "camellia-256-ecb" */
-&(nid_objs[768]),/* "camellia-256-ofb" */
-&(nid_objs[443]),/* "caseIgnoreIA5StringSyntax" */
-&(nid_objs[108]),/* "cast5-cbc" */
-&(nid_objs[110]),/* "cast5-cfb" */
-&(nid_objs[109]),/* "cast5-ecb" */
-&(nid_objs[111]),/* "cast5-ofb" */
-&(nid_objs[152]),/* "certBag" */
-&(nid_objs[677]),/* "certicom-arc" */
-&(nid_objs[517]),/* "certificate extensions" */
-&(nid_objs[883]),/* "certificateRevocationList" */
-&(nid_objs[54]),/* "challengePassword" */
-&(nid_objs[407]),/* "characteristic-two-field" */
-&(nid_objs[395]),/* "clearance" */
-&(nid_objs[633]),/* "cleartext track 2" */
-&(nid_objs[13]),/* "commonName" */
-&(nid_objs[513]),/* "content types" */
-&(nid_objs[50]),/* "contentType" */
-&(nid_objs[53]),/* "countersignature" */
-&(nid_objs[14]),/* "countryName" */
-&(nid_objs[153]),/* "crlBag" */
-&(nid_objs[884]),/* "crossCertificatePair" */
-&(nid_objs[806]),/* "cryptocom" */
-&(nid_objs[805]),/* "cryptopro" */
-&(nid_objs[500]),/* "dITRedirect" */
-&(nid_objs[451]),/* "dNSDomain" */
-&(nid_objs[495]),/* "dSAQuality" */
-&(nid_objs[434]),/* "data" */
-&(nid_objs[390]),/* "dcObject" */
-&(nid_objs[891]),/* "deltaRevocationList" */
-&(nid_objs[31]),/* "des-cbc" */
-&(nid_objs[643]),/* "des-cdmf" */
-&(nid_objs[30]),/* "des-cfb" */
-&(nid_objs[656]),/* "des-cfb1" */
-&(nid_objs[657]),/* "des-cfb8" */
-&(nid_objs[29]),/* "des-ecb" */
-&(nid_objs[32]),/* "des-ede" */
-&(nid_objs[43]),/* "des-ede-cbc" */
-&(nid_objs[60]),/* "des-ede-cfb" */
-&(nid_objs[62]),/* "des-ede-ofb" */
-&(nid_objs[33]),/* "des-ede3" */
-&(nid_objs[44]),/* "des-ede3-cbc" */
-&(nid_objs[61]),/* "des-ede3-cfb" */
-&(nid_objs[658]),/* "des-ede3-cfb1" */
-&(nid_objs[659]),/* "des-ede3-cfb8" */
-&(nid_objs[63]),/* "des-ede3-ofb" */
-&(nid_objs[45]),/* "des-ofb" */
-&(nid_objs[107]),/* "description" */
-&(nid_objs[871]),/* "destinationIndicator" */
-&(nid_objs[80]),/* "desx-cbc" */
-&(nid_objs[28]),/* "dhKeyAgreement" */
-&(nid_objs[11]),/* "directory services (X.500)" */
-&(nid_objs[378]),/* "directory services - algorithms" */
-&(nid_objs[887]),/* "distinguishedName" */
-&(nid_objs[892]),/* "dmdName" */
-&(nid_objs[174]),/* "dnQualifier" */
-&(nid_objs[447]),/* "document" */
-&(nid_objs[471]),/* "documentAuthor" */
-&(nid_objs[468]),/* "documentIdentifier" */
-&(nid_objs[472]),/* "documentLocation" */
-&(nid_objs[502]),/* "documentPublisher" */
-&(nid_objs[449]),/* "documentSeries" */
-&(nid_objs[469]),/* "documentTitle" */
-&(nid_objs[470]),/* "documentVersion" */
-&(nid_objs[380]),/* "dod" */
-&(nid_objs[391]),/* "domainComponent" */
-&(nid_objs[452]),/* "domainRelatedObject" */
-&(nid_objs[116]),/* "dsaEncryption" */
-&(nid_objs[67]),/* "dsaEncryption-old" */
-&(nid_objs[66]),/* "dsaWithSHA" */
-&(nid_objs[113]),/* "dsaWithSHA1" */
-&(nid_objs[70]),/* "dsaWithSHA1-old" */
-&(nid_objs[802]),/* "dsa_with_SHA224" */
-&(nid_objs[803]),/* "dsa_with_SHA256" */
-&(nid_objs[297]),/* "dvcs" */
-&(nid_objs[791]),/* "ecdsa-with-Recommended" */
-&(nid_objs[416]),/* "ecdsa-with-SHA1" */
-&(nid_objs[793]),/* "ecdsa-with-SHA224" */
-&(nid_objs[794]),/* "ecdsa-with-SHA256" */
-&(nid_objs[795]),/* "ecdsa-with-SHA384" */
-&(nid_objs[796]),/* "ecdsa-with-SHA512" */
-&(nid_objs[792]),/* "ecdsa-with-Specified" */
-&(nid_objs[48]),/* "emailAddress" */
-&(nid_objs[632]),/* "encrypted track 2" */
-&(nid_objs[885]),/* "enhancedSearchGuide" */
-&(nid_objs[56]),/* "extendedCertificateAttributes" */
-&(nid_objs[867]),/* "facsimileTelephoneNumber" */
-&(nid_objs[462]),/* "favouriteDrink" */
-&(nid_objs[453]),/* "friendlyCountry" */
-&(nid_objs[490]),/* "friendlyCountryName" */
-&(nid_objs[156]),/* "friendlyName" */
-&(nid_objs[631]),/* "generate cryptogram" */
-&(nid_objs[509]),/* "generationQualifier" */
-&(nid_objs[601]),/* "generic cryptogram" */
-&(nid_objs[99]),/* "givenName" */
-&(nid_objs[814]),/* "gost89-cnt" */
-&(nid_objs[855]),/* "hmac" */
-&(nid_objs[780]),/* "hmac-md5" */
-&(nid_objs[781]),/* "hmac-sha1" */
-&(nid_objs[797]),/* "hmacWithMD5" */
-&(nid_objs[163]),/* "hmacWithSHA1" */
-&(nid_objs[798]),/* "hmacWithSHA224" */
-&(nid_objs[799]),/* "hmacWithSHA256" */
-&(nid_objs[800]),/* "hmacWithSHA384" */
-&(nid_objs[801]),/* "hmacWithSHA512" */
-&(nid_objs[486]),/* "homePostalAddress" */
-&(nid_objs[473]),/* "homeTelephoneNumber" */
-&(nid_objs[466]),/* "host" */
-&(nid_objs[889]),/* "houseIdentifier" */
-&(nid_objs[442]),/* "iA5StringSyntax" */
-&(nid_objs[381]),/* "iana" */
-&(nid_objs[824]),/* "id-Gost28147-89-CryptoPro-A-ParamSet" */
-&(nid_objs[825]),/* "id-Gost28147-89-CryptoPro-B-ParamSet" */
-&(nid_objs[826]),/* "id-Gost28147-89-CryptoPro-C-ParamSet" */
-&(nid_objs[827]),/* "id-Gost28147-89-CryptoPro-D-ParamSet" */
-&(nid_objs[819]),/* "id-Gost28147-89-CryptoPro-KeyMeshing" */
-&(nid_objs[829]),/* "id-Gost28147-89-CryptoPro-Oscar-1-0-ParamSet" */
-&(nid_objs[828]),/* "id-Gost28147-89-CryptoPro-Oscar-1-1-ParamSet" */
-&(nid_objs[830]),/* "id-Gost28147-89-CryptoPro-RIC-1-ParamSet" */
-&(nid_objs[820]),/* "id-Gost28147-89-None-KeyMeshing" */
-&(nid_objs[823]),/* "id-Gost28147-89-TestParamSet" */
-&(nid_objs[840]),/* "id-GostR3410-2001-CryptoPro-A-ParamSet" */
-&(nid_objs[841]),/* "id-GostR3410-2001-CryptoPro-B-ParamSet" */
-&(nid_objs[842]),/* "id-GostR3410-2001-CryptoPro-C-ParamSet" */
-&(nid_objs[843]),/* "id-GostR3410-2001-CryptoPro-XchA-ParamSet" */
-&(nid_objs[844]),/* "id-GostR3410-2001-CryptoPro-XchB-ParamSet" */
-&(nid_objs[839]),/* "id-GostR3410-2001-TestParamSet" */
-&(nid_objs[832]),/* "id-GostR3410-94-CryptoPro-A-ParamSet" */
-&(nid_objs[833]),/* "id-GostR3410-94-CryptoPro-B-ParamSet" */
-&(nid_objs[834]),/* "id-GostR3410-94-CryptoPro-C-ParamSet" */
-&(nid_objs[835]),/* "id-GostR3410-94-CryptoPro-D-ParamSet" */
-&(nid_objs[836]),/* "id-GostR3410-94-CryptoPro-XchA-ParamSet" */
-&(nid_objs[837]),/* "id-GostR3410-94-CryptoPro-XchB-ParamSet" */
-&(nid_objs[838]),/* "id-GostR3410-94-CryptoPro-XchC-ParamSet" */
-&(nid_objs[831]),/* "id-GostR3410-94-TestParamSet" */
-&(nid_objs[845]),/* "id-GostR3410-94-a" */
-&(nid_objs[846]),/* "id-GostR3410-94-aBis" */
-&(nid_objs[847]),/* "id-GostR3410-94-b" */
-&(nid_objs[848]),/* "id-GostR3410-94-bBis" */
-&(nid_objs[822]),/* "id-GostR3411-94-CryptoProParamSet" */
-&(nid_objs[821]),/* "id-GostR3411-94-TestParamSet" */
-&(nid_objs[266]),/* "id-aca" */
-&(nid_objs[355]),/* "id-aca-accessIdentity" */
-&(nid_objs[354]),/* "id-aca-authenticationInfo" */
-&(nid_objs[356]),/* "id-aca-chargingIdentity" */
-&(nid_objs[399]),/* "id-aca-encAttrs" */
-&(nid_objs[357]),/* "id-aca-group" */
-&(nid_objs[358]),/* "id-aca-role" */
-&(nid_objs[176]),/* "id-ad" */
-&(nid_objs[788]),/* "id-aes128-wrap" */
-&(nid_objs[789]),/* "id-aes192-wrap" */
-&(nid_objs[790]),/* "id-aes256-wrap" */
-&(nid_objs[262]),/* "id-alg" */
-&(nid_objs[323]),/* "id-alg-des40" */
-&(nid_objs[326]),/* "id-alg-dh-pop" */
-&(nid_objs[325]),/* "id-alg-dh-sig-hmac-sha1" */
-&(nid_objs[324]),/* "id-alg-noSignature" */
-&(nid_objs[268]),/* "id-cct" */
-&(nid_objs[361]),/* "id-cct-PKIData" */
-&(nid_objs[362]),/* "id-cct-PKIResponse" */
-&(nid_objs[360]),/* "id-cct-crs" */
-&(nid_objs[81]),/* "id-ce" */
-&(nid_objs[680]),/* "id-characteristic-two-basis" */
-&(nid_objs[263]),/* "id-cmc" */
-&(nid_objs[334]),/* "id-cmc-addExtensions" */
-&(nid_objs[346]),/* "id-cmc-confirmCertAcceptance" */
-&(nid_objs[330]),/* "id-cmc-dataReturn" */
-&(nid_objs[336]),/* "id-cmc-decryptedPOP" */
-&(nid_objs[335]),/* "id-cmc-encryptedPOP" */
-&(nid_objs[339]),/* "id-cmc-getCRL" */
-&(nid_objs[338]),/* "id-cmc-getCert" */
-&(nid_objs[328]),/* "id-cmc-identification" */
-&(nid_objs[329]),/* "id-cmc-identityProof" */
-&(nid_objs[337]),/* "id-cmc-lraPOPWitness" */
-&(nid_objs[344]),/* "id-cmc-popLinkRandom" */
-&(nid_objs[345]),/* "id-cmc-popLinkWitness" */
-&(nid_objs[343]),/* "id-cmc-queryPending" */
-&(nid_objs[333]),/* "id-cmc-recipientNonce" */
-&(nid_objs[341]),/* "id-cmc-regInfo" */
-&(nid_objs[342]),/* "id-cmc-responseInfo" */
-&(nid_objs[340]),/* "id-cmc-revokeRequest" */
-&(nid_objs[332]),/* "id-cmc-senderNonce" */
-&(nid_objs[327]),/* "id-cmc-statusInfo" */
-&(nid_objs[331]),/* "id-cmc-transactionId" */
-&(nid_objs[787]),/* "id-ct-asciiTextWithCRLF" */
-&(nid_objs[408]),/* "id-ecPublicKey" */
-&(nid_objs[508]),/* "id-hex-multipart-message" */
-&(nid_objs[507]),/* "id-hex-partial-message" */
-&(nid_objs[260]),/* "id-it" */
-&(nid_objs[302]),/* "id-it-caKeyUpdateInfo" */
-&(nid_objs[298]),/* "id-it-caProtEncCert" */
-&(nid_objs[311]),/* "id-it-confirmWaitTime" */
-&(nid_objs[303]),/* "id-it-currentCRL" */
-&(nid_objs[300]),/* "id-it-encKeyPairTypes" */
-&(nid_objs[310]),/* "id-it-implicitConfirm" */
-&(nid_objs[308]),/* "id-it-keyPairParamRep" */
-&(nid_objs[307]),/* "id-it-keyPairParamReq" */
-&(nid_objs[312]),/* "id-it-origPKIMessage" */
-&(nid_objs[301]),/* "id-it-preferredSymmAlg" */
-&(nid_objs[309]),/* "id-it-revPassphrase" */
-&(nid_objs[299]),/* "id-it-signKeyPairTypes" */
-&(nid_objs[305]),/* "id-it-subscriptionRequest" */
-&(nid_objs[306]),/* "id-it-subscriptionResponse" */
-&(nid_objs[784]),/* "id-it-suppLangTags" */
-&(nid_objs[304]),/* "id-it-unsupportedOIDs" */
-&(nid_objs[128]),/* "id-kp" */
-&(nid_objs[280]),/* "id-mod-attribute-cert" */
-&(nid_objs[274]),/* "id-mod-cmc" */
-&(nid_objs[277]),/* "id-mod-cmp" */
-&(nid_objs[284]),/* "id-mod-cmp2000" */
-&(nid_objs[273]),/* "id-mod-crmf" */
-&(nid_objs[283]),/* "id-mod-dvcs" */
-&(nid_objs[275]),/* "id-mod-kea-profile-88" */
-&(nid_objs[276]),/* "id-mod-kea-profile-93" */
-&(nid_objs[282]),/* "id-mod-ocsp" */
-&(nid_objs[278]),/* "id-mod-qualified-cert-88" */
-&(nid_objs[279]),/* "id-mod-qualified-cert-93" */
-&(nid_objs[281]),/* "id-mod-timestamp-protocol" */
-&(nid_objs[264]),/* "id-on" */
-&(nid_objs[347]),/* "id-on-personalData" */
-&(nid_objs[265]),/* "id-pda" */
-&(nid_objs[352]),/* "id-pda-countryOfCitizenship" */
-&(nid_objs[353]),/* "id-pda-countryOfResidence" */
-&(nid_objs[348]),/* "id-pda-dateOfBirth" */
-&(nid_objs[351]),/* "id-pda-gender" */
-&(nid_objs[349]),/* "id-pda-placeOfBirth" */
-&(nid_objs[175]),/* "id-pe" */
-&(nid_objs[261]),/* "id-pkip" */
-&(nid_objs[258]),/* "id-pkix-mod" */
-&(nid_objs[269]),/* "id-pkix1-explicit-88" */
-&(nid_objs[271]),/* "id-pkix1-explicit-93" */
-&(nid_objs[270]),/* "id-pkix1-implicit-88" */
-&(nid_objs[272]),/* "id-pkix1-implicit-93" */
-&(nid_objs[662]),/* "id-ppl" */
-&(nid_objs[267]),/* "id-qcs" */
-&(nid_objs[359]),/* "id-qcs-pkixQCSyntax-v1" */
-&(nid_objs[259]),/* "id-qt" */
-&(nid_objs[313]),/* "id-regCtrl" */
-&(nid_objs[316]),/* "id-regCtrl-authenticator" */
-&(nid_objs[319]),/* "id-regCtrl-oldCertID" */
-&(nid_objs[318]),/* "id-regCtrl-pkiArchiveOptions" */
-&(nid_objs[317]),/* "id-regCtrl-pkiPublicationInfo" */
-&(nid_objs[320]),/* "id-regCtrl-protocolEncrKey" */
-&(nid_objs[315]),/* "id-regCtrl-regToken" */
-&(nid_objs[314]),/* "id-regInfo" */
-&(nid_objs[322]),/* "id-regInfo-certReq" */
-&(nid_objs[321]),/* "id-regInfo-utf8Pairs" */
-&(nid_objs[191]),/* "id-smime-aa" */
-&(nid_objs[215]),/* "id-smime-aa-contentHint" */
-&(nid_objs[218]),/* "id-smime-aa-contentIdentifier" */
-&(nid_objs[221]),/* "id-smime-aa-contentReference" */
-&(nid_objs[240]),/* "id-smime-aa-dvcs-dvc" */
-&(nid_objs[217]),/* "id-smime-aa-encapContentType" */
-&(nid_objs[222]),/* "id-smime-aa-encrypKeyPref" */
-&(nid_objs[220]),/* "id-smime-aa-equivalentLabels" */
-&(nid_objs[232]),/* "id-smime-aa-ets-CertificateRefs" */
-&(nid_objs[233]),/* "id-smime-aa-ets-RevocationRefs" */
-&(nid_objs[238]),/* "id-smime-aa-ets-archiveTimeStamp" */
-&(nid_objs[237]),/* "id-smime-aa-ets-certCRLTimestamp" */
-&(nid_objs[234]),/* "id-smime-aa-ets-certValues" */
-&(nid_objs[227]),/* "id-smime-aa-ets-commitmentType" */
-&(nid_objs[231]),/* "id-smime-aa-ets-contentTimestamp" */
-&(nid_objs[236]),/* "id-smime-aa-ets-escTimeStamp" */
-&(nid_objs[230]),/* "id-smime-aa-ets-otherSigCert" */
-&(nid_objs[235]),/* "id-smime-aa-ets-revocationValues" */
-&(nid_objs[226]),/* "id-smime-aa-ets-sigPolicyId" */
-&(nid_objs[229]),/* "id-smime-aa-ets-signerAttr" */
-&(nid_objs[228]),/* "id-smime-aa-ets-signerLocation" */
-&(nid_objs[219]),/* "id-smime-aa-macValue" */
-&(nid_objs[214]),/* "id-smime-aa-mlExpandHistory" */
-&(nid_objs[216]),/* "id-smime-aa-msgSigDigest" */
-&(nid_objs[212]),/* "id-smime-aa-receiptRequest" */
-&(nid_objs[213]),/* "id-smime-aa-securityLabel" */
-&(nid_objs[239]),/* "id-smime-aa-signatureType" */
-&(nid_objs[223]),/* "id-smime-aa-signingCertificate" */
-&(nid_objs[224]),/* "id-smime-aa-smimeEncryptCerts" */
-&(nid_objs[225]),/* "id-smime-aa-timeStampToken" */
-&(nid_objs[192]),/* "id-smime-alg" */
-&(nid_objs[243]),/* "id-smime-alg-3DESwrap" */
-&(nid_objs[246]),/* "id-smime-alg-CMS3DESwrap" */
-&(nid_objs[247]),/* "id-smime-alg-CMSRC2wrap" */
-&(nid_objs[245]),/* "id-smime-alg-ESDH" */
-&(nid_objs[241]),/* "id-smime-alg-ESDHwith3DES" */
-&(nid_objs[242]),/* "id-smime-alg-ESDHwithRC2" */
-&(nid_objs[244]),/* "id-smime-alg-RC2wrap" */
-&(nid_objs[193]),/* "id-smime-cd" */
-&(nid_objs[248]),/* "id-smime-cd-ldap" */
-&(nid_objs[190]),/* "id-smime-ct" */
-&(nid_objs[210]),/* "id-smime-ct-DVCSRequestData" */
-&(nid_objs[211]),/* "id-smime-ct-DVCSResponseData" */
-&(nid_objs[208]),/* "id-smime-ct-TDTInfo" */
-&(nid_objs[207]),/* "id-smime-ct-TSTInfo" */
-&(nid_objs[205]),/* "id-smime-ct-authData" */
-&(nid_objs[786]),/* "id-smime-ct-compressedData" */
-&(nid_objs[209]),/* "id-smime-ct-contentInfo" */
-&(nid_objs[206]),/* "id-smime-ct-publishCert" */
-&(nid_objs[204]),/* "id-smime-ct-receipt" */
-&(nid_objs[195]),/* "id-smime-cti" */
-&(nid_objs[255]),/* "id-smime-cti-ets-proofOfApproval" */
-&(nid_objs[256]),/* "id-smime-cti-ets-proofOfCreation" */
-&(nid_objs[253]),/* "id-smime-cti-ets-proofOfDelivery" */
-&(nid_objs[251]),/* "id-smime-cti-ets-proofOfOrigin" */
-&(nid_objs[252]),/* "id-smime-cti-ets-proofOfReceipt" */
-&(nid_objs[254]),/* "id-smime-cti-ets-proofOfSender" */
-&(nid_objs[189]),/* "id-smime-mod" */
-&(nid_objs[196]),/* "id-smime-mod-cms" */
-&(nid_objs[197]),/* "id-smime-mod-ess" */
-&(nid_objs[202]),/* "id-smime-mod-ets-eSigPolicy-88" */
-&(nid_objs[203]),/* "id-smime-mod-ets-eSigPolicy-97" */
-&(nid_objs[200]),/* "id-smime-mod-ets-eSignature-88" */
-&(nid_objs[201]),/* "id-smime-mod-ets-eSignature-97" */
-&(nid_objs[199]),/* "id-smime-mod-msg-v3" */
-&(nid_objs[198]),/* "id-smime-mod-oid" */
-&(nid_objs[194]),/* "id-smime-spq" */
-&(nid_objs[250]),/* "id-smime-spq-ets-sqt-unotice" */
-&(nid_objs[249]),/* "id-smime-spq-ets-sqt-uri" */
-&(nid_objs[34]),/* "idea-cbc" */
-&(nid_objs[35]),/* "idea-cfb" */
-&(nid_objs[36]),/* "idea-ecb" */
-&(nid_objs[46]),/* "idea-ofb" */
-&(nid_objs[676]),/* "identified-organization" */
-&(nid_objs[461]),/* "info" */
-&(nid_objs[101]),/* "initials" */
-&(nid_objs[869]),/* "internationaliSDNNumber" */
-&(nid_objs[749]),/* "ipsec3" */
-&(nid_objs[750]),/* "ipsec4" */
-&(nid_objs[181]),/* "iso" */
-&(nid_objs[623]),/* "issuer capabilities" */
-&(nid_objs[645]),/* "itu-t" */
-&(nid_objs[492]),/* "janetMailbox" */
-&(nid_objs[646]),/* "joint-iso-itu-t" */
-&(nid_objs[150]),/* "keyBag" */
-&(nid_objs[773]),/* "kisa" */
-&(nid_objs[477]),/* "lastModifiedBy" */
-&(nid_objs[476]),/* "lastModifiedTime" */
-&(nid_objs[157]),/* "localKeyID" */
-&(nid_objs[15]),/* "localityName" */
-&(nid_objs[480]),/* "mXRecord" */
-&(nid_objs[493]),/* "mailPreferenceOption" */
-&(nid_objs[467]),/* "manager" */
-&(nid_objs[ 3]),/* "md2" */
-&(nid_objs[ 7]),/* "md2WithRSAEncryption" */
-&(nid_objs[257]),/* "md4" */
-&(nid_objs[396]),/* "md4WithRSAEncryption" */
-&(nid_objs[ 4]),/* "md5" */
-&(nid_objs[114]),/* "md5-sha1" */
-&(nid_objs[104]),/* "md5WithRSA" */
-&(nid_objs[ 8]),/* "md5WithRSAEncryption" */
-&(nid_objs[95]),/* "mdc2" */
-&(nid_objs[96]),/* "mdc2WithRSA" */
-&(nid_objs[875]),/* "member" */
-&(nid_objs[602]),/* "merchant initiated auth" */
-&(nid_objs[514]),/* "message extensions" */
-&(nid_objs[51]),/* "messageDigest" */
-&(nid_objs[506]),/* "mime-mhs-bodies" */
-&(nid_objs[505]),/* "mime-mhs-headings" */
-&(nid_objs[488]),/* "mobileTelephoneNumber" */
-&(nid_objs[481]),/* "nSRecord" */
-&(nid_objs[173]),/* "name" */
-&(nid_objs[681]),/* "onBasis" */
-&(nid_objs[379]),/* "org" */
-&(nid_objs[17]),/* "organizationName" */
-&(nid_objs[491]),/* "organizationalStatus" */
-&(nid_objs[18]),/* "organizationalUnitName" */
-&(nid_objs[475]),/* "otherMailbox" */
-&(nid_objs[876]),/* "owner" */
-&(nid_objs[489]),/* "pagerTelephoneNumber" */
-&(nid_objs[782]),/* "password based MAC" */
-&(nid_objs[374]),/* "path" */
-&(nid_objs[621]),/* "payment gateway capabilities" */
-&(nid_objs[ 9]),/* "pbeWithMD2AndDES-CBC" */
-&(nid_objs[168]),/* "pbeWithMD2AndRC2-CBC" */
-&(nid_objs[112]),/* "pbeWithMD5AndCast5CBC" */
-&(nid_objs[10]),/* "pbeWithMD5AndDES-CBC" */
-&(nid_objs[169]),/* "pbeWithMD5AndRC2-CBC" */
-&(nid_objs[148]),/* "pbeWithSHA1And128BitRC2-CBC" */
-&(nid_objs[144]),/* "pbeWithSHA1And128BitRC4" */
-&(nid_objs[147]),/* "pbeWithSHA1And2-KeyTripleDES-CBC" */
-&(nid_objs[146]),/* "pbeWithSHA1And3-KeyTripleDES-CBC" */
-&(nid_objs[149]),/* "pbeWithSHA1And40BitRC2-CBC" */
-&(nid_objs[145]),/* "pbeWithSHA1And40BitRC4" */
-&(nid_objs[170]),/* "pbeWithSHA1AndDES-CBC" */
-&(nid_objs[68]),/* "pbeWithSHA1AndRC2-CBC" */
-&(nid_objs[499]),/* "personalSignature" */
-&(nid_objs[487]),/* "personalTitle" */
-&(nid_objs[464]),/* "photo" */
-&(nid_objs[863]),/* "physicalDeliveryOfficeName" */
-&(nid_objs[437]),/* "pilot" */
-&(nid_objs[439]),/* "pilotAttributeSyntax" */
-&(nid_objs[438]),/* "pilotAttributeType" */
-&(nid_objs[479]),/* "pilotAttributeType27" */
-&(nid_objs[456]),/* "pilotDSA" */
-&(nid_objs[441]),/* "pilotGroups" */
-&(nid_objs[444]),/* "pilotObject" */
-&(nid_objs[440]),/* "pilotObjectClass" */
-&(nid_objs[455]),/* "pilotOrganization" */
-&(nid_objs[445]),/* "pilotPerson" */
-&(nid_objs[186]),/* "pkcs1" */
-&(nid_objs[27]),/* "pkcs3" */
-&(nid_objs[187]),/* "pkcs5" */
-&(nid_objs[20]),/* "pkcs7" */
-&(nid_objs[21]),/* "pkcs7-data" */
-&(nid_objs[25]),/* "pkcs7-digestData" */
-&(nid_objs[26]),/* "pkcs7-encryptedData" */
-&(nid_objs[23]),/* "pkcs7-envelopedData" */
-&(nid_objs[24]),/* "pkcs7-signedAndEnvelopedData" */
-&(nid_objs[22]),/* "pkcs7-signedData" */
-&(nid_objs[151]),/* "pkcs8ShroudedKeyBag" */
-&(nid_objs[47]),/* "pkcs9" */
-&(nid_objs[862]),/* "postOfficeBox" */
-&(nid_objs[861]),/* "postalAddress" */
-&(nid_objs[661]),/* "postalCode" */
-&(nid_objs[683]),/* "ppBasis" */
-&(nid_objs[872]),/* "preferredDeliveryMethod" */
-&(nid_objs[873]),/* "presentationAddress" */
-&(nid_objs[406]),/* "prime-field" */
-&(nid_objs[409]),/* "prime192v1" */
-&(nid_objs[410]),/* "prime192v2" */
-&(nid_objs[411]),/* "prime192v3" */
-&(nid_objs[412]),/* "prime239v1" */
-&(nid_objs[413]),/* "prime239v2" */
-&(nid_objs[414]),/* "prime239v3" */
-&(nid_objs[415]),/* "prime256v1" */
-&(nid_objs[886]),/* "protocolInformation" */
-&(nid_objs[510]),/* "pseudonym" */
-&(nid_objs[435]),/* "pss" */
-&(nid_objs[286]),/* "qcStatements" */
-&(nid_objs[457]),/* "qualityLabelledData" */
-&(nid_objs[450]),/* "rFC822localPart" */
-&(nid_objs[98]),/* "rc2-40-cbc" */
-&(nid_objs[166]),/* "rc2-64-cbc" */
-&(nid_objs[37]),/* "rc2-cbc" */
-&(nid_objs[39]),/* "rc2-cfb" */
-&(nid_objs[38]),/* "rc2-ecb" */
-&(nid_objs[40]),/* "rc2-ofb" */
-&(nid_objs[ 5]),/* "rc4" */
-&(nid_objs[97]),/* "rc4-40" */
-&(nid_objs[120]),/* "rc5-cbc" */
-&(nid_objs[122]),/* "rc5-cfb" */
-&(nid_objs[121]),/* "rc5-ecb" */
-&(nid_objs[123]),/* "rc5-ofb" */
-&(nid_objs[870]),/* "registeredAddress" */
-&(nid_objs[460]),/* "rfc822Mailbox" */
-&(nid_objs[117]),/* "ripemd160" */
-&(nid_objs[119]),/* "ripemd160WithRSA" */
-&(nid_objs[400]),/* "role" */
-&(nid_objs[877]),/* "roleOccupant" */
-&(nid_objs[448]),/* "room" */
-&(nid_objs[463]),/* "roomNumber" */
-&(nid_objs[19]),/* "rsa" */
-&(nid_objs[ 6]),/* "rsaEncryption" */
-&(nid_objs[644]),/* "rsaOAEPEncryptionSET" */
-&(nid_objs[377]),/* "rsaSignature" */
-&(nid_objs[124]),/* "run length compression" */
-&(nid_objs[482]),/* "sOARecord" */
-&(nid_objs[155]),/* "safeContentsBag" */
-&(nid_objs[291]),/* "sbgp-autonomousSysNum" */
-&(nid_objs[290]),/* "sbgp-ipAddrBlock" */
-&(nid_objs[292]),/* "sbgp-routerIdentifier" */
-&(nid_objs[159]),/* "sdsiCertificate" */
-&(nid_objs[859]),/* "searchGuide" */
-&(nid_objs[704]),/* "secp112r1" */
-&(nid_objs[705]),/* "secp112r2" */
-&(nid_objs[706]),/* "secp128r1" */
-&(nid_objs[707]),/* "secp128r2" */
-&(nid_objs[708]),/* "secp160k1" */
-&(nid_objs[709]),/* "secp160r1" */
-&(nid_objs[710]),/* "secp160r2" */
-&(nid_objs[711]),/* "secp192k1" */
-&(nid_objs[712]),/* "secp224k1" */
-&(nid_objs[713]),/* "secp224r1" */
-&(nid_objs[714]),/* "secp256k1" */
-&(nid_objs[715]),/* "secp384r1" */
-&(nid_objs[716]),/* "secp521r1" */
-&(nid_objs[154]),/* "secretBag" */
-&(nid_objs[474]),/* "secretary" */
-&(nid_objs[717]),/* "sect113r1" */
-&(nid_objs[718]),/* "sect113r2" */
-&(nid_objs[719]),/* "sect131r1" */
-&(nid_objs[720]),/* "sect131r2" */
-&(nid_objs[721]),/* "sect163k1" */
-&(nid_objs[722]),/* "sect163r1" */
-&(nid_objs[723]),/* "sect163r2" */
-&(nid_objs[724]),/* "sect193r1" */
-&(nid_objs[725]),/* "sect193r2" */
-&(nid_objs[726]),/* "sect233k1" */
-&(nid_objs[727]),/* "sect233r1" */
-&(nid_objs[728]),/* "sect239k1" */
-&(nid_objs[729]),/* "sect283k1" */
-&(nid_objs[730]),/* "sect283r1" */
-&(nid_objs[731]),/* "sect409k1" */
-&(nid_objs[732]),/* "sect409r1" */
-&(nid_objs[733]),/* "sect571k1" */
-&(nid_objs[734]),/* "sect571r1" */
-&(nid_objs[635]),/* "secure device signature" */
-&(nid_objs[878]),/* "seeAlso" */
-&(nid_objs[777]),/* "seed-cbc" */
-&(nid_objs[779]),/* "seed-cfb" */
-&(nid_objs[776]),/* "seed-ecb" */
-&(nid_objs[778]),/* "seed-ofb" */
-&(nid_objs[105]),/* "serialNumber" */
-&(nid_objs[625]),/* "set-addPolicy" */
-&(nid_objs[515]),/* "set-attr" */
-&(nid_objs[518]),/* "set-brand" */
-&(nid_objs[638]),/* "set-brand-AmericanExpress" */
-&(nid_objs[637]),/* "set-brand-Diners" */
-&(nid_objs[636]),/* "set-brand-IATA-ATA" */
-&(nid_objs[639]),/* "set-brand-JCB" */
-&(nid_objs[641]),/* "set-brand-MasterCard" */
-&(nid_objs[642]),/* "set-brand-Novus" */
-&(nid_objs[640]),/* "set-brand-Visa" */
-&(nid_objs[516]),/* "set-policy" */
-&(nid_objs[607]),/* "set-policy-root" */
-&(nid_objs[624]),/* "set-rootKeyThumb" */
-&(nid_objs[620]),/* "setAttr-Cert" */
-&(nid_objs[628]),/* "setAttr-IssCap-CVM" */
-&(nid_objs[630]),/* "setAttr-IssCap-Sig" */
-&(nid_objs[629]),/* "setAttr-IssCap-T2" */
-&(nid_objs[627]),/* "setAttr-Token-B0Prime" */
-&(nid_objs[626]),/* "setAttr-Token-EMV" */
-&(nid_objs[622]),/* "setAttr-TokenType" */
-&(nid_objs[619]),/* "setCext-IssuerCapabilities" */
-&(nid_objs[615]),/* "setCext-PGWYcapabilities" */
-&(nid_objs[616]),/* "setCext-TokenIdentifier" */
-&(nid_objs[618]),/* "setCext-TokenType" */
-&(nid_objs[617]),/* "setCext-Track2Data" */
-&(nid_objs[611]),/* "setCext-cCertRequired" */
-&(nid_objs[609]),/* "setCext-certType" */
-&(nid_objs[608]),/* "setCext-hashedRoot" */
-&(nid_objs[610]),/* "setCext-merchData" */
-&(nid_objs[613]),/* "setCext-setExt" */
-&(nid_objs[614]),/* "setCext-setQualf" */
-&(nid_objs[612]),/* "setCext-tunneling" */
-&(nid_objs[540]),/* "setct-AcqCardCodeMsg" */
-&(nid_objs[576]),/* "setct-AcqCardCodeMsgTBE" */
-&(nid_objs[570]),/* "setct-AuthReqTBE" */
-&(nid_objs[534]),/* "setct-AuthReqTBS" */
-&(nid_objs[527]),/* "setct-AuthResBaggage" */
-&(nid_objs[571]),/* "setct-AuthResTBE" */
-&(nid_objs[572]),/* "setct-AuthResTBEX" */
-&(nid_objs[535]),/* "setct-AuthResTBS" */
-&(nid_objs[536]),/* "setct-AuthResTBSX" */
-&(nid_objs[528]),/* "setct-AuthRevReqBaggage" */
-&(nid_objs[577]),/* "setct-AuthRevReqTBE" */
-&(nid_objs[541]),/* "setct-AuthRevReqTBS" */
-&(nid_objs[529]),/* "setct-AuthRevResBaggage" */
-&(nid_objs[542]),/* "setct-AuthRevResData" */
-&(nid_objs[578]),/* "setct-AuthRevResTBE" */
-&(nid_objs[579]),/* "setct-AuthRevResTBEB" */
-&(nid_objs[543]),/* "setct-AuthRevResTBS" */
-&(nid_objs[573]),/* "setct-AuthTokenTBE" */
-&(nid_objs[537]),/* "setct-AuthTokenTBS" */
-&(nid_objs[600]),/* "setct-BCIDistributionTBS" */
-&(nid_objs[558]),/* "setct-BatchAdminReqData" */
-&(nid_objs[592]),/* "setct-BatchAdminReqTBE" */
-&(nid_objs[559]),/* "setct-BatchAdminResData" */
-&(nid_objs[593]),/* "setct-BatchAdminResTBE" */
-&(nid_objs[599]),/* "setct-CRLNotificationResTBS" */
-&(nid_objs[598]),/* "setct-CRLNotificationTBS" */
-&(nid_objs[580]),/* "setct-CapReqTBE" */
-&(nid_objs[581]),/* "setct-CapReqTBEX" */
-&(nid_objs[544]),/* "setct-CapReqTBS" */
-&(nid_objs[545]),/* "setct-CapReqTBSX" */
-&(nid_objs[546]),/* "setct-CapResData" */
-&(nid_objs[582]),/* "setct-CapResTBE" */
-&(nid_objs[583]),/* "setct-CapRevReqTBE" */
-&(nid_objs[584]),/* "setct-CapRevReqTBEX" */
-&(nid_objs[547]),/* "setct-CapRevReqTBS" */
-&(nid_objs[548]),/* "setct-CapRevReqTBSX" */
-&(nid_objs[549]),/* "setct-CapRevResData" */
-&(nid_objs[585]),/* "setct-CapRevResTBE" */
-&(nid_objs[538]),/* "setct-CapTokenData" */
-&(nid_objs[530]),/* "setct-CapTokenSeq" */
-&(nid_objs[574]),/* "setct-CapTokenTBE" */
-&(nid_objs[575]),/* "setct-CapTokenTBEX" */
-&(nid_objs[539]),/* "setct-CapTokenTBS" */
-&(nid_objs[560]),/* "setct-CardCInitResTBS" */
-&(nid_objs[566]),/* "setct-CertInqReqTBS" */
-&(nid_objs[563]),/* "setct-CertReqData" */
-&(nid_objs[595]),/* "setct-CertReqTBE" */
-&(nid_objs[596]),/* "setct-CertReqTBEX" */
-&(nid_objs[564]),/* "setct-CertReqTBS" */
-&(nid_objs[565]),/* "setct-CertResData" */
-&(nid_objs[597]),/* "setct-CertResTBE" */
-&(nid_objs[586]),/* "setct-CredReqTBE" */
-&(nid_objs[587]),/* "setct-CredReqTBEX" */
-&(nid_objs[550]),/* "setct-CredReqTBS" */
-&(nid_objs[551]),/* "setct-CredReqTBSX" */
-&(nid_objs[552]),/* "setct-CredResData" */
-&(nid_objs[588]),/* "setct-CredResTBE" */
-&(nid_objs[589]),/* "setct-CredRevReqTBE" */
-&(nid_objs[590]),/* "setct-CredRevReqTBEX" */
-&(nid_objs[553]),/* "setct-CredRevReqTBS" */
-&(nid_objs[554]),/* "setct-CredRevReqTBSX" */
-&(nid_objs[555]),/* "setct-CredRevResData" */
-&(nid_objs[591]),/* "setct-CredRevResTBE" */
-&(nid_objs[567]),/* "setct-ErrorTBS" */
-&(nid_objs[526]),/* "setct-HODInput" */
-&(nid_objs[561]),/* "setct-MeAqCInitResTBS" */
-&(nid_objs[522]),/* "setct-OIData" */
-&(nid_objs[519]),/* "setct-PANData" */
-&(nid_objs[521]),/* "setct-PANOnly" */
-&(nid_objs[520]),/* "setct-PANToken" */
-&(nid_objs[556]),/* "setct-PCertReqData" */
-&(nid_objs[557]),/* "setct-PCertResTBS" */
-&(nid_objs[523]),/* "setct-PI" */
-&(nid_objs[532]),/* "setct-PI-TBS" */
-&(nid_objs[524]),/* "setct-PIData" */
-&(nid_objs[525]),/* "setct-PIDataUnsigned" */
-&(nid_objs[568]),/* "setct-PIDualSignedTBE" */
-&(nid_objs[569]),/* "setct-PIUnsignedTBE" */
-&(nid_objs[531]),/* "setct-PInitResData" */
-&(nid_objs[533]),/* "setct-PResData" */
-&(nid_objs[594]),/* "setct-RegFormReqTBE" */
-&(nid_objs[562]),/* "setct-RegFormResTBS" */
-&(nid_objs[604]),/* "setext-pinAny" */
-&(nid_objs[603]),/* "setext-pinSecure" */
-&(nid_objs[605]),/* "setext-track2" */
-&(nid_objs[41]),/* "sha" */
-&(nid_objs[64]),/* "sha1" */
-&(nid_objs[115]),/* "sha1WithRSA" */
-&(nid_objs[65]),/* "sha1WithRSAEncryption" */
-&(nid_objs[675]),/* "sha224" */
-&(nid_objs[671]),/* "sha224WithRSAEncryption" */
-&(nid_objs[672]),/* "sha256" */
-&(nid_objs[668]),/* "sha256WithRSAEncryption" */
-&(nid_objs[673]),/* "sha384" */
-&(nid_objs[669]),/* "sha384WithRSAEncryption" */
-&(nid_objs[674]),/* "sha512" */
-&(nid_objs[670]),/* "sha512WithRSAEncryption" */
-&(nid_objs[42]),/* "shaWithRSAEncryption" */
-&(nid_objs[52]),/* "signingTime" */
-&(nid_objs[454]),/* "simpleSecurityObject" */
-&(nid_objs[496]),/* "singleLevelQuality" */
-&(nid_objs[16]),/* "stateOrProvinceName" */
-&(nid_objs[660]),/* "streetAddress" */
-&(nid_objs[498]),/* "subtreeMaximumQuality" */
-&(nid_objs[497]),/* "subtreeMinimumQuality" */
-&(nid_objs[890]),/* "supportedAlgorithms" */
-&(nid_objs[874]),/* "supportedApplicationContext" */
-&(nid_objs[100]),/* "surname" */
-&(nid_objs[864]),/* "telephoneNumber" */
-&(nid_objs[866]),/* "teletexTerminalIdentifier" */
-&(nid_objs[865]),/* "telexNumber" */
-&(nid_objs[459]),/* "textEncodedORAddress" */
-&(nid_objs[293]),/* "textNotice" */
-&(nid_objs[106]),/* "title" */
-&(nid_objs[682]),/* "tpBasis" */
-&(nid_objs[436]),/* "ucl" */
-&(nid_objs[ 0]),/* "undefined" */
-&(nid_objs[888]),/* "uniqueMember" */
-&(nid_objs[55]),/* "unstructuredAddress" */
-&(nid_objs[49]),/* "unstructuredName" */
-&(nid_objs[880]),/* "userCertificate" */
-&(nid_objs[465]),/* "userClass" */
-&(nid_objs[458]),/* "userId" */
-&(nid_objs[879]),/* "userPassword" */
-&(nid_objs[373]),/* "valid" */
-&(nid_objs[678]),/* "wap" */
-&(nid_objs[679]),/* "wap-wsg" */
-&(nid_objs[735]),/* "wap-wsg-idm-ecid-wtls1" */
-&(nid_objs[743]),/* "wap-wsg-idm-ecid-wtls10" */
-&(nid_objs[744]),/* "wap-wsg-idm-ecid-wtls11" */
-&(nid_objs[745]),/* "wap-wsg-idm-ecid-wtls12" */
-&(nid_objs[736]),/* "wap-wsg-idm-ecid-wtls3" */
-&(nid_objs[737]),/* "wap-wsg-idm-ecid-wtls4" */
-&(nid_objs[738]),/* "wap-wsg-idm-ecid-wtls5" */
-&(nid_objs[739]),/* "wap-wsg-idm-ecid-wtls6" */
-&(nid_objs[740]),/* "wap-wsg-idm-ecid-wtls7" */
-&(nid_objs[741]),/* "wap-wsg-idm-ecid-wtls8" */
-&(nid_objs[742]),/* "wap-wsg-idm-ecid-wtls9" */
-&(nid_objs[804]),/* "whirlpool" */
-&(nid_objs[868]),/* "x121Address" */
-&(nid_objs[503]),/* "x500UniqueIdentifier" */
-&(nid_objs[158]),/* "x509Certificate" */
-&(nid_objs[160]),/* "x509Crl" */
-&(nid_objs[125]),/* "zlib compression" */
+static const unsigned int ln_objs[NUM_LN]={
+363,	/* "AD Time Stamping" */
+405,	/* "ANSI X9.62" */
+368,	/* "Acceptable OCSP Responses" */
+910,	/* "Any Extended Key Usage" */
+664,	/* "Any language" */
+177,	/* "Authority Information Access" */
+365,	/* "Basic OCSP Response" */
+285,	/* "Biometric Info" */
+179,	/* "CA Issuers" */
+785,	/* "CA Repository" */
+131,	/* "Code Signing" */
+783,	/* "Diffie-Hellman based MAC" */
+382,	/* "Directory" */
+392,	/* "Domain" */
+132,	/* "E-mail Protection" */
+389,	/* "Enterprises" */
+384,	/* "Experimental" */
+372,	/* "Extended OCSP Status" */
+172,	/* "Extension Request" */
+813,	/* "GOST 28147-89" */
+849,	/* "GOST 28147-89 Cryptocom ParamSet" */
+815,	/* "GOST 28147-89 MAC" */
+851,	/* "GOST 34.10-2001 Cryptocom" */
+850,	/* "GOST 34.10-94 Cryptocom" */
+811,	/* "GOST R 34.10-2001" */
+817,	/* "GOST R 34.10-2001 DH" */
+812,	/* "GOST R 34.10-94" */
+818,	/* "GOST R 34.10-94 DH" */
+809,	/* "GOST R 34.11-94" */
+816,	/* "GOST R 34.11-94 PRF" */
+807,	/* "GOST R 34.11-94 with GOST R 34.10-2001" */
+853,	/* "GOST R 34.11-94 with GOST R 34.10-2001 Cryptocom" */
+808,	/* "GOST R 34.11-94 with GOST R 34.10-94" */
+852,	/* "GOST R 34.11-94 with GOST R 34.10-94 Cryptocom" */
+854,	/* "GOST R 3410-2001 Parameter Set Cryptocom" */
+810,	/* "HMAC GOST 34.11-94" */
+432,	/* "Hold Instruction Call Issuer" */
+430,	/* "Hold Instruction Code" */
+431,	/* "Hold Instruction None" */
+433,	/* "Hold Instruction Reject" */
+634,	/* "ICC or token signature" */
+294,	/* "IPSec End System" */
+295,	/* "IPSec Tunnel" */
+296,	/* "IPSec User" */
+182,	/* "ISO Member Body" */
+183,	/* "ISO US Member Body" */
+667,	/* "Independent" */
+665,	/* "Inherit all" */
+647,	/* "International Organizations" */
+142,	/* "Invalidity Date" */
+504,	/* "MIME MHS" */
+388,	/* "Mail" */
+383,	/* "Management" */
+417,	/* "Microsoft CSP Name" */
+135,	/* "Microsoft Commercial Code Signing" */
+138,	/* "Microsoft Encrypted File System" */
+171,	/* "Microsoft Extension Request" */
+134,	/* "Microsoft Individual Code Signing" */
+856,	/* "Microsoft Local Key set" */
+137,	/* "Microsoft Server Gated Crypto" */
+648,	/* "Microsoft Smartcardlogin" */
+136,	/* "Microsoft Trust List Signing" */
+649,	/* "Microsoft Universal Principal Name" */
+393,	/* "NULL" */
+404,	/* "NULL" */
+72,	/* "Netscape Base Url" */
+76,	/* "Netscape CA Policy Url" */
+74,	/* "Netscape CA Revocation Url" */
+71,	/* "Netscape Cert Type" */
+58,	/* "Netscape Certificate Extension" */
+79,	/* "Netscape Certificate Sequence" */
+78,	/* "Netscape Comment" */
+57,	/* "Netscape Communications Corp." */
+59,	/* "Netscape Data Type" */
+75,	/* "Netscape Renewal Url" */
+73,	/* "Netscape Revocation Url" */
+77,	/* "Netscape SSL Server Name" */
+139,	/* "Netscape Server Gated Crypto" */
+178,	/* "OCSP" */
+370,	/* "OCSP Archive Cutoff" */
+367,	/* "OCSP CRL ID" */
+369,	/* "OCSP No Check" */
+366,	/* "OCSP Nonce" */
+371,	/* "OCSP Service Locator" */
+180,	/* "OCSP Signing" */
+161,	/* "PBES2" */
+69,	/* "PBKDF2" */
+162,	/* "PBMAC1" */
+127,	/* "PKIX" */
+858,	/* "Permanent Identifier" */
+164,	/* "Policy Qualifier CPS" */
+165,	/* "Policy Qualifier User Notice" */
+385,	/* "Private" */
+663,	/* "Proxy Certificate Information" */
+ 1,	/* "RSA Data Security, Inc." */
+ 2,	/* "RSA Data Security, Inc. PKCS" */
+188,	/* "S/MIME" */
+167,	/* "S/MIME Capabilities" */
+387,	/* "SNMPv2" */
+512,	/* "Secure Electronic Transactions" */
+386,	/* "Security" */
+394,	/* "Selected Attribute Types" */
+143,	/* "Strong Extranet ID" */
+398,	/* "Subject Information Access" */
+130,	/* "TLS Web Client Authentication" */
+129,	/* "TLS Web Server Authentication" */
+133,	/* "Time Stamping" */
+375,	/* "Trust Root" */
+12,	/* "X509" */
+402,	/* "X509v3 AC Targeting" */
+746,	/* "X509v3 Any Policy" */
+90,	/* "X509v3 Authority Key Identifier" */
+87,	/* "X509v3 Basic Constraints" */
+103,	/* "X509v3 CRL Distribution Points" */
+88,	/* "X509v3 CRL Number" */
+141,	/* "X509v3 CRL Reason Code" */
+771,	/* "X509v3 Certificate Issuer" */
+89,	/* "X509v3 Certificate Policies" */
+140,	/* "X509v3 Delta CRL Indicator" */
+126,	/* "X509v3 Extended Key Usage" */
+857,	/* "X509v3 Freshest CRL" */
+748,	/* "X509v3 Inhibit Any Policy" */
+86,	/* "X509v3 Issuer Alternative Name" */
+770,	/* "X509v3 Issuing Distrubution Point" */
+83,	/* "X509v3 Key Usage" */
+666,	/* "X509v3 Name Constraints" */
+403,	/* "X509v3 No Revocation Available" */
+401,	/* "X509v3 Policy Constraints" */
+747,	/* "X509v3 Policy Mappings" */
+84,	/* "X509v3 Private Key Usage Period" */
+85,	/* "X509v3 Subject Alternative Name" */
+769,	/* "X509v3 Subject Directory Attributes" */
+82,	/* "X509v3 Subject Key Identifier" */
+184,	/* "X9.57" */
+185,	/* "X9.57 CM ?" */
+478,	/* "aRecord" */
+289,	/* "aaControls" */
+287,	/* "ac-auditEntity" */
+397,	/* "ac-proxying" */
+288,	/* "ac-targeting" */
+446,	/* "account" */
+364,	/* "ad dvcs" */
+606,	/* "additional verification" */
+419,	/* "aes-128-cbc" */
+916,	/* "aes-128-cbc-hmac-sha1" */
+896,	/* "aes-128-ccm" */
+421,	/* "aes-128-cfb" */
+650,	/* "aes-128-cfb1" */
+653,	/* "aes-128-cfb8" */
+904,	/* "aes-128-ctr" */
+418,	/* "aes-128-ecb" */
+895,	/* "aes-128-gcm" */
+420,	/* "aes-128-ofb" */
+913,	/* "aes-128-xts" */
+423,	/* "aes-192-cbc" */
+917,	/* "aes-192-cbc-hmac-sha1" */
+899,	/* "aes-192-ccm" */
+425,	/* "aes-192-cfb" */
+651,	/* "aes-192-cfb1" */
+654,	/* "aes-192-cfb8" */
+905,	/* "aes-192-ctr" */
+422,	/* "aes-192-ecb" */
+898,	/* "aes-192-gcm" */
+424,	/* "aes-192-ofb" */
+427,	/* "aes-256-cbc" */
+918,	/* "aes-256-cbc-hmac-sha1" */
+902,	/* "aes-256-ccm" */
+429,	/* "aes-256-cfb" */
+652,	/* "aes-256-cfb1" */
+655,	/* "aes-256-cfb8" */
+906,	/* "aes-256-ctr" */
+426,	/* "aes-256-ecb" */
+901,	/* "aes-256-gcm" */
+428,	/* "aes-256-ofb" */
+914,	/* "aes-256-xts" */
+376,	/* "algorithm" */
+484,	/* "associatedDomain" */
+485,	/* "associatedName" */
+501,	/* "audio" */
+882,	/* "authorityRevocationList" */
+91,	/* "bf-cbc" */
+93,	/* "bf-cfb" */
+92,	/* "bf-ecb" */
+94,	/* "bf-ofb" */
+494,	/* "buildingName" */
+860,	/* "businessCategory" */
+691,	/* "c2onb191v4" */
+692,	/* "c2onb191v5" */
+697,	/* "c2onb239v4" */
+698,	/* "c2onb239v5" */
+684,	/* "c2pnb163v1" */
+685,	/* "c2pnb163v2" */
+686,	/* "c2pnb163v3" */
+687,	/* "c2pnb176v1" */
+693,	/* "c2pnb208w1" */
+699,	/* "c2pnb272w1" */
+700,	/* "c2pnb304w1" */
+702,	/* "c2pnb368w1" */
+688,	/* "c2tnb191v1" */
+689,	/* "c2tnb191v2" */
+690,	/* "c2tnb191v3" */
+694,	/* "c2tnb239v1" */
+695,	/* "c2tnb239v2" */
+696,	/* "c2tnb239v3" */
+701,	/* "c2tnb359v1" */
+703,	/* "c2tnb431r1" */
+881,	/* "cACertificate" */
+483,	/* "cNAMERecord" */
+751,	/* "camellia-128-cbc" */
+757,	/* "camellia-128-cfb" */
+760,	/* "camellia-128-cfb1" */
+763,	/* "camellia-128-cfb8" */
+754,	/* "camellia-128-ecb" */
+766,	/* "camellia-128-ofb" */
+752,	/* "camellia-192-cbc" */
+758,	/* "camellia-192-cfb" */
+761,	/* "camellia-192-cfb1" */
+764,	/* "camellia-192-cfb8" */
+755,	/* "camellia-192-ecb" */
+767,	/* "camellia-192-ofb" */
+753,	/* "camellia-256-cbc" */
+759,	/* "camellia-256-cfb" */
+762,	/* "camellia-256-cfb1" */
+765,	/* "camellia-256-cfb8" */
+756,	/* "camellia-256-ecb" */
+768,	/* "camellia-256-ofb" */
+443,	/* "caseIgnoreIA5StringSyntax" */
+108,	/* "cast5-cbc" */
+110,	/* "cast5-cfb" */
+109,	/* "cast5-ecb" */
+111,	/* "cast5-ofb" */
+152,	/* "certBag" */
+677,	/* "certicom-arc" */
+517,	/* "certificate extensions" */
+883,	/* "certificateRevocationList" */
+54,	/* "challengePassword" */
+407,	/* "characteristic-two-field" */
+395,	/* "clearance" */
+633,	/* "cleartext track 2" */
+894,	/* "cmac" */
+13,	/* "commonName" */
+513,	/* "content types" */
+50,	/* "contentType" */
+53,	/* "countersignature" */
+14,	/* "countryName" */
+153,	/* "crlBag" */
+884,	/* "crossCertificatePair" */
+806,	/* "cryptocom" */
+805,	/* "cryptopro" */
+500,	/* "dITRedirect" */
+451,	/* "dNSDomain" */
+495,	/* "dSAQuality" */
+434,	/* "data" */
+390,	/* "dcObject" */
+891,	/* "deltaRevocationList" */
+31,	/* "des-cbc" */
+643,	/* "des-cdmf" */
+30,	/* "des-cfb" */
+656,	/* "des-cfb1" */
+657,	/* "des-cfb8" */
+29,	/* "des-ecb" */
+32,	/* "des-ede" */
+43,	/* "des-ede-cbc" */
+60,	/* "des-ede-cfb" */
+62,	/* "des-ede-ofb" */
+33,	/* "des-ede3" */
+44,	/* "des-ede3-cbc" */
+61,	/* "des-ede3-cfb" */
+658,	/* "des-ede3-cfb1" */
+659,	/* "des-ede3-cfb8" */
+63,	/* "des-ede3-ofb" */
+45,	/* "des-ofb" */
+107,	/* "description" */
+871,	/* "destinationIndicator" */
+80,	/* "desx-cbc" */
+28,	/* "dhKeyAgreement" */
+11,	/* "directory services (X.500)" */
+378,	/* "directory services - algorithms" */
+887,	/* "distinguishedName" */
+892,	/* "dmdName" */
+174,	/* "dnQualifier" */
+447,	/* "document" */
+471,	/* "documentAuthor" */
+468,	/* "documentIdentifier" */
+472,	/* "documentLocation" */
+502,	/* "documentPublisher" */
+449,	/* "documentSeries" */
+469,	/* "documentTitle" */
+470,	/* "documentVersion" */
+380,	/* "dod" */
+391,	/* "domainComponent" */
+452,	/* "domainRelatedObject" */
+116,	/* "dsaEncryption" */
+67,	/* "dsaEncryption-old" */
+66,	/* "dsaWithSHA" */
+113,	/* "dsaWithSHA1" */
+70,	/* "dsaWithSHA1-old" */
+802,	/* "dsa_with_SHA224" */
+803,	/* "dsa_with_SHA256" */
+297,	/* "dvcs" */
+791,	/* "ecdsa-with-Recommended" */
+416,	/* "ecdsa-with-SHA1" */
+793,	/* "ecdsa-with-SHA224" */
+794,	/* "ecdsa-with-SHA256" */
+795,	/* "ecdsa-with-SHA384" */
+796,	/* "ecdsa-with-SHA512" */
+792,	/* "ecdsa-with-Specified" */
+48,	/* "emailAddress" */
+632,	/* "encrypted track 2" */
+885,	/* "enhancedSearchGuide" */
+56,	/* "extendedCertificateAttributes" */
+867,	/* "facsimileTelephoneNumber" */
+462,	/* "favouriteDrink" */
+453,	/* "friendlyCountry" */
+490,	/* "friendlyCountryName" */
+156,	/* "friendlyName" */
+631,	/* "generate cryptogram" */
+509,	/* "generationQualifier" */
+601,	/* "generic cryptogram" */
+99,	/* "givenName" */
+814,	/* "gost89-cnt" */
+855,	/* "hmac" */
+780,	/* "hmac-md5" */
+781,	/* "hmac-sha1" */
+797,	/* "hmacWithMD5" */
+163,	/* "hmacWithSHA1" */
+798,	/* "hmacWithSHA224" */
+799,	/* "hmacWithSHA256" */
+800,	/* "hmacWithSHA384" */
+801,	/* "hmacWithSHA512" */
+486,	/* "homePostalAddress" */
+473,	/* "homeTelephoneNumber" */
+466,	/* "host" */
+889,	/* "houseIdentifier" */
+442,	/* "iA5StringSyntax" */
+381,	/* "iana" */
+824,	/* "id-Gost28147-89-CryptoPro-A-ParamSet" */
+825,	/* "id-Gost28147-89-CryptoPro-B-ParamSet" */
+826,	/* "id-Gost28147-89-CryptoPro-C-ParamSet" */
+827,	/* "id-Gost28147-89-CryptoPro-D-ParamSet" */
+819,	/* "id-Gost28147-89-CryptoPro-KeyMeshing" */
+829,	/* "id-Gost28147-89-CryptoPro-Oscar-1-0-ParamSet" */
+828,	/* "id-Gost28147-89-CryptoPro-Oscar-1-1-ParamSet" */
+830,	/* "id-Gost28147-89-CryptoPro-RIC-1-ParamSet" */
+820,	/* "id-Gost28147-89-None-KeyMeshing" */
+823,	/* "id-Gost28147-89-TestParamSet" */
+840,	/* "id-GostR3410-2001-CryptoPro-A-ParamSet" */
+841,	/* "id-GostR3410-2001-CryptoPro-B-ParamSet" */
+842,	/* "id-GostR3410-2001-CryptoPro-C-ParamSet" */
+843,	/* "id-GostR3410-2001-CryptoPro-XchA-ParamSet" */
+844,	/* "id-GostR3410-2001-CryptoPro-XchB-ParamSet" */
+839,	/* "id-GostR3410-2001-TestParamSet" */
+832,	/* "id-GostR3410-94-CryptoPro-A-ParamSet" */
+833,	/* "id-GostR3410-94-CryptoPro-B-ParamSet" */
+834,	/* "id-GostR3410-94-CryptoPro-C-ParamSet" */
+835,	/* "id-GostR3410-94-CryptoPro-D-ParamSet" */
+836,	/* "id-GostR3410-94-CryptoPro-XchA-ParamSet" */
+837,	/* "id-GostR3410-94-CryptoPro-XchB-ParamSet" */
+838,	/* "id-GostR3410-94-CryptoPro-XchC-ParamSet" */
+831,	/* "id-GostR3410-94-TestParamSet" */
+845,	/* "id-GostR3410-94-a" */
+846,	/* "id-GostR3410-94-aBis" */
+847,	/* "id-GostR3410-94-b" */
+848,	/* "id-GostR3410-94-bBis" */
+822,	/* "id-GostR3411-94-CryptoProParamSet" */
+821,	/* "id-GostR3411-94-TestParamSet" */
+266,	/* "id-aca" */
+355,	/* "id-aca-accessIdentity" */
+354,	/* "id-aca-authenticationInfo" */
+356,	/* "id-aca-chargingIdentity" */
+399,	/* "id-aca-encAttrs" */
+357,	/* "id-aca-group" */
+358,	/* "id-aca-role" */
+176,	/* "id-ad" */
+788,	/* "id-aes128-wrap" */
+897,	/* "id-aes128-wrap-pad" */
+789,	/* "id-aes192-wrap" */
+900,	/* "id-aes192-wrap-pad" */
+790,	/* "id-aes256-wrap" */
+903,	/* "id-aes256-wrap-pad" */
+262,	/* "id-alg" */
+893,	/* "id-alg-PWRI-KEK" */
+323,	/* "id-alg-des40" */
+326,	/* "id-alg-dh-pop" */
+325,	/* "id-alg-dh-sig-hmac-sha1" */
+324,	/* "id-alg-noSignature" */
+907,	/* "id-camellia128-wrap" */
+908,	/* "id-camellia192-wrap" */
+909,	/* "id-camellia256-wrap" */
+268,	/* "id-cct" */
+361,	/* "id-cct-PKIData" */
+362,	/* "id-cct-PKIResponse" */
+360,	/* "id-cct-crs" */
+81,	/* "id-ce" */
+680,	/* "id-characteristic-two-basis" */
+263,	/* "id-cmc" */
+334,	/* "id-cmc-addExtensions" */
+346,	/* "id-cmc-confirmCertAcceptance" */
+330,	/* "id-cmc-dataReturn" */
+336,	/* "id-cmc-decryptedPOP" */
+335,	/* "id-cmc-encryptedPOP" */
+339,	/* "id-cmc-getCRL" */
+338,	/* "id-cmc-getCert" */
+328,	/* "id-cmc-identification" */
+329,	/* "id-cmc-identityProof" */
+337,	/* "id-cmc-lraPOPWitness" */
+344,	/* "id-cmc-popLinkRandom" */
+345,	/* "id-cmc-popLinkWitness" */
+343,	/* "id-cmc-queryPending" */
+333,	/* "id-cmc-recipientNonce" */
+341,	/* "id-cmc-regInfo" */
+342,	/* "id-cmc-responseInfo" */
+340,	/* "id-cmc-revokeRequest" */
+332,	/* "id-cmc-senderNonce" */
+327,	/* "id-cmc-statusInfo" */
+331,	/* "id-cmc-transactionId" */
+787,	/* "id-ct-asciiTextWithCRLF" */
+408,	/* "id-ecPublicKey" */
+508,	/* "id-hex-multipart-message" */
+507,	/* "id-hex-partial-message" */
+260,	/* "id-it" */
+302,	/* "id-it-caKeyUpdateInfo" */
+298,	/* "id-it-caProtEncCert" */
+311,	/* "id-it-confirmWaitTime" */
+303,	/* "id-it-currentCRL" */
+300,	/* "id-it-encKeyPairTypes" */
+310,	/* "id-it-implicitConfirm" */
+308,	/* "id-it-keyPairParamRep" */
+307,	/* "id-it-keyPairParamReq" */
+312,	/* "id-it-origPKIMessage" */
+301,	/* "id-it-preferredSymmAlg" */
+309,	/* "id-it-revPassphrase" */
+299,	/* "id-it-signKeyPairTypes" */
+305,	/* "id-it-subscriptionRequest" */
+306,	/* "id-it-subscriptionResponse" */
+784,	/* "id-it-suppLangTags" */
+304,	/* "id-it-unsupportedOIDs" */
+128,	/* "id-kp" */
+280,	/* "id-mod-attribute-cert" */
+274,	/* "id-mod-cmc" */
+277,	/* "id-mod-cmp" */
+284,	/* "id-mod-cmp2000" */
+273,	/* "id-mod-crmf" */
+283,	/* "id-mod-dvcs" */
+275,	/* "id-mod-kea-profile-88" */
+276,	/* "id-mod-kea-profile-93" */
+282,	/* "id-mod-ocsp" */
+278,	/* "id-mod-qualified-cert-88" */
+279,	/* "id-mod-qualified-cert-93" */
+281,	/* "id-mod-timestamp-protocol" */
+264,	/* "id-on" */
+347,	/* "id-on-personalData" */
+265,	/* "id-pda" */
+352,	/* "id-pda-countryOfCitizenship" */
+353,	/* "id-pda-countryOfResidence" */
+348,	/* "id-pda-dateOfBirth" */
+351,	/* "id-pda-gender" */
+349,	/* "id-pda-placeOfBirth" */
+175,	/* "id-pe" */
+261,	/* "id-pkip" */
+258,	/* "id-pkix-mod" */
+269,	/* "id-pkix1-explicit-88" */
+271,	/* "id-pkix1-explicit-93" */
+270,	/* "id-pkix1-implicit-88" */
+272,	/* "id-pkix1-implicit-93" */
+662,	/* "id-ppl" */
+267,	/* "id-qcs" */
+359,	/* "id-qcs-pkixQCSyntax-v1" */
+259,	/* "id-qt" */
+313,	/* "id-regCtrl" */
+316,	/* "id-regCtrl-authenticator" */
+319,	/* "id-regCtrl-oldCertID" */
+318,	/* "id-regCtrl-pkiArchiveOptions" */
+317,	/* "id-regCtrl-pkiPublicationInfo" */
+320,	/* "id-regCtrl-protocolEncrKey" */
+315,	/* "id-regCtrl-regToken" */
+314,	/* "id-regInfo" */
+322,	/* "id-regInfo-certReq" */
+321,	/* "id-regInfo-utf8Pairs" */
+191,	/* "id-smime-aa" */
+215,	/* "id-smime-aa-contentHint" */
+218,	/* "id-smime-aa-contentIdentifier" */
+221,	/* "id-smime-aa-contentReference" */
+240,	/* "id-smime-aa-dvcs-dvc" */
+217,	/* "id-smime-aa-encapContentType" */
+222,	/* "id-smime-aa-encrypKeyPref" */
+220,	/* "id-smime-aa-equivalentLabels" */
+232,	/* "id-smime-aa-ets-CertificateRefs" */
+233,	/* "id-smime-aa-ets-RevocationRefs" */
+238,	/* "id-smime-aa-ets-archiveTimeStamp" */
+237,	/* "id-smime-aa-ets-certCRLTimestamp" */
+234,	/* "id-smime-aa-ets-certValues" */
+227,	/* "id-smime-aa-ets-commitmentType" */
+231,	/* "id-smime-aa-ets-contentTimestamp" */
+236,	/* "id-smime-aa-ets-escTimeStamp" */
+230,	/* "id-smime-aa-ets-otherSigCert" */
+235,	/* "id-smime-aa-ets-revocationValues" */
+226,	/* "id-smime-aa-ets-sigPolicyId" */
+229,	/* "id-smime-aa-ets-signerAttr" */
+228,	/* "id-smime-aa-ets-signerLocation" */
+219,	/* "id-smime-aa-macValue" */
+214,	/* "id-smime-aa-mlExpandHistory" */
+216,	/* "id-smime-aa-msgSigDigest" */
+212,	/* "id-smime-aa-receiptRequest" */
+213,	/* "id-smime-aa-securityLabel" */
+239,	/* "id-smime-aa-signatureType" */
+223,	/* "id-smime-aa-signingCertificate" */
+224,	/* "id-smime-aa-smimeEncryptCerts" */
+225,	/* "id-smime-aa-timeStampToken" */
+192,	/* "id-smime-alg" */
+243,	/* "id-smime-alg-3DESwrap" */
+246,	/* "id-smime-alg-CMS3DESwrap" */
+247,	/* "id-smime-alg-CMSRC2wrap" */
+245,	/* "id-smime-alg-ESDH" */
+241,	/* "id-smime-alg-ESDHwith3DES" */
+242,	/* "id-smime-alg-ESDHwithRC2" */
+244,	/* "id-smime-alg-RC2wrap" */
+193,	/* "id-smime-cd" */
+248,	/* "id-smime-cd-ldap" */
+190,	/* "id-smime-ct" */
+210,	/* "id-smime-ct-DVCSRequestData" */
+211,	/* "id-smime-ct-DVCSResponseData" */
+208,	/* "id-smime-ct-TDTInfo" */
+207,	/* "id-smime-ct-TSTInfo" */
+205,	/* "id-smime-ct-authData" */
+786,	/* "id-smime-ct-compressedData" */
+209,	/* "id-smime-ct-contentInfo" */
+206,	/* "id-smime-ct-publishCert" */
+204,	/* "id-smime-ct-receipt" */
+195,	/* "id-smime-cti" */
+255,	/* "id-smime-cti-ets-proofOfApproval" */
+256,	/* "id-smime-cti-ets-proofOfCreation" */
+253,	/* "id-smime-cti-ets-proofOfDelivery" */
+251,	/* "id-smime-cti-ets-proofOfOrigin" */
+252,	/* "id-smime-cti-ets-proofOfReceipt" */
+254,	/* "id-smime-cti-ets-proofOfSender" */
+189,	/* "id-smime-mod" */
+196,	/* "id-smime-mod-cms" */
+197,	/* "id-smime-mod-ess" */
+202,	/* "id-smime-mod-ets-eSigPolicy-88" */
+203,	/* "id-smime-mod-ets-eSigPolicy-97" */
+200,	/* "id-smime-mod-ets-eSignature-88" */
+201,	/* "id-smime-mod-ets-eSignature-97" */
+199,	/* "id-smime-mod-msg-v3" */
+198,	/* "id-smime-mod-oid" */
+194,	/* "id-smime-spq" */
+250,	/* "id-smime-spq-ets-sqt-unotice" */
+249,	/* "id-smime-spq-ets-sqt-uri" */
+34,	/* "idea-cbc" */
+35,	/* "idea-cfb" */
+36,	/* "idea-ecb" */
+46,	/* "idea-ofb" */
+676,	/* "identified-organization" */
+461,	/* "info" */
+101,	/* "initials" */
+869,	/* "internationaliSDNNumber" */
+749,	/* "ipsec3" */
+750,	/* "ipsec4" */
+181,	/* "iso" */
+623,	/* "issuer capabilities" */
+645,	/* "itu-t" */
+492,	/* "janetMailbox" */
+646,	/* "joint-iso-itu-t" */
+150,	/* "keyBag" */
+773,	/* "kisa" */
+477,	/* "lastModifiedBy" */
+476,	/* "lastModifiedTime" */
+157,	/* "localKeyID" */
+15,	/* "localityName" */
+480,	/* "mXRecord" */
+493,	/* "mailPreferenceOption" */
+467,	/* "manager" */
+ 3,	/* "md2" */
+ 7,	/* "md2WithRSAEncryption" */
+257,	/* "md4" */
+396,	/* "md4WithRSAEncryption" */
+ 4,	/* "md5" */
+114,	/* "md5-sha1" */
+104,	/* "md5WithRSA" */
+ 8,	/* "md5WithRSAEncryption" */
+95,	/* "mdc2" */
+96,	/* "mdc2WithRSA" */
+875,	/* "member" */
+602,	/* "merchant initiated auth" */
+514,	/* "message extensions" */
+51,	/* "messageDigest" */
+911,	/* "mgf1" */
+506,	/* "mime-mhs-bodies" */
+505,	/* "mime-mhs-headings" */
+488,	/* "mobileTelephoneNumber" */
+481,	/* "nSRecord" */
+173,	/* "name" */
+681,	/* "onBasis" */
+379,	/* "org" */
+17,	/* "organizationName" */
+491,	/* "organizationalStatus" */
+18,	/* "organizationalUnitName" */
+475,	/* "otherMailbox" */
+876,	/* "owner" */
+489,	/* "pagerTelephoneNumber" */
+782,	/* "password based MAC" */
+374,	/* "path" */
+621,	/* "payment gateway capabilities" */
+ 9,	/* "pbeWithMD2AndDES-CBC" */
+168,	/* "pbeWithMD2AndRC2-CBC" */
+112,	/* "pbeWithMD5AndCast5CBC" */
+10,	/* "pbeWithMD5AndDES-CBC" */
+169,	/* "pbeWithMD5AndRC2-CBC" */
+148,	/* "pbeWithSHA1And128BitRC2-CBC" */
+144,	/* "pbeWithSHA1And128BitRC4" */
+147,	/* "pbeWithSHA1And2-KeyTripleDES-CBC" */
+146,	/* "pbeWithSHA1And3-KeyTripleDES-CBC" */
+149,	/* "pbeWithSHA1And40BitRC2-CBC" */
+145,	/* "pbeWithSHA1And40BitRC4" */
+170,	/* "pbeWithSHA1AndDES-CBC" */
+68,	/* "pbeWithSHA1AndRC2-CBC" */
+499,	/* "personalSignature" */
+487,	/* "personalTitle" */
+464,	/* "photo" */
+863,	/* "physicalDeliveryOfficeName" */
+437,	/* "pilot" */
+439,	/* "pilotAttributeSyntax" */
+438,	/* "pilotAttributeType" */
+479,	/* "pilotAttributeType27" */
+456,	/* "pilotDSA" */
+441,	/* "pilotGroups" */
+444,	/* "pilotObject" */
+440,	/* "pilotObjectClass" */
+455,	/* "pilotOrganization" */
+445,	/* "pilotPerson" */
+186,	/* "pkcs1" */
+27,	/* "pkcs3" */
+187,	/* "pkcs5" */
+20,	/* "pkcs7" */
+21,	/* "pkcs7-data" */
+25,	/* "pkcs7-digestData" */
+26,	/* "pkcs7-encryptedData" */
+23,	/* "pkcs7-envelopedData" */
+24,	/* "pkcs7-signedAndEnvelopedData" */
+22,	/* "pkcs7-signedData" */
+151,	/* "pkcs8ShroudedKeyBag" */
+47,	/* "pkcs9" */
+862,	/* "postOfficeBox" */
+861,	/* "postalAddress" */
+661,	/* "postalCode" */
+683,	/* "ppBasis" */
+872,	/* "preferredDeliveryMethod" */
+873,	/* "presentationAddress" */
+406,	/* "prime-field" */
+409,	/* "prime192v1" */
+410,	/* "prime192v2" */
+411,	/* "prime192v3" */
+412,	/* "prime239v1" */
+413,	/* "prime239v2" */
+414,	/* "prime239v3" */
+415,	/* "prime256v1" */
+886,	/* "protocolInformation" */
+510,	/* "pseudonym" */
+435,	/* "pss" */
+286,	/* "qcStatements" */
+457,	/* "qualityLabelledData" */
+450,	/* "rFC822localPart" */
+98,	/* "rc2-40-cbc" */
+166,	/* "rc2-64-cbc" */
+37,	/* "rc2-cbc" */
+39,	/* "rc2-cfb" */
+38,	/* "rc2-ecb" */
+40,	/* "rc2-ofb" */
+ 5,	/* "rc4" */
+97,	/* "rc4-40" */
+915,	/* "rc4-hmac-md5" */
+120,	/* "rc5-cbc" */
+122,	/* "rc5-cfb" */
+121,	/* "rc5-ecb" */
+123,	/* "rc5-ofb" */
+870,	/* "registeredAddress" */
+460,	/* "rfc822Mailbox" */
+117,	/* "ripemd160" */
+119,	/* "ripemd160WithRSA" */
+400,	/* "role" */
+877,	/* "roleOccupant" */
+448,	/* "room" */
+463,	/* "roomNumber" */
+19,	/* "rsa" */
+ 6,	/* "rsaEncryption" */
+644,	/* "rsaOAEPEncryptionSET" */
+377,	/* "rsaSignature" */
+919,	/* "rsaesOaep" */
+912,	/* "rsassaPss" */
+124,	/* "run length compression" */
+482,	/* "sOARecord" */
+155,	/* "safeContentsBag" */
+291,	/* "sbgp-autonomousSysNum" */
+290,	/* "sbgp-ipAddrBlock" */
+292,	/* "sbgp-routerIdentifier" */
+159,	/* "sdsiCertificate" */
+859,	/* "searchGuide" */
+704,	/* "secp112r1" */
+705,	/* "secp112r2" */
+706,	/* "secp128r1" */
+707,	/* "secp128r2" */
+708,	/* "secp160k1" */
+709,	/* "secp160r1" */
+710,	/* "secp160r2" */
+711,	/* "secp192k1" */
+712,	/* "secp224k1" */
+713,	/* "secp224r1" */
+714,	/* "secp256k1" */
+715,	/* "secp384r1" */
+716,	/* "secp521r1" */
+154,	/* "secretBag" */
+474,	/* "secretary" */
+717,	/* "sect113r1" */
+718,	/* "sect113r2" */
+719,	/* "sect131r1" */
+720,	/* "sect131r2" */
+721,	/* "sect163k1" */
+722,	/* "sect163r1" */
+723,	/* "sect163r2" */
+724,	/* "sect193r1" */
+725,	/* "sect193r2" */
+726,	/* "sect233k1" */
+727,	/* "sect233r1" */
+728,	/* "sect239k1" */
+729,	/* "sect283k1" */
+730,	/* "sect283r1" */
+731,	/* "sect409k1" */
+732,	/* "sect409r1" */
+733,	/* "sect571k1" */
+734,	/* "sect571r1" */
+635,	/* "secure device signature" */
+878,	/* "seeAlso" */
+777,	/* "seed-cbc" */
+779,	/* "seed-cfb" */
+776,	/* "seed-ecb" */
+778,	/* "seed-ofb" */
+105,	/* "serialNumber" */
+625,	/* "set-addPolicy" */
+515,	/* "set-attr" */
+518,	/* "set-brand" */
+638,	/* "set-brand-AmericanExpress" */
+637,	/* "set-brand-Diners" */
+636,	/* "set-brand-IATA-ATA" */
+639,	/* "set-brand-JCB" */
+641,	/* "set-brand-MasterCard" */
+642,	/* "set-brand-Novus" */
+640,	/* "set-brand-Visa" */
+516,	/* "set-policy" */
+607,	/* "set-policy-root" */
+624,	/* "set-rootKeyThumb" */
+620,	/* "setAttr-Cert" */
+628,	/* "setAttr-IssCap-CVM" */
+630,	/* "setAttr-IssCap-Sig" */
+629,	/* "setAttr-IssCap-T2" */
+627,	/* "setAttr-Token-B0Prime" */
+626,	/* "setAttr-Token-EMV" */
+622,	/* "setAttr-TokenType" */
+619,	/* "setCext-IssuerCapabilities" */
+615,	/* "setCext-PGWYcapabilities" */
+616,	/* "setCext-TokenIdentifier" */
+618,	/* "setCext-TokenType" */
+617,	/* "setCext-Track2Data" */
+611,	/* "setCext-cCertRequired" */
+609,	/* "setCext-certType" */
+608,	/* "setCext-hashedRoot" */
+610,	/* "setCext-merchData" */
+613,	/* "setCext-setExt" */
+614,	/* "setCext-setQualf" */
+612,	/* "setCext-tunneling" */
+540,	/* "setct-AcqCardCodeMsg" */
+576,	/* "setct-AcqCardCodeMsgTBE" */
+570,	/* "setct-AuthReqTBE" */
+534,	/* "setct-AuthReqTBS" */
+527,	/* "setct-AuthResBaggage" */
+571,	/* "setct-AuthResTBE" */
+572,	/* "setct-AuthResTBEX" */
+535,	/* "setct-AuthResTBS" */
+536,	/* "setct-AuthResTBSX" */
+528,	/* "setct-AuthRevReqBaggage" */
+577,	/* "setct-AuthRevReqTBE" */
+541,	/* "setct-AuthRevReqTBS" */
+529,	/* "setct-AuthRevResBaggage" */
+542,	/* "setct-AuthRevResData" */
+578,	/* "setct-AuthRevResTBE" */
+579,	/* "setct-AuthRevResTBEB" */
+543,	/* "setct-AuthRevResTBS" */
+573,	/* "setct-AuthTokenTBE" */
+537,	/* "setct-AuthTokenTBS" */
+600,	/* "setct-BCIDistributionTBS" */
+558,	/* "setct-BatchAdminReqData" */
+592,	/* "setct-BatchAdminReqTBE" */
+559,	/* "setct-BatchAdminResData" */
+593,	/* "setct-BatchAdminResTBE" */
+599,	/* "setct-CRLNotificationResTBS" */
+598,	/* "setct-CRLNotificationTBS" */
+580,	/* "setct-CapReqTBE" */
+581,	/* "setct-CapReqTBEX" */
+544,	/* "setct-CapReqTBS" */
+545,	/* "setct-CapReqTBSX" */
+546,	/* "setct-CapResData" */
+582,	/* "setct-CapResTBE" */
+583,	/* "setct-CapRevReqTBE" */
+584,	/* "setct-CapRevReqTBEX" */
+547,	/* "setct-CapRevReqTBS" */
+548,	/* "setct-CapRevReqTBSX" */
+549,	/* "setct-CapRevResData" */
+585,	/* "setct-CapRevResTBE" */
+538,	/* "setct-CapTokenData" */
+530,	/* "setct-CapTokenSeq" */
+574,	/* "setct-CapTokenTBE" */
+575,	/* "setct-CapTokenTBEX" */
+539,	/* "setct-CapTokenTBS" */
+560,	/* "setct-CardCInitResTBS" */
+566,	/* "setct-CertInqReqTBS" */
+563,	/* "setct-CertReqData" */
+595,	/* "setct-CertReqTBE" */
+596,	/* "setct-CertReqTBEX" */
+564,	/* "setct-CertReqTBS" */
+565,	/* "setct-CertResData" */
+597,	/* "setct-CertResTBE" */
+586,	/* "setct-CredReqTBE" */
+587,	/* "setct-CredReqTBEX" */
+550,	/* "setct-CredReqTBS" */
+551,	/* "setct-CredReqTBSX" */
+552,	/* "setct-CredResData" */
+588,	/* "setct-CredResTBE" */
+589,	/* "setct-CredRevReqTBE" */
+590,	/* "setct-CredRevReqTBEX" */
+553,	/* "setct-CredRevReqTBS" */
+554,	/* "setct-CredRevReqTBSX" */
+555,	/* "setct-CredRevResData" */
+591,	/* "setct-CredRevResTBE" */
+567,	/* "setct-ErrorTBS" */
+526,	/* "setct-HODInput" */
+561,	/* "setct-MeAqCInitResTBS" */
+522,	/* "setct-OIData" */
+519,	/* "setct-PANData" */
+521,	/* "setct-PANOnly" */
+520,	/* "setct-PANToken" */
+556,	/* "setct-PCertReqData" */
+557,	/* "setct-PCertResTBS" */
+523,	/* "setct-PI" */
+532,	/* "setct-PI-TBS" */
+524,	/* "setct-PIData" */
+525,	/* "setct-PIDataUnsigned" */
+568,	/* "setct-PIDualSignedTBE" */
+569,	/* "setct-PIUnsignedTBE" */
+531,	/* "setct-PInitResData" */
+533,	/* "setct-PResData" */
+594,	/* "setct-RegFormReqTBE" */
+562,	/* "setct-RegFormResTBS" */
+604,	/* "setext-pinAny" */
+603,	/* "setext-pinSecure" */
+605,	/* "setext-track2" */
+41,	/* "sha" */
+64,	/* "sha1" */
+115,	/* "sha1WithRSA" */
+65,	/* "sha1WithRSAEncryption" */
+675,	/* "sha224" */
+671,	/* "sha224WithRSAEncryption" */
+672,	/* "sha256" */
+668,	/* "sha256WithRSAEncryption" */
+673,	/* "sha384" */
+669,	/* "sha384WithRSAEncryption" */
+674,	/* "sha512" */
+670,	/* "sha512WithRSAEncryption" */
+42,	/* "shaWithRSAEncryption" */
+52,	/* "signingTime" */
+454,	/* "simpleSecurityObject" */
+496,	/* "singleLevelQuality" */
+16,	/* "stateOrProvinceName" */
+660,	/* "streetAddress" */
+498,	/* "subtreeMaximumQuality" */
+497,	/* "subtreeMinimumQuality" */
+890,	/* "supportedAlgorithms" */
+874,	/* "supportedApplicationContext" */
+100,	/* "surname" */
+864,	/* "telephoneNumber" */
+866,	/* "teletexTerminalIdentifier" */
+865,	/* "telexNumber" */
+459,	/* "textEncodedORAddress" */
+293,	/* "textNotice" */
+106,	/* "title" */
+682,	/* "tpBasis" */
+436,	/* "ucl" */
+ 0,	/* "undefined" */
+888,	/* "uniqueMember" */
+55,	/* "unstructuredAddress" */
+49,	/* "unstructuredName" */
+880,	/* "userCertificate" */
+465,	/* "userClass" */
+458,	/* "userId" */
+879,	/* "userPassword" */
+373,	/* "valid" */
+678,	/* "wap" */
+679,	/* "wap-wsg" */
+735,	/* "wap-wsg-idm-ecid-wtls1" */
+743,	/* "wap-wsg-idm-ecid-wtls10" */
+744,	/* "wap-wsg-idm-ecid-wtls11" */
+745,	/* "wap-wsg-idm-ecid-wtls12" */
+736,	/* "wap-wsg-idm-ecid-wtls3" */
+737,	/* "wap-wsg-idm-ecid-wtls4" */
+738,	/* "wap-wsg-idm-ecid-wtls5" */
+739,	/* "wap-wsg-idm-ecid-wtls6" */
+740,	/* "wap-wsg-idm-ecid-wtls7" */
+741,	/* "wap-wsg-idm-ecid-wtls8" */
+742,	/* "wap-wsg-idm-ecid-wtls9" */
+804,	/* "whirlpool" */
+868,	/* "x121Address" */
+503,	/* "x500UniqueIdentifier" */
+158,	/* "x509Certificate" */
+160,	/* "x509Crl" */
+125,	/* "zlib compression" */
 };
 
-static ASN1_OBJECT *obj_objs[NUM_OBJ]={
-&(nid_objs[ 0]),/* OBJ_undef                        0 */
-&(nid_objs[393]),/* OBJ_joint_iso_ccitt              OBJ_joint_iso_itu_t */
-&(nid_objs[404]),/* OBJ_ccitt                        OBJ_itu_t */
-&(nid_objs[645]),/* OBJ_itu_t                        0 */
-&(nid_objs[434]),/* OBJ_data                         0 9 */
-&(nid_objs[181]),/* OBJ_iso                          1 */
-&(nid_objs[182]),/* OBJ_member_body                  1 2 */
-&(nid_objs[379]),/* OBJ_org                          1 3 */
-&(nid_objs[676]),/* OBJ_identified_organization      1 3 */
-&(nid_objs[646]),/* OBJ_joint_iso_itu_t              2 */
-&(nid_objs[11]),/* OBJ_X500                         2 5 */
-&(nid_objs[647]),/* OBJ_international_organizations  2 23 */
-&(nid_objs[380]),/* OBJ_dod                          1 3 6 */
-&(nid_objs[12]),/* OBJ_X509                         2 5 4 */
-&(nid_objs[378]),/* OBJ_X500algorithms               2 5 8 */
-&(nid_objs[81]),/* OBJ_id_ce                        2 5 29 */
-&(nid_objs[512]),/* OBJ_id_set                       2 23 42 */
-&(nid_objs[678]),/* OBJ_wap                          2 23 43 */
-&(nid_objs[435]),/* OBJ_pss                          0 9 2342 */
-&(nid_objs[183]),/* OBJ_ISO_US                       1 2 840 */
-&(nid_objs[381]),/* OBJ_iana                         1 3 6 1 */
-&(nid_objs[677]),/* OBJ_certicom_arc                 1 3 132 */
-&(nid_objs[394]),/* OBJ_selected_attribute_types     2 5 1 5 */
-&(nid_objs[13]),/* OBJ_commonName                   2 5 4 3 */
-&(nid_objs[100]),/* OBJ_surname                      2 5 4 4 */
-&(nid_objs[105]),/* OBJ_serialNumber                 2 5 4 5 */
-&(nid_objs[14]),/* OBJ_countryName                  2 5 4 6 */
-&(nid_objs[15]),/* OBJ_localityName                 2 5 4 7 */
-&(nid_objs[16]),/* OBJ_stateOrProvinceName          2 5 4 8 */
-&(nid_objs[660]),/* OBJ_streetAddress                2 5 4 9 */
-&(nid_objs[17]),/* OBJ_organizationName             2 5 4 10 */
-&(nid_objs[18]),/* OBJ_organizationalUnitName       2 5 4 11 */
-&(nid_objs[106]),/* OBJ_title                        2 5 4 12 */
-&(nid_objs[107]),/* OBJ_description                  2 5 4 13 */
-&(nid_objs[859]),/* OBJ_searchGuide                  2 5 4 14 */
-&(nid_objs[860]),/* OBJ_businessCategory             2 5 4 15 */
-&(nid_objs[861]),/* OBJ_postalAddress                2 5 4 16 */
-&(nid_objs[661]),/* OBJ_postalCode                   2 5 4 17 */
-&(nid_objs[862]),/* OBJ_postOfficeBox                2 5 4 18 */
-&(nid_objs[863]),/* OBJ_physicalDeliveryOfficeName   2 5 4 19 */
-&(nid_objs[864]),/* OBJ_telephoneNumber              2 5 4 20 */
-&(nid_objs[865]),/* OBJ_telexNumber                  2 5 4 21 */
-&(nid_objs[866]),/* OBJ_teletexTerminalIdentifier    2 5 4 22 */
-&(nid_objs[867]),/* OBJ_facsimileTelephoneNumber     2 5 4 23 */
-&(nid_objs[868]),/* OBJ_x121Address                  2 5 4 24 */
-&(nid_objs[869]),/* OBJ_internationaliSDNNumber      2 5 4 25 */
-&(nid_objs[870]),/* OBJ_registeredAddress            2 5 4 26 */
-&(nid_objs[871]),/* OBJ_destinationIndicator         2 5 4 27 */
-&(nid_objs[872]),/* OBJ_preferredDeliveryMethod      2 5 4 28 */
-&(nid_objs[873]),/* OBJ_presentationAddress          2 5 4 29 */
-&(nid_objs[874]),/* OBJ_supportedApplicationContext  2 5 4 30 */
-&(nid_objs[875]),/* OBJ_member                       2 5 4 31 */
-&(nid_objs[876]),/* OBJ_owner                        2 5 4 32 */
-&(nid_objs[877]),/* OBJ_roleOccupant                 2 5 4 33 */
-&(nid_objs[878]),/* OBJ_seeAlso                      2 5 4 34 */
-&(nid_objs[879]),/* OBJ_userPassword                 2 5 4 35 */
-&(nid_objs[880]),/* OBJ_userCertificate              2 5 4 36 */
-&(nid_objs[881]),/* OBJ_cACertificate                2 5 4 37 */
-&(nid_objs[882]),/* OBJ_authorityRevocationList      2 5 4 38 */
-&(nid_objs[883]),/* OBJ_certificateRevocationList    2 5 4 39 */
-&(nid_objs[884]),/* OBJ_crossCertificatePair         2 5 4 40 */
-&(nid_objs[173]),/* OBJ_name                         2 5 4 41 */
-&(nid_objs[99]),/* OBJ_givenName                    2 5 4 42 */
-&(nid_objs[101]),/* OBJ_initials                     2 5 4 43 */
-&(nid_objs[509]),/* OBJ_generationQualifier          2 5 4 44 */
-&(nid_objs[503]),/* OBJ_x500UniqueIdentifier         2 5 4 45 */
-&(nid_objs[174]),/* OBJ_dnQualifier                  2 5 4 46 */
-&(nid_objs[885]),/* OBJ_enhancedSearchGuide          2 5 4 47 */
-&(nid_objs[886]),/* OBJ_protocolInformation          2 5 4 48 */
-&(nid_objs[887]),/* OBJ_distinguishedName            2 5 4 49 */
-&(nid_objs[888]),/* OBJ_uniqueMember                 2 5 4 50 */
-&(nid_objs[889]),/* OBJ_houseIdentifier              2 5 4 51 */
-&(nid_objs[890]),/* OBJ_supportedAlgorithms          2 5 4 52 */
-&(nid_objs[891]),/* OBJ_deltaRevocationList          2 5 4 53 */
-&(nid_objs[892]),/* OBJ_dmdName                      2 5 4 54 */
-&(nid_objs[510]),/* OBJ_pseudonym                    2 5 4 65 */
-&(nid_objs[400]),/* OBJ_role                         2 5 4 72 */
-&(nid_objs[769]),/* OBJ_subject_directory_attributes 2 5 29 9 */
-&(nid_objs[82]),/* OBJ_subject_key_identifier       2 5 29 14 */
-&(nid_objs[83]),/* OBJ_key_usage                    2 5 29 15 */
-&(nid_objs[84]),/* OBJ_private_key_usage_period     2 5 29 16 */
-&(nid_objs[85]),/* OBJ_subject_alt_name             2 5 29 17 */
-&(nid_objs[86]),/* OBJ_issuer_alt_name              2 5 29 18 */
-&(nid_objs[87]),/* OBJ_basic_constraints            2 5 29 19 */
-&(nid_objs[88]),/* OBJ_crl_number                   2 5 29 20 */
-&(nid_objs[141]),/* OBJ_crl_reason                   2 5 29 21 */
-&(nid_objs[430]),/* OBJ_hold_instruction_code        2 5 29 23 */
-&(nid_objs[142]),/* OBJ_invalidity_date              2 5 29 24 */
-&(nid_objs[140]),/* OBJ_delta_crl                    2 5 29 27 */
-&(nid_objs[770]),/* OBJ_issuing_distribution_point   2 5 29 28 */
-&(nid_objs[771]),/* OBJ_certificate_issuer           2 5 29 29 */
-&(nid_objs[666]),/* OBJ_name_constraints             2 5 29 30 */
-&(nid_objs[103]),/* OBJ_crl_distribution_points      2 5 29 31 */
-&(nid_objs[89]),/* OBJ_certificate_policies         2 5 29 32 */
-&(nid_objs[747]),/* OBJ_policy_mappings              2 5 29 33 */
-&(nid_objs[90]),/* OBJ_authority_key_identifier     2 5 29 35 */
-&(nid_objs[401]),/* OBJ_policy_constraints           2 5 29 36 */
-&(nid_objs[126]),/* OBJ_ext_key_usage                2 5 29 37 */
-&(nid_objs[857]),/* OBJ_freshest_crl                 2 5 29 46 */
-&(nid_objs[748]),/* OBJ_inhibit_any_policy           2 5 29 54 */
-&(nid_objs[402]),/* OBJ_target_information           2 5 29 55 */
-&(nid_objs[403]),/* OBJ_no_rev_avail                 2 5 29 56 */
-&(nid_objs[513]),/* OBJ_set_ctype                    2 23 42 0 */
-&(nid_objs[514]),/* OBJ_set_msgExt                   2 23 42 1 */
-&(nid_objs[515]),/* OBJ_set_attr                     2 23 42 3 */
-&(nid_objs[516]),/* OBJ_set_policy                   2 23 42 5 */
-&(nid_objs[517]),/* OBJ_set_certExt                  2 23 42 7 */
-&(nid_objs[518]),/* OBJ_set_brand                    2 23 42 8 */
-&(nid_objs[679]),/* OBJ_wap_wsg                      2 23 43 1 */
-&(nid_objs[382]),/* OBJ_Directory                    1 3 6 1 1 */
-&(nid_objs[383]),/* OBJ_Management                   1 3 6 1 2 */
-&(nid_objs[384]),/* OBJ_Experimental                 1 3 6 1 3 */
-&(nid_objs[385]),/* OBJ_Private                      1 3 6 1 4 */
-&(nid_objs[386]),/* OBJ_Security                     1 3 6 1 5 */
-&(nid_objs[387]),/* OBJ_SNMPv2                       1 3 6 1 6 */
-&(nid_objs[388]),/* OBJ_Mail                         1 3 6 1 7 */
-&(nid_objs[376]),/* OBJ_algorithm                    1 3 14 3 2 */
-&(nid_objs[395]),/* OBJ_clearance                    2 5 1 5 55 */
-&(nid_objs[19]),/* OBJ_rsa                          2 5 8 1 1 */
-&(nid_objs[96]),/* OBJ_mdc2WithRSA                  2 5 8 3 100 */
-&(nid_objs[95]),/* OBJ_mdc2                         2 5 8 3 101 */
-&(nid_objs[746]),/* OBJ_any_policy                   2 5 29 32 0 */
-&(nid_objs[519]),/* OBJ_setct_PANData                2 23 42 0 0 */
-&(nid_objs[520]),/* OBJ_setct_PANToken               2 23 42 0 1 */
-&(nid_objs[521]),/* OBJ_setct_PANOnly                2 23 42 0 2 */
-&(nid_objs[522]),/* OBJ_setct_OIData                 2 23 42 0 3 */
-&(nid_objs[523]),/* OBJ_setct_PI                     2 23 42 0 4 */
-&(nid_objs[524]),/* OBJ_setct_PIData                 2 23 42 0 5 */
-&(nid_objs[525]),/* OBJ_setct_PIDataUnsigned         2 23 42 0 6 */
-&(nid_objs[526]),/* OBJ_setct_HODInput               2 23 42 0 7 */
-&(nid_objs[527]),/* OBJ_setct_AuthResBaggage         2 23 42 0 8 */
-&(nid_objs[528]),/* OBJ_setct_AuthRevReqBaggage      2 23 42 0 9 */
-&(nid_objs[529]),/* OBJ_setct_AuthRevResBaggage      2 23 42 0 10 */
-&(nid_objs[530]),/* OBJ_setct_CapTokenSeq            2 23 42 0 11 */
-&(nid_objs[531]),/* OBJ_setct_PInitResData           2 23 42 0 12 */
-&(nid_objs[532]),/* OBJ_setct_PI_TBS                 2 23 42 0 13 */
-&(nid_objs[533]),/* OBJ_setct_PResData               2 23 42 0 14 */
-&(nid_objs[534]),/* OBJ_setct_AuthReqTBS             2 23 42 0 16 */
-&(nid_objs[535]),/* OBJ_setct_AuthResTBS             2 23 42 0 17 */
-&(nid_objs[536]),/* OBJ_setct_AuthResTBSX            2 23 42 0 18 */
-&(nid_objs[537]),/* OBJ_setct_AuthTokenTBS           2 23 42 0 19 */
-&(nid_objs[538]),/* OBJ_setct_CapTokenData           2 23 42 0 20 */
-&(nid_objs[539]),/* OBJ_setct_CapTokenTBS            2 23 42 0 21 */
-&(nid_objs[540]),/* OBJ_setct_AcqCardCodeMsg         2 23 42 0 22 */
-&(nid_objs[541]),/* OBJ_setct_AuthRevReqTBS          2 23 42 0 23 */
-&(nid_objs[542]),/* OBJ_setct_AuthRevResData         2 23 42 0 24 */
-&(nid_objs[543]),/* OBJ_setct_AuthRevResTBS          2 23 42 0 25 */
-&(nid_objs[544]),/* OBJ_setct_CapReqTBS              2 23 42 0 26 */
-&(nid_objs[545]),/* OBJ_setct_CapReqTBSX             2 23 42 0 27 */
-&(nid_objs[546]),/* OBJ_setct_CapResData             2 23 42 0 28 */
-&(nid_objs[547]),/* OBJ_setct_CapRevReqTBS           2 23 42 0 29 */
-&(nid_objs[548]),/* OBJ_setct_CapRevReqTBSX          2 23 42 0 30 */
-&(nid_objs[549]),/* OBJ_setct_CapRevResData          2 23 42 0 31 */
-&(nid_objs[550]),/* OBJ_setct_CredReqTBS             2 23 42 0 32 */
-&(nid_objs[551]),/* OBJ_setct_CredReqTBSX            2 23 42 0 33 */
-&(nid_objs[552]),/* OBJ_setct_CredResData            2 23 42 0 34 */
-&(nid_objs[553]),/* OBJ_setct_CredRevReqTBS          2 23 42 0 35 */
-&(nid_objs[554]),/* OBJ_setct_CredRevReqTBSX         2 23 42 0 36 */
-&(nid_objs[555]),/* OBJ_setct_CredRevResData         2 23 42 0 37 */
-&(nid_objs[556]),/* OBJ_setct_PCertReqData           2 23 42 0 38 */
-&(nid_objs[557]),/* OBJ_setct_PCertResTBS            2 23 42 0 39 */
-&(nid_objs[558]),/* OBJ_setct_BatchAdminReqData      2 23 42 0 40 */
-&(nid_objs[559]),/* OBJ_setct_BatchAdminResData      2 23 42 0 41 */
-&(nid_objs[560]),/* OBJ_setct_CardCInitResTBS        2 23 42 0 42 */
-&(nid_objs[561]),/* OBJ_setct_MeAqCInitResTBS        2 23 42 0 43 */
-&(nid_objs[562]),/* OBJ_setct_RegFormResTBS          2 23 42 0 44 */
-&(nid_objs[563]),/* OBJ_setct_CertReqData            2 23 42 0 45 */
-&(nid_objs[564]),/* OBJ_setct_CertReqTBS             2 23 42 0 46 */
-&(nid_objs[565]),/* OBJ_setct_CertResData            2 23 42 0 47 */
-&(nid_objs[566]),/* OBJ_setct_CertInqReqTBS          2 23 42 0 48 */
-&(nid_objs[567]),/* OBJ_setct_ErrorTBS               2 23 42 0 49 */
-&(nid_objs[568]),/* OBJ_setct_PIDualSignedTBE        2 23 42 0 50 */
-&(nid_objs[569]),/* OBJ_setct_PIUnsignedTBE          2 23 42 0 51 */
-&(nid_objs[570]),/* OBJ_setct_AuthReqTBE             2 23 42 0 52 */
-&(nid_objs[571]),/* OBJ_setct_AuthResTBE             2 23 42 0 53 */
-&(nid_objs[572]),/* OBJ_setct_AuthResTBEX            2 23 42 0 54 */
-&(nid_objs[573]),/* OBJ_setct_AuthTokenTBE           2 23 42 0 55 */
-&(nid_objs[574]),/* OBJ_setct_CapTokenTBE            2 23 42 0 56 */
-&(nid_objs[575]),/* OBJ_setct_CapTokenTBEX           2 23 42 0 57 */
-&(nid_objs[576]),/* OBJ_setct_AcqCardCodeMsgTBE      2 23 42 0 58 */
-&(nid_objs[577]),/* OBJ_setct_AuthRevReqTBE          2 23 42 0 59 */
-&(nid_objs[578]),/* OBJ_setct_AuthRevResTBE          2 23 42 0 60 */
-&(nid_objs[579]),/* OBJ_setct_AuthRevResTBEB         2 23 42 0 61 */
-&(nid_objs[580]),/* OBJ_setct_CapReqTBE              2 23 42 0 62 */
-&(nid_objs[581]),/* OBJ_setct_CapReqTBEX             2 23 42 0 63 */
-&(nid_objs[582]),/* OBJ_setct_CapResTBE              2 23 42 0 64 */
-&(nid_objs[583]),/* OBJ_setct_CapRevReqTBE           2 23 42 0 65 */
-&(nid_objs[584]),/* OBJ_setct_CapRevReqTBEX          2 23 42 0 66 */
-&(nid_objs[585]),/* OBJ_setct_CapRevResTBE           2 23 42 0 67 */
-&(nid_objs[586]),/* OBJ_setct_CredReqTBE             2 23 42 0 68 */
-&(nid_objs[587]),/* OBJ_setct_CredReqTBEX            2 23 42 0 69 */
-&(nid_objs[588]),/* OBJ_setct_CredResTBE             2 23 42 0 70 */
-&(nid_objs[589]),/* OBJ_setct_CredRevReqTBE          2 23 42 0 71 */
-&(nid_objs[590]),/* OBJ_setct_CredRevReqTBEX         2 23 42 0 72 */
-&(nid_objs[591]),/* OBJ_setct_CredRevResTBE          2 23 42 0 73 */
-&(nid_objs[592]),/* OBJ_setct_BatchAdminReqTBE       2 23 42 0 74 */
-&(nid_objs[593]),/* OBJ_setct_BatchAdminResTBE       2 23 42 0 75 */
-&(nid_objs[594]),/* OBJ_setct_RegFormReqTBE          2 23 42 0 76 */
-&(nid_objs[595]),/* OBJ_setct_CertReqTBE             2 23 42 0 77 */
-&(nid_objs[596]),/* OBJ_setct_CertReqTBEX            2 23 42 0 78 */
-&(nid_objs[597]),/* OBJ_setct_CertResTBE             2 23 42 0 79 */
-&(nid_objs[598]),/* OBJ_setct_CRLNotificationTBS     2 23 42 0 80 */
-&(nid_objs[599]),/* OBJ_setct_CRLNotificationResTBS  2 23 42 0 81 */
-&(nid_objs[600]),/* OBJ_setct_BCIDistributionTBS     2 23 42 0 82 */
-&(nid_objs[601]),/* OBJ_setext_genCrypt              2 23 42 1 1 */
-&(nid_objs[602]),/* OBJ_setext_miAuth                2 23 42 1 3 */
-&(nid_objs[603]),/* OBJ_setext_pinSecure             2 23 42 1 4 */
-&(nid_objs[604]),/* OBJ_setext_pinAny                2 23 42 1 5 */
-&(nid_objs[605]),/* OBJ_setext_track2                2 23 42 1 7 */
-&(nid_objs[606]),/* OBJ_setext_cv                    2 23 42 1 8 */
-&(nid_objs[620]),/* OBJ_setAttr_Cert                 2 23 42 3 0 */
-&(nid_objs[621]),/* OBJ_setAttr_PGWYcap              2 23 42 3 1 */
-&(nid_objs[622]),/* OBJ_setAttr_TokenType            2 23 42 3 2 */
-&(nid_objs[623]),/* OBJ_setAttr_IssCap               2 23 42 3 3 */
-&(nid_objs[607]),/* OBJ_set_policy_root              2 23 42 5 0 */
-&(nid_objs[608]),/* OBJ_setCext_hashedRoot           2 23 42 7 0 */
-&(nid_objs[609]),/* OBJ_setCext_certType             2 23 42 7 1 */
-&(nid_objs[610]),/* OBJ_setCext_merchData            2 23 42 7 2 */
-&(nid_objs[611]),/* OBJ_setCext_cCertRequired        2 23 42 7 3 */
-&(nid_objs[612]),/* OBJ_setCext_tunneling            2 23 42 7 4 */
-&(nid_objs[613]),/* OBJ_setCext_setExt               2 23 42 7 5 */
-&(nid_objs[614]),/* OBJ_setCext_setQualf             2 23 42 7 6 */
-&(nid_objs[615]),/* OBJ_setCext_PGWYcapabilities     2 23 42 7 7 */
-&(nid_objs[616]),/* OBJ_setCext_TokenIdentifier      2 23 42 7 8 */
-&(nid_objs[617]),/* OBJ_setCext_Track2Data           2 23 42 7 9 */
-&(nid_objs[618]),/* OBJ_setCext_TokenType            2 23 42 7 10 */
-&(nid_objs[619]),/* OBJ_setCext_IssuerCapabilities   2 23 42 7 11 */
-&(nid_objs[636]),/* OBJ_set_brand_IATA_ATA           2 23 42 8 1 */
-&(nid_objs[640]),/* OBJ_set_brand_Visa               2 23 42 8 4 */
-&(nid_objs[641]),/* OBJ_set_brand_MasterCard         2 23 42 8 5 */
-&(nid_objs[637]),/* OBJ_set_brand_Diners             2 23 42 8 30 */
-&(nid_objs[638]),/* OBJ_set_brand_AmericanExpress    2 23 42 8 34 */
-&(nid_objs[639]),/* OBJ_set_brand_JCB                2 23 42 8 35 */
-&(nid_objs[805]),/* OBJ_cryptopro                    1 2 643 2 2 */
-&(nid_objs[806]),/* OBJ_cryptocom                    1 2 643 2 9 */
-&(nid_objs[184]),/* OBJ_X9_57                        1 2 840 10040 */
-&(nid_objs[405]),/* OBJ_ansi_X9_62                   1 2 840 10045 */
-&(nid_objs[389]),/* OBJ_Enterprises                  1 3 6 1 4 1 */
-&(nid_objs[504]),/* OBJ_mime_mhs                     1 3 6 1 7 1 */
-&(nid_objs[104]),/* OBJ_md5WithRSA                   1 3 14 3 2 3 */
-&(nid_objs[29]),/* OBJ_des_ecb                      1 3 14 3 2 6 */
-&(nid_objs[31]),/* OBJ_des_cbc                      1 3 14 3 2 7 */
-&(nid_objs[45]),/* OBJ_des_ofb64                    1 3 14 3 2 8 */
-&(nid_objs[30]),/* OBJ_des_cfb64                    1 3 14 3 2 9 */
-&(nid_objs[377]),/* OBJ_rsaSignature                 1 3 14 3 2 11 */
-&(nid_objs[67]),/* OBJ_dsa_2                        1 3 14 3 2 12 */
-&(nid_objs[66]),/* OBJ_dsaWithSHA                   1 3 14 3 2 13 */
-&(nid_objs[42]),/* OBJ_shaWithRSAEncryption         1 3 14 3 2 15 */
-&(nid_objs[32]),/* OBJ_des_ede_ecb                  1 3 14 3 2 17 */
-&(nid_objs[41]),/* OBJ_sha                          1 3 14 3 2 18 */
-&(nid_objs[64]),/* OBJ_sha1                         1 3 14 3 2 26 */
-&(nid_objs[70]),/* OBJ_dsaWithSHA1_2                1 3 14 3 2 27 */
-&(nid_objs[115]),/* OBJ_sha1WithRSA                  1 3 14 3 2 29 */
-&(nid_objs[117]),/* OBJ_ripemd160                    1 3 36 3 2 1 */
-&(nid_objs[143]),/* OBJ_sxnet                        1 3 101 1 4 1 */
-&(nid_objs[721]),/* OBJ_sect163k1                    1 3 132 0 1 */
-&(nid_objs[722]),/* OBJ_sect163r1                    1 3 132 0 2 */
-&(nid_objs[728]),/* OBJ_sect239k1                    1 3 132 0 3 */
-&(nid_objs[717]),/* OBJ_sect113r1                    1 3 132 0 4 */
-&(nid_objs[718]),/* OBJ_sect113r2                    1 3 132 0 5 */
-&(nid_objs[704]),/* OBJ_secp112r1                    1 3 132 0 6 */
-&(nid_objs[705]),/* OBJ_secp112r2                    1 3 132 0 7 */
-&(nid_objs[709]),/* OBJ_secp160r1                    1 3 132 0 8 */
-&(nid_objs[708]),/* OBJ_secp160k1                    1 3 132 0 9 */
-&(nid_objs[714]),/* OBJ_secp256k1                    1 3 132 0 10 */
-&(nid_objs[723]),/* OBJ_sect163r2                    1 3 132 0 15 */
-&(nid_objs[729]),/* OBJ_sect283k1                    1 3 132 0 16 */
-&(nid_objs[730]),/* OBJ_sect283r1                    1 3 132 0 17 */
-&(nid_objs[719]),/* OBJ_sect131r1                    1 3 132 0 22 */
-&(nid_objs[720]),/* OBJ_sect131r2                    1 3 132 0 23 */
-&(nid_objs[724]),/* OBJ_sect193r1                    1 3 132 0 24 */
-&(nid_objs[725]),/* OBJ_sect193r2                    1 3 132 0 25 */
-&(nid_objs[726]),/* OBJ_sect233k1                    1 3 132 0 26 */
-&(nid_objs[727]),/* OBJ_sect233r1                    1 3 132 0 27 */
-&(nid_objs[706]),/* OBJ_secp128r1                    1 3 132 0 28 */
-&(nid_objs[707]),/* OBJ_secp128r2                    1 3 132 0 29 */
-&(nid_objs[710]),/* OBJ_secp160r2                    1 3 132 0 30 */
-&(nid_objs[711]),/* OBJ_secp192k1                    1 3 132 0 31 */
-&(nid_objs[712]),/* OBJ_secp224k1                    1 3 132 0 32 */
-&(nid_objs[713]),/* OBJ_secp224r1                    1 3 132 0 33 */
-&(nid_objs[715]),/* OBJ_secp384r1                    1 3 132 0 34 */
-&(nid_objs[716]),/* OBJ_secp521r1                    1 3 132 0 35 */
-&(nid_objs[731]),/* OBJ_sect409k1                    1 3 132 0 36 */
-&(nid_objs[732]),/* OBJ_sect409r1                    1 3 132 0 37 */
-&(nid_objs[733]),/* OBJ_sect571k1                    1 3 132 0 38 */
-&(nid_objs[734]),/* OBJ_sect571r1                    1 3 132 0 39 */
-&(nid_objs[624]),/* OBJ_set_rootKeyThumb             2 23 42 3 0 0 */
-&(nid_objs[625]),/* OBJ_set_addPolicy                2 23 42 3 0 1 */
-&(nid_objs[626]),/* OBJ_setAttr_Token_EMV            2 23 42 3 2 1 */
-&(nid_objs[627]),/* OBJ_setAttr_Token_B0Prime        2 23 42 3 2 2 */
-&(nid_objs[628]),/* OBJ_setAttr_IssCap_CVM           2 23 42 3 3 3 */
-&(nid_objs[629]),/* OBJ_setAttr_IssCap_T2            2 23 42 3 3 4 */
-&(nid_objs[630]),/* OBJ_setAttr_IssCap_Sig           2 23 42 3 3 5 */
-&(nid_objs[642]),/* OBJ_set_brand_Novus              2 23 42 8 6011 */
-&(nid_objs[735]),/* OBJ_wap_wsg_idm_ecid_wtls1       2 23 43 1 4 1 */
-&(nid_objs[736]),/* OBJ_wap_wsg_idm_ecid_wtls3       2 23 43 1 4 3 */
-&(nid_objs[737]),/* OBJ_wap_wsg_idm_ecid_wtls4       2 23 43 1 4 4 */
-&(nid_objs[738]),/* OBJ_wap_wsg_idm_ecid_wtls5       2 23 43 1 4 5 */
-&(nid_objs[739]),/* OBJ_wap_wsg_idm_ecid_wtls6       2 23 43 1 4 6 */
-&(nid_objs[740]),/* OBJ_wap_wsg_idm_ecid_wtls7       2 23 43 1 4 7 */
-&(nid_objs[741]),/* OBJ_wap_wsg_idm_ecid_wtls8       2 23 43 1 4 8 */
-&(nid_objs[742]),/* OBJ_wap_wsg_idm_ecid_wtls9       2 23 43 1 4 9 */
-&(nid_objs[743]),/* OBJ_wap_wsg_idm_ecid_wtls10      2 23 43 1 4 10 */
-&(nid_objs[744]),/* OBJ_wap_wsg_idm_ecid_wtls11      2 23 43 1 4 11 */
-&(nid_objs[745]),/* OBJ_wap_wsg_idm_ecid_wtls12      2 23 43 1 4 12 */
-&(nid_objs[804]),/* OBJ_whirlpool                    1 0 10118 3 0 55 */
-&(nid_objs[124]),/* OBJ_rle_compression              1 1 1 1 666 1 */
-&(nid_objs[773]),/* OBJ_kisa                         1 2 410 200004 */
-&(nid_objs[807]),/* OBJ_id_GostR3411_94_with_GostR3410_2001 1 2 643 2 2 3 */
-&(nid_objs[808]),/* OBJ_id_GostR3411_94_with_GostR3410_94 1 2 643 2 2 4 */
-&(nid_objs[809]),/* OBJ_id_GostR3411_94              1 2 643 2 2 9 */
-&(nid_objs[810]),/* OBJ_id_HMACGostR3411_94          1 2 643 2 2 10 */
-&(nid_objs[811]),/* OBJ_id_GostR3410_2001            1 2 643 2 2 19 */
-&(nid_objs[812]),/* OBJ_id_GostR3410_94              1 2 643 2 2 20 */
-&(nid_objs[813]),/* OBJ_id_Gost28147_89              1 2 643 2 2 21 */
-&(nid_objs[815]),/* OBJ_id_Gost28147_89_MAC          1 2 643 2 2 22 */
-&(nid_objs[816]),/* OBJ_id_GostR3411_94_prf          1 2 643 2 2 23 */
-&(nid_objs[817]),/* OBJ_id_GostR3410_2001DH          1 2 643 2 2 98 */
-&(nid_objs[818]),/* OBJ_id_GostR3410_94DH            1 2 643 2 2 99 */
-&(nid_objs[ 1]),/* OBJ_rsadsi                       1 2 840 113549 */
-&(nid_objs[185]),/* OBJ_X9cm                         1 2 840 10040 4 */
-&(nid_objs[127]),/* OBJ_id_pkix                      1 3 6 1 5 5 7 */
-&(nid_objs[505]),/* OBJ_mime_mhs_headings            1 3 6 1 7 1 1 */
-&(nid_objs[506]),/* OBJ_mime_mhs_bodies              1 3 6 1 7 1 2 */
-&(nid_objs[119]),/* OBJ_ripemd160WithRSA             1 3 36 3 3 1 2 */
-&(nid_objs[631]),/* OBJ_setAttr_GenCryptgrm          2 23 42 3 3 3 1 */
-&(nid_objs[632]),/* OBJ_setAttr_T2Enc                2 23 42 3 3 4 1 */
-&(nid_objs[633]),/* OBJ_setAttr_T2cleartxt           2 23 42 3 3 4 2 */
-&(nid_objs[634]),/* OBJ_setAttr_TokICCsig            2 23 42 3 3 5 1 */
-&(nid_objs[635]),/* OBJ_setAttr_SecDevSig            2 23 42 3 3 5 2 */
-&(nid_objs[436]),/* OBJ_ucl                          0 9 2342 19200300 */
-&(nid_objs[820]),/* OBJ_id_Gost28147_89_None_KeyMeshing 1 2 643 2 2 14 0 */
-&(nid_objs[819]),/* OBJ_id_Gost28147_89_CryptoPro_KeyMeshing 1 2 643 2 2 14 1 */
-&(nid_objs[845]),/* OBJ_id_GostR3410_94_a            1 2 643 2 2 20 1 */
-&(nid_objs[846]),/* OBJ_id_GostR3410_94_aBis         1 2 643 2 2 20 2 */
-&(nid_objs[847]),/* OBJ_id_GostR3410_94_b            1 2 643 2 2 20 3 */
-&(nid_objs[848]),/* OBJ_id_GostR3410_94_bBis         1 2 643 2 2 20 4 */
-&(nid_objs[821]),/* OBJ_id_GostR3411_94_TestParamSet 1 2 643 2 2 30 0 */
-&(nid_objs[822]),/* OBJ_id_GostR3411_94_CryptoProParamSet 1 2 643 2 2 30 1 */
-&(nid_objs[823]),/* OBJ_id_Gost28147_89_TestParamSet 1 2 643 2 2 31 0 */
-&(nid_objs[824]),/* OBJ_id_Gost28147_89_CryptoPro_A_ParamSet 1 2 643 2 2 31 1 */
-&(nid_objs[825]),/* OBJ_id_Gost28147_89_CryptoPro_B_ParamSet 1 2 643 2 2 31 2 */
-&(nid_objs[826]),/* OBJ_id_Gost28147_89_CryptoPro_C_ParamSet 1 2 643 2 2 31 3 */
-&(nid_objs[827]),/* OBJ_id_Gost28147_89_CryptoPro_D_ParamSet 1 2 643 2 2 31 4 */
-&(nid_objs[828]),/* OBJ_id_Gost28147_89_CryptoPro_Oscar_1_1_ParamSet 1 2 643 2 2 31 5 */
-&(nid_objs[829]),/* OBJ_id_Gost28147_89_CryptoPro_Oscar_1_0_ParamSet 1 2 643 2 2 31 6 */
-&(nid_objs[830]),/* OBJ_id_Gost28147_89_CryptoPro_RIC_1_ParamSet 1 2 643 2 2 31 7 */
-&(nid_objs[831]),/* OBJ_id_GostR3410_94_TestParamSet 1 2 643 2 2 32 0 */
-&(nid_objs[832]),/* OBJ_id_GostR3410_94_CryptoPro_A_ParamSet 1 2 643 2 2 32 2 */
-&(nid_objs[833]),/* OBJ_id_GostR3410_94_CryptoPro_B_ParamSet 1 2 643 2 2 32 3 */
-&(nid_objs[834]),/* OBJ_id_GostR3410_94_CryptoPro_C_ParamSet 1 2 643 2 2 32 4 */
-&(nid_objs[835]),/* OBJ_id_GostR3410_94_CryptoPro_D_ParamSet 1 2 643 2 2 32 5 */
-&(nid_objs[836]),/* OBJ_id_GostR3410_94_CryptoPro_XchA_ParamSet 1 2 643 2 2 33 1 */
-&(nid_objs[837]),/* OBJ_id_GostR3410_94_CryptoPro_XchB_ParamSet 1 2 643 2 2 33 2 */
-&(nid_objs[838]),/* OBJ_id_GostR3410_94_CryptoPro_XchC_ParamSet 1 2 643 2 2 33 3 */
-&(nid_objs[839]),/* OBJ_id_GostR3410_2001_TestParamSet 1 2 643 2 2 35 0 */
-&(nid_objs[840]),/* OBJ_id_GostR3410_2001_CryptoPro_A_ParamSet 1 2 643 2 2 35 1 */
-&(nid_objs[841]),/* OBJ_id_GostR3410_2001_CryptoPro_B_ParamSet 1 2 643 2 2 35 2 */
-&(nid_objs[842]),/* OBJ_id_GostR3410_2001_CryptoPro_C_ParamSet 1 2 643 2 2 35 3 */
-&(nid_objs[843]),/* OBJ_id_GostR3410_2001_CryptoPro_XchA_ParamSet 1 2 643 2 2 36 0 */
-&(nid_objs[844]),/* OBJ_id_GostR3410_2001_CryptoPro_XchB_ParamSet 1 2 643 2 2 36 1 */
-&(nid_objs[ 2]),/* OBJ_pkcs                         1 2 840 113549 1 */
-&(nid_objs[431]),/* OBJ_hold_instruction_none        1 2 840 10040 2 1 */
-&(nid_objs[432]),/* OBJ_hold_instruction_call_issuer 1 2 840 10040 2 2 */
-&(nid_objs[433]),/* OBJ_hold_instruction_reject      1 2 840 10040 2 3 */
-&(nid_objs[116]),/* OBJ_dsa                          1 2 840 10040 4 1 */
-&(nid_objs[113]),/* OBJ_dsaWithSHA1                  1 2 840 10040 4 3 */
-&(nid_objs[406]),/* OBJ_X9_62_prime_field            1 2 840 10045 1 1 */
-&(nid_objs[407]),/* OBJ_X9_62_characteristic_two_field 1 2 840 10045 1 2 */
-&(nid_objs[408]),/* OBJ_X9_62_id_ecPublicKey         1 2 840 10045 2 1 */
-&(nid_objs[416]),/* OBJ_ecdsa_with_SHA1              1 2 840 10045 4 1 */
-&(nid_objs[791]),/* OBJ_ecdsa_with_Recommended       1 2 840 10045 4 2 */
-&(nid_objs[792]),/* OBJ_ecdsa_with_Specified         1 2 840 10045 4 3 */
-&(nid_objs[258]),/* OBJ_id_pkix_mod                  1 3 6 1 5 5 7 0 */
-&(nid_objs[175]),/* OBJ_id_pe                        1 3 6 1 5 5 7 1 */
-&(nid_objs[259]),/* OBJ_id_qt                        1 3 6 1 5 5 7 2 */
-&(nid_objs[128]),/* OBJ_id_kp                        1 3 6 1 5 5 7 3 */
-&(nid_objs[260]),/* OBJ_id_it                        1 3 6 1 5 5 7 4 */
-&(nid_objs[261]),/* OBJ_id_pkip                      1 3 6 1 5 5 7 5 */
-&(nid_objs[262]),/* OBJ_id_alg                       1 3 6 1 5 5 7 6 */
-&(nid_objs[263]),/* OBJ_id_cmc                       1 3 6 1 5 5 7 7 */
-&(nid_objs[264]),/* OBJ_id_on                        1 3 6 1 5 5 7 8 */
-&(nid_objs[265]),/* OBJ_id_pda                       1 3 6 1 5 5 7 9 */
-&(nid_objs[266]),/* OBJ_id_aca                       1 3 6 1 5 5 7 10 */
-&(nid_objs[267]),/* OBJ_id_qcs                       1 3 6 1 5 5 7 11 */
-&(nid_objs[268]),/* OBJ_id_cct                       1 3 6 1 5 5 7 12 */
-&(nid_objs[662]),/* OBJ_id_ppl                       1 3 6 1 5 5 7 21 */
-&(nid_objs[176]),/* OBJ_id_ad                        1 3 6 1 5 5 7 48 */
-&(nid_objs[507]),/* OBJ_id_hex_partial_message       1 3 6 1 7 1 1 1 */
-&(nid_objs[508]),/* OBJ_id_hex_multipart_message     1 3 6 1 7 1 1 2 */
-&(nid_objs[57]),/* OBJ_netscape                     2 16 840 1 113730 */
-&(nid_objs[754]),/* OBJ_camellia_128_ecb             0 3 4401 5 3 1 9 1 */
-&(nid_objs[766]),/* OBJ_camellia_128_ofb128          0 3 4401 5 3 1 9 3 */
-&(nid_objs[757]),/* OBJ_camellia_128_cfb128          0 3 4401 5 3 1 9 4 */
-&(nid_objs[755]),/* OBJ_camellia_192_ecb             0 3 4401 5 3 1 9 21 */
-&(nid_objs[767]),/* OBJ_camellia_192_ofb128          0 3 4401 5 3 1 9 23 */
-&(nid_objs[758]),/* OBJ_camellia_192_cfb128          0 3 4401 5 3 1 9 24 */
-&(nid_objs[756]),/* OBJ_camellia_256_ecb             0 3 4401 5 3 1 9 41 */
-&(nid_objs[768]),/* OBJ_camellia_256_ofb128          0 3 4401 5 3 1 9 43 */
-&(nid_objs[759]),/* OBJ_camellia_256_cfb128          0 3 4401 5 3 1 9 44 */
-&(nid_objs[437]),/* OBJ_pilot                        0 9 2342 19200300 100 */
-&(nid_objs[776]),/* OBJ_seed_ecb                     1 2 410 200004 1 3 */
-&(nid_objs[777]),/* OBJ_seed_cbc                     1 2 410 200004 1 4 */
-&(nid_objs[779]),/* OBJ_seed_cfb128                  1 2 410 200004 1 5 */
-&(nid_objs[778]),/* OBJ_seed_ofb128                  1 2 410 200004 1 6 */
-&(nid_objs[852]),/* OBJ_id_GostR3411_94_with_GostR3410_94_cc 1 2 643 2 9 1 3 3 */
-&(nid_objs[853]),/* OBJ_id_GostR3411_94_with_GostR3410_2001_cc 1 2 643 2 9 1 3 4 */
-&(nid_objs[850]),/* OBJ_id_GostR3410_94_cc           1 2 643 2 9 1 5 3 */
-&(nid_objs[851]),/* OBJ_id_GostR3410_2001_cc         1 2 643 2 9 1 5 4 */
-&(nid_objs[849]),/* OBJ_id_Gost28147_89_cc           1 2 643 2 9 1 6 1 */
-&(nid_objs[854]),/* OBJ_id_GostR3410_2001_ParamSet_cc 1 2 643 2 9 1 8 1 */
-&(nid_objs[186]),/* OBJ_pkcs1                        1 2 840 113549 1 1 */
-&(nid_objs[27]),/* OBJ_pkcs3                        1 2 840 113549 1 3 */
-&(nid_objs[187]),/* OBJ_pkcs5                        1 2 840 113549 1 5 */
-&(nid_objs[20]),/* OBJ_pkcs7                        1 2 840 113549 1 7 */
-&(nid_objs[47]),/* OBJ_pkcs9                        1 2 840 113549 1 9 */
-&(nid_objs[ 3]),/* OBJ_md2                          1 2 840 113549 2 2 */
-&(nid_objs[257]),/* OBJ_md4                          1 2 840 113549 2 4 */
-&(nid_objs[ 4]),/* OBJ_md5                          1 2 840 113549 2 5 */
-&(nid_objs[797]),/* OBJ_hmacWithMD5                  1 2 840 113549 2 6 */
-&(nid_objs[163]),/* OBJ_hmacWithSHA1                 1 2 840 113549 2 7 */
-&(nid_objs[798]),/* OBJ_hmacWithSHA224               1 2 840 113549 2 8 */
-&(nid_objs[799]),/* OBJ_hmacWithSHA256               1 2 840 113549 2 9 */
-&(nid_objs[800]),/* OBJ_hmacWithSHA384               1 2 840 113549 2 10 */
-&(nid_objs[801]),/* OBJ_hmacWithSHA512               1 2 840 113549 2 11 */
-&(nid_objs[37]),/* OBJ_rc2_cbc                      1 2 840 113549 3 2 */
-&(nid_objs[ 5]),/* OBJ_rc4                          1 2 840 113549 3 4 */
-&(nid_objs[44]),/* OBJ_des_ede3_cbc                 1 2 840 113549 3 7 */
-&(nid_objs[120]),/* OBJ_rc5_cbc                      1 2 840 113549 3 8 */
-&(nid_objs[643]),/* OBJ_des_cdmf                     1 2 840 113549 3 10 */
-&(nid_objs[680]),/* OBJ_X9_62_id_characteristic_two_basis 1 2 840 10045 1 2 3 */
-&(nid_objs[684]),/* OBJ_X9_62_c2pnb163v1             1 2 840 10045 3 0 1 */
-&(nid_objs[685]),/* OBJ_X9_62_c2pnb163v2             1 2 840 10045 3 0 2 */
-&(nid_objs[686]),/* OBJ_X9_62_c2pnb163v3             1 2 840 10045 3 0 3 */
-&(nid_objs[687]),/* OBJ_X9_62_c2pnb176v1             1 2 840 10045 3 0 4 */
-&(nid_objs[688]),/* OBJ_X9_62_c2tnb191v1             1 2 840 10045 3 0 5 */
-&(nid_objs[689]),/* OBJ_X9_62_c2tnb191v2             1 2 840 10045 3 0 6 */
-&(nid_objs[690]),/* OBJ_X9_62_c2tnb191v3             1 2 840 10045 3 0 7 */
-&(nid_objs[691]),/* OBJ_X9_62_c2onb191v4             1 2 840 10045 3 0 8 */
-&(nid_objs[692]),/* OBJ_X9_62_c2onb191v5             1 2 840 10045 3 0 9 */
-&(nid_objs[693]),/* OBJ_X9_62_c2pnb208w1             1 2 840 10045 3 0 10 */
-&(nid_objs[694]),/* OBJ_X9_62_c2tnb239v1             1 2 840 10045 3 0 11 */
-&(nid_objs[695]),/* OBJ_X9_62_c2tnb239v2             1 2 840 10045 3 0 12 */
-&(nid_objs[696]),/* OBJ_X9_62_c2tnb239v3             1 2 840 10045 3 0 13 */
-&(nid_objs[697]),/* OBJ_X9_62_c2onb239v4             1 2 840 10045 3 0 14 */
-&(nid_objs[698]),/* OBJ_X9_62_c2onb239v5             1 2 840 10045 3 0 15 */
-&(nid_objs[699]),/* OBJ_X9_62_c2pnb272w1             1 2 840 10045 3 0 16 */
-&(nid_objs[700]),/* OBJ_X9_62_c2pnb304w1             1 2 840 10045 3 0 17 */
-&(nid_objs[701]),/* OBJ_X9_62_c2tnb359v1             1 2 840 10045 3 0 18 */
-&(nid_objs[702]),/* OBJ_X9_62_c2pnb368w1             1 2 840 10045 3 0 19 */
-&(nid_objs[703]),/* OBJ_X9_62_c2tnb431r1             1 2 840 10045 3 0 20 */
-&(nid_objs[409]),/* OBJ_X9_62_prime192v1             1 2 840 10045 3 1 1 */
-&(nid_objs[410]),/* OBJ_X9_62_prime192v2             1 2 840 10045 3 1 2 */
-&(nid_objs[411]),/* OBJ_X9_62_prime192v3             1 2 840 10045 3 1 3 */
-&(nid_objs[412]),/* OBJ_X9_62_prime239v1             1 2 840 10045 3 1 4 */
-&(nid_objs[413]),/* OBJ_X9_62_prime239v2             1 2 840 10045 3 1 5 */
-&(nid_objs[414]),/* OBJ_X9_62_prime239v3             1 2 840 10045 3 1 6 */
-&(nid_objs[415]),/* OBJ_X9_62_prime256v1             1 2 840 10045 3 1 7 */
-&(nid_objs[793]),/* OBJ_ecdsa_with_SHA224            1 2 840 10045 4 3 1 */
-&(nid_objs[794]),/* OBJ_ecdsa_with_SHA256            1 2 840 10045 4 3 2 */
-&(nid_objs[795]),/* OBJ_ecdsa_with_SHA384            1 2 840 10045 4 3 3 */
-&(nid_objs[796]),/* OBJ_ecdsa_with_SHA512            1 2 840 10045 4 3 4 */
-&(nid_objs[269]),/* OBJ_id_pkix1_explicit_88         1 3 6 1 5 5 7 0 1 */
-&(nid_objs[270]),/* OBJ_id_pkix1_implicit_88         1 3 6 1 5 5 7 0 2 */
-&(nid_objs[271]),/* OBJ_id_pkix1_explicit_93         1 3 6 1 5 5 7 0 3 */
-&(nid_objs[272]),/* OBJ_id_pkix1_implicit_93         1 3 6 1 5 5 7 0 4 */
-&(nid_objs[273]),/* OBJ_id_mod_crmf                  1 3 6 1 5 5 7 0 5 */
-&(nid_objs[274]),/* OBJ_id_mod_cmc                   1 3 6 1 5 5 7 0 6 */
-&(nid_objs[275]),/* OBJ_id_mod_kea_profile_88        1 3 6 1 5 5 7 0 7 */
-&(nid_objs[276]),/* OBJ_id_mod_kea_profile_93        1 3 6 1 5 5 7 0 8 */
-&(nid_objs[277]),/* OBJ_id_mod_cmp                   1 3 6 1 5 5 7 0 9 */
-&(nid_objs[278]),/* OBJ_id_mod_qualified_cert_88     1 3 6 1 5 5 7 0 10 */
-&(nid_objs[279]),/* OBJ_id_mod_qualified_cert_93     1 3 6 1 5 5 7 0 11 */
-&(nid_objs[280]),/* OBJ_id_mod_attribute_cert        1 3 6 1 5 5 7 0 12 */
-&(nid_objs[281]),/* OBJ_id_mod_timestamp_protocol    1 3 6 1 5 5 7 0 13 */
-&(nid_objs[282]),/* OBJ_id_mod_ocsp                  1 3 6 1 5 5 7 0 14 */
-&(nid_objs[283]),/* OBJ_id_mod_dvcs                  1 3 6 1 5 5 7 0 15 */
-&(nid_objs[284]),/* OBJ_id_mod_cmp2000               1 3 6 1 5 5 7 0 16 */
-&(nid_objs[177]),/* OBJ_info_access                  1 3 6 1 5 5 7 1 1 */
-&(nid_objs[285]),/* OBJ_biometricInfo                1 3 6 1 5 5 7 1 2 */
-&(nid_objs[286]),/* OBJ_qcStatements                 1 3 6 1 5 5 7 1 3 */
-&(nid_objs[287]),/* OBJ_ac_auditEntity               1 3 6 1 5 5 7 1 4 */
-&(nid_objs[288]),/* OBJ_ac_targeting                 1 3 6 1 5 5 7 1 5 */
-&(nid_objs[289]),/* OBJ_aaControls                   1 3 6 1 5 5 7 1 6 */
-&(nid_objs[290]),/* OBJ_sbgp_ipAddrBlock             1 3 6 1 5 5 7 1 7 */
-&(nid_objs[291]),/* OBJ_sbgp_autonomousSysNum        1 3 6 1 5 5 7 1 8 */
-&(nid_objs[292]),/* OBJ_sbgp_routerIdentifier        1 3 6 1 5 5 7 1 9 */
-&(nid_objs[397]),/* OBJ_ac_proxying                  1 3 6 1 5 5 7 1 10 */
-&(nid_objs[398]),/* OBJ_sinfo_access                 1 3 6 1 5 5 7 1 11 */
-&(nid_objs[663]),/* OBJ_proxyCertInfo                1 3 6 1 5 5 7 1 14 */
-&(nid_objs[164]),/* OBJ_id_qt_cps                    1 3 6 1 5 5 7 2 1 */
-&(nid_objs[165]),/* OBJ_id_qt_unotice                1 3 6 1 5 5 7 2 2 */
-&(nid_objs[293]),/* OBJ_textNotice                   1 3 6 1 5 5 7 2 3 */
-&(nid_objs[129]),/* OBJ_server_auth                  1 3 6 1 5 5 7 3 1 */
-&(nid_objs[130]),/* OBJ_client_auth                  1 3 6 1 5 5 7 3 2 */
-&(nid_objs[131]),/* OBJ_code_sign                    1 3 6 1 5 5 7 3 3 */
-&(nid_objs[132]),/* OBJ_email_protect                1 3 6 1 5 5 7 3 4 */
-&(nid_objs[294]),/* OBJ_ipsecEndSystem               1 3 6 1 5 5 7 3 5 */
-&(nid_objs[295]),/* OBJ_ipsecTunnel                  1 3 6 1 5 5 7 3 6 */
-&(nid_objs[296]),/* OBJ_ipsecUser                    1 3 6 1 5 5 7 3 7 */
-&(nid_objs[133]),/* OBJ_time_stamp                   1 3 6 1 5 5 7 3 8 */
-&(nid_objs[180]),/* OBJ_OCSP_sign                    1 3 6 1 5 5 7 3 9 */
-&(nid_objs[297]),/* OBJ_dvcs                         1 3 6 1 5 5 7 3 10 */
-&(nid_objs[298]),/* OBJ_id_it_caProtEncCert          1 3 6 1 5 5 7 4 1 */
-&(nid_objs[299]),/* OBJ_id_it_signKeyPairTypes       1 3 6 1 5 5 7 4 2 */
-&(nid_objs[300]),/* OBJ_id_it_encKeyPairTypes        1 3 6 1 5 5 7 4 3 */
-&(nid_objs[301]),/* OBJ_id_it_preferredSymmAlg       1 3 6 1 5 5 7 4 4 */
-&(nid_objs[302]),/* OBJ_id_it_caKeyUpdateInfo        1 3 6 1 5 5 7 4 5 */
-&(nid_objs[303]),/* OBJ_id_it_currentCRL             1 3 6 1 5 5 7 4 6 */
-&(nid_objs[304]),/* OBJ_id_it_unsupportedOIDs        1 3 6 1 5 5 7 4 7 */
-&(nid_objs[305]),/* OBJ_id_it_subscriptionRequest    1 3 6 1 5 5 7 4 8 */
-&(nid_objs[306]),/* OBJ_id_it_subscriptionResponse   1 3 6 1 5 5 7 4 9 */
-&(nid_objs[307]),/* OBJ_id_it_keyPairParamReq        1 3 6 1 5 5 7 4 10 */
-&(nid_objs[308]),/* OBJ_id_it_keyPairParamRep        1 3 6 1 5 5 7 4 11 */
-&(nid_objs[309]),/* OBJ_id_it_revPassphrase          1 3 6 1 5 5 7 4 12 */
-&(nid_objs[310]),/* OBJ_id_it_implicitConfirm        1 3 6 1 5 5 7 4 13 */
-&(nid_objs[311]),/* OBJ_id_it_confirmWaitTime        1 3 6 1 5 5 7 4 14 */
-&(nid_objs[312]),/* OBJ_id_it_origPKIMessage         1 3 6 1 5 5 7 4 15 */
-&(nid_objs[784]),/* OBJ_id_it_suppLangTags           1 3 6 1 5 5 7 4 16 */
-&(nid_objs[313]),/* OBJ_id_regCtrl                   1 3 6 1 5 5 7 5 1 */
-&(nid_objs[314]),/* OBJ_id_regInfo                   1 3 6 1 5 5 7 5 2 */
-&(nid_objs[323]),/* OBJ_id_alg_des40                 1 3 6 1 5 5 7 6 1 */
-&(nid_objs[324]),/* OBJ_id_alg_noSignature           1 3 6 1 5 5 7 6 2 */
-&(nid_objs[325]),/* OBJ_id_alg_dh_sig_hmac_sha1      1 3 6 1 5 5 7 6 3 */
-&(nid_objs[326]),/* OBJ_id_alg_dh_pop                1 3 6 1 5 5 7 6 4 */
-&(nid_objs[327]),/* OBJ_id_cmc_statusInfo            1 3 6 1 5 5 7 7 1 */
-&(nid_objs[328]),/* OBJ_id_cmc_identification        1 3 6 1 5 5 7 7 2 */
-&(nid_objs[329]),/* OBJ_id_cmc_identityProof         1 3 6 1 5 5 7 7 3 */
-&(nid_objs[330]),/* OBJ_id_cmc_dataReturn            1 3 6 1 5 5 7 7 4 */
-&(nid_objs[331]),/* OBJ_id_cmc_transactionId         1 3 6 1 5 5 7 7 5 */
-&(nid_objs[332]),/* OBJ_id_cmc_senderNonce           1 3 6 1 5 5 7 7 6 */
-&(nid_objs[333]),/* OBJ_id_cmc_recipientNonce        1 3 6 1 5 5 7 7 7 */
-&(nid_objs[334]),/* OBJ_id_cmc_addExtensions         1 3 6 1 5 5 7 7 8 */
-&(nid_objs[335]),/* OBJ_id_cmc_encryptedPOP          1 3 6 1 5 5 7 7 9 */
-&(nid_objs[336]),/* OBJ_id_cmc_decryptedPOP          1 3 6 1 5 5 7 7 10 */
-&(nid_objs[337]),/* OBJ_id_cmc_lraPOPWitness         1 3 6 1 5 5 7 7 11 */
-&(nid_objs[338]),/* OBJ_id_cmc_getCert               1 3 6 1 5 5 7 7 15 */
-&(nid_objs[339]),/* OBJ_id_cmc_getCRL                1 3 6 1 5 5 7 7 16 */
-&(nid_objs[340]),/* OBJ_id_cmc_revokeRequest         1 3 6 1 5 5 7 7 17 */
-&(nid_objs[341]),/* OBJ_id_cmc_regInfo               1 3 6 1 5 5 7 7 18 */
-&(nid_objs[342]),/* OBJ_id_cmc_responseInfo          1 3 6 1 5 5 7 7 19 */
-&(nid_objs[343]),/* OBJ_id_cmc_queryPending          1 3 6 1 5 5 7 7 21 */
-&(nid_objs[344]),/* OBJ_id_cmc_popLinkRandom         1 3 6 1 5 5 7 7 22 */
-&(nid_objs[345]),/* OBJ_id_cmc_popLinkWitness        1 3 6 1 5 5 7 7 23 */
-&(nid_objs[346]),/* OBJ_id_cmc_confirmCertAcceptance 1 3 6 1 5 5 7 7 24 */
-&(nid_objs[347]),/* OBJ_id_on_personalData           1 3 6 1 5 5 7 8 1 */
-&(nid_objs[858]),/* OBJ_id_on_permanentIdentifier    1 3 6 1 5 5 7 8 3 */
-&(nid_objs[348]),/* OBJ_id_pda_dateOfBirth           1 3 6 1 5 5 7 9 1 */
-&(nid_objs[349]),/* OBJ_id_pda_placeOfBirth          1 3 6 1 5 5 7 9 2 */
-&(nid_objs[351]),/* OBJ_id_pda_gender                1 3 6 1 5 5 7 9 3 */
-&(nid_objs[352]),/* OBJ_id_pda_countryOfCitizenship  1 3 6 1 5 5 7 9 4 */
-&(nid_objs[353]),/* OBJ_id_pda_countryOfResidence    1 3 6 1 5 5 7 9 5 */
-&(nid_objs[354]),/* OBJ_id_aca_authenticationInfo    1 3 6 1 5 5 7 10 1 */
-&(nid_objs[355]),/* OBJ_id_aca_accessIdentity        1 3 6 1 5 5 7 10 2 */
-&(nid_objs[356]),/* OBJ_id_aca_chargingIdentity      1 3 6 1 5 5 7 10 3 */
-&(nid_objs[357]),/* OBJ_id_aca_group                 1 3 6 1 5 5 7 10 4 */
-&(nid_objs[358]),/* OBJ_id_aca_role                  1 3 6 1 5 5 7 10 5 */
-&(nid_objs[399]),/* OBJ_id_aca_encAttrs              1 3 6 1 5 5 7 10 6 */
-&(nid_objs[359]),/* OBJ_id_qcs_pkixQCSyntax_v1       1 3 6 1 5 5 7 11 1 */
-&(nid_objs[360]),/* OBJ_id_cct_crs                   1 3 6 1 5 5 7 12 1 */
-&(nid_objs[361]),/* OBJ_id_cct_PKIData               1 3 6 1 5 5 7 12 2 */
-&(nid_objs[362]),/* OBJ_id_cct_PKIResponse           1 3 6 1 5 5 7 12 3 */
-&(nid_objs[664]),/* OBJ_id_ppl_anyLanguage           1 3 6 1 5 5 7 21 0 */
-&(nid_objs[665]),/* OBJ_id_ppl_inheritAll            1 3 6 1 5 5 7 21 1 */
-&(nid_objs[667]),/* OBJ_Independent                  1 3 6 1 5 5 7 21 2 */
-&(nid_objs[178]),/* OBJ_ad_OCSP                      1 3 6 1 5 5 7 48 1 */
-&(nid_objs[179]),/* OBJ_ad_ca_issuers                1 3 6 1 5 5 7 48 2 */
-&(nid_objs[363]),/* OBJ_ad_timeStamping              1 3 6 1 5 5 7 48 3 */
-&(nid_objs[364]),/* OBJ_ad_dvcs                      1 3 6 1 5 5 7 48 4 */
-&(nid_objs[785]),/* OBJ_caRepository                 1 3 6 1 5 5 7 48 5 */
-&(nid_objs[780]),/* OBJ_hmac_md5                     1 3 6 1 5 5 8 1 1 */
-&(nid_objs[781]),/* OBJ_hmac_sha1                    1 3 6 1 5 5 8 1 2 */
-&(nid_objs[58]),/* OBJ_netscape_cert_extension      2 16 840 1 113730 1 */
-&(nid_objs[59]),/* OBJ_netscape_data_type           2 16 840 1 113730 2 */
-&(nid_objs[438]),/* OBJ_pilotAttributeType           0 9 2342 19200300 100 1 */
-&(nid_objs[439]),/* OBJ_pilotAttributeSyntax         0 9 2342 19200300 100 3 */
-&(nid_objs[440]),/* OBJ_pilotObjectClass             0 9 2342 19200300 100 4 */
-&(nid_objs[441]),/* OBJ_pilotGroups                  0 9 2342 19200300 100 10 */
-&(nid_objs[108]),/* OBJ_cast5_cbc                    1 2 840 113533 7 66 10 */
-&(nid_objs[112]),/* OBJ_pbeWithMD5AndCast5_CBC       1 2 840 113533 7 66 12 */
-&(nid_objs[782]),/* OBJ_id_PasswordBasedMAC          1 2 840 113533 7 66 13 */
-&(nid_objs[783]),/* OBJ_id_DHBasedMac                1 2 840 113533 7 66 30 */
-&(nid_objs[ 6]),/* OBJ_rsaEncryption                1 2 840 113549 1 1 1 */
-&(nid_objs[ 7]),/* OBJ_md2WithRSAEncryption         1 2 840 113549 1 1 2 */
-&(nid_objs[396]),/* OBJ_md4WithRSAEncryption         1 2 840 113549 1 1 3 */
-&(nid_objs[ 8]),/* OBJ_md5WithRSAEncryption         1 2 840 113549 1 1 4 */
-&(nid_objs[65]),/* OBJ_sha1WithRSAEncryption        1 2 840 113549 1 1 5 */
-&(nid_objs[644]),/* OBJ_rsaOAEPEncryptionSET         1 2 840 113549 1 1 6 */
-&(nid_objs[668]),/* OBJ_sha256WithRSAEncryption      1 2 840 113549 1 1 11 */
-&(nid_objs[669]),/* OBJ_sha384WithRSAEncryption      1 2 840 113549 1 1 12 */
-&(nid_objs[670]),/* OBJ_sha512WithRSAEncryption      1 2 840 113549 1 1 13 */
-&(nid_objs[671]),/* OBJ_sha224WithRSAEncryption      1 2 840 113549 1 1 14 */
-&(nid_objs[28]),/* OBJ_dhKeyAgreement               1 2 840 113549 1 3 1 */
-&(nid_objs[ 9]),/* OBJ_pbeWithMD2AndDES_CBC         1 2 840 113549 1 5 1 */
-&(nid_objs[10]),/* OBJ_pbeWithMD5AndDES_CBC         1 2 840 113549 1 5 3 */
-&(nid_objs[168]),/* OBJ_pbeWithMD2AndRC2_CBC         1 2 840 113549 1 5 4 */
-&(nid_objs[169]),/* OBJ_pbeWithMD5AndRC2_CBC         1 2 840 113549 1 5 6 */
-&(nid_objs[170]),/* OBJ_pbeWithSHA1AndDES_CBC        1 2 840 113549 1 5 10 */
-&(nid_objs[68]),/* OBJ_pbeWithSHA1AndRC2_CBC        1 2 840 113549 1 5 11 */
-&(nid_objs[69]),/* OBJ_id_pbkdf2                    1 2 840 113549 1 5 12 */
-&(nid_objs[161]),/* OBJ_pbes2                        1 2 840 113549 1 5 13 */
-&(nid_objs[162]),/* OBJ_pbmac1                       1 2 840 113549 1 5 14 */
-&(nid_objs[21]),/* OBJ_pkcs7_data                   1 2 840 113549 1 7 1 */
-&(nid_objs[22]),/* OBJ_pkcs7_signed                 1 2 840 113549 1 7 2 */
-&(nid_objs[23]),/* OBJ_pkcs7_enveloped              1 2 840 113549 1 7 3 */
-&(nid_objs[24]),/* OBJ_pkcs7_signedAndEnveloped     1 2 840 113549 1 7 4 */
-&(nid_objs[25]),/* OBJ_pkcs7_digest                 1 2 840 113549 1 7 5 */
-&(nid_objs[26]),/* OBJ_pkcs7_encrypted              1 2 840 113549 1 7 6 */
-&(nid_objs[48]),/* OBJ_pkcs9_emailAddress           1 2 840 113549 1 9 1 */
-&(nid_objs[49]),/* OBJ_pkcs9_unstructuredName       1 2 840 113549 1 9 2 */
-&(nid_objs[50]),/* OBJ_pkcs9_contentType            1 2 840 113549 1 9 3 */
-&(nid_objs[51]),/* OBJ_pkcs9_messageDigest          1 2 840 113549 1 9 4 */
-&(nid_objs[52]),/* OBJ_pkcs9_signingTime            1 2 840 113549 1 9 5 */
-&(nid_objs[53]),/* OBJ_pkcs9_countersignature       1 2 840 113549 1 9 6 */
-&(nid_objs[54]),/* OBJ_pkcs9_challengePassword      1 2 840 113549 1 9 7 */
-&(nid_objs[55]),/* OBJ_pkcs9_unstructuredAddress    1 2 840 113549 1 9 8 */
-&(nid_objs[56]),/* OBJ_pkcs9_extCertAttributes      1 2 840 113549 1 9 9 */
-&(nid_objs[172]),/* OBJ_ext_req                      1 2 840 113549 1 9 14 */
-&(nid_objs[167]),/* OBJ_SMIMECapabilities            1 2 840 113549 1 9 15 */
-&(nid_objs[188]),/* OBJ_SMIME                        1 2 840 113549 1 9 16 */
-&(nid_objs[156]),/* OBJ_friendlyName                 1 2 840 113549 1 9 20 */
-&(nid_objs[157]),/* OBJ_localKeyID                   1 2 840 113549 1 9 21 */
-&(nid_objs[681]),/* OBJ_X9_62_onBasis                1 2 840 10045 1 2 3 1 */
-&(nid_objs[682]),/* OBJ_X9_62_tpBasis                1 2 840 10045 1 2 3 2 */
-&(nid_objs[683]),/* OBJ_X9_62_ppBasis                1 2 840 10045 1 2 3 3 */
-&(nid_objs[417]),/* OBJ_ms_csp_name                  1 3 6 1 4 1 311 17 1 */
-&(nid_objs[856]),/* OBJ_LocalKeySet                  1 3 6 1 4 1 311 17 2 */
-&(nid_objs[390]),/* OBJ_dcObject                     1 3 6 1 4 1 1466 344 */
-&(nid_objs[91]),/* OBJ_bf_cbc                       1 3 6 1 4 1 3029 1 2 */
-&(nid_objs[315]),/* OBJ_id_regCtrl_regToken          1 3 6 1 5 5 7 5 1 1 */
-&(nid_objs[316]),/* OBJ_id_regCtrl_authenticator     1 3 6 1 5 5 7 5 1 2 */
-&(nid_objs[317]),/* OBJ_id_regCtrl_pkiPublicationInfo 1 3 6 1 5 5 7 5 1 3 */
-&(nid_objs[318]),/* OBJ_id_regCtrl_pkiArchiveOptions 1 3 6 1 5 5 7 5 1 4 */
-&(nid_objs[319]),/* OBJ_id_regCtrl_oldCertID         1 3 6 1 5 5 7 5 1 5 */
-&(nid_objs[320]),/* OBJ_id_regCtrl_protocolEncrKey   1 3 6 1 5 5 7 5 1 6 */
-&(nid_objs[321]),/* OBJ_id_regInfo_utf8Pairs         1 3 6 1 5 5 7 5 2 1 */
-&(nid_objs[322]),/* OBJ_id_regInfo_certReq           1 3 6 1 5 5 7 5 2 2 */
-&(nid_objs[365]),/* OBJ_id_pkix_OCSP_basic           1 3 6 1 5 5 7 48 1 1 */
-&(nid_objs[366]),/* OBJ_id_pkix_OCSP_Nonce           1 3 6 1 5 5 7 48 1 2 */
-&(nid_objs[367]),/* OBJ_id_pkix_OCSP_CrlID           1 3 6 1 5 5 7 48 1 3 */
-&(nid_objs[368]),/* OBJ_id_pkix_OCSP_acceptableResponses 1 3 6 1 5 5 7 48 1 4 */
-&(nid_objs[369]),/* OBJ_id_pkix_OCSP_noCheck         1 3 6 1 5 5 7 48 1 5 */
-&(nid_objs[370]),/* OBJ_id_pkix_OCSP_archiveCutoff   1 3 6 1 5 5 7 48 1 6 */
-&(nid_objs[371]),/* OBJ_id_pkix_OCSP_serviceLocator  1 3 6 1 5 5 7 48 1 7 */
-&(nid_objs[372]),/* OBJ_id_pkix_OCSP_extendedStatus  1 3 6 1 5 5 7 48 1 8 */
-&(nid_objs[373]),/* OBJ_id_pkix_OCSP_valid           1 3 6 1 5 5 7 48 1 9 */
-&(nid_objs[374]),/* OBJ_id_pkix_OCSP_path            1 3 6 1 5 5 7 48 1 10 */
-&(nid_objs[375]),/* OBJ_id_pkix_OCSP_trustRoot       1 3 6 1 5 5 7 48 1 11 */
-&(nid_objs[418]),/* OBJ_aes_128_ecb                  2 16 840 1 101 3 4 1 1 */
-&(nid_objs[419]),/* OBJ_aes_128_cbc                  2 16 840 1 101 3 4 1 2 */
-&(nid_objs[420]),/* OBJ_aes_128_ofb128               2 16 840 1 101 3 4 1 3 */
-&(nid_objs[421]),/* OBJ_aes_128_cfb128               2 16 840 1 101 3 4 1 4 */
-&(nid_objs[788]),/* OBJ_id_aes128_wrap               2 16 840 1 101 3 4 1 5 */
-&(nid_objs[422]),/* OBJ_aes_192_ecb                  2 16 840 1 101 3 4 1 21 */
-&(nid_objs[423]),/* OBJ_aes_192_cbc                  2 16 840 1 101 3 4 1 22 */
-&(nid_objs[424]),/* OBJ_aes_192_ofb128               2 16 840 1 101 3 4 1 23 */
-&(nid_objs[425]),/* OBJ_aes_192_cfb128               2 16 840 1 101 3 4 1 24 */
-&(nid_objs[789]),/* OBJ_id_aes192_wrap               2 16 840 1 101 3 4 1 25 */
-&(nid_objs[426]),/* OBJ_aes_256_ecb                  2 16 840 1 101 3 4 1 41 */
-&(nid_objs[427]),/* OBJ_aes_256_cbc                  2 16 840 1 101 3 4 1 42 */
-&(nid_objs[428]),/* OBJ_aes_256_ofb128               2 16 840 1 101 3 4 1 43 */
-&(nid_objs[429]),/* OBJ_aes_256_cfb128               2 16 840 1 101 3 4 1 44 */
-&(nid_objs[790]),/* OBJ_id_aes256_wrap               2 16 840 1 101 3 4 1 45 */
-&(nid_objs[672]),/* OBJ_sha256                       2 16 840 1 101 3 4 2 1 */
-&(nid_objs[673]),/* OBJ_sha384                       2 16 840 1 101 3 4 2 2 */
-&(nid_objs[674]),/* OBJ_sha512                       2 16 840 1 101 3 4 2 3 */
-&(nid_objs[675]),/* OBJ_sha224                       2 16 840 1 101 3 4 2 4 */
-&(nid_objs[802]),/* OBJ_dsa_with_SHA224              2 16 840 1 101 3 4 3 1 */
-&(nid_objs[803]),/* OBJ_dsa_with_SHA256              2 16 840 1 101 3 4 3 2 */
-&(nid_objs[71]),/* OBJ_netscape_cert_type           2 16 840 1 113730 1 1 */
-&(nid_objs[72]),/* OBJ_netscape_base_url            2 16 840 1 113730 1 2 */
-&(nid_objs[73]),/* OBJ_netscape_revocation_url      2 16 840 1 113730 1 3 */
-&(nid_objs[74]),/* OBJ_netscape_ca_revocation_url   2 16 840 1 113730 1 4 */
-&(nid_objs[75]),/* OBJ_netscape_renewal_url         2 16 840 1 113730 1 7 */
-&(nid_objs[76]),/* OBJ_netscape_ca_policy_url       2 16 840 1 113730 1 8 */
-&(nid_objs[77]),/* OBJ_netscape_ssl_server_name     2 16 840 1 113730 1 12 */
-&(nid_objs[78]),/* OBJ_netscape_comment             2 16 840 1 113730 1 13 */
-&(nid_objs[79]),/* OBJ_netscape_cert_sequence       2 16 840 1 113730 2 5 */
-&(nid_objs[139]),/* OBJ_ns_sgc                       2 16 840 1 113730 4 1 */
-&(nid_objs[458]),/* OBJ_userId                       0 9 2342 19200300 100 1 1 */
-&(nid_objs[459]),/* OBJ_textEncodedORAddress         0 9 2342 19200300 100 1 2 */
-&(nid_objs[460]),/* OBJ_rfc822Mailbox                0 9 2342 19200300 100 1 3 */
-&(nid_objs[461]),/* OBJ_info                         0 9 2342 19200300 100 1 4 */
-&(nid_objs[462]),/* OBJ_favouriteDrink               0 9 2342 19200300 100 1 5 */
-&(nid_objs[463]),/* OBJ_roomNumber                   0 9 2342 19200300 100 1 6 */
-&(nid_objs[464]),/* OBJ_photo                        0 9 2342 19200300 100 1 7 */
-&(nid_objs[465]),/* OBJ_userClass                    0 9 2342 19200300 100 1 8 */
-&(nid_objs[466]),/* OBJ_host                         0 9 2342 19200300 100 1 9 */
-&(nid_objs[467]),/* OBJ_manager                      0 9 2342 19200300 100 1 10 */
-&(nid_objs[468]),/* OBJ_documentIdentifier           0 9 2342 19200300 100 1 11 */
-&(nid_objs[469]),/* OBJ_documentTitle                0 9 2342 19200300 100 1 12 */
-&(nid_objs[470]),/* OBJ_documentVersion              0 9 2342 19200300 100 1 13 */
-&(nid_objs[471]),/* OBJ_documentAuthor               0 9 2342 19200300 100 1 14 */
-&(nid_objs[472]),/* OBJ_documentLocation             0 9 2342 19200300 100 1 15 */
-&(nid_objs[473]),/* OBJ_homeTelephoneNumber          0 9 2342 19200300 100 1 20 */
-&(nid_objs[474]),/* OBJ_secretary                    0 9 2342 19200300 100 1 21 */
-&(nid_objs[475]),/* OBJ_otherMailbox                 0 9 2342 19200300 100 1 22 */
-&(nid_objs[476]),/* OBJ_lastModifiedTime             0 9 2342 19200300 100 1 23 */
-&(nid_objs[477]),/* OBJ_lastModifiedBy               0 9 2342 19200300 100 1 24 */
-&(nid_objs[391]),/* OBJ_domainComponent              0 9 2342 19200300 100 1 25 */
-&(nid_objs[478]),/* OBJ_aRecord                      0 9 2342 19200300 100 1 26 */
-&(nid_objs[479]),/* OBJ_pilotAttributeType27         0 9 2342 19200300 100 1 27 */
-&(nid_objs[480]),/* OBJ_mXRecord                     0 9 2342 19200300 100 1 28 */
-&(nid_objs[481]),/* OBJ_nSRecord                     0 9 2342 19200300 100 1 29 */
-&(nid_objs[482]),/* OBJ_sOARecord                    0 9 2342 19200300 100 1 30 */
-&(nid_objs[483]),/* OBJ_cNAMERecord                  0 9 2342 19200300 100 1 31 */
-&(nid_objs[484]),/* OBJ_associatedDomain             0 9 2342 19200300 100 1 37 */
-&(nid_objs[485]),/* OBJ_associatedName               0 9 2342 19200300 100 1 38 */
-&(nid_objs[486]),/* OBJ_homePostalAddress            0 9 2342 19200300 100 1 39 */
-&(nid_objs[487]),/* OBJ_personalTitle                0 9 2342 19200300 100 1 40 */
-&(nid_objs[488]),/* OBJ_mobileTelephoneNumber        0 9 2342 19200300 100 1 41 */
-&(nid_objs[489]),/* OBJ_pagerTelephoneNumber         0 9 2342 19200300 100 1 42 */
-&(nid_objs[490]),/* OBJ_friendlyCountryName          0 9 2342 19200300 100 1 43 */
-&(nid_objs[491]),/* OBJ_organizationalStatus         0 9 2342 19200300 100 1 45 */
-&(nid_objs[492]),/* OBJ_janetMailbox                 0 9 2342 19200300 100 1 46 */
-&(nid_objs[493]),/* OBJ_mailPreferenceOption         0 9 2342 19200300 100 1 47 */
-&(nid_objs[494]),/* OBJ_buildingName                 0 9 2342 19200300 100 1 48 */
-&(nid_objs[495]),/* OBJ_dSAQuality                   0 9 2342 19200300 100 1 49 */
-&(nid_objs[496]),/* OBJ_singleLevelQuality           0 9 2342 19200300 100 1 50 */
-&(nid_objs[497]),/* OBJ_subtreeMinimumQuality        0 9 2342 19200300 100 1 51 */
-&(nid_objs[498]),/* OBJ_subtreeMaximumQuality        0 9 2342 19200300 100 1 52 */
-&(nid_objs[499]),/* OBJ_personalSignature            0 9 2342 19200300 100 1 53 */
-&(nid_objs[500]),/* OBJ_dITRedirect                  0 9 2342 19200300 100 1 54 */
-&(nid_objs[501]),/* OBJ_audio                        0 9 2342 19200300 100 1 55 */
-&(nid_objs[502]),/* OBJ_documentPublisher            0 9 2342 19200300 100 1 56 */
-&(nid_objs[442]),/* OBJ_iA5StringSyntax              0 9 2342 19200300 100 3 4 */
-&(nid_objs[443]),/* OBJ_caseIgnoreIA5StringSyntax    0 9 2342 19200300 100 3 5 */
-&(nid_objs[444]),/* OBJ_pilotObject                  0 9 2342 19200300 100 4 3 */
-&(nid_objs[445]),/* OBJ_pilotPerson                  0 9 2342 19200300 100 4 4 */
-&(nid_objs[446]),/* OBJ_account                      0 9 2342 19200300 100 4 5 */
-&(nid_objs[447]),/* OBJ_document                     0 9 2342 19200300 100 4 6 */
-&(nid_objs[448]),/* OBJ_room                         0 9 2342 19200300 100 4 7 */
-&(nid_objs[449]),/* OBJ_documentSeries               0 9 2342 19200300 100 4 9 */
-&(nid_objs[392]),/* OBJ_Domain                       0 9 2342 19200300 100 4 13 */
-&(nid_objs[450]),/* OBJ_rFC822localPart              0 9 2342 19200300 100 4 14 */
-&(nid_objs[451]),/* OBJ_dNSDomain                    0 9 2342 19200300 100 4 15 */
-&(nid_objs[452]),/* OBJ_domainRelatedObject          0 9 2342 19200300 100 4 17 */
-&(nid_objs[453]),/* OBJ_friendlyCountry              0 9 2342 19200300 100 4 18 */
-&(nid_objs[454]),/* OBJ_simpleSecurityObject         0 9 2342 19200300 100 4 19 */
-&(nid_objs[455]),/* OBJ_pilotOrganization            0 9 2342 19200300 100 4 20 */
-&(nid_objs[456]),/* OBJ_pilotDSA                     0 9 2342 19200300 100 4 21 */
-&(nid_objs[457]),/* OBJ_qualityLabelledData          0 9 2342 19200300 100 4 22 */
-&(nid_objs[189]),/* OBJ_id_smime_mod                 1 2 840 113549 1 9 16 0 */
-&(nid_objs[190]),/* OBJ_id_smime_ct                  1 2 840 113549 1 9 16 1 */
-&(nid_objs[191]),/* OBJ_id_smime_aa                  1 2 840 113549 1 9 16 2 */
-&(nid_objs[192]),/* OBJ_id_smime_alg                 1 2 840 113549 1 9 16 3 */
-&(nid_objs[193]),/* OBJ_id_smime_cd                  1 2 840 113549 1 9 16 4 */
-&(nid_objs[194]),/* OBJ_id_smime_spq                 1 2 840 113549 1 9 16 5 */
-&(nid_objs[195]),/* OBJ_id_smime_cti                 1 2 840 113549 1 9 16 6 */
-&(nid_objs[158]),/* OBJ_x509Certificate              1 2 840 113549 1 9 22 1 */
-&(nid_objs[159]),/* OBJ_sdsiCertificate              1 2 840 113549 1 9 22 2 */
-&(nid_objs[160]),/* OBJ_x509Crl                      1 2 840 113549 1 9 23 1 */
-&(nid_objs[144]),/* OBJ_pbe_WithSHA1And128BitRC4     1 2 840 113549 1 12 1 1 */
-&(nid_objs[145]),/* OBJ_pbe_WithSHA1And40BitRC4      1 2 840 113549 1 12 1 2 */
-&(nid_objs[146]),/* OBJ_pbe_WithSHA1And3_Key_TripleDES_CBC 1 2 840 113549 1 12 1 3 */
-&(nid_objs[147]),/* OBJ_pbe_WithSHA1And2_Key_TripleDES_CBC 1 2 840 113549 1 12 1 4 */
-&(nid_objs[148]),/* OBJ_pbe_WithSHA1And128BitRC2_CBC 1 2 840 113549 1 12 1 5 */
-&(nid_objs[149]),/* OBJ_pbe_WithSHA1And40BitRC2_CBC  1 2 840 113549 1 12 1 6 */
-&(nid_objs[171]),/* OBJ_ms_ext_req                   1 3 6 1 4 1 311 2 1 14 */
-&(nid_objs[134]),/* OBJ_ms_code_ind                  1 3 6 1 4 1 311 2 1 21 */
-&(nid_objs[135]),/* OBJ_ms_code_com                  1 3 6 1 4 1 311 2 1 22 */
-&(nid_objs[136]),/* OBJ_ms_ctl_sign                  1 3 6 1 4 1 311 10 3 1 */
-&(nid_objs[137]),/* OBJ_ms_sgc                       1 3 6 1 4 1 311 10 3 3 */
-&(nid_objs[138]),/* OBJ_ms_efs                       1 3 6 1 4 1 311 10 3 4 */
-&(nid_objs[648]),/* OBJ_ms_smartcard_login           1 3 6 1 4 1 311 20 2 2 */
-&(nid_objs[649]),/* OBJ_ms_upn                       1 3 6 1 4 1 311 20 2 3 */
-&(nid_objs[751]),/* OBJ_camellia_128_cbc             1 2 392 200011 61 1 1 1 2 */
-&(nid_objs[752]),/* OBJ_camellia_192_cbc             1 2 392 200011 61 1 1 1 3 */
-&(nid_objs[753]),/* OBJ_camellia_256_cbc             1 2 392 200011 61 1 1 1 4 */
-&(nid_objs[196]),/* OBJ_id_smime_mod_cms             1 2 840 113549 1 9 16 0 1 */
-&(nid_objs[197]),/* OBJ_id_smime_mod_ess             1 2 840 113549 1 9 16 0 2 */
-&(nid_objs[198]),/* OBJ_id_smime_mod_oid             1 2 840 113549 1 9 16 0 3 */
-&(nid_objs[199]),/* OBJ_id_smime_mod_msg_v3          1 2 840 113549 1 9 16 0 4 */
-&(nid_objs[200]),/* OBJ_id_smime_mod_ets_eSignature_88 1 2 840 113549 1 9 16 0 5 */
-&(nid_objs[201]),/* OBJ_id_smime_mod_ets_eSignature_97 1 2 840 113549 1 9 16 0 6 */
-&(nid_objs[202]),/* OBJ_id_smime_mod_ets_eSigPolicy_88 1 2 840 113549 1 9 16 0 7 */
-&(nid_objs[203]),/* OBJ_id_smime_mod_ets_eSigPolicy_97 1 2 840 113549 1 9 16 0 8 */
-&(nid_objs[204]),/* OBJ_id_smime_ct_receipt          1 2 840 113549 1 9 16 1 1 */
-&(nid_objs[205]),/* OBJ_id_smime_ct_authData         1 2 840 113549 1 9 16 1 2 */
-&(nid_objs[206]),/* OBJ_id_smime_ct_publishCert      1 2 840 113549 1 9 16 1 3 */
-&(nid_objs[207]),/* OBJ_id_smime_ct_TSTInfo          1 2 840 113549 1 9 16 1 4 */
-&(nid_objs[208]),/* OBJ_id_smime_ct_TDTInfo          1 2 840 113549 1 9 16 1 5 */
-&(nid_objs[209]),/* OBJ_id_smime_ct_contentInfo      1 2 840 113549 1 9 16 1 6 */
-&(nid_objs[210]),/* OBJ_id_smime_ct_DVCSRequestData  1 2 840 113549 1 9 16 1 7 */
-&(nid_objs[211]),/* OBJ_id_smime_ct_DVCSResponseData 1 2 840 113549 1 9 16 1 8 */
-&(nid_objs[786]),/* OBJ_id_smime_ct_compressedData   1 2 840 113549 1 9 16 1 9 */
-&(nid_objs[787]),/* OBJ_id_ct_asciiTextWithCRLF      1 2 840 113549 1 9 16 1 27 */
-&(nid_objs[212]),/* OBJ_id_smime_aa_receiptRequest   1 2 840 113549 1 9 16 2 1 */
-&(nid_objs[213]),/* OBJ_id_smime_aa_securityLabel    1 2 840 113549 1 9 16 2 2 */
-&(nid_objs[214]),/* OBJ_id_smime_aa_mlExpandHistory  1 2 840 113549 1 9 16 2 3 */
-&(nid_objs[215]),/* OBJ_id_smime_aa_contentHint      1 2 840 113549 1 9 16 2 4 */
-&(nid_objs[216]),/* OBJ_id_smime_aa_msgSigDigest     1 2 840 113549 1 9 16 2 5 */
-&(nid_objs[217]),/* OBJ_id_smime_aa_encapContentType 1 2 840 113549 1 9 16 2 6 */
-&(nid_objs[218]),/* OBJ_id_smime_aa_contentIdentifier 1 2 840 113549 1 9 16 2 7 */
-&(nid_objs[219]),/* OBJ_id_smime_aa_macValue         1 2 840 113549 1 9 16 2 8 */
-&(nid_objs[220]),/* OBJ_id_smime_aa_equivalentLabels 1 2 840 113549 1 9 16 2 9 */
-&(nid_objs[221]),/* OBJ_id_smime_aa_contentReference 1 2 840 113549 1 9 16 2 10 */
-&(nid_objs[222]),/* OBJ_id_smime_aa_encrypKeyPref    1 2 840 113549 1 9 16 2 11 */
-&(nid_objs[223]),/* OBJ_id_smime_aa_signingCertificate 1 2 840 113549 1 9 16 2 12 */
-&(nid_objs[224]),/* OBJ_id_smime_aa_smimeEncryptCerts 1 2 840 113549 1 9 16 2 13 */
-&(nid_objs[225]),/* OBJ_id_smime_aa_timeStampToken   1 2 840 113549 1 9 16 2 14 */
-&(nid_objs[226]),/* OBJ_id_smime_aa_ets_sigPolicyId  1 2 840 113549 1 9 16 2 15 */
-&(nid_objs[227]),/* OBJ_id_smime_aa_ets_commitmentType 1 2 840 113549 1 9 16 2 16 */
-&(nid_objs[228]),/* OBJ_id_smime_aa_ets_signerLocation 1 2 840 113549 1 9 16 2 17 */
-&(nid_objs[229]),/* OBJ_id_smime_aa_ets_signerAttr   1 2 840 113549 1 9 16 2 18 */
-&(nid_objs[230]),/* OBJ_id_smime_aa_ets_otherSigCert 1 2 840 113549 1 9 16 2 19 */
-&(nid_objs[231]),/* OBJ_id_smime_aa_ets_contentTimestamp 1 2 840 113549 1 9 16 2 20 */
-&(nid_objs[232]),/* OBJ_id_smime_aa_ets_CertificateRefs 1 2 840 113549 1 9 16 2 21 */
-&(nid_objs[233]),/* OBJ_id_smime_aa_ets_RevocationRefs 1 2 840 113549 1 9 16 2 22 */
-&(nid_objs[234]),/* OBJ_id_smime_aa_ets_certValues   1 2 840 113549 1 9 16 2 23 */
-&(nid_objs[235]),/* OBJ_id_smime_aa_ets_revocationValues 1 2 840 113549 1 9 16 2 24 */
-&(nid_objs[236]),/* OBJ_id_smime_aa_ets_escTimeStamp 1 2 840 113549 1 9 16 2 25 */
-&(nid_objs[237]),/* OBJ_id_smime_aa_ets_certCRLTimestamp 1 2 840 113549 1 9 16 2 26 */
-&(nid_objs[238]),/* OBJ_id_smime_aa_ets_archiveTimeStamp 1 2 840 113549 1 9 16 2 27 */
-&(nid_objs[239]),/* OBJ_id_smime_aa_signatureType    1 2 840 113549 1 9 16 2 28 */
-&(nid_objs[240]),/* OBJ_id_smime_aa_dvcs_dvc         1 2 840 113549 1 9 16 2 29 */
-&(nid_objs[241]),/* OBJ_id_smime_alg_ESDHwith3DES    1 2 840 113549 1 9 16 3 1 */
-&(nid_objs[242]),/* OBJ_id_smime_alg_ESDHwithRC2     1 2 840 113549 1 9 16 3 2 */
-&(nid_objs[243]),/* OBJ_id_smime_alg_3DESwrap        1 2 840 113549 1 9 16 3 3 */
-&(nid_objs[244]),/* OBJ_id_smime_alg_RC2wrap         1 2 840 113549 1 9 16 3 4 */
-&(nid_objs[245]),/* OBJ_id_smime_alg_ESDH            1 2 840 113549 1 9 16 3 5 */
-&(nid_objs[246]),/* OBJ_id_smime_alg_CMS3DESwrap     1 2 840 113549 1 9 16 3 6 */
-&(nid_objs[247]),/* OBJ_id_smime_alg_CMSRC2wrap      1 2 840 113549 1 9 16 3 7 */
-&(nid_objs[125]),/* OBJ_zlib_compression             1 2 840 113549 1 9 16 3 8 */
-&(nid_objs[248]),/* OBJ_id_smime_cd_ldap             1 2 840 113549 1 9 16 4 1 */
-&(nid_objs[249]),/* OBJ_id_smime_spq_ets_sqt_uri     1 2 840 113549 1 9 16 5 1 */
-&(nid_objs[250]),/* OBJ_id_smime_spq_ets_sqt_unotice 1 2 840 113549 1 9 16 5 2 */
-&(nid_objs[251]),/* OBJ_id_smime_cti_ets_proofOfOrigin 1 2 840 113549 1 9 16 6 1 */
-&(nid_objs[252]),/* OBJ_id_smime_cti_ets_proofOfReceipt 1 2 840 113549 1 9 16 6 2 */
-&(nid_objs[253]),/* OBJ_id_smime_cti_ets_proofOfDelivery 1 2 840 113549 1 9 16 6 3 */
-&(nid_objs[254]),/* OBJ_id_smime_cti_ets_proofOfSender 1 2 840 113549 1 9 16 6 4 */
-&(nid_objs[255]),/* OBJ_id_smime_cti_ets_proofOfApproval 1 2 840 113549 1 9 16 6 5 */
-&(nid_objs[256]),/* OBJ_id_smime_cti_ets_proofOfCreation 1 2 840 113549 1 9 16 6 6 */
-&(nid_objs[150]),/* OBJ_keyBag                       1 2 840 113549 1 12 10 1 1 */
-&(nid_objs[151]),/* OBJ_pkcs8ShroudedKeyBag          1 2 840 113549 1 12 10 1 2 */
-&(nid_objs[152]),/* OBJ_certBag                      1 2 840 113549 1 12 10 1 3 */
-&(nid_objs[153]),/* OBJ_crlBag                       1 2 840 113549 1 12 10 1 4 */
-&(nid_objs[154]),/* OBJ_secretBag                    1 2 840 113549 1 12 10 1 5 */
-&(nid_objs[155]),/* OBJ_safeContentsBag              1 2 840 113549 1 12 10 1 6 */
-&(nid_objs[34]),/* OBJ_idea_cbc                     1 3 6 1 4 1 188 7 1 1 2 */
+static const unsigned int obj_objs[NUM_OBJ]={
+ 0,	/* OBJ_undef                        0 */
+393,	/* OBJ_joint_iso_ccitt              OBJ_joint_iso_itu_t */
+404,	/* OBJ_ccitt                        OBJ_itu_t */
+645,	/* OBJ_itu_t                        0 */
+434,	/* OBJ_data                         0 9 */
+181,	/* OBJ_iso                          1 */
+182,	/* OBJ_member_body                  1 2 */
+379,	/* OBJ_org                          1 3 */
+676,	/* OBJ_identified_organization      1 3 */
+646,	/* OBJ_joint_iso_itu_t              2 */
+11,	/* OBJ_X500                         2 5 */
+647,	/* OBJ_international_organizations  2 23 */
+380,	/* OBJ_dod                          1 3 6 */
+12,	/* OBJ_X509                         2 5 4 */
+378,	/* OBJ_X500algorithms               2 5 8 */
+81,	/* OBJ_id_ce                        2 5 29 */
+512,	/* OBJ_id_set                       2 23 42 */
+678,	/* OBJ_wap                          2 23 43 */
+435,	/* OBJ_pss                          0 9 2342 */
+183,	/* OBJ_ISO_US                       1 2 840 */
+381,	/* OBJ_iana                         1 3 6 1 */
+677,	/* OBJ_certicom_arc                 1 3 132 */
+394,	/* OBJ_selected_attribute_types     2 5 1 5 */
+13,	/* OBJ_commonName                   2 5 4 3 */
+100,	/* OBJ_surname                      2 5 4 4 */
+105,	/* OBJ_serialNumber                 2 5 4 5 */
+14,	/* OBJ_countryName                  2 5 4 6 */
+15,	/* OBJ_localityName                 2 5 4 7 */
+16,	/* OBJ_stateOrProvinceName          2 5 4 8 */
+660,	/* OBJ_streetAddress                2 5 4 9 */
+17,	/* OBJ_organizationName             2 5 4 10 */
+18,	/* OBJ_organizationalUnitName       2 5 4 11 */
+106,	/* OBJ_title                        2 5 4 12 */
+107,	/* OBJ_description                  2 5 4 13 */
+859,	/* OBJ_searchGuide                  2 5 4 14 */
+860,	/* OBJ_businessCategory             2 5 4 15 */
+861,	/* OBJ_postalAddress                2 5 4 16 */
+661,	/* OBJ_postalCode                   2 5 4 17 */
+862,	/* OBJ_postOfficeBox                2 5 4 18 */
+863,	/* OBJ_physicalDeliveryOfficeName   2 5 4 19 */
+864,	/* OBJ_telephoneNumber              2 5 4 20 */
+865,	/* OBJ_telexNumber                  2 5 4 21 */
+866,	/* OBJ_teletexTerminalIdentifier    2 5 4 22 */
+867,	/* OBJ_facsimileTelephoneNumber     2 5 4 23 */
+868,	/* OBJ_x121Address                  2 5 4 24 */
+869,	/* OBJ_internationaliSDNNumber      2 5 4 25 */
+870,	/* OBJ_registeredAddress            2 5 4 26 */
+871,	/* OBJ_destinationIndicator         2 5 4 27 */
+872,	/* OBJ_preferredDeliveryMethod      2 5 4 28 */
+873,	/* OBJ_presentationAddress          2 5 4 29 */
+874,	/* OBJ_supportedApplicationContext  2 5 4 30 */
+875,	/* OBJ_member                       2 5 4 31 */
+876,	/* OBJ_owner                        2 5 4 32 */
+877,	/* OBJ_roleOccupant                 2 5 4 33 */
+878,	/* OBJ_seeAlso                      2 5 4 34 */
+879,	/* OBJ_userPassword                 2 5 4 35 */
+880,	/* OBJ_userCertificate              2 5 4 36 */
+881,	/* OBJ_cACertificate                2 5 4 37 */
+882,	/* OBJ_authorityRevocationList      2 5 4 38 */
+883,	/* OBJ_certificateRevocationList    2 5 4 39 */
+884,	/* OBJ_crossCertificatePair         2 5 4 40 */
+173,	/* OBJ_name                         2 5 4 41 */
+99,	/* OBJ_givenName                    2 5 4 42 */
+101,	/* OBJ_initials                     2 5 4 43 */
+509,	/* OBJ_generationQualifier          2 5 4 44 */
+503,	/* OBJ_x500UniqueIdentifier         2 5 4 45 */
+174,	/* OBJ_dnQualifier                  2 5 4 46 */
+885,	/* OBJ_enhancedSearchGuide          2 5 4 47 */
+886,	/* OBJ_protocolInformation          2 5 4 48 */
+887,	/* OBJ_distinguishedName            2 5 4 49 */
+888,	/* OBJ_uniqueMember                 2 5 4 50 */
+889,	/* OBJ_houseIdentifier              2 5 4 51 */
+890,	/* OBJ_supportedAlgorithms          2 5 4 52 */
+891,	/* OBJ_deltaRevocationList          2 5 4 53 */
+892,	/* OBJ_dmdName                      2 5 4 54 */
+510,	/* OBJ_pseudonym                    2 5 4 65 */
+400,	/* OBJ_role                         2 5 4 72 */
+769,	/* OBJ_subject_directory_attributes 2 5 29 9 */
+82,	/* OBJ_subject_key_identifier       2 5 29 14 */
+83,	/* OBJ_key_usage                    2 5 29 15 */
+84,	/* OBJ_private_key_usage_period     2 5 29 16 */
+85,	/* OBJ_subject_alt_name             2 5 29 17 */
+86,	/* OBJ_issuer_alt_name              2 5 29 18 */
+87,	/* OBJ_basic_constraints            2 5 29 19 */
+88,	/* OBJ_crl_number                   2 5 29 20 */
+141,	/* OBJ_crl_reason                   2 5 29 21 */
+430,	/* OBJ_hold_instruction_code        2 5 29 23 */
+142,	/* OBJ_invalidity_date              2 5 29 24 */
+140,	/* OBJ_delta_crl                    2 5 29 27 */
+770,	/* OBJ_issuing_distribution_point   2 5 29 28 */
+771,	/* OBJ_certificate_issuer           2 5 29 29 */
+666,	/* OBJ_name_constraints             2 5 29 30 */
+103,	/* OBJ_crl_distribution_points      2 5 29 31 */
+89,	/* OBJ_certificate_policies         2 5 29 32 */
+747,	/* OBJ_policy_mappings              2 5 29 33 */
+90,	/* OBJ_authority_key_identifier     2 5 29 35 */
+401,	/* OBJ_policy_constraints           2 5 29 36 */
+126,	/* OBJ_ext_key_usage                2 5 29 37 */
+857,	/* OBJ_freshest_crl                 2 5 29 46 */
+748,	/* OBJ_inhibit_any_policy           2 5 29 54 */
+402,	/* OBJ_target_information           2 5 29 55 */
+403,	/* OBJ_no_rev_avail                 2 5 29 56 */
+513,	/* OBJ_set_ctype                    2 23 42 0 */
+514,	/* OBJ_set_msgExt                   2 23 42 1 */
+515,	/* OBJ_set_attr                     2 23 42 3 */
+516,	/* OBJ_set_policy                   2 23 42 5 */
+517,	/* OBJ_set_certExt                  2 23 42 7 */
+518,	/* OBJ_set_brand                    2 23 42 8 */
+679,	/* OBJ_wap_wsg                      2 23 43 1 */
+382,	/* OBJ_Directory                    1 3 6 1 1 */
+383,	/* OBJ_Management                   1 3 6 1 2 */
+384,	/* OBJ_Experimental                 1 3 6 1 3 */
+385,	/* OBJ_Private                      1 3 6 1 4 */
+386,	/* OBJ_Security                     1 3 6 1 5 */
+387,	/* OBJ_SNMPv2                       1 3 6 1 6 */
+388,	/* OBJ_Mail                         1 3 6 1 7 */
+376,	/* OBJ_algorithm                    1 3 14 3 2 */
+395,	/* OBJ_clearance                    2 5 1 5 55 */
+19,	/* OBJ_rsa                          2 5 8 1 1 */
+96,	/* OBJ_mdc2WithRSA                  2 5 8 3 100 */
+95,	/* OBJ_mdc2                         2 5 8 3 101 */
+746,	/* OBJ_any_policy                   2 5 29 32 0 */
+910,	/* OBJ_anyExtendedKeyUsage          2 5 29 37 0 */
+519,	/* OBJ_setct_PANData                2 23 42 0 0 */
+520,	/* OBJ_setct_PANToken               2 23 42 0 1 */
+521,	/* OBJ_setct_PANOnly                2 23 42 0 2 */
+522,	/* OBJ_setct_OIData                 2 23 42 0 3 */
+523,	/* OBJ_setct_PI                     2 23 42 0 4 */
+524,	/* OBJ_setct_PIData                 2 23 42 0 5 */
+525,	/* OBJ_setct_PIDataUnsigned         2 23 42 0 6 */
+526,	/* OBJ_setct_HODInput               2 23 42 0 7 */
+527,	/* OBJ_setct_AuthResBaggage         2 23 42 0 8 */
+528,	/* OBJ_setct_AuthRevReqBaggage      2 23 42 0 9 */
+529,	/* OBJ_setct_AuthRevResBaggage      2 23 42 0 10 */
+530,	/* OBJ_setct_CapTokenSeq            2 23 42 0 11 */
+531,	/* OBJ_setct_PInitResData           2 23 42 0 12 */
+532,	/* OBJ_setct_PI_TBS                 2 23 42 0 13 */
+533,	/* OBJ_setct_PResData               2 23 42 0 14 */
+534,	/* OBJ_setct_AuthReqTBS             2 23 42 0 16 */
+535,	/* OBJ_setct_AuthResTBS             2 23 42 0 17 */
+536,	/* OBJ_setct_AuthResTBSX            2 23 42 0 18 */
+537,	/* OBJ_setct_AuthTokenTBS           2 23 42 0 19 */
+538,	/* OBJ_setct_CapTokenData           2 23 42 0 20 */
+539,	/* OBJ_setct_CapTokenTBS            2 23 42 0 21 */
+540,	/* OBJ_setct_AcqCardCodeMsg         2 23 42 0 22 */
+541,	/* OBJ_setct_AuthRevReqTBS          2 23 42 0 23 */
+542,	/* OBJ_setct_AuthRevResData         2 23 42 0 24 */
+543,	/* OBJ_setct_AuthRevResTBS          2 23 42 0 25 */
+544,	/* OBJ_setct_CapReqTBS              2 23 42 0 26 */
+545,	/* OBJ_setct_CapReqTBSX             2 23 42 0 27 */
+546,	/* OBJ_setct_CapResData             2 23 42 0 28 */
+547,	/* OBJ_setct_CapRevReqTBS           2 23 42 0 29 */
+548,	/* OBJ_setct_CapRevReqTBSX          2 23 42 0 30 */
+549,	/* OBJ_setct_CapRevResData          2 23 42 0 31 */
+550,	/* OBJ_setct_CredReqTBS             2 23 42 0 32 */
+551,	/* OBJ_setct_CredReqTBSX            2 23 42 0 33 */
+552,	/* OBJ_setct_CredResData            2 23 42 0 34 */
+553,	/* OBJ_setct_CredRevReqTBS          2 23 42 0 35 */
+554,	/* OBJ_setct_CredRevReqTBSX         2 23 42 0 36 */
+555,	/* OBJ_setct_CredRevResData         2 23 42 0 37 */
+556,	/* OBJ_setct_PCertReqData           2 23 42 0 38 */
+557,	/* OBJ_setct_PCertResTBS            2 23 42 0 39 */
+558,	/* OBJ_setct_BatchAdminReqData      2 23 42 0 40 */
+559,	/* OBJ_setct_BatchAdminResData      2 23 42 0 41 */
+560,	/* OBJ_setct_CardCInitResTBS        2 23 42 0 42 */
+561,	/* OBJ_setct_MeAqCInitResTBS        2 23 42 0 43 */
+562,	/* OBJ_setct_RegFormResTBS          2 23 42 0 44 */
+563,	/* OBJ_setct_CertReqData            2 23 42 0 45 */
+564,	/* OBJ_setct_CertReqTBS             2 23 42 0 46 */
+565,	/* OBJ_setct_CertResData            2 23 42 0 47 */
+566,	/* OBJ_setct_CertInqReqTBS          2 23 42 0 48 */
+567,	/* OBJ_setct_ErrorTBS               2 23 42 0 49 */
+568,	/* OBJ_setct_PIDualSignedTBE        2 23 42 0 50 */
+569,	/* OBJ_setct_PIUnsignedTBE          2 23 42 0 51 */
+570,	/* OBJ_setct_AuthReqTBE             2 23 42 0 52 */
+571,	/* OBJ_setct_AuthResTBE             2 23 42 0 53 */
+572,	/* OBJ_setct_AuthResTBEX            2 23 42 0 54 */
+573,	/* OBJ_setct_AuthTokenTBE           2 23 42 0 55 */
+574,	/* OBJ_setct_CapTokenTBE            2 23 42 0 56 */
+575,	/* OBJ_setct_CapTokenTBEX           2 23 42 0 57 */
+576,	/* OBJ_setct_AcqCardCodeMsgTBE      2 23 42 0 58 */
+577,	/* OBJ_setct_AuthRevReqTBE          2 23 42 0 59 */
+578,	/* OBJ_setct_AuthRevResTBE          2 23 42 0 60 */
+579,	/* OBJ_setct_AuthRevResTBEB         2 23 42 0 61 */
+580,	/* OBJ_setct_CapReqTBE              2 23 42 0 62 */
+581,	/* OBJ_setct_CapReqTBEX             2 23 42 0 63 */
+582,	/* OBJ_setct_CapResTBE              2 23 42 0 64 */
+583,	/* OBJ_setct_CapRevReqTBE           2 23 42 0 65 */
+584,	/* OBJ_setct_CapRevReqTBEX          2 23 42 0 66 */
+585,	/* OBJ_setct_CapRevResTBE           2 23 42 0 67 */
+586,	/* OBJ_setct_CredReqTBE             2 23 42 0 68 */
+587,	/* OBJ_setct_CredReqTBEX            2 23 42 0 69 */
+588,	/* OBJ_setct_CredResTBE             2 23 42 0 70 */
+589,	/* OBJ_setct_CredRevReqTBE          2 23 42 0 71 */
+590,	/* OBJ_setct_CredRevReqTBEX         2 23 42 0 72 */
+591,	/* OBJ_setct_CredRevResTBE          2 23 42 0 73 */
+592,	/* OBJ_setct_BatchAdminReqTBE       2 23 42 0 74 */
+593,	/* OBJ_setct_BatchAdminResTBE       2 23 42 0 75 */
+594,	/* OBJ_setct_RegFormReqTBE          2 23 42 0 76 */
+595,	/* OBJ_setct_CertReqTBE             2 23 42 0 77 */
+596,	/* OBJ_setct_CertReqTBEX            2 23 42 0 78 */
+597,	/* OBJ_setct_CertResTBE             2 23 42 0 79 */
+598,	/* OBJ_setct_CRLNotificationTBS     2 23 42 0 80 */
+599,	/* OBJ_setct_CRLNotificationResTBS  2 23 42 0 81 */
+600,	/* OBJ_setct_BCIDistributionTBS     2 23 42 0 82 */
+601,	/* OBJ_setext_genCrypt              2 23 42 1 1 */
+602,	/* OBJ_setext_miAuth                2 23 42 1 3 */
+603,	/* OBJ_setext_pinSecure             2 23 42 1 4 */
+604,	/* OBJ_setext_pinAny                2 23 42 1 5 */
+605,	/* OBJ_setext_track2                2 23 42 1 7 */
+606,	/* OBJ_setext_cv                    2 23 42 1 8 */
+620,	/* OBJ_setAttr_Cert                 2 23 42 3 0 */
+621,	/* OBJ_setAttr_PGWYcap              2 23 42 3 1 */
+622,	/* OBJ_setAttr_TokenType            2 23 42 3 2 */
+623,	/* OBJ_setAttr_IssCap               2 23 42 3 3 */
+607,	/* OBJ_set_policy_root              2 23 42 5 0 */
+608,	/* OBJ_setCext_hashedRoot           2 23 42 7 0 */
+609,	/* OBJ_setCext_certType             2 23 42 7 1 */
+610,	/* OBJ_setCext_merchData            2 23 42 7 2 */
+611,	/* OBJ_setCext_cCertRequired        2 23 42 7 3 */
+612,	/* OBJ_setCext_tunneling            2 23 42 7 4 */
+613,	/* OBJ_setCext_setExt               2 23 42 7 5 */
+614,	/* OBJ_setCext_setQualf             2 23 42 7 6 */
+615,	/* OBJ_setCext_PGWYcapabilities     2 23 42 7 7 */
+616,	/* OBJ_setCext_TokenIdentifier      2 23 42 7 8 */
+617,	/* OBJ_setCext_Track2Data           2 23 42 7 9 */
+618,	/* OBJ_setCext_TokenType            2 23 42 7 10 */
+619,	/* OBJ_setCext_IssuerCapabilities   2 23 42 7 11 */
+636,	/* OBJ_set_brand_IATA_ATA           2 23 42 8 1 */
+640,	/* OBJ_set_brand_Visa               2 23 42 8 4 */
+641,	/* OBJ_set_brand_MasterCard         2 23 42 8 5 */
+637,	/* OBJ_set_brand_Diners             2 23 42 8 30 */
+638,	/* OBJ_set_brand_AmericanExpress    2 23 42 8 34 */
+639,	/* OBJ_set_brand_JCB                2 23 42 8 35 */
+805,	/* OBJ_cryptopro                    1 2 643 2 2 */
+806,	/* OBJ_cryptocom                    1 2 643 2 9 */
+184,	/* OBJ_X9_57                        1 2 840 10040 */
+405,	/* OBJ_ansi_X9_62                   1 2 840 10045 */
+389,	/* OBJ_Enterprises                  1 3 6 1 4 1 */
+504,	/* OBJ_mime_mhs                     1 3 6 1 7 1 */
+104,	/* OBJ_md5WithRSA                   1 3 14 3 2 3 */
+29,	/* OBJ_des_ecb                      1 3 14 3 2 6 */
+31,	/* OBJ_des_cbc                      1 3 14 3 2 7 */
+45,	/* OBJ_des_ofb64                    1 3 14 3 2 8 */
+30,	/* OBJ_des_cfb64                    1 3 14 3 2 9 */
+377,	/* OBJ_rsaSignature                 1 3 14 3 2 11 */
+67,	/* OBJ_dsa_2                        1 3 14 3 2 12 */
+66,	/* OBJ_dsaWithSHA                   1 3 14 3 2 13 */
+42,	/* OBJ_shaWithRSAEncryption         1 3 14 3 2 15 */
+32,	/* OBJ_des_ede_ecb                  1 3 14 3 2 17 */
+41,	/* OBJ_sha                          1 3 14 3 2 18 */
+64,	/* OBJ_sha1                         1 3 14 3 2 26 */
+70,	/* OBJ_dsaWithSHA1_2                1 3 14 3 2 27 */
+115,	/* OBJ_sha1WithRSA                  1 3 14 3 2 29 */
+117,	/* OBJ_ripemd160                    1 3 36 3 2 1 */
+143,	/* OBJ_sxnet                        1 3 101 1 4 1 */
+721,	/* OBJ_sect163k1                    1 3 132 0 1 */
+722,	/* OBJ_sect163r1                    1 3 132 0 2 */
+728,	/* OBJ_sect239k1                    1 3 132 0 3 */
+717,	/* OBJ_sect113r1                    1 3 132 0 4 */
+718,	/* OBJ_sect113r2                    1 3 132 0 5 */
+704,	/* OBJ_secp112r1                    1 3 132 0 6 */
+705,	/* OBJ_secp112r2                    1 3 132 0 7 */
+709,	/* OBJ_secp160r1                    1 3 132 0 8 */
+708,	/* OBJ_secp160k1                    1 3 132 0 9 */
+714,	/* OBJ_secp256k1                    1 3 132 0 10 */
+723,	/* OBJ_sect163r2                    1 3 132 0 15 */
+729,	/* OBJ_sect283k1                    1 3 132 0 16 */
+730,	/* OBJ_sect283r1                    1 3 132 0 17 */
+719,	/* OBJ_sect131r1                    1 3 132 0 22 */
+720,	/* OBJ_sect131r2                    1 3 132 0 23 */
+724,	/* OBJ_sect193r1                    1 3 132 0 24 */
+725,	/* OBJ_sect193r2                    1 3 132 0 25 */
+726,	/* OBJ_sect233k1                    1 3 132 0 26 */
+727,	/* OBJ_sect233r1                    1 3 132 0 27 */
+706,	/* OBJ_secp128r1                    1 3 132 0 28 */
+707,	/* OBJ_secp128r2                    1 3 132 0 29 */
+710,	/* OBJ_secp160r2                    1 3 132 0 30 */
+711,	/* OBJ_secp192k1                    1 3 132 0 31 */
+712,	/* OBJ_secp224k1                    1 3 132 0 32 */
+713,	/* OBJ_secp224r1                    1 3 132 0 33 */
+715,	/* OBJ_secp384r1                    1 3 132 0 34 */
+716,	/* OBJ_secp521r1                    1 3 132 0 35 */
+731,	/* OBJ_sect409k1                    1 3 132 0 36 */
+732,	/* OBJ_sect409r1                    1 3 132 0 37 */
+733,	/* OBJ_sect571k1                    1 3 132 0 38 */
+734,	/* OBJ_sect571r1                    1 3 132 0 39 */
+624,	/* OBJ_set_rootKeyThumb             2 23 42 3 0 0 */
+625,	/* OBJ_set_addPolicy                2 23 42 3 0 1 */
+626,	/* OBJ_setAttr_Token_EMV            2 23 42 3 2 1 */
+627,	/* OBJ_setAttr_Token_B0Prime        2 23 42 3 2 2 */
+628,	/* OBJ_setAttr_IssCap_CVM           2 23 42 3 3 3 */
+629,	/* OBJ_setAttr_IssCap_T2            2 23 42 3 3 4 */
+630,	/* OBJ_setAttr_IssCap_Sig           2 23 42 3 3 5 */
+642,	/* OBJ_set_brand_Novus              2 23 42 8 6011 */
+735,	/* OBJ_wap_wsg_idm_ecid_wtls1       2 23 43 1 4 1 */
+736,	/* OBJ_wap_wsg_idm_ecid_wtls3       2 23 43 1 4 3 */
+737,	/* OBJ_wap_wsg_idm_ecid_wtls4       2 23 43 1 4 4 */
+738,	/* OBJ_wap_wsg_idm_ecid_wtls5       2 23 43 1 4 5 */
+739,	/* OBJ_wap_wsg_idm_ecid_wtls6       2 23 43 1 4 6 */
+740,	/* OBJ_wap_wsg_idm_ecid_wtls7       2 23 43 1 4 7 */
+741,	/* OBJ_wap_wsg_idm_ecid_wtls8       2 23 43 1 4 8 */
+742,	/* OBJ_wap_wsg_idm_ecid_wtls9       2 23 43 1 4 9 */
+743,	/* OBJ_wap_wsg_idm_ecid_wtls10      2 23 43 1 4 10 */
+744,	/* OBJ_wap_wsg_idm_ecid_wtls11      2 23 43 1 4 11 */
+745,	/* OBJ_wap_wsg_idm_ecid_wtls12      2 23 43 1 4 12 */
+804,	/* OBJ_whirlpool                    1 0 10118 3 0 55 */
+124,	/* OBJ_rle_compression              1 1 1 1 666 1 */
+773,	/* OBJ_kisa                         1 2 410 200004 */
+807,	/* OBJ_id_GostR3411_94_with_GostR3410_2001 1 2 643 2 2 3 */
+808,	/* OBJ_id_GostR3411_94_with_GostR3410_94 1 2 643 2 2 4 */
+809,	/* OBJ_id_GostR3411_94              1 2 643 2 2 9 */
+810,	/* OBJ_id_HMACGostR3411_94          1 2 643 2 2 10 */
+811,	/* OBJ_id_GostR3410_2001            1 2 643 2 2 19 */
+812,	/* OBJ_id_GostR3410_94              1 2 643 2 2 20 */
+813,	/* OBJ_id_Gost28147_89              1 2 643 2 2 21 */
+815,	/* OBJ_id_Gost28147_89_MAC          1 2 643 2 2 22 */
+816,	/* OBJ_id_GostR3411_94_prf          1 2 643 2 2 23 */
+817,	/* OBJ_id_GostR3410_2001DH          1 2 643 2 2 98 */
+818,	/* OBJ_id_GostR3410_94DH            1 2 643 2 2 99 */
+ 1,	/* OBJ_rsadsi                       1 2 840 113549 */
+185,	/* OBJ_X9cm                         1 2 840 10040 4 */
+127,	/* OBJ_id_pkix                      1 3 6 1 5 5 7 */
+505,	/* OBJ_mime_mhs_headings            1 3 6 1 7 1 1 */
+506,	/* OBJ_mime_mhs_bodies              1 3 6 1 7 1 2 */
+119,	/* OBJ_ripemd160WithRSA             1 3 36 3 3 1 2 */
+631,	/* OBJ_setAttr_GenCryptgrm          2 23 42 3 3 3 1 */
+632,	/* OBJ_setAttr_T2Enc                2 23 42 3 3 4 1 */
+633,	/* OBJ_setAttr_T2cleartxt           2 23 42 3 3 4 2 */
+634,	/* OBJ_setAttr_TokICCsig            2 23 42 3 3 5 1 */
+635,	/* OBJ_setAttr_SecDevSig            2 23 42 3 3 5 2 */
+436,	/* OBJ_ucl                          0 9 2342 19200300 */
+820,	/* OBJ_id_Gost28147_89_None_KeyMeshing 1 2 643 2 2 14 0 */
+819,	/* OBJ_id_Gost28147_89_CryptoPro_KeyMeshing 1 2 643 2 2 14 1 */
+845,	/* OBJ_id_GostR3410_94_a            1 2 643 2 2 20 1 */
+846,	/* OBJ_id_GostR3410_94_aBis         1 2 643 2 2 20 2 */
+847,	/* OBJ_id_GostR3410_94_b            1 2 643 2 2 20 3 */
+848,	/* OBJ_id_GostR3410_94_bBis         1 2 643 2 2 20 4 */
+821,	/* OBJ_id_GostR3411_94_TestParamSet 1 2 643 2 2 30 0 */
+822,	/* OBJ_id_GostR3411_94_CryptoProParamSet 1 2 643 2 2 30 1 */
+823,	/* OBJ_id_Gost28147_89_TestParamSet 1 2 643 2 2 31 0 */
+824,	/* OBJ_id_Gost28147_89_CryptoPro_A_ParamSet 1 2 643 2 2 31 1 */
+825,	/* OBJ_id_Gost28147_89_CryptoPro_B_ParamSet 1 2 643 2 2 31 2 */
+826,	/* OBJ_id_Gost28147_89_CryptoPro_C_ParamSet 1 2 643 2 2 31 3 */
+827,	/* OBJ_id_Gost28147_89_CryptoPro_D_ParamSet 1 2 643 2 2 31 4 */
+828,	/* OBJ_id_Gost28147_89_CryptoPro_Oscar_1_1_ParamSet 1 2 643 2 2 31 5 */
+829,	/* OBJ_id_Gost28147_89_CryptoPro_Oscar_1_0_ParamSet 1 2 643 2 2 31 6 */
+830,	/* OBJ_id_Gost28147_89_CryptoPro_RIC_1_ParamSet 1 2 643 2 2 31 7 */
+831,	/* OBJ_id_GostR3410_94_TestParamSet 1 2 643 2 2 32 0 */
+832,	/* OBJ_id_GostR3410_94_CryptoPro_A_ParamSet 1 2 643 2 2 32 2 */
+833,	/* OBJ_id_GostR3410_94_CryptoPro_B_ParamSet 1 2 643 2 2 32 3 */
+834,	/* OBJ_id_GostR3410_94_CryptoPro_C_ParamSet 1 2 643 2 2 32 4 */
+835,	/* OBJ_id_GostR3410_94_CryptoPro_D_ParamSet 1 2 643 2 2 32 5 */
+836,	/* OBJ_id_GostR3410_94_CryptoPro_XchA_ParamSet 1 2 643 2 2 33 1 */
+837,	/* OBJ_id_GostR3410_94_CryptoPro_XchB_ParamSet 1 2 643 2 2 33 2 */
+838,	/* OBJ_id_GostR3410_94_CryptoPro_XchC_ParamSet 1 2 643 2 2 33 3 */
+839,	/* OBJ_id_GostR3410_2001_TestParamSet 1 2 643 2 2 35 0 */
+840,	/* OBJ_id_GostR3410_2001_CryptoPro_A_ParamSet 1 2 643 2 2 35 1 */
+841,	/* OBJ_id_GostR3410_2001_CryptoPro_B_ParamSet 1 2 643 2 2 35 2 */
+842,	/* OBJ_id_GostR3410_2001_CryptoPro_C_ParamSet 1 2 643 2 2 35 3 */
+843,	/* OBJ_id_GostR3410_2001_CryptoPro_XchA_ParamSet 1 2 643 2 2 36 0 */
+844,	/* OBJ_id_GostR3410_2001_CryptoPro_XchB_ParamSet 1 2 643 2 2 36 1 */
+ 2,	/* OBJ_pkcs                         1 2 840 113549 1 */
+431,	/* OBJ_hold_instruction_none        1 2 840 10040 2 1 */
+432,	/* OBJ_hold_instruction_call_issuer 1 2 840 10040 2 2 */
+433,	/* OBJ_hold_instruction_reject      1 2 840 10040 2 3 */
+116,	/* OBJ_dsa                          1 2 840 10040 4 1 */
+113,	/* OBJ_dsaWithSHA1                  1 2 840 10040 4 3 */
+406,	/* OBJ_X9_62_prime_field            1 2 840 10045 1 1 */
+407,	/* OBJ_X9_62_characteristic_two_field 1 2 840 10045 1 2 */
+408,	/* OBJ_X9_62_id_ecPublicKey         1 2 840 10045 2 1 */
+416,	/* OBJ_ecdsa_with_SHA1              1 2 840 10045 4 1 */
+791,	/* OBJ_ecdsa_with_Recommended       1 2 840 10045 4 2 */
+792,	/* OBJ_ecdsa_with_Specified         1 2 840 10045 4 3 */
+258,	/* OBJ_id_pkix_mod                  1 3 6 1 5 5 7 0 */
+175,	/* OBJ_id_pe                        1 3 6 1 5 5 7 1 */
+259,	/* OBJ_id_qt                        1 3 6 1 5 5 7 2 */
+128,	/* OBJ_id_kp                        1 3 6 1 5 5 7 3 */
+260,	/* OBJ_id_it                        1 3 6 1 5 5 7 4 */
+261,	/* OBJ_id_pkip                      1 3 6 1 5 5 7 5 */
+262,	/* OBJ_id_alg                       1 3 6 1 5 5 7 6 */
+263,	/* OBJ_id_cmc                       1 3 6 1 5 5 7 7 */
+264,	/* OBJ_id_on                        1 3 6 1 5 5 7 8 */
+265,	/* OBJ_id_pda                       1 3 6 1 5 5 7 9 */
+266,	/* OBJ_id_aca                       1 3 6 1 5 5 7 10 */
+267,	/* OBJ_id_qcs                       1 3 6 1 5 5 7 11 */
+268,	/* OBJ_id_cct                       1 3 6 1 5 5 7 12 */
+662,	/* OBJ_id_ppl                       1 3 6 1 5 5 7 21 */
+176,	/* OBJ_id_ad                        1 3 6 1 5 5 7 48 */
+507,	/* OBJ_id_hex_partial_message       1 3 6 1 7 1 1 1 */
+508,	/* OBJ_id_hex_multipart_message     1 3 6 1 7 1 1 2 */
+57,	/* OBJ_netscape                     2 16 840 1 113730 */
+754,	/* OBJ_camellia_128_ecb             0 3 4401 5 3 1 9 1 */
+766,	/* OBJ_camellia_128_ofb128          0 3 4401 5 3 1 9 3 */
+757,	/* OBJ_camellia_128_cfb128          0 3 4401 5 3 1 9 4 */
+755,	/* OBJ_camellia_192_ecb             0 3 4401 5 3 1 9 21 */
+767,	/* OBJ_camellia_192_ofb128          0 3 4401 5 3 1 9 23 */
+758,	/* OBJ_camellia_192_cfb128          0 3 4401 5 3 1 9 24 */
+756,	/* OBJ_camellia_256_ecb             0 3 4401 5 3 1 9 41 */
+768,	/* OBJ_camellia_256_ofb128          0 3 4401 5 3 1 9 43 */
+759,	/* OBJ_camellia_256_cfb128          0 3 4401 5 3 1 9 44 */
+437,	/* OBJ_pilot                        0 9 2342 19200300 100 */
+776,	/* OBJ_seed_ecb                     1 2 410 200004 1 3 */
+777,	/* OBJ_seed_cbc                     1 2 410 200004 1 4 */
+779,	/* OBJ_seed_cfb128                  1 2 410 200004 1 5 */
+778,	/* OBJ_seed_ofb128                  1 2 410 200004 1 6 */
+852,	/* OBJ_id_GostR3411_94_with_GostR3410_94_cc 1 2 643 2 9 1 3 3 */
+853,	/* OBJ_id_GostR3411_94_with_GostR3410_2001_cc 1 2 643 2 9 1 3 4 */
+850,	/* OBJ_id_GostR3410_94_cc           1 2 643 2 9 1 5 3 */
+851,	/* OBJ_id_GostR3410_2001_cc         1 2 643 2 9 1 5 4 */
+849,	/* OBJ_id_Gost28147_89_cc           1 2 643 2 9 1 6 1 */
+854,	/* OBJ_id_GostR3410_2001_ParamSet_cc 1 2 643 2 9 1 8 1 */
+186,	/* OBJ_pkcs1                        1 2 840 113549 1 1 */
+27,	/* OBJ_pkcs3                        1 2 840 113549 1 3 */
+187,	/* OBJ_pkcs5                        1 2 840 113549 1 5 */
+20,	/* OBJ_pkcs7                        1 2 840 113549 1 7 */
+47,	/* OBJ_pkcs9                        1 2 840 113549 1 9 */
+ 3,	/* OBJ_md2                          1 2 840 113549 2 2 */
+257,	/* OBJ_md4                          1 2 840 113549 2 4 */
+ 4,	/* OBJ_md5                          1 2 840 113549 2 5 */
+797,	/* OBJ_hmacWithMD5                  1 2 840 113549 2 6 */
+163,	/* OBJ_hmacWithSHA1                 1 2 840 113549 2 7 */
+798,	/* OBJ_hmacWithSHA224               1 2 840 113549 2 8 */
+799,	/* OBJ_hmacWithSHA256               1 2 840 113549 2 9 */
+800,	/* OBJ_hmacWithSHA384               1 2 840 113549 2 10 */
+801,	/* OBJ_hmacWithSHA512               1 2 840 113549 2 11 */
+37,	/* OBJ_rc2_cbc                      1 2 840 113549 3 2 */
+ 5,	/* OBJ_rc4                          1 2 840 113549 3 4 */
+44,	/* OBJ_des_ede3_cbc                 1 2 840 113549 3 7 */
+120,	/* OBJ_rc5_cbc                      1 2 840 113549 3 8 */
+643,	/* OBJ_des_cdmf                     1 2 840 113549 3 10 */
+680,	/* OBJ_X9_62_id_characteristic_two_basis 1 2 840 10045 1 2 3 */
+684,	/* OBJ_X9_62_c2pnb163v1             1 2 840 10045 3 0 1 */
+685,	/* OBJ_X9_62_c2pnb163v2             1 2 840 10045 3 0 2 */
+686,	/* OBJ_X9_62_c2pnb163v3             1 2 840 10045 3 0 3 */
+687,	/* OBJ_X9_62_c2pnb176v1             1 2 840 10045 3 0 4 */
+688,	/* OBJ_X9_62_c2tnb191v1             1 2 840 10045 3 0 5 */
+689,	/* OBJ_X9_62_c2tnb191v2             1 2 840 10045 3 0 6 */
+690,	/* OBJ_X9_62_c2tnb191v3             1 2 840 10045 3 0 7 */
+691,	/* OBJ_X9_62_c2onb191v4             1 2 840 10045 3 0 8 */
+692,	/* OBJ_X9_62_c2onb191v5             1 2 840 10045 3 0 9 */
+693,	/* OBJ_X9_62_c2pnb208w1             1 2 840 10045 3 0 10 */
+694,	/* OBJ_X9_62_c2tnb239v1             1 2 840 10045 3 0 11 */
+695,	/* OBJ_X9_62_c2tnb239v2             1 2 840 10045 3 0 12 */
+696,	/* OBJ_X9_62_c2tnb239v3             1 2 840 10045 3 0 13 */
+697,	/* OBJ_X9_62_c2onb239v4             1 2 840 10045 3 0 14 */
+698,	/* OBJ_X9_62_c2onb239v5             1 2 840 10045 3 0 15 */
+699,	/* OBJ_X9_62_c2pnb272w1             1 2 840 10045 3 0 16 */
+700,	/* OBJ_X9_62_c2pnb304w1             1 2 840 10045 3 0 17 */
+701,	/* OBJ_X9_62_c2tnb359v1             1 2 840 10045 3 0 18 */
+702,	/* OBJ_X9_62_c2pnb368w1             1 2 840 10045 3 0 19 */
+703,	/* OBJ_X9_62_c2tnb431r1             1 2 840 10045 3 0 20 */
+409,	/* OBJ_X9_62_prime192v1             1 2 840 10045 3 1 1 */
+410,	/* OBJ_X9_62_prime192v2             1 2 840 10045 3 1 2 */
+411,	/* OBJ_X9_62_prime192v3             1 2 840 10045 3 1 3 */
+412,	/* OBJ_X9_62_prime239v1             1 2 840 10045 3 1 4 */
+413,	/* OBJ_X9_62_prime239v2             1 2 840 10045 3 1 5 */
+414,	/* OBJ_X9_62_prime239v3             1 2 840 10045 3 1 6 */
+415,	/* OBJ_X9_62_prime256v1             1 2 840 10045 3 1 7 */
+793,	/* OBJ_ecdsa_with_SHA224            1 2 840 10045 4 3 1 */
+794,	/* OBJ_ecdsa_with_SHA256            1 2 840 10045 4 3 2 */
+795,	/* OBJ_ecdsa_with_SHA384            1 2 840 10045 4 3 3 */
+796,	/* OBJ_ecdsa_with_SHA512            1 2 840 10045 4 3 4 */
+269,	/* OBJ_id_pkix1_explicit_88         1 3 6 1 5 5 7 0 1 */
+270,	/* OBJ_id_pkix1_implicit_88         1 3 6 1 5 5 7 0 2 */
+271,	/* OBJ_id_pkix1_explicit_93         1 3 6 1 5 5 7 0 3 */
+272,	/* OBJ_id_pkix1_implicit_93         1 3 6 1 5 5 7 0 4 */
+273,	/* OBJ_id_mod_crmf                  1 3 6 1 5 5 7 0 5 */
+274,	/* OBJ_id_mod_cmc                   1 3 6 1 5 5 7 0 6 */
+275,	/* OBJ_id_mod_kea_profile_88        1 3 6 1 5 5 7 0 7 */
+276,	/* OBJ_id_mod_kea_profile_93        1 3 6 1 5 5 7 0 8 */
+277,	/* OBJ_id_mod_cmp                   1 3 6 1 5 5 7 0 9 */
+278,	/* OBJ_id_mod_qualified_cert_88     1 3 6 1 5 5 7 0 10 */
+279,	/* OBJ_id_mod_qualified_cert_93     1 3 6 1 5 5 7 0 11 */
+280,	/* OBJ_id_mod_attribute_cert        1 3 6 1 5 5 7 0 12 */
+281,	/* OBJ_id_mod_timestamp_protocol    1 3 6 1 5 5 7 0 13 */
+282,	/* OBJ_id_mod_ocsp                  1 3 6 1 5 5 7 0 14 */
+283,	/* OBJ_id_mod_dvcs                  1 3 6 1 5 5 7 0 15 */
+284,	/* OBJ_id_mod_cmp2000               1 3 6 1 5 5 7 0 16 */
+177,	/* OBJ_info_access                  1 3 6 1 5 5 7 1 1 */
+285,	/* OBJ_biometricInfo                1 3 6 1 5 5 7 1 2 */
+286,	/* OBJ_qcStatements                 1 3 6 1 5 5 7 1 3 */
+287,	/* OBJ_ac_auditEntity               1 3 6 1 5 5 7 1 4 */
+288,	/* OBJ_ac_targeting                 1 3 6 1 5 5 7 1 5 */
+289,	/* OBJ_aaControls                   1 3 6 1 5 5 7 1 6 */
+290,	/* OBJ_sbgp_ipAddrBlock             1 3 6 1 5 5 7 1 7 */
+291,	/* OBJ_sbgp_autonomousSysNum        1 3 6 1 5 5 7 1 8 */
+292,	/* OBJ_sbgp_routerIdentifier        1 3 6 1 5 5 7 1 9 */
+397,	/* OBJ_ac_proxying                  1 3 6 1 5 5 7 1 10 */
+398,	/* OBJ_sinfo_access                 1 3 6 1 5 5 7 1 11 */
+663,	/* OBJ_proxyCertInfo                1 3 6 1 5 5 7 1 14 */
+164,	/* OBJ_id_qt_cps                    1 3 6 1 5 5 7 2 1 */
+165,	/* OBJ_id_qt_unotice                1 3 6 1 5 5 7 2 2 */
+293,	/* OBJ_textNotice                   1 3 6 1 5 5 7 2 3 */
+129,	/* OBJ_server_auth                  1 3 6 1 5 5 7 3 1 */
+130,	/* OBJ_client_auth                  1 3 6 1 5 5 7 3 2 */
+131,	/* OBJ_code_sign                    1 3 6 1 5 5 7 3 3 */
+132,	/* OBJ_email_protect                1 3 6 1 5 5 7 3 4 */
+294,	/* OBJ_ipsecEndSystem               1 3 6 1 5 5 7 3 5 */
+295,	/* OBJ_ipsecTunnel                  1 3 6 1 5 5 7 3 6 */
+296,	/* OBJ_ipsecUser                    1 3 6 1 5 5 7 3 7 */
+133,	/* OBJ_time_stamp                   1 3 6 1 5 5 7 3 8 */
+180,	/* OBJ_OCSP_sign                    1 3 6 1 5 5 7 3 9 */
+297,	/* OBJ_dvcs                         1 3 6 1 5 5 7 3 10 */
+298,	/* OBJ_id_it_caProtEncCert          1 3 6 1 5 5 7 4 1 */
+299,	/* OBJ_id_it_signKeyPairTypes       1 3 6 1 5 5 7 4 2 */
+300,	/* OBJ_id_it_encKeyPairTypes        1 3 6 1 5 5 7 4 3 */
+301,	/* OBJ_id_it_preferredSymmAlg       1 3 6 1 5 5 7 4 4 */
+302,	/* OBJ_id_it_caKeyUpdateInfo        1 3 6 1 5 5 7 4 5 */
+303,	/* OBJ_id_it_currentCRL             1 3 6 1 5 5 7 4 6 */
+304,	/* OBJ_id_it_unsupportedOIDs        1 3 6 1 5 5 7 4 7 */
+305,	/* OBJ_id_it_subscriptionRequest    1 3 6 1 5 5 7 4 8 */
+306,	/* OBJ_id_it_subscriptionResponse   1 3 6 1 5 5 7 4 9 */
+307,	/* OBJ_id_it_keyPairParamReq        1 3 6 1 5 5 7 4 10 */
+308,	/* OBJ_id_it_keyPairParamRep        1 3 6 1 5 5 7 4 11 */
+309,	/* OBJ_id_it_revPassphrase          1 3 6 1 5 5 7 4 12 */
+310,	/* OBJ_id_it_implicitConfirm        1 3 6 1 5 5 7 4 13 */
+311,	/* OBJ_id_it_confirmWaitTime        1 3 6 1 5 5 7 4 14 */
+312,	/* OBJ_id_it_origPKIMessage         1 3 6 1 5 5 7 4 15 */
+784,	/* OBJ_id_it_suppLangTags           1 3 6 1 5 5 7 4 16 */
+313,	/* OBJ_id_regCtrl                   1 3 6 1 5 5 7 5 1 */
+314,	/* OBJ_id_regInfo                   1 3 6 1 5 5 7 5 2 */
+323,	/* OBJ_id_alg_des40                 1 3 6 1 5 5 7 6 1 */
+324,	/* OBJ_id_alg_noSignature           1 3 6 1 5 5 7 6 2 */
+325,	/* OBJ_id_alg_dh_sig_hmac_sha1      1 3 6 1 5 5 7 6 3 */
+326,	/* OBJ_id_alg_dh_pop                1 3 6 1 5 5 7 6 4 */
+327,	/* OBJ_id_cmc_statusInfo            1 3 6 1 5 5 7 7 1 */
+328,	/* OBJ_id_cmc_identification        1 3 6 1 5 5 7 7 2 */
+329,	/* OBJ_id_cmc_identityProof         1 3 6 1 5 5 7 7 3 */
+330,	/* OBJ_id_cmc_dataReturn            1 3 6 1 5 5 7 7 4 */
+331,	/* OBJ_id_cmc_transactionId         1 3 6 1 5 5 7 7 5 */
+332,	/* OBJ_id_cmc_senderNonce           1 3 6 1 5 5 7 7 6 */
+333,	/* OBJ_id_cmc_recipientNonce        1 3 6 1 5 5 7 7 7 */
+334,	/* OBJ_id_cmc_addExtensions         1 3 6 1 5 5 7 7 8 */
+335,	/* OBJ_id_cmc_encryptedPOP          1 3 6 1 5 5 7 7 9 */
+336,	/* OBJ_id_cmc_decryptedPOP          1 3 6 1 5 5 7 7 10 */
+337,	/* OBJ_id_cmc_lraPOPWitness         1 3 6 1 5 5 7 7 11 */
+338,	/* OBJ_id_cmc_getCert               1 3 6 1 5 5 7 7 15 */
+339,	/* OBJ_id_cmc_getCRL                1 3 6 1 5 5 7 7 16 */
+340,	/* OBJ_id_cmc_revokeRequest         1 3 6 1 5 5 7 7 17 */
+341,	/* OBJ_id_cmc_regInfo               1 3 6 1 5 5 7 7 18 */
+342,	/* OBJ_id_cmc_responseInfo          1 3 6 1 5 5 7 7 19 */
+343,	/* OBJ_id_cmc_queryPending          1 3 6 1 5 5 7 7 21 */
+344,	/* OBJ_id_cmc_popLinkRandom         1 3 6 1 5 5 7 7 22 */
+345,	/* OBJ_id_cmc_popLinkWitness        1 3 6 1 5 5 7 7 23 */
+346,	/* OBJ_id_cmc_confirmCertAcceptance 1 3 6 1 5 5 7 7 24 */
+347,	/* OBJ_id_on_personalData           1 3 6 1 5 5 7 8 1 */
+858,	/* OBJ_id_on_permanentIdentifier    1 3 6 1 5 5 7 8 3 */
+348,	/* OBJ_id_pda_dateOfBirth           1 3 6 1 5 5 7 9 1 */
+349,	/* OBJ_id_pda_placeOfBirth          1 3 6 1 5 5 7 9 2 */
+351,	/* OBJ_id_pda_gender                1 3 6 1 5 5 7 9 3 */
+352,	/* OBJ_id_pda_countryOfCitizenship  1 3 6 1 5 5 7 9 4 */
+353,	/* OBJ_id_pda_countryOfResidence    1 3 6 1 5 5 7 9 5 */
+354,	/* OBJ_id_aca_authenticationInfo    1 3 6 1 5 5 7 10 1 */
+355,	/* OBJ_id_aca_accessIdentity        1 3 6 1 5 5 7 10 2 */
+356,	/* OBJ_id_aca_chargingIdentity      1 3 6 1 5 5 7 10 3 */
+357,	/* OBJ_id_aca_group                 1 3 6 1 5 5 7 10 4 */
+358,	/* OBJ_id_aca_role                  1 3 6 1 5 5 7 10 5 */
+399,	/* OBJ_id_aca_encAttrs              1 3 6 1 5 5 7 10 6 */
+359,	/* OBJ_id_qcs_pkixQCSyntax_v1       1 3 6 1 5 5 7 11 1 */
+360,	/* OBJ_id_cct_crs                   1 3 6 1 5 5 7 12 1 */
+361,	/* OBJ_id_cct_PKIData               1 3 6 1 5 5 7 12 2 */
+362,	/* OBJ_id_cct_PKIResponse           1 3 6 1 5 5 7 12 3 */
+664,	/* OBJ_id_ppl_anyLanguage           1 3 6 1 5 5 7 21 0 */
+665,	/* OBJ_id_ppl_inheritAll            1 3 6 1 5 5 7 21 1 */
+667,	/* OBJ_Independent                  1 3 6 1 5 5 7 21 2 */
+178,	/* OBJ_ad_OCSP                      1 3 6 1 5 5 7 48 1 */
+179,	/* OBJ_ad_ca_issuers                1 3 6 1 5 5 7 48 2 */
+363,	/* OBJ_ad_timeStamping              1 3 6 1 5 5 7 48 3 */
+364,	/* OBJ_ad_dvcs                      1 3 6 1 5 5 7 48 4 */
+785,	/* OBJ_caRepository                 1 3 6 1 5 5 7 48 5 */
+780,	/* OBJ_hmac_md5                     1 3 6 1 5 5 8 1 1 */
+781,	/* OBJ_hmac_sha1                    1 3 6 1 5 5 8 1 2 */
+58,	/* OBJ_netscape_cert_extension      2 16 840 1 113730 1 */
+59,	/* OBJ_netscape_data_type           2 16 840 1 113730 2 */
+438,	/* OBJ_pilotAttributeType           0 9 2342 19200300 100 1 */
+439,	/* OBJ_pilotAttributeSyntax         0 9 2342 19200300 100 3 */
+440,	/* OBJ_pilotObjectClass             0 9 2342 19200300 100 4 */
+441,	/* OBJ_pilotGroups                  0 9 2342 19200300 100 10 */
+108,	/* OBJ_cast5_cbc                    1 2 840 113533 7 66 10 */
+112,	/* OBJ_pbeWithMD5AndCast5_CBC       1 2 840 113533 7 66 12 */
+782,	/* OBJ_id_PasswordBasedMAC          1 2 840 113533 7 66 13 */
+783,	/* OBJ_id_DHBasedMac                1 2 840 113533 7 66 30 */
+ 6,	/* OBJ_rsaEncryption                1 2 840 113549 1 1 1 */
+ 7,	/* OBJ_md2WithRSAEncryption         1 2 840 113549 1 1 2 */
+396,	/* OBJ_md4WithRSAEncryption         1 2 840 113549 1 1 3 */
+ 8,	/* OBJ_md5WithRSAEncryption         1 2 840 113549 1 1 4 */
+65,	/* OBJ_sha1WithRSAEncryption        1 2 840 113549 1 1 5 */
+644,	/* OBJ_rsaOAEPEncryptionSET         1 2 840 113549 1 1 6 */
+919,	/* OBJ_rsaesOaep                    1 2 840 113549 1 1 7 */
+911,	/* OBJ_mgf1                         1 2 840 113549 1 1 8 */
+912,	/* OBJ_rsassaPss                    1 2 840 113549 1 1 10 */
+668,	/* OBJ_sha256WithRSAEncryption      1 2 840 113549 1 1 11 */
+669,	/* OBJ_sha384WithRSAEncryption      1 2 840 113549 1 1 12 */
+670,	/* OBJ_sha512WithRSAEncryption      1 2 840 113549 1 1 13 */
+671,	/* OBJ_sha224WithRSAEncryption      1 2 840 113549 1 1 14 */
+28,	/* OBJ_dhKeyAgreement               1 2 840 113549 1 3 1 */
+ 9,	/* OBJ_pbeWithMD2AndDES_CBC         1 2 840 113549 1 5 1 */
+10,	/* OBJ_pbeWithMD5AndDES_CBC         1 2 840 113549 1 5 3 */
+168,	/* OBJ_pbeWithMD2AndRC2_CBC         1 2 840 113549 1 5 4 */
+169,	/* OBJ_pbeWithMD5AndRC2_CBC         1 2 840 113549 1 5 6 */
+170,	/* OBJ_pbeWithSHA1AndDES_CBC        1 2 840 113549 1 5 10 */
+68,	/* OBJ_pbeWithSHA1AndRC2_CBC        1 2 840 113549 1 5 11 */
+69,	/* OBJ_id_pbkdf2                    1 2 840 113549 1 5 12 */
+161,	/* OBJ_pbes2                        1 2 840 113549 1 5 13 */
+162,	/* OBJ_pbmac1                       1 2 840 113549 1 5 14 */
+21,	/* OBJ_pkcs7_data                   1 2 840 113549 1 7 1 */
+22,	/* OBJ_pkcs7_signed                 1 2 840 113549 1 7 2 */
+23,	/* OBJ_pkcs7_enveloped              1 2 840 113549 1 7 3 */
+24,	/* OBJ_pkcs7_signedAndEnveloped     1 2 840 113549 1 7 4 */
+25,	/* OBJ_pkcs7_digest                 1 2 840 113549 1 7 5 */
+26,	/* OBJ_pkcs7_encrypted              1 2 840 113549 1 7 6 */
+48,	/* OBJ_pkcs9_emailAddress           1 2 840 113549 1 9 1 */
+49,	/* OBJ_pkcs9_unstructuredName       1 2 840 113549 1 9 2 */
+50,	/* OBJ_pkcs9_contentType            1 2 840 113549 1 9 3 */
+51,	/* OBJ_pkcs9_messageDigest          1 2 840 113549 1 9 4 */
+52,	/* OBJ_pkcs9_signingTime            1 2 840 113549 1 9 5 */
+53,	/* OBJ_pkcs9_countersignature       1 2 840 113549 1 9 6 */
+54,	/* OBJ_pkcs9_challengePassword      1 2 840 113549 1 9 7 */
+55,	/* OBJ_pkcs9_unstructuredAddress    1 2 840 113549 1 9 8 */
+56,	/* OBJ_pkcs9_extCertAttributes      1 2 840 113549 1 9 9 */
+172,	/* OBJ_ext_req                      1 2 840 113549 1 9 14 */
+167,	/* OBJ_SMIMECapabilities            1 2 840 113549 1 9 15 */
+188,	/* OBJ_SMIME                        1 2 840 113549 1 9 16 */
+156,	/* OBJ_friendlyName                 1 2 840 113549 1 9 20 */
+157,	/* OBJ_localKeyID                   1 2 840 113549 1 9 21 */
+681,	/* OBJ_X9_62_onBasis                1 2 840 10045 1 2 3 1 */
+682,	/* OBJ_X9_62_tpBasis                1 2 840 10045 1 2 3 2 */
+683,	/* OBJ_X9_62_ppBasis                1 2 840 10045 1 2 3 3 */
+417,	/* OBJ_ms_csp_name                  1 3 6 1 4 1 311 17 1 */
+856,	/* OBJ_LocalKeySet                  1 3 6 1 4 1 311 17 2 */
+390,	/* OBJ_dcObject                     1 3 6 1 4 1 1466 344 */
+91,	/* OBJ_bf_cbc                       1 3 6 1 4 1 3029 1 2 */
+315,	/* OBJ_id_regCtrl_regToken          1 3 6 1 5 5 7 5 1 1 */
+316,	/* OBJ_id_regCtrl_authenticator     1 3 6 1 5 5 7 5 1 2 */
+317,	/* OBJ_id_regCtrl_pkiPublicationInfo 1 3 6 1 5 5 7 5 1 3 */
+318,	/* OBJ_id_regCtrl_pkiArchiveOptions 1 3 6 1 5 5 7 5 1 4 */
+319,	/* OBJ_id_regCtrl_oldCertID         1 3 6 1 5 5 7 5 1 5 */
+320,	/* OBJ_id_regCtrl_protocolEncrKey   1 3 6 1 5 5 7 5 1 6 */
+321,	/* OBJ_id_regInfo_utf8Pairs         1 3 6 1 5 5 7 5 2 1 */
+322,	/* OBJ_id_regInfo_certReq           1 3 6 1 5 5 7 5 2 2 */
+365,	/* OBJ_id_pkix_OCSP_basic           1 3 6 1 5 5 7 48 1 1 */
+366,	/* OBJ_id_pkix_OCSP_Nonce           1 3 6 1 5 5 7 48 1 2 */
+367,	/* OBJ_id_pkix_OCSP_CrlID           1 3 6 1 5 5 7 48 1 3 */
+368,	/* OBJ_id_pkix_OCSP_acceptableResponses 1 3 6 1 5 5 7 48 1 4 */
+369,	/* OBJ_id_pkix_OCSP_noCheck         1 3 6 1 5 5 7 48 1 5 */
+370,	/* OBJ_id_pkix_OCSP_archiveCutoff   1 3 6 1 5 5 7 48 1 6 */
+371,	/* OBJ_id_pkix_OCSP_serviceLocator  1 3 6 1 5 5 7 48 1 7 */
+372,	/* OBJ_id_pkix_OCSP_extendedStatus  1 3 6 1 5 5 7 48 1 8 */
+373,	/* OBJ_id_pkix_OCSP_valid           1 3 6 1 5 5 7 48 1 9 */
+374,	/* OBJ_id_pkix_OCSP_path            1 3 6 1 5 5 7 48 1 10 */
+375,	/* OBJ_id_pkix_OCSP_trustRoot       1 3 6 1 5 5 7 48 1 11 */
+418,	/* OBJ_aes_128_ecb                  2 16 840 1 101 3 4 1 1 */
+419,	/* OBJ_aes_128_cbc                  2 16 840 1 101 3 4 1 2 */
+420,	/* OBJ_aes_128_ofb128               2 16 840 1 101 3 4 1 3 */
+421,	/* OBJ_aes_128_cfb128               2 16 840 1 101 3 4 1 4 */
+788,	/* OBJ_id_aes128_wrap               2 16 840 1 101 3 4 1 5 */
+895,	/* OBJ_aes_128_gcm                  2 16 840 1 101 3 4 1 6 */
+896,	/* OBJ_aes_128_ccm                  2 16 840 1 101 3 4 1 7 */
+897,	/* OBJ_id_aes128_wrap_pad           2 16 840 1 101 3 4 1 8 */
+422,	/* OBJ_aes_192_ecb                  2 16 840 1 101 3 4 1 21 */
+423,	/* OBJ_aes_192_cbc                  2 16 840 1 101 3 4 1 22 */
+424,	/* OBJ_aes_192_ofb128               2 16 840 1 101 3 4 1 23 */
+425,	/* OBJ_aes_192_cfb128               2 16 840 1 101 3 4 1 24 */
+789,	/* OBJ_id_aes192_wrap               2 16 840 1 101 3 4 1 25 */
+898,	/* OBJ_aes_192_gcm                  2 16 840 1 101 3 4 1 26 */
+899,	/* OBJ_aes_192_ccm                  2 16 840 1 101 3 4 1 27 */
+900,	/* OBJ_id_aes192_wrap_pad           2 16 840 1 101 3 4 1 28 */
+426,	/* OBJ_aes_256_ecb                  2 16 840 1 101 3 4 1 41 */
+427,	/* OBJ_aes_256_cbc                  2 16 840 1 101 3 4 1 42 */
+428,	/* OBJ_aes_256_ofb128               2 16 840 1 101 3 4 1 43 */
+429,	/* OBJ_aes_256_cfb128               2 16 840 1 101 3 4 1 44 */
+790,	/* OBJ_id_aes256_wrap               2 16 840 1 101 3 4 1 45 */
+901,	/* OBJ_aes_256_gcm                  2 16 840 1 101 3 4 1 46 */
+902,	/* OBJ_aes_256_ccm                  2 16 840 1 101 3 4 1 47 */
+903,	/* OBJ_id_aes256_wrap_pad           2 16 840 1 101 3 4 1 48 */
+672,	/* OBJ_sha256                       2 16 840 1 101 3 4 2 1 */
+673,	/* OBJ_sha384                       2 16 840 1 101 3 4 2 2 */
+674,	/* OBJ_sha512                       2 16 840 1 101 3 4 2 3 */
+675,	/* OBJ_sha224                       2 16 840 1 101 3 4 2 4 */
+802,	/* OBJ_dsa_with_SHA224              2 16 840 1 101 3 4 3 1 */
+803,	/* OBJ_dsa_with_SHA256              2 16 840 1 101 3 4 3 2 */
+71,	/* OBJ_netscape_cert_type           2 16 840 1 113730 1 1 */
+72,	/* OBJ_netscape_base_url            2 16 840 1 113730 1 2 */
+73,	/* OBJ_netscape_revocation_url      2 16 840 1 113730 1 3 */
+74,	/* OBJ_netscape_ca_revocation_url   2 16 840 1 113730 1 4 */
+75,	/* OBJ_netscape_renewal_url         2 16 840 1 113730 1 7 */
+76,	/* OBJ_netscape_ca_policy_url       2 16 840 1 113730 1 8 */
+77,	/* OBJ_netscape_ssl_server_name     2 16 840 1 113730 1 12 */
+78,	/* OBJ_netscape_comment             2 16 840 1 113730 1 13 */
+79,	/* OBJ_netscape_cert_sequence       2 16 840 1 113730 2 5 */
+139,	/* OBJ_ns_sgc                       2 16 840 1 113730 4 1 */
+458,	/* OBJ_userId                       0 9 2342 19200300 100 1 1 */
+459,	/* OBJ_textEncodedORAddress         0 9 2342 19200300 100 1 2 */
+460,	/* OBJ_rfc822Mailbox                0 9 2342 19200300 100 1 3 */
+461,	/* OBJ_info                         0 9 2342 19200300 100 1 4 */
+462,	/* OBJ_favouriteDrink               0 9 2342 19200300 100 1 5 */
+463,	/* OBJ_roomNumber                   0 9 2342 19200300 100 1 6 */
+464,	/* OBJ_photo                        0 9 2342 19200300 100 1 7 */
+465,	/* OBJ_userClass                    0 9 2342 19200300 100 1 8 */
+466,	/* OBJ_host                         0 9 2342 19200300 100 1 9 */
+467,	/* OBJ_manager                      0 9 2342 19200300 100 1 10 */
+468,	/* OBJ_documentIdentifier           0 9 2342 19200300 100 1 11 */
+469,	/* OBJ_documentTitle                0 9 2342 19200300 100 1 12 */
+470,	/* OBJ_documentVersion              0 9 2342 19200300 100 1 13 */
+471,	/* OBJ_documentAuthor               0 9 2342 19200300 100 1 14 */
+472,	/* OBJ_documentLocation             0 9 2342 19200300 100 1 15 */
+473,	/* OBJ_homeTelephoneNumber          0 9 2342 19200300 100 1 20 */
+474,	/* OBJ_secretary                    0 9 2342 19200300 100 1 21 */
+475,	/* OBJ_otherMailbox                 0 9 2342 19200300 100 1 22 */
+476,	/* OBJ_lastModifiedTime             0 9 2342 19200300 100 1 23 */
+477,	/* OBJ_lastModifiedBy               0 9 2342 19200300 100 1 24 */
+391,	/* OBJ_domainComponent              0 9 2342 19200300 100 1 25 */
+478,	/* OBJ_aRecord                      0 9 2342 19200300 100 1 26 */
+479,	/* OBJ_pilotAttributeType27         0 9 2342 19200300 100 1 27 */
+480,	/* OBJ_mXRecord                     0 9 2342 19200300 100 1 28 */
+481,	/* OBJ_nSRecord                     0 9 2342 19200300 100 1 29 */
+482,	/* OBJ_sOARecord                    0 9 2342 19200300 100 1 30 */
+483,	/* OBJ_cNAMERecord                  0 9 2342 19200300 100 1 31 */
+484,	/* OBJ_associatedDomain             0 9 2342 19200300 100 1 37 */
+485,	/* OBJ_associatedName               0 9 2342 19200300 100 1 38 */
+486,	/* OBJ_homePostalAddress            0 9 2342 19200300 100 1 39 */
+487,	/* OBJ_personalTitle                0 9 2342 19200300 100 1 40 */
+488,	/* OBJ_mobileTelephoneNumber        0 9 2342 19200300 100 1 41 */
+489,	/* OBJ_pagerTelephoneNumber         0 9 2342 19200300 100 1 42 */
+490,	/* OBJ_friendlyCountryName          0 9 2342 19200300 100 1 43 */
+491,	/* OBJ_organizationalStatus         0 9 2342 19200300 100 1 45 */
+492,	/* OBJ_janetMailbox                 0 9 2342 19200300 100 1 46 */
+493,	/* OBJ_mailPreferenceOption         0 9 2342 19200300 100 1 47 */
+494,	/* OBJ_buildingName                 0 9 2342 19200300 100 1 48 */
+495,	/* OBJ_dSAQuality                   0 9 2342 19200300 100 1 49 */
+496,	/* OBJ_singleLevelQuality           0 9 2342 19200300 100 1 50 */
+497,	/* OBJ_subtreeMinimumQuality        0 9 2342 19200300 100 1 51 */
+498,	/* OBJ_subtreeMaximumQuality        0 9 2342 19200300 100 1 52 */
+499,	/* OBJ_personalSignature            0 9 2342 19200300 100 1 53 */
+500,	/* OBJ_dITRedirect                  0 9 2342 19200300 100 1 54 */
+501,	/* OBJ_audio                        0 9 2342 19200300 100 1 55 */
+502,	/* OBJ_documentPublisher            0 9 2342 19200300 100 1 56 */
+442,	/* OBJ_iA5StringSyntax              0 9 2342 19200300 100 3 4 */
+443,	/* OBJ_caseIgnoreIA5StringSyntax    0 9 2342 19200300 100 3 5 */
+444,	/* OBJ_pilotObject                  0 9 2342 19200300 100 4 3 */
+445,	/* OBJ_pilotPerson                  0 9 2342 19200300 100 4 4 */
+446,	/* OBJ_account                      0 9 2342 19200300 100 4 5 */
+447,	/* OBJ_document                     0 9 2342 19200300 100 4 6 */
+448,	/* OBJ_room                         0 9 2342 19200300 100 4 7 */
+449,	/* OBJ_documentSeries               0 9 2342 19200300 100 4 9 */
+392,	/* OBJ_Domain                       0 9 2342 19200300 100 4 13 */
+450,	/* OBJ_rFC822localPart              0 9 2342 19200300 100 4 14 */
+451,	/* OBJ_dNSDomain                    0 9 2342 19200300 100 4 15 */
+452,	/* OBJ_domainRelatedObject          0 9 2342 19200300 100 4 17 */
+453,	/* OBJ_friendlyCountry              0 9 2342 19200300 100 4 18 */
+454,	/* OBJ_simpleSecurityObject         0 9 2342 19200300 100 4 19 */
+455,	/* OBJ_pilotOrganization            0 9 2342 19200300 100 4 20 */
+456,	/* OBJ_pilotDSA                     0 9 2342 19200300 100 4 21 */
+457,	/* OBJ_qualityLabelledData          0 9 2342 19200300 100 4 22 */
+189,	/* OBJ_id_smime_mod                 1 2 840 113549 1 9 16 0 */
+190,	/* OBJ_id_smime_ct                  1 2 840 113549 1 9 16 1 */
+191,	/* OBJ_id_smime_aa                  1 2 840 113549 1 9 16 2 */
+192,	/* OBJ_id_smime_alg                 1 2 840 113549 1 9 16 3 */
+193,	/* OBJ_id_smime_cd                  1 2 840 113549 1 9 16 4 */
+194,	/* OBJ_id_smime_spq                 1 2 840 113549 1 9 16 5 */
+195,	/* OBJ_id_smime_cti                 1 2 840 113549 1 9 16 6 */
+158,	/* OBJ_x509Certificate              1 2 840 113549 1 9 22 1 */
+159,	/* OBJ_sdsiCertificate              1 2 840 113549 1 9 22 2 */
+160,	/* OBJ_x509Crl                      1 2 840 113549 1 9 23 1 */
+144,	/* OBJ_pbe_WithSHA1And128BitRC4     1 2 840 113549 1 12 1 1 */
+145,	/* OBJ_pbe_WithSHA1And40BitRC4      1 2 840 113549 1 12 1 2 */
+146,	/* OBJ_pbe_WithSHA1And3_Key_TripleDES_CBC 1 2 840 113549 1 12 1 3 */
+147,	/* OBJ_pbe_WithSHA1And2_Key_TripleDES_CBC 1 2 840 113549 1 12 1 4 */
+148,	/* OBJ_pbe_WithSHA1And128BitRC2_CBC 1 2 840 113549 1 12 1 5 */
+149,	/* OBJ_pbe_WithSHA1And40BitRC2_CBC  1 2 840 113549 1 12 1 6 */
+171,	/* OBJ_ms_ext_req                   1 3 6 1 4 1 311 2 1 14 */
+134,	/* OBJ_ms_code_ind                  1 3 6 1 4 1 311 2 1 21 */
+135,	/* OBJ_ms_code_com                  1 3 6 1 4 1 311 2 1 22 */
+136,	/* OBJ_ms_ctl_sign                  1 3 6 1 4 1 311 10 3 1 */
+137,	/* OBJ_ms_sgc                       1 3 6 1 4 1 311 10 3 3 */
+138,	/* OBJ_ms_efs                       1 3 6 1 4 1 311 10 3 4 */
+648,	/* OBJ_ms_smartcard_login           1 3 6 1 4 1 311 20 2 2 */
+649,	/* OBJ_ms_upn                       1 3 6 1 4 1 311 20 2 3 */
+751,	/* OBJ_camellia_128_cbc             1 2 392 200011 61 1 1 1 2 */
+752,	/* OBJ_camellia_192_cbc             1 2 392 200011 61 1 1 1 3 */
+753,	/* OBJ_camellia_256_cbc             1 2 392 200011 61 1 1 1 4 */
+907,	/* OBJ_id_camellia128_wrap          1 2 392 200011 61 1 1 3 2 */
+908,	/* OBJ_id_camellia192_wrap          1 2 392 200011 61 1 1 3 3 */
+909,	/* OBJ_id_camellia256_wrap          1 2 392 200011 61 1 1 3 4 */
+196,	/* OBJ_id_smime_mod_cms             1 2 840 113549 1 9 16 0 1 */
+197,	/* OBJ_id_smime_mod_ess             1 2 840 113549 1 9 16 0 2 */
+198,	/* OBJ_id_smime_mod_oid             1 2 840 113549 1 9 16 0 3 */
+199,	/* OBJ_id_smime_mod_msg_v3          1 2 840 113549 1 9 16 0 4 */
+200,	/* OBJ_id_smime_mod_ets_eSignature_88 1 2 840 113549 1 9 16 0 5 */
+201,	/* OBJ_id_smime_mod_ets_eSignature_97 1 2 840 113549 1 9 16 0 6 */
+202,	/* OBJ_id_smime_mod_ets_eSigPolicy_88 1 2 840 113549 1 9 16 0 7 */
+203,	/* OBJ_id_smime_mod_ets_eSigPolicy_97 1 2 840 113549 1 9 16 0 8 */
+204,	/* OBJ_id_smime_ct_receipt          1 2 840 113549 1 9 16 1 1 */
+205,	/* OBJ_id_smime_ct_authData         1 2 840 113549 1 9 16 1 2 */
+206,	/* OBJ_id_smime_ct_publishCert      1 2 840 113549 1 9 16 1 3 */
+207,	/* OBJ_id_smime_ct_TSTInfo          1 2 840 113549 1 9 16 1 4 */
+208,	/* OBJ_id_smime_ct_TDTInfo          1 2 840 113549 1 9 16 1 5 */
+209,	/* OBJ_id_smime_ct_contentInfo      1 2 840 113549 1 9 16 1 6 */
+210,	/* OBJ_id_smime_ct_DVCSRequestData  1 2 840 113549 1 9 16 1 7 */
+211,	/* OBJ_id_smime_ct_DVCSResponseData 1 2 840 113549 1 9 16 1 8 */
+786,	/* OBJ_id_smime_ct_compressedData   1 2 840 113549 1 9 16 1 9 */
+787,	/* OBJ_id_ct_asciiTextWithCRLF      1 2 840 113549 1 9 16 1 27 */
+212,	/* OBJ_id_smime_aa_receiptRequest   1 2 840 113549 1 9 16 2 1 */
+213,	/* OBJ_id_smime_aa_securityLabel    1 2 840 113549 1 9 16 2 2 */
+214,	/* OBJ_id_smime_aa_mlExpandHistory  1 2 840 113549 1 9 16 2 3 */
+215,	/* OBJ_id_smime_aa_contentHint      1 2 840 113549 1 9 16 2 4 */
+216,	/* OBJ_id_smime_aa_msgSigDigest     1 2 840 113549 1 9 16 2 5 */
+217,	/* OBJ_id_smime_aa_encapContentType 1 2 840 113549 1 9 16 2 6 */
+218,	/* OBJ_id_smime_aa_contentIdentifier 1 2 840 113549 1 9 16 2 7 */
+219,	/* OBJ_id_smime_aa_macValue         1 2 840 113549 1 9 16 2 8 */
+220,	/* OBJ_id_smime_aa_equivalentLabels 1 2 840 113549 1 9 16 2 9 */
+221,	/* OBJ_id_smime_aa_contentReference 1 2 840 113549 1 9 16 2 10 */
+222,	/* OBJ_id_smime_aa_encrypKeyPref    1 2 840 113549 1 9 16 2 11 */
+223,	/* OBJ_id_smime_aa_signingCertificate 1 2 840 113549 1 9 16 2 12 */
+224,	/* OBJ_id_smime_aa_smimeEncryptCerts 1 2 840 113549 1 9 16 2 13 */
+225,	/* OBJ_id_smime_aa_timeStampToken   1 2 840 113549 1 9 16 2 14 */
+226,	/* OBJ_id_smime_aa_ets_sigPolicyId  1 2 840 113549 1 9 16 2 15 */
+227,	/* OBJ_id_smime_aa_ets_commitmentType 1 2 840 113549 1 9 16 2 16 */
+228,	/* OBJ_id_smime_aa_ets_signerLocation 1 2 840 113549 1 9 16 2 17 */
+229,	/* OBJ_id_smime_aa_ets_signerAttr   1 2 840 113549 1 9 16 2 18 */
+230,	/* OBJ_id_smime_aa_ets_otherSigCert 1 2 840 113549 1 9 16 2 19 */
+231,	/* OBJ_id_smime_aa_ets_contentTimestamp 1 2 840 113549 1 9 16 2 20 */
+232,	/* OBJ_id_smime_aa_ets_CertificateRefs 1 2 840 113549 1 9 16 2 21 */
+233,	/* OBJ_id_smime_aa_ets_RevocationRefs 1 2 840 113549 1 9 16 2 22 */
+234,	/* OBJ_id_smime_aa_ets_certValues   1 2 840 113549 1 9 16 2 23 */
+235,	/* OBJ_id_smime_aa_ets_revocationValues 1 2 840 113549 1 9 16 2 24 */
+236,	/* OBJ_id_smime_aa_ets_escTimeStamp 1 2 840 113549 1 9 16 2 25 */
+237,	/* OBJ_id_smime_aa_ets_certCRLTimestamp 1 2 840 113549 1 9 16 2 26 */
+238,	/* OBJ_id_smime_aa_ets_archiveTimeStamp 1 2 840 113549 1 9 16 2 27 */
+239,	/* OBJ_id_smime_aa_signatureType    1 2 840 113549 1 9 16 2 28 */
+240,	/* OBJ_id_smime_aa_dvcs_dvc         1 2 840 113549 1 9 16 2 29 */
+241,	/* OBJ_id_smime_alg_ESDHwith3DES    1 2 840 113549 1 9 16 3 1 */
+242,	/* OBJ_id_smime_alg_ESDHwithRC2     1 2 840 113549 1 9 16 3 2 */
+243,	/* OBJ_id_smime_alg_3DESwrap        1 2 840 113549 1 9 16 3 3 */
+244,	/* OBJ_id_smime_alg_RC2wrap         1 2 840 113549 1 9 16 3 4 */
+245,	/* OBJ_id_smime_alg_ESDH            1 2 840 113549 1 9 16 3 5 */
+246,	/* OBJ_id_smime_alg_CMS3DESwrap     1 2 840 113549 1 9 16 3 6 */
+247,	/* OBJ_id_smime_alg_CMSRC2wrap      1 2 840 113549 1 9 16 3 7 */
+125,	/* OBJ_zlib_compression             1 2 840 113549 1 9 16 3 8 */
+893,	/* OBJ_id_alg_PWRI_KEK              1 2 840 113549 1 9 16 3 9 */
+248,	/* OBJ_id_smime_cd_ldap             1 2 840 113549 1 9 16 4 1 */
+249,	/* OBJ_id_smime_spq_ets_sqt_uri     1 2 840 113549 1 9 16 5 1 */
+250,	/* OBJ_id_smime_spq_ets_sqt_unotice 1 2 840 113549 1 9 16 5 2 */
+251,	/* OBJ_id_smime_cti_ets_proofOfOrigin 1 2 840 113549 1 9 16 6 1 */
+252,	/* OBJ_id_smime_cti_ets_proofOfReceipt 1 2 840 113549 1 9 16 6 2 */
+253,	/* OBJ_id_smime_cti_ets_proofOfDelivery 1 2 840 113549 1 9 16 6 3 */
+254,	/* OBJ_id_smime_cti_ets_proofOfSender 1 2 840 113549 1 9 16 6 4 */
+255,	/* OBJ_id_smime_cti_ets_proofOfApproval 1 2 840 113549 1 9 16 6 5 */
+256,	/* OBJ_id_smime_cti_ets_proofOfCreation 1 2 840 113549 1 9 16 6 6 */
+150,	/* OBJ_keyBag                       1 2 840 113549 1 12 10 1 1 */
+151,	/* OBJ_pkcs8ShroudedKeyBag          1 2 840 113549 1 12 10 1 2 */
+152,	/* OBJ_certBag                      1 2 840 113549 1 12 10 1 3 */
+153,	/* OBJ_crlBag                       1 2 840 113549 1 12 10 1 4 */
+154,	/* OBJ_secretBag                    1 2 840 113549 1 12 10 1 5 */
+155,	/* OBJ_safeContentsBag              1 2 840 113549 1 12 10 1 6 */
+34,	/* OBJ_idea_cbc                     1 3 6 1 4 1 188 7 1 1 2 */
 };
 
diff --git a/crypto/objects/obj_dat.pl b/crypto/objects/obj_dat.pl
index 7de2f77afdae..c67f71c3277b 100644
--- a/crypto/objects/obj_dat.pl
+++ b/crypto/objects/obj_dat.pl
@@ -2,9 +2,7 @@
 
 # fixes bug in floating point emulation on sparc64 when
 # this script produces off-by-one output on sparc64
-eval 'use integer;';
-
-print STDERR "Warning: perl module integer not found.\n" if ($@);
+use integer;
 
 sub obj_cmp
 	{
@@ -150,13 +148,13 @@ for ($i=0; $i<$n; $i++)
 @a=grep(defined($sn{$nid{$_}}),0 .. $n);
 foreach (sort { $sn{$nid{$a}} cmp $sn{$nid{$b}} } @a)
 	{
-	push(@sn,sprintf("&(nid_objs[%2d]),/* \"$sn{$nid{$_}}\" */\n",$_));
+	push(@sn,sprintf("%2d,\t/* \"$sn{$nid{$_}}\" */\n",$_));
 	}
 
 @a=grep(defined($ln{$nid{$_}}),0 .. $n);
 foreach (sort { $ln{$nid{$a}} cmp $ln{$nid{$b}} } @a)
 	{
-	push(@ln,sprintf("&(nid_objs[%2d]),/* \"$ln{$nid{$_}}\" */\n",$_));
+	push(@ln,sprintf("%2d,\t/* \"$ln{$nid{$_}}\" */\n",$_));
 	}
 
 @a=grep(defined($obj{$nid{$_}}),0 .. $n);
@@ -166,7 +164,7 @@ foreach (sort obj_cmp @a)
 	$v=$objd{$m};
 	$v =~ s/L//g;
 	$v =~ s/,/ /g;
-	push(@ob,sprintf("&(nid_objs[%2d]),/* %-32s %s */\n",$_,$m,$v));
+	push(@ob,sprintf("%2d,\t/* %-32s %s */\n",$_,$m,$v));
 	}
 
 print OUT <<'EOF';
@@ -241,11 +239,11 @@ printf OUT "#define NUM_SN %d\n",$#sn+1;
 printf OUT "#define NUM_LN %d\n",$#ln+1;
 printf OUT "#define NUM_OBJ %d\n\n",$#ob+1;
 
-printf OUT "static unsigned char lvalues[%d]={\n",$lvalues+1;
+printf OUT "static const unsigned char lvalues[%d]={\n",$lvalues+1;
 print OUT @lvalues;
 print OUT "};\n\n";
 
-printf OUT "static ASN1_OBJECT nid_objs[NUM_NID]={\n";
+printf OUT "static const ASN1_OBJECT nid_objs[NUM_NID]={\n";
 foreach (@out)
 	{
 	if (length($_) > 75)
@@ -269,15 +267,15 @@ foreach (@out)
 	}
 print  OUT "};\n\n";
 
-printf OUT "static ASN1_OBJECT *sn_objs[NUM_SN]={\n";
+printf OUT "static const unsigned int sn_objs[NUM_SN]={\n";
 print  OUT @sn;
 print  OUT "};\n\n";
 
-printf OUT "static ASN1_OBJECT *ln_objs[NUM_LN]={\n";
+printf OUT "static const unsigned int ln_objs[NUM_LN]={\n";
 print  OUT @ln;
 print  OUT "};\n\n";
 
-printf OUT "static ASN1_OBJECT *obj_objs[NUM_OBJ]={\n";
+printf OUT "static const unsigned int obj_objs[NUM_OBJ]={\n";
 print  OUT @ob;
 print  OUT "};\n\n";
 
diff --git a/crypto/objects/obj_err.c b/crypto/objects/obj_err.c
index 12b48850c6bc..2e7a034c3fd5 100644
--- a/crypto/objects/obj_err.c
+++ b/crypto/objects/obj_err.c
@@ -1,6 +1,6 @@
 /* crypto/objects/obj_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
diff --git a/crypto/objects/obj_lib.c b/crypto/objects/obj_lib.c
index 706fa0b0e78c..23e9d48cdf3d 100644
--- a/crypto/objects/obj_lib.c
+++ b/crypto/objects/obj_lib.c
@@ -66,7 +66,8 @@ ASN1_OBJECT *OBJ_dup(const ASN1_OBJECT *o)
 	{
 	ASN1_OBJECT *r;
 	int i;
-	char *ln=NULL;
+	char *ln=NULL,*sn=NULL;
+	unsigned char *data=NULL;
 
 	if (o == NULL) return(NULL);
 	if (!(o->flags & ASN1_OBJECT_FLAG_DYNAMIC))
@@ -79,42 +80,42 @@ ASN1_OBJECT *OBJ_dup(const ASN1_OBJECT *o)
 		OBJerr(OBJ_F_OBJ_DUP,ERR_R_ASN1_LIB);
 		return(NULL);
 		}
-	r->data=OPENSSL_malloc(o->length);
-	if (r->data == NULL)
+	data=OPENSSL_malloc(o->length);
+	if (data == NULL)
 		goto err;
 	if (o->data != NULL)
-		memcpy(r->data,o->data,o->length);
+		memcpy(data,o->data,o->length);
+	/* once data attached to object it remains const */
+	r->data = data;
 	r->length=o->length;
 	r->nid=o->nid;
 	r->ln=r->sn=NULL;
 	if (o->ln != NULL)
 		{
 		i=strlen(o->ln)+1;
-		r->ln=ln=OPENSSL_malloc(i);
-		if (r->ln == NULL) goto err;
+		ln=OPENSSL_malloc(i);
+		if (ln == NULL) goto err;
 		memcpy(ln,o->ln,i);
+		r->ln=ln;
 		}
 
 	if (o->sn != NULL)
 		{
-		char *s;
-
 		i=strlen(o->sn)+1;
-		r->sn=s=OPENSSL_malloc(i);
-		if (r->sn == NULL) goto err;
-		memcpy(s,o->sn,i);
+		sn=OPENSSL_malloc(i);
+		if (sn == NULL) goto err;
+		memcpy(sn,o->sn,i);
+		r->sn=sn;
 		}
 	r->flags=o->flags|(ASN1_OBJECT_FLAG_DYNAMIC|
 		ASN1_OBJECT_FLAG_DYNAMIC_STRINGS|ASN1_OBJECT_FLAG_DYNAMIC_DATA);
 	return(r);
 err:
 	OBJerr(OBJ_F_OBJ_DUP,ERR_R_MALLOC_FAILURE);
-	if (r != NULL)
-		{
-		if (ln != NULL) OPENSSL_free(ln);
-		if (r->data != NULL) OPENSSL_free(r->data);
-		OPENSSL_free(r);
-		}
+	if (ln != NULL)		OPENSSL_free(ln);
+	if (sn != NULL)		OPENSSL_free(sn);
+	if (data != NULL)	OPENSSL_free(data);
+	if (r != NULL)		OPENSSL_free(r);
 	return(NULL);
 	}
 
diff --git a/crypto/objects/obj_mac.h b/crypto/objects/obj_mac.h
index 282f11a8a88e..b5ea7cdab4f8 100644
--- a/crypto/objects/obj_mac.h
+++ b/crypto/objects/obj_mac.h
@@ -580,6 +580,21 @@
 #define NID_sha1WithRSAEncryption		65
 #define OBJ_sha1WithRSAEncryption		OBJ_pkcs1,5L
 
+#define SN_rsaesOaep		"RSAES-OAEP"
+#define LN_rsaesOaep		"rsaesOaep"
+#define NID_rsaesOaep		919
+#define OBJ_rsaesOaep		OBJ_pkcs1,7L
+
+#define SN_mgf1		"MGF1"
+#define LN_mgf1		"mgf1"
+#define NID_mgf1		911
+#define OBJ_mgf1		OBJ_pkcs1,8L
+
+#define SN_rsassaPss		"RSASSA-PSS"
+#define LN_rsassaPss		"rsassaPss"
+#define NID_rsassaPss		912
+#define OBJ_rsassaPss		OBJ_pkcs1,10L
+
 #define SN_sha256WithRSAEncryption		"RSA-SHA256"
 #define LN_sha256WithRSAEncryption		"sha256WithRSAEncryption"
 #define NID_sha256WithRSAEncryption		668
@@ -981,6 +996,10 @@
 #define NID_id_smime_alg_CMSRC2wrap		247
 #define OBJ_id_smime_alg_CMSRC2wrap		OBJ_id_smime_alg,7L
 
+#define SN_id_alg_PWRI_KEK		"id-alg-PWRI-KEK"
+#define NID_id_alg_PWRI_KEK		893
+#define OBJ_id_alg_PWRI_KEK		OBJ_id_smime_alg,9L
+
 #define SN_id_smime_cd_ldap		"id-smime-cd-ldap"
 #define NID_id_smime_cd_ldap		248
 #define OBJ_id_smime_cd_ldap		OBJ_id_smime_cd,1L
@@ -2399,6 +2418,11 @@
 #define NID_no_rev_avail		403
 #define OBJ_no_rev_avail		OBJ_id_ce,56L
 
+#define SN_anyExtendedKeyUsage		"anyExtendedKeyUsage"
+#define LN_anyExtendedKeyUsage		"Any Extended Key Usage"
+#define NID_anyExtendedKeyUsage		910
+#define OBJ_anyExtendedKeyUsage		OBJ_ext_key_usage,0L
+
 #define SN_netscape		"Netscape"
 #define LN_netscape		"Netscape Communications Corp."
 #define NID_netscape		57
@@ -2586,6 +2610,24 @@
 #define NID_aes_128_cfb128		421
 #define OBJ_aes_128_cfb128		OBJ_aes,4L
 
+#define SN_id_aes128_wrap		"id-aes128-wrap"
+#define NID_id_aes128_wrap		788
+#define OBJ_id_aes128_wrap		OBJ_aes,5L
+
+#define SN_aes_128_gcm		"id-aes128-GCM"
+#define LN_aes_128_gcm		"aes-128-gcm"
+#define NID_aes_128_gcm		895
+#define OBJ_aes_128_gcm		OBJ_aes,6L
+
+#define SN_aes_128_ccm		"id-aes128-CCM"
+#define LN_aes_128_ccm		"aes-128-ccm"
+#define NID_aes_128_ccm		896
+#define OBJ_aes_128_ccm		OBJ_aes,7L
+
+#define SN_id_aes128_wrap_pad		"id-aes128-wrap-pad"
+#define NID_id_aes128_wrap_pad		897
+#define OBJ_id_aes128_wrap_pad		OBJ_aes,8L
+
 #define SN_aes_192_ecb		"AES-192-ECB"
 #define LN_aes_192_ecb		"aes-192-ecb"
 #define NID_aes_192_ecb		422
@@ -2606,6 +2648,24 @@
 #define NID_aes_192_cfb128		425
 #define OBJ_aes_192_cfb128		OBJ_aes,24L
 
+#define SN_id_aes192_wrap		"id-aes192-wrap"
+#define NID_id_aes192_wrap		789
+#define OBJ_id_aes192_wrap		OBJ_aes,25L
+
+#define SN_aes_192_gcm		"id-aes192-GCM"
+#define LN_aes_192_gcm		"aes-192-gcm"
+#define NID_aes_192_gcm		898
+#define OBJ_aes_192_gcm		OBJ_aes,26L
+
+#define SN_aes_192_ccm		"id-aes192-CCM"
+#define LN_aes_192_ccm		"aes-192-ccm"
+#define NID_aes_192_ccm		899
+#define OBJ_aes_192_ccm		OBJ_aes,27L
+
+#define SN_id_aes192_wrap_pad		"id-aes192-wrap-pad"
+#define NID_id_aes192_wrap_pad		900
+#define OBJ_id_aes192_wrap_pad		OBJ_aes,28L
+
 #define SN_aes_256_ecb		"AES-256-ECB"
 #define LN_aes_256_ecb		"aes-256-ecb"
 #define NID_aes_256_ecb		426
@@ -2626,6 +2686,24 @@
 #define NID_aes_256_cfb128		429
 #define OBJ_aes_256_cfb128		OBJ_aes,44L
 
+#define SN_id_aes256_wrap		"id-aes256-wrap"
+#define NID_id_aes256_wrap		790
+#define OBJ_id_aes256_wrap		OBJ_aes,45L
+
+#define SN_aes_256_gcm		"id-aes256-GCM"
+#define LN_aes_256_gcm		"aes-256-gcm"
+#define NID_aes_256_gcm		901
+#define OBJ_aes_256_gcm		OBJ_aes,46L
+
+#define SN_aes_256_ccm		"id-aes256-CCM"
+#define LN_aes_256_ccm		"aes-256-ccm"
+#define NID_aes_256_ccm		902
+#define OBJ_aes_256_ccm		OBJ_aes,47L
+
+#define SN_id_aes256_wrap_pad		"id-aes256-wrap-pad"
+#define NID_id_aes256_wrap_pad		903
+#define OBJ_id_aes256_wrap_pad		OBJ_aes,48L
+
 #define SN_aes_128_cfb1		"AES-128-CFB1"
 #define LN_aes_128_cfb1		"aes-128-cfb1"
 #define NID_aes_128_cfb1		650
@@ -2650,6 +2728,26 @@
 #define LN_aes_256_cfb8		"aes-256-cfb8"
 #define NID_aes_256_cfb8		655
 
+#define SN_aes_128_ctr		"AES-128-CTR"
+#define LN_aes_128_ctr		"aes-128-ctr"
+#define NID_aes_128_ctr		904
+
+#define SN_aes_192_ctr		"AES-192-CTR"
+#define LN_aes_192_ctr		"aes-192-ctr"
+#define NID_aes_192_ctr		905
+
+#define SN_aes_256_ctr		"AES-256-CTR"
+#define LN_aes_256_ctr		"aes-256-ctr"
+#define NID_aes_256_ctr		906
+
+#define SN_aes_128_xts		"AES-128-XTS"
+#define LN_aes_128_xts		"aes-128-xts"
+#define NID_aes_128_xts		913
+
+#define SN_aes_256_xts		"AES-256-XTS"
+#define LN_aes_256_xts		"aes-256-xts"
+#define NID_aes_256_xts		914
+
 #define SN_des_cfb1		"DES-CFB1"
 #define LN_des_cfb1		"des-cfb1"
 #define NID_des_cfb1		656
@@ -2666,18 +2764,6 @@
 #define LN_des_ede3_cfb8		"des-ede3-cfb8"
 #define NID_des_ede3_cfb8		659
 
-#define SN_id_aes128_wrap		"id-aes128-wrap"
-#define NID_id_aes128_wrap		788
-#define OBJ_id_aes128_wrap		OBJ_aes,5L
-
-#define SN_id_aes192_wrap		"id-aes192-wrap"
-#define NID_id_aes192_wrap		789
-#define OBJ_id_aes192_wrap		OBJ_aes,25L
-
-#define SN_id_aes256_wrap		"id-aes256-wrap"
-#define NID_id_aes256_wrap		790
-#define OBJ_id_aes256_wrap		OBJ_aes,45L
-
 #define OBJ_nist_hashalgs		OBJ_nistAlgorithms,2L
 
 #define SN_sha256		"SHA256"
@@ -3810,6 +3896,18 @@
 #define NID_camellia_256_cbc		753
 #define OBJ_camellia_256_cbc		1L,2L,392L,200011L,61L,1L,1L,1L,4L
 
+#define SN_id_camellia128_wrap		"id-camellia128-wrap"
+#define NID_id_camellia128_wrap		907
+#define OBJ_id_camellia128_wrap		1L,2L,392L,200011L,61L,1L,1L,3L,2L
+
+#define SN_id_camellia192_wrap		"id-camellia192-wrap"
+#define NID_id_camellia192_wrap		908
+#define OBJ_id_camellia192_wrap		1L,2L,392L,200011L,61L,1L,1L,3L,3L
+
+#define SN_id_camellia256_wrap		"id-camellia256-wrap"
+#define NID_id_camellia256_wrap		909
+#define OBJ_id_camellia256_wrap		1L,2L,392L,200011L,61L,1L,1L,3L,4L
+
 #define OBJ_ntt_ds		0L,3L,4401L,5L
 
 #define OBJ_camellia		OBJ_ntt_ds,3L,1L,9L
@@ -3912,3 +4010,23 @@
 #define LN_hmac		"hmac"
 #define NID_hmac		855
 
+#define SN_cmac		"CMAC"
+#define LN_cmac		"cmac"
+#define NID_cmac		894
+
+#define SN_rc4_hmac_md5		"RC4-HMAC-MD5"
+#define LN_rc4_hmac_md5		"rc4-hmac-md5"
+#define NID_rc4_hmac_md5		915
+
+#define SN_aes_128_cbc_hmac_sha1		"AES-128-CBC-HMAC-SHA1"
+#define LN_aes_128_cbc_hmac_sha1		"aes-128-cbc-hmac-sha1"
+#define NID_aes_128_cbc_hmac_sha1		916
+
+#define SN_aes_192_cbc_hmac_sha1		"AES-192-CBC-HMAC-SHA1"
+#define LN_aes_192_cbc_hmac_sha1		"aes-192-cbc-hmac-sha1"
+#define NID_aes_192_cbc_hmac_sha1		917
+
+#define SN_aes_256_cbc_hmac_sha1		"AES-256-CBC-HMAC-SHA1"
+#define LN_aes_256_cbc_hmac_sha1		"aes-256-cbc-hmac-sha1"
+#define NID_aes_256_cbc_hmac_sha1		918
+
diff --git a/crypto/objects/obj_mac.num b/crypto/objects/obj_mac.num
index 8c50aac27f11..1d0a7c802daf 100644
--- a/crypto/objects/obj_mac.num
+++ b/crypto/objects/obj_mac.num
@@ -890,3 +890,30 @@ houseIdentifier		889
 supportedAlgorithms		890
 deltaRevocationList		891
 dmdName		892
+id_alg_PWRI_KEK		893
+cmac		894
+aes_128_gcm		895
+aes_128_ccm		896
+id_aes128_wrap_pad		897
+aes_192_gcm		898
+aes_192_ccm		899
+id_aes192_wrap_pad		900
+aes_256_gcm		901
+aes_256_ccm		902
+id_aes256_wrap_pad		903
+aes_128_ctr		904
+aes_192_ctr		905
+aes_256_ctr		906
+id_camellia128_wrap		907
+id_camellia192_wrap		908
+id_camellia256_wrap		909
+anyExtendedKeyUsage		910
+mgf1		911
+rsassaPss		912
+aes_128_xts		913
+aes_256_xts		914
+rc4_hmac_md5		915
+aes_128_cbc_hmac_sha1		916
+aes_192_cbc_hmac_sha1		917
+aes_256_cbc_hmac_sha1		918
+rsaesOaep		919
diff --git a/crypto/objects/obj_xref.c b/crypto/objects/obj_xref.c
new file mode 100644
index 000000000000..9f744bcededb
--- /dev/null
+++ b/crypto/objects/obj_xref.c
@@ -0,0 +1,234 @@
+/* crypto/objects/obj_xref.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <openssl/objects.h>
+#include "obj_xref.h"
+
+DECLARE_STACK_OF(nid_triple)
+STACK_OF(nid_triple) *sig_app, *sigx_app;
+
+static int sig_cmp(const nid_triple *a, const nid_triple *b)
+	{
+	return a->sign_id - b->sign_id;
+	}
+
+DECLARE_OBJ_BSEARCH_CMP_FN(nid_triple, nid_triple, sig);
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(nid_triple, nid_triple, sig);
+
+static int sig_sk_cmp(const nid_triple * const *a, const nid_triple * const *b)
+	{
+	return (*a)->sign_id - (*b)->sign_id;
+	}
+
+DECLARE_OBJ_BSEARCH_CMP_FN(const nid_triple *, const nid_triple *, sigx);
+
+static int sigx_cmp(const nid_triple * const *a, const nid_triple * const *b)
+	{
+	int ret;
+	ret = (*a)->hash_id - (*b)->hash_id;
+	if (ret)
+		return ret;
+	return (*a)->pkey_id - (*b)->pkey_id;
+	}
+
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(const nid_triple *, const nid_triple *, sigx);
+
+int OBJ_find_sigid_algs(int signid, int *pdig_nid, int *ppkey_nid)
+	{
+	nid_triple tmp;
+	const nid_triple *rv = NULL;
+	tmp.sign_id = signid;
+
+	if (sig_app)
+		{
+		int idx = sk_nid_triple_find(sig_app, &tmp);
+		if (idx >= 0)
+			rv = sk_nid_triple_value(sig_app, idx);
+		}
+
+#ifndef OBJ_XREF_TEST2
+	if (rv == NULL)
+		{
+		rv = OBJ_bsearch_sig(&tmp, sigoid_srt,
+				 sizeof(sigoid_srt) / sizeof(nid_triple));
+		}
+#endif
+	if (rv == NULL)
+		return 0;
+	if (pdig_nid)
+		*pdig_nid = rv->hash_id;
+	if (ppkey_nid)
+		*ppkey_nid = rv->pkey_id;
+	return 1;
+	}
+
+int OBJ_find_sigid_by_algs(int *psignid, int dig_nid, int pkey_nid)
+	{
+	nid_triple tmp;
+	const nid_triple *t=&tmp;
+	const nid_triple **rv = NULL;
+
+	tmp.hash_id = dig_nid;
+	tmp.pkey_id = pkey_nid;
+
+	if (sigx_app)
+		{
+		int idx = sk_nid_triple_find(sigx_app, &tmp);
+		if (idx >= 0)
+			{
+			t = sk_nid_triple_value(sigx_app, idx);
+			rv = &t;
+			}
+		}
+
+#ifndef OBJ_XREF_TEST2
+	if (rv == NULL)
+		{
+		rv = OBJ_bsearch_sigx(&t, sigoid_srt_xref,
+				 sizeof(sigoid_srt_xref) / sizeof(nid_triple *)
+				 );
+		}
+#endif
+	if (rv == NULL)
+		return 0;
+	if (psignid)
+		*psignid = (*rv)->sign_id;
+	return 1;
+	}
+
+int OBJ_add_sigid(int signid, int dig_id, int pkey_id)
+	{
+	nid_triple *ntr;
+	if (!sig_app)
+		sig_app = sk_nid_triple_new(sig_sk_cmp);
+	if (!sig_app)
+		return 0;
+	if (!sigx_app)
+		sigx_app = sk_nid_triple_new(sigx_cmp);
+	if (!sigx_app)
+		return 0;
+	ntr = OPENSSL_malloc(sizeof(int) * 3);
+	if (!ntr)
+		return 0;
+	ntr->sign_id = signid;
+	ntr->hash_id = dig_id;
+	ntr->pkey_id = pkey_id;
+
+	if (!sk_nid_triple_push(sig_app, ntr))
+		{
+		OPENSSL_free(ntr);
+		return 0;
+		}
+
+	if (!sk_nid_triple_push(sigx_app, ntr))
+		return 0;
+
+	sk_nid_triple_sort(sig_app);
+	sk_nid_triple_sort(sigx_app);
+
+	return 1;
+	}
+
+static void sid_free(nid_triple *tt)
+	{
+	OPENSSL_free(tt);
+	}
+
+void OBJ_sigid_free(void)
+	{
+	if (sig_app)
+		{
+		sk_nid_triple_pop_free(sig_app, sid_free);
+		sig_app = NULL;
+		}
+	if (sigx_app)
+		{
+		sk_nid_triple_free(sigx_app);
+		sigx_app = NULL;
+		}
+	}
+		
+#ifdef OBJ_XREF_TEST
+
+main()
+	{
+	int n1, n2, n3;
+
+	int i, rv;
+#ifdef OBJ_XREF_TEST2
+	for (i = 0; i <	sizeof(sigoid_srt) / sizeof(nid_triple); i++)
+		{
+		OBJ_add_sigid(sigoid_srt[i][0], sigoid_srt[i][1],
+				sigoid_srt[i][2]);
+		}
+#endif
+
+	for (i = 0; i <	sizeof(sigoid_srt) / sizeof(nid_triple); i++)
+		{
+		n1 = sigoid_srt[i][0];
+		rv = OBJ_find_sigid_algs(n1, &n2, &n3);
+		printf("Forward: %d, %s %s %s\n", rv,
+			OBJ_nid2ln(n1), OBJ_nid2ln(n2), OBJ_nid2ln(n3));
+		n1=0;
+		rv = OBJ_find_sigid_by_algs(&n1, n2, n3);
+		printf("Reverse: %d, %s %s %s\n", rv,
+			OBJ_nid2ln(n1), OBJ_nid2ln(n2), OBJ_nid2ln(n3));
+		}
+	}
+	
+#endif
diff --git a/crypto/objects/obj_xref.h b/crypto/objects/obj_xref.h
new file mode 100644
index 000000000000..e23938c29604
--- /dev/null
+++ b/crypto/objects/obj_xref.h
@@ -0,0 +1,77 @@
+/* AUTOGENERATED BY objxref.pl, DO NOT EDIT */
+
+typedef struct
+	{
+	int sign_id;
+	int hash_id;
+	int pkey_id;
+	} nid_triple;
+
+static const nid_triple sigoid_srt[] =
+	{
+	{NID_md2WithRSAEncryption, NID_md2, NID_rsaEncryption},
+	{NID_md5WithRSAEncryption, NID_md5, NID_rsaEncryption},
+	{NID_shaWithRSAEncryption, NID_sha, NID_rsaEncryption},
+	{NID_sha1WithRSAEncryption, NID_sha1, NID_rsaEncryption},
+	{NID_dsaWithSHA, NID_sha, NID_dsa},
+	{NID_dsaWithSHA1_2, NID_sha1, NID_dsa_2},
+	{NID_mdc2WithRSA, NID_mdc2, NID_rsaEncryption},
+	{NID_md5WithRSA, NID_md5, NID_rsa},
+	{NID_dsaWithSHA1, NID_sha1, NID_dsa},
+	{NID_sha1WithRSA, NID_sha1, NID_rsa},
+	{NID_ripemd160WithRSA, NID_ripemd160, NID_rsaEncryption},
+	{NID_md4WithRSAEncryption, NID_md4, NID_rsaEncryption},
+	{NID_ecdsa_with_SHA1, NID_sha1, NID_X9_62_id_ecPublicKey},
+	{NID_sha256WithRSAEncryption, NID_sha256, NID_rsaEncryption},
+	{NID_sha384WithRSAEncryption, NID_sha384, NID_rsaEncryption},
+	{NID_sha512WithRSAEncryption, NID_sha512, NID_rsaEncryption},
+	{NID_sha224WithRSAEncryption, NID_sha224, NID_rsaEncryption},
+	{NID_ecdsa_with_Recommended, NID_undef, NID_X9_62_id_ecPublicKey},
+	{NID_ecdsa_with_Specified, NID_undef, NID_X9_62_id_ecPublicKey},
+	{NID_ecdsa_with_SHA224, NID_sha224, NID_X9_62_id_ecPublicKey},
+	{NID_ecdsa_with_SHA256, NID_sha256, NID_X9_62_id_ecPublicKey},
+	{NID_ecdsa_with_SHA384, NID_sha384, NID_X9_62_id_ecPublicKey},
+	{NID_ecdsa_with_SHA512, NID_sha512, NID_X9_62_id_ecPublicKey},
+	{NID_dsa_with_SHA224, NID_sha224, NID_dsa},
+	{NID_dsa_with_SHA256, NID_sha256, NID_dsa},
+	{NID_id_GostR3411_94_with_GostR3410_2001, NID_id_GostR3411_94, NID_id_GostR3410_2001},
+	{NID_id_GostR3411_94_with_GostR3410_94, NID_id_GostR3411_94, NID_id_GostR3410_94},
+	{NID_id_GostR3411_94_with_GostR3410_94_cc, NID_id_GostR3411_94, NID_id_GostR3410_94_cc},
+	{NID_id_GostR3411_94_with_GostR3410_2001_cc, NID_id_GostR3411_94, NID_id_GostR3410_2001_cc},
+	{NID_rsassaPss, NID_undef, NID_rsaEncryption},
+	};
+
+static const nid_triple * const sigoid_srt_xref[] =
+	{
+	&sigoid_srt[29],
+	&sigoid_srt[17],
+	&sigoid_srt[18],
+	&sigoid_srt[0],
+	&sigoid_srt[1],
+	&sigoid_srt[7],
+	&sigoid_srt[2],
+	&sigoid_srt[4],
+	&sigoid_srt[3],
+	&sigoid_srt[9],
+	&sigoid_srt[5],
+	&sigoid_srt[8],
+	&sigoid_srt[12],
+	&sigoid_srt[6],
+	&sigoid_srt[10],
+	&sigoid_srt[11],
+	&sigoid_srt[13],
+	&sigoid_srt[24],
+	&sigoid_srt[20],
+	&sigoid_srt[14],
+	&sigoid_srt[21],
+	&sigoid_srt[15],
+	&sigoid_srt[22],
+	&sigoid_srt[16],
+	&sigoid_srt[23],
+	&sigoid_srt[19],
+	&sigoid_srt[25],
+	&sigoid_srt[26],
+	&sigoid_srt[27],
+	&sigoid_srt[28],
+	};
+
diff --git a/crypto/objects/obj_xref.txt b/crypto/objects/obj_xref.txt
new file mode 100644
index 000000000000..cb917182ee2f
--- /dev/null
+++ b/crypto/objects/obj_xref.txt
@@ -0,0 +1,46 @@
+# OID cross reference table.
+# Links signatures OIDs to their corresponding public key algorithms
+# and digests.
+
+md2WithRSAEncryption	md2	rsaEncryption
+md5WithRSAEncryption	md5	rsaEncryption
+shaWithRSAEncryption	sha	rsaEncryption
+sha1WithRSAEncryption	sha1	rsaEncryption
+md4WithRSAEncryption	md4	rsaEncryption
+sha256WithRSAEncryption sha256	rsaEncryption
+sha384WithRSAEncryption	sha384	rsaEncryption
+sha512WithRSAEncryption	sha512	rsaEncryption
+sha224WithRSAEncryption	sha224	rsaEncryption
+mdc2WithRSA		mdc2	rsaEncryption
+ripemd160WithRSA	ripemd160 rsaEncryption
+# For PSS the digest algorithm can vary and depends on the included
+# AlgorithmIdentifier. The digest "undef" indicates the public key
+# method should handle this explicitly.
+rsassaPss		undef	rsaEncryption
+
+# Alternative deprecated OIDs. By using the older "rsa" OID this
+# type will be recognized by not normally used.
+
+md5WithRSA		md5	rsa
+sha1WithRSA		sha1	rsa
+
+dsaWithSHA		sha	dsa
+dsaWithSHA1		sha1	dsa
+
+dsaWithSHA1_2		sha1	dsa_2
+
+ecdsa_with_SHA1		sha1	X9_62_id_ecPublicKey
+ecdsa_with_SHA224	sha224	X9_62_id_ecPublicKey
+ecdsa_with_SHA256	sha256	X9_62_id_ecPublicKey
+ecdsa_with_SHA384	sha384	X9_62_id_ecPublicKey
+ecdsa_with_SHA512	sha512	X9_62_id_ecPublicKey
+ecdsa_with_Recommended	undef	X9_62_id_ecPublicKey
+ecdsa_with_Specified	undef	X9_62_id_ecPublicKey
+
+dsa_with_SHA224		sha224	dsa
+dsa_with_SHA256		sha256	dsa
+
+id_GostR3411_94_with_GostR3410_2001	id_GostR3411_94 id_GostR3410_2001
+id_GostR3411_94_with_GostR3410_94	id_GostR3411_94 id_GostR3410_94
+id_GostR3411_94_with_GostR3410_94_cc	id_GostR3411_94 id_GostR3410_94_cc
+id_GostR3411_94_with_GostR3410_2001_cc	id_GostR3411_94 id_GostR3410_2001_cc
diff --git a/crypto/objects/objects.h b/crypto/objects/objects.h
index 7242f76fb0f1..bd0ee52feb22 100644
--- a/crypto/objects/objects.h
+++ b/crypto/objects/objects.h
@@ -1011,10 +1011,91 @@ int		OBJ_txt2nid(const char *s);
 int		OBJ_ln2nid(const char *s);
 int		OBJ_sn2nid(const char *s);
 int		OBJ_cmp(const ASN1_OBJECT *a,const ASN1_OBJECT *b);
-const char *	OBJ_bsearch(const char *key,const char *base,int num,int size,
-	int (*cmp)(const void *, const void *));
-const char *	OBJ_bsearch_ex(const char *key,const char *base,int num,
-	int size, int (*cmp)(const void *, const void *), int flags);
+const void *	OBJ_bsearch_(const void *key,const void *base,int num,int size,
+			     int (*cmp)(const void *, const void *));
+const void *	OBJ_bsearch_ex_(const void *key,const void *base,int num,
+				int size,
+				int (*cmp)(const void *, const void *),
+				int flags);
+
+#define _DECLARE_OBJ_BSEARCH_CMP_FN(scope, type1, type2, nm)	\
+  static int nm##_cmp_BSEARCH_CMP_FN(const void *, const void *); \
+  static int nm##_cmp(type1 const *, type2 const *); \
+  scope type2 * OBJ_bsearch_##nm(type1 *key, type2 const *base, int num)
+
+#define DECLARE_OBJ_BSEARCH_CMP_FN(type1, type2, cmp)	\
+  _DECLARE_OBJ_BSEARCH_CMP_FN(static, type1, type2, cmp)
+#define DECLARE_OBJ_BSEARCH_GLOBAL_CMP_FN(type1, type2, nm)	\
+  type2 * OBJ_bsearch_##nm(type1 *key, type2 const *base, int num)
+
+/*
+ * Unsolved problem: if a type is actually a pointer type, like
+ * nid_triple is, then its impossible to get a const where you need
+ * it. Consider:
+ *
+ * typedef int nid_triple[3];
+ * const void *a_;
+ * const nid_triple const *a = a_;
+ *
+ * The assignement discards a const because what you really want is:
+ *
+ * const int const * const *a = a_;
+ *
+ * But if you do that, you lose the fact that a is an array of 3 ints,
+ * which breaks comparison functions.
+ *
+ * Thus we end up having to cast, sadly, or unpack the
+ * declarations. Or, as I finally did in this case, delcare nid_triple
+ * to be a struct, which it should have been in the first place.
+ *
+ * Ben, August 2008.
+ *
+ * Also, strictly speaking not all types need be const, but handling
+ * the non-constness means a lot of complication, and in practice
+ * comparison routines do always not touch their arguments.
+ */
+
+#define IMPLEMENT_OBJ_BSEARCH_CMP_FN(type1, type2, nm)	\
+  static int nm##_cmp_BSEARCH_CMP_FN(const void *a_, const void *b_)	\
+      { \
+      type1 const *a = a_; \
+      type2 const *b = b_; \
+      return nm##_cmp(a,b); \
+      } \
+  static type2 *OBJ_bsearch_##nm(type1 *key, type2 const *base, int num) \
+      { \
+      return (type2 *)OBJ_bsearch_(key, base, num, sizeof(type2), \
+					nm##_cmp_BSEARCH_CMP_FN); \
+      } \
+      extern void dummy_prototype(void)
+
+#define IMPLEMENT_OBJ_BSEARCH_GLOBAL_CMP_FN(type1, type2, nm)	\
+  static int nm##_cmp_BSEARCH_CMP_FN(const void *a_, const void *b_)	\
+      { \
+      type1 const *a = a_; \
+      type2 const *b = b_; \
+      return nm##_cmp(a,b); \
+      } \
+  type2 *OBJ_bsearch_##nm(type1 *key, type2 const *base, int num) \
+      { \
+      return (type2 *)OBJ_bsearch_(key, base, num, sizeof(type2), \
+					nm##_cmp_BSEARCH_CMP_FN); \
+      } \
+      extern void dummy_prototype(void)
+
+#define OBJ_bsearch(type1,key,type2,base,num,cmp)			       \
+  ((type2 *)OBJ_bsearch_(CHECKED_PTR_OF(type1,key),CHECKED_PTR_OF(type2,base), \
+			 num,sizeof(type2),				\
+			 ((void)CHECKED_PTR_OF(type1,cmp##_type_1),	\
+			  (void)CHECKED_PTR_OF(type2,cmp##_type_2),	\
+			  cmp##_BSEARCH_CMP_FN)))
+
+#define OBJ_bsearch_ex(type1,key,type2,base,num,cmp,flags)			\
+  ((type2 *)OBJ_bsearch_ex_(CHECKED_PTR_OF(type1,key),CHECKED_PTR_OF(type2,base), \
+			 num,sizeof(type2),				\
+			 ((void)CHECKED_PTR_OF(type1,cmp##_type_1),	\
+			  (void)type_2=CHECKED_PTR_OF(type2,cmp##_type_2), \
+			  cmp##_BSEARCH_CMP_FN)),flags)
 
 int		OBJ_new_nid(int num);
 int		OBJ_add_object(const ASN1_OBJECT *obj);
@@ -1022,6 +1103,14 @@ int		OBJ_create(const char *oid,const char *sn,const char *ln);
 void		OBJ_cleanup(void );
 int		OBJ_create_objects(BIO *in);
 
+int OBJ_find_sigid_algs(int signid, int *pdig_nid, int *ppkey_nid);
+int OBJ_find_sigid_by_algs(int *psignid, int dig_nid, int pkey_nid);
+int OBJ_add_sigid(int signid, int dig_id, int pkey_id);
+void OBJ_sigid_free(void);
+
+extern int obj_cleanup_defer;
+void check_defer(int nid);
+
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
  * made after this point may be overwritten when the script is next run.
diff --git a/crypto/objects/objects.pl b/crypto/objects/objects.pl
index 76c06cc8f9a2..15c00bbd5269 100644
--- a/crypto/objects/objects.pl
+++ b/crypto/objects/objects.pl
@@ -14,6 +14,8 @@ while(<NUMIN>)
 	$Cname =~ s/^X//;
 	if (defined($nidn{$mynum}))
 		{ die "$ARGV[1]:$o:There's already an object with NID ",$mynum," on line ",$order{$mynum},"\n"; }
+	if (defined($nid{$Cname}))
+		{ die "$ARGV[1]:$o:There's already an object with name ",$Cname," on line ",$order{$nid{$Cname}},"\n"; }
 	$nid{$Cname} = $mynum;
 	$nidn{$mynum} = $Cname;
 	$order{$mynum} = $o;
@@ -102,6 +104,7 @@ while (<IN>)
 		$max_nid++;
 		$nid{$Cname} = $max_nid;
 		$nidn{$max_nid} = $Cname;
+print STDERR "Added OID $Cname\n";
 		}
 	$Cname="";
 	}
diff --git a/crypto/objects/objects.txt b/crypto/objects/objects.txt
index e61fe60cbf2b..d3bfad72a2ac 100644
--- a/crypto/objects/objects.txt
+++ b/crypto/objects/objects.txt
@@ -166,6 +166,10 @@ pkcs1 3			: RSA-MD4		: md4WithRSAEncryption
 pkcs1 4			: RSA-MD5		: md5WithRSAEncryption
 pkcs1 5			: RSA-SHA1		: sha1WithRSAEncryption
 # According to PKCS #1 version 2.1
+pkcs1 7			: RSAES-OAEP		: rsaesOaep
+pkcs1 8			: MGF1			: mgf1
+pkcs1 10		: RSASSA-PSS		: rsassaPss
+
 pkcs1 11		: RSA-SHA256		: sha256WithRSAEncryption
 pkcs1 12		: RSA-SHA384		: sha384WithRSAEncryption
 pkcs1 13		: RSA-SHA512		: sha512WithRSAEncryption
@@ -299,6 +303,7 @@ id-smime-alg 4		: id-smime-alg-RC2wrap
 id-smime-alg 5		: id-smime-alg-ESDH
 id-smime-alg 6		: id-smime-alg-CMS3DESwrap
 id-smime-alg 7		: id-smime-alg-CMSRC2wrap
+id-smime-alg 9		: id-alg-PWRI-KEK
 
 # S/MIME Certificate Distribution
 id-smime-cd 1		: id-smime-cd-ldap
@@ -770,6 +775,10 @@ id-ce 55		: targetInformation	: X509v3 AC Targeting
 !Cname no-rev-avail
 id-ce 56		: noRevAvail		: X509v3 No Revocation Available
 
+# From RFC5280
+ext-key-usage 0		: anyExtendedKeyUsage	: Any Extended Key Usage
+
+
 !Cname netscape
 2 16 840 1 113730	: Netscape		: Netscape Communications Corp.
 !Cname netscape-cert-extension
@@ -846,6 +855,10 @@ aes 2			: AES-128-CBC		: aes-128-cbc
 aes 3			: AES-128-OFB		: aes-128-ofb
 !Cname aes-128-cfb128
 aes 4			: AES-128-CFB		: aes-128-cfb
+aes 5			: id-aes128-wrap
+aes 6			: id-aes128-GCM		: aes-128-gcm
+aes 7			: id-aes128-CCM		: aes-128-ccm
+aes 8			: id-aes128-wrap-pad
 
 aes 21			: AES-192-ECB		: aes-192-ecb
 aes 22			: AES-192-CBC		: aes-192-cbc
@@ -853,6 +866,10 @@ aes 22			: AES-192-CBC		: aes-192-cbc
 aes 23			: AES-192-OFB		: aes-192-ofb
 !Cname aes-192-cfb128
 aes 24			: AES-192-CFB		: aes-192-cfb
+aes 25			: id-aes192-wrap
+aes 26			: id-aes192-GCM		: aes-192-gcm
+aes 27			: id-aes192-CCM		: aes-192-ccm
+aes 28			: id-aes192-wrap-pad
 
 aes 41			: AES-256-ECB		: aes-256-ecb
 aes 42			: AES-256-CBC		: aes-256-cbc
@@ -860,6 +877,10 @@ aes 42			: AES-256-CBC		: aes-256-cbc
 aes 43			: AES-256-OFB		: aes-256-ofb
 !Cname aes-256-cfb128
 aes 44			: AES-256-CFB		: aes-256-cfb
+aes 45			: id-aes256-wrap
+aes 46			: id-aes256-GCM		: aes-256-gcm
+aes 47			: id-aes256-CCM		: aes-256-ccm
+aes 48			: id-aes256-wrap-pad
 
 # There are no OIDs for these modes...
 
@@ -869,15 +890,16 @@ aes 44			: AES-256-CFB		: aes-256-cfb
 			: AES-128-CFB8		: aes-128-cfb8
 			: AES-192-CFB8		: aes-192-cfb8
 			: AES-256-CFB8		: aes-256-cfb8
+			: AES-128-CTR		: aes-128-ctr
+			: AES-192-CTR		: aes-192-ctr
+			: AES-256-CTR		: aes-256-ctr
+			: AES-128-XTS		: aes-128-xts
+			: AES-256-XTS		: aes-256-xts
 			: DES-CFB1		: des-cfb1
 			: DES-CFB8		: des-cfb8
 			: DES-EDE3-CFB1		: des-ede3-cfb1
 			: DES-EDE3-CFB8		: des-ede3-cfb8
 
-aes 5			: id-aes128-wrap 
-aes 25			: id-aes192-wrap 
-aes 45			: id-aes256-wrap 
-
 # OIDs for SHA224, SHA256, SHA385 and SHA512, according to x9.84.
 !Alias nist_hashalgs nistAlgorithms 2
 nist_hashalgs 1		: SHA256		: sha256
@@ -1211,6 +1233,9 @@ cryptocom 1 8 1		: id-GostR3410-2001-ParamSet-cc : GOST R 3410-2001 Parameter Se
 1 2 392 200011 61 1 1 1 2 : CAMELLIA-128-CBC		: camellia-128-cbc
 1 2 392 200011 61 1 1 1 3 : CAMELLIA-192-CBC		: camellia-192-cbc
 1 2 392 200011 61 1 1 1 4 : CAMELLIA-256-CBC		: camellia-256-cbc
+1 2 392 200011 61 1 1 3 2 : id-camellia128-wrap
+1 2 392 200011 61 1 1 3 3 : id-camellia192-wrap
+1 2 392 200011 61 1 1 3 4 : id-camellia256-wrap
 
 # Definitions for Camellia cipher - ECB, CFB, OFB MODE
 
@@ -1257,3 +1282,11 @@ kisa 1 6                : SEED-OFB      : seed-ofb
 # There is no OID that just denotes "HMAC" oddly enough...
 
 			: HMAC				: hmac
+# Nor CMAC either
+			: CMAC				: cmac
+
+# Synthetic composite ciphersuites
+			: RC4-HMAC-MD5			: rc4-hmac-md5
+			: AES-128-CBC-HMAC-SHA1		: aes-128-cbc-hmac-sha1
+			: AES-192-CBC-HMAC-SHA1		: aes-192-cbc-hmac-sha1
+			: AES-256-CBC-HMAC-SHA1		: aes-256-cbc-hmac-sha1
diff --git a/crypto/objects/objxref.pl b/crypto/objects/objxref.pl
new file mode 100755
index 000000000000..731d3ae22c27
--- /dev/null
+++ b/crypto/objects/objxref.pl
@@ -0,0 +1,107 @@
+#!/usr/local/bin/perl
+
+use strict;
+
+my %xref_tbl;
+my %oid_tbl;
+
+my ($mac_file, $xref_file) = @ARGV;
+
+open(IN, $mac_file) || die "Can't open $mac_file";
+
+# Read in OID nid values for a lookup table.
+
+while (<IN>)
+	{
+	chomp;
+	my ($name, $num) = /^(\S+)\s+(\S+)$/;
+	$oid_tbl{$name} = $num;
+	}
+close IN;
+
+open(IN, $xref_file) || die "Can't open $xref_file";
+
+my $ln = 1;
+
+while (<IN>)
+	{
+	chomp;
+	s/#.*$//;
+	next if (/^\S*$/);
+	my ($xr, $p1, $p2) = /^(\S+)\s+(\S+)\s+(\S+)/;
+	check_oid($xr);
+	check_oid($p1);
+	check_oid($p2);
+	$xref_tbl{$xr} = [$p1, $p2, $ln];
+	}
+
+my @xrkeys = keys %xref_tbl;
+
+my @srt1 = sort { $oid_tbl{$a} <=> $oid_tbl{$b}} @xrkeys;
+
+for(my $i = 0; $i <= $#srt1; $i++)
+	{
+	$xref_tbl{$srt1[$i]}[2] = $i;
+	}
+
+my @srt2 = sort
+	{
+	my$ap1 = $oid_tbl{$xref_tbl{$a}[0]};
+	my$bp1 = $oid_tbl{$xref_tbl{$b}[0]};
+	return $ap1 - $bp1 if ($ap1 != $bp1);
+	my$ap2 = $oid_tbl{$xref_tbl{$a}[1]};
+	my$bp2 = $oid_tbl{$xref_tbl{$b}[1]};
+
+	return $ap2 - $bp2;
+	} @xrkeys;
+
+my $pname = $0;
+
+$pname =~ s|^.[^/]/||;
+
+print <<EOF;
+/* AUTOGENERATED BY $pname, DO NOT EDIT */
+
+typedef struct
+	{
+	int sign_id;
+	int hash_id;
+	int pkey_id;
+	} nid_triple;
+
+static const nid_triple sigoid_srt[] =
+	{
+EOF
+
+foreach (@srt1)
+	{
+	my $xr = $_;
+	my ($p1, $p2) = @{$xref_tbl{$_}};
+	print "\t{NID_$xr, NID_$p1, NID_$p2},\n";
+	}
+
+print "\t};";
+print <<EOF;
+
+
+static const nid_triple * const sigoid_srt_xref[] =
+	{
+EOF
+
+foreach (@srt2)
+	{
+	my $x = $xref_tbl{$_}[2];
+	print "\t\&sigoid_srt\[$x\],\n";
+	}
+
+print "\t};\n\n";
+
+sub check_oid
+	{
+	my ($chk) = @_;
+	if (!exists $oid_tbl{$chk})
+		{
+		die "Not Found \"$chk\"\n";
+		}
+	}
+
diff --git a/crypto/ocsp/Makefile b/crypto/ocsp/Makefile
index 30a00b337264..60c414cf4da6 100644
--- a/crypto/ocsp/Makefile
+++ b/crypto/ocsp/Makefile
@@ -36,7 +36,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -82,10 +82,9 @@ ocsp_asn.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 ocsp_asn.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 ocsp_asn.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 ocsp_asn.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-ocsp_asn.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-ocsp_asn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-ocsp_asn.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
-ocsp_asn.o: ../../include/openssl/opensslconf.h
+ocsp_asn.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ocsp_asn.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ocsp_asn.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
 ocsp_asn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ocsp_asn.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 ocsp_asn.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -98,25 +97,24 @@ ocsp_cl.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 ocsp_cl.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 ocsp_cl.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 ocsp_cl.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-ocsp_cl.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-ocsp_cl.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-ocsp_cl.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
-ocsp_cl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-ocsp_cl.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
-ocsp_cl.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
-ocsp_cl.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-ocsp_cl.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-ocsp_cl.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-ocsp_cl.o: ../../include/openssl/x509v3.h ../cryptlib.h ocsp_cl.c
+ocsp_cl.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ocsp_cl.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
+ocsp_cl.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+ocsp_cl.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
+ocsp_cl.o: ../../include/openssl/pem2.h ../../include/openssl/pkcs7.h
+ocsp_cl.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
+ocsp_cl.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ocsp_cl.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+ocsp_cl.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+ocsp_cl.o: ../cryptlib.h ocsp_cl.c
 ocsp_err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 ocsp_err.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 ocsp_err.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 ocsp_err.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 ocsp_err.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-ocsp_err.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-ocsp_err.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-ocsp_err.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
-ocsp_err.o: ../../include/openssl/opensslconf.h
+ocsp_err.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ocsp_err.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ocsp_err.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
 ocsp_err.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ocsp_err.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 ocsp_err.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -129,9 +127,9 @@ ocsp_ext.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 ocsp_ext.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 ocsp_ext.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 ocsp_ext.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-ocsp_ext.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-ocsp_ext.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-ocsp_ext.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
+ocsp_ext.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ocsp_ext.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
+ocsp_ext.o: ../../include/openssl/opensslconf.h
 ocsp_ext.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ocsp_ext.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
 ocsp_ext.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -144,22 +142,21 @@ ocsp_ht.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 ocsp_ht.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 ocsp_ht.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 ocsp_ht.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-ocsp_ht.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-ocsp_ht.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-ocsp_ht.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
-ocsp_ht.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-ocsp_ht.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-ocsp_ht.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-ocsp_ht.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-ocsp_ht.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-ocsp_ht.o: ocsp_ht.c
+ocsp_ht.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ocsp_ht.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
+ocsp_ht.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+ocsp_ht.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ocsp_ht.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+ocsp_ht.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+ocsp_ht.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ocsp_ht.o: ../../include/openssl/x509v3.h ocsp_ht.c
 ocsp_lib.o: ../../e_os.h ../../include/openssl/asn1.h
-ocsp_lib.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
-ocsp_lib.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
-ocsp_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
-ocsp_lib.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-ocsp_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-ocsp_lib.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+ocsp_lib.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+ocsp_lib.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
+ocsp_lib.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+ocsp_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+ocsp_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+ocsp_lib.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
 ocsp_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
 ocsp_lib.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
 ocsp_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
@@ -174,10 +171,9 @@ ocsp_prn.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 ocsp_prn.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 ocsp_prn.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 ocsp_prn.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-ocsp_prn.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-ocsp_prn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-ocsp_prn.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
-ocsp_prn.o: ../../include/openssl/opensslconf.h
+ocsp_prn.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ocsp_prn.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ocsp_prn.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
 ocsp_prn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ocsp_prn.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
 ocsp_prn.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
@@ -191,9 +187,9 @@ ocsp_srv.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 ocsp_srv.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 ocsp_srv.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 ocsp_srv.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-ocsp_srv.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-ocsp_srv.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-ocsp_srv.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
+ocsp_srv.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ocsp_srv.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
+ocsp_srv.o: ../../include/openssl/opensslconf.h
 ocsp_srv.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ocsp_srv.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
 ocsp_srv.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
@@ -206,10 +202,9 @@ ocsp_vfy.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 ocsp_vfy.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 ocsp_vfy.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 ocsp_vfy.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-ocsp_vfy.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-ocsp_vfy.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-ocsp_vfy.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
-ocsp_vfy.o: ../../include/openssl/opensslconf.h
+ocsp_vfy.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ocsp_vfy.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ocsp_vfy.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
 ocsp_vfy.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 ocsp_vfy.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 ocsp_vfy.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
diff --git a/crypto/ocsp/ocsp.h b/crypto/ocsp/ocsp.h
index a0577a717ef4..31e45744bac4 100644
--- a/crypto/ocsp/ocsp.h
+++ b/crypto/ocsp/ocsp.h
@@ -64,6 +64,7 @@
 #ifndef HEADER_OCSP_H
 #define HEADER_OCSP_H
 
+#include <openssl/ossl_typ.h>
 #include <openssl/x509.h>
 #include <openssl/x509v3.h>
 #include <openssl/safestack.h>
@@ -394,17 +395,20 @@ typedef struct ocsp_service_locator_st
 #define ASN1_BIT_STRING_digest(data,type,md,len) \
 	ASN1_item_digest(ASN1_ITEM_rptr(ASN1_BIT_STRING),type,data,md,len)
 
-#define OCSP_CERTID_dup(cid) ASN1_dup_of(OCSP_CERTID,i2d_OCSP_CERTID,d2i_OCSP_CERTID,cid)
-
 #define OCSP_CERTSTATUS_dup(cs)\
                 (OCSP_CERTSTATUS*)ASN1_dup((int(*)())i2d_OCSP_CERTSTATUS,\
 		(char *(*)())d2i_OCSP_CERTSTATUS,(char *)(cs))
 
+OCSP_CERTID *OCSP_CERTID_dup(OCSP_CERTID *id);
+
 OCSP_RESPONSE *OCSP_sendreq_bio(BIO *b, char *path, OCSP_REQUEST *req);
 OCSP_REQ_CTX *OCSP_sendreq_new(BIO *io, char *path, OCSP_REQUEST *req,
 								int maxline);
 int OCSP_sendreq_nbio(OCSP_RESPONSE **presp, OCSP_REQ_CTX *rctx);
 void OCSP_REQ_CTX_free(OCSP_REQ_CTX *rctx);
+int OCSP_REQ_CTX_set1_req(OCSP_REQ_CTX *rctx, OCSP_REQUEST *req);
+int OCSP_REQ_CTX_add1_header(OCSP_REQ_CTX *rctx,
+		const char *name, const char *value);
 
 OCSP_CERTID *OCSP_cert_to_id(const EVP_MD *dgst, X509 *subject, X509 *issuer);
 
@@ -474,11 +478,6 @@ int OCSP_basic_sign(OCSP_BASICRESP *brsp,
 			X509 *signer, EVP_PKEY *key, const EVP_MD *dgst,
 			STACK_OF(X509) *certs, unsigned long flags);
 
-ASN1_STRING *ASN1_STRING_encode(ASN1_STRING *s, i2d_of_void *i2d,
-				void *data, STACK_OF(ASN1_OBJECT) *sk);
-#define ASN1_STRING_encode_of(type,s,i2d,data,sk) \
-	ASN1_STRING_encode(s, CHECKED_I2D_OF(type, i2d), data, sk)
-
 X509_EXTENSION *OCSP_crlID_new(char *url, long *n, char *tim);
 
 X509_EXTENSION *OCSP_accept_responses_new(char **oids);
@@ -547,9 +546,9 @@ DECLARE_ASN1_FUNCTIONS(OCSP_REQINFO)
 DECLARE_ASN1_FUNCTIONS(OCSP_CRLID)
 DECLARE_ASN1_FUNCTIONS(OCSP_SERVICELOC)
 
-char *OCSP_response_status_str(long s);
-char *OCSP_cert_status_str(long s);
-char *OCSP_crl_reason_str(long s);
+const char *OCSP_response_status_str(long s);
+const char *OCSP_cert_status_str(long s);
+const char *OCSP_crl_reason_str(long s);
 
 int OCSP_REQUEST_print(BIO *bp, OCSP_REQUEST* a, unsigned long flags);
 int OCSP_RESPONSE_print(BIO *bp, OCSP_RESPONSE* o, unsigned long flags);
@@ -582,7 +581,8 @@ void ERR_load_OCSP_strings(void);
 #define OCSP_F_OCSP_REQUEST_VERIFY			 116
 #define OCSP_F_OCSP_RESPONSE_GET1_BASIC			 111
 #define OCSP_F_OCSP_SENDREQ_BIO				 112
-#define OCSP_F_PARSE_HTTP_LINE1				 117
+#define OCSP_F_OCSP_SENDREQ_NBIO			 117
+#define OCSP_F_PARSE_HTTP_LINE1				 118
 #define OCSP_F_REQUEST_VERIFY				 113
 
 /* Reason codes. */
diff --git a/crypto/ocsp/ocsp_cl.c b/crypto/ocsp/ocsp_cl.c
index 17bab5fc59c7..9c14d9da273f 100755
--- a/crypto/ocsp/ocsp_cl.c
+++ b/crypto/ocsp/ocsp_cl.c
@@ -155,7 +155,6 @@ int OCSP_request_sign(OCSP_REQUEST   *req,
 			goto err;
 
 	if (!(req->optionalSignature = sig = OCSP_SIGNATURE_new())) goto err;
-	if (!dgst) dgst = EVP_sha1();
 	if (key)
 		{
 		if (!X509_check_private_key(signer, key))
diff --git a/crypto/ocsp/ocsp_err.c b/crypto/ocsp/ocsp_err.c
index d2f2e79f444c..0cedcea68204 100644
--- a/crypto/ocsp/ocsp_err.c
+++ b/crypto/ocsp/ocsp_err.c
@@ -1,6 +1,6 @@
 /* crypto/ocsp/ocsp_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -86,6 +86,7 @@ static ERR_STRING_DATA OCSP_str_functs[]=
 {ERR_FUNC(OCSP_F_OCSP_REQUEST_VERIFY),	"OCSP_request_verify"},
 {ERR_FUNC(OCSP_F_OCSP_RESPONSE_GET1_BASIC),	"OCSP_response_get1_basic"},
 {ERR_FUNC(OCSP_F_OCSP_SENDREQ_BIO),	"OCSP_sendreq_bio"},
+{ERR_FUNC(OCSP_F_OCSP_SENDREQ_NBIO),	"OCSP_sendreq_nbio"},
 {ERR_FUNC(OCSP_F_PARSE_HTTP_LINE1),	"PARSE_HTTP_LINE1"},
 {ERR_FUNC(OCSP_F_REQUEST_VERIFY),	"REQUEST_VERIFY"},
 {0,NULL}
diff --git a/crypto/ocsp/ocsp_ext.c b/crypto/ocsp/ocsp_ext.c
index 815cc29d58fe..ec884cb08f44 100755
--- a/crypto/ocsp/ocsp_ext.c
+++ b/crypto/ocsp/ocsp_ext.c
@@ -264,7 +264,7 @@ int OCSP_SINGLERESP_add_ext(OCSP_SINGLERESP *x, X509_EXTENSION *ex, int loc)
 	}
 
 /* also CRL Entry Extensions */
-
+#if 0
 ASN1_STRING *ASN1_STRING_encode(ASN1_STRING *s, i2d_of_void *i2d,
 				void *data, STACK_OF(ASN1_OBJECT) *sk)
         {
@@ -305,6 +305,7 @@ err:
 	if (b) OPENSSL_free(b);
 	return NULL;
 	}
+#endif
 
 /* Nonce handling functions */
 
@@ -442,17 +443,10 @@ X509_EXTENSION *OCSP_crlID_new(char *url, long *n, char *tim)
 		if (!(ASN1_GENERALIZEDTIME_set_string(cid->crlTime, tim))) 
 		        goto err;
 		}
-	if (!(x = X509_EXTENSION_new())) goto err;
-	if (!(x->object = OBJ_nid2obj(NID_id_pkix_OCSP_CrlID))) goto err;
-	if (!(ASN1_STRING_encode_of(OCSP_CRLID,x->value,i2d_OCSP_CRLID,cid,
-				    NULL)))
-	        goto err;
-	OCSP_CRLID_free(cid);
-	return x;
+	x = X509V3_EXT_i2d(NID_id_pkix_OCSP_CrlID, 0, cid);
 err:
-	if (x) X509_EXTENSION_free(x);
 	if (cid) OCSP_CRLID_free(cid);
-	return NULL;
+	return x;
 	}
 
 /*   AcceptableResponses ::= SEQUENCE OF OBJECT IDENTIFIER */
@@ -470,18 +464,10 @@ X509_EXTENSION *OCSP_accept_responses_new(char **oids)
 		        sk_ASN1_OBJECT_push(sk, o);
 		oids++;
 		}
-	if (!(x = X509_EXTENSION_new())) goto err;
-	if (!(x->object = OBJ_nid2obj(NID_id_pkix_OCSP_acceptableResponses)))
-		goto err;
-	if (!(ASN1_STRING_encode_of(ASN1_OBJECT,x->value,i2d_ASN1_OBJECT,NULL,
-				    sk)))
-	        goto err;
-	sk_ASN1_OBJECT_pop_free(sk, ASN1_OBJECT_free);
-	return x;
+	x = X509V3_EXT_i2d(NID_id_pkix_OCSP_acceptableResponses, 0, sk);
 err:
-	if (x) X509_EXTENSION_free(x);
 	if (sk) sk_ASN1_OBJECT_pop_free(sk, ASN1_OBJECT_free);
-	return NULL;
+	return x;
         }
 
 /*  ArchiveCutoff ::= GeneralizedTime */
@@ -492,16 +478,10 @@ X509_EXTENSION *OCSP_archive_cutoff_new(char* tim)
 
 	if (!(gt = ASN1_GENERALIZEDTIME_new())) goto err;
 	if (!(ASN1_GENERALIZEDTIME_set_string(gt, tim))) goto err;
-	if (!(x = X509_EXTENSION_new())) goto err;
-	if (!(x->object=OBJ_nid2obj(NID_id_pkix_OCSP_archiveCutoff)))goto err;
-	if (!(ASN1_STRING_encode_of(ASN1_GENERALIZEDTIME,x->value,
-				    i2d_ASN1_GENERALIZEDTIME,gt,NULL))) goto err;
-	ASN1_GENERALIZEDTIME_free(gt);
-	return x;
+	x = X509V3_EXT_i2d(NID_id_pkix_OCSP_archiveCutoff, 0, gt);
 err:
 	if (gt) ASN1_GENERALIZEDTIME_free(gt);
-	if (x) X509_EXTENSION_free(x);
-	return NULL;
+	return x;
 	}
 
 /* per ACCESS_DESCRIPTION parameter are oids, of which there are currently
@@ -530,16 +510,9 @@ X509_EXTENSION *OCSP_url_svcloc_new(X509_NAME* issuer, char **urls)
 		if (!sk_ACCESS_DESCRIPTION_push(sloc->locator, ad)) goto err;
 		urls++;
 		}
-	if (!(x = X509_EXTENSION_new())) goto err;
-	if (!(x->object = OBJ_nid2obj(NID_id_pkix_OCSP_serviceLocator))) 
-	        goto err;
-	if (!(ASN1_STRING_encode_of(OCSP_SERVICELOC,x->value,
-				    i2d_OCSP_SERVICELOC,sloc,NULL))) goto err;
-	OCSP_SERVICELOC_free(sloc);
-	return x;
+	x = X509V3_EXT_i2d(NID_id_pkix_OCSP_serviceLocator, 0, sloc);
 err:
-	if (x) X509_EXTENSION_free(x);
 	if (sloc) OCSP_SERVICELOC_free(sloc);
-	return NULL;
+	return x;
 	}
 
diff --git a/crypto/ocsp/ocsp_ht.c b/crypto/ocsp/ocsp_ht.c
index 92aba08c8fbd..af5fc16691d4 100644
--- a/crypto/ocsp/ocsp_ht.c
+++ b/crypto/ocsp/ocsp_ht.c
@@ -118,39 +118,65 @@ void OCSP_REQ_CTX_free(OCSP_REQ_CTX *rctx)
 	OPENSSL_free(rctx);
 	}
 
-OCSP_REQ_CTX *OCSP_sendreq_new(BIO *io, char *path, OCSP_REQUEST *req,
-								int maxline)
+int OCSP_REQ_CTX_set1_req(OCSP_REQ_CTX *rctx, OCSP_REQUEST *req)
 	{
-	static char post_hdr[] = "POST %s HTTP/1.0\r\n"
+	static const char req_hdr[] =
 	"Content-Type: application/ocsp-request\r\n"
 	"Content-Length: %d\r\n\r\n";
+        if (BIO_printf(rctx->mem, req_hdr, i2d_OCSP_REQUEST(req, NULL)) <= 0)
+		return 0;
+        if (i2d_OCSP_REQUEST_bio(rctx->mem, req) <= 0)
+		return 0;
+	rctx->state = OHS_ASN1_WRITE;
+	rctx->asn1_len = BIO_get_mem_data(rctx->mem, NULL);
+	return 1;
+	}
+
+int OCSP_REQ_CTX_add1_header(OCSP_REQ_CTX *rctx,
+		const char *name, const char *value)
+	{
+	if (!name)
+		return 0;
+	if (BIO_puts(rctx->mem, name) <= 0)
+		return 0;
+	if (value)
+		{
+		if (BIO_write(rctx->mem, ": ", 2) != 2)
+			return 0;
+		if (BIO_puts(rctx->mem, value) <= 0)
+			return 0;
+		}
+	if (BIO_write(rctx->mem, "\r\n", 2) != 2)
+		return 0;
+	return 1;
+	}
+
+OCSP_REQ_CTX *OCSP_sendreq_new(BIO *io, char *path, OCSP_REQUEST *req,
+								int maxline)
+	{
+	static const char post_hdr[] = "POST %s HTTP/1.0\r\n";
 
 	OCSP_REQ_CTX *rctx;
 	rctx = OPENSSL_malloc(sizeof(OCSP_REQ_CTX));
-	rctx->state = OHS_FIRSTLINE;
+	rctx->state = OHS_ERROR;
 	rctx->mem = BIO_new(BIO_s_mem());
 	rctx->io = io;
+	rctx->asn1_len = 0;
 	if (maxline > 0)
 		rctx->iobuflen = maxline;
 	else
 		rctx->iobuflen = OCSP_MAX_LINE_LEN;
 	rctx->iobuf = OPENSSL_malloc(rctx->iobuflen);
+	if (!rctx->iobuf)
+		return 0;
 	if (!path)
 		path = "/";
 
-        if (BIO_printf(rctx->mem, post_hdr, path,
-				i2d_OCSP_REQUEST(req, NULL)) <= 0)
-		{
-		rctx->state = OHS_ERROR;
+        if (BIO_printf(rctx->mem, post_hdr, path) <= 0)
 		return 0;
-		}
-        if (i2d_OCSP_REQUEST_bio(rctx->mem, req) <= 0)
-		{
-		rctx->state = OHS_ERROR;
+
+	if (req && !OCSP_REQ_CTX_set1_req(rctx, req))
 		return 0;
-		}
-	rctx->state = OHS_ASN1_WRITE;
-	rctx->asn1_len = BIO_get_mem_data(rctx->mem, NULL);
 
 	return rctx;
 	}
diff --git a/crypto/ocsp/ocsp_lib.c b/crypto/ocsp/ocsp_lib.c
index 441ccb7a9e4e..a94dc838eecb 100755
--- a/crypto/ocsp/ocsp_lib.c
+++ b/crypto/ocsp/ocsp_lib.c
@@ -69,6 +69,7 @@
 #include <openssl/pem.h>
 #include <openssl/x509v3.h>
 #include <openssl/ocsp.h>
+#include <openssl/asn1t.h>
 
 /* Convert a certificate and its issuer to an OCSP_CERTID */
 
@@ -123,7 +124,8 @@ OCSP_CERTID *OCSP_cert_id_new(const EVP_MD *dgst,
 	if (!(ASN1_OCTET_STRING_set(cid->issuerNameHash, md, i))) goto err;
 
 	/* Calculate the issuerKey hash, excluding tag and length */
-	EVP_Digest(issuerKey->data, issuerKey->length, md, &i, dgst, NULL);
+	if (!EVP_Digest(issuerKey->data, issuerKey->length, md, &i, dgst, NULL))
+		goto err;
 
 	if (!(ASN1_OCTET_STRING_set(cid->issuerKeyHash, md, i))) goto err;
 
@@ -260,3 +262,5 @@ int OCSP_parse_url(char *url, char **phost, char **pport, char **ppath, int *pss
 	return 0;
 
 	}
+
+IMPLEMENT_ASN1_DUP_FUNCTION(OCSP_CERTID)
diff --git a/crypto/ocsp/ocsp_prn.c b/crypto/ocsp/ocsp_prn.c
index b8b7871d0a8c..87608ff399b1 100644
--- a/crypto/ocsp/ocsp_prn.c
+++ b/crypto/ocsp/ocsp_prn.c
@@ -85,21 +85,21 @@ static int ocsp_certid_print(BIO *bp, OCSP_CERTID* a, int indent)
 typedef struct
 	{
 	long t;
-	char *m;
+	const char *m;
 	} OCSP_TBLSTR;
 
-static char *table2string(long s, OCSP_TBLSTR *ts, int len)
+static const char *table2string(long s, const OCSP_TBLSTR *ts, int len)
 {
-	OCSP_TBLSTR *p;
+	const OCSP_TBLSTR *p;
 	for (p=ts; p < ts + len; p++)
 	        if (p->t == s)
 		         return p->m;
 	return "(UNKNOWN)";
 }
 
-char *OCSP_response_status_str(long s)
+const char *OCSP_response_status_str(long s)
         {
-	static OCSP_TBLSTR rstat_tbl[] = {
+	static const OCSP_TBLSTR rstat_tbl[] = {
 	        { OCSP_RESPONSE_STATUS_SUCCESSFUL, "successful" },
 	        { OCSP_RESPONSE_STATUS_MALFORMEDREQUEST, "malformedrequest" },
 	        { OCSP_RESPONSE_STATUS_INTERNALERROR, "internalerror" },
@@ -109,18 +109,18 @@ char *OCSP_response_status_str(long s)
 	return table2string(s, rstat_tbl, 6);
 	} 
 
-char *OCSP_cert_status_str(long s)
+const char *OCSP_cert_status_str(long s)
         {
-	static OCSP_TBLSTR cstat_tbl[] = {
+	static const OCSP_TBLSTR cstat_tbl[] = {
 	        { V_OCSP_CERTSTATUS_GOOD, "good" },
 	        { V_OCSP_CERTSTATUS_REVOKED, "revoked" },
 	        { V_OCSP_CERTSTATUS_UNKNOWN, "unknown" } };
 	return table2string(s, cstat_tbl, 3);
 	} 
 
-char *OCSP_crl_reason_str(long s)
+const char *OCSP_crl_reason_str(long s)
         {
-	OCSP_TBLSTR reason_tbl[] = {
+	static const OCSP_TBLSTR reason_tbl[] = {
 	  { OCSP_REVOKED_STATUS_UNSPECIFIED, "unspecified" },
           { OCSP_REVOKED_STATUS_KEYCOMPROMISE, "keyCompromise" },
           { OCSP_REVOKED_STATUS_CACOMPROMISE, "cACompromise" },
diff --git a/crypto/ocsp/ocsp_vfy.c b/crypto/ocsp/ocsp_vfy.c
index 4a0c3870d83a..415d67e61cfd 100644
--- a/crypto/ocsp/ocsp_vfy.c
+++ b/crypto/ocsp/ocsp_vfy.c
@@ -308,6 +308,8 @@ static int ocsp_match_issuerid(X509 *cert, OCSP_CERTID *cid,
 			}
 
 		mdlen = EVP_MD_size(dgst);
+		if (mdlen < 0)
+		    return -1;
 		if ((cid->issuerNameHash->length != mdlen) ||
 		   (cid->issuerKeyHash->length != mdlen))
 			return 0;
@@ -316,7 +318,7 @@ static int ocsp_match_issuerid(X509 *cert, OCSP_CERTID *cid,
 			return -1;
 		if (memcmp(md, cid->issuerNameHash->data, mdlen))
 			return 0;
-		X509_pubkey_digest(cert, EVP_sha1(), md, NULL);
+		X509_pubkey_digest(cert, dgst, md, NULL);
 		if (memcmp(md, cid->issuerKeyHash->data, mdlen))
 			return 0;
 
diff --git a/crypto/opensslconf.h b/crypto/opensslconf.h
index 60505327d386..b18f4da49661 100644
--- a/crypto/opensslconf.h
+++ b/crypto/opensslconf.h
@@ -5,14 +5,8 @@
 #ifndef OPENSSL_DOING_MAKEDEPEND
 
 
-#ifndef OPENSSL_NO_CAMELLIA
-# define OPENSSL_NO_CAMELLIA
-#endif
-#ifndef OPENSSL_NO_CAPIENG
-# define OPENSSL_NO_CAPIENG
-#endif
-#ifndef OPENSSL_NO_CMS
-# define OPENSSL_NO_CMS
+#ifndef OPENSSL_NO_EC_NISTP_64_GCC_128
+# define OPENSSL_NO_EC_NISTP_64_GCC_128
 #endif
 #ifndef OPENSSL_NO_GMP
 # define OPENSSL_NO_GMP
@@ -23,8 +17,8 @@
 #ifndef OPENSSL_NO_KRB5
 # define OPENSSL_NO_KRB5
 #endif
-#ifndef OPENSSL_NO_MDC2
-# define OPENSSL_NO_MDC2
+#ifndef OPENSSL_NO_MD2
+# define OPENSSL_NO_MD2
 #endif
 #ifndef OPENSSL_NO_RC5
 # define OPENSSL_NO_RC5
@@ -32,8 +26,11 @@
 #ifndef OPENSSL_NO_RFC3779
 # define OPENSSL_NO_RFC3779
 #endif
-#ifndef OPENSSL_NO_SEED
-# define OPENSSL_NO_SEED
+#ifndef OPENSSL_NO_SCTP
+# define OPENSSL_NO_SCTP
+#endif
+#ifndef OPENSSL_NO_STORE
+# define OPENSSL_NO_STORE
 #endif
 
 #endif /* OPENSSL_DOING_MAKEDEPEND */
@@ -47,14 +44,8 @@
    who haven't had the time to do the appropriate changes in their
    applications.  */
 #ifdef OPENSSL_ALGORITHM_DEFINES
-# if defined(OPENSSL_NO_CAMELLIA) && !defined(NO_CAMELLIA)
-#  define NO_CAMELLIA
-# endif
-# if defined(OPENSSL_NO_CAPIENG) && !defined(NO_CAPIENG)
-#  define NO_CAPIENG
-# endif
-# if defined(OPENSSL_NO_CMS) && !defined(NO_CMS)
-#  define NO_CMS
+# if defined(OPENSSL_NO_EC_NISTP_64_GCC_128) && !defined(NO_EC_NISTP_64_GCC_128)
+#  define NO_EC_NISTP_64_GCC_128
 # endif
 # if defined(OPENSSL_NO_GMP) && !defined(NO_GMP)
 #  define NO_GMP
@@ -65,8 +56,8 @@
 # if defined(OPENSSL_NO_KRB5) && !defined(NO_KRB5)
 #  define NO_KRB5
 # endif
-# if defined(OPENSSL_NO_MDC2) && !defined(NO_MDC2)
-#  define NO_MDC2
+# if defined(OPENSSL_NO_MD2) && !defined(NO_MD2)
+#  define NO_MD2
 # endif
 # if defined(OPENSSL_NO_RC5) && !defined(NO_RC5)
 #  define NO_RC5
@@ -74,28 +65,16 @@
 # if defined(OPENSSL_NO_RFC3779) && !defined(NO_RFC3779)
 #  define NO_RFC3779
 # endif
-# if defined(OPENSSL_NO_SEED) && !defined(NO_SEED)
-#  define NO_SEED
+# if defined(OPENSSL_NO_SCTP) && !defined(NO_SCTP)
+#  define NO_SCTP
+# endif
+# if defined(OPENSSL_NO_STORE) && !defined(NO_STORE)
+#  define NO_STORE
 # endif
 #endif
 
 /* crypto/opensslconf.h.in */
 
-#ifdef OPENSSL_DOING_MAKEDEPEND
-
-/* Include any symbols here that have to be explicitly set to enable a feature
- * that should be visible to makedepend.
- *
- * [Our "make depend" doesn't actually look at this, we use actual build settings
- * instead; we want to make it easy to remove subdirectories with disabled algorithms.]
- */
-
-#ifndef OPENSSL_FIPS
-#define OPENSSL_FIPS
-#endif
-
-#endif
-
 /* Generate 80386 code? */
 #undef I386_ONLY
 
@@ -159,14 +138,9 @@
 /* Should we define BN_DIV2W here? */
 
 /* Only one for the following should be defined */
-/* The prime number generation stuff may not work when
- * EIGHT_BIT but I don't care since I've only used this mode
- * for debuging the bignum libraries */
 #undef SIXTY_FOUR_BIT_LONG
 #undef SIXTY_FOUR_BIT
 #define THIRTY_TWO_BIT
-#undef SIXTEEN_BIT
-#undef EIGHT_BIT
 #endif
 
 #if defined(HEADER_RC4_LOCL_H) && !defined(CONFIG_HEADER_RC4_LOCL_H)
diff --git a/crypto/opensslconf.h.in b/crypto/opensslconf.h.in
index 1c77f03c3dc3..97e37455634d 100644
--- a/crypto/opensslconf.h.in
+++ b/crypto/opensslconf.h.in
@@ -1,20 +1,5 @@
 /* crypto/opensslconf.h.in */
 
-#ifdef OPENSSL_DOING_MAKEDEPEND
-
-/* Include any symbols here that have to be explicitly set to enable a feature
- * that should be visible to makedepend.
- *
- * [Our "make depend" doesn't actually look at this, we use actual build settings
- * instead; we want to make it easy to remove subdirectories with disabled algorithms.]
- */
-
-#ifndef OPENSSL_FIPS
-#define OPENSSL_FIPS
-#endif
-
-#endif
-
 /* Generate 80386 code? */
 #undef I386_ONLY
 
@@ -78,14 +63,9 @@
 /* Should we define BN_DIV2W here? */
 
 /* Only one for the following should be defined */
-/* The prime number generation stuff may not work when
- * EIGHT_BIT but I don't care since I've only used this mode
- * for debuging the bignum libraries */
 #undef SIXTY_FOUR_BIT_LONG
 #undef SIXTY_FOUR_BIT
 #define THIRTY_TWO_BIT
-#undef SIXTEEN_BIT
-#undef EIGHT_BIT
 #endif
 
 #if defined(HEADER_RC4_LOCL_H) && !defined(CONFIG_HEADER_RC4_LOCL_H)
diff --git a/crypto/opensslv.h b/crypto/opensslv.h
index 516940c887c4..71be3590af47 100644
--- a/crypto/opensslv.h
+++ b/crypto/opensslv.h
@@ -12,7 +12,7 @@
  * 0.9.3-beta2    0x00903002 (same as ...beta2-dev)
  * 0.9.3	  0x0090300f
  * 0.9.3a	  0x0090301f
- * 0.9.4	  0x0090400f
+ * 0.9.4 	  0x0090400f
  * 1.2.3z	  0x102031af
  *
  * For continuity reasons (because 0.9.5 is already out, and is coded
@@ -25,11 +25,11 @@
  * (Prior to 0.9.5a beta1, a different scheme was used: MMNNFFRBB for
  *  major minor fix final patch/beta)
  */
-#define OPENSSL_VERSION_NUMBER	0x0090818fL
+#define OPENSSL_VERSION_NUMBER	0x1000103fL
 #ifdef OPENSSL_FIPS
-#define OPENSSL_VERSION_TEXT	"OpenSSL 0.9.8x-fips 10 May 2012"
+#define OPENSSL_VERSION_TEXT	"OpenSSL 1.0.1c-fips 10 May 2012"
 #else
-#define OPENSSL_VERSION_TEXT	"OpenSSL 0.9.8x 10 May 2012"
+#define OPENSSL_VERSION_TEXT	"OpenSSL 1.0.1c 10 May 2012"
 #endif
 #define OPENSSL_VERSION_PTEXT	" part of " OPENSSL_VERSION_TEXT
 
@@ -83,7 +83,7 @@
  * should only keep the versions that are binary compatible with the current.
  */
 #define SHLIB_VERSION_HISTORY ""
-#define SHLIB_VERSION_NUMBER "0.9.8"
+#define SHLIB_VERSION_NUMBER "1.0.0"
 
 
 #endif /* HEADER_OPENSSLV_H */
diff --git a/crypto/ossl_typ.h b/crypto/ossl_typ.h
index 0e7a38088035..ea9227f6f95a 100644
--- a/crypto/ossl_typ.h
+++ b/crypto/ossl_typ.h
@@ -91,10 +91,14 @@ typedef struct asn1_string_st ASN1_TIME;
 typedef struct asn1_string_st ASN1_GENERALIZEDTIME;
 typedef struct asn1_string_st ASN1_VISIBLESTRING;
 typedef struct asn1_string_st ASN1_UTF8STRING;
+typedef struct asn1_string_st ASN1_STRING;
 typedef int ASN1_BOOLEAN;
 typedef int ASN1_NULL;
 #endif
 
+typedef struct ASN1_ITEM_st ASN1_ITEM;
+typedef struct asn1_pctx_st ASN1_PCTX;
+
 #ifdef OPENSSL_SYS_WIN32
 #undef X509_NAME
 #undef X509_EXTENSIONS
@@ -122,6 +126,11 @@ typedef struct env_md_st EVP_MD;
 typedef struct env_md_ctx_st EVP_MD_CTX;
 typedef struct evp_pkey_st EVP_PKEY;
 
+typedef struct evp_pkey_asn1_method_st EVP_PKEY_ASN1_METHOD;
+
+typedef struct evp_pkey_method_st EVP_PKEY_METHOD;
+typedef struct evp_pkey_ctx_st EVP_PKEY_CTX;
+
 typedef struct dh_st DH;
 typedef struct dh_method DH_METHOD;
 
@@ -139,11 +148,14 @@ typedef struct ecdsa_method ECDSA_METHOD;
 typedef struct x509_st X509;
 typedef struct X509_algor_st X509_ALGOR;
 typedef struct X509_crl_st X509_CRL;
+typedef struct x509_crl_method_st X509_CRL_METHOD;
+typedef struct x509_revoked_st X509_REVOKED;
 typedef struct X509_name_st X509_NAME;
+typedef struct X509_pubkey_st X509_PUBKEY;
 typedef struct x509_store_st X509_STORE;
 typedef struct x509_store_ctx_st X509_STORE_CTX;
-typedef struct ssl_st SSL;
-typedef struct ssl_ctx_st SSL_CTX;
+
+typedef struct pkcs8_priv_key_info_st PKCS8_PRIV_KEY_INFO;
 
 typedef struct v3_ext_ctx X509V3_CTX;
 typedef struct conf_st CONF;
@@ -157,12 +169,19 @@ typedef struct ui_method_st UI_METHOD;
 typedef struct st_ERR_FNS ERR_FNS;
 
 typedef struct engine_st ENGINE;
+typedef struct ssl_st SSL;
+typedef struct ssl_ctx_st SSL_CTX;
 
 typedef struct X509_POLICY_NODE_st X509_POLICY_NODE;
 typedef struct X509_POLICY_LEVEL_st X509_POLICY_LEVEL;
 typedef struct X509_POLICY_TREE_st X509_POLICY_TREE;
 typedef struct X509_POLICY_CACHE_st X509_POLICY_CACHE;
 
+typedef struct AUTHORITY_KEYID_st AUTHORITY_KEYID;
+typedef struct DIST_POINT_st DIST_POINT;
+typedef struct ISSUING_DIST_POINT_st ISSUING_DIST_POINT;
+typedef struct NAME_CONSTRAINTS_st NAME_CONSTRAINTS;
+
   /* If placed in pkcs12.h, we end up with a circular depency with pkcs7.h */
 #define DECLARE_PKCS12_STACK_OF(type) /* Nothing */
 #define IMPLEMENT_PKCS12_STACK_OF(type) /* Nothing */
diff --git a/crypto/pariscid.pl b/crypto/pariscid.pl
new file mode 100755
index 000000000000..477ec9b87dd7
--- /dev/null
+++ b/crypto/pariscid.pl
@@ -0,0 +1,224 @@
+#!/usr/bin/env perl
+
+$flavour = shift;
+$output = shift;
+open STDOUT,">$output";
+
+if ($flavour =~ /64/) {
+	$LEVEL		="2.0W";
+	$SIZE_T		=8;
+	$ST		="std";
+} else {
+	$LEVEL		="1.1";
+	$SIZE_T		=4;
+	$ST		="stw";
+}
+
+$rp="%r2";
+$sp="%r30";
+$rv="%r28";
+
+$code=<<___;
+	.LEVEL	$LEVEL
+	.SPACE	\$TEXT\$
+	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY
+
+	.EXPORT	OPENSSL_cpuid_setup,ENTRY
+	.ALIGN	8
+OPENSSL_cpuid_setup
+	.PROC
+	.CALLINFO	NO_CALLS
+	.ENTRY
+	bv	($rp)
+	.EXIT
+	nop
+	.PROCEND
+
+	.EXPORT	OPENSSL_rdtsc,ENTRY
+	.ALIGN	8
+OPENSSL_rdtsc
+	.PROC
+	.CALLINFO	NO_CALLS
+	.ENTRY
+	mfctl	%cr16,$rv
+	bv	($rp)
+	.EXIT
+	nop
+	.PROCEND
+
+	.EXPORT	OPENSSL_wipe_cpu,ENTRY
+	.ALIGN	8
+OPENSSL_wipe_cpu
+	.PROC
+	.CALLINFO	NO_CALLS
+	.ENTRY
+	xor		%r0,%r0,%r1
+	fcpy,dbl	%fr0,%fr4
+	xor		%r0,%r0,%r19
+	fcpy,dbl	%fr0,%fr5
+	xor		%r0,%r0,%r20
+	fcpy,dbl	%fr0,%fr6
+	xor		%r0,%r0,%r21
+	fcpy,dbl	%fr0,%fr7
+	xor		%r0,%r0,%r22
+	fcpy,dbl	%fr0,%fr8
+	xor		%r0,%r0,%r23
+	fcpy,dbl	%fr0,%fr9
+	xor		%r0,%r0,%r24
+	fcpy,dbl	%fr0,%fr10
+	xor		%r0,%r0,%r25
+	fcpy,dbl	%fr0,%fr11
+	xor		%r0,%r0,%r26
+	fcpy,dbl	%fr0,%fr22
+	xor		%r0,%r0,%r29
+	fcpy,dbl	%fr0,%fr23
+	xor		%r0,%r0,%r31
+	fcpy,dbl	%fr0,%fr24
+	fcpy,dbl	%fr0,%fr25
+	fcpy,dbl	%fr0,%fr26
+	fcpy,dbl	%fr0,%fr27
+	fcpy,dbl	%fr0,%fr28
+	fcpy,dbl	%fr0,%fr29
+	fcpy,dbl	%fr0,%fr30
+	fcpy,dbl	%fr0,%fr31
+	bv		($rp)
+	.EXIT
+	ldo		0($sp),$rv
+	.PROCEND
+___
+{
+my $inp="%r26";
+my $len="%r25";
+
+$code.=<<___;
+	.EXPORT	OPENSSL_cleanse,ENTRY,ARGW0=GR,ARGW1=GR
+	.ALIGN	8
+OPENSSL_cleanse
+	.PROC
+	.CALLINFO	NO_CALLS
+	.ENTRY
+	cmpib,*=	0,$len,Ldone
+	nop
+	cmpib,*>>=	15,$len,Little
+	ldi		$SIZE_T-1,%r1
+
+Lalign
+	and,*<>		$inp,%r1,%r28
+	b,n		Laligned
+	stb		%r0,0($inp)
+	ldo		-1($len),$len
+	b		Lalign
+	ldo		1($inp),$inp
+
+Laligned
+	andcm		$len,%r1,%r28
+Lot
+	$ST		%r0,0($inp)
+	addib,*<>	-$SIZE_T,%r28,Lot
+	ldo		$SIZE_T($inp),$inp
+
+	and,*<>		$len,%r1,$len
+	b,n		Ldone
+Little
+	stb		%r0,0($inp)
+	addib,*<>	-1,$len,Little
+	ldo		1($inp),$inp
+Ldone
+	bv		($rp)
+	.EXIT
+	nop
+	.PROCEND
+___
+}
+{
+my ($out,$cnt,$max)=("%r26","%r25","%r24");
+my ($tick,$lasttick)=("%r23","%r22");
+my ($diff,$lastdiff)=("%r21","%r20");
+
+$code.=<<___;
+	.EXPORT	OPENSSL_instrument_bus,ENTRY,ARGW0=GR,ARGW1=GR
+	.ALIGN	8
+OPENSSL_instrument_bus
+	.PROC
+	.CALLINFO	NO_CALLS
+	.ENTRY
+	copy		$cnt,$rv
+	mfctl		%cr16,$tick
+	copy		$tick,$lasttick
+	ldi		0,$diff
+
+	fdc		0($out)
+	ldw		0($out),$tick
+	add		$diff,$tick,$tick
+	stw		$tick,0($out)
+Loop
+	mfctl		%cr16,$tick
+	sub		$tick,$lasttick,$diff
+	copy		$tick,$lasttick
+
+	fdc		0($out)
+	ldw		0($out),$tick
+	add		$diff,$tick,$tick
+	stw		$tick,0($out)
+
+	addib,<>	-1,$cnt,Loop
+	addi		4,$out,$out
+
+	bv		($rp)
+	.EXIT
+	sub		$rv,$cnt,$rv
+	.PROCEND
+
+	.EXPORT	OPENSSL_instrument_bus2,ENTRY,ARGW0=GR,ARGW1=GR
+	.ALIGN	8
+OPENSSL_instrument_bus2
+	.PROC
+	.CALLINFO	NO_CALLS
+	.ENTRY
+	copy		$cnt,$rv
+	sub		%r0,$cnt,$cnt
+
+	mfctl		%cr16,$tick
+	copy		$tick,$lasttick
+	ldi		0,$diff
+
+	fdc		0($out)
+	ldw		0($out),$tick
+	add		$diff,$tick,$tick
+	stw		$tick,0($out)
+
+	mfctl		%cr16,$tick
+	sub		$tick,$lasttick,$diff
+	copy		$tick,$lasttick
+Loop2
+	copy		$diff,$lastdiff
+	fdc		0($out)
+	ldw		0($out),$tick
+	add		$diff,$tick,$tick
+	stw		$tick,0($out)
+
+	addib,=		-1,$max,Ldone2
+	nop
+
+	mfctl		%cr16,$tick
+	sub		$tick,$lasttick,$diff
+	copy		$tick,$lasttick
+	cmpclr,<>	$lastdiff,$diff,$tick
+	ldi		1,$tick
+
+	ldi		1,%r1
+	xor		%r1,$tick,$tick
+	addb,<>		$tick,$cnt,Loop2
+	shladd,l	$tick,2,$out,$out
+Ldone2
+	bv		($rp)
+	.EXIT
+	add		$rv,$cnt,$rv
+	.PROCEND
+___
+}
+$code =~ s/cmpib,\*/comib,/gm if ($SIZE_T==4);
+$code =~ s/,\*/,/gm if ($SIZE_T==4);
+print $code;
+close STDOUT;
+
diff --git a/crypto/pem/Makefile b/crypto/pem/Makefile
index 669f36612c4e..2cc7801529e0 100644
--- a/crypto/pem/Makefile
+++ b/crypto/pem/Makefile
@@ -18,10 +18,10 @@ APPS=
 
 LIB=$(TOP)/libcrypto.a
 LIBSRC= pem_sign.c pem_seal.c pem_info.c pem_lib.c pem_all.c pem_err.c \
-	pem_x509.c pem_xaux.c pem_oth.c pem_pk8.c pem_pkey.c
+	pem_x509.c pem_xaux.c pem_oth.c pem_pk8.c pem_pkey.c pvkfmt.c
 
 LIBOBJ=	pem_sign.o pem_seal.o pem_info.o pem_lib.o pem_all.o pem_err.o \
-	pem_x509.o pem_xaux.o pem_oth.o pem_pk8.o pem_pkey.o
+	pem_x509.o pem_xaux.o pem_oth.o pem_pk8.o pem_pkey.o pvkfmt.o
 
 SRC= $(LIBSRC)
 
@@ -36,7 +36,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -83,39 +83,36 @@ pem_all.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
 pem_all.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 pem_all.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pem_all.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pem_all.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-pem_all.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-pem_all.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-pem_all.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-pem_all.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
-pem_all.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
-pem_all.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-pem_all.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-pem_all.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-pem_all.o: ../cryptlib.h pem_all.c
+pem_all.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pem_all.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pem_all.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+pem_all.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
+pem_all.o: ../../include/openssl/pem2.h ../../include/openssl/pkcs7.h
+pem_all.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+pem_all.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+pem_all.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+pem_all.o: ../../include/openssl/x509_vfy.h ../cryptlib.h pem_all.c
 pem_err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 pem_err.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 pem_err.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pem_err.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pem_err.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pem_err.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pem_err.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pem_err.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-pem_err.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
-pem_err.o: ../../include/openssl/pem2.h ../../include/openssl/pkcs7.h
-pem_err.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-pem_err.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-pem_err.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-pem_err.o: pem_err.c
+pem_err.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pem_err.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pem_err.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+pem_err.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
+pem_err.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+pem_err.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+pem_err.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+pem_err.o: ../../include/openssl/x509_vfy.h pem_err.c
 pem_info.o: ../../e_os.h ../../include/openssl/asn1.h
 pem_info.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pem_info.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
 pem_info.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pem_info.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pem_info.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pem_info.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pem_info.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pem_info.o: ../../include/openssl/opensslconf.h
+pem_info.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pem_info.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 pem_info.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pem_info.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
 pem_info.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
@@ -128,8 +125,8 @@ pem_lib.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pem_lib.o: ../../include/openssl/crypto.h ../../include/openssl/des.h
 pem_lib.o: ../../include/openssl/des_old.h ../../include/openssl/e_os2.h
 pem_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
-pem_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pem_lib.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
+pem_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
+pem_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
 pem_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
 pem_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 pem_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
@@ -139,43 +136,43 @@ pem_lib.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
 pem_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 pem_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
 pem_lib.o: ../../include/openssl/ui_compat.h ../../include/openssl/x509.h
-pem_lib.o: ../../include/openssl/x509_vfy.h ../cryptlib.h pem_lib.c
+pem_lib.o: ../../include/openssl/x509_vfy.h ../asn1/asn1_locl.h ../cryptlib.h
+pem_lib.o: pem_lib.c
 pem_oth.o: ../../e_os.h ../../include/openssl/asn1.h
 pem_oth.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pem_oth.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pem_oth.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pem_oth.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pem_oth.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-pem_oth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-pem_oth.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-pem_oth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-pem_oth.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
-pem_oth.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
-pem_oth.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-pem_oth.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-pem_oth.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-pem_oth.o: ../cryptlib.h pem_oth.c
+pem_oth.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pem_oth.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pem_oth.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+pem_oth.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
+pem_oth.o: ../../include/openssl/pem2.h ../../include/openssl/pkcs7.h
+pem_oth.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
+pem_oth.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+pem_oth.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+pem_oth.o: ../../include/openssl/x509_vfy.h ../cryptlib.h pem_oth.c
 pem_pk8.o: ../../e_os.h ../../include/openssl/asn1.h
 pem_pk8.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pem_pk8.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pem_pk8.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pem_pk8.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pem_pk8.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-pem_pk8.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-pem_pk8.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-pem_pk8.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-pem_pk8.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
-pem_pk8.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
-pem_pk8.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-pem_pk8.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-pem_pk8.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-pem_pk8.o: ../../include/openssl/x509_vfy.h ../cryptlib.h pem_pk8.c
+pem_pk8.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pem_pk8.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pem_pk8.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+pem_pk8.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
+pem_pk8.o: ../../include/openssl/pem2.h ../../include/openssl/pkcs12.h
+pem_pk8.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+pem_pk8.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+pem_pk8.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+pem_pk8.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+pem_pk8.o: ../cryptlib.h pem_pk8.c
 pem_pkey.o: ../../e_os.h ../../include/openssl/asn1.h
 pem_pkey.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pem_pkey.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pem_pkey.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
-pem_pkey.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pem_pkey.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
+pem_pkey.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
+pem_pkey.o: ../../include/openssl/err.h ../../include/openssl/evp.h
 pem_pkey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
 pem_pkey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 pem_pkey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
@@ -184,15 +181,16 @@ pem_pkey.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
 pem_pkey.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
 pem_pkey.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 pem_pkey.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-pem_pkey.o: ../../include/openssl/x509_vfy.h ../cryptlib.h pem_pkey.c
+pem_pkey.o: ../../include/openssl/x509_vfy.h ../asn1/asn1_locl.h ../cryptlib.h
+pem_pkey.o: pem_pkey.c
 pem_seal.o: ../../e_os.h ../../include/openssl/asn1.h
 pem_seal.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pem_seal.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pem_seal.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pem_seal.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pem_seal.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-pem_seal.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-pem_seal.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pem_seal.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pem_seal.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pem_seal.o: ../../include/openssl/opensslconf.h
 pem_seal.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pem_seal.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
 pem_seal.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
@@ -205,9 +203,9 @@ pem_sign.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pem_sign.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pem_sign.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pem_sign.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pem_sign.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-pem_sign.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-pem_sign.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pem_sign.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pem_sign.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pem_sign.o: ../../include/openssl/opensslconf.h
 pem_sign.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pem_sign.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
 pem_sign.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
@@ -220,9 +218,9 @@ pem_x509.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pem_x509.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pem_x509.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pem_x509.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pem_x509.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-pem_x509.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-pem_x509.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pem_x509.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pem_x509.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pem_x509.o: ../../include/openssl/opensslconf.h
 pem_x509.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pem_x509.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
 pem_x509.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
@@ -234,12 +232,27 @@ pem_xaux.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pem_xaux.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pem_xaux.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pem_xaux.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pem_xaux.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-pem_xaux.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-pem_xaux.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pem_xaux.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pem_xaux.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pem_xaux.o: ../../include/openssl/opensslconf.h
 pem_xaux.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pem_xaux.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
 pem_xaux.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 pem_xaux.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 pem_xaux.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 pem_xaux.o: ../../include/openssl/x509_vfy.h ../cryptlib.h pem_xaux.c
+pvkfmt.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
+pvkfmt.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
+pvkfmt.o: ../../include/openssl/crypto.h ../../include/openssl/dsa.h
+pvkfmt.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+pvkfmt.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+pvkfmt.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+pvkfmt.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pvkfmt.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pvkfmt.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+pvkfmt.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
+pvkfmt.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+pvkfmt.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+pvkfmt.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+pvkfmt.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+pvkfmt.o: ../../include/openssl/x509_vfy.h ../cryptlib.h pvkfmt.c
diff --git a/crypto/pem/pem.h b/crypto/pem/pem.h
index 6c193f1cbfa4..8a6ababe3a84 100644
--- a/crypto/pem/pem.h
+++ b/crypto/pem/pem.h
@@ -134,6 +134,7 @@ extern "C" {
 #define PEM_STRING_ECDSA_PUBLIC "ECDSA PUBLIC KEY"
 #define PEM_STRING_ECPARAMETERS "EC PARAMETERS"
 #define PEM_STRING_ECPRIVATEKEY	"EC PRIVATE KEY"
+#define PEM_STRING_PARAMETERS	"PARAMETERS"
 #define PEM_STRING_CMS		"CMS"
 
   /* Note that this structure is initialised by PEM_SealInit and cleaned up
@@ -183,11 +184,8 @@ typedef struct pem_ctx_st
 	int num_recipient;
 	PEM_USER **recipient;
 
-#ifndef OPENSSL_NO_STACK
-	STACK *x509_chain;	/* certificate chain */
-#else
-	char *x509_chain;	/* certificate chain */
-#endif
+	/* XXX(ben): don#t think this is used! 
+		STACK *x509_chain;	/ * certificate chain */
 	EVP_MD *md;		/* signature type */
 
 	int md_enc;		/* is the md encrypted or not? */
@@ -224,28 +222,19 @@ typedef struct pem_ctx_st
 #define IMPLEMENT_PEM_read_fp(name, type, str, asn1) \
 type *PEM_read_##name(FILE *fp, type **x, pem_password_cb *cb, void *u)\
 { \
-    return (type*)PEM_ASN1_read(CHECKED_D2I_OF(type, d2i_##asn1), \
-				str, fp, \
-				CHECKED_PPTR_OF(type, x), \
-				cb, u); \
+return PEM_ASN1_read((d2i_of_void *)d2i_##asn1, str,fp,(void **)x,cb,u); \
 } 
 
 #define IMPLEMENT_PEM_write_fp(name, type, str, asn1) \
 int PEM_write_##name(FILE *fp, type *x) \
 { \
-    return PEM_ASN1_write(CHECKED_I2D_OF(type, i2d_##asn1), \
-			  str, fp, \
-			  CHECKED_PTR_OF(type, x), \
-			  NULL, NULL, 0, NULL, NULL); \
+return PEM_ASN1_write((i2d_of_void *)i2d_##asn1,str,fp,x,NULL,NULL,0,NULL,NULL); \
 }
 
 #define IMPLEMENT_PEM_write_fp_const(name, type, str, asn1) \
 int PEM_write_##name(FILE *fp, const type *x) \
 { \
-    return PEM_ASN1_write(CHECKED_I2D_OF(const type, i2d_##asn1), \
-			  str, fp, \
-			  CHECKED_PTR_OF(const type, x), \
-			  NULL, NULL, 0, NULL, NULL); \
+return PEM_ASN1_write((i2d_of_void *)i2d_##asn1,str,fp,(void *)x,NULL,NULL,0,NULL,NULL); \
 }
 
 #define IMPLEMENT_PEM_write_cb_fp(name, type, str, asn1) \
@@ -253,10 +242,7 @@ int PEM_write_##name(FILE *fp, type *x, const EVP_CIPHER *enc, \
 	     unsigned char *kstr, int klen, pem_password_cb *cb, \
 		  void *u) \
 	{ \
-	    return PEM_ASN1_write(CHECKED_I2D_OF(type, i2d_##asn1), \
-				  str, fp, \
-				  CHECKED_PTR_OF(type, x), \
-				  enc, kstr, klen, cb, u); \
+	return PEM_ASN1_write((i2d_of_void *)i2d_##asn1,str,fp,x,enc,kstr,klen,cb,u); \
 	}
 
 #define IMPLEMENT_PEM_write_cb_fp_const(name, type, str, asn1) \
@@ -264,10 +250,7 @@ int PEM_write_##name(FILE *fp, type *x, const EVP_CIPHER *enc, \
 	     unsigned char *kstr, int klen, pem_password_cb *cb, \
 		  void *u) \
 	{ \
-	    return PEM_ASN1_write(CHECKED_I2D_OF(const type, i2d_##asn1), \
-				  str, fp, \
-				  CHECKED_PTR_OF(const type, x), \
-				  enc, kstr, klen, cb, u); \
+	return PEM_ASN1_write((i2d_of_void *)i2d_##asn1,str,fp,x,enc,kstr,klen,cb,u); \
 	}
 
 #endif
@@ -275,48 +258,33 @@ int PEM_write_##name(FILE *fp, type *x, const EVP_CIPHER *enc, \
 #define IMPLEMENT_PEM_read_bio(name, type, str, asn1) \
 type *PEM_read_bio_##name(BIO *bp, type **x, pem_password_cb *cb, void *u)\
 { \
-    return (type*)PEM_ASN1_read_bio(CHECKED_D2I_OF(type, d2i_##asn1), \
-				    str, bp, \
-				    CHECKED_PPTR_OF(type, x), \
-				    cb, u); \
+return PEM_ASN1_read_bio((d2i_of_void *)d2i_##asn1, str,bp,(void **)x,cb,u); \
 }
 
 #define IMPLEMENT_PEM_write_bio(name, type, str, asn1) \
 int PEM_write_bio_##name(BIO *bp, type *x) \
 { \
-    return PEM_ASN1_write_bio(CHECKED_I2D_OF(type, i2d_##asn1), \
-			      str, bp, \
-			      CHECKED_PTR_OF(type, x), \
-			      NULL, NULL, 0, NULL, NULL); \
+return PEM_ASN1_write_bio((i2d_of_void *)i2d_##asn1,str,bp,x,NULL,NULL,0,NULL,NULL); \
 }
 
 #define IMPLEMENT_PEM_write_bio_const(name, type, str, asn1) \
 int PEM_write_bio_##name(BIO *bp, const type *x) \
 { \
-    return PEM_ASN1_write_bio(CHECKED_I2D_OF(const type, i2d_##asn1), \
-			      str, bp, \
-			      CHECKED_PTR_OF(const type, x), \
-			      NULL, NULL, 0, NULL, NULL); \
+return PEM_ASN1_write_bio((i2d_of_void *)i2d_##asn1,str,bp,(void *)x,NULL,NULL,0,NULL,NULL); \
 }
 
 #define IMPLEMENT_PEM_write_cb_bio(name, type, str, asn1) \
 int PEM_write_bio_##name(BIO *bp, type *x, const EVP_CIPHER *enc, \
 	     unsigned char *kstr, int klen, pem_password_cb *cb, void *u) \
 	{ \
-	    return PEM_ASN1_write_bio(CHECKED_I2D_OF(type, i2d_##asn1), \
-				      str, bp, \
-				      CHECKED_PTR_OF(type, x), \
-				      enc, kstr, klen, cb, u); \
+	return PEM_ASN1_write_bio((i2d_of_void *)i2d_##asn1,str,bp,x,enc,kstr,klen,cb,u); \
 	}
 
 #define IMPLEMENT_PEM_write_cb_bio_const(name, type, str, asn1) \
 int PEM_write_bio_##name(BIO *bp, type *x, const EVP_CIPHER *enc, \
 	     unsigned char *kstr, int klen, pem_password_cb *cb, void *u) \
 	{ \
-	    return PEM_ASN1_write_bio(CHECKED_I2D_OF(const type, i2d_##asn1), \
-				      str, bp, \
-				      CHECKED_PTR_OF(const type, x), \
-				      enc, kstr, klen, cb, u); \
+	return PEM_ASN1_write_bio((i2d_of_void *)i2d_##asn1,str,bp,(void *)x,enc,kstr,klen,cb,u); \
 	}
 
 #define IMPLEMENT_PEM_write(name, type, str, asn1) \
@@ -353,11 +321,10 @@ int PEM_write_bio_##name(BIO *bp, type *x, const EVP_CIPHER *enc, \
 
 /* These are the same except they are for the declarations */
 
-#if defined(OPENSSL_SYS_WIN16) || defined(OPENSSL_NO_FP_API)
+#if defined(OPENSSL_NO_FP_API)
 
 #define DECLARE_PEM_read_fp(name, type) /**/
 #define DECLARE_PEM_write_fp(name, type) /**/
-#define DECLARE_PEM_write_fp_const(name, type) /**/
 #define DECLARE_PEM_write_cb_fp(name, type) /**/
 
 #else
@@ -428,138 +395,6 @@ int PEM_write_bio_##name(BIO *bp, type *x, const EVP_CIPHER *enc, \
 	DECLARE_PEM_read(name, type) \
 	DECLARE_PEM_write_cb(name, type)
 
-#ifdef SSLEAY_MACROS
-
-#define PEM_write_SSL_SESSION(fp,x) \
-		PEM_ASN1_write((int (*)())i2d_SSL_SESSION, \
-			PEM_STRING_SSL_SESSION,fp, (char *)x, NULL,NULL,0,NULL,NULL)
-#define PEM_write_X509(fp,x) \
-		PEM_ASN1_write((int (*)())i2d_X509,PEM_STRING_X509,fp, \
-			(char *)x, NULL,NULL,0,NULL,NULL)
-#define PEM_write_X509_REQ(fp,x) PEM_ASN1_write( \
-		(int (*)())i2d_X509_REQ,PEM_STRING_X509_REQ,fp,(char *)x, \
-			NULL,NULL,0,NULL,NULL)
-#define PEM_write_X509_CRL(fp,x) \
-		PEM_ASN1_write((int (*)())i2d_X509_CRL,PEM_STRING_X509_CRL, \
-			fp,(char *)x, NULL,NULL,0,NULL,NULL)
-#define	PEM_write_RSAPrivateKey(fp,x,enc,kstr,klen,cb,u) \
-		PEM_ASN1_write((int (*)())i2d_RSAPrivateKey,PEM_STRING_RSA,fp,\
-			(char *)x,enc,kstr,klen,cb,u)
-#define	PEM_write_RSAPublicKey(fp,x) \
-		PEM_ASN1_write((int (*)())i2d_RSAPublicKey,\
-			PEM_STRING_RSA_PUBLIC,fp,(char *)x,NULL,NULL,0,NULL,NULL)
-#define	PEM_write_DSAPrivateKey(fp,x,enc,kstr,klen,cb,u) \
-		PEM_ASN1_write((int (*)())i2d_DSAPrivateKey,PEM_STRING_DSA,fp,\
-			(char *)x,enc,kstr,klen,cb,u)
-#define	PEM_write_PrivateKey(bp,x,enc,kstr,klen,cb,u) \
-		PEM_ASN1_write((int (*)())i2d_PrivateKey,\
-		(((x)->type == EVP_PKEY_DSA)?PEM_STRING_DSA:PEM_STRING_RSA),\
-			bp,(char *)x,enc,kstr,klen,cb,u)
-#define PEM_write_PKCS7(fp,x) \
-		PEM_ASN1_write((int (*)())i2d_PKCS7,PEM_STRING_PKCS7,fp, \
-			(char *)x, NULL,NULL,0,NULL,NULL)
-#define PEM_write_DHparams(fp,x) \
-		PEM_ASN1_write((int (*)())i2d_DHparams,PEM_STRING_DHPARAMS,fp,\
-			(char *)x,NULL,NULL,0,NULL,NULL)
-
-#define PEM_write_NETSCAPE_CERT_SEQUENCE(fp,x) \
-                PEM_ASN1_write((int (*)())i2d_NETSCAPE_CERT_SEQUENCE, \
-			PEM_STRING_X509,fp, \
-                        (char *)x, NULL,NULL,0,NULL,NULL)
-
-#define	PEM_read_SSL_SESSION(fp,x,cb,u) (SSL_SESSION *)PEM_ASN1_read( \
-	(char *(*)())d2i_SSL_SESSION,PEM_STRING_SSL_SESSION,fp,(char **)x,cb,u)
-#define	PEM_read_X509(fp,x,cb,u) (X509 *)PEM_ASN1_read( \
-	(char *(*)())d2i_X509,PEM_STRING_X509,fp,(char **)x,cb,u)
-#define	PEM_read_X509_REQ(fp,x,cb,u) (X509_REQ *)PEM_ASN1_read( \
-	(char *(*)())d2i_X509_REQ,PEM_STRING_X509_REQ,fp,(char **)x,cb,u)
-#define	PEM_read_X509_CRL(fp,x,cb,u) (X509_CRL *)PEM_ASN1_read( \
-	(char *(*)())d2i_X509_CRL,PEM_STRING_X509_CRL,fp,(char **)x,cb,u)
-#define	PEM_read_RSAPrivateKey(fp,x,cb,u) (RSA *)PEM_ASN1_read( \
-	(char *(*)())d2i_RSAPrivateKey,PEM_STRING_RSA,fp,(char **)x,cb,u)
-#define	PEM_read_RSAPublicKey(fp,x,cb,u) (RSA *)PEM_ASN1_read( \
-	(char *(*)())d2i_RSAPublicKey,PEM_STRING_RSA_PUBLIC,fp,(char **)x,cb,u)
-#define	PEM_read_DSAPrivateKey(fp,x,cb,u) (DSA *)PEM_ASN1_read( \
-	(char *(*)())d2i_DSAPrivateKey,PEM_STRING_DSA,fp,(char **)x,cb,u)
-#define	PEM_read_PrivateKey(fp,x,cb,u) (EVP_PKEY *)PEM_ASN1_read( \
-	(char *(*)())d2i_PrivateKey,PEM_STRING_EVP_PKEY,fp,(char **)x,cb,u)
-#define	PEM_read_PKCS7(fp,x,cb,u) (PKCS7 *)PEM_ASN1_read( \
-	(char *(*)())d2i_PKCS7,PEM_STRING_PKCS7,fp,(char **)x,cb,u)
-#define	PEM_read_DHparams(fp,x,cb,u) (DH *)PEM_ASN1_read( \
-	(char *(*)())d2i_DHparams,PEM_STRING_DHPARAMS,fp,(char **)x,cb,u)
-
-#define PEM_read_NETSCAPE_CERT_SEQUENCE(fp,x,cb,u) \
-		(NETSCAPE_CERT_SEQUENCE *)PEM_ASN1_read( \
-        (char *(*)())d2i_NETSCAPE_CERT_SEQUENCE,PEM_STRING_X509,fp,\
-							(char **)x,cb,u)
-
-#define PEM_write_bio_X509(bp,x) \
-		PEM_ASN1_write_bio((int (*)())i2d_X509,PEM_STRING_X509,bp, \
-			(char *)x, NULL,NULL,0,NULL,NULL)
-#define PEM_write_bio_X509_REQ(bp,x) PEM_ASN1_write_bio( \
-		(int (*)())i2d_X509_REQ,PEM_STRING_X509_REQ,bp,(char *)x, \
-			NULL,NULL,0,NULL,NULL)
-#define PEM_write_bio_X509_CRL(bp,x) \
-		PEM_ASN1_write_bio((int (*)())i2d_X509_CRL,PEM_STRING_X509_CRL,\
-			bp,(char *)x, NULL,NULL,0,NULL,NULL)
-#define	PEM_write_bio_RSAPrivateKey(bp,x,enc,kstr,klen,cb,u) \
-		PEM_ASN1_write_bio((int (*)())i2d_RSAPrivateKey,PEM_STRING_RSA,\
-			bp,(char *)x,enc,kstr,klen,cb,u)
-#define	PEM_write_bio_RSAPublicKey(bp,x) \
-		PEM_ASN1_write_bio((int (*)())i2d_RSAPublicKey, \
-			PEM_STRING_RSA_PUBLIC,\
-			bp,(char *)x,NULL,NULL,0,NULL,NULL)
-#define	PEM_write_bio_DSAPrivateKey(bp,x,enc,kstr,klen,cb,u) \
-		PEM_ASN1_write_bio((int (*)())i2d_DSAPrivateKey,PEM_STRING_DSA,\
-			bp,(char *)x,enc,kstr,klen,cb,u)
-#define	PEM_write_bio_PrivateKey(bp,x,enc,kstr,klen,cb,u) \
-		PEM_ASN1_write_bio((int (*)())i2d_PrivateKey,\
-		(((x)->type == EVP_PKEY_DSA)?PEM_STRING_DSA:PEM_STRING_RSA),\
-			bp,(char *)x,enc,kstr,klen,cb,u)
-#define PEM_write_bio_PKCS7(bp,x) \
-		PEM_ASN1_write_bio((int (*)())i2d_PKCS7,PEM_STRING_PKCS7,bp, \
-			(char *)x, NULL,NULL,0,NULL,NULL)
-#define PEM_write_bio_DHparams(bp,x) \
-		PEM_ASN1_write_bio((int (*)())i2d_DHparams,PEM_STRING_DHPARAMS,\
-			bp,(char *)x,NULL,NULL,0,NULL,NULL)
-#define PEM_write_bio_DSAparams(bp,x) \
-		PEM_ASN1_write_bio((int (*)())i2d_DSAparams, \
-			PEM_STRING_DSAPARAMS,bp,(char *)x,NULL,NULL,0,NULL,NULL)
-
-#define PEM_write_bio_NETSCAPE_CERT_SEQUENCE(bp,x) \
-                PEM_ASN1_write_bio((int (*)())i2d_NETSCAPE_CERT_SEQUENCE, \
-			PEM_STRING_X509,bp, \
-                        (char *)x, NULL,NULL,0,NULL,NULL)
-
-#define	PEM_read_bio_X509(bp,x,cb,u) (X509 *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_X509,PEM_STRING_X509,bp,(char **)x,cb,u)
-#define	PEM_read_bio_X509_REQ(bp,x,cb,u) (X509_REQ *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_X509_REQ,PEM_STRING_X509_REQ,bp,(char **)x,cb,u)
-#define	PEM_read_bio_X509_CRL(bp,x,cb,u) (X509_CRL *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_X509_CRL,PEM_STRING_X509_CRL,bp,(char **)x,cb,u)
-#define	PEM_read_bio_RSAPrivateKey(bp,x,cb,u) (RSA *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_RSAPrivateKey,PEM_STRING_RSA,bp,(char **)x,cb,u)
-#define	PEM_read_bio_RSAPublicKey(bp,x,cb,u) (RSA *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_RSAPublicKey,PEM_STRING_RSA_PUBLIC,bp,(char **)x,cb,u)
-#define	PEM_read_bio_DSAPrivateKey(bp,x,cb,u) (DSA *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_DSAPrivateKey,PEM_STRING_DSA,bp,(char **)x,cb,u)
-#define	PEM_read_bio_PrivateKey(bp,x,cb,u) (EVP_PKEY *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_PrivateKey,PEM_STRING_EVP_PKEY,bp,(char **)x,cb,u)
-
-#define	PEM_read_bio_PKCS7(bp,x,cb,u) (PKCS7 *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_PKCS7,PEM_STRING_PKCS7,bp,(char **)x,cb,u)
-#define	PEM_read_bio_DHparams(bp,x,cb,u) (DH *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_DHparams,PEM_STRING_DHPARAMS,bp,(char **)x,cb,u)
-#define	PEM_read_bio_DSAparams(bp,x,cb,u) (DSA *)PEM_ASN1_read_bio( \
-	(char *(*)())d2i_DSAparams,PEM_STRING_DSAPARAMS,bp,(char **)x,cb,u)
-
-#define PEM_read_bio_NETSCAPE_CERT_SEQUENCE(bp,x,cb,u) \
-		(NETSCAPE_CERT_SEQUENCE *)PEM_ASN1_read_bio( \
-        (char *(*)())d2i_NETSCAPE_CERT_SEQUENCE,PEM_STRING_X509,bp,\
-							(char **)x,cb,u)
-
-#endif
-
 #if 1
 /* "userdata": new with OpenSSL 0.9.4 */
 typedef int pem_password_cb(char *buf, int size, int rwflag, void *userdata);
@@ -581,40 +416,25 @@ int PEM_bytes_read_bio(unsigned char **pdata, long *plen, char **pnm, const char
 	     pem_password_cb *cb, void *u);
 void *	PEM_ASN1_read_bio(d2i_of_void *d2i, const char *name, BIO *bp,
 			  void **x, pem_password_cb *cb, void *u);
-
-#define PEM_ASN1_read_bio_of(type,d2i,name,bp,x,cb,u) \
-    ((type*)PEM_ASN1_read_bio(CHECKED_D2I_OF(type, d2i), \
-			      name, bp,			\
-			      CHECKED_PPTR_OF(type, x), \
-			      cb, u))
-
-int	PEM_ASN1_write_bio(i2d_of_void *i2d,const char *name,BIO *bp,char *x,
+int	PEM_ASN1_write_bio(i2d_of_void *i2d,const char *name,BIO *bp, void *x,
 			   const EVP_CIPHER *enc,unsigned char *kstr,int klen,
 			   pem_password_cb *cb, void *u);
 
-#define PEM_ASN1_write_bio_of(type,i2d,name,bp,x,enc,kstr,klen,cb,u) \
-    (PEM_ASN1_write_bio(CHECKED_I2D_OF(type, i2d), \
-			name, bp,		   \
-			CHECKED_PTR_OF(type, x), \
-			enc, kstr, klen, cb, u))
-
 STACK_OF(X509_INFO) *	PEM_X509_INFO_read_bio(BIO *bp, STACK_OF(X509_INFO) *sk, pem_password_cb *cb, void *u);
 int	PEM_X509_INFO_write_bio(BIO *bp,X509_INFO *xi, EVP_CIPHER *enc,
 		unsigned char *kstr, int klen, pem_password_cb *cd, void *u);
 #endif
 
-#ifndef OPENSSL_SYS_WIN16
 int	PEM_read(FILE *fp, char **name, char **header,
 		unsigned char **data,long *len);
 int	PEM_write(FILE *fp,char *name,char *hdr,unsigned char *data,long len);
 void *  PEM_ASN1_read(d2i_of_void *d2i, const char *name, FILE *fp, void **x,
 		      pem_password_cb *cb, void *u);
 int	PEM_ASN1_write(i2d_of_void *i2d,const char *name,FILE *fp,
-		       char *x,const EVP_CIPHER *enc,unsigned char *kstr,
+		       void *x,const EVP_CIPHER *enc,unsigned char *kstr,
 		       int klen,pem_password_cb *callback, void *u);
 STACK_OF(X509_INFO) *	PEM_X509_INFO_read(FILE *fp, STACK_OF(X509_INFO) *sk,
 	pem_password_cb *cb, void *u);
-#endif
 
 int	PEM_SealInit(PEM_ENCODE_SEAL_CTX *ctx, EVP_CIPHER *type,
 		EVP_MD *md_type, unsigned char **ek, int *ekl,
@@ -633,7 +453,6 @@ int	PEM_def_callback(char *buf, int num, int w, void *key);
 void	PEM_proc_type(char *buf, int type);
 void	PEM_dek_info(char *buf, const char *type, int len, char *str);
 
-#ifndef SSLEAY_MACROS
 
 #include <openssl/symhacks.h>
 
@@ -719,7 +538,21 @@ EVP_PKEY *d2i_PKCS8PrivateKey_fp(FILE *fp, EVP_PKEY **x, pem_password_cb *cb, vo
 int PEM_write_PKCS8PrivateKey(FILE *fp,EVP_PKEY *x,const EVP_CIPHER *enc,
 			      char *kstr,int klen, pem_password_cb *cd, void *u);
 
-#endif /* SSLEAY_MACROS */
+EVP_PKEY *PEM_read_bio_Parameters(BIO *bp, EVP_PKEY **x);
+int PEM_write_bio_Parameters(BIO *bp, EVP_PKEY *x);
+
+
+EVP_PKEY *b2i_PrivateKey(const unsigned char **in, long length);
+EVP_PKEY *b2i_PublicKey(const unsigned char **in, long length);
+EVP_PKEY *b2i_PrivateKey_bio(BIO *in);
+EVP_PKEY *b2i_PublicKey_bio(BIO *in);
+int i2b_PrivateKey_bio(BIO *out, EVP_PKEY *pk);
+int i2b_PublicKey_bio(BIO *out, EVP_PKEY *pk);
+#ifndef OPENSSL_NO_RC4
+EVP_PKEY *b2i_PVK_bio(BIO *in, pem_password_cb *cb, void *u);
+int i2b_PVK_bio(BIO *out, EVP_PKEY *pk, int enclevel,
+		pem_password_cb *cb, void *u);
+#endif
 
 
 /* BEGIN ERROR CODES */
@@ -731,10 +564,22 @@ void ERR_load_PEM_strings(void);
 /* Error codes for the PEM functions. */
 
 /* Function codes. */
+#define PEM_F_B2I_DSS					 127
+#define PEM_F_B2I_PVK_BIO				 128
+#define PEM_F_B2I_RSA					 129
+#define PEM_F_CHECK_BITLEN_DSA				 130
+#define PEM_F_CHECK_BITLEN_RSA				 131
 #define PEM_F_D2I_PKCS8PRIVATEKEY_BIO			 120
 #define PEM_F_D2I_PKCS8PRIVATEKEY_FP			 121
+#define PEM_F_DO_B2I					 132
+#define PEM_F_DO_B2I_BIO				 133
+#define PEM_F_DO_BLOB_HEADER				 134
 #define PEM_F_DO_PK8PKEY				 126
 #define PEM_F_DO_PK8PKEY_FP				 125
+#define PEM_F_DO_PVK_BODY				 135
+#define PEM_F_DO_PVK_HEADER				 136
+#define PEM_F_I2B_PVK					 137
+#define PEM_F_I2B_PVK_BIO				 138
 #define PEM_F_LOAD_IV					 101
 #define PEM_F_PEM_ASN1_READ				 102
 #define PEM_F_PEM_ASN1_READ_BIO				 103
@@ -747,6 +592,7 @@ void ERR_load_PEM_strings(void);
 #define PEM_F_PEM_PK8PKEY				 119
 #define PEM_F_PEM_READ					 108
 #define PEM_F_PEM_READ_BIO				 109
+#define PEM_F_PEM_READ_BIO_PARAMETERS			 140
 #define PEM_F_PEM_READ_BIO_PRIVATEKEY			 123
 #define PEM_F_PEM_READ_PRIVATEKEY			 124
 #define PEM_F_PEM_SEALFINAL				 110
@@ -754,6 +600,7 @@ void ERR_load_PEM_strings(void);
 #define PEM_F_PEM_SIGNFINAL				 112
 #define PEM_F_PEM_WRITE					 113
 #define PEM_F_PEM_WRITE_BIO				 114
+#define PEM_F_PEM_WRITE_PRIVATEKEY			 139
 #define PEM_F_PEM_X509_INFO_READ			 115
 #define PEM_F_PEM_X509_INFO_READ_BIO			 116
 #define PEM_F_PEM_X509_INFO_WRITE_BIO			 117
@@ -763,18 +610,30 @@ void ERR_load_PEM_strings(void);
 #define PEM_R_BAD_DECRYPT				 101
 #define PEM_R_BAD_END_LINE				 102
 #define PEM_R_BAD_IV_CHARS				 103
+#define PEM_R_BAD_MAGIC_NUMBER				 116
 #define PEM_R_BAD_PASSWORD_READ				 104
+#define PEM_R_BAD_VERSION_NUMBER			 117
+#define PEM_R_BIO_WRITE_FAILURE				 118
+#define PEM_R_CIPHER_IS_NULL				 127
 #define PEM_R_ERROR_CONVERTING_PRIVATE_KEY		 115
+#define PEM_R_EXPECTING_PRIVATE_KEY_BLOB		 119
+#define PEM_R_EXPECTING_PUBLIC_KEY_BLOB			 120
+#define PEM_R_INCONSISTENT_HEADER			 121
+#define PEM_R_KEYBLOB_HEADER_PARSE_ERROR		 122
+#define PEM_R_KEYBLOB_TOO_SHORT				 123
 #define PEM_R_NOT_DEK_INFO				 105
 #define PEM_R_NOT_ENCRYPTED				 106
 #define PEM_R_NOT_PROC_TYPE				 107
 #define PEM_R_NO_START_LINE				 108
 #define PEM_R_PROBLEMS_GETTING_PASSWORD			 109
 #define PEM_R_PUBLIC_KEY_NO_RSA				 110
+#define PEM_R_PVK_DATA_TOO_SHORT			 124
+#define PEM_R_PVK_TOO_SHORT				 125
 #define PEM_R_READ_KEY					 111
 #define PEM_R_SHORT_HEADER				 112
 #define PEM_R_UNSUPPORTED_CIPHER			 113
 #define PEM_R_UNSUPPORTED_ENCRYPTION			 114
+#define PEM_R_UNSUPPORTED_KEY_COMPONENTS		 126
 
 #ifdef  __cplusplus
 }
diff --git a/crypto/pem/pem_all.c b/crypto/pem/pem_all.c
index 69dd19bf2e76..3e7a6093ad82 100644
--- a/crypto/pem/pem_all.c
+++ b/crypto/pem/pem_all.c
@@ -110,7 +110,6 @@
  */
 
 #include <stdio.h>
-#undef SSLEAY_MACROS
 #include "cryptlib.h"
 #include <openssl/bio.h>
 #include <openssl/evp.h>
@@ -194,49 +193,7 @@ RSA *PEM_read_RSAPrivateKey(FILE *fp, RSA **rsa, pem_password_cb *cb,
 
 #endif
 
-#ifdef OPENSSL_FIPS
-
-int PEM_write_bio_RSAPrivateKey(BIO *bp, RSA *x, const EVP_CIPHER *enc,
-                                               unsigned char *kstr, int klen,
-                                               pem_password_cb *cb, void *u)
-{
-	EVP_PKEY *k;
-	int ret;
-	k = EVP_PKEY_new();
-	if (!k)
-		return 0;
-	EVP_PKEY_set1_RSA(k, x);
-
-	ret = PEM_write_bio_PrivateKey(bp, k, enc, kstr, klen, cb, u);
-	EVP_PKEY_free(k);
-	return ret;
-}
-
-#ifndef OPENSSL_NO_FP_API
-int PEM_write_RSAPrivateKey(FILE *fp, RSA *x, const EVP_CIPHER *enc,
-                                               unsigned char *kstr, int klen,
-                                               pem_password_cb *cb, void *u)
-{
-	EVP_PKEY *k;
-	int ret;
-	k = EVP_PKEY_new();
-	if (!k)
-		return 0;
-
-	EVP_PKEY_set1_RSA(k, x);
-
-	ret = PEM_write_PrivateKey(fp, k, enc, kstr, klen, cb, u);
-	EVP_PKEY_free(k);
-	return ret;
-}
-#endif
-
-#else
-
 IMPLEMENT_PEM_write_cb_const(RSAPrivateKey, RSA, PEM_STRING_RSA, RSAPrivateKey)
-
-#endif
-
 IMPLEMENT_PEM_rw_const(RSAPublicKey, RSA, PEM_STRING_RSA_PUBLIC, RSAPublicKey)
 IMPLEMENT_PEM_rw(RSA_PUBKEY, RSA, PEM_STRING_PUBLIC, RSA_PUBKEY)
 
@@ -263,50 +220,10 @@ DSA *PEM_read_bio_DSAPrivateKey(BIO *bp, DSA **dsa, pem_password_cb *cb,
 {
 	EVP_PKEY *pktmp;
 	pktmp = PEM_read_bio_PrivateKey(bp, NULL, cb, u);
-	return pkey_get_dsa(pktmp, dsa);
+	return pkey_get_dsa(pktmp, dsa);	/* will free pktmp */
 }
 
-#ifdef OPENSSL_FIPS
-
-int PEM_write_bio_DSAPrivateKey(BIO *bp, DSA *x, const EVP_CIPHER *enc,
-                                               unsigned char *kstr, int klen,
-                                               pem_password_cb *cb, void *u)
-{
-	EVP_PKEY *k;
-	int ret;
-	k = EVP_PKEY_new();
-	if (!k)
-		return 0;
-	EVP_PKEY_set1_DSA(k, x);
-
-	ret = PEM_write_bio_PrivateKey(bp, k, enc, kstr, klen, cb, u);
-	EVP_PKEY_free(k);
-	return ret;
-}
-
-#ifndef OPENSSL_NO_FP_API
-int PEM_write_DSAPrivateKey(FILE *fp, DSA *x, const EVP_CIPHER *enc,
-                                               unsigned char *kstr, int klen,
-                                               pem_password_cb *cb, void *u)
-{
-	EVP_PKEY *k;
-	int ret;
-	k = EVP_PKEY_new();
-	if (!k)
-		return 0;
-	EVP_PKEY_set1_DSA(k, x);
-	ret = PEM_write_PrivateKey(fp, k, enc, kstr, klen, cb, u);
-	EVP_PKEY_free(k);
-	return ret;
-}
-#endif
-
-#else
-
 IMPLEMENT_PEM_write_cb_const(DSAPrivateKey, DSA, PEM_STRING_DSA, DSAPrivateKey)
-
-#endif
-
 IMPLEMENT_PEM_rw(DSA_PUBKEY, DSA, PEM_STRING_PUBLIC, DSA_PUBKEY)
 
 #ifndef OPENSSL_NO_FP_API
@@ -316,7 +233,7 @@ DSA *PEM_read_DSAPrivateKey(FILE *fp, DSA **dsa, pem_password_cb *cb,
 {
 	EVP_PKEY *pktmp;
 	pktmp = PEM_read_PrivateKey(fp, NULL, cb, u);
-	return pkey_get_dsa(pktmp, dsa);
+	return pkey_get_dsa(pktmp, dsa);	/* will free pktmp */
 }
 
 #endif
@@ -347,54 +264,13 @@ EC_KEY *PEM_read_bio_ECPrivateKey(BIO *bp, EC_KEY **key, pem_password_cb *cb,
 {
 	EVP_PKEY *pktmp;
 	pktmp = PEM_read_bio_PrivateKey(bp, NULL, cb, u);
-	return pkey_get_eckey(pktmp, key);
+	return pkey_get_eckey(pktmp, key);	/* will free pktmp */
 }
 
 IMPLEMENT_PEM_rw_const(ECPKParameters, EC_GROUP, PEM_STRING_ECPARAMETERS, ECPKParameters)
 
-
-
-#ifdef OPENSSL_FIPS
-
-int PEM_write_bio_ECPrivateKey(BIO *bp, EC_KEY *x, const EVP_CIPHER *enc,
-                                               unsigned char *kstr, int klen,
-                                               pem_password_cb *cb, void *u)
-{
-	EVP_PKEY *k;
-	int ret;
-	k = EVP_PKEY_new();
-	if (!k)
-		return 0;
-	EVP_PKEY_set1_EC_KEY(k, x);
-
-	ret = PEM_write_bio_PrivateKey(bp, k, enc, kstr, klen, cb, u);
-	EVP_PKEY_free(k);
-	return ret;
-}
-
-#ifndef OPENSSL_NO_FP_API
-int PEM_write_ECPrivateKey(FILE *fp, EC_KEY *x, const EVP_CIPHER *enc,
-                                               unsigned char *kstr, int klen,
-                                               pem_password_cb *cb, void *u)
-{
-	EVP_PKEY *k;
-	int ret;
-	k = EVP_PKEY_new();
-	if (!k)
-		return 0;
-	EVP_PKEY_set1_EC_KEY(k, x);
-	ret = PEM_write_PrivateKey(fp, k, enc, kstr, klen, cb, u);
-	EVP_PKEY_free(k);
-	return ret;
-}
-#endif
-
-#else
-
 IMPLEMENT_PEM_write_cb(ECPrivateKey, EC_KEY, PEM_STRING_ECPRIVATEKEY, ECPrivateKey)
 
-#endif
-
 IMPLEMENT_PEM_rw(EC_PUBKEY, EC_KEY, PEM_STRING_PUBLIC, EC_PUBKEY)
 
 #ifndef OPENSSL_NO_FP_API
@@ -404,7 +280,7 @@ EC_KEY *PEM_read_ECPrivateKey(FILE *fp, EC_KEY **eckey, pem_password_cb *cb,
 {
 	EVP_PKEY *pktmp;
 	pktmp = PEM_read_PrivateKey(fp, NULL, cb, u);
-	return pkey_get_eckey(pktmp, eckey);
+	return pkey_get_eckey(pktmp, eckey);	/* will free pktmp */
 }
 
 #endif
@@ -417,66 +293,4 @@ IMPLEMENT_PEM_rw_const(DHparams, DH, PEM_STRING_DHPARAMS, DHparams)
 
 #endif
 
-
-/* The PrivateKey case is not that straightforward.
- *   IMPLEMENT_PEM_rw_cb(PrivateKey, EVP_PKEY, PEM_STRING_EVP_PKEY, PrivateKey)
- * does not work, RSA and DSA keys have specific strings.
- * (When reading, parameter PEM_STRING_EVP_PKEY is a wildcard for anything
- * appropriate.)
- */
-
-#ifdef OPENSSL_FIPS
-
-static const char *pkey_str(EVP_PKEY *x)
-	{
-	switch (x->type)
-		{
-		case EVP_PKEY_RSA:
-		return PEM_STRING_RSA;
-
-		case EVP_PKEY_DSA:
-		return PEM_STRING_DSA;
-
-		case EVP_PKEY_EC:
-		return PEM_STRING_ECPRIVATEKEY;
-
-		default:
-		return NULL;
-		}
-	}
-
-
-int PEM_write_bio_PrivateKey(BIO *bp, EVP_PKEY *x, const EVP_CIPHER *enc,
-                                               unsigned char *kstr, int klen,
-                                               pem_password_cb *cb, void *u)
-	{
-		if (FIPS_mode())
-			return PEM_write_bio_PKCS8PrivateKey(bp, x, enc,
-						(char *)kstr, klen, cb, u);
-		else
-                	return PEM_ASN1_write_bio((i2d_of_void *)i2d_PrivateKey,
-			pkey_str(x), bp,(char *)x,enc,kstr,klen,cb,u);
-	}
-
-#ifndef OPENSSL_NO_FP_API
-int PEM_write_PrivateKey(FILE *fp, EVP_PKEY *x, const EVP_CIPHER *enc,
-                                               unsigned char *kstr, int klen,
-                                               pem_password_cb *cb, void *u)
-	{
-		if (FIPS_mode())
-			return PEM_write_PKCS8PrivateKey(fp, x, enc,
-						(char *)kstr, klen, cb, u);
-		else
-                	return PEM_ASN1_write((i2d_of_void *)i2d_PrivateKey,
-			pkey_str(x), fp,(char *)x,enc,kstr,klen,cb,u);
-	}
-#endif
-
-#else
-IMPLEMENT_PEM_write_cb(PrivateKey, EVP_PKEY, ((x->type == EVP_PKEY_DSA)?PEM_STRING_DSA:\
-			(x->type == EVP_PKEY_RSA)?PEM_STRING_RSA:PEM_STRING_ECPRIVATEKEY), PrivateKey)
-
-#endif
-
 IMPLEMENT_PEM_rw(PUBKEY, EVP_PKEY, PEM_STRING_PUBLIC, PUBKEY)
-
diff --git a/crypto/pem/pem_err.c b/crypto/pem/pem_err.c
index 3133563d77b7..d644aeedd4f5 100644
--- a/crypto/pem/pem_err.c
+++ b/crypto/pem/pem_err.c
@@ -1,6 +1,6 @@
 /* crypto/pem/pem_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,10 +70,22 @@
 
 static ERR_STRING_DATA PEM_str_functs[]=
 	{
+{ERR_FUNC(PEM_F_B2I_DSS),	"B2I_DSS"},
+{ERR_FUNC(PEM_F_B2I_PVK_BIO),	"b2i_PVK_bio"},
+{ERR_FUNC(PEM_F_B2I_RSA),	"B2I_RSA"},
+{ERR_FUNC(PEM_F_CHECK_BITLEN_DSA),	"CHECK_BITLEN_DSA"},
+{ERR_FUNC(PEM_F_CHECK_BITLEN_RSA),	"CHECK_BITLEN_RSA"},
 {ERR_FUNC(PEM_F_D2I_PKCS8PRIVATEKEY_BIO),	"d2i_PKCS8PrivateKey_bio"},
 {ERR_FUNC(PEM_F_D2I_PKCS8PRIVATEKEY_FP),	"d2i_PKCS8PrivateKey_fp"},
+{ERR_FUNC(PEM_F_DO_B2I),	"DO_B2I"},
+{ERR_FUNC(PEM_F_DO_B2I_BIO),	"DO_B2I_BIO"},
+{ERR_FUNC(PEM_F_DO_BLOB_HEADER),	"DO_BLOB_HEADER"},
 {ERR_FUNC(PEM_F_DO_PK8PKEY),	"DO_PK8PKEY"},
 {ERR_FUNC(PEM_F_DO_PK8PKEY_FP),	"DO_PK8PKEY_FP"},
+{ERR_FUNC(PEM_F_DO_PVK_BODY),	"DO_PVK_BODY"},
+{ERR_FUNC(PEM_F_DO_PVK_HEADER),	"DO_PVK_HEADER"},
+{ERR_FUNC(PEM_F_I2B_PVK),	"I2B_PVK"},
+{ERR_FUNC(PEM_F_I2B_PVK_BIO),	"i2b_PVK_bio"},
 {ERR_FUNC(PEM_F_LOAD_IV),	"LOAD_IV"},
 {ERR_FUNC(PEM_F_PEM_ASN1_READ),	"PEM_ASN1_read"},
 {ERR_FUNC(PEM_F_PEM_ASN1_READ_BIO),	"PEM_ASN1_read_bio"},
@@ -86,6 +98,7 @@ static ERR_STRING_DATA PEM_str_functs[]=
 {ERR_FUNC(PEM_F_PEM_PK8PKEY),	"PEM_PK8PKEY"},
 {ERR_FUNC(PEM_F_PEM_READ),	"PEM_read"},
 {ERR_FUNC(PEM_F_PEM_READ_BIO),	"PEM_read_bio"},
+{ERR_FUNC(PEM_F_PEM_READ_BIO_PARAMETERS),	"PEM_read_bio_Parameters"},
 {ERR_FUNC(PEM_F_PEM_READ_BIO_PRIVATEKEY),	"PEM_READ_BIO_PRIVATEKEY"},
 {ERR_FUNC(PEM_F_PEM_READ_PRIVATEKEY),	"PEM_READ_PRIVATEKEY"},
 {ERR_FUNC(PEM_F_PEM_SEALFINAL),	"PEM_SealFinal"},
@@ -93,6 +106,7 @@ static ERR_STRING_DATA PEM_str_functs[]=
 {ERR_FUNC(PEM_F_PEM_SIGNFINAL),	"PEM_SignFinal"},
 {ERR_FUNC(PEM_F_PEM_WRITE),	"PEM_write"},
 {ERR_FUNC(PEM_F_PEM_WRITE_BIO),	"PEM_write_bio"},
+{ERR_FUNC(PEM_F_PEM_WRITE_PRIVATEKEY),	"PEM_WRITE_PRIVATEKEY"},
 {ERR_FUNC(PEM_F_PEM_X509_INFO_READ),	"PEM_X509_INFO_read"},
 {ERR_FUNC(PEM_F_PEM_X509_INFO_READ_BIO),	"PEM_X509_INFO_read_bio"},
 {ERR_FUNC(PEM_F_PEM_X509_INFO_WRITE_BIO),	"PEM_X509_INFO_write_bio"},
@@ -105,18 +119,30 @@ static ERR_STRING_DATA PEM_str_reasons[]=
 {ERR_REASON(PEM_R_BAD_DECRYPT)           ,"bad decrypt"},
 {ERR_REASON(PEM_R_BAD_END_LINE)          ,"bad end line"},
 {ERR_REASON(PEM_R_BAD_IV_CHARS)          ,"bad iv chars"},
+{ERR_REASON(PEM_R_BAD_MAGIC_NUMBER)      ,"bad magic number"},
 {ERR_REASON(PEM_R_BAD_PASSWORD_READ)     ,"bad password read"},
+{ERR_REASON(PEM_R_BAD_VERSION_NUMBER)    ,"bad version number"},
+{ERR_REASON(PEM_R_BIO_WRITE_FAILURE)     ,"bio write failure"},
+{ERR_REASON(PEM_R_CIPHER_IS_NULL)        ,"cipher is null"},
 {ERR_REASON(PEM_R_ERROR_CONVERTING_PRIVATE_KEY),"error converting private key"},
+{ERR_REASON(PEM_R_EXPECTING_PRIVATE_KEY_BLOB),"expecting private key blob"},
+{ERR_REASON(PEM_R_EXPECTING_PUBLIC_KEY_BLOB),"expecting public key blob"},
+{ERR_REASON(PEM_R_INCONSISTENT_HEADER)   ,"inconsistent header"},
+{ERR_REASON(PEM_R_KEYBLOB_HEADER_PARSE_ERROR),"keyblob header parse error"},
+{ERR_REASON(PEM_R_KEYBLOB_TOO_SHORT)     ,"keyblob too short"},
 {ERR_REASON(PEM_R_NOT_DEK_INFO)          ,"not dek info"},
 {ERR_REASON(PEM_R_NOT_ENCRYPTED)         ,"not encrypted"},
 {ERR_REASON(PEM_R_NOT_PROC_TYPE)         ,"not proc type"},
 {ERR_REASON(PEM_R_NO_START_LINE)         ,"no start line"},
 {ERR_REASON(PEM_R_PROBLEMS_GETTING_PASSWORD),"problems getting password"},
 {ERR_REASON(PEM_R_PUBLIC_KEY_NO_RSA)     ,"public key no rsa"},
+{ERR_REASON(PEM_R_PVK_DATA_TOO_SHORT)    ,"pvk data too short"},
+{ERR_REASON(PEM_R_PVK_TOO_SHORT)         ,"pvk too short"},
 {ERR_REASON(PEM_R_READ_KEY)              ,"read key"},
 {ERR_REASON(PEM_R_SHORT_HEADER)          ,"short header"},
 {ERR_REASON(PEM_R_UNSUPPORTED_CIPHER)    ,"unsupported cipher"},
 {ERR_REASON(PEM_R_UNSUPPORTED_ENCRYPTION),"unsupported encryption"},
+{ERR_REASON(PEM_R_UNSUPPORTED_KEY_COMPONENTS),"unsupported key components"},
 {0,NULL}
 	};
 
diff --git a/crypto/pem/pem_info.c b/crypto/pem/pem_info.c
index 3a273f6f70b7..1b2be527edde 100644
--- a/crypto/pem/pem_info.c
+++ b/crypto/pem/pem_info.c
@@ -98,8 +98,8 @@ STACK_OF(X509_INFO) *PEM_X509_INFO_read_bio(BIO *bp, STACK_OF(X509_INFO) *sk, pe
 	long len,error=0;
 	int ok=0;
 	STACK_OF(X509_INFO) *ret=NULL;
-	unsigned int i,raw;
-	d2i_of_void *d2i;
+	unsigned int i,raw,ptype;
+	d2i_of_void *d2i = 0;
 
 	if (sk == NULL)
 		{
@@ -116,6 +116,7 @@ STACK_OF(X509_INFO) *PEM_X509_INFO_read_bio(BIO *bp, STACK_OF(X509_INFO) *sk, pe
 	for (;;)
 		{
 		raw=0;
+		ptype = 0;
 		i=PEM_read_bio(bp,&name,&header,&data,&len);
 		if (i == 0)
 			{
@@ -166,7 +167,6 @@ start:
 #ifndef OPENSSL_NO_RSA
 			if (strcmp(name,PEM_STRING_RSA) == 0)
 			{
-			d2i=(D2I_OF(void))d2i_RSAPrivateKey;
 			if (xi->x_pkey != NULL) 
 				{
 				if (!sk_X509_INFO_push(ret,xi)) goto err;
@@ -178,10 +178,8 @@ start:
 			xi->enc_len=0;
 
 			xi->x_pkey=X509_PKEY_new();
-			if ((xi->x_pkey->dec_pkey=EVP_PKEY_new()) == NULL)
-				goto err;
-			xi->x_pkey->dec_pkey->type=EVP_PKEY_RSA;
-			pp=&(xi->x_pkey->dec_pkey->pkey.rsa);
+			ptype=EVP_PKEY_RSA;
+			pp=&xi->x_pkey->dec_pkey;
 			if ((int)strlen(header) > 10) /* assume encrypted */
 				raw=1;
 			}
@@ -202,10 +200,8 @@ start:
 			xi->enc_len=0;
 
 			xi->x_pkey=X509_PKEY_new();
-			if ((xi->x_pkey->dec_pkey=EVP_PKEY_new()) == NULL)
-				goto err;
-			xi->x_pkey->dec_pkey->type=EVP_PKEY_DSA;
-			pp=&xi->x_pkey->dec_pkey->pkey.dsa;
+			ptype = EVP_PKEY_DSA;
+			pp=&xi->x_pkey->dec_pkey;
 			if ((int)strlen(header) > 10) /* assume encrypted */
 				raw=1;
 			}
@@ -226,10 +222,8 @@ start:
  			xi->enc_len=0;
  
  			xi->x_pkey=X509_PKEY_new();
- 			if ((xi->x_pkey->dec_pkey=EVP_PKEY_new()) == NULL)
- 				goto err;
- 			xi->x_pkey->dec_pkey->type=EVP_PKEY_EC;
- 			pp=&(xi->x_pkey->dec_pkey->pkey.ec);
+			ptype = EVP_PKEY_EC;
+ 			pp=&xi->x_pkey->dec_pkey;
  			if ((int)strlen(header) > 10) /* assume encrypted */
  				raw=1;
 			}
@@ -251,7 +245,15 @@ start:
 				if (!PEM_do_header(&cipher,data,&len,cb,u))
 					goto err;
 				p=data;
-				if (d2i(pp,&p,len) == NULL)
+				if (ptype)
+					{
+					if (!d2i_PrivateKey(ptype, pp, &p, len))
+						{
+						PEMerr(PEM_F_PEM_X509_INFO_READ_BIO,ERR_R_ASN1_LIB);
+						goto err;
+						}
+					}
+				else if (d2i(pp,&p,len) == NULL)
 					{
 					PEMerr(PEM_F_PEM_X509_INFO_READ_BIO,ERR_R_ASN1_LIB);
 					goto err;
@@ -337,6 +339,12 @@ int PEM_X509_INFO_write_bio(BIO *bp, X509_INFO *xi, EVP_CIPHER *enc,
 		{
 		if ( (xi->enc_data!=NULL) && (xi->enc_len>0) )
 			{
+			if (enc == NULL)
+				{
+				PEMerr(PEM_F_PEM_X509_INFO_WRITE_BIO,PEM_R_CIPHER_IS_NULL);
+				goto err;
+				}
+
 			/* copy from weirdo names into more normal things */
 			iv=xi->enc_cipher.iv;
 			data=(unsigned char *)xi->enc_data;
diff --git a/crypto/pem/pem_lib.c b/crypto/pem/pem_lib.c
index 22bb791bcee4..cfc89a992155 100644
--- a/crypto/pem/pem_lib.c
+++ b/crypto/pem/pem_lib.c
@@ -57,6 +57,7 @@
  */
 
 #include <stdio.h>
+#include <ctype.h>
 #include "cryptlib.h"
 #include <openssl/buffer.h>
 #include <openssl/objects.h>
@@ -65,9 +66,13 @@
 #include <openssl/x509.h>
 #include <openssl/pem.h>
 #include <openssl/pkcs12.h>
+#include "asn1_locl.h"
 #ifndef OPENSSL_NO_DES
 #include <openssl/des.h>
 #endif
+#ifndef OPENSSL_NO_ENGINE
+#include <openssl/engine.h>
+#endif
 
 const char PEM_version[]="PEM" OPENSSL_VERSION_PTEXT;
 
@@ -75,6 +80,7 @@ const char PEM_version[]="PEM" OPENSSL_VERSION_PTEXT;
 
 static int load_iv(char **fromp,unsigned char *to, int num);
 static int check_pem(const char *nm, const char *name);
+int pem_check_suffix(const char *pem_str, const char *suffix);
 
 int PEM_def_callback(char *buf, int num, int w, void *key)
 	{
@@ -99,7 +105,7 @@ int PEM_def_callback(char *buf, int num, int w, void *key)
 
 	for (;;)
 		{
-		i=EVP_read_pw_string(buf,num,prompt,w);
+		i=EVP_read_pw_string_min(buf,MIN_LENGTH,num,prompt,w);
 		if (i != 0)
 			{
 			PEMerr(PEM_F_PEM_DEF_CALLBACK,PEM_R_PROBLEMS_GETTING_PASSWORD);
@@ -183,20 +189,54 @@ static int check_pem(const char *nm, const char *name)
 
 	/* Make PEM_STRING_EVP_PKEY match any private key */
 
-	if(!strcmp(nm,PEM_STRING_PKCS8) &&
-		!strcmp(name,PEM_STRING_EVP_PKEY)) return 1;
-
-	if(!strcmp(nm,PEM_STRING_PKCS8INF) &&
-		 !strcmp(name,PEM_STRING_EVP_PKEY)) return 1;
-
-	if(!strcmp(nm,PEM_STRING_RSA) &&
-		!strcmp(name,PEM_STRING_EVP_PKEY)) return 1;
+	if(!strcmp(name,PEM_STRING_EVP_PKEY))
+		{
+		int slen;
+		const EVP_PKEY_ASN1_METHOD *ameth;
+		if(!strcmp(nm,PEM_STRING_PKCS8))
+			return 1;
+		if(!strcmp(nm,PEM_STRING_PKCS8INF))
+			return 1;
+		slen = pem_check_suffix(nm, "PRIVATE KEY"); 
+		if (slen > 0)
+			{
+			/* NB: ENGINE implementations wont contain
+			 * a deprecated old private key decode function
+			 * so don't look for them.
+			 */
+			ameth = EVP_PKEY_asn1_find_str(NULL, nm, slen);
+			if (ameth && ameth->old_priv_decode)
+				return 1;
+			}
+		return 0;
+		}
 
-	if(!strcmp(nm,PEM_STRING_DSA) &&
-		 !strcmp(name,PEM_STRING_EVP_PKEY)) return 1;
+	if(!strcmp(name,PEM_STRING_PARAMETERS))
+		{
+		int slen;
+		const EVP_PKEY_ASN1_METHOD *ameth;
+		slen = pem_check_suffix(nm, "PARAMETERS"); 
+		if (slen > 0)
+			{
+			ENGINE *e;
+			ameth = EVP_PKEY_asn1_find_str(&e, nm, slen);
+			if (ameth)
+				{
+				int r;
+				if (ameth->param_decode)
+					r = 1;
+				else
+					r = 0;
+#ifndef OPENSSL_NO_ENGINE
+				if (e)
+					ENGINE_finish(e);
+#endif
+				return r;
+				}
+			}
+		return 0;
+		}
 
- 	if(!strcmp(nm,PEM_STRING_ECPRIVATEKEY) &&
- 		 !strcmp(name,PEM_STRING_EVP_PKEY)) return 1;
 	/* Permit older strings */
 
 	if(!strcmp(nm,PEM_STRING_X509_OLD) &&
@@ -219,6 +259,14 @@ static int check_pem(const char *nm, const char *name)
 	if(!strcmp(nm, PEM_STRING_PKCS7_SIGNED) &&
 		!strcmp(name, PEM_STRING_PKCS7)) return 1;
 
+#ifndef OPENSSL_NO_CMS
+	if(!strcmp(nm, PEM_STRING_X509) &&
+		!strcmp(name, PEM_STRING_CMS)) return 1;
+	/* Allow CMS to be read from PKCS#7 headers */
+	if(!strcmp(nm, PEM_STRING_PKCS7) &&
+		!strcmp(name, PEM_STRING_CMS)) return 1;
+#endif
+
 	return 0;
 }
 
@@ -264,7 +312,7 @@ err:
 
 #ifndef OPENSSL_NO_FP_API
 int PEM_ASN1_write(i2d_of_void *i2d, const char *name, FILE *fp,
-		   char *x, const EVP_CIPHER *enc, unsigned char *kstr,
+		   void *x, const EVP_CIPHER *enc, unsigned char *kstr,
 		   int klen, pem_password_cb *callback, void *u)
         {
         BIO *b;
@@ -283,7 +331,7 @@ int PEM_ASN1_write(i2d_of_void *i2d, const char *name, FILE *fp,
 #endif
 
 int PEM_ASN1_write_bio(i2d_of_void *i2d, const char *name, BIO *bp,
-		       char *x, const EVP_CIPHER *enc, unsigned char *kstr,
+		       void *x, const EVP_CIPHER *enc, unsigned char *kstr,
 		       int klen, pem_password_cb *callback, void *u)
 	{
 	EVP_CIPHER_CTX ctx;
@@ -780,3 +828,25 @@ err:
 	BUF_MEM_free(dataB);
 	return(0);
 	}
+
+/* Check pem string and return prefix length.
+ * If for example the pem_str == "RSA PRIVATE KEY" and suffix = "PRIVATE KEY"
+ * the return value is 3 for the string "RSA".
+ */
+
+int pem_check_suffix(const char *pem_str, const char *suffix)
+	{
+	int pem_len = strlen(pem_str);
+	int suffix_len = strlen(suffix);
+	const char *p;
+	if (suffix_len + 1 >= pem_len)
+		return 0;
+	p = pem_str + pem_len - suffix_len;
+	if (strcmp(p, suffix))
+		return 0;
+	p--;
+	if (*p != ' ')
+		return 0;
+	return p - pem_str;
+	}
+
diff --git a/crypto/pem/pem_pkey.c b/crypto/pem/pem_pkey.c
index 4da4c31ce5fa..8ecf24903bc2 100644
--- a/crypto/pem/pem_pkey.c
+++ b/crypto/pem/pem_pkey.c
@@ -65,7 +65,12 @@
 #include <openssl/x509.h>
 #include <openssl/pkcs12.h>
 #include <openssl/pem.h>
+#ifndef OPENSSL_NO_ENGINE
+#include <openssl/engine.h>
+#endif
+#include "asn1_locl.h"
 
+int pem_check_suffix(const char *pem_str, const char *suffix);
 
 EVP_PKEY *PEM_read_bio_PrivateKey(BIO *bp, EVP_PKEY **x, pem_password_cb *cb, void *u)
 	{
@@ -73,19 +78,14 @@ EVP_PKEY *PEM_read_bio_PrivateKey(BIO *bp, EVP_PKEY **x, pem_password_cb *cb, vo
 	const unsigned char *p=NULL;
 	unsigned char *data=NULL;
 	long len;
+	int slen;
 	EVP_PKEY *ret=NULL;
 
 	if (!PEM_bytes_read_bio(&data, &len, &nm, PEM_STRING_EVP_PKEY, bp, cb, u))
 		return NULL;
 	p = data;
 
-	if (strcmp(nm,PEM_STRING_RSA) == 0)
-		ret=d2i_PrivateKey(EVP_PKEY_RSA,x,&p,len);
-	else if (strcmp(nm,PEM_STRING_DSA) == 0)
-		ret=d2i_PrivateKey(EVP_PKEY_DSA,x,&p,len);
-	else if (strcmp(nm,PEM_STRING_ECPRIVATEKEY) == 0)
-		ret=d2i_PrivateKey(EVP_PKEY_EC,x,&p,len);
-	else if (strcmp(nm,PEM_STRING_PKCS8INF) == 0) {
+	if (strcmp(nm,PEM_STRING_PKCS8INF) == 0) {
 		PKCS8_PRIV_KEY_INFO *p8inf;
 		p8inf=d2i_PKCS8_PRIV_KEY_INFO(NULL, &p, len);
 		if(!p8inf) goto p8err;
@@ -119,7 +119,14 @@ EVP_PKEY *PEM_read_bio_PrivateKey(BIO *bp, EVP_PKEY **x, pem_password_cb *cb, vo
 			*x = ret;
 		}
 		PKCS8_PRIV_KEY_INFO_free(p8inf);
-	}
+	} else if ((slen = pem_check_suffix(nm, "PRIVATE KEY")) > 0)
+		{
+		const EVP_PKEY_ASN1_METHOD *ameth;
+		ameth = EVP_PKEY_asn1_find_str(NULL, nm, slen);
+		if (!ameth || !ameth->old_priv_decode)
+			goto p8err;
+		ret=d2i_PrivateKey(ameth->pkey_id,x,&p,len);
+		}
 p8err:
 	if (ret == NULL)
 		PEMerr(PEM_F_PEM_READ_BIO_PRIVATEKEY,ERR_R_ASN1_LIB);
@@ -130,6 +137,74 @@ err:
 	return(ret);
 	}
 
+int PEM_write_bio_PrivateKey(BIO *bp, EVP_PKEY *x, const EVP_CIPHER *enc,
+                                               unsigned char *kstr, int klen,
+                                               pem_password_cb *cb, void *u)
+	{
+	char pem_str[80];
+	if (!x->ameth || x->ameth->priv_encode)
+		return PEM_write_bio_PKCS8PrivateKey(bp, x, enc,
+							(char *)kstr, klen,
+							cb, u);
+
+	BIO_snprintf(pem_str, 80, "%s PRIVATE KEY", x->ameth->pem_str);
+	return PEM_ASN1_write_bio((i2d_of_void *)i2d_PrivateKey,
+				pem_str,bp,x,enc,kstr,klen,cb,u);
+	}
+
+EVP_PKEY *PEM_read_bio_Parameters(BIO *bp, EVP_PKEY **x)
+	{
+	char *nm=NULL;
+	const unsigned char *p=NULL;
+	unsigned char *data=NULL;
+	long len;
+	int slen;
+	EVP_PKEY *ret=NULL;
+
+	if (!PEM_bytes_read_bio(&data, &len, &nm, PEM_STRING_PARAMETERS,
+								bp, 0, NULL))
+		return NULL;
+	p = data;
+
+	if ((slen = pem_check_suffix(nm, "PARAMETERS")) > 0)
+		{
+		ret = EVP_PKEY_new();
+		if (!ret)
+			goto err;
+		if (!EVP_PKEY_set_type_str(ret, nm, slen)
+			|| !ret->ameth->param_decode
+			|| !ret->ameth->param_decode(ret, &p, len))
+			{
+			EVP_PKEY_free(ret);
+			ret = NULL;
+			goto err;
+			}
+		if(x)
+			{
+			if(*x) EVP_PKEY_free((EVP_PKEY *)*x);
+			*x = ret;
+			}
+		}
+err:
+	if (ret == NULL)
+		PEMerr(PEM_F_PEM_READ_BIO_PARAMETERS,ERR_R_ASN1_LIB);
+	OPENSSL_free(nm);
+	OPENSSL_free(data);
+	return(ret);
+	}
+
+int PEM_write_bio_Parameters(BIO *bp, EVP_PKEY *x)
+	{
+	char pem_str[80];
+	if (!x->ameth || !x->ameth->param_encode)
+		return 0;
+
+	BIO_snprintf(pem_str, 80, "%s PARAMETERS", x->ameth->pem_str);
+	return PEM_ASN1_write_bio(
+		(i2d_of_void *)x->ameth->param_encode,
+				pem_str,bp,x,NULL,NULL,0,0,NULL);
+	}
+
 #ifndef OPENSSL_NO_FP_API
 EVP_PKEY *PEM_read_PrivateKey(FILE *fp, EVP_PKEY **x, pem_password_cb *cb, void *u)
 	{
@@ -146,4 +221,22 @@ EVP_PKEY *PEM_read_PrivateKey(FILE *fp, EVP_PKEY **x, pem_password_cb *cb, void
         BIO_free(b);
         return(ret);
 	}
+
+int PEM_write_PrivateKey(FILE *fp, EVP_PKEY *x, const EVP_CIPHER *enc,
+                                               unsigned char *kstr, int klen,
+                                               pem_password_cb *cb, void *u)
+	{
+        BIO *b;
+        int ret;
+
+        if ((b=BIO_new_fp(fp, BIO_NOCLOSE)) == NULL)
+		{
+		PEMerr(PEM_F_PEM_WRITE_PRIVATEKEY,ERR_R_BUF_LIB);
+                return 0;
+		}
+        ret=PEM_write_bio_PrivateKey(b, x, enc, kstr, klen, cb, u);
+        BIO_free(b);
+        return ret;
+	}
+
 #endif
diff --git a/crypto/pem/pem_x509.c b/crypto/pem/pem_x509.c
index 3f709f13e65f..b531057dc9a4 100644
--- a/crypto/pem/pem_x509.c
+++ b/crypto/pem/pem_x509.c
@@ -57,7 +57,6 @@
  */
 
 #include <stdio.h>
-#undef SSLEAY_MACROS
 #include "cryptlib.h"
 #include <openssl/bio.h>
 #include <openssl/evp.h>
diff --git a/crypto/pem/pem_xaux.c b/crypto/pem/pem_xaux.c
index 7cc74910095d..328f7962002c 100644
--- a/crypto/pem/pem_xaux.c
+++ b/crypto/pem/pem_xaux.c
@@ -57,7 +57,6 @@
  */
 
 #include <stdio.h>
-#undef SSLEAY_MACROS
 #include "cryptlib.h"
 #include <openssl/bio.h>
 #include <openssl/evp.h>
diff --git a/crypto/pem/pvkfmt.c b/crypto/pem/pvkfmt.c
new file mode 100644
index 000000000000..b1bf71a5daad
--- /dev/null
+++ b/crypto/pem/pvkfmt.c
@@ -0,0 +1,950 @@
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2005.
+ */
+/* ====================================================================
+ * Copyright (c) 2005 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+/* Support for PVK format keys and related structures (such a PUBLICKEYBLOB
+ * and PRIVATEKEYBLOB).
+ */
+
+#include "cryptlib.h"
+#include <openssl/pem.h>
+#include <openssl/rand.h>
+#include <openssl/bn.h>
+#if !defined(OPENSSL_NO_RSA) && !defined(OPENSSL_NO_DSA)
+#include <openssl/dsa.h>
+#include <openssl/rsa.h>
+
+/* Utility function: read a DWORD (4 byte unsigned integer) in little endian
+ * format
+ */
+
+static unsigned int read_ledword(const unsigned char **in)
+	{
+	const unsigned char *p = *in;
+	unsigned int ret;
+	ret = *p++;
+	ret |= (*p++ << 8);
+	ret |= (*p++ << 16);
+	ret |= (*p++ << 24);
+	*in = p;
+	return ret;
+	}
+
+/* Read a BIGNUM in little endian format. The docs say that this should take up 
+ * bitlen/8 bytes.
+ */
+
+static int read_lebn(const unsigned char **in, unsigned int nbyte, BIGNUM **r)
+	{
+	const unsigned char *p;
+	unsigned char *tmpbuf, *q;
+	unsigned int i;
+	p = *in + nbyte - 1;
+	tmpbuf = OPENSSL_malloc(nbyte);
+	if (!tmpbuf)
+		return 0;
+	q = tmpbuf;
+	for (i = 0; i < nbyte; i++)
+		*q++ = *p--;
+	*r = BN_bin2bn(tmpbuf, nbyte, NULL);
+	OPENSSL_free(tmpbuf);
+	if (*r)
+		{
+		*in += nbyte;
+		return 1;
+		}
+	else
+		return 0;
+	}
+
+
+/* Convert private key blob to EVP_PKEY: RSA and DSA keys supported */
+
+#define MS_PUBLICKEYBLOB	0x6
+#define MS_PRIVATEKEYBLOB	0x7
+#define MS_RSA1MAGIC		0x31415352L
+#define MS_RSA2MAGIC		0x32415352L
+#define MS_DSS1MAGIC		0x31535344L
+#define MS_DSS2MAGIC		0x32535344L
+
+#define MS_KEYALG_RSA_KEYX	0xa400
+#define MS_KEYALG_DSS_SIGN	0x2200
+
+#define MS_KEYTYPE_KEYX		0x1
+#define MS_KEYTYPE_SIGN		0x2
+
+/* The PVK file magic number: seems to spell out "bobsfile", who is Bob? */
+#define MS_PVKMAGIC		0xb0b5f11eL
+/* Salt length for PVK files */
+#define PVK_SALTLEN		0x10
+
+static EVP_PKEY *b2i_rsa(const unsigned char **in, unsigned int length,
+						unsigned int bitlen, int ispub);
+static EVP_PKEY *b2i_dss(const unsigned char **in, unsigned int length,
+						unsigned int bitlen, int ispub);
+
+static int do_blob_header(const unsigned char **in, unsigned int length,
+				unsigned int *pmagic, unsigned int *pbitlen,
+				int *pisdss, int *pispub)
+	{
+	const unsigned char *p = *in;
+	if (length < 16)
+		return 0;
+	/* bType */
+	if (*p == MS_PUBLICKEYBLOB)
+		{
+		if (*pispub == 0)
+			{
+			PEMerr(PEM_F_DO_BLOB_HEADER,
+					PEM_R_EXPECTING_PRIVATE_KEY_BLOB);
+			return 0;
+			}
+		*pispub = 1;
+		}
+	else if (*p == MS_PRIVATEKEYBLOB)
+		{
+		if (*pispub == 1)
+			{
+			PEMerr(PEM_F_DO_BLOB_HEADER,
+					PEM_R_EXPECTING_PUBLIC_KEY_BLOB);
+			return 0;
+			}
+		*pispub = 0;
+		}
+	else
+		return 0;
+	p++;
+	/* Version */
+	if (*p++ != 0x2)
+		{
+		PEMerr(PEM_F_DO_BLOB_HEADER, PEM_R_BAD_VERSION_NUMBER);
+		return 0;
+		}
+	/* Ignore reserved, aiKeyAlg */
+	p+= 6;
+	*pmagic = read_ledword(&p);
+	*pbitlen = read_ledword(&p);
+	*pisdss = 0;
+	switch (*pmagic)
+		{
+
+		case MS_DSS1MAGIC:
+		*pisdss = 1;
+		case MS_RSA1MAGIC:
+		if (*pispub == 0)
+			{
+			PEMerr(PEM_F_DO_BLOB_HEADER,
+					PEM_R_EXPECTING_PRIVATE_KEY_BLOB);
+			return 0;
+			}
+		break;
+
+		case MS_DSS2MAGIC:
+		*pisdss = 1;
+		case MS_RSA2MAGIC:
+		if (*pispub == 1)
+			{
+			PEMerr(PEM_F_DO_BLOB_HEADER,
+					PEM_R_EXPECTING_PUBLIC_KEY_BLOB);
+			return 0;
+			}
+		break;
+
+		default:
+		PEMerr(PEM_F_DO_BLOB_HEADER, PEM_R_BAD_MAGIC_NUMBER);
+		return -1;
+		}
+	*in = p;
+	return 1;
+	}
+
+static unsigned int blob_length(unsigned bitlen, int isdss, int ispub)
+	{
+	unsigned int nbyte, hnbyte;
+	nbyte = (bitlen + 7) >> 3;
+	hnbyte = (bitlen + 15) >> 4;
+	if (isdss)
+		{
+
+		/* Expected length: 20 for q + 3 components bitlen each + 24
+		 * for seed structure.
+		 */
+		if (ispub)
+			return  44 + 3 * nbyte;
+		/* Expected length: 20 for q, priv, 2 bitlen components + 24
+		 * for seed structure.
+		 */
+		else
+			return 64 + 2 * nbyte;
+		}
+	else
+		{
+		/* Expected length: 4 for 'e' + 'n' */
+		if (ispub)
+			return 4 + nbyte;
+		else
+		/* Expected length: 4 for 'e' and 7 other components.
+		 * 2 components are bitlen size, 5 are bitlen/2
+		 */
+			return 4 + 2*nbyte + 5*hnbyte;
+		}
+
+	}
+
+static EVP_PKEY *do_b2i(const unsigned char **in, unsigned int length,
+								int ispub)
+	{
+	const unsigned char *p = *in;
+	unsigned int bitlen, magic;
+	int isdss;
+	if (do_blob_header(&p, length, &magic, &bitlen, &isdss, &ispub) <= 0)
+		{
+		PEMerr(PEM_F_DO_B2I, PEM_R_KEYBLOB_HEADER_PARSE_ERROR);
+		return NULL;
+		}
+	length -= 16;
+	if (length < blob_length(bitlen, isdss, ispub))
+		{
+		PEMerr(PEM_F_DO_B2I, PEM_R_KEYBLOB_TOO_SHORT);
+		return NULL;
+		}
+	if (isdss)
+		return b2i_dss(&p, length, bitlen, ispub);
+	else
+		return b2i_rsa(&p, length, bitlen, ispub);
+	}
+
+static EVP_PKEY *do_b2i_bio(BIO *in, int ispub)
+	{
+	const unsigned char *p;
+	unsigned char hdr_buf[16], *buf = NULL;
+	unsigned int bitlen, magic, length;
+	int isdss;
+	EVP_PKEY *ret = NULL;
+	if (BIO_read(in, hdr_buf, 16) != 16)
+		{
+		PEMerr(PEM_F_DO_B2I_BIO, PEM_R_KEYBLOB_TOO_SHORT);
+		return NULL;
+		}
+	p = hdr_buf;
+	if (do_blob_header(&p, 16, &magic, &bitlen, &isdss, &ispub) <= 0)
+		return NULL;
+
+	length = blob_length(bitlen, isdss, ispub);
+	buf = OPENSSL_malloc(length);
+	if (!buf)
+		{
+		PEMerr(PEM_F_DO_B2I_BIO, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+	p = buf;
+	if (BIO_read(in, buf, length) != (int)length)
+		{
+		PEMerr(PEM_F_DO_B2I_BIO, PEM_R_KEYBLOB_TOO_SHORT);
+		goto err;
+		}
+
+	if (isdss)
+		ret = b2i_dss(&p, length, bitlen, ispub);
+	else
+		ret = b2i_rsa(&p, length, bitlen, ispub);
+
+	err:
+	if (buf)
+		OPENSSL_free(buf);
+	return ret;
+	}
+
+static EVP_PKEY *b2i_dss(const unsigned char **in, unsigned int length,
+						unsigned int bitlen, int ispub)
+	{
+	const unsigned char *p = *in;
+	EVP_PKEY *ret = NULL;
+	DSA *dsa = NULL;
+	BN_CTX *ctx = NULL;
+	unsigned int nbyte;
+	nbyte = (bitlen + 7) >> 3;
+
+	dsa = DSA_new();
+	ret = EVP_PKEY_new();
+	if (!dsa || !ret)
+		goto memerr;
+	if (!read_lebn(&p, nbyte, &dsa->p))
+		goto memerr;
+	if (!read_lebn(&p, 20, &dsa->q))
+		goto memerr;
+	if (!read_lebn(&p, nbyte, &dsa->g))
+		goto memerr;
+	if (ispub)
+		{
+		if (!read_lebn(&p, nbyte, &dsa->pub_key))
+			goto memerr;
+		}
+	else
+		{
+		if (!read_lebn(&p, 20, &dsa->priv_key))
+			goto memerr;
+		/* Calculate public key */
+		if (!(dsa->pub_key = BN_new()))
+			goto memerr;
+		if (!(ctx = BN_CTX_new()))
+			goto memerr;
+			
+		if (!BN_mod_exp(dsa->pub_key, dsa->g,
+						 dsa->priv_key, dsa->p, ctx))
+			
+			goto memerr;
+		BN_CTX_free(ctx);
+		}
+
+	EVP_PKEY_set1_DSA(ret, dsa);
+	DSA_free(dsa);
+	*in = p;
+	return ret;
+
+	memerr:
+	PEMerr(PEM_F_B2I_DSS, ERR_R_MALLOC_FAILURE);
+	if (dsa)
+		DSA_free(dsa);
+	if (ret)
+		EVP_PKEY_free(ret);
+	if (ctx)
+		BN_CTX_free(ctx);
+	return NULL;
+	}
+
+static EVP_PKEY *b2i_rsa(const unsigned char **in, unsigned int length,
+						unsigned int bitlen, int ispub)
+		
+	{
+	const unsigned char *p = *in;
+	EVP_PKEY *ret = NULL;
+	RSA *rsa = NULL;
+	unsigned int nbyte, hnbyte;
+	nbyte = (bitlen + 7) >> 3;
+	hnbyte = (bitlen + 15) >> 4;
+	rsa = RSA_new();
+	ret = EVP_PKEY_new();
+	if (!rsa || !ret)
+		goto memerr;
+	rsa->e = BN_new();
+	if (!rsa->e)
+		goto memerr;
+	if (!BN_set_word(rsa->e, read_ledword(&p)))
+		goto memerr;
+	if (!read_lebn(&p, nbyte, &rsa->n))
+		goto memerr;
+	if (!ispub)
+		{
+		if (!read_lebn(&p, hnbyte, &rsa->p))
+			goto memerr;
+		if (!read_lebn(&p, hnbyte, &rsa->q))
+			goto memerr;
+		if (!read_lebn(&p, hnbyte, &rsa->dmp1))
+			goto memerr;
+		if (!read_lebn(&p, hnbyte, &rsa->dmq1))
+			goto memerr;
+		if (!read_lebn(&p, hnbyte, &rsa->iqmp))
+			goto memerr;
+		if (!read_lebn(&p, nbyte, &rsa->d))
+			goto memerr;
+		}
+
+	EVP_PKEY_set1_RSA(ret, rsa);
+	RSA_free(rsa);
+	*in = p;
+	return ret;
+	memerr:
+	PEMerr(PEM_F_B2I_RSA, ERR_R_MALLOC_FAILURE);
+	if (rsa)
+		RSA_free(rsa);
+	if (ret)
+		EVP_PKEY_free(ret);
+	return NULL;
+	}
+
+EVP_PKEY *b2i_PrivateKey(const unsigned char **in, long length)
+	{
+	return do_b2i(in, length, 0);
+	}
+
+EVP_PKEY *b2i_PublicKey(const unsigned char **in, long length)
+	{
+	return do_b2i(in, length, 1);
+	}
+
+
+EVP_PKEY *b2i_PrivateKey_bio(BIO *in)
+	{
+	return do_b2i_bio(in, 0);
+	}
+
+EVP_PKEY *b2i_PublicKey_bio(BIO *in)
+	{
+	return do_b2i_bio(in, 1);
+	}
+
+static void write_ledword(unsigned char **out, unsigned int dw)
+	{
+	unsigned char *p = *out;
+	*p++ = dw & 0xff;
+	*p++ = (dw>>8) & 0xff;
+	*p++ = (dw>>16) & 0xff;
+	*p++ = (dw>>24) & 0xff;
+	*out = p;
+	}
+
+static void write_lebn(unsigned char **out, const BIGNUM *bn, int len)
+	{
+	int nb, i;
+	unsigned char *p = *out, *q, c;
+	nb = BN_num_bytes(bn);
+	BN_bn2bin(bn, p);
+	q = p + nb - 1;
+	/* In place byte order reversal */
+	for (i = 0; i < nb/2; i++)
+		{
+		c = *p;
+		*p++ = *q;
+		*q-- = c;
+		}
+	*out += nb;
+	/* Pad with zeroes if we have to */
+	if (len > 0)
+		{
+		len -= nb;
+		if (len > 0)
+			{
+			memset(*out, 0, len);
+			*out += len;
+			}
+		}
+	}
+
+
+static int check_bitlen_rsa(RSA *rsa, int ispub, unsigned int *magic);
+static int check_bitlen_dsa(DSA *dsa, int ispub, unsigned int *magic);
+
+static void write_rsa(unsigned char **out, RSA *rsa, int ispub);
+static void write_dsa(unsigned char **out, DSA *dsa, int ispub);
+	
+static int do_i2b(unsigned char **out, EVP_PKEY *pk, int ispub)
+	{
+	unsigned char *p;
+	unsigned int bitlen, magic = 0, keyalg;
+	int outlen, noinc = 0;
+	if (pk->type == EVP_PKEY_DSA)
+		{
+		bitlen = check_bitlen_dsa(pk->pkey.dsa, ispub, &magic);
+		keyalg = MS_KEYALG_DSS_SIGN;
+		}
+	else if (pk->type == EVP_PKEY_RSA)
+		{
+		bitlen = check_bitlen_rsa(pk->pkey.rsa, ispub, &magic);
+		keyalg = MS_KEYALG_RSA_KEYX;
+		}
+	else
+		return -1;
+	if (bitlen == 0)
+		return -1;
+	outlen = 16 + blob_length(bitlen,
+			keyalg == MS_KEYALG_DSS_SIGN ? 1 : 0, ispub);
+	if (out == NULL)
+		return outlen;
+	if (*out)
+		p = *out;
+	else
+		{
+		p = OPENSSL_malloc(outlen);
+		if (!p)
+			return -1;
+		*out = p;
+		noinc = 1;
+		}
+	if (ispub)
+		*p++ = MS_PUBLICKEYBLOB;
+	else
+		*p++ = MS_PRIVATEKEYBLOB;
+	*p++ = 0x2;
+	*p++ = 0;
+	*p++ = 0;
+	write_ledword(&p, keyalg);
+	write_ledword(&p, magic);
+	write_ledword(&p, bitlen);
+	if (keyalg == MS_KEYALG_DSS_SIGN)
+		write_dsa(&p, pk->pkey.dsa, ispub);
+	else
+		write_rsa(&p, pk->pkey.rsa, ispub);
+	if (!noinc)
+		*out += outlen;
+	return outlen;
+	}
+
+static int do_i2b_bio(BIO *out, EVP_PKEY *pk, int ispub)
+	{
+	unsigned char *tmp = NULL;
+	int outlen, wrlen;
+	outlen = do_i2b(&tmp, pk, ispub);
+	if (outlen < 0)
+		return -1;
+	wrlen = BIO_write(out, tmp, outlen);
+	OPENSSL_free(tmp);
+	if (wrlen == outlen)
+		return outlen;
+	return -1;
+	}
+
+static int check_bitlen_dsa(DSA *dsa, int ispub, unsigned int *pmagic)
+	{
+	int bitlen;
+	bitlen = BN_num_bits(dsa->p);
+	if ((bitlen & 7) || (BN_num_bits(dsa->q) != 160)
+		|| (BN_num_bits(dsa->g) > bitlen))
+		goto badkey;
+	if (ispub)
+		{
+		if (BN_num_bits(dsa->pub_key) > bitlen)
+			goto badkey;
+		*pmagic = MS_DSS1MAGIC;
+		}
+	else
+		{
+		if (BN_num_bits(dsa->priv_key) > 160)
+			goto badkey;
+		*pmagic = MS_DSS2MAGIC;
+		}
+	
+	return bitlen;
+	badkey:
+	PEMerr(PEM_F_CHECK_BITLEN_DSA, PEM_R_UNSUPPORTED_KEY_COMPONENTS);
+	return 0;
+	}
+
+static int check_bitlen_rsa(RSA *rsa, int ispub, unsigned int *pmagic)
+	{
+	int nbyte, hnbyte, bitlen;
+	if (BN_num_bits(rsa->e) > 32)
+		goto badkey;
+	bitlen = BN_num_bits(rsa->n);
+	nbyte = BN_num_bytes(rsa->n);
+	hnbyte = (BN_num_bits(rsa->n) + 15) >> 4;
+	if (ispub)
+		{
+		*pmagic = MS_RSA1MAGIC;
+		return bitlen;
+		}
+	else
+	{
+		*pmagic = MS_RSA2MAGIC;
+		/* For private key each component must fit within nbyte or
+		 * hnbyte.
+		 */
+		if (BN_num_bytes(rsa->d) > nbyte)
+			goto badkey;
+		if ((BN_num_bytes(rsa->iqmp) > hnbyte)
+			|| (BN_num_bytes(rsa->p) > hnbyte)
+			|| (BN_num_bytes(rsa->q) > hnbyte)
+			|| (BN_num_bytes(rsa->dmp1) > hnbyte)
+			|| (BN_num_bytes(rsa->dmq1) > hnbyte))
+			goto badkey;
+	}
+	return bitlen;
+	badkey:
+	PEMerr(PEM_F_CHECK_BITLEN_RSA, PEM_R_UNSUPPORTED_KEY_COMPONENTS);
+	return 0;
+	}
+
+
+static void write_rsa(unsigned char **out, RSA *rsa, int ispub)
+	{
+	int nbyte, hnbyte;
+	nbyte = BN_num_bytes(rsa->n);
+	hnbyte = (BN_num_bits(rsa->n) + 15) >> 4;
+	write_lebn(out, rsa->e, 4);
+	write_lebn(out, rsa->n, -1);
+	if (ispub)
+		return;
+	write_lebn(out, rsa->p, hnbyte);
+	write_lebn(out, rsa->q, hnbyte);
+	write_lebn(out, rsa->dmp1, hnbyte);
+	write_lebn(out, rsa->dmq1, hnbyte);
+	write_lebn(out, rsa->iqmp, hnbyte);
+	write_lebn(out, rsa->d, nbyte);
+	}
+
+	
+static void write_dsa(unsigned char **out, DSA *dsa, int ispub)
+	{
+	int nbyte;
+	nbyte = BN_num_bytes(dsa->p);
+	write_lebn(out, dsa->p, nbyte);
+	write_lebn(out, dsa->q, 20);
+	write_lebn(out, dsa->g, nbyte);
+	if (ispub)
+		write_lebn(out, dsa->pub_key, nbyte);
+	else
+		write_lebn(out, dsa->priv_key, 20);
+	/* Set "invalid" for seed structure values */
+	memset(*out, 0xff, 24);
+	*out += 24;
+	return;
+	}
+	
+
+int i2b_PrivateKey_bio(BIO *out, EVP_PKEY *pk)
+	{
+	return do_i2b_bio(out, pk, 0);
+	}
+
+int i2b_PublicKey_bio(BIO *out, EVP_PKEY *pk)
+	{
+	return do_i2b_bio(out, pk, 1);
+	}
+
+#ifndef OPENSSL_NO_RC4
+
+static int do_PVK_header(const unsigned char **in, unsigned int length,
+		int skip_magic,
+	       	unsigned int *psaltlen, unsigned int *pkeylen)
+		
+	{
+	const unsigned char *p = *in;
+	unsigned int pvk_magic, is_encrypted;
+	if (skip_magic)
+		{
+		if (length < 20)
+			{
+			PEMerr(PEM_F_DO_PVK_HEADER, PEM_R_PVK_TOO_SHORT);
+			return 0;
+			}
+		length -= 20;
+		}
+	else
+		{
+		if (length < 24)
+			{
+			PEMerr(PEM_F_DO_PVK_HEADER, PEM_R_PVK_TOO_SHORT);
+			return 0;
+			}
+		length -= 24;
+		pvk_magic = read_ledword(&p);
+		if (pvk_magic != MS_PVKMAGIC)
+			{
+			PEMerr(PEM_F_DO_PVK_HEADER, PEM_R_BAD_MAGIC_NUMBER);
+			return 0;
+			}
+		}
+	/* Skip reserved */
+	p += 4;
+	/*keytype = */read_ledword(&p);
+	is_encrypted = read_ledword(&p);
+	*psaltlen = read_ledword(&p);
+	*pkeylen = read_ledword(&p);
+
+	if (is_encrypted && !*psaltlen)
+		{
+		PEMerr(PEM_F_DO_PVK_HEADER, PEM_R_INCONSISTENT_HEADER);
+		return 0;
+		}
+
+	*in = p;
+	return 1;
+	}
+
+static int derive_pvk_key(unsigned char *key, 
+			const unsigned char *salt, unsigned int saltlen,
+			const unsigned char *pass, int passlen)
+	{
+	EVP_MD_CTX mctx;
+	int rv = 1;
+	EVP_MD_CTX_init(&mctx);
+	if (!EVP_DigestInit_ex(&mctx, EVP_sha1(), NULL)
+		|| !EVP_DigestUpdate(&mctx, salt, saltlen)
+		|| !EVP_DigestUpdate(&mctx, pass, passlen)
+		|| !EVP_DigestFinal_ex(&mctx, key, NULL))
+			rv = 0;
+
+	EVP_MD_CTX_cleanup(&mctx);
+	return rv;
+	}
+	
+
+static EVP_PKEY *do_PVK_body(const unsigned char **in,
+		unsigned int saltlen, unsigned int keylen,
+		pem_password_cb *cb, void *u)
+	{
+	EVP_PKEY *ret = NULL;
+	const unsigned char *p = *in;
+	unsigned int magic;
+	unsigned char *enctmp = NULL, *q;
+	EVP_CIPHER_CTX cctx;
+	EVP_CIPHER_CTX_init(&cctx);
+	if (saltlen)
+		{
+		char psbuf[PEM_BUFSIZE];
+		unsigned char keybuf[20];
+		int enctmplen, inlen;
+		if (cb)
+			inlen=cb(psbuf,PEM_BUFSIZE,0,u);
+		else
+			inlen=PEM_def_callback(psbuf,PEM_BUFSIZE,0,u);
+		if (inlen <= 0)
+			{
+			PEMerr(PEM_F_DO_PVK_BODY,PEM_R_BAD_PASSWORD_READ);
+			return NULL;
+			}
+		enctmp = OPENSSL_malloc(keylen + 8);
+		if (!enctmp)
+			{
+			PEMerr(PEM_F_DO_PVK_BODY, ERR_R_MALLOC_FAILURE);
+			return NULL;
+			}
+		if (!derive_pvk_key(keybuf, p, saltlen,
+			    (unsigned char *)psbuf, inlen))
+			return NULL;
+		p += saltlen;
+		/* Copy BLOBHEADER across, decrypt rest */
+		memcpy(enctmp, p, 8);
+		p += 8;
+		inlen = keylen - 8;
+		q = enctmp + 8;
+		if (!EVP_DecryptInit_ex(&cctx, EVP_rc4(), NULL, keybuf, NULL))
+			goto err;
+		if (!EVP_DecryptUpdate(&cctx, q, &enctmplen, p, inlen))
+			goto err;
+		if (!EVP_DecryptFinal_ex(&cctx, q + enctmplen, &enctmplen))
+			goto err;
+		magic = read_ledword((const unsigned char **)&q);
+		if (magic != MS_RSA2MAGIC && magic != MS_DSS2MAGIC)
+			{
+			q = enctmp + 8;
+			memset(keybuf + 5, 0, 11);
+			if (!EVP_DecryptInit_ex(&cctx, EVP_rc4(), NULL, keybuf,
+								NULL))
+				goto err;
+			OPENSSL_cleanse(keybuf, 20);
+			if (!EVP_DecryptUpdate(&cctx, q, &enctmplen, p, inlen))
+				goto err;
+			if (!EVP_DecryptFinal_ex(&cctx, q + enctmplen,
+								&enctmplen))
+				goto err;
+			magic = read_ledword((const unsigned char **)&q);
+			if (magic != MS_RSA2MAGIC && magic != MS_DSS2MAGIC)
+				{
+				PEMerr(PEM_F_DO_PVK_BODY, PEM_R_BAD_DECRYPT);
+				goto err;
+				}
+			}
+		else
+			OPENSSL_cleanse(keybuf, 20);
+		p = enctmp;
+		}
+
+	ret = b2i_PrivateKey(&p, keylen);
+	err:
+	EVP_CIPHER_CTX_cleanup(&cctx);
+	if (enctmp && saltlen)
+		OPENSSL_free(enctmp);
+	return ret;
+	}
+
+
+EVP_PKEY *b2i_PVK_bio(BIO *in, pem_password_cb *cb, void *u)
+	{
+	unsigned char pvk_hdr[24], *buf = NULL;
+	const unsigned char *p;
+	int buflen;
+	EVP_PKEY *ret = NULL;
+	unsigned int saltlen, keylen;
+	if (BIO_read(in, pvk_hdr, 24) != 24)
+		{
+		PEMerr(PEM_F_B2I_PVK_BIO, PEM_R_PVK_DATA_TOO_SHORT);
+		return NULL;
+		}
+	p = pvk_hdr;
+
+	if (!do_PVK_header(&p, 24, 0, &saltlen, &keylen))
+		return 0;
+	buflen = (int) keylen + saltlen;
+	buf = OPENSSL_malloc(buflen);
+	if (!buf)
+		{
+		PEMerr(PEM_F_B2I_PVK_BIO, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	p = buf;
+	if (BIO_read(in, buf, buflen) != buflen)
+		{
+		PEMerr(PEM_F_B2I_PVK_BIO, PEM_R_PVK_DATA_TOO_SHORT);
+		goto err;
+		}
+	ret = do_PVK_body(&p, saltlen, keylen, cb, u);
+
+	err:
+	if (buf)
+		{
+		OPENSSL_cleanse(buf, buflen);
+		OPENSSL_free(buf);
+		}
+	return ret;
+	}
+
+	
+	
+static int i2b_PVK(unsigned char **out, EVP_PKEY*pk, int enclevel,
+		pem_password_cb *cb, void *u)
+	{
+	int outlen = 24, pklen;
+	unsigned char *p, *salt = NULL;
+	EVP_CIPHER_CTX cctx;
+	EVP_CIPHER_CTX_init(&cctx);
+	if (enclevel)
+		outlen += PVK_SALTLEN;
+	pklen = do_i2b(NULL, pk, 0);
+	if (pklen < 0)
+		return -1;
+	outlen += pklen;
+	if (!out)
+		return outlen;
+	if (*out)
+		p = *out;
+	else
+		{
+		p = OPENSSL_malloc(outlen);
+		if (!p)
+			{
+			PEMerr(PEM_F_I2B_PVK,ERR_R_MALLOC_FAILURE);
+			return -1;
+			}
+		*out = p;
+		}
+
+	write_ledword(&p, MS_PVKMAGIC);
+	write_ledword(&p, 0);
+	if (pk->type == EVP_PKEY_DSA)
+		write_ledword(&p, MS_KEYTYPE_SIGN);
+	else
+		write_ledword(&p, MS_KEYTYPE_KEYX);
+	write_ledword(&p, enclevel ? 1 : 0);
+	write_ledword(&p, enclevel ? PVK_SALTLEN: 0);
+	write_ledword(&p, pklen);
+	if (enclevel)
+		{
+		if (RAND_bytes(p, PVK_SALTLEN) <= 0)
+			goto error;
+		salt = p;
+		p += PVK_SALTLEN;
+		}
+	do_i2b(&p, pk, 0);
+	if (enclevel == 0)
+		return outlen;
+	else
+		{
+		char psbuf[PEM_BUFSIZE];
+		unsigned char keybuf[20];
+		int enctmplen, inlen;
+		if (cb)
+			inlen=cb(psbuf,PEM_BUFSIZE,1,u);
+		else
+			inlen=PEM_def_callback(psbuf,PEM_BUFSIZE,1,u);
+		if (inlen <= 0)
+			{
+			PEMerr(PEM_F_I2B_PVK,PEM_R_BAD_PASSWORD_READ);
+			goto error;
+			}
+		if (!derive_pvk_key(keybuf, salt, PVK_SALTLEN,
+			    (unsigned char *)psbuf, inlen))
+			goto error;
+		if (enclevel == 1)
+			memset(keybuf + 5, 0, 11);
+		p = salt + PVK_SALTLEN + 8;
+		if (!EVP_EncryptInit_ex(&cctx, EVP_rc4(), NULL, keybuf, NULL))
+			goto error;
+		OPENSSL_cleanse(keybuf, 20);
+		if (!EVP_DecryptUpdate(&cctx, p, &enctmplen, p, pklen - 8))
+			goto error;
+		if (!EVP_DecryptFinal_ex(&cctx, p + enctmplen, &enctmplen))
+			goto error;
+		}
+	EVP_CIPHER_CTX_cleanup(&cctx);
+	return outlen;
+
+	error:
+	EVP_CIPHER_CTX_cleanup(&cctx);
+	return -1;
+	}
+
+int i2b_PVK_bio(BIO *out, EVP_PKEY *pk, int enclevel,
+		pem_password_cb *cb, void *u)
+	{
+	unsigned char *tmp = NULL;
+	int outlen, wrlen;
+	outlen = i2b_PVK(&tmp, pk, enclevel, cb, u);
+	if (outlen < 0)
+		return -1;
+	wrlen = BIO_write(out, tmp, outlen);
+	OPENSSL_free(tmp);
+	if (wrlen == outlen)
+		{
+		PEMerr(PEM_F_I2B_PVK_BIO, PEM_R_BIO_WRITE_FAILURE);
+		return outlen;
+		}
+	return -1;
+	}
+
+#endif
+
+#endif
diff --git a/crypto/perlasm/ppc-xlate.pl b/crypto/perlasm/ppc-xlate.pl
new file mode 100755
index 000000000000..a3edd982b664
--- /dev/null
+++ b/crypto/perlasm/ppc-xlate.pl
@@ -0,0 +1,159 @@
+#!/usr/bin/env perl
+
+# PowerPC assembler distiller by <appro>.
+
+my $flavour = shift;
+my $output = shift;
+open STDOUT,">$output" || die "can't open $output: $!";
+
+my %GLOBALS;
+my $dotinlocallabels=($flavour=~/linux/)?1:0;
+
+################################################################
+# directives which need special treatment on different platforms
+################################################################
+my $globl = sub {
+    my $junk = shift;
+    my $name = shift;
+    my $global = \$GLOBALS{$name};
+    my $ret;
+
+    $name =~ s|^[\.\_]||;
+ 
+    SWITCH: for ($flavour) {
+	/aix/		&& do { $name = ".$name";
+				last;
+			      };
+	/osx/		&& do { $name = "_$name";
+				last;
+			      };
+	/linux.*32/	&& do {	$ret .= ".globl	$name\n";
+				$ret .= ".type	$name,\@function";
+				last;
+			      };
+	/linux.*64/	&& do {	$ret .= ".globl	$name\n";
+				$ret .= ".type	$name,\@function\n";
+				$ret .= ".section	\".opd\",\"aw\"\n";
+				$ret .= ".align	3\n";
+				$ret .= "$name:\n";
+				$ret .= ".quad	.$name,.TOC.\@tocbase,0\n";
+				$ret .= ".size	$name,24\n";
+				$ret .= ".previous\n";
+
+				$name = ".$name";
+				last;
+			      };
+    }
+
+    $ret = ".globl	$name" if (!$ret);
+    $$global = $name;
+    $ret;
+};
+my $text = sub {
+    ($flavour =~ /aix/) ? ".csect" : ".text";
+};
+my $machine = sub {
+    my $junk = shift;
+    my $arch = shift;
+    if ($flavour =~ /osx/)
+    {	$arch =~ s/\"//g;
+	$arch = ($flavour=~/64/) ? "ppc970-64" : "ppc970" if ($arch eq "any");
+    }
+    ".machine	$arch";
+};
+my $size = sub {
+    if ($flavour =~ /linux.*32/)
+    {	shift;
+	".size	" . join(",",@_);
+    }
+    else
+    {	"";	}
+};
+my $asciz = sub {
+    shift;
+    my $line = join(",",@_);
+    if ($line =~ /^"(.*)"$/)
+    {	".byte	" . join(",",unpack("C*",$1),0) . "\n.align	2";	}
+    else
+    {	"";	}
+};
+
+################################################################
+# simplified mnemonics not handled by at least one assembler
+################################################################
+my $cmplw = sub {
+    my $f = shift;
+    my $cr = 0; $cr = shift if ($#_>1);
+    # Some out-of-date 32-bit GNU assembler just can't handle cmplw...
+    ($flavour =~ /linux.*32/) ?
+	"	.long	".sprintf "0x%x",31<<26|$cr<<23|$_[0]<<16|$_[1]<<11|64 :
+	"	cmplw	".join(',',$cr,@_);
+};
+my $bdnz = sub {
+    my $f = shift;
+    my $bo = $f=~/[\+\-]/ ? 16+9 : 16;	# optional "to be taken" hint
+    "	bc	$bo,0,".shift;
+} if ($flavour!~/linux/);
+my $bltlr = sub {
+    my $f = shift;
+    my $bo = $f=~/\-/ ? 12+2 : 12;	# optional "not to be taken" hint
+    ($flavour =~ /linux/) ?		# GNU as doesn't allow most recent hints
+	"	.long	".sprintf "0x%x",19<<26|$bo<<21|16<<1 :
+	"	bclr	$bo,0";
+};
+my $bnelr = sub {
+    my $f = shift;
+    my $bo = $f=~/\-/ ? 4+2 : 4;	# optional "not to be taken" hint
+    ($flavour =~ /linux/) ?		# GNU as doesn't allow most recent hints
+	"	.long	".sprintf "0x%x",19<<26|$bo<<21|2<<16|16<<1 :
+	"	bclr	$bo,2";
+};
+my $beqlr = sub {
+    my $f = shift;
+    my $bo = $f=~/-/ ? 12+2 : 12;	# optional "not to be taken" hint
+    ($flavour =~ /linux/) ?		# GNU as doesn't allow most recent hints
+	"	.long	".sprintf "0x%X",19<<26|$bo<<21|2<<16|16<<1 :
+	"	bclr	$bo,2";
+};
+# GNU assembler can't handle extrdi rA,rS,16,48, or when sum of last two
+# arguments is 64, with "operand out of range" error.
+my $extrdi = sub {
+    my ($f,$ra,$rs,$n,$b) = @_;
+    $b = ($b+$n)&63; $n = 64-$n;
+    "	rldicl	$ra,$rs,$b,$n";
+};
+
+while($line=<>) {
+
+    $line =~ s|[#!;].*$||;	# get rid of asm-style comments...
+    $line =~ s|/\*.*\*/||;	# ... and C-style comments...
+    $line =~ s|^\s+||;		# ... and skip white spaces in beginning...
+    $line =~ s|\s+$||;		# ... and at the end
+
+    {
+	$line =~ s|\b\.L(\w+)|L$1|g;	# common denominator for Locallabel
+	$line =~ s|\bL(\w+)|\.L$1|g	if ($dotinlocallabels);
+    }
+
+    {
+	$line =~ s|(^[\.\w]+)\:\s*||;
+	my $label = $1;
+	printf "%s:",($GLOBALS{$label} or $label) if ($label);
+    }
+
+    {
+	$line =~ s|^\s*(\.?)(\w+)([\.\+\-]?)\s*||;
+	my $c = $1; $c = "\t" if ($c eq "");
+	my $mnemonic = $2;
+	my $f = $3;
+	my $opcode = eval("\$$mnemonic");
+	$line =~ s|\bc?[rf]([0-9]+)\b|$1|g if ($c ne "." and $flavour !~ /osx/);
+	if (ref($opcode) eq 'CODE') { $line = &$opcode($f,split(',',$line)); }
+	elsif ($mnemonic)           { $line = $c.$mnemonic.$f."\t".$line; }
+    }
+
+    print $line if ($line);
+    print "\n";
+}
+
+close STDOUT;
diff --git a/crypto/perlasm/x86_64-xlate.pl b/crypto/perlasm/x86_64-xlate.pl
index fe348b99ad26..56d9b64b6fb3 100755
--- a/crypto/perlasm/x86_64-xlate.pl
+++ b/crypto/perlasm/x86_64-xlate.pl
@@ -1,6 +1,6 @@
 #!/usr/bin/env perl
 
-# Ascetic x86_64 AT&T to MASM assembler translator by <appro>.
+# Ascetic x86_64 AT&T to MASM/NASM assembler translator by <appro>.
 #
 # Why AT&T to MASM and not vice versa? Several reasons. Because AT&T
 # format is way easier to parse. Because it's simpler to "gear" from
@@ -20,12 +20,11 @@
 # Currently recognized limitations:
 #
 # - can't use multiple ops per line;
-# - indirect calls and jumps are not supported;
 #
 # Dual-ABI styling rules.
 #
-# 1. Adhere to Unix register and stack layout [see the end for
-#    explanation].
+# 1. Adhere to Unix register and stack layout [see cross-reference
+#    ABI "card" at the end for explanation].
 # 2. Forget about "red zone," stick to more traditional blended
 #    stack frame allocation. If volatile storage is actually required
 #    that is. If not, just leave the stack as is.
@@ -42,36 +41,64 @@
 # 6. Don't use [or hand-code with .byte] "rep ret." "ret" mnemonic is
 #    required to identify the spots, where to inject Win64 epilogue!
 #    But on the pros, it's then prefixed with rep automatically:-)
-# 7. Due to MASM limitations [and certain general counter-intuitivity
-#    of ip-relative addressing] generation of position-independent
-#    code is assisted by synthetic directive, .picmeup, which puts
-#    address of the *next* instruction into target register.
+# 7. Stick to explicit ip-relative addressing. If you have to use
+#    GOTPCREL addressing, stick to mov symbol@GOTPCREL(%rip),%r??.
+#    Both are recognized and translated to proper Win64 addressing
+#    modes. To support legacy code a synthetic directive, .picmeup,
+#    is implemented. It puts address of the *next* instruction into
+#    target register, e.g.:
 #
-#    Example 1:
 #		.picmeup	%rax
 #		lea		.Label-.(%rax),%rax
-#    Example 2:
-#		.picmeup	%rcx
-#	.Lpic_point:
-#		...
-#		lea		.Label-.Lpic_point(%rcx),%rbp
-
-my $output = shift;
+#
+# 8. In order to provide for structured exception handling unified
+#    Win64 prologue copies %rsp value to %rax. For further details
+#    see SEH paragraph at the end.
+# 9. .init segment is allowed to contain calls to functions only.
+# a. If function accepts more than 4 arguments *and* >4th argument
+#    is declared as non 64-bit value, do clear its upper part.
+
+my $flavour = shift;
+my $output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
 
-{ my ($stddev,$stdino,@junk)=stat(STDOUT);
-  my ($outdev,$outino,@junk)=stat($output);
+open STDOUT,">$output" || die "can't open $output: $!"
+	if (defined($output));
 
-    open STDOUT,">$output" || die "can't open $output: $!"
-	if ($stddev!=$outdev || $stdino!=$outino);
-}
+my $gas=1;	$gas=0 if ($output =~ /\.asm$/);
+my $elf=1;	$elf=0 if (!$gas);
+my $win64=0;
+my $prefix="";
+my $decor=".L";
 
 my $masmref=8 + 50727*2**-32;	# 8.00.50727 shipped with VS2005
-my $masm=$masmref if ($output =~ /\.asm/);
-if ($masm && `ml64 2>&1` =~ m/Version ([0-9]+)\.([0-9]+)(\.([0-9]+))?/)
-{   $masm=$1 + $2*2**-16 + $4*2**-32;   }
+my $masm=0;
+my $PTR=" PTR";
+
+my $nasmref=2.03;
+my $nasm=0;
+
+if    ($flavour eq "mingw64")	{ $gas=1; $elf=0; $win64=1;
+				  $prefix=`echo __USER_LABEL_PREFIX__ | $ENV{CC} -E -P -`;
+				  chomp($prefix);
+				}
+elsif ($flavour eq "macosx")	{ $gas=1; $elf=0; $prefix="_"; $decor="L\$"; }
+elsif ($flavour eq "masm")	{ $gas=0; $elf=0; $masm=$masmref; $win64=1; $decor="\$L\$"; }
+elsif ($flavour eq "nasm")	{ $gas=0; $elf=0; $nasm=$nasmref; $win64=1; $decor="\$L\$"; $PTR=""; }
+elsif (!$gas)
+{   if ($ENV{ASM} =~ m/nasm/ && `nasm -v` =~ m/version ([0-9]+)\.([0-9]+)/i)
+    {	$nasm = $1 + $2*0.01; $PTR="";  }
+    elsif (`ml64 2>&1` =~ m/Version ([0-9]+)\.([0-9]+)(\.([0-9]+))?/)
+    {	$masm = $1 + $2*2**-16 + $4*2**-32;   }
+    die "no assembler found on %PATH" if (!($nasm || $masm));
+    $win64=1;
+    $elf=0;
+    $decor="\$L\$";
+}
 
 my $current_segment;
 my $current_function;
+my %globals;
 
 { package opcode;	# pick up opcodes
     sub re {
@@ -85,11 +112,17 @@ my $current_function;
 	    $line = substr($line,@+[0]); $line =~ s/^\s+//;
 
 	    undef $self->{sz};
-	    if ($self->{op} =~ /^(movz)b.*/) {	# movz is pain...
+	    if ($self->{op} =~ /^(movz)x?([bw]).*/) {	# movz is pain...
 		$self->{op} = $1;
-		$self->{sz} = "b";
-	    } elsif ($self->{op} =~ /call/) {
-		$self->{sz} = ""
+		$self->{sz} = $2;
+	    } elsif ($self->{op} =~ /call|jmp/) {
+		$self->{sz} = "";
+	    } elsif ($self->{op} =~ /^p/ && $' !~ /^(ush|op|insrw)/) { # SSEn
+		$self->{sz} = "";
+	    } elsif ($self->{op} =~ /^v/) { # VEX
+		$self->{sz} = "";
+	    } elsif ($self->{op} =~ /movq/ && $line =~ /%xmm/) {
+		$self->{sz} = "";
 	    } elsif ($self->{op} =~ /([a-z]{3,})([qlwb])$/) {
 		$self->{op} = $1;
 		$self->{sz} = $2;
@@ -105,13 +138,20 @@ my $current_function;
     }
     sub out {
 	my $self = shift;
-	if (!$masm) {
+	if ($gas) {
 	    if ($self->{op} eq "movz") {	# movz is pain...
 		sprintf "%s%s%s",$self->{op},$self->{sz},shift;
 	    } elsif ($self->{op} =~ /^set/) { 
 		"$self->{op}";
 	    } elsif ($self->{op} eq "ret") {
-	    	".byte	0xf3,0xc3";
+		my $epilogue = "";
+		if ($win64 && $current_function->{abi} eq "svr4") {
+		    $epilogue = "movq	8(%rsp),%rdi\n\t" .
+				"movq	16(%rsp),%rsi\n\t";
+		}
+	    	$epilogue . ".byte	0xf3,0xc3";
+	    } elsif ($self->{op} eq "call" && !$elf && $current_segment eq ".init") {
+		".p2align\t3\n\t.quad";
 	    } else {
 		"$self->{op}$self->{sz}";
 	    }
@@ -119,15 +159,25 @@ my $current_function;
 	    $self->{op} =~ s/^movz/movzx/;
 	    if ($self->{op} eq "ret") {
 		$self->{op} = "";
-		if ($current_function->{abi} eq "svr4") {
-		    $self->{op} = "mov	rdi,QWORD PTR 8[rsp]\t;WIN64 epilogue\n\t".
-				  "mov	rsi,QWORD PTR 16[rsp]\n\t";
+		if ($win64 && $current_function->{abi} eq "svr4") {
+		    $self->{op} = "mov	rdi,QWORD${PTR}[8+rsp]\t;WIN64 epilogue\n\t".
+				  "mov	rsi,QWORD${PTR}[16+rsp]\n\t";
 	    	}
 		$self->{op} .= "DB\t0F3h,0C3h\t\t;repret";
-	    }
+	    } elsif ($self->{op} =~ /^(pop|push)f/) {
+		$self->{op} .= $self->{sz};
+	    } elsif ($self->{op} eq "call" && $current_segment eq ".CRT\$XCU") {
+		$self->{op} = "\tDQ";
+	    } 
 	    $self->{op};
 	}
     }
+    sub mnemonic {
+	my $self=shift;
+	my $op=shift;
+	$self->{op}=$op if (defined($op));
+	$self->{op};
+    }
 }
 { package const;	# pick up constants, which start with $
     sub re {
@@ -145,14 +195,15 @@ my $current_function;
     sub out {
     	my $self = shift;
 
-	if (!$masm) {
+	if ($gas) {
 	    # Solaris /usr/ccs/bin/as can't handle multiplications
 	    # in $self->{value}
-	    $self->{value} =~ s/(?<![0-9a-f])(0[x0-9a-f]+)/oct($1)/egi;
+	    $self->{value} =~ s/(?<![\w\$\.])(0x?[0-9a-f]+)/oct($1)/egi;
 	    $self->{value} =~ s/([0-9]+\s*[\*\/\%]\s*[0-9]+)/eval($1)/eg;
 	    sprintf "\$%s",$self->{value};
 	} else {
-	    $self->{value} =~ s/0x([0-9a-f]+)/0$1h/ig;
+	    $self->{value} =~ s/(0b[0-1]+)/oct($1)/eig;
+	    $self->{value} =~ s/0x([0-9a-f]+)/0$1h/ig if ($masm);
 	    sprintf "%s",$self->{value};
 	}
     }
@@ -163,13 +214,19 @@ my $current_function;
 	local	*line = shift;
 	undef	$ret;
 
-	if ($line =~ /^([^\(,]*)\(([%\w,]+)\)/) {
-	    $self->{label} = $1;
-	    ($self->{base},$self->{index},$self->{scale})=split(/,/,$2);
+	# optional * ---vvv--- appears in indirect jmp/call
+	if ($line =~ /^(\*?)([^\(,]*)\(([%\w,]+)\)/) {
+	    $self->{asterisk} = $1;
+	    $self->{label} = $2;
+	    ($self->{base},$self->{index},$self->{scale})=split(/,/,$3);
 	    $self->{scale} = 1 if (!defined($self->{scale}));
 	    $ret = $self;
 	    $line = substr($line,@+[0]); $line =~ s/^\s+//;
 
+	    if ($win64 && $self->{label} =~ s/\@GOTPCREL//) {
+		die if (opcode->mnemonic() ne "mov");
+		opcode->mnemonic("lea");
+	    }
 	    $self->{base}  =~ s/^%//;
 	    $self->{index} =~ s/^%// if (defined($self->{index}));
 	}
@@ -180,44 +237,54 @@ my $current_function;
     	my $self = shift;
 	my $sz = shift;
 
+	$self->{label} =~ s/([_a-z][_a-z0-9]*)/$globals{$1} or $1/gei;
+	$self->{label} =~ s/\.L/$decor/g;
+
 	# Silently convert all EAs to 64-bit. This is required for
 	# elder GNU assembler and results in more compact code,
 	# *but* most importantly AES module depends on this feature!
 	$self->{index} =~ s/^[er](.?[0-9xpi])[d]?$/r\1/;
 	$self->{base}  =~ s/^[er](.?[0-9xpi])[d]?$/r\1/;
 
-	if (!$masm) {
-	    # Solaris /usr/ccs/bin/as can't handle multiplications
-	    # in $self->{label}
-	    use integer;
-	    $self->{label} =~ s/(?<![0-9a-f])(0[x0-9a-f]+)/oct($1)/egi;
-	    $self->{label} =~ s/([0-9]+\s*[\*\/\%]\s*[0-9]+)/eval($1)/eg;
-	    $self->{label} =~ s/([0-9]+)/$1<<32>>32/eg;
+	# Solaris /usr/ccs/bin/as can't handle multiplications
+	# in $self->{label}, new gas requires sign extension...
+	use integer;
+	$self->{label} =~ s/(?<![\w\$\.])(0x?[0-9a-f]+)/oct($1)/egi;
+	$self->{label} =~ s/([0-9]+\s*[\*\/\%]\s*[0-9]+)/eval($1)/eg;
+	$self->{label} =~ s/([0-9]+)/$1<<32>>32/eg;
+
+	if ($gas) {
+	    $self->{label} =~ s/^___imp_/__imp__/   if ($flavour eq "mingw64");
 
 	    if (defined($self->{index})) {
-		sprintf "%s(%%%s,%%%s,%d)",
-					$self->{label},$self->{base},
+		sprintf "%s%s(%s,%%%s,%d)",$self->{asterisk},
+					$self->{label},
+					$self->{base}?"%$self->{base}":"",
 					$self->{index},$self->{scale};
 	    } else {
-		sprintf "%s(%%%s)",	$self->{label},$self->{base};
+		sprintf "%s%s(%%%s)",	$self->{asterisk},$self->{label},$self->{base};
 	    }
 	} else {
-	    %szmap = ( b=>"BYTE", w=>"WORD", l=>"DWORD", q=>"QWORD" );
+	    %szmap = (	b=>"BYTE$PTR", w=>"WORD$PTR", l=>"DWORD$PTR",
+	    		q=>"QWORD$PTR",o=>"OWORD$PTR",x=>"XMMWORD$PTR" );
 
 	    $self->{label} =~ s/\./\$/g;
-	    $self->{label} =~ s/0x([0-9a-f]+)/0$1h/ig;
+	    $self->{label} =~ s/(?<![\w\$\.])0x([0-9a-f]+)/0$1h/ig;
 	    $self->{label} = "($self->{label})" if ($self->{label} =~ /[\*\+\-\/]/);
+	    $sz="q" if ($self->{asterisk} || opcode->mnemonic() eq "movq");
+	    $sz="l" if (opcode->mnemonic() eq "movd");
 
 	    if (defined($self->{index})) {
-		sprintf "%s PTR %s[%s*%d+%s]",$szmap{$sz},
-					$self->{label},
+		sprintf "%s[%s%s*%d%s]",$szmap{$sz},
+					$self->{label}?"$self->{label}+":"",
 					$self->{index},$self->{scale},
-					$self->{base};
+					$self->{base}?"+$self->{base}":"";
 	    } elsif ($self->{base} eq "rip") {
-		sprintf "%s PTR %s",$szmap{$sz},$self->{label};
+		sprintf "%s[%s]",$szmap{$sz},$self->{label};
 	    } else {
-		sprintf "%s PTR %s[%s]",$szmap{$sz},
-					$self->{label},$self->{base};
+		sprintf "%s[%s%s]",$szmap{$sz},
+					$self->{label}?"$self->{label}+":"",
+					$self->{base};
 	    }
 	}
     }
@@ -229,9 +296,11 @@ my $current_function;
 	local	*line = shift;
 	undef	$ret;
 
-	if ($line =~ /^%(\w+)/) {
+	# optional * ---vvv--- appears in indirect jmp/call
+	if ($line =~ /^(\*?)%(\w+)/) {
 	    bless $self,$class;
-	    $self->{value} = $1;
+	    $self->{asterisk} = $1;
+	    $self->{value} = $2;
 	    $ret = $self;
 	    $line = substr($line,@+[0]); $line =~ s/^\s+//;
 	}
@@ -254,7 +323,8 @@ my $current_function;
     }
     sub out {
     	my $self = shift;
-	sprintf $masm?"%s":"%%%s",$self->{value};
+	if ($gas)	{ sprintf "%s%%%s",$self->{asterisk},$self->{value}; }
+	else		{ $self->{value}; }
     }
 }
 { package label;	# pick up labels, which end with :
@@ -263,37 +333,63 @@ my $current_function;
 	local	*line = shift;
 	undef	$ret;
 
-	if ($line =~ /(^[\.\w]+\:)/) {
+	if ($line =~ /(^[\.\w]+)\:/) {
 	    $self->{value} = $1;
 	    $ret = $self;
 	    $line = substr($line,@+[0]); $line =~ s/^\s+//;
 
-	    $self->{value} =~ s/\.L/\$L/ if ($masm);
+	    $self->{value} =~ s/^\.L/$decor/;
 	}
 	$ret;
     }
     sub out {
 	my $self = shift;
 
-	if (!$masm) {
-	    $self->{value};
-	} elsif ($self->{value} ne "$current_function->{name}:") {
-	    $self->{value};
-	} elsif ($current_function->{abi} eq "svr4") {
-	    my $func =	"$current_function->{name}	PROC\n".
-			"	mov	QWORD PTR 8[rsp],rdi\t;WIN64 prologue\n".
-			"	mov	QWORD PTR 16[rsp],rsi\n";
+	if ($gas) {
+	    my $func = ($globals{$self->{value}} or $self->{value}) . ":";
+	    if ($win64	&&
+			$current_function->{name} eq $self->{value} &&
+			$current_function->{abi} eq "svr4") {
+		$func .= "\n";
+		$func .= "	movq	%rdi,8(%rsp)\n";
+		$func .= "	movq	%rsi,16(%rsp)\n";
+		$func .= "	movq	%rsp,%rax\n";
+		$func .= "${decor}SEH_begin_$current_function->{name}:\n";
+		my $narg = $current_function->{narg};
+		$narg=6 if (!defined($narg));
+		$func .= "	movq	%rcx,%rdi\n" if ($narg>0);
+		$func .= "	movq	%rdx,%rsi\n" if ($narg>1);
+		$func .= "	movq	%r8,%rdx\n"  if ($narg>2);
+		$func .= "	movq	%r9,%rcx\n"  if ($narg>3);
+		$func .= "	movq	40(%rsp),%r8\n" if ($narg>4);
+		$func .= "	movq	48(%rsp),%r9\n" if ($narg>5);
+	    }
+	    $func;
+	} elsif ($self->{value} ne "$current_function->{name}") {
+	    $self->{value} .= ":" if ($masm && $ret!~m/^\$/);
+	    $self->{value} . ":";
+	} elsif ($win64 && $current_function->{abi} eq "svr4") {
+	    my $func =	"$current_function->{name}" .
+			($nasm ? ":" : "\tPROC $current_function->{scope}") .
+			"\n";
+	    $func .= "	mov	QWORD${PTR}[8+rsp],rdi\t;WIN64 prologue\n";
+	    $func .= "	mov	QWORD${PTR}[16+rsp],rsi\n";
+	    $func .= "	mov	rax,rsp\n";
+	    $func .= "${decor}SEH_begin_$current_function->{name}:";
+	    $func .= ":" if ($masm);
+	    $func .= "\n";
 	    my $narg = $current_function->{narg};
 	    $narg=6 if (!defined($narg));
 	    $func .= "	mov	rdi,rcx\n" if ($narg>0);
 	    $func .= "	mov	rsi,rdx\n" if ($narg>1);
 	    $func .= "	mov	rdx,r8\n"  if ($narg>2);
 	    $func .= "	mov	rcx,r9\n"  if ($narg>3);
-	    $func .= "	mov	r8,QWORD PTR 40[rsp]\n" if ($narg>4);
-	    $func .= "	mov	r9,QWORD PTR 48[rsp]\n" if ($narg>5);
+	    $func .= "	mov	r8,QWORD${PTR}[40+rsp]\n" if ($narg>4);
+	    $func .= "	mov	r9,QWORD${PTR}[48+rsp]\n" if ($narg>5);
 	    $func .= "\n";
 	} else {
-	   "$current_function->{name}	PROC";
+	   "$current_function->{name}".
+			($nasm ? ":" : "\tPROC $current_function->{scope}");
 	}
     }
 }
@@ -308,13 +404,19 @@ my $current_function;
 	    $ret = $self;
 	    $line = substr($line,@+[0]); $line =~ s/^\s+//;
 
-	    $self->{value} =~ s/\.L/\$L/g if ($masm);
+	    $self->{value} =~ s/\@PLT// if (!$elf);
+	    $self->{value} =~ s/([_a-z][_a-z0-9]*)/$globals{$1} or $1/gei;
+	    $self->{value} =~ s/\.L/$decor/g;
 	}
 	$ret;
     }
     sub out {
 	my $self = shift;
-	$self->{value};
+	if ($nasm && opcode->mnemonic()=~m/^j/) {
+	    "NEAR ".$self->{value};
+	} else {
+	    $self->{value};
+	}
     }
 }
 { package directive;	# pick up directives, which start with .
@@ -334,89 +436,205 @@ my $current_function;
 			"%r14"=>0x01358d4c,	"%r15"=>0x013d8d4c	);
 
 	if ($line =~ /^\s*(\.\w+)/) {
-	    if (!$masm) {
-		$self->{value} = $1;
-		$line =~ s/\@abi\-omnipotent/\@function/;
-		$line =~ s/\@function.*/\@function/;
-		if ($line =~ /\.picmeup\s+(%r[\w]+)/i) {
-		    $self->{value} = sprintf "\t.long\t0x%x,0x90000000",$opcode{$1};
-		} elsif ($line =~ /\.asciz\s+"(.*)"$/) {
-		    $self->{value} = ".byte\t".join(",",unpack("C*",$1),0);
-		} elsif ($line =~ /\.extern/) {
-		    $self->{value} = ""; # swallow extern
-		} else {
-		    $self->{value} = $line;
-		}
-		$line = "";
-		return $self;
-	    }
-
 	    $dir = $1;
 	    $ret = $self;
 	    undef $self->{value};
 	    $line = substr($line,@+[0]); $line =~ s/^\s+//;
+
 	    SWITCH: for ($dir) {
-		/\.(text)/
-			    && do { my $v=undef;
-				    $v="$current_segment\tENDS\n" if ($current_segment);
-				    $current_segment = "_$1\$";
-				    $current_segment =~ tr/[a-z]/[A-Z]/;
-				    $v.="$current_segment\tSEGMENT ";
-				    $v.=$masm>=$masmref ? "ALIGN(64)" : "PAGE";
-				    $v.=" 'CODE'";
-				    $self->{value} = $v;
+		/\.picmeup/ && do { if ($line =~ /(%r[\w]+)/i) {
+			    		$dir="\t.long";
+					$line=sprintf "0x%x,0x90000000",$opcode{$1};
+				    }
+				    last;
+				  };
+		/\.global|\.globl|\.extern/
+			    && do { $globals{$line} = $prefix . $line;
+				    $line = $globals{$line} if ($prefix);
 				    last;
 				  };
-		/\.extern/  && do { $self->{value} = "EXTRN\t".$line.":BYTE"; last;  };
-		/\.globl/   && do { $self->{value} = "PUBLIC\t".$line; last; };
 		/\.type/    && do { ($sym,$type,$narg) = split(',',$line);
 				    if ($type eq "\@function") {
 					undef $current_function;
 					$current_function->{name} = $sym;
 					$current_function->{abi}  = "svr4";
 					$current_function->{narg} = $narg;
+					$current_function->{scope} = defined($globals{$sym})?"PUBLIC":"PRIVATE";
 				    } elsif ($type eq "\@abi-omnipotent") {
 					undef $current_function;
 					$current_function->{name} = $sym;
+					$current_function->{scope} = defined($globals{$sym})?"PUBLIC":"PRIVATE";
+				    }
+				    $line =~ s/\@abi\-omnipotent/\@function/;
+				    $line =~ s/\@function.*/\@function/;
+				    last;
+				  };
+		/\.asciz/   && do { if ($line =~ /^"(.*)"$/) {
+					$dir  = ".byte";
+					$line = join(",",unpack("C*",$1),0);
+				    }
+				    last;
+				  };
+		/\.rva|\.long|\.quad/
+			    && do { $line =~ s/([_a-z][_a-z0-9]*)/$globals{$1} or $1/gei;
+				    $line =~ s/\.L/$decor/g;
+				    last;
+				  };
+	    }
+
+	    if ($gas) {
+		$self->{value} = $dir . "\t" . $line;
+
+		if ($dir =~ /\.extern/) {
+		    $self->{value} = ""; # swallow extern
+		} elsif (!$elf && $dir =~ /\.type/) {
+		    $self->{value} = "";
+		    $self->{value} = ".def\t" . ($globals{$1} or $1) . ";\t" .
+				(defined($globals{$1})?".scl 2;":".scl 3;") .
+				"\t.type 32;\t.endef"
+				if ($win64 && $line =~ /([^,]+),\@function/);
+		} elsif (!$elf && $dir =~ /\.size/) {
+		    $self->{value} = "";
+		    if (defined($current_function)) {
+			$self->{value} .= "${decor}SEH_end_$current_function->{name}:"
+				if ($win64 && $current_function->{abi} eq "svr4");
+			undef $current_function;
+		    }
+		} elsif (!$elf && $dir =~ /\.align/) {
+		    $self->{value} = ".p2align\t" . (log($line)/log(2));
+		} elsif ($dir eq ".section") {
+		    $current_segment=$line;
+		    if (!$elf && $current_segment eq ".init") {
+			if	($flavour eq "macosx")	{ $self->{value} = ".mod_init_func"; }
+			elsif	($flavour eq "mingw64")	{ $self->{value} = ".section\t.ctors"; }
+		    }
+		} elsif ($dir =~ /\.(text|data)/) {
+		    $current_segment=".$1";
+		} elsif ($dir =~ /\.hidden/) {
+		    if    ($flavour eq "macosx")  { $self->{value} = ".private_extern\t$prefix$line"; }
+		    elsif ($flavour eq "mingw64") { $self->{value} = ""; }
+		} elsif ($dir =~ /\.comm/) {
+		    $self->{value} = "$dir\t$prefix$line";
+		    $self->{value} =~ s|,([0-9]+),([0-9]+)$|",$1,".log($2)/log(2)|e if ($flavour eq "macosx");
+		}
+		$line = "";
+		return $self;
+	    }
+
+	    # non-gas case or nasm/masm
+	    SWITCH: for ($dir) {
+		/\.text/    && do { my $v=undef;
+				    if ($nasm) {
+					$v="section	.text code align=64\n";
+				    } else {
+					$v="$current_segment\tENDS\n" if ($current_segment);
+					$current_segment = ".text\$";
+					$v.="$current_segment\tSEGMENT ";
+					$v.=$masm>=$masmref ? "ALIGN(64)" : "PAGE";
+					$v.=" 'CODE'";
+				    }
+				    $self->{value} = $v;
+				    last;
+				  };
+		/\.data/    && do { my $v=undef;
+				    if ($nasm) {
+					$v="section	.data data align=8\n";
+				    } else {
+					$v="$current_segment\tENDS\n" if ($current_segment);
+					$current_segment = "_DATA";
+					$v.="$current_segment\tSEGMENT";
+				    }
+				    $self->{value} = $v;
+				    last;
+				  };
+		/\.section/ && do { my $v=undef;
+				    $line =~ s/([^,]*).*/$1/;
+				    $line = ".CRT\$XCU" if ($line eq ".init");
+				    if ($nasm) {
+					$v="section	$line";
+					if ($line=~/\.([px])data/) {
+					    $v.=" rdata align=";
+					    $v.=$1 eq "p"? 4 : 8;
+					} elsif ($line=~/\.CRT\$/i) {
+					    $v.=" rdata align=8";
+					}
+				    } else {
+					$v="$current_segment\tENDS\n" if ($current_segment);
+					$v.="$line\tSEGMENT";
+					if ($line=~/\.([px])data/) {
+					    $v.=" READONLY";
+					    $v.=" ALIGN(".($1 eq "p" ? 4 : 8).")" if ($masm>=$masmref);
+					} elsif ($line=~/\.CRT\$/i) {
+					    $v.=" READONLY ";
+					    $v.=$masm>=$masmref ? "ALIGN(8)" : "DWORD";
+					}
 				    }
+				    $current_segment = $line;
+				    $self->{value} = $v;
+				    last;
+				  };
+		/\.extern/  && do { $self->{value}  = "EXTERN\t".$line;
+				    $self->{value} .= ":NEAR" if ($masm);
+				    last;
+				  };
+		/\.globl|.global/
+			    && do { $self->{value}  = $masm?"PUBLIC":"global";
+				    $self->{value} .= "\t".$line;
 				    last;
 				  };
 		/\.size/    && do { if (defined($current_function)) {
-					$self->{value}="$current_function->{name}\tENDP";
+					undef $self->{value};
+					if ($current_function->{abi} eq "svr4") {
+					    $self->{value}="${decor}SEH_end_$current_function->{name}:";
+					    $self->{value}.=":\n" if($masm);
+					}
+					$self->{value}.="$current_function->{name}\tENDP" if($masm && $current_function->{name});
 					undef $current_function;
 				    }
 				    last;
 				  };
 		/\.align/   && do { $self->{value} = "ALIGN\t".$line; last; };
-		/\.(byte|value|long|quad)/
-			    && do { my @arr = split(',',$line);
-				    my $sz  = substr($1,0,1);
+		/\.(value|long|rva|quad)/
+			    && do { my $sz  = substr($1,0,1);
+				    my @arr = split(/,\s*/,$line);
 				    my $last = pop(@arr);
 				    my $conv = sub  {	my $var=shift;
-							if ($var=~s/0x([0-9a-f]+)/0$1h/i) { $var; }
-							else { sprintf"0%Xh",$var; }
+							$var=~s/^(0b[0-1]+)/oct($1)/eig;
+							$var=~s/^0x([0-9a-f]+)/0$1h/ig if ($masm);
+							if ($sz eq "D" && ($current_segment=~/.[px]data/ || $dir eq ".rva"))
+							{ $var=~s/([_a-z\$\@][_a-z0-9\$\@]*)/$nasm?"$1 wrt ..imagebase":"imagerel $1"/egi; }
+							$var;
 						    };  
 
-				    $sz =~ tr/bvlq/BWDQ/;
+				    $sz =~ tr/bvlrq/BWDDQ/;
 				    $self->{value} = "\tD$sz\t";
 				    for (@arr) { $self->{value} .= &$conv($_).","; }
 				    $self->{value} .= &$conv($last);
 				    last;
 				  };
-		/\.picmeup/ && do { $self->{value} = sprintf"\tDD\t 0%Xh,090000000h",$opcode{$line};
-				    last;
-				  };
-		/\.asciz/   && do { if ($line =~ /^"(.*)"$/) {
-					my @str=unpack("C*",$1);
-					push @str,0;
-					while ($#str>15) {
-					    $self->{value}.="DB\t"
-						.join(",",@str[0..15])."\n";
-					    foreach (0..15) { shift @str; }
-					}
+		/\.byte/    && do { my @str=split(/,\s*/,$line);
+				    map(s/(0b[0-1]+)/oct($1)/eig,@str);
+				    map(s/0x([0-9a-f]+)/0$1h/ig,@str) if ($masm);	
+				    while ($#str>15) {
 					$self->{value}.="DB\t"
+						.join(",",@str[0..15])."\n";
+					foreach (0..15) { shift @str; }
+				    }
+				    $self->{value}.="DB\t"
 						.join(",",@str) if (@str);
+				    last;
+				  };
+		/\.comm/    && do { my @str=split(/,\s*/,$line);
+				    my $v=undef;
+				    if ($nasm) {
+					$v.="common	$prefix@str[0] @str[1]";
+				    } else {
+					$v="$current_segment\tENDS\n" if ($current_segment);
+					$current_segment = "_DATA";
+					$v.="$current_segment\tSEGMENT\n";
+					$v.="COMM	@str[0]:DWORD:".@str[1]/4;
 				    }
+				    $self->{value} = $v;
 				    last;
 				  };
 	    }
@@ -431,6 +649,139 @@ my $current_function;
     }
 }
 
+sub rex {
+ local *opcode=shift;
+ my ($dst,$src,$rex)=@_;
+
+   $rex|=0x04 if($dst>=8);
+   $rex|=0x01 if($src>=8);
+   push @opcode,($rex|0x40) if ($rex);
+}
+
+# older gas and ml64 don't handle SSE>2 instructions
+my %regrm = (	"%eax"=>0, "%ecx"=>1, "%edx"=>2, "%ebx"=>3,
+		"%esp"=>4, "%ebp"=>5, "%esi"=>6, "%edi"=>7	);
+
+my $movq = sub {	# elderly gas can't handle inter-register movq
+  my $arg = shift;
+  my @opcode=(0x66);
+    if ($arg =~ /%xmm([0-9]+),\s*%r(\w+)/) {
+	my ($src,$dst)=($1,$2);
+	if ($dst !~ /[0-9]+/)	{ $dst = $regrm{"%e$dst"}; }
+	rex(\@opcode,$src,$dst,0x8);
+	push @opcode,0x0f,0x7e;
+	push @opcode,0xc0|(($src&7)<<3)|($dst&7);	# ModR/M
+	@opcode;
+    } elsif ($arg =~ /%r(\w+),\s*%xmm([0-9]+)/) {
+	my ($src,$dst)=($2,$1);
+	if ($dst !~ /[0-9]+/)	{ $dst = $regrm{"%e$dst"}; }
+	rex(\@opcode,$src,$dst,0x8);
+	push @opcode,0x0f,0x6e;
+	push @opcode,0xc0|(($src&7)<<3)|($dst&7);	# ModR/M
+	@opcode;
+    } else {
+	();
+    }
+};
+
+my $pextrd = sub {
+    if (shift =~ /\$([0-9]+),\s*%xmm([0-9]+),\s*(%\w+)/) {
+      my @opcode=(0x66);
+	$imm=$1;
+	$src=$2;
+	$dst=$3;
+	if ($dst =~ /%r([0-9]+)d/)	{ $dst = $1; }
+	elsif ($dst =~ /%e/)		{ $dst = $regrm{$dst}; }
+	rex(\@opcode,$src,$dst);
+	push @opcode,0x0f,0x3a,0x16;
+	push @opcode,0xc0|(($src&7)<<3)|($dst&7);	# ModR/M
+	push @opcode,$imm;
+	@opcode;
+    } else {
+	();
+    }
+};
+
+my $pinsrd = sub {
+    if (shift =~ /\$([0-9]+),\s*(%\w+),\s*%xmm([0-9]+)/) {
+      my @opcode=(0x66);
+	$imm=$1;
+	$src=$2;
+	$dst=$3;
+	if ($src =~ /%r([0-9]+)/)	{ $src = $1; }
+	elsif ($src =~ /%e/)		{ $src = $regrm{$src}; }
+	rex(\@opcode,$dst,$src);
+	push @opcode,0x0f,0x3a,0x22;
+	push @opcode,0xc0|(($dst&7)<<3)|($src&7);	# ModR/M
+	push @opcode,$imm;
+	@opcode;
+    } else {
+	();
+    }
+};
+
+my $pshufb = sub {
+    if (shift =~ /%xmm([0-9]+),\s*%xmm([0-9]+)/) {
+      my @opcode=(0x66);
+	rex(\@opcode,$2,$1);
+	push @opcode,0x0f,0x38,0x00;
+	push @opcode,0xc0|($1&7)|(($2&7)<<3);		# ModR/M
+	@opcode;
+    } else {
+	();
+    }
+};
+
+my $palignr = sub {
+    if (shift =~ /\$([0-9]+),\s*%xmm([0-9]+),\s*%xmm([0-9]+)/) {
+      my @opcode=(0x66);
+	rex(\@opcode,$3,$2);
+	push @opcode,0x0f,0x3a,0x0f;
+	push @opcode,0xc0|($2&7)|(($3&7)<<3);		# ModR/M
+	push @opcode,$1;
+	@opcode;
+    } else {
+	();
+    }
+};
+
+my $pclmulqdq = sub {
+    if (shift =~ /\$([x0-9a-f]+),\s*%xmm([0-9]+),\s*%xmm([0-9]+)/) {
+      my @opcode=(0x66);
+	rex(\@opcode,$3,$2);
+	push @opcode,0x0f,0x3a,0x44;
+	push @opcode,0xc0|($2&7)|(($3&7)<<3);		# ModR/M
+	my $c=$1;
+	push @opcode,$c=~/^0/?oct($c):$c;
+	@opcode;
+    } else {
+	();
+    }
+};
+
+my $rdrand = sub {
+    if (shift =~ /%[er](\w+)/) {
+      my @opcode=();
+      my $dst=$1;
+	if ($dst !~ /[0-9]+/) { $dst = $regrm{"%e$dst"}; }
+	rex(\@opcode,0,$1,8);
+	push @opcode,0x0f,0xc7,0xf0|($dst&7);
+	@opcode;
+    } else {
+	();
+    }
+};
+
+if ($nasm) {
+    print <<___;
+default	rel
+%define XMMWORD
+___
+} elsif ($masm) {
+    print <<___;
+OPTION	DOTNAME
+___
+}
 while($line=<>) {
 
     chomp($line);
@@ -441,43 +792,57 @@ while($line=<>) {
 
     undef $label;
     undef $opcode;
-    undef $dst;
-    undef $src;
-    undef $sz;
+    undef @args;
 
     if ($label=label->re(\$line))	{ print $label->out(); }
 
     if (directive->re(\$line)) {
 	printf "%s",directive->out();
-    } elsif ($opcode=opcode->re(\$line)) { ARGUMENT: {
+    } elsif ($opcode=opcode->re(\$line)) {
+	my $asm = eval("\$".$opcode->mnemonic());
+	undef @bytes;
+	
+	if ((ref($asm) eq 'CODE') && scalar(@bytes=&$asm($line))) {
+	    print $gas?".byte\t":"DB\t",join(',',@bytes),"\n";
+	    next;
+	}
 
-	if ($src=register->re(\$line))	{ opcode->size($src->size()); }
-	elsif ($src=const->re(\$line))	{ }
-	elsif ($src=ea->re(\$line))	{ }
-	elsif ($src=expr->re(\$line))	{ }
+	ARGUMENT: while (1) {
+	my $arg;
 
-	last ARGUMENT if ($line !~ /^,/);
+	if ($arg=register->re(\$line))	{ opcode->size($arg->size()); }
+	elsif ($arg=const->re(\$line))	{ }
+	elsif ($arg=ea->re(\$line))	{ }
+	elsif ($arg=expr->re(\$line))	{ }
+	else				{ last ARGUMENT; }
 
-	$line = substr($line,1); $line =~ s/^\s+//;
+	push @args,$arg;
 
-	if ($dst=register->re(\$line))	{ opcode->size($dst->size()); }
-	elsif ($dst=const->re(\$line))	{ }
-	elsif ($dst=ea->re(\$line))	{ }
+	last ARGUMENT if ($line !~ /^,/);
 
+	$line =~ s/^,\s*//;
 	} # ARGUMENT:
 
-	$sz=opcode->size();
+	if ($#args>=0) {
+	    my $insn;
+	    my $sz=opcode->size();
 
-	if (defined($dst)) {
-	    if (!$masm) {
-		printf "\t%s\t%s,%s",	$opcode->out($dst->size()),
-					$src->out($sz),$dst->out($sz);
+	    if ($gas) {
+		$insn = $opcode->out($#args>=1?$args[$#args]->size():$sz);
+		@args = map($_->out($sz),@args);
+		printf "\t%s\t%s",$insn,join(",",@args);
 	    } else {
-		printf "\t%s\t%s,%s",	$opcode->out(),
-					$dst->out($sz),$src->out($sz);
+		$insn = $opcode->out();
+		foreach (@args) {
+		    my $arg = $_->out();
+		    # $insn.=$sz compensates for movq, pinsrw, ...
+		    if ($arg =~ /^xmm[0-9]+$/) { $insn.=$sz; $sz="x" if(!$sz); last; }
+		    if ($arg =~ /^mm[0-9]+$/)  { $insn.=$sz; $sz="q" if(!$sz); last; }
+		}
+		@args = reverse(@args);
+		undef $sz if ($nasm && $opcode->mnemonic() eq "lea");
+		printf "\t%s\t%s",$insn,join(",",map($_->out($sz),@args));
 	    }
-	} elsif (defined($src)) {
-	    printf "\t%s\t%s",$opcode->out(),$src->out($sz);
 	} else {
 	    printf "\t%s",$opcode->out();
 	}
@@ -486,11 +851,12 @@ while($line=<>) {
     print $line,"\n";
 }
 
-print "\n$current_segment\tENDS\nEND\n" if ($masm);
+print "\n$current_segment\tENDS\n"	if ($current_segment && $masm);
+print "END\n"				if ($masm);
 
 close STDOUT;
 
-#################################################
+#################################################
 # Cross-reference x86_64 ABI "card"
 #
 # 		Unix		Win64
@@ -554,3 +920,161 @@ close STDOUT;
 #	movq	16(%rsp),%rsi
 # endif
 #	ret
+#
+#################################################
+# Win64 SEH, Structured Exception Handling.
+#
+# Unlike on Unix systems(*) lack of Win64 stack unwinding information
+# has undesired side-effect at run-time: if an exception is raised in
+# assembler subroutine such as those in question (basically we're
+# referring to segmentation violations caused by malformed input
+# parameters), the application is briskly terminated without invoking
+# any exception handlers, most notably without generating memory dump
+# or any user notification whatsoever. This poses a problem. It's
+# possible to address it by registering custom language-specific
+# handler that would restore processor context to the state at
+# subroutine entry point and return "exception is not handled, keep
+# unwinding" code. Writing such handler can be a challenge... But it's
+# doable, though requires certain coding convention. Consider following
+# snippet:
+#
+# .type	function,@function
+# function:
+#	movq	%rsp,%rax	# copy rsp to volatile register
+#	pushq	%r15		# save non-volatile registers
+#	pushq	%rbx
+#	pushq	%rbp
+#	movq	%rsp,%r11
+#	subq	%rdi,%r11	# prepare [variable] stack frame
+#	andq	$-64,%r11
+#	movq	%rax,0(%r11)	# check for exceptions
+#	movq	%r11,%rsp	# allocate [variable] stack frame
+#	movq	%rax,0(%rsp)	# save original rsp value
+# magic_point:
+#	...
+#	movq	0(%rsp),%rcx	# pull original rsp value
+#	movq	-24(%rcx),%rbp	# restore non-volatile registers
+#	movq	-16(%rcx),%rbx
+#	movq	-8(%rcx),%r15
+#	movq	%rcx,%rsp	# restore original rsp
+#	ret
+# .size function,.-function
+#
+# The key is that up to magic_point copy of original rsp value remains
+# in chosen volatile register and no non-volatile register, except for
+# rsp, is modified. While past magic_point rsp remains constant till
+# the very end of the function. In this case custom language-specific
+# exception handler would look like this:
+#
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+# {	ULONG64 *rsp = (ULONG64 *)context->Rax;
+#	if (context->Rip >= magic_point)
+#	{   rsp = ((ULONG64 **)context->Rsp)[0];
+#	    context->Rbp = rsp[-3];
+#	    context->Rbx = rsp[-2];
+#	    context->R15 = rsp[-1];
+#	}
+#	context->Rsp = (ULONG64)rsp;
+#	context->Rdi = rsp[1];
+#	context->Rsi = rsp[2];
+#
+#	memcpy (disp->ContextRecord,context,sizeof(CONTEXT));
+#	RtlVirtualUnwind(UNW_FLAG_NHANDLER,disp->ImageBase,
+#		dips->ControlPc,disp->FunctionEntry,disp->ContextRecord,
+#		&disp->HandlerData,&disp->EstablisherFrame,NULL);
+#	return ExceptionContinueSearch;
+# }
+#
+# It's appropriate to implement this handler in assembler, directly in
+# function's module. In order to do that one has to know members'
+# offsets in CONTEXT and DISPATCHER_CONTEXT structures and some constant
+# values. Here they are:
+#
+#	CONTEXT.Rax				120
+#	CONTEXT.Rcx				128
+#	CONTEXT.Rdx				136
+#	CONTEXT.Rbx				144
+#	CONTEXT.Rsp				152
+#	CONTEXT.Rbp				160
+#	CONTEXT.Rsi				168
+#	CONTEXT.Rdi				176
+#	CONTEXT.R8				184
+#	CONTEXT.R9				192
+#	CONTEXT.R10				200
+#	CONTEXT.R11				208
+#	CONTEXT.R12				216
+#	CONTEXT.R13				224
+#	CONTEXT.R14				232
+#	CONTEXT.R15				240
+#	CONTEXT.Rip				248
+#	CONTEXT.Xmm6				512
+#	sizeof(CONTEXT)				1232
+#	DISPATCHER_CONTEXT.ControlPc		0
+#	DISPATCHER_CONTEXT.ImageBase		8
+#	DISPATCHER_CONTEXT.FunctionEntry	16
+#	DISPATCHER_CONTEXT.EstablisherFrame	24
+#	DISPATCHER_CONTEXT.TargetIp		32
+#	DISPATCHER_CONTEXT.ContextRecord	40
+#	DISPATCHER_CONTEXT.LanguageHandler	48
+#	DISPATCHER_CONTEXT.HandlerData		56
+#	UNW_FLAG_NHANDLER			0
+#	ExceptionContinueSearch			1
+#
+# In order to tie the handler to the function one has to compose
+# couple of structures: one for .xdata segment and one for .pdata.
+#
+# UNWIND_INFO structure for .xdata segment would be
+#
+# function_unwind_info:
+#	.byte	9,0,0,0
+#	.rva	handler
+#
+# This structure designates exception handler for a function with
+# zero-length prologue, no stack frame or frame register.
+#
+# To facilitate composing of .pdata structures, auto-generated "gear"
+# prologue copies rsp value to rax and denotes next instruction with
+# .LSEH_begin_{function_name} label. This essentially defines the SEH
+# styling rule mentioned in the beginning. Position of this label is
+# chosen in such manner that possible exceptions raised in the "gear"
+# prologue would be accounted to caller and unwound from latter's frame.
+# End of function is marked with respective .LSEH_end_{function_name}
+# label. To summarize, .pdata segment would contain
+#
+#	.rva	.LSEH_begin_function
+#	.rva	.LSEH_end_function
+#	.rva	function_unwind_info
+#
+# Reference to functon_unwind_info from .xdata segment is the anchor.
+# In case you wonder why references are 32-bit .rvas and not 64-bit
+# .quads. References put into these two segments are required to be
+# *relative* to the base address of the current binary module, a.k.a.
+# image base. No Win64 module, be it .exe or .dll, can be larger than
+# 2GB and thus such relative references can be and are accommodated in
+# 32 bits.
+#
+# Having reviewed the example function code, one can argue that "movq
+# %rsp,%rax" above is redundant. It is not! Keep in mind that on Unix
+# rax would contain an undefined value. If this "offends" you, use
+# another register and refrain from modifying rax till magic_point is
+# reached, i.e. as if it was a non-volatile register. If more registers
+# are required prior [variable] frame setup is completed, note that
+# nobody says that you can have only one "magic point." You can
+# "liberate" non-volatile registers by denoting last stack off-load
+# instruction and reflecting it in finer grade unwind logic in handler.
+# After all, isn't it why it's called *language-specific* handler...
+#
+# Attentive reader can notice that exceptions would be mishandled in
+# auto-generated "gear" epilogue. Well, exception effectively can't
+# occur there, because if memory area used by it was subject to
+# segmentation violation, then it would be raised upon call to the
+# function (and as already mentioned be accounted to caller, which is
+# not a problem). If you're still not comfortable, then define tail
+# "magic point" just prior ret instruction and have handler treat it...
+#
+# (*)	Note that we're talking about run-time, not debug-time. Lack of
+#	unwind information makes debugging hard on both Windows and
+#	Unix. "Unlike" referes to the fact that on Unix signal handler
+#	will always be invoked, core dumped and appropriate exit code
+#	returned to parent (for user notification).
diff --git a/crypto/perlasm/x86asm.pl b/crypto/perlasm/x86asm.pl
index 5979122158fd..eb543db2f66e 100644
--- a/crypto/perlasm/x86asm.pl
+++ b/crypto/perlasm/x86asm.pl
@@ -1,130 +1,260 @@
-#!/usr/local/bin/perl
+#!/usr/bin/env perl
 
 # require 'x86asm.pl';
-# &asm_init("cpp","des-586.pl");
-# XXX
-# XXX
-# main'asm_finish
-
-sub main'asm_finish
-	{
-	&file_end();
-	&asm_finish_cpp() if $cpp;
-	print &asm_get_output();
-	}
-
-sub main'asm_init
-	{
-	($type,$fn,$i386)=@_;
-	$filename=$fn;
-
-	$elf=$cpp=$coff=$aout=$win32=$netware=$mwerks=0;
-	if (	($type eq "elf"))
-		{ $elf=1; require "x86unix.pl"; }
-	elsif (	($type eq "a.out"))
-		{ $aout=1; require "x86unix.pl"; }
-	elsif (	($type eq "coff" or $type eq "gaswin"))
-		{ $coff=1; require "x86unix.pl"; }
-	elsif (	($type eq "cpp"))
-		{ $cpp=1; require "x86unix.pl"; }
-	elsif (	($type eq "win32"))
-		{ $win32=1; require "x86ms.pl"; }
-	elsif (	($type eq "win32n"))
-		{ $win32=1; require "x86nasm.pl"; }
-	elsif (	($type eq "nw-nasm"))
-		{ $netware=1; require "x86nasm.pl"; }
-	elsif (	($type eq "nw-mwasm"))
-		{ $netware=1; $mwerks=1; require "x86nasm.pl"; }
-	else
-		{
-		print STDERR <<"EOF";
+# &asm_init(<flavor>,"des-586.pl"[,$i386only]);
+# &function_begin("foo");
+# ...
+# &function_end("foo");
+# &asm_finish
+
+$out=();
+$i386=0;
+
+# AUTOLOAD is this context has quite unpleasant side effect, namely
+# that typos in function calls effectively go to assembler output,
+# but on the pros side we don't have to implement one subroutine per
+# each opcode...
+sub ::AUTOLOAD
+{ my $opcode = $AUTOLOAD;
+
+    die "more than 4 arguments passed to $opcode" if ($#_>3);
+
+    $opcode =~ s/.*:://;
+    if    ($opcode =~ /^push/) { $stack+=4; }
+    elsif ($opcode =~ /^pop/)  { $stack-=4; }
+
+    &generic($opcode,@_) or die "undefined subroutine \&$AUTOLOAD";
+}
+
+sub ::emit
+{ my $opcode=shift;
+
+    if ($#_==-1)    { push(@out,"\t$opcode\n");				}
+    else            { push(@out,"\t$opcode\t".join(',',@_)."\n");	}
+}
+
+sub ::LB
+{   $_[0] =~ m/^e?([a-d])x$/o or die "$_[0] does not have a 'low byte'";
+  $1."l";
+}
+sub ::HB
+{   $_[0] =~ m/^e?([a-d])x$/o or die "$_[0] does not have a 'high byte'";
+  $1."h";
+}
+sub ::stack_push{ my $num=$_[0]*4; $stack+=$num; &sub("esp",$num);	}
+sub ::stack_pop	{ my $num=$_[0]*4; $stack-=$num; &add("esp",$num);	}
+sub ::blindpop	{ &pop($_[0]); $stack+=4;				}
+sub ::wparam	{ &DWP($stack+4*$_[0],"esp");				}
+sub ::swtmp	{ &DWP(4*$_[0],"esp");					}
+
+sub ::bswap
+{   if ($i386)	# emulate bswap for i386
+    {	&comment("bswap @_");
+	&xchg(&HB(@_),&LB(@_));
+	&ror (@_,16);
+	&xchg(&HB(@_),&LB(@_));
+    }
+    else
+    {	&generic("bswap",@_);	}
+}
+# These are made-up opcodes introduced over the years essentially
+# by ignorance, just alias them to real ones...
+sub ::movb	{ &mov(@_);	}
+sub ::xorb	{ &xor(@_);	}
+sub ::rotl	{ &rol(@_);	}
+sub ::rotr	{ &ror(@_);	}
+sub ::exch	{ &xchg(@_);	}
+sub ::halt	{ &hlt;		}
+sub ::movz	{ &movzx(@_);	}
+sub ::pushf	{ &pushfd;	}
+sub ::popf	{ &popfd;	}
+
+# 3 argument instructions
+sub ::movq
+{ my($p1,$p2,$optimize)=@_;
+
+    if ($optimize && $p1=~/^mm[0-7]$/ && $p2=~/^mm[0-7]$/)
+    # movq between mmx registers can sink Intel CPUs
+    {	&::pshufw($p1,$p2,0xe4);		}
+    else
+    {	&::generic("movq",@_);			}
+}
+
+# SSE>2 instructions
+my %regrm = (	"eax"=>0, "ecx"=>1, "edx"=>2, "ebx"=>3,
+		"esp"=>4, "ebp"=>5, "esi"=>6, "edi"=>7	);
+sub ::pextrd
+{ my($dst,$src,$imm)=@_;
+    if ("$dst:$src" =~ /(e[a-dsd][ixp]):xmm([0-7])/)
+    {	&::data_byte(0x66,0x0f,0x3a,0x16,0xc0|($2<<3)|$regrm{$1},$imm);	}
+    else
+    {	&::generic("pextrd",@_);		}
+}
+
+sub ::pinsrd
+{ my($dst,$src,$imm)=@_;
+    if ("$dst:$src" =~ /xmm([0-7]):(e[a-dsd][ixp])/)
+    {	&::data_byte(0x66,0x0f,0x3a,0x22,0xc0|($1<<3)|$regrm{$2},$imm);	}
+    else
+    {	&::generic("pinsrd",@_);		}
+}
+
+sub ::pshufb
+{ my($dst,$src)=@_;
+    if ("$dst:$src" =~ /xmm([0-7]):xmm([0-7])/)
+    {	&data_byte(0x66,0x0f,0x38,0x00,0xc0|($1<<3)|$2);	}
+    else
+    {	&::generic("pshufb",@_);		}
+}
+
+sub ::palignr
+{ my($dst,$src,$imm)=@_;
+    if ("$dst:$src" =~ /xmm([0-7]):xmm([0-7])/)
+    {	&::data_byte(0x66,0x0f,0x3a,0x0f,0xc0|($1<<3)|$2,$imm);	}
+    else
+    {	&::generic("palignr",@_);		}
+}
+
+sub ::pclmulqdq
+{ my($dst,$src,$imm)=@_;
+    if ("$dst:$src" =~ /xmm([0-7]):xmm([0-7])/)
+    {	&::data_byte(0x66,0x0f,0x3a,0x44,0xc0|($1<<3)|$2,$imm);	}
+    else
+    {	&::generic("pclmulqdq",@_);		}
+}
+
+sub ::rdrand
+{ my ($dst)=@_;
+    if ($dst =~ /(e[a-dsd][ixp])/)
+    {	&::data_byte(0x0f,0xc7,0xf0|$regrm{$dst});	}
+    else
+    {	&::generic("rdrand",@_);	}
+}
+
+# label management
+$lbdecor="L";		# local label decoration, set by package
+$label="000";
+
+sub ::islabel		# see is argument is a known label
+{ my $i;
+    foreach $i (values %label) { return $i if ($i eq $_[0]); }
+  $label{$_[0]};	# can be undef
+}
+
+sub ::label		# instantiate a function-scope label
+{   if (!defined($label{$_[0]}))
+    {	$label{$_[0]}="${lbdecor}${label}${_[0]}"; $label++;   }
+  $label{$_[0]};
+}
+
+sub ::LABEL		# instantiate a file-scope label
+{   $label{$_[0]}=$_[1] if (!defined($label{$_[0]}));
+  $label{$_[0]};
+}
+
+sub ::static_label	{ &::LABEL($_[0],$lbdecor.$_[0]); }
+
+sub ::set_label_B	{ push(@out,"@_:\n"); }
+sub ::set_label
+{ my $label=&::label($_[0]);
+    &::align($_[1]) if ($_[1]>1);
+    &::set_label_B($label);
+  $label;
+}
+
+sub ::wipe_labels	# wipes function-scope labels
+{   foreach $i (keys %label)
+    {	delete $label{$i} if ($label{$i} =~ /^\Q${lbdecor}\E[0-9]{3}/);	}
+}
+
+# subroutine management
+sub ::function_begin
+{   &function_begin_B(@_);
+    $stack=4;
+    &push("ebp");
+    &push("ebx");
+    &push("esi");
+    &push("edi");
+}
+
+sub ::function_end
+{   &pop("edi");
+    &pop("esi");
+    &pop("ebx");
+    &pop("ebp");
+    &ret();
+    &function_end_B(@_);
+    $stack=0;
+    &wipe_labels();
+}
+
+sub ::function_end_A
+{   &pop("edi");
+    &pop("esi");
+    &pop("ebx");
+    &pop("ebp");
+    &ret();
+    $stack+=16;	# readjust esp as if we didn't pop anything
+}
+
+sub ::asciz
+{ my @str=unpack("C*",shift);
+    push @str,0;
+    while ($#str>15) {
+	&data_byte(@str[0..15]);
+	foreach (0..15) { shift @str; }
+    }
+    &data_byte(@str) if (@str);
+}
+
+sub ::asm_finish
+{   &file_end();
+    print @out;
+}
+
+sub ::asm_init
+{ my ($type,$fn,$cpu)=@_;
+
+    $filename=$fn;
+    $i386=$cpu;
+
+    $elf=$cpp=$coff=$aout=$macosx=$win32=$netware=$mwerks=$android=0;
+    if    (($type eq "elf"))
+    {	$elf=1;			require "x86gas.pl";	}
+    elsif (($type eq "a\.out"))
+    {	$aout=1;		require "x86gas.pl";	}
+    elsif (($type eq "coff" or $type eq "gaswin"))
+    {	$coff=1;		require "x86gas.pl";	}
+    elsif (($type eq "win32n"))
+    {	$win32=1;		require "x86nasm.pl";	}
+    elsif (($type eq "nw-nasm"))
+    {	$netware=1;		require "x86nasm.pl";	}
+    #elsif (($type eq "nw-mwasm"))
+    #{	$netware=1; $mwerks=1;	require "x86nasm.pl";	}
+    elsif (($type eq "win32"))
+    {	$win32=1;		require "x86masm.pl";	}
+    elsif (($type eq "macosx"))
+    {	$aout=1; $macosx=1;	require "x86gas.pl";	}
+    elsif (($type eq "android"))
+    {	$elf=1; $android=1;	require "x86gas.pl";	}
+    else
+    {	print STDERR <<"EOF";
 Pick one target type from
 	elf	- Linux, FreeBSD, Solaris x86, etc.
-	a.out	- OpenBSD, DJGPP, etc.
+	a.out	- DJGPP, elder OpenBSD, etc.
 	coff	- GAS/COFF such as Win32 targets
-	win32	- Windows 95/Windows NT
 	win32n	- Windows 95/Windows NT NASM format
 	nw-nasm - NetWare NASM format
-	nw-mwasm- NetWare Metrowerks Assembler
-EOF
-		exit(1);
-		}
-
-	$pic=0;
-	for (@ARGV) {	$pic=1 if (/\-[fK]PIC/i);	}
-
-	&asm_init_output();
-
-&comment("Don't even think of reading this code");
-&comment("It was automatically generated by $filename");
-&comment("Which is a perl program used to generate the x86 assember for");
-&comment("any of ELF, a.out, COFF, Win32, ...");
-&comment("eric <eay\@cryptsoft.com>");
-&comment("");
-
-	$filename =~ s/\.pl$//;
-	&file($filename);
-	}
-
-sub asm_finish_cpp
-	{
-	return unless $cpp;
-
-	local($tmp,$i);
-	foreach $i (&get_labels())
-		{
-		$tmp.="#define $i _$i\n";
-		}
-	print <<"EOF";
-/* Run the C pre-processor over this file with one of the following defined
- * ELF - elf object files,
- * OUT - a.out object files,
- * BSDI - BSDI style a.out object files
- * SOL - Solaris style elf
- */
-
-#define TYPE(a,b)       .type   a,b
-#define SIZE(a,b)       .size   a,b
-
-#if defined(OUT) || (defined(BSDI) && !defined(ELF))
-$tmp
-#endif
-
-#ifdef OUT
-#define OK	1
-#define ALIGN	4
-#if defined(__CYGWIN__) || defined(__DJGPP__) || (__MINGW32__)
-#undef SIZE
-#undef TYPE
-#define SIZE(a,b)
-#define TYPE(a,b)	.def a; .scl 2; .type 32; .endef
-#endif /* __CYGWIN || __DJGPP */
-#endif
-
-#if defined(BSDI) && !defined(ELF)
-#define OK              1
-#define ALIGN           4
-#undef SIZE
-#undef TYPE
-#define SIZE(a,b)
-#define TYPE(a,b)
-#endif
-
-#if defined(ELF) || defined(SOL)
-#define OK              1
-#define ALIGN           16
-#endif
-
-#ifndef OK
-You need to define one of
-ELF - elf systems - linux-elf, NetBSD and DG-UX
-OUT - a.out systems - linux-a.out and FreeBSD
-SOL - solaris systems, which are elf with strange comment lines
-BSDI - a.out with a very primative version of as.
-#endif
-
-/* Let the Assembler begin :-) */
+	macosx	- Mac OS X
 EOF
-	}
+	exit(1);
+    }
+
+    $pic=0;
+    for (@ARGV) { $pic=1 if (/\-[fK]PIC/i); }
+
+    $filename =~ s/\.pl$//;
+    &file($filename);
+}
 
 1;
diff --git a/crypto/perlasm/x86gas.pl b/crypto/perlasm/x86gas.pl
new file mode 100755
index 000000000000..682a3a3163e2
--- /dev/null
+++ b/crypto/perlasm/x86gas.pl
@@ -0,0 +1,253 @@
+#!/usr/bin/env perl
+
+package x86gas;
+
+*out=\@::out;
+
+$::lbdecor=$::aout?"L":".L";		# local label decoration
+$nmdecor=($::aout or $::coff)?"_":"";	# external name decoration
+
+$initseg="";
+
+$align=16;
+$align=log($align)/log(2) if ($::aout);
+$com_start="#" if ($::aout or $::coff);
+
+sub opsize()
+{ my $reg=shift;
+    if    ($reg =~ m/^%e/o)		{ "l"; }
+    elsif ($reg =~ m/^%[a-d][hl]$/o)	{ "b"; }
+    elsif ($reg =~ m/^%[xm]/o)		{ undef; }
+    else				{ "w"; }
+}
+
+# swap arguments;
+# expand opcode with size suffix;
+# prefix numeric constants with $;
+sub ::generic
+{ my($opcode,@arg)=@_;
+  my($suffix,$dst,$src);
+
+    @arg=reverse(@arg);
+
+    for (@arg)
+    {	s/^(\*?)(e?[a-dsixphl]{2})$/$1%$2/o;	# gp registers
+	s/^([xy]?mm[0-7])$/%$1/o;		# xmm/mmx registers
+	s/^(\-?[0-9]+)$/\$$1/o;			# constants
+	s/^(\-?0x[0-9a-f]+)$/\$$1/o;		# constants
+    }
+
+    $dst = $arg[$#arg]		if ($#arg>=0);
+    $src = $arg[$#arg-1]	if ($#arg>=1);
+    if    ($dst =~ m/^%/o)	{ $suffix=&opsize($dst); }
+    elsif ($src =~ m/^%/o)	{ $suffix=&opsize($src); }
+    else			{ $suffix="l";           }
+    undef $suffix if ($dst =~ m/^%[xm]/o || $src =~ m/^%[xm]/o);
+
+    if ($#_==0)				{ &::emit($opcode);		}
+    elsif ($#_==1 && $opcode =~ m/^(call|clflush|j|loop|set)/o)
+					{ &::emit($opcode,@arg);	}
+    else				{ &::emit($opcode.$suffix,@arg);}
+
+  1;
+}
+#
+# opcodes not covered by ::generic above, mostly inconsistent namings...
+#
+sub ::movzx	{ &::movzb(@_);			}
+sub ::pushfd	{ &::pushfl;			}
+sub ::popfd	{ &::popfl;			}
+sub ::cpuid	{ &::emit(".byte\t0x0f,0xa2");	}
+sub ::rdtsc	{ &::emit(".byte\t0x0f,0x31");	}
+
+sub ::call	{ &::emit("call",(&::islabel($_[0]) or "$nmdecor$_[0]")); }
+sub ::call_ptr	{ &::generic("call","*$_[0]");	}
+sub ::jmp_ptr	{ &::generic("jmp","*$_[0]");	}
+
+*::bswap = sub	{ &::emit("bswap","%$_[0]");	} if (!$::i386);
+
+sub ::DWP
+{ my($addr,$reg1,$reg2,$idx)=@_;
+  my $ret="";
+
+    $addr =~ s/^\s+//;
+    # prepend global references with optional underscore
+    $addr =~ s/^([^\+\-0-9][^\+\-]*)/&::islabel($1) or "$nmdecor$1"/ige;
+
+    $reg1 = "%$reg1" if ($reg1);
+    $reg2 = "%$reg2" if ($reg2);
+
+    $ret .= $addr if (($addr ne "") && ($addr ne 0));
+
+    if ($reg2)
+    {	$idx!= 0 or $idx=1;
+	$ret .= "($reg1,$reg2,$idx)";
+    }
+    elsif ($reg1)
+    {	$ret .= "($reg1)";	}
+
+  $ret;
+}
+sub ::QWP	{ &::DWP(@_);	}
+sub ::BP	{ &::DWP(@_);	}
+sub ::WP	{ &::DWP(@_);	}
+sub ::BC	{ @_;		}
+sub ::DWC	{ @_;		}
+
+sub ::file
+{   push(@out,".file\t\"$_[0].s\"\n.text\n");	}
+
+sub ::function_begin_B
+{ my $func=shift;
+  my $global=($func !~ /^_/);
+  my $begin="${::lbdecor}_${func}_begin";
+
+    &::LABEL($func,$global?"$begin":"$nmdecor$func");
+    $func=$nmdecor.$func;
+
+    push(@out,".globl\t$func\n")	if ($global);
+    if ($::coff)
+    {	push(@out,".def\t$func;\t.scl\t".(3-$global).";\t.type\t32;\t.endef\n"); }
+    elsif (($::aout and !$::pic) or $::macosx)
+    { }
+    else
+    {	push(@out,".type	$func,\@function\n"); }
+    push(@out,".align\t$align\n");
+    push(@out,"$func:\n");
+    push(@out,"$begin:\n")		if ($global);
+    $::stack=4;
+}
+
+sub ::function_end_B
+{ my $func=shift;
+    push(@out,".size\t$nmdecor$func,.-".&::LABEL($func)."\n") if ($::elf);
+    $::stack=0;
+    &::wipe_labels();
+}
+
+sub ::comment
+	{
+	if (!defined($com_start) or $::elf)
+		{	# Regarding $::elf above...
+			# GNU and SVR4 as'es use different comment delimiters,
+		push(@out,"\n");	# so we just skip ELF comments...
+		return;
+		}
+	foreach (@_)
+		{
+		if (/^\s*$/)
+			{ push(@out,"\n"); }
+		else
+			{ push(@out,"\t$com_start $_ $com_end\n"); }
+		}
+	}
+
+sub ::external_label
+{   foreach(@_) { &::LABEL($_,$nmdecor.$_); }   }
+
+sub ::public_label
+{   push(@out,".globl\t".&::LABEL($_[0],$nmdecor.$_[0])."\n");   }
+
+sub ::file_end
+{   if ($::macosx)
+    {	if (%non_lazy_ptr)
+    	{   push(@out,".section __IMPORT,__pointers,non_lazy_symbol_pointers\n");
+	    foreach $i (keys %non_lazy_ptr)
+	    {	push(@out,"$non_lazy_ptr{$i}:\n.indirect_symbol\t$i\n.long\t0\n");   }
+	}
+    }
+    if (grep {/\b${nmdecor}OPENSSL_ia32cap_P\b/i} @out) {
+	my $tmp=".comm\t${nmdecor}OPENSSL_ia32cap_P,8";
+	if ($::macosx)	{ push (@out,"$tmp,2\n"); }
+	elsif ($::elf)	{ push (@out,"$tmp,4\n"); }
+	else		{ push (@out,"$tmp\n"); }
+    }
+    push(@out,$initseg) if ($initseg);
+}
+
+sub ::data_byte	{   push(@out,".byte\t".join(',',@_)."\n");   }
+sub ::data_short{   push(@out,".value\t".join(',',@_)."\n");  }
+sub ::data_word {   push(@out,".long\t".join(',',@_)."\n");   }
+
+sub ::align
+{ my $val=$_[0],$p2,$i;
+    if ($::aout)
+    {	for ($p2=0;$val!=0;$val>>=1) { $p2++; }
+	$val=$p2-1;
+	$val.=",0x90";
+    }
+    push(@out,".align\t$val\n");
+}
+
+sub ::picmeup
+{ my($dst,$sym,$base,$reflabel)=@_;
+
+    if (($::pic && ($::elf || $::aout)) || $::macosx)
+    {	if (!defined($base))
+	{   &::call(&::label("PIC_me_up"));
+	    &::set_label("PIC_me_up");
+	    &::blindpop($dst);
+	    $base=$dst;
+	    $reflabel=&::label("PIC_me_up");
+	}
+	if ($::macosx)
+	{   my $indirect=&::static_label("$nmdecor$sym\$non_lazy_ptr");
+	    &::mov($dst,&::DWP("$indirect-$reflabel",$base));
+	    $non_lazy_ptr{"$nmdecor$sym"}=$indirect;
+	}
+	else
+	{   &::lea($dst,&::DWP("_GLOBAL_OFFSET_TABLE_+[.-$reflabel]",
+			    $base));
+	    &::mov($dst,&::DWP("$sym\@GOT",$dst));
+	}
+    }
+    else
+    {	&::lea($dst,&::DWP($sym));	}
+}
+
+sub ::initseg
+{ my $f=$nmdecor.shift;
+
+    if ($::android)
+    {	$initseg.=<<___;
+.section	.init_array
+.align	4
+.long	$f
+___
+    }
+    elsif ($::elf)
+    {	$initseg.=<<___;
+.section	.init
+	call	$f
+___
+    }
+    elsif ($::coff)
+    {   $initseg.=<<___;	# applies to both Cygwin and Mingw
+.section	.ctors
+.long	$f
+___
+    }
+    elsif ($::macosx)
+    {	$initseg.=<<___;
+.mod_init_func
+.align 2
+.long   $f
+___
+    }
+    elsif ($::aout)
+    {	my $ctor="${nmdecor}_GLOBAL_\$I\$$f";
+	$initseg.=".text\n";
+	$initseg.=".type	$ctor,\@function\n" if ($::pic);
+	$initseg.=<<___;	# OpenBSD way...
+.globl	$ctor
+.align	2
+$ctor:
+	jmp	$f
+___
+    }
+}
+
+sub ::dataseg
+{   push(@out,".data\n");   }
+
+1;
diff --git a/crypto/perlasm/x86masm.pl b/crypto/perlasm/x86masm.pl
new file mode 100755
index 000000000000..96b1b73e1a3a
--- /dev/null
+++ b/crypto/perlasm/x86masm.pl
@@ -0,0 +1,197 @@
+#!/usr/bin/env perl
+
+package x86masm;
+
+*out=\@::out;
+
+$::lbdecor="\$L";	# local label decoration
+$nmdecor="_";		# external name decoration
+
+$initseg="";
+$segment="";
+
+sub ::generic
+{ my ($opcode,@arg)=@_;
+
+    # fix hexadecimal constants
+    for (@arg) { s/(?<![\w\$\.])0x([0-9a-f]+)/0$1h/oi; }
+
+    if ($opcode =~ /lea/ && @arg[1] =~ s/.*PTR\s+(\(.*\))$/OFFSET $1/)	# no []
+    {	$opcode="mov";	}
+    elsif ($opcode !~ /movq/)
+    {	# fix xmm references
+	$arg[0] =~ s/\b[A-Z]+WORD\s+PTR/XMMWORD PTR/i if ($arg[1]=~/\bxmm[0-7]\b/i);
+	$arg[1] =~ s/\b[A-Z]+WORD\s+PTR/XMMWORD PTR/i if ($arg[0]=~/\bxmm[0-7]\b/i);
+    }
+
+    &::emit($opcode,@arg);
+  1;
+}
+#
+# opcodes not covered by ::generic above, mostly inconsistent namings...
+#
+sub ::call	{ &::emit("call",(&::islabel($_[0]) or "$nmdecor$_[0]")); }
+sub ::call_ptr	{ &::emit("call",@_);	}
+sub ::jmp_ptr	{ &::emit("jmp",@_);	}
+
+sub get_mem
+{ my($size,$addr,$reg1,$reg2,$idx)=@_;
+  my($post,$ret);
+
+    $ret .= "$size PTR " if ($size ne "");
+
+    $addr =~ s/^\s+//;
+    # prepend global references with optional underscore
+    $addr =~ s/^([^\+\-0-9][^\+\-]*)/&::islabel($1) or "$nmdecor$1"/ige;
+    # put address arithmetic expression in parenthesis
+    $addr="($addr)" if ($addr =~ /^.+[\-\+].+$/);
+
+    if (($addr ne "") && ($addr ne 0))
+    {	if ($addr !~ /^-/)	{ $ret .= "$addr";  }
+	else			{ $post=$addr;      }
+    }
+    $ret .= "[";
+
+    if ($reg2 ne "")
+    {	$idx!=0 or $idx=1;
+	$ret .= "$reg2*$idx";
+	$ret .= "+$reg1" if ($reg1 ne "");
+    }
+    else
+    {	$ret .= "$reg1";   }
+
+    $ret .= "$post]";
+    $ret =~ s/\+\]/]/; # in case $addr was the only argument
+    $ret =~ s/\[\s*\]//;
+
+  $ret;
+}
+sub ::BP	{ &get_mem("BYTE",@_);  }
+sub ::WP	{ &get_mem("WORD",@_);	}
+sub ::DWP	{ &get_mem("DWORD",@_); }
+sub ::QWP	{ &get_mem("QWORD",@_); }
+sub ::BC	{ "@_";  }
+sub ::DWC	{ "@_"; }
+
+sub ::file
+{ my $tmp=<<___;
+TITLE	$_[0].asm
+IF \@Version LT 800
+ECHO MASM version 8.00 or later is strongly recommended.
+ENDIF
+.486
+.MODEL	FLAT
+OPTION	DOTNAME
+IF \@Version LT 800
+.text\$	SEGMENT PAGE 'CODE'
+ELSE
+.text\$	SEGMENT ALIGN(64) 'CODE'
+ENDIF
+___
+    push(@out,$tmp);
+    $segment = ".text\$";
+}
+
+sub ::function_begin_B
+{ my $func=shift;
+  my $global=($func !~ /^_/);
+  my $begin="${::lbdecor}_${func}_begin";
+
+    &::LABEL($func,$global?"$begin":"$nmdecor$func");
+    $func="ALIGN\t16\n".$nmdecor.$func."\tPROC";
+
+    if ($global)    { $func.=" PUBLIC\n${begin}::\n"; }
+    else	    { $func.=" PRIVATE\n";            }
+    push(@out,$func);
+    $::stack=4;
+}
+sub ::function_end_B
+{ my $func=shift;
+
+    push(@out,"$nmdecor$func ENDP\n");
+    $::stack=0;
+    &::wipe_labels();
+}
+
+sub ::file_end
+{ my $xmmheader=<<___;
+.686
+.XMM
+IF \@Version LT 800
+XMMWORD STRUCT 16
+DQ	2 dup (?)
+XMMWORD	ENDS
+ENDIF
+___
+    if (grep {/\b[x]?mm[0-7]\b/i} @out) {
+	grep {s/\.[3-7]86/$xmmheader/} @out;
+    }
+
+    push(@out,"$segment	ENDS\n");
+
+    if (grep {/\b${nmdecor}OPENSSL_ia32cap_P\b/i} @out)
+    {	my $comm=<<___;
+.bss	SEGMENT 'BSS'
+COMM	${nmdecor}OPENSSL_ia32cap_P:QWORD
+.bss	ENDS
+___
+	# comment out OPENSSL_ia32cap_P declarations
+	grep {s/(^EXTERN\s+${nmdecor}OPENSSL_ia32cap_P)/\;$1/} @out;
+	push (@out,$comm);
+    }
+    push (@out,$initseg) if ($initseg);
+    push (@out,"END\n");
+}
+
+sub ::comment {   foreach (@_) { push(@out,"\t; $_\n"); }   }
+
+*::set_label_B = sub
+{ my $l=shift; push(@out,$l.($l=~/^\Q${::lbdecor}\E[0-9]{3}/?":\n":"::\n")); };
+
+sub ::external_label
+{   foreach(@_)
+    {	push(@out, "EXTERN\t".&::LABEL($_,$nmdecor.$_).":NEAR\n");   }
+}
+
+sub ::public_label
+{   push(@out,"PUBLIC\t".&::LABEL($_[0],$nmdecor.$_[0])."\n");   }
+
+sub ::data_byte
+{   push(@out,("DB\t").join(',',@_)."\n");	}
+
+sub ::data_short
+{   push(@out,("DW\t").join(',',@_)."\n");	}
+
+sub ::data_word
+{   push(@out,("DD\t").join(',',@_)."\n");	}
+
+sub ::align
+{   push(@out,"ALIGN\t$_[0]\n");	}
+
+sub ::picmeup
+{ my($dst,$sym)=@_;
+    &::lea($dst,&::DWP($sym));
+}
+
+sub ::initseg
+{ my $f=$nmdecor.shift;
+
+    $initseg.=<<___;
+.CRT\$XCU	SEGMENT DWORD PUBLIC 'DATA'
+EXTERN	$f:NEAR
+DD	$f
+.CRT\$XCU	ENDS
+___
+}
+
+sub ::dataseg
+{   push(@out,"$segment\tENDS\n_DATA\tSEGMENT\n"); $segment="_DATA";   }
+
+sub ::safeseh
+{ my $nm=shift;
+    push(@out,"IF \@Version GE 710\n");
+    push(@out,".SAFESEH	".&::LABEL($nm,$nmdecor.$nm)."\n");
+    push(@out,"ENDIF\n");
+}
+
+1;
diff --git a/crypto/perlasm/x86ms.pl b/crypto/perlasm/x86ms.pl
deleted file mode 100644
index a0be2934c207..000000000000
--- a/crypto/perlasm/x86ms.pl
+++ /dev/null
@@ -1,472 +0,0 @@
-#!/usr/local/bin/perl
-
-package x86ms;
-
-$label="L000";
-
-%lb=(	'eax',	'al',
-	'ebx',	'bl',
-	'ecx',	'cl',
-	'edx',	'dl',
-	'ax',	'al',
-	'bx',	'bl',
-	'cx',	'cl',
-	'dx',	'dl',
-	);
-
-%hb=(	'eax',	'ah',
-	'ebx',	'bh',
-	'ecx',	'ch',
-	'edx',	'dh',
-	'ax',	'ah',
-	'bx',	'bh',
-	'cx',	'ch',
-	'dx',	'dh',
-	);
-
-sub main'asm_init_output { @out=(); }
-sub main'asm_get_output { return(@out); }
-sub main'get_labels { return(@labels); }
-sub main'external_label
-{
-	push(@labels,@_);
-	foreach (@_) {
-		push(@out, "EXTRN\t_$_:DWORD\n");
-	}
-}
-
-sub main'LB
-	{
-	(defined($lb{$_[0]})) || die "$_[0] does not have a 'low byte'\n";
-	return($lb{$_[0]});
-	}
-
-sub main'HB
-	{
-	(defined($hb{$_[0]})) || die "$_[0] does not have a 'high byte'\n";
-	return($hb{$_[0]});
-	}
-
-sub main'BP
-	{
-	&get_mem("BYTE",@_);
-	}
-
-sub main'DWP
-	{
-	&get_mem("DWORD",@_);
-	}
-
-sub main'QWP
-	{
-	&get_mem("QWORD",@_);
-	}
-
-sub main'BC
-	{
-	return @_;
-	}
-
-sub main'DWC
-	{
-	return @_;
-	}
-
-sub main'stack_push
-	{
-	local($num)=@_;
-	$stack+=$num*4;
-	&main'sub("esp",$num*4);
-	}
-
-sub main'stack_pop
-	{
-	local($num)=@_;
-	$stack-=$num*4;
-	&main'add("esp",$num*4);
-	}
-
-sub get_mem
-	{
-	local($size,$addr,$reg1,$reg2,$idx)=@_;
-	local($t,$post);
-	local($ret)="$size PTR ";
-
-	$addr =~ s/^\s+//;
-	if ($addr =~ /^(.+)\+(.+)$/)
-		{
-		$reg2=&conv($1);
-		$addr="_$2";
-		}
-	elsif ($addr =~ /^[_a-z][_a-z0-9]*$/i)
-		{
-		$addr="_$addr";
-		}
-
-	if ($addr =~ /^.+\-.+$/) { $addr="($addr)"; }
-
-	$reg1="$regs{$reg1}" if defined($regs{$reg1});
-	$reg2="$regs{$reg2}" if defined($regs{$reg2});
-	if (($addr ne "") && ($addr ne 0))
-		{
-		if ($addr !~ /^-/)
-			{ $ret.=$addr; }
-		else	{ $post=$addr; }
-		}
-	if ($reg2 ne "")
-		{
-		$t="";
-		$t="*$idx" if ($idx != 0);
-		$reg1="+".$reg1 if ("$reg1$post" ne "");
-		$ret.="[$reg2$t$reg1$post]";
-		}
-	else
-		{
-		$ret.="[$reg1$post]"
-		}
-	$ret =~ s/\[\]//;	# in case $addr was the only argument
-	return($ret);
-	}
-
-sub main'mov	{ &out2("mov",@_); }
-sub main'movb	{ &out2("mov",@_); }
-sub main'and	{ &out2("and",@_); }
-sub main'or	{ &out2("or",@_); }
-sub main'shl	{ &out2("shl",@_); }
-sub main'shr	{ &out2("shr",@_); }
-sub main'xor	{ &out2("xor",@_); }
-sub main'xorb	{ &out2("xor",@_); }
-sub main'add	{ &out2("add",@_); }
-sub main'adc	{ &out2("adc",@_); }
-sub main'sub	{ &out2("sub",@_); }
-sub main'sbb	{ &out2("sbb",@_); }
-sub main'rotl	{ &out2("rol",@_); }
-sub main'rotr	{ &out2("ror",@_); }
-sub main'exch	{ &out2("xchg",@_); }
-sub main'cmp	{ &out2("cmp",@_); }
-sub main'lea	{ &out2("lea",@_); }
-sub main'mul	{ &out1("mul",@_); }
-sub main'imul	{ &out2("imul",@_); }
-sub main'div	{ &out1("div",@_); }
-sub main'dec	{ &out1("dec",@_); }
-sub main'inc	{ &out1("inc",@_); }
-sub main'jmp	{ &out1("jmp",@_); }
-sub main'jmp_ptr { &out1p("jmp",@_); }
-sub main'je	{ &out1("je",@_); }
-sub main'jle	{ &out1("jle",@_); }
-sub main'jz	{ &out1("jz",@_); }
-sub main'jge	{ &out1("jge",@_); }
-sub main'jl	{ &out1("jl",@_); }
-sub main'ja	{ &out1("ja",@_); }
-sub main'jae	{ &out1("jae",@_); }
-sub main'jb	{ &out1("jb",@_); }
-sub main'jbe	{ &out1("jbe",@_); }
-sub main'jc	{ &out1("jc",@_); }
-sub main'jnc	{ &out1("jnc",@_); }
-sub main'jnz	{ &out1("jnz",@_); }
-sub main'jne	{ &out1("jne",@_); }
-sub main'jno	{ &out1("jno",@_); }
-sub main'push	{ &out1("push",@_); $stack+=4; }
-sub main'pop	{ &out1("pop",@_); $stack-=4; }
-sub main'pushf	{ &out0("pushfd"); $stack+=4; }
-sub main'popf	{ &out0("popfd"); $stack-=4; }
-sub main'bswap	{ &out1("bswap",@_); &using486(); }
-sub main'not	{ &out1("not",@_); }
-sub main'call	{ &out1("call",($_[0]=~/^\$L/?'':'_').$_[0]); }
-sub main'call_ptr { &out1p("call",@_); }
-sub main'ret	{ &out0("ret"); }
-sub main'nop	{ &out0("nop"); }
-sub main'test	{ &out2("test",@_); }
-sub main'bt	{ &out2("bt",@_); }
-sub main'leave	{ &out0("leave"); }
-sub main'cpuid  { &out0("DW\t0A20Fh"); }
-sub main'rdtsc  { &out0("DW\t0310Fh"); }
-sub main'halt	{ &out0("hlt"); }
-sub main'movz	{ &out2("movzx",@_); }
-sub main'neg	{ &out1("neg",@_); }
-sub main'cld	{ &out0("cld"); }
-
-# SSE2
-sub main'emms	{ &out0("emms"); }
-sub main'movd	{ &out2("movd",@_); }
-sub main'movq	{ &out2("movq",@_); }
-sub main'movdqu	{ &out2("movdqu",@_); }
-sub main'movdqa	{ &out2("movdqa",@_); }
-sub main'movdq2q{ &out2("movdq2q",@_); }
-sub main'movq2dq{ &out2("movq2dq",@_); }
-sub main'paddq	{ &out2("paddq",@_); }
-sub main'pmuludq{ &out2("pmuludq",@_); }
-sub main'psrlq	{ &out2("psrlq",@_); }
-sub main'psllq	{ &out2("psllq",@_); }
-sub main'pxor	{ &out2("pxor",@_); }
-sub main'por	{ &out2("por",@_); }
-sub main'pand	{ &out2("pand",@_); }
-
-sub out2
-	{
-	local($name,$p1,$p2)=@_;
-	local($l,$t,$line);
-
-	$line="\t$name\t";
-	$t=&conv($p1).",";
-	$l=length($t);
-	$line.="$t";
-	$l=4-($l+9)/8;
-	$line.="\t" x $l;
-	$line.=&conv($p2);
-	if ($line=~/\bxmm[0-7]\b/i) { $line=~s/\b[A-Z]+WORD\s+PTR/XMMWORD PTR/i; }
-	push(@out,$line."\n");
-	}
-
-sub out0
-	{
-	local($name)=@_;
-
-	push(@out,"\t$name\n");
-	}
-
-sub out1
-	{
-	local($name,$p1)=@_;
-	local($l,$t);
-
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
-
-sub conv
-	{
-	local($p)=@_;
-
-	$p =~ s/0x([0-9A-Fa-f]+)/0$1h/;
-	return $p;
-	}
-
-sub using486
-	{
-	return if $using486;
-	$using486++;
-	grep(s/\.386/\.486/,@out);
-	}
-
-sub main'file
-	{
-	local($file)=@_;
-
-	local($tmp)=<<"EOF";
-	TITLE	$file.asm
-        .386
-.model	FLAT
-_TEXT\$	SEGMENT PAGE 'CODE'
-
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'function_begin
-	{
-	local($func,$extra)=@_;
-
-	push(@labels,$func);
-
-	local($tmp)=<<"EOF";
-PUBLIC	_$func
-$extra
-_$func PROC NEAR
-	push	ebp
-	push	ebx
-	push	esi
-	push	edi
-EOF
-	push(@out,$tmp);
-	$stack=20;
-	}
-
-sub main'function_begin_B
-	{
-	local($func,$extra)=@_;
-
-	local($tmp)=<<"EOF";
-PUBLIC	_$func
-$extra
-_$func PROC NEAR
-EOF
-	push(@out,$tmp);
-	$stack=4;
-	}
-
-sub main'function_end
-	{
-	local($func)=@_;
-
-	local($tmp)=<<"EOF";
-	pop	edi
-	pop	esi
-	pop	ebx
-	pop	ebp
-	ret
-_$func ENDP
-EOF
-	push(@out,$tmp);
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_B
-	{
-	local($func)=@_;
-
-	local($tmp)=<<"EOF";
-_$func ENDP
-EOF
-	push(@out,$tmp);
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_A
-	{
-	local($func)=@_;
-
-	local($tmp)=<<"EOF";
-	pop	edi
-	pop	esi
-	pop	ebx
-	pop	ebp
-	ret
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'file_end
-	{
-	# try to detect if SSE2 or MMX extensions were used...
-	my $xmmheader=<<___;
-.686
-.XMM
-IF \@Version LT 800
-XMMWORD STRUCT 16
-	DQ  2 dup (?)
-XMMWORD ENDS
-ENDIF
-___
-	if (grep {/\b[x]?mm[0-7]\b/i} @out) {
-		grep {s/\.[3-7]86/$xmmheader/} @out;
-		}
-	push(@out,"_TEXT\$	ENDS\n");
-	push(@out,"END\n");
-	}
-
-sub main'wparam
-	{
-	local($num)=@_;
-
-	return(&main'DWP($stack+$num*4,"esp","",0));
-	}
-
-sub main'swtmp
-	{
-	return(&main'DWP($_[0]*4,"esp","",0));
-	}
-
-# Should use swtmp, which is above esp.  Linix can trash the stack above esp
-#sub main'wtmp
-#	{
-#	local($num)=@_;
-#
-#	return(&main'DWP(-(($num+1)*4),"esp","",0));
-#	}
-
-sub main'comment
-	{
-	foreach (@_)
-		{
-		push(@out,"\t; $_\n");
-		}
-	}
-
-sub main'public_label
-	{
-	$label{$_[0]}="_$_[0]"	if (!defined($label{$_[0]}));
-	push(@out,"PUBLIC\t$label{$_[0]}\n");
-	}
-
-sub main'label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="\$${label}${_[0]}";
-		$label++;
-		}
-	return($label{$_[0]});
-	}
-
-sub main'set_label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="\$${label}${_[0]}";
-		$label++;
-		}
-	if ($_[1]!=0 && $_[1]>1)
-		{
-		main'align($_[1]);
-		}
-	if((defined $_[2]) && ($_[2] == 1))
-		{
-		push(@out,"$label{$_[0]}::\n");
-		}
-	elsif ($label{$_[0]} !~ /^\$/)
-		{
-		push(@out,"$label{$_[0]}\tLABEL PTR\n");
-		}
-	else
-		{
-		push(@out,"$label{$_[0]}:\n");
-		}
-	}
-
-sub main'data_byte
-	{
-	push(@out,"\tDB\t".join(',',@_)."\n");
-	}
-
-sub main'data_word
-	{
-	push(@out,"\tDD\t".join(',',@_)."\n");
-	}
-
-sub main'align
-	{
-	push(@out,"\tALIGN\t$_[0]\n");
-	}
-
-sub out1p
-	{
-	local($name,$p1)=@_;
-	local($l,$t);
-
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
-
-sub main'picmeup
-	{
-	local($dst,$sym)=@_;
-	&main'lea($dst,&main'DWP($sym));
-	}
-
-sub main'blindpop { &out1("pop",@_); }
-
-sub main'initseg 
-	{
-	local($f)=@_;
-	local($tmp)=<<___;
-OPTION	DOTNAME
-.CRT\$XCU	SEGMENT DWORD PUBLIC 'DATA'
-EXTRN	_$f:NEAR
-DD	_$f
-.CRT\$XCU	ENDS
-___
-	push(@out,$tmp);
-	}
-
-1;
diff --git a/crypto/perlasm/x86nasm.pl b/crypto/perlasm/x86nasm.pl
index fa38f89c09fc..ca2511c9eb91 100644
--- a/crypto/perlasm/x86nasm.pl
+++ b/crypto/perlasm/x86nasm.pl
@@ -1,455 +1,177 @@
-#!/usr/local/bin/perl
+#!/usr/bin/env perl
 
 package x86nasm;
 
-$label="L000";
-$under=($main'netware)?'':'_';
+*out=\@::out;
 
-%lb=(	'eax',	'al',
-	'ebx',	'bl',
-	'ecx',	'cl',
-	'edx',	'dl',
-	'ax',	'al',
-	'bx',	'bl',
-	'cx',	'cl',
-	'dx',	'dl',
-	);
+$::lbdecor="L\$";		# local label decoration
+$nmdecor=$::netware?"":"_";	# external name decoration
+$drdecor=$::mwerks?".":"";	# directive decoration
 
-%hb=(	'eax',	'ah',
-	'ebx',	'bh',
-	'ecx',	'ch',
-	'edx',	'dh',
-	'ax',	'ah',
-	'bx',	'bh',
-	'cx',	'ch',
-	'dx',	'dh',
-	);
+$initseg="";
 
-sub main'asm_init_output { @out=(); }
-sub main'asm_get_output { return(@out); }
-sub main'get_labels { return(@labels); }
+sub ::generic
+{ my $opcode=shift;
+  my $tmp;
 
-sub main'external_label
-{
-	push(@labels,@_);
-	foreach (@_) {
-		push(@out,".") if ($main'mwerks);
-		push(@out, "extern\t${under}$_\n");
-	}
+    if (!$::mwerks)
+    {   if    ($opcode =~ m/^j/o && $#_==0) # optimize jumps
+	{   $_[0] = "NEAR $_[0]";   	}
+	elsif ($opcode eq "lea" && $#_==1)  # wipe storage qualifier from lea
+	{   $_[1] =~ s/^[^\[]*\[/\[/o;	}
+	elsif ($opcode eq "clflush" && $#_==0)
+	{   $_[0] =~ s/^[^\[]*\[/\[/o;	}
+    }
+    &::emit($opcode,@_);
+  1;
 }
-
-sub main'LB
-	{
-	(defined($lb{$_[0]})) || die "$_[0] does not have a 'low byte'\n";
-	return($lb{$_[0]});
-	}
-
-sub main'HB
-	{
-	(defined($hb{$_[0]})) || die "$_[0] does not have a 'high byte'\n";
-	return($hb{$_[0]});
-	}
-
-sub main'BP
-	{
-	&get_mem("BYTE",@_);
-	}
-
-sub main'DWP
-	{
-	&get_mem("DWORD",@_);
-	}
-
-sub main'QWP
-	{
-	&get_mem("",@_);
-	}
-
-sub main'BC
-	{
-	return (($main'mwerks)?"":"BYTE ")."@_";
-	}
-
-sub main'DWC
-	{
-	return (($main'mwerks)?"":"DWORD ")."@_";
-	}
-
-sub main'stack_push
-	{
-	my($num)=@_;
-	$stack+=$num*4;
-	&main'sub("esp",$num*4);
-	}
-
-sub main'stack_pop
-	{
-	my($num)=@_;
-	$stack-=$num*4;
-	&main'add("esp",$num*4);
-	}
+#
+# opcodes not covered by ::generic above, mostly inconsistent namings...
+#
+sub ::call	{ &::emit("call",(&::islabel($_[0]) or "$nmdecor$_[0]")); }
+sub ::call_ptr	{ &::emit("call",@_);	}
+sub ::jmp_ptr	{ &::emit("jmp",@_);	}
 
 sub get_mem
-	{
-	my($size,$addr,$reg1,$reg2,$idx)=@_;
-	my($t,$post);
-	my($ret)=$size;
-	if ($ret ne "")
-		{
-		$ret .= " PTR" if ($main'mwerks);
-		$ret .= " ";
-		}
-	$ret .= "[";
-	$addr =~ s/^\s+//;
-	if ($addr =~ /^(.+)\+(.+)$/)
-		{
-		$reg2=&conv($1);
-		$addr="$under$2";
-		}
-	elsif ($addr =~ /^[_a-z][_a-z0-9]*$/i)
-		{
-		$addr="$under$addr";
-		}
-
-	if ($addr =~ /^.+\-.+$/) { $addr="($addr)"; }
-
-	$reg1="$regs{$reg1}" if defined($regs{$reg1});
-	$reg2="$regs{$reg2}" if defined($regs{$reg2});
-	if (($addr ne "") && ($addr ne 0))
-		{
-		if ($addr !~ /^-/)
-			{ $ret.="${addr}+"; }
-		else	{ $post=$addr; }
-		}
-	if ($reg2 ne "")
-		{
-		$t="";
-		$t="*$idx" if ($idx != 0);
-		$reg1="+".$reg1 if ("$reg1$post" ne "");
-		$ret.="$reg2$t$reg1$post]";
-		}
-	else
-		{
-		$ret.="$reg1$post]"
-		}
-	$ret =~ s/\+\]/]/; # in case $addr was the only argument
-	return($ret);
-	}
-
-sub main'mov	{ &out2("mov",@_); }
-sub main'movb	{ &out2("mov",@_); }
-sub main'and	{ &out2("and",@_); }
-sub main'or	{ &out2("or",@_); }
-sub main'shl	{ &out2("shl",@_); }
-sub main'shr	{ &out2("shr",@_); }
-sub main'xor	{ &out2("xor",@_); }
-sub main'xorb	{ &out2("xor",@_); }
-sub main'add	{ &out2("add",@_); }
-sub main'adc	{ &out2("adc",@_); }
-sub main'sub	{ &out2("sub",@_); }
-sub main'sbb	{ &out2("sbb",@_); }
-sub main'rotl	{ &out2("rol",@_); }
-sub main'rotr	{ &out2("ror",@_); }
-sub main'exch	{ &out2("xchg",@_); }
-sub main'cmp	{ &out2("cmp",@_); }
-sub main'lea	{ &out2("lea",@_); }
-sub main'mul	{ &out1("mul",@_); }
-sub main'imul	{ &out2("imul",@_); }
-sub main'div	{ &out1("div",@_); }
-sub main'dec	{ &out1("dec",@_); }
-sub main'inc	{ &out1("inc",@_); }
-sub main'jmp	{ &out1("jmp",@_); }
-sub main'jmp_ptr { &out1p("jmp",@_); }
-
-# This is a bit of a kludge: declare all branches as NEAR.
-$near=($main'mwerks)?'':'NEAR';
-sub main'je	{ &out1("je $near",@_); }
-sub main'jle	{ &out1("jle $near",@_); }
-sub main'jz	{ &out1("jz $near",@_); }
-sub main'jge	{ &out1("jge $near",@_); }
-sub main'jl	{ &out1("jl $near",@_); }
-sub main'ja	{ &out1("ja $near",@_); }
-sub main'jae	{ &out1("jae $near",@_); }
-sub main'jb	{ &out1("jb $near",@_); }
-sub main'jbe	{ &out1("jbe $near",@_); }
-sub main'jc	{ &out1("jc $near",@_); }
-sub main'jnc	{ &out1("jnc $near",@_); }
-sub main'jnz	{ &out1("jnz $near",@_); }
-sub main'jne	{ &out1("jne $near",@_); }
-sub main'jno	{ &out1("jno $near",@_); }
-
-sub main'push	{ &out1("push",@_); $stack+=4; }
-sub main'pop	{ &out1("pop",@_); $stack-=4; }
-sub main'pushf	{ &out0("pushfd"); $stack+=4; }
-sub main'popf	{ &out0("popfd"); $stack-=4; }
-sub main'bswap	{ &out1("bswap",@_); &using486(); }
-sub main'not	{ &out1("not",@_); }
-sub main'call	{ &out1("call",($_[0]=~/^\@L/?'':$under).$_[0]); }
-sub main'call_ptr { &out1p("call",@_); }
-sub main'ret	{ &out0("ret"); }
-sub main'nop	{ &out0("nop"); }
-sub main'test	{ &out2("test",@_); }
-sub main'bt	{ &out2("bt",@_); }
-sub main'leave	{ &out0("leave"); }
-sub main'cpuid	{ &out0("cpuid"); }
-sub main'rdtsc	{ &out0("rdtsc"); }
-sub main'halt	{ &out0("hlt"); }
-sub main'movz	{ &out2("movzx",@_); }
-sub main'neg	{ &out1("neg",@_); }
-sub main'cld	{ &out0("cld"); }
-
-# SSE2
-sub main'emms	{ &out0("emms"); }
-sub main'movd	{ &out2("movd",@_); }
-sub main'movq	{ &out2("movq",@_); }
-sub main'movdqu	{ &out2("movdqu",@_); }
-sub main'movdqa	{ &out2("movdqa",@_); }
-sub main'movdq2q{ &out2("movdq2q",@_); }
-sub main'movq2dq{ &out2("movq2dq",@_); }
-sub main'paddq	{ &out2("paddq",@_); }
-sub main'pmuludq{ &out2("pmuludq",@_); }
-sub main'psrlq	{ &out2("psrlq",@_); }
-sub main'psllq	{ &out2("psllq",@_); }
-sub main'pxor	{ &out2("pxor",@_); }
-sub main'por	{ &out2("por",@_); }
-sub main'pand	{ &out2("pand",@_); }
-
-sub out2
-	{
-	my($name,$p1,$p2)=@_;
-	my($l,$t);
-
-	push(@out,"\t$name\t");
-	if (!$main'mwerks and $name eq "lea")
-		{
-		$p1 =~ s/^[^\[]*\[/\[/;
-		$p2 =~ s/^[^\[]*\[/\[/;
-		}
-	$t=&conv($p1).",";
-	$l=length($t);
-	push(@out,$t);
-	$l=4-($l+9)/8;
-	push(@out,"\t" x $l);
-	push(@out,&conv($p2));
-	push(@out,"\n");
-	}
-
-sub out0
-	{
-	my($name)=@_;
-
-	push(@out,"\t$name\n");
-	}
-
-sub out1
-	{
-	my($name,$p1)=@_;
-	my($l,$t);
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
-
-sub conv
-	{
-	my($p)=@_;
-	$p =~ s/0x([0-9A-Fa-f]+)/0$1h/;
-	return $p;
-	}
-
-sub using486
-	{
-	return if $using486;
-	$using486++;
-	grep(s/\.386/\.486/,@out);
-	}
-
-sub main'file
-	{
-	if ($main'mwerks)	{ push(@out,".section\t.text\n"); }
-	else	{
-		local $tmp=<<___;
-%ifdef __omf__
-section	code	use32 class=code
+{ my($size,$addr,$reg1,$reg2,$idx)=@_;
+  my($post,$ret);
+
+    if ($size ne "")
+    {	$ret .= "$size";
+	$ret .= " PTR" if ($::mwerks);
+	$ret .= " ";
+    }
+    $ret .= "[";
+
+    $addr =~ s/^\s+//;
+    # prepend global references with optional underscore
+    $addr =~ s/^([^\+\-0-9][^\+\-]*)/::islabel($1) or "$nmdecor$1"/ige;
+    # put address arithmetic expression in parenthesis
+    $addr="($addr)" if ($addr =~ /^.+[\-\+].+$/);
+
+    if (($addr ne "") && ($addr ne 0))
+    {	if ($addr !~ /^-/)	{ $ret .= "$addr+"; }
+	else			{ $post=$addr;      }
+    }
+
+    if ($reg2 ne "")
+    {	$idx!=0 or $idx=1;
+	$ret .= "$reg2*$idx";
+	$ret .= "+$reg1" if ($reg1 ne "");
+    }
+    else
+    {	$ret .= "$reg1";   }
+
+    $ret .= "$post]";
+    $ret =~ s/\+\]/]/; # in case $addr was the only argument
+
+  $ret;
+}
+sub ::BP	{ &get_mem("BYTE",@_);  }
+sub ::DWP	{ &get_mem("DWORD",@_); }
+sub ::WP	{ &get_mem("WORD",@_);	}
+sub ::QWP	{ &get_mem("",@_);      }
+sub ::BC	{ (($::mwerks)?"":"BYTE ")."@_";  }
+sub ::DWC	{ (($::mwerks)?"":"DWORD ")."@_"; }
+
+sub ::file
+{   if ($::mwerks)	{ push(@out,".section\t.text,64\n"); }
+    else
+    { my $tmp=<<___;
+%ifidn __OUTPUT_FORMAT__,obj
+section	code	use32 class=code align=64
+%elifidn __OUTPUT_FORMAT__,win32
+\$\@feat.00 equ 1
+section	.text	code align=64
 %else
-section	.text
+section	.text	code
 %endif
 ___
-		push(@out,$tmp);
-		}
-	}
-
-sub main'function_begin
-	{
-	my($func,$extra)=@_;
-
-	push(@labels,$func);
-	push(@out,".") if ($main'mwerks);
-	my($tmp)=<<"EOF";
-global	$under$func
-$under$func:
-	push	ebp
-	push	ebx
-	push	esi
-	push	edi
-EOF
-	push(@out,$tmp);
-	$stack=20;
-	}
-
-sub main'function_begin_B
-	{
-	my($func,$extra)=@_;
-	push(@out,".") if ($main'mwerks);
-	my($tmp)=<<"EOF";
-global	$under$func
-$under$func:
-EOF
-	push(@out,$tmp);
-	$stack=4;
-	}
-
-sub main'function_end
-	{
-	my($func)=@_;
-
-	my($tmp)=<<"EOF";
-	pop	edi
-	pop	esi
-	pop	ebx
-	pop	ebp
-	ret
-EOF
 	push(@out,$tmp);
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_B
-	{
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_A
-	{
-	my($func)=@_;
-
-	my($tmp)=<<"EOF";
-	pop	edi
-	pop	esi
-	pop	ebx
-	pop	ebp
-	ret
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'file_end
-	{
-	}
-
-sub main'wparam
-	{
-	my($num)=@_;
+    }
+}
 
-	return(&main'DWP($stack+$num*4,"esp","",0));
-	}
+sub ::function_begin_B
+{ my $func=shift;
+  my $global=($func !~ /^_/);
+  my $begin="${::lbdecor}_${func}_begin";
 
-sub main'swtmp
-	{
-	return(&main'DWP($_[0]*4,"esp","",0));
-	}
+    $begin =~ s/^\@/./ if ($::mwerks);	# the torture never stops
 
-# Should use swtmp, which is above esp.  Linix can trash the stack above esp
-#sub main'wtmp
-#	{
-#	my($num)=@_;
-#
-#	return(&main'DWP(-(($num+1)*4),"esp","",0));
-#	}
+    &::LABEL($func,$global?"$begin":"$nmdecor$func");
+    $func=$nmdecor.$func;
 
-sub main'comment
-	{
-	foreach (@_)
-		{
-		push(@out,"\t; $_\n");
-		}
-	}
+    push(@out,"${drdecor}global	$func\n")	if ($global);
+    push(@out,"${drdecor}align	16\n");
+    push(@out,"$func:\n");
+    push(@out,"$begin:\n")			if ($global);
+    $::stack=4;
+}
 
-sub main'public_label
-	{
-	$label{$_[0]}="${under}${_[0]}"	if (!defined($label{$_[0]}));
-	push(@out,".") if ($main'mwerks);
-	push(@out,"global\t$label{$_[0]}\n");
-	}
+sub ::function_end_B
+{   $::stack=0;
+    &::wipe_labels();
+}
 
-sub main'label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="\@${label}${_[0]}";
-		$label++;
-		}
-	return($label{$_[0]});
-	}
+sub ::file_end
+{   if (grep {/\b${nmdecor}OPENSSL_ia32cap_P\b/i} @out)
+    {	my $comm=<<___;
+${drdecor}segment	.bss
+${drdecor}common	${nmdecor}OPENSSL_ia32cap_P 8
+___
+	# comment out OPENSSL_ia32cap_P declarations
+	grep {s/(^extern\s+${nmdecor}OPENSSL_ia32cap_P)/\;$1/} @out;
+	push (@out,$comm)
+    }
+    push (@out,$initseg) if ($initseg);		
+}
 
-sub main'set_label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="\@${label}${_[0]}";
-		$label++;
-		}
-	if ($_[1]!=0 && $_[1]>1)
-		{
-		main'align($_[1]);
-		}
-	push(@out,"$label{$_[0]}:\n");
-	}
+sub ::comment {   foreach (@_) { push(@out,"\t; $_\n"); }   }
 
-sub main'data_byte
-	{
-	push(@out,(($main'mwerks)?".byte\t":"DB\t").join(',',@_)."\n");
-	}
+sub ::external_label
+{   foreach(@_)
+    {	push(@out,"${drdecor}extern\t".&::LABEL($_,$nmdecor.$_)."\n");   }
+}
 
-sub main'data_word
-	{
-	push(@out,(($main'mwerks)?".long\t":"DD\t").join(',',@_)."\n");
-	}
+sub ::public_label
+{   push(@out,"${drdecor}global\t".&::LABEL($_[0],$nmdecor.$_[0])."\n");  }
 
-sub main'align
-	{
-	push(@out,".") if ($main'mwerks);
-	push(@out,"align\t$_[0]\n");
-	}
+sub ::data_byte
+{   push(@out,(($::mwerks)?".byte\t":"db\t").join(',',@_)."\n");	}
+sub ::data_short
+{   push(@out,(($::mwerks)?".word\t":"dw\t").join(',',@_)."\n");	}
+sub ::data_word
+{   push(@out,(($::mwerks)?".long\t":"dd\t").join(',',@_)."\n");	}
 
-sub out1p
-	{
-	my($name,$p1)=@_;
-	my($l,$t);
+sub ::align
+{   push(@out,"${drdecor}align\t$_[0]\n");	}
 
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
+sub ::picmeup
+{ my($dst,$sym)=@_;
+    &::lea($dst,&::DWP($sym));
+}
 
-sub main'picmeup
-	{
-	local($dst,$sym)=@_;
-	&main'lea($dst,&main'DWP($sym));
-	}
+sub ::initseg
+{ my $f=$nmdecor.shift;
+    if ($::win32)
+    {	$initseg=<<___;
+segment	.CRT\$XCU data align=4
+extern	$f
+dd	$f
+___
+    }
+}
 
-sub main'blindpop { &out1("pop",@_); }
+sub ::dataseg
+{   if ($mwerks)	{ push(@out,".section\t.data,4\n");   }
+    else		{ push(@out,"section\t.data align=4\n"); }
+}
 
-sub main'initseg
-	{
-	local($f)=@_;
-	if ($main'win32)
-		{
-		local($tmp)=<<___;
-segment	.CRT\$XCU data
-extern	$under$f
-DD	$under$f
-___
-		push(@out,$tmp);
-		}
-	}
+sub ::safeseh
+{ my $nm=shift;
+    push(@out,"%if	__NASM_VERSION_ID__ >= 0x02030000\n");
+    push(@out,"safeseh	".&::LABEL($nm,$nmdecor.$nm)."\n");
+    push(@out,"%endif\n");
+}
 
 1;
diff --git a/crypto/perlasm/x86unix.pl b/crypto/perlasm/x86unix.pl
deleted file mode 100644
index a4c947165e53..000000000000
--- a/crypto/perlasm/x86unix.pl
+++ /dev/null
@@ -1,725 +0,0 @@
-#!/usr/local/bin/perl
-
-package x86unix;	# GAS actually...
-
-$label="L000";
-$const="";
-$constl=0;
-
-$align=($main'aout)?"4":"16";
-$under=($main'aout or $main'coff)?"_":"";
-$dot=($main'aout)?"":".";
-$com_start="#" if ($main'aout or $main'coff);
-
-sub main'asm_init_output { @out=(); }
-sub main'asm_get_output { return(@out); }
-sub main'get_labels { return(@labels); }
-sub main'external_label { push(@labels,@_); }
-
-if ($main'cpp)
-	{
-	$align="ALIGN";
-	$under="";
-	$com_start='/*';
-	$com_end='*/';
-	}
-
-%lb=(	'eax',	'%al',
-	'ebx',	'%bl',
-	'ecx',	'%cl',
-	'edx',	'%dl',
-	'ax',	'%al',
-	'bx',	'%bl',
-	'cx',	'%cl',
-	'dx',	'%dl',
-	);
-
-%hb=(	'eax',	'%ah',
-	'ebx',	'%bh',
-	'ecx',	'%ch',
-	'edx',	'%dh',
-	'ax',	'%ah',
-	'bx',	'%bh',
-	'cx',	'%ch',
-	'dx',	'%dh',
-	);
-
-%regs=(	'eax',	'%eax',
-	'ebx',	'%ebx',
-	'ecx',	'%ecx',
-	'edx',	'%edx',
-	'esi',	'%esi',
-	'edi',	'%edi',
-	'ebp',	'%ebp',
-	'esp',	'%esp',
-
-	'mm0',	'%mm0',
-	'mm1',	'%mm1',
-	'mm2',	'%mm2',
-	'mm3',	'%mm3',
-	'mm4',	'%mm4',
-	'mm5',	'%mm5',
-	'mm6',	'%mm6',
-	'mm7',	'%mm7',
-
-	'xmm0',	'%xmm0',
-	'xmm1',	'%xmm1',
-	'xmm2',	'%xmm2',
-	'xmm3',	'%xmm3',
-	'xmm4',	'%xmm4',
-	'xmm5',	'%xmm5',
-	'xmm6',	'%xmm6',
-	'xmm7',	'%xmm7',
-	);
-
-%reg_val=(
-	'eax',	0x00,
-	'ebx',	0x03,
-	'ecx',	0x01,
-	'edx',	0x02,
-	'esi',	0x06,
-	'edi',	0x07,
-	'ebp',	0x05,
-	'esp',	0x04,
-	);
-
-sub main'LB
-	{
-	(defined($lb{$_[0]})) || die "$_[0] does not have a 'low byte'\n";
-	return($lb{$_[0]});
-	}
-
-sub main'HB
-	{
-	(defined($hb{$_[0]})) || die "$_[0] does not have a 'high byte'\n";
-	return($hb{$_[0]});
-	}
-
-sub main'DWP
-	{
-	local($addr,$reg1,$reg2,$idx)=@_;
-
-	$ret="";
-	$addr =~ s/(^|[+ \t])([A-Za-z_]+[A-Za-z0-9_]+)($|[+ \t])/$1$under$2$3/;
-	$reg1="$regs{$reg1}" if defined($regs{$reg1});
-	$reg2="$regs{$reg2}" if defined($regs{$reg2});
-	$ret.=$addr if ($addr ne "") && ($addr ne 0);
-	if ($reg2 ne "")
-		{
-		if($idx ne "" && $idx != 0)
-		    { $ret.="($reg1,$reg2,$idx)"; }
-		else
-		    { $ret.="($reg1,$reg2)"; }
-	        }
-	elsif ($reg1 ne "")
-		{ $ret.="($reg1)" }
-	return($ret);
-	}
-
-sub main'QWP
-	{
-	return(&main'DWP(@_));
-	}
-
-sub main'BP
-	{
-	return(&main'DWP(@_));
-	}
-
-sub main'BC
-	{
-	return @_;
-	}
-
-sub main'DWC
-	{
-	return @_;
-	}
-
-#sub main'BP
-#	{
-#	local($addr,$reg1,$reg2,$idx)=@_;
-#
-#	$ret="";
-#
-#	$addr =~ s/(^|[+ \t])([A-Za-z_]+)($|[+ \t])/$1$under$2$3/;
-#	$reg1="$regs{$reg1}" if defined($regs{$reg1});
-#	$reg2="$regs{$reg2}" if defined($regs{$reg2});
-#	$ret.=$addr if ($addr ne "") && ($addr ne 0);
-#	if ($reg2 ne "")
-#		{ $ret.="($reg1,$reg2,$idx)"; }
-#	else
-#		{ $ret.="($reg1)" }
-#	return($ret);
-#	}
-
-sub main'mov	{ &out2("movl",@_); }
-sub main'movb	{ &out2("movb",@_); }
-sub main'and	{ &out2("andl",@_); }
-sub main'or	{ &out2("orl",@_); }
-sub main'shl	{ &out2("sall",@_); }
-sub main'shr	{ &out2("shrl",@_); }
-sub main'xor	{ &out2("xorl",@_); }
-sub main'xorb	{ &out2("xorb",@_); }
-sub main'add	{ &out2($_[0]=~/%[a-d][lh]/?"addb":"addl",@_); }
-sub main'adc	{ &out2("adcl",@_); }
-sub main'sub	{ &out2("subl",@_); }
-sub main'sbb	{ &out2("sbbl",@_); }
-sub main'rotl	{ &out2("roll",@_); }
-sub main'rotr	{ &out2("rorl",@_); }
-sub main'exch	{ &out2($_[0]=~/%[a-d][lh]/?"xchgb":"xchgl",@_); }
-sub main'cmp	{ &out2("cmpl",@_); }
-sub main'lea	{ &out2("leal",@_); }
-sub main'mul	{ &out1("mull",@_); }
-sub main'imul	{ &out2("imull",@_); }
-sub main'div	{ &out1("divl",@_); }
-sub main'jmp	{ &out1("jmp",@_); }
-sub main'jmp_ptr { &out1p("jmp",@_); }
-sub main'je	{ &out1("je",@_); }
-sub main'jle	{ &out1("jle",@_); }
-sub main'jne	{ &out1("jne",@_); }
-sub main'jnz	{ &out1("jnz",@_); }
-sub main'jz	{ &out1("jz",@_); }
-sub main'jge	{ &out1("jge",@_); }
-sub main'jl	{ &out1("jl",@_); }
-sub main'ja	{ &out1("ja",@_); }
-sub main'jae	{ &out1("jae",@_); }
-sub main'jb	{ &out1("jb",@_); }
-sub main'jbe	{ &out1("jbe",@_); }
-sub main'jc	{ &out1("jc",@_); }
-sub main'jnc	{ &out1("jnc",@_); }
-sub main'jno	{ &out1("jno",@_); }
-sub main'dec	{ &out1("decl",@_); }
-sub main'inc	{ &out1($_[0]=~/%[a-d][hl]/?"incb":"incl",@_); }
-sub main'push	{ &out1("pushl",@_); $stack+=4; }
-sub main'pop	{ &out1("popl",@_); $stack-=4; }
-sub main'pushf	{ &out0("pushfl"); $stack+=4; }
-sub main'popf	{ &out0("popfl"); $stack-=4; }
-sub main'not	{ &out1("notl",@_); }
-sub main'call	{	my $pre=$under;
-			foreach $i (%label)
-			{ if ($label{$i} eq $_[0]) { $pre=''; last; } }
-			&out1("call",$pre.$_[0]);
-		}
-sub main'call_ptr { &out1p("call",@_); }
-sub main'ret	{ &out0("ret"); }
-sub main'nop	{ &out0("nop"); }
-sub main'test	{ &out2("testl",@_); }
-sub main'bt	{ &out2("btl",@_); }
-sub main'leave	{ &out0("leave"); }
-sub main'cpuid	{ &out0(".byte\t0x0f,0xa2"); }
-sub main'rdtsc	{ &out0(".byte\t0x0f,0x31"); }
-sub main'halt	{ &out0("hlt"); }
-sub main'movz	{ &out2("movzbl",@_); }
-sub main'neg	{ &out1("negl",@_); }
-sub main'cld	{ &out0("cld"); }
-
-# SSE2
-sub main'emms	{ &out0("emms"); }
-sub main'movd	{ &out2("movd",@_); }
-sub main'movdqu	{ &out2("movdqu",@_); }
-sub main'movdqa	{ &out2("movdqa",@_); }
-sub main'movdq2q{ &out2("movdq2q",@_); }
-sub main'movq2dq{ &out2("movq2dq",@_); }
-sub main'paddq	{ &out2("paddq",@_); }
-sub main'pmuludq{ &out2("pmuludq",@_); }
-sub main'psrlq	{ &out2("psrlq",@_); }
-sub main'psllq	{ &out2("psllq",@_); }
-sub main'pxor	{ &out2("pxor",@_); }
-sub main'por	{ &out2("por",@_); }
-sub main'pand	{ &out2("pand",@_); }
-sub main'movq	{
-	local($p1,$p2,$optimize)=@_;
-	if ($optimize && $p1=~/^mm[0-7]$/ && $p2=~/^mm[0-7]$/)
-		# movq between mmx registers can sink Intel CPUs
-		{	push(@out,"\tpshufw\t\$0xe4,%$p2,%$p1\n");	}
-	else	{	&out2("movq",@_);				}
-	}
-
-# The bswapl instruction is new for the 486. Emulate if i386.
-sub main'bswap
-	{
-	if ($main'i386)
-		{
-		&main'comment("bswapl @_");
-		&main'exch(main'HB(@_),main'LB(@_));
-		&main'rotr(@_,16);
-		&main'exch(main'HB(@_),main'LB(@_));
-		}
-	else
-		{
-		&out1("bswapl",@_);
-		}
-	}
-
-sub out2
-	{
-	local($name,$p1,$p2)=@_;
-	local($l,$ll,$t);
-	local(%special)=(	"roll",0xD1C0,"rorl",0xD1C8,
-				"rcll",0xD1D0,"rcrl",0xD1D8,
-				"shll",0xD1E0,"shrl",0xD1E8,
-				"sarl",0xD1F8);
-	
-	if ((defined($special{$name})) && defined($regs{$p1}) && ($p2 == 1))
-		{
-		$op=$special{$name}|$reg_val{$p1};
-		$tmp1=sprintf(".byte %d\n",($op>>8)&0xff);
-		$tmp2=sprintf(".byte %d\t",$op     &0xff);
-		push(@out,$tmp1);
-		push(@out,$tmp2);
-
-		$p2=&conv($p2);
-		$p1=&conv($p1);
-		&main'comment("$name $p2 $p1");
-		return;
-		}
-
-	push(@out,"\t$name\t");
-	$t=&conv($p2).",";
-	$l=length($t);
-	push(@out,$t);
-	$ll=4-($l+9)/8;
-	$tmp1=sprintf("\t" x $ll);
-	push(@out,$tmp1);
-	push(@out,&conv($p1)."\n");
-	}
-
-sub out1
-	{
-	local($name,$p1)=@_;
-	local($l,$t);
-	local(%special)=("bswapl",0x0FC8);
-
-	if ((defined($special{$name})) && defined($regs{$p1}))
-		{
-		$op=$special{$name}|$reg_val{$p1};
-		$tmp1=sprintf(".byte %d\n",($op>>8)&0xff);
-		$tmp2=sprintf(".byte %d\t",$op     &0xff);
-		push(@out,$tmp1);
-		push(@out,$tmp2);
-
-		$p2=&conv($p2);
-		$p1=&conv($p1);
-		&main'comment("$name $p2 $p1");
-		return;
-		}
-
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
-
-sub out1p
-	{
-	local($name,$p1)=@_;
-	local($l,$t);
-
-	push(@out,"\t$name\t*".&conv($p1)."\n");
-	}
-
-sub out0
-	{
-	push(@out,"\t$_[0]\n");
-	}
-
-sub conv
-	{
-	local($p)=@_;
-
-#	$p =~ s/0x([0-9A-Fa-f]+)/0$1h/;
-
-	$p=$regs{$p} if (defined($regs{$p}));
-
-	$p =~ s/^(-{0,1}[0-9A-Fa-f]+)$/\$$1/;
-	$p =~ s/^(0x[0-9A-Fa-f]+)$/\$$1/;
-	return $p;
-	}
-
-sub main'file
-	{
-	local($file)=@_;
-
-	local($tmp)=<<"EOF";
-	.file	"$file.s"
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'function_begin
-	{
-	local($func)=@_;
-
-	&main'external_label($func);
-	$func=$under.$func;
-
-	local($tmp)=<<"EOF";
-.text
-.globl	$func
-EOF
-	push(@out,$tmp);
-	if ($main'cpp)
-		{ $tmp=push(@out,"TYPE($func,\@function)\n"); }
-	elsif ($main'coff)
-		{ $tmp=push(@out,".def\t$func;\t.scl\t2;\t.type\t32;\t.endef\n"); }
-	elsif ($main'aout and !$main'pic)
-		{ }
-	else	{ $tmp=push(@out,".type\t$func,\@function\n"); }
-	push(@out,".align\t$align\n");
-	push(@out,"$func:\n");
-	$tmp=<<"EOF";
-	pushl	%ebp
-	pushl	%ebx
-	pushl	%esi
-	pushl	%edi
-
-EOF
-	push(@out,$tmp);
-	$stack=20;
-	}
-
-sub main'function_begin_B
-	{
-	local($func,$extra)=@_;
-
-	&main'external_label($func);
-	$func=$under.$func;
-
-	local($tmp)=<<"EOF";
-.text
-.globl	$func
-EOF
-	push(@out,$tmp);
-	if ($main'cpp)
-		{ push(@out,"TYPE($func,\@function)\n"); }
-	elsif ($main'coff)
-		{ $tmp=push(@out,".def\t$func;\t.scl\t2;\t.type\t32;\t.endef\n"); }
-	elsif ($main'aout and !$main'pic)
-		{ }
-	else	{ push(@out,".type	$func,\@function\n"); }
-	push(@out,".align\t$align\n");
-	push(@out,"$func:\n");
-	$stack=4;
-	}
-
-sub main'function_end
-	{
-	local($func)=@_;
-
-	$func=$under.$func;
-
-	local($tmp)=<<"EOF";
-	popl	%edi
-	popl	%esi
-	popl	%ebx
-	popl	%ebp
-	ret
-${dot}L_${func}_end:
-EOF
-	push(@out,$tmp);
-
-	if ($main'cpp)
-		{ push(@out,"SIZE($func,${dot}L_${func}_end-$func)\n"); }
-	elsif ($main'coff or $main'aout)
-                { }
-	else	{ push(@out,".size\t$func,${dot}L_${func}_end-$func\n"); }
-	push(@out,".ident	\"$func\"\n");
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_A
-	{
-	local($func)=@_;
-
-	local($tmp)=<<"EOF";
-	popl	%edi
-	popl	%esi
-	popl	%ebx
-	popl	%ebp
-	ret
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'function_end_B
-	{
-	local($func)=@_;
-
-	$func=$under.$func;
-
-	push(@out,"${dot}L_${func}_end:\n");
-	if ($main'cpp)
-		{ push(@out,"SIZE($func,${dot}L_${func}_end-$func)\n"); }
-        elsif ($main'coff or $main'aout)
-                { }
-	else	{ push(@out,".size\t$func,${dot}L_${func}_end-$func\n"); }
-	push(@out,".ident	\"$func\"\n");
-	$stack=0;
-	%label=();
-	}
-
-sub main'wparam
-	{
-	local($num)=@_;
-
-	return(&main'DWP($stack+$num*4,"esp","",0));
-	}
-
-sub main'stack_push
-	{
-	local($num)=@_;
-	$stack+=$num*4;
-	&main'sub("esp",$num*4);
-	}
-
-sub main'stack_pop
-	{
-	local($num)=@_;
-	$stack-=$num*4;
-	&main'add("esp",$num*4);
-	}
-
-sub main'swtmp
-	{
-	return(&main'DWP($_[0]*4,"esp","",0));
-	}
-
-# Should use swtmp, which is above esp.  Linix can trash the stack above esp
-#sub main'wtmp
-#	{
-#	local($num)=@_;
-#
-#	return(&main'DWP(-($num+1)*4,"esp","",0));
-#	}
-
-sub main'comment
-	{
-	if (!defined($com_start) or $main'elf)
-		{	# Regarding $main'elf above...
-			# GNU and SVR4 as'es use different comment delimiters,
-		push(@out,"\n");	# so we just skip ELF comments...
-		return;
-		}
-	foreach (@_)
-		{
-		if (/^\s*$/)
-			{ push(@out,"\n"); }
-		else
-			{ push(@out,"\t$com_start $_ $com_end\n"); }
-		}
-	}
-
-sub main'public_label
-	{
-	$label{$_[0]}="${under}${_[0]}"	if (!defined($label{$_[0]}));
-	push(@out,".globl\t$label{$_[0]}\n");
-	}
-
-sub main'label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="${dot}${label}${_[0]}";
-		$label++;
-		}
-	return($label{$_[0]});
-	}
-
-sub main'set_label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="${dot}${label}${_[0]}";
-		$label++;
-		}
-	if ($_[1]!=0)
-		{
-		if ($_[1]>1)	{ main'align($_[1]);		}
-		else		{ push(@out,".align $align\n");	}
-		}
-	push(@out,"$label{$_[0]}:\n");
-	}
-
-sub main'file_end
-	{
-	# try to detect if SSE2 or MMX extensions were used on ELF platform...
-	if ($main'elf && grep {/\b%[x]*mm[0-7]\b|OPENSSL_ia32cap_P\b/i} @out) {
-		local($tmp);
-
-		push (@out,"\n.section\t.bss\n");
-		push (@out,".comm\t${under}OPENSSL_ia32cap_P,4,4\n");
-
-		return;
-	}
-
-	if ($const ne "")
-		{
-		push(@out,".section .rodata\n");
-		push(@out,$const);
-		$const="";
-		}
-	}
-
-sub main'data_byte
-	{
-	push(@out,"\t.byte\t".join(',',@_)."\n");
-	}
-
-sub main'data_word
-	{
-	push(@out,"\t.long\t".join(',',@_)."\n");
-	}
-
-sub main'align
-	{
-	my $val=$_[0],$p2,$i;
-	if ($main'aout) {
-		for ($p2=0;$val!=0;$val>>=1) { $p2++; }
-		$val=$p2-1;
-		$val.=",0x90";
-	}
-	push(@out,".align\t$val\n");
-	}
-
-# debug output functions: puts, putx, printf
-
-sub main'puts
-	{
-	&pushvars();
-	&main'push('$Lstring' . ++$constl);
-	&main'call('puts');
-	$stack-=4;
-	&main'add("esp",4);
-	&popvars();
-
-	$const .= "Lstring$constl:\n\t.string \"@_[0]\"\n";
-	}
-
-sub main'putx
-	{
-	&pushvars();
-	&main'push($_[0]);
-	&main'push('$Lstring' . ++$constl);
-	&main'call('printf');
-	&main'add("esp",8);
-	$stack-=8;
-	&popvars();
-
-	$const .= "Lstring$constl:\n\t.string \"\%X\"\n";
-	}
-
-sub main'printf
-	{
-	$ostack = $stack;
-	&pushvars();
-	for ($i = @_ - 1; $i >= 0; $i--)
-		{
-		if ($i == 0) # change this to support %s format strings
-			{
-			&main'push('$Lstring' . ++$constl);
-			$const .= "Lstring$constl:\n\t.string \"@_[$i]\"\n";
-			}
-		else
-			{
-			if ($_[$i] =~ /([0-9]*)\(%esp\)/)
-				{
-				&main'push(($1 + $stack - $ostack) . '(%esp)');
-				}
-			else
-				{
-				&main'push($_[$i]);
-				}
-			}
-		}
-	&main'call('printf');
-	$stack-=4*@_;
-	&main'add("esp",4*@_);
-	&popvars();
-	}
-
-sub pushvars
-	{
-	&main'pushf();
-	&main'push("edx");
-	&main'push("ecx");
-	&main'push("eax");
-	}
-
-sub popvars
-	{
-	&main'pop("eax");
-	&main'pop("ecx");
-	&main'pop("edx");
-	&main'popf();
-	}
-
-sub main'picmeup
-	{
-	local($dst,$sym)=@_;
-	if ($main'cpp)
-		{
-		local($tmp)=<<___;
-#if (defined(ELF) || defined(SOL)) && defined(PIC)
-	call	1f
-1:	popl	$regs{$dst}
-	addl	\$_GLOBAL_OFFSET_TABLE_+[.-1b],$regs{$dst}
-	movl	$sym\@GOT($regs{$dst}),$regs{$dst}
-#else
-	leal	$sym,$regs{$dst}
-#endif
-___
-		push(@out,$tmp);
-		}
-	elsif ($main'pic && ($main'elf || $main'aout))
-		{
-		&main'call(&main'label("PIC_me_up"));
-		&main'set_label("PIC_me_up");
-		&main'blindpop($dst);
-		&main'add($dst,"\$${under}_GLOBAL_OFFSET_TABLE_+[.-".
-				&main'label("PIC_me_up") . "]");
-		&main'mov($dst,&main'DWP($under.$sym."\@GOT",$dst));
-		}
-	else
-		{
-		&main'lea($dst,&main'DWP($sym));
-		}
-	}
-
-sub main'blindpop { &out1("popl",@_); }
-
-sub main'initseg
-	{
-	local($f)=@_;
-	local($tmp);
-	if ($main'elf)
-		{
-		$tmp=<<___;
-.section	.init
-	call	$under$f
-	jmp	.Linitalign
-.align	$align
-.Linitalign:
-___
-		}
-	elsif ($main'coff)
-		{
-		$tmp=<<___;	# applies to both Cygwin and Mingw
-.section	.ctors
-.long	$under$f
-___
-		}
-	elsif ($main'aout)
-		{
-		local($ctor)="${under}_GLOBAL_\$I\$$f";
-		$tmp=".text\n";
-		$tmp.=".type	$ctor,\@function\n" if ($main'pic);
-		$tmp.=<<___;	# OpenBSD way...
-.globl	$ctor
-.align	2
-$ctor:
-	jmp	$under$f
-___
-		}
-	push(@out,$tmp) if ($tmp);
-	}
-
-1;
diff --git a/crypto/pkcs12/Makefile b/crypto/pkcs12/Makefile
index eed226b30d31..3a7498fe7ad9 100644
--- a/crypto/pkcs12/Makefile
+++ b/crypto/pkcs12/Makefile
@@ -39,7 +39,7 @@ test:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -85,37 +85,36 @@ p12_add.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_add.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_add.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_add.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_add.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_add.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_add.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-p12_add.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-p12_add.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
-p12_add.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p12_add.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p12_add.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p12_add.o: ../cryptlib.h p12_add.c
+p12_add.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_add.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_add.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+p12_add.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
+p12_add.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+p12_add.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p12_add.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p12_add.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p12_add.c
 p12_asn.o: ../../e_os.h ../../include/openssl/asn1.h
 p12_asn.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 p12_asn.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p12_asn.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p12_asn.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p12_asn.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p12_asn.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p12_asn.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p12_asn.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-p12_asn.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
-p12_asn.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-p12_asn.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-p12_asn.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-p12_asn.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p12_asn.c
+p12_asn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p12_asn.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p12_asn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+p12_asn.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
+p12_asn.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+p12_asn.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+p12_asn.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+p12_asn.o: ../cryptlib.h p12_asn.c
 p12_attr.o: ../../e_os.h ../../include/openssl/asn1.h
 p12_attr.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_attr.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_attr.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_attr.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_attr.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_attr.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_attr.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p12_attr.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_attr.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_attr.o: ../../include/openssl/opensslconf.h
 p12_attr.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p12_attr.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
 p12_attr.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -127,9 +126,9 @@ p12_crpt.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_crpt.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_crpt.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_crpt.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_crpt.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_crpt.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_crpt.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p12_crpt.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_crpt.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_crpt.o: ../../include/openssl/opensslconf.h
 p12_crpt.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p12_crpt.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
 p12_crpt.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -141,23 +140,22 @@ p12_crt.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_crt.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_crt.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_crt.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_crt.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_crt.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_crt.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-p12_crt.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-p12_crt.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
-p12_crt.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p12_crt.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p12_crt.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p12_crt.o: ../cryptlib.h p12_crt.c
+p12_crt.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_crt.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_crt.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+p12_crt.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
+p12_crt.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+p12_crt.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p12_crt.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p12_crt.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p12_crt.c
 p12_decr.o: ../../e_os.h ../../include/openssl/asn1.h
 p12_decr.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_decr.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_decr.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_decr.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_decr.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_decr.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_decr.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p12_decr.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_decr.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_decr.o: ../../include/openssl/opensslconf.h
 p12_decr.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p12_decr.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
 p12_decr.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -169,9 +167,9 @@ p12_init.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_init.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_init.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_init.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_init.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_init.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_init.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p12_init.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_init.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_init.o: ../../include/openssl/opensslconf.h
 p12_init.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p12_init.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
 p12_init.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -184,22 +182,22 @@ p12_key.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p12_key.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p12_key.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p12_key.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p12_key.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p12_key.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p12_key.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-p12_key.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
-p12_key.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-p12_key.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-p12_key.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-p12_key.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p12_key.c
+p12_key.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p12_key.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p12_key.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+p12_key.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
+p12_key.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+p12_key.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+p12_key.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+p12_key.o: ../cryptlib.h p12_key.c
 p12_kiss.o: ../../e_os.h ../../include/openssl/asn1.h
 p12_kiss.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_kiss.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_kiss.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_kiss.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_kiss.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_kiss.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_kiss.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+p12_kiss.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_kiss.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_kiss.o: ../../include/openssl/opensslconf.h
 p12_kiss.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p12_kiss.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
 p12_kiss.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -211,10 +209,9 @@ p12_mutl.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_mutl.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_mutl.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_mutl.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_mutl.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_mutl.o: ../../include/openssl/hmac.h ../../include/openssl/lhash.h
-p12_mutl.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p12_mutl.o: ../../include/openssl/opensslconf.h
+p12_mutl.o: ../../include/openssl/evp.h ../../include/openssl/hmac.h
+p12_mutl.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p12_mutl.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 p12_mutl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p12_mutl.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
 p12_mutl.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
@@ -226,9 +223,8 @@ p12_npas.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 p12_npas.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 p12_npas.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 p12_npas.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-p12_npas.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-p12_npas.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-p12_npas.o: ../../include/openssl/opensslconf.h
+p12_npas.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+p12_npas.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 p12_npas.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 p12_npas.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
 p12_npas.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
@@ -241,53 +237,50 @@ p12_p8d.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_p8d.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_p8d.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_p8d.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_p8d.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_p8d.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_p8d.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-p12_p8d.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-p12_p8d.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
-p12_p8d.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p12_p8d.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p12_p8d.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p12_p8d.o: ../cryptlib.h p12_p8d.c
+p12_p8d.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_p8d.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_p8d.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+p12_p8d.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
+p12_p8d.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+p12_p8d.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p12_p8d.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p12_p8d.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p12_p8d.c
 p12_p8e.o: ../../e_os.h ../../include/openssl/asn1.h
 p12_p8e.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_p8e.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_p8e.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_p8e.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_p8e.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_p8e.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_p8e.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-p12_p8e.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-p12_p8e.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
-p12_p8e.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p12_p8e.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p12_p8e.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p12_p8e.o: ../cryptlib.h p12_p8e.c
+p12_p8e.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_p8e.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_p8e.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+p12_p8e.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
+p12_p8e.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+p12_p8e.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p12_p8e.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p12_p8e.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p12_p8e.c
 p12_utl.o: ../../e_os.h ../../include/openssl/asn1.h
 p12_utl.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 p12_utl.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 p12_utl.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 p12_utl.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-p12_utl.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-p12_utl.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-p12_utl.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-p12_utl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-p12_utl.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
-p12_utl.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-p12_utl.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-p12_utl.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-p12_utl.o: ../cryptlib.h p12_utl.c
+p12_utl.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+p12_utl.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+p12_utl.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+p12_utl.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
+p12_utl.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+p12_utl.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+p12_utl.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+p12_utl.o: ../../include/openssl/x509_vfy.h ../cryptlib.h p12_utl.c
 pk12err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 pk12err.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 pk12err.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pk12err.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pk12err.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pk12err.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pk12err.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pk12err.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-pk12err.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs12.h
-pk12err.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-pk12err.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-pk12err.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-pk12err.o: ../../include/openssl/x509_vfy.h pk12err.c
+pk12err.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pk12err.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pk12err.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+pk12err.o: ../../include/openssl/pkcs12.h ../../include/openssl/pkcs7.h
+pk12err.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+pk12err.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+pk12err.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+pk12err.o: pk12err.c
diff --git a/crypto/pkcs12/p12_add.c b/crypto/pkcs12/p12_add.c
index 1f3e378f5cba..27ac5facfa4c 100644
--- a/crypto/pkcs12/p12_add.c
+++ b/crypto/pkcs12/p12_add.c
@@ -106,6 +106,7 @@ PKCS12_SAFEBAG *PKCS12_MAKE_SHKEYBAG(int pbe_nid, const char *pass,
 	     PKCS8_PRIV_KEY_INFO *p8)
 {
 	PKCS12_SAFEBAG *bag;
+	const EVP_CIPHER *pbe_ciph;
 
 	/* Set up the safe bag */
 	if (!(bag = PKCS12_SAFEBAG_new())) {
@@ -114,8 +115,14 @@ PKCS12_SAFEBAG *PKCS12_MAKE_SHKEYBAG(int pbe_nid, const char *pass,
 	}
 
 	bag->type = OBJ_nid2obj(NID_pkcs8ShroudedKeyBag);
+
+	pbe_ciph = EVP_get_cipherbynid(pbe_nid);
+
+	if (pbe_ciph)
+		pbe_nid = -1;
+
 	if (!(bag->value.shkeybag = 
-	  PKCS8_encrypt(pbe_nid, NULL, pass, passlen, salt, saltlen, iter,
+	  PKCS8_encrypt(pbe_nid, pbe_ciph, pass, passlen, salt, saltlen, iter,
 									 p8))) {
 		PKCS12err(PKCS12_F_PKCS12_MAKE_SHKEYBAG, ERR_R_MALLOC_FAILURE);
 		return NULL;
@@ -164,6 +171,7 @@ PKCS7 *PKCS12_pack_p7encdata(int pbe_nid, const char *pass, int passlen,
 {
 	PKCS7 *p7;
 	X509_ALGOR *pbe;
+	const EVP_CIPHER *pbe_ciph;
 	if (!(p7 = PKCS7_new())) {
 		PKCS12err(PKCS12_F_PKCS12_PACK_P7ENCDATA, ERR_R_MALLOC_FAILURE);
 		return NULL;
@@ -173,7 +181,15 @@ PKCS7 *PKCS12_pack_p7encdata(int pbe_nid, const char *pass, int passlen,
 				PKCS12_R_ERROR_SETTING_ENCRYPTED_DATA_TYPE);
 		return NULL;
 	}
-	if (!(pbe = PKCS5_pbe_set(pbe_nid, iter, salt, saltlen))) {
+
+	pbe_ciph = EVP_get_cipherbynid(pbe_nid);
+
+	if (pbe_ciph)
+		pbe = PKCS5_pbe2_set(pbe_ciph, iter, salt, saltlen);
+	else
+		pbe = PKCS5_pbe_set(pbe_nid, iter, salt, saltlen);
+
+	if (!pbe) {
 		PKCS12err(PKCS12_F_PKCS12_PACK_P7ENCDATA, ERR_R_MALLOC_FAILURE);
 		return NULL;
 	}
diff --git a/crypto/pkcs12/p12_attr.c b/crypto/pkcs12/p12_attr.c
index 856933d978d1..e4d9c2564703 100644
--- a/crypto/pkcs12/p12_attr.c
+++ b/crypto/pkcs12/p12_attr.c
@@ -60,12 +60,6 @@
 #include "cryptlib.h"
 #include <openssl/pkcs12.h>
 
-#ifdef OPENSSL_SYS_NETWARE
-/* Rename these functions to avoid name clashes on NetWare OS */
-#define uni2asc OPENSSL_uni2asc
-#define asc2uni OPENSSL_asc2uni
-#endif
-
 /* Add a local keyid to a safebag */
 
 int PKCS12_add_localkeyid(PKCS12_SAFEBAG *bag, unsigned char *name,
@@ -145,7 +139,7 @@ char *PKCS12_get_friendlyname(PKCS12_SAFEBAG *bag)
 	ASN1_TYPE *atype;
 	if (!(atype = PKCS12_get_attr(bag, NID_friendlyName))) return NULL;
 	if (atype->type != V_ASN1_BMPSTRING) return NULL;
-	return uni2asc(atype->value.bmpstring->data,
+	return OPENSSL_uni2asc(atype->value.bmpstring->data,
 				 atype->value.bmpstring->length);
 }
 
diff --git a/crypto/pkcs12/p12_crpt.c b/crypto/pkcs12/p12_crpt.c
index f8b952e27eed..b71d07b4d022 100644
--- a/crypto/pkcs12/p12_crpt.c
+++ b/crypto/pkcs12/p12_crpt.c
@@ -60,28 +60,10 @@
 #include "cryptlib.h"
 #include <openssl/pkcs12.h>
 
-/* PKCS#12 specific PBE functions */
+/* PKCS#12 PBE algorithms now in static table */
 
 void PKCS12_PBE_add(void)
 {
-#ifndef OPENSSL_NO_RC4
-EVP_PBE_alg_add(NID_pbe_WithSHA1And128BitRC4, EVP_rc4(), EVP_sha1(),
-							 PKCS12_PBE_keyivgen);
-EVP_PBE_alg_add(NID_pbe_WithSHA1And40BitRC4, EVP_rc4_40(), EVP_sha1(),
-							 PKCS12_PBE_keyivgen);
-#endif
-#ifndef OPENSSL_NO_DES
-EVP_PBE_alg_add(NID_pbe_WithSHA1And3_Key_TripleDES_CBC,
-		 	EVP_des_ede3_cbc(), EVP_sha1(), PKCS12_PBE_keyivgen);
-EVP_PBE_alg_add(NID_pbe_WithSHA1And2_Key_TripleDES_CBC, 
-			EVP_des_ede_cbc(), EVP_sha1(), PKCS12_PBE_keyivgen);
-#endif
-#ifndef OPENSSL_NO_RC2
-EVP_PBE_alg_add(NID_pbe_WithSHA1And128BitRC2_CBC, EVP_rc2_cbc(),
-					EVP_sha1(), PKCS12_PBE_keyivgen);
-EVP_PBE_alg_add(NID_pbe_WithSHA1And40BitRC2_CBC, EVP_rc2_40_cbc(),
-					EVP_sha1(), PKCS12_PBE_keyivgen);
-#endif
 }
 
 int PKCS12_PBE_keyivgen(EVP_CIPHER_CTX *ctx, const char *pass, int passlen,
diff --git a/crypto/pkcs12/p12_crt.c b/crypto/pkcs12/p12_crt.c
index 9522342fa5e1..96b131defa0c 100644
--- a/crypto/pkcs12/p12_crt.c
+++ b/crypto/pkcs12/p12_crt.c
@@ -59,10 +59,6 @@
 #include <stdio.h>
 #include "cryptlib.h"
 #include <openssl/pkcs12.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 
 
 static int pkcs12_add_bag(STACK_OF(PKCS12_SAFEBAG) **pbags, PKCS12_SAFEBAG *bag);
@@ -94,14 +90,7 @@ PKCS12 *PKCS12_create(char *pass, char *name, EVP_PKEY *pkey, X509 *cert,
 
 	/* Set defaults */
 	if (!nid_cert)
-		{
-#ifdef OPENSSL_FIPS
-		if (FIPS_mode())
-			nid_cert = NID_pbe_WithSHA1And3_Key_TripleDES_CBC;
-		else
-#endif
 		nid_cert = NID_pbe_WithSHA1And40BitRC2_CBC;
-		}
 	if (!nid_key)
 		nid_key = NID_pbe_WithSHA1And3_Key_TripleDES_CBC;
 	if (!iter)
diff --git a/crypto/pkcs12/p12_decr.c b/crypto/pkcs12/p12_decr.c
index ba77dbbe32c5..9d3557e8d7e0 100644
--- a/crypto/pkcs12/p12_decr.c
+++ b/crypto/pkcs12/p12_decr.c
@@ -89,7 +89,14 @@ unsigned char * PKCS12_pbe_crypt(X509_ALGOR *algor, const char *pass,
 		goto err;
 	}
 
-	EVP_CipherUpdate(&ctx, out, &i, in, inlen);
+	if (!EVP_CipherUpdate(&ctx, out, &i, in, inlen))
+		{
+		OPENSSL_free(out);
+		out = NULL;
+		PKCS12err(PKCS12_F_PKCS12_PBE_CRYPT,ERR_R_EVP_LIB);
+		goto err;
+		}
+
 	outlen = i;
 	if(!EVP_CipherFinal_ex(&ctx, out + i, &i)) {
 		OPENSSL_free(out);
diff --git a/crypto/pkcs12/p12_key.c b/crypto/pkcs12/p12_key.c
index 03cbcd8ab3ec..c55c7b60b343 100644
--- a/crypto/pkcs12/p12_key.c
+++ b/crypto/pkcs12/p12_key.c
@@ -69,12 +69,6 @@ extern BIO *bio_err;
 void h__dump (unsigned char *p, int len);
 #endif
 
-#ifdef OPENSSL_SYS_NETWARE
-/* Rename these functions to avoid name clashes on NetWare OS */
-#define uni2asc OPENSSL_uni2asc
-#define asc2uni OPENSSL_asc2uni
-#endif
-
 /* PKCS12 compatible key/IV generation */
 #ifndef min
 #define min(a,b) ((a) < (b) ? (a) : (b))
@@ -87,15 +81,18 @@ int PKCS12_key_gen_asc(const char *pass, int passlen, unsigned char *salt,
 	int ret;
 	unsigned char *unipass;
 	int uniplen;
+
 	if(!pass) {
 		unipass = NULL;
 		uniplen = 0;
-	} else if (!asc2uni(pass, passlen, &unipass, &uniplen)) {
+	} else if (!OPENSSL_asc2uni(pass, passlen, &unipass, &uniplen)) {
 		PKCS12err(PKCS12_F_PKCS12_KEY_GEN_ASC,ERR_R_MALLOC_FAILURE);
 		return 0;
 	}
 	ret = PKCS12_key_gen_uni(unipass, uniplen, salt, saltlen,
 						 id, iter, n, out, md_type);
+	if (ret <= 0)
+	    return 0;
 	if(unipass) {
 		OPENSSL_cleanse(unipass, uniplen);	/* Clear password from memory */
 		OPENSSL_free(unipass);
@@ -136,6 +133,8 @@ int PKCS12_key_gen_uni(unsigned char *pass, int passlen, unsigned char *salt,
 #endif
 	v = EVP_MD_block_size (md_type);
 	u = EVP_MD_size (md_type);
+	if (u < 0)
+	    return 0;
 	D = OPENSSL_malloc (v);
 	Ai = OPENSSL_malloc (u);
 	B = OPENSSL_malloc (v + 1);
@@ -153,14 +152,16 @@ int PKCS12_key_gen_uni(unsigned char *pass, int passlen, unsigned char *salt,
 	for (i = 0; i < Slen; i++) *p++ = salt[i % saltlen];
 	for (i = 0; i < Plen; i++) *p++ = pass[i % passlen];
 	for (;;) {
-		EVP_DigestInit_ex(&ctx, md_type, NULL);
-		EVP_DigestUpdate(&ctx, D, v);
-		EVP_DigestUpdate(&ctx, I, Ilen);
-		EVP_DigestFinal_ex(&ctx, Ai, NULL);
+		if (!EVP_DigestInit_ex(&ctx, md_type, NULL)
+			|| !EVP_DigestUpdate(&ctx, D, v)
+			|| !EVP_DigestUpdate(&ctx, I, Ilen)
+			|| !EVP_DigestFinal_ex(&ctx, Ai, NULL))
+			goto err;
 		for (j = 1; j < iter; j++) {
-			EVP_DigestInit_ex(&ctx, md_type, NULL);
-			EVP_DigestUpdate(&ctx, Ai, u);
-			EVP_DigestFinal_ex(&ctx, Ai, NULL);
+			if (!EVP_DigestInit_ex(&ctx, md_type, NULL)
+				|| !EVP_DigestUpdate(&ctx, Ai, u)
+				|| !EVP_DigestFinal_ex(&ctx, Ai, NULL))
+			goto err;
 		}
 		memcpy (out, Ai, min (n, u));
 		if (u >= n) {
diff --git a/crypto/pkcs12/p12_kiss.c b/crypto/pkcs12/p12_kiss.c
index 5c4c6ec98889..206b1b0b18ab 100644
--- a/crypto/pkcs12/p12_kiss.c
+++ b/crypto/pkcs12/p12_kiss.c
@@ -63,16 +63,13 @@
 /* Simplified PKCS#12 routines */
 
 static int parse_pk12( PKCS12 *p12, const char *pass, int passlen,
-		EVP_PKEY **pkey, X509 **cert, STACK_OF(X509) **ca);
+		EVP_PKEY **pkey, STACK_OF(X509) *ocerts);
 
 static int parse_bags( STACK_OF(PKCS12_SAFEBAG) *bags, const char *pass,
-		       int passlen, EVP_PKEY **pkey, X509 **cert,
-		       STACK_OF(X509) **ca, ASN1_OCTET_STRING **keyid,
-		       char *keymatch);
+		       int passlen, EVP_PKEY **pkey, STACK_OF(X509) *ocerts);
 
 static int parse_bag( PKCS12_SAFEBAG *bag, const char *pass, int passlen,
-			EVP_PKEY **pkey, X509 **cert, STACK_OF(X509) **ca,
-			ASN1_OCTET_STRING **keyid, char *keymatch);
+			EVP_PKEY **pkey, STACK_OF(X509) *ocerts);
 
 /* Parse and decrypt a PKCS#12 structure returning user key, user cert
  * and other (CA) certs. Note either ca should be NULL, *ca should be NULL,
@@ -83,24 +80,20 @@ static int parse_bag( PKCS12_SAFEBAG *bag, const char *pass, int passlen,
 int PKCS12_parse(PKCS12 *p12, const char *pass, EVP_PKEY **pkey, X509 **cert,
 	     STACK_OF(X509) **ca)
 {
-
+	STACK_OF(X509) *ocerts = NULL;
+	X509 *x = NULL;
 	/* Check for NULL PKCS12 structure */
 
-	if(!p12) {
+	if(!p12)
+		{
 		PKCS12err(PKCS12_F_PKCS12_PARSE,PKCS12_R_INVALID_NULL_PKCS12_POINTER);
 		return 0;
-	}
-
-	/* Allocate stack for ca certificates if needed */
-	if ((ca != NULL) && (*ca == NULL)) {
-		if (!(*ca = sk_X509_new_null())) {
-			PKCS12err(PKCS12_F_PKCS12_PARSE,ERR_R_MALLOC_FAILURE);
-			return 0;
 		}
-	}
 
-	if(pkey) *pkey = NULL;
-	if(cert) *cert = NULL;
+	if(pkey)
+		*pkey = NULL;
+	if(cert)
+		*cert = NULL;
 
 	/* Check the mac */
 
@@ -122,19 +115,61 @@ int PKCS12_parse(PKCS12 *p12, const char *pass, EVP_PKEY **pkey, X509 **cert,
 		goto err;
 	}
 
-	if (!parse_pk12 (p12, pass, -1, pkey, cert, ca))
+	/* Allocate stack for other certificates */
+	ocerts = sk_X509_new_null();
+
+	if (!ocerts)
+		{
+		PKCS12err(PKCS12_F_PKCS12_PARSE,ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+
+	if (!parse_pk12 (p12, pass, -1, pkey, ocerts))
 		{
 		PKCS12err(PKCS12_F_PKCS12_PARSE,PKCS12_R_PARSE_ERROR);
 		goto err;
 		}
 
+	while ((x = sk_X509_pop(ocerts)))
+		{
+		if (pkey && *pkey && cert && !*cert)
+			{
+			if (X509_check_private_key(x, *pkey))
+				{
+				*cert = x;
+				x = NULL;
+				}
+			}
+
+		if (ca && x)
+			{
+			if (!*ca)
+				*ca = sk_X509_new_null();
+			if (!*ca)
+				goto err;
+			if (!sk_X509_push(*ca, x))
+				goto err;
+			x = NULL;
+			}
+		if (x)
+			X509_free(x);
+		}
+
+	if (ocerts)
+		sk_X509_pop_free(ocerts, X509_free);
+
 	return 1;
 
  err:
 
-	if (pkey && *pkey) EVP_PKEY_free(*pkey);
-	if (cert && *cert) X509_free(*cert);
-	if (ca) sk_X509_pop_free(*ca, X509_free);
+	if (pkey && *pkey)
+		EVP_PKEY_free(*pkey);
+	if (cert && *cert)
+		X509_free(*cert);
+	if (x)
+		X509_free(x);
+	if (ocerts)
+		sk_X509_pop_free(ocerts, X509_free);
 	return 0;
 
 }
@@ -142,15 +177,13 @@ int PKCS12_parse(PKCS12 *p12, const char *pass, EVP_PKEY **pkey, X509 **cert,
 /* Parse the outer PKCS#12 structure */
 
 static int parse_pk12(PKCS12 *p12, const char *pass, int passlen,
-	     EVP_PKEY **pkey, X509 **cert, STACK_OF(X509) **ca)
+	     EVP_PKEY **pkey, STACK_OF(X509) *ocerts)
 {
 	STACK_OF(PKCS7) *asafes;
 	STACK_OF(PKCS12_SAFEBAG) *bags;
 	int i, bagnid;
 	PKCS7 *p7;
-	ASN1_OCTET_STRING *keyid = NULL;
 
-	char keymatch = 0;
 	if (!(asafes = PKCS12_unpack_authsafes (p12))) return 0;
 	for (i = 0; i < sk_PKCS7_num (asafes); i++) {
 		p7 = sk_PKCS7_value (asafes, i);
@@ -164,8 +197,7 @@ static int parse_pk12(PKCS12 *p12, const char *pass, int passlen,
 			sk_PKCS7_pop_free(asafes, PKCS7_free);
 			return 0;
 		}
-	    	if (!parse_bags(bags, pass, passlen, pkey, cert, ca,
-							 &keyid, &keymatch)) {
+	    	if (!parse_bags(bags, pass, passlen, pkey, ocerts)) {
 			sk_PKCS12_SAFEBAG_pop_free(bags, PKCS12_SAFEBAG_free);
 			sk_PKCS7_pop_free(asafes, PKCS7_free);
 			return 0;
@@ -173,89 +205,65 @@ static int parse_pk12(PKCS12 *p12, const char *pass, int passlen,
 		sk_PKCS12_SAFEBAG_pop_free(bags, PKCS12_SAFEBAG_free);
 	}
 	sk_PKCS7_pop_free(asafes, PKCS7_free);
-	if (keyid) M_ASN1_OCTET_STRING_free(keyid);
 	return 1;
 }
 
 
 static int parse_bags(STACK_OF(PKCS12_SAFEBAG) *bags, const char *pass,
-		      int passlen, EVP_PKEY **pkey, X509 **cert,
-		      STACK_OF(X509) **ca, ASN1_OCTET_STRING **keyid,
-		      char *keymatch)
+		      int passlen, EVP_PKEY **pkey, STACK_OF(X509) *ocerts)
 {
 	int i;
 	for (i = 0; i < sk_PKCS12_SAFEBAG_num(bags); i++) {
 		if (!parse_bag(sk_PKCS12_SAFEBAG_value (bags, i),
-			 pass, passlen, pkey, cert, ca, keyid,
-							 keymatch)) return 0;
+				 pass, passlen, pkey, ocerts))
+			return 0;
 	}
 	return 1;
 }
 
-#define MATCH_KEY  0x1
-#define MATCH_CERT 0x2
-#define MATCH_ALL  0x3
-
 static int parse_bag(PKCS12_SAFEBAG *bag, const char *pass, int passlen,
-		     EVP_PKEY **pkey, X509 **cert, STACK_OF(X509) **ca,
-		     ASN1_OCTET_STRING **keyid,
-		     char *keymatch)
+		     EVP_PKEY **pkey, STACK_OF(X509) *ocerts)
 {
 	PKCS8_PRIV_KEY_INFO *p8;
 	X509 *x509;
-	ASN1_OCTET_STRING *lkey = NULL, *ckid = NULL;
 	ASN1_TYPE *attrib;
 	ASN1_BMPSTRING *fname = NULL;
+	ASN1_OCTET_STRING *lkid = NULL;
 
 	if ((attrib = PKCS12_get_attr (bag, NID_friendlyName)))
 		fname = attrib->value.bmpstring;
 
-	if ((attrib = PKCS12_get_attr (bag, NID_localKeyID))) {
-		lkey = attrib->value.octet_string;
-		ckid = lkey;
-	}
+	if ((attrib = PKCS12_get_attr (bag, NID_localKeyID)))
+		lkid = attrib->value.octet_string;
 
-	/* Check for any local key id matching (if needed) */
-	if (lkey && ((*keymatch & MATCH_ALL) != MATCH_ALL)) {
-		if (*keyid) {
-			if (M_ASN1_OCTET_STRING_cmp(*keyid, lkey)) lkey = NULL;
-		} else {
-			if (!(*keyid = M_ASN1_OCTET_STRING_dup(lkey))) {
-				PKCS12err(PKCS12_F_PARSE_BAG,ERR_R_MALLOC_FAILURE);
-				return 0;
-		    }
-		}
-	}
-	
 	switch (M_PKCS12_bag_type(bag))
 	{
 	case NID_keyBag:
-		if (!lkey || !pkey) return 1;	
-		if (!(*pkey = EVP_PKCS82PKEY(bag->value.keybag))) return 0;
-		*keymatch |= MATCH_KEY;
+		if (!pkey || *pkey)
+			return 1;	
+		if (!(*pkey = EVP_PKCS82PKEY(bag->value.keybag)))
+			return 0;
 	break;
 
 	case NID_pkcs8ShroudedKeyBag:
-		if (!lkey || !pkey) return 1;	
+		if (!pkey || *pkey)
+			return 1;	
 		if (!(p8 = PKCS12_decrypt_skey(bag, pass, passlen)))
 				return 0;
 		*pkey = EVP_PKCS82PKEY(p8);
 		PKCS8_PRIV_KEY_INFO_free(p8);
 		if (!(*pkey)) return 0;
-		*keymatch |= MATCH_KEY;
 	break;
 
 	case NID_certBag:
 		if (M_PKCS12_cert_bag_type(bag) != NID_x509Certificate )
-								 return 1;
-		if (!(x509 = PKCS12_certbag2x509(bag))) return 0;
-		if(ckid)
+			return 1;
+		if (!(x509 = PKCS12_certbag2x509(bag)))
+			return 0;
+		if(lkid && !X509_keyid_set1(x509, lkid->data, lkid->length))
 			{
-			if (!X509_keyid_set1(x509, ckid->data, ckid->length))
-				{
-				X509_free(x509);
-				return 0;
-				}
+			X509_free(x509);
+			return 0;
 			}
 		if(fname) {
 			int len, r;
@@ -272,20 +280,17 @@ static int parse_bag(PKCS12_SAFEBAG *bag, const char *pass, int passlen,
 			}
 		}
 
+		if(!sk_X509_push(ocerts, x509))
+			{
+			X509_free(x509);
+			return 0;
+			}
 
-		if (lkey) {
-			*keymatch |= MATCH_CERT;
-			if (cert) *cert = x509;
-			else X509_free(x509);
-		} else {
-			if(ca) sk_X509_push (*ca, x509);
-			else X509_free(x509);
-		}
 	break;
 
 	case NID_safeContentsBag:
 		return parse_bags(bag->value.safes, pass, passlen,
-			 		pkey, cert, ca, keyid, keymatch);
+			 		pkey, ocerts);
 	break;
 
 	default:
diff --git a/crypto/pkcs12/p12_mutl.c b/crypto/pkcs12/p12_mutl.c
index 70bfef6e5d16..96de1bd11e7c 100644
--- a/crypto/pkcs12/p12_mutl.c
+++ b/crypto/pkcs12/p12_mutl.c
@@ -71,6 +71,7 @@ int PKCS12_gen_mac(PKCS12 *p12, const char *pass, int passlen,
 	HMAC_CTX hmac;
 	unsigned char key[EVP_MAX_MD_SIZE], *salt;
 	int saltlen, iter;
+	int md_size;
 
 	if (!PKCS7_type_is_data(p12->authsafes))
 		{
@@ -87,16 +88,23 @@ int PKCS12_gen_mac(PKCS12 *p12, const char *pass, int passlen,
 		PKCS12err(PKCS12_F_PKCS12_GEN_MAC,PKCS12_R_UNKNOWN_DIGEST_ALGORITHM);
 		return 0;
 	}
+	md_size = EVP_MD_size(md_type);
+	if (md_size < 0)
+	    return 0;
 	if(!PKCS12_key_gen (pass, passlen, salt, saltlen, PKCS12_MAC_ID, iter,
-				 EVP_MD_size(md_type), key, md_type)) {
+				 md_size, key, md_type)) {
 		PKCS12err(PKCS12_F_PKCS12_GEN_MAC,PKCS12_R_KEY_GEN_ERROR);
 		return 0;
 	}
 	HMAC_CTX_init(&hmac);
-	HMAC_Init_ex(&hmac, key, EVP_MD_size(md_type), md_type, NULL);
-    	HMAC_Update(&hmac, p12->authsafes->d.data->data,
-					 p12->authsafes->d.data->length);
-    	HMAC_Final(&hmac, mac, maclen);
+	if (!HMAC_Init_ex(&hmac, key, md_size, md_type, NULL)
+    		|| !HMAC_Update(&hmac, p12->authsafes->d.data->data,
+					 p12->authsafes->d.data->length)
+    		|| !HMAC_Final(&hmac, mac, maclen))
+		{
+    		HMAC_CTX_cleanup(&hmac);
+		return 0;
+		}
     	HMAC_CTX_cleanup(&hmac);
 	return 1;
 }
diff --git a/crypto/pkcs12/p12_utl.c b/crypto/pkcs12/p12_utl.c
index 2edbf905c58a..59c6f453f6e5 100644
--- a/crypto/pkcs12/p12_utl.c
+++ b/crypto/pkcs12/p12_utl.c
@@ -60,15 +60,9 @@
 #include "cryptlib.h"
 #include <openssl/pkcs12.h>
 
-#ifdef OPENSSL_SYS_NETWARE
-/* Rename these functions to avoid name clashes on NetWare OS */
-#define uni2asc OPENSSL_uni2asc
-#define asc2uni OPENSSL_asc2uni
-#endif
-
 /* Cheap and nasty Unicode stuff */
 
-unsigned char *asc2uni(const char *asc, int asclen, unsigned char **uni, int *unilen)
+unsigned char *OPENSSL_asc2uni(const char *asc, int asclen, unsigned char **uni, int *unilen)
 {
 	int ulen, i;
 	unsigned char *unitmp;
@@ -87,7 +81,7 @@ unsigned char *asc2uni(const char *asc, int asclen, unsigned char **uni, int *un
 	return unitmp;
 }
 
-char *uni2asc(unsigned char *uni, int unilen)
+char *OPENSSL_uni2asc(unsigned char *uni, int unilen)
 {
 	int asclen, i;
 	char *asctmp;
diff --git a/crypto/pkcs12/pk12err.c b/crypto/pkcs12/pk12err.c
index 07a1fb6907b1..f6ddf2df12eb 100644
--- a/crypto/pkcs12/pk12err.c
+++ b/crypto/pkcs12/pk12err.c
@@ -1,6 +1,6 @@
 /* crypto/pkcs12/pk12err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
diff --git a/crypto/pkcs12/pkcs12.h b/crypto/pkcs12/pkcs12.h
index 78317fb1fa56..b17eb9f42b8e 100644
--- a/crypto/pkcs12/pkcs12.h
+++ b/crypto/pkcs12/pkcs12.h
@@ -108,8 +108,6 @@ PKCS12_MAC_DATA *mac;
 PKCS7 *authsafes;
 } PKCS12;
 
-PREDECLARE_STACK_OF(PKCS12_SAFEBAG)
-
 typedef struct {
 ASN1_OBJECT *type;
 union {
@@ -232,14 +230,9 @@ int PKCS12_set_mac(PKCS12 *p12, const char *pass, int passlen,
 		   const EVP_MD *md_type);
 int PKCS12_setup_mac(PKCS12 *p12, int iter, unsigned char *salt,
 					 int saltlen, const EVP_MD *md_type);
-#if defined(NETWARE) || defined(OPENSSL_SYS_NETWARE)
-/* Rename these functions to avoid name clashes on NetWare OS */
 unsigned char *OPENSSL_asc2uni(const char *asc, int asclen, unsigned char **uni, int *unilen);
 char *OPENSSL_uni2asc(unsigned char *uni, int unilen);
-#else
-unsigned char *asc2uni(const char *asc, int asclen, unsigned char **uni, int *unilen);
-char *uni2asc(unsigned char *uni, int unilen);
-#endif
+
 DECLARE_ASN1_FUNCTIONS(PKCS12)
 DECLARE_ASN1_FUNCTIONS(PKCS12_MAC_DATA)
 DECLARE_ASN1_FUNCTIONS(PKCS12_SAFEBAG)
diff --git a/crypto/pkcs7/Makefile b/crypto/pkcs7/Makefile
index 790d8edf36e8..56dc6823d123 100644
--- a/crypto/pkcs7/Makefile
+++ b/crypto/pkcs7/Makefile
@@ -21,9 +21,9 @@ APPS=
 
 LIB=$(TOP)/libcrypto.a
 LIBSRC=	pk7_asn1.c pk7_lib.c pkcs7err.c pk7_doit.c pk7_smime.c pk7_attr.c \
-	pk7_mime.c
+	pk7_mime.c bio_pk7.c
 LIBOBJ= pk7_asn1.o pk7_lib.o pkcs7err.o pk7_doit.o pk7_smime.o pk7_attr.o \
-	pk7_mime.o
+	pk7_mime.o bio_pk7.o
 
 SRC= $(LIBSRC)
 
@@ -54,7 +54,7 @@ verify: verify.o example.o lib
 	$(CC) $(CFLAGS) -o verify verify.o $(PEX_LIBS) example.o $(LIB) $(EX_LIBS)
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -95,26 +95,31 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
+bio_pk7.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+bio_pk7.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+bio_pk7.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+bio_pk7.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+bio_pk7.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+bio_pk7.o: ../../include/openssl/symhacks.h bio_pk7.c
 pk7_asn1.o: ../../e_os.h ../../include/openssl/asn1.h
 pk7_asn1.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 pk7_asn1.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 pk7_asn1.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pk7_asn1.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pk7_asn1.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pk7_asn1.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pk7_asn1.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pk7_asn1.o: ../../include/openssl/opensslconf.h
+pk7_asn1.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pk7_asn1.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 pk7_asn1.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pk7_asn1.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 pk7_asn1.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 pk7_asn1.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 pk7_asn1.o: ../../include/openssl/x509_vfy.h ../cryptlib.h pk7_asn1.c
-pk7_attr.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
-pk7_attr.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-pk7_attr.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
-pk7_attr.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-pk7_attr.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pk7_attr.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+pk7_attr.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
+pk7_attr.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+pk7_attr.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+pk7_attr.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+pk7_attr.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+pk7_attr.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
 pk7_attr.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
 pk7_attr.o: ../../include/openssl/opensslconf.h
 pk7_attr.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
@@ -129,9 +134,8 @@ pk7_doit.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 pk7_doit.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pk7_doit.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pk7_doit.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pk7_doit.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pk7_doit.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pk7_doit.o: ../../include/openssl/opensslconf.h
+pk7_doit.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pk7_doit.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 pk7_doit.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pk7_doit.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
 pk7_doit.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -143,22 +147,22 @@ pk7_lib.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pk7_lib.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pk7_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pk7_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pk7_lib.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-pk7_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-pk7_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-pk7_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-pk7_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-pk7_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-pk7_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-pk7_lib.o: ../../include/openssl/x509_vfy.h ../cryptlib.h pk7_lib.c
+pk7_lib.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pk7_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pk7_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+pk7_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+pk7_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+pk7_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+pk7_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+pk7_lib.o: ../asn1/asn1_locl.h ../cryptlib.h pk7_lib.c
 pk7_mime.o: ../../e_os.h ../../include/openssl/asn1.h
 pk7_mime.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pk7_mime.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pk7_mime.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pk7_mime.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-pk7_mime.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-pk7_mime.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-pk7_mime.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pk7_mime.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+pk7_mime.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pk7_mime.o: ../../include/openssl/opensslconf.h
 pk7_mime.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pk7_mime.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
 pk7_mime.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -171,8 +175,8 @@ pk7_smime.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 pk7_smime.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pk7_smime.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pk7_smime.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pk7_smime.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pk7_smime.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pk7_smime.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pk7_smime.o: ../../include/openssl/objects.h
 pk7_smime.o: ../../include/openssl/opensslconf.h
 pk7_smime.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pk7_smime.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
diff --git a/crypto/rc4/rc4_fblk.c b/crypto/pkcs7/bio_pk7.c
index 1b2a42979ba5..c8d06d6cdc83 100644
--- a/crypto/rc4/rc4_fblk.c
+++ b/crypto/pkcs7/bio_pk7.c
@@ -1,4 +1,4 @@
-/* crypto/rc4/rc4_fblk.c */
+/* bio_pk7.c */
 /* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
  * project.
  */
@@ -49,27 +49,21 @@
  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
  * OF THE POSSIBILITY OF SUCH DAMAGE.
  * ====================================================================
+ *
  */
 
+#include <openssl/asn1.h>
+#include <openssl/pkcs7.h>
+#include <openssl/bio.h>
 
-#include <openssl/rc4.h>
-#include "rc4_locl.h"
-#include <openssl/opensslv.h>
-#include <openssl/crypto.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
+#ifndef OPENSSL_SYSNAME_NETWARE
+#include <memory.h>
 #endif
+#include <stdio.h>
 
-/* FIPS mode blocking for RC4 has to be done separately since RC4_set_key
- * may be implemented in an assembly language file.
- */
+/* Streaming encode support for PKCS#7 */
 
-#ifdef OPENSSL_FIPS
-void RC4_set_key(RC4_KEY *key, int len, const unsigned char *data)
+BIO *BIO_new_PKCS7(BIO *out, PKCS7 *p7) 
 	{
-	if (FIPS_mode())
-		FIPS_BAD_ABORT(RC4)
-	private_RC4_set_key(key, len, data);
+	return BIO_new_NDEF(out, (ASN1_VALUE *)p7, ASN1_ITEM_rptr(PKCS7));
 	}
-#endif
-
diff --git a/crypto/pkcs7/pk7_asn1.c b/crypto/pkcs7/pk7_asn1.c
index 1f70d3138677..b7ec2883cb32 100644
--- a/crypto/pkcs7/pk7_asn1.c
+++ b/crypto/pkcs7/pk7_asn1.c
@@ -77,10 +77,39 @@ ASN1_ADB(PKCS7) = {
 	ADB_ENTRY(NID_pkcs7_encrypted, ASN1_NDEF_EXP_OPT(PKCS7, d.encrypted, PKCS7_ENCRYPT, 0))
 } ASN1_ADB_END(PKCS7, 0, type, 0, &p7default_tt, NULL);
 
-ASN1_NDEF_SEQUENCE(PKCS7) = {
+/* PKCS#7 streaming support */
+static int pk7_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+							void *exarg)
+{
+	ASN1_STREAM_ARG *sarg = exarg;
+	PKCS7 **pp7 = (PKCS7 **)pval;
+
+	switch(operation)
+		{
+
+		case ASN1_OP_STREAM_PRE:
+		if (PKCS7_stream(&sarg->boundary, *pp7) <= 0)
+			return 0;
+		case ASN1_OP_DETACHED_PRE:
+		sarg->ndef_bio = PKCS7_dataInit(*pp7, sarg->out);
+		if (!sarg->ndef_bio)
+			return 0;
+		break;
+
+		case ASN1_OP_STREAM_POST:
+		case ASN1_OP_DETACHED_POST:
+		if (PKCS7_dataFinal(*pp7, sarg->ndef_bio) <= 0)
+			return 0;
+		break;
+
+		}
+	return 1;
+}
+
+ASN1_NDEF_SEQUENCE_cb(PKCS7, pk7_cb) = {
 	ASN1_SIMPLE(PKCS7, type, ASN1_OBJECT),
 	ASN1_ADB_OBJECT(PKCS7)
-}ASN1_NDEF_SEQUENCE_END(PKCS7)
+}ASN1_NDEF_SEQUENCE_END_cb(PKCS7, PKCS7)
 
 IMPLEMENT_ASN1_FUNCTIONS(PKCS7)
 IMPLEMENT_ASN1_NDEF_FUNCTION(PKCS7)
@@ -98,7 +127,8 @@ ASN1_NDEF_SEQUENCE(PKCS7_SIGNED) = {
 IMPLEMENT_ASN1_FUNCTIONS(PKCS7_SIGNED)
 
 /* Minor tweak to operation: free up EVP_PKEY */
-static int si_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int si_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+							void *exarg)
 {
 	if(operation == ASN1_OP_FREE_POST) {
 		PKCS7_SIGNER_INFO *si = (PKCS7_SIGNER_INFO *)*pval;
@@ -140,7 +170,8 @@ ASN1_NDEF_SEQUENCE(PKCS7_ENVELOPE) = {
 IMPLEMENT_ASN1_FUNCTIONS(PKCS7_ENVELOPE)
 
 /* Minor tweak to operation: free up X509 */
-static int ri_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int ri_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+								void *exarg)
 {
 	if(operation == ASN1_OP_FREE_POST) {
 		PKCS7_RECIP_INFO *ri = (PKCS7_RECIP_INFO *)*pval;
@@ -161,7 +192,7 @@ IMPLEMENT_ASN1_FUNCTIONS(PKCS7_RECIP_INFO)
 ASN1_NDEF_SEQUENCE(PKCS7_ENC_CONTENT) = {
 	ASN1_SIMPLE(PKCS7_ENC_CONTENT, content_type, ASN1_OBJECT),
 	ASN1_SIMPLE(PKCS7_ENC_CONTENT, algorithm, X509_ALGOR),
-	ASN1_IMP_OPT(PKCS7_ENC_CONTENT, enc_data, ASN1_OCTET_STRING, 0)
+	ASN1_IMP_OPT(PKCS7_ENC_CONTENT, enc_data, ASN1_OCTET_STRING_NDEF, 0)
 } ASN1_NDEF_SEQUENCE_END(PKCS7_ENC_CONTENT)
 
 IMPLEMENT_ASN1_FUNCTIONS(PKCS7_ENC_CONTENT)
@@ -212,3 +243,5 @@ ASN1_ITEM_TEMPLATE(PKCS7_ATTR_VERIFY) =
 	ASN1_EX_TEMPLATE_TYPE(ASN1_TFLG_SEQUENCE_OF | ASN1_TFLG_IMPTAG | ASN1_TFLG_UNIVERSAL,
 				V_ASN1_SET, PKCS7_ATTRIBUTES, X509_ATTRIBUTE)
 ASN1_ITEM_TEMPLATE_END(PKCS7_ATTR_VERIFY)
+
+IMPLEMENT_ASN1_PRINT_FUNCTION(PKCS7)
diff --git a/crypto/pkcs7/pk7_attr.c b/crypto/pkcs7/pk7_attr.c
index d54971716913..a97db512107f 100644
--- a/crypto/pkcs7/pk7_attr.c
+++ b/crypto/pkcs7/pk7_attr.c
@@ -60,6 +60,7 @@
 #include <stdlib.h>
 #include <openssl/bio.h>
 #include <openssl/asn1.h>
+#include <openssl/asn1t.h>
 #include <openssl/pem.h>
 #include <openssl/pkcs7.h>
 #include <openssl/x509.h>
@@ -68,27 +69,12 @@
 int PKCS7_add_attrib_smimecap(PKCS7_SIGNER_INFO *si, STACK_OF(X509_ALGOR) *cap)
 {
 	ASN1_STRING *seq;
-	unsigned char *p, *pp;
-	int len;
-	len=i2d_ASN1_SET_OF_X509_ALGOR(cap,NULL,i2d_X509_ALGOR,
-				       V_ASN1_SEQUENCE,V_ASN1_UNIVERSAL,
-				       IS_SEQUENCE);
-	if(!(pp=(unsigned char *)OPENSSL_malloc(len))) {
-		PKCS7err(PKCS7_F_PKCS7_ADD_ATTRIB_SMIMECAP,ERR_R_MALLOC_FAILURE);
-		return 0;
-	}
-	p=pp;
-	i2d_ASN1_SET_OF_X509_ALGOR(cap,&p,i2d_X509_ALGOR, V_ASN1_SEQUENCE,
-				   V_ASN1_UNIVERSAL, IS_SEQUENCE);
 	if(!(seq = ASN1_STRING_new())) {
 		PKCS7err(PKCS7_F_PKCS7_ADD_ATTRIB_SMIMECAP,ERR_R_MALLOC_FAILURE);
 		return 0;
 	}
-	if(!ASN1_STRING_set (seq, pp, len)) {
-		PKCS7err(PKCS7_F_PKCS7_ADD_ATTRIB_SMIMECAP,ERR_R_MALLOC_FAILURE);
-		return 0;
-	}
-	OPENSSL_free (pp);
+	seq->length = ASN1_item_i2d((ASN1_VALUE *)cap,&seq->data,
+				ASN1_ITEM_rptr(X509_ALGORS));
         return PKCS7_add_signed_attribute(si, NID_SMIMECapabilities,
 							V_ASN1_SEQUENCE, seq);
 }
@@ -102,10 +88,9 @@ STACK_OF(X509_ALGOR) *PKCS7_get_smimecap(PKCS7_SIGNER_INFO *si)
 	if (!cap || (cap->type != V_ASN1_SEQUENCE))
 		return NULL;
 	p = cap->value.sequence->data;
-	return d2i_ASN1_SET_OF_X509_ALGOR(NULL, &p,
-					  cap->value.sequence->length,
-					  d2i_X509_ALGOR, X509_ALGOR_free,
-					  V_ASN1_SEQUENCE, V_ASN1_UNIVERSAL);
+	return (STACK_OF(X509_ALGOR) *)
+		ASN1_item_d2i(NULL, &p, cap->value.sequence->length,
+				ASN1_ITEM_rptr(X509_ALGORS));
 	}
 
 /* Basic smime-capabilities OID and optional integer arg */
@@ -139,3 +124,42 @@ int PKCS7_simple_smimecap(STACK_OF(X509_ALGOR) *sk, int nid, int arg)
 	sk_X509_ALGOR_push (sk, alg);
 	return 1;
 }
+
+int PKCS7_add_attrib_content_type(PKCS7_SIGNER_INFO *si, ASN1_OBJECT *coid)
+	{
+	if (PKCS7_get_signed_attribute(si, NID_pkcs9_contentType))
+		return 0;
+	if (!coid)
+		coid = OBJ_nid2obj(NID_pkcs7_data);
+	return PKCS7_add_signed_attribute(si, NID_pkcs9_contentType,
+				V_ASN1_OBJECT, coid);
+	}
+
+int PKCS7_add0_attrib_signing_time(PKCS7_SIGNER_INFO *si, ASN1_TIME *t)
+	{
+	if (!t && !(t=X509_gmtime_adj(NULL,0)))
+		{
+		PKCS7err(PKCS7_F_PKCS7_ADD0_ATTRIB_SIGNING_TIME,
+				ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	return PKCS7_add_signed_attribute(si, NID_pkcs9_signingTime,
+						V_ASN1_UTCTIME, t);
+	}
+
+int PKCS7_add1_attrib_digest(PKCS7_SIGNER_INFO *si,
+				const unsigned char *md, int mdlen)
+	{
+	ASN1_OCTET_STRING *os;
+	os = ASN1_OCTET_STRING_new();
+	if (!os)
+		return 0;
+	if (!ASN1_STRING_set(os, md, mdlen)
+		|| !PKCS7_add_signed_attribute(si, NID_pkcs9_messageDigest,
+						V_ASN1_OCTET_STRING, os))
+		{
+		ASN1_OCTET_STRING_free(os);
+		return 0;
+		}
+	return 1;
+	}
diff --git a/crypto/pkcs7/pk7_doit.c b/crypto/pkcs7/pk7_doit.c
index 8b3024e77407..77fda3b82a07 100644
--- a/crypto/pkcs7/pk7_doit.c
+++ b/crypto/pkcs7/pk7_doit.c
@@ -138,6 +138,128 @@ static int PKCS7_bio_add_digest(BIO **pbio, X509_ALGOR *alg)
 
 	}
 
+static int pkcs7_encode_rinfo(PKCS7_RECIP_INFO *ri,
+					unsigned char *key, int keylen)
+	{
+	EVP_PKEY_CTX *pctx = NULL;
+	EVP_PKEY *pkey = NULL;
+	unsigned char *ek = NULL;
+	int ret = 0;
+	size_t eklen;
+
+	pkey = X509_get_pubkey(ri->cert);
+
+	if (!pkey)
+		return 0;
+
+	pctx = EVP_PKEY_CTX_new(pkey, NULL);
+	if (!pctx)
+		return 0;
+
+	if (EVP_PKEY_encrypt_init(pctx) <= 0)
+		goto err;
+
+	if (EVP_PKEY_CTX_ctrl(pctx, -1, EVP_PKEY_OP_ENCRYPT,
+				EVP_PKEY_CTRL_PKCS7_ENCRYPT, 0, ri) <= 0)
+		{
+		PKCS7err(PKCS7_F_PKCS7_ENCODE_RINFO, PKCS7_R_CTRL_ERROR);
+		goto err;
+		}
+
+	if (EVP_PKEY_encrypt(pctx, NULL, &eklen, key, keylen) <= 0)
+		goto err;
+
+	ek = OPENSSL_malloc(eklen);
+
+	if (ek == NULL)
+		{
+		PKCS7err(PKCS7_F_PKCS7_ENCODE_RINFO, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	if (EVP_PKEY_encrypt(pctx, ek, &eklen, key, keylen) <= 0)
+		goto err;
+
+	ASN1_STRING_set0(ri->enc_key, ek, eklen);
+	ek = NULL;
+
+	ret = 1;
+
+	err:
+	if (pkey)
+		EVP_PKEY_free(pkey);
+	if (pctx)
+		EVP_PKEY_CTX_free(pctx);
+	if (ek)
+		OPENSSL_free(ek);
+	return ret;
+
+	}
+
+
+static int pkcs7_decrypt_rinfo(unsigned char **pek, int *peklen,
+			       PKCS7_RECIP_INFO *ri, EVP_PKEY *pkey)
+	{
+	EVP_PKEY_CTX *pctx = NULL;
+	unsigned char *ek = NULL;
+	size_t eklen;
+
+	int ret = -1;
+
+	pctx = EVP_PKEY_CTX_new(pkey, NULL);
+	if (!pctx)
+		return -1;
+
+	if (EVP_PKEY_decrypt_init(pctx) <= 0)
+		goto err;
+
+	if (EVP_PKEY_CTX_ctrl(pctx, -1, EVP_PKEY_OP_DECRYPT,
+				EVP_PKEY_CTRL_PKCS7_DECRYPT, 0, ri) <= 0)
+		{
+		PKCS7err(PKCS7_F_PKCS7_DECRYPT_RINFO, PKCS7_R_CTRL_ERROR);
+		goto err;
+		}
+
+	if (EVP_PKEY_decrypt(pctx, NULL, &eklen,
+				ri->enc_key->data, ri->enc_key->length) <= 0)
+		goto err;
+
+	ek = OPENSSL_malloc(eklen);
+
+	if (ek == NULL)
+		{
+		PKCS7err(PKCS7_F_PKCS7_DECRYPT_RINFO, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	if (EVP_PKEY_decrypt(pctx, ek, &eklen,
+				ri->enc_key->data, ri->enc_key->length) <= 0)
+		{
+		ret = 0;
+		PKCS7err(PKCS7_F_PKCS7_DECRYPT_RINFO, ERR_R_EVP_LIB);
+		goto err;
+		}
+
+	ret = 1;
+
+	if (*pek)
+		{
+		OPENSSL_cleanse(*pek, *peklen);
+		OPENSSL_free(*pek);
+		}
+
+	*pek = ek;
+	*peklen = eklen;
+
+	err:
+	if (pctx)
+		EVP_PKEY_CTX_free(pctx);
+	if (!ret && ek)
+		OPENSSL_free(ek);
+
+	return ret;
+	}
+
 BIO *PKCS7_dataInit(PKCS7 *p7, BIO *bio)
 	{
 	int i;
@@ -148,7 +270,6 @@ BIO *PKCS7_dataInit(PKCS7 *p7, BIO *bio)
 	STACK_OF(PKCS7_RECIP_INFO) *rsk=NULL;
 	X509_ALGOR *xalg=NULL;
 	PKCS7_RECIP_INFO *ri=NULL;
-	EVP_PKEY *pkey;
 	ASN1_OCTET_STRING *os=NULL;
 
 	i=OBJ_obj2nid(p7->type);
@@ -187,6 +308,8 @@ BIO *PKCS7_dataInit(PKCS7 *p7, BIO *bio)
 		xa = p7->d.digest->md;
 		os = PKCS7_get_octet_string(p7->d.digest->contents);
 		break;
+	case NID_pkcs7_data:
+		break;
 	default:
 		PKCS7err(PKCS7_F_PKCS7_DATAINIT,PKCS7_R_UNSUPPORTED_CONTENT_TYPE);
 	        goto err;
@@ -204,8 +327,6 @@ BIO *PKCS7_dataInit(PKCS7 *p7, BIO *bio)
 		unsigned char key[EVP_MAX_KEY_LENGTH];
 		unsigned char iv[EVP_MAX_IV_LENGTH];
 		int keylen,ivlen;
-		int jj,max;
-		unsigned char *tmp;
 		EVP_CIPHER_CTX *ctx;
 
 		if ((btmp=BIO_new(BIO_f_cipher())) == NULL)
@@ -234,52 +355,16 @@ BIO *PKCS7_dataInit(PKCS7 *p7, BIO *bio)
 					goto err;
 			}
 			if(EVP_CIPHER_param_to_asn1(ctx, xalg->parameter) < 0)
-			       goto err;
+				goto err;
 		}
 
 		/* Lets do the pub key stuff :-) */
-		max=0;
 		for (i=0; i<sk_PKCS7_RECIP_INFO_num(rsk); i++)
 			{
 			ri=sk_PKCS7_RECIP_INFO_value(rsk,i);
-			if (ri->cert == NULL)
-				{
-				PKCS7err(PKCS7_F_PKCS7_DATAINIT,PKCS7_R_MISSING_CERIPEND_INFO);
+			if (pkcs7_encode_rinfo(ri, key, keylen) <= 0)
 				goto err;
-				}
-			if ((pkey=X509_get_pubkey(ri->cert)) == NULL)
-				goto err;
-			jj=EVP_PKEY_size(pkey);
-			EVP_PKEY_free(pkey);
-			if (max < jj) max=jj;
-			}
-		if ((tmp=(unsigned char *)OPENSSL_malloc(max)) == NULL)
-			{
-			PKCS7err(PKCS7_F_PKCS7_DATAINIT,ERR_R_MALLOC_FAILURE);
-			goto err;
 			}
-		for (i=0; i<sk_PKCS7_RECIP_INFO_num(rsk); i++)
-			{
-			ri=sk_PKCS7_RECIP_INFO_value(rsk,i);
-			if ((pkey=X509_get_pubkey(ri->cert)) == NULL)
-				goto err;
-			jj=EVP_PKEY_encrypt(tmp,key,keylen,pkey);
-			EVP_PKEY_free(pkey);
-			if (jj <= 0)
-				{
-				PKCS7err(PKCS7_F_PKCS7_DATAINIT,ERR_R_EVP_LIB);
-				OPENSSL_free(tmp);
-				goto err;
-				}
-			if (!M_ASN1_OCTET_STRING_set(ri->enc_key,tmp,jj))
-				{
-				PKCS7err(PKCS7_F_PKCS7_DATAINIT,
-					ERR_R_MALLOC_FAILURE);
-				OPENSSL_free(tmp);
-				goto err;
-				}
-			}
-		OPENSSL_free(tmp);
 		OPENSSL_cleanse(key, keylen);
 
 		if (out == NULL)
@@ -303,7 +388,10 @@ BIO *PKCS7_dataInit(PKCS7 *p7, BIO *bio)
 			BIO_set_mem_eof_return(bio,0);
 			}
 		}
-	BIO_push(out,bio);
+	if (out)
+		BIO_push(out,bio);
+	else
+		out = bio;
 	bio=NULL;
 	if (0)
 		{
@@ -333,7 +421,6 @@ BIO *PKCS7_dataDecode(PKCS7 *p7, EVP_PKEY *pkey, BIO *in_bio, X509 *pcert)
 	{
 	int i,j;
 	BIO *out=NULL,*btmp=NULL,*etmp=NULL,*bio=NULL;
-	unsigned char *tmp=NULL;
 	X509_ALGOR *xa;
 	ASN1_OCTET_STRING *data_body=NULL;
 	const EVP_MD *evp_md;
@@ -343,6 +430,8 @@ BIO *PKCS7_dataDecode(PKCS7 *p7, EVP_PKEY *pkey, BIO *in_bio, X509 *pcert)
 	STACK_OF(X509_ALGOR) *md_sk=NULL;
 	STACK_OF(PKCS7_RECIP_INFO) *rsk=NULL;
 	PKCS7_RECIP_INFO *ri=NULL;
+       unsigned char *ek = NULL, *tkey = NULL;
+       int eklen = 0, tkeylen = 0;
 
 	i=OBJ_obj2nid(p7->type);
 	p7->state=PKCS7_S_HEADER;
@@ -420,9 +509,6 @@ BIO *PKCS7_dataDecode(PKCS7 *p7, EVP_PKEY *pkey, BIO *in_bio, X509 *pcert)
 		int max;
 		X509_OBJECT ret;
 #endif
-		unsigned char *tkey = NULL;
-		int tkeylen;
-		int jj;
 
 		if ((etmp=BIO_new(BIO_f_cipher())) == NULL)
 			{
@@ -437,67 +523,45 @@ BIO *PKCS7_dataDecode(PKCS7 *p7, EVP_PKEY *pkey, BIO *in_bio, X509 *pcert)
 		 * (if any)
 		 */
 
-		if (pcert) {
-			for (i=0; i<sk_PKCS7_RECIP_INFO_num(rsk); i++) {
+		if (pcert)
+			{
+			for (i=0; i<sk_PKCS7_RECIP_INFO_num(rsk); i++)
+				{
 				ri=sk_PKCS7_RECIP_INFO_value(rsk,i);
 				if (!pkcs7_cmp_ri(ri, pcert))
 					break;
 				ri=NULL;
-			}
-			if (ri == NULL) {
+				}
+			if (ri == NULL)
+				{
 				PKCS7err(PKCS7_F_PKCS7_DATADECODE,
 				      PKCS7_R_NO_RECIPIENT_MATCHES_CERTIFICATE);
 				goto err;
-			}
-		}
-
-		jj=EVP_PKEY_size(pkey);
-		tmp=(unsigned char *)OPENSSL_malloc(jj+10);
-		if (tmp == NULL)
-			{
-			PKCS7err(PKCS7_F_PKCS7_DATADECODE,ERR_R_MALLOC_FAILURE);
-			goto err;
+				}
 			}
 
 		/* If we haven't got a certificate try each ri in turn */
-
 		if (pcert == NULL)
 			{
-			/* Temporary storage in case EVP_PKEY_decrypt
-			 * overwrites output buffer on error.
-			 */
-			unsigned char *tmp2;
-			tmp2 = OPENSSL_malloc(jj);
-			if (!tmp2)
-				goto err;
-			jj = -1;
-			/* Always attempt to decrypt all cases to avoid
-			 * leaking timing information about a successful
-			 * decrypt.
+			/* Always attempt to decrypt all rinfo even
+			 * after sucess as a defence against MMA timing
+			 * attacks.
 			 */
 			for (i=0; i<sk_PKCS7_RECIP_INFO_num(rsk); i++)
 				{
-				int tret;
 				ri=sk_PKCS7_RECIP_INFO_value(rsk,i);
-				tret=EVP_PKEY_decrypt(tmp2,
-					M_ASN1_STRING_data(ri->enc_key),
-					M_ASN1_STRING_length(ri->enc_key),
-						pkey);
-				if (tret > 0)
-					{
-					memcpy(tmp, tmp2, tret);
-					OPENSSL_cleanse(tmp2, tret);
-					jj = tret;
-					}
+				
+				if (pkcs7_decrypt_rinfo(&ek, &eklen,
+							ri, pkey) < 0)
+					goto err;
 				ERR_clear_error();
 				}
-			OPENSSL_free(tmp2);
 			}
 		else
 			{
-			jj=EVP_PKEY_decrypt(tmp,
-				M_ASN1_STRING_data(ri->enc_key),
-				M_ASN1_STRING_length(ri->enc_key), pkey);
+			/* Only exit on fatal errors, not decrypt failure */
+			if (pkcs7_decrypt_rinfo(&ek, &eklen, ri, pkey) < 0)
+				goto err;
 			ERR_clear_error();
 			}
 
@@ -507,47 +571,51 @@ BIO *PKCS7_dataDecode(PKCS7 *p7, EVP_PKEY *pkey, BIO *in_bio, X509 *pcert)
 			goto err;
 		if (EVP_CIPHER_asn1_to_param(evp_ctx,enc_alg->parameter) < 0)
 			goto err;
-		/* Generate random key to counter MMA */
+		/* Generate random key as MMA defence */
 		tkeylen = EVP_CIPHER_CTX_key_length(evp_ctx);
 		tkey = OPENSSL_malloc(tkeylen);
 		if (!tkey)
 			goto err;
 		if (EVP_CIPHER_CTX_rand_key(evp_ctx, tkey) <= 0)
 			goto err;
-		/* If we have no key use random key */
-		if (jj <= 0)
+		if (ek == NULL)
 			{
-			OPENSSL_free(tmp);
-			jj = tkeylen;
-			tmp = tkey;
+			ek = tkey;
+			eklen = tkeylen;
 			tkey = NULL;
 			}
 
-		if (jj != tkeylen) {
+		if (eklen != EVP_CIPHER_CTX_key_length(evp_ctx)) {
 			/* Some S/MIME clients don't use the same key
 			 * and effective key length. The key length is
 			 * determined by the size of the decrypted RSA key.
 			 */
-			if(!EVP_CIPHER_CTX_set_key_length(evp_ctx, jj))
+			if(!EVP_CIPHER_CTX_set_key_length(evp_ctx, eklen))
 				{
-				/* As MMA defence use random key instead */
-				OPENSSL_cleanse(tmp, jj);
-				OPENSSL_free(tmp);
-				jj = tkeylen;
-				tmp = tkey;
+				/* Use random key as MMA defence */
+				OPENSSL_cleanse(ek, eklen);
+				OPENSSL_free(ek);
+				ek = tkey;
+				eklen = tkeylen;
 				tkey = NULL;
 				}
 		} 
+		/* Clear errors so we don't leak information useful in MMA */
 		ERR_clear_error();
-		if (EVP_CipherInit_ex(evp_ctx,NULL,NULL,tmp,NULL,0) <= 0)
+		if (EVP_CipherInit_ex(evp_ctx,NULL,NULL,ek,NULL,0) <= 0)
 			goto err;
 
-		OPENSSL_cleanse(tmp,jj);
-
+		if (ek)
+			{
+			OPENSSL_cleanse(ek,eklen);
+			OPENSSL_free(ek);
+                       ek = NULL;
+			}
 		if (tkey)
 			{
-			OPENSSL_cleanse(tkey, tkeylen);
+			OPENSSL_cleanse(tkey,tkeylen);
 			OPENSSL_free(tkey);
+                       tkey = NULL;
 			}
 
 		if (out == NULL)
@@ -590,14 +658,22 @@ BIO *PKCS7_dataDecode(PKCS7 *p7, EVP_PKEY *pkey, BIO *in_bio, X509 *pcert)
 	if (0)
 		{
 err:
+               if (ek)
+                       {
+                       OPENSSL_cleanse(ek,eklen);
+                       OPENSSL_free(ek);
+                       }
+               if (tkey)
+                       {
+                       OPENSSL_cleanse(tkey,tkeylen);
+                       OPENSSL_free(tkey);
+                       }
 		if (out != NULL) BIO_free_all(out);
 		if (btmp != NULL) BIO_free_all(btmp);
 		if (etmp != NULL) BIO_free_all(etmp);
 		if (bio != NULL) BIO_free_all(bio);
 		out=NULL;
 		}
-	if (tmp != NULL)
-		OPENSSL_free(tmp);
 	return(out);
 	}
 
@@ -624,13 +700,47 @@ static BIO *PKCS7_find_digest(EVP_MD_CTX **pmd, BIO *bio, int nid)
 	return NULL;
 	}
 
+static int do_pkcs7_signed_attrib(PKCS7_SIGNER_INFO *si, EVP_MD_CTX *mctx)
+	{
+	unsigned char md_data[EVP_MAX_MD_SIZE];
+	unsigned int md_len;
+
+	/* Add signing time if not already present */
+	if (!PKCS7_get_signed_attribute(si, NID_pkcs9_signingTime))
+		{
+		if (!PKCS7_add0_attrib_signing_time(si, NULL))
+			{
+			PKCS7err(PKCS7_F_DO_PKCS7_SIGNED_ATTRIB,
+					ERR_R_MALLOC_FAILURE);
+			return 0;
+			}
+		}
+
+	/* Add digest */
+	if (!EVP_DigestFinal_ex(mctx, md_data,&md_len))
+		{
+		PKCS7err(PKCS7_F_DO_PKCS7_SIGNED_ATTRIB, ERR_R_EVP_LIB);
+		return 0;
+		}
+	if (!PKCS7_add1_attrib_digest(si, md_data, md_len))
+		{
+		PKCS7err(PKCS7_F_DO_PKCS7_SIGNED_ATTRIB, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+
+	/* Now sign the attributes */
+	if (!PKCS7_SIGNER_INFO_sign(si))
+			return 0;
+
+	return 1;
+	}
+	
+				
 int PKCS7_dataFinal(PKCS7 *p7, BIO *bio)
 	{
 	int ret=0;
 	int i,j;
 	BIO *btmp;
-	BUF_MEM *buf_mem=NULL;
-	BUF_MEM *buf=NULL;
 	PKCS7_SIGNER_INFO *si;
 	EVP_MD_CTX *mdc,ctx_tmp;
 	STACK_OF(X509_ATTRIBUTE) *sk;
@@ -643,24 +753,37 @@ int PKCS7_dataFinal(PKCS7 *p7, BIO *bio)
 
 	switch (i)
 		{
+	case NID_pkcs7_data:
+		os = p7->d.data;
+		break;
 	case NID_pkcs7_signedAndEnveloped:
 		/* XXXXXXXXXXXXXXXX */
 		si_sk=p7->d.signed_and_enveloped->signer_info;
-		if (!(os=M_ASN1_OCTET_STRING_new()))
+		os = p7->d.signed_and_enveloped->enc_data->enc_data;
+		if (!os)
 			{
-			PKCS7err(PKCS7_F_PKCS7_DATAFINAL,ERR_R_MALLOC_FAILURE);
-			goto err;
+			os=M_ASN1_OCTET_STRING_new();
+			if (!os)
+				{
+				PKCS7err(PKCS7_F_PKCS7_DATAFINAL,ERR_R_MALLOC_FAILURE);
+				goto err;
+				}
+			p7->d.signed_and_enveloped->enc_data->enc_data=os;
 			}
-		p7->d.signed_and_enveloped->enc_data->enc_data=os;
 		break;
 	case NID_pkcs7_enveloped:
 		/* XXXXXXXXXXXXXXXX */
-		if (!(os=M_ASN1_OCTET_STRING_new()))
+		os = p7->d.enveloped->enc_data->enc_data;
+		if (!os)
 			{
-			PKCS7err(PKCS7_F_PKCS7_DATAFINAL,ERR_R_MALLOC_FAILURE);
-			goto err;
+			os=M_ASN1_OCTET_STRING_new();
+			if (!os)
+				{
+				PKCS7err(PKCS7_F_PKCS7_DATAFINAL,ERR_R_MALLOC_FAILURE);
+				goto err;
+				}
+			p7->d.enveloped->enc_data->enc_data=os;
 			}
-		p7->d.enveloped->enc_data->enc_data=os;
 		break;
 	case NID_pkcs7_signed:
 		si_sk=p7->d.sign->signer_info;
@@ -682,21 +805,20 @@ int PKCS7_dataFinal(PKCS7 *p7, BIO *bio)
 			}
 		break;
 
+	default:
+		PKCS7err(PKCS7_F_PKCS7_DATAFINAL,PKCS7_R_UNSUPPORTED_CONTENT_TYPE);
+	        goto err;
 		}
 
 	if (si_sk != NULL)
 		{
-		if ((buf=BUF_MEM_new()) == NULL)
-			{
-			PKCS7err(PKCS7_F_PKCS7_DATAFINAL,ERR_R_BIO_LIB);
-			goto err;
-			}
 		for (i=0; i<sk_PKCS7_SIGNER_INFO_num(si_sk); i++)
 			{
 			si=sk_PKCS7_SIGNER_INFO_value(si_sk,i);
-			if (si->pkey == NULL) continue;
+			if (si->pkey == NULL)
+				continue;
 
-			j=OBJ_obj2nid(si->digest_alg->algorithm);
+			j = OBJ_obj2nid(si->digest_alg->algorithm);
 
 			btmp=bio;
 
@@ -707,98 +829,35 @@ int PKCS7_dataFinal(PKCS7 *p7, BIO *bio)
 
 			/* We now have the EVP_MD_CTX, lets do the
 			 * signing. */
-			EVP_MD_CTX_copy_ex(&ctx_tmp,mdc);
-			if (!BUF_MEM_grow_clean(buf,EVP_PKEY_size(si->pkey)))
-				{
-				PKCS7err(PKCS7_F_PKCS7_DATAFINAL,ERR_R_BIO_LIB);
+			if (!EVP_MD_CTX_copy_ex(&ctx_tmp,mdc))
 				goto err;
-				}
 
 			sk=si->auth_attr;
 
 			/* If there are attributes, we add the digest
 			 * attribute and only sign the attributes */
-			if ((sk != NULL) && (sk_X509_ATTRIBUTE_num(sk) != 0))
+			if (sk_X509_ATTRIBUTE_num(sk) > 0)
 				{
-				unsigned char md_data[EVP_MAX_MD_SIZE], *abuf=NULL;
-				unsigned int md_len, alen;
-				ASN1_OCTET_STRING *digest;
-				ASN1_UTCTIME *sign_time;
-				const EVP_MD *md_tmp;
-
-				/* Add signing time if not already present */
-				if (!PKCS7_get_signed_attribute(si,
-							NID_pkcs9_signingTime))
-					{
-					if (!(sign_time=X509_gmtime_adj(NULL,0)))
-						{
-						PKCS7err(PKCS7_F_PKCS7_DATAFINAL,
-							ERR_R_MALLOC_FAILURE);
-						goto err;
-						}
-					if (!PKCS7_add_signed_attribute(si,
-						NID_pkcs9_signingTime,
-						V_ASN1_UTCTIME,sign_time))
-						{
-						M_ASN1_UTCTIME_free(sign_time);
-						goto err;
-						}
-					}
-
-				/* Add digest */
-				md_tmp=EVP_MD_CTX_md(&ctx_tmp);
-				EVP_DigestFinal_ex(&ctx_tmp,md_data,&md_len);
-				if (!(digest=M_ASN1_OCTET_STRING_new()))
-					{
-					PKCS7err(PKCS7_F_PKCS7_DATAFINAL,
-						ERR_R_MALLOC_FAILURE);
+				if (!do_pkcs7_signed_attrib(si, &ctx_tmp))
 					goto err;
-					}
-				if (!M_ASN1_OCTET_STRING_set(digest,md_data,
-								md_len))
-					{
-					PKCS7err(PKCS7_F_PKCS7_DATAFINAL,
-						ERR_R_MALLOC_FAILURE);
-					M_ASN1_OCTET_STRING_free(digest);
+				}
+			else
+				{
+				unsigned char *abuf = NULL;
+				unsigned int abuflen;
+				abuflen = EVP_PKEY_size(si->pkey);
+				abuf = OPENSSL_malloc(abuflen);
+				if (!abuf)
 					goto err;
-					}
-				if (!PKCS7_add_signed_attribute(si,
-					NID_pkcs9_messageDigest,
-					V_ASN1_OCTET_STRING,digest))
+
+				if (!EVP_SignFinal(&ctx_tmp, abuf, &abuflen,
+							si->pkey))
 					{
-					M_ASN1_OCTET_STRING_free(digest);
+					PKCS7err(PKCS7_F_PKCS7_DATAFINAL,
+							ERR_R_EVP_LIB);
 					goto err;
 					}
-
-				/* Now sign the attributes */
-				EVP_SignInit_ex(&ctx_tmp,md_tmp,NULL);
-				alen = ASN1_item_i2d((ASN1_VALUE *)sk,&abuf,
-							ASN1_ITEM_rptr(PKCS7_ATTR_SIGN));
-				if(!abuf) goto err;
-				EVP_SignUpdate(&ctx_tmp,abuf,alen);
-				OPENSSL_free(abuf);
-				}
-
-#ifndef OPENSSL_NO_DSA
-			if (si->pkey->type == EVP_PKEY_DSA)
-				ctx_tmp.digest=EVP_dss1();
-#endif
-#ifndef OPENSSL_NO_ECDSA
- 			if (si->pkey->type == EVP_PKEY_EC)
- 				ctx_tmp.digest=EVP_ecdsa();
-#endif
-
-			if (!EVP_SignFinal(&ctx_tmp,(unsigned char *)buf->data,
-				(unsigned int *)&buf->length,si->pkey))
-				{
-				PKCS7err(PKCS7_F_PKCS7_DATAFINAL,ERR_R_EVP_LIB);
-				goto err;
-				}
-			if (!ASN1_STRING_set(si->enc_digest,
-				(unsigned char *)buf->data,buf->length))
-				{
-				PKCS7err(PKCS7_F_PKCS7_DATAFINAL,ERR_R_ASN1_LIB);
-				goto err;
+				ASN1_STRING_set0(si->enc_digest, abuf, abuflen);
 				}
 			}
 		}
@@ -809,38 +868,95 @@ int PKCS7_dataFinal(PKCS7 *p7, BIO *bio)
 		if (!PKCS7_find_digest(&mdc, bio,
 				OBJ_obj2nid(p7->d.digest->md->algorithm)))
 			goto err;
-		EVP_DigestFinal_ex(mdc,md_data,&md_len);
+		if (!EVP_DigestFinal_ex(mdc,md_data,&md_len))
+			goto err;
 		M_ASN1_OCTET_STRING_set(p7->d.digest->digest, md_data, md_len);
 		}
 
-	if (!PKCS7_is_detached(p7))
+	if (!PKCS7_is_detached(p7) && !(os->flags & ASN1_STRING_FLAG_NDEF))
 		{
+		char *cont;
+		long contlen;
 		btmp=BIO_find_type(bio,BIO_TYPE_MEM);
 		if (btmp == NULL)
 			{
 			PKCS7err(PKCS7_F_PKCS7_DATAFINAL,PKCS7_R_UNABLE_TO_FIND_MEM_BIO);
 			goto err;
 			}
-		BIO_get_mem_ptr(btmp,&buf_mem);
+		contlen = BIO_get_mem_data(btmp, &cont);
 		/* Mark the BIO read only then we can use its copy of the data
 		 * instead of making an extra copy.
 		 */
 		BIO_set_flags(btmp, BIO_FLAGS_MEM_RDONLY);
 		BIO_set_mem_eof_return(btmp, 0);
-		os->data = (unsigned char *)buf_mem->data;
-		os->length = buf_mem->length;
-#if 0
-		M_ASN1_OCTET_STRING_set(os,
-			(unsigned char *)buf_mem->data,buf_mem->length);
-#endif
+		ASN1_STRING_set0(os, (unsigned char *)cont, contlen);
 		}
 	ret=1;
 err:
 	EVP_MD_CTX_cleanup(&ctx_tmp);
-	if (buf != NULL) BUF_MEM_free(buf);
 	return(ret);
 	}
 
+int PKCS7_SIGNER_INFO_sign(PKCS7_SIGNER_INFO *si)
+	{
+	EVP_MD_CTX mctx;
+	EVP_PKEY_CTX *pctx;
+	unsigned char *abuf = NULL;
+	int alen;
+	size_t siglen;
+	const EVP_MD *md = NULL;
+
+	md = EVP_get_digestbyobj(si->digest_alg->algorithm);
+	if (md == NULL)
+		return 0;
+
+	EVP_MD_CTX_init(&mctx);
+	if (EVP_DigestSignInit(&mctx, &pctx, md,NULL, si->pkey) <= 0)
+		goto err;
+
+	if (EVP_PKEY_CTX_ctrl(pctx, -1, EVP_PKEY_OP_SIGN,
+				EVP_PKEY_CTRL_PKCS7_SIGN, 0, si) <= 0)
+		{
+		PKCS7err(PKCS7_F_PKCS7_SIGNER_INFO_SIGN, PKCS7_R_CTRL_ERROR);
+		goto err;
+		}
+
+	alen = ASN1_item_i2d((ASN1_VALUE *)si->auth_attr,&abuf,
+				ASN1_ITEM_rptr(PKCS7_ATTR_SIGN));
+	if(!abuf)
+		goto err;
+	if (EVP_DigestSignUpdate(&mctx,abuf,alen) <= 0)
+		goto err;
+	OPENSSL_free(abuf);
+	if (EVP_DigestSignFinal(&mctx, NULL, &siglen) <= 0)
+		goto err;
+	abuf = OPENSSL_malloc(siglen);
+	if(!abuf)
+		goto err;
+	if (EVP_DigestSignFinal(&mctx, abuf, &siglen) <= 0)
+		goto err;
+
+	if (EVP_PKEY_CTX_ctrl(pctx, -1, EVP_PKEY_OP_SIGN,
+				EVP_PKEY_CTRL_PKCS7_SIGN, 1, si) <= 0)
+		{
+		PKCS7err(PKCS7_F_PKCS7_SIGNER_INFO_SIGN, PKCS7_R_CTRL_ERROR);
+		goto err;
+		}
+
+	EVP_MD_CTX_cleanup(&mctx);
+
+	ASN1_STRING_set0(si->enc_digest, abuf, siglen);
+
+	return 1;
+
+	err:
+	if (abuf)
+		OPENSSL_free(abuf);
+	EVP_MD_CTX_cleanup(&mctx);
+	return 0;
+
+	}
+
 int PKCS7_dataVerify(X509_STORE *cert_store, X509_STORE_CTX *ctx, BIO *bio,
 	     PKCS7 *p7, PKCS7_SIGNER_INFO *si)
 	{
@@ -946,16 +1062,19 @@ int PKCS7_signatureVerify(BIO *bio, PKCS7 *p7, PKCS7_SIGNER_INFO *si,
 
 	/* mdc is the digest ctx that we want, unless there are attributes,
 	 * in which case the digest is the signed attributes */
-	EVP_MD_CTX_copy_ex(&mdc_tmp,mdc);
+	if (!EVP_MD_CTX_copy_ex(&mdc_tmp,mdc))
+		goto err;
 
 	sk=si->auth_attr;
 	if ((sk != NULL) && (sk_X509_ATTRIBUTE_num(sk) != 0))
 		{
 		unsigned char md_dat[EVP_MAX_MD_SIZE], *abuf = NULL;
-                unsigned int md_len, alen;
+                unsigned int md_len;
+		int alen;
 		ASN1_OCTET_STRING *message_digest;
 
-		EVP_DigestFinal_ex(&mdc_tmp,md_dat,&md_len);
+		if (!EVP_DigestFinal_ex(&mdc_tmp,md_dat,&md_len))
+			goto err;
 		message_digest=PKCS7_digest_from_attributes(sk);
 		if (!message_digest)
 			{
@@ -980,11 +1099,19 @@ for (ii=0; ii<md_len; ii++) printf("%02X",md_dat[ii]); printf(" calc\n");
 			goto err;
 			}
 
-		EVP_VerifyInit_ex(&mdc_tmp,EVP_get_digestbynid(md_type), NULL);
+		if (!EVP_VerifyInit_ex(&mdc_tmp,EVP_get_digestbynid(md_type), NULL))
+			goto err;
 
 		alen = ASN1_item_i2d((ASN1_VALUE *)sk, &abuf,
 						ASN1_ITEM_rptr(PKCS7_ATTR_VERIFY));
-		EVP_VerifyUpdate(&mdc_tmp, abuf, alen);
+		if (alen <= 0) 
+			{
+			PKCS7err(PKCS7_F_PKCS7_SIGNATUREVERIFY,ERR_R_ASN1_LIB);
+			ret = -1;
+			goto err;
+			}
+		if (!EVP_VerifyUpdate(&mdc_tmp, abuf, alen))
+			goto err;
 
 		OPENSSL_free(abuf);
 		}
@@ -996,12 +1123,6 @@ for (ii=0; ii<md_len; ii++) printf("%02X",md_dat[ii]); printf(" calc\n");
 		ret = -1;
 		goto err;
 		}
-#ifndef OPENSSL_NO_DSA
-	if(pkey->type == EVP_PKEY_DSA) mdc_tmp.digest=EVP_dss1();
-#endif
-#ifndef OPENSSL_NO_ECDSA
-	if (pkey->type == EVP_PKEY_EC) mdc_tmp.digest=EVP_ecdsa();
-#endif
 
 	i=EVP_VerifyFinal(&mdc_tmp,os->data,os->length, pkey);
 	EVP_PKEY_free(pkey);
@@ -1137,8 +1258,9 @@ static int add_attribute(STACK_OF(X509_ATTRIBUTE) **sk, int nid, int atrtype,
 
 	if (*sk == NULL)
 		{
-		if (!(*sk = sk_X509_ATTRIBUTE_new_null()))
-			return 0;
+		*sk = sk_X509_ATTRIBUTE_new_null();
+		if (*sk == NULL)
+			return 0;	
 new_attrib:
 		if (!(attr=X509_ATTRIBUTE_create(nid,atrtype,value)))
 			return 0;
diff --git a/crypto/pkcs7/pk7_lib.c b/crypto/pkcs7/pk7_lib.c
index 898cddad0797..d411269b50bd 100644
--- a/crypto/pkcs7/pk7_lib.c
+++ b/crypto/pkcs7/pk7_lib.c
@@ -60,6 +60,7 @@
 #include "cryptlib.h"
 #include <openssl/objects.h>
 #include <openssl/x509.h>
+#include "asn1_locl.h"
 
 long PKCS7_ctrl(PKCS7 *p7, int cmd, long larg, char *parg)
 	{
@@ -314,7 +315,7 @@ int PKCS7_add_certificate(PKCS7 *p7, X509 *x509)
 		*sk=sk_X509_new_null();
 	if (*sk == NULL)
 		{
-		PKCS7err(PKCS7_F_PKCS7_ADD_CERTIFICATE,ERR_R_MALLOC_FAILURE);
+		PKCS7err(PKCS7_F_PKCS7_ADD_CERTIFICATE, ERR_R_MALLOC_FAILURE);
 		return 0;
 		}
 	CRYPTO_add(&x509->references,1,CRYPTO_LOCK_X509);
@@ -365,13 +366,8 @@ int PKCS7_add_crl(PKCS7 *p7, X509_CRL *crl)
 int PKCS7_SIGNER_INFO_set(PKCS7_SIGNER_INFO *p7i, X509 *x509, EVP_PKEY *pkey,
 	     const EVP_MD *dgst)
 	{
-	int nid;
-	char is_dsa;
+	int ret;
 
-	if (pkey->type == EVP_PKEY_DSA || pkey->type == EVP_PKEY_EC)
-		is_dsa = 1;
-	else
-		is_dsa = 0;
 	/* We now need to add another PKCS7_SIGNER_INFO entry */
 	if (!ASN1_INTEGER_set(p7i->version,1))
 		goto err;
@@ -391,65 +387,55 @@ int PKCS7_SIGNER_INFO_set(PKCS7_SIGNER_INFO *p7i, X509 *x509, EVP_PKEY *pkey,
 	p7i->pkey=pkey;
 
 	/* Set the algorithms */
-	if (is_dsa) p7i->digest_alg->algorithm=OBJ_nid2obj(NID_sha1);
-	else	
-		p7i->digest_alg->algorithm=OBJ_nid2obj(EVP_MD_type(dgst));
 
-	if (p7i->digest_alg->parameter != NULL)
-		ASN1_TYPE_free(p7i->digest_alg->parameter);
-	if ((p7i->digest_alg->parameter=ASN1_TYPE_new()) == NULL)
-		goto err;
-	p7i->digest_alg->parameter->type=V_ASN1_NULL;
+	X509_ALGOR_set0(p7i->digest_alg, OBJ_nid2obj(EVP_MD_type(dgst)),
+				V_ASN1_NULL, NULL);
 
-	if (p7i->digest_enc_alg->parameter != NULL)
-		ASN1_TYPE_free(p7i->digest_enc_alg->parameter);
-	nid = EVP_PKEY_type(pkey->type);
-	if (nid == EVP_PKEY_RSA)
+	if (pkey->ameth && pkey->ameth->pkey_ctrl)
 		{
-		p7i->digest_enc_alg->algorithm=OBJ_nid2obj(NID_rsaEncryption);
-		if (!(p7i->digest_enc_alg->parameter=ASN1_TYPE_new()))
-			goto err;
-		p7i->digest_enc_alg->parameter->type=V_ASN1_NULL;
-		}
-	else if (nid == EVP_PKEY_DSA)
-		{
-#if 1
-		/* use 'dsaEncryption' OID for compatibility with other software
-		 * (PKCS #7 v1.5 does specify how to handle DSA) ... */
-		p7i->digest_enc_alg->algorithm=OBJ_nid2obj(NID_dsa);
-#else
-		/* ... although the 'dsaWithSHA1' OID (as required by RFC 2630 for CMS)
-		 * would make more sense. */
-		p7i->digest_enc_alg->algorithm=OBJ_nid2obj(NID_dsaWithSHA1);
-#endif
-		p7i->digest_enc_alg->parameter = NULL; /* special case for DSA: omit 'parameter'! */
-		}
-	else if (nid == EVP_PKEY_EC)
-		{
-		p7i->digest_enc_alg->algorithm=OBJ_nid2obj(NID_ecdsa_with_SHA1);
-		if (!(p7i->digest_enc_alg->parameter=ASN1_TYPE_new()))
-			goto err;
-		p7i->digest_enc_alg->parameter->type=V_ASN1_NULL;
+		ret = pkey->ameth->pkey_ctrl(pkey, ASN1_PKEY_CTRL_PKCS7_SIGN,
+						0, p7i);
+		if (ret > 0)
+			return 1;
+		if (ret != -2)
+			{
+			PKCS7err(PKCS7_F_PKCS7_SIGNER_INFO_SET,
+					PKCS7_R_SIGNING_CTRL_FAILURE);
+			return 0;
+			}
 		}
-	else
-		return(0);
-
-	return(1);
+	PKCS7err(PKCS7_F_PKCS7_SIGNER_INFO_SET,
+			PKCS7_R_SIGNING_NOT_SUPPORTED_FOR_THIS_KEY_TYPE);
 err:
-	return(0);
+	return 0;
 	}
 
 PKCS7_SIGNER_INFO *PKCS7_add_signature(PKCS7 *p7, X509 *x509, EVP_PKEY *pkey,
 	     const EVP_MD *dgst)
 	{
-	PKCS7_SIGNER_INFO *si;
+	PKCS7_SIGNER_INFO *si = NULL;
+
+	if (dgst == NULL)
+		{
+		int def_nid;
+		if (EVP_PKEY_get_default_digest_nid(pkey, &def_nid) <= 0)
+			goto err;
+		dgst = EVP_get_digestbynid(def_nid);
+		if (dgst == NULL)
+			{
+			PKCS7err(PKCS7_F_PKCS7_ADD_SIGNATURE,
+						PKCS7_R_NO_DEFAULT_DIGEST);
+			goto err;
+			}
+		}
 
 	if ((si=PKCS7_SIGNER_INFO_new()) == NULL) goto err;
 	if (!PKCS7_SIGNER_INFO_set(si,x509,pkey,dgst)) goto err;
 	if (!PKCS7_add_signer(p7,si)) goto err;
 	return(si);
 err:
-	PKCS7_SIGNER_INFO_free(si);
+	if (si)
+		PKCS7_SIGNER_INFO_free(si);
 	return(NULL);
 	}
 
@@ -485,6 +471,23 @@ STACK_OF(PKCS7_SIGNER_INFO) *PKCS7_get_signer_info(PKCS7 *p7)
 		return(NULL);
 	}
 
+void PKCS7_SIGNER_INFO_get0_algs(PKCS7_SIGNER_INFO *si, EVP_PKEY **pk,
+					X509_ALGOR **pdig, X509_ALGOR **psig)
+	{
+	if (pk)
+		*pk = si->pkey;
+	if (pdig)
+		*pdig = si->digest_alg;
+	if (psig)
+		*psig = si->digest_enc_alg;
+	}
+
+void PKCS7_RECIP_INFO_get0_alg(PKCS7_RECIP_INFO *ri, X509_ALGOR **penc)
+	{
+	if (penc)
+		*penc = ri->key_enc_algor;
+	}
+
 PKCS7_RECIP_INFO *PKCS7_add_recipient(PKCS7 *p7, X509 *x509)
 	{
 	PKCS7_RECIP_INFO *ri;
@@ -492,10 +495,11 @@ PKCS7_RECIP_INFO *PKCS7_add_recipient(PKCS7 *p7, X509 *x509)
 	if ((ri=PKCS7_RECIP_INFO_new()) == NULL) goto err;
 	if (!PKCS7_RECIP_INFO_set(ri,x509)) goto err;
 	if (!PKCS7_add_recipient_info(p7,ri)) goto err;
-	return(ri);
+	return ri;
 err:
-	PKCS7_RECIP_INFO_free(ri);
-	return(NULL);
+	if (ri)
+		PKCS7_RECIP_INFO_free(ri);
+	return NULL;
 	}
 
 int PKCS7_add_recipient_info(PKCS7 *p7, PKCS7_RECIP_INFO *ri)
@@ -524,6 +528,8 @@ int PKCS7_add_recipient_info(PKCS7 *p7, PKCS7_RECIP_INFO *ri)
 
 int PKCS7_RECIP_INFO_set(PKCS7_RECIP_INFO *p7i, X509 *x509)
 	{
+	int ret;
+	EVP_PKEY *pkey = NULL;
 	if (!ASN1_INTEGER_set(p7i->version,0))
 		return 0;
 	if (!X509_NAME_set(&p7i->issuer_and_serial->issuer,
@@ -535,14 +541,41 @@ int PKCS7_RECIP_INFO_set(PKCS7_RECIP_INFO *p7i, X509 *x509)
 		M_ASN1_INTEGER_dup(X509_get_serialNumber(x509))))
 		return 0;
 
-	X509_ALGOR_free(p7i->key_enc_algor);
-	if (!(p7i->key_enc_algor= X509_ALGOR_dup(x509->cert_info->key->algor)))
-		return 0;
+	pkey = X509_get_pubkey(x509);
+
+	if (!pkey || !pkey->ameth || !pkey->ameth->pkey_ctrl)
+		{
+		PKCS7err(PKCS7_F_PKCS7_RECIP_INFO_SET,
+			PKCS7_R_ENCRYPTION_NOT_SUPPORTED_FOR_THIS_KEY_TYPE);
+		goto err;
+		}
+
+	ret = pkey->ameth->pkey_ctrl(pkey, ASN1_PKEY_CTRL_PKCS7_ENCRYPT,
+						0, p7i);
+	if (ret == -2)
+		{
+		PKCS7err(PKCS7_F_PKCS7_RECIP_INFO_SET,
+			PKCS7_R_ENCRYPTION_NOT_SUPPORTED_FOR_THIS_KEY_TYPE);
+		goto err;
+		}
+	if (ret <= 0)
+		{
+		PKCS7err(PKCS7_F_PKCS7_RECIP_INFO_SET,
+				PKCS7_R_ENCRYPTION_CTRL_FAILURE);
+		goto err;
+		}
+
+	EVP_PKEY_free(pkey);
 
 	CRYPTO_add(&x509->references,1,CRYPTO_LOCK_X509);
 	p7i->cert=x509;
 
-	return(1);
+	return 1;
+
+	err:
+	if (pkey)
+		EVP_PKEY_free(pkey);
+	return 0;
 	}
 
 X509 *PKCS7_cert_from_signer_info(PKCS7 *p7, PKCS7_SIGNER_INFO *si)
@@ -585,3 +618,48 @@ int PKCS7_set_cipher(PKCS7 *p7, const EVP_CIPHER *cipher)
 	return 1;
 	}
 
+int PKCS7_stream(unsigned char ***boundary, PKCS7 *p7)
+	{
+	ASN1_OCTET_STRING *os = NULL;
+
+	switch (OBJ_obj2nid(p7->type))
+		{
+		case NID_pkcs7_data:
+		os = p7->d.data;
+		break;
+
+		case NID_pkcs7_signedAndEnveloped:
+		os = p7->d.signed_and_enveloped->enc_data->enc_data;
+		if (os == NULL)
+			{
+			os=M_ASN1_OCTET_STRING_new();
+			p7->d.signed_and_enveloped->enc_data->enc_data=os;
+			}
+		break;
+
+		case NID_pkcs7_enveloped:
+		os = p7->d.enveloped->enc_data->enc_data;
+		if (os == NULL)
+			{
+			os=M_ASN1_OCTET_STRING_new();
+			p7->d.enveloped->enc_data->enc_data=os;
+			}
+		break;
+
+		case NID_pkcs7_signed:
+		os=p7->d.sign->contents->d.data;
+		break;
+
+		default:
+		os = NULL;
+		break;
+		}
+	
+	if (os == NULL)
+		return 0;
+
+	os->flags |= ASN1_STRING_FLAG_NDEF;
+	*boundary = &os->data;
+
+	return 1;
+	}
diff --git a/crypto/pkcs7/pk7_mime.c b/crypto/pkcs7/pk7_mime.c
index 831b47d230b7..938f79a646cb 100644
--- a/crypto/pkcs7/pk7_mime.c
+++ b/crypto/pkcs7/pk7_mime.c
@@ -59,54 +59,19 @@
 #include <openssl/x509.h>
 #include <openssl/asn1.h>
 
-/* PKCS#7 wrappers round generalised MIME routines */
+/* PKCS#7 wrappers round generalised stream and MIME routines */
 
-PKCS7 *SMIME_read_PKCS7(BIO *bio, BIO **bcont)
+int i2d_PKCS7_bio_stream(BIO *out, PKCS7 *p7, BIO *in, int flags)
 	{
-	return (PKCS7 *)SMIME_read_ASN1(bio, bcont, ASN1_ITEM_rptr(PKCS7));
+	return i2d_ASN1_bio_stream(out, (ASN1_VALUE *)p7, in, flags,
+					ASN1_ITEM_rptr(PKCS7));
 	}
 
-/* Callback for int_smime_write_ASN1 */
-
-static int pk7_output_data(BIO *out, BIO *data, ASN1_VALUE *val, int flags,
-					const ASN1_ITEM *it)
+int PEM_write_bio_PKCS7_stream(BIO *out, PKCS7 *p7, BIO *in, int flags)
 	{
-	PKCS7 *p7 = (PKCS7 *)val;
-	BIO *tmpbio, *p7bio;
-
-	if (!(flags & SMIME_DETACHED))
-		{
-		SMIME_crlf_copy(data, out, flags);
-		return 1;
-		}
-
-	/* Let PKCS7 code prepend any needed BIOs */
-
-	p7bio = PKCS7_dataInit(p7, out);
-
-	if (!p7bio)
-		return 0;
-
-	/* Copy data across, passing through filter BIOs for processing */
-	SMIME_crlf_copy(data, p7bio, flags);
-
-	/* Finalize structure */
-	if (PKCS7_dataFinal(p7, p7bio) <= 0)
-		goto err;
-
-	err:
-
-	/* Now remove any digests prepended to the BIO */
-
-	while (p7bio != out)
-		{
-		tmpbio = BIO_pop(p7bio);
-		BIO_free(p7bio);
-		p7bio = tmpbio;
-		}
-
-	return 1;
-
+	return PEM_write_bio_ASN1_stream(out, (ASN1_VALUE *) p7, in, flags,
+						"PKCS7",
+						ASN1_ITEM_rptr(PKCS7));
 	}
 
 int SMIME_write_PKCS7(BIO *bio, PKCS7 *p7, BIO *data, int flags)
@@ -118,8 +83,15 @@ int SMIME_write_PKCS7(BIO *bio, PKCS7 *p7, BIO *data, int flags)
 	else
 		mdalgs = NULL;
 
-	return int_smime_write_ASN1(bio, (ASN1_VALUE *)p7, data, flags,
+	flags ^= SMIME_OLDMIME;
+
+
+	return SMIME_write_ASN1(bio, (ASN1_VALUE *)p7, data, flags,
 					ctype_nid, NID_undef, mdalgs,
-					pk7_output_data,
 					ASN1_ITEM_rptr(PKCS7));	
 	}
+
+PKCS7 *SMIME_read_PKCS7(BIO *bio, BIO **bcont)
+	{
+	return (PKCS7 *)SMIME_read_ASN1(bio, bcont, ASN1_ITEM_rptr(PKCS7));
+	}
diff --git a/crypto/pkcs7/pk7_smime.c b/crypto/pkcs7/pk7_smime.c
index 6b458af3d795..a5104f8d05af 100644
--- a/crypto/pkcs7/pk7_smime.c
+++ b/crypto/pkcs7/pk7_smime.c
@@ -63,24 +63,19 @@
 #include <openssl/x509.h>
 #include <openssl/x509v3.h>
 
+static int pkcs7_copy_existing_digest(PKCS7 *p7, PKCS7_SIGNER_INFO *si);
+
 PKCS7 *PKCS7_sign(X509 *signcert, EVP_PKEY *pkey, STACK_OF(X509) *certs,
 		  BIO *data, int flags)
 {
-	PKCS7 *p7 = NULL;
-	PKCS7_SIGNER_INFO *si;
-	BIO *p7bio = NULL;
-	STACK_OF(X509_ALGOR) *smcap = NULL;
+	PKCS7 *p7;
 	int i;
 
-	if(!X509_check_private_key(signcert, pkey)) {
-		PKCS7err(PKCS7_F_PKCS7_SIGN,PKCS7_R_PRIVATE_KEY_DOES_NOT_MATCH_CERTIFICATE);
-                return NULL;
-	}
-
-	if(!(p7 = PKCS7_new())) {
+	if(!(p7 = PKCS7_new()))
+		{
 		PKCS7err(PKCS7_F_PKCS7_SIGN,ERR_R_MALLOC_FAILURE);
 		return NULL;
-	}
+		}
 
 	if (!PKCS7_set_type(p7, NID_pkcs7_signed))
 		goto err;
@@ -88,82 +83,185 @@ PKCS7 *PKCS7_sign(X509 *signcert, EVP_PKEY *pkey, STACK_OF(X509) *certs,
 	if (!PKCS7_content_new(p7, NID_pkcs7_data))
 		goto err;
 
-	if (!(si = PKCS7_add_signature(p7,signcert,pkey,EVP_sha1()))) {
-		PKCS7err(PKCS7_F_PKCS7_SIGN,PKCS7_R_PKCS7_ADD_SIGNATURE_ERROR);
+    	if (pkey && !PKCS7_sign_add_signer(p7, signcert, pkey, NULL, flags))
+		{
+		PKCS7err(PKCS7_F_PKCS7_SIGN,PKCS7_R_PKCS7_ADD_SIGNER_ERROR);
 		goto err;
-	}
+		}
 
-	if(!(flags & PKCS7_NOCERTS)) {
-		if (!PKCS7_add_certificate(p7, signcert))
-			goto err;
-		if(certs) for(i = 0; i < sk_X509_num(certs); i++)
+	if(!(flags & PKCS7_NOCERTS))
+		{
+		for(i = 0; i < sk_X509_num(certs); i++)
+			{
 			if (!PKCS7_add_certificate(p7, sk_X509_value(certs, i)))
 				goto err;
-	}
+			}
+		}
 
-	if(!(flags & PKCS7_NOATTR)) {
-		if (!PKCS7_add_signed_attribute(si, NID_pkcs9_contentType,
-				V_ASN1_OBJECT, OBJ_nid2obj(NID_pkcs7_data)))
-			goto err;
-		/* Add SMIMECapabilities */
-		if(!(flags & PKCS7_NOSMIMECAP))
+	if(flags & PKCS7_DETACHED)
+		PKCS7_set_detached(p7, 1);
+
+	if (flags & (PKCS7_STREAM|PKCS7_PARTIAL))
+		return p7;
+
+	if (PKCS7_final(p7, data, flags))
+		return p7;
+
+	err:
+	PKCS7_free(p7);
+	return NULL;
+}
+
+int PKCS7_final(PKCS7 *p7, BIO *data, int flags)
+	{
+	BIO *p7bio;
+	int ret = 0;
+	if (!(p7bio = PKCS7_dataInit(p7, NULL)))
 		{
-		if(!(smcap = sk_X509_ALGOR_new_null())) {
-			PKCS7err(PKCS7_F_PKCS7_SIGN,ERR_R_MALLOC_FAILURE);
-			goto err;
-		}
-#ifndef OPENSSL_NO_DES
-		if (!PKCS7_simple_smimecap (smcap, NID_des_ede3_cbc, -1))
-			goto err;
-#endif
-#ifndef OPENSSL_NO_RC2
-		if (!PKCS7_simple_smimecap (smcap, NID_rc2_cbc, 128))
-			goto err;
-		if (!PKCS7_simple_smimecap (smcap, NID_rc2_cbc, 64))
-			goto err;
-#endif
-#ifndef OPENSSL_NO_DES
-		if (!PKCS7_simple_smimecap (smcap, NID_des_cbc, -1))
-			goto err;
-#endif
-#ifndef OPENSSL_NO_RC2
-		if (!PKCS7_simple_smimecap (smcap, NID_rc2_cbc, 40))
-			goto err;
-#endif
-		if (!PKCS7_add_attrib_smimecap (si, smcap))
-			goto err;
-		sk_X509_ALGOR_pop_free(smcap, X509_ALGOR_free);
-		smcap = NULL;
+		PKCS7err(PKCS7_F_PKCS7_FINAL,ERR_R_MALLOC_FAILURE);
+		return 0;
 		}
-	}
 
-	if(flags & PKCS7_DETACHED)PKCS7_set_detached(p7, 1);
+	SMIME_crlf_copy(data, p7bio, flags);
 
-	if (flags & PKCS7_STREAM)
-		return p7;
+	(void)BIO_flush(p7bio);
 
 
-	if (!(p7bio = PKCS7_dataInit(p7, NULL))) {
-		PKCS7err(PKCS7_F_PKCS7_SIGN,ERR_R_MALLOC_FAILURE);
+        if (!PKCS7_dataFinal(p7,p7bio))
+		{
+		PKCS7err(PKCS7_F_PKCS7_FINAL,PKCS7_R_PKCS7_DATASIGN);
 		goto err;
+		}
+
+	ret = 1;
+
+	err:
+	BIO_free_all(p7bio);
+
+	return ret;
+
 	}
 
-	SMIME_crlf_copy(data, p7bio, flags);
+/* Check to see if a cipher exists and if so add S/MIME capabilities */
 
+static int add_cipher_smcap(STACK_OF(X509_ALGOR) *sk, int nid, int arg)
+	{
+	if (EVP_get_cipherbynid(nid))
+		return PKCS7_simple_smimecap(sk, nid, arg);
+	return 1;
+	}
 
-	if (!PKCS7_dataFinal(p7,p7bio)) {
-		PKCS7err(PKCS7_F_PKCS7_SIGN,PKCS7_R_PKCS7_DATASIGN);
-		goto err;
+static int add_digest_smcap(STACK_OF(X509_ALGOR) *sk, int nid, int arg)
+	{
+	if (EVP_get_digestbynid(nid))
+		return PKCS7_simple_smimecap(sk, nid, arg);
+	return 1;
 	}
 
-	BIO_free_all(p7bio);
-	return p7;
-err:
-	sk_X509_ALGOR_pop_free(smcap, X509_ALGOR_free);
-	BIO_free_all(p7bio);
-	PKCS7_free(p7);
+PKCS7_SIGNER_INFO *PKCS7_sign_add_signer(PKCS7 *p7, X509 *signcert,
+					EVP_PKEY *pkey, const EVP_MD *md,
+					int flags)
+	{
+	PKCS7_SIGNER_INFO *si = NULL;
+	STACK_OF(X509_ALGOR) *smcap = NULL;
+	if(!X509_check_private_key(signcert, pkey))
+		{
+		PKCS7err(PKCS7_F_PKCS7_SIGN_ADD_SIGNER,
+			PKCS7_R_PRIVATE_KEY_DOES_NOT_MATCH_CERTIFICATE);
+                return NULL;
+		}
+
+    	if (!(si = PKCS7_add_signature(p7,signcert,pkey, md)))
+		{
+		PKCS7err(PKCS7_F_PKCS7_SIGN_ADD_SIGNER,
+				PKCS7_R_PKCS7_ADD_SIGNATURE_ERROR);
+		return NULL;
+		}
+
+	if(!(flags & PKCS7_NOCERTS))
+		{
+		if (!PKCS7_add_certificate(p7, signcert))
+			goto err;
+		}
+
+	if(!(flags & PKCS7_NOATTR))
+		{
+		if (!PKCS7_add_attrib_content_type(si, NULL))
+			goto err;
+		/* Add SMIMECapabilities */
+		if(!(flags & PKCS7_NOSMIMECAP))
+			{
+			if(!(smcap = sk_X509_ALGOR_new_null()))
+				{
+				PKCS7err(PKCS7_F_PKCS7_SIGN_ADD_SIGNER,
+					ERR_R_MALLOC_FAILURE);
+				goto err;
+				}
+			if (!add_cipher_smcap(smcap, NID_aes_256_cbc, -1)
+			|| !add_digest_smcap(smcap, NID_id_GostR3411_94, -1)
+			|| !add_cipher_smcap(smcap, NID_id_Gost28147_89, -1)
+				|| !add_cipher_smcap(smcap, NID_aes_192_cbc, -1)
+				|| !add_cipher_smcap(smcap, NID_aes_128_cbc, -1)
+			|| !add_cipher_smcap(smcap, NID_des_ede3_cbc, -1)
+				|| !add_cipher_smcap(smcap, NID_rc2_cbc, 128)
+				|| !add_cipher_smcap(smcap, NID_rc2_cbc, 64)
+				|| !add_cipher_smcap(smcap, NID_des_cbc, -1)
+				|| !add_cipher_smcap(smcap, NID_rc2_cbc, 40)
+				|| !PKCS7_add_attrib_smimecap (si, smcap))
+				goto err;
+			sk_X509_ALGOR_pop_free(smcap, X509_ALGOR_free);
+			smcap = NULL;
+			}
+		if (flags & PKCS7_REUSE_DIGEST)
+			{
+			if (!pkcs7_copy_existing_digest(p7, si))
+				goto err;
+			if (!(flags & PKCS7_PARTIAL) &&
+					!PKCS7_SIGNER_INFO_sign(si))
+				goto err;
+			}
+		}
+	return si;
+	err:
+	if (smcap)
+		sk_X509_ALGOR_pop_free(smcap, X509_ALGOR_free);
 	return NULL;
-}
+	}
+
+/* Search for a digest matching SignerInfo digest type and if found
+ * copy across.
+ */
+
+static int pkcs7_copy_existing_digest(PKCS7 *p7, PKCS7_SIGNER_INFO *si)
+	{
+	int i;
+	STACK_OF(PKCS7_SIGNER_INFO) *sinfos;
+	PKCS7_SIGNER_INFO *sitmp;
+	ASN1_OCTET_STRING *osdig = NULL;
+	sinfos = PKCS7_get_signer_info(p7);
+	for (i = 0; i < sk_PKCS7_SIGNER_INFO_num(sinfos); i++)
+		{
+		sitmp = sk_PKCS7_SIGNER_INFO_value(sinfos, i);
+		if (si == sitmp)
+			break;
+		if (sk_X509_ATTRIBUTE_num(sitmp->auth_attr) <= 0)
+			continue;
+		if (!OBJ_cmp(si->digest_alg->algorithm,
+				sitmp->digest_alg->algorithm))
+			{
+			osdig = PKCS7_digest_from_attributes(sitmp->auth_attr);
+			break;
+			}
+
+		}
+
+	if (osdig)
+		return PKCS7_add1_attrib_digest(si, osdig->data, osdig->length);
+
+	PKCS7err(PKCS7_F_PKCS7_COPY_EXISTING_DIGEST,
+			PKCS7_R_NO_MATCHING_DIGEST_TYPE_FOUND);
+	return 0;
+	}
 
 int PKCS7_verify(PKCS7 *p7, STACK_OF(X509) *certs, X509_STORE *store,
 					BIO *indata, BIO *out, int flags)
@@ -354,7 +452,7 @@ STACK_OF(X509) *PKCS7_get0_signers(PKCS7 *p7, STACK_OF(X509) *certs, int flags)
 
 	if(sk_PKCS7_SIGNER_INFO_num(sinfos) <= 0) {
 		PKCS7err(PKCS7_F_PKCS7_GET0_SIGNERS,PKCS7_R_NO_SIGNERS);
-		return NULL;
+		return 0;
 	}
 
 	if(!(signers = sk_X509_new_null())) {
@@ -377,12 +475,12 @@ STACK_OF(X509) *PKCS7_get0_signers(PKCS7 *p7, STACK_OF(X509) *certs, int flags)
 	    if (!signer) {
 			PKCS7err(PKCS7_F_PKCS7_GET0_SIGNERS,PKCS7_R_SIGNER_CERTIFICATE_NOT_FOUND);
 			sk_X509_free(signers);
-			return NULL;
+			return 0;
 	    }
 
 	    if (!sk_X509_push(signers, signer)) {
-			sk_X509_free(signers);
-			return NULL;
+		sk_X509_free(signers);
+		return NULL;
 	    }
 	}
 	return signers;
@@ -405,7 +503,7 @@ PKCS7 *PKCS7_encrypt(STACK_OF(X509) *certs, BIO *in, const EVP_CIPHER *cipher,
 
 	if (!PKCS7_set_type(p7, NID_pkcs7_enveloped))
 		goto err;
-	if(!PKCS7_set_cipher(p7, cipher)) {
+	if (!PKCS7_set_cipher(p7, cipher)) {
 		PKCS7err(PKCS7_F_PKCS7_ENCRYPT,PKCS7_R_ERROR_SETTING_CIPHER);
 		goto err;
 	}
@@ -419,22 +517,11 @@ PKCS7 *PKCS7_encrypt(STACK_OF(X509) *certs, BIO *in, const EVP_CIPHER *cipher,
 		}
 	}
 
-	if(!(p7bio = PKCS7_dataInit(p7, NULL))) {
-		PKCS7err(PKCS7_F_PKCS7_ENCRYPT,ERR_R_MALLOC_FAILURE);
-		goto err;
-	}
-
-	SMIME_crlf_copy(in, p7bio, flags);
-
-	(void)BIO_flush(p7bio);
-
-        if (!PKCS7_dataFinal(p7,p7bio)) {
-		PKCS7err(PKCS7_F_PKCS7_ENCRYPT,PKCS7_R_PKCS7_DATAFINAL_ERROR);
-		goto err;
-	}
-        BIO_free_all(p7bio);
+	if (flags & PKCS7_STREAM)
+		return p7;
 
-	return p7;
+	if (PKCS7_final(p7, in, flags))
+		return p7;
 
 	err:
 
diff --git a/crypto/pkcs7/pkcs7.h b/crypto/pkcs7/pkcs7.h
index cc092d262dc3..e4d443193c43 100644
--- a/crypto/pkcs7/pkcs7.h
+++ b/crypto/pkcs7/pkcs7.h
@@ -232,6 +232,9 @@ DECLARE_PKCS12_STACK_OF(PKCS7)
 #define PKCS7_type_is_signedAndEnveloped(a) \
 		(OBJ_obj2nid((a)->type) == NID_pkcs7_signedAndEnveloped)
 #define PKCS7_type_is_data(a)   (OBJ_obj2nid((a)->type) == NID_pkcs7_data)
+#define PKCS7_type_is_digest(a)   (OBJ_obj2nid((a)->type) == NID_pkcs7_digest)
+#define PKCS7_type_is_encrypted(a) \
+		(OBJ_obj2nid((a)->type) == NID_pkcs7_encrypted)
 
 #define PKCS7_type_is_digest(a)   (OBJ_obj2nid((a)->type) == NID_pkcs7_digest)
 
@@ -242,14 +245,6 @@ DECLARE_PKCS12_STACK_OF(PKCS7)
 
 #define PKCS7_is_detached(p7) (PKCS7_type_is_signed(p7) && PKCS7_get_detached(p7))
 
-#ifdef SSLEAY_MACROS
-#ifndef PKCS7_ISSUER_AND_SERIAL_digest
-#define PKCS7_ISSUER_AND_SERIAL_digest(data,type,md,len) \
-        ASN1_digest((int (*)())i2d_PKCS7_ISSUER_AND_SERIAL,type,\
-	                (char *)data,md,len)
-#endif
-#endif
-
 /* S/MIME related flags */
 
 #define PKCS7_TEXT		0x1
@@ -266,6 +261,8 @@ DECLARE_PKCS12_STACK_OF(PKCS7)
 #define PKCS7_CRLFEOL		0x800
 #define PKCS7_STREAM		0x1000
 #define PKCS7_NOCRL		0x2000
+#define PKCS7_PARTIAL		0x4000
+#define PKCS7_REUSE_DIGEST	0x8000
 
 /* Flags: for compatibility with older code */
 
@@ -281,7 +278,6 @@ DECLARE_PKCS12_STACK_OF(PKCS7)
 
 DECLARE_ASN1_FUNCTIONS(PKCS7_ISSUER_AND_SERIAL)
 
-#ifndef SSLEAY_MACROS
 int PKCS7_ISSUER_AND_SERIAL_digest(PKCS7_ISSUER_AND_SERIAL *data,const EVP_MD *type,
 	unsigned char *md,unsigned int *len);
 #ifndef OPENSSL_NO_FP_API
@@ -291,7 +287,8 @@ int i2d_PKCS7_fp(FILE *fp,PKCS7 *p7);
 PKCS7 *PKCS7_dup(PKCS7 *p7);
 PKCS7 *d2i_PKCS7_bio(BIO *bp,PKCS7 **p7);
 int i2d_PKCS7_bio(BIO *bp,PKCS7 *p7);
-#endif
+int i2d_PKCS7_bio_stream(BIO *out, PKCS7 *p7, BIO *in, int flags);
+int PEM_write_bio_PKCS7_stream(BIO *out, PKCS7 *p7, BIO *in, int flags);
 
 DECLARE_ASN1_FUNCTIONS(PKCS7_SIGNER_INFO)
 DECLARE_ASN1_FUNCTIONS(PKCS7_RECIP_INFO)
@@ -307,6 +304,7 @@ DECLARE_ASN1_ITEM(PKCS7_ATTR_SIGN)
 DECLARE_ASN1_ITEM(PKCS7_ATTR_VERIFY)
 
 DECLARE_ASN1_NDEF_FUNCTION(PKCS7)
+DECLARE_ASN1_PRINT_FUNCTION(PKCS7)
 
 long PKCS7_ctrl(PKCS7 *p7, int cmd, long larg, char *parg);
 
@@ -315,6 +313,7 @@ int PKCS7_set0_type_other(PKCS7 *p7, int type, ASN1_TYPE *other);
 int PKCS7_set_content(PKCS7 *p7, PKCS7 *p7_data);
 int PKCS7_SIGNER_INFO_set(PKCS7_SIGNER_INFO *p7i, X509 *x509, EVP_PKEY *pkey,
 	const EVP_MD *dgst);
+int PKCS7_SIGNER_INFO_sign(PKCS7_SIGNER_INFO *si);
 int PKCS7_add_signer(PKCS7 *p7, PKCS7_SIGNER_INFO *p7i);
 int PKCS7_add_certificate(PKCS7 *p7, X509 *x509);
 int PKCS7_add_crl(PKCS7 *p7, X509_CRL *x509);
@@ -336,9 +335,13 @@ int PKCS7_set_digest(PKCS7 *p7, const EVP_MD *md);
 STACK_OF(PKCS7_SIGNER_INFO) *PKCS7_get_signer_info(PKCS7 *p7);
 
 PKCS7_RECIP_INFO *PKCS7_add_recipient(PKCS7 *p7, X509 *x509);
+void PKCS7_SIGNER_INFO_get0_algs(PKCS7_SIGNER_INFO *si, EVP_PKEY **pk,
+					X509_ALGOR **pdig, X509_ALGOR **psig);
+void PKCS7_RECIP_INFO_get0_alg(PKCS7_RECIP_INFO *ri, X509_ALGOR **penc);
 int PKCS7_add_recipient_info(PKCS7 *p7, PKCS7_RECIP_INFO *ri);
 int PKCS7_RECIP_INFO_set(PKCS7_RECIP_INFO *p7i, X509 *x509);
 int PKCS7_set_cipher(PKCS7 *p7, const EVP_CIPHER *cipher);
+int PKCS7_stream(unsigned char ***boundary, PKCS7 *p7);
 
 PKCS7_ISSUER_AND_SERIAL *PKCS7_get_issuer_and_serial(PKCS7 *p7, int idx);
 ASN1_OCTET_STRING *PKCS7_digest_from_attributes(STACK_OF(X509_ATTRIBUTE) *sk);
@@ -355,6 +358,12 @@ int PKCS7_set_attributes(PKCS7_SIGNER_INFO *p7si,STACK_OF(X509_ATTRIBUTE) *sk);
 
 PKCS7 *PKCS7_sign(X509 *signcert, EVP_PKEY *pkey, STACK_OF(X509) *certs,
 							BIO *data, int flags);
+
+PKCS7_SIGNER_INFO *PKCS7_sign_add_signer(PKCS7 *p7,
+			X509 *signcert, EVP_PKEY *pkey, const EVP_MD *md,
+			int flags);
+
+int PKCS7_final(PKCS7 *p7, BIO *data, int flags);
 int PKCS7_verify(PKCS7 *p7, STACK_OF(X509) *certs, X509_STORE *store,
 					BIO *indata, BIO *out, int flags);
 STACK_OF(X509) *PKCS7_get0_signers(PKCS7 *p7, STACK_OF(X509) *certs, int flags);
@@ -367,10 +376,16 @@ int PKCS7_add_attrib_smimecap(PKCS7_SIGNER_INFO *si,
 STACK_OF(X509_ALGOR) *PKCS7_get_smimecap(PKCS7_SIGNER_INFO *si);
 int PKCS7_simple_smimecap(STACK_OF(X509_ALGOR) *sk, int nid, int arg);
 
+int PKCS7_add_attrib_content_type(PKCS7_SIGNER_INFO *si, ASN1_OBJECT *coid);
+int PKCS7_add0_attrib_signing_time(PKCS7_SIGNER_INFO *si, ASN1_TIME *t);
+int PKCS7_add1_attrib_digest(PKCS7_SIGNER_INFO *si,
+				const unsigned char *md, int mdlen);
+
 int SMIME_write_PKCS7(BIO *bio, PKCS7 *p7, BIO *data, int flags);
 PKCS7 *SMIME_read_PKCS7(BIO *bio, BIO **bcont);
-int SMIME_crlf_copy(BIO *in, BIO *out, int flags);
-int SMIME_text(BIO *in, BIO *out);
+
+BIO *BIO_new_PKCS7(BIO *out, PKCS7 *p7);
+
 
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
@@ -383,12 +398,17 @@ void ERR_load_PKCS7_strings(void);
 /* Function codes. */
 #define PKCS7_F_B64_READ_PKCS7				 120
 #define PKCS7_F_B64_WRITE_PKCS7				 121
+#define PKCS7_F_DO_PKCS7_SIGNED_ATTRIB			 136
+#define PKCS7_F_I2D_PKCS7_BIO_STREAM			 140
+#define PKCS7_F_PKCS7_ADD0_ATTRIB_SIGNING_TIME		 135
 #define PKCS7_F_PKCS7_ADD_ATTRIB_SMIMECAP		 118
 #define PKCS7_F_PKCS7_ADD_CERTIFICATE			 100
 #define PKCS7_F_PKCS7_ADD_CRL				 101
 #define PKCS7_F_PKCS7_ADD_RECIPIENT_INFO		 102
+#define PKCS7_F_PKCS7_ADD_SIGNATURE			 131
 #define PKCS7_F_PKCS7_ADD_SIGNER			 103
 #define PKCS7_F_PKCS7_BIO_ADD_DIGEST			 125
+#define PKCS7_F_PKCS7_COPY_EXISTING_DIGEST		 138
 #define PKCS7_F_PKCS7_CTRL				 104
 #define PKCS7_F_PKCS7_DATADECODE			 112
 #define PKCS7_F_PKCS7_DATAFINAL				 128
@@ -396,15 +416,22 @@ void ERR_load_PKCS7_strings(void);
 #define PKCS7_F_PKCS7_DATASIGN				 106
 #define PKCS7_F_PKCS7_DATAVERIFY			 107
 #define PKCS7_F_PKCS7_DECRYPT				 114
+#define PKCS7_F_PKCS7_DECRYPT_RINFO			 133
+#define PKCS7_F_PKCS7_ENCODE_RINFO			 132
 #define PKCS7_F_PKCS7_ENCRYPT				 115
+#define PKCS7_F_PKCS7_FINAL				 134
 #define PKCS7_F_PKCS7_FIND_DIGEST			 127
 #define PKCS7_F_PKCS7_GET0_SIGNERS			 124
+#define PKCS7_F_PKCS7_RECIP_INFO_SET			 130
 #define PKCS7_F_PKCS7_SET_CIPHER			 108
 #define PKCS7_F_PKCS7_SET_CONTENT			 109
 #define PKCS7_F_PKCS7_SET_DIGEST			 126
 #define PKCS7_F_PKCS7_SET_TYPE				 110
 #define PKCS7_F_PKCS7_SIGN				 116
 #define PKCS7_F_PKCS7_SIGNATUREVERIFY			 113
+#define PKCS7_F_PKCS7_SIGNER_INFO_SET			 129
+#define PKCS7_F_PKCS7_SIGNER_INFO_SIGN			 139
+#define PKCS7_F_PKCS7_SIGN_ADD_SIGNER			 137
 #define PKCS7_F_PKCS7_SIMPLE_SMIMECAP			 119
 #define PKCS7_F_PKCS7_VERIFY				 117
 #define PKCS7_F_SMIME_READ_PKCS7			 122
@@ -415,10 +442,13 @@ void ERR_load_PKCS7_strings(void);
 #define PKCS7_R_CIPHER_HAS_NO_OBJECT_IDENTIFIER		 144
 #define PKCS7_R_CIPHER_NOT_INITIALIZED			 116
 #define PKCS7_R_CONTENT_AND_DATA_PRESENT		 118
+#define PKCS7_R_CTRL_ERROR				 152
 #define PKCS7_R_DECODE_ERROR				 130
 #define PKCS7_R_DECRYPTED_KEY_IS_WRONG_LENGTH		 100
 #define PKCS7_R_DECRYPT_ERROR				 119
 #define PKCS7_R_DIGEST_FAILURE				 101
+#define PKCS7_R_ENCRYPTION_CTRL_FAILURE			 149
+#define PKCS7_R_ENCRYPTION_NOT_SUPPORTED_FOR_THIS_KEY_TYPE 150
 #define PKCS7_R_ERROR_ADDING_RECIPIENT			 120
 #define PKCS7_R_ERROR_SETTING_CIPHER			 121
 #define PKCS7_R_INVALID_MIME_TYPE			 131
@@ -429,6 +459,8 @@ void ERR_load_PKCS7_strings(void);
 #define PKCS7_R_MISSING_CERIPEND_INFO			 103
 #define PKCS7_R_NO_CONTENT				 122
 #define PKCS7_R_NO_CONTENT_TYPE				 135
+#define PKCS7_R_NO_DEFAULT_DIGEST			 151
+#define PKCS7_R_NO_MATCHING_DIGEST_TYPE_FOUND		 154
 #define PKCS7_R_NO_MULTIPART_BODY_FAILURE		 136
 #define PKCS7_R_NO_MULTIPART_BOUNDARY			 137
 #define PKCS7_R_NO_RECIPIENT_MATCHES_CERTIFICATE	 115
@@ -438,6 +470,7 @@ void ERR_load_PKCS7_strings(void);
 #define PKCS7_R_NO_SIG_CONTENT_TYPE			 138
 #define PKCS7_R_OPERATION_NOT_SUPPORTED_ON_THIS_TYPE	 104
 #define PKCS7_R_PKCS7_ADD_SIGNATURE_ERROR		 124
+#define PKCS7_R_PKCS7_ADD_SIGNER_ERROR			 153
 #define PKCS7_R_PKCS7_DATAFINAL				 126
 #define PKCS7_R_PKCS7_DATAFINAL_ERROR			 125
 #define PKCS7_R_PKCS7_DATASIGN				 145
@@ -446,6 +479,8 @@ void ERR_load_PKCS7_strings(void);
 #define PKCS7_R_PRIVATE_KEY_DOES_NOT_MATCH_CERTIFICATE	 127
 #define PKCS7_R_SIGNATURE_FAILURE			 105
 #define PKCS7_R_SIGNER_CERTIFICATE_NOT_FOUND		 128
+#define PKCS7_R_SIGNING_CTRL_FAILURE			 147
+#define PKCS7_R_SIGNING_NOT_SUPPORTED_FOR_THIS_KEY_TYPE	 148
 #define PKCS7_R_SIG_INVALID_MIME_TYPE			 141
 #define PKCS7_R_SMIME_TEXT_ERROR			 129
 #define PKCS7_R_UNABLE_TO_FIND_CERTIFICATE		 106
diff --git a/crypto/pkcs7/pkcs7err.c b/crypto/pkcs7/pkcs7err.c
index c0e3d4cd3336..d0af32a265df 100644
--- a/crypto/pkcs7/pkcs7err.c
+++ b/crypto/pkcs7/pkcs7err.c
@@ -1,6 +1,6 @@
 /* crypto/pkcs7/pkcs7err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -72,12 +72,17 @@ static ERR_STRING_DATA PKCS7_str_functs[]=
 	{
 {ERR_FUNC(PKCS7_F_B64_READ_PKCS7),	"B64_READ_PKCS7"},
 {ERR_FUNC(PKCS7_F_B64_WRITE_PKCS7),	"B64_WRITE_PKCS7"},
+{ERR_FUNC(PKCS7_F_DO_PKCS7_SIGNED_ATTRIB),	"DO_PKCS7_SIGNED_ATTRIB"},
+{ERR_FUNC(PKCS7_F_I2D_PKCS7_BIO_STREAM),	"i2d_PKCS7_bio_stream"},
+{ERR_FUNC(PKCS7_F_PKCS7_ADD0_ATTRIB_SIGNING_TIME),	"PKCS7_add0_attrib_signing_time"},
 {ERR_FUNC(PKCS7_F_PKCS7_ADD_ATTRIB_SMIMECAP),	"PKCS7_add_attrib_smimecap"},
 {ERR_FUNC(PKCS7_F_PKCS7_ADD_CERTIFICATE),	"PKCS7_add_certificate"},
 {ERR_FUNC(PKCS7_F_PKCS7_ADD_CRL),	"PKCS7_add_crl"},
 {ERR_FUNC(PKCS7_F_PKCS7_ADD_RECIPIENT_INFO),	"PKCS7_add_recipient_info"},
+{ERR_FUNC(PKCS7_F_PKCS7_ADD_SIGNATURE),	"PKCS7_add_signature"},
 {ERR_FUNC(PKCS7_F_PKCS7_ADD_SIGNER),	"PKCS7_add_signer"},
 {ERR_FUNC(PKCS7_F_PKCS7_BIO_ADD_DIGEST),	"PKCS7_BIO_ADD_DIGEST"},
+{ERR_FUNC(PKCS7_F_PKCS7_COPY_EXISTING_DIGEST),	"PKCS7_COPY_EXISTING_DIGEST"},
 {ERR_FUNC(PKCS7_F_PKCS7_CTRL),	"PKCS7_ctrl"},
 {ERR_FUNC(PKCS7_F_PKCS7_DATADECODE),	"PKCS7_dataDecode"},
 {ERR_FUNC(PKCS7_F_PKCS7_DATAFINAL),	"PKCS7_dataFinal"},
@@ -85,15 +90,22 @@ static ERR_STRING_DATA PKCS7_str_functs[]=
 {ERR_FUNC(PKCS7_F_PKCS7_DATASIGN),	"PKCS7_DATASIGN"},
 {ERR_FUNC(PKCS7_F_PKCS7_DATAVERIFY),	"PKCS7_dataVerify"},
 {ERR_FUNC(PKCS7_F_PKCS7_DECRYPT),	"PKCS7_decrypt"},
+{ERR_FUNC(PKCS7_F_PKCS7_DECRYPT_RINFO),	"PKCS7_DECRYPT_RINFO"},
+{ERR_FUNC(PKCS7_F_PKCS7_ENCODE_RINFO),	"PKCS7_ENCODE_RINFO"},
 {ERR_FUNC(PKCS7_F_PKCS7_ENCRYPT),	"PKCS7_encrypt"},
+{ERR_FUNC(PKCS7_F_PKCS7_FINAL),	"PKCS7_final"},
 {ERR_FUNC(PKCS7_F_PKCS7_FIND_DIGEST),	"PKCS7_FIND_DIGEST"},
 {ERR_FUNC(PKCS7_F_PKCS7_GET0_SIGNERS),	"PKCS7_get0_signers"},
+{ERR_FUNC(PKCS7_F_PKCS7_RECIP_INFO_SET),	"PKCS7_RECIP_INFO_set"},
 {ERR_FUNC(PKCS7_F_PKCS7_SET_CIPHER),	"PKCS7_set_cipher"},
 {ERR_FUNC(PKCS7_F_PKCS7_SET_CONTENT),	"PKCS7_set_content"},
 {ERR_FUNC(PKCS7_F_PKCS7_SET_DIGEST),	"PKCS7_set_digest"},
 {ERR_FUNC(PKCS7_F_PKCS7_SET_TYPE),	"PKCS7_set_type"},
 {ERR_FUNC(PKCS7_F_PKCS7_SIGN),	"PKCS7_sign"},
 {ERR_FUNC(PKCS7_F_PKCS7_SIGNATUREVERIFY),	"PKCS7_signatureVerify"},
+{ERR_FUNC(PKCS7_F_PKCS7_SIGNER_INFO_SET),	"PKCS7_SIGNER_INFO_set"},
+{ERR_FUNC(PKCS7_F_PKCS7_SIGNER_INFO_SIGN),	"PKCS7_SIGNER_INFO_sign"},
+{ERR_FUNC(PKCS7_F_PKCS7_SIGN_ADD_SIGNER),	"PKCS7_sign_add_signer"},
 {ERR_FUNC(PKCS7_F_PKCS7_SIMPLE_SMIMECAP),	"PKCS7_simple_smimecap"},
 {ERR_FUNC(PKCS7_F_PKCS7_VERIFY),	"PKCS7_verify"},
 {ERR_FUNC(PKCS7_F_SMIME_READ_PKCS7),	"SMIME_read_PKCS7"},
@@ -107,10 +119,13 @@ static ERR_STRING_DATA PKCS7_str_reasons[]=
 {ERR_REASON(PKCS7_R_CIPHER_HAS_NO_OBJECT_IDENTIFIER),"cipher has no object identifier"},
 {ERR_REASON(PKCS7_R_CIPHER_NOT_INITIALIZED),"cipher not initialized"},
 {ERR_REASON(PKCS7_R_CONTENT_AND_DATA_PRESENT),"content and data present"},
+{ERR_REASON(PKCS7_R_CTRL_ERROR)          ,"ctrl error"},
 {ERR_REASON(PKCS7_R_DECODE_ERROR)        ,"decode error"},
 {ERR_REASON(PKCS7_R_DECRYPTED_KEY_IS_WRONG_LENGTH),"decrypted key is wrong length"},
 {ERR_REASON(PKCS7_R_DECRYPT_ERROR)       ,"decrypt error"},
 {ERR_REASON(PKCS7_R_DIGEST_FAILURE)      ,"digest failure"},
+{ERR_REASON(PKCS7_R_ENCRYPTION_CTRL_FAILURE),"encryption ctrl failure"},
+{ERR_REASON(PKCS7_R_ENCRYPTION_NOT_SUPPORTED_FOR_THIS_KEY_TYPE),"encryption not supported for this key type"},
 {ERR_REASON(PKCS7_R_ERROR_ADDING_RECIPIENT),"error adding recipient"},
 {ERR_REASON(PKCS7_R_ERROR_SETTING_CIPHER),"error setting cipher"},
 {ERR_REASON(PKCS7_R_INVALID_MIME_TYPE)   ,"invalid mime type"},
@@ -121,6 +136,8 @@ static ERR_STRING_DATA PKCS7_str_reasons[]=
 {ERR_REASON(PKCS7_R_MISSING_CERIPEND_INFO),"missing ceripend info"},
 {ERR_REASON(PKCS7_R_NO_CONTENT)          ,"no content"},
 {ERR_REASON(PKCS7_R_NO_CONTENT_TYPE)     ,"no content type"},
+{ERR_REASON(PKCS7_R_NO_DEFAULT_DIGEST)   ,"no default digest"},
+{ERR_REASON(PKCS7_R_NO_MATCHING_DIGEST_TYPE_FOUND),"no matching digest type found"},
 {ERR_REASON(PKCS7_R_NO_MULTIPART_BODY_FAILURE),"no multipart body failure"},
 {ERR_REASON(PKCS7_R_NO_MULTIPART_BOUNDARY),"no multipart boundary"},
 {ERR_REASON(PKCS7_R_NO_RECIPIENT_MATCHES_CERTIFICATE),"no recipient matches certificate"},
@@ -130,6 +147,7 @@ static ERR_STRING_DATA PKCS7_str_reasons[]=
 {ERR_REASON(PKCS7_R_NO_SIG_CONTENT_TYPE) ,"no sig content type"},
 {ERR_REASON(PKCS7_R_OPERATION_NOT_SUPPORTED_ON_THIS_TYPE),"operation not supported on this type"},
 {ERR_REASON(PKCS7_R_PKCS7_ADD_SIGNATURE_ERROR),"pkcs7 add signature error"},
+{ERR_REASON(PKCS7_R_PKCS7_ADD_SIGNER_ERROR),"pkcs7 add signer error"},
 {ERR_REASON(PKCS7_R_PKCS7_DATAFINAL)     ,"pkcs7 datafinal"},
 {ERR_REASON(PKCS7_R_PKCS7_DATAFINAL_ERROR),"pkcs7 datafinal error"},
 {ERR_REASON(PKCS7_R_PKCS7_DATASIGN)      ,"pkcs7 datasign"},
@@ -138,6 +156,8 @@ static ERR_STRING_DATA PKCS7_str_reasons[]=
 {ERR_REASON(PKCS7_R_PRIVATE_KEY_DOES_NOT_MATCH_CERTIFICATE),"private key does not match certificate"},
 {ERR_REASON(PKCS7_R_SIGNATURE_FAILURE)   ,"signature failure"},
 {ERR_REASON(PKCS7_R_SIGNER_CERTIFICATE_NOT_FOUND),"signer certificate not found"},
+{ERR_REASON(PKCS7_R_SIGNING_CTRL_FAILURE),"signing ctrl failure"},
+{ERR_REASON(PKCS7_R_SIGNING_NOT_SUPPORTED_FOR_THIS_KEY_TYPE),"signing not supported for this key type"},
 {ERR_REASON(PKCS7_R_SIG_INVALID_MIME_TYPE),"sig invalid mime type"},
 {ERR_REASON(PKCS7_R_SMIME_TEXT_ERROR)    ,"smime text error"},
 {ERR_REASON(PKCS7_R_UNABLE_TO_FIND_CERTIFICATE),"unable to find certificate"},
diff --git a/crypto/ppccap.c b/crypto/ppccap.c
new file mode 100644
index 000000000000..ab89ccaa12c8
--- /dev/null
+++ b/crypto/ppccap.c
@@ -0,0 +1,115 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <setjmp.h>
+#include <signal.h>
+#include <crypto.h>
+#include <openssl/bn.h>
+
+#define PPC_FPU64	(1<<0)
+#define PPC_ALTIVEC	(1<<1)
+
+static int OPENSSL_ppccap_P = 0;
+
+static sigset_t all_masked;
+
+#ifdef OPENSSL_BN_ASM_MONT
+int bn_mul_mont(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np, const BN_ULONG *n0, int num)
+	{
+	int bn_mul_mont_fpu64(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np, const BN_ULONG *n0, int num);
+	int bn_mul_mont_int(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np, const BN_ULONG *n0, int num);
+
+	if (sizeof(size_t)==4)
+		{
+#if (defined(__APPLE__) && defined(__MACH__))
+		if (num>=8 && (num&3)==0 && (OPENSSL_ppccap_P&PPC_FPU64))
+			return bn_mul_mont_fpu64(rp,ap,bp,np,n0,num);
+#else
+		/* boundary of 32 was experimentally determined on
+		   Linux 2.6.22, might have to be adjusted on AIX... */
+		if (num>=32 && (num&3)==0 && (OPENSSL_ppccap_P&PPC_FPU64))
+			{
+			sigset_t oset;
+			int ret;
+
+			sigprocmask(SIG_SETMASK,&all_masked,&oset);
+			ret=bn_mul_mont_fpu64(rp,ap,bp,np,n0,num);
+			sigprocmask(SIG_SETMASK,&oset,NULL);
+
+			return ret;
+			}
+#endif
+		}
+	else if ((OPENSSL_ppccap_P&PPC_FPU64))
+		/* this is a "must" on POWER6, but run-time detection
+		 * is not implemented yet... */
+		return bn_mul_mont_fpu64(rp,ap,bp,np,n0,num);
+
+	return bn_mul_mont_int(rp,ap,bp,np,n0,num);
+	}
+#endif
+
+static sigjmp_buf ill_jmp;
+static void ill_handler (int sig) { siglongjmp(ill_jmp,sig); }
+
+void OPENSSL_ppc64_probe(void);
+
+void OPENSSL_cpuid_setup(void)
+	{
+	char *e;
+	struct sigaction	ill_oact,ill_act;
+	sigset_t		oset;
+	static int trigger=0;
+
+	if (trigger) return;
+	trigger=1;
+ 
+	sigfillset(&all_masked);
+	sigdelset(&all_masked,SIGILL);
+	sigdelset(&all_masked,SIGTRAP);
+#ifdef SIGEMT
+	sigdelset(&all_masked,SIGEMT);
+#endif
+	sigdelset(&all_masked,SIGFPE);
+	sigdelset(&all_masked,SIGBUS);
+	sigdelset(&all_masked,SIGSEGV);
+
+	if ((e=getenv("OPENSSL_ppccap")))
+		{
+		OPENSSL_ppccap_P=strtoul(e,NULL,0);
+		return;
+		}
+
+	OPENSSL_ppccap_P = 0;
+
+	memset(&ill_act,0,sizeof(ill_act));
+	ill_act.sa_handler = ill_handler;
+	ill_act.sa_mask    = all_masked;
+
+	sigprocmask(SIG_SETMASK,&ill_act.sa_mask,&oset);
+	sigaction(SIGILL,&ill_act,&ill_oact);
+
+	if (sizeof(size_t)==4)
+		{
+		if (sigsetjmp(ill_jmp,1) == 0)
+			{
+			OPENSSL_ppc64_probe();
+			OPENSSL_ppccap_P |= PPC_FPU64;
+			}
+		}
+	else
+		{
+		/*
+		 * Wanted code detecting POWER6 CPU and setting PPC_FPU64
+		 */
+		}
+
+	if (sigsetjmp(ill_jmp,1) == 0)
+		{
+		OPENSSL_altivec_probe();
+		OPENSSL_ppccap_P |= PPC_ALTIVEC;
+		}
+
+	sigaction (SIGILL,&ill_oact,NULL);
+	sigprocmask(SIG_SETMASK,&oset,NULL);
+	}
diff --git a/crypto/ppccpuid.pl b/crypto/ppccpuid.pl
new file mode 100755
index 000000000000..4ba736a1d1bd
--- /dev/null
+++ b/crypto/ppccpuid.pl
@@ -0,0 +1,132 @@
+#!/usr/bin/env perl
+
+$flavour = shift;
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour ".shift || die "can't call $xlate: $!";
+
+if ($flavour=~/64/) {
+    $CMPLI="cmpldi";
+    $SHRLI="srdi";
+    $SIGNX="extsw";
+} else {
+    $CMPLI="cmplwi";
+    $SHRLI="srwi";
+    $SIGNX="mr";
+}
+
+$code=<<___;
+.machine	"any"
+.text
+
+.globl	.OPENSSL_ppc64_probe
+.align	4
+.OPENSSL_ppc64_probe:
+	fcfid	f1,f1
+	extrdi	r0,r0,32,0
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+
+.globl	.OPENSSL_altivec_probe
+.align	4
+.OPENSSL_altivec_probe:
+	.long	0x10000484	# vor	v0,v0,v0
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+
+.globl	.OPENSSL_wipe_cpu
+.align	4
+.OPENSSL_wipe_cpu:
+	xor	r0,r0,r0
+	fmr	f0,f31
+	fmr	f1,f31
+	fmr	f2,f31
+	mr	r3,r1
+	fmr	f3,f31
+	xor	r4,r4,r4
+	fmr	f4,f31
+	xor	r5,r5,r5
+	fmr	f5,f31
+	xor	r6,r6,r6
+	fmr	f6,f31
+	xor	r7,r7,r7
+	fmr	f7,f31
+	xor	r8,r8,r8
+	fmr	f8,f31
+	xor	r9,r9,r9
+	fmr	f9,f31
+	xor	r10,r10,r10
+	fmr	f10,f31
+	xor	r11,r11,r11
+	fmr	f11,f31
+	xor	r12,r12,r12
+	fmr	f12,f31
+	fmr	f13,f31
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+
+.globl	.OPENSSL_atomic_add
+.align	4
+.OPENSSL_atomic_add:
+Ladd:	lwarx	r5,0,r3
+	add	r0,r4,r5
+	stwcx.	r0,0,r3
+	bne-	Ladd
+	$SIGNX	r3,r0
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,2,0
+	.long	0
+
+.globl	.OPENSSL_rdtsc
+.align	4
+.OPENSSL_rdtsc:
+	mftb	r3
+	mftbu	r4
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+
+.globl	.OPENSSL_cleanse
+.align	4
+.OPENSSL_cleanse:
+	$CMPLI	r4,7
+	li	r0,0
+	bge	Lot
+	$CMPLI	r4,0
+	beqlr-
+Little:	mtctr	r4
+	stb	r0,0(r3)
+	addi	r3,r3,1
+	bdnz	\$-8
+	blr
+Lot:	andi.	r5,r3,3
+	beq	Laligned
+	stb	r0,0(r3)
+	subi	r4,r4,1
+	addi	r3,r3,1
+	b	Lot
+Laligned:
+	$SHRLI	r5,r4,2
+	mtctr	r5
+	stw	r0,0(r3)
+	addi	r3,r3,4
+	bdnz	\$-8
+	andi.	r4,r4,3
+	bne	Little
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,2,0
+	.long	0
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/pqueue/Makefile b/crypto/pqueue/Makefile
index 36bfc349aabc..fb36a0c876e3 100644
--- a/crypto/pqueue/Makefile
+++ b/crypto/pqueue/Makefile
@@ -22,7 +22,7 @@ LIBOBJ=pqueue.o
 
 SRC= $(LIBSRC)
 
-EXHEADER= pqueue.h pq_compat.h
+EXHEADER= pqueue.h
 HEADER=	$(EXHEADER)
 
 ALL=    $(GENERAL) $(SRC) $(HEADER)
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -79,6 +79,5 @@ pqueue.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 pqueue.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
 pqueue.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 pqueue.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-pqueue.o: ../../include/openssl/pq_compat.h ../../include/openssl/safestack.h
-pqueue.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-pqueue.o: ../cryptlib.h pqueue.c pqueue.h
+pqueue.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+pqueue.o: ../../include/openssl/symhacks.h ../cryptlib.h pqueue.c pqueue.h
diff --git a/crypto/pqueue/pq_compat.h b/crypto/pqueue/pq_compat.h
deleted file mode 100644
index 7b2c32725cd0..000000000000
--- a/crypto/pqueue/pq_compat.h
+++ /dev/null
@@ -1,152 +0,0 @@
-/* crypto/pqueue/pqueue_compat.h */
-/* 
- * DTLS implementation written by Nagendra Modadugu
- * (nagendra@cs.stanford.edu) for the OpenSSL project 2005.  
- */
-/* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    openssl-core@OpenSSL.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
- */
-
-#ifndef HEADER_PQ_COMPAT_H
-#define HEADER_PQ_COMPAT_H
-
-#include <openssl/opensslconf.h>
-#include <openssl/bn.h>
-
-/* 
- * The purpose of this header file is for supporting 64-bit integer
- * manipulation on 32-bit (and lower) machines.  Currently the only
- * such environment is VMS, Utrix and those with smaller default integer
- * sizes than 32 bits.  For all such environment, we fall back to using
- * BIGNUM.  We may need to fine tune the conditions for systems that
- * are incorrectly configured.
- *
- * The only clients of this code are (1) pqueue for priority, and
- * (2) DTLS, for sequence number manipulation.
- */
-
-#if (defined(THIRTY_TWO_BIT) && !defined(BN_LLONG)) || defined(SIXTEEN_BIT) || defined(EIGHT_BIT)
-
-#define PQ_64BIT_IS_INTEGER 0
-#define PQ_64BIT_IS_BIGNUM 1
-
-#define PQ_64BIT     BIGNUM
-#define PQ_64BIT_CTX BN_CTX
-
-#define pq_64bit_init(x)           BN_init(x)
-#define pq_64bit_free(x)           BN_free(x)
-
-#define pq_64bit_ctx_new(ctx)      BN_CTX_new()
-#define pq_64bit_ctx_free(x)       BN_CTX_free(x)
-
-#define pq_64bit_assign(x, y)      BN_copy(x, y)
-#define pq_64bit_assign_word(x, y) BN_set_word(x, y)
-#define pq_64bit_gt(x, y)          BN_ucmp(x, y) >= 1 ? 1 : 0
-#define pq_64bit_eq(x, y)          BN_ucmp(x, y) == 0 ? 1 : 0
-#define pq_64bit_add_word(x, w)    BN_add_word(x, w)
-#define pq_64bit_sub(r, x, y)      BN_sub(r, x, y)
-#define pq_64bit_sub_word(x, w)    BN_sub_word(x, w)
-#define pq_64bit_mod(r, x, n, ctx) BN_mod(r, x, n, ctx)
-
-#define pq_64bit_bin2num(bn, bytes, len)   BN_bin2bn(bytes, len, bn)
-#define pq_64bit_num2bin(bn, bytes)        BN_bn2bin(bn, bytes)
-#define pq_64bit_get_word(x)               BN_get_word(x)
-#define pq_64bit_is_bit_set(x, offset)     BN_is_bit_set(x, offset)
-#define pq_64bit_lshift(r, x, shift)       BN_lshift(r, x, shift)
-#define pq_64bit_set_bit(x, num)           BN_set_bit(x, num)
-#define pq_64bit_get_length(x)             BN_num_bits((x))
-
-#else
-
-#define PQ_64BIT_IS_INTEGER 1
-#define PQ_64BIT_IS_BIGNUM 0
-
-#if defined(SIXTY_FOUR_BIT)
-#define PQ_64BIT BN_ULONG
-#define PQ_64BIT_PRINT "%lld"
-#elif defined(SIXTY_FOUR_BIT_LONG)
-#define PQ_64BIT BN_ULONG
-#define PQ_64BIT_PRINT "%ld"
-#elif defined(THIRTY_TWO_BIT)
-#define PQ_64BIT BN_ULLONG
-#define PQ_64BIT_PRINT "%lld"
-#endif
-
-#define PQ_64BIT_CTX      void
-
-#define pq_64bit_init(x)
-#define pq_64bit_free(x)
-#define pq_64bit_ctx_new(ctx)        (ctx)
-#define pq_64bit_ctx_free(x)
-
-#define pq_64bit_assign(x, y)        (*(x) = *(y))
-#define pq_64bit_assign_word(x, y)   (*(x) = y)
-#define pq_64bit_gt(x, y)	         (*(x) > *(y))
-#define pq_64bit_eq(x, y)            (*(x) == *(y))
-#define pq_64bit_add_word(x, w)      (*(x) = (*(x) + (w)))
-#define pq_64bit_sub(r, x, y)        (*(r) = (*(x) - *(y)))
-#define pq_64bit_sub_word(x, w)      (*(x) = (*(x) - (w)))
-#define pq_64bit_mod(r, x, n, ctx)
-
-#define pq_64bit_bin2num(num, bytes, len) bytes_to_long_long(bytes, num)
-#define pq_64bit_num2bin(num, bytes)      long_long_to_bytes(num, bytes)
-#define pq_64bit_get_word(x)              *(x)
-#define pq_64bit_lshift(r, x, shift)      (*(r) = (*(x) << (shift)))
-#define pq_64bit_set_bit(x, num)          do { \
-                                              PQ_64BIT mask = 1; \
-                                              mask = mask << (num); \
-                                              *(x) |= mask; \
-                                          } while(0)
-#endif /* OPENSSL_SYS_VMS */
-
-#endif
diff --git a/crypto/pqueue/pqueue.c b/crypto/pqueue/pqueue.c
index 8ebba8a8564d..eab13a125013 100644
--- a/crypto/pqueue/pqueue.c
+++ b/crypto/pqueue/pqueue.c
@@ -68,13 +68,12 @@ typedef struct _pqueue
 	} pqueue_s;
 
 pitem *
-pitem_new(PQ_64BIT priority, void *data)
+pitem_new(unsigned char *prio64be, void *data)
 	{
 	pitem *item = (pitem *) OPENSSL_malloc(sizeof(pitem));
 	if (item == NULL) return NULL;
 
-	pq_64bit_init(&(item->priority));
-	pq_64bit_assign(&item->priority, &priority);
+	memcpy(item->priority,prio64be,sizeof(item->priority));
 
 	item->data = data;
 	item->next = NULL;
@@ -87,7 +86,6 @@ pitem_free(pitem *item)
 	{
 	if (item == NULL) return;
 
-	pq_64bit_free(&(item->priority));
 	OPENSSL_free(item);
 	}
 
@@ -124,7 +122,10 @@ pqueue_insert(pqueue_s *pq, pitem *item)
 		next != NULL;
 		curr = next, next = next->next)
 		{
-		if (pq_64bit_gt(&(next->priority), &(item->priority)))
+		/* we can compare 64-bit value in big-endian encoding
+		 * with memcmp:-) */
+		int cmp = memcmp(next->priority, item->priority,8);
+		if (cmp > 0)		/* next > item */
 			{
 			item->next = next;
 
@@ -135,8 +136,8 @@ pqueue_insert(pqueue_s *pq, pitem *item)
 
 			return item;
 			}
-		/* duplicates not allowed */
-		if (pq_64bit_eq(&(item->priority), &(next->priority)))
+		
+		else if (cmp == 0)	/* duplicates not allowed */
 			return NULL;
 		}
 
@@ -164,7 +165,7 @@ pqueue_pop(pqueue_s *pq)
 	}
 
 pitem *
-pqueue_find(pqueue_s *pq, PQ_64BIT priority)
+pqueue_find(pqueue_s *pq, unsigned char *prio64be)
 	{
 	pitem *next;
 	pitem *found = NULL;
@@ -174,7 +175,7 @@ pqueue_find(pqueue_s *pq, PQ_64BIT priority)
 
 	for ( next = pq->items; next->next != NULL; next = next->next)
 		{
-		if ( pq_64bit_eq(&(next->priority), &priority))
+		if ( memcmp(next->priority, prio64be,8) == 0)
 			{
 			found = next;
 			break;
@@ -182,16 +183,22 @@ pqueue_find(pqueue_s *pq, PQ_64BIT priority)
 		}
 	
 	/* check the one last node */
-	if ( pq_64bit_eq(&(next->priority), &priority))
+	if ( memcmp(next->priority, prio64be,8) ==0)
 		found = next;
 
 	if ( ! found)
 		return NULL;
 
+#if 0 /* find works in peek mode */
+	if ( prev == NULL)
+		pq->items = next->next;
+	else
+		prev->next = next->next;
+#endif
+
 	return found;
 	}
 
-#if PQ_64BIT_IS_INTEGER
 void
 pqueue_print(pqueue_s *pq)
 	{
@@ -199,11 +206,14 @@ pqueue_print(pqueue_s *pq)
 
 	while(item != NULL)
 		{
-		printf("item\t" PQ_64BIT_PRINT "\n", item->priority);
+		printf("item\t%02x%02x%02x%02x%02x%02x%02x%02x\n",
+			item->priority[0],item->priority[1],
+			item->priority[2],item->priority[3],
+			item->priority[4],item->priority[5],
+			item->priority[6],item->priority[7]);
 		item = item->next;
 		}
 	}
-#endif
 
 pitem *
 pqueue_iterator(pqueue_s *pq)
diff --git a/crypto/pqueue/pqueue.h b/crypto/pqueue/pqueue.h
index 16c4072681c2..87fc9037c8f3 100644
--- a/crypto/pqueue/pqueue.h
+++ b/crypto/pqueue/pqueue.h
@@ -64,20 +64,18 @@
 #include <stdlib.h>
 #include <string.h>
 
-#include <openssl/pq_compat.h>
-
 typedef struct _pqueue *pqueue;
 
 typedef struct _pitem
 	{
-	PQ_64BIT priority;
+	unsigned char priority[8]; /* 64-bit value in big-endian encoding */
 	void *data;
 	struct _pitem *next;
 	} pitem;
 
 typedef struct _pitem *piterator;
 
-pitem *pitem_new(PQ_64BIT priority, void *data);
+pitem *pitem_new(unsigned char *prio64be, void *data);
 void   pitem_free(pitem *item);
 
 pqueue pqueue_new(void);
@@ -86,7 +84,7 @@ void   pqueue_free(pqueue pq);
 pitem *pqueue_insert(pqueue pq, pitem *item);
 pitem *pqueue_peek(pqueue pq);
 pitem *pqueue_pop(pqueue pq);
-pitem *pqueue_find(pqueue pq, PQ_64BIT priority);
+pitem *pqueue_find(pqueue pq, unsigned char *prio64be);
 pitem *pqueue_iterator(pqueue pq);
 pitem *pqueue_next(piterator *iter);
 
diff --git a/crypto/rand/Makefile b/crypto/rand/Makefile
index 30794305cb39..27694aa66498 100644
--- a/crypto/rand/Makefile
+++ b/crypto/rand/Makefile
@@ -17,9 +17,9 @@ TEST= randtest.c
 APPS=
 
 LIB=$(TOP)/libcrypto.a
-LIBSRC=md_rand.c randfile.c rand_lib.c rand_eng.c rand_err.c rand_egd.c \
+LIBSRC=md_rand.c randfile.c rand_lib.c rand_err.c rand_egd.c \
 	rand_win.c rand_unix.c rand_os2.c rand_nw.c
-LIBOBJ=md_rand.o randfile.o rand_lib.o rand_eng.o rand_err.o rand_egd.o \
+LIBOBJ=md_rand.o randfile.o rand_lib.o rand_err.o rand_egd.o \
 	rand_win.o rand_unix.o rand_os2.o rand_nw.o
 
 SRC= $(LIBSRC)
@@ -35,7 +35,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -79,34 +79,17 @@ clean:
 md_rand.o: ../../e_os.h ../../include/openssl/asn1.h
 md_rand.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
 md_rand.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-md_rand.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-md_rand.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-md_rand.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-md_rand.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-md_rand.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
-md_rand.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-md_rand.o: ../../include/openssl/symhacks.h md_rand.c rand_lcl.h
+md_rand.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+md_rand.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+md_rand.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+md_rand.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
+md_rand.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+md_rand.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+md_rand.o: md_rand.c rand_lcl.h
 rand_egd.o: ../../include/openssl/buffer.h ../../include/openssl/e_os2.h
 rand_egd.o: ../../include/openssl/opensslconf.h
 rand_egd.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
 rand_egd.o: rand_egd.c
-rand_eng.o: ../../e_os.h ../../include/openssl/asn1.h
-rand_eng.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
-rand_eng.o: ../../include/openssl/crypto.h ../../include/openssl/des.h
-rand_eng.o: ../../include/openssl/des_old.h ../../include/openssl/e_os2.h
-rand_eng.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
-rand_eng.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
-rand_eng.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-rand_eng.o: ../../include/openssl/fips.h ../../include/openssl/fips_rand.h
-rand_eng.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-rand_eng.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-rand_eng.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-rand_eng.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
-rand_eng.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-rand_eng.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-rand_eng.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
-rand_eng.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-rand_eng.o: ../cryptlib.h rand_eng.c rand_lcl.h
 rand_err.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
 rand_err.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
 rand_err.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
@@ -116,39 +99,34 @@ rand_err.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 rand_err.o: rand_err.c
 rand_lib.o: ../../e_os.h ../../include/openssl/asn1.h
 rand_lib.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
-rand_lib.o: ../../include/openssl/crypto.h ../../include/openssl/des.h
-rand_lib.o: ../../include/openssl/des_old.h ../../include/openssl/e_os2.h
+rand_lib.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 rand_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 rand_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 rand_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-rand_lib.o: ../../include/openssl/fips.h ../../include/openssl/fips_rand.h
 rand_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
 rand_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 rand_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rand_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
 rand_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 rand_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-rand_lib.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
 rand_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-rand_lib.o: ../cryptlib.h rand_lcl.h rand_lib.c
+rand_lib.o: ../cryptlib.h rand_lib.c
 rand_nw.o: ../../e_os.h ../../include/openssl/asn1.h
 rand_nw.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 rand_nw.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 rand_nw.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-rand_nw.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-rand_nw.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-rand_nw.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-rand_nw.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
-rand_nw.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-rand_nw.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-rand_nw.o: ../cryptlib.h rand_lcl.h rand_nw.c
+rand_nw.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+rand_nw.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+rand_nw.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+rand_nw.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
+rand_nw.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+rand_nw.o: ../../include/openssl/symhacks.h ../cryptlib.h rand_lcl.h rand_nw.c
 rand_os2.o: ../../e_os.h ../../include/openssl/asn1.h
 rand_os2.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 rand_os2.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 rand_os2.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-rand_os2.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-rand_os2.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-rand_os2.o: ../../include/openssl/opensslconf.h
+rand_os2.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+rand_os2.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 rand_os2.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rand_os2.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
 rand_os2.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -158,8 +136,8 @@ rand_unix.o: ../../e_os.h ../../include/openssl/asn1.h
 rand_unix.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 rand_unix.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 rand_unix.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-rand_unix.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-rand_unix.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+rand_unix.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+rand_unix.o: ../../include/openssl/objects.h
 rand_unix.o: ../../include/openssl/opensslconf.h
 rand_unix.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rand_unix.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
@@ -170,9 +148,8 @@ rand_win.o: ../../e_os.h ../../include/openssl/asn1.h
 rand_win.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 rand_win.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 rand_win.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-rand_win.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-rand_win.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-rand_win.o: ../../include/openssl/opensslconf.h
+rand_win.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+rand_win.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 rand_win.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rand_win.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
 rand_win.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
diff --git a/crypto/rand/md_rand.c b/crypto/rand/md_rand.c
index 0f8dd3e00faf..fcdd3f2a845f 100644
--- a/crypto/rand/md_rand.c
+++ b/crypto/rand/md_rand.c
@@ -109,6 +109,8 @@
  *
  */
 
+#define OPENSSL_FIPSEVP
+
 #ifdef MD_RAND_DEBUG
 # ifndef NDEBUG
 #   define NDEBUG
@@ -126,10 +128,6 @@
 
 #include <openssl/crypto.h>
 #include <openssl/err.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 
 #ifdef BN_DEBUG
 # define PREDICT
@@ -149,7 +147,7 @@ static unsigned int crypto_lock_rand = 0; /* may be set only when a thread
                                            * holds CRYPTO_LOCK_RAND
                                            * (to prevent double locking) */
 /* access to lockin_thread is synchronized by CRYPTO_LOCK_RAND2 */
-static unsigned long locking_thread = 0; /* valid iff crypto_lock_rand is set */
+static CRYPTO_THREADID locking_threadid; /* valid iff crypto_lock_rand is set */
 
 
 #ifdef PREDICT
@@ -161,13 +159,14 @@ const char RAND_version[]="RAND" OPENSSL_VERSION_PTEXT;
 static void ssleay_rand_cleanup(void);
 static void ssleay_rand_seed(const void *buf, int num);
 static void ssleay_rand_add(const void *buf, int num, double add_entropy);
-static int ssleay_rand_bytes(unsigned char *buf, int num);
+static int ssleay_rand_bytes(unsigned char *buf, int num, int pseudo);
+static int ssleay_rand_nopseudo_bytes(unsigned char *buf, int num);
 static int ssleay_rand_pseudo_bytes(unsigned char *buf, int num);
 static int ssleay_rand_status(void);
 
 RAND_METHOD rand_ssleay_meth={
 	ssleay_rand_seed,
-	ssleay_rand_bytes,
+	ssleay_rand_nopseudo_bytes,
 	ssleay_rand_cleanup,
 	ssleay_rand_add,
 	ssleay_rand_pseudo_bytes,
@@ -217,8 +216,10 @@ static void ssleay_rand_add(const void *buf, int num, double add)
 	/* check if we already have the lock */
 	if (crypto_lock_rand)
 		{
+		CRYPTO_THREADID cur;
+		CRYPTO_THREADID_current(&cur);
 		CRYPTO_r_lock(CRYPTO_LOCK_RAND2);
-		do_not_lock = (locking_thread == CRYPTO_thread_id());
+		do_not_lock = !CRYPTO_THREADID_cmp(&locking_threadid, &cur);
 		CRYPTO_r_unlock(CRYPTO_LOCK_RAND2);
 		}
 	else
@@ -274,8 +275,16 @@ static void ssleay_rand_add(const void *buf, int num, double add)
 			}
 		else
 			MD_Update(&m,&(state[st_idx]),j);
-			
+
+		/* DO NOT REMOVE THE FOLLOWING CALL TO MD_Update()! */
 		MD_Update(&m,buf,j);
+		/* We know that line may cause programs such as
+		   purify and valgrind to complain about use of
+		   uninitialized data.  The problem is not, it's
+		   with the caller.  Removing that line will make
+		   sure you get really bad randomness and thereby
+		   other problems such as very insecure keys. */
+
 		MD_Update(&m,(unsigned char *)&(md_c[0]),sizeof(md_c));
 		MD_Final(&m,local_md);
 		md_c[1]++;
@@ -322,7 +331,7 @@ static void ssleay_rand_seed(const void *buf, int num)
 	ssleay_rand_add(buf, num, (double)num);
 	}
 
-static int ssleay_rand_bytes(unsigned char *buf, int num)
+static int ssleay_rand_bytes(unsigned char *buf, int num, int pseudo)
 	{
 	static volatile int stirred_pool = 0;
 	int i,j,k,st_num,st_idx;
@@ -336,14 +345,6 @@ static int ssleay_rand_bytes(unsigned char *buf, int num)
 #endif
 	int do_stir_pool = 0;
 
-#ifdef OPENSSL_FIPS
-	if(FIPS_mode())
-	    {
-	    FIPSerr(FIPS_F_SSLEAY_RAND_BYTES,FIPS_R_NON_FIPS_METHOD);
-	    return 0;
-	    }
-#endif
-
 #ifdef PREDICT
 	if (rand_predictable)
 		{
@@ -384,7 +385,7 @@ static int ssleay_rand_bytes(unsigned char *buf, int num)
 
 	/* prevent ssleay_rand_bytes() from trying to obtain the lock again */
 	CRYPTO_w_lock(CRYPTO_LOCK_RAND2);
-	locking_thread = CRYPTO_thread_id();
+	CRYPTO_THREADID_current(&locking_threadid);
 	CRYPTO_w_unlock(CRYPTO_LOCK_RAND2);
 	crypto_lock_rand = 1;
 
@@ -476,9 +477,18 @@ static int ssleay_rand_bytes(unsigned char *buf, int num)
 #endif
 		MD_Update(&m,local_md,MD_DIGEST_LENGTH);
 		MD_Update(&m,(unsigned char *)&(md_c[0]),sizeof(md_c));
-#ifndef PURIFY
-		MD_Update(&m,buf,j); /* purify complains */
+
+#ifndef PURIFY /* purify complains */
+		/* The following line uses the supplied buffer as a small
+		 * source of entropy: since this buffer is often uninitialised
+		 * it may cause programs such as purify or valgrind to
+		 * complain. So for those builds it is not used: the removal
+		 * of such a small source of entropy has negligible impact on
+		 * security.
+		 */
+		MD_Update(&m,buf,j);
 #endif
+
 		k=(st_idx+MD_DIGEST_LENGTH/2)-st_num;
 		if (k > 0)
 			{
@@ -510,7 +520,9 @@ static int ssleay_rand_bytes(unsigned char *buf, int num)
 	EVP_MD_CTX_cleanup(&m);
 	if (ok)
 		return(1);
-	else
+	else if (pseudo)
+		return 0;
+	else 
 		{
 		RANDerr(RAND_F_SSLEAY_RAND_BYTES,RAND_R_PRNG_NOT_SEEDED);
 		ERR_add_error_data(1, "You need to read the OpenSSL FAQ, "
@@ -519,35 +531,31 @@ static int ssleay_rand_bytes(unsigned char *buf, int num)
 		}
 	}
 
+static int ssleay_rand_nopseudo_bytes(unsigned char *buf, int num)
+	{
+	return ssleay_rand_bytes(buf, num, 0);
+	}
+
 /* pseudo-random bytes that are guaranteed to be unique but not
    unpredictable */
 static int ssleay_rand_pseudo_bytes(unsigned char *buf, int num) 
 	{
-	int ret;
-	unsigned long err;
-
-	ret = RAND_bytes(buf, num);
-	if (ret == 0)
-		{
-		err = ERR_peek_error();
-		if (ERR_GET_LIB(err) == ERR_LIB_RAND &&
-		    ERR_GET_REASON(err) == RAND_R_PRNG_NOT_SEEDED)
-			ERR_clear_error();
-		}
-	return (ret);
+	return ssleay_rand_bytes(buf, num, 1);
 	}
 
 static int ssleay_rand_status(void)
 	{
+	CRYPTO_THREADID cur;
 	int ret;
 	int do_not_lock;
 
+	CRYPTO_THREADID_current(&cur);
 	/* check if we already have the lock
 	 * (could happen if a RAND_poll() implementation calls RAND_status()) */
 	if (crypto_lock_rand)
 		{
 		CRYPTO_r_lock(CRYPTO_LOCK_RAND2);
-		do_not_lock = (locking_thread == CRYPTO_thread_id());
+		do_not_lock = !CRYPTO_THREADID_cmp(&locking_threadid, &cur);
 		CRYPTO_r_unlock(CRYPTO_LOCK_RAND2);
 		}
 	else
@@ -559,7 +567,7 @@ static int ssleay_rand_status(void)
 		
 		/* prevent ssleay_rand_bytes() from trying to obtain the lock again */
 		CRYPTO_w_lock(CRYPTO_LOCK_RAND2);
-		locking_thread = CRYPTO_thread_id();
+		CRYPTO_THREADID_cpy(&locking_threadid, &cur);
 		CRYPTO_w_unlock(CRYPTO_LOCK_RAND2);
 		crypto_lock_rand = 1;
 		}
diff --git a/crypto/rand/rand.h b/crypto/rand/rand.h
index ea89153cba34..dc8fcf94c5aa 100644
--- a/crypto/rand/rand.h
+++ b/crypto/rand/rand.h
@@ -72,7 +72,7 @@ extern "C" {
 #endif
 
 #if defined(OPENSSL_FIPS)
-#define FIPS_RAND_SIZE_T int
+#define FIPS_RAND_SIZE_T size_t
 #endif
 
 /* Already defined in ossl_typ.h */
@@ -111,15 +111,6 @@ int RAND_query_egd_bytes(const char *path, unsigned char *buf, int bytes);
 int RAND_egd(const char *path);
 int RAND_egd_bytes(const char *path,int bytes);
 int RAND_poll(void);
-#ifndef OPENSSL_NO_ENGINE
-#ifdef OPENSSL_FIPS
-void int_RAND_init_engine_callbacks(void);
-void int_RAND_set_callbacks(
-	int (*set_rand_func)(const RAND_METHOD *meth,
-						const RAND_METHOD **pmeth),
-	const RAND_METHOD *(*get_rand_func)(const RAND_METHOD **pmeth));
-#endif
-#endif
 
 #if defined(OPENSSL_SYS_WINDOWS) || defined(OPENSSL_SYS_WIN32)
 
@@ -128,6 +119,11 @@ int RAND_event(UINT, WPARAM, LPARAM);
 
 #endif
 
+#ifdef OPENSSL_FIPS
+void RAND_set_fips_drbg_type(int type, int flags);
+int RAND_init_fips(void);
+#endif
+
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
  * made after this point may be overwritten when the script is next run.
@@ -137,29 +133,15 @@ void ERR_load_RAND_strings(void);
 /* Error codes for the RAND functions. */
 
 /* Function codes. */
-#define RAND_F_ENG_RAND_GET_RAND_METHOD			 108
-#define RAND_F_FIPS_RAND				 103
-#define RAND_F_FIPS_RAND_BYTES				 102
-#define RAND_F_FIPS_RAND_GET_RAND_METHOD		 109
-#define RAND_F_FIPS_RAND_SET_DT				 106
-#define RAND_F_FIPS_SET_DT				 104
-#define RAND_F_FIPS_SET_PRNG_SEED			 107
-#define RAND_F_FIPS_SET_TEST_MODE			 105
 #define RAND_F_RAND_GET_RAND_METHOD			 101
+#define RAND_F_RAND_INIT_FIPS				 102
 #define RAND_F_SSLEAY_RAND_BYTES			 100
 
 /* Reason codes. */
-#define RAND_R_NON_FIPS_METHOD				 105
-#define RAND_R_NOT_IN_TEST_MODE				 106
-#define RAND_R_NO_KEY_SET				 107
-#define RAND_R_PRNG_ASKING_FOR_TOO_MUCH			 101
-#define RAND_R_PRNG_ERROR				 108
-#define RAND_R_PRNG_KEYED				 109
-#define RAND_R_PRNG_NOT_REKEYED				 102
-#define RAND_R_PRNG_NOT_RESEEDED			 103
+#define RAND_R_ERROR_INITIALISING_DRBG			 102
+#define RAND_R_ERROR_INSTANTIATING_DRBG			 103
+#define RAND_R_NO_FIPS_RANDOM_METHOD_SET		 101
 #define RAND_R_PRNG_NOT_SEEDED				 100
-#define RAND_R_PRNG_SEED_MUST_NOT_MATCH_KEY		 110
-#define RAND_R_PRNG_STUCK				 104
 
 #ifdef  __cplusplus
 }
diff --git a/crypto/rand/rand_egd.c b/crypto/rand/rand_egd.c
index 50bce6cabae8..d53b916ebee8 100644
--- a/crypto/rand/rand_egd.c
+++ b/crypto/rand/rand_egd.c
@@ -95,7 +95,7 @@
  *   RAND_egd() is a wrapper for RAND_egd_bytes() with numbytes=255.
  */
 
-#if defined(OPENSSL_SYS_WIN32) || defined(OPENSSL_SYS_VMS) || defined(OPENSSL_SYS_MSDOS) || defined(OPENSSL_SYS_VXWORKS) || defined(OPENSSL_SYS_NETWARE) || defined(OPENSSL_SYS_VOS)
+#if defined(OPENSSL_SYS_WIN32) || defined(OPENSSL_SYS_VMS) || defined(OPENSSL_SYS_MSDOS) || defined(OPENSSL_SYS_VXWORKS) || defined(OPENSSL_SYS_NETWARE) || defined(OPENSSL_SYS_VOS) || defined(OPENSSL_SYS_BEOS)
 int RAND_query_egd_bytes(const char *path, unsigned char *buf, int bytes)
 	{
 	return(-1);
diff --git a/crypto/rand/rand_eng.c b/crypto/rand/rand_eng.c
deleted file mode 100644
index 1669cef43c49..000000000000
--- a/crypto/rand/rand_eng.c
+++ /dev/null
@@ -1,152 +0,0 @@
-/* crypto/rand/rand_lib.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#include <stdio.h>
-#include <time.h>
-#include "cryptlib.h"
-#include "rand_lcl.h"
-#include <openssl/rand.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#include <openssl/fips_rand.h>
-#endif
-
-#ifndef OPENSSL_NO_ENGINE
-#include <openssl/engine.h>
-#endif
-
-#if defined(OPENSSL_FIPS) && !defined(OPENSSL_NO_ENGINE)
-
-/* non-NULL if default_RAND_meth is ENGINE-provided */
-static ENGINE *funct_ref =NULL;
-
-int eng_RAND_set_rand_method(const RAND_METHOD *meth, const RAND_METHOD **pmeth)
-	{
-	if(funct_ref)
-		{
-		ENGINE_finish(funct_ref);
-		funct_ref = NULL;
-		}
-	*pmeth = meth;
-	return 1;
-	}
-
-const RAND_METHOD *eng_RAND_get_rand_method(const RAND_METHOD **pmeth)
-	{
-	if (!*pmeth)
-		{
-		ENGINE *e = ENGINE_get_default_RAND();
-		if(e)
-			{
-			*pmeth = ENGINE_get_RAND(e);
-			if(!*pmeth)
-				{
-				ENGINE_finish(e);
-				e = NULL;
-				}
-			}
-		if(e)
-			funct_ref = e;
-		else
-			if(FIPS_mode())
-				*pmeth=FIPS_rand_method();
-			else
-			*pmeth = RAND_SSLeay();
-		}
-
-	if(FIPS_mode()
-		&& *pmeth != FIPS_rand_check())
-	    {
-	    RANDerr(RAND_F_ENG_RAND_GET_RAND_METHOD,RAND_R_NON_FIPS_METHOD);
-	    return 0;
-	    }
-
-	return *pmeth;
-	}
-
-int RAND_set_rand_engine(ENGINE *engine)
-	{
-	const RAND_METHOD *tmp_meth = NULL;
-	if(engine)
-		{
-		if(!ENGINE_init(engine))
-			return 0;
-		tmp_meth = ENGINE_get_RAND(engine);
-		if(!tmp_meth)
-			{
-			ENGINE_finish(engine);
-			return 0;
-			}
-		}
-	/* This function releases any prior ENGINE so call it first */
-	RAND_set_rand_method(tmp_meth);
-	funct_ref = engine;
-	return 1;
-	}
-
-void int_RAND_init_engine_callbacks(void)
-	{
-	static int done = 0;
-	if (done)
-		return;
-	int_RAND_set_callbacks(eng_RAND_set_rand_method,
-				 eng_RAND_get_rand_method);
-	done = 1;
-	}
-
-#endif
diff --git a/crypto/rand/rand_err.c b/crypto/rand/rand_err.c
index 829fb44d77ea..b8586c8f4a9c 100644
--- a/crypto/rand/rand_err.c
+++ b/crypto/rand/rand_err.c
@@ -1,6 +1,6 @@
 /* crypto/rand/rand_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,32 +70,18 @@
 
 static ERR_STRING_DATA RAND_str_functs[]=
 	{
-{ERR_FUNC(RAND_F_ENG_RAND_GET_RAND_METHOD),	"ENG_RAND_GET_RAND_METHOD"},
-{ERR_FUNC(RAND_F_FIPS_RAND),	"FIPS_RAND"},
-{ERR_FUNC(RAND_F_FIPS_RAND_BYTES),	"FIPS_RAND_BYTES"},
-{ERR_FUNC(RAND_F_FIPS_RAND_GET_RAND_METHOD),	"FIPS_RAND_GET_RAND_METHOD"},
-{ERR_FUNC(RAND_F_FIPS_RAND_SET_DT),	"FIPS_RAND_SET_DT"},
-{ERR_FUNC(RAND_F_FIPS_SET_DT),	"FIPS_SET_DT"},
-{ERR_FUNC(RAND_F_FIPS_SET_PRNG_SEED),	"FIPS_SET_PRNG_SEED"},
-{ERR_FUNC(RAND_F_FIPS_SET_TEST_MODE),	"FIPS_SET_TEST_MODE"},
 {ERR_FUNC(RAND_F_RAND_GET_RAND_METHOD),	"RAND_get_rand_method"},
+{ERR_FUNC(RAND_F_RAND_INIT_FIPS),	"RAND_init_fips"},
 {ERR_FUNC(RAND_F_SSLEAY_RAND_BYTES),	"SSLEAY_RAND_BYTES"},
 {0,NULL}
 	};
 
 static ERR_STRING_DATA RAND_str_reasons[]=
 	{
-{ERR_REASON(RAND_R_NON_FIPS_METHOD)      ,"non fips method"},
-{ERR_REASON(RAND_R_NOT_IN_TEST_MODE)     ,"not in test mode"},
-{ERR_REASON(RAND_R_NO_KEY_SET)           ,"no key set"},
-{ERR_REASON(RAND_R_PRNG_ASKING_FOR_TOO_MUCH),"prng asking for too much"},
-{ERR_REASON(RAND_R_PRNG_ERROR)           ,"prng error"},
-{ERR_REASON(RAND_R_PRNG_KEYED)           ,"prng keyed"},
-{ERR_REASON(RAND_R_PRNG_NOT_REKEYED)     ,"prng not rekeyed"},
-{ERR_REASON(RAND_R_PRNG_NOT_RESEEDED)    ,"prng not reseeded"},
+{ERR_REASON(RAND_R_ERROR_INITIALISING_DRBG),"error initialising drbg"},
+{ERR_REASON(RAND_R_ERROR_INSTANTIATING_DRBG),"error instantiating drbg"},
+{ERR_REASON(RAND_R_NO_FIPS_RANDOM_METHOD_SET),"no fips random method set"},
 {ERR_REASON(RAND_R_PRNG_NOT_SEEDED)      ,"PRNG not seeded"},
-{ERR_REASON(RAND_R_PRNG_SEED_MUST_NOT_MATCH_KEY),"prng seed must not match key"},
-{ERR_REASON(RAND_R_PRNG_STUCK)           ,"prng stuck"},
 {0,NULL}
 	};
 
diff --git a/crypto/rand/rand_lcl.h b/crypto/rand/rand_lcl.h
index 18cc9b1e4afc..618a8ec899ce 100755
--- a/crypto/rand/rand_lcl.h
+++ b/crypto/rand/rand_lcl.h
@@ -154,16 +154,5 @@
 #define	MD(a,b,c)		EVP_Digest(a,b,c,NULL,EVP_md2(), NULL)
 #endif
 
-#ifndef OPENSSL_NO_ENGINE
-void int_RAND_set_callbacks(
-	int (*set_rand_func)(const RAND_METHOD *meth,
-						const RAND_METHOD **pmeth),
-	const RAND_METHOD *(*get_rand_func)
-						(const RAND_METHOD **pmeth));
-int eng_RAND_set_rand_method(const RAND_METHOD *meth,
-				const RAND_METHOD **pmeth);
-const RAND_METHOD *eng_RAND_get_rand_method(const RAND_METHOD **pmeth);
-#endif
-
 
 #endif
diff --git a/crypto/rand/rand_lib.c b/crypto/rand/rand_lib.c
index da6b4e0e862f..daf1dab97395 100644
--- a/crypto/rand/rand_lib.c
+++ b/crypto/rand/rand_lib.c
@@ -60,82 +60,21 @@
 #include <time.h>
 #include "cryptlib.h"
 #include <openssl/rand.h>
-#include "rand_lcl.h"
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#include <openssl/fips_rand.h>
-#endif
 
 #ifndef OPENSSL_NO_ENGINE
 #include <openssl/engine.h>
 #endif
 
-static const RAND_METHOD *default_RAND_meth = NULL;
-
 #ifdef OPENSSL_FIPS
-
-static int fips_RAND_set_rand_method(const RAND_METHOD *meth,
-					const RAND_METHOD **pmeth)
-	{
-	*pmeth = meth;
-	return 1;
-	}
-
-static const RAND_METHOD *fips_RAND_get_rand_method(const RAND_METHOD **pmeth)
-	{
-	if (!*pmeth)
-		{
-		if(FIPS_mode())
-			*pmeth=FIPS_rand_method();
-		else
-			*pmeth = RAND_SSLeay();
-		}
-
-	if(FIPS_mode()
-		&& *pmeth != FIPS_rand_check())
-	    {
-	    RANDerr(RAND_F_FIPS_RAND_GET_RAND_METHOD,RAND_R_NON_FIPS_METHOD);
-	    return 0;
-	    }
-
-	return *pmeth;
-	}
-
-static int (*RAND_set_rand_method_func)(const RAND_METHOD *meth,
-						const RAND_METHOD **pmeth)
-	= fips_RAND_set_rand_method;
-static const RAND_METHOD *(*RAND_get_rand_method_func)
-						(const RAND_METHOD **pmeth)
-	= fips_RAND_get_rand_method;
-
-#ifndef OPENSSL_NO_ENGINE
-void int_RAND_set_callbacks(
-	int (*set_rand_func)(const RAND_METHOD *meth,
-						const RAND_METHOD **pmeth),
-	const RAND_METHOD *(*get_rand_func)
-						(const RAND_METHOD **pmeth))
-	{
-	RAND_set_rand_method_func = set_rand_func;
-	RAND_get_rand_method_func = get_rand_func;
-	}
+#include <openssl/fips.h>
+#include <openssl/fips_rand.h>
 #endif
 
-int RAND_set_rand_method(const RAND_METHOD *meth)
-	{
-	return RAND_set_rand_method_func(meth, &default_RAND_meth);
-	}
-
-const RAND_METHOD *RAND_get_rand_method(void)
-	{
-	return RAND_get_rand_method_func(&default_RAND_meth);
-	}
-
-#else
-
 #ifndef OPENSSL_NO_ENGINE
 /* non-NULL if default_RAND_meth is ENGINE-provided */
 static ENGINE *funct_ref =NULL;
 #endif
+static const RAND_METHOD *default_RAND_meth = NULL;
 
 int RAND_set_rand_method(const RAND_METHOD *meth)
 	{
@@ -196,8 +135,6 @@ int RAND_set_rand_engine(ENGINE *engine)
 	}
 #endif
 
-#endif
-
 void RAND_cleanup(void)
 	{
 	const RAND_METHOD *meth = RAND_get_rand_method();
@@ -243,3 +180,116 @@ int RAND_status(void)
 		return meth->status();
 	return 0;
 	}
+
+#ifdef OPENSSL_FIPS
+
+/* FIPS DRBG initialisation code. This sets up the DRBG for use by the
+ * rest of OpenSSL. 
+ */
+
+/* Entropy gatherer: use standard OpenSSL PRNG to seed (this will gather
+ * entropy internally through RAND_poll().
+ */
+
+static size_t drbg_get_entropy(DRBG_CTX *ctx, unsigned char **pout,
+                                int entropy, size_t min_len, size_t max_len)
+        {
+	/* Round up request to multiple of block size */
+	min_len = ((min_len + 19) / 20) * 20;
+	*pout = OPENSSL_malloc(min_len);
+	if (!*pout)
+		return 0;
+	if (RAND_SSLeay()->bytes(*pout, min_len) <= 0)
+		{
+		OPENSSL_free(*pout);
+		*pout = NULL;
+		return 0;
+		}
+        return min_len;
+        }
+
+static void drbg_free_entropy(DRBG_CTX *ctx, unsigned char *out, size_t olen)
+	{
+	OPENSSL_cleanse(out, olen);
+	OPENSSL_free(out);
+	}
+
+/* Set "additional input" when generating random data. This uses the
+ * current PID, a time value and a counter.
+ */
+
+static size_t drbg_get_adin(DRBG_CTX *ctx, unsigned char **pout)
+    	{
+	/* Use of static variables is OK as this happens under a lock */
+	static unsigned char buf[16];
+	static unsigned long counter;
+	FIPS_get_timevec(buf, &counter);
+	*pout = buf;
+	return sizeof(buf);
+	}
+
+/* RAND_add() and RAND_seed() pass through to OpenSSL PRNG so it is 
+ * correctly seeded by RAND_poll().
+ */
+
+static int drbg_rand_add(DRBG_CTX *ctx, const void *in, int inlen,
+				double entropy)
+	{
+	RAND_SSLeay()->add(in, inlen, entropy);
+	return 1;
+	}
+
+static int drbg_rand_seed(DRBG_CTX *ctx, const void *in, int inlen)
+	{
+	RAND_SSLeay()->seed(in, inlen);
+	return 1;
+	}
+
+#ifndef OPENSSL_DRBG_DEFAULT_TYPE
+#define OPENSSL_DRBG_DEFAULT_TYPE	NID_aes_256_ctr
+#endif
+#ifndef OPENSSL_DRBG_DEFAULT_FLAGS
+#define OPENSSL_DRBG_DEFAULT_FLAGS	DRBG_FLAG_CTR_USE_DF
+#endif 
+
+static int fips_drbg_type = OPENSSL_DRBG_DEFAULT_TYPE;
+static int fips_drbg_flags = OPENSSL_DRBG_DEFAULT_FLAGS;
+
+void RAND_set_fips_drbg_type(int type, int flags)
+	{
+	fips_drbg_type = type;
+	fips_drbg_flags = flags;
+	}
+
+int RAND_init_fips(void)
+	{
+	DRBG_CTX *dctx;
+	size_t plen;
+	unsigned char pers[32], *p;
+	dctx = FIPS_get_default_drbg();
+        if (FIPS_drbg_init(dctx, fips_drbg_type, fips_drbg_flags) <= 0)
+		{
+		RANDerr(RAND_F_RAND_INIT_FIPS, RAND_R_ERROR_INITIALISING_DRBG);
+		return 0;
+		}
+		
+        FIPS_drbg_set_callbacks(dctx,
+				drbg_get_entropy, drbg_free_entropy, 20,
+				drbg_get_entropy, drbg_free_entropy);
+	FIPS_drbg_set_rand_callbacks(dctx, drbg_get_adin, 0,
+					drbg_rand_seed, drbg_rand_add);
+	/* Personalisation string: a string followed by date time vector */
+	strcpy((char *)pers, "OpenSSL DRBG2.0");
+	plen = drbg_get_adin(dctx, &p);
+	memcpy(pers + 16, p, plen);
+
+        if (FIPS_drbg_instantiate(dctx, pers, sizeof(pers)) <= 0)
+		{
+		RANDerr(RAND_F_RAND_INIT_FIPS, RAND_R_ERROR_INSTANTIATING_DRBG);
+		return 0;
+		}
+        FIPS_rand_set_method(FIPS_drbg_method());
+	return 1;
+	}
+
+#endif
diff --git a/crypto/rand/rand_nw.c b/crypto/rand/rand_nw.c
deleted file mode 100644
index 8d5b8d2e3217..000000000000
--- a/crypto/rand/rand_nw.c
+++ /dev/null
@@ -1,183 +0,0 @@
-/* crypto/rand/rand_nw.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-/* ====================================================================
- * Copyright (c) 1998-2000 The OpenSSL Project.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    openssl-core@openssl.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
- */
-
-#include "cryptlib.h"
-#include <openssl/rand.h>
-#include "rand_lcl.h"
-
-#if defined (OPENSSL_SYS_NETWARE)
-
-#if defined(NETWARE_LIBC)
-#include <nks/thread.h>
-#else
-#include <nwthread.h>
-#endif
-
-extern int GetProcessSwitchCount(void);
-#if !defined(NETWARE_LIBC) || (CURRENT_NDK_THRESHOLD < 509220000)
-extern void *RunningProcess; /* declare here same as found in newer NDKs */
-extern unsigned long GetSuperHighResolutionTimer(void);
-#endif
-
-   /* the FAQ indicates we need to provide at least 20 bytes (160 bits) of seed
-   */
-int RAND_poll(void)
-{
-   unsigned long l;
-   unsigned long tsc;
-   int i; 
-
-      /* There are several options to gather miscellaneous data
-       * but for now we will loop checking the time stamp counter (rdtsc) and
-       * the SuperHighResolutionTimer.  Each iteration will collect 8 bytes
-       * of data but it is treated as only 1 byte of entropy.  The call to
-       * ThreadSwitchWithDelay() will introduce additional variability into
-       * the data returned by rdtsc.
-       *
-       * Applications can agument the seed material by adding additional
-       * stuff with RAND_add() and should probably do so.
-      */
-   l = GetProcessSwitchCount();
-   RAND_add(&l,sizeof(l),1);
-   
-   /* need to cast the void* to unsigned long here */
-   l = (unsigned long)RunningProcess;
-   RAND_add(&l,sizeof(l),1);
-
-   for( i=2; i<ENTROPY_NEEDED; i++)
-   {
-#ifdef __MWERKS__
-      asm 
-      {
-         rdtsc
-         mov tsc, eax        
-      }
-#elif defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
-      asm volatile("rdtsc":"=a"(tsc)::"edx");
-#endif
-
-      RAND_add(&tsc, sizeof(tsc), 1);
-
-      l = GetSuperHighResolutionTimer();
-      RAND_add(&l, sizeof(l), 0);
-
-# if defined(NETWARE_LIBC)
-      NXThreadYield();
-# else /* NETWARE_CLIB */
-      ThreadSwitchWithDelay();
-# endif
-   }
-
-   return 1;
-}
-
-#endif 
-
diff --git a/crypto/rand/rand_os2.c b/crypto/rand/rand_os2.c
deleted file mode 100644
index c3e36d4e5e33..000000000000
--- a/crypto/rand/rand_os2.c
+++ /dev/null
@@ -1,147 +0,0 @@
-/* crypto/rand/rand_os2.c */
-/* ====================================================================
- * Copyright (c) 1998-2000 The OpenSSL Project.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    openssl-core@openssl.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
- */
-
-#include "cryptlib.h"
-#include <openssl/rand.h>
-#include "rand_lcl.h"
-
-#ifdef OPENSSL_SYS_OS2
-
-#define INCL_DOSPROCESS
-#define INCL_DOSPROFILE
-#define INCL_DOSMISC
-#define INCL_DOSMODULEMGR
-#include <os2.h>
-
-#define   CMD_KI_RDCNT    (0x63)
-
-typedef struct _CPUUTIL {
-    ULONG ulTimeLow;            /* Low 32 bits of time stamp      */
-    ULONG ulTimeHigh;           /* High 32 bits of time stamp     */
-    ULONG ulIdleLow;            /* Low 32 bits of idle time       */
-    ULONG ulIdleHigh;           /* High 32 bits of idle time      */
-    ULONG ulBusyLow;            /* Low 32 bits of busy time       */
-    ULONG ulBusyHigh;           /* High 32 bits of busy time      */
-    ULONG ulIntrLow;            /* Low 32 bits of interrupt time  */
-    ULONG ulIntrHigh;           /* High 32 bits of interrupt time */
-} CPUUTIL;
-
-APIRET APIENTRY(*DosPerfSysCall) (ULONG ulCommand, ULONG ulParm1, ULONG ulParm2, ULONG ulParm3) = NULL;
-APIRET APIENTRY(*DosQuerySysState) (ULONG func, ULONG arg1, ULONG pid, ULONG _res_, PVOID buf, ULONG bufsz) = NULL;
-HMODULE hDoscalls = 0;
-
-int RAND_poll(void)
-{
-    char failed_module[20];
-    QWORD qwTime;
-    ULONG SysVars[QSV_FOREGROUND_PROCESS];
-
-    if (hDoscalls == 0) {
-        ULONG rc = DosLoadModule(failed_module, sizeof(failed_module), "DOSCALLS", &hDoscalls);
-
-        if (rc == 0) {
-            rc = DosQueryProcAddr(hDoscalls, 976, NULL, (PFN *)&DosPerfSysCall);
-
-            if (rc)
-                DosPerfSysCall = NULL;
-
-            rc = DosQueryProcAddr(hDoscalls, 368, NULL, (PFN *)&DosQuerySysState);
-
-            if (rc)
-                DosQuerySysState = NULL;
-        }
-    }
-
-    /* Sample the hi-res timer, runs at around 1.1 MHz */
-    DosTmrQueryTime(&qwTime);
-    RAND_add(&qwTime, sizeof(qwTime), 2);
-
-    /* Sample a bunch of system variables, includes various process & memory statistics */
-    DosQuerySysInfo(1, QSV_FOREGROUND_PROCESS, SysVars, sizeof(SysVars));
-    RAND_add(SysVars, sizeof(SysVars), 4);
-
-    /* If available, sample CPU registers that count at CPU MHz
-     * Only fairly new CPUs (PPro & K6 onwards) & OS/2 versions support this
-     */
-    if (DosPerfSysCall) {
-        CPUUTIL util;
-
-        if (DosPerfSysCall(CMD_KI_RDCNT, (ULONG)&util, 0, 0) == 0) {
-            RAND_add(&util, sizeof(util), 10);
-        }
-        else {
-            DosPerfSysCall = NULL;
-        }
-    }
-
-    /* DosQuerySysState() gives us a huge quantity of process, thread, memory & handle stats */
-    if (DosQuerySysState) {
-        char *buffer = OPENSSL_malloc(256 * 1024);
-
-        if (DosQuerySysState(0x1F, 0, 0, 0, buffer, 256 * 1024) == 0) {
-            /* First 4 bytes in buffer is a pointer to the thread count
-             * there should be at least 1 byte of entropy per thread
-             */
-            RAND_add(buffer, 256 * 1024, **(ULONG **)buffer);
-        }
-
-        OPENSSL_free(buffer);
-        return 1;
-    }
-
-    return 0;
-}
-
-#endif /* OPENSSL_SYS_OS2 */
diff --git a/crypto/rand/rand_unix.c b/crypto/rand/rand_unix.c
index 71b98ec2121d..e3a65571c8df 100644
--- a/crypto/rand/rand_unix.c
+++ b/crypto/rand/rand_unix.c
@@ -133,7 +133,90 @@
 # define FD_SETSIZE (8*sizeof(fd_set))
 #endif
 
-#ifdef __OpenBSD__
+#if defined(OPENSSL_SYS_VOS)
+
+/* The following algorithm repeatedly samples the real-time clock
+   (RTC) to generate a sequence of unpredictable data.  The algorithm
+   relies upon the uneven execution speed of the code (due to factors
+   such as cache misses, interrupts, bus activity, and scheduling) and
+   upon the rather large relative difference between the speed of the
+   clock and the rate at which it can be read.
+
+   If this code is ported to an environment where execution speed is
+   more constant or where the RTC ticks at a much slower rate, or the
+   clock can be read with fewer instructions, it is likely that the
+   results would be far more predictable.
+
+   As a precaution, we generate 4 times the minimum required amount of
+   seed data.  */
+
+int RAND_poll(void)
+{
+	short int code;
+	gid_t curr_gid;
+	pid_t curr_pid;
+	uid_t curr_uid;
+	int i, k;
+	struct timespec ts;
+	unsigned char v;
+
+#ifdef OPENSSL_SYS_VOS_HPPA
+	long duration;
+	extern void s$sleep (long *_duration, short int *_code);
+#else
+#ifdef OPENSSL_SYS_VOS_IA32
+	long long duration;
+	extern void s$sleep2 (long long *_duration, short int *_code);
+#else
+#error "Unsupported Platform."
+#endif /* OPENSSL_SYS_VOS_IA32 */
+#endif /* OPENSSL_SYS_VOS_HPPA */
+
+	/* Seed with the gid, pid, and uid, to ensure *some*
+	   variation between different processes.  */
+
+	curr_gid = getgid();
+	RAND_add (&curr_gid, sizeof curr_gid, 1);
+	curr_gid = 0;
+
+	curr_pid = getpid();
+	RAND_add (&curr_pid, sizeof curr_pid, 1);
+	curr_pid = 0;
+
+	curr_uid = getuid();
+	RAND_add (&curr_uid, sizeof curr_uid, 1);
+	curr_uid = 0;
+
+	for (i=0; i<(ENTROPY_NEEDED*4); i++)
+	{
+		/* burn some cpu; hope for interrupts, cache
+		   collisions, bus interference, etc.  */
+		for (k=0; k<99; k++)
+			ts.tv_nsec = random ();
+
+#ifdef OPENSSL_SYS_VOS_HPPA
+		/* sleep for 1/1024 of a second (976 us).  */
+		duration = 1;
+		s$sleep (&duration, &code);
+#else
+#ifdef OPENSSL_SYS_VOS_IA32
+		/* sleep for 1/65536 of a second (15 us).  */
+		duration = 1;
+		s$sleep2 (&duration, &code);
+#endif /* OPENSSL_SYS_VOS_IA32 */
+#endif /* OPENSSL_SYS_VOS_HPPA */
+
+		/* get wall clock time.  */
+		clock_gettime (CLOCK_REALTIME, &ts);
+
+		/* take 8 bits */
+		v = (unsigned char) (ts.tv_nsec % 256);
+		RAND_add (&v, sizeof v, 1);
+		v = 0;
+	}
+	return 1;
+}
+#elif defined __OpenBSD__
 int RAND_poll(void)
 {
 	u_int32_t rnd = 0, i;
@@ -163,7 +246,7 @@ int RAND_poll(void)
 	static const char *randomfiles[] = { DEVRANDOM };
 	struct stat randomstats[sizeof(randomfiles)/sizeof(randomfiles[0])];
 	int fd;
-	size_t i;
+	unsigned int i;
 #endif
 #ifdef DEVRANDOM_EGD
 	static const char *egdsockets[] = { DEVRANDOM_EGD, NULL };
@@ -176,7 +259,8 @@ int RAND_poll(void)
 	 * have this. Use /dev/urandom if you can as /dev/random may block
 	 * if it runs out of random entries.  */
 
-	for (i=0; i<sizeof(randomfiles)/sizeof(randomfiles[0]) && n < ENTROPY_NEEDED; i++)
+	for (i = 0; (i < sizeof(randomfiles)/sizeof(randomfiles[0])) &&
+			(n < ENTROPY_NEEDED); i++)
 		{
 		if ((fd = open(randomfiles[i], O_RDONLY
 #ifdef O_NONBLOCK
@@ -193,7 +277,7 @@ int RAND_poll(void)
 			{
 			int usec = 10*1000; /* spend 10ms on each file */
 			int r;
-			size_t j;
+			unsigned int j;
 			struct stat *st=&randomstats[i];
 
 			/* Avoid using same input... Used to be O_NOFOLLOW
@@ -211,7 +295,12 @@ int RAND_poll(void)
 				{
 				int try_read = 0;
 
-#if defined(OPENSSL_SYS_LINUX)
+#if defined(OPENSSL_SYS_BEOS_R5)
+				/* select() is broken in BeOS R5, so we simply
+				 *  try to read something and snooze if we couldn't */
+				try_read = 1;
+
+#elif defined(OPENSSL_SYS_LINUX)
 				/* use poll() */
 				struct pollfd pset;
 				
@@ -258,6 +347,10 @@ int RAND_poll(void)
 					r = read(fd,(unsigned char *)tmpbuf+n, ENTROPY_NEEDED-n);
 					if (r > 0)
 						n += r;
+#if defined(OPENSSL_SYS_BEOS_R5)
+					if (r == 0)
+						snooze(t.tv_usec);
+#endif
 					}
 				else
 					r = -1;
@@ -311,6 +404,14 @@ int RAND_poll(void)
 	l=time(NULL);
 	RAND_add(&l,sizeof(l),0.0);
 
+#if defined(OPENSSL_SYS_BEOS)
+	{
+	system_info sysInfo;
+	get_system_info(&sysInfo);
+	RAND_add(&sysInfo,sizeof(sysInfo),0);
+	}
+#endif
+
 #if defined(DEVRANDOM) || defined(DEVRANDOM_EGD)
 	return 1;
 #else
diff --git a/crypto/rand/rand_win.c b/crypto/rand/rand_win.c
deleted file mode 100644
index 5d134e186bb0..000000000000
--- a/crypto/rand/rand_win.c
+++ /dev/null
@@ -1,807 +0,0 @@
-/* crypto/rand/rand_win.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-/* ====================================================================
- * Copyright (c) 1998-2000 The OpenSSL Project.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
- *
- * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    openssl-core@openssl.org.
- *
- * 5. Products derived from this software may not be called "OpenSSL"
- *    nor may "OpenSSL" appear in their names without prior written
- *    permission of the OpenSSL Project.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the OpenSSL Project
- *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
- *
- * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
- */
-
-#include "cryptlib.h"
-#include <openssl/rand.h>
-#include "rand_lcl.h"
-
-#if defined(OPENSSL_SYS_WINDOWS) || defined(OPENSSL_SYS_WIN32)
-#include <windows.h>
-#ifndef _WIN32_WINNT
-# define _WIN32_WINNT 0x0400
-#endif
-#include <wincrypt.h>
-#include <tlhelp32.h>
-
-/* Limit the time spent walking through the heap, processes, threads and modules to
-   a maximum of 1000 miliseconds each, unless CryptoGenRandom failed */
-#define MAXDELAY 1000
-
-/* Intel hardware RNG CSP -- available from
- * http://developer.intel.com/design/security/rng/redist_license.htm
- */
-#define PROV_INTEL_SEC 22
-#define INTEL_DEF_PROV L"Intel Hardware Cryptographic Service Provider"
-
-static void readtimer(void);
-static void readscreen(void);
-
-/* It appears like CURSORINFO, PCURSORINFO and LPCURSORINFO are only defined
-   when WINVER is 0x0500 and up, which currently only happens on Win2000.
-   Unfortunately, those are typedefs, so they're a little bit difficult to
-   detect properly.  On the other hand, the macro CURSOR_SHOWING is defined
-   within the same conditional, so it can be use to detect the absence of said
-   typedefs. */
-
-#ifndef CURSOR_SHOWING
-/*
- * Information about the global cursor.
- */
-typedef struct tagCURSORINFO
-{
-    DWORD   cbSize;
-    DWORD   flags;
-    HCURSOR hCursor;
-    POINT   ptScreenPos;
-} CURSORINFO, *PCURSORINFO, *LPCURSORINFO;
-
-#define CURSOR_SHOWING     0x00000001
-#endif /* CURSOR_SHOWING */
-
-#if !defined(OPENSSL_SYS_WINCE)
-typedef BOOL (WINAPI *CRYPTACQUIRECONTEXTW)(HCRYPTPROV *, LPCWSTR, LPCWSTR,
-				    DWORD, DWORD);
-typedef BOOL (WINAPI *CRYPTGENRANDOM)(HCRYPTPROV, DWORD, BYTE *);
-typedef BOOL (WINAPI *CRYPTRELEASECONTEXT)(HCRYPTPROV, DWORD);
-
-typedef HWND (WINAPI *GETFOREGROUNDWINDOW)(VOID);
-typedef BOOL (WINAPI *GETCURSORINFO)(PCURSORINFO);
-typedef DWORD (WINAPI *GETQUEUESTATUS)(UINT);
-
-typedef HANDLE (WINAPI *CREATETOOLHELP32SNAPSHOT)(DWORD, DWORD);
-typedef BOOL (WINAPI *CLOSETOOLHELP32SNAPSHOT)(HANDLE);
-typedef BOOL (WINAPI *HEAP32FIRST)(LPHEAPENTRY32, DWORD, size_t);
-typedef BOOL (WINAPI *HEAP32NEXT)(LPHEAPENTRY32);
-typedef BOOL (WINAPI *HEAP32LIST)(HANDLE, LPHEAPLIST32);
-typedef BOOL (WINAPI *PROCESS32)(HANDLE, LPPROCESSENTRY32);
-typedef BOOL (WINAPI *THREAD32)(HANDLE, LPTHREADENTRY32);
-typedef BOOL (WINAPI *MODULE32)(HANDLE, LPMODULEENTRY32);
-
-#include <lmcons.h>
-#include <lmstats.h>
-#if 1 /* The NET API is Unicode only.  It requires the use of the UNICODE
-       * macro.  When UNICODE is defined LPTSTR becomes LPWSTR.  LMSTR was
-       * was added to the Platform SDK to allow the NET API to be used in
-       * non-Unicode applications provided that Unicode strings were still
-       * used for input.  LMSTR is defined as LPWSTR.
-       */
-typedef NET_API_STATUS (NET_API_FUNCTION * NETSTATGET)
-        (LPWSTR, LPWSTR, DWORD, DWORD, LPBYTE*);
-typedef NET_API_STATUS (NET_API_FUNCTION * NETFREE)(LPBYTE);
-#endif /* 1 */
-#endif /* !OPENSSL_SYS_WINCE */
-
-int RAND_poll(void)
-{
-	MEMORYSTATUS m;
-	HCRYPTPROV hProvider = 0;
-	DWORD w;
-	int good = 0;
-
-	/* Determine the OS version we are on so we can turn off things 
-	 * that do not work properly.
-	 */
-        OSVERSIONINFO osverinfo ;
-        osverinfo.dwOSVersionInfoSize = sizeof(OSVERSIONINFO) ;
-        GetVersionEx( &osverinfo ) ;
-
-#if defined(OPENSSL_SYS_WINCE)
-# if defined(_WIN32_WCE) && _WIN32_WCE>=300
-/* Even though MSDN says _WIN32_WCE>=210, it doesn't seem to be available
- * in commonly available implementations prior 300... */
-	{
-	BYTE buf[64];
-	/* poll the CryptoAPI PRNG */
-	/* The CryptoAPI returns sizeof(buf) bytes of randomness */
-	if (CryptAcquireContextW(&hProvider, NULL, NULL, PROV_RSA_FULL,
-				CRYPT_VERIFYCONTEXT))
-		{
-		if (CryptGenRandom(hProvider, sizeof(buf), buf))
-			RAND_add(buf, sizeof(buf), sizeof(buf));
-		CryptReleaseContext(hProvider, 0); 
-		}
-	}
-# endif
-#else	/* OPENSSL_SYS_WINCE */
-	/*
-	 * None of below libraries are present on Windows CE, which is
-	 * why we #ifndef the whole section. This also excuses us from
-	 * handling the GetProcAddress issue. The trouble is that in
-	 * real Win32 API GetProcAddress is available in ANSI flavor
-	 * only. In WinCE on the other hand GetProcAddress is a macro
-	 * most commonly defined as GetProcAddressW, which accepts
-	 * Unicode argument. If we were to call GetProcAddress under
-	 * WinCE, I'd recommend to either redefine GetProcAddress as
-	 * GetProcAddressA (there seem to be one in common CE spec) or
-	 * implement own shim routine, which would accept ANSI argument
-	 * and expand it to Unicode.
-	 */
-	{
-	/* load functions dynamically - not available on all systems */
-	HMODULE advapi = LoadLibrary(TEXT("ADVAPI32.DLL"));
-	HMODULE kernel = LoadLibrary(TEXT("KERNEL32.DLL"));
-	HMODULE user = NULL;
-	HMODULE netapi = LoadLibrary(TEXT("NETAPI32.DLL"));
-	CRYPTACQUIRECONTEXTW acquire = NULL;
-	CRYPTGENRANDOM gen = NULL;
-	CRYPTRELEASECONTEXT release = NULL;
-	NETSTATGET netstatget = NULL;
-	NETFREE netfree = NULL;
-	BYTE buf[64];
-
-	if (netapi)
-		{
-		netstatget = (NETSTATGET) GetProcAddress(netapi,"NetStatisticsGet");
-		netfree = (NETFREE) GetProcAddress(netapi,"NetApiBufferFree");
-		}
-
-	if (netstatget && netfree)
-		{
-		LPBYTE outbuf;
-		/* NetStatisticsGet() is a Unicode only function
- 		 * STAT_WORKSTATION_0 contains 45 fields and STAT_SERVER_0
-		 * contains 17 fields.  We treat each field as a source of
-		 * one byte of entropy.
-                 */
-
-		if (netstatget(NULL, L"LanmanWorkstation", 0, 0, &outbuf) == 0)
-			{
-			RAND_add(outbuf, sizeof(STAT_WORKSTATION_0), 45);
-			netfree(outbuf);
-			}
-		if (netstatget(NULL, L"LanmanServer", 0, 0, &outbuf) == 0)
-			{
-			RAND_add(outbuf, sizeof(STAT_SERVER_0), 17);
-			netfree(outbuf);
-			}
-		}
-
-	if (netapi)
-		FreeLibrary(netapi);
-
-        /* It appears like this can cause an exception deep within ADVAPI32.DLL
-         * at random times on Windows 2000.  Reported by Jeffrey Altman.  
-         * Only use it on NT.
-	 */
-	/* Wolfgang Marczy <WMarczy@topcall.co.at> reports that
-	 * the RegQueryValueEx call below can hang on NT4.0 (SP6).
-	 * So we don't use this at all for now. */
-#if 0
-        if ( osverinfo.dwPlatformId == VER_PLATFORM_WIN32_NT &&
-		osverinfo.dwMajorVersion < 5)
-		{
-		/* Read Performance Statistics from NT/2000 registry
-		 * The size of the performance data can vary from call
-		 * to call so we must guess the size of the buffer to use
-		 * and increase its size if we get an ERROR_MORE_DATA
-		 * return instead of ERROR_SUCCESS.
-		 */
-		LONG   rc=ERROR_MORE_DATA;
-		char * buf=NULL;
-		DWORD bufsz=0;
-		DWORD length;
-
-		while (rc == ERROR_MORE_DATA)
-			{
-			buf = realloc(buf,bufsz+8192);
-			if (!buf)
-				break;
-			bufsz += 8192;
-
-			length = bufsz;
-			rc = RegQueryValueEx(HKEY_PERFORMANCE_DATA, TEXT("Global"),
-				NULL, NULL, buf, &length);
-			}
-		if (rc == ERROR_SUCCESS)
-			{
-                        /* For entropy count assume only least significant
-			 * byte of each DWORD is random.
-			 */
-			RAND_add(&length, sizeof(length), 0);
-			RAND_add(buf, length, length / 4.0);
-
-			/* Close the Registry Key to allow Windows to cleanup/close
-			 * the open handle
-			 * Note: The 'HKEY_PERFORMANCE_DATA' key is implicitly opened
-			 *       when the RegQueryValueEx above is done.  However, if
-			 *       it is not explicitly closed, it can cause disk
-			 *       partition manipulation problems.
-			 */
-			RegCloseKey(HKEY_PERFORMANCE_DATA);
-			}
-		if (buf)
-			free(buf);
-		}
-#endif
-
-	if (advapi)
-		{
-		/*
-		 * If it's available, then it's available in both ANSI
-		 * and UNICODE flavors even in Win9x, documentation says.
-		 * We favor Unicode...
-		 */
-		acquire = (CRYPTACQUIRECONTEXTW) GetProcAddress(advapi,
-			"CryptAcquireContextW");
-		gen = (CRYPTGENRANDOM) GetProcAddress(advapi,
-			"CryptGenRandom");
-		release = (CRYPTRELEASECONTEXT) GetProcAddress(advapi,
-			"CryptReleaseContext");
-		}
-
-	if (acquire && gen && release)
-		{
-		/* poll the CryptoAPI PRNG */
-                /* The CryptoAPI returns sizeof(buf) bytes of randomness */
-		if (acquire(&hProvider, NULL, NULL, PROV_RSA_FULL,
-			CRYPT_VERIFYCONTEXT))
-			{
-			if (gen(hProvider, sizeof(buf), buf) != 0)
-				{
-				RAND_add(buf, sizeof(buf), 0);
-				good = 1;
-#if 0
-				printf("randomness from PROV_RSA_FULL\n");
-#endif
-				}
-			release(hProvider, 0); 
-			}
-		
-		/* poll the Pentium PRG with CryptoAPI */
-		if (acquire(&hProvider, 0, INTEL_DEF_PROV, PROV_INTEL_SEC, 0))
-			{
-			if (gen(hProvider, sizeof(buf), buf) != 0)
-				{
-				RAND_add(buf, sizeof(buf), sizeof(buf));
-				good = 1;
-#if 0
-				printf("randomness from PROV_INTEL_SEC\n");
-#endif
-				}
-			release(hProvider, 0);
-			}
-		}
-
-        if (advapi)
-		FreeLibrary(advapi);
-
-	if ((osverinfo.dwPlatformId != VER_PLATFORM_WIN32_NT ||
-	     !OPENSSL_isservice()) &&
-	    (user = LoadLibrary(TEXT("USER32.DLL"))))
-		{
-		GETCURSORINFO cursor;
-		GETFOREGROUNDWINDOW win;
-		GETQUEUESTATUS queue;
-
-		win = (GETFOREGROUNDWINDOW) GetProcAddress(user, "GetForegroundWindow");
-		cursor = (GETCURSORINFO) GetProcAddress(user, "GetCursorInfo");
-		queue = (GETQUEUESTATUS) GetProcAddress(user, "GetQueueStatus");
-
-		if (win)
-			{
-			/* window handle */
-			HWND h = win();
-			RAND_add(&h, sizeof(h), 0);
-			}
-		if (cursor)
-			{
-			/* unfortunately, its not safe to call GetCursorInfo()
-			 * on NT4 even though it exists in SP3 (or SP6) and
-			 * higher.
-			 */
-			if ( osverinfo.dwPlatformId == VER_PLATFORM_WIN32_NT &&
-				osverinfo.dwMajorVersion < 5)
-				cursor = 0;
-			}
-		if (cursor)
-			{
-			/* cursor position */
-                        /* assume 2 bytes of entropy */
-			CURSORINFO ci;
-			ci.cbSize = sizeof(CURSORINFO);
-			if (cursor(&ci))
-				RAND_add(&ci, ci.cbSize, 2);
-			}
-
-		if (queue)
-			{
-			/* message queue status */
-                        /* assume 1 byte of entropy */
-			w = queue(QS_ALLEVENTS);
-			RAND_add(&w, sizeof(w), 1);
-			}
-
-		FreeLibrary(user);
-		}
-
-	/* Toolhelp32 snapshot: enumerate processes, threads, modules and heap
-	 * http://msdn.microsoft.com/library/psdk/winbase/toolhelp_5pfd.htm
-	 * (Win 9x and 2000 only, not available on NT)
-	 *
-	 * This seeding method was proposed in Peter Gutmann, Software
-	 * Generation of Practically Strong Random Numbers,
-	 * http://www.usenix.org/publications/library/proceedings/sec98/gutmann.html
-	 * revised version at http://www.cryptoengines.com/~peter/06_random.pdf
-	 * (The assignment of entropy estimates below is arbitrary, but based
-	 * on Peter's analysis the full poll appears to be safe. Additional
-	 * interactive seeding is encouraged.)
-	 */
-
-	if (kernel)
-		{
-		CREATETOOLHELP32SNAPSHOT snap;
-		CLOSETOOLHELP32SNAPSHOT close_snap;
-		HANDLE handle;
-
-		HEAP32FIRST heap_first;
-		HEAP32NEXT heap_next;
-		HEAP32LIST heaplist_first, heaplist_next;
-		PROCESS32 process_first, process_next;
-		THREAD32 thread_first, thread_next;
-		MODULE32 module_first, module_next;
-
-		HEAPLIST32 hlist;
-		HEAPENTRY32 hentry;
-		PROCESSENTRY32 p;
-		THREADENTRY32 t;
-		MODULEENTRY32 m;
-		DWORD starttime = 0;
-
-		snap = (CREATETOOLHELP32SNAPSHOT)
-			GetProcAddress(kernel, "CreateToolhelp32Snapshot");
-		close_snap = (CLOSETOOLHELP32SNAPSHOT)
-			GetProcAddress(kernel, "CloseToolhelp32Snapshot");
-		heap_first = (HEAP32FIRST) GetProcAddress(kernel, "Heap32First");
-		heap_next = (HEAP32NEXT) GetProcAddress(kernel, "Heap32Next");
-		heaplist_first = (HEAP32LIST) GetProcAddress(kernel, "Heap32ListFirst");
-		heaplist_next = (HEAP32LIST) GetProcAddress(kernel, "Heap32ListNext");
-		process_first = (PROCESS32) GetProcAddress(kernel, "Process32First");
-		process_next = (PROCESS32) GetProcAddress(kernel, "Process32Next");
-		thread_first = (THREAD32) GetProcAddress(kernel, "Thread32First");
-		thread_next = (THREAD32) GetProcAddress(kernel, "Thread32Next");
-		module_first = (MODULE32) GetProcAddress(kernel, "Module32First");
-		module_next = (MODULE32) GetProcAddress(kernel, "Module32Next");
-
-		if (snap && heap_first && heap_next && heaplist_first &&
-			heaplist_next && process_first && process_next &&
-			thread_first && thread_next && module_first &&
-			module_next && (handle = snap(TH32CS_SNAPALL,0))
-			!= INVALID_HANDLE_VALUE)
-			{
-			/* heap list and heap walking */
-                        /* HEAPLIST32 contains 3 fields that will change with
-                         * each entry.  Consider each field a source of 1 byte
-                         * of entropy.
-                         * HEAPENTRY32 contains 5 fields that will change with 
-                         * each entry.  Consider each field a source of 1 byte
-                         * of entropy.
-                         */
-			ZeroMemory(&hlist, sizeof(HEAPLIST32));
-			hlist.dwSize = sizeof(HEAPLIST32);		
-			if (good) starttime = GetTickCount();
-#ifdef _MSC_VER
-			if (heaplist_first(handle, &hlist))
-				{
-				/*
-				   following discussion on dev ML, exception on WinCE (or other Win
-				   platform) is theoretically of unknown origin; prevent infinite
-				   loop here when this theoretical case occurs; otherwise cope with
-				   the expected (MSDN documented) exception-throwing behaviour of
-				   Heap32Next() on WinCE.
-
-				   based on patch in original message by Tanguy Fautré (2009/03/02)
-			           Subject: RAND_poll() and CreateToolhelp32Snapshot() stability
-			     */
-				int ex_cnt_limit = 42; 
-				do
-					{
-					RAND_add(&hlist, hlist.dwSize, 3);
-					__try
-						{
-						ZeroMemory(&hentry, sizeof(HEAPENTRY32));
-					hentry.dwSize = sizeof(HEAPENTRY32);
-					if (heap_first(&hentry,
-						hlist.th32ProcessID,
-						hlist.th32HeapID))
-						{
-						int entrycnt = 80;
-						do
-							RAND_add(&hentry,
-								hentry.dwSize, 5);
-						while (heap_next(&hentry)
-						&& (!good || (GetTickCount()-starttime)<MAXDELAY)
-							&& --entrycnt > 0);
-						}
-						}
-					__except (EXCEPTION_EXECUTE_HANDLER)
-						{
-							/* ignore access violations when walking the heap list */
-							ex_cnt_limit--;
-						}
-					} while (heaplist_next(handle, &hlist) 
-						&& (!good || (GetTickCount()-starttime)<MAXDELAY)
-						&& ex_cnt_limit > 0);
-				}
-
-#else
-			if (heaplist_first(handle, &hlist))
-				{
-				do
-					{
-					RAND_add(&hlist, hlist.dwSize, 3);
-					hentry.dwSize = sizeof(HEAPENTRY32);
-					if (heap_first(&hentry,
-						hlist.th32ProcessID,
-						hlist.th32HeapID))
-						{
-						int entrycnt = 80;
-						do
-							RAND_add(&hentry,
-								hentry.dwSize, 5);
-						while (heap_next(&hentry)
-							&& --entrycnt > 0);
-						}
-					} while (heaplist_next(handle, &hlist) 
-						&& (!good || (GetTickCount()-starttime)<MAXDELAY));
-				}
-#endif
-
-			/* process walking */
-                        /* PROCESSENTRY32 contains 9 fields that will change
-                         * with each entry.  Consider each field a source of
-                         * 1 byte of entropy.
-                         */
-			p.dwSize = sizeof(PROCESSENTRY32);
-		
-			if (good) starttime = GetTickCount();
-			if (process_first(handle, &p))
-				do
-					RAND_add(&p, p.dwSize, 9);
-				while (process_next(handle, &p) && (!good || (GetTickCount()-starttime)<MAXDELAY));
-
-			/* thread walking */
-                        /* THREADENTRY32 contains 6 fields that will change
-                         * with each entry.  Consider each field a source of
-                         * 1 byte of entropy.
-                         */
-			t.dwSize = sizeof(THREADENTRY32);
-			if (good) starttime = GetTickCount();
-			if (thread_first(handle, &t))
-				do
-					RAND_add(&t, t.dwSize, 6);
-				while (thread_next(handle, &t) && (!good || (GetTickCount()-starttime)<MAXDELAY));
-
-			/* module walking */
-                        /* MODULEENTRY32 contains 9 fields that will change
-                         * with each entry.  Consider each field a source of
-                         * 1 byte of entropy.
-                         */
-			m.dwSize = sizeof(MODULEENTRY32);
-			if (good) starttime = GetTickCount();
-			if (module_first(handle, &m))
-				do
-					RAND_add(&m, m.dwSize, 9);
-				while (module_next(handle, &m)
-					       	&& (!good || (GetTickCount()-starttime)<MAXDELAY));
-			if (close_snap)
-				close_snap(handle);
-			else
-				CloseHandle(handle);
-
-			}
-
-		FreeLibrary(kernel);
-		}
-	}
-#endif /* !OPENSSL_SYS_WINCE */
-
-	/* timer data */
-	readtimer();
-	
-	/* memory usage statistics */
-	GlobalMemoryStatus(&m);
-	RAND_add(&m, sizeof(m), 1);
-
-	/* process ID */
-	w = GetCurrentProcessId();
-	RAND_add(&w, sizeof(w), 1);
-
-#if 0
-	printf("Exiting RAND_poll\n");
-#endif
-
-	return(1);
-}
-
-int RAND_event(UINT iMsg, WPARAM wParam, LPARAM lParam)
-        {
-        double add_entropy=0;
-
-        switch (iMsg)
-                {
-        case WM_KEYDOWN:
-                        {
-                        static WPARAM key;
-                        if (key != wParam)
-                                add_entropy = 0.05;
-                        key = wParam;
-                        }
-                        break;
-	case WM_MOUSEMOVE:
-                        {
-                        static int lastx,lasty,lastdx,lastdy;
-                        int x,y,dx,dy;
-
-                        x=LOWORD(lParam);
-                        y=HIWORD(lParam);
-                        dx=lastx-x;
-                        dy=lasty-y;
-                        if (dx != 0 && dy != 0 && dx-lastdx != 0 && dy-lastdy != 0)
-                                add_entropy=.2;
-                        lastx=x, lasty=y;
-                        lastdx=dx, lastdy=dy;
-                        }
-		break;
-		}
-
-	readtimer();
-        RAND_add(&iMsg, sizeof(iMsg), add_entropy);
-	RAND_add(&wParam, sizeof(wParam), 0);
-	RAND_add(&lParam, sizeof(lParam), 0);
- 
-	return (RAND_status());
-	}
-
-
-void RAND_screen(void) /* function available for backward compatibility */
-{
-	RAND_poll();
-	readscreen();
-}
-
-
-/* feed timing information to the PRNG */
-static void readtimer(void)
-{
-	DWORD w;
-	LARGE_INTEGER l;
-	static int have_perfc = 1;
-#if defined(_MSC_VER) && defined(_M_X86)
-	static int have_tsc = 1;
-	DWORD cyclecount;
-
-	if (have_tsc) {
-	  __try {
-	    __asm {
-	      _emit 0x0f
-	      _emit 0x31
-	      mov cyclecount, eax
-	      }
-	    RAND_add(&cyclecount, sizeof(cyclecount), 1);
-	  } __except(EXCEPTION_EXECUTE_HANDLER) {
-	    have_tsc = 0;
-	  }
-	}
-#else
-# define have_tsc 0
-#endif
-
-	if (have_perfc) {
-	  if (QueryPerformanceCounter(&l) == 0)
-	    have_perfc = 0;
-	  else
-	    RAND_add(&l, sizeof(l), 0);
-	}
-
-	if (!have_tsc && !have_perfc) {
-	  w = GetTickCount();
-	  RAND_add(&w, sizeof(w), 0);
-	}
-}
-
-/* feed screen contents to PRNG */
-/*****************************************************************************
- *
- * Created 960901 by Gertjan van Oosten, gertjan@West.NL, West Consulting B.V.
- *
- * Code adapted from
- * <URL:http://support.microsoft.com/default.aspx?scid=kb;[LN];97193>;
- * the original copyright message is:
- *
- *   (C) Copyright Microsoft Corp. 1993.  All rights reserved.
- *
- *   You have a royalty-free right to use, modify, reproduce and
- *   distribute the Sample Files (and/or any modified version) in
- *   any way you find useful, provided that you agree that
- *   Microsoft has no warranty obligations or liability for any
- *   Sample Application Files which are modified.
- */
-
-static void readscreen(void)
-{
-#if !defined(OPENSSL_SYS_WINCE) && !defined(OPENSSL_SYS_WIN32_CYGWIN)
-  HDC		hScrDC;		/* screen DC */
-  HDC		hMemDC;		/* memory DC */
-  HBITMAP	hBitmap;	/* handle for our bitmap */
-  HBITMAP	hOldBitmap;	/* handle for previous bitmap */
-  BITMAP	bm;		/* bitmap properties */
-  unsigned int	size;		/* size of bitmap */
-  char		*bmbits;	/* contents of bitmap */
-  int		w;		/* screen width */
-  int		h;		/* screen height */
-  int		y;		/* y-coordinate of screen lines to grab */
-  int		n = 16;		/* number of screen lines to grab at a time */
-
-  if (GetVersion() < 0x80000000 && OPENSSL_isservice()>0)
-    return;
-
-  /* Create a screen DC and a memory DC compatible to screen DC */
-  hScrDC = CreateDC(TEXT("DISPLAY"), NULL, NULL, NULL);
-  hMemDC = CreateCompatibleDC(hScrDC);
-
-  /* Get screen resolution */
-  w = GetDeviceCaps(hScrDC, HORZRES);
-  h = GetDeviceCaps(hScrDC, VERTRES);
-
-  /* Create a bitmap compatible with the screen DC */
-  hBitmap = CreateCompatibleBitmap(hScrDC, w, n);
-
-  /* Select new bitmap into memory DC */
-  hOldBitmap = SelectObject(hMemDC, hBitmap);
-
-  /* Get bitmap properties */
-  GetObject(hBitmap, sizeof(BITMAP), (LPSTR)&bm);
-  size = (unsigned int)bm.bmWidthBytes * bm.bmHeight * bm.bmPlanes;
-
-  bmbits = OPENSSL_malloc(size);
-  if (bmbits) {
-    /* Now go through the whole screen, repeatedly grabbing n lines */
-    for (y = 0; y < h-n; y += n)
-    	{
-	unsigned char md[MD_DIGEST_LENGTH];
-
-	/* Bitblt screen DC to memory DC */
-	BitBlt(hMemDC, 0, 0, w, n, hScrDC, 0, y, SRCCOPY);
-
-	/* Copy bitmap bits from memory DC to bmbits */
-	GetBitmapBits(hBitmap, size, bmbits);
-
-	/* Get the hash of the bitmap */
-	MD(bmbits,size,md);
-
-	/* Seed the random generator with the hash value */
-	RAND_add(md, MD_DIGEST_LENGTH, 0);
-	}
-
-    OPENSSL_free(bmbits);
-  }
-
-  /* Select old bitmap back into memory DC */
-  hBitmap = SelectObject(hMemDC, hOldBitmap);
-
-  /* Clean up */
-  DeleteObject(hBitmap);
-  DeleteDC(hMemDC);
-  DeleteDC(hScrDC);
-#endif /* !OPENSSL_SYS_WINCE */
-}
-
-#endif
diff --git a/crypto/rand/randfile.c b/crypto/rand/randfile.c
index 181056847e14..030e07f41826 100644
--- a/crypto/rand/randfile.c
+++ b/crypto/rand/randfile.c
@@ -75,9 +75,7 @@
 #ifndef NO_SYS_TYPES_H
 # include <sys/types.h>
 #endif
-#ifdef MAC_OS_pre_X
-# include <stat.h>
-#else
+#ifndef OPENSSL_NO_POSIX_IO
 # include <sys/stat.h>
 #endif
 
@@ -111,12 +109,15 @@ int RAND_load_file(const char *file, long bytes)
 	 * if bytes == -1, read complete file. */
 
 	MS_STATIC unsigned char buf[BUFSIZE];
+#ifndef OPENSSL_NO_POSIX_IO
 	struct stat sb;
+#endif
 	int i,ret=0,n;
 	FILE *in;
 
 	if (file == NULL) return(0);
 
+#ifndef OPENSSL_NO_POSIX_IO
 #ifdef PURIFY
 	/* struct stat can have padding and unused fields that may not be
 	 * initialized in the call to stat(). We need to clear the entire
@@ -125,9 +126,9 @@ int RAND_load_file(const char *file, long bytes)
 	 */
 	memset(&sb, 0, sizeof(sb));
 #endif
-
 	if (stat(file,&sb) < 0) return(0);
 	RAND_add(&sb,sizeof(sb),0.0);
+#endif
 	if (bytes == 0) return(ret);
 
 #ifdef OPENSSL_SYS_VMS
@@ -136,14 +137,16 @@ int RAND_load_file(const char *file, long bytes)
 	in=fopen(file,"rb");
 #endif
 	if (in == NULL) goto err;
-#if defined(S_ISBLK) && defined(S_ISCHR)
-	if (S_ISBLK(sb.st_mode) || S_ISCHR(sb.st_mode)) {
+#if defined(S_IFBLK) && defined(S_IFCHR) && !defined(OPENSSL_NO_POSIX_IO)
+	if (sb.st_mode & (S_IFBLK | S_IFCHR)) {
 	  /* this file is a device. we don't want read an infinite number
 	   * of bytes from a random device, nor do we want to use buffered
 	   * I/O because we will waste system entropy. 
 	   */
 	  bytes = (bytes == -1) ? 2048 : bytes; /* ok, is 2048 enough? */
+#ifndef OPENSSL_NO_SETVBUF_IONBF
 	  setvbuf(in, NULL, _IONBF, 0); /* don't do buffered reads */
+#endif /* ndef OPENSSL_NO_SETVBUF_IONBF */
 	}
 #endif
 	for (;;)
@@ -179,6 +182,7 @@ int RAND_write_file(const char *file)
 	int i,ret=0,rand_err=0;
 	FILE *out = NULL;
 	int n;
+#ifndef OPENSSL_NO_POSIX_IO
 	struct stat sb;
 	
 	i=stat(file,&sb);
@@ -194,14 +198,16 @@ int RAND_write_file(const char *file)
 	  }
 #endif
 	}
+#endif
 
-#if defined(O_CREAT) && !defined(OPENSSL_SYS_WIN32) && !defined(OPENSSL_SYS_VMS)
+#if defined(O_CREAT) && !defined(OPENSSL_NO_POSIX_IO) && !defined(OPENSSL_SYS_VMS)
 	{
-	/* For some reason Win32 can't write to files created this way */
-	
+#ifndef O_BINARY
+#define O_BINARY 0
+#endif
 	/* chmod(..., 0600) is too late to protect the file,
 	 * permissions should be restrictive from the start */
-	int fd = open(file, O_CREAT, 0600);
+	int fd = open(file, O_WRONLY|O_CREAT|O_BINARY, 0600);
 	if (fd != -1)
 		out = fdopen(fd, "wb");
 	}
@@ -266,7 +272,6 @@ const char *RAND_file_name(char *buf, size_t size)
 	{
 	char *s=NULL;
 #ifdef __OpenBSD__
-	int ok = 0;
 	struct stat sb;
 #endif
 
@@ -294,9 +299,6 @@ const char *RAND_file_name(char *buf, size_t size)
 			BUF_strlcat(buf,"/",size);
 #endif
 			BUF_strlcat(buf,RFILE,size);
-#ifdef __OpenBSD__
-			ok = 1;
-#endif
 			}
 		else
 		  	buf[0] = '\0'; /* no file name */
@@ -310,7 +312,7 @@ const char *RAND_file_name(char *buf, size_t size)
 	 * to something hopefully decent if that isn't available. 
 	 */
 
-	if (!ok)
+	if (!buf[0])
 		if (BUF_strlcpy(buf,"/dev/arandom",size) >= size) {
 			return(NULL);
 		}	
diff --git a/crypto/rc2/Makefile b/crypto/rc2/Makefile
index 4b6292b65f59..8a9d49ab5eb8 100644
--- a/crypto/rc2/Makefile
+++ b/crypto/rc2/Makefile
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -79,7 +79,7 @@ rc2_cbc.o: rc2_cbc.c rc2_locl.h
 rc2_ecb.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
 rc2_ecb.o: ../../include/openssl/rc2.h rc2_ecb.c rc2_locl.h
 rc2_skey.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
-rc2_skey.o: ../../include/openssl/fips.h ../../include/openssl/opensslconf.h
+rc2_skey.o: ../../include/openssl/opensslconf.h
 rc2_skey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rc2_skey.o: ../../include/openssl/rc2.h ../../include/openssl/safestack.h
 rc2_skey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
diff --git a/crypto/rc2/rc2_skey.c b/crypto/rc2/rc2_skey.c
index 4e000e5b992a..6668ac011f0d 100644
--- a/crypto/rc2/rc2_skey.c
+++ b/crypto/rc2/rc2_skey.c
@@ -56,15 +56,11 @@
  * [including the GNU Public Licence.]
  */
 
-#include <openssl/rc2.h>
 #include <openssl/crypto.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/rc2.h>
 #include "rc2_locl.h"
 
-static unsigned char key_table[256]={
+static const unsigned char key_table[256]={
 	0xd9,0x78,0xf9,0xc4,0x19,0xdd,0xb5,0xed,0x28,0xe9,0xfd,0x79,
 	0x4a,0xa0,0xd8,0x9d,0xc6,0x7e,0x37,0x83,0x2b,0x76,0x53,0x8e,
 	0x62,0x4c,0x64,0x88,0x44,0x8b,0xfb,0xa2,0x17,0x9a,0x59,0xf5,
@@ -99,18 +95,13 @@ static unsigned char key_table[256]={
  * BSAFE uses the 'retarded' version.  What I previously shipped is
  * the same as specifying 1024 for the 'bits' parameter.  Bsafe uses
  * a version where the bits parameter is the same as len*8 */
-
-#ifdef OPENSSL_FIPS
 void RC2_set_key(RC2_KEY *key, int len, const unsigned char *data, int bits)
+#ifdef OPENSSL_FIPS
 	{
-	if (FIPS_mode())
-		FIPS_BAD_ABORT(RC2)
+	fips_cipher_abort(RC2);
 	private_RC2_set_key(key, len, data, bits);
 	}
-void private_RC2_set_key(RC2_KEY *key, int len, const unsigned char *data,
-								int bits)
-#else
-void RC2_set_key(RC2_KEY *key, int len, const unsigned char *data, int bits)
+void private_RC2_set_key(RC2_KEY *key, int len, const unsigned char *data, int bits)
 #endif
 	{
 	int i,j;
diff --git a/crypto/rc4/Makefile b/crypto/rc4/Makefile
index f0bd7678fc92..1614d479619f 100644
--- a/crypto/rc4/Makefile
+++ b/crypto/rc4/Makefile
@@ -21,8 +21,8 @@ TEST=rc4test.c
 APPS=
 
 LIB=$(TOP)/libcrypto.a
-LIBSRC=rc4_skey.c rc4_enc.c rc4_fblk.c
-LIBOBJ=$(RC4_ENC) rc4_fblk.o
+LIBSRC=rc4_skey.c rc4_enc.c rc4_utl.c
+LIBOBJ=$(RC4_ENC) rc4_utl.o
 
 SRC= $(LIBSRC)
 
@@ -37,29 +37,34 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-# ELF
-rx86-elf.s: asm/rc4-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) rc4-586.pl elf $(CFLAGS) > ../$@)
-# COFF
-rx86-cof.s: asm/rc4-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) rc4-586.pl coff $(CFLAGS) > ../$@)
-# a.out
-rx86-out.s: asm/rc4-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) rc4-586.pl a.out $(CFLAGS) > ../$@)
+rc4-586.s:	asm/rc4-586.pl ../perlasm/x86asm.pl
+	$(PERL) asm/rc4-586.pl $(PERLASM_SCHEME) $(CFLAGS) > $@
 
-rc4-x86_64.s: asm/rc4-x86_64.pl;	$(PERL) asm/rc4-x86_64.pl $@
+rc4-x86_64.s: asm/rc4-x86_64.pl
+	$(PERL) asm/rc4-x86_64.pl $(PERLASM_SCHEME) > $@
+rc4-md5-x86_64.s:	asm/rc4-md5-x86_64.pl
+	$(PERL) asm/rc4-md5-x86_64.pl $(PERLASM_SCHEME) > $@
 
-rc4-ia64.s: asm/rc4-ia64.S
+rc4-ia64.S: asm/rc4-ia64.pl
+	$(PERL) asm/rc4-ia64.pl $(CFLAGS) > $@
+
+rc4-parisc.s:	asm/rc4-parisc.pl
+	$(PERL) asm/rc4-parisc.pl $(PERLASM_SCHEME) $@
+
+rc4-ia64.s: rc4-ia64.S
 	@case `awk '/^#define RC4_INT/{print$$NF}' $(TOP)/include/openssl/opensslconf.h` in \
-	int)	set -x; $(CC) $(CFLAGS) -DSZ=4 -E asm/rc4-ia64.S > $@ ;; \
-	char)	set -x; $(CC) $(CFLAGS) -DSZ=1 -E asm/rc4-ia64.S > $@ ;; \
+	int)	set -x; $(CC) $(CFLAGS) -DSZ=4 -E rc4-ia64.S > $@ ;; \
+	char)	set -x; $(CC) $(CFLAGS) -DSZ=1 -E rc4-ia64.S > $@ ;; \
 	*)	exit 1 ;; \
 	esac
 
+# GNU make "catch all"
+rc4-%.s:	asm/rc4-%.pl;	$(PERL) $< $(PERLASM_SCHEME) $@
+
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
 
@@ -105,21 +110,16 @@ rc4_enc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rc4_enc.o: ../../include/openssl/rc4.h ../../include/openssl/safestack.h
 rc4_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 rc4_enc.o: ../cryptlib.h rc4_enc.c rc4_locl.h
-rc4_fblk.o: ../../e_os.h ../../include/openssl/bio.h
-rc4_fblk.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-rc4_fblk.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-rc4_fblk.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-rc4_fblk.o: ../../include/openssl/opensslconf.h
-rc4_fblk.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-rc4_fblk.o: ../../include/openssl/rc4.h ../../include/openssl/safestack.h
-rc4_fblk.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-rc4_fblk.o: ../cryptlib.h rc4_fblk.c rc4_locl.h
 rc4_skey.o: ../../e_os.h ../../include/openssl/bio.h
 rc4_skey.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 rc4_skey.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-rc4_skey.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-rc4_skey.o: ../../include/openssl/opensslconf.h
+rc4_skey.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 rc4_skey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rc4_skey.o: ../../include/openssl/rc4.h ../../include/openssl/safestack.h
 rc4_skey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 rc4_skey.o: ../cryptlib.h rc4_locl.h rc4_skey.c
+rc4_utl.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+rc4_utl.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+rc4_utl.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rc4.h
+rc4_utl.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
+rc4_utl.o: ../../include/openssl/symhacks.h rc4_utl.c
diff --git a/crypto/rc4/asm/rc4-586.pl b/crypto/rc4/asm/rc4-586.pl
index ef7eee766cb9..5c9ac6ad286e 100644
--- a/crypto/rc4/asm/rc4-586.pl
+++ b/crypto/rc4/asm/rc4-586.pl
@@ -1,14 +1,21 @@
-#!/usr/local/bin/perl
+#!/usr/bin/env perl
+
+# ====================================================================
+# [Re]written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
 
 # At some point it became apparent that the original SSLeay RC4
-# assembler implementation performs suboptimaly on latest IA-32
+# assembler implementation performs suboptimally on latest IA-32
 # microarchitectures. After re-tuning performance has changed as
 # following:
 #
-# Pentium	+0%
-# Pentium III	+17%
-# AMD		+52%(*)
-# P4		+180%(**)
+# Pentium	-10%
+# Pentium III	+12%
+# AMD		+50%(*)
+# P4		+250%(**)
 #
 # (*)	This number is actually a trade-off:-) It's possible to
 #	achieve	+72%, but at the cost of -48% off PIII performance.
@@ -17,214 +24,387 @@
 #	For reference! This code delivers ~80% of rc4-amd64.pl
 #	performance on the same Opteron machine.
 # (**)	This number requires compressed key schedule set up by
-#	RC4_set_key and therefore doesn't apply to 0.9.7 [option for
-#	compressed key schedule is implemented in 0.9.8 and later,
-#	see commentary section in rc4_skey.c for further details].
+#	RC4_set_key [see commentary below for further details].
 #
 #					<appro@fy.chalmers.se>
 
-push(@INC,"perlasm","../../perlasm");
+# May 2011
+#
+# Optimize for Core2 and Westmere [and incidentally Opteron]. Current
+# performance in cycles per processed byte (less is better) and
+# improvement relative to previous version of this module is:
+#
+# Pentium	10.2			# original numbers
+# Pentium III	7.8(*)
+# Intel P4	7.5
+#
+# Opteron	6.1/+20%		# new MMX numbers
+# Core2		5.3/+67%(**)
+# Westmere	5.1/+94%(**)
+# Sandy Bridge	5.0/+8%
+# Atom		12.6/+6%
+#
+# (*)	PIII can actually deliver 6.6 cycles per byte with MMX code,
+#	but this specific code performs poorly on Core2. And vice
+#	versa, below MMX/SSE code delivering 5.8/7.1 on Core2 performs
+#	poorly on PIII, at 8.0/14.5:-( As PIII is not a "hot" CPU
+#	[anymore], I chose to discard PIII-specific code path and opt
+#	for original IALU-only code, which is why MMX/SSE code path
+#	is guarded by SSE2 bit (see below), not MMX/SSE.
+# (**)	Performance vs. block size on Core2 and Westmere had a maximum
+#	at ... 64 bytes block size. And it was quite a maximum, 40-60%
+#	in comparison to largest 8KB block size. Above improvement
+#	coefficients are for the largest block size.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],"rc4-586.pl");
 
-$x="eax";
-$y="ebx";
+$xx="eax";
+$yy="ebx";
 $tx="ecx";
 $ty="edx";
-$in="esi";
-$out="edi";
-$d="ebp";
-
-&RC4("RC4");
-
-&asm_finish();
-
-sub RC4_loop
-	{
-	local($n,$p,$char)=@_;
-
-	&comment("Round $n");
-
-	if ($char)
-		{
-		if ($p >= 0)
-			{
-			 &mov($ty,	&swtmp(2));
-			&cmp($ty,	$in);
-			 &jbe(&label("finished"));
-			&inc($in);
-			}
-		else
-			{
-			&add($ty,	8);
-			 &inc($in);
-			&cmp($ty,	$in);
-			 &jb(&label("finished"));
-			&mov(&swtmp(2),	$ty);
-			}
-		}
-	# Moved out
-	# &mov(	$tx,		&DWP(0,$d,$x,4)) if $p < 0;
-
-	&add(	&LB($y),	&LB($tx));
-	&mov(	$ty,		&DWP(0,$d,$y,4));
-	 # XXX
-	&mov(	&DWP(0,$d,$x,4),$ty);
-	 &add(	$ty,		$tx);
-	&mov(	&DWP(0,$d,$y,4),$tx);
-	 &and(	$ty,		0xff);
-	 &inc(	&LB($x));			# NEXT ROUND
-	&mov(	$tx,		&DWP(0,$d,$x,4)) if $p < 1; # NEXT ROUND
-	 &mov(	$ty,		&DWP(0,$d,$ty,4));
-
-	if (!$char)
-		{
-		#moved up into last round
-		if ($p >= 1)
-			{
-			&add(	$out,	8)
-			}
-		&movb(	&BP($n,"esp","",0),	&LB($ty));
-		}
-	else
-		{
-		# Note in+=8 has occured
-		&movb(	&HB($ty),	&BP(-1,$in,"",0));
-		 # XXX
-		&xorb(&LB($ty),		&HB($ty));
-		 # XXX
-		&movb(&BP($n,$out,"",0),&LB($ty));
-		}
+$inp="esi";
+$out="ebp";
+$dat="edi";
+
+sub RC4_loop {
+  my $i=shift;
+  my $func = ($i==0)?*mov:*or;
+
+	&add	(&LB($yy),&LB($tx));
+	&mov	($ty,&DWP(0,$dat,$yy,4));
+	&mov	(&DWP(0,$dat,$yy,4),$tx);
+	&mov	(&DWP(0,$dat,$xx,4),$ty);
+	&add	($ty,$tx);
+	&inc	(&LB($xx));
+	&and	($ty,0xff);
+	&ror	($out,8)	if ($i!=0);
+	if ($i<3) {
+	  &mov	($tx,&DWP(0,$dat,$xx,4));
+	} else {
+	  &mov	($tx,&wparam(3));	# reload [re-biased] out
 	}
-
-
-sub RC4
-	{
-	local($name)=@_;
-
-	&function_begin_B($name,"");
-
-	&mov($ty,&wparam(1));		# len
-	&cmp($ty,0);
-	&jne(&label("proceed"));
-	&ret();
-	&set_label("proceed");
-
-	&comment("");
-
-	&push("ebp");
-	 &push("ebx");
-	&push("esi");
-	 &xor(	$x,	$x);		# avoid partial register stalls
-	&push("edi");
-	 &xor(	$y,	$y);		# avoid partial register stalls
-	&mov(	$d,	&wparam(0));	# key
-	 &mov(	$in,	&wparam(2));
-
-	&movb(	&LB($x),	&BP(0,$d,"",1));
-	 &movb(	&LB($y),	&BP(4,$d,"",1));
-
-	&mov(	$out,	&wparam(3));
-	 &inc(	&LB($x));
-
-	&stack_push(3);	# 3 temp variables
-	 &add(	$d,	8);
-
-	# detect compressed schedule, see commentary section in rc4_skey.c...
-	# in 0.9.7 context ~50 bytes below RC4_CHAR label remain redundant,
-	# as compressed key schedule is set up in 0.9.8 and later.
-	&cmp(&DWP(256,$d),-1);
-	&je(&label("RC4_CHAR"));
-
-	 &lea(	$ty,	&DWP(-8,$ty,$in));
-
-	# check for 0 length input
-
-	 &mov(	&swtmp(2),	$ty);	# this is now address to exit at
-	&mov(	$tx,	&DWP(0,$d,$x,4));
-
-	 &cmp(	$ty,	$in);
-	&jb(	&label("end")); # less than 8 bytes
-
-	&set_label("start");
-
-	# filling DELAY SLOT
-	&add(	$in,	8);
-
-	&RC4_loop(0,-1,0);
-	&RC4_loop(1,0,0);
-	&RC4_loop(2,0,0);
-	&RC4_loop(3,0,0);
-	&RC4_loop(4,0,0);
-	&RC4_loop(5,0,0);
-	&RC4_loop(6,0,0);
-	&RC4_loop(7,1,0);
-	
-	&comment("apply the cipher text");
-	# xor the cipher data with input
-
-	#&add(	$out,	8); #moved up into last round
-
-	&mov(	$tx,	&swtmp(0));
-	 &mov(	$ty,	&DWP(-8,$in,"",0));
-	&xor(	$tx,	$ty);
-	 &mov(	$ty,	&DWP(-4,$in,"",0)); 
-	&mov(	&DWP(-8,$out,"",0),	$tx);
-	 &mov(	$tx,	&swtmp(1));
-	&xor(	$tx,	$ty);
-	 &mov(	$ty,	&swtmp(2));	# load end ptr;
-	&mov(	&DWP(-4,$out,"",0),	$tx);
-	 &mov(	$tx,		&DWP(0,$d,$x,4));
-	&cmp($in,	$ty);
-	 &jbe(&label("start"));
-
-	&set_label("end");
-
-	# There is quite a bit of extra crap in RC4_loop() for this
-	# first round
-	&RC4_loop(0,-1,1);
-	&RC4_loop(1,0,1);
-	&RC4_loop(2,0,1);
-	&RC4_loop(3,0,1);
-	&RC4_loop(4,0,1);
-	&RC4_loop(5,0,1);
-	&RC4_loop(6,1,1);
-
-	&jmp(&label("finished"));
-
-	&align(16);
-	# this is essentially Intel P4 specific codepath, see rc4_skey.c,
-	# and is engaged in 0.9.8 and later context...
-	&set_label("RC4_CHAR");
-
-	&lea	($ty,&DWP(0,$in,$ty));
-	&mov	(&swtmp(2),$ty);
-	&movz	($tx,&BP(0,$d,$x));
-
+	&$func	($out,&DWP(0,$dat,$ty,4));
+}
+
+if ($alt=0) {
+  # >20% faster on Atom and Sandy Bridge[!], 8% faster on Opteron,
+  # but ~40% slower on Core2 and Westmere... Attempt to add movz
+  # brings down Opteron by 25%, Atom and Sandy Bridge by 15%, yet
+  # on Core2 with movz it's almost 20% slower than below alternative
+  # code... Yes, it's a total mess...
+  my @XX=($xx,$out);
+  $RC4_loop_mmx = sub {		# SSE actually...
+    my $i=shift;
+    my $j=$i<=0?0:$i>>1;
+    my $mm=$i<=0?"mm0":"mm".($i&1);
+
+	&add	(&LB($yy),&LB($tx));
+	&lea	(@XX[1],&DWP(1,@XX[0]));
+	&pxor	("mm2","mm0")				if ($i==0);
+	&psllq	("mm1",8)				if ($i==0);
+	&and	(@XX[1],0xff);
+	&pxor	("mm0","mm0")				if ($i<=0);
+	&mov	($ty,&DWP(0,$dat,$yy,4));
+	&mov	(&DWP(0,$dat,$yy,4),$tx);
+	&pxor	("mm1","mm2")				if ($i==0);
+	&mov	(&DWP(0,$dat,$XX[0],4),$ty);
+	&add	(&LB($ty),&LB($tx));
+	&movd	(@XX[0],"mm7")				if ($i==0);
+	&mov	($tx,&DWP(0,$dat,@XX[1],4));
+	&pxor	("mm1","mm1")				if ($i==1);
+	&movq	("mm2",&QWP(0,$inp))			if ($i==1);
+	&movq	(&QWP(-8,(@XX[0],$inp)),"mm1")		if ($i==0);
+	&pinsrw	($mm,&DWP(0,$dat,$ty,4),$j);
+
+	push	(@XX,shift(@XX))			if ($i>=0);
+  }
+} else {
+  # Using pinsrw here improves performane on Intel CPUs by 2-3%, but
+  # brings down AMD by 7%...
+  $RC4_loop_mmx = sub {
+    my $i=shift;
+
+	&add	(&LB($yy),&LB($tx));
+	&psllq	("mm1",8*(($i-1)&7))			if (abs($i)!=1);
+	&mov	($ty,&DWP(0,$dat,$yy,4));
+	&mov	(&DWP(0,$dat,$yy,4),$tx);
+	&mov	(&DWP(0,$dat,$xx,4),$ty);
+	&inc	($xx);
+	&add	($ty,$tx);
+	&movz	($xx,&LB($xx));				# (*)
+	&movz	($ty,&LB($ty));				# (*)
+	&pxor	("mm2",$i==1?"mm0":"mm1")		if ($i>=0);
+	&movq	("mm0",&QWP(0,$inp))			if ($i<=0);
+	&movq	(&QWP(-8,($out,$inp)),"mm2")		if ($i==0);
+	&mov	($tx,&DWP(0,$dat,$xx,4));
+	&movd	($i>0?"mm1":"mm2",&DWP(0,$dat,$ty,4));
+
+	# (*)	This is the key to Core2 and Westmere performance.
+	#	Whithout movz out-of-order execution logic confuses
+	#	itself and fails to reorder loads and stores. Problem
+	#	appears to be fixed in Sandy Bridge...
+  }
+}
+
+&external_label("OPENSSL_ia32cap_P");
+
+# void RC4(RC4_KEY *key,size_t len,const unsigned char *inp,unsigned char *out);
+&function_begin("RC4");
+	&mov	($dat,&wparam(0));	# load key schedule pointer
+	&mov	($ty, &wparam(1));	# load len
+	&mov	($inp,&wparam(2));	# load inp
+	&mov	($out,&wparam(3));	# load out
+
+	&xor	($xx,$xx);		# avoid partial register stalls
+	&xor	($yy,$yy);
+
+	&cmp	($ty,0);		# safety net
+	&je	(&label("abort"));
+
+	&mov	(&LB($xx),&BP(0,$dat));	# load key->x
+	&mov	(&LB($yy),&BP(4,$dat));	# load key->y
+	&add	($dat,8);
+
+	&lea	($tx,&DWP(0,$inp,$ty));
+	&sub	($out,$inp);		# re-bias out
+	&mov	(&wparam(1),$tx);	# save input+len
+
+	&inc	(&LB($xx));
+
+	# detect compressed key schedule...
+	&cmp	(&DWP(256,$dat),-1);
+	&je	(&label("RC4_CHAR"));
+
+	&mov	($tx,&DWP(0,$dat,$xx,4));
+
+	&and	($ty,-4);		# how many 4-byte chunks?
+	&jz	(&label("loop1"));
+
+	&test	($ty,-8);
+	&mov	(&wparam(3),$out);	# $out as accumulator in these loops
+	&jz	(&label("go4loop4"));
+
+	&picmeup($out,"OPENSSL_ia32cap_P");
+	&bt	(&DWP(0,$out),26);	# check SSE2 bit [could have been MMX]
+	&jnc	(&label("go4loop4"));
+
+	&mov	($out,&wparam(3))	if (!$alt);
+	&movd	("mm7",&wparam(3))	if ($alt);
+	&and	($ty,-8);
+	&lea	($ty,&DWP(-8,$inp,$ty));
+	&mov	(&DWP(-4,$dat),$ty);	# save input+(len/8)*8-8
+
+	&$RC4_loop_mmx(-1);
+	&jmp(&label("loop_mmx_enter"));
+
+	&set_label("loop_mmx",16);
+		&$RC4_loop_mmx(0);
+	&set_label("loop_mmx_enter");
+		for 	($i=1;$i<8;$i++) { &$RC4_loop_mmx($i); }
+		&mov	($ty,$yy);
+		&xor	($yy,$yy);		# this is second key to Core2
+		&mov	(&LB($yy),&LB($ty));	# and Westmere performance...
+		&cmp	($inp,&DWP(-4,$dat));
+		&lea	($inp,&DWP(8,$inp));
+	&jb	(&label("loop_mmx"));
+
+    if ($alt) {
+	&movd	($out,"mm7");
+	&pxor	("mm2","mm0");
+	&psllq	("mm1",8);
+	&pxor	("mm1","mm2");
+	&movq	(&QWP(-8,$out,$inp),"mm1");
+    } else {
+	&psllq	("mm1",56);
+	&pxor	("mm2","mm1");
+	&movq	(&QWP(-8,$out,$inp),"mm2");
+    }
+	&emms	();
+
+	&cmp	($inp,&wparam(1));	# compare to input+len
+	&je	(&label("done"));
+	&jmp	(&label("loop1"));
+
+&set_label("go4loop4",16);
+	&lea	($ty,&DWP(-4,$inp,$ty));
+	&mov	(&wparam(2),$ty);	# save input+(len/4)*4-4
+
+	&set_label("loop4");
+		for ($i=0;$i<4;$i++) { RC4_loop($i); }
+		&ror	($out,8);
+		&xor	($out,&DWP(0,$inp));
+		&cmp	($inp,&wparam(2));	# compare to input+(len/4)*4-4
+		&mov	(&DWP(0,$tx,$inp),$out);# $tx holds re-biased out here
+		&lea	($inp,&DWP(4,$inp));
+		&mov	($tx,&DWP(0,$dat,$xx,4));
+	&jb	(&label("loop4"));
+
+	&cmp	($inp,&wparam(1));	# compare to input+len
+	&je	(&label("done"));
+	&mov	($out,&wparam(3));	# restore $out
+
+	&set_label("loop1",16);
+		&add	(&LB($yy),&LB($tx));
+		&mov	($ty,&DWP(0,$dat,$yy,4));
+		&mov	(&DWP(0,$dat,$yy,4),$tx);
+		&mov	(&DWP(0,$dat,$xx,4),$ty);
+		&add	($ty,$tx);
+		&inc	(&LB($xx));
+		&and	($ty,0xff);
+		&mov	($ty,&DWP(0,$dat,$ty,4));
+		&xor	(&LB($ty),&BP(0,$inp));
+		&lea	($inp,&DWP(1,$inp));
+		&mov	($tx,&DWP(0,$dat,$xx,4));
+		&cmp	($inp,&wparam(1));	# compare to input+len
+		&mov	(&BP(-1,$out,$inp),&LB($ty));
+	&jb	(&label("loop1"));
+
+	&jmp	(&label("done"));
+
+# this is essentially Intel P4 specific codepath...
+&set_label("RC4_CHAR",16);
+	&movz	($tx,&BP(0,$dat,$xx));
 	# strangely enough unrolled loop performs over 20% slower...
-	&set_label("RC4_CHAR_loop");
-		&add	(&LB($y),&LB($tx));
-		&movz	($ty,&BP(0,$d,$y));
-		&movb	(&BP(0,$d,$y),&LB($tx));
-		&movb	(&BP(0,$d,$x),&LB($ty));
+	&set_label("cloop1");
+		&add	(&LB($yy),&LB($tx));
+		&movz	($ty,&BP(0,$dat,$yy));
+		&mov	(&BP(0,$dat,$yy),&LB($tx));
+		&mov	(&BP(0,$dat,$xx),&LB($ty));
 		&add	(&LB($ty),&LB($tx));
-		&movz	($ty,&BP(0,$d,$ty));
-		&add	(&LB($x),1);
-		&xorb	(&LB($ty),&BP(0,$in));
-		&lea	($in,&DWP(1,$in));
-		&movz	($tx,&BP(0,$d,$x));
-		&cmp	($in,&swtmp(2));
-		&movb	(&BP(0,$out),&LB($ty));
-		&lea	($out,&DWP(1,$out));
-	&jb	(&label("RC4_CHAR_loop"));
-
-	&set_label("finished");
-	&dec(	$x);
-	 &stack_pop(3);
-	&movb(	&BP(-4,$d,"",0),&LB($y));
-	 &movb(	&BP(-8,$d,"",0),&LB($x));
-
-	&function_end($name);
-	}
+		&movz	($ty,&BP(0,$dat,$ty));
+		&add	(&LB($xx),1);
+		&xor	(&LB($ty),&BP(0,$inp));
+		&lea	($inp,&DWP(1,$inp));
+		&movz	($tx,&BP(0,$dat,$xx));
+		&cmp	($inp,&wparam(1));
+		&mov	(&BP(-1,$out,$inp),&LB($ty));
+	&jb	(&label("cloop1"));
+
+&set_label("done");
+	&dec	(&LB($xx));
+	&mov	(&DWP(-4,$dat),$yy);		# save key->y
+	&mov	(&BP(-8,$dat),&LB($xx));	# save key->x
+&set_label("abort");
+&function_end("RC4");
+
+########################################################################
+
+$inp="esi";
+$out="edi";
+$idi="ebp";
+$ido="ecx";
+$idx="edx";
+
+# void RC4_set_key(RC4_KEY *key,int len,const unsigned char *data);
+&function_begin("private_RC4_set_key");
+	&mov	($out,&wparam(0));		# load key
+	&mov	($idi,&wparam(1));		# load len
+	&mov	($inp,&wparam(2));		# load data
+	&picmeup($idx,"OPENSSL_ia32cap_P");
+
+	&lea	($out,&DWP(2*4,$out));		# &key->data
+	&lea	($inp,&DWP(0,$inp,$idi));	# $inp to point at the end
+	&neg	($idi);
+	&xor	("eax","eax");
+	&mov	(&DWP(-4,$out),$idi);		# borrow key->y
+
+	&bt	(&DWP(0,$idx),20);		# check for bit#20
+	&jc	(&label("c1stloop"));
+
+&set_label("w1stloop",16);
+	&mov	(&DWP(0,$out,"eax",4),"eax");	# key->data[i]=i;
+	&add	(&LB("eax"),1);			# i++;
+	&jnc	(&label("w1stloop"));
+
+	&xor	($ido,$ido);
+	&xor	($idx,$idx);
+
+&set_label("w2ndloop",16);
+	&mov	("eax",&DWP(0,$out,$ido,4));
+	&add	(&LB($idx),&BP(0,$inp,$idi));
+	&add	(&LB($idx),&LB("eax"));
+	&add	($idi,1);
+	&mov	("ebx",&DWP(0,$out,$idx,4));
+	&jnz	(&label("wnowrap"));
+	  &mov	($idi,&DWP(-4,$out));
+	&set_label("wnowrap");
+	&mov	(&DWP(0,$out,$idx,4),"eax");
+	&mov	(&DWP(0,$out,$ido,4),"ebx");
+	&add	(&LB($ido),1);
+	&jnc	(&label("w2ndloop"));
+&jmp	(&label("exit"));
+
+# Unlike all other x86 [and x86_64] implementations, Intel P4 core
+# [including EM64T] was found to perform poorly with above "32-bit" key
+# schedule, a.k.a. RC4_INT. Performance improvement for IA-32 hand-coded
+# assembler turned out to be 3.5x if re-coded for compressed 8-bit one,
+# a.k.a. RC4_CHAR! It's however inappropriate to just switch to 8-bit
+# schedule for x86[_64], because non-P4 implementations suffer from
+# significant performance losses then, e.g. PIII exhibits >2x
+# deterioration, and so does Opteron. In order to assure optimal
+# all-round performance, we detect P4 at run-time and set up compressed
+# key schedule, which is recognized by RC4 procedure.
+
+&set_label("c1stloop",16);
+	&mov	(&BP(0,$out,"eax"),&LB("eax"));	# key->data[i]=i;
+	&add	(&LB("eax"),1);			# i++;
+	&jnc	(&label("c1stloop"));
+
+	&xor	($ido,$ido);
+	&xor	($idx,$idx);
+	&xor	("ebx","ebx");
+
+&set_label("c2ndloop",16);
+	&mov	(&LB("eax"),&BP(0,$out,$ido));
+	&add	(&LB($idx),&BP(0,$inp,$idi));
+	&add	(&LB($idx),&LB("eax"));
+	&add	($idi,1);
+	&mov	(&LB("ebx"),&BP(0,$out,$idx));
+	&jnz	(&label("cnowrap"));
+	  &mov	($idi,&DWP(-4,$out));
+	&set_label("cnowrap");
+	&mov	(&BP(0,$out,$idx),&LB("eax"));
+	&mov	(&BP(0,$out,$ido),&LB("ebx"));
+	&add	(&LB($ido),1);
+	&jnc	(&label("c2ndloop"));
+
+	&mov	(&DWP(256,$out),-1);		# mark schedule as compressed
+
+&set_label("exit");
+	&xor	("eax","eax");
+	&mov	(&DWP(-8,$out),"eax");		# key->x=0;
+	&mov	(&DWP(-4,$out),"eax");		# key->y=0;
+&function_end("private_RC4_set_key");
+
+# const char *RC4_options(void);
+&function_begin_B("RC4_options");
+	&call	(&label("pic_point"));
+&set_label("pic_point");
+	&blindpop("eax");
+	&lea	("eax",&DWP(&label("opts")."-".&label("pic_point"),"eax"));
+	&picmeup("edx","OPENSSL_ia32cap_P");
+	&mov	("edx",&DWP(0,"edx"));
+	&bt	("edx",20);
+	&jc	(&label("1xchar"));
+	&bt	("edx",26);
+	&jnc	(&label("ret"));
+	&add	("eax",25);
+	&ret	();
+&set_label("1xchar");
+	&add	("eax",12);
+&set_label("ret");
+	&ret	();
+&set_label("opts",64);
+&asciz	("rc4(4x,int)");
+&asciz	("rc4(1x,char)");
+&asciz	("rc4(8x,mmx)");
+&asciz	("RC4 for x86, CRYPTOGAMS by <appro\@openssl.org>");
+&align	(64);
+&function_end_B("RC4_options");
+
+&asm_finish();
 
diff --git a/crypto/rc4/asm/rc4-ia64.S b/crypto/rc4/asm/rc4-ia64.S
deleted file mode 100644
index 8210c47d049d..000000000000
--- a/crypto/rc4/asm/rc4-ia64.S
+++ /dev/null
@@ -1,159 +0,0 @@
-// ====================================================================
-// Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
-// project.
-//
-// Rights for redistribution and usage in source and binary forms are
-// granted according to the OpenSSL license. Warranty of any kind is
-// disclaimed.
-// ====================================================================
-
-.ident  "rc4-ia64.S, Version 2.0"
-.ident  "IA-64 ISA artwork by Andy Polyakov <appro@fy.chalmers.se>"
-
-// What's wrong with compiler generated code? Because of the nature of
-// C language, compiler doesn't [dare to] reorder load and stores. But
-// being memory-bound, RC4 should benefit from reorder [on in-order-
-// execution core such as IA-64]. But what can we reorder? At the very
-// least we can safely reorder references to key schedule in respect
-// to input and output streams. Secondly, from the first [close] glance
-// it appeared that it's possible to pull up some references to
-// elements of the key schedule itself. Original rationale ["prior
-// loads are not safe only for "degenerated" key schedule, when some
-// elements equal to the same value"] was kind of sloppy. I should have
-// formulated as it really was: if we assume that pulling up reference
-// to key[x+1] is not safe, then it would mean that key schedule would
-// "degenerate," which is never the case. The problem is that this
-// holds true in respect to references to key[x], but not to key[y].
-// Legitimate "collisions" do occur within every 256^2 bytes window.
-// Fortunately there're enough free instruction slots to keep prior
-// reference to key[x+1], detect "collision" and compensate for it.
-// All this without sacrificing a single clock cycle:-) Throughput is
-// ~210MBps on 900MHz CPU, which is is >3x faster than gcc generated
-// code and +30% - if compared to HP-UX C. Unrolling loop below should
-// give >30% on top of that...
-
-.text
-.explicit
-
-#if defined(_HPUX_SOURCE) && !defined(_LP64)
-# define ADDP	addp4
-#else
-# define ADDP	add
-#endif
-
-#ifndef SZ
-#define SZ	4	// this is set to sizeof(RC4_INT)
-#endif
-// SZ==4 seems to be optimal. At least SZ==8 is not any faster, not for
-// assembler implementation, while SZ==1 code is ~30% slower.
-#if SZ==1	// RC4_INT is unsigned char
-# define	LDKEY	ld1
-# define	STKEY	st1
-# define	OFF	0
-#elif SZ==4	// RC4_INT is unsigned int
-# define	LDKEY	ld4
-# define	STKEY	st4
-# define	OFF	2
-#elif SZ==8	// RC4_INT is unsigned long
-# define	LDKEY	ld8
-# define	STKEY	st8
-# define	OFF	3
-#endif
-
-out=r8;		// [expanded] output pointer
-inp=r9;		// [expanded] output pointer
-prsave=r10;
-key=r28;	// [expanded] pointer to RC4_KEY
-ksch=r29;	// (key->data+255)[&~(sizeof(key->data)-1)]
-xx=r30;
-yy=r31;
-
-// void RC4(RC4_KEY *key,size_t len,const void *inp,void *out);
-.global	RC4#
-.proc	RC4#
-.align	32
-.skip	16
-RC4:
-	.prologue
-	.save   ar.pfs,r2
-{ .mii;	alloc	r2=ar.pfs,4,12,0,16
-	.save	pr,prsave
-	mov	prsave=pr
-	ADDP	key=0,in0		};;
-{ .mib;	cmp.eq	p6,p0=0,in1			// len==0?
-	.save	ar.lc,r3
-	mov	r3=ar.lc
-(p6)	br.ret.spnt.many	b0	};;	// emergency exit
-
-	.body
-	.rotr	dat[4],key_x[4],tx[2],rnd[2],key_y[2],ty[1];
-
-{ .mib;	LDKEY	xx=[key],SZ			// load key->x
-	add	in1=-1,in1			// adjust len for loop counter
-	nop.b	0			}
-{ .mib;	ADDP	inp=0,in2
-	ADDP	out=0,in3
-	brp.loop.imp	.Ltop,.Lexit-16	};;
-{ .mmi;	LDKEY	yy=[key]			// load key->y
-	add	ksch=SZ,key
-	mov	ar.lc=in1		}
-{ .mmi;	mov	key_y[1]=r0			// guarantee inequality
-						// in first iteration
-	add	xx=1,xx
-	mov	pr.rot=1<<16		};;
-{ .mii;	nop.m	0
-	dep	key_x[1]=xx,r0,OFF,8
-	mov	ar.ec=3			};;	// note that epilogue counter
-						// is off by 1. I compensate
-						// for this at exit...
-.Ltop:
-// The loop is scheduled for 4*(n+2) spin-rate on Itanium 2, which
-// theoretically gives asymptotic performance of clock frequency
-// divided by 4 bytes per seconds, or 400MBps on 1.6GHz CPU. This is
-// for sizeof(RC4_INT)==4. For smaller RC4_INT STKEY inadvertently
-// splits the last bundle and you end up with 5*n spin-rate:-(
-// Originally the loop was scheduled for 3*n and relied on key
-// schedule to be aligned at 256*sizeof(RC4_INT) boundary. But
-// *(out++)=dat, which maps to st1, had same effect [inadvertent
-// bundle split] and holded the loop back. Rescheduling for 4*n
-// made it possible to eliminate dependence on specific alignment
-// and allow OpenSSH keep "abusing" our API. Reaching for 3*n would
-// require unrolling, sticking to variable shift instruction for
-// collecting output [to avoid starvation for integer shifter] and
-// copying of key schedule to controlled place in stack [so that
-// deposit instruction can serve as substitute for whole
-// key->data+((x&255)<<log2(sizeof(key->data[0])))]...
-{ .mmi;	(p19)	st1	[out]=dat[3],1			// *(out++)=dat
-	(p16)	add	xx=1,xx				// x++
-	(p18)	dep	rnd[1]=rnd[1],r0,OFF,8	}	// ((tx+ty)&255)<<OFF
-{ .mmi;	(p16)	add	key_x[1]=ksch,key_x[1]		// &key[xx&255]
-	(p17)	add	key_y[1]=ksch,key_y[1]	};;	// &key[yy&255]	
-{ .mmi;	(p16)	LDKEY	tx[0]=[key_x[1]]		// tx=key[xx]
-	(p17)	LDKEY	ty[0]=[key_y[1]]		// ty=key[yy]	
-	(p16)	dep	key_x[0]=xx,r0,OFF,8	}	// (xx&255)<<OFF
-{ .mmi;	(p18)	add	rnd[1]=ksch,rnd[1]		// &key[(tx+ty)&255]
-	(p16)	cmp.ne.unc p20,p21=key_x[1],key_y[1] };;
-{ .mmi;	(p18)	LDKEY	rnd[1]=[rnd[1]]			// rnd=key[(tx+ty)&255]
-	(p16)	ld1	dat[0]=[inp],1		}	// dat=*(inp++)
-.pred.rel	"mutex",p20,p21
-{ .mmi;	(p21)	add	yy=yy,tx[1]			// (p16)
-	(p20)	add	yy=yy,tx[0]			// (p16) y+=tx
-	(p21)	mov	tx[0]=tx[1]		};;	// (p16)
-{ .mmi;	(p17)	STKEY	[key_y[1]]=tx[1]		// key[yy]=tx
-	(p17)	STKEY	[key_x[2]]=ty[0]		// key[xx]=ty
-	(p16)	dep	key_y[0]=yy,r0,OFF,8	}	// &key[yy&255]
-{ .mmb;	(p17)	add	rnd[0]=tx[1],ty[0]		// tx+=ty
-	(p18)	xor	dat[2]=dat[2],rnd[1]		// dat^=rnd
-	br.ctop.sptk	.Ltop			};;
-.Lexit:
-{ .mib;	STKEY	[key]=yy,-SZ			// save key->y
-	mov	pr=prsave,0x1ffff
-	nop.b	0			}
-{ .mib;	st1	[out]=dat[3],1			// compensate for truncated
-						// epilogue counter
-	add	xx=-1,xx
-	nop.b	0			};;
-{ .mib;	STKEY	[key]=xx			// save key->x
-	mov	ar.lc=r3
-	br.ret.sptk.many	b0	};;
-.endp	RC4#
diff --git a/crypto/rc4/asm/rc4-ia64.pl b/crypto/rc4/asm/rc4-ia64.pl
new file mode 100755
index 000000000000..49cd5b5e6945
--- /dev/null
+++ b/crypto/rc4/asm/rc4-ia64.pl
@@ -0,0 +1,755 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by David Mosberger <David.Mosberger@acm.org> based on the
+# Itanium optimized Crypto code which was released by HP Labs at
+# http://www.hpl.hp.com/research/linux/crypto/.
+#
+# Copyright (c) 2005 Hewlett-Packard Development Company, L.P.
+#
+# Permission is hereby granted, free of charge, to any person obtaining
+# a copy of this software and associated documentation files (the
+# "Software"), to deal in the Software without restriction, including
+# without limitation the rights to use, copy, modify, merge, publish,
+# distribute, sublicense, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to
+# the following conditions:
+#
+# The above copyright notice and this permission notice shall be
+# included in all copies or substantial portions of the Software.
+
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
+# LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+# OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+# WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.  */
+
+
+
+# This is a little helper program which generates a software-pipelined
+# for RC4 encryption.  The basic algorithm looks like this:
+#
+#   for (counter = 0; counter < len; ++counter)
+#     {
+#       in = inp[counter];
+#       SI = S[I];
+#       J = (SI + J) & 0xff;
+#       SJ = S[J];
+#       T = (SI + SJ) & 0xff;
+#       S[I] = SJ, S[J] = SI;
+#       ST = S[T];
+#       outp[counter] = in ^ ST;
+#       I = (I + 1) & 0xff;
+#     }
+#
+# Pipelining this loop isn't easy, because the stores to the S[] array
+# need to be observed in the right order.  The loop generated by the
+# code below has the following pipeline diagram:
+#
+#      cycle
+#     | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |10 |11 |12 |13 |14 |15 |16 |17 |
+# iter
+#   1: xxx LDI xxx xxx xxx LDJ xxx SWP xxx LDT xxx xxx
+#   2:             xxx LDI xxx xxx xxx LDJ xxx SWP xxx LDT xxx xxx
+#   3:                         xxx LDI xxx xxx xxx LDJ xxx SWP xxx LDT xxx xxx
+#
+#   where:
+# 	LDI = load of S[I]
+# 	LDJ = load of S[J]
+# 	SWP = swap of S[I] and S[J]
+# 	LDT = load of S[T]
+#
+# Note that in the above diagram, the major trouble-spot is that LDI
+# of the 2nd iteration is performed BEFORE the SWP of the first
+# iteration.  Fortunately, this is easy to detect (I of the 1st
+# iteration will be equal to J of the 2nd iteration) and when this
+# happens, we simply forward the proper value from the 1st iteration
+# to the 2nd one.  The proper value in this case is simply the value
+# of S[I] from the first iteration (thanks to the fact that SWP
+# simply swaps the contents of S[I] and S[J]).
+#
+# Another potential trouble-spot is in cycle 7, where SWP of the 1st
+# iteration issues at the same time as the LDI of the 3rd iteration.
+# However, thanks to IA-64 execution semantics, this can be taken
+# care of simply by placing LDI later in the instruction-group than
+# SWP.  IA-64 CPUs will automatically forward the value if they
+# detect that the SWP and LDI are accessing the same memory-location.
+
+# The core-loop that can be pipelined then looks like this (annotated
+# with McKinley/Madison issue port & latency numbers, assuming L1
+# cache hits for the most part):
+
+# operation:	    instruction:		    issue-ports:  latency
+# ------------------  -----------------------------   ------------- -------
+
+# Data = *inp++       ld1 data = [inp], 1             M0-M1         1 cyc     c0
+#                     shladd Iptr = I, KeyTable, 3    M0-M3, I0, I1 1 cyc
+# I = (I + 1) & 0xff  padd1 nextI = I, one            M0-M3, I0, I1 3 cyc
+#                     ;;
+# SI = S[I]           ld8 SI = [Iptr]                 M0-M1         1 cyc     c1 * after SWAP!
+#                     ;;
+#                     cmp.eq.unc pBypass = I, J                                  * after J is valid!
+# J = SI + J          add J = J, SI                   M0-M3, I0, I1 1 cyc     c2
+#                     (pBypass) br.cond.spnt Bypass
+#                     ;;
+# ---------------------------------------------------------------------------------------
+# J = J & 0xff        zxt1 J = J                      I0, I1, 1 cyc           c3
+#                     ;;
+#                     shladd Jptr = J, KeyTable, 3    M0-M3, I0, I1 1 cyc     c4
+#                     ;;
+# SJ = S[J]           ld8 SJ = [Jptr]                 M0-M1         1 cyc     c5
+#                     ;;
+# ---------------------------------------------------------------------------------------
+# T = (SI + SJ)       add T = SI, SJ                  M0-M3, I0, I1 1 cyc     c6
+#                     ;;
+# T = T & 0xff        zxt1 T = T                      I0, I1        1 cyc
+# S[I] = SJ           st8 [Iptr] = SJ                 M2-M3                   c7
+# S[J] = SI           st8 [Jptr] = SI                 M2-M3
+#                     ;;
+#                     shladd Tptr = T, KeyTable, 3    M0-M3, I0, I1 1 cyc     c8
+#                     ;;
+# ---------------------------------------------------------------------------------------
+# T = S[T]            ld8 T = [Tptr]                  M0-M1         1 cyc     c9
+#                     ;;
+# data ^= T           xor data = data, T              M0-M3, I0, I1 1 cyc     c10
+#                     ;;
+# *out++ = Data ^ T   dep word = word, data, 8, POS   I0, I1        1 cyc     c11
+#                     ;;
+# ---------------------------------------------------------------------------------------
+
+# There are several points worth making here:
+
+#   - Note that due to the bypass/forwarding-path, the first two
+#     phases of the loop are strangly mingled together.  In
+#     particular, note that the first stage of the pipeline is
+#     using the value of "J", as calculated by the second stage.
+#   - Each bundle-pair will have exactly 6 instructions.
+#   - Pipelined, the loop can execute in 3 cycles/iteration and
+#     4 stages.  However, McKinley/Madison can issue "st1" to
+#     the same bank at a rate of at most one per 4 cycles.  Thus,
+#     instead of storing each byte, we accumulate them in a word
+#     and then write them back at once with a single "st8" (this
+#     implies that the setup code needs to ensure that the output
+#     buffer is properly aligned, if need be, by encoding the
+#     first few bytes separately).
+#   - There is no space for a "br.ctop" instruction.  For this
+#     reason we can't use module-loop support in IA-64 and have
+#     to do a traditional, purely software-pipelined loop.
+#   - We can't replace any of the remaining "add/zxt1" pairs with
+#     "padd1" because the latency for that instruction is too high
+#     and would push the loop to the point where more bypasses
+#     would be needed, which we don't have space for.
+#   - The above loop runs at around 3.26 cycles/byte, or roughly
+#     440 MByte/sec on a 1.5GHz Madison.  This is well below the
+#     system bus bandwidth and hence with judicious use of
+#     "lfetch" this loop can run at (almost) peak speed even when
+#     the input and output data reside in memory.  The
+#     max. latency that can be tolerated is (PREFETCH_DISTANCE *
+#     L2_LINE_SIZE * 3 cyc), or about 384 cycles assuming (at
+#     least) 1-ahead prefetching of 128 byte cache-lines.  Note
+#     that we do NOT prefetch into L1, since that would only
+#     interfere with the S[] table values stored there.  This is
+#     acceptable because there is a 10 cycle latency between
+#     load and first use of the input data.
+#   - We use a branch to out-of-line bypass-code of cycle-pressure:
+#     we calculate the next J, check for the need to activate the
+#     bypass path, and activate the bypass path ALL IN THE SAME
+#     CYCLE.  If we didn't have these constraints, we could do
+#     the bypass with a simple conditional move instruction.
+#     Fortunately, the bypass paths get activated relatively
+#     infrequently, so the extra branches don't cost all that much
+#     (about 0.04 cycles/byte, measured on a 16396 byte file with
+#     random input data).
+#
+
+$phases = 4;		# number of stages/phases in the pipelined-loop
+$unroll_count = 6;	# number of times we unrolled it
+$pComI = (1 << 0);
+$pComJ = (1 << 1);
+$pComT = (1 << 2);
+$pOut  = (1 << 3);
+
+$NData = 4;
+$NIP = 3;
+$NJP = 2;
+$NI = 2;
+$NSI = 3;
+$NSJ = 2;
+$NT = 2;
+$NOutWord = 2;
+
+#
+# $threshold is the minimum length before we attempt to use the
+# big software-pipelined loop.  It MUST be greater-or-equal
+# to:
+#  		PHASES * (UNROLL_COUNT + 1) + 7
+#
+# The "+ 7" comes from the fact we may have to encode up to
+#   7 bytes separately before the output pointer is aligned.
+#
+$threshold = (3 * ($phases * ($unroll_count + 1)) + 7);
+
+sub I {
+    local *code = shift;
+    local $format = shift;
+    $code .= sprintf ("\t\t".$format."\n", @_);
+}
+
+sub P {
+    local *code = shift;
+    local $format = shift;
+    $code .= sprintf ($format."\n", @_);
+}
+
+sub STOP {
+    local *code = shift;
+    $code .=<<___;
+		;;
+___
+}
+
+sub emit_body {
+    local *c = shift;
+    local *bypass = shift;
+    local ($iteration, $p) = @_;
+
+    local $i0 = $iteration;
+    local $i1 = $iteration - 1;
+    local $i2 = $iteration - 2;
+    local $i3 = $iteration - 3;
+    local $iw0 = ($iteration - 3) / 8;
+    local $iw1 = ($iteration > 3) ? ($iteration - 4) / 8 : 1;
+    local $byte_num = ($iteration - 3) % 8;
+    local $label = $iteration + 1;
+    local $pAny = ($p & 0xf) == 0xf;
+    local $pByp = (($p & $pComI) && ($iteration > 0));
+
+    $c.=<<___;
+//////////////////////////////////////////////////
+___
+
+    if (($p & 0xf) == 0) {
+	$c.="#ifdef HOST_IS_BIG_ENDIAN\n";
+	&I(\$c,"shr.u	OutWord[%u] = OutWord[%u], 32;;",
+				$iw1 % $NOutWord, $iw1 % $NOutWord);
+	$c.="#endif\n";
+	&I(\$c, "st4 [OutPtr] = OutWord[%u], 4", $iw1 % $NOutWord);
+	return;
+    }
+
+    # Cycle 0
+    &I(\$c, "{ .mmi")					      if ($pAny);
+    &I(\$c, "ld1    Data[%u] = [InPtr], 1", $i0 % $NData)     if ($p & $pComI);
+    &I(\$c, "padd1  I[%u] = One, I[%u]", $i0 % $NI, $i1 % $NI)if ($p & $pComI);
+    &I(\$c, "zxt1   J = J")				      if ($p & $pComJ);
+    &I(\$c, "}")					      if ($pAny);
+    &I(\$c, "{ .mmi")					      if ($pAny);
+    &I(\$c, "LKEY   T[%u] = [T[%u]]", $i1 % $NT, $i1 % $NT)   if ($p & $pOut);
+    &I(\$c, "add    T[%u] = SI[%u], SJ[%u]",
+       $i0 % $NT, $i2 % $NSI, $i1 % $NSJ)		      if ($p & $pComT);
+    &I(\$c, "KEYADDR(IPr[%u], I[%u])", $i0 % $NIP, $i1 % $NI) if ($p & $pComI);
+    &I(\$c, "}")					      if ($pAny);
+    &STOP(\$c);
+
+    # Cycle 1
+    &I(\$c, "{ .mmi")					      if ($pAny);
+    &I(\$c, "SKEY   [IPr[%u]] = SJ[%u]", $i2 % $NIP, $i1%$NSJ)if ($p & $pComT);
+    &I(\$c, "SKEY   [JP[%u]] = SI[%u]", $i1 % $NJP, $i2%$NSI) if ($p & $pComT);
+    &I(\$c, "zxt1   T[%u] = T[%u]", $i0 % $NT, $i0 % $NT)     if ($p & $pComT);
+    &I(\$c, "}")					      if ($pAny);
+    &I(\$c, "{ .mmi")					      if ($pAny);
+    &I(\$c, "LKEY   SI[%u] = [IPr[%u]]", $i0 % $NSI, $i0%$NIP)if ($p & $pComI);
+    &I(\$c, "KEYADDR(JP[%u], J)", $i0 % $NJP)		      if ($p & $pComJ);
+    &I(\$c, "xor    Data[%u] = Data[%u], T[%u]",
+       $i3 % $NData, $i3 % $NData, $i1 % $NT)		      if ($p & $pOut);
+    &I(\$c, "}")					      if ($pAny);
+    &STOP(\$c);
+
+    # Cycle 2
+    &I(\$c, "{ .mmi")					      if ($pAny);
+    &I(\$c, "LKEY   SJ[%u] = [JP[%u]]", $i0 % $NSJ, $i0%$NJP) if ($p & $pComJ);
+    &I(\$c, "cmp.eq pBypass, p0 = I[%u], J", $i1 % $NI)	      if ($pByp);
+    &I(\$c, "dep OutWord[%u] = Data[%u], OutWord[%u], BYTE_POS(%u), 8",
+       $iw0%$NOutWord, $i3%$NData, $iw1%$NOutWord, $byte_num) if ($p & $pOut);
+    &I(\$c, "}")					      if ($pAny);
+    &I(\$c, "{ .mmb")					      if ($pAny);
+    &I(\$c, "add    J = J, SI[%u]", $i0 % $NSI)		      if ($p & $pComI);
+    &I(\$c, "KEYADDR(T[%u], T[%u])", $i0 % $NT, $i0 % $NT)    if ($p & $pComT);
+    &P(\$c, "(pBypass)\tbr.cond.spnt.many .rc4Bypass%u",$label)if ($pByp);
+    &I(\$c, "}") if ($pAny);
+    &STOP(\$c);
+
+    &P(\$c, ".rc4Resume%u:", $label)			      if ($pByp);
+    if ($byte_num == 0 && $iteration >= $phases) {
+	&I(\$c, "st8 [OutPtr] = OutWord[%u], 8",
+	   $iw1 % $NOutWord)				      if ($p & $pOut);
+	if ($iteration == (1 + $unroll_count) * $phases - 1) {
+	    if ($unroll_count == 6) {
+		&I(\$c, "mov OutWord[%u] = OutWord[%u]",
+		   $iw1 % $NOutWord, $iw0 % $NOutWord);
+	    }
+	    &I(\$c, "lfetch.nt1 [InPrefetch], %u",
+	       $unroll_count * $phases);
+	    &I(\$c, "lfetch.excl.nt1 [OutPrefetch], %u",
+	       $unroll_count * $phases);
+	    &I(\$c, "br.cloop.sptk.few .rc4Loop");
+	}
+    }
+
+    if ($pByp) {
+	&P(\$bypass, ".rc4Bypass%u:", $label);
+	&I(\$bypass, "sub J = J, SI[%u]", $i0 % $NSI);
+	&I(\$bypass, "nop 0");
+	&I(\$bypass, "nop 0");
+	&I(\$bypass, ";;");
+	&I(\$bypass, "add J = J, SI[%u]", $i1 % $NSI);
+	&I(\$bypass, "mov SI[%u] = SI[%u]", $i0 % $NSI, $i1 % $NSI);
+	&I(\$bypass, "br.sptk.many .rc4Resume%u\n", $label);
+	&I(\$bypass, ";;");
+    }
+}
+
+$code=<<___;
+.ident \"rc4-ia64.s, version 3.0\"
+.ident \"Copyright (c) 2005 Hewlett-Packard Development Company, L.P.\"
+
+#define LCSave		r8
+#define PRSave		r9
+
+/* Inputs become invalid once rotation begins!  */
+
+#define StateTable	in0
+#define DataLen		in1
+#define InputBuffer	in2
+#define OutputBuffer	in3
+
+#define KTable		r14
+#define J		r15
+#define InPtr		r16
+#define OutPtr		r17
+#define InPrefetch	r18
+#define OutPrefetch	r19
+#define One		r20
+#define LoopCount	r21
+#define Remainder	r22
+#define IFinal		r23
+#define EndPtr		r24
+
+#define tmp0		r25
+#define tmp1		r26
+
+#define pBypass		p6
+#define pDone		p7
+#define pSmall		p8
+#define pAligned	p9
+#define pUnaligned	p10
+
+#define pComputeI	pPhase[0]
+#define pComputeJ	pPhase[1]
+#define pComputeT	pPhase[2]
+#define pOutput		pPhase[3]
+
+#define RetVal		r8
+#define L_OK		p7
+#define L_NOK		p8
+
+#define	_NINPUTS	4
+#define	_NOUTPUT	0
+
+#define	_NROTATE	24
+#define	_NLOCALS	(_NROTATE - _NINPUTS - _NOUTPUT)
+
+#ifndef SZ
+# define SZ	4	// this must be set to sizeof(RC4_INT)
+#endif
+
+#if SZ == 1
+# define LKEY			ld1
+# define SKEY			st1
+# define KEYADDR(dst, i)	add dst = i, KTable
+#elif SZ == 2
+# define LKEY			ld2
+# define SKEY			st2
+# define KEYADDR(dst, i)	shladd dst = i, 1, KTable
+#elif SZ == 4
+# define LKEY			ld4
+# define SKEY			st4
+# define KEYADDR(dst, i)	shladd dst = i, 2, KTable
+#else
+# define LKEY			ld8
+# define SKEY			st8
+# define KEYADDR(dst, i)	shladd dst = i, 3, KTable
+#endif
+
+#if defined(_HPUX_SOURCE) && !defined(_LP64)
+# define ADDP	addp4
+#else
+# define ADDP	add
+#endif
+
+/* Define a macro for the bit number of the n-th byte: */
+
+#if defined(_HPUX_SOURCE) || defined(B_ENDIAN)
+# define HOST_IS_BIG_ENDIAN
+# define BYTE_POS(n)	(56 - (8 * (n)))
+#else
+# define BYTE_POS(n)	(8 * (n))
+#endif
+
+/*
+   We must perform the first phase of the pipeline explicitly since
+   we will always load from the stable the first time. The br.cexit
+   will never be taken since regardless of the number of bytes because
+   the epilogue count is 4.
+*/
+/* MODSCHED_RC4 macro was split to _PROLOGUE and _LOOP, because HP-UX
+   assembler failed on original macro with syntax error. <appro> */
+#define MODSCHED_RC4_PROLOGUE						   \\
+	{								   \\
+				ld1		Data[0] = [InPtr], 1;	   \\
+				add		IFinal = 1, I[1];	   \\
+				KEYADDR(IPr[0], I[1]);			   \\
+	} ;;								   \\
+	{								   \\
+				LKEY		SI[0] = [IPr[0]];	   \\
+				mov		pr.rot = 0x10000;	   \\
+				mov		ar.ec = 4;		   \\
+	} ;;								   \\
+	{								   \\
+				add		J = J, SI[0];		   \\
+				zxt1		I[0] = IFinal;		   \\
+				br.cexit.spnt.few .+16; /* never taken */  \\
+	} ;;
+#define MODSCHED_RC4_LOOP(label)					   \\
+label:									   \\
+	{	.mmi;							   \\
+		(pComputeI)	ld1		Data[0] = [InPtr], 1;	   \\
+		(pComputeI)	add		IFinal = 1, I[1];	   \\
+		(pComputeJ)	zxt1		J = J;			   \\
+	}{	.mmi;							   \\
+		(pOutput)	LKEY		T[1] = [T[1]];		   \\
+		(pComputeT)	add		T[0] = SI[2], SJ[1];	   \\
+		(pComputeI)	KEYADDR(IPr[0], I[1]);			   \\
+	} ;;								   \\
+	{	.mmi;							   \\
+		(pComputeT)	SKEY		[IPr[2]] = SJ[1];	   \\
+		(pComputeT)	SKEY		[JP[1]] = SI[2];	   \\
+		(pComputeT)	zxt1		T[0] = T[0];		   \\
+	}{	.mmi;							   \\
+		(pComputeI)	LKEY		SI[0] = [IPr[0]];	   \\
+		(pComputeJ)	KEYADDR(JP[0], J);			   \\
+		(pComputeI)	cmp.eq.unc	pBypass, p0 = I[1], J;	   \\
+	} ;;								   \\
+	{	.mmi;							   \\
+		(pComputeJ)	LKEY		SJ[0] = [JP[0]];	   \\
+		(pOutput)	xor		Data[3] = Data[3], T[1];   \\
+				nop		0x0;			   \\
+	}{	.mmi;							   \\
+		(pComputeT)	KEYADDR(T[0], T[0]);			   \\
+		(pBypass)	mov		SI[0] = SI[1];		   \\
+		(pComputeI)	zxt1		I[0] = IFinal;		   \\
+	} ;;								   \\
+	{	.mmb;							   \\
+		(pOutput)	st1		[OutPtr] = Data[3], 1;	   \\
+		(pComputeI)	add		J = J, SI[0];		   \\
+				br.ctop.sptk.few label;			   \\
+	} ;;
+
+	.text
+
+	.align	32
+
+	.type	RC4, \@function
+	.global	RC4
+
+	.proc	RC4
+	.prologue
+
+RC4:
+	{
+	  	.mmi
+		alloc	r2 = ar.pfs, _NINPUTS, _NLOCALS, _NOUTPUT, _NROTATE
+
+		.rotr Data[4], I[2], IPr[3], SI[3], JP[2], SJ[2], T[2], \\
+		      OutWord[2]
+		.rotp pPhase[4]
+
+		ADDP		InPrefetch = 0, InputBuffer
+		ADDP		KTable = 0, StateTable
+	}
+	{
+		.mmi
+		ADDP		InPtr = 0, InputBuffer
+		ADDP		OutPtr = 0, OutputBuffer
+		mov		RetVal = r0
+	}
+	;;
+	{
+		.mmi
+		lfetch.nt1	[InPrefetch], 0x80
+		ADDP		OutPrefetch = 0, OutputBuffer
+	}
+	{               // Return 0 if the input length is nonsensical
+        	.mib
+		ADDP		StateTable = 0, StateTable
+        	cmp.ge.unc  	L_NOK, L_OK = r0, DataLen
+	(L_NOK) br.ret.sptk.few rp
+	}
+	;;
+	{
+        	.mib
+        	cmp.eq.or  	L_NOK, L_OK = r0, InPtr
+        	cmp.eq.or  	L_NOK, L_OK = r0, OutPtr
+		nop		0x0
+	}
+	{
+		.mib
+        	cmp.eq.or  	L_NOK, L_OK = r0, StateTable
+		nop		0x0
+	(L_NOK) br.ret.sptk.few rp
+	}
+	;;
+		LKEY		I[1] = [KTable], SZ
+/* Prefetch the state-table. It contains 256 elements of size SZ */
+
+#if SZ == 1
+		ADDP		tmp0 = 1*128, StateTable
+#elif SZ == 2
+		ADDP		tmp0 = 3*128, StateTable
+		ADDP		tmp1 = 2*128, StateTable
+#elif SZ == 4
+		ADDP		tmp0 = 7*128, StateTable
+		ADDP		tmp1 = 6*128, StateTable
+#elif SZ == 8
+		ADDP		tmp0 = 15*128, StateTable
+		ADDP		tmp1 = 14*128, StateTable
+#endif
+		;;
+#if SZ >= 8
+		lfetch.fault.nt1		[tmp0], -256	// 15
+		lfetch.fault.nt1		[tmp1], -256;;
+		lfetch.fault.nt1		[tmp0], -256	// 13
+		lfetch.fault.nt1		[tmp1], -256;;
+		lfetch.fault.nt1		[tmp0], -256	// 11
+		lfetch.fault.nt1		[tmp1], -256;;
+		lfetch.fault.nt1		[tmp0], -256	//  9
+		lfetch.fault.nt1		[tmp1], -256;;
+#endif
+#if SZ >= 4
+		lfetch.fault.nt1		[tmp0], -256	//  7
+		lfetch.fault.nt1		[tmp1], -256;;
+		lfetch.fault.nt1		[tmp0], -256	//  5
+		lfetch.fault.nt1		[tmp1], -256;;
+#endif
+#if SZ >= 2
+		lfetch.fault.nt1		[tmp0], -256	//  3
+		lfetch.fault.nt1		[tmp1], -256;;
+#endif
+	{
+		.mii
+		lfetch.fault.nt1		[tmp0]		//  1
+		add		I[1]=1,I[1];;
+		zxt1		I[1]=I[1]
+	}
+	{
+		.mmi
+		lfetch.nt1	[InPrefetch], 0x80
+		lfetch.excl.nt1	[OutPrefetch], 0x80
+		.save		pr, PRSave
+		mov		PRSave = pr
+	} ;;
+	{
+		.mmi
+		lfetch.excl.nt1	[OutPrefetch], 0x80
+		LKEY		J = [KTable], SZ
+		ADDP		EndPtr = DataLen, InPtr
+	}  ;;
+	{
+		.mmi
+		ADDP		EndPtr = -1, EndPtr	// Make it point to
+							// last data byte.
+		mov		One = 1
+		.save		ar.lc, LCSave
+		mov		LCSave = ar.lc
+		.body
+	} ;;
+	{
+		.mmb
+		sub		Remainder = 0, OutPtr
+		cmp.gtu		pSmall, p0 = $threshold, DataLen
+(pSmall)	br.cond.dpnt	.rc4Remainder		// Data too small for
+							// big loop.
+	} ;;
+	{
+		.mmi
+		and		Remainder = 0x7, Remainder
+		;;
+		cmp.eq		pAligned, pUnaligned = Remainder, r0
+		nop		0x0
+	} ;;
+	{
+		.mmb
+.pred.rel	"mutex",pUnaligned,pAligned
+(pUnaligned)	add		Remainder = -1, Remainder
+(pAligned)	sub		Remainder = EndPtr, InPtr
+(pAligned)	br.cond.dptk.many .rc4Aligned
+	} ;;
+	{
+		.mmi
+		nop		0x0
+		nop		0x0
+		mov.i		ar.lc = Remainder
+	}
+
+/* Do the initial few bytes via the compact, modulo-scheduled loop
+   until the output pointer is 8-byte-aligned.  */
+
+		MODSCHED_RC4_PROLOGUE
+		MODSCHED_RC4_LOOP(.RC4AlignLoop)
+
+	{
+		.mib
+		sub		Remainder = EndPtr, InPtr
+		zxt1		IFinal = IFinal
+		clrrrb				// Clear CFM.rrb.pr so
+		;;				// next "mov pr.rot = N"
+						// does the right thing.
+	}
+	{
+		.mmi
+		mov		I[1] = IFinal
+		nop		0x0
+		nop		0x0
+	} ;;
+
+
+.rc4Aligned:
+
+/*
+   Unrolled loop count = (Remainder - ($unroll_count+1)*$phases)/($unroll_count*$phases)
+ */
+
+	{
+		.mlx
+		add	LoopCount = 1 - ($unroll_count + 1)*$phases, Remainder
+		movl		Remainder = 0xaaaaaaaaaaaaaaab
+	} ;;
+	{
+		.mmi
+		setf.sig	f6 = LoopCount		// M2, M3	6 cyc
+		setf.sig	f7 = Remainder		// M2, M3	6 cyc
+		nop		0x0
+	} ;;
+	{
+		.mfb
+		nop		0x0
+		xmpy.hu		f6 = f6, f7
+		nop		0x0
+	} ;;
+	{
+		.mmi
+		getf.sig	LoopCount = f6;;	// M2		5 cyc
+		nop		0x0
+		shr.u		LoopCount = LoopCount, 4
+	} ;;
+	{
+		.mmi
+		nop		0x0
+		nop		0x0
+		mov.i		ar.lc = LoopCount
+	} ;;
+
+/* Now comes the unrolled loop: */
+
+.rc4Prologue:
+___
+
+$iteration = 0;
+
+# Generate the prologue:
+$predicates = 1;
+for ($i = 0; $i < $phases; ++$i) {
+    &emit_body (\$code, \$bypass, $iteration++, $predicates);
+    $predicates = ($predicates << 1) | 1;
+}
+
+$code.=<<___;
+.rc4Loop:
+___
+
+# Generate the body:
+for ($i = 0; $i < $unroll_count*$phases; ++$i) {
+    &emit_body (\$code, \$bypass, $iteration++, $predicates);
+}
+
+$code.=<<___;
+.rc4Epilogue:
+___
+
+# Generate the epilogue:
+for ($i = 0; $i < $phases; ++$i) {
+    $predicates <<= 1;
+    &emit_body (\$code, \$bypass, $iteration++, $predicates);
+}
+
+$code.=<<___;
+	{
+		.mmi
+		lfetch.nt1	[EndPtr]	// fetch line with last byte
+		mov		IFinal = I[1]
+		nop		0x0
+	}
+
+.rc4Remainder:
+	{
+		.mmi
+		sub		Remainder = EndPtr, InPtr	// Calculate
+								// # of bytes
+								// left - 1
+		nop		0x0
+		nop		0x0
+	} ;;
+	{
+		.mib
+		cmp.eq		pDone, p0 = -1, Remainder // done already?
+		mov.i		ar.lc = Remainder
+(pDone)		br.cond.dptk.few .rc4Complete
+	}
+
+/* Do the remaining bytes via the compact, modulo-scheduled loop */
+
+		MODSCHED_RC4_PROLOGUE
+		MODSCHED_RC4_LOOP(.RC4RestLoop)
+
+.rc4Complete:
+	{
+		.mmi
+		add		KTable = -SZ, KTable
+		add		IFinal = -1, IFinal
+		mov		ar.lc = LCSave
+	} ;;
+	{
+		.mii
+		SKEY		[KTable] = J,-SZ
+		zxt1		IFinal = IFinal
+		mov		pr = PRSave, 0x1FFFF
+	} ;;
+	{
+		.mib
+		SKEY		[KTable] = IFinal
+		add		RetVal = 1, r0
+		br.ret.sptk.few	rp
+	} ;;
+___
+
+# Last but not least, emit the code for the bypass-code of the unrolled loop:
+
+$code.=$bypass;
+
+$code.=<<___;
+	.endp RC4
+___
+
+print $code;
diff --git a/crypto/rc4/asm/rc4-md5-x86_64.pl b/crypto/rc4/asm/rc4-md5-x86_64.pl
new file mode 100755
index 000000000000..7f684092d40a
--- /dev/null
+++ b/crypto/rc4/asm/rc4-md5-x86_64.pl
@@ -0,0 +1,631 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# June 2011
+#
+# This is RC4+MD5 "stitch" implementation. The idea, as spelled in
+# http://download.intel.com/design/intarch/papers/323686.pdf, is that
+# since both algorithms exhibit instruction-level parallelism, ILP,
+# below theoretical maximum, interleaving them would allow to utilize
+# processor resources better and achieve better performance. RC4
+# instruction sequence is virtually identical to rc4-x86_64.pl, which
+# is heavily based on submission by Maxim Perminov, Maxim Locktyukhin
+# and Jim Guilford of Intel. MD5 is fresh implementation aiming to
+# minimize register usage, which was used as "main thread" with RC4
+# weaved into it, one RC4 round per one MD5 round. In addition to the
+# stiched subroutine the script can generate standalone replacement
+# md5_block_asm_data_order and RC4. Below are performance numbers in
+# cycles per processed byte, less is better, for these the standalone
+# subroutines, sum of them, and stitched one:
+#
+#		RC4	MD5	RC4+MD5	stitch	gain
+# Opteron	6.5(*)	5.4	11.9	7.0	+70%(*)
+# Core2		6.5	5.8	12.3	7.7	+60%
+# Westmere	4.3	5.2	9.5	7.0	+36%
+# Sandy Bridge	4.2	5.5	9.7	6.8	+43%
+# Atom		9.3	6.5	15.8	11.1	+42%
+#
+# (*)	rc4-x86_64.pl delivers 5.3 on Opteron, so real improvement
+#	is +53%...
+
+my ($rc4,$md5)=(1,1);	# what to generate?
+my $D="#" if (!$md5);	# if set to "#", MD5 is stitched into RC4(),
+			# but its result is discarded. Idea here is
+			# to be able to use 'openssl speed rc4' for
+			# benchmarking the stitched subroutine... 
+
+my $flavour = shift;
+my $output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+my $win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; my $dir=$1; my $xlate;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+my ($dat,$in0,$out,$ctx,$inp,$len, $func,$nargs);
+
+if ($rc4 && !$md5) {
+  ($dat,$len,$in0,$out) = ("%rdi","%rsi","%rdx","%rcx");
+  $func="RC4";				$nargs=4;
+} elsif ($md5 && !$rc4) {
+  ($ctx,$inp,$len) = ("%rdi","%rsi","%rdx");
+  $func="md5_block_asm_data_order";	$nargs=3;
+} else {
+  ($dat,$in0,$out,$ctx,$inp,$len) = ("%rdi","%rsi","%rdx","%rcx","%r8","%r9");
+  $func="rc4_md5_enc";			$nargs=6;
+  # void rc4_md5_enc(
+  #		RC4_KEY *key,		#
+  #		const void *in0,	# RC4 input
+  #		void *out,		# RC4 output
+  #		MD5_CTX *ctx,		#
+  #		const void *inp,	# MD5 input
+  #		size_t len);		# number of 64-byte blocks
+}
+
+my @K=(	0xd76aa478,0xe8c7b756,0x242070db,0xc1bdceee,
+	0xf57c0faf,0x4787c62a,0xa8304613,0xfd469501,
+	0x698098d8,0x8b44f7af,0xffff5bb1,0x895cd7be,
+	0x6b901122,0xfd987193,0xa679438e,0x49b40821,
+
+	0xf61e2562,0xc040b340,0x265e5a51,0xe9b6c7aa,
+	0xd62f105d,0x02441453,0xd8a1e681,0xe7d3fbc8,
+	0x21e1cde6,0xc33707d6,0xf4d50d87,0x455a14ed,
+	0xa9e3e905,0xfcefa3f8,0x676f02d9,0x8d2a4c8a,
+
+	0xfffa3942,0x8771f681,0x6d9d6122,0xfde5380c,
+	0xa4beea44,0x4bdecfa9,0xf6bb4b60,0xbebfbc70,
+	0x289b7ec6,0xeaa127fa,0xd4ef3085,0x04881d05,
+	0xd9d4d039,0xe6db99e5,0x1fa27cf8,0xc4ac5665,
+
+	0xf4292244,0x432aff97,0xab9423a7,0xfc93a039,
+	0x655b59c3,0x8f0ccc92,0xffeff47d,0x85845dd1,
+	0x6fa87e4f,0xfe2ce6e0,0xa3014314,0x4e0811a1,
+	0xf7537e82,0xbd3af235,0x2ad7d2bb,0xeb86d391	);
+
+my @V=("%r8d","%r9d","%r10d","%r11d");	# MD5 registers
+my $tmp="%r12d";
+
+my @XX=("%rbp","%rsi");			# RC4 registers
+my @TX=("%rax","%rbx");
+my $YY="%rcx";
+my $TY="%rdx";
+
+my $MOD=32;				# 16, 32 or 64
+
+$code.=<<___;
+.text
+.align 16
+
+.globl	$func
+.type	$func,\@function,$nargs
+$func:
+	cmp	\$0,$len
+	je	.Labort
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	sub	\$40,%rsp
+.Lbody:
+___
+if ($rc4) {
+$code.=<<___;
+$D#md5#	mov	$ctx,%r11		# reassign arguments
+	mov	$len,%r12
+	mov	$in0,%r13
+	mov	$out,%r14
+$D#md5#	mov	$inp,%r15
+___
+    $ctx="%r11"	if ($md5);		# reassign arguments
+    $len="%r12";
+    $in0="%r13";
+    $out="%r14";
+    $inp="%r15"	if ($md5);
+    $inp=$in0	if (!$md5);
+$code.=<<___;
+	xor	$XX[0],$XX[0]
+	xor	$YY,$YY
+
+	lea	8($dat),$dat
+	mov	-8($dat),$XX[0]#b
+	mov	-4($dat),$YY#b
+
+	inc	$XX[0]#b
+	sub	$in0,$out
+	movl	($dat,$XX[0],4),$TX[0]#d
+___
+$code.=<<___ if (!$md5);
+	xor	$TX[1],$TX[1]
+	test	\$-128,$len
+	jz	.Loop1
+	sub	$XX[0],$TX[1]
+	and	\$`$MOD-1`,$TX[1]
+	jz	.Loop${MOD}_is_hot
+	sub	$TX[1],$len
+.Loop${MOD}_warmup:
+	add	$TX[0]#b,$YY#b
+	movl	($dat,$YY,4),$TY#d
+	movl	$TX[0]#d,($dat,$YY,4)
+	movl	$TY#d,($dat,$XX[0],4)
+	add	$TY#b,$TX[0]#b
+	inc	$XX[0]#b
+	movl	($dat,$TX[0],4),$TY#d
+	movl	($dat,$XX[0],4),$TX[0]#d
+	xorb	($in0),$TY#b
+	movb	$TY#b,($out,$in0)
+	lea	1($in0),$in0
+	dec	$TX[1]
+	jnz	.Loop${MOD}_warmup
+
+	mov	$YY,$TX[1]
+	xor	$YY,$YY
+	mov	$TX[1]#b,$YY#b
+
+.Loop${MOD}_is_hot:
+	mov	$len,32(%rsp)		# save original $len
+	shr	\$6,$len		# number of 64-byte blocks
+___
+  if ($D && !$md5) {			# stitch in dummy MD5
+    $md5=1;
+    $ctx="%r11";
+    $inp="%r15";
+    $code.=<<___;
+	mov	%rsp,$ctx
+	mov	$in0,$inp
+___
+  }
+}
+$code.=<<___;
+#rc4#	add	$TX[0]#b,$YY#b
+#rc4#	lea	($dat,$XX[0],4),$XX[1]
+	shl	\$6,$len
+	add	$inp,$len		# pointer to the end of input
+	mov	$len,16(%rsp)
+
+#md5#	mov	$ctx,24(%rsp)		# save pointer to MD5_CTX
+#md5#	mov	0*4($ctx),$V[0]		# load current hash value from MD5_CTX
+#md5#	mov	1*4($ctx),$V[1]
+#md5#	mov	2*4($ctx),$V[2]
+#md5#	mov	3*4($ctx),$V[3]
+	jmp	.Loop
+
+.align	16
+.Loop:
+#md5#	mov	$V[0],0*4(%rsp)		# put aside current hash value
+#md5#	mov	$V[1],1*4(%rsp)
+#md5#	mov	$V[2],2*4(%rsp)
+#md5#	mov	$V[3],$tmp		# forward reference
+#md5#	mov	$V[3],3*4(%rsp)
+___
+
+sub R0 {
+  my ($i,$a,$b,$c,$d)=@_;
+  my @rot0=(7,12,17,22);
+  my $j=$i%16;
+  my $k=$i%$MOD;
+  my $xmm="%xmm".($j&1);
+    $code.="	movdqu	($in0),%xmm2\n"		if ($rc4 && $j==15);
+    $code.="	add	\$$MOD,$XX[0]#b\n"	if ($rc4 && $j==15 && $k==$MOD-1);
+    $code.="	pxor	$xmm,$xmm\n"		if ($rc4 && $j<=1);
+    $code.=<<___;
+#rc4#	movl	($dat,$YY,4),$TY#d
+#md5#	xor	$c,$tmp
+#rc4#	movl	$TX[0]#d,($dat,$YY,4)
+#md5#	and	$b,$tmp
+#md5#	add	4*`$j`($inp),$a
+#rc4#	add	$TY#b,$TX[0]#b
+#rc4#	movl	`4*(($k+1)%$MOD)`(`$k==$MOD-1?"$dat,$XX[0],4":"$XX[1]"`),$TX[1]#d
+#md5#	add	\$$K[$i],$a
+#md5#	xor	$d,$tmp
+#rc4#	movz	$TX[0]#b,$TX[0]#d
+#rc4#	movl	$TY#d,4*$k($XX[1])
+#md5#	add	$tmp,$a
+#rc4#	add	$TX[1]#b,$YY#b
+#md5#	rol	\$$rot0[$j%4],$a
+#md5#	mov	`$j==15?"$b":"$c"`,$tmp		# forward reference
+#rc4#	pinsrw	\$`($j>>1)&7`,($dat,$TX[0],4),$xmm\n
+#md5#	add	$b,$a
+___
+    $code.=<<___ if ($rc4 && $j==15 && $k==$MOD-1);
+	mov	$YY,$XX[1]
+	xor	$YY,$YY				# keyword to partial register
+	mov	$XX[1]#b,$YY#b
+	lea	($dat,$XX[0],4),$XX[1]
+___
+    $code.=<<___ if ($rc4 && $j==15);
+	psllq	\$8,%xmm1
+	pxor	%xmm0,%xmm2
+	pxor	%xmm1,%xmm2
+___
+}
+sub R1 {
+  my ($i,$a,$b,$c,$d)=@_;
+  my @rot1=(5,9,14,20);
+  my $j=$i%16;
+  my $k=$i%$MOD;
+  my $xmm="%xmm".($j&1);
+    $code.="	movdqu	16($in0),%xmm3\n"	if ($rc4 && $j==15);
+    $code.="	add	\$$MOD,$XX[0]#b\n"	if ($rc4 && $j==15 && $k==$MOD-1);
+    $code.="	pxor	$xmm,$xmm\n"		if ($rc4 && $j<=1);
+    $code.=<<___;
+#rc4#	movl	($dat,$YY,4),$TY#d
+#md5#	xor	$b,$tmp
+#rc4#	movl	$TX[0]#d,($dat,$YY,4)
+#md5#	and	$d,$tmp
+#md5#	add	4*`((1+5*$j)%16)`($inp),$a
+#rc4#	add	$TY#b,$TX[0]#b
+#rc4#	movl	`4*(($k+1)%$MOD)`(`$k==$MOD-1?"$dat,$XX[0],4":"$XX[1]"`),$TX[1]#d
+#md5#	add	\$$K[$i],$a
+#md5#	xor	$c,$tmp
+#rc4#	movz	$TX[0]#b,$TX[0]#d
+#rc4#	movl	$TY#d,4*$k($XX[1])
+#md5#	add	$tmp,$a
+#rc4#	add	$TX[1]#b,$YY#b
+#md5#	rol	\$$rot1[$j%4],$a
+#md5#	mov	`$j==15?"$c":"$b"`,$tmp		# forward reference
+#rc4#	pinsrw	\$`($j>>1)&7`,($dat,$TX[0],4),$xmm\n
+#md5#	add	$b,$a
+___
+    $code.=<<___ if ($rc4 && $j==15 && $k==$MOD-1);
+	mov	$YY,$XX[1]
+	xor	$YY,$YY				# keyword to partial register
+	mov	$XX[1]#b,$YY#b
+	lea	($dat,$XX[0],4),$XX[1]
+___
+    $code.=<<___ if ($rc4 && $j==15);
+	psllq	\$8,%xmm1
+	pxor	%xmm0,%xmm3
+	pxor	%xmm1,%xmm3
+___
+}
+sub R2 {
+  my ($i,$a,$b,$c,$d)=@_;
+  my @rot2=(4,11,16,23);
+  my $j=$i%16;
+  my $k=$i%$MOD;
+  my $xmm="%xmm".($j&1);
+    $code.="	movdqu	32($in0),%xmm4\n"	if ($rc4 && $j==15);
+    $code.="	add	\$$MOD,$XX[0]#b\n"	if ($rc4 && $j==15 && $k==$MOD-1);
+    $code.="	pxor	$xmm,$xmm\n"		if ($rc4 && $j<=1);
+    $code.=<<___;
+#rc4#	movl	($dat,$YY,4),$TY#d
+#md5#	xor	$c,$tmp
+#rc4#	movl	$TX[0]#d,($dat,$YY,4)
+#md5#	xor	$b,$tmp
+#md5#	add	4*`((5+3*$j)%16)`($inp),$a
+#rc4#	add	$TY#b,$TX[0]#b
+#rc4#	movl	`4*(($k+1)%$MOD)`(`$k==$MOD-1?"$dat,$XX[0],4":"$XX[1]"`),$TX[1]#d
+#md5#	add	\$$K[$i],$a
+#rc4#	movz	$TX[0]#b,$TX[0]#d
+#md5#	add	$tmp,$a
+#rc4#	movl	$TY#d,4*$k($XX[1])
+#rc4#	add	$TX[1]#b,$YY#b
+#md5#	rol	\$$rot2[$j%4],$a
+#md5#	mov	`$j==15?"\\\$-1":"$c"`,$tmp	# forward reference
+#rc4#	pinsrw	\$`($j>>1)&7`,($dat,$TX[0],4),$xmm\n
+#md5#	add	$b,$a
+___
+    $code.=<<___ if ($rc4 && $j==15 && $k==$MOD-1);
+	mov	$YY,$XX[1]
+	xor	$YY,$YY				# keyword to partial register
+	mov	$XX[1]#b,$YY#b
+	lea	($dat,$XX[0],4),$XX[1]
+___
+    $code.=<<___ if ($rc4 && $j==15);
+	psllq	\$8,%xmm1
+	pxor	%xmm0,%xmm4
+	pxor	%xmm1,%xmm4
+___
+}
+sub R3 {
+  my ($i,$a,$b,$c,$d)=@_;
+  my @rot3=(6,10,15,21);
+  my $j=$i%16;
+  my $k=$i%$MOD;
+  my $xmm="%xmm".($j&1);
+    $code.="	movdqu	48($in0),%xmm5\n"	if ($rc4 && $j==15);
+    $code.="	add	\$$MOD,$XX[0]#b\n"	if ($rc4 && $j==15 && $k==$MOD-1);
+    $code.="	pxor	$xmm,$xmm\n"		if ($rc4 && $j<=1);
+    $code.=<<___;
+#rc4#	movl	($dat,$YY,4),$TY#d
+#md5#	xor	$d,$tmp
+#rc4#	movl	$TX[0]#d,($dat,$YY,4)
+#md5#	or	$b,$tmp
+#md5#	add	4*`((7*$j)%16)`($inp),$a
+#rc4#	add	$TY#b,$TX[0]#b
+#rc4#	movl	`4*(($k+1)%$MOD)`(`$k==$MOD-1?"$dat,$XX[0],4":"$XX[1]"`),$TX[1]#d
+#md5#	add	\$$K[$i],$a
+#rc4#	movz	$TX[0]#b,$TX[0]#d
+#md5#	xor	$c,$tmp
+#rc4#	movl	$TY#d,4*$k($XX[1])
+#md5#	add	$tmp,$a
+#rc4#	add	$TX[1]#b,$YY#b
+#md5#	rol	\$$rot3[$j%4],$a
+#md5#	mov	\$-1,$tmp			# forward reference
+#rc4#	pinsrw	\$`($j>>1)&7`,($dat,$TX[0],4),$xmm\n
+#md5#	add	$b,$a
+___
+    $code.=<<___ if ($rc4 && $j==15);
+	mov	$XX[0],$XX[1]
+	xor	$XX[0],$XX[0]			# keyword to partial register
+	mov	$XX[1]#b,$XX[0]#b
+	mov	$YY,$XX[1]
+	xor	$YY,$YY				# keyword to partial register
+	mov	$XX[1]#b,$YY#b
+	lea	($dat,$XX[0],4),$XX[1]
+	psllq	\$8,%xmm1
+	pxor	%xmm0,%xmm5
+	pxor	%xmm1,%xmm5
+___
+}
+
+my $i=0;
+for(;$i<16;$i++) { R0($i,@V); unshift(@V,pop(@V)); push(@TX,shift(@TX)); }
+for(;$i<32;$i++) { R1($i,@V); unshift(@V,pop(@V)); push(@TX,shift(@TX)); }
+for(;$i<48;$i++) { R2($i,@V); unshift(@V,pop(@V)); push(@TX,shift(@TX)); }
+for(;$i<64;$i++) { R3($i,@V); unshift(@V,pop(@V)); push(@TX,shift(@TX)); }
+
+$code.=<<___;
+#md5#	add	0*4(%rsp),$V[0]		# accumulate hash value
+#md5#	add	1*4(%rsp),$V[1]
+#md5#	add	2*4(%rsp),$V[2]
+#md5#	add	3*4(%rsp),$V[3]
+
+#rc4#	movdqu	%xmm2,($out,$in0)	# write RC4 output
+#rc4#	movdqu	%xmm3,16($out,$in0)
+#rc4#	movdqu	%xmm4,32($out,$in0)
+#rc4#	movdqu	%xmm5,48($out,$in0)
+#md5#	lea	64($inp),$inp
+#rc4#	lea	64($in0),$in0
+	cmp	16(%rsp),$inp		# are we done?
+	jb	.Loop
+
+#md5#	mov	24(%rsp),$len		# restore pointer to MD5_CTX
+#rc4#	sub	$TX[0]#b,$YY#b		# correct $YY
+#md5#	mov	$V[0],0*4($len)		# write MD5_CTX
+#md5#	mov	$V[1],1*4($len)
+#md5#	mov	$V[2],2*4($len)
+#md5#	mov	$V[3],3*4($len)
+___
+$code.=<<___ if ($rc4 && (!$md5 || $D));
+	mov	32(%rsp),$len		# restore original $len
+	and	\$63,$len		# remaining bytes
+	jnz	.Loop1
+	jmp	.Ldone
+	
+.align	16
+.Loop1:
+	add	$TX[0]#b,$YY#b
+	movl	($dat,$YY,4),$TY#d
+	movl	$TX[0]#d,($dat,$YY,4)
+	movl	$TY#d,($dat,$XX[0],4)
+	add	$TY#b,$TX[0]#b
+	inc	$XX[0]#b
+	movl	($dat,$TX[0],4),$TY#d
+	movl	($dat,$XX[0],4),$TX[0]#d
+	xorb	($in0),$TY#b
+	movb	$TY#b,($out,$in0)
+	lea	1($in0),$in0
+	dec	$len
+	jnz	.Loop1
+
+.Ldone:
+___
+$code.=<<___;
+#rc4#	sub	\$1,$XX[0]#b
+#rc4#	movl	$XX[0]#d,-8($dat)
+#rc4#	movl	$YY#d,-4($dat)
+
+	mov	40(%rsp),%r15
+	mov	48(%rsp),%r14
+	mov	56(%rsp),%r13
+	mov	64(%rsp),%r12
+	mov	72(%rsp),%rbp
+	mov	80(%rsp),%rbx
+	lea	88(%rsp),%rsp
+.Lepilogue:
+.Labort:
+	ret
+.size $func,.-$func
+___
+
+if ($rc4 && $D) {	# sole purpose of this section is to provide
+			# option to use the generated module as drop-in
+			# replacement for rc4-x86_64.pl for debugging
+			# and testing purposes...
+my ($idx,$ido)=("%r8","%r9");
+my ($dat,$len,$inp)=("%rdi","%rsi","%rdx");
+
+$code.=<<___;
+.globl	RC4_set_key
+.type	RC4_set_key,\@function,3
+.align	16
+RC4_set_key:
+	lea	8($dat),$dat
+	lea	($inp,$len),$inp
+	neg	$len
+	mov	$len,%rcx
+	xor	%eax,%eax
+	xor	$ido,$ido
+	xor	%r10,%r10
+	xor	%r11,%r11
+	jmp	.Lw1stloop
+
+.align	16
+.Lw1stloop:
+	mov	%eax,($dat,%rax,4)
+	add	\$1,%al
+	jnc	.Lw1stloop
+
+	xor	$ido,$ido
+	xor	$idx,$idx
+.align	16
+.Lw2ndloop:
+	mov	($dat,$ido,4),%r10d
+	add	($inp,$len,1),$idx#b
+	add	%r10b,$idx#b
+	add	\$1,$len
+	mov	($dat,$idx,4),%r11d
+	cmovz	%rcx,$len
+	mov	%r10d,($dat,$idx,4)
+	mov	%r11d,($dat,$ido,4)
+	add	\$1,$ido#b
+	jnc	.Lw2ndloop
+
+	xor	%eax,%eax
+	mov	%eax,-8($dat)
+	mov	%eax,-4($dat)
+	ret
+.size	RC4_set_key,.-RC4_set_key
+
+.globl	RC4_options
+.type	RC4_options,\@abi-omnipotent
+.align	16
+RC4_options:
+	lea	.Lopts(%rip),%rax
+	ret
+.align	64
+.Lopts:
+.asciz	"rc4(64x,int)"
+.align	64
+.size	RC4_options,.-RC4_options
+___
+}
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+my $rec="%rcx";
+my $frame="%rdx";
+my $context="%r8";
+my $disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lbody(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lbody
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lepilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lepilogue
+	jae	.Lin_prologue
+
+	mov	40(%rax),%r15
+	mov	48(%rax),%r14
+	mov	56(%rax),%r13
+	mov	64(%rax),%r12
+	mov	72(%rax),%rbp
+	mov	80(%rax),%rbx
+	lea	88(%rax),%rax
+
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R12
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	se_handler,.-se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_$func
+	.rva	.LSEH_end_$func
+	.rva	.LSEH_info_$func
+
+.section	.xdata
+.align	8
+.LSEH_info_$func:
+	.byte	9,0,0,0
+	.rva	se_handler
+___
+}
+
+sub reg_part {
+my ($reg,$conv)=@_;
+    if ($reg =~ /%r[0-9]+/)     { $reg .= $conv; }
+    elsif ($conv eq "b")        { $reg =~ s/%[er]([^x]+)x?/%$1l/;       }
+    elsif ($conv eq "w")        { $reg =~ s/%[er](.+)/%$1/;             }
+    elsif ($conv eq "d")        { $reg =~ s/%[er](.+)/%e$1/;            }
+    return $reg;
+}
+
+$code =~ s/(%[a-z0-9]+)#([bwd])/reg_part($1,$2)/gem;
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/pinsrw\s+\$0,/movd	/gm;
+
+$code =~ s/#md5#//gm	if ($md5);
+$code =~ s/#rc4#//gm	if ($rc4);
+
+print $code;
+
+close STDOUT;
diff --git a/crypto/rc4/asm/rc4-parisc.pl b/crypto/rc4/asm/rc4-parisc.pl
new file mode 100755
index 000000000000..9165067080ef
--- /dev/null
+++ b/crypto/rc4/asm/rc4-parisc.pl
@@ -0,0 +1,313 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# RC4 for PA-RISC.
+
+# June 2009.
+#
+# Performance is 33% better than gcc 3.2 generated code on PA-7100LC.
+# For reference, [4x] unrolled loop is >40% faster than folded one.
+# It's possible to unroll loop 8 times on PA-RISC 2.0, but improvement
+# is believed to be not sufficient to justify the effort...
+#
+# Special thanks to polarhome.com for providing HP-UX account.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+
+$flavour = shift;
+$output = shift;
+open STDOUT,">$output";
+
+if ($flavour =~ /64/) {
+	$LEVEL		="2.0W";
+	$SIZE_T		=8;
+	$FRAME_MARKER	=80;
+	$SAVED_RP	=16;
+	$PUSH		="std";
+	$PUSHMA		="std,ma";
+	$POP		="ldd";
+	$POPMB		="ldd,mb";
+} else {
+	$LEVEL		="1.0";
+	$SIZE_T		=4;
+	$FRAME_MARKER	=48;
+	$SAVED_RP	=20;
+	$PUSH		="stw";
+	$PUSHMA		="stwm";
+	$POP		="ldw";
+	$POPMB		="ldwm";
+}
+
+$FRAME=4*$SIZE_T+$FRAME_MARKER;	# 4 saved regs + frame marker
+				#                [+ argument transfer]
+$SZ=1;				# defaults to RC4_CHAR
+if (open CONF,"<${dir}../../opensslconf.h") {
+    while(<CONF>) {
+	if (m/#\s*define\s+RC4_INT\s+(.*)/) {
+	    $SZ = ($1=~/char$/) ? 1 : 4;
+	    last;
+	}
+    }
+    close CONF;
+}
+
+if ($SZ==1) {	# RC4_CHAR
+    $LD="ldb";
+    $LDX="ldbx";
+    $MKX="addl";
+    $ST="stb";
+} else {	# RC4_INT (~5% faster than RC4_CHAR on PA-7100LC)
+    $LD="ldw";
+    $LDX="ldwx,s";
+    $MKX="sh2addl";
+    $ST="stw";
+}
+
+$key="%r26";
+$len="%r25";
+$inp="%r24";
+$out="%r23";
+
+@XX=("%r19","%r20");
+@TX=("%r21","%r22");
+$YY="%r28";
+$TY="%r29";
+
+$acc="%r1";
+$ix="%r2";
+$iy="%r3";
+$dat0="%r4";
+$dat1="%r5";
+$rem="%r6";
+$mask="%r31";
+
+sub unrolledloopbody {
+for ($i=0;$i<4;$i++) {
+$code.=<<___;
+	ldo	1($XX[0]),$XX[1]
+	`sprintf("$LDX	%$TY(%$key),%$dat1") if ($i>0)`	
+	and	$mask,$XX[1],$XX[1]
+	$LDX	$YY($key),$TY
+	$MKX	$YY,$key,$ix
+	$LDX	$XX[1]($key),$TX[1]
+	$MKX	$XX[0],$key,$iy
+	$ST	$TX[0],0($ix)
+	comclr,<> $XX[1],$YY,%r0	; conditional
+	copy	$TX[0],$TX[1]		; move
+	`sprintf("%sdep	%$dat1,%d,8,%$acc",$i==1?"z":"",8*($i-1)+7) if ($i>0)`
+	$ST	$TY,0($iy)
+	addl	$TX[0],$TY,$TY
+	addl	$TX[1],$YY,$YY
+	and	$mask,$TY,$TY
+	and	$mask,$YY,$YY
+___
+push(@TX,shift(@TX)); push(@XX,shift(@XX));	# "rotate" registers
+} }
+
+sub foldedloop {
+my ($label,$count)=@_;
+$code.=<<___;
+$label
+	$MKX	$YY,$key,$iy
+	$LDX	$YY($key),$TY
+	$MKX	$XX[0],$key,$ix
+	$ST	$TX[0],0($iy)
+	ldo	1($XX[0]),$XX[0]
+	$ST	$TY,0($ix)
+	addl	$TX[0],$TY,$TY
+	ldbx	$inp($out),$dat1
+	and	$mask,$TY,$TY
+	and	$mask,$XX[0],$XX[0]
+	$LDX	$TY($key),$acc
+	$LDX	$XX[0]($key),$TX[0]
+	ldo	1($out),$out
+	xor	$dat1,$acc,$acc
+	addl	$TX[0],$YY,$YY
+	stb	$acc,-1($out)
+	addib,<> -1,$count,$label	; $count is always small
+	and	$mask,$YY,$YY
+___
+}
+
+$code=<<___;
+	.LEVEL	$LEVEL
+	.SPACE	\$TEXT\$
+	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY
+
+	.EXPORT	RC4,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR,ARGW3=GR
+RC4
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-4*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=6
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+
+	cmpib,*= 0,$len,L\$abort
+	sub	$inp,$out,$inp		; distance between $inp and $out
+
+	$LD	`0*$SZ`($key),$XX[0]
+	$LD	`1*$SZ`($key),$YY
+	ldo	`2*$SZ`($key),$key
+
+	ldi	0xff,$mask
+	ldi	3,$dat0		
+
+	ldo	1($XX[0]),$XX[0]	; warm up loop
+	and	$mask,$XX[0],$XX[0]
+	$LDX	$XX[0]($key),$TX[0]
+	addl	$TX[0],$YY,$YY
+	cmpib,*>>= 6,$len,L\$oop1	; is $len large enough to bother?
+	and	$mask,$YY,$YY
+
+	and,<>	$out,$dat0,$rem		; is $out aligned?
+	b	L\$alignedout
+	subi	4,$rem,$rem
+	sub	$len,$rem,$len
+___
+&foldedloop("L\$alignout",$rem);	# process till $out is aligned
+
+$code.=<<___;
+L\$alignedout				; $len is at least 4 here
+	and,<>	$inp,$dat0,$acc		; is $inp aligned?
+	b	L\$oop4
+	sub	$inp,$acc,$rem		; align $inp
+
+	sh3addl	$acc,%r0,$acc
+	subi	32,$acc,$acc
+	mtctl	$acc,%cr11		; load %sar with vshd align factor
+	ldwx	$rem($out),$dat0
+	ldo	4($rem),$rem
+L\$oop4misalignedinp
+___
+&unrolledloopbody();
+$code.=<<___;
+	$LDX	$TY($key),$ix
+	ldwx	$rem($out),$dat1
+	ldo	-4($len),$len
+	or	$ix,$acc,$acc		; last piece, no need to dep
+	vshd	$dat0,$dat1,$iy		; align data
+	copy	$dat1,$dat0
+	xor	$iy,$acc,$acc
+	stw	$acc,0($out)
+	cmpib,*<< 3,$len,L\$oop4misalignedinp
+	ldo	4($out),$out
+	cmpib,*= 0,$len,L\$done
+	nop
+	b	L\$oop1
+	nop
+
+	.ALIGN	8
+L\$oop4
+___
+&unrolledloopbody();
+$code.=<<___;
+	$LDX	$TY($key),$ix
+	ldwx	$inp($out),$dat0
+	ldo	-4($len),$len
+	or	$ix,$acc,$acc		; last piece, no need to dep
+	xor	$dat0,$acc,$acc
+	stw	$acc,0($out)
+	cmpib,*<< 3,$len,L\$oop4
+	ldo	4($out),$out
+	cmpib,*= 0,$len,L\$done
+	nop
+___
+&foldedloop("L\$oop1",$len);
+$code.=<<___;
+L\$done
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2
+	ldo	-1($XX[0]),$XX[0]	; chill out loop
+	sub	$YY,$TX[0],$YY
+	and	$mask,$XX[0],$XX[0]
+	and	$mask,$YY,$YY
+	$ST	$XX[0],`-2*$SZ`($key)
+	$ST	$YY,`-1*$SZ`($key)
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+L\$abort
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+___
+
+$code.=<<___;
+
+	.EXPORT	private_RC4_set_key,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR
+	.ALIGN	8
+private_RC4_set_key
+	.PROC
+	.CALLINFO	NO_CALLS
+	.ENTRY
+	$ST	%r0,`0*$SZ`($key)
+	$ST	%r0,`1*$SZ`($key)
+	ldo	`2*$SZ`($key),$key
+	copy	%r0,@XX[0]
+L\$1st
+	$ST	@XX[0],0($key)
+	ldo	1(@XX[0]),@XX[0]
+	bb,>=	@XX[0],`31-8`,L\$1st	; @XX[0]<256
+	ldo	$SZ($key),$key
+
+	ldo	`-256*$SZ`($key),$key	; rewind $key
+	addl	$len,$inp,$inp		; $inp to point at the end
+	sub	%r0,$len,%r23		; inverse index
+	copy	%r0,@XX[0]
+	copy	%r0,@XX[1]
+	ldi	0xff,$mask
+
+L\$2nd
+	$LDX	@XX[0]($key),@TX[0]
+	ldbx	%r23($inp),@TX[1]
+	addi,nuv 1,%r23,%r23		; increment and conditional
+	sub	%r0,$len,%r23		; inverse index
+	addl	@TX[0],@XX[1],@XX[1]
+	addl	@TX[1],@XX[1],@XX[1]
+	and	$mask,@XX[1],@XX[1]
+	$MKX	@XX[0],$key,$TY
+	$LDX	@XX[1]($key),@TX[1]
+	$MKX	@XX[1],$key,$YY
+	ldo	1(@XX[0]),@XX[0]
+	$ST	@TX[0],0($YY)
+	bb,>=	@XX[0],`31-8`,L\$2nd	; @XX[0]<256
+	$ST	@TX[1],0($TY)
+
+	bv,n	(%r2)
+	.EXIT
+	nop
+	.PROCEND
+
+	.EXPORT	RC4_options,ENTRY
+	.ALIGN	8
+RC4_options
+	.PROC
+	.CALLINFO	NO_CALLS
+	.ENTRY
+	blr	%r0,%r28
+	ldi	3,%r1
+L\$pic
+	andcm	%r28,%r1,%r28
+	bv	(%r2)
+	.EXIT
+	ldo	L\$opts-L\$pic(%r28),%r28
+	.PROCEND
+	.ALIGN	8
+L\$opts
+	.STRINGZ "rc4(4x,`$SZ==1?"char":"int"`)"
+	.STRINGZ "RC4 for PA-RISC, CRYPTOGAMS by <appro\@openssl.org>"
+___
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/cmpib,\*/comib,/gm if ($SIZE_T==4);
+
+print $code;
+close STDOUT;
diff --git a/crypto/rc4/asm/rc4-s390x.pl b/crypto/rc4/asm/rc4-s390x.pl
new file mode 100755
index 000000000000..7528ece13c30
--- /dev/null
+++ b/crypto/rc4/asm/rc4-s390x.pl
@@ -0,0 +1,234 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# February 2009
+#
+# Performance is 2x of gcc 3.4.6 on z10. Coding "secret" is to
+# "cluster" Address Generation Interlocks, so that one pipeline stall
+# resolves several dependencies.
+
+# November 2010.
+#
+# Adapt for -m31 build. If kernel supports what's called "highgprs"
+# feature on Linux [see /proc/cpuinfo], it's possible to use 64-bit
+# instructions and achieve "64-bit" performance even in 31-bit legacy
+# application context. The feature is not specific to any particular
+# processor, as long as it's "z-CPU". Latter implies that the code
+# remains z/Architecture specific. On z990 it was measured to perform
+# 50% better than code generated by gcc 4.3.
+
+$flavour = shift;
+
+if ($flavour =~ /3[12]/) {
+	$SIZE_T=4;
+	$g="";
+} else {
+	$SIZE_T=8;
+	$g="g";
+}
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$rp="%r14";
+$sp="%r15";
+$code=<<___;
+.text
+
+___
+
+# void RC4(RC4_KEY *key,size_t len,const void *inp,void *out)
+{
+$acc="%r0";
+$cnt="%r1";
+$key="%r2";
+$len="%r3";
+$inp="%r4";
+$out="%r5";
+
+@XX=("%r6","%r7");
+@TX=("%r8","%r9");
+$YY="%r10";
+$TY="%r11";
+
+$code.=<<___;
+.globl	RC4
+.type	RC4,\@function
+.align	64
+RC4:
+	stm${g}	%r6,%r11,6*$SIZE_T($sp)
+___
+$code.=<<___ if ($flavour =~ /3[12]/);
+	llgfr	$len,$len
+___
+$code.=<<___;
+	llgc	$XX[0],0($key)
+	llgc	$YY,1($key)
+	la	$XX[0],1($XX[0])
+	nill	$XX[0],0xff
+	srlg	$cnt,$len,3
+	ltgr	$cnt,$cnt
+	llgc	$TX[0],2($XX[0],$key)
+	jz	.Lshort
+	j	.Loop8
+
+.align	64
+.Loop8:
+___
+for ($i=0;$i<8;$i++) {
+$code.=<<___;
+	la	$YY,0($YY,$TX[0])	# $i
+	nill	$YY,255
+	la	$XX[1],1($XX[0])
+	nill	$XX[1],255
+___
+$code.=<<___ if ($i==1);
+	llgc	$acc,2($TY,$key)
+___
+$code.=<<___ if ($i>1);
+	sllg	$acc,$acc,8
+	ic	$acc,2($TY,$key)
+___
+$code.=<<___;
+	llgc	$TY,2($YY,$key)
+	stc	$TX[0],2($YY,$key)
+	llgc	$TX[1],2($XX[1],$key)
+	stc	$TY,2($XX[0],$key)
+	cr	$XX[1],$YY
+	jne	.Lcmov$i
+	la	$TX[1],0($TX[0])
+.Lcmov$i:
+	la	$TY,0($TY,$TX[0])
+	nill	$TY,255
+___
+push(@TX,shift(@TX)); push(@XX,shift(@XX));     # "rotate" registers
+}
+
+$code.=<<___;
+	lg	$TX[1],0($inp)
+	sllg	$acc,$acc,8
+	la	$inp,8($inp)
+	ic	$acc,2($TY,$key)
+	xgr	$acc,$TX[1]
+	stg	$acc,0($out)
+	la	$out,8($out)
+	brctg	$cnt,.Loop8
+
+.Lshort:
+	lghi	$acc,7
+	ngr	$len,$acc
+	jz	.Lexit
+	j	.Loop1
+
+.align	16
+.Loop1:
+	la	$YY,0($YY,$TX[0])
+	nill	$YY,255
+	llgc	$TY,2($YY,$key)
+	stc	$TX[0],2($YY,$key)
+	stc	$TY,2($XX[0],$key)
+	ar	$TY,$TX[0]
+	ahi	$XX[0],1
+	nill	$TY,255
+	nill	$XX[0],255
+	llgc	$acc,0($inp)
+	la	$inp,1($inp)
+	llgc	$TY,2($TY,$key)
+	llgc	$TX[0],2($XX[0],$key)
+	xr	$acc,$TY
+	stc	$acc,0($out)
+	la	$out,1($out)
+	brct	$len,.Loop1
+
+.Lexit:
+	ahi	$XX[0],-1
+	stc	$XX[0],0($key)
+	stc	$YY,1($key)
+	lm${g}	%r6,%r11,6*$SIZE_T($sp)
+	br	$rp
+.size	RC4,.-RC4
+.string	"RC4 for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+
+___
+}
+
+# void RC4_set_key(RC4_KEY *key,unsigned int len,const void *inp)
+{
+$cnt="%r0";
+$idx="%r1";
+$key="%r2";
+$len="%r3";
+$inp="%r4";
+$acc="%r5";
+$dat="%r6";
+$ikey="%r7";
+$iinp="%r8";
+
+$code.=<<___;
+.globl	private_RC4_set_key
+.type	private_RC4_set_key,\@function
+.align	64
+private_RC4_set_key:
+	stm${g}	%r6,%r8,6*$SIZE_T($sp)
+	lhi	$cnt,256
+	la	$idx,0(%r0)
+	sth	$idx,0($key)
+.align	4
+.L1stloop:
+	stc	$idx,2($idx,$key)
+	la	$idx,1($idx)
+	brct	$cnt,.L1stloop
+
+	lghi	$ikey,-256
+	lr	$cnt,$len
+	la	$iinp,0(%r0)
+	la	$idx,0(%r0)
+.align	16
+.L2ndloop:
+	llgc	$acc,2+256($ikey,$key)
+	llgc	$dat,0($iinp,$inp)
+	la	$idx,0($idx,$acc)
+	la	$ikey,1($ikey)
+	la	$idx,0($idx,$dat)
+	nill	$idx,255
+	la	$iinp,1($iinp)
+	tml	$ikey,255
+	llgc	$dat,2($idx,$key)
+	stc	$dat,2+256-1($ikey,$key)
+	stc	$acc,2($idx,$key)
+	jz	.Ldone
+	brct	$cnt,.L2ndloop
+	lr	$cnt,$len
+	la	$iinp,0(%r0)
+	j	.L2ndloop
+.Ldone:
+	lm${g}	%r6,%r8,6*$SIZE_T($sp)
+	br	$rp
+.size	private_RC4_set_key,.-private_RC4_set_key
+
+___
+}
+
+# const char *RC4_options()
+$code.=<<___;
+.globl	RC4_options
+.type	RC4_options,\@function
+.align	16
+RC4_options:
+	larl	%r2,.Loptions
+	br	%r14
+.size	RC4_options,.-RC4_options
+.section	.rodata
+.Loptions:
+.align	8
+.string	"rc4(8x,char)"
+___
+
+print $code;
+close STDOUT;	# force flush
diff --git a/crypto/rc4/asm/rc4-x86_64.pl b/crypto/rc4/asm/rc4-x86_64.pl
index 53ce20ea92e5..d6eac205e98e 100755
--- a/crypto/rc4/asm/rc4-x86_64.pl
+++ b/crypto/rc4/asm/rc4-x86_64.pl
@@ -7,6 +7,8 @@
 # details see http://www.openssl.org/~appro/cryptogams/.
 # ====================================================================
 #
+# July 2004
+#
 # 2.22x RC4 tune-up:-) It should be noted though that my hand [as in
 # "hand-coded assembler"] doesn't stand for the whole improvement
 # coefficient. It turned out that eliminating RC4_CHAR from config
@@ -19,6 +21,8 @@
 # to operate on partial registers, it turned out to be the best bet.
 # At least for AMD... How IA32E would perform remains to be seen...
 
+# November 2004
+#
 # As was shown by Marc Bevand reordering of couple of load operations
 # results in even higher performance gain of 3.3x:-) At least on
 # Opteron... For reference, 1x in this case is RC4_CHAR C-code
@@ -26,6 +30,8 @@
 # Latter means that if you want to *estimate* what to expect from
 # *your* Opteron, then multiply 54 by 3.3 and clock frequency in GHz.
 
+# November 2004
+#
 # Intel P4 EM64T core was found to run the AMD64 code really slow...
 # The only way to achieve comparable performance on P4 was to keep
 # RC4_CHAR. Kind of ironic, huh? As it's apparently impossible to
@@ -33,10 +39,14 @@
 # on either AMD and Intel platforms, I implement both cases. See
 # rc4_skey.c for further details...
 
+# April 2005
+#
 # P4 EM64T core appears to be "allergic" to 64-bit inc/dec. Replacing 
 # those with add/sub results in 50% performance improvement of folded
 # loop...
 
+# May 2005
+#
 # As was shown by Zou Nanhai loop unrolling can improve Intel EM64T
 # performance by >30% [unlike P4 32-bit case that is]. But this is
 # provided that loads are reordered even more aggressively! Both code
@@ -50,6 +60,8 @@
 # is not implemented, then this final RC4_CHAR code-path should be
 # preferred, as it provides better *all-round* performance].
 
+# March 2007
+#
 # Intel Core2 was observed to perform poorly on both code paths:-( It
 # apparently suffers from some kind of partial register stall, which
 # occurs in 64-bit mode only [as virtually identical 32-bit loop was
@@ -58,27 +70,59 @@
 # fit for Core2 and therefore the code was modified to skip cloop8 on
 # this CPU.
 
-$output=shift;
+# May 2010
+#
+# Intel Westmere was observed to perform suboptimally. Adding yet
+# another movzb to cloop1 improved performance by almost 50%! Core2
+# performance is improved too, but nominally...
+
+# May 2011
+#
+# The only code path that was not modified is P4-specific one. Non-P4
+# Intel code path optimization is heavily based on submission by Maxim
+# Perminov, Maxim Locktyukhin and Jim Guilford of Intel. I've used
+# some of the ideas even in attempt to optmize the original RC4_INT
+# code path... Current performance in cycles per processed byte (less
+# is better) and improvement coefficients relative to previous
+# version of this module are:
+#
+# Opteron	5.3/+0%(*)
+# P4		6.5
+# Core2		6.2/+15%(**)
+# Westmere	4.2/+60%
+# Sandy Bridge	4.2/+120%
+# Atom		9.3/+80%
+#
+# (*)	But corresponding loop has less instructions, which should have
+#	positive effect on upcoming Bulldozer, which has one less ALU.
+#	For reference, Intel code runs at 6.8 cpb rate on Opteron.
+# (**)	Note that Core2 result is ~15% lower than corresponding result
+#	for 32-bit code, meaning that it's possible to improve it,
+#	but more than likely at the cost of the others (see rc4-586.pl
+#	to get the idea)...
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
 
 $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
 ( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
 ( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
 die "can't locate x86_64-xlate.pl";
 
-open STDOUT,"| $^X $xlate $output";
+open STDOUT,"| $^X $xlate $flavour $output";
 
 $dat="%rdi";	    # arg1
 $len="%rsi";	    # arg2
 $inp="%rdx";	    # arg3
 $out="%rcx";	    # arg4
 
-@XX=("%r8","%r10");
-@TX=("%r9","%r11");
-$YY="%r12";
-$TY="%r13";
-
+{
 $code=<<___;
 .text
+.extern	OPENSSL_ia32cap_P
 
 .globl	RC4
 .type	RC4,\@function,4
@@ -87,62 +131,181 @@ RC4:	or	$len,$len
 	jne	.Lentry
 	ret
 .Lentry:
+	push	%rbx
 	push	%r12
 	push	%r13
+.Lprologue:
+	mov	$len,%r11
+	mov	$inp,%r12
+	mov	$out,%r13
+___
+my $len="%r11";		# reassign input arguments
+my $inp="%r12";
+my $out="%r13";
 
-	add	\$8,$dat
-	movl	-8($dat),$XX[0]#d
-	movl	-4($dat),$YY#d
+my @XX=("%r10","%rsi");
+my @TX=("%rax","%rbx");
+my $YY="%rcx";
+my $TY="%rdx";
+
+$code.=<<___;
+	xor	$XX[0],$XX[0]
+	xor	$YY,$YY
+
+	lea	8($dat),$dat
+	mov	-8($dat),$XX[0]#b
+	mov	-4($dat),$YY#b
 	cmpl	\$-1,256($dat)
 	je	.LRC4_CHAR
+	mov	OPENSSL_ia32cap_P(%rip),%r8d
+	xor	$TX[1],$TX[1]
 	inc	$XX[0]#b
+	sub	$XX[0],$TX[1]
+	sub	$inp,$out
 	movl	($dat,$XX[0],4),$TX[0]#d
-	test	\$-8,$len
+	test	\$-16,$len
 	jz	.Lloop1
-	jmp	.Lloop8
+	bt	\$30,%r8d	# Intel CPU?
+	jc	.Lintel
+	and	\$7,$TX[1]
+	lea	1($XX[0]),$XX[1]
+	jz	.Loop8
+	sub	$TX[1],$len
+.Loop8_warmup:
+	add	$TX[0]#b,$YY#b
+	movl	($dat,$YY,4),$TY#d
+	movl	$TX[0]#d,($dat,$YY,4)
+	movl	$TY#d,($dat,$XX[0],4)
+	add	$TY#b,$TX[0]#b
+	inc	$XX[0]#b
+	movl	($dat,$TX[0],4),$TY#d
+	movl	($dat,$XX[0],4),$TX[0]#d
+	xorb	($inp),$TY#b
+	movb	$TY#b,($out,$inp)
+	lea	1($inp),$inp
+	dec	$TX[1]
+	jnz	.Loop8_warmup
+
+	lea	1($XX[0]),$XX[1]
+	jmp	.Loop8
 .align	16
-.Lloop8:
+.Loop8:
 ___
 for ($i=0;$i<8;$i++) {
+$code.=<<___ if ($i==7);
+	add	\$8,$XX[1]#b
+___
 $code.=<<___;
 	add	$TX[0]#b,$YY#b
-	mov	$XX[0],$XX[1]
 	movl	($dat,$YY,4),$TY#d
-	ror	\$8,%rax			# ror is redundant when $i=0
-	inc	$XX[1]#b
-	movl	($dat,$XX[1],4),$TX[1]#d
-	cmp	$XX[1],$YY
 	movl	$TX[0]#d,($dat,$YY,4)
-	cmove	$TX[0],$TX[1]
-	movl	$TY#d,($dat,$XX[0],4)
+	movl	`4*($i==7?-1:$i)`($dat,$XX[1],4),$TX[1]#d
+	ror	\$8,%r8				# ror is redundant when $i=0
+	movl	$TY#d,4*$i($dat,$XX[0],4)
 	add	$TX[0]#b,$TY#b
-	movb	($dat,$TY,4),%al
+	movb	($dat,$TY,4),%r8b
 ___
-push(@TX,shift(@TX)); push(@XX,shift(@XX));	# "rotate" registers
+push(@TX,shift(@TX)); #push(@XX,shift(@XX));	# "rotate" registers
 }
 $code.=<<___;
-	ror	\$8,%rax
+	add	\$8,$XX[0]#b
+	ror	\$8,%r8
 	sub	\$8,$len
 
-	xor	($inp),%rax
-	add	\$8,$inp
-	mov	%rax,($out)
-	add	\$8,$out
+	xor	($inp),%r8
+	mov	%r8,($out,$inp)
+	lea	8($inp),$inp
 
 	test	\$-8,$len
-	jnz	.Lloop8
+	jnz	.Loop8
 	cmp	\$0,$len
 	jne	.Lloop1
+	jmp	.Lexit
+
+.align	16
+.Lintel:
+	test	\$-32,$len
+	jz	.Lloop1
+	and	\$15,$TX[1]
+	jz	.Loop16_is_hot
+	sub	$TX[1],$len
+.Loop16_warmup:
+	add	$TX[0]#b,$YY#b
+	movl	($dat,$YY,4),$TY#d
+	movl	$TX[0]#d,($dat,$YY,4)
+	movl	$TY#d,($dat,$XX[0],4)
+	add	$TY#b,$TX[0]#b
+	inc	$XX[0]#b
+	movl	($dat,$TX[0],4),$TY#d
+	movl	($dat,$XX[0],4),$TX[0]#d
+	xorb	($inp),$TY#b
+	movb	$TY#b,($out,$inp)
+	lea	1($inp),$inp
+	dec	$TX[1]
+	jnz	.Loop16_warmup
+
+	mov	$YY,$TX[1]
+	xor	$YY,$YY
+	mov	$TX[1]#b,$YY#b
+
+.Loop16_is_hot:
+	lea	($dat,$XX[0],4),$XX[1]
 ___
+sub RC4_loop {
+  my $i=shift;
+  my $j=$i<0?0:$i;
+  my $xmm="%xmm".($j&1);
+
+    $code.="	add	\$16,$XX[0]#b\n"		if ($i==15);
+    $code.="	movdqu	($inp),%xmm2\n"			if ($i==15);
+    $code.="	add	$TX[0]#b,$YY#b\n"		if ($i<=0);
+    $code.="	movl	($dat,$YY,4),$TY#d\n";
+    $code.="	pxor	%xmm0,%xmm2\n"			if ($i==0);
+    $code.="	psllq	\$8,%xmm1\n"			if ($i==0);
+    $code.="	pxor	$xmm,$xmm\n"			if ($i<=1);
+    $code.="	movl	$TX[0]#d,($dat,$YY,4)\n";
+    $code.="	add	$TY#b,$TX[0]#b\n";
+    $code.="	movl	`4*($j+1)`($XX[1]),$TX[1]#d\n"	if ($i<15);
+    $code.="	movz	$TX[0]#b,$TX[0]#d\n";
+    $code.="	movl	$TY#d,4*$j($XX[1])\n";
+    $code.="	pxor	%xmm1,%xmm2\n"			if ($i==0);
+    $code.="	lea	($dat,$XX[0],4),$XX[1]\n"	if ($i==15);
+    $code.="	add	$TX[1]#b,$YY#b\n"		if ($i<15);
+    $code.="	pinsrw	\$`($j>>1)&7`,($dat,$TX[0],4),$xmm\n";
+    $code.="	movdqu	%xmm2,($out,$inp)\n"		if ($i==0);
+    $code.="	lea	16($inp),$inp\n"		if ($i==0);
+    $code.="	movl	($XX[1]),$TX[1]#d\n"		if ($i==15);
+}
+	RC4_loop(-1);
 $code.=<<___;
-.Lexit:
-	sub	\$1,$XX[0]#b
-	movl	$XX[0]#d,-8($dat)
-	movl	$YY#d,-4($dat)
+	jmp	.Loop16_enter
+.align	16
+.Loop16:
+___
+
+for ($i=0;$i<16;$i++) {
+    $code.=".Loop16_enter:\n"		if ($i==1);
+	RC4_loop($i);
+	push(@TX,shift(@TX)); 		# "rotate" registers
+}
+$code.=<<___;
+	mov	$YY,$TX[1]
+	xor	$YY,$YY			# keyword to partial register
+	sub	\$16,$len
+	mov	$TX[1]#b,$YY#b
+	test	\$-16,$len
+	jnz	.Loop16
+
+	psllq	\$8,%xmm1
+	pxor	%xmm0,%xmm2
+	pxor	%xmm1,%xmm2
+	movdqu	%xmm2,($out,$inp)
+	lea	16($inp),$inp
+
+	cmp	\$0,$len
+	jne	.Lloop1
+	jmp	.Lexit
 
-	pop	%r13
-	pop	%r12
-	ret
 .align	16
 .Lloop1:
 	add	$TX[0]#b,$YY#b
@@ -154,9 +317,8 @@ $code.=<<___;
 	movl	($dat,$TX[0],4),$TY#d
 	movl	($dat,$XX[0],4),$TX[0]#d
 	xorb	($inp),$TY#b
-	inc	$inp
-	movb	$TY#b,($out)
-	inc	$out
+	movb	$TY#b,($out,$inp)
+	lea	1($inp),$inp
 	dec	$len
 	jnz	.Lloop1
 	jmp	.Lexit
@@ -167,14 +329,11 @@ $code.=<<___;
 	movzb	($dat,$XX[0]),$TX[0]#d
 	test	\$-8,$len
 	jz	.Lcloop1
-	cmpl	\$0,260($dat)
-	jnz	.Lcloop1
-	push	%rbx
 	jmp	.Lcloop8
 .align	16
 .Lcloop8:
-	mov	($inp),%eax
-	mov	4($inp),%ebx
+	mov	($inp),%r8d
+	mov	4($inp),%r9d
 ___
 # unroll 2x4-wise, because 64-bit rotates kill Intel P4...
 for ($i=0;$i<4;$i++) {
@@ -191,8 +350,8 @@ $code.=<<___;
 	mov	$TX[0],$TX[1]
 .Lcmov$i:
 	add	$TX[0]#b,$TY#b
-	xor	($dat,$TY),%al
-	ror	\$8,%eax
+	xor	($dat,$TY),%r8b
+	ror	\$8,%r8d
 ___
 push(@TX,shift(@TX)); push(@XX,shift(@XX));	# "rotate" registers
 }
@@ -210,21 +369,20 @@ $code.=<<___;
 	mov	$TX[0],$TX[1]
 .Lcmov$i:
 	add	$TX[0]#b,$TY#b
-	xor	($dat,$TY),%bl
-	ror	\$8,%ebx
+	xor	($dat,$TY),%r9b
+	ror	\$8,%r9d
 ___
 push(@TX,shift(@TX)); push(@XX,shift(@XX));	# "rotate" registers
 }
 $code.=<<___;
 	lea	-8($len),$len
-	mov	%eax,($out)
+	mov	%r8d,($out)
 	lea	8($inp),$inp
-	mov	%ebx,4($out)
+	mov	%r9d,4($out)
 	lea	8($out),$out
 
 	test	\$-8,$len
 	jnz	.Lcloop8
-	pop	%rbx
 	cmp	\$0,$len
 	jne	.Lcloop1
 	jmp	.Lexit
@@ -233,6 +391,7 @@ $code.=<<___;
 .align	16
 .Lcloop1:
 	add	$TX[0]#b,$YY#b
+	movzb	$YY#b,$YY#d
 	movzb	($dat,$YY),$TY#d
 	movb	$TX[0]#b,($dat,$YY)
 	movb	$TY#b,($dat,$XX[0])
@@ -249,18 +408,31 @@ $code.=<<___;
 	sub	\$1,$len
 	jnz	.Lcloop1
 	jmp	.Lexit
+
+.align	16
+.Lexit:
+	sub	\$1,$XX[0]#b
+	movl	$XX[0]#d,-8($dat)
+	movl	$YY#d,-4($dat)
+
+	mov	(%rsp),%r13
+	mov	8(%rsp),%r12
+	mov	16(%rsp),%rbx
+	add	\$24,%rsp
+.Lepilogue:
+	ret
 .size	RC4,.-RC4
 ___
+}
 
 $idx="%r8";
 $ido="%r9";
 
 $code.=<<___;
-.extern	OPENSSL_ia32cap_P
-.globl	RC4_set_key
-.type	RC4_set_key,\@function,3
+.globl	private_RC4_set_key
+.type	private_RC4_set_key,\@function,3
 .align	16
-RC4_set_key:
+private_RC4_set_key:
 	lea	8($dat),$dat
 	lea	($inp,$len),$inp
 	neg	$len
@@ -271,12 +443,9 @@ RC4_set_key:
 	xor	%r11,%r11
 
 	mov	OPENSSL_ia32cap_P(%rip),$idx#d
-	bt	\$20,$idx#d
-	jnc	.Lw1stloop
-	bt	\$30,$idx#d
-	setc	$ido#b
-	mov	$ido#d,260($dat)
-	jmp	.Lc1stloop
+	bt	\$20,$idx#d	# RC4_CHAR?
+	jc	.Lc1stloop
+	jmp	.Lw1stloop
 
 .align	16
 .Lw1stloop:
@@ -330,36 +499,177 @@ RC4_set_key:
 	mov	%eax,-8($dat)
 	mov	%eax,-4($dat)
 	ret
-.size	RC4_set_key,.-RC4_set_key
+.size	private_RC4_set_key,.-private_RC4_set_key
 
 .globl	RC4_options
-.type	RC4_options,\@function,0
+.type	RC4_options,\@abi-omnipotent
 .align	16
 RC4_options:
-	.picmeup %rax
-	lea	.Lopts-.(%rax),%rax
+	lea	.Lopts(%rip),%rax
 	mov	OPENSSL_ia32cap_P(%rip),%edx
 	bt	\$20,%edx
-	jnc	.Ldone
-	add	\$12,%rax
+	jc	.L8xchar
 	bt	\$30,%edx
 	jnc	.Ldone
-	add	\$13,%rax
+	add	\$25,%rax
+	ret
+.L8xchar:
+	add	\$12,%rax
 .Ldone:
 	ret
 .align	64
 .Lopts:
 .asciz	"rc4(8x,int)"
 .asciz	"rc4(8x,char)"
-.asciz	"rc4(1x,char)"
+.asciz	"rc4(16x,int)"
 .asciz	"RC4 for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
 .align	64
 .size	RC4_options,.-RC4_options
 ___
 
-$code =~ s/#([bwd])/$1/gm;
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	stream_se_handler,\@abi-omnipotent
+.align	16
+stream_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lprologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lepilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lin_prologue
+
+	lea	24(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%r12
+	mov	-24(%rax),%r13
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	jmp	.Lcommon_seh_exit
+.size	stream_se_handler,.-stream_se_handler
+
+.type	key_se_handler,\@abi-omnipotent
+.align	16
+key_se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	152($context),%rax	# pull context->Rsp
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+.Lcommon_seh_exit:
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	key_se_handler,.-key_se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_RC4
+	.rva	.LSEH_end_RC4
+	.rva	.LSEH_info_RC4
+
+	.rva	.LSEH_begin_private_RC4_set_key
+	.rva	.LSEH_end_private_RC4_set_key
+	.rva	.LSEH_info_private_RC4_set_key
+
+.section	.xdata
+.align	8
+.LSEH_info_RC4:
+	.byte	9,0,0,0
+	.rva	stream_se_handler
+.LSEH_info_private_RC4_set_key:
+	.byte	9,0,0,0
+	.rva	key_se_handler
+___
+}
+
+sub reg_part {
+my ($reg,$conv)=@_;
+    if ($reg =~ /%r[0-9]+/)	{ $reg .= $conv; }
+    elsif ($conv eq "b")	{ $reg =~ s/%[er]([^x]+)x?/%$1l/;	}
+    elsif ($conv eq "w")	{ $reg =~ s/%[er](.+)/%$1/;		}
+    elsif ($conv eq "d")	{ $reg =~ s/%[er](.+)/%e$1/;		}
+    return $reg;
+}
 
-$code =~ s/RC4_set_key/private_RC4_set_key/g if ($ENV{FIPSCANLIB} ne "");
+$code =~ s/(%[a-z0-9]+)#([bwd])/reg_part($1,$2)/gem;
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
 
 print $code;
 
diff --git a/crypto/rc4/rc4.h b/crypto/rc4/rc4.h
index 2d8620d33b91..88ceb46bc54d 100644
--- a/crypto/rc4/rc4.h
+++ b/crypto/rc4/rc4.h
@@ -64,6 +64,8 @@
 #error RC4 is disabled.
 #endif
 
+#include <stddef.h>
+
 #ifdef  __cplusplus
 extern "C" {
 #endif
@@ -76,11 +78,9 @@ typedef struct rc4_key_st
 
  
 const char *RC4_options(void);
-#ifdef OPENSSL_FIPS
-void private_RC4_set_key(RC4_KEY *key, int len, const unsigned char *data);
-#endif
 void RC4_set_key(RC4_KEY *key, int len, const unsigned char *data);
-void RC4(RC4_KEY *key, unsigned long len, const unsigned char *indata,
+void private_RC4_set_key(RC4_KEY *key, int len, const unsigned char *data);
+void RC4(RC4_KEY *key, size_t len, const unsigned char *indata,
 		unsigned char *outdata);
 
 #ifdef  __cplusplus
diff --git a/crypto/rc4/rc4_enc.c b/crypto/rc4/rc4_enc.c
index 0660ea60a25e..8c4fc6c7a3da 100644
--- a/crypto/rc4/rc4_enc.c
+++ b/crypto/rc4/rc4_enc.c
@@ -67,12 +67,12 @@
  * Date: Wed, 14 Sep 1994 06:35:31 GMT
  */
 
-void RC4(RC4_KEY *key, unsigned long len, const unsigned char *indata,
+void RC4(RC4_KEY *key, size_t len, const unsigned char *indata,
 	     unsigned char *outdata)
 	{
         register RC4_INT *d;
         register RC4_INT x,y,tx,ty;
-	int i;
+	size_t i;
         
         x=key->x;     
         y=key->y;     
@@ -120,8 +120,8 @@ void RC4(RC4_KEY *key, unsigned long len, const unsigned char *indata,
 			(RC4_CHUNK)d[(tx+ty)&0xff]\
 			)
 
-	if ( ( ((unsigned long)indata  & (sizeof(RC4_CHUNK)-1)) | 
-	       ((unsigned long)outdata & (sizeof(RC4_CHUNK)-1)) ) == 0 )
+	if ( ( ((size_t)indata  & (sizeof(RC4_CHUNK)-1)) | 
+	       ((size_t)outdata & (sizeof(RC4_CHUNK)-1)) ) == 0 )
 		{
 		RC4_CHUNK ichunk,otp;
 		const union { long one; char little; } is_endian = {1};
@@ -157,7 +157,7 @@ void RC4(RC4_KEY *key, unsigned long len, const unsigned char *indata,
 		if (!is_endian.little)
 			{	/* BIG-ENDIAN CASE */
 # define BESHFT(c)	(((sizeof(RC4_CHUNK)-(c)-1)*8)&(sizeof(RC4_CHUNK)*8-1))
-			for (;len&~(sizeof(RC4_CHUNK)-1);len-=sizeof(RC4_CHUNK))
+			for (;len&(0-sizeof(RC4_CHUNK));len-=sizeof(RC4_CHUNK))
 				{
 				ichunk  = *(RC4_CHUNK *)indata;
 				otp  = RC4_STEP<<BESHFT(0);
@@ -210,7 +210,7 @@ void RC4(RC4_KEY *key, unsigned long len, const unsigned char *indata,
 		else
 			{	/* LITTLE-ENDIAN CASE */
 # define LESHFT(c)	(((c)*8)&(sizeof(RC4_CHUNK)*8-1))
-			for (;len&~(sizeof(RC4_CHUNK)-1);len-=sizeof(RC4_CHUNK))
+			for (;len&(0-sizeof(RC4_CHUNK));len-=sizeof(RC4_CHUNK))
 				{
 				ichunk  = *(RC4_CHUNK *)indata;
 				otp  = RC4_STEP;
@@ -276,7 +276,7 @@ void RC4(RC4_KEY *key, unsigned long len, const unsigned char *indata,
 #define RC4_LOOP(a,b,i)	LOOP(a[i],b[i])
 #endif
 
-	i=(int)(len>>3L);
+	i=len>>3;
 	if (i)
 		{
 		for (;;)
@@ -296,7 +296,7 @@ void RC4(RC4_KEY *key, unsigned long len, const unsigned char *indata,
 			if (--i == 0) break;
 			}
 		}
-	i=(int)len&0x07;
+	i=len&0x07;
 	if (i)
 		{
 		for (;;)
diff --git a/crypto/rc4/rc4_skey.c b/crypto/rc4/rc4_skey.c
index d1dc912b2414..fda27636e7b3 100644
--- a/crypto/rc4/rc4_skey.c
+++ b/crypto/rc4/rc4_skey.c
@@ -59,11 +59,6 @@
 #include <openssl/rc4.h>
 #include "rc4_locl.h"
 #include <openssl/opensslv.h>
-#include <openssl/crypto.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 
 const char RC4_version[]="RC4" OPENSSL_VERSION_PTEXT;
 
@@ -90,11 +85,7 @@ const char *RC4_options(void)
  * Date: Wed, 14 Sep 1994 06:35:31 GMT
  */
 
-#ifdef OPENSSL_FIPS
 void private_RC4_set_key(RC4_KEY *key, int len, const unsigned char *data)
-#else
-void RC4_set_key(RC4_KEY *key, int len, const unsigned char *data)
-#endif
 	{
         register RC4_INT tmp;
         register int id1,id2;
@@ -113,46 +104,6 @@ void RC4_set_key(RC4_KEY *key, int len, const unsigned char *data)
 		d[(n)]=d[id2]; \
 		d[id2]=tmp; }
 
-#if defined(OPENSSL_CPUID_OBJ) && !defined(OPENSSL_NO_ASM)
-# if	defined(__i386)   || defined(__i386__)   || defined(_M_IX86) || \
-	defined(__INTEL__) || \
-	defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64)
-	if (sizeof(RC4_INT) > 1) {
-		/*
-		 * Unlike all other x86 [and x86_64] implementations,
-		 * Intel P4 core [including EM64T] was found to perform
-		 * poorly with wider RC4_INT. Performance improvement
-		 * for IA-32 hand-coded assembler turned out to be 2.8x
-		 * if re-coded for RC4_CHAR! It's however inappropriate
-		 * to just switch to RC4_CHAR for x86[_64], as non-P4
-		 * implementations suffer from significant performance
-		 * losses then, e.g. PIII exhibits >2x deterioration,
-		 * and so does Opteron. In order to assure optimal
-		 * all-round performance, we detect P4 at run-time by
-		 * checking upon reserved bit 20 in CPU capability
-		 * vector and set up compressed key schedule, which is
-		 * recognized by correspondingly updated assembler
-		 * module... Bit 20 is set up by OPENSSL_ia32_cpuid.
-		 *
-		 *				<appro@fy.chalmers.se>
-		 */
-#ifdef OPENSSL_FIPS
-		unsigned long *ia32cap_ptr = OPENSSL_ia32cap_loc();
-		if (ia32cap_ptr && (*ia32cap_ptr & (1<<20))) {
-#else
-		if (OPENSSL_ia32cap_P & (1<<20)) {
-#endif
-			unsigned char *cp=(unsigned char *)d;
-
-			for (i=0;i<256;i++) cp[i]=i;
-			for (i=0;i<256;i++) SK_LOOP(cp,i);
-			/* mark schedule as compressed! */
-			d[256/sizeof(RC4_INT)]=-1;
-			return;
-		}
-	}
-# endif
-#endif
 	for (i=0; i < 256; i++) d[i]=i;
 	for (i=0; i < 256; i+=4)
 		{
diff --git a/crypto/rc4/rc4_utl.c b/crypto/rc4/rc4_utl.c
new file mode 100644
index 000000000000..ab3f02fe6a99
--- /dev/null
+++ b/crypto/rc4/rc4_utl.c
@@ -0,0 +1,62 @@
+/* crypto/rc4/rc4_utl.c -*- mode:C; c-file-style: "eay" -*- */
+/* ====================================================================
+ * Copyright (c) 2011 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ */
+
+#include <openssl/opensslv.h>
+#include <openssl/crypto.h>
+#include <openssl/rc4.h>
+
+void RC4_set_key(RC4_KEY *key, int len, const unsigned char *data)
+	{
+#ifdef OPENSSL_FIPS
+	fips_cipher_abort(RC4);
+#endif
+	private_RC4_set_key(key, len, data);
+	}
diff --git a/crypto/rc4/rc4s.cpp b/crypto/rc4/rc4s.cpp
deleted file mode 100644
index 3814fde9972f..000000000000
--- a/crypto/rc4/rc4s.cpp
+++ /dev/null
@@ -1,73 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/rc4.h>
-
-void main(int argc,char *argv[])
-	{
-	unsigned char buffer[1024];
-	RC4_KEY ctx;
-	unsigned long s1,s2,e1,e2;
-	unsigned char k[16];
-	unsigned long data[2];
-	unsigned char iv[8];
-	int i,num=64,numm;
-	int j=0;
-
-	if (argc >= 2)
-		num=atoi(argv[1]);
-
-	if (num == 0) num=256;
-	if (num > 1024-16) num=1024-16;
-	numm=num+8;
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<10; i++) /**/
-			{
-			RC4(&ctx,numm,buffer,buffer);
-			GetTSC(s1);
-			RC4(&ctx,numm,buffer,buffer);
-			GetTSC(e1);
-			GetTSC(s2);
-			RC4(&ctx,num,buffer,buffer);
-			GetTSC(e2);
-			RC4(&ctx,num,buffer,buffer);
-			}
-
-		printf("RC4 (%d bytes) %d %d (%d) - 8 bytes\n",num,
-			e1-s1,e2-s2,(e1-s1)-(e2-s2));
-		}
-	}
-
diff --git a/crypto/rc4/rc4test.c b/crypto/rc4/rc4test.c
index 54b597fa266e..4312605ccbca 100644
--- a/crypto/rc4/rc4test.c
+++ b/crypto/rc4/rc4test.c
@@ -114,12 +114,18 @@ static unsigned char output[7][30]={
 
 int main(int argc, char *argv[])
 	{
-	int err=0;
-	unsigned int i, j;
+	int i,err=0;
+	int j;
 	unsigned char *p;
 	RC4_KEY key;
 	unsigned char obuf[512];
 
+#if !defined(OPENSSL_PIC)
+	void OPENSSL_cpuid_setup(void);
+
+	OPENSSL_cpuid_setup();
+#endif
+
 	for (i=0; i<6; i++)
 		{
 		RC4_set_key(&key,keys[i][0],&(keys[i][1]));
@@ -129,12 +135,12 @@ int main(int argc, char *argv[])
 			{
 			printf("error calculating RC4\n");
 			printf("output:");
-			for (j=0; j<data_len[i]+1U; j++)
+			for (j=0; j<data_len[i]+1; j++)
 				printf(" %02x",obuf[j]);
 			printf("\n");
 			printf("expect:");
 			p= &(output[i][0]);
-			for (j=0; j<data_len[i]+1U; j++)
+			for (j=0; j<data_len[i]+1; j++)
 				printf(" %02x",*(p++));
 			printf("\n");
 			err++;
@@ -180,12 +186,12 @@ int main(int argc, char *argv[])
 			{
 			printf("error in RC4 multi-call processing\n");
 			printf("output:");
-			for (j=0; j<data_len[3]+1U; j++)
+			for (j=0; j<data_len[3]+1; j++)
 				printf(" %02x",obuf[j]);
 			printf("\n");
 			printf("expect:");
 			p= &(output[3][0]);
-			for (j=0; j<data_len[3]+1U; j++)
+			for (j=0; j<data_len[3]+1; j++)
 				printf(" %02x",*(p++));
 			err++;
 			}
@@ -216,11 +222,11 @@ int main(int argc, char *argv[])
 		if (memcmp(md,expected,sizeof(md))) {
 			printf("error in RC4 bulk test\n");
 			printf("output:");
-			for (j=0; j<sizeof(md); j++)
+			for (j=0; j<(int)sizeof(md); j++)
 				printf(" %02x",md[j]);
 			printf("\n");
 			printf("expect:");
-			for (j=0; j<sizeof(md); j++)
+			for (j=0; j<(int)sizeof(md); j++)
 				printf(" %02x",expected[j]);
 			printf("\n");
 			err++;
diff --git a/crypto/rc5/Makefile b/crypto/rc5/Makefile
index b4e21c9bb23c..8a8b00eb89e6 100644
--- a/crypto/rc5/Makefile
+++ b/crypto/rc5/Makefile
@@ -12,8 +12,6 @@ MAKEFILE=	Makefile
 AR=		ar r
 
 RC5_ENC=		rc5_enc.o
-# or use
-#DES_ENC=	r586-elf.o
 
 CFLAGS= $(INCLUDES) $(CFLAG)
 ASFLAGS= $(INCLUDES) $(ASFLAG)
@@ -40,19 +38,12 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-# ELF
-r586-elf.s: asm/rc5-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) rc5-586.pl elf $(CFLAGS) > ../$@)
-# COFF
-r586-cof.s: asm/rc5-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) rc5-586.pl coff $(CFLAGS) > ../$@)
-# a.out
-r586-out.s: asm/rc5-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
-	(cd asm; $(PERL) rc5-586.pl a.out $(CFLAGS) > ../$@)
+rc5-586.s: asm/rc5-586.pl ../perlasm/x86asm.pl ../perlasm/cbc.pl
+	$(PERL) asm/rc5-586.pl $(PERLASM_SCHEME) $(CFLAGS) > $@
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
diff --git a/crypto/rc5/asm/rc5-586.pl b/crypto/rc5/asm/rc5-586.pl
index edff1d1e64af..61ac6effc69c 100644
--- a/crypto/rc5/asm/rc5-586.pl
+++ b/crypto/rc5/asm/rc5-586.pl
@@ -1,6 +1,7 @@
 #!/usr/local/bin/perl
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 require "cbc.pl";
 
diff --git a/crypto/rc5/rc5.h b/crypto/rc5/rc5.h
index f73a2a02a451..4b3c153b5038 100644
--- a/crypto/rc5/rc5.h
+++ b/crypto/rc5/rc5.h
@@ -94,10 +94,7 @@ typedef struct rc5_key_st
 	RC5_32_INT data[2*(RC5_16_ROUNDS+1)];
 	} RC5_32_KEY;
 
-#ifdef OPENSSL_FIPS 
-void private_RC5_32_set_key(RC5_32_KEY *key, int len, const unsigned char *data,
-	int rounds);
-#endif
+ 
 void RC5_32_set_key(RC5_32_KEY *key, int len, const unsigned char *data,
 	int rounds);
 void RC5_32_ecb_encrypt(const unsigned char *in,unsigned char *out,RC5_32_KEY *key,
diff --git a/crypto/rc5/rc5_locl.h b/crypto/rc5/rc5_locl.h
index 282dd3882234..d337f73fad9b 100644
--- a/crypto/rc5/rc5_locl.h
+++ b/crypto/rc5/rc5_locl.h
@@ -154,14 +154,14 @@
 #  define ROTATE_l32(a,n)	({ register unsigned int ret;	\
 					asm ("roll %%cl,%0"	\
 						: "=r"(ret)	\
-						: "c"(n),"0"(a)	\
+						: "c"(n),"0"((unsigned int)(a))	\
 						: "cc");	\
 					ret;			\
 				})
 #  define ROTATE_r32(a,n)	({ register unsigned int ret;	\
 					asm ("rorl %%cl,%0"	\
 						: "=r"(ret)	\
-						: "c"(n),"0"(a)	\
+						: "c"(n),"0"((unsigned int)(a))	\
 						: "cc");	\
 					ret;			\
 				})
diff --git a/crypto/rc5/rc5_skey.c b/crypto/rc5/rc5_skey.c
index 247fa6500ce8..a2e00a41c550 100644
--- a/crypto/rc5/rc5_skey.c
+++ b/crypto/rc5/rc5_skey.c
@@ -56,28 +56,11 @@
  * [including the GNU Public Licence.]
  */
 
-#include <openssl/crypto.h>
 #include <openssl/rc5.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 #include "rc5_locl.h"
 
-#ifdef OPENSSL_FIPS
-void RC5_32_set_key(RC5_32_KEY *key, int len, const unsigned char *data,
-		    int rounds)
-	{
-	if (FIPS_mode())
-		FIPS_BAD_ABORT(RC5)
-	private_RC5_32_set_key(key, len, data, rounds);
-	}
-void private_RC5_32_set_key(RC5_32_KEY *key, int len, const unsigned char *data,
-		    int rounds)
-#else
 void RC5_32_set_key(RC5_32_KEY *key, int len, const unsigned char *data,
 		    int rounds)
-#endif
 	{
 	RC5_32_INT L[64],l,ll,A,B,*S,k;
 	int i,j,m,c,t,ii,jj;
diff --git a/crypto/rc5/rc5s.cpp b/crypto/rc5/rc5s.cpp
deleted file mode 100644
index 1c5518bc8045..000000000000
--- a/crypto/rc5/rc5s.cpp
+++ /dev/null
@@ -1,70 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/rc5.h>
-
-void main(int argc,char *argv[])
-	{
-	RC5_32_KEY key;
-	unsigned long s1,s2,e1,e2;
-	unsigned long data[2];
-	int i,j;
-	static unsigned char d[16]={0x01,0x23,0x45,0x67,0x89,0xAB,0xCD,0xEF};
-
-	RC5_32_set_key(&key, 16,d,12);
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<1000; i++) /**/
-			{
-			RC5_32_encrypt(&data[0],&key);
-			GetTSC(s1);
-			RC5_32_encrypt(&data[0],&key);
-			RC5_32_encrypt(&data[0],&key);
-			RC5_32_encrypt(&data[0],&key);
-			GetTSC(e1);
-			GetTSC(s2);
-			RC5_32_encrypt(&data[0],&key);
-			RC5_32_encrypt(&data[0],&key);
-			RC5_32_encrypt(&data[0],&key);
-			RC5_32_encrypt(&data[0],&key);
-			GetTSC(e2);
-			RC5_32_encrypt(&data[0],&key);
-			}
-
-		printf("cast %d %d (%d)\n",
-			e1-s1,e2-s2,((e2-s2)-(e1-s1)));
-		}
-	}
-
diff --git a/crypto/ripemd/Makefile b/crypto/ripemd/Makefile
index 6145f1369922..25140b2a73e2 100644
--- a/crypto/ripemd/Makefile
+++ b/crypto/ripemd/Makefile
@@ -38,19 +38,12 @@ top:
 all:    lib
 
 lib:    $(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-# ELF
-rm86-elf.s: asm/rmd-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) rmd-586.pl elf $(CFLAGS) > ../$@)
-# COFF
-rm86-cof.s: asm/rmd-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) rmd-586.pl coff $(CFLAGS) > ../$@)
-# a.out
-rm86-out.s: asm/rmd-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) rmd-586.pl a.out $(CFLAGS) > ../$@)
+rmd-586.s:	asm/rmd-586.pl ../perlasm/x86asm.pl
+	$(PERL) asm/rmd-586.pl $(PERLASM_SCHEME) $(CFLAGS) > $@
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
@@ -89,9 +82,7 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
-rmd_dgst.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
-rmd_dgst.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-rmd_dgst.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+rmd_dgst.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 rmd_dgst.o: ../../include/openssl/opensslconf.h
 rmd_dgst.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rmd_dgst.o: ../../include/openssl/ripemd.h ../../include/openssl/safestack.h
diff --git a/crypto/ripemd/asm/rips.cpp b/crypto/ripemd/asm/rips.cpp
deleted file mode 100644
index f7a13677a921..000000000000
--- a/crypto/ripemd/asm/rips.cpp
+++ /dev/null
@@ -1,82 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/ripemd.h>
-
-#define ripemd160_block_x86 ripemd160_block_asm_host_order
-
-extern "C" {
-void ripemd160_block_x86(RIPEMD160_CTX *ctx, unsigned char *buffer,int num);
-}
-
-void main(int argc,char *argv[])
-	{
-	unsigned char buffer[64*256];
-	RIPEMD160_CTX ctx;
-	unsigned long s1,s2,e1,e2;
-	unsigned char k[16];
-	unsigned long data[2];
-	unsigned char iv[8];
-	int i,num=0,numm;
-	int j=0;
-
-	if (argc >= 2)
-		num=atoi(argv[1]);
-
-	if (num == 0) num=16;
-	if (num > 250) num=16;
-	numm=num+2;
-#if 0
-	num*=64;
-	numm*=64;
-#endif
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<10; i++) /**/
-			{
-			ripemd160_block_x86(&ctx,buffer,numm);
-			GetTSC(s1);
-			ripemd160_block_x86(&ctx,buffer,numm);
-			GetTSC(e1);
-			GetTSC(s2);
-			ripemd160_block_x86(&ctx,buffer,num);
-			GetTSC(e2);
-			ripemd160_block_x86(&ctx,buffer,num);
-			}
-		printf("ripemd160 (%d bytes) %d %d (%.2f)\n",num*64,
-			e1-s1,e2-s2,(double)((e1-s1)-(e2-s2))/2);
-		}
-	}
-
diff --git a/crypto/ripemd/asm/rmd-586.pl b/crypto/ripemd/asm/rmd-586.pl
index 4f3c4c967f51..e8b2bc2db2de 100644
--- a/crypto/ripemd/asm/rmd-586.pl
+++ b/crypto/ripemd/asm/rmd-586.pl
@@ -5,7 +5,8 @@
 
 $normal=0;
 
-push(@INC,"perlasm","../../perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],$0);
diff --git a/crypto/ripemd/ripemd.h b/crypto/ripemd/ripemd.h
index 3b6d04386d4e..189bd8c90e92 100644
--- a/crypto/ripemd/ripemd.h
+++ b/crypto/ripemd/ripemd.h
@@ -70,7 +70,7 @@ extern "C" {
 #error RIPEMD is disabled.
 #endif
 
-#if defined(OPENSSL_SYS_WIN16) || defined(__LP32__)
+#if defined(__LP32__)
 #define RIPEMD160_LONG unsigned long
 #elif defined(OPENSSL_SYS_CRAY) || defined(__ILP64__)
 #define RIPEMD160_LONG unsigned long
@@ -90,6 +90,7 @@ typedef struct RIPEMD160state_st
 	RIPEMD160_LONG data[RIPEMD160_LBLOCK];
 	unsigned int   num;
 	} RIPEMD160_CTX;
+
 #ifdef OPENSSL_FIPS
 int private_RIPEMD160_Init(RIPEMD160_CTX *c);
 #endif
diff --git a/crypto/ripemd/rmd_dgst.c b/crypto/ripemd/rmd_dgst.c
index ead11d075ac9..63f0d983f7b4 100644
--- a/crypto/ripemd/rmd_dgst.c
+++ b/crypto/ripemd/rmd_dgst.c
@@ -59,11 +59,7 @@
 #include <stdio.h>
 #include "rmd_locl.h"
 #include <openssl/opensslv.h>
-#include <openssl/err.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
+#include <openssl/crypto.h>
 
 const char RMD160_version[]="RIPE-MD160" OPENSSL_VERSION_PTEXT;
 
@@ -74,16 +70,14 @@ const char RMD160_version[]="RIPE-MD160" OPENSSL_VERSION_PTEXT;
      void ripemd160_block(RIPEMD160_CTX *c, unsigned long *p,size_t num);
 #  endif
 
-FIPS_NON_FIPS_MD_Init(RIPEMD160)
+fips_md_init(RIPEMD160)
 	{
+	memset (c,0,sizeof(*c));
 	c->A=RIPEMD160_A;
 	c->B=RIPEMD160_B;
 	c->C=RIPEMD160_C;
 	c->D=RIPEMD160_D;
 	c->E=RIPEMD160_E;
-	c->Nl=0;
-	c->Nh=0;
-	c->num=0;
 	return 1;
 	}
 
diff --git a/crypto/ripemd/rmd_locl.h b/crypto/ripemd/rmd_locl.h
index ce12a8000ea0..f14b346e6622 100644
--- a/crypto/ripemd/rmd_locl.h
+++ b/crypto/ripemd/rmd_locl.h
@@ -72,7 +72,7 @@
  */
 #ifdef RMD160_ASM
 # if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__)
-#  define ripemd160_block_host_order ripemd160_block_asm_data_order
+#  define ripemd160_block_data_order ripemd160_block_asm_data_order
 # endif
 #endif
 
diff --git a/crypto/rsa/Makefile b/crypto/rsa/Makefile
index 7b1fd6428c91..f798d2f74932 100644
--- a/crypto/rsa/Makefile
+++ b/crypto/rsa/Makefile
@@ -19,10 +19,12 @@ APPS=
 LIB=$(TOP)/libcrypto.a
 LIBSRC= rsa_eay.c rsa_gen.c rsa_lib.c rsa_sign.c rsa_saos.c rsa_err.c \
 	rsa_pk1.c rsa_ssl.c rsa_none.c rsa_oaep.c rsa_chk.c rsa_null.c \
-	rsa_pss.c rsa_x931.c rsa_x931g.c rsa_asn1.c rsa_depr.c rsa_eng.c
+	rsa_pss.c rsa_x931.c rsa_asn1.c rsa_depr.c rsa_ameth.c rsa_prn.c \
+	rsa_pmeth.c rsa_crpt.c
 LIBOBJ= rsa_eay.o rsa_gen.o rsa_lib.o rsa_sign.o rsa_saos.o rsa_err.o \
 	rsa_pk1.o rsa_ssl.o rsa_none.o rsa_oaep.o rsa_chk.o rsa_null.o \
-	rsa_pss.o rsa_x931.o rsa_x931g.o rsa_asn1.o rsa_depr.o rsa_eng.o
+	rsa_pss.o rsa_x931.o rsa_asn1.o rsa_depr.o rsa_ameth.o rsa_prn.o \
+	rsa_pmeth.o rsa_crpt.o
 
 SRC= $(LIBSRC)
 
@@ -37,7 +39,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -78,15 +80,36 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
+rsa_ameth.o: ../../e_os.h ../../include/openssl/asn1.h
+rsa_ameth.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+rsa_ameth.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
+rsa_ameth.o: ../../include/openssl/cms.h ../../include/openssl/crypto.h
+rsa_ameth.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+rsa_ameth.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+rsa_ameth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+rsa_ameth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+rsa_ameth.o: ../../include/openssl/objects.h
+rsa_ameth.o: ../../include/openssl/opensslconf.h
+rsa_ameth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+rsa_ameth.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+rsa_ameth.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+rsa_ameth.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+rsa_ameth.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+rsa_ameth.o: ../asn1/asn1_locl.h ../cryptlib.h rsa_ameth.c
 rsa_asn1.o: ../../e_os.h ../../include/openssl/asn1.h
 rsa_asn1.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 rsa_asn1.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
 rsa_asn1.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
-rsa_asn1.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
+rsa_asn1.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+rsa_asn1.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+rsa_asn1.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+rsa_asn1.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
 rsa_asn1.o: ../../include/openssl/opensslconf.h
 rsa_asn1.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-rsa_asn1.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+rsa_asn1.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+rsa_asn1.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 rsa_asn1.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+rsa_asn1.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
 rsa_asn1.o: ../cryptlib.h rsa_asn1.c
 rsa_chk.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 rsa_chk.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
@@ -96,6 +119,21 @@ rsa_chk.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rsa_chk.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
 rsa_chk.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 rsa_chk.o: rsa_chk.c
+rsa_crpt.o: ../../e_os.h ../../include/openssl/asn1.h
+rsa_crpt.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+rsa_crpt.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+rsa_crpt.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+rsa_crpt.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+rsa_crpt.o: ../../include/openssl/engine.h ../../include/openssl/err.h
+rsa_crpt.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+rsa_crpt.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+rsa_crpt.o: ../../include/openssl/opensslconf.h
+rsa_crpt.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+rsa_crpt.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
+rsa_crpt.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+rsa_crpt.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+rsa_crpt.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+rsa_crpt.o: ../../include/openssl/x509_vfy.h ../cryptlib.h rsa_crpt.c
 rsa_depr.o: ../../e_os.h ../../include/openssl/asn1.h
 rsa_depr.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 rsa_depr.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
@@ -114,21 +152,6 @@ rsa_eay.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rsa_eay.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
 rsa_eay.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 rsa_eay.o: ../../include/openssl/symhacks.h ../cryptlib.h rsa_eay.c
-rsa_eng.o: ../../e_os.h ../../include/openssl/asn1.h
-rsa_eng.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
-rsa_eng.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-rsa_eng.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
-rsa_eng.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-rsa_eng.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-rsa_eng.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-rsa_eng.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-rsa_eng.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-rsa_eng.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-rsa_eng.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
-rsa_eng.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
-rsa_eng.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-rsa_eng.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-rsa_eng.o: ../../include/openssl/x509_vfy.h ../cryptlib.h rsa_eng.c
 rsa_err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 rsa_err.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 rsa_err.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
@@ -151,15 +174,15 @@ rsa_lib.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 rsa_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 rsa_lib.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 rsa_lib.o: ../../include/openssl/engine.h ../../include/openssl/err.h
-rsa_lib.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-rsa_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-rsa_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-rsa_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-rsa_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/rand.h
-rsa_lib.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
-rsa_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-rsa_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-rsa_lib.o: ../../include/openssl/x509_vfy.h ../cryptlib.h rsa_lib.c
+rsa_lib.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+rsa_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+rsa_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+rsa_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+rsa_lib.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
+rsa_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+rsa_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+rsa_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+rsa_lib.o: ../cryptlib.h rsa_lib.c
 rsa_none.o: ../../e_os.h ../../include/openssl/asn1.h
 rsa_none.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 rsa_none.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
@@ -182,9 +205,9 @@ rsa_oaep.o: ../../e_os.h ../../include/openssl/asn1.h
 rsa_oaep.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 rsa_oaep.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 rsa_oaep.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-rsa_oaep.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-rsa_oaep.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-rsa_oaep.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+rsa_oaep.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+rsa_oaep.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+rsa_oaep.o: ../../include/openssl/opensslconf.h
 rsa_oaep.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rsa_oaep.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
 rsa_oaep.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -199,27 +222,51 @@ rsa_pk1.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rsa_pk1.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
 rsa_pk1.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 rsa_pk1.o: ../../include/openssl/symhacks.h ../cryptlib.h rsa_pk1.c
+rsa_pmeth.o: ../../e_os.h ../../include/openssl/asn1.h
+rsa_pmeth.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+rsa_pmeth.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
+rsa_pmeth.o: ../../include/openssl/cms.h ../../include/openssl/crypto.h
+rsa_pmeth.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+rsa_pmeth.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+rsa_pmeth.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+rsa_pmeth.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+rsa_pmeth.o: ../../include/openssl/objects.h
+rsa_pmeth.o: ../../include/openssl/opensslconf.h
+rsa_pmeth.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+rsa_pmeth.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+rsa_pmeth.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+rsa_pmeth.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+rsa_pmeth.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+rsa_pmeth.o: ../cryptlib.h ../evp/evp_locl.h rsa_locl.h rsa_pmeth.c
+rsa_prn.o: ../../e_os.h ../../include/openssl/asn1.h
+rsa_prn.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+rsa_prn.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+rsa_prn.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+rsa_prn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+rsa_prn.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+rsa_prn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+rsa_prn.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+rsa_prn.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+rsa_prn.o: ../cryptlib.h rsa_prn.c
 rsa_pss.o: ../../e_os.h ../../include/openssl/asn1.h
 rsa_pss.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 rsa_pss.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 rsa_pss.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-rsa_pss.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-rsa_pss.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-rsa_pss.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-rsa_pss.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-rsa_pss.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
-rsa_pss.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-rsa_pss.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-rsa_pss.o: ../cryptlib.h rsa_pss.c
+rsa_pss.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+rsa_pss.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+rsa_pss.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+rsa_pss.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
+rsa_pss.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+rsa_pss.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+rsa_pss.o: ../../include/openssl/symhacks.h ../cryptlib.h rsa_pss.c
 rsa_saos.o: ../../e_os.h ../../include/openssl/asn1.h
 rsa_saos.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 rsa_saos.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 rsa_saos.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 rsa_saos.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 rsa_saos.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-rsa_saos.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-rsa_saos.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-rsa_saos.o: ../../include/openssl/opensslconf.h
+rsa_saos.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+rsa_saos.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 rsa_saos.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rsa_saos.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
 rsa_saos.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
@@ -232,15 +279,14 @@ rsa_sign.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 rsa_sign.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 rsa_sign.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 rsa_sign.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-rsa_sign.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-rsa_sign.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-rsa_sign.o: ../../include/openssl/opensslconf.h
+rsa_sign.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+rsa_sign.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 rsa_sign.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rsa_sign.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
 rsa_sign.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 rsa_sign.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 rsa_sign.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-rsa_sign.o: ../cryptlib.h rsa_sign.c
+rsa_sign.o: ../cryptlib.h rsa_locl.h rsa_sign.c
 rsa_ssl.o: ../../e_os.h ../../include/openssl/asn1.h
 rsa_ssl.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
 rsa_ssl.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
@@ -260,11 +306,3 @@ rsa_x931.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 rsa_x931.o: ../../include/openssl/rand.h ../../include/openssl/rsa.h
 rsa_x931.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 rsa_x931.o: ../../include/openssl/symhacks.h ../cryptlib.h rsa_x931.c
-rsa_x931g.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
-rsa_x931g.o: ../../include/openssl/bn.h ../../include/openssl/crypto.h
-rsa_x931g.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-rsa_x931g.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
-rsa_x931g.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-rsa_x931g.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
-rsa_x931g.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-rsa_x931g.o: rsa_x931g.c
diff --git a/crypto/rsa/rsa.h b/crypto/rsa/rsa.h
index 5bb932ae15ff..4814a2fc15e8 100644
--- a/crypto/rsa/rsa.h
+++ b/crypto/rsa/rsa.h
@@ -74,25 +74,6 @@
 #error RSA is disabled.
 #endif
 
-/* If this flag is set the RSA method is FIPS compliant and can be used
- * in FIPS mode. This is set in the validated module method. If an
- * application sets this flag in its own methods it is its reposibility
- * to ensure the result is compliant.
- */
-
-#define RSA_FLAG_FIPS_METHOD			0x0400
-
-/* If this flag is set the operations normally disabled in FIPS mode are
- * permitted it is then the applications responsibility to ensure that the
- * usage is compliant.
- */
-
-#define RSA_FLAG_NON_FIPS_ALLOW			0x0400
-
-#ifdef OPENSSL_FIPS
-#define FIPS_RSA_SIZE_T	int
-#endif
-
 #ifdef  __cplusplus
 extern "C" {
 #endif
@@ -136,7 +117,8 @@ struct rsa_meth_st
 		unsigned char *sigret, unsigned int *siglen, const RSA *rsa);
 	int (*rsa_verify)(int dtype,
 		const unsigned char *m, unsigned int m_length,
-		unsigned char *sigbuf, unsigned int siglen, const RSA *rsa);
+		const unsigned char *sigbuf, unsigned int siglen,
+								const RSA *rsa);
 /* If this callback is NULL, the builtin software RSA key-gen will be used. This
  * is for behavioural compatibility whilst the code gets rewired, but one day
  * it would be nice to assume there are no such things as "builtin software"
@@ -182,8 +164,6 @@ struct rsa_st
 # define OPENSSL_RSA_MAX_MODULUS_BITS	16384
 #endif
 
-#define OPENSSL_RSA_FIPS_MIN_MODULUS_BITS 1024
-
 #ifndef OPENSSL_RSA_SMALL_MODULUS_BITS
 # define OPENSSL_RSA_SMALL_MODULUS_BITS	3072
 #endif
@@ -238,11 +218,60 @@ struct rsa_st
 #endif
 
 
+#define EVP_PKEY_CTX_set_rsa_padding(ctx, pad) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_RSA, -1, EVP_PKEY_CTRL_RSA_PADDING, \
+				pad, NULL)
+
+#define EVP_PKEY_CTX_get_rsa_padding(ctx, ppad) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_RSA, -1, \
+				EVP_PKEY_CTRL_GET_RSA_PADDING, 0, ppad)
+
+#define EVP_PKEY_CTX_set_rsa_pss_saltlen(ctx, len) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_RSA, \
+				(EVP_PKEY_OP_SIGN|EVP_PKEY_OP_VERIFY), \
+				EVP_PKEY_CTRL_RSA_PSS_SALTLEN, \
+				len, NULL)
+
+#define EVP_PKEY_CTX_get_rsa_pss_saltlen(ctx, plen) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_RSA, \
+				(EVP_PKEY_OP_SIGN|EVP_PKEY_OP_VERIFY), \
+				EVP_PKEY_CTRL_GET_RSA_PSS_SALTLEN, \
+				0, plen)
+
+#define EVP_PKEY_CTX_set_rsa_keygen_bits(ctx, bits) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_RSA, EVP_PKEY_OP_KEYGEN, \
+				EVP_PKEY_CTRL_RSA_KEYGEN_BITS, bits, NULL)
+
+#define EVP_PKEY_CTX_set_rsa_keygen_pubexp(ctx, pubexp) \
+	EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_RSA, EVP_PKEY_OP_KEYGEN, \
+				EVP_PKEY_CTRL_RSA_KEYGEN_PUBEXP, 0, pubexp)
+
+#define	 EVP_PKEY_CTX_set_rsa_mgf1_md(ctx, md)	\
+		EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_RSA, EVP_PKEY_OP_TYPE_SIG,  \
+				EVP_PKEY_CTRL_RSA_MGF1_MD, 0, (void *)md)
+
+#define	 EVP_PKEY_CTX_get_rsa_mgf1_md(ctx, pmd)	\
+		EVP_PKEY_CTX_ctrl(ctx, EVP_PKEY_RSA, EVP_PKEY_OP_TYPE_SIG,  \
+				EVP_PKEY_CTRL_GET_RSA_MGF1_MD, 0, (void *)pmd)
+
+#define EVP_PKEY_CTRL_RSA_PADDING	(EVP_PKEY_ALG_CTRL + 1)
+#define EVP_PKEY_CTRL_RSA_PSS_SALTLEN	(EVP_PKEY_ALG_CTRL + 2)
+
+#define EVP_PKEY_CTRL_RSA_KEYGEN_BITS	(EVP_PKEY_ALG_CTRL + 3)
+#define EVP_PKEY_CTRL_RSA_KEYGEN_PUBEXP	(EVP_PKEY_ALG_CTRL + 4)
+#define EVP_PKEY_CTRL_RSA_MGF1_MD	(EVP_PKEY_ALG_CTRL + 5)
+
+#define EVP_PKEY_CTRL_GET_RSA_PADDING		(EVP_PKEY_ALG_CTRL + 6)
+#define EVP_PKEY_CTRL_GET_RSA_PSS_SALTLEN	(EVP_PKEY_ALG_CTRL + 7)
+#define EVP_PKEY_CTRL_GET_RSA_MGF1_MD		(EVP_PKEY_ALG_CTRL + 8)
+
 #define RSA_PKCS1_PADDING	1
 #define RSA_SSLV23_PADDING	2
 #define RSA_NO_PADDING		3
 #define RSA_PKCS1_OAEP_PADDING	4
 #define RSA_X931_PADDING	5
+/* EVP_PKEY_ only */
+#define RSA_PKCS1_PSS_PADDING	6
 
 #define RSA_PKCS1_PADDING_SIZE	11
 
@@ -261,11 +290,6 @@ RSA *	RSA_generate_key(int bits, unsigned long e,void
 
 /* New version */
 int	RSA_generate_key_ex(RSA *rsa, int bits, BIGNUM *e, BN_GENCB *cb);
-int RSA_X931_derive_ex(RSA *rsa, BIGNUM *p1, BIGNUM *p2, BIGNUM *q1, BIGNUM *q2,
-			const BIGNUM *Xp1, const BIGNUM *Xp2, const BIGNUM *Xp,
-			const BIGNUM *Xq1, const BIGNUM *Xq2, const BIGNUM *Xq,
-			const BIGNUM *e, BN_GENCB *cb);
-int RSA_X931_generate_key_ex(RSA *rsa, int bits, const BIGNUM *e, BN_GENCB *cb);
 
 int	RSA_check_key(const RSA *);
 	/* next 4 return -1 on error */
@@ -283,11 +307,6 @@ int	RSA_up_ref(RSA *r);
 
 int	RSA_flags(const RSA *r);
 
-#ifdef OPENSSL_FIPS
-RSA *FIPS_rsa_new(void);
-void FIPS_rsa_free(RSA *r);
-#endif
-
 void RSA_set_default_method(const RSA_METHOD *meth);
 const RSA_METHOD *RSA_get_default_method(void);
 const RSA_METHOD *RSA_get_method(const RSA *rsa);
@@ -304,6 +323,16 @@ const RSA_METHOD *RSA_null_method(void);
 DECLARE_ASN1_ENCODE_FUNCTIONS_const(RSA, RSAPublicKey)
 DECLARE_ASN1_ENCODE_FUNCTIONS_const(RSA, RSAPrivateKey)
 
+typedef struct rsa_pss_params_st
+	{
+	X509_ALGOR *hashAlgorithm;
+	X509_ALGOR *maskGenAlgorithm;
+	ASN1_INTEGER *saltLength;
+	ASN1_INTEGER *trailerField;
+	} RSA_PSS_PARAMS;
+
+DECLARE_ASN1_FUNCTIONS(RSA_PSS_PARAMS)
+
 #ifndef OPENSSL_NO_FP_API
 int	RSA_print_fp(FILE *fp, const RSA *r,int offset);
 #endif
@@ -333,7 +362,7 @@ RSA *d2i_Netscape_RSA(RSA **a, const unsigned char **pp, long length,
 int RSA_sign(int type, const unsigned char *m, unsigned int m_length,
 	unsigned char *sigret, unsigned int *siglen, RSA *rsa);
 int RSA_verify(int type, const unsigned char *m, unsigned int m_length,
-	unsigned char *sigbuf, unsigned int siglen, RSA *rsa);
+	const unsigned char *sigbuf, unsigned int siglen, RSA *rsa);
 
 /* The following 2 function sign and verify a ASN1_OCTET_STRING
  * object inside PKCS#1 padded RSA encryption */
@@ -384,6 +413,14 @@ int RSA_padding_add_PKCS1_PSS(RSA *rsa, unsigned char *EM,
 			const unsigned char *mHash,
 			const EVP_MD *Hash, int sLen);
 
+int RSA_verify_PKCS1_PSS_mgf1(RSA *rsa, const unsigned char *mHash,
+			const EVP_MD *Hash, const EVP_MD *mgf1Hash, 
+			const unsigned char *EM, int sLen);
+
+int RSA_padding_add_PKCS1_PSS_mgf1(RSA *rsa, unsigned char *EM,
+			const unsigned char *mHash,
+			const EVP_MD *Hash, const EVP_MD *mgf1Hash, int sLen);
+
 int RSA_get_ex_new_index(long argl, void *argp, CRYPTO_EX_new *new_func,
 	CRYPTO_EX_dup *dup_func, CRYPTO_EX_free *free_func);
 int RSA_set_ex_data(RSA *r,int idx,void *arg);
@@ -392,6 +429,25 @@ void *RSA_get_ex_data(const RSA *r, int idx);
 RSA *RSAPublicKey_dup(RSA *rsa);
 RSA *RSAPrivateKey_dup(RSA *rsa);
 
+/* If this flag is set the RSA method is FIPS compliant and can be used
+ * in FIPS mode. This is set in the validated module method. If an
+ * application sets this flag in its own methods it is its responsibility
+ * to ensure the result is compliant.
+ */
+
+#define RSA_FLAG_FIPS_METHOD			0x0400
+
+/* If this flag is set the operations normally disabled in FIPS mode are
+ * permitted it is then the applications responsibility to ensure that the
+ * usage is compliant.
+ */
+
+#define RSA_FLAG_NON_FIPS_ALLOW			0x0400
+/* Application has decided PRNG is good enough to generate a key: don't
+ * check.
+ */
+#define RSA_FLAG_CHECKED			0x0800
+
 /* BEGIN ERROR CODES */
 /* The following lines are auto generated by the script mkerr.pl. Any changes
  * made after this point may be overwritten when the script is next run.
@@ -401,9 +457,16 @@ void ERR_load_RSA_strings(void);
 /* Error codes for the RSA functions. */
 
 /* Function codes. */
-#define RSA_F_FIPS_RSA_SIGN				 140
-#define RSA_F_FIPS_RSA_VERIFY				 141
+#define RSA_F_CHECK_PADDING_MD				 140
+#define RSA_F_DO_RSA_PRINT				 146
+#define RSA_F_INT_RSA_VERIFY				 145
 #define RSA_F_MEMORY_LOCK				 100
+#define RSA_F_OLD_RSA_PRIV_DECODE			 147
+#define RSA_F_PKEY_RSA_CTRL				 143
+#define RSA_F_PKEY_RSA_CTRL_STR				 144
+#define RSA_F_PKEY_RSA_SIGN				 142
+#define RSA_F_PKEY_RSA_VERIFY				 154
+#define RSA_F_PKEY_RSA_VERIFYRECOVER			 141
 #define RSA_F_RSA_BUILTIN_KEYGEN			 129
 #define RSA_F_RSA_CHECK_KEY				 123
 #define RSA_F_RSA_EAY_PRIVATE_DECRYPT			 101
@@ -411,6 +474,8 @@ void ERR_load_RSA_strings(void);
 #define RSA_F_RSA_EAY_PUBLIC_DECRYPT			 103
 #define RSA_F_RSA_EAY_PUBLIC_ENCRYPT			 104
 #define RSA_F_RSA_GENERATE_KEY				 105
+#define RSA_F_RSA_GENERATE_KEY_EX			 155
+#define RSA_F_RSA_ITEM_VERIFY				 156
 #define RSA_F_RSA_MEMORY_LOCK				 130
 #define RSA_F_RSA_NEW_METHOD				 106
 #define RSA_F_RSA_NULL					 124
@@ -422,6 +487,7 @@ void ERR_load_RSA_strings(void);
 #define RSA_F_RSA_PADDING_ADD_NONE			 107
 #define RSA_F_RSA_PADDING_ADD_PKCS1_OAEP		 121
 #define RSA_F_RSA_PADDING_ADD_PKCS1_PSS			 125
+#define RSA_F_RSA_PADDING_ADD_PKCS1_PSS_MGF1		 148
 #define RSA_F_RSA_PADDING_ADD_PKCS1_TYPE_1		 108
 #define RSA_F_RSA_PADDING_ADD_PKCS1_TYPE_2		 109
 #define RSA_F_RSA_PADDING_ADD_SSLV23			 110
@@ -434,16 +500,20 @@ void ERR_load_RSA_strings(void);
 #define RSA_F_RSA_PADDING_CHECK_X931			 128
 #define RSA_F_RSA_PRINT					 115
 #define RSA_F_RSA_PRINT_FP				 116
-#define RSA_F_RSA_PRIVATE_ENCRYPT			 137
-#define RSA_F_RSA_PUBLIC_DECRYPT			 138
+#define RSA_F_RSA_PRIVATE_DECRYPT			 150
+#define RSA_F_RSA_PRIVATE_ENCRYPT			 151
+#define RSA_F_RSA_PRIV_DECODE				 137
+#define RSA_F_RSA_PRIV_ENCODE				 138
+#define RSA_F_RSA_PUBLIC_DECRYPT			 152
+#define RSA_F_RSA_PUBLIC_ENCRYPT			 153
+#define RSA_F_RSA_PUB_DECODE				 139
 #define RSA_F_RSA_SETUP_BLINDING			 136
-#define RSA_F_RSA_SET_DEFAULT_METHOD			 139
-#define RSA_F_RSA_SET_METHOD				 142
 #define RSA_F_RSA_SIGN					 117
 #define RSA_F_RSA_SIGN_ASN1_OCTET_STRING		 118
 #define RSA_F_RSA_VERIFY				 119
 #define RSA_F_RSA_VERIFY_ASN1_OCTET_STRING		 120
 #define RSA_F_RSA_VERIFY_PKCS1_PSS			 126
+#define RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1			 149
 
 /* Reason codes. */
 #define RSA_R_ALGORITHM_MISMATCH			 100
@@ -464,20 +534,30 @@ void ERR_load_RSA_strings(void);
 #define RSA_R_DMQ1_NOT_CONGRUENT_TO_D			 125
 #define RSA_R_D_E_NOT_CONGRUENT_TO_1			 123
 #define RSA_R_FIRST_OCTET_INVALID			 133
+#define RSA_R_ILLEGAL_OR_UNSUPPORTED_PADDING_MODE	 144
+#define RSA_R_INVALID_DIGEST_LENGTH			 143
 #define RSA_R_INVALID_HEADER				 137
+#define RSA_R_INVALID_KEYBITS				 145
 #define RSA_R_INVALID_MESSAGE_LENGTH			 131
+#define RSA_R_INVALID_MGF1_MD				 156
 #define RSA_R_INVALID_PADDING				 138
+#define RSA_R_INVALID_PADDING_MODE			 141
+#define RSA_R_INVALID_PSS_PARAMETERS			 149
+#define RSA_R_INVALID_PSS_SALTLEN			 146
+#define RSA_R_INVALID_SALT_LENGTH			 150
 #define RSA_R_INVALID_TRAILER				 139
+#define RSA_R_INVALID_X931_DIGEST			 142
 #define RSA_R_IQMP_NOT_INVERSE_OF_Q			 126
 #define RSA_R_KEY_SIZE_TOO_SMALL			 120
 #define RSA_R_LAST_OCTET_INVALID			 134
 #define RSA_R_MODULUS_TOO_LARGE				 105
-#define RSA_R_NON_FIPS_METHOD				 141
+#define RSA_R_NON_FIPS_RSA_METHOD			 157
 #define RSA_R_NO_PUBLIC_EXPONENT			 140
 #define RSA_R_NULL_BEFORE_BLOCK_MISSING			 113
 #define RSA_R_N_DOES_NOT_EQUAL_P_Q			 127
 #define RSA_R_OAEP_DECODING_ERROR			 121
-#define RSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE	 142
+#define RSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE	 158
+#define RSA_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE	 148
 #define RSA_R_PADDING_CHECK_FAILED			 114
 #define RSA_R_P_NOT_PRIME				 128
 #define RSA_R_Q_NOT_PRIME				 129
@@ -487,7 +567,13 @@ void ERR_load_RSA_strings(void);
 #define RSA_R_SSLV3_ROLLBACK_ATTACK			 115
 #define RSA_R_THE_ASN1_OBJECT_IDENTIFIER_IS_NOT_KNOWN_FOR_THIS_MD 116
 #define RSA_R_UNKNOWN_ALGORITHM_TYPE			 117
+#define RSA_R_UNKNOWN_MASK_DIGEST			 151
 #define RSA_R_UNKNOWN_PADDING_TYPE			 118
+#define RSA_R_UNKNOWN_PSS_DIGEST			 152
+#define RSA_R_UNSUPPORTED_MASK_ALGORITHM		 153
+#define RSA_R_UNSUPPORTED_MASK_PARAMETER		 154
+#define RSA_R_UNSUPPORTED_SIGNATURE_TYPE		 155
+#define RSA_R_VALUE_MISSING				 147
 #define RSA_R_WRONG_SIGNATURE_LENGTH			 119
 
 #ifdef  __cplusplus
diff --git a/crypto/rsa/rsa_ameth.c b/crypto/rsa/rsa_ameth.c
new file mode 100644
index 000000000000..2460910ab278
--- /dev/null
+++ b/crypto/rsa/rsa_ameth.c
@@ -0,0 +1,698 @@
+/* crypto/rsa/rsa_ameth.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/asn1t.h>
+#include <openssl/x509.h>
+#include <openssl/rsa.h>
+#include <openssl/bn.h>
+#ifndef OPENSSL_NO_CMS
+#include <openssl/cms.h>
+#endif
+#include "asn1_locl.h"
+
+static int rsa_pub_encode(X509_PUBKEY *pk, const EVP_PKEY *pkey)
+	{
+	unsigned char *penc = NULL;
+	int penclen;
+	penclen = i2d_RSAPublicKey(pkey->pkey.rsa, &penc);
+	if (penclen <= 0)
+		return 0;
+	if (X509_PUBKEY_set0_param(pk, OBJ_nid2obj(EVP_PKEY_RSA),
+				V_ASN1_NULL, NULL, penc, penclen))
+		return 1;
+
+	OPENSSL_free(penc);
+	return 0;
+	}
+
+static int rsa_pub_decode(EVP_PKEY *pkey, X509_PUBKEY *pubkey)
+	{
+	const unsigned char *p;
+	int pklen;
+	RSA *rsa = NULL;
+	if (!X509_PUBKEY_get0_param(NULL, &p, &pklen, NULL, pubkey))
+		return 0;
+	if (!(rsa = d2i_RSAPublicKey(NULL, &p, pklen)))
+		{
+		RSAerr(RSA_F_RSA_PUB_DECODE, ERR_R_RSA_LIB);
+		return 0;
+		}
+	EVP_PKEY_assign_RSA (pkey, rsa);
+	return 1;
+	}
+
+static int rsa_pub_cmp(const EVP_PKEY *a, const EVP_PKEY *b)
+	{
+	if (BN_cmp(b->pkey.rsa->n,a->pkey.rsa->n) != 0
+		|| BN_cmp(b->pkey.rsa->e,a->pkey.rsa->e) != 0)
+			return 0;
+	return 1;
+	}
+
+static int old_rsa_priv_decode(EVP_PKEY *pkey,
+					const unsigned char **pder, int derlen)
+	{
+	RSA *rsa;
+	if (!(rsa = d2i_RSAPrivateKey (NULL, pder, derlen)))
+		{
+		RSAerr(RSA_F_OLD_RSA_PRIV_DECODE, ERR_R_RSA_LIB);
+		return 0;
+		}
+	EVP_PKEY_assign_RSA(pkey, rsa);
+	return 1;
+	}
+
+static int old_rsa_priv_encode(const EVP_PKEY *pkey, unsigned char **pder)
+	{
+	return i2d_RSAPrivateKey(pkey->pkey.rsa, pder);
+	}
+
+static int rsa_priv_encode(PKCS8_PRIV_KEY_INFO *p8, const EVP_PKEY *pkey)
+	{
+	unsigned char *rk = NULL;
+	int rklen;
+	rklen = i2d_RSAPrivateKey(pkey->pkey.rsa, &rk);
+
+	if (rklen <= 0)
+		{
+		RSAerr(RSA_F_RSA_PRIV_ENCODE,ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+
+	if (!PKCS8_pkey_set0(p8, OBJ_nid2obj(NID_rsaEncryption), 0,
+				V_ASN1_NULL, NULL, rk, rklen))
+		{
+		RSAerr(RSA_F_RSA_PRIV_ENCODE,ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+
+	return 1;
+	}
+
+static int rsa_priv_decode(EVP_PKEY *pkey, PKCS8_PRIV_KEY_INFO *p8)
+	{
+	const unsigned char *p;
+	int pklen;
+	if (!PKCS8_pkey_get0(NULL, &p, &pklen, NULL, p8))
+		return 0;
+	return old_rsa_priv_decode(pkey, &p, pklen);
+	}
+
+static int int_rsa_size(const EVP_PKEY *pkey)
+	{
+	return RSA_size(pkey->pkey.rsa);
+	}
+
+static int rsa_bits(const EVP_PKEY *pkey)
+	{
+	return BN_num_bits(pkey->pkey.rsa->n);
+	}
+
+static void int_rsa_free(EVP_PKEY *pkey)
+	{
+	RSA_free(pkey->pkey.rsa);
+	}
+
+
+static void update_buflen(const BIGNUM *b, size_t *pbuflen)
+	{
+	size_t i;
+	if (!b)
+		return;
+	if (*pbuflen < (i = (size_t)BN_num_bytes(b)))
+			*pbuflen = i;
+	}
+
+static int do_rsa_print(BIO *bp, const RSA *x, int off, int priv)
+	{
+	char *str;
+	const char *s;
+	unsigned char *m=NULL;
+	int ret=0, mod_len = 0;
+	size_t buf_len=0;
+
+	update_buflen(x->n, &buf_len);
+	update_buflen(x->e, &buf_len);
+
+	if (priv)
+		{
+		update_buflen(x->d, &buf_len);
+		update_buflen(x->p, &buf_len);
+		update_buflen(x->q, &buf_len);
+		update_buflen(x->dmp1, &buf_len);
+		update_buflen(x->dmq1, &buf_len);
+		update_buflen(x->iqmp, &buf_len);
+		}
+
+	m=(unsigned char *)OPENSSL_malloc(buf_len+10);
+	if (m == NULL)
+		{
+		RSAerr(RSA_F_DO_RSA_PRINT,ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	if (x->n != NULL)
+		mod_len = BN_num_bits(x->n);
+
+	if(!BIO_indent(bp,off,128))
+		goto err;
+
+	if (priv && x->d)
+		{
+		if (BIO_printf(bp,"Private-Key: (%d bit)\n", mod_len)
+			<= 0) goto err;
+		str = "modulus:";
+		s = "publicExponent:";
+		}
+	else
+		{
+		if (BIO_printf(bp,"Public-Key: (%d bit)\n", mod_len)
+			<= 0) goto err;
+		str = "Modulus:";
+		s= "Exponent:";
+		}
+	if (!ASN1_bn_print(bp,str,x->n,m,off)) goto err;
+	if (!ASN1_bn_print(bp,s,x->e,m,off))
+		goto err;
+	if (priv)
+		{
+		if (!ASN1_bn_print(bp,"privateExponent:",x->d,m,off))
+			goto err;
+		if (!ASN1_bn_print(bp,"prime1:",x->p,m,off))
+			goto err;
+		if (!ASN1_bn_print(bp,"prime2:",x->q,m,off))
+			goto err;
+		if (!ASN1_bn_print(bp,"exponent1:",x->dmp1,m,off))
+			goto err;
+		if (!ASN1_bn_print(bp,"exponent2:",x->dmq1,m,off))
+			goto err;
+		if (!ASN1_bn_print(bp,"coefficient:",x->iqmp,m,off))
+			goto err;
+		}
+	ret=1;
+err:
+	if (m != NULL) OPENSSL_free(m);
+	return(ret);
+	}
+
+static int rsa_pub_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_rsa_print(bp, pkey->pkey.rsa, indent, 0);
+	}
+
+
+static int rsa_priv_print(BIO *bp, const EVP_PKEY *pkey, int indent,
+							ASN1_PCTX *ctx)
+	{
+	return do_rsa_print(bp, pkey->pkey.rsa, indent, 1);
+	}
+
+static RSA_PSS_PARAMS *rsa_pss_decode(const X509_ALGOR *alg,
+					X509_ALGOR **pmaskHash)
+	{
+	const unsigned char *p;
+	int plen;
+	RSA_PSS_PARAMS *pss;
+
+	*pmaskHash = NULL;
+
+	if (!alg->parameter || alg->parameter->type != V_ASN1_SEQUENCE)
+		return NULL;
+	p = alg->parameter->value.sequence->data;
+	plen = alg->parameter->value.sequence->length;
+	pss = d2i_RSA_PSS_PARAMS(NULL, &p, plen);
+
+	if (!pss)
+		return NULL;
+	
+	if (pss->maskGenAlgorithm)
+		{
+		ASN1_TYPE *param = pss->maskGenAlgorithm->parameter;
+		if (OBJ_obj2nid(pss->maskGenAlgorithm->algorithm) == NID_mgf1
+			&& param->type == V_ASN1_SEQUENCE)
+			{
+			p = param->value.sequence->data;
+			plen = param->value.sequence->length;
+			*pmaskHash = d2i_X509_ALGOR(NULL, &p, plen);
+			}
+		}
+
+	return pss;
+	}
+
+static int rsa_pss_param_print(BIO *bp, RSA_PSS_PARAMS *pss, 
+				X509_ALGOR *maskHash, int indent)
+	{
+	int rv = 0;
+	if (!pss)
+		{
+		if (BIO_puts(bp, " (INVALID PSS PARAMETERS)\n") <= 0)
+			return 0;
+		return 1;
+		}
+	if (BIO_puts(bp, "\n") <= 0)
+		goto err;
+	if (!BIO_indent(bp, indent, 128))
+		goto err;
+	if (BIO_puts(bp, "Hash Algorithm: ") <= 0)
+		goto err;
+
+	if (pss->hashAlgorithm)
+		{
+		if (i2a_ASN1_OBJECT(bp, pss->hashAlgorithm->algorithm) <= 0)
+			goto err;
+		}
+	else if (BIO_puts(bp, "sha1 (default)") <= 0)
+		goto err;
+
+	if (BIO_puts(bp, "\n") <= 0)
+		goto err;
+
+	if (!BIO_indent(bp, indent, 128))
+		goto err;
+
+	if (BIO_puts(bp, "Mask Algorithm: ") <= 0)
+			goto err;
+	if (pss->maskGenAlgorithm)
+		{
+		if (i2a_ASN1_OBJECT(bp, pss->maskGenAlgorithm->algorithm) <= 0)
+			goto err;
+		if (BIO_puts(bp, " with ") <= 0)
+			goto err;
+		if (maskHash)
+			{
+			if (i2a_ASN1_OBJECT(bp, maskHash->algorithm) <= 0)
+			goto err;
+			}
+		else if (BIO_puts(bp, "INVALID") <= 0)
+			goto err;
+		}
+	else if (BIO_puts(bp, "mgf1 with sha1 (default)") <= 0)
+		goto err;
+	BIO_puts(bp, "\n");
+
+	if (!BIO_indent(bp, indent, 128))
+		goto err;
+	if (BIO_puts(bp, "Salt Length: ") <= 0)
+			goto err;
+	if (pss->saltLength)
+		{
+		if (i2a_ASN1_INTEGER(bp, pss->saltLength) <= 0)
+			goto err;
+		}
+	else if (BIO_puts(bp, "20 (default)") <= 0)
+		goto err;
+	BIO_puts(bp, "\n");
+
+	if (!BIO_indent(bp, indent, 128))
+		goto err;
+	if (BIO_puts(bp, "Trailer Field: ") <= 0)
+			goto err;
+	if (pss->trailerField)
+		{
+		if (i2a_ASN1_INTEGER(bp, pss->trailerField) <= 0)
+			goto err;
+		}
+	else if (BIO_puts(bp, "0xbc (default)") <= 0)
+		goto err;
+	BIO_puts(bp, "\n");
+	
+	rv = 1;
+
+	err:
+	return rv;
+
+	}
+
+static int rsa_sig_print(BIO *bp, const X509_ALGOR *sigalg,
+					const ASN1_STRING *sig,
+					int indent, ASN1_PCTX *pctx)
+	{
+	if (OBJ_obj2nid(sigalg->algorithm) == NID_rsassaPss)
+		{
+		int rv;
+		RSA_PSS_PARAMS *pss;
+		X509_ALGOR *maskHash;
+		pss = rsa_pss_decode(sigalg, &maskHash);
+		rv = rsa_pss_param_print(bp, pss, maskHash, indent);
+		if (pss)
+			RSA_PSS_PARAMS_free(pss);
+		if (maskHash)
+			X509_ALGOR_free(maskHash);
+		if (!rv)
+			return 0;
+		}
+	else if (!sig && BIO_puts(bp, "\n") <= 0)
+		return 0;
+	if (sig)
+		return X509_signature_dump(bp, sig, indent);
+	return 1;
+	}
+
+static int rsa_pkey_ctrl(EVP_PKEY *pkey, int op, long arg1, void *arg2)
+	{
+	X509_ALGOR *alg = NULL;
+	switch (op)
+		{
+
+		case ASN1_PKEY_CTRL_PKCS7_SIGN:
+		if (arg1 == 0)
+			PKCS7_SIGNER_INFO_get0_algs(arg2, NULL, NULL, &alg);
+		break;
+
+		case ASN1_PKEY_CTRL_PKCS7_ENCRYPT:
+		if (arg1 == 0)
+			PKCS7_RECIP_INFO_get0_alg(arg2, &alg);
+		break;
+#ifndef OPENSSL_NO_CMS
+		case ASN1_PKEY_CTRL_CMS_SIGN:
+		if (arg1 == 0)
+			CMS_SignerInfo_get0_algs(arg2, NULL, NULL, NULL, &alg);
+		break;
+
+		case ASN1_PKEY_CTRL_CMS_ENVELOPE:
+		if (arg1 == 0)
+			CMS_RecipientInfo_ktri_get0_algs(arg2, NULL, NULL, &alg);
+		break;
+#endif
+
+		case ASN1_PKEY_CTRL_DEFAULT_MD_NID:
+		*(int *)arg2 = NID_sha1;
+		return 1;
+
+		default:
+		return -2;
+
+		}
+
+	if (alg)
+		X509_ALGOR_set0(alg, OBJ_nid2obj(NID_rsaEncryption),
+							V_ASN1_NULL, 0);
+
+	return 1;
+
+	}
+
+/* Customised RSA item verification routine. This is called 
+ * when a signature is encountered requiring special handling. We 
+ * currently only handle PSS.
+ */
+
+
+static int rsa_item_verify(EVP_MD_CTX *ctx, const ASN1_ITEM *it, void *asn,
+			X509_ALGOR *sigalg, ASN1_BIT_STRING *sig,
+			EVP_PKEY *pkey)
+	{
+	int rv = -1;
+	int saltlen;
+	const EVP_MD *mgf1md = NULL, *md = NULL;
+	RSA_PSS_PARAMS *pss;
+	X509_ALGOR *maskHash;
+	EVP_PKEY_CTX *pkctx;
+	/* Sanity check: make sure it is PSS */
+	if (OBJ_obj2nid(sigalg->algorithm) != NID_rsassaPss)
+		{
+		RSAerr(RSA_F_RSA_ITEM_VERIFY, RSA_R_UNSUPPORTED_SIGNATURE_TYPE);
+		return -1;
+		}
+	/* Decode PSS parameters */
+	pss = rsa_pss_decode(sigalg, &maskHash);
+
+	if (pss == NULL)
+		{
+		RSAerr(RSA_F_RSA_ITEM_VERIFY, RSA_R_INVALID_PSS_PARAMETERS);
+		goto err;
+		}
+	/* Check mask and lookup mask hash algorithm */
+	if (pss->maskGenAlgorithm)
+		{
+		if (OBJ_obj2nid(pss->maskGenAlgorithm->algorithm) != NID_mgf1)
+			{
+			RSAerr(RSA_F_RSA_ITEM_VERIFY, RSA_R_UNSUPPORTED_MASK_ALGORITHM);
+			goto err;
+			}
+		if (!maskHash)
+			{
+			RSAerr(RSA_F_RSA_ITEM_VERIFY, RSA_R_UNSUPPORTED_MASK_PARAMETER);
+			goto err;
+			}
+		mgf1md = EVP_get_digestbyobj(maskHash->algorithm);
+		if (mgf1md == NULL)
+			{
+			RSAerr(RSA_F_RSA_ITEM_VERIFY, RSA_R_UNKNOWN_MASK_DIGEST);
+			goto err;
+			}
+		}
+	else
+		mgf1md = EVP_sha1();
+
+	if (pss->hashAlgorithm)
+		{
+		md = EVP_get_digestbyobj(pss->hashAlgorithm->algorithm);
+		if (md == NULL)
+			{
+			RSAerr(RSA_F_RSA_ITEM_VERIFY, RSA_R_UNKNOWN_PSS_DIGEST);
+			goto err;
+			}
+		}
+	else
+		md = EVP_sha1();
+
+	if (pss->saltLength)
+		{
+		saltlen = ASN1_INTEGER_get(pss->saltLength);
+
+		/* Could perform more salt length sanity checks but the main
+		 * RSA routines will trap other invalid values anyway.
+		 */
+		if (saltlen < 0)
+			{
+			RSAerr(RSA_F_RSA_ITEM_VERIFY, RSA_R_INVALID_SALT_LENGTH);
+			goto err;
+			}
+		}
+	else
+		saltlen = 20;
+
+	/* low-level routines support only trailer field 0xbc (value 1)
+	 * and PKCS#1 says we should reject any other value anyway.
+	 */
+	if (pss->trailerField && ASN1_INTEGER_get(pss->trailerField) != 1)
+		{
+		RSAerr(RSA_F_RSA_ITEM_VERIFY, RSA_R_INVALID_TRAILER);
+		goto err;
+		}
+
+	/* We have all parameters now set up context */
+
+	if (!EVP_DigestVerifyInit(ctx, &pkctx, md, NULL, pkey))
+		goto err;
+
+	if (EVP_PKEY_CTX_set_rsa_padding(pkctx, RSA_PKCS1_PSS_PADDING) <= 0)
+		goto err;
+
+	if (EVP_PKEY_CTX_set_rsa_pss_saltlen(pkctx, saltlen) <= 0)
+		goto err;
+
+	if (EVP_PKEY_CTX_set_rsa_mgf1_md(pkctx, mgf1md) <= 0)
+		goto err;
+	/* Carry on */
+	rv = 2;
+
+	err:
+	RSA_PSS_PARAMS_free(pss);
+	if (maskHash)
+		X509_ALGOR_free(maskHash);
+	return rv;
+	}
+
+static int rsa_item_sign(EVP_MD_CTX *ctx, const ASN1_ITEM *it, void *asn,
+				X509_ALGOR *alg1, X509_ALGOR *alg2, 
+				ASN1_BIT_STRING *sig)
+	{
+	int pad_mode;
+	EVP_PKEY_CTX *pkctx = ctx->pctx;
+	if (EVP_PKEY_CTX_get_rsa_padding(pkctx, &pad_mode) <= 0)
+		return 0;
+	if (pad_mode == RSA_PKCS1_PADDING)
+		return 2;
+	if (pad_mode == RSA_PKCS1_PSS_PADDING)
+		{
+		const EVP_MD *sigmd, *mgf1md;
+		RSA_PSS_PARAMS *pss = NULL;
+		X509_ALGOR *mgf1alg = NULL;
+		ASN1_STRING *os1 = NULL, *os2 = NULL;
+		EVP_PKEY *pk = EVP_PKEY_CTX_get0_pkey(pkctx);
+		int saltlen, rv = 0;
+		sigmd = EVP_MD_CTX_md(ctx);
+		if (EVP_PKEY_CTX_get_rsa_mgf1_md(pkctx, &mgf1md) <= 0)
+			goto err;
+		if (!EVP_PKEY_CTX_get_rsa_pss_saltlen(pkctx, &saltlen))
+			goto err;
+		if (saltlen == -1)
+			saltlen = EVP_MD_size(sigmd);
+		else if (saltlen == -2)
+			{
+			saltlen = EVP_PKEY_size(pk) - EVP_MD_size(sigmd) - 2;
+			if (((EVP_PKEY_bits(pk) - 1) & 0x7) == 0)
+				saltlen--;
+			}
+		pss = RSA_PSS_PARAMS_new();
+		if (!pss)
+			goto err;
+		if (saltlen != 20)
+			{
+			pss->saltLength = ASN1_INTEGER_new();
+			if (!pss->saltLength)
+				goto err;
+			if (!ASN1_INTEGER_set(pss->saltLength, saltlen))
+				goto err;
+			}
+		if (EVP_MD_type(sigmd) != NID_sha1)
+			{
+			pss->hashAlgorithm = X509_ALGOR_new();
+			if (!pss->hashAlgorithm)
+				goto err;
+			X509_ALGOR_set_md(pss->hashAlgorithm, sigmd);
+			}
+		if (EVP_MD_type(mgf1md) != NID_sha1)
+			{
+			ASN1_STRING *stmp = NULL;
+			/* need to embed algorithm ID inside another */
+			mgf1alg = X509_ALGOR_new();
+			X509_ALGOR_set_md(mgf1alg, mgf1md);
+			if (!ASN1_item_pack(mgf1alg, ASN1_ITEM_rptr(X509_ALGOR),
+									&stmp))
+					goto err;
+			pss->maskGenAlgorithm = X509_ALGOR_new();
+			if (!pss->maskGenAlgorithm)
+				goto err;
+			X509_ALGOR_set0(pss->maskGenAlgorithm,
+					OBJ_nid2obj(NID_mgf1),
+					V_ASN1_SEQUENCE, stmp);
+			}
+		/* Finally create string with pss parameter encoding. */
+		if (!ASN1_item_pack(pss, ASN1_ITEM_rptr(RSA_PSS_PARAMS), &os1))
+			goto err;
+		if (alg2)
+			{
+			os2 = ASN1_STRING_dup(os1);
+			if (!os2)
+				goto err;
+			X509_ALGOR_set0(alg2, OBJ_nid2obj(NID_rsassaPss),
+						V_ASN1_SEQUENCE, os2);
+			}
+		X509_ALGOR_set0(alg1, OBJ_nid2obj(NID_rsassaPss),
+					V_ASN1_SEQUENCE, os1);
+		os1 = os2 = NULL;
+		rv = 3;
+		err:
+		if (mgf1alg)
+			X509_ALGOR_free(mgf1alg);
+		if (pss)
+			RSA_PSS_PARAMS_free(pss);
+		if (os1)
+			ASN1_STRING_free(os1);
+		return rv;
+		
+		}
+	return 2;
+	}
+
+const EVP_PKEY_ASN1_METHOD rsa_asn1_meths[] = 
+	{
+		{
+		EVP_PKEY_RSA,
+		EVP_PKEY_RSA,
+		ASN1_PKEY_SIGPARAM_NULL,
+
+		"RSA",
+		"OpenSSL RSA method",
+
+		rsa_pub_decode,
+		rsa_pub_encode,
+		rsa_pub_cmp,
+		rsa_pub_print,
+
+		rsa_priv_decode,
+		rsa_priv_encode,
+		rsa_priv_print,
+
+		int_rsa_size,
+		rsa_bits,
+
+		0,0,0,0,0,0,
+
+		rsa_sig_print,
+		int_rsa_free,
+		rsa_pkey_ctrl,
+		old_rsa_priv_decode,
+		old_rsa_priv_encode,
+		rsa_item_verify,
+		rsa_item_sign
+		},
+
+		{
+		EVP_PKEY_RSA2,
+		EVP_PKEY_RSA,
+		ASN1_PKEY_ALIAS
+		}
+	};
diff --git a/crypto/rsa/rsa_asn1.c b/crypto/rsa/rsa_asn1.c
index 6e8a803e8145..6ed5de3db4d4 100644
--- a/crypto/rsa/rsa_asn1.c
+++ b/crypto/rsa/rsa_asn1.c
@@ -3,7 +3,7 @@
  * project 2000.
  */
 /* ====================================================================
- * Copyright (c) 2000 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 2000-2005 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -60,21 +60,12 @@
 #include "cryptlib.h"
 #include <openssl/bn.h>
 #include <openssl/rsa.h>
+#include <openssl/x509.h>
 #include <openssl/asn1t.h>
 
-static ASN1_METHOD method={
-        (I2D_OF(void))     i2d_RSAPrivateKey,
-        (D2I_OF(void))     d2i_RSAPrivateKey,
-        (void *(*)(void))  RSA_new,
-        (void (*)(void *)) RSA_free};
-
-ASN1_METHOD *RSAPrivateKey_asn1_meth(void)
-	{
-	return(&method);
-	}
-
 /* Override the default free and new methods */
-static int rsa_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it)
+static int rsa_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+								void *exarg)
 {
 	if(operation == ASN1_OP_NEW_PRE) {
 		*pval = (ASN1_VALUE *)RSA_new();
@@ -106,6 +97,15 @@ ASN1_SEQUENCE_cb(RSAPublicKey, rsa_cb) = {
 	ASN1_SIMPLE(RSA, e, BIGNUM),
 } ASN1_SEQUENCE_END_cb(RSA, RSAPublicKey)
 
+ASN1_SEQUENCE(RSA_PSS_PARAMS) = {
+	ASN1_EXP_OPT(RSA_PSS_PARAMS, hashAlgorithm, X509_ALGOR,0),
+	ASN1_EXP_OPT(RSA_PSS_PARAMS, maskGenAlgorithm, X509_ALGOR,1),
+	ASN1_EXP_OPT(RSA_PSS_PARAMS, saltLength, ASN1_INTEGER,2),
+	ASN1_EXP_OPT(RSA_PSS_PARAMS, trailerField, ASN1_INTEGER,3)
+} ASN1_SEQUENCE_END(RSA_PSS_PARAMS)
+
+IMPLEMENT_ASN1_FUNCTIONS(RSA_PSS_PARAMS)
+
 IMPLEMENT_ASN1_ENCODE_FUNCTIONS_const_fname(RSA, RSAPrivateKey, RSAPrivateKey)
 
 IMPLEMENT_ASN1_ENCODE_FUNCTIONS_const_fname(RSA, RSAPublicKey, RSAPublicKey)
diff --git a/crypto/rsa/rsa_crpt.c b/crypto/rsa/rsa_crpt.c
new file mode 100644
index 000000000000..d3e44785dcfb
--- /dev/null
+++ b/crypto/rsa/rsa_crpt.c
@@ -0,0 +1,257 @@
+/* crypto/rsa/rsa_lib.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <stdio.h>
+#include <openssl/crypto.h>
+#include "cryptlib.h"
+#include <openssl/lhash.h>
+#include <openssl/bn.h>
+#include <openssl/rsa.h>
+#include <openssl/rand.h>
+#ifndef OPENSSL_NO_ENGINE
+#include <openssl/engine.h>
+#endif
+
+int RSA_size(const RSA *r)
+	{
+	return(BN_num_bytes(r->n));
+	}
+
+int RSA_public_encrypt(int flen, const unsigned char *from, unsigned char *to,
+	     RSA *rsa, int padding)
+	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(rsa->meth->flags & RSA_FLAG_FIPS_METHOD)
+			&& !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
+		{
+		RSAerr(RSA_F_RSA_PUBLIC_ENCRYPT, RSA_R_NON_FIPS_RSA_METHOD);
+		return -1;
+		}
+#endif
+	return(rsa->meth->rsa_pub_enc(flen, from, to, rsa, padding));
+	}
+
+int RSA_private_encrypt(int flen, const unsigned char *from, unsigned char *to,
+	     RSA *rsa, int padding)
+	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(rsa->meth->flags & RSA_FLAG_FIPS_METHOD)
+			&& !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
+		{
+		RSAerr(RSA_F_RSA_PRIVATE_ENCRYPT, RSA_R_NON_FIPS_RSA_METHOD);
+		return -1;
+		}
+#endif
+	return(rsa->meth->rsa_priv_enc(flen, from, to, rsa, padding));
+	}
+
+int RSA_private_decrypt(int flen, const unsigned char *from, unsigned char *to,
+	     RSA *rsa, int padding)
+	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(rsa->meth->flags & RSA_FLAG_FIPS_METHOD)
+			&& !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
+		{
+		RSAerr(RSA_F_RSA_PRIVATE_DECRYPT, RSA_R_NON_FIPS_RSA_METHOD);
+		return -1;
+		}
+#endif
+	return(rsa->meth->rsa_priv_dec(flen, from, to, rsa, padding));
+	}
+
+int RSA_public_decrypt(int flen, const unsigned char *from, unsigned char *to,
+	     RSA *rsa, int padding)
+	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(rsa->meth->flags & RSA_FLAG_FIPS_METHOD)
+			&& !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
+		{
+		RSAerr(RSA_F_RSA_PUBLIC_DECRYPT, RSA_R_NON_FIPS_RSA_METHOD);
+		return -1;
+		}
+#endif
+	return(rsa->meth->rsa_pub_dec(flen, from, to, rsa, padding));
+	}
+
+int RSA_flags(const RSA *r)
+	{
+	return((r == NULL)?0:r->meth->flags);
+	}
+
+void RSA_blinding_off(RSA *rsa)
+	{
+	if (rsa->blinding != NULL)
+		{
+		BN_BLINDING_free(rsa->blinding);
+		rsa->blinding=NULL;
+		}
+	rsa->flags &= ~RSA_FLAG_BLINDING;
+	rsa->flags |= RSA_FLAG_NO_BLINDING;
+	}
+
+int RSA_blinding_on(RSA *rsa, BN_CTX *ctx)
+	{
+	int ret=0;
+
+	if (rsa->blinding != NULL)
+		RSA_blinding_off(rsa);
+
+	rsa->blinding = RSA_setup_blinding(rsa, ctx);
+	if (rsa->blinding == NULL)
+		goto err;
+
+	rsa->flags |= RSA_FLAG_BLINDING;
+	rsa->flags &= ~RSA_FLAG_NO_BLINDING;
+	ret=1;
+err:
+	return(ret);
+	}
+
+static BIGNUM *rsa_get_public_exp(const BIGNUM *d, const BIGNUM *p,
+	const BIGNUM *q, BN_CTX *ctx)
+{
+	BIGNUM *ret = NULL, *r0, *r1, *r2;
+
+	if (d == NULL || p == NULL || q == NULL)
+		return NULL;
+
+	BN_CTX_start(ctx);
+	r0 = BN_CTX_get(ctx);
+	r1 = BN_CTX_get(ctx);
+	r2 = BN_CTX_get(ctx);
+	if (r2 == NULL)
+		goto err;
+
+	if (!BN_sub(r1, p, BN_value_one())) goto err;
+	if (!BN_sub(r2, q, BN_value_one())) goto err;
+	if (!BN_mul(r0, r1, r2, ctx)) goto err;
+
+	ret = BN_mod_inverse(NULL, d, r0, ctx);
+err:
+	BN_CTX_end(ctx);
+	return ret;
+}
+
+BN_BLINDING *RSA_setup_blinding(RSA *rsa, BN_CTX *in_ctx)
+{
+	BIGNUM local_n;
+	BIGNUM *e,*n;
+	BN_CTX *ctx;
+	BN_BLINDING *ret = NULL;
+
+	if (in_ctx == NULL)
+		{
+		if ((ctx = BN_CTX_new()) == NULL) return 0;
+		}
+	else
+		ctx = in_ctx;
+
+	BN_CTX_start(ctx);
+	e  = BN_CTX_get(ctx);
+	if (e == NULL)
+		{
+		RSAerr(RSA_F_RSA_SETUP_BLINDING, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	if (rsa->e == NULL)
+		{
+		e = rsa_get_public_exp(rsa->d, rsa->p, rsa->q, ctx);
+		if (e == NULL)
+			{
+			RSAerr(RSA_F_RSA_SETUP_BLINDING, RSA_R_NO_PUBLIC_EXPONENT);
+			goto err;
+			}
+		}
+	else
+		e = rsa->e;
+
+	
+	if ((RAND_status() == 0) && rsa->d != NULL && rsa->d->d != NULL)
+		{
+		/* if PRNG is not properly seeded, resort to secret
+		 * exponent as unpredictable seed */
+		RAND_add(rsa->d->d, rsa->d->dmax * sizeof rsa->d->d[0], 0.0);
+		}
+
+	if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+		{
+		/* Set BN_FLG_CONSTTIME flag */
+		n = &local_n;
+		BN_with_flags(n, rsa->n, BN_FLG_CONSTTIME);
+		}
+	else
+		n = rsa->n;
+
+	ret = BN_BLINDING_create_param(NULL, e, n, ctx,
+			rsa->meth->bn_mod_exp, rsa->_method_mod_n);
+	if (ret == NULL)
+		{
+		RSAerr(RSA_F_RSA_SETUP_BLINDING, ERR_R_BN_LIB);
+		goto err;
+		}
+	CRYPTO_THREADID_current(BN_BLINDING_thread_id(ret));
+err:
+	BN_CTX_end(ctx);
+	if (in_ctx == NULL)
+		BN_CTX_free(ctx);
+	if(rsa->e == NULL)
+		BN_free(e);
+
+	return ret;
+}
diff --git a/crypto/rsa/rsa_eay.c b/crypto/rsa/rsa_eay.c
index d477f08c4cac..2e1ddd48d35b 100644
--- a/crypto/rsa/rsa_eay.c
+++ b/crypto/rsa/rsa_eay.c
@@ -115,7 +115,7 @@
 #include <openssl/rsa.h>
 #include <openssl/rand.h>
 
-#if !defined(RSA_NULL) && !defined(OPENSSL_FIPS)
+#ifndef RSA_NULL
 
 static int RSA_eay_public_encrypt(int flen, const unsigned char *from,
 		unsigned char *to, RSA *rsa,int padding);
@@ -256,6 +256,7 @@ static BN_BLINDING *rsa_get_blinding(RSA *rsa, int *local, BN_CTX *ctx)
 {
 	BN_BLINDING *ret;
 	int got_write_lock = 0;
+	CRYPTO_THREADID cur;
 
 	CRYPTO_r_lock(CRYPTO_LOCK_RSA);
 
@@ -273,7 +274,8 @@ static BN_BLINDING *rsa_get_blinding(RSA *rsa, int *local, BN_CTX *ctx)
 	if (ret == NULL)
 		goto err;
 
-	if (BN_BLINDING_get_thread_id(ret) == CRYPTO_thread_id())
+	CRYPTO_THREADID_current(&cur);
+	if (!CRYPTO_THREADID_cmp(&cur, BN_BLINDING_thread_id(ret)))
 		{
 		/* rsa->blinding is ours! */
 
diff --git a/crypto/rsa/rsa_eng.c b/crypto/rsa/rsa_eng.c
deleted file mode 100644
index 2f21ddbe74c5..000000000000
--- a/crypto/rsa/rsa_eng.c
+++ /dev/null
@@ -1,357 +0,0 @@
-/* crypto/rsa/rsa_lib.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#include <stdio.h>
-#include <openssl/crypto.h>
-#include "cryptlib.h"
-#include <openssl/lhash.h>
-#include <openssl/bn.h>
-#include <openssl/rsa.h>
-#include <openssl/rand.h>
-#ifndef OPENSSL_NO_ENGINE
-#include <openssl/engine.h>
-#endif
-
-const char RSA_version[]="RSA" OPENSSL_VERSION_PTEXT;
-
-static const RSA_METHOD *default_RSA_meth=NULL;
-
-RSA *RSA_new(void)
-	{
-	RSA *r=RSA_new_method(NULL);
-
-	return r;
-	}
-
-void RSA_set_default_method(const RSA_METHOD *meth)
-	{
-#ifdef OPENSSL_FIPS
-	if (FIPS_mode() && !(meth->flags & RSA_FLAG_FIPS_METHOD))
-		{
-		RSAerr(RSA_F_RSA_SET_DEFAULT_METHOD, RSA_R_NON_FIPS_METHOD);
-		return;
-		}
-#endif
-	default_RSA_meth = meth;
-	}
-
-const RSA_METHOD *RSA_get_default_method(void)
-	{
-	if (default_RSA_meth == NULL)
-		{
-#ifdef RSA_NULL
-		default_RSA_meth=RSA_null_method();
-#else
-#if 0 /* was: #ifdef RSAref */
-		default_RSA_meth=RSA_PKCS1_RSAref();
-#else
-		default_RSA_meth=RSA_PKCS1_SSLeay();
-#endif
-#endif
-		}
-
-	return default_RSA_meth;
-	}
-
-const RSA_METHOD *RSA_get_method(const RSA *rsa)
-	{
-	return rsa->meth;
-	}
-
-int RSA_set_method(RSA *rsa, const RSA_METHOD *meth)
-	{
-	/* NB: The caller is specifically setting a method, so it's not up to us
-	 * to deal with which ENGINE it comes from. */
-	const RSA_METHOD *mtmp;
-#ifdef OPENSSL_FIPS
-	if (FIPS_mode() && !(meth->flags & RSA_FLAG_FIPS_METHOD))
-		{
-		RSAerr(RSA_F_RSA_SET_METHOD, RSA_R_NON_FIPS_METHOD);
-		return 0;
-		}
-#endif
-	mtmp = rsa->meth;
-	if (mtmp->finish) mtmp->finish(rsa);
-#ifndef OPENSSL_NO_ENGINE
-	if (rsa->engine)
-		{
-		ENGINE_finish(rsa->engine);
-		rsa->engine = NULL;
-		}
-#endif
-	rsa->meth = meth;
-	if (meth->init) meth->init(rsa);
-	return 1;
-	}
-
-RSA *RSA_new_method(ENGINE *engine)
-	{
-	RSA *ret;
-
-	ret=(RSA *)OPENSSL_malloc(sizeof(RSA));
-	if (ret == NULL)
-		{
-		RSAerr(RSA_F_RSA_NEW_METHOD,ERR_R_MALLOC_FAILURE);
-		return NULL;
-		}
-
-	ret->meth = RSA_get_default_method();
-#ifndef OPENSSL_NO_ENGINE
-	if (engine)
-		{
-		if (!ENGINE_init(engine))
-			{
-			RSAerr(RSA_F_RSA_NEW_METHOD, ERR_R_ENGINE_LIB);
-			OPENSSL_free(ret);
-			return NULL;
-			}
-		ret->engine = engine;
-		}
-	else
-		ret->engine = ENGINE_get_default_RSA();
-	if(ret->engine)
-		{
-		ret->meth = ENGINE_get_RSA(ret->engine);
-		if(!ret->meth)
-			{
-			RSAerr(RSA_F_RSA_NEW_METHOD,
-				ERR_R_ENGINE_LIB);
-			ENGINE_finish(ret->engine);
-			OPENSSL_free(ret);
-			return NULL;
-			}
-		}
-#endif
-#ifdef OPENSSL_FIPS
-	if (FIPS_mode() && !(ret->meth->flags & RSA_FLAG_FIPS_METHOD))
-		{
-		RSAerr(RSA_F_RSA_NEW_METHOD, RSA_R_NON_FIPS_METHOD);
-#ifndef OPENSSL_NO_ENGINE
-		if (ret->engine)
-			ENGINE_finish(ret->engine);
-#endif
-		OPENSSL_free(ret);
-		return NULL;
-		}
-#endif
-
-	ret->pad=0;
-	ret->version=0;
-	ret->n=NULL;
-	ret->e=NULL;
-	ret->d=NULL;
-	ret->p=NULL;
-	ret->q=NULL;
-	ret->dmp1=NULL;
-	ret->dmq1=NULL;
-	ret->iqmp=NULL;
-	ret->references=1;
-	ret->_method_mod_n=NULL;
-	ret->_method_mod_p=NULL;
-	ret->_method_mod_q=NULL;
-	ret->blinding=NULL;
-	ret->mt_blinding=NULL;
-	ret->bignum_data=NULL;
-	ret->flags=ret->meth->flags & ~RSA_FLAG_NON_FIPS_ALLOW;
-	if (!CRYPTO_new_ex_data(CRYPTO_EX_INDEX_RSA, ret, &ret->ex_data))
-		{
-#ifndef OPENSSL_NO_ENGINE
-	if (ret->engine)
-		ENGINE_finish(ret->engine);
-#endif
-		OPENSSL_free(ret);
-		return(NULL);
-		}
-
-	if ((ret->meth->init != NULL) && !ret->meth->init(ret))
-		{
-#ifndef OPENSSL_NO_ENGINE
-		if (ret->engine)
-			ENGINE_finish(ret->engine);
-#endif
-		CRYPTO_free_ex_data(CRYPTO_EX_INDEX_RSA, ret, &ret->ex_data);
-		OPENSSL_free(ret);
-		ret=NULL;
-		}
-	return(ret);
-	}
-
-void RSA_free(RSA *r)
-	{
-	int i;
-
-	if (r == NULL) return;
-
-	i=CRYPTO_add(&r->references,-1,CRYPTO_LOCK_RSA);
-#ifdef REF_PRINT
-	REF_PRINT("RSA",r);
-#endif
-	if (i > 0) return;
-#ifdef REF_CHECK
-	if (i < 0)
-		{
-		fprintf(stderr,"RSA_free, bad reference count\n");
-		abort();
-		}
-#endif
-
-	if (r->meth->finish)
-		r->meth->finish(r);
-#ifndef OPENSSL_NO_ENGINE
-	if (r->engine)
-		ENGINE_finish(r->engine);
-#endif
-
-	CRYPTO_free_ex_data(CRYPTO_EX_INDEX_RSA, r, &r->ex_data);
-
-	if (r->n != NULL) BN_clear_free(r->n);
-	if (r->e != NULL) BN_clear_free(r->e);
-	if (r->d != NULL) BN_clear_free(r->d);
-	if (r->p != NULL) BN_clear_free(r->p);
-	if (r->q != NULL) BN_clear_free(r->q);
-	if (r->dmp1 != NULL) BN_clear_free(r->dmp1);
-	if (r->dmq1 != NULL) BN_clear_free(r->dmq1);
-	if (r->iqmp != NULL) BN_clear_free(r->iqmp);
-	if (r->blinding != NULL) BN_BLINDING_free(r->blinding);
-	if (r->mt_blinding != NULL) BN_BLINDING_free(r->mt_blinding);
-	if (r->bignum_data != NULL) OPENSSL_free_locked(r->bignum_data);
-	OPENSSL_free(r);
-	}
-
-int RSA_up_ref(RSA *r)
-	{
-	int i = CRYPTO_add(&r->references, 1, CRYPTO_LOCK_RSA);
-#ifdef REF_PRINT
-	REF_PRINT("RSA",r);
-#endif
-#ifdef REF_CHECK
-	if (i < 2)
-		{
-		fprintf(stderr, "RSA_up_ref, bad reference count\n");
-		abort();
-		}
-#endif
-	return ((i > 1) ? 1 : 0);
-	}
-
-int RSA_get_ex_new_index(long argl, void *argp, CRYPTO_EX_new *new_func,
-	     CRYPTO_EX_dup *dup_func, CRYPTO_EX_free *free_func)
-        {
-	return CRYPTO_get_ex_new_index(CRYPTO_EX_INDEX_RSA, argl, argp,
-				new_func, dup_func, free_func);
-        }
-
-int RSA_set_ex_data(RSA *r, int idx, void *arg)
-	{
-	return(CRYPTO_set_ex_data(&r->ex_data,idx,arg));
-	}
-
-void *RSA_get_ex_data(const RSA *r, int idx)
-	{
-	return(CRYPTO_get_ex_data(&r->ex_data,idx));
-	}
-
-int RSA_flags(const RSA *r)
-	{
-	return((r == NULL)?0:r->meth->flags);
-	}
-
-int RSA_memory_lock(RSA *r)
-	{
-	int i,j,k,off;
-	char *p;
-	BIGNUM *bn,**t[6],*b;
-	BN_ULONG *ul;
-
-	if (r->d == NULL) return(1);
-	t[0]= &r->d;
-	t[1]= &r->p;
-	t[2]= &r->q;
-	t[3]= &r->dmp1;
-	t[4]= &r->dmq1;
-	t[5]= &r->iqmp;
-	k=sizeof(BIGNUM)*6;
-	off=k/sizeof(BN_ULONG)+1;
-	j=1;
-	for (i=0; i<6; i++)
-		j+= (*t[i])->top;
-	if ((p=OPENSSL_malloc_locked((off+j)*sizeof(BN_ULONG))) == NULL)
-		{
-		RSAerr(RSA_F_RSA_MEMORY_LOCK,ERR_R_MALLOC_FAILURE);
-		return(0);
-		}
-	bn=(BIGNUM *)p;
-	ul=(BN_ULONG *)&(p[off]);
-	for (i=0; i<6; i++)
-		{
-		b= *(t[i]);
-		*(t[i])= &(bn[i]);
-		memcpy((char *)&(bn[i]),(char *)b,sizeof(BIGNUM));
-		bn[i].flags=BN_FLG_STATIC_DATA;
-		bn[i].d=ul;
-		memcpy((char *)ul,b->d,sizeof(BN_ULONG)*b->top);
-		ul+=b->top;
-		BN_clear_free(b);
-		}
-	
-	/* I should fix this so it can still be done */
-	r->flags&= ~(RSA_FLAG_CACHE_PRIVATE|RSA_FLAG_CACHE_PUBLIC);
-
-	r->bignum_data=p;
-	return(1);
-	}
diff --git a/crypto/rsa/rsa_err.c b/crypto/rsa/rsa_err.c
index 501f5ea38942..46e0bf99806c 100644
--- a/crypto/rsa/rsa_err.c
+++ b/crypto/rsa/rsa_err.c
@@ -1,6 +1,6 @@
 /* crypto/rsa/rsa_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,9 +70,16 @@
 
 static ERR_STRING_DATA RSA_str_functs[]=
 	{
-{ERR_FUNC(RSA_F_FIPS_RSA_SIGN),	"FIPS_RSA_SIGN"},
-{ERR_FUNC(RSA_F_FIPS_RSA_VERIFY),	"FIPS_RSA_VERIFY"},
+{ERR_FUNC(RSA_F_CHECK_PADDING_MD),	"CHECK_PADDING_MD"},
+{ERR_FUNC(RSA_F_DO_RSA_PRINT),	"DO_RSA_PRINT"},
+{ERR_FUNC(RSA_F_INT_RSA_VERIFY),	"INT_RSA_VERIFY"},
 {ERR_FUNC(RSA_F_MEMORY_LOCK),	"MEMORY_LOCK"},
+{ERR_FUNC(RSA_F_OLD_RSA_PRIV_DECODE),	"OLD_RSA_PRIV_DECODE"},
+{ERR_FUNC(RSA_F_PKEY_RSA_CTRL),	"PKEY_RSA_CTRL"},
+{ERR_FUNC(RSA_F_PKEY_RSA_CTRL_STR),	"PKEY_RSA_CTRL_STR"},
+{ERR_FUNC(RSA_F_PKEY_RSA_SIGN),	"PKEY_RSA_SIGN"},
+{ERR_FUNC(RSA_F_PKEY_RSA_VERIFY),	"PKEY_RSA_VERIFY"},
+{ERR_FUNC(RSA_F_PKEY_RSA_VERIFYRECOVER),	"PKEY_RSA_VERIFYRECOVER"},
 {ERR_FUNC(RSA_F_RSA_BUILTIN_KEYGEN),	"RSA_BUILTIN_KEYGEN"},
 {ERR_FUNC(RSA_F_RSA_CHECK_KEY),	"RSA_check_key"},
 {ERR_FUNC(RSA_F_RSA_EAY_PRIVATE_DECRYPT),	"RSA_EAY_PRIVATE_DECRYPT"},
@@ -80,6 +87,8 @@ static ERR_STRING_DATA RSA_str_functs[]=
 {ERR_FUNC(RSA_F_RSA_EAY_PUBLIC_DECRYPT),	"RSA_EAY_PUBLIC_DECRYPT"},
 {ERR_FUNC(RSA_F_RSA_EAY_PUBLIC_ENCRYPT),	"RSA_EAY_PUBLIC_ENCRYPT"},
 {ERR_FUNC(RSA_F_RSA_GENERATE_KEY),	"RSA_generate_key"},
+{ERR_FUNC(RSA_F_RSA_GENERATE_KEY_EX),	"RSA_generate_key_ex"},
+{ERR_FUNC(RSA_F_RSA_ITEM_VERIFY),	"RSA_ITEM_VERIFY"},
 {ERR_FUNC(RSA_F_RSA_MEMORY_LOCK),	"RSA_memory_lock"},
 {ERR_FUNC(RSA_F_RSA_NEW_METHOD),	"RSA_new_method"},
 {ERR_FUNC(RSA_F_RSA_NULL),	"RSA_NULL"},
@@ -91,6 +100,7 @@ static ERR_STRING_DATA RSA_str_functs[]=
 {ERR_FUNC(RSA_F_RSA_PADDING_ADD_NONE),	"RSA_padding_add_none"},
 {ERR_FUNC(RSA_F_RSA_PADDING_ADD_PKCS1_OAEP),	"RSA_padding_add_PKCS1_OAEP"},
 {ERR_FUNC(RSA_F_RSA_PADDING_ADD_PKCS1_PSS),	"RSA_padding_add_PKCS1_PSS"},
+{ERR_FUNC(RSA_F_RSA_PADDING_ADD_PKCS1_PSS_MGF1),	"RSA_padding_add_PKCS1_PSS_mgf1"},
 {ERR_FUNC(RSA_F_RSA_PADDING_ADD_PKCS1_TYPE_1),	"RSA_padding_add_PKCS1_type_1"},
 {ERR_FUNC(RSA_F_RSA_PADDING_ADD_PKCS1_TYPE_2),	"RSA_padding_add_PKCS1_type_2"},
 {ERR_FUNC(RSA_F_RSA_PADDING_ADD_SSLV23),	"RSA_padding_add_SSLv23"},
@@ -103,16 +113,20 @@ static ERR_STRING_DATA RSA_str_functs[]=
 {ERR_FUNC(RSA_F_RSA_PADDING_CHECK_X931),	"RSA_padding_check_X931"},
 {ERR_FUNC(RSA_F_RSA_PRINT),	"RSA_print"},
 {ERR_FUNC(RSA_F_RSA_PRINT_FP),	"RSA_print_fp"},
+{ERR_FUNC(RSA_F_RSA_PRIVATE_DECRYPT),	"RSA_private_decrypt"},
 {ERR_FUNC(RSA_F_RSA_PRIVATE_ENCRYPT),	"RSA_private_encrypt"},
+{ERR_FUNC(RSA_F_RSA_PRIV_DECODE),	"RSA_PRIV_DECODE"},
+{ERR_FUNC(RSA_F_RSA_PRIV_ENCODE),	"RSA_PRIV_ENCODE"},
 {ERR_FUNC(RSA_F_RSA_PUBLIC_DECRYPT),	"RSA_public_decrypt"},
+{ERR_FUNC(RSA_F_RSA_PUBLIC_ENCRYPT),	"RSA_public_encrypt"},
+{ERR_FUNC(RSA_F_RSA_PUB_DECODE),	"RSA_PUB_DECODE"},
 {ERR_FUNC(RSA_F_RSA_SETUP_BLINDING),	"RSA_setup_blinding"},
-{ERR_FUNC(RSA_F_RSA_SET_DEFAULT_METHOD),	"RSA_set_default_method"},
-{ERR_FUNC(RSA_F_RSA_SET_METHOD),	"RSA_set_method"},
 {ERR_FUNC(RSA_F_RSA_SIGN),	"RSA_sign"},
 {ERR_FUNC(RSA_F_RSA_SIGN_ASN1_OCTET_STRING),	"RSA_sign_ASN1_OCTET_STRING"},
 {ERR_FUNC(RSA_F_RSA_VERIFY),	"RSA_verify"},
 {ERR_FUNC(RSA_F_RSA_VERIFY_ASN1_OCTET_STRING),	"RSA_verify_ASN1_OCTET_STRING"},
 {ERR_FUNC(RSA_F_RSA_VERIFY_PKCS1_PSS),	"RSA_verify_PKCS1_PSS"},
+{ERR_FUNC(RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1),	"RSA_verify_PKCS1_PSS_mgf1"},
 {0,NULL}
 	};
 
@@ -136,20 +150,30 @@ static ERR_STRING_DATA RSA_str_reasons[]=
 {ERR_REASON(RSA_R_DMQ1_NOT_CONGRUENT_TO_D),"dmq1 not congruent to d"},
 {ERR_REASON(RSA_R_D_E_NOT_CONGRUENT_TO_1),"d e not congruent to 1"},
 {ERR_REASON(RSA_R_FIRST_OCTET_INVALID)   ,"first octet invalid"},
+{ERR_REASON(RSA_R_ILLEGAL_OR_UNSUPPORTED_PADDING_MODE),"illegal or unsupported padding mode"},
+{ERR_REASON(RSA_R_INVALID_DIGEST_LENGTH) ,"invalid digest length"},
 {ERR_REASON(RSA_R_INVALID_HEADER)        ,"invalid header"},
+{ERR_REASON(RSA_R_INVALID_KEYBITS)       ,"invalid keybits"},
 {ERR_REASON(RSA_R_INVALID_MESSAGE_LENGTH),"invalid message length"},
+{ERR_REASON(RSA_R_INVALID_MGF1_MD)       ,"invalid mgf1 md"},
 {ERR_REASON(RSA_R_INVALID_PADDING)       ,"invalid padding"},
+{ERR_REASON(RSA_R_INVALID_PADDING_MODE)  ,"invalid padding mode"},
+{ERR_REASON(RSA_R_INVALID_PSS_PARAMETERS),"invalid pss parameters"},
+{ERR_REASON(RSA_R_INVALID_PSS_SALTLEN)   ,"invalid pss saltlen"},
+{ERR_REASON(RSA_R_INVALID_SALT_LENGTH)   ,"invalid salt length"},
 {ERR_REASON(RSA_R_INVALID_TRAILER)       ,"invalid trailer"},
+{ERR_REASON(RSA_R_INVALID_X931_DIGEST)   ,"invalid x931 digest"},
 {ERR_REASON(RSA_R_IQMP_NOT_INVERSE_OF_Q) ,"iqmp not inverse of q"},
 {ERR_REASON(RSA_R_KEY_SIZE_TOO_SMALL)    ,"key size too small"},
 {ERR_REASON(RSA_R_LAST_OCTET_INVALID)    ,"last octet invalid"},
 {ERR_REASON(RSA_R_MODULUS_TOO_LARGE)     ,"modulus too large"},
-{ERR_REASON(RSA_R_NON_FIPS_METHOD)       ,"non fips method"},
+{ERR_REASON(RSA_R_NON_FIPS_RSA_METHOD)   ,"non fips rsa method"},
 {ERR_REASON(RSA_R_NO_PUBLIC_EXPONENT)    ,"no public exponent"},
 {ERR_REASON(RSA_R_NULL_BEFORE_BLOCK_MISSING),"null before block missing"},
 {ERR_REASON(RSA_R_N_DOES_NOT_EQUAL_P_Q)  ,"n does not equal p q"},
 {ERR_REASON(RSA_R_OAEP_DECODING_ERROR)   ,"oaep decoding error"},
 {ERR_REASON(RSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE),"operation not allowed in fips mode"},
+{ERR_REASON(RSA_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE),"operation not supported for this keytype"},
 {ERR_REASON(RSA_R_PADDING_CHECK_FAILED)  ,"padding check failed"},
 {ERR_REASON(RSA_R_P_NOT_PRIME)           ,"p not prime"},
 {ERR_REASON(RSA_R_Q_NOT_PRIME)           ,"q not prime"},
@@ -159,7 +183,13 @@ static ERR_STRING_DATA RSA_str_reasons[]=
 {ERR_REASON(RSA_R_SSLV3_ROLLBACK_ATTACK) ,"sslv3 rollback attack"},
 {ERR_REASON(RSA_R_THE_ASN1_OBJECT_IDENTIFIER_IS_NOT_KNOWN_FOR_THIS_MD),"the asn1 object identifier is not known for this md"},
 {ERR_REASON(RSA_R_UNKNOWN_ALGORITHM_TYPE),"unknown algorithm type"},
+{ERR_REASON(RSA_R_UNKNOWN_MASK_DIGEST)   ,"unknown mask digest"},
 {ERR_REASON(RSA_R_UNKNOWN_PADDING_TYPE)  ,"unknown padding type"},
+{ERR_REASON(RSA_R_UNKNOWN_PSS_DIGEST)    ,"unknown pss digest"},
+{ERR_REASON(RSA_R_UNSUPPORTED_MASK_ALGORITHM),"unsupported mask algorithm"},
+{ERR_REASON(RSA_R_UNSUPPORTED_MASK_PARAMETER),"unsupported mask parameter"},
+{ERR_REASON(RSA_R_UNSUPPORTED_SIGNATURE_TYPE),"unsupported signature type"},
+{ERR_REASON(RSA_R_VALUE_MISSING)         ,"value missing"},
 {ERR_REASON(RSA_R_WRONG_SIGNATURE_LENGTH),"wrong signature length"},
 {0,NULL}
 	};
diff --git a/crypto/rsa/rsa_gen.c b/crypto/rsa/rsa_gen.c
index 41278f83c68d..42290cce66cb 100644
--- a/crypto/rsa/rsa_gen.c
+++ b/crypto/rsa/rsa_gen.c
@@ -67,8 +67,9 @@
 #include "cryptlib.h"
 #include <openssl/bn.h>
 #include <openssl/rsa.h>
-
-#ifndef OPENSSL_FIPS
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
 
 static int rsa_builtin_keygen(RSA *rsa, int bits, BIGNUM *e_value, BN_GENCB *cb);
 
@@ -79,8 +80,20 @@ static int rsa_builtin_keygen(RSA *rsa, int bits, BIGNUM *e_value, BN_GENCB *cb)
  * now just because key-generation is part of RSA_METHOD. */
 int RSA_generate_key_ex(RSA *rsa, int bits, BIGNUM *e_value, BN_GENCB *cb)
 	{
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(rsa->meth->flags & RSA_FLAG_FIPS_METHOD)
+			&& !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
+		{
+		RSAerr(RSA_F_RSA_GENERATE_KEY_EX, RSA_R_NON_FIPS_RSA_METHOD);
+		return 0;
+		}
+#endif
 	if(rsa->meth->rsa_keygen)
 		return rsa->meth->rsa_keygen(rsa, bits, e_value, cb);
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode())
+		return FIPS_rsa_generate_key_ex(rsa, bits, e_value, cb);
+#endif
 	return rsa_builtin_keygen(rsa, bits, e_value, cb);
 	}
 
@@ -219,4 +232,3 @@ err:
 	return ok;
 	}
 
-#endif
diff --git a/crypto/rsa/rsa_lib.c b/crypto/rsa/rsa_lib.c
index 5714841f4cab..c95ceafc824d 100644
--- a/crypto/rsa/rsa_lib.c
+++ b/crypto/rsa/rsa_lib.c
@@ -67,168 +67,267 @@
 #include <openssl/engine.h>
 #endif
 
-int RSA_public_encrypt(int flen, const unsigned char *from, unsigned char *to,
-	     RSA *rsa, int padding)
-	{
-	return(rsa->meth->rsa_pub_enc(flen, from, to, rsa, padding));
-	}
-
-int RSA_private_encrypt(int flen, const unsigned char *from, unsigned char *to,
-	     RSA *rsa, int padding)
-	{
 #ifdef OPENSSL_FIPS
-	if(FIPS_mode() && !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
-		{
-		RSAerr(RSA_F_RSA_PRIVATE_ENCRYPT, RSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
-		return 0;
-		}
+#include <openssl/fips.h>
 #endif
-	return(rsa->meth->rsa_priv_enc(flen, from, to, rsa, padding));
+
+const char RSA_version[]="RSA" OPENSSL_VERSION_PTEXT;
+
+static const RSA_METHOD *default_RSA_meth=NULL;
+
+RSA *RSA_new(void)
+	{
+	RSA *r=RSA_new_method(NULL);
+
+	return r;
 	}
 
-int RSA_private_decrypt(int flen, const unsigned char *from, unsigned char *to,
-	     RSA *rsa, int padding)
+void RSA_set_default_method(const RSA_METHOD *meth)
 	{
-	return(rsa->meth->rsa_priv_dec(flen, from, to, rsa, padding));
+	default_RSA_meth = meth;
 	}
 
-int RSA_public_decrypt(int flen, const unsigned char *from, unsigned char *to,
-	     RSA *rsa, int padding)
+const RSA_METHOD *RSA_get_default_method(void)
 	{
-#ifdef OPENSSL_FIPS
-	if(FIPS_mode() && !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
+	if (default_RSA_meth == NULL)
 		{
-		RSAerr(RSA_F_RSA_PUBLIC_DECRYPT, RSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
-		return 0;
-		}
+#ifdef OPENSSL_FIPS
+		if (FIPS_mode())
+			return FIPS_rsa_pkcs1_ssleay();
+		else
+			return RSA_PKCS1_SSLeay();
+#else
+#ifdef RSA_NULL
+		default_RSA_meth=RSA_null_method();
+#else
+		default_RSA_meth=RSA_PKCS1_SSLeay();
+#endif
 #endif
-	return(rsa->meth->rsa_pub_dec(flen, from, to, rsa, padding));
+		}
+
+	return default_RSA_meth;
 	}
 
-int RSA_size(const RSA *r)
+const RSA_METHOD *RSA_get_method(const RSA *rsa)
 	{
-	return(BN_num_bytes(r->n));
+	return rsa->meth;
 	}
 
-void RSA_blinding_off(RSA *rsa)
+int RSA_set_method(RSA *rsa, const RSA_METHOD *meth)
 	{
-	if (rsa->blinding != NULL)
+	/* NB: The caller is specifically setting a method, so it's not up to us
+	 * to deal with which ENGINE it comes from. */
+	const RSA_METHOD *mtmp;
+	mtmp = rsa->meth;
+	if (mtmp->finish) mtmp->finish(rsa);
+#ifndef OPENSSL_NO_ENGINE
+	if (rsa->engine)
 		{
-		BN_BLINDING_free(rsa->blinding);
-		rsa->blinding=NULL;
+		ENGINE_finish(rsa->engine);
+		rsa->engine = NULL;
 		}
-	rsa->flags &= ~RSA_FLAG_BLINDING;
-	rsa->flags |= RSA_FLAG_NO_BLINDING;
+#endif
+	rsa->meth = meth;
+	if (meth->init) meth->init(rsa);
+	return 1;
 	}
 
-int RSA_blinding_on(RSA *rsa, BN_CTX *ctx)
+RSA *RSA_new_method(ENGINE *engine)
 	{
-	int ret=0;
-
-	if (rsa->blinding != NULL)
-		RSA_blinding_off(rsa);
+	RSA *ret;
 
-	rsa->blinding = RSA_setup_blinding(rsa, ctx);
-	if (rsa->blinding == NULL)
-		goto err;
-
-	rsa->flags |= RSA_FLAG_BLINDING;
-	rsa->flags &= ~RSA_FLAG_NO_BLINDING;
-	ret=1;
-err:
-	return(ret);
-	}
-
-static BIGNUM *rsa_get_public_exp(const BIGNUM *d, const BIGNUM *p,
-	const BIGNUM *q, BN_CTX *ctx)
-{
-	BIGNUM *ret = NULL, *r0, *r1, *r2;
-
-	if (d == NULL || p == NULL || q == NULL)
+	ret=(RSA *)OPENSSL_malloc(sizeof(RSA));
+	if (ret == NULL)
+		{
+		RSAerr(RSA_F_RSA_NEW_METHOD,ERR_R_MALLOC_FAILURE);
 		return NULL;
+		}
 
-	BN_CTX_start(ctx);
-	r0 = BN_CTX_get(ctx);
-	r1 = BN_CTX_get(ctx);
-	r2 = BN_CTX_get(ctx);
-	if (r2 == NULL)
-		goto err;
-
-	if (!BN_sub(r1, p, BN_value_one())) goto err;
-	if (!BN_sub(r2, q, BN_value_one())) goto err;
-	if (!BN_mul(r0, r1, r2, ctx)) goto err;
-
-	ret = BN_mod_inverse(NULL, d, r0, ctx);
-err:
-	BN_CTX_end(ctx);
-	return ret;
-}
-
-BN_BLINDING *RSA_setup_blinding(RSA *rsa, BN_CTX *in_ctx)
-{
-	BIGNUM local_n;
-	BIGNUM *e,*n;
-	BN_CTX *ctx;
-	BN_BLINDING *ret = NULL;
-
-	if (in_ctx == NULL)
+	ret->meth = RSA_get_default_method();
+#ifndef OPENSSL_NO_ENGINE
+	if (engine)
 		{
-		if ((ctx = BN_CTX_new()) == NULL) return 0;
+		if (!ENGINE_init(engine))
+			{
+			RSAerr(RSA_F_RSA_NEW_METHOD, ERR_R_ENGINE_LIB);
+			OPENSSL_free(ret);
+			return NULL;
+			}
+		ret->engine = engine;
 		}
 	else
-		ctx = in_ctx;
+		ret->engine = ENGINE_get_default_RSA();
+	if(ret->engine)
+		{
+		ret->meth = ENGINE_get_RSA(ret->engine);
+		if(!ret->meth)
+			{
+			RSAerr(RSA_F_RSA_NEW_METHOD,
+				ERR_R_ENGINE_LIB);
+			ENGINE_finish(ret->engine);
+			OPENSSL_free(ret);
+			return NULL;
+			}
+		}
+#endif
 
-	BN_CTX_start(ctx);
-	e  = BN_CTX_get(ctx);
-	if (e == NULL)
+	ret->pad=0;
+	ret->version=0;
+	ret->n=NULL;
+	ret->e=NULL;
+	ret->d=NULL;
+	ret->p=NULL;
+	ret->q=NULL;
+	ret->dmp1=NULL;
+	ret->dmq1=NULL;
+	ret->iqmp=NULL;
+	ret->references=1;
+	ret->_method_mod_n=NULL;
+	ret->_method_mod_p=NULL;
+	ret->_method_mod_q=NULL;
+	ret->blinding=NULL;
+	ret->mt_blinding=NULL;
+	ret->bignum_data=NULL;
+	ret->flags=ret->meth->flags & ~RSA_FLAG_NON_FIPS_ALLOW;
+	if (!CRYPTO_new_ex_data(CRYPTO_EX_INDEX_RSA, ret, &ret->ex_data))
 		{
-		RSAerr(RSA_F_RSA_SETUP_BLINDING, ERR_R_MALLOC_FAILURE);
-		goto err;
+#ifndef OPENSSL_NO_ENGINE
+	if (ret->engine)
+		ENGINE_finish(ret->engine);
+#endif
+		OPENSSL_free(ret);
+		return(NULL);
 		}
 
-	if (rsa->e == NULL)
+	if ((ret->meth->init != NULL) && !ret->meth->init(ret))
 		{
-		e = rsa_get_public_exp(rsa->d, rsa->p, rsa->q, ctx);
-		if (e == NULL)
-			{
-			RSAerr(RSA_F_RSA_SETUP_BLINDING, RSA_R_NO_PUBLIC_EXPONENT);
-			goto err;
-			}
+#ifndef OPENSSL_NO_ENGINE
+		if (ret->engine)
+			ENGINE_finish(ret->engine);
+#endif
+		CRYPTO_free_ex_data(CRYPTO_EX_INDEX_RSA, ret, &ret->ex_data);
+		OPENSSL_free(ret);
+		ret=NULL;
 		}
-	else
-		e = rsa->e;
+	return(ret);
+	}
 
-	
-	if ((RAND_status() == 0) && rsa->d != NULL && rsa->d->d != NULL)
+void RSA_free(RSA *r)
+	{
+	int i;
+
+	if (r == NULL) return;
+
+	i=CRYPTO_add(&r->references,-1,CRYPTO_LOCK_RSA);
+#ifdef REF_PRINT
+	REF_PRINT("RSA",r);
+#endif
+	if (i > 0) return;
+#ifdef REF_CHECK
+	if (i < 0)
 		{
-		/* if PRNG is not properly seeded, resort to secret
-		 * exponent as unpredictable seed */
-		RAND_add(rsa->d->d, rsa->d->dmax * sizeof rsa->d->d[0], 0.0);
+		fprintf(stderr,"RSA_free, bad reference count\n");
+		abort();
 		}
+#endif
+
+	if (r->meth->finish)
+		r->meth->finish(r);
+#ifndef OPENSSL_NO_ENGINE
+	if (r->engine)
+		ENGINE_finish(r->engine);
+#endif
 
-	if (!(rsa->flags & RSA_FLAG_NO_CONSTTIME))
+	CRYPTO_free_ex_data(CRYPTO_EX_INDEX_RSA, r, &r->ex_data);
+
+	if (r->n != NULL) BN_clear_free(r->n);
+	if (r->e != NULL) BN_clear_free(r->e);
+	if (r->d != NULL) BN_clear_free(r->d);
+	if (r->p != NULL) BN_clear_free(r->p);
+	if (r->q != NULL) BN_clear_free(r->q);
+	if (r->dmp1 != NULL) BN_clear_free(r->dmp1);
+	if (r->dmq1 != NULL) BN_clear_free(r->dmq1);
+	if (r->iqmp != NULL) BN_clear_free(r->iqmp);
+	if (r->blinding != NULL) BN_BLINDING_free(r->blinding);
+	if (r->mt_blinding != NULL) BN_BLINDING_free(r->mt_blinding);
+	if (r->bignum_data != NULL) OPENSSL_free_locked(r->bignum_data);
+	OPENSSL_free(r);
+	}
+
+int RSA_up_ref(RSA *r)
+	{
+	int i = CRYPTO_add(&r->references, 1, CRYPTO_LOCK_RSA);
+#ifdef REF_PRINT
+	REF_PRINT("RSA",r);
+#endif
+#ifdef REF_CHECK
+	if (i < 2)
 		{
-		/* Set BN_FLG_CONSTTIME flag */
-		n = &local_n;
-		BN_with_flags(n, rsa->n, BN_FLG_CONSTTIME);
+		fprintf(stderr, "RSA_up_ref, bad reference count\n");
+		abort();
 		}
-	else
-		n = rsa->n;
+#endif
+	return ((i > 1) ? 1 : 0);
+	}
 
-	ret = BN_BLINDING_create_param(NULL, e, n, ctx,
-			rsa->meth->bn_mod_exp, rsa->_method_mod_n);
-	if (ret == NULL)
+int RSA_get_ex_new_index(long argl, void *argp, CRYPTO_EX_new *new_func,
+	     CRYPTO_EX_dup *dup_func, CRYPTO_EX_free *free_func)
+        {
+	return CRYPTO_get_ex_new_index(CRYPTO_EX_INDEX_RSA, argl, argp,
+				new_func, dup_func, free_func);
+        }
+
+int RSA_set_ex_data(RSA *r, int idx, void *arg)
+	{
+	return(CRYPTO_set_ex_data(&r->ex_data,idx,arg));
+	}
+
+void *RSA_get_ex_data(const RSA *r, int idx)
+	{
+	return(CRYPTO_get_ex_data(&r->ex_data,idx));
+	}
+
+int RSA_memory_lock(RSA *r)
+	{
+	int i,j,k,off;
+	char *p;
+	BIGNUM *bn,**t[6],*b;
+	BN_ULONG *ul;
+
+	if (r->d == NULL) return(1);
+	t[0]= &r->d;
+	t[1]= &r->p;
+	t[2]= &r->q;
+	t[3]= &r->dmp1;
+	t[4]= &r->dmq1;
+	t[5]= &r->iqmp;
+	k=sizeof(BIGNUM)*6;
+	off=k/sizeof(BN_ULONG)+1;
+	j=1;
+	for (i=0; i<6; i++)
+		j+= (*t[i])->top;
+	if ((p=OPENSSL_malloc_locked((off+j)*sizeof(BN_ULONG))) == NULL)
 		{
-		RSAerr(RSA_F_RSA_SETUP_BLINDING, ERR_R_BN_LIB);
-		goto err;
+		RSAerr(RSA_F_RSA_MEMORY_LOCK,ERR_R_MALLOC_FAILURE);
+		return(0);
 		}
-	BN_BLINDING_set_thread_id(ret, CRYPTO_thread_id());
-err:
-	BN_CTX_end(ctx);
-	if (in_ctx == NULL)
-		BN_CTX_free(ctx);
-	if(rsa->e == NULL)
-		BN_free(e);
-
-	return ret;
-}
+	bn=(BIGNUM *)p;
+	ul=(BN_ULONG *)&(p[off]);
+	for (i=0; i<6; i++)
+		{
+		b= *(t[i]);
+		*(t[i])= &(bn[i]);
+		memcpy((char *)&(bn[i]),(char *)b,sizeof(BIGNUM));
+		bn[i].flags=BN_FLG_STATIC_DATA;
+		bn[i].d=ul;
+		memcpy((char *)ul,b->d,sizeof(BN_ULONG)*b->top);
+		ul+=b->top;
+		BN_clear_free(b);
+		}
+	
+	/* I should fix this so it can still be done */
+	r->flags&= ~(RSA_FLAG_CACHE_PRIVATE|RSA_FLAG_CACHE_PUBLIC);
+
+	r->bignum_data=p;
+	return(1);
+	}
diff --git a/crypto/rsa/rsa_locl.h b/crypto/rsa/rsa_locl.h
new file mode 100644
index 000000000000..f5d2d5662843
--- /dev/null
+++ b/crypto/rsa/rsa_locl.h
@@ -0,0 +1,4 @@
+extern int int_rsa_verify(int dtype, const unsigned char *m, unsigned int m_len,
+		unsigned char *rm, size_t *prm_len,
+		const unsigned char *sigbuf, size_t siglen,
+		RSA *rsa);
diff --git a/crypto/rsa/rsa_oaep.c b/crypto/rsa/rsa_oaep.c
index 546ae5fcb2ed..553d212ebe9c 100644
--- a/crypto/rsa/rsa_oaep.c
+++ b/crypto/rsa/rsa_oaep.c
@@ -28,7 +28,7 @@
 #include <openssl/rand.h>
 #include <openssl/sha.h>
 
-int MGF1(unsigned char *mask, long len,
+static int MGF1(unsigned char *mask, long len,
 	const unsigned char *seed, long seedlen);
 
 int RSA_padding_add_PKCS1_OAEP(unsigned char *to, int tlen,
@@ -56,7 +56,8 @@ int RSA_padding_add_PKCS1_OAEP(unsigned char *to, int tlen,
 	seed = to + 1;
 	db = to + SHA_DIGEST_LENGTH + 1;
 
-	EVP_Digest((void *)param, plen, db, NULL, EVP_sha1(), NULL);
+	if (!EVP_Digest((void *)param, plen, db, NULL, EVP_sha1(), NULL))
+		return 0;
 	memset(db + SHA_DIGEST_LENGTH, 0,
 		emlen - flen - 2 * SHA_DIGEST_LENGTH - 1);
 	db[emlen - flen - SHA_DIGEST_LENGTH - 1] = 0x01;
@@ -76,11 +77,13 @@ int RSA_padding_add_PKCS1_OAEP(unsigned char *to, int tlen,
 		return 0;
 		}
 
-	MGF1(dbmask, emlen - SHA_DIGEST_LENGTH, seed, SHA_DIGEST_LENGTH);
+	if (MGF1(dbmask, emlen - SHA_DIGEST_LENGTH, seed, SHA_DIGEST_LENGTH) < 0)
+		return 0;
 	for (i = 0; i < emlen - SHA_DIGEST_LENGTH; i++)
 		db[i] ^= dbmask[i];
 
-	MGF1(seedmask, SHA_DIGEST_LENGTH, db, emlen - SHA_DIGEST_LENGTH);
+	if (MGF1(seedmask, SHA_DIGEST_LENGTH, db, emlen - SHA_DIGEST_LENGTH) < 0)
+		return 0;
 	for (i = 0; i < SHA_DIGEST_LENGTH; i++)
 		seed[i] ^= seedmask[i];
 
@@ -133,15 +136,18 @@ int RSA_padding_check_PKCS1_OAEP(unsigned char *to, int tlen,
 
 	maskeddb = padded_from + SHA_DIGEST_LENGTH;
 
-	MGF1(seed, SHA_DIGEST_LENGTH, maskeddb, dblen);
+	if (MGF1(seed, SHA_DIGEST_LENGTH, maskeddb, dblen))
+		return -1;
 	for (i = 0; i < SHA_DIGEST_LENGTH; i++)
 		seed[i] ^= padded_from[i];
   
-	MGF1(db, dblen, seed, SHA_DIGEST_LENGTH);
+	if (MGF1(db, dblen, seed, SHA_DIGEST_LENGTH))
+		return -1;
 	for (i = 0; i < dblen; i++)
 		db[i] ^= maskeddb[i];
 
-	EVP_Digest((void *)param, plen, phash, NULL, EVP_sha1(), NULL);
+	if (!EVP_Digest((void *)param, plen, phash, NULL, EVP_sha1(), NULL))
+		return -1;
 
 	if (memcmp(db, phash, SHA_DIGEST_LENGTH) != 0 || bad)
 		goto decoding_err;
@@ -185,35 +191,44 @@ int PKCS1_MGF1(unsigned char *mask, long len,
 	EVP_MD_CTX c;
 	unsigned char md[EVP_MAX_MD_SIZE];
 	int mdlen;
+	int rv = -1;
 
 	EVP_MD_CTX_init(&c);
-	mdlen = M_EVP_MD_size(dgst);
+	mdlen = EVP_MD_size(dgst);
+	if (mdlen < 0)
+		goto err;
 	for (i = 0; outlen < len; i++)
 		{
 		cnt[0] = (unsigned char)((i >> 24) & 255);
 		cnt[1] = (unsigned char)((i >> 16) & 255);
 		cnt[2] = (unsigned char)((i >> 8)) & 255;
 		cnt[3] = (unsigned char)(i & 255);
-		EVP_DigestInit_ex(&c,dgst, NULL);
-		EVP_DigestUpdate(&c, seed, seedlen);
-		EVP_DigestUpdate(&c, cnt, 4);
+		if (!EVP_DigestInit_ex(&c,dgst, NULL)
+			|| !EVP_DigestUpdate(&c, seed, seedlen)
+			|| !EVP_DigestUpdate(&c, cnt, 4))
+			goto err;
 		if (outlen + mdlen <= len)
 			{
-			EVP_DigestFinal_ex(&c, mask + outlen, NULL);
+			if (!EVP_DigestFinal_ex(&c, mask + outlen, NULL))
+				goto err;
 			outlen += mdlen;
 			}
 		else
 			{
-			EVP_DigestFinal_ex(&c, md, NULL);
+			if (!EVP_DigestFinal_ex(&c, md, NULL))
+				goto err;
 			memcpy(mask + outlen, md, len - outlen);
 			outlen = len;
 			}
 		}
+	rv = 0;
+	err:
 	EVP_MD_CTX_cleanup(&c);
-	return 0;
+	return rv;
 	}
 
-int MGF1(unsigned char *mask, long len, const unsigned char *seed, long seedlen)
+static int MGF1(unsigned char *mask, long len, const unsigned char *seed,
+		 long seedlen)
 	{
 	return PKCS1_MGF1(mask, len, seed, seedlen, EVP_sha1());
 	}
diff --git a/crypto/rsa/rsa_pmeth.c b/crypto/rsa/rsa_pmeth.c
new file mode 100644
index 000000000000..5b2ecf56adc7
--- /dev/null
+++ b/crypto/rsa/rsa_pmeth.c
@@ -0,0 +1,723 @@
+/* crypto/rsa/rsa_pmeth.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/asn1t.h>
+#include <openssl/x509.h>
+#include <openssl/rsa.h>
+#include <openssl/bn.h>
+#include <openssl/evp.h>
+#ifndef OPENSSL_NO_CMS
+#include <openssl/cms.h>
+#endif
+#ifdef OPENSSL_FIPS
+#include <openssl/fips.h>
+#endif
+#include "evp_locl.h"
+#include "rsa_locl.h"
+
+/* RSA pkey context structure */
+
+typedef struct
+	{
+	/* Key gen parameters */
+	int nbits;
+	BIGNUM *pub_exp;
+	/* Keygen callback info */
+	int gentmp[2];
+	/* RSA padding mode */
+	int pad_mode;
+	/* message digest */
+	const EVP_MD *md;
+	/* message digest for MGF1 */
+	const EVP_MD *mgf1md;
+	/* PSS/OAEP salt length */
+	int saltlen;
+	/* Temp buffer */
+	unsigned char *tbuf;
+	} RSA_PKEY_CTX;
+
+static int pkey_rsa_init(EVP_PKEY_CTX *ctx)
+	{
+	RSA_PKEY_CTX *rctx;
+	rctx = OPENSSL_malloc(sizeof(RSA_PKEY_CTX));
+	if (!rctx)
+		return 0;
+	rctx->nbits = 1024;
+	rctx->pub_exp = NULL;
+	rctx->pad_mode = RSA_PKCS1_PADDING;
+	rctx->md = NULL;
+	rctx->mgf1md = NULL;
+	rctx->tbuf = NULL;
+
+	rctx->saltlen = -2;
+
+	ctx->data = rctx;
+	ctx->keygen_info = rctx->gentmp;
+	ctx->keygen_info_count = 2;
+	
+	return 1;
+	}
+
+static int pkey_rsa_copy(EVP_PKEY_CTX *dst, EVP_PKEY_CTX *src)
+	{
+	RSA_PKEY_CTX *dctx, *sctx;
+	if (!pkey_rsa_init(dst))
+		return 0;
+       	sctx = src->data;
+	dctx = dst->data;
+	dctx->nbits = sctx->nbits;
+	if (sctx->pub_exp)
+		{
+		dctx->pub_exp = BN_dup(sctx->pub_exp);
+		if (!dctx->pub_exp)
+			return 0;
+		}
+	dctx->pad_mode = sctx->pad_mode;
+	dctx->md = sctx->md;
+	return 1;
+	}
+
+static int setup_tbuf(RSA_PKEY_CTX *ctx, EVP_PKEY_CTX *pk)
+	{
+	if (ctx->tbuf)
+		return 1;
+	ctx->tbuf = OPENSSL_malloc(EVP_PKEY_size(pk->pkey));
+	if (!ctx->tbuf)
+		return 0;
+	return 1;
+	}
+
+static void pkey_rsa_cleanup(EVP_PKEY_CTX *ctx)
+	{
+	RSA_PKEY_CTX *rctx = ctx->data;
+	if (rctx)
+		{
+		if (rctx->pub_exp)
+			BN_free(rctx->pub_exp);
+		if (rctx->tbuf)
+			OPENSSL_free(rctx->tbuf);
+		OPENSSL_free(rctx);
+		}
+	}
+#ifdef OPENSSL_FIPS
+/* FIP checker. Return value indicates status of context parameters:
+ * 1  : redirect to FIPS.
+ * 0  : don't redirect to FIPS.
+ * -1 : illegal operation in FIPS mode.
+ */
+
+static int pkey_fips_check_ctx(EVP_PKEY_CTX *ctx)
+	{
+	RSA_PKEY_CTX *rctx = ctx->data;
+	RSA *rsa = ctx->pkey->pkey.rsa;
+	int rv = -1;
+	if (!FIPS_mode())
+		return 0;
+	if (rsa->flags & RSA_FLAG_NON_FIPS_ALLOW)
+		rv = 0;
+	if (!(rsa->meth->flags & RSA_FLAG_FIPS_METHOD) && rv)
+		return -1;
+	if (rctx->md && !(rctx->md->flags & EVP_MD_FLAG_FIPS))
+		return rv;
+	if (rctx->mgf1md && !(rctx->mgf1md->flags & EVP_MD_FLAG_FIPS))
+		return rv;
+	return 1;
+	}
+#endif
+
+static int pkey_rsa_sign(EVP_PKEY_CTX *ctx, unsigned char *sig, size_t *siglen,
+					const unsigned char *tbs, size_t tbslen)
+	{
+	int ret;
+	RSA_PKEY_CTX *rctx = ctx->data;
+	RSA *rsa = ctx->pkey->pkey.rsa;
+
+#ifdef OPENSSL_FIPS
+	ret = pkey_fips_check_ctx(ctx);
+	if (ret < 0)
+		{
+		RSAerr(RSA_F_PKEY_RSA_SIGN, RSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
+		return -1;
+		}
+#endif
+
+	if (rctx->md)
+		{
+		if (tbslen != (size_t)EVP_MD_size(rctx->md))
+			{
+			RSAerr(RSA_F_PKEY_RSA_SIGN,
+					RSA_R_INVALID_DIGEST_LENGTH);
+			return -1;
+			}
+#ifdef OPENSSL_FIPS
+		if (ret > 0)
+			{
+			unsigned int slen;
+			ret = FIPS_rsa_sign_digest(rsa, tbs, tbslen, rctx->md,
+							rctx->pad_mode,
+							rctx->saltlen,
+							rctx->mgf1md,
+							sig, &slen);
+			if (ret > 0)
+				*siglen = slen;
+			else
+				*siglen = 0;
+			return ret;
+			}
+#endif
+
+		if (EVP_MD_type(rctx->md) == NID_mdc2)
+			{
+			unsigned int sltmp;
+			if (rctx->pad_mode != RSA_PKCS1_PADDING)
+				return -1;
+			ret = RSA_sign_ASN1_OCTET_STRING(NID_mdc2,
+						tbs, tbslen, sig, &sltmp, rsa);
+
+			if (ret <= 0)
+				return ret;
+			ret = sltmp;
+			}
+		else if (rctx->pad_mode == RSA_X931_PADDING)
+			{
+			if (!setup_tbuf(rctx, ctx))
+				return -1;
+			memcpy(rctx->tbuf, tbs, tbslen);
+			rctx->tbuf[tbslen] =
+				RSA_X931_hash_id(EVP_MD_type(rctx->md));
+			ret = RSA_private_encrypt(tbslen + 1, rctx->tbuf,
+						sig, rsa, RSA_X931_PADDING);
+			}
+		else if (rctx->pad_mode == RSA_PKCS1_PADDING)
+			{
+			unsigned int sltmp;
+			ret = RSA_sign(EVP_MD_type(rctx->md),
+						tbs, tbslen, sig, &sltmp, rsa);
+			if (ret <= 0)
+				return ret;
+			ret = sltmp;
+			}
+		else if (rctx->pad_mode == RSA_PKCS1_PSS_PADDING)
+			{
+			if (!setup_tbuf(rctx, ctx))
+				return -1;
+			if (!RSA_padding_add_PKCS1_PSS_mgf1(rsa,
+						rctx->tbuf, tbs,
+						rctx->md, rctx->mgf1md,
+						rctx->saltlen))
+				return -1;
+			ret = RSA_private_encrypt(RSA_size(rsa), rctx->tbuf,
+						sig, rsa, RSA_NO_PADDING);
+			}
+		else
+			return -1;
+		}
+	else
+		ret = RSA_private_encrypt(tbslen, tbs, sig, ctx->pkey->pkey.rsa,
+							rctx->pad_mode);
+	if (ret < 0)
+		return ret;
+	*siglen = ret;
+	return 1;
+	}
+
+
+static int pkey_rsa_verifyrecover(EVP_PKEY_CTX *ctx,
+					unsigned char *rout, size_t *routlen,
+					const unsigned char *sig, size_t siglen)
+	{
+	int ret;
+	RSA_PKEY_CTX *rctx = ctx->data;
+
+	if (rctx->md)
+		{
+		if (rctx->pad_mode == RSA_X931_PADDING)
+			{
+			if (!setup_tbuf(rctx, ctx))
+				return -1;
+			ret = RSA_public_decrypt(siglen, sig,
+						rctx->tbuf, ctx->pkey->pkey.rsa,
+						RSA_X931_PADDING);
+			if (ret < 1)
+				return 0;
+			ret--;
+			if (rctx->tbuf[ret] !=
+				RSA_X931_hash_id(EVP_MD_type(rctx->md)))
+				{
+				RSAerr(RSA_F_PKEY_RSA_VERIFYRECOVER,
+						RSA_R_ALGORITHM_MISMATCH);
+				return 0;
+				}
+			if (ret != EVP_MD_size(rctx->md))
+				{
+				RSAerr(RSA_F_PKEY_RSA_VERIFYRECOVER,
+					RSA_R_INVALID_DIGEST_LENGTH);
+				return 0;
+				}
+			if (rout)
+				memcpy(rout, rctx->tbuf, ret);
+			}
+		else if (rctx->pad_mode == RSA_PKCS1_PADDING)
+			{
+			size_t sltmp;
+			ret = int_rsa_verify(EVP_MD_type(rctx->md),
+						NULL, 0, rout, &sltmp,
+					sig, siglen, ctx->pkey->pkey.rsa);
+			if (ret <= 0)
+				return 0;
+			ret = sltmp;
+			}
+		else
+			return -1;
+		}
+	else
+		ret = RSA_public_decrypt(siglen, sig, rout, ctx->pkey->pkey.rsa,
+							rctx->pad_mode);
+	if (ret < 0)
+		return ret;
+	*routlen = ret;
+	return 1;
+	}
+
+static int pkey_rsa_verify(EVP_PKEY_CTX *ctx,
+					const unsigned char *sig, size_t siglen,
+					const unsigned char *tbs, size_t tbslen)
+	{
+	RSA_PKEY_CTX *rctx = ctx->data;
+	RSA *rsa = ctx->pkey->pkey.rsa;
+	size_t rslen;
+#ifdef OPENSSL_FIPS
+	int rv;
+	rv = pkey_fips_check_ctx(ctx);
+	if (rv < 0)
+		{
+		RSAerr(RSA_F_PKEY_RSA_VERIFY, RSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
+		return -1;
+		}
+#endif
+	if (rctx->md)
+		{
+#ifdef OPENSSL_FIPS
+		if (rv > 0)
+			{
+			return FIPS_rsa_verify_digest(rsa,
+							tbs, tbslen,
+							rctx->md,
+							rctx->pad_mode,
+							rctx->saltlen,
+							rctx->mgf1md,
+							sig, siglen);
+							
+			}
+#endif
+		if (rctx->pad_mode == RSA_PKCS1_PADDING)
+			return RSA_verify(EVP_MD_type(rctx->md), tbs, tbslen,
+					sig, siglen, rsa);
+		if (rctx->pad_mode == RSA_X931_PADDING)
+			{
+			if (pkey_rsa_verifyrecover(ctx, NULL, &rslen,
+					sig, siglen) <= 0)
+				return 0;
+			}
+		else if (rctx->pad_mode == RSA_PKCS1_PSS_PADDING)
+			{
+			int ret;
+			if (!setup_tbuf(rctx, ctx))
+				return -1;
+			ret = RSA_public_decrypt(siglen, sig, rctx->tbuf,
+							rsa, RSA_NO_PADDING);
+			if (ret <= 0)
+				return 0;
+			ret = RSA_verify_PKCS1_PSS_mgf1(rsa, tbs,
+						rctx->md, rctx->mgf1md,
+						rctx->tbuf, rctx->saltlen);
+			if (ret <= 0)
+				return 0;
+			return 1;
+			}
+		else
+			return -1;
+		}
+	else
+		{
+		if (!setup_tbuf(rctx, ctx))
+			return -1;
+		rslen = RSA_public_decrypt(siglen, sig, rctx->tbuf,
+						rsa, rctx->pad_mode);
+		if (rslen == 0)
+			return 0;
+		}
+
+	if ((rslen != tbslen) || memcmp(tbs, rctx->tbuf, rslen))
+		return 0;
+
+	return 1;
+			
+	}
+	
+
+static int pkey_rsa_encrypt(EVP_PKEY_CTX *ctx,
+					unsigned char *out, size_t *outlen,
+					const unsigned char *in, size_t inlen)
+	{
+	int ret;
+	RSA_PKEY_CTX *rctx = ctx->data;
+	ret = RSA_public_encrypt(inlen, in, out, ctx->pkey->pkey.rsa,
+							rctx->pad_mode);
+	if (ret < 0)
+		return ret;
+	*outlen = ret;
+	return 1;
+	}
+
+static int pkey_rsa_decrypt(EVP_PKEY_CTX *ctx,
+					unsigned char *out, size_t *outlen,
+					const unsigned char *in, size_t inlen)
+	{
+	int ret;
+	RSA_PKEY_CTX *rctx = ctx->data;
+	ret = RSA_private_decrypt(inlen, in, out, ctx->pkey->pkey.rsa,
+							rctx->pad_mode);
+	if (ret < 0)
+		return ret;
+	*outlen = ret;
+	return 1;
+	}
+
+static int check_padding_md(const EVP_MD *md, int padding)
+	{
+	if (!md)
+		return 1;
+
+	if (padding == RSA_NO_PADDING)
+		{
+		RSAerr(RSA_F_CHECK_PADDING_MD, RSA_R_INVALID_PADDING_MODE);
+		return 0;
+		}
+
+	if (padding == RSA_X931_PADDING)
+		{
+		if (RSA_X931_hash_id(EVP_MD_type(md)) == -1)
+			{
+			RSAerr(RSA_F_CHECK_PADDING_MD,
+						RSA_R_INVALID_X931_DIGEST);
+			return 0;
+			}
+		return 1;
+		}
+
+	return 1;
+	}
+			
+
+static int pkey_rsa_ctrl(EVP_PKEY_CTX *ctx, int type, int p1, void *p2)
+	{
+	RSA_PKEY_CTX *rctx = ctx->data;
+	switch (type)
+		{
+		case EVP_PKEY_CTRL_RSA_PADDING:
+		if ((p1 >= RSA_PKCS1_PADDING) && (p1 <= RSA_PKCS1_PSS_PADDING))
+			{
+			if (!check_padding_md(rctx->md, p1))
+				return 0;
+			if (p1 == RSA_PKCS1_PSS_PADDING) 
+				{
+				if (!(ctx->operation &
+				     (EVP_PKEY_OP_SIGN | EVP_PKEY_OP_VERIFY)))
+					goto bad_pad;
+				if (!rctx->md)
+					rctx->md = EVP_sha1();
+				}
+			if (p1 == RSA_PKCS1_OAEP_PADDING) 
+				{
+				if (!(ctx->operation & EVP_PKEY_OP_TYPE_CRYPT))
+					goto bad_pad;
+				if (!rctx->md)
+					rctx->md = EVP_sha1();
+				}
+			rctx->pad_mode = p1;
+			return 1;
+			}
+		bad_pad:
+		RSAerr(RSA_F_PKEY_RSA_CTRL,
+				RSA_R_ILLEGAL_OR_UNSUPPORTED_PADDING_MODE);
+		return -2;
+
+		case EVP_PKEY_CTRL_GET_RSA_PADDING:
+		*(int *)p2 = rctx->pad_mode;
+		return 1;
+
+		case EVP_PKEY_CTRL_RSA_PSS_SALTLEN:
+		case EVP_PKEY_CTRL_GET_RSA_PSS_SALTLEN:
+		if (rctx->pad_mode != RSA_PKCS1_PSS_PADDING)
+			{
+			RSAerr(RSA_F_PKEY_RSA_CTRL, RSA_R_INVALID_PSS_SALTLEN);
+			return -2;
+			}
+		if (type == EVP_PKEY_CTRL_GET_RSA_PSS_SALTLEN)
+			*(int *)p2 = rctx->saltlen;
+		else
+			{
+			if (p1 < -2)
+				return -2;
+			rctx->saltlen = p1;
+			}
+		return 1;
+
+		case EVP_PKEY_CTRL_RSA_KEYGEN_BITS:
+		if (p1 < 256)
+			{
+			RSAerr(RSA_F_PKEY_RSA_CTRL, RSA_R_INVALID_KEYBITS);
+			return -2;
+			}
+		rctx->nbits = p1;
+		return 1;
+
+		case EVP_PKEY_CTRL_RSA_KEYGEN_PUBEXP:
+		if (!p2)
+			return -2;
+		rctx->pub_exp = p2;
+		return 1;
+
+		case EVP_PKEY_CTRL_MD:
+		if (!check_padding_md(p2, rctx->pad_mode))
+			return 0;
+		rctx->md = p2;
+		return 1;
+
+		case EVP_PKEY_CTRL_RSA_MGF1_MD:
+		case EVP_PKEY_CTRL_GET_RSA_MGF1_MD:
+		if (rctx->pad_mode != RSA_PKCS1_PSS_PADDING)
+			{
+			RSAerr(RSA_F_PKEY_RSA_CTRL, RSA_R_INVALID_MGF1_MD);
+			return -2;
+			}
+		if (type == EVP_PKEY_CTRL_GET_RSA_MGF1_MD)
+			{
+			if (rctx->mgf1md)
+				*(const EVP_MD **)p2 = rctx->mgf1md;
+			else
+				*(const EVP_MD **)p2 = rctx->md;
+			}
+		else
+			rctx->mgf1md = p2;
+		return 1;
+
+		case EVP_PKEY_CTRL_DIGESTINIT:
+		case EVP_PKEY_CTRL_PKCS7_ENCRYPT:
+		case EVP_PKEY_CTRL_PKCS7_DECRYPT:
+		case EVP_PKEY_CTRL_PKCS7_SIGN:
+		return 1;
+#ifndef OPENSSL_NO_CMS
+		case EVP_PKEY_CTRL_CMS_DECRYPT:
+		{
+		X509_ALGOR *alg = NULL;
+		ASN1_OBJECT *encalg = NULL;
+		if (p2)
+			CMS_RecipientInfo_ktri_get0_algs(p2, NULL, NULL, &alg);
+		if (alg)
+			X509_ALGOR_get0(&encalg, NULL, NULL, alg);
+		if (encalg && OBJ_obj2nid(encalg) == NID_rsaesOaep)
+			rctx->pad_mode = RSA_PKCS1_OAEP_PADDING;
+		}
+		case EVP_PKEY_CTRL_CMS_ENCRYPT:
+		case EVP_PKEY_CTRL_CMS_SIGN:
+		return 1;
+#endif
+		case EVP_PKEY_CTRL_PEER_KEY:
+			RSAerr(RSA_F_PKEY_RSA_CTRL,
+			RSA_R_OPERATION_NOT_SUPPORTED_FOR_THIS_KEYTYPE);
+			return -2;	
+
+		default:
+		return -2;
+
+		}
+	}
+			
+static int pkey_rsa_ctrl_str(EVP_PKEY_CTX *ctx,
+			const char *type, const char *value)
+	{
+	if (!value)
+		{
+		RSAerr(RSA_F_PKEY_RSA_CTRL_STR, RSA_R_VALUE_MISSING);
+		return 0;
+		}
+	if (!strcmp(type, "rsa_padding_mode"))
+		{
+		int pm;
+		if (!strcmp(value, "pkcs1"))
+			pm = RSA_PKCS1_PADDING;
+		else if (!strcmp(value, "sslv23"))
+			pm = RSA_SSLV23_PADDING;
+		else if (!strcmp(value, "none"))
+			pm = RSA_NO_PADDING;
+		else if (!strcmp(value, "oeap"))
+			pm = RSA_PKCS1_OAEP_PADDING;
+		else if (!strcmp(value, "x931"))
+			pm = RSA_X931_PADDING;
+		else if (!strcmp(value, "pss"))
+			pm = RSA_PKCS1_PSS_PADDING;
+		else
+			{
+			RSAerr(RSA_F_PKEY_RSA_CTRL_STR,
+						RSA_R_UNKNOWN_PADDING_TYPE);
+			return -2;
+			}
+		return EVP_PKEY_CTX_set_rsa_padding(ctx, pm);
+		}
+
+	if (!strcmp(type, "rsa_pss_saltlen"))
+		{
+		int saltlen;
+		saltlen = atoi(value);
+		return EVP_PKEY_CTX_set_rsa_pss_saltlen(ctx, saltlen);
+		}
+
+	if (!strcmp(type, "rsa_keygen_bits"))
+		{
+		int nbits;
+		nbits = atoi(value);
+		return EVP_PKEY_CTX_set_rsa_keygen_bits(ctx, nbits);
+		}
+
+	if (!strcmp(type, "rsa_keygen_pubexp"))
+		{
+		int ret;
+		BIGNUM *pubexp = NULL;
+		if (!BN_asc2bn(&pubexp, value))
+			return 0;
+		ret = EVP_PKEY_CTX_set_rsa_keygen_pubexp(ctx, pubexp);
+		if (ret <= 0)
+			BN_free(pubexp);
+		return ret;
+		}
+
+	return -2;
+	}
+
+static int pkey_rsa_keygen(EVP_PKEY_CTX *ctx, EVP_PKEY *pkey)
+	{
+	RSA *rsa = NULL;
+	RSA_PKEY_CTX *rctx = ctx->data;
+	BN_GENCB *pcb, cb;
+	int ret;
+	if (!rctx->pub_exp)
+		{
+		rctx->pub_exp = BN_new();
+		if (!rctx->pub_exp || !BN_set_word(rctx->pub_exp, RSA_F4))
+			return 0;
+		}
+	rsa = RSA_new();
+	if (!rsa)
+		return 0;
+	if (ctx->pkey_gencb)
+		{
+		pcb = &cb;
+		evp_pkey_set_cb_translate(pcb, ctx);
+		}
+	else
+		pcb = NULL;
+	ret = RSA_generate_key_ex(rsa, rctx->nbits, rctx->pub_exp, pcb);
+	if (ret > 0)
+		EVP_PKEY_assign_RSA(pkey, rsa);
+	else
+		RSA_free(rsa);
+	return ret;
+	}
+
+const EVP_PKEY_METHOD rsa_pkey_meth = 
+	{
+	EVP_PKEY_RSA,
+	EVP_PKEY_FLAG_AUTOARGLEN,
+	pkey_rsa_init,
+	pkey_rsa_copy,
+	pkey_rsa_cleanup,
+
+	0,0,
+
+	0,
+	pkey_rsa_keygen,
+
+	0,
+	pkey_rsa_sign,
+
+	0,
+	pkey_rsa_verify,
+
+	0,
+	pkey_rsa_verifyrecover,
+
+
+	0,0,0,0,
+
+	0,
+	pkey_rsa_encrypt,
+
+	0,
+	pkey_rsa_decrypt,
+
+	0,0,
+
+	pkey_rsa_ctrl,
+	pkey_rsa_ctrl_str
+
+
+	};
diff --git a/crypto/rsa/rsa_prn.c b/crypto/rsa/rsa_prn.c
new file mode 100644
index 000000000000..224db0fae52a
--- /dev/null
+++ b/crypto/rsa/rsa_prn.c
@@ -0,0 +1,93 @@
+/* crypto/rsa/rsa_prn.c */
+/* Written by Dr Stephen N Henson (steve@openssl.org) for the OpenSSL
+ * project 2006.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/rsa.h>
+#include <openssl/evp.h>
+
+#ifndef OPENSSL_NO_FP_API
+int RSA_print_fp(FILE *fp, const RSA *x, int off)
+	{
+	BIO *b;
+	int ret;
+
+	if ((b=BIO_new(BIO_s_file())) == NULL)
+		{
+		RSAerr(RSA_F_RSA_PRINT_FP,ERR_R_BUF_LIB);
+		return(0);
+		}
+	BIO_set_fp(b,fp,BIO_NOCLOSE);
+	ret=RSA_print(b,x,off);
+	BIO_free(b);
+	return(ret);
+	}
+#endif
+
+int RSA_print(BIO *bp, const RSA *x, int off)
+	{
+	EVP_PKEY *pk;
+	int ret;
+	pk = EVP_PKEY_new();
+	if (!pk || !EVP_PKEY_set1_RSA(pk, (RSA *)x))
+		return 0;
+	ret = EVP_PKEY_print_private(bp, pk, off, NULL);
+	EVP_PKEY_free(pk);
+	return ret;
+	}
+
diff --git a/crypto/rsa/rsa_pss.c b/crypto/rsa/rsa_pss.c
index 2bda491a091c..5f9f533d0ce5 100644
--- a/crypto/rsa/rsa_pss.c
+++ b/crypto/rsa/rsa_pss.c
@@ -73,6 +73,13 @@ static const unsigned char zeroes[] = {0,0,0,0,0,0,0,0};
 int RSA_verify_PKCS1_PSS(RSA *rsa, const unsigned char *mHash,
 			const EVP_MD *Hash, const unsigned char *EM, int sLen)
 	{
+	return RSA_verify_PKCS1_PSS_mgf1(rsa, mHash, Hash, NULL, EM, sLen);
+	}
+
+int RSA_verify_PKCS1_PSS_mgf1(RSA *rsa, const unsigned char *mHash,
+			const EVP_MD *Hash, const EVP_MD *mgf1Hash,
+			const unsigned char *EM, int sLen)
+	{
 	int i;
 	int ret = 0;
 	int hLen, maskedDBLen, MSBits, emLen;
@@ -80,8 +87,14 @@ int RSA_verify_PKCS1_PSS(RSA *rsa, const unsigned char *mHash,
 	unsigned char *DB = NULL;
 	EVP_MD_CTX ctx;
 	unsigned char H_[EVP_MAX_MD_SIZE];
+	EVP_MD_CTX_init(&ctx);
 
-	hLen = M_EVP_MD_size(Hash);
+	if (mgf1Hash == NULL)
+		mgf1Hash = Hash;
+
+	hLen = EVP_MD_size(Hash);
+	if (hLen < 0)
+		goto err;
 	/*
 	 * Negative sLen has special meanings:
 	 *	-1	sLen == hLen
@@ -92,7 +105,7 @@ int RSA_verify_PKCS1_PSS(RSA *rsa, const unsigned char *mHash,
 	else if (sLen == -2)	sLen = -2;
 	else if (sLen < -2)
 		{
-		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS, RSA_R_SLEN_CHECK_FAILED);
+		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1, RSA_R_SLEN_CHECK_FAILED);
 		goto err;
 		}
 
@@ -100,7 +113,7 @@ int RSA_verify_PKCS1_PSS(RSA *rsa, const unsigned char *mHash,
 	emLen = RSA_size(rsa);
 	if (EM[0] & (0xFF << MSBits))
 		{
-		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS, RSA_R_FIRST_OCTET_INVALID);
+		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1, RSA_R_FIRST_OCTET_INVALID);
 		goto err;
 		}
 	if (MSBits == 0)
@@ -110,12 +123,12 @@ int RSA_verify_PKCS1_PSS(RSA *rsa, const unsigned char *mHash,
 		}
 	if (emLen < (hLen + sLen + 2)) /* sLen can be small negative */
 		{
-		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS, RSA_R_DATA_TOO_LARGE);
+		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1, RSA_R_DATA_TOO_LARGE);
 		goto err;
 		}
 	if (EM[emLen - 1] != 0xbc)
 		{
-		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS, RSA_R_LAST_OCTET_INVALID);
+		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1, RSA_R_LAST_OCTET_INVALID);
 		goto err;
 		}
 	maskedDBLen = emLen - hLen - 1;
@@ -123,10 +136,11 @@ int RSA_verify_PKCS1_PSS(RSA *rsa, const unsigned char *mHash,
 	DB = OPENSSL_malloc(maskedDBLen);
 	if (!DB)
 		{
-		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS, ERR_R_MALLOC_FAILURE);
+		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1, ERR_R_MALLOC_FAILURE);
 		goto err;
 		}
-	PKCS1_MGF1(DB, maskedDBLen, H, hLen, Hash);
+	if (PKCS1_MGF1(DB, maskedDBLen, H, hLen, mgf1Hash) < 0)
+		goto err;
 	for (i = 0; i < maskedDBLen; i++)
 		DB[i] ^= EM[i];
 	if (MSBits)
@@ -134,25 +148,28 @@ int RSA_verify_PKCS1_PSS(RSA *rsa, const unsigned char *mHash,
 	for (i = 0; DB[i] == 0 && i < (maskedDBLen-1); i++) ;
 	if (DB[i++] != 0x1)
 		{
-		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS, RSA_R_SLEN_RECOVERY_FAILED);
+		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1, RSA_R_SLEN_RECOVERY_FAILED);
 		goto err;
 		}
 	if (sLen >= 0 && (maskedDBLen - i) != sLen)
 		{
-		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS, RSA_R_SLEN_CHECK_FAILED);
+		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1, RSA_R_SLEN_CHECK_FAILED);
 		goto err;
 		}
-	EVP_MD_CTX_init(&ctx);
-	EVP_DigestInit_ex(&ctx, Hash, NULL);
-	EVP_DigestUpdate(&ctx, zeroes, sizeof zeroes);
-	EVP_DigestUpdate(&ctx, mHash, hLen);
+	if (!EVP_DigestInit_ex(&ctx, Hash, NULL)
+		|| !EVP_DigestUpdate(&ctx, zeroes, sizeof zeroes)
+		|| !EVP_DigestUpdate(&ctx, mHash, hLen))
+		goto err;
 	if (maskedDBLen - i)
-		EVP_DigestUpdate(&ctx, DB + i, maskedDBLen - i);
-	EVP_DigestFinal(&ctx, H_, NULL);
-	EVP_MD_CTX_cleanup(&ctx);
+		{
+		if (!EVP_DigestUpdate(&ctx, DB + i, maskedDBLen - i))
+			goto err;
+		}
+	if (!EVP_DigestFinal_ex(&ctx, H_, NULL))
+		goto err;
 	if (memcmp(H_, H, hLen))
 		{
-		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS, RSA_R_BAD_SIGNATURE);
+		RSAerr(RSA_F_RSA_VERIFY_PKCS1_PSS_MGF1, RSA_R_BAD_SIGNATURE);
 		ret = 0;
 		}
 	else 
@@ -161,6 +178,7 @@ int RSA_verify_PKCS1_PSS(RSA *rsa, const unsigned char *mHash,
 	err:
 	if (DB)
 		OPENSSL_free(DB);
+	EVP_MD_CTX_cleanup(&ctx);
 
 	return ret;
 
@@ -170,13 +188,25 @@ int RSA_padding_add_PKCS1_PSS(RSA *rsa, unsigned char *EM,
 			const unsigned char *mHash,
 			const EVP_MD *Hash, int sLen)
 	{
+	return RSA_padding_add_PKCS1_PSS_mgf1(rsa, EM, mHash, Hash, NULL, sLen);
+	}
+
+int RSA_padding_add_PKCS1_PSS_mgf1(RSA *rsa, unsigned char *EM,
+			const unsigned char *mHash,
+			const EVP_MD *Hash, const EVP_MD *mgf1Hash, int sLen)
+	{
 	int i;
 	int ret = 0;
 	int hLen, maskedDBLen, MSBits, emLen;
 	unsigned char *H, *salt = NULL, *p;
 	EVP_MD_CTX ctx;
 
-	hLen = M_EVP_MD_size(Hash);
+	if (mgf1Hash == NULL)
+		mgf1Hash = Hash;
+
+	hLen = EVP_MD_size(Hash);
+	if (hLen < 0)
+		goto err;
 	/*
 	 * Negative sLen has special meanings:
 	 *	-1	sLen == hLen
@@ -187,7 +217,7 @@ int RSA_padding_add_PKCS1_PSS(RSA *rsa, unsigned char *EM,
 	else if (sLen == -2)	sLen = -2;
 	else if (sLen < -2)
 		{
-		RSAerr(RSA_F_RSA_PADDING_ADD_PKCS1_PSS, RSA_R_SLEN_CHECK_FAILED);
+		RSAerr(RSA_F_RSA_PADDING_ADD_PKCS1_PSS_MGF1, RSA_R_SLEN_CHECK_FAILED);
 		goto err;
 		}
 
@@ -204,8 +234,7 @@ int RSA_padding_add_PKCS1_PSS(RSA *rsa, unsigned char *EM,
 		}
 	else if (emLen < (hLen + sLen + 2))
 		{
-		RSAerr(RSA_F_RSA_PADDING_ADD_PKCS1_PSS,
-		   RSA_R_DATA_TOO_LARGE_FOR_KEY_SIZE);
+		RSAerr(RSA_F_RSA_PADDING_ADD_PKCS1_PSS_MGF1,RSA_R_DATA_TOO_LARGE_FOR_KEY_SIZE);
 		goto err;
 		}
 	if (sLen > 0)
@@ -213,8 +242,7 @@ int RSA_padding_add_PKCS1_PSS(RSA *rsa, unsigned char *EM,
 		salt = OPENSSL_malloc(sLen);
 		if (!salt)
 			{
-			RSAerr(RSA_F_RSA_PADDING_ADD_PKCS1_PSS,
-		   		ERR_R_MALLOC_FAILURE);
+			RSAerr(RSA_F_RSA_PADDING_ADD_PKCS1_PSS_MGF1,ERR_R_MALLOC_FAILURE);
 			goto err;
 			}
 		if (RAND_bytes(salt, sLen) <= 0)
@@ -223,16 +251,19 @@ int RSA_padding_add_PKCS1_PSS(RSA *rsa, unsigned char *EM,
 	maskedDBLen = emLen - hLen - 1;
 	H = EM + maskedDBLen;
 	EVP_MD_CTX_init(&ctx);
-	EVP_DigestInit_ex(&ctx, Hash, NULL);
-	EVP_DigestUpdate(&ctx, zeroes, sizeof zeroes);
-	EVP_DigestUpdate(&ctx, mHash, hLen);
-	if (sLen)
-		EVP_DigestUpdate(&ctx, salt, sLen);
-	EVP_DigestFinal(&ctx, H, NULL);
+	if (!EVP_DigestInit_ex(&ctx, Hash, NULL)
+		|| !EVP_DigestUpdate(&ctx, zeroes, sizeof zeroes)
+		|| !EVP_DigestUpdate(&ctx, mHash, hLen))
+		goto err;
+	if (sLen && !EVP_DigestUpdate(&ctx, salt, sLen))
+		goto err;
+	if (!EVP_DigestFinal_ex(&ctx, H, NULL))
+		goto err;
 	EVP_MD_CTX_cleanup(&ctx);
 
 	/* Generate dbMask in place then perform XOR on it */
-	PKCS1_MGF1(EM, maskedDBLen, H, hLen, Hash);
+	if (PKCS1_MGF1(EM, maskedDBLen, H, hLen, mgf1Hash))
+		goto err;
 
 	p = EM;
 
diff --git a/crypto/rsa/rsa_sign.c b/crypto/rsa/rsa_sign.c
index 743dfd76501e..b6f6037ae002 100644
--- a/crypto/rsa/rsa_sign.c
+++ b/crypto/rsa/rsa_sign.c
@@ -62,6 +62,7 @@
 #include <openssl/rsa.h>
 #include <openssl/objects.h>
 #include <openssl/x509.h>
+#include "rsa_locl.h"
 
 /* Size of an SSL signature: MD5+SHA1 */
 #define SSL_SIG_LENGTH	36
@@ -76,6 +77,14 @@ int RSA_sign(int type, const unsigned char *m, unsigned int m_len,
 	const unsigned char *s = NULL;
 	X509_ALGOR algor;
 	ASN1_OCTET_STRING digest;
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(rsa->meth->flags & RSA_FLAG_FIPS_METHOD)
+			&& !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
+		{
+		RSAerr(RSA_F_RSA_SIGN, RSA_R_NON_FIPS_RSA_METHOD);
+		return 0;
+		}
+#endif
 	if((rsa->flags & RSA_FLAG_SIGN_VER) && rsa->meth->rsa_sign)
 		{
 		return rsa->meth->rsa_sign(type, m, m_len,
@@ -90,14 +99,6 @@ int RSA_sign(int type, const unsigned char *m, unsigned int m_len,
 		i = SSL_SIG_LENGTH;
 		s = m;
 	} else {
-	/* NB: in FIPS mode block anything that isn't a TLS signature */
-#ifdef OPENSSL_FIPS
-		if(FIPS_mode() && !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
-			{
-			RSAerr(RSA_F_RSA_SIGN, RSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
-			return 0;
-			}
-#endif
 		sig.algor= &algor;
 		sig.algor->algorithm=OBJ_nid2obj(type);
 		if (sig.algor->algorithm == NULL)
@@ -137,12 +138,7 @@ int RSA_sign(int type, const unsigned char *m, unsigned int m_len,
 		i2d_X509_SIG(&sig,&p);
 		s=tmps;
 	}
-#ifdef OPENSSL_FIPS
-	/* Bypass algorithm blocking: this is allowed if we get this far */
-	i=rsa->meth->rsa_priv_enc(i,s,sigret,rsa,RSA_PKCS1_PADDING);
-#else
 	i=RSA_private_encrypt(i,s,sigret,rsa,RSA_PKCS1_PADDING);
-#endif
 	if (i <= 0)
 		ret=0;
 	else
@@ -155,58 +151,75 @@ int RSA_sign(int type, const unsigned char *m, unsigned int m_len,
 	return(ret);
 	}
 
-int RSA_verify(int dtype, const unsigned char *m, unsigned int m_len,
-	     unsigned char *sigbuf, unsigned int siglen, RSA *rsa)
+int int_rsa_verify(int dtype, const unsigned char *m,
+			  unsigned int m_len,
+			  unsigned char *rm, size_t *prm_len,
+			  const unsigned char *sigbuf, size_t siglen,
+			  RSA *rsa)
 	{
 	int i,ret=0,sigtype;
 	unsigned char *s;
 	X509_SIG *sig=NULL;
 
+#ifdef OPENSSL_FIPS
+	if (FIPS_mode() && !(rsa->meth->flags & RSA_FLAG_FIPS_METHOD)
+			&& !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
+		{
+		RSAerr(RSA_F_INT_RSA_VERIFY, RSA_R_NON_FIPS_RSA_METHOD);
+		return 0;
+		}
+#endif
+
 	if (siglen != (unsigned int)RSA_size(rsa))
 		{
-		RSAerr(RSA_F_RSA_VERIFY,RSA_R_WRONG_SIGNATURE_LENGTH);
+		RSAerr(RSA_F_INT_RSA_VERIFY,RSA_R_WRONG_SIGNATURE_LENGTH);
 		return(0);
 		}
 
-	if((rsa->flags & RSA_FLAG_SIGN_VER) && rsa->meth->rsa_verify)
+	if((dtype == NID_md5_sha1) && rm)
 		{
-		return rsa->meth->rsa_verify(dtype, m, m_len,
-			sigbuf, siglen, rsa);
+		i = RSA_public_decrypt((int)siglen,
+					sigbuf,rm,rsa,RSA_PKCS1_PADDING);
+		if (i <= 0)
+			return 0;
+		*prm_len = i;
+		return 1;
 		}
 
 	s=(unsigned char *)OPENSSL_malloc((unsigned int)siglen);
 	if (s == NULL)
 		{
-		RSAerr(RSA_F_RSA_VERIFY,ERR_R_MALLOC_FAILURE);
+		RSAerr(RSA_F_INT_RSA_VERIFY,ERR_R_MALLOC_FAILURE);
 		goto err;
 		}
-	if(dtype == NID_md5_sha1)
-		{
-		if (m_len != SSL_SIG_LENGTH)
-			{
-			RSAerr(RSA_F_RSA_VERIFY,RSA_R_INVALID_MESSAGE_LENGTH);
+	if((dtype == NID_md5_sha1) && (m_len != SSL_SIG_LENGTH) ) {
+			RSAerr(RSA_F_INT_RSA_VERIFY,RSA_R_INVALID_MESSAGE_LENGTH);
 			goto err;
-			}
-		}
-	/* NB: in FIPS mode block anything that isn't a TLS signature */
-#ifdef OPENSSL_FIPS
-	else if(FIPS_mode() && !(rsa->flags & RSA_FLAG_NON_FIPS_ALLOW))
-		{
-		RSAerr(RSA_F_RSA_VERIFY, RSA_R_OPERATION_NOT_ALLOWED_IN_FIPS_MODE);
-		return 0;
-		}
-	/* Bypass algorithm blocking: this is allowed */
-	i=rsa->meth->rsa_pub_dec((int)siglen,sigbuf,s,rsa,RSA_PKCS1_PADDING);
-#else
+	}
 	i=RSA_public_decrypt((int)siglen,sigbuf,s,rsa,RSA_PKCS1_PADDING);
-#endif
 
 	if (i <= 0) goto err;
+	/* Oddball MDC2 case: signature can be OCTET STRING.
+	 * check for correct tag and length octets.
+	 */
+	if (dtype == NID_mdc2 && i == 18 && s[0] == 0x04 && s[1] == 0x10)
+		{
+		if (rm)
+			{
+			memcpy(rm, s + 2, 16);
+			*prm_len = 16;
+			ret = 1;
+			}
+		else if(memcmp(m, s + 2, 16))
+			RSAerr(RSA_F_INT_RSA_VERIFY,RSA_R_BAD_SIGNATURE);
+		else
+			ret = 1;
+		}
 
 	/* Special case: SSL signature */
 	if(dtype == NID_md5_sha1) {
 		if((i != SSL_SIG_LENGTH) || memcmp(s, m, SSL_SIG_LENGTH))
-				RSAerr(RSA_F_RSA_VERIFY,RSA_R_BAD_SIGNATURE);
+				RSAerr(RSA_F_INT_RSA_VERIFY,RSA_R_BAD_SIGNATURE);
 		else ret = 1;
 	} else {
 		const unsigned char *p=s;
@@ -217,7 +230,7 @@ int RSA_verify(int dtype, const unsigned char *m, unsigned int m_len,
 		/* Excess data can be used to create forgeries */
 		if(p != s+i)
 			{
-			RSAerr(RSA_F_RSA_VERIFY,RSA_R_BAD_SIGNATURE);
+			RSAerr(RSA_F_INT_RSA_VERIFY,RSA_R_BAD_SIGNATURE);
 			goto err;
 			}
 
@@ -226,7 +239,7 @@ int RSA_verify(int dtype, const unsigned char *m, unsigned int m_len,
 		if(sig->algor->parameter
 		   && ASN1_TYPE_get(sig->algor->parameter) != V_ASN1_NULL)
 			{
-			RSAerr(RSA_F_RSA_VERIFY,RSA_R_BAD_SIGNATURE);
+			RSAerr(RSA_F_INT_RSA_VERIFY,RSA_R_BAD_SIGNATURE);
 			goto err;
 			}
 
@@ -252,15 +265,30 @@ int RSA_verify(int dtype, const unsigned char *m, unsigned int m_len,
 				}
 			else
 				{
-				RSAerr(RSA_F_RSA_VERIFY,
+				RSAerr(RSA_F_INT_RSA_VERIFY,
 						RSA_R_ALGORITHM_MISMATCH);
 				goto err;
 				}
 			}
-		if (	((unsigned int)sig->digest->length != m_len) ||
+		if (rm)
+			{
+			const EVP_MD *md;
+			md = EVP_get_digestbynid(dtype);
+			if (md && (EVP_MD_size(md) != sig->digest->length))
+				RSAerr(RSA_F_INT_RSA_VERIFY,
+						RSA_R_INVALID_DIGEST_LENGTH);
+			else
+				{
+				memcpy(rm, sig->digest->data,
+							sig->digest->length);
+				*prm_len = sig->digest->length;
+				ret = 1;
+				}
+			}
+		else if (((unsigned int)sig->digest->length != m_len) ||
 			(memcmp(m,sig->digest->data,m_len) != 0))
 			{
-			RSAerr(RSA_F_RSA_VERIFY,RSA_R_BAD_SIGNATURE);
+			RSAerr(RSA_F_INT_RSA_VERIFY,RSA_R_BAD_SIGNATURE);
 			}
 		else
 			ret=1;
@@ -275,3 +303,16 @@ err:
 	return(ret);
 	}
 
+int RSA_verify(int dtype, const unsigned char *m, unsigned int m_len,
+		const unsigned char *sigbuf, unsigned int siglen,
+		RSA *rsa)
+	{
+
+	if((rsa->flags & RSA_FLAG_SIGN_VER) && rsa->meth->rsa_verify)
+		{
+		return rsa->meth->rsa_verify(dtype, m, m_len,
+			sigbuf, siglen, rsa);
+		}
+
+	return int_rsa_verify(dtype, m, m_len, NULL, NULL, sigbuf, siglen, rsa);
+	}
diff --git a/crypto/rsa/rsa_test.c b/crypto/rsa/rsa_test.c
index 4080de8bcf99..c8705a0f6e59 100644
--- a/crypto/rsa/rsa_test.c
+++ b/crypto/rsa/rsa_test.c
@@ -328,7 +328,7 @@ int main(int argc, char *argv[])
 	}
 
     CRYPTO_cleanup_all_ex_data();
-    ERR_remove_state(0);
+    ERR_remove_thread_state(NULL);
 
     CRYPTO_mem_leaks_fp(stderr);
 
diff --git a/crypto/rsa/rsa_x931g.c b/crypto/rsa/rsa_x931g.c
deleted file mode 100644
index bf94f8be7a03..000000000000
--- a/crypto/rsa/rsa_x931g.c
+++ /dev/null
@@ -1,255 +0,0 @@
-/* crypto/rsa/rsa_gen.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-#include <stdio.h>
-#include <string.h>
-#include <time.h>
-#include <openssl/err.h>
-#include <openssl/bn.h>
-#include <openssl/rsa.h>
-
-#ifndef OPENSSL_FIPS
-
-/* X9.31 RSA key derivation and generation */
-
-int RSA_X931_derive_ex(RSA *rsa, BIGNUM *p1, BIGNUM *p2, BIGNUM *q1, BIGNUM *q2,
-			const BIGNUM *Xp1, const BIGNUM *Xp2, const BIGNUM *Xp,
-			const BIGNUM *Xq1, const BIGNUM *Xq2, const BIGNUM *Xq,
-			const BIGNUM *e, BN_GENCB *cb)
-	{
-	BIGNUM *r0=NULL,*r1=NULL,*r2=NULL,*r3=NULL;
-	BN_CTX *ctx=NULL,*ctx2=NULL;
-
-	if (!rsa) 
-		goto err;
-
-	ctx = BN_CTX_new();
-	if (!ctx) 
-		goto err;
-	BN_CTX_start(ctx);
-
-	r0 = BN_CTX_get(ctx);
-	r1 = BN_CTX_get(ctx);
-	r2 = BN_CTX_get(ctx);
-	r3 = BN_CTX_get(ctx);
-
-	if (r3 == NULL)
-		goto err;
-	if (!rsa->e)
-		{
-		rsa->e = BN_dup(e);
-		if (!rsa->e)
-			goto err;
-		}
-	else
-		e = rsa->e;
-
-	/* If not all parameters present only calculate what we can.
-	 * This allows test programs to output selective parameters.
-	 */
-
-	if (Xp && !rsa->p)
-		{
-		rsa->p = BN_new();
-		if (!rsa->p)
-			goto err;
-
-		if (!BN_X931_derive_prime_ex(rsa->p, p1, p2,
-					Xp, Xp1, Xp2, e, ctx, cb))
-			goto err;
-		}
-
-	if (Xq && !rsa->q)
-		{
-		rsa->q = BN_new();
-		if (!rsa->q)
-			goto err;
-		if (!BN_X931_derive_prime_ex(rsa->q, q1, q2,
-					Xq, Xq1, Xq2, e, ctx, cb))
-			goto err;
-		}
-
-	if (!rsa->p || !rsa->q)
-		{
-		BN_CTX_end(ctx);
-		BN_CTX_free(ctx);
-		return 2;
-		}
-
-	/* Since both primes are set we can now calculate all remaining 
-	 * components.
-	 */
-
-	/* calculate n */
-	rsa->n=BN_new();
-	if (rsa->n == NULL)
-		goto err;
-	if (!BN_mul(rsa->n,rsa->p,rsa->q,ctx))
-		goto err;
-
-	/* calculate d */
-	if (!BN_sub(r1,rsa->p,BN_value_one()))
-		goto err;	/* p-1 */
-	if (!BN_sub(r2,rsa->q,BN_value_one()))
-		goto err;	/* q-1 */
-	if (!BN_mul(r0,r1,r2,ctx))
-		goto err;	/* (p-1)(q-1) */
-
-	if (!BN_gcd(r3, r1, r2, ctx))
-		goto err;
-
-	if (!BN_div(r0, NULL, r0, r3, ctx))
-		goto err;	/* LCM((p-1)(q-1)) */
-
-	ctx2 = BN_CTX_new();
-	if (!ctx2)
-		goto err;
-
-	rsa->d=BN_mod_inverse(NULL,rsa->e,r0,ctx2);	/* d */
-	if (rsa->d == NULL)
-		goto err;
-
-	/* calculate d mod (p-1) */
-	rsa->dmp1=BN_new();
-	if (rsa->dmp1 == NULL)
-		goto err;
-	if (!BN_mod(rsa->dmp1,rsa->d,r1,ctx))
-		goto err;
-
-	/* calculate d mod (q-1) */
-	rsa->dmq1=BN_new();
-	if (rsa->dmq1 == NULL)
-		goto err;
-	if (!BN_mod(rsa->dmq1,rsa->d,r2,ctx))
-		goto err;
-
-	/* calculate inverse of q mod p */
-	rsa->iqmp=BN_mod_inverse(NULL,rsa->q,rsa->p,ctx2);
-
-	err:
-	if (ctx)
-		{
-		BN_CTX_end(ctx);
-		BN_CTX_free(ctx);
-		}
-	if (ctx2)
-		BN_CTX_free(ctx2);
-	/* If this is set all calls successful */
-	if (rsa && rsa->iqmp != NULL)
-		return 1;
-
-	return 0;
-
-	}
-
-int RSA_X931_generate_key_ex(RSA *rsa, int bits, const BIGNUM *e, BN_GENCB *cb)
-	{
-	int ok = 0;
-	BIGNUM *Xp = NULL, *Xq = NULL;
-	BN_CTX *ctx = NULL;
-	
-	ctx = BN_CTX_new();
-	if (!ctx)
-		goto error;
-
-	BN_CTX_start(ctx);
-	Xp = BN_CTX_get(ctx);
-	Xq = BN_CTX_get(ctx);
-	if (!BN_X931_generate_Xpq(Xp, Xq, bits, ctx))
-		goto error;
-
-	rsa->p = BN_new();
-	rsa->q = BN_new();
-	if (!rsa->p || !rsa->q)
-		goto error;
-
-	/* Generate two primes from Xp, Xq */
-
-	if (!BN_X931_generate_prime_ex(rsa->p, NULL, NULL, NULL, NULL, Xp,
-					e, ctx, cb))
-		goto error;
-
-	if (!BN_X931_generate_prime_ex(rsa->q, NULL, NULL, NULL, NULL, Xq,
-					e, ctx, cb))
-		goto error;
-
-	/* Since rsa->p and rsa->q are valid this call will just derive
-	 * remaining RSA components.
-	 */
-
-	if (!RSA_X931_derive_ex(rsa, NULL, NULL, NULL, NULL,
-				NULL, NULL, NULL, NULL, NULL, NULL, e, cb))
-		goto error;
-
-	ok = 1;
-
-	error:
-	if (ctx)
-		{
-		BN_CTX_end(ctx);
-		BN_CTX_free(ctx);
-		}
-
-	if (ok)
-		return 1;
-
-	return 0;
-
-	}
-
-#endif
diff --git a/crypto/s390xcap.c b/crypto/s390xcap.c
new file mode 100644
index 000000000000..f2e94ef47e83
--- /dev/null
+++ b/crypto/s390xcap.c
@@ -0,0 +1,37 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <setjmp.h>
+#include <signal.h>
+
+extern unsigned long OPENSSL_s390xcap_P[];
+
+static sigjmp_buf ill_jmp;
+static void ill_handler (int sig) { siglongjmp(ill_jmp,sig); }
+
+unsigned long OPENSSL_s390x_facilities(void);
+
+void OPENSSL_cpuid_setup(void)
+	{
+	sigset_t oset;
+	struct sigaction ill_act,oact;
+
+	if (OPENSSL_s390xcap_P[0]) return;
+
+	OPENSSL_s390xcap_P[0] = 1UL<<(8*sizeof(unsigned long)-1);
+
+	memset(&ill_act,0,sizeof(ill_act));
+	ill_act.sa_handler = ill_handler;
+	sigfillset(&ill_act.sa_mask);
+	sigdelset(&ill_act.sa_mask,SIGILL);
+	sigdelset(&ill_act.sa_mask,SIGTRAP);
+	sigprocmask(SIG_SETMASK,&ill_act.sa_mask,&oset);
+	sigaction (SIGILL,&ill_act,&oact);
+
+	/* protection against missing store-facility-list-extended */
+	if (sigsetjmp(ill_jmp,1) == 0)
+		OPENSSL_s390x_facilities();
+
+	sigaction (SIGILL,&oact,NULL);
+	sigprocmask(SIG_SETMASK,&oset,NULL);
+	}
diff --git a/crypto/s390xcpuid.S b/crypto/s390xcpuid.S
new file mode 100644
index 000000000000..06815347e6a3
--- /dev/null
+++ b/crypto/s390xcpuid.S
@@ -0,0 +1,99 @@
+.text
+
+.globl	OPENSSL_s390x_facilities
+.type	OPENSSL_s390x_facilities,@function
+.align	16
+OPENSSL_s390x_facilities:
+	lghi	%r0,0
+	larl	%r2,OPENSSL_s390xcap_P
+	stg	%r0,8(%r2)
+	.long	0xb2b02000	# stfle	0(%r2)
+	brc	8,.Ldone
+	lghi	%r0,1
+	.long	0xb2b02000	# stfle 0(%r2)
+.Ldone:
+	lg	%r2,0(%r2)
+	br	%r14
+.size	OPENSSL_s390x_facilities,.-OPENSSL_s390x_facilities
+
+.globl	OPENSSL_rdtsc
+.type	OPENSSL_rdtsc,@function
+.align	16
+OPENSSL_rdtsc:
+	stck	16(%r15)
+	lg	%r2,16(%r15)
+	br	%r14
+.size	OPENSSL_rdtsc,.-OPENSSL_rdtsc
+
+.globl	OPENSSL_atomic_add
+.type	OPENSSL_atomic_add,@function
+.align	16
+OPENSSL_atomic_add:
+	l	%r1,0(%r2)
+.Lspin:	lr	%r0,%r1
+	ar	%r0,%r3
+	cs	%r1,%r0,0(%r2)
+	brc	4,.Lspin
+	lgfr	%r2,%r0		# OpenSSL expects the new value
+	br	%r14
+.size	OPENSSL_atomic_add,.-OPENSSL_atomic_add
+
+.globl	OPENSSL_wipe_cpu
+.type	OPENSSL_wipe_cpu,@function
+.align	16
+OPENSSL_wipe_cpu:
+	xgr	%r0,%r0
+	xgr	%r1,%r1
+	lgr	%r2,%r15
+	xgr	%r3,%r3
+	xgr	%r4,%r4
+	lzdr	%f0
+	lzdr	%f1
+	lzdr	%f2
+	lzdr	%f3
+	lzdr	%f4
+	lzdr	%f5
+	lzdr	%f6
+	lzdr	%f7
+	br	%r14
+.size	OPENSSL_wipe_cpu,.-OPENSSL_wipe_cpu
+
+.globl	OPENSSL_cleanse
+.type	OPENSSL_cleanse,@function
+.align	16
+OPENSSL_cleanse:
+#if !defined(__s390x__) && !defined(__s390x)
+	llgfr	%r3,%r3
+#endif
+	lghi	%r4,15
+	lghi	%r0,0
+	clgr	%r3,%r4
+	jh	.Lot
+	clgr	%r3,%r0
+	bcr	8,%r14
+.Little:
+	stc	%r0,0(%r2)
+	la	%r2,1(%r2)
+	brctg	%r3,.Little
+	br	%r14
+.align	4
+.Lot:	tmll	%r2,7
+	jz	.Laligned
+	stc	%r0,0(%r2)
+	la	%r2,1(%r2)
+	brctg	%r3,.Lot
+.Laligned:
+	srlg	%r4,%r3,3
+.Loop:	stg	%r0,0(%r2)
+	la	%r2,8(%r2)
+	brctg	%r4,.Loop
+	lghi	%r4,7
+	ngr	%r3,%r4
+	jnz	.Little
+	br	%r14
+.size	OPENSSL_cleanse,.-OPENSSL_cleanse
+
+.section	.init
+	brasl	%r14,OPENSSL_cpuid_setup
+
+.comm	OPENSSL_s390xcap_P,16,8
diff --git a/crypto/seed/Makefile b/crypto/seed/Makefile
index ffaeb84218db..4bc55e491649 100644
--- a/crypto/seed/Makefile
+++ b/crypto/seed/Makefile
@@ -34,7 +34,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -75,13 +75,32 @@ clean:
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.
 
-seed.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-seed.o: ../../include/openssl/seed.h seed.c seed_locl.h
-seed_cbc.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-seed_cbc.o: ../../include/openssl/seed.h seed_cbc.c seed_locl.h
-seed_cfb.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-seed_cfb.o: ../../include/openssl/seed.h seed_cfb.c seed_locl.h
-seed_ecb.o: ../../include/openssl/opensslconf.h ../../include/openssl/seed.h
+seed.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+seed.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+seed.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+seed.o: ../../include/openssl/seed.h ../../include/openssl/stack.h
+seed.o: ../../include/openssl/symhacks.h seed.c seed_locl.h
+seed_cbc.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+seed_cbc.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+seed_cbc.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+seed_cbc.o: ../../include/openssl/safestack.h ../../include/openssl/seed.h
+seed_cbc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+seed_cbc.o: seed_cbc.c
+seed_cfb.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+seed_cfb.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+seed_cfb.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+seed_cfb.o: ../../include/openssl/safestack.h ../../include/openssl/seed.h
+seed_cfb.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+seed_cfb.o: seed_cfb.c
+seed_ecb.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+seed_ecb.o: ../../include/openssl/opensslconf.h
+seed_ecb.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+seed_ecb.o: ../../include/openssl/safestack.h ../../include/openssl/seed.h
+seed_ecb.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 seed_ecb.o: seed_ecb.c
-seed_ofb.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-seed_ofb.o: ../../include/openssl/seed.h seed_locl.h seed_ofb.c
+seed_ofb.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+seed_ofb.o: ../../include/openssl/modes.h ../../include/openssl/opensslconf.h
+seed_ofb.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+seed_ofb.o: ../../include/openssl/safestack.h ../../include/openssl/seed.h
+seed_ofb.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+seed_ofb.o: seed_ofb.c
diff --git a/crypto/seed/seed.c b/crypto/seed/seed.c
index 125dd7d66f6e..3e675a8d7551 100644
--- a/crypto/seed/seed.c
+++ b/crypto/seed/seed.c
@@ -32,10 +32,15 @@
 #include <memory.h>
 #endif
 
+#include <openssl/crypto.h>
 #include <openssl/seed.h>
 #include "seed_locl.h"
 
-static seed_word SS[4][256] = {	{
+#ifdef SS	/* can get defined on Solaris by inclusion of <stdlib.h> */
+#undef SS
+#endif
+
+static const seed_word SS[4][256] = {	{
 	0x2989a1a8, 0x05858184, 0x16c6d2d4, 0x13c3d3d0, 0x14445054, 0x1d0d111c, 0x2c8ca0ac, 0x25052124,
 	0x1d4d515c, 0x03434340, 0x18081018, 0x1e0e121c, 0x11415150, 0x3cccf0fc, 0x0acac2c8, 0x23436360,
 	0x28082028, 0x04444044, 0x20002020, 0x1d8d919c, 0x20c0e0e0, 0x22c2e2e0, 0x08c8c0c8, 0x17071314,
@@ -187,8 +192,19 @@ static seed_word SS[4][256] = {	{
 #define KC14    0xde6e678d
 #define KC15    0xbcdccf1b
 
-
+#if defined(OPENSSL_SMALL_FOOTPRINT)
+static const seed_word KC[] = {
+	KC0,	KC1,	KC2,	KC3,	KC4,	KC5,	KC6,	KC7,
+	KC8,	KC9,	KC10,	KC11,	KC12,	KC13,	KC14,	KC15	};
+#endif
 void SEED_set_key(const unsigned char rawkey[SEED_KEY_LENGTH], SEED_KEY_SCHEDULE *ks)
+#ifdef OPENSSL_FIPS
+	{
+	fips_cipher_abort(SEED);
+	private_SEED_set_key(rawkey, ks);
+	}
+void private_SEED_set_key(const unsigned char rawkey[SEED_KEY_LENGTH], SEED_KEY_SCHEDULE *ks)
+#endif
 {
 	seed_word x1, x2, x3, x4;
 	seed_word t0, t1;
@@ -201,6 +217,8 @@ void SEED_set_key(const unsigned char rawkey[SEED_KEY_LENGTH], SEED_KEY_SCHEDULE
 	t0 = (x1 + x3 - KC0) & 0xffffffff;
 	t1 = (x2 - x4 + KC0) & 0xffffffff;                     KEYUPDATE_TEMP(t0, t1, &ks->data[0]);
 	KEYSCHEDULE_UPDATE1(t0, t1, x1, x2, x3, x4, KC1);      KEYUPDATE_TEMP(t0, t1, &ks->data[2]);
+
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
 	KEYSCHEDULE_UPDATE0(t0, t1, x1, x2, x3, x4, KC2);      KEYUPDATE_TEMP(t0, t1, &ks->data[4]);
 	KEYSCHEDULE_UPDATE1(t0, t1, x1, x2, x3, x4, KC3);      KEYUPDATE_TEMP(t0, t1, &ks->data[6]);
 	KEYSCHEDULE_UPDATE0(t0, t1, x1, x2, x3, x4, KC4);      KEYUPDATE_TEMP(t0, t1, &ks->data[8]);
@@ -215,6 +233,17 @@ void SEED_set_key(const unsigned char rawkey[SEED_KEY_LENGTH], SEED_KEY_SCHEDULE
 	KEYSCHEDULE_UPDATE1(t0, t1, x1, x2, x3, x4, KC13);     KEYUPDATE_TEMP(t0, t1, &ks->data[26]);
 	KEYSCHEDULE_UPDATE0(t0, t1, x1, x2, x3, x4, KC14);     KEYUPDATE_TEMP(t0, t1, &ks->data[28]);
 	KEYSCHEDULE_UPDATE1(t0, t1, x1, x2, x3, x4, KC15);     KEYUPDATE_TEMP(t0, t1, &ks->data[30]);
+#else
+	{
+	    int i;
+	    for (i=2; i<16; i+=2) {
+		KEYSCHEDULE_UPDATE0(t0, t1, x1, x2, x3, x4, KC[i]);
+		KEYUPDATE_TEMP(t0, t1, &ks->data[i*2]);
+		KEYSCHEDULE_UPDATE1(t0, t1, x1, x2, x3, x4, KC[i+1]);
+		KEYUPDATE_TEMP(t0, t1, &ks->data[i*2+2]);
+	    }
+	}
+#endif
 }
 
 void SEED_encrypt(const unsigned char s[SEED_BLOCK_SIZE], unsigned char d[SEED_BLOCK_SIZE], const SEED_KEY_SCHEDULE *ks)
@@ -226,7 +255,8 @@ void SEED_encrypt(const unsigned char s[SEED_BLOCK_SIZE], unsigned char d[SEED_B
 	char2word(s+4,  x2);
 	char2word(s+8,  x3);
 	char2word(s+12, x4);
-	
+
+#if !defined(OPENSSL_SMALL_FOOTPRINT)	
 	E_SEED(t0, t1, x1, x2, x3, x4, 0);
 	E_SEED(t0, t1, x3, x4, x1, x2, 2);
 	E_SEED(t0, t1, x1, x2, x3, x4, 4);
@@ -243,6 +273,15 @@ void SEED_encrypt(const unsigned char s[SEED_BLOCK_SIZE], unsigned char d[SEED_B
 	E_SEED(t0, t1, x3, x4, x1, x2, 26);
 	E_SEED(t0, t1, x1, x2, x3, x4, 28);
 	E_SEED(t0, t1, x3, x4, x1, x2, 30);
+#else
+	{
+	    int i;
+	    for (i=0;i<30;i+=4) {
+		E_SEED(t0,t1,x1,x2,x3,x4,i);
+		E_SEED(t0,t1,x3,x4,x1,x2,i+2);
+	    }
+	}
+#endif
 
 	word2char(x3, d);
 	word2char(x4, d+4);
@@ -259,7 +298,8 @@ void SEED_decrypt(const unsigned char s[SEED_BLOCK_SIZE], unsigned char d[SEED_B
 	char2word(s+4,  x2);
 	char2word(s+8,  x3);
 	char2word(s+12, x4);
-	
+
+#if !defined(OPENSSL_SMALL_FOOTPRINT)
 	E_SEED(t0, t1, x1, x2, x3, x4, 30);
 	E_SEED(t0, t1, x3, x4, x1, x2, 28);
 	E_SEED(t0, t1, x1, x2, x3, x4, 26);
@@ -276,6 +316,16 @@ void SEED_decrypt(const unsigned char s[SEED_BLOCK_SIZE], unsigned char d[SEED_B
 	E_SEED(t0, t1, x3, x4, x1, x2, 4);
 	E_SEED(t0, t1, x1, x2, x3, x4, 2);
 	E_SEED(t0, t1, x3, x4, x1, x2, 0);
+#else
+	{
+	    int i;
+	    for (i=30; i>0; i-=4) {
+		E_SEED(t0, t1, x1, x2, x3, x4, i);
+		E_SEED(t0, t1, x3, x4, x1, x2, i-2);
+
+	    }
+	}
+#endif
 
 	word2char(x3, d);
 	word2char(x4, d+4);
diff --git a/crypto/seed/seed.h b/crypto/seed/seed.h
index 427915ed9a9f..c50fdd360733 100644
--- a/crypto/seed/seed.h
+++ b/crypto/seed/seed.h
@@ -82,6 +82,8 @@
 #define HEADER_SEED_H
 
 #include <openssl/opensslconf.h>
+#include <openssl/e_os2.h>
+#include <openssl/crypto.h>
 
 #ifdef OPENSSL_NO_SEED
 #error SEED is disabled.
@@ -114,7 +116,9 @@ typedef struct seed_key_st {
 #endif
 } SEED_KEY_SCHEDULE;
 
-
+#ifdef OPENSSL_FIPS
+void private_SEED_set_key(const unsigned char rawkey[SEED_KEY_LENGTH], SEED_KEY_SCHEDULE *ks);
+#endif
 void SEED_set_key(const unsigned char rawkey[SEED_KEY_LENGTH], SEED_KEY_SCHEDULE *ks);
 
 void SEED_encrypt(const unsigned char s[SEED_BLOCK_SIZE], unsigned char d[SEED_BLOCK_SIZE], const SEED_KEY_SCHEDULE *ks);
diff --git a/crypto/seed/seed_cbc.c b/crypto/seed/seed_cbc.c
index 4f718ccb44e6..6c3f9b527af7 100644
--- a/crypto/seed/seed_cbc.c
+++ b/crypto/seed/seed_cbc.c
@@ -49,81 +49,15 @@
  *
  */
 
-#include "seed_locl.h"
-#include <string.h>
+#include <openssl/seed.h>
+#include <openssl/modes.h>
 
 void SEED_cbc_encrypt(const unsigned char *in, unsigned char *out,
                       size_t len, const SEED_KEY_SCHEDULE *ks,
                       unsigned char ivec[SEED_BLOCK_SIZE], int enc)
 	{
-	size_t n;
-	unsigned char tmp[SEED_BLOCK_SIZE];
-	const unsigned char *iv = ivec;
-
 	if (enc)
-		{
-		while (len >= SEED_BLOCK_SIZE)
-			{
-			for (n = 0; n < SEED_BLOCK_SIZE; ++n)
-				out[n] = in[n] ^ iv[n];
-			SEED_encrypt(out, out, ks);
-			iv = out;
-			len -= SEED_BLOCK_SIZE;
-			in  += SEED_BLOCK_SIZE;
-			out += SEED_BLOCK_SIZE;
-			}
-		if (len)
-			{
-			for (n = 0; n < len; ++n)
-				out[n] = in[n] ^ iv[n];
-			for (n = len; n < SEED_BLOCK_SIZE; ++n)
-				out[n] = iv[n];
-			SEED_encrypt(out, out, ks);
-			iv = out;
-			}
-		memcpy(ivec, iv, SEED_BLOCK_SIZE);
-		}
-	else if (in != out) /* decrypt */
-		{
-		while (len >= SEED_BLOCK_SIZE)
-			{
-			SEED_decrypt(in, out, ks);
-			for (n = 0; n < SEED_BLOCK_SIZE; ++n)
-				out[n] ^= iv[n];
-			iv = in;
-			len -= SEED_BLOCK_SIZE;
-			in  += SEED_BLOCK_SIZE;
-			out += SEED_BLOCK_SIZE;
-			}
-		if (len)
-			{
-			SEED_decrypt(in, tmp, ks);
-			for (n = 0; n < len; ++n)
-				out[n] = tmp[n] ^ iv[n];
-			iv = in;
-			}
-		memcpy(ivec, iv, SEED_BLOCK_SIZE);
-		}
-	else /* decrypt, overlap */
-		{
-		while (len >= SEED_BLOCK_SIZE)
-			{
-			memcpy(tmp, in, SEED_BLOCK_SIZE);
-			SEED_decrypt(in, out, ks);
-			for (n = 0; n < SEED_BLOCK_SIZE; ++n)
-				out[n] ^= ivec[n];
-			memcpy(ivec, tmp, SEED_BLOCK_SIZE);
-			len -= SEED_BLOCK_SIZE;
-			in  += SEED_BLOCK_SIZE;
-			out += SEED_BLOCK_SIZE;
-			}
-		if (len)
-			{
-			memcpy(tmp, in, SEED_BLOCK_SIZE);
-			SEED_decrypt(tmp, tmp, ks);
-			for (n = 0; n < len; ++n)
-				out[n] = tmp[n] ^ ivec[n];
-			memcpy(ivec, tmp, SEED_BLOCK_SIZE);
-			}
-		}
+		CRYPTO_cbc128_encrypt(in,out,len,ks,ivec,(block128_f)SEED_encrypt);
+	else
+		CRYPTO_cbc128_decrypt(in,out,len,ks,ivec,(block128_f)SEED_decrypt);
 	}
diff --git a/crypto/seed/seed_cfb.c b/crypto/seed/seed_cfb.c
index 07d878a78889..694597dd06e1 100644
--- a/crypto/seed/seed_cfb.c
+++ b/crypto/seed/seed_cfb.c
@@ -105,40 +105,12 @@
  * [including the GNU Public Licence.]
  */
 
-#include "seed_locl.h"
-#include <string.h>
+#include <openssl/seed.h>
+#include <openssl/modes.h>
 
 void SEED_cfb128_encrypt(const unsigned char *in, unsigned char *out,
                          size_t len, const SEED_KEY_SCHEDULE *ks,
                          unsigned char ivec[SEED_BLOCK_SIZE], int *num, int enc)
 	{
-	int n;
-	unsigned char c;
-
-	n = *num;
-
-	if (enc)
-		{
-		while (len--)
-			{
-			if (n == 0)
-				SEED_encrypt(ivec, ivec, ks);
-			ivec[n] = *(out++) = *(in++) ^ ivec[n];
-			n = (n+1) % SEED_BLOCK_SIZE;
-			}
-		}
-	else
-		{
-		while (len--)
-			{
-			if (n == 0)
-				SEED_encrypt(ivec, ivec, ks);
-			c = *(in);
-			*(out++) = *(in++) ^ ivec[n];
-			ivec[n] = c;
-			n = (n+1) % SEED_BLOCK_SIZE;
-			}
-		}
-
-	*num = n;
+	CRYPTO_cfb128_encrypt(in,out,len,ks,ivec,num,enc,(block128_f)SEED_encrypt);
 	}
diff --git a/crypto/seed/seed_ofb.c b/crypto/seed/seed_ofb.c
index e2f3f57a38cd..3c8ba33bb9f9 100644
--- a/crypto/seed/seed_ofb.c
+++ b/crypto/seed/seed_ofb.c
@@ -105,24 +105,12 @@
  * [including the GNU Public Licence.]
  */
 
-#include "seed_locl.h"
-#include <string.h>
+#include <openssl/seed.h>
+#include <openssl/modes.h>
 
 void SEED_ofb128_encrypt(const unsigned char *in, unsigned char *out,
                          size_t len, const SEED_KEY_SCHEDULE *ks,
                          unsigned char ivec[SEED_BLOCK_SIZE], int *num)
 	{
-	int n;
-
-	n = *num;
-	
-	while (len--)
-		{
-		if (n == 0)
-			SEED_encrypt(ivec, ivec, ks);
-		*(out++) = *(in++) ^ ivec[n];
-		n = (n+1) % SEED_BLOCK_SIZE;
-		}
-
-	*num = n;
+	CRYPTO_ofb128_encrypt(in,out,len,ks,ivec,num,(block128_f)SEED_encrypt);
 	}
diff --git a/crypto/sha/Makefile b/crypto/sha/Makefile
index f4741b9ee63a..6d191d3936ee 100644
--- a/crypto/sha/Makefile
+++ b/crypto/sha/Makefile
@@ -38,25 +38,16 @@ top:
 all:    lib
 
 lib:    $(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
-# ELF
-sx86-elf.s: asm/sha1-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) sha1-586.pl elf $(CFLAGS) $(PROCESSOR) > ../$@)
-s512sse2-elf.s:	asm/sha512-sse2.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) sha512-sse2.pl elf $(CFLAGS) $(PROCESSOR) > ../$@)
-# COFF
-sx86-cof.s: asm/sha1-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) sha1-586.pl coff $(CFLAGS) $(PROCESSOR) > ../$@)
-s512sse2-cof.s:     asm/sha512-sse2.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) sha512-sse2.pl coff $(CFLAGS) $(PROCESSOR) > ../$@)
-# a.out
-sx86-out.s: asm/sha1-586.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) sha1-586.pl a.out $(CFLAGS) $(PROCESSOR) > ../$@)
-s512sse2-out.s:     asm/sha512-sse2.pl ../perlasm/x86asm.pl
-	(cd asm; $(PERL) sha512-sse2.pl a.out $(CFLAGS) $(PROCESSOR) > ../$@)
+sha1-586.s:	asm/sha1-586.pl ../perlasm/x86asm.pl
+	$(PERL) asm/sha1-586.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+sha256-586.s:	asm/sha256-586.pl ../perlasm/x86asm.pl
+	$(PERL) asm/sha256-586.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+sha512-586.s:	asm/sha512-586.pl ../perlasm/x86asm.pl
+	$(PERL) asm/sha512-586.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
 
 sha1-ia64.s:   asm/sha1-ia64.pl
 	(cd asm; $(PERL) sha1-ia64.pl ../$@ $(CFLAGS))
@@ -65,10 +56,40 @@ sha256-ia64.s: asm/sha512-ia64.pl
 sha512-ia64.s: asm/sha512-ia64.pl
 	(cd asm; $(PERL) sha512-ia64.pl ../$@ $(CFLAGS))
 
+sha256-armv4.S: asm/sha256-armv4.pl
+	$(PERL) $< $(PERLASM_SCHEME) $@
+
+sha1-alpha.s:	asm/sha1-alpha.pl
+	$(PERL) $< | $(CC) -E - | tee $@ > /dev/null
+
 # Solaris make has to be explicitly told
-sha1-x86_64.s:	asm/sha1-x86_64.pl;	$(PERL) asm/sha1-x86_64.pl $@
-sha256-x86_64.s:asm/sha512-x86_64.pl;	$(PERL) asm/sha512-x86_64.pl $@
-sha512-x86_64.s:asm/sha512-x86_64.pl;	$(PERL) asm/sha512-x86_64.pl $@
+sha1-x86_64.s:	asm/sha1-x86_64.pl;	$(PERL) asm/sha1-x86_64.pl $(PERLASM_SCHEME) > $@
+sha256-x86_64.s:asm/sha512-x86_64.pl;	$(PERL) asm/sha512-x86_64.pl $(PERLASM_SCHEME) $@
+sha512-x86_64.s:asm/sha512-x86_64.pl;	$(PERL) asm/sha512-x86_64.pl $(PERLASM_SCHEME) $@
+sha1-sparcv9.s:	asm/sha1-sparcv9.pl;	$(PERL) asm/sha1-sparcv9.pl $@ $(CFLAGS)
+sha256-sparcv9.s:asm/sha512-sparcv9.pl;	$(PERL) asm/sha512-sparcv9.pl $@ $(CFLAGS)
+sha512-sparcv9.s:asm/sha512-sparcv9.pl;	$(PERL) asm/sha512-sparcv9.pl $@ $(CFLAGS)
+
+sha1-ppc.s:	asm/sha1-ppc.pl;	$(PERL) asm/sha1-ppc.pl $(PERLASM_SCHEME) $@
+sha256-ppc.s:	asm/sha512-ppc.pl;	$(PERL) asm/sha512-ppc.pl $(PERLASM_SCHEME) $@
+sha512-ppc.s:	asm/sha512-ppc.pl;	$(PERL) asm/sha512-ppc.pl $(PERLASM_SCHEME) $@
+
+sha1-parisc.s:	asm/sha1-parisc.pl;	$(PERL) asm/sha1-parisc.pl $(PERLASM_SCHEME) $@
+sha256-parisc.s:asm/sha512-parisc.pl;	$(PERL) asm/sha512-parisc.pl $(PERLASM_SCHEME) $@
+sha512-parisc.s:asm/sha512-parisc.pl;	$(PERL) asm/sha512-parisc.pl $(PERLASM_SCHEME) $@
+
+sha1-mips.S:	asm/sha1-mips.pl;	$(PERL) asm/sha1-mips.pl $(PERLASM_SCHEME) $@
+sha256-mips.S:	asm/sha512-mips.pl;	$(PERL) asm/sha512-mips.pl $(PERLASM_SCHEME) $@
+sha512-mips.S:	asm/sha512-mips.pl;	$(PERL) asm/sha512-mips.pl $(PERLASM_SCHEME) $@
+
+# GNU make "catch all"
+sha1-%.S:	asm/sha1-%.pl;		$(PERL) $< $(PERLASM_SCHEME) $@
+sha256-%.S:	asm/sha512-%.pl;	$(PERL) $< $(PERLASM_SCHEME) $@
+sha512-%.S:	asm/sha512-%.pl;	$(PERL) $< $(PERLASM_SCHEME) $@
+
+sha1-armv4-large.o:	sha1-armv4-large.S
+sha256-armv4.o:		sha256-armv4.S
+sha512-armv4.o:		sha512-armv4.S
 
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
@@ -113,27 +134,26 @@ sha1_one.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 sha1_one.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 sha1_one.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 sha1_one.o: sha1_one.c
-sha1dgst.o: ../../include/openssl/e_os2.h ../../include/openssl/fips.h
+sha1dgst.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 sha1dgst.o: ../../include/openssl/opensslconf.h
-sha1dgst.o: ../../include/openssl/opensslv.h ../../include/openssl/sha.h
+sha1dgst.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+sha1dgst.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+sha1dgst.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 sha1dgst.o: ../md32_common.h sha1dgst.c sha_locl.h
 sha256.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
-sha256.o: ../../include/openssl/fips.h ../../include/openssl/opensslconf.h
-sha256.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-sha256.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-sha256.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-sha256.o: ../md32_common.h sha256.c
+sha256.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+sha256.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+sha256.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+sha256.o: ../../include/openssl/symhacks.h ../md32_common.h sha256.c
 sha512.o: ../../e_os.h ../../include/openssl/bio.h
 sha512.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 sha512.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-sha512.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-sha512.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-sha512.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
-sha512.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-sha512.o: ../../include/openssl/symhacks.h ../cryptlib.h sha512.c
-sha_dgst.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
-sha_dgst.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
-sha_dgst.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+sha512.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
+sha512.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+sha512.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+sha512.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+sha512.o: ../cryptlib.h sha512.c
+sha_dgst.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 sha_dgst.o: ../../include/openssl/opensslconf.h
 sha_dgst.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 sha_dgst.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
diff --git a/crypto/sha/asm/sha1-586.pl b/crypto/sha/asm/sha1-586.pl
index a787dd37da42..1084d227fe06 100644
--- a/crypto/sha/asm/sha1-586.pl
+++ b/crypto/sha/asm/sha1-586.pl
@@ -12,6 +12,8 @@
 # commentary below], and in 2006 the rest was rewritten in order to
 # gain freedom to liberate licensing terms.
 
+# January, September 2004.
+#
 # It was noted that Intel IA-32 C compiler generates code which
 # performs ~30% *faster* on P4 CPU than original *hand-coded*
 # SHA1 assembler implementation. To address this problem (and
@@ -31,12 +33,92 @@
 # ----------------------------------------------------------------
 #					<appro@fy.chalmers.se>
 
+# August 2009.
+#
+# George Spelvin has tipped that F_40_59(b,c,d) can be rewritten as
+# '(c&d) + (b&(c^d))', which allows to accumulate partial results
+# and lighten "pressure" on scratch registers. This resulted in
+# >12% performance improvement on contemporary AMD cores (with no
+# degradation on other CPUs:-). Also, the code was revised to maximize
+# "distance" between instructions producing input to 'lea' instruction
+# and the 'lea' instruction itself, which is essential for Intel Atom
+# core and resulted in ~15% improvement.
+
+# October 2010.
+#
+# Add SSSE3, Supplemental[!] SSE3, implementation. The idea behind it
+# is to offload message schedule denoted by Wt in NIST specification,
+# or Xupdate in OpenSSL source, to SIMD unit. The idea is not novel,
+# and in SSE2 context was first explored by Dean Gaudet in 2004, see
+# http://arctic.org/~dean/crypto/sha1.html. Since then several things
+# have changed that made it interesting again:
+#
+# a) XMM units became faster and wider;
+# b) instruction set became more versatile;
+# c) an important observation was made by Max Locktykhin, which made
+#    it possible to reduce amount of instructions required to perform
+#    the operation in question, for further details see
+#    http://software.intel.com/en-us/articles/improving-the-performance-of-the-secure-hash-algorithm-1/.
+
+# April 2011.
+#
+# Add AVX code path, probably most controversial... The thing is that
+# switch to AVX alone improves performance by as little as 4% in
+# comparison to SSSE3 code path. But below result doesn't look like
+# 4% improvement... Trouble is that Sandy Bridge decodes 'ro[rl]' as
+# pair of µ-ops, and it's the additional µ-ops, two per round, that
+# make it run slower than Core2 and Westmere. But 'sh[rl]d' is decoded
+# as single µ-op by Sandy Bridge and it's replacing 'ro[rl]' with
+# equivalent 'sh[rl]d' that is responsible for the impressive 5.1
+# cycles per processed byte. But 'sh[rl]d' is not something that used
+# to be fast, nor does it appear to be fast in upcoming Bulldozer
+# [according to its optimization manual]. Which is why AVX code path
+# is guarded by *both* AVX and synthetic bit denoting Intel CPUs.
+# One can argue that it's unfair to AMD, but without 'sh[rl]d' it
+# makes no sense to keep the AVX code path. If somebody feels that
+# strongly, it's probably more appropriate to discuss possibility of
+# using vector rotate XOP on AMD...
+
+######################################################################
+# Current performance is summarized in following table. Numbers are
+# CPU clock cycles spent to process single byte (less is better).
+#
+#		x86		SSSE3		AVX
+# Pentium	15.7		-
+# PIII		11.5		-
+# P4		10.6		-
+# AMD K8	7.1		-
+# Core2		7.3		6.1/+20%	-
+# Atom		12.5		9.5(*)/+32%	-
+# Westmere	7.3		5.6/+30%	-
+# Sandy Bridge	8.8		6.2/+40%	5.1(**)/+70%
+#
+# (*)	Loop is 1056 instructions long and expected result is ~8.25.
+#	It remains mystery [to me] why ILP is limited to 1.7.
+#
+# (**)	As per above comment, the result is for AVX *plus* sh[rl]d.
+
 $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
 push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],"sha1-586.pl",$ARGV[$#ARGV] eq "386");
 
+$xmm=$ymm=0;
+for (@ARGV) { $xmm=1 if (/-DOPENSSL_IA32_SSE2/); }
+
+$ymm=1 if ($xmm &&
+		`$ENV{CC} -Wa,-v -c -o /dev/null -x assembler /dev/null 2>&1`
+			=~ /GNU assembler version ([2-9]\.[0-9]+)/ &&
+		$1>=2.19);	# first version supporting AVX
+
+$ymm=1 if ($xmm && !$ymm && $ARGV[0] eq "win32n" && 
+		`nasm -v 2>&1` =~ /NASM version ([2-9]\.[0-9]+)/ &&
+		$1>=2.03);	# first version supporting AVX
+
+&external_label("OPENSSL_ia32cap_P") if ($xmm);
+
+
 $A="eax";
 $B="ebx";
 $C="ecx";
@@ -47,6 +129,10 @@ $tmp1="ebp";
 
 @V=($A,$B,$C,$D,$E,$T);
 
+$alt=0;	# 1 denotes alternative IALU implementation, which performs
+	# 8% *worse* on P4, same on Westmere and Atom, 2% better on
+	# Sandy Bridge...
+
 sub BODY_00_15
 	{
 	local($n,$a,$b,$c,$d,$e,$f)=@_;
@@ -59,16 +145,18 @@ sub BODY_00_15
 	&rotl($tmp1,5);			# tmp1=ROTATE(a,5)
 	 &xor($f,$d);
 	&add($tmp1,$e);			# tmp1+=e;
-	 &and($f,$b);
-	&mov($e,&swtmp($n%16));		# e becomes volatile and is loaded
+	 &mov($e,&swtmp($n%16));	# e becomes volatile and is loaded
 	 				# with xi, also note that e becomes
 					# f in next round...
-	 &xor($f,$d);			# f holds F_00_19(b,c,d)
+	&and($f,$b);
 	&rotr($b,2);			# b=ROTATE(b,30)
-	 &lea($tmp1,&DWP(0x5a827999,$tmp1,$e));	# tmp1+=K_00_19+xi
+	 &xor($f,$d);			# f holds F_00_19(b,c,d)
+	&lea($tmp1,&DWP(0x5a827999,$tmp1,$e));	# tmp1+=K_00_19+xi
 
-	if ($n==15) { &add($f,$tmp1); }	# f+=tmp1
+	if ($n==15) { &mov($e,&swtmp(($n+1)%16));# pre-fetch f for next round
+		      &add($f,$tmp1); }	# f+=tmp1
 	else        { &add($tmp1,$f); }	# f becomes a in next round
+	&mov($tmp1,$a)			if ($alt && $n==15);
 	}
 
 sub BODY_16_19
@@ -77,22 +165,41 @@ sub BODY_16_19
 
 	&comment("16_19 $n");
 
-	&mov($f,&swtmp($n%16));		# f to hold Xupdate(xi,xa,xb,xc,xd)
-	 &mov($tmp1,$c);		# tmp1 to hold F_00_19(b,c,d)
-	&xor($f,&swtmp(($n+2)%16));
-	 &xor($tmp1,$d);
-	&xor($f,&swtmp(($n+8)%16));
-	 &and($tmp1,$b);		# tmp1 holds F_00_19(b,c,d)
-	&rotr($b,2);			# b=ROTATE(b,30)
+if ($alt) {
+	&xor($c,$d);
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&and($tmp1,$c);			# tmp1 to hold F_00_19(b,c,d), b&=c^d
+	 &xor($f,&swtmp(($n+8)%16));
+	&xor($tmp1,$d);			# tmp1=F_00_19(b,c,d)
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
+	&rotl($f,1);			# f=ROTATE(f,1)
+	 &add($e,$tmp1);		# e+=F_00_19(b,c,d)
+	&xor($c,$d);			# restore $c
+	 &mov($tmp1,$a);		# b in next round
+	&rotr($b,$n==16?2:7);		# b=ROTATE(b,30)
+	 &mov(&swtmp($n%16),$f);	# xi=f
+	&rotl($a,5);			# ROTATE(a,5)
+	 &lea($f,&DWP(0x5a827999,$f,$e));# f+=F_00_19(b,c,d)+e
+	&mov($e,&swtmp(($n+1)%16));	# pre-fetch f for next round
+	 &add($f,$a);			# f+=ROTATE(a,5)
+} else {
+	&mov($tmp1,$c);			# tmp1 to hold F_00_19(b,c,d)
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&xor($tmp1,$d);
+	 &xor($f,&swtmp(($n+8)%16));
+	&and($tmp1,$b);
 	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
 	&rotl($f,1);			# f=ROTATE(f,1)
 	 &xor($tmp1,$d);		# tmp1=F_00_19(b,c,d)
-	&mov(&swtmp($n%16),$f);		# xi=f
-	&lea($f,&DWP(0x5a827999,$f,$e));# f+=K_00_19+e
-	 &mov($e,$a);			# e becomes volatile
-	&rotl($e,5);			# e=ROTATE(a,5)
-	 &add($f,$tmp1);		# f+=F_00_19(b,c,d)
-	&add($f,$e);			# f+=ROTATE(a,5)
+	&add($e,$tmp1);			# e+=F_00_19(b,c,d)
+	 &mov($tmp1,$a);
+	&rotr($b,2);			# b=ROTATE(b,30)
+	 &mov(&swtmp($n%16),$f);	# xi=f
+	&rotl($tmp1,5);			# ROTATE(a,5)
+	 &lea($f,&DWP(0x5a827999,$f,$e));# f+=F_00_19(b,c,d)+e
+	&mov($e,&swtmp(($n+1)%16));	# pre-fetch f for next round
+	 &add($f,$tmp1);		# f+=ROTATE(a,5)
+}
 	}
 
 sub BODY_20_39
@@ -102,21 +209,41 @@ sub BODY_20_39
 
 	&comment("20_39 $n");
 
+if ($alt) {
+	&xor($tmp1,$c);			# tmp1 to hold F_20_39(b,c,d), b^=c
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&xor($tmp1,$d);			# tmp1 holds F_20_39(b,c,d)
+	 &xor($f,&swtmp(($n+8)%16));
+	&add($e,$tmp1);			# e+=F_20_39(b,c,d)
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
+	&rotl($f,1);			# f=ROTATE(f,1)
+	 &mov($tmp1,$a);		# b in next round
+	&rotr($b,7);			# b=ROTATE(b,30)
+	 &mov(&swtmp($n%16),$f)		if($n<77);# xi=f
+	&rotl($a,5);			# ROTATE(a,5)
+	 &xor($b,$c)			if($n==39);# warm up for BODY_40_59
+	&and($tmp1,$b)			if($n==39);
+	 &lea($f,&DWP($K,$f,$e));	# f+=e+K_XX_YY
+	&mov($e,&swtmp(($n+1)%16))	if($n<79);# pre-fetch f for next round
+	 &add($f,$a);			# f+=ROTATE(a,5)
+	&rotr($a,5)			if ($n==79);
+} else {
 	&mov($tmp1,$b);			# tmp1 to hold F_20_39(b,c,d)
-	 &mov($f,&swtmp($n%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
-	&rotr($b,2);			# b=ROTATE(b,30)
-	 &xor($f,&swtmp(($n+2)%16));
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
 	&xor($tmp1,$c);
 	 &xor($f,&swtmp(($n+8)%16));
 	&xor($tmp1,$d);			# tmp1 holds F_20_39(b,c,d)
 	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
 	&rotl($f,1);			# f=ROTATE(f,1)
-	 &add($tmp1,$e);
-	&mov(&swtmp($n%16),$f);		# xi=f
-	 &mov($e,$a);			# e becomes volatile
-	&rotl($e,5);			# e=ROTATE(a,5)
-	 &lea($f,&DWP($K,$f,$tmp1));	# f+=K_20_39+e
-	&add($f,$e);			# f+=ROTATE(a,5)
+	 &add($e,$tmp1);		# e+=F_20_39(b,c,d)
+	&rotr($b,2);			# b=ROTATE(b,30)
+	 &mov($tmp1,$a);
+	&rotl($tmp1,5);			# ROTATE(a,5)
+	 &mov(&swtmp($n%16),$f) if($n<77);# xi=f
+	&lea($f,&DWP($K,$f,$e));	# f+=e+K_XX_YY
+	 &mov($e,&swtmp(($n+1)%16)) if($n<79);# pre-fetch f for next round
+	&add($f,$tmp1);			# f+=ROTATE(a,5)
+}
 	}
 
 sub BODY_40_59
@@ -125,41 +252,86 @@ sub BODY_40_59
 
 	&comment("40_59 $n");
 
-	&mov($f,&swtmp($n%16));		# f to hold Xupdate(xi,xa,xb,xc,xd)
-	 &mov($tmp1,&swtmp(($n+2)%16));
-	&xor($f,$tmp1);
-	 &mov($tmp1,&swtmp(($n+8)%16));
-	&xor($f,$tmp1);
-	 &mov($tmp1,&swtmp(($n+13)%16));
-	&xor($f,$tmp1);			# f holds xa^xb^xc^xd
-	 &mov($tmp1,$b);		# tmp1 to hold F_40_59(b,c,d)
+if ($alt) {
+	&add($e,$tmp1);			# e+=b&(c^d)
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&mov($tmp1,$d);
+	 &xor($f,&swtmp(($n+8)%16));
+	&xor($c,$d);			# restore $c
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
 	&rotl($f,1);			# f=ROTATE(f,1)
-	 &or($tmp1,$c);
-	&mov(&swtmp($n%16),$f);		# xi=f
-	 &and($tmp1,$d);
-	&lea($f,&DWP(0x8f1bbcdc,$f,$e));# f+=K_40_59+e
-	 &mov($e,$b);			# e becomes volatile and is used
-					# to calculate F_40_59(b,c,d)
+	 &and($tmp1,$c);
+	&rotr($b,7);			# b=ROTATE(b,30)
+	 &add($e,$tmp1);		# e+=c&d
+	&mov($tmp1,$a);			# b in next round
+	 &mov(&swtmp($n%16),$f);	# xi=f
+	&rotl($a,5);			# ROTATE(a,5)
+	 &xor($b,$c)			if ($n<59);
+	&and($tmp1,$b)			if ($n<59);# tmp1 to hold F_40_59(b,c,d)
+	 &lea($f,&DWP(0x8f1bbcdc,$f,$e));# f+=K_40_59+e+(b&(c^d))
+	&mov($e,&swtmp(($n+1)%16));	# pre-fetch f for next round
+	 &add($f,$a);			# f+=ROTATE(a,5)
+} else {
+	&mov($tmp1,$c);			# tmp1 to hold F_40_59(b,c,d)
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&xor($tmp1,$d);
+	 &xor($f,&swtmp(($n+8)%16));
+	&and($tmp1,$b);
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
+	&rotl($f,1);			# f=ROTATE(f,1)
+	 &add($tmp1,$e);		# b&(c^d)+=e
 	&rotr($b,2);			# b=ROTATE(b,30)
-	 &and($e,$c);
-	&or($tmp1,$e);			# tmp1 holds F_40_59(b,c,d)		
-	 &mov($e,$a);
-	&rotl($e,5);			# e=ROTATE(a,5)
-	 &add($f,$tmp1);		# f+=tmp1;
+	 &mov($e,$a);			# e becomes volatile
+	&rotl($e,5);			# ROTATE(a,5)
+	 &mov(&swtmp($n%16),$f);	# xi=f
+	&lea($f,&DWP(0x8f1bbcdc,$f,$tmp1));# f+=K_40_59+e+(b&(c^d))
+	 &mov($tmp1,$c);
 	&add($f,$e);			# f+=ROTATE(a,5)
+	 &and($tmp1,$d);
+	&mov($e,&swtmp(($n+1)%16));	# pre-fetch f for next round
+	 &add($f,$tmp1);		# f+=c&d
+}
 	}
 
 &function_begin("sha1_block_data_order");
+if ($xmm) {
+  &static_label("ssse3_shortcut");
+  &static_label("avx_shortcut")		if ($ymm);
+  &static_label("K_XX_XX");
+
+	&call	(&label("pic_point"));	# make it PIC!
+  &set_label("pic_point");
+	&blindpop($tmp1);
+	&picmeup($T,"OPENSSL_ia32cap_P",$tmp1,&label("pic_point"));
+	&lea	($tmp1,&DWP(&label("K_XX_XX")."-".&label("pic_point"),$tmp1));
+
+	&mov	($A,&DWP(0,$T));
+	&mov	($D,&DWP(4,$T));
+	&test	($D,1<<9);		# check SSSE3 bit
+	&jz	(&label("x86"));
+	&test	($A,1<<24);		# check FXSR bit
+	&jz	(&label("x86"));
+	if ($ymm) {
+		&and	($D,1<<28);		# mask AVX bit
+		&and	($A,1<<30);		# mask "Intel CPU" bit
+		&or	($A,$D);
+		&cmp	($A,1<<28|1<<30);
+		&je	(&label("avx_shortcut"));
+	}
+	&jmp	(&label("ssse3_shortcut"));
+  &set_label("x86",16);
+}
 	&mov($tmp1,&wparam(0));	# SHA_CTX *c
 	&mov($T,&wparam(1));	# const void *input
 	&mov($A,&wparam(2));	# size_t num
-	&stack_push(16);	# allocate X[16]
+	&stack_push(16+3);	# allocate X[16]
 	&shl($A,6);
 	&add($A,$T);
 	&mov(&wparam(2),$A);	# pointer beyond the end of input
 	&mov($E,&DWP(16,$tmp1));# pre-load E
+	&jmp(&label("loop"));
 
-	&set_label("loop",16);
+&set_label("loop",16);
 
 	# copy input chunk to X, but reversing byte order!
 	for ($i=0; $i<16; $i+=4)
@@ -213,7 +385,845 @@ sub BODY_40_59
 	&mov(&DWP(16,$tmp1),$C);
 	&jb(&label("loop"));
 
-	&stack_pop(16);
+	&stack_pop(16+3);
 &function_end("sha1_block_data_order");
 
+if ($xmm) {
+######################################################################
+# The SSSE3 implementation.
+#
+# %xmm[0-7] are used as ring @X[] buffer containing quadruples of last
+# 32 elements of the message schedule or Xupdate outputs. First 4
+# quadruples are simply byte-swapped input, next 4 are calculated
+# according to method originally suggested by Dean Gaudet (modulo
+# being implemented in SSSE3). Once 8 quadruples or 32 elements are
+# collected, it switches to routine proposed by Max Locktyukhin.
+#
+# Calculations inevitably require temporary reqisters, and there are
+# no %xmm registers left to spare. For this reason part of the ring
+# buffer, X[2..4] to be specific, is offloaded to 3 quadriples ring
+# buffer on the stack. Keep in mind that X[2] is alias X[-6], X[3] -
+# X[-5], and X[4] - X[-4]...
+#
+# Another notable optimization is aggressive stack frame compression
+# aiming to minimize amount of 9-byte instructions...
+#
+# Yet another notable optimization is "jumping" $B variable. It means
+# that there is no register permanently allocated for $B value. This
+# allowed to eliminate one instruction from body_20_39...
+#
+my $Xi=4;			# 4xSIMD Xupdate round, start pre-seeded
+my @X=map("xmm$_",(4..7,0..3));	# pre-seeded for $Xi=4
+my @V=($A,$B,$C,$D,$E);
+my $j=0;			# hash round
+my @T=($T,$tmp1);
+my $inp;
+
+my $_rol=sub { &rol(@_) };
+my $_ror=sub { &ror(@_) };
+
+&function_begin("_sha1_block_data_order_ssse3");
+	&call	(&label("pic_point"));	# make it PIC!
+	&set_label("pic_point");
+	&blindpop($tmp1);
+	&lea	($tmp1,&DWP(&label("K_XX_XX")."-".&label("pic_point"),$tmp1));
+&set_label("ssse3_shortcut");
+
+	&movdqa	(@X[3],&QWP(0,$tmp1));		# K_00_19
+	&movdqa	(@X[4],&QWP(16,$tmp1));		# K_20_39
+	&movdqa	(@X[5],&QWP(32,$tmp1));		# K_40_59
+	&movdqa	(@X[6],&QWP(48,$tmp1));		# K_60_79
+	&movdqa	(@X[2],&QWP(64,$tmp1));		# pbswap mask
+
+	&mov	($E,&wparam(0));		# load argument block
+	&mov	($inp=@T[1],&wparam(1));
+	&mov	($D,&wparam(2));
+	&mov	(@T[0],"esp");
+
+	# stack frame layout
+	#
+	# +0	X[0]+K	X[1]+K	X[2]+K	X[3]+K	# XMM->IALU xfer area
+	#	X[4]+K	X[5]+K	X[6]+K	X[7]+K
+	#	X[8]+K	X[9]+K	X[10]+K	X[11]+K
+	#	X[12]+K	X[13]+K	X[14]+K	X[15]+K
+	#
+	# +64	X[0]	X[1]	X[2]	X[3]	# XMM->XMM backtrace area
+	#	X[4]	X[5]	X[6]	X[7]
+	#	X[8]	X[9]	X[10]	X[11]	# even borrowed for K_00_19
+	#
+	# +112	K_20_39	K_20_39	K_20_39	K_20_39	# constants
+	#	K_40_59	K_40_59	K_40_59	K_40_59
+	#	K_60_79	K_60_79	K_60_79	K_60_79
+	#	K_00_19	K_00_19	K_00_19	K_00_19
+	#	pbswap mask
+	#
+	# +192	ctx				# argument block
+	# +196	inp
+	# +200	end
+	# +204	esp
+	&sub	("esp",208);
+	&and	("esp",-64);
+
+	&movdqa	(&QWP(112+0,"esp"),@X[4]);	# copy constants
+	&movdqa	(&QWP(112+16,"esp"),@X[5]);
+	&movdqa	(&QWP(112+32,"esp"),@X[6]);
+	&shl	($D,6);				# len*64
+	&movdqa	(&QWP(112+48,"esp"),@X[3]);
+	&add	($D,$inp);			# end of input
+	&movdqa	(&QWP(112+64,"esp"),@X[2]);
+	&add	($inp,64);
+	&mov	(&DWP(192+0,"esp"),$E);		# save argument block
+	&mov	(&DWP(192+4,"esp"),$inp);
+	&mov	(&DWP(192+8,"esp"),$D);
+	&mov	(&DWP(192+12,"esp"),@T[0]);	# save original %esp
+
+	&mov	($A,&DWP(0,$E));		# load context
+	&mov	($B,&DWP(4,$E));
+	&mov	($C,&DWP(8,$E));
+	&mov	($D,&DWP(12,$E));
+	&mov	($E,&DWP(16,$E));
+	&mov	(@T[0],$B);			# magic seed
+
+	&movdqu	(@X[-4&7],&QWP(-64,$inp));	# load input to %xmm[0-3]
+	&movdqu	(@X[-3&7],&QWP(-48,$inp));
+	&movdqu	(@X[-2&7],&QWP(-32,$inp));
+	&movdqu	(@X[-1&7],&QWP(-16,$inp));
+	&pshufb	(@X[-4&7],@X[2]);		# byte swap
+	&pshufb	(@X[-3&7],@X[2]);
+	&pshufb	(@X[-2&7],@X[2]);
+	&movdqa	(&QWP(112-16,"esp"),@X[3]);	# borrow last backtrace slot
+	&pshufb	(@X[-1&7],@X[2]);
+	&paddd	(@X[-4&7],@X[3]);		# add K_00_19
+	&paddd	(@X[-3&7],@X[3]);
+	&paddd	(@X[-2&7],@X[3]);
+	&movdqa	(&QWP(0,"esp"),@X[-4&7]);	# X[]+K xfer to IALU
+	&psubd	(@X[-4&7],@X[3]);		# restore X[]
+	&movdqa	(&QWP(0+16,"esp"),@X[-3&7]);
+	&psubd	(@X[-3&7],@X[3]);
+	&movdqa	(&QWP(0+32,"esp"),@X[-2&7]);
+	&psubd	(@X[-2&7],@X[3]);
+	&movdqa	(@X[0],@X[-3&7]);
+	&jmp	(&label("loop"));
+
+######################################################################
+# SSE instruction sequence is first broken to groups of indepentent
+# instructions, independent in respect to their inputs and shifter
+# (not all architectures have more than one). Then IALU instructions
+# are "knitted in" between the SSE groups. Distance is maintained for
+# SSE latency of 2 in hope that it fits better upcoming AMD Bulldozer
+# [which allegedly also implements SSSE3]...
+#
+# Temporary registers usage. X[2] is volatile at the entry and at the
+# end is restored from backtrace ring buffer. X[3] is expected to
+# contain current K_XX_XX constant and is used to caclulate X[-1]+K
+# from previous round, it becomes volatile the moment the value is
+# saved to stack for transfer to IALU. X[4] becomes volatile whenever
+# X[-4] is accumulated and offloaded to backtrace ring buffer, at the
+# end it is loaded with next K_XX_XX [which becomes X[3] in next
+# round]...
+#
+sub Xupdate_ssse3_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&palignr(@X[0],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	&movdqa	(@X[2],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &paddd	(@X[3],@X[-1&7]);
+	  &movdqa	(&QWP(64+16*(($Xi-4)%3),"esp"),@X[-4&7]);# save X[] to backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psrldq	(@X[2],4);		# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[2],@X[-2&7]);	# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@X[2]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&movdqa	(@X[4],@X[0]);
+	&movdqa	(@X[2],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslldq	(@X[4],12);		# "X[0]"<<96, extract one dword
+	&paddd	(@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@X[2],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(@X[3],@X[4]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@X[4],30);
+	&por	(@X[0],@X[2]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(@X[2],&QWP(64+16*(($Xi-6)%3),"esp")) if ($Xi>5);	# restore X[] from backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslld	(@X[3],2);
+	&pxor	(@X[0],@X[4]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(@X[4],&QWP(112-16+16*(($Xi)/5),"esp"));	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@X[3]);		# "X[0]"^=("X[0]"<<96)<<<2
+	  &movdqa	(@X[1],@X[-2&7])	if ($Xi<7);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+}
+
+sub Xupdate_ssse3_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&movdqa	(@X[2],@X[-1&7])	if ($Xi==8);
+	 eval(shift(@insns));		# body_20_39
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"="X[-32]"^"X[-16]"
+	&palignr(@X[2],@X[-2&7],8);	# compose "X[-6]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&pxor	(@X[0],@X[-7&7]);	# "X[0]"^="X[-28]"
+	  &movdqa	(&QWP(64+16*(($Xi-4)%3),"esp"),@X[-4&7]);	# save X[] to backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 if ($Xi%5) {
+	  &movdqa	(@X[4],@X[3]);	# "perpetuate" K_XX_XX...
+	 } else {			# ... or load next one
+	  &movdqa	(@X[4],&QWP(112-16+16*($Xi/5),"esp"));
+	 }
+	  &paddd	(@X[3],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@X[2]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&movdqa	(@X[2],@X[0]);
+	  &movdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pslld	(@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	&psrld	(@X[2],30);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&por	(@X[0],@X[2]);		# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &movdqa	(@X[2],&QWP(64+16*(($Xi-6)%3),"esp")) if($Xi<19);	# restore X[] from backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	  &movdqa	(@X[3],@X[0])	if ($Xi<19);
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+}
+
+sub Xuplast_ssse3_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &paddd	(@X[3],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &movdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&mov	($inp=@T[1],&DWP(192+4,"esp"));
+	&cmp	($inp,&DWP(192+8,"esp"));
+	&je	(&label("done"));
+
+	&movdqa	(@X[3],&QWP(112+48,"esp"));	# K_00_19
+	&movdqa	(@X[2],&QWP(112+64,"esp"));	# pbswap mask
+	&movdqu	(@X[-4&7],&QWP(0,$inp));	# load input
+	&movdqu	(@X[-3&7],&QWP(16,$inp));
+	&movdqu	(@X[-2&7],&QWP(32,$inp));
+	&movdqu	(@X[-1&7],&QWP(48,$inp));
+	&add	($inp,64);
+	&pshufb	(@X[-4&7],@X[2]);		# byte swap
+	&mov	(&DWP(192+4,"esp"),$inp);
+	&movdqa	(&QWP(112-16,"esp"),@X[3]);	# borrow last backtrace slot
+
+  $Xi=0;
+}
+
+sub Xloop_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pshufb	(@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&paddd	(@X[($Xi-4)&7],@X[3]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(&QWP(0+16*$Xi,"esp"),@X[($Xi-4)&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psubd	(@X[($Xi-4)&7],@X[3]);
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+sub body_00_19 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,&DWP(4*($j&15),"esp"));',	# X[]+K xfer
+	'&xor	($c,$d);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&xor	($c,$d);',	# restore $c
+	'&xor	(@T[0],$d);',
+	'&add	($e,$a);',
+	'&$_ror	($b,$j?7:2);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'$j++; unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+sub body_20_39 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,&DWP(4*($j++&15),"esp"));',	# X[]+K xfer
+	'&xor	(@T[0],$d);',	# ($b^$d)
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&xor	(@T[0],$c);',	# ($b^$d^$c)
+	'&add	($e,$a);',
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+sub body_40_59 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&mov	(@T[1],$c);',
+	'&xor	($c,$d);',
+	'&add	($e,&DWP(4*($j++&15),"esp"));',	# X[]+K xfer
+	'&and	(@T[1],$d);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[1]);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&add	($e,@T[0]);',
+	'&xor	($c,$d);',	# restore $c
+	'&add	($e,$a);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+&set_label("loop",16);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xuplast_ssse3_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+
+	&mov	(@T[1],&DWP(192,"esp"));	# update context
+	&add	($A,&DWP(0,@T[1]));
+	&add	(@T[0],&DWP(4,@T[1]));		# $b
+	&add	($C,&DWP(8,@T[1]));
+	&mov	(&DWP(0,@T[1]),$A);
+	&add	($D,&DWP(12,@T[1]));
+	&mov	(&DWP(4,@T[1]),@T[0]);
+	&add	($E,&DWP(16,@T[1]));
+	&mov	(&DWP(8,@T[1]),$C);
+	&mov	($B,@T[0]);
+	&mov	(&DWP(12,@T[1]),$D);
+	&mov	(&DWP(16,@T[1]),$E);
+	&movdqa	(@X[0],@X[-3&7]);
+
+	&jmp	(&label("loop"));
+
+&set_label("done",16);		$j=$saved_j; @V=@saved_V;
+
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+
+	&mov	(@T[1],&DWP(192,"esp"));	# update context
+	&add	($A,&DWP(0,@T[1]));
+	&mov	("esp",&DWP(192+12,"esp"));	# restore %esp
+	&add	(@T[0],&DWP(4,@T[1]));		# $b
+	&add	($C,&DWP(8,@T[1]));
+	&mov	(&DWP(0,@T[1]),$A);
+	&add	($D,&DWP(12,@T[1]));
+	&mov	(&DWP(4,@T[1]),@T[0]);
+	&add	($E,&DWP(16,@T[1]));
+	&mov	(&DWP(8,@T[1]),$C);
+	&mov	(&DWP(12,@T[1]),$D);
+	&mov	(&DWP(16,@T[1]),$E);
+
+&function_end("_sha1_block_data_order_ssse3");
+
+if ($ymm) {
+my $Xi=4;			# 4xSIMD Xupdate round, start pre-seeded
+my @X=map("xmm$_",(4..7,0..3));	# pre-seeded for $Xi=4
+my @V=($A,$B,$C,$D,$E);
+my $j=0;			# hash round
+my @T=($T,$tmp1);
+my $inp;
+
+my $_rol=sub { &shld(@_[0],@_) };
+my $_ror=sub { &shrd(@_[0],@_) };
+
+&function_begin("_sha1_block_data_order_avx");
+	&call	(&label("pic_point"));	# make it PIC!
+	&set_label("pic_point");
+	&blindpop($tmp1);
+	&lea	($tmp1,&DWP(&label("K_XX_XX")."-".&label("pic_point"),$tmp1));
+&set_label("avx_shortcut");
+	&vzeroall();
+
+	&vmovdqa(@X[3],&QWP(0,$tmp1));		# K_00_19
+	&vmovdqa(@X[4],&QWP(16,$tmp1));		# K_20_39
+	&vmovdqa(@X[5],&QWP(32,$tmp1));		# K_40_59
+	&vmovdqa(@X[6],&QWP(48,$tmp1));		# K_60_79
+	&vmovdqa(@X[2],&QWP(64,$tmp1));		# pbswap mask
+
+	&mov	($E,&wparam(0));		# load argument block
+	&mov	($inp=@T[1],&wparam(1));
+	&mov	($D,&wparam(2));
+	&mov	(@T[0],"esp");
+
+	# stack frame layout
+	#
+	# +0	X[0]+K	X[1]+K	X[2]+K	X[3]+K	# XMM->IALU xfer area
+	#	X[4]+K	X[5]+K	X[6]+K	X[7]+K
+	#	X[8]+K	X[9]+K	X[10]+K	X[11]+K
+	#	X[12]+K	X[13]+K	X[14]+K	X[15]+K
+	#
+	# +64	X[0]	X[1]	X[2]	X[3]	# XMM->XMM backtrace area
+	#	X[4]	X[5]	X[6]	X[7]
+	#	X[8]	X[9]	X[10]	X[11]	# even borrowed for K_00_19
+	#
+	# +112	K_20_39	K_20_39	K_20_39	K_20_39	# constants
+	#	K_40_59	K_40_59	K_40_59	K_40_59
+	#	K_60_79	K_60_79	K_60_79	K_60_79
+	#	K_00_19	K_00_19	K_00_19	K_00_19
+	#	pbswap mask
+	#
+	# +192	ctx				# argument block
+	# +196	inp
+	# +200	end
+	# +204	esp
+	&sub	("esp",208);
+	&and	("esp",-64);
+
+	&vmovdqa(&QWP(112+0,"esp"),@X[4]);	# copy constants
+	&vmovdqa(&QWP(112+16,"esp"),@X[5]);
+	&vmovdqa(&QWP(112+32,"esp"),@X[6]);
+	&shl	($D,6);				# len*64
+	&vmovdqa(&QWP(112+48,"esp"),@X[3]);
+	&add	($D,$inp);			# end of input
+	&vmovdqa(&QWP(112+64,"esp"),@X[2]);
+	&add	($inp,64);
+	&mov	(&DWP(192+0,"esp"),$E);		# save argument block
+	&mov	(&DWP(192+4,"esp"),$inp);
+	&mov	(&DWP(192+8,"esp"),$D);
+	&mov	(&DWP(192+12,"esp"),@T[0]);	# save original %esp
+
+	&mov	($A,&DWP(0,$E));		# load context
+	&mov	($B,&DWP(4,$E));
+	&mov	($C,&DWP(8,$E));
+	&mov	($D,&DWP(12,$E));
+	&mov	($E,&DWP(16,$E));
+	&mov	(@T[0],$B);			# magic seed
+
+	&vmovdqu(@X[-4&7],&QWP(-64,$inp));	# load input to %xmm[0-3]
+	&vmovdqu(@X[-3&7],&QWP(-48,$inp));
+	&vmovdqu(@X[-2&7],&QWP(-32,$inp));
+	&vmovdqu(@X[-1&7],&QWP(-16,$inp));
+	&vpshufb(@X[-4&7],@X[-4&7],@X[2]);	# byte swap
+	&vpshufb(@X[-3&7],@X[-3&7],@X[2]);
+	&vpshufb(@X[-2&7],@X[-2&7],@X[2]);
+	&vmovdqa(&QWP(112-16,"esp"),@X[3]);	# borrow last backtrace slot
+	&vpshufb(@X[-1&7],@X[-1&7],@X[2]);
+	&vpaddd	(@X[0],@X[-4&7],@X[3]);		# add K_00_19
+	&vpaddd	(@X[1],@X[-3&7],@X[3]);
+	&vpaddd	(@X[2],@X[-2&7],@X[3]);
+	&vmovdqa(&QWP(0,"esp"),@X[0]);		# X[]+K xfer to IALU
+	&vmovdqa(&QWP(0+16,"esp"),@X[1]);
+	&vmovdqa(&QWP(0+32,"esp"),@X[2]);
+	&jmp	(&label("loop"));
+
+sub Xupdate_avx_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpalignr(@X[0],@X[-3&7],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &vpaddd	(@X[3],@X[3],@X[-1&7]);
+	  &vmovdqa	(&QWP(64+16*(($Xi-4)%3),"esp"),@X[-4&7]);# save X[] to backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpsrldq(@X[2],@X[-1&7],4);		# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpxor	(@X[0],@X[0],@X[-4&7]);		# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[2],@X[2],@X[-2&7]);		# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@X[2]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@X[2],@X[0],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslldq(@X[4],@X[0],12);		# "X[0]"<<96, extract one dword
+	&vpaddd	(@X[0],@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@X[3],@X[4],30);
+	&vpor	(@X[0],@X[0],@X[2]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslld	(@X[4],@X[4],2);
+	  &vmovdqa	(@X[2],&QWP(64+16*(($Xi-6)%3),"esp")) if ($Xi>5);	# restore X[] from backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpxor	(@X[0],@X[0],@X[3]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@X[4]);		# "X[0]"^=("X[0]"<<96)<<<2
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(@X[4],&QWP(112-16+16*(($Xi)/5),"esp"));	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+}
+
+sub Xupdate_avx_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&vpalignr(@X[2],@X[-1&7],@X[-2&7],8);	# compose "X[-6]"
+	&vpxor	(@X[0],@X[0],@X[-4&7]);	# "X[0]"="X[-32]"^"X[-16]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpxor	(@X[0],@X[0],@X[-7&7]);	# "X[0]"^="X[-28]"
+	  &vmovdqa	(&QWP(64+16*(($Xi-4)%3),"esp"),@X[-4&7]);	# save X[] to backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 if ($Xi%5) {
+	  &vmovdqa	(@X[4],@X[3]);	# "perpetuate" K_XX_XX...
+	 } else {			# ... or load next one
+	  &vmovdqa	(@X[4],&QWP(112-16+16*($Xi/5),"esp"));
+	 }
+	  &vpaddd	(@X[3],@X[3],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@X[2]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpsrld	(@X[2],@X[0],30);
+	  &vmovdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpslld	(@X[0],@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpor	(@X[0],@X[0],@X[2]);	# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &vmovdqa	(@X[2],&QWP(64+16*(($Xi-6)%3),"esp")) if($Xi<19);	# restore X[] from backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+}
+
+sub Xuplast_avx_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &vpaddd	(@X[3],@X[3],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &vmovdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&mov	($inp=@T[1],&DWP(192+4,"esp"));
+	&cmp	($inp,&DWP(192+8,"esp"));
+	&je	(&label("done"));
+
+	&vmovdqa(@X[3],&QWP(112+48,"esp"));	# K_00_19
+	&vmovdqa(@X[2],&QWP(112+64,"esp"));	# pbswap mask
+	&vmovdqu(@X[-4&7],&QWP(0,$inp));	# load input
+	&vmovdqu(@X[-3&7],&QWP(16,$inp));
+	&vmovdqu(@X[-2&7],&QWP(32,$inp));
+	&vmovdqu(@X[-1&7],&QWP(48,$inp));
+	&add	($inp,64);
+	&vpshufb(@X[-4&7],@X[-4&7],@X[2]);		# byte swap
+	&mov	(&DWP(192+4,"esp"),$inp);
+	&vmovdqa(&QWP(112-16,"esp"),@X[3]);	# borrow last backtrace slot
+
+  $Xi=0;
+}
+
+sub Xloop_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpshufb	(@X[($Xi-3)&7],@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpaddd	(@X[$Xi&7],@X[($Xi-4)&7],@X[3]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vmovdqa	(&QWP(0+16*$Xi,"esp"),@X[$Xi&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+&set_label("loop",16);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xuplast_avx_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+
+	&mov	(@T[1],&DWP(192,"esp"));	# update context
+	&add	($A,&DWP(0,@T[1]));
+	&add	(@T[0],&DWP(4,@T[1]));		# $b
+	&add	($C,&DWP(8,@T[1]));
+	&mov	(&DWP(0,@T[1]),$A);
+	&add	($D,&DWP(12,@T[1]));
+	&mov	(&DWP(4,@T[1]),@T[0]);
+	&add	($E,&DWP(16,@T[1]));
+	&mov	(&DWP(8,@T[1]),$C);
+	&mov	($B,@T[0]);
+	&mov	(&DWP(12,@T[1]),$D);
+	&mov	(&DWP(16,@T[1]),$E);
+
+	&jmp	(&label("loop"));
+
+&set_label("done",16);		$j=$saved_j; @V=@saved_V;
+
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+
+	&vzeroall();
+
+	&mov	(@T[1],&DWP(192,"esp"));	# update context
+	&add	($A,&DWP(0,@T[1]));
+	&mov	("esp",&DWP(192+12,"esp"));	# restore %esp
+	&add	(@T[0],&DWP(4,@T[1]));		# $b
+	&add	($C,&DWP(8,@T[1]));
+	&mov	(&DWP(0,@T[1]),$A);
+	&add	($D,&DWP(12,@T[1]));
+	&mov	(&DWP(4,@T[1]),@T[0]);
+	&add	($E,&DWP(16,@T[1]));
+	&mov	(&DWP(8,@T[1]),$C);
+	&mov	(&DWP(12,@T[1]),$D);
+	&mov	(&DWP(16,@T[1]),$E);
+&function_end("_sha1_block_data_order_avx");
+}
+&set_label("K_XX_XX",64);
+&data_word(0x5a827999,0x5a827999,0x5a827999,0x5a827999);	# K_00_19
+&data_word(0x6ed9eba1,0x6ed9eba1,0x6ed9eba1,0x6ed9eba1);	# K_20_39
+&data_word(0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc);	# K_40_59
+&data_word(0xca62c1d6,0xca62c1d6,0xca62c1d6,0xca62c1d6);	# K_60_79
+&data_word(0x00010203,0x04050607,0x08090a0b,0x0c0d0e0f);	# pbswap mask
+}
+&asciz("SHA1 block transform for x86, CRYPTOGAMS by <appro\@openssl.org>");
+
 &asm_finish();
diff --git a/crypto/sha/asm/sha1-armv4-large.pl b/crypto/sha/asm/sha1-armv4-large.pl
new file mode 100755
index 000000000000..fe8207f77f8c
--- /dev/null
+++ b/crypto/sha/asm/sha1-armv4-large.pl
@@ -0,0 +1,248 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# sha1_block procedure for ARMv4.
+#
+# January 2007.
+
+# Size/performance trade-off
+# ====================================================================
+# impl		size in bytes	comp cycles[*]	measured performance
+# ====================================================================
+# thumb		304		3212		4420
+# armv4-small	392/+29%	1958/+64%	2250/+96%
+# armv4-compact	740/+89%	1552/+26%	1840/+22%
+# armv4-large	1420/+92%	1307/+19%	1370/+34%[***]
+# full unroll	~5100/+260%	~1260/+4%	~1300/+5%
+# ====================================================================
+# thumb		= same as 'small' but in Thumb instructions[**] and
+#		  with recurring code in two private functions;
+# small		= detached Xload/update, loops are folded;
+# compact	= detached Xload/update, 5x unroll;
+# large		= interleaved Xload/update, 5x unroll;
+# full unroll	= interleaved Xload/update, full unroll, estimated[!];
+#
+# [*]	Manually counted instructions in "grand" loop body. Measured
+#	performance is affected by prologue and epilogue overhead,
+#	i-cache availability, branch penalties, etc.
+# [**]	While each Thumb instruction is twice smaller, they are not as
+#	diverse as ARM ones: e.g., there are only two arithmetic
+#	instructions with 3 arguments, no [fixed] rotate, addressing
+#	modes are limited. As result it takes more instructions to do
+#	the same job in Thumb, therefore the code is never twice as
+#	small and always slower.
+# [***]	which is also ~35% better than compiler generated code. Dual-
+#	issue Cortex A8 core was measured to process input block in
+#	~990 cycles.
+
+# August 2010.
+#
+# Rescheduling for dual-issue pipeline resulted in 13% improvement on
+# Cortex A8 core and in absolute terms ~870 cycles per input block
+# [or 13.6 cycles per byte].
+
+# February 2011.
+#
+# Profiler-assisted and platform-specific optimization resulted in 10%
+# improvement on Cortex A8 core and 12.2 cycles per byte.
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$ctx="r0";
+$inp="r1";
+$len="r2";
+$a="r3";
+$b="r4";
+$c="r5";
+$d="r6";
+$e="r7";
+$K="r8";
+$t0="r9";
+$t1="r10";
+$t2="r11";
+$t3="r12";
+$Xi="r14";
+@V=($a,$b,$c,$d,$e);
+
+sub Xupdate {
+my ($a,$b,$c,$d,$e,$opt1,$opt2)=@_;
+$code.=<<___;
+	ldr	$t0,[$Xi,#15*4]
+	ldr	$t1,[$Xi,#13*4]
+	ldr	$t2,[$Xi,#7*4]
+	add	$e,$K,$e,ror#2			@ E+=K_xx_xx
+	ldr	$t3,[$Xi,#2*4]
+	eor	$t0,$t0,$t1
+	eor	$t2,$t2,$t3			@ 1 cycle stall
+	eor	$t1,$c,$d			@ F_xx_xx
+	mov	$t0,$t0,ror#31
+	add	$e,$e,$a,ror#27			@ E+=ROR(A,27)
+	eor	$t0,$t0,$t2,ror#31
+	str	$t0,[$Xi,#-4]!
+	$opt1					@ F_xx_xx
+	$opt2					@ F_xx_xx
+	add	$e,$e,$t0			@ E+=X[i]
+___
+}
+
+sub BODY_00_15 {
+my ($a,$b,$c,$d,$e)=@_;
+$code.=<<___;
+#if __ARM_ARCH__<7
+	ldrb	$t1,[$inp,#2]
+	ldrb	$t0,[$inp,#3]
+	ldrb	$t2,[$inp,#1]
+	add	$e,$K,$e,ror#2			@ E+=K_00_19
+	ldrb	$t3,[$inp],#4
+	orr	$t0,$t0,$t1,lsl#8
+	eor	$t1,$c,$d			@ F_xx_xx
+	orr	$t0,$t0,$t2,lsl#16
+	add	$e,$e,$a,ror#27			@ E+=ROR(A,27)
+	orr	$t0,$t0,$t3,lsl#24
+#else
+	ldr	$t0,[$inp],#4			@ handles unaligned
+	add	$e,$K,$e,ror#2			@ E+=K_00_19
+	eor	$t1,$c,$d			@ F_xx_xx
+	add	$e,$e,$a,ror#27			@ E+=ROR(A,27)
+#ifdef __ARMEL__
+	rev	$t0,$t0				@ byte swap
+#endif
+#endif
+	and	$t1,$b,$t1,ror#2
+	add	$e,$e,$t0			@ E+=X[i]
+	eor	$t1,$t1,$d,ror#2		@ F_00_19(B,C,D)
+	str	$t0,[$Xi,#-4]!
+	add	$e,$e,$t1			@ E+=F_00_19(B,C,D)
+___
+}
+
+sub BODY_16_19 {
+my ($a,$b,$c,$d,$e)=@_;
+	&Xupdate(@_,"and $t1,$b,$t1,ror#2");
+$code.=<<___;
+	eor	$t1,$t1,$d,ror#2		@ F_00_19(B,C,D)
+	add	$e,$e,$t1			@ E+=F_00_19(B,C,D)
+___
+}
+
+sub BODY_20_39 {
+my ($a,$b,$c,$d,$e)=@_;
+	&Xupdate(@_,"eor $t1,$b,$t1,ror#2");
+$code.=<<___;
+	add	$e,$e,$t1			@ E+=F_20_39(B,C,D)
+___
+}
+
+sub BODY_40_59 {
+my ($a,$b,$c,$d,$e)=@_;
+	&Xupdate(@_,"and $t1,$b,$t1,ror#2","and $t2,$c,$d");
+$code.=<<___;
+	add	$e,$e,$t1			@ E+=F_40_59(B,C,D)
+	add	$e,$e,$t2,ror#2
+___
+}
+
+$code=<<___;
+#include "arm_arch.h"
+
+.text
+
+.global	sha1_block_data_order
+.type	sha1_block_data_order,%function
+
+.align	2
+sha1_block_data_order:
+	stmdb	sp!,{r4-r12,lr}
+	add	$len,$inp,$len,lsl#6	@ $len to point at the end of $inp
+	ldmia	$ctx,{$a,$b,$c,$d,$e}
+.Lloop:
+	ldr	$K,.LK_00_19
+	mov	$Xi,sp
+	sub	sp,sp,#15*4
+	mov	$c,$c,ror#30
+	mov	$d,$d,ror#30
+	mov	$e,$e,ror#30		@ [6]
+.L_00_15:
+___
+for($i=0;$i<5;$i++) {
+	&BODY_00_15(@V);	unshift(@V,pop(@V));
+}
+$code.=<<___;
+	teq	$Xi,sp
+	bne	.L_00_15		@ [((11+4)*5+2)*3]
+___
+	&BODY_00_15(@V);	unshift(@V,pop(@V));
+	&BODY_16_19(@V);	unshift(@V,pop(@V));
+	&BODY_16_19(@V);	unshift(@V,pop(@V));
+	&BODY_16_19(@V);	unshift(@V,pop(@V));
+	&BODY_16_19(@V);	unshift(@V,pop(@V));
+$code.=<<___;
+
+	ldr	$K,.LK_20_39		@ [+15+16*4]
+	sub	sp,sp,#25*4
+	cmn	sp,#0			@ [+3], clear carry to denote 20_39
+.L_20_39_or_60_79:
+___
+for($i=0;$i<5;$i++) {
+	&BODY_20_39(@V);	unshift(@V,pop(@V));
+}
+$code.=<<___;
+	teq	$Xi,sp			@ preserve carry
+	bne	.L_20_39_or_60_79	@ [+((12+3)*5+2)*4]
+	bcs	.L_done			@ [+((12+3)*5+2)*4], spare 300 bytes
+
+	ldr	$K,.LK_40_59
+	sub	sp,sp,#20*4		@ [+2]
+.L_40_59:
+___
+for($i=0;$i<5;$i++) {
+	&BODY_40_59(@V);	unshift(@V,pop(@V));
+}
+$code.=<<___;
+	teq	$Xi,sp
+	bne	.L_40_59		@ [+((12+5)*5+2)*4]
+
+	ldr	$K,.LK_60_79
+	sub	sp,sp,#20*4
+	cmp	sp,#0			@ set carry to denote 60_79
+	b	.L_20_39_or_60_79	@ [+4], spare 300 bytes
+.L_done:
+	add	sp,sp,#80*4		@ "deallocate" stack frame
+	ldmia	$ctx,{$K,$t0,$t1,$t2,$t3}
+	add	$a,$K,$a
+	add	$b,$t0,$b
+	add	$c,$t1,$c,ror#2
+	add	$d,$t2,$d,ror#2
+	add	$e,$t3,$e,ror#2
+	stmia	$ctx,{$a,$b,$c,$d,$e}
+	teq	$inp,$len
+	bne	.Lloop			@ [+18], total 1307
+
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r12,pc}
+#else
+	ldmia	sp!,{r4-r12,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.align	2
+.LK_00_19:	.word	0x5a827999
+.LK_20_39:	.word	0x6ed9eba1
+.LK_40_59:	.word	0x8f1bbcdc
+.LK_60_79:	.word	0xca62c1d6
+.size	sha1_block_data_order,.-sha1_block_data_order
+.asciz	"SHA1 block transform for ARMv4, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+___
+
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
+print $code;
+close STDOUT; # enforce flush
diff --git a/crypto/sha/asm/sha1-ia64.pl b/crypto/sha/asm/sha1-ia64.pl
index 51c4f47ecbdc..db28f0805a11 100644
--- a/crypto/sha/asm/sha1-ia64.pl
+++ b/crypto/sha/asm/sha1-ia64.pl
@@ -15,7 +15,7 @@
 # is >50% better than HP C and >2x better than gcc.
 
 $code=<<___;
-.ident  \"sha1-ia64.s, version 1.2\"
+.ident  \"sha1-ia64.s, version 1.3\"
 .ident  \"IA-64 ISA artwork by Andy Polyakov <appro\@fy.chalmers.se>\"
 .explicit
 
@@ -26,14 +26,10 @@ if ($^O eq "hpux") {
     $ADDP="addp4";
     for (@ARGV) { $ADDP="add" if (/[\+DD|\-mlp]64/); }
 } else { $ADDP="add"; }
-for (@ARGV) {	$big_endian=1 if (/\-DB_ENDIAN/);
-		$big_endian=0 if (/\-DL_ENDIAN/);   }
-if (!defined($big_endian))
-	    {	$big_endian=(unpack('L',pack('N',1))==1);   }
 
 #$human=1;
 if ($human) {	# useful for visual code auditing...
-	($A,$B,$C,$D,$E,$T)   = ("A","B","C","D","E","T");
+	($A,$B,$C,$D,$E)   = ("A","B","C","D","E");
 	($h0,$h1,$h2,$h3,$h4) = ("h0","h1","h2","h3","h4");
 	($K_00_19, $K_20_39, $K_40_59, $K_60_79) =
 	    (	"K_00_19","K_20_39","K_40_59","K_60_79"	);
@@ -41,47 +37,50 @@ if ($human) {	# useful for visual code auditing...
 		"X8", "X9","X10","X11","X12","X13","X14","X15"	);
 }
 else {
-	($A,$B,$C,$D,$E,$T)   = ("loc0","loc1","loc2","loc3","loc4","loc5");
-	($h0,$h1,$h2,$h3,$h4) = ("loc6","loc7","loc8","loc9","loc10");
+	($A,$B,$C,$D,$E)   =    ("loc0","loc1","loc2","loc3","loc4");
+	($h0,$h1,$h2,$h3,$h4) = ("loc5","loc6","loc7","loc8","loc9");
 	($K_00_19, $K_20_39, $K_40_59, $K_60_79) =
-	    (	"r14", "r15", "loc11", "loc12"	);
+	    (	"r14", "r15", "loc10", "loc11"	);
 	@X= (	"r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
 		"r24", "r25", "r26", "r27", "r28", "r29", "r30", "r31"	);
 }
 
 sub BODY_00_15 {
 local	*code=shift;
-local	($i,$a,$b,$c,$d,$e,$f)=@_;
+my	($i,$a,$b,$c,$d,$e)=@_;
+my	$j=$i+1;
+my	$Xn=@X[$j%16];
 
 $code.=<<___ if ($i==0);
-{ .mmi;	ld1	$X[$i&0xf]=[inp],2	    // MSB
+{ .mmi;	ld1	$X[$i]=[inp],2		    // MSB
 	ld1	tmp2=[tmp3],2		};;
 { .mmi;	ld1	tmp0=[inp],2
 	ld1	tmp4=[tmp3],2		    // LSB
-	dep	$X[$i&0xf]=$X[$i&0xf],tmp2,8,8	};;
+	dep	$X[$i]=$X[$i],tmp2,8,8	};;
 ___
 if ($i<15) {
 	$code.=<<___;
-{ .mmi;	ld1	$X[($i+1)&0xf]=[inp],2	    // +1
+{ .mmi;	ld1	$Xn=[inp],2		    // forward Xload
+	nop.m	0x0
 	dep	tmp1=tmp0,tmp4,8,8	};;
-{ .mmi;	ld1	tmp2=[tmp3],2		    // +1
+{ .mmi;	ld1	tmp2=[tmp3],2		    // forward Xload
 	and	tmp4=$c,$b
-	dep	$X[$i&0xf]=$X[$i&0xf],tmp1,16,16	} //;;
-{ .mmi;	andcm	tmp1=$d,$b
-	add	tmp0=$e,$K_00_19
+	dep	$X[$i]=$X[$i],tmp1,16,16} //;;
+{ .mmi;	add	$e=$e,$K_00_19		    // e+=K_00_19
+	andcm	tmp1=$d,$b
 	dep.z	tmp5=$a,5,27		};; // a<<5
-{ .mmi;	or	tmp4=tmp4,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
-	add	$f=tmp0,$X[$i&0xf]	    // f=xi+e+K_00_19
+{ .mmi;	add	$e=$e,$X[$i]		    // e+=Xload
+	or	tmp4=tmp4,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
 	extr.u	tmp1=$a,27,5		};; // a>>27
-{ .mmi;	ld1	tmp0=[inp],2		    // +1
-	add	$f=$f,tmp4		    // f+=F_00_19(b,c,d)
+{ .mmi;	ld1	tmp0=[inp],2		    // forward Xload
+	add	$e=$e,tmp4		    // e+=F_00_19(b,c,d)
 	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
-{ .mmi;	ld1	tmp4=[tmp3],2		    // +1
+{ .mmi;	ld1	tmp4=[tmp3],2		    // forward Xload
 	or	tmp5=tmp1,tmp5		    // ROTATE(a,5)
 	mux2	tmp6=$a,0x44		};; // see b in next iteration
-{ .mii;	add	$f=$f,tmp5		    // f+=ROTATE(a,5)
-	dep	$X[($i+1)&0xf]=$X[($i+1)&0xf],tmp2,8,8	// +1
-	mux2	$X[$i&0xf]=$X[$i&0xf],0x44	} //;;
+{ .mii;	add	$e=$e,tmp5		    // e+=ROTATE(a,5)
+	dep	$Xn=$Xn,tmp2,8,8	    // forward Xload
+	mux2	$X[$i]=$X[$i],0x44	} //;;
 
 ___
 	}
@@ -89,24 +88,24 @@ else	{
 	$code.=<<___;
 { .mii;	and	tmp3=$c,$b
 	dep	tmp1=tmp0,tmp4,8,8;;
-	dep	$X[$i&0xf]=$X[$i&0xf],tmp1,16,16	} //;;
-{ .mmi;	andcm	tmp1=$d,$b
-	add	tmp0=$e,$K_00_19
+	dep	$X[$i]=$X[$i],tmp1,16,16} //;;
+{ .mmi;	add	$e=$e,$K_00_19		    // e+=K_00_19
+	andcm	tmp1=$d,$b
 	dep.z	tmp5=$a,5,27		};; // a<<5
-{ .mmi;	or	tmp4=tmp3,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
-	add	$f=tmp0,$X[$i&0xf]	    // f=xi+e+K_00_19
+{ .mmi;	add	$e=$e,$X[$i]		    // e+=Xupdate
+	or	tmp4=tmp3,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
 	extr.u	tmp1=$a,27,5		}   // a>>27
-{ .mmi;	xor	tmp2=$X[($i+0+1)&0xf],$X[($i+2+1)&0xf]	// +1
-	xor	tmp3=$X[($i+8+1)&0xf],$X[($i+13+1)&0xf] // +1
+{ .mmi;	xor	$Xn=$Xn,$X[($j+2)%16]	    // forward Xupdate
+	xor	tmp3=$X[($j+8)%16],$X[($j+13)%16] // forward Xupdate
 	nop.i	0			};;
-{ .mmi;	add	$f=$f,tmp4		    // f+=F_00_19(b,c,d)
-	xor	tmp2=tmp2,tmp3		    // +1
+{ .mmi;	add	$e=$e,tmp4		    // e+=F_00_19(b,c,d)
+	xor	$Xn=$Xn,tmp3		    // forward Xupdate
 	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
 { .mmi; or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
 	mux2	tmp6=$a,0x44		};; // see b in next iteration
-{ .mii;	add	$f=$f,tmp1		    // f+=ROTATE(a,5)
-	shrp	$e=tmp2,tmp2,31		    // f+1=ROTATE(x[0]^x[2]^x[8]^x[13],1)
-	mux2	$X[$i&0xf]=$X[$i&0xf],0x44  };;
+{ .mii;	add	$e=$e,tmp1		    // e+=ROTATE(a,5)
+	shrp	$Xn=$Xn,$Xn,31		    // ROTATE(x[0]^x[2]^x[8]^x[13],1)
+	mux2	$X[$i]=$X[$i],0x44	};;
 
 ___
 	}
@@ -114,27 +113,28 @@ ___
 
 sub BODY_16_19 {
 local	*code=shift;
-local	($i,$a,$b,$c,$d,$e,$f)=@_;
+my	($i,$a,$b,$c,$d,$e)=@_;
+my	$j=$i+1;
+my	$Xn=@X[$j%16];
 
 $code.=<<___;
-{ .mmi;	mov	$X[$i&0xf]=$f		    // Xupdate
-	and	tmp0=$c,$b
+{ .mib;	add	$e=$e,$K_00_19		    // e+=K_00_19
 	dep.z	tmp5=$a,5,27		}   // a<<5
-{ .mmi;	andcm	tmp1=$d,$b
-	add	tmp4=$e,$K_00_19	};;
-{ .mmi;	or	tmp0=tmp0,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
-	add	$f=$f,tmp4		    // f+=e+K_00_19
+{ .mib;	andcm	tmp1=$d,$b
+	and	tmp0=$c,$b		};;
+{ .mmi;	add	$e=$e,$X[$i%16]		    // e+=Xupdate
+	or	tmp0=tmp0,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
 	extr.u	tmp1=$a,27,5		}   // a>>27
-{ .mmi;	xor	tmp2=$X[($i+0+1)&0xf],$X[($i+2+1)&0xf]	// +1
-	xor	tmp3=$X[($i+8+1)&0xf],$X[($i+13+1)&0xf]	// +1
+{ .mmi;	xor	$Xn=$Xn,$X[($j+2)%16]	    // forward Xupdate
+	xor	tmp3=$X[($j+8)%16],$X[($j+13)%16]	// forward Xupdate
 	nop.i	0			};;
-{ .mmi;	add	$f=$f,tmp0		    // f+=F_00_19(b,c,d)
-	xor	tmp2=tmp2,tmp3		    // +1
+{ .mmi;	add	$e=$e,tmp0		    // f+=F_00_19(b,c,d)
+	xor	$Xn=$Xn,tmp3		    // forward Xupdate
 	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
 { .mmi;	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
 	mux2	tmp6=$a,0x44		};; // see b in next iteration
-{ .mii;	add	$f=$f,tmp1		    // f+=ROTATE(a,5)
-	shrp	$e=tmp2,tmp2,31		    // f+1=ROTATE(x[0]^x[2]^x[8]^x[13],1)
+{ .mii;	add	$e=$e,tmp1		    // e+=ROTATE(a,5)
+	shrp	$Xn=$Xn,$Xn,31		    // ROTATE(x[0]^x[2]^x[8]^x[13],1)
 	nop.i	0			};;
 
 ___
@@ -142,49 +142,47 @@ ___
 
 sub BODY_20_39 {
 local	*code=shift;
-local	($i,$a,$b,$c,$d,$e,$f,$Konst)=@_;
+my	($i,$a,$b,$c,$d,$e,$Konst)=@_;
 	$Konst = $K_20_39 if (!defined($Konst));
+my	$j=$i+1;
+my	$Xn=@X[$j%16];
 
 if ($i<79) {
 $code.=<<___;
-{ .mib;	mov	$X[$i&0xf]=$f		    // Xupdate
+{ .mib;	add	$e=$e,$Konst		    // e+=K_XX_XX
 	dep.z	tmp5=$a,5,27		}   // a<<5
 { .mib;	xor	tmp0=$c,$b
-	add	tmp4=$e,$Konst		};;
-{ .mmi;	xor	tmp0=tmp0,$d		    // F_20_39(b,c,d)=b^c^d
-	add	$f=$f,tmp4		    // f+=e+K_20_39
+	xor	$Xn=$Xn,$X[($j+2)%16]	};; // forward Xupdate
+{ .mib;	add	$e=$e,$X[$i%16]		    // e+=Xupdate
 	extr.u	tmp1=$a,27,5		}   // a>>27
-{ .mmi;	xor	tmp2=$X[($i+0+1)&0xf],$X[($i+2+1)&0xf]	// +1
-	xor	tmp3=$X[($i+8+1)&0xf],$X[($i+13+1)&0xf]	// +1
-	nop.i	0			};;
-{ .mmi;	add	$f=$f,tmp0		    // f+=F_20_39(b,c,d)
-	xor	tmp2=tmp2,tmp3		    // +1
+{ .mib;	xor	tmp0=tmp0,$d		    // F_20_39(b,c,d)=b^c^d
+	xor	$Xn=$Xn,$X[($j+8)%16]	};; // forward Xupdate
+{ .mmi;	add	$e=$e,tmp0		    // e+=F_20_39(b,c,d)
+	xor	$Xn=$Xn,$X[($j+13)%16]	    // forward Xupdate
 	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
 { .mmi;	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
 	mux2	tmp6=$a,0x44		};; // see b in next iteration
-{ .mii;	add	$f=$f,tmp1		    // f+=ROTATE(a,5)
-	shrp	$e=tmp2,tmp2,31		    // f+1=ROTATE(x[0]^x[2]^x[8]^x[13],1)
+{ .mii;	add	$e=$e,tmp1		    // e+=ROTATE(a,5)
+	shrp	$Xn=$Xn,$Xn,31		    // ROTATE(x[0]^x[2]^x[8]^x[13],1)
 	nop.i	0			};;
 
 ___
 }
 else {
 $code.=<<___;
-{ .mib;	mov	$X[$i&0xf]=$f		    // Xupdate
+{ .mib;	add	$e=$e,$Konst		    // e+=K_60_79
 	dep.z	tmp5=$a,5,27		}   // a<<5
 { .mib;	xor	tmp0=$c,$b
-	add	tmp4=$e,$Konst		};;
-{ .mib;	xor	tmp0=tmp0,$d		    // F_20_39(b,c,d)=b^c^d
-	extr.u	tmp1=$a,27,5		}   // a>>27
-{ .mib;	add	$f=$f,tmp4		    // f+=e+K_20_39
 	add	$h1=$h1,$a		};; // wrap up
-{ .mmi;	add	$f=$f,tmp0		    // f+=F_20_39(b,c,d)
-	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30) ;;?
-{ .mmi;	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
+{ .mib;	add	$e=$e,$X[$i%16]		    // e+=Xupdate
+	extr.u	tmp1=$a,27,5		}   // a>>27
+{ .mib;	xor	tmp0=tmp0,$d		    // F_20_39(b,c,d)=b^c^d
 	add	$h3=$h3,$c		};; // wrap up
-{ .mib;	add	tmp3=1,inp		    // used in unaligned codepath
-	add	$f=$f,tmp1		}   // f+=ROTATE(a,5)
-{ .mib;	add	$h2=$h2,$b		    // wrap up
+{ .mmi;	add	$e=$e,tmp0		    // e+=F_20_39(b,c,d)
+	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
+	shrp	$b=tmp6,tmp6,2		};; // b=ROTATE(b,30) ;;?
+{ .mmi;	add	$e=$e,tmp1		    // e+=ROTATE(a,5)
+	add	tmp3=1,inp		    // used in unaligned codepath
 	add	$h4=$h4,$d		};; // wrap up
 
 ___
@@ -193,29 +191,29 @@ ___
 
 sub BODY_40_59 {
 local	*code=shift;
-local	($i,$a,$b,$c,$d,$e,$f)=@_;
+my	($i,$a,$b,$c,$d,$e)=@_;
+my	$j=$i+1;
+my	$Xn=@X[$j%16];
 
 $code.=<<___;
-{ .mmi;	mov	$X[$i&0xf]=$f		    // Xupdate
-	and	tmp0=$c,$b
+{ .mib;	add	$e=$e,$K_40_59		    // e+=K_40_59
 	dep.z	tmp5=$a,5,27		}   // a<<5
-{ .mmi;	and	tmp1=$d,$b
-	add	tmp4=$e,$K_40_59	};;
-{ .mmi;	or	tmp0=tmp0,tmp1		    // (b&c)|(b&d)
-	add	$f=$f,tmp4		    // f+=e+K_40_59
+{ .mib;	and	tmp1=$c,$d
+	xor	tmp0=$c,$d		};;
+{ .mmi;	add	$e=$e,$X[$i%16]		    // e+=Xupdate
+	add	tmp5=tmp5,tmp1		    // a<<5+(c&d)
 	extr.u	tmp1=$a,27,5		}   // a>>27
-{ .mmi;	and	tmp4=$c,$d
-	xor	tmp2=$X[($i+0+1)&0xf],$X[($i+2+1)&0xf]	// +1
-	xor	tmp3=$X[($i+8+1)&0xf],$X[($i+13+1)&0xf]	// +1
-	};;
-{ .mmi;	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
-	xor	tmp2=tmp2,tmp3		    // +1
+{ .mmi;	and	tmp0=tmp0,$b
+	xor	$Xn=$Xn,$X[($j+2)%16]	    // forward Xupdate
+	xor	tmp3=$X[($j+8)%16],$X[($j+13)%16] };;	// forward Xupdate
+{ .mmi;	add	$e=$e,tmp0		    // e+=b&(c^d)
+	add	tmp5=tmp5,tmp1		    // ROTATE(a,5)+(c&d)
 	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
-{ .mmi;	or	tmp0=tmp0,tmp4		    // F_40_59(b,c,d)=(b&c)|(b&d)|(c&d)
+{ .mmi;	xor	$Xn=$Xn,tmp3
 	mux2	tmp6=$a,0x44		};; // see b in next iteration
-{ .mii;	add	$f=$f,tmp0		    // f+=F_40_59(b,c,d)
-	shrp	$e=tmp2,tmp2,31;;	    // f+1=ROTATE(x[0]^x[2]^x[8]^x[13],1)
-	add	$f=$f,tmp1		};; // f+=ROTATE(a,5)
+{ .mii;	add	$e=$e,tmp5		    // e+=ROTATE(a,5)+(c&d)
+	shrp	$Xn=$Xn,$Xn,31		    // ROTATE(x[0]^x[2]^x[8]^x[13],1)
+	nop.i	0x0			};;
 
 ___
 }
@@ -237,7 +235,7 @@ inp=r33;	// in1
 .align	32
 sha1_block_data_order:
 	.prologue
-{ .mmi;	alloc	tmp1=ar.pfs,3,15,0,0
+{ .mmi;	alloc	tmp1=ar.pfs,3,14,0,0
 	$ADDP	tmp0=4,ctx
 	.save	ar.lc,r3
 	mov	r3=ar.lc		}
@@ -245,8 +243,8 @@ sha1_block_data_order:
 	$ADDP	inp=0,inp
 	mov	r2=pr			};;
 tmp4=in2;
-tmp5=loc13;
-tmp6=loc14;
+tmp5=loc12;
+tmp6=loc13;
 	.body
 { .mlx;	ld4	$h0=[ctx],8
 	movl	$K_00_19=0x5a827999	}
@@ -273,7 +271,7 @@ tmp6=loc14;
 
 ___
 
-{ my $i,@V=($A,$B,$C,$D,$E,$T);
+{ my $i,@V=($A,$B,$C,$D,$E);
 
 	for($i=0;$i<16;$i++)	{ &BODY_00_15(\$code,$i,@V); unshift(@V,pop(@V)); }
 	for(;$i<20;$i++)	{ &BODY_16_19(\$code,$i,@V); unshift(@V,pop(@V)); }
@@ -281,12 +279,12 @@ ___
 	for(;$i<60;$i++)	{ &BODY_40_59(\$code,$i,@V); unshift(@V,pop(@V)); }
 	for(;$i<80;$i++)	{ &BODY_60_79(\$code,$i,@V); unshift(@V,pop(@V)); }
 
-	(($V[5] eq $D) and ($V[0] eq $E)) or die;	# double-check
+	(($V[0] eq $A) and ($V[4] eq $E)) or die;	# double-check
 }
 
 $code.=<<___;
-{ .mmb;	add	$h0=$h0,$E
-	nop.m	0
+{ .mmb;	add	$h0=$h0,$A
+	add	$h2=$h2,$C
 	br.ctop.dptk.many	.Ldtop	};;
 .Ldend:
 { .mmi;	add	tmp0=4,ctx
diff --git a/crypto/sha/asm/sha1-mips.pl b/crypto/sha/asm/sha1-mips.pl
new file mode 100755
index 000000000000..f1a702f38f53
--- /dev/null
+++ b/crypto/sha/asm/sha1-mips.pl
@@ -0,0 +1,354 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA1 block procedure for MIPS.
+
+# Performance improvement is 30% on unaligned input. The "secret" is
+# to deploy lwl/lwr pair to load unaligned input. One could have
+# vectorized Xupdate on MIPSIII/IV, but the goal was to code MIPS32-
+# compatible subroutine. There is room for minor optimization on
+# little-endian platforms...
+
+######################################################################
+# There is a number of MIPS ABI in use, O32 and N32/64 are most
+# widely used. Then there is a new contender: NUBI. It appears that if
+# one picks the latter, it's possible to arrange code in ABI neutral
+# manner. Therefore let's stick to NUBI register layout:
+#
+($zero,$at,$t0,$t1,$t2)=map("\$$_",(0..2,24,25));
+($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7,$s8,$s9,$s10,$s11)=map("\$$_",(12..23));
+($gp,$tp,$sp,$fp,$ra)=map("\$$_",(3,28..31));
+#
+# The return value is placed in $a0. Following coding rules facilitate
+# interoperability:
+#
+# - never ever touch $tp, "thread pointer", former $gp;
+# - copy return value to $t0, former $v0 [or to $a0 if you're adapting
+#   old code];
+# - on O32 populate $a4-$a7 with 'lw $aN,4*N($sp)' if necessary;
+#
+# For reference here is register layout for N32/64 MIPS ABIs:
+#
+# ($zero,$at,$v0,$v1)=map("\$$_",(0..3));
+# ($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+# ($t0,$t1,$t2,$t3,$t8,$t9)=map("\$$_",(12..15,24,25));
+# ($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7)=map("\$$_",(16..23));
+# ($gp,$sp,$fp,$ra)=map("\$$_",(28..31));
+#
+$flavour = shift; # supported flavours are o32,n32,64,nubi32,nubi64
+
+if ($flavour =~ /64|n32/i) {
+	$PTR_ADD="dadd";	# incidentally works even on n32
+	$PTR_SUB="dsub";	# incidentally works even on n32
+	$REG_S="sd";
+	$REG_L="ld";
+	$PTR_SLL="dsll";	# incidentally works even on n32
+	$SZREG=8;
+} else {
+	$PTR_ADD="add";
+	$PTR_SUB="sub";
+	$REG_S="sw";
+	$REG_L="lw";
+	$PTR_SLL="sll";
+	$SZREG=4;
+}
+#
+# <appro@openssl.org>
+#
+######################################################################
+
+$big_endian=(`echo MIPSEL | $ENV{CC} -E -P -`=~/MIPSEL/)?1:0;
+
+for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);   }
+open STDOUT,">$output";
+
+if (!defined($big_endian))
+            {   $big_endian=(unpack('L',pack('N',1))==1);   }
+
+# offsets of the Most and Least Significant Bytes
+$MSB=$big_endian?0:3;
+$LSB=3&~$MSB;
+
+@X=map("\$$_",(8..23));	# a4-a7,s0-s11
+
+$ctx=$a0;
+$inp=$a1;
+$num=$a2;
+$A="\$1";
+$B="\$2";
+$C="\$3";
+$D="\$7";
+$E="\$24";	@V=($A,$B,$C,$D,$E);
+$t0="\$25";
+$t1=$num;	# $num is offloaded to stack
+$t2="\$30";	# fp
+$K="\$31";	# ra
+
+sub BODY_00_14 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___	if (!$big_endian);
+	srl	$t0,@X[$i],24	# byte swap($i)
+	srl	$t1,@X[$i],8
+	andi	$t2,@X[$i],0xFF00
+	sll	@X[$i],@X[$i],24
+	andi	$t1,0xFF00
+	sll	$t2,$t2,8
+	or	@X[$i],$t0
+	or	$t1,$t2
+	or	@X[$i],$t1
+___
+$code.=<<___;
+	 lwl	@X[$j],$j*4+$MSB($inp)
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	 lwr	@X[$j],$j*4+$LSB($inp)
+	srl	$t1,$a,27
+	addu	$e,$t0
+	xor	$t0,$c,$d
+	addu	$e,$t1
+	sll	$t2,$b,30
+	and	$t0,$b
+	srl	$b,$b,2
+	xor	$t0,$d
+	addu	$e,@X[$i]
+	or	$b,$t2
+	addu	$e,$t0
+___
+}
+
+sub BODY_15_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+
+$code.=<<___	if (!$big_endian && $i==15);
+	srl	$t0,@X[$i],24	# byte swap($i)
+	srl	$t1,@X[$i],8
+	andi	$t2,@X[$i],0xFF00
+	sll	@X[$i],@X[$i],24
+	andi	$t1,0xFF00
+	sll	$t2,$t2,8
+	or	@X[$i],$t0
+	or	@X[$i],$t1
+	or	@X[$i],$t2
+___
+$code.=<<___;
+	 xor	@X[$j%16],@X[($j+2)%16]
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	srl	$t1,$a,27
+	addu	$e,$t0
+	 xor	@X[$j%16],@X[($j+8)%16]
+	xor	$t0,$c,$d
+	addu	$e,$t1
+	 xor	@X[$j%16],@X[($j+13)%16]
+	sll	$t2,$b,30
+	and	$t0,$b
+	 srl	$t1,@X[$j%16],31
+	 addu	@X[$j%16],@X[$j%16]
+	srl	$b,$b,2
+	xor	$t0,$d
+	 or	@X[$j%16],$t1
+	addu	$e,@X[$i%16]
+	or	$b,$t2
+	addu	$e,$t0
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<79);
+	 xor	@X[$j%16],@X[($j+2)%16]
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	srl	$t1,$a,27
+	addu	$e,$t0
+	 xor	@X[$j%16],@X[($j+8)%16]
+	xor	$t0,$c,$d
+	addu	$e,$t1
+	 xor	@X[$j%16],@X[($j+13)%16]
+	sll	$t2,$b,30
+	xor	$t0,$b
+	 srl	$t1,@X[$j%16],31
+	 addu	@X[$j%16],@X[$j%16]
+	srl	$b,$b,2
+	addu	$e,@X[$i%16]
+	 or	@X[$j%16],$t1
+	or	$b,$t2
+	addu	$e,$t0
+___
+$code.=<<___ if ($i==79);
+	 lw	@X[0],0($ctx)
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	 lw	@X[1],4($ctx)
+	srl	$t1,$a,27
+	addu	$e,$t0
+	 lw	@X[2],8($ctx)
+	xor	$t0,$c,$d
+	addu	$e,$t1
+	 lw	@X[3],12($ctx)
+	sll	$t2,$b,30
+	xor	$t0,$b
+	 lw	@X[4],16($ctx)
+	srl	$b,$b,2
+	addu	$e,@X[$i%16]
+	or	$b,$t2
+	addu	$e,$t0
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<79);
+	 xor	@X[$j%16],@X[($j+2)%16]
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	srl	$t1,$a,27
+	addu	$e,$t0
+	 xor	@X[$j%16],@X[($j+8)%16]
+	and	$t0,$c,$d
+	addu	$e,$t1
+	 xor	@X[$j%16],@X[($j+13)%16]
+	sll	$t2,$b,30
+	addu	$e,$t0
+	 srl	$t1,@X[$j%16],31
+	xor	$t0,$c,$d
+	 addu	@X[$j%16],@X[$j%16]
+	and	$t0,$b
+	srl	$b,$b,2
+	 or	@X[$j%16],$t1
+	addu	$e,@X[$i%16]
+	or	$b,$t2
+	addu	$e,$t0
+___
+}
+
+$FRAMESIZE=16;	# large enough to accomodate NUBI saved registers
+$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0xc0fff008 : 0xc0ff0000;
+
+$code=<<___;
+#ifdef OPENSSL_FIPSCANISTER
+# include <openssl/fipssyms.h>
+#endif
+
+.text
+
+.set	noat
+.set	noreorder
+.align	5
+.globl	sha1_block_data_order
+.ent	sha1_block_data_order
+sha1_block_data_order:
+	.frame	$sp,$FRAMESIZE*$SZREG,$ra
+	.mask	$SAVED_REGS_MASK,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,$FRAMESIZE*$SZREG
+	$REG_S	$ra,($FRAMESIZE-1)*$SZREG($sp)
+	$REG_S	$fp,($FRAMESIZE-2)*$SZREG($sp)
+	$REG_S	$s11,($FRAMESIZE-3)*$SZREG($sp)
+	$REG_S	$s10,($FRAMESIZE-4)*$SZREG($sp)
+	$REG_S	$s9,($FRAMESIZE-5)*$SZREG($sp)
+	$REG_S	$s8,($FRAMESIZE-6)*$SZREG($sp)
+	$REG_S	$s7,($FRAMESIZE-7)*$SZREG($sp)
+	$REG_S	$s6,($FRAMESIZE-8)*$SZREG($sp)
+	$REG_S	$s5,($FRAMESIZE-9)*$SZREG($sp)
+	$REG_S	$s4,($FRAMESIZE-10)*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);	# optimize non-nubi prologue
+	$REG_S	$s3,($FRAMESIZE-11)*$SZREG($sp)
+	$REG_S	$s2,($FRAMESIZE-12)*$SZREG($sp)
+	$REG_S	$s1,($FRAMESIZE-13)*$SZREG($sp)
+	$REG_S	$s0,($FRAMESIZE-14)*$SZREG($sp)
+	$REG_S	$gp,($FRAMESIZE-15)*$SZREG($sp)
+___
+$code.=<<___;
+	$PTR_SLL $num,6
+	$PTR_ADD $num,$inp
+	$REG_S	$num,0($sp)
+	lw	$A,0($ctx)
+	lw	$B,4($ctx)
+	lw	$C,8($ctx)
+	lw	$D,12($ctx)
+	b	.Loop
+	lw	$E,16($ctx)
+.align	4
+.Loop:
+	.set	reorder
+	lwl	@X[0],$MSB($inp)
+	lui	$K,0x5a82
+	lwr	@X[0],$LSB($inp)
+	ori	$K,0x7999	# K_00_19
+___
+for ($i=0;$i<15;$i++)	{ &BODY_00_14($i,@V); unshift(@V,pop(@V)); }
+for (;$i<20;$i++)	{ &BODY_15_19($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	lui	$K,0x6ed9
+	ori	$K,0xeba1	# K_20_39
+___
+for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	lui	$K,0x8f1b
+	ori	$K,0xbcdc	# K_40_59
+___
+for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	lui	$K,0xca62
+	ori	$K,0xc1d6	# K_60_79
+___
+for (;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	$PTR_ADD $inp,64
+	$REG_L	$num,0($sp)
+
+	addu	$A,$X[0]
+	addu	$B,$X[1]
+	sw	$A,0($ctx)
+	addu	$C,$X[2]
+	addu	$D,$X[3]
+	sw	$B,4($ctx)
+	addu	$E,$X[4]
+	sw	$C,8($ctx)
+	sw	$D,12($ctx)
+	sw	$E,16($ctx)
+	.set	noreorder
+	bne	$inp,$num,.Loop
+	nop
+
+	.set	noreorder
+	$REG_L	$ra,($FRAMESIZE-1)*$SZREG($sp)
+	$REG_L	$fp,($FRAMESIZE-2)*$SZREG($sp)
+	$REG_L	$s11,($FRAMESIZE-3)*$SZREG($sp)
+	$REG_L	$s10,($FRAMESIZE-4)*$SZREG($sp)
+	$REG_L	$s9,($FRAMESIZE-5)*$SZREG($sp)
+	$REG_L	$s8,($FRAMESIZE-6)*$SZREG($sp)
+	$REG_L	$s7,($FRAMESIZE-7)*$SZREG($sp)
+	$REG_L	$s6,($FRAMESIZE-8)*$SZREG($sp)
+	$REG_L	$s5,($FRAMESIZE-9)*$SZREG($sp)
+	$REG_L	$s4,($FRAMESIZE-10)*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$s3,($FRAMESIZE-11)*$SZREG($sp)
+	$REG_L	$s2,($FRAMESIZE-12)*$SZREG($sp)
+	$REG_L	$s1,($FRAMESIZE-13)*$SZREG($sp)
+	$REG_L	$s0,($FRAMESIZE-14)*$SZREG($sp)
+	$REG_L	$gp,($FRAMESIZE-15)*$SZREG($sp)
+___
+$code.=<<___;
+	jr	$ra
+	$PTR_ADD $sp,$FRAMESIZE*$SZREG
+.end	sha1_block_data_order
+.rdata
+.asciiz	"SHA1 for MIPS, CRYPTOGAMS by <appro\@openssl.org>"
+___
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha1-parisc.pl b/crypto/sha/asm/sha1-parisc.pl
new file mode 100755
index 000000000000..6d7bf495b20a
--- /dev/null
+++ b/crypto/sha/asm/sha1-parisc.pl
@@ -0,0 +1,259 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA1 block procedure for PA-RISC.
+
+# June 2009.
+#
+# On PA-7100LC performance is >30% better than gcc 3.2 generated code
+# for aligned input and >50% better for unaligned. Compared to vendor
+# compiler on PA-8600 it's almost 60% faster in 64-bit build and just
+# few percent faster in 32-bit one (this for aligned input, data for
+# unaligned input is not available).
+#
+# Special thanks to polarhome.com for providing HP-UX account.
+
+$flavour = shift;
+$output = shift;
+open STDOUT,">$output";
+
+if ($flavour =~ /64/) {
+	$LEVEL		="2.0W";
+	$SIZE_T		=8;
+	$FRAME_MARKER	=80;
+	$SAVED_RP	=16;
+	$PUSH		="std";
+	$PUSHMA		="std,ma";
+	$POP		="ldd";
+	$POPMB		="ldd,mb";
+} else {
+	$LEVEL		="1.0";
+	$SIZE_T		=4;
+	$FRAME_MARKER	=48;
+	$SAVED_RP	=20;
+	$PUSH		="stw";
+	$PUSHMA		="stwm";
+	$POP		="ldw";
+	$POPMB		="ldwm";
+}
+
+$FRAME=14*$SIZE_T+$FRAME_MARKER;# 14 saved regs + frame marker
+				#                 [+ argument transfer]
+$ctx="%r26";		# arg0
+$inp="%r25";		# arg1
+$num="%r24";		# arg2
+
+$t0="%r28";
+$t1="%r29";
+$K="%r31";
+
+@X=("%r1", "%r2", "%r3", "%r4", "%r5", "%r6", "%r7", "%r8",
+    "%r9", "%r10","%r11","%r12","%r13","%r14","%r15","%r16",$t0);
+
+@V=($A,$B,$C,$D,$E)=("%r19","%r20","%r21","%r22","%r23");
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<15);
+	addl	$K,$e,$e	; $i
+	shd	$a,$a,27,$t1
+	addl	@X[$i],$e,$e
+	and	$c,$b,$t0
+	addl	$t1,$e,$e
+	andcm	$d,$b,$t1
+	shd	$b,$b,2,$b
+	or	$t1,$t0,$t0
+	addl	$t0,$e,$e
+___
+$code.=<<___ if ($i>=15);	# with forward Xupdate
+	addl	$K,$e,$e	; $i
+	shd	$a,$a,27,$t1
+	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]
+	addl	@X[$i%16],$e,$e
+	and	$c,$b,$t0
+	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
+	addl	$t1,$e,$e
+	andcm	$d,$b,$t1
+	shd	$b,$b,2,$b
+	or	$t1,$t0,$t0
+	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
+	add	$t0,$e,$e
+	shd	@X[$j%16],@X[$j%16],31,@X[$j%16]
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<79);
+	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]	; $i
+	addl	$K,$e,$e
+	shd	$a,$a,27,$t1
+	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
+	addl	@X[$i%16],$e,$e
+	xor	$b,$c,$t0
+	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
+	addl	$t1,$e,$e
+	shd	$b,$b,2,$b
+	xor	$d,$t0,$t0
+	shd	@X[$j%16],@X[$j%16],31,@X[$j%16]
+	addl	$t0,$e,$e
+___
+$code.=<<___ if ($i==79);	# with context load
+	ldw	0($ctx),@X[0]	; $i
+	addl	$K,$e,$e
+	shd	$a,$a,27,$t1
+	ldw	4($ctx),@X[1]
+	addl	@X[$i%16],$e,$e
+	xor	$b,$c,$t0
+	ldw	8($ctx),@X[2]
+	addl	$t1,$e,$e
+	shd	$b,$b,2,$b
+	xor	$d,$t0,$t0
+	ldw	12($ctx),@X[3]
+	addl	$t0,$e,$e
+	ldw	16($ctx),@X[4]
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___;
+	shd	$a,$a,27,$t1	; $i
+	addl	$K,$e,$e
+	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]
+	xor	$d,$c,$t0
+	addl	@X[$i%16],$e,$e
+	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
+	and	$b,$t0,$t0
+	addl	$t1,$e,$e
+	shd	$b,$b,2,$b
+	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
+	addl	$t0,$e,$e
+	and	$d,$c,$t1
+	shd	@X[$j%16],@X[$j%16],31,@X[$j%16]
+	addl	$t1,$e,$e
+___
+}
+
+$code=<<___;
+	.LEVEL	$LEVEL
+	.SPACE	\$TEXT\$
+	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY
+
+	.EXPORT	sha1_block_data_order,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR
+sha1_block_data_order
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-14*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=16
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
+	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
+	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
+	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
+	$PUSH	%r11,`-$FRAME+8*$SIZE_T`(%sp)
+	$PUSH	%r12,`-$FRAME+9*$SIZE_T`(%sp)
+	$PUSH	%r13,`-$FRAME+10*$SIZE_T`(%sp)
+	$PUSH	%r14,`-$FRAME+11*$SIZE_T`(%sp)
+	$PUSH	%r15,`-$FRAME+12*$SIZE_T`(%sp)
+	$PUSH	%r16,`-$FRAME+13*$SIZE_T`(%sp)
+
+	ldw	0($ctx),$A
+	ldw	4($ctx),$B
+	ldw	8($ctx),$C
+	ldw	12($ctx),$D
+	ldw	16($ctx),$E
+
+	extru	$inp,31,2,$t0		; t0=inp&3;
+	sh3addl	$t0,%r0,$t0		; t0*=8;
+	subi	32,$t0,$t0		; t0=32-t0;
+	mtctl	$t0,%cr11		; %sar=t0;
+
+L\$oop
+	ldi	3,$t0
+	andcm	$inp,$t0,$t0		; 64-bit neutral
+___
+	for ($i=0;$i<15;$i++) {		# load input block
+	$code.="\tldw	`4*$i`($t0),@X[$i]\n";		}
+$code.=<<___;
+	cmpb,*=	$inp,$t0,L\$aligned
+	ldw	60($t0),@X[15]
+	ldw	64($t0),@X[16]
+___
+	for ($i=0;$i<16;$i++) {		# align input
+	$code.="\tvshd	@X[$i],@X[$i+1],@X[$i]\n";	}
+$code.=<<___;
+L\$aligned
+	ldil	L'0x5a827000,$K		; K_00_19
+	ldo	0x999($K),$K
+___
+for ($i=0;$i<20;$i++)   { &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	ldil	L'0x6ed9e000,$K		; K_20_39
+	ldo	0xba1($K),$K
+___
+
+for (;$i<40;$i++)       { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	ldil	L'0x8f1bb000,$K		; K_40_59
+	ldo	0xcdc($K),$K
+___
+
+for (;$i<60;$i++)       { &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	ldil	L'0xca62c000,$K		; K_60_79
+	ldo	0x1d6($K),$K
+___
+for (;$i<80;$i++)       { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+
+$code.=<<___;
+	addl	@X[0],$A,$A
+	addl	@X[1],$B,$B
+	addl	@X[2],$C,$C
+	addl	@X[3],$D,$D
+	addl	@X[4],$E,$E
+	stw	$A,0($ctx)
+	stw	$B,4($ctx)
+	stw	$C,8($ctx)
+	stw	$D,12($ctx)
+	stw	$E,16($ctx)
+	addib,*<> -1,$num,L\$oop
+	ldo	64($inp),$inp
+
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2	; standard epilogue
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
+	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
+	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
+	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
+	$POP	`-$FRAME+8*$SIZE_T`(%sp),%r11
+	$POP	`-$FRAME+9*$SIZE_T`(%sp),%r12
+	$POP	`-$FRAME+10*$SIZE_T`(%sp),%r13
+	$POP	`-$FRAME+11*$SIZE_T`(%sp),%r14
+	$POP	`-$FRAME+12*$SIZE_T`(%sp),%r15
+	$POP	`-$FRAME+13*$SIZE_T`(%sp),%r16
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+	.STRINGZ "SHA1 block transform for PA-RISC, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/,\*/,/gm if ($SIZE_T==4);
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha1-ppc.pl b/crypto/sha/asm/sha1-ppc.pl
new file mode 100755
index 000000000000..2140dd2f8dd6
--- /dev/null
+++ b/crypto/sha/asm/sha1-ppc.pl
@@ -0,0 +1,326 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# I let hardware handle unaligned input(*), except on page boundaries
+# (see below for details). Otherwise straightforward implementation
+# with X vector in register bank. The module is big-endian [which is
+# not big deal as there're no little-endian targets left around].
+#
+# (*) this means that this module is inappropriate for PPC403? Does
+#     anybody know if pre-POWER3 can sustain unaligned load?
+
+# 			-m64	-m32
+# ----------------------------------
+# PPC970,gcc-4.0.0	+76%	+59%
+# Power6,xlc-7		+68%	+33%
+
+$flavour = shift;
+
+if ($flavour =~ /64/) {
+	$SIZE_T	=8;
+	$LRSAVE	=2*$SIZE_T;
+	$UCMP	="cmpld";
+	$STU	="stdu";
+	$POP	="ld";
+	$PUSH	="std";
+} elsif ($flavour =~ /32/) {
+	$SIZE_T	=4;
+	$LRSAVE	=$SIZE_T;
+	$UCMP	="cmplw";
+	$STU	="stwu";
+	$POP	="lwz";
+	$PUSH	="stw";
+} else { die "nonsense $flavour"; }
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour ".shift || die "can't call $xlate: $!";
+
+$FRAME=24*$SIZE_T+64;
+$LOCALS=6*$SIZE_T;
+
+$K  ="r0";
+$sp ="r1";
+$toc="r2";
+$ctx="r3";
+$inp="r4";
+$num="r5";
+$t0 ="r15";
+$t1 ="r6";
+
+$A  ="r7";
+$B  ="r8";
+$C  ="r9";
+$D  ="r10";
+$E  ="r11";
+$T  ="r12";
+
+@V=($A,$B,$C,$D,$E,$T);
+@X=("r16","r17","r18","r19","r20","r21","r22","r23",
+    "r24","r25","r26","r27","r28","r29","r30","r31");
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i==0);
+	lwz	@X[$i],`$i*4`($inp)
+___
+$code.=<<___ if ($i<15);
+	lwz	@X[$j],`$j*4`($inp)
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	add	$f,$f,@X[$i]
+	and	$t0,$c,$b
+	add	$f,$f,$e
+	andc	$t1,$d,$b
+	rotlwi	$b,$b,30
+	or	$t0,$t0,$t1
+	add	$f,$f,$t0
+___
+$code.=<<___ if ($i>=15);
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	xor	@X[$j%16],@X[$j%16],@X[($j+2)%16]
+	add	$f,$f,@X[$i%16]
+	and	$t0,$c,$b
+	xor	@X[$j%16],@X[$j%16],@X[($j+8)%16]
+	add	$f,$f,$e
+	andc	$t1,$d,$b
+	rotlwi	$b,$b,30
+	or	$t0,$t0,$t1
+	xor	@X[$j%16],@X[$j%16],@X[($j+13)%16]
+	add	$f,$f,$t0
+	rotlwi	@X[$j%16],@X[$j%16],1
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<79);
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	xor	@X[$j%16],@X[$j%16],@X[($j+2)%16]
+	add	$f,$f,@X[$i%16]
+	xor	$t0,$b,$c
+	xor	@X[$j%16],@X[$j%16],@X[($j+8)%16]
+	add	$f,$f,$e
+	rotlwi	$b,$b,30
+	xor	$t0,$t0,$d
+	xor	@X[$j%16],@X[$j%16],@X[($j+13)%16]
+	add	$f,$f,$t0
+	rotlwi	@X[$j%16],@X[$j%16],1
+___
+$code.=<<___ if ($i==79);
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	lwz	r16,0($ctx)
+	add	$f,$f,@X[$i%16]
+	xor	$t0,$b,$c
+	lwz	r17,4($ctx)
+	add	$f,$f,$e
+	rotlwi	$b,$b,30
+	lwz	r18,8($ctx)
+	xor	$t0,$t0,$d
+	lwz	r19,12($ctx)
+	add	$f,$f,$t0
+	lwz	r20,16($ctx)
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my $j=$i+1;
+$code.=<<___;
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	xor	@X[$j%16],@X[$j%16],@X[($j+2)%16]
+	add	$f,$f,@X[$i%16]
+	and	$t0,$b,$c
+	xor	@X[$j%16],@X[$j%16],@X[($j+8)%16]
+	add	$f,$f,$e
+	or	$t1,$b,$c
+	rotlwi	$b,$b,30
+	xor	@X[$j%16],@X[$j%16],@X[($j+13)%16]
+	and	$t1,$t1,$d
+	or	$t0,$t0,$t1
+	rotlwi	@X[$j%16],@X[$j%16],1
+	add	$f,$f,$t0
+___
+}
+
+$code=<<___;
+.machine	"any"
+.text
+
+.globl	.sha1_block_data_order
+.align	4
+.sha1_block_data_order:
+	$STU	$sp,-$FRAME($sp)
+	mflr	r0
+	$PUSH	r15,`$FRAME-$SIZE_T*17`($sp)
+	$PUSH	r16,`$FRAME-$SIZE_T*16`($sp)
+	$PUSH	r17,`$FRAME-$SIZE_T*15`($sp)
+	$PUSH	r18,`$FRAME-$SIZE_T*14`($sp)
+	$PUSH	r19,`$FRAME-$SIZE_T*13`($sp)
+	$PUSH	r20,`$FRAME-$SIZE_T*12`($sp)
+	$PUSH	r21,`$FRAME-$SIZE_T*11`($sp)
+	$PUSH	r22,`$FRAME-$SIZE_T*10`($sp)
+	$PUSH	r23,`$FRAME-$SIZE_T*9`($sp)
+	$PUSH	r24,`$FRAME-$SIZE_T*8`($sp)
+	$PUSH	r25,`$FRAME-$SIZE_T*7`($sp)
+	$PUSH	r26,`$FRAME-$SIZE_T*6`($sp)
+	$PUSH	r27,`$FRAME-$SIZE_T*5`($sp)
+	$PUSH	r28,`$FRAME-$SIZE_T*4`($sp)
+	$PUSH	r29,`$FRAME-$SIZE_T*3`($sp)
+	$PUSH	r30,`$FRAME-$SIZE_T*2`($sp)
+	$PUSH	r31,`$FRAME-$SIZE_T*1`($sp)
+	$PUSH	r0,`$FRAME+$LRSAVE`($sp)
+	lwz	$A,0($ctx)
+	lwz	$B,4($ctx)
+	lwz	$C,8($ctx)
+	lwz	$D,12($ctx)
+	lwz	$E,16($ctx)
+	andi.	r0,$inp,3
+	bne	Lunaligned
+Laligned:
+	mtctr	$num
+	bl	Lsha1_block_private
+	b	Ldone
+
+; PowerPC specification allows an implementation to be ill-behaved
+; upon unaligned access which crosses page boundary. "Better safe
+; than sorry" principle makes me treat it specially. But I don't
+; look for particular offending word, but rather for 64-byte input
+; block which crosses the boundary. Once found that block is aligned
+; and hashed separately...
+.align	4
+Lunaligned:
+	subfic	$t1,$inp,4096
+	andi.	$t1,$t1,4095	; distance to closest page boundary
+	srwi.	$t1,$t1,6	; t1/=64
+	beq	Lcross_page
+	$UCMP	$num,$t1
+	ble-	Laligned	; didn't cross the page boundary
+	mtctr	$t1
+	subfc	$num,$t1,$num
+	bl	Lsha1_block_private
+Lcross_page:
+	li	$t1,16
+	mtctr	$t1
+	addi	r20,$sp,$LOCALS	; spot within the frame
+Lmemcpy:
+	lbz	r16,0($inp)
+	lbz	r17,1($inp)
+	lbz	r18,2($inp)
+	lbz	r19,3($inp)
+	addi	$inp,$inp,4
+	stb	r16,0(r20)
+	stb	r17,1(r20)
+	stb	r18,2(r20)
+	stb	r19,3(r20)
+	addi	r20,r20,4
+	bdnz	Lmemcpy
+
+	$PUSH	$inp,`$FRAME-$SIZE_T*18`($sp)
+	li	$t1,1
+	addi	$inp,$sp,$LOCALS
+	mtctr	$t1
+	bl	Lsha1_block_private
+	$POP	$inp,`$FRAME-$SIZE_T*18`($sp)
+	addic.	$num,$num,-1
+	bne-	Lunaligned
+
+Ldone:
+	$POP	r0,`$FRAME+$LRSAVE`($sp)
+	$POP	r15,`$FRAME-$SIZE_T*17`($sp)
+	$POP	r16,`$FRAME-$SIZE_T*16`($sp)
+	$POP	r17,`$FRAME-$SIZE_T*15`($sp)
+	$POP	r18,`$FRAME-$SIZE_T*14`($sp)
+	$POP	r19,`$FRAME-$SIZE_T*13`($sp)
+	$POP	r20,`$FRAME-$SIZE_T*12`($sp)
+	$POP	r21,`$FRAME-$SIZE_T*11`($sp)
+	$POP	r22,`$FRAME-$SIZE_T*10`($sp)
+	$POP	r23,`$FRAME-$SIZE_T*9`($sp)
+	$POP	r24,`$FRAME-$SIZE_T*8`($sp)
+	$POP	r25,`$FRAME-$SIZE_T*7`($sp)
+	$POP	r26,`$FRAME-$SIZE_T*6`($sp)
+	$POP	r27,`$FRAME-$SIZE_T*5`($sp)
+	$POP	r28,`$FRAME-$SIZE_T*4`($sp)
+	$POP	r29,`$FRAME-$SIZE_T*3`($sp)
+	$POP	r30,`$FRAME-$SIZE_T*2`($sp)
+	$POP	r31,`$FRAME-$SIZE_T*1`($sp)
+	mtlr	r0
+	addi	$sp,$sp,$FRAME
+	blr
+	.long	0
+	.byte	0,12,4,1,0x80,18,3,0
+	.long	0
+___
+
+# This is private block function, which uses tailored calling
+# interface, namely upon entry SHA_CTX is pre-loaded to given
+# registers and counter register contains amount of chunks to
+# digest...
+$code.=<<___;
+.align	4
+Lsha1_block_private:
+___
+$code.=<<___;	# load K_00_19
+	lis	$K,0x5a82
+	ori	$K,$K,0x7999
+___
+for($i=0;$i<20;$i++)	{ &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;	# load K_20_39
+	lis	$K,0x6ed9
+	ori	$K,$K,0xeba1
+___
+for(;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;	# load K_40_59
+	lis	$K,0x8f1b
+	ori	$K,$K,0xbcdc
+___
+for(;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;	# load K_60_79
+	lis	$K,0xca62
+	ori	$K,$K,0xc1d6
+___
+for(;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	add	r16,r16,$E
+	add	r17,r17,$T
+	add	r18,r18,$A
+	add	r19,r19,$B
+	add	r20,r20,$C
+	stw	r16,0($ctx)
+	mr	$A,r16
+	stw	r17,4($ctx)
+	mr	$B,r17
+	stw	r18,8($ctx)
+	mr	$C,r18
+	stw	r19,12($ctx)
+	mr	$D,r19
+	stw	r20,16($ctx)
+	mr	$E,r20
+	addi	$inp,$inp,`16*4`
+	bdnz-	Lsha1_block_private
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+___
+$code.=<<___;
+.asciz	"SHA1 block transform for PPC, CRYPTOGAMS by <appro\@fy.chalmers.se>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha1-s390x.pl b/crypto/sha/asm/sha1-s390x.pl
new file mode 100755
index 000000000000..9193dda45eff
--- /dev/null
+++ b/crypto/sha/asm/sha1-s390x.pl
@@ -0,0 +1,246 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA1 block procedure for s390x.
+
+# April 2007.
+#
+# Performance is >30% better than gcc 3.3 generated code. But the real
+# twist is that SHA1 hardware support is detected and utilized. In
+# which case performance can reach further >4.5x for larger chunks.
+
+# January 2009.
+#
+# Optimize Xupdate for amount of memory references and reschedule
+# instructions to favour dual-issue z10 pipeline. On z10 hardware is
+# "only" ~2.3x faster than software.
+
+# November 2010.
+#
+# Adapt for -m31 build. If kernel supports what's called "highgprs"
+# feature on Linux [see /proc/cpuinfo], it's possible to use 64-bit
+# instructions and achieve "64-bit" performance even in 31-bit legacy
+# application context. The feature is not specific to any particular
+# processor, as long as it's "z-CPU". Latter implies that the code
+# remains z/Architecture specific.
+
+$kimdfunc=1;	# magic function code for kimd instruction
+
+$flavour = shift;
+
+if ($flavour =~ /3[12]/) {
+	$SIZE_T=4;
+	$g="";
+} else {
+	$SIZE_T=8;
+	$g="g";
+}
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$K_00_39="%r0"; $K=$K_00_39;
+$K_40_79="%r1";
+$ctx="%r2";	$prefetch="%r2";
+$inp="%r3";
+$len="%r4";
+
+$A="%r5";
+$B="%r6";
+$C="%r7";
+$D="%r8";
+$E="%r9";	@V=($A,$B,$C,$D,$E);
+$t0="%r10";
+$t1="%r11";
+@X=("%r12","%r13","%r14");
+$sp="%r15";
+
+$stdframe=16*$SIZE_T+4*8;
+$frame=$stdframe+16*4;
+
+sub Xupdate {
+my $i=shift;
+
+$code.=<<___ if ($i==15);
+	lg	$prefetch,$stdframe($sp)	### Xupdate(16) warm-up
+	lr	$X[0],$X[2]
+___
+return if ($i&1);	# Xupdate is vectorized and executed every 2nd cycle
+$code.=<<___ if ($i<16);
+	lg	$X[0],`$i*4`($inp)	### Xload($i)
+	rllg	$X[1],$X[0],32
+___
+$code.=<<___ if ($i>=16);
+	xgr	$X[0],$prefetch		### Xupdate($i)
+	lg	$prefetch,`$stdframe+4*(($i+2)%16)`($sp)
+	xg	$X[0],`$stdframe+4*(($i+8)%16)`($sp)
+	xgr	$X[0],$prefetch
+	rll	$X[0],$X[0],1
+	rllg	$X[1],$X[0],32
+	rll	$X[1],$X[1],1
+	rllg	$X[0],$X[1],32
+	lr	$X[2],$X[1]		# feedback
+___
+$code.=<<___ if ($i<=70);
+	stg	$X[0],`$stdframe+4*($i%16)`($sp)
+___
+unshift(@X,pop(@X));
+}
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=$X[1];
+
+	&Xupdate($i);
+$code.=<<___;
+	alr	$e,$K		### $i
+	rll	$t1,$a,5
+	lr	$t0,$d
+	xr	$t0,$c
+	alr	$e,$t1
+	nr	$t0,$b
+	alr	$e,$xi
+	xr	$t0,$d
+	rll	$b,$b,30
+	alr	$e,$t0
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=$X[1];
+
+	&Xupdate($i);
+$code.=<<___;
+	alr	$e,$K		### $i
+	rll	$t1,$a,5
+	lr	$t0,$b
+	alr	$e,$t1
+	xr	$t0,$c
+	alr	$e,$xi
+	xr	$t0,$d
+	rll	$b,$b,30
+	alr	$e,$t0
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=$X[1];
+
+	&Xupdate($i);
+$code.=<<___;
+	alr	$e,$K		### $i
+	rll	$t1,$a,5
+	lr	$t0,$b
+	alr	$e,$t1
+	or	$t0,$c
+	lr	$t1,$b
+	nr	$t0,$d
+	nr	$t1,$c
+	alr	$e,$xi
+	or	$t0,$t1
+	rll	$b,$b,30
+	alr	$e,$t0
+___
+}
+
+$code.=<<___;
+.text
+.align	64
+.type	Ktable,\@object
+Ktable: .long	0x5a827999,0x6ed9eba1,0x8f1bbcdc,0xca62c1d6
+	.skip	48	#.long	0,0,0,0,0,0,0,0,0,0,0,0
+.size	Ktable,.-Ktable
+.globl	sha1_block_data_order
+.type	sha1_block_data_order,\@function
+sha1_block_data_order:
+___
+$code.=<<___ if ($kimdfunc);
+	larl	%r1,OPENSSL_s390xcap_P
+	lg	%r0,0(%r1)
+	tmhl	%r0,0x4000	# check for message-security assist
+	jz	.Lsoftware
+	lghi	%r0,0
+	la	%r1,`2*$SIZE_T`($sp)
+	.long	0xb93e0002	# kimd %r0,%r2
+	lg	%r0,`2*$SIZE_T`($sp)
+	tmhh	%r0,`0x8000>>$kimdfunc`
+	jz	.Lsoftware
+	lghi	%r0,$kimdfunc
+	lgr	%r1,$ctx
+	lgr	%r2,$inp
+	sllg	%r3,$len,6
+	.long	0xb93e0002	# kimd %r0,%r2
+	brc	1,.-4		# pay attention to "partial completion"
+	br	%r14
+.align	16
+.Lsoftware:
+___
+$code.=<<___;
+	lghi	%r1,-$frame
+	st${g}	$ctx,`2*$SIZE_T`($sp)
+	stm${g}	%r6,%r15,`6*$SIZE_T`($sp)
+	lgr	%r0,$sp
+	la	$sp,0(%r1,$sp)
+	st${g}	%r0,0($sp)
+
+	larl	$t0,Ktable
+	llgf	$A,0($ctx)
+	llgf	$B,4($ctx)
+	llgf	$C,8($ctx)
+	llgf	$D,12($ctx)
+	llgf	$E,16($ctx)
+
+	lg	$K_00_39,0($t0)
+	lg	$K_40_79,8($t0)
+
+.Lloop:
+	rllg	$K_00_39,$K_00_39,32
+___
+for ($i=0;$i<20;$i++)	{ &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	rllg	$K_00_39,$K_00_39,32
+___
+for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;	$K=$K_40_79;
+	rllg	$K_40_79,$K_40_79,32
+___
+for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	rllg	$K_40_79,$K_40_79,32
+___
+for (;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+
+	l${g}	$ctx,`$frame+2*$SIZE_T`($sp)
+	la	$inp,64($inp)
+	al	$A,0($ctx)
+	al	$B,4($ctx)
+	al	$C,8($ctx)
+	al	$D,12($ctx)
+	al	$E,16($ctx)
+	st	$A,0($ctx)
+	st	$B,4($ctx)
+	st	$C,8($ctx)
+	st	$D,12($ctx)
+	st	$E,16($ctx)
+	brct${g} $len,.Lloop
+
+	lm${g}	%r6,%r15,`$frame+6*$SIZE_T`($sp)
+	br	%r14
+.size	sha1_block_data_order,.-sha1_block_data_order
+.string	"SHA1 block transform for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+.comm	OPENSSL_s390xcap_P,16,8
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha1-sparcv9.pl b/crypto/sha/asm/sha1-sparcv9.pl
new file mode 100755
index 000000000000..5c161cecd696
--- /dev/null
+++ b/crypto/sha/asm/sha1-sparcv9.pl
@@ -0,0 +1,284 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# Performance improvement is not really impressive on pre-T1 CPU: +8%
+# over Sun C and +25% over gcc [3.3]. While on T1, a.k.a. Niagara, it
+# turned to be 40% faster than 64-bit code generated by Sun C 5.8 and
+# >2x than 64-bit code generated by gcc 3.4. And there is a gimmick.
+# X[16] vector is packed to 8 64-bit registers and as result nothing
+# is spilled on stack. In addition input data is loaded in compact
+# instruction sequence, thus minimizing the window when the code is
+# subject to [inter-thread] cache-thrashing hazard. The goal is to
+# ensure scalability on UltraSPARC T1, or rather to avoid decay when
+# amount of active threads exceeds the number of physical cores.
+
+$bits=32;
+for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+if ($bits==64)	{ $bias=2047; $frame=192; }
+else		{ $bias=0;    $frame=112; }
+
+$output=shift;
+open STDOUT,">$output";
+
+@X=("%o0","%o1","%o2","%o3","%o4","%o5","%g1","%o7");
+$rot1m="%g2";
+$tmp64="%g3";
+$Xi="%g4";
+$A="%l0";
+$B="%l1";
+$C="%l2";
+$D="%l3";
+$E="%l4";
+@V=($A,$B,$C,$D,$E);
+$K_00_19="%l5";
+$K_20_39="%l6";
+$K_40_59="%l7";
+$K_60_79="%g5";
+@K=($K_00_19,$K_20_39,$K_40_59,$K_60_79);
+
+$ctx="%i0";
+$inp="%i1";
+$len="%i2";
+$tmp0="%i3";
+$tmp1="%i4";
+$tmp2="%i5";
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=($i&1)?@X[($i/2)%8]:$Xi;
+
+$code.=<<___;
+	sll	$a,5,$tmp0		!! $i
+	add	@K[$i/20],$e,$e
+	srl	$a,27,$tmp1
+	add	$tmp0,$e,$e
+	and	$c,$b,$tmp0
+	add	$tmp1,$e,$e
+	sll	$b,30,$tmp2
+	andn	$d,$b,$tmp1
+	srl	$b,2,$b
+	or	$tmp1,$tmp0,$tmp1
+	or	$tmp2,$b,$b
+	add	$xi,$e,$e
+___
+if ($i&1 && $i<15) {
+	$code.=
+	"	srlx	@X[(($i+1)/2)%8],32,$Xi\n";
+}
+$code.=<<___;
+	add	$tmp1,$e,$e
+___
+}
+
+sub Xupdate {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i/2;
+
+if ($i&1) {
+$code.=<<___;
+	sll	$a,5,$tmp0		!! $i
+	add	@K[$i/20],$e,$e
+	srl	$a,27,$tmp1
+___
+} else {
+$code.=<<___;
+	sllx	@X[($j+6)%8],32,$Xi	! Xupdate($i)
+	xor	@X[($j+1)%8],@X[$j%8],@X[$j%8]
+	srlx	@X[($j+7)%8],32,$tmp1
+	xor	@X[($j+4)%8],@X[$j%8],@X[$j%8]
+	sll	$a,5,$tmp0		!! $i
+	or	$tmp1,$Xi,$Xi
+	add	@K[$i/20],$e,$e		!!
+	xor	$Xi,@X[$j%8],@X[$j%8]
+	srlx	@X[$j%8],31,$Xi
+	add	@X[$j%8],@X[$j%8],@X[$j%8]
+	and	$Xi,$rot1m,$Xi
+	andn	@X[$j%8],$rot1m,@X[$j%8]
+	srl	$a,27,$tmp1		!!
+	or	$Xi,@X[$j%8],@X[$j%8]
+___
+}
+}
+
+sub BODY_16_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+
+	&Xupdate(@_);
+    if ($i&1) {
+	$xi=@X[($i/2)%8];
+    } else {
+	$xi=$Xi;
+	$code.="\tsrlx	@X[($i/2)%8],32,$xi\n";
+    }
+$code.=<<___;
+	add	$tmp0,$e,$e		!!
+	and	$c,$b,$tmp0
+	add	$tmp1,$e,$e
+	sll	$b,30,$tmp2
+	add	$xi,$e,$e
+	andn	$d,$b,$tmp1
+	srl	$b,2,$b
+	or	$tmp1,$tmp0,$tmp1
+	or	$tmp2,$b,$b
+	add	$tmp1,$e,$e
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi;
+	&Xupdate(@_);
+    if ($i&1) {
+	$xi=@X[($i/2)%8];
+    } else {
+	$xi=$Xi;
+	$code.="\tsrlx	@X[($i/2)%8],32,$xi\n";
+    }
+$code.=<<___;
+	add	$tmp0,$e,$e		!!
+	xor	$c,$b,$tmp0
+	add	$tmp1,$e,$e
+	sll	$b,30,$tmp2
+	xor	$d,$tmp0,$tmp1
+	srl	$b,2,$b
+	add	$tmp1,$e,$e
+	or	$tmp2,$b,$b
+	add	$xi,$e,$e
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi;
+	&Xupdate(@_);
+    if ($i&1) {
+	$xi=@X[($i/2)%8];
+    } else {
+	$xi=$Xi;
+	$code.="\tsrlx	@X[($i/2)%8],32,$xi\n";
+    }
+$code.=<<___;
+	add	$tmp0,$e,$e		!!
+	and	$c,$b,$tmp0
+	add	$tmp1,$e,$e
+	sll	$b,30,$tmp2
+	or	$c,$b,$tmp1
+	srl	$b,2,$b
+	and	$d,$tmp1,$tmp1
+	add	$xi,$e,$e
+	or	$tmp1,$tmp0,$tmp1
+	or	$tmp2,$b,$b
+	add	$tmp1,$e,$e
+___
+}
+
+$code.=<<___ if ($bits==64);
+.register	%g2,#scratch
+.register	%g3,#scratch
+___
+$code.=<<___;
+.section	".text",#alloc,#execinstr
+
+.align	32
+.globl	sha1_block_data_order
+sha1_block_data_order:
+	save	%sp,-$frame,%sp
+	sllx	$len,6,$len
+	add	$inp,$len,$len
+
+	or	%g0,1,$rot1m
+	sllx	$rot1m,32,$rot1m
+	or	$rot1m,1,$rot1m
+
+	ld	[$ctx+0],$A
+	ld	[$ctx+4],$B
+	ld	[$ctx+8],$C
+	ld	[$ctx+12],$D
+	ld	[$ctx+16],$E
+	andn	$inp,7,$tmp0
+
+	sethi	%hi(0x5a827999),$K_00_19
+	or	$K_00_19,%lo(0x5a827999),$K_00_19
+	sethi	%hi(0x6ed9eba1),$K_20_39
+	or	$K_20_39,%lo(0x6ed9eba1),$K_20_39
+	sethi	%hi(0x8f1bbcdc),$K_40_59
+	or	$K_40_59,%lo(0x8f1bbcdc),$K_40_59
+	sethi	%hi(0xca62c1d6),$K_60_79
+	or	$K_60_79,%lo(0xca62c1d6),$K_60_79
+
+.Lloop:
+	ldx	[$tmp0+0],@X[0]
+	ldx	[$tmp0+16],@X[2]
+	ldx	[$tmp0+32],@X[4]
+	ldx	[$tmp0+48],@X[6]
+	and	$inp,7,$tmp1
+	ldx	[$tmp0+8],@X[1]
+	sll	$tmp1,3,$tmp1
+	ldx	[$tmp0+24],@X[3]
+	subcc	%g0,$tmp1,$tmp2	! should be 64-$tmp1, but -$tmp1 works too
+	ldx	[$tmp0+40],@X[5]
+	bz,pt	%icc,.Laligned
+	ldx	[$tmp0+56],@X[7]
+
+	sllx	@X[0],$tmp1,@X[0]
+	ldx	[$tmp0+64],$tmp64
+___
+for($i=0;$i<7;$i++)
+{   $code.=<<___;
+	srlx	@X[$i+1],$tmp2,$Xi
+	sllx	@X[$i+1],$tmp1,@X[$i+1]
+	or	$Xi,@X[$i],@X[$i]
+___
+}
+$code.=<<___;
+	srlx	$tmp64,$tmp2,$tmp64
+	or	$tmp64,@X[7],@X[7]
+.Laligned:
+	srlx	@X[0],32,$Xi
+___
+for ($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+for (;$i<20;$i++)	{ &BODY_16_19($i,@V); unshift(@V,pop(@V)); }
+for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+for (;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+
+	ld	[$ctx+0],@X[0]
+	ld	[$ctx+4],@X[1]
+	ld	[$ctx+8],@X[2]
+	ld	[$ctx+12],@X[3]
+	add	$inp,64,$inp
+	ld	[$ctx+16],@X[4]
+	cmp	$inp,$len
+
+	add	$A,@X[0],$A
+	st	$A,[$ctx+0]
+	add	$B,@X[1],$B
+	st	$B,[$ctx+4]
+	add	$C,@X[2],$C
+	st	$C,[$ctx+8]
+	add	$D,@X[3],$D
+	st	$D,[$ctx+12]
+	add	$E,@X[4],$E
+	st	$E,[$ctx+16]
+
+	bne	`$bits==64?"%xcc":"%icc"`,.Lloop
+	andn	$inp,7,$tmp0
+
+	ret
+	restore
+.type	sha1_block_data_order,#function
+.size	sha1_block_data_order,(.-sha1_block_data_order)
+.asciz	"SHA1 block transform for SPARCv9, CRYPTOGAMS by <appro\@openssl.org>"
+.align	4
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha1-sparcv9a.pl b/crypto/sha/asm/sha1-sparcv9a.pl
new file mode 100755
index 000000000000..85e8d68086bb
--- /dev/null
+++ b/crypto/sha/asm/sha1-sparcv9a.pl
@@ -0,0 +1,601 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# January 2009
+#
+# Provided that UltraSPARC VIS instructions are pipe-lined(*) and
+# pairable(*) with IALU ones, offloading of Xupdate to the UltraSPARC
+# Graphic Unit would make it possible to achieve higher instruction-
+# level parallelism, ILP, and thus higher performance. It should be
+# explicitly noted that ILP is the keyword, and it means that this
+# code would be unsuitable for cores like UltraSPARC-Tx. The idea is
+# not really novel, Sun had VIS-powered implementation for a while.
+# Unlike Sun's implementation this one can process multiple unaligned
+# input blocks, and as such works as drop-in replacement for OpenSSL
+# sha1_block_data_order. Performance improvement was measured to be
+# 40% over pure IALU sha1-sparcv9.pl on UltraSPARC-IIi, but 12% on
+# UltraSPARC-III. See below for discussion...
+#
+# The module does not present direct interest for OpenSSL, because
+# it doesn't provide better performance on contemporary SPARCv9 CPUs,
+# UltraSPARC-Tx and SPARC64-V[II] to be specific. Those who feel they
+# absolutely must score on UltraSPARC-I-IV can simply replace
+# crypto/sha/asm/sha1-sparcv9.pl with this module.
+#
+# (*)	"Pipe-lined" means that even if it takes several cycles to
+#	complete, next instruction using same functional unit [but not
+#	depending on the result of the current instruction] can start
+#	execution without having to wait for the unit. "Pairable"
+#	means that two [or more] independent instructions can be
+#	issued at the very same time.
+
+$bits=32;
+for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+if ($bits==64)	{ $bias=2047; $frame=192; }
+else		{ $bias=0;    $frame=112; }
+
+$output=shift;
+open STDOUT,">$output";
+
+$ctx="%i0";
+$inp="%i1";
+$len="%i2";
+$tmp0="%i3";
+$tmp1="%i4";
+$tmp2="%i5";
+$tmp3="%g5";
+
+$base="%g1";
+$align="%g4";
+$Xfer="%o5";
+$nXfer=$tmp3;
+$Xi="%o7";
+
+$A="%l0";
+$B="%l1";
+$C="%l2";
+$D="%l3";
+$E="%l4";
+@V=($A,$B,$C,$D,$E);
+
+$Actx="%o0";
+$Bctx="%o1";
+$Cctx="%o2";
+$Dctx="%o3";
+$Ectx="%o4";
+
+$fmul="%f32";
+$VK_00_19="%f34";
+$VK_20_39="%f36";
+$VK_40_59="%f38";
+$VK_60_79="%f40";
+@VK=($VK_00_19,$VK_20_39,$VK_40_59,$VK_60_79);
+@X=("%f0", "%f1", "%f2", "%f3", "%f4", "%f5", "%f6", "%f7",
+    "%f8", "%f9","%f10","%f11","%f12","%f13","%f14","%f15","%f16");
+
+# This is reference 2x-parallelized VIS-powered Xupdate procedure. It
+# covers even K_NN_MM addition...
+sub Xupdate {
+my ($i)=@_;
+my $K=@VK[($i+16)/20];
+my $j=($i+16)%16;
+
+#	[ provided that GSR.alignaddr_offset is 5, $mul contains
+#	  0x100ULL<<32|0x100 value and K_NN_MM are pre-loaded to
+#	  chosen registers... ]
+$code.=<<___;
+	fxors		@X[($j+13)%16],@X[$j],@X[$j]	!-1/-1/-1:X[0]^=X[13]
+	fxors		@X[($j+14)%16],@X[$j+1],@X[$j+1]! 0/ 0/ 0:X[1]^=X[14]
+	fxor		@X[($j+2)%16],@X[($j+8)%16],%f18! 1/ 1/ 1:Tmp=X[2,3]^X[8,9]
+	fxor		%f18,@X[$j],@X[$j]		! 2/ 4/ 3:X[0,1]^=X[2,3]^X[8,9]
+	faligndata	@X[$j],@X[$j],%f18		! 3/ 7/ 5:Tmp=X[0,1]>>>24
+	fpadd32		@X[$j],@X[$j],@X[$j]		! 4/ 8/ 6:X[0,1]<<=1
+	fmul8ulx16	%f18,$fmul,%f18			! 5/10/ 7:Tmp>>=7, Tmp&=1
+	![fxors		%f15,%f2,%f2]
+	for		%f18,@X[$j],@X[$j]		! 8/14/10:X[0,1]|=Tmp
+	![fxors		%f0,%f3,%f3]			!10/17/12:X[0] dependency
+	fpadd32		$K,@X[$j],%f20
+	std		%f20,[$Xfer+`4*$j`]
+___
+# The numbers delimited with slash are the earliest possible dispatch
+# cycles for given instruction assuming 1 cycle latency for simple VIS
+# instructions, such as on UltraSPARC-I&II, 3 cycles latency, such as
+# on UltraSPARC-III&IV, and 2 cycles latency(*), respectively. Being
+# 2x-parallelized the procedure is "worth" 5, 8.5 or 6 ticks per SHA1
+# round. As [long as] FPU/VIS instructions are perfectly pairable with
+# IALU ones, the round timing is defined by the maximum between VIS
+# and IALU timings. The latter varies from round to round and averages
+# out at 6.25 ticks. This means that USI&II should operate at IALU
+# rate, while USIII&IV - at VIS rate. This explains why performance
+# improvement varies among processors. Well, given that pure IALU
+# sha1-sparcv9.pl module exhibits virtually uniform performance of
+# ~9.3 cycles per SHA1 round. Timings mentioned above are theoretical
+# lower limits. Real-life performance was measured to be 6.6 cycles
+# per SHA1 round on USIIi and 8.3 on USIII. The latter is lower than
+# half-round VIS timing, because there are 16 Xupdate-free rounds,
+# which "push down" average theoretical timing to 8 cycles...
+
+# (*)	SPARC64-V[II] was originally believed to have 2 cycles VIS
+#	latency. Well, it might have, but it doesn't have dedicated
+#	VIS-unit. Instead, VIS instructions are executed by other
+#	functional units, ones used here - by IALU. This doesn't
+#	improve effective ILP...
+}
+
+# The reference Xupdate procedure is then "strained" over *pairs* of
+# BODY_NN_MM and kind of modulo-scheduled in respect to X[n]^=X[n+13]
+# and K_NN_MM addition. It's "running" 15 rounds ahead, which leaves
+# plenty of room to amortize for read-after-write hazard, as well as
+# to fetch and align input for the next spin. The VIS instructions are
+# scheduled for latency of 2 cycles, because there are not enough IALU
+# instructions to schedule for latency of 3, while scheduling for 1
+# would give no gain on USI&II anyway.
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i&~1;
+my $k=($j+16+2)%16;	# ahead reference
+my $l=($j+16-2)%16;	# behind reference
+my $K=@VK[($j+16-2)/20];
+
+$j=($j+16)%16;
+
+$code.=<<___ if (!($i&1));
+	sll		$a,5,$tmp0			!! $i
+	and		$c,$b,$tmp3
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fxors		@X[($j+14)%16],@X[$j+1],@X[$j+1]! 0/ 0/ 0:X[1]^=X[14]
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fxor		@X[($j+2)%16],@X[($j+8)%16],%f18! 1/ 1/ 1:Tmp=X[2,3]^X[8,9]
+	sll		$b,30,$tmp2
+	add		$tmp1,$e,$e
+	andn		$d,$b,$tmp1
+	add		$Xi,$e,$e
+	 fxor		%f18,@X[$j],@X[$j]		! 2/ 4/ 3:X[0,1]^=X[2,3]^X[8,9]
+	srl		$b,2,$b
+	or		$tmp1,$tmp3,$tmp1
+	or		$tmp2,$b,$b
+	add		$tmp1,$e,$e
+	 faligndata	@X[$j],@X[$j],%f18		! 3/ 7/ 5:Tmp=X[0,1]>>>24
+___
+$code.=<<___ if ($i&1);
+	sll		$a,5,$tmp0			!! $i
+	and		$c,$b,$tmp3
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fpadd32	@X[$j],@X[$j],@X[$j]		! 4/ 8/ 6:X[0,1]<<=1
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fmul8ulx16	%f18,$fmul,%f18			! 5/10/ 7:Tmp>>=7, Tmp&=1
+	sll		$b,30,$tmp2
+	add		$tmp1,$e,$e
+	 fpadd32	$K,@X[$l],%f20			!
+	andn		$d,$b,$tmp1
+	add		$Xi,$e,$e
+	 fxors		@X[($k+13)%16],@X[$k],@X[$k]	!-1/-1/-1:X[0]^=X[13]
+	srl		$b,2,$b
+	or		$tmp1,$tmp3,$tmp1
+	 fxor		%f18,@X[$j],@X[$j]		! 8/14/10:X[0,1]|=Tmp
+	or		$tmp2,$b,$b
+	add		$tmp1,$e,$e
+___
+$code.=<<___ if ($i&1 && $i>=2);
+	 std		%f20,[$Xfer+`4*$l`]		!
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i&~1;
+my $k=($j+16+2)%16;	# ahead reference
+my $l=($j+16-2)%16;	# behind reference
+my $K=@VK[($j+16-2)/20];
+
+$j=($j+16)%16;
+
+$code.=<<___ if (!($i&1) && $i<64);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fxors		@X[($j+14)%16],@X[$j+1],@X[$j+1]! 0/ 0/ 0:X[1]^=X[14]
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fxor		@X[($j+2)%16],@X[($j+8)%16],%f18! 1/ 1/ 1:Tmp=X[2,3]^X[8,9]
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	 fxor		%f18,@X[$j],@X[$j]		! 2/ 4/ 3:X[0,1]^=X[2,3]^X[8,9]
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+	 faligndata	@X[$j],@X[$j],%f18		! 3/ 7/ 5:Tmp=X[0,1]>>>24
+___
+$code.=<<___ if ($i&1 && $i<64);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fpadd32	@X[$j],@X[$j],@X[$j]		! 4/ 8/ 6:X[0,1]<<=1
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fmul8ulx16	%f18,$fmul,%f18			! 5/10/ 7:Tmp>>=7, Tmp&=1
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	 fpadd32	$K,@X[$l],%f20			!
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	 fxors		@X[($k+13)%16],@X[$k],@X[$k]	!-1/-1/-1:X[0]^=X[13]
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	 fxor		%f18,@X[$j],@X[$j]		! 8/14/10:X[0,1]|=Tmp
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+	 std		%f20,[$Xfer+`4*$l`]		!
+___
+$code.=<<___ if ($i==64);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fpadd32	$K,@X[$l],%f20
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	 std		%f20,[$Xfer+`4*$l`]
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+___
+$code.=<<___ if ($i>64);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i&~1;
+my $k=($j+16+2)%16;	# ahead reference
+my $l=($j+16-2)%16;	# behind reference
+my $K=@VK[($j+16-2)/20];
+
+$j=($j+16)%16;
+
+$code.=<<___ if (!($i&1));
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fxors		@X[($j+14)%16],@X[$j+1],@X[$j+1]! 0/ 0/ 0:X[1]^=X[14]
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fxor		@X[($j+2)%16],@X[($j+8)%16],%f18! 1/ 1/ 1:Tmp=X[2,3]^X[8,9]
+	and		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	or		$c,$b,$tmp1
+	 fxor		%f18,@X[$j],@X[$j]		! 2/ 4/ 3:X[0,1]^=X[2,3]^X[8,9]
+	srl		$b,2,$b
+	and		$d,$tmp1,$tmp1
+	add		$Xi,$e,$e
+	or		$tmp1,$tmp0,$tmp1
+	 faligndata	@X[$j],@X[$j],%f18		! 3/ 7/ 5:Tmp=X[0,1]>>>24
+	or		$tmp2,$b,$b
+	add		$tmp1,$e,$e
+	 fpadd32	@X[$j],@X[$j],@X[$j]		! 4/ 8/ 6:X[0,1]<<=1
+___
+$code.=<<___ if ($i&1);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fmul8ulx16	%f18,$fmul,%f18			! 5/10/ 7:Tmp>>=7, Tmp&=1
+	and		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	 fpadd32	$K,@X[$l],%f20			!
+	sll		$b,30,$tmp2
+	or		$c,$b,$tmp1
+	 fxors		@X[($k+13)%16],@X[$k],@X[$k]	!-1/-1/-1:X[0]^=X[13]
+	srl		$b,2,$b
+	and		$d,$tmp1,$tmp1
+	 fxor		%f18,@X[$j],@X[$j]		! 8/14/10:X[0,1]|=Tmp
+	add		$Xi,$e,$e
+	or		$tmp1,$tmp0,$tmp1
+	or		$tmp2,$b,$b
+	add		$tmp1,$e,$e
+	 std		%f20,[$Xfer+`4*$l`]		!
+___
+}
+
+# If there is more data to process, then we pre-fetch the data for
+# next iteration in last ten rounds...
+sub BODY_70_79 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i&~1;
+my $m=($i%8)*2;
+
+$j=($j+16)%16;
+
+$code.=<<___ if ($i==70);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 ldd		[$inp+64],@X[0]
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+
+	and		$inp,-64,$nXfer
+	inc		64,$inp
+	and		$nXfer,255,$nXfer
+	alignaddr	%g0,$align,%g0
+	add		$base,$nXfer,$nXfer
+___
+$code.=<<___ if ($i==71);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+___
+$code.=<<___ if ($i>=72);
+	 faligndata	@X[$m],@X[$m+2],@X[$m]
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	 fpadd32	$VK_00_19,@X[$m],%f20
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+___
+$code.=<<___ if ($i<77);
+	 ldd		[$inp+`8*($i+1-70)`],@X[2*($i+1-70)]
+___
+$code.=<<___ if ($i==77);	# redundant if $inp was aligned
+	 add		$align,63,$tmp0
+	 and		$tmp0,-8,$tmp0
+	 ldd		[$inp+$tmp0],@X[16]
+___
+$code.=<<___ if ($i>=72);
+	 std		%f20,[$nXfer+`4*$m`]
+___
+}
+
+$code.=<<___;
+.section	".text",#alloc,#execinstr
+
+.align	64
+vis_const:
+.long	0x5a827999,0x5a827999	! K_00_19
+.long	0x6ed9eba1,0x6ed9eba1	! K_20_39
+.long	0x8f1bbcdc,0x8f1bbcdc	! K_40_59
+.long	0xca62c1d6,0xca62c1d6	! K_60_79
+.long	0x00000100,0x00000100
+.align	64
+.type	vis_const,#object
+.size	vis_const,(.-vis_const)
+
+.globl	sha1_block_data_order
+sha1_block_data_order:
+	save	%sp,-$frame,%sp
+	add	%fp,$bias-256,$base
+
+1:	call	.+8
+	add	%o7,vis_const-1b,$tmp0
+
+	ldd	[$tmp0+0],$VK_00_19
+	ldd	[$tmp0+8],$VK_20_39
+	ldd	[$tmp0+16],$VK_40_59
+	ldd	[$tmp0+24],$VK_60_79
+	ldd	[$tmp0+32],$fmul
+
+	ld	[$ctx+0],$Actx
+	and	$base,-256,$base
+	ld	[$ctx+4],$Bctx
+	sub	$base,$bias+$frame,%sp
+	ld	[$ctx+8],$Cctx
+	and	$inp,7,$align
+	ld	[$ctx+12],$Dctx
+	and	$inp,-8,$inp
+	ld	[$ctx+16],$Ectx
+
+	! X[16] is maintained in FP register bank
+	alignaddr	%g0,$align,%g0
+	ldd		[$inp+0],@X[0]
+	sub		$inp,-64,$Xfer
+	ldd		[$inp+8],@X[2]
+	and		$Xfer,-64,$Xfer
+	ldd		[$inp+16],@X[4]
+	and		$Xfer,255,$Xfer
+	ldd		[$inp+24],@X[6]
+	add		$base,$Xfer,$Xfer
+	ldd		[$inp+32],@X[8]
+	ldd		[$inp+40],@X[10]
+	ldd		[$inp+48],@X[12]
+	brz,pt		$align,.Laligned
+	ldd		[$inp+56],@X[14]
+
+	ldd		[$inp+64],@X[16]
+	faligndata	@X[0],@X[2],@X[0]
+	faligndata	@X[2],@X[4],@X[2]
+	faligndata	@X[4],@X[6],@X[4]
+	faligndata	@X[6],@X[8],@X[6]
+	faligndata	@X[8],@X[10],@X[8]
+	faligndata	@X[10],@X[12],@X[10]
+	faligndata	@X[12],@X[14],@X[12]
+	faligndata	@X[14],@X[16],@X[14]
+
+.Laligned:
+	mov		5,$tmp0
+	dec		1,$len
+	alignaddr	%g0,$tmp0,%g0
+	fpadd32		$VK_00_19,@X[0],%f16
+	fpadd32		$VK_00_19,@X[2],%f18
+	fpadd32		$VK_00_19,@X[4],%f20
+	fpadd32		$VK_00_19,@X[6],%f22
+	fpadd32		$VK_00_19,@X[8],%f24
+	fpadd32		$VK_00_19,@X[10],%f26
+	fpadd32		$VK_00_19,@X[12],%f28
+	fpadd32		$VK_00_19,@X[14],%f30
+	std		%f16,[$Xfer+0]
+	mov		$Actx,$A
+	std		%f18,[$Xfer+8]
+	mov		$Bctx,$B
+	std		%f20,[$Xfer+16]
+	mov		$Cctx,$C
+	std		%f22,[$Xfer+24]
+	mov		$Dctx,$D
+	std		%f24,[$Xfer+32]
+	mov		$Ectx,$E
+	std		%f26,[$Xfer+40]
+	fxors		@X[13],@X[0],@X[0]
+	std		%f28,[$Xfer+48]
+	ba		.Loop
+	std		%f30,[$Xfer+56]
+.align	32
+.Loop:
+___
+for ($i=0;$i<20;$i++)	{ &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+for (;$i<70;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	tst		$len
+	bz,pn		`$bits==32?"%icc":"%xcc"`,.Ltail
+	nop
+___
+for (;$i<80;$i++)	{ &BODY_70_79($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	add		$A,$Actx,$Actx
+	add		$B,$Bctx,$Bctx
+	add		$C,$Cctx,$Cctx
+	add		$D,$Dctx,$Dctx
+	add		$E,$Ectx,$Ectx
+	mov		5,$tmp0
+	fxors		@X[13],@X[0],@X[0]
+	mov		$Actx,$A
+	mov		$Bctx,$B
+	mov		$Cctx,$C
+	mov		$Dctx,$D
+	mov		$Ectx,$E
+	alignaddr	%g0,$tmp0,%g0	
+	dec		1,$len
+	ba		.Loop
+	mov		$nXfer,$Xfer
+
+.align	32
+.Ltail:
+___
+for($i=70;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	add	$A,$Actx,$Actx
+	add	$B,$Bctx,$Bctx
+	add	$C,$Cctx,$Cctx
+	add	$D,$Dctx,$Dctx
+	add	$E,$Ectx,$Ectx
+
+	st	$Actx,[$ctx+0]
+	st	$Bctx,[$ctx+4]
+	st	$Cctx,[$ctx+8]
+	st	$Dctx,[$ctx+12]
+	st	$Ectx,[$ctx+16]
+
+	ret
+	restore
+.type	sha1_block_data_order,#function
+.size	sha1_block_data_order,(.-sha1_block_data_order)
+.asciz	"SHA1 block transform for SPARCv9a, CRYPTOGAMS by <appro\@openssl.org>"
+.align	4
+___
+
+# Purpose of these subroutines is to explicitly encode VIS instructions,
+# so that one can compile the module without having to specify VIS
+# extentions on compiler command line, e.g. -xarch=v9 vs. -xarch=v9a.
+# Idea is to reserve for option to produce "universal" binary and let
+# programmer detect if current CPU is VIS capable at run-time.
+sub unvis {
+my ($mnemonic,$rs1,$rs2,$rd)=@_;
+my $ref,$opf;
+my %visopf = (	"fmul8ulx16"	=> 0x037,
+		"faligndata"	=> 0x048,
+		"fpadd32"	=> 0x052,
+		"fxor"		=> 0x06c,
+		"fxors"		=> 0x06d	);
+
+    $ref = "$mnemonic\t$rs1,$rs2,$rd";
+
+    if ($opf=$visopf{$mnemonic}) {
+	foreach ($rs1,$rs2,$rd) {
+	    return $ref if (!/%f([0-9]{1,2})/);
+	    $_=$1;
+	    if ($1>=32) {
+		return $ref if ($1&1);
+		# re-encode for upper double register addressing
+		$_=($1|$1>>5)&31;
+	    }
+	}
+
+	return	sprintf ".word\t0x%08x !%s",
+			0x81b00000|$rd<<25|$rs1<<14|$opf<<5|$rs2,
+			$ref;
+    } else {
+	return $ref;
+    }
+}
+sub unalignaddr {
+my ($mnemonic,$rs1,$rs2,$rd)=@_;
+my %bias = ( "g" => 0, "o" => 8, "l" => 16, "i" => 24 );
+my $ref="$mnemonic\t$rs1,$rs2,$rd";
+
+    foreach ($rs1,$rs2,$rd) {
+	if (/%([goli])([0-7])/)	{ $_=$bias{$1}+$2; }
+	else			{ return $ref; }
+    }
+    return  sprintf ".word\t0x%08x !%s",
+		    0x81b00300|$rd<<25|$rs1<<14|$rs2,
+		    $ref;
+}
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/\b(f[^\s]*)\s+(%f[0-9]{1,2}),(%f[0-9]{1,2}),(%f[0-9]{1,2})/
+		&unvis($1,$2,$3,$4)
+	  /gem;
+$code =~ s/\b(alignaddr)\s+(%[goli][0-7]),(%[goli][0-7]),(%[goli][0-7])/
+		&unalignaddr($1,$2,$3,$4)
+	  /gem;
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha1-thumb.pl b/crypto/sha/asm/sha1-thumb.pl
new file mode 100755
index 000000000000..7c9ea9b0296c
--- /dev/null
+++ b/crypto/sha/asm/sha1-thumb.pl
@@ -0,0 +1,259 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# sha1_block for Thumb.
+#
+# January 2007.
+#
+# The code does not present direct interest to OpenSSL, because of low
+# performance. Its purpose is to establish _size_ benchmark. Pretty
+# useless one I must say, because 30% or 88 bytes larger ARMv4 code
+# [avialable on demand] is almost _twice_ as fast. It should also be
+# noted that in-lining of .Lcommon and .Lrotate improves performance
+# by over 40%, while code increases by only 10% or 32 bytes. But once
+# again, the goal was to establish _size_ benchmark, not performance.
+
+$output=shift;
+open STDOUT,">$output";
+
+$inline=0;
+#$cheat_on_binutils=1;
+
+$t0="r0";
+$t1="r1";
+$t2="r2";
+$a="r3";
+$b="r4";
+$c="r5";
+$d="r6";
+$e="r7";
+$K="r8";	# "upper" registers can be used in add/sub and mov insns
+$ctx="r9";
+$inp="r10";
+$len="r11";
+$Xi="r12";
+
+sub common {
+<<___;
+	sub	$t0,#4
+	ldr	$t1,[$t0]
+	add	$e,$K			@ E+=K_xx_xx
+	lsl	$t2,$a,#5
+	add	$t2,$e
+	lsr	$e,$a,#27
+	add	$t2,$e			@ E+=ROR(A,27)
+	add	$t2,$t1			@ E+=X[i]
+___
+}
+sub rotate {
+<<___;
+	mov	$e,$d			@ E=D
+	mov	$d,$c			@ D=C
+	lsl	$c,$b,#30
+	lsr	$b,$b,#2
+	orr	$c,$b			@ C=ROR(B,2)
+	mov	$b,$a			@ B=A
+	add	$a,$t2,$t1		@ A=E+F_xx_xx(B,C,D)
+___
+}
+
+sub BODY_00_19 {
+$code.=$inline?&common():"\tbl	.Lcommon\n";
+$code.=<<___;
+	mov	$t1,$c
+	eor	$t1,$d
+	and	$t1,$b
+	eor	$t1,$d			@ F_00_19(B,C,D)
+___
+$code.=$inline?&rotate():"\tbl	.Lrotate\n";
+}
+
+sub BODY_20_39 {
+$code.=$inline?&common():"\tbl	.Lcommon\n";
+$code.=<<___;
+	mov	$t1,$b
+	eor	$t1,$c
+	eor	$t1,$d			@ F_20_39(B,C,D)
+___
+$code.=$inline?&rotate():"\tbl	.Lrotate\n";
+}
+
+sub BODY_40_59 {
+$code.=$inline?&common():"\tbl	.Lcommon\n";
+$code.=<<___;
+	mov	$t1,$b
+	and	$t1,$c
+	mov	$e,$b
+	orr	$e,$c
+	and	$e,$d
+	orr	$t1,$e			@ F_40_59(B,C,D)
+___
+$code.=$inline?&rotate():"\tbl	.Lrotate\n";
+}
+
+$code=<<___;
+.text
+.code	16
+
+.global	sha1_block_data_order
+.type	sha1_block_data_order,%function
+
+.align	2
+sha1_block_data_order:
+___
+if ($cheat_on_binutils) {
+$code.=<<___;
+.code	32
+	add	r3,pc,#1
+	bx	r3			@ switch to Thumb ISA
+.code	16
+___
+}
+$code.=<<___;
+	push	{r4-r7}
+	mov	r3,r8
+	mov	r4,r9
+	mov	r5,r10
+	mov	r6,r11
+	mov	r7,r12
+	push	{r3-r7,lr}
+	lsl	r2,#6
+	mov	$ctx,r0			@ save context
+	mov	$inp,r1			@ save inp
+	mov	$len,r2			@ save len
+	add	$len,$inp		@ $len to point at inp end
+
+.Lloop:
+	mov	$Xi,sp
+	mov	$t2,sp
+	sub	$t2,#16*4		@ [3]
+.LXload:
+	ldrb	$a,[$t1,#0]		@ $t1 is r1 and holds inp
+	ldrb	$b,[$t1,#1]
+	ldrb	$c,[$t1,#2]
+	ldrb	$d,[$t1,#3]
+	lsl	$a,#24
+	lsl	$b,#16
+	lsl	$c,#8
+	orr	$a,$b
+	orr	$a,$c
+	orr	$a,$d
+	add	$t1,#4
+	push	{$a}
+	cmp	sp,$t2
+	bne	.LXload			@ [+14*16]
+
+	mov	$inp,$t1		@ update $inp
+	sub	$t2,#32*4
+	sub	$t2,#32*4
+	mov	$e,#31			@ [+4]
+.LXupdate:
+	ldr	$a,[sp,#15*4]
+	ldr	$b,[sp,#13*4]
+	ldr	$c,[sp,#7*4]
+	ldr	$d,[sp,#2*4]
+	eor	$a,$b
+	eor	$a,$c
+	eor	$a,$d
+	ror	$a,$e
+	push	{$a}
+	cmp	sp,$t2
+	bne	.LXupdate		@ [+(11+1)*64]
+
+	ldmia	$t0!,{$a,$b,$c,$d,$e}	@ $t0 is r0 and holds ctx
+	mov	$t0,$Xi
+
+	ldr	$t2,.LK_00_19
+	mov	$t1,$t0
+	sub	$t1,#20*4
+	mov	$Xi,$t1
+	mov	$K,$t2			@ [+7+4]
+.L_00_19:
+___
+	&BODY_00_19();
+$code.=<<___;
+	cmp	$Xi,$t0
+	bne	.L_00_19		@ [+(2+9+4+2+8+2)*20]
+
+	ldr	$t2,.LK_20_39
+	mov	$t1,$t0
+	sub	$t1,#20*4
+	mov	$Xi,$t1
+	mov	$K,$t2			@ [+5]
+.L_20_39_or_60_79:
+___
+	&BODY_20_39();
+$code.=<<___;
+	cmp	$Xi,$t0
+	bne	.L_20_39_or_60_79	@ [+(2+9+3+2+8+2)*20*2]
+	cmp	sp,$t0
+	beq	.Ldone			@ [+2]
+
+	ldr	$t2,.LK_40_59
+	mov	$t1,$t0
+	sub	$t1,#20*4
+	mov	$Xi,$t1
+	mov	$K,$t2			@ [+5]
+.L_40_59:
+___
+	&BODY_40_59();
+$code.=<<___;
+	cmp	$Xi,$t0
+	bne	.L_40_59		@ [+(2+9+6+2+8+2)*20]
+
+	ldr	$t2,.LK_60_79
+	mov	$Xi,sp
+	mov	$K,$t2
+	b	.L_20_39_or_60_79	@ [+4]
+.Ldone:
+	mov	$t0,$ctx
+	ldr	$t1,[$t0,#0]
+	ldr	$t2,[$t0,#4]
+	add	$a,$t1
+	ldr	$t1,[$t0,#8]
+	add	$b,$t2
+	ldr	$t2,[$t0,#12]
+	add	$c,$t1
+	ldr	$t1,[$t0,#16]
+	add	$d,$t2
+	add	$e,$t1
+	stmia	$t0!,{$a,$b,$c,$d,$e}	@ [+20]
+
+	add	sp,#80*4		@ deallocate stack frame
+	mov	$t0,$ctx		@ restore ctx
+	mov	$t1,$inp		@ restore inp
+	cmp	$t1,$len
+	beq	.Lexit
+	b	.Lloop			@ [+6] total 3212 cycles
+.Lexit:
+	pop	{r2-r7}
+	mov	r8,r2
+	mov	r9,r3
+	mov	r10,r4
+	mov	r11,r5
+	mov	r12,r6
+	mov	lr,r7
+	pop	{r4-r7}
+	bx	lr
+.align	2
+___
+$code.=".Lcommon:\n".&common()."\tmov	pc,lr\n" if (!$inline);
+$code.=".Lrotate:\n".&rotate()."\tmov	pc,lr\n" if (!$inline);
+$code.=<<___;
+.align	2
+.LK_00_19:	.word	0x5a827999
+.LK_20_39:	.word	0x6ed9eba1
+.LK_40_59:	.word	0x8f1bbcdc
+.LK_60_79:	.word	0xca62c1d6
+.size	sha1_block_data_order,.-sha1_block_data_order
+.asciz	"SHA1 block transform for Thumb, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+print $code;
+close STDOUT; # enforce flush
diff --git a/crypto/sha/asm/sha1-x86_64.pl b/crypto/sha/asm/sha1-x86_64.pl
index f7ed67a72609..f27c1e3fb031 100755
--- a/crypto/sha/asm/sha1-x86_64.pl
+++ b/crypto/sha/asm/sha1-x86_64.pl
@@ -16,7 +16,7 @@
 # There was suggestion to mechanically translate 32-bit code, but I
 # dismissed it, reasoning that x86_64 offers enough register bank
 # capacity to fully utilize SHA-1 parallelism. Therefore this fresh
-# implementation:-) However! While 64-bit code does performs better
+# implementation:-) However! While 64-bit code does perform better
 # on Opteron, I failed to beat 32-bit assembler on EM64T core. Well,
 # x86_64 does offer larger *addressable* bank, but out-of-order core
 # reaches for even more registers through dynamic aliasing, and EM64T
@@ -29,14 +29,60 @@
 # Xeon P4	+65%		+0%		9.9
 # Core2		+60%		+10%		7.0
 
-$output=shift;
+# August 2009.
+#
+# The code was revised to minimize code size and to maximize
+# "distance" between instructions producing input to 'lea'
+# instruction and the 'lea' instruction itself, which is essential
+# for Intel Atom core.
+
+# October 2010.
+#
+# Add SSSE3, Supplemental[!] SSE3, implementation. The idea behind it
+# is to offload message schedule denoted by Wt in NIST specification,
+# or Xupdate in OpenSSL source, to SIMD unit. See sha1-586.pl module
+# for background and implementation details. The only difference from
+# 32-bit code is that 64-bit code doesn't have to spill @X[] elements
+# to free temporary registers.
+
+# April 2011.
+#
+# Add AVX code path. See sha1-586.pl for further information.
+
+######################################################################
+# Current performance is summarized in following table. Numbers are
+# CPU clock cycles spent to process single byte (less is better).
+#
+#		x86_64		SSSE3		AVX
+# P4		9.8		-
+# Opteron	6.6		-
+# Core2		6.7		6.1/+10%	-
+# Atom		11.0		9.7/+13%	-
+# Westmere	7.1		5.6/+27%	-
+# Sandy Bridge	7.9		6.3/+25%	5.2/+51%
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
 
 $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
 ( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
 ( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
 die "can't locate x86_64-xlate.pl";
 
-open STDOUT,"| $^X $xlate $output";
+$avx=1 if (`$ENV{CC} -Wa,-v -c -o /dev/null -x assembler /dev/null 2>&1`
+		=~ /GNU assembler version ([2-9]\.[0-9]+)/ &&
+	   $1>=2.19);
+$avx=1 if (!$avx && $win64 && ($flavour =~ /nasm/ || $ENV{ASM} =~ /nasm/) &&
+	   `nasm -v 2>&1` =~ /NASM version ([2-9]\.[0-9]+)/ &&
+	   $1>=2.09);
+$avx=1 if (!$avx && $win64 && ($flavour =~ /masm/ || $ENV{ASM} =~ /ml64/) &&
+	   `ml64 2>&1` =~ /Version ([0-9]+)\./ &&
+	   $1>=10);
+
+open STDOUT,"| $^X $xlate $flavour $output";
 
 $ctx="%rdi";	# 1st arg
 $inp="%rsi";	# 2nd arg
@@ -47,193 +93,1165 @@ $ctx="%r8";
 $inp="%r9";
 $num="%r10";
 
-$xi="%eax";
-$t0="%ebx";
-$t1="%ecx";
-$A="%edx";
-$B="%esi";
-$C="%edi";
-$D="%ebp";
-$E="%r11d";
-$T="%r12d";
-
-@V=($A,$B,$C,$D,$E,$T);
+$t0="%eax";
+$t1="%ebx";
+$t2="%ecx";
+@xi=("%edx","%ebp");
+$A="%esi";
+$B="%edi";
+$C="%r11d";
+$D="%r12d";
+$E="%r13d";
 
-sub PROLOGUE {
-my $func=shift;
-$code.=<<___;
-.globl	$func
-.type	$func,\@function,3
-.align	16
-$func:
-	push	%rbx
-	push	%rbp
-	push	%r12
-	mov	%rsp,%rax
-	mov	%rdi,$ctx	# reassigned argument
-	sub	\$`8+16*4`,%rsp
-	mov	%rsi,$inp	# reassigned argument
-	and	\$-64,%rsp
-	mov	%rdx,$num	# reassigned argument
-	mov	%rax,`16*4`(%rsp)
-
-	mov	0($ctx),$A
-	mov	4($ctx),$B
-	mov	8($ctx),$C
-	mov	12($ctx),$D
-	mov	16($ctx),$E
-___
-}
-
-sub EPILOGUE {
-my $func=shift;
-$code.=<<___;
-	mov	`16*4`(%rsp),%rsp
-	pop	%r12
-	pop	%rbp
-	pop	%rbx
-	ret
-.size	$func,.-$func
-___
-}
+@V=($A,$B,$C,$D,$E);
 
 sub BODY_00_19 {
-my ($i,$a,$b,$c,$d,$e,$f,$host)=@_;
+my ($i,$a,$b,$c,$d,$e)=@_;
 my $j=$i+1;
 $code.=<<___ if ($i==0);
-	mov	`4*$i`($inp),$xi	
-	`"bswap	$xi"	if(!defined($host))`
-	mov	$xi,`4*$i`(%rsp)
+	mov	`4*$i`($inp),$xi[0]
+	bswap	$xi[0]
+	mov	$xi[0],`4*$i`(%rsp)
 ___
 $code.=<<___ if ($i<15);
-	lea	0x5a827999($xi,$e),$f
 	mov	$c,$t0
-	mov	`4*$j`($inp),$xi
-	mov	$a,$e
+	mov	`4*$j`($inp),$xi[1]
+	mov	$a,$t2
 	xor	$d,$t0
-	`"bswap	$xi"	if(!defined($host))`	
-	rol	\$5,$e
+	bswap	$xi[1]
+	rol	\$5,$t2
+	lea	0x5a827999($xi[0],$e),$e
 	and	$b,$t0
-	mov	$xi,`4*$j`(%rsp)
-	add	$e,$f
+	mov	$xi[1],`4*$j`(%rsp)
+	add	$t2,$e
 	xor	$d,$t0
 	rol	\$30,$b
-	add	$t0,$f
+	add	$t0,$e
 ___
 $code.=<<___ if ($i>=15);
-	lea	0x5a827999($xi,$e),$f
-	mov	`4*($j%16)`(%rsp),$xi
+	mov	`4*($j%16)`(%rsp),$xi[1]
 	mov	$c,$t0
-	mov	$a,$e
-	xor	`4*(($j+2)%16)`(%rsp),$xi
+	mov	$a,$t2
+	xor	`4*(($j+2)%16)`(%rsp),$xi[1]
 	xor	$d,$t0
-	rol	\$5,$e
-	xor	`4*(($j+8)%16)`(%rsp),$xi
+	rol	\$5,$t2
+	xor	`4*(($j+8)%16)`(%rsp),$xi[1]
 	and	$b,$t0
-	add	$e,$f
-	xor	`4*(($j+13)%16)`(%rsp),$xi
+	lea	0x5a827999($xi[0],$e),$e
+	xor	`4*(($j+13)%16)`(%rsp),$xi[1]
 	xor	$d,$t0
+	rol	\$1,$xi[1]
+	add	$t2,$e
 	rol	\$30,$b
-	add	$t0,$f
-	rol	\$1,$xi
-	mov	$xi,`4*($j%16)`(%rsp)
+	mov	$xi[1],`4*($j%16)`(%rsp)
+	add	$t0,$e
 ___
+unshift(@xi,pop(@xi));
 }
 
 sub BODY_20_39 {
-my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my ($i,$a,$b,$c,$d,$e)=@_;
 my $j=$i+1;
 my $K=($i<40)?0x6ed9eba1:0xca62c1d6;
 $code.=<<___ if ($i<79);
-	lea	$K($xi,$e),$f
-	mov	`4*($j%16)`(%rsp),$xi
+	mov	`4*($j%16)`(%rsp),$xi[1]
 	mov	$c,$t0
-	mov	$a,$e
-	xor	`4*(($j+2)%16)`(%rsp),$xi
+	mov	$a,$t2
+	xor	`4*(($j+2)%16)`(%rsp),$xi[1]
 	xor	$b,$t0
-	rol	\$5,$e
-	xor	`4*(($j+8)%16)`(%rsp),$xi
+	rol	\$5,$t2
+	lea	$K($xi[0],$e),$e
+	xor	`4*(($j+8)%16)`(%rsp),$xi[1]
 	xor	$d,$t0
-	add	$e,$f
-	xor	`4*(($j+13)%16)`(%rsp),$xi
+	add	$t2,$e
+	xor	`4*(($j+13)%16)`(%rsp),$xi[1]
 	rol	\$30,$b
-	add	$t0,$f
-	rol	\$1,$xi
+	add	$t0,$e
+	rol	\$1,$xi[1]
 ___
 $code.=<<___ if ($i<76);
-	mov	$xi,`4*($j%16)`(%rsp)
+	mov	$xi[1],`4*($j%16)`(%rsp)
 ___
 $code.=<<___ if ($i==79);
-	lea	$K($xi,$e),$f
 	mov	$c,$t0
-	mov	$a,$e
+	mov	$a,$t2
 	xor	$b,$t0
-	rol	\$5,$e
+	lea	$K($xi[0],$e),$e
+	rol	\$5,$t2
 	xor	$d,$t0
-	add	$e,$f
+	add	$t2,$e
 	rol	\$30,$b
-	add	$t0,$f
+	add	$t0,$e
 ___
+unshift(@xi,pop(@xi));
 }
 
 sub BODY_40_59 {
-my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my ($i,$a,$b,$c,$d,$e)=@_;
 my $j=$i+1;
 $code.=<<___;
-	lea	0x8f1bbcdc($xi,$e),$f
-	mov	`4*($j%16)`(%rsp),$xi
-	mov	$b,$t0
-	mov	$b,$t1
-	xor	`4*(($j+2)%16)`(%rsp),$xi
-	mov	$a,$e
-	and	$c,$t0
-	xor	`4*(($j+8)%16)`(%rsp),$xi
-	or	$c,$t1
-	rol	\$5,$e
-	xor	`4*(($j+13)%16)`(%rsp),$xi
-	and	$d,$t1
-	add	$e,$f
-	rol	\$1,$xi
-	or	$t1,$t0
+	mov	`4*($j%16)`(%rsp),$xi[1]
+	mov	$c,$t0
+	mov	$c,$t1
+	xor	`4*(($j+2)%16)`(%rsp),$xi[1]
+	and	$d,$t0
+	mov	$a,$t2
+	xor	`4*(($j+8)%16)`(%rsp),$xi[1]
+	xor	$d,$t1
+	lea	0x8f1bbcdc($xi[0],$e),$e
+	rol	\$5,$t2
+	xor	`4*(($j+13)%16)`(%rsp),$xi[1]
+	add	$t0,$e
+	and	$b,$t1
+	rol	\$1,$xi[1]
+	add	$t1,$e
 	rol	\$30,$b
-	mov	$xi,`4*($j%16)`(%rsp)
-	add	$t0,$f
+	mov	$xi[1],`4*($j%16)`(%rsp)
+	add	$t2,$e
 ___
+unshift(@xi,pop(@xi));
 }
 
-$code=".text\n";
+$code.=<<___;
+.text
+.extern	OPENSSL_ia32cap_P
+
+.globl	sha1_block_data_order
+.type	sha1_block_data_order,\@function,3
+.align	16
+sha1_block_data_order:
+	mov	OPENSSL_ia32cap_P+0(%rip),%r9d
+	mov	OPENSSL_ia32cap_P+4(%rip),%r8d
+	test	\$`1<<9`,%r8d		# check SSSE3 bit
+	jz	.Lialu
+___
+$code.=<<___ if ($avx);
+	and	\$`1<<28`,%r8d		# mask AVX bit
+	and	\$`1<<30`,%r9d		# mask "Intel CPU" bit
+	or	%r9d,%r8d
+	cmp	\$`1<<28|1<<30`,%r8d
+	je	_avx_shortcut
+___
+$code.=<<___;
+	jmp	_ssse3_shortcut
 
-&PROLOGUE("sha1_block_data_order");
-$code.=".align	4\n.Lloop:\n";
+.align	16
+.Lialu:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	mov	%rsp,%r11
+	mov	%rdi,$ctx	# reassigned argument
+	sub	\$`8+16*4`,%rsp
+	mov	%rsi,$inp	# reassigned argument
+	and	\$-64,%rsp
+	mov	%rdx,$num	# reassigned argument
+	mov	%r11,`16*4`(%rsp)
+.Lprologue:
+
+	mov	0($ctx),$A
+	mov	4($ctx),$B
+	mov	8($ctx),$C
+	mov	12($ctx),$D
+	mov	16($ctx),$E
+	jmp	.Lloop
+
+.align	16
+.Lloop:
+___
 for($i=0;$i<20;$i++)	{ &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
 for(;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
 for(;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
 for(;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
 $code.=<<___;
-	add	0($ctx),$E
-	add	4($ctx),$T
-	add	8($ctx),$A
-	add	12($ctx),$B
-	add	16($ctx),$C
-	mov	$E,0($ctx)
-	mov	$T,4($ctx)
-	mov	$A,8($ctx)
-	mov	$B,12($ctx)
-	mov	$C,16($ctx)
-
-	xchg	$E,$A	# mov	$E,$A
-	xchg	$T,$B	# mov	$T,$B
-	xchg	$E,$C	# mov	$A,$C
-	xchg	$T,$D	# mov	$B,$D
-			# mov	$C,$E
-	lea	`16*4`($inp),$inp
+	add	0($ctx),$A
+	add	4($ctx),$B
+	add	8($ctx),$C
+	add	12($ctx),$D
+	add	16($ctx),$E
+	mov	$A,0($ctx)
+	mov	$B,4($ctx)
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+
 	sub	\$1,$num
+	lea	`16*4`($inp),$inp
 	jnz	.Lloop
+
+	mov	`16*4`(%rsp),%rsi
+	mov	(%rsi),%r13
+	mov	8(%rsi),%r12
+	mov	16(%rsi),%rbp
+	mov	24(%rsi),%rbx
+	lea	32(%rsi),%rsp
+.Lepilogue:
+	ret
+.size	sha1_block_data_order,.-sha1_block_data_order
+___
+{{{
+my $Xi=4;
+my @X=map("%xmm$_",(4..7,0..3));
+my @Tx=map("%xmm$_",(8..10));
+my @V=($A,$B,$C,$D,$E)=("%eax","%ebx","%ecx","%edx","%ebp");	# size optimization
+my @T=("%esi","%edi");
+my $j=0;
+my $K_XX_XX="%r11";
+
+my $_rol=sub { &rol(@_) };
+my $_ror=sub { &ror(@_) };
+
+$code.=<<___;
+.type	sha1_block_data_order_ssse3,\@function,3
+.align	16
+sha1_block_data_order_ssse3:
+_ssse3_shortcut:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	lea	`-64-($win64?5*16:0)`(%rsp),%rsp
+___
+$code.=<<___ if ($win64);
+	movaps	%xmm6,64+0(%rsp)
+	movaps	%xmm7,64+16(%rsp)
+	movaps	%xmm8,64+32(%rsp)
+	movaps	%xmm9,64+48(%rsp)
+	movaps	%xmm10,64+64(%rsp)
+.Lprologue_ssse3:
+___
+$code.=<<___;
+	mov	%rdi,$ctx	# reassigned argument
+	mov	%rsi,$inp	# reassigned argument
+	mov	%rdx,$num	# reassigned argument
+
+	shl	\$6,$num
+	add	$inp,$num
+	lea	K_XX_XX(%rip),$K_XX_XX
+
+	mov	0($ctx),$A		# load context
+	mov	4($ctx),$B
+	mov	8($ctx),$C
+	mov	12($ctx),$D
+	mov	$B,@T[0]		# magic seed
+	mov	16($ctx),$E
+
+	movdqa	64($K_XX_XX),@X[2]	# pbswap mask
+	movdqa	0($K_XX_XX),@Tx[1]	# K_00_19
+	movdqu	0($inp),@X[-4&7]	# load input to %xmm[0-3]
+	movdqu	16($inp),@X[-3&7]
+	movdqu	32($inp),@X[-2&7]
+	movdqu	48($inp),@X[-1&7]
+	pshufb	@X[2],@X[-4&7]		# byte swap
+	add	\$64,$inp
+	pshufb	@X[2],@X[-3&7]
+	pshufb	@X[2],@X[-2&7]
+	pshufb	@X[2],@X[-1&7]
+	paddd	@Tx[1],@X[-4&7]		# add K_00_19
+	paddd	@Tx[1],@X[-3&7]
+	paddd	@Tx[1],@X[-2&7]
+	movdqa	@X[-4&7],0(%rsp)	# X[]+K xfer to IALU
+	psubd	@Tx[1],@X[-4&7]		# restore X[]
+	movdqa	@X[-3&7],16(%rsp)
+	psubd	@Tx[1],@X[-3&7]
+	movdqa	@X[-2&7],32(%rsp)
+	psubd	@Tx[1],@X[-2&7]
+	jmp	.Loop_ssse3
+___
+
+sub AUTOLOAD()		# thunk [simplified] 32-bit style perlasm
+{ my $opcode = $AUTOLOAD; $opcode =~ s/.*:://;
+  my $arg = pop;
+    $arg = "\$$arg" if ($arg*1 eq $arg);
+    $code .= "\t$opcode\t".join(',',$arg,reverse @_)."\n";
+}
+
+sub Xupdate_ssse3_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&movdqa	(@X[0],@X[-3&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(@Tx[0],@X[-1&7]);
+	&palignr(@X[0],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &paddd	(@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psrldq	(@Tx[0],4);		# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@Tx[0],@X[-2&7]);	# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@Tx[0]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&movdqa	(@Tx[2],@X[0]);
+	&movdqa	(@Tx[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslldq	(@Tx[2],12);		# "X[0]"<<96, extract one dword
+	&paddd	(@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@Tx[0],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(@Tx[1],@Tx[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@Tx[2],30);
+	&por	(@X[0],@Tx[0]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslld	(@Tx[1],2);
+	&pxor	(@X[0],@Tx[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(@Tx[2],eval(16*(($Xi)/5))."($K_XX_XX)");	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@Tx[1]);		# "X[0]"^=("X[0]">>96)<<<2
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xupdate_ssse3_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&movdqa	(@Tx[0],@X[-1&7])	if ($Xi==8);
+	 eval(shift(@insns));		# body_20_39
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"="X[-32]"^"X[-16]"
+	&palignr(@Tx[0],@X[-2&7],8);	# compose "X[-6]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&pxor	(@X[0],@X[-7&7]);	# "X[0]"^="X[-28]"
+	 eval(shift(@insns));
+	 eval(shift(@insns))	if (@insns[0] !~ /&ro[rl]/);
+	if ($Xi%5) {
+	  &movdqa	(@Tx[2],@Tx[1]);# "perpetuate" K_XX_XX...
+	} else {			# ... or load next one
+	  &movdqa	(@Tx[2],eval(16*($Xi/5))."($K_XX_XX)");
+	}
+	  &paddd	(@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@Tx[0]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&movdqa	(@Tx[0],@X[0]);
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pslld	(@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	&psrld	(@Tx[0],30);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&por	(@X[0],@Tx[0]);		# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &movdqa	(@Tx[1],@X[0])	if ($Xi<19);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xuplast_ssse3_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &paddd	(@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&cmp	($inp,$num);
+	&je	(".Ldone_ssse3");
+
+	unshift(@Tx,pop(@Tx));
+
+	&movdqa	(@X[2],"64($K_XX_XX)");		# pbswap mask
+	&movdqa	(@Tx[1],"0($K_XX_XX)");		# K_00_19
+	&movdqu	(@X[-4&7],"0($inp)");		# load input
+	&movdqu	(@X[-3&7],"16($inp)");
+	&movdqu	(@X[-2&7],"32($inp)");
+	&movdqu	(@X[-1&7],"48($inp)");
+	&pshufb	(@X[-4&7],@X[2]);		# byte swap
+	&add	($inp,64);
+
+  $Xi=0;
+}
+
+sub Xloop_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pshufb	(@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&paddd	(@X[($Xi-4)&7],@Tx[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(eval(16*$Xi)."(%rsp)",@X[($Xi-4)&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psubd	(@X[($Xi-4)&7],@Tx[1]);
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+sub body_00_19 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,eval(4*($j&15))."(%rsp)");',	# X[]+K xfer
+	'&xor	($c,$d);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&xor	($c,$d);',	# restore $c
+	'&xor	(@T[0],$d);',
+	'&add	($e,$a);',
+	'&$_ror	($b,$j?7:2);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'$j++; unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+sub body_20_39 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,eval(4*($j++&15))."(%rsp)");',	# X[]+K xfer
+	'&xor	(@T[0],$d);',	# ($b^$d)
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&xor	(@T[0],$c);',	# ($b^$d^$c)
+	'&add	($e,$a);',
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+sub body_40_59 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&mov	(@T[1],$c);',
+	'&xor	($c,$d);',
+	'&add	($e,eval(4*($j++&15))."(%rsp)");',	# X[]+K xfer
+	'&and	(@T[1],$d);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[1]);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&add	($e,@T[0]);',
+	'&xor	($c,$d);',	# restore $c
+	'&add	($e,$a);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+$code.=<<___;
+.align	16
+.Loop_ssse3:
+___
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xuplast_ssse3_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+
+$code.=<<___;
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	add	12($ctx),$D
+	mov	$A,0($ctx)
+	add	16($ctx),$E
+	mov	@T[0],4($ctx)
+	mov	@T[0],$B			# magic seed
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+	jmp	.Loop_ssse3
+
+.align	16
+.Ldone_ssse3:
+___
+				$j=$saved_j; @V=@saved_V;
+
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+
+$code.=<<___;
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	mov	$A,0($ctx)
+	add	12($ctx),$D
+	mov	@T[0],4($ctx)
+	add	16($ctx),$E
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+___
+$code.=<<___ if ($win64);
+	movaps	64+0(%rsp),%xmm6
+	movaps	64+16(%rsp),%xmm7
+	movaps	64+32(%rsp),%xmm8
+	movaps	64+48(%rsp),%xmm9
+	movaps	64+64(%rsp),%xmm10
+___
+$code.=<<___;
+	lea	`64+($win64?5*16:0)`(%rsp),%rsi
+	mov	0(%rsi),%r12
+	mov	8(%rsi),%rbp
+	mov	16(%rsi),%rbx
+	lea	24(%rsi),%rsp
+.Lepilogue_ssse3:
+	ret
+.size	sha1_block_data_order_ssse3,.-sha1_block_data_order_ssse3
+___
+
+if ($avx) {
+my $Xi=4;
+my @X=map("%xmm$_",(4..7,0..3));
+my @Tx=map("%xmm$_",(8..10));
+my @V=($A,$B,$C,$D,$E)=("%eax","%ebx","%ecx","%edx","%ebp");	# size optimization
+my @T=("%esi","%edi");
+my $j=0;
+my $K_XX_XX="%r11";
+
+my $_rol=sub { &shld(@_[0],@_) };
+my $_ror=sub { &shrd(@_[0],@_) };
+
+$code.=<<___;
+.type	sha1_block_data_order_avx,\@function,3
+.align	16
+sha1_block_data_order_avx:
+_avx_shortcut:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	lea	`-64-($win64?5*16:0)`(%rsp),%rsp
+___
+$code.=<<___ if ($win64);
+	movaps	%xmm6,64+0(%rsp)
+	movaps	%xmm7,64+16(%rsp)
+	movaps	%xmm8,64+32(%rsp)
+	movaps	%xmm9,64+48(%rsp)
+	movaps	%xmm10,64+64(%rsp)
+.Lprologue_avx:
+___
+$code.=<<___;
+	mov	%rdi,$ctx	# reassigned argument
+	mov	%rsi,$inp	# reassigned argument
+	mov	%rdx,$num	# reassigned argument
+	vzeroall
+
+	shl	\$6,$num
+	add	$inp,$num
+	lea	K_XX_XX(%rip),$K_XX_XX
+
+	mov	0($ctx),$A		# load context
+	mov	4($ctx),$B
+	mov	8($ctx),$C
+	mov	12($ctx),$D
+	mov	$B,@T[0]		# magic seed
+	mov	16($ctx),$E
+
+	vmovdqa	64($K_XX_XX),@X[2]	# pbswap mask
+	vmovdqa	0($K_XX_XX),@Tx[1]	# K_00_19
+	vmovdqu	0($inp),@X[-4&7]	# load input to %xmm[0-3]
+	vmovdqu	16($inp),@X[-3&7]
+	vmovdqu	32($inp),@X[-2&7]
+	vmovdqu	48($inp),@X[-1&7]
+	vpshufb	@X[2],@X[-4&7],@X[-4&7]	# byte swap
+	add	\$64,$inp
+	vpshufb	@X[2],@X[-3&7],@X[-3&7]
+	vpshufb	@X[2],@X[-2&7],@X[-2&7]
+	vpshufb	@X[2],@X[-1&7],@X[-1&7]
+	vpaddd	@Tx[1],@X[-4&7],@X[0]	# add K_00_19
+	vpaddd	@Tx[1],@X[-3&7],@X[1]
+	vpaddd	@Tx[1],@X[-2&7],@X[2]
+	vmovdqa	@X[0],0(%rsp)		# X[]+K xfer to IALU
+	vmovdqa	@X[1],16(%rsp)
+	vmovdqa	@X[2],32(%rsp)
+	jmp	.Loop_avx
+___
+
+sub Xupdate_avx_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpalignr(@X[0],@X[-3&7],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &vpaddd	(@Tx[1],@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpsrldq(@Tx[0],@X[-1&7],4);	# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpxor	(@X[0],@X[0],@X[-4&7]);		# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@Tx[0],@Tx[0],@X[-2&7]);	# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@Tx[0]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@Tx[0],@X[0],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslldq(@Tx[2],@X[0],12);		# "X[0]"<<96, extract one dword
+	&vpaddd	(@X[0],@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@Tx[1],@Tx[2],30);
+	&vpor	(@X[0],@X[0],@Tx[0]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslld	(@Tx[2],@Tx[2],2);
+	&vpxor	(@X[0],@X[0],@Tx[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@Tx[2]);		# "X[0]"^=("X[0]">>96)<<<2
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(@Tx[2],eval(16*(($Xi)/5))."($K_XX_XX)");	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xupdate_avx_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&vpalignr(@Tx[0],@X[-1&7],@X[-2&7],8);	# compose "X[-6]"
+	&vpxor	(@X[0],@X[0],@X[-4&7]);		# "X[0]"="X[-32]"^"X[-16]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpxor	(@X[0],@X[0],@X[-7&7]);		# "X[0]"^="X[-28]"
+	 eval(shift(@insns));
+	 eval(shift(@insns))	if (@insns[0] !~ /&ro[rl]/);
+	if ($Xi%5) {
+	  &vmovdqa	(@Tx[2],@Tx[1]);# "perpetuate" K_XX_XX...
+	} else {			# ... or load next one
+	  &vmovdqa	(@Tx[2],eval(16*($Xi/5))."($K_XX_XX)");
+	}
+	  &vpaddd	(@Tx[1],@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@Tx[0]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpsrld	(@Tx[0],@X[0],30);
+	  &vmovdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpslld	(@X[0],@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpor	(@X[0],@X[0],@Tx[0]);		# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &vmovdqa	(@Tx[1],@X[0])	if ($Xi<19);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xuplast_avx_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &vpaddd	(@Tx[1],@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&cmp	($inp,$num);
+	&je	(".Ldone_avx");
+
+	unshift(@Tx,pop(@Tx));
+
+	&vmovdqa(@X[2],"64($K_XX_XX)");		# pbswap mask
+	&vmovdqa(@Tx[1],"0($K_XX_XX)");		# K_00_19
+	&vmovdqu(@X[-4&7],"0($inp)");		# load input
+	&vmovdqu(@X[-3&7],"16($inp)");
+	&vmovdqu(@X[-2&7],"32($inp)");
+	&vmovdqu(@X[-1&7],"48($inp)");
+	&vpshufb(@X[-4&7],@X[-4&7],@X[2]);	# byte swap
+	&add	($inp,64);
+
+  $Xi=0;
+}
+
+sub Xloop_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpshufb(@X[($Xi-3)&7],@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpaddd	(@X[$Xi&7],@X[($Xi-4)&7],@Tx[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vmovdqa(eval(16*$Xi)."(%rsp)",@X[$Xi&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+$code.=<<___;
+.align	16
+.Loop_avx:
 ___
-&EPILOGUE("sha1_block_data_order");
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xuplast_avx_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+
+$code.=<<___;
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	add	12($ctx),$D
+	mov	$A,0($ctx)
+	add	16($ctx),$E
+	mov	@T[0],4($ctx)
+	mov	@T[0],$B			# magic seed
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+	jmp	.Loop_avx
+
+.align	16
+.Ldone_avx:
+___
+				$j=$saved_j; @V=@saved_V;
+
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+
+$code.=<<___;
+	vzeroall
+
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	mov	$A,0($ctx)
+	add	12($ctx),$D
+	mov	@T[0],4($ctx)
+	add	16($ctx),$E
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+___
+$code.=<<___ if ($win64);
+	movaps	64+0(%rsp),%xmm6
+	movaps	64+16(%rsp),%xmm7
+	movaps	64+32(%rsp),%xmm8
+	movaps	64+48(%rsp),%xmm9
+	movaps	64+64(%rsp),%xmm10
+___
+$code.=<<___;
+	lea	`64+($win64?5*16:0)`(%rsp),%rsi
+	mov	0(%rsi),%r12
+	mov	8(%rsi),%rbp
+	mov	16(%rsi),%rbx
+	lea	24(%rsi),%rsp
+.Lepilogue_avx:
+	ret
+.size	sha1_block_data_order_avx,.-sha1_block_data_order_avx
+___
+}
+$code.=<<___;
+.align	64
+K_XX_XX:
+.long	0x5a827999,0x5a827999,0x5a827999,0x5a827999	# K_00_19
+.long	0x6ed9eba1,0x6ed9eba1,0x6ed9eba1,0x6ed9eba1	# K_20_39
+.long	0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc	# K_40_59
+.long	0xca62c1d6,0xca62c1d6,0xca62c1d6,0xca62c1d6	# K_60_79
+.long	0x00010203,0x04050607,0x08090a0b,0x0c0d0e0f	# pbswap mask
+___
+}}}
 $code.=<<___;
 .asciz	"SHA1 block transform for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+.align	64
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lprologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lprologue
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lepilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lepilogue
+	jae	.Lcommon_seh_tail
+
+	mov	`16*4`(%rax),%rax	# pull saved stack pointer
+	lea	32(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+
+	jmp	.Lcommon_seh_tail
+.size	se_handler,.-se_handler
+
+.type	ssse3_handler,\@abi-omnipotent
+.align	16
+ssse3_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lcommon_seh_tail
+
+	lea	64(%rax),%rsi
+	lea	512($context),%rdi	# &context.Xmm6
+	mov	\$10,%ecx
+	.long	0xa548f3fc		# cld; rep movsq
+	lea	`24+64+5*16`(%rax),%rax	# adjust stack pointer
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore cotnext->R12
+
+.Lcommon_seh_tail:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	ssse3_handler,.-ssse3_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_sha1_block_data_order
+	.rva	.LSEH_end_sha1_block_data_order
+	.rva	.LSEH_info_sha1_block_data_order
+	.rva	.LSEH_begin_sha1_block_data_order_ssse3
+	.rva	.LSEH_end_sha1_block_data_order_ssse3
+	.rva	.LSEH_info_sha1_block_data_order_ssse3
+___
+$code.=<<___ if ($avx);
+	.rva	.LSEH_begin_sha1_block_data_order_avx
+	.rva	.LSEH_end_sha1_block_data_order_avx
+	.rva	.LSEH_info_sha1_block_data_order_avx
 ___
+$code.=<<___;
+.section	.xdata
+.align	8
+.LSEH_info_sha1_block_data_order:
+	.byte	9,0,0,0
+	.rva	se_handler
+.LSEH_info_sha1_block_data_order_ssse3:
+	.byte	9,0,0,0
+	.rva	ssse3_handler
+	.rva	.Lprologue_ssse3,.Lepilogue_ssse3	# HandlerData[]
+___
+$code.=<<___ if ($avx);
+.LSEH_info_sha1_block_data_order_avx:
+	.byte	9,0,0,0
+	.rva	ssse3_handler
+	.rva	.Lprologue_avx,.Lepilogue_avx		# HandlerData[]
+___
+}
 
 ####################################################################
 
diff --git a/crypto/sha/asm/sha256-586.pl b/crypto/sha/asm/sha256-586.pl
new file mode 100755
index 000000000000..928ec53123bf
--- /dev/null
+++ b/crypto/sha/asm/sha256-586.pl
@@ -0,0 +1,249 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# SHA256 block transform for x86. September 2007.
+#
+# Performance in clock cycles per processed byte (less is better):
+#
+#		Pentium	PIII	P4	AMD K8	Core2
+# gcc		46	36	41	27	26
+# icc		57	33	38	25	23	
+# x86 asm	40	30	33	20	18
+# x86_64 asm(*)	-	-	21	16	16
+#
+# (*) x86_64 assembler performance is presented for reference
+#     purposes.
+#
+# Performance improvement over compiler generated code varies from
+# 10% to 40% [see above]. Not very impressive on some µ-archs, but
+# it's 5 times smaller and optimizies amount of writes.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"sha512-586.pl",$ARGV[$#ARGV] eq "386");
+
+$A="eax";
+$E="edx";
+$T="ebx";
+$Aoff=&DWP(0,"esp");
+$Boff=&DWP(4,"esp");
+$Coff=&DWP(8,"esp");
+$Doff=&DWP(12,"esp");
+$Eoff=&DWP(16,"esp");
+$Foff=&DWP(20,"esp");
+$Goff=&DWP(24,"esp");
+$Hoff=&DWP(28,"esp");
+$Xoff=&DWP(32,"esp");
+$K256="ebp";
+
+sub BODY_00_15() {
+    my $in_16_63=shift;
+
+	&mov	("ecx",$E);
+	 &add	($T,"edi")			if ($in_16_63);	# T += sigma1(X[-2])
+	&ror	("ecx",25-11);
+	 &mov	("esi",$Foff);
+	&xor	("ecx",$E);
+	&ror	("ecx",11-6);
+	 &mov	(&DWP(4*(8+15),"esp"),$T)	if ($in_16_63);	# save X[0]
+	&xor	("ecx",$E);
+	&ror	("ecx",6);	# Sigma1(e)
+	 &mov	("edi",$Goff);
+	&add	($T,"ecx");	# T += Sigma1(e)
+
+	&xor	("esi","edi");
+	 &mov	($Eoff,$E);	# modulo-scheduled
+	 &mov	("ecx",$A);
+	&and	("esi",$E);
+	 &mov	($E,$Doff);	# e becomes d, which is e in next iteration
+	&xor	("esi","edi");	# Ch(e,f,g)
+	 &mov	("edi",$A);
+	&add	($T,"esi");	# T += Ch(e,f,g)
+
+	&ror	("ecx",22-13);
+	 &add	($T,$Hoff);	# T += h
+	&xor	("ecx",$A);
+	&ror	("ecx",13-2);
+	 &mov	("esi",$Boff);
+	&xor	("ecx",$A);
+	&ror	("ecx",2);	# Sigma0(a)
+	 &add	($E,$T);	# d += T
+	 &mov	("edi",$Coff);
+
+	&add	($T,"ecx");	# T += Sigma0(a)
+	 &mov	($Aoff,$A);	# modulo-scheduled
+
+	&mov	("ecx",$A);
+	 &sub	("esp",4);
+	&or	($A,"esi");	# a becomes h, which is a in next iteration
+	&and	("ecx","esi");
+	&and	($A,"edi");
+	 &mov	("esi",&DWP(0,$K256));
+	&or	($A,"ecx");	# h=Maj(a,b,c)
+
+	&add	($K256,4);
+	&add	($A,$T);	# h += T
+	 &mov	($T,&DWP(4*(8+15+16-1),"esp"))	if ($in_16_63);	# preload T
+	&add	($E,"esi");	# d += K256[i]
+	&add	($A,"esi");	# h += K256[i]
+}
+
+&function_begin("sha256_block_data_order");
+	&mov	("esi",wparam(0));	# ctx
+	&mov	("edi",wparam(1));	# inp
+	&mov	("eax",wparam(2));	# num
+	&mov	("ebx","esp");		# saved sp
+
+	&call	(&label("pic_point"));	# make it PIC!
+&set_label("pic_point");
+	&blindpop($K256);
+	&lea	($K256,&DWP(&label("K256")."-".&label("pic_point"),$K256));
+
+	&sub	("esp",16);
+	&and	("esp",-64);
+
+	&shl	("eax",6);
+	&add	("eax","edi");
+	&mov	(&DWP(0,"esp"),"esi");	# ctx
+	&mov	(&DWP(4,"esp"),"edi");	# inp
+	&mov	(&DWP(8,"esp"),"eax");	# inp+num*128
+	&mov	(&DWP(12,"esp"),"ebx");	# saved sp
+
+&set_label("loop",16);
+    # copy input block to stack reversing byte and dword order
+    for($i=0;$i<4;$i++) {
+	&mov	("eax",&DWP($i*16+0,"edi"));
+	&mov	("ebx",&DWP($i*16+4,"edi"));
+	&mov	("ecx",&DWP($i*16+8,"edi"));
+	&mov	("edx",&DWP($i*16+12,"edi"));
+	&bswap	("eax");
+	&bswap	("ebx");
+	&bswap	("ecx");
+	&bswap	("edx");
+	&push	("eax");
+	&push	("ebx");
+	&push	("ecx");
+	&push	("edx");
+    }
+	&add	("edi",64);
+	&sub	("esp",4*8);		# place for A,B,C,D,E,F,G,H
+	&mov	(&DWP(4*(8+16)+4,"esp"),"edi");
+
+	# copy ctx->h[0-7] to A,B,C,D,E,F,G,H on stack
+	&mov	($A,&DWP(0,"esi"));
+	&mov	("ebx",&DWP(4,"esi"));
+	&mov	("ecx",&DWP(8,"esi"));
+	&mov	("edi",&DWP(12,"esi"));
+	# &mov	($Aoff,$A);
+	&mov	($Boff,"ebx");
+	&mov	($Coff,"ecx");
+	&mov	($Doff,"edi");
+	&mov	($E,&DWP(16,"esi"));	
+	&mov	("ebx",&DWP(20,"esi"));
+	&mov	("ecx",&DWP(24,"esi"));
+	&mov	("edi",&DWP(28,"esi"));
+	# &mov	($Eoff,$E);
+	&mov	($Foff,"ebx");
+	&mov	($Goff,"ecx");
+	&mov	($Hoff,"edi");
+
+&set_label("00_15",16);
+	&mov	($T,&DWP(4*(8+15),"esp"));
+
+	&BODY_00_15();
+
+	&cmp	("esi",0xc19bf174);
+	&jne	(&label("00_15"));
+
+	&mov	($T,&DWP(4*(8+15+16-1),"esp"));	# preloaded in BODY_00_15(1)
+&set_label("16_63",16);
+	&mov	("esi",$T);
+	 &mov	("ecx",&DWP(4*(8+15+16-14),"esp"));
+	&ror	("esi",18-7);
+	 &mov	("edi","ecx");
+	&xor	("esi",$T);
+	&ror	("esi",7);
+	&shr	($T,3);
+
+	&ror	("edi",19-17);
+	 &xor	($T,"esi");			# T = sigma0(X[-15])
+	&xor	("edi","ecx");
+	&ror	("edi",17);
+	&shr	("ecx",10);
+	 &add	($T,&DWP(4*(8+15+16),"esp"));	# T += X[-16]
+	&xor	("edi","ecx");			# sigma1(X[-2])
+
+	 &add	($T,&DWP(4*(8+15+16-9),"esp"));	# T += X[-7]
+	# &add	($T,"edi");			# T += sigma1(X[-2])
+	# &mov	(&DWP(4*(8+15),"esp"),$T);	# save X[0]
+
+	&BODY_00_15(1);
+
+	&cmp	("esi",0xc67178f2);
+	&jne	(&label("16_63"));
+
+	&mov	("esi",&DWP(4*(8+16+64)+0,"esp"));#ctx
+	# &mov	($A,$Aoff);
+	&mov	("ebx",$Boff);
+	&mov	("ecx",$Coff);
+	&mov	("edi",$Doff);
+	&add	($A,&DWP(0,"esi"));
+	&add	("ebx",&DWP(4,"esi"));
+	&add	("ecx",&DWP(8,"esi"));
+	&add	("edi",&DWP(12,"esi"));
+	&mov	(&DWP(0,"esi"),$A);
+	&mov	(&DWP(4,"esi"),"ebx");
+	&mov	(&DWP(8,"esi"),"ecx");
+	&mov	(&DWP(12,"esi"),"edi");
+	# &mov	($E,$Eoff);
+	&mov	("eax",$Foff);
+	&mov	("ebx",$Goff);
+	&mov	("ecx",$Hoff);
+	&mov	("edi",&DWP(4*(8+16+64)+4,"esp"));#inp
+	&add	($E,&DWP(16,"esi"));
+	&add	("eax",&DWP(20,"esi"));
+	&add	("ebx",&DWP(24,"esi"));
+	&add	("ecx",&DWP(28,"esi"));
+	&mov	(&DWP(16,"esi"),$E);
+	&mov	(&DWP(20,"esi"),"eax");
+	&mov	(&DWP(24,"esi"),"ebx");
+	&mov	(&DWP(28,"esi"),"ecx");
+
+	&add	("esp",4*(8+16+64));		# destroy frame
+	&sub	($K256,4*64);			# rewind K
+
+	&cmp	("edi",&DWP(8,"esp"));		# are we done yet?
+	&jb	(&label("loop"));
+
+	&mov	("esp",&DWP(12,"esp"));		# restore sp
+&function_end_A();
+
+&set_label("K256",64);	# Yes! I keep it in the code segment!
+	&data_word(0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5);
+	&data_word(0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5);
+	&data_word(0xd807aa98,0x12835b01,0x243185be,0x550c7dc3);
+	&data_word(0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174);
+	&data_word(0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc);
+	&data_word(0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da);
+	&data_word(0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7);
+	&data_word(0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967);
+	&data_word(0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13);
+	&data_word(0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85);
+	&data_word(0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3);
+	&data_word(0xd192e819,0xd6990624,0xf40e3585,0x106aa070);
+	&data_word(0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5);
+	&data_word(0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3);
+	&data_word(0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208);
+	&data_word(0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2);
+&function_end_B("sha256_block_data_order");
+&asciz("SHA256 block transform for x86, CRYPTOGAMS by <appro\@openssl.org>");
+
+&asm_finish();
diff --git a/crypto/sha/asm/sha256-armv4.pl b/crypto/sha/asm/sha256-armv4.pl
new file mode 100755
index 000000000000..9c84e8d93c30
--- /dev/null
+++ b/crypto/sha/asm/sha256-armv4.pl
@@ -0,0 +1,211 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA256 block procedure for ARMv4. May 2007.
+
+# Performance is ~2x better than gcc 3.4 generated code and in "abso-
+# lute" terms is ~2250 cycles per 64-byte block or ~35 cycles per
+# byte [on single-issue Xscale PXA250 core].
+
+# July 2010.
+#
+# Rescheduling for dual-issue pipeline resulted in 22% improvement on
+# Cortex A8 core and ~20 cycles per processed byte.
+
+# February 2011.
+#
+# Profiler-assisted and platform-specific optimization resulted in 16%
+# improvement on Cortex A8 core and ~17 cycles per processed byte.
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$ctx="r0";	$t0="r0";
+$inp="r1";	$t3="r1";
+$len="r2";	$t1="r2";
+$T1="r3";
+$A="r4";
+$B="r5";
+$C="r6";
+$D="r7";
+$E="r8";
+$F="r9";
+$G="r10";
+$H="r11";
+@V=($A,$B,$C,$D,$E,$F,$G,$H);
+$t2="r12";
+$Ktbl="r14";
+
+@Sigma0=( 2,13,22);
+@Sigma1=( 6,11,25);
+@sigma0=( 7,18, 3);
+@sigma1=(17,19,10);
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___ if ($i<16);
+#if __ARM_ARCH__>=7
+	ldr	$T1,[$inp],#4
+#else
+	ldrb	$T1,[$inp,#3]			@ $i
+	ldrb	$t2,[$inp,#2]
+	ldrb	$t1,[$inp,#1]
+	ldrb	$t0,[$inp],#4
+	orr	$T1,$T1,$t2,lsl#8
+	orr	$T1,$T1,$t1,lsl#16
+	orr	$T1,$T1,$t0,lsl#24
+#endif
+___
+$code.=<<___;
+	mov	$t0,$e,ror#$Sigma1[0]
+	ldr	$t2,[$Ktbl],#4			@ *K256++
+	eor	$t0,$t0,$e,ror#$Sigma1[1]
+	eor	$t1,$f,$g
+#if $i>=16
+	add	$T1,$T1,$t3			@ from BODY_16_xx
+#elif __ARM_ARCH__>=7 && defined(__ARMEL__)
+	rev	$T1,$T1
+#endif
+#if $i==15
+	str	$inp,[sp,#17*4]			@ leave room for $t3
+#endif
+	eor	$t0,$t0,$e,ror#$Sigma1[2]	@ Sigma1(e)
+	and	$t1,$t1,$e
+	str	$T1,[sp,#`$i%16`*4]
+	add	$T1,$T1,$t0
+	eor	$t1,$t1,$g			@ Ch(e,f,g)
+	add	$T1,$T1,$h
+	mov	$h,$a,ror#$Sigma0[0]
+	add	$T1,$T1,$t1
+	eor	$h,$h,$a,ror#$Sigma0[1]
+	add	$T1,$T1,$t2
+	eor	$h,$h,$a,ror#$Sigma0[2]		@ Sigma0(a)
+#if $i>=15
+	ldr	$t3,[sp,#`($i+2)%16`*4]		@ from BODY_16_xx
+#endif
+	orr	$t0,$a,$b
+	and	$t1,$a,$b
+	and	$t0,$t0,$c
+	add	$h,$h,$T1
+	orr	$t0,$t0,$t1			@ Maj(a,b,c)
+	add	$d,$d,$T1
+	add	$h,$h,$t0
+___
+}
+
+sub BODY_16_XX {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	@ ldr	$t3,[sp,#`($i+1)%16`*4]		@ $i
+	ldr	$t2,[sp,#`($i+14)%16`*4]
+	mov	$t0,$t3,ror#$sigma0[0]
+	ldr	$T1,[sp,#`($i+0)%16`*4]
+	eor	$t0,$t0,$t3,ror#$sigma0[1]
+	ldr	$t1,[sp,#`($i+9)%16`*4]
+	eor	$t0,$t0,$t3,lsr#$sigma0[2]	@ sigma0(X[i+1])
+	mov	$t3,$t2,ror#$sigma1[0]
+	add	$T1,$T1,$t0
+	eor	$t3,$t3,$t2,ror#$sigma1[1]
+	add	$T1,$T1,$t1
+	eor	$t3,$t3,$t2,lsr#$sigma1[2]	@ sigma1(X[i+14])
+	@ add	$T1,$T1,$t3
+___
+	&BODY_00_15(@_);
+}
+
+$code=<<___;
+#include "arm_arch.h"
+
+.text
+.code	32
+
+.type	K256,%object
+.align	5
+K256:
+.word	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+.word	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+.word	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+.word	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+.word	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+.word	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+.word	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+.word	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+.word	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+.word	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+.word	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+.word	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+.word	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+.word	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+.word	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+.word	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+.size	K256,.-K256
+
+.global	sha256_block_data_order
+.type	sha256_block_data_order,%function
+sha256_block_data_order:
+	sub	r3,pc,#8		@ sha256_block_data_order
+	add	$len,$inp,$len,lsl#6	@ len to point at the end of inp
+	stmdb	sp!,{$ctx,$inp,$len,r4-r11,lr}
+	ldmia	$ctx,{$A,$B,$C,$D,$E,$F,$G,$H}
+	sub	$Ktbl,r3,#256		@ K256
+	sub	sp,sp,#16*4		@ alloca(X[16])
+.Loop:
+___
+for($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=".Lrounds_16_xx:\n";
+for (;$i<32;$i++)	{ &BODY_16_XX($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	and	$t2,$t2,#0xff
+	cmp	$t2,#0xf2
+	bne	.Lrounds_16_xx
+
+	ldr	$T1,[sp,#16*4]		@ pull ctx
+	ldr	$t0,[$T1,#0]
+	ldr	$t1,[$T1,#4]
+	ldr	$t2,[$T1,#8]
+	add	$A,$A,$t0
+	ldr	$t0,[$T1,#12]
+	add	$B,$B,$t1
+	ldr	$t1,[$T1,#16]
+	add	$C,$C,$t2
+	ldr	$t2,[$T1,#20]
+	add	$D,$D,$t0
+	ldr	$t0,[$T1,#24]
+	add	$E,$E,$t1
+	ldr	$t1,[$T1,#28]
+	add	$F,$F,$t2
+	ldr	$inp,[sp,#17*4]		@ pull inp
+	ldr	$t2,[sp,#18*4]		@ pull inp+len
+	add	$G,$G,$t0
+	add	$H,$H,$t1
+	stmia	$T1,{$A,$B,$C,$D,$E,$F,$G,$H}
+	cmp	$inp,$t2
+	sub	$Ktbl,$Ktbl,#256	@ rewind Ktbl
+	bne	.Loop
+
+	add	sp,sp,#`16+3`*4	@ destroy frame
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r11,pc}
+#else
+	ldmia	sp!,{r4-r11,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.size   sha256_block_data_order,.-sha256_block_data_order
+.asciz  "SHA256 block transform for ARMv4, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
+print $code;
+close STDOUT; # enforce flush
diff --git a/crypto/sha/asm/sha512-586.pl b/crypto/sha/asm/sha512-586.pl
new file mode 100755
index 000000000000..5b9f3337add3
--- /dev/null
+++ b/crypto/sha/asm/sha512-586.pl
@@ -0,0 +1,644 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# SHA512 block transform for x86. September 2007.
+#
+# Performance in clock cycles per processed byte (less is better):
+#
+#		Pentium	PIII	P4	AMD K8	Core2
+# gcc		100	75	116	54	66
+# icc		97	77	95	55	57
+# x86 asm	61	56	82	36	40
+# SSE2 asm	-	-	38	24	20
+# x86_64 asm(*)	-	-	30	10.0	10.5
+#
+# (*) x86_64 assembler performance is presented for reference
+#     purposes.
+#
+# IALU code-path is optimized for elder Pentiums. On vanilla Pentium
+# performance improvement over compiler generated code reaches ~60%,
+# while on PIII - ~35%. On newer µ-archs improvement varies from 15%
+# to 50%, but it's less important as they are expected to execute SSE2
+# code-path, which is commonly ~2-3x faster [than compiler generated
+# code]. SSE2 code-path is as fast as original sha512-sse2.pl, even
+# though it does not use 128-bit operations. The latter means that
+# SSE2-aware kernel is no longer required to execute the code. Another
+# difference is that new code optimizes amount of writes, but at the
+# cost of increased data cache "footprint" by 1/2KB.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"sha512-586.pl",$ARGV[$#ARGV] eq "386");
+
+$sse2=0;
+for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); }
+
+&external_label("OPENSSL_ia32cap_P") if ($sse2);
+
+$Tlo=&DWP(0,"esp");	$Thi=&DWP(4,"esp");
+$Alo=&DWP(8,"esp");	$Ahi=&DWP(8+4,"esp");
+$Blo=&DWP(16,"esp");	$Bhi=&DWP(16+4,"esp");
+$Clo=&DWP(24,"esp");	$Chi=&DWP(24+4,"esp");
+$Dlo=&DWP(32,"esp");	$Dhi=&DWP(32+4,"esp");
+$Elo=&DWP(40,"esp");	$Ehi=&DWP(40+4,"esp");
+$Flo=&DWP(48,"esp");	$Fhi=&DWP(48+4,"esp");
+$Glo=&DWP(56,"esp");	$Ghi=&DWP(56+4,"esp");
+$Hlo=&DWP(64,"esp");	$Hhi=&DWP(64+4,"esp");
+$K512="ebp";
+
+$Asse2=&QWP(0,"esp");
+$Bsse2=&QWP(8,"esp");
+$Csse2=&QWP(16,"esp");
+$Dsse2=&QWP(24,"esp");
+$Esse2=&QWP(32,"esp");
+$Fsse2=&QWP(40,"esp");
+$Gsse2=&QWP(48,"esp");
+$Hsse2=&QWP(56,"esp");
+
+$A="mm0";	# B-D and
+$E="mm4";	# F-H are commonly loaded to respectively mm1-mm3 and
+		# mm5-mm7, but it's done on on-demand basis...
+
+sub BODY_00_15_sse2 {
+    my $prefetch=shift;
+
+	&movq	("mm5",$Fsse2);			# load f
+	&movq	("mm6",$Gsse2);			# load g
+	&movq	("mm7",$Hsse2);			# load h
+
+	&movq	("mm1",$E);			# %mm1 is sliding right
+	&movq	("mm2",$E);			# %mm2 is sliding left
+	&psrlq	("mm1",14);
+	&movq	($Esse2,$E);			# modulo-scheduled save e
+	&psllq	("mm2",23);
+	&movq	("mm3","mm1");			# %mm3 is T1
+	&psrlq	("mm1",4);
+	&pxor	("mm3","mm2");
+	&psllq	("mm2",23);
+	&pxor	("mm3","mm1");
+	&psrlq	("mm1",23);
+	&pxor	("mm3","mm2");
+	&psllq	("mm2",4);
+	&pxor	("mm3","mm1");
+	&paddq	("mm7",QWP(0,$K512));		# h+=K512[i]
+	&pxor	("mm3","mm2");			# T1=Sigma1_512(e)
+
+	&pxor	("mm5","mm6");			# f^=g
+	&movq	("mm1",$Bsse2);			# load b
+	&pand	("mm5",$E);			# f&=e
+	&movq	("mm2",$Csse2);			# load c
+	&pxor	("mm5","mm6");			# f^=g
+	&movq	($E,$Dsse2);			# e = load d
+	&paddq	("mm3","mm5");			# T1+=Ch(e,f,g)
+	&movq	(&QWP(0,"esp"),$A);		# modulo-scheduled save a
+	&paddq	("mm3","mm7");			# T1+=h
+
+	&movq	("mm5",$A);			# %mm5 is sliding right
+	&movq	("mm6",$A);			# %mm6 is sliding left
+	&paddq	("mm3",&QWP(8*9,"esp"));	# T1+=X[0]
+	&psrlq	("mm5",28);
+	&paddq	($E,"mm3");			# e += T1
+	&psllq	("mm6",25);
+	&movq	("mm7","mm5");			# %mm7 is T2
+	&psrlq	("mm5",6);
+	&pxor	("mm7","mm6");
+	&psllq	("mm6",5);
+	&pxor	("mm7","mm5");
+	&psrlq	("mm5",5);
+	&pxor	("mm7","mm6");
+	&psllq	("mm6",6);
+	&pxor	("mm7","mm5");
+	&sub	("esp",8);
+	&pxor	("mm7","mm6");			# T2=Sigma0_512(a)
+
+	&movq	("mm5",$A);			# %mm5=a
+	&por	($A,"mm2");			# a=a|c
+	&movq	("mm6",&QWP(8*(9+16-14),"esp"))	if ($prefetch);
+	&pand	("mm5","mm2");			# %mm5=a&c
+	&pand	($A,"mm1");			# a=(a|c)&b
+	&movq	("mm2",&QWP(8*(9+16-1),"esp"))	if ($prefetch);
+	&por	("mm5",$A);			# %mm5=(a&c)|((a|c)&b)
+	&paddq	("mm7","mm5");			# T2+=Maj(a,b,c)
+	&movq	($A,"mm3");			# a=T1
+
+	&mov	(&LB("edx"),&BP(0,$K512));
+	&paddq	($A,"mm7");			# a+=T2
+	&add	($K512,8);
+}
+
+sub BODY_00_15_x86 {
+	#define Sigma1(x)	(ROTR((x),14) ^ ROTR((x),18)  ^ ROTR((x),41))
+	#	LO		lo>>14^hi<<18 ^ lo>>18^hi<<14 ^ hi>>9^lo<<23
+	#	HI		hi>>14^lo<<18 ^ hi>>18^lo<<14 ^ lo>>9^hi<<23
+	&mov	("ecx",$Elo);
+	&mov	("edx",$Ehi);
+	&mov	("esi","ecx");
+
+	&shr	("ecx",9)	# lo>>9
+	&mov	("edi","edx");
+	&shr	("edx",9)	# hi>>9
+	&mov	("ebx","ecx");
+	&shl	("esi",14);	# lo<<14
+	&mov	("eax","edx");
+	&shl	("edi",14);	# hi<<14
+	&xor	("ebx","esi");
+
+	&shr	("ecx",14-9);	# lo>>14
+	&xor	("eax","edi");
+	&shr	("edx",14-9);	# hi>>14
+	&xor	("eax","ecx");
+	&shl	("esi",18-14);	# lo<<18
+	&xor	("ebx","edx");
+	&shl	("edi",18-14);	# hi<<18
+	&xor	("ebx","esi");
+
+	&shr	("ecx",18-14);	# lo>>18
+	&xor	("eax","edi");
+	&shr	("edx",18-14);	# hi>>18
+	&xor	("eax","ecx");
+	&shl	("esi",23-18);	# lo<<23
+	&xor	("ebx","edx");
+	&shl	("edi",23-18);	# hi<<23
+	&xor	("eax","esi");
+	&xor	("ebx","edi");			# T1 = Sigma1(e)
+
+	&mov	("ecx",$Flo);
+	&mov	("edx",$Fhi);
+	&mov	("esi",$Glo);
+	&mov	("edi",$Ghi);
+	 &add	("eax",$Hlo);
+	 &adc	("ebx",$Hhi);			# T1 += h
+	&xor	("ecx","esi");
+	&xor	("edx","edi");
+	&and	("ecx",$Elo);
+	&and	("edx",$Ehi);
+	 &add	("eax",&DWP(8*(9+15)+0,"esp"));
+	 &adc	("ebx",&DWP(8*(9+15)+4,"esp"));	# T1 += X[0]
+	&xor	("ecx","esi");
+	&xor	("edx","edi");			# Ch(e,f,g) = (f^g)&e)^g
+
+	&mov	("esi",&DWP(0,$K512));
+	&mov	("edi",&DWP(4,$K512));		# K[i]
+	&add	("eax","ecx");
+	&adc	("ebx","edx");			# T1 += Ch(e,f,g)
+	&mov	("ecx",$Dlo);
+	&mov	("edx",$Dhi);
+	&add	("eax","esi");
+	&adc	("ebx","edi");			# T1 += K[i]
+	&mov	($Tlo,"eax");
+	&mov	($Thi,"ebx");			# put T1 away
+	&add	("eax","ecx");
+	&adc	("ebx","edx");			# d += T1
+
+	#define Sigma0(x)	(ROTR((x),28) ^ ROTR((x),34) ^ ROTR((x),39))
+	#	LO		lo>>28^hi<<4  ^ hi>>2^lo<<30 ^ hi>>7^lo<<25
+	#	HI		hi>>28^lo<<4  ^ lo>>2^hi<<30 ^ lo>>7^hi<<25
+	&mov	("ecx",$Alo);
+	&mov	("edx",$Ahi);
+	&mov	($Dlo,"eax");
+	&mov	($Dhi,"ebx");
+	&mov	("esi","ecx");
+
+	&shr	("ecx",2)	# lo>>2
+	&mov	("edi","edx");
+	&shr	("edx",2)	# hi>>2
+	&mov	("ebx","ecx");
+	&shl	("esi",4);	# lo<<4
+	&mov	("eax","edx");
+	&shl	("edi",4);	# hi<<4
+	&xor	("ebx","esi");
+
+	&shr	("ecx",7-2);	# lo>>7
+	&xor	("eax","edi");
+	&shr	("edx",7-2);	# hi>>7
+	&xor	("ebx","ecx");
+	&shl	("esi",25-4);	# lo<<25
+	&xor	("eax","edx");
+	&shl	("edi",25-4);	# hi<<25
+	&xor	("eax","esi");
+
+	&shr	("ecx",28-7);	# lo>>28
+	&xor	("ebx","edi");
+	&shr	("edx",28-7);	# hi>>28
+	&xor	("eax","ecx");
+	&shl	("esi",30-25);	# lo<<30
+	&xor	("ebx","edx");
+	&shl	("edi",30-25);	# hi<<30
+	&xor	("eax","esi");
+	&xor	("ebx","edi");			# Sigma0(a)
+
+	&mov	("ecx",$Alo);
+	&mov	("edx",$Ahi);
+	&mov	("esi",$Blo);
+	&mov	("edi",$Bhi);
+	&add	("eax",$Tlo);
+	&adc	("ebx",$Thi);			# T1 = Sigma0(a)+T1
+	&or	("ecx","esi");
+	&or	("edx","edi");
+	&and	("ecx",$Clo);
+	&and	("edx",$Chi);
+	&and	("esi",$Alo);
+	&and	("edi",$Ahi);
+	&or	("ecx","esi");
+	&or	("edx","edi");			# Maj(a,b,c) = ((a|b)&c)|(a&b)
+
+	&add	("eax","ecx");
+	&adc	("ebx","edx");			# T1 += Maj(a,b,c)
+	&mov	($Tlo,"eax");
+	&mov	($Thi,"ebx");
+
+	&mov	(&LB("edx"),&BP(0,$K512));	# pre-fetch LSB of *K
+	&sub	("esp",8);
+	&lea	($K512,&DWP(8,$K512));		# K++
+}
+
+
+&function_begin("sha512_block_data_order");
+	&mov	("esi",wparam(0));	# ctx
+	&mov	("edi",wparam(1));	# inp
+	&mov	("eax",wparam(2));	# num
+	&mov	("ebx","esp");		# saved sp
+
+	&call	(&label("pic_point"));	# make it PIC!
+&set_label("pic_point");
+	&blindpop($K512);
+	&lea	($K512,&DWP(&label("K512")."-".&label("pic_point"),$K512));
+
+	&sub	("esp",16);
+	&and	("esp",-64);
+
+	&shl	("eax",7);
+	&add	("eax","edi");
+	&mov	(&DWP(0,"esp"),"esi");	# ctx
+	&mov	(&DWP(4,"esp"),"edi");	# inp
+	&mov	(&DWP(8,"esp"),"eax");	# inp+num*128
+	&mov	(&DWP(12,"esp"),"ebx");	# saved sp
+
+if ($sse2) {
+	&picmeup("edx","OPENSSL_ia32cap_P",$K512,&label("K512"));
+	&bt	(&DWP(0,"edx"),26);
+	&jnc	(&label("loop_x86"));
+
+	# load ctx->h[0-7]
+	&movq	($A,&QWP(0,"esi"));
+	&movq	("mm1",&QWP(8,"esi"));
+	&movq	("mm2",&QWP(16,"esi"));
+	&movq	("mm3",&QWP(24,"esi"));
+	&movq	($E,&QWP(32,"esi"));
+	&movq	("mm5",&QWP(40,"esi"));
+	&movq	("mm6",&QWP(48,"esi"));
+	&movq	("mm7",&QWP(56,"esi"));
+	&sub	("esp",8*10);
+
+&set_label("loop_sse2",16);
+	# &movq	($Asse2,$A);
+	&movq	($Bsse2,"mm1");
+	&movq	($Csse2,"mm2");
+	&movq	($Dsse2,"mm3");
+	# &movq	($Esse2,$E);
+	&movq	($Fsse2,"mm5");
+	&movq	($Gsse2,"mm6");
+	&movq	($Hsse2,"mm7");
+
+	&mov	("ecx",&DWP(0,"edi"));
+	&mov	("edx",&DWP(4,"edi"));
+	&add	("edi",8);
+	&bswap	("ecx");
+	&bswap	("edx");
+	&mov	(&DWP(8*9+4,"esp"),"ecx");
+	&mov	(&DWP(8*9+0,"esp"),"edx");
+
+&set_label("00_14_sse2",16);
+	&mov	("eax",&DWP(0,"edi"));
+	&mov	("ebx",&DWP(4,"edi"));
+	&add	("edi",8);
+	&bswap	("eax");
+	&bswap	("ebx");
+	&mov	(&DWP(8*8+4,"esp"),"eax");
+	&mov	(&DWP(8*8+0,"esp"),"ebx");
+
+	&BODY_00_15_sse2();
+
+	&cmp	(&LB("edx"),0x35);
+	&jne	(&label("00_14_sse2"));
+
+	&BODY_00_15_sse2(1);
+
+&set_label("16_79_sse2",16);
+	#&movq	("mm2",&QWP(8*(9+16-1),"esp"));	#prefetched in BODY_00_15 
+	#&movq	("mm6",&QWP(8*(9+16-14),"esp"));
+	&movq	("mm1","mm2");
+
+	&psrlq	("mm2",1);
+	&movq	("mm7","mm6");
+	&psrlq	("mm6",6);
+	&movq	("mm3","mm2");
+
+	&psrlq	("mm2",7-1);
+	&movq	("mm5","mm6");
+	&psrlq	("mm6",19-6);
+	&pxor	("mm3","mm2");
+
+	&psrlq	("mm2",8-7);
+	&pxor	("mm5","mm6");
+	&psrlq	("mm6",61-19);
+	&pxor	("mm3","mm2");
+
+	&movq	("mm2",&QWP(8*(9+16),"esp"));
+
+	&psllq	("mm1",56);
+	&pxor	("mm5","mm6");
+	&psllq	("mm7",3);
+	&pxor	("mm3","mm1");
+
+	&paddq	("mm2",&QWP(8*(9+16-9),"esp"));
+
+	&psllq	("mm1",63-56);
+	&pxor	("mm5","mm7");
+	&psllq	("mm7",45-3);
+	&pxor	("mm3","mm1");
+	&pxor	("mm5","mm7");
+
+	&paddq	("mm3","mm5");
+	&paddq	("mm3","mm2");
+	&movq	(&QWP(8*9,"esp"),"mm3");
+
+	&BODY_00_15_sse2(1);
+
+	&cmp	(&LB("edx"),0x17);
+	&jne	(&label("16_79_sse2"));
+
+	# &movq	($A,$Asse2);
+	&movq	("mm1",$Bsse2);
+	&movq	("mm2",$Csse2);
+	&movq	("mm3",$Dsse2);
+	# &movq	($E,$Esse2);
+	&movq	("mm5",$Fsse2);
+	&movq	("mm6",$Gsse2);
+	&movq	("mm7",$Hsse2);
+
+	&paddq	($A,&QWP(0,"esi"));
+	&paddq	("mm1",&QWP(8,"esi"));
+	&paddq	("mm2",&QWP(16,"esi"));
+	&paddq	("mm3",&QWP(24,"esi"));
+	&paddq	($E,&QWP(32,"esi"));
+	&paddq	("mm5",&QWP(40,"esi"));
+	&paddq	("mm6",&QWP(48,"esi"));
+	&paddq	("mm7",&QWP(56,"esi"));
+
+	&movq	(&QWP(0,"esi"),$A);
+	&movq	(&QWP(8,"esi"),"mm1");
+	&movq	(&QWP(16,"esi"),"mm2");
+	&movq	(&QWP(24,"esi"),"mm3");
+	&movq	(&QWP(32,"esi"),$E);
+	&movq	(&QWP(40,"esi"),"mm5");
+	&movq	(&QWP(48,"esi"),"mm6");
+	&movq	(&QWP(56,"esi"),"mm7");
+
+	&add	("esp",8*80);			# destroy frame
+	&sub	($K512,8*80);			# rewind K
+
+	&cmp	("edi",&DWP(8*10+8,"esp"));	# are we done yet?
+	&jb	(&label("loop_sse2"));
+
+	&emms	();
+	&mov	("esp",&DWP(8*10+12,"esp"));	# restore sp
+&function_end_A();
+}
+&set_label("loop_x86",16);
+    # copy input block to stack reversing byte and qword order
+    for ($i=0;$i<8;$i++) {
+	&mov	("eax",&DWP($i*16+0,"edi"));
+	&mov	("ebx",&DWP($i*16+4,"edi"));
+	&mov	("ecx",&DWP($i*16+8,"edi"));
+	&mov	("edx",&DWP($i*16+12,"edi"));
+	&bswap	("eax");
+	&bswap	("ebx");
+	&bswap	("ecx");
+	&bswap	("edx");
+	&push	("eax");
+	&push	("ebx");
+	&push	("ecx");
+	&push	("edx");
+    }
+	&add	("edi",128);
+	&sub	("esp",9*8);		# place for T,A,B,C,D,E,F,G,H
+	&mov	(&DWP(8*(9+16)+4,"esp"),"edi");
+
+	# copy ctx->h[0-7] to A,B,C,D,E,F,G,H on stack
+	&lea	("edi",&DWP(8,"esp"));
+	&mov	("ecx",16);
+	&data_word(0xA5F3F689);		# rep movsd
+
+&set_label("00_15_x86",16);
+	&BODY_00_15_x86();
+
+	&cmp	(&LB("edx"),0x94);
+	&jne	(&label("00_15_x86"));
+
+&set_label("16_79_x86",16);
+	#define sigma0(x)	(ROTR((x),1)  ^ ROTR((x),8)  ^ ((x)>>7))
+	#	LO		lo>>1^hi<<31  ^ lo>>8^hi<<24 ^ lo>>7^hi<<25
+	#	HI		hi>>1^lo<<31  ^ hi>>8^lo<<24 ^ hi>>7
+	&mov	("ecx",&DWP(8*(9+15+16-1)+0,"esp"));
+	&mov	("edx",&DWP(8*(9+15+16-1)+4,"esp"));
+	&mov	("esi","ecx");
+
+	&shr	("ecx",1)	# lo>>1
+	&mov	("edi","edx");
+	&shr	("edx",1)	# hi>>1
+	&mov	("eax","ecx");
+	&shl	("esi",24);	# lo<<24
+	&mov	("ebx","edx");
+	&shl	("edi",24);	# hi<<24
+	&xor	("ebx","esi");
+
+	&shr	("ecx",7-1);	# lo>>7
+	&xor	("eax","edi");
+	&shr	("edx",7-1);	# hi>>7
+	&xor	("eax","ecx");
+	&shl	("esi",31-24);	# lo<<31
+	&xor	("ebx","edx");
+	&shl	("edi",25-24);	# hi<<25
+	&xor	("ebx","esi");
+
+	&shr	("ecx",8-7);	# lo>>8
+	&xor	("eax","edi");
+	&shr	("edx",8-7);	# hi>>8
+	&xor	("eax","ecx");
+	&shl	("edi",31-25);	# hi<<31
+	&xor	("ebx","edx");
+	&xor	("eax","edi");			# T1 = sigma0(X[-15])
+
+	&mov	(&DWP(0,"esp"),"eax");
+	&mov	(&DWP(4,"esp"),"ebx");		# put T1 away
+
+	#define sigma1(x)	(ROTR((x),19) ^ ROTR((x),61) ^ ((x)>>6))
+	#	LO		lo>>19^hi<<13 ^ hi>>29^lo<<3 ^ lo>>6^hi<<26
+	#	HI		hi>>19^lo<<13 ^ lo>>29^hi<<3 ^ hi>>6
+	&mov	("ecx",&DWP(8*(9+15+16-14)+0,"esp"));
+	&mov	("edx",&DWP(8*(9+15+16-14)+4,"esp"));
+	&mov	("esi","ecx");
+
+	&shr	("ecx",6)	# lo>>6
+	&mov	("edi","edx");
+	&shr	("edx",6)	# hi>>6
+	&mov	("eax","ecx");
+	&shl	("esi",3);	# lo<<3
+	&mov	("ebx","edx");
+	&shl	("edi",3);	# hi<<3
+	&xor	("eax","esi");
+
+	&shr	("ecx",19-6);	# lo>>19
+	&xor	("ebx","edi");
+	&shr	("edx",19-6);	# hi>>19
+	&xor	("eax","ecx");
+	&shl	("esi",13-3);	# lo<<13
+	&xor	("ebx","edx");
+	&shl	("edi",13-3);	# hi<<13
+	&xor	("ebx","esi");
+
+	&shr	("ecx",29-19);	# lo>>29
+	&xor	("eax","edi");
+	&shr	("edx",29-19);	# hi>>29
+	&xor	("ebx","ecx");
+	&shl	("edi",26-13);	# hi<<26
+	&xor	("eax","edx");
+	&xor	("eax","edi");			# sigma1(X[-2])
+
+	&mov	("ecx",&DWP(8*(9+15+16)+0,"esp"));
+	&mov	("edx",&DWP(8*(9+15+16)+4,"esp"));
+	&add	("eax",&DWP(0,"esp"));
+	&adc	("ebx",&DWP(4,"esp"));		# T1 = sigma1(X[-2])+T1
+	&mov	("esi",&DWP(8*(9+15+16-9)+0,"esp"));
+	&mov	("edi",&DWP(8*(9+15+16-9)+4,"esp"));
+	&add	("eax","ecx");
+	&adc	("ebx","edx");			# T1 += X[-16]
+	&add	("eax","esi");
+	&adc	("ebx","edi");			# T1 += X[-7]
+	&mov	(&DWP(8*(9+15)+0,"esp"),"eax");
+	&mov	(&DWP(8*(9+15)+4,"esp"),"ebx");	# save X[0]
+
+	&BODY_00_15_x86();
+
+	&cmp	(&LB("edx"),0x17);
+	&jne	(&label("16_79_x86"));
+
+	&mov	("esi",&DWP(8*(9+16+80)+0,"esp"));# ctx
+	&mov	("edi",&DWP(8*(9+16+80)+4,"esp"));# inp
+    for($i=0;$i<4;$i++) {
+	&mov	("eax",&DWP($i*16+0,"esi"));
+	&mov	("ebx",&DWP($i*16+4,"esi"));
+	&mov	("ecx",&DWP($i*16+8,"esi"));
+	&mov	("edx",&DWP($i*16+12,"esi"));
+	&add	("eax",&DWP(8+($i*16)+0,"esp"));
+	&adc	("ebx",&DWP(8+($i*16)+4,"esp"));
+	&mov	(&DWP($i*16+0,"esi"),"eax");
+	&mov	(&DWP($i*16+4,"esi"),"ebx");
+	&add	("ecx",&DWP(8+($i*16)+8,"esp"));
+	&adc	("edx",&DWP(8+($i*16)+12,"esp"));
+	&mov	(&DWP($i*16+8,"esi"),"ecx");
+	&mov	(&DWP($i*16+12,"esi"),"edx");
+    }
+	&add	("esp",8*(9+16+80));		# destroy frame
+	&sub	($K512,8*80);			# rewind K
+
+	&cmp	("edi",&DWP(8,"esp"));		# are we done yet?
+	&jb	(&label("loop_x86"));
+
+	&mov	("esp",&DWP(12,"esp"));		# restore sp
+&function_end_A();
+
+&set_label("K512",64);	# Yes! I keep it in the code segment!
+	&data_word(0xd728ae22,0x428a2f98);	# u64
+	&data_word(0x23ef65cd,0x71374491);	# u64
+	&data_word(0xec4d3b2f,0xb5c0fbcf);	# u64
+	&data_word(0x8189dbbc,0xe9b5dba5);	# u64
+	&data_word(0xf348b538,0x3956c25b);	# u64
+	&data_word(0xb605d019,0x59f111f1);	# u64
+	&data_word(0xaf194f9b,0x923f82a4);	# u64
+	&data_word(0xda6d8118,0xab1c5ed5);	# u64
+	&data_word(0xa3030242,0xd807aa98);	# u64
+	&data_word(0x45706fbe,0x12835b01);	# u64
+	&data_word(0x4ee4b28c,0x243185be);	# u64
+	&data_word(0xd5ffb4e2,0x550c7dc3);	# u64
+	&data_word(0xf27b896f,0x72be5d74);	# u64
+	&data_word(0x3b1696b1,0x80deb1fe);	# u64
+	&data_word(0x25c71235,0x9bdc06a7);	# u64
+	&data_word(0xcf692694,0xc19bf174);	# u64
+	&data_word(0x9ef14ad2,0xe49b69c1);	# u64
+	&data_word(0x384f25e3,0xefbe4786);	# u64
+	&data_word(0x8b8cd5b5,0x0fc19dc6);	# u64
+	&data_word(0x77ac9c65,0x240ca1cc);	# u64
+	&data_word(0x592b0275,0x2de92c6f);	# u64
+	&data_word(0x6ea6e483,0x4a7484aa);	# u64
+	&data_word(0xbd41fbd4,0x5cb0a9dc);	# u64
+	&data_word(0x831153b5,0x76f988da);	# u64
+	&data_word(0xee66dfab,0x983e5152);	# u64
+	&data_word(0x2db43210,0xa831c66d);	# u64
+	&data_word(0x98fb213f,0xb00327c8);	# u64
+	&data_word(0xbeef0ee4,0xbf597fc7);	# u64
+	&data_word(0x3da88fc2,0xc6e00bf3);	# u64
+	&data_word(0x930aa725,0xd5a79147);	# u64
+	&data_word(0xe003826f,0x06ca6351);	# u64
+	&data_word(0x0a0e6e70,0x14292967);	# u64
+	&data_word(0x46d22ffc,0x27b70a85);	# u64
+	&data_word(0x5c26c926,0x2e1b2138);	# u64
+	&data_word(0x5ac42aed,0x4d2c6dfc);	# u64
+	&data_word(0x9d95b3df,0x53380d13);	# u64
+	&data_word(0x8baf63de,0x650a7354);	# u64
+	&data_word(0x3c77b2a8,0x766a0abb);	# u64
+	&data_word(0x47edaee6,0x81c2c92e);	# u64
+	&data_word(0x1482353b,0x92722c85);	# u64
+	&data_word(0x4cf10364,0xa2bfe8a1);	# u64
+	&data_word(0xbc423001,0xa81a664b);	# u64
+	&data_word(0xd0f89791,0xc24b8b70);	# u64
+	&data_word(0x0654be30,0xc76c51a3);	# u64
+	&data_word(0xd6ef5218,0xd192e819);	# u64
+	&data_word(0x5565a910,0xd6990624);	# u64
+	&data_word(0x5771202a,0xf40e3585);	# u64
+	&data_word(0x32bbd1b8,0x106aa070);	# u64
+	&data_word(0xb8d2d0c8,0x19a4c116);	# u64
+	&data_word(0x5141ab53,0x1e376c08);	# u64
+	&data_word(0xdf8eeb99,0x2748774c);	# u64
+	&data_word(0xe19b48a8,0x34b0bcb5);	# u64
+	&data_word(0xc5c95a63,0x391c0cb3);	# u64
+	&data_word(0xe3418acb,0x4ed8aa4a);	# u64
+	&data_word(0x7763e373,0x5b9cca4f);	# u64
+	&data_word(0xd6b2b8a3,0x682e6ff3);	# u64
+	&data_word(0x5defb2fc,0x748f82ee);	# u64
+	&data_word(0x43172f60,0x78a5636f);	# u64
+	&data_word(0xa1f0ab72,0x84c87814);	# u64
+	&data_word(0x1a6439ec,0x8cc70208);	# u64
+	&data_word(0x23631e28,0x90befffa);	# u64
+	&data_word(0xde82bde9,0xa4506ceb);	# u64
+	&data_word(0xb2c67915,0xbef9a3f7);	# u64
+	&data_word(0xe372532b,0xc67178f2);	# u64
+	&data_word(0xea26619c,0xca273ece);	# u64
+	&data_word(0x21c0c207,0xd186b8c7);	# u64
+	&data_word(0xcde0eb1e,0xeada7dd6);	# u64
+	&data_word(0xee6ed178,0xf57d4f7f);	# u64
+	&data_word(0x72176fba,0x06f067aa);	# u64
+	&data_word(0xa2c898a6,0x0a637dc5);	# u64
+	&data_word(0xbef90dae,0x113f9804);	# u64
+	&data_word(0x131c471b,0x1b710b35);	# u64
+	&data_word(0x23047d84,0x28db77f5);	# u64
+	&data_word(0x40c72493,0x32caab7b);	# u64
+	&data_word(0x15c9bebc,0x3c9ebe0a);	# u64
+	&data_word(0x9c100d4c,0x431d67c4);	# u64
+	&data_word(0xcb3e42b6,0x4cc5d4be);	# u64
+	&data_word(0xfc657e2a,0x597f299c);	# u64
+	&data_word(0x3ad6faec,0x5fcb6fab);	# u64
+	&data_word(0x4a475817,0x6c44198c);	# u64
+&function_end_B("sha512_block_data_order");
+&asciz("SHA512 block transform for x86, CRYPTOGAMS by <appro\@openssl.org>");
+
+&asm_finish();
diff --git a/crypto/sha/asm/sha512-armv4.pl b/crypto/sha/asm/sha512-armv4.pl
new file mode 100755
index 000000000000..7faf37b14790
--- /dev/null
+++ b/crypto/sha/asm/sha512-armv4.pl
@@ -0,0 +1,582 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA512 block procedure for ARMv4. September 2007.
+
+# This code is ~4.5 (four and a half) times faster than code generated
+# by gcc 3.4 and it spends ~72 clock cycles per byte [on single-issue
+# Xscale PXA250 core].
+#
+# July 2010.
+#
+# Rescheduling for dual-issue pipeline resulted in 6% improvement on
+# Cortex A8 core and ~40 cycles per processed byte.
+
+# February 2011.
+#
+# Profiler-assisted and platform-specific optimization resulted in 7%
+# improvement on Coxtex A8 core and ~38 cycles per byte.
+
+# March 2011.
+#
+# Add NEON implementation. On Cortex A8 it was measured to process
+# one byte in 25.5 cycles or 47% faster than integer-only code.
+
+# Byte order [in]dependence. =========================================
+#
+# Originally caller was expected to maintain specific *dword* order in
+# h[0-7], namely with most significant dword at *lower* address, which
+# was reflected in below two parameters as 0 and 4. Now caller is
+# expected to maintain native byte order for whole 64-bit values.
+$hi="HI";
+$lo="LO";
+# ====================================================================
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$ctx="r0";	# parameter block
+$inp="r1";
+$len="r2";
+
+$Tlo="r3";
+$Thi="r4";
+$Alo="r5";
+$Ahi="r6";
+$Elo="r7";
+$Ehi="r8";
+$t0="r9";
+$t1="r10";
+$t2="r11";
+$t3="r12";
+############	r13 is stack pointer
+$Ktbl="r14";
+############	r15 is program counter
+
+$Aoff=8*0;
+$Boff=8*1;
+$Coff=8*2;
+$Doff=8*3;
+$Eoff=8*4;
+$Foff=8*5;
+$Goff=8*6;
+$Hoff=8*7;
+$Xoff=8*8;
+
+sub BODY_00_15() {
+my $magic = shift;
+$code.=<<___;
+	@ Sigma1(x)	(ROTR((x),14) ^ ROTR((x),18)  ^ ROTR((x),41))
+	@ LO		lo>>14^hi<<18 ^ lo>>18^hi<<14 ^ hi>>9^lo<<23
+	@ HI		hi>>14^lo<<18 ^ hi>>18^lo<<14 ^ lo>>9^hi<<23
+	mov	$t0,$Elo,lsr#14
+	str	$Tlo,[sp,#$Xoff+0]
+	mov	$t1,$Ehi,lsr#14
+	str	$Thi,[sp,#$Xoff+4]
+	eor	$t0,$t0,$Ehi,lsl#18
+	ldr	$t2,[sp,#$Hoff+0]	@ h.lo
+	eor	$t1,$t1,$Elo,lsl#18
+	ldr	$t3,[sp,#$Hoff+4]	@ h.hi
+	eor	$t0,$t0,$Elo,lsr#18
+	eor	$t1,$t1,$Ehi,lsr#18
+	eor	$t0,$t0,$Ehi,lsl#14
+	eor	$t1,$t1,$Elo,lsl#14
+	eor	$t0,$t0,$Ehi,lsr#9
+	eor	$t1,$t1,$Elo,lsr#9
+	eor	$t0,$t0,$Elo,lsl#23
+	eor	$t1,$t1,$Ehi,lsl#23	@ Sigma1(e)
+	adds	$Tlo,$Tlo,$t0
+	ldr	$t0,[sp,#$Foff+0]	@ f.lo
+	adc	$Thi,$Thi,$t1		@ T += Sigma1(e)
+	ldr	$t1,[sp,#$Foff+4]	@ f.hi
+	adds	$Tlo,$Tlo,$t2
+	ldr	$t2,[sp,#$Goff+0]	@ g.lo
+	adc	$Thi,$Thi,$t3		@ T += h
+	ldr	$t3,[sp,#$Goff+4]	@ g.hi
+
+	eor	$t0,$t0,$t2
+	str	$Elo,[sp,#$Eoff+0]
+	eor	$t1,$t1,$t3
+	str	$Ehi,[sp,#$Eoff+4]
+	and	$t0,$t0,$Elo
+	str	$Alo,[sp,#$Aoff+0]
+	and	$t1,$t1,$Ehi
+	str	$Ahi,[sp,#$Aoff+4]
+	eor	$t0,$t0,$t2
+	ldr	$t2,[$Ktbl,#$lo]	@ K[i].lo
+	eor	$t1,$t1,$t3		@ Ch(e,f,g)
+	ldr	$t3,[$Ktbl,#$hi]	@ K[i].hi
+
+	adds	$Tlo,$Tlo,$t0
+	ldr	$Elo,[sp,#$Doff+0]	@ d.lo
+	adc	$Thi,$Thi,$t1		@ T += Ch(e,f,g)
+	ldr	$Ehi,[sp,#$Doff+4]	@ d.hi
+	adds	$Tlo,$Tlo,$t2
+	and	$t0,$t2,#0xff
+	adc	$Thi,$Thi,$t3		@ T += K[i]
+	adds	$Elo,$Elo,$Tlo
+	ldr	$t2,[sp,#$Boff+0]	@ b.lo
+	adc	$Ehi,$Ehi,$Thi		@ d += T
+	teq	$t0,#$magic
+
+	ldr	$t3,[sp,#$Coff+0]	@ c.lo
+	orreq	$Ktbl,$Ktbl,#1
+	@ Sigma0(x)	(ROTR((x),28) ^ ROTR((x),34) ^ ROTR((x),39))
+	@ LO		lo>>28^hi<<4  ^ hi>>2^lo<<30 ^ hi>>7^lo<<25
+	@ HI		hi>>28^lo<<4  ^ lo>>2^hi<<30 ^ lo>>7^hi<<25
+	mov	$t0,$Alo,lsr#28
+	mov	$t1,$Ahi,lsr#28
+	eor	$t0,$t0,$Ahi,lsl#4
+	eor	$t1,$t1,$Alo,lsl#4
+	eor	$t0,$t0,$Ahi,lsr#2
+	eor	$t1,$t1,$Alo,lsr#2
+	eor	$t0,$t0,$Alo,lsl#30
+	eor	$t1,$t1,$Ahi,lsl#30
+	eor	$t0,$t0,$Ahi,lsr#7
+	eor	$t1,$t1,$Alo,lsr#7
+	eor	$t0,$t0,$Alo,lsl#25
+	eor	$t1,$t1,$Ahi,lsl#25	@ Sigma0(a)
+	adds	$Tlo,$Tlo,$t0
+	and	$t0,$Alo,$t2
+	adc	$Thi,$Thi,$t1		@ T += Sigma0(a)
+
+	ldr	$t1,[sp,#$Boff+4]	@ b.hi
+	orr	$Alo,$Alo,$t2
+	ldr	$t2,[sp,#$Coff+4]	@ c.hi
+	and	$Alo,$Alo,$t3
+	and	$t3,$Ahi,$t1
+	orr	$Ahi,$Ahi,$t1
+	orr	$Alo,$Alo,$t0		@ Maj(a,b,c).lo
+	and	$Ahi,$Ahi,$t2
+	adds	$Alo,$Alo,$Tlo
+	orr	$Ahi,$Ahi,$t3		@ Maj(a,b,c).hi
+	sub	sp,sp,#8
+	adc	$Ahi,$Ahi,$Thi		@ h += T
+	tst	$Ktbl,#1
+	add	$Ktbl,$Ktbl,#8
+___
+}
+$code=<<___;
+#include "arm_arch.h"
+#ifdef __ARMEL__
+# define LO 0
+# define HI 4
+# define WORD64(hi0,lo0,hi1,lo1)	.word	lo0,hi0, lo1,hi1
+#else
+# define HI 0
+# define LO 4
+# define WORD64(hi0,lo0,hi1,lo1)	.word	hi0,lo0, hi1,lo1
+#endif
+
+.text
+.code	32
+.type	K512,%object
+.align	5
+K512:
+WORD64(0x428a2f98,0xd728ae22, 0x71374491,0x23ef65cd)
+WORD64(0xb5c0fbcf,0xec4d3b2f, 0xe9b5dba5,0x8189dbbc)
+WORD64(0x3956c25b,0xf348b538, 0x59f111f1,0xb605d019)
+WORD64(0x923f82a4,0xaf194f9b, 0xab1c5ed5,0xda6d8118)
+WORD64(0xd807aa98,0xa3030242, 0x12835b01,0x45706fbe)
+WORD64(0x243185be,0x4ee4b28c, 0x550c7dc3,0xd5ffb4e2)
+WORD64(0x72be5d74,0xf27b896f, 0x80deb1fe,0x3b1696b1)
+WORD64(0x9bdc06a7,0x25c71235, 0xc19bf174,0xcf692694)
+WORD64(0xe49b69c1,0x9ef14ad2, 0xefbe4786,0x384f25e3)
+WORD64(0x0fc19dc6,0x8b8cd5b5, 0x240ca1cc,0x77ac9c65)
+WORD64(0x2de92c6f,0x592b0275, 0x4a7484aa,0x6ea6e483)
+WORD64(0x5cb0a9dc,0xbd41fbd4, 0x76f988da,0x831153b5)
+WORD64(0x983e5152,0xee66dfab, 0xa831c66d,0x2db43210)
+WORD64(0xb00327c8,0x98fb213f, 0xbf597fc7,0xbeef0ee4)
+WORD64(0xc6e00bf3,0x3da88fc2, 0xd5a79147,0x930aa725)
+WORD64(0x06ca6351,0xe003826f, 0x14292967,0x0a0e6e70)
+WORD64(0x27b70a85,0x46d22ffc, 0x2e1b2138,0x5c26c926)
+WORD64(0x4d2c6dfc,0x5ac42aed, 0x53380d13,0x9d95b3df)
+WORD64(0x650a7354,0x8baf63de, 0x766a0abb,0x3c77b2a8)
+WORD64(0x81c2c92e,0x47edaee6, 0x92722c85,0x1482353b)
+WORD64(0xa2bfe8a1,0x4cf10364, 0xa81a664b,0xbc423001)
+WORD64(0xc24b8b70,0xd0f89791, 0xc76c51a3,0x0654be30)
+WORD64(0xd192e819,0xd6ef5218, 0xd6990624,0x5565a910)
+WORD64(0xf40e3585,0x5771202a, 0x106aa070,0x32bbd1b8)
+WORD64(0x19a4c116,0xb8d2d0c8, 0x1e376c08,0x5141ab53)
+WORD64(0x2748774c,0xdf8eeb99, 0x34b0bcb5,0xe19b48a8)
+WORD64(0x391c0cb3,0xc5c95a63, 0x4ed8aa4a,0xe3418acb)
+WORD64(0x5b9cca4f,0x7763e373, 0x682e6ff3,0xd6b2b8a3)
+WORD64(0x748f82ee,0x5defb2fc, 0x78a5636f,0x43172f60)
+WORD64(0x84c87814,0xa1f0ab72, 0x8cc70208,0x1a6439ec)
+WORD64(0x90befffa,0x23631e28, 0xa4506ceb,0xde82bde9)
+WORD64(0xbef9a3f7,0xb2c67915, 0xc67178f2,0xe372532b)
+WORD64(0xca273ece,0xea26619c, 0xd186b8c7,0x21c0c207)
+WORD64(0xeada7dd6,0xcde0eb1e, 0xf57d4f7f,0xee6ed178)
+WORD64(0x06f067aa,0x72176fba, 0x0a637dc5,0xa2c898a6)
+WORD64(0x113f9804,0xbef90dae, 0x1b710b35,0x131c471b)
+WORD64(0x28db77f5,0x23047d84, 0x32caab7b,0x40c72493)
+WORD64(0x3c9ebe0a,0x15c9bebc, 0x431d67c4,0x9c100d4c)
+WORD64(0x4cc5d4be,0xcb3e42b6, 0x597f299c,0xfc657e2a)
+WORD64(0x5fcb6fab,0x3ad6faec, 0x6c44198c,0x4a475817)
+.size	K512,.-K512
+.LOPENSSL_armcap:
+.word	OPENSSL_armcap_P-sha512_block_data_order
+.skip	32-4
+
+.global	sha512_block_data_order
+.type	sha512_block_data_order,%function
+sha512_block_data_order:
+	sub	r3,pc,#8		@ sha512_block_data_order
+	add	$len,$inp,$len,lsl#7	@ len to point at the end of inp
+#if __ARM_ARCH__>=7
+	ldr	r12,.LOPENSSL_armcap
+	ldr	r12,[r3,r12]		@ OPENSSL_armcap_P
+	tst	r12,#1
+	bne	.LNEON
+#endif
+	stmdb	sp!,{r4-r12,lr}
+	sub	$Ktbl,r3,#672		@ K512
+	sub	sp,sp,#9*8
+
+	ldr	$Elo,[$ctx,#$Eoff+$lo]
+	ldr	$Ehi,[$ctx,#$Eoff+$hi]
+	ldr	$t0, [$ctx,#$Goff+$lo]
+	ldr	$t1, [$ctx,#$Goff+$hi]
+	ldr	$t2, [$ctx,#$Hoff+$lo]
+	ldr	$t3, [$ctx,#$Hoff+$hi]
+.Loop:
+	str	$t0, [sp,#$Goff+0]
+	str	$t1, [sp,#$Goff+4]
+	str	$t2, [sp,#$Hoff+0]
+	str	$t3, [sp,#$Hoff+4]
+	ldr	$Alo,[$ctx,#$Aoff+$lo]
+	ldr	$Ahi,[$ctx,#$Aoff+$hi]
+	ldr	$Tlo,[$ctx,#$Boff+$lo]
+	ldr	$Thi,[$ctx,#$Boff+$hi]
+	ldr	$t0, [$ctx,#$Coff+$lo]
+	ldr	$t1, [$ctx,#$Coff+$hi]
+	ldr	$t2, [$ctx,#$Doff+$lo]
+	ldr	$t3, [$ctx,#$Doff+$hi]
+	str	$Tlo,[sp,#$Boff+0]
+	str	$Thi,[sp,#$Boff+4]
+	str	$t0, [sp,#$Coff+0]
+	str	$t1, [sp,#$Coff+4]
+	str	$t2, [sp,#$Doff+0]
+	str	$t3, [sp,#$Doff+4]
+	ldr	$Tlo,[$ctx,#$Foff+$lo]
+	ldr	$Thi,[$ctx,#$Foff+$hi]
+	str	$Tlo,[sp,#$Foff+0]
+	str	$Thi,[sp,#$Foff+4]
+
+.L00_15:
+#if __ARM_ARCH__<7
+	ldrb	$Tlo,[$inp,#7]
+	ldrb	$t0, [$inp,#6]
+	ldrb	$t1, [$inp,#5]
+	ldrb	$t2, [$inp,#4]
+	ldrb	$Thi,[$inp,#3]
+	ldrb	$t3, [$inp,#2]
+	orr	$Tlo,$Tlo,$t0,lsl#8
+	ldrb	$t0, [$inp,#1]
+	orr	$Tlo,$Tlo,$t1,lsl#16
+	ldrb	$t1, [$inp],#8
+	orr	$Tlo,$Tlo,$t2,lsl#24
+	orr	$Thi,$Thi,$t3,lsl#8
+	orr	$Thi,$Thi,$t0,lsl#16
+	orr	$Thi,$Thi,$t1,lsl#24
+#else
+	ldr	$Tlo,[$inp,#4]
+	ldr	$Thi,[$inp],#8
+#ifdef __ARMEL__
+	rev	$Tlo,$Tlo
+	rev	$Thi,$Thi
+#endif
+#endif
+___
+	&BODY_00_15(0x94);
+$code.=<<___;
+	tst	$Ktbl,#1
+	beq	.L00_15
+	ldr	$t0,[sp,#`$Xoff+8*(16-1)`+0]
+	ldr	$t1,[sp,#`$Xoff+8*(16-1)`+4]
+	bic	$Ktbl,$Ktbl,#1
+.L16_79:
+	@ sigma0(x)	(ROTR((x),1)  ^ ROTR((x),8)  ^ ((x)>>7))
+	@ LO		lo>>1^hi<<31  ^ lo>>8^hi<<24 ^ lo>>7^hi<<25
+	@ HI		hi>>1^lo<<31  ^ hi>>8^lo<<24 ^ hi>>7
+	mov	$Tlo,$t0,lsr#1
+	ldr	$t2,[sp,#`$Xoff+8*(16-14)`+0]
+	mov	$Thi,$t1,lsr#1
+	ldr	$t3,[sp,#`$Xoff+8*(16-14)`+4]
+	eor	$Tlo,$Tlo,$t1,lsl#31
+	eor	$Thi,$Thi,$t0,lsl#31
+	eor	$Tlo,$Tlo,$t0,lsr#8
+	eor	$Thi,$Thi,$t1,lsr#8
+	eor	$Tlo,$Tlo,$t1,lsl#24
+	eor	$Thi,$Thi,$t0,lsl#24
+	eor	$Tlo,$Tlo,$t0,lsr#7
+	eor	$Thi,$Thi,$t1,lsr#7
+	eor	$Tlo,$Tlo,$t1,lsl#25
+
+	@ sigma1(x)	(ROTR((x),19) ^ ROTR((x),61) ^ ((x)>>6))
+	@ LO		lo>>19^hi<<13 ^ hi>>29^lo<<3 ^ lo>>6^hi<<26
+	@ HI		hi>>19^lo<<13 ^ lo>>29^hi<<3 ^ hi>>6
+	mov	$t0,$t2,lsr#19
+	mov	$t1,$t3,lsr#19
+	eor	$t0,$t0,$t3,lsl#13
+	eor	$t1,$t1,$t2,lsl#13
+	eor	$t0,$t0,$t3,lsr#29
+	eor	$t1,$t1,$t2,lsr#29
+	eor	$t0,$t0,$t2,lsl#3
+	eor	$t1,$t1,$t3,lsl#3
+	eor	$t0,$t0,$t2,lsr#6
+	eor	$t1,$t1,$t3,lsr#6
+	ldr	$t2,[sp,#`$Xoff+8*(16-9)`+0]
+	eor	$t0,$t0,$t3,lsl#26
+
+	ldr	$t3,[sp,#`$Xoff+8*(16-9)`+4]
+	adds	$Tlo,$Tlo,$t0
+	ldr	$t0,[sp,#`$Xoff+8*16`+0]
+	adc	$Thi,$Thi,$t1
+
+	ldr	$t1,[sp,#`$Xoff+8*16`+4]
+	adds	$Tlo,$Tlo,$t2
+	adc	$Thi,$Thi,$t3
+	adds	$Tlo,$Tlo,$t0
+	adc	$Thi,$Thi,$t1
+___
+	&BODY_00_15(0x17);
+$code.=<<___;
+	ldreq	$t0,[sp,#`$Xoff+8*(16-1)`+0]
+	ldreq	$t1,[sp,#`$Xoff+8*(16-1)`+4]
+	beq	.L16_79
+	bic	$Ktbl,$Ktbl,#1
+
+	ldr	$Tlo,[sp,#$Boff+0]
+	ldr	$Thi,[sp,#$Boff+4]
+	ldr	$t0, [$ctx,#$Aoff+$lo]
+	ldr	$t1, [$ctx,#$Aoff+$hi]
+	ldr	$t2, [$ctx,#$Boff+$lo]
+	ldr	$t3, [$ctx,#$Boff+$hi]
+	adds	$t0,$Alo,$t0
+	str	$t0, [$ctx,#$Aoff+$lo]
+	adc	$t1,$Ahi,$t1
+	str	$t1, [$ctx,#$Aoff+$hi]
+	adds	$t2,$Tlo,$t2
+	str	$t2, [$ctx,#$Boff+$lo]
+	adc	$t3,$Thi,$t3
+	str	$t3, [$ctx,#$Boff+$hi]
+
+	ldr	$Alo,[sp,#$Coff+0]
+	ldr	$Ahi,[sp,#$Coff+4]
+	ldr	$Tlo,[sp,#$Doff+0]
+	ldr	$Thi,[sp,#$Doff+4]
+	ldr	$t0, [$ctx,#$Coff+$lo]
+	ldr	$t1, [$ctx,#$Coff+$hi]
+	ldr	$t2, [$ctx,#$Doff+$lo]
+	ldr	$t3, [$ctx,#$Doff+$hi]
+	adds	$t0,$Alo,$t0
+	str	$t0, [$ctx,#$Coff+$lo]
+	adc	$t1,$Ahi,$t1
+	str	$t1, [$ctx,#$Coff+$hi]
+	adds	$t2,$Tlo,$t2
+	str	$t2, [$ctx,#$Doff+$lo]
+	adc	$t3,$Thi,$t3
+	str	$t3, [$ctx,#$Doff+$hi]
+
+	ldr	$Tlo,[sp,#$Foff+0]
+	ldr	$Thi,[sp,#$Foff+4]
+	ldr	$t0, [$ctx,#$Eoff+$lo]
+	ldr	$t1, [$ctx,#$Eoff+$hi]
+	ldr	$t2, [$ctx,#$Foff+$lo]
+	ldr	$t3, [$ctx,#$Foff+$hi]
+	adds	$Elo,$Elo,$t0
+	str	$Elo,[$ctx,#$Eoff+$lo]
+	adc	$Ehi,$Ehi,$t1
+	str	$Ehi,[$ctx,#$Eoff+$hi]
+	adds	$t2,$Tlo,$t2
+	str	$t2, [$ctx,#$Foff+$lo]
+	adc	$t3,$Thi,$t3
+	str	$t3, [$ctx,#$Foff+$hi]
+
+	ldr	$Alo,[sp,#$Goff+0]
+	ldr	$Ahi,[sp,#$Goff+4]
+	ldr	$Tlo,[sp,#$Hoff+0]
+	ldr	$Thi,[sp,#$Hoff+4]
+	ldr	$t0, [$ctx,#$Goff+$lo]
+	ldr	$t1, [$ctx,#$Goff+$hi]
+	ldr	$t2, [$ctx,#$Hoff+$lo]
+	ldr	$t3, [$ctx,#$Hoff+$hi]
+	adds	$t0,$Alo,$t0
+	str	$t0, [$ctx,#$Goff+$lo]
+	adc	$t1,$Ahi,$t1
+	str	$t1, [$ctx,#$Goff+$hi]
+	adds	$t2,$Tlo,$t2
+	str	$t2, [$ctx,#$Hoff+$lo]
+	adc	$t3,$Thi,$t3
+	str	$t3, [$ctx,#$Hoff+$hi]
+
+	add	sp,sp,#640
+	sub	$Ktbl,$Ktbl,#640
+
+	teq	$inp,$len
+	bne	.Loop
+
+	add	sp,sp,#8*9		@ destroy frame
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r12,pc}
+#else
+	ldmia	sp!,{r4-r12,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+___
+
+{
+my @Sigma0=(28,34,39);
+my @Sigma1=(14,18,41);
+my @sigma0=(1, 8, 7);
+my @sigma1=(19,61,6);
+
+my $Ktbl="r3";
+my $cnt="r12";	# volatile register known as ip, intra-procedure-call scratch
+
+my @X=map("d$_",(0..15));
+my @V=($A,$B,$C,$D,$E,$F,$G,$H)=map("d$_",(16..23));
+
+sub NEON_00_15() {
+my $i=shift;
+my ($a,$b,$c,$d,$e,$f,$g,$h)=@_;
+my ($t0,$t1,$t2,$T1,$K,$Ch,$Maj)=map("d$_",(24..31));	# temps
+
+$code.=<<___ if ($i<16 || $i&1);
+	vshr.u64	$t0,$e,#@Sigma1[0]	@ $i
+#if $i<16
+	vld1.64		{@X[$i%16]},[$inp]!	@ handles unaligned
+#endif
+	vshr.u64	$t1,$e,#@Sigma1[1]
+	vshr.u64	$t2,$e,#@Sigma1[2]
+___
+$code.=<<___;
+	vld1.64		{$K},[$Ktbl,:64]!	@ K[i++]
+	vsli.64		$t0,$e,#`64-@Sigma1[0]`
+	vsli.64		$t1,$e,#`64-@Sigma1[1]`
+	vsli.64		$t2,$e,#`64-@Sigma1[2]`
+#if $i<16 && defined(__ARMEL__)
+	vrev64.8	@X[$i],@X[$i]
+#endif
+	vadd.i64	$T1,$K,$h
+	veor		$Ch,$f,$g
+	veor		$t0,$t1
+	vand		$Ch,$e
+	veor		$t0,$t2			@ Sigma1(e)
+	veor		$Ch,$g			@ Ch(e,f,g)
+	vadd.i64	$T1,$t0
+	vshr.u64	$t0,$a,#@Sigma0[0]
+	vadd.i64	$T1,$Ch
+	vshr.u64	$t1,$a,#@Sigma0[1]
+	vshr.u64	$t2,$a,#@Sigma0[2]
+	vsli.64		$t0,$a,#`64-@Sigma0[0]`
+	vsli.64		$t1,$a,#`64-@Sigma0[1]`
+	vsli.64		$t2,$a,#`64-@Sigma0[2]`
+	vadd.i64	$T1,@X[$i%16]
+	vorr		$Maj,$a,$c
+	vand		$Ch,$a,$c
+	veor		$h,$t0,$t1
+	vand		$Maj,$b
+	veor		$h,$t2			@ Sigma0(a)
+	vorr		$Maj,$Ch		@ Maj(a,b,c)
+	vadd.i64	$h,$T1
+	vadd.i64	$d,$T1
+	vadd.i64	$h,$Maj
+___
+}
+
+sub NEON_16_79() {
+my $i=shift;
+
+if ($i&1)	{ &NEON_00_15($i,@_); return; }
+
+# 2x-vectorized, therefore runs every 2nd round
+my @X=map("q$_",(0..7));			# view @X as 128-bit vector
+my ($t0,$t1,$s0,$s1) = map("q$_",(12..15));	# temps
+my ($d0,$d1,$d2) = map("d$_",(24..26));		# temps from NEON_00_15
+my $e=@_[4];					# $e from NEON_00_15
+$i /= 2;
+$code.=<<___;
+	vshr.u64	$t0,@X[($i+7)%8],#@sigma1[0]
+	vshr.u64	$t1,@X[($i+7)%8],#@sigma1[1]
+	vshr.u64	$s1,@X[($i+7)%8],#@sigma1[2]
+	vsli.64		$t0,@X[($i+7)%8],#`64-@sigma1[0]`
+	vext.8		$s0,@X[$i%8],@X[($i+1)%8],#8	@ X[i+1]
+	vsli.64		$t1,@X[($i+7)%8],#`64-@sigma1[1]`
+	veor		$s1,$t0
+	vshr.u64	$t0,$s0,#@sigma0[0]
+	veor		$s1,$t1				@ sigma1(X[i+14])
+	vshr.u64	$t1,$s0,#@sigma0[1]
+	vadd.i64	@X[$i%8],$s1
+	vshr.u64	$s1,$s0,#@sigma0[2]
+	vsli.64		$t0,$s0,#`64-@sigma0[0]`
+	vsli.64		$t1,$s0,#`64-@sigma0[1]`
+	vext.8		$s0,@X[($i+4)%8],@X[($i+5)%8],#8	@ X[i+9]
+	veor		$s1,$t0
+	vshr.u64	$d0,$e,#@Sigma1[0]		@ from NEON_00_15
+	vadd.i64	@X[$i%8],$s0
+	vshr.u64	$d1,$e,#@Sigma1[1]		@ from NEON_00_15
+	veor		$s1,$t1				@ sigma0(X[i+1])
+	vshr.u64	$d2,$e,#@Sigma1[2]		@ from NEON_00_15
+	vadd.i64	@X[$i%8],$s1
+___
+	&NEON_00_15(2*$i,@_);
+}
+
+$code.=<<___;
+#if __ARM_ARCH__>=7
+.fpu	neon
+
+.align	4
+.LNEON:
+	dmb				@ errata #451034 on early Cortex A8
+	vstmdb	sp!,{d8-d15}		@ ABI specification says so
+	sub	$Ktbl,r3,#672		@ K512
+	vldmia	$ctx,{$A-$H}		@ load context
+.Loop_neon:
+___
+for($i=0;$i<16;$i++)	{ &NEON_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	mov		$cnt,#4
+.L16_79_neon:
+	subs		$cnt,#1
+___
+for(;$i<32;$i++)	{ &NEON_16_79($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	bne		.L16_79_neon
+
+	vldmia		$ctx,{d24-d31}	@ load context to temp
+	vadd.i64	q8,q12		@ vectorized accumulate
+	vadd.i64	q9,q13
+	vadd.i64	q10,q14
+	vadd.i64	q11,q15
+	vstmia		$ctx,{$A-$H}	@ save context
+	teq		$inp,$len
+	sub		$Ktbl,#640	@ rewind K512
+	bne		.Loop_neon
+
+	vldmia	sp!,{d8-d15}		@ epilogue
+	bx	lr
+#endif
+___
+}
+$code.=<<___;
+.size	sha512_block_data_order,.-sha512_block_data_order
+.asciz	"SHA512 block transform for ARMv4/NEON, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+.comm	OPENSSL_armcap_P,4,4
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
+print $code;
+close STDOUT; # enforce flush
diff --git a/crypto/sha/asm/sha512-mips.pl b/crypto/sha/asm/sha512-mips.pl
new file mode 100755
index 000000000000..ba5b250890e0
--- /dev/null
+++ b/crypto/sha/asm/sha512-mips.pl
@@ -0,0 +1,455 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA2 block procedures for MIPS.
+
+# October 2010.
+#
+# SHA256 performance improvement on MIPS R5000 CPU is ~27% over gcc-
+# generated code in o32 build and ~55% in n32/64 build. SHA512 [which
+# for now can only be compiled for MIPS64 ISA] improvement is modest
+# ~17%, but it comes for free, because it's same instruction sequence.
+# Improvement coefficients are for aligned input.
+
+######################################################################
+# There is a number of MIPS ABI in use, O32 and N32/64 are most
+# widely used. Then there is a new contender: NUBI. It appears that if
+# one picks the latter, it's possible to arrange code in ABI neutral
+# manner. Therefore let's stick to NUBI register layout:
+#
+($zero,$at,$t0,$t1,$t2)=map("\$$_",(0..2,24,25));
+($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7,$s8,$s9,$s10,$s11)=map("\$$_",(12..23));
+($gp,$tp,$sp,$fp,$ra)=map("\$$_",(3,28..31));
+#
+# The return value is placed in $a0. Following coding rules facilitate
+# interoperability:
+#
+# - never ever touch $tp, "thread pointer", former $gp [o32 can be
+#   excluded from the rule, because it's specified volatile];
+# - copy return value to $t0, former $v0 [or to $a0 if you're adapting
+#   old code];
+# - on O32 populate $a4-$a7 with 'lw $aN,4*N($sp)' if necessary;
+#
+# For reference here is register layout for N32/64 MIPS ABIs:
+#
+# ($zero,$at,$v0,$v1)=map("\$$_",(0..3));
+# ($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+# ($t0,$t1,$t2,$t3,$t8,$t9)=map("\$$_",(12..15,24,25));
+# ($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7)=map("\$$_",(16..23));
+# ($gp,$sp,$fp,$ra)=map("\$$_",(28..31));
+#
+$flavour = shift; # supported flavours are o32,n32,64,nubi32,nubi64
+
+if ($flavour =~ /64|n32/i) {
+	$PTR_ADD="dadd";	# incidentally works even on n32
+	$PTR_SUB="dsub";	# incidentally works even on n32
+	$REG_S="sd";
+	$REG_L="ld";
+	$PTR_SLL="dsll";	# incidentally works even on n32
+	$SZREG=8;
+} else {
+	$PTR_ADD="add";
+	$PTR_SUB="sub";
+	$REG_S="sw";
+	$REG_L="lw";
+	$PTR_SLL="sll";
+	$SZREG=4;
+}
+$pf = ($flavour =~ /nubi/i) ? $t0 : $t2;
+#
+# <appro@openssl.org>
+#
+######################################################################
+
+$big_endian=(`echo MIPSEL | $ENV{CC} -E -P -`=~/MIPSEL/)?1:0;
+
+for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);	}
+open STDOUT,">$output";
+
+if (!defined($big_endian)) { $big_endian=(unpack('L',pack('N',1))==1); }
+
+if ($output =~ /512/) {
+	$label="512";
+	$SZ=8;
+	$LD="ld";		# load from memory
+	$ST="sd";		# store to memory
+	$SLL="dsll";		# shift left logical
+	$SRL="dsrl";		# shift right logical
+	$ADDU="daddu";
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=( 7, 1, 8);	# right shift first
+	@sigma1=( 6,19,61);	# right shift first
+	$lastK=0x817;
+	$rounds=80;
+} else {
+	$label="256";
+	$SZ=4;
+	$LD="lw";		# load from memory
+	$ST="sw";		# store to memory
+	$SLL="sll";		# shift left logical
+	$SRL="srl";		# shift right logical
+	$ADDU="addu";
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 3, 7,18);	# right shift first
+	@sigma1=(10,17,19);	# right shift first
+	$lastK=0x8f2;
+	$rounds=64;
+}
+
+$MSB = $big_endian ? 0 : ($SZ-1);
+$LSB = ($SZ-1)&~$MSB;
+
+@V=($A,$B,$C,$D,$E,$F,$G,$H)=map("\$$_",(1,2,3,7,24,25,30,31));
+@X=map("\$$_",(8..23));
+
+$ctx=$a0;
+$inp=$a1;
+$len=$a2;	$Ktbl=$len;
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+my ($T1,$tmp0,$tmp1,$tmp2)=(@X[4],@X[5],@X[6],@X[7]);
+
+$code.=<<___ if ($i<15);
+	${LD}l	@X[1],`($i+1)*$SZ+$MSB`($inp)
+	${LD}r	@X[1],`($i+1)*$SZ+$LSB`($inp)
+___
+$code.=<<___	if (!$big_endian && $i<16 && $SZ==4);
+	srl	$tmp0,@X[0],24		# byte swap($i)
+	srl	$tmp1,@X[0],8
+	andi	$tmp2,@X[0],0xFF00
+	sll	@X[0],@X[0],24
+	andi	$tmp1,0xFF00
+	sll	$tmp2,$tmp2,8
+	or	@X[0],$tmp0
+	or	$tmp1,$tmp2
+	or	@X[0],$tmp1
+___
+$code.=<<___	if (!$big_endian && $i<16 && $SZ==8);
+	ori	$tmp0,$zero,0xFF
+	dsll	$tmp2,$tmp0,32
+	or	$tmp0,$tmp2		# 0x000000FF000000FF
+	and	$tmp1,@X[0],$tmp0	# byte swap($i)
+	dsrl	$tmp2,@X[0],24
+	dsll	$tmp1,24
+	and	$tmp2,$tmp0
+	dsll	$tmp0,8			# 0x0000FF000000FF00
+	or	$tmp1,$tmp2
+	and	$tmp2,@X[0],$tmp0
+	dsrl	@X[0],8
+	dsll	$tmp2,8
+	and	@X[0],$tmp0
+	or	$tmp1,$tmp2
+	or	@X[0],$tmp1
+	dsrl	$tmp1,@X[0],32
+	dsll	@X[0],32
+	or	@X[0],$tmp1
+___
+$code.=<<___;
+	$ADDU	$T1,$X[0],$h			# $i
+	$SRL	$h,$e,@Sigma1[0]
+	xor	$tmp2,$f,$g
+	$SLL	$tmp1,$e,`$SZ*8-@Sigma1[2]`
+	and	$tmp2,$e
+	$SRL	$tmp0,$e,@Sigma1[1]
+	xor	$h,$tmp1
+	$SLL	$tmp1,$e,`$SZ*8-@Sigma1[1]`
+	xor	$h,$tmp0
+	$SRL	$tmp0,$e,@Sigma1[2]
+	xor	$h,$tmp1
+	$SLL	$tmp1,$e,`$SZ*8-@Sigma1[0]`
+	xor	$h,$tmp0
+	xor	$tmp2,$g			# Ch(e,f,g)
+	xor	$tmp0,$tmp1,$h			# Sigma1(e)
+
+	$SRL	$h,$a,@Sigma0[0]
+	$ADDU	$T1,$tmp2
+	$LD	$tmp2,`$i*$SZ`($Ktbl)		# K[$i]
+	$SLL	$tmp1,$a,`$SZ*8-@Sigma0[2]`
+	$ADDU	$T1,$tmp0
+	$SRL	$tmp0,$a,@Sigma0[1]
+	xor	$h,$tmp1
+	$SLL	$tmp1,$a,`$SZ*8-@Sigma0[1]`
+	xor	$h,$tmp0
+	$SRL	$tmp0,$a,@Sigma0[2]
+	xor	$h,$tmp1
+	$SLL	$tmp1,$a,`$SZ*8-@Sigma0[0]`
+	xor	$h,$tmp0
+	$ST	@X[0],`($i%16)*$SZ`($sp)	# offload to ring buffer
+	xor	$h,$tmp1			# Sigma0(a)
+
+	or	$tmp0,$a,$b
+	and	$tmp1,$a,$b
+	and	$tmp0,$c
+	or	$tmp1,$tmp0			# Maj(a,b,c)
+	$ADDU	$T1,$tmp2			# +=K[$i]
+	$ADDU	$h,$tmp1
+
+	$ADDU	$d,$T1
+	$ADDU	$h,$T1
+___
+$code.=<<___ if ($i>=13);
+	$LD	@X[3],`(($i+3)%16)*$SZ`($sp)	# prefetch from ring buffer
+___
+}
+
+sub BODY_16_XX {
+my $i=@_[0];
+my ($tmp0,$tmp1,$tmp2,$tmp3)=(@X[4],@X[5],@X[6],@X[7]);
+
+$code.=<<___;
+	$SRL	$tmp2,@X[1],@sigma0[0]		# Xupdate($i)
+	$ADDU	@X[0],@X[9]			# +=X[i+9]
+	$SLL	$tmp1,@X[1],`$SZ*8-@sigma0[2]`
+	$SRL	$tmp0,@X[1],@sigma0[1]
+	xor	$tmp2,$tmp1
+	$SLL	$tmp1,`@sigma0[2]-@sigma0[1]`
+	xor	$tmp2,$tmp0
+	$SRL	$tmp0,@X[1],@sigma0[2]
+	xor	$tmp2,$tmp1
+
+	$SRL	$tmp3,@X[14],@sigma1[0]
+	xor	$tmp2,$tmp0			# sigma0(X[i+1])
+	$SLL	$tmp1,@X[14],`$SZ*8-@sigma1[2]`
+	$ADDU	@X[0],$tmp2
+	$SRL	$tmp0,@X[14],@sigma1[1]
+	xor	$tmp3,$tmp1
+	$SLL	$tmp1,`@sigma1[2]-@sigma1[1]`
+	xor	$tmp3,$tmp0
+	$SRL	$tmp0,@X[14],@sigma1[2]
+	xor	$tmp3,$tmp1
+
+	xor	$tmp3,$tmp0			# sigma1(X[i+14])
+	$ADDU	@X[0],$tmp3
+___
+	&BODY_00_15(@_);
+}
+
+$FRAMESIZE=16*$SZ+16*$SZREG;
+$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0xc0fff008 : 0xc0ff0000;
+
+$code.=<<___;
+#ifdef OPENSSL_FIPSCANISTER
+# include <openssl/fipssyms.h>
+#endif
+
+.text
+.set	noat
+#if !defined(__vxworks) || defined(__pic__)
+.option	pic2
+#endif
+
+.align	5
+.globl	sha${label}_block_data_order
+.ent	sha${label}_block_data_order
+sha${label}_block_data_order:
+	.frame	$sp,$FRAMESIZE,$ra
+	.mask	$SAVED_REGS_MASK,-$SZREG
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /o32/i);	# o32 PIC-ification
+	.cpload	$pf
+___
+$code.=<<___;
+	$PTR_SUB $sp,$FRAMESIZE
+	$REG_S	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_S	$fp,$FRAMESIZE-2*$SZREG($sp)
+	$REG_S	$s11,$FRAMESIZE-3*$SZREG($sp)
+	$REG_S	$s10,$FRAMESIZE-4*$SZREG($sp)
+	$REG_S	$s9,$FRAMESIZE-5*$SZREG($sp)
+	$REG_S	$s8,$FRAMESIZE-6*$SZREG($sp)
+	$REG_S	$s7,$FRAMESIZE-7*$SZREG($sp)
+	$REG_S	$s6,$FRAMESIZE-8*$SZREG($sp)
+	$REG_S	$s5,$FRAMESIZE-9*$SZREG($sp)
+	$REG_S	$s4,$FRAMESIZE-10*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);	# optimize non-nubi prologue
+	$REG_S	$s3,$FRAMESIZE-11*$SZREG($sp)
+	$REG_S	$s2,$FRAMESIZE-12*$SZREG($sp)
+	$REG_S	$s1,$FRAMESIZE-13*$SZREG($sp)
+	$REG_S	$s0,$FRAMESIZE-14*$SZREG($sp)
+	$REG_S	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___;
+	$PTR_SLL @X[15],$len,`log(16*$SZ)/log(2)`
+___
+$code.=<<___ if ($flavour !~ /o32/i);	# non-o32 PIC-ification
+	.cplocal	$Ktbl
+	.cpsetup	$pf,$zero,sha${label}_block_data_order
+___
+$code.=<<___;
+	.set	reorder
+	la	$Ktbl,K${label}		# PIC-ified 'load address'
+
+	$LD	$A,0*$SZ($ctx)		# load context
+	$LD	$B,1*$SZ($ctx)
+	$LD	$C,2*$SZ($ctx)
+	$LD	$D,3*$SZ($ctx)
+	$LD	$E,4*$SZ($ctx)
+	$LD	$F,5*$SZ($ctx)
+	$LD	$G,6*$SZ($ctx)
+	$LD	$H,7*$SZ($ctx)
+
+	$PTR_ADD @X[15],$inp		# pointer to the end of input
+	$REG_S	@X[15],16*$SZ($sp)
+	b	.Loop
+
+.align	5
+.Loop:
+	${LD}l	@X[0],$MSB($inp)
+	${LD}r	@X[0],$LSB($inp)
+___
+for ($i=0;$i<16;$i++)
+{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); push(@X,shift(@X)); }
+$code.=<<___;
+	b	.L16_xx
+.align	4
+.L16_xx:
+___
+for (;$i<32;$i++)
+{ &BODY_16_XX($i,@V); unshift(@V,pop(@V)); push(@X,shift(@X)); }
+$code.=<<___;
+	and	@X[6],0xfff
+	li	@X[7],$lastK
+	.set	noreorder
+	bne	@X[6],@X[7],.L16_xx
+	$PTR_ADD $Ktbl,16*$SZ		# Ktbl+=16
+
+	$REG_L	@X[15],16*$SZ($sp)	# restore pointer to the end of input
+	$LD	@X[0],0*$SZ($ctx)
+	$LD	@X[1],1*$SZ($ctx)
+	$LD	@X[2],2*$SZ($ctx)
+	$PTR_ADD $inp,16*$SZ
+	$LD	@X[3],3*$SZ($ctx)
+	$ADDU	$A,@X[0]
+	$LD	@X[4],4*$SZ($ctx)
+	$ADDU	$B,@X[1]
+	$LD	@X[5],5*$SZ($ctx)
+	$ADDU	$C,@X[2]
+	$LD	@X[6],6*$SZ($ctx)
+	$ADDU	$D,@X[3]
+	$LD	@X[7],7*$SZ($ctx)
+	$ADDU	$E,@X[4]
+	$ST	$A,0*$SZ($ctx)
+	$ADDU	$F,@X[5]
+	$ST	$B,1*$SZ($ctx)
+	$ADDU	$G,@X[6]
+	$ST	$C,2*$SZ($ctx)
+	$ADDU	$H,@X[7]
+	$ST	$D,3*$SZ($ctx)
+	$ST	$E,4*$SZ($ctx)
+	$ST	$F,5*$SZ($ctx)
+	$ST	$G,6*$SZ($ctx)
+	$ST	$H,7*$SZ($ctx)
+
+	bnel	$inp,@X[15],.Loop
+	$PTR_SUB $Ktbl,`($rounds-16)*$SZ`	# rewind $Ktbl
+
+	$REG_L	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_L	$fp,$FRAMESIZE-2*$SZREG($sp)
+	$REG_L	$s11,$FRAMESIZE-3*$SZREG($sp)
+	$REG_L	$s10,$FRAMESIZE-4*$SZREG($sp)
+	$REG_L	$s9,$FRAMESIZE-5*$SZREG($sp)
+	$REG_L	$s8,$FRAMESIZE-6*$SZREG($sp)
+	$REG_L	$s7,$FRAMESIZE-7*$SZREG($sp)
+	$REG_L	$s6,$FRAMESIZE-8*$SZREG($sp)
+	$REG_L	$s5,$FRAMESIZE-9*$SZREG($sp)
+	$REG_L	$s4,$FRAMESIZE-10*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$s3,$FRAMESIZE-11*$SZREG($sp)
+	$REG_L	$s2,$FRAMESIZE-12*$SZREG($sp)
+	$REG_L	$s1,$FRAMESIZE-13*$SZREG($sp)
+	$REG_L	$s0,$FRAMESIZE-14*$SZREG($sp)
+	$REG_L	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___;
+	jr	$ra
+	$PTR_ADD $sp,$FRAMESIZE
+.end	sha${label}_block_data_order
+
+.rdata
+.align	5
+K${label}:
+___
+if ($SZ==4) {
+$code.=<<___;
+	.word	0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5
+	.word	0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5
+	.word	0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3
+	.word	0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174
+	.word	0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc
+	.word	0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da
+	.word	0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7
+	.word	0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967
+	.word	0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13
+	.word	0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85
+	.word	0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3
+	.word	0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070
+	.word	0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5
+	.word	0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3
+	.word	0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208
+	.word	0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
+___
+} else {
+$code.=<<___;
+	.dword	0x428a2f98d728ae22, 0x7137449123ef65cd
+	.dword	0xb5c0fbcfec4d3b2f, 0xe9b5dba58189dbbc
+	.dword	0x3956c25bf348b538, 0x59f111f1b605d019
+	.dword	0x923f82a4af194f9b, 0xab1c5ed5da6d8118
+	.dword	0xd807aa98a3030242, 0x12835b0145706fbe
+	.dword	0x243185be4ee4b28c, 0x550c7dc3d5ffb4e2
+	.dword	0x72be5d74f27b896f, 0x80deb1fe3b1696b1
+	.dword	0x9bdc06a725c71235, 0xc19bf174cf692694
+	.dword	0xe49b69c19ef14ad2, 0xefbe4786384f25e3
+	.dword	0x0fc19dc68b8cd5b5, 0x240ca1cc77ac9c65
+	.dword	0x2de92c6f592b0275, 0x4a7484aa6ea6e483
+	.dword	0x5cb0a9dcbd41fbd4, 0x76f988da831153b5
+	.dword	0x983e5152ee66dfab, 0xa831c66d2db43210
+	.dword	0xb00327c898fb213f, 0xbf597fc7beef0ee4
+	.dword	0xc6e00bf33da88fc2, 0xd5a79147930aa725
+	.dword	0x06ca6351e003826f, 0x142929670a0e6e70
+	.dword	0x27b70a8546d22ffc, 0x2e1b21385c26c926
+	.dword	0x4d2c6dfc5ac42aed, 0x53380d139d95b3df
+	.dword	0x650a73548baf63de, 0x766a0abb3c77b2a8
+	.dword	0x81c2c92e47edaee6, 0x92722c851482353b
+	.dword	0xa2bfe8a14cf10364, 0xa81a664bbc423001
+	.dword	0xc24b8b70d0f89791, 0xc76c51a30654be30
+	.dword	0xd192e819d6ef5218, 0xd69906245565a910
+	.dword	0xf40e35855771202a, 0x106aa07032bbd1b8
+	.dword	0x19a4c116b8d2d0c8, 0x1e376c085141ab53
+	.dword	0x2748774cdf8eeb99, 0x34b0bcb5e19b48a8
+	.dword	0x391c0cb3c5c95a63, 0x4ed8aa4ae3418acb
+	.dword	0x5b9cca4f7763e373, 0x682e6ff3d6b2b8a3
+	.dword	0x748f82ee5defb2fc, 0x78a5636f43172f60
+	.dword	0x84c87814a1f0ab72, 0x8cc702081a6439ec
+	.dword	0x90befffa23631e28, 0xa4506cebde82bde9
+	.dword	0xbef9a3f7b2c67915, 0xc67178f2e372532b
+	.dword	0xca273eceea26619c, 0xd186b8c721c0c207
+	.dword	0xeada7dd6cde0eb1e, 0xf57d4f7fee6ed178
+	.dword	0x06f067aa72176fba, 0x0a637dc5a2c898a6
+	.dword	0x113f9804bef90dae, 0x1b710b35131c471b
+	.dword	0x28db77f523047d84, 0x32caab7b40c72493
+	.dword	0x3c9ebe0a15c9bebc, 0x431d67c49c100d4c
+	.dword	0x4cc5d4becb3e42b6, 0x597f299cfc657e2a
+	.dword	0x5fcb6fab3ad6faec, 0x6c44198c4a475817
+___
+}
+$code.=<<___;
+.asciiz	"SHA${label} for MIPS, CRYPTOGAMS by <appro\@openssl.org>"
+.align	5
+
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha512-parisc.pl b/crypto/sha/asm/sha512-parisc.pl
new file mode 100755
index 000000000000..e24ee58ae97f
--- /dev/null
+++ b/crypto/sha/asm/sha512-parisc.pl
@@ -0,0 +1,791 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA256/512 block procedure for PA-RISC.
+
+# June 2009.
+#
+# SHA256 performance is >75% better than gcc 3.2 generated code on
+# PA-7100LC. Compared to code generated by vendor compiler this
+# implementation is almost 70% faster in 64-bit build, but delivers
+# virtually same performance in 32-bit build on PA-8600.
+#
+# SHA512 performance is >2.9x better than gcc 3.2 generated code on
+# PA-7100LC, PA-RISC 1.1 processor. Then implementation detects if the
+# code is executed on PA-RISC 2.0 processor and switches to 64-bit
+# code path delivering adequate peformance even in "blended" 32-bit
+# build. Though 64-bit code is not any faster than code generated by
+# vendor compiler on PA-8600...
+#
+# Special thanks to polarhome.com for providing HP-UX account.
+
+$flavour = shift;
+$output = shift;
+open STDOUT,">$output";
+
+if ($flavour =~ /64/) {
+	$LEVEL		="2.0W";
+	$SIZE_T		=8;
+	$FRAME_MARKER	=80;
+	$SAVED_RP	=16;
+	$PUSH		="std";
+	$PUSHMA		="std,ma";
+	$POP		="ldd";
+	$POPMB		="ldd,mb";
+} else {
+	$LEVEL		="1.0";
+	$SIZE_T		=4;
+	$FRAME_MARKER	=48;
+	$SAVED_RP	=20;
+	$PUSH		="stw";
+	$PUSHMA		="stwm";
+	$POP		="ldw";
+	$POPMB		="ldwm";
+}
+
+if ($output =~ /512/) {
+	$func="sha512_block_data_order";
+	$SZ=8;
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=(1,  8, 7);
+	@sigma1=(19,61, 6);
+	$rounds=80;
+	$LAST10BITS=0x017;
+	$LD="ldd";
+	$LDM="ldd,ma";
+	$ST="std";
+} else {
+	$func="sha256_block_data_order";
+	$SZ=4;
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 7,18, 3);
+	@sigma1=(17,19,10);
+	$rounds=64;
+	$LAST10BITS=0x0f2;
+	$LD="ldw";
+	$LDM="ldwm";
+	$ST="stw";
+}
+
+$FRAME=16*$SIZE_T+$FRAME_MARKER;# 16 saved regs + frame marker
+				#                 [+ argument transfer]
+$XOFF=16*$SZ+32;		# local variables
+$FRAME+=$XOFF;
+$XOFF+=$FRAME_MARKER;		# distance between %sp and local variables
+
+$ctx="%r26";	# zapped by $a0
+$inp="%r25";	# zapped by $a1
+$num="%r24";	# zapped by $t0
+
+$a0 ="%r26";
+$a1 ="%r25";
+$t0 ="%r24";
+$t1 ="%r29";
+$Tbl="%r31";
+
+@V=($A,$B,$C,$D,$E,$F,$G,$H)=("%r17","%r18","%r19","%r20","%r21","%r22","%r23","%r28");
+
+@X=("%r1", "%r2", "%r3", "%r4", "%r5", "%r6", "%r7", "%r8",
+    "%r9", "%r10","%r11","%r12","%r13","%r14","%r15","%r16",$inp);
+
+sub ROUND_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+$code.=<<___;
+	_ror	$e,$Sigma1[0],$a0
+	and	$f,$e,$t0
+	_ror	$e,$Sigma1[1],$a1
+	addl	$t1,$h,$h
+	andcm	$g,$e,$t1
+	xor	$a1,$a0,$a0
+	_ror	$a1,`$Sigma1[2]-$Sigma1[1]`,$a1
+	or	$t0,$t1,$t1		; Ch(e,f,g)
+	addl	@X[$i%16],$h,$h
+	xor	$a0,$a1,$a1		; Sigma1(e)
+	addl	$t1,$h,$h
+	_ror	$a,$Sigma0[0],$a0
+	addl	$a1,$h,$h
+
+	_ror	$a,$Sigma0[1],$a1
+	and	$a,$b,$t0
+	and	$a,$c,$t1
+	xor	$a1,$a0,$a0
+	_ror	$a1,`$Sigma0[2]-$Sigma0[1]`,$a1
+	xor	$t1,$t0,$t0
+	and	$b,$c,$t1
+	xor	$a0,$a1,$a1		; Sigma0(a)
+	addl	$h,$d,$d
+	xor	$t1,$t0,$t0		; Maj(a,b,c)
+	`"$LDM	$SZ($Tbl),$t1" if ($i<15)`
+	addl	$a1,$h,$h
+	addl	$t0,$h,$h
+
+___
+}
+
+sub ROUND_16_xx {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+$i-=16;
+$code.=<<___;
+	_ror	@X[($i+1)%16],$sigma0[0],$a0
+	_ror	@X[($i+1)%16],$sigma0[1],$a1
+	addl	@X[($i+9)%16],@X[$i],@X[$i]
+	_ror	@X[($i+14)%16],$sigma1[0],$t0
+	_ror	@X[($i+14)%16],$sigma1[1],$t1
+	xor	$a1,$a0,$a0
+	_shr	@X[($i+1)%16],$sigma0[2],$a1
+	xor	$t1,$t0,$t0
+	_shr	@X[($i+14)%16],$sigma1[2],$t1
+	xor	$a1,$a0,$a0		; sigma0(X[(i+1)&0x0f])
+	xor	$t1,$t0,$t0		; sigma1(X[(i+14)&0x0f])
+	$LDM	$SZ($Tbl),$t1
+	addl	$a0,@X[$i],@X[$i]
+	addl	$t0,@X[$i],@X[$i]
+___
+$code.=<<___ if ($i==15);
+	extru	$t1,31,10,$a1
+	comiclr,<> $LAST10BITS,$a1,%r0
+	ldo	1($Tbl),$Tbl		; signal end of $Tbl
+___
+&ROUND_00_15($i+16,$a,$b,$c,$d,$e,$f,$g,$h);
+}
+
+$code=<<___;
+	.LEVEL	$LEVEL
+	.SPACE	\$TEXT\$
+	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY
+
+	.ALIGN	64
+L\$table
+___
+$code.=<<___ if ($SZ==8);
+	.WORD	0x428a2f98,0xd728ae22,0x71374491,0x23ef65cd
+	.WORD	0xb5c0fbcf,0xec4d3b2f,0xe9b5dba5,0x8189dbbc
+	.WORD	0x3956c25b,0xf348b538,0x59f111f1,0xb605d019
+	.WORD	0x923f82a4,0xaf194f9b,0xab1c5ed5,0xda6d8118
+	.WORD	0xd807aa98,0xa3030242,0x12835b01,0x45706fbe
+	.WORD	0x243185be,0x4ee4b28c,0x550c7dc3,0xd5ffb4e2
+	.WORD	0x72be5d74,0xf27b896f,0x80deb1fe,0x3b1696b1
+	.WORD	0x9bdc06a7,0x25c71235,0xc19bf174,0xcf692694
+	.WORD	0xe49b69c1,0x9ef14ad2,0xefbe4786,0x384f25e3
+	.WORD	0x0fc19dc6,0x8b8cd5b5,0x240ca1cc,0x77ac9c65
+	.WORD	0x2de92c6f,0x592b0275,0x4a7484aa,0x6ea6e483
+	.WORD	0x5cb0a9dc,0xbd41fbd4,0x76f988da,0x831153b5
+	.WORD	0x983e5152,0xee66dfab,0xa831c66d,0x2db43210
+	.WORD	0xb00327c8,0x98fb213f,0xbf597fc7,0xbeef0ee4
+	.WORD	0xc6e00bf3,0x3da88fc2,0xd5a79147,0x930aa725
+	.WORD	0x06ca6351,0xe003826f,0x14292967,0x0a0e6e70
+	.WORD	0x27b70a85,0x46d22ffc,0x2e1b2138,0x5c26c926
+	.WORD	0x4d2c6dfc,0x5ac42aed,0x53380d13,0x9d95b3df
+	.WORD	0x650a7354,0x8baf63de,0x766a0abb,0x3c77b2a8
+	.WORD	0x81c2c92e,0x47edaee6,0x92722c85,0x1482353b
+	.WORD	0xa2bfe8a1,0x4cf10364,0xa81a664b,0xbc423001
+	.WORD	0xc24b8b70,0xd0f89791,0xc76c51a3,0x0654be30
+	.WORD	0xd192e819,0xd6ef5218,0xd6990624,0x5565a910
+	.WORD	0xf40e3585,0x5771202a,0x106aa070,0x32bbd1b8
+	.WORD	0x19a4c116,0xb8d2d0c8,0x1e376c08,0x5141ab53
+	.WORD	0x2748774c,0xdf8eeb99,0x34b0bcb5,0xe19b48a8
+	.WORD	0x391c0cb3,0xc5c95a63,0x4ed8aa4a,0xe3418acb
+	.WORD	0x5b9cca4f,0x7763e373,0x682e6ff3,0xd6b2b8a3
+	.WORD	0x748f82ee,0x5defb2fc,0x78a5636f,0x43172f60
+	.WORD	0x84c87814,0xa1f0ab72,0x8cc70208,0x1a6439ec
+	.WORD	0x90befffa,0x23631e28,0xa4506ceb,0xde82bde9
+	.WORD	0xbef9a3f7,0xb2c67915,0xc67178f2,0xe372532b
+	.WORD	0xca273ece,0xea26619c,0xd186b8c7,0x21c0c207
+	.WORD	0xeada7dd6,0xcde0eb1e,0xf57d4f7f,0xee6ed178
+	.WORD	0x06f067aa,0x72176fba,0x0a637dc5,0xa2c898a6
+	.WORD	0x113f9804,0xbef90dae,0x1b710b35,0x131c471b
+	.WORD	0x28db77f5,0x23047d84,0x32caab7b,0x40c72493
+	.WORD	0x3c9ebe0a,0x15c9bebc,0x431d67c4,0x9c100d4c
+	.WORD	0x4cc5d4be,0xcb3e42b6,0x597f299c,0xfc657e2a
+	.WORD	0x5fcb6fab,0x3ad6faec,0x6c44198c,0x4a475817
+___
+$code.=<<___ if ($SZ==4);
+	.WORD	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.WORD	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.WORD	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.WORD	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.WORD	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.WORD	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.WORD	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.WORD	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.WORD	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.WORD	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.WORD	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.WORD	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.WORD	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.WORD	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.WORD	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.WORD	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+___
+$code.=<<___;
+
+	.EXPORT	$func,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR
+	.ALIGN	64
+$func
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-16*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=18
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
+	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
+	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
+	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
+	$PUSH	%r11,`-$FRAME+8*$SIZE_T`(%sp)
+	$PUSH	%r12,`-$FRAME+9*$SIZE_T`(%sp)
+	$PUSH	%r13,`-$FRAME+10*$SIZE_T`(%sp)
+	$PUSH	%r14,`-$FRAME+11*$SIZE_T`(%sp)
+	$PUSH	%r15,`-$FRAME+12*$SIZE_T`(%sp)
+	$PUSH	%r16,`-$FRAME+13*$SIZE_T`(%sp)
+	$PUSH	%r17,`-$FRAME+14*$SIZE_T`(%sp)
+	$PUSH	%r18,`-$FRAME+15*$SIZE_T`(%sp)
+
+	_shl	$num,`log(16*$SZ)/log(2)`,$num
+	addl	$inp,$num,$num		; $num to point at the end of $inp
+
+	$PUSH	$num,`-$FRAME_MARKER-4*$SIZE_T`(%sp)	; save arguments
+	$PUSH	$inp,`-$FRAME_MARKER-3*$SIZE_T`(%sp)
+	$PUSH	$ctx,`-$FRAME_MARKER-2*$SIZE_T`(%sp)
+
+	blr	%r0,$Tbl
+	ldi	3,$t1
+L\$pic
+	andcm	$Tbl,$t1,$Tbl		; wipe privilege level
+	ldo	L\$table-L\$pic($Tbl),$Tbl
+___
+$code.=<<___ if ($SZ==8 && $SIZE_T==4);
+	ldi	31,$t1
+	mtctl	$t1,%cr11
+	extrd,u,*= $t1,%sar,1,$t1	; executes on PA-RISC 1.0
+	b	L\$parisc1
+	nop
+___
+$code.=<<___;
+	$LD	`0*$SZ`($ctx),$A	; load context
+	$LD	`1*$SZ`($ctx),$B
+	$LD	`2*$SZ`($ctx),$C
+	$LD	`3*$SZ`($ctx),$D
+	$LD	`4*$SZ`($ctx),$E
+	$LD	`5*$SZ`($ctx),$F
+	$LD	`6*$SZ`($ctx),$G
+	$LD	`7*$SZ`($ctx),$H
+
+	extru	$inp,31,`log($SZ)/log(2)`,$t0
+	sh3addl	$t0,%r0,$t0
+	subi	`8*$SZ`,$t0,$t0
+	mtctl	$t0,%cr11		; load %sar with align factor
+
+L\$oop
+	ldi	`$SZ-1`,$t0
+	$LDM	$SZ($Tbl),$t1
+	andcm	$inp,$t0,$t0		; align $inp
+___
+	for ($i=0;$i<15;$i++) {		# load input block
+	$code.="\t$LD	`$SZ*$i`($t0),@X[$i]\n";		}
+$code.=<<___;
+	cmpb,*=	$inp,$t0,L\$aligned
+	$LD	`$SZ*15`($t0),@X[15]
+	$LD	`$SZ*16`($t0),@X[16]
+___
+	for ($i=0;$i<16;$i++) {		# align data
+	$code.="\t_align	@X[$i],@X[$i+1],@X[$i]\n";	}
+$code.=<<___;
+L\$aligned
+	nop	; otherwise /usr/ccs/bin/as is confused by below .WORD
+___
+
+for($i=0;$i<16;$i++)	{ &ROUND_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+L\$rounds
+	nop	; otherwise /usr/ccs/bin/as is confused by below .WORD
+___
+for(;$i<32;$i++)	{ &ROUND_16_xx($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	bb,>=	$Tbl,31,L\$rounds	; end of $Tbl signalled?
+	nop
+
+	$POP	`-$FRAME_MARKER-2*$SIZE_T`(%sp),$ctx	; restore arguments
+	$POP	`-$FRAME_MARKER-3*$SIZE_T`(%sp),$inp
+	$POP	`-$FRAME_MARKER-4*$SIZE_T`(%sp),$num
+	ldo	`-$rounds*$SZ-1`($Tbl),$Tbl		; rewind $Tbl
+
+	$LD	`0*$SZ`($ctx),@X[0]	; load context
+	$LD	`1*$SZ`($ctx),@X[1]
+	$LD	`2*$SZ`($ctx),@X[2]
+	$LD	`3*$SZ`($ctx),@X[3]
+	$LD	`4*$SZ`($ctx),@X[4]
+	$LD	`5*$SZ`($ctx),@X[5]
+	addl	@X[0],$A,$A
+	$LD	`6*$SZ`($ctx),@X[6]
+	addl	@X[1],$B,$B
+	$LD	`7*$SZ`($ctx),@X[7]
+	ldo	`16*$SZ`($inp),$inp	; advance $inp
+
+	$ST	$A,`0*$SZ`($ctx)	; save context
+	addl	@X[2],$C,$C
+	$ST	$B,`1*$SZ`($ctx)
+	addl	@X[3],$D,$D
+	$ST	$C,`2*$SZ`($ctx)
+	addl	@X[4],$E,$E
+	$ST	$D,`3*$SZ`($ctx)
+	addl	@X[5],$F,$F
+	$ST	$E,`4*$SZ`($ctx)
+	addl	@X[6],$G,$G
+	$ST	$F,`5*$SZ`($ctx)
+	addl	@X[7],$H,$H
+	$ST	$G,`6*$SZ`($ctx)
+	$ST	$H,`7*$SZ`($ctx)
+
+	cmpb,*<>,n $inp,$num,L\$oop
+	$PUSH	$inp,`-$FRAME_MARKER-3*$SIZE_T`(%sp)	; save $inp
+___
+if ($SZ==8 && $SIZE_T==4)	# SHA512 for 32-bit PA-RISC 1.0
+{{
+$code.=<<___;
+	b	L\$done
+	nop
+
+	.ALIGN	64
+L\$parisc1
+___
+
+@V=(  $Ahi,  $Alo,  $Bhi,  $Blo,  $Chi,  $Clo,  $Dhi,  $Dlo,
+      $Ehi,  $Elo,  $Fhi,  $Flo,  $Ghi,  $Glo,  $Hhi,  $Hlo) = 
+   ( "%r1", "%r2", "%r3", "%r4", "%r5", "%r6", "%r7", "%r8",
+     "%r9","%r10","%r11","%r12","%r13","%r14","%r15","%r16");
+$a0 ="%r17";
+$a1 ="%r18";
+$a2 ="%r19";
+$a3 ="%r20";
+$t0 ="%r21";
+$t1 ="%r22";
+$t2 ="%r28";
+$t3 ="%r29";
+$Tbl="%r31";
+
+@X=("%r23","%r24","%r25","%r26");	# zaps $num,$inp,$ctx
+
+sub ROUND_00_15_pa1 {
+my ($i,$ahi,$alo,$bhi,$blo,$chi,$clo,$dhi,$dlo,
+       $ehi,$elo,$fhi,$flo,$ghi,$glo,$hhi,$hlo,$flag)=@_;
+my ($Xhi,$Xlo,$Xnhi,$Xnlo) = @X;
+
+$code.=<<___ if (!$flag);
+	ldw	`-$XOFF+8*(($i+1)%16)`(%sp),$Xnhi
+	ldw	`-$XOFF+8*(($i+1)%16)+4`(%sp),$Xnlo	; load X[i+1]
+___
+$code.=<<___;
+	shd	$ehi,$elo,$Sigma1[0],$t0
+	 add	$Xlo,$hlo,$hlo
+	shd	$elo,$ehi,$Sigma1[0],$t1
+	 addc	$Xhi,$hhi,$hhi		; h += X[i]
+	shd	$ehi,$elo,$Sigma1[1],$t2
+	 ldwm	8($Tbl),$Xhi
+	shd	$elo,$ehi,$Sigma1[1],$t3
+	 ldw	-4($Tbl),$Xlo		; load K[i]
+	xor	$t2,$t0,$t0
+	xor	$t3,$t1,$t1
+	 and	$flo,$elo,$a0
+	 and	$fhi,$ehi,$a1
+	shd	$ehi,$elo,$Sigma1[2],$t2
+	 andcm	$glo,$elo,$a2
+	shd	$elo,$ehi,$Sigma1[2],$t3
+	 andcm	$ghi,$ehi,$a3
+	xor	$t2,$t0,$t0
+	xor	$t3,$t1,$t1		; Sigma1(e)
+	add	$Xlo,$hlo,$hlo
+	 xor	$a2,$a0,$a0
+	addc	$Xhi,$hhi,$hhi		; h += K[i]
+	 xor	$a3,$a1,$a1		; Ch(e,f,g)
+
+	 add	$t0,$hlo,$hlo
+	shd	$ahi,$alo,$Sigma0[0],$t0
+	 addc	$t1,$hhi,$hhi		; h += Sigma1(e)
+	shd	$alo,$ahi,$Sigma0[0],$t1	
+	 add	$a0,$hlo,$hlo
+	shd	$ahi,$alo,$Sigma0[1],$t2
+	 addc	$a1,$hhi,$hhi		; h += Ch(e,f,g)
+	shd	$alo,$ahi,$Sigma0[1],$t3
+
+	xor	$t2,$t0,$t0
+	xor	$t3,$t1,$t1
+	shd	$ahi,$alo,$Sigma0[2],$t2
+	and	$alo,$blo,$a0
+	shd	$alo,$ahi,$Sigma0[2],$t3
+	and	$ahi,$bhi,$a1
+	xor	$t2,$t0,$t0
+	xor	$t3,$t1,$t1		; Sigma0(a)
+
+	and	$alo,$clo,$a2
+	and	$ahi,$chi,$a3
+	xor	$a2,$a0,$a0
+	 add	$hlo,$dlo,$dlo
+	xor	$a3,$a1,$a1
+	 addc	$hhi,$dhi,$dhi		; d += h
+	and	$blo,$clo,$a2
+	 add	$t0,$hlo,$hlo
+	and	$bhi,$chi,$a3
+	 addc	$t1,$hhi,$hhi		; h += Sigma0(a)
+	xor	$a2,$a0,$a0
+	 add	$a0,$hlo,$hlo
+	xor	$a3,$a1,$a1		; Maj(a,b,c)
+	 addc	$a1,$hhi,$hhi		; h += Maj(a,b,c)
+
+___
+$code.=<<___ if ($i==15 && $flag);
+	extru	$Xlo,31,10,$Xlo
+	comiclr,= $LAST10BITS,$Xlo,%r0
+	b	L\$rounds_pa1
+	nop
+___
+push(@X,shift(@X)); push(@X,shift(@X));
+}
+
+sub ROUND_16_xx_pa1 {
+my ($Xhi,$Xlo,$Xnhi,$Xnlo) = @X;
+my ($i)=shift;
+$i-=16;
+$code.=<<___;
+	ldw	`-$XOFF+8*(($i+1)%16)`(%sp),$Xnhi
+	ldw	`-$XOFF+8*(($i+1)%16)+4`(%sp),$Xnlo	; load X[i+1]
+	ldw	`-$XOFF+8*(($i+9)%16)`(%sp),$a1
+	ldw	`-$XOFF+8*(($i+9)%16)+4`(%sp),$a0	; load X[i+9]
+	ldw	`-$XOFF+8*(($i+14)%16)`(%sp),$a3
+	ldw	`-$XOFF+8*(($i+14)%16)+4`(%sp),$a2	; load X[i+14]
+	shd	$Xnhi,$Xnlo,$sigma0[0],$t0
+	shd	$Xnlo,$Xnhi,$sigma0[0],$t1
+	 add	$a0,$Xlo,$Xlo
+	shd	$Xnhi,$Xnlo,$sigma0[1],$t2
+	 addc	$a1,$Xhi,$Xhi
+	shd	$Xnlo,$Xnhi,$sigma0[1],$t3
+	xor	$t2,$t0,$t0
+	shd	$Xnhi,$Xnlo,$sigma0[2],$t2
+	xor	$t3,$t1,$t1
+	extru	$Xnhi,`31-$sigma0[2]`,`32-$sigma0[2]`,$t3
+	xor	$t2,$t0,$t0
+	 shd	$a3,$a2,$sigma1[0],$a0
+	xor	$t3,$t1,$t1		; sigma0(X[i+1)&0x0f])
+	 shd	$a2,$a3,$sigma1[0],$a1
+	add	$t0,$Xlo,$Xlo
+	 shd	$a3,$a2,$sigma1[1],$t2
+	addc	$t1,$Xhi,$Xhi
+	 shd	$a2,$a3,$sigma1[1],$t3
+	xor	$t2,$a0,$a0
+	shd	$a3,$a2,$sigma1[2],$t2
+	xor	$t3,$a1,$a1
+	extru	$a3,`31-$sigma1[2]`,`32-$sigma1[2]`,$t3
+	xor	$t2,$a0,$a0
+	xor	$t3,$a1,$a1		; sigma0(X[i+14)&0x0f])
+	add	$a0,$Xlo,$Xlo
+	addc	$a1,$Xhi,$Xhi
+
+	stw	$Xhi,`-$XOFF+8*($i%16)`(%sp)
+	stw	$Xlo,`-$XOFF+8*($i%16)+4`(%sp)
+___
+&ROUND_00_15_pa1($i,@_,1);
+}
+$code.=<<___;
+	ldw	`0*4`($ctx),$Ahi		; load context
+	ldw	`1*4`($ctx),$Alo
+	ldw	`2*4`($ctx),$Bhi
+	ldw	`3*4`($ctx),$Blo
+	ldw	`4*4`($ctx),$Chi
+	ldw	`5*4`($ctx),$Clo
+	ldw	`6*4`($ctx),$Dhi
+	ldw	`7*4`($ctx),$Dlo
+	ldw	`8*4`($ctx),$Ehi
+	ldw	`9*4`($ctx),$Elo
+	ldw	`10*4`($ctx),$Fhi
+	ldw	`11*4`($ctx),$Flo
+	ldw	`12*4`($ctx),$Ghi
+	ldw	`13*4`($ctx),$Glo
+	ldw	`14*4`($ctx),$Hhi
+	ldw	`15*4`($ctx),$Hlo
+
+	extru	$inp,31,2,$t0
+	sh3addl	$t0,%r0,$t0
+	subi	32,$t0,$t0
+	mtctl	$t0,%cr11		; load %sar with align factor
+
+L\$oop_pa1
+	extru	$inp,31,2,$a3
+	comib,=	0,$a3,L\$aligned_pa1
+	sub	$inp,$a3,$inp
+
+	ldw	`0*4`($inp),$X[0]
+	ldw	`1*4`($inp),$X[1]
+	ldw	`2*4`($inp),$t2
+	ldw	`3*4`($inp),$t3
+	ldw	`4*4`($inp),$a0
+	ldw	`5*4`($inp),$a1
+	ldw	`6*4`($inp),$a2
+	ldw	`7*4`($inp),$a3
+	vshd	$X[0],$X[1],$X[0]
+	vshd	$X[1],$t2,$X[1]
+	stw	$X[0],`-$XOFF+0*4`(%sp)
+	ldw	`8*4`($inp),$t0
+	vshd	$t2,$t3,$t2
+	stw	$X[1],`-$XOFF+1*4`(%sp)
+	ldw	`9*4`($inp),$t1
+	vshd	$t3,$a0,$t3
+___
+{
+my @t=($t2,$t3,$a0,$a1,$a2,$a3,$t0,$t1);
+for ($i=2;$i<=(128/4-8);$i++) {
+$code.=<<___;
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+	ldw	`(8+$i)*4`($inp),$t[0]
+	vshd	$t[1],$t[2],$t[1]
+___
+push(@t,shift(@t));
+}
+for (;$i<(128/4-1);$i++) {
+$code.=<<___;
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+	vshd	$t[1],$t[2],$t[1]
+___
+push(@t,shift(@t));
+}
+$code.=<<___;
+	b	L\$collected_pa1
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+
+___
+}
+$code.=<<___;
+L\$aligned_pa1
+	ldw	`0*4`($inp),$X[0]
+	ldw	`1*4`($inp),$X[1]
+	ldw	`2*4`($inp),$t2
+	ldw	`3*4`($inp),$t3
+	ldw	`4*4`($inp),$a0
+	ldw	`5*4`($inp),$a1
+	ldw	`6*4`($inp),$a2
+	ldw	`7*4`($inp),$a3
+	stw	$X[0],`-$XOFF+0*4`(%sp)
+	ldw	`8*4`($inp),$t0
+	stw	$X[1],`-$XOFF+1*4`(%sp)
+	ldw	`9*4`($inp),$t1
+___
+{
+my @t=($t2,$t3,$a0,$a1,$a2,$a3,$t0,$t1);
+for ($i=2;$i<(128/4-8);$i++) {
+$code.=<<___;
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+	ldw	`(8+$i)*4`($inp),$t[0]
+___
+push(@t,shift(@t));
+}
+for (;$i<128/4;$i++) {
+$code.=<<___;
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+___
+push(@t,shift(@t));
+}
+$code.="L\$collected_pa1\n";
+}
+
+for($i=0;$i<16;$i++)	{ &ROUND_00_15_pa1($i,@V); unshift(@V,pop(@V)); unshift(@V,pop(@V)); }
+$code.="L\$rounds_pa1\n";
+for(;$i<32;$i++)	{ &ROUND_16_xx_pa1($i,@V); unshift(@V,pop(@V)); unshift(@V,pop(@V)); }
+
+$code.=<<___;
+	$POP	`-$FRAME_MARKER-2*$SIZE_T`(%sp),$ctx	; restore arguments
+	$POP	`-$FRAME_MARKER-3*$SIZE_T`(%sp),$inp
+	$POP	`-$FRAME_MARKER-4*$SIZE_T`(%sp),$num
+	ldo	`-$rounds*$SZ`($Tbl),$Tbl		; rewind $Tbl
+
+	ldw	`0*4`($ctx),$t1		; update context
+	ldw	`1*4`($ctx),$t0
+	ldw	`2*4`($ctx),$t3
+	ldw	`3*4`($ctx),$t2
+	ldw	`4*4`($ctx),$a1
+	ldw	`5*4`($ctx),$a0
+	ldw	`6*4`($ctx),$a3
+	add	$t0,$Alo,$Alo
+	ldw	`7*4`($ctx),$a2
+	addc	$t1,$Ahi,$Ahi
+	ldw	`8*4`($ctx),$t1
+	add	$t2,$Blo,$Blo
+	ldw	`9*4`($ctx),$t0
+	addc	$t3,$Bhi,$Bhi
+	ldw	`10*4`($ctx),$t3
+	add	$a0,$Clo,$Clo
+	ldw	`11*4`($ctx),$t2
+	addc	$a1,$Chi,$Chi
+	ldw	`12*4`($ctx),$a1
+	add	$a2,$Dlo,$Dlo
+	ldw	`13*4`($ctx),$a0
+	addc	$a3,$Dhi,$Dhi
+	ldw	`14*4`($ctx),$a3
+	add	$t0,$Elo,$Elo
+	ldw	`15*4`($ctx),$a2
+	addc	$t1,$Ehi,$Ehi
+	stw	$Ahi,`0*4`($ctx)
+	add	$t2,$Flo,$Flo
+	stw	$Alo,`1*4`($ctx)
+	addc	$t3,$Fhi,$Fhi
+	stw	$Bhi,`2*4`($ctx)
+	add	$a0,$Glo,$Glo
+	stw	$Blo,`3*4`($ctx)
+	addc	$a1,$Ghi,$Ghi
+	stw	$Chi,`4*4`($ctx)
+	add	$a2,$Hlo,$Hlo
+	stw	$Clo,`5*4`($ctx)
+	addc	$a3,$Hhi,$Hhi
+	stw	$Dhi,`6*4`($ctx)
+	ldo	`16*$SZ`($inp),$inp	; advance $inp
+	stw	$Dlo,`7*4`($ctx)
+	stw	$Ehi,`8*4`($ctx)
+	stw	$Elo,`9*4`($ctx)
+	stw	$Fhi,`10*4`($ctx)
+	stw	$Flo,`11*4`($ctx)
+	stw	$Ghi,`12*4`($ctx)
+	stw	$Glo,`13*4`($ctx)
+	stw	$Hhi,`14*4`($ctx)
+	comb,=	$inp,$num,L\$done
+	stw	$Hlo,`15*4`($ctx)
+	b	L\$oop_pa1
+	$PUSH	$inp,`-$FRAME_MARKER-3*$SIZE_T`(%sp)	; save $inp
+L\$done
+___
+}}
+$code.=<<___;
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2		; standard epilogue
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
+	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
+	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
+	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
+	$POP	`-$FRAME+8*$SIZE_T`(%sp),%r11
+	$POP	`-$FRAME+9*$SIZE_T`(%sp),%r12
+	$POP	`-$FRAME+10*$SIZE_T`(%sp),%r13
+	$POP	`-$FRAME+11*$SIZE_T`(%sp),%r14
+	$POP	`-$FRAME+12*$SIZE_T`(%sp),%r15
+	$POP	`-$FRAME+13*$SIZE_T`(%sp),%r16
+	$POP	`-$FRAME+14*$SIZE_T`(%sp),%r17
+	$POP	`-$FRAME+15*$SIZE_T`(%sp),%r18
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+	.STRINGZ "SHA`64*$SZ` block transform for PA-RISC, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+# Explicitly encode PA-RISC 2.0 instructions used in this module, so
+# that it can be compiled with .LEVEL 1.0. It should be noted that I
+# wouldn't have to do this, if GNU assembler understood .ALLOW 2.0
+# directive...
+
+my $ldd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "ldd$mod\t$args";
+
+    if ($args =~ /(\-?[0-9]+)\(%r([0-9]+)\),%r([0-9]+)/) # format 3 suffices
+    {	my $opcode=(0x14<<26)|($2<<21)|($3<<16)|(($1&0x1FF8)<<1)|(($1>>13)&1);
+	$opcode|=(1<<3) if ($mod =~ /^,m/);
+	$opcode|=(1<<2) if ($mod =~ /^,mb/);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $std = sub {
+  my ($mod,$args) = @_;
+  my $orig = "std$mod\t$args";
+
+    if ($args =~ /%r([0-9]+),(\-?[0-9]+)\(%r([0-9]+)\)/) # format 3 suffices
+    {	my $opcode=(0x1c<<26)|($3<<21)|($1<<16)|(($2&0x1FF8)<<1)|(($2>>13)&1);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $extrd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "extrd$mod\t$args";
+
+    # I only have ",u" completer, it's implicitly encoded...
+    if ($args =~ /%r([0-9]+),([0-9]+),([0-9]+),%r([0-9]+)/)	# format 15
+    {	my $opcode=(0x36<<26)|($1<<21)|($4<<16);
+	my $len=32-$3;
+	$opcode |= (($2&0x20)<<6)|(($2&0x1f)<<5);		# encode pos
+	$opcode |= (($len&0x20)<<7)|($len&0x1f);		# encode len
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    elsif ($args =~ /%r([0-9]+),%sar,([0-9]+),%r([0-9]+)/)	# format 12
+    {	my $opcode=(0x34<<26)|($1<<21)|($3<<16)|(2<<11)|(1<<9);
+	my $len=32-$2;
+	$opcode |= (($len&0x20)<<3)|($len&0x1f);		# encode len
+	$opcode |= (1<<13) if ($mod =~ /,\**=/);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $shrpd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "shrpd$mod\t$args";
+
+    if ($args =~ /%r([0-9]+),%r([0-9]+),([0-9]+),%r([0-9]+)/)	# format 14
+    {	my $opcode=(0x34<<26)|($2<<21)|($1<<16)|(1<<10)|$4;
+	my $cpos=63-$3;
+	$opcode |= (($cpos&0x20)<<6)|(($cpos&0x1f)<<5);		# encode sa
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    elsif ($args =~ /%r([0-9]+),%r([0-9]+),%sar,%r([0-9]+)/)	# format 11
+    {	sprintf "\t.WORD\t0x%08x\t; %s",
+		(0x34<<26)|($2<<21)|($1<<16)|(1<<9)|$3,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+sub assemble {
+  my ($mnemonic,$mod,$args)=@_;
+  my $opcode = eval("\$$mnemonic");
+
+    ref($opcode) eq 'CODE' ? &$opcode($mod,$args) : "\t$mnemonic$mod\t$args";
+}
+
+foreach (split("\n",$code)) {
+	s/\`([^\`]*)\`/eval $1/ge;
+
+	s/shd\s+(%r[0-9]+),(%r[0-9]+),([0-9]+)/
+		$3>31 ? sprintf("shd\t%$2,%$1,%d",$3-32)	# rotation for >=32
+		:       sprintf("shd\t%$1,%$2,%d",$3)/e			or
+	# translate made up instructons: _ror, _shr, _align, _shl
+	s/_ror(\s+)(%r[0-9]+),/
+		($SZ==4 ? "shd" : "shrpd")."$1$2,$2,"/e			or
+
+	s/_shr(\s+%r[0-9]+),([0-9]+),/
+		$SZ==4 ? sprintf("extru%s,%d,%d,",$1,31-$2,32-$2)
+		:        sprintf("extrd,u%s,%d,%d,",$1,63-$2,64-$2)/e	or
+
+	s/_align(\s+%r[0-9]+,%r[0-9]+),/
+		($SZ==4 ? "vshd$1," : "shrpd$1,%sar,")/e		or
+
+	s/_shl(\s+%r[0-9]+),([0-9]+),/
+		$SIZE_T==4 ? sprintf("zdep%s,%d,%d,",$1,31-$2,32-$2)
+		:            sprintf("depd,z%s,%d,%d,",$1,63-$2,64-$2)/e;
+
+	s/^\s+([a-z]+)([\S]*)\s+([\S]*)/&assemble($1,$2,$3)/e if ($SIZE_T==4);
+
+	s/cmpb,\*/comb,/ if ($SIZE_T==4);
+
+	print $_,"\n";
+}
+
+close STDOUT;
diff --git a/crypto/sha/asm/sha512-ppc.pl b/crypto/sha/asm/sha512-ppc.pl
new file mode 100755
index 000000000000..6b44a68e599e
--- /dev/null
+++ b/crypto/sha/asm/sha512-ppc.pl
@@ -0,0 +1,460 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# I let hardware handle unaligned input, except on page boundaries
+# (see below for details). Otherwise straightforward implementation
+# with X vector in register bank. The module is big-endian [which is
+# not big deal as there're no little-endian targets left around].
+
+#			sha256		|	sha512
+# 			-m64	-m32	|	-m64	-m32
+# --------------------------------------+-----------------------
+# PPC970,gcc-4.0.0	+50%	+38%	|	+40%	+410%(*)
+# Power6,xlc-7		+150%	+90%	|	+100%	+430%(*)
+#
+# (*)	64-bit code in 32-bit application context, which actually is
+#	on TODO list. It should be noted that for safe deployment in
+#	32-bit *mutli-threaded* context asyncronous signals should be
+#	blocked upon entry to SHA512 block routine. This is because
+#	32-bit signaling procedure invalidates upper halves of GPRs.
+#	Context switch procedure preserves them, but not signaling:-(
+
+# Second version is true multi-thread safe. Trouble with the original
+# version was that it was using thread local storage pointer register.
+# Well, it scrupulously preserved it, but the problem would arise the
+# moment asynchronous signal was delivered and signal handler would
+# dereference the TLS pointer. While it's never the case in openssl
+# application or test suite, we have to respect this scenario and not
+# use TLS pointer register. Alternative would be to require caller to
+# block signals prior calling this routine. For the record, in 32-bit
+# context R2 serves as TLS pointer, while in 64-bit context - R13.
+
+$flavour=shift;
+$output =shift;
+
+if ($flavour =~ /64/) {
+	$SIZE_T=8;
+	$LRSAVE=2*$SIZE_T;
+	$STU="stdu";
+	$UCMP="cmpld";
+	$SHL="sldi";
+	$POP="ld";
+	$PUSH="std";
+} elsif ($flavour =~ /32/) {
+	$SIZE_T=4;
+	$LRSAVE=$SIZE_T;
+	$STU="stwu";
+	$UCMP="cmplw";
+	$SHL="slwi";
+	$POP="lwz";
+	$PUSH="stw";
+} else { die "nonsense $flavour"; }
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output" || die "can't call $xlate: $!";
+
+if ($output =~ /512/) {
+	$func="sha512_block_data_order";
+	$SZ=8;
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=(1,  8, 7);
+	@sigma1=(19,61, 6);
+	$rounds=80;
+	$LD="ld";
+	$ST="std";
+	$ROR="rotrdi";
+	$SHR="srdi";
+} else {
+	$func="sha256_block_data_order";
+	$SZ=4;
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 7,18, 3);
+	@sigma1=(17,19,10);
+	$rounds=64;
+	$LD="lwz";
+	$ST="stw";
+	$ROR="rotrwi";
+	$SHR="srwi";
+}
+
+$FRAME=32*$SIZE_T+16*$SZ;
+$LOCALS=6*$SIZE_T;
+
+$sp ="r1";
+$toc="r2";
+$ctx="r3";	# zapped by $a0
+$inp="r4";	# zapped by $a1
+$num="r5";	# zapped by $t0
+
+$T  ="r0";
+$a0 ="r3";
+$a1 ="r4";
+$t0 ="r5";
+$t1 ="r6";
+$Tbl="r7";
+
+$A  ="r8";
+$B  ="r9";
+$C  ="r10";
+$D  ="r11";
+$E  ="r12";
+$F  ="r13";	$F="r2" if ($SIZE_T==8);# reassigned to exempt TLS pointer
+$G  ="r14";
+$H  ="r15";
+
+@V=($A,$B,$C,$D,$E,$F,$G,$H);
+@X=("r16","r17","r18","r19","r20","r21","r22","r23",
+    "r24","r25","r26","r27","r28","r29","r30","r31");
+
+$inp="r31";	# reassigned $inp! aliases with @X[15]
+
+sub ROUND_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+$code.=<<___;
+	$LD	$T,`$i*$SZ`($Tbl)
+	$ROR	$a0,$e,$Sigma1[0]
+	$ROR	$a1,$e,$Sigma1[1]
+	and	$t0,$f,$e
+	andc	$t1,$g,$e
+	add	$T,$T,$h
+	xor	$a0,$a0,$a1
+	$ROR	$a1,$a1,`$Sigma1[2]-$Sigma1[1]`
+	or	$t0,$t0,$t1		; Ch(e,f,g)
+	add	$T,$T,@X[$i]
+	xor	$a0,$a0,$a1		; Sigma1(e)
+	add	$T,$T,$t0
+	add	$T,$T,$a0
+
+	$ROR	$a0,$a,$Sigma0[0]
+	$ROR	$a1,$a,$Sigma0[1]
+	and	$t0,$a,$b
+	and	$t1,$a,$c
+	xor	$a0,$a0,$a1
+	$ROR	$a1,$a1,`$Sigma0[2]-$Sigma0[1]`
+	xor	$t0,$t0,$t1
+	and	$t1,$b,$c
+	xor	$a0,$a0,$a1		; Sigma0(a)
+	add	$d,$d,$T
+	xor	$t0,$t0,$t1		; Maj(a,b,c)
+	add	$h,$T,$a0
+	add	$h,$h,$t0
+
+___
+}
+
+sub ROUND_16_xx {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+$i-=16;
+$code.=<<___;
+	$ROR	$a0,@X[($i+1)%16],$sigma0[0]
+	$ROR	$a1,@X[($i+1)%16],$sigma0[1]
+	$ROR	$t0,@X[($i+14)%16],$sigma1[0]
+	$ROR	$t1,@X[($i+14)%16],$sigma1[1]
+	xor	$a0,$a0,$a1
+	$SHR	$a1,@X[($i+1)%16],$sigma0[2]
+	xor	$t0,$t0,$t1
+	$SHR	$t1,@X[($i+14)%16],$sigma1[2]
+	add	@X[$i],@X[$i],@X[($i+9)%16]
+	xor	$a0,$a0,$a1		; sigma0(X[(i+1)&0x0f])
+	xor	$t0,$t0,$t1		; sigma1(X[(i+14)&0x0f])
+	add	@X[$i],@X[$i],$a0
+	add	@X[$i],@X[$i],$t0
+___
+&ROUND_00_15($i,$a,$b,$c,$d,$e,$f,$g,$h);
+}
+
+$code=<<___;
+.machine	"any"
+.text
+
+.globl	$func
+.align	6
+$func:
+	$STU	$sp,-$FRAME($sp)
+	mflr	r0
+	$SHL	$num,$num,`log(16*$SZ)/log(2)`
+
+	$PUSH	$ctx,`$FRAME-$SIZE_T*22`($sp)
+
+	$PUSH	$toc,`$FRAME-$SIZE_T*20`($sp)
+	$PUSH	r13,`$FRAME-$SIZE_T*19`($sp)
+	$PUSH	r14,`$FRAME-$SIZE_T*18`($sp)
+	$PUSH	r15,`$FRAME-$SIZE_T*17`($sp)
+	$PUSH	r16,`$FRAME-$SIZE_T*16`($sp)
+	$PUSH	r17,`$FRAME-$SIZE_T*15`($sp)
+	$PUSH	r18,`$FRAME-$SIZE_T*14`($sp)
+	$PUSH	r19,`$FRAME-$SIZE_T*13`($sp)
+	$PUSH	r20,`$FRAME-$SIZE_T*12`($sp)
+	$PUSH	r21,`$FRAME-$SIZE_T*11`($sp)
+	$PUSH	r22,`$FRAME-$SIZE_T*10`($sp)
+	$PUSH	r23,`$FRAME-$SIZE_T*9`($sp)
+	$PUSH	r24,`$FRAME-$SIZE_T*8`($sp)
+	$PUSH	r25,`$FRAME-$SIZE_T*7`($sp)
+	$PUSH	r26,`$FRAME-$SIZE_T*6`($sp)
+	$PUSH	r27,`$FRAME-$SIZE_T*5`($sp)
+	$PUSH	r28,`$FRAME-$SIZE_T*4`($sp)
+	$PUSH	r29,`$FRAME-$SIZE_T*3`($sp)
+	$PUSH	r30,`$FRAME-$SIZE_T*2`($sp)
+	$PUSH	r31,`$FRAME-$SIZE_T*1`($sp)
+	$PUSH	r0,`$FRAME+$LRSAVE`($sp)
+
+	$LD	$A,`0*$SZ`($ctx)
+	mr	$inp,r4				; incarnate $inp
+	$LD	$B,`1*$SZ`($ctx)
+	$LD	$C,`2*$SZ`($ctx)
+	$LD	$D,`3*$SZ`($ctx)
+	$LD	$E,`4*$SZ`($ctx)
+	$LD	$F,`5*$SZ`($ctx)
+	$LD	$G,`6*$SZ`($ctx)
+	$LD	$H,`7*$SZ`($ctx)
+
+	bl	LPICmeup
+LPICedup:
+	andi.	r0,$inp,3
+	bne	Lunaligned
+Laligned:
+	add	$num,$inp,$num
+	$PUSH	$num,`$FRAME-$SIZE_T*24`($sp)	; end pointer
+	$PUSH	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
+	bl	Lsha2_block_private
+	b	Ldone
+
+; PowerPC specification allows an implementation to be ill-behaved
+; upon unaligned access which crosses page boundary. "Better safe
+; than sorry" principle makes me treat it specially. But I don't
+; look for particular offending word, but rather for the input
+; block which crosses the boundary. Once found that block is aligned
+; and hashed separately...
+.align	4
+Lunaligned:
+	subfic	$t1,$inp,4096
+	andi.	$t1,$t1,`4096-16*$SZ`	; distance to closest page boundary
+	beq	Lcross_page
+	$UCMP	$num,$t1
+	ble-	Laligned		; didn't cross the page boundary
+	subfc	$num,$t1,$num
+	add	$t1,$inp,$t1
+	$PUSH	$num,`$FRAME-$SIZE_T*25`($sp)	; save real remaining num
+	$PUSH	$t1,`$FRAME-$SIZE_T*24`($sp)	; intermediate end pointer
+	$PUSH	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
+	bl	Lsha2_block_private
+	; $inp equals to the intermediate end pointer here
+	$POP	$num,`$FRAME-$SIZE_T*25`($sp)	; restore real remaining num
+Lcross_page:
+	li	$t1,`16*$SZ/4`
+	mtctr	$t1
+	addi	r20,$sp,$LOCALS			; aligned spot below the frame
+Lmemcpy:
+	lbz	r16,0($inp)
+	lbz	r17,1($inp)
+	lbz	r18,2($inp)
+	lbz	r19,3($inp)
+	addi	$inp,$inp,4
+	stb	r16,0(r20)
+	stb	r17,1(r20)
+	stb	r18,2(r20)
+	stb	r19,3(r20)
+	addi	r20,r20,4
+	bdnz	Lmemcpy
+
+	$PUSH	$inp,`$FRAME-$SIZE_T*26`($sp)	; save real inp
+	addi	$t1,$sp,`$LOCALS+16*$SZ`	; fictitious end pointer
+	addi	$inp,$sp,$LOCALS		; fictitious inp pointer
+	$PUSH	$num,`$FRAME-$SIZE_T*25`($sp)	; save real num
+	$PUSH	$t1,`$FRAME-$SIZE_T*24`($sp)	; end pointer
+	$PUSH	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
+	bl	Lsha2_block_private
+	$POP	$inp,`$FRAME-$SIZE_T*26`($sp)	; restore real inp
+	$POP	$num,`$FRAME-$SIZE_T*25`($sp)	; restore real num
+	addic.	$num,$num,`-16*$SZ`		; num--
+	bne-	Lunaligned
+
+Ldone:
+	$POP	r0,`$FRAME+$LRSAVE`($sp)
+	$POP	$toc,`$FRAME-$SIZE_T*20`($sp)
+	$POP	r13,`$FRAME-$SIZE_T*19`($sp)
+	$POP	r14,`$FRAME-$SIZE_T*18`($sp)
+	$POP	r15,`$FRAME-$SIZE_T*17`($sp)
+	$POP	r16,`$FRAME-$SIZE_T*16`($sp)
+	$POP	r17,`$FRAME-$SIZE_T*15`($sp)
+	$POP	r18,`$FRAME-$SIZE_T*14`($sp)
+	$POP	r19,`$FRAME-$SIZE_T*13`($sp)
+	$POP	r20,`$FRAME-$SIZE_T*12`($sp)
+	$POP	r21,`$FRAME-$SIZE_T*11`($sp)
+	$POP	r22,`$FRAME-$SIZE_T*10`($sp)
+	$POP	r23,`$FRAME-$SIZE_T*9`($sp)
+	$POP	r24,`$FRAME-$SIZE_T*8`($sp)
+	$POP	r25,`$FRAME-$SIZE_T*7`($sp)
+	$POP	r26,`$FRAME-$SIZE_T*6`($sp)
+	$POP	r27,`$FRAME-$SIZE_T*5`($sp)
+	$POP	r28,`$FRAME-$SIZE_T*4`($sp)
+	$POP	r29,`$FRAME-$SIZE_T*3`($sp)
+	$POP	r30,`$FRAME-$SIZE_T*2`($sp)
+	$POP	r31,`$FRAME-$SIZE_T*1`($sp)
+	mtlr	r0
+	addi	$sp,$sp,$FRAME
+	blr
+	.long	0
+	.byte	0,12,4,1,0x80,18,3,0
+	.long	0
+
+.align	4
+Lsha2_block_private:
+___
+for($i=0;$i<16;$i++) {
+$code.=<<___ if ($SZ==4);
+	lwz	@X[$i],`$i*$SZ`($inp)
+___
+# 64-bit loads are split to 2x32-bit ones, as CPU can't handle
+# unaligned 64-bit loads, only 32-bit ones...
+$code.=<<___ if ($SZ==8);
+	lwz	$t0,`$i*$SZ`($inp)
+	lwz	@X[$i],`$i*$SZ+4`($inp)
+	insrdi	@X[$i],$t0,32,0
+___
+	&ROUND_00_15($i,@V);
+	unshift(@V,pop(@V));
+}
+$code.=<<___;
+	li	$T,`$rounds/16-1`
+	mtctr	$T
+.align	4
+Lrounds:
+	addi	$Tbl,$Tbl,`16*$SZ`
+___
+for(;$i<32;$i++) {
+	&ROUND_16_xx($i,@V);
+	unshift(@V,pop(@V));
+}
+$code.=<<___;
+	bdnz-	Lrounds
+
+	$POP	$ctx,`$FRAME-$SIZE_T*22`($sp)
+	$POP	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
+	$POP	$num,`$FRAME-$SIZE_T*24`($sp)	; end pointer
+	subi	$Tbl,$Tbl,`($rounds-16)*$SZ`	; rewind Tbl
+
+	$LD	r16,`0*$SZ`($ctx)
+	$LD	r17,`1*$SZ`($ctx)
+	$LD	r18,`2*$SZ`($ctx)
+	$LD	r19,`3*$SZ`($ctx)
+	$LD	r20,`4*$SZ`($ctx)
+	$LD	r21,`5*$SZ`($ctx)
+	$LD	r22,`6*$SZ`($ctx)
+	addi	$inp,$inp,`16*$SZ`		; advance inp
+	$LD	r23,`7*$SZ`($ctx)
+	add	$A,$A,r16
+	add	$B,$B,r17
+	$PUSH	$inp,`$FRAME-$SIZE_T*23`($sp)
+	add	$C,$C,r18
+	$ST	$A,`0*$SZ`($ctx)
+	add	$D,$D,r19
+	$ST	$B,`1*$SZ`($ctx)
+	add	$E,$E,r20
+	$ST	$C,`2*$SZ`($ctx)
+	add	$F,$F,r21
+	$ST	$D,`3*$SZ`($ctx)
+	add	$G,$G,r22
+	$ST	$E,`4*$SZ`($ctx)
+	add	$H,$H,r23
+	$ST	$F,`5*$SZ`($ctx)
+	$ST	$G,`6*$SZ`($ctx)
+	$UCMP	$inp,$num
+	$ST	$H,`7*$SZ`($ctx)
+	bne	Lsha2_block_private
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+___
+
+# Ugly hack here, because PPC assembler syntax seem to vary too
+# much from platforms to platform...
+$code.=<<___;
+.align	6
+LPICmeup:
+	mflr	r0
+	bcl	20,31,\$+4
+	mflr	$Tbl	; vvvvvv "distance" between . and 1st data entry
+	addi	$Tbl,$Tbl,`64-8`
+	mtlr	r0
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+	.space	`64-9*4`
+___
+$code.=<<___ if ($SZ==8);
+	.long	0x428a2f98,0xd728ae22,0x71374491,0x23ef65cd
+	.long	0xb5c0fbcf,0xec4d3b2f,0xe9b5dba5,0x8189dbbc
+	.long	0x3956c25b,0xf348b538,0x59f111f1,0xb605d019
+	.long	0x923f82a4,0xaf194f9b,0xab1c5ed5,0xda6d8118
+	.long	0xd807aa98,0xa3030242,0x12835b01,0x45706fbe
+	.long	0x243185be,0x4ee4b28c,0x550c7dc3,0xd5ffb4e2
+	.long	0x72be5d74,0xf27b896f,0x80deb1fe,0x3b1696b1
+	.long	0x9bdc06a7,0x25c71235,0xc19bf174,0xcf692694
+	.long	0xe49b69c1,0x9ef14ad2,0xefbe4786,0x384f25e3
+	.long	0x0fc19dc6,0x8b8cd5b5,0x240ca1cc,0x77ac9c65
+	.long	0x2de92c6f,0x592b0275,0x4a7484aa,0x6ea6e483
+	.long	0x5cb0a9dc,0xbd41fbd4,0x76f988da,0x831153b5
+	.long	0x983e5152,0xee66dfab,0xa831c66d,0x2db43210
+	.long	0xb00327c8,0x98fb213f,0xbf597fc7,0xbeef0ee4
+	.long	0xc6e00bf3,0x3da88fc2,0xd5a79147,0x930aa725
+	.long	0x06ca6351,0xe003826f,0x14292967,0x0a0e6e70
+	.long	0x27b70a85,0x46d22ffc,0x2e1b2138,0x5c26c926
+	.long	0x4d2c6dfc,0x5ac42aed,0x53380d13,0x9d95b3df
+	.long	0x650a7354,0x8baf63de,0x766a0abb,0x3c77b2a8
+	.long	0x81c2c92e,0x47edaee6,0x92722c85,0x1482353b
+	.long	0xa2bfe8a1,0x4cf10364,0xa81a664b,0xbc423001
+	.long	0xc24b8b70,0xd0f89791,0xc76c51a3,0x0654be30
+	.long	0xd192e819,0xd6ef5218,0xd6990624,0x5565a910
+	.long	0xf40e3585,0x5771202a,0x106aa070,0x32bbd1b8
+	.long	0x19a4c116,0xb8d2d0c8,0x1e376c08,0x5141ab53
+	.long	0x2748774c,0xdf8eeb99,0x34b0bcb5,0xe19b48a8
+	.long	0x391c0cb3,0xc5c95a63,0x4ed8aa4a,0xe3418acb
+	.long	0x5b9cca4f,0x7763e373,0x682e6ff3,0xd6b2b8a3
+	.long	0x748f82ee,0x5defb2fc,0x78a5636f,0x43172f60
+	.long	0x84c87814,0xa1f0ab72,0x8cc70208,0x1a6439ec
+	.long	0x90befffa,0x23631e28,0xa4506ceb,0xde82bde9
+	.long	0xbef9a3f7,0xb2c67915,0xc67178f2,0xe372532b
+	.long	0xca273ece,0xea26619c,0xd186b8c7,0x21c0c207
+	.long	0xeada7dd6,0xcde0eb1e,0xf57d4f7f,0xee6ed178
+	.long	0x06f067aa,0x72176fba,0x0a637dc5,0xa2c898a6
+	.long	0x113f9804,0xbef90dae,0x1b710b35,0x131c471b
+	.long	0x28db77f5,0x23047d84,0x32caab7b,0x40c72493
+	.long	0x3c9ebe0a,0x15c9bebc,0x431d67c4,0x9c100d4c
+	.long	0x4cc5d4be,0xcb3e42b6,0x597f299c,0xfc657e2a
+	.long	0x5fcb6fab,0x3ad6faec,0x6c44198c,0x4a475817
+___
+$code.=<<___ if ($SZ==4);
+	.long	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.long	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.long	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.long	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.long	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.long	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.long	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.long	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.long	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.long	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.long	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.long	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.long	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.long	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.long	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.long	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha512-s390x.pl b/crypto/sha/asm/sha512-s390x.pl
new file mode 100755
index 000000000000..079a3fc78ab4
--- /dev/null
+++ b/crypto/sha/asm/sha512-s390x.pl
@@ -0,0 +1,322 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA256/512 block procedures for s390x.
+
+# April 2007.
+#
+# sha256_block_data_order is reportedly >3 times faster than gcc 3.3
+# generated code (must be a bug in compiler, as improvement is
+# "pathologically" high, in particular in comparison to other SHA
+# modules). But the real twist is that it detects if hardware support
+# for SHA256 is available and in such case utilizes it. Then the
+# performance can reach >6.5x of assembler one for larger chunks.
+#
+# sha512_block_data_order is ~70% faster than gcc 3.3 generated code.
+
+# January 2009.
+#
+# Add support for hardware SHA512 and reschedule instructions to
+# favour dual-issue z10 pipeline. Hardware SHA256/512 is ~4.7x faster
+# than software.
+
+# November 2010.
+#
+# Adapt for -m31 build. If kernel supports what's called "highgprs"
+# feature on Linux [see /proc/cpuinfo], it's possible to use 64-bit
+# instructions and achieve "64-bit" performance even in 31-bit legacy
+# application context. The feature is not specific to any particular
+# processor, as long as it's "z-CPU". Latter implies that the code
+# remains z/Architecture specific. On z900 SHA256 was measured to
+# perform 2.4x and SHA512 - 13x better than code generated by gcc 4.3.
+
+$flavour = shift;
+
+if ($flavour =~ /3[12]/) {
+	$SIZE_T=4;
+	$g="";
+} else {
+	$SIZE_T=8;
+	$g="g";
+}
+
+$t0="%r0";
+$t1="%r1";
+$ctx="%r2";	$t2="%r2";
+$inp="%r3";
+$len="%r4";	# used as index in inner loop
+
+$A="%r5";
+$B="%r6";
+$C="%r7";
+$D="%r8";
+$E="%r9";
+$F="%r10";
+$G="%r11";
+$H="%r12";	@V=($A,$B,$C,$D,$E,$F,$G,$H);
+$tbl="%r13";
+$T1="%r14";
+$sp="%r15";
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+if ($output =~ /512/) {
+	$label="512";
+	$SZ=8;
+	$LD="lg";	# load from memory
+	$ST="stg";	# store to memory
+	$ADD="alg";	# add with memory operand
+	$ROT="rllg";	# rotate left
+	$SHR="srlg";	# logical right shift [see even at the end]
+	@Sigma0=(25,30,36);
+	@Sigma1=(23,46,50);
+	@sigma0=(56,63, 7);
+	@sigma1=( 3,45, 6);
+	$rounds=80;
+	$kimdfunc=3;	# 0 means unknown/unsupported/unimplemented/disabled
+} else {
+	$label="256";
+	$SZ=4;
+	$LD="llgf";	# load from memory
+	$ST="st";	# store to memory
+	$ADD="al";	# add with memory operand
+	$ROT="rll";	# rotate left
+	$SHR="srl";	# logical right shift
+	@Sigma0=(10,19,30);
+	@Sigma1=( 7,21,26);
+	@sigma0=(14,25, 3);
+	@sigma1=(13,15,10);
+	$rounds=64;
+	$kimdfunc=2;	# magic function code for kimd instruction
+}
+$Func="sha${label}_block_data_order";
+$Table="K${label}";
+$stdframe=16*$SIZE_T+4*8;
+$frame=$stdframe+16*$SZ;
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___ if ($i<16);
+	$LD	$T1,`$i*$SZ`($inp)	### $i
+___
+$code.=<<___;
+	$ROT	$t0,$e,$Sigma1[0]
+	$ROT	$t1,$e,$Sigma1[1]
+	 lgr	$t2,$f
+	xgr	$t0,$t1
+	$ROT	$t1,$t1,`$Sigma1[2]-$Sigma1[1]`
+	 xgr	$t2,$g
+	$ST	$T1,`$stdframe+$SZ*($i%16)`($sp)
+	xgr	$t0,$t1			# Sigma1(e)
+	algr	$T1,$h			# T1+=h
+	 ngr	$t2,$e
+	 lgr	$t1,$a
+	algr	$T1,$t0			# T1+=Sigma1(e)
+	$ROT	$h,$a,$Sigma0[0]
+	 xgr	$t2,$g			# Ch(e,f,g)
+	$ADD	$T1,`$i*$SZ`($len,$tbl)	# T1+=K[i]
+	$ROT	$t0,$a,$Sigma0[1]
+	algr	$T1,$t2			# T1+=Ch(e,f,g)
+	 ogr	$t1,$b
+	xgr	$h,$t0
+	 lgr	$t2,$a
+	 ngr	$t1,$c
+	$ROT	$t0,$t0,`$Sigma0[2]-$Sigma0[1]`
+	xgr	$h,$t0			# h=Sigma0(a)
+	 ngr	$t2,$b
+	algr	$h,$T1			# h+=T1
+	 ogr	$t2,$t1			# Maj(a,b,c)
+	algr	$d,$T1			# d+=T1
+	algr	$h,$t2			# h+=Maj(a,b,c)
+___
+}
+
+sub BODY_16_XX {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	$LD	$T1,`$stdframe+$SZ*(($i+1)%16)`($sp)	### $i
+	$LD	$t1,`$stdframe+$SZ*(($i+14)%16)`($sp)
+	$ROT	$t0,$T1,$sigma0[0]
+	$SHR	$T1,$sigma0[2]
+	$ROT	$t2,$t0,`$sigma0[1]-$sigma0[0]`
+	xgr	$T1,$t0
+	$ROT	$t0,$t1,$sigma1[0]
+	xgr	$T1,$t2					# sigma0(X[i+1])
+	$SHR	$t1,$sigma1[2]
+	$ADD	$T1,`$stdframe+$SZ*($i%16)`($sp)	# +=X[i]
+	xgr	$t1,$t0
+	$ROT	$t0,$t0,`$sigma1[1]-$sigma1[0]`
+	$ADD	$T1,`$stdframe+$SZ*(($i+9)%16)`($sp)	# +=X[i+9]
+	xgr	$t1,$t0				# sigma1(X[i+14])
+	algr	$T1,$t1				# +=sigma1(X[i+14])
+___
+	&BODY_00_15(@_);
+}
+
+$code.=<<___;
+.text
+.align	64
+.type	$Table,\@object
+$Table:
+___
+$code.=<<___ if ($SZ==4);
+	.long	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.long	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.long	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.long	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.long	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.long	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.long	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.long	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.long	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.long	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.long	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.long	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.long	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.long	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.long	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.long	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+___
+$code.=<<___ if ($SZ==8);
+	.quad	0x428a2f98d728ae22,0x7137449123ef65cd
+	.quad	0xb5c0fbcfec4d3b2f,0xe9b5dba58189dbbc
+	.quad	0x3956c25bf348b538,0x59f111f1b605d019
+	.quad	0x923f82a4af194f9b,0xab1c5ed5da6d8118
+	.quad	0xd807aa98a3030242,0x12835b0145706fbe
+	.quad	0x243185be4ee4b28c,0x550c7dc3d5ffb4e2
+	.quad	0x72be5d74f27b896f,0x80deb1fe3b1696b1
+	.quad	0x9bdc06a725c71235,0xc19bf174cf692694
+	.quad	0xe49b69c19ef14ad2,0xefbe4786384f25e3
+	.quad	0x0fc19dc68b8cd5b5,0x240ca1cc77ac9c65
+	.quad	0x2de92c6f592b0275,0x4a7484aa6ea6e483
+	.quad	0x5cb0a9dcbd41fbd4,0x76f988da831153b5
+	.quad	0x983e5152ee66dfab,0xa831c66d2db43210
+	.quad	0xb00327c898fb213f,0xbf597fc7beef0ee4
+	.quad	0xc6e00bf33da88fc2,0xd5a79147930aa725
+	.quad	0x06ca6351e003826f,0x142929670a0e6e70
+	.quad	0x27b70a8546d22ffc,0x2e1b21385c26c926
+	.quad	0x4d2c6dfc5ac42aed,0x53380d139d95b3df
+	.quad	0x650a73548baf63de,0x766a0abb3c77b2a8
+	.quad	0x81c2c92e47edaee6,0x92722c851482353b
+	.quad	0xa2bfe8a14cf10364,0xa81a664bbc423001
+	.quad	0xc24b8b70d0f89791,0xc76c51a30654be30
+	.quad	0xd192e819d6ef5218,0xd69906245565a910
+	.quad	0xf40e35855771202a,0x106aa07032bbd1b8
+	.quad	0x19a4c116b8d2d0c8,0x1e376c085141ab53
+	.quad	0x2748774cdf8eeb99,0x34b0bcb5e19b48a8
+	.quad	0x391c0cb3c5c95a63,0x4ed8aa4ae3418acb
+	.quad	0x5b9cca4f7763e373,0x682e6ff3d6b2b8a3
+	.quad	0x748f82ee5defb2fc,0x78a5636f43172f60
+	.quad	0x84c87814a1f0ab72,0x8cc702081a6439ec
+	.quad	0x90befffa23631e28,0xa4506cebde82bde9
+	.quad	0xbef9a3f7b2c67915,0xc67178f2e372532b
+	.quad	0xca273eceea26619c,0xd186b8c721c0c207
+	.quad	0xeada7dd6cde0eb1e,0xf57d4f7fee6ed178
+	.quad	0x06f067aa72176fba,0x0a637dc5a2c898a6
+	.quad	0x113f9804bef90dae,0x1b710b35131c471b
+	.quad	0x28db77f523047d84,0x32caab7b40c72493
+	.quad	0x3c9ebe0a15c9bebc,0x431d67c49c100d4c
+	.quad	0x4cc5d4becb3e42b6,0x597f299cfc657e2a
+	.quad	0x5fcb6fab3ad6faec,0x6c44198c4a475817
+___
+$code.=<<___;
+.size	$Table,.-$Table
+.globl	$Func
+.type	$Func,\@function
+$Func:
+	sllg	$len,$len,`log(16*$SZ)/log(2)`
+___
+$code.=<<___ if ($kimdfunc);
+	larl	%r1,OPENSSL_s390xcap_P
+	lg	%r0,0(%r1)
+	tmhl	%r0,0x4000	# check for message-security assist
+	jz	.Lsoftware
+	lghi	%r0,0
+	la	%r1,`2*$SIZE_T`($sp)
+	.long	0xb93e0002	# kimd %r0,%r2
+	lg	%r0,`2*$SIZE_T`($sp)
+	tmhh	%r0,`0x8000>>$kimdfunc`
+	jz	.Lsoftware
+	lghi	%r0,$kimdfunc
+	lgr	%r1,$ctx
+	lgr	%r2,$inp
+	lgr	%r3,$len
+	.long	0xb93e0002	# kimd %r0,%r2
+	brc	1,.-4		# pay attention to "partial completion"
+	br	%r14
+.align	16
+.Lsoftware:
+___
+$code.=<<___;
+	lghi	%r1,-$frame
+	la	$len,0($len,$inp)
+	stm${g}	$ctx,%r15,`2*$SIZE_T`($sp)
+	lgr	%r0,$sp
+	la	$sp,0(%r1,$sp)
+	st${g}	%r0,0($sp)
+
+	larl	$tbl,$Table
+	$LD	$A,`0*$SZ`($ctx)
+	$LD	$B,`1*$SZ`($ctx)
+	$LD	$C,`2*$SZ`($ctx)
+	$LD	$D,`3*$SZ`($ctx)
+	$LD	$E,`4*$SZ`($ctx)
+	$LD	$F,`5*$SZ`($ctx)
+	$LD	$G,`6*$SZ`($ctx)
+	$LD	$H,`7*$SZ`($ctx)
+
+.Lloop:
+	lghi	$len,0
+___
+for ($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=".Lrounds_16_xx:\n";
+for (;$i<32;$i++)	{ &BODY_16_XX($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	aghi	$len,`16*$SZ`
+	lghi	$t0,`($rounds-16)*$SZ`
+	clgr	$len,$t0
+	jne	.Lrounds_16_xx
+
+	l${g}	$ctx,`$frame+2*$SIZE_T`($sp)
+	la	$inp,`16*$SZ`($inp)
+	$ADD	$A,`0*$SZ`($ctx)
+	$ADD	$B,`1*$SZ`($ctx)
+	$ADD	$C,`2*$SZ`($ctx)
+	$ADD	$D,`3*$SZ`($ctx)
+	$ADD	$E,`4*$SZ`($ctx)
+	$ADD	$F,`5*$SZ`($ctx)
+	$ADD	$G,`6*$SZ`($ctx)
+	$ADD	$H,`7*$SZ`($ctx)
+	$ST	$A,`0*$SZ`($ctx)
+	$ST	$B,`1*$SZ`($ctx)
+	$ST	$C,`2*$SZ`($ctx)
+	$ST	$D,`3*$SZ`($ctx)
+	$ST	$E,`4*$SZ`($ctx)
+	$ST	$F,`5*$SZ`($ctx)
+	$ST	$G,`6*$SZ`($ctx)
+	$ST	$H,`7*$SZ`($ctx)
+	cl${g}	$inp,`$frame+4*$SIZE_T`($sp)
+	jne	.Lloop
+
+	lm${g}	%r6,%r15,`$frame+6*$SIZE_T`($sp)	
+	br	%r14
+.size	$Func,.-$Func
+.string	"SHA${label} block transform for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+.comm	OPENSSL_s390xcap_P,16,8
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+# unlike 32-bit shift 64-bit one takes three arguments
+$code =~ s/(srlg\s+)(%r[0-9]+),/$1$2,$2,/gm;
+
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha512-sparcv9.pl b/crypto/sha/asm/sha512-sparcv9.pl
new file mode 100755
index 000000000000..585740789e63
--- /dev/null
+++ b/crypto/sha/asm/sha512-sparcv9.pl
@@ -0,0 +1,594 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA256 performance improvement over compiler generated code varies
+# from 40% for Sun C [32-bit build] to 70% for gcc [3.3, 64-bit
+# build]. Just like in SHA1 module I aim to ensure scalability on
+# UltraSPARC T1 by packing X[16] to 8 64-bit registers.
+
+# SHA512 on pre-T1 UltraSPARC.
+#
+# Performance is >75% better than 64-bit code generated by Sun C and
+# over 2x than 32-bit code. X[16] resides on stack, but access to it
+# is scheduled for L2 latency and staged through 32 least significant
+# bits of %l0-%l7. The latter is done to achieve 32-/64-bit ABI
+# duality. Nevetheless it's ~40% faster than SHA256, which is pretty
+# good [optimal coefficient is 50%].
+#
+# SHA512 on UltraSPARC T1.
+#
+# It's not any faster than 64-bit code generated by Sun C 5.8. This is
+# because 64-bit code generator has the advantage of using 64-bit
+# loads(*) to access X[16], which I consciously traded for 32-/64-bit
+# ABI duality [as per above]. But it surpasses 32-bit Sun C generated
+# code by 60%, not to mention that it doesn't suffer from severe decay
+# when running 4 times physical cores threads and that it leaves gcc
+# [3.4] behind by over 4x factor! If compared to SHA256, single thread
+# performance is only 10% better, but overall throughput for maximum
+# amount of threads for given CPU exceeds corresponding one of SHA256
+# by 30% [again, optimal coefficient is 50%].
+#
+# (*)	Unlike pre-T1 UltraSPARC loads on T1 are executed strictly
+#	in-order, i.e. load instruction has to complete prior next
+#	instruction in given thread is executed, even if the latter is
+#	not dependent on load result! This means that on T1 two 32-bit
+#	loads are always slower than one 64-bit load. Once again this
+#	is unlike pre-T1 UltraSPARC, where, if scheduled appropriately,
+#	2x32-bit loads can be as fast as 1x64-bit ones.
+
+$bits=32;
+for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+if ($bits==64)	{ $bias=2047; $frame=192; }
+else		{ $bias=0;    $frame=112; }
+
+$output=shift;
+open STDOUT,">$output";
+
+if ($output =~ /512/) {
+	$label="512";
+	$SZ=8;
+	$LD="ldx";		# load from memory
+	$ST="stx";		# store to memory
+	$SLL="sllx";		# shift left logical
+	$SRL="srlx";		# shift right logical
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=( 7, 1, 8);	# right shift first
+	@sigma1=( 6,19,61);	# right shift first
+	$lastK=0x817;
+	$rounds=80;
+	$align=4;
+
+	$locals=16*$SZ;		# X[16]
+
+	$A="%o0";
+	$B="%o1";
+	$C="%o2";
+	$D="%o3";
+	$E="%o4";
+	$F="%o5";
+	$G="%g1";
+	$H="%o7";
+	@V=($A,$B,$C,$D,$E,$F,$G,$H);
+} else {
+	$label="256";
+	$SZ=4;
+	$LD="ld";		# load from memory
+	$ST="st";		# store to memory
+	$SLL="sll";		# shift left logical
+	$SRL="srl";		# shift right logical
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 3, 7,18);	# right shift first
+	@sigma1=(10,17,19);	# right shift first
+	$lastK=0x8f2;
+	$rounds=64;
+	$align=8;
+
+	$locals=0;		# X[16] is register resident
+	@X=("%o0","%o1","%o2","%o3","%o4","%o5","%g1","%o7");
+	
+	$A="%l0";
+	$B="%l1";
+	$C="%l2";
+	$D="%l3";
+	$E="%l4";
+	$F="%l5";
+	$G="%l6";
+	$H="%l7";
+	@V=($A,$B,$C,$D,$E,$F,$G,$H);
+}
+$T1="%g2";
+$tmp0="%g3";
+$tmp1="%g4";
+$tmp2="%g5";
+
+$ctx="%i0";
+$inp="%i1";
+$len="%i2";
+$Ktbl="%i3";
+$tmp31="%i4";
+$tmp32="%i5";
+
+########### SHA256
+$Xload = sub {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+
+    if ($i==0) {
+$code.=<<___;
+	ldx	[$inp+0],@X[0]
+	ldx	[$inp+16],@X[2]
+	ldx	[$inp+32],@X[4]
+	ldx	[$inp+48],@X[6]
+	ldx	[$inp+8],@X[1]
+	ldx	[$inp+24],@X[3]
+	subcc	%g0,$tmp31,$tmp32 ! should be 64-$tmp31, but -$tmp31 works too
+	ldx	[$inp+40],@X[5]
+	bz,pt	%icc,.Laligned
+	ldx	[$inp+56],@X[7]
+
+	sllx	@X[0],$tmp31,@X[0]
+	ldx	[$inp+64],$T1
+___
+for($j=0;$j<7;$j++)
+{   $code.=<<___;
+	srlx	@X[$j+1],$tmp32,$tmp1
+	sllx	@X[$j+1],$tmp31,@X[$j+1]
+	or	$tmp1,@X[$j],@X[$j]
+___
+}
+$code.=<<___;
+	srlx	$T1,$tmp32,$T1
+	or	$T1,@X[7],@X[7]
+.Laligned:
+___
+    }
+
+    if ($i&1) {
+	$code.="\tadd	@X[$i/2],$h,$T1\n";
+    } else {
+	$code.="\tsrlx	@X[$i/2],32,$T1\n\tadd	$h,$T1,$T1\n";
+    }
+} if ($SZ==4);
+
+########### SHA512
+$Xload = sub {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+my @pair=("%l".eval(($i*2)%8),"%l".eval(($i*2)%8+1),"%l".eval((($i+1)*2)%8));
+
+$code.=<<___ if ($i==0);
+	ld	[$inp+0],%l0
+	ld	[$inp+4],%l1
+	ld	[$inp+8],%l2
+	ld	[$inp+12],%l3
+	ld	[$inp+16],%l4
+	ld	[$inp+20],%l5
+	ld	[$inp+24],%l6
+	ld	[$inp+28],%l7
+___
+$code.=<<___ if ($i<15);
+	sllx	@pair[1],$tmp31,$tmp2	! Xload($i)
+	add	$tmp31,32,$tmp0
+	sllx	@pair[0],$tmp0,$tmp1
+	`"ld	[$inp+".eval(32+0+$i*8)."],@pair[0]"	if ($i<12)`
+	srlx	@pair[2],$tmp32,@pair[1]
+	or	$tmp1,$tmp2,$tmp2
+	or	@pair[1],$tmp2,$tmp2
+	`"ld	[$inp+".eval(32+4+$i*8)."],@pair[1]"	if ($i<12)`
+	add	$h,$tmp2,$T1
+	$ST	$tmp2,[%sp+`$bias+$frame+$i*$SZ`]
+___
+$code.=<<___ if ($i==12);
+	brnz,a	$tmp31,.+8
+	ld	[$inp+128],%l0
+___
+$code.=<<___ if ($i==15);
+	ld	[%sp+`$bias+$frame+(($i+1+1)%16)*$SZ+0`],%l2
+	sllx	@pair[1],$tmp31,$tmp2	! Xload($i)
+	add	$tmp31,32,$tmp0
+	ld	[%sp+`$bias+$frame+(($i+1+1)%16)*$SZ+4`],%l3
+	sllx	@pair[0],$tmp0,$tmp1
+	ld	[%sp+`$bias+$frame+(($i+1+9)%16)*$SZ+0`],%l4
+	srlx	@pair[2],$tmp32,@pair[1]
+	or	$tmp1,$tmp2,$tmp2
+	ld	[%sp+`$bias+$frame+(($i+1+9)%16)*$SZ+4`],%l5
+	or	@pair[1],$tmp2,$tmp2
+	ld	[%sp+`$bias+$frame+(($i+1+14)%16)*$SZ+0`],%l6
+	add	$h,$tmp2,$T1
+	$ST	$tmp2,[%sp+`$bias+$frame+$i*$SZ`]
+	ld	[%sp+`$bias+$frame+(($i+1+14)%16)*$SZ+4`],%l7
+	ld	[%sp+`$bias+$frame+(($i+1+0)%16)*$SZ+0`],%l0
+	ld	[%sp+`$bias+$frame+(($i+1+0)%16)*$SZ+4`],%l1
+___
+} if ($SZ==8);
+
+########### common
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+
+    if ($i<16) {
+	&$Xload(@_);
+    } else {
+	$code.="\tadd	$h,$T1,$T1\n";
+    }
+
+$code.=<<___;
+	$SRL	$e,@Sigma1[0],$h	!! $i
+	xor	$f,$g,$tmp2
+	$SLL	$e,`$SZ*8-@Sigma1[2]`,$tmp1
+	and	$e,$tmp2,$tmp2
+	$SRL	$e,@Sigma1[1],$tmp0
+	xor	$tmp1,$h,$h
+	$SLL	$e,`$SZ*8-@Sigma1[1]`,$tmp1
+	xor	$tmp0,$h,$h
+	$SRL	$e,@Sigma1[2],$tmp0
+	xor	$tmp1,$h,$h
+	$SLL	$e,`$SZ*8-@Sigma1[0]`,$tmp1
+	xor	$tmp0,$h,$h
+	xor	$g,$tmp2,$tmp2		! Ch(e,f,g)
+	xor	$tmp1,$h,$tmp0		! Sigma1(e)
+
+	$SRL	$a,@Sigma0[0],$h
+	add	$tmp2,$T1,$T1
+	$LD	[$Ktbl+`$i*$SZ`],$tmp2	! K[$i]
+	$SLL	$a,`$SZ*8-@Sigma0[2]`,$tmp1
+	add	$tmp0,$T1,$T1
+	$SRL	$a,@Sigma0[1],$tmp0
+	xor	$tmp1,$h,$h
+	$SLL	$a,`$SZ*8-@Sigma0[1]`,$tmp1
+	xor	$tmp0,$h,$h
+	$SRL	$a,@Sigma0[2],$tmp0
+	xor	$tmp1,$h,$h	
+	$SLL	$a,`$SZ*8-@Sigma0[0]`,$tmp1
+	xor	$tmp0,$h,$h
+	xor	$tmp1,$h,$h		! Sigma0(a)
+
+	or	$a,$b,$tmp0
+	and	$a,$b,$tmp1
+	and	$c,$tmp0,$tmp0
+	or	$tmp0,$tmp1,$tmp1	! Maj(a,b,c)
+	add	$tmp2,$T1,$T1		! +=K[$i]
+	add	$tmp1,$h,$h
+
+	add	$T1,$d,$d
+	add	$T1,$h,$h
+___
+}
+
+########### SHA256
+$BODY_16_XX = sub {
+my $i=@_[0];
+my $xi;
+
+    if ($i&1) {
+	$xi=$tmp32;
+	$code.="\tsrlx	@X[(($i+1)/2)%8],32,$xi\n";
+    } else {
+	$xi=@X[(($i+1)/2)%8];
+    }
+$code.=<<___;
+	srl	$xi,@sigma0[0],$T1		!! Xupdate($i)
+	sll	$xi,`32-@sigma0[2]`,$tmp1
+	srl	$xi,@sigma0[1],$tmp0
+	xor	$tmp1,$T1,$T1
+	sll	$tmp1,`@sigma0[2]-@sigma0[1]`,$tmp1
+	xor	$tmp0,$T1,$T1
+	srl	$xi,@sigma0[2],$tmp0
+	xor	$tmp1,$T1,$T1
+___
+    if ($i&1) {
+	$xi=@X[(($i+14)/2)%8];
+    } else {
+	$xi=$tmp32;
+	$code.="\tsrlx	@X[(($i+14)/2)%8],32,$xi\n";
+    }
+$code.=<<___;
+	srl	$xi,@sigma1[0],$tmp2
+	xor	$tmp0,$T1,$T1			! T1=sigma0(X[i+1])
+	sll	$xi,`32-@sigma1[2]`,$tmp1
+	srl	$xi,@sigma1[1],$tmp0
+	xor	$tmp1,$tmp2,$tmp2
+	sll	$tmp1,`@sigma1[2]-@sigma1[1]`,$tmp1
+	xor	$tmp0,$tmp2,$tmp2
+	srl	$xi,@sigma1[2],$tmp0
+	xor	$tmp1,$tmp2,$tmp2
+___
+    if ($i&1) {
+	$xi=@X[($i/2)%8];
+$code.=<<___;
+	srlx	@X[(($i+9)/2)%8],32,$tmp1	! X[i+9]
+	xor	$tmp0,$tmp2,$tmp2		! sigma1(X[i+14])
+	srl	@X[($i/2)%8],0,$tmp0
+	add	$tmp2,$tmp1,$tmp1
+	add	$xi,$T1,$T1			! +=X[i]
+	xor	$tmp0,@X[($i/2)%8],@X[($i/2)%8]
+	add	$tmp1,$T1,$T1
+
+	srl	$T1,0,$T1
+	or	$T1,@X[($i/2)%8],@X[($i/2)%8]
+___
+    } else {
+	$xi=@X[(($i+9)/2)%8];
+$code.=<<___;
+	srlx	@X[($i/2)%8],32,$tmp1		! X[i]
+	xor	$tmp0,$tmp2,$tmp2		! sigma1(X[i+14])
+	add	$xi,$T1,$T1			! +=X[i+9]
+	add	$tmp2,$tmp1,$tmp1
+	srl	@X[($i/2)%8],0,@X[($i/2)%8]
+	add	$tmp1,$T1,$T1
+
+	sllx	$T1,32,$tmp0
+	or	$tmp0,@X[($i/2)%8],@X[($i/2)%8]
+___
+    }
+    &BODY_00_15(@_);
+} if ($SZ==4);
+
+########### SHA512
+$BODY_16_XX = sub {
+my $i=@_[0];
+my @pair=("%l".eval(($i*2)%8),"%l".eval(($i*2)%8+1));
+
+$code.=<<___;
+	sllx	%l2,32,$tmp0		!! Xupdate($i)
+	or	%l3,$tmp0,$tmp0
+
+	srlx	$tmp0,@sigma0[0],$T1
+	ld	[%sp+`$bias+$frame+(($i+1+1)%16)*$SZ+0`],%l2
+	sllx	$tmp0,`64-@sigma0[2]`,$tmp1
+	ld	[%sp+`$bias+$frame+(($i+1+1)%16)*$SZ+4`],%l3
+	srlx	$tmp0,@sigma0[1],$tmp0
+	xor	$tmp1,$T1,$T1
+	sllx	$tmp1,`@sigma0[2]-@sigma0[1]`,$tmp1
+	xor	$tmp0,$T1,$T1
+	srlx	$tmp0,`@sigma0[2]-@sigma0[1]`,$tmp0
+	xor	$tmp1,$T1,$T1
+	sllx	%l6,32,$tmp2
+	xor	$tmp0,$T1,$T1		! sigma0(X[$i+1])
+	or	%l7,$tmp2,$tmp2
+
+	srlx	$tmp2,@sigma1[0],$tmp1
+	ld	[%sp+`$bias+$frame+(($i+1+14)%16)*$SZ+0`],%l6
+	sllx	$tmp2,`64-@sigma1[2]`,$tmp0
+	ld	[%sp+`$bias+$frame+(($i+1+14)%16)*$SZ+4`],%l7
+	srlx	$tmp2,@sigma1[1],$tmp2
+	xor	$tmp0,$tmp1,$tmp1
+	sllx	$tmp0,`@sigma1[2]-@sigma1[1]`,$tmp0
+	xor	$tmp2,$tmp1,$tmp1
+	srlx	$tmp2,`@sigma1[2]-@sigma1[1]`,$tmp2
+	xor	$tmp0,$tmp1,$tmp1
+	sllx	%l4,32,$tmp0
+	xor	$tmp2,$tmp1,$tmp1	! sigma1(X[$i+14])
+	ld	[%sp+`$bias+$frame+(($i+1+9)%16)*$SZ+0`],%l4
+	or	%l5,$tmp0,$tmp0
+	ld	[%sp+`$bias+$frame+(($i+1+9)%16)*$SZ+4`],%l5
+
+	sllx	%l0,32,$tmp2
+	add	$tmp1,$T1,$T1
+	ld	[%sp+`$bias+$frame+(($i+1+0)%16)*$SZ+0`],%l0
+	or	%l1,$tmp2,$tmp2
+	add	$tmp0,$T1,$T1		! +=X[$i+9]
+	ld	[%sp+`$bias+$frame+(($i+1+0)%16)*$SZ+4`],%l1
+	add	$tmp2,$T1,$T1		! +=X[$i]
+	$ST	$T1,[%sp+`$bias+$frame+($i%16)*$SZ`]
+___
+    &BODY_00_15(@_);
+} if ($SZ==8);
+
+$code.=<<___ if ($bits==64);
+.register	%g2,#scratch
+.register	%g3,#scratch
+___
+$code.=<<___;
+.section	".text",#alloc,#execinstr
+
+.align	64
+K${label}:
+.type	K${label},#object
+___
+if ($SZ==4) {
+$code.=<<___;
+	.long	0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5
+	.long	0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5
+	.long	0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3
+	.long	0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174
+	.long	0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc
+	.long	0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da
+	.long	0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7
+	.long	0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967
+	.long	0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13
+	.long	0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85
+	.long	0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3
+	.long	0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070
+	.long	0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5
+	.long	0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3
+	.long	0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208
+	.long	0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
+___
+} else {
+$code.=<<___;
+	.long	0x428a2f98,0xd728ae22, 0x71374491,0x23ef65cd
+	.long	0xb5c0fbcf,0xec4d3b2f, 0xe9b5dba5,0x8189dbbc
+	.long	0x3956c25b,0xf348b538, 0x59f111f1,0xb605d019
+	.long	0x923f82a4,0xaf194f9b, 0xab1c5ed5,0xda6d8118
+	.long	0xd807aa98,0xa3030242, 0x12835b01,0x45706fbe
+	.long	0x243185be,0x4ee4b28c, 0x550c7dc3,0xd5ffb4e2
+	.long	0x72be5d74,0xf27b896f, 0x80deb1fe,0x3b1696b1
+	.long	0x9bdc06a7,0x25c71235, 0xc19bf174,0xcf692694
+	.long	0xe49b69c1,0x9ef14ad2, 0xefbe4786,0x384f25e3
+	.long	0x0fc19dc6,0x8b8cd5b5, 0x240ca1cc,0x77ac9c65
+	.long	0x2de92c6f,0x592b0275, 0x4a7484aa,0x6ea6e483
+	.long	0x5cb0a9dc,0xbd41fbd4, 0x76f988da,0x831153b5
+	.long	0x983e5152,0xee66dfab, 0xa831c66d,0x2db43210
+	.long	0xb00327c8,0x98fb213f, 0xbf597fc7,0xbeef0ee4
+	.long	0xc6e00bf3,0x3da88fc2, 0xd5a79147,0x930aa725
+	.long	0x06ca6351,0xe003826f, 0x14292967,0x0a0e6e70
+	.long	0x27b70a85,0x46d22ffc, 0x2e1b2138,0x5c26c926
+	.long	0x4d2c6dfc,0x5ac42aed, 0x53380d13,0x9d95b3df
+	.long	0x650a7354,0x8baf63de, 0x766a0abb,0x3c77b2a8
+	.long	0x81c2c92e,0x47edaee6, 0x92722c85,0x1482353b
+	.long	0xa2bfe8a1,0x4cf10364, 0xa81a664b,0xbc423001
+	.long	0xc24b8b70,0xd0f89791, 0xc76c51a3,0x0654be30
+	.long	0xd192e819,0xd6ef5218, 0xd6990624,0x5565a910
+	.long	0xf40e3585,0x5771202a, 0x106aa070,0x32bbd1b8
+	.long	0x19a4c116,0xb8d2d0c8, 0x1e376c08,0x5141ab53
+	.long	0x2748774c,0xdf8eeb99, 0x34b0bcb5,0xe19b48a8
+	.long	0x391c0cb3,0xc5c95a63, 0x4ed8aa4a,0xe3418acb
+	.long	0x5b9cca4f,0x7763e373, 0x682e6ff3,0xd6b2b8a3
+	.long	0x748f82ee,0x5defb2fc, 0x78a5636f,0x43172f60
+	.long	0x84c87814,0xa1f0ab72, 0x8cc70208,0x1a6439ec
+	.long	0x90befffa,0x23631e28, 0xa4506ceb,0xde82bde9
+	.long	0xbef9a3f7,0xb2c67915, 0xc67178f2,0xe372532b
+	.long	0xca273ece,0xea26619c, 0xd186b8c7,0x21c0c207
+	.long	0xeada7dd6,0xcde0eb1e, 0xf57d4f7f,0xee6ed178
+	.long	0x06f067aa,0x72176fba, 0x0a637dc5,0xa2c898a6
+	.long	0x113f9804,0xbef90dae, 0x1b710b35,0x131c471b
+	.long	0x28db77f5,0x23047d84, 0x32caab7b,0x40c72493
+	.long	0x3c9ebe0a,0x15c9bebc, 0x431d67c4,0x9c100d4c
+	.long	0x4cc5d4be,0xcb3e42b6, 0x597f299c,0xfc657e2a
+	.long	0x5fcb6fab,0x3ad6faec, 0x6c44198c,0x4a475817
+___
+}
+$code.=<<___;
+.size	K${label},.-K${label}
+.globl	sha${label}_block_data_order
+sha${label}_block_data_order:
+	save	%sp,`-$frame-$locals`,%sp
+	and	$inp,`$align-1`,$tmp31
+	sllx	$len,`log(16*$SZ)/log(2)`,$len
+	andn	$inp,`$align-1`,$inp
+	sll	$tmp31,3,$tmp31
+	add	$inp,$len,$len
+___
+$code.=<<___ if ($SZ==8); # SHA512
+	mov	32,$tmp32
+	sub	$tmp32,$tmp31,$tmp32
+___
+$code.=<<___;
+.Lpic:	call	.+8
+	add	%o7,K${label}-.Lpic,$Ktbl
+
+	$LD	[$ctx+`0*$SZ`],$A
+	$LD	[$ctx+`1*$SZ`],$B
+	$LD	[$ctx+`2*$SZ`],$C
+	$LD	[$ctx+`3*$SZ`],$D
+	$LD	[$ctx+`4*$SZ`],$E
+	$LD	[$ctx+`5*$SZ`],$F
+	$LD	[$ctx+`6*$SZ`],$G
+	$LD	[$ctx+`7*$SZ`],$H
+
+.Lloop:
+___
+for ($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=".L16_xx:\n";
+for (;$i<32;$i++)	{ &$BODY_16_XX($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	and	$tmp2,0xfff,$tmp2
+	cmp	$tmp2,$lastK
+	bne	.L16_xx
+	add	$Ktbl,`16*$SZ`,$Ktbl	! Ktbl+=16
+
+___
+$code.=<<___ if ($SZ==4); # SHA256
+	$LD	[$ctx+`0*$SZ`],@X[0]
+	$LD	[$ctx+`1*$SZ`],@X[1]
+	$LD	[$ctx+`2*$SZ`],@X[2]
+	$LD	[$ctx+`3*$SZ`],@X[3]
+	$LD	[$ctx+`4*$SZ`],@X[4]
+	$LD	[$ctx+`5*$SZ`],@X[5]
+	$LD	[$ctx+`6*$SZ`],@X[6]
+	$LD	[$ctx+`7*$SZ`],@X[7]
+
+	add	$A,@X[0],$A
+	$ST	$A,[$ctx+`0*$SZ`]
+	add	$B,@X[1],$B
+	$ST	$B,[$ctx+`1*$SZ`]
+	add	$C,@X[2],$C
+	$ST	$C,[$ctx+`2*$SZ`]
+	add	$D,@X[3],$D
+	$ST	$D,[$ctx+`3*$SZ`]
+	add	$E,@X[4],$E
+	$ST	$E,[$ctx+`4*$SZ`]
+	add	$F,@X[5],$F
+	$ST	$F,[$ctx+`5*$SZ`]
+	add	$G,@X[6],$G
+	$ST	$G,[$ctx+`6*$SZ`]
+	add	$H,@X[7],$H
+	$ST	$H,[$ctx+`7*$SZ`]
+___
+$code.=<<___ if ($SZ==8); # SHA512
+	ld	[$ctx+`0*$SZ+0`],%l0
+	ld	[$ctx+`0*$SZ+4`],%l1
+	ld	[$ctx+`1*$SZ+0`],%l2
+	ld	[$ctx+`1*$SZ+4`],%l3
+	ld	[$ctx+`2*$SZ+0`],%l4
+	ld	[$ctx+`2*$SZ+4`],%l5
+	ld	[$ctx+`3*$SZ+0`],%l6
+
+	sllx	%l0,32,$tmp0
+	ld	[$ctx+`3*$SZ+4`],%l7
+	sllx	%l2,32,$tmp1
+	or	%l1,$tmp0,$tmp0
+	or	%l3,$tmp1,$tmp1
+	add	$tmp0,$A,$A
+	add	$tmp1,$B,$B
+	$ST	$A,[$ctx+`0*$SZ`]
+	sllx	%l4,32,$tmp2
+	$ST	$B,[$ctx+`1*$SZ`]
+	sllx	%l6,32,$T1
+	or	%l5,$tmp2,$tmp2
+	or	%l7,$T1,$T1
+	add	$tmp2,$C,$C
+	$ST	$C,[$ctx+`2*$SZ`]
+	add	$T1,$D,$D
+	$ST	$D,[$ctx+`3*$SZ`]
+
+	ld	[$ctx+`4*$SZ+0`],%l0
+	ld	[$ctx+`4*$SZ+4`],%l1
+	ld	[$ctx+`5*$SZ+0`],%l2
+	ld	[$ctx+`5*$SZ+4`],%l3
+	ld	[$ctx+`6*$SZ+0`],%l4
+	ld	[$ctx+`6*$SZ+4`],%l5
+	ld	[$ctx+`7*$SZ+0`],%l6
+
+	sllx	%l0,32,$tmp0
+	ld	[$ctx+`7*$SZ+4`],%l7
+	sllx	%l2,32,$tmp1
+	or	%l1,$tmp0,$tmp0
+	or	%l3,$tmp1,$tmp1
+	add	$tmp0,$E,$E
+	add	$tmp1,$F,$F
+	$ST	$E,[$ctx+`4*$SZ`]
+	sllx	%l4,32,$tmp2
+	$ST	$F,[$ctx+`5*$SZ`]
+	sllx	%l6,32,$T1
+	or	%l5,$tmp2,$tmp2
+	or	%l7,$T1,$T1
+	add	$tmp2,$G,$G
+	$ST	$G,[$ctx+`6*$SZ`]
+	add	$T1,$H,$H
+	$ST	$H,[$ctx+`7*$SZ`]
+___
+$code.=<<___;
+	add	$inp,`16*$SZ`,$inp		! advance inp
+	cmp	$inp,$len
+	bne	`$bits==64?"%xcc":"%icc"`,.Lloop
+	sub	$Ktbl,`($rounds-16)*$SZ`,$Ktbl	! rewind Ktbl
+
+	ret
+	restore
+.type	sha${label}_block_data_order,#function
+.size	sha${label}_block_data_order,(.-sha${label}_block_data_order)
+.asciz	"SHA${label} block transform for SPARCv9, CRYPTOGAMS by <appro\@openssl.org>"
+.align	4
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/sha/asm/sha512-sse2.pl b/crypto/sha/asm/sha512-sse2.pl
deleted file mode 100644
index 10902bf673d5..000000000000
--- a/crypto/sha/asm/sha512-sse2.pl
+++ /dev/null
@@ -1,404 +0,0 @@
-#!/usr/bin/env perl
-#
-# ====================================================================
-# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
-# project. Rights for redistribution and usage in source and binary
-# forms are granted according to the OpenSSL license.
-# ====================================================================
-#
-# SHA512_Transform_SSE2.
-#
-# As the name suggests, this is an IA-32 SSE2 implementation of
-# SHA512_Transform. Motivating factor for the undertaken effort was that
-# SHA512 was observed to *consistently* perform *significantly* poorer
-# than SHA256 [2x and slower is common] on 32-bit platforms. On 64-bit
-# platforms on the other hand SHA512 tend to outperform SHA256 [~50%
-# seem to be common improvement factor]. All this is perfectly natural,
-# as SHA512 is a 64-bit algorithm. But isn't IA-32 SSE2 essentially
-# a 64-bit instruction set? Is it rich enough to implement SHA512?
-# If answer was "no," then you wouldn't have been reading this...
-#
-# Throughput performance in MBps (larger is better):
-#
-#		2.4GHz P4	1.4GHz AMD32	1.4GHz AMD64(*)
-# SHA256/gcc(*)	54		43		59
-# SHA512/gcc	17		23		92
-# SHA512/sse2	61(**)		57(**)
-# SHA512/icc	26		28
-# SHA256/icc(*)	65		54
-#
-# (*)	AMD64 and SHA256 numbers are presented mostly for amusement or
-#	reference purposes.
-# (**)	I.e. it gives ~2-3x speed-up if compared with compiler generated
-#	code. One can argue that hand-coded *non*-SSE2 implementation
-#	would perform better than compiler generated one as well, and
-#	that comparison is therefore not exactly fair. Well, as SHA512
-#	puts enormous pressure on IA-32 GP register bank, I reckon that
-#	hand-coded version wouldn't perform significantly better than
-#	one compiled with icc, ~20% perhaps... So that this code would
-#	still outperform it with distinguishing marginal. But feel free
-#	to prove me wrong:-)
-#						<appro@fy.chalmers.se>
-push(@INC,"perlasm","../../perlasm");
-require "x86asm.pl";
-
-&asm_init($ARGV[0],"sha512-sse2.pl",$ARGV[$#ARGV] eq "386");
-
-$K512="esi";	# K512[80] table, found at the end...
-#$W512="esp";	# $W512 is not just W512[16]: it comprises *two* copies
-		# of W512[16] and a copy of A-H variables...
-$W512_SZ=8*(16+16+8);	# see above...
-#$Kidx="ebx";	# index in K512 table, advances from 0 to 80...
-$Widx="edx";	# index in W512, wraps around at 16...
-$data="edi";	# 16 qwords of input data...
-$A="mm0";	# B-D and
-$E="mm1";	# F-H are allocated dynamically...
-$Aoff=256+0;	# A-H offsets relative to $W512...
-$Boff=256+8;
-$Coff=256+16;
-$Doff=256+24;
-$Eoff=256+32;
-$Foff=256+40;
-$Goff=256+48;
-$Hoff=256+56;
-
-sub SHA2_ROUND()
-{ local ($kidx,$widx)=@_;
-
-	# One can argue that one could reorder instructions for better
-	# performance. Well, I tried and it doesn't seem to make any
-	# noticeable difference. Modern out-of-order execution cores
-	# reorder instructions to their liking in either case and they
-	# apparently do decent job. So we can keep the code more
-	# readable/regular/comprehensible:-)
-
-	# I adhere to 64-bit %mmX registers in order to avoid/not care
-	# about #GP exceptions on misaligned 128-bit access, most
-	# notably in paddq with memory operand. Not to mention that
-	# SSE2 intructions operating on %mmX can be scheduled every
-	# cycle [and not every second one if operating on %xmmN].
-
-	&movq	("mm4",&QWP($Foff,$W512));	# load f
-	&movq	("mm5",&QWP($Goff,$W512));	# load g
-	&movq	("mm6",&QWP($Hoff,$W512));	# load h
-
-	&movq	("mm2",$E);			# %mm2 is sliding right
-	&movq	("mm3",$E);			# %mm3 is sliding left
-	&psrlq	("mm2",14);
-	&psllq	("mm3",23);
-	&movq	("mm7","mm2");			# %mm7 is T1
-	&pxor	("mm7","mm3");
-	&psrlq	("mm2",4);
-	&psllq	("mm3",23);
-	&pxor	("mm7","mm2");
-	&pxor	("mm7","mm3");
-	&psrlq	("mm2",23);
-	&psllq	("mm3",4);
-	&pxor	("mm7","mm2");
-	&pxor	("mm7","mm3");			# T1=Sigma1_512(e)
-
-	&movq	(&QWP($Foff,$W512),$E);		# f = e
-	&movq	(&QWP($Goff,$W512),"mm4");	# g = f
-	&movq	(&QWP($Hoff,$W512),"mm5");	# h = g
-
-	&pxor	("mm4","mm5");			# f^=g
-	&pand	("mm4",$E);			# f&=e
-	&pxor	("mm4","mm5");			# f^=g
-	&paddq	("mm7","mm4");			# T1+=Ch(e,f,g)
-
-	&movq	("mm2",&QWP($Boff,$W512));	# load b
-	&movq	("mm3",&QWP($Coff,$W512));	# load c
-	&movq	($E,&QWP($Doff,$W512));		# e = d
-
-	&paddq	("mm7","mm6");			# T1+=h
-	&paddq	("mm7",&QWP(0,$K512,$kidx,8));	# T1+=K512[i]
-	&paddq	("mm7",&QWP(0,$W512,$widx,8));	# T1+=W512[i]
-	&paddq	($E,"mm7");			# e += T1
-
-	&movq	("mm4",$A);			# %mm4 is sliding right
-	&movq	("mm5",$A);			# %mm5 is sliding left
-	&psrlq	("mm4",28);
-	&psllq	("mm5",25);
-	&movq	("mm6","mm4");			# %mm6 is T2
-	&pxor	("mm6","mm5");
-	&psrlq	("mm4",6);
-	&psllq	("mm5",5);
-	&pxor	("mm6","mm4");
-	&pxor	("mm6","mm5");
-	&psrlq	("mm4",5);
-	&psllq	("mm5",6);
-	&pxor	("mm6","mm4");
-	&pxor	("mm6","mm5");			# T2=Sigma0_512(a)
-
-	&movq	(&QWP($Boff,$W512),$A);		# b = a
-	&movq	(&QWP($Coff,$W512),"mm2");	# c = b
-	&movq	(&QWP($Doff,$W512),"mm3");	# d = c
-
-	&movq	("mm4",$A);			# %mm4=a
-	&por	($A,"mm3");			# a=a|c
-	&pand	("mm4","mm3");			# %mm4=a&c
-	&pand	($A,"mm2");			# a=(a|c)&b
-	&por	("mm4",$A);			# %mm4=(a&c)|((a|c)&b)
-	&paddq	("mm6","mm4");			# T2+=Maj(a,b,c)
-
-	&movq	($A,"mm7");			# a=T1
-	&paddq	($A,"mm6");			# a+=T2
-}
-
-$func="sha512_block_sse2";
-
-&function_begin_B($func);
-	if (0) {# Caller is expected to check if it's appropriate to
-		# call this routine. Below 3 lines are retained for
-		# debugging purposes...
-		&picmeup("eax","OPENSSL_ia32cap");
-		&bt	(&DWP(0,"eax"),26);
-		&jnc	("SHA512_Transform");
-	}
-
-	&push	("ebp");
-	&mov	("ebp","esp");
-	&push	("ebx");
-	&push	("esi");
-	&push	("edi");
-
-	&mov	($Widx,&DWP(8,"ebp"));		# A-H state, 1st arg
-	&mov	($data,&DWP(12,"ebp"));		# input data, 2nd arg
-	&call	(&label("pic_point"));		# make it PIC!
-&set_label("pic_point");
-	&blindpop($K512);
-	&lea	($K512,&DWP(&label("K512")."-".&label("pic_point"),$K512));
-
-	$W512 = "esp";			# start using %esp as W512
-	&sub	($W512,$W512_SZ);
-	&and	($W512,-16);		# ensure 128-bit alignment
-
-	# make private copy of A-H
-	#     v assume the worst and stick to unaligned load
-	&movdqu	("xmm0",&QWP(0,$Widx));
-	&movdqu	("xmm1",&QWP(16,$Widx));
-	&movdqu	("xmm2",&QWP(32,$Widx));
-	&movdqu	("xmm3",&QWP(48,$Widx));
-
-&align(8);
-&set_label("_chunk_loop");
-
-	&movdqa	(&QWP($Aoff,$W512),"xmm0");	# a,b
-	&movdqa	(&QWP($Coff,$W512),"xmm1");	# c,d
-	&movdqa	(&QWP($Eoff,$W512),"xmm2");	# e,f
-	&movdqa	(&QWP($Goff,$W512),"xmm3");	# g,h
-
-	&xor	($Widx,$Widx);
-
-	&movdq2q($A,"xmm0");			# load a
-	&movdq2q($E,"xmm2");			# load e
-
-	# Why aren't loops unrolled? It makes sense to unroll if
-	# execution time for loop body is comparable with branch
-	# penalties and/or if whole data-set resides in register bank.
-	# Neither is case here... Well, it would be possible to
-	# eliminate few store operations, but it would hardly affect
-	# so to say stop-watch performance, as there is a lot of
-	# available memory slots to fill. It will only relieve some
-	# pressure off memory bus...
-
-	# flip input stream byte order...
-	&mov	("eax",&DWP(0,$data,$Widx,8));
-	&mov	("ebx",&DWP(4,$data,$Widx,8));
-	&bswap	("eax");
-	&bswap	("ebx");
-	&mov	(&DWP(0,$W512,$Widx,8),"ebx");		# W512[i]
-	&mov	(&DWP(4,$W512,$Widx,8),"eax");
-	&mov	(&DWP(128+0,$W512,$Widx,8),"ebx");	# copy of W512[i]
-	&mov	(&DWP(128+4,$W512,$Widx,8),"eax");
-
-&align(8);
-&set_label("_1st_loop");		# 0-15
-	# flip input stream byte order...
-	&mov	("eax",&DWP(0+8,$data,$Widx,8));
-	&mov	("ebx",&DWP(4+8,$data,$Widx,8));
-	&bswap	("eax");
-	&bswap	("ebx");
-	&mov	(&DWP(0+8,$W512,$Widx,8),"ebx");	# W512[i]
-	&mov	(&DWP(4+8,$W512,$Widx,8),"eax");
-	&mov	(&DWP(128+0+8,$W512,$Widx,8),"ebx");	# copy of W512[i]
-	&mov	(&DWP(128+4+8,$W512,$Widx,8),"eax");
-&set_label("_1st_looplet");
-	&SHA2_ROUND($Widx,$Widx); &inc($Widx);
-
-&cmp	($Widx,15)
-&jl	(&label("_1st_loop"));
-&je	(&label("_1st_looplet"));	# playing similar trick on 2nd loop
-					# does not improve performance...
-
-	$Kidx = "ebx";			# start using %ebx as Kidx
-	&mov	($Kidx,$Widx);
-
-&align(8);
-&set_label("_2nd_loop");		# 16-79
-	&and($Widx,0xf);
-
-	# 128-bit fragment! I update W512[i] and W512[i+1] in
-	# parallel:-) Note that I refer to W512[(i&0xf)+N] and not to
-	# W512[(i+N)&0xf]! This is exactly what I maintain the second
-	# copy of W512[16] for...
-	&movdqu	("xmm0",&QWP(8*1,$W512,$Widx,8));	# s0=W512[i+1]
-	&movdqa	("xmm2","xmm0");		# %xmm2 is sliding right
-	&movdqa	("xmm3","xmm0");		# %xmm3 is sliding left
-	&psrlq	("xmm2",1);
-	&psllq	("xmm3",56);
-	&movdqa	("xmm0","xmm2");
-	&pxor	("xmm0","xmm3");
-	&psrlq	("xmm2",6);
-	&psllq	("xmm3",7);
-	&pxor	("xmm0","xmm2");
-	&pxor	("xmm0","xmm3");
-	&psrlq	("xmm2",1);
-	&pxor	("xmm0","xmm2");		# s0 = sigma0_512(s0);
-
-	&movdqa	("xmm1",&QWP(8*14,$W512,$Widx,8));	# s1=W512[i+14]
-	&movdqa	("xmm4","xmm1");		# %xmm4 is sliding right
-	&movdqa	("xmm5","xmm1");		# %xmm5 is sliding left
-	&psrlq	("xmm4",6);
-	&psllq	("xmm5",3);
-	&movdqa	("xmm1","xmm4");
-	&pxor	("xmm1","xmm5");
-	&psrlq	("xmm4",13);
-	&psllq	("xmm5",42);
-	&pxor	("xmm1","xmm4");
-	&pxor	("xmm1","xmm5");
-	&psrlq	("xmm4",42);
-	&pxor	("xmm1","xmm4");		# s1 = sigma1_512(s1);
-
-	#     + have to explictly load W512[i+9] as it's not 128-bit
-	#     v	aligned and paddq would throw an exception...
-	&movdqu	("xmm6",&QWP(8*9,$W512,$Widx,8));
-	&paddq	("xmm0","xmm1");		# s0 += s1
-	&paddq	("xmm0","xmm6");		# s0 += W512[i+9]
-	&paddq	("xmm0",&QWP(0,$W512,$Widx,8));	# s0 += W512[i]
-
-	&movdqa	(&QWP(0,$W512,$Widx,8),"xmm0");		# W512[i] = s0
-	&movdqa	(&QWP(16*8,$W512,$Widx,8),"xmm0");	# copy of W512[i]
-
-	# as the above fragment was 128-bit, we "owe" 2 rounds...
-	&SHA2_ROUND($Kidx,$Widx); &inc($Kidx); &inc($Widx);
-	&SHA2_ROUND($Kidx,$Widx); &inc($Kidx); &inc($Widx);
-
-&cmp	($Kidx,80);
-&jl	(&label("_2nd_loop"));
-
-	# update A-H state
-	&mov	($Widx,&DWP(8,"ebp"));		# A-H state, 1st arg
-	&movq	(&QWP($Aoff,$W512),$A);		# write out a
-	&movq	(&QWP($Eoff,$W512),$E);		# write out e
-	&movdqu	("xmm0",&QWP(0,$Widx));
-	&movdqu	("xmm1",&QWP(16,$Widx));
-	&movdqu	("xmm2",&QWP(32,$Widx));
-	&movdqu	("xmm3",&QWP(48,$Widx));
-	&paddq	("xmm0",&QWP($Aoff,$W512));	# 128-bit additions...
-	&paddq	("xmm1",&QWP($Coff,$W512));
-	&paddq	("xmm2",&QWP($Eoff,$W512));
-	&paddq	("xmm3",&QWP($Goff,$W512));
-	&movdqu	(&QWP(0,$Widx),"xmm0");
-	&movdqu	(&QWP(16,$Widx),"xmm1");
-	&movdqu	(&QWP(32,$Widx),"xmm2");
-	&movdqu	(&QWP(48,$Widx),"xmm3");
-
-&add	($data,16*8);				# advance input data pointer
-&dec	(&DWP(16,"ebp"));			# decrement 3rd arg
-&jnz	(&label("_chunk_loop"));
-
-	# epilogue
-	&emms	();	# required for at least ELF and Win32 ABIs
-	&mov	("edi",&DWP(-12,"ebp"));
-	&mov	("esi",&DWP(-8,"ebp"));
-	&mov	("ebx",&DWP(-4,"ebp"));
-	&leave	();
-&ret	();
-
-&align(64);
-&set_label("K512");	# Yes! I keep it in the code segment!
-	&data_word(0xd728ae22,0x428a2f98);	# u64
-	&data_word(0x23ef65cd,0x71374491);	# u64
-	&data_word(0xec4d3b2f,0xb5c0fbcf);	# u64
-	&data_word(0x8189dbbc,0xe9b5dba5);	# u64
-	&data_word(0xf348b538,0x3956c25b);	# u64
-	&data_word(0xb605d019,0x59f111f1);	# u64
-	&data_word(0xaf194f9b,0x923f82a4);	# u64
-	&data_word(0xda6d8118,0xab1c5ed5);	# u64
-	&data_word(0xa3030242,0xd807aa98);	# u64
-	&data_word(0x45706fbe,0x12835b01);	# u64
-	&data_word(0x4ee4b28c,0x243185be);	# u64
-	&data_word(0xd5ffb4e2,0x550c7dc3);	# u64
-	&data_word(0xf27b896f,0x72be5d74);	# u64
-	&data_word(0x3b1696b1,0x80deb1fe);	# u64
-	&data_word(0x25c71235,0x9bdc06a7);	# u64
-	&data_word(0xcf692694,0xc19bf174);	# u64
-	&data_word(0x9ef14ad2,0xe49b69c1);	# u64
-	&data_word(0x384f25e3,0xefbe4786);	# u64
-	&data_word(0x8b8cd5b5,0x0fc19dc6);	# u64
-	&data_word(0x77ac9c65,0x240ca1cc);	# u64
-	&data_word(0x592b0275,0x2de92c6f);	# u64
-	&data_word(0x6ea6e483,0x4a7484aa);	# u64
-	&data_word(0xbd41fbd4,0x5cb0a9dc);	# u64
-	&data_word(0x831153b5,0x76f988da);	# u64
-	&data_word(0xee66dfab,0x983e5152);	# u64
-	&data_word(0x2db43210,0xa831c66d);	# u64
-	&data_word(0x98fb213f,0xb00327c8);	# u64
-	&data_word(0xbeef0ee4,0xbf597fc7);	# u64
-	&data_word(0x3da88fc2,0xc6e00bf3);	# u64
-	&data_word(0x930aa725,0xd5a79147);	# u64
-	&data_word(0xe003826f,0x06ca6351);	# u64
-	&data_word(0x0a0e6e70,0x14292967);	# u64
-	&data_word(0x46d22ffc,0x27b70a85);	# u64
-	&data_word(0x5c26c926,0x2e1b2138);	# u64
-	&data_word(0x5ac42aed,0x4d2c6dfc);	# u64
-	&data_word(0x9d95b3df,0x53380d13);	# u64
-	&data_word(0x8baf63de,0x650a7354);	# u64
-	&data_word(0x3c77b2a8,0x766a0abb);	# u64
-	&data_word(0x47edaee6,0x81c2c92e);	# u64
-	&data_word(0x1482353b,0x92722c85);	# u64
-	&data_word(0x4cf10364,0xa2bfe8a1);	# u64
-	&data_word(0xbc423001,0xa81a664b);	# u64
-	&data_word(0xd0f89791,0xc24b8b70);	# u64
-	&data_word(0x0654be30,0xc76c51a3);	# u64
-	&data_word(0xd6ef5218,0xd192e819);	# u64
-	&data_word(0x5565a910,0xd6990624);	# u64
-	&data_word(0x5771202a,0xf40e3585);	# u64
-	&data_word(0x32bbd1b8,0x106aa070);	# u64
-	&data_word(0xb8d2d0c8,0x19a4c116);	# u64
-	&data_word(0x5141ab53,0x1e376c08);	# u64
-	&data_word(0xdf8eeb99,0x2748774c);	# u64
-	&data_word(0xe19b48a8,0x34b0bcb5);	# u64
-	&data_word(0xc5c95a63,0x391c0cb3);	# u64
-	&data_word(0xe3418acb,0x4ed8aa4a);	# u64
-	&data_word(0x7763e373,0x5b9cca4f);	# u64
-	&data_word(0xd6b2b8a3,0x682e6ff3);	# u64
-	&data_word(0x5defb2fc,0x748f82ee);	# u64
-	&data_word(0x43172f60,0x78a5636f);	# u64
-	&data_word(0xa1f0ab72,0x84c87814);	# u64
-	&data_word(0x1a6439ec,0x8cc70208);	# u64
-	&data_word(0x23631e28,0x90befffa);	# u64
-	&data_word(0xde82bde9,0xa4506ceb);	# u64
-	&data_word(0xb2c67915,0xbef9a3f7);	# u64
-	&data_word(0xe372532b,0xc67178f2);	# u64
-	&data_word(0xea26619c,0xca273ece);	# u64
-	&data_word(0x21c0c207,0xd186b8c7);	# u64
-	&data_word(0xcde0eb1e,0xeada7dd6);	# u64
-	&data_word(0xee6ed178,0xf57d4f7f);	# u64
-	&data_word(0x72176fba,0x06f067aa);	# u64
-	&data_word(0xa2c898a6,0x0a637dc5);	# u64
-	&data_word(0xbef90dae,0x113f9804);	# u64
-	&data_word(0x131c471b,0x1b710b35);	# u64
-	&data_word(0x23047d84,0x28db77f5);	# u64
-	&data_word(0x40c72493,0x32caab7b);	# u64
-	&data_word(0x15c9bebc,0x3c9ebe0a);	# u64
-	&data_word(0x9c100d4c,0x431d67c4);	# u64
-	&data_word(0xcb3e42b6,0x4cc5d4be);	# u64
-	&data_word(0xfc657e2a,0x597f299c);	# u64
-	&data_word(0x3ad6faec,0x5fcb6fab);	# u64
-	&data_word(0x4a475817,0x6c44198c);	# u64
-
-&function_end_B($func);
-
-&asm_finish();
diff --git a/crypto/sha/asm/sha512-x86_64.pl b/crypto/sha/asm/sha512-x86_64.pl
index b6252d31eca2..f611a2d898e2 100755
--- a/crypto/sha/asm/sha512-x86_64.pl
+++ b/crypto/sha/asm/sha512-x86_64.pl
@@ -40,14 +40,18 @@
 # sha256_block:-( This is presumably because 64-bit shifts/rotates
 # apparently are not atomic instructions, but implemented in microcode.
 
-$output=shift;
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
 
 $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
 ( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
 ( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
 die "can't locate x86_64-xlate.pl";
 
-open STDOUT,"| $^X $xlate $output";
+open STDOUT,"| $^X $xlate $flavour $output";
 
 if ($output =~ /512/) {
 	$func="sha512_block_data_order";
@@ -91,50 +95,44 @@ sub ROUND_00_15()
 { my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
 
 $code.=<<___;
-	mov	$e,$a0
-	mov	$e,$a1
+	ror	\$`$Sigma1[2]-$Sigma1[1]`,$a0
 	mov	$f,$a2
+	mov	$T1,`$SZ*($i&0xf)`(%rsp)
 
-	ror	\$$Sigma1[0],$a0
-	ror	\$$Sigma1[1],$a1
+	ror	\$`$Sigma0[2]-$Sigma0[1]`,$a1
+	xor	$e,$a0
 	xor	$g,$a2			# f^g
 
-	xor	$a1,$a0
-	ror	\$`$Sigma1[2]-$Sigma1[1]`,$a1
+	ror	\$`$Sigma1[1]-$Sigma1[0]`,$a0
+	add	$h,$T1			# T1+=h
+	xor	$a,$a1
+
+	add	($Tbl,$round,$SZ),$T1	# T1+=K[round]
 	and	$e,$a2			# (f^g)&e
-	mov	$T1,`$SZ*($i&0xf)`(%rsp)
+	mov	$b,$h
 
-	xor	$a1,$a0			# Sigma1(e)
+	ror	\$`$Sigma0[1]-$Sigma0[0]`,$a1
+	xor	$e,$a0
 	xor	$g,$a2			# Ch(e,f,g)=((f^g)&e)^g
-	add	$h,$T1			# T1+=h
-
-	mov	$a,$h
-	add	$a0,$T1			# T1+=Sigma1(e)
 
+	xor	$c,$h			# b^c
+	xor	$a,$a1
 	add	$a2,$T1			# T1+=Ch(e,f,g)
-	mov	$a,$a0
-	mov	$a,$a1
+	mov	$b,$a2
 
-	ror	\$$Sigma0[0],$h
-	ror	\$$Sigma0[1],$a0
-	mov	$a,$a2
-	add	($Tbl,$round,$SZ),$T1	# T1+=K[round]
+	ror	\$$Sigma1[0],$a0	# Sigma1(e)
+	and	$a,$h			# h=(b^c)&a
+	and	$c,$a2			# b&c
 
-	xor	$a0,$h
-	ror	\$`$Sigma0[2]-$Sigma0[1]`,$a0
-	or	$c,$a1			# a|c
+	ror	\$$Sigma0[0],$a1	# Sigma0(a)
+	add	$a0,$T1			# T1+=Sigma1(e)
+	add	$a2,$h			# h+=b&c (completes +=Maj(a,b,c)
 
-	xor	$a0,$h			# h=Sigma0(a)
-	and	$c,$a2			# a&c
 	add	$T1,$d			# d+=T1
-
-	and	$b,$a1			# (a|c)&b
 	add	$T1,$h			# h+=T1
-
-	or	$a2,$a1			# Maj(a,b,c)=((a|c)&b)|(a&c)
 	lea	1($round),$round	# round++
+	add	$a1,$h			# h+=Sigma0(a)
 
-	add	$a1,$h			# h+=Maj(a,b,c)
 ___
 }
 
@@ -143,32 +141,30 @@ sub ROUND_16_XX()
 
 $code.=<<___;
 	mov	`$SZ*(($i+1)&0xf)`(%rsp),$a0
-	mov	`$SZ*(($i+14)&0xf)`(%rsp),$T1
-
-	mov	$a0,$a2
+	mov	`$SZ*(($i+14)&0xf)`(%rsp),$a1
+	mov	$a0,$T1
+	mov	$a1,$a2
 
+	ror	\$`$sigma0[1]-$sigma0[0]`,$T1
+	xor	$a0,$T1
 	shr	\$$sigma0[2],$a0
-	ror	\$$sigma0[0],$a2
-
-	xor	$a2,$a0
-	ror	\$`$sigma0[1]-$sigma0[0]`,$a2
-
-	xor	$a2,$a0			# sigma0(X[(i+1)&0xf])
-	mov	$T1,$a1
 
-	shr	\$$sigma1[2],$T1
-	ror	\$$sigma1[0],$a1
+	ror	\$$sigma0[0],$T1
+	xor	$T1,$a0			# sigma0(X[(i+1)&0xf])
+	mov	`$SZ*(($i+9)&0xf)`(%rsp),$T1
 
-	xor	$a1,$T1
-	ror	\$`$sigma1[1]-$sigma1[0]`,$a1
-
-	xor	$a1,$T1			# sigma1(X[(i+14)&0xf])
+	ror	\$`$sigma1[1]-$sigma1[0]`,$a2
+	xor	$a1,$a2
+	shr	\$$sigma1[2],$a1
 
+	ror	\$$sigma1[0],$a2
 	add	$a0,$T1
-
-	add	`$SZ*(($i+9)&0xf)`(%rsp),$T1
+	xor	$a2,$a1			# sigma1(X[(i+14)&0xf])
 
 	add	`$SZ*($i&0xf)`(%rsp),$T1
+	mov	$e,$a0
+	add	$a1,$T1
+	mov	$a,$a1
 ___
 	&ROUND_00_15(@_);
 }
@@ -186,7 +182,7 @@ $func:
 	push	%r13
 	push	%r14
 	push	%r15
-	mov	%rsp,%rbp		# copy %rsp
+	mov	%rsp,%r11		# copy %rsp
 	shl	\$4,%rdx		# num*16
 	sub	\$$framesz,%rsp
 	lea	($inp,%rdx,$SZ),%rdx	# inp+num*16*$SZ
@@ -194,10 +190,10 @@ $func:
 	mov	$ctx,$_ctx		# save ctx, 1st arg
 	mov	$inp,$_inp		# save inp, 2nd arh
 	mov	%rdx,$_end		# save end pointer, "3rd" arg
-	mov	%rbp,$_rsp		# save copy of %rsp
+	mov	%r11,$_rsp		# save copy of %rsp
+.Lprologue:
 
-	.picmeup $Tbl
-	lea	$TABLE-.($Tbl),$Tbl
+	lea	$TABLE(%rip),$Tbl
 
 	mov	$SZ*0($ctx),$A
 	mov	$SZ*1($ctx),$B
@@ -215,6 +211,8 @@ $func:
 ___
 	for($i=0;$i<16;$i++) {
 		$code.="	mov	$SZ*$i($inp),$T1\n";
+		$code.="	mov	@ROT[4],$a0\n";
+		$code.="	mov	@ROT[0],$a1\n";
 		$code.="	bswap	$T1\n";
 		&ROUND_00_15($i,@ROT);
 		unshift(@ROT,pop(@ROT));
@@ -257,14 +255,15 @@ $code.=<<___;
 	mov	$H,$SZ*7($ctx)
 	jb	.Lloop
 
-	mov	$_rsp,%rsp
-	pop	%r15
-	pop	%r14
-	pop	%r13
-	pop	%r12
-	pop	%rbp
-	pop	%rbx
-
+	mov	$_rsp,%rsi
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lepilogue:
 	ret
 .size	$func,.-$func
 ___
@@ -339,6 +338,113 @@ $TABLE:
 ___
 }
 
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lprologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lprologue
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lepilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lepilogue
+	jae	.Lin_prologue
+
+	mov	16*$SZ+3*8(%rax),%rax	# pull $_rsp
+	lea	48(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	-40(%rax),%r14
+	mov	-48(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	se_handler,.-se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_$func
+	.rva	.LSEH_end_$func
+	.rva	.LSEH_info_$func
+
+.section	.xdata
+.align	8
+.LSEH_info_$func:
+	.byte	9,0,0,0
+	.rva	se_handler
+___
+}
+
 $code =~ s/\`([^\`]*)\`/eval $1/gem;
 print $code;
 close STDOUT;
diff --git a/crypto/sha/sha.h b/crypto/sha/sha.h
index 47a2c29f6627..8a6bf4bbbb1d 100644
--- a/crypto/sha/sha.h
+++ b/crypto/sha/sha.h
@@ -81,7 +81,7 @@ extern "C" {
  * !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  */
 
-#if defined(OPENSSL_SYS_WIN16) || defined(__LP32__)
+#if defined(__LP32__)
 #define SHA_LONG unsigned long
 #elif defined(OPENSSL_SYS_CRAY) || defined(__ILP64__)
 #define SHA_LONG unsigned long
@@ -116,6 +116,9 @@ unsigned char *SHA(const unsigned char *d, size_t n, unsigned char *md);
 void SHA_Transform(SHA_CTX *c, const unsigned char *data);
 #endif
 #ifndef OPENSSL_NO_SHA1
+#ifdef OPENSSL_FIPS
+int private_SHA1_Init(SHA_CTX *c);
+#endif
 int SHA1_Init(SHA_CTX *c);
 int SHA1_Update(SHA_CTX *c, const void *data, size_t len);
 int SHA1_Final(unsigned char *md, SHA_CTX *c);
@@ -138,6 +141,10 @@ typedef struct SHA256state_st
 	} SHA256_CTX;
 
 #ifndef OPENSSL_NO_SHA256
+#ifdef OPENSSL_FIPS
+int private_SHA224_Init(SHA256_CTX *c);
+int private_SHA256_Init(SHA256_CTX *c);
+#endif
 int SHA224_Init(SHA256_CTX *c);
 int SHA224_Update(SHA256_CTX *c, const void *data, size_t len);
 int SHA224_Final(unsigned char *md, SHA256_CTX *c);
@@ -185,6 +192,10 @@ typedef struct SHA512state_st
 #endif
 
 #ifndef OPENSSL_NO_SHA512
+#ifdef OPENSSL_FIPS
+int private_SHA384_Init(SHA512_CTX *c);
+int private_SHA512_Init(SHA512_CTX *c);
+#endif
 int SHA384_Init(SHA512_CTX *c);
 int SHA384_Update(SHA512_CTX *c, const void *data, size_t len);
 int SHA384_Final(unsigned char *md, SHA512_CTX *c);
diff --git a/crypto/sha/sha1_one.c b/crypto/sha/sha1_one.c
index 4831174198e7..7c65b60276cd 100644
--- a/crypto/sha/sha1_one.c
+++ b/crypto/sha/sha1_one.c
@@ -61,7 +61,7 @@
 #include <openssl/sha.h>
 #include <openssl/crypto.h>
 
-#if !defined(OPENSSL_NO_SHA1)
+#ifndef OPENSSL_NO_SHA1
 unsigned char *SHA1(const unsigned char *d, size_t n, unsigned char *md)
 	{
 	SHA_CTX c;
diff --git a/crypto/sha/sha1dgst.c b/crypto/sha/sha1dgst.c
index d31f0781a06a..81219af08834 100644
--- a/crypto/sha/sha1dgst.c
+++ b/crypto/sha/sha1dgst.c
@@ -57,16 +57,13 @@
  */
 
 #include <openssl/opensslconf.h>
+#include <openssl/crypto.h>
 #if !defined(OPENSSL_NO_SHA1) && !defined(OPENSSL_NO_SHA)
 
 #undef  SHA_0
 #define SHA_1
 
 #include <openssl/opensslv.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 
 const char SHA1_version[]="SHA1" OPENSSL_VERSION_PTEXT;
 
diff --git a/crypto/sha/sha1s.cpp b/crypto/sha/sha1s.cpp
deleted file mode 100644
index af23d1e0f218..000000000000
--- a/crypto/sha/sha1s.cpp
+++ /dev/null
@@ -1,82 +0,0 @@
-//
-// gettsc.inl
-//
-// gives access to the Pentium's (secret) cycle counter
-//
-// This software was written by Leonard Janke (janke@unixg.ubc.ca)
-// in 1996-7 and is entered, by him, into the public domain.
-
-#if defined(__WATCOMC__)
-void GetTSC(unsigned long&);
-#pragma aux GetTSC = 0x0f 0x31 "mov [edi], eax" parm [edi] modify [edx eax];
-#elif defined(__GNUC__)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  asm volatile(".byte 15, 49\n\t"
-	       : "=eax" (tsc)
-	       :
-	       : "%edx", "%eax");
-}
-#elif defined(_MSC_VER)
-inline
-void GetTSC(unsigned long& tsc)
-{
-  unsigned long a;
-  __asm _emit 0fh
-  __asm _emit 31h
-  __asm mov a, eax;
-  tsc=a;
-}
-#endif      
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <openssl/sha.h>
-
-#define sha1_block_x86 sha1_block_asm_data_order
-extern "C" {
-void sha1_block_x86(SHA_CTX *ctx, unsigned char *buffer,int num);
-}
-
-void main(int argc,char *argv[])
-	{
-	unsigned char buffer[64*256];
-	SHA_CTX ctx;
-	unsigned long s1,s2,e1,e2;
-	unsigned char k[16];
-	unsigned long data[2];
-	unsigned char iv[8];
-	int i,num=0,numm;
-	int j=0;
-
-	if (argc >= 2)
-		num=atoi(argv[1]);
-
-	if (num == 0) num=16;
-	if (num > 250) num=16;
-	numm=num+2;
-#if 0
-	num*=64;
-	numm*=64;
-#endif
-
-	for (j=0; j<6; j++)
-		{
-		for (i=0; i<10; i++) /**/
-			{
-			sha1_block_x86(&ctx,buffer,numm);
-			GetTSC(s1);
-			sha1_block_x86(&ctx,buffer,numm);
-			GetTSC(e1);
-			GetTSC(s2);
-			sha1_block_x86(&ctx,buffer,num);
-			GetTSC(e2);
-			sha1_block_x86(&ctx,buffer,num);
-			}
-
-		printf("sha1 (%d bytes) %d %d (%.2f)\n",num*64,
-			e1-s1,e2-s2,(double)((e1-s1)-(e2-s2))/2);
-		}
-	}
-
diff --git a/crypto/sha/sha256.c b/crypto/sha/sha256.c
index 3256a83e98e1..f88d3d6dadba 100644
--- a/crypto/sha/sha256.c
+++ b/crypto/sha/sha256.c
@@ -12,39 +12,29 @@
 
 #include <openssl/crypto.h>
 #include <openssl/sha.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 #include <openssl/opensslv.h>
 
 const char SHA256_version[]="SHA-256" OPENSSL_VERSION_PTEXT;
 
-int SHA224_Init (SHA256_CTX *c)
+fips_md_init_ctx(SHA224, SHA256)
 	{
-#ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
+	memset (c,0,sizeof(*c));
 	c->h[0]=0xc1059ed8UL;	c->h[1]=0x367cd507UL;
 	c->h[2]=0x3070dd17UL;	c->h[3]=0xf70e5939UL;
 	c->h[4]=0xffc00b31UL;	c->h[5]=0x68581511UL;
 	c->h[6]=0x64f98fa7UL;	c->h[7]=0xbefa4fa4UL;
-	c->Nl=0;	c->Nh=0;
-	c->num=0;	c->md_len=SHA224_DIGEST_LENGTH;
+	c->md_len=SHA224_DIGEST_LENGTH;
 	return 1;
 	}
 
-int SHA256_Init (SHA256_CTX *c)
+fips_md_init(SHA256)
 	{
-#ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
+	memset (c,0,sizeof(*c));
 	c->h[0]=0x6a09e667UL;	c->h[1]=0xbb67ae85UL;
 	c->h[2]=0x3c6ef372UL;	c->h[3]=0xa54ff53aUL;
 	c->h[4]=0x510e527fUL;	c->h[5]=0x9b05688cUL;
 	c->h[6]=0x1f83d9abUL;	c->h[7]=0x5be0cd19UL;
-	c->Nl=0;	c->Nh=0;
-	c->num=0;	c->md_len=SHA256_DIGEST_LENGTH;
+	c->md_len=SHA256_DIGEST_LENGTH;
 	return 1;
 	}
 
@@ -94,21 +84,21 @@ int SHA224_Final (unsigned char *md, SHA256_CTX *c)
  */
 #define	HASH_MAKE_STRING(c,s)	do {	\
 	unsigned long ll;		\
-	unsigned int  xn;		\
+	unsigned int  nn;		\
 	switch ((c)->md_len)		\
 	{   case SHA224_DIGEST_LENGTH:	\
-		for (xn=0;xn<SHA224_DIGEST_LENGTH/4;xn++)	\
-		{   ll=(c)->h[xn]; HOST_l2c(ll,(s));   }	\
+		for (nn=0;nn<SHA224_DIGEST_LENGTH/4;nn++)	\
+		{   ll=(c)->h[nn]; HOST_l2c(ll,(s));   }	\
 		break;			\
 	    case SHA256_DIGEST_LENGTH:	\
-		for (xn=0;xn<SHA256_DIGEST_LENGTH/4;xn++)	\
-		{   ll=(c)->h[xn]; HOST_l2c(ll,(s));   }	\
+		for (nn=0;nn<SHA256_DIGEST_LENGTH/4;nn++)	\
+		{   ll=(c)->h[nn]; HOST_l2c(ll,(s));   }	\
 		break;			\
 	    default:			\
 		if ((c)->md_len > SHA256_DIGEST_LENGTH)	\
 		    return 0;				\
-		for (xn=0;xn<(c)->md_len/4;xn++)		\
-		{   ll=(c)->h[xn]; HOST_l2c(ll,(s));   }	\
+		for (nn=0;nn<(c)->md_len/4;nn++)		\
+		{   ll=(c)->h[nn]; HOST_l2c(ll,(s));   }	\
 		break;			\
 	}				\
 	} while (0)
diff --git a/crypto/sha/sha512.c b/crypto/sha/sha512.c
index 9e91bcad0425..50dd7dc74437 100644
--- a/crypto/sha/sha512.c
+++ b/crypto/sha/sha512.c
@@ -5,10 +5,6 @@
  * ====================================================================
  */
 #include <openssl/opensslconf.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
 #if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA512)
 /*
  * IMPLEMENTATION NOTES.
@@ -63,11 +59,8 @@ const char SHA512_version[]="SHA-512" OPENSSL_VERSION_PTEXT;
 #define SHA512_BLOCK_CAN_MANAGE_UNALIGNED_DATA
 #endif
 
-int SHA384_Init (SHA512_CTX *c)
+fips_md_init_ctx(SHA384, SHA512)
 	{
-#ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
 	c->h[0]=U64(0xcbbb9d5dc1059ed8);
 	c->h[1]=U64(0x629a292a367cd507);
 	c->h[2]=U64(0x9159015a3070dd17);
@@ -76,16 +69,14 @@ int SHA384_Init (SHA512_CTX *c)
 	c->h[5]=U64(0x8eb44a8768581511);
 	c->h[6]=U64(0xdb0c2e0d64f98fa7);
 	c->h[7]=U64(0x47b5481dbefa4fa4);
+
         c->Nl=0;        c->Nh=0;
         c->num=0;       c->md_len=SHA384_DIGEST_LENGTH;
         return 1;
 	}
 
-int SHA512_Init (SHA512_CTX *c)
+fips_md_init(SHA512)
 	{
-#ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
 	c->h[0]=U64(0x6a09e667f3bcc908);
 	c->h[1]=U64(0xbb67ae8584caa73b);
 	c->h[2]=U64(0x3c6ef372fe94f82b);
@@ -94,6 +85,7 @@ int SHA512_Init (SHA512_CTX *c)
 	c->h[5]=U64(0x9b05688c2b3e6c1f);
 	c->h[6]=U64(0x1f83d9abfb41bd6b);
 	c->h[7]=U64(0x5be0cd19137e2179);
+
         c->Nl=0;        c->Nh=0;
         c->num=0;       c->md_len=SHA512_DIGEST_LENGTH;
         return 1;
@@ -204,7 +196,7 @@ int SHA512_Update (SHA512_CTX *c, const void *_data, size_t len)
 
 		if (len < n)
 			{
-			memcpy (p+c->num,data,len), c->num += len;
+			memcpy (p+c->num,data,len), c->num += (unsigned int)len;
 			return 1;
 			}
 		else	{
@@ -314,7 +306,7 @@ static const SHA_LONG64 K512[80] = {
 #ifndef PEDANTIC
 # if defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
 #  if defined(__x86_64) || defined(__x86_64__)
-#   define ROTR(a,n)	({ unsigned long ret;		\
+#   define ROTR(a,n)	({ SHA_LONG64 ret;		\
 				asm ("rorq %1,%0"	\
 				: "=r"(ret)		\
 				: "J"(n),"0"(a)		\
@@ -337,20 +329,21 @@ static const SHA_LONG64 K512[80] = {
 				((SHA_LONG64)hi)<<32|lo;	})
 #   else
 #    define PULL64(x) ({ const unsigned int *p=(const unsigned int *)(&(x));\
-			 unsigned int hi=p[0],lo=p[1];			\
+			 unsigned int hi=p[0],lo=p[1];		\
 				asm ("bswapl %0; bswapl %1;"	\
 				: "=r"(lo),"=r"(hi)		\
 				: "0"(lo),"1"(hi));		\
 				((SHA_LONG64)hi)<<32|lo;	})
 #   endif
 #  elif (defined(_ARCH_PPC) && defined(__64BIT__)) || defined(_ARCH_PPC64)
-#   define ROTR(a,n)	({ unsigned long ret;		\
+#   define ROTR(a,n)	({ SHA_LONG64 ret;		\
 				asm ("rotrdi %0,%1,%2"	\
 				: "=r"(ret)		\
 				: "r"(a),"K"(n)); ret;	})
 #  endif
 # elif defined(_MSC_VER)
 #  if defined(_WIN64)	/* applies to both IA-64 and AMD64 */
+#   pragma intrinsic(_rotr64)
 #   define ROTR(a,n)	_rotr64((a),n)
 #  endif
 #  if defined(_M_IX86) && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
@@ -398,15 +391,66 @@ static const SHA_LONG64 K512[80] = {
 #define Ch(x,y,z)	(((x) & (y)) ^ ((~(x)) & (z)))
 #define Maj(x,y,z)	(((x) & (y)) ^ ((x) & (z)) ^ ((y) & (z)))
 
-#if defined(OPENSSL_IA32_SSE2) && !defined(OPENSSL_NO_ASM) && !defined(I386_ONLY)
-#define	GO_FOR_SSE2(ctx,in,num)		do {		\
-	void	sha512_block_sse2(void *,const void *,size_t);	\
-	if (!(OPENSSL_ia32cap_P & (1<<26))) break;	\
-	sha512_block_sse2(ctx->h,in,num); return;	\
-					} while (0)
+
+#if defined(__i386) || defined(__i386__) || defined(_M_IX86)
+/*
+ * This code should give better results on 32-bit CPU with less than
+ * ~24 registers, both size and performance wise...
+ */
+static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num)
+	{
+	const SHA_LONG64 *W=in;
+	SHA_LONG64	A,E,T;
+	SHA_LONG64	X[9+80],*F;
+	int i;
+
+			while (num--) {
+
+	F    = X+80;
+	A    = ctx->h[0];	F[1] = ctx->h[1];
+	F[2] = ctx->h[2];	F[3] = ctx->h[3];
+	E    = ctx->h[4];	F[5] = ctx->h[5];
+	F[6] = ctx->h[6];	F[7] = ctx->h[7];
+
+	for (i=0;i<16;i++,F--)
+		{
+#ifdef B_ENDIAN
+		T = W[i];
+#else
+		T = PULL64(W[i]);
 #endif
+		F[0] = A;
+		F[4] = E;
+		F[8] = T;
+		T   += F[7] + Sigma1(E) + Ch(E,F[5],F[6]) + K512[i];
+		E    = F[3] + T;
+		A    = T + Sigma0(A) + Maj(A,F[1],F[2]);
+		}
+
+	for (;i<80;i++,F--)
+		{
+		T    = sigma0(F[8+16-1]);
+		T   += sigma1(F[8+16-14]);
+		T   += F[8+16] + F[8+16-9];
+
+		F[0] = A;
+		F[4] = E;
+		F[8] = T;
+		T   += F[7] + Sigma1(E) + Ch(E,F[5],F[6]) + K512[i];
+		E    = F[3] + T;
+		A    = T + Sigma0(A) + Maj(A,F[1],F[2]);
+		}
+
+	ctx->h[0] += A;		ctx->h[1] += F[1];
+	ctx->h[2] += F[2];	ctx->h[3] += F[3];
+	ctx->h[4] += E;		ctx->h[5] += F[5];
+	ctx->h[6] += F[6];	ctx->h[7] += F[7];
+
+			W+=SHA_LBLOCK;
+			}
+	}
 
-#ifdef OPENSSL_SMALL_FOOTPRINT
+#elif defined(OPENSSL_SMALL_FOOTPRINT)
 
 static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num)
 	{
@@ -415,10 +459,6 @@ static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num
 	SHA_LONG64	X[16];
 	int i;
 
-#ifdef GO_FOR_SSE2
-	GO_FOR_SSE2(ctx,in,num);
-#endif
-
 			while (num--) {
 
 	a = ctx->h[0];	b = ctx->h[1];	c = ctx->h[2];	d = ctx->h[3];
@@ -463,11 +503,11 @@ static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num
 	h = Sigma0(a) + Maj(a,b,c);			\
 	d += T1;	h += T1;		} while (0)
 
-#define	ROUND_16_80(i,a,b,c,d,e,f,g,h,X)	do {	\
-	s0 = X[(i+1)&0x0f];	s0 = sigma0(s0);	\
-	s1 = X[(i+14)&0x0f];	s1 = sigma1(s1);	\
-	T1 = X[(i)&0x0f] += s0 + s1 + X[(i+9)&0x0f];	\
-	ROUND_00_15(i,a,b,c,d,e,f,g,h);		} while (0)
+#define	ROUND_16_80(i,j,a,b,c,d,e,f,g,h,X)	do {	\
+	s0 = X[(j+1)&0x0f];	s0 = sigma0(s0);	\
+	s1 = X[(j+14)&0x0f];	s1 = sigma1(s1);	\
+	T1 = X[(j)&0x0f] += s0 + s1 + X[(j+9)&0x0f];	\
+	ROUND_00_15(i+j,a,b,c,d,e,f,g,h);		} while (0)
 
 static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num)
 	{
@@ -476,10 +516,6 @@ static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num
 	SHA_LONG64	X[16];
 	int i;
 
-#ifdef GO_FOR_SSE2
-	GO_FOR_SSE2(ctx,in,num);
-#endif
-
 			while (num--) {
 
 	a = ctx->h[0];	b = ctx->h[1];	c = ctx->h[2];	d = ctx->h[3];
@@ -521,16 +557,24 @@ static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num
 	T1 = X[15] = PULL64(W[15]);	ROUND_00_15(15,b,c,d,e,f,g,h,a);
 #endif
 
-	for (i=16;i<80;i+=8)
+	for (i=16;i<80;i+=16)
 		{
-		ROUND_16_80(i+0,a,b,c,d,e,f,g,h,X);
-		ROUND_16_80(i+1,h,a,b,c,d,e,f,g,X);
-		ROUND_16_80(i+2,g,h,a,b,c,d,e,f,X);
-		ROUND_16_80(i+3,f,g,h,a,b,c,d,e,X);
-		ROUND_16_80(i+4,e,f,g,h,a,b,c,d,X);
-		ROUND_16_80(i+5,d,e,f,g,h,a,b,c,X);
-		ROUND_16_80(i+6,c,d,e,f,g,h,a,b,X);
-		ROUND_16_80(i+7,b,c,d,e,f,g,h,a,X);
+		ROUND_16_80(i, 0,a,b,c,d,e,f,g,h,X);
+		ROUND_16_80(i, 1,h,a,b,c,d,e,f,g,X);
+		ROUND_16_80(i, 2,g,h,a,b,c,d,e,f,X);
+		ROUND_16_80(i, 3,f,g,h,a,b,c,d,e,X);
+		ROUND_16_80(i, 4,e,f,g,h,a,b,c,d,X);
+		ROUND_16_80(i, 5,d,e,f,g,h,a,b,c,X);
+		ROUND_16_80(i, 6,c,d,e,f,g,h,a,b,X);
+		ROUND_16_80(i, 7,b,c,d,e,f,g,h,a,X);
+		ROUND_16_80(i, 8,a,b,c,d,e,f,g,h,X);
+		ROUND_16_80(i, 9,h,a,b,c,d,e,f,g,X);
+		ROUND_16_80(i,10,g,h,a,b,c,d,e,f,X);
+		ROUND_16_80(i,11,f,g,h,a,b,c,d,e,X);
+		ROUND_16_80(i,12,e,f,g,h,a,b,c,d,X);
+		ROUND_16_80(i,13,d,e,f,g,h,a,b,c,X);
+		ROUND_16_80(i,14,c,d,e,f,g,h,a,b,X);
+		ROUND_16_80(i,15,b,c,d,e,f,g,h,a,X);
 		}
 
 	ctx->h[0] += a;	ctx->h[1] += b;	ctx->h[2] += c;	ctx->h[3] += d;
@@ -544,13 +588,10 @@ static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num
 
 #endif /* SHA512_ASM */
 
-#else /* OPENSSL_NO_SHA512 */
-
-/* Sensitive compilers ("Compaq C V6.4-005 on OpenVMS VAX V7.3", for
- * example) dislike a statement-free file, complaining:
- * "%CC-W-EMPTYFILE, Source file does not contain any declarations."
- */
+#else /* !OPENSSL_NO_SHA512 */
 
-int sha512_dummy();
+#if defined(PEDANTIC) || defined(__DECC) || defined(OPENSSL_SYS_MACOSX)
+static void *dummy=&dummy;
+#endif
 
-#endif /* OPENSSL_NO_SHA512 */
+#endif /* !OPENSSL_NO_SHA512 */
diff --git a/crypto/sha/sha_dgst.c b/crypto/sha/sha_dgst.c
index 598f4d721af9..c946ad827d58 100644
--- a/crypto/sha/sha_dgst.c
+++ b/crypto/sha/sha_dgst.c
@@ -58,11 +58,6 @@
 
 #include <openssl/opensslconf.h>
 #include <openssl/crypto.h>
-#ifdef OPENSSL_FIPS
-#include <openssl/fips.h>
-#endif
-
-#include <openssl/err.h>
 #if !defined(OPENSSL_NO_SHA0) && !defined(OPENSSL_NO_SHA)
 
 #undef  SHA_1
diff --git a/crypto/sha/sha_locl.h b/crypto/sha/sha_locl.h
index da46ddfe7943..7a0c3ca8d8ce 100644
--- a/crypto/sha/sha_locl.h
+++ b/crypto/sha/sha_locl.h
@@ -122,23 +122,18 @@ void sha1_block_data_order (SHA_CTX *c, const void *p,size_t num);
 #define INIT_DATA_h3 0x10325476UL
 #define INIT_DATA_h4 0xc3d2e1f0UL
 
-#if defined(SHA_0) && defined(OPENSSL_FIPS)
-FIPS_NON_FIPS_MD_Init(SHA)
+#ifdef SHA_0
+fips_md_init(SHA)
 #else
-int HASH_INIT (SHA_CTX *c)
+fips_md_init_ctx(SHA1, SHA)
 #endif
 	{
-#if defined(SHA_1) && defined(OPENSSL_FIPS)
-	FIPS_selftest_check();
-#endif
+	memset (c,0,sizeof(*c));
 	c->h0=INIT_DATA_h0;
 	c->h1=INIT_DATA_h1;
 	c->h2=INIT_DATA_h2;
 	c->h3=INIT_DATA_h3;
 	c->h4=INIT_DATA_h4;
-	c->Nl=0;
-	c->Nh=0;
-	c->num=0;
 	return 1;
 	}
 
diff --git a/crypto/sha/shatest.c b/crypto/sha/shatest.c
index ed0fe06a7bee..27614646d154 100644
--- a/crypto/sha/shatest.c
+++ b/crypto/sha/shatest.c
@@ -123,9 +123,9 @@ int main(int argc, char *argv[])
 	i=1;
 	while (*P != NULL)
 		{
-		EVP_Digest(*P,strlen((char *)*P),md,NULL,EVP_sha(), NULL);
+		EVP_Digest(*P,strlen(*P),md,NULL,EVP_sha(), NULL);
 		p=pt(md);
-		if (strcmp(p,(char *)*R) != 0)
+		if (strcmp(p,*R) != 0)
 			{
 			printf("error calculating SHA on '%s'\n",*P);
 			printf("got %s instead of %s\n",p,*R);
diff --git a/crypto/sparccpuid.S b/crypto/sparccpuid.S
index c17350fc89e0..ae61f7f5cedb 100644
--- a/crypto/sparccpuid.S
+++ b/crypto/sparccpuid.S
@@ -34,7 +34,8 @@ OPENSSL_wipe_cpu:
 	nop
 	call	.PIC.zero.up
 	mov	.zero-(.-4),%o0
-	ldd	[%o0],%f0
+	ld	[%o0],%f0
+	ld	[%o0],%f1
 
 	subcc	%g0,1,%o0
 	! Following is V9 "rd %ccr,%o0" instruction. However! V8
@@ -166,6 +167,7 @@ walk_reg_wins:
 
 .global	OPENSSL_atomic_add
 .type	OPENSSL_atomic_add,#function
+.align	32
 OPENSSL_atomic_add:
 #ifndef ABI64
 	subcc	%g0,1,%o2
@@ -177,7 +179,7 @@ OPENSSL_atomic_add:
 	ba	.enter
 	nop
 #ifdef __sun
-! Note that you don't have to link with libthread to call thr_yield,
+! Note that you do not have to link with libthread to call thr_yield,
 ! as libc provides a stub, which is overloaded the moment you link
 ! with *either* libpthread or libthread...
 #define	YIELD_CPU	thr_yield
@@ -213,27 +215,188 @@ OPENSSL_atomic_add:
 	sra	%o0,%g0,%o0	! we return signed int, remember?
 .size	OPENSSL_atomic_add,.-OPENSSL_atomic_add
 
-.global	OPENSSL_rdtsc
+.global	_sparcv9_rdtick
+.align	32
+_sparcv9_rdtick:
 	subcc	%g0,1,%o0
 	.word	0x91408000	!rd	%ccr,%o0
 	cmp	%o0,0x99
-	bne	.notsc
+	bne	.notick
 	xor	%o0,%o0,%o0
-	save	%sp,FRAME-16,%sp
-	mov	513,%o0		!SI_PLATFORM
-	add	%sp,BIAS+16,%o1
-	call	sysinfo
-	mov	256,%o2
-
-	add	%sp,BIAS-16,%o1
-	ld	[%o1],%l0
-	ld	[%o1+4],%l1
-	ld	[%o1+8],%l2
-	mov	%lo('SUNW'),%l3
-	ret
-	restore
-.notsc:
+	.word	0x91410000	!rd	%tick,%o0
+	retl
+	.word	0x93323020	!srlx	%o0,32,%o1
+.notick:
+	retl
+	xor	%o1,%o1,%o1
+.type	_sparcv9_rdtick,#function
+.size	_sparcv9_rdtick,.-_sparcv9_rdtick
+
+.global	_sparcv9_vis1_probe
+.align	8
+_sparcv9_vis1_probe:
+	.word	0x81b00d80	!fxor	%f0,%f0,%f0
+	add	%sp,BIAS+2,%o1
+	retl
+	.word	0xc19a5a40	!ldda	[%o1]ASI_FP16_P,%f0
+.type	_sparcv9_vis1_probe,#function
+.size	_sparcv9_vis1_probe,.-_sparcv9_vis1_probe
+
+! Probe and instrument VIS1 instruction. Output is number of cycles it
+! takes to execute rdtick and pair of VIS1 instructions. US-Tx VIS unit
+! is slow (documented to be 6 cycles on T2) and the core is in-order
+! single-issue, it should be possible to distinguish Tx reliably...
+! Observed return values are:
+!
+!	UltraSPARC IIe		7
+!	UltraSPARC III		7
+!	UltraSPARC T1		24
+!
+! Numbers for T2 and SPARC64 V-VII are more than welcomed.
+!
+! It would be possible to detect specifically US-T1 by instrumenting
+! fmul8ulx16, which is emulated on T1 and as such accounts for quite
+! a lot of %tick-s, couple of thousand on Linux...
+.global	_sparcv9_vis1_instrument
+.align	8
+_sparcv9_vis1_instrument:
+	.word	0x91410000	!rd	%tick,%o0
+	.word	0x81b00d80	!fxor	%f0,%f0,%f0
+	.word	0x85b08d82	!fxor	%f2,%f2,%f2
+	.word	0x93410000	!rd	%tick,%o1
+	.word	0x81b00d80	!fxor	%f0,%f0,%f0
+	.word	0x85b08d82	!fxor	%f2,%f2,%f2
+	.word	0x95410000	!rd	%tick,%o2
+	.word	0x81b00d80	!fxor	%f0,%f0,%f0
+	.word	0x85b08d82	!fxor	%f2,%f2,%f2
+	.word	0x97410000	!rd	%tick,%o3
+	.word	0x81b00d80	!fxor	%f0,%f0,%f0
+	.word	0x85b08d82	!fxor	%f2,%f2,%f2
+	.word	0x99410000	!rd	%tick,%o4
+
+	! calculate intervals
+	sub	%o1,%o0,%o0
+	sub	%o2,%o1,%o1
+	sub	%o3,%o2,%o2
+	sub	%o4,%o3,%o3
+
+	! find minumum value
+	cmp	%o0,%o1
+	.word	0x38680002	!bgu,a	%xcc,.+8
+	mov	%o1,%o0
+	cmp	%o0,%o2
+	.word	0x38680002	!bgu,a	%xcc,.+8
+	mov	%o2,%o0
+	cmp	%o0,%o3
+	.word	0x38680002	!bgu,a	%xcc,.+8
+	mov	%o3,%o0
+
+	retl
+	nop
+.type	_sparcv9_vis1_instrument,#function
+.size	_sparcv9_vis1_instrument,.-_sparcv9_vis1_instrument
+
+.global	_sparcv9_vis2_probe
+.align	8
+_sparcv9_vis2_probe:
+	retl
+	.word	0x81b00980	!bshuffle	%f0,%f0,%f0
+.type	_sparcv9_vis2_probe,#function
+.size	_sparcv9_vis2_probe,.-_sparcv9_vis2_probe
+
+.global	_sparcv9_fmadd_probe
+.align	8
+_sparcv9_fmadd_probe:
+	.word	0x81b00d80	!fxor	%f0,%f0,%f0
+	.word	0x85b08d82	!fxor	%f2,%f2,%f2
 	retl
+	.word	0x81b80440	!fmaddd	%f0,%f0,%f2,%f0
+.type	_sparcv9_fmadd_probe,#function
+.size	_sparcv9_fmadd_probe,.-_sparcv9_fmadd_probe
+
+.global	OPENSSL_cleanse
+.align	32
+OPENSSL_cleanse:
+	cmp	%o1,14
+	nop
+#ifdef ABI64
+	bgu	%xcc,.Lot
+#else
+	bgu	.Lot
+#endif
+	cmp	%o1,0
+	bne	.Little
+	nop
+	retl
+	nop
+
+.Little:
+	stb	%g0,[%o0]
+	subcc	%o1,1,%o1
+	bnz	.Little
+	add	%o0,1,%o0
+	retl
+	nop
+.align	32
+.Lot:
+#ifndef ABI64
+	subcc	%g0,1,%g1
+	! see above for explanation
+	.word	0x83408000	!rd	%ccr,%g1
+	cmp	%g1,0x99
+	bne	.v8lot
+	nop
+#endif
+
+.v9lot:	andcc	%o0,7,%g0
+	bz	.v9aligned
+	nop
+	stb	%g0,[%o0]
+	sub	%o1,1,%o1
+	ba	.v9lot
+	add	%o0,1,%o0
+.align	16,0x01000000
+.v9aligned:
+	.word	0xc0720000	!stx	%g0,[%o0]
+	sub	%o1,8,%o1
+	andcc	%o1,-8,%g0
+#ifdef ABI64
+	.word	0x126ffffd	!bnz	%xcc,.v9aligned
+#else
+	.word	0x124ffffd	!bnz	%icc,.v9aligned
+#endif
+	add	%o0,8,%o0
+
+	cmp	%o1,0
+	bne	.Little
+	nop
+	retl
+	nop
+#ifndef ABI64
+.v8lot:	andcc	%o0,3,%g0
+	bz	.v8aligned
+	nop
+	stb	%g0,[%o0]
+	sub	%o1,1,%o1
+	ba	.v8lot
+	add	%o0,1,%o0
+	nop
+.v8aligned:
+	st	%g0,[%o0]
+	sub	%o1,4,%o1
+	andcc	%o1,-4,%g0
+	bnz	.v8aligned
+	add	%o0,4,%o0
+
+	cmp	%o1,0
+	bne	.Little
+	nop
+	retl
+	nop
+#endif
+.type	OPENSSL_cleanse,#function
+.size	OPENSSL_cleanse,.-OPENSSL_cleanse
+
+.section	".init",#alloc,#execinstr
+	call	OPENSSL_cpuid_setup
 	nop
-.type	OPENSSL_rdtsc,#function
-.size	OPENSSL_rdtsc,.-OPENSSL_atomic_add
diff --git a/crypto/sparcv9cap.c b/crypto/sparcv9cap.c
new file mode 100644
index 000000000000..43b3ac6f81c6
--- /dev/null
+++ b/crypto/sparcv9cap.c
@@ -0,0 +1,237 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <setjmp.h>
+#include <signal.h>
+#include <sys/time.h>
+#include <openssl/bn.h>
+
+#define SPARCV9_TICK_PRIVILEGED	(1<<0)
+#define SPARCV9_PREFER_FPU	(1<<1)
+#define SPARCV9_VIS1		(1<<2)
+#define SPARCV9_VIS2		(1<<3)	/* reserved */
+#define SPARCV9_FMADD		(1<<4)	/* reserved for SPARC64 V */
+
+static int OPENSSL_sparcv9cap_P=SPARCV9_TICK_PRIVILEGED;
+
+int bn_mul_mont(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np,const BN_ULONG *n0, int num)
+	{
+	int bn_mul_mont_fpu(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np,const BN_ULONG *n0, int num);
+	int bn_mul_mont_int(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np,const BN_ULONG *n0, int num);
+
+	if (num>=8 && !(num&1) &&
+	    (OPENSSL_sparcv9cap_P&(SPARCV9_PREFER_FPU|SPARCV9_VIS1)) ==
+		(SPARCV9_PREFER_FPU|SPARCV9_VIS1))
+		return bn_mul_mont_fpu(rp,ap,bp,np,n0,num);
+	else
+		return bn_mul_mont_int(rp,ap,bp,np,n0,num);
+	}
+
+unsigned long	_sparcv9_rdtick(void);
+void		_sparcv9_vis1_probe(void);
+unsigned long	_sparcv9_vis1_instrument(void);
+void		_sparcv9_vis2_probe(void);
+void		_sparcv9_fmadd_probe(void);
+
+unsigned long OPENSSL_rdtsc(void)
+	{
+	if (OPENSSL_sparcv9cap_P&SPARCV9_TICK_PRIVILEGED)
+#if defined(__sun) && defined(__SVR4)
+		return gethrtime();
+#else
+		return 0;
+#endif
+	else
+		return _sparcv9_rdtick();
+	}
+
+#if 0 && defined(__sun) && defined(__SVR4)
+/* This code path is disabled, because of incompatibility of
+ * libdevinfo.so.1 and libmalloc.so.1 (see below for details)
+ */
+#include <malloc.h>
+#include <dlfcn.h>
+#include <libdevinfo.h>
+#include <sys/systeminfo.h>
+
+typedef di_node_t (*di_init_t)(const char *,uint_t);
+typedef void      (*di_fini_t)(di_node_t);
+typedef char *    (*di_node_name_t)(di_node_t);
+typedef int       (*di_walk_node_t)(di_node_t,uint_t,di_node_name_t,int (*)(di_node_t,di_node_name_t));
+
+#define DLLINK(h,name) (name=(name##_t)dlsym((h),#name))
+
+static int walk_nodename(di_node_t node, di_node_name_t di_node_name)
+	{
+	char *name = (*di_node_name)(node);
+
+	/* This is expected to catch all UltraSPARC flavors prior T1 */
+	if (!strcmp (name,"SUNW,UltraSPARC") ||
+	    !strncmp(name,"SUNW,UltraSPARC-I",17))  /* covers II,III,IV */
+		{
+		OPENSSL_sparcv9cap_P |= SPARCV9_PREFER_FPU|SPARCV9_VIS1;
+
+		/* %tick is privileged only on UltraSPARC-I/II, but not IIe */
+		if (name[14]!='\0' && name[17]!='\0' && name[18]!='\0')
+			OPENSSL_sparcv9cap_P &= ~SPARCV9_TICK_PRIVILEGED;
+
+		return DI_WALK_TERMINATE;
+		}
+	/* This is expected to catch remaining UltraSPARCs, such as T1 */
+	else if (!strncmp(name,"SUNW,UltraSPARC",15))
+		{
+		OPENSSL_sparcv9cap_P &= ~SPARCV9_TICK_PRIVILEGED;
+
+		return DI_WALK_TERMINATE;
+		}
+
+	return DI_WALK_CONTINUE;
+	}
+
+void OPENSSL_cpuid_setup(void)
+	{
+	void *h;
+	char *e,si[256];
+	static int trigger=0;
+
+	if (trigger) return;
+	trigger=1;
+
+	if ((e=getenv("OPENSSL_sparcv9cap")))
+		{
+		OPENSSL_sparcv9cap_P=strtoul(e,NULL,0);
+		return;
+		}
+
+	if (sysinfo(SI_MACHINE,si,sizeof(si))>0)
+		{
+		if (strcmp(si,"sun4v"))
+			/* FPU is preferred for all CPUs, but US-T1/2 */
+			OPENSSL_sparcv9cap_P |= SPARCV9_PREFER_FPU;
+		}
+
+	if (sysinfo(SI_ISALIST,si,sizeof(si))>0)
+		{
+		if (strstr(si,"+vis"))
+			OPENSSL_sparcv9cap_P |= SPARCV9_VIS1;
+		if (strstr(si,"+vis2"))
+			{
+			OPENSSL_sparcv9cap_P |= SPARCV9_VIS2;
+			OPENSSL_sparcv9cap_P &= ~SPARCV9_TICK_PRIVILEGED;
+			return;
+			}
+		}
+#ifdef M_KEEP
+	/*
+	 * Solaris libdevinfo.so.1 is effectively incomatible with
+	 * libmalloc.so.1. Specifically, if application is linked with
+	 * -lmalloc, it crashes upon startup with SIGSEGV in
+	 * free(3LIBMALLOC) called by di_fini. Prior call to
+	 * mallopt(M_KEEP,0) somehow helps... But not always...
+	 */
+	if ((h = dlopen(NULL,RTLD_LAZY)))
+		{
+		union { void *p; int (*f)(int,int); } sym;
+		if ((sym.p = dlsym(h,"mallopt"))) (*sym.f)(M_KEEP,0);
+		dlclose(h);
+		}
+#endif
+	if ((h = dlopen("libdevinfo.so.1",RTLD_LAZY))) do
+		{
+		di_init_t	di_init;
+		di_fini_t	di_fini;
+		di_walk_node_t	di_walk_node;
+		di_node_name_t	di_node_name;
+		di_node_t	root_node;
+
+		if (!DLLINK(h,di_init))		break;
+		if (!DLLINK(h,di_fini))		break;
+		if (!DLLINK(h,di_walk_node))	break;
+		if (!DLLINK(h,di_node_name))	break;
+
+		if ((root_node = (*di_init)("/",DINFOSUBTREE))!=DI_NODE_NIL)
+			{
+			(*di_walk_node)(root_node,DI_WALK_SIBFIRST,
+					di_node_name,walk_nodename);
+			(*di_fini)(root_node);
+			}
+		} while(0);
+
+	if (h) dlclose(h);
+	}
+
+#else
+
+static sigjmp_buf common_jmp;
+static void common_handler(int sig) { siglongjmp(common_jmp,sig); }
+
+void OPENSSL_cpuid_setup(void)
+	{
+	char *e;
+	struct sigaction	common_act,ill_oact,bus_oact;
+	sigset_t		all_masked,oset;
+	static int trigger=0;
+
+	if (trigger) return;
+	trigger=1;
+ 
+	if ((e=getenv("OPENSSL_sparcv9cap")))
+		{
+		OPENSSL_sparcv9cap_P=strtoul(e,NULL,0);
+		return;
+		}
+
+	/* Initial value, fits UltraSPARC-I&II... */
+	OPENSSL_sparcv9cap_P = SPARCV9_PREFER_FPU|SPARCV9_TICK_PRIVILEGED;
+
+	sigfillset(&all_masked);
+	sigdelset(&all_masked,SIGILL);
+	sigdelset(&all_masked,SIGTRAP);
+#ifdef SIGEMT
+	sigdelset(&all_masked,SIGEMT);
+#endif
+	sigdelset(&all_masked,SIGFPE);
+	sigdelset(&all_masked,SIGBUS);
+	sigdelset(&all_masked,SIGSEGV);
+	sigprocmask(SIG_SETMASK,&all_masked,&oset);
+
+	memset(&common_act,0,sizeof(common_act));
+	common_act.sa_handler = common_handler;
+	common_act.sa_mask    = all_masked;
+
+	sigaction(SIGILL,&common_act,&ill_oact);
+	sigaction(SIGBUS,&common_act,&bus_oact);/* T1 fails 16-bit ldda [on Linux] */
+
+	if (sigsetjmp(common_jmp,1) == 0)
+		{
+		_sparcv9_rdtick();
+		OPENSSL_sparcv9cap_P &= ~SPARCV9_TICK_PRIVILEGED;
+		}
+
+	if (sigsetjmp(common_jmp,1) == 0)
+		{
+		_sparcv9_vis1_probe();
+		OPENSSL_sparcv9cap_P |= SPARCV9_VIS1;
+		/* detect UltraSPARC-Tx, see sparccpud.S for details... */
+		if (_sparcv9_vis1_instrument() >= 12)
+			OPENSSL_sparcv9cap_P &= ~(SPARCV9_VIS1|SPARCV9_PREFER_FPU);
+		else
+			{
+			_sparcv9_vis2_probe();
+			OPENSSL_sparcv9cap_P |= SPARCV9_VIS2;
+			}
+		}
+
+	if (sigsetjmp(common_jmp,1) == 0)
+		{
+		_sparcv9_fmadd_probe();
+		OPENSSL_sparcv9cap_P |= SPARCV9_FMADD;
+		}
+
+	sigaction(SIGBUS,&bus_oact,NULL);
+	sigaction(SIGILL,&ill_oact,NULL);
+
+	sigprocmask(SIG_SETMASK,&oset,NULL);
+	}
+
+#endif
diff --git a/crypto/srp/Makefile b/crypto/srp/Makefile
new file mode 100644
index 000000000000..41859d46fa7f
--- /dev/null
+++ b/crypto/srp/Makefile
@@ -0,0 +1,98 @@
+DIR=	srp
+TOP=	../..
+CC=	cc
+INCLUDES= -I.. -I$(TOP) -I../../include
+CFLAG=-g
+INSTALL_PREFIX=
+OPENSSLDIR=     /usr/local/ssl
+INSTALLTOP=/usr/local/ssl
+MAKE=		make -f Makefile.ssl
+MAKEDEPPROG=	makedepend
+MAKEDEPEND=	$(TOP)/util/domd $(TOP) -MD $(MAKEDEPPROG)
+MAKEFILE=	Makefile.ssl
+AR=		ar r
+
+CFLAGS= $(INCLUDES) $(CFLAG)
+
+GENERAL=Makefile
+TEST=srptest.c
+APPS=
+
+LIB=$(TOP)/libcrypto.a
+LIBSRC=srp_lib.c srp_vfy.c
+LIBOBJ=srp_lib.o srp_vfy.o
+
+SRC= $(LIBSRC)
+
+EXHEADER= srp.h
+HEADER=	$(EXHEADER)
+
+top:
+	(cd ../..; $(MAKE) DIRS=crypto SDIRS=$(DIR) sub_all)
+
+all:	lib
+
+lib:	$(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
+	$(RANLIB) $(LIB) || echo Never mind.
+	@touch lib
+
+links:
+	@$(PERL) $(TOP)/util/mklink.pl ../../include/openssl $(EXHEADER)
+	@$(PERL) $(TOP)/util/mklink.pl ../../test $(TEST)
+	@$(PERL) $(TOP)/util/mklink.pl ../../apps $(APPS)
+
+install:
+	@[ -n "$(INSTALLTOP)" ] # should be set by top Makefile...
+	@headerlist="$(EXHEADER)"; for i in $$headerlist ; \
+	do  \
+	(cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i; \
+	chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i ); \
+	done;
+
+tags:
+	ctags $(SRC)
+
+tests:
+
+srptest: top srptest.c $(LIB)
+	$(CC) $(CFLAGS) -Wall -Werror -g -o srptest srptest.c $(LIB)
+
+lint:
+	lint -DLINT $(INCLUDES) $(SRC)>fluff
+
+depend:
+	$(MAKEDEPEND) -- $(CFLAG) $(INCLUDES) $(DEPFLAG) -- $(PROGS) $(LIBSRC)
+
+dclean:
+	$(PERL) -pe 'if (/^# DO NOT DELETE THIS LINE/) {print; exit(0);}' $(MAKEFILE) >Makefile.new
+	mv -f Makefile.new $(MAKEFILE)
+
+clean:
+	rm -f *.o *.obj lib tags core .pure .nfs* *.old *.bak fluff
+
+# DO NOT DELETE THIS LINE -- make depend depends on it.
+
+srp_lib.o: ../../e_os.h ../../include/openssl/asn1.h
+srp_lib.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+srp_lib.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+srp_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
+srp_lib.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+srp_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+srp_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+srp_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+srp_lib.o: ../../include/openssl/sha.h ../../include/openssl/srp.h
+srp_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+srp_lib.o: ../cryptlib.h srp_grps.h srp_lcl.h srp_lib.c
+srp_vfy.o: ../../e_os.h ../../include/openssl/asn1.h
+srp_vfy.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+srp_vfy.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+srp_vfy.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
+srp_vfy.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+srp_vfy.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+srp_vfy.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+srp_vfy.o: ../../include/openssl/ossl_typ.h ../../include/openssl/rand.h
+srp_vfy.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+srp_vfy.o: ../../include/openssl/srp.h ../../include/openssl/stack.h
+srp_vfy.o: ../../include/openssl/symhacks.h ../../include/openssl/txt_db.h
+srp_vfy.o: ../cryptlib.h srp_lcl.h srp_vfy.c
diff --git a/crypto/srp/srp.h b/crypto/srp/srp.h
new file mode 100644
index 000000000000..7ec7825cade0
--- /dev/null
+++ b/crypto/srp/srp.h
@@ -0,0 +1,172 @@
+/* crypto/srp/srp.h */
+/* Written by Christophe Renou (christophe.renou@edelweb.fr) with 
+ * the precious help of Peter Sylvester (peter.sylvester@edelweb.fr) 
+ * for the EdelKey project and contributed to the OpenSSL project 2004.
+ */
+/* ====================================================================
+ * Copyright (c) 2004 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+#ifndef __SRP_H__
+#define __SRP_H__
+
+#ifndef OPENSSL_NO_SRP
+
+#include <stdio.h>
+#include <string.h>
+
+#ifdef  __cplusplus
+extern "C" {
+#endif
+
+#include <openssl/safestack.h>
+#include <openssl/bn.h>
+#include <openssl/crypto.h>
+
+typedef struct SRP_gN_cache_st
+	{
+	char *b64_bn;
+	BIGNUM *bn;
+	} SRP_gN_cache;
+
+
+DECLARE_STACK_OF(SRP_gN_cache)
+
+typedef struct SRP_user_pwd_st
+	{
+	char *id;
+	BIGNUM *s;
+	BIGNUM *v;
+	const BIGNUM *g;
+	const BIGNUM *N;
+	char *info;
+	} SRP_user_pwd;
+
+DECLARE_STACK_OF(SRP_user_pwd)
+
+typedef struct SRP_VBASE_st
+	{
+	STACK_OF(SRP_user_pwd) *users_pwd;
+	STACK_OF(SRP_gN_cache) *gN_cache;
+/* to simulate a user */
+	char *seed_key;
+	BIGNUM *default_g;
+	BIGNUM *default_N;
+	} SRP_VBASE;
+
+
+/*Structure interne pour retenir les couples N et g*/
+typedef struct SRP_gN_st
+	{
+	char *id;
+	BIGNUM *g;
+	BIGNUM *N;
+	} SRP_gN;
+
+DECLARE_STACK_OF(SRP_gN)
+
+SRP_VBASE *SRP_VBASE_new(char *seed_key);
+int SRP_VBASE_free(SRP_VBASE *vb);
+int SRP_VBASE_init(SRP_VBASE *vb, char * verifier_file);
+SRP_user_pwd *SRP_VBASE_get_by_user(SRP_VBASE *vb, char *username);
+char *SRP_create_verifier(const char *user, const char *pass, char **salt,
+			  char **verifier, const char *N, const char *g);
+int SRP_create_verifier_BN(const char *user, const char *pass, BIGNUM **salt, BIGNUM **verifier, BIGNUM *N, BIGNUM *g);
+
+
+#define SRP_NO_ERROR 0
+#define SRP_ERR_VBASE_INCOMPLETE_FILE 1
+#define SRP_ERR_VBASE_BN_LIB 2
+#define SRP_ERR_OPEN_FILE 3
+#define SRP_ERR_MEMORY 4
+
+#define DB_srptype	0
+#define DB_srpverifier	1
+#define DB_srpsalt 	2
+#define DB_srpid	3              
+#define DB_srpgN	4       
+#define DB_srpinfo	5 
+#undef  DB_NUMBER      
+#define DB_NUMBER       6
+
+#define DB_SRP_INDEX	'I'
+#define DB_SRP_VALID	'V'
+#define DB_SRP_REVOKED	'R'
+#define DB_SRP_MODIF	'v'
+
+
+/* see srp.c */
+char * SRP_check_known_gN_param(BIGNUM* g, BIGNUM* N); 
+SRP_gN *SRP_get_default_gN(const char * id) ;
+
+/* server side .... */
+BIGNUM *SRP_Calc_server_key(BIGNUM *A, BIGNUM *v, BIGNUM *u, BIGNUM *b, BIGNUM *N);
+BIGNUM *SRP_Calc_B(BIGNUM *b, BIGNUM *N, BIGNUM *g, BIGNUM *v);
+int SRP_Verify_A_mod_N(BIGNUM *A, BIGNUM *N);
+BIGNUM *SRP_Calc_u(BIGNUM *A, BIGNUM *B, BIGNUM *N) ;
+
+
+
+/* client side .... */
+BIGNUM *SRP_Calc_x(BIGNUM *s, const char *user, const char *pass);
+BIGNUM *SRP_Calc_A(BIGNUM *a, BIGNUM *N, BIGNUM *g);
+BIGNUM *SRP_Calc_client_key(BIGNUM *N, BIGNUM *B, BIGNUM *g, BIGNUM *x, BIGNUM *a, BIGNUM *u);
+int SRP_Verify_B_mod_N(BIGNUM *B, BIGNUM *N);
+
+#define SRP_MINIMAL_N 1024
+
+#ifdef  __cplusplus
+}
+#endif
+
+#endif
+#endif
diff --git a/crypto/srp/srp_grps.h b/crypto/srp/srp_grps.h
new file mode 100644
index 000000000000..d77c9fff4ba1
--- /dev/null
+++ b/crypto/srp/srp_grps.h
@@ -0,0 +1,517 @@
+/* start of generated data */
+
+static BN_ULONG bn_group_1024_value[] = {
+	bn_pack4(9FC6,1D2F,C0EB,06E3),
+	bn_pack4(FD51,38FE,8376,435B),
+	bn_pack4(2FD4,CBF4,976E,AA9A),
+	bn_pack4(68ED,BC3C,0572,6CC0),
+	bn_pack4(C529,F566,660E,57EC),
+	bn_pack4(8255,9B29,7BCF,1885),
+	bn_pack4(CE8E,F4AD,69B1,5D49),
+	bn_pack4(5DC7,D7B4,6154,D6B6),
+	bn_pack4(8E49,5C1D,6089,DAD1),
+	bn_pack4(E0D5,D8E2,50B9,8BE4),
+	bn_pack4(383B,4813,D692,C6E0),
+	bn_pack4(D674,DF74,96EA,81D3),
+	bn_pack4(9EA2,314C,9C25,6576),
+	bn_pack4(6072,6187,75FF,3C0B),
+	bn_pack4(9C33,F80A,FA8F,C5E8),
+	bn_pack4(EEAF,0AB9,ADB3,8DD6)
+};
+static BIGNUM bn_group_1024 = {
+	bn_group_1024_value,
+	(sizeof bn_group_1024_value)/sizeof(BN_ULONG),
+	(sizeof bn_group_1024_value)/sizeof(BN_ULONG),
+	0,
+	BN_FLG_STATIC_DATA
+};
+
+static BN_ULONG bn_group_1536_value[] = {
+	bn_pack4(CF76,E3FE,D135,F9BB),
+	bn_pack4(1518,0F93,499A,234D),
+	bn_pack4(8CE7,A28C,2442,C6F3),
+	bn_pack4(5A02,1FFF,5E91,479E),
+	bn_pack4(7F8A,2FE9,B8B5,292E),
+	bn_pack4(837C,264A,E3A9,BEB8),
+	bn_pack4(E442,734A,F7CC,B7AE),
+	bn_pack4(6577,2E43,7D6C,7F8C),
+	bn_pack4(DB2F,D53D,24B7,C486),
+	bn_pack4(6EDF,0195,3934,9627),
+	bn_pack4(158B,FD3E,2B9C,8CF5),
+	bn_pack4(764E,3F4B,53DD,9DA1),
+	bn_pack4(4754,8381,DBC5,B1FC),
+	bn_pack4(9B60,9E0B,E3BA,B63D),
+	bn_pack4(8134,B1C8,B979,8914),
+	bn_pack4(DF02,8A7C,EC67,F0D0),
+	bn_pack4(80B6,55BB,9A22,E8DC),
+	bn_pack4(1558,903B,A0D0,F843),
+	bn_pack4(51C6,A94B,E460,7A29),
+	bn_pack4(5F4F,5F55,6E27,CBDE),
+	bn_pack4(BEEE,A961,4B19,CC4D),
+	bn_pack4(DBA5,1DF4,99AC,4C80),
+	bn_pack4(B1F1,2A86,17A4,7BBB),
+	bn_pack4(9DEF,3CAF,B939,277A)
+};
+static BIGNUM bn_group_1536 = {
+	bn_group_1536_value,
+	(sizeof bn_group_1536_value)/sizeof(BN_ULONG),
+	(sizeof bn_group_1536_value)/sizeof(BN_ULONG),
+	0,
+	BN_FLG_STATIC_DATA
+};
+
+static BN_ULONG bn_group_2048_value[] = {
+	bn_pack4(0FA7,111F,9E4A,FF73),
+	bn_pack4(9B65,E372,FCD6,8EF2),
+	bn_pack4(35DE,236D,525F,5475),
+	bn_pack4(94B5,C803,D89F,7AE4),
+	bn_pack4(71AE,35F8,E9DB,FBB6),
+	bn_pack4(2A56,98F3,A8D0,C382),
+	bn_pack4(9CCC,041C,7BC3,08D8),
+	bn_pack4(AF87,4E73,03CE,5329),
+	bn_pack4(6160,2790,04E5,7AE6),
+	bn_pack4(032C,FBDB,F52F,B378),
+	bn_pack4(5EA7,7A27,75D2,ECFA),
+	bn_pack4(5445,23B5,24B0,D57D),
+	bn_pack4(5B9D,32E6,88F8,7748),
+	bn_pack4(F1D2,B907,8717,461A),
+	bn_pack4(76BD,207A,436C,6481),
+	bn_pack4(CA97,B43A,23FB,8016),
+	bn_pack4(1D28,1E44,6B14,773B),
+	bn_pack4(7359,D041,D5C3,3EA7),
+	bn_pack4(A80D,740A,DBF4,FF74),
+	bn_pack4(55F9,7993,EC97,5EEA),
+	bn_pack4(2918,A996,2F0B,93B8),
+	bn_pack4(661A,05FB,D5FA,AAE8),
+	bn_pack4(CF60,9517,9A16,3AB3),
+	bn_pack4(E808,3969,EDB7,67B0),
+	bn_pack4(CD7F,48A9,DA04,FD50),
+	bn_pack4(D523,12AB,4B03,310D),
+	bn_pack4(8193,E075,7767,A13D),
+	bn_pack4(A373,29CB,B4A0,99ED),
+	bn_pack4(FC31,9294,3DB5,6050),
+	bn_pack4(AF72,B665,1987,EE07),
+	bn_pack4(F166,DE5E,1389,582F),
+	bn_pack4(AC6B,DB41,324A,9A9B)
+};
+static BIGNUM bn_group_2048 = {
+	bn_group_2048_value,
+	(sizeof bn_group_2048_value)/sizeof(BN_ULONG),
+	(sizeof bn_group_2048_value)/sizeof(BN_ULONG),
+	0,
+	BN_FLG_STATIC_DATA
+};
+
+static BN_ULONG bn_group_3072_value[] = {
+	bn_pack4(FFFF,FFFF,FFFF,FFFF),
+	bn_pack4(4B82,D120,A93A,D2CA),
+	bn_pack4(43DB,5BFC,E0FD,108E),
+	bn_pack4(08E2,4FA0,74E5,AB31),
+	bn_pack4(7709,88C0,BAD9,46E2),
+	bn_pack4(BBE1,1757,7A61,5D6C),
+	bn_pack4(521F,2B18,177B,200C),
+	bn_pack4(D876,0273,3EC8,6A64),
+	bn_pack4(F12F,FA06,D98A,0864),
+	bn_pack4(CEE3,D226,1AD2,EE6B),
+	bn_pack4(1E8C,94E0,4A25,619D),
+	bn_pack4(ABF5,AE8C,DB09,33D7),
+	bn_pack4(B397,0F85,A6E1,E4C7),
+	bn_pack4(8AEA,7157,5D06,0C7D),
+	bn_pack4(ECFB,8504,58DB,EF0A),
+	bn_pack4(A855,21AB,DF1C,BA64),
+	bn_pack4(AD33,170D,0450,7A33),
+	bn_pack4(1572,8E5A,8AAA,C42D),
+	bn_pack4(15D2,2618,98FA,0510),
+	bn_pack4(3995,497C,EA95,6AE5),
+	bn_pack4(DE2B,CBF6,9558,1718),
+	bn_pack4(B5C5,5DF0,6F4C,52C9),
+	bn_pack4(9B27,83A2,EC07,A28F),
+	bn_pack4(E39E,772C,180E,8603),
+	bn_pack4(3290,5E46,2E36,CE3B),
+	bn_pack4(F174,6C08,CA18,217C),
+	bn_pack4(670C,354E,4ABC,9804),
+	bn_pack4(9ED5,2907,7096,966D),
+	bn_pack4(1C62,F356,2085,52BB),
+	bn_pack4(8365,5D23,DCA3,AD96),
+	bn_pack4(6916,3FA8,FD24,CF5F),
+	bn_pack4(98DA,4836,1C55,D39A),
+	bn_pack4(C200,7CB8,A163,BF05),
+	bn_pack4(4928,6651,ECE4,5B3D),
+	bn_pack4(AE9F,2411,7C4B,1FE6),
+	bn_pack4(EE38,6BFB,5A89,9FA5),
+	bn_pack4(0BFF,5CB6,F406,B7ED),
+	bn_pack4(F44C,42E9,A637,ED6B),
+	bn_pack4(E485,B576,625E,7EC6),
+	bn_pack4(4FE1,356D,6D51,C245),
+	bn_pack4(302B,0A6D,F25F,1437),
+	bn_pack4(EF95,19B3,CD3A,431B),
+	bn_pack4(514A,0879,8E34,04DD),
+	bn_pack4(020B,BEA6,3B13,9B22),
+	bn_pack4(2902,4E08,8A67,CC74),
+	bn_pack4(C4C6,628B,80DC,1CD1),
+	bn_pack4(C90F,DAA2,2168,C234),
+	bn_pack4(FFFF,FFFF,FFFF,FFFF)
+};
+static BIGNUM bn_group_3072 = {
+	bn_group_3072_value,
+	(sizeof bn_group_3072_value)/sizeof(BN_ULONG),
+	(sizeof bn_group_3072_value)/sizeof(BN_ULONG),
+	0,
+	BN_FLG_STATIC_DATA
+};
+
+static BN_ULONG bn_group_4096_value[] = {
+	bn_pack4(FFFF,FFFF,FFFF,FFFF),
+	bn_pack4(4DF4,35C9,3406,3199),
+	bn_pack4(86FF,B7DC,90A6,C08F),
+	bn_pack4(93B4,EA98,8D8F,DDC1),
+	bn_pack4(D006,9127,D5B0,5AA9),
+	bn_pack4(B81B,DD76,2170,481C),
+	bn_pack4(1F61,2970,CEE2,D7AF),
+	bn_pack4(233B,A186,515B,E7ED),
+	bn_pack4(99B2,964F,A090,C3A2),
+	bn_pack4(287C,5947,4E6B,C05D),
+	bn_pack4(2E8E,FC14,1FBE,CAA6),
+	bn_pack4(DBBB,C2DB,04DE,8EF9),
+	bn_pack4(2583,E9CA,2AD4,4CE8),
+	bn_pack4(1A94,6834,B615,0BDA),
+	bn_pack4(99C3,2718,6AF4,E23C),
+	bn_pack4(8871,9A10,BDBA,5B26),
+	bn_pack4(1A72,3C12,A787,E6D7),
+	bn_pack4(4B82,D120,A921,0801),
+	bn_pack4(43DB,5BFC,E0FD,108E),
+	bn_pack4(08E2,4FA0,74E5,AB31),
+	bn_pack4(7709,88C0,BAD9,46E2),
+	bn_pack4(BBE1,1757,7A61,5D6C),
+	bn_pack4(521F,2B18,177B,200C),
+	bn_pack4(D876,0273,3EC8,6A64),
+	bn_pack4(F12F,FA06,D98A,0864),
+	bn_pack4(CEE3,D226,1AD2,EE6B),
+	bn_pack4(1E8C,94E0,4A25,619D),
+	bn_pack4(ABF5,AE8C,DB09,33D7),
+	bn_pack4(B397,0F85,A6E1,E4C7),
+	bn_pack4(8AEA,7157,5D06,0C7D),
+	bn_pack4(ECFB,8504,58DB,EF0A),
+	bn_pack4(A855,21AB,DF1C,BA64),
+	bn_pack4(AD33,170D,0450,7A33),
+	bn_pack4(1572,8E5A,8AAA,C42D),
+	bn_pack4(15D2,2618,98FA,0510),
+	bn_pack4(3995,497C,EA95,6AE5),
+	bn_pack4(DE2B,CBF6,9558,1718),
+	bn_pack4(B5C5,5DF0,6F4C,52C9),
+	bn_pack4(9B27,83A2,EC07,A28F),
+	bn_pack4(E39E,772C,180E,8603),
+	bn_pack4(3290,5E46,2E36,CE3B),
+	bn_pack4(F174,6C08,CA18,217C),
+	bn_pack4(670C,354E,4ABC,9804),
+	bn_pack4(9ED5,2907,7096,966D),
+	bn_pack4(1C62,F356,2085,52BB),
+	bn_pack4(8365,5D23,DCA3,AD96),
+	bn_pack4(6916,3FA8,FD24,CF5F),
+	bn_pack4(98DA,4836,1C55,D39A),
+	bn_pack4(C200,7CB8,A163,BF05),
+	bn_pack4(4928,6651,ECE4,5B3D),
+	bn_pack4(AE9F,2411,7C4B,1FE6),
+	bn_pack4(EE38,6BFB,5A89,9FA5),
+	bn_pack4(0BFF,5CB6,F406,B7ED),
+	bn_pack4(F44C,42E9,A637,ED6B),
+	bn_pack4(E485,B576,625E,7EC6),
+	bn_pack4(4FE1,356D,6D51,C245),
+	bn_pack4(302B,0A6D,F25F,1437),
+	bn_pack4(EF95,19B3,CD3A,431B),
+	bn_pack4(514A,0879,8E34,04DD),
+	bn_pack4(020B,BEA6,3B13,9B22),
+	bn_pack4(2902,4E08,8A67,CC74),
+	bn_pack4(C4C6,628B,80DC,1CD1),
+	bn_pack4(C90F,DAA2,2168,C234),
+	bn_pack4(FFFF,FFFF,FFFF,FFFF)
+};
+static BIGNUM bn_group_4096 = {
+	bn_group_4096_value,
+	(sizeof bn_group_4096_value)/sizeof(BN_ULONG),
+	(sizeof bn_group_4096_value)/sizeof(BN_ULONG),
+	0,
+	BN_FLG_STATIC_DATA
+};
+
+static BN_ULONG bn_group_6144_value[] = {
+	bn_pack4(FFFF,FFFF,FFFF,FFFF),
+	bn_pack4(E694,F91E,6DCC,4024),
+	bn_pack4(12BF,2D5B,0B74,74D6),
+	bn_pack4(043E,8F66,3F48,60EE),
+	bn_pack4(387F,E8D7,6E3C,0468),
+	bn_pack4(DA56,C9EC,2EF2,9632),
+	bn_pack4(EB19,CCB1,A313,D55C),
+	bn_pack4(F550,AA3D,8A1F,BFF0),
+	bn_pack4(06A1,D58B,B7C5,DA76),
+	bn_pack4(A797,15EE,F29B,E328),
+	bn_pack4(14CC,5ED2,0F80,37E0),
+	bn_pack4(CC8F,6D7E,BF48,E1D8),
+	bn_pack4(4BD4,07B2,2B41,54AA),
+	bn_pack4(0F1D,45B7,FF58,5AC5),
+	bn_pack4(23A9,7A7E,36CC,88BE),
+	bn_pack4(59E7,C97F,BEC7,E8F3),
+	bn_pack4(B5A8,4031,900B,1C9E),
+	bn_pack4(D55E,702F,4698,0C82),
+	bn_pack4(F482,D7CE,6E74,FEF6),
+	bn_pack4(F032,EA15,D172,1D03),
+	bn_pack4(5983,CA01,C64B,92EC),
+	bn_pack4(6FB8,F401,378C,D2BF),
+	bn_pack4(3320,5151,2BD7,AF42),
+	bn_pack4(DB7F,1447,E6CC,254B),
+	bn_pack4(44CE,6CBA,CED4,BB1B),
+	bn_pack4(DA3E,DBEB,CF9B,14ED),
+	bn_pack4(1797,27B0,865A,8918),
+	bn_pack4(B06A,53ED,9027,D831),
+	bn_pack4(E5DB,382F,4130,01AE),
+	bn_pack4(F8FF,9406,AD9E,530E),
+	bn_pack4(C975,1E76,3DBA,37BD),
+	bn_pack4(C1D4,DCB2,6026,46DE),
+	bn_pack4(36C3,FAB4,D27C,7026),
+	bn_pack4(4DF4,35C9,3402,8492),
+	bn_pack4(86FF,B7DC,90A6,C08F),
+	bn_pack4(93B4,EA98,8D8F,DDC1),
+	bn_pack4(D006,9127,D5B0,5AA9),
+	bn_pack4(B81B,DD76,2170,481C),
+	bn_pack4(1F61,2970,CEE2,D7AF),
+	bn_pack4(233B,A186,515B,E7ED),
+	bn_pack4(99B2,964F,A090,C3A2),
+	bn_pack4(287C,5947,4E6B,C05D),
+	bn_pack4(2E8E,FC14,1FBE,CAA6),
+	bn_pack4(DBBB,C2DB,04DE,8EF9),
+	bn_pack4(2583,E9CA,2AD4,4CE8),
+	bn_pack4(1A94,6834,B615,0BDA),
+	bn_pack4(99C3,2718,6AF4,E23C),
+	bn_pack4(8871,9A10,BDBA,5B26),
+	bn_pack4(1A72,3C12,A787,E6D7),
+	bn_pack4(4B82,D120,A921,0801),
+	bn_pack4(43DB,5BFC,E0FD,108E),
+	bn_pack4(08E2,4FA0,74E5,AB31),
+	bn_pack4(7709,88C0,BAD9,46E2),
+	bn_pack4(BBE1,1757,7A61,5D6C),
+	bn_pack4(521F,2B18,177B,200C),
+	bn_pack4(D876,0273,3EC8,6A64),
+	bn_pack4(F12F,FA06,D98A,0864),
+	bn_pack4(CEE3,D226,1AD2,EE6B),
+	bn_pack4(1E8C,94E0,4A25,619D),
+	bn_pack4(ABF5,AE8C,DB09,33D7),
+	bn_pack4(B397,0F85,A6E1,E4C7),
+	bn_pack4(8AEA,7157,5D06,0C7D),
+	bn_pack4(ECFB,8504,58DB,EF0A),
+	bn_pack4(A855,21AB,DF1C,BA64),
+	bn_pack4(AD33,170D,0450,7A33),
+	bn_pack4(1572,8E5A,8AAA,C42D),
+	bn_pack4(15D2,2618,98FA,0510),
+	bn_pack4(3995,497C,EA95,6AE5),
+	bn_pack4(DE2B,CBF6,9558,1718),
+	bn_pack4(B5C5,5DF0,6F4C,52C9),
+	bn_pack4(9B27,83A2,EC07,A28F),
+	bn_pack4(E39E,772C,180E,8603),
+	bn_pack4(3290,5E46,2E36,CE3B),
+	bn_pack4(F174,6C08,CA18,217C),
+	bn_pack4(670C,354E,4ABC,9804),
+	bn_pack4(9ED5,2907,7096,966D),
+	bn_pack4(1C62,F356,2085,52BB),
+	bn_pack4(8365,5D23,DCA3,AD96),
+	bn_pack4(6916,3FA8,FD24,CF5F),
+	bn_pack4(98DA,4836,1C55,D39A),
+	bn_pack4(C200,7CB8,A163,BF05),
+	bn_pack4(4928,6651,ECE4,5B3D),
+	bn_pack4(AE9F,2411,7C4B,1FE6),
+	bn_pack4(EE38,6BFB,5A89,9FA5),
+	bn_pack4(0BFF,5CB6,F406,B7ED),
+	bn_pack4(F44C,42E9,A637,ED6B),
+	bn_pack4(E485,B576,625E,7EC6),
+	bn_pack4(4FE1,356D,6D51,C245),
+	bn_pack4(302B,0A6D,F25F,1437),
+	bn_pack4(EF95,19B3,CD3A,431B),
+	bn_pack4(514A,0879,8E34,04DD),
+	bn_pack4(020B,BEA6,3B13,9B22),
+	bn_pack4(2902,4E08,8A67,CC74),
+	bn_pack4(C4C6,628B,80DC,1CD1),
+	bn_pack4(C90F,DAA2,2168,C234),
+	bn_pack4(FFFF,FFFF,FFFF,FFFF)
+};
+static BIGNUM bn_group_6144 = {
+	bn_group_6144_value,
+	(sizeof bn_group_6144_value)/sizeof(BN_ULONG),
+	(sizeof bn_group_6144_value)/sizeof(BN_ULONG),
+	0,
+	BN_FLG_STATIC_DATA
+};
+
+static BN_ULONG bn_group_8192_value[] = {
+	bn_pack4(FFFF,FFFF,FFFF,FFFF),
+	bn_pack4(60C9,80DD,98ED,D3DF),
+	bn_pack4(C81F,56E8,80B9,6E71),
+	bn_pack4(9E30,50E2,7656,94DF),
+	bn_pack4(9558,E447,5677,E9AA),
+	bn_pack4(C919,0DA6,FC02,6E47),
+	bn_pack4(889A,002E,D5EE,382B),
+	bn_pack4(4009,438B,481C,6CD7),
+	bn_pack4(3590,46F4,EB87,9F92),
+	bn_pack4(FAF3,6BC3,1ECF,A268),
+	bn_pack4(B1D5,10BD,7EE7,4D73),
+	bn_pack4(F9AB,4819,5DED,7EA1),
+	bn_pack4(64F3,1CC5,0846,851D),
+	bn_pack4(4597,E899,A025,5DC1),
+	bn_pack4(DF31,0EE0,74AB,6A36),
+	bn_pack4(6D2A,13F8,3F44,F82D),
+	bn_pack4(062B,3CF5,B3A2,78A6),
+	bn_pack4(7968,3303,ED5B,DD3A),
+	bn_pack4(FA9D,4B7F,A2C0,87E8),
+	bn_pack4(4BCB,C886,2F83,85DD),
+	bn_pack4(3473,FC64,6CEA,306B),
+	bn_pack4(13EB,57A8,1A23,F0C7),
+	bn_pack4(2222,2E04,A403,7C07),
+	bn_pack4(E3FD,B8BE,FC84,8AD9),
+	bn_pack4(238F,16CB,E39D,652D),
+	bn_pack4(3423,B474,2BF1,C978),
+	bn_pack4(3AAB,639C,5AE4,F568),
+	bn_pack4(2576,F693,6BA4,2466),
+	bn_pack4(741F,A7BF,8AFC,47ED),
+	bn_pack4(3BC8,32B6,8D9D,D300),
+	bn_pack4(D8BE,C4D0,73B9,31BA),
+	bn_pack4(3877,7CB6,A932,DF8C),
+	bn_pack4(74A3,926F,12FE,E5E4),
+	bn_pack4(E694,F91E,6DBE,1159),
+	bn_pack4(12BF,2D5B,0B74,74D6),
+	bn_pack4(043E,8F66,3F48,60EE),
+	bn_pack4(387F,E8D7,6E3C,0468),
+	bn_pack4(DA56,C9EC,2EF2,9632),
+	bn_pack4(EB19,CCB1,A313,D55C),
+	bn_pack4(F550,AA3D,8A1F,BFF0),
+	bn_pack4(06A1,D58B,B7C5,DA76),
+	bn_pack4(A797,15EE,F29B,E328),
+	bn_pack4(14CC,5ED2,0F80,37E0),
+	bn_pack4(CC8F,6D7E,BF48,E1D8),
+	bn_pack4(4BD4,07B2,2B41,54AA),
+	bn_pack4(0F1D,45B7,FF58,5AC5),
+	bn_pack4(23A9,7A7E,36CC,88BE),
+	bn_pack4(59E7,C97F,BEC7,E8F3),
+	bn_pack4(B5A8,4031,900B,1C9E),
+	bn_pack4(D55E,702F,4698,0C82),
+	bn_pack4(F482,D7CE,6E74,FEF6),
+	bn_pack4(F032,EA15,D172,1D03),
+	bn_pack4(5983,CA01,C64B,92EC),
+	bn_pack4(6FB8,F401,378C,D2BF),
+	bn_pack4(3320,5151,2BD7,AF42),
+	bn_pack4(DB7F,1447,E6CC,254B),
+	bn_pack4(44CE,6CBA,CED4,BB1B),
+	bn_pack4(DA3E,DBEB,CF9B,14ED),
+	bn_pack4(1797,27B0,865A,8918),
+	bn_pack4(B06A,53ED,9027,D831),
+	bn_pack4(E5DB,382F,4130,01AE),
+	bn_pack4(F8FF,9406,AD9E,530E),
+	bn_pack4(C975,1E76,3DBA,37BD),
+	bn_pack4(C1D4,DCB2,6026,46DE),
+	bn_pack4(36C3,FAB4,D27C,7026),
+	bn_pack4(4DF4,35C9,3402,8492),
+	bn_pack4(86FF,B7DC,90A6,C08F),
+	bn_pack4(93B4,EA98,8D8F,DDC1),
+	bn_pack4(D006,9127,D5B0,5AA9),
+	bn_pack4(B81B,DD76,2170,481C),
+	bn_pack4(1F61,2970,CEE2,D7AF),
+	bn_pack4(233B,A186,515B,E7ED),
+	bn_pack4(99B2,964F,A090,C3A2),
+	bn_pack4(287C,5947,4E6B,C05D),
+	bn_pack4(2E8E,FC14,1FBE,CAA6),
+	bn_pack4(DBBB,C2DB,04DE,8EF9),
+	bn_pack4(2583,E9CA,2AD4,4CE8),
+	bn_pack4(1A94,6834,B615,0BDA),
+	bn_pack4(99C3,2718,6AF4,E23C),
+	bn_pack4(8871,9A10,BDBA,5B26),
+	bn_pack4(1A72,3C12,A787,E6D7),
+	bn_pack4(4B82,D120,A921,0801),
+	bn_pack4(43DB,5BFC,E0FD,108E),
+	bn_pack4(08E2,4FA0,74E5,AB31),
+	bn_pack4(7709,88C0,BAD9,46E2),
+	bn_pack4(BBE1,1757,7A61,5D6C),
+	bn_pack4(521F,2B18,177B,200C),
+	bn_pack4(D876,0273,3EC8,6A64),
+	bn_pack4(F12F,FA06,D98A,0864),
+	bn_pack4(CEE3,D226,1AD2,EE6B),
+	bn_pack4(1E8C,94E0,4A25,619D),
+	bn_pack4(ABF5,AE8C,DB09,33D7),
+	bn_pack4(B397,0F85,A6E1,E4C7),
+	bn_pack4(8AEA,7157,5D06,0C7D),
+	bn_pack4(ECFB,8504,58DB,EF0A),
+	bn_pack4(A855,21AB,DF1C,BA64),
+	bn_pack4(AD33,170D,0450,7A33),
+	bn_pack4(1572,8E5A,8AAA,C42D),
+	bn_pack4(15D2,2618,98FA,0510),
+	bn_pack4(3995,497C,EA95,6AE5),
+	bn_pack4(DE2B,CBF6,9558,1718),
+	bn_pack4(B5C5,5DF0,6F4C,52C9),
+	bn_pack4(9B27,83A2,EC07,A28F),
+	bn_pack4(E39E,772C,180E,8603),
+	bn_pack4(3290,5E46,2E36,CE3B),
+	bn_pack4(F174,6C08,CA18,217C),
+	bn_pack4(670C,354E,4ABC,9804),
+	bn_pack4(9ED5,2907,7096,966D),
+	bn_pack4(1C62,F356,2085,52BB),
+	bn_pack4(8365,5D23,DCA3,AD96),
+	bn_pack4(6916,3FA8,FD24,CF5F),
+	bn_pack4(98DA,4836,1C55,D39A),
+	bn_pack4(C200,7CB8,A163,BF05),
+	bn_pack4(4928,6651,ECE4,5B3D),
+	bn_pack4(AE9F,2411,7C4B,1FE6),
+	bn_pack4(EE38,6BFB,5A89,9FA5),
+	bn_pack4(0BFF,5CB6,F406,B7ED),
+	bn_pack4(F44C,42E9,A637,ED6B),
+	bn_pack4(E485,B576,625E,7EC6),
+	bn_pack4(4FE1,356D,6D51,C245),
+	bn_pack4(302B,0A6D,F25F,1437),
+	bn_pack4(EF95,19B3,CD3A,431B),
+	bn_pack4(514A,0879,8E34,04DD),
+	bn_pack4(020B,BEA6,3B13,9B22),
+	bn_pack4(2902,4E08,8A67,CC74),
+	bn_pack4(C4C6,628B,80DC,1CD1),
+	bn_pack4(C90F,DAA2,2168,C234),
+	bn_pack4(FFFF,FFFF,FFFF,FFFF)
+};
+static BIGNUM bn_group_8192 = {
+	bn_group_8192_value,
+	(sizeof bn_group_8192_value)/sizeof(BN_ULONG),
+	(sizeof bn_group_8192_value)/sizeof(BN_ULONG),
+	0,
+	BN_FLG_STATIC_DATA
+};
+
+static BN_ULONG bn_generator_19_value[] = {19} ;
+static BIGNUM bn_generator_19 = {
+	bn_generator_19_value,
+	1,
+	1,
+	0,
+	BN_FLG_STATIC_DATA
+};
+static BN_ULONG bn_generator_5_value[] = {5} ;
+static BIGNUM bn_generator_5 = {
+	bn_generator_5_value,
+	1,
+	1,
+	0,
+	BN_FLG_STATIC_DATA
+};
+static BN_ULONG bn_generator_2_value[] = {2} ;
+static BIGNUM bn_generator_2 = {
+	bn_generator_2_value,
+	1,
+	1,
+	0,
+	BN_FLG_STATIC_DATA
+};
+
+static SRP_gN knowngN[] = {
+	{"8192",&bn_generator_19 , &bn_group_8192},
+	{"6144",&bn_generator_5 , &bn_group_6144},
+	{"4096",&bn_generator_5 , &bn_group_4096},
+	{"3072",&bn_generator_5 , &bn_group_3072},
+	{"2048",&bn_generator_2 , &bn_group_2048},
+	{"1536",&bn_generator_2 , &bn_group_1536},
+	{"1024",&bn_generator_2 , &bn_group_1024},
+};
+#define KNOWN_GN_NUMBER sizeof(knowngN) / sizeof(SRP_gN)
+
+/* end of generated data */
diff --git a/crypto/srp/srp_lcl.h b/crypto/srp/srp_lcl.h
new file mode 100644
index 000000000000..42bda3f148f3
--- /dev/null
+++ b/crypto/srp/srp_lcl.h
@@ -0,0 +1,83 @@
+/* crypto/srp/srp_lcl.h */
+/* Written by Peter Sylvester (peter.sylvester@edelweb.fr)  
+ * for the EdelKey project and contributed to the OpenSSL project 2004.
+ */
+/* ====================================================================
+ * Copyright (c) 2004 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+#ifndef HEADER_SRP_LCL_H
+#define HEADER_SRP_LCL_H
+
+#include <openssl/srp.h>
+#include <openssl/sha.h>
+
+#if 0
+#define srp_bn_print(a) {fprintf(stderr, #a "="); BN_print_fp(stderr,a); \
+   fprintf(stderr,"\n");}
+#else
+#define   srp_bn_print(a)
+#endif
+
+
+
+#ifdef  __cplusplus
+extern "C" {
+#endif
+
+
+
+#ifdef  __cplusplus
+}
+#endif
+
+#endif
diff --git a/crypto/srp/srp_lib.c b/crypto/srp/srp_lib.c
new file mode 100644
index 000000000000..92cea98dcd6a
--- /dev/null
+++ b/crypto/srp/srp_lib.c
@@ -0,0 +1,357 @@
+/* crypto/srp/srp_lib.c */
+/* Written by Christophe Renou (christophe.renou@edelweb.fr) with 
+ * the precious help of Peter Sylvester (peter.sylvester@edelweb.fr) 
+ * for the EdelKey project and contributed to the OpenSSL project 2004.
+ */
+/* ====================================================================
+ * Copyright (c) 2004 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+#ifndef OPENSSL_NO_SRP
+#include "cryptlib.h"
+#include "srp_lcl.h"
+#include <openssl/srp.h>
+#include <openssl/evp.h>
+
+#if (BN_BYTES == 8)
+#define bn_pack4(a1,a2,a3,a4) 0x##a1##a2##a3##a4##ul
+#endif
+#if (BN_BYTES == 4)
+#define bn_pack4(a1,a2,a3,a4)  0x##a3##a4##ul, 0x##a1##a2##ul
+#endif
+#if (BN_BYTES == 2)
+#define bn_pack4(a1,a2,a3,a4) 0x##a4##u,0x##a3##u,0x##a2##u,0x##a1##u
+#endif
+
+
+#include "srp_grps.h"
+
+static BIGNUM *srp_Calc_k(BIGNUM *N, BIGNUM *g)
+	{
+	/* k = SHA1(N | PAD(g)) -- tls-srp draft 8 */
+
+	unsigned char digest[SHA_DIGEST_LENGTH];
+	unsigned char *tmp;
+	EVP_MD_CTX ctxt;
+	int longg ;
+	int longN = BN_num_bytes(N);
+
+	if ((tmp = OPENSSL_malloc(longN)) == NULL)
+		return NULL;
+	BN_bn2bin(N,tmp) ;
+
+	EVP_MD_CTX_init(&ctxt);
+	EVP_DigestInit_ex(&ctxt, EVP_sha1(), NULL);
+	EVP_DigestUpdate(&ctxt, tmp, longN);
+
+	memset(tmp, 0, longN);
+	longg = BN_bn2bin(g,tmp) ;
+        /* use the zeros behind to pad on left */
+	EVP_DigestUpdate(&ctxt, tmp + longg, longN-longg);
+	EVP_DigestUpdate(&ctxt, tmp, longg);
+	OPENSSL_free(tmp);
+
+	EVP_DigestFinal_ex(&ctxt, digest, NULL);
+	EVP_MD_CTX_cleanup(&ctxt);
+	return BN_bin2bn(digest, sizeof(digest), NULL);	
+	}
+
+BIGNUM *SRP_Calc_u(BIGNUM *A, BIGNUM *B, BIGNUM *N)
+	{
+	/* k = SHA1(PAD(A) || PAD(B) ) -- tls-srp draft 8 */
+
+	BIGNUM *u;	
+	unsigned char cu[SHA_DIGEST_LENGTH];
+	unsigned char *cAB;
+	EVP_MD_CTX ctxt;
+	int longN;  
+	if ((A == NULL) ||(B == NULL) || (N == NULL))
+		return NULL;
+
+	longN= BN_num_bytes(N);
+
+	if ((cAB = OPENSSL_malloc(2*longN)) == NULL) 
+		return NULL;
+
+	memset(cAB, 0, longN);
+
+	EVP_MD_CTX_init(&ctxt);
+	EVP_DigestInit_ex(&ctxt, EVP_sha1(), NULL);
+	EVP_DigestUpdate(&ctxt, cAB + BN_bn2bin(A,cAB+longN), longN);
+	EVP_DigestUpdate(&ctxt, cAB + BN_bn2bin(B,cAB+longN), longN);
+	OPENSSL_free(cAB);
+	EVP_DigestFinal_ex(&ctxt, cu, NULL);
+	EVP_MD_CTX_cleanup(&ctxt);
+
+	if (!(u = BN_bin2bn(cu, sizeof(cu), NULL)))
+		return NULL;
+	if (!BN_is_zero(u))
+		return u;
+	BN_free(u);
+	return NULL;
+}
+
+BIGNUM *SRP_Calc_server_key(BIGNUM *A, BIGNUM *v, BIGNUM *u, BIGNUM *b, BIGNUM *N)
+	{
+	BIGNUM *tmp = NULL, *S = NULL;
+	BN_CTX *bn_ctx; 
+	
+	if (u == NULL || A == NULL || v == NULL || b == NULL || N == NULL)
+		return NULL; 
+
+	if ((bn_ctx = BN_CTX_new()) == NULL ||
+		(tmp = BN_new()) == NULL ||
+		(S = BN_new()) == NULL )
+		goto err;
+
+	/* S = (A*v**u) ** b */ 
+
+	if (!BN_mod_exp(tmp,v,u,N,bn_ctx))
+		goto err;
+	if (!BN_mod_mul(tmp,A,tmp,N,bn_ctx))
+		goto err;
+	if (!BN_mod_exp(S,tmp,b,N,bn_ctx))
+		goto err;
+err:
+	BN_CTX_free(bn_ctx);
+	BN_clear_free(tmp);
+	return S;
+	}
+
+BIGNUM *SRP_Calc_B(BIGNUM *b, BIGNUM *N, BIGNUM *g, BIGNUM *v)
+	{
+	BIGNUM  *kv = NULL, *gb = NULL;
+	BIGNUM *B = NULL, *k = NULL;
+	BN_CTX *bn_ctx;
+
+	if (b == NULL || N == NULL || g == NULL || v == NULL ||
+		(bn_ctx = BN_CTX_new()) == NULL)
+		return NULL; 
+
+	if ( (kv = BN_new()) == NULL ||
+		(gb = BN_new()) == NULL ||
+		(B = BN_new())== NULL)
+		goto err;
+
+	/* B = g**b + k*v */
+
+	if (!BN_mod_exp(gb,g,b,N,bn_ctx) ||
+	   !(k = srp_Calc_k(N,g)) ||
+	   !BN_mod_mul(kv,v,k,N,bn_ctx) || 
+	   !BN_mod_add(B,gb,kv,N,bn_ctx))
+		{
+		BN_free(B);
+		B = NULL;
+		}
+err:
+	BN_CTX_free(bn_ctx);
+	BN_clear_free(kv);
+	BN_clear_free(gb);
+	BN_free(k); 
+	return B;
+	}
+
+BIGNUM *SRP_Calc_x(BIGNUM *s, const char *user, const char *pass)
+	{
+	unsigned char dig[SHA_DIGEST_LENGTH];
+	EVP_MD_CTX ctxt;
+	unsigned char *cs;
+
+	if ((s == NULL) ||
+		(user == NULL) ||
+		(pass == NULL))
+		return NULL;
+
+	if ((cs = OPENSSL_malloc(BN_num_bytes(s))) == NULL)
+		return NULL;
+
+	EVP_MD_CTX_init(&ctxt);
+	EVP_DigestInit_ex(&ctxt, EVP_sha1(), NULL);
+	EVP_DigestUpdate(&ctxt, user, strlen(user));
+	EVP_DigestUpdate(&ctxt, ":", 1);
+	EVP_DigestUpdate(&ctxt, pass, strlen(pass));
+	EVP_DigestFinal_ex(&ctxt, dig, NULL);
+
+	EVP_DigestInit_ex(&ctxt, EVP_sha1(), NULL);
+	BN_bn2bin(s,cs);
+	EVP_DigestUpdate(&ctxt, cs, BN_num_bytes(s));
+	OPENSSL_free(cs);
+	EVP_DigestUpdate(&ctxt, dig, sizeof(dig));
+	EVP_DigestFinal_ex(&ctxt, dig, NULL);
+	EVP_MD_CTX_cleanup(&ctxt);
+
+	return BN_bin2bn(dig, sizeof(dig), NULL);
+	}
+
+BIGNUM *SRP_Calc_A(BIGNUM *a, BIGNUM *N, BIGNUM *g)
+	{
+	BN_CTX *bn_ctx; 
+	BIGNUM * A = NULL;
+
+	if (a == NULL || N == NULL || g == NULL ||
+		(bn_ctx = BN_CTX_new()) == NULL) 
+		return NULL;
+
+	if ((A = BN_new()) != NULL &&
+	   !BN_mod_exp(A,g,a,N,bn_ctx))
+		{
+		BN_free(A);
+		A = NULL;
+		}
+	BN_CTX_free(bn_ctx);
+	return A;
+	}
+
+
+BIGNUM *SRP_Calc_client_key(BIGNUM *N, BIGNUM *B, BIGNUM *g, BIGNUM *x, BIGNUM *a, BIGNUM *u)
+	{
+	BIGNUM *tmp = NULL, *tmp2 = NULL, *tmp3 = NULL , *k = NULL, *K = NULL;
+	BN_CTX *bn_ctx;
+
+	if (u == NULL || B == NULL || N == NULL || g == NULL || x == NULL || a == NULL ||
+		(bn_ctx = BN_CTX_new()) == NULL)
+		return NULL; 
+
+	if ((tmp = BN_new()) == NULL ||
+		(tmp2 = BN_new())== NULL ||
+		(tmp3 = BN_new())== NULL ||
+		(K = BN_new()) == NULL)
+		goto err;
+	
+	if (!BN_mod_exp(tmp,g,x,N,bn_ctx))
+		goto err;
+	if (!(k = srp_Calc_k(N,g)))
+		goto err;
+	if (!BN_mod_mul(tmp2,tmp,k,N,bn_ctx))
+		goto err;
+	if (!BN_mod_sub(tmp,B,tmp2,N,bn_ctx))
+		goto err;
+
+	if (!BN_mod_mul(tmp3,u,x,N,bn_ctx))
+		goto err;
+	if (!BN_mod_add(tmp2,a,tmp3,N,bn_ctx))
+		goto err;
+	if (!BN_mod_exp(K,tmp,tmp2,N,bn_ctx))
+		goto err;
+
+err :
+	BN_CTX_free(bn_ctx);
+	BN_clear_free(tmp);
+	BN_clear_free(tmp2);
+	BN_clear_free(tmp3);
+	BN_free(k);
+	return K;	
+	}
+
+int SRP_Verify_B_mod_N(BIGNUM *B, BIGNUM *N)
+	{
+	BIGNUM *r;
+	BN_CTX *bn_ctx; 
+	int ret = 0;
+
+	if (B == NULL || N == NULL ||
+		(bn_ctx = BN_CTX_new()) == NULL)
+		return 0;
+
+	if ((r = BN_new()) == NULL)
+		goto err;
+	/* Checks if B % N == 0 */
+	if (!BN_nnmod(r,B,N,bn_ctx))
+		goto err;
+	ret = !BN_is_zero(r);
+err:
+	BN_CTX_free(bn_ctx);
+	BN_free(r);
+	return ret;
+	}
+
+int SRP_Verify_A_mod_N(BIGNUM *A, BIGNUM *N)
+	{
+	/* Checks if A % N == 0 */
+	return SRP_Verify_B_mod_N(A,N) ;
+	}
+
+
+/* Check if G and N are kwown parameters. 
+   The values have been generated from the ietf-tls-srp draft version 8
+*/
+char *SRP_check_known_gN_param(BIGNUM *g, BIGNUM *N)
+	{
+	size_t i;
+	if ((g == NULL) || (N == NULL))
+		return 0;
+
+	srp_bn_print(g);
+	srp_bn_print(N);
+
+	for(i = 0; i < KNOWN_GN_NUMBER; i++)
+		{
+		if (BN_cmp(knowngN[i].g, g) == 0 && BN_cmp(knowngN[i].N, N) == 0) 
+			return knowngN[i].id;
+		}
+	return NULL;
+	}
+
+SRP_gN *SRP_get_default_gN(const char *id)
+	{
+	size_t i;
+
+	if (id == NULL) 
+		return knowngN;
+	for(i = 0; i < KNOWN_GN_NUMBER; i++)
+		{
+		if (strcmp(knowngN[i].id, id)==0)
+			return knowngN + i;
+		}
+	return NULL;
+	}
+#endif
diff --git a/crypto/srp/srp_vfy.c b/crypto/srp/srp_vfy.c
new file mode 100644
index 000000000000..c8be907d7f7f
--- /dev/null
+++ b/crypto/srp/srp_vfy.c
@@ -0,0 +1,657 @@
+/* crypto/srp/srp_vfy.c */
+/* Written by Christophe Renou (christophe.renou@edelweb.fr) with 
+ * the precious help of Peter Sylvester (peter.sylvester@edelweb.fr) 
+ * for the EdelKey project and contributed to the OpenSSL project 2004.
+ */
+/* ====================================================================
+ * Copyright (c) 2004 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+#ifndef OPENSSL_NO_SRP
+#include "cryptlib.h"
+#include "srp_lcl.h"
+#include <openssl/srp.h>
+#include <openssl/evp.h>
+#include <openssl/buffer.h>
+#include <openssl/rand.h>
+#include <openssl/txt_db.h>
+
+#define SRP_RANDOM_SALT_LEN 20
+#define MAX_LEN 2500
+
+static char b64table[] =
+  "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz./";
+
+/* the following two conversion routines have been inspired by code from Stanford */ 
+
+/*
+ * Convert a base64 string into raw byte array representation.
+ */
+static int t_fromb64(unsigned char *a, const char *src)
+	{
+	char *loc;
+	int i, j;
+	int size;
+
+	while(*src && (*src == ' ' || *src == '\t' || *src == '\n'))
+		++src;
+	size = strlen(src);
+	i = 0;
+	while(i < size)
+		{
+		loc = strchr(b64table, src[i]);
+		if(loc == (char *) 0) break;
+		else a[i] = loc - b64table;
+		++i;
+		}
+	size = i;
+	i = size - 1;
+	j = size;
+	while(1)
+		{
+		a[j] = a[i];
+		if(--i < 0) break;
+		a[j] |= (a[i] & 3) << 6;
+		--j;
+		a[j] = (unsigned char) ((a[i] & 0x3c) >> 2);
+		if(--i < 0) break;
+		a[j] |= (a[i] & 0xf) << 4;
+		--j;
+		a[j] = (unsigned char) ((a[i] & 0x30) >> 4);
+		if(--i < 0) break;
+		a[j] |= (a[i] << 2);
+
+		a[--j] = 0;
+		if(--i < 0) break;
+		}
+	while(a[j] == 0 && j <= size) ++j;
+	i = 0;
+	while (j <= size) a[i++] = a[j++];
+	return i;
+	}
+
+
+/*
+ * Convert a raw byte string into a null-terminated base64 ASCII string.
+ */
+static char *t_tob64(char *dst, const unsigned char *src, int size)
+	{
+	int c, pos = size % 3;
+	unsigned char b0 = 0, b1 = 0, b2 = 0, notleading = 0;
+	char *olddst = dst;
+
+	switch(pos)
+		{
+	case 1:
+		b2 = src[0];
+		break;
+	case 2:
+		b1 = src[0];
+		b2 = src[1];
+		break;
+		}
+
+	while(1)
+		{
+		c = (b0 & 0xfc) >> 2;
+		if(notleading || c != 0)
+			{
+			*dst++ = b64table[c];
+			notleading = 1;
+			}
+		c = ((b0 & 3) << 4) | ((b1 & 0xf0) >> 4);
+		if(notleading || c != 0)
+			{
+			*dst++ = b64table[c];
+			notleading = 1;
+			}
+		c = ((b1 & 0xf) << 2) | ((b2 & 0xc0) >> 6);
+		if(notleading || c != 0)
+			{
+			*dst++ = b64table[c];
+			notleading = 1;
+			}
+		c = b2 & 0x3f;
+		if(notleading || c != 0)
+			{
+			*dst++ = b64table[c];
+			notleading = 1;
+			}
+		if(pos >= size) break;
+		else
+			{
+			b0 = src[pos++];
+			b1 = src[pos++];
+			b2 = src[pos++];
+			}
+		}
+
+	*dst++ = '\0';
+	return olddst;
+	}
+
+static void SRP_user_pwd_free(SRP_user_pwd *user_pwd)
+	{
+	if (user_pwd == NULL) 
+		return;
+	BN_free(user_pwd->s);
+	BN_clear_free(user_pwd->v);
+	OPENSSL_free(user_pwd->id);
+	OPENSSL_free(user_pwd->info);
+	OPENSSL_free(user_pwd);
+	}
+
+static SRP_user_pwd *SRP_user_pwd_new()
+	{
+	SRP_user_pwd *ret = OPENSSL_malloc(sizeof(SRP_user_pwd));
+	if (ret == NULL)
+		return NULL;								
+	ret->N = NULL;
+	ret->g = NULL;	
+	ret->s = NULL;
+	ret->v = NULL;
+	ret->id = NULL ;
+	ret->info = NULL;
+	return ret;
+	}
+
+static void SRP_user_pwd_set_gN(SRP_user_pwd *vinfo, const BIGNUM *g,
+				const BIGNUM *N)
+	{
+	vinfo->N = N;
+	vinfo->g = g;	
+	}
+
+static int SRP_user_pwd_set_ids(SRP_user_pwd *vinfo, const char *id,
+				const char *info)
+	{
+	if (id != NULL && NULL == (vinfo->id = BUF_strdup(id)))
+		return 0;
+	return (info == NULL || NULL != (vinfo->info = BUF_strdup(info))) ;
+	}
+
+static int SRP_user_pwd_set_sv(SRP_user_pwd *vinfo, const char *s,
+			       const char *v)
+	{
+	unsigned char tmp[MAX_LEN];
+	int len;
+
+	if (strlen(s) > MAX_LEN || strlen(v) > MAX_LEN) 
+		return 0; 
+	len = t_fromb64(tmp, v);
+	if (NULL == (vinfo->v = BN_bin2bn(tmp, len, NULL)) )
+		return 0;
+	len = t_fromb64(tmp, s);
+	return ((vinfo->s = BN_bin2bn(tmp, len, NULL)) != NULL) ;
+	}
+
+static int SRP_user_pwd_set_sv_BN(SRP_user_pwd *vinfo, BIGNUM *s, BIGNUM *v)
+	{
+	vinfo->v = v;
+	vinfo->s = s;
+	return (vinfo->s != NULL && vinfo->v != NULL) ;
+	}
+
+SRP_VBASE *SRP_VBASE_new(char *seed_key)
+	{
+	SRP_VBASE *vb = (SRP_VBASE *) OPENSSL_malloc(sizeof(SRP_VBASE));
+
+	if (vb == NULL)
+		return NULL;
+	if (!(vb->users_pwd = sk_SRP_user_pwd_new_null()) ||
+		!(vb->gN_cache = sk_SRP_gN_cache_new_null()))
+		{
+		OPENSSL_free(vb);
+		return NULL;
+		}
+	vb->default_g = NULL;
+	vb->default_N = NULL;
+	vb->seed_key = NULL;
+	if ((seed_key != NULL) && 
+		(vb->seed_key = BUF_strdup(seed_key)) == NULL)
+		{
+		sk_SRP_user_pwd_free(vb->users_pwd);
+		sk_SRP_gN_cache_free(vb->gN_cache);
+		OPENSSL_free(vb);
+		return NULL;
+		}
+	return vb;
+	}
+
+
+int SRP_VBASE_free(SRP_VBASE *vb)
+	{
+	sk_SRP_user_pwd_pop_free(vb->users_pwd,SRP_user_pwd_free);
+	sk_SRP_gN_cache_free(vb->gN_cache);
+	OPENSSL_free(vb->seed_key);
+	OPENSSL_free(vb);
+	return 0;
+	}
+
+
+static SRP_gN_cache *SRP_gN_new_init(const char *ch)
+	{
+	unsigned char tmp[MAX_LEN];
+	int len;
+
+	SRP_gN_cache *newgN = (SRP_gN_cache *)OPENSSL_malloc(sizeof(SRP_gN_cache));
+	if (newgN == NULL)
+		return NULL;
+
+	if ((newgN->b64_bn = BUF_strdup(ch)) == NULL)
+		goto err;
+
+	len = t_fromb64(tmp, ch);
+	if ((newgN->bn = BN_bin2bn(tmp, len, NULL)))
+		return newgN;
+
+	OPENSSL_free(newgN->b64_bn);
+err:
+	OPENSSL_free(newgN);
+	return NULL;
+	}
+
+
+static void SRP_gN_free(SRP_gN_cache *gN_cache)
+	{
+	if (gN_cache == NULL)
+		return;
+	OPENSSL_free(gN_cache->b64_bn);
+	BN_free(gN_cache->bn);
+	OPENSSL_free(gN_cache);
+	}
+
+static SRP_gN *SRP_get_gN_by_id(const char *id, STACK_OF(SRP_gN) *gN_tab)
+	{
+	int i;
+
+	SRP_gN *gN;
+	if (gN_tab != NULL) 
+	for(i = 0; i < sk_SRP_gN_num(gN_tab); i++)
+		{
+		gN = sk_SRP_gN_value(gN_tab, i);
+		if (gN && (id == NULL || strcmp(gN->id,id)==0))
+			return gN;
+		}
+	
+	return SRP_get_default_gN(id);
+	}
+
+static BIGNUM *SRP_gN_place_bn(STACK_OF(SRP_gN_cache) *gN_cache, char *ch)
+	{
+	int i;
+	if (gN_cache == NULL)
+		return NULL;
+
+	/* search if we have already one... */
+	for(i = 0; i < sk_SRP_gN_cache_num(gN_cache); i++)
+		{
+		SRP_gN_cache *cache = sk_SRP_gN_cache_value(gN_cache, i);
+		if (strcmp(cache->b64_bn,ch)==0)
+			return cache->bn;
+		}
+		{		/* it is the first time that we find it */
+		SRP_gN_cache *newgN = SRP_gN_new_init(ch);
+		if (newgN)
+			{
+			if (sk_SRP_gN_cache_insert(gN_cache,newgN,0)>0)
+				return newgN->bn;
+			SRP_gN_free(newgN);
+			}
+		}
+	return NULL;
+	}
+
+/* this function parses verifier file. Format is:
+ * string(index):base64(N):base64(g):0
+ * string(username):base64(v):base64(salt):int(index)
+ */
+
+
+int SRP_VBASE_init(SRP_VBASE *vb, char *verifier_file)
+	{
+	int error_code ;
+	STACK_OF(SRP_gN) *SRP_gN_tab = sk_SRP_gN_new_null();
+	char *last_index = NULL;
+	int i;
+	char **pp;
+
+	SRP_gN *gN = NULL;
+	SRP_user_pwd *user_pwd = NULL ;
+
+	TXT_DB *tmpdb = NULL;
+	BIO *in = BIO_new(BIO_s_file());
+
+	error_code = SRP_ERR_OPEN_FILE;
+
+	if (in == NULL || BIO_read_filename(in,verifier_file) <= 0)
+		goto err;
+
+	error_code = SRP_ERR_VBASE_INCOMPLETE_FILE;
+
+	if ((tmpdb =TXT_DB_read(in,DB_NUMBER)) == NULL)
+		goto err;
+
+	error_code = SRP_ERR_MEMORY;
+
+
+	if (vb->seed_key)
+		{
+		last_index = SRP_get_default_gN(NULL)->id;
+		}
+	for (i = 0; i < sk_OPENSSL_PSTRING_num(tmpdb->data); i++)
+		{
+		pp = (char **)sk_OPENSSL_PSTRING_value(tmpdb->data,i);
+		if (pp[DB_srptype][0] == DB_SRP_INDEX)
+			{
+			/*we add this couple in the internal Stack */
+
+			if ((gN = (SRP_gN *)OPENSSL_malloc(sizeof(SRP_gN))) == NULL) 
+ 				goto err;
+
+			if  (!(gN->id = BUF_strdup(pp[DB_srpid]))
+	                ||  !(gN->N = SRP_gN_place_bn(vb->gN_cache,pp[DB_srpverifier]))
+			||  !(gN->g = SRP_gN_place_bn(vb->gN_cache,pp[DB_srpsalt]))
+			||  sk_SRP_gN_insert(SRP_gN_tab,gN,0) == 0)
+				goto err;
+
+			gN = NULL;
+
+			if (vb->seed_key != NULL)
+				{
+				last_index = pp[DB_srpid];
+				}
+			}
+		else if (pp[DB_srptype][0] == DB_SRP_VALID)
+			{
+			/* it is a user .... */
+			SRP_gN *lgN;
+			if ((lgN = SRP_get_gN_by_id(pp[DB_srpgN],SRP_gN_tab))!=NULL)
+				{
+				error_code = SRP_ERR_MEMORY;
+				if ((user_pwd = SRP_user_pwd_new()) == NULL) 
+					goto err;
+				
+				SRP_user_pwd_set_gN(user_pwd,lgN->g,lgN->N);
+				if (!SRP_user_pwd_set_ids(user_pwd, pp[DB_srpid],pp[DB_srpinfo]))
+					goto err;
+				
+				error_code = SRP_ERR_VBASE_BN_LIB;
+				if (!SRP_user_pwd_set_sv(user_pwd, pp[DB_srpsalt],pp[DB_srpverifier]))
+					goto err;
+
+				if (sk_SRP_user_pwd_insert(vb->users_pwd, user_pwd, 0) == 0)
+					goto err;
+				user_pwd = NULL; /* abandon responsability */
+				}
+			}
+		}
+	
+	if (last_index != NULL)
+		{
+		/* this means that we want to simulate a default user */
+
+		if (((gN = SRP_get_gN_by_id(last_index,SRP_gN_tab))==NULL))
+			{
+			error_code = SRP_ERR_VBASE_BN_LIB;
+			goto err;
+			}
+		vb->default_g = gN->g ;
+		vb->default_N = gN->N ;
+		gN = NULL ;
+		}
+	error_code = SRP_NO_ERROR;
+
+ err:
+	/* there may be still some leaks to fix, if this fails, the application terminates most likely */
+
+	if (gN != NULL)
+		{
+		OPENSSL_free(gN->id);
+		OPENSSL_free(gN);
+		}
+
+	SRP_user_pwd_free(user_pwd);
+
+	if (tmpdb) TXT_DB_free(tmpdb);
+	if (in) BIO_free_all(in);
+
+	sk_SRP_gN_free(SRP_gN_tab);
+
+	return error_code;
+
+	}
+
+
+SRP_user_pwd *SRP_VBASE_get_by_user(SRP_VBASE *vb, char *username)
+	{
+	int i;
+	SRP_user_pwd *user;
+	unsigned char digv[SHA_DIGEST_LENGTH];
+	unsigned char digs[SHA_DIGEST_LENGTH];
+	EVP_MD_CTX ctxt;
+
+	if (vb == NULL)
+		return NULL;
+	for(i = 0; i < sk_SRP_user_pwd_num(vb->users_pwd); i++)
+		{
+		user = sk_SRP_user_pwd_value(vb->users_pwd, i);
+		if (strcmp(user->id,username)==0)
+			return user;
+		}
+	if ((vb->seed_key == NULL) ||
+		(vb->default_g == NULL) ||
+		(vb->default_N == NULL))
+		return NULL;
+
+/* if the user is unknown we set parameters as well if we have a seed_key */
+
+	if ((user = SRP_user_pwd_new()) == NULL) 
+		return NULL;
+
+	SRP_user_pwd_set_gN(user,vb->default_g,vb->default_N);
+				
+	if (!SRP_user_pwd_set_ids(user,username,NULL))
+		goto err;
+		
+	RAND_pseudo_bytes(digv, SHA_DIGEST_LENGTH);
+	EVP_MD_CTX_init(&ctxt);
+	EVP_DigestInit_ex(&ctxt, EVP_sha1(), NULL);
+	EVP_DigestUpdate(&ctxt, vb->seed_key, strlen(vb->seed_key));
+	EVP_DigestUpdate(&ctxt, username, strlen(username));
+	EVP_DigestFinal_ex(&ctxt, digs, NULL);
+	EVP_MD_CTX_cleanup(&ctxt);
+	if (SRP_user_pwd_set_sv_BN(user, BN_bin2bn(digs,SHA_DIGEST_LENGTH,NULL), BN_bin2bn(digv,SHA_DIGEST_LENGTH, NULL))) 
+		return user;
+
+err:    SRP_user_pwd_free(user);
+	return NULL;
+	}
+
+
+/*
+   create a verifier (*salt,*verifier,g and N are in base64)
+*/
+char *SRP_create_verifier(const char *user, const char *pass, char **salt,
+			  char **verifier, const char *N, const char *g)
+	{
+	int len;
+	char * result=NULL;
+	char *vf;
+	BIGNUM *N_bn = NULL, *g_bn = NULL, *s = NULL, *v = NULL;
+	unsigned char tmp[MAX_LEN];
+	unsigned char tmp2[MAX_LEN];
+	char * defgNid = NULL;
+
+	if ((user == NULL)||
+		(pass == NULL)||
+		(salt == NULL)||
+		(verifier == NULL))
+		goto err;
+
+	if (N)
+		{
+		if (!(len = t_fromb64(tmp, N))) goto err;
+		N_bn = BN_bin2bn(tmp, len, NULL);
+		if (!(len = t_fromb64(tmp, g))) goto err;
+		g_bn = BN_bin2bn(tmp, len, NULL);
+		defgNid = "*";
+		}
+	else
+		{ 
+		SRP_gN * gN = SRP_get_gN_by_id(g, NULL) ;
+		if (gN == NULL)
+			goto err;
+		N_bn = gN->N;
+		g_bn = gN->g;
+		defgNid = gN->id;
+		}
+
+	if (*salt == NULL)
+		{
+		RAND_pseudo_bytes(tmp2, SRP_RANDOM_SALT_LEN);
+
+		s = BN_bin2bn(tmp2, SRP_RANDOM_SALT_LEN, NULL);
+		}
+	else
+		{
+		if (!(len = t_fromb64(tmp2, *salt)))
+			goto err;
+		s = BN_bin2bn(tmp2, len, NULL);
+		}
+
+
+	if(!SRP_create_verifier_BN(user, pass, &s, &v, N_bn, g_bn)) goto err;
+
+	BN_bn2bin(v,tmp);
+	if (((vf = OPENSSL_malloc(BN_num_bytes(v)*2)) == NULL))
+		goto err;
+	t_tob64(vf, tmp, BN_num_bytes(v));
+
+	*verifier = vf;
+	if (*salt == NULL)
+		{
+		char *tmp_salt;
+		if ((tmp_salt = (char *)OPENSSL_malloc(SRP_RANDOM_SALT_LEN * 2)) == NULL)
+			{
+			OPENSSL_free(vf);
+			goto err;
+			}
+		t_tob64(tmp_salt, tmp2, SRP_RANDOM_SALT_LEN);
+		*salt = tmp_salt;
+		}
+
+	result=defgNid;
+
+err:
+	if(N)
+		{
+		BN_free(N_bn);
+		BN_free(g_bn);
+		}
+	return result;
+	}
+
+/*
+   create a verifier (*salt,*verifier,g and N are BIGNUMs)
+*/
+int SRP_create_verifier_BN(const char *user, const char *pass, BIGNUM **salt, BIGNUM **verifier, BIGNUM *N, BIGNUM *g)
+	{
+	int result=0;
+	BIGNUM *x = NULL;
+	BN_CTX *bn_ctx = BN_CTX_new();
+	unsigned char tmp2[MAX_LEN];
+
+	if ((user == NULL)||
+		(pass == NULL)||
+		(salt == NULL)||
+		(verifier == NULL)||
+		(N == NULL)||
+		(g == NULL)||
+		(bn_ctx == NULL))
+		goto err;
+
+	srp_bn_print(N);
+	srp_bn_print(g);
+
+	if (*salt == NULL)
+		{
+		RAND_pseudo_bytes(tmp2, SRP_RANDOM_SALT_LEN);
+
+		*salt = BN_bin2bn(tmp2,SRP_RANDOM_SALT_LEN,NULL);
+		}
+
+	x = SRP_Calc_x(*salt,user,pass);
+
+	*verifier = BN_new();
+	if(*verifier == NULL) goto err;
+
+	if (!BN_mod_exp(*verifier,g,x,N,bn_ctx))
+		{
+		BN_clear_free(*verifier);
+		goto err;
+		}
+
+	srp_bn_print(*verifier);
+
+	result=1;
+
+err:
+
+	BN_clear_free(x);
+	BN_CTX_free(bn_ctx);
+	return result;
+	}
+
+
+
+#endif
diff --git a/crypto/srp/srptest.c b/crypto/srp/srptest.c
new file mode 100644
index 000000000000..04b66b454416
--- /dev/null
+++ b/crypto/srp/srptest.c
@@ -0,0 +1,162 @@
+#include <openssl/opensslconf.h>
+#ifdef OPENSSL_NO_SRP
+
+#include <stdio.h>
+
+int main(int argc, char *argv[])
+	{
+	printf("No SRP support\n");
+	return(0);
+	}
+
+#else
+
+#include <openssl/srp.h>
+#include <openssl/rand.h>
+#include <openssl/err.h>
+
+static void showbn(const char *name, const BIGNUM *bn)
+	{
+	fputs(name, stdout);
+	fputs(" = ", stdout);
+	BN_print_fp(stdout, bn);
+	putc('\n', stdout);
+	}
+
+#define RANDOM_SIZE 32	/* use 256 bits on each side */
+
+static int run_srp(const char *username, const char *client_pass, const char *server_pass)
+	{
+	int ret=-1;
+	BIGNUM *s = NULL;
+	BIGNUM *v = NULL;
+	BIGNUM *a = NULL;
+	BIGNUM *b = NULL;
+	BIGNUM *u = NULL;
+	BIGNUM *x = NULL;
+	BIGNUM *Apub = NULL;
+	BIGNUM *Bpub = NULL;
+	BIGNUM *Kclient = NULL;
+	BIGNUM *Kserver = NULL;
+	unsigned char rand_tmp[RANDOM_SIZE];
+	/* use builtin 1024-bit params */
+	SRP_gN *GN = SRP_get_default_gN("1024");
+
+	if(GN == NULL)
+		{
+		fprintf(stderr, "Failed to get SRP parameters\n");
+		return -1;
+		}
+	/* Set up server's password entry */
+	if(!SRP_create_verifier_BN(username, server_pass, &s, &v, GN->N, GN->g))
+		{
+		fprintf(stderr, "Failed to create SRP verifier\n");
+		return -1;
+		}
+
+	showbn("N", GN->N);
+	showbn("g", GN->g);
+	showbn("Salt", s);
+	showbn("Verifier", v);
+
+	/* Server random */
+	RAND_pseudo_bytes(rand_tmp, sizeof(rand_tmp));
+	b = BN_bin2bn(rand_tmp, sizeof(rand_tmp), NULL);
+	/* TODO - check b != 0 */
+	showbn("b", b);
+
+	/* Server's first message */
+	Bpub = SRP_Calc_B(b, GN->N, GN->g, v);
+	showbn("B", Bpub);
+
+	if(!SRP_Verify_B_mod_N(Bpub, GN->N))
+		{
+		fprintf(stderr, "Invalid B\n");
+		return -1;
+		}
+
+	/* Client random */
+	RAND_pseudo_bytes(rand_tmp, sizeof(rand_tmp));
+	a = BN_bin2bn(rand_tmp, sizeof(rand_tmp), NULL);
+	/* TODO - check a != 0 */
+	showbn("a", a);
+
+	/* Client's response */
+	Apub = SRP_Calc_A(a, GN->N, GN->g);
+	showbn("A", Apub);
+
+	if(!SRP_Verify_A_mod_N(Apub, GN->N))
+		{
+		fprintf(stderr, "Invalid A\n");
+		return -1;
+		}
+
+	/* Both sides calculate u */
+	u = SRP_Calc_u(Apub, Bpub, GN->N);
+
+	/* Client's key */
+	x = SRP_Calc_x(s, username, client_pass);
+	Kclient = SRP_Calc_client_key(GN->N, Bpub, GN->g, x, a, u);
+	showbn("Client's key", Kclient);
+
+	/* Server's key */
+	Kserver = SRP_Calc_server_key(Apub, v, u, b, GN->N);
+	showbn("Server's key", Kserver);
+
+	if(BN_cmp(Kclient, Kserver) == 0)
+		{
+		ret = 0;
+		}
+	else
+		{
+		fprintf(stderr, "Keys mismatch\n");
+		ret = 1;
+		}
+
+	BN_clear_free(Kclient);
+	BN_clear_free(Kserver);
+	BN_clear_free(x);
+	BN_free(u);
+	BN_free(Apub);
+	BN_clear_free(a);
+	BN_free(Bpub);
+	BN_clear_free(b);
+	BN_free(s);
+	BN_clear_free(v);
+
+	return ret;
+	}
+
+int main(int argc, char **argv)
+	{
+	BIO *bio_err;
+	bio_err = BIO_new_fp(stderr, BIO_NOCLOSE);
+
+	CRYPTO_malloc_debug_init();
+	CRYPTO_dbg_set_options(V_CRYPTO_MDEBUG_ALL);
+	CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON);
+
+	ERR_load_crypto_strings();
+
+	/* "Negative" test, expect a mismatch */
+	if(run_srp("alice", "password1", "password2") == 0)
+		{
+		fprintf(stderr, "Mismatched SRP run failed\n");
+		return 1;
+		}
+
+	/* "Positive" test, should pass */
+	if(run_srp("alice", "password", "password") != 0)
+		{
+		fprintf(stderr, "Plain SRP run failed\n");
+		return 1;
+		}
+
+	CRYPTO_cleanup_all_ex_data();
+	ERR_remove_thread_state(NULL);
+	ERR_free_strings();
+	CRYPTO_mem_leaks(bio_err);
+
+	return 0;
+	}
+#endif
diff --git a/crypto/stack/Makefile b/crypto/stack/Makefile
index 489a77b93c43..5327692ac895 100644
--- a/crypto/stack/Makefile
+++ b/crypto/stack/Makefile
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
diff --git a/crypto/stack/safestack.h b/crypto/stack/safestack.h
index b59c6409e4a5..ea3aa0d800ec 100644
--- a/crypto/stack/safestack.h
+++ b/crypto/stack/safestack.h
@@ -57,18 +57,27 @@
 
 #include <openssl/stack.h>
 
-#ifdef DEBUG_SAFESTACK
-
 #ifndef CHECKED_PTR_OF
 #define CHECKED_PTR_OF(type, p) \
     ((void*) (1 ? p : (type*)0))
 #endif
 
+/* In C++ we get problems because an explicit cast is needed from (void *)
+ * we use CHECKED_STACK_OF to ensure the correct type is passed in the macros
+ * below. 
+ */
+
+#define CHECKED_STACK_OF(type, p) \
+    ((_STACK*) (1 ? p : (STACK_OF(type)*)0))
+
 #define CHECKED_SK_FREE_FUNC(type, p) \
     ((void (*)(void *)) ((1 ? p : (void (*)(type *))0)))
 
+#define CHECKED_SK_FREE_FUNC2(type, p) \
+    ((void (*)(void *)) ((1 ? p : (void (*)(type))0)))
+
 #define CHECKED_SK_CMP_FUNC(type, p) \
-    ((int (*)(const char * const *, const char * const *)) \
+    ((int (*)(const void *, const void *)) \
 	((1 ? p : (int (*)(const type * const *, const type * const *))0)))
 
 #define STACK_OF(type) struct stack_st_##type
@@ -77,11 +86,51 @@
 #define DECLARE_STACK_OF(type) \
 STACK_OF(type) \
     { \
-    STACK stack; \
+    _STACK stack; \
+    };
+#define DECLARE_SPECIAL_STACK_OF(type, type2) \
+STACK_OF(type) \
+    { \
+    _STACK stack; \
     };
 
 #define IMPLEMENT_STACK_OF(type) /* nada (obsolete in new safestack approach)*/
 
+
+/* Strings are special: normally an lhash entry will point to a single
+ * (somewhat) mutable object. In the case of strings:
+ *
+ * a) Instead of a single char, there is an array of chars, NUL-terminated.
+ * b) The string may have be immutable.
+ *
+ * So, they need their own declarations. Especially important for
+ * type-checking tools, such as Deputy.
+ *
+o * In practice, however, it appears to be hard to have a const
+ * string. For now, I'm settling for dealing with the fact it is a
+ * string at all.
+ */
+typedef char *OPENSSL_STRING;
+
+typedef const char *OPENSSL_CSTRING;
+
+/* Confusingly, LHASH_OF(STRING) deals with char ** throughout, but
+ * STACK_OF(STRING) is really more like STACK_OF(char), only, as
+ * mentioned above, instead of a single char each entry is a
+ * NUL-terminated array of chars. So, we have to implement STRING
+ * specially for STACK_OF. This is dealt with in the autogenerated
+ * macros below.
+ */
+
+DECLARE_SPECIAL_STACK_OF(OPENSSL_STRING, char)
+
+/* Similarly, we sometimes use a block of characters, NOT
+ * nul-terminated. These should also be distinguished from "normal"
+ * stacks. */
+
+typedef void *OPENSSL_BLOCK;
+DECLARE_SPECIAL_STACK_OF(OPENSSL_BLOCK, void)
+
 /* SKM_sk_... stack macros are internal to safestack.h:
  * never use them directly, use sk_<type>_... instead */
 #define SKM_sk_new(type, cmp) \
@@ -89,52 +138,56 @@ STACK_OF(type) \
 #define SKM_sk_new_null(type) \
 	((STACK_OF(type) *)sk_new_null())
 #define SKM_sk_free(type, st) \
-	sk_free(CHECKED_PTR_OF(STACK_OF(type), st))
+	sk_free(CHECKED_STACK_OF(type, st))
 #define SKM_sk_num(type, st) \
-	sk_num(CHECKED_PTR_OF(STACK_OF(type), st))
+	sk_num(CHECKED_STACK_OF(type, st))
 #define SKM_sk_value(type, st,i) \
-	((type *)sk_value(CHECKED_PTR_OF(STACK_OF(type), st), i))
+	((type *)sk_value(CHECKED_STACK_OF(type, st), i))
 #define SKM_sk_set(type, st,i,val) \
-	sk_set(CHECKED_PTR_OF(STACK_OF(type), st), i, CHECKED_PTR_OF(type, val))
+	sk_set(CHECKED_STACK_OF(type, st), i, CHECKED_PTR_OF(type, val))
 #define SKM_sk_zero(type, st) \
-	sk_zero(CHECKED_PTR_OF(STACK_OF(type), st))
-#define SKM_sk_push(type, st,val) \
-	sk_push(CHECKED_PTR_OF(STACK_OF(type), st), CHECKED_PTR_OF(type, val))
-#define SKM_sk_unshift(type, st,val) \
-	sk_unshift(CHECKED_PTR_OF(STACK_OF(type), st), CHECKED_PTR_OF(type, val))
-#define SKM_sk_find(type, st,val) \
-	sk_find(CHECKED_PTR_OF(STACK_OF(type), st), CHECKED_PTR_OF(type, val))
-#define SKM_sk_delete(type, st,i) \
-	(type *)sk_delete(CHECKED_PTR_OF(STACK_OF(type), st), i)
-#define SKM_sk_delete_ptr(type, st,ptr) \
-	(type *)sk_delete_ptr(CHECKED_PTR_OF(STACK_OF(type), st), CHECKED_PTR_OF(type, ptr))
-#define SKM_sk_insert(type, st,val,i) \
-	sk_insert(CHECKED_PTR_OF(STACK_OF(type), st), CHECKED_PTR_OF(type, val), i)
-#define SKM_sk_set_cmp_func(type, st,cmp) \
+	sk_zero(CHECKED_STACK_OF(type, st))
+#define SKM_sk_push(type, st, val) \
+	sk_push(CHECKED_STACK_OF(type, st), CHECKED_PTR_OF(type, val))
+#define SKM_sk_unshift(type, st, val) \
+	sk_unshift(CHECKED_STACK_OF(type, st), CHECKED_PTR_OF(type, val))
+#define SKM_sk_find(type, st, val) \
+	sk_find(CHECKED_STACK_OF(type, st), CHECKED_PTR_OF(type, val))
+#define SKM_sk_find_ex(type, st, val) \
+	sk_find_ex(CHECKED_STACK_OF(type, st), \
+		   CHECKED_PTR_OF(type, val))
+#define SKM_sk_delete(type, st, i) \
+	(type *)sk_delete(CHECKED_STACK_OF(type, st), i)
+#define SKM_sk_delete_ptr(type, st, ptr) \
+	(type *)sk_delete_ptr(CHECKED_STACK_OF(type, st), CHECKED_PTR_OF(type, ptr))
+#define SKM_sk_insert(type, st,val, i) \
+	sk_insert(CHECKED_STACK_OF(type, st), CHECKED_PTR_OF(type, val), i)
+#define SKM_sk_set_cmp_func(type, st, cmp) \
 	((int (*)(const type * const *,const type * const *)) \
-	sk_set_cmp_func(CHECKED_PTR_OF(STACK_OF(type), st), CHECKED_SK_CMP_FUNC(type, cmp)))
+	sk_set_cmp_func(CHECKED_STACK_OF(type, st), CHECKED_SK_CMP_FUNC(type, cmp)))
 #define SKM_sk_dup(type, st) \
-	(STACK_OF(type) *)sk_dup(CHECKED_PTR_OF(STACK_OF(type), st))
-#define SKM_sk_pop_free(type, st,free_func) \
-	sk_pop_free(CHECKED_PTR_OF(STACK_OF(type), st), CHECKED_SK_FREE_FUNC(type, free_func))
+	(STACK_OF(type) *)sk_dup(CHECKED_STACK_OF(type, st))
+#define SKM_sk_pop_free(type, st, free_func) \
+	sk_pop_free(CHECKED_STACK_OF(type, st), CHECKED_SK_FREE_FUNC(type, free_func))
 #define SKM_sk_shift(type, st) \
-	(type *)sk_shift(CHECKED_PTR_OF(STACK_OF(type), st))
+	(type *)sk_shift(CHECKED_STACK_OF(type, st))
 #define SKM_sk_pop(type, st) \
-	(type *)sk_pop(CHECKED_PTR_OF(STACK_OF(type), st))
+	(type *)sk_pop(CHECKED_STACK_OF(type, st))
 #define SKM_sk_sort(type, st) \
-	sk_sort(CHECKED_PTR_OF(STACK_OF(type), st))
+	sk_sort(CHECKED_STACK_OF(type, st))
 #define SKM_sk_is_sorted(type, st) \
-	sk_is_sorted(CHECKED_PTR_OF(STACK_OF(type), st))
+	sk_is_sorted(CHECKED_STACK_OF(type, st))
 
 #define	SKM_ASN1_SET_OF_d2i(type, st, pp, length, d2i_func, free_func, ex_tag, ex_class) \
-	(STACK_OF(type) *)d2i_ASN1_SET(CHECKED_PTR_OF(STACK_OF(type)*, st), \
+  (STACK_OF(type) *)d2i_ASN1_SET( \
+				(STACK_OF(OPENSSL_BLOCK) **)CHECKED_PTR_OF(STACK_OF(type)*, st), \
 				pp, length, \
 				CHECKED_D2I_OF(type, d2i_func), \
 				CHECKED_SK_FREE_FUNC(type, free_func), \
 				ex_tag, ex_class)
 
 #define	SKM_ASN1_SET_OF_i2d(type, st, pp, i2d_func, ex_tag, ex_class, is_set) \
-	i2d_ASN1_SET(CHECKED_PTR_OF(STACK_OF(type), st), pp, \
+  i2d_ASN1_SET((STACK_OF(OPENSSL_BLOCK) *)CHECKED_STACK_OF(type, st), pp, \
 				CHECKED_I2D_OF(type, i2d_func), \
 				ex_tag, ex_class, is_set)
 
@@ -151,72 +204,8 @@ STACK_OF(type) \
 				CHECKED_SK_FREE_FUNC(type, free_func), \
 				pass, passlen, oct, seq)
 
-#else
-
-#define STACK_OF(type) STACK
-#define PREDECLARE_STACK_OF(type) /* nada */
-#define DECLARE_STACK_OF(type)    /* nada */
-#define IMPLEMENT_STACK_OF(type)  /* nada */
-
-#define SKM_sk_new(type, cmp) \
-	sk_new((int (*)(const char * const *, const char * const *))(cmp))
-#define SKM_sk_new_null(type) \
-	sk_new_null()
-#define SKM_sk_free(type, st) \
-	sk_free(st)
-#define SKM_sk_num(type, st) \
-	sk_num(st)
-#define SKM_sk_value(type, st,i) \
-	((type *)sk_value(st, i))
-#define SKM_sk_set(type, st,i,val) \
-	((type *)sk_set(st, i,(char *)val))
-#define SKM_sk_zero(type, st) \
-	sk_zero(st)
-#define SKM_sk_push(type, st,val) \
-	sk_push(st, (char *)val)
-#define SKM_sk_unshift(type, st,val) \
-	sk_unshift(st, (char *)val)
-#define SKM_sk_find(type, st,val) \
-	sk_find(st, (char *)val)
-#define SKM_sk_delete(type, st,i) \
-	((type *)sk_delete(st, i))
-#define SKM_sk_delete_ptr(type, st,ptr) \
-	((type *)sk_delete_ptr(st,(char *)ptr))
-#define SKM_sk_insert(type, st,val,i) \
-	sk_insert(st, (char *)val, i)
-#define SKM_sk_set_cmp_func(type, st,cmp) \
-	((int (*)(const type * const *,const type * const *)) \
-	sk_set_cmp_func(st, (int (*)(const char * const *, const char * const *))(cmp)))
-#define SKM_sk_dup(type, st) \
-	sk_dup(st)
-#define SKM_sk_pop_free(type, st,free_func) \
-	sk_pop_free(st, (void (*)(void *))free_func)
-#define SKM_sk_shift(type, st) \
-	((type *)sk_shift(st))
-#define SKM_sk_pop(type, st) \
-	((type *)sk_pop(st))
-#define SKM_sk_sort(type, st) \
-	sk_sort(st)
-#define SKM_sk_is_sorted(type, st) \
-	sk_is_sorted(st)
-
-#define	SKM_ASN1_SET_OF_d2i(type, st, pp, length, d2i_func, free_func, ex_tag, ex_class) \
-	d2i_ASN1_SET(st,pp,length, (void *(*)(void ** ,const unsigned char ** ,long))d2i_func, (void (*)(void *))free_func, ex_tag,ex_class)
-#define	SKM_ASN1_SET_OF_i2d(type, st, pp, i2d_func, ex_tag, ex_class, is_set) \
-	i2d_ASN1_SET(st,pp,(int (*)(void *, unsigned char **))i2d_func,ex_tag,ex_class,is_set)
-
-#define	SKM_ASN1_seq_pack(type, st, i2d_func, buf, len) \
-	ASN1_seq_pack(st, (int (*)(void *, unsigned char **))i2d_func, buf, len)
-#define	SKM_ASN1_seq_unpack(type, buf, len, d2i_func, free_func) \
-	ASN1_seq_unpack(buf,len,(void *(*)(void **,const unsigned char **,long))d2i_func, (void(*)(void *))free_func)
-
-#define SKM_PKCS12_decrypt_d2i(type, algor, d2i_func, free_func, pass, passlen, oct, seq) \
-	((STACK *)PKCS12_decrypt_d2i(algor,(char *(*)())d2i_func, (void(*)(void *))free_func,pass,passlen,oct,seq))
-
-#endif
-
 /* This block of defines is updated by util/mkstack.pl, please do not touch! */
-#define sk_ACCESS_DESCRIPTION_new(st) SKM_sk_new(ACCESS_DESCRIPTION, (st))
+#define sk_ACCESS_DESCRIPTION_new(cmp) SKM_sk_new(ACCESS_DESCRIPTION, (cmp))
 #define sk_ACCESS_DESCRIPTION_new_null() SKM_sk_new_null(ACCESS_DESCRIPTION)
 #define sk_ACCESS_DESCRIPTION_free(st) SKM_sk_free(ACCESS_DESCRIPTION, (st))
 #define sk_ACCESS_DESCRIPTION_num(st) SKM_sk_num(ACCESS_DESCRIPTION, (st))
@@ -238,7 +227,7 @@ STACK_OF(type) \
 #define sk_ACCESS_DESCRIPTION_sort(st) SKM_sk_sort(ACCESS_DESCRIPTION, (st))
 #define sk_ACCESS_DESCRIPTION_is_sorted(st) SKM_sk_is_sorted(ACCESS_DESCRIPTION, (st))
 
-#define sk_ASIdOrRange_new(st) SKM_sk_new(ASIdOrRange, (st))
+#define sk_ASIdOrRange_new(cmp) SKM_sk_new(ASIdOrRange, (cmp))
 #define sk_ASIdOrRange_new_null() SKM_sk_new_null(ASIdOrRange)
 #define sk_ASIdOrRange_free(st) SKM_sk_free(ASIdOrRange, (st))
 #define sk_ASIdOrRange_num(st) SKM_sk_num(ASIdOrRange, (st))
@@ -260,7 +249,7 @@ STACK_OF(type) \
 #define sk_ASIdOrRange_sort(st) SKM_sk_sort(ASIdOrRange, (st))
 #define sk_ASIdOrRange_is_sorted(st) SKM_sk_is_sorted(ASIdOrRange, (st))
 
-#define sk_ASN1_GENERALSTRING_new(st) SKM_sk_new(ASN1_GENERALSTRING, (st))
+#define sk_ASN1_GENERALSTRING_new(cmp) SKM_sk_new(ASN1_GENERALSTRING, (cmp))
 #define sk_ASN1_GENERALSTRING_new_null() SKM_sk_new_null(ASN1_GENERALSTRING)
 #define sk_ASN1_GENERALSTRING_free(st) SKM_sk_free(ASN1_GENERALSTRING, (st))
 #define sk_ASN1_GENERALSTRING_num(st) SKM_sk_num(ASN1_GENERALSTRING, (st))
@@ -282,7 +271,7 @@ STACK_OF(type) \
 #define sk_ASN1_GENERALSTRING_sort(st) SKM_sk_sort(ASN1_GENERALSTRING, (st))
 #define sk_ASN1_GENERALSTRING_is_sorted(st) SKM_sk_is_sorted(ASN1_GENERALSTRING, (st))
 
-#define sk_ASN1_INTEGER_new(st) SKM_sk_new(ASN1_INTEGER, (st))
+#define sk_ASN1_INTEGER_new(cmp) SKM_sk_new(ASN1_INTEGER, (cmp))
 #define sk_ASN1_INTEGER_new_null() SKM_sk_new_null(ASN1_INTEGER)
 #define sk_ASN1_INTEGER_free(st) SKM_sk_free(ASN1_INTEGER, (st))
 #define sk_ASN1_INTEGER_num(st) SKM_sk_num(ASN1_INTEGER, (st))
@@ -304,7 +293,7 @@ STACK_OF(type) \
 #define sk_ASN1_INTEGER_sort(st) SKM_sk_sort(ASN1_INTEGER, (st))
 #define sk_ASN1_INTEGER_is_sorted(st) SKM_sk_is_sorted(ASN1_INTEGER, (st))
 
-#define sk_ASN1_OBJECT_new(st) SKM_sk_new(ASN1_OBJECT, (st))
+#define sk_ASN1_OBJECT_new(cmp) SKM_sk_new(ASN1_OBJECT, (cmp))
 #define sk_ASN1_OBJECT_new_null() SKM_sk_new_null(ASN1_OBJECT)
 #define sk_ASN1_OBJECT_free(st) SKM_sk_free(ASN1_OBJECT, (st))
 #define sk_ASN1_OBJECT_num(st) SKM_sk_num(ASN1_OBJECT, (st))
@@ -326,7 +315,7 @@ STACK_OF(type) \
 #define sk_ASN1_OBJECT_sort(st) SKM_sk_sort(ASN1_OBJECT, (st))
 #define sk_ASN1_OBJECT_is_sorted(st) SKM_sk_is_sorted(ASN1_OBJECT, (st))
 
-#define sk_ASN1_STRING_TABLE_new(st) SKM_sk_new(ASN1_STRING_TABLE, (st))
+#define sk_ASN1_STRING_TABLE_new(cmp) SKM_sk_new(ASN1_STRING_TABLE, (cmp))
 #define sk_ASN1_STRING_TABLE_new_null() SKM_sk_new_null(ASN1_STRING_TABLE)
 #define sk_ASN1_STRING_TABLE_free(st) SKM_sk_free(ASN1_STRING_TABLE, (st))
 #define sk_ASN1_STRING_TABLE_num(st) SKM_sk_num(ASN1_STRING_TABLE, (st))
@@ -348,7 +337,7 @@ STACK_OF(type) \
 #define sk_ASN1_STRING_TABLE_sort(st) SKM_sk_sort(ASN1_STRING_TABLE, (st))
 #define sk_ASN1_STRING_TABLE_is_sorted(st) SKM_sk_is_sorted(ASN1_STRING_TABLE, (st))
 
-#define sk_ASN1_TYPE_new(st) SKM_sk_new(ASN1_TYPE, (st))
+#define sk_ASN1_TYPE_new(cmp) SKM_sk_new(ASN1_TYPE, (cmp))
 #define sk_ASN1_TYPE_new_null() SKM_sk_new_null(ASN1_TYPE)
 #define sk_ASN1_TYPE_free(st) SKM_sk_free(ASN1_TYPE, (st))
 #define sk_ASN1_TYPE_num(st) SKM_sk_num(ASN1_TYPE, (st))
@@ -370,7 +359,29 @@ STACK_OF(type) \
 #define sk_ASN1_TYPE_sort(st) SKM_sk_sort(ASN1_TYPE, (st))
 #define sk_ASN1_TYPE_is_sorted(st) SKM_sk_is_sorted(ASN1_TYPE, (st))
 
-#define sk_ASN1_VALUE_new(st) SKM_sk_new(ASN1_VALUE, (st))
+#define sk_ASN1_UTF8STRING_new(cmp) SKM_sk_new(ASN1_UTF8STRING, (cmp))
+#define sk_ASN1_UTF8STRING_new_null() SKM_sk_new_null(ASN1_UTF8STRING)
+#define sk_ASN1_UTF8STRING_free(st) SKM_sk_free(ASN1_UTF8STRING, (st))
+#define sk_ASN1_UTF8STRING_num(st) SKM_sk_num(ASN1_UTF8STRING, (st))
+#define sk_ASN1_UTF8STRING_value(st, i) SKM_sk_value(ASN1_UTF8STRING, (st), (i))
+#define sk_ASN1_UTF8STRING_set(st, i, val) SKM_sk_set(ASN1_UTF8STRING, (st), (i), (val))
+#define sk_ASN1_UTF8STRING_zero(st) SKM_sk_zero(ASN1_UTF8STRING, (st))
+#define sk_ASN1_UTF8STRING_push(st, val) SKM_sk_push(ASN1_UTF8STRING, (st), (val))
+#define sk_ASN1_UTF8STRING_unshift(st, val) SKM_sk_unshift(ASN1_UTF8STRING, (st), (val))
+#define sk_ASN1_UTF8STRING_find(st, val) SKM_sk_find(ASN1_UTF8STRING, (st), (val))
+#define sk_ASN1_UTF8STRING_find_ex(st, val) SKM_sk_find_ex(ASN1_UTF8STRING, (st), (val))
+#define sk_ASN1_UTF8STRING_delete(st, i) SKM_sk_delete(ASN1_UTF8STRING, (st), (i))
+#define sk_ASN1_UTF8STRING_delete_ptr(st, ptr) SKM_sk_delete_ptr(ASN1_UTF8STRING, (st), (ptr))
+#define sk_ASN1_UTF8STRING_insert(st, val, i) SKM_sk_insert(ASN1_UTF8STRING, (st), (val), (i))
+#define sk_ASN1_UTF8STRING_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(ASN1_UTF8STRING, (st), (cmp))
+#define sk_ASN1_UTF8STRING_dup(st) SKM_sk_dup(ASN1_UTF8STRING, st)
+#define sk_ASN1_UTF8STRING_pop_free(st, free_func) SKM_sk_pop_free(ASN1_UTF8STRING, (st), (free_func))
+#define sk_ASN1_UTF8STRING_shift(st) SKM_sk_shift(ASN1_UTF8STRING, (st))
+#define sk_ASN1_UTF8STRING_pop(st) SKM_sk_pop(ASN1_UTF8STRING, (st))
+#define sk_ASN1_UTF8STRING_sort(st) SKM_sk_sort(ASN1_UTF8STRING, (st))
+#define sk_ASN1_UTF8STRING_is_sorted(st) SKM_sk_is_sorted(ASN1_UTF8STRING, (st))
+
+#define sk_ASN1_VALUE_new(cmp) SKM_sk_new(ASN1_VALUE, (cmp))
 #define sk_ASN1_VALUE_new_null() SKM_sk_new_null(ASN1_VALUE)
 #define sk_ASN1_VALUE_free(st) SKM_sk_free(ASN1_VALUE, (st))
 #define sk_ASN1_VALUE_num(st) SKM_sk_num(ASN1_VALUE, (st))
@@ -392,7 +403,7 @@ STACK_OF(type) \
 #define sk_ASN1_VALUE_sort(st) SKM_sk_sort(ASN1_VALUE, (st))
 #define sk_ASN1_VALUE_is_sorted(st) SKM_sk_is_sorted(ASN1_VALUE, (st))
 
-#define sk_BIO_new(st) SKM_sk_new(BIO, (st))
+#define sk_BIO_new(cmp) SKM_sk_new(BIO, (cmp))
 #define sk_BIO_new_null() SKM_sk_new_null(BIO)
 #define sk_BIO_free(st) SKM_sk_free(BIO, (st))
 #define sk_BIO_num(st) SKM_sk_num(BIO, (st))
@@ -414,7 +425,51 @@ STACK_OF(type) \
 #define sk_BIO_sort(st) SKM_sk_sort(BIO, (st))
 #define sk_BIO_is_sorted(st) SKM_sk_is_sorted(BIO, (st))
 
-#define sk_CMS_CertificateChoices_new(st) SKM_sk_new(CMS_CertificateChoices, (st))
+#define sk_BY_DIR_ENTRY_new(cmp) SKM_sk_new(BY_DIR_ENTRY, (cmp))
+#define sk_BY_DIR_ENTRY_new_null() SKM_sk_new_null(BY_DIR_ENTRY)
+#define sk_BY_DIR_ENTRY_free(st) SKM_sk_free(BY_DIR_ENTRY, (st))
+#define sk_BY_DIR_ENTRY_num(st) SKM_sk_num(BY_DIR_ENTRY, (st))
+#define sk_BY_DIR_ENTRY_value(st, i) SKM_sk_value(BY_DIR_ENTRY, (st), (i))
+#define sk_BY_DIR_ENTRY_set(st, i, val) SKM_sk_set(BY_DIR_ENTRY, (st), (i), (val))
+#define sk_BY_DIR_ENTRY_zero(st) SKM_sk_zero(BY_DIR_ENTRY, (st))
+#define sk_BY_DIR_ENTRY_push(st, val) SKM_sk_push(BY_DIR_ENTRY, (st), (val))
+#define sk_BY_DIR_ENTRY_unshift(st, val) SKM_sk_unshift(BY_DIR_ENTRY, (st), (val))
+#define sk_BY_DIR_ENTRY_find(st, val) SKM_sk_find(BY_DIR_ENTRY, (st), (val))
+#define sk_BY_DIR_ENTRY_find_ex(st, val) SKM_sk_find_ex(BY_DIR_ENTRY, (st), (val))
+#define sk_BY_DIR_ENTRY_delete(st, i) SKM_sk_delete(BY_DIR_ENTRY, (st), (i))
+#define sk_BY_DIR_ENTRY_delete_ptr(st, ptr) SKM_sk_delete_ptr(BY_DIR_ENTRY, (st), (ptr))
+#define sk_BY_DIR_ENTRY_insert(st, val, i) SKM_sk_insert(BY_DIR_ENTRY, (st), (val), (i))
+#define sk_BY_DIR_ENTRY_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(BY_DIR_ENTRY, (st), (cmp))
+#define sk_BY_DIR_ENTRY_dup(st) SKM_sk_dup(BY_DIR_ENTRY, st)
+#define sk_BY_DIR_ENTRY_pop_free(st, free_func) SKM_sk_pop_free(BY_DIR_ENTRY, (st), (free_func))
+#define sk_BY_DIR_ENTRY_shift(st) SKM_sk_shift(BY_DIR_ENTRY, (st))
+#define sk_BY_DIR_ENTRY_pop(st) SKM_sk_pop(BY_DIR_ENTRY, (st))
+#define sk_BY_DIR_ENTRY_sort(st) SKM_sk_sort(BY_DIR_ENTRY, (st))
+#define sk_BY_DIR_ENTRY_is_sorted(st) SKM_sk_is_sorted(BY_DIR_ENTRY, (st))
+
+#define sk_BY_DIR_HASH_new(cmp) SKM_sk_new(BY_DIR_HASH, (cmp))
+#define sk_BY_DIR_HASH_new_null() SKM_sk_new_null(BY_DIR_HASH)
+#define sk_BY_DIR_HASH_free(st) SKM_sk_free(BY_DIR_HASH, (st))
+#define sk_BY_DIR_HASH_num(st) SKM_sk_num(BY_DIR_HASH, (st))
+#define sk_BY_DIR_HASH_value(st, i) SKM_sk_value(BY_DIR_HASH, (st), (i))
+#define sk_BY_DIR_HASH_set(st, i, val) SKM_sk_set(BY_DIR_HASH, (st), (i), (val))
+#define sk_BY_DIR_HASH_zero(st) SKM_sk_zero(BY_DIR_HASH, (st))
+#define sk_BY_DIR_HASH_push(st, val) SKM_sk_push(BY_DIR_HASH, (st), (val))
+#define sk_BY_DIR_HASH_unshift(st, val) SKM_sk_unshift(BY_DIR_HASH, (st), (val))
+#define sk_BY_DIR_HASH_find(st, val) SKM_sk_find(BY_DIR_HASH, (st), (val))
+#define sk_BY_DIR_HASH_find_ex(st, val) SKM_sk_find_ex(BY_DIR_HASH, (st), (val))
+#define sk_BY_DIR_HASH_delete(st, i) SKM_sk_delete(BY_DIR_HASH, (st), (i))
+#define sk_BY_DIR_HASH_delete_ptr(st, ptr) SKM_sk_delete_ptr(BY_DIR_HASH, (st), (ptr))
+#define sk_BY_DIR_HASH_insert(st, val, i) SKM_sk_insert(BY_DIR_HASH, (st), (val), (i))
+#define sk_BY_DIR_HASH_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(BY_DIR_HASH, (st), (cmp))
+#define sk_BY_DIR_HASH_dup(st) SKM_sk_dup(BY_DIR_HASH, st)
+#define sk_BY_DIR_HASH_pop_free(st, free_func) SKM_sk_pop_free(BY_DIR_HASH, (st), (free_func))
+#define sk_BY_DIR_HASH_shift(st) SKM_sk_shift(BY_DIR_HASH, (st))
+#define sk_BY_DIR_HASH_pop(st) SKM_sk_pop(BY_DIR_HASH, (st))
+#define sk_BY_DIR_HASH_sort(st) SKM_sk_sort(BY_DIR_HASH, (st))
+#define sk_BY_DIR_HASH_is_sorted(st) SKM_sk_is_sorted(BY_DIR_HASH, (st))
+
+#define sk_CMS_CertificateChoices_new(cmp) SKM_sk_new(CMS_CertificateChoices, (cmp))
 #define sk_CMS_CertificateChoices_new_null() SKM_sk_new_null(CMS_CertificateChoices)
 #define sk_CMS_CertificateChoices_free(st) SKM_sk_free(CMS_CertificateChoices, (st))
 #define sk_CMS_CertificateChoices_num(st) SKM_sk_num(CMS_CertificateChoices, (st))
@@ -436,7 +491,7 @@ STACK_OF(type) \
 #define sk_CMS_CertificateChoices_sort(st) SKM_sk_sort(CMS_CertificateChoices, (st))
 #define sk_CMS_CertificateChoices_is_sorted(st) SKM_sk_is_sorted(CMS_CertificateChoices, (st))
 
-#define sk_CMS_RecipientInfo_new(st) SKM_sk_new(CMS_RecipientInfo, (st))
+#define sk_CMS_RecipientInfo_new(cmp) SKM_sk_new(CMS_RecipientInfo, (cmp))
 #define sk_CMS_RecipientInfo_new_null() SKM_sk_new_null(CMS_RecipientInfo)
 #define sk_CMS_RecipientInfo_free(st) SKM_sk_free(CMS_RecipientInfo, (st))
 #define sk_CMS_RecipientInfo_num(st) SKM_sk_num(CMS_RecipientInfo, (st))
@@ -458,7 +513,7 @@ STACK_OF(type) \
 #define sk_CMS_RecipientInfo_sort(st) SKM_sk_sort(CMS_RecipientInfo, (st))
 #define sk_CMS_RecipientInfo_is_sorted(st) SKM_sk_is_sorted(CMS_RecipientInfo, (st))
 
-#define sk_CMS_RevocationInfoChoice_new(st) SKM_sk_new(CMS_RevocationInfoChoice, (st))
+#define sk_CMS_RevocationInfoChoice_new(cmp) SKM_sk_new(CMS_RevocationInfoChoice, (cmp))
 #define sk_CMS_RevocationInfoChoice_new_null() SKM_sk_new_null(CMS_RevocationInfoChoice)
 #define sk_CMS_RevocationInfoChoice_free(st) SKM_sk_free(CMS_RevocationInfoChoice, (st))
 #define sk_CMS_RevocationInfoChoice_num(st) SKM_sk_num(CMS_RevocationInfoChoice, (st))
@@ -480,7 +535,7 @@ STACK_OF(type) \
 #define sk_CMS_RevocationInfoChoice_sort(st) SKM_sk_sort(CMS_RevocationInfoChoice, (st))
 #define sk_CMS_RevocationInfoChoice_is_sorted(st) SKM_sk_is_sorted(CMS_RevocationInfoChoice, (st))
 
-#define sk_CMS_SignerInfo_new(st) SKM_sk_new(CMS_SignerInfo, (st))
+#define sk_CMS_SignerInfo_new(cmp) SKM_sk_new(CMS_SignerInfo, (cmp))
 #define sk_CMS_SignerInfo_new_null() SKM_sk_new_null(CMS_SignerInfo)
 #define sk_CMS_SignerInfo_free(st) SKM_sk_free(CMS_SignerInfo, (st))
 #define sk_CMS_SignerInfo_num(st) SKM_sk_num(CMS_SignerInfo, (st))
@@ -502,7 +557,7 @@ STACK_OF(type) \
 #define sk_CMS_SignerInfo_sort(st) SKM_sk_sort(CMS_SignerInfo, (st))
 #define sk_CMS_SignerInfo_is_sorted(st) SKM_sk_is_sorted(CMS_SignerInfo, (st))
 
-#define sk_CONF_IMODULE_new(st) SKM_sk_new(CONF_IMODULE, (st))
+#define sk_CONF_IMODULE_new(cmp) SKM_sk_new(CONF_IMODULE, (cmp))
 #define sk_CONF_IMODULE_new_null() SKM_sk_new_null(CONF_IMODULE)
 #define sk_CONF_IMODULE_free(st) SKM_sk_free(CONF_IMODULE, (st))
 #define sk_CONF_IMODULE_num(st) SKM_sk_num(CONF_IMODULE, (st))
@@ -524,7 +579,7 @@ STACK_OF(type) \
 #define sk_CONF_IMODULE_sort(st) SKM_sk_sort(CONF_IMODULE, (st))
 #define sk_CONF_IMODULE_is_sorted(st) SKM_sk_is_sorted(CONF_IMODULE, (st))
 
-#define sk_CONF_MODULE_new(st) SKM_sk_new(CONF_MODULE, (st))
+#define sk_CONF_MODULE_new(cmp) SKM_sk_new(CONF_MODULE, (cmp))
 #define sk_CONF_MODULE_new_null() SKM_sk_new_null(CONF_MODULE)
 #define sk_CONF_MODULE_free(st) SKM_sk_free(CONF_MODULE, (st))
 #define sk_CONF_MODULE_num(st) SKM_sk_num(CONF_MODULE, (st))
@@ -546,7 +601,7 @@ STACK_OF(type) \
 #define sk_CONF_MODULE_sort(st) SKM_sk_sort(CONF_MODULE, (st))
 #define sk_CONF_MODULE_is_sorted(st) SKM_sk_is_sorted(CONF_MODULE, (st))
 
-#define sk_CONF_VALUE_new(st) SKM_sk_new(CONF_VALUE, (st))
+#define sk_CONF_VALUE_new(cmp) SKM_sk_new(CONF_VALUE, (cmp))
 #define sk_CONF_VALUE_new_null() SKM_sk_new_null(CONF_VALUE)
 #define sk_CONF_VALUE_free(st) SKM_sk_free(CONF_VALUE, (st))
 #define sk_CONF_VALUE_num(st) SKM_sk_num(CONF_VALUE, (st))
@@ -568,7 +623,7 @@ STACK_OF(type) \
 #define sk_CONF_VALUE_sort(st) SKM_sk_sort(CONF_VALUE, (st))
 #define sk_CONF_VALUE_is_sorted(st) SKM_sk_is_sorted(CONF_VALUE, (st))
 
-#define sk_CRYPTO_EX_DATA_FUNCS_new(st) SKM_sk_new(CRYPTO_EX_DATA_FUNCS, (st))
+#define sk_CRYPTO_EX_DATA_FUNCS_new(cmp) SKM_sk_new(CRYPTO_EX_DATA_FUNCS, (cmp))
 #define sk_CRYPTO_EX_DATA_FUNCS_new_null() SKM_sk_new_null(CRYPTO_EX_DATA_FUNCS)
 #define sk_CRYPTO_EX_DATA_FUNCS_free(st) SKM_sk_free(CRYPTO_EX_DATA_FUNCS, (st))
 #define sk_CRYPTO_EX_DATA_FUNCS_num(st) SKM_sk_num(CRYPTO_EX_DATA_FUNCS, (st))
@@ -590,7 +645,7 @@ STACK_OF(type) \
 #define sk_CRYPTO_EX_DATA_FUNCS_sort(st) SKM_sk_sort(CRYPTO_EX_DATA_FUNCS, (st))
 #define sk_CRYPTO_EX_DATA_FUNCS_is_sorted(st) SKM_sk_is_sorted(CRYPTO_EX_DATA_FUNCS, (st))
 
-#define sk_CRYPTO_dynlock_new(st) SKM_sk_new(CRYPTO_dynlock, (st))
+#define sk_CRYPTO_dynlock_new(cmp) SKM_sk_new(CRYPTO_dynlock, (cmp))
 #define sk_CRYPTO_dynlock_new_null() SKM_sk_new_null(CRYPTO_dynlock)
 #define sk_CRYPTO_dynlock_free(st) SKM_sk_free(CRYPTO_dynlock, (st))
 #define sk_CRYPTO_dynlock_num(st) SKM_sk_num(CRYPTO_dynlock, (st))
@@ -612,7 +667,7 @@ STACK_OF(type) \
 #define sk_CRYPTO_dynlock_sort(st) SKM_sk_sort(CRYPTO_dynlock, (st))
 #define sk_CRYPTO_dynlock_is_sorted(st) SKM_sk_is_sorted(CRYPTO_dynlock, (st))
 
-#define sk_DIST_POINT_new(st) SKM_sk_new(DIST_POINT, (st))
+#define sk_DIST_POINT_new(cmp) SKM_sk_new(DIST_POINT, (cmp))
 #define sk_DIST_POINT_new_null() SKM_sk_new_null(DIST_POINT)
 #define sk_DIST_POINT_free(st) SKM_sk_free(DIST_POINT, (st))
 #define sk_DIST_POINT_num(st) SKM_sk_num(DIST_POINT, (st))
@@ -634,7 +689,7 @@ STACK_OF(type) \
 #define sk_DIST_POINT_sort(st) SKM_sk_sort(DIST_POINT, (st))
 #define sk_DIST_POINT_is_sorted(st) SKM_sk_is_sorted(DIST_POINT, (st))
 
-#define sk_ENGINE_new(st) SKM_sk_new(ENGINE, (st))
+#define sk_ENGINE_new(cmp) SKM_sk_new(ENGINE, (cmp))
 #define sk_ENGINE_new_null() SKM_sk_new_null(ENGINE)
 #define sk_ENGINE_free(st) SKM_sk_free(ENGINE, (st))
 #define sk_ENGINE_num(st) SKM_sk_num(ENGINE, (st))
@@ -656,7 +711,7 @@ STACK_OF(type) \
 #define sk_ENGINE_sort(st) SKM_sk_sort(ENGINE, (st))
 #define sk_ENGINE_is_sorted(st) SKM_sk_is_sorted(ENGINE, (st))
 
-#define sk_ENGINE_CLEANUP_ITEM_new(st) SKM_sk_new(ENGINE_CLEANUP_ITEM, (st))
+#define sk_ENGINE_CLEANUP_ITEM_new(cmp) SKM_sk_new(ENGINE_CLEANUP_ITEM, (cmp))
 #define sk_ENGINE_CLEANUP_ITEM_new_null() SKM_sk_new_null(ENGINE_CLEANUP_ITEM)
 #define sk_ENGINE_CLEANUP_ITEM_free(st) SKM_sk_free(ENGINE_CLEANUP_ITEM, (st))
 #define sk_ENGINE_CLEANUP_ITEM_num(st) SKM_sk_num(ENGINE_CLEANUP_ITEM, (st))
@@ -678,7 +733,117 @@ STACK_OF(type) \
 #define sk_ENGINE_CLEANUP_ITEM_sort(st) SKM_sk_sort(ENGINE_CLEANUP_ITEM, (st))
 #define sk_ENGINE_CLEANUP_ITEM_is_sorted(st) SKM_sk_is_sorted(ENGINE_CLEANUP_ITEM, (st))
 
-#define sk_GENERAL_NAME_new(st) SKM_sk_new(GENERAL_NAME, (st))
+#define sk_ESS_CERT_ID_new(cmp) SKM_sk_new(ESS_CERT_ID, (cmp))
+#define sk_ESS_CERT_ID_new_null() SKM_sk_new_null(ESS_CERT_ID)
+#define sk_ESS_CERT_ID_free(st) SKM_sk_free(ESS_CERT_ID, (st))
+#define sk_ESS_CERT_ID_num(st) SKM_sk_num(ESS_CERT_ID, (st))
+#define sk_ESS_CERT_ID_value(st, i) SKM_sk_value(ESS_CERT_ID, (st), (i))
+#define sk_ESS_CERT_ID_set(st, i, val) SKM_sk_set(ESS_CERT_ID, (st), (i), (val))
+#define sk_ESS_CERT_ID_zero(st) SKM_sk_zero(ESS_CERT_ID, (st))
+#define sk_ESS_CERT_ID_push(st, val) SKM_sk_push(ESS_CERT_ID, (st), (val))
+#define sk_ESS_CERT_ID_unshift(st, val) SKM_sk_unshift(ESS_CERT_ID, (st), (val))
+#define sk_ESS_CERT_ID_find(st, val) SKM_sk_find(ESS_CERT_ID, (st), (val))
+#define sk_ESS_CERT_ID_find_ex(st, val) SKM_sk_find_ex(ESS_CERT_ID, (st), (val))
+#define sk_ESS_CERT_ID_delete(st, i) SKM_sk_delete(ESS_CERT_ID, (st), (i))
+#define sk_ESS_CERT_ID_delete_ptr(st, ptr) SKM_sk_delete_ptr(ESS_CERT_ID, (st), (ptr))
+#define sk_ESS_CERT_ID_insert(st, val, i) SKM_sk_insert(ESS_CERT_ID, (st), (val), (i))
+#define sk_ESS_CERT_ID_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(ESS_CERT_ID, (st), (cmp))
+#define sk_ESS_CERT_ID_dup(st) SKM_sk_dup(ESS_CERT_ID, st)
+#define sk_ESS_CERT_ID_pop_free(st, free_func) SKM_sk_pop_free(ESS_CERT_ID, (st), (free_func))
+#define sk_ESS_CERT_ID_shift(st) SKM_sk_shift(ESS_CERT_ID, (st))
+#define sk_ESS_CERT_ID_pop(st) SKM_sk_pop(ESS_CERT_ID, (st))
+#define sk_ESS_CERT_ID_sort(st) SKM_sk_sort(ESS_CERT_ID, (st))
+#define sk_ESS_CERT_ID_is_sorted(st) SKM_sk_is_sorted(ESS_CERT_ID, (st))
+
+#define sk_EVP_MD_new(cmp) SKM_sk_new(EVP_MD, (cmp))
+#define sk_EVP_MD_new_null() SKM_sk_new_null(EVP_MD)
+#define sk_EVP_MD_free(st) SKM_sk_free(EVP_MD, (st))
+#define sk_EVP_MD_num(st) SKM_sk_num(EVP_MD, (st))
+#define sk_EVP_MD_value(st, i) SKM_sk_value(EVP_MD, (st), (i))
+#define sk_EVP_MD_set(st, i, val) SKM_sk_set(EVP_MD, (st), (i), (val))
+#define sk_EVP_MD_zero(st) SKM_sk_zero(EVP_MD, (st))
+#define sk_EVP_MD_push(st, val) SKM_sk_push(EVP_MD, (st), (val))
+#define sk_EVP_MD_unshift(st, val) SKM_sk_unshift(EVP_MD, (st), (val))
+#define sk_EVP_MD_find(st, val) SKM_sk_find(EVP_MD, (st), (val))
+#define sk_EVP_MD_find_ex(st, val) SKM_sk_find_ex(EVP_MD, (st), (val))
+#define sk_EVP_MD_delete(st, i) SKM_sk_delete(EVP_MD, (st), (i))
+#define sk_EVP_MD_delete_ptr(st, ptr) SKM_sk_delete_ptr(EVP_MD, (st), (ptr))
+#define sk_EVP_MD_insert(st, val, i) SKM_sk_insert(EVP_MD, (st), (val), (i))
+#define sk_EVP_MD_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(EVP_MD, (st), (cmp))
+#define sk_EVP_MD_dup(st) SKM_sk_dup(EVP_MD, st)
+#define sk_EVP_MD_pop_free(st, free_func) SKM_sk_pop_free(EVP_MD, (st), (free_func))
+#define sk_EVP_MD_shift(st) SKM_sk_shift(EVP_MD, (st))
+#define sk_EVP_MD_pop(st) SKM_sk_pop(EVP_MD, (st))
+#define sk_EVP_MD_sort(st) SKM_sk_sort(EVP_MD, (st))
+#define sk_EVP_MD_is_sorted(st) SKM_sk_is_sorted(EVP_MD, (st))
+
+#define sk_EVP_PBE_CTL_new(cmp) SKM_sk_new(EVP_PBE_CTL, (cmp))
+#define sk_EVP_PBE_CTL_new_null() SKM_sk_new_null(EVP_PBE_CTL)
+#define sk_EVP_PBE_CTL_free(st) SKM_sk_free(EVP_PBE_CTL, (st))
+#define sk_EVP_PBE_CTL_num(st) SKM_sk_num(EVP_PBE_CTL, (st))
+#define sk_EVP_PBE_CTL_value(st, i) SKM_sk_value(EVP_PBE_CTL, (st), (i))
+#define sk_EVP_PBE_CTL_set(st, i, val) SKM_sk_set(EVP_PBE_CTL, (st), (i), (val))
+#define sk_EVP_PBE_CTL_zero(st) SKM_sk_zero(EVP_PBE_CTL, (st))
+#define sk_EVP_PBE_CTL_push(st, val) SKM_sk_push(EVP_PBE_CTL, (st), (val))
+#define sk_EVP_PBE_CTL_unshift(st, val) SKM_sk_unshift(EVP_PBE_CTL, (st), (val))
+#define sk_EVP_PBE_CTL_find(st, val) SKM_sk_find(EVP_PBE_CTL, (st), (val))
+#define sk_EVP_PBE_CTL_find_ex(st, val) SKM_sk_find_ex(EVP_PBE_CTL, (st), (val))
+#define sk_EVP_PBE_CTL_delete(st, i) SKM_sk_delete(EVP_PBE_CTL, (st), (i))
+#define sk_EVP_PBE_CTL_delete_ptr(st, ptr) SKM_sk_delete_ptr(EVP_PBE_CTL, (st), (ptr))
+#define sk_EVP_PBE_CTL_insert(st, val, i) SKM_sk_insert(EVP_PBE_CTL, (st), (val), (i))
+#define sk_EVP_PBE_CTL_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(EVP_PBE_CTL, (st), (cmp))
+#define sk_EVP_PBE_CTL_dup(st) SKM_sk_dup(EVP_PBE_CTL, st)
+#define sk_EVP_PBE_CTL_pop_free(st, free_func) SKM_sk_pop_free(EVP_PBE_CTL, (st), (free_func))
+#define sk_EVP_PBE_CTL_shift(st) SKM_sk_shift(EVP_PBE_CTL, (st))
+#define sk_EVP_PBE_CTL_pop(st) SKM_sk_pop(EVP_PBE_CTL, (st))
+#define sk_EVP_PBE_CTL_sort(st) SKM_sk_sort(EVP_PBE_CTL, (st))
+#define sk_EVP_PBE_CTL_is_sorted(st) SKM_sk_is_sorted(EVP_PBE_CTL, (st))
+
+#define sk_EVP_PKEY_ASN1_METHOD_new(cmp) SKM_sk_new(EVP_PKEY_ASN1_METHOD, (cmp))
+#define sk_EVP_PKEY_ASN1_METHOD_new_null() SKM_sk_new_null(EVP_PKEY_ASN1_METHOD)
+#define sk_EVP_PKEY_ASN1_METHOD_free(st) SKM_sk_free(EVP_PKEY_ASN1_METHOD, (st))
+#define sk_EVP_PKEY_ASN1_METHOD_num(st) SKM_sk_num(EVP_PKEY_ASN1_METHOD, (st))
+#define sk_EVP_PKEY_ASN1_METHOD_value(st, i) SKM_sk_value(EVP_PKEY_ASN1_METHOD, (st), (i))
+#define sk_EVP_PKEY_ASN1_METHOD_set(st, i, val) SKM_sk_set(EVP_PKEY_ASN1_METHOD, (st), (i), (val))
+#define sk_EVP_PKEY_ASN1_METHOD_zero(st) SKM_sk_zero(EVP_PKEY_ASN1_METHOD, (st))
+#define sk_EVP_PKEY_ASN1_METHOD_push(st, val) SKM_sk_push(EVP_PKEY_ASN1_METHOD, (st), (val))
+#define sk_EVP_PKEY_ASN1_METHOD_unshift(st, val) SKM_sk_unshift(EVP_PKEY_ASN1_METHOD, (st), (val))
+#define sk_EVP_PKEY_ASN1_METHOD_find(st, val) SKM_sk_find(EVP_PKEY_ASN1_METHOD, (st), (val))
+#define sk_EVP_PKEY_ASN1_METHOD_find_ex(st, val) SKM_sk_find_ex(EVP_PKEY_ASN1_METHOD, (st), (val))
+#define sk_EVP_PKEY_ASN1_METHOD_delete(st, i) SKM_sk_delete(EVP_PKEY_ASN1_METHOD, (st), (i))
+#define sk_EVP_PKEY_ASN1_METHOD_delete_ptr(st, ptr) SKM_sk_delete_ptr(EVP_PKEY_ASN1_METHOD, (st), (ptr))
+#define sk_EVP_PKEY_ASN1_METHOD_insert(st, val, i) SKM_sk_insert(EVP_PKEY_ASN1_METHOD, (st), (val), (i))
+#define sk_EVP_PKEY_ASN1_METHOD_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(EVP_PKEY_ASN1_METHOD, (st), (cmp))
+#define sk_EVP_PKEY_ASN1_METHOD_dup(st) SKM_sk_dup(EVP_PKEY_ASN1_METHOD, st)
+#define sk_EVP_PKEY_ASN1_METHOD_pop_free(st, free_func) SKM_sk_pop_free(EVP_PKEY_ASN1_METHOD, (st), (free_func))
+#define sk_EVP_PKEY_ASN1_METHOD_shift(st) SKM_sk_shift(EVP_PKEY_ASN1_METHOD, (st))
+#define sk_EVP_PKEY_ASN1_METHOD_pop(st) SKM_sk_pop(EVP_PKEY_ASN1_METHOD, (st))
+#define sk_EVP_PKEY_ASN1_METHOD_sort(st) SKM_sk_sort(EVP_PKEY_ASN1_METHOD, (st))
+#define sk_EVP_PKEY_ASN1_METHOD_is_sorted(st) SKM_sk_is_sorted(EVP_PKEY_ASN1_METHOD, (st))
+
+#define sk_EVP_PKEY_METHOD_new(cmp) SKM_sk_new(EVP_PKEY_METHOD, (cmp))
+#define sk_EVP_PKEY_METHOD_new_null() SKM_sk_new_null(EVP_PKEY_METHOD)
+#define sk_EVP_PKEY_METHOD_free(st) SKM_sk_free(EVP_PKEY_METHOD, (st))
+#define sk_EVP_PKEY_METHOD_num(st) SKM_sk_num(EVP_PKEY_METHOD, (st))
+#define sk_EVP_PKEY_METHOD_value(st, i) SKM_sk_value(EVP_PKEY_METHOD, (st), (i))
+#define sk_EVP_PKEY_METHOD_set(st, i, val) SKM_sk_set(EVP_PKEY_METHOD, (st), (i), (val))
+#define sk_EVP_PKEY_METHOD_zero(st) SKM_sk_zero(EVP_PKEY_METHOD, (st))
+#define sk_EVP_PKEY_METHOD_push(st, val) SKM_sk_push(EVP_PKEY_METHOD, (st), (val))
+#define sk_EVP_PKEY_METHOD_unshift(st, val) SKM_sk_unshift(EVP_PKEY_METHOD, (st), (val))
+#define sk_EVP_PKEY_METHOD_find(st, val) SKM_sk_find(EVP_PKEY_METHOD, (st), (val))
+#define sk_EVP_PKEY_METHOD_find_ex(st, val) SKM_sk_find_ex(EVP_PKEY_METHOD, (st), (val))
+#define sk_EVP_PKEY_METHOD_delete(st, i) SKM_sk_delete(EVP_PKEY_METHOD, (st), (i))
+#define sk_EVP_PKEY_METHOD_delete_ptr(st, ptr) SKM_sk_delete_ptr(EVP_PKEY_METHOD, (st), (ptr))
+#define sk_EVP_PKEY_METHOD_insert(st, val, i) SKM_sk_insert(EVP_PKEY_METHOD, (st), (val), (i))
+#define sk_EVP_PKEY_METHOD_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(EVP_PKEY_METHOD, (st), (cmp))
+#define sk_EVP_PKEY_METHOD_dup(st) SKM_sk_dup(EVP_PKEY_METHOD, st)
+#define sk_EVP_PKEY_METHOD_pop_free(st, free_func) SKM_sk_pop_free(EVP_PKEY_METHOD, (st), (free_func))
+#define sk_EVP_PKEY_METHOD_shift(st) SKM_sk_shift(EVP_PKEY_METHOD, (st))
+#define sk_EVP_PKEY_METHOD_pop(st) SKM_sk_pop(EVP_PKEY_METHOD, (st))
+#define sk_EVP_PKEY_METHOD_sort(st) SKM_sk_sort(EVP_PKEY_METHOD, (st))
+#define sk_EVP_PKEY_METHOD_is_sorted(st) SKM_sk_is_sorted(EVP_PKEY_METHOD, (st))
+
+#define sk_GENERAL_NAME_new(cmp) SKM_sk_new(GENERAL_NAME, (cmp))
 #define sk_GENERAL_NAME_new_null() SKM_sk_new_null(GENERAL_NAME)
 #define sk_GENERAL_NAME_free(st) SKM_sk_free(GENERAL_NAME, (st))
 #define sk_GENERAL_NAME_num(st) SKM_sk_num(GENERAL_NAME, (st))
@@ -700,7 +865,7 @@ STACK_OF(type) \
 #define sk_GENERAL_NAME_sort(st) SKM_sk_sort(GENERAL_NAME, (st))
 #define sk_GENERAL_NAME_is_sorted(st) SKM_sk_is_sorted(GENERAL_NAME, (st))
 
-#define sk_GENERAL_NAMES_new(st) SKM_sk_new(GENERAL_NAMES, (st))
+#define sk_GENERAL_NAMES_new(cmp) SKM_sk_new(GENERAL_NAMES, (cmp))
 #define sk_GENERAL_NAMES_new_null() SKM_sk_new_null(GENERAL_NAMES)
 #define sk_GENERAL_NAMES_free(st) SKM_sk_free(GENERAL_NAMES, (st))
 #define sk_GENERAL_NAMES_num(st) SKM_sk_num(GENERAL_NAMES, (st))
@@ -722,7 +887,7 @@ STACK_OF(type) \
 #define sk_GENERAL_NAMES_sort(st) SKM_sk_sort(GENERAL_NAMES, (st))
 #define sk_GENERAL_NAMES_is_sorted(st) SKM_sk_is_sorted(GENERAL_NAMES, (st))
 
-#define sk_GENERAL_SUBTREE_new(st) SKM_sk_new(GENERAL_SUBTREE, (st))
+#define sk_GENERAL_SUBTREE_new(cmp) SKM_sk_new(GENERAL_SUBTREE, (cmp))
 #define sk_GENERAL_SUBTREE_new_null() SKM_sk_new_null(GENERAL_SUBTREE)
 #define sk_GENERAL_SUBTREE_free(st) SKM_sk_free(GENERAL_SUBTREE, (st))
 #define sk_GENERAL_SUBTREE_num(st) SKM_sk_num(GENERAL_SUBTREE, (st))
@@ -744,7 +909,7 @@ STACK_OF(type) \
 #define sk_GENERAL_SUBTREE_sort(st) SKM_sk_sort(GENERAL_SUBTREE, (st))
 #define sk_GENERAL_SUBTREE_is_sorted(st) SKM_sk_is_sorted(GENERAL_SUBTREE, (st))
 
-#define sk_IPAddressFamily_new(st) SKM_sk_new(IPAddressFamily, (st))
+#define sk_IPAddressFamily_new(cmp) SKM_sk_new(IPAddressFamily, (cmp))
 #define sk_IPAddressFamily_new_null() SKM_sk_new_null(IPAddressFamily)
 #define sk_IPAddressFamily_free(st) SKM_sk_free(IPAddressFamily, (st))
 #define sk_IPAddressFamily_num(st) SKM_sk_num(IPAddressFamily, (st))
@@ -766,7 +931,7 @@ STACK_OF(type) \
 #define sk_IPAddressFamily_sort(st) SKM_sk_sort(IPAddressFamily, (st))
 #define sk_IPAddressFamily_is_sorted(st) SKM_sk_is_sorted(IPAddressFamily, (st))
 
-#define sk_IPAddressOrRange_new(st) SKM_sk_new(IPAddressOrRange, (st))
+#define sk_IPAddressOrRange_new(cmp) SKM_sk_new(IPAddressOrRange, (cmp))
 #define sk_IPAddressOrRange_new_null() SKM_sk_new_null(IPAddressOrRange)
 #define sk_IPAddressOrRange_free(st) SKM_sk_free(IPAddressOrRange, (st))
 #define sk_IPAddressOrRange_num(st) SKM_sk_num(IPAddressOrRange, (st))
@@ -788,7 +953,7 @@ STACK_OF(type) \
 #define sk_IPAddressOrRange_sort(st) SKM_sk_sort(IPAddressOrRange, (st))
 #define sk_IPAddressOrRange_is_sorted(st) SKM_sk_is_sorted(IPAddressOrRange, (st))
 
-#define sk_KRB5_APREQBODY_new(st) SKM_sk_new(KRB5_APREQBODY, (st))
+#define sk_KRB5_APREQBODY_new(cmp) SKM_sk_new(KRB5_APREQBODY, (cmp))
 #define sk_KRB5_APREQBODY_new_null() SKM_sk_new_null(KRB5_APREQBODY)
 #define sk_KRB5_APREQBODY_free(st) SKM_sk_free(KRB5_APREQBODY, (st))
 #define sk_KRB5_APREQBODY_num(st) SKM_sk_num(KRB5_APREQBODY, (st))
@@ -810,7 +975,7 @@ STACK_OF(type) \
 #define sk_KRB5_APREQBODY_sort(st) SKM_sk_sort(KRB5_APREQBODY, (st))
 #define sk_KRB5_APREQBODY_is_sorted(st) SKM_sk_is_sorted(KRB5_APREQBODY, (st))
 
-#define sk_KRB5_AUTHDATA_new(st) SKM_sk_new(KRB5_AUTHDATA, (st))
+#define sk_KRB5_AUTHDATA_new(cmp) SKM_sk_new(KRB5_AUTHDATA, (cmp))
 #define sk_KRB5_AUTHDATA_new_null() SKM_sk_new_null(KRB5_AUTHDATA)
 #define sk_KRB5_AUTHDATA_free(st) SKM_sk_free(KRB5_AUTHDATA, (st))
 #define sk_KRB5_AUTHDATA_num(st) SKM_sk_num(KRB5_AUTHDATA, (st))
@@ -832,7 +997,7 @@ STACK_OF(type) \
 #define sk_KRB5_AUTHDATA_sort(st) SKM_sk_sort(KRB5_AUTHDATA, (st))
 #define sk_KRB5_AUTHDATA_is_sorted(st) SKM_sk_is_sorted(KRB5_AUTHDATA, (st))
 
-#define sk_KRB5_AUTHENTBODY_new(st) SKM_sk_new(KRB5_AUTHENTBODY, (st))
+#define sk_KRB5_AUTHENTBODY_new(cmp) SKM_sk_new(KRB5_AUTHENTBODY, (cmp))
 #define sk_KRB5_AUTHENTBODY_new_null() SKM_sk_new_null(KRB5_AUTHENTBODY)
 #define sk_KRB5_AUTHENTBODY_free(st) SKM_sk_free(KRB5_AUTHENTBODY, (st))
 #define sk_KRB5_AUTHENTBODY_num(st) SKM_sk_num(KRB5_AUTHENTBODY, (st))
@@ -854,7 +1019,7 @@ STACK_OF(type) \
 #define sk_KRB5_AUTHENTBODY_sort(st) SKM_sk_sort(KRB5_AUTHENTBODY, (st))
 #define sk_KRB5_AUTHENTBODY_is_sorted(st) SKM_sk_is_sorted(KRB5_AUTHENTBODY, (st))
 
-#define sk_KRB5_CHECKSUM_new(st) SKM_sk_new(KRB5_CHECKSUM, (st))
+#define sk_KRB5_CHECKSUM_new(cmp) SKM_sk_new(KRB5_CHECKSUM, (cmp))
 #define sk_KRB5_CHECKSUM_new_null() SKM_sk_new_null(KRB5_CHECKSUM)
 #define sk_KRB5_CHECKSUM_free(st) SKM_sk_free(KRB5_CHECKSUM, (st))
 #define sk_KRB5_CHECKSUM_num(st) SKM_sk_num(KRB5_CHECKSUM, (st))
@@ -876,7 +1041,7 @@ STACK_OF(type) \
 #define sk_KRB5_CHECKSUM_sort(st) SKM_sk_sort(KRB5_CHECKSUM, (st))
 #define sk_KRB5_CHECKSUM_is_sorted(st) SKM_sk_is_sorted(KRB5_CHECKSUM, (st))
 
-#define sk_KRB5_ENCDATA_new(st) SKM_sk_new(KRB5_ENCDATA, (st))
+#define sk_KRB5_ENCDATA_new(cmp) SKM_sk_new(KRB5_ENCDATA, (cmp))
 #define sk_KRB5_ENCDATA_new_null() SKM_sk_new_null(KRB5_ENCDATA)
 #define sk_KRB5_ENCDATA_free(st) SKM_sk_free(KRB5_ENCDATA, (st))
 #define sk_KRB5_ENCDATA_num(st) SKM_sk_num(KRB5_ENCDATA, (st))
@@ -898,7 +1063,7 @@ STACK_OF(type) \
 #define sk_KRB5_ENCDATA_sort(st) SKM_sk_sort(KRB5_ENCDATA, (st))
 #define sk_KRB5_ENCDATA_is_sorted(st) SKM_sk_is_sorted(KRB5_ENCDATA, (st))
 
-#define sk_KRB5_ENCKEY_new(st) SKM_sk_new(KRB5_ENCKEY, (st))
+#define sk_KRB5_ENCKEY_new(cmp) SKM_sk_new(KRB5_ENCKEY, (cmp))
 #define sk_KRB5_ENCKEY_new_null() SKM_sk_new_null(KRB5_ENCKEY)
 #define sk_KRB5_ENCKEY_free(st) SKM_sk_free(KRB5_ENCKEY, (st))
 #define sk_KRB5_ENCKEY_num(st) SKM_sk_num(KRB5_ENCKEY, (st))
@@ -920,7 +1085,7 @@ STACK_OF(type) \
 #define sk_KRB5_ENCKEY_sort(st) SKM_sk_sort(KRB5_ENCKEY, (st))
 #define sk_KRB5_ENCKEY_is_sorted(st) SKM_sk_is_sorted(KRB5_ENCKEY, (st))
 
-#define sk_KRB5_PRINCNAME_new(st) SKM_sk_new(KRB5_PRINCNAME, (st))
+#define sk_KRB5_PRINCNAME_new(cmp) SKM_sk_new(KRB5_PRINCNAME, (cmp))
 #define sk_KRB5_PRINCNAME_new_null() SKM_sk_new_null(KRB5_PRINCNAME)
 #define sk_KRB5_PRINCNAME_free(st) SKM_sk_free(KRB5_PRINCNAME, (st))
 #define sk_KRB5_PRINCNAME_num(st) SKM_sk_num(KRB5_PRINCNAME, (st))
@@ -942,7 +1107,7 @@ STACK_OF(type) \
 #define sk_KRB5_PRINCNAME_sort(st) SKM_sk_sort(KRB5_PRINCNAME, (st))
 #define sk_KRB5_PRINCNAME_is_sorted(st) SKM_sk_is_sorted(KRB5_PRINCNAME, (st))
 
-#define sk_KRB5_TKTBODY_new(st) SKM_sk_new(KRB5_TKTBODY, (st))
+#define sk_KRB5_TKTBODY_new(cmp) SKM_sk_new(KRB5_TKTBODY, (cmp))
 #define sk_KRB5_TKTBODY_new_null() SKM_sk_new_null(KRB5_TKTBODY)
 #define sk_KRB5_TKTBODY_free(st) SKM_sk_free(KRB5_TKTBODY, (st))
 #define sk_KRB5_TKTBODY_num(st) SKM_sk_num(KRB5_TKTBODY, (st))
@@ -964,7 +1129,29 @@ STACK_OF(type) \
 #define sk_KRB5_TKTBODY_sort(st) SKM_sk_sort(KRB5_TKTBODY, (st))
 #define sk_KRB5_TKTBODY_is_sorted(st) SKM_sk_is_sorted(KRB5_TKTBODY, (st))
 
-#define sk_MIME_HEADER_new(st) SKM_sk_new(MIME_HEADER, (st))
+#define sk_MEM_OBJECT_DATA_new(cmp) SKM_sk_new(MEM_OBJECT_DATA, (cmp))
+#define sk_MEM_OBJECT_DATA_new_null() SKM_sk_new_null(MEM_OBJECT_DATA)
+#define sk_MEM_OBJECT_DATA_free(st) SKM_sk_free(MEM_OBJECT_DATA, (st))
+#define sk_MEM_OBJECT_DATA_num(st) SKM_sk_num(MEM_OBJECT_DATA, (st))
+#define sk_MEM_OBJECT_DATA_value(st, i) SKM_sk_value(MEM_OBJECT_DATA, (st), (i))
+#define sk_MEM_OBJECT_DATA_set(st, i, val) SKM_sk_set(MEM_OBJECT_DATA, (st), (i), (val))
+#define sk_MEM_OBJECT_DATA_zero(st) SKM_sk_zero(MEM_OBJECT_DATA, (st))
+#define sk_MEM_OBJECT_DATA_push(st, val) SKM_sk_push(MEM_OBJECT_DATA, (st), (val))
+#define sk_MEM_OBJECT_DATA_unshift(st, val) SKM_sk_unshift(MEM_OBJECT_DATA, (st), (val))
+#define sk_MEM_OBJECT_DATA_find(st, val) SKM_sk_find(MEM_OBJECT_DATA, (st), (val))
+#define sk_MEM_OBJECT_DATA_find_ex(st, val) SKM_sk_find_ex(MEM_OBJECT_DATA, (st), (val))
+#define sk_MEM_OBJECT_DATA_delete(st, i) SKM_sk_delete(MEM_OBJECT_DATA, (st), (i))
+#define sk_MEM_OBJECT_DATA_delete_ptr(st, ptr) SKM_sk_delete_ptr(MEM_OBJECT_DATA, (st), (ptr))
+#define sk_MEM_OBJECT_DATA_insert(st, val, i) SKM_sk_insert(MEM_OBJECT_DATA, (st), (val), (i))
+#define sk_MEM_OBJECT_DATA_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(MEM_OBJECT_DATA, (st), (cmp))
+#define sk_MEM_OBJECT_DATA_dup(st) SKM_sk_dup(MEM_OBJECT_DATA, st)
+#define sk_MEM_OBJECT_DATA_pop_free(st, free_func) SKM_sk_pop_free(MEM_OBJECT_DATA, (st), (free_func))
+#define sk_MEM_OBJECT_DATA_shift(st) SKM_sk_shift(MEM_OBJECT_DATA, (st))
+#define sk_MEM_OBJECT_DATA_pop(st) SKM_sk_pop(MEM_OBJECT_DATA, (st))
+#define sk_MEM_OBJECT_DATA_sort(st) SKM_sk_sort(MEM_OBJECT_DATA, (st))
+#define sk_MEM_OBJECT_DATA_is_sorted(st) SKM_sk_is_sorted(MEM_OBJECT_DATA, (st))
+
+#define sk_MIME_HEADER_new(cmp) SKM_sk_new(MIME_HEADER, (cmp))
 #define sk_MIME_HEADER_new_null() SKM_sk_new_null(MIME_HEADER)
 #define sk_MIME_HEADER_free(st) SKM_sk_free(MIME_HEADER, (st))
 #define sk_MIME_HEADER_num(st) SKM_sk_num(MIME_HEADER, (st))
@@ -986,7 +1173,7 @@ STACK_OF(type) \
 #define sk_MIME_HEADER_sort(st) SKM_sk_sort(MIME_HEADER, (st))
 #define sk_MIME_HEADER_is_sorted(st) SKM_sk_is_sorted(MIME_HEADER, (st))
 
-#define sk_MIME_PARAM_new(st) SKM_sk_new(MIME_PARAM, (st))
+#define sk_MIME_PARAM_new(cmp) SKM_sk_new(MIME_PARAM, (cmp))
 #define sk_MIME_PARAM_new_null() SKM_sk_new_null(MIME_PARAM)
 #define sk_MIME_PARAM_free(st) SKM_sk_free(MIME_PARAM, (st))
 #define sk_MIME_PARAM_num(st) SKM_sk_num(MIME_PARAM, (st))
@@ -1008,7 +1195,7 @@ STACK_OF(type) \
 #define sk_MIME_PARAM_sort(st) SKM_sk_sort(MIME_PARAM, (st))
 #define sk_MIME_PARAM_is_sorted(st) SKM_sk_is_sorted(MIME_PARAM, (st))
 
-#define sk_NAME_FUNCS_new(st) SKM_sk_new(NAME_FUNCS, (st))
+#define sk_NAME_FUNCS_new(cmp) SKM_sk_new(NAME_FUNCS, (cmp))
 #define sk_NAME_FUNCS_new_null() SKM_sk_new_null(NAME_FUNCS)
 #define sk_NAME_FUNCS_free(st) SKM_sk_free(NAME_FUNCS, (st))
 #define sk_NAME_FUNCS_num(st) SKM_sk_num(NAME_FUNCS, (st))
@@ -1030,7 +1217,7 @@ STACK_OF(type) \
 #define sk_NAME_FUNCS_sort(st) SKM_sk_sort(NAME_FUNCS, (st))
 #define sk_NAME_FUNCS_is_sorted(st) SKM_sk_is_sorted(NAME_FUNCS, (st))
 
-#define sk_OCSP_CERTID_new(st) SKM_sk_new(OCSP_CERTID, (st))
+#define sk_OCSP_CERTID_new(cmp) SKM_sk_new(OCSP_CERTID, (cmp))
 #define sk_OCSP_CERTID_new_null() SKM_sk_new_null(OCSP_CERTID)
 #define sk_OCSP_CERTID_free(st) SKM_sk_free(OCSP_CERTID, (st))
 #define sk_OCSP_CERTID_num(st) SKM_sk_num(OCSP_CERTID, (st))
@@ -1052,7 +1239,7 @@ STACK_OF(type) \
 #define sk_OCSP_CERTID_sort(st) SKM_sk_sort(OCSP_CERTID, (st))
 #define sk_OCSP_CERTID_is_sorted(st) SKM_sk_is_sorted(OCSP_CERTID, (st))
 
-#define sk_OCSP_ONEREQ_new(st) SKM_sk_new(OCSP_ONEREQ, (st))
+#define sk_OCSP_ONEREQ_new(cmp) SKM_sk_new(OCSP_ONEREQ, (cmp))
 #define sk_OCSP_ONEREQ_new_null() SKM_sk_new_null(OCSP_ONEREQ)
 #define sk_OCSP_ONEREQ_free(st) SKM_sk_free(OCSP_ONEREQ, (st))
 #define sk_OCSP_ONEREQ_num(st) SKM_sk_num(OCSP_ONEREQ, (st))
@@ -1074,7 +1261,7 @@ STACK_OF(type) \
 #define sk_OCSP_ONEREQ_sort(st) SKM_sk_sort(OCSP_ONEREQ, (st))
 #define sk_OCSP_ONEREQ_is_sorted(st) SKM_sk_is_sorted(OCSP_ONEREQ, (st))
 
-#define sk_OCSP_RESPID_new(st) SKM_sk_new(OCSP_RESPID, (st))
+#define sk_OCSP_RESPID_new(cmp) SKM_sk_new(OCSP_RESPID, (cmp))
 #define sk_OCSP_RESPID_new_null() SKM_sk_new_null(OCSP_RESPID)
 #define sk_OCSP_RESPID_free(st) SKM_sk_free(OCSP_RESPID, (st))
 #define sk_OCSP_RESPID_num(st) SKM_sk_num(OCSP_RESPID, (st))
@@ -1096,7 +1283,7 @@ STACK_OF(type) \
 #define sk_OCSP_RESPID_sort(st) SKM_sk_sort(OCSP_RESPID, (st))
 #define sk_OCSP_RESPID_is_sorted(st) SKM_sk_is_sorted(OCSP_RESPID, (st))
 
-#define sk_OCSP_SINGLERESP_new(st) SKM_sk_new(OCSP_SINGLERESP, (st))
+#define sk_OCSP_SINGLERESP_new(cmp) SKM_sk_new(OCSP_SINGLERESP, (cmp))
 #define sk_OCSP_SINGLERESP_new_null() SKM_sk_new_null(OCSP_SINGLERESP)
 #define sk_OCSP_SINGLERESP_free(st) SKM_sk_free(OCSP_SINGLERESP, (st))
 #define sk_OCSP_SINGLERESP_num(st) SKM_sk_num(OCSP_SINGLERESP, (st))
@@ -1118,7 +1305,7 @@ STACK_OF(type) \
 #define sk_OCSP_SINGLERESP_sort(st) SKM_sk_sort(OCSP_SINGLERESP, (st))
 #define sk_OCSP_SINGLERESP_is_sorted(st) SKM_sk_is_sorted(OCSP_SINGLERESP, (st))
 
-#define sk_PKCS12_SAFEBAG_new(st) SKM_sk_new(PKCS12_SAFEBAG, (st))
+#define sk_PKCS12_SAFEBAG_new(cmp) SKM_sk_new(PKCS12_SAFEBAG, (cmp))
 #define sk_PKCS12_SAFEBAG_new_null() SKM_sk_new_null(PKCS12_SAFEBAG)
 #define sk_PKCS12_SAFEBAG_free(st) SKM_sk_free(PKCS12_SAFEBAG, (st))
 #define sk_PKCS12_SAFEBAG_num(st) SKM_sk_num(PKCS12_SAFEBAG, (st))
@@ -1140,7 +1327,7 @@ STACK_OF(type) \
 #define sk_PKCS12_SAFEBAG_sort(st) SKM_sk_sort(PKCS12_SAFEBAG, (st))
 #define sk_PKCS12_SAFEBAG_is_sorted(st) SKM_sk_is_sorted(PKCS12_SAFEBAG, (st))
 
-#define sk_PKCS7_new(st) SKM_sk_new(PKCS7, (st))
+#define sk_PKCS7_new(cmp) SKM_sk_new(PKCS7, (cmp))
 #define sk_PKCS7_new_null() SKM_sk_new_null(PKCS7)
 #define sk_PKCS7_free(st) SKM_sk_free(PKCS7, (st))
 #define sk_PKCS7_num(st) SKM_sk_num(PKCS7, (st))
@@ -1162,7 +1349,7 @@ STACK_OF(type) \
 #define sk_PKCS7_sort(st) SKM_sk_sort(PKCS7, (st))
 #define sk_PKCS7_is_sorted(st) SKM_sk_is_sorted(PKCS7, (st))
 
-#define sk_PKCS7_RECIP_INFO_new(st) SKM_sk_new(PKCS7_RECIP_INFO, (st))
+#define sk_PKCS7_RECIP_INFO_new(cmp) SKM_sk_new(PKCS7_RECIP_INFO, (cmp))
 #define sk_PKCS7_RECIP_INFO_new_null() SKM_sk_new_null(PKCS7_RECIP_INFO)
 #define sk_PKCS7_RECIP_INFO_free(st) SKM_sk_free(PKCS7_RECIP_INFO, (st))
 #define sk_PKCS7_RECIP_INFO_num(st) SKM_sk_num(PKCS7_RECIP_INFO, (st))
@@ -1184,7 +1371,7 @@ STACK_OF(type) \
 #define sk_PKCS7_RECIP_INFO_sort(st) SKM_sk_sort(PKCS7_RECIP_INFO, (st))
 #define sk_PKCS7_RECIP_INFO_is_sorted(st) SKM_sk_is_sorted(PKCS7_RECIP_INFO, (st))
 
-#define sk_PKCS7_SIGNER_INFO_new(st) SKM_sk_new(PKCS7_SIGNER_INFO, (st))
+#define sk_PKCS7_SIGNER_INFO_new(cmp) SKM_sk_new(PKCS7_SIGNER_INFO, (cmp))
 #define sk_PKCS7_SIGNER_INFO_new_null() SKM_sk_new_null(PKCS7_SIGNER_INFO)
 #define sk_PKCS7_SIGNER_INFO_free(st) SKM_sk_free(PKCS7_SIGNER_INFO, (st))
 #define sk_PKCS7_SIGNER_INFO_num(st) SKM_sk_num(PKCS7_SIGNER_INFO, (st))
@@ -1206,7 +1393,7 @@ STACK_OF(type) \
 #define sk_PKCS7_SIGNER_INFO_sort(st) SKM_sk_sort(PKCS7_SIGNER_INFO, (st))
 #define sk_PKCS7_SIGNER_INFO_is_sorted(st) SKM_sk_is_sorted(PKCS7_SIGNER_INFO, (st))
 
-#define sk_POLICYINFO_new(st) SKM_sk_new(POLICYINFO, (st))
+#define sk_POLICYINFO_new(cmp) SKM_sk_new(POLICYINFO, (cmp))
 #define sk_POLICYINFO_new_null() SKM_sk_new_null(POLICYINFO)
 #define sk_POLICYINFO_free(st) SKM_sk_free(POLICYINFO, (st))
 #define sk_POLICYINFO_num(st) SKM_sk_num(POLICYINFO, (st))
@@ -1228,7 +1415,7 @@ STACK_OF(type) \
 #define sk_POLICYINFO_sort(st) SKM_sk_sort(POLICYINFO, (st))
 #define sk_POLICYINFO_is_sorted(st) SKM_sk_is_sorted(POLICYINFO, (st))
 
-#define sk_POLICYQUALINFO_new(st) SKM_sk_new(POLICYQUALINFO, (st))
+#define sk_POLICYQUALINFO_new(cmp) SKM_sk_new(POLICYQUALINFO, (cmp))
 #define sk_POLICYQUALINFO_new_null() SKM_sk_new_null(POLICYQUALINFO)
 #define sk_POLICYQUALINFO_free(st) SKM_sk_free(POLICYQUALINFO, (st))
 #define sk_POLICYQUALINFO_num(st) SKM_sk_num(POLICYQUALINFO, (st))
@@ -1250,7 +1437,7 @@ STACK_OF(type) \
 #define sk_POLICYQUALINFO_sort(st) SKM_sk_sort(POLICYQUALINFO, (st))
 #define sk_POLICYQUALINFO_is_sorted(st) SKM_sk_is_sorted(POLICYQUALINFO, (st))
 
-#define sk_POLICY_MAPPING_new(st) SKM_sk_new(POLICY_MAPPING, (st))
+#define sk_POLICY_MAPPING_new(cmp) SKM_sk_new(POLICY_MAPPING, (cmp))
 #define sk_POLICY_MAPPING_new_null() SKM_sk_new_null(POLICY_MAPPING)
 #define sk_POLICY_MAPPING_free(st) SKM_sk_free(POLICY_MAPPING, (st))
 #define sk_POLICY_MAPPING_num(st) SKM_sk_num(POLICY_MAPPING, (st))
@@ -1272,7 +1459,95 @@ STACK_OF(type) \
 #define sk_POLICY_MAPPING_sort(st) SKM_sk_sort(POLICY_MAPPING, (st))
 #define sk_POLICY_MAPPING_is_sorted(st) SKM_sk_is_sorted(POLICY_MAPPING, (st))
 
-#define sk_SSL_CIPHER_new(st) SKM_sk_new(SSL_CIPHER, (st))
+#define sk_SRP_gN_new(cmp) SKM_sk_new(SRP_gN, (cmp))
+#define sk_SRP_gN_new_null() SKM_sk_new_null(SRP_gN)
+#define sk_SRP_gN_free(st) SKM_sk_free(SRP_gN, (st))
+#define sk_SRP_gN_num(st) SKM_sk_num(SRP_gN, (st))
+#define sk_SRP_gN_value(st, i) SKM_sk_value(SRP_gN, (st), (i))
+#define sk_SRP_gN_set(st, i, val) SKM_sk_set(SRP_gN, (st), (i), (val))
+#define sk_SRP_gN_zero(st) SKM_sk_zero(SRP_gN, (st))
+#define sk_SRP_gN_push(st, val) SKM_sk_push(SRP_gN, (st), (val))
+#define sk_SRP_gN_unshift(st, val) SKM_sk_unshift(SRP_gN, (st), (val))
+#define sk_SRP_gN_find(st, val) SKM_sk_find(SRP_gN, (st), (val))
+#define sk_SRP_gN_find_ex(st, val) SKM_sk_find_ex(SRP_gN, (st), (val))
+#define sk_SRP_gN_delete(st, i) SKM_sk_delete(SRP_gN, (st), (i))
+#define sk_SRP_gN_delete_ptr(st, ptr) SKM_sk_delete_ptr(SRP_gN, (st), (ptr))
+#define sk_SRP_gN_insert(st, val, i) SKM_sk_insert(SRP_gN, (st), (val), (i))
+#define sk_SRP_gN_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(SRP_gN, (st), (cmp))
+#define sk_SRP_gN_dup(st) SKM_sk_dup(SRP_gN, st)
+#define sk_SRP_gN_pop_free(st, free_func) SKM_sk_pop_free(SRP_gN, (st), (free_func))
+#define sk_SRP_gN_shift(st) SKM_sk_shift(SRP_gN, (st))
+#define sk_SRP_gN_pop(st) SKM_sk_pop(SRP_gN, (st))
+#define sk_SRP_gN_sort(st) SKM_sk_sort(SRP_gN, (st))
+#define sk_SRP_gN_is_sorted(st) SKM_sk_is_sorted(SRP_gN, (st))
+
+#define sk_SRP_gN_cache_new(cmp) SKM_sk_new(SRP_gN_cache, (cmp))
+#define sk_SRP_gN_cache_new_null() SKM_sk_new_null(SRP_gN_cache)
+#define sk_SRP_gN_cache_free(st) SKM_sk_free(SRP_gN_cache, (st))
+#define sk_SRP_gN_cache_num(st) SKM_sk_num(SRP_gN_cache, (st))
+#define sk_SRP_gN_cache_value(st, i) SKM_sk_value(SRP_gN_cache, (st), (i))
+#define sk_SRP_gN_cache_set(st, i, val) SKM_sk_set(SRP_gN_cache, (st), (i), (val))
+#define sk_SRP_gN_cache_zero(st) SKM_sk_zero(SRP_gN_cache, (st))
+#define sk_SRP_gN_cache_push(st, val) SKM_sk_push(SRP_gN_cache, (st), (val))
+#define sk_SRP_gN_cache_unshift(st, val) SKM_sk_unshift(SRP_gN_cache, (st), (val))
+#define sk_SRP_gN_cache_find(st, val) SKM_sk_find(SRP_gN_cache, (st), (val))
+#define sk_SRP_gN_cache_find_ex(st, val) SKM_sk_find_ex(SRP_gN_cache, (st), (val))
+#define sk_SRP_gN_cache_delete(st, i) SKM_sk_delete(SRP_gN_cache, (st), (i))
+#define sk_SRP_gN_cache_delete_ptr(st, ptr) SKM_sk_delete_ptr(SRP_gN_cache, (st), (ptr))
+#define sk_SRP_gN_cache_insert(st, val, i) SKM_sk_insert(SRP_gN_cache, (st), (val), (i))
+#define sk_SRP_gN_cache_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(SRP_gN_cache, (st), (cmp))
+#define sk_SRP_gN_cache_dup(st) SKM_sk_dup(SRP_gN_cache, st)
+#define sk_SRP_gN_cache_pop_free(st, free_func) SKM_sk_pop_free(SRP_gN_cache, (st), (free_func))
+#define sk_SRP_gN_cache_shift(st) SKM_sk_shift(SRP_gN_cache, (st))
+#define sk_SRP_gN_cache_pop(st) SKM_sk_pop(SRP_gN_cache, (st))
+#define sk_SRP_gN_cache_sort(st) SKM_sk_sort(SRP_gN_cache, (st))
+#define sk_SRP_gN_cache_is_sorted(st) SKM_sk_is_sorted(SRP_gN_cache, (st))
+
+#define sk_SRP_user_pwd_new(cmp) SKM_sk_new(SRP_user_pwd, (cmp))
+#define sk_SRP_user_pwd_new_null() SKM_sk_new_null(SRP_user_pwd)
+#define sk_SRP_user_pwd_free(st) SKM_sk_free(SRP_user_pwd, (st))
+#define sk_SRP_user_pwd_num(st) SKM_sk_num(SRP_user_pwd, (st))
+#define sk_SRP_user_pwd_value(st, i) SKM_sk_value(SRP_user_pwd, (st), (i))
+#define sk_SRP_user_pwd_set(st, i, val) SKM_sk_set(SRP_user_pwd, (st), (i), (val))
+#define sk_SRP_user_pwd_zero(st) SKM_sk_zero(SRP_user_pwd, (st))
+#define sk_SRP_user_pwd_push(st, val) SKM_sk_push(SRP_user_pwd, (st), (val))
+#define sk_SRP_user_pwd_unshift(st, val) SKM_sk_unshift(SRP_user_pwd, (st), (val))
+#define sk_SRP_user_pwd_find(st, val) SKM_sk_find(SRP_user_pwd, (st), (val))
+#define sk_SRP_user_pwd_find_ex(st, val) SKM_sk_find_ex(SRP_user_pwd, (st), (val))
+#define sk_SRP_user_pwd_delete(st, i) SKM_sk_delete(SRP_user_pwd, (st), (i))
+#define sk_SRP_user_pwd_delete_ptr(st, ptr) SKM_sk_delete_ptr(SRP_user_pwd, (st), (ptr))
+#define sk_SRP_user_pwd_insert(st, val, i) SKM_sk_insert(SRP_user_pwd, (st), (val), (i))
+#define sk_SRP_user_pwd_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(SRP_user_pwd, (st), (cmp))
+#define sk_SRP_user_pwd_dup(st) SKM_sk_dup(SRP_user_pwd, st)
+#define sk_SRP_user_pwd_pop_free(st, free_func) SKM_sk_pop_free(SRP_user_pwd, (st), (free_func))
+#define sk_SRP_user_pwd_shift(st) SKM_sk_shift(SRP_user_pwd, (st))
+#define sk_SRP_user_pwd_pop(st) SKM_sk_pop(SRP_user_pwd, (st))
+#define sk_SRP_user_pwd_sort(st) SKM_sk_sort(SRP_user_pwd, (st))
+#define sk_SRP_user_pwd_is_sorted(st) SKM_sk_is_sorted(SRP_user_pwd, (st))
+
+#define sk_SRTP_PROTECTION_PROFILE_new(cmp) SKM_sk_new(SRTP_PROTECTION_PROFILE, (cmp))
+#define sk_SRTP_PROTECTION_PROFILE_new_null() SKM_sk_new_null(SRTP_PROTECTION_PROFILE)
+#define sk_SRTP_PROTECTION_PROFILE_free(st) SKM_sk_free(SRTP_PROTECTION_PROFILE, (st))
+#define sk_SRTP_PROTECTION_PROFILE_num(st) SKM_sk_num(SRTP_PROTECTION_PROFILE, (st))
+#define sk_SRTP_PROTECTION_PROFILE_value(st, i) SKM_sk_value(SRTP_PROTECTION_PROFILE, (st), (i))
+#define sk_SRTP_PROTECTION_PROFILE_set(st, i, val) SKM_sk_set(SRTP_PROTECTION_PROFILE, (st), (i), (val))
+#define sk_SRTP_PROTECTION_PROFILE_zero(st) SKM_sk_zero(SRTP_PROTECTION_PROFILE, (st))
+#define sk_SRTP_PROTECTION_PROFILE_push(st, val) SKM_sk_push(SRTP_PROTECTION_PROFILE, (st), (val))
+#define sk_SRTP_PROTECTION_PROFILE_unshift(st, val) SKM_sk_unshift(SRTP_PROTECTION_PROFILE, (st), (val))
+#define sk_SRTP_PROTECTION_PROFILE_find(st, val) SKM_sk_find(SRTP_PROTECTION_PROFILE, (st), (val))
+#define sk_SRTP_PROTECTION_PROFILE_find_ex(st, val) SKM_sk_find_ex(SRTP_PROTECTION_PROFILE, (st), (val))
+#define sk_SRTP_PROTECTION_PROFILE_delete(st, i) SKM_sk_delete(SRTP_PROTECTION_PROFILE, (st), (i))
+#define sk_SRTP_PROTECTION_PROFILE_delete_ptr(st, ptr) SKM_sk_delete_ptr(SRTP_PROTECTION_PROFILE, (st), (ptr))
+#define sk_SRTP_PROTECTION_PROFILE_insert(st, val, i) SKM_sk_insert(SRTP_PROTECTION_PROFILE, (st), (val), (i))
+#define sk_SRTP_PROTECTION_PROFILE_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(SRTP_PROTECTION_PROFILE, (st), (cmp))
+#define sk_SRTP_PROTECTION_PROFILE_dup(st) SKM_sk_dup(SRTP_PROTECTION_PROFILE, st)
+#define sk_SRTP_PROTECTION_PROFILE_pop_free(st, free_func) SKM_sk_pop_free(SRTP_PROTECTION_PROFILE, (st), (free_func))
+#define sk_SRTP_PROTECTION_PROFILE_shift(st) SKM_sk_shift(SRTP_PROTECTION_PROFILE, (st))
+#define sk_SRTP_PROTECTION_PROFILE_pop(st) SKM_sk_pop(SRTP_PROTECTION_PROFILE, (st))
+#define sk_SRTP_PROTECTION_PROFILE_sort(st) SKM_sk_sort(SRTP_PROTECTION_PROFILE, (st))
+#define sk_SRTP_PROTECTION_PROFILE_is_sorted(st) SKM_sk_is_sorted(SRTP_PROTECTION_PROFILE, (st))
+
+#define sk_SSL_CIPHER_new(cmp) SKM_sk_new(SSL_CIPHER, (cmp))
 #define sk_SSL_CIPHER_new_null() SKM_sk_new_null(SSL_CIPHER)
 #define sk_SSL_CIPHER_free(st) SKM_sk_free(SSL_CIPHER, (st))
 #define sk_SSL_CIPHER_num(st) SKM_sk_num(SSL_CIPHER, (st))
@@ -1294,7 +1569,7 @@ STACK_OF(type) \
 #define sk_SSL_CIPHER_sort(st) SKM_sk_sort(SSL_CIPHER, (st))
 #define sk_SSL_CIPHER_is_sorted(st) SKM_sk_is_sorted(SSL_CIPHER, (st))
 
-#define sk_SSL_COMP_new(st) SKM_sk_new(SSL_COMP, (st))
+#define sk_SSL_COMP_new(cmp) SKM_sk_new(SSL_COMP, (cmp))
 #define sk_SSL_COMP_new_null() SKM_sk_new_null(SSL_COMP)
 #define sk_SSL_COMP_free(st) SKM_sk_free(SSL_COMP, (st))
 #define sk_SSL_COMP_num(st) SKM_sk_num(SSL_COMP, (st))
@@ -1316,7 +1591,51 @@ STACK_OF(type) \
 #define sk_SSL_COMP_sort(st) SKM_sk_sort(SSL_COMP, (st))
 #define sk_SSL_COMP_is_sorted(st) SKM_sk_is_sorted(SSL_COMP, (st))
 
-#define sk_STORE_OBJECT_new(st) SKM_sk_new(STORE_OBJECT, (st))
+#define sk_STACK_OF_X509_NAME_ENTRY_new(cmp) SKM_sk_new(STACK_OF_X509_NAME_ENTRY, (cmp))
+#define sk_STACK_OF_X509_NAME_ENTRY_new_null() SKM_sk_new_null(STACK_OF_X509_NAME_ENTRY)
+#define sk_STACK_OF_X509_NAME_ENTRY_free(st) SKM_sk_free(STACK_OF_X509_NAME_ENTRY, (st))
+#define sk_STACK_OF_X509_NAME_ENTRY_num(st) SKM_sk_num(STACK_OF_X509_NAME_ENTRY, (st))
+#define sk_STACK_OF_X509_NAME_ENTRY_value(st, i) SKM_sk_value(STACK_OF_X509_NAME_ENTRY, (st), (i))
+#define sk_STACK_OF_X509_NAME_ENTRY_set(st, i, val) SKM_sk_set(STACK_OF_X509_NAME_ENTRY, (st), (i), (val))
+#define sk_STACK_OF_X509_NAME_ENTRY_zero(st) SKM_sk_zero(STACK_OF_X509_NAME_ENTRY, (st))
+#define sk_STACK_OF_X509_NAME_ENTRY_push(st, val) SKM_sk_push(STACK_OF_X509_NAME_ENTRY, (st), (val))
+#define sk_STACK_OF_X509_NAME_ENTRY_unshift(st, val) SKM_sk_unshift(STACK_OF_X509_NAME_ENTRY, (st), (val))
+#define sk_STACK_OF_X509_NAME_ENTRY_find(st, val) SKM_sk_find(STACK_OF_X509_NAME_ENTRY, (st), (val))
+#define sk_STACK_OF_X509_NAME_ENTRY_find_ex(st, val) SKM_sk_find_ex(STACK_OF_X509_NAME_ENTRY, (st), (val))
+#define sk_STACK_OF_X509_NAME_ENTRY_delete(st, i) SKM_sk_delete(STACK_OF_X509_NAME_ENTRY, (st), (i))
+#define sk_STACK_OF_X509_NAME_ENTRY_delete_ptr(st, ptr) SKM_sk_delete_ptr(STACK_OF_X509_NAME_ENTRY, (st), (ptr))
+#define sk_STACK_OF_X509_NAME_ENTRY_insert(st, val, i) SKM_sk_insert(STACK_OF_X509_NAME_ENTRY, (st), (val), (i))
+#define sk_STACK_OF_X509_NAME_ENTRY_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(STACK_OF_X509_NAME_ENTRY, (st), (cmp))
+#define sk_STACK_OF_X509_NAME_ENTRY_dup(st) SKM_sk_dup(STACK_OF_X509_NAME_ENTRY, st)
+#define sk_STACK_OF_X509_NAME_ENTRY_pop_free(st, free_func) SKM_sk_pop_free(STACK_OF_X509_NAME_ENTRY, (st), (free_func))
+#define sk_STACK_OF_X509_NAME_ENTRY_shift(st) SKM_sk_shift(STACK_OF_X509_NAME_ENTRY, (st))
+#define sk_STACK_OF_X509_NAME_ENTRY_pop(st) SKM_sk_pop(STACK_OF_X509_NAME_ENTRY, (st))
+#define sk_STACK_OF_X509_NAME_ENTRY_sort(st) SKM_sk_sort(STACK_OF_X509_NAME_ENTRY, (st))
+#define sk_STACK_OF_X509_NAME_ENTRY_is_sorted(st) SKM_sk_is_sorted(STACK_OF_X509_NAME_ENTRY, (st))
+
+#define sk_STORE_ATTR_INFO_new(cmp) SKM_sk_new(STORE_ATTR_INFO, (cmp))
+#define sk_STORE_ATTR_INFO_new_null() SKM_sk_new_null(STORE_ATTR_INFO)
+#define sk_STORE_ATTR_INFO_free(st) SKM_sk_free(STORE_ATTR_INFO, (st))
+#define sk_STORE_ATTR_INFO_num(st) SKM_sk_num(STORE_ATTR_INFO, (st))
+#define sk_STORE_ATTR_INFO_value(st, i) SKM_sk_value(STORE_ATTR_INFO, (st), (i))
+#define sk_STORE_ATTR_INFO_set(st, i, val) SKM_sk_set(STORE_ATTR_INFO, (st), (i), (val))
+#define sk_STORE_ATTR_INFO_zero(st) SKM_sk_zero(STORE_ATTR_INFO, (st))
+#define sk_STORE_ATTR_INFO_push(st, val) SKM_sk_push(STORE_ATTR_INFO, (st), (val))
+#define sk_STORE_ATTR_INFO_unshift(st, val) SKM_sk_unshift(STORE_ATTR_INFO, (st), (val))
+#define sk_STORE_ATTR_INFO_find(st, val) SKM_sk_find(STORE_ATTR_INFO, (st), (val))
+#define sk_STORE_ATTR_INFO_find_ex(st, val) SKM_sk_find_ex(STORE_ATTR_INFO, (st), (val))
+#define sk_STORE_ATTR_INFO_delete(st, i) SKM_sk_delete(STORE_ATTR_INFO, (st), (i))
+#define sk_STORE_ATTR_INFO_delete_ptr(st, ptr) SKM_sk_delete_ptr(STORE_ATTR_INFO, (st), (ptr))
+#define sk_STORE_ATTR_INFO_insert(st, val, i) SKM_sk_insert(STORE_ATTR_INFO, (st), (val), (i))
+#define sk_STORE_ATTR_INFO_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(STORE_ATTR_INFO, (st), (cmp))
+#define sk_STORE_ATTR_INFO_dup(st) SKM_sk_dup(STORE_ATTR_INFO, st)
+#define sk_STORE_ATTR_INFO_pop_free(st, free_func) SKM_sk_pop_free(STORE_ATTR_INFO, (st), (free_func))
+#define sk_STORE_ATTR_INFO_shift(st) SKM_sk_shift(STORE_ATTR_INFO, (st))
+#define sk_STORE_ATTR_INFO_pop(st) SKM_sk_pop(STORE_ATTR_INFO, (st))
+#define sk_STORE_ATTR_INFO_sort(st) SKM_sk_sort(STORE_ATTR_INFO, (st))
+#define sk_STORE_ATTR_INFO_is_sorted(st) SKM_sk_is_sorted(STORE_ATTR_INFO, (st))
+
+#define sk_STORE_OBJECT_new(cmp) SKM_sk_new(STORE_OBJECT, (cmp))
 #define sk_STORE_OBJECT_new_null() SKM_sk_new_null(STORE_OBJECT)
 #define sk_STORE_OBJECT_free(st) SKM_sk_free(STORE_OBJECT, (st))
 #define sk_STORE_OBJECT_num(st) SKM_sk_num(STORE_OBJECT, (st))
@@ -1338,7 +1657,7 @@ STACK_OF(type) \
 #define sk_STORE_OBJECT_sort(st) SKM_sk_sort(STORE_OBJECT, (st))
 #define sk_STORE_OBJECT_is_sorted(st) SKM_sk_is_sorted(STORE_OBJECT, (st))
 
-#define sk_SXNETID_new(st) SKM_sk_new(SXNETID, (st))
+#define sk_SXNETID_new(cmp) SKM_sk_new(SXNETID, (cmp))
 #define sk_SXNETID_new_null() SKM_sk_new_null(SXNETID)
 #define sk_SXNETID_free(st) SKM_sk_free(SXNETID, (st))
 #define sk_SXNETID_num(st) SKM_sk_num(SXNETID, (st))
@@ -1360,7 +1679,7 @@ STACK_OF(type) \
 #define sk_SXNETID_sort(st) SKM_sk_sort(SXNETID, (st))
 #define sk_SXNETID_is_sorted(st) SKM_sk_is_sorted(SXNETID, (st))
 
-#define sk_UI_STRING_new(st) SKM_sk_new(UI_STRING, (st))
+#define sk_UI_STRING_new(cmp) SKM_sk_new(UI_STRING, (cmp))
 #define sk_UI_STRING_new_null() SKM_sk_new_null(UI_STRING)
 #define sk_UI_STRING_free(st) SKM_sk_free(UI_STRING, (st))
 #define sk_UI_STRING_num(st) SKM_sk_num(UI_STRING, (st))
@@ -1382,7 +1701,7 @@ STACK_OF(type) \
 #define sk_UI_STRING_sort(st) SKM_sk_sort(UI_STRING, (st))
 #define sk_UI_STRING_is_sorted(st) SKM_sk_is_sorted(UI_STRING, (st))
 
-#define sk_X509_new(st) SKM_sk_new(X509, (st))
+#define sk_X509_new(cmp) SKM_sk_new(X509, (cmp))
 #define sk_X509_new_null() SKM_sk_new_null(X509)
 #define sk_X509_free(st) SKM_sk_free(X509, (st))
 #define sk_X509_num(st) SKM_sk_num(X509, (st))
@@ -1404,7 +1723,7 @@ STACK_OF(type) \
 #define sk_X509_sort(st) SKM_sk_sort(X509, (st))
 #define sk_X509_is_sorted(st) SKM_sk_is_sorted(X509, (st))
 
-#define sk_X509V3_EXT_METHOD_new(st) SKM_sk_new(X509V3_EXT_METHOD, (st))
+#define sk_X509V3_EXT_METHOD_new(cmp) SKM_sk_new(X509V3_EXT_METHOD, (cmp))
 #define sk_X509V3_EXT_METHOD_new_null() SKM_sk_new_null(X509V3_EXT_METHOD)
 #define sk_X509V3_EXT_METHOD_free(st) SKM_sk_free(X509V3_EXT_METHOD, (st))
 #define sk_X509V3_EXT_METHOD_num(st) SKM_sk_num(X509V3_EXT_METHOD, (st))
@@ -1426,7 +1745,7 @@ STACK_OF(type) \
 #define sk_X509V3_EXT_METHOD_sort(st) SKM_sk_sort(X509V3_EXT_METHOD, (st))
 #define sk_X509V3_EXT_METHOD_is_sorted(st) SKM_sk_is_sorted(X509V3_EXT_METHOD, (st))
 
-#define sk_X509_ALGOR_new(st) SKM_sk_new(X509_ALGOR, (st))
+#define sk_X509_ALGOR_new(cmp) SKM_sk_new(X509_ALGOR, (cmp))
 #define sk_X509_ALGOR_new_null() SKM_sk_new_null(X509_ALGOR)
 #define sk_X509_ALGOR_free(st) SKM_sk_free(X509_ALGOR, (st))
 #define sk_X509_ALGOR_num(st) SKM_sk_num(X509_ALGOR, (st))
@@ -1448,7 +1767,7 @@ STACK_OF(type) \
 #define sk_X509_ALGOR_sort(st) SKM_sk_sort(X509_ALGOR, (st))
 #define sk_X509_ALGOR_is_sorted(st) SKM_sk_is_sorted(X509_ALGOR, (st))
 
-#define sk_X509_ATTRIBUTE_new(st) SKM_sk_new(X509_ATTRIBUTE, (st))
+#define sk_X509_ATTRIBUTE_new(cmp) SKM_sk_new(X509_ATTRIBUTE, (cmp))
 #define sk_X509_ATTRIBUTE_new_null() SKM_sk_new_null(X509_ATTRIBUTE)
 #define sk_X509_ATTRIBUTE_free(st) SKM_sk_free(X509_ATTRIBUTE, (st))
 #define sk_X509_ATTRIBUTE_num(st) SKM_sk_num(X509_ATTRIBUTE, (st))
@@ -1470,7 +1789,7 @@ STACK_OF(type) \
 #define sk_X509_ATTRIBUTE_sort(st) SKM_sk_sort(X509_ATTRIBUTE, (st))
 #define sk_X509_ATTRIBUTE_is_sorted(st) SKM_sk_is_sorted(X509_ATTRIBUTE, (st))
 
-#define sk_X509_CRL_new(st) SKM_sk_new(X509_CRL, (st))
+#define sk_X509_CRL_new(cmp) SKM_sk_new(X509_CRL, (cmp))
 #define sk_X509_CRL_new_null() SKM_sk_new_null(X509_CRL)
 #define sk_X509_CRL_free(st) SKM_sk_free(X509_CRL, (st))
 #define sk_X509_CRL_num(st) SKM_sk_num(X509_CRL, (st))
@@ -1492,7 +1811,7 @@ STACK_OF(type) \
 #define sk_X509_CRL_sort(st) SKM_sk_sort(X509_CRL, (st))
 #define sk_X509_CRL_is_sorted(st) SKM_sk_is_sorted(X509_CRL, (st))
 
-#define sk_X509_EXTENSION_new(st) SKM_sk_new(X509_EXTENSION, (st))
+#define sk_X509_EXTENSION_new(cmp) SKM_sk_new(X509_EXTENSION, (cmp))
 #define sk_X509_EXTENSION_new_null() SKM_sk_new_null(X509_EXTENSION)
 #define sk_X509_EXTENSION_free(st) SKM_sk_free(X509_EXTENSION, (st))
 #define sk_X509_EXTENSION_num(st) SKM_sk_num(X509_EXTENSION, (st))
@@ -1514,7 +1833,7 @@ STACK_OF(type) \
 #define sk_X509_EXTENSION_sort(st) SKM_sk_sort(X509_EXTENSION, (st))
 #define sk_X509_EXTENSION_is_sorted(st) SKM_sk_is_sorted(X509_EXTENSION, (st))
 
-#define sk_X509_INFO_new(st) SKM_sk_new(X509_INFO, (st))
+#define sk_X509_INFO_new(cmp) SKM_sk_new(X509_INFO, (cmp))
 #define sk_X509_INFO_new_null() SKM_sk_new_null(X509_INFO)
 #define sk_X509_INFO_free(st) SKM_sk_free(X509_INFO, (st))
 #define sk_X509_INFO_num(st) SKM_sk_num(X509_INFO, (st))
@@ -1536,7 +1855,7 @@ STACK_OF(type) \
 #define sk_X509_INFO_sort(st) SKM_sk_sort(X509_INFO, (st))
 #define sk_X509_INFO_is_sorted(st) SKM_sk_is_sorted(X509_INFO, (st))
 
-#define sk_X509_LOOKUP_new(st) SKM_sk_new(X509_LOOKUP, (st))
+#define sk_X509_LOOKUP_new(cmp) SKM_sk_new(X509_LOOKUP, (cmp))
 #define sk_X509_LOOKUP_new_null() SKM_sk_new_null(X509_LOOKUP)
 #define sk_X509_LOOKUP_free(st) SKM_sk_free(X509_LOOKUP, (st))
 #define sk_X509_LOOKUP_num(st) SKM_sk_num(X509_LOOKUP, (st))
@@ -1558,7 +1877,7 @@ STACK_OF(type) \
 #define sk_X509_LOOKUP_sort(st) SKM_sk_sort(X509_LOOKUP, (st))
 #define sk_X509_LOOKUP_is_sorted(st) SKM_sk_is_sorted(X509_LOOKUP, (st))
 
-#define sk_X509_NAME_new(st) SKM_sk_new(X509_NAME, (st))
+#define sk_X509_NAME_new(cmp) SKM_sk_new(X509_NAME, (cmp))
 #define sk_X509_NAME_new_null() SKM_sk_new_null(X509_NAME)
 #define sk_X509_NAME_free(st) SKM_sk_free(X509_NAME, (st))
 #define sk_X509_NAME_num(st) SKM_sk_num(X509_NAME, (st))
@@ -1580,7 +1899,7 @@ STACK_OF(type) \
 #define sk_X509_NAME_sort(st) SKM_sk_sort(X509_NAME, (st))
 #define sk_X509_NAME_is_sorted(st) SKM_sk_is_sorted(X509_NAME, (st))
 
-#define sk_X509_NAME_ENTRY_new(st) SKM_sk_new(X509_NAME_ENTRY, (st))
+#define sk_X509_NAME_ENTRY_new(cmp) SKM_sk_new(X509_NAME_ENTRY, (cmp))
 #define sk_X509_NAME_ENTRY_new_null() SKM_sk_new_null(X509_NAME_ENTRY)
 #define sk_X509_NAME_ENTRY_free(st) SKM_sk_free(X509_NAME_ENTRY, (st))
 #define sk_X509_NAME_ENTRY_num(st) SKM_sk_num(X509_NAME_ENTRY, (st))
@@ -1602,7 +1921,7 @@ STACK_OF(type) \
 #define sk_X509_NAME_ENTRY_sort(st) SKM_sk_sort(X509_NAME_ENTRY, (st))
 #define sk_X509_NAME_ENTRY_is_sorted(st) SKM_sk_is_sorted(X509_NAME_ENTRY, (st))
 
-#define sk_X509_OBJECT_new(st) SKM_sk_new(X509_OBJECT, (st))
+#define sk_X509_OBJECT_new(cmp) SKM_sk_new(X509_OBJECT, (cmp))
 #define sk_X509_OBJECT_new_null() SKM_sk_new_null(X509_OBJECT)
 #define sk_X509_OBJECT_free(st) SKM_sk_free(X509_OBJECT, (st))
 #define sk_X509_OBJECT_num(st) SKM_sk_num(X509_OBJECT, (st))
@@ -1624,7 +1943,7 @@ STACK_OF(type) \
 #define sk_X509_OBJECT_sort(st) SKM_sk_sort(X509_OBJECT, (st))
 #define sk_X509_OBJECT_is_sorted(st) SKM_sk_is_sorted(X509_OBJECT, (st))
 
-#define sk_X509_POLICY_DATA_new(st) SKM_sk_new(X509_POLICY_DATA, (st))
+#define sk_X509_POLICY_DATA_new(cmp) SKM_sk_new(X509_POLICY_DATA, (cmp))
 #define sk_X509_POLICY_DATA_new_null() SKM_sk_new_null(X509_POLICY_DATA)
 #define sk_X509_POLICY_DATA_free(st) SKM_sk_free(X509_POLICY_DATA, (st))
 #define sk_X509_POLICY_DATA_num(st) SKM_sk_num(X509_POLICY_DATA, (st))
@@ -1646,7 +1965,7 @@ STACK_OF(type) \
 #define sk_X509_POLICY_DATA_sort(st) SKM_sk_sort(X509_POLICY_DATA, (st))
 #define sk_X509_POLICY_DATA_is_sorted(st) SKM_sk_is_sorted(X509_POLICY_DATA, (st))
 
-#define sk_X509_POLICY_NODE_new(st) SKM_sk_new(X509_POLICY_NODE, (st))
+#define sk_X509_POLICY_NODE_new(cmp) SKM_sk_new(X509_POLICY_NODE, (cmp))
 #define sk_X509_POLICY_NODE_new_null() SKM_sk_new_null(X509_POLICY_NODE)
 #define sk_X509_POLICY_NODE_free(st) SKM_sk_free(X509_POLICY_NODE, (st))
 #define sk_X509_POLICY_NODE_num(st) SKM_sk_num(X509_POLICY_NODE, (st))
@@ -1668,29 +1987,7 @@ STACK_OF(type) \
 #define sk_X509_POLICY_NODE_sort(st) SKM_sk_sort(X509_POLICY_NODE, (st))
 #define sk_X509_POLICY_NODE_is_sorted(st) SKM_sk_is_sorted(X509_POLICY_NODE, (st))
 
-#define sk_X509_POLICY_REF_new(st) SKM_sk_new(X509_POLICY_REF, (st))
-#define sk_X509_POLICY_REF_new_null() SKM_sk_new_null(X509_POLICY_REF)
-#define sk_X509_POLICY_REF_free(st) SKM_sk_free(X509_POLICY_REF, (st))
-#define sk_X509_POLICY_REF_num(st) SKM_sk_num(X509_POLICY_REF, (st))
-#define sk_X509_POLICY_REF_value(st, i) SKM_sk_value(X509_POLICY_REF, (st), (i))
-#define sk_X509_POLICY_REF_set(st, i, val) SKM_sk_set(X509_POLICY_REF, (st), (i), (val))
-#define sk_X509_POLICY_REF_zero(st) SKM_sk_zero(X509_POLICY_REF, (st))
-#define sk_X509_POLICY_REF_push(st, val) SKM_sk_push(X509_POLICY_REF, (st), (val))
-#define sk_X509_POLICY_REF_unshift(st, val) SKM_sk_unshift(X509_POLICY_REF, (st), (val))
-#define sk_X509_POLICY_REF_find(st, val) SKM_sk_find(X509_POLICY_REF, (st), (val))
-#define sk_X509_POLICY_REF_find_ex(st, val) SKM_sk_find_ex(X509_POLICY_REF, (st), (val))
-#define sk_X509_POLICY_REF_delete(st, i) SKM_sk_delete(X509_POLICY_REF, (st), (i))
-#define sk_X509_POLICY_REF_delete_ptr(st, ptr) SKM_sk_delete_ptr(X509_POLICY_REF, (st), (ptr))
-#define sk_X509_POLICY_REF_insert(st, val, i) SKM_sk_insert(X509_POLICY_REF, (st), (val), (i))
-#define sk_X509_POLICY_REF_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(X509_POLICY_REF, (st), (cmp))
-#define sk_X509_POLICY_REF_dup(st) SKM_sk_dup(X509_POLICY_REF, st)
-#define sk_X509_POLICY_REF_pop_free(st, free_func) SKM_sk_pop_free(X509_POLICY_REF, (st), (free_func))
-#define sk_X509_POLICY_REF_shift(st) SKM_sk_shift(X509_POLICY_REF, (st))
-#define sk_X509_POLICY_REF_pop(st) SKM_sk_pop(X509_POLICY_REF, (st))
-#define sk_X509_POLICY_REF_sort(st) SKM_sk_sort(X509_POLICY_REF, (st))
-#define sk_X509_POLICY_REF_is_sorted(st) SKM_sk_is_sorted(X509_POLICY_REF, (st))
-
-#define sk_X509_PURPOSE_new(st) SKM_sk_new(X509_PURPOSE, (st))
+#define sk_X509_PURPOSE_new(cmp) SKM_sk_new(X509_PURPOSE, (cmp))
 #define sk_X509_PURPOSE_new_null() SKM_sk_new_null(X509_PURPOSE)
 #define sk_X509_PURPOSE_free(st) SKM_sk_free(X509_PURPOSE, (st))
 #define sk_X509_PURPOSE_num(st) SKM_sk_num(X509_PURPOSE, (st))
@@ -1712,7 +2009,7 @@ STACK_OF(type) \
 #define sk_X509_PURPOSE_sort(st) SKM_sk_sort(X509_PURPOSE, (st))
 #define sk_X509_PURPOSE_is_sorted(st) SKM_sk_is_sorted(X509_PURPOSE, (st))
 
-#define sk_X509_REVOKED_new(st) SKM_sk_new(X509_REVOKED, (st))
+#define sk_X509_REVOKED_new(cmp) SKM_sk_new(X509_REVOKED, (cmp))
 #define sk_X509_REVOKED_new_null() SKM_sk_new_null(X509_REVOKED)
 #define sk_X509_REVOKED_free(st) SKM_sk_free(X509_REVOKED, (st))
 #define sk_X509_REVOKED_num(st) SKM_sk_num(X509_REVOKED, (st))
@@ -1734,7 +2031,7 @@ STACK_OF(type) \
 #define sk_X509_REVOKED_sort(st) SKM_sk_sort(X509_REVOKED, (st))
 #define sk_X509_REVOKED_is_sorted(st) SKM_sk_is_sorted(X509_REVOKED, (st))
 
-#define sk_X509_TRUST_new(st) SKM_sk_new(X509_TRUST, (st))
+#define sk_X509_TRUST_new(cmp) SKM_sk_new(X509_TRUST, (cmp))
 #define sk_X509_TRUST_new_null() SKM_sk_new_null(X509_TRUST)
 #define sk_X509_TRUST_free(st) SKM_sk_free(X509_TRUST, (st))
 #define sk_X509_TRUST_num(st) SKM_sk_num(X509_TRUST, (st))
@@ -1756,7 +2053,7 @@ STACK_OF(type) \
 #define sk_X509_TRUST_sort(st) SKM_sk_sort(X509_TRUST, (st))
 #define sk_X509_TRUST_is_sorted(st) SKM_sk_is_sorted(X509_TRUST, (st))
 
-#define sk_X509_VERIFY_PARAM_new(st) SKM_sk_new(X509_VERIFY_PARAM, (st))
+#define sk_X509_VERIFY_PARAM_new(cmp) SKM_sk_new(X509_VERIFY_PARAM, (cmp))
 #define sk_X509_VERIFY_PARAM_new_null() SKM_sk_new_null(X509_VERIFY_PARAM)
 #define sk_X509_VERIFY_PARAM_free(st) SKM_sk_free(X509_VERIFY_PARAM, (st))
 #define sk_X509_VERIFY_PARAM_num(st) SKM_sk_num(X509_VERIFY_PARAM, (st))
@@ -1778,6 +2075,125 @@ STACK_OF(type) \
 #define sk_X509_VERIFY_PARAM_sort(st) SKM_sk_sort(X509_VERIFY_PARAM, (st))
 #define sk_X509_VERIFY_PARAM_is_sorted(st) SKM_sk_is_sorted(X509_VERIFY_PARAM, (st))
 
+#define sk_nid_triple_new(cmp) SKM_sk_new(nid_triple, (cmp))
+#define sk_nid_triple_new_null() SKM_sk_new_null(nid_triple)
+#define sk_nid_triple_free(st) SKM_sk_free(nid_triple, (st))
+#define sk_nid_triple_num(st) SKM_sk_num(nid_triple, (st))
+#define sk_nid_triple_value(st, i) SKM_sk_value(nid_triple, (st), (i))
+#define sk_nid_triple_set(st, i, val) SKM_sk_set(nid_triple, (st), (i), (val))
+#define sk_nid_triple_zero(st) SKM_sk_zero(nid_triple, (st))
+#define sk_nid_triple_push(st, val) SKM_sk_push(nid_triple, (st), (val))
+#define sk_nid_triple_unshift(st, val) SKM_sk_unshift(nid_triple, (st), (val))
+#define sk_nid_triple_find(st, val) SKM_sk_find(nid_triple, (st), (val))
+#define sk_nid_triple_find_ex(st, val) SKM_sk_find_ex(nid_triple, (st), (val))
+#define sk_nid_triple_delete(st, i) SKM_sk_delete(nid_triple, (st), (i))
+#define sk_nid_triple_delete_ptr(st, ptr) SKM_sk_delete_ptr(nid_triple, (st), (ptr))
+#define sk_nid_triple_insert(st, val, i) SKM_sk_insert(nid_triple, (st), (val), (i))
+#define sk_nid_triple_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(nid_triple, (st), (cmp))
+#define sk_nid_triple_dup(st) SKM_sk_dup(nid_triple, st)
+#define sk_nid_triple_pop_free(st, free_func) SKM_sk_pop_free(nid_triple, (st), (free_func))
+#define sk_nid_triple_shift(st) SKM_sk_shift(nid_triple, (st))
+#define sk_nid_triple_pop(st) SKM_sk_pop(nid_triple, (st))
+#define sk_nid_triple_sort(st) SKM_sk_sort(nid_triple, (st))
+#define sk_nid_triple_is_sorted(st) SKM_sk_is_sorted(nid_triple, (st))
+
+#define sk_void_new(cmp) SKM_sk_new(void, (cmp))
+#define sk_void_new_null() SKM_sk_new_null(void)
+#define sk_void_free(st) SKM_sk_free(void, (st))
+#define sk_void_num(st) SKM_sk_num(void, (st))
+#define sk_void_value(st, i) SKM_sk_value(void, (st), (i))
+#define sk_void_set(st, i, val) SKM_sk_set(void, (st), (i), (val))
+#define sk_void_zero(st) SKM_sk_zero(void, (st))
+#define sk_void_push(st, val) SKM_sk_push(void, (st), (val))
+#define sk_void_unshift(st, val) SKM_sk_unshift(void, (st), (val))
+#define sk_void_find(st, val) SKM_sk_find(void, (st), (val))
+#define sk_void_find_ex(st, val) SKM_sk_find_ex(void, (st), (val))
+#define sk_void_delete(st, i) SKM_sk_delete(void, (st), (i))
+#define sk_void_delete_ptr(st, ptr) SKM_sk_delete_ptr(void, (st), (ptr))
+#define sk_void_insert(st, val, i) SKM_sk_insert(void, (st), (val), (i))
+#define sk_void_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(void, (st), (cmp))
+#define sk_void_dup(st) SKM_sk_dup(void, st)
+#define sk_void_pop_free(st, free_func) SKM_sk_pop_free(void, (st), (free_func))
+#define sk_void_shift(st) SKM_sk_shift(void, (st))
+#define sk_void_pop(st) SKM_sk_pop(void, (st))
+#define sk_void_sort(st) SKM_sk_sort(void, (st))
+#define sk_void_is_sorted(st) SKM_sk_is_sorted(void, (st))
+
+#define sk_OPENSSL_STRING_new(cmp) ((STACK_OF(OPENSSL_STRING) *)sk_new(CHECKED_SK_CMP_FUNC(char, cmp)))
+#define sk_OPENSSL_STRING_new_null() ((STACK_OF(OPENSSL_STRING) *)sk_new_null())
+#define sk_OPENSSL_STRING_push(st, val) sk_push(CHECKED_STACK_OF(OPENSSL_STRING, st), CHECKED_PTR_OF(char, val))
+#define sk_OPENSSL_STRING_find(st, val) sk_find(CHECKED_STACK_OF(OPENSSL_STRING, st), CHECKED_PTR_OF(char, val))
+#define sk_OPENSSL_STRING_value(st, i) ((OPENSSL_STRING)sk_value(CHECKED_STACK_OF(OPENSSL_STRING, st), i))
+#define sk_OPENSSL_STRING_num(st) SKM_sk_num(OPENSSL_STRING, st)
+#define sk_OPENSSL_STRING_pop_free(st, free_func) sk_pop_free(CHECKED_STACK_OF(OPENSSL_STRING, st), CHECKED_SK_FREE_FUNC2(OPENSSL_STRING, free_func))
+#define sk_OPENSSL_STRING_insert(st, val, i) sk_insert(CHECKED_STACK_OF(OPENSSL_STRING, st), CHECKED_PTR_OF(char, val), i)
+#define sk_OPENSSL_STRING_free(st) SKM_sk_free(OPENSSL_STRING, st)
+#define sk_OPENSSL_STRING_set(st, i, val) sk_set(CHECKED_STACK_OF(OPENSSL_STRING, st), i, CHECKED_PTR_OF(char, val))
+#define sk_OPENSSL_STRING_zero(st) SKM_sk_zero(OPENSSL_STRING, (st))
+#define sk_OPENSSL_STRING_unshift(st, val) sk_unshift(CHECKED_STACK_OF(OPENSSL_STRING, st), CHECKED_PTR_OF(char, val))
+#define sk_OPENSSL_STRING_find_ex(st, val) sk_find_ex((_STACK *)CHECKED_CONST_PTR_OF(STACK_OF(OPENSSL_STRING), st), CHECKED_CONST_PTR_OF(char, val))
+#define sk_OPENSSL_STRING_delete(st, i) SKM_sk_delete(OPENSSL_STRING, (st), (i))
+#define sk_OPENSSL_STRING_delete_ptr(st, ptr) (OPENSSL_STRING *)sk_delete_ptr(CHECKED_STACK_OF(OPENSSL_STRING, st), CHECKED_PTR_OF(char, ptr))
+#define sk_OPENSSL_STRING_set_cmp_func(st, cmp)  \
+	((int (*)(const char * const *,const char * const *)) \
+	sk_set_cmp_func(CHECKED_STACK_OF(OPENSSL_STRING, st), CHECKED_SK_CMP_FUNC(char, cmp)))
+#define sk_OPENSSL_STRING_dup(st) SKM_sk_dup(OPENSSL_STRING, st)
+#define sk_OPENSSL_STRING_shift(st) SKM_sk_shift(OPENSSL_STRING, (st))
+#define sk_OPENSSL_STRING_pop(st) (char *)sk_pop(CHECKED_STACK_OF(OPENSSL_STRING, st))
+#define sk_OPENSSL_STRING_sort(st) SKM_sk_sort(OPENSSL_STRING, (st))
+#define sk_OPENSSL_STRING_is_sorted(st) SKM_sk_is_sorted(OPENSSL_STRING, (st))
+
+
+#define sk_OPENSSL_BLOCK_new(cmp) ((STACK_OF(OPENSSL_BLOCK) *)sk_new(CHECKED_SK_CMP_FUNC(void, cmp)))
+#define sk_OPENSSL_BLOCK_new_null() ((STACK_OF(OPENSSL_BLOCK) *)sk_new_null())
+#define sk_OPENSSL_BLOCK_push(st, val) sk_push(CHECKED_STACK_OF(OPENSSL_BLOCK, st), CHECKED_PTR_OF(void, val))
+#define sk_OPENSSL_BLOCK_find(st, val) sk_find(CHECKED_STACK_OF(OPENSSL_BLOCK, st), CHECKED_PTR_OF(void, val))
+#define sk_OPENSSL_BLOCK_value(st, i) ((OPENSSL_BLOCK)sk_value(CHECKED_STACK_OF(OPENSSL_BLOCK, st), i))
+#define sk_OPENSSL_BLOCK_num(st) SKM_sk_num(OPENSSL_BLOCK, st)
+#define sk_OPENSSL_BLOCK_pop_free(st, free_func) sk_pop_free(CHECKED_STACK_OF(OPENSSL_BLOCK, st), CHECKED_SK_FREE_FUNC2(OPENSSL_BLOCK, free_func))
+#define sk_OPENSSL_BLOCK_insert(st, val, i) sk_insert(CHECKED_STACK_OF(OPENSSL_BLOCK, st), CHECKED_PTR_OF(void, val), i)
+#define sk_OPENSSL_BLOCK_free(st) SKM_sk_free(OPENSSL_BLOCK, st)
+#define sk_OPENSSL_BLOCK_set(st, i, val) sk_set(CHECKED_STACK_OF(OPENSSL_BLOCK, st), i, CHECKED_PTR_OF(void, val))
+#define sk_OPENSSL_BLOCK_zero(st) SKM_sk_zero(OPENSSL_BLOCK, (st))
+#define sk_OPENSSL_BLOCK_unshift(st, val) sk_unshift(CHECKED_STACK_OF(OPENSSL_BLOCK, st), CHECKED_PTR_OF(void, val))
+#define sk_OPENSSL_BLOCK_find_ex(st, val) sk_find_ex((_STACK *)CHECKED_CONST_PTR_OF(STACK_OF(OPENSSL_BLOCK), st), CHECKED_CONST_PTR_OF(void, val))
+#define sk_OPENSSL_BLOCK_delete(st, i) SKM_sk_delete(OPENSSL_BLOCK, (st), (i))
+#define sk_OPENSSL_BLOCK_delete_ptr(st, ptr) (OPENSSL_BLOCK *)sk_delete_ptr(CHECKED_STACK_OF(OPENSSL_BLOCK, st), CHECKED_PTR_OF(void, ptr))
+#define sk_OPENSSL_BLOCK_set_cmp_func(st, cmp)  \
+	((int (*)(const void * const *,const void * const *)) \
+	sk_set_cmp_func(CHECKED_STACK_OF(OPENSSL_BLOCK, st), CHECKED_SK_CMP_FUNC(void, cmp)))
+#define sk_OPENSSL_BLOCK_dup(st) SKM_sk_dup(OPENSSL_BLOCK, st)
+#define sk_OPENSSL_BLOCK_shift(st) SKM_sk_shift(OPENSSL_BLOCK, (st))
+#define sk_OPENSSL_BLOCK_pop(st) (void *)sk_pop(CHECKED_STACK_OF(OPENSSL_BLOCK, st))
+#define sk_OPENSSL_BLOCK_sort(st) SKM_sk_sort(OPENSSL_BLOCK, (st))
+#define sk_OPENSSL_BLOCK_is_sorted(st) SKM_sk_is_sorted(OPENSSL_BLOCK, (st))
+
+
+#define sk_OPENSSL_PSTRING_new(cmp) ((STACK_OF(OPENSSL_PSTRING) *)sk_new(CHECKED_SK_CMP_FUNC(OPENSSL_STRING, cmp)))
+#define sk_OPENSSL_PSTRING_new_null() ((STACK_OF(OPENSSL_PSTRING) *)sk_new_null())
+#define sk_OPENSSL_PSTRING_push(st, val) sk_push(CHECKED_STACK_OF(OPENSSL_PSTRING, st), CHECKED_PTR_OF(OPENSSL_STRING, val))
+#define sk_OPENSSL_PSTRING_find(st, val) sk_find(CHECKED_STACK_OF(OPENSSL_PSTRING, st), CHECKED_PTR_OF(OPENSSL_STRING, val))
+#define sk_OPENSSL_PSTRING_value(st, i) ((OPENSSL_PSTRING)sk_value(CHECKED_STACK_OF(OPENSSL_PSTRING, st), i))
+#define sk_OPENSSL_PSTRING_num(st) SKM_sk_num(OPENSSL_PSTRING, st)
+#define sk_OPENSSL_PSTRING_pop_free(st, free_func) sk_pop_free(CHECKED_STACK_OF(OPENSSL_PSTRING, st), CHECKED_SK_FREE_FUNC2(OPENSSL_PSTRING, free_func))
+#define sk_OPENSSL_PSTRING_insert(st, val, i) sk_insert(CHECKED_STACK_OF(OPENSSL_PSTRING, st), CHECKED_PTR_OF(OPENSSL_STRING, val), i)
+#define sk_OPENSSL_PSTRING_free(st) SKM_sk_free(OPENSSL_PSTRING, st)
+#define sk_OPENSSL_PSTRING_set(st, i, val) sk_set(CHECKED_STACK_OF(OPENSSL_PSTRING, st), i, CHECKED_PTR_OF(OPENSSL_STRING, val))
+#define sk_OPENSSL_PSTRING_zero(st) SKM_sk_zero(OPENSSL_PSTRING, (st))
+#define sk_OPENSSL_PSTRING_unshift(st, val) sk_unshift(CHECKED_STACK_OF(OPENSSL_PSTRING, st), CHECKED_PTR_OF(OPENSSL_STRING, val))
+#define sk_OPENSSL_PSTRING_find_ex(st, val) sk_find_ex((_STACK *)CHECKED_CONST_PTR_OF(STACK_OF(OPENSSL_PSTRING), st), CHECKED_CONST_PTR_OF(OPENSSL_STRING, val))
+#define sk_OPENSSL_PSTRING_delete(st, i) SKM_sk_delete(OPENSSL_PSTRING, (st), (i))
+#define sk_OPENSSL_PSTRING_delete_ptr(st, ptr) (OPENSSL_PSTRING *)sk_delete_ptr(CHECKED_STACK_OF(OPENSSL_PSTRING, st), CHECKED_PTR_OF(OPENSSL_STRING, ptr))
+#define sk_OPENSSL_PSTRING_set_cmp_func(st, cmp)  \
+	((int (*)(const OPENSSL_STRING * const *,const OPENSSL_STRING * const *)) \
+	sk_set_cmp_func(CHECKED_STACK_OF(OPENSSL_PSTRING, st), CHECKED_SK_CMP_FUNC(OPENSSL_STRING, cmp)))
+#define sk_OPENSSL_PSTRING_dup(st) SKM_sk_dup(OPENSSL_PSTRING, st)
+#define sk_OPENSSL_PSTRING_shift(st) SKM_sk_shift(OPENSSL_PSTRING, (st))
+#define sk_OPENSSL_PSTRING_pop(st) (OPENSSL_STRING *)sk_pop(CHECKED_STACK_OF(OPENSSL_PSTRING, st))
+#define sk_OPENSSL_PSTRING_sort(st) SKM_sk_sort(OPENSSL_PSTRING, (st))
+#define sk_OPENSSL_PSTRING_is_sorted(st) SKM_sk_is_sorted(OPENSSL_PSTRING, (st))
+
+
 #define d2i_ASN1_SET_OF_ACCESS_DESCRIPTION(st, pp, length, d2i_func, free_func, ex_tag, ex_class) \
 	SKM_ASN1_SET_OF_d2i(ACCESS_DESCRIPTION, (st), (pp), (length), (d2i_func), (free_func), (ex_tag), (ex_class)) 
 #define i2d_ASN1_SET_OF_ACCESS_DESCRIPTION(st, pp, i2d_func, ex_tag, ex_class, is_set) \
@@ -1814,6 +2230,15 @@ STACK_OF(type) \
 #define ASN1_seq_unpack_ASN1_TYPE(buf, len, d2i_func, free_func) \
 	SKM_ASN1_seq_unpack(ASN1_TYPE, (buf), (len), (d2i_func), (free_func))
 
+#define d2i_ASN1_SET_OF_ASN1_UTF8STRING(st, pp, length, d2i_func, free_func, ex_tag, ex_class) \
+	SKM_ASN1_SET_OF_d2i(ASN1_UTF8STRING, (st), (pp), (length), (d2i_func), (free_func), (ex_tag), (ex_class)) 
+#define i2d_ASN1_SET_OF_ASN1_UTF8STRING(st, pp, i2d_func, ex_tag, ex_class, is_set) \
+	SKM_ASN1_SET_OF_i2d(ASN1_UTF8STRING, (st), (pp), (i2d_func), (ex_tag), (ex_class), (is_set))
+#define ASN1_seq_pack_ASN1_UTF8STRING(st, i2d_func, buf, len) \
+	SKM_ASN1_seq_pack(ASN1_UTF8STRING, (st), (i2d_func), (buf), (len))
+#define ASN1_seq_unpack_ASN1_UTF8STRING(buf, len, d2i_func, free_func) \
+	SKM_ASN1_seq_unpack(ASN1_UTF8STRING, (buf), (len), (d2i_func), (free_func))
+
 #define d2i_ASN1_SET_OF_DIST_POINT(st, pp, length, d2i_func, free_func, ex_tag, ex_class) \
 	SKM_ASN1_SET_OF_d2i(DIST_POINT, (st), (pp), (length), (d2i_func), (free_func), (ex_tag), (ex_class)) 
 #define i2d_ASN1_SET_OF_DIST_POINT(st, pp, i2d_func, ex_tag, ex_class, is_set) \
@@ -1823,6 +2248,24 @@ STACK_OF(type) \
 #define ASN1_seq_unpack_DIST_POINT(buf, len, d2i_func, free_func) \
 	SKM_ASN1_seq_unpack(DIST_POINT, (buf), (len), (d2i_func), (free_func))
 
+#define d2i_ASN1_SET_OF_ESS_CERT_ID(st, pp, length, d2i_func, free_func, ex_tag, ex_class) \
+	SKM_ASN1_SET_OF_d2i(ESS_CERT_ID, (st), (pp), (length), (d2i_func), (free_func), (ex_tag), (ex_class)) 
+#define i2d_ASN1_SET_OF_ESS_CERT_ID(st, pp, i2d_func, ex_tag, ex_class, is_set) \
+	SKM_ASN1_SET_OF_i2d(ESS_CERT_ID, (st), (pp), (i2d_func), (ex_tag), (ex_class), (is_set))
+#define ASN1_seq_pack_ESS_CERT_ID(st, i2d_func, buf, len) \
+	SKM_ASN1_seq_pack(ESS_CERT_ID, (st), (i2d_func), (buf), (len))
+#define ASN1_seq_unpack_ESS_CERT_ID(buf, len, d2i_func, free_func) \
+	SKM_ASN1_seq_unpack(ESS_CERT_ID, (buf), (len), (d2i_func), (free_func))
+
+#define d2i_ASN1_SET_OF_EVP_MD(st, pp, length, d2i_func, free_func, ex_tag, ex_class) \
+	SKM_ASN1_SET_OF_d2i(EVP_MD, (st), (pp), (length), (d2i_func), (free_func), (ex_tag), (ex_class)) 
+#define i2d_ASN1_SET_OF_EVP_MD(st, pp, i2d_func, ex_tag, ex_class, is_set) \
+	SKM_ASN1_SET_OF_i2d(EVP_MD, (st), (pp), (i2d_func), (ex_tag), (ex_class), (is_set))
+#define ASN1_seq_pack_EVP_MD(st, i2d_func, buf, len) \
+	SKM_ASN1_seq_pack(EVP_MD, (st), (i2d_func), (buf), (len))
+#define ASN1_seq_unpack_EVP_MD(buf, len, d2i_func, free_func) \
+	SKM_ASN1_seq_unpack(EVP_MD, (buf), (len), (d2i_func), (free_func))
+
 #define d2i_ASN1_SET_OF_GENERAL_NAME(st, pp, length, d2i_func, free_func, ex_tag, ex_class) \
 	SKM_ASN1_SET_OF_d2i(GENERAL_NAME, (st), (pp), (length), (d2i_func), (free_func), (ex_tag), (ex_class)) 
 #define i2d_ASN1_SET_OF_GENERAL_NAME(st, pp, i2d_func, ex_tag, ex_class, is_set) \
@@ -1981,6 +2424,240 @@ STACK_OF(type) \
 
 #define PKCS12_decrypt_d2i_PKCS7(algor, d2i_func, free_func, pass, passlen, oct, seq) \
 	SKM_PKCS12_decrypt_d2i(PKCS7, (algor), (d2i_func), (free_func), (pass), (passlen), (oct), (seq))
+
+#define lh_ADDED_OBJ_new() LHM_lh_new(ADDED_OBJ,added_obj)
+#define lh_ADDED_OBJ_insert(lh,inst) LHM_lh_insert(ADDED_OBJ,lh,inst)
+#define lh_ADDED_OBJ_retrieve(lh,inst) LHM_lh_retrieve(ADDED_OBJ,lh,inst)
+#define lh_ADDED_OBJ_delete(lh,inst) LHM_lh_delete(ADDED_OBJ,lh,inst)
+#define lh_ADDED_OBJ_doall(lh,fn) LHM_lh_doall(ADDED_OBJ,lh,fn)
+#define lh_ADDED_OBJ_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(ADDED_OBJ,lh,fn,arg_type,arg)
+#define lh_ADDED_OBJ_error(lh) LHM_lh_error(ADDED_OBJ,lh)
+#define lh_ADDED_OBJ_num_items(lh) LHM_lh_num_items(ADDED_OBJ,lh)
+#define lh_ADDED_OBJ_down_load(lh) LHM_lh_down_load(ADDED_OBJ,lh)
+#define lh_ADDED_OBJ_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(ADDED_OBJ,lh,out)
+#define lh_ADDED_OBJ_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(ADDED_OBJ,lh,out)
+#define lh_ADDED_OBJ_stats_bio(lh,out) \
+  LHM_lh_stats_bio(ADDED_OBJ,lh,out)
+#define lh_ADDED_OBJ_free(lh) LHM_lh_free(ADDED_OBJ,lh)
+
+#define lh_APP_INFO_new() LHM_lh_new(APP_INFO,app_info)
+#define lh_APP_INFO_insert(lh,inst) LHM_lh_insert(APP_INFO,lh,inst)
+#define lh_APP_INFO_retrieve(lh,inst) LHM_lh_retrieve(APP_INFO,lh,inst)
+#define lh_APP_INFO_delete(lh,inst) LHM_lh_delete(APP_INFO,lh,inst)
+#define lh_APP_INFO_doall(lh,fn) LHM_lh_doall(APP_INFO,lh,fn)
+#define lh_APP_INFO_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(APP_INFO,lh,fn,arg_type,arg)
+#define lh_APP_INFO_error(lh) LHM_lh_error(APP_INFO,lh)
+#define lh_APP_INFO_num_items(lh) LHM_lh_num_items(APP_INFO,lh)
+#define lh_APP_INFO_down_load(lh) LHM_lh_down_load(APP_INFO,lh)
+#define lh_APP_INFO_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(APP_INFO,lh,out)
+#define lh_APP_INFO_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(APP_INFO,lh,out)
+#define lh_APP_INFO_stats_bio(lh,out) \
+  LHM_lh_stats_bio(APP_INFO,lh,out)
+#define lh_APP_INFO_free(lh) LHM_lh_free(APP_INFO,lh)
+
+#define lh_CONF_VALUE_new() LHM_lh_new(CONF_VALUE,conf_value)
+#define lh_CONF_VALUE_insert(lh,inst) LHM_lh_insert(CONF_VALUE,lh,inst)
+#define lh_CONF_VALUE_retrieve(lh,inst) LHM_lh_retrieve(CONF_VALUE,lh,inst)
+#define lh_CONF_VALUE_delete(lh,inst) LHM_lh_delete(CONF_VALUE,lh,inst)
+#define lh_CONF_VALUE_doall(lh,fn) LHM_lh_doall(CONF_VALUE,lh,fn)
+#define lh_CONF_VALUE_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(CONF_VALUE,lh,fn,arg_type,arg)
+#define lh_CONF_VALUE_error(lh) LHM_lh_error(CONF_VALUE,lh)
+#define lh_CONF_VALUE_num_items(lh) LHM_lh_num_items(CONF_VALUE,lh)
+#define lh_CONF_VALUE_down_load(lh) LHM_lh_down_load(CONF_VALUE,lh)
+#define lh_CONF_VALUE_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(CONF_VALUE,lh,out)
+#define lh_CONF_VALUE_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(CONF_VALUE,lh,out)
+#define lh_CONF_VALUE_stats_bio(lh,out) \
+  LHM_lh_stats_bio(CONF_VALUE,lh,out)
+#define lh_CONF_VALUE_free(lh) LHM_lh_free(CONF_VALUE,lh)
+
+#define lh_ENGINE_PILE_new() LHM_lh_new(ENGINE_PILE,engine_pile)
+#define lh_ENGINE_PILE_insert(lh,inst) LHM_lh_insert(ENGINE_PILE,lh,inst)
+#define lh_ENGINE_PILE_retrieve(lh,inst) LHM_lh_retrieve(ENGINE_PILE,lh,inst)
+#define lh_ENGINE_PILE_delete(lh,inst) LHM_lh_delete(ENGINE_PILE,lh,inst)
+#define lh_ENGINE_PILE_doall(lh,fn) LHM_lh_doall(ENGINE_PILE,lh,fn)
+#define lh_ENGINE_PILE_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(ENGINE_PILE,lh,fn,arg_type,arg)
+#define lh_ENGINE_PILE_error(lh) LHM_lh_error(ENGINE_PILE,lh)
+#define lh_ENGINE_PILE_num_items(lh) LHM_lh_num_items(ENGINE_PILE,lh)
+#define lh_ENGINE_PILE_down_load(lh) LHM_lh_down_load(ENGINE_PILE,lh)
+#define lh_ENGINE_PILE_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(ENGINE_PILE,lh,out)
+#define lh_ENGINE_PILE_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(ENGINE_PILE,lh,out)
+#define lh_ENGINE_PILE_stats_bio(lh,out) \
+  LHM_lh_stats_bio(ENGINE_PILE,lh,out)
+#define lh_ENGINE_PILE_free(lh) LHM_lh_free(ENGINE_PILE,lh)
+
+#define lh_ERR_STATE_new() LHM_lh_new(ERR_STATE,err_state)
+#define lh_ERR_STATE_insert(lh,inst) LHM_lh_insert(ERR_STATE,lh,inst)
+#define lh_ERR_STATE_retrieve(lh,inst) LHM_lh_retrieve(ERR_STATE,lh,inst)
+#define lh_ERR_STATE_delete(lh,inst) LHM_lh_delete(ERR_STATE,lh,inst)
+#define lh_ERR_STATE_doall(lh,fn) LHM_lh_doall(ERR_STATE,lh,fn)
+#define lh_ERR_STATE_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(ERR_STATE,lh,fn,arg_type,arg)
+#define lh_ERR_STATE_error(lh) LHM_lh_error(ERR_STATE,lh)
+#define lh_ERR_STATE_num_items(lh) LHM_lh_num_items(ERR_STATE,lh)
+#define lh_ERR_STATE_down_load(lh) LHM_lh_down_load(ERR_STATE,lh)
+#define lh_ERR_STATE_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(ERR_STATE,lh,out)
+#define lh_ERR_STATE_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(ERR_STATE,lh,out)
+#define lh_ERR_STATE_stats_bio(lh,out) \
+  LHM_lh_stats_bio(ERR_STATE,lh,out)
+#define lh_ERR_STATE_free(lh) LHM_lh_free(ERR_STATE,lh)
+
+#define lh_ERR_STRING_DATA_new() LHM_lh_new(ERR_STRING_DATA,err_string_data)
+#define lh_ERR_STRING_DATA_insert(lh,inst) LHM_lh_insert(ERR_STRING_DATA,lh,inst)
+#define lh_ERR_STRING_DATA_retrieve(lh,inst) LHM_lh_retrieve(ERR_STRING_DATA,lh,inst)
+#define lh_ERR_STRING_DATA_delete(lh,inst) LHM_lh_delete(ERR_STRING_DATA,lh,inst)
+#define lh_ERR_STRING_DATA_doall(lh,fn) LHM_lh_doall(ERR_STRING_DATA,lh,fn)
+#define lh_ERR_STRING_DATA_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(ERR_STRING_DATA,lh,fn,arg_type,arg)
+#define lh_ERR_STRING_DATA_error(lh) LHM_lh_error(ERR_STRING_DATA,lh)
+#define lh_ERR_STRING_DATA_num_items(lh) LHM_lh_num_items(ERR_STRING_DATA,lh)
+#define lh_ERR_STRING_DATA_down_load(lh) LHM_lh_down_load(ERR_STRING_DATA,lh)
+#define lh_ERR_STRING_DATA_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(ERR_STRING_DATA,lh,out)
+#define lh_ERR_STRING_DATA_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(ERR_STRING_DATA,lh,out)
+#define lh_ERR_STRING_DATA_stats_bio(lh,out) \
+  LHM_lh_stats_bio(ERR_STRING_DATA,lh,out)
+#define lh_ERR_STRING_DATA_free(lh) LHM_lh_free(ERR_STRING_DATA,lh)
+
+#define lh_EX_CLASS_ITEM_new() LHM_lh_new(EX_CLASS_ITEM,ex_class_item)
+#define lh_EX_CLASS_ITEM_insert(lh,inst) LHM_lh_insert(EX_CLASS_ITEM,lh,inst)
+#define lh_EX_CLASS_ITEM_retrieve(lh,inst) LHM_lh_retrieve(EX_CLASS_ITEM,lh,inst)
+#define lh_EX_CLASS_ITEM_delete(lh,inst) LHM_lh_delete(EX_CLASS_ITEM,lh,inst)
+#define lh_EX_CLASS_ITEM_doall(lh,fn) LHM_lh_doall(EX_CLASS_ITEM,lh,fn)
+#define lh_EX_CLASS_ITEM_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(EX_CLASS_ITEM,lh,fn,arg_type,arg)
+#define lh_EX_CLASS_ITEM_error(lh) LHM_lh_error(EX_CLASS_ITEM,lh)
+#define lh_EX_CLASS_ITEM_num_items(lh) LHM_lh_num_items(EX_CLASS_ITEM,lh)
+#define lh_EX_CLASS_ITEM_down_load(lh) LHM_lh_down_load(EX_CLASS_ITEM,lh)
+#define lh_EX_CLASS_ITEM_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(EX_CLASS_ITEM,lh,out)
+#define lh_EX_CLASS_ITEM_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(EX_CLASS_ITEM,lh,out)
+#define lh_EX_CLASS_ITEM_stats_bio(lh,out) \
+  LHM_lh_stats_bio(EX_CLASS_ITEM,lh,out)
+#define lh_EX_CLASS_ITEM_free(lh) LHM_lh_free(EX_CLASS_ITEM,lh)
+
+#define lh_FUNCTION_new() LHM_lh_new(FUNCTION,function)
+#define lh_FUNCTION_insert(lh,inst) LHM_lh_insert(FUNCTION,lh,inst)
+#define lh_FUNCTION_retrieve(lh,inst) LHM_lh_retrieve(FUNCTION,lh,inst)
+#define lh_FUNCTION_delete(lh,inst) LHM_lh_delete(FUNCTION,lh,inst)
+#define lh_FUNCTION_doall(lh,fn) LHM_lh_doall(FUNCTION,lh,fn)
+#define lh_FUNCTION_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(FUNCTION,lh,fn,arg_type,arg)
+#define lh_FUNCTION_error(lh) LHM_lh_error(FUNCTION,lh)
+#define lh_FUNCTION_num_items(lh) LHM_lh_num_items(FUNCTION,lh)
+#define lh_FUNCTION_down_load(lh) LHM_lh_down_load(FUNCTION,lh)
+#define lh_FUNCTION_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(FUNCTION,lh,out)
+#define lh_FUNCTION_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(FUNCTION,lh,out)
+#define lh_FUNCTION_stats_bio(lh,out) \
+  LHM_lh_stats_bio(FUNCTION,lh,out)
+#define lh_FUNCTION_free(lh) LHM_lh_free(FUNCTION,lh)
+
+#define lh_MEM_new() LHM_lh_new(MEM,mem)
+#define lh_MEM_insert(lh,inst) LHM_lh_insert(MEM,lh,inst)
+#define lh_MEM_retrieve(lh,inst) LHM_lh_retrieve(MEM,lh,inst)
+#define lh_MEM_delete(lh,inst) LHM_lh_delete(MEM,lh,inst)
+#define lh_MEM_doall(lh,fn) LHM_lh_doall(MEM,lh,fn)
+#define lh_MEM_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(MEM,lh,fn,arg_type,arg)
+#define lh_MEM_error(lh) LHM_lh_error(MEM,lh)
+#define lh_MEM_num_items(lh) LHM_lh_num_items(MEM,lh)
+#define lh_MEM_down_load(lh) LHM_lh_down_load(MEM,lh)
+#define lh_MEM_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(MEM,lh,out)
+#define lh_MEM_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(MEM,lh,out)
+#define lh_MEM_stats_bio(lh,out) \
+  LHM_lh_stats_bio(MEM,lh,out)
+#define lh_MEM_free(lh) LHM_lh_free(MEM,lh)
+
+#define lh_OBJ_NAME_new() LHM_lh_new(OBJ_NAME,obj_name)
+#define lh_OBJ_NAME_insert(lh,inst) LHM_lh_insert(OBJ_NAME,lh,inst)
+#define lh_OBJ_NAME_retrieve(lh,inst) LHM_lh_retrieve(OBJ_NAME,lh,inst)
+#define lh_OBJ_NAME_delete(lh,inst) LHM_lh_delete(OBJ_NAME,lh,inst)
+#define lh_OBJ_NAME_doall(lh,fn) LHM_lh_doall(OBJ_NAME,lh,fn)
+#define lh_OBJ_NAME_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(OBJ_NAME,lh,fn,arg_type,arg)
+#define lh_OBJ_NAME_error(lh) LHM_lh_error(OBJ_NAME,lh)
+#define lh_OBJ_NAME_num_items(lh) LHM_lh_num_items(OBJ_NAME,lh)
+#define lh_OBJ_NAME_down_load(lh) LHM_lh_down_load(OBJ_NAME,lh)
+#define lh_OBJ_NAME_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(OBJ_NAME,lh,out)
+#define lh_OBJ_NAME_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(OBJ_NAME,lh,out)
+#define lh_OBJ_NAME_stats_bio(lh,out) \
+  LHM_lh_stats_bio(OBJ_NAME,lh,out)
+#define lh_OBJ_NAME_free(lh) LHM_lh_free(OBJ_NAME,lh)
+
+#define lh_OPENSSL_CSTRING_new() LHM_lh_new(OPENSSL_CSTRING,openssl_cstring)
+#define lh_OPENSSL_CSTRING_insert(lh,inst) LHM_lh_insert(OPENSSL_CSTRING,lh,inst)
+#define lh_OPENSSL_CSTRING_retrieve(lh,inst) LHM_lh_retrieve(OPENSSL_CSTRING,lh,inst)
+#define lh_OPENSSL_CSTRING_delete(lh,inst) LHM_lh_delete(OPENSSL_CSTRING,lh,inst)
+#define lh_OPENSSL_CSTRING_doall(lh,fn) LHM_lh_doall(OPENSSL_CSTRING,lh,fn)
+#define lh_OPENSSL_CSTRING_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(OPENSSL_CSTRING,lh,fn,arg_type,arg)
+#define lh_OPENSSL_CSTRING_error(lh) LHM_lh_error(OPENSSL_CSTRING,lh)
+#define lh_OPENSSL_CSTRING_num_items(lh) LHM_lh_num_items(OPENSSL_CSTRING,lh)
+#define lh_OPENSSL_CSTRING_down_load(lh) LHM_lh_down_load(OPENSSL_CSTRING,lh)
+#define lh_OPENSSL_CSTRING_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(OPENSSL_CSTRING,lh,out)
+#define lh_OPENSSL_CSTRING_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(OPENSSL_CSTRING,lh,out)
+#define lh_OPENSSL_CSTRING_stats_bio(lh,out) \
+  LHM_lh_stats_bio(OPENSSL_CSTRING,lh,out)
+#define lh_OPENSSL_CSTRING_free(lh) LHM_lh_free(OPENSSL_CSTRING,lh)
+
+#define lh_OPENSSL_STRING_new() LHM_lh_new(OPENSSL_STRING,openssl_string)
+#define lh_OPENSSL_STRING_insert(lh,inst) LHM_lh_insert(OPENSSL_STRING,lh,inst)
+#define lh_OPENSSL_STRING_retrieve(lh,inst) LHM_lh_retrieve(OPENSSL_STRING,lh,inst)
+#define lh_OPENSSL_STRING_delete(lh,inst) LHM_lh_delete(OPENSSL_STRING,lh,inst)
+#define lh_OPENSSL_STRING_doall(lh,fn) LHM_lh_doall(OPENSSL_STRING,lh,fn)
+#define lh_OPENSSL_STRING_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(OPENSSL_STRING,lh,fn,arg_type,arg)
+#define lh_OPENSSL_STRING_error(lh) LHM_lh_error(OPENSSL_STRING,lh)
+#define lh_OPENSSL_STRING_num_items(lh) LHM_lh_num_items(OPENSSL_STRING,lh)
+#define lh_OPENSSL_STRING_down_load(lh) LHM_lh_down_load(OPENSSL_STRING,lh)
+#define lh_OPENSSL_STRING_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(OPENSSL_STRING,lh,out)
+#define lh_OPENSSL_STRING_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(OPENSSL_STRING,lh,out)
+#define lh_OPENSSL_STRING_stats_bio(lh,out) \
+  LHM_lh_stats_bio(OPENSSL_STRING,lh,out)
+#define lh_OPENSSL_STRING_free(lh) LHM_lh_free(OPENSSL_STRING,lh)
+
+#define lh_SSL_SESSION_new() LHM_lh_new(SSL_SESSION,ssl_session)
+#define lh_SSL_SESSION_insert(lh,inst) LHM_lh_insert(SSL_SESSION,lh,inst)
+#define lh_SSL_SESSION_retrieve(lh,inst) LHM_lh_retrieve(SSL_SESSION,lh,inst)
+#define lh_SSL_SESSION_delete(lh,inst) LHM_lh_delete(SSL_SESSION,lh,inst)
+#define lh_SSL_SESSION_doall(lh,fn) LHM_lh_doall(SSL_SESSION,lh,fn)
+#define lh_SSL_SESSION_doall_arg(lh,fn,arg_type,arg) \
+  LHM_lh_doall_arg(SSL_SESSION,lh,fn,arg_type,arg)
+#define lh_SSL_SESSION_error(lh) LHM_lh_error(SSL_SESSION,lh)
+#define lh_SSL_SESSION_num_items(lh) LHM_lh_num_items(SSL_SESSION,lh)
+#define lh_SSL_SESSION_down_load(lh) LHM_lh_down_load(SSL_SESSION,lh)
+#define lh_SSL_SESSION_node_stats_bio(lh,out) \
+  LHM_lh_node_stats_bio(SSL_SESSION,lh,out)
+#define lh_SSL_SESSION_node_usage_stats_bio(lh,out) \
+  LHM_lh_node_usage_stats_bio(SSL_SESSION,lh,out)
+#define lh_SSL_SESSION_stats_bio(lh,out) \
+  LHM_lh_stats_bio(SSL_SESSION,lh,out)
+#define lh_SSL_SESSION_free(lh) LHM_lh_free(SSL_SESSION,lh)
 /* End of util/mkstack.pl block, you may now edit :-) */
 
 #endif /* !defined HEADER_SAFESTACK_H */
diff --git a/crypto/stack/stack.c b/crypto/stack/stack.c
index 378bd7c7968c..76cf1a116897 100644
--- a/crypto/stack/stack.c
+++ b/crypto/stack/stack.c
@@ -77,10 +77,10 @@ const char STACK_version[]="Stack" OPENSSL_VERSION_PTEXT;
 
 #include <errno.h>
 
-int (*sk_set_cmp_func(STACK *sk, int (*c)(const char * const *,const char * const *)))
-		(const char * const *, const char * const *)
+int (*sk_set_cmp_func(_STACK *sk, int (*c)(const void *, const void *)))
+		(const void *, const void *)
 	{
-	int (*old)(const char * const *,const char * const *)=sk->comp;
+	int (*old)(const void *,const void *)=sk->comp;
 
 	if (sk->comp != c)
 		sk->sorted=0;
@@ -89,9 +89,9 @@ int (*sk_set_cmp_func(STACK *sk, int (*c)(const char * const *,const char * cons
 	return old;
 	}
 
-STACK *sk_dup(STACK *sk)
+_STACK *sk_dup(_STACK *sk)
 	{
-	STACK *ret;
+	_STACK *ret;
 	char **s;
 
 	if ((ret=sk_new(sk->comp)) == NULL) goto err;
@@ -112,19 +112,19 @@ err:
 	return(NULL);
 	}
 
-STACK *sk_new_null(void)
+_STACK *sk_new_null(void)
 	{
-	return sk_new((int (*)(const char * const *, const char * const *))0);
+	return sk_new((int (*)(const void *, const void *))0);
 	}
 
-STACK *sk_new(int (*c)(const char * const *, const char * const *))
+_STACK *sk_new(int (*c)(const void *, const void *))
 	{
-	STACK *ret;
+	_STACK *ret;
 	int i;
 
-	if ((ret=(STACK *)OPENSSL_malloc(sizeof(STACK))) == NULL)
+	if ((ret=OPENSSL_malloc(sizeof(_STACK))) == NULL)
 		goto err;
-	if ((ret->data=(char **)OPENSSL_malloc(sizeof(char *)*MIN_NODES)) == NULL)
+	if ((ret->data=OPENSSL_malloc(sizeof(char *)*MIN_NODES)) == NULL)
 		goto err;
 	for (i=0; i<MIN_NODES; i++)
 		ret->data[i]=NULL;
@@ -139,14 +139,14 @@ err:
 	return(NULL);
 	}
 
-int sk_insert(STACK *st, char *data, int loc)
+int sk_insert(_STACK *st, void *data, int loc)
 	{
 	char **s;
 
 	if(st == NULL) return 0;
 	if (st->num_alloc <= st->num+1)
 		{
-		s=(char **)OPENSSL_realloc((char *)st->data,
+		s=OPENSSL_realloc((char *)st->data,
 			(unsigned int)sizeof(char *)*st->num_alloc*2);
 		if (s == NULL)
 			return(0);
@@ -160,14 +160,14 @@ int sk_insert(STACK *st, char *data, int loc)
 		int i;
 		char **f,**t;
 
-		f=(char **)st->data;
-		t=(char **)&(st->data[1]);
+		f=st->data;
+		t=&(st->data[1]);
 		for (i=st->num; i>=loc; i--)
 			t[i]=f[i];
 			
 #ifdef undef /* no memmove on sunos :-( */
-		memmove( (char *)&(st->data[loc+1]),
-			(char *)&(st->data[loc]),
+		memmove(&(st->data[loc+1]),
+			&(st->data[loc]),
 			sizeof(char *)*(st->num-loc));
 #endif
 		st->data[loc]=data;
@@ -177,7 +177,7 @@ int sk_insert(STACK *st, char *data, int loc)
 	return(st->num);
 	}
 
-char *sk_delete_ptr(STACK *st, char *p)
+void *sk_delete_ptr(_STACK *st, void *p)
 	{
 	int i;
 
@@ -187,7 +187,7 @@ char *sk_delete_ptr(STACK *st, char *p)
 	return(NULL);
 	}
 
-char *sk_delete(STACK *st, int loc)
+void *sk_delete(_STACK *st, int loc)
 	{
 	char *ret;
 	int i,j;
@@ -210,11 +210,11 @@ char *sk_delete(STACK *st, int loc)
 	return(ret);
 	}
 
-static int internal_find(STACK *st, char *data, int ret_val_options)
+static int internal_find(_STACK *st, void *data, int ret_val_options)
 	{
-	char **r;
+	const void * const *r;
 	int i;
-	int (*comp_func)(const void *,const void *);
+
 	if(st == NULL) return -1;
 
 	if (st->comp == NULL)
@@ -226,53 +226,46 @@ static int internal_find(STACK *st, char *data, int ret_val_options)
 		}
 	sk_sort(st);
 	if (data == NULL) return(-1);
-	/* This (and the "qsort" below) are the two places in OpenSSL
-	 * where we need to convert from our standard (type **,type **)
-	 * compare callback type to the (void *,void *) type required by
-	 * bsearch. However, the "data" it is being called(back) with are
-	 * not (type *) pointers, but the *pointers* to (type *) pointers,
-	 * so we get our extra level of pointer dereferencing that way. */
-	comp_func=(int (*)(const void *,const void *))(st->comp);
-	r=(char **)OBJ_bsearch_ex((char *)&data,(char *)st->data,
-		st->num,sizeof(char *),comp_func,ret_val_options);
+	r=OBJ_bsearch_ex_(&data,st->data,st->num,sizeof(void *),st->comp,
+			  ret_val_options);
 	if (r == NULL) return(-1);
-	return((int)(r-st->data));
+	return (int)((char **)r-st->data);
 	}
 
-int sk_find(STACK *st, char *data)
+int sk_find(_STACK *st, void *data)
 	{
 	return internal_find(st, data, OBJ_BSEARCH_FIRST_VALUE_ON_MATCH);
 	}
-int sk_find_ex(STACK *st, char *data)
+int sk_find_ex(_STACK *st, void *data)
 	{
 	return internal_find(st, data, OBJ_BSEARCH_VALUE_ON_NOMATCH);
 	}
 
-int sk_push(STACK *st, char *data)
+int sk_push(_STACK *st, void *data)
 	{
 	return(sk_insert(st,data,st->num));
 	}
 
-int sk_unshift(STACK *st, char *data)
+int sk_unshift(_STACK *st, void *data)
 	{
 	return(sk_insert(st,data,0));
 	}
 
-char *sk_shift(STACK *st)
+void *sk_shift(_STACK *st)
 	{
 	if (st == NULL) return(NULL);
 	if (st->num <= 0) return(NULL);
 	return(sk_delete(st,0));
 	}
 
-char *sk_pop(STACK *st)
+void *sk_pop(_STACK *st)
 	{
 	if (st == NULL) return(NULL);
 	if (st->num <= 0) return(NULL);
 	return(sk_delete(st,st->num-1));
 	}
 
-void sk_zero(STACK *st)
+void sk_zero(_STACK *st)
 	{
 	if (st == NULL) return;
 	if (st->num <= 0) return;
@@ -280,7 +273,7 @@ void sk_zero(STACK *st)
 	st->num=0;
 	}
 
-void sk_pop_free(STACK *st, void (*func)(void *))
+void sk_pop_free(_STACK *st, void (*func)(void *))
 	{
 	int i;
 
@@ -291,32 +284,32 @@ void sk_pop_free(STACK *st, void (*func)(void *))
 	sk_free(st);
 	}
 
-void sk_free(STACK *st)
+void sk_free(_STACK *st)
 	{
 	if (st == NULL) return;
 	if (st->data != NULL) OPENSSL_free(st->data);
 	OPENSSL_free(st);
 	}
 
-int sk_num(const STACK *st)
+int sk_num(const _STACK *st)
 {
 	if(st == NULL) return -1;
 	return st->num;
 }
 
-char *sk_value(const STACK *st, int i)
+void *sk_value(const _STACK *st, int i)
 {
 	if(!st || (i < 0) || (i >= st->num)) return NULL;
 	return st->data[i];
 }
 
-char *sk_set(STACK *st, int i, char *value)
+void *sk_set(_STACK *st, int i, void *value)
 {
 	if(!st || (i < 0) || (i >= st->num)) return NULL;
 	return (st->data[i] = value);
 }
 
-void sk_sort(STACK *st)
+void sk_sort(_STACK *st)
 	{
 	if (st && !st->sorted)
 		{
@@ -333,7 +326,7 @@ void sk_sort(STACK *st)
 		}
 	}
 
-int sk_is_sorted(const STACK *st)
+int sk_is_sorted(const _STACK *st)
 	{
 	if (!st)
 		return 1;
diff --git a/crypto/stack/stack.h b/crypto/stack/stack.h
index 5cbb116a8b2f..ce35e554eb9e 100644
--- a/crypto/stack/stack.h
+++ b/crypto/stack/stack.h
@@ -70,37 +70,36 @@ typedef struct stack_st
 	int sorted;
 
 	int num_alloc;
-	int (*comp)(const char * const *, const char * const *);
-	} STACK;
+	int (*comp)(const void *, const void *);
+	} _STACK;  /* Use STACK_OF(...) instead */
 
 #define M_sk_num(sk)		((sk) ? (sk)->num:-1)
 #define M_sk_value(sk,n)	((sk) ? (sk)->data[n] : NULL)
 
-int sk_num(const STACK *);
-char *sk_value(const STACK *, int);
+int sk_num(const _STACK *);
+void *sk_value(const _STACK *, int);
 
-char *sk_set(STACK *, int, char *);
+void *sk_set(_STACK *, int, void *);
 
-STACK *sk_new(int (*cmp)(const char * const *, const char * const *));
-STACK *sk_new_null(void);
-void sk_free(STACK *);
-void sk_pop_free(STACK *st, void (*func)(void *));
-int sk_insert(STACK *sk,char *data,int where);
-char *sk_delete(STACK *st,int loc);
-char *sk_delete_ptr(STACK *st, char *p);
-int sk_find(STACK *st,char *data);
-int sk_find_ex(STACK *st,char *data);
-int sk_push(STACK *st,char *data);
-int sk_unshift(STACK *st,char *data);
-char *sk_shift(STACK *st);
-char *sk_pop(STACK *st);
-void sk_zero(STACK *st);
-int (*sk_set_cmp_func(STACK *sk, int (*c)(const char * const *,
-			const char * const *)))
-			(const char * const *, const char * const *);
-STACK *sk_dup(STACK *st);
-void sk_sort(STACK *st);
-int sk_is_sorted(const STACK *st);
+_STACK *sk_new(int (*cmp)(const void *, const void *));
+_STACK *sk_new_null(void);
+void sk_free(_STACK *);
+void sk_pop_free(_STACK *st, void (*func)(void *));
+int sk_insert(_STACK *sk, void *data, int where);
+void *sk_delete(_STACK *st, int loc);
+void *sk_delete_ptr(_STACK *st, void *p);
+int sk_find(_STACK *st, void *data);
+int sk_find_ex(_STACK *st, void *data);
+int sk_push(_STACK *st, void *data);
+int sk_unshift(_STACK *st, void *data);
+void *sk_shift(_STACK *st);
+void *sk_pop(_STACK *st);
+void sk_zero(_STACK *st);
+int (*sk_set_cmp_func(_STACK *sk, int (*c)(const void *, const void *)))
+	(const void *, const void *);
+_STACK *sk_dup(_STACK *st);
+void sk_sort(_STACK *st);
+int sk_is_sorted(const _STACK *st);
 
 #ifdef  __cplusplus
 }
diff --git a/crypto/store/Makefile b/crypto/store/Makefile
index c9f5d001a39b..0dcfd7857a37 100644
--- a/crypto/store/Makefile
+++ b/crypto/store/Makefile
@@ -35,7 +35,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -89,14 +89,14 @@ str_lib.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 str_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 str_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
 str_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-str_lib.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-str_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-str_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-str_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-str_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-str_lib.o: ../../include/openssl/stack.h ../../include/openssl/store.h
-str_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-str_lib.o: ../../include/openssl/x509_vfy.h str_lib.c str_locl.h
+str_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+str_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+str_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+str_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+str_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+str_lib.o: ../../include/openssl/store.h ../../include/openssl/symhacks.h
+str_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+str_lib.o: str_lib.c str_locl.h
 str_mem.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
 str_mem.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
 str_mem.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
diff --git a/crypto/store/store.h b/crypto/store/store.h
index 64583377a94a..0a28c7d5a20e 100644
--- a/crypto/store/store.h
+++ b/crypto/store/store.h
@@ -59,6 +59,12 @@
 #ifndef HEADER_STORE_H
 #define HEADER_STORE_H
 
+#include <openssl/opensslconf.h>
+
+#ifdef OPENSSL_NO_STORE
+#error STORE is disabled.
+#endif
+
 #include <openssl/ossl_typ.h>
 #ifndef OPENSSL_NO_DEPRECATED
 #include <openssl/evp.h>
@@ -408,7 +414,8 @@ int STORE_ATTR_INFO_modify_number(STORE_ATTR_INFO *attrs, STORE_ATTR_TYPES code,
 
 /* Compare on basis of a bit pattern formed by the STORE_ATTR_TYPES values
    in each contained attribute. */
-int STORE_ATTR_INFO_compare(STORE_ATTR_INFO *a, STORE_ATTR_INFO *b);
+int STORE_ATTR_INFO_compare(const STORE_ATTR_INFO * const *a,
+			    const STORE_ATTR_INFO * const *b);
 /* Check if the set of attributes in a is within the range of attributes
    set in b. */
 int STORE_ATTR_INFO_in_range(STORE_ATTR_INFO *a, STORE_ATTR_INFO *b);
diff --git a/crypto/store/str_err.c b/crypto/store/str_err.c
index 6fee6498220a..924edf050581 100644
--- a/crypto/store/str_err.c
+++ b/crypto/store/str_err.c
@@ -1,6 +1,6 @@
 /* crypto/store/str_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
diff --git a/crypto/store/str_lib.c b/crypto/store/str_lib.c
index 32ae5bd3957d..f1dbcbd0e0c7 100644
--- a/crypto/store/str_lib.c
+++ b/crypto/store/str_lib.c
@@ -1670,7 +1670,7 @@ int STORE_parse_attrs_endp(void *handle)
 	}
 
 static int attr_info_compare_compute_range(
-	unsigned char *abits, unsigned char *bbits,
+	const unsigned char *abits, const unsigned char *bbits,
 	unsigned int *alowp, unsigned int *ahighp,
 	unsigned int *blowp, unsigned int *bhighp)
 	{
@@ -1739,13 +1739,15 @@ static int attr_info_compare_compute_range(
 	return res;
 	}
 
-int STORE_ATTR_INFO_compare(STORE_ATTR_INFO *a, STORE_ATTR_INFO *b)
+int STORE_ATTR_INFO_compare(const STORE_ATTR_INFO * const *a,
+			    const STORE_ATTR_INFO * const *b)
 	{
 	if (a == b) return 0;
 	if (!a) return -1;
 	if (!b) return 1;
-	return attr_info_compare_compute_range(a->set, b->set, 0, 0, 0, 0);
+	return attr_info_compare_compute_range((*a)->set, (*b)->set, 0, 0, 0, 0);
 	}
+
 int STORE_ATTR_INFO_in_range(STORE_ATTR_INFO *a, STORE_ATTR_INFO *b)
 	{
 	unsigned int alow, ahigh, blow, bhigh;
@@ -1759,6 +1761,7 @@ int STORE_ATTR_INFO_in_range(STORE_ATTR_INFO *a, STORE_ATTR_INFO *b)
 		return 1;
 	return 0;
 	}
+
 int STORE_ATTR_INFO_in(STORE_ATTR_INFO *a, STORE_ATTR_INFO *b)
 	{
 	unsigned char *abits, *bbits;
@@ -1776,6 +1779,7 @@ int STORE_ATTR_INFO_in(STORE_ATTR_INFO *a, STORE_ATTR_INFO *b)
 		}
 	return 1;
 	}
+
 int STORE_ATTR_INFO_in_ex(STORE_ATTR_INFO *a, STORE_ATTR_INFO *b)
 	{
 	STORE_ATTR_TYPES i;
diff --git a/crypto/store/str_mem.c b/crypto/store/str_mem.c
index 527757ae0990..8ac4f7e55c70 100644
--- a/crypto/store/str_mem.c
+++ b/crypto/store/str_mem.c
@@ -76,30 +76,35 @@
    attribute type code).
 */
 
-struct mem_object_data_st
+typedef struct mem_object_data_st
 	{
 	STORE_OBJECT *object;
 	STORE_ATTR_INFO *attr_info;
 	int references;
-	};
+	} MEM_OBJECT_DATA;
 
+DECLARE_STACK_OF(MEM_OBJECT_DATA)
 struct mem_data_st
 	{
-	STACK *data;		/* A stack of mem_object_data_st,
-				   sorted with STORE_ATTR_INFO_compare(). */
+	STACK_OF(MEM_OBJECT_DATA) *data; /* sorted with
+					  * STORE_ATTR_INFO_compare(). */
 	unsigned int compute_components : 1; /* Currently unused, but can
 						be used to add attributes
 						from parts of the data. */
 	};
 
+DECLARE_STACK_OF(STORE_ATTR_INFO)
 struct mem_ctx_st
 	{
 	int type;		/* The type we're searching for */
-	STACK *search_attributes; /* Sets of attributes to search for.
-				     Each element is a STORE_ATTR_INFO. */
-	int search_index;	/* which of the search attributes we found a match
-				   for, -1 when we still haven't found any */
-	int index;		/* -1 as long as we're searching for the first */
+	STACK_OF(STORE_ATTR_INFO) *search_attributes; /* Sets of
+				     attributes to search for.  Each
+				     element is a STORE_ATTR_INFO. */
+	int search_index;	/* which of the search attributes we
+				   found a match for, -1 when we still
+				   haven't found any */
+	int index;		/* -1 as long as we're searching for
+                                    the first */
 	};
 
 static int mem_init(STORE *s);
@@ -240,7 +245,7 @@ static void *mem_list_start(STORE *s, STORE_OBJECT_TYPES type,
 		if (context->search_attributes == NULL)
 			{
 			context->search_attributes =
-				sk_new((int (*)(const char * const *, const char * const *))STORE_ATTR_INFO_compare);
+				sk_STORE_ATTR_INFO_new(STORE_ATTR_INFO_compare);
 			if (!context->search_attributes)
 				{
 				STOREerr(STORE_F_MEM_LIST_START,
@@ -248,7 +253,7 @@ static void *mem_list_start(STORE *s, STORE_OBJECT_TYPES type,
 				goto err;
 				}
 			}
-		sk_push(context->search_attributes,(char *)attrs);
+		sk_STORE_ATTR_INFO_push(context->search_attributes,attrs);
 		}
 	if (!STORE_parse_attrs_endp(attribute_context))
 		goto err;
@@ -284,11 +289,14 @@ static STORE_OBJECT *mem_list_next(STORE *s, void *handle)
 
 	if (context->search_index == -1)
 		{
-		for (i = 0; i < sk_num(context->search_attributes); i++)
+		for (i = 0;
+		     i < sk_STORE_ATTR_INFO_num(context->search_attributes);
+		     i++)
 			{
-			key.attr_info =
-				(STORE_ATTR_INFO *)sk_value(context->search_attributes, i);
-			srch = sk_find_ex(store->data, (char *)&key);
+			key.attr_info
+			  = sk_STORE_ATTR_INFO_value(context->search_attributes,
+						     i);
+			srch = sk_MEM_OBJECT_DATA_find_ex(store->data, &key);
 
 			if (srch >= 0)
 				{
@@ -301,21 +309,20 @@ static STORE_OBJECT *mem_list_next(STORE *s, void *handle)
 		return NULL;
 	
 	key.attr_info =
-		(STORE_ATTR_INFO *)sk_value(context->search_attributes,
-			context->search_index);
+		sk_STORE_ATTR_INFO_value(context->search_attributes,
+					 context->search_index);
 	for(srch = context->search_index;
-	    srch < sk_num(store->data)
+	    srch < sk_MEM_OBJECT_DATA_num(store->data)
 		    && STORE_ATTR_INFO_in_range(key.attr_info,
-			    (STORE_ATTR_INFO *)sk_value(store->data, srch))
+			    sk_MEM_OBJECT_DATA_value(store->data, srch)->attr_info)
 		    && !(cres = STORE_ATTR_INFO_in_ex(key.attr_info,
-				 (STORE_ATTR_INFO *)sk_value(store->data, srch)));
+				 sk_MEM_OBJECT_DATA_value(store->data, srch)->attr_info));
 	    srch++)
 		;
 
 	context->search_index = srch;
 	if (cres)
-		return ((struct mem_object_data_st *)sk_value(store->data,
-				srch))->object;
+		return (sk_MEM_OBJECT_DATA_value(store->data, srch))->object;
 	return NULL;
 	}
 static int mem_list_end(STORE *s, void *handle)
@@ -328,7 +335,7 @@ static int mem_list_end(STORE *s, void *handle)
 		return 0;
 		}
 	if (context && context->search_attributes)
-		sk_free(context->search_attributes);
+		sk_STORE_ATTR_INFO_free(context->search_attributes);
 	if (context) OPENSSL_free(context);
 	return 1;
 	}
@@ -337,7 +344,8 @@ static int mem_list_endp(STORE *s, void *handle)
 	struct mem_ctx_st *context = (struct mem_ctx_st *)handle;
 
 	if (!context
-		|| context->search_index == sk_num(context->search_attributes))
+	    || context->search_index
+	       == sk_STORE_ATTR_INFO_num(context->search_attributes))
 		return 1;
 	return 0;
 	}
diff --git a/crypto/symhacks.h b/crypto/symhacks.h
index 0114093c1ee7..403f592dcd69 100644
--- a/crypto/symhacks.h
+++ b/crypto/symhacks.h
@@ -67,10 +67,6 @@
    incompatibilities. */
 #ifdef OPENSSL_SYS_VMS
 
-/* Hack a long name in crypto/cryptlib.c */
-#undef int_CRYPTO_set_do_dynlock_callback
-#define int_CRYPTO_set_do_dynlock_callback	int_CRYPTO_set_do_dynlock_cb
-
 /* Hack a long name in crypto/ex_data.c */
 #undef CRYPTO_get_ex_data_implementation
 #define CRYPTO_get_ex_data_implementation	CRYPTO_get_ex_data_impl
@@ -151,9 +147,9 @@
 #undef CRYPTO_set_dynlock_create_callback
 #define CRYPTO_set_dynlock_create_callback      CRYPTO_set_dynlock_create_cb
 #undef CRYPTO_set_dynlock_lock_callback
-#define CRYPTO_set_dynlock_lock_callback        CRYPTO_set_dynlock_lock_cb
+#define CRYPTO_set_dynlock_lock_callback	CRYPTO_set_dynlock_lock_cb
 #undef CRYPTO_get_dynlock_lock_callback
-#define CRYPTO_get_dynlock_lock_callback        CRYPTO_get_dynlock_lock_cb
+#define CRYPTO_get_dynlock_lock_callback	CRYPTO_get_dynlock_lock_cb
 #undef CRYPTO_get_dynlock_destroy_callback
 #define CRYPTO_get_dynlock_destroy_callback     CRYPTO_get_dynlock_destroy_cb
 #undef CRYPTO_get_dynlock_create_callback
@@ -165,7 +161,7 @@
 
 /* Hack some long SSL names */
 #undef SSL_CTX_set_default_verify_paths
-#define SSL_CTX_set_default_verify_paths        SSL_CTX_set_def_verify_paths
+#define SSL_CTX_set_default_verify_paths	SSL_CTX_set_def_verify_paths
 #undef SSL_get_ex_data_X509_STORE_CTX_idx
 #define SSL_get_ex_data_X509_STORE_CTX_idx      SSL_get_ex_d_X509_STORE_CTX_idx
 #undef SSL_add_file_cert_subjects_to_stack
@@ -175,12 +171,11 @@
 #undef SSL_CTX_use_certificate_chain_file
 #define SSL_CTX_use_certificate_chain_file      SSL_CTX_use_cert_chain_file
 #undef SSL_CTX_set_cert_verify_callback
-#define SSL_CTX_set_cert_verify_callback        SSL_CTX_set_cert_verify_cb
+#define SSL_CTX_set_cert_verify_callback	SSL_CTX_set_cert_verify_cb
 #undef SSL_CTX_set_default_passwd_cb_userdata
 #define SSL_CTX_set_default_passwd_cb_userdata  SSL_CTX_set_def_passwd_cb_ud
 #undef SSL_COMP_get_compression_methods
 #define SSL_COMP_get_compression_methods	SSL_COMP_get_compress_methods
-
 #undef ssl_add_clienthello_renegotiate_ext
 #define ssl_add_clienthello_renegotiate_ext	ssl_add_clienthello_reneg_ext
 #undef ssl_add_serverhello_renegotiate_ext
@@ -189,6 +184,26 @@
 #define ssl_parse_clienthello_renegotiate_ext	ssl_parse_clienthello_reneg_ext
 #undef ssl_parse_serverhello_renegotiate_ext
 #define ssl_parse_serverhello_renegotiate_ext	ssl_parse_serverhello_reneg_ext
+#undef SSL_srp_server_param_with_username
+#define SSL_srp_server_param_with_username	SSL_srp_server_param_with_un
+#undef SSL_CTX_set_srp_client_pwd_callback
+#define SSL_CTX_set_srp_client_pwd_callback	SSL_CTX_set_srp_client_pwd_cb
+#undef SSL_CTX_set_srp_verify_param_callback
+#define SSL_CTX_set_srp_verify_param_callback	SSL_CTX_set_srp_vfy_param_cb
+#undef SSL_CTX_set_srp_username_callback
+#define SSL_CTX_set_srp_username_callback	SSL_CTX_set_srp_un_cb
+#undef ssl_add_clienthello_use_srtp_ext
+#define ssl_add_clienthello_use_srtp_ext ssl_add_clihello_use_srtp_ext
+#undef ssl_add_serverhello_use_srtp_ext
+#define ssl_add_serverhello_use_srtp_ext ssl_add_serhello_use_srtp_ext
+#undef ssl_parse_clienthello_use_srtp_ext
+#define ssl_parse_clienthello_use_srtp_ext ssl_parse_clihello_use_srtp_ext
+#undef ssl_parse_serverhello_use_srtp_ext
+#define ssl_parse_serverhello_use_srtp_ext ssl_parse_serhello_use_srtp_ext
+#undef SSL_CTX_set_next_protos_advertised_cb
+#define SSL_CTX_set_next_protos_advertised_cb SSL_CTX_set_next_protos_adv_cb
+#undef SSL_CTX_set_next_proto_select_cb
+#define SSL_CTX_set_next_proto_select_cb SSL_CTX_set_next_proto_sel_cb
 
 /* Hack some long ENGINE names */
 #undef ENGINE_get_default_BN_mod_exp_crt
@@ -196,9 +211,17 @@
 #undef ENGINE_set_default_BN_mod_exp_crt
 #define ENGINE_set_default_BN_mod_exp_crt	ENGINE_set_def_BN_mod_exp_crt
 #undef ENGINE_set_load_privkey_function
-#define ENGINE_set_load_privkey_function        ENGINE_set_load_privkey_fn
+#define ENGINE_set_load_privkey_function	ENGINE_set_load_privkey_fn
 #undef ENGINE_get_load_privkey_function
-#define ENGINE_get_load_privkey_function        ENGINE_get_load_privkey_fn
+#define ENGINE_get_load_privkey_function	ENGINE_get_load_privkey_fn
+#undef ENGINE_unregister_pkey_asn1_meths
+#define ENGINE_unregister_pkey_asn1_meths	ENGINE_unreg_pkey_asn1_meths
+#undef ENGINE_register_all_pkey_asn1_meths
+#define ENGINE_register_all_pkey_asn1_meths	ENGINE_reg_all_pkey_asn1_meths
+#undef ENGINE_set_default_pkey_asn1_meths
+#define ENGINE_set_default_pkey_asn1_meths	ENGINE_set_def_pkey_asn1_meths
+#undef ENGINE_get_pkey_asn1_meth_engine
+#define ENGINE_get_pkey_asn1_meth_engine	ENGINE_get_pkey_asn1_meth_eng
 #undef ENGINE_set_load_ssl_client_cert_function
 #define ENGINE_set_load_ssl_client_cert_function \
 						ENGINE_set_ld_ssl_clnt_cert_fn
@@ -207,7 +230,7 @@
 
 /* Hack some long OCSP names */
 #undef OCSP_REQUEST_get_ext_by_critical
-#define OCSP_REQUEST_get_ext_by_critical        OCSP_REQUEST_get_ext_by_crit
+#define OCSP_REQUEST_get_ext_by_critical	OCSP_REQUEST_get_ext_by_crit
 #undef OCSP_BASICRESP_get_ext_by_critical
 #define OCSP_BASICRESP_get_ext_by_critical      OCSP_BASICRESP_get_ext_by_crit
 #undef OCSP_SINGLERESP_get_ext_by_critical
@@ -224,6 +247,8 @@
 #define OPENSSL_add_all_algorithms_noconf	OPENSSL_add_all_algo_noconf
 #undef OPENSSL_add_all_algorithms_conf
 #define OPENSSL_add_all_algorithms_conf		OPENSSL_add_all_algo_conf
+#undef EVP_PKEY_meth_set_verify_recover
+#define EVP_PKEY_meth_set_verify_recover	EVP_PKEY_meth_set_vrfy_recover
 
 /* Hack some long EC names */
 #undef EC_GROUP_set_point_conversion_form
@@ -232,6 +257,9 @@
 #define EC_GROUP_get_point_conversion_form	EC_GROUP_get_point_conv_form
 #undef EC_GROUP_clear_free_all_extra_data
 #define EC_GROUP_clear_free_all_extra_data	EC_GROUP_clr_free_all_xtra_data
+#undef EC_KEY_set_public_key_affine_coordinates
+#define EC_KEY_set_public_key_affine_coordinates \
+						EC_KEY_set_pub_key_aff_coords
 #undef EC_POINT_set_Jprojective_coordinates_GFp
 #define EC_POINT_set_Jprojective_coordinates_GFp \
                                                 EC_POINT_set_Jproj_coords_GFp
@@ -252,15 +280,15 @@
 #define EC_POINT_set_compressed_coordinates_GF2m \
                                                 EC_POINT_set_compr_coords_GF2m
 #undef ec_GF2m_simple_group_clear_finish
-#define ec_GF2m_simple_group_clear_finish        ec_GF2m_simple_grp_clr_finish
+#define ec_GF2m_simple_group_clear_finish	ec_GF2m_simple_grp_clr_finish
 #undef ec_GF2m_simple_group_check_discriminant
 #define ec_GF2m_simple_group_check_discriminant	ec_GF2m_simple_grp_chk_discrim
 #undef ec_GF2m_simple_point_clear_finish
-#define ec_GF2m_simple_point_clear_finish        ec_GF2m_simple_pt_clr_finish
+#define ec_GF2m_simple_point_clear_finish	ec_GF2m_simple_pt_clr_finish
 #undef ec_GF2m_simple_point_set_to_infinity
-#define ec_GF2m_simple_point_set_to_infinity     ec_GF2m_simple_pt_set_to_inf
+#define ec_GF2m_simple_point_set_to_infinity	ec_GF2m_simple_pt_set_to_inf
 #undef ec_GF2m_simple_points_make_affine
-#define ec_GF2m_simple_points_make_affine        ec_GF2m_simple_pts_make_affine
+#define ec_GF2m_simple_points_make_affine	ec_GF2m_simple_pts_make_affine
 #undef ec_GF2m_simple_point_set_affine_coordinates
 #define ec_GF2m_simple_point_set_affine_coordinates \
                                                 ec_GF2m_smp_pt_set_af_coords
@@ -275,19 +303,19 @@
 #undef ec_GFp_simple_group_get_curve_GFp
 #define ec_GFp_simple_group_get_curve_GFp       ec_GFp_simple_grp_get_curve_GFp
 #undef ec_GFp_simple_group_clear_finish
-#define ec_GFp_simple_group_clear_finish        ec_GFp_simple_grp_clear_finish
+#define ec_GFp_simple_group_clear_finish	ec_GFp_simple_grp_clear_finish
 #undef ec_GFp_simple_group_set_generator
 #define ec_GFp_simple_group_set_generator       ec_GFp_simple_grp_set_generator
 #undef ec_GFp_simple_group_get0_generator
 #define ec_GFp_simple_group_get0_generator      ec_GFp_simple_grp_gt0_generator
 #undef ec_GFp_simple_group_get_cofactor
-#define ec_GFp_simple_group_get_cofactor        ec_GFp_simple_grp_get_cofactor
+#define ec_GFp_simple_group_get_cofactor	ec_GFp_simple_grp_get_cofactor
 #undef ec_GFp_simple_point_clear_finish
-#define ec_GFp_simple_point_clear_finish        ec_GFp_simple_pt_clear_finish
+#define ec_GFp_simple_point_clear_finish	ec_GFp_simple_pt_clear_finish
 #undef ec_GFp_simple_point_set_to_infinity
 #define ec_GFp_simple_point_set_to_infinity     ec_GFp_simple_pt_set_to_inf
 #undef ec_GFp_simple_points_make_affine
-#define ec_GFp_simple_points_make_affine        ec_GFp_simple_pts_make_affine
+#define ec_GFp_simple_points_make_affine	ec_GFp_simple_pts_make_affine
 #undef ec_GFp_simple_group_get_curve_GFp
 #define ec_GFp_simple_group_get_curve_GFp       ec_GFp_simple_grp_get_curve_GFp
 #undef ec_GFp_simple_set_Jprojective_coordinates_GFp
@@ -367,6 +395,14 @@
 #undef STORE_method_get_unlock_store_function
 #define STORE_method_get_unlock_store_function	STORE_meth_get_unlock_store_fn
 
+/* Hack some long TS names */
+#undef TS_RESP_CTX_set_status_info_cond
+#define TS_RESP_CTX_set_status_info_cond	TS_RESP_CTX_set_stat_info_cond
+#undef TS_RESP_CTX_set_clock_precision_digits
+#define TS_RESP_CTX_set_clock_precision_digits	TS_RESP_CTX_set_clk_prec_digits
+#undef TS_CONF_set_clock_precision_digits
+#define TS_CONF_set_clock_precision_digits	TS_CONF_set_clk_prec_digits
+
 /* Hack some long CMS names */
 #undef CMS_RecipientInfo_ktri_get0_algs
 #define CMS_RecipientInfo_ktri_get0_algs	CMS_RecipInfo_ktri_get0_algs
@@ -385,24 +421,36 @@
 #undef dtls1_retransmit_buffered_messages
 #define dtls1_retransmit_buffered_messages	dtls1_retransmit_buffered_msgs
 
+/* Hack some long SRP names */
+#undef SRP_generate_server_master_secret
+#define SRP_generate_server_master_secret	SRP_gen_server_master_secret
+#undef SRP_generate_client_master_secret
+#define SRP_generate_client_master_secret	SRP_gen_client_master_secret
+
+/* Hack some long UI names */
+#undef UI_method_get_prompt_constructor
+#define UI_method_get_prompt_constructor	UI_method_get_prompt_constructr
+#undef UI_method_set_prompt_constructor
+#define UI_method_set_prompt_constructor	UI_method_set_prompt_constructr
+
 #endif /* defined OPENSSL_SYS_VMS */
 
 
-/* Case insensiteve linking causes problems.... */
-#if defined(OPENSSL_SYS_WIN16) || defined(OPENSSL_SYS_VMS) || defined(OPENSSL_SYS_OS2)
+/* Case insensitive linking causes problems.... */
+#if defined(OPENSSL_SYS_VMS) || defined(OPENSSL_SYS_OS2)
 #undef ERR_load_CRYPTO_strings
 #define ERR_load_CRYPTO_strings			ERR_load_CRYPTOlib_strings
 #undef OCSP_crlID_new
-#define OCSP_crlID_new                          OCSP_crlID2_new
+#define OCSP_crlID_new				OCSP_crlID2_new
 
 #undef d2i_ECPARAMETERS
-#define d2i_ECPARAMETERS                        d2i_UC_ECPARAMETERS
+#define d2i_ECPARAMETERS			d2i_UC_ECPARAMETERS
 #undef i2d_ECPARAMETERS
-#define i2d_ECPARAMETERS                        i2d_UC_ECPARAMETERS
+#define i2d_ECPARAMETERS			i2d_UC_ECPARAMETERS
 #undef d2i_ECPKPARAMETERS
-#define d2i_ECPKPARAMETERS                      d2i_UC_ECPKPARAMETERS
+#define d2i_ECPKPARAMETERS			d2i_UC_ECPKPARAMETERS
 #undef i2d_ECPKPARAMETERS
-#define i2d_ECPKPARAMETERS                      i2d_UC_ECPKPARAMETERS
+#define i2d_ECPKPARAMETERS			i2d_UC_ECPKPARAMETERS
 
 /* These functions do not seem to exist!  However, I'm paranoid...
    Original command in x509v3.h:
@@ -411,19 +459,19 @@
    hide them a little, by giving them an extra 'o' at the
    beginning of the name... */
 #undef X509v3_cleanup_extensions
-#define X509v3_cleanup_extensions               oX509v3_cleanup_extensions
+#define X509v3_cleanup_extensions		oX509v3_cleanup_extensions
 #undef X509v3_add_extension
-#define X509v3_add_extension                    oX509v3_add_extension
+#define X509v3_add_extension			oX509v3_add_extension
 #undef X509v3_add_netscape_extensions
-#define X509v3_add_netscape_extensions          oX509v3_add_netscape_extensions
+#define X509v3_add_netscape_extensions		oX509v3_add_netscape_extensions
 #undef X509v3_add_standard_extensions
-#define X509v3_add_standard_extensions          oX509v3_add_standard_extensions
+#define X509v3_add_standard_extensions		oX509v3_add_standard_extensions
 
+/* This one clashes with CMS_data_create */
+#undef cms_Data_create
+#define cms_Data_create				priv_cms_Data_create
 
 #endif
 
 
 #endif /* ! defined HEADER_VMS_IDHACKS_H */
-/* This one clashes with CMS_data_create */
-#undef cms_Data_create
-#define cms_Data_create				priv_cms_Data_create
diff --git a/crypto/threads/mttest.c b/crypto/threads/mttest.c
index f6f3df4b6aa8..eba7aa8a6ebc 100644
--- a/crypto/threads/mttest.c
+++ b/crypto/threads/mttest.c
@@ -117,11 +117,13 @@ void solaris_locking_callback(int mode,int type,char *file,int line);
 void win32_locking_callback(int mode,int type,char *file,int line);
 void pthreads_locking_callback(int mode,int type,char *file,int line);
 void netware_locking_callback(int mode,int type,char *file,int line);
+void beos_locking_callback(int mode,int type,const char *file,int line);
 
 unsigned long irix_thread_id(void );
 unsigned long solaris_thread_id(void );
 unsigned long pthreads_thread_id(void );
 unsigned long netware_thread_id(void );
+unsigned long beos_thread_id(void );
 
 #if defined(OPENSSL_SYS_NETWARE)
 static MPKMutex *lock_cs;
@@ -1209,3 +1211,100 @@ unsigned long netware_thread_id(void)
    return(ret);
 }
 #endif /* NETWARE */
+
+#ifdef BEOS_THREADS
+
+#include <Locker.h>
+
+static BLocker** lock_cs;
+static long* lock_count;
+
+void thread_setup(void)
+	{
+	int i;
+
+	lock_cs=(BLocker**)OPENSSL_malloc(CRYPTO_num_locks() * sizeof(BLocker*));
+	lock_count=(long*)OPENSSL_malloc(CRYPTO_num_locks() * sizeof(long));
+	for (i=0; i<CRYPTO_num_locks(); i++)
+		{
+		lock_count[i]=0;
+		lock_cs[i] = new BLocker(CRYPTO_get_lock_name(i));
+		}
+
+	CRYPTO_set_id_callback((unsigned long (*)())beos_thread_id);
+	CRYPTO_set_locking_callback(beos_locking_callback);
+	}
+
+void thread_cleanup(void)
+	{
+	int i;
+
+	CRYPTO_set_locking_callback(NULL);
+	fprintf(stderr,"cleanup\n");
+	for (i=0; i<CRYPTO_num_locks(); i++)
+		{
+		delete lock_cs[i];
+		fprintf(stderr,"%8ld:%s\n",lock_count[i],
+			CRYPTO_get_lock_name(i));
+		}
+	OPENSSL_free(lock_cs);
+	OPENSSL_free(lock_count);
+
+	fprintf(stderr,"done cleanup\n");
+	}
+
+void beos_locking_callback(int mode, int type, const char *file, int line)
+    {
+#if 0
+	fprintf(stderr,"thread=%4d mode=%s lock=%s %s:%d\n",
+		CRYPTO_thread_id(),
+		(mode&CRYPTO_LOCK)?"l":"u",
+		(type&CRYPTO_READ)?"r":"w",file,line);
+#endif
+	if (mode & CRYPTO_LOCK)
+		{
+		lock_cs[type]->Lock();
+		lock_count[type]++;
+		}
+	else
+		{
+		lock_cs[type]->Unlock();
+		}
+	}
+
+void do_threads(SSL_CTX *s_ctx, SSL_CTX *c_ctx)
+	{
+	SSL_CTX *ssl_ctx[2];
+	thread_id thread_ctx[MAX_THREAD_NUMBER];
+	int i;
+
+	ssl_ctx[0]=s_ctx;
+	ssl_ctx[1]=c_ctx;
+
+	for (i=0; i<thread_number; i++)
+		{
+		thread_ctx[i] = spawn_thread((thread_func)ndoit,
+			NULL, B_NORMAL_PRIORITY, (void *)ssl_ctx);
+		resume_thread(thread_ctx[i]);
+		}
+
+	printf("waiting...\n");
+	for (i=0; i<thread_number; i++)
+		{
+		status_t result;
+		wait_for_thread(thread_ctx[i], &result);
+		}
+
+	printf("beos threads done (%d,%d)\n",
+		s_ctx->references,c_ctx->references);
+	}
+
+unsigned long beos_thread_id(void)
+	{
+	unsigned long ret;
+
+	ret=(unsigned long)find_thread(NULL);
+	return(ret);
+	}
+
+#endif /* BEOS_THREADS */
diff --git a/crypto/tmdiff.c b/crypto/tmdiff.c
deleted file mode 100644
index 1c6e052ac98d..000000000000
--- a/crypto/tmdiff.c
+++ /dev/null
@@ -1,260 +0,0 @@
-/* crypto/tmdiff.c */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-#include <stdio.h>
-#include <stdlib.h>
-#include "cryptlib.h"
-#include <openssl/tmdiff.h>
-#if !defined(OPENSSL_SYS_MSDOS)
-#include OPENSSL_UNISTD
-#endif
-
-#ifdef TIMEB
-#undef OPENSSL_SYS_WIN32
-#undef TIMES
-#endif
-
-#if !defined(OPENSSL_SYS_MSDOS) && !defined(OPENSSL_SYS_WIN32) && !(defined(OPENSSL_SYS_VMS) || defined(__DECC)) && !defined(OPENSSL_SYS_MACOSX_RHAPSODY) && !defined(OPENSSL_SYS_VXWORKS)
-# define TIMES
-#endif
-
-#ifdef OPENSSL_SYS_NETWARE
-#undef TIMES
-#endif
-
-#if !defined(_IRIX) || defined (OPENSSL_SYS_NETWARE)
-#  include <time.h>
-#endif
-#ifdef TIMES
-#  include <sys/types.h>
-#  include <sys/times.h>
-#endif
-
-/* Depending on the VMS version, the tms structure is perhaps defined.
-   The __TMS macro will show if it was.  If it wasn't defined, we should
-   undefine TIMES, since that tells the rest of the program how things
-   should be handled.				-- Richard Levitte */
-#if defined(OPENSSL_SYS_VMS_DECC) && !defined(__TMS)
-#undef TIMES
-#endif
-
-#if defined(sun) || defined(__ultrix)
-#define _POSIX_SOURCE
-#include <limits.h>
-#include <sys/param.h>
-#endif
-
-#if !defined(TIMES) && !defined(OPENSSL_SYS_VXWORKS) && !defined(OPENSSL_SYS_NETWARE)
-#include <sys/timeb.h>
-#endif
-
-#ifdef OPENSSL_SYS_WIN32
-#include <windows.h>
-#endif
-
-/* The following if from times(3) man page.  It may need to be changed */
-#ifndef HZ
-# if defined(_SC_CLK_TCK) \
-     && (!defined(OPENSSL_SYS_VMS) || __CTRL_VER >= 70000000)
-/* #  define HZ ((double)sysconf(_SC_CLK_TCK)) */
-#  define HZ sysconf(_SC_CLK_TCK)
-# else
-#  ifndef CLK_TCK
-#   ifndef _BSD_CLK_TCK_ /* FreeBSD hack */
-#    define HZ  100.0
-#   else /* _BSD_CLK_TCK_ */
-#    define HZ ((double)_BSD_CLK_TCK_)
-#   endif
-#  else /* CLK_TCK */
-#   define HZ ((double)CLK_TCK)
-#  endif
-# endif
-#endif
-
-struct ms_tm
-	{
-#ifdef TIMES
-	struct tms ms_tms;
-#else
-#  ifdef OPENSSL_SYS_WIN32
-	HANDLE thread_id;
-	FILETIME ms_win32;
-#  elif defined (OPENSSL_SYS_NETWARE)
-   clock_t ms_clock;
-#  else
-#    ifdef OPENSSL_SYS_VXWORKS
-          unsigned long ticks;
-#    else
-	struct timeb ms_timeb;
-#    endif
-#  endif
-#endif
-	};
-
-MS_TM *ms_time_new(void)
-	{
-	MS_TM *ret;
-
-	ret=(MS_TM *)OPENSSL_malloc(sizeof(MS_TM));
-	if (ret == NULL)
-		return(NULL);
-	memset(ret,0,sizeof(MS_TM));
-#ifdef OPENSSL_SYS_WIN32
-	ret->thread_id=GetCurrentThread();
-#endif
-	return ret;
-	}
-
-void ms_time_free(MS_TM *a)
-	{
-	if (a != NULL)
-		OPENSSL_free(a);
-	}
-
-void ms_time_get(MS_TM *tm)
-	{
-#ifdef OPENSSL_SYS_WIN32
-	FILETIME tmpa,tmpb,tmpc;
-#endif
-
-#ifdef TIMES
-	times(&tm->ms_tms);
-#else
-#  ifdef OPENSSL_SYS_WIN32
-	GetThreadTimes(tm->thread_id,&tmpa,&tmpb,&tmpc,&(tm->ms_win32));
-#  elif defined (OPENSSL_SYS_NETWARE)
-   tm->ms_clock = clock();
-#  else
-#    ifdef OPENSSL_SYS_VXWORKS
-        tm->ticks = tickGet();
-#    else
-	ftime(&tm->ms_timeb);
-#    endif
-#  endif
-#endif
-	}
-
-double ms_time_diff(MS_TM *a, MS_TM *b)
-	{
-	double ret;
-
-#ifdef TIMES
-	ret = HZ;
-	ret = (b->ms_tms.tms_utime-a->ms_tms.tms_utime) / ret;
-#else
-# ifdef OPENSSL_SYS_WIN32
-	{
-#ifdef __GNUC__
-	signed long long la,lb;
-#else
-	signed _int64 la,lb;
-#endif
-	la=a->ms_win32.dwHighDateTime;
-	lb=b->ms_win32.dwHighDateTime;
-	la<<=32;
-	lb<<=32;
-	la+=a->ms_win32.dwLowDateTime;
-	lb+=b->ms_win32.dwLowDateTime;
-	ret=((double)(lb-la))/1e7;
-	}
-# elif defined (OPENSSL_SYS_NETWARE)
-    ret= (double)(b->ms_clock - a->ms_clock);
-# else
-#  ifdef OPENSSL_SYS_VXWORKS
-        ret = (double)(b->ticks - a->ticks) / (double)sysClkRateGet();
-#  else
-	ret=	 (double)(b->ms_timeb.time-a->ms_timeb.time)+
-		(((double)b->ms_timeb.millitm)-
-		((double)a->ms_timeb.millitm))/1000.0;
-#  endif
-# endif
-#endif
-	return((ret < 0.0000001)?0.0000001:ret);
-	}
-
-int ms_time_cmp(const MS_TM *a, const MS_TM *b)
-	{
-	double d;
-	int ret;
-
-#ifdef TIMES
-	d = HZ;
-	d = (b->ms_tms.tms_utime-a->ms_tms.tms_utime) / d;
-#else
-# ifdef OPENSSL_SYS_WIN32
-	d =(b->ms_win32.dwHighDateTime&0x000fffff)*10+b->ms_win32.dwLowDateTime/1e7;
-	d-=(a->ms_win32.dwHighDateTime&0x000fffff)*10+a->ms_win32.dwLowDateTime/1e7;
-# elif defined (OPENSSL_SYS_NETWARE)
-    d= (double)(b->ms_clock - a->ms_clock);
-# else
-#  ifdef OPENSSL_SYS_VXWORKS
-        d = (b->ticks - a->ticks);
-#  else
-	d=	 (double)(b->ms_timeb.time-a->ms_timeb.time)+
-		(((double)b->ms_timeb.millitm)-(double)a->ms_timeb.millitm)/1000.0;
-#  endif
-# endif
-#endif
-	if (d == 0.0)
-		ret=0;
-	else if (d < 0)
-		ret= -1;
-	else
-		ret=1;
-	return(ret);
-	}
-
diff --git a/crypto/tmdiff.h b/crypto/tmdiff.h
deleted file mode 100644
index af5c41c64992..000000000000
--- a/crypto/tmdiff.h
+++ /dev/null
@@ -1,93 +0,0 @@
-/* crypto/tmdiff.h */
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- * 
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- * 
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from 
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- * 
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- * 
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.]
- */
-
-/* Header for dynamic hash table routines
- * Author - Eric Young
- */
-/* ... erm yeah, "dynamic hash tables" you say?
- * 
- * And what would dynamic hash tables have to do with any of this code *now*?
- * AFAICS, this code is only referenced by crypto/bn/exp.c which is an unused
- * file that I doubt compiles any more. speed.c is the only thing that could
- * use this (and it has nothing to do with hash tables), yet it instead has its
- * own duplication of all this stuff and looks, if anything, more complete. See
- * the corresponding note in apps/speed.c.
- * The Bemused - Geoff
- */
-
-#ifndef HEADER_TMDIFF_H
-#define HEADER_TMDIFF_H
-
-#ifdef  __cplusplus
-extern "C" {
-#endif
-
-typedef struct ms_tm MS_TM;
-
-MS_TM *ms_time_new(void );
-void ms_time_free(MS_TM *a);
-void ms_time_get(MS_TM *a);
-double ms_time_diff(MS_TM *start, MS_TM *end);
-int ms_time_cmp(const MS_TM *ap, const MS_TM *bp);
-
-#ifdef  __cplusplus
-}
-#endif
-
-#endif
-
diff --git a/crypto/ts/Makefile b/crypto/ts/Makefile
new file mode 100644
index 000000000000..c18234555be3
--- /dev/null
+++ b/crypto/ts/Makefile
@@ -0,0 +1,269 @@
+#
+# SSLeay/crypto/ts/Makefile
+#
+
+DIR=	ts
+TOP=	../..
+CC=	cc
+INCLUDES= -I.. -I../../include
+CFLAG = -g
+INSTALL_PREFIX=
+OPENSSLDIR=     /usr/local/ssl
+INSTALLTOP=/usr/local/ssl
+MAKEDEPPROG=	makedepend
+MAKEDEPEND=	$(TOP)/util/domd $(TOP) -MD $(MAKEDEPPROG)
+MAKEFILE=	Makefile
+AR=		ar r
+
+PEX_LIBS=
+EX_LIBS=
+ 
+CFLAGS= $(INCLUDES) $(CFLAG)
+
+GENERAL= Makefile
+TEST=
+APPS=
+
+LIB=$(TOP)/libcrypto.a
+LIBSRC=	ts_err.c ts_req_utils.c ts_req_print.c ts_rsp_utils.c ts_rsp_print.c \
+	ts_rsp_sign.c ts_rsp_verify.c ts_verify_ctx.c ts_lib.c ts_conf.c \
+	ts_asn1.c
+LIBOBJ= ts_err.o ts_req_utils.o ts_req_print.o ts_rsp_utils.o ts_rsp_print.o \
+	ts_rsp_sign.o ts_rsp_verify.o ts_verify_ctx.o ts_lib.o ts_conf.o \
+	ts_asn1.o
+
+SRC= $(LIBSRC)
+
+EXHEADER= ts.h
+HEADER=	$(EXHEADER)
+
+ALL=    $(GENERAL) $(SRC) $(HEADER)
+
+top:
+	(cd ../..; $(MAKE) DIRS=crypto SDIRS=$(DIR) sub_all)
+
+test:
+
+all:	lib
+
+lib:	$(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
+	$(RANLIB) $(LIB) || echo Never mind.
+	@touch lib
+
+files:
+	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
+
+links:
+	@$(PERL) $(TOP)/util/mklink.pl ../../include/openssl $(EXHEADER)
+	@$(PERL) $(TOP)/util/mklink.pl ../../test $(TEST)
+	@$(PERL) $(TOP)/util/mklink.pl ../../apps $(APPS)
+
+install:
+	@[ -n "$(INSTALLTOP)" ] # should be set by top Makefile...
+	@headerlist="$(EXHEADER)"; for i in $$headerlist ; \
+	do  \
+	(cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i; \
+	chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i ); \
+	done;
+
+tags:
+	ctags $(SRC)
+
+lint:
+	lint -DLINT $(INCLUDES) $(SRC)>fluff
+
+depend:
+	$(MAKEDEPEND) -- $(CFLAG) $(INCLUDES) $(DEPFLAG) -- $(LIBSRC)
+
+dclean:
+	$(PERL) -pe 'if (/^# DO NOT DELETE THIS LINE/) {print; exit(0);}' $(MAKEFILE) >Makefile.new
+	mv -f Makefile.new $(MAKEFILE)
+
+clean:
+	rm -f *.o *.obj lib tags core .pure .nfs* *.old *.bak fluff enc dec sign verify
+
+# DO NOT DELETE THIS LINE -- make depend depends on it.
+
+ts_asn1.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
+ts_asn1.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+ts_asn1.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+ts_asn1.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
+ts_asn1.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ts_asn1.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ts_asn1.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+ts_asn1.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ts_asn1.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+ts_asn1.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ts_asn1.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+ts_asn1.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+ts_asn1.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+ts_asn1.o: ../../include/openssl/ts.h ../../include/openssl/x509.h
+ts_asn1.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+ts_asn1.o: ts_asn1.c
+ts_conf.o: ../../e_os.h ../../include/openssl/asn1.h
+ts_conf.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+ts_conf.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+ts_conf.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
+ts_conf.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ts_conf.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ts_conf.o: ../../include/openssl/engine.h ../../include/openssl/err.h
+ts_conf.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ts_conf.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ts_conf.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+ts_conf.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
+ts_conf.o: ../../include/openssl/pem2.h ../../include/openssl/pkcs7.h
+ts_conf.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+ts_conf.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ts_conf.o: ../../include/openssl/symhacks.h ../../include/openssl/ts.h
+ts_conf.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ts_conf.o: ../../include/openssl/x509v3.h ../cryptlib.h ts_conf.c
+ts_err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+ts_err.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
+ts_err.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
+ts_err.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
+ts_err.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+ts_err.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+ts_err.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ts_err.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ts_err.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+ts_err.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ts_err.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+ts_err.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ts_err.o: ../../include/openssl/symhacks.h ../../include/openssl/ts.h
+ts_err.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ts_err.o: ../../include/openssl/x509v3.h ts_err.c
+ts_lib.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
+ts_lib.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
+ts_lib.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+ts_lib.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
+ts_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ts_lib.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ts_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+ts_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ts_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+ts_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+ts_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
+ts_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+ts_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+ts_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ts_lib.o: ../../include/openssl/x509v3.h ../cryptlib.h ts.h ts_lib.c
+ts_req_print.o: ../../e_os.h ../../include/openssl/asn1.h
+ts_req_print.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+ts_req_print.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
+ts_req_print.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
+ts_req_print.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
+ts_req_print.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+ts_req_print.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+ts_req_print.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ts_req_print.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ts_req_print.o: ../../include/openssl/opensslconf.h
+ts_req_print.o: ../../include/openssl/opensslv.h
+ts_req_print.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ts_req_print.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+ts_req_print.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ts_req_print.o: ../../include/openssl/symhacks.h ../../include/openssl/ts.h
+ts_req_print.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ts_req_print.o: ../../include/openssl/x509v3.h ../cryptlib.h ts_req_print.c
+ts_req_utils.o: ../../e_os.h ../../include/openssl/asn1.h
+ts_req_utils.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+ts_req_utils.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+ts_req_utils.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
+ts_req_utils.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ts_req_utils.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ts_req_utils.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+ts_req_utils.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ts_req_utils.o: ../../include/openssl/objects.h
+ts_req_utils.o: ../../include/openssl/opensslconf.h
+ts_req_utils.o: ../../include/openssl/opensslv.h
+ts_req_utils.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ts_req_utils.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+ts_req_utils.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ts_req_utils.o: ../../include/openssl/symhacks.h ../../include/openssl/ts.h
+ts_req_utils.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ts_req_utils.o: ../../include/openssl/x509v3.h ../cryptlib.h ts_req_utils.c
+ts_rsp_print.o: ../../e_os.h ../../include/openssl/asn1.h
+ts_rsp_print.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
+ts_rsp_print.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
+ts_rsp_print.o: ../../include/openssl/crypto.h ../../include/openssl/dh.h
+ts_rsp_print.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
+ts_rsp_print.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+ts_rsp_print.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+ts_rsp_print.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+ts_rsp_print.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+ts_rsp_print.o: ../../include/openssl/opensslconf.h
+ts_rsp_print.o: ../../include/openssl/opensslv.h
+ts_rsp_print.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ts_rsp_print.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+ts_rsp_print.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ts_rsp_print.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+ts_rsp_print.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+ts_rsp_print.o: ../cryptlib.h ts.h ts_rsp_print.c
+ts_rsp_sign.o: ../../e_os.h ../../include/openssl/asn1.h
+ts_rsp_sign.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+ts_rsp_sign.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+ts_rsp_sign.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
+ts_rsp_sign.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ts_rsp_sign.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ts_rsp_sign.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+ts_rsp_sign.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ts_rsp_sign.o: ../../include/openssl/objects.h
+ts_rsp_sign.o: ../../include/openssl/opensslconf.h
+ts_rsp_sign.o: ../../include/openssl/opensslv.h
+ts_rsp_sign.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ts_rsp_sign.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+ts_rsp_sign.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ts_rsp_sign.o: ../../include/openssl/symhacks.h ../../include/openssl/ts.h
+ts_rsp_sign.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ts_rsp_sign.o: ../../include/openssl/x509v3.h ../cryptlib.h ts_rsp_sign.c
+ts_rsp_utils.o: ../../e_os.h ../../include/openssl/asn1.h
+ts_rsp_utils.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+ts_rsp_utils.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+ts_rsp_utils.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
+ts_rsp_utils.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ts_rsp_utils.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ts_rsp_utils.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+ts_rsp_utils.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ts_rsp_utils.o: ../../include/openssl/objects.h
+ts_rsp_utils.o: ../../include/openssl/opensslconf.h
+ts_rsp_utils.o: ../../include/openssl/opensslv.h
+ts_rsp_utils.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ts_rsp_utils.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+ts_rsp_utils.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ts_rsp_utils.o: ../../include/openssl/symhacks.h ../../include/openssl/ts.h
+ts_rsp_utils.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ts_rsp_utils.o: ../../include/openssl/x509v3.h ../cryptlib.h ts_rsp_utils.c
+ts_rsp_verify.o: ../../e_os.h ../../include/openssl/asn1.h
+ts_rsp_verify.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+ts_rsp_verify.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+ts_rsp_verify.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
+ts_rsp_verify.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ts_rsp_verify.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ts_rsp_verify.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+ts_rsp_verify.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ts_rsp_verify.o: ../../include/openssl/objects.h
+ts_rsp_verify.o: ../../include/openssl/opensslconf.h
+ts_rsp_verify.o: ../../include/openssl/opensslv.h
+ts_rsp_verify.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ts_rsp_verify.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+ts_rsp_verify.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ts_rsp_verify.o: ../../include/openssl/symhacks.h ../../include/openssl/ts.h
+ts_rsp_verify.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ts_rsp_verify.o: ../../include/openssl/x509v3.h ../cryptlib.h ts_rsp_verify.c
+ts_verify_ctx.o: ../../e_os.h ../../include/openssl/asn1.h
+ts_verify_ctx.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+ts_verify_ctx.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+ts_verify_ctx.o: ../../include/openssl/dh.h ../../include/openssl/dsa.h
+ts_verify_ctx.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
+ts_verify_ctx.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
+ts_verify_ctx.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+ts_verify_ctx.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+ts_verify_ctx.o: ../../include/openssl/objects.h
+ts_verify_ctx.o: ../../include/openssl/opensslconf.h
+ts_verify_ctx.o: ../../include/openssl/opensslv.h
+ts_verify_ctx.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+ts_verify_ctx.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+ts_verify_ctx.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+ts_verify_ctx.o: ../../include/openssl/symhacks.h ../../include/openssl/ts.h
+ts_verify_ctx.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+ts_verify_ctx.o: ../../include/openssl/x509v3.h ../cryptlib.h ts_verify_ctx.c
diff --git a/crypto/ts/ts.h b/crypto/ts/ts.h
new file mode 100644
index 000000000000..c2448e3c3be8
--- /dev/null
+++ b/crypto/ts/ts.h
@@ -0,0 +1,858 @@
+/* crypto/ts/ts.h */
+/* Written by Zoltan Glozik (zglozik@opentsa.org) for the OpenSSL
+ * project 2002, 2003, 2004.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#ifndef HEADER_TS_H
+#define HEADER_TS_H
+
+#include <openssl/opensslconf.h>
+#include <openssl/symhacks.h>
+#ifndef OPENSSL_NO_BUFFER
+#include <openssl/buffer.h>
+#endif
+#ifndef OPENSSL_NO_EVP
+#include <openssl/evp.h>
+#endif
+#ifndef OPENSSL_NO_BIO
+#include <openssl/bio.h>
+#endif
+#include <openssl/stack.h>
+#include <openssl/asn1.h>
+#include <openssl/safestack.h>
+
+#ifndef OPENSSL_NO_RSA
+#include <openssl/rsa.h>
+#endif
+
+#ifndef OPENSSL_NO_DSA
+#include <openssl/dsa.h>
+#endif
+
+#ifndef OPENSSL_NO_DH
+#include <openssl/dh.h>
+#endif
+
+#ifdef  __cplusplus
+extern "C" {
+#endif
+
+#ifdef WIN32
+/* Under Win32 this is defined in wincrypt.h */
+#undef X509_NAME
+#endif
+
+#include <openssl/x509.h>
+#include <openssl/x509v3.h>
+
+/*
+MessageImprint ::= SEQUENCE  {
+     hashAlgorithm                AlgorithmIdentifier,
+     hashedMessage                OCTET STRING  }
+*/
+
+typedef struct TS_msg_imprint_st
+	{
+	X509_ALGOR *hash_algo;
+	ASN1_OCTET_STRING *hashed_msg;
+	} TS_MSG_IMPRINT;
+
+/*
+TimeStampReq ::= SEQUENCE  {
+   version                  INTEGER  { v1(1) },
+   messageImprint           MessageImprint,
+     --a hash algorithm OID and the hash value of the data to be
+     --time-stamped
+   reqPolicy                TSAPolicyId                OPTIONAL,
+   nonce                    INTEGER                    OPTIONAL,
+   certReq                  BOOLEAN                    DEFAULT FALSE,
+   extensions               [0] IMPLICIT Extensions    OPTIONAL  }
+*/
+
+typedef struct TS_req_st
+	{
+	ASN1_INTEGER *version;
+	TS_MSG_IMPRINT *msg_imprint;
+	ASN1_OBJECT *policy_id;		/* OPTIONAL */
+	ASN1_INTEGER *nonce;		/* OPTIONAL */
+	ASN1_BOOLEAN cert_req;		/* DEFAULT FALSE */
+	STACK_OF(X509_EXTENSION) *extensions;	/* [0] OPTIONAL */
+	} TS_REQ;
+
+/*
+Accuracy ::= SEQUENCE {
+                seconds        INTEGER           OPTIONAL,
+                millis     [0] INTEGER  (1..999) OPTIONAL,
+                micros     [1] INTEGER  (1..999) OPTIONAL  }
+*/
+
+typedef struct TS_accuracy_st
+	{
+	ASN1_INTEGER *seconds;
+	ASN1_INTEGER *millis;
+	ASN1_INTEGER *micros;
+	} TS_ACCURACY;
+
+/*
+TSTInfo ::= SEQUENCE  {
+    version                      INTEGER  { v1(1) },
+    policy                       TSAPolicyId,
+    messageImprint               MessageImprint,
+      -- MUST have the same value as the similar field in
+      -- TimeStampReq
+    serialNumber                 INTEGER,
+     -- Time-Stamping users MUST be ready to accommodate integers
+     -- up to 160 bits.
+    genTime                      GeneralizedTime,
+    accuracy                     Accuracy                 OPTIONAL,
+    ordering                     BOOLEAN             DEFAULT FALSE,
+    nonce                        INTEGER                  OPTIONAL,
+      -- MUST be present if the similar field was present
+      -- in TimeStampReq.  In that case it MUST have the same value.
+    tsa                          [0] GeneralName          OPTIONAL,
+    extensions                   [1] IMPLICIT Extensions  OPTIONAL   }
+*/
+
+typedef struct TS_tst_info_st
+	{
+	ASN1_INTEGER *version;
+	ASN1_OBJECT *policy_id;
+	TS_MSG_IMPRINT *msg_imprint;
+	ASN1_INTEGER *serial;
+	ASN1_GENERALIZEDTIME *time;
+	TS_ACCURACY *accuracy;
+	ASN1_BOOLEAN ordering;
+	ASN1_INTEGER *nonce;
+	GENERAL_NAME *tsa;
+	STACK_OF(X509_EXTENSION) *extensions;
+	} TS_TST_INFO;	
+
+/*
+PKIStatusInfo ::= SEQUENCE {
+    status        PKIStatus,
+    statusString  PKIFreeText     OPTIONAL,
+    failInfo      PKIFailureInfo  OPTIONAL  }
+
+From RFC 1510 - section 3.1.1:
+PKIFreeText ::= SEQUENCE SIZE (1..MAX) OF UTF8String
+	-- text encoded as UTF-8 String (note:  each UTF8String SHOULD
+	-- include an RFC 1766 language tag to indicate the language
+	-- of the contained text)
+*/
+
+/* Possible values for status. See ts_resp_print.c && ts_resp_verify.c. */
+
+#define	TS_STATUS_GRANTED			0
+#define	TS_STATUS_GRANTED_WITH_MODS		1
+#define	TS_STATUS_REJECTION			2
+#define	TS_STATUS_WAITING			3
+#define	TS_STATUS_REVOCATION_WARNING		4
+#define	TS_STATUS_REVOCATION_NOTIFICATION	5
+
+/* Possible values for failure_info. See ts_resp_print.c && ts_resp_verify.c */
+
+#define	TS_INFO_BAD_ALG			0
+#define	TS_INFO_BAD_REQUEST		2
+#define	TS_INFO_BAD_DATA_FORMAT		5
+#define	TS_INFO_TIME_NOT_AVAILABLE	14
+#define	TS_INFO_UNACCEPTED_POLICY	15
+#define	TS_INFO_UNACCEPTED_EXTENSION	16
+#define	TS_INFO_ADD_INFO_NOT_AVAILABLE	17
+#define	TS_INFO_SYSTEM_FAILURE		25
+
+typedef struct TS_status_info_st
+	{
+	ASN1_INTEGER *status;
+	STACK_OF(ASN1_UTF8STRING) *text;
+	ASN1_BIT_STRING *failure_info;
+	} TS_STATUS_INFO;
+
+DECLARE_STACK_OF(ASN1_UTF8STRING)
+DECLARE_ASN1_SET_OF(ASN1_UTF8STRING)
+
+/*
+TimeStampResp ::= SEQUENCE  {
+     status                  PKIStatusInfo,
+     timeStampToken          TimeStampToken     OPTIONAL }
+*/
+
+typedef struct TS_resp_st
+	{
+	TS_STATUS_INFO *status_info;
+	PKCS7 *token;
+	TS_TST_INFO *tst_info;
+	} TS_RESP;
+
+/* The structure below would belong to the ESS component. */
+
+/*
+IssuerSerial ::= SEQUENCE {
+	issuer                   GeneralNames,
+	serialNumber             CertificateSerialNumber
+	}
+*/
+
+typedef struct ESS_issuer_serial
+	{
+	STACK_OF(GENERAL_NAME)	*issuer;
+	ASN1_INTEGER		*serial;
+	} ESS_ISSUER_SERIAL;
+
+/*
+ESSCertID ::=  SEQUENCE {
+        certHash                 Hash,
+        issuerSerial             IssuerSerial OPTIONAL
+}
+*/
+
+typedef struct ESS_cert_id
+	{
+	ASN1_OCTET_STRING *hash;	/* Always SHA-1 digest. */
+	ESS_ISSUER_SERIAL *issuer_serial;
+	} ESS_CERT_ID;
+
+DECLARE_STACK_OF(ESS_CERT_ID)
+DECLARE_ASN1_SET_OF(ESS_CERT_ID)
+
+/*
+SigningCertificate ::=  SEQUENCE {
+       certs        SEQUENCE OF ESSCertID,
+       policies     SEQUENCE OF PolicyInformation OPTIONAL
+}
+*/
+
+typedef struct ESS_signing_cert
+	{
+	STACK_OF(ESS_CERT_ID) *cert_ids;
+	STACK_OF(POLICYINFO) *policy_info;
+	} ESS_SIGNING_CERT;
+
+
+TS_REQ	*TS_REQ_new(void);
+void	TS_REQ_free(TS_REQ *a);
+int	i2d_TS_REQ(const TS_REQ *a, unsigned char **pp);
+TS_REQ	*d2i_TS_REQ(TS_REQ **a, const unsigned char **pp, long length);
+
+TS_REQ	*TS_REQ_dup(TS_REQ *a);
+
+TS_REQ	*d2i_TS_REQ_fp(FILE *fp, TS_REQ **a);
+int	i2d_TS_REQ_fp(FILE *fp, TS_REQ *a);
+TS_REQ	*d2i_TS_REQ_bio(BIO *fp, TS_REQ **a);
+int	i2d_TS_REQ_bio(BIO *fp, TS_REQ *a);
+
+TS_MSG_IMPRINT	*TS_MSG_IMPRINT_new(void);
+void		TS_MSG_IMPRINT_free(TS_MSG_IMPRINT *a);
+int		i2d_TS_MSG_IMPRINT(const TS_MSG_IMPRINT *a, unsigned char **pp);
+TS_MSG_IMPRINT	*d2i_TS_MSG_IMPRINT(TS_MSG_IMPRINT **a,
+				    const unsigned char **pp, long length);
+
+TS_MSG_IMPRINT	*TS_MSG_IMPRINT_dup(TS_MSG_IMPRINT *a);
+
+TS_MSG_IMPRINT	*d2i_TS_MSG_IMPRINT_fp(FILE *fp, TS_MSG_IMPRINT **a);
+int		i2d_TS_MSG_IMPRINT_fp(FILE *fp, TS_MSG_IMPRINT *a);
+TS_MSG_IMPRINT	*d2i_TS_MSG_IMPRINT_bio(BIO *fp, TS_MSG_IMPRINT **a);
+int		i2d_TS_MSG_IMPRINT_bio(BIO *fp, TS_MSG_IMPRINT *a);
+
+TS_RESP	*TS_RESP_new(void);
+void	TS_RESP_free(TS_RESP *a);
+int	i2d_TS_RESP(const TS_RESP *a, unsigned char **pp);
+TS_RESP	*d2i_TS_RESP(TS_RESP **a, const unsigned char **pp, long length);
+TS_TST_INFO *PKCS7_to_TS_TST_INFO(PKCS7 *token);
+TS_RESP	*TS_RESP_dup(TS_RESP *a);
+
+TS_RESP	*d2i_TS_RESP_fp(FILE *fp, TS_RESP **a);
+int	i2d_TS_RESP_fp(FILE *fp, TS_RESP *a);
+TS_RESP	*d2i_TS_RESP_bio(BIO *fp, TS_RESP **a);
+int	i2d_TS_RESP_bio(BIO *fp, TS_RESP *a);
+
+TS_STATUS_INFO	*TS_STATUS_INFO_new(void);
+void		TS_STATUS_INFO_free(TS_STATUS_INFO *a);
+int		i2d_TS_STATUS_INFO(const TS_STATUS_INFO *a, unsigned char **pp);
+TS_STATUS_INFO	*d2i_TS_STATUS_INFO(TS_STATUS_INFO **a, 
+				    const unsigned char **pp, long length);
+TS_STATUS_INFO	*TS_STATUS_INFO_dup(TS_STATUS_INFO *a);
+
+TS_TST_INFO	*TS_TST_INFO_new(void);
+void		TS_TST_INFO_free(TS_TST_INFO *a);
+int		i2d_TS_TST_INFO(const TS_TST_INFO *a, unsigned char **pp);
+TS_TST_INFO	*d2i_TS_TST_INFO(TS_TST_INFO **a, const unsigned char **pp,
+				    long length);
+TS_TST_INFO	*TS_TST_INFO_dup(TS_TST_INFO *a);
+
+TS_TST_INFO	*d2i_TS_TST_INFO_fp(FILE *fp, TS_TST_INFO **a);
+int		i2d_TS_TST_INFO_fp(FILE *fp, TS_TST_INFO *a);
+TS_TST_INFO	*d2i_TS_TST_INFO_bio(BIO *fp, TS_TST_INFO **a);
+int		i2d_TS_TST_INFO_bio(BIO *fp, TS_TST_INFO *a);
+
+TS_ACCURACY	*TS_ACCURACY_new(void);
+void		TS_ACCURACY_free(TS_ACCURACY *a);
+int		i2d_TS_ACCURACY(const TS_ACCURACY *a, unsigned char **pp);
+TS_ACCURACY	*d2i_TS_ACCURACY(TS_ACCURACY **a, const unsigned char **pp,
+				    long length);
+TS_ACCURACY	*TS_ACCURACY_dup(TS_ACCURACY *a);
+
+ESS_ISSUER_SERIAL *ESS_ISSUER_SERIAL_new(void);
+void		  ESS_ISSUER_SERIAL_free(ESS_ISSUER_SERIAL *a);
+int		  i2d_ESS_ISSUER_SERIAL(const ESS_ISSUER_SERIAL *a,
+					unsigned char **pp);
+ESS_ISSUER_SERIAL *d2i_ESS_ISSUER_SERIAL(ESS_ISSUER_SERIAL **a,
+					 const unsigned char **pp, long length);
+ESS_ISSUER_SERIAL *ESS_ISSUER_SERIAL_dup(ESS_ISSUER_SERIAL *a);
+
+ESS_CERT_ID	*ESS_CERT_ID_new(void);
+void		ESS_CERT_ID_free(ESS_CERT_ID *a);
+int		i2d_ESS_CERT_ID(const ESS_CERT_ID *a, unsigned char **pp);
+ESS_CERT_ID	*d2i_ESS_CERT_ID(ESS_CERT_ID **a, const unsigned char **pp,
+				 long length);
+ESS_CERT_ID	*ESS_CERT_ID_dup(ESS_CERT_ID *a);
+
+ESS_SIGNING_CERT *ESS_SIGNING_CERT_new(void);
+void		 ESS_SIGNING_CERT_free(ESS_SIGNING_CERT *a);
+int		 i2d_ESS_SIGNING_CERT(const ESS_SIGNING_CERT *a, 
+				      unsigned char **pp);
+ESS_SIGNING_CERT *d2i_ESS_SIGNING_CERT(ESS_SIGNING_CERT **a,
+				       const unsigned char **pp, long length);
+ESS_SIGNING_CERT *ESS_SIGNING_CERT_dup(ESS_SIGNING_CERT *a);
+
+void ERR_load_TS_strings(void);
+
+int TS_REQ_set_version(TS_REQ *a, long version);
+long TS_REQ_get_version(const TS_REQ *a);
+
+int TS_REQ_set_msg_imprint(TS_REQ *a, TS_MSG_IMPRINT *msg_imprint);
+TS_MSG_IMPRINT *TS_REQ_get_msg_imprint(TS_REQ *a);
+
+int TS_MSG_IMPRINT_set_algo(TS_MSG_IMPRINT *a, X509_ALGOR *alg);
+X509_ALGOR *TS_MSG_IMPRINT_get_algo(TS_MSG_IMPRINT *a);
+
+int TS_MSG_IMPRINT_set_msg(TS_MSG_IMPRINT *a, unsigned char *d, int len);
+ASN1_OCTET_STRING *TS_MSG_IMPRINT_get_msg(TS_MSG_IMPRINT *a);
+
+int TS_REQ_set_policy_id(TS_REQ *a, ASN1_OBJECT *policy);
+ASN1_OBJECT *TS_REQ_get_policy_id(TS_REQ *a);
+
+int TS_REQ_set_nonce(TS_REQ *a, const ASN1_INTEGER *nonce);
+const ASN1_INTEGER *TS_REQ_get_nonce(const TS_REQ *a);
+
+int TS_REQ_set_cert_req(TS_REQ *a, int cert_req);
+int TS_REQ_get_cert_req(const TS_REQ *a);
+
+STACK_OF(X509_EXTENSION) *TS_REQ_get_exts(TS_REQ *a);
+void TS_REQ_ext_free(TS_REQ *a);
+int TS_REQ_get_ext_count(TS_REQ *a);
+int TS_REQ_get_ext_by_NID(TS_REQ *a, int nid, int lastpos);
+int TS_REQ_get_ext_by_OBJ(TS_REQ *a, ASN1_OBJECT *obj, int lastpos);
+int TS_REQ_get_ext_by_critical(TS_REQ *a, int crit, int lastpos);
+X509_EXTENSION *TS_REQ_get_ext(TS_REQ *a, int loc);
+X509_EXTENSION *TS_REQ_delete_ext(TS_REQ *a, int loc);
+int TS_REQ_add_ext(TS_REQ *a, X509_EXTENSION *ex, int loc);
+void *TS_REQ_get_ext_d2i(TS_REQ *a, int nid, int *crit, int *idx);
+
+/* Function declarations for TS_REQ defined in ts/ts_req_print.c */
+
+int TS_REQ_print_bio(BIO *bio, TS_REQ *a);
+
+/* Function declarations for TS_RESP defined in ts/ts_resp_utils.c */
+
+int TS_RESP_set_status_info(TS_RESP *a, TS_STATUS_INFO *info);
+TS_STATUS_INFO *TS_RESP_get_status_info(TS_RESP *a);
+
+/* Caller loses ownership of PKCS7 and TS_TST_INFO objects. */
+void TS_RESP_set_tst_info(TS_RESP *a, PKCS7 *p7, TS_TST_INFO *tst_info);
+PKCS7 *TS_RESP_get_token(TS_RESP *a);
+TS_TST_INFO *TS_RESP_get_tst_info(TS_RESP *a);
+
+int TS_TST_INFO_set_version(TS_TST_INFO *a, long version);
+long TS_TST_INFO_get_version(const TS_TST_INFO *a);
+
+int TS_TST_INFO_set_policy_id(TS_TST_INFO *a, ASN1_OBJECT *policy_id);
+ASN1_OBJECT *TS_TST_INFO_get_policy_id(TS_TST_INFO *a);
+
+int TS_TST_INFO_set_msg_imprint(TS_TST_INFO *a, TS_MSG_IMPRINT *msg_imprint);
+TS_MSG_IMPRINT *TS_TST_INFO_get_msg_imprint(TS_TST_INFO *a);
+
+int TS_TST_INFO_set_serial(TS_TST_INFO *a, const ASN1_INTEGER *serial);
+const ASN1_INTEGER *TS_TST_INFO_get_serial(const TS_TST_INFO *a);
+
+int TS_TST_INFO_set_time(TS_TST_INFO *a, const ASN1_GENERALIZEDTIME *gtime);
+const ASN1_GENERALIZEDTIME *TS_TST_INFO_get_time(const TS_TST_INFO *a);
+
+int TS_TST_INFO_set_accuracy(TS_TST_INFO *a, TS_ACCURACY *accuracy);
+TS_ACCURACY *TS_TST_INFO_get_accuracy(TS_TST_INFO *a);
+
+int TS_ACCURACY_set_seconds(TS_ACCURACY *a, const ASN1_INTEGER *seconds);
+const ASN1_INTEGER *TS_ACCURACY_get_seconds(const TS_ACCURACY *a);
+
+int TS_ACCURACY_set_millis(TS_ACCURACY *a, const ASN1_INTEGER *millis);
+const ASN1_INTEGER *TS_ACCURACY_get_millis(const TS_ACCURACY *a);
+
+int TS_ACCURACY_set_micros(TS_ACCURACY *a, const ASN1_INTEGER *micros);
+const ASN1_INTEGER *TS_ACCURACY_get_micros(const TS_ACCURACY *a);
+
+int TS_TST_INFO_set_ordering(TS_TST_INFO *a, int ordering);
+int TS_TST_INFO_get_ordering(const TS_TST_INFO *a);
+
+int TS_TST_INFO_set_nonce(TS_TST_INFO *a, const ASN1_INTEGER *nonce);
+const ASN1_INTEGER *TS_TST_INFO_get_nonce(const TS_TST_INFO *a);
+
+int TS_TST_INFO_set_tsa(TS_TST_INFO *a, GENERAL_NAME *tsa);
+GENERAL_NAME *TS_TST_INFO_get_tsa(TS_TST_INFO *a);
+
+STACK_OF(X509_EXTENSION) *TS_TST_INFO_get_exts(TS_TST_INFO *a);
+void TS_TST_INFO_ext_free(TS_TST_INFO *a);
+int TS_TST_INFO_get_ext_count(TS_TST_INFO *a);
+int TS_TST_INFO_get_ext_by_NID(TS_TST_INFO *a, int nid, int lastpos);
+int TS_TST_INFO_get_ext_by_OBJ(TS_TST_INFO *a, ASN1_OBJECT *obj, int lastpos);
+int TS_TST_INFO_get_ext_by_critical(TS_TST_INFO *a, int crit, int lastpos);
+X509_EXTENSION *TS_TST_INFO_get_ext(TS_TST_INFO *a, int loc);
+X509_EXTENSION *TS_TST_INFO_delete_ext(TS_TST_INFO *a, int loc);
+int TS_TST_INFO_add_ext(TS_TST_INFO *a, X509_EXTENSION *ex, int loc);
+void *TS_TST_INFO_get_ext_d2i(TS_TST_INFO *a, int nid, int *crit, int *idx);
+
+/* Declarations related to response generation, defined in ts/ts_resp_sign.c. */
+
+/* Optional flags for response generation. */
+
+/* Don't include the TSA name in response. */
+#define	TS_TSA_NAME		0x01
+
+/* Set ordering to true in response. */
+#define	TS_ORDERING		0x02
+
+/*
+ * Include the signer certificate and the other specified certificates in
+ * the ESS signing certificate attribute beside the PKCS7 signed data.
+ * Only the signer certificates is included by default.
+ */
+#define	TS_ESS_CERT_ID_CHAIN	0x04
+
+/* Forward declaration. */
+struct TS_resp_ctx;
+
+/* This must return a unique number less than 160 bits long. */
+typedef ASN1_INTEGER *(*TS_serial_cb)(struct TS_resp_ctx *, void *);
+
+/* This must return the seconds and microseconds since Jan 1, 1970 in
+   the sec and usec variables allocated by the caller. 
+   Return non-zero for success and zero for failure. */
+typedef	int (*TS_time_cb)(struct TS_resp_ctx *, void *, long *sec, long *usec);
+
+/* This must process the given extension.
+ * It can modify the TS_TST_INFO object of the context.
+ * Return values: !0 (processed), 0 (error, it must set the 
+ * status info/failure info of the response).
+ */
+typedef	int (*TS_extension_cb)(struct TS_resp_ctx *, X509_EXTENSION *, void *);
+
+typedef struct TS_resp_ctx
+	{
+	X509		*signer_cert;
+	EVP_PKEY	*signer_key;
+	STACK_OF(X509)	*certs;	/* Certs to include in signed data. */
+	STACK_OF(ASN1_OBJECT)	*policies;	/* Acceptable policies. */
+	ASN1_OBJECT	*default_policy; /* It may appear in policies, too. */
+	STACK_OF(EVP_MD)	*mds;	/* Acceptable message digests. */
+	ASN1_INTEGER	*seconds;	/* accuracy, 0 means not specified. */
+	ASN1_INTEGER	*millis;	/* accuracy, 0 means not specified. */
+	ASN1_INTEGER	*micros;	/* accuracy, 0 means not specified. */
+	unsigned	clock_precision_digits; /* fraction of seconds in
+						   time stamp token. */
+	unsigned	flags;		/* Optional info, see values above. */
+
+	/* Callback functions. */
+	TS_serial_cb serial_cb;
+	void *serial_cb_data;	/* User data for serial_cb. */
+	
+	TS_time_cb time_cb;
+	void *time_cb_data;	/* User data for time_cb. */
+	
+	TS_extension_cb extension_cb;
+	void *extension_cb_data;	/* User data for extension_cb. */
+
+	/* These members are used only while creating the response. */
+	TS_REQ		*request;
+	TS_RESP		*response;
+	TS_TST_INFO	*tst_info;
+	} TS_RESP_CTX;
+
+DECLARE_STACK_OF(EVP_MD)
+DECLARE_ASN1_SET_OF(EVP_MD)
+
+/* Creates a response context that can be used for generating responses. */
+TS_RESP_CTX *TS_RESP_CTX_new(void);
+void TS_RESP_CTX_free(TS_RESP_CTX *ctx);
+
+/* This parameter must be set. */
+int TS_RESP_CTX_set_signer_cert(TS_RESP_CTX *ctx, X509 *signer);
+
+/* This parameter must be set. */
+int TS_RESP_CTX_set_signer_key(TS_RESP_CTX *ctx, EVP_PKEY *key);
+
+/* This parameter must be set. */
+int TS_RESP_CTX_set_def_policy(TS_RESP_CTX *ctx, ASN1_OBJECT *def_policy);
+
+/* No additional certs are included in the response by default. */
+int TS_RESP_CTX_set_certs(TS_RESP_CTX *ctx, STACK_OF(X509) *certs);
+
+/* Adds a new acceptable policy, only the default policy 
+   is accepted by default. */
+int TS_RESP_CTX_add_policy(TS_RESP_CTX *ctx, ASN1_OBJECT *policy);
+
+/* Adds a new acceptable message digest. Note that no message digests 
+   are accepted by default. The md argument is shared with the caller. */
+int TS_RESP_CTX_add_md(TS_RESP_CTX *ctx, const EVP_MD *md);
+
+/* Accuracy is not included by default. */
+int TS_RESP_CTX_set_accuracy(TS_RESP_CTX *ctx,
+			     int secs, int millis, int micros);
+
+/* Clock precision digits, i.e. the number of decimal digits: 
+   '0' means sec, '3' msec, '6' usec, and so on. Default is 0. */ 
+int TS_RESP_CTX_set_clock_precision_digits(TS_RESP_CTX *ctx,
+					   unsigned clock_precision_digits);
+/* At most we accept usec precision. */	
+#define TS_MAX_CLOCK_PRECISION_DIGITS	6
+
+/* No flags are set by default. */
+void TS_RESP_CTX_add_flags(TS_RESP_CTX *ctx, int flags);
+
+/* Default callback always returns a constant. */
+void TS_RESP_CTX_set_serial_cb(TS_RESP_CTX *ctx, TS_serial_cb cb, void *data);
+
+/* Default callback uses the gettimeofday() and gmtime() system calls. */
+void TS_RESP_CTX_set_time_cb(TS_RESP_CTX *ctx, TS_time_cb cb, void *data);
+
+/* Default callback rejects all extensions. The extension callback is called 
+ * when the TS_TST_INFO object is already set up and not signed yet. */
+/* FIXME: extension handling is not tested yet. */
+void TS_RESP_CTX_set_extension_cb(TS_RESP_CTX *ctx, 
+				  TS_extension_cb cb, void *data);
+
+/* The following methods can be used in the callbacks. */
+int TS_RESP_CTX_set_status_info(TS_RESP_CTX *ctx, 
+				int status, const char *text);
+
+/* Sets the status info only if it is still TS_STATUS_GRANTED. */
+int TS_RESP_CTX_set_status_info_cond(TS_RESP_CTX *ctx, 
+				     int status, const char *text);
+
+int TS_RESP_CTX_add_failure_info(TS_RESP_CTX *ctx, int failure);
+
+/* The get methods below can be used in the extension callback. */
+TS_REQ *TS_RESP_CTX_get_request(TS_RESP_CTX *ctx);
+
+TS_TST_INFO *TS_RESP_CTX_get_tst_info(TS_RESP_CTX *ctx);
+
+/* 
+ * Creates the signed TS_TST_INFO and puts it in TS_RESP.
+ * In case of errors it sets the status info properly.
+ * Returns NULL only in case of memory allocation/fatal error.
+ */
+TS_RESP *TS_RESP_create_response(TS_RESP_CTX *ctx, BIO *req_bio);
+
+/*
+ * Declarations related to response verification,
+ * they are defined in ts/ts_resp_verify.c.
+ */
+
+int TS_RESP_verify_signature(PKCS7 *token, STACK_OF(X509) *certs,
+			     X509_STORE *store, X509 **signer_out);
+
+/* Context structure for the generic verify method. */
+
+/* Verify the signer's certificate and the signature of the response. */
+#define	TS_VFY_SIGNATURE	(1u << 0)
+/* Verify the version number of the response. */
+#define	TS_VFY_VERSION		(1u << 1)
+/* Verify if the policy supplied by the user matches the policy of the TSA. */
+#define	TS_VFY_POLICY		(1u << 2)
+/* Verify the message imprint provided by the user. This flag should not be
+   specified with TS_VFY_DATA. */
+#define	TS_VFY_IMPRINT		(1u << 3)
+/* Verify the message imprint computed by the verify method from the user
+   provided data and the MD algorithm of the response. This flag should not be
+   specified with TS_VFY_IMPRINT. */
+#define	TS_VFY_DATA		(1u << 4)
+/* Verify the nonce value. */
+#define	TS_VFY_NONCE		(1u << 5)
+/* Verify if the TSA name field matches the signer certificate. */
+#define	TS_VFY_SIGNER		(1u << 6)
+/* Verify if the TSA name field equals to the user provided name. */
+#define	TS_VFY_TSA_NAME		(1u << 7)
+
+/* You can use the following convenience constants. */
+#define	TS_VFY_ALL_IMPRINT	(TS_VFY_SIGNATURE	\
+				 | TS_VFY_VERSION	\
+				 | TS_VFY_POLICY	\
+				 | TS_VFY_IMPRINT	\
+				 | TS_VFY_NONCE		\
+				 | TS_VFY_SIGNER	\
+				 | TS_VFY_TSA_NAME)
+#define	TS_VFY_ALL_DATA		(TS_VFY_SIGNATURE	\
+				 | TS_VFY_VERSION	\
+				 | TS_VFY_POLICY	\
+				 | TS_VFY_DATA		\
+				 | TS_VFY_NONCE		\
+				 | TS_VFY_SIGNER	\
+				 | TS_VFY_TSA_NAME)
+
+typedef struct TS_verify_ctx
+	{
+	/* Set this to the union of TS_VFY_... flags you want to carry out. */
+	unsigned	flags;
+
+	/* Must be set only with TS_VFY_SIGNATURE. certs is optional. */
+	X509_STORE	*store;
+	STACK_OF(X509)	*certs;
+
+	/* Must be set only with TS_VFY_POLICY. */
+	ASN1_OBJECT	*policy;
+
+	/* Must be set only with TS_VFY_IMPRINT. If md_alg is NULL, 
+	   the algorithm from the response is used. */
+	X509_ALGOR	*md_alg;
+	unsigned char	*imprint;
+	unsigned	imprint_len;
+
+	/* Must be set only with TS_VFY_DATA. */
+	BIO		*data;
+
+	/* Must be set only with TS_VFY_TSA_NAME. */
+	ASN1_INTEGER	*nonce;
+
+	/* Must be set only with TS_VFY_TSA_NAME. */
+	GENERAL_NAME	*tsa_name;
+	} TS_VERIFY_CTX;
+
+int TS_RESP_verify_response(TS_VERIFY_CTX *ctx, TS_RESP *response);
+int TS_RESP_verify_token(TS_VERIFY_CTX *ctx, PKCS7 *token);
+
+/*
+ * Declarations related to response verification context,
+ * they are defined in ts/ts_verify_ctx.c.
+ */
+
+/* Set all fields to zero. */
+TS_VERIFY_CTX *TS_VERIFY_CTX_new(void);
+void TS_VERIFY_CTX_init(TS_VERIFY_CTX *ctx);
+void TS_VERIFY_CTX_free(TS_VERIFY_CTX *ctx);
+void TS_VERIFY_CTX_cleanup(TS_VERIFY_CTX *ctx);
+
+/* 
+ * If ctx is NULL, it allocates and returns a new object, otherwise
+ * it returns ctx. It initialises all the members as follows:
+ * flags = TS_VFY_ALL_IMPRINT & ~(TS_VFY_TSA_NAME | TS_VFY_SIGNATURE)
+ * certs = NULL
+ * store = NULL
+ * policy = policy from the request or NULL if absent (in this case
+ *	TS_VFY_POLICY is cleared from flags as well)
+ * md_alg = MD algorithm from request
+ * imprint, imprint_len = imprint from request
+ * data = NULL
+ * nonce, nonce_len = nonce from the request or NULL if absent (in this case
+ * 	TS_VFY_NONCE is cleared from flags as well)
+ * tsa_name = NULL
+ * Important: after calling this method TS_VFY_SIGNATURE should be added!
+ */
+TS_VERIFY_CTX *TS_REQ_to_TS_VERIFY_CTX(TS_REQ *req, TS_VERIFY_CTX *ctx);
+
+/* Function declarations for TS_RESP defined in ts/ts_resp_print.c */
+
+int TS_RESP_print_bio(BIO *bio, TS_RESP *a);
+int TS_STATUS_INFO_print_bio(BIO *bio, TS_STATUS_INFO *a);
+int TS_TST_INFO_print_bio(BIO *bio, TS_TST_INFO *a);
+
+/* Common utility functions defined in ts/ts_lib.c */
+
+int TS_ASN1_INTEGER_print_bio(BIO *bio, const ASN1_INTEGER *num);
+int TS_OBJ_print_bio(BIO *bio, const ASN1_OBJECT *obj);
+int TS_ext_print_bio(BIO *bio, const STACK_OF(X509_EXTENSION) *extensions);
+int TS_X509_ALGOR_print_bio(BIO *bio, const X509_ALGOR *alg);
+int TS_MSG_IMPRINT_print_bio(BIO *bio, TS_MSG_IMPRINT *msg);
+
+/* Function declarations for handling configuration options,
+   defined in ts/ts_conf.c */
+
+X509 *TS_CONF_load_cert(const char *file);
+STACK_OF(X509) *TS_CONF_load_certs(const char *file);
+EVP_PKEY *TS_CONF_load_key(const char *file, const char *pass);
+const char *TS_CONF_get_tsa_section(CONF *conf, const char *section);
+int TS_CONF_set_serial(CONF *conf, const char *section, TS_serial_cb cb,
+		       TS_RESP_CTX *ctx);
+int TS_CONF_set_crypto_device(CONF *conf, const char *section,
+			      const char *device);
+int TS_CONF_set_default_engine(const char *name);
+int TS_CONF_set_signer_cert(CONF *conf, const char *section,
+			    const char *cert, TS_RESP_CTX *ctx);
+int TS_CONF_set_certs(CONF *conf, const char *section, const char *certs,
+		      TS_RESP_CTX *ctx);
+int TS_CONF_set_signer_key(CONF *conf, const char *section,
+			   const char *key, const char *pass, TS_RESP_CTX *ctx);
+int TS_CONF_set_def_policy(CONF *conf, const char *section,
+			   const char *policy, TS_RESP_CTX *ctx);
+int TS_CONF_set_policies(CONF *conf, const char *section, TS_RESP_CTX *ctx);
+int TS_CONF_set_digests(CONF *conf, const char *section, TS_RESP_CTX *ctx);
+int TS_CONF_set_accuracy(CONF *conf, const char *section, TS_RESP_CTX *ctx);
+int TS_CONF_set_clock_precision_digits(CONF *conf, const char *section,
+				       TS_RESP_CTX *ctx);
+int TS_CONF_set_ordering(CONF *conf, const char *section, TS_RESP_CTX *ctx);
+int TS_CONF_set_tsa_name(CONF *conf, const char *section, TS_RESP_CTX *ctx);
+int TS_CONF_set_ess_cert_id_chain(CONF *conf, const char *section,
+				  TS_RESP_CTX *ctx);
+
+/* -------------------------------------------------- */
+/* BEGIN ERROR CODES */
+/* The following lines are auto generated by the script mkerr.pl. Any changes
+ * made after this point may be overwritten when the script is next run.
+ */
+void ERR_load_TS_strings(void);
+
+/* Error codes for the TS functions. */
+
+/* Function codes. */
+#define TS_F_D2I_TS_RESP				 147
+#define TS_F_DEF_SERIAL_CB				 110
+#define TS_F_DEF_TIME_CB				 111
+#define TS_F_ESS_ADD_SIGNING_CERT			 112
+#define TS_F_ESS_CERT_ID_NEW_INIT			 113
+#define TS_F_ESS_SIGNING_CERT_NEW_INIT			 114
+#define TS_F_INT_TS_RESP_VERIFY_TOKEN			 149
+#define TS_F_PKCS7_TO_TS_TST_INFO			 148
+#define TS_F_TS_ACCURACY_SET_MICROS			 115
+#define TS_F_TS_ACCURACY_SET_MILLIS			 116
+#define TS_F_TS_ACCURACY_SET_SECONDS			 117
+#define TS_F_TS_CHECK_IMPRINTS				 100
+#define TS_F_TS_CHECK_NONCES				 101
+#define TS_F_TS_CHECK_POLICY				 102
+#define TS_F_TS_CHECK_SIGNING_CERTS			 103
+#define TS_F_TS_CHECK_STATUS_INFO			 104
+#define TS_F_TS_COMPUTE_IMPRINT				 145
+#define TS_F_TS_CONF_SET_DEFAULT_ENGINE			 146
+#define TS_F_TS_GET_STATUS_TEXT				 105
+#define TS_F_TS_MSG_IMPRINT_SET_ALGO			 118
+#define TS_F_TS_REQ_SET_MSG_IMPRINT			 119
+#define TS_F_TS_REQ_SET_NONCE				 120
+#define TS_F_TS_REQ_SET_POLICY_ID			 121
+#define TS_F_TS_RESP_CREATE_RESPONSE			 122
+#define TS_F_TS_RESP_CREATE_TST_INFO			 123
+#define TS_F_TS_RESP_CTX_ADD_FAILURE_INFO		 124
+#define TS_F_TS_RESP_CTX_ADD_MD				 125
+#define TS_F_TS_RESP_CTX_ADD_POLICY			 126
+#define TS_F_TS_RESP_CTX_NEW				 127
+#define TS_F_TS_RESP_CTX_SET_ACCURACY			 128
+#define TS_F_TS_RESP_CTX_SET_CERTS			 129
+#define TS_F_TS_RESP_CTX_SET_DEF_POLICY			 130
+#define TS_F_TS_RESP_CTX_SET_SIGNER_CERT		 131
+#define TS_F_TS_RESP_CTX_SET_STATUS_INFO		 132
+#define TS_F_TS_RESP_GET_POLICY				 133
+#define TS_F_TS_RESP_SET_GENTIME_WITH_PRECISION		 134
+#define TS_F_TS_RESP_SET_STATUS_INFO			 135
+#define TS_F_TS_RESP_SET_TST_INFO			 150
+#define TS_F_TS_RESP_SIGN				 136
+#define TS_F_TS_RESP_VERIFY_SIGNATURE			 106
+#define TS_F_TS_RESP_VERIFY_TOKEN			 107
+#define TS_F_TS_TST_INFO_SET_ACCURACY			 137
+#define TS_F_TS_TST_INFO_SET_MSG_IMPRINT		 138
+#define TS_F_TS_TST_INFO_SET_NONCE			 139
+#define TS_F_TS_TST_INFO_SET_POLICY_ID			 140
+#define TS_F_TS_TST_INFO_SET_SERIAL			 141
+#define TS_F_TS_TST_INFO_SET_TIME			 142
+#define TS_F_TS_TST_INFO_SET_TSA			 143
+#define TS_F_TS_VERIFY					 108
+#define TS_F_TS_VERIFY_CERT				 109
+#define TS_F_TS_VERIFY_CTX_NEW				 144
+
+/* Reason codes. */
+#define TS_R_BAD_PKCS7_TYPE				 132
+#define TS_R_BAD_TYPE					 133
+#define TS_R_CERTIFICATE_VERIFY_ERROR			 100
+#define TS_R_COULD_NOT_SET_ENGINE			 127
+#define TS_R_COULD_NOT_SET_TIME				 115
+#define TS_R_D2I_TS_RESP_INT_FAILED			 128
+#define TS_R_DETACHED_CONTENT				 134
+#define TS_R_ESS_ADD_SIGNING_CERT_ERROR			 116
+#define TS_R_ESS_SIGNING_CERTIFICATE_ERROR		 101
+#define TS_R_INVALID_NULL_POINTER			 102
+#define TS_R_INVALID_SIGNER_CERTIFICATE_PURPOSE		 117
+#define TS_R_MESSAGE_IMPRINT_MISMATCH			 103
+#define TS_R_NONCE_MISMATCH				 104
+#define TS_R_NONCE_NOT_RETURNED				 105
+#define TS_R_NO_CONTENT					 106
+#define TS_R_NO_TIME_STAMP_TOKEN			 107
+#define TS_R_PKCS7_ADD_SIGNATURE_ERROR			 118
+#define TS_R_PKCS7_ADD_SIGNED_ATTR_ERROR		 119
+#define TS_R_PKCS7_TO_TS_TST_INFO_FAILED		 129
+#define TS_R_POLICY_MISMATCH				 108
+#define TS_R_PRIVATE_KEY_DOES_NOT_MATCH_CERTIFICATE	 120
+#define TS_R_RESPONSE_SETUP_ERROR			 121
+#define TS_R_SIGNATURE_FAILURE				 109
+#define TS_R_THERE_MUST_BE_ONE_SIGNER			 110
+#define TS_R_TIME_SYSCALL_ERROR				 122
+#define TS_R_TOKEN_NOT_PRESENT				 130
+#define TS_R_TOKEN_PRESENT				 131
+#define TS_R_TSA_NAME_MISMATCH				 111
+#define TS_R_TSA_UNTRUSTED				 112
+#define TS_R_TST_INFO_SETUP_ERROR			 123
+#define TS_R_TS_DATASIGN				 124
+#define TS_R_UNACCEPTABLE_POLICY			 125
+#define TS_R_UNSUPPORTED_MD_ALGORITHM			 126
+#define TS_R_UNSUPPORTED_VERSION			 113
+#define TS_R_WRONG_CONTENT_TYPE				 114
+
+#ifdef  __cplusplus
+}
+#endif
+#endif
diff --git a/crypto/ts/ts_asn1.c b/crypto/ts/ts_asn1.c
new file mode 100644
index 000000000000..40b730c5e2ea
--- /dev/null
+++ b/crypto/ts/ts_asn1.c
@@ -0,0 +1,322 @@
+/* crypto/ts/ts_asn1.c */
+/* Written by Nils Larsch for the OpenSSL project 2004.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <openssl/ts.h>
+#include <openssl/err.h>
+#include <openssl/asn1t.h>
+
+ASN1_SEQUENCE(TS_MSG_IMPRINT) = {
+	ASN1_SIMPLE(TS_MSG_IMPRINT, hash_algo, X509_ALGOR),
+	ASN1_SIMPLE(TS_MSG_IMPRINT, hashed_msg, ASN1_OCTET_STRING)
+} ASN1_SEQUENCE_END(TS_MSG_IMPRINT)
+
+IMPLEMENT_ASN1_FUNCTIONS_const(TS_MSG_IMPRINT)
+IMPLEMENT_ASN1_DUP_FUNCTION(TS_MSG_IMPRINT)
+#ifndef OPENSSL_NO_BIO
+TS_MSG_IMPRINT *d2i_TS_MSG_IMPRINT_bio(BIO *bp, TS_MSG_IMPRINT **a)
+	{
+	return ASN1_d2i_bio_of(TS_MSG_IMPRINT, TS_MSG_IMPRINT_new, d2i_TS_MSG_IMPRINT, bp, a);
+	}
+
+int i2d_TS_MSG_IMPRINT_bio(BIO *bp, TS_MSG_IMPRINT *a)
+{
+	return ASN1_i2d_bio_of_const(TS_MSG_IMPRINT, i2d_TS_MSG_IMPRINT, bp, a);
+}
+#endif
+#ifndef OPENSSL_NO_FP_API
+TS_MSG_IMPRINT *d2i_TS_MSG_IMPRINT_fp(FILE *fp, TS_MSG_IMPRINT **a)
+	{
+	return ASN1_d2i_fp_of(TS_MSG_IMPRINT, TS_MSG_IMPRINT_new, d2i_TS_MSG_IMPRINT, fp, a);
+	}
+
+int i2d_TS_MSG_IMPRINT_fp(FILE *fp, TS_MSG_IMPRINT *a)
+	{
+	return ASN1_i2d_fp_of_const(TS_MSG_IMPRINT, i2d_TS_MSG_IMPRINT, fp, a);
+	}
+#endif
+
+ASN1_SEQUENCE(TS_REQ) = {
+	ASN1_SIMPLE(TS_REQ, version, ASN1_INTEGER),
+	ASN1_SIMPLE(TS_REQ, msg_imprint, TS_MSG_IMPRINT),
+	ASN1_OPT(TS_REQ, policy_id, ASN1_OBJECT),
+	ASN1_OPT(TS_REQ, nonce, ASN1_INTEGER),
+	ASN1_OPT(TS_REQ, cert_req, ASN1_FBOOLEAN),
+	ASN1_IMP_SEQUENCE_OF_OPT(TS_REQ, extensions, X509_EXTENSION, 0)
+} ASN1_SEQUENCE_END(TS_REQ)
+
+IMPLEMENT_ASN1_FUNCTIONS_const(TS_REQ)
+IMPLEMENT_ASN1_DUP_FUNCTION(TS_REQ)
+#ifndef OPENSSL_NO_BIO
+TS_REQ *d2i_TS_REQ_bio(BIO *bp, TS_REQ **a)
+	{
+	return ASN1_d2i_bio_of(TS_REQ, TS_REQ_new, d2i_TS_REQ, bp, a);
+	}
+
+int i2d_TS_REQ_bio(BIO *bp, TS_REQ *a)
+	{
+	return ASN1_i2d_bio_of_const(TS_REQ, i2d_TS_REQ, bp, a);
+	}
+#endif
+#ifndef OPENSSL_NO_FP_API
+TS_REQ *d2i_TS_REQ_fp(FILE *fp, TS_REQ **a)
+	{
+	return ASN1_d2i_fp_of(TS_REQ, TS_REQ_new, d2i_TS_REQ, fp, a);
+	}
+
+int i2d_TS_REQ_fp(FILE *fp, TS_REQ *a)
+	{
+	return ASN1_i2d_fp_of_const(TS_REQ, i2d_TS_REQ, fp, a);
+	}
+#endif
+
+ASN1_SEQUENCE(TS_ACCURACY) = {
+	ASN1_OPT(TS_ACCURACY, seconds, ASN1_INTEGER),
+	ASN1_IMP_OPT(TS_ACCURACY, millis, ASN1_INTEGER, 0),
+	ASN1_IMP_OPT(TS_ACCURACY, micros, ASN1_INTEGER, 1)
+} ASN1_SEQUENCE_END(TS_ACCURACY)
+
+IMPLEMENT_ASN1_FUNCTIONS_const(TS_ACCURACY)
+IMPLEMENT_ASN1_DUP_FUNCTION(TS_ACCURACY)
+
+ASN1_SEQUENCE(TS_TST_INFO) = {
+	ASN1_SIMPLE(TS_TST_INFO, version, ASN1_INTEGER),
+	ASN1_SIMPLE(TS_TST_INFO, policy_id, ASN1_OBJECT),
+	ASN1_SIMPLE(TS_TST_INFO, msg_imprint, TS_MSG_IMPRINT),
+	ASN1_SIMPLE(TS_TST_INFO, serial, ASN1_INTEGER),
+	ASN1_SIMPLE(TS_TST_INFO, time, ASN1_GENERALIZEDTIME),
+	ASN1_OPT(TS_TST_INFO, accuracy, TS_ACCURACY),
+	ASN1_OPT(TS_TST_INFO, ordering, ASN1_FBOOLEAN),
+	ASN1_OPT(TS_TST_INFO, nonce, ASN1_INTEGER),
+	ASN1_EXP_OPT(TS_TST_INFO, tsa, GENERAL_NAME, 0),
+	ASN1_IMP_SEQUENCE_OF_OPT(TS_TST_INFO, extensions, X509_EXTENSION, 1)
+} ASN1_SEQUENCE_END(TS_TST_INFO)
+
+IMPLEMENT_ASN1_FUNCTIONS_const(TS_TST_INFO)
+IMPLEMENT_ASN1_DUP_FUNCTION(TS_TST_INFO)
+#ifndef OPENSSL_NO_BIO
+TS_TST_INFO *d2i_TS_TST_INFO_bio(BIO *bp, TS_TST_INFO **a)
+	{
+	return ASN1_d2i_bio_of(TS_TST_INFO, TS_TST_INFO_new, d2i_TS_TST_INFO, bp, a);
+	}
+
+int i2d_TS_TST_INFO_bio(BIO *bp, TS_TST_INFO *a)
+	{
+	return ASN1_i2d_bio_of_const(TS_TST_INFO, i2d_TS_TST_INFO, bp, a);
+	}
+#endif
+#ifndef OPENSSL_NO_FP_API
+TS_TST_INFO *d2i_TS_TST_INFO_fp(FILE *fp, TS_TST_INFO **a)
+	{
+	return ASN1_d2i_fp_of(TS_TST_INFO, TS_TST_INFO_new, d2i_TS_TST_INFO, fp, a);
+	}
+
+int i2d_TS_TST_INFO_fp(FILE *fp, TS_TST_INFO *a)
+	{
+	return ASN1_i2d_fp_of_const(TS_TST_INFO, i2d_TS_TST_INFO, fp, a);
+	}
+#endif
+
+ASN1_SEQUENCE(TS_STATUS_INFO) = {
+	ASN1_SIMPLE(TS_STATUS_INFO, status, ASN1_INTEGER),
+	ASN1_SEQUENCE_OF_OPT(TS_STATUS_INFO, text, ASN1_UTF8STRING),
+	ASN1_OPT(TS_STATUS_INFO, failure_info, ASN1_BIT_STRING)
+} ASN1_SEQUENCE_END(TS_STATUS_INFO)
+
+IMPLEMENT_ASN1_FUNCTIONS_const(TS_STATUS_INFO)
+IMPLEMENT_ASN1_DUP_FUNCTION(TS_STATUS_INFO)
+
+static int ts_resp_set_tst_info(TS_RESP *a)
+{
+	long    status;
+
+	status = ASN1_INTEGER_get(a->status_info->status);
+
+	if (a->token) {
+		if (status != 0 && status != 1) {
+			TSerr(TS_F_TS_RESP_SET_TST_INFO, TS_R_TOKEN_PRESENT);
+			return 0;
+		}
+		if (a->tst_info != NULL)
+			TS_TST_INFO_free(a->tst_info);
+		a->tst_info = PKCS7_to_TS_TST_INFO(a->token);
+		if (!a->tst_info) {
+			TSerr(TS_F_TS_RESP_SET_TST_INFO, TS_R_PKCS7_TO_TS_TST_INFO_FAILED);
+			return 0;
+		}
+	} else if (status == 0 || status == 1) {
+		TSerr(TS_F_TS_RESP_SET_TST_INFO, TS_R_TOKEN_NOT_PRESENT);
+		return 0;
+	}
+
+	return 1;
+}
+
+static int ts_resp_cb(int op, ASN1_VALUE **pval, const ASN1_ITEM *it,
+	void *exarg)
+{
+	TS_RESP *ts_resp = (TS_RESP *)*pval;
+	if (op == ASN1_OP_NEW_POST) {
+		ts_resp->tst_info = NULL;
+	} else if (op == ASN1_OP_FREE_POST) {
+		if (ts_resp->tst_info != NULL)
+			TS_TST_INFO_free(ts_resp->tst_info);
+	} else if (op == ASN1_OP_D2I_POST) {
+		if (ts_resp_set_tst_info(ts_resp) == 0)
+			return 0;
+	}
+	return 1;
+}
+
+ASN1_SEQUENCE_cb(TS_RESP, ts_resp_cb) = {
+	ASN1_SIMPLE(TS_RESP, status_info, TS_STATUS_INFO),
+	ASN1_OPT(TS_RESP, token, PKCS7),
+} ASN1_SEQUENCE_END_cb(TS_RESP, TS_RESP)
+
+IMPLEMENT_ASN1_FUNCTIONS_const(TS_RESP)
+IMPLEMENT_ASN1_DUP_FUNCTION(TS_RESP)
+#ifndef OPENSSL_NO_BIO
+TS_RESP *d2i_TS_RESP_bio(BIO *bp, TS_RESP **a)
+	{
+	return ASN1_d2i_bio_of(TS_RESP, TS_RESP_new, d2i_TS_RESP, bp, a);
+	}
+
+int i2d_TS_RESP_bio(BIO *bp, TS_RESP *a)
+	{
+	return ASN1_i2d_bio_of_const(TS_RESP, i2d_TS_RESP, bp, a);
+	}
+#endif
+#ifndef OPENSSL_NO_FP_API
+TS_RESP *d2i_TS_RESP_fp(FILE *fp, TS_RESP **a)
+	{
+	return ASN1_d2i_fp_of(TS_RESP, TS_RESP_new, d2i_TS_RESP, fp, a);
+	}
+
+int i2d_TS_RESP_fp(FILE *fp, TS_RESP *a)
+	{
+	return ASN1_i2d_fp_of_const(TS_RESP, i2d_TS_RESP, fp, a);
+	}
+#endif
+
+ASN1_SEQUENCE(ESS_ISSUER_SERIAL) = {
+	ASN1_SEQUENCE_OF(ESS_ISSUER_SERIAL, issuer, GENERAL_NAME),
+	ASN1_SIMPLE(ESS_ISSUER_SERIAL, serial, ASN1_INTEGER)
+} ASN1_SEQUENCE_END(ESS_ISSUER_SERIAL)
+
+IMPLEMENT_ASN1_FUNCTIONS_const(ESS_ISSUER_SERIAL)
+IMPLEMENT_ASN1_DUP_FUNCTION(ESS_ISSUER_SERIAL)
+
+ASN1_SEQUENCE(ESS_CERT_ID) = {
+	ASN1_SIMPLE(ESS_CERT_ID, hash, ASN1_OCTET_STRING),
+	ASN1_OPT(ESS_CERT_ID, issuer_serial, ESS_ISSUER_SERIAL)
+} ASN1_SEQUENCE_END(ESS_CERT_ID)
+
+IMPLEMENT_ASN1_FUNCTIONS_const(ESS_CERT_ID)
+IMPLEMENT_ASN1_DUP_FUNCTION(ESS_CERT_ID)
+
+ASN1_SEQUENCE(ESS_SIGNING_CERT) = {
+	ASN1_SEQUENCE_OF(ESS_SIGNING_CERT, cert_ids, ESS_CERT_ID),
+	ASN1_SEQUENCE_OF_OPT(ESS_SIGNING_CERT, policy_info, POLICYINFO)
+} ASN1_SEQUENCE_END(ESS_SIGNING_CERT)
+
+IMPLEMENT_ASN1_FUNCTIONS_const(ESS_SIGNING_CERT)
+IMPLEMENT_ASN1_DUP_FUNCTION(ESS_SIGNING_CERT)
+
+/* Getting encapsulated TS_TST_INFO object from PKCS7. */
+TS_TST_INFO *PKCS7_to_TS_TST_INFO(PKCS7 *token)
+{
+	PKCS7_SIGNED *pkcs7_signed;
+	PKCS7 *enveloped;
+	ASN1_TYPE *tst_info_wrapper;
+	ASN1_OCTET_STRING *tst_info_der;
+	const unsigned char *p;
+
+	if (!PKCS7_type_is_signed(token))
+		{
+		TSerr(TS_F_PKCS7_TO_TS_TST_INFO, TS_R_BAD_PKCS7_TYPE);
+		return NULL;
+		}
+
+	/* Content must be present. */
+	if (PKCS7_get_detached(token))
+		{
+		TSerr(TS_F_PKCS7_TO_TS_TST_INFO, TS_R_DETACHED_CONTENT);
+		return NULL;
+		}
+
+	/* We have a signed data with content. */
+	pkcs7_signed = token->d.sign;
+	enveloped = pkcs7_signed->contents;
+	if (OBJ_obj2nid(enveloped->type) != NID_id_smime_ct_TSTInfo)
+		{
+		TSerr(TS_F_PKCS7_TO_TS_TST_INFO, TS_R_BAD_PKCS7_TYPE);
+		return NULL;
+		}
+
+	/* We have a DER encoded TST_INFO as the signed data. */
+	tst_info_wrapper = enveloped->d.other;
+	if (tst_info_wrapper->type != V_ASN1_OCTET_STRING)
+		{
+		TSerr(TS_F_PKCS7_TO_TS_TST_INFO, TS_R_BAD_TYPE);
+		return NULL;
+		}
+
+	/* We have the correct ASN1_OCTET_STRING type. */
+	tst_info_der = tst_info_wrapper->value.octet_string;
+	/* At last, decode the TST_INFO. */
+	p = tst_info_der->data;
+	return d2i_TS_TST_INFO(NULL, &p, tst_info_der->length);
+}
diff --git a/crypto/ts/ts_conf.c b/crypto/ts/ts_conf.c
new file mode 100644
index 000000000000..c39be76f2843
--- /dev/null
+++ b/crypto/ts/ts_conf.c
@@ -0,0 +1,507 @@
+/* crypto/ts/ts_conf.c */
+/* Written by Zoltan Glozik (zglozik@stones.com) for the OpenSSL
+ * project 2002.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <string.h>
+
+#include <openssl/crypto.h>
+#include "cryptlib.h"
+#include <openssl/pem.h>
+#ifndef OPENSSL_NO_ENGINE
+#include <openssl/engine.h>
+#endif
+#include <openssl/ts.h>
+
+/* Macro definitions for the configuration file. */
+
+#define	BASE_SECTION			"tsa"
+#define	ENV_DEFAULT_TSA			"default_tsa"
+#define	ENV_SERIAL			"serial"
+#define ENV_CRYPTO_DEVICE		"crypto_device"
+#define	ENV_SIGNER_CERT			"signer_cert"
+#define	ENV_CERTS			"certs"
+#define	ENV_SIGNER_KEY			"signer_key"
+#define	ENV_DEFAULT_POLICY		"default_policy"
+#define	ENV_OTHER_POLICIES		"other_policies"
+#define	ENV_DIGESTS			"digests"
+#define	ENV_ACCURACY			"accuracy"
+#define	ENV_ORDERING			"ordering"
+#define	ENV_TSA_NAME			"tsa_name"
+#define	ENV_ESS_CERT_ID_CHAIN		"ess_cert_id_chain"
+#define	ENV_VALUE_SECS			"secs"
+#define	ENV_VALUE_MILLISECS		"millisecs"
+#define	ENV_VALUE_MICROSECS		"microsecs"
+#define	ENV_CLOCK_PRECISION_DIGITS	"clock_precision_digits" 
+#define	ENV_VALUE_YES			"yes"
+#define	ENV_VALUE_NO			"no"
+
+/* Function definitions for certificate and key loading. */
+
+X509 *TS_CONF_load_cert(const char *file)
+	{
+	BIO *cert = NULL;
+	X509 *x = NULL;
+
+	if ((cert = BIO_new_file(file, "r")) == NULL) goto end;
+	x = PEM_read_bio_X509_AUX(cert, NULL, NULL, NULL);
+end:
+	if (x == NULL)
+		fprintf(stderr, "unable to load certificate: %s\n", file);
+	BIO_free(cert);
+	return x;
+	}
+
+STACK_OF(X509) *TS_CONF_load_certs(const char *file)
+	{
+	BIO *certs = NULL;
+	STACK_OF(X509) *othercerts = NULL;
+	STACK_OF(X509_INFO) *allcerts = NULL;
+	int i;
+
+	if (!(certs = BIO_new_file(file, "r"))) goto end;
+
+	if (!(othercerts = sk_X509_new_null())) goto end;
+	allcerts = PEM_X509_INFO_read_bio(certs, NULL, NULL, NULL);
+	for(i = 0; i < sk_X509_INFO_num(allcerts); i++)
+		{
+		X509_INFO *xi = sk_X509_INFO_value(allcerts, i);
+		if (xi->x509)
+			{
+			sk_X509_push(othercerts, xi->x509);
+			xi->x509 = NULL;
+			}
+		}
+end:
+	if (othercerts == NULL)
+		fprintf(stderr, "unable to load certificates: %s\n", file);
+	sk_X509_INFO_pop_free(allcerts, X509_INFO_free);
+	BIO_free(certs);
+	return othercerts;
+	}
+
+EVP_PKEY *TS_CONF_load_key(const char *file, const char *pass)
+	{
+	BIO *key = NULL;
+	EVP_PKEY *pkey = NULL;
+
+	if (!(key = BIO_new_file(file, "r"))) goto end;
+	pkey = PEM_read_bio_PrivateKey(key, NULL, NULL, (char *) pass);
+ end:
+	if (pkey == NULL)
+		fprintf(stderr, "unable to load private key: %s\n", file);
+	BIO_free(key);
+	return pkey;
+	}
+
+/* Function definitions for handling configuration options. */
+
+static void TS_CONF_lookup_fail(const char *name, const char *tag)
+	{
+	fprintf(stderr, "variable lookup failed for %s::%s\n", name, tag);
+	}
+
+static void TS_CONF_invalid(const char *name, const char *tag)
+	{
+	fprintf(stderr, "invalid variable value for %s::%s\n", name, tag);
+	}
+
+const char *TS_CONF_get_tsa_section(CONF *conf, const char *section)
+	{
+	if (!section)
+		{
+		section = NCONF_get_string(conf, BASE_SECTION, ENV_DEFAULT_TSA);
+		if (!section)
+			TS_CONF_lookup_fail(BASE_SECTION, ENV_DEFAULT_TSA);
+		}
+	return section;
+	}
+
+int TS_CONF_set_serial(CONF *conf, const char *section, TS_serial_cb cb,
+		       TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	char *serial = NCONF_get_string(conf, section, ENV_SERIAL);
+	if (!serial)
+		{
+		TS_CONF_lookup_fail(section, ENV_SERIAL);
+		goto err;
+		}
+	TS_RESP_CTX_set_serial_cb(ctx, cb, serial);
+
+	ret = 1;
+ err:
+	return ret;
+	}
+
+#ifndef OPENSSL_NO_ENGINE
+
+int TS_CONF_set_crypto_device(CONF *conf, const char *section,
+			      const char *device)
+	{
+	int ret = 0;
+	
+	if (!device)
+		device = NCONF_get_string(conf, section,
+					  ENV_CRYPTO_DEVICE);
+
+	if (device && !TS_CONF_set_default_engine(device))
+		{
+		TS_CONF_invalid(section, ENV_CRYPTO_DEVICE);
+		goto err;
+		}
+	ret = 1;
+ err:
+	return ret;
+	}
+
+int TS_CONF_set_default_engine(const char *name)
+	{
+	ENGINE *e = NULL;
+	int ret = 0;
+
+	/* Leave the default if builtin specified. */
+	if (strcmp(name, "builtin") == 0) return 1;
+
+	if (!(e = ENGINE_by_id(name))) goto err;
+	/* Enable the use of the NCipher HSM for forked children. */
+	if (strcmp(name, "chil") == 0) 
+		ENGINE_ctrl(e, ENGINE_CTRL_CHIL_SET_FORKCHECK, 1, 0, 0);
+	/* All the operations are going to be carried out by the engine. */
+	if (!ENGINE_set_default(e, ENGINE_METHOD_ALL)) goto err;
+	ret = 1;
+ err:
+	if (!ret)
+		{
+		TSerr(TS_F_TS_CONF_SET_DEFAULT_ENGINE, 
+		      TS_R_COULD_NOT_SET_ENGINE);
+		ERR_add_error_data(2, "engine:", name);
+		}
+	if (e) ENGINE_free(e);
+	return ret;
+	}
+
+#endif
+
+int TS_CONF_set_signer_cert(CONF *conf, const char *section,
+			    const char *cert, TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	X509 *cert_obj = NULL;
+	if (!cert) 
+		cert = NCONF_get_string(conf, section, ENV_SIGNER_CERT);
+	if (!cert)
+		{
+		TS_CONF_lookup_fail(section, ENV_SIGNER_CERT);
+		goto err;
+		}
+	if (!(cert_obj = TS_CONF_load_cert(cert)))
+		goto err;
+	if (!TS_RESP_CTX_set_signer_cert(ctx, cert_obj))
+		goto err;
+
+	ret = 1;
+ err:
+	X509_free(cert_obj);
+	return ret;
+	}
+
+int TS_CONF_set_certs(CONF *conf, const char *section, const char *certs,
+		      TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	STACK_OF(X509) *certs_obj = NULL;
+	if (!certs) 
+		certs = NCONF_get_string(conf, section, ENV_CERTS);
+	/* Certificate chain is optional. */
+	if (!certs) goto end;
+	if (!(certs_obj = TS_CONF_load_certs(certs))) goto err;
+	if (!TS_RESP_CTX_set_certs(ctx, certs_obj)) goto err;
+ end:
+	ret = 1;
+ err:
+	sk_X509_pop_free(certs_obj, X509_free);
+	return ret;
+	}
+
+int TS_CONF_set_signer_key(CONF *conf, const char *section,
+			   const char *key, const char *pass,
+			   TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	EVP_PKEY *key_obj = NULL;
+	if (!key) 
+		key = NCONF_get_string(conf, section, ENV_SIGNER_KEY);
+	if (!key)
+		{
+		TS_CONF_lookup_fail(section, ENV_SIGNER_KEY);
+		goto err;
+		}
+	if (!(key_obj = TS_CONF_load_key(key, pass))) goto err;
+	if (!TS_RESP_CTX_set_signer_key(ctx, key_obj)) goto err;
+
+	ret = 1;
+ err:
+	EVP_PKEY_free(key_obj);
+	return ret;
+	}
+
+int TS_CONF_set_def_policy(CONF *conf, const char *section,
+			   const char *policy, TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	ASN1_OBJECT *policy_obj = NULL;
+	if (!policy) 
+		policy = NCONF_get_string(conf, section, 
+					  ENV_DEFAULT_POLICY);
+	if (!policy)
+		{
+		TS_CONF_lookup_fail(section, ENV_DEFAULT_POLICY);
+		goto err;
+		}
+	if (!(policy_obj = OBJ_txt2obj(policy, 0)))
+		{
+		TS_CONF_invalid(section, ENV_DEFAULT_POLICY);
+		goto err;
+		}
+	if (!TS_RESP_CTX_set_def_policy(ctx, policy_obj))
+		goto err;
+
+	ret = 1;
+ err:
+	ASN1_OBJECT_free(policy_obj);
+	return ret;
+	}
+
+int TS_CONF_set_policies(CONF *conf, const char *section,
+			 TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	int i;
+	STACK_OF(CONF_VALUE) *list = NULL;
+	char *policies = NCONF_get_string(conf, section, 
+					  ENV_OTHER_POLICIES);
+	/* If no other policy is specified, that's fine. */
+	if (policies && !(list = X509V3_parse_list(policies)))
+		{
+		TS_CONF_invalid(section, ENV_OTHER_POLICIES);
+		goto err;
+		}
+	for (i = 0; i < sk_CONF_VALUE_num(list); ++i)
+		{
+		CONF_VALUE *val = sk_CONF_VALUE_value(list, i);
+		const char *extval = val->value ? val->value : val->name;
+		ASN1_OBJECT *objtmp;
+		if (!(objtmp = OBJ_txt2obj(extval, 0)))
+			{
+			TS_CONF_invalid(section, ENV_OTHER_POLICIES);
+			goto err;
+			}
+		if (!TS_RESP_CTX_add_policy(ctx, objtmp))
+			goto err;
+		ASN1_OBJECT_free(objtmp);
+		}
+
+	ret = 1;
+ err:
+	sk_CONF_VALUE_pop_free(list, X509V3_conf_free);
+	return ret;
+	}
+
+int TS_CONF_set_digests(CONF *conf, const char *section,
+			TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	int i;
+	STACK_OF(CONF_VALUE) *list = NULL;
+	char *digests = NCONF_get_string(conf, section, ENV_DIGESTS);
+	if (!digests)
+		{
+		TS_CONF_lookup_fail(section, ENV_DIGESTS);
+		goto err;
+		}
+	if (!(list = X509V3_parse_list(digests)))
+		{
+		TS_CONF_invalid(section, ENV_DIGESTS);
+		goto err;
+		}
+	if (sk_CONF_VALUE_num(list) == 0)
+		{
+		TS_CONF_invalid(section, ENV_DIGESTS);
+		goto err;
+		}
+	for (i = 0; i < sk_CONF_VALUE_num(list); ++i)
+		{
+		CONF_VALUE *val = sk_CONF_VALUE_value(list, i);
+		const char *extval = val->value ? val->value : val->name;
+		const EVP_MD *md;
+		if (!(md = EVP_get_digestbyname(extval)))
+			{
+			TS_CONF_invalid(section, ENV_DIGESTS);
+			goto err;
+			}
+		if (!TS_RESP_CTX_add_md(ctx, md))
+			goto err;
+		}
+
+	ret = 1;
+ err:
+	sk_CONF_VALUE_pop_free(list, X509V3_conf_free);
+	return ret;
+	}
+
+int TS_CONF_set_accuracy(CONF *conf, const char *section, TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	int i;
+	int secs = 0, millis = 0, micros = 0;
+	STACK_OF(CONF_VALUE) *list = NULL;
+	char *accuracy = NCONF_get_string(conf, section, ENV_ACCURACY);
+
+	if (accuracy && !(list = X509V3_parse_list(accuracy)))
+		{
+		TS_CONF_invalid(section, ENV_ACCURACY);
+		goto err;
+		}
+	for (i = 0; i < sk_CONF_VALUE_num(list); ++i)
+		{
+		CONF_VALUE *val = sk_CONF_VALUE_value(list, i);
+		if (strcmp(val->name, ENV_VALUE_SECS) == 0) 
+			{
+			if (val->value) secs = atoi(val->value);
+			}
+		else if (strcmp(val->name, ENV_VALUE_MILLISECS) == 0)
+			{
+			if (val->value) millis = atoi(val->value);
+			}
+		else if (strcmp(val->name, ENV_VALUE_MICROSECS) == 0)
+			{
+			if (val->value) micros = atoi(val->value);
+			}
+		else
+			{
+			TS_CONF_invalid(section, ENV_ACCURACY);
+			goto err;
+			}
+		}
+	if (!TS_RESP_CTX_set_accuracy(ctx, secs, millis, micros))
+		goto err;
+
+	ret = 1;
+ err:
+	sk_CONF_VALUE_pop_free(list, X509V3_conf_free);
+	return ret;
+	}
+
+int TS_CONF_set_clock_precision_digits(CONF *conf, const char *section,
+				       TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	long digits = 0;
+	
+	/* If not specified, set the default value to 0, i.e. sec  precision */
+	if (!NCONF_get_number_e(conf, section, ENV_CLOCK_PRECISION_DIGITS,
+				&digits))
+		digits = 0;
+	if (digits < 0 || digits > TS_MAX_CLOCK_PRECISION_DIGITS)
+		{
+		TS_CONF_invalid(section, ENV_CLOCK_PRECISION_DIGITS);
+		goto err;
+		}
+
+	if (!TS_RESP_CTX_set_clock_precision_digits(ctx, digits))
+		goto err;
+
+	return 1;
+ err:
+	return ret;
+	}
+
+static int TS_CONF_add_flag(CONF *conf, const char *section, const char *field,
+			    int flag, TS_RESP_CTX *ctx)
+	{
+	/* Default is false. */
+	const char *value = NCONF_get_string(conf, section, field);
+	if (value)
+		{
+		if (strcmp(value, ENV_VALUE_YES) == 0)
+			TS_RESP_CTX_add_flags(ctx, flag);
+		else if (strcmp(value, ENV_VALUE_NO) != 0)
+			{
+			TS_CONF_invalid(section, field);
+			return 0;
+			}
+		}
+
+	return 1;
+	}
+
+int TS_CONF_set_ordering(CONF *conf, const char *section, TS_RESP_CTX *ctx)
+	{
+	return TS_CONF_add_flag(conf, section, ENV_ORDERING, TS_ORDERING, ctx);
+	}
+
+int TS_CONF_set_tsa_name(CONF *conf, const char *section, TS_RESP_CTX *ctx)
+	{
+	return TS_CONF_add_flag(conf, section, ENV_TSA_NAME, TS_TSA_NAME, ctx);
+	}
+
+int TS_CONF_set_ess_cert_id_chain(CONF *conf, const char *section,
+				  TS_RESP_CTX *ctx)
+	{
+	return TS_CONF_add_flag(conf, section, ENV_ESS_CERT_ID_CHAIN, 
+				TS_ESS_CERT_ID_CHAIN, ctx);
+	}
diff --git a/crypto/ts/ts_err.c b/crypto/ts/ts_err.c
new file mode 100644
index 000000000000..a08b0ffa234c
--- /dev/null
+++ b/crypto/ts/ts_err.c
@@ -0,0 +1,179 @@
+/* crypto/ts/ts_err.c */
+/* ====================================================================
+ * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+/* NOTE: this file was auto generated by the mkerr.pl script: any changes
+ * made to it will be overwritten when the script next updates this file,
+ * only reason strings will be preserved.
+ */
+
+#include <stdio.h>
+#include <openssl/err.h>
+#include <openssl/ts.h>
+
+/* BEGIN ERROR CODES */
+#ifndef OPENSSL_NO_ERR
+
+#define ERR_FUNC(func) ERR_PACK(ERR_LIB_TS,func,0)
+#define ERR_REASON(reason) ERR_PACK(ERR_LIB_TS,0,reason)
+
+static ERR_STRING_DATA TS_str_functs[]=
+	{
+{ERR_FUNC(TS_F_D2I_TS_RESP),	"d2i_TS_RESP"},
+{ERR_FUNC(TS_F_DEF_SERIAL_CB),	"DEF_SERIAL_CB"},
+{ERR_FUNC(TS_F_DEF_TIME_CB),	"DEF_TIME_CB"},
+{ERR_FUNC(TS_F_ESS_ADD_SIGNING_CERT),	"ESS_ADD_SIGNING_CERT"},
+{ERR_FUNC(TS_F_ESS_CERT_ID_NEW_INIT),	"ESS_CERT_ID_NEW_INIT"},
+{ERR_FUNC(TS_F_ESS_SIGNING_CERT_NEW_INIT),	"ESS_SIGNING_CERT_NEW_INIT"},
+{ERR_FUNC(TS_F_INT_TS_RESP_VERIFY_TOKEN),	"INT_TS_RESP_VERIFY_TOKEN"},
+{ERR_FUNC(TS_F_PKCS7_TO_TS_TST_INFO),	"PKCS7_to_TS_TST_INFO"},
+{ERR_FUNC(TS_F_TS_ACCURACY_SET_MICROS),	"TS_ACCURACY_set_micros"},
+{ERR_FUNC(TS_F_TS_ACCURACY_SET_MILLIS),	"TS_ACCURACY_set_millis"},
+{ERR_FUNC(TS_F_TS_ACCURACY_SET_SECONDS),	"TS_ACCURACY_set_seconds"},
+{ERR_FUNC(TS_F_TS_CHECK_IMPRINTS),	"TS_CHECK_IMPRINTS"},
+{ERR_FUNC(TS_F_TS_CHECK_NONCES),	"TS_CHECK_NONCES"},
+{ERR_FUNC(TS_F_TS_CHECK_POLICY),	"TS_CHECK_POLICY"},
+{ERR_FUNC(TS_F_TS_CHECK_SIGNING_CERTS),	"TS_CHECK_SIGNING_CERTS"},
+{ERR_FUNC(TS_F_TS_CHECK_STATUS_INFO),	"TS_CHECK_STATUS_INFO"},
+{ERR_FUNC(TS_F_TS_COMPUTE_IMPRINT),	"TS_COMPUTE_IMPRINT"},
+{ERR_FUNC(TS_F_TS_CONF_SET_DEFAULT_ENGINE),	"TS_CONF_set_default_engine"},
+{ERR_FUNC(TS_F_TS_GET_STATUS_TEXT),	"TS_GET_STATUS_TEXT"},
+{ERR_FUNC(TS_F_TS_MSG_IMPRINT_SET_ALGO),	"TS_MSG_IMPRINT_set_algo"},
+{ERR_FUNC(TS_F_TS_REQ_SET_MSG_IMPRINT),	"TS_REQ_set_msg_imprint"},
+{ERR_FUNC(TS_F_TS_REQ_SET_NONCE),	"TS_REQ_set_nonce"},
+{ERR_FUNC(TS_F_TS_REQ_SET_POLICY_ID),	"TS_REQ_set_policy_id"},
+{ERR_FUNC(TS_F_TS_RESP_CREATE_RESPONSE),	"TS_RESP_create_response"},
+{ERR_FUNC(TS_F_TS_RESP_CREATE_TST_INFO),	"TS_RESP_CREATE_TST_INFO"},
+{ERR_FUNC(TS_F_TS_RESP_CTX_ADD_FAILURE_INFO),	"TS_RESP_CTX_add_failure_info"},
+{ERR_FUNC(TS_F_TS_RESP_CTX_ADD_MD),	"TS_RESP_CTX_add_md"},
+{ERR_FUNC(TS_F_TS_RESP_CTX_ADD_POLICY),	"TS_RESP_CTX_add_policy"},
+{ERR_FUNC(TS_F_TS_RESP_CTX_NEW),	"TS_RESP_CTX_new"},
+{ERR_FUNC(TS_F_TS_RESP_CTX_SET_ACCURACY),	"TS_RESP_CTX_set_accuracy"},
+{ERR_FUNC(TS_F_TS_RESP_CTX_SET_CERTS),	"TS_RESP_CTX_set_certs"},
+{ERR_FUNC(TS_F_TS_RESP_CTX_SET_DEF_POLICY),	"TS_RESP_CTX_set_def_policy"},
+{ERR_FUNC(TS_F_TS_RESP_CTX_SET_SIGNER_CERT),	"TS_RESP_CTX_set_signer_cert"},
+{ERR_FUNC(TS_F_TS_RESP_CTX_SET_STATUS_INFO),	"TS_RESP_CTX_set_status_info"},
+{ERR_FUNC(TS_F_TS_RESP_GET_POLICY),	"TS_RESP_GET_POLICY"},
+{ERR_FUNC(TS_F_TS_RESP_SET_GENTIME_WITH_PRECISION),	"TS_RESP_SET_GENTIME_WITH_PRECISION"},
+{ERR_FUNC(TS_F_TS_RESP_SET_STATUS_INFO),	"TS_RESP_set_status_info"},
+{ERR_FUNC(TS_F_TS_RESP_SET_TST_INFO),	"TS_RESP_set_tst_info"},
+{ERR_FUNC(TS_F_TS_RESP_SIGN),	"TS_RESP_SIGN"},
+{ERR_FUNC(TS_F_TS_RESP_VERIFY_SIGNATURE),	"TS_RESP_verify_signature"},
+{ERR_FUNC(TS_F_TS_RESP_VERIFY_TOKEN),	"TS_RESP_verify_token"},
+{ERR_FUNC(TS_F_TS_TST_INFO_SET_ACCURACY),	"TS_TST_INFO_set_accuracy"},
+{ERR_FUNC(TS_F_TS_TST_INFO_SET_MSG_IMPRINT),	"TS_TST_INFO_set_msg_imprint"},
+{ERR_FUNC(TS_F_TS_TST_INFO_SET_NONCE),	"TS_TST_INFO_set_nonce"},
+{ERR_FUNC(TS_F_TS_TST_INFO_SET_POLICY_ID),	"TS_TST_INFO_set_policy_id"},
+{ERR_FUNC(TS_F_TS_TST_INFO_SET_SERIAL),	"TS_TST_INFO_set_serial"},
+{ERR_FUNC(TS_F_TS_TST_INFO_SET_TIME),	"TS_TST_INFO_set_time"},
+{ERR_FUNC(TS_F_TS_TST_INFO_SET_TSA),	"TS_TST_INFO_set_tsa"},
+{ERR_FUNC(TS_F_TS_VERIFY),	"TS_VERIFY"},
+{ERR_FUNC(TS_F_TS_VERIFY_CERT),	"TS_VERIFY_CERT"},
+{ERR_FUNC(TS_F_TS_VERIFY_CTX_NEW),	"TS_VERIFY_CTX_new"},
+{0,NULL}
+	};
+
+static ERR_STRING_DATA TS_str_reasons[]=
+	{
+{ERR_REASON(TS_R_BAD_PKCS7_TYPE)         ,"bad pkcs7 type"},
+{ERR_REASON(TS_R_BAD_TYPE)               ,"bad type"},
+{ERR_REASON(TS_R_CERTIFICATE_VERIFY_ERROR),"certificate verify error"},
+{ERR_REASON(TS_R_COULD_NOT_SET_ENGINE)   ,"could not set engine"},
+{ERR_REASON(TS_R_COULD_NOT_SET_TIME)     ,"could not set time"},
+{ERR_REASON(TS_R_D2I_TS_RESP_INT_FAILED) ,"d2i ts resp int failed"},
+{ERR_REASON(TS_R_DETACHED_CONTENT)       ,"detached content"},
+{ERR_REASON(TS_R_ESS_ADD_SIGNING_CERT_ERROR),"ess add signing cert error"},
+{ERR_REASON(TS_R_ESS_SIGNING_CERTIFICATE_ERROR),"ess signing certificate error"},
+{ERR_REASON(TS_R_INVALID_NULL_POINTER)   ,"invalid null pointer"},
+{ERR_REASON(TS_R_INVALID_SIGNER_CERTIFICATE_PURPOSE),"invalid signer certificate purpose"},
+{ERR_REASON(TS_R_MESSAGE_IMPRINT_MISMATCH),"message imprint mismatch"},
+{ERR_REASON(TS_R_NONCE_MISMATCH)         ,"nonce mismatch"},
+{ERR_REASON(TS_R_NONCE_NOT_RETURNED)     ,"nonce not returned"},
+{ERR_REASON(TS_R_NO_CONTENT)             ,"no content"},
+{ERR_REASON(TS_R_NO_TIME_STAMP_TOKEN)    ,"no time stamp token"},
+{ERR_REASON(TS_R_PKCS7_ADD_SIGNATURE_ERROR),"pkcs7 add signature error"},
+{ERR_REASON(TS_R_PKCS7_ADD_SIGNED_ATTR_ERROR),"pkcs7 add signed attr error"},
+{ERR_REASON(TS_R_PKCS7_TO_TS_TST_INFO_FAILED),"pkcs7 to ts tst info failed"},
+{ERR_REASON(TS_R_POLICY_MISMATCH)        ,"policy mismatch"},
+{ERR_REASON(TS_R_PRIVATE_KEY_DOES_NOT_MATCH_CERTIFICATE),"private key does not match certificate"},
+{ERR_REASON(TS_R_RESPONSE_SETUP_ERROR)   ,"response setup error"},
+{ERR_REASON(TS_R_SIGNATURE_FAILURE)      ,"signature failure"},
+{ERR_REASON(TS_R_THERE_MUST_BE_ONE_SIGNER),"there must be one signer"},
+{ERR_REASON(TS_R_TIME_SYSCALL_ERROR)     ,"time syscall error"},
+{ERR_REASON(TS_R_TOKEN_NOT_PRESENT)      ,"token not present"},
+{ERR_REASON(TS_R_TOKEN_PRESENT)          ,"token present"},
+{ERR_REASON(TS_R_TSA_NAME_MISMATCH)      ,"tsa name mismatch"},
+{ERR_REASON(TS_R_TSA_UNTRUSTED)          ,"tsa untrusted"},
+{ERR_REASON(TS_R_TST_INFO_SETUP_ERROR)   ,"tst info setup error"},
+{ERR_REASON(TS_R_TS_DATASIGN)            ,"ts datasign"},
+{ERR_REASON(TS_R_UNACCEPTABLE_POLICY)    ,"unacceptable policy"},
+{ERR_REASON(TS_R_UNSUPPORTED_MD_ALGORITHM),"unsupported md algorithm"},
+{ERR_REASON(TS_R_UNSUPPORTED_VERSION)    ,"unsupported version"},
+{ERR_REASON(TS_R_WRONG_CONTENT_TYPE)     ,"wrong content type"},
+{0,NULL}
+	};
+
+#endif
+
+void ERR_load_TS_strings(void)
+	{
+#ifndef OPENSSL_NO_ERR
+
+	if (ERR_func_error_string(TS_str_functs[0].error) == NULL)
+		{
+		ERR_load_strings(0,TS_str_functs);
+		ERR_load_strings(0,TS_str_reasons);
+		}
+#endif
+	}
diff --git a/crypto/ts/ts_lib.c b/crypto/ts/ts_lib.c
new file mode 100644
index 000000000000..e8608dbf711e
--- /dev/null
+++ b/crypto/ts/ts_lib.c
@@ -0,0 +1,145 @@
+/* crypto/ts/ts_lib.c */
+/* Written by Zoltan Glozik (zglozik@stones.com) for the OpenSSL
+ * project 2002.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/bn.h>
+#include <openssl/x509v3.h>
+#include "ts.h"
+
+/* Local function declarations. */
+
+/* Function definitions. */
+
+int TS_ASN1_INTEGER_print_bio(BIO *bio, const ASN1_INTEGER *num)
+	{
+	BIGNUM num_bn;
+	int result = 0;
+	char *hex;
+
+	BN_init(&num_bn);
+	ASN1_INTEGER_to_BN(num, &num_bn);
+	if ((hex = BN_bn2hex(&num_bn))) 
+		{
+		result = BIO_write(bio, "0x", 2) > 0;
+		result = result && BIO_write(bio, hex, strlen(hex)) > 0;
+		OPENSSL_free(hex);
+		}
+	BN_free(&num_bn);
+
+	return result;
+	}
+
+int TS_OBJ_print_bio(BIO *bio, const ASN1_OBJECT *obj)
+	{
+	char obj_txt[128];
+
+	int len = OBJ_obj2txt(obj_txt, sizeof(obj_txt), obj, 0);
+	BIO_write(bio, obj_txt, len);
+	BIO_write(bio, "\n", 1);
+
+	return 1;
+	}
+
+int TS_ext_print_bio(BIO *bio, const STACK_OF(X509_EXTENSION) *extensions)
+	{
+	int i, critical, n;
+	X509_EXTENSION *ex;
+	ASN1_OBJECT *obj;
+
+	BIO_printf(bio, "Extensions:\n");
+	n = X509v3_get_ext_count(extensions);
+	for (i = 0; i < n; i++)
+		{
+		ex = X509v3_get_ext(extensions, i);
+		obj = X509_EXTENSION_get_object(ex);
+		i2a_ASN1_OBJECT(bio, obj);
+		critical = X509_EXTENSION_get_critical(ex);
+		BIO_printf(bio, ": %s\n", critical ? "critical" : "");
+		if (!X509V3_EXT_print(bio, ex, 0, 4))
+			{
+			BIO_printf(bio, "%4s", "");
+			M_ASN1_OCTET_STRING_print(bio, ex->value);
+			}
+		BIO_write(bio, "\n", 1);
+		}
+
+	return 1;
+	}
+
+int TS_X509_ALGOR_print_bio(BIO *bio, const X509_ALGOR *alg)
+	{
+	int i = OBJ_obj2nid(alg->algorithm);
+	return BIO_printf(bio, "Hash Algorithm: %s\n",
+		(i == NID_undef) ? "UNKNOWN" : OBJ_nid2ln(i));
+	}
+
+int TS_MSG_IMPRINT_print_bio(BIO *bio, TS_MSG_IMPRINT *a)
+	{
+	const ASN1_OCTET_STRING *msg;
+
+	TS_X509_ALGOR_print_bio(bio, TS_MSG_IMPRINT_get_algo(a));
+
+	BIO_printf(bio, "Message data:\n");
+	msg = TS_MSG_IMPRINT_get_msg(a);
+	BIO_dump_indent(bio, (const char *)M_ASN1_STRING_data(msg), 
+			M_ASN1_STRING_length(msg), 4);
+
+	return 1;
+	}
diff --git a/crypto/ts/ts_req_print.c b/crypto/ts/ts_req_print.c
new file mode 100644
index 000000000000..eba12c382408
--- /dev/null
+++ b/crypto/ts/ts_req_print.c
@@ -0,0 +1,102 @@
+/* crypto/ts/ts_req_print.c */
+/* Written by Zoltan Glozik (zglozik@stones.com) for the OpenSSL
+ * project 2002.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/bn.h>
+#include <openssl/x509v3.h>
+#include <openssl/ts.h>
+
+/* Function definitions. */
+
+int TS_REQ_print_bio(BIO *bio, TS_REQ *a)
+	{
+	int v;
+	ASN1_OBJECT *policy_id;
+	const ASN1_INTEGER *nonce;
+
+	if (a == NULL) return 0;
+
+	v = TS_REQ_get_version(a);
+	BIO_printf(bio, "Version: %d\n", v);
+
+	TS_MSG_IMPRINT_print_bio(bio, TS_REQ_get_msg_imprint(a));
+
+	BIO_printf(bio, "Policy OID: ");
+	policy_id = TS_REQ_get_policy_id(a);
+	if (policy_id == NULL)
+		BIO_printf(bio, "unspecified\n");
+	else	
+		TS_OBJ_print_bio(bio, policy_id);
+
+	BIO_printf(bio, "Nonce: ");
+	nonce = TS_REQ_get_nonce(a);
+	if (nonce == NULL)
+		BIO_printf(bio, "unspecified");
+	else
+		TS_ASN1_INTEGER_print_bio(bio, nonce);
+	BIO_write(bio, "\n", 1);
+
+	BIO_printf(bio, "Certificate required: %s\n", 
+		   TS_REQ_get_cert_req(a) ? "yes" : "no");
+
+	TS_ext_print_bio(bio, TS_REQ_get_exts(a));
+
+	return 1;
+	}
diff --git a/crypto/ts/ts_req_utils.c b/crypto/ts/ts_req_utils.c
new file mode 100644
index 000000000000..43280c15874e
--- /dev/null
+++ b/crypto/ts/ts_req_utils.c
@@ -0,0 +1,234 @@
+/* crypto/ts/ts_req_utils.c */
+/* Written by Zoltan Glozik (zglozik@stones.com) for the OpenSSL
+ * project 2002.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/x509v3.h>
+#include <openssl/ts.h>
+
+int TS_REQ_set_version(TS_REQ *a, long version)
+	{
+	return ASN1_INTEGER_set(a->version, version);
+	}
+
+long TS_REQ_get_version(const TS_REQ *a)
+	{
+	return ASN1_INTEGER_get(a->version);
+	}
+
+int TS_REQ_set_msg_imprint(TS_REQ *a, TS_MSG_IMPRINT *msg_imprint)
+	{
+	TS_MSG_IMPRINT *new_msg_imprint;
+
+	if (a->msg_imprint == msg_imprint)
+		return 1;
+	new_msg_imprint = TS_MSG_IMPRINT_dup(msg_imprint);
+	if (new_msg_imprint == NULL)
+		{
+		TSerr(TS_F_TS_REQ_SET_MSG_IMPRINT, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	TS_MSG_IMPRINT_free(a->msg_imprint);
+	a->msg_imprint = new_msg_imprint;
+	return 1;
+	}
+
+TS_MSG_IMPRINT *TS_REQ_get_msg_imprint(TS_REQ *a)
+	{
+	return a->msg_imprint;
+	}
+
+int TS_MSG_IMPRINT_set_algo(TS_MSG_IMPRINT *a, X509_ALGOR *alg)
+	{
+	X509_ALGOR *new_alg;
+
+	if (a->hash_algo == alg)
+		return 1;
+	new_alg = X509_ALGOR_dup(alg);
+	if (new_alg == NULL)
+		{
+		TSerr(TS_F_TS_MSG_IMPRINT_SET_ALGO, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	X509_ALGOR_free(a->hash_algo);
+	a->hash_algo = new_alg;
+	return 1;
+	}
+
+X509_ALGOR *TS_MSG_IMPRINT_get_algo(TS_MSG_IMPRINT *a)
+	{
+	return a->hash_algo;
+	}
+
+int TS_MSG_IMPRINT_set_msg(TS_MSG_IMPRINT *a, unsigned char *d, int len)
+	{
+	return ASN1_OCTET_STRING_set(a->hashed_msg, d, len);
+	}
+
+ASN1_OCTET_STRING *TS_MSG_IMPRINT_get_msg(TS_MSG_IMPRINT *a)
+	{
+	return a->hashed_msg;
+	}
+
+int TS_REQ_set_policy_id(TS_REQ *a, ASN1_OBJECT *policy)
+	{
+	ASN1_OBJECT *new_policy;
+
+	if (a->policy_id == policy)
+		return 1;
+	new_policy = OBJ_dup(policy);
+	if (new_policy == NULL)
+		{
+		TSerr(TS_F_TS_REQ_SET_POLICY_ID, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	ASN1_OBJECT_free(a->policy_id);
+	a->policy_id = new_policy;
+	return 1;
+	}
+
+ASN1_OBJECT *TS_REQ_get_policy_id(TS_REQ *a)
+	{
+	return a->policy_id;
+	}
+
+int TS_REQ_set_nonce(TS_REQ *a, const ASN1_INTEGER *nonce)
+	{
+	ASN1_INTEGER *new_nonce;
+
+	if (a->nonce == nonce)
+		return 1;
+	new_nonce = ASN1_INTEGER_dup(nonce);
+	if (new_nonce == NULL)
+		{
+		TSerr(TS_F_TS_REQ_SET_NONCE, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	ASN1_INTEGER_free(a->nonce);
+	a->nonce = new_nonce;
+	return 1;
+	}
+
+const ASN1_INTEGER *TS_REQ_get_nonce(const TS_REQ *a)
+	{
+	return a->nonce;
+	}
+
+int TS_REQ_set_cert_req(TS_REQ *a, int cert_req)
+	{
+	a->cert_req = cert_req ? 0xFF : 0x00;
+	return 1;
+	}
+
+int TS_REQ_get_cert_req(const TS_REQ *a)
+	{
+	return a->cert_req ? 1 : 0;
+	}
+
+STACK_OF(X509_EXTENSION) *TS_REQ_get_exts(TS_REQ *a)
+	{
+	return a->extensions;
+	}
+
+void TS_REQ_ext_free(TS_REQ *a)
+	{
+	if (!a) return;
+	sk_X509_EXTENSION_pop_free(a->extensions, X509_EXTENSION_free);
+	a->extensions = NULL;
+	}
+
+int TS_REQ_get_ext_count(TS_REQ *a)
+	{
+	return X509v3_get_ext_count(a->extensions);
+	}
+
+int TS_REQ_get_ext_by_NID(TS_REQ *a, int nid, int lastpos)
+	{
+	return X509v3_get_ext_by_NID(a->extensions, nid, lastpos);
+	}
+
+int TS_REQ_get_ext_by_OBJ(TS_REQ *a, ASN1_OBJECT *obj, int lastpos)
+	{
+	return X509v3_get_ext_by_OBJ(a->extensions, obj, lastpos);
+	}
+
+int TS_REQ_get_ext_by_critical(TS_REQ *a, int crit, int lastpos)
+	{
+	return X509v3_get_ext_by_critical(a->extensions, crit, lastpos);
+	}
+
+X509_EXTENSION *TS_REQ_get_ext(TS_REQ *a, int loc)
+	{
+	return X509v3_get_ext(a->extensions,loc);
+	}
+
+X509_EXTENSION *TS_REQ_delete_ext(TS_REQ *a, int loc)
+	{
+	return X509v3_delete_ext(a->extensions,loc);
+	}
+
+int TS_REQ_add_ext(TS_REQ *a, X509_EXTENSION *ex, int loc)
+	{
+	return X509v3_add_ext(&a->extensions,ex,loc) != NULL;
+	}
+
+void *TS_REQ_get_ext_d2i(TS_REQ *a, int nid, int *crit, int *idx)
+	{
+	return X509V3_get_d2i(a->extensions, nid, crit, idx);
+	}
diff --git a/crypto/ts/ts_rsp_print.c b/crypto/ts/ts_rsp_print.c
new file mode 100644
index 000000000000..21062517ba98
--- /dev/null
+++ b/crypto/ts/ts_rsp_print.c
@@ -0,0 +1,287 @@
+/* crypto/ts/ts_resp_print.c */
+/* Written by Zoltan Glozik (zglozik@stones.com) for the OpenSSL
+ * project 2002.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/bn.h>
+#include <openssl/x509v3.h>
+#include "ts.h"
+
+struct status_map_st
+	{
+	int bit;
+	const char *text;
+	};
+
+/* Local function declarations. */
+
+static int TS_status_map_print(BIO *bio, struct status_map_st *a,
+			       ASN1_BIT_STRING *v);
+static int TS_ACCURACY_print_bio(BIO *bio, const TS_ACCURACY *accuracy);
+
+/* Function definitions. */
+
+int TS_RESP_print_bio(BIO *bio, TS_RESP *a)
+	{
+	TS_TST_INFO *tst_info;
+
+	BIO_printf(bio, "Status info:\n");
+	TS_STATUS_INFO_print_bio(bio, TS_RESP_get_status_info(a));
+
+	BIO_printf(bio, "\nTST info:\n");
+	tst_info = TS_RESP_get_tst_info(a);
+	if (tst_info != NULL)
+		TS_TST_INFO_print_bio(bio, TS_RESP_get_tst_info(a));
+	else
+		BIO_printf(bio, "Not included.\n");
+		
+	return 1;
+	}
+
+int TS_STATUS_INFO_print_bio(BIO *bio, TS_STATUS_INFO *a)
+	{
+	static const char *status_map[] =
+		{
+		"Granted.",
+		"Granted with modifications.",
+		"Rejected.",
+		"Waiting.",
+		"Revocation warning.",
+		"Revoked."
+		};
+	static struct status_map_st failure_map[] =
+		{
+		{ TS_INFO_BAD_ALG,
+		"unrecognized or unsupported algorithm identifier" },
+		{ TS_INFO_BAD_REQUEST,
+		"transaction not permitted or supported" },
+		{ TS_INFO_BAD_DATA_FORMAT,
+		"the data submitted has the wrong format" },
+		{ TS_INFO_TIME_NOT_AVAILABLE,
+		"the TSA's time source is not available" },
+		{ TS_INFO_UNACCEPTED_POLICY,
+		"the requested TSA policy is not supported by the TSA" },
+		{ TS_INFO_UNACCEPTED_EXTENSION,
+		"the requested extension is not supported by the TSA" },
+		{ TS_INFO_ADD_INFO_NOT_AVAILABLE,
+		"the additional information requested could not be understood "
+		"or is not available" },
+		{ TS_INFO_SYSTEM_FAILURE,
+		"the request cannot be handled due to system failure" },
+		{ -1, NULL }
+		};
+	long status;
+	int i, lines = 0;
+
+	/* Printing status code. */
+	BIO_printf(bio, "Status: ");
+	status = ASN1_INTEGER_get(a->status);
+	if (0 <= status && status < (long)(sizeof(status_map)/sizeof(status_map[0])))
+		BIO_printf(bio, "%s\n", status_map[status]);
+	else
+		BIO_printf(bio, "out of bounds\n");
+	
+	/* Printing status description. */
+	BIO_printf(bio, "Status description: ");
+	for (i = 0; i < sk_ASN1_UTF8STRING_num(a->text); ++i)
+		{
+		if (i > 0)
+			BIO_puts(bio, "\t");
+		ASN1_STRING_print_ex(bio, sk_ASN1_UTF8STRING_value(a->text, i),
+				     0);
+		BIO_puts(bio, "\n");
+		}
+	if (i == 0)
+		BIO_printf(bio, "unspecified\n");
+
+	/* Printing failure information. */
+	BIO_printf(bio, "Failure info: ");
+	if (a->failure_info != NULL)
+		lines = TS_status_map_print(bio, failure_map,
+					    a->failure_info);
+	if (lines == 0)
+		BIO_printf(bio, "unspecified");
+	BIO_printf(bio, "\n");
+
+	return 1;
+	}
+
+static int TS_status_map_print(BIO *bio, struct status_map_st *a,
+			       ASN1_BIT_STRING *v)
+	{
+	int lines = 0;
+
+	for (; a->bit >= 0; ++a)
+		{
+		if (ASN1_BIT_STRING_get_bit(v, a->bit))
+			{
+			if (++lines > 1)
+				BIO_printf(bio, ", ");
+			BIO_printf(bio, "%s", a->text);
+			}
+		}
+
+	return lines;
+	}
+
+int TS_TST_INFO_print_bio(BIO *bio, TS_TST_INFO *a)
+	{
+	int v;
+	ASN1_OBJECT *policy_id;
+	const ASN1_INTEGER *serial;
+	const ASN1_GENERALIZEDTIME *gtime;
+	TS_ACCURACY *accuracy;
+	const ASN1_INTEGER *nonce;
+	GENERAL_NAME *tsa_name;
+
+	if (a == NULL) return 0;
+
+	/* Print version. */
+	v = TS_TST_INFO_get_version(a);
+	BIO_printf(bio, "Version: %d\n", v);
+
+	/* Print policy id. */
+	BIO_printf(bio, "Policy OID: ");
+	policy_id = TS_TST_INFO_get_policy_id(a);
+	TS_OBJ_print_bio(bio, policy_id);
+
+	/* Print message imprint. */
+	TS_MSG_IMPRINT_print_bio(bio, TS_TST_INFO_get_msg_imprint(a));
+
+	/* Print serial number. */
+	BIO_printf(bio, "Serial number: ");
+	serial = TS_TST_INFO_get_serial(a);
+	if (serial == NULL)
+		BIO_printf(bio, "unspecified");
+	else
+		TS_ASN1_INTEGER_print_bio(bio, serial);
+	BIO_write(bio, "\n", 1);
+
+	/* Print time stamp. */
+	BIO_printf(bio, "Time stamp: ");
+	gtime = TS_TST_INFO_get_time(a);
+	ASN1_GENERALIZEDTIME_print(bio, gtime);
+	BIO_write(bio, "\n", 1);
+
+	/* Print accuracy. */
+	BIO_printf(bio, "Accuracy: ");
+	accuracy = TS_TST_INFO_get_accuracy(a);
+	if (accuracy == NULL)
+		BIO_printf(bio, "unspecified");
+	else
+		TS_ACCURACY_print_bio(bio, accuracy);
+	BIO_write(bio, "\n", 1);
+
+	/* Print ordering. */
+	BIO_printf(bio, "Ordering: %s\n", 
+		   TS_TST_INFO_get_ordering(a) ? "yes" : "no");
+
+	/* Print nonce. */
+	BIO_printf(bio, "Nonce: ");
+	nonce = TS_TST_INFO_get_nonce(a);
+	if (nonce == NULL)
+		BIO_printf(bio, "unspecified");
+	else
+		TS_ASN1_INTEGER_print_bio(bio, nonce);
+	BIO_write(bio, "\n", 1);
+
+	/* Print TSA name. */
+	BIO_printf(bio, "TSA: ");
+	tsa_name = TS_TST_INFO_get_tsa(a);
+	if (tsa_name == NULL)
+		BIO_printf(bio, "unspecified");
+	else
+		{
+		STACK_OF(CONF_VALUE) *nval;
+		if ((nval = i2v_GENERAL_NAME(NULL, tsa_name, NULL)))
+			X509V3_EXT_val_prn(bio, nval, 0, 0);
+		sk_CONF_VALUE_pop_free(nval, X509V3_conf_free);
+		}
+	BIO_write(bio, "\n", 1);
+
+	/* Print extensions. */
+	TS_ext_print_bio(bio, TS_TST_INFO_get_exts(a));
+
+	return 1;
+	}
+
+static int TS_ACCURACY_print_bio(BIO *bio, const TS_ACCURACY *accuracy)
+	{
+	const ASN1_INTEGER *seconds = TS_ACCURACY_get_seconds(accuracy);
+	const ASN1_INTEGER *millis = TS_ACCURACY_get_millis(accuracy);
+	const ASN1_INTEGER *micros = TS_ACCURACY_get_micros(accuracy);
+
+	if (seconds != NULL)
+		TS_ASN1_INTEGER_print_bio(bio, seconds);
+	else
+		BIO_printf(bio, "unspecified");
+	BIO_printf(bio, " seconds, ");
+	if (millis != NULL)
+		TS_ASN1_INTEGER_print_bio(bio, millis);
+	else
+		BIO_printf(bio, "unspecified");
+	BIO_printf(bio, " millis, ");
+	if (micros != NULL)
+		TS_ASN1_INTEGER_print_bio(bio, micros);
+	else
+		BIO_printf(bio, "unspecified");
+	BIO_printf(bio, " micros");
+
+	return 1;
+	}
diff --git a/crypto/ts/ts_rsp_sign.c b/crypto/ts/ts_rsp_sign.c
new file mode 100644
index 000000000000..b0f023c9d2f8
--- /dev/null
+++ b/crypto/ts/ts_rsp_sign.c
@@ -0,0 +1,1020 @@
+/* crypto/ts/ts_resp_sign.c */
+/* Written by Zoltan Glozik (zglozik@stones.com) for the OpenSSL
+ * project 2002.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include "cryptlib.h"
+
+#if defined(OPENSSL_SYS_UNIX)
+#include <sys/time.h>
+#endif
+
+#include <openssl/objects.h>
+#include <openssl/ts.h>
+#include <openssl/pkcs7.h>
+
+/* Private function declarations. */
+
+static ASN1_INTEGER *def_serial_cb(struct TS_resp_ctx *, void *);
+static int def_time_cb(struct TS_resp_ctx *, void *, long *sec, long *usec);
+static int def_extension_cb(struct TS_resp_ctx *, X509_EXTENSION *, void *);
+
+static void TS_RESP_CTX_init(TS_RESP_CTX *ctx);
+static void TS_RESP_CTX_cleanup(TS_RESP_CTX *ctx);
+static int TS_RESP_check_request(TS_RESP_CTX *ctx);
+static ASN1_OBJECT *TS_RESP_get_policy(TS_RESP_CTX *ctx);
+static TS_TST_INFO *TS_RESP_create_tst_info(TS_RESP_CTX *ctx, 
+					    ASN1_OBJECT *policy);
+static int TS_RESP_process_extensions(TS_RESP_CTX *ctx);
+static int TS_RESP_sign(TS_RESP_CTX *ctx);
+
+static ESS_SIGNING_CERT *ESS_SIGNING_CERT_new_init(X509 *signcert, 
+						   STACK_OF(X509) *certs);
+static ESS_CERT_ID *ESS_CERT_ID_new_init(X509 *cert, int issuer_needed);
+static int TS_TST_INFO_content_new(PKCS7 *p7);
+static int ESS_add_signing_cert(PKCS7_SIGNER_INFO *si, ESS_SIGNING_CERT *sc);
+
+static ASN1_GENERALIZEDTIME *TS_RESP_set_genTime_with_precision(
+	ASN1_GENERALIZEDTIME *, long, long, unsigned);
+
+/* Default callbacks for response generation. */
+
+static ASN1_INTEGER *def_serial_cb(struct TS_resp_ctx *ctx, void *data)
+	{
+	ASN1_INTEGER *serial = ASN1_INTEGER_new();
+	if (!serial) goto err;
+	if (!ASN1_INTEGER_set(serial, 1)) goto err;
+	return serial;
+ err:
+	TSerr(TS_F_DEF_SERIAL_CB, ERR_R_MALLOC_FAILURE);
+	TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+				    "Error during serial number generation.");
+	return NULL;
+	}
+
+#if defined(OPENSSL_SYS_UNIX)
+
+/* Use the gettimeofday function call. */
+static int def_time_cb(struct TS_resp_ctx *ctx, void *data, 
+		       long *sec, long *usec)
+	{
+	struct timeval tv;
+	if (gettimeofday(&tv, NULL) != 0) 
+		{
+		TSerr(TS_F_DEF_TIME_CB, TS_R_TIME_SYSCALL_ERROR);
+		TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+					    "Time is not available.");
+		TS_RESP_CTX_add_failure_info(ctx, TS_INFO_TIME_NOT_AVAILABLE);
+		return 0;
+		}
+	/* Return time to caller. */
+	*sec = tv.tv_sec;
+	*usec = tv.tv_usec;
+
+	return 1;
+	}
+
+#else
+
+/* Use the time function call that provides only seconds precision. */
+static int def_time_cb(struct TS_resp_ctx *ctx, void *data, 
+		       long *sec, long *usec)
+	{
+	time_t t;
+	if (time(&t) == (time_t) -1)
+		{
+		TSerr(TS_F_DEF_TIME_CB, TS_R_TIME_SYSCALL_ERROR);
+		TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+					    "Time is not available.");
+		TS_RESP_CTX_add_failure_info(ctx, TS_INFO_TIME_NOT_AVAILABLE);
+		return 0;
+		}
+	/* Return time to caller, only second precision. */
+	*sec = (long) t;
+	*usec = 0;
+
+	return 1;
+	}
+
+#endif
+
+static int def_extension_cb(struct TS_resp_ctx *ctx, X509_EXTENSION *ext,
+			    void *data)
+	{
+	/* No extensions are processed here. */
+	TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+				    "Unsupported extension.");
+	TS_RESP_CTX_add_failure_info(ctx, TS_INFO_UNACCEPTED_EXTENSION);
+	return 0;
+	}
+
+/* TS_RESP_CTX management functions. */
+
+TS_RESP_CTX *TS_RESP_CTX_new()
+	{
+	TS_RESP_CTX *ctx;
+
+	if (!(ctx = (TS_RESP_CTX *) OPENSSL_malloc(sizeof(TS_RESP_CTX))))
+		{
+		TSerr(TS_F_TS_RESP_CTX_NEW, ERR_R_MALLOC_FAILURE);
+		return NULL;
+		}
+	memset(ctx, 0, sizeof(TS_RESP_CTX));
+
+	/* Setting default callbacks. */
+	ctx->serial_cb = def_serial_cb;
+	ctx->time_cb = def_time_cb;
+	ctx->extension_cb = def_extension_cb;
+
+	return ctx;
+	}
+
+void TS_RESP_CTX_free(TS_RESP_CTX *ctx)
+	{
+	if (!ctx) return;
+
+	X509_free(ctx->signer_cert);
+	EVP_PKEY_free(ctx->signer_key);
+	sk_X509_pop_free(ctx->certs, X509_free);
+	sk_ASN1_OBJECT_pop_free(ctx->policies, ASN1_OBJECT_free);
+	ASN1_OBJECT_free(ctx->default_policy);
+	sk_EVP_MD_free(ctx->mds);	/* No EVP_MD_free method exists. */
+	ASN1_INTEGER_free(ctx->seconds);
+	ASN1_INTEGER_free(ctx->millis);
+	ASN1_INTEGER_free(ctx->micros);
+	OPENSSL_free(ctx);
+	}
+
+int TS_RESP_CTX_set_signer_cert(TS_RESP_CTX *ctx, X509 *signer)
+	{
+	if (X509_check_purpose(signer, X509_PURPOSE_TIMESTAMP_SIGN, 0) != 1)
+		{
+		TSerr(TS_F_TS_RESP_CTX_SET_SIGNER_CERT, 
+		      TS_R_INVALID_SIGNER_CERTIFICATE_PURPOSE);
+		return 0;
+		}
+	if (ctx->signer_cert) X509_free(ctx->signer_cert);
+	ctx->signer_cert = signer;
+	CRYPTO_add(&ctx->signer_cert->references, +1, CRYPTO_LOCK_X509);
+	return 1;
+	}
+
+int TS_RESP_CTX_set_signer_key(TS_RESP_CTX *ctx, EVP_PKEY *key)
+	{
+	if (ctx->signer_key) EVP_PKEY_free(ctx->signer_key);
+	ctx->signer_key = key;
+	CRYPTO_add(&ctx->signer_key->references, +1, CRYPTO_LOCK_EVP_PKEY);
+
+	return 1;
+	}
+
+int TS_RESP_CTX_set_def_policy(TS_RESP_CTX *ctx, ASN1_OBJECT *def_policy)
+	{
+	if (ctx->default_policy) ASN1_OBJECT_free(ctx->default_policy);
+	if (!(ctx->default_policy = OBJ_dup(def_policy))) goto err;
+	return 1;
+ err:
+	TSerr(TS_F_TS_RESP_CTX_SET_DEF_POLICY, ERR_R_MALLOC_FAILURE);
+	return 0;
+	}
+
+int TS_RESP_CTX_set_certs(TS_RESP_CTX *ctx, STACK_OF(X509) *certs)
+	{
+	int i;
+
+	if (ctx->certs)
+		{
+		sk_X509_pop_free(ctx->certs, X509_free);
+		ctx->certs = NULL;
+		}
+	if (!certs) return 1;
+	if (!(ctx->certs = sk_X509_dup(certs))) 
+		{
+		TSerr(TS_F_TS_RESP_CTX_SET_CERTS, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	for (i = 0; i < sk_X509_num(ctx->certs); ++i)
+		{
+		X509 *cert = sk_X509_value(ctx->certs, i);
+		CRYPTO_add(&cert->references, +1, CRYPTO_LOCK_X509);
+		}
+
+	return 1;
+	}
+
+int TS_RESP_CTX_add_policy(TS_RESP_CTX *ctx, ASN1_OBJECT *policy)
+	{
+	ASN1_OBJECT *copy = NULL;
+
+	/* Create new policy stack if necessary. */
+	if (!ctx->policies && !(ctx->policies = sk_ASN1_OBJECT_new_null())) 
+		goto err;
+	if (!(copy = OBJ_dup(policy))) goto err;
+	if (!sk_ASN1_OBJECT_push(ctx->policies, copy)) goto err;
+
+	return 1;
+ err:
+	TSerr(TS_F_TS_RESP_CTX_ADD_POLICY, ERR_R_MALLOC_FAILURE);
+	ASN1_OBJECT_free(copy);
+	return 0;
+	}
+
+int TS_RESP_CTX_add_md(TS_RESP_CTX *ctx, const EVP_MD *md)
+	{
+	/* Create new md stack if necessary. */
+	if (!ctx->mds && !(ctx->mds = sk_EVP_MD_new_null())) 
+		goto err;
+	/* Add the shared md, no copy needed. */
+	if (!sk_EVP_MD_push(ctx->mds, (EVP_MD *)md)) goto err;
+
+	return 1;
+ err:
+	TSerr(TS_F_TS_RESP_CTX_ADD_MD, ERR_R_MALLOC_FAILURE);
+	return 0;
+	}
+
+#define TS_RESP_CTX_accuracy_free(ctx)		\
+	ASN1_INTEGER_free(ctx->seconds);	\
+	ctx->seconds = NULL;			\
+	ASN1_INTEGER_free(ctx->millis);		\
+	ctx->millis = NULL;			\
+	ASN1_INTEGER_free(ctx->micros);		\
+	ctx->micros = NULL;
+
+int TS_RESP_CTX_set_accuracy(TS_RESP_CTX *ctx, 
+			     int secs, int millis, int micros)
+	{
+
+	TS_RESP_CTX_accuracy_free(ctx);
+	if (secs && (!(ctx->seconds = ASN1_INTEGER_new())
+		     || !ASN1_INTEGER_set(ctx->seconds, secs)))
+		goto err;
+	if (millis && (!(ctx->millis = ASN1_INTEGER_new())
+		       || !ASN1_INTEGER_set(ctx->millis, millis)))
+		goto err;
+	if (micros && (!(ctx->micros = ASN1_INTEGER_new())
+		       || !ASN1_INTEGER_set(ctx->micros, micros)))
+		goto err;
+
+	return 1;
+ err:
+	TS_RESP_CTX_accuracy_free(ctx);
+	TSerr(TS_F_TS_RESP_CTX_SET_ACCURACY, ERR_R_MALLOC_FAILURE);
+	return 0;
+	}
+
+void TS_RESP_CTX_add_flags(TS_RESP_CTX *ctx, int flags)
+	{
+	ctx->flags |= flags;
+	}
+
+void TS_RESP_CTX_set_serial_cb(TS_RESP_CTX *ctx, TS_serial_cb cb, void *data)
+	{
+	ctx->serial_cb = cb;
+	ctx->serial_cb_data = data;
+	}
+
+void TS_RESP_CTX_set_time_cb(TS_RESP_CTX *ctx, TS_time_cb cb, void *data)
+	{
+	ctx->time_cb = cb;
+	ctx->time_cb_data = data;
+	}
+
+void TS_RESP_CTX_set_extension_cb(TS_RESP_CTX *ctx, 
+				  TS_extension_cb cb, void *data)
+	{
+	ctx->extension_cb = cb;
+	ctx->extension_cb_data = data;
+	}
+
+int TS_RESP_CTX_set_status_info(TS_RESP_CTX *ctx, 
+				int status, const char *text)
+	{
+	TS_STATUS_INFO *si = NULL;
+	ASN1_UTF8STRING *utf8_text = NULL;
+	int ret = 0;
+
+	if (!(si = TS_STATUS_INFO_new())) goto err;
+	if (!ASN1_INTEGER_set(si->status, status)) goto err;
+	if (text)
+		{
+		if (!(utf8_text = ASN1_UTF8STRING_new())
+		    || !ASN1_STRING_set(utf8_text, text, strlen(text)))
+			goto err;
+		if (!si->text && !(si->text = sk_ASN1_UTF8STRING_new_null()))
+			goto err;
+		if (!sk_ASN1_UTF8STRING_push(si->text, utf8_text)) goto err;
+		utf8_text = NULL;	/* Ownership is lost. */
+		}
+	if (!TS_RESP_set_status_info(ctx->response, si)) goto err;
+	ret = 1;
+ err:
+	if (!ret)
+		TSerr(TS_F_TS_RESP_CTX_SET_STATUS_INFO, ERR_R_MALLOC_FAILURE);
+	TS_STATUS_INFO_free(si);
+	ASN1_UTF8STRING_free(utf8_text);
+	return ret;
+	}
+
+int TS_RESP_CTX_set_status_info_cond(TS_RESP_CTX *ctx, 
+				     int status, const char *text)
+	{
+	int ret = 1;
+	TS_STATUS_INFO *si = TS_RESP_get_status_info(ctx->response);
+
+	if (ASN1_INTEGER_get(si->status) == TS_STATUS_GRANTED)
+		{
+		/* Status has not been set, set it now. */
+		ret = TS_RESP_CTX_set_status_info(ctx, status, text);
+		}
+	return ret;
+	}
+
+int TS_RESP_CTX_add_failure_info(TS_RESP_CTX *ctx, int failure)
+	{
+	TS_STATUS_INFO *si = TS_RESP_get_status_info(ctx->response);
+	if (!si->failure_info && !(si->failure_info = ASN1_BIT_STRING_new()))
+		goto err;
+	if (!ASN1_BIT_STRING_set_bit(si->failure_info, failure, 1))
+		goto err;
+	return 1;
+ err:
+	TSerr(TS_F_TS_RESP_CTX_ADD_FAILURE_INFO, ERR_R_MALLOC_FAILURE);
+	return 0;
+	}
+
+TS_REQ *TS_RESP_CTX_get_request(TS_RESP_CTX *ctx)
+	{
+	return ctx->request;
+	}
+
+TS_TST_INFO *TS_RESP_CTX_get_tst_info(TS_RESP_CTX *ctx)
+	{
+	return ctx->tst_info;
+	}
+
+int TS_RESP_CTX_set_clock_precision_digits(TS_RESP_CTX *ctx, unsigned precision)
+       {
+       if (precision > TS_MAX_CLOCK_PRECISION_DIGITS)
+	       return 0;
+       ctx->clock_precision_digits = precision;
+       return 1;
+       }
+
+/* Main entry method of the response generation. */
+TS_RESP *TS_RESP_create_response(TS_RESP_CTX *ctx, BIO *req_bio)
+	{
+	ASN1_OBJECT *policy;
+	TS_RESP *response;
+	int result = 0;
+
+	TS_RESP_CTX_init(ctx);
+
+	/* Creating the response object. */
+	if (!(ctx->response = TS_RESP_new())) 
+		{
+		TSerr(TS_F_TS_RESP_CREATE_RESPONSE, ERR_R_MALLOC_FAILURE);
+		goto end;
+		}
+
+	/* Parsing DER request. */
+	if (!(ctx->request = d2i_TS_REQ_bio(req_bio, NULL)))
+		{
+		TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+					    "Bad request format or "
+					    "system error.");
+		TS_RESP_CTX_add_failure_info(ctx, TS_INFO_BAD_DATA_FORMAT);
+		goto end;
+		}
+
+	/* Setting default status info. */
+	if (!TS_RESP_CTX_set_status_info(ctx, TS_STATUS_GRANTED, NULL))
+		goto end;
+
+	/* Checking the request format. */
+	if (!TS_RESP_check_request(ctx)) goto end;
+
+	/* Checking acceptable policies. */
+	if (!(policy = TS_RESP_get_policy(ctx))) goto end;
+
+	/* Creating the TS_TST_INFO object. */
+	if (!(ctx->tst_info = TS_RESP_create_tst_info(ctx, policy)))
+		goto end;
+
+	/* Processing extensions. */
+	if (!TS_RESP_process_extensions(ctx)) goto end;
+
+	/* Generating the signature. */
+	if (!TS_RESP_sign(ctx)) goto end;
+
+	/* Everything was successful. */
+	result = 1;
+ end:
+	if (!result)
+		{
+		TSerr(TS_F_TS_RESP_CREATE_RESPONSE, TS_R_RESPONSE_SETUP_ERROR);
+		if (ctx->response != NULL)
+			{
+			if (TS_RESP_CTX_set_status_info_cond(ctx,
+				TS_STATUS_REJECTION, "Error during response "
+				"generation.") == 0)
+				{
+				TS_RESP_free(ctx->response);
+				ctx->response = NULL;
+				}
+			}
+		}
+	response = ctx->response;
+	ctx->response = NULL;	/* Ownership will be returned to caller. */
+	TS_RESP_CTX_cleanup(ctx);
+	return response;
+	}
+
+/* Initializes the variable part of the context. */
+static void TS_RESP_CTX_init(TS_RESP_CTX *ctx)
+	{
+	ctx->request = NULL;
+	ctx->response = NULL;
+	ctx->tst_info = NULL;
+	}
+
+/* Cleans up the variable part of the context. */
+static void TS_RESP_CTX_cleanup(TS_RESP_CTX *ctx)
+	{
+	TS_REQ_free(ctx->request);
+	ctx->request = NULL;
+	TS_RESP_free(ctx->response);
+	ctx->response = NULL;
+	TS_TST_INFO_free(ctx->tst_info);
+	ctx->tst_info = NULL;
+	}
+
+/* Checks the format and content of the request. */
+static int TS_RESP_check_request(TS_RESP_CTX *ctx)
+	{
+	TS_REQ *request = ctx->request;
+	TS_MSG_IMPRINT *msg_imprint;
+	X509_ALGOR *md_alg;
+	int md_alg_id;
+	const ASN1_OCTET_STRING *digest;
+	EVP_MD *md = NULL;
+	int i;
+
+	/* Checking request version. */
+	if (TS_REQ_get_version(request) != 1)
+		{
+		TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+					    "Bad request version.");
+		TS_RESP_CTX_add_failure_info(ctx, TS_INFO_BAD_REQUEST);
+		return 0;
+		}
+
+	/* Checking message digest algorithm. */
+	msg_imprint = TS_REQ_get_msg_imprint(request);
+	md_alg = TS_MSG_IMPRINT_get_algo(msg_imprint);
+	md_alg_id = OBJ_obj2nid(md_alg->algorithm);
+	for (i = 0; !md && i < sk_EVP_MD_num(ctx->mds); ++i)
+		{
+		EVP_MD *current_md = sk_EVP_MD_value(ctx->mds, i);
+		if (md_alg_id == EVP_MD_type(current_md))
+			md = current_md;
+		}
+	if (!md)
+		{
+		TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+					    "Message digest algorithm is "
+					    "not supported.");
+		TS_RESP_CTX_add_failure_info(ctx, TS_INFO_BAD_ALG);
+		return 0;
+		}
+
+	/* No message digest takes parameter. */
+	if (md_alg->parameter 
+	    && ASN1_TYPE_get(md_alg->parameter) != V_ASN1_NULL)
+		{
+		TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+					    "Superfluous message digest "
+					    "parameter.");
+		TS_RESP_CTX_add_failure_info(ctx, TS_INFO_BAD_ALG);
+		return 0;
+		}
+	/* Checking message digest size. */
+	digest = TS_MSG_IMPRINT_get_msg(msg_imprint);
+	if (digest->length != EVP_MD_size(md))
+		{
+		TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+					    "Bad message digest.");
+		TS_RESP_CTX_add_failure_info(ctx, TS_INFO_BAD_DATA_FORMAT);
+		return 0;
+		}
+
+	return 1;
+	}
+
+/* Returns the TSA policy based on the requested and acceptable policies. */
+static ASN1_OBJECT *TS_RESP_get_policy(TS_RESP_CTX *ctx)
+	{
+	ASN1_OBJECT *requested = TS_REQ_get_policy_id(ctx->request);
+	ASN1_OBJECT *policy = NULL;
+	int i;
+
+	if (ctx->default_policy == NULL)
+		{
+		TSerr(TS_F_TS_RESP_GET_POLICY, TS_R_INVALID_NULL_POINTER);
+		return NULL;
+		}
+	/* Return the default policy if none is requested or the default is
+	   requested. */
+	if (!requested || !OBJ_cmp(requested, ctx->default_policy))
+		policy = ctx->default_policy;
+
+	/* Check if the policy is acceptable. */
+	for (i = 0; !policy && i < sk_ASN1_OBJECT_num(ctx->policies); ++i)
+		{
+		ASN1_OBJECT *current = sk_ASN1_OBJECT_value(ctx->policies, i);
+		if (!OBJ_cmp(requested, current))
+			policy = current;
+		}
+	if (!policy)
+		{
+		TSerr(TS_F_TS_RESP_GET_POLICY, TS_R_UNACCEPTABLE_POLICY);
+		TS_RESP_CTX_set_status_info(ctx, TS_STATUS_REJECTION,
+					    "Requested policy is not "
+					    "supported.");
+		TS_RESP_CTX_add_failure_info(ctx, TS_INFO_UNACCEPTED_POLICY);
+		}
+	return policy;
+	}
+
+/* Creates the TS_TST_INFO object based on the settings of the context. */
+static TS_TST_INFO *TS_RESP_create_tst_info(TS_RESP_CTX *ctx,
+					    ASN1_OBJECT *policy)
+	{
+	int result = 0;
+	TS_TST_INFO *tst_info = NULL;
+	ASN1_INTEGER *serial = NULL;
+	ASN1_GENERALIZEDTIME *asn1_time = NULL;
+	long sec, usec;
+	TS_ACCURACY *accuracy = NULL;
+	const ASN1_INTEGER *nonce;
+	GENERAL_NAME *tsa_name = NULL;
+
+	if (!(tst_info = TS_TST_INFO_new())) goto end;
+	if (!TS_TST_INFO_set_version(tst_info, 1)) goto end;
+	if (!TS_TST_INFO_set_policy_id(tst_info, policy)) goto end;
+	if (!TS_TST_INFO_set_msg_imprint(tst_info, ctx->request->msg_imprint))
+		goto end;
+	if (!(serial = (*ctx->serial_cb)(ctx, ctx->serial_cb_data))
+	    || !TS_TST_INFO_set_serial(tst_info, serial))
+		goto end;
+	if (!(*ctx->time_cb)(ctx, ctx->time_cb_data, &sec, &usec)
+            || !(asn1_time = TS_RESP_set_genTime_with_precision(NULL, 
+					sec, usec, 
+					ctx->clock_precision_digits))
+	    || !TS_TST_INFO_set_time(tst_info, asn1_time))
+		goto end;
+
+	/* Setting accuracy if needed. */
+	if ((ctx->seconds || ctx->millis || ctx->micros) 
+	    && !(accuracy = TS_ACCURACY_new()))
+		goto end;
+
+	if (ctx->seconds && !TS_ACCURACY_set_seconds(accuracy, ctx->seconds))
+		goto end;
+	if (ctx->millis && !TS_ACCURACY_set_millis(accuracy, ctx->millis))
+		goto end;
+	if (ctx->micros && !TS_ACCURACY_set_micros(accuracy, ctx->micros))
+		goto end;
+	if (accuracy && !TS_TST_INFO_set_accuracy(tst_info, accuracy)) 
+		goto end;
+
+	/* Setting ordering. */
+	if ((ctx->flags & TS_ORDERING) 
+	    && !TS_TST_INFO_set_ordering(tst_info, 1))
+		goto end;
+	
+	/* Setting nonce if needed. */
+	if ((nonce = TS_REQ_get_nonce(ctx->request)) != NULL
+	    && !TS_TST_INFO_set_nonce(tst_info, nonce))
+		goto end;
+
+	/* Setting TSA name to subject of signer certificate. */
+	if (ctx->flags & TS_TSA_NAME)
+		{
+		if (!(tsa_name = GENERAL_NAME_new())) goto end;
+		tsa_name->type = GEN_DIRNAME;
+		tsa_name->d.dirn = 
+			X509_NAME_dup(ctx->signer_cert->cert_info->subject);
+		if (!tsa_name->d.dirn) goto end;
+		if (!TS_TST_INFO_set_tsa(tst_info, tsa_name)) goto end;
+		}
+
+	result = 1;
+ end:
+	if (!result)
+		{
+		TS_TST_INFO_free(tst_info);
+		tst_info = NULL;
+		TSerr(TS_F_TS_RESP_CREATE_TST_INFO, TS_R_TST_INFO_SETUP_ERROR);
+		TS_RESP_CTX_set_status_info_cond(ctx, TS_STATUS_REJECTION,
+						 "Error during TSTInfo "
+						 "generation.");
+		}
+	GENERAL_NAME_free(tsa_name);
+	TS_ACCURACY_free(accuracy);
+	ASN1_GENERALIZEDTIME_free(asn1_time);
+	ASN1_INTEGER_free(serial);
+	
+	return tst_info;
+	}
+
+/* Processing the extensions of the request. */
+static int TS_RESP_process_extensions(TS_RESP_CTX *ctx)
+	{
+	STACK_OF(X509_EXTENSION) *exts = TS_REQ_get_exts(ctx->request);
+	int i;
+	int ok = 1;
+
+	for (i = 0; ok && i < sk_X509_EXTENSION_num(exts); ++i)
+		{
+		X509_EXTENSION *ext = sk_X509_EXTENSION_value(exts, i);
+		/* XXXXX The last argument was previously
+		   (void *)ctx->extension_cb, but ISO C doesn't permit
+		   converting a function pointer to void *.  For lack of
+		   better information, I'm placing a NULL there instead.
+		   The callback can pick its own address out from the ctx
+		   anyway...
+		*/
+		ok = (*ctx->extension_cb)(ctx, ext, NULL);
+		}
+
+	return ok;
+	}
+
+/* Functions for signing the TS_TST_INFO structure of the context. */
+static int TS_RESP_sign(TS_RESP_CTX *ctx)
+	{
+	int ret = 0;
+	PKCS7 *p7 = NULL;
+	PKCS7_SIGNER_INFO *si;
+	STACK_OF(X509) *certs;	/* Certificates to include in sc. */
+	ESS_SIGNING_CERT *sc = NULL;
+	ASN1_OBJECT *oid;
+	BIO *p7bio = NULL;
+	int i;
+
+	/* Check if signcert and pkey match. */
+	if (!X509_check_private_key(ctx->signer_cert, ctx->signer_key)) {
+		TSerr(TS_F_TS_RESP_SIGN, 
+		      TS_R_PRIVATE_KEY_DOES_NOT_MATCH_CERTIFICATE);
+                goto err;
+	}
+
+	/* Create a new PKCS7 signed object. */
+	if (!(p7 = PKCS7_new())) {
+		TSerr(TS_F_TS_RESP_SIGN, ERR_R_MALLOC_FAILURE);
+		goto err;
+	}
+	if (!PKCS7_set_type(p7, NID_pkcs7_signed)) goto err;
+
+	/* Force SignedData version to be 3 instead of the default 1. */
+	if (!ASN1_INTEGER_set(p7->d.sign->version, 3)) goto err;
+
+	/* Add signer certificate and optional certificate chain. */
+	if (TS_REQ_get_cert_req(ctx->request))
+		{
+		PKCS7_add_certificate(p7, ctx->signer_cert);
+		if (ctx->certs)
+			{
+			for(i = 0; i < sk_X509_num(ctx->certs); ++i) 
+				{
+				X509 *cert = sk_X509_value(ctx->certs, i);
+				PKCS7_add_certificate(p7, cert);
+				}
+			}
+		}
+
+	/* Add a new signer info. */
+    	if (!(si = PKCS7_add_signature(p7, ctx->signer_cert, 
+				       ctx->signer_key, EVP_sha1())))
+		{
+		TSerr(TS_F_TS_RESP_SIGN, TS_R_PKCS7_ADD_SIGNATURE_ERROR);
+		goto err;
+		}
+
+	/* Add content type signed attribute to the signer info. */
+	oid = OBJ_nid2obj(NID_id_smime_ct_TSTInfo);
+	if (!PKCS7_add_signed_attribute(si, NID_pkcs9_contentType,
+					V_ASN1_OBJECT, oid))
+		{
+		TSerr(TS_F_TS_RESP_SIGN, TS_R_PKCS7_ADD_SIGNED_ATTR_ERROR);
+		goto err;
+		}
+
+	/* Create the ESS SigningCertificate attribute which contains 
+	   the signer certificate id and optionally the certificate chain. */
+	certs = ctx->flags & TS_ESS_CERT_ID_CHAIN ? ctx->certs : NULL;
+	if (!(sc = ESS_SIGNING_CERT_new_init(ctx->signer_cert, certs)))
+		goto err;
+
+	/* Add SigningCertificate signed attribute to the signer info. */
+	if (!ESS_add_signing_cert(si, sc))
+		{
+		TSerr(TS_F_TS_RESP_SIGN, TS_R_ESS_ADD_SIGNING_CERT_ERROR);
+		goto err;
+		}	
+
+	/* Add a new empty NID_id_smime_ct_TSTInfo encapsulated content. */
+	if (!TS_TST_INFO_content_new(p7)) goto err;
+
+	/* Add the DER encoded tst_info to the PKCS7 structure. */
+	if (!(p7bio = PKCS7_dataInit(p7, NULL))) {
+		TSerr(TS_F_TS_RESP_SIGN, ERR_R_MALLOC_FAILURE);
+		goto err;
+	}
+
+	/* Convert tst_info to DER. */
+	if (!i2d_TS_TST_INFO_bio(p7bio, ctx->tst_info))
+		{
+		TSerr(TS_F_TS_RESP_SIGN, TS_R_TS_DATASIGN);
+		goto err;
+		}
+
+	/* Create the signature and add it to the signer info. */
+        if (!PKCS7_dataFinal(p7, p7bio))
+		{
+		TSerr(TS_F_TS_RESP_SIGN, TS_R_TS_DATASIGN);
+		goto err;
+		}
+
+	/* Set new PKCS7 and TST_INFO objects. */
+	TS_RESP_set_tst_info(ctx->response, p7, ctx->tst_info);
+	p7 = NULL;		/* Ownership is lost. */
+	ctx->tst_info = NULL;	/* Ownership is lost. */
+
+	ret = 1;
+ err:
+	if (!ret)
+		TS_RESP_CTX_set_status_info_cond(ctx, TS_STATUS_REJECTION,
+						 "Error during signature "
+						 "generation.");
+	BIO_free_all(p7bio);
+	ESS_SIGNING_CERT_free(sc);
+	PKCS7_free(p7);
+	return ret;
+	}
+
+static ESS_SIGNING_CERT *ESS_SIGNING_CERT_new_init(X509 *signcert, 
+						   STACK_OF(X509) *certs)
+	{
+	ESS_CERT_ID *cid;
+	ESS_SIGNING_CERT *sc = NULL;
+	int i;
+
+	/* Creating the ESS_CERT_ID stack. */
+	if (!(sc = ESS_SIGNING_CERT_new())) goto err;
+	if (!sc->cert_ids && !(sc->cert_ids = sk_ESS_CERT_ID_new_null()))
+		goto err;
+
+	/* Adding the signing certificate id. */
+	if (!(cid = ESS_CERT_ID_new_init(signcert, 0))
+	    || !sk_ESS_CERT_ID_push(sc->cert_ids, cid))
+		goto err;
+	/* Adding the certificate chain ids. */
+	for (i = 0; i < sk_X509_num(certs); ++i)
+		{
+		X509 *cert = sk_X509_value(certs, i);
+		if (!(cid = ESS_CERT_ID_new_init(cert, 1))
+		    || !sk_ESS_CERT_ID_push(sc->cert_ids, cid))
+			goto err;
+		}
+
+	return sc;
+err:
+	ESS_SIGNING_CERT_free(sc);
+	TSerr(TS_F_ESS_SIGNING_CERT_NEW_INIT, ERR_R_MALLOC_FAILURE);
+	return NULL;
+	}
+
+static ESS_CERT_ID *ESS_CERT_ID_new_init(X509 *cert, int issuer_needed)
+	{
+	ESS_CERT_ID *cid = NULL;
+	GENERAL_NAME *name = NULL;
+	
+	/* Recompute SHA1 hash of certificate if necessary (side effect). */
+	X509_check_purpose(cert, -1, 0);
+
+	if (!(cid = ESS_CERT_ID_new())) goto err;
+	if (!ASN1_OCTET_STRING_set(cid->hash, cert->sha1_hash,
+				   sizeof(cert->sha1_hash)))
+		goto err;
+
+	/* Setting the issuer/serial if requested. */
+	if (issuer_needed)
+		{
+		/* Creating issuer/serial structure. */
+		if (!cid->issuer_serial
+		    && !(cid->issuer_serial = ESS_ISSUER_SERIAL_new()))
+			goto err;
+		/* Creating general name from the certificate issuer. */
+		if (!(name = GENERAL_NAME_new())) goto err;
+		name->type = GEN_DIRNAME;
+		if (!(name->d.dirn = X509_NAME_dup(cert->cert_info->issuer))) 
+			goto err;
+		if (!sk_GENERAL_NAME_push(cid->issuer_serial->issuer, name)) 
+			goto err;
+		name = NULL;	/* Ownership is lost. */
+		/* Setting the serial number. */
+		ASN1_INTEGER_free(cid->issuer_serial->serial);
+		if (!(cid->issuer_serial->serial = 
+		      ASN1_INTEGER_dup(cert->cert_info->serialNumber)))
+			goto err;
+		}
+
+	return cid;
+err:
+	GENERAL_NAME_free(name);
+	ESS_CERT_ID_free(cid);
+	TSerr(TS_F_ESS_CERT_ID_NEW_INIT, ERR_R_MALLOC_FAILURE);
+	return NULL;
+	}
+
+static int TS_TST_INFO_content_new(PKCS7 *p7)
+	{
+	PKCS7 *ret = NULL;
+	ASN1_OCTET_STRING *octet_string = NULL;
+
+	/* Create new encapsulated NID_id_smime_ct_TSTInfo content. */
+	if (!(ret = PKCS7_new())) goto err;
+	if (!(ret->d.other = ASN1_TYPE_new())) goto err;
+	ret->type = OBJ_nid2obj(NID_id_smime_ct_TSTInfo);
+	if (!(octet_string = ASN1_OCTET_STRING_new())) goto err;
+	ASN1_TYPE_set(ret->d.other, V_ASN1_OCTET_STRING, octet_string);
+	octet_string = NULL;
+
+	/* Add encapsulated content to signed PKCS7 structure. */
+	if (!PKCS7_set_content(p7, ret)) goto err;
+
+	return 1;
+ err:
+	ASN1_OCTET_STRING_free(octet_string);
+	PKCS7_free(ret);
+	return 0;
+	}
+
+static int ESS_add_signing_cert(PKCS7_SIGNER_INFO *si, ESS_SIGNING_CERT *sc)
+	{
+	ASN1_STRING *seq = NULL;
+	unsigned char *p, *pp = NULL;
+	int len;
+
+	len = i2d_ESS_SIGNING_CERT(sc, NULL);
+	if (!(pp = (unsigned char *) OPENSSL_malloc(len)))
+		{
+		TSerr(TS_F_ESS_ADD_SIGNING_CERT, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+	p = pp;
+	i2d_ESS_SIGNING_CERT(sc, &p);
+	if (!(seq = ASN1_STRING_new()) || !ASN1_STRING_set(seq, pp, len))
+		{
+		TSerr(TS_F_ESS_ADD_SIGNING_CERT, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+	OPENSSL_free(pp); pp = NULL;
+	return PKCS7_add_signed_attribute(si, 
+					  NID_id_smime_aa_signingCertificate,
+					  V_ASN1_SEQUENCE, seq);
+ err:
+	ASN1_STRING_free(seq);
+	OPENSSL_free(pp);
+
+	return 0;
+	}
+
+
+static ASN1_GENERALIZEDTIME *
+TS_RESP_set_genTime_with_precision(ASN1_GENERALIZEDTIME *asn1_time, 
+				   long sec, long usec, unsigned precision)
+	{
+	time_t time_sec = (time_t) sec;
+	struct tm *tm = NULL;	
+	char genTime_str[17 + TS_MAX_CLOCK_PRECISION_DIGITS];
+	char *p = genTime_str;
+	char *p_end = genTime_str + sizeof(genTime_str);
+
+	if (precision > TS_MAX_CLOCK_PRECISION_DIGITS)
+		goto err;
+
+	
+	if (!(tm = gmtime(&time_sec)))
+		goto err;
+
+	/* 
+	 * Put "genTime_str" in GeneralizedTime format.  We work around the 
+	 * restrictions imposed by rfc3280 (i.e. "GeneralizedTime values MUST 
+	 * NOT include fractional seconds") and OpenSSL related functions to 
+	 * meet the rfc3161 requirement: "GeneralizedTime syntax can include 
+	 * fraction-of-second details". 
+	 */                   
+	p += BIO_snprintf(p, p_end - p,
+			  "%04d%02d%02d%02d%02d%02d",
+			  tm->tm_year + 1900, tm->tm_mon + 1, tm->tm_mday, 
+			  tm->tm_hour, tm->tm_min, tm->tm_sec);
+	if (precision > 0)
+	{
+		/* Add fraction of seconds (leave space for dot and null). */
+		BIO_snprintf(p, 2 + precision, ".%ld", usec);
+		/* We cannot use the snprintf return value, 
+		   because it might have been truncated. */
+		p += strlen(p);
+
+		/* To make things a bit harder, X.690 | ISO/IEC 8825-1 provides
+		   the following restrictions for a DER-encoding, which OpenSSL
+		   (specifically ASN1_GENERALIZEDTIME_check() function) doesn't 
+		   support:
+		   "The encoding MUST terminate with a "Z" (which means "Zulu" 
+		   time). The decimal point element, if present, MUST be the 
+		   point option ".". The fractional-seconds elements, 
+		   if present, MUST omit all trailing 0's; 
+		   if the elements correspond to 0, they MUST be wholly
+		   omitted, and the decimal point element also MUST be
+		   omitted." */
+		/* Remove trailing zeros. The dot guarantees the exit
+		   condition of this loop even if all the digits are zero. */
+		while (*--p == '0')
+			/* empty */;
+		/* p points to either the dot or the last non-zero digit. */
+		if (*p != '.') ++p;
+		}
+	/* Add the trailing Z and the terminating null. */
+	*p++ = 'Z';
+	*p++ = '\0';
+
+	/* Now call OpenSSL to check and set our genTime value */
+	if (!asn1_time && !(asn1_time = M_ASN1_GENERALIZEDTIME_new()))
+		goto err;
+	if (!ASN1_GENERALIZEDTIME_set_string(asn1_time, genTime_str))
+		{
+		ASN1_GENERALIZEDTIME_free(asn1_time);
+		goto err;
+		}
+
+	return asn1_time;
+ err:
+	TSerr(TS_F_TS_RESP_SET_GENTIME_WITH_PRECISION, TS_R_COULD_NOT_SET_TIME);
+	return NULL;
+	}
diff --git a/crypto/ts/ts_rsp_utils.c b/crypto/ts/ts_rsp_utils.c
new file mode 100644
index 000000000000..401c1fdc51d0
--- /dev/null
+++ b/crypto/ts/ts_rsp_utils.c
@@ -0,0 +1,409 @@
+/* crypto/ts/ts_resp_utils.c */
+/* Written by Zoltan Glozik (zglozik@stones.com) for the OpenSSL
+ * project 2002.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/ts.h>
+#include <openssl/pkcs7.h>
+
+/* Function definitions. */
+
+int TS_RESP_set_status_info(TS_RESP *a, TS_STATUS_INFO *status_info)
+	{
+	TS_STATUS_INFO *new_status_info;
+
+	if (a->status_info == status_info)
+		return 1;
+	new_status_info = TS_STATUS_INFO_dup(status_info);
+	if (new_status_info == NULL)
+		{
+		TSerr(TS_F_TS_RESP_SET_STATUS_INFO, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	TS_STATUS_INFO_free(a->status_info);
+	a->status_info = new_status_info;
+
+	return 1;
+	}
+
+TS_STATUS_INFO *TS_RESP_get_status_info(TS_RESP *a)
+	{
+	return a->status_info;
+	}
+
+/* Caller loses ownership of PKCS7 and TS_TST_INFO objects. */
+void TS_RESP_set_tst_info(TS_RESP *a, PKCS7 *p7, TS_TST_INFO *tst_info)
+	{
+	/* Set new PKCS7 and TST_INFO objects. */
+	PKCS7_free(a->token);
+	a->token = p7;
+	TS_TST_INFO_free(a->tst_info);
+	a->tst_info = tst_info;
+	}
+
+PKCS7 *TS_RESP_get_token(TS_RESP *a)
+	{
+	return a->token;
+	}
+
+TS_TST_INFO *TS_RESP_get_tst_info(TS_RESP *a)
+	{
+	return a->tst_info;
+	}
+
+int TS_TST_INFO_set_version(TS_TST_INFO *a, long version)
+	{
+	return ASN1_INTEGER_set(a->version, version);
+	}
+
+long TS_TST_INFO_get_version(const TS_TST_INFO *a)
+	{
+	return ASN1_INTEGER_get(a->version);
+	}
+
+int TS_TST_INFO_set_policy_id(TS_TST_INFO *a, ASN1_OBJECT *policy)
+	{
+	ASN1_OBJECT *new_policy;
+
+	if (a->policy_id == policy)
+		return 1;
+	new_policy = OBJ_dup(policy);
+	if (new_policy == NULL)
+		{
+		TSerr(TS_F_TS_TST_INFO_SET_POLICY_ID, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	ASN1_OBJECT_free(a->policy_id);
+	a->policy_id = new_policy;
+	return 1;
+	}
+
+ASN1_OBJECT *TS_TST_INFO_get_policy_id(TS_TST_INFO *a)
+	{
+	return a->policy_id;
+	}
+
+int TS_TST_INFO_set_msg_imprint(TS_TST_INFO *a, TS_MSG_IMPRINT *msg_imprint)
+	{
+	TS_MSG_IMPRINT *new_msg_imprint;
+
+	if (a->msg_imprint == msg_imprint)
+		return 1;
+	new_msg_imprint = TS_MSG_IMPRINT_dup(msg_imprint);
+	if (new_msg_imprint == NULL)
+		{
+		TSerr(TS_F_TS_TST_INFO_SET_MSG_IMPRINT, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	TS_MSG_IMPRINT_free(a->msg_imprint);
+	a->msg_imprint = new_msg_imprint;
+	return 1;
+	}
+
+TS_MSG_IMPRINT *TS_TST_INFO_get_msg_imprint(TS_TST_INFO *a)
+	{
+	return a->msg_imprint;
+	}
+
+int TS_TST_INFO_set_serial(TS_TST_INFO *a, const ASN1_INTEGER *serial)
+	{
+	ASN1_INTEGER *new_serial;
+
+	if (a->serial == serial)
+		return 1;
+	new_serial = ASN1_INTEGER_dup(serial);
+	if (new_serial == NULL)
+		{
+		TSerr(TS_F_TS_TST_INFO_SET_SERIAL, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	ASN1_INTEGER_free(a->serial);
+	a->serial = new_serial;
+	return 1;
+	}
+
+const ASN1_INTEGER *TS_TST_INFO_get_serial(const TS_TST_INFO *a)
+	{
+	return a->serial;
+	}
+
+int TS_TST_INFO_set_time(TS_TST_INFO *a, const ASN1_GENERALIZEDTIME *gtime)
+	{
+	ASN1_GENERALIZEDTIME *new_time;
+
+	if (a->time == gtime)
+		return 1;
+	new_time = M_ASN1_GENERALIZEDTIME_dup(gtime);
+	if (new_time == NULL)
+		{
+		TSerr(TS_F_TS_TST_INFO_SET_TIME, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	ASN1_GENERALIZEDTIME_free(a->time);
+	a->time = new_time;
+	return 1;
+	}
+
+const ASN1_GENERALIZEDTIME *TS_TST_INFO_get_time(const TS_TST_INFO *a)
+	{
+	return a->time;
+	}
+
+int TS_TST_INFO_set_accuracy(TS_TST_INFO *a, TS_ACCURACY *accuracy)
+	{
+	TS_ACCURACY *new_accuracy;
+
+	if (a->accuracy == accuracy)
+		return 1;
+	new_accuracy = TS_ACCURACY_dup(accuracy);
+	if (new_accuracy == NULL)
+		{
+		TSerr(TS_F_TS_TST_INFO_SET_ACCURACY, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	TS_ACCURACY_free(a->accuracy);
+	a->accuracy = new_accuracy;
+	return 1;
+	}
+
+TS_ACCURACY *TS_TST_INFO_get_accuracy(TS_TST_INFO *a)
+	{
+	return a->accuracy;
+	}
+
+int TS_ACCURACY_set_seconds(TS_ACCURACY *a, const ASN1_INTEGER *seconds)
+	{
+	ASN1_INTEGER *new_seconds;
+
+	if (a->seconds == seconds)
+		return 1;
+	new_seconds = ASN1_INTEGER_dup(seconds);
+	if (new_seconds == NULL)
+		{
+		TSerr(TS_F_TS_ACCURACY_SET_SECONDS, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	ASN1_INTEGER_free(a->seconds);
+	a->seconds = new_seconds;
+	return 1;
+	}
+
+const ASN1_INTEGER *TS_ACCURACY_get_seconds(const TS_ACCURACY *a)
+	{
+	return a->seconds;
+	}
+
+int TS_ACCURACY_set_millis(TS_ACCURACY *a, const ASN1_INTEGER *millis)
+	{
+	ASN1_INTEGER *new_millis = NULL;
+
+	if (a->millis == millis)
+		return 1;
+	if (millis != NULL)
+		{
+		new_millis = ASN1_INTEGER_dup(millis);
+		if (new_millis == NULL)
+			{
+			TSerr(TS_F_TS_ACCURACY_SET_MILLIS, 
+			      ERR_R_MALLOC_FAILURE);
+			return 0;
+			}
+		}
+	ASN1_INTEGER_free(a->millis);
+	a->millis = new_millis;
+	return 1;
+	}
+
+const ASN1_INTEGER *TS_ACCURACY_get_millis(const TS_ACCURACY *a)
+	{
+	return a->millis;
+	}
+
+int TS_ACCURACY_set_micros(TS_ACCURACY *a, const ASN1_INTEGER *micros)
+	{
+	ASN1_INTEGER *new_micros = NULL;
+
+	if (a->micros == micros)
+		return 1;
+	if (micros != NULL)
+		{
+		new_micros = ASN1_INTEGER_dup(micros);
+		if (new_micros == NULL)
+			{
+			TSerr(TS_F_TS_ACCURACY_SET_MICROS, 
+			      ERR_R_MALLOC_FAILURE);
+			return 0;
+			}
+		}
+	ASN1_INTEGER_free(a->micros);
+	a->micros = new_micros;
+	return 1;
+	}
+
+const ASN1_INTEGER *TS_ACCURACY_get_micros(const TS_ACCURACY *a)
+	{
+	return a->micros;
+	}
+
+int TS_TST_INFO_set_ordering(TS_TST_INFO *a, int ordering)
+	{
+	a->ordering = ordering ? 0xFF : 0x00;
+	return 1;
+	}
+
+int TS_TST_INFO_get_ordering(const TS_TST_INFO *a)
+	{
+	return a->ordering ? 1 : 0;
+	}
+
+int TS_TST_INFO_set_nonce(TS_TST_INFO *a, const ASN1_INTEGER *nonce)
+	{
+	ASN1_INTEGER *new_nonce;
+
+	if (a->nonce == nonce)
+		return 1;
+	new_nonce = ASN1_INTEGER_dup(nonce);
+	if (new_nonce == NULL)
+		{
+		TSerr(TS_F_TS_TST_INFO_SET_NONCE, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	ASN1_INTEGER_free(a->nonce);
+	a->nonce = new_nonce;
+	return 1;
+	}
+
+const ASN1_INTEGER *TS_TST_INFO_get_nonce(const TS_TST_INFO *a)
+	{
+	return a->nonce;
+	}
+
+int TS_TST_INFO_set_tsa(TS_TST_INFO *a, GENERAL_NAME *tsa)
+	{
+	GENERAL_NAME *new_tsa;
+
+	if (a->tsa == tsa)
+		return 1;
+	new_tsa = GENERAL_NAME_dup(tsa);
+	if (new_tsa == NULL)
+		{
+		TSerr(TS_F_TS_TST_INFO_SET_TSA, ERR_R_MALLOC_FAILURE);
+		return 0;
+		}
+	GENERAL_NAME_free(a->tsa);
+	a->tsa = new_tsa;
+	return 1;
+	}
+
+GENERAL_NAME *TS_TST_INFO_get_tsa(TS_TST_INFO *a)
+	{
+	return a->tsa;
+	}
+
+STACK_OF(X509_EXTENSION) *TS_TST_INFO_get_exts(TS_TST_INFO *a)
+	{
+	return a->extensions;
+	}
+
+void TS_TST_INFO_ext_free(TS_TST_INFO *a)
+	{
+	if (!a) return;
+	sk_X509_EXTENSION_pop_free(a->extensions, X509_EXTENSION_free);
+	a->extensions = NULL;
+	}
+
+int TS_TST_INFO_get_ext_count(TS_TST_INFO *a)
+	{
+	return X509v3_get_ext_count(a->extensions);
+	}
+
+int TS_TST_INFO_get_ext_by_NID(TS_TST_INFO *a, int nid, int lastpos)
+	{
+	return X509v3_get_ext_by_NID(a->extensions, nid, lastpos);
+	}
+
+int TS_TST_INFO_get_ext_by_OBJ(TS_TST_INFO *a, ASN1_OBJECT *obj, int lastpos)
+	{
+	return X509v3_get_ext_by_OBJ(a->extensions, obj, lastpos);
+	}
+
+int TS_TST_INFO_get_ext_by_critical(TS_TST_INFO *a, int crit, int lastpos)
+	{
+	return X509v3_get_ext_by_critical(a->extensions, crit, lastpos);
+	}
+
+X509_EXTENSION *TS_TST_INFO_get_ext(TS_TST_INFO *a, int loc)
+	{
+	return X509v3_get_ext(a->extensions,loc);
+	}
+
+X509_EXTENSION *TS_TST_INFO_delete_ext(TS_TST_INFO *a, int loc)
+	{
+	return X509v3_delete_ext(a->extensions,loc);
+	}
+
+int TS_TST_INFO_add_ext(TS_TST_INFO *a, X509_EXTENSION *ex, int loc)
+	{
+	return X509v3_add_ext(&a->extensions,ex,loc) != NULL;
+	}
+
+void *TS_TST_INFO_get_ext_d2i(TS_TST_INFO *a, int nid, int *crit, int *idx)
+	{
+	return X509V3_get_d2i(a->extensions, nid, crit, idx);
+	}
diff --git a/crypto/ts/ts_rsp_verify.c b/crypto/ts/ts_rsp_verify.c
new file mode 100644
index 000000000000..afe16afbe454
--- /dev/null
+++ b/crypto/ts/ts_rsp_verify.c
@@ -0,0 +1,728 @@
+/* crypto/ts/ts_resp_verify.c */
+/* Written by Zoltan Glozik (zglozik@stones.com) for the OpenSSL
+ * project 2002.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/ts.h>
+#include <openssl/pkcs7.h>
+
+/* Private function declarations. */
+
+static int TS_verify_cert(X509_STORE *store, STACK_OF(X509) *untrusted,
+			  X509 *signer, STACK_OF(X509) **chain);
+static int TS_check_signing_certs(PKCS7_SIGNER_INFO *si, STACK_OF(X509) *chain);
+static ESS_SIGNING_CERT *ESS_get_signing_cert(PKCS7_SIGNER_INFO *si);
+static int TS_find_cert(STACK_OF(ESS_CERT_ID) *cert_ids, X509 *cert);
+static int TS_issuer_serial_cmp(ESS_ISSUER_SERIAL *is, X509_CINF *cinfo);
+static int int_TS_RESP_verify_token(TS_VERIFY_CTX *ctx, 
+				 PKCS7 *token, TS_TST_INFO *tst_info);
+static int TS_check_status_info(TS_RESP *response);
+static char *TS_get_status_text(STACK_OF(ASN1_UTF8STRING) *text);
+static int TS_check_policy(ASN1_OBJECT *req_oid, TS_TST_INFO *tst_info);
+static int TS_compute_imprint(BIO *data, TS_TST_INFO *tst_info,
+			      X509_ALGOR **md_alg, 
+			      unsigned char **imprint, unsigned *imprint_len);
+static int TS_check_imprints(X509_ALGOR *algor_a, 
+			     unsigned char *imprint_a, unsigned len_a,
+			     TS_TST_INFO *tst_info);
+static int TS_check_nonces(const ASN1_INTEGER *a, TS_TST_INFO *tst_info);
+static int TS_check_signer_name(GENERAL_NAME *tsa_name, X509 *signer);
+static int TS_find_name(STACK_OF(GENERAL_NAME) *gen_names, GENERAL_NAME *name);
+
+/*
+ * Local mapping between response codes and descriptions.
+ * Don't forget to change TS_STATUS_BUF_SIZE when modifying 
+ * the elements of this array.
+ */
+static const char *TS_status_text[] =
+	{ "granted",
+	  "grantedWithMods",
+	  "rejection",
+	  "waiting",
+	  "revocationWarning",
+	  "revocationNotification" };
+
+#define TS_STATUS_TEXT_SIZE	(sizeof(TS_status_text)/sizeof(*TS_status_text))
+
+/*
+ * This must be greater or equal to the sum of the strings in TS_status_text
+ * plus the number of its elements.
+ */
+#define TS_STATUS_BUF_SIZE	256
+
+static struct
+	{
+	int code;
+	const char *text;
+	} TS_failure_info[] =
+		{ { TS_INFO_BAD_ALG, "badAlg" },
+		  { TS_INFO_BAD_REQUEST, "badRequest" },
+		  { TS_INFO_BAD_DATA_FORMAT, "badDataFormat" },
+		  { TS_INFO_TIME_NOT_AVAILABLE, "timeNotAvailable" },
+		  { TS_INFO_UNACCEPTED_POLICY, "unacceptedPolicy" },
+		  { TS_INFO_UNACCEPTED_EXTENSION, "unacceptedExtension" },
+		  { TS_INFO_ADD_INFO_NOT_AVAILABLE, "addInfoNotAvailable" },
+		  { TS_INFO_SYSTEM_FAILURE, "systemFailure" } };
+
+#define TS_FAILURE_INFO_SIZE	(sizeof(TS_failure_info) / \
+				sizeof(*TS_failure_info))
+
+/* Functions for verifying a signed TS_TST_INFO structure. */
+
+/*
+ * This function carries out the following tasks:
+ *	- Checks if there is one and only one signer.
+ *	- Search for the signing certificate in 'certs' and in the response.
+ *	- Check the extended key usage and key usage fields of the signer
+ *	certificate (done by the path validation).
+ *	- Build and validate the certificate path.
+ *	- Check if the certificate path meets the requirements of the
+ *	SigningCertificate ESS signed attribute.
+ *	- Verify the signature value.
+ *	- Returns the signer certificate in 'signer', if 'signer' is not NULL.
+ */
+int TS_RESP_verify_signature(PKCS7 *token, STACK_OF(X509) *certs,
+			     X509_STORE *store, X509 **signer_out)
+	{
+	STACK_OF(PKCS7_SIGNER_INFO) *sinfos = NULL;
+	PKCS7_SIGNER_INFO *si;
+	STACK_OF(X509) *signers = NULL;
+	X509	*signer;
+	STACK_OF(X509) *chain = NULL;
+	char	buf[4096];
+	int	i, j = 0, ret = 0;
+	BIO	*p7bio = NULL;
+
+	/* Some sanity checks first. */
+	if (!token)
+		{
+		TSerr(TS_F_TS_RESP_VERIFY_SIGNATURE, TS_R_INVALID_NULL_POINTER);
+		goto err;
+		}
+
+	/* Check for the correct content type */
+	if(!PKCS7_type_is_signed(token))
+		{
+		TSerr(TS_F_TS_RESP_VERIFY_SIGNATURE, TS_R_WRONG_CONTENT_TYPE);
+		goto err;
+		}
+
+	/* Check if there is one and only one signer. */
+	sinfos = PKCS7_get_signer_info(token);
+	if (!sinfos || sk_PKCS7_SIGNER_INFO_num(sinfos) != 1)
+		{
+		TSerr(TS_F_TS_RESP_VERIFY_SIGNATURE,
+		      TS_R_THERE_MUST_BE_ONE_SIGNER);
+		goto err;
+		}
+	si = sk_PKCS7_SIGNER_INFO_value(sinfos, 0);
+
+	/* Check for no content: no data to verify signature. */
+	if (PKCS7_get_detached(token))
+		{
+		TSerr(TS_F_TS_RESP_VERIFY_SIGNATURE, TS_R_NO_CONTENT);
+		goto err;
+		}
+	
+	/* Get hold of the signer certificate, search only internal
+	   certificates if it was requested. */
+	signers = PKCS7_get0_signers(token, certs, 0);
+	if (!signers || sk_X509_num(signers) != 1) goto err;
+	signer = sk_X509_value(signers, 0);
+
+	/* Now verify the certificate. */
+	if (!TS_verify_cert(store, certs, signer, &chain)) goto err;
+
+	/* Check if the signer certificate is consistent with the
+	   ESS extension. */
+	if (!TS_check_signing_certs(si, chain)) goto err;
+
+	/* Creating the message digest. */
+	p7bio = PKCS7_dataInit(token, NULL);
+
+	/* We now have to 'read' from p7bio to calculate digests etc. */
+	while ((i = BIO_read(p7bio,buf,sizeof(buf))) > 0);
+
+	/* Verifying the signature. */
+	j = PKCS7_signatureVerify(p7bio, token, si, signer);
+	if (j <= 0)
+		{
+		TSerr(TS_F_TS_RESP_VERIFY_SIGNATURE, TS_R_SIGNATURE_FAILURE);
+		goto err;
+		}
+
+	/* Return the signer certificate if needed. */
+	if (signer_out)
+		{
+		*signer_out = signer;
+		CRYPTO_add(&signer->references, 1, CRYPTO_LOCK_X509);
+		}
+
+	ret = 1;
+
+ err:
+	BIO_free_all(p7bio);
+	sk_X509_pop_free(chain, X509_free);
+	sk_X509_free(signers);
+
+	return ret;
+	}
+
+/*
+ * The certificate chain is returned in chain. Caller is responsible for
+ * freeing the vector.
+ */
+static int TS_verify_cert(X509_STORE *store, STACK_OF(X509) *untrusted,
+			  X509 *signer, STACK_OF(X509) **chain)
+	{
+	X509_STORE_CTX	cert_ctx;
+	int i;
+	int ret = 1;
+
+	/* chain is an out argument. */
+	*chain = NULL;
+	X509_STORE_CTX_init(&cert_ctx, store, signer, untrusted);
+	X509_STORE_CTX_set_purpose(&cert_ctx, X509_PURPOSE_TIMESTAMP_SIGN);
+	i = X509_verify_cert(&cert_ctx);
+	if (i <= 0)
+		{
+		int j = X509_STORE_CTX_get_error(&cert_ctx);
+		TSerr(TS_F_TS_VERIFY_CERT, TS_R_CERTIFICATE_VERIFY_ERROR);
+		ERR_add_error_data(2, "Verify error:",
+				   X509_verify_cert_error_string(j));
+		ret = 0;
+		}
+	else
+		{
+		/* Get a copy of the certificate chain. */
+		*chain = X509_STORE_CTX_get1_chain(&cert_ctx);
+		}
+
+	X509_STORE_CTX_cleanup(&cert_ctx);
+
+	return ret;
+	}
+
+static int TS_check_signing_certs(PKCS7_SIGNER_INFO *si, STACK_OF(X509) *chain)
+	{
+	ESS_SIGNING_CERT *ss = ESS_get_signing_cert(si);
+	STACK_OF(ESS_CERT_ID) *cert_ids = NULL;
+	X509 *cert;
+	int i = 0;
+	int ret = 0;
+
+	if (!ss) goto err;
+	cert_ids = ss->cert_ids;
+	/* The signer certificate must be the first in cert_ids. */
+	cert = sk_X509_value(chain, 0);
+	if (TS_find_cert(cert_ids, cert) != 0) goto err;
+	
+	/* Check the other certificates of the chain if there are more
+	   than one certificate ids in cert_ids. */
+	if (sk_ESS_CERT_ID_num(cert_ids) > 1)
+		{
+		/* All the certificates of the chain must be in cert_ids. */
+		for (i = 1; i < sk_X509_num(chain); ++i)
+			{
+			cert = sk_X509_value(chain, i);
+			if (TS_find_cert(cert_ids, cert) < 0) goto err;
+			}
+		}
+	ret = 1;
+ err:
+	if (!ret)
+		TSerr(TS_F_TS_CHECK_SIGNING_CERTS, 
+		      TS_R_ESS_SIGNING_CERTIFICATE_ERROR);
+	ESS_SIGNING_CERT_free(ss);
+	return ret;
+	}
+
+static ESS_SIGNING_CERT *ESS_get_signing_cert(PKCS7_SIGNER_INFO *si)
+	{
+	ASN1_TYPE *attr;
+	const unsigned char *p;
+	attr = PKCS7_get_signed_attribute(si, 
+					  NID_id_smime_aa_signingCertificate);
+	if (!attr) return NULL;
+	p = attr->value.sequence->data;
+	return d2i_ESS_SIGNING_CERT(NULL, &p, attr->value.sequence->length);
+	}
+
+/* Returns < 0 if certificate is not found, certificate index otherwise. */
+static int TS_find_cert(STACK_OF(ESS_CERT_ID) *cert_ids, X509 *cert)
+	{
+	int i;
+
+	if (!cert_ids || !cert) return -1;
+
+	/* Recompute SHA1 hash of certificate if necessary (side effect). */
+	X509_check_purpose(cert, -1, 0);
+
+	/* Look for cert in the cert_ids vector. */
+	for (i = 0; i < sk_ESS_CERT_ID_num(cert_ids); ++i)
+		{
+		ESS_CERT_ID *cid = sk_ESS_CERT_ID_value(cert_ids, i);
+
+		/* Check the SHA-1 hash first. */
+		if (cid->hash->length == sizeof(cert->sha1_hash)
+		    && !memcmp(cid->hash->data, cert->sha1_hash,
+			       sizeof(cert->sha1_hash)))
+			{
+			/* Check the issuer/serial as well if specified. */
+			ESS_ISSUER_SERIAL *is = cid->issuer_serial;
+			if (!is || !TS_issuer_serial_cmp(is, cert->cert_info))
+				return i;
+			}
+		}
+	
+	return -1;
+	}
+
+static int TS_issuer_serial_cmp(ESS_ISSUER_SERIAL *is, X509_CINF *cinfo)
+	{
+	GENERAL_NAME *issuer;
+
+	if (!is || !cinfo || sk_GENERAL_NAME_num(is->issuer) != 1) return -1;
+
+	/* Check the issuer first. It must be a directory name. */
+	issuer = sk_GENERAL_NAME_value(is->issuer, 0);
+	if (issuer->type != GEN_DIRNAME 
+	    || X509_NAME_cmp(issuer->d.dirn, cinfo->issuer))
+		return -1;
+
+	/* Check the serial number, too. */
+	if (ASN1_INTEGER_cmp(is->serial, cinfo->serialNumber))
+		return -1;
+
+	return 0;
+	}
+
+/*
+ * Verifies whether 'response' contains a valid response with regards 
+ * to the settings of the context:
+ *	- Gives an error message if the TS_TST_INFO is not present.
+ *	- Calls _TS_RESP_verify_token to verify the token content.
+ */
+int TS_RESP_verify_response(TS_VERIFY_CTX *ctx, TS_RESP *response)
+	{
+	PKCS7 *token = TS_RESP_get_token(response);
+	TS_TST_INFO *tst_info = TS_RESP_get_tst_info(response);
+	int ret = 0;
+
+	/* Check if we have a successful TS_TST_INFO object in place. */
+	if (!TS_check_status_info(response)) goto err;
+
+	/* Check the contents of the time stamp token. */
+	if (!int_TS_RESP_verify_token(ctx, token, tst_info))
+		goto err;
+
+	ret = 1;
+ err:
+	return ret;
+	}
+
+/*
+ * Tries to extract a TS_TST_INFO structure from the PKCS7 token and
+ * calls the internal int_TS_RESP_verify_token function for verifying it.
+ */
+int TS_RESP_verify_token(TS_VERIFY_CTX *ctx, PKCS7 *token)
+	{
+	TS_TST_INFO *tst_info = PKCS7_to_TS_TST_INFO(token);
+	int ret = 0;
+	if (tst_info)
+		{
+		ret = int_TS_RESP_verify_token(ctx, token, tst_info);
+		TS_TST_INFO_free(tst_info);
+		}
+	return ret;
+	}
+
+/*
+ * Verifies whether the 'token' contains a valid time stamp token 
+ * with regards to the settings of the context. Only those checks are
+ * carried out that are specified in the context:
+ *	- Verifies the signature of the TS_TST_INFO.
+ *	- Checks the version number of the response.
+ *	- Check if the requested and returned policies math.
+ *	- Check if the message imprints are the same.
+ *	- Check if the nonces are the same.
+ *	- Check if the TSA name matches the signer.
+ *	- Check if the TSA name is the expected TSA.
+ */
+static int int_TS_RESP_verify_token(TS_VERIFY_CTX *ctx, 
+				 PKCS7 *token, TS_TST_INFO *tst_info)
+	{
+	X509 *signer = NULL;
+	GENERAL_NAME *tsa_name = TS_TST_INFO_get_tsa(tst_info);
+	X509_ALGOR *md_alg = NULL;
+	unsigned char *imprint = NULL;
+	unsigned imprint_len = 0;
+	int ret = 0;
+
+	/* Verify the signature. */
+	if ((ctx->flags & TS_VFY_SIGNATURE)
+	    && !TS_RESP_verify_signature(token, ctx->certs, ctx->store,
+					 &signer))
+		goto err;
+	
+	/* Check version number of response. */
+	if ((ctx->flags & TS_VFY_VERSION)
+	    && TS_TST_INFO_get_version(tst_info) != 1)
+		{
+		TSerr(TS_F_INT_TS_RESP_VERIFY_TOKEN, TS_R_UNSUPPORTED_VERSION);
+		goto err;
+		}
+
+	/* Check policies. */
+	if ((ctx->flags & TS_VFY_POLICY)
+	    && !TS_check_policy(ctx->policy, tst_info))
+		goto err;
+	
+	/* Check message imprints. */
+	if ((ctx->flags & TS_VFY_IMPRINT)
+	    && !TS_check_imprints(ctx->md_alg, ctx->imprint, ctx->imprint_len,
+				  tst_info)) 
+		goto err;
+
+	/* Compute and check message imprints. */
+	if ((ctx->flags & TS_VFY_DATA)
+	    && (!TS_compute_imprint(ctx->data, tst_info,
+				    &md_alg, &imprint, &imprint_len)
+	    || !TS_check_imprints(md_alg, imprint, imprint_len, tst_info)))
+		goto err;
+
+	/* Check nonces. */
+	if ((ctx->flags & TS_VFY_NONCE)
+	    && !TS_check_nonces(ctx->nonce, tst_info))
+		goto err;
+
+	/* Check whether TSA name and signer certificate match. */
+	if ((ctx->flags & TS_VFY_SIGNER)
+	    && tsa_name && !TS_check_signer_name(tsa_name, signer))
+		{
+		TSerr(TS_F_INT_TS_RESP_VERIFY_TOKEN, TS_R_TSA_NAME_MISMATCH);
+		goto err;
+		}
+
+	/* Check whether the TSA is the expected one. */
+	if ((ctx->flags & TS_VFY_TSA_NAME)
+	    && !TS_check_signer_name(ctx->tsa_name, signer))
+		{
+		TSerr(TS_F_INT_TS_RESP_VERIFY_TOKEN, TS_R_TSA_UNTRUSTED);
+		goto err;
+		}
+
+	ret = 1;
+ err:
+	X509_free(signer);
+	X509_ALGOR_free(md_alg);
+	OPENSSL_free(imprint);
+	return ret;
+	}
+
+static int TS_check_status_info(TS_RESP *response)
+	{
+	TS_STATUS_INFO *info = TS_RESP_get_status_info(response);
+	long status = ASN1_INTEGER_get(info->status);
+	const char *status_text = NULL;
+	char *embedded_status_text = NULL;
+	char failure_text[TS_STATUS_BUF_SIZE] = "";
+
+	/* Check if everything went fine. */
+	if (status == 0 || status == 1) return 1;
+
+	/* There was an error, get the description in status_text. */
+	if (0 <= status && status < (long)TS_STATUS_TEXT_SIZE)
+		status_text = TS_status_text[status];
+	else
+		status_text = "unknown code";
+
+	/* Set the embedded_status_text to the returned description. */
+	if (sk_ASN1_UTF8STRING_num(info->text) > 0
+	    && !(embedded_status_text = TS_get_status_text(info->text)))
+		return 0;
+	
+	/* Filling in failure_text with the failure information. */
+	if (info->failure_info)
+		{
+		int i;
+		int first = 1;
+		for (i = 0; i < (int)TS_FAILURE_INFO_SIZE; ++i)
+			{
+			if (ASN1_BIT_STRING_get_bit(info->failure_info,
+						    TS_failure_info[i].code))
+				{
+				if (!first)
+					strcpy(failure_text, ",");
+				else
+					first = 0;
+				strcat(failure_text, TS_failure_info[i].text);
+				}
+			}
+		}
+	if (failure_text[0] == '\0')
+		strcpy(failure_text, "unspecified");
+
+	/* Making up the error string. */
+	TSerr(TS_F_TS_CHECK_STATUS_INFO, TS_R_NO_TIME_STAMP_TOKEN);
+	ERR_add_error_data(6,
+			   "status code: ", status_text,
+			   ", status text: ", embedded_status_text ? 
+			   embedded_status_text : "unspecified",
+			   ", failure codes: ", failure_text);
+	OPENSSL_free(embedded_status_text);
+
+	return 0;
+	}
+
+static char *TS_get_status_text(STACK_OF(ASN1_UTF8STRING) *text)
+	{
+	int i;
+	unsigned int length = 0;
+	char *result = NULL;
+	char *p;
+
+	/* Determine length first. */
+	for (i = 0; i < sk_ASN1_UTF8STRING_num(text); ++i)
+		{
+		ASN1_UTF8STRING *current = sk_ASN1_UTF8STRING_value(text, i);
+		length += ASN1_STRING_length(current);
+		length += 1;	/* separator character */
+		}
+	/* Allocate memory (closing '\0' included). */
+	if (!(result = OPENSSL_malloc(length)))
+		{
+		TSerr(TS_F_TS_GET_STATUS_TEXT, ERR_R_MALLOC_FAILURE);
+		return NULL;
+		}
+	/* Concatenate the descriptions. */
+	for (i = 0, p = result; i < sk_ASN1_UTF8STRING_num(text); ++i)
+		{
+		ASN1_UTF8STRING *current = sk_ASN1_UTF8STRING_value(text, i);
+		length = ASN1_STRING_length(current);
+		if (i > 0) *p++ = '/';
+		strncpy(p, (const char *)ASN1_STRING_data(current), length);
+		p += length;
+		}
+	/* We do have space for this, too. */
+	*p = '\0';
+	
+	return result;
+	}
+
+static int TS_check_policy(ASN1_OBJECT *req_oid, TS_TST_INFO *tst_info)
+	{
+	ASN1_OBJECT *resp_oid = TS_TST_INFO_get_policy_id(tst_info);
+
+	if (OBJ_cmp(req_oid, resp_oid) != 0)
+		{
+		TSerr(TS_F_TS_CHECK_POLICY, TS_R_POLICY_MISMATCH);
+		return 0;
+		}
+
+	return 1;
+	}
+
+static int TS_compute_imprint(BIO *data, TS_TST_INFO *tst_info,
+			      X509_ALGOR **md_alg, 
+			      unsigned char **imprint, unsigned *imprint_len)
+	{
+	TS_MSG_IMPRINT *msg_imprint = TS_TST_INFO_get_msg_imprint(tst_info);
+	X509_ALGOR *md_alg_resp = TS_MSG_IMPRINT_get_algo(msg_imprint);
+	const EVP_MD *md;
+	EVP_MD_CTX md_ctx;
+	unsigned char buffer[4096];
+	int length;
+
+	*md_alg = NULL;
+	*imprint = NULL;
+
+	/* Return the MD algorithm of the response. */
+	if (!(*md_alg = X509_ALGOR_dup(md_alg_resp))) goto err;
+
+	/* Getting the MD object. */
+	if (!(md = EVP_get_digestbyobj((*md_alg)->algorithm)))
+		{
+		TSerr(TS_F_TS_COMPUTE_IMPRINT, TS_R_UNSUPPORTED_MD_ALGORITHM);
+		goto err;
+		}
+
+	/* Compute message digest. */
+	length = EVP_MD_size(md);
+	if (length < 0)
+	    goto err;
+	*imprint_len = length;
+	if (!(*imprint = OPENSSL_malloc(*imprint_len))) 
+		{
+		TSerr(TS_F_TS_COMPUTE_IMPRINT, ERR_R_MALLOC_FAILURE);
+		goto err;
+		}
+
+	if (!EVP_DigestInit(&md_ctx, md))
+		goto err;
+	while ((length = BIO_read(data, buffer, sizeof(buffer))) > 0)
+		{
+		if (!EVP_DigestUpdate(&md_ctx, buffer, length))
+			goto err;
+		}
+	if (!EVP_DigestFinal(&md_ctx, *imprint, NULL))
+		goto err;
+
+	return 1;
+ err:
+	X509_ALGOR_free(*md_alg);
+	OPENSSL_free(*imprint);
+	*imprint_len = 0;
+	return 0;
+	}
+
+static int TS_check_imprints(X509_ALGOR *algor_a, 
+			     unsigned char *imprint_a, unsigned len_a,
+			     TS_TST_INFO *tst_info)
+	{
+	TS_MSG_IMPRINT *b = TS_TST_INFO_get_msg_imprint(tst_info);
+	X509_ALGOR *algor_b = TS_MSG_IMPRINT_get_algo(b);
+	int ret = 0;
+
+	/* algor_a is optional. */
+	if (algor_a)
+		{
+		/* Compare algorithm OIDs. */
+		if (OBJ_cmp(algor_a->algorithm, algor_b->algorithm)) goto err;
+
+		/* The parameter must be NULL in both. */
+		if ((algor_a->parameter 
+		     && ASN1_TYPE_get(algor_a->parameter) != V_ASN1_NULL)
+		    || (algor_b->parameter
+			&& ASN1_TYPE_get(algor_b->parameter) != V_ASN1_NULL))
+			goto err;
+		}
+
+	/* Compare octet strings. */
+	ret = len_a == (unsigned) ASN1_STRING_length(b->hashed_msg) &&
+		memcmp(imprint_a, ASN1_STRING_data(b->hashed_msg), len_a) == 0;
+ err:
+	if (!ret)
+		TSerr(TS_F_TS_CHECK_IMPRINTS, TS_R_MESSAGE_IMPRINT_MISMATCH);
+	return ret;
+	}
+
+static int TS_check_nonces(const ASN1_INTEGER *a, TS_TST_INFO *tst_info)
+	{
+	const ASN1_INTEGER *b = TS_TST_INFO_get_nonce(tst_info);
+
+	/* Error if nonce is missing. */
+	if (!b)
+		{
+		TSerr(TS_F_TS_CHECK_NONCES, TS_R_NONCE_NOT_RETURNED);
+		return 0;
+		}
+
+	/* No error if a nonce is returned without being requested. */
+	if (ASN1_INTEGER_cmp(a, b) != 0)
+		{
+		TSerr(TS_F_TS_CHECK_NONCES, TS_R_NONCE_MISMATCH);
+		return 0;
+		}
+
+	return 1;
+	}
+
+/* Check if the specified TSA name matches either the subject
+   or one of the subject alternative names of the TSA certificate. */
+static int TS_check_signer_name(GENERAL_NAME *tsa_name, X509 *signer)
+	{
+	STACK_OF(GENERAL_NAME) *gen_names = NULL;
+	int idx = -1;
+	int found = 0;
+
+	/* Check the subject name first. */
+	if (tsa_name->type == GEN_DIRNAME 
+	    && X509_name_cmp(tsa_name->d.dirn, signer->cert_info->subject) == 0)
+		return 1;
+
+	/* Check all the alternative names. */
+	gen_names = X509_get_ext_d2i(signer, NID_subject_alt_name,
+				     NULL, &idx);
+	while (gen_names != NULL
+	       && !(found = TS_find_name(gen_names, tsa_name) >= 0))
+		{
+		/* Get the next subject alternative name,
+		   although there should be no more than one. */
+		GENERAL_NAMES_free(gen_names);
+		gen_names = X509_get_ext_d2i(signer, NID_subject_alt_name,
+					     NULL, &idx);
+		}
+	if (gen_names) GENERAL_NAMES_free(gen_names);
+	
+	return found;
+	}
+
+/* Returns 1 if name is in gen_names, 0 otherwise. */
+static int TS_find_name(STACK_OF(GENERAL_NAME) *gen_names, GENERAL_NAME *name)
+	{
+	int i, found;
+	for (i = 0, found = 0; !found && i < sk_GENERAL_NAME_num(gen_names);
+	     ++i)
+		{
+		GENERAL_NAME *current = sk_GENERAL_NAME_value(gen_names, i);
+		found = GENERAL_NAME_cmp(current, name) == 0;
+		}
+	return found ? i - 1 : -1;
+	}
diff --git a/crypto/ts/ts_verify_ctx.c b/crypto/ts/ts_verify_ctx.c
new file mode 100644
index 000000000000..609b7735d420
--- /dev/null
+++ b/crypto/ts/ts_verify_ctx.c
@@ -0,0 +1,159 @@
+/* crypto/ts/ts_verify_ctx.c */
+/* Written by Zoltan Glozik (zglozik@stones.com) for the OpenSSL
+ * project 2003.
+ */
+/* ====================================================================
+ * Copyright (c) 2006 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer. 
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.OpenSSL.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    licensing@OpenSSL.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.OpenSSL.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include "cryptlib.h"
+#include <openssl/objects.h>
+#include <openssl/ts.h>
+
+TS_VERIFY_CTX *TS_VERIFY_CTX_new(void)
+	{
+	TS_VERIFY_CTX *ctx = 
+		(TS_VERIFY_CTX *) OPENSSL_malloc(sizeof(TS_VERIFY_CTX));
+	if (ctx)
+		memset(ctx, 0, sizeof(TS_VERIFY_CTX));
+	else
+		TSerr(TS_F_TS_VERIFY_CTX_NEW, ERR_R_MALLOC_FAILURE);
+	return ctx;
+	}
+
+void TS_VERIFY_CTX_init(TS_VERIFY_CTX *ctx)
+	{
+	OPENSSL_assert(ctx != NULL);
+	memset(ctx, 0, sizeof(TS_VERIFY_CTX));
+	}
+
+void TS_VERIFY_CTX_free(TS_VERIFY_CTX *ctx)
+	{
+	if (!ctx) return;
+
+	TS_VERIFY_CTX_cleanup(ctx);
+	OPENSSL_free(ctx);
+	}
+
+void TS_VERIFY_CTX_cleanup(TS_VERIFY_CTX *ctx)
+	{
+	if (!ctx) return;
+
+	X509_STORE_free(ctx->store);
+	sk_X509_pop_free(ctx->certs, X509_free);
+
+	ASN1_OBJECT_free(ctx->policy);
+
+	X509_ALGOR_free(ctx->md_alg);
+	OPENSSL_free(ctx->imprint);
+	
+	BIO_free_all(ctx->data);
+
+	ASN1_INTEGER_free(ctx->nonce);
+
+	GENERAL_NAME_free(ctx->tsa_name);
+
+	TS_VERIFY_CTX_init(ctx);
+	}
+
+TS_VERIFY_CTX *TS_REQ_to_TS_VERIFY_CTX(TS_REQ *req, TS_VERIFY_CTX *ctx)
+	{
+	TS_VERIFY_CTX *ret = ctx;
+	ASN1_OBJECT *policy;
+	TS_MSG_IMPRINT *imprint;
+	X509_ALGOR *md_alg;
+	ASN1_OCTET_STRING *msg;
+	const ASN1_INTEGER *nonce;
+
+	OPENSSL_assert(req != NULL);
+	if (ret)
+		TS_VERIFY_CTX_cleanup(ret);
+	else
+		if (!(ret = TS_VERIFY_CTX_new())) return NULL;
+
+	/* Setting flags. */
+	ret->flags = TS_VFY_ALL_IMPRINT & ~(TS_VFY_TSA_NAME | TS_VFY_SIGNATURE);
+
+	/* Setting policy. */
+	if ((policy = TS_REQ_get_policy_id(req)) != NULL)
+		{
+		if (!(ret->policy = OBJ_dup(policy))) goto err;
+		}
+	else
+		ret->flags &= ~TS_VFY_POLICY;
+
+	/* Setting md_alg, imprint and imprint_len. */
+	imprint = TS_REQ_get_msg_imprint(req);
+	md_alg = TS_MSG_IMPRINT_get_algo(imprint);
+	if (!(ret->md_alg = X509_ALGOR_dup(md_alg))) goto err;
+	msg = TS_MSG_IMPRINT_get_msg(imprint);
+	ret->imprint_len = ASN1_STRING_length(msg);
+	if (!(ret->imprint = OPENSSL_malloc(ret->imprint_len))) goto err;
+	memcpy(ret->imprint, ASN1_STRING_data(msg), ret->imprint_len);
+
+	/* Setting nonce. */
+	if ((nonce = TS_REQ_get_nonce(req)) != NULL)
+		{
+		if (!(ret->nonce = ASN1_INTEGER_dup(nonce))) goto err;
+		}
+	else
+		ret->flags &= ~TS_VFY_NONCE;
+
+	return ret;
+ err:
+	if (ctx)
+		TS_VERIFY_CTX_cleanup(ctx);
+	else
+		TS_VERIFY_CTX_free(ret);
+	return NULL;
+	}
diff --git a/crypto/txt_db/Makefile b/crypto/txt_db/Makefile
index 87e57b49f646..e6f30331d8dd 100644
--- a/crypto/txt_db/Makefile
+++ b/crypto/txt_db/Makefile
@@ -33,7 +33,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
diff --git a/crypto/txt_db/txt_db.c b/crypto/txt_db/txt_db.c
index 3ed5f72ee94b..6f2ce3b5a499 100644
--- a/crypto/txt_db/txt_db.c
+++ b/crypto/txt_db/txt_db.c
@@ -77,22 +77,23 @@ TXT_DB *TXT_DB_read(BIO *in, int num)
 	int i,add,n;
 	int size=BUFSIZE;
 	int offset=0;
-	char *p,**pp,*f;
+	char *p,*f;
+	OPENSSL_STRING *pp;
 	BUF_MEM *buf=NULL;
 
 	if ((buf=BUF_MEM_new()) == NULL) goto err;
 	if (!BUF_MEM_grow(buf,size)) goto err;
 
-	if ((ret=(TXT_DB *)OPENSSL_malloc(sizeof(TXT_DB))) == NULL)
+	if ((ret=OPENSSL_malloc(sizeof(TXT_DB))) == NULL)
 		goto err;
 	ret->num_fields=num;
 	ret->index=NULL;
 	ret->qual=NULL;
-	if ((ret->data=sk_new_null()) == NULL)
+	if ((ret->data=sk_OPENSSL_PSTRING_new_null()) == NULL)
 		goto err;
-	if ((ret->index=(LHASH **)OPENSSL_malloc(sizeof(LHASH *)*num)) == NULL)
+	if ((ret->index=OPENSSL_malloc(sizeof(*ret->index)*num)) == NULL)
 		goto err;
-	if ((ret->qual=(int (**)(char **))OPENSSL_malloc(sizeof(int (**)(char **))*num)) == NULL)
+	if ((ret->qual=OPENSSL_malloc(sizeof(*(ret->qual))*num)) == NULL)
 		goto err;
 	for (i=0; i<num; i++)
 		{
@@ -122,7 +123,7 @@ TXT_DB *TXT_DB_read(BIO *in, int num)
 		else
 			{
 			buf->data[offset-1]='\0'; /* blat the '\n' */
-			if (!(p=(char *)OPENSSL_malloc(add+offset))) goto err;
+			if (!(p=OPENSSL_malloc(add+offset))) goto err;
 			offset=0;
 			}
 		pp=(char **)p;
@@ -155,16 +156,16 @@ TXT_DB *TXT_DB_read(BIO *in, int num)
 		*(p++)='\0';
 		if ((n != num) || (*f != '\0'))
 			{
-#if !defined(OPENSSL_NO_STDIO) && !defined(OPENSSL_SYS_WIN16)	/* temporaty fix :-( */
+#if !defined(OPENSSL_NO_STDIO) && !defined(OPENSSL_SYS_WIN16)	/* temporary fix :-( */
 			fprintf(stderr,"wrong number of fields on line %ld (looking for field %d, got %d, '%s' left)\n",ln,num,n,f);
 #endif
 			er=2;
 			goto err;
 			}
 		pp[n]=p;
-		if (!sk_push(ret->data,(char *)pp))
+		if (!sk_OPENSSL_PSTRING_push(ret->data,pp))
 			{
-#if !defined(OPENSSL_NO_STDIO) && !defined(OPENSSL_SYS_WIN16)	/* temporaty fix :-( */
+#if !defined(OPENSSL_NO_STDIO) && !defined(OPENSSL_SYS_WIN16)	/* temporary fix :-( */
 			fprintf(stderr,"failure in sk_push\n");
 #endif
 			er=2;
@@ -181,7 +182,7 @@ err:
 #endif
 		if (ret != NULL)
 			{
-			if (ret->data != NULL) sk_free(ret->data);
+			if (ret->data != NULL) sk_OPENSSL_PSTRING_free(ret->data);
 			if (ret->index != NULL) OPENSSL_free(ret->index);
 			if (ret->qual != NULL) OPENSSL_free(ret->qual);
 			if (ret != NULL) OPENSSL_free(ret);
@@ -192,10 +193,10 @@ err:
 		return(ret);
 	}
 
-char **TXT_DB_get_by_index(TXT_DB *db, int idx, char **value)
+OPENSSL_STRING *TXT_DB_get_by_index(TXT_DB *db, int idx, OPENSSL_STRING *value)
 	{
-	char **ret;
-	LHASH *lh;
+	OPENSSL_STRING *ret;
+	LHASH_OF(OPENSSL_STRING) *lh;
 
 	if (idx >= db->num_fields)
 		{
@@ -208,16 +209,16 @@ char **TXT_DB_get_by_index(TXT_DB *db, int idx, char **value)
 		db->error=DB_ERROR_NO_INDEX;
 		return(NULL);
 		}
-	ret=(char **)lh_retrieve(lh,value);
+	ret=lh_OPENSSL_STRING_retrieve(lh,value);
 	db->error=DB_ERROR_OK;
 	return(ret);
 	}
 
-int TXT_DB_create_index(TXT_DB *db, int field, int (*qual)(char **),
-		LHASH_HASH_FN_TYPE hash, LHASH_COMP_FN_TYPE cmp)
+int TXT_DB_create_index(TXT_DB *db, int field, int (*qual)(OPENSSL_STRING *),
+			LHASH_HASH_FN_TYPE hash, LHASH_COMP_FN_TYPE cmp)
 	{
-	LHASH *idx;
-	char **r;
+	LHASH_OF(OPENSSL_STRING) *idx;
+	OPENSSL_STRING *r;
 	int i,n;
 
 	if (field >= db->num_fields)
@@ -225,26 +226,27 @@ int TXT_DB_create_index(TXT_DB *db, int field, int (*qual)(char **),
 		db->error=DB_ERROR_INDEX_OUT_OF_RANGE;
 		return(0);
 		}
-	if ((idx=lh_new(hash,cmp)) == NULL)
+	/* FIXME: we lose type checking at this point */
+	if ((idx=(LHASH_OF(OPENSSL_STRING) *)lh_new(hash,cmp)) == NULL)
 		{
 		db->error=DB_ERROR_MALLOC;
 		return(0);
 		}
-	n=sk_num(db->data);
+	n=sk_OPENSSL_PSTRING_num(db->data);
 	for (i=0; i<n; i++)
 		{
-		r=(char **)sk_value(db->data,i);
+		r=sk_OPENSSL_PSTRING_value(db->data,i);
 		if ((qual != NULL) && (qual(r) == 0)) continue;
-		if ((r=lh_insert(idx,r)) != NULL)
+		if ((r=lh_OPENSSL_STRING_insert(idx,r)) != NULL)
 			{
 			db->error=DB_ERROR_INDEX_CLASH;
-			db->arg1=sk_find(db->data,(char *)r);
+			db->arg1=sk_OPENSSL_PSTRING_find(db->data,r);
 			db->arg2=i;
-			lh_free(idx);
+			lh_OPENSSL_STRING_free(idx);
 			return(0);
 			}
 		}
-	if (db->index[field] != NULL) lh_free(db->index[field]);
+	if (db->index[field] != NULL) lh_OPENSSL_STRING_free(db->index[field]);
 	db->index[field]=idx;
 	db->qual[field]=qual;
 	return(1);
@@ -259,11 +261,11 @@ long TXT_DB_write(BIO *out, TXT_DB *db)
 
 	if ((buf=BUF_MEM_new()) == NULL)
 		goto err;
-	n=sk_num(db->data);
+	n=sk_OPENSSL_PSTRING_num(db->data);
 	nn=db->num_fields;
 	for (i=0; i<n; i++)
 		{
-		pp=(char **)sk_value(db->data,i);
+		pp=sk_OPENSSL_PSTRING_value(db->data,i);
 
 		l=0;
 		for (j=0; j<nn; j++)
@@ -298,10 +300,10 @@ err:
 	return(ret);
 	}
 
-int TXT_DB_insert(TXT_DB *db, char **row)
+int TXT_DB_insert(TXT_DB *db, OPENSSL_STRING *row)
 	{
 	int i;
-	char **r;
+	OPENSSL_STRING *r;
 
 	for (i=0; i<db->num_fields; i++)
 		{
@@ -309,7 +311,7 @@ int TXT_DB_insert(TXT_DB *db, char **row)
 			{
 			if ((db->qual[i] != NULL) &&
 				(db->qual[i](row) == 0)) continue;
-			r=(char **)lh_retrieve(db->index[i],row);
+			r=lh_OPENSSL_STRING_retrieve(db->index[i],row);
 			if (r != NULL)
 				{
 				db->error=DB_ERROR_INDEX_CLASH;
@@ -320,7 +322,7 @@ int TXT_DB_insert(TXT_DB *db, char **row)
 			}
 		}
 	/* We have passed the index checks, now just append and insert */
-	if (!sk_push(db->data,(char *)row))
+	if (!sk_OPENSSL_PSTRING_push(db->data,row))
 		{
 		db->error=DB_ERROR_MALLOC;
 		goto err;
@@ -332,7 +334,7 @@ int TXT_DB_insert(TXT_DB *db, char **row)
 			{
 			if ((db->qual[i] != NULL) &&
 				(db->qual[i](row) == 0)) continue;
-			lh_insert(db->index[i],row);
+			(void)lh_OPENSSL_STRING_insert(db->index[i],row);
 			}
 		}
 	return(1);
@@ -351,18 +353,18 @@ void TXT_DB_free(TXT_DB *db)
 	if (db->index != NULL)
 		{
 		for (i=db->num_fields-1; i>=0; i--)
-			if (db->index[i] != NULL) lh_free(db->index[i]);
+			if (db->index[i] != NULL) lh_OPENSSL_STRING_free(db->index[i]);
 		OPENSSL_free(db->index);
 		}
 	if (db->qual != NULL)
 		OPENSSL_free(db->qual);
 	if (db->data != NULL)
 		{
-		for (i=sk_num(db->data)-1; i>=0; i--)
+		for (i=sk_OPENSSL_PSTRING_num(db->data)-1; i>=0; i--)
 			{
 			/* check if any 'fields' have been allocated
 			 * from outside of the initial block */
-			p=(char **)sk_value(db->data,i);
+			p=sk_OPENSSL_PSTRING_value(db->data,i);
 			max=p[db->num_fields]; /* last address */
 			if (max == NULL) /* new row */
 				{
@@ -378,9 +380,9 @@ void TXT_DB_free(TXT_DB *db)
 						OPENSSL_free(p[n]);
 					}
 				}
-			OPENSSL_free(sk_value(db->data,i));
+			OPENSSL_free(sk_OPENSSL_PSTRING_value(db->data,i));
 			}
-		sk_free(db->data);
+		sk_OPENSSL_PSTRING_free(db->data);
 		}
 	OPENSSL_free(db);
 	}
diff --git a/crypto/txt_db/txt_db.h b/crypto/txt_db/txt_db.h
index 307e1ba23fc2..6abe435bc8fa 100644
--- a/crypto/txt_db/txt_db.h
+++ b/crypto/txt_db/txt_db.h
@@ -77,16 +77,19 @@
 extern "C" {
 #endif
 
+typedef OPENSSL_STRING *OPENSSL_PSTRING;
+DECLARE_SPECIAL_STACK_OF(OPENSSL_PSTRING, OPENSSL_STRING)
+
 typedef struct txt_db_st
 	{
 	int num_fields;
-	STACK /* char ** */ *data;
-	LHASH **index;
-	int (**qual)(char **);
+	STACK_OF(OPENSSL_PSTRING) *data;
+	LHASH_OF(OPENSSL_STRING) **index;
+	int (**qual)(OPENSSL_STRING *);
 	long error;
 	long arg1;
 	long arg2;
-	char **arg_row;
+	OPENSSL_STRING *arg_row;
 	} TXT_DB;
 
 #ifndef OPENSSL_NO_BIO
@@ -96,11 +99,11 @@ long TXT_DB_write(BIO *out, TXT_DB *db);
 TXT_DB *TXT_DB_read(char *in, int num);
 long TXT_DB_write(char *out, TXT_DB *db);
 #endif
-int TXT_DB_create_index(TXT_DB *db,int field,int (*qual)(char **),
-		LHASH_HASH_FN_TYPE hash, LHASH_COMP_FN_TYPE cmp);
+int TXT_DB_create_index(TXT_DB *db,int field,int (*qual)(OPENSSL_STRING *),
+			LHASH_HASH_FN_TYPE hash, LHASH_COMP_FN_TYPE cmp);
 void TXT_DB_free(TXT_DB *db);
-char **TXT_DB_get_by_index(TXT_DB *db, int idx, char **value);
-int TXT_DB_insert(TXT_DB *db,char **value);
+OPENSSL_STRING *TXT_DB_get_by_index(TXT_DB *db, int idx, OPENSSL_STRING *value);
+int TXT_DB_insert(TXT_DB *db, OPENSSL_STRING *value);
 
 #ifdef  __cplusplus
 }
diff --git a/crypto/ui/Makefile b/crypto/ui/Makefile
index 4755e206f667..a685659fb4c8 100644
--- a/crypto/ui/Makefile
+++ b/crypto/ui/Makefile
@@ -37,7 +37,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
diff --git a/crypto/ui/ui.h b/crypto/ui/ui.h
index 018296412b18..bd78aa413f8b 100644
--- a/crypto/ui/ui.h
+++ b/crypto/ui/ui.h
@@ -287,8 +287,8 @@ UI_METHOD *UI_OpenSSL(void);
 /* The UI_STRING type is the data structure that contains all the needed info
    about a string or a prompt, including test data for a verification prompt.
 */
-DECLARE_STACK_OF(UI_STRING)
 typedef struct ui_string_st UI_STRING;
+DECLARE_STACK_OF(UI_STRING)
 
 /* The different types of strings that are currently supported.
    This is only needed by method authors. */
@@ -310,11 +310,13 @@ int UI_method_set_writer(UI_METHOD *method, int (*writer)(UI *ui, UI_STRING *uis
 int UI_method_set_flusher(UI_METHOD *method, int (*flusher)(UI *ui));
 int UI_method_set_reader(UI_METHOD *method, int (*reader)(UI *ui, UI_STRING *uis));
 int UI_method_set_closer(UI_METHOD *method, int (*closer)(UI *ui));
+int UI_method_set_prompt_constructor(UI_METHOD *method, char *(*prompt_constructor)(UI* ui, const char* object_desc, const char* object_name));
 int (*UI_method_get_opener(UI_METHOD *method))(UI*);
 int (*UI_method_get_writer(UI_METHOD *method))(UI*,UI_STRING*);
 int (*UI_method_get_flusher(UI_METHOD *method))(UI*);
 int (*UI_method_get_reader(UI_METHOD *method))(UI*,UI_STRING*);
 int (*UI_method_get_closer(UI_METHOD *method))(UI*);
+char * (*UI_method_get_prompt_constructor(UI_METHOD *method))(UI*, const char*, const char*);
 
 /* The following functions are helpers for method writers to access relevant
    data from a UI_STRING. */
diff --git a/crypto/ui/ui_err.c b/crypto/ui/ui_err.c
index 786bd0dbc3b9..a6b96299a0f4 100644
--- a/crypto/ui/ui_err.c
+++ b/crypto/ui/ui_err.c
@@ -1,6 +1,6 @@
 /* crypto/ui/ui_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
diff --git a/crypto/ui/ui_lib.c b/crypto/ui/ui_lib.c
index ac0100808f5d..a8abc270642a 100644
--- a/crypto/ui/ui_lib.c
+++ b/crypto/ui/ui_lib.c
@@ -693,6 +693,17 @@ int UI_method_set_closer(UI_METHOD *method, int (*closer)(UI *ui))
 		return -1;
 	}
 
+int UI_method_set_prompt_constructor(UI_METHOD *method, char *(*prompt_constructor)(UI* ui, const char* object_desc, const char* object_name))
+	{
+	if (method)
+		{
+		method->ui_construct_prompt = prompt_constructor;
+		return 0;
+		}
+	else
+		return -1;
+	}
+
 int (*UI_method_get_opener(UI_METHOD *method))(UI*)
 	{
 	if (method)
@@ -733,6 +744,14 @@ int (*UI_method_get_closer(UI_METHOD *method))(UI*)
 		return NULL;
 	}
 
+char* (*UI_method_get_prompt_constructor(UI_METHOD *method))(UI*, const char*, const char*)
+	{
+	if (method)
+		return method->ui_construct_prompt;
+	else
+		return NULL;
+	}
+
 enum UI_string_types UI_get_string_type(UI_STRING *uis)
 	{
 	if (!uis)
diff --git a/crypto/ui/ui_openssl.c b/crypto/ui/ui_openssl.c
index 06270f09cd42..5832a73cf59a 100644
--- a/crypto/ui/ui_openssl.c
+++ b/crypto/ui/ui_openssl.c
@@ -122,7 +122,9 @@
  * sigaction and fileno included. -pedantic would be more appropriate for
  * the intended purposes, but we can't prevent users from adding -ansi.
  */
-#define _POSIX_C_SOURCE 1
+#if !defined(_POSIX_C_SOURCE) && defined(OPENSSL_SYS_VMS)
+#define _POSIX_C_SOURCE 2
+#endif
 #include <signal.h>
 #include <stdio.h>
 #include <string.h>
@@ -476,7 +478,7 @@ static int open_console(UI *ui)
 	CRYPTO_w_lock(CRYPTO_LOCK_UI);
 	is_a_tty = 1;
 
-#if defined(OPENSSL_SYS_MACINTOSH_CLASSIC) || defined(OPENSSL_SYS_VXWORKS) || defined(OPENSSL_SYS_NETWARE)
+#if defined(OPENSSL_SYS_MACINTOSH_CLASSIC) || defined(OPENSSL_SYS_VXWORKS) || defined(OPENSSL_SYS_NETWARE) || defined(OPENSSL_SYS_BEOS)
 	tty_in=stdin;
 	tty_out=stderr;
 #else
diff --git a/crypto/vms_rms.h b/crypto/vms_rms.h
new file mode 100755
index 000000000000..00a00d993fc8
--- /dev/null
+++ b/crypto/vms_rms.h
@@ -0,0 +1,51 @@
+
+#ifdef NAML$C_MAXRSS
+
+# define CC_RMS_NAMX cc$rms_naml
+# define FAB_NAMX fab$l_naml
+# define FAB_OR_NAML( fab, naml) naml
+# define FAB_OR_NAML_DNA naml$l_long_defname
+# define FAB_OR_NAML_DNS naml$l_long_defname_size
+# define FAB_OR_NAML_FNA naml$l_long_filename
+# define FAB_OR_NAML_FNS naml$l_long_filename_size
+# define NAMX_ESA naml$l_long_expand
+# define NAMX_ESL naml$l_long_expand_size
+# define NAMX_ESS naml$l_long_expand_alloc
+# define NAMX_NOP naml$b_nop
+# define SET_NAMX_NO_SHORT_UPCASE( nam) nam.naml$v_no_short_upcase = 1
+
+# if __INITIAL_POINTER_SIZE == 64
+#  define NAMX_DNA_FNA_SET(fab) fab.fab$l_dna = (__char_ptr32) -1; \
+   fab.fab$l_fna = (__char_ptr32) -1;
+# else /* __INITIAL_POINTER_SIZE == 64 */
+#  define NAMX_DNA_FNA_SET(fab) fab.fab$l_dna = (char *) -1; \
+   fab.fab$l_fna = (char *) -1;
+# endif /* __INITIAL_POINTER_SIZE == 64 [else] */
+
+# define NAMX_MAXRSS NAML$C_MAXRSS
+# define NAMX_STRUCT NAML
+
+#else /* def NAML$C_MAXRSS */
+
+# define CC_RMS_NAMX cc$rms_nam
+# define FAB_NAMX fab$l_nam
+# define FAB_OR_NAML( fab, naml) fab
+# define FAB_OR_NAML_DNA fab$l_dna
+# define FAB_OR_NAML_DNS fab$b_dns
+# define FAB_OR_NAML_FNA fab$l_fna
+# define FAB_OR_NAML_FNS fab$b_fns
+# define NAMX_ESA nam$l_esa
+# define NAMX_ESL nam$b_esl
+# define NAMX_ESS nam$b_ess
+# define NAMX_NOP nam$b_nop
+# define NAMX_DNA_FNA_SET(fab)
+# define NAMX_MAXRSS NAM$C_MAXRSS
+# define NAMX_STRUCT NAM
+# ifdef NAM$M_NO_SHORT_UPCASE
+#  define SET_NAMX_NO_SHORT_UPCASE( nam) naml.naml$v_no_short_upcase = 1
+# else /* def NAM$M_NO_SHORT_UPCASE */
+#  define SET_NAMX_NO_SHORT_UPCASE( nam)
+# endif /* def NAM$M_NO_SHORT_UPCASE [else] */
+
+#endif /* def NAML$C_MAXRSS [else] */
+
diff --git a/crypto/whrlpool/Makefile b/crypto/whrlpool/Makefile
new file mode 100644
index 000000000000..f4d46e4d17b7
--- /dev/null
+++ b/crypto/whrlpool/Makefile
@@ -0,0 +1,96 @@
+#
+# crypto/whrlpool/Makefile
+#
+
+DIR=	whrlpool
+TOP=	../..
+CC=	cc
+CPP=	$(CC) -E
+INCLUDES=
+CFLAG=-g
+MAKEFILE=	Makefile
+AR=		ar r
+
+WP_ASM_OBJ=wp_block.o
+
+CFLAGS= $(INCLUDES) $(CFLAG)
+ASFLAGS= $(INCLUDES) $(ASFLAG)
+AFLAGS= $(ASFLAGS)
+
+GENERAL=Makefile
+TEST=wp_test.c
+APPS=
+
+LIB=$(TOP)/libcrypto.a
+LIBSRC=wp_dgst.c wp_block.c
+LIBOBJ=wp_dgst.o $(WP_ASM_OBJ)
+
+SRC= $(LIBSRC)
+
+EXHEADER= whrlpool.h
+HEADER= wp_locl.h $(EXHEADER)
+
+ALL=    $(GENERAL) $(SRC) $(HEADER)
+
+top:
+	(cd ../..; $(MAKE) DIRS=crypto SDIRS=$(DIR) sub_all)
+
+all:	lib
+
+lib:	$(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
+	$(RANLIB) $(LIB) || echo Never mind.
+	@touch lib
+
+wp-mmx.s:	asm/wp-mmx.pl ../perlasm/x86asm.pl
+	$(PERL) asm/wp-mmx.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
+
+wp-x86_64.s: asm/wp-x86_64.pl
+	$(PERL) asm/wp-x86_64.pl $(PERLASM_SCHEME) > $@
+
+$(LIBOBJ): $(LIBSRC)
+
+files:
+	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
+
+links:
+	@$(PERL) $(TOP)/util/mklink.pl ../../include/openssl $(EXHEADER)
+	@$(PERL) $(TOP)/util/mklink.pl ../../test $(TEST)
+	@$(PERL) $(TOP)/util/mklink.pl ../../apps $(APPS)
+
+install:
+	@[ -n "$(INSTALLTOP)" ] # should be set by top Makefile...
+	@headerlist="$(EXHEADER)"; for i in $$headerlist ; \
+	do  \
+	(cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i; \
+	chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/include/openssl/$$i ); \
+	done;
+
+tags:
+	ctags $(SRC)
+
+tests:
+
+lint:
+	lint -DLINT $(INCLUDES) $(SRC)>fluff
+
+depend:
+	@[ -n "$(MAKEDEPEND)" ] # should be set by upper Makefile...
+	$(MAKEDEPEND) -- $(CFLAG) $(INCLUDES) $(DEPFLAG) -- $(PROGS) $(LIBSRC)
+
+dclean:
+	$(PERL) -pe 'if (/^# DO NOT DELETE THIS LINE/) {print; exit(0);}' $(MAKEFILE) >Makefile.new
+	mv -f Makefile.new $(MAKEFILE)
+
+clean:
+	rm -f *.s *.o *.obj lib tags core .pure .nfs* *.old *.bak fluff
+
+# DO NOT DELETE THIS LINE -- make depend depends on it.
+
+wp_block.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
+wp_block.o: ../../include/openssl/whrlpool.h wp_block.c wp_locl.h
+wp_dgst.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+wp_dgst.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+wp_dgst.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+wp_dgst.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+wp_dgst.o: ../../include/openssl/whrlpool.h wp_dgst.c wp_locl.h
diff --git a/crypto/whrlpool/asm/wp-mmx.pl b/crypto/whrlpool/asm/wp-mmx.pl
new file mode 100755
index 000000000000..32cf16380b54
--- /dev/null
+++ b/crypto/whrlpool/asm/wp-mmx.pl
@@ -0,0 +1,493 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. Rights for redistribution and usage in source and binary
+# forms are granted according to the OpenSSL license.
+# ====================================================================
+#
+# whirlpool_block_mmx implementation.
+#
+*SCALE=\(2); # 2 or 8, that is the question:-) Value of 8 results
+# in 16KB large table, which is tough on L1 cache, but eliminates
+# unaligned references to it. Value of 2 results in 4KB table, but
+# 7/8 of references to it are unaligned. AMD cores seem to be
+# allergic to the latter, while Intel ones - to former [see the
+# table]. I stick to value of 2 for two reasons: 1. smaller table
+# minimizes cache trashing and thus mitigates the hazard of side-
+# channel leakage similar to AES cache-timing one; 2. performance
+# gap among different µ-archs is smaller.
+#
+# Performance table lists rounded amounts of CPU cycles spent by
+# whirlpool_block_mmx routine on single 64 byte input block, i.e.
+# smaller is better and asymptotic throughput can be estimated by
+# multiplying 64 by CPU clock frequency and dividing by relevant
+# value from the given table:
+#
+#		$SCALE=2/8	icc8	gcc3	
+# Intel P4	3200/4600	4600(*)	6400
+# Intel PIII	2900/3000	4900	5400
+# AMD K[78]	2500/1800	9900	8200(**)
+#
+# (*)	I've sketched even non-MMX assembler, but for the record
+#	I've failed to beat the Intel compiler on P4, without using
+#	MMX that is...
+# (**)	... on AMD on the other hand non-MMX assembler was observed
+#	to perform significantly better, but I figured this MMX
+#	implementation is even faster anyway, so why bother? As for
+#	pre-MMX AMD core[s], the improvement coefficient is more
+#	than likely to vary anyway and I don't know how. But the
+#	least I know is that gcc-generated code compiled with
+#	-DL_ENDIAN and -DOPENSSL_SMALL_FOOTPRINT [see C module for
+#	details] and optimized for Pentium was observed to perform
+#	*better* on Pentium 100 than unrolled non-MMX assembler
+#	loop... So we just say that I don't know if maintaining
+#	non-MMX implementation would actually pay off, but till
+#	opposite is proved "unlikely" is assumed.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"wp-mmx.pl");
+
+sub L()  { &data_byte(@_); }
+sub LL()
+{	if	($SCALE==2)	{ &data_byte(@_); &data_byte(@_); }
+	elsif	($SCALE==8)	{ for ($i=0;$i<8;$i++) {
+					&data_byte(@_);
+					unshift(@_,pop(@_));
+				  }
+				}
+	else			{ die "unvalid SCALE value"; }
+}
+
+sub scale()
+{	if	($SCALE==2)	{ &lea(@_[0],&DWP(0,@_[1],@_[1])); }
+	elsif	($SCALE==8)	{ &lea(@_[0],&DWP(0,"",@_[1],8));  }
+	else			{ die "unvalid SCALE value";       }
+}
+
+sub row()
+{	if	($SCALE==2)	{ ((8-shift)&7); }
+	elsif	($SCALE==8)	{ (8*shift);     }
+	else			{ die "unvalid SCALE value"; }
+}
+
+$tbl="ebp";
+@mm=("mm0","mm1","mm2","mm3","mm4","mm5","mm6","mm7");
+
+&function_begin_B("whirlpool_block_mmx");
+	&push	("ebp");
+	&push	("ebx");
+	&push	("esi");
+	&push	("edi");
+
+	&mov	("esi",&wparam(0));		# hash value
+	&mov	("edi",&wparam(1));		# input data stream
+	&mov	("ebp",&wparam(2));		# number of chunks in input
+
+	&mov	("eax","esp");			# copy stack pointer
+	&sub	("esp",128+20);			# allocate frame
+	&and	("esp",-64);			# align for cache-line
+
+	&lea	("ebx",&DWP(128,"esp"));
+	&mov	(&DWP(0,"ebx"),"esi");		# save parameter block
+	&mov	(&DWP(4,"ebx"),"edi");
+	&mov	(&DWP(8,"ebx"),"ebp");
+	&mov	(&DWP(16,"ebx"),"eax");		# saved stack pointer
+
+	&call	(&label("pic_point"));
+&set_label("pic_point");
+	&blindpop($tbl);
+	&lea	($tbl,&DWP(&label("table")."-".&label("pic_point"),$tbl));
+
+	&xor	("ecx","ecx");
+	&xor	("edx","edx");
+
+	for($i=0;$i<8;$i++) { &movq(@mm[$i],&QWP($i*8,"esi")); }    # L=H
+&set_label("outerloop");
+	for($i=0;$i<8;$i++) { &movq(&QWP($i*8,"esp"),@mm[$i]); }    # K=L
+	for($i=0;$i<8;$i++) { &pxor(@mm[$i],&QWP($i*8,"edi")); }    # L^=inp
+	for($i=0;$i<8;$i++) { &movq(&QWP(64+$i*8,"esp"),@mm[$i]); } # S=L
+
+	&xor	("esi","esi");
+	&mov	(&DWP(12,"ebx"),"esi");		# zero round counter
+
+&set_label("round",16);
+	&movq	(@mm[0],&QWP(2048*$SCALE,$tbl,"esi",8));	# rc[r]
+	&mov	("eax",&DWP(0,"esp"));
+	&mov	("ebx",&DWP(4,"esp"));
+for($i=0;$i<8;$i++) {
+    my $func = ($i==0)? movq : pxor;
+	&movb	(&LB("ecx"),&LB("eax"));
+	&movb	(&LB("edx"),&HB("eax"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&shr	("eax",16);
+	&pxor	(@mm[0],&QWP(&row(0),$tbl,"esi",8));
+	&$func	(@mm[1],&QWP(&row(1),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("eax"));
+	&movb	(&LB("edx"),&HB("eax"));
+	&mov	("eax",&DWP(($i+1)*8,"esp"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&$func	(@mm[2],&QWP(&row(2),$tbl,"esi",8));
+	&$func	(@mm[3],&QWP(&row(3),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("ebx"));
+	&movb	(&LB("edx"),&HB("ebx"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&shr	("ebx",16);
+	&$func	(@mm[4],&QWP(&row(4),$tbl,"esi",8));
+	&$func	(@mm[5],&QWP(&row(5),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("ebx"));
+	&movb	(&LB("edx"),&HB("ebx"));
+	&mov	("ebx",&DWP(($i+1)*8+4,"esp"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&$func	(@mm[6],&QWP(&row(6),$tbl,"esi",8));
+	&$func	(@mm[7],&QWP(&row(7),$tbl,"edi",8));
+    push(@mm,shift(@mm));
+}
+
+	for($i=0;$i<8;$i++) { &movq(&QWP($i*8,"esp"),@mm[$i]); }    # K=L
+
+for($i=0;$i<8;$i++) {
+	&movb	(&LB("ecx"),&LB("eax"));
+	&movb	(&LB("edx"),&HB("eax"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&shr	("eax",16);
+	&pxor	(@mm[0],&QWP(&row(0),$tbl,"esi",8));
+	&pxor	(@mm[1],&QWP(&row(1),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("eax"));
+	&movb	(&LB("edx"),&HB("eax"));
+	&mov	("eax",&DWP(64+($i+1)*8,"esp"))		if ($i<7);
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&pxor	(@mm[2],&QWP(&row(2),$tbl,"esi",8));
+	&pxor	(@mm[3],&QWP(&row(3),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("ebx"));
+	&movb	(&LB("edx"),&HB("ebx"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&shr	("ebx",16);
+	&pxor	(@mm[4],&QWP(&row(4),$tbl,"esi",8));
+	&pxor	(@mm[5],&QWP(&row(5),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("ebx"));
+	&movb	(&LB("edx"),&HB("ebx"));
+	&mov	("ebx",&DWP(64+($i+1)*8+4,"esp"))	if ($i<7);
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&pxor	(@mm[6],&QWP(&row(6),$tbl,"esi",8));
+	&pxor	(@mm[7],&QWP(&row(7),$tbl,"edi",8));
+    push(@mm,shift(@mm));
+}
+	&lea	("ebx",&DWP(128,"esp"));
+	&mov	("esi",&DWP(12,"ebx"));		# pull round counter
+	&add	("esi",1);
+	&cmp	("esi",10);
+	&je	(&label("roundsdone"));
+
+	&mov	(&DWP(12,"ebx"),"esi");		# update round counter
+	for($i=0;$i<8;$i++) { &movq(&QWP(64+$i*8,"esp"),@mm[$i]); } # S=L
+	&jmp	(&label("round"));
+
+&set_label("roundsdone",16);
+	&mov	("esi",&DWP(0,"ebx"));		# reload argument block
+	&mov	("edi",&DWP(4,"ebx"));
+	&mov	("eax",&DWP(8,"ebx"));
+
+	for($i=0;$i<8;$i++) { &pxor(@mm[$i],&QWP($i*8,"edi")); }    # L^=inp
+	for($i=0;$i<8;$i++) { &pxor(@mm[$i],&QWP($i*8,"esi")); }    # L^=H
+	for($i=0;$i<8;$i++) { &movq(&QWP($i*8,"esi"),@mm[$i]); }    # H=L
+
+	&lea	("edi",&DWP(64,"edi"));		# inp+=64
+	&sub	("eax",1);			# num--
+	&jz	(&label("alldone"));
+	&mov	(&DWP(4,"ebx"),"edi");		# update argument block
+	&mov	(&DWP(8,"ebx"),"eax");
+	&jmp	(&label("outerloop"));
+
+&set_label("alldone");
+	&emms	();
+	&mov	("esp",&DWP(16,"ebx"));		# restore saved stack pointer
+	&pop	("edi");
+	&pop	("esi");
+	&pop	("ebx");
+	&pop	("ebp");
+	&ret	();
+
+&align(64);
+&set_label("table");
+	&LL(0x18,0x18,0x60,0x18,0xc0,0x78,0x30,0xd8);
+	&LL(0x23,0x23,0x8c,0x23,0x05,0xaf,0x46,0x26);
+	&LL(0xc6,0xc6,0x3f,0xc6,0x7e,0xf9,0x91,0xb8);
+	&LL(0xe8,0xe8,0x87,0xe8,0x13,0x6f,0xcd,0xfb);
+	&LL(0x87,0x87,0x26,0x87,0x4c,0xa1,0x13,0xcb);
+	&LL(0xb8,0xb8,0xda,0xb8,0xa9,0x62,0x6d,0x11);
+	&LL(0x01,0x01,0x04,0x01,0x08,0x05,0x02,0x09);
+	&LL(0x4f,0x4f,0x21,0x4f,0x42,0x6e,0x9e,0x0d);
+	&LL(0x36,0x36,0xd8,0x36,0xad,0xee,0x6c,0x9b);
+	&LL(0xa6,0xa6,0xa2,0xa6,0x59,0x04,0x51,0xff);
+	&LL(0xd2,0xd2,0x6f,0xd2,0xde,0xbd,0xb9,0x0c);
+	&LL(0xf5,0xf5,0xf3,0xf5,0xfb,0x06,0xf7,0x0e);
+	&LL(0x79,0x79,0xf9,0x79,0xef,0x80,0xf2,0x96);
+	&LL(0x6f,0x6f,0xa1,0x6f,0x5f,0xce,0xde,0x30);
+	&LL(0x91,0x91,0x7e,0x91,0xfc,0xef,0x3f,0x6d);
+	&LL(0x52,0x52,0x55,0x52,0xaa,0x07,0xa4,0xf8);
+	&LL(0x60,0x60,0x9d,0x60,0x27,0xfd,0xc0,0x47);
+	&LL(0xbc,0xbc,0xca,0xbc,0x89,0x76,0x65,0x35);
+	&LL(0x9b,0x9b,0x56,0x9b,0xac,0xcd,0x2b,0x37);
+	&LL(0x8e,0x8e,0x02,0x8e,0x04,0x8c,0x01,0x8a);
+	&LL(0xa3,0xa3,0xb6,0xa3,0x71,0x15,0x5b,0xd2);
+	&LL(0x0c,0x0c,0x30,0x0c,0x60,0x3c,0x18,0x6c);
+	&LL(0x7b,0x7b,0xf1,0x7b,0xff,0x8a,0xf6,0x84);
+	&LL(0x35,0x35,0xd4,0x35,0xb5,0xe1,0x6a,0x80);
+	&LL(0x1d,0x1d,0x74,0x1d,0xe8,0x69,0x3a,0xf5);
+	&LL(0xe0,0xe0,0xa7,0xe0,0x53,0x47,0xdd,0xb3);
+	&LL(0xd7,0xd7,0x7b,0xd7,0xf6,0xac,0xb3,0x21);
+	&LL(0xc2,0xc2,0x2f,0xc2,0x5e,0xed,0x99,0x9c);
+	&LL(0x2e,0x2e,0xb8,0x2e,0x6d,0x96,0x5c,0x43);
+	&LL(0x4b,0x4b,0x31,0x4b,0x62,0x7a,0x96,0x29);
+	&LL(0xfe,0xfe,0xdf,0xfe,0xa3,0x21,0xe1,0x5d);
+	&LL(0x57,0x57,0x41,0x57,0x82,0x16,0xae,0xd5);
+	&LL(0x15,0x15,0x54,0x15,0xa8,0x41,0x2a,0xbd);
+	&LL(0x77,0x77,0xc1,0x77,0x9f,0xb6,0xee,0xe8);
+	&LL(0x37,0x37,0xdc,0x37,0xa5,0xeb,0x6e,0x92);
+	&LL(0xe5,0xe5,0xb3,0xe5,0x7b,0x56,0xd7,0x9e);
+	&LL(0x9f,0x9f,0x46,0x9f,0x8c,0xd9,0x23,0x13);
+	&LL(0xf0,0xf0,0xe7,0xf0,0xd3,0x17,0xfd,0x23);
+	&LL(0x4a,0x4a,0x35,0x4a,0x6a,0x7f,0x94,0x20);
+	&LL(0xda,0xda,0x4f,0xda,0x9e,0x95,0xa9,0x44);
+	&LL(0x58,0x58,0x7d,0x58,0xfa,0x25,0xb0,0xa2);
+	&LL(0xc9,0xc9,0x03,0xc9,0x06,0xca,0x8f,0xcf);
+	&LL(0x29,0x29,0xa4,0x29,0x55,0x8d,0x52,0x7c);
+	&LL(0x0a,0x0a,0x28,0x0a,0x50,0x22,0x14,0x5a);
+	&LL(0xb1,0xb1,0xfe,0xb1,0xe1,0x4f,0x7f,0x50);
+	&LL(0xa0,0xa0,0xba,0xa0,0x69,0x1a,0x5d,0xc9);
+	&LL(0x6b,0x6b,0xb1,0x6b,0x7f,0xda,0xd6,0x14);
+	&LL(0x85,0x85,0x2e,0x85,0x5c,0xab,0x17,0xd9);
+	&LL(0xbd,0xbd,0xce,0xbd,0x81,0x73,0x67,0x3c);
+	&LL(0x5d,0x5d,0x69,0x5d,0xd2,0x34,0xba,0x8f);
+	&LL(0x10,0x10,0x40,0x10,0x80,0x50,0x20,0x90);
+	&LL(0xf4,0xf4,0xf7,0xf4,0xf3,0x03,0xf5,0x07);
+	&LL(0xcb,0xcb,0x0b,0xcb,0x16,0xc0,0x8b,0xdd);
+	&LL(0x3e,0x3e,0xf8,0x3e,0xed,0xc6,0x7c,0xd3);
+	&LL(0x05,0x05,0x14,0x05,0x28,0x11,0x0a,0x2d);
+	&LL(0x67,0x67,0x81,0x67,0x1f,0xe6,0xce,0x78);
+	&LL(0xe4,0xe4,0xb7,0xe4,0x73,0x53,0xd5,0x97);
+	&LL(0x27,0x27,0x9c,0x27,0x25,0xbb,0x4e,0x02);
+	&LL(0x41,0x41,0x19,0x41,0x32,0x58,0x82,0x73);
+	&LL(0x8b,0x8b,0x16,0x8b,0x2c,0x9d,0x0b,0xa7);
+	&LL(0xa7,0xa7,0xa6,0xa7,0x51,0x01,0x53,0xf6);
+	&LL(0x7d,0x7d,0xe9,0x7d,0xcf,0x94,0xfa,0xb2);
+	&LL(0x95,0x95,0x6e,0x95,0xdc,0xfb,0x37,0x49);
+	&LL(0xd8,0xd8,0x47,0xd8,0x8e,0x9f,0xad,0x56);
+	&LL(0xfb,0xfb,0xcb,0xfb,0x8b,0x30,0xeb,0x70);
+	&LL(0xee,0xee,0x9f,0xee,0x23,0x71,0xc1,0xcd);
+	&LL(0x7c,0x7c,0xed,0x7c,0xc7,0x91,0xf8,0xbb);
+	&LL(0x66,0x66,0x85,0x66,0x17,0xe3,0xcc,0x71);
+	&LL(0xdd,0xdd,0x53,0xdd,0xa6,0x8e,0xa7,0x7b);
+	&LL(0x17,0x17,0x5c,0x17,0xb8,0x4b,0x2e,0xaf);
+	&LL(0x47,0x47,0x01,0x47,0x02,0x46,0x8e,0x45);
+	&LL(0x9e,0x9e,0x42,0x9e,0x84,0xdc,0x21,0x1a);
+	&LL(0xca,0xca,0x0f,0xca,0x1e,0xc5,0x89,0xd4);
+	&LL(0x2d,0x2d,0xb4,0x2d,0x75,0x99,0x5a,0x58);
+	&LL(0xbf,0xbf,0xc6,0xbf,0x91,0x79,0x63,0x2e);
+	&LL(0x07,0x07,0x1c,0x07,0x38,0x1b,0x0e,0x3f);
+	&LL(0xad,0xad,0x8e,0xad,0x01,0x23,0x47,0xac);
+	&LL(0x5a,0x5a,0x75,0x5a,0xea,0x2f,0xb4,0xb0);
+	&LL(0x83,0x83,0x36,0x83,0x6c,0xb5,0x1b,0xef);
+	&LL(0x33,0x33,0xcc,0x33,0x85,0xff,0x66,0xb6);
+	&LL(0x63,0x63,0x91,0x63,0x3f,0xf2,0xc6,0x5c);
+	&LL(0x02,0x02,0x08,0x02,0x10,0x0a,0x04,0x12);
+	&LL(0xaa,0xaa,0x92,0xaa,0x39,0x38,0x49,0x93);
+	&LL(0x71,0x71,0xd9,0x71,0xaf,0xa8,0xe2,0xde);
+	&LL(0xc8,0xc8,0x07,0xc8,0x0e,0xcf,0x8d,0xc6);
+	&LL(0x19,0x19,0x64,0x19,0xc8,0x7d,0x32,0xd1);
+	&LL(0x49,0x49,0x39,0x49,0x72,0x70,0x92,0x3b);
+	&LL(0xd9,0xd9,0x43,0xd9,0x86,0x9a,0xaf,0x5f);
+	&LL(0xf2,0xf2,0xef,0xf2,0xc3,0x1d,0xf9,0x31);
+	&LL(0xe3,0xe3,0xab,0xe3,0x4b,0x48,0xdb,0xa8);
+	&LL(0x5b,0x5b,0x71,0x5b,0xe2,0x2a,0xb6,0xb9);
+	&LL(0x88,0x88,0x1a,0x88,0x34,0x92,0x0d,0xbc);
+	&LL(0x9a,0x9a,0x52,0x9a,0xa4,0xc8,0x29,0x3e);
+	&LL(0x26,0x26,0x98,0x26,0x2d,0xbe,0x4c,0x0b);
+	&LL(0x32,0x32,0xc8,0x32,0x8d,0xfa,0x64,0xbf);
+	&LL(0xb0,0xb0,0xfa,0xb0,0xe9,0x4a,0x7d,0x59);
+	&LL(0xe9,0xe9,0x83,0xe9,0x1b,0x6a,0xcf,0xf2);
+	&LL(0x0f,0x0f,0x3c,0x0f,0x78,0x33,0x1e,0x77);
+	&LL(0xd5,0xd5,0x73,0xd5,0xe6,0xa6,0xb7,0x33);
+	&LL(0x80,0x80,0x3a,0x80,0x74,0xba,0x1d,0xf4);
+	&LL(0xbe,0xbe,0xc2,0xbe,0x99,0x7c,0x61,0x27);
+	&LL(0xcd,0xcd,0x13,0xcd,0x26,0xde,0x87,0xeb);
+	&LL(0x34,0x34,0xd0,0x34,0xbd,0xe4,0x68,0x89);
+	&LL(0x48,0x48,0x3d,0x48,0x7a,0x75,0x90,0x32);
+	&LL(0xff,0xff,0xdb,0xff,0xab,0x24,0xe3,0x54);
+	&LL(0x7a,0x7a,0xf5,0x7a,0xf7,0x8f,0xf4,0x8d);
+	&LL(0x90,0x90,0x7a,0x90,0xf4,0xea,0x3d,0x64);
+	&LL(0x5f,0x5f,0x61,0x5f,0xc2,0x3e,0xbe,0x9d);
+	&LL(0x20,0x20,0x80,0x20,0x1d,0xa0,0x40,0x3d);
+	&LL(0x68,0x68,0xbd,0x68,0x67,0xd5,0xd0,0x0f);
+	&LL(0x1a,0x1a,0x68,0x1a,0xd0,0x72,0x34,0xca);
+	&LL(0xae,0xae,0x82,0xae,0x19,0x2c,0x41,0xb7);
+	&LL(0xb4,0xb4,0xea,0xb4,0xc9,0x5e,0x75,0x7d);
+	&LL(0x54,0x54,0x4d,0x54,0x9a,0x19,0xa8,0xce);
+	&LL(0x93,0x93,0x76,0x93,0xec,0xe5,0x3b,0x7f);
+	&LL(0x22,0x22,0x88,0x22,0x0d,0xaa,0x44,0x2f);
+	&LL(0x64,0x64,0x8d,0x64,0x07,0xe9,0xc8,0x63);
+	&LL(0xf1,0xf1,0xe3,0xf1,0xdb,0x12,0xff,0x2a);
+	&LL(0x73,0x73,0xd1,0x73,0xbf,0xa2,0xe6,0xcc);
+	&LL(0x12,0x12,0x48,0x12,0x90,0x5a,0x24,0x82);
+	&LL(0x40,0x40,0x1d,0x40,0x3a,0x5d,0x80,0x7a);
+	&LL(0x08,0x08,0x20,0x08,0x40,0x28,0x10,0x48);
+	&LL(0xc3,0xc3,0x2b,0xc3,0x56,0xe8,0x9b,0x95);
+	&LL(0xec,0xec,0x97,0xec,0x33,0x7b,0xc5,0xdf);
+	&LL(0xdb,0xdb,0x4b,0xdb,0x96,0x90,0xab,0x4d);
+	&LL(0xa1,0xa1,0xbe,0xa1,0x61,0x1f,0x5f,0xc0);
+	&LL(0x8d,0x8d,0x0e,0x8d,0x1c,0x83,0x07,0x91);
+	&LL(0x3d,0x3d,0xf4,0x3d,0xf5,0xc9,0x7a,0xc8);
+	&LL(0x97,0x97,0x66,0x97,0xcc,0xf1,0x33,0x5b);
+	&LL(0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00);
+	&LL(0xcf,0xcf,0x1b,0xcf,0x36,0xd4,0x83,0xf9);
+	&LL(0x2b,0x2b,0xac,0x2b,0x45,0x87,0x56,0x6e);
+	&LL(0x76,0x76,0xc5,0x76,0x97,0xb3,0xec,0xe1);
+	&LL(0x82,0x82,0x32,0x82,0x64,0xb0,0x19,0xe6);
+	&LL(0xd6,0xd6,0x7f,0xd6,0xfe,0xa9,0xb1,0x28);
+	&LL(0x1b,0x1b,0x6c,0x1b,0xd8,0x77,0x36,0xc3);
+	&LL(0xb5,0xb5,0xee,0xb5,0xc1,0x5b,0x77,0x74);
+	&LL(0xaf,0xaf,0x86,0xaf,0x11,0x29,0x43,0xbe);
+	&LL(0x6a,0x6a,0xb5,0x6a,0x77,0xdf,0xd4,0x1d);
+	&LL(0x50,0x50,0x5d,0x50,0xba,0x0d,0xa0,0xea);
+	&LL(0x45,0x45,0x09,0x45,0x12,0x4c,0x8a,0x57);
+	&LL(0xf3,0xf3,0xeb,0xf3,0xcb,0x18,0xfb,0x38);
+	&LL(0x30,0x30,0xc0,0x30,0x9d,0xf0,0x60,0xad);
+	&LL(0xef,0xef,0x9b,0xef,0x2b,0x74,0xc3,0xc4);
+	&LL(0x3f,0x3f,0xfc,0x3f,0xe5,0xc3,0x7e,0xda);
+	&LL(0x55,0x55,0x49,0x55,0x92,0x1c,0xaa,0xc7);
+	&LL(0xa2,0xa2,0xb2,0xa2,0x79,0x10,0x59,0xdb);
+	&LL(0xea,0xea,0x8f,0xea,0x03,0x65,0xc9,0xe9);
+	&LL(0x65,0x65,0x89,0x65,0x0f,0xec,0xca,0x6a);
+	&LL(0xba,0xba,0xd2,0xba,0xb9,0x68,0x69,0x03);
+	&LL(0x2f,0x2f,0xbc,0x2f,0x65,0x93,0x5e,0x4a);
+	&LL(0xc0,0xc0,0x27,0xc0,0x4e,0xe7,0x9d,0x8e);
+	&LL(0xde,0xde,0x5f,0xde,0xbe,0x81,0xa1,0x60);
+	&LL(0x1c,0x1c,0x70,0x1c,0xe0,0x6c,0x38,0xfc);
+	&LL(0xfd,0xfd,0xd3,0xfd,0xbb,0x2e,0xe7,0x46);
+	&LL(0x4d,0x4d,0x29,0x4d,0x52,0x64,0x9a,0x1f);
+	&LL(0x92,0x92,0x72,0x92,0xe4,0xe0,0x39,0x76);
+	&LL(0x75,0x75,0xc9,0x75,0x8f,0xbc,0xea,0xfa);
+	&LL(0x06,0x06,0x18,0x06,0x30,0x1e,0x0c,0x36);
+	&LL(0x8a,0x8a,0x12,0x8a,0x24,0x98,0x09,0xae);
+	&LL(0xb2,0xb2,0xf2,0xb2,0xf9,0x40,0x79,0x4b);
+	&LL(0xe6,0xe6,0xbf,0xe6,0x63,0x59,0xd1,0x85);
+	&LL(0x0e,0x0e,0x38,0x0e,0x70,0x36,0x1c,0x7e);
+	&LL(0x1f,0x1f,0x7c,0x1f,0xf8,0x63,0x3e,0xe7);
+	&LL(0x62,0x62,0x95,0x62,0x37,0xf7,0xc4,0x55);
+	&LL(0xd4,0xd4,0x77,0xd4,0xee,0xa3,0xb5,0x3a);
+	&LL(0xa8,0xa8,0x9a,0xa8,0x29,0x32,0x4d,0x81);
+	&LL(0x96,0x96,0x62,0x96,0xc4,0xf4,0x31,0x52);
+	&LL(0xf9,0xf9,0xc3,0xf9,0x9b,0x3a,0xef,0x62);
+	&LL(0xc5,0xc5,0x33,0xc5,0x66,0xf6,0x97,0xa3);
+	&LL(0x25,0x25,0x94,0x25,0x35,0xb1,0x4a,0x10);
+	&LL(0x59,0x59,0x79,0x59,0xf2,0x20,0xb2,0xab);
+	&LL(0x84,0x84,0x2a,0x84,0x54,0xae,0x15,0xd0);
+	&LL(0x72,0x72,0xd5,0x72,0xb7,0xa7,0xe4,0xc5);
+	&LL(0x39,0x39,0xe4,0x39,0xd5,0xdd,0x72,0xec);
+	&LL(0x4c,0x4c,0x2d,0x4c,0x5a,0x61,0x98,0x16);
+	&LL(0x5e,0x5e,0x65,0x5e,0xca,0x3b,0xbc,0x94);
+	&LL(0x78,0x78,0xfd,0x78,0xe7,0x85,0xf0,0x9f);
+	&LL(0x38,0x38,0xe0,0x38,0xdd,0xd8,0x70,0xe5);
+	&LL(0x8c,0x8c,0x0a,0x8c,0x14,0x86,0x05,0x98);
+	&LL(0xd1,0xd1,0x63,0xd1,0xc6,0xb2,0xbf,0x17);
+	&LL(0xa5,0xa5,0xae,0xa5,0x41,0x0b,0x57,0xe4);
+	&LL(0xe2,0xe2,0xaf,0xe2,0x43,0x4d,0xd9,0xa1);
+	&LL(0x61,0x61,0x99,0x61,0x2f,0xf8,0xc2,0x4e);
+	&LL(0xb3,0xb3,0xf6,0xb3,0xf1,0x45,0x7b,0x42);
+	&LL(0x21,0x21,0x84,0x21,0x15,0xa5,0x42,0x34);
+	&LL(0x9c,0x9c,0x4a,0x9c,0x94,0xd6,0x25,0x08);
+	&LL(0x1e,0x1e,0x78,0x1e,0xf0,0x66,0x3c,0xee);
+	&LL(0x43,0x43,0x11,0x43,0x22,0x52,0x86,0x61);
+	&LL(0xc7,0xc7,0x3b,0xc7,0x76,0xfc,0x93,0xb1);
+	&LL(0xfc,0xfc,0xd7,0xfc,0xb3,0x2b,0xe5,0x4f);
+	&LL(0x04,0x04,0x10,0x04,0x20,0x14,0x08,0x24);
+	&LL(0x51,0x51,0x59,0x51,0xb2,0x08,0xa2,0xe3);
+	&LL(0x99,0x99,0x5e,0x99,0xbc,0xc7,0x2f,0x25);
+	&LL(0x6d,0x6d,0xa9,0x6d,0x4f,0xc4,0xda,0x22);
+	&LL(0x0d,0x0d,0x34,0x0d,0x68,0x39,0x1a,0x65);
+	&LL(0xfa,0xfa,0xcf,0xfa,0x83,0x35,0xe9,0x79);
+	&LL(0xdf,0xdf,0x5b,0xdf,0xb6,0x84,0xa3,0x69);
+	&LL(0x7e,0x7e,0xe5,0x7e,0xd7,0x9b,0xfc,0xa9);
+	&LL(0x24,0x24,0x90,0x24,0x3d,0xb4,0x48,0x19);
+	&LL(0x3b,0x3b,0xec,0x3b,0xc5,0xd7,0x76,0xfe);
+	&LL(0xab,0xab,0x96,0xab,0x31,0x3d,0x4b,0x9a);
+	&LL(0xce,0xce,0x1f,0xce,0x3e,0xd1,0x81,0xf0);
+	&LL(0x11,0x11,0x44,0x11,0x88,0x55,0x22,0x99);
+	&LL(0x8f,0x8f,0x06,0x8f,0x0c,0x89,0x03,0x83);
+	&LL(0x4e,0x4e,0x25,0x4e,0x4a,0x6b,0x9c,0x04);
+	&LL(0xb7,0xb7,0xe6,0xb7,0xd1,0x51,0x73,0x66);
+	&LL(0xeb,0xeb,0x8b,0xeb,0x0b,0x60,0xcb,0xe0);
+	&LL(0x3c,0x3c,0xf0,0x3c,0xfd,0xcc,0x78,0xc1);
+	&LL(0x81,0x81,0x3e,0x81,0x7c,0xbf,0x1f,0xfd);
+	&LL(0x94,0x94,0x6a,0x94,0xd4,0xfe,0x35,0x40);
+	&LL(0xf7,0xf7,0xfb,0xf7,0xeb,0x0c,0xf3,0x1c);
+	&LL(0xb9,0xb9,0xde,0xb9,0xa1,0x67,0x6f,0x18);
+	&LL(0x13,0x13,0x4c,0x13,0x98,0x5f,0x26,0x8b);
+	&LL(0x2c,0x2c,0xb0,0x2c,0x7d,0x9c,0x58,0x51);
+	&LL(0xd3,0xd3,0x6b,0xd3,0xd6,0xb8,0xbb,0x05);
+	&LL(0xe7,0xe7,0xbb,0xe7,0x6b,0x5c,0xd3,0x8c);
+	&LL(0x6e,0x6e,0xa5,0x6e,0x57,0xcb,0xdc,0x39);
+	&LL(0xc4,0xc4,0x37,0xc4,0x6e,0xf3,0x95,0xaa);
+	&LL(0x03,0x03,0x0c,0x03,0x18,0x0f,0x06,0x1b);
+	&LL(0x56,0x56,0x45,0x56,0x8a,0x13,0xac,0xdc);
+	&LL(0x44,0x44,0x0d,0x44,0x1a,0x49,0x88,0x5e);
+	&LL(0x7f,0x7f,0xe1,0x7f,0xdf,0x9e,0xfe,0xa0);
+	&LL(0xa9,0xa9,0x9e,0xa9,0x21,0x37,0x4f,0x88);
+	&LL(0x2a,0x2a,0xa8,0x2a,0x4d,0x82,0x54,0x67);
+	&LL(0xbb,0xbb,0xd6,0xbb,0xb1,0x6d,0x6b,0x0a);
+	&LL(0xc1,0xc1,0x23,0xc1,0x46,0xe2,0x9f,0x87);
+	&LL(0x53,0x53,0x51,0x53,0xa2,0x02,0xa6,0xf1);
+	&LL(0xdc,0xdc,0x57,0xdc,0xae,0x8b,0xa5,0x72);
+	&LL(0x0b,0x0b,0x2c,0x0b,0x58,0x27,0x16,0x53);
+	&LL(0x9d,0x9d,0x4e,0x9d,0x9c,0xd3,0x27,0x01);
+	&LL(0x6c,0x6c,0xad,0x6c,0x47,0xc1,0xd8,0x2b);
+	&LL(0x31,0x31,0xc4,0x31,0x95,0xf5,0x62,0xa4);
+	&LL(0x74,0x74,0xcd,0x74,0x87,0xb9,0xe8,0xf3);
+	&LL(0xf6,0xf6,0xff,0xf6,0xe3,0x09,0xf1,0x15);
+	&LL(0x46,0x46,0x05,0x46,0x0a,0x43,0x8c,0x4c);
+	&LL(0xac,0xac,0x8a,0xac,0x09,0x26,0x45,0xa5);
+	&LL(0x89,0x89,0x1e,0x89,0x3c,0x97,0x0f,0xb5);
+	&LL(0x14,0x14,0x50,0x14,0xa0,0x44,0x28,0xb4);
+	&LL(0xe1,0xe1,0xa3,0xe1,0x5b,0x42,0xdf,0xba);
+	&LL(0x16,0x16,0x58,0x16,0xb0,0x4e,0x2c,0xa6);
+	&LL(0x3a,0x3a,0xe8,0x3a,0xcd,0xd2,0x74,0xf7);
+	&LL(0x69,0x69,0xb9,0x69,0x6f,0xd0,0xd2,0x06);
+	&LL(0x09,0x09,0x24,0x09,0x48,0x2d,0x12,0x41);
+	&LL(0x70,0x70,0xdd,0x70,0xa7,0xad,0xe0,0xd7);
+	&LL(0xb6,0xb6,0xe2,0xb6,0xd9,0x54,0x71,0x6f);
+	&LL(0xd0,0xd0,0x67,0xd0,0xce,0xb7,0xbd,0x1e);
+	&LL(0xed,0xed,0x93,0xed,0x3b,0x7e,0xc7,0xd6);
+	&LL(0xcc,0xcc,0x17,0xcc,0x2e,0xdb,0x85,0xe2);
+	&LL(0x42,0x42,0x15,0x42,0x2a,0x57,0x84,0x68);
+	&LL(0x98,0x98,0x5a,0x98,0xb4,0xc2,0x2d,0x2c);
+	&LL(0xa4,0xa4,0xaa,0xa4,0x49,0x0e,0x55,0xed);
+	&LL(0x28,0x28,0xa0,0x28,0x5d,0x88,0x50,0x75);
+	&LL(0x5c,0x5c,0x6d,0x5c,0xda,0x31,0xb8,0x86);
+	&LL(0xf8,0xf8,0xc7,0xf8,0x93,0x3f,0xed,0x6b);
+	&LL(0x86,0x86,0x22,0x86,0x44,0xa4,0x11,0xc2);
+
+	&L(0x18,0x23,0xc6,0xe8,0x87,0xb8,0x01,0x4f);	# rc[ROUNDS]
+	&L(0x36,0xa6,0xd2,0xf5,0x79,0x6f,0x91,0x52);
+	&L(0x60,0xbc,0x9b,0x8e,0xa3,0x0c,0x7b,0x35);
+	&L(0x1d,0xe0,0xd7,0xc2,0x2e,0x4b,0xfe,0x57);
+	&L(0x15,0x77,0x37,0xe5,0x9f,0xf0,0x4a,0xda);
+	&L(0x58,0xc9,0x29,0x0a,0xb1,0xa0,0x6b,0x85);
+	&L(0xbd,0x5d,0x10,0xf4,0xcb,0x3e,0x05,0x67);
+	&L(0xe4,0x27,0x41,0x8b,0xa7,0x7d,0x95,0xd8);
+	&L(0xfb,0xee,0x7c,0x66,0xdd,0x17,0x47,0x9e);
+	&L(0xca,0x2d,0xbf,0x07,0xad,0x5a,0x83,0x33);
+
+&function_end_B("whirlpool_block_mmx");
+&asm_finish(); 
diff --git a/crypto/whrlpool/asm/wp-x86_64.pl b/crypto/whrlpool/asm/wp-x86_64.pl
new file mode 100755
index 000000000000..87c0843dc1d3
--- /dev/null
+++ b/crypto/whrlpool/asm/wp-x86_64.pl
@@ -0,0 +1,589 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. Rights for redistribution and usage in source and binary
+# forms are granted according to the OpenSSL license.
+# ====================================================================
+#
+# whirlpool_block for x86_64.
+#
+# 2500 cycles per 64-byte input block on AMD64, which is *identical*
+# to 32-bit MMX version executed on same CPU. So why did I bother?
+# Well, it's faster than gcc 3.3.2 generated code by over 50%, and
+# over 80% faster than PathScale 1.4, an "ambitious" commercial
+# compiler. Furthermore it surpasses gcc 3.4.3 by 170% and Sun Studio
+# 10 - by 360%[!]... What is it with x86_64 compilers? It's not the
+# first example when they fail to generate more optimal code, when
+# I believe they had *all* chances to...
+#
+# Note that register and stack frame layout are virtually identical
+# to 32-bit MMX version, except that %r8-15 are used instead of
+# %mm0-8. You can even notice that K[i] and S[i] are loaded to
+# %eax:%ebx as pair of 32-bit values and not as single 64-bit one.
+# This is done in order to avoid 64-bit shift penalties on Intel
+# EM64T core. Speaking of which! I bet it's possible to improve
+# Opteron performance by compressing the table to 2KB and replacing
+# unaligned references with complementary rotations [which would
+# incidentally replace lea instructions], but it would definitely
+# just "kill" EM64T, because it has only 1 shifter/rotator [against
+# 3 on Opteron] and which is *unacceptably* slow with 64-bit
+# operand.
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; my $dir=$1; my $xlate;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+sub L() { $code.=".byte	".join(',',@_)."\n"; }
+sub LL(){ $code.=".byte	".join(',',@_).",".join(',',@_)."\n"; }
+
+@mm=("%r8","%r9","%r10","%r11","%r12","%r13","%r14","%r15");
+
+$func="whirlpool_block";
+$table=".Ltable";
+
+$code=<<___;
+.text
+
+.globl	$func
+.type	$func,\@function,3
+.align	16
+$func:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+
+	mov	%rsp,%r11
+	sub	\$128+40,%rsp
+	and	\$-64,%rsp
+
+	lea	128(%rsp),%r10
+	mov	%rdi,0(%r10)		# save parameter block
+	mov	%rsi,8(%r10)
+	mov	%rdx,16(%r10)
+	mov	%r11,32(%r10)		# saved stack pointer
+.Lprologue:
+
+	mov	%r10,%rbx
+	lea	$table(%rip),%rbp
+
+	xor	%rcx,%rcx
+	xor	%rdx,%rdx
+___
+for($i=0;$i<8;$i++) { $code.="mov $i*8(%rdi),@mm[$i]\n"; }	# L=H
+$code.=".Louterloop:\n";
+for($i=0;$i<8;$i++) { $code.="mov @mm[$i],$i*8(%rsp)\n"; }	# K=L
+for($i=0;$i<8;$i++) { $code.="xor $i*8(%rsi),@mm[$i]\n"; }	# L^=inp
+for($i=0;$i<8;$i++) { $code.="mov @mm[$i],64+$i*8(%rsp)\n"; }	# S=L
+$code.=<<___;
+	xor	%rsi,%rsi
+	mov	%rsi,24(%rbx)		# zero round counter
+.align	16
+.Lround:
+	mov	4096(%rbp,%rsi,8),@mm[0]	# rc[r]
+	mov	0(%rsp),%eax
+	mov	4(%rsp),%ebx
+___
+for($i=0;$i<8;$i++) {
+    my $func = ($i==0)? "mov" : "xor";
+    $code.=<<___;
+	mov	%al,%cl
+	mov	%ah,%dl
+	lea	(%rcx,%rcx),%rsi
+	lea	(%rdx,%rdx),%rdi
+	shr	\$16,%eax
+	xor	0(%rbp,%rsi,8),@mm[0]
+	$func	7(%rbp,%rdi,8),@mm[1]
+	mov	%al,%cl
+	mov	%ah,%dl
+	mov	$i*8+8(%rsp),%eax		# ($i+1)*8
+	lea	(%rcx,%rcx),%rsi
+	lea	(%rdx,%rdx),%rdi
+	$func	6(%rbp,%rsi,8),@mm[2]
+	$func	5(%rbp,%rdi,8),@mm[3]
+	mov	%bl,%cl
+	mov	%bh,%dl
+	lea	(%rcx,%rcx),%rsi
+	lea	(%rdx,%rdx),%rdi
+	shr	\$16,%ebx
+	$func	4(%rbp,%rsi,8),@mm[4]
+	$func	3(%rbp,%rdi,8),@mm[5]
+	mov	%bl,%cl
+	mov	%bh,%dl
+	mov	$i*8+8+4(%rsp),%ebx		# ($i+1)*8+4
+	lea	(%rcx,%rcx),%rsi
+	lea	(%rdx,%rdx),%rdi
+	$func	2(%rbp,%rsi,8),@mm[6]
+	$func	1(%rbp,%rdi,8),@mm[7]
+___
+    push(@mm,shift(@mm));
+}
+for($i=0;$i<8;$i++) { $code.="mov @mm[$i],$i*8(%rsp)\n"; }	# K=L
+for($i=0;$i<8;$i++) {
+    $code.=<<___;
+	mov	%al,%cl
+	mov	%ah,%dl
+	lea	(%rcx,%rcx),%rsi
+	lea	(%rdx,%rdx),%rdi
+	shr	\$16,%eax
+	xor	0(%rbp,%rsi,8),@mm[0]
+	xor	7(%rbp,%rdi,8),@mm[1]
+	mov	%al,%cl
+	mov	%ah,%dl
+	`"mov	64+$i*8+8(%rsp),%eax"	if($i<7);`	# 64+($i+1)*8
+	lea	(%rcx,%rcx),%rsi
+	lea	(%rdx,%rdx),%rdi
+	xor	6(%rbp,%rsi,8),@mm[2]
+	xor	5(%rbp,%rdi,8),@mm[3]
+	mov	%bl,%cl
+	mov	%bh,%dl
+	lea	(%rcx,%rcx),%rsi
+	lea	(%rdx,%rdx),%rdi
+	shr	\$16,%ebx
+	xor	4(%rbp,%rsi,8),@mm[4]
+	xor	3(%rbp,%rdi,8),@mm[5]
+	mov	%bl,%cl
+	mov	%bh,%dl
+	`"mov	64+$i*8+8+4(%rsp),%ebx"	if($i<7);`	# 64+($i+1)*8+4
+	lea	(%rcx,%rcx),%rsi
+	lea	(%rdx,%rdx),%rdi
+	xor	2(%rbp,%rsi,8),@mm[6]
+	xor	1(%rbp,%rdi,8),@mm[7]
+___
+    push(@mm,shift(@mm));
+}
+$code.=<<___;
+	lea	128(%rsp),%rbx
+	mov	24(%rbx),%rsi		# pull round counter
+	add	\$1,%rsi
+	cmp	\$10,%rsi
+	je	.Lroundsdone
+
+	mov	%rsi,24(%rbx)		# update round counter
+___
+for($i=0;$i<8;$i++) { $code.="mov @mm[$i],64+$i*8(%rsp)\n"; }	# S=L
+$code.=<<___;
+	jmp	.Lround
+.align	16
+.Lroundsdone:
+	mov	0(%rbx),%rdi		# reload argument block
+	mov	8(%rbx),%rsi
+	mov	16(%rbx),%rax
+___
+for($i=0;$i<8;$i++) { $code.="xor $i*8(%rsi),@mm[$i]\n"; }	# L^=inp
+for($i=0;$i<8;$i++) { $code.="xor $i*8(%rdi),@mm[$i]\n"; }	# L^=H
+for($i=0;$i<8;$i++) { $code.="mov @mm[$i],$i*8(%rdi)\n"; }	# H=L
+$code.=<<___;
+	lea	64(%rsi),%rsi		# inp+=64
+	sub	\$1,%rax		# num--
+	jz	.Lalldone
+	mov	%rsi,8(%rbx)		# update parameter block
+	mov	%rax,16(%rbx)
+	jmp	.Louterloop
+.Lalldone:
+	mov	32(%rbx),%rsi		# restore saved pointer
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lepilogue:
+	ret
+.size	$func,.-$func
+
+.align	64
+.type	$table,\@object
+$table:
+___
+	&LL(0x18,0x18,0x60,0x18,0xc0,0x78,0x30,0xd8);
+	&LL(0x23,0x23,0x8c,0x23,0x05,0xaf,0x46,0x26);
+	&LL(0xc6,0xc6,0x3f,0xc6,0x7e,0xf9,0x91,0xb8);
+	&LL(0xe8,0xe8,0x87,0xe8,0x13,0x6f,0xcd,0xfb);
+	&LL(0x87,0x87,0x26,0x87,0x4c,0xa1,0x13,0xcb);
+	&LL(0xb8,0xb8,0xda,0xb8,0xa9,0x62,0x6d,0x11);
+	&LL(0x01,0x01,0x04,0x01,0x08,0x05,0x02,0x09);
+	&LL(0x4f,0x4f,0x21,0x4f,0x42,0x6e,0x9e,0x0d);
+	&LL(0x36,0x36,0xd8,0x36,0xad,0xee,0x6c,0x9b);
+	&LL(0xa6,0xa6,0xa2,0xa6,0x59,0x04,0x51,0xff);
+	&LL(0xd2,0xd2,0x6f,0xd2,0xde,0xbd,0xb9,0x0c);
+	&LL(0xf5,0xf5,0xf3,0xf5,0xfb,0x06,0xf7,0x0e);
+	&LL(0x79,0x79,0xf9,0x79,0xef,0x80,0xf2,0x96);
+	&LL(0x6f,0x6f,0xa1,0x6f,0x5f,0xce,0xde,0x30);
+	&LL(0x91,0x91,0x7e,0x91,0xfc,0xef,0x3f,0x6d);
+	&LL(0x52,0x52,0x55,0x52,0xaa,0x07,0xa4,0xf8);
+	&LL(0x60,0x60,0x9d,0x60,0x27,0xfd,0xc0,0x47);
+	&LL(0xbc,0xbc,0xca,0xbc,0x89,0x76,0x65,0x35);
+	&LL(0x9b,0x9b,0x56,0x9b,0xac,0xcd,0x2b,0x37);
+	&LL(0x8e,0x8e,0x02,0x8e,0x04,0x8c,0x01,0x8a);
+	&LL(0xa3,0xa3,0xb6,0xa3,0x71,0x15,0x5b,0xd2);
+	&LL(0x0c,0x0c,0x30,0x0c,0x60,0x3c,0x18,0x6c);
+	&LL(0x7b,0x7b,0xf1,0x7b,0xff,0x8a,0xf6,0x84);
+	&LL(0x35,0x35,0xd4,0x35,0xb5,0xe1,0x6a,0x80);
+	&LL(0x1d,0x1d,0x74,0x1d,0xe8,0x69,0x3a,0xf5);
+	&LL(0xe0,0xe0,0xa7,0xe0,0x53,0x47,0xdd,0xb3);
+	&LL(0xd7,0xd7,0x7b,0xd7,0xf6,0xac,0xb3,0x21);
+	&LL(0xc2,0xc2,0x2f,0xc2,0x5e,0xed,0x99,0x9c);
+	&LL(0x2e,0x2e,0xb8,0x2e,0x6d,0x96,0x5c,0x43);
+	&LL(0x4b,0x4b,0x31,0x4b,0x62,0x7a,0x96,0x29);
+	&LL(0xfe,0xfe,0xdf,0xfe,0xa3,0x21,0xe1,0x5d);
+	&LL(0x57,0x57,0x41,0x57,0x82,0x16,0xae,0xd5);
+	&LL(0x15,0x15,0x54,0x15,0xa8,0x41,0x2a,0xbd);
+	&LL(0x77,0x77,0xc1,0x77,0x9f,0xb6,0xee,0xe8);
+	&LL(0x37,0x37,0xdc,0x37,0xa5,0xeb,0x6e,0x92);
+	&LL(0xe5,0xe5,0xb3,0xe5,0x7b,0x56,0xd7,0x9e);
+	&LL(0x9f,0x9f,0x46,0x9f,0x8c,0xd9,0x23,0x13);
+	&LL(0xf0,0xf0,0xe7,0xf0,0xd3,0x17,0xfd,0x23);
+	&LL(0x4a,0x4a,0x35,0x4a,0x6a,0x7f,0x94,0x20);
+	&LL(0xda,0xda,0x4f,0xda,0x9e,0x95,0xa9,0x44);
+	&LL(0x58,0x58,0x7d,0x58,0xfa,0x25,0xb0,0xa2);
+	&LL(0xc9,0xc9,0x03,0xc9,0x06,0xca,0x8f,0xcf);
+	&LL(0x29,0x29,0xa4,0x29,0x55,0x8d,0x52,0x7c);
+	&LL(0x0a,0x0a,0x28,0x0a,0x50,0x22,0x14,0x5a);
+	&LL(0xb1,0xb1,0xfe,0xb1,0xe1,0x4f,0x7f,0x50);
+	&LL(0xa0,0xa0,0xba,0xa0,0x69,0x1a,0x5d,0xc9);
+	&LL(0x6b,0x6b,0xb1,0x6b,0x7f,0xda,0xd6,0x14);
+	&LL(0x85,0x85,0x2e,0x85,0x5c,0xab,0x17,0xd9);
+	&LL(0xbd,0xbd,0xce,0xbd,0x81,0x73,0x67,0x3c);
+	&LL(0x5d,0x5d,0x69,0x5d,0xd2,0x34,0xba,0x8f);
+	&LL(0x10,0x10,0x40,0x10,0x80,0x50,0x20,0x90);
+	&LL(0xf4,0xf4,0xf7,0xf4,0xf3,0x03,0xf5,0x07);
+	&LL(0xcb,0xcb,0x0b,0xcb,0x16,0xc0,0x8b,0xdd);
+	&LL(0x3e,0x3e,0xf8,0x3e,0xed,0xc6,0x7c,0xd3);
+	&LL(0x05,0x05,0x14,0x05,0x28,0x11,0x0a,0x2d);
+	&LL(0x67,0x67,0x81,0x67,0x1f,0xe6,0xce,0x78);
+	&LL(0xe4,0xe4,0xb7,0xe4,0x73,0x53,0xd5,0x97);
+	&LL(0x27,0x27,0x9c,0x27,0x25,0xbb,0x4e,0x02);
+	&LL(0x41,0x41,0x19,0x41,0x32,0x58,0x82,0x73);
+	&LL(0x8b,0x8b,0x16,0x8b,0x2c,0x9d,0x0b,0xa7);
+	&LL(0xa7,0xa7,0xa6,0xa7,0x51,0x01,0x53,0xf6);
+	&LL(0x7d,0x7d,0xe9,0x7d,0xcf,0x94,0xfa,0xb2);
+	&LL(0x95,0x95,0x6e,0x95,0xdc,0xfb,0x37,0x49);
+	&LL(0xd8,0xd8,0x47,0xd8,0x8e,0x9f,0xad,0x56);
+	&LL(0xfb,0xfb,0xcb,0xfb,0x8b,0x30,0xeb,0x70);
+	&LL(0xee,0xee,0x9f,0xee,0x23,0x71,0xc1,0xcd);
+	&LL(0x7c,0x7c,0xed,0x7c,0xc7,0x91,0xf8,0xbb);
+	&LL(0x66,0x66,0x85,0x66,0x17,0xe3,0xcc,0x71);
+	&LL(0xdd,0xdd,0x53,0xdd,0xa6,0x8e,0xa7,0x7b);
+	&LL(0x17,0x17,0x5c,0x17,0xb8,0x4b,0x2e,0xaf);
+	&LL(0x47,0x47,0x01,0x47,0x02,0x46,0x8e,0x45);
+	&LL(0x9e,0x9e,0x42,0x9e,0x84,0xdc,0x21,0x1a);
+	&LL(0xca,0xca,0x0f,0xca,0x1e,0xc5,0x89,0xd4);
+	&LL(0x2d,0x2d,0xb4,0x2d,0x75,0x99,0x5a,0x58);
+	&LL(0xbf,0xbf,0xc6,0xbf,0x91,0x79,0x63,0x2e);
+	&LL(0x07,0x07,0x1c,0x07,0x38,0x1b,0x0e,0x3f);
+	&LL(0xad,0xad,0x8e,0xad,0x01,0x23,0x47,0xac);
+	&LL(0x5a,0x5a,0x75,0x5a,0xea,0x2f,0xb4,0xb0);
+	&LL(0x83,0x83,0x36,0x83,0x6c,0xb5,0x1b,0xef);
+	&LL(0x33,0x33,0xcc,0x33,0x85,0xff,0x66,0xb6);
+	&LL(0x63,0x63,0x91,0x63,0x3f,0xf2,0xc6,0x5c);
+	&LL(0x02,0x02,0x08,0x02,0x10,0x0a,0x04,0x12);
+	&LL(0xaa,0xaa,0x92,0xaa,0x39,0x38,0x49,0x93);
+	&LL(0x71,0x71,0xd9,0x71,0xaf,0xa8,0xe2,0xde);
+	&LL(0xc8,0xc8,0x07,0xc8,0x0e,0xcf,0x8d,0xc6);
+	&LL(0x19,0x19,0x64,0x19,0xc8,0x7d,0x32,0xd1);
+	&LL(0x49,0x49,0x39,0x49,0x72,0x70,0x92,0x3b);
+	&LL(0xd9,0xd9,0x43,0xd9,0x86,0x9a,0xaf,0x5f);
+	&LL(0xf2,0xf2,0xef,0xf2,0xc3,0x1d,0xf9,0x31);
+	&LL(0xe3,0xe3,0xab,0xe3,0x4b,0x48,0xdb,0xa8);
+	&LL(0x5b,0x5b,0x71,0x5b,0xe2,0x2a,0xb6,0xb9);
+	&LL(0x88,0x88,0x1a,0x88,0x34,0x92,0x0d,0xbc);
+	&LL(0x9a,0x9a,0x52,0x9a,0xa4,0xc8,0x29,0x3e);
+	&LL(0x26,0x26,0x98,0x26,0x2d,0xbe,0x4c,0x0b);
+	&LL(0x32,0x32,0xc8,0x32,0x8d,0xfa,0x64,0xbf);
+	&LL(0xb0,0xb0,0xfa,0xb0,0xe9,0x4a,0x7d,0x59);
+	&LL(0xe9,0xe9,0x83,0xe9,0x1b,0x6a,0xcf,0xf2);
+	&LL(0x0f,0x0f,0x3c,0x0f,0x78,0x33,0x1e,0x77);
+	&LL(0xd5,0xd5,0x73,0xd5,0xe6,0xa6,0xb7,0x33);
+	&LL(0x80,0x80,0x3a,0x80,0x74,0xba,0x1d,0xf4);
+	&LL(0xbe,0xbe,0xc2,0xbe,0x99,0x7c,0x61,0x27);
+	&LL(0xcd,0xcd,0x13,0xcd,0x26,0xde,0x87,0xeb);
+	&LL(0x34,0x34,0xd0,0x34,0xbd,0xe4,0x68,0x89);
+	&LL(0x48,0x48,0x3d,0x48,0x7a,0x75,0x90,0x32);
+	&LL(0xff,0xff,0xdb,0xff,0xab,0x24,0xe3,0x54);
+	&LL(0x7a,0x7a,0xf5,0x7a,0xf7,0x8f,0xf4,0x8d);
+	&LL(0x90,0x90,0x7a,0x90,0xf4,0xea,0x3d,0x64);
+	&LL(0x5f,0x5f,0x61,0x5f,0xc2,0x3e,0xbe,0x9d);
+	&LL(0x20,0x20,0x80,0x20,0x1d,0xa0,0x40,0x3d);
+	&LL(0x68,0x68,0xbd,0x68,0x67,0xd5,0xd0,0x0f);
+	&LL(0x1a,0x1a,0x68,0x1a,0xd0,0x72,0x34,0xca);
+	&LL(0xae,0xae,0x82,0xae,0x19,0x2c,0x41,0xb7);
+	&LL(0xb4,0xb4,0xea,0xb4,0xc9,0x5e,0x75,0x7d);
+	&LL(0x54,0x54,0x4d,0x54,0x9a,0x19,0xa8,0xce);
+	&LL(0x93,0x93,0x76,0x93,0xec,0xe5,0x3b,0x7f);
+	&LL(0x22,0x22,0x88,0x22,0x0d,0xaa,0x44,0x2f);
+	&LL(0x64,0x64,0x8d,0x64,0x07,0xe9,0xc8,0x63);
+	&LL(0xf1,0xf1,0xe3,0xf1,0xdb,0x12,0xff,0x2a);
+	&LL(0x73,0x73,0xd1,0x73,0xbf,0xa2,0xe6,0xcc);
+	&LL(0x12,0x12,0x48,0x12,0x90,0x5a,0x24,0x82);
+	&LL(0x40,0x40,0x1d,0x40,0x3a,0x5d,0x80,0x7a);
+	&LL(0x08,0x08,0x20,0x08,0x40,0x28,0x10,0x48);
+	&LL(0xc3,0xc3,0x2b,0xc3,0x56,0xe8,0x9b,0x95);
+	&LL(0xec,0xec,0x97,0xec,0x33,0x7b,0xc5,0xdf);
+	&LL(0xdb,0xdb,0x4b,0xdb,0x96,0x90,0xab,0x4d);
+	&LL(0xa1,0xa1,0xbe,0xa1,0x61,0x1f,0x5f,0xc0);
+	&LL(0x8d,0x8d,0x0e,0x8d,0x1c,0x83,0x07,0x91);
+	&LL(0x3d,0x3d,0xf4,0x3d,0xf5,0xc9,0x7a,0xc8);
+	&LL(0x97,0x97,0x66,0x97,0xcc,0xf1,0x33,0x5b);
+	&LL(0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00);
+	&LL(0xcf,0xcf,0x1b,0xcf,0x36,0xd4,0x83,0xf9);
+	&LL(0x2b,0x2b,0xac,0x2b,0x45,0x87,0x56,0x6e);
+	&LL(0x76,0x76,0xc5,0x76,0x97,0xb3,0xec,0xe1);
+	&LL(0x82,0x82,0x32,0x82,0x64,0xb0,0x19,0xe6);
+	&LL(0xd6,0xd6,0x7f,0xd6,0xfe,0xa9,0xb1,0x28);
+	&LL(0x1b,0x1b,0x6c,0x1b,0xd8,0x77,0x36,0xc3);
+	&LL(0xb5,0xb5,0xee,0xb5,0xc1,0x5b,0x77,0x74);
+	&LL(0xaf,0xaf,0x86,0xaf,0x11,0x29,0x43,0xbe);
+	&LL(0x6a,0x6a,0xb5,0x6a,0x77,0xdf,0xd4,0x1d);
+	&LL(0x50,0x50,0x5d,0x50,0xba,0x0d,0xa0,0xea);
+	&LL(0x45,0x45,0x09,0x45,0x12,0x4c,0x8a,0x57);
+	&LL(0xf3,0xf3,0xeb,0xf3,0xcb,0x18,0xfb,0x38);
+	&LL(0x30,0x30,0xc0,0x30,0x9d,0xf0,0x60,0xad);
+	&LL(0xef,0xef,0x9b,0xef,0x2b,0x74,0xc3,0xc4);
+	&LL(0x3f,0x3f,0xfc,0x3f,0xe5,0xc3,0x7e,0xda);
+	&LL(0x55,0x55,0x49,0x55,0x92,0x1c,0xaa,0xc7);
+	&LL(0xa2,0xa2,0xb2,0xa2,0x79,0x10,0x59,0xdb);
+	&LL(0xea,0xea,0x8f,0xea,0x03,0x65,0xc9,0xe9);
+	&LL(0x65,0x65,0x89,0x65,0x0f,0xec,0xca,0x6a);
+	&LL(0xba,0xba,0xd2,0xba,0xb9,0x68,0x69,0x03);
+	&LL(0x2f,0x2f,0xbc,0x2f,0x65,0x93,0x5e,0x4a);
+	&LL(0xc0,0xc0,0x27,0xc0,0x4e,0xe7,0x9d,0x8e);
+	&LL(0xde,0xde,0x5f,0xde,0xbe,0x81,0xa1,0x60);
+	&LL(0x1c,0x1c,0x70,0x1c,0xe0,0x6c,0x38,0xfc);
+	&LL(0xfd,0xfd,0xd3,0xfd,0xbb,0x2e,0xe7,0x46);
+	&LL(0x4d,0x4d,0x29,0x4d,0x52,0x64,0x9a,0x1f);
+	&LL(0x92,0x92,0x72,0x92,0xe4,0xe0,0x39,0x76);
+	&LL(0x75,0x75,0xc9,0x75,0x8f,0xbc,0xea,0xfa);
+	&LL(0x06,0x06,0x18,0x06,0x30,0x1e,0x0c,0x36);
+	&LL(0x8a,0x8a,0x12,0x8a,0x24,0x98,0x09,0xae);
+	&LL(0xb2,0xb2,0xf2,0xb2,0xf9,0x40,0x79,0x4b);
+	&LL(0xe6,0xe6,0xbf,0xe6,0x63,0x59,0xd1,0x85);
+	&LL(0x0e,0x0e,0x38,0x0e,0x70,0x36,0x1c,0x7e);
+	&LL(0x1f,0x1f,0x7c,0x1f,0xf8,0x63,0x3e,0xe7);
+	&LL(0x62,0x62,0x95,0x62,0x37,0xf7,0xc4,0x55);
+	&LL(0xd4,0xd4,0x77,0xd4,0xee,0xa3,0xb5,0x3a);
+	&LL(0xa8,0xa8,0x9a,0xa8,0x29,0x32,0x4d,0x81);
+	&LL(0x96,0x96,0x62,0x96,0xc4,0xf4,0x31,0x52);
+	&LL(0xf9,0xf9,0xc3,0xf9,0x9b,0x3a,0xef,0x62);
+	&LL(0xc5,0xc5,0x33,0xc5,0x66,0xf6,0x97,0xa3);
+	&LL(0x25,0x25,0x94,0x25,0x35,0xb1,0x4a,0x10);
+	&LL(0x59,0x59,0x79,0x59,0xf2,0x20,0xb2,0xab);
+	&LL(0x84,0x84,0x2a,0x84,0x54,0xae,0x15,0xd0);
+	&LL(0x72,0x72,0xd5,0x72,0xb7,0xa7,0xe4,0xc5);
+	&LL(0x39,0x39,0xe4,0x39,0xd5,0xdd,0x72,0xec);
+	&LL(0x4c,0x4c,0x2d,0x4c,0x5a,0x61,0x98,0x16);
+	&LL(0x5e,0x5e,0x65,0x5e,0xca,0x3b,0xbc,0x94);
+	&LL(0x78,0x78,0xfd,0x78,0xe7,0x85,0xf0,0x9f);
+	&LL(0x38,0x38,0xe0,0x38,0xdd,0xd8,0x70,0xe5);
+	&LL(0x8c,0x8c,0x0a,0x8c,0x14,0x86,0x05,0x98);
+	&LL(0xd1,0xd1,0x63,0xd1,0xc6,0xb2,0xbf,0x17);
+	&LL(0xa5,0xa5,0xae,0xa5,0x41,0x0b,0x57,0xe4);
+	&LL(0xe2,0xe2,0xaf,0xe2,0x43,0x4d,0xd9,0xa1);
+	&LL(0x61,0x61,0x99,0x61,0x2f,0xf8,0xc2,0x4e);
+	&LL(0xb3,0xb3,0xf6,0xb3,0xf1,0x45,0x7b,0x42);
+	&LL(0x21,0x21,0x84,0x21,0x15,0xa5,0x42,0x34);
+	&LL(0x9c,0x9c,0x4a,0x9c,0x94,0xd6,0x25,0x08);
+	&LL(0x1e,0x1e,0x78,0x1e,0xf0,0x66,0x3c,0xee);
+	&LL(0x43,0x43,0x11,0x43,0x22,0x52,0x86,0x61);
+	&LL(0xc7,0xc7,0x3b,0xc7,0x76,0xfc,0x93,0xb1);
+	&LL(0xfc,0xfc,0xd7,0xfc,0xb3,0x2b,0xe5,0x4f);
+	&LL(0x04,0x04,0x10,0x04,0x20,0x14,0x08,0x24);
+	&LL(0x51,0x51,0x59,0x51,0xb2,0x08,0xa2,0xe3);
+	&LL(0x99,0x99,0x5e,0x99,0xbc,0xc7,0x2f,0x25);
+	&LL(0x6d,0x6d,0xa9,0x6d,0x4f,0xc4,0xda,0x22);
+	&LL(0x0d,0x0d,0x34,0x0d,0x68,0x39,0x1a,0x65);
+	&LL(0xfa,0xfa,0xcf,0xfa,0x83,0x35,0xe9,0x79);
+	&LL(0xdf,0xdf,0x5b,0xdf,0xb6,0x84,0xa3,0x69);
+	&LL(0x7e,0x7e,0xe5,0x7e,0xd7,0x9b,0xfc,0xa9);
+	&LL(0x24,0x24,0x90,0x24,0x3d,0xb4,0x48,0x19);
+	&LL(0x3b,0x3b,0xec,0x3b,0xc5,0xd7,0x76,0xfe);
+	&LL(0xab,0xab,0x96,0xab,0x31,0x3d,0x4b,0x9a);
+	&LL(0xce,0xce,0x1f,0xce,0x3e,0xd1,0x81,0xf0);
+	&LL(0x11,0x11,0x44,0x11,0x88,0x55,0x22,0x99);
+	&LL(0x8f,0x8f,0x06,0x8f,0x0c,0x89,0x03,0x83);
+	&LL(0x4e,0x4e,0x25,0x4e,0x4a,0x6b,0x9c,0x04);
+	&LL(0xb7,0xb7,0xe6,0xb7,0xd1,0x51,0x73,0x66);
+	&LL(0xeb,0xeb,0x8b,0xeb,0x0b,0x60,0xcb,0xe0);
+	&LL(0x3c,0x3c,0xf0,0x3c,0xfd,0xcc,0x78,0xc1);
+	&LL(0x81,0x81,0x3e,0x81,0x7c,0xbf,0x1f,0xfd);
+	&LL(0x94,0x94,0x6a,0x94,0xd4,0xfe,0x35,0x40);
+	&LL(0xf7,0xf7,0xfb,0xf7,0xeb,0x0c,0xf3,0x1c);
+	&LL(0xb9,0xb9,0xde,0xb9,0xa1,0x67,0x6f,0x18);
+	&LL(0x13,0x13,0x4c,0x13,0x98,0x5f,0x26,0x8b);
+	&LL(0x2c,0x2c,0xb0,0x2c,0x7d,0x9c,0x58,0x51);
+	&LL(0xd3,0xd3,0x6b,0xd3,0xd6,0xb8,0xbb,0x05);
+	&LL(0xe7,0xe7,0xbb,0xe7,0x6b,0x5c,0xd3,0x8c);
+	&LL(0x6e,0x6e,0xa5,0x6e,0x57,0xcb,0xdc,0x39);
+	&LL(0xc4,0xc4,0x37,0xc4,0x6e,0xf3,0x95,0xaa);
+	&LL(0x03,0x03,0x0c,0x03,0x18,0x0f,0x06,0x1b);
+	&LL(0x56,0x56,0x45,0x56,0x8a,0x13,0xac,0xdc);
+	&LL(0x44,0x44,0x0d,0x44,0x1a,0x49,0x88,0x5e);
+	&LL(0x7f,0x7f,0xe1,0x7f,0xdf,0x9e,0xfe,0xa0);
+	&LL(0xa9,0xa9,0x9e,0xa9,0x21,0x37,0x4f,0x88);
+	&LL(0x2a,0x2a,0xa8,0x2a,0x4d,0x82,0x54,0x67);
+	&LL(0xbb,0xbb,0xd6,0xbb,0xb1,0x6d,0x6b,0x0a);
+	&LL(0xc1,0xc1,0x23,0xc1,0x46,0xe2,0x9f,0x87);
+	&LL(0x53,0x53,0x51,0x53,0xa2,0x02,0xa6,0xf1);
+	&LL(0xdc,0xdc,0x57,0xdc,0xae,0x8b,0xa5,0x72);
+	&LL(0x0b,0x0b,0x2c,0x0b,0x58,0x27,0x16,0x53);
+	&LL(0x9d,0x9d,0x4e,0x9d,0x9c,0xd3,0x27,0x01);
+	&LL(0x6c,0x6c,0xad,0x6c,0x47,0xc1,0xd8,0x2b);
+	&LL(0x31,0x31,0xc4,0x31,0x95,0xf5,0x62,0xa4);
+	&LL(0x74,0x74,0xcd,0x74,0x87,0xb9,0xe8,0xf3);
+	&LL(0xf6,0xf6,0xff,0xf6,0xe3,0x09,0xf1,0x15);
+	&LL(0x46,0x46,0x05,0x46,0x0a,0x43,0x8c,0x4c);
+	&LL(0xac,0xac,0x8a,0xac,0x09,0x26,0x45,0xa5);
+	&LL(0x89,0x89,0x1e,0x89,0x3c,0x97,0x0f,0xb5);
+	&LL(0x14,0x14,0x50,0x14,0xa0,0x44,0x28,0xb4);
+	&LL(0xe1,0xe1,0xa3,0xe1,0x5b,0x42,0xdf,0xba);
+	&LL(0x16,0x16,0x58,0x16,0xb0,0x4e,0x2c,0xa6);
+	&LL(0x3a,0x3a,0xe8,0x3a,0xcd,0xd2,0x74,0xf7);
+	&LL(0x69,0x69,0xb9,0x69,0x6f,0xd0,0xd2,0x06);
+	&LL(0x09,0x09,0x24,0x09,0x48,0x2d,0x12,0x41);
+	&LL(0x70,0x70,0xdd,0x70,0xa7,0xad,0xe0,0xd7);
+	&LL(0xb6,0xb6,0xe2,0xb6,0xd9,0x54,0x71,0x6f);
+	&LL(0xd0,0xd0,0x67,0xd0,0xce,0xb7,0xbd,0x1e);
+	&LL(0xed,0xed,0x93,0xed,0x3b,0x7e,0xc7,0xd6);
+	&LL(0xcc,0xcc,0x17,0xcc,0x2e,0xdb,0x85,0xe2);
+	&LL(0x42,0x42,0x15,0x42,0x2a,0x57,0x84,0x68);
+	&LL(0x98,0x98,0x5a,0x98,0xb4,0xc2,0x2d,0x2c);
+	&LL(0xa4,0xa4,0xaa,0xa4,0x49,0x0e,0x55,0xed);
+	&LL(0x28,0x28,0xa0,0x28,0x5d,0x88,0x50,0x75);
+	&LL(0x5c,0x5c,0x6d,0x5c,0xda,0x31,0xb8,0x86);
+	&LL(0xf8,0xf8,0xc7,0xf8,0x93,0x3f,0xed,0x6b);
+	&LL(0x86,0x86,0x22,0x86,0x44,0xa4,0x11,0xc2);
+
+	&L(0x18,0x23,0xc6,0xe8,0x87,0xb8,0x01,0x4f);	# rc[ROUNDS]
+	&L(0x36,0xa6,0xd2,0xf5,0x79,0x6f,0x91,0x52);
+	&L(0x60,0xbc,0x9b,0x8e,0xa3,0x0c,0x7b,0x35);
+	&L(0x1d,0xe0,0xd7,0xc2,0x2e,0x4b,0xfe,0x57);
+	&L(0x15,0x77,0x37,0xe5,0x9f,0xf0,0x4a,0xda);
+	&L(0x58,0xc9,0x29,0x0a,0xb1,0xa0,0x6b,0x85);
+	&L(0xbd,0x5d,0x10,0xf4,0xcb,0x3e,0x05,0x67);
+	&L(0xe4,0x27,0x41,0x8b,0xa7,0x7d,0x95,0xd8);
+	&L(0xfb,0xee,0x7c,0x66,0xdd,0x17,0x47,0x9e);
+	&L(0xca,0x2d,0xbf,0x07,0xad,0x5a,0x83,0x33);
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lprologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lprologue
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lepilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lepilogue
+	jae	.Lin_prologue
+
+	mov	128+32(%rax),%rax	# pull saved stack pointer
+	lea	48(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	-40(%rax),%r14
+	mov	-48(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	se_handler,.-se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_$func
+	.rva	.LSEH_end_$func
+	.rva	.LSEH_info_$func
+
+.section	.xdata
+.align	8
+.LSEH_info_$func:
+	.byte	9,0,0,0
+	.rva	se_handler
+___
+}
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/crypto/whrlpool/whrlpool.h b/crypto/whrlpool/whrlpool.h
new file mode 100644
index 000000000000..9e01f5b07663
--- /dev/null
+++ b/crypto/whrlpool/whrlpool.h
@@ -0,0 +1,41 @@
+#ifndef HEADER_WHRLPOOL_H
+#define HEADER_WHRLPOOL_H
+
+#include <openssl/e_os2.h>
+#include <stddef.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#define WHIRLPOOL_DIGEST_LENGTH	(512/8)
+#define WHIRLPOOL_BBLOCK	512
+#define WHIRLPOOL_COUNTER	(256/8)
+
+typedef struct	{
+	union	{
+		unsigned char	c[WHIRLPOOL_DIGEST_LENGTH];
+		/* double q is here to ensure 64-bit alignment */
+		double		q[WHIRLPOOL_DIGEST_LENGTH/sizeof(double)];
+		}	H;
+	unsigned char	data[WHIRLPOOL_BBLOCK/8];
+	unsigned int	bitoff;
+	size_t		bitlen[WHIRLPOOL_COUNTER/sizeof(size_t)];
+	} WHIRLPOOL_CTX;
+
+#ifndef OPENSSL_NO_WHIRLPOOL
+#ifdef OPENSSL_FIPS
+int private_WHIRLPOOL_Init(WHIRLPOOL_CTX *c);
+#endif
+int WHIRLPOOL_Init	(WHIRLPOOL_CTX *c);
+int WHIRLPOOL_Update	(WHIRLPOOL_CTX *c,const void *inp,size_t bytes);
+void WHIRLPOOL_BitUpdate(WHIRLPOOL_CTX *c,const void *inp,size_t bits);
+int WHIRLPOOL_Final	(unsigned char *md,WHIRLPOOL_CTX *c);
+unsigned char *WHIRLPOOL(const void *inp,size_t bytes,unsigned char *md);
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif
diff --git a/crypto/whrlpool/wp_block.c b/crypto/whrlpool/wp_block.c
new file mode 100644
index 000000000000..824ed1827c48
--- /dev/null
+++ b/crypto/whrlpool/wp_block.c
@@ -0,0 +1,655 @@
+/**
+ * The Whirlpool hashing function.
+ *
+ * <P>
+ * <b>References</b>
+ *
+ * <P>
+ * The Whirlpool algorithm was developed by
+ * <a href="mailto:pbarreto@scopus.com.br">Paulo S. L. M. Barreto</a> and
+ * <a href="mailto:vincent.rijmen@cryptomathic.com">Vincent Rijmen</a>.
+ *
+ * See
+ *      P.S.L.M. Barreto, V. Rijmen,
+ *      ``The Whirlpool hashing function,''
+ *      NESSIE submission, 2000 (tweaked version, 2001),
+ *      <https://www.cosic.esat.kuleuven.ac.be/nessie/workshop/submissions/whirlpool.zip>
+ *
+ * Based on "@version 3.0 (2003.03.12)" by Paulo S.L.M. Barreto and
+ * Vincent Rijmen. Lookup "reference implementations" on
+ * <http://planeta.terra.com.br/informatica/paulobarreto/>
+ *
+ * =============================================================================
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ''AS IS'' AND ANY EXPRESS
+ * OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
+ * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+ * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+#include "wp_locl.h"
+#include <string.h>
+
+typedef unsigned char		u8;
+#if (defined(_WIN32) || defined(_WIN64)) && !defined(__MINGW32)
+typedef unsigned __int64	u64;
+#elif defined(__arch64__)
+typedef unsigned long		u64;
+#else
+typedef unsigned long long	u64;
+#endif
+
+#define ROUNDS	10
+
+#define STRICT_ALIGNMENT
+#if defined(__i386) || defined(__i386__) || \
+    defined(__x86_64) || defined(__x86_64__) || \
+    defined(_M_IX86) || defined(_M_AMD64) || defined(_M_X64)
+/* Well, formally there're couple of other architectures, which permit
+ * unaligned loads, specifically those not crossing cache lines, IA-64
+ * and PowerPC... */
+#  undef STRICT_ALIGNMENT
+#endif
+
+#undef SMALL_REGISTER_BANK
+#if defined(__i386) || defined(__i386__) || defined(_M_IX86)
+#  define SMALL_REGISTER_BANK
+#  if defined(WHIRLPOOL_ASM)
+#    ifndef OPENSSL_SMALL_FOOTPRINT
+#      define OPENSSL_SMALL_FOOTPRINT	/* it appears that for elder non-MMX
+					   CPUs this is actually faster! */
+#    endif
+#    define GO_FOR_MMX(ctx,inp,num)	do {			\
+	extern unsigned int OPENSSL_ia32cap_P[];		\
+	void whirlpool_block_mmx(void *,const void *,size_t);	\
+	if (!(OPENSSL_ia32cap_P[0] & (1<<23)))	break;		\
+        whirlpool_block_mmx(ctx->H.c,inp,num);	return;		\
+					} while (0)
+#  endif
+#endif
+
+#undef ROTATE
+#if defined(_MSC_VER)
+#  if defined(_WIN64)	/* applies to both IA-64 and AMD64 */
+#    pragma intrinsic(_rotl64)
+#    define ROTATE(a,n)	_rotl64((a),n)
+#  endif
+#elif defined(__GNUC__) && __GNUC__>=2
+#  if defined(__x86_64) || defined(__x86_64__)
+#    if defined(L_ENDIAN)
+#      define ROTATE(a,n)	({ u64 ret; asm ("rolq %1,%0"	\
+				   : "=r"(ret) : "J"(n),"0"(a) : "cc"); ret; })
+#    elif defined(B_ENDIAN)
+       /* Most will argue that x86_64 is always little-endian. Well,
+        * yes, but then we have stratus.com who has modified gcc to
+	* "emulate" big-endian on x86. Is there evidence that they
+	* [or somebody else] won't do same for x86_64? Naturally no.
+	* And this line is waiting ready for that brave soul:-) */
+#      define ROTATE(a,n)	({ u64 ret; asm ("rorq %1,%0"	\
+				   : "=r"(ret) : "J"(n),"0"(a) : "cc"); ret; })
+#    endif
+#  elif defined(__ia64) || defined(__ia64__)
+#    if defined(L_ENDIAN)
+#      define ROTATE(a,n)	({ u64 ret; asm ("shrp %0=%1,%1,%2"	\
+				   : "=r"(ret) : "r"(a),"M"(64-(n))); ret; })
+#    elif defined(B_ENDIAN)
+#      define ROTATE(a,n)	({ u64 ret; asm ("shrp %0=%1,%1,%2"	\
+				   : "=r"(ret) : "r"(a),"M"(n)); ret; })
+#    endif
+#  endif
+#endif
+
+#if defined(OPENSSL_SMALL_FOOTPRINT)
+#  if !defined(ROTATE)
+#    if defined(L_ENDIAN)	/* little-endians have to rotate left */
+#      define ROTATE(i,n)	((i)<<(n) ^ (i)>>(64-n))
+#    elif defined(B_ENDIAN)	/* big-endians have to rotate right */
+#      define ROTATE(i,n)	((i)>>(n) ^ (i)<<(64-n))
+#    endif
+#  endif
+#  if defined(ROTATE) && !defined(STRICT_ALIGNMENT)
+#    define STRICT_ALIGNMENT	/* ensure smallest table size */
+#  endif
+#endif
+
+/*
+ * Table size depends on STRICT_ALIGNMENT and whether or not endian-
+ * specific ROTATE macro is defined. If STRICT_ALIGNMENT is not
+ * defined, which is normally the case on x86[_64] CPUs, the table is
+ * 4KB large unconditionally. Otherwise if ROTATE is defined, the
+ * table is 2KB large, and otherwise - 16KB. 2KB table requires a
+ * whole bunch of additional rotations, but I'm willing to "trade,"
+ * because 16KB table certainly trashes L1 cache. I wish all CPUs
+ * could handle unaligned load as 4KB table doesn't trash the cache,
+ * nor does it require additional rotations.
+ */
+/*
+ * Note that every Cn macro expands as two loads: one byte load and
+ * one quadword load. One can argue that that many single-byte loads
+ * is too excessive, as one could load a quadword and "milk" it for
+ * eight 8-bit values instead. Well, yes, but in order to do so *and*
+ * avoid excessive loads you have to accomodate a handful of 64-bit
+ * values in the register bank and issue a bunch of shifts and mask.
+ * It's a tradeoff: loads vs. shift and mask in big register bank[!].
+ * On most CPUs eight single-byte loads are faster and I let other
+ * ones to depend on smart compiler to fold byte loads if beneficial.
+ * Hand-coded assembler would be another alternative:-)
+ */
+#ifdef STRICT_ALIGNMENT
+#  if defined(ROTATE)
+#    define N	1
+#    define LL(c0,c1,c2,c3,c4,c5,c6,c7)	c0,c1,c2,c3,c4,c5,c6,c7
+#    define C0(K,i)	(Cx.q[K.c[(i)*8+0]])
+#    define C1(K,i)	ROTATE(Cx.q[K.c[(i)*8+1]],8)
+#    define C2(K,i)	ROTATE(Cx.q[K.c[(i)*8+2]],16)
+#    define C3(K,i)	ROTATE(Cx.q[K.c[(i)*8+3]],24)
+#    define C4(K,i)	ROTATE(Cx.q[K.c[(i)*8+4]],32)
+#    define C5(K,i)	ROTATE(Cx.q[K.c[(i)*8+5]],40)
+#    define C6(K,i)	ROTATE(Cx.q[K.c[(i)*8+6]],48)
+#    define C7(K,i)	ROTATE(Cx.q[K.c[(i)*8+7]],56)
+#  else
+#    define N	8
+#    define LL(c0,c1,c2,c3,c4,c5,c6,c7)	c0,c1,c2,c3,c4,c5,c6,c7, \
+					c7,c0,c1,c2,c3,c4,c5,c6, \
+					c6,c7,c0,c1,c2,c3,c4,c5, \
+					c5,c6,c7,c0,c1,c2,c3,c4, \
+					c4,c5,c6,c7,c0,c1,c2,c3, \
+					c3,c4,c5,c6,c7,c0,c1,c2, \
+					c2,c3,c4,c5,c6,c7,c0,c1, \
+					c1,c2,c3,c4,c5,c6,c7,c0
+#    define C0(K,i)	(Cx.q[0+8*K.c[(i)*8+0]])
+#    define C1(K,i)	(Cx.q[1+8*K.c[(i)*8+1]])
+#    define C2(K,i)	(Cx.q[2+8*K.c[(i)*8+2]])
+#    define C3(K,i)	(Cx.q[3+8*K.c[(i)*8+3]])
+#    define C4(K,i)	(Cx.q[4+8*K.c[(i)*8+4]])
+#    define C5(K,i)	(Cx.q[5+8*K.c[(i)*8+5]])
+#    define C6(K,i)	(Cx.q[6+8*K.c[(i)*8+6]])
+#    define C7(K,i)	(Cx.q[7+8*K.c[(i)*8+7]])
+#  endif
+#else
+#  define N	2
+#  define LL(c0,c1,c2,c3,c4,c5,c6,c7)	c0,c1,c2,c3,c4,c5,c6,c7, \
+					c0,c1,c2,c3,c4,c5,c6,c7
+#  define C0(K,i)	(((u64*)(Cx.c+0))[2*K.c[(i)*8+0]])
+#  define C1(K,i)	(((u64*)(Cx.c+7))[2*K.c[(i)*8+1]])
+#  define C2(K,i)	(((u64*)(Cx.c+6))[2*K.c[(i)*8+2]])
+#  define C3(K,i)	(((u64*)(Cx.c+5))[2*K.c[(i)*8+3]])
+#  define C4(K,i)	(((u64*)(Cx.c+4))[2*K.c[(i)*8+4]])
+#  define C5(K,i)	(((u64*)(Cx.c+3))[2*K.c[(i)*8+5]])
+#  define C6(K,i)	(((u64*)(Cx.c+2))[2*K.c[(i)*8+6]])
+#  define C7(K,i)	(((u64*)(Cx.c+1))[2*K.c[(i)*8+7]])
+#endif
+
+static const
+union	{
+	u8	c[(256*N+ROUNDS)*sizeof(u64)];
+	u64	q[(256*N+ROUNDS)];
+	} Cx = { {
+	/* Note endian-neutral representation:-) */
+	LL(0x18,0x18,0x60,0x18,0xc0,0x78,0x30,0xd8),
+	LL(0x23,0x23,0x8c,0x23,0x05,0xaf,0x46,0x26),
+	LL(0xc6,0xc6,0x3f,0xc6,0x7e,0xf9,0x91,0xb8),
+	LL(0xe8,0xe8,0x87,0xe8,0x13,0x6f,0xcd,0xfb),
+	LL(0x87,0x87,0x26,0x87,0x4c,0xa1,0x13,0xcb),
+	LL(0xb8,0xb8,0xda,0xb8,0xa9,0x62,0x6d,0x11),
+	LL(0x01,0x01,0x04,0x01,0x08,0x05,0x02,0x09),
+	LL(0x4f,0x4f,0x21,0x4f,0x42,0x6e,0x9e,0x0d),
+	LL(0x36,0x36,0xd8,0x36,0xad,0xee,0x6c,0x9b),
+	LL(0xa6,0xa6,0xa2,0xa6,0x59,0x04,0x51,0xff),
+	LL(0xd2,0xd2,0x6f,0xd2,0xde,0xbd,0xb9,0x0c),
+	LL(0xf5,0xf5,0xf3,0xf5,0xfb,0x06,0xf7,0x0e),
+	LL(0x79,0x79,0xf9,0x79,0xef,0x80,0xf2,0x96),
+	LL(0x6f,0x6f,0xa1,0x6f,0x5f,0xce,0xde,0x30),
+	LL(0x91,0x91,0x7e,0x91,0xfc,0xef,0x3f,0x6d),
+	LL(0x52,0x52,0x55,0x52,0xaa,0x07,0xa4,0xf8),
+	LL(0x60,0x60,0x9d,0x60,0x27,0xfd,0xc0,0x47),
+	LL(0xbc,0xbc,0xca,0xbc,0x89,0x76,0x65,0x35),
+	LL(0x9b,0x9b,0x56,0x9b,0xac,0xcd,0x2b,0x37),
+	LL(0x8e,0x8e,0x02,0x8e,0x04,0x8c,0x01,0x8a),
+	LL(0xa3,0xa3,0xb6,0xa3,0x71,0x15,0x5b,0xd2),
+	LL(0x0c,0x0c,0x30,0x0c,0x60,0x3c,0x18,0x6c),
+	LL(0x7b,0x7b,0xf1,0x7b,0xff,0x8a,0xf6,0x84),
+	LL(0x35,0x35,0xd4,0x35,0xb5,0xe1,0x6a,0x80),
+	LL(0x1d,0x1d,0x74,0x1d,0xe8,0x69,0x3a,0xf5),
+	LL(0xe0,0xe0,0xa7,0xe0,0x53,0x47,0xdd,0xb3),
+	LL(0xd7,0xd7,0x7b,0xd7,0xf6,0xac,0xb3,0x21),
+	LL(0xc2,0xc2,0x2f,0xc2,0x5e,0xed,0x99,0x9c),
+	LL(0x2e,0x2e,0xb8,0x2e,0x6d,0x96,0x5c,0x43),
+	LL(0x4b,0x4b,0x31,0x4b,0x62,0x7a,0x96,0x29),
+	LL(0xfe,0xfe,0xdf,0xfe,0xa3,0x21,0xe1,0x5d),
+	LL(0x57,0x57,0x41,0x57,0x82,0x16,0xae,0xd5),
+	LL(0x15,0x15,0x54,0x15,0xa8,0x41,0x2a,0xbd),
+	LL(0x77,0x77,0xc1,0x77,0x9f,0xb6,0xee,0xe8),
+	LL(0x37,0x37,0xdc,0x37,0xa5,0xeb,0x6e,0x92),
+	LL(0xe5,0xe5,0xb3,0xe5,0x7b,0x56,0xd7,0x9e),
+	LL(0x9f,0x9f,0x46,0x9f,0x8c,0xd9,0x23,0x13),
+	LL(0xf0,0xf0,0xe7,0xf0,0xd3,0x17,0xfd,0x23),
+	LL(0x4a,0x4a,0x35,0x4a,0x6a,0x7f,0x94,0x20),
+	LL(0xda,0xda,0x4f,0xda,0x9e,0x95,0xa9,0x44),
+	LL(0x58,0x58,0x7d,0x58,0xfa,0x25,0xb0,0xa2),
+	LL(0xc9,0xc9,0x03,0xc9,0x06,0xca,0x8f,0xcf),
+	LL(0x29,0x29,0xa4,0x29,0x55,0x8d,0x52,0x7c),
+	LL(0x0a,0x0a,0x28,0x0a,0x50,0x22,0x14,0x5a),
+	LL(0xb1,0xb1,0xfe,0xb1,0xe1,0x4f,0x7f,0x50),
+	LL(0xa0,0xa0,0xba,0xa0,0x69,0x1a,0x5d,0xc9),
+	LL(0x6b,0x6b,0xb1,0x6b,0x7f,0xda,0xd6,0x14),
+	LL(0x85,0x85,0x2e,0x85,0x5c,0xab,0x17,0xd9),
+	LL(0xbd,0xbd,0xce,0xbd,0x81,0x73,0x67,0x3c),
+	LL(0x5d,0x5d,0x69,0x5d,0xd2,0x34,0xba,0x8f),
+	LL(0x10,0x10,0x40,0x10,0x80,0x50,0x20,0x90),
+	LL(0xf4,0xf4,0xf7,0xf4,0xf3,0x03,0xf5,0x07),
+	LL(0xcb,0xcb,0x0b,0xcb,0x16,0xc0,0x8b,0xdd),
+	LL(0x3e,0x3e,0xf8,0x3e,0xed,0xc6,0x7c,0xd3),
+	LL(0x05,0x05,0x14,0x05,0x28,0x11,0x0a,0x2d),
+	LL(0x67,0x67,0x81,0x67,0x1f,0xe6,0xce,0x78),
+	LL(0xe4,0xe4,0xb7,0xe4,0x73,0x53,0xd5,0x97),
+	LL(0x27,0x27,0x9c,0x27,0x25,0xbb,0x4e,0x02),
+	LL(0x41,0x41,0x19,0x41,0x32,0x58,0x82,0x73),
+	LL(0x8b,0x8b,0x16,0x8b,0x2c,0x9d,0x0b,0xa7),
+	LL(0xa7,0xa7,0xa6,0xa7,0x51,0x01,0x53,0xf6),
+	LL(0x7d,0x7d,0xe9,0x7d,0xcf,0x94,0xfa,0xb2),
+	LL(0x95,0x95,0x6e,0x95,0xdc,0xfb,0x37,0x49),
+	LL(0xd8,0xd8,0x47,0xd8,0x8e,0x9f,0xad,0x56),
+	LL(0xfb,0xfb,0xcb,0xfb,0x8b,0x30,0xeb,0x70),
+	LL(0xee,0xee,0x9f,0xee,0x23,0x71,0xc1,0xcd),
+	LL(0x7c,0x7c,0xed,0x7c,0xc7,0x91,0xf8,0xbb),
+	LL(0x66,0x66,0x85,0x66,0x17,0xe3,0xcc,0x71),
+	LL(0xdd,0xdd,0x53,0xdd,0xa6,0x8e,0xa7,0x7b),
+	LL(0x17,0x17,0x5c,0x17,0xb8,0x4b,0x2e,0xaf),
+	LL(0x47,0x47,0x01,0x47,0x02,0x46,0x8e,0x45),
+	LL(0x9e,0x9e,0x42,0x9e,0x84,0xdc,0x21,0x1a),
+	LL(0xca,0xca,0x0f,0xca,0x1e,0xc5,0x89,0xd4),
+	LL(0x2d,0x2d,0xb4,0x2d,0x75,0x99,0x5a,0x58),
+	LL(0xbf,0xbf,0xc6,0xbf,0x91,0x79,0x63,0x2e),
+	LL(0x07,0x07,0x1c,0x07,0x38,0x1b,0x0e,0x3f),
+	LL(0xad,0xad,0x8e,0xad,0x01,0x23,0x47,0xac),
+	LL(0x5a,0x5a,0x75,0x5a,0xea,0x2f,0xb4,0xb0),
+	LL(0x83,0x83,0x36,0x83,0x6c,0xb5,0x1b,0xef),
+	LL(0x33,0x33,0xcc,0x33,0x85,0xff,0x66,0xb6),
+	LL(0x63,0x63,0x91,0x63,0x3f,0xf2,0xc6,0x5c),
+	LL(0x02,0x02,0x08,0x02,0x10,0x0a,0x04,0x12),
+	LL(0xaa,0xaa,0x92,0xaa,0x39,0x38,0x49,0x93),
+	LL(0x71,0x71,0xd9,0x71,0xaf,0xa8,0xe2,0xde),
+	LL(0xc8,0xc8,0x07,0xc8,0x0e,0xcf,0x8d,0xc6),
+	LL(0x19,0x19,0x64,0x19,0xc8,0x7d,0x32,0xd1),
+	LL(0x49,0x49,0x39,0x49,0x72,0x70,0x92,0x3b),
+	LL(0xd9,0xd9,0x43,0xd9,0x86,0x9a,0xaf,0x5f),
+	LL(0xf2,0xf2,0xef,0xf2,0xc3,0x1d,0xf9,0x31),
+	LL(0xe3,0xe3,0xab,0xe3,0x4b,0x48,0xdb,0xa8),
+	LL(0x5b,0x5b,0x71,0x5b,0xe2,0x2a,0xb6,0xb9),
+	LL(0x88,0x88,0x1a,0x88,0x34,0x92,0x0d,0xbc),
+	LL(0x9a,0x9a,0x52,0x9a,0xa4,0xc8,0x29,0x3e),
+	LL(0x26,0x26,0x98,0x26,0x2d,0xbe,0x4c,0x0b),
+	LL(0x32,0x32,0xc8,0x32,0x8d,0xfa,0x64,0xbf),
+	LL(0xb0,0xb0,0xfa,0xb0,0xe9,0x4a,0x7d,0x59),
+	LL(0xe9,0xe9,0x83,0xe9,0x1b,0x6a,0xcf,0xf2),
+	LL(0x0f,0x0f,0x3c,0x0f,0x78,0x33,0x1e,0x77),
+	LL(0xd5,0xd5,0x73,0xd5,0xe6,0xa6,0xb7,0x33),
+	LL(0x80,0x80,0x3a,0x80,0x74,0xba,0x1d,0xf4),
+	LL(0xbe,0xbe,0xc2,0xbe,0x99,0x7c,0x61,0x27),
+	LL(0xcd,0xcd,0x13,0xcd,0x26,0xde,0x87,0xeb),
+	LL(0x34,0x34,0xd0,0x34,0xbd,0xe4,0x68,0x89),
+	LL(0x48,0x48,0x3d,0x48,0x7a,0x75,0x90,0x32),
+	LL(0xff,0xff,0xdb,0xff,0xab,0x24,0xe3,0x54),
+	LL(0x7a,0x7a,0xf5,0x7a,0xf7,0x8f,0xf4,0x8d),
+	LL(0x90,0x90,0x7a,0x90,0xf4,0xea,0x3d,0x64),
+	LL(0x5f,0x5f,0x61,0x5f,0xc2,0x3e,0xbe,0x9d),
+	LL(0x20,0x20,0x80,0x20,0x1d,0xa0,0x40,0x3d),
+	LL(0x68,0x68,0xbd,0x68,0x67,0xd5,0xd0,0x0f),
+	LL(0x1a,0x1a,0x68,0x1a,0xd0,0x72,0x34,0xca),
+	LL(0xae,0xae,0x82,0xae,0x19,0x2c,0x41,0xb7),
+	LL(0xb4,0xb4,0xea,0xb4,0xc9,0x5e,0x75,0x7d),
+	LL(0x54,0x54,0x4d,0x54,0x9a,0x19,0xa8,0xce),
+	LL(0x93,0x93,0x76,0x93,0xec,0xe5,0x3b,0x7f),
+	LL(0x22,0x22,0x88,0x22,0x0d,0xaa,0x44,0x2f),
+	LL(0x64,0x64,0x8d,0x64,0x07,0xe9,0xc8,0x63),
+	LL(0xf1,0xf1,0xe3,0xf1,0xdb,0x12,0xff,0x2a),
+	LL(0x73,0x73,0xd1,0x73,0xbf,0xa2,0xe6,0xcc),
+	LL(0x12,0x12,0x48,0x12,0x90,0x5a,0x24,0x82),
+	LL(0x40,0x40,0x1d,0x40,0x3a,0x5d,0x80,0x7a),
+	LL(0x08,0x08,0x20,0x08,0x40,0x28,0x10,0x48),
+	LL(0xc3,0xc3,0x2b,0xc3,0x56,0xe8,0x9b,0x95),
+	LL(0xec,0xec,0x97,0xec,0x33,0x7b,0xc5,0xdf),
+	LL(0xdb,0xdb,0x4b,0xdb,0x96,0x90,0xab,0x4d),
+	LL(0xa1,0xa1,0xbe,0xa1,0x61,0x1f,0x5f,0xc0),
+	LL(0x8d,0x8d,0x0e,0x8d,0x1c,0x83,0x07,0x91),
+	LL(0x3d,0x3d,0xf4,0x3d,0xf5,0xc9,0x7a,0xc8),
+	LL(0x97,0x97,0x66,0x97,0xcc,0xf1,0x33,0x5b),
+	LL(0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00),
+	LL(0xcf,0xcf,0x1b,0xcf,0x36,0xd4,0x83,0xf9),
+	LL(0x2b,0x2b,0xac,0x2b,0x45,0x87,0x56,0x6e),
+	LL(0x76,0x76,0xc5,0x76,0x97,0xb3,0xec,0xe1),
+	LL(0x82,0x82,0x32,0x82,0x64,0xb0,0x19,0xe6),
+	LL(0xd6,0xd6,0x7f,0xd6,0xfe,0xa9,0xb1,0x28),
+	LL(0x1b,0x1b,0x6c,0x1b,0xd8,0x77,0x36,0xc3),
+	LL(0xb5,0xb5,0xee,0xb5,0xc1,0x5b,0x77,0x74),
+	LL(0xaf,0xaf,0x86,0xaf,0x11,0x29,0x43,0xbe),
+	LL(0x6a,0x6a,0xb5,0x6a,0x77,0xdf,0xd4,0x1d),
+	LL(0x50,0x50,0x5d,0x50,0xba,0x0d,0xa0,0xea),
+	LL(0x45,0x45,0x09,0x45,0x12,0x4c,0x8a,0x57),
+	LL(0xf3,0xf3,0xeb,0xf3,0xcb,0x18,0xfb,0x38),
+	LL(0x30,0x30,0xc0,0x30,0x9d,0xf0,0x60,0xad),
+	LL(0xef,0xef,0x9b,0xef,0x2b,0x74,0xc3,0xc4),
+	LL(0x3f,0x3f,0xfc,0x3f,0xe5,0xc3,0x7e,0xda),
+	LL(0x55,0x55,0x49,0x55,0x92,0x1c,0xaa,0xc7),
+	LL(0xa2,0xa2,0xb2,0xa2,0x79,0x10,0x59,0xdb),
+	LL(0xea,0xea,0x8f,0xea,0x03,0x65,0xc9,0xe9),
+	LL(0x65,0x65,0x89,0x65,0x0f,0xec,0xca,0x6a),
+	LL(0xba,0xba,0xd2,0xba,0xb9,0x68,0x69,0x03),
+	LL(0x2f,0x2f,0xbc,0x2f,0x65,0x93,0x5e,0x4a),
+	LL(0xc0,0xc0,0x27,0xc0,0x4e,0xe7,0x9d,0x8e),
+	LL(0xde,0xde,0x5f,0xde,0xbe,0x81,0xa1,0x60),
+	LL(0x1c,0x1c,0x70,0x1c,0xe0,0x6c,0x38,0xfc),
+	LL(0xfd,0xfd,0xd3,0xfd,0xbb,0x2e,0xe7,0x46),
+	LL(0x4d,0x4d,0x29,0x4d,0x52,0x64,0x9a,0x1f),
+	LL(0x92,0x92,0x72,0x92,0xe4,0xe0,0x39,0x76),
+	LL(0x75,0x75,0xc9,0x75,0x8f,0xbc,0xea,0xfa),
+	LL(0x06,0x06,0x18,0x06,0x30,0x1e,0x0c,0x36),
+	LL(0x8a,0x8a,0x12,0x8a,0x24,0x98,0x09,0xae),
+	LL(0xb2,0xb2,0xf2,0xb2,0xf9,0x40,0x79,0x4b),
+	LL(0xe6,0xe6,0xbf,0xe6,0x63,0x59,0xd1,0x85),
+	LL(0x0e,0x0e,0x38,0x0e,0x70,0x36,0x1c,0x7e),
+	LL(0x1f,0x1f,0x7c,0x1f,0xf8,0x63,0x3e,0xe7),
+	LL(0x62,0x62,0x95,0x62,0x37,0xf7,0xc4,0x55),
+	LL(0xd4,0xd4,0x77,0xd4,0xee,0xa3,0xb5,0x3a),
+	LL(0xa8,0xa8,0x9a,0xa8,0x29,0x32,0x4d,0x81),
+	LL(0x96,0x96,0x62,0x96,0xc4,0xf4,0x31,0x52),
+	LL(0xf9,0xf9,0xc3,0xf9,0x9b,0x3a,0xef,0x62),
+	LL(0xc5,0xc5,0x33,0xc5,0x66,0xf6,0x97,0xa3),
+	LL(0x25,0x25,0x94,0x25,0x35,0xb1,0x4a,0x10),
+	LL(0x59,0x59,0x79,0x59,0xf2,0x20,0xb2,0xab),
+	LL(0x84,0x84,0x2a,0x84,0x54,0xae,0x15,0xd0),
+	LL(0x72,0x72,0xd5,0x72,0xb7,0xa7,0xe4,0xc5),
+	LL(0x39,0x39,0xe4,0x39,0xd5,0xdd,0x72,0xec),
+	LL(0x4c,0x4c,0x2d,0x4c,0x5a,0x61,0x98,0x16),
+	LL(0x5e,0x5e,0x65,0x5e,0xca,0x3b,0xbc,0x94),
+	LL(0x78,0x78,0xfd,0x78,0xe7,0x85,0xf0,0x9f),
+	LL(0x38,0x38,0xe0,0x38,0xdd,0xd8,0x70,0xe5),
+	LL(0x8c,0x8c,0x0a,0x8c,0x14,0x86,0x05,0x98),
+	LL(0xd1,0xd1,0x63,0xd1,0xc6,0xb2,0xbf,0x17),
+	LL(0xa5,0xa5,0xae,0xa5,0x41,0x0b,0x57,0xe4),
+	LL(0xe2,0xe2,0xaf,0xe2,0x43,0x4d,0xd9,0xa1),
+	LL(0x61,0x61,0x99,0x61,0x2f,0xf8,0xc2,0x4e),
+	LL(0xb3,0xb3,0xf6,0xb3,0xf1,0x45,0x7b,0x42),
+	LL(0x21,0x21,0x84,0x21,0x15,0xa5,0x42,0x34),
+	LL(0x9c,0x9c,0x4a,0x9c,0x94,0xd6,0x25,0x08),
+	LL(0x1e,0x1e,0x78,0x1e,0xf0,0x66,0x3c,0xee),
+	LL(0x43,0x43,0x11,0x43,0x22,0x52,0x86,0x61),
+	LL(0xc7,0xc7,0x3b,0xc7,0x76,0xfc,0x93,0xb1),
+	LL(0xfc,0xfc,0xd7,0xfc,0xb3,0x2b,0xe5,0x4f),
+	LL(0x04,0x04,0x10,0x04,0x20,0x14,0x08,0x24),
+	LL(0x51,0x51,0x59,0x51,0xb2,0x08,0xa2,0xe3),
+	LL(0x99,0x99,0x5e,0x99,0xbc,0xc7,0x2f,0x25),
+	LL(0x6d,0x6d,0xa9,0x6d,0x4f,0xc4,0xda,0x22),
+	LL(0x0d,0x0d,0x34,0x0d,0x68,0x39,0x1a,0x65),
+	LL(0xfa,0xfa,0xcf,0xfa,0x83,0x35,0xe9,0x79),
+	LL(0xdf,0xdf,0x5b,0xdf,0xb6,0x84,0xa3,0x69),
+	LL(0x7e,0x7e,0xe5,0x7e,0xd7,0x9b,0xfc,0xa9),
+	LL(0x24,0x24,0x90,0x24,0x3d,0xb4,0x48,0x19),
+	LL(0x3b,0x3b,0xec,0x3b,0xc5,0xd7,0x76,0xfe),
+	LL(0xab,0xab,0x96,0xab,0x31,0x3d,0x4b,0x9a),
+	LL(0xce,0xce,0x1f,0xce,0x3e,0xd1,0x81,0xf0),
+	LL(0x11,0x11,0x44,0x11,0x88,0x55,0x22,0x99),
+	LL(0x8f,0x8f,0x06,0x8f,0x0c,0x89,0x03,0x83),
+	LL(0x4e,0x4e,0x25,0x4e,0x4a,0x6b,0x9c,0x04),
+	LL(0xb7,0xb7,0xe6,0xb7,0xd1,0x51,0x73,0x66),
+	LL(0xeb,0xeb,0x8b,0xeb,0x0b,0x60,0xcb,0xe0),
+	LL(0x3c,0x3c,0xf0,0x3c,0xfd,0xcc,0x78,0xc1),
+	LL(0x81,0x81,0x3e,0x81,0x7c,0xbf,0x1f,0xfd),
+	LL(0x94,0x94,0x6a,0x94,0xd4,0xfe,0x35,0x40),
+	LL(0xf7,0xf7,0xfb,0xf7,0xeb,0x0c,0xf3,0x1c),
+	LL(0xb9,0xb9,0xde,0xb9,0xa1,0x67,0x6f,0x18),
+	LL(0x13,0x13,0x4c,0x13,0x98,0x5f,0x26,0x8b),
+	LL(0x2c,0x2c,0xb0,0x2c,0x7d,0x9c,0x58,0x51),
+	LL(0xd3,0xd3,0x6b,0xd3,0xd6,0xb8,0xbb,0x05),
+	LL(0xe7,0xe7,0xbb,0xe7,0x6b,0x5c,0xd3,0x8c),
+	LL(0x6e,0x6e,0xa5,0x6e,0x57,0xcb,0xdc,0x39),
+	LL(0xc4,0xc4,0x37,0xc4,0x6e,0xf3,0x95,0xaa),
+	LL(0x03,0x03,0x0c,0x03,0x18,0x0f,0x06,0x1b),
+	LL(0x56,0x56,0x45,0x56,0x8a,0x13,0xac,0xdc),
+	LL(0x44,0x44,0x0d,0x44,0x1a,0x49,0x88,0x5e),
+	LL(0x7f,0x7f,0xe1,0x7f,0xdf,0x9e,0xfe,0xa0),
+	LL(0xa9,0xa9,0x9e,0xa9,0x21,0x37,0x4f,0x88),
+	LL(0x2a,0x2a,0xa8,0x2a,0x4d,0x82,0x54,0x67),
+	LL(0xbb,0xbb,0xd6,0xbb,0xb1,0x6d,0x6b,0x0a),
+	LL(0xc1,0xc1,0x23,0xc1,0x46,0xe2,0x9f,0x87),
+	LL(0x53,0x53,0x51,0x53,0xa2,0x02,0xa6,0xf1),
+	LL(0xdc,0xdc,0x57,0xdc,0xae,0x8b,0xa5,0x72),
+	LL(0x0b,0x0b,0x2c,0x0b,0x58,0x27,0x16,0x53),
+	LL(0x9d,0x9d,0x4e,0x9d,0x9c,0xd3,0x27,0x01),
+	LL(0x6c,0x6c,0xad,0x6c,0x47,0xc1,0xd8,0x2b),
+	LL(0x31,0x31,0xc4,0x31,0x95,0xf5,0x62,0xa4),
+	LL(0x74,0x74,0xcd,0x74,0x87,0xb9,0xe8,0xf3),
+	LL(0xf6,0xf6,0xff,0xf6,0xe3,0x09,0xf1,0x15),
+	LL(0x46,0x46,0x05,0x46,0x0a,0x43,0x8c,0x4c),
+	LL(0xac,0xac,0x8a,0xac,0x09,0x26,0x45,0xa5),
+	LL(0x89,0x89,0x1e,0x89,0x3c,0x97,0x0f,0xb5),
+	LL(0x14,0x14,0x50,0x14,0xa0,0x44,0x28,0xb4),
+	LL(0xe1,0xe1,0xa3,0xe1,0x5b,0x42,0xdf,0xba),
+	LL(0x16,0x16,0x58,0x16,0xb0,0x4e,0x2c,0xa6),
+	LL(0x3a,0x3a,0xe8,0x3a,0xcd,0xd2,0x74,0xf7),
+	LL(0x69,0x69,0xb9,0x69,0x6f,0xd0,0xd2,0x06),
+	LL(0x09,0x09,0x24,0x09,0x48,0x2d,0x12,0x41),
+	LL(0x70,0x70,0xdd,0x70,0xa7,0xad,0xe0,0xd7),
+	LL(0xb6,0xb6,0xe2,0xb6,0xd9,0x54,0x71,0x6f),
+	LL(0xd0,0xd0,0x67,0xd0,0xce,0xb7,0xbd,0x1e),
+	LL(0xed,0xed,0x93,0xed,0x3b,0x7e,0xc7,0xd6),
+	LL(0xcc,0xcc,0x17,0xcc,0x2e,0xdb,0x85,0xe2),
+	LL(0x42,0x42,0x15,0x42,0x2a,0x57,0x84,0x68),
+	LL(0x98,0x98,0x5a,0x98,0xb4,0xc2,0x2d,0x2c),
+	LL(0xa4,0xa4,0xaa,0xa4,0x49,0x0e,0x55,0xed),
+	LL(0x28,0x28,0xa0,0x28,0x5d,0x88,0x50,0x75),
+	LL(0x5c,0x5c,0x6d,0x5c,0xda,0x31,0xb8,0x86),
+	LL(0xf8,0xf8,0xc7,0xf8,0x93,0x3f,0xed,0x6b),
+	LL(0x86,0x86,0x22,0x86,0x44,0xa4,0x11,0xc2),
+#define RC	(&(Cx.q[256*N]))
+	0x18,0x23,0xc6,0xe8,0x87,0xb8,0x01,0x4f,	/* rc[ROUNDS] */
+	0x36,0xa6,0xd2,0xf5,0x79,0x6f,0x91,0x52,
+	0x60,0xbc,0x9b,0x8e,0xa3,0x0c,0x7b,0x35,
+	0x1d,0xe0,0xd7,0xc2,0x2e,0x4b,0xfe,0x57,
+	0x15,0x77,0x37,0xe5,0x9f,0xf0,0x4a,0xda,
+	0x58,0xc9,0x29,0x0a,0xb1,0xa0,0x6b,0x85,
+	0xbd,0x5d,0x10,0xf4,0xcb,0x3e,0x05,0x67,
+	0xe4,0x27,0x41,0x8b,0xa7,0x7d,0x95,0xd8,
+	0xfb,0xee,0x7c,0x66,0xdd,0x17,0x47,0x9e,
+	0xca,0x2d,0xbf,0x07,0xad,0x5a,0x83,0x33
+	}
+};
+
+void whirlpool_block(WHIRLPOOL_CTX *ctx,const void *inp,size_t n)
+	{
+	int	r;
+	const u8 *p=inp;
+	union	{ u64 q[8]; u8 c[64]; } S,K,*H=(void *)ctx->H.q;
+
+#ifdef GO_FOR_MMX
+	GO_FOR_MMX(ctx,inp,n);
+#endif
+							do {
+#ifdef OPENSSL_SMALL_FOOTPRINT
+	u64	L[8];
+	int	i;
+
+	for (i=0;i<64;i++)	S.c[i] = (K.c[i] = H->c[i]) ^ p[i];
+	for (r=0;r<ROUNDS;r++)
+		{
+		for (i=0;i<8;i++)
+			{
+			L[i]  = i ? 0 : RC[r];
+			L[i] ^=	C0(K,i)       ^ C1(K,(i-1)&7) ^
+				C2(K,(i-2)&7) ^ C3(K,(i-3)&7) ^
+				C4(K,(i-4)&7) ^ C5(K,(i-5)&7) ^
+				C6(K,(i-6)&7) ^ C7(K,(i-7)&7);
+			}
+		memcpy (K.q,L,64);
+		for (i=0;i<8;i++)
+			{
+			L[i] ^= C0(S,i)       ^ C1(S,(i-1)&7) ^
+				C2(S,(i-2)&7) ^ C3(S,(i-3)&7) ^
+				C4(S,(i-4)&7) ^ C5(S,(i-5)&7) ^
+				C6(S,(i-6)&7) ^ C7(S,(i-7)&7);
+			}
+		memcpy (S.q,L,64);
+		}
+	for (i=0;i<64;i++)	H->c[i] ^= S.c[i] ^ p[i];
+#else
+	u64	L0,L1,L2,L3,L4,L5,L6,L7;
+
+#ifdef STRICT_ALIGNMENT
+	if ((size_t)p & 7)
+		{
+		memcpy (S.c,p,64);
+		S.q[0] ^= (K.q[0] = H->q[0]);
+		S.q[1] ^= (K.q[1] = H->q[1]);
+		S.q[2] ^= (K.q[2] = H->q[2]);
+		S.q[3] ^= (K.q[3] = H->q[3]);
+		S.q[4] ^= (K.q[4] = H->q[4]);
+		S.q[5] ^= (K.q[5] = H->q[5]);
+		S.q[6] ^= (K.q[6] = H->q[6]);
+		S.q[7] ^= (K.q[7] = H->q[7]);
+		}
+	else
+#endif
+		{
+		const u64 *pa = (const u64*)p;
+		S.q[0] = (K.q[0] = H->q[0]) ^ pa[0];
+		S.q[1] = (K.q[1] = H->q[1]) ^ pa[1];
+		S.q[2] = (K.q[2] = H->q[2]) ^ pa[2];
+		S.q[3] = (K.q[3] = H->q[3]) ^ pa[3];
+		S.q[4] = (K.q[4] = H->q[4]) ^ pa[4];
+		S.q[5] = (K.q[5] = H->q[5]) ^ pa[5];
+		S.q[6] = (K.q[6] = H->q[6]) ^ pa[6];
+		S.q[7] = (K.q[7] = H->q[7]) ^ pa[7];
+		}
+
+	for(r=0;r<ROUNDS;r++)
+		{
+#ifdef SMALL_REGISTER_BANK
+		L0 =	C0(K,0) ^ C1(K,7) ^ C2(K,6) ^ C3(K,5) ^
+			C4(K,4) ^ C5(K,3) ^ C6(K,2) ^ C7(K,1) ^ RC[r];
+		L1 =	C0(K,1) ^ C1(K,0) ^ C2(K,7) ^ C3(K,6) ^
+			C4(K,5) ^ C5(K,4) ^ C6(K,3) ^ C7(K,2);
+		L2 =	C0(K,2) ^ C1(K,1) ^ C2(K,0) ^ C3(K,7) ^
+			C4(K,6) ^ C5(K,5) ^ C6(K,4) ^ C7(K,3);
+		L3 =	C0(K,3) ^ C1(K,2) ^ C2(K,1) ^ C3(K,0) ^
+			C4(K,7) ^ C5(K,6) ^ C6(K,5) ^ C7(K,4);
+		L4 =	C0(K,4) ^ C1(K,3) ^ C2(K,2) ^ C3(K,1) ^
+			C4(K,0) ^ C5(K,7) ^ C6(K,6) ^ C7(K,5);
+		L5 =	C0(K,5) ^ C1(K,4) ^ C2(K,3) ^ C3(K,2) ^
+			C4(K,1) ^ C5(K,0) ^ C6(K,7) ^ C7(K,6);
+		L6 =	C0(K,6) ^ C1(K,5) ^ C2(K,4) ^ C3(K,3) ^
+			C4(K,2) ^ C5(K,1) ^ C6(K,0) ^ C7(K,7);
+		L7 =	C0(K,7) ^ C1(K,6) ^ C2(K,5) ^ C3(K,4) ^
+			C4(K,3) ^ C5(K,2) ^ C6(K,1) ^ C7(K,0);
+
+		K.q[0] = L0; K.q[1] = L1; K.q[2] = L2; K.q[3] = L3;
+		K.q[4] = L4; K.q[5] = L5; K.q[6] = L6; K.q[7] = L7;
+
+		L0 ^=	C0(S,0) ^ C1(S,7) ^ C2(S,6) ^ C3(S,5) ^
+			C4(S,4) ^ C5(S,3) ^ C6(S,2) ^ C7(S,1);
+		L1 ^=	C0(S,1) ^ C1(S,0) ^ C2(S,7) ^ C3(S,6) ^
+			C4(S,5) ^ C5(S,4) ^ C6(S,3) ^ C7(S,2);
+		L2 ^=	C0(S,2) ^ C1(S,1) ^ C2(S,0) ^ C3(S,7) ^
+			C4(S,6) ^ C5(S,5) ^ C6(S,4) ^ C7(S,3);
+		L3 ^=	C0(S,3) ^ C1(S,2) ^ C2(S,1) ^ C3(S,0) ^
+			C4(S,7) ^ C5(S,6) ^ C6(S,5) ^ C7(S,4);
+		L4 ^=	C0(S,4) ^ C1(S,3) ^ C2(S,2) ^ C3(S,1) ^
+			C4(S,0) ^ C5(S,7) ^ C6(S,6) ^ C7(S,5);
+		L5 ^=	C0(S,5) ^ C1(S,4) ^ C2(S,3) ^ C3(S,2) ^
+			C4(S,1) ^ C5(S,0) ^ C6(S,7) ^ C7(S,6);
+		L6 ^=	C0(S,6) ^ C1(S,5) ^ C2(S,4) ^ C3(S,3) ^
+			C4(S,2) ^ C5(S,1) ^ C6(S,0) ^ C7(S,7);
+		L7 ^=	C0(S,7) ^ C1(S,6) ^ C2(S,5) ^ C3(S,4) ^
+			C4(S,3) ^ C5(S,2) ^ C6(S,1) ^ C7(S,0);
+
+		S.q[0] = L0; S.q[1] = L1; S.q[2] = L2; S.q[3] = L3;
+		S.q[4] = L4; S.q[5] = L5; S.q[6] = L6; S.q[7] = L7;
+#else
+		L0  = C0(K,0); L1  = C1(K,0); L2  = C2(K,0); L3  = C3(K,0);
+		L4  = C4(K,0); L5  = C5(K,0); L6  = C6(K,0); L7  = C7(K,0);
+		L0 ^= RC[r];
+
+		L1 ^= C0(K,1); L2 ^= C1(K,1); L3 ^= C2(K,1); L4 ^= C3(K,1);
+		L5 ^= C4(K,1); L6 ^= C5(K,1); L7 ^= C6(K,1); L0 ^= C7(K,1);
+
+		L2 ^= C0(K,2); L3 ^= C1(K,2); L4 ^= C2(K,2); L5 ^= C3(K,2);
+		L6 ^= C4(K,2); L7 ^= C5(K,2); L0 ^= C6(K,2); L1 ^= C7(K,2);
+
+		L3 ^= C0(K,3); L4 ^= C1(K,3); L5 ^= C2(K,3); L6 ^= C3(K,3);
+		L7 ^= C4(K,3); L0 ^= C5(K,3); L1 ^= C6(K,3); L2 ^= C7(K,3);
+
+		L4 ^= C0(K,4); L5 ^= C1(K,4); L6 ^= C2(K,4); L7 ^= C3(K,4);
+		L0 ^= C4(K,4); L1 ^= C5(K,4); L2 ^= C6(K,4); L3 ^= C7(K,4);
+
+		L5 ^= C0(K,5); L6 ^= C1(K,5); L7 ^= C2(K,5); L0 ^= C3(K,5);
+		L1 ^= C4(K,5); L2 ^= C5(K,5); L3 ^= C6(K,5); L4 ^= C7(K,5);
+
+		L6 ^= C0(K,6); L7 ^= C1(K,6); L0 ^= C2(K,6); L1 ^= C3(K,6);
+		L2 ^= C4(K,6); L3 ^= C5(K,6); L4 ^= C6(K,6); L5 ^= C7(K,6);
+
+		L7 ^= C0(K,7); L0 ^= C1(K,7); L1 ^= C2(K,7); L2 ^= C3(K,7);
+		L3 ^= C4(K,7); L4 ^= C5(K,7); L5 ^= C6(K,7); L6 ^= C7(K,7);
+
+		K.q[0] = L0; K.q[1] = L1; K.q[2] = L2; K.q[3] = L3;
+		K.q[4] = L4; K.q[5] = L5; K.q[6] = L6; K.q[7] = L7;
+
+		L0 ^= C0(S,0); L1 ^= C1(S,0); L2 ^= C2(S,0); L3 ^= C3(S,0);
+		L4 ^= C4(S,0); L5 ^= C5(S,0); L6 ^= C6(S,0); L7 ^= C7(S,0);
+
+		L1 ^= C0(S,1); L2 ^= C1(S,1); L3 ^= C2(S,1); L4 ^= C3(S,1);
+		L5 ^= C4(S,1); L6 ^= C5(S,1); L7 ^= C6(S,1); L0 ^= C7(S,1);
+
+		L2 ^= C0(S,2); L3 ^= C1(S,2); L4 ^= C2(S,2); L5 ^= C3(S,2);
+		L6 ^= C4(S,2); L7 ^= C5(S,2); L0 ^= C6(S,2); L1 ^= C7(S,2);
+
+		L3 ^= C0(S,3); L4 ^= C1(S,3); L5 ^= C2(S,3); L6 ^= C3(S,3);
+		L7 ^= C4(S,3); L0 ^= C5(S,3); L1 ^= C6(S,3); L2 ^= C7(S,3);
+
+		L4 ^= C0(S,4); L5 ^= C1(S,4); L6 ^= C2(S,4); L7 ^= C3(S,4);
+		L0 ^= C4(S,4); L1 ^= C5(S,4); L2 ^= C6(S,4); L3 ^= C7(S,4);
+
+		L5 ^= C0(S,5); L6 ^= C1(S,5); L7 ^= C2(S,5); L0 ^= C3(S,5);
+		L1 ^= C4(S,5); L2 ^= C5(S,5); L3 ^= C6(S,5); L4 ^= C7(S,5);
+
+		L6 ^= C0(S,6); L7 ^= C1(S,6); L0 ^= C2(S,6); L1 ^= C3(S,6);
+		L2 ^= C4(S,6); L3 ^= C5(S,6); L4 ^= C6(S,6); L5 ^= C7(S,6);
+
+		L7 ^= C0(S,7); L0 ^= C1(S,7); L1 ^= C2(S,7); L2 ^= C3(S,7);
+		L3 ^= C4(S,7); L4 ^= C5(S,7); L5 ^= C6(S,7); L6 ^= C7(S,7);
+
+		S.q[0] = L0; S.q[1] = L1; S.q[2] = L2; S.q[3] = L3;
+		S.q[4] = L4; S.q[5] = L5; S.q[6] = L6; S.q[7] = L7;
+#endif
+		}
+
+#ifdef STRICT_ALIGNMENT
+	if ((size_t)p & 7)
+		{
+		int i;
+		for(i=0;i<64;i++)	H->c[i] ^= S.c[i] ^ p[i];
+		}
+	else
+#endif
+		{
+		const u64 *pa=(const u64 *)p;
+		H->q[0] ^= S.q[0] ^ pa[0];
+		H->q[1] ^= S.q[1] ^ pa[1];
+		H->q[2] ^= S.q[2] ^ pa[2];
+		H->q[3] ^= S.q[3] ^ pa[3];
+		H->q[4] ^= S.q[4] ^ pa[4];
+		H->q[5] ^= S.q[5] ^ pa[5];
+		H->q[6] ^= S.q[6] ^ pa[6];
+		H->q[7] ^= S.q[7] ^ pa[7];
+		}
+#endif
+							p += 64;
+							} while(--n);
+	}
diff --git a/crypto/whrlpool/wp_dgst.c b/crypto/whrlpool/wp_dgst.c
new file mode 100644
index 000000000000..7e28bef51d0a
--- /dev/null
+++ b/crypto/whrlpool/wp_dgst.c
@@ -0,0 +1,265 @@
+/**
+ * The Whirlpool hashing function.
+ *
+ * <P>
+ * <b>References</b>
+ *
+ * <P>
+ * The Whirlpool algorithm was developed by
+ * <a href="mailto:pbarreto@scopus.com.br">Paulo S. L. M. Barreto</a> and
+ * <a href="mailto:vincent.rijmen@cryptomathic.com">Vincent Rijmen</a>.
+ *
+ * See
+ *      P.S.L.M. Barreto, V. Rijmen,
+ *      ``The Whirlpool hashing function,''
+ *      NESSIE submission, 2000 (tweaked version, 2001),
+ *      <https://www.cosic.esat.kuleuven.ac.be/nessie/workshop/submissions/whirlpool.zip>
+ *
+ * Based on "@version 3.0 (2003.03.12)" by Paulo S.L.M. Barreto and
+ * Vincent Rijmen. Lookup "reference implementations" on
+ * <http://planeta.terra.com.br/informatica/paulobarreto/>
+ *
+ * =============================================================================
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ''AS IS'' AND ANY EXPRESS
+ * OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
+ * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+ * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * OpenSSL-specific implementation notes.
+ *
+ * WHIRLPOOL_Update as well as one-stroke WHIRLPOOL both expect
+ * number of *bytes* as input length argument. Bit-oriented routine
+ * as specified by authors is called WHIRLPOOL_BitUpdate[!] and
+ * does not have one-stroke counterpart.
+ *
+ * WHIRLPOOL_BitUpdate implements byte-oriented loop, essentially
+ * to serve WHIRLPOOL_Update. This is done for performance.
+ *
+ * Unlike authors' reference implementation, block processing
+ * routine whirlpool_block is designed to operate on multi-block
+ * input. This is done for perfomance.
+ */
+
+#include "wp_locl.h"
+#include <openssl/crypto.h>
+#include <string.h>
+
+fips_md_init(WHIRLPOOL)
+	{
+	memset (c,0,sizeof(*c));
+	return(1);
+	}
+
+int WHIRLPOOL_Update	(WHIRLPOOL_CTX *c,const void *_inp,size_t bytes)
+	{
+	/* Well, largest suitable chunk size actually is
+	 * (1<<(sizeof(size_t)*8-3))-64, but below number
+	 * is large enough for not to care about excessive
+	 * calls to WHIRLPOOL_BitUpdate... */
+	size_t chunk = ((size_t)1)<<(sizeof(size_t)*8-4);
+	const unsigned char *inp = _inp;
+
+	while (bytes>=chunk)
+		{
+		WHIRLPOOL_BitUpdate(c,inp,chunk*8);
+		bytes -= chunk;
+		inp   += chunk;
+		}
+	if (bytes)
+		WHIRLPOOL_BitUpdate(c,inp,bytes*8);
+
+	return(1);
+	}
+
+void WHIRLPOOL_BitUpdate(WHIRLPOOL_CTX *c,const void *_inp,size_t bits)
+	{
+	size_t		n;
+	unsigned int	bitoff = c->bitoff,
+			bitrem = bitoff%8,
+			inpgap = (8-(unsigned int)bits%8)&7;
+	const unsigned char *inp=_inp;
+
+	/* This 256-bit increment procedure relies on the size_t
+	 * being natural size of CPU register, so that we don't
+	 * have to mask the value in order to detect overflows. */
+	c->bitlen[0] += bits;
+	if (c->bitlen[0] < bits)	/* overflow */
+		{
+		n = 1;
+		do  { 	c->bitlen[n]++;
+		    } while(c->bitlen[n]==0
+		   	    && ++n<(WHIRLPOOL_COUNTER/sizeof(size_t)));
+		}
+
+#ifndef OPENSSL_SMALL_FOOTPRINT
+	reconsider:
+	if (inpgap==0 && bitrem==0)	/* byte-oriented loop */
+		{
+		while (bits)
+			{
+			if (bitoff==0 && (n=bits/WHIRLPOOL_BBLOCK))
+				{
+				whirlpool_block(c,inp,n);
+				inp  += n*WHIRLPOOL_BBLOCK/8;
+				bits %= WHIRLPOOL_BBLOCK;
+				}
+			else
+				{
+				unsigned int byteoff = bitoff/8;
+
+				bitrem = WHIRLPOOL_BBLOCK - bitoff;/* re-use bitrem */
+				if (bits >= bitrem)
+					{
+					bits -= bitrem;
+					bitrem /= 8;
+					memcpy(c->data+byteoff,inp,bitrem);
+					inp  += bitrem;
+					whirlpool_block(c,c->data,1);
+					bitoff = 0;
+					}
+				else
+					{
+					memcpy(c->data+byteoff,inp,bits/8);
+					bitoff += (unsigned int)bits;
+					bits = 0;
+					}
+				c->bitoff = bitoff;
+				}
+			}
+		}
+	else				/* bit-oriented loop */
+#endif
+		{
+		/*
+			   inp
+			   |
+			   +-------+-------+-------
+			      |||||||||||||||||||||
+			   +-------+-------+-------
+		+-------+-------+-------+-------+-------
+		||||||||||||||				c->data
+		+-------+-------+-------+-------+-------
+			|
+			c->bitoff/8
+		*/
+		while (bits)
+			{
+			unsigned int	byteoff	= bitoff/8;
+			unsigned char	b;
+
+#ifndef OPENSSL_SMALL_FOOTPRINT
+			if (bitrem==inpgap)
+				{
+				c->data[byteoff++] |= inp[0] & (0xff>>inpgap);
+				inpgap = 8-inpgap;
+				bitoff += inpgap;  bitrem = 0;	/* bitoff%8 */
+				bits   -= inpgap;  inpgap = 0;	/* bits%8   */
+				inp++;
+				if (bitoff==WHIRLPOOL_BBLOCK)
+					{
+					whirlpool_block(c,c->data,1);
+					bitoff = 0;
+					}
+				c->bitoff = bitoff;
+				goto reconsider;
+				}
+			else
+#endif
+			if (bits>=8)
+				{
+				b  = ((inp[0]<<inpgap) | (inp[1]>>(8-inpgap)));
+				b &= 0xff;
+				if (bitrem)	c->data[byteoff++] |= b>>bitrem;
+				else		c->data[byteoff++]  = b;
+				bitoff += 8;
+				bits   -= 8;
+				inp++;
+				if (bitoff>=WHIRLPOOL_BBLOCK)
+					{
+					whirlpool_block(c,c->data,1);
+					byteoff  = 0;
+					bitoff  %= WHIRLPOOL_BBLOCK;
+					}
+				if (bitrem)	c->data[byteoff] = b<<(8-bitrem);
+				}
+			else	/* remaining less than 8 bits */
+				{
+				b = (inp[0]<<inpgap)&0xff;
+				if (bitrem)	c->data[byteoff++] |= b>>bitrem;
+				else		c->data[byteoff++]  = b;
+				bitoff += (unsigned int)bits;
+				if (bitoff==WHIRLPOOL_BBLOCK)
+					{
+					whirlpool_block(c,c->data,1);
+					byteoff  = 0;
+			        	bitoff  %= WHIRLPOOL_BBLOCK;
+					}
+				if (bitrem)	c->data[byteoff] = b<<(8-bitrem);
+				bits = 0;
+				}
+			c->bitoff = bitoff;
+			}
+		}
+	}
+
+int WHIRLPOOL_Final	(unsigned char *md,WHIRLPOOL_CTX *c)
+	{
+	unsigned int	bitoff  = c->bitoff,
+			byteoff = bitoff/8;
+	size_t		i,j,v;
+	unsigned char  *p;
+
+	bitoff %= 8;
+	if (bitoff)	c->data[byteoff] |= 0x80>>bitoff;
+	else		c->data[byteoff]  = 0x80;
+	byteoff++;
+
+	/* pad with zeros */
+	if (byteoff > (WHIRLPOOL_BBLOCK/8-WHIRLPOOL_COUNTER))
+		{
+		if (byteoff<WHIRLPOOL_BBLOCK/8)
+			memset(&c->data[byteoff],0,WHIRLPOOL_BBLOCK/8-byteoff);
+		whirlpool_block(c,c->data,1);
+		byteoff = 0;
+		}
+	if (byteoff < (WHIRLPOOL_BBLOCK/8-WHIRLPOOL_COUNTER))
+		memset(&c->data[byteoff],0,
+			(WHIRLPOOL_BBLOCK/8-WHIRLPOOL_COUNTER)-byteoff);
+	/* smash 256-bit c->bitlen in big-endian order */
+	p = &c->data[WHIRLPOOL_BBLOCK/8-1];	/* last byte in c->data */
+	for(i=0;i<WHIRLPOOL_COUNTER/sizeof(size_t);i++)
+		for(v=c->bitlen[i],j=0;j<sizeof(size_t);j++,v>>=8)
+			*p-- = (unsigned char)(v&0xff);
+
+	whirlpool_block(c,c->data,1);
+
+	if (md)	{
+		memcpy(md,c->H.c,WHIRLPOOL_DIGEST_LENGTH);
+		memset(c,0,sizeof(*c));
+		return(1);
+		}
+	return(0);
+	}
+
+unsigned char *WHIRLPOOL(const void *inp, size_t bytes,unsigned char *md)
+	{
+	WHIRLPOOL_CTX ctx;
+	static unsigned char m[WHIRLPOOL_DIGEST_LENGTH];
+
+	if (md == NULL) md=m;
+	WHIRLPOOL_Init(&ctx);
+	WHIRLPOOL_Update(&ctx,inp,bytes);
+	WHIRLPOOL_Final(md,&ctx);
+	return(md);
+	}
diff --git a/crypto/whrlpool/wp_locl.h b/crypto/whrlpool/wp_locl.h
new file mode 100644
index 000000000000..94e56a39f18a
--- /dev/null
+++ b/crypto/whrlpool/wp_locl.h
@@ -0,0 +1,3 @@
+#include <openssl/whrlpool.h>
+
+void whirlpool_block(WHIRLPOOL_CTX *,const void *,size_t);
diff --git a/crypto/whrlpool/wp_test.c b/crypto/whrlpool/wp_test.c
new file mode 100644
index 000000000000..c68c2c62cabd
--- /dev/null
+++ b/crypto/whrlpool/wp_test.c
@@ -0,0 +1,228 @@
+/* ====================================================================
+ * Copyright (c) 2005 The OpenSSL Project.  All rights reserved.
+ * ====================================================================
+ */
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+
+#include <openssl/whrlpool.h>
+#include <openssl/crypto.h>
+
+#if defined(OPENSSL_NO_WHIRLPOOL)
+int main(int argc, char *argv[])
+{
+    printf("No Whirlpool support\n");
+    return(0);
+}
+#else
+
+/* ISO/IEC 10118-3 test vector set */
+unsigned char iso_test_1[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x19,0xFA,0x61,0xD7,0x55,0x22,0xA4,0x66,
+	0x9B,0x44,0xE3,0x9C,0x1D,0x2E,0x17,0x26,
+	0xC5,0x30,0x23,0x21,0x30,0xD4,0x07,0xF8,
+	0x9A,0xFE,0xE0,0x96,0x49,0x97,0xF7,0xA7,
+	0x3E,0x83,0xBE,0x69,0x8B,0x28,0x8F,0xEB,
+	0xCF,0x88,0xE3,0xE0,0x3C,0x4F,0x07,0x57,
+	0xEA,0x89,0x64,0xE5,0x9B,0x63,0xD9,0x37,
+	0x08,0xB1,0x38,0xCC,0x42,0xA6,0x6E,0xB3 };
+
+unsigned char iso_test_2[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x8A,0xCA,0x26,0x02,0x79,0x2A,0xEC,0x6F,
+	0x11,0xA6,0x72,0x06,0x53,0x1F,0xB7,0xD7,
+	0xF0,0xDF,0xF5,0x94,0x13,0x14,0x5E,0x69,
+	0x73,0xC4,0x50,0x01,0xD0,0x08,0x7B,0x42,
+	0xD1,0x1B,0xC6,0x45,0x41,0x3A,0xEF,0xF6,
+	0x3A,0x42,0x39,0x1A,0x39,0x14,0x5A,0x59,
+	0x1A,0x92,0x20,0x0D,0x56,0x01,0x95,0xE5,
+	0x3B,0x47,0x85,0x84,0xFD,0xAE,0x23,0x1A };
+
+unsigned char iso_test_3[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x4E,0x24,0x48,0xA4,0xC6,0xF4,0x86,0xBB,
+	0x16,0xB6,0x56,0x2C,0x73,0xB4,0x02,0x0B,
+	0xF3,0x04,0x3E,0x3A,0x73,0x1B,0xCE,0x72,
+	0x1A,0xE1,0xB3,0x03,0xD9,0x7E,0x6D,0x4C,
+	0x71,0x81,0xEE,0xBD,0xB6,0xC5,0x7E,0x27,
+	0x7D,0x0E,0x34,0x95,0x71,0x14,0xCB,0xD6,
+	0xC7,0x97,0xFC,0x9D,0x95,0xD8,0xB5,0x82,
+	0xD2,0x25,0x29,0x20,0x76,0xD4,0xEE,0xF5 };
+
+unsigned char iso_test_4[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x37,0x8C,0x84,0xA4,0x12,0x6E,0x2D,0xC6,
+	0xE5,0x6D,0xCC,0x74,0x58,0x37,0x7A,0xAC,
+	0x83,0x8D,0x00,0x03,0x22,0x30,0xF5,0x3C,
+	0xE1,0xF5,0x70,0x0C,0x0F,0xFB,0x4D,0x3B,
+	0x84,0x21,0x55,0x76,0x59,0xEF,0x55,0xC1,
+	0x06,0xB4,0xB5,0x2A,0xC5,0xA4,0xAA,0xA6,
+	0x92,0xED,0x92,0x00,0x52,0x83,0x8F,0x33,
+	0x62,0xE8,0x6D,0xBD,0x37,0xA8,0x90,0x3E };
+
+unsigned char iso_test_5[WHIRLPOOL_DIGEST_LENGTH] = {
+	0xF1,0xD7,0x54,0x66,0x26,0x36,0xFF,0xE9,
+	0x2C,0x82,0xEB,0xB9,0x21,0x2A,0x48,0x4A,
+	0x8D,0x38,0x63,0x1E,0xAD,0x42,0x38,0xF5,
+	0x44,0x2E,0xE1,0x3B,0x80,0x54,0xE4,0x1B,
+	0x08,0xBF,0x2A,0x92,0x51,0xC3,0x0B,0x6A,
+	0x0B,0x8A,0xAE,0x86,0x17,0x7A,0xB4,0xA6,
+	0xF6,0x8F,0x67,0x3E,0x72,0x07,0x86,0x5D,
+	0x5D,0x98,0x19,0xA3,0xDB,0xA4,0xEB,0x3B };
+
+unsigned char iso_test_6[WHIRLPOOL_DIGEST_LENGTH] = {
+	0xDC,0x37,0xE0,0x08,0xCF,0x9E,0xE6,0x9B,
+	0xF1,0x1F,0x00,0xED,0x9A,0xBA,0x26,0x90,
+	0x1D,0xD7,0xC2,0x8C,0xDE,0xC0,0x66,0xCC,
+	0x6A,0xF4,0x2E,0x40,0xF8,0x2F,0x3A,0x1E,
+	0x08,0xEB,0xA2,0x66,0x29,0x12,0x9D,0x8F,
+	0xB7,0xCB,0x57,0x21,0x1B,0x92,0x81,0xA6,
+	0x55,0x17,0xCC,0x87,0x9D,0x7B,0x96,0x21,
+	0x42,0xC6,0x5F,0x5A,0x7A,0xF0,0x14,0x67 };
+
+unsigned char iso_test_7[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x46,0x6E,0xF1,0x8B,0xAB,0xB0,0x15,0x4D,
+	0x25,0xB9,0xD3,0x8A,0x64,0x14,0xF5,0xC0,
+	0x87,0x84,0x37,0x2B,0xCC,0xB2,0x04,0xD6,
+	0x54,0x9C,0x4A,0xFA,0xDB,0x60,0x14,0x29,
+	0x4D,0x5B,0xD8,0xDF,0x2A,0x6C,0x44,0xE5,
+	0x38,0xCD,0x04,0x7B,0x26,0x81,0xA5,0x1A,
+	0x2C,0x60,0x48,0x1E,0x88,0xC5,0xA2,0x0B,
+	0x2C,0x2A,0x80,0xCF,0x3A,0x9A,0x08,0x3B };
+
+unsigned char iso_test_8[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x2A,0x98,0x7E,0xA4,0x0F,0x91,0x70,0x61,
+	0xF5,0xD6,0xF0,0xA0,0xE4,0x64,0x4F,0x48,
+	0x8A,0x7A,0x5A,0x52,0xDE,0xEE,0x65,0x62,
+	0x07,0xC5,0x62,0xF9,0x88,0xE9,0x5C,0x69,
+	0x16,0xBD,0xC8,0x03,0x1B,0xC5,0xBE,0x1B,
+	0x7B,0x94,0x76,0x39,0xFE,0x05,0x0B,0x56,
+	0x93,0x9B,0xAA,0xA0,0xAD,0xFF,0x9A,0xE6,
+	0x74,0x5B,0x7B,0x18,0x1C,0x3B,0xE3,0xFD };
+
+unsigned char iso_test_9[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x0C,0x99,0x00,0x5B,0xEB,0x57,0xEF,0xF5,
+	0x0A,0x7C,0xF0,0x05,0x56,0x0D,0xDF,0x5D,
+	0x29,0x05,0x7F,0xD8,0x6B,0x20,0xBF,0xD6,
+	0x2D,0xEC,0xA0,0xF1,0xCC,0xEA,0x4A,0xF5,
+	0x1F,0xC1,0x54,0x90,0xED,0xDC,0x47,0xAF,
+	0x32,0xBB,0x2B,0x66,0xC3,0x4F,0xF9,0xAD,
+	0x8C,0x60,0x08,0xAD,0x67,0x7F,0x77,0x12,
+	0x69,0x53,0xB2,0x26,0xE4,0xED,0x8B,0x01 };
+
+int main (int argc,char *argv[])
+{ unsigned char md[WHIRLPOOL_DIGEST_LENGTH];
+  int		i;
+  WHIRLPOOL_CTX	ctx;
+
+#ifdef OPENSSL_IA32_SSE2
+    /* Alternative to this is to call OpenSSL_add_all_algorithms...
+     * The below code is retained exclusively for debugging purposes. */
+    { char      *env;
+
+	if ((env=getenv("OPENSSL_ia32cap")))
+	    OPENSSL_ia32cap = strtoul (env,NULL,0);
+    }
+#endif
+
+    fprintf(stdout,"Testing Whirlpool ");
+
+    WHIRLPOOL("",0,md);
+    if (memcmp(md,iso_test_1,sizeof(iso_test_1)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 1 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("a",1,md);
+    if (memcmp(md,iso_test_2,sizeof(iso_test_2)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 2 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("abc",3,md);
+    if (memcmp(md,iso_test_3,sizeof(iso_test_3)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 3 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("message digest",14,md);
+    if (memcmp(md,iso_test_4,sizeof(iso_test_4)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 4 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("abcdefghijklmnopqrstuvwxyz",26,md);
+    if (memcmp(md,iso_test_5,sizeof(iso_test_5)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 5 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL(	"ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+		"abcdefghijklmnopqrstuvwxyz"
+		"0123456789",62,md);
+    if (memcmp(md,iso_test_6,sizeof(iso_test_6)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 6 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL(	"1234567890""1234567890""1234567890""1234567890"
+		"1234567890""1234567890""1234567890""1234567890",80,md);
+    if (memcmp(md,iso_test_7,sizeof(iso_test_7)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 7 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("abcdbcdecdefdefgefghfghighijhijk",32,md);
+    if (memcmp(md,iso_test_8,sizeof(iso_test_8)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 8 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+ 
+    WHIRLPOOL_Init (&ctx);
+    for (i=0;i<1000000;i+=288)
+	WHIRLPOOL_Update (&ctx,	"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa",
+				(1000000-i)<288?1000000-i:288);
+    WHIRLPOOL_Final (md,&ctx);
+    if (memcmp(md,iso_test_9,sizeof(iso_test_9)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 9 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    fprintf(stdout," passed.\n"); fflush(stdout);
+
+  return 0;
+}
+#endif
diff --git a/crypto/x509/Makefile b/crypto/x509/Makefile
index 464752b159a0..72c82278f43d 100644
--- a/crypto/x509/Makefile
+++ b/crypto/x509/Makefile
@@ -43,12 +43,12 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
 files:
-	$(PERL) $(TOP)/util/files.pl Makefile.ssl >> $(TOP)/MINFO
+	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
 
 links:
 	@$(PERL) $(TOP)/util/mklink.pl ../../include/openssl $(EXHEADER)
@@ -89,37 +89,35 @@ by_dir.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 by_dir.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 by_dir.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 by_dir.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-by_dir.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-by_dir.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-by_dir.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-by_dir.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-by_dir.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-by_dir.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-by_dir.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-by_dir.o: ../cryptlib.h by_dir.c
+by_dir.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+by_dir.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+by_dir.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+by_dir.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+by_dir.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+by_dir.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+by_dir.o: ../../include/openssl/x509_vfy.h ../cryptlib.h by_dir.c
 by_file.o: ../../e_os.h ../../include/openssl/asn1.h
 by_file.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 by_file.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 by_file.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 by_file.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-by_file.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-by_file.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-by_file.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-by_file.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-by_file.o: ../../include/openssl/pem.h ../../include/openssl/pem2.h
-by_file.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-by_file.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-by_file.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-by_file.o: ../../include/openssl/x509_vfy.h ../cryptlib.h by_file.c
+by_file.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+by_file.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+by_file.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+by_file.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pem.h
+by_file.o: ../../include/openssl/pem2.h ../../include/openssl/pkcs7.h
+by_file.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+by_file.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+by_file.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+by_file.o: ../cryptlib.h by_file.c
 x509_att.o: ../../e_os.h ../../include/openssl/asn1.h
 x509_att.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509_att.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 x509_att.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_att.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_att.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_att.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_att.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_att.o: ../../include/openssl/opensslconf.h
+x509_att.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_att.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 x509_att.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_att.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_att.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -132,9 +130,8 @@ x509_cmp.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 x509_cmp.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_cmp.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_cmp.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_cmp.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_cmp.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_cmp.o: ../../include/openssl/opensslconf.h
+x509_cmp.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_cmp.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 x509_cmp.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_cmp.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_cmp.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -146,22 +143,22 @@ x509_d2.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509_d2.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509_d2.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509_d2.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509_d2.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509_d2.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509_d2.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-x509_d2.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-x509_d2.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-x509_d2.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-x509_d2.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-x509_d2.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x509_d2.c
+x509_d2.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509_d2.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509_d2.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+x509_d2.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+x509_d2.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+x509_d2.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+x509_d2.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+x509_d2.o: ../cryptlib.h x509_d2.c
 x509_def.o: ../../e_os.h ../../include/openssl/asn1.h
 x509_def.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509_def.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509_def.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509_def.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509_def.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509_def.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509_def.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509_def.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509_def.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509_def.o: ../../include/openssl/opensslconf.h
 x509_def.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_def.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_def.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -172,9 +169,8 @@ x509_err.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x509_err.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_err.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_err.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_err.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_err.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_err.o: ../../include/openssl/opensslconf.h
+x509_err.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_err.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 x509_err.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_err.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_err.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -186,9 +182,8 @@ x509_ext.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 x509_ext.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_ext.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_ext.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_ext.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_ext.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_ext.o: ../../include/openssl/opensslconf.h
+x509_ext.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_ext.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 x509_ext.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_ext.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_ext.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -201,22 +196,22 @@ x509_lu.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 x509_lu.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_lu.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_lu.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_lu.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_lu.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_lu.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x509_lu.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x509_lu.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x509_lu.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x509_lu.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x509_lu.o: ../../include/openssl/x509v3.h ../cryptlib.h x509_lu.c
+x509_lu.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_lu.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509_lu.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x509_lu.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x509_lu.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x509_lu.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x509_lu.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+x509_lu.o: ../cryptlib.h x509_lu.c
 x509_obj.o: ../../e_os.h ../../include/openssl/asn1.h
 x509_obj.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509_obj.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509_obj.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509_obj.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509_obj.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509_obj.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509_obj.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509_obj.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509_obj.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509_obj.o: ../../include/openssl/opensslconf.h
 x509_obj.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_obj.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_obj.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -228,21 +223,20 @@ x509_r2x.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x509_r2x.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_r2x.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_r2x.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_r2x.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_r2x.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_r2x.o: ../../include/openssl/opensslconf.h
+x509_r2x.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_r2x.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 x509_r2x.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_r2x.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_r2x.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
 x509_r2x.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
 x509_r2x.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x509_r2x.c
 x509_req.o: ../../e_os.h ../../include/openssl/asn1.h
-x509_req.o: ../../include/openssl/bio.h ../../include/openssl/bn.h
-x509_req.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
-x509_req.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
-x509_req.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-x509_req.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_req.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+x509_req.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
+x509_req.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
+x509_req.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+x509_req.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+x509_req.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+x509_req.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
 x509_req.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
 x509_req.o: ../../include/openssl/opensslconf.h
 x509_req.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
@@ -256,9 +250,9 @@ x509_set.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509_set.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509_set.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509_set.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509_set.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509_set.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509_set.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509_set.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509_set.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509_set.o: ../../include/openssl/opensslconf.h
 x509_set.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_set.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_set.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -270,9 +264,8 @@ x509_trs.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 x509_trs.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_trs.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_trs.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_trs.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_trs.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_trs.o: ../../include/openssl/opensslconf.h
+x509_trs.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_trs.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 x509_trs.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_trs.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_trs.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -284,9 +277,9 @@ x509_txt.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509_txt.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509_txt.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509_txt.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509_txt.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509_txt.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509_txt.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509_txt.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509_txt.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509_txt.o: ../../include/openssl/opensslconf.h
 x509_txt.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_txt.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_txt.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -298,23 +291,22 @@ x509_v3.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 x509_v3.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_v3.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_v3.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_v3.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_v3.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_v3.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-x509_v3.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-x509_v3.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x509_v3.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x509_v3.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x509_v3.o: ../../include/openssl/x509v3.h ../cryptlib.h x509_v3.c
+x509_v3.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_v3.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509_v3.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+x509_v3.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+x509_v3.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x509_v3.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x509_v3.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+x509_v3.o: ../cryptlib.h x509_v3.c
 x509_vfy.o: ../../e_os.h ../../include/openssl/asn1.h
 x509_vfy.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509_vfy.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 x509_vfy.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_vfy.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_vfy.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_vfy.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_vfy.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_vfy.o: ../../include/openssl/opensslconf.h
+x509_vfy.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_vfy.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 x509_vfy.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_vfy.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_vfy.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -327,9 +319,8 @@ x509_vpm.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 x509_vpm.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 x509_vpm.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 x509_vpm.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-x509_vpm.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-x509_vpm.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-x509_vpm.o: ../../include/openssl/opensslconf.h
+x509_vpm.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+x509_vpm.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 x509_vpm.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509_vpm.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509_vpm.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -341,9 +332,9 @@ x509cset.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509cset.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509cset.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509cset.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509cset.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509cset.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509cset.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509cset.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509cset.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509cset.o: ../../include/openssl/opensslconf.h
 x509cset.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509cset.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509cset.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -354,9 +345,9 @@ x509name.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509name.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509name.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509name.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509name.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509name.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509name.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509name.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509name.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509name.o: ../../include/openssl/opensslconf.h
 x509name.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509name.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509name.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -367,9 +358,9 @@ x509rset.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509rset.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509rset.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509rset.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509rset.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509rset.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509rset.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509rset.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509rset.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509rset.o: ../../include/openssl/opensslconf.h
 x509rset.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509rset.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509rset.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -380,9 +371,9 @@ x509spki.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509spki.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509spki.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509spki.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509spki.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509spki.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509spki.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509spki.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509spki.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509spki.o: ../../include/openssl/opensslconf.h
 x509spki.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509spki.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509spki.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -393,9 +384,9 @@ x509type.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 x509type.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 x509type.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x509type.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x509type.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x509type.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x509type.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+x509type.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x509type.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x509type.o: ../../include/openssl/opensslconf.h
 x509type.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 x509type.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 x509type.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -406,12 +397,11 @@ x_all.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 x_all.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
 x_all.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 x_all.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-x_all.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-x_all.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-x_all.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-x_all.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-x_all.o: ../../include/openssl/pkcs7.h ../../include/openssl/rsa.h
-x_all.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-x_all.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-x_all.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-x_all.o: ../cryptlib.h x_all.c
+x_all.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+x_all.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+x_all.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+x_all.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+x_all.o: ../../include/openssl/rsa.h ../../include/openssl/safestack.h
+x_all.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+x_all.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+x_all.o: ../../include/openssl/x509_vfy.h ../cryptlib.h x_all.c
diff --git a/crypto/x509/by_dir.c b/crypto/x509/by_dir.c
index b3acd80f25b1..27ca5150c195 100644
--- a/crypto/x509/by_dir.c
+++ b/crypto/x509/by_dir.c
@@ -65,28 +65,36 @@
 #ifndef NO_SYS_TYPES_H
 # include <sys/types.h>
 #endif
-#ifdef MAC_OS_pre_X
-# include <stat.h>
-#else
+#ifndef OPENSSL_NO_POSIX_IO
 # include <sys/stat.h>
 #endif
 
 #include <openssl/lhash.h>
 #include <openssl/x509.h>
 
-#ifdef _WIN32
-#define stat	_stat
-#endif
+
+typedef struct lookup_dir_hashes_st
+	{
+	unsigned long hash;
+	int suffix;
+	} BY_DIR_HASH;
+
+typedef struct lookup_dir_entry_st
+	{
+	char *dir;
+	int dir_type;
+	STACK_OF(BY_DIR_HASH) *hashes;
+	} BY_DIR_ENTRY;
 
 typedef struct lookup_dir_st
 	{
 	BUF_MEM *buffer;
-	int num_dirs;
-	char **dirs;
-	int *dirs_type;
-	int num_dirs_alloced;
+	STACK_OF(BY_DIR_ENTRY) *dirs;
 	} BY_DIR;
 
+DECLARE_STACK_OF(BY_DIR_HASH)
+DECLARE_STACK_OF(BY_DIR_ENTRY)
+
 static int dir_ctrl(X509_LOOKUP *ctx, int cmd, const char *argp, long argl,
 	char **ret);
 static int new_dir(X509_LOOKUP *lu);
@@ -127,7 +135,7 @@ static int dir_ctrl(X509_LOOKUP *ctx, int cmd, const char *argp, long argl,
 	case X509_L_ADD_DIR:
 		if (argl == X509_FILETYPE_DEFAULT)
 			{
-			dir=(char *)Getenv(X509_get_default_cert_dir_env());
+			dir=(char *)getenv(X509_get_default_cert_dir_env());
 			if (dir)
 				ret=add_cert_dir(ld,dir,X509_FILETYPE_PEM);
 			else
@@ -156,34 +164,51 @@ static int new_dir(X509_LOOKUP *lu)
 		OPENSSL_free(a);
 		return(0);
 		}
-	a->num_dirs=0;
 	a->dirs=NULL;
-	a->dirs_type=NULL;
-	a->num_dirs_alloced=0;
 	lu->method_data=(char *)a;
 	return(1);
 	}
 
+static void by_dir_hash_free(BY_DIR_HASH *hash)
+	{
+	OPENSSL_free(hash);
+	}
+
+static int by_dir_hash_cmp(const BY_DIR_HASH * const *a,
+			const BY_DIR_HASH * const *b)
+	{
+	if ((*a)->hash > (*b)->hash)
+		return 1;
+	if ((*a)->hash < (*b)->hash)
+		return -1;
+	return 0;
+	}
+
+static void by_dir_entry_free(BY_DIR_ENTRY *ent)
+	{
+	if (ent->dir)
+		OPENSSL_free(ent->dir);
+	if (ent->hashes)
+		sk_BY_DIR_HASH_pop_free(ent->hashes, by_dir_hash_free);
+	OPENSSL_free(ent);
+	}
+
 static void free_dir(X509_LOOKUP *lu)
 	{
 	BY_DIR *a;
-	int i;
 
 	a=(BY_DIR *)lu->method_data;
-	for (i=0; i<a->num_dirs; i++)
-		if (a->dirs[i] != NULL) OPENSSL_free(a->dirs[i]);
-	if (a->dirs != NULL) OPENSSL_free(a->dirs);
-	if (a->dirs_type != NULL) OPENSSL_free(a->dirs_type);
-	if (a->buffer != NULL) BUF_MEM_free(a->buffer);
+	if (a->dirs != NULL)
+		sk_BY_DIR_ENTRY_pop_free(a->dirs, by_dir_entry_free);
+	if (a->buffer != NULL)
+		BUF_MEM_free(a->buffer);
 	OPENSSL_free(a);
 	}
 
 static int add_cert_dir(BY_DIR *ctx, const char *dir, int type)
 	{
 	int j,len;
-	int *ip;
 	const char *s,*ss,*p;
-	char **pp;
 
 	if (dir == NULL || !*dir)
 	    {
@@ -197,49 +222,52 @@ static int add_cert_dir(BY_DIR *ctx, const char *dir, int type)
 		{
 		if ((*p == LIST_SEPARATOR_CHAR) || (*p == '\0'))
 			{
+			BY_DIR_ENTRY *ent;
 			ss=s;
 			s=p+1;
 			len=(int)(p-ss);
 			if (len == 0) continue;
-			for (j=0; j<ctx->num_dirs; j++)
-				if (strlen(ctx->dirs[j]) == (size_t)len &&
-				    strncmp(ctx->dirs[j],ss,(unsigned int)len) == 0)
+			for (j=0; j < sk_BY_DIR_ENTRY_num(ctx->dirs); j++)
+				{
+				ent = sk_BY_DIR_ENTRY_value(ctx->dirs, j);
+				if (strlen(ent->dir) == (size_t)len &&
+				    strncmp(ent->dir,ss,(unsigned int)len) == 0)
 					break;
-			if (j<ctx->num_dirs)
+				}
+			if (j < sk_BY_DIR_ENTRY_num(ctx->dirs))
 				continue;
-			if (ctx->num_dirs_alloced < (ctx->num_dirs+1))
+			if (ctx->dirs == NULL)
 				{
-				ctx->num_dirs_alloced+=10;
-				pp=(char **)OPENSSL_malloc(ctx->num_dirs_alloced*
-					sizeof(char *));
-				ip=(int *)OPENSSL_malloc(ctx->num_dirs_alloced*
-					sizeof(int));
-				if ((pp == NULL) || (ip == NULL))
+				ctx->dirs = sk_BY_DIR_ENTRY_new_null();
+				if (!ctx->dirs)
 					{
 					X509err(X509_F_ADD_CERT_DIR,ERR_R_MALLOC_FAILURE);
-					return(0);
+					return 0;
 					}
-				memcpy(pp,ctx->dirs,(ctx->num_dirs_alloced-10)*
-					sizeof(char *));
-				memcpy(ip,ctx->dirs_type,(ctx->num_dirs_alloced-10)*
-					sizeof(int));
-				if (ctx->dirs != NULL)
-					OPENSSL_free(ctx->dirs);
-				if (ctx->dirs_type != NULL)
-					OPENSSL_free(ctx->dirs_type);
-				ctx->dirs=pp;
-				ctx->dirs_type=ip;
 				}
-			ctx->dirs_type[ctx->num_dirs]=type;
-			ctx->dirs[ctx->num_dirs]=(char *)OPENSSL_malloc((unsigned int)len+1);
-			if (ctx->dirs[ctx->num_dirs] == NULL) return(0);
-			strncpy(ctx->dirs[ctx->num_dirs],ss,(unsigned int)len);
-			ctx->dirs[ctx->num_dirs][len]='\0';
-			ctx->num_dirs++;
+			ent = OPENSSL_malloc(sizeof(BY_DIR_ENTRY));
+			if (!ent)
+				return 0;
+			ent->dir_type = type;
+			ent->hashes = sk_BY_DIR_HASH_new(by_dir_hash_cmp);
+			ent->dir = OPENSSL_malloc((unsigned int)len+1);
+			if (!ent->dir || !ent->hashes)
+				{
+				by_dir_entry_free(ent);
+				return 0;
+				}
+			strncpy(ent->dir,ss,(unsigned int)len);
+			ent->dir[len] = '\0';
+			if (!sk_BY_DIR_ENTRY_push(ctx->dirs, ent))
+				{
+				by_dir_entry_free(ent);
+				return 0;
+				}
 			}
-		if (*p == '\0') break;
+		if (*p == '\0')
+			break;
 		}
-	return(1);
+	return 1;
 	}
 
 static int get_cert_by_subject(X509_LOOKUP *xl, int type, X509_NAME *name,
@@ -260,7 +288,6 @@ static int get_cert_by_subject(X509_LOOKUP *xl, int type, X509_NAME *name,
 	int i,j,k;
 	unsigned long h;
 	BUF_MEM *b=NULL;
-	struct stat st;
 	X509_OBJECT stmp,*tmp;
 	const char *postfix="";
 
@@ -296,20 +323,45 @@ static int get_cert_by_subject(X509_LOOKUP *xl, int type, X509_NAME *name,
 	ctx=(BY_DIR *)xl->method_data;
 
 	h=X509_NAME_hash(name);
-	for (i=0; i<ctx->num_dirs; i++)
+	for (i=0; i < sk_BY_DIR_ENTRY_num(ctx->dirs); i++)
 		{
-		j=strlen(ctx->dirs[i])+1+8+6+1+1;
+		BY_DIR_ENTRY *ent;
+		int idx;
+		BY_DIR_HASH htmp, *hent;
+		ent = sk_BY_DIR_ENTRY_value(ctx->dirs, i);
+		j=strlen(ent->dir)+1+8+6+1+1;
 		if (!BUF_MEM_grow(b,j))
 			{
 			X509err(X509_F_GET_CERT_BY_SUBJECT,ERR_R_MALLOC_FAILURE);
 			goto finish;
 			}
-		k=0;
+		if (type == X509_LU_CRL && ent->hashes)
+			{
+			htmp.hash = h;
+			CRYPTO_r_lock(CRYPTO_LOCK_X509_STORE);
+			idx = sk_BY_DIR_HASH_find(ent->hashes, &htmp);
+			if (idx >= 0)
+				{
+				hent = sk_BY_DIR_HASH_value(ent->hashes, idx);
+				k = hent->suffix;
+				}
+			else
+				{
+				hent = NULL;
+				k=0;
+				}
+			CRYPTO_r_unlock(CRYPTO_LOCK_X509_STORE);
+			}
+		else
+			{
+			k = 0;
+			hent = NULL;
+			}
 		for (;;)
 			{
 			char c = '/';
 #ifdef OPENSSL_SYS_VMS
-			c = ctx->dirs[i][strlen(ctx->dirs[i])-1];
+			c = ent->dir[strlen(ent->dir)-1];
 			if (c != ':' && c != '>' && c != ']')
 				{
 				/* If no separator is present, we assume the
@@ -330,32 +382,40 @@ static int get_cert_by_subject(X509_LOOKUP *xl, int type, X509_NAME *name,
 				/* This is special.  When c == '\0', no
 				   directory separator should be added. */
 				BIO_snprintf(b->data,b->max,
-					"%s%08lx.%s%d",ctx->dirs[i],h,
+					"%s%08lx.%s%d",ent->dir,h,
 					postfix,k);
 				}
 			else
 				{
 				BIO_snprintf(b->data,b->max,
-					"%s%c%08lx.%s%d",ctx->dirs[i],c,h,
+					"%s%c%08lx.%s%d",ent->dir,c,h,
 					postfix,k);
 				}
-			k++;
+#ifndef OPENSSL_NO_POSIX_IO
+#ifdef _WIN32
+#define stat _stat
+#endif
+			{
+			struct stat st;
 			if (stat(b->data,&st) < 0)
 				break;
+			}
+#endif
 			/* found one. */
 			if (type == X509_LU_X509)
 				{
 				if ((X509_load_cert_file(xl,b->data,
-					ctx->dirs_type[i])) == 0)
+					ent->dir_type)) == 0)
 					break;
 				}
 			else if (type == X509_LU_CRL)
 				{
 				if ((X509_load_crl_file(xl,b->data,
-					ctx->dirs_type[i])) == 0)
+					ent->dir_type)) == 0)
 					break;
 				}
 			/* else case will caught higher up */
+			k++;
 			}
 
 		/* we have added it to the cache so now pull
@@ -366,6 +426,43 @@ static int get_cert_by_subject(X509_LOOKUP *xl, int type, X509_NAME *name,
 		else tmp = NULL;
 		CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
 
+
+		/* If a CRL, update the last file suffix added for this */
+
+		if (type == X509_LU_CRL)
+			{
+			CRYPTO_w_lock(CRYPTO_LOCK_X509_STORE);
+			/* Look for entry again in case another thread added
+			 * an entry first.
+			 */
+			if (!hent)
+				{
+				htmp.hash = h;
+				idx = sk_BY_DIR_HASH_find(ent->hashes, &htmp);
+				if (idx >= 0)
+					hent =
+					 sk_BY_DIR_HASH_value(ent->hashes, idx);
+				}
+			if (!hent)
+				{
+				hent = OPENSSL_malloc(sizeof(BY_DIR_HASH));
+				hent->hash = h;
+				hent->suffix = k;
+				if (!sk_BY_DIR_HASH_push(ent->hashes, hent))
+					{
+					CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+					OPENSSL_free(hent);
+					ok = 0;
+					goto finish;
+					}
+				}
+			else if (hent->suffix < k)
+				hent->suffix = k;
+
+			CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+
+			}
+
 		if (tmp != NULL)
 			{
 			ok=1;
diff --git a/crypto/x509/by_file.c b/crypto/x509/by_file.c
index a5e0d4aefa1a..57b08ee09458 100644
--- a/crypto/x509/by_file.c
+++ b/crypto/x509/by_file.c
@@ -100,7 +100,7 @@ static int by_file_ctrl(X509_LOOKUP *ctx, int cmd, const char *argp, long argl,
 	case X509_L_FILE_LOAD:
 		if (argl == X509_FILETYPE_DEFAULT)
 			{
-			file = (char *)Getenv(X509_get_default_cert_file_env());
+			file = (char *)getenv(X509_get_default_cert_file_env());
 			if (file)
 				ok = (X509_load_cert_crl_file(ctx,file,
 					      X509_FILETYPE_PEM) != 0);
diff --git a/crypto/x509/x509.h b/crypto/x509/x509.h
index c34689a1b389..092dd7450d3b 100644
--- a/crypto/x509/x509.h
+++ b/crypto/x509/x509.h
@@ -157,12 +157,12 @@ typedef struct X509_val_st
 	ASN1_TIME *notAfter;
 	} X509_VAL;
 
-typedef struct X509_pubkey_st
+struct X509_pubkey_st
 	{
 	X509_ALGOR *algor;
 	ASN1_BIT_STRING *public_key;
 	EVP_PKEY *pkey;
-	} X509_PUBKEY;
+	};
 
 typedef struct X509_sig_st
 	{
@@ -191,7 +191,9 @@ struct X509_name_st
 #else
 	char *bytes;
 #endif
-	unsigned long hash; /* Keep the hash around for lookups */
+/*	unsigned long hash; Keep the hash around for lookups */
+	unsigned char *canon_enc;
+	int canon_enclen;
 	} /* X509_NAME */;
 
 DECLARE_STACK_OF(X509_NAME)
@@ -291,8 +293,11 @@ struct x509_st
 	unsigned long ex_xkusage;
 	unsigned long ex_nscert;
 	ASN1_OCTET_STRING *skid;
-	struct AUTHORITY_KEYID_st *akid;
+	AUTHORITY_KEYID *akid;
 	X509_POLICY_CACHE *policy_cache;
+	STACK_OF(DIST_POINT) *crldp;
+	STACK_OF(GENERAL_NAME) *altname;
+	NAME_CONSTRAINTS *nc;
 #ifndef OPENSSL_NO_RFC3779
 	STACK_OF(IPAddressFamily) *rfc3779_addr;
 	struct ASIdentifiers_st *rfc3779_asid;
@@ -335,10 +340,11 @@ typedef struct x509_cert_pair_st {
 #define X509_TRUST_OBJECT_SIGN	5
 #define X509_TRUST_OCSP_SIGN	6
 #define X509_TRUST_OCSP_REQUEST	7
+#define X509_TRUST_TSA		8
 
 /* Keep these up to date! */
 #define X509_TRUST_MIN		1
-#define X509_TRUST_MAX		7
+#define X509_TRUST_MAX		8
 
 
 /* trust_flags values */
@@ -425,13 +431,17 @@ typedef struct x509_cert_pair_st {
 			XN_FLAG_FN_LN | \
 			XN_FLAG_FN_ALIGN)
 
-typedef struct X509_revoked_st
+struct x509_revoked_st
 	{
 	ASN1_INTEGER *serialNumber;
 	ASN1_TIME *revocationDate;
 	STACK_OF(X509_EXTENSION) /* optional */ *extensions;
+	/* Set up if indirect CRL */
+	STACK_OF(GENERAL_NAME) *issuer;
+	/* Revocation reason */
+	int reason;
 	int sequence; /* load sequence */
-	} X509_REVOKED;
+	};
 
 DECLARE_STACK_OF(X509_REVOKED)
 DECLARE_ASN1_SET_OF(X509_REVOKED)
@@ -455,6 +465,22 @@ struct X509_crl_st
 	X509_ALGOR *sig_alg;
 	ASN1_BIT_STRING *signature;
 	int references;
+	int flags;
+	/* Copies of various extensions */
+	AUTHORITY_KEYID *akid;
+	ISSUING_DIST_POINT *idp;
+	/* Convenient breakdown of IDP */
+	int idp_flags;
+	int idp_reasons;
+	/* CRL and base CRL numbers for delta processing */
+	ASN1_INTEGER *crl_number;
+	ASN1_INTEGER *base_crl_number;
+#ifndef OPENSSL_NO_SHA
+	unsigned char sha1_hash[SHA_DIGEST_LENGTH];
+#endif
+	STACK_OF(GENERAL_NAMES) *issuers;
+	const X509_CRL_METHOD *meth;
+	void *meth_data;
 	} /* X509_CRL */;
 
 DECLARE_STACK_OF(X509_CRL)
@@ -553,18 +579,19 @@ X509_ALGOR *prf;
 
 /* PKCS#8 private key info structure */
 
-typedef struct pkcs8_priv_key_info_st
+struct pkcs8_priv_key_info_st
         {
         int broken;     /* Flag for various broken formats */
 #define PKCS8_OK		0
 #define PKCS8_NO_OCTET		1
 #define PKCS8_EMBEDDED_PARAM	2
 #define PKCS8_NS_DB		3
+#define PKCS8_NEG_PRIVKEY	4
         ASN1_INTEGER *version;
         X509_ALGOR *pkeyalg;
         ASN1_TYPE *pkey; /* Should be OCTET STRING but some are broken */
         STACK_OF(X509_ATTRIBUTE) *attributes;
-        } PKCS8_PRIV_KEY_INFO;
+        };
 
 #ifdef  __cplusplus
 }
@@ -577,151 +604,6 @@ typedef struct pkcs8_priv_key_info_st
 extern "C" {
 #endif
 
-#ifdef SSLEAY_MACROS
-#define X509_verify(a,r) ASN1_verify((int (*)())i2d_X509_CINF,a->sig_alg,\
-	a->signature,(char *)a->cert_info,r)
-#define X509_REQ_verify(a,r) ASN1_verify((int (*)())i2d_X509_REQ_INFO, \
-	a->sig_alg,a->signature,(char *)a->req_info,r)
-#define X509_CRL_verify(a,r) ASN1_verify((int (*)())i2d_X509_CRL_INFO, \
-	a->sig_alg, a->signature,(char *)a->crl,r)
-
-#define X509_sign(x,pkey,md) \
-	ASN1_sign((int (*)())i2d_X509_CINF, x->cert_info->signature, \
-		x->sig_alg, x->signature, (char *)x->cert_info,pkey,md)
-#define X509_REQ_sign(x,pkey,md) \
-	ASN1_sign((int (*)())i2d_X509_REQ_INFO,x->sig_alg, NULL, \
-		x->signature, (char *)x->req_info,pkey,md)
-#define X509_CRL_sign(x,pkey,md) \
-	ASN1_sign((int (*)())i2d_X509_CRL_INFO,x->crl->sig_alg,x->sig_alg, \
-		x->signature, (char *)x->crl,pkey,md)
-#define NETSCAPE_SPKI_sign(x,pkey,md) \
-	ASN1_sign((int (*)())i2d_NETSCAPE_SPKAC, x->sig_algor,NULL, \
-		x->signature, (char *)x->spkac,pkey,md)
-
-#define X509_dup(x509) (X509 *)ASN1_dup((int (*)())i2d_X509, \
-		(char *(*)())d2i_X509,(char *)x509)
-#define X509_ATTRIBUTE_dup(xa) (X509_ATTRIBUTE *)ASN1_dup(\
-		(int (*)())i2d_X509_ATTRIBUTE, \
-		(char *(*)())d2i_X509_ATTRIBUTE,(char *)xa)
-#define X509_EXTENSION_dup(ex) (X509_EXTENSION *)ASN1_dup( \
-		(int (*)())i2d_X509_EXTENSION, \
-		(char *(*)())d2i_X509_EXTENSION,(char *)ex)
-#define d2i_X509_fp(fp,x509) (X509 *)ASN1_d2i_fp((char *(*)())X509_new, \
-		(char *(*)())d2i_X509, (fp),(unsigned char **)(x509))
-#define i2d_X509_fp(fp,x509) ASN1_i2d_fp(i2d_X509,fp,(unsigned char *)x509)
-#define d2i_X509_bio(bp,x509) (X509 *)ASN1_d2i_bio((char *(*)())X509_new, \
-		(char *(*)())d2i_X509, (bp),(unsigned char **)(x509))
-#define i2d_X509_bio(bp,x509) ASN1_i2d_bio(i2d_X509,bp,(unsigned char *)x509)
-
-#define X509_CRL_dup(crl) (X509_CRL *)ASN1_dup((int (*)())i2d_X509_CRL, \
-		(char *(*)())d2i_X509_CRL,(char *)crl)
-#define d2i_X509_CRL_fp(fp,crl) (X509_CRL *)ASN1_d2i_fp((char *(*)()) \
-		X509_CRL_new,(char *(*)())d2i_X509_CRL, (fp),\
-		(unsigned char **)(crl))
-#define i2d_X509_CRL_fp(fp,crl) ASN1_i2d_fp(i2d_X509_CRL,fp,\
-		(unsigned char *)crl)
-#define d2i_X509_CRL_bio(bp,crl) (X509_CRL *)ASN1_d2i_bio((char *(*)()) \
-		X509_CRL_new,(char *(*)())d2i_X509_CRL, (bp),\
-		(unsigned char **)(crl))
-#define i2d_X509_CRL_bio(bp,crl) ASN1_i2d_bio(i2d_X509_CRL,bp,\
-		(unsigned char *)crl)
-
-#define PKCS7_dup(p7) (PKCS7 *)ASN1_dup((int (*)())i2d_PKCS7, \
-		(char *(*)())d2i_PKCS7,(char *)p7)
-#define d2i_PKCS7_fp(fp,p7) (PKCS7 *)ASN1_d2i_fp((char *(*)()) \
-		PKCS7_new,(char *(*)())d2i_PKCS7, (fp),\
-		(unsigned char **)(p7))
-#define i2d_PKCS7_fp(fp,p7) ASN1_i2d_fp(i2d_PKCS7,fp,\
-		(unsigned char *)p7)
-#define d2i_PKCS7_bio(bp,p7) (PKCS7 *)ASN1_d2i_bio((char *(*)()) \
-		PKCS7_new,(char *(*)())d2i_PKCS7, (bp),\
-		(unsigned char **)(p7))
-#define i2d_PKCS7_bio(bp,p7) ASN1_i2d_bio(i2d_PKCS7,bp,\
-		(unsigned char *)p7)
-
-#define X509_REQ_dup(req) (X509_REQ *)ASN1_dup((int (*)())i2d_X509_REQ, \
-		(char *(*)())d2i_X509_REQ,(char *)req)
-#define d2i_X509_REQ_fp(fp,req) (X509_REQ *)ASN1_d2i_fp((char *(*)())\
-		X509_REQ_new, (char *(*)())d2i_X509_REQ, (fp),\
-		(unsigned char **)(req))
-#define i2d_X509_REQ_fp(fp,req) ASN1_i2d_fp(i2d_X509_REQ,fp,\
-		(unsigned char *)req)
-#define d2i_X509_REQ_bio(bp,req) (X509_REQ *)ASN1_d2i_bio((char *(*)())\
-		X509_REQ_new, (char *(*)())d2i_X509_REQ, (bp),\
-		(unsigned char **)(req))
-#define i2d_X509_REQ_bio(bp,req) ASN1_i2d_bio(i2d_X509_REQ,bp,\
-		(unsigned char *)req)
-
-#define RSAPublicKey_dup(rsa) (RSA *)ASN1_dup((int (*)())i2d_RSAPublicKey, \
-		(char *(*)())d2i_RSAPublicKey,(char *)rsa)
-#define RSAPrivateKey_dup(rsa) (RSA *)ASN1_dup((int (*)())i2d_RSAPrivateKey, \
-		(char *(*)())d2i_RSAPrivateKey,(char *)rsa)
-
-#define d2i_RSAPrivateKey_fp(fp,rsa) (RSA *)ASN1_d2i_fp((char *(*)())\
-		RSA_new,(char *(*)())d2i_RSAPrivateKey, (fp), \
-		(unsigned char **)(rsa))
-#define i2d_RSAPrivateKey_fp(fp,rsa) ASN1_i2d_fp(i2d_RSAPrivateKey,fp, \
-		(unsigned char *)rsa)
-#define d2i_RSAPrivateKey_bio(bp,rsa) (RSA *)ASN1_d2i_bio((char *(*)())\
-		RSA_new,(char *(*)())d2i_RSAPrivateKey, (bp), \
-		(unsigned char **)(rsa))
-#define i2d_RSAPrivateKey_bio(bp,rsa) ASN1_i2d_bio(i2d_RSAPrivateKey,bp, \
-		(unsigned char *)rsa)
-
-#define d2i_RSAPublicKey_fp(fp,rsa) (RSA *)ASN1_d2i_fp((char *(*)())\
-		RSA_new,(char *(*)())d2i_RSAPublicKey, (fp), \
-		(unsigned char **)(rsa))
-#define i2d_RSAPublicKey_fp(fp,rsa) ASN1_i2d_fp(i2d_RSAPublicKey,fp, \
-		(unsigned char *)rsa)
-#define d2i_RSAPublicKey_bio(bp,rsa) (RSA *)ASN1_d2i_bio((char *(*)())\
-		RSA_new,(char *(*)())d2i_RSAPublicKey, (bp), \
-		(unsigned char **)(rsa))
-#define i2d_RSAPublicKey_bio(bp,rsa) ASN1_i2d_bio(i2d_RSAPublicKey,bp, \
-		(unsigned char *)rsa)
-
-#define d2i_DSAPrivateKey_fp(fp,dsa) (DSA *)ASN1_d2i_fp((char *(*)())\
-		DSA_new,(char *(*)())d2i_DSAPrivateKey, (fp), \
-		(unsigned char **)(dsa))
-#define i2d_DSAPrivateKey_fp(fp,dsa) ASN1_i2d_fp(i2d_DSAPrivateKey,fp, \
-		(unsigned char *)dsa)
-#define d2i_DSAPrivateKey_bio(bp,dsa) (DSA *)ASN1_d2i_bio((char *(*)())\
-		DSA_new,(char *(*)())d2i_DSAPrivateKey, (bp), \
-		(unsigned char **)(dsa))
-#define i2d_DSAPrivateKey_bio(bp,dsa) ASN1_i2d_bio(i2d_DSAPrivateKey,bp, \
-		(unsigned char *)dsa)
-
-#define d2i_ECPrivateKey_fp(fp,ecdsa) (EC_KEY *)ASN1_d2i_fp((char *(*)())\
-		EC_KEY_new,(char *(*)())d2i_ECPrivateKey, (fp), \
-		(unsigned char **)(ecdsa))
-#define i2d_ECPrivateKey_fp(fp,ecdsa) ASN1_i2d_fp(i2d_ECPrivateKey,fp, \
-		(unsigned char *)ecdsa)
-#define d2i_ECPrivateKey_bio(bp,ecdsa) (EC_KEY *)ASN1_d2i_bio((char *(*)())\
-		EC_KEY_new,(char *(*)())d2i_ECPrivateKey, (bp), \
-		(unsigned char **)(ecdsa))
-#define i2d_ECPrivateKey_bio(bp,ecdsa) ASN1_i2d_bio(i2d_ECPrivateKey,bp, \
-		(unsigned char *)ecdsa)
-
-#define X509_ALGOR_dup(xn) (X509_ALGOR *)ASN1_dup((int (*)())i2d_X509_ALGOR,\
-		(char *(*)())d2i_X509_ALGOR,(char *)xn)
-
-#define X509_NAME_dup(xn) (X509_NAME *)ASN1_dup((int (*)())i2d_X509_NAME, \
-		(char *(*)())d2i_X509_NAME,(char *)xn)
-#define X509_NAME_ENTRY_dup(ne) (X509_NAME_ENTRY *)ASN1_dup( \
-		(int (*)())i2d_X509_NAME_ENTRY, \
-		(char *(*)())d2i_X509_NAME_ENTRY,\
-		(char *)ne)
-
-#define X509_digest(data,type,md,len) \
-	ASN1_digest((int (*)())i2d_X509,type,(char *)data,md,len)
-#define X509_NAME_digest(data,type,md,len) \
-	ASN1_digest((int (*)())i2d_X509_NAME,type,(char *)data,md,len)
-#ifndef PKCS7_ISSUER_AND_SERIAL_digest
-#define PKCS7_ISSUER_AND_SERIAL_digest(data,type,md,len) \
-	ASN1_digest((int (*)())i2d_PKCS7_ISSUER_AND_SERIAL,type,\
-		(char *)data,md,len)
-#endif
-#endif
-
 #define X509_EXT_PACK_UNKNOWN	1
 #define X509_EXT_PACK_STRING	2
 
@@ -742,6 +624,18 @@ extern "C" {
 #define		X509_CRL_get_issuer(x) ((x)->crl->issuer)
 #define		X509_CRL_get_REVOKED(x) ((x)->crl->revoked)
 
+void X509_CRL_set_default_method(const X509_CRL_METHOD *meth);
+X509_CRL_METHOD *X509_CRL_METHOD_new(
+	int (*crl_init)(X509_CRL *crl),
+	int (*crl_free)(X509_CRL *crl),
+	int (*crl_lookup)(X509_CRL *crl, X509_REVOKED **ret,
+				ASN1_INTEGER *ser, X509_NAME *issuer),
+	int (*crl_verify)(X509_CRL *crl, EVP_PKEY *pk));
+void X509_CRL_METHOD_free(X509_CRL_METHOD *m);
+
+void X509_CRL_set_meth_data(X509_CRL *crl, void *dat);
+void *X509_CRL_get_meth_data(X509_CRL *crl);
+
 /* This one is only used so that a binary form can output, as in
  * i2d_X509_NAME(X509_get_X509_PUBKEY(x),&buf) */
 #define 	X509_get_X509_PUBKEY(x) ((x)->cert_info->key)
@@ -749,7 +643,6 @@ extern "C" {
 
 const char *X509_verify_cert_error_string(long n);
 
-#ifndef SSLEAY_MACROS
 #ifndef OPENSSL_NO_EVP
 int X509_verify(X509 *a, EVP_PKEY *r);
 
@@ -764,11 +657,15 @@ int NETSCAPE_SPKI_set_pubkey(NETSCAPE_SPKI *x, EVP_PKEY *pkey);
 
 int NETSCAPE_SPKI_print(BIO *out, NETSCAPE_SPKI *spki);
 
+int X509_signature_dump(BIO *bp,const ASN1_STRING *sig, int indent);
 int X509_signature_print(BIO *bp,X509_ALGOR *alg, ASN1_STRING *sig);
 
 int X509_sign(X509 *x, EVP_PKEY *pkey, const EVP_MD *md);
+int X509_sign_ctx(X509 *x, EVP_MD_CTX *ctx);
 int X509_REQ_sign(X509_REQ *x, EVP_PKEY *pkey, const EVP_MD *md);
+int X509_REQ_sign_ctx(X509_REQ *x, EVP_MD_CTX *ctx);
 int X509_CRL_sign(X509_CRL *x, EVP_PKEY *pkey, const EVP_MD *md);
+int X509_CRL_sign_ctx(X509_CRL *x, EVP_MD_CTX *ctx);
 int NETSCAPE_SPKI_sign(NETSCAPE_SPKI *x, EVP_PKEY *pkey, const EVP_MD *md);
 
 int X509_pubkey_digest(const X509 *data,const EVP_MD *type,
@@ -870,15 +767,16 @@ X509_ALGOR *X509_ALGOR_dup(X509_ALGOR *xn);
 int X509_ALGOR_set0(X509_ALGOR *alg, ASN1_OBJECT *aobj, int ptype, void *pval);
 void X509_ALGOR_get0(ASN1_OBJECT **paobj, int *pptype, void **ppval,
 						X509_ALGOR *algor);
+void X509_ALGOR_set_md(X509_ALGOR *alg, const EVP_MD *md);
 
 X509_NAME *X509_NAME_dup(X509_NAME *xn);
 X509_NAME_ENTRY *X509_NAME_ENTRY_dup(X509_NAME_ENTRY *ne);
 
-#endif /* !SSLEAY_MACROS */
-
-int		X509_cmp_time(ASN1_TIME *s, time_t *t);
-int		X509_cmp_current_time(ASN1_TIME *s);
+int		X509_cmp_time(const ASN1_TIME *s, time_t *t);
+int		X509_cmp_current_time(const ASN1_TIME *s);
 ASN1_TIME *	X509_time_adj(ASN1_TIME *s, long adj, time_t *t);
+ASN1_TIME *	X509_time_adj_ex(ASN1_TIME *s,
+				int offset_day, long offset_sec, time_t *t);
 ASN1_TIME *	X509_gmtime_adj(ASN1_TIME *s, long adj);
 
 const char *	X509_get_default_cert_area(void );
@@ -966,6 +864,9 @@ DECLARE_ASN1_FUNCTIONS(X509_CRL_INFO)
 DECLARE_ASN1_FUNCTIONS(X509_CRL)
 
 int X509_CRL_add0_revoked(X509_CRL *crl, X509_REVOKED *rev);
+int X509_CRL_get0_by_serial(X509_CRL *crl,
+		X509_REVOKED **ret, ASN1_INTEGER *serial);
+int X509_CRL_get0_by_cert(X509_CRL *crl, X509_REVOKED **ret, X509 *x);
 
 X509_PKEY *	X509_PKEY_new(void );
 void		X509_PKEY_free(X509_PKEY *a);
@@ -1000,6 +901,9 @@ int ASN1_item_verify(const ASN1_ITEM *it, X509_ALGOR *algor1,
 int ASN1_item_sign(const ASN1_ITEM *it, X509_ALGOR *algor1, X509_ALGOR *algor2,
 	ASN1_BIT_STRING *signature,
 	void *data, EVP_PKEY *pkey, const EVP_MD *type);
+int ASN1_item_sign_ctx(const ASN1_ITEM *it,
+		X509_ALGOR *algor1, X509_ALGOR *algor2,
+	     	ASN1_BIT_STRING *signature, void *asn, EVP_MD_CTX *ctx);
 #endif
 
 int 		X509_set_version(X509 *x,long version);
@@ -1009,8 +913,8 @@ int 		X509_set_issuer_name(X509 *x, X509_NAME *name);
 X509_NAME *	X509_get_issuer_name(X509 *a);
 int 		X509_set_subject_name(X509 *x, X509_NAME *name);
 X509_NAME *	X509_get_subject_name(X509 *a);
-int 		X509_set_notBefore(X509 *x, ASN1_TIME *tm);
-int 		X509_set_notAfter(X509 *x, ASN1_TIME *tm);
+int 		X509_set_notBefore(X509 *x, const ASN1_TIME *tm);
+int 		X509_set_notAfter(X509 *x, const ASN1_TIME *tm);
 int 		X509_set_pubkey(X509 *x, EVP_PKEY *pkey);
 EVP_PKEY *	X509_get_pubkey(X509 *x);
 ASN1_BIT_STRING * X509_get0_pubkey_bitstr(const X509 *x);
@@ -1047,8 +951,8 @@ int X509_REQ_add1_attr_by_txt(X509_REQ *req,
 
 int X509_CRL_set_version(X509_CRL *x, long version);
 int X509_CRL_set_issuer_name(X509_CRL *x, X509_NAME *name);
-int X509_CRL_set_lastUpdate(X509_CRL *x, ASN1_TIME *tm);
-int X509_CRL_set_nextUpdate(X509_CRL *x, ASN1_TIME *tm);
+int X509_CRL_set_lastUpdate(X509_CRL *x, const ASN1_TIME *tm);
+int X509_CRL_set_nextUpdate(X509_CRL *x, const ASN1_TIME *tm);
 int X509_CRL_sort(X509_CRL *crl);
 
 int X509_REVOKED_set_serialNumber(X509_REVOKED *x, ASN1_INTEGER *serial);
@@ -1067,11 +971,18 @@ unsigned long	X509_issuer_name_hash(X509 *a);
 int		X509_subject_name_cmp(const X509 *a, const X509 *b);
 unsigned long	X509_subject_name_hash(X509 *x);
 
+#ifndef OPENSSL_NO_MD5
+unsigned long	X509_issuer_name_hash_old(X509 *a);
+unsigned long	X509_subject_name_hash_old(X509 *x);
+#endif
+
 int		X509_cmp(const X509 *a, const X509 *b);
 int		X509_NAME_cmp(const X509_NAME *a, const X509_NAME *b);
 unsigned long	X509_NAME_hash(X509_NAME *x);
+unsigned long	X509_NAME_hash_old(X509_NAME *x);
 
 int		X509_CRL_cmp(const X509_CRL *a, const X509_CRL *b);
+int		X509_CRL_match(const X509_CRL *a, const X509_CRL *b);
 #ifndef OPENSSL_NO_FP_API
 int		X509_print_ex_fp(FILE *bp,X509 *x, unsigned long nmflag, unsigned long cflag);
 int		X509_print_fp(FILE *bp,X509 *x);
@@ -1247,9 +1158,19 @@ DECLARE_ASN1_FUNCTIONS(PBEPARAM)
 DECLARE_ASN1_FUNCTIONS(PBE2PARAM)
 DECLARE_ASN1_FUNCTIONS(PBKDF2PARAM)
 
-X509_ALGOR *PKCS5_pbe_set(int alg, int iter, unsigned char *salt, int saltlen);
+int PKCS5_pbe_set0_algor(X509_ALGOR *algor, int alg, int iter,
+				const unsigned char *salt, int saltlen);
+
+X509_ALGOR *PKCS5_pbe_set(int alg, int iter,
+				const unsigned char *salt, int saltlen);
 X509_ALGOR *PKCS5_pbe2_set(const EVP_CIPHER *cipher, int iter,
 					 unsigned char *salt, int saltlen);
+X509_ALGOR *PKCS5_pbe2_set_iv(const EVP_CIPHER *cipher, int iter,
+				 unsigned char *salt, int saltlen,
+				 unsigned char *aiv, int prf_nid);
+
+X509_ALGOR *PKCS5_pbkdf2_set(int iter, unsigned char *salt, int saltlen,
+				int prf_nid, int keylen);
 
 /* PKCS#8 utilities */
 
@@ -1260,6 +1181,22 @@ PKCS8_PRIV_KEY_INFO *EVP_PKEY2PKCS8(EVP_PKEY *pkey);
 PKCS8_PRIV_KEY_INFO *EVP_PKEY2PKCS8_broken(EVP_PKEY *pkey, int broken);
 PKCS8_PRIV_KEY_INFO *PKCS8_set_broken(PKCS8_PRIV_KEY_INFO *p8, int broken);
 
+int PKCS8_pkey_set0(PKCS8_PRIV_KEY_INFO *priv, ASN1_OBJECT *aobj,
+			int version, int ptype, void *pval,
+				unsigned char *penc, int penclen);
+int PKCS8_pkey_get0(ASN1_OBJECT **ppkalg,
+		const unsigned char **pk, int *ppklen,
+		X509_ALGOR **pa,
+		PKCS8_PRIV_KEY_INFO *p8);
+
+int X509_PUBKEY_set0_param(X509_PUBKEY *pub, ASN1_OBJECT *aobj,
+					int ptype, void *pval,
+					unsigned char *penc, int penclen);
+int X509_PUBKEY_get0_param(ASN1_OBJECT **ppkalg,
+		const unsigned char **pk, int *ppklen,
+		X509_ALGOR **pa,
+		X509_PUBKEY *pub);
+
 int X509_check_trust(X509 *x, int id, int flags);
 int X509_TRUST_get_count(void);
 X509_TRUST * X509_TRUST_get0(int idx);
@@ -1339,7 +1276,10 @@ void ERR_load_X509_strings(void);
 #define X509_R_KEY_VALUES_MISMATCH			 116
 #define X509_R_LOADING_CERT_DIR				 103
 #define X509_R_LOADING_DEFAULTS				 104
+#define X509_R_METHOD_NOT_SUPPORTED			 124
 #define X509_R_NO_CERT_SET_FOR_US_TO_VERIFY		 105
+#define X509_R_PUBLIC_KEY_DECODE_ERROR			 125
+#define X509_R_PUBLIC_KEY_ENCODE_ERROR			 126
 #define X509_R_SHOULD_RETRY				 106
 #define X509_R_UNABLE_TO_FIND_PARAMETERS_IN_CHAIN	 107
 #define X509_R_UNABLE_TO_GET_CERTS_PUBLIC_KEY		 108
diff --git a/crypto/x509/x509_cmp.c b/crypto/x509/x509_cmp.c
index 2faf92514a25..7c2aaee2e9a2 100644
--- a/crypto/x509/x509_cmp.c
+++ b/crypto/x509/x509_cmp.c
@@ -87,15 +87,20 @@ unsigned long X509_issuer_and_serial_hash(X509 *a)
 	EVP_MD_CTX_init(&ctx);
 	f=X509_NAME_oneline(a->cert_info->issuer,NULL,0);
 	ret=strlen(f);
-	EVP_DigestInit_ex(&ctx, EVP_md5(), NULL);
-	EVP_DigestUpdate(&ctx,(unsigned char *)f,ret);
+	if (!EVP_DigestInit_ex(&ctx, EVP_md5(), NULL))
+		goto err;
+	if (!EVP_DigestUpdate(&ctx,(unsigned char *)f,ret))
+		goto err;
 	OPENSSL_free(f);
-	EVP_DigestUpdate(&ctx,(unsigned char *)a->cert_info->serialNumber->data,
-		(unsigned long)a->cert_info->serialNumber->length);
-	EVP_DigestFinal_ex(&ctx,&(md[0]),NULL);
+	if(!EVP_DigestUpdate(&ctx,(unsigned char *)a->cert_info->serialNumber->data,
+		(unsigned long)a->cert_info->serialNumber->length))
+		goto err;
+	if (!EVP_DigestFinal_ex(&ctx,&(md[0]),NULL))
+		goto err;
 	ret=(	((unsigned long)md[0]     )|((unsigned long)md[1]<<8L)|
 		((unsigned long)md[2]<<16L)|((unsigned long)md[3]<<24L)
 		)&0xffffffffL;
+	err:
 	EVP_MD_CTX_cleanup(&ctx);
 	return(ret);
 	}
@@ -116,6 +121,13 @@ int X509_CRL_cmp(const X509_CRL *a, const X509_CRL *b)
 	return(X509_NAME_cmp(a->crl->issuer,b->crl->issuer));
 	}
 
+#ifndef OPENSSL_NO_SHA
+int X509_CRL_match(const X509_CRL *a, const X509_CRL *b)
+	{
+	return memcmp(a->sha1_hash, b->sha1_hash, 20);
+	}
+#endif
+
 X509_NAME *X509_get_issuer_name(X509 *a)
 	{
 	return(a->cert_info->issuer);
@@ -126,6 +138,13 @@ unsigned long X509_issuer_name_hash(X509 *x)
 	return(X509_NAME_hash(x->cert_info->issuer));
 	}
 
+#ifndef OPENSSL_NO_MD5
+unsigned long X509_issuer_name_hash_old(X509 *x)
+	{
+	return(X509_NAME_hash_old(x->cert_info->issuer));
+	}
+#endif
+
 X509_NAME *X509_get_subject_name(X509 *a)
 	{
 	return(a->cert_info->subject);
@@ -141,6 +160,13 @@ unsigned long X509_subject_name_hash(X509 *x)
 	return(X509_NAME_hash(x->cert_info->subject));
 	}
 
+#ifndef OPENSSL_NO_MD5
+unsigned long X509_subject_name_hash_old(X509 *x)
+	{
+	return(X509_NAME_hash_old(x->cert_info->subject));
+	}
+#endif
+
 #ifndef OPENSSL_NO_SHA
 /* Compare two certificates: they must be identical for
  * this to work. NB: Although "cmp" operations are generally
@@ -162,168 +188,62 @@ int X509_cmp(const X509 *a, const X509 *b)
 #endif
 
 
-/* Case insensitive string comparision */
-static int nocase_cmp(const ASN1_STRING *a, const ASN1_STRING *b)
-{
-	int i;
-
-	if (a->length != b->length)
-		return (a->length - b->length);
-
-	for (i=0; i<a->length; i++)
-	{
-		int ca, cb;
-
-		ca = tolower(a->data[i]);
-		cb = tolower(b->data[i]);
-
-		if (ca != cb)
-			return(ca-cb);
-	}
-	return 0;
-}
-
-/* Case insensitive string comparision with space normalization 
- * Space normalization - ignore leading, trailing spaces, 
- *       multiple spaces between characters are replaced by single space  
- */
-static int nocase_spacenorm_cmp(const ASN1_STRING *a, const ASN1_STRING *b)
-{
-	unsigned char *pa = NULL, *pb = NULL;
-	int la, lb;
-	
-	la = a->length;
-	lb = b->length;
-	pa = a->data;
-	pb = b->data;
-
-	/* skip leading spaces */
-	while (la > 0 && isspace(*pa))
-	{
-		la--;
-		pa++;
-	}
-	while (lb > 0 && isspace(*pb))
-	{
-		lb--;
-		pb++;
-	}
-
-	/* skip trailing spaces */
-	while (la > 0 && isspace(pa[la-1]))
-		la--;
-	while (lb > 0 && isspace(pb[lb-1]))
-		lb--;
-
-	/* compare strings with space normalization */
-	while (la > 0 && lb > 0)
+int X509_NAME_cmp(const X509_NAME *a, const X509_NAME *b)
 	{
-		int ca, cb;
-
-		/* compare character */
-		ca = tolower(*pa);
-		cb = tolower(*pb);
-		if (ca != cb)
-			return (ca - cb);
+	int ret;
 
-		pa++; pb++;
-		la--; lb--;
+	/* Ensure canonical encoding is present and up to date */
 
-		if (la <= 0 || lb <= 0)
-			break;
+	if (!a->canon_enc || a->modified)
+		{
+		ret = i2d_X509_NAME((X509_NAME *)a, NULL);
+		if (ret < 0)
+			return -2;
+		}
 
-		/* is white space next character ? */
-		if (isspace(*pa) && isspace(*pb))
+	if (!b->canon_enc || b->modified)
 		{
-			/* skip remaining white spaces */
-			while (la > 0 && isspace(*pa))
-			{
-				la--;
-				pa++;
-			}
-			while (lb > 0 && isspace(*pb))
-			{
-				lb--;
-				pb++;
-			}
+		ret = i2d_X509_NAME((X509_NAME *)b, NULL);
+		if (ret < 0)
+			return -2;
 		}
-	}
-	if (la > 0 || lb > 0)
-		return la - lb;
 
-	return 0;
-}
+	ret = a->canon_enclen - b->canon_enclen;
 
-static int asn1_string_memcmp(ASN1_STRING *a, ASN1_STRING *b)
-	{
-	int j;
-	j = a->length - b->length;
-	if (j)
-		return j;
-	return memcmp(a->data, b->data, a->length);
-	}
+	if (ret)
+		return ret;
 
-#define STR_TYPE_CMP (B_ASN1_PRINTABLESTRING|B_ASN1_T61STRING|B_ASN1_UTF8STRING)
+	return memcmp(a->canon_enc, b->canon_enc, a->canon_enclen);
 
-int X509_NAME_cmp(const X509_NAME *a, const X509_NAME *b)
-	{
-	int i,j;
-	X509_NAME_ENTRY *na,*nb;
+	}
 
-	unsigned long nabit, nbbit;
+unsigned long X509_NAME_hash(X509_NAME *x)
+	{
+	unsigned long ret=0;
+	unsigned char md[SHA_DIGEST_LENGTH];
 
-	j = sk_X509_NAME_ENTRY_num(a->entries)
-		  - sk_X509_NAME_ENTRY_num(b->entries);
-	if (j)
-		return j;
-	for (i=sk_X509_NAME_ENTRY_num(a->entries)-1; i>=0; i--)
-		{
-		na=sk_X509_NAME_ENTRY_value(a->entries,i);
-		nb=sk_X509_NAME_ENTRY_value(b->entries,i);
-		j=na->value->type-nb->value->type;
-		if (j)
-			{
-			nabit = ASN1_tag2bit(na->value->type);
-			nbbit = ASN1_tag2bit(nb->value->type);
-			if (!(nabit & STR_TYPE_CMP) ||
-				!(nbbit & STR_TYPE_CMP))
-				return j;
-			if (!asn1_string_memcmp(na->value, nb->value))
-				j = 0;
-			}
-		else if (na->value->type == V_ASN1_PRINTABLESTRING)
-			j=nocase_spacenorm_cmp(na->value, nb->value);
-		else if (na->value->type == V_ASN1_IA5STRING
-			&& OBJ_obj2nid(na->object) == NID_pkcs9_emailAddress)
-			j=nocase_cmp(na->value, nb->value);
-		else
-			j = asn1_string_memcmp(na->value, nb->value);
-		if (j) return(j);
-		j=na->set-nb->set;
-		if (j) return(j);
-		}
+	/* Make sure X509_NAME structure contains valid cached encoding */
+	i2d_X509_NAME(x,NULL);
+	if (!EVP_Digest(x->canon_enc, x->canon_enclen, md, NULL, EVP_sha1(),
+		NULL))
+		return 0;
 
-	/* We will check the object types after checking the values
-	 * since the values will more often be different than the object
-	 * types. */
-	for (i=sk_X509_NAME_ENTRY_num(a->entries)-1; i>=0; i--)
-		{
-		na=sk_X509_NAME_ENTRY_value(a->entries,i);
-		nb=sk_X509_NAME_ENTRY_value(b->entries,i);
-		j=OBJ_cmp(na->object,nb->object);
-		if (j) return(j);
-		}
-	return(0);
+	ret=(	((unsigned long)md[0]     )|((unsigned long)md[1]<<8L)|
+		((unsigned long)md[2]<<16L)|((unsigned long)md[3]<<24L)
+		)&0xffffffffL;
+	return(ret);
 	}
 
+
 #ifndef OPENSSL_NO_MD5
 /* I now DER encode the name and hash it.  Since I cache the DER encoding,
  * this is reasonably efficient. */
-unsigned long X509_NAME_hash(X509_NAME *x)
+
+unsigned long X509_NAME_hash_old(X509_NAME *x)
 	{
+	EVP_MD_CTX md_ctx;
 	unsigned long ret=0;
 	unsigned char md[16];
-	EVP_MD_CTX md_ctx;
 
 	/* Make sure X509_NAME structure contains valid cached encoding */
 	i2d_X509_NAME(x,NULL);
@@ -393,14 +313,19 @@ ASN1_BIT_STRING *X509_get0_pubkey_bitstr(const X509 *x)
 
 int X509_check_private_key(X509 *x, EVP_PKEY *k)
 	{
-	EVP_PKEY *xk=NULL;
-	int ok=0;
+	EVP_PKEY *xk;
+	int ret;
 
 	xk=X509_get_pubkey(x);
-	switch (EVP_PKEY_cmp(xk, k))
+
+	if (xk)
+		ret = EVP_PKEY_cmp(xk, k);
+	else
+		ret = -2;
+
+	switch (ret)
 		{
 	case 1:
-		ok=1;
 		break;
 	case 0:
 		X509err(X509_F_X509_CHECK_PRIVATE_KEY,X509_R_KEY_VALUES_MISMATCH);
@@ -409,24 +334,11 @@ int X509_check_private_key(X509 *x, EVP_PKEY *k)
 		X509err(X509_F_X509_CHECK_PRIVATE_KEY,X509_R_KEY_TYPE_MISMATCH);
 		break;
 	case -2:
-#ifndef OPENSSL_NO_EC
-		if (k->type == EVP_PKEY_EC)
-			{
-			X509err(X509_F_X509_CHECK_PRIVATE_KEY, ERR_R_EC_LIB);
-			break;
-			}
-#endif
-#ifndef OPENSSL_NO_DH
-		if (k->type == EVP_PKEY_DH)
-			{
-			/* No idea */
-			X509err(X509_F_X509_CHECK_PRIVATE_KEY,X509_R_CANT_CHECK_DH_KEY);
-			break;
-			}
-#endif
 	        X509err(X509_F_X509_CHECK_PRIVATE_KEY,X509_R_UNKNOWN_KEY_TYPE);
 		}
-
-	EVP_PKEY_free(xk);
-	return(ok);
+	if (xk)
+		EVP_PKEY_free(xk);
+	if (ret > 0)
+		return 1;
+	return 0;
 	}
diff --git a/crypto/x509/x509_err.c b/crypto/x509/x509_err.c
index fb377292dac5..a01402f41673 100644
--- a/crypto/x509/x509_err.c
+++ b/crypto/x509/x509_err.c
@@ -1,6 +1,6 @@
 /* crypto/x509/x509_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2006 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -132,7 +132,10 @@ static ERR_STRING_DATA X509_str_reasons[]=
 {ERR_REASON(X509_R_KEY_VALUES_MISMATCH)  ,"key values mismatch"},
 {ERR_REASON(X509_R_LOADING_CERT_DIR)     ,"loading cert dir"},
 {ERR_REASON(X509_R_LOADING_DEFAULTS)     ,"loading defaults"},
+{ERR_REASON(X509_R_METHOD_NOT_SUPPORTED) ,"method not supported"},
 {ERR_REASON(X509_R_NO_CERT_SET_FOR_US_TO_VERIFY),"no cert set for us to verify"},
+{ERR_REASON(X509_R_PUBLIC_KEY_DECODE_ERROR),"public key decode error"},
+{ERR_REASON(X509_R_PUBLIC_KEY_ENCODE_ERROR),"public key encode error"},
 {ERR_REASON(X509_R_SHOULD_RETRY)         ,"should retry"},
 {ERR_REASON(X509_R_UNABLE_TO_FIND_PARAMETERS_IN_CHAIN),"unable to find parameters in chain"},
 {ERR_REASON(X509_R_UNABLE_TO_GET_CERTS_PUBLIC_KEY),"unable to get certs public key"},
diff --git a/crypto/x509/x509_lu.c b/crypto/x509/x509_lu.c
index b486171868a6..38525a8cddbc 100644
--- a/crypto/x509/x509_lu.c
+++ b/crypto/x509/x509_lu.c
@@ -87,7 +87,7 @@ void X509_LOOKUP_free(X509_LOOKUP *ctx)
 	if (ctx == NULL) return;
 	if (	(ctx->method != NULL) &&
 		(ctx->method->free != NULL))
-		ctx->method->free(ctx);
+		(*ctx->method->free)(ctx);
 	OPENSSL_free(ctx);
 	}
 
@@ -196,6 +196,8 @@ X509_STORE *X509_STORE_new(void)
 	ret->get_crl = 0;
 	ret->check_crl = 0;
 	ret->cert_crl = 0;
+	ret->lookup_certs = 0;
+	ret->lookup_crls = 0;
 	ret->cleanup = 0;
 
 	if (!CRYPTO_new_ex_data(CRYPTO_EX_INDEX_X509_STORE, ret, &ret->ex_data))
@@ -296,7 +298,7 @@ int X509_STORE_get_by_subject(X509_STORE_CTX *vs, int type, X509_NAME *name,
 	tmp=X509_OBJECT_retrieve_by_subject(ctx->objs,type,name);
 	CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
 
-	if (tmp == NULL)
+	if (tmp == NULL || type == X509_LU_CRL)
 		{
 		for (i=vs->current_method; i<sk_X509_LOOKUP_num(ctx->get_cert_methods); i++)
 			{
@@ -421,14 +423,15 @@ void X509_OBJECT_free_contents(X509_OBJECT *a)
 		}
 	}
 
-int X509_OBJECT_idx_by_subject(STACK_OF(X509_OBJECT) *h, int type,
-	     X509_NAME *name)
+static int x509_object_idx_cnt(STACK_OF(X509_OBJECT) *h, int type,
+	     X509_NAME *name, int *pnmatch)
 	{
 	X509_OBJECT stmp;
 	X509 x509_s;
 	X509_CINF cinf_s;
 	X509_CRL crl_s;
 	X509_CRL_INFO crl_info_s;
+	int idx;
 
 	stmp.type=type;
 	switch (type)
@@ -448,7 +451,29 @@ int X509_OBJECT_idx_by_subject(STACK_OF(X509_OBJECT) *h, int type,
 		return -1;
 		}
 
-	return sk_X509_OBJECT_find(h,&stmp);
+	idx = sk_X509_OBJECT_find(h,&stmp);
+	if (idx >= 0 && pnmatch)
+		{
+		int tidx;
+		const X509_OBJECT *tobj, *pstmp;
+		*pnmatch = 1;
+		pstmp = &stmp;
+		for (tidx = idx + 1; tidx < sk_X509_OBJECT_num(h); tidx++)
+			{
+			tobj = sk_X509_OBJECT_value(h, tidx);
+			if (x509_object_cmp(&tobj, &pstmp))
+				break;
+			(*pnmatch)++;
+			}
+		}
+	return idx;
+	}
+
+
+int X509_OBJECT_idx_by_subject(STACK_OF(X509_OBJECT) *h, int type,
+	     X509_NAME *name)
+	{
+	return x509_object_idx_cnt(h, type, name, NULL);
 	}
 
 X509_OBJECT *X509_OBJECT_retrieve_by_subject(STACK_OF(X509_OBJECT) *h, int type,
@@ -460,19 +485,125 @@ X509_OBJECT *X509_OBJECT_retrieve_by_subject(STACK_OF(X509_OBJECT) *h, int type,
 	return sk_X509_OBJECT_value(h, idx);
 	}
 
+STACK_OF(X509)* X509_STORE_get1_certs(X509_STORE_CTX *ctx, X509_NAME *nm)
+	{
+	int i, idx, cnt;
+	STACK_OF(X509) *sk;
+	X509 *x;
+	X509_OBJECT *obj;
+	sk = sk_X509_new_null();
+	CRYPTO_w_lock(CRYPTO_LOCK_X509_STORE);
+	idx = x509_object_idx_cnt(ctx->ctx->objs, X509_LU_X509, nm, &cnt);
+	if (idx < 0)
+		{
+		/* Nothing found in cache: do lookup to possibly add new
+		 * objects to cache
+		 */
+		X509_OBJECT xobj;
+		CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+		if (!X509_STORE_get_by_subject(ctx, X509_LU_X509, nm, &xobj))
+			{
+			sk_X509_free(sk);
+			return NULL;
+			}
+		X509_OBJECT_free_contents(&xobj);
+		CRYPTO_w_lock(CRYPTO_LOCK_X509_STORE);
+		idx = x509_object_idx_cnt(ctx->ctx->objs,X509_LU_X509,nm, &cnt);
+		if (idx < 0)
+			{
+			CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+			sk_X509_free(sk);
+			return NULL;
+			}
+		}
+	for (i = 0; i < cnt; i++, idx++)
+		{
+		obj = sk_X509_OBJECT_value(ctx->ctx->objs, idx);
+		x = obj->data.x509;
+		CRYPTO_add(&x->references, 1, CRYPTO_LOCK_X509);
+		if (!sk_X509_push(sk, x))
+			{
+			CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+			X509_free(x);
+			sk_X509_pop_free(sk, X509_free);
+			return NULL;
+			}
+		}
+	CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+	return sk;
+
+	}
+
+STACK_OF(X509_CRL)* X509_STORE_get1_crls(X509_STORE_CTX *ctx, X509_NAME *nm)
+	{
+	int i, idx, cnt;
+	STACK_OF(X509_CRL) *sk;
+	X509_CRL *x;
+	X509_OBJECT *obj, xobj;
+	sk = sk_X509_CRL_new_null();
+	CRYPTO_w_lock(CRYPTO_LOCK_X509_STORE);
+	/* Check cache first */
+	idx = x509_object_idx_cnt(ctx->ctx->objs, X509_LU_CRL, nm, &cnt);
+
+	/* Always do lookup to possibly add new CRLs to cache
+	 */
+	CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+	if (!X509_STORE_get_by_subject(ctx, X509_LU_CRL, nm, &xobj))
+		{
+		sk_X509_CRL_free(sk);
+		return NULL;
+		}
+	X509_OBJECT_free_contents(&xobj);
+	CRYPTO_w_lock(CRYPTO_LOCK_X509_STORE);
+	idx = x509_object_idx_cnt(ctx->ctx->objs,X509_LU_CRL, nm, &cnt);
+	if (idx < 0)
+		{
+		CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+		sk_X509_CRL_free(sk);
+		return NULL;
+		}
+
+	for (i = 0; i < cnt; i++, idx++)
+		{
+		obj = sk_X509_OBJECT_value(ctx->ctx->objs, idx);
+		x = obj->data.crl;
+		CRYPTO_add(&x->references, 1, CRYPTO_LOCK_X509_CRL);
+		if (!sk_X509_CRL_push(sk, x))
+			{
+			CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+			X509_CRL_free(x);
+			sk_X509_CRL_pop_free(sk, X509_CRL_free);
+			return NULL;
+			}
+		}
+	CRYPTO_w_unlock(CRYPTO_LOCK_X509_STORE);
+	return sk;
+	}
+
 X509_OBJECT *X509_OBJECT_retrieve_match(STACK_OF(X509_OBJECT) *h, X509_OBJECT *x)
 	{
 	int idx, i;
 	X509_OBJECT *obj;
 	idx = sk_X509_OBJECT_find(h, x);
 	if (idx == -1) return NULL;
-	if (x->type != X509_LU_X509) return sk_X509_OBJECT_value(h, idx);
+	if ((x->type != X509_LU_X509) && (x->type != X509_LU_CRL))
+		return sk_X509_OBJECT_value(h, idx);
 	for (i = idx; i < sk_X509_OBJECT_num(h); i++)
 		{
 		obj = sk_X509_OBJECT_value(h, i);
 		if (x509_object_cmp((const X509_OBJECT **)&obj, (const X509_OBJECT **)&x))
 			return NULL;
-		if ((x->type != X509_LU_X509) || !X509_cmp(obj->data.x509, x->data.x509))
+		if (x->type == X509_LU_X509)
+			{
+			if (!X509_cmp(obj->data.x509, x->data.x509))
+				return obj;
+			}
+		else if (x->type == X509_LU_CRL)
+			{
+			if (!X509_CRL_match(obj->data.crl, x->data.crl))
+				return obj;
+			}
+		else
 			return obj;
 		}
 	return NULL;
@@ -575,5 +706,11 @@ int X509_STORE_set1_param(X509_STORE *ctx, X509_VERIFY_PARAM *param)
 	return X509_VERIFY_PARAM_set1(ctx->param, param);
 	}
 
+void X509_STORE_set_verify_cb(X509_STORE *ctx,
+				  int (*verify_cb)(int, X509_STORE_CTX *))
+	{
+	ctx->verify_cb = verify_cb;
+	}
+
 IMPLEMENT_STACK_OF(X509_LOOKUP)
 IMPLEMENT_STACK_OF(X509_OBJECT)
diff --git a/crypto/x509/x509_obj.c b/crypto/x509/x509_obj.c
index 1e718f76eb2d..21fed9f838d6 100644
--- a/crypto/x509/x509_obj.c
+++ b/crypto/x509/x509_obj.c
@@ -72,7 +72,7 @@ int i;
 	char *p;
 	unsigned char *q;
 	BUF_MEM *b=NULL;
-	static char hex[17]="0123456789ABCDEF";
+	static const char hex[17]="0123456789ABCDEF";
 	int gs_doit[4];
 	char tmp_buf[80];
 #ifdef CHARSET_EBCDIC
diff --git a/crypto/x509/x509_req.c b/crypto/x509/x509_req.c
index 3872e1fb64a5..48183dc00cb8 100644
--- a/crypto/x509/x509_req.c
+++ b/crypto/x509/x509_req.c
@@ -61,6 +61,7 @@
 #include <openssl/bn.h>
 #include <openssl/evp.h>
 #include <openssl/asn1.h>
+#include <openssl/asn1t.h>
 #include <openssl/x509.h>
 #include <openssl/objects.h>
 #include <openssl/buffer.h>
@@ -205,10 +206,9 @@ STACK_OF(X509_EXTENSION) *X509_REQ_get_extensions(X509_REQ *req)
 	if(!ext || (ext->type != V_ASN1_SEQUENCE))
 		return NULL;
 	p = ext->value.sequence->data;
-	return d2i_ASN1_SET_OF_X509_EXTENSION(NULL, &p,
-			ext->value.sequence->length,
-			d2i_X509_EXTENSION, X509_EXTENSION_free,
-			V_ASN1_SEQUENCE, V_ASN1_UNIVERSAL);
+	return (STACK_OF(X509_EXTENSION) *)
+		ASN1_item_d2i(NULL, &p, ext->value.sequence->length,
+				ASN1_ITEM_rptr(X509_EXTENSIONS));
 }
 
 /* Add a STACK_OF extensions to a certificate request: allow alternative OIDs
@@ -218,8 +218,6 @@ STACK_OF(X509_EXTENSION) *X509_REQ_get_extensions(X509_REQ *req)
 int X509_REQ_add_extensions_nid(X509_REQ *req, STACK_OF(X509_EXTENSION) *exts,
 				int nid)
 {
-	unsigned char *p = NULL, *q;
-	long len;
 	ASN1_TYPE *at = NULL;
 	X509_ATTRIBUTE *attr = NULL;
 	if(!(at = ASN1_TYPE_new()) ||
@@ -227,15 +225,10 @@ int X509_REQ_add_extensions_nid(X509_REQ *req, STACK_OF(X509_EXTENSION) *exts,
 
 	at->type = V_ASN1_SEQUENCE;
 	/* Generate encoding of extensions */
-	len = i2d_ASN1_SET_OF_X509_EXTENSION(exts, NULL, i2d_X509_EXTENSION,
-			V_ASN1_SEQUENCE, V_ASN1_UNIVERSAL, IS_SEQUENCE);
-	if(!(p = OPENSSL_malloc(len))) goto err;
-	q = p;
-	i2d_ASN1_SET_OF_X509_EXTENSION(exts, &q, i2d_X509_EXTENSION,
-			V_ASN1_SEQUENCE, V_ASN1_UNIVERSAL, IS_SEQUENCE);
-	at->value.sequence->data = p;
-	p = NULL;
-	at->value.sequence->length = len;
+	at->value.sequence->length = 
+			ASN1_item_i2d((ASN1_VALUE *)exts,
+				&at->value.sequence->data,
+				ASN1_ITEM_rptr(X509_EXTENSIONS));
 	if(!(attr = X509_ATTRIBUTE_new())) goto err;
 	if(!(attr->value.set = sk_ASN1_TYPE_new_null())) goto err;
 	if(!sk_ASN1_TYPE_push(attr->value.set, at)) goto err;
@@ -250,7 +243,6 @@ int X509_REQ_add_extensions_nid(X509_REQ *req, STACK_OF(X509_EXTENSION) *exts,
 	if(!sk_X509_ATTRIBUTE_push(req->req_info->attributes, attr)) goto err;
 	return 1;
 	err:
-	if(p) OPENSSL_free(p);
 	X509_ATTRIBUTE_free(attr);
 	ASN1_TYPE_free(at);
 	return 0;
diff --git a/crypto/x509/x509_set.c b/crypto/x509/x509_set.c
index aaf61ca062b9..4b94fc58477f 100644
--- a/crypto/x509/x509_set.c
+++ b/crypto/x509/x509_set.c
@@ -104,7 +104,7 @@ int X509_set_subject_name(X509 *x, X509_NAME *name)
 	return(X509_NAME_set(&x->cert_info->subject,name));
 	}
 
-int X509_set_notBefore(X509 *x, ASN1_TIME *tm)
+int X509_set_notBefore(X509 *x, const ASN1_TIME *tm)
 	{
 	ASN1_TIME *in;
 
@@ -122,7 +122,7 @@ int X509_set_notBefore(X509 *x, ASN1_TIME *tm)
 	return(in != NULL);
 	}
 
-int X509_set_notAfter(X509 *x, ASN1_TIME *tm)
+int X509_set_notAfter(X509 *x, const ASN1_TIME *tm)
 	{
 	ASN1_TIME *in;
 
diff --git a/crypto/x509/x509_trs.c b/crypto/x509/x509_trs.c
index ed1870058591..a6cb9c8b1b37 100644
--- a/crypto/x509/x509_trs.c
+++ b/crypto/x509/x509_trs.c
@@ -84,7 +84,8 @@ static X509_TRUST trstandard[] = {
 {X509_TRUST_EMAIL, 0, trust_1oidany, "S/MIME email", NID_email_protect, NULL},
 {X509_TRUST_OBJECT_SIGN, 0, trust_1oidany, "Object Signer", NID_code_sign, NULL},
 {X509_TRUST_OCSP_SIGN, 0, trust_1oid, "OCSP responder", NID_OCSP_sign, NULL},
-{X509_TRUST_OCSP_REQUEST, 0, trust_1oid, "OCSP request", NID_ad_OCSP, NULL}
+{X509_TRUST_OCSP_REQUEST, 0, trust_1oid, "OCSP request", NID_ad_OCSP, NULL},
+{X509_TRUST_TSA, 0, trust_1oidany, "TSA server", NID_time_stamp, NULL}
 };
 
 #define X509_TRUST_COUNT	(sizeof(trstandard)/sizeof(X509_TRUST))
diff --git a/crypto/x509/x509_txt.c b/crypto/x509/x509_txt.c
index 73a8ec726fee..c44f753c462c 100644
--- a/crypto/x509/x509_txt.c
+++ b/crypto/x509/x509_txt.c
@@ -162,8 +162,28 @@ const char *X509_verify_cert_error_string(long n)
 		return("invalid or inconsistent certificate policy extension");
 	case X509_V_ERR_NO_EXPLICIT_POLICY:
 		return("no explicit policy");
-	case X509_V_ERR_UNNESTED_RESOURCE:
-		return("RFC 3779 resource not subset of parent's resources");
+	case X509_V_ERR_DIFFERENT_CRL_SCOPE:
+	return("Different CRL scope");
+	case X509_V_ERR_UNSUPPORTED_EXTENSION_FEATURE:
+	return("Unsupported extension feature");
+ 	case X509_V_ERR_UNNESTED_RESOURCE:
+ 		return("RFC 3779 resource not subset of parent's resources");
+
+	case X509_V_ERR_PERMITTED_VIOLATION:
+		return("permitted subtree violation");
+	case X509_V_ERR_EXCLUDED_VIOLATION:
+		return("excluded subtree violation");
+	case X509_V_ERR_SUBTREE_MINMAX:
+		return("name constraints minimum and maximum not supported");
+	case X509_V_ERR_UNSUPPORTED_CONSTRAINT_TYPE:
+		return("unsupported name constraint type");
+	case X509_V_ERR_UNSUPPORTED_CONSTRAINT_SYNTAX:
+		return("unsupported or invalid name constraint syntax");
+	case X509_V_ERR_UNSUPPORTED_NAME_SYNTAX:
+		return("unsupported or invalid name syntax");
+	case X509_V_ERR_CRL_PATH_VALIDATION_ERROR:
+		return("CRL path validation error");
+
 	default:
 		BIO_snprintf(buf,sizeof buf,"error number %ld",n);
 		return(buf);
diff --git a/crypto/x509/x509_vfy.c b/crypto/x509/x509_vfy.c
index b92758480548..b0779db0231b 100644
--- a/crypto/x509/x509_vfy.c
+++ b/crypto/x509/x509_vfy.c
@@ -70,14 +70,70 @@
 #include <openssl/x509v3.h>
 #include <openssl/objects.h>
 
+/* CRL score values */
+
+/* No unhandled critical extensions */
+
+#define CRL_SCORE_NOCRITICAL	0x100
+
+/* certificate is within CRL scope */
+
+#define CRL_SCORE_SCOPE		0x080
+
+/* CRL times valid */
+
+#define CRL_SCORE_TIME		0x040
+
+/* Issuer name matches certificate */
+
+#define CRL_SCORE_ISSUER_NAME	0x020
+
+/* If this score or above CRL is probably valid */
+
+#define CRL_SCORE_VALID (CRL_SCORE_NOCRITICAL|CRL_SCORE_TIME|CRL_SCORE_SCOPE)
+
+/* CRL issuer is certificate issuer */
+
+#define CRL_SCORE_ISSUER_CERT	0x018
+
+/* CRL issuer is on certificate path */
+
+#define CRL_SCORE_SAME_PATH	0x008
+
+/* CRL issuer matches CRL AKID */
+
+#define CRL_SCORE_AKID		0x004
+
+/* Have a delta CRL with valid times */
+
+#define CRL_SCORE_TIME_DELTA	0x002
+
 static int null_callback(int ok,X509_STORE_CTX *e);
 static int check_issued(X509_STORE_CTX *ctx, X509 *x, X509 *issuer);
 static X509 *find_issuer(X509_STORE_CTX *ctx, STACK_OF(X509) *sk, X509 *x);
 static int check_chain_extensions(X509_STORE_CTX *ctx);
+static int check_name_constraints(X509_STORE_CTX *ctx);
 static int check_trust(X509_STORE_CTX *ctx);
 static int check_revocation(X509_STORE_CTX *ctx);
 static int check_cert(X509_STORE_CTX *ctx);
 static int check_policy(X509_STORE_CTX *ctx);
+
+static int get_crl_score(X509_STORE_CTX *ctx, X509 **pissuer,
+			unsigned int *preasons,
+			X509_CRL *crl, X509 *x);
+static int get_crl_delta(X509_STORE_CTX *ctx,
+				X509_CRL **pcrl, X509_CRL **pdcrl, X509 *x);
+static void get_delta_sk(X509_STORE_CTX *ctx, X509_CRL **dcrl, int *pcrl_score,
+			X509_CRL *base, STACK_OF(X509_CRL) *crls);
+static void crl_akid_check(X509_STORE_CTX *ctx, X509_CRL *crl,
+				X509 **pissuer, int *pcrl_score);
+static int crl_crldp_check(X509 *x, X509_CRL *crl, int crl_score,
+				unsigned int *preasons);
+static int check_crl_path(X509_STORE_CTX *ctx, X509 *x);
+static int check_crl_chain(X509_STORE_CTX *ctx,
+			STACK_OF(X509) *cert_path,
+			STACK_OF(X509) *crl_path);
+
 static int internal_verify(X509_STORE_CTX *ctx);
 const char X509_version[]="X.509" OPENSSL_VERSION_PTEXT;
 
@@ -284,6 +340,12 @@ int X509_verify_cert(X509_STORE_CTX *ctx)
 
 	if (!ok) goto end;
 
+	/* Check name constraints */
+
+	ok = check_name_constraints(ctx);
+	
+	if (!ok) goto end;
+
 	/* The chain extensions are OK: check trust */
 
 	if (param->trust > 0) ok = check_trust(ctx);
@@ -393,8 +455,8 @@ static int check_chain_extensions(X509_STORE_CTX *ctx)
 	X509 *x;
 	int (*cb)(int xok,X509_STORE_CTX *xctx);
 	int proxy_path_length = 0;
-	int allow_proxy_certs =
-		!!(ctx->param->flags & X509_V_FLAG_ALLOW_PROXY_CERTS);
+	int purpose;
+	int allow_proxy_certs;
 	cb=ctx->verify_cb;
 
 	/* must_be_ca can have 1 of 3 values:
@@ -407,10 +469,22 @@ static int check_chain_extensions(X509_STORE_CTX *ctx)
 	*/
 	must_be_ca = -1;
 
-	/* A hack to keep people who don't want to modify their software
-	   happy */
-	if (getenv("OPENSSL_ALLOW_PROXY_CERTS"))
-		allow_proxy_certs = 1;
+	/* CRL path validation */
+	if (ctx->parent)
+		{
+		allow_proxy_certs = 0;
+		purpose = X509_PURPOSE_CRL_SIGN;
+		}
+	else
+		{
+		allow_proxy_certs =
+			!!(ctx->param->flags & X509_V_FLAG_ALLOW_PROXY_CERTS);
+		/* A hack to keep people who don't want to modify their
+		   software happy */
+		if (getenv("OPENSSL_ALLOW_PROXY_CERTS"))
+			allow_proxy_certs = 1;
+		purpose = ctx->param->purpose;
+		}
 
 	/* Check all untrusted certificates */
 	for (i = 0; i < ctx->last_untrusted; i++)
@@ -477,8 +551,7 @@ static int check_chain_extensions(X509_STORE_CTX *ctx)
 			}
 		if (ctx->param->purpose > 0)
 			{
-			ret = X509_check_purpose(x, ctx->param->purpose,
-				must_be_ca > 0);
+			ret = X509_check_purpose(x, purpose, must_be_ca > 0);
 			if ((ret == 0)
 				|| ((ctx->param->flags & X509_V_FLAG_X509_STRICT)
 					&& (ret != 1)))
@@ -531,6 +604,42 @@ static int check_chain_extensions(X509_STORE_CTX *ctx)
 #endif
 }
 
+static int check_name_constraints(X509_STORE_CTX *ctx)
+	{
+	X509 *x;
+	int i, j, rv;
+	/* Check name constraints for all certificates */
+	for (i = sk_X509_num(ctx->chain) - 1; i >= 0; i--)
+		{
+		x = sk_X509_value(ctx->chain, i);
+		/* Ignore self issued certs unless last in chain */
+		if (i && (x->ex_flags & EXFLAG_SI))
+			continue;
+		/* Check against constraints for all certificates higher in
+		 * chain including trust anchor. Trust anchor not strictly
+		 * speaking needed but if it includes constraints it is to be
+		 * assumed it expects them to be obeyed.
+		 */
+		for (j = sk_X509_num(ctx->chain) - 1; j > i; j--)
+			{
+			NAME_CONSTRAINTS *nc = sk_X509_value(ctx->chain, j)->nc;
+			if (nc)
+				{
+				rv = NAME_CONSTRAINTS_check(x, nc);
+				if (rv != X509_V_OK)
+					{
+					ctx->error = rv;
+					ctx->error_depth = i;
+					ctx->current_cert = x;
+					if (!ctx->verify_cb(0,ctx))
+						return 0;
+					}
+				}
+			}
+		}
+	return 1;
+	}
+
 static int check_trust(X509_STORE_CTX *ctx)
 {
 #ifdef OPENSSL_NO_CHAIN_VERIFY
@@ -565,7 +674,12 @@ static int check_revocation(X509_STORE_CTX *ctx)
 	if (ctx->param->flags & X509_V_FLAG_CRL_CHECK_ALL)
 		last = sk_X509_num(ctx->chain) - 1;
 	else
+		{
+		/* If checking CRL paths this isn't the EE certificate */
+		if (ctx->parent)
+			return 1;
 		last = 0;
+		}
 	for(i = 0; i <= last; i++)
 		{
 		ctx->error_depth = i;
@@ -577,30 +691,66 @@ static int check_revocation(X509_STORE_CTX *ctx)
 
 static int check_cert(X509_STORE_CTX *ctx)
 	{
-	X509_CRL *crl = NULL;
+	X509_CRL *crl = NULL, *dcrl = NULL;
 	X509 *x;
 	int ok, cnum;
 	cnum = ctx->error_depth;
 	x = sk_X509_value(ctx->chain, cnum);
 	ctx->current_cert = x;
-	/* Try to retrieve relevant CRL */
-	ok = ctx->get_crl(ctx, &crl, x);
-	/* If error looking up CRL, nothing we can do except
-	 * notify callback
-	 */
-	if(!ok)
+	ctx->current_issuer = NULL;
+	ctx->current_crl_score = 0;
+	ctx->current_reasons = 0;
+	while (ctx->current_reasons != CRLDP_ALL_REASONS)
 		{
-		ctx->error = X509_V_ERR_UNABLE_TO_GET_CRL;
-		ok = ctx->verify_cb(0, ctx);
-		goto err;
+		/* Try to retrieve relevant CRL */
+		if (ctx->get_crl)
+			ok = ctx->get_crl(ctx, &crl, x);
+		else
+			ok = get_crl_delta(ctx, &crl, &dcrl, x);
+		/* If error looking up CRL, nothing we can do except
+		 * notify callback
+		 */
+		if(!ok)
+			{
+			ctx->error = X509_V_ERR_UNABLE_TO_GET_CRL;
+			ok = ctx->verify_cb(0, ctx);
+			goto err;
+			}
+		ctx->current_crl = crl;
+		ok = ctx->check_crl(ctx, crl);
+		if (!ok)
+			goto err;
+
+		if (dcrl)
+			{
+			ok = ctx->check_crl(ctx, dcrl);
+			if (!ok)
+				goto err;
+			ok = ctx->cert_crl(ctx, dcrl, x);
+			if (!ok)
+				goto err;
+			}
+		else
+			ok = 1;
+
+		/* Don't look in full CRL if delta reason is removefromCRL */
+		if (ok != 2)
+			{
+			ok = ctx->cert_crl(ctx, crl, x);
+			if (!ok)
+				goto err;
+			}
+
+		X509_CRL_free(crl);
+		X509_CRL_free(dcrl);
+		crl = NULL;
+		dcrl = NULL;
 		}
-	ctx->current_crl = crl;
-	ok = ctx->check_crl(ctx, crl);
-	if (!ok) goto err;
-	ok = ctx->cert_crl(ctx, crl, x);
 	err:
-	ctx->current_crl = NULL;
 	X509_CRL_free(crl);
+	X509_CRL_free(dcrl);
+
+	ctx->current_crl = NULL;
 	return ok;
 
 	}
@@ -611,7 +761,8 @@ static int check_crl_time(X509_STORE_CTX *ctx, X509_CRL *crl, int notify)
 	{
 	time_t *ptime;
 	int i;
-	ctx->current_crl = crl;
+	if (notify)
+		ctx->current_crl = crl;
 	if (ctx->param->flags & X509_V_FLAG_USE_CHECK_TIME)
 		ptime = &ctx->param->check_time;
 	else
@@ -620,15 +771,19 @@ static int check_crl_time(X509_STORE_CTX *ctx, X509_CRL *crl, int notify)
 	i=X509_cmp_time(X509_CRL_get_lastUpdate(crl), ptime);
 	if (i == 0)
 		{
+		if (!notify)
+			return 0;
 		ctx->error=X509_V_ERR_ERROR_IN_CRL_LAST_UPDATE_FIELD;
-		if (!notify || !ctx->verify_cb(0, ctx))
+		if (!ctx->verify_cb(0, ctx))
 			return 0;
 		}
 
 	if (i > 0)
 		{
+		if (!notify)
+			return 0;
 		ctx->error=X509_V_ERR_CRL_NOT_YET_VALID;
-		if (!notify || !ctx->verify_cb(0, ctx))
+		if (!ctx->verify_cb(0, ctx))
 			return 0;
 		}
 
@@ -638,92 +793,545 @@ static int check_crl_time(X509_STORE_CTX *ctx, X509_CRL *crl, int notify)
 
 		if (i == 0)
 			{
+			if (!notify)
+				return 0;
 			ctx->error=X509_V_ERR_ERROR_IN_CRL_NEXT_UPDATE_FIELD;
-			if (!notify || !ctx->verify_cb(0, ctx))
+			if (!ctx->verify_cb(0, ctx))
 				return 0;
 			}
-
-		if (i < 0)
+		/* Ignore expiry of base CRL is delta is valid */
+		if ((i < 0) && !(ctx->current_crl_score & CRL_SCORE_TIME_DELTA))
 			{
+			if (!notify)
+				return 0;
 			ctx->error=X509_V_ERR_CRL_HAS_EXPIRED;
-			if (!notify || !ctx->verify_cb(0, ctx))
+			if (!ctx->verify_cb(0, ctx))
 				return 0;
 			}
 		}
 
-	ctx->current_crl = NULL;
+	if (notify)
+		ctx->current_crl = NULL;
 
 	return 1;
 	}
 
-/* Lookup CRLs from the supplied list. Look for matching isser name
- * and validity. If we can't find a valid CRL return the last one
- * with matching name. This gives more meaningful error codes. Otherwise
- * we'd get a CRL not found error if a CRL existed with matching name but
- * was invalid.
- */
-
-static int get_crl_sk(X509_STORE_CTX *ctx, X509_CRL **pcrl,
-			X509_NAME *nm, STACK_OF(X509_CRL) *crls)
+static int get_crl_sk(X509_STORE_CTX *ctx, X509_CRL **pcrl, X509_CRL **pdcrl,
+			X509 **pissuer, int *pscore, unsigned int *preasons,
+			STACK_OF(X509_CRL) *crls)
 	{
-	int i;
+	int i, crl_score, best_score = *pscore;
+	unsigned int reasons, best_reasons = 0;
+	X509 *x = ctx->current_cert;
 	X509_CRL *crl, *best_crl = NULL;
+	X509 *crl_issuer = NULL, *best_crl_issuer = NULL;
+
 	for (i = 0; i < sk_X509_CRL_num(crls); i++)
 		{
 		crl = sk_X509_CRL_value(crls, i);
-		if (X509_NAME_cmp(nm, X509_CRL_get_issuer(crl)))
-			continue;
-		if (check_crl_time(ctx, crl, 0))
+		reasons = *preasons;
+		crl_score = get_crl_score(ctx, &crl_issuer, &reasons, crl, x);
+
+		if (crl_score > best_score)
 			{
-			*pcrl = crl;
-			CRYPTO_add(&crl->references, 1, CRYPTO_LOCK_X509);
-			return 1;
+			best_crl = crl;
+			best_crl_issuer = crl_issuer;
+			best_score = crl_score;
+			best_reasons = reasons;
 			}
-		best_crl = crl;
 		}
+
 	if (best_crl)
 		{
+		if (*pcrl)
+			X509_CRL_free(*pcrl);
 		*pcrl = best_crl;
-		CRYPTO_add(&best_crl->references, 1, CRYPTO_LOCK_X509);
+		*pissuer = best_crl_issuer;
+		*pscore = best_score;
+		*preasons = best_reasons;
+		CRYPTO_add(&best_crl->references, 1, CRYPTO_LOCK_X509_CRL);
+		if (*pdcrl)
+			{
+			X509_CRL_free(*pdcrl);
+			*pdcrl = NULL;
+			}
+		get_delta_sk(ctx, pdcrl, pscore, best_crl, crls);
 		}
-		
+
+	if (best_score >= CRL_SCORE_VALID)
+		return 1;
+
 	return 0;
 	}
 
-/* Retrieve CRL corresponding to certificate: currently just a
- * subject lookup: maybe use AKID later...
+/* Compare two CRL extensions for delta checking purposes. They should be
+ * both present or both absent. If both present all fields must be identical.
  */
-static int get_crl(X509_STORE_CTX *ctx, X509_CRL **pcrl, X509 *x)
+
+static int crl_extension_match(X509_CRL *a, X509_CRL *b, int nid)
 	{
-	int ok;
-	X509_CRL *crl = NULL;
-	X509_OBJECT xobj;
-	X509_NAME *nm;
-	nm = X509_get_issuer_name(x);
-	ok = get_crl_sk(ctx, &crl, nm, ctx->crls);
-	if (ok)
+	ASN1_OCTET_STRING *exta, *extb;
+	int i;
+	i = X509_CRL_get_ext_by_NID(a, nid, 0);
+	if (i >= 0)
 		{
-		*pcrl = crl;
+		/* Can't have multiple occurrences */
+		if (X509_CRL_get_ext_by_NID(a, nid, i) != -1)
+			return 0;
+		exta = X509_EXTENSION_get_data(X509_CRL_get_ext(a, i));
+		}
+	else
+		exta = NULL;
+
+	i = X509_CRL_get_ext_by_NID(b, nid, 0);
+
+	if (i >= 0)
+		{
+
+		if (X509_CRL_get_ext_by_NID(b, nid, i) != -1)
+			return 0;
+		extb = X509_EXTENSION_get_data(X509_CRL_get_ext(b, i));
+		}
+	else
+		extb = NULL;
+
+	if (!exta && !extb)
 		return 1;
+
+	if (!exta || !extb)
+		return 0;
+
+
+	if (ASN1_OCTET_STRING_cmp(exta, extb))
+		return 0;
+
+	return 1;
+	}
+
+/* See if a base and delta are compatible */
+
+static int check_delta_base(X509_CRL *delta, X509_CRL *base)
+	{
+	/* Delta CRL must be a delta */
+	if (!delta->base_crl_number)
+			return 0;
+	/* Base must have a CRL number */
+	if (!base->crl_number)
+			return 0;
+	/* Issuer names must match */
+	if (X509_NAME_cmp(X509_CRL_get_issuer(base),
+				X509_CRL_get_issuer(delta)))
+		return 0;
+	/* AKID and IDP must match */
+	if (!crl_extension_match(delta, base, NID_authority_key_identifier))
+			return 0;
+	if (!crl_extension_match(delta, base, NID_issuing_distribution_point))
+			return 0;
+	/* Delta CRL base number must not exceed Full CRL number. */
+	if (ASN1_INTEGER_cmp(delta->base_crl_number, base->crl_number) > 0)
+			return 0;
+	/* Delta CRL number must exceed full CRL number */
+	if (ASN1_INTEGER_cmp(delta->crl_number, base->crl_number) > 0)
+			return 1;
+	return 0;
+	}
+
+/* For a given base CRL find a delta... maybe extend to delta scoring
+ * or retrieve a chain of deltas...
+ */
+
+static void get_delta_sk(X509_STORE_CTX *ctx, X509_CRL **dcrl, int *pscore,
+			X509_CRL *base, STACK_OF(X509_CRL) *crls)
+	{
+	X509_CRL *delta;
+	int i;
+	if (!(ctx->param->flags & X509_V_FLAG_USE_DELTAS))
+		return;
+	if (!((ctx->current_cert->ex_flags | base->flags) & EXFLAG_FRESHEST))
+		return;
+	for (i = 0; i < sk_X509_CRL_num(crls); i++)
+		{
+		delta = sk_X509_CRL_value(crls, i);
+		if (check_delta_base(delta, base))
+			{
+			if (check_crl_time(ctx, delta, 0))
+				*pscore |= CRL_SCORE_TIME_DELTA;
+			CRYPTO_add(&delta->references, 1, CRYPTO_LOCK_X509_CRL);
+			*dcrl = delta;
+			return;
+			}
+		}
+	*dcrl = NULL;
+	}
+
+/* For a given CRL return how suitable it is for the supplied certificate 'x'.
+ * The return value is a mask of several criteria.
+ * If the issuer is not the certificate issuer this is returned in *pissuer.
+ * The reasons mask is also used to determine if the CRL is suitable: if
+ * no new reasons the CRL is rejected, otherwise reasons is updated.
+ */
+
+static int get_crl_score(X509_STORE_CTX *ctx, X509 **pissuer,
+			unsigned int *preasons,
+			X509_CRL *crl, X509 *x)
+	{
+
+	int crl_score = 0;
+	unsigned int tmp_reasons = *preasons, crl_reasons;
+
+	/* First see if we can reject CRL straight away */
+
+	/* Invalid IDP cannot be processed */
+	if (crl->idp_flags & IDP_INVALID)
+		return 0;
+	/* Reason codes or indirect CRLs need extended CRL support */
+	if (!(ctx->param->flags & X509_V_FLAG_EXTENDED_CRL_SUPPORT))
+		{
+		if (crl->idp_flags & (IDP_INDIRECT | IDP_REASONS))
+			return 0;
+		}
+	else if (crl->idp_flags & IDP_REASONS)
+		{
+		/* If no new reasons reject */
+		if (!(crl->idp_reasons & ~tmp_reasons))
+			return 0;
+		}
+	/* Don't process deltas at this stage */
+	else if (crl->base_crl_number)
+		return 0;
+	/* If issuer name doesn't match certificate need indirect CRL */
+	if (X509_NAME_cmp(X509_get_issuer_name(x), X509_CRL_get_issuer(crl)))
+		{
+		if (!(crl->idp_flags & IDP_INDIRECT))
+			return 0;
+		}
+	else
+		crl_score |= CRL_SCORE_ISSUER_NAME;
+
+	if (!(crl->flags & EXFLAG_CRITICAL))
+		crl_score |= CRL_SCORE_NOCRITICAL;
+
+	/* Check expiry */
+	if (check_crl_time(ctx, crl, 0))
+		crl_score |= CRL_SCORE_TIME;
+
+	/* Check authority key ID and locate certificate issuer */
+	crl_akid_check(ctx, crl, pissuer, &crl_score);
+
+	/* If we can't locate certificate issuer at this point forget it */
+
+	if (!(crl_score & CRL_SCORE_AKID))
+		return 0;
+
+	/* Check cert for matching CRL distribution points */
+
+	if (crl_crldp_check(x, crl, crl_score, &crl_reasons))
+		{
+		/* If no new reasons reject */
+		if (!(crl_reasons & ~tmp_reasons))
+			return 0;
+		tmp_reasons |= crl_reasons;
+		crl_score |= CRL_SCORE_SCOPE;
 		}
 
-	ok = X509_STORE_get_by_subject(ctx, X509_LU_CRL, nm, &xobj);
+	*preasons = tmp_reasons;
+
+	return crl_score;
+
+	}
 
-	if (!ok)
+static void crl_akid_check(X509_STORE_CTX *ctx, X509_CRL *crl,
+				X509 **pissuer, int *pcrl_score)
+	{
+	X509 *crl_issuer = NULL;
+	X509_NAME *cnm = X509_CRL_get_issuer(crl);
+	int cidx = ctx->error_depth;
+	int i;
+
+	if (cidx != sk_X509_num(ctx->chain) - 1)
+		cidx++;
+
+	crl_issuer = sk_X509_value(ctx->chain, cidx);
+
+	if (X509_check_akid(crl_issuer, crl->akid) == X509_V_OK)
 		{
-		/* If we got a near match from get_crl_sk use that */
-		if (crl)
+		if (*pcrl_score & CRL_SCORE_ISSUER_NAME)
 			{
-			*pcrl = crl;
-			return 1;
+			*pcrl_score |= CRL_SCORE_AKID|CRL_SCORE_ISSUER_CERT;
+			*pissuer = crl_issuer;
+			return;
+			}
+		}
+
+	for (cidx++; cidx < sk_X509_num(ctx->chain); cidx++)
+		{
+		crl_issuer = sk_X509_value(ctx->chain, cidx);
+		if (X509_NAME_cmp(X509_get_subject_name(crl_issuer), cnm))
+			continue;
+		if (X509_check_akid(crl_issuer, crl->akid) == X509_V_OK)
+			{
+			*pcrl_score |= CRL_SCORE_AKID|CRL_SCORE_SAME_PATH;
+			*pissuer = crl_issuer;
+			return;
+			}
+		}
+
+	/* Anything else needs extended CRL support */
+
+	if (!(ctx->param->flags & X509_V_FLAG_EXTENDED_CRL_SUPPORT))
+		return;
+
+	/* Otherwise the CRL issuer is not on the path. Look for it in the
+	 * set of untrusted certificates.
+	 */
+	for (i = 0; i < sk_X509_num(ctx->untrusted); i++)
+		{
+		crl_issuer = sk_X509_value(ctx->untrusted, i);
+		if (X509_NAME_cmp(X509_get_subject_name(crl_issuer), cnm))
+			continue;
+		if (X509_check_akid(crl_issuer, crl->akid) == X509_V_OK)
+			{
+			*pissuer = crl_issuer;
+			*pcrl_score |= CRL_SCORE_AKID;
+			return;
+			}
+		}
+	}
+
+/* Check the path of a CRL issuer certificate. This creates a new
+ * X509_STORE_CTX and populates it with most of the parameters from the
+ * parent. This could be optimised somewhat since a lot of path checking
+ * will be duplicated by the parent, but this will rarely be used in 
+ * practice.
+ */
+
+static int check_crl_path(X509_STORE_CTX *ctx, X509 *x)
+	{
+	X509_STORE_CTX crl_ctx;
+	int ret;
+	/* Don't allow recursive CRL path validation */
+	if (ctx->parent)
+		return 0;
+	if (!X509_STORE_CTX_init(&crl_ctx, ctx->ctx, x, ctx->untrusted))
+		return -1;
+
+	crl_ctx.crls = ctx->crls;
+	/* Copy verify params across */
+	X509_STORE_CTX_set0_param(&crl_ctx, ctx->param);
+
+	crl_ctx.parent = ctx;
+	crl_ctx.verify_cb = ctx->verify_cb;
+
+	/* Verify CRL issuer */
+	ret = X509_verify_cert(&crl_ctx);
+
+	if (ret <= 0)
+		goto err;
+
+	/* Check chain is acceptable */
+
+	ret = check_crl_chain(ctx, ctx->chain, crl_ctx.chain);
+	err:
+	X509_STORE_CTX_cleanup(&crl_ctx);
+	return ret;
+	}
+
+/* RFC3280 says nothing about the relationship between CRL path
+ * and certificate path, which could lead to situations where a
+ * certificate could be revoked or validated by a CA not authorised
+ * to do so. RFC5280 is more strict and states that the two paths must
+ * end in the same trust anchor, though some discussions remain...
+ * until this is resolved we use the RFC5280 version
+ */
+
+static int check_crl_chain(X509_STORE_CTX *ctx,
+			STACK_OF(X509) *cert_path,
+			STACK_OF(X509) *crl_path)
+	{
+	X509 *cert_ta, *crl_ta;
+	cert_ta = sk_X509_value(cert_path, sk_X509_num(cert_path) - 1);
+	crl_ta = sk_X509_value(crl_path, sk_X509_num(crl_path) - 1);
+	if (!X509_cmp(cert_ta, crl_ta))
+		return 1;
+	return 0;
+	}
+
+/* Check for match between two dist point names: three separate cases.
+ * 1. Both are relative names and compare X509_NAME types.
+ * 2. One full, one relative. Compare X509_NAME to GENERAL_NAMES.
+ * 3. Both are full names and compare two GENERAL_NAMES.
+ * 4. One is NULL: automatic match.
+ */
+
+
+static int idp_check_dp(DIST_POINT_NAME *a, DIST_POINT_NAME *b)
+	{
+	X509_NAME *nm = NULL;
+	GENERAL_NAMES *gens = NULL;
+	GENERAL_NAME *gena, *genb;
+	int i, j;
+	if (!a || !b)
+		return 1;
+	if (a->type == 1)
+		{
+		if (!a->dpname)
+			return 0;
+		/* Case 1: two X509_NAME */
+		if (b->type == 1)
+			{
+			if (!b->dpname)
+				return 0;
+			if (!X509_NAME_cmp(a->dpname, b->dpname))
+				return 1;
+			else
+				return 0;
 			}
+		/* Case 2: set name and GENERAL_NAMES appropriately */
+		nm = a->dpname;
+		gens = b->name.fullname;
+		}
+	else if (b->type == 1)
+		{
+		if (!b->dpname)
+			return 0;
+		/* Case 2: set name and GENERAL_NAMES appropriately */
+		gens = a->name.fullname;
+		nm = b->dpname;
+		}
+
+	/* Handle case 2 with one GENERAL_NAMES and one X509_NAME */
+	if (nm)
+		{
+		for (i = 0; i < sk_GENERAL_NAME_num(gens); i++)
+			{
+			gena = sk_GENERAL_NAME_value(gens, i);	
+			if (gena->type != GEN_DIRNAME)
+				continue;
+			if (!X509_NAME_cmp(nm, gena->d.directoryName))
+				return 1;
+			}
+		return 0;
+		}
+
+	/* Else case 3: two GENERAL_NAMES */
+
+	for (i = 0; i < sk_GENERAL_NAME_num(a->name.fullname); i++)
+		{
+		gena = sk_GENERAL_NAME_value(a->name.fullname, i);
+		for (j = 0; j < sk_GENERAL_NAME_num(b->name.fullname); j++)
+			{
+			genb = sk_GENERAL_NAME_value(b->name.fullname, j);
+			if (!GENERAL_NAME_cmp(gena, genb))
+				return 1;
+			}
+		}
+
+	return 0;
+
+	}
+
+static int crldp_check_crlissuer(DIST_POINT *dp, X509_CRL *crl, int crl_score)
+	{
+	int i;
+	X509_NAME *nm = X509_CRL_get_issuer(crl);
+	/* If no CRLissuer return is successful iff don't need a match */
+	if (!dp->CRLissuer)
+		return !!(crl_score & CRL_SCORE_ISSUER_NAME);
+	for (i = 0; i < sk_GENERAL_NAME_num(dp->CRLissuer); i++)
+		{
+		GENERAL_NAME *gen = sk_GENERAL_NAME_value(dp->CRLissuer, i);
+		if (gen->type != GEN_DIRNAME)
+			continue;
+		if (!X509_NAME_cmp(gen->d.directoryName, nm))
+			return 1;
+		}
+	return 0;
+	}
+
+/* Check CRLDP and IDP */
+
+static int crl_crldp_check(X509 *x, X509_CRL *crl, int crl_score,
+				unsigned int *preasons)
+	{
+	int i;
+	if (crl->idp_flags & IDP_ONLYATTR)
 		return 0;
+	if (x->ex_flags & EXFLAG_CA)
+		{
+		if (crl->idp_flags & IDP_ONLYUSER)
+			return 0;
+		}
+	else
+		{
+		if (crl->idp_flags & IDP_ONLYCA)
+			return 0;
+		}
+	*preasons = crl->idp_reasons;
+	for (i = 0; i < sk_DIST_POINT_num(x->crldp); i++)
+		{
+		DIST_POINT *dp = sk_DIST_POINT_value(x->crldp, i);
+		if (crldp_check_crlissuer(dp, crl, crl_score))
+			{
+			if (!crl->idp ||
+			     idp_check_dp(dp->distpoint, crl->idp->distpoint))
+				{
+				*preasons &= dp->dp_reasons;
+				return 1;
+				}
+			}
 		}
+	if ((!crl->idp || !crl->idp->distpoint) && (crl_score & CRL_SCORE_ISSUER_NAME))
+		return 1;
+	return 0;
+	}
+
+/* Retrieve CRL corresponding to current certificate.
+ * If deltas enabled try to find a delta CRL too
+ */
+	
+static int get_crl_delta(X509_STORE_CTX *ctx,
+				X509_CRL **pcrl, X509_CRL **pdcrl, X509 *x)
+	{
+	int ok;
+	X509 *issuer = NULL;
+	int crl_score = 0;
+	unsigned int reasons;
+	X509_CRL *crl = NULL, *dcrl = NULL;
+	STACK_OF(X509_CRL) *skcrl;
+	X509_NAME *nm = X509_get_issuer_name(x);
+	reasons = ctx->current_reasons;
+	ok = get_crl_sk(ctx, &crl, &dcrl, 
+				&issuer, &crl_score, &reasons, ctx->crls);
+
+	if (ok)
+		goto done;
+
+	/* Lookup CRLs from store */
 
-	*pcrl = xobj.data.crl;
+	skcrl = ctx->lookup_crls(ctx, nm);
+
+	/* If no CRLs found and a near match from get_crl_sk use that */
+	if (!skcrl && crl)
+		goto done;
+
+	get_crl_sk(ctx, &crl, &dcrl, &issuer, &crl_score, &reasons, skcrl);
+
+	sk_X509_CRL_pop_free(skcrl, X509_CRL_free);
+
+	done:
+
+	/* If we got any kind of CRL use it and return success */
 	if (crl)
-		X509_CRL_free(crl);
-	return 1;
+		{
+		ctx->current_issuer = issuer;
+		ctx->current_crl_score = crl_score;
+		ctx->current_reasons = reasons;
+		*pcrl = crl;
+		*pdcrl = dcrl;
+		return 1;
+		}
+
+	return 0;
 	}
 
 /* Check CRL validity */
@@ -734,10 +1342,14 @@ static int check_crl(X509_STORE_CTX *ctx, X509_CRL *crl)
 	int ok = 0, chnum, cnum;
 	cnum = ctx->error_depth;
 	chnum = sk_X509_num(ctx->chain) - 1;
-	/* Find CRL issuer: if not last certificate then issuer
+	/* if we have an alternative CRL issuer cert use that */
+	if (ctx->current_issuer)
+		issuer = ctx->current_issuer;
+
+	/* Else find CRL issuer: if not last certificate then issuer
 	 * is next certificate in chain.
 	 */
-	if(cnum < chnum)
+	else if (cnum < chnum)
 		issuer = sk_X509_value(ctx->chain, cnum + 1);
 	else
 		{
@@ -753,13 +1365,52 @@ static int check_crl(X509_STORE_CTX *ctx, X509_CRL *crl)
 
 	if(issuer)
 		{
-		/* Check for cRLSign bit if keyUsage present */
-		if ((issuer->ex_flags & EXFLAG_KUSAGE) &&
-			!(issuer->ex_kusage & KU_CRL_SIGN))
+		/* Skip most tests for deltas because they have already
+		 * been done
+		 */
+		if (!crl->base_crl_number)
 			{
-			ctx->error = X509_V_ERR_KEYUSAGE_NO_CRL_SIGN;
-			ok = ctx->verify_cb(0, ctx);
-			if(!ok) goto err;
+			/* Check for cRLSign bit if keyUsage present */
+			if ((issuer->ex_flags & EXFLAG_KUSAGE) &&
+				!(issuer->ex_kusage & KU_CRL_SIGN))
+				{
+				ctx->error = X509_V_ERR_KEYUSAGE_NO_CRL_SIGN;
+				ok = ctx->verify_cb(0, ctx);
+				if(!ok) goto err;
+				}
+
+			if (!(ctx->current_crl_score & CRL_SCORE_SCOPE))
+				{
+				ctx->error = X509_V_ERR_DIFFERENT_CRL_SCOPE;
+				ok = ctx->verify_cb(0, ctx);
+				if(!ok) goto err;
+				}
+
+			if (!(ctx->current_crl_score & CRL_SCORE_SAME_PATH))
+				{
+				if (check_crl_path(ctx, ctx->current_issuer) <= 0)
+					{
+					ctx->error = X509_V_ERR_CRL_PATH_VALIDATION_ERROR;
+					ok = ctx->verify_cb(0, ctx);
+					if(!ok) goto err;
+					}
+				}
+
+			if (crl->idp_flags & IDP_INVALID)
+				{
+				ctx->error = X509_V_ERR_INVALID_EXTENSION;
+				ok = ctx->verify_cb(0, ctx);
+				if(!ok) goto err;
+				}
+
+
+			}
+
+		if (!(ctx->current_crl_score & CRL_SCORE_TIME))
+			{
+			ok = check_crl_time(ctx, crl, 1);
+			if (!ok)
+				goto err;
 			}
 
 		/* Attempt to get issuer certificate public key */
@@ -783,10 +1434,6 @@ static int check_crl(X509_STORE_CTX *ctx, X509_CRL *crl)
 			}
 		}
 
-	ok = check_crl_time(ctx, crl, 1);
-	if (!ok)
-		goto err;
-
 	ok = 1;
 
 	err:
@@ -797,62 +1444,43 @@ static int check_crl(X509_STORE_CTX *ctx, X509_CRL *crl)
 /* Check certificate against CRL */
 static int cert_crl(X509_STORE_CTX *ctx, X509_CRL *crl, X509 *x)
 	{
-	int idx, ok;
-	X509_REVOKED rtmp;
-	STACK_OF(X509_EXTENSION) *exts;
-	X509_EXTENSION *ext;
-	/* Look for serial number of certificate in CRL */
-	rtmp.serialNumber = X509_get_serialNumber(x);
-	/* Sort revoked into serial number order if not already sorted.
-	 * Do this under a lock to avoid race condition.
- 	 */
-	if (!sk_X509_REVOKED_is_sorted(crl->crl->revoked))
+	int ok;
+	X509_REVOKED *rev;
+	/* The rules changed for this... previously if a CRL contained
+	 * unhandled critical extensions it could still be used to indicate
+	 * a certificate was revoked. This has since been changed since 
+	 * critical extension can change the meaning of CRL entries.
+	 */
+	if (crl->flags & EXFLAG_CRITICAL)
 		{
-		CRYPTO_w_lock(CRYPTO_LOCK_X509_CRL);
-		sk_X509_REVOKED_sort(crl->crl->revoked);
-		CRYPTO_w_unlock(CRYPTO_LOCK_X509_CRL);
+		if (ctx->param->flags & X509_V_FLAG_IGNORE_CRITICAL)
+			return 1;
+		ctx->error = X509_V_ERR_UNHANDLED_CRITICAL_CRL_EXTENSION;
+		ok = ctx->verify_cb(0, ctx);
+		if(!ok)
+			return 0;
 		}
-	idx = sk_X509_REVOKED_find(crl->crl->revoked, &rtmp);
-	/* If found assume revoked: want something cleverer than
-	 * this to handle entry extensions in V2 CRLs.
+	/* Look for serial number of certificate in CRL
+	 * If found make sure reason is not removeFromCRL.
 	 */
-	if(idx >= 0)
+	if (X509_CRL_get0_by_cert(crl, &rev, x))
 		{
+		if (rev->reason == CRL_REASON_REMOVE_FROM_CRL)
+			return 2;
 		ctx->error = X509_V_ERR_CERT_REVOKED;
 		ok = ctx->verify_cb(0, ctx);
-		if (!ok) return 0;
+		if (!ok)
+			return 0;
 		}
 
-	if (ctx->param->flags & X509_V_FLAG_IGNORE_CRITICAL)
-		return 1;
-
-	/* See if we have any critical CRL extensions: since we
-	 * currently don't handle any CRL extensions the CRL must be
-	 * rejected. 
-	 * This code accesses the X509_CRL structure directly: applications
-	 * shouldn't do this.
-	 */
-
-	exts = crl->crl->extensions;
-
-	for (idx = 0; idx < sk_X509_EXTENSION_num(exts); idx++)
-		{
-		ext = sk_X509_EXTENSION_value(exts, idx);
-		if (ext->critical > 0)
-			{
-			ctx->error =
-				X509_V_ERR_UNHANDLED_CRITICAL_CRL_EXTENSION;
-			ok = ctx->verify_cb(0, ctx);
-			if(!ok) return 0;
-			break;
-			}
-		}
 	return 1;
 	}
 
 static int check_policy(X509_STORE_CTX *ctx)
 	{
 	int ret;
+	if (ctx->parent)
+		return 1;
 	ret = X509_policy_check(&ctx->tree, &ctx->explicit_policy, ctx->chain,
 				ctx->param->policies, ctx->param->flags);
 	if (ret == 0)
@@ -875,7 +1503,8 @@ static int check_policy(X509_STORE_CTX *ctx)
 				continue;
 			ctx->current_cert = x;
 			ctx->error = X509_V_ERR_INVALID_POLICY_EXTENSION;
-			ret = ctx->verify_cb(0, ctx);
+			if(!ctx->verify_cb(0, ctx))
+				return 0;
 			}
 		return 1;
 		}
@@ -1034,12 +1663,12 @@ end:
 	return ok;
 	}
 
-int X509_cmp_current_time(ASN1_TIME *ctm)
+int X509_cmp_current_time(const ASN1_TIME *ctm)
 {
 	return X509_cmp_time(ctm, NULL);
 }
 
-int X509_cmp_time(ASN1_TIME *ctm, time_t *cmp_time)
+int X509_cmp_time(const ASN1_TIME *ctm, time_t *cmp_time)
 	{
 	char *str;
 	ASN1_TIME atm;
@@ -1094,6 +1723,7 @@ int X509_cmp_time(ASN1_TIME *ctm, time_t *cmp_time)
 			offset= -offset;
 		}
 	atm.type=ctm->type;
+	atm.flags = 0;
 	atm.length=sizeof(buff2);
 	atm.data=(unsigned char *)buff2;
 
@@ -1122,19 +1752,28 @@ ASN1_TIME *X509_gmtime_adj(ASN1_TIME *s, long adj)
 	return X509_time_adj(s, adj, NULL);
 }
 
-ASN1_TIME *X509_time_adj(ASN1_TIME *s, long adj, time_t *in_tm)
+ASN1_TIME *X509_time_adj(ASN1_TIME *s, long offset_sec, time_t *in_tm)
+	{
+	return X509_time_adj_ex(s, 0, offset_sec, in_tm);
+	}
+
+ASN1_TIME *X509_time_adj_ex(ASN1_TIME *s,
+				int offset_day, long offset_sec, time_t *in_tm)
 	{
 	time_t t;
-	int type = -1;
 
 	if (in_tm) t = *in_tm;
 	else time(&t);
 
-	t+=adj;
-	if (s) type = s->type;
-	if (type == V_ASN1_UTCTIME) return ASN1_UTCTIME_set(s,t);
-	if (type == V_ASN1_GENERALIZEDTIME) return ASN1_GENERALIZEDTIME_set(s, t);
-	return ASN1_TIME_set(s, t);
+	if (s && !(s->flags & ASN1_STRING_FLAG_MSTRING))
+		{
+		if (s->type == V_ASN1_UTCTIME)
+			return ASN1_UTCTIME_adj(s,t, offset_day, offset_sec);
+		if (s->type == V_ASN1_GENERALIZEDTIME)
+			return ASN1_GENERALIZEDTIME_adj(s, t, offset_day,
+								offset_sec);
+		}
+	return ASN1_TIME_adj(s, t, offset_day, offset_sec);
 	}
 
 int X509_get_pubkey_parameters(EVP_PKEY *pkey, STACK_OF(X509) *chain)
@@ -1237,6 +1876,21 @@ STACK_OF(X509) *X509_STORE_CTX_get1_chain(X509_STORE_CTX *ctx)
 	return chain;
 	}
 
+X509 *X509_STORE_CTX_get0_current_issuer(X509_STORE_CTX *ctx)
+	{
+	return ctx->current_issuer;
+	}
+
+X509_CRL *X509_STORE_CTX_get0_current_crl(X509_STORE_CTX *ctx)
+	{
+	return ctx->current_crl;
+	}
+
+X509_STORE_CTX *X509_STORE_CTX_get0_parent_ctx(X509_STORE_CTX *ctx)
+	{
+	return ctx->parent;
+	}
+
 void X509_STORE_CTX_set_cert(X509_STORE_CTX *ctx, X509 *x)
 	{
 	ctx->cert=x;
@@ -1357,7 +2011,11 @@ int X509_STORE_CTX_init(X509_STORE_CTX *ctx, X509_STORE *store, X509 *x509,
 	ctx->error_depth=0;
 	ctx->current_cert=NULL;
 	ctx->current_issuer=NULL;
+	ctx->current_crl=NULL;
+	ctx->current_crl_score=0;
+	ctx->current_reasons=0;
 	ctx->tree = NULL;
+	ctx->parent = NULL;
 
 	ctx->param = X509_VERIFY_PARAM_new();
 
@@ -1423,7 +2081,7 @@ int X509_STORE_CTX_init(X509_STORE_CTX *ctx, X509_STORE *store, X509 *x509,
 	if (store && store->get_crl)
 		ctx->get_crl = store->get_crl;
 	else
-		ctx->get_crl = get_crl;
+		ctx->get_crl = NULL;
 
 	if (store && store->check_crl)
 		ctx->check_crl = store->check_crl;
@@ -1435,6 +2093,16 @@ int X509_STORE_CTX_init(X509_STORE_CTX *ctx, X509_STORE *store, X509 *x509,
 	else
 		ctx->cert_crl = cert_crl;
 
+	if (store && store->lookup_certs)
+		ctx->lookup_certs = store->lookup_certs;
+	else
+		ctx->lookup_certs = X509_STORE_get1_certs;
+
+	if (store && store->lookup_crls)
+		ctx->lookup_crls = store->lookup_crls;
+	else
+		ctx->lookup_crls = X509_STORE_get1_crls;
+
 	ctx->check_policy = check_policy;
 
 
@@ -1467,7 +2135,8 @@ void X509_STORE_CTX_cleanup(X509_STORE_CTX *ctx)
 	if (ctx->cleanup) ctx->cleanup(ctx);
 	if (ctx->param != NULL)
 		{
-		X509_VERIFY_PARAM_free(ctx->param);
+		if (ctx->parent == NULL)
+			X509_VERIFY_PARAM_free(ctx->param);
 		ctx->param=NULL;
 		}
 	if (ctx->tree != NULL)
diff --git a/crypto/x509/x509_vfy.h b/crypto/x509/x509_vfy.h
index 86ae35f69d5f..fe09b30aaa62 100644
--- a/crypto/x509/x509_vfy.h
+++ b/crypto/x509/x509_vfy.h
@@ -77,6 +77,7 @@
 extern "C" {
 #endif
 
+#if 0
 /* Outer object */
 typedef struct x509_hash_dir_st
 	{
@@ -85,6 +86,7 @@ typedef struct x509_hash_dir_st
 	int *dirs_type;
 	int num_dirs_alloced;
 	} X509_HASH_DIR_CTX;
+#endif
 
 typedef struct x509_file_st
 	{
@@ -198,6 +200,8 @@ struct x509_store_st
 	int (*get_crl)(X509_STORE_CTX *ctx, X509_CRL **crl, X509 *x); /* retrieve CRL */
 	int (*check_crl)(X509_STORE_CTX *ctx, X509_CRL *crl); /* Check CRL validity */
 	int (*cert_crl)(X509_STORE_CTX *ctx, X509_CRL *crl, X509 *x); /* Check certificate against CRL */
+	STACK_OF(X509) * (*lookup_certs)(X509_STORE_CTX *ctx, X509_NAME *nm);
+	STACK_OF(X509_CRL) * (*lookup_crls)(X509_STORE_CTX *ctx, X509_NAME *nm);
 	int (*cleanup)(X509_STORE_CTX *ctx);
 
 	CRYPTO_EX_DATA ex_data;
@@ -246,6 +250,8 @@ struct x509_store_ctx_st      /* X509_STORE_CTX */
 	int (*check_crl)(X509_STORE_CTX *ctx, X509_CRL *crl); /* Check CRL validity */
 	int (*cert_crl)(X509_STORE_CTX *ctx, X509_CRL *crl, X509 *x); /* Check certificate against CRL */
 	int (*check_policy)(X509_STORE_CTX *ctx);
+	STACK_OF(X509) * (*lookup_certs)(X509_STORE_CTX *ctx, X509_NAME *nm);
+	STACK_OF(X509_CRL) * (*lookup_crls)(X509_STORE_CTX *ctx, X509_NAME *nm);
 	int (*cleanup)(X509_STORE_CTX *ctx);
 
 	/* The following is built up */
@@ -263,6 +269,11 @@ struct x509_store_ctx_st      /* X509_STORE_CTX */
 	X509 *current_issuer;	/* cert currently being tested as valid issuer */
 	X509_CRL *current_crl;	/* current CRL */
 
+	int current_crl_score;  /* score of current CRL */
+	unsigned int current_reasons;  /* Reason mask */
+
+	X509_STORE_CTX *parent; /* For CRL path validation: parent context */
+
 	CRYPTO_EX_DATA ex_data;
 	} /* X509_STORE_CTX */;
 
@@ -330,8 +341,18 @@ void X509_STORE_CTX_set_depth(X509_STORE_CTX *ctx, int depth);
 #define		X509_V_ERR_INVALID_EXTENSION			41
 #define		X509_V_ERR_INVALID_POLICY_EXTENSION		42
 #define		X509_V_ERR_NO_EXPLICIT_POLICY			43
+#define		X509_V_ERR_DIFFERENT_CRL_SCOPE			44
+#define		X509_V_ERR_UNSUPPORTED_EXTENSION_FEATURE	45
+
+#define		X509_V_ERR_UNNESTED_RESOURCE			46
 
-#define		X509_V_ERR_UNNESTED_RESOURCE			44
+#define		X509_V_ERR_PERMITTED_VIOLATION			47
+#define		X509_V_ERR_EXCLUDED_VIOLATION			48
+#define		X509_V_ERR_SUBTREE_MINMAX			49
+#define		X509_V_ERR_UNSUPPORTED_CONSTRAINT_TYPE		51
+#define		X509_V_ERR_UNSUPPORTED_CONSTRAINT_SYNTAX	52
+#define		X509_V_ERR_UNSUPPORTED_NAME_SYNTAX		53
+#define		X509_V_ERR_CRL_PATH_VALIDATION_ERROR		54
 
 /* The application is not happy */
 #define		X509_V_ERR_APPLICATION_VERIFICATION		50
@@ -362,10 +383,14 @@ void X509_STORE_CTX_set_depth(X509_STORE_CTX *ctx, int depth);
 #define X509_V_FLAG_INHIBIT_MAP			0x400
 /* Notify callback that policy is OK */
 #define X509_V_FLAG_NOTIFY_POLICY		0x800
-
+/* Extended CRL features such as indirect CRLs, alternate CRL signing keys */
+#define X509_V_FLAG_EXTENDED_CRL_SUPPORT	0x1000
+/* Delta CRL support */
+#define X509_V_FLAG_USE_DELTAS			0x2000
 /* Check selfsigned CA signature */
 #define X509_V_FLAG_CHECK_SS_SIGNATURE		0x4000
 
+
 #define X509_VP_FLAG_DEFAULT			0x1
 #define X509_VP_FLAG_OVERWRITE			0x2
 #define X509_VP_FLAG_RESET_FLAGS		0x4
@@ -387,11 +412,16 @@ void X509_OBJECT_free_contents(X509_OBJECT *a);
 X509_STORE *X509_STORE_new(void );
 void X509_STORE_free(X509_STORE *v);
 
+STACK_OF(X509)* X509_STORE_get1_certs(X509_STORE_CTX *st, X509_NAME *nm);
+STACK_OF(X509_CRL)* X509_STORE_get1_crls(X509_STORE_CTX *st, X509_NAME *nm);
 int X509_STORE_set_flags(X509_STORE *ctx, unsigned long flags);
 int X509_STORE_set_purpose(X509_STORE *ctx, int purpose);
 int X509_STORE_set_trust(X509_STORE *ctx, int trust);
 int X509_STORE_set1_param(X509_STORE *ctx, X509_VERIFY_PARAM *pm);
 
+void X509_STORE_set_verify_cb(X509_STORE *ctx,
+				  int (*verify_cb)(int, X509_STORE_CTX *));
+
 X509_STORE_CTX *X509_STORE_CTX_new(void);
 
 int X509_STORE_CTX_get1_issuer(X509 **issuer, X509_STORE_CTX *ctx, X509 *x);
@@ -450,6 +480,9 @@ int	X509_STORE_CTX_get_error(X509_STORE_CTX *ctx);
 void	X509_STORE_CTX_set_error(X509_STORE_CTX *ctx,int s);
 int	X509_STORE_CTX_get_error_depth(X509_STORE_CTX *ctx);
 X509 *	X509_STORE_CTX_get_current_cert(X509_STORE_CTX *ctx);
+X509 *X509_STORE_CTX_get0_current_issuer(X509_STORE_CTX *ctx);
+X509_CRL *X509_STORE_CTX_get0_current_crl(X509_STORE_CTX *ctx);
+X509_STORE_CTX *X509_STORE_CTX_get0_parent_ctx(X509_STORE_CTX *ctx);
 STACK_OF(X509) *X509_STORE_CTX_get_chain(X509_STORE_CTX *ctx);
 STACK_OF(X509) *X509_STORE_CTX_get1_chain(X509_STORE_CTX *ctx);
 void	X509_STORE_CTX_set_cert(X509_STORE_CTX *c,X509 *x);
diff --git a/crypto/x509/x509_vpm.c b/crypto/x509/x509_vpm.c
index 01c5541e2e88..dfd89d89faf0 100644
--- a/crypto/x509/x509_vpm.c
+++ b/crypto/x509/x509_vpm.c
@@ -74,6 +74,7 @@ static void x509_verify_param_zero(X509_VERIFY_PARAM *param)
 	param->name = NULL;
 	param->purpose = 0;
 	param->trust = 0;
+	/*param->inh_flags = X509_VP_FLAG_DEFAULT;*/
 	param->inh_flags = 0;
 	param->flags = 0;
 	param->depth = -1;
@@ -328,7 +329,7 @@ static const X509_VERIFY_PARAM default_table[] = {
 	NULL		/* policies */
 	},
 	{
-	"pkcs7",			/* S/MIME signing parameters */
+	"pkcs7",			/* S/MIME sign parameters */
 	0,				/* Check time */
 	0,				/* internal flags */
 	0,				/* flags */
@@ -338,7 +339,7 @@ static const X509_VERIFY_PARAM default_table[] = {
 	NULL				/* policies */
 	},
 	{
-	"smime_sign",			/* S/MIME signing parameters */
+	"smime_sign",			/* S/MIME sign parameters */
 	0,				/* Check time */
 	0,				/* internal flags */
 	0,				/* flags */
@@ -370,12 +371,17 @@ static const X509_VERIFY_PARAM default_table[] = {
 
 static STACK_OF(X509_VERIFY_PARAM) *param_table = NULL;
 
-static int table_cmp(const void *pa, const void *pb)
+static int table_cmp(const X509_VERIFY_PARAM *a, const X509_VERIFY_PARAM *b)
+
 	{
-	const X509_VERIFY_PARAM *a = pa, *b = pb;
 	return strcmp(a->name, b->name);
 	}
 
+DECLARE_OBJ_BSEARCH_CMP_FN(X509_VERIFY_PARAM, X509_VERIFY_PARAM,
+			   table);
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(X509_VERIFY_PARAM, X509_VERIFY_PARAM,
+			     table);
+
 static int param_cmp(const X509_VERIFY_PARAM * const *a,
 			const X509_VERIFY_PARAM * const *b)
 	{
@@ -411,6 +417,7 @@ const X509_VERIFY_PARAM *X509_VERIFY_PARAM_lookup(const char *name)
 	{
 	int idx;
 	X509_VERIFY_PARAM pm;
+
 	pm.name = (char *)name;
 	if (param_table)
 		{
@@ -418,11 +425,8 @@ const X509_VERIFY_PARAM *X509_VERIFY_PARAM_lookup(const char *name)
 		if (idx != -1)
 			return sk_X509_VERIFY_PARAM_value(param_table, idx);
 		}
-	return (const X509_VERIFY_PARAM *) OBJ_bsearch((char *)&pm,
-				(char *)&default_table,
-				sizeof(default_table)/sizeof(X509_VERIFY_PARAM),
-				sizeof(X509_VERIFY_PARAM),
-				table_cmp);
+	return OBJ_bsearch_table(&pm, default_table,
+			   sizeof(default_table)/sizeof(X509_VERIFY_PARAM));
 	}
 
 void X509_VERIFY_PARAM_table_cleanup(void)
diff --git a/crypto/x509/x509cset.c b/crypto/x509/x509cset.c
index 7f4004b2911a..3109defb0b2d 100644
--- a/crypto/x509/x509cset.c
+++ b/crypto/x509/x509cset.c
@@ -81,7 +81,7 @@ int X509_CRL_set_issuer_name(X509_CRL *x, X509_NAME *name)
 	}
 
 
-int X509_CRL_set_lastUpdate(X509_CRL *x, ASN1_TIME *tm)
+int X509_CRL_set_lastUpdate(X509_CRL *x, const ASN1_TIME *tm)
 	{
 	ASN1_TIME *in;
 
@@ -99,7 +99,7 @@ int X509_CRL_set_lastUpdate(X509_CRL *x, ASN1_TIME *tm)
 	return(in != NULL);
 	}
 
-int X509_CRL_set_nextUpdate(X509_CRL *x, ASN1_TIME *tm)
+int X509_CRL_set_nextUpdate(X509_CRL *x, const ASN1_TIME *tm)
 	{
 	ASN1_TIME *in;
 
diff --git a/crypto/x509/x509name.c b/crypto/x509/x509name.c
index 068abfe5f045..27bc4dc9a3e3 100644
--- a/crypto/x509/x509name.c
+++ b/crypto/x509/x509name.c
@@ -356,7 +356,7 @@ int X509_NAME_ENTRY_set_data(X509_NAME_ENTRY *ne, int type,
 		return ASN1_STRING_set_by_NID(&ne->value, bytes,
 						len, type,
 					OBJ_obj2nid(ne->object)) ? 1 : 0;
-	if (len < 0) len=strlen((char *)bytes);
+	if (len < 0) len=strlen((const char *)bytes);
 	i=ASN1_STRING_set(ne->value,bytes,len);
 	if (!i) return(0);
 	if (type != V_ASN1_UNDEF)
diff --git a/crypto/x509/x509type.c b/crypto/x509/x509type.c
index 2cd994c5b0f5..9702ec53103f 100644
--- a/crypto/x509/x509type.c
+++ b/crypto/x509/x509type.c
@@ -91,25 +91,35 @@ int X509_certificate_type(X509 *x, EVP_PKEY *pkey)
 		break;
 	case EVP_PKEY_DH:
 		ret=EVP_PK_DH|EVP_PKT_EXCH;
+		break;	
+	case NID_id_GostR3410_94:
+	case NID_id_GostR3410_2001:
+		ret=EVP_PKT_EXCH|EVP_PKT_SIGN;
 		break;
 	default:
 		break;
 		}
 
-	i=X509_get_signature_type(x);
-	switch (i)
+	i=OBJ_obj2nid(x->sig_alg->algorithm);
+	if (i && OBJ_find_sigid_algs(i, NULL, &i))
 		{
-	case EVP_PKEY_RSA:
-		ret|=EVP_PKS_RSA;
-		break;
-	case EVP_PKEY_DSA:
-		ret|=EVP_PKS_DSA;
-		break;
-	case EVP_PKEY_EC:
-		ret|=EVP_PKS_EC;
-		break;
-	default:
-		break;
+
+		switch (i)
+			{
+		case NID_rsaEncryption:
+		case NID_rsa:
+			ret|=EVP_PKS_RSA;
+			break;
+		case NID_dsa:
+		case NID_dsa_2:
+			ret|=EVP_PKS_DSA;
+			break;
+		case NID_X9_62_id_ecPublicKey:
+			ret|=EVP_PKS_EC;
+			break;
+		default:
+			break;
+			}
 		}
 
 	if (EVP_PKEY_size(pk) <= 1024/8)/* /8 because it's 1024 bits we look
diff --git a/crypto/x509/x_all.c b/crypto/x509/x_all.c
index c7b07f7c9d58..b94aeeb873fd 100644
--- a/crypto/x509/x_all.c
+++ b/crypto/x509/x_all.c
@@ -57,7 +57,6 @@
  */
 
 #include <stdio.h>
-#undef SSLEAY_MACROS
 #include <openssl/stack.h>
 #include "cryptlib.h"
 #include <openssl/buffer.h>
@@ -83,12 +82,6 @@ int X509_REQ_verify(X509_REQ *a, EVP_PKEY *r)
 		a->sig_alg,a->signature,a->req_info,r));
 	}
 
-int X509_CRL_verify(X509_CRL *a, EVP_PKEY *r)
-	{
-	return(ASN1_item_verify(ASN1_ITEM_rptr(X509_CRL_INFO),
-		a->sig_alg, a->signature,a->crl,r));
-	}
-
 int NETSCAPE_SPKI_verify(NETSCAPE_SPKI *a, EVP_PKEY *r)
 	{
 	return(ASN1_item_verify(ASN1_ITEM_rptr(NETSCAPE_SPKAC),
@@ -102,12 +95,25 @@ int X509_sign(X509 *x, EVP_PKEY *pkey, const EVP_MD *md)
 		x->sig_alg, x->signature, x->cert_info,pkey,md));
 	}
 
+int X509_sign_ctx(X509 *x, EVP_MD_CTX *ctx)
+	{
+	return ASN1_item_sign_ctx(ASN1_ITEM_rptr(X509_CINF),
+		x->cert_info->signature,
+		x->sig_alg, x->signature, x->cert_info, ctx);
+	}
+
 int X509_REQ_sign(X509_REQ *x, EVP_PKEY *pkey, const EVP_MD *md)
 	{
 	return(ASN1_item_sign(ASN1_ITEM_rptr(X509_REQ_INFO),x->sig_alg, NULL,
 		x->signature, x->req_info,pkey,md));
 	}
 
+int X509_REQ_sign_ctx(X509_REQ *x, EVP_MD_CTX *ctx)
+	{
+	return ASN1_item_sign_ctx(ASN1_ITEM_rptr(X509_REQ_INFO),
+		x->sig_alg, NULL, x->signature, x->req_info, ctx);
+	}
+
 int X509_CRL_sign(X509_CRL *x, EVP_PKEY *pkey, const EVP_MD *md)
 	{
 	x->crl->enc.modified = 1;
@@ -115,6 +121,12 @@ int X509_CRL_sign(X509_CRL *x, EVP_PKEY *pkey, const EVP_MD *md)
 		x->sig_alg, x->signature, x->crl,pkey,md));
 	}
 
+int X509_CRL_sign_ctx(X509_CRL *x, EVP_MD_CTX *ctx)
+	{
+	return ASN1_item_sign_ctx(ASN1_ITEM_rptr(X509_CRL_INFO),
+		x->crl->sig_alg, x->sig_alg, x->signature, x->crl, ctx);
+	}
+
 int NETSCAPE_SPKI_sign(NETSCAPE_SPKI *x, EVP_PKEY *pkey, const EVP_MD *md)
 	{
 	return(ASN1_item_sign(ASN1_ITEM_rptr(NETSCAPE_SPKAC), x->sig_algor,NULL,
diff --git a/crypto/x509v3/Makefile b/crypto/x509v3/Makefile
index e71dc42f9f71..556ef351bf88 100644
--- a/crypto/x509v3/Makefile
+++ b/crypto/x509v3/Makefile
@@ -43,7 +43,7 @@ top:
 all:	lib
 
 lib:	$(LIBOBJ)
-	$(ARX) $(LIB) $(LIBOBJ)
+	$(AR) $(LIB) $(LIBOBJ)
 	$(RANLIB) $(LIB) || echo Never mind.
 	@touch lib
 
@@ -90,8 +90,8 @@ pcy_cache.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 pcy_cache.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pcy_cache.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pcy_cache.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pcy_cache.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pcy_cache.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+pcy_cache.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pcy_cache.o: ../../include/openssl/objects.h
 pcy_cache.o: ../../include/openssl/opensslconf.h
 pcy_cache.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pcy_cache.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
@@ -105,9 +105,8 @@ pcy_data.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 pcy_data.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pcy_data.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pcy_data.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pcy_data.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pcy_data.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pcy_data.o: ../../include/openssl/opensslconf.h
+pcy_data.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pcy_data.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 pcy_data.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pcy_data.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 pcy_data.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -120,36 +119,35 @@ pcy_lib.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 pcy_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pcy_lib.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pcy_lib.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pcy_lib.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pcy_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pcy_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-pcy_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-pcy_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-pcy_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-pcy_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-pcy_lib.o: ../../include/openssl/x509v3.h ../cryptlib.h pcy_int.h pcy_lib.c
+pcy_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pcy_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pcy_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+pcy_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+pcy_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+pcy_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+pcy_lib.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+pcy_lib.o: ../cryptlib.h pcy_int.h pcy_lib.c
 pcy_map.o: ../../e_os.h ../../include/openssl/asn1.h
 pcy_map.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 pcy_map.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 pcy_map.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pcy_map.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pcy_map.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pcy_map.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pcy_map.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pcy_map.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-pcy_map.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-pcy_map.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-pcy_map.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-pcy_map.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-pcy_map.o: ../../include/openssl/x509v3.h ../cryptlib.h pcy_int.h pcy_map.c
+pcy_map.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pcy_map.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+pcy_map.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+pcy_map.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+pcy_map.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+pcy_map.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+pcy_map.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+pcy_map.o: ../cryptlib.h pcy_int.h pcy_map.c
 pcy_node.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 pcy_node.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 pcy_node.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 pcy_node.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 pcy_node.o: ../../include/openssl/ecdsa.h ../../include/openssl/evp.h
-pcy_node.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pcy_node.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pcy_node.o: ../../include/openssl/opensslconf.h
+pcy_node.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pcy_node.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 pcy_node.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pcy_node.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 pcy_node.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -162,9 +160,8 @@ pcy_tree.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 pcy_tree.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 pcy_tree.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 pcy_tree.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-pcy_tree.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-pcy_tree.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-pcy_tree.o: ../../include/openssl/opensslconf.h
+pcy_tree.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+pcy_tree.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 pcy_tree.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 pcy_tree.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 pcy_tree.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -177,39 +174,37 @@ v3_addr.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_addr.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_addr.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_addr.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_addr.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_addr.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_addr.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_addr.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_addr.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_addr.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_addr.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_addr.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_addr.o: ../cryptlib.h v3_addr.c
+v3_addr.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_addr.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_addr.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_addr.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_addr.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_addr.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_addr.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_addr.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_addr.c
 v3_akey.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_akey.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_akey.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_akey.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_akey.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_akey.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_akey.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_akey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_akey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_akey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_akey.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_akey.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_akey.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_akey.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_akey.o: ../cryptlib.h v3_akey.c
+v3_akey.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_akey.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_akey.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_akey.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_akey.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_akey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_akey.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_akey.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_akey.c
 v3_akeya.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_akeya.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_akeya.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_akeya.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_akeya.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_akeya.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_akeya.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_akeya.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_akeya.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_akeya.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_akeya.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_akeya.o: ../../include/openssl/opensslconf.h
 v3_akeya.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 v3_akeya.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 v3_akeya.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -221,15 +216,14 @@ v3_alt.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_alt.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_alt.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_alt.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_alt.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_alt.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_alt.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_alt.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_alt.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_alt.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_alt.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_alt.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_alt.o: ../cryptlib.h v3_alt.c
+v3_alt.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_alt.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_alt.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_alt.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_alt.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_alt.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_alt.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_alt.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_alt.c
 v3_asid.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_asid.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_asid.o: ../../include/openssl/bn.h ../../include/openssl/buffer.h
@@ -237,23 +231,23 @@ v3_asid.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 v3_asid.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 v3_asid.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 v3_asid.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-v3_asid.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-v3_asid.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-v3_asid.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-v3_asid.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-v3_asid.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-v3_asid.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-v3_asid.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-v3_asid.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_asid.c
+v3_asid.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+v3_asid.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_asid.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+v3_asid.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+v3_asid.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+v3_asid.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+v3_asid.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+v3_asid.o: ../cryptlib.h v3_asid.c
 v3_bcons.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_bcons.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_bcons.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_bcons.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_bcons.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_bcons.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_bcons.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_bcons.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_bcons.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_bcons.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_bcons.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_bcons.o: ../../include/openssl/opensslconf.h
 v3_bcons.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 v3_bcons.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 v3_bcons.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -266,9 +260,8 @@ v3_bitst.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 v3_bitst.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 v3_bitst.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 v3_bitst.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-v3_bitst.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-v3_bitst.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-v3_bitst.o: ../../include/openssl/opensslconf.h
+v3_bitst.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+v3_bitst.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 v3_bitst.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 v3_bitst.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 v3_bitst.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -281,23 +274,23 @@ v3_conf.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 v3_conf.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 v3_conf.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 v3_conf.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-v3_conf.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-v3_conf.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-v3_conf.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-v3_conf.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-v3_conf.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-v3_conf.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-v3_conf.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-v3_conf.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_conf.c
+v3_conf.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+v3_conf.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_conf.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+v3_conf.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+v3_conf.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+v3_conf.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+v3_conf.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+v3_conf.o: ../cryptlib.h v3_conf.c
 v3_cpols.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_cpols.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_cpols.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_cpols.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_cpols.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_cpols.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_cpols.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_cpols.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_cpols.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_cpols.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_cpols.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_cpols.o: ../../include/openssl/opensslconf.h
 v3_cpols.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 v3_cpols.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 v3_cpols.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -310,38 +303,37 @@ v3_crld.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_crld.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_crld.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_crld.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_crld.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_crld.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_crld.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_crld.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_crld.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_crld.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_crld.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_crld.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_crld.o: ../cryptlib.h v3_crld.c
+v3_crld.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_crld.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_crld.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_crld.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_crld.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_crld.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_crld.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_crld.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_crld.c
 v3_enum.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_enum.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 v3_enum.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 v3_enum.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 v3_enum.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 v3_enum.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-v3_enum.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-v3_enum.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-v3_enum.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-v3_enum.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-v3_enum.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-v3_enum.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-v3_enum.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-v3_enum.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_enum.c
+v3_enum.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+v3_enum.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_enum.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+v3_enum.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+v3_enum.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+v3_enum.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+v3_enum.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+v3_enum.o: ../cryptlib.h v3_enum.c
 v3_extku.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_extku.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_extku.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_extku.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_extku.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_extku.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_extku.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_extku.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_extku.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_extku.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_extku.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_extku.o: ../../include/openssl/opensslconf.h
 v3_extku.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 v3_extku.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 v3_extku.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -354,81 +346,76 @@ v3_genn.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_genn.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_genn.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_genn.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_genn.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_genn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_genn.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_genn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_genn.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_genn.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_genn.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_genn.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_genn.o: ../cryptlib.h v3_genn.c
+v3_genn.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_genn.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_genn.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_genn.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_genn.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_genn.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_genn.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_genn.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_genn.c
 v3_ia5.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 v3_ia5.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_ia5.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_ia5.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_ia5.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_ia5.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_ia5.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_ia5.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_ia5.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_ia5.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_ia5.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_ia5.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_ia5.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_ia5.o: ../cryptlib.h v3_ia5.c
+v3_ia5.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_ia5.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_ia5.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_ia5.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_ia5.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_ia5.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_ia5.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_ia5.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_ia5.c
 v3_info.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_info.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_info.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_info.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_info.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_info.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_info.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_info.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_info.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_info.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_info.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_info.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_info.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_info.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_info.o: ../cryptlib.h v3_info.c
+v3_info.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_info.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_info.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_info.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_info.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_info.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_info.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_info.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_info.c
 v3_int.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 v3_int.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_int.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_int.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_int.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_int.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_int.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_int.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_int.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_int.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_int.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_int.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_int.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_int.o: ../cryptlib.h v3_int.c
+v3_int.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_int.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_int.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_int.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_int.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_int.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_int.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_int.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_int.c
 v3_lib.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 v3_lib.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_lib.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_lib.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_lib.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_lib.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_lib.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_lib.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_lib.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_lib.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_lib.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_lib.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_lib.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_lib.o: ../cryptlib.h ext_dat.h v3_lib.c
+v3_lib.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_lib.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_lib.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_lib.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_lib.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_lib.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_lib.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_lib.o: ../../include/openssl/x509v3.h ../cryptlib.h ext_dat.h v3_lib.c
 v3_ncons.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_ncons.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_ncons.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_ncons.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_ncons.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_ncons.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_ncons.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_ncons.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_ncons.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_ncons.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_ncons.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_ncons.o: ../../include/openssl/opensslconf.h
 v3_ncons.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 v3_ncons.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 v3_ncons.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -441,52 +428,49 @@ v3_ocsp.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 v3_ocsp.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 v3_ocsp.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 v3_ocsp.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-v3_ocsp.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-v3_ocsp.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-v3_ocsp.o: ../../include/openssl/ocsp.h ../../include/openssl/opensslconf.h
-v3_ocsp.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_ocsp.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_ocsp.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_ocsp.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_ocsp.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_ocsp.o: ../cryptlib.h v3_ocsp.c
+v3_ocsp.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+v3_ocsp.o: ../../include/openssl/objects.h ../../include/openssl/ocsp.h
+v3_ocsp.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_ocsp.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_ocsp.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_ocsp.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_ocsp.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_ocsp.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_ocsp.c
 v3_pci.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 v3_pci.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_pci.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_pci.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_pci.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_pci.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_pci.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_pci.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_pci.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_pci.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_pci.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_pci.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_pci.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_pci.o: ../cryptlib.h v3_pci.c
+v3_pci.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_pci.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_pci.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_pci.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_pci.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_pci.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_pci.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_pci.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_pci.c
 v3_pcia.o: ../../include/openssl/asn1.h ../../include/openssl/asn1t.h
 v3_pcia.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 v3_pcia.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 v3_pcia.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 v3_pcia.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
-v3_pcia.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_pcia.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_pcia.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_pcia.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_pcia.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_pcia.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_pcia.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_pcia.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_pcia.o: v3_pcia.c
+v3_pcia.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_pcia.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_pcia.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_pcia.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_pcia.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_pcia.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_pcia.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_pcia.o: ../../include/openssl/x509v3.h v3_pcia.c
 v3_pcons.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_pcons.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_pcons.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_pcons.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_pcons.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_pcons.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_pcons.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_pcons.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_pcons.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_pcons.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_pcons.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_pcons.o: ../../include/openssl/opensslconf.h
 v3_pcons.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 v3_pcons.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 v3_pcons.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -499,24 +483,23 @@ v3_pku.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_pku.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_pku.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_pku.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_pku.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_pku.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_pku.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_pku.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_pku.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_pku.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_pku.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_pku.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_pku.o: ../cryptlib.h v3_pku.c
+v3_pku.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_pku.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_pku.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_pku.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_pku.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_pku.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_pku.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_pku.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_pku.c
 v3_pmaps.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_pmaps.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_pmaps.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_pmaps.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_pmaps.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_pmaps.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_pmaps.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_pmaps.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_pmaps.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_pmaps.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_pmaps.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_pmaps.o: ../../include/openssl/opensslconf.h
 v3_pmaps.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 v3_pmaps.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 v3_pmaps.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -528,52 +511,51 @@ v3_prn.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_prn.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_prn.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_prn.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_prn.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_prn.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_prn.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3_prn.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3_prn.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3_prn.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3_prn.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3_prn.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3_prn.o: ../cryptlib.h v3_prn.c
+v3_prn.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_prn.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_prn.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3_prn.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3_prn.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3_prn.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3_prn.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3_prn.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_prn.c
 v3_purp.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_purp.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 v3_purp.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 v3_purp.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 v3_purp.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 v3_purp.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-v3_purp.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-v3_purp.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-v3_purp.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-v3_purp.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-v3_purp.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-v3_purp.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-v3_purp.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-v3_purp.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_purp.c
+v3_purp.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+v3_purp.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_purp.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+v3_purp.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+v3_purp.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+v3_purp.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+v3_purp.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+v3_purp.o: ../cryptlib.h v3_purp.c
 v3_skey.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_skey.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 v3_skey.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 v3_skey.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 v3_skey.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 v3_skey.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-v3_skey.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-v3_skey.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-v3_skey.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-v3_skey.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-v3_skey.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-v3_skey.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-v3_skey.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-v3_skey.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_skey.c
+v3_skey.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+v3_skey.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_skey.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+v3_skey.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+v3_skey.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+v3_skey.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+v3_skey.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+v3_skey.o: ../cryptlib.h v3_skey.c
 v3_sxnet.o: ../../e_os.h ../../include/openssl/asn1.h
 v3_sxnet.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 v3_sxnet.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3_sxnet.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3_sxnet.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3_sxnet.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3_sxnet.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3_sxnet.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3_sxnet.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_sxnet.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3_sxnet.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3_sxnet.o: ../../include/openssl/opensslconf.h
 v3_sxnet.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 v3_sxnet.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
 v3_sxnet.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
@@ -586,25 +568,24 @@ v3_utl.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
 v3_utl.o: ../../include/openssl/e_os2.h ../../include/openssl/ec.h
 v3_utl.o: ../../include/openssl/ecdh.h ../../include/openssl/ecdsa.h
 v3_utl.o: ../../include/openssl/err.h ../../include/openssl/evp.h
-v3_utl.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
-v3_utl.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
-v3_utl.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
-v3_utl.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
-v3_utl.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
-v3_utl.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-v3_utl.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
-v3_utl.o: ../../include/openssl/x509v3.h ../cryptlib.h v3_utl.c
+v3_utl.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+v3_utl.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+v3_utl.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+v3_utl.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+v3_utl.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+v3_utl.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+v3_utl.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+v3_utl.o: ../cryptlib.h v3_utl.c
 v3err.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
 v3err.o: ../../include/openssl/buffer.h ../../include/openssl/conf.h
 v3err.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 v3err.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
 v3err.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
-v3err.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
-v3err.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
-v3err.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
-v3err.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
-v3err.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
-v3err.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
-v3err.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
-v3err.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
-v3err.o: v3err.c
+v3err.o: ../../include/openssl/evp.h ../../include/openssl/lhash.h
+v3err.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+v3err.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+v3err.o: ../../include/openssl/ossl_typ.h ../../include/openssl/pkcs7.h
+v3err.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
+v3err.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+v3err.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
+v3err.o: ../../include/openssl/x509v3.h v3err.c
diff --git a/crypto/x509v3/ext_dat.h b/crypto/x509v3/ext_dat.h
index 3eaec46f8a6d..76daee6fcde6 100644
--- a/crypto/x509v3/ext_dat.h
+++ b/crypto/x509v3/ext_dat.h
@@ -61,21 +61,19 @@ extern X509V3_EXT_METHOD v3_bcons, v3_nscert, v3_key_usage, v3_ext_ku;
 extern X509V3_EXT_METHOD v3_pkey_usage_period, v3_sxnet, v3_info, v3_sinfo;
 extern X509V3_EXT_METHOD v3_ns_ia5_list[], v3_alt[], v3_skey_id, v3_akey_id;
 extern X509V3_EXT_METHOD v3_crl_num, v3_crl_reason, v3_crl_invdate;
-extern X509V3_EXT_METHOD v3_delta_crl, v3_cpols, v3_crld;
+extern X509V3_EXT_METHOD v3_delta_crl, v3_cpols, v3_crld, v3_freshest_crl;
 extern X509V3_EXT_METHOD v3_ocsp_nonce, v3_ocsp_accresp, v3_ocsp_acutoff;
 extern X509V3_EXT_METHOD v3_ocsp_crlid, v3_ocsp_nocheck, v3_ocsp_serviceloc;
 extern X509V3_EXT_METHOD v3_crl_hold, v3_pci;
 extern X509V3_EXT_METHOD v3_policy_mappings, v3_policy_constraints;
-extern X509V3_EXT_METHOD v3_name_constraints, v3_inhibit_anyp;
-#ifndef OPENSSL_NO_RFC3779
+extern X509V3_EXT_METHOD v3_name_constraints, v3_inhibit_anyp, v3_idp;
 extern X509V3_EXT_METHOD v3_addr, v3_asid;
-#endif
 
 /* This table will be searched using OBJ_bsearch so it *must* kept in
  * order of the ext_nid values.
  */
 
-static X509V3_EXT_METHOD *standard_exts[] = {
+static const X509V3_EXT_METHOD *standard_exts[] = {
 &v3_nscert,
 &v3_ns_ia5_list[0],
 &v3_ns_ia5_list[1],
@@ -122,7 +120,10 @@ static X509V3_EXT_METHOD *standard_exts[] = {
 &v3_pci,
 &v3_name_constraints,
 &v3_policy_mappings,
-&v3_inhibit_anyp
+&v3_inhibit_anyp,
+&v3_idp,
+&v3_alt[2],
+&v3_freshest_crl,
 };
 
 /* Number of standard extensions */
diff --git a/crypto/x509v3/pcy_cache.c b/crypto/x509v3/pcy_cache.c
index 1030931b7136..172b7e7ee4f4 100644
--- a/crypto/x509v3/pcy_cache.c
+++ b/crypto/x509v3/pcy_cache.c
@@ -139,7 +139,6 @@ static int policy_cache_new(X509 *x)
 		return 0;
 	cache->anyPolicy = NULL;
 	cache->data = NULL;
-	cache->maps = NULL;
 	cache->any_skip = -1;
 	cache->explicit_skip = -1;
 	cache->map_skip = -1;
diff --git a/crypto/x509v3/pcy_data.c b/crypto/x509v3/pcy_data.c
index fb392b901f0d..3444b0319505 100644
--- a/crypto/x509v3/pcy_data.c
+++ b/crypto/x509v3/pcy_data.c
@@ -82,17 +82,21 @@ void policy_data_free(X509_POLICY_DATA *data)
  * another source.
  */
 
-X509_POLICY_DATA *policy_data_new(POLICYINFO *policy, ASN1_OBJECT *id, int crit)
+X509_POLICY_DATA *policy_data_new(POLICYINFO *policy,
+					const ASN1_OBJECT *cid, int crit)
 	{
 	X509_POLICY_DATA *ret;
-	if (!policy && !id)
+	ASN1_OBJECT *id;
+	if (!policy && !cid)
 		return NULL;
-	if (id)
+	if (cid)
 		{
-		id = OBJ_dup(id);
+		id = OBJ_dup(cid);
 		if (!id)
 			return NULL;
 		}
+	else
+		id = NULL;
 	ret = OPENSSL_malloc(sizeof(X509_POLICY_DATA));
 	if (!ret)
 		return NULL;
diff --git a/crypto/x509v3/pcy_int.h b/crypto/x509v3/pcy_int.h
index 3780de4fcdbd..ccff92846e40 100644
--- a/crypto/x509v3/pcy_int.h
+++ b/crypto/x509v3/pcy_int.h
@@ -56,12 +56,10 @@
  *
  */
 
-DECLARE_STACK_OF(X509_POLICY_DATA)
-DECLARE_STACK_OF(X509_POLICY_REF)
-DECLARE_STACK_OF(X509_POLICY_NODE)
 
 typedef struct X509_POLICY_DATA_st X509_POLICY_DATA;
-typedef struct X509_POLICY_REF_st X509_POLICY_REF;
+
+DECLARE_STACK_OF(X509_POLICY_DATA)
 
 /* Internal structures */
 
@@ -110,16 +108,6 @@ struct X509_POLICY_DATA_st
 
 #define POLICY_DATA_FLAG_CRITICAL		0x10
 
-/* This structure is an entry from a table of mapped policies which
- * cross reference the policy it refers to.
- */
-
-struct X509_POLICY_REF_st
-	{
-	ASN1_OBJECT *subjectDomainPolicy;
-	const X509_POLICY_DATA *data;
-	};
-
 /* This structure is cached with a certificate */
 
 struct X509_POLICY_CACHE_st {
@@ -127,8 +115,6 @@ struct X509_POLICY_CACHE_st {
 	X509_POLICY_DATA *anyPolicy;
 	/* other policy data */
 	STACK_OF(X509_POLICY_DATA) *data;
-	/* If policyMappings extension present a table of mapped policies */
-	STACK_OF(X509_POLICY_REF) *maps;
 	/* If InhibitAnyPolicy present this is its value or -1 if absent. */
 	long any_skip;
 	/* If policyConstraints and requireExplicitPolicy present this is its
@@ -193,7 +179,7 @@ struct X509_POLICY_TREE_st
 
 /* Internal functions */
 
-X509_POLICY_DATA *policy_data_new(POLICYINFO *policy, ASN1_OBJECT *id,
+X509_POLICY_DATA *policy_data_new(POLICYINFO *policy, const ASN1_OBJECT *id,
 								int crit);
 void policy_data_free(X509_POLICY_DATA *data);
 
@@ -209,15 +195,18 @@ void policy_cache_init(void);
 void policy_cache_free(X509_POLICY_CACHE *cache);
 
 X509_POLICY_NODE *level_find_node(const X509_POLICY_LEVEL *level,
+					const X509_POLICY_NODE *parent,	
 					const ASN1_OBJECT *id);
 
 X509_POLICY_NODE *tree_find_sk(STACK_OF(X509_POLICY_NODE) *sk,
 						const ASN1_OBJECT *id);
 
 X509_POLICY_NODE *level_add_node(X509_POLICY_LEVEL *level,
-			X509_POLICY_DATA *data,
+			const X509_POLICY_DATA *data,
 			X509_POLICY_NODE *parent,
 			X509_POLICY_TREE *tree);
 void policy_node_free(X509_POLICY_NODE *node);
+int policy_node_match(const X509_POLICY_LEVEL *lvl,
+		      const X509_POLICY_NODE *node, const ASN1_OBJECT *oid);
 
 const X509_POLICY_CACHE *policy_cache_set(X509 *x);
diff --git a/crypto/x509v3/pcy_map.c b/crypto/x509v3/pcy_map.c
index acd2ede6f34c..21163b529d4b 100644
--- a/crypto/x509v3/pcy_map.c
+++ b/crypto/x509v3/pcy_map.c
@@ -62,29 +62,6 @@
 
 #include "pcy_int.h"
 
-static int ref_cmp(const X509_POLICY_REF * const *a,
-			const X509_POLICY_REF * const *b)
-	{
-	return OBJ_cmp((*a)->subjectDomainPolicy, (*b)->subjectDomainPolicy);
-	}
-
-static void policy_map_free(X509_POLICY_REF *map)
-	{
-	OPENSSL_free(map);
-	}
-
-static X509_POLICY_REF *policy_map_find(X509_POLICY_CACHE *cache, ASN1_OBJECT *id)
-	{
-	X509_POLICY_REF tmp;
-	int idx;
-	tmp.subjectDomainPolicy = id;
-
-	idx = sk_X509_POLICY_REF_find(cache->maps, &tmp);
-	if (idx == -1)
-		return NULL;
-	return sk_X509_POLICY_REF_value(cache->maps, idx);
-	}
-
 /* Set policy mapping entries in cache.
  * Note: this modifies the passed POLICY_MAPPINGS structure
  */
@@ -92,8 +69,6 @@ static X509_POLICY_REF *policy_map_find(X509_POLICY_CACHE *cache, ASN1_OBJECT *i
 int policy_cache_set_mapping(X509 *x, POLICY_MAPPINGS *maps)
 	{
 	POLICY_MAPPING *map;
-	X509_POLICY_REF *ref = NULL;
-	ASN1_OBJECT *subjectDomainPolicyRef;
 	X509_POLICY_DATA *data;
 	X509_POLICY_CACHE *cache = x->policy_cache;
 	int i;
@@ -103,7 +78,6 @@ int policy_cache_set_mapping(X509 *x, POLICY_MAPPINGS *maps)
 		ret = -1;
 		goto bad_mapping;
 		}
-	cache->maps = sk_X509_POLICY_REF_new(ref_cmp);
 	for (i = 0; i < sk_POLICY_MAPPING_num(maps); i++)
 		{
 		map = sk_POLICY_MAPPING_value(maps, i);
@@ -115,13 +89,6 @@ int policy_cache_set_mapping(X509 *x, POLICY_MAPPINGS *maps)
 			goto bad_mapping;
 			}
 
-		/* If we've already mapped from this OID bad mapping */
-		if (policy_map_find(cache, map->subjectDomainPolicy) != NULL)
-			{
-			ret = -1;
-			goto bad_mapping;
-			}
-
 		/* Attempt to find matching policy data */
 		data = policy_cache_find_data(cache, map->issuerDomainPolicy);
 		/* If we don't have anyPolicy can't map */
@@ -137,7 +104,7 @@ int policy_cache_set_mapping(X509 *x, POLICY_MAPPINGS *maps)
 			if (!data)
 				goto bad_mapping;
 			data->qualifier_set = cache->anyPolicy->qualifier_set;
-			map->issuerDomainPolicy = NULL;
+			/*map->issuerDomainPolicy = NULL;*/
 			data->flags |= POLICY_DATA_FLAG_MAPPED_ANY;
 			data->flags |= POLICY_DATA_FLAG_SHARED_QUALIFIERS;
 			if (!sk_X509_POLICY_DATA_push(cache->data, data))
@@ -148,26 +115,10 @@ int policy_cache_set_mapping(X509 *x, POLICY_MAPPINGS *maps)
 			}
 		else
 			data->flags |= POLICY_DATA_FLAG_MAPPED;
-
 		if (!sk_ASN1_OBJECT_push(data->expected_policy_set, 
 						map->subjectDomainPolicy))
 			goto bad_mapping;
-                /* map->subjectDomainPolicy will be freed when
-                 * cache->data is freed. Set it to NULL to avoid double-free. */
-                subjectDomainPolicyRef = map->subjectDomainPolicy;
-                map->subjectDomainPolicy = NULL;
-		
-		ref = OPENSSL_malloc(sizeof(X509_POLICY_REF));
-		if (!ref)
-			goto bad_mapping;
-
-		ref->subjectDomainPolicy = subjectDomainPolicyRef;
-		ref->data = data;
-
-		if (!sk_X509_POLICY_REF_push(cache->maps, ref))
-			goto bad_mapping;
-
-		ref = NULL;
+		map->subjectDomainPolicy = NULL;
 
 		}
 
@@ -175,13 +126,6 @@ int policy_cache_set_mapping(X509 *x, POLICY_MAPPINGS *maps)
 	bad_mapping:
 	if (ret == -1)
 		x->ex_flags |= EXFLAG_INVALID_POLICY;
-	if (ref)
-		policy_map_free(ref);
-	if (ret <= 0)
-		{
-		sk_X509_POLICY_REF_pop_free(cache->maps, policy_map_free);
-		cache->maps = NULL;
-		}
 	sk_POLICY_MAPPING_pop_free(maps, POLICY_MAPPING_free);
 	return ret;
 
diff --git a/crypto/x509v3/pcy_node.c b/crypto/x509v3/pcy_node.c
index 6587cb05abad..bd1e7f1ae8b7 100644
--- a/crypto/x509v3/pcy_node.c
+++ b/crypto/x509v3/pcy_node.c
@@ -92,13 +92,25 @@ X509_POLICY_NODE *tree_find_sk(STACK_OF(X509_POLICY_NODE) *nodes,
 	}
 
 X509_POLICY_NODE *level_find_node(const X509_POLICY_LEVEL *level,
+					const X509_POLICY_NODE *parent,	
 					const ASN1_OBJECT *id)
 	{
-	return tree_find_sk(level->nodes, id);
+	X509_POLICY_NODE *node;
+	int i;
+	for (i = 0; i < sk_X509_POLICY_NODE_num(level->nodes); i++)
+		{
+		node = sk_X509_POLICY_NODE_value(level->nodes, i);
+		if (node->parent == parent)
+			{
+			if (!OBJ_cmp(node->data->valid_policy, id))
+				return node;
+			}
+		}
+	return NULL;
 	}
 
 X509_POLICY_NODE *level_add_node(X509_POLICY_LEVEL *level,
-			X509_POLICY_DATA *data,
+			const X509_POLICY_DATA *data,
 			X509_POLICY_NODE *parent,
 			X509_POLICY_TREE *tree)
 	{
@@ -155,4 +167,31 @@ void policy_node_free(X509_POLICY_NODE *node)
 	OPENSSL_free(node);
 	}
 
+/* See if a policy node matches a policy OID. If mapping enabled look through
+ * expected policy set otherwise just valid policy.
+ */
+
+int policy_node_match(const X509_POLICY_LEVEL *lvl,
+		      const X509_POLICY_NODE *node, const ASN1_OBJECT *oid)
+	{
+	int i;
+	ASN1_OBJECT *policy_oid;
+	const X509_POLICY_DATA *x = node->data;
+
+	if (	    (lvl->flags & X509_V_FLAG_INHIBIT_MAP)
+		|| !(x->flags & POLICY_DATA_FLAG_MAP_MASK))
+		{
+		if (!OBJ_cmp(x->valid_policy, oid))
+			return 1;
+		return 0;
+		}
+
+	for (i = 0; i < sk_ASN1_OBJECT_num(x->expected_policy_set); i++)
+		{
+		policy_oid = sk_ASN1_OBJECT_value(x->expected_policy_set, i);
+		if (!OBJ_cmp(policy_oid, oid))
+			return 1;
+		}
+	return 0;
 
+	}
diff --git a/crypto/x509v3/pcy_tree.c b/crypto/x509v3/pcy_tree.c
index 92ad0a2b3991..bb9777348f8b 100644
--- a/crypto/x509v3/pcy_tree.c
+++ b/crypto/x509v3/pcy_tree.c
@@ -62,6 +62,75 @@
 
 #include "pcy_int.h"
 
+/* Enable this to print out the complete policy tree at various point during
+ * evaluation.
+ */
+
+/*#define OPENSSL_POLICY_DEBUG*/
+
+#ifdef OPENSSL_POLICY_DEBUG
+
+static void expected_print(BIO *err, X509_POLICY_LEVEL *lev,
+				X509_POLICY_NODE *node, int indent)
+	{
+	if (	    (lev->flags & X509_V_FLAG_INHIBIT_MAP)
+		|| !(node->data->flags & POLICY_DATA_FLAG_MAP_MASK))
+		BIO_puts(err, "  Not Mapped\n");
+	else
+		{
+		int i;
+		STACK_OF(ASN1_OBJECT) *pset = node->data->expected_policy_set;
+		ASN1_OBJECT *oid;
+		BIO_puts(err, "  Expected: ");
+		for (i = 0; i < sk_ASN1_OBJECT_num(pset); i++)
+			{
+			oid = sk_ASN1_OBJECT_value(pset, i);
+			if (i)
+				BIO_puts(err, ", ");
+			i2a_ASN1_OBJECT(err, oid);
+			}
+		BIO_puts(err, "\n");
+		}
+	}
+
+static void tree_print(char *str, X509_POLICY_TREE *tree,
+			X509_POLICY_LEVEL *curr)
+	{
+	X509_POLICY_LEVEL *plev;
+	X509_POLICY_NODE *node;
+	int i;
+	BIO *err;
+	err = BIO_new_fp(stderr, BIO_NOCLOSE);
+	if (!curr)
+		curr = tree->levels + tree->nlevel;
+	else
+		curr++;
+	BIO_printf(err, "Level print after %s\n", str);
+	BIO_printf(err, "Printing Up to Level %ld\n", curr - tree->levels);
+	for (plev = tree->levels; plev != curr; plev++)
+		{
+		BIO_printf(err, "Level %ld, flags = %x\n",
+				plev - tree->levels, plev->flags);
+		for (i = 0; i < sk_X509_POLICY_NODE_num(plev->nodes); i++)
+			{
+			node = sk_X509_POLICY_NODE_value(plev->nodes, i);
+			X509_POLICY_NODE_print(err, node, 2);
+			expected_print(err, plev, node, 2);
+			BIO_printf(err, "  Flags: %x\n", node->data->flags);
+			}
+		if (plev->anyPolicy)
+			X509_POLICY_NODE_print(err, plev->anyPolicy, 2);
+		}
+
+	BIO_free(err);
+
+	}
+#else
+
+#define tree_print(a,b,c) /* */
+
+#endif
+
 /* Initialize policy tree. Return values:
  *  0 Some internal error occured.
  * -1 Inconsistent or invalid extensions in certificates.
@@ -87,8 +156,10 @@ static int tree_init(X509_POLICY_TREE **ptree, STACK_OF(X509) *certs,
 	*ptree = NULL;
 	n = sk_X509_num(certs);
 
+#if 0
 	/* Disable policy mapping for now... */
 	flags |= X509_V_FLAG_INHIBIT_MAP;
+#endif
 
 	if (flags & X509_V_FLAG_EXPLICIT_POLICY)
 		explicit_policy = 0;
@@ -184,7 +255,6 @@ static int tree_init(X509_POLICY_TREE **ptree, STACK_OF(X509) *certs,
 		level++;
 		x = sk_X509_value(certs, i);
 		cache = policy_cache_set(x);
-
 		CRYPTO_add(&x->references, 1, CRYPTO_LOCK_X509);
 		level->cert = x;
 
@@ -213,13 +283,13 @@ static int tree_init(X509_POLICY_TREE **ptree, STACK_OF(X509) *certs,
 			level->flags |= X509_V_FLAG_INHIBIT_MAP;
 		else
 			{
-			map_skip--;
+			if (!(x->ex_flags & EXFLAG_SI))
+				map_skip--;
 			if ((cache->map_skip >= 0)
 				&& (cache->map_skip < map_skip))
 				map_skip = cache->map_skip;
 			}
 
-
 		}
 
 	*ptree = tree;
@@ -237,7 +307,32 @@ static int tree_init(X509_POLICY_TREE **ptree, STACK_OF(X509) *certs,
 
 	}
 
-/* This corresponds to RFC3280 XXXX XXXXX:
+static int tree_link_matching_nodes(X509_POLICY_LEVEL *curr,
+				const X509_POLICY_DATA *data)
+	{
+	X509_POLICY_LEVEL *last = curr - 1;
+	X509_POLICY_NODE *node;
+	int i, matched = 0;
+	/* Iterate through all in nodes linking matches */
+	for (i = 0; i < sk_X509_POLICY_NODE_num(last->nodes); i++)
+		{
+		node = sk_X509_POLICY_NODE_value(last->nodes, i);
+		if (policy_node_match(last, node, data->valid_policy))
+			{
+			if (!level_add_node(curr, data, node, NULL))
+				return 0;
+			matched = 1;
+			}
+		}
+	if (!matched && last->anyPolicy)
+		{
+		if (!level_add_node(curr, data, last->anyPolicy, NULL))
+			return 0;
+		}
+	return 1;
+	}
+
+/* This corresponds to RFC3280 6.1.3(d)(1):
  * link any data from CertificatePolicies onto matching parent
  * or anyPolicy if no match.
  */
@@ -246,10 +341,8 @@ static int tree_link_nodes(X509_POLICY_LEVEL *curr,
 				const X509_POLICY_CACHE *cache)
 	{
 	int i;
-	X509_POLICY_LEVEL *last;
 	X509_POLICY_DATA *data;
-	X509_POLICY_NODE *parent;
-	last = curr - 1;
+
 	for (i = 0; i < sk_X509_POLICY_DATA_num(cache->data); i++)
 		{
 		data = sk_X509_POLICY_DATA_value(cache->data, i);
@@ -261,40 +354,109 @@ static int tree_link_nodes(X509_POLICY_LEVEL *curr,
 		 * link because then it will have the mapping flags
 		 * right and we can prune it later.
 		 */
+#if 0
 		if ((data->flags & POLICY_DATA_FLAG_MAPPED_ANY)
 			&& !(curr->flags & X509_V_FLAG_INHIBIT_ANY))
 			continue;
-		/* Look for matching node in parent */
-		parent = level_find_node(last, data->valid_policy);
-		/* If no match link to anyPolicy */
-		if (!parent)
-			parent = last->anyPolicy;
-		if (parent && !level_add_node(curr, data, parent, NULL))
+#endif
+		/* Look for matching nodes in previous level */
+		if (!tree_link_matching_nodes(curr, data))
 				return 0;
 		}
 	return 1;
 	}
 
-/* This corresponds to RFC3280 XXXX XXXXX:
+/* This corresponds to RFC3280 6.1.3(d)(2):
  * Create new data for any unmatched policies in the parent and link
  * to anyPolicy.
  */
 
+static int tree_add_unmatched(X509_POLICY_LEVEL *curr,
+			const X509_POLICY_CACHE *cache,
+			const ASN1_OBJECT *id,
+			X509_POLICY_NODE *node,
+			X509_POLICY_TREE *tree)
+	{
+	X509_POLICY_DATA *data;
+	if (id == NULL)
+		id = node->data->valid_policy;
+	/* Create a new node with qualifiers from anyPolicy and
+	 * id from unmatched node.
+	 */
+	data = policy_data_new(NULL, id, node_critical(node));
+
+	if (data == NULL)
+		return 0;
+	/* Curr may not have anyPolicy */
+	data->qualifier_set = cache->anyPolicy->qualifier_set;
+	data->flags |= POLICY_DATA_FLAG_SHARED_QUALIFIERS;
+	if (!level_add_node(curr, data, node, tree))
+		{
+		policy_data_free(data);
+		return 0;
+		}
+
+	return 1;
+	}
+
+static int tree_link_unmatched(X509_POLICY_LEVEL *curr,
+			const X509_POLICY_CACHE *cache,
+			X509_POLICY_NODE *node,
+			X509_POLICY_TREE *tree)
+	{
+	const X509_POLICY_LEVEL *last = curr - 1;
+	int i;
+
+	if (	    (last->flags & X509_V_FLAG_INHIBIT_MAP)
+		|| !(node->data->flags & POLICY_DATA_FLAG_MAPPED))
+		{
+		/* If no policy mapping: matched if one child present */
+		if (node->nchild)
+			return 1;
+		if (!tree_add_unmatched(curr, cache, NULL, node, tree))
+			return 0;
+		/* Add it */
+		}
+	else
+		{
+		/* If mapping: matched if one child per expected policy set */
+		STACK_OF(ASN1_OBJECT) *expset = node->data->expected_policy_set;
+		if (node->nchild == sk_ASN1_OBJECT_num(expset))
+			return 1;
+		/* Locate unmatched nodes */
+		for (i = 0; i < sk_ASN1_OBJECT_num(expset); i++)
+			{
+			ASN1_OBJECT *oid = sk_ASN1_OBJECT_value(expset, i);
+			if (level_find_node(curr, node, oid))
+				continue;
+			if (!tree_add_unmatched(curr, cache, oid, node, tree))
+				return 0;
+			}
+
+		}
+
+	return 1;
+
+	}
+
 static int tree_link_any(X509_POLICY_LEVEL *curr,
 			const X509_POLICY_CACHE *cache,
 			X509_POLICY_TREE *tree)
 	{
 	int i;
-	X509_POLICY_DATA *data;
+	/*X509_POLICY_DATA *data;*/
 	X509_POLICY_NODE *node;
-	X509_POLICY_LEVEL *last;
-
-	last = curr - 1;
+	X509_POLICY_LEVEL *last = curr - 1;
 
 	for (i = 0; i < sk_X509_POLICY_NODE_num(last->nodes); i++)
 		{
 		node = sk_X509_POLICY_NODE_value(last->nodes, i);
 
+		if (!tree_link_unmatched(curr, cache, node, tree))
+			return 0;
+
+#if 0
+
 		/* Skip any node with any children: we only want unmathced
 		 * nodes.
 		 *
@@ -303,6 +465,7 @@ static int tree_link_any(X509_POLICY_LEVEL *curr,
 		 */
 		if (node->nchild)
 			continue;
+
 		/* Create a new node with qualifiers from anyPolicy and
 		 * id from unmatched node.
 		 */
@@ -319,6 +482,9 @@ static int tree_link_any(X509_POLICY_LEVEL *curr,
 			policy_data_free(data);
 			return 0;
 			}
+
+#endif
+
 		}
 	/* Finally add link to anyPolicy */
 	if (last->anyPolicy)
@@ -337,30 +503,36 @@ static int tree_link_any(X509_POLICY_LEVEL *curr,
 
 static int tree_prune(X509_POLICY_TREE *tree, X509_POLICY_LEVEL *curr)
 	{
+	STACK_OF(X509_POLICY_NODE) *nodes;
 	X509_POLICY_NODE *node;
 	int i;
-	for (i = sk_X509_POLICY_NODE_num(curr->nodes) - 1; i >= 0; i--)
+	nodes = curr->nodes;
+	if (curr->flags & X509_V_FLAG_INHIBIT_MAP)
 		{
-		node = sk_X509_POLICY_NODE_value(curr->nodes, i);
-		/* Delete any mapped data: see RFC3280 XXXX */
-		if (node->data->flags & POLICY_DATA_FLAG_MAP_MASK)
+		for (i = sk_X509_POLICY_NODE_num(nodes) - 1; i >= 0; i--)
 			{
-			node->parent->nchild--;
-			OPENSSL_free(node);
-			(void)sk_X509_POLICY_NODE_delete(curr->nodes, i);
+			node = sk_X509_POLICY_NODE_value(nodes, i);
+			/* Delete any mapped data: see RFC3280 XXXX */
+			if (node->data->flags & POLICY_DATA_FLAG_MAP_MASK)
+				{
+				node->parent->nchild--;
+				OPENSSL_free(node);
+				(void)sk_X509_POLICY_NODE_delete(nodes,i);
+				}
 			}
 		}
 
 	for(;;)	{
 		--curr;
-		for (i = sk_X509_POLICY_NODE_num(curr->nodes) - 1; i >= 0; i--)
+		nodes = curr->nodes;
+		for (i = sk_X509_POLICY_NODE_num(nodes) - 1; i >= 0; i--)
 			{
-			node = sk_X509_POLICY_NODE_value(curr->nodes, i);
+			node = sk_X509_POLICY_NODE_value(nodes, i);
 			if (node->nchild == 0)
 				{
 				node->parent->nchild--;
 				OPENSSL_free(node);
-				(void)sk_X509_POLICY_NODE_delete(curr->nodes, i);
+				(void)sk_X509_POLICY_NODE_delete(nodes, i);
 				}
 			}
 		if (curr->anyPolicy && !curr->anyPolicy->nchild)
@@ -536,6 +708,7 @@ static int tree_evaluate(X509_POLICY_TREE *tree)
 		if (!(curr->flags & X509_V_FLAG_INHIBIT_ANY)
 			&& !tree_link_any(curr, cache, tree))
 			return 0;
+	tree_print("before tree_prune()", tree, curr);
 		ret = tree_prune(tree, curr);
 		if (ret != 1)
 			return ret;
@@ -604,7 +777,6 @@ int X509_policy_check(X509_POLICY_TREE **ptree, int *pexplicit_policy,
 	*pexplicit_policy = 0;
 	ret = tree_init(&tree, certs, flags);
 
-
 	switch (ret)
 		{
 
@@ -612,7 +784,7 @@ int X509_policy_check(X509_POLICY_TREE **ptree, int *pexplicit_policy,
 		case 2:
 		return 1;
 
-                /* Some internal error */
+		/* Some internal error */
 		case -1:
 		return -1;
 
@@ -650,6 +822,8 @@ int X509_policy_check(X509_POLICY_TREE **ptree, int *pexplicit_policy,
 	if (!tree) goto error;
 	ret = tree_evaluate(tree);
 
+	tree_print("tree_evaluate()", tree, NULL);
+
 	if (ret <= 0)
 		goto error;
 
@@ -695,3 +869,4 @@ int X509_policy_check(X509_POLICY_TREE **ptree, int *pexplicit_policy,
 	return 0;
 
 	}
+
diff --git a/crypto/x509v3/v3_addr.c b/crypto/x509v3/v3_addr.c
index c0e1d2d142e0..df46a4983be2 100644
--- a/crypto/x509v3/v3_addr.c
+++ b/crypto/x509v3/v3_addr.c
@@ -242,7 +242,7 @@ static int i2r_IPAddressOrRanges(BIO *out,
 /*
  * i2r handler for an IPAddrBlocks extension.
  */
-static int i2r_IPAddrBlocks(X509V3_EXT_METHOD *method,
+static int i2r_IPAddrBlocks(const X509V3_EXT_METHOD *method,
 			    void *ext,
 			    BIO *out,
 			    int indent)
@@ -327,8 +327,7 @@ static int IPAddressOrRange_cmp(const IPAddressOrRange *a,
 				const int length)
 {
   unsigned char addr_a[ADDR_RAW_BUF_LEN], addr_b[ADDR_RAW_BUF_LEN];
-  int prefixlen_a = 0;
-  int prefixlen_b = 0;
+  int prefixlen_a = 0, prefixlen_b = 0;
   int r;
 
   switch (a->type) {
@@ -613,10 +612,10 @@ static IPAddressOrRanges *make_prefix_or_range(IPAddrBlocks *addr,
     return NULL;
   switch (afi) {
   case IANA_AFI_IPV4:
-    (void)sk_IPAddressOrRange_set_cmp_func(aors, v4IPAddressOrRange_cmp);
+    (void) sk_IPAddressOrRange_set_cmp_func(aors, v4IPAddressOrRange_cmp);
     break;
   case IANA_AFI_IPV6:
-    (void)sk_IPAddressOrRange_set_cmp_func(aors, v6IPAddressOrRange_cmp);
+    (void) sk_IPAddressOrRange_set_cmp_func(aors, v6IPAddressOrRange_cmp);
     break;
   }
   f->ipAddressChoice->type = IPAddressChoice_addressesOrRanges;
@@ -885,8 +884,8 @@ static int IPAddressOrRanges_canonize(IPAddressOrRanges *aors,
       IPAddressOrRange *merged;
       if (!make_addressRange(&merged, a_min, b_max, length))
 	return 0;
-      sk_IPAddressOrRange_set(aors, i, merged);
-      (void)sk_IPAddressOrRange_delete(aors, i + 1);
+      (void) sk_IPAddressOrRange_set(aors, i, merged);
+      (void) sk_IPAddressOrRange_delete(aors, i + 1);
       IPAddressOrRange_free(a);
       IPAddressOrRange_free(b);
       --i;
@@ -924,7 +923,7 @@ int v3_addr_canonize(IPAddrBlocks *addr)
 				    v3_addr_get_afi(f)))
       return 0;
   }
-  (void)sk_IPAddressFamily_set_cmp_func(addr, IPAddressFamily_cmp);
+  (void) sk_IPAddressFamily_set_cmp_func(addr, IPAddressFamily_cmp);
   sk_IPAddressFamily_sort(addr);
   OPENSSL_assert(v3_addr_is_canonical(addr));
   return 1;
@@ -933,7 +932,7 @@ int v3_addr_canonize(IPAddrBlocks *addr)
 /*
  * v2i handler for the IPAddrBlocks extension.
  */
-static void *v2i_IPAddrBlocks(struct v3_ext_method *method,
+static void *v2i_IPAddrBlocks(const struct v3_ext_method *method,
 			      struct v3_ext_ctx *ctx,
 			      STACK_OF(CONF_VALUE) *values)
 {
@@ -1176,7 +1175,7 @@ int v3_addr_subset(IPAddrBlocks *a, IPAddrBlocks *b)
     return 1;
   if (b == NULL || v3_addr_inherits(a) || v3_addr_inherits(b))
     return 0;
-  (void)sk_IPAddressFamily_set_cmp_func(b, IPAddressFamily_cmp);
+  (void) sk_IPAddressFamily_set_cmp_func(b, IPAddressFamily_cmp);
   for (i = 0; i < sk_IPAddressFamily_num(a); i++) {
     IPAddressFamily *fa = sk_IPAddressFamily_value(a, i);
     int j = sk_IPAddressFamily_find(b, fa);
@@ -1218,7 +1217,7 @@ static int v3_addr_validate_path_internal(X509_STORE_CTX *ctx,
 {
   IPAddrBlocks *child = NULL;
   int i, j, ret = 1;
-  X509 *x = NULL;
+  X509 *x;
 
   OPENSSL_assert(chain != NULL && sk_X509_num(chain) > 0);
   OPENSSL_assert(ctx != NULL || ext != NULL);
@@ -1231,6 +1230,7 @@ static int v3_addr_validate_path_internal(X509_STORE_CTX *ctx,
    */
   if (ext != NULL) {
     i = -1;
+    x = NULL;
   } else {
     i = 0;
     x = sk_X509_value(chain, i);
@@ -1240,7 +1240,7 @@ static int v3_addr_validate_path_internal(X509_STORE_CTX *ctx,
   }
   if (!v3_addr_is_canonical(ext))
     validation_err(X509_V_ERR_INVALID_EXTENSION);
-  (void)sk_IPAddressFamily_set_cmp_func(ext, IPAddressFamily_cmp);
+  (void) sk_IPAddressFamily_set_cmp_func(ext, IPAddressFamily_cmp);
   if ((child = sk_IPAddressFamily_dup(ext)) == NULL) {
     X509V3err(X509V3_F_V3_ADDR_VALIDATE_PATH_INTERNAL, ERR_R_MALLOC_FAILURE);
     ret = 0;
@@ -1266,7 +1266,7 @@ static int v3_addr_validate_path_internal(X509_STORE_CTX *ctx,
       }
       continue;
     }
-    (void)sk_IPAddressFamily_set_cmp_func(x->rfc3779_addr, IPAddressFamily_cmp);
+    (void) sk_IPAddressFamily_set_cmp_func(x->rfc3779_addr, IPAddressFamily_cmp);
     for (j = 0; j < sk_IPAddressFamily_num(child); j++) {
       IPAddressFamily *fc = sk_IPAddressFamily_value(child, j);
       int k = sk_IPAddressFamily_find(x->rfc3779_addr, fc);
@@ -1293,6 +1293,7 @@ static int v3_addr_validate_path_internal(X509_STORE_CTX *ctx,
   /*
    * Trust anchor can't inherit.
    */
+  OPENSSL_assert(x != NULL);
   if (x->rfc3779_addr != NULL) {
     for (j = 0; j < sk_IPAddressFamily_num(x->rfc3779_addr); j++) {
       IPAddressFamily *fp = sk_IPAddressFamily_value(x->rfc3779_addr, j);
diff --git a/crypto/x509v3/v3_alt.c b/crypto/x509v3/v3_alt.c
index 69244e4976a5..d29d94338ec5 100644
--- a/crypto/x509v3/v3_alt.c
+++ b/crypto/x509v3/v3_alt.c
@@ -82,6 +82,12 @@ NULL, NULL, NULL},
 (X509V3_EXT_I2V)i2v_GENERAL_NAMES,
 (X509V3_EXT_V2I)v2i_issuer_alt,
 NULL, NULL, NULL},
+
+{ NID_certificate_issuer, 0, ASN1_ITEM_ref(GENERAL_NAMES),
+0,0,0,0,
+0,0,
+(X509V3_EXT_I2V)i2v_GENERAL_NAMES,
+NULL, NULL, NULL, NULL},
 };
 
 STACK_OF(CONF_VALUE) *i2v_GENERAL_NAMES(X509V3_EXT_METHOD *method,
@@ -387,8 +393,8 @@ static int copy_email(X509V3_CTX *ctx, GENERAL_NAMES *gens, int move_p)
 	
 }
 
-GENERAL_NAMES *v2i_GENERAL_NAMES(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
+GENERAL_NAMES *v2i_GENERAL_NAMES(const X509V3_EXT_METHOD *method,
+				 X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
 {
 	GENERAL_NAME *gen;
 	GENERAL_NAMES *gens = NULL;
@@ -409,28 +415,22 @@ GENERAL_NAMES *v2i_GENERAL_NAMES(X509V3_EXT_METHOD *method,
 	return NULL;
 }
 
-GENERAL_NAME *v2i_GENERAL_NAME(X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
-							 CONF_VALUE *cnf)
+GENERAL_NAME *v2i_GENERAL_NAME(const X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
+			       CONF_VALUE *cnf)
 	{
 	return v2i_GENERAL_NAME_ex(NULL, method, ctx, cnf, 0);
 	}
 
-GENERAL_NAME *v2i_GENERAL_NAME_ex(GENERAL_NAME *out,
-				X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
-						 CONF_VALUE *cnf, int is_nc)
+GENERAL_NAME *a2i_GENERAL_NAME(GENERAL_NAME *out,
+			       const X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
+			       int gen_type, char *value, int is_nc)
 	{
 	char is_string = 0;
-	int type;
 	GENERAL_NAME *gen = NULL;
 
-	char *name, *value;
-
-	name = cnf->name;
-	value = cnf->value;
-
 	if(!value)
 		{
-		X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,X509V3_R_MISSING_VALUE);
+		X509V3err(X509V3_F_A2I_GENERAL_NAME,X509V3_R_MISSING_VALUE);
 		return NULL;
 		}
 
@@ -441,74 +441,62 @@ GENERAL_NAME *v2i_GENERAL_NAME_ex(GENERAL_NAME *out,
 		gen = GENERAL_NAME_new();
 		if(gen == NULL)
 			{
-			X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,ERR_R_MALLOC_FAILURE);
+			X509V3err(X509V3_F_A2I_GENERAL_NAME,ERR_R_MALLOC_FAILURE);
 			return NULL;
 			}
 		}
 
-	if(!name_cmp(name, "email"))
-		{
-		is_string = 1;
-		type = GEN_EMAIL;
-		}
-	else if(!name_cmp(name, "URI"))
-		{
-		is_string = 1;
-		type = GEN_URI;
-		}
-	else if(!name_cmp(name, "DNS"))
+	switch (gen_type)
 		{
+		case GEN_URI:
+		case GEN_EMAIL:
+		case GEN_DNS:
 		is_string = 1;
-		type = GEN_DNS;
-		}
-	else if(!name_cmp(name, "RID"))
+		break;
+		
+		case GEN_RID:
 		{
 		ASN1_OBJECT *obj;
 		if(!(obj = OBJ_txt2obj(value,0)))
 			{
-			X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,X509V3_R_BAD_OBJECT);
+			X509V3err(X509V3_F_A2I_GENERAL_NAME,X509V3_R_BAD_OBJECT);
 			ERR_add_error_data(2, "value=", value);
 			goto err;
 			}
 		gen->d.rid = obj;
-		type = GEN_RID;
 		}
-	else if(!name_cmp(name, "IP"))
-		{
+		break;
+
+		case GEN_IPADD:
 		if (is_nc)
 			gen->d.ip = a2i_IPADDRESS_NC(value);
 		else
 			gen->d.ip = a2i_IPADDRESS(value);
 		if(gen->d.ip == NULL)
 			{
-			X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,X509V3_R_BAD_IP_ADDRESS);
+			X509V3err(X509V3_F_A2I_GENERAL_NAME,X509V3_R_BAD_IP_ADDRESS);
 			ERR_add_error_data(2, "value=", value);
 			goto err;
 			}
-		type = GEN_IPADD;
-		}
-	else if(!name_cmp(name, "dirName"))
-		{
-		type = GEN_DIRNAME;
+		break;
+
+		case GEN_DIRNAME:
 		if (!do_dirname(gen, value, ctx))
 			{
-			X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,X509V3_R_DIRNAME_ERROR);
+			X509V3err(X509V3_F_A2I_GENERAL_NAME,X509V3_R_DIRNAME_ERROR);
 			goto err;
 			}
-		}
-	else if(!name_cmp(name, "otherName"))
-		{
+		break;
+
+		case GEN_OTHERNAME:
 		if (!do_othername(gen, value, ctx))
 			{
-			X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,X509V3_R_OTHERNAME_ERROR);
+			X509V3err(X509V3_F_A2I_GENERAL_NAME,X509V3_R_OTHERNAME_ERROR);
 			goto err;
 			}
-		type = GEN_OTHERNAME;
-		}
-	else
-		{
-		X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,X509V3_R_UNSUPPORTED_OPTION);
-		ERR_add_error_data(2, "name=", name);
+		break;
+		default:
+		X509V3err(X509V3_F_A2I_GENERAL_NAME,X509V3_R_UNSUPPORTED_TYPE);
 		goto err;
 		}
 
@@ -518,12 +506,12 @@ GENERAL_NAME *v2i_GENERAL_NAME_ex(GENERAL_NAME *out,
 			      !ASN1_STRING_set(gen->d.ia5, (unsigned char*)value,
 					       strlen(value)))
 			{
-			X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,ERR_R_MALLOC_FAILURE);
+			X509V3err(X509V3_F_A2I_GENERAL_NAME,ERR_R_MALLOC_FAILURE);
 			goto err;
 			}
 		}
 
-	gen->type = type;
+	gen->type = gen_type;
 
 	return gen;
 
@@ -533,6 +521,48 @@ GENERAL_NAME *v2i_GENERAL_NAME_ex(GENERAL_NAME *out,
 	return NULL;
 	}
 
+GENERAL_NAME *v2i_GENERAL_NAME_ex(GENERAL_NAME *out,
+				  const X509V3_EXT_METHOD *method,
+				  X509V3_CTX *ctx, CONF_VALUE *cnf, int is_nc)
+	{
+	int type;
+
+	char *name, *value;
+
+	name = cnf->name;
+	value = cnf->value;
+
+	if(!value)
+		{
+		X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,X509V3_R_MISSING_VALUE);
+		return NULL;
+		}
+
+	if(!name_cmp(name, "email"))
+		type = GEN_EMAIL;
+	else if(!name_cmp(name, "URI"))
+		type = GEN_URI;
+	else if(!name_cmp(name, "DNS"))
+		type = GEN_DNS;
+	else if(!name_cmp(name, "RID"))
+		type = GEN_RID;
+	else if(!name_cmp(name, "IP"))
+		type = GEN_IPADD;
+	else if(!name_cmp(name, "dirName"))
+		type = GEN_DIRNAME;
+	else if(!name_cmp(name, "otherName"))
+		type = GEN_OTHERNAME;
+	else
+		{
+		X509V3err(X509V3_F_V2I_GENERAL_NAME_EX,X509V3_R_UNSUPPORTED_OPTION);
+		ERR_add_error_data(2, "name=", name);
+		return NULL;
+		}
+
+	return a2i_GENERAL_NAME(out, method, ctx, type, value, is_nc);
+
+	}
+
 static int do_othername(GENERAL_NAME *gen, char *value, X509V3_CTX *ctx)
 	{
 	char *objtmp = NULL, *p;
@@ -578,7 +608,6 @@ static int do_dirname(GENERAL_NAME *gen, char *value, X509V3_CTX *ctx)
 	if (!ret)
 		X509_NAME_free(nm);
 	gen->d.dirn = nm;
-
 	X509V3_section_free(ctx, sk);
 		
 	return ret;
diff --git a/crypto/x509v3/v3_asid.c b/crypto/x509v3/v3_asid.c
index 11aad0b8b438..1587e8ed7267 100644
--- a/crypto/x509v3/v3_asid.c
+++ b/crypto/x509v3/v3_asid.c
@@ -151,7 +151,7 @@ static int i2r_ASIdentifierChoice(BIO *out,
 /*
  * i2r method for an ASIdentifier extension.
  */
-static int i2r_ASIdentifiers(X509V3_EXT_METHOD *method,
+static int i2r_ASIdentifiers(const X509V3_EXT_METHOD *method,
 			     void *ext,
 			     BIO *out,
 			     int indent)
@@ -538,7 +538,7 @@ int v3_asid_canonize(ASIdentifiers *asid)
 /*
  * v2i method for an ASIdentifier extension.
  */
-static void *v2i_ASIdentifiers(struct v3_ext_method *method,
+static void *v2i_ASIdentifiers(const struct v3_ext_method *method,
 			       struct v3_ext_ctx *ctx,
 			       STACK_OF(CONF_VALUE) *values)
 {
@@ -753,7 +753,7 @@ static int v3_asid_validate_path_internal(X509_STORE_CTX *ctx,
 {
   ASIdOrRanges *child_as = NULL, *child_rdi = NULL;
   int i, ret = 1, inherit_as = 0, inherit_rdi = 0;
-  X509 *x = NULL;
+  X509 *x;
 
   OPENSSL_assert(chain != NULL && sk_X509_num(chain) > 0);
   OPENSSL_assert(ctx != NULL || ext != NULL);
@@ -766,6 +766,7 @@ static int v3_asid_validate_path_internal(X509_STORE_CTX *ctx,
    */
   if (ext != NULL) {
     i = -1;
+    x = NULL;
   } else {
     i = 0;
     x = sk_X509_value(chain, i);
@@ -845,6 +846,7 @@ static int v3_asid_validate_path_internal(X509_STORE_CTX *ctx,
   /*
    * Trust anchor can't inherit.
    */
+  OPENSSL_assert(x != NULL);
   if (x->rfc3779_asid != NULL) {
     if (x->rfc3779_asid->asnum != NULL &&
 	x->rfc3779_asid->asnum->type == ASIdentifierChoice_inherit)
diff --git a/crypto/x509v3/v3_conf.c b/crypto/x509v3/v3_conf.c
index 11eb6b7fd5f7..6730f9a6ee96 100644
--- a/crypto/x509v3/v3_conf.c
+++ b/crypto/x509v3/v3_conf.c
@@ -72,14 +72,14 @@ static X509_EXTENSION *do_ext_nconf(CONF *conf, X509V3_CTX *ctx, int ext_nid, in
 static X509_EXTENSION *v3_generic_extension(const char *ext, char *value, int crit, int type, X509V3_CTX *ctx);
 static char *conf_lhash_get_string(void *db, char *section, char *value);
 static STACK_OF(CONF_VALUE) *conf_lhash_get_section(void *db, char *section);
-static X509_EXTENSION *do_ext_i2d(X509V3_EXT_METHOD *method, int ext_nid,
-						 int crit, void *ext_struc);
+static X509_EXTENSION *do_ext_i2d(const X509V3_EXT_METHOD *method, int ext_nid,
+				  int crit, void *ext_struc);
 static unsigned char *generic_asn1(char *value, X509V3_CTX *ctx, long *ext_len);
 /* CONF *conf:  Config file    */
 /* char *name:  Name    */
 /* char *value:  Value    */
 X509_EXTENSION *X509V3_EXT_nconf(CONF *conf, X509V3_CTX *ctx, char *name,
-	     char *value)
+				 char *value)
 	{
 	int crit;
 	int ext_type;
@@ -99,7 +99,7 @@ X509_EXTENSION *X509V3_EXT_nconf(CONF *conf, X509V3_CTX *ctx, char *name,
 /* CONF *conf:  Config file    */
 /* char *value:  Value    */
 X509_EXTENSION *X509V3_EXT_nconf_nid(CONF *conf, X509V3_CTX *ctx, int ext_nid,
-	     char *value)
+				     char *value)
 	{
 	int crit;
 	int ext_type;
@@ -113,9 +113,9 @@ X509_EXTENSION *X509V3_EXT_nconf_nid(CONF *conf, X509V3_CTX *ctx, int ext_nid,
 /* CONF *conf:  Config file    */
 /* char *value:  Value    */
 static X509_EXTENSION *do_ext_nconf(CONF *conf, X509V3_CTX *ctx, int ext_nid,
-	     int crit, char *value)
+				    int crit, char *value)
 	{
-	X509V3_EXT_METHOD *method;
+	const X509V3_EXT_METHOD *method;
 	X509_EXTENSION *ext;
 	STACK_OF(CONF_VALUE) *nval;
 	void *ext_struc;
@@ -172,8 +172,8 @@ static X509_EXTENSION *do_ext_nconf(CONF *conf, X509V3_CTX *ctx, int ext_nid,
 
 	}
 
-static X509_EXTENSION *do_ext_i2d(X509V3_EXT_METHOD *method, int ext_nid,
-						 int crit, void *ext_struc)
+static X509_EXTENSION *do_ext_i2d(const X509V3_EXT_METHOD *method, int ext_nid,
+				  int crit, void *ext_struc)
 	{
 	unsigned char *ext_der;
 	int ext_len;
@@ -214,7 +214,7 @@ static X509_EXTENSION *do_ext_i2d(X509V3_EXT_METHOD *method, int ext_nid,
 
 X509_EXTENSION *X509V3_EXT_i2d(int ext_nid, int crit, void *ext_struc)
 	{
-	X509V3_EXT_METHOD *method;
+	const X509V3_EXT_METHOD *method;
 	if (!(method = X509V3_EXT_get_nid(ext_nid))) {
 		X509V3err(X509V3_F_X509V3_EXT_I2D,X509V3_R_UNKNOWN_EXTENSION);
 		return NULL;
@@ -258,7 +258,8 @@ static int v3_check_generic(char **value)
 
 /* Create a generic extension: for now just handle DER type */
 static X509_EXTENSION *v3_generic_extension(const char *ext, char *value,
-	     int crit, int gen_type, X509V3_CTX *ctx)
+					    int crit, int gen_type,
+					    X509V3_CTX *ctx)
 	{
 	unsigned char *ext_der=NULL;
 	long ext_len;
@@ -322,7 +323,7 @@ static unsigned char *generic_asn1(char *value, X509V3_CTX *ctx, long *ext_len)
 
 
 int X509V3_EXT_add_nconf_sk(CONF *conf, X509V3_CTX *ctx, char *section,
-	     STACK_OF(X509_EXTENSION) **sk)
+			    STACK_OF(X509_EXTENSION) **sk)
 	{
 	X509_EXTENSION *ext;
 	STACK_OF(CONF_VALUE) *nval;
@@ -343,7 +344,7 @@ int X509V3_EXT_add_nconf_sk(CONF *conf, X509V3_CTX *ctx, char *section,
 /* Convenience functions to add extensions to a certificate, CRL and request */
 
 int X509V3_EXT_add_nconf(CONF *conf, X509V3_CTX *ctx, char *section,
-	     X509 *cert)
+			 X509 *cert)
 	{
 	STACK_OF(X509_EXTENSION) **sk = NULL;
 	if (cert)
@@ -354,7 +355,7 @@ int X509V3_EXT_add_nconf(CONF *conf, X509V3_CTX *ctx, char *section,
 /* Same as above but for a CRL */
 
 int X509V3_EXT_CRL_add_nconf(CONF *conf, X509V3_CTX *ctx, char *section,
-	     X509_CRL *crl)
+			     X509_CRL *crl)
 	{
 	STACK_OF(X509_EXTENSION) **sk = NULL;
 	if (crl)
@@ -443,7 +444,7 @@ void X509V3_set_nconf(X509V3_CTX *ctx, CONF *conf)
 	}
 
 void X509V3_set_ctx(X509V3_CTX *ctx, X509 *issuer, X509 *subj, X509_REQ *req,
-	     X509_CRL *crl, int flags)
+		    X509_CRL *crl, int flags)
 	{
 	ctx->issuer_cert = issuer;
 	ctx->subject_cert = subj;
@@ -454,8 +455,8 @@ void X509V3_set_ctx(X509V3_CTX *ctx, X509 *issuer, X509 *subj, X509_REQ *req,
 
 /* Old conf compatibility functions */
 
-X509_EXTENSION *X509V3_EXT_conf(LHASH *conf, X509V3_CTX *ctx, char *name,
-	     char *value)
+X509_EXTENSION *X509V3_EXT_conf(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+				char *name, char *value)
 	{
 	CONF ctmp;
 	CONF_set_nconf(&ctmp, conf);
@@ -464,8 +465,8 @@ X509_EXTENSION *X509V3_EXT_conf(LHASH *conf, X509V3_CTX *ctx, char *name,
 
 /* LHASH *conf:  Config file    */
 /* char *value:  Value    */
-X509_EXTENSION *X509V3_EXT_conf_nid(LHASH *conf, X509V3_CTX *ctx, int ext_nid,
-	     char *value)
+X509_EXTENSION *X509V3_EXT_conf_nid(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+				    int ext_nid, char *value)
 	{
 	CONF ctmp;
 	CONF_set_nconf(&ctmp, conf);
@@ -489,14 +490,14 @@ NULL,
 NULL
 };
 
-void X509V3_set_conf_lhash(X509V3_CTX *ctx, LHASH *lhash)
+void X509V3_set_conf_lhash(X509V3_CTX *ctx, LHASH_OF(CONF_VALUE) *lhash)
 	{
 	ctx->db_meth = &conf_lhash_method;
 	ctx->db = lhash;
 	}
 
-int X509V3_EXT_add_conf(LHASH *conf, X509V3_CTX *ctx, char *section,
-	     X509 *cert)
+int X509V3_EXT_add_conf(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+			char *section, X509 *cert)
 	{
 	CONF ctmp;
 	CONF_set_nconf(&ctmp, conf);
@@ -505,8 +506,8 @@ int X509V3_EXT_add_conf(LHASH *conf, X509V3_CTX *ctx, char *section,
 
 /* Same as above but for a CRL */
 
-int X509V3_EXT_CRL_add_conf(LHASH *conf, X509V3_CTX *ctx, char *section,
-	     X509_CRL *crl)
+int X509V3_EXT_CRL_add_conf(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+			    char *section, X509_CRL *crl)
 	{
 	CONF ctmp;
 	CONF_set_nconf(&ctmp, conf);
@@ -515,8 +516,8 @@ int X509V3_EXT_CRL_add_conf(LHASH *conf, X509V3_CTX *ctx, char *section,
 
 /* Add extensions to certificate request */
 
-int X509V3_EXT_REQ_add_conf(LHASH *conf, X509V3_CTX *ctx, char *section,
-	     X509_REQ *req)
+int X509V3_EXT_REQ_add_conf(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+			    char *section, X509_REQ *req)
 	{
 	CONF ctmp;
 	CONF_set_nconf(&ctmp, conf);
diff --git a/crypto/x509v3/v3_cpols.c b/crypto/x509v3/v3_cpols.c
index ad0506d75c8e..1f0798b9468f 100644
--- a/crypto/x509v3/v3_cpols.c
+++ b/crypto/x509v3/v3_cpols.c
@@ -450,5 +450,8 @@ void X509_POLICY_NODE_print(BIO *out, X509_POLICY_NODE *node, int indent)
 	else
 		BIO_printf(out, "%*sNo Qualifiers\n", indent + 2, "");
 	}
-	
+
+
 IMPLEMENT_STACK_OF(X509_POLICY_NODE)
+IMPLEMENT_STACK_OF(X509_POLICY_DATA)
+
diff --git a/crypto/x509v3/v3_crld.c b/crypto/x509v3/v3_crld.c
index 181a8977b124..790a6dd03280 100644
--- a/crypto/x509v3/v3_crld.c
+++ b/crypto/x509v3/v3_crld.c
@@ -3,7 +3,7 @@
  * project 1999.
  */
 /* ====================================================================
- * Copyright (c) 1999 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2008 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -63,45 +63,254 @@
 #include <openssl/asn1t.h>
 #include <openssl/x509v3.h>
 
-static STACK_OF(CONF_VALUE) *i2v_crld(X509V3_EXT_METHOD *method,
-		STACK_OF(DIST_POINT) *crld, STACK_OF(CONF_VALUE) *extlist);
-static STACK_OF(DIST_POINT) *v2i_crld(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval);
-
-const X509V3_EXT_METHOD v3_crld = {
-NID_crl_distribution_points, X509V3_EXT_MULTILINE, ASN1_ITEM_ref(CRL_DIST_POINTS),
-0,0,0,0,
-0,0,
-(X509V3_EXT_I2V)i2v_crld,
-(X509V3_EXT_V2I)v2i_crld,
-0,0,
-NULL
+static void *v2i_crld(const X509V3_EXT_METHOD *method,
+		      X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval);
+static int i2r_crldp(const X509V3_EXT_METHOD *method, void *pcrldp, BIO *out,
+		     int indent);
+
+const X509V3_EXT_METHOD v3_crld =
+	{
+	NID_crl_distribution_points, 0, ASN1_ITEM_ref(CRL_DIST_POINTS),
+	0,0,0,0,
+	0,0,
+	0,
+	v2i_crld,
+	i2r_crldp,0,
+	NULL
+	};
+
+const X509V3_EXT_METHOD v3_freshest_crl =
+	{
+	NID_freshest_crl, 0, ASN1_ITEM_ref(CRL_DIST_POINTS),
+	0,0,0,0,
+	0,0,
+	0,
+	v2i_crld,
+	i2r_crldp,0,
+	NULL
+	};
+
+static STACK_OF(GENERAL_NAME) *gnames_from_sectname(X509V3_CTX *ctx, char *sect)
+	{
+	STACK_OF(CONF_VALUE) *gnsect;
+	STACK_OF(GENERAL_NAME) *gens;
+	if (*sect == '@')
+		gnsect = X509V3_get_section(ctx, sect + 1);
+	else
+		gnsect = X509V3_parse_list(sect);
+	if (!gnsect)
+		{
+		X509V3err(X509V3_F_GNAMES_FROM_SECTNAME,
+						X509V3_R_SECTION_NOT_FOUND);
+		return NULL;
+		}
+	gens = v2i_GENERAL_NAMES(NULL, ctx, gnsect);
+	if (*sect == '@')
+		X509V3_section_free(ctx, gnsect);
+	else
+		sk_CONF_VALUE_pop_free(gnsect, X509V3_conf_free);
+	return gens;
+	}
+
+static int set_dist_point_name(DIST_POINT_NAME **pdp, X509V3_CTX *ctx,
+							CONF_VALUE *cnf)
+	{
+	STACK_OF(GENERAL_NAME) *fnm = NULL;
+	STACK_OF(X509_NAME_ENTRY) *rnm = NULL;
+	if (!strncmp(cnf->name, "fullname", 9))
+		{
+		fnm = gnames_from_sectname(ctx, cnf->value);
+		if (!fnm)
+			goto err;
+		}
+	else if (!strcmp(cnf->name, "relativename"))
+		{
+		int ret;
+		STACK_OF(CONF_VALUE) *dnsect;
+		X509_NAME *nm;
+		nm = X509_NAME_new();
+		if (!nm)
+			return -1;
+		dnsect = X509V3_get_section(ctx, cnf->value);
+		if (!dnsect)
+			{
+			X509V3err(X509V3_F_SET_DIST_POINT_NAME,
+						X509V3_R_SECTION_NOT_FOUND);
+			return -1;
+			}
+		ret = X509V3_NAME_from_section(nm, dnsect, MBSTRING_ASC);
+		X509V3_section_free(ctx, dnsect);
+		rnm = nm->entries;
+		nm->entries = NULL;
+		X509_NAME_free(nm);
+		if (!ret || sk_X509_NAME_ENTRY_num(rnm) <= 0)
+			goto err;
+		/* Since its a name fragment can't have more than one
+		 * RDNSequence
+		 */
+		if (sk_X509_NAME_ENTRY_value(rnm,
+				sk_X509_NAME_ENTRY_num(rnm) - 1)->set)
+			{
+			X509V3err(X509V3_F_SET_DIST_POINT_NAME,
+						X509V3_R_INVALID_MULTIPLE_RDNS);
+			goto err;
+			}
+		}
+	else
+		return 0;
+
+	if (*pdp)
+		{
+		X509V3err(X509V3_F_SET_DIST_POINT_NAME,
+						X509V3_R_DISTPOINT_ALREADY_SET);
+		goto err;
+		}
+
+	*pdp = DIST_POINT_NAME_new();
+	if (!*pdp)
+		goto err;
+	if (fnm)
+		{
+		(*pdp)->type = 0;
+		(*pdp)->name.fullname = fnm;
+		}
+	else
+		{
+		(*pdp)->type = 1;
+		(*pdp)->name.relativename = rnm;
+		}
+
+	return 1;
+		
+	err:
+	if (fnm)
+		sk_GENERAL_NAME_pop_free(fnm, GENERAL_NAME_free);
+	if (rnm)
+		sk_X509_NAME_ENTRY_pop_free(rnm, X509_NAME_ENTRY_free);
+	return -1;
+	}
+
+static const BIT_STRING_BITNAME reason_flags[] = {
+{0, "Unused", "unused"},
+{1, "Key Compromise", "keyCompromise"},
+{2, "CA Compromise", "CACompromise"},
+{3, "Affiliation Changed", "affiliationChanged"},
+{4, "Superseded", "superseded"},
+{5, "Cessation Of Operation", "cessationOfOperation"},
+{6, "Certificate Hold", "certificateHold"},
+{7, "Privilege Withdrawn", "privilegeWithdrawn"},
+{8, "AA Compromise", "AACompromise"},
+{-1, NULL, NULL}
 };
 
-static STACK_OF(CONF_VALUE) *i2v_crld(X509V3_EXT_METHOD *method,
-			STACK_OF(DIST_POINT) *crld, STACK_OF(CONF_VALUE) *exts)
-{
-	DIST_POINT *point;
+static int set_reasons(ASN1_BIT_STRING **preas, char *value)
+	{
+	STACK_OF(CONF_VALUE) *rsk = NULL;
+	const BIT_STRING_BITNAME *pbn;
+	const char *bnam;
+	int i, ret = 0;
+	rsk = X509V3_parse_list(value);
+	if (!rsk)
+		return 0;
+	if (*preas)
+		return 0;
+	for (i = 0; i < sk_CONF_VALUE_num(rsk); i++)
+		{
+		bnam = sk_CONF_VALUE_value(rsk, i)->name;
+		if (!*preas)
+			{
+			*preas = ASN1_BIT_STRING_new();
+			if (!*preas)
+				goto err;
+			}
+		for (pbn = reason_flags; pbn->lname; pbn++)
+			{
+			if (!strcmp(pbn->sname, bnam))
+				{
+				if (!ASN1_BIT_STRING_set_bit(*preas,
+							pbn->bitnum, 1))
+					goto err;
+				break;
+				}
+			}
+		if (!pbn->lname)
+			goto err;
+		}
+	ret = 1;
+
+	err:
+	sk_CONF_VALUE_pop_free(rsk, X509V3_conf_free);
+	return ret;
+	}
+
+static int print_reasons(BIO *out, const char *rname,
+			ASN1_BIT_STRING *rflags, int indent)
+	{
+	int first = 1;
+	const BIT_STRING_BITNAME *pbn;
+	BIO_printf(out, "%*s%s:\n%*s", indent, "", rname, indent + 2, "");
+	for (pbn = reason_flags; pbn->lname; pbn++)
+		{
+		if (ASN1_BIT_STRING_get_bit(rflags, pbn->bitnum))
+			{
+			if (first)
+				first = 0;
+			else
+				BIO_puts(out, ", ");
+			BIO_puts(out, pbn->lname);
+			}
+		}
+	if (first)
+		BIO_puts(out, "<EMPTY>\n");
+	else
+		BIO_puts(out, "\n");
+	return 1;
+	}
+
+static DIST_POINT *crldp_from_section(X509V3_CTX *ctx,
+						STACK_OF(CONF_VALUE) *nval)
+	{
 	int i;
-	for(i = 0; i < sk_DIST_POINT_num(crld); i++) {
-		point = sk_DIST_POINT_value(crld, i);
-		if(point->distpoint) {
-			if(point->distpoint->type == 0)
-				exts = i2v_GENERAL_NAMES(NULL,
-					 point->distpoint->name.fullname, exts);
-		        else X509V3_add_value("RelativeName","<UNSUPPORTED>", &exts);
+	CONF_VALUE *cnf;
+	DIST_POINT *point = NULL;
+	point = DIST_POINT_new();
+	if (!point)
+		goto err;
+	for(i = 0; i < sk_CONF_VALUE_num(nval); i++)
+		{
+		int ret;
+		cnf = sk_CONF_VALUE_value(nval, i);
+		ret = set_dist_point_name(&point->distpoint, ctx, cnf);
+		if (ret > 0)
+			continue;
+		if (ret < 0)
+			goto err;
+		if (!strcmp(cnf->name, "reasons"))
+			{
+			if (!set_reasons(&point->reasons, cnf->value))
+				goto err;
+			}
+		else if (!strcmp(cnf->name, "CRLissuer"))
+			{
+			point->CRLissuer =
+				gnames_from_sectname(ctx, cnf->value);
+			if (!point->CRLissuer)
+				goto err;
+			}
 		}
-		if(point->reasons) 
-			X509V3_add_value("reasons","<UNSUPPORTED>", &exts);
-		if(point->CRLissuer)
-			X509V3_add_value("CRLissuer","<UNSUPPORTED>", &exts);
+
+	return point;
+			
+
+	err:
+	if (point)
+		DIST_POINT_free(point);
+	return NULL;
 	}
-	return exts;
-}
 
-static STACK_OF(DIST_POINT) *v2i_crld(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
-{
+static void *v2i_crld(const X509V3_EXT_METHOD *method,
+		      X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
+	{
 	STACK_OF(DIST_POINT) *crld = NULL;
 	GENERAL_NAMES *gens = NULL;
 	GENERAL_NAME *gen = NULL;
@@ -111,19 +320,44 @@ static STACK_OF(DIST_POINT) *v2i_crld(X509V3_EXT_METHOD *method,
 	for(i = 0; i < sk_CONF_VALUE_num(nval); i++) {
 		DIST_POINT *point;
 		cnf = sk_CONF_VALUE_value(nval, i);
-		if(!(gen = v2i_GENERAL_NAME(method, ctx, cnf))) goto err; 
-		if(!(gens = GENERAL_NAMES_new())) goto merr;
-		if(!sk_GENERAL_NAME_push(gens, gen)) goto merr;
-		gen = NULL;
-		if(!(point = DIST_POINT_new())) goto merr;
-		if(!sk_DIST_POINT_push(crld, point)) {
-			DIST_POINT_free(point);
-			goto merr;
-		}
-		if(!(point->distpoint = DIST_POINT_NAME_new())) goto merr;
-		point->distpoint->name.fullname = gens;
-		point->distpoint->type = 0;
-		gens = NULL;
+		if (!cnf->value)
+			{
+			STACK_OF(CONF_VALUE) *dpsect;
+			dpsect = X509V3_get_section(ctx, cnf->name);
+			if (!dpsect)
+				goto err;
+			point = crldp_from_section(ctx, dpsect);
+			X509V3_section_free(ctx, dpsect);
+			if (!point)
+				goto err;
+			if(!sk_DIST_POINT_push(crld, point))
+				{
+				DIST_POINT_free(point);
+				goto merr;
+				}
+			}
+		else
+			{
+			if(!(gen = v2i_GENERAL_NAME(method, ctx, cnf)))
+				goto err; 
+			if(!(gens = GENERAL_NAMES_new()))
+				goto merr;
+			if(!sk_GENERAL_NAME_push(gens, gen))
+				goto merr;
+			gen = NULL;
+			if(!(point = DIST_POINT_new()))
+				goto merr;
+			if(!sk_DIST_POINT_push(crld, point))
+				{
+				DIST_POINT_free(point);
+				goto merr;
+				}
+			if(!(point->distpoint = DIST_POINT_NAME_new()))
+				goto merr;
+			point->distpoint->name.fullname = gens;
+			point->distpoint->type = 0;
+			gens = NULL;
+			}
 	}
 	return crld;
 
@@ -139,11 +373,31 @@ static STACK_OF(DIST_POINT) *v2i_crld(X509V3_EXT_METHOD *method,
 IMPLEMENT_STACK_OF(DIST_POINT)
 IMPLEMENT_ASN1_SET_OF(DIST_POINT)
 
+static int dpn_cb(int operation, ASN1_VALUE **pval, const ASN1_ITEM *it,
+								void *exarg)
+	{
+	DIST_POINT_NAME *dpn = (DIST_POINT_NAME *)*pval;
+
+	switch(operation)
+		{
+		case ASN1_OP_NEW_POST:
+		dpn->dpname = NULL;
+		break;
+
+		case ASN1_OP_FREE_POST:
+		if (dpn->dpname)
+			X509_NAME_free(dpn->dpname);
+		break;
+		}
+	return 1;
+	}
+
 
-ASN1_CHOICE(DIST_POINT_NAME) = {
+ASN1_CHOICE_cb(DIST_POINT_NAME, dpn_cb) = {
 	ASN1_IMP_SEQUENCE_OF(DIST_POINT_NAME, name.fullname, GENERAL_NAME, 0),
 	ASN1_IMP_SET_OF(DIST_POINT_NAME, name.relativename, X509_NAME_ENTRY, 1)
-} ASN1_CHOICE_END(DIST_POINT_NAME)
+} ASN1_CHOICE_END_cb(DIST_POINT_NAME, DIST_POINT_NAME, type)
+
 
 IMPLEMENT_ASN1_FUNCTIONS(DIST_POINT_NAME)
 
@@ -160,3 +414,203 @@ ASN1_ITEM_TEMPLATE(CRL_DIST_POINTS) =
 ASN1_ITEM_TEMPLATE_END(CRL_DIST_POINTS)
 
 IMPLEMENT_ASN1_FUNCTIONS(CRL_DIST_POINTS)
+
+ASN1_SEQUENCE(ISSUING_DIST_POINT) = {
+	ASN1_EXP_OPT(ISSUING_DIST_POINT, distpoint, DIST_POINT_NAME, 0),
+	ASN1_IMP_OPT(ISSUING_DIST_POINT, onlyuser, ASN1_FBOOLEAN, 1),
+	ASN1_IMP_OPT(ISSUING_DIST_POINT, onlyCA, ASN1_FBOOLEAN, 2),
+	ASN1_IMP_OPT(ISSUING_DIST_POINT, onlysomereasons, ASN1_BIT_STRING, 3),
+	ASN1_IMP_OPT(ISSUING_DIST_POINT, indirectCRL, ASN1_FBOOLEAN, 4),
+	ASN1_IMP_OPT(ISSUING_DIST_POINT, onlyattr, ASN1_FBOOLEAN, 5)
+} ASN1_SEQUENCE_END(ISSUING_DIST_POINT)
+
+IMPLEMENT_ASN1_FUNCTIONS(ISSUING_DIST_POINT)
+
+static int i2r_idp(const X509V3_EXT_METHOD *method, void *pidp, BIO *out,
+		   int indent);
+static void *v2i_idp(const X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
+		     STACK_OF(CONF_VALUE) *nval);
+
+const X509V3_EXT_METHOD v3_idp =
+	{
+	NID_issuing_distribution_point, X509V3_EXT_MULTILINE,
+	ASN1_ITEM_ref(ISSUING_DIST_POINT),
+	0,0,0,0,
+	0,0,
+	0,
+	v2i_idp,
+	i2r_idp,0,
+	NULL
+	};
+
+static void *v2i_idp(const X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
+		     STACK_OF(CONF_VALUE) *nval)
+	{
+	ISSUING_DIST_POINT *idp = NULL;
+	CONF_VALUE *cnf;
+	char *name, *val;
+	int i, ret;
+	idp = ISSUING_DIST_POINT_new();
+	if (!idp)
+		goto merr;
+	for(i = 0; i < sk_CONF_VALUE_num(nval); i++)
+		{
+		cnf = sk_CONF_VALUE_value(nval, i);
+		name = cnf->name;
+		val = cnf->value;
+		ret = set_dist_point_name(&idp->distpoint, ctx, cnf);
+		if (ret > 0)
+			continue;
+		if (ret < 0)
+			goto err;
+		if (!strcmp(name, "onlyuser"))
+			{
+			if (!X509V3_get_value_bool(cnf, &idp->onlyuser))
+				goto err;
+			}
+		else if (!strcmp(name, "onlyCA"))
+			{
+			if (!X509V3_get_value_bool(cnf, &idp->onlyCA))
+				goto err;
+			}
+		else if (!strcmp(name, "onlyAA"))
+			{
+			if (!X509V3_get_value_bool(cnf, &idp->onlyattr))
+				goto err;
+			}
+		else if (!strcmp(name, "indirectCRL"))
+			{
+			if (!X509V3_get_value_bool(cnf, &idp->indirectCRL))
+				goto err;
+			}
+		else if (!strcmp(name, "onlysomereasons"))
+			{
+			if (!set_reasons(&idp->onlysomereasons, val))
+				goto err;
+			}
+		else
+			{
+                        X509V3err(X509V3_F_V2I_IDP, X509V3_R_INVALID_NAME);
+                        X509V3_conf_err(cnf);
+                        goto err;
+			}
+		}
+	return idp;
+
+	merr:
+	X509V3err(X509V3_F_V2I_IDP,ERR_R_MALLOC_FAILURE);
+	err:
+	ISSUING_DIST_POINT_free(idp);
+	return NULL;
+	}
+
+static int print_gens(BIO *out, STACK_OF(GENERAL_NAME) *gens, int indent)
+	{
+	int i;
+	for (i = 0; i < sk_GENERAL_NAME_num(gens); i++)
+		{
+		BIO_printf(out, "%*s", indent + 2, "");
+		GENERAL_NAME_print(out, sk_GENERAL_NAME_value(gens, i));
+		BIO_puts(out, "\n");
+		}
+	return 1;
+	}
+
+static int print_distpoint(BIO *out, DIST_POINT_NAME *dpn, int indent)
+	{
+	if (dpn->type == 0)
+		{
+		BIO_printf(out, "%*sFull Name:\n", indent, "");
+		print_gens(out, dpn->name.fullname, indent);
+		}
+	else
+		{
+		X509_NAME ntmp;
+		ntmp.entries = dpn->name.relativename;
+		BIO_printf(out, "%*sRelative Name:\n%*s",
+						indent, "", indent + 2, "");
+		X509_NAME_print_ex(out, &ntmp, 0, XN_FLAG_ONELINE);
+		BIO_puts(out, "\n");
+		}
+	return 1;
+	}
+
+static int i2r_idp(const X509V3_EXT_METHOD *method, void *pidp, BIO *out,
+		   int indent)
+	{
+	ISSUING_DIST_POINT *idp = pidp;
+	if (idp->distpoint)
+		print_distpoint(out, idp->distpoint, indent);
+	if (idp->onlyuser > 0)
+		BIO_printf(out, "%*sOnly User Certificates\n", indent, "");
+	if (idp->onlyCA > 0)
+		BIO_printf(out, "%*sOnly CA Certificates\n", indent, "");
+	if (idp->indirectCRL > 0)
+		BIO_printf(out, "%*sIndirect CRL\n", indent, "");
+	if (idp->onlysomereasons)
+		print_reasons(out, "Only Some Reasons", 
+				idp->onlysomereasons, indent);
+	if (idp->onlyattr > 0)
+		BIO_printf(out, "%*sOnly Attribute Certificates\n", indent, "");
+	if (!idp->distpoint && (idp->onlyuser <= 0) && (idp->onlyCA <= 0)
+		&& (idp->indirectCRL <= 0) && !idp->onlysomereasons
+		&& (idp->onlyattr <= 0))
+		BIO_printf(out, "%*s<EMPTY>\n", indent, "");
+		
+	return 1;
+	}
+
+static int i2r_crldp(const X509V3_EXT_METHOD *method, void *pcrldp, BIO *out,
+		     int indent)
+	{
+	STACK_OF(DIST_POINT) *crld = pcrldp;
+	DIST_POINT *point;
+	int i;
+	for(i = 0; i < sk_DIST_POINT_num(crld); i++)
+		{
+		BIO_puts(out, "\n");
+		point = sk_DIST_POINT_value(crld, i);
+		if(point->distpoint)
+			print_distpoint(out, point->distpoint, indent);
+		if(point->reasons) 
+			print_reasons(out, "Reasons", point->reasons,
+								indent);
+		if(point->CRLissuer)
+			{
+			BIO_printf(out, "%*sCRL Issuer:\n", indent, "");
+			print_gens(out, point->CRLissuer, indent);
+			}
+		}
+	return 1;
+	}
+
+int DIST_POINT_set_dpname(DIST_POINT_NAME *dpn, X509_NAME *iname)
+	{
+	int i;
+	STACK_OF(X509_NAME_ENTRY) *frag;
+	X509_NAME_ENTRY *ne;
+	if (!dpn || (dpn->type != 1))
+		return 1;
+	frag = dpn->name.relativename;
+	dpn->dpname = X509_NAME_dup(iname);
+	if (!dpn->dpname)
+		return 0;
+	for (i = 0; i < sk_X509_NAME_ENTRY_num(frag); i++)
+		{
+		ne = sk_X509_NAME_ENTRY_value(frag, i);
+		if (!X509_NAME_add_entry(dpn->dpname, ne, -1, i ? 0 : 1))
+			{
+			X509_NAME_free(dpn->dpname);
+			dpn->dpname = NULL;
+			return 0;
+			}
+		}
+	/* generate cached encoding of name */
+	if (i2d_X509_NAME(dpn->dpname, NULL) < 0)
+		{
+		X509_NAME_free(dpn->dpname);
+		dpn->dpname = NULL;
+		return 0;
+		}
+	return 1;
+	}
diff --git a/crypto/x509v3/v3_enum.c b/crypto/x509v3/v3_enum.c
index 36576eaa4d1a..c0575e368d31 100644
--- a/crypto/x509v3/v3_enum.c
+++ b/crypto/x509v3/v3_enum.c
@@ -61,14 +61,17 @@
 #include <openssl/x509v3.h>
 
 static ENUMERATED_NAMES crl_reasons[] = {
-{0, "Unspecified", "unspecified"},
-{1, "Key Compromise", "keyCompromise"},
-{2, "CA Compromise", "CACompromise"},
-{3, "Affiliation Changed", "affiliationChanged"},
-{4, "Superseded", "superseded"},
-{5, "Cessation Of Operation", "cessationOfOperation"},
-{6, "Certificate Hold", "certificateHold"},
-{8, "Remove From CRL", "removeFromCRL"},
+{CRL_REASON_UNSPECIFIED, 	 "Unspecified", "unspecified"},
+{CRL_REASON_KEY_COMPROMISE,	 "Key Compromise", "keyCompromise"},
+{CRL_REASON_CA_COMPROMISE,	 "CA Compromise", "CACompromise"},
+{CRL_REASON_AFFILIATION_CHANGED, "Affiliation Changed", "affiliationChanged"},
+{CRL_REASON_SUPERSEDED, 	 "Superseded", "superseded"},
+{CRL_REASON_CESSATION_OF_OPERATION,
+			"Cessation Of Operation", "cessationOfOperation"},
+{CRL_REASON_CERTIFICATE_HOLD,	 "Certificate Hold", "certificateHold"},
+{CRL_REASON_REMOVE_FROM_CRL,	 "Remove From CRL", "removeFromCRL"},
+{CRL_REASON_PRIVILEGE_WITHDRAWN, "Privilege Withdrawn", "privilegeWithdrawn"},
+{CRL_REASON_AA_COMPROMISE,	 "AA Compromise", "AACompromise"},
 {-1, NULL, NULL}
 };
 
diff --git a/crypto/x509v3/v3_extku.c b/crypto/x509v3/v3_extku.c
index c0d14500ed87..1c6653275700 100644
--- a/crypto/x509v3/v3_extku.c
+++ b/crypto/x509v3/v3_extku.c
@@ -63,9 +63,10 @@
 #include <openssl/conf.h>
 #include <openssl/x509v3.h>
 
-static void *v2i_EXTENDED_KEY_USAGE(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval);
-static STACK_OF(CONF_VALUE) *i2v_EXTENDED_KEY_USAGE(X509V3_EXT_METHOD *method,
+static void *v2i_EXTENDED_KEY_USAGE(const X509V3_EXT_METHOD *method,
+				    X509V3_CTX *ctx,
+				    STACK_OF(CONF_VALUE) *nval);
+static STACK_OF(CONF_VALUE) *i2v_EXTENDED_KEY_USAGE(const X509V3_EXT_METHOD *method,
 		void *eku, STACK_OF(CONF_VALUE) *extlist);
 
 const X509V3_EXT_METHOD v3_ext_ku = {
@@ -97,8 +98,9 @@ ASN1_ITEM_TEMPLATE_END(EXTENDED_KEY_USAGE)
 
 IMPLEMENT_ASN1_FUNCTIONS(EXTENDED_KEY_USAGE)
 
-static STACK_OF(CONF_VALUE) *i2v_EXTENDED_KEY_USAGE(X509V3_EXT_METHOD *method,
-		void *a, STACK_OF(CONF_VALUE) *ext_list)
+static STACK_OF(CONF_VALUE) *
+  i2v_EXTENDED_KEY_USAGE(const X509V3_EXT_METHOD *method, void *a,
+			 STACK_OF(CONF_VALUE) *ext_list)
 {
 	EXTENDED_KEY_USAGE *eku = a;
 	int i;
@@ -112,8 +114,8 @@ static STACK_OF(CONF_VALUE) *i2v_EXTENDED_KEY_USAGE(X509V3_EXT_METHOD *method,
 	return ext_list;
 }
 
-static void *v2i_EXTENDED_KEY_USAGE(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
+static void *v2i_EXTENDED_KEY_USAGE(const X509V3_EXT_METHOD *method,
+				    X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
 {
 	EXTENDED_KEY_USAGE *extku;
 	char *extval;
diff --git a/crypto/x509v3/v3_genn.c b/crypto/x509v3/v3_genn.c
index 84b4b1c8813f..b6283573013e 100644
--- a/crypto/x509v3/v3_genn.c
+++ b/crypto/x509v3/v3_genn.c
@@ -3,7 +3,7 @@
  * project 1999.
  */
 /* ====================================================================
- * Copyright (c) 1999 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2008 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -99,3 +99,154 @@ ASN1_ITEM_TEMPLATE(GENERAL_NAMES) =
 ASN1_ITEM_TEMPLATE_END(GENERAL_NAMES)
 
 IMPLEMENT_ASN1_FUNCTIONS(GENERAL_NAMES)
+
+GENERAL_NAME *GENERAL_NAME_dup(GENERAL_NAME *a)
+	{
+	return (GENERAL_NAME *) ASN1_dup((i2d_of_void *) i2d_GENERAL_NAME,
+					 (d2i_of_void *) d2i_GENERAL_NAME,
+					 (char *) a);
+	}
+
+/* Returns 0 if they are equal, != 0 otherwise. */
+int GENERAL_NAME_cmp(GENERAL_NAME *a, GENERAL_NAME *b)
+	{
+	int result = -1;
+
+	if (!a || !b || a->type != b->type) return -1;
+	switch(a->type)
+		{
+	case GEN_X400:
+	case GEN_EDIPARTY:
+		result = ASN1_TYPE_cmp(a->d.other, b->d.other);
+		break;
+
+	case GEN_OTHERNAME:
+		result = OTHERNAME_cmp(a->d.otherName, b->d.otherName);
+		break;
+
+	case GEN_EMAIL:
+	case GEN_DNS:
+	case GEN_URI:
+		result = ASN1_STRING_cmp(a->d.ia5, b->d.ia5);
+		break;
+
+	case GEN_DIRNAME:
+		result = X509_NAME_cmp(a->d.dirn, b->d.dirn);
+		break;
+
+	case GEN_IPADD:
+		result = ASN1_OCTET_STRING_cmp(a->d.ip, b->d.ip);
+		break;
+	
+	case GEN_RID:
+		result = OBJ_cmp(a->d.rid, b->d.rid);
+		break;
+		}
+	return result;
+	}
+
+/* Returns 0 if they are equal, != 0 otherwise. */
+int OTHERNAME_cmp(OTHERNAME *a, OTHERNAME *b)
+	{
+	int result = -1;
+
+	if (!a || !b) return -1;
+	/* Check their type first. */
+	if ((result = OBJ_cmp(a->type_id, b->type_id)) != 0)
+		return result;
+	/* Check the value. */
+	result = ASN1_TYPE_cmp(a->value, b->value);
+	return result;
+	}
+
+void GENERAL_NAME_set0_value(GENERAL_NAME *a, int type, void *value)
+	{
+	switch(type)
+		{
+	case GEN_X400:
+	case GEN_EDIPARTY:
+		a->d.other = value;
+		break;
+
+	case GEN_OTHERNAME:
+		a->d.otherName = value;
+		break;
+
+	case GEN_EMAIL:
+	case GEN_DNS:
+	case GEN_URI:
+		a->d.ia5 = value;
+		break;
+
+	case GEN_DIRNAME:
+		a->d.dirn = value;
+		break;
+
+	case GEN_IPADD:
+		a->d.ip = value;
+		break;
+	
+	case GEN_RID:
+		a->d.rid = value;
+		break;
+		}
+	a->type = type;
+	}
+
+void *GENERAL_NAME_get0_value(GENERAL_NAME *a, int *ptype)
+	{
+	if (ptype)
+		*ptype = a->type;
+	switch(a->type)
+		{
+	case GEN_X400:
+	case GEN_EDIPARTY:
+		return a->d.other;
+
+	case GEN_OTHERNAME:
+		return a->d.otherName;
+
+	case GEN_EMAIL:
+	case GEN_DNS:
+	case GEN_URI:
+		return a->d.ia5;
+
+	case GEN_DIRNAME:
+		return a->d.dirn;
+
+	case GEN_IPADD:
+		return a->d.ip;
+	
+	case GEN_RID:
+		return a->d.rid;
+
+	default:
+		return NULL;
+		}
+	}
+
+int GENERAL_NAME_set0_othername(GENERAL_NAME *gen,
+				ASN1_OBJECT *oid, ASN1_TYPE *value)
+	{
+	OTHERNAME *oth;
+	oth = OTHERNAME_new();
+	if (!oth)
+		return 0;
+	oth->type_id = oid;
+	oth->value = value;
+	GENERAL_NAME_set0_value(gen, GEN_OTHERNAME, oth);
+	return 1;
+	}
+
+int GENERAL_NAME_get0_otherName(GENERAL_NAME *gen, 
+				ASN1_OBJECT **poid, ASN1_TYPE **pvalue)
+	{
+	if (gen->type != GEN_OTHERNAME)
+		return 0;
+	if (poid)
+		*poid = gen->d.otherName->type_id;
+	if (pvalue)
+		*pvalue = gen->d.otherName->value;
+	return 1;
+	}
+
diff --git a/crypto/x509v3/v3_lib.c b/crypto/x509v3/v3_lib.c
index df3a48f43eff..0f1e1d4422b1 100644
--- a/crypto/x509v3/v3_lib.c
+++ b/crypto/x509v3/v3_lib.c
@@ -84,20 +84,24 @@ int X509V3_EXT_add(X509V3_EXT_METHOD *ext)
 }
 
 static int ext_cmp(const X509V3_EXT_METHOD * const *a,
-		const X509V3_EXT_METHOD * const *b)
+		   const X509V3_EXT_METHOD * const *b)
 {
 	return ((*a)->ext_nid - (*b)->ext_nid);
 }
 
-X509V3_EXT_METHOD *X509V3_EXT_get_nid(int nid)
+DECLARE_OBJ_BSEARCH_CMP_FN(const X509V3_EXT_METHOD *, const X509V3_EXT_METHOD *,
+			   ext);
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(const X509V3_EXT_METHOD *,
+			     const X509V3_EXT_METHOD *, ext);
+
+const X509V3_EXT_METHOD *X509V3_EXT_get_nid(int nid)
 {
-	X509V3_EXT_METHOD tmp, *t = &tmp, **ret;
+	X509V3_EXT_METHOD tmp;
+	const X509V3_EXT_METHOD *t = &tmp, * const *ret;
 	int idx;
 	if(nid < 0) return NULL;
 	tmp.ext_nid = nid;
-	ret = (X509V3_EXT_METHOD **) OBJ_bsearch((char *)&t,
-			(char *)standard_exts, STANDARD_EXTENSION_COUNT,
-			sizeof(X509V3_EXT_METHOD *), (int (*)(const void *, const void *))ext_cmp);
+	ret = OBJ_bsearch_ext(&t, standard_exts, STANDARD_EXTENSION_COUNT);
 	if(ret) return *ret;
 	if(!ext_list) return NULL;
 	idx = sk_X509V3_EXT_METHOD_find(ext_list, &tmp);
@@ -105,7 +109,7 @@ X509V3_EXT_METHOD *X509V3_EXT_get_nid(int nid)
 	return sk_X509V3_EXT_METHOD_value(ext_list, idx);
 }
 
-X509V3_EXT_METHOD *X509V3_EXT_get(X509_EXTENSION *ext)
+const X509V3_EXT_METHOD *X509V3_EXT_get(X509_EXTENSION *ext)
 {
 	int nid;
 	if((nid = OBJ_obj2nid(ext->object)) == NID_undef) return NULL;
@@ -122,7 +126,9 @@ int X509V3_EXT_add_list(X509V3_EXT_METHOD *extlist)
 
 int X509V3_EXT_add_alias(int nid_to, int nid_from)
 {
-	X509V3_EXT_METHOD *ext, *tmpext;
+	const X509V3_EXT_METHOD *ext;
+	X509V3_EXT_METHOD *tmpext;
+
 	if(!(ext = X509V3_EXT_get_nid(nid_from))) {
 		X509V3err(X509V3_F_X509V3_EXT_ADD_ALIAS,X509V3_R_EXTENSION_NOT_FOUND);
 		return 0;
@@ -161,7 +167,7 @@ int X509V3_add_standard_extensions(void)
 
 void *X509V3_EXT_d2i(X509_EXTENSION *ext)
 {
-	X509V3_EXT_METHOD *method;
+	const X509V3_EXT_METHOD *method;
 	const unsigned char *p;
 
 	if(!(method = X509V3_EXT_get(ext))) return NULL;
diff --git a/crypto/x509v3/v3_ncons.c b/crypto/x509v3/v3_ncons.c
index 624fe7e23bf4..a01dc64dd22e 100644
--- a/crypto/x509v3/v3_ncons.c
+++ b/crypto/x509v3/v3_ncons.c
@@ -63,15 +63,22 @@
 #include <openssl/conf.h>
 #include <openssl/x509v3.h>
 
-static void *v2i_NAME_CONSTRAINTS(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval);
-static int i2r_NAME_CONSTRAINTS(X509V3_EXT_METHOD *method, 
+static void *v2i_NAME_CONSTRAINTS(const X509V3_EXT_METHOD *method,
+				  X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval);
+static int i2r_NAME_CONSTRAINTS(const X509V3_EXT_METHOD *method, 
 				void *a, BIO *bp, int ind);
-static int do_i2r_name_constraints(X509V3_EXT_METHOD *method,
-				STACK_OF(GENERAL_SUBTREE) *trees,
-					BIO *bp, int ind, char *name);
+static int do_i2r_name_constraints(const X509V3_EXT_METHOD *method,
+				   STACK_OF(GENERAL_SUBTREE) *trees,
+				   BIO *bp, int ind, char *name);
 static int print_nc_ipadd(BIO *bp, ASN1_OCTET_STRING *ip);
 
+static int nc_match(GENERAL_NAME *gen, NAME_CONSTRAINTS *nc);
+static int nc_match_single(GENERAL_NAME *sub, GENERAL_NAME *gen);
+static int nc_dn(X509_NAME *sub, X509_NAME *nm);
+static int nc_dns(ASN1_IA5STRING *sub, ASN1_IA5STRING *dns);
+static int nc_email(ASN1_IA5STRING *sub, ASN1_IA5STRING *eml);
+static int nc_uri(ASN1_IA5STRING *uri, ASN1_IA5STRING *base);
+
 const X509V3_EXT_METHOD v3_name_constraints = {
 	NID_name_constraints, 0,
 	ASN1_ITEM_ref(NAME_CONSTRAINTS),
@@ -99,8 +106,8 @@ ASN1_SEQUENCE(NAME_CONSTRAINTS) = {
 IMPLEMENT_ASN1_ALLOC_FUNCTIONS(GENERAL_SUBTREE)
 IMPLEMENT_ASN1_ALLOC_FUNCTIONS(NAME_CONSTRAINTS)
 
-static void *v2i_NAME_CONSTRAINTS(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
+static void *v2i_NAME_CONSTRAINTS(const X509V3_EXT_METHOD *method,
+				  X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
 	{
 	int i;
 	CONF_VALUE tval, *val;
@@ -155,8 +162,8 @@ static void *v2i_NAME_CONSTRAINTS(X509V3_EXT_METHOD *method,
 
 	
 
-static int i2r_NAME_CONSTRAINTS(X509V3_EXT_METHOD *method,
-				void *a, BIO *bp, int ind)
+static int i2r_NAME_CONSTRAINTS(const X509V3_EXT_METHOD *method, void *a,
+				BIO *bp, int ind)
 	{
 	NAME_CONSTRAINTS *ncons = a;
 	do_i2r_name_constraints(method, ncons->permittedSubtrees,
@@ -166,9 +173,9 @@ static int i2r_NAME_CONSTRAINTS(X509V3_EXT_METHOD *method,
 	return 1;
 	}
 
-static int do_i2r_name_constraints(X509V3_EXT_METHOD *method,
-				STACK_OF(GENERAL_SUBTREE) *trees,
-					BIO *bp, int ind, char *name)
+static int do_i2r_name_constraints(const X509V3_EXT_METHOD *method,
+				   STACK_OF(GENERAL_SUBTREE) *trees,
+				   BIO *bp, int ind, char *name)
 	{
 	GENERAL_SUBTREE *tree;
 	int i;
@@ -217,3 +224,282 @@ static int print_nc_ipadd(BIO *bp, ASN1_OCTET_STRING *ip)
 	return 1;
 	}
 
+/* Check a certificate conforms to a specified set of constraints.
+ * Return values:
+ *  X509_V_OK: All constraints obeyed.
+ *  X509_V_ERR_PERMITTED_VIOLATION: Permitted subtree violation.
+ *  X509_V_ERR_EXCLUDED_VIOLATION: Excluded subtree violation.
+ *  X509_V_ERR_SUBTREE_MINMAX: Min or max values present and matching type.
+ *  X509_V_ERR_UNSUPPORTED_CONSTRAINT_TYPE:  Unsupported constraint type.
+ *  X509_V_ERR_UNSUPPORTED_CONSTRAINT_SYNTAX: bad unsupported constraint syntax.
+ *  X509_V_ERR_UNSUPPORTED_NAME_SYNTAX: bad or unsupported syntax of name
+
+ */
+
+int NAME_CONSTRAINTS_check(X509 *x, NAME_CONSTRAINTS *nc)
+	{
+	int r, i;
+	X509_NAME *nm;
+
+	nm = X509_get_subject_name(x);
+
+	if (X509_NAME_entry_count(nm) > 0)
+		{
+		GENERAL_NAME gntmp;
+		gntmp.type = GEN_DIRNAME;
+		gntmp.d.directoryName = nm;
+
+		r = nc_match(&gntmp, nc);
+
+		if (r != X509_V_OK)
+			return r;
+
+		gntmp.type = GEN_EMAIL;
+
+
+		/* Process any email address attributes in subject name */
+
+		for (i = -1;;)
+			{
+			X509_NAME_ENTRY *ne;
+			i = X509_NAME_get_index_by_NID(nm,
+						       NID_pkcs9_emailAddress,
+						       i);
+			if (i == -1)
+				break;
+			ne = X509_NAME_get_entry(nm, i);
+			gntmp.d.rfc822Name = X509_NAME_ENTRY_get_data(ne);
+			if (gntmp.d.rfc822Name->type != V_ASN1_IA5STRING)
+				return X509_V_ERR_UNSUPPORTED_NAME_SYNTAX;
+
+			r = nc_match(&gntmp, nc);
+
+			if (r != X509_V_OK)
+				return r;
+			}
+		
+		}
+
+	for (i = 0; i < sk_GENERAL_NAME_num(x->altname); i++)
+		{
+		GENERAL_NAME *gen = sk_GENERAL_NAME_value(x->altname, i);
+		r = nc_match(gen, nc);
+		if (r != X509_V_OK)
+			return r;
+		}
+
+	return X509_V_OK;
+
+	}
+
+static int nc_match(GENERAL_NAME *gen, NAME_CONSTRAINTS *nc)
+	{
+	GENERAL_SUBTREE *sub;
+	int i, r, match = 0;
+
+	/* Permitted subtrees: if any subtrees exist of matching the type
+	 * at least one subtree must match.
+	 */
+
+	for (i = 0; i < sk_GENERAL_SUBTREE_num(nc->permittedSubtrees); i++)
+		{
+		sub = sk_GENERAL_SUBTREE_value(nc->permittedSubtrees, i);
+		if (gen->type != sub->base->type)
+			continue;
+		if (sub->minimum || sub->maximum)
+			return X509_V_ERR_SUBTREE_MINMAX;
+		/* If we already have a match don't bother trying any more */
+		if (match == 2)
+			continue;
+		if (match == 0)
+			match = 1;
+		r = nc_match_single(gen, sub->base);
+		if (r == X509_V_OK)
+			match = 2;
+		else if (r != X509_V_ERR_PERMITTED_VIOLATION)
+			return r;
+		}
+
+	if (match == 1)
+		return X509_V_ERR_PERMITTED_VIOLATION;
+
+	/* Excluded subtrees: must not match any of these */
+
+	for (i = 0; i < sk_GENERAL_SUBTREE_num(nc->excludedSubtrees); i++)
+		{
+		sub = sk_GENERAL_SUBTREE_value(nc->excludedSubtrees, i);
+		if (gen->type != sub->base->type)
+			continue;
+		if (sub->minimum || sub->maximum)
+			return X509_V_ERR_SUBTREE_MINMAX;
+
+		r = nc_match_single(gen, sub->base);
+		if (r == X509_V_OK)
+			return X509_V_ERR_EXCLUDED_VIOLATION;
+		else if (r != X509_V_ERR_PERMITTED_VIOLATION)
+			return r;
+
+		}
+
+	return X509_V_OK;
+
+	}
+
+static int nc_match_single(GENERAL_NAME *gen, GENERAL_NAME *base)
+	{
+	switch(base->type)
+		{
+		case GEN_DIRNAME:
+		return nc_dn(gen->d.directoryName, base->d.directoryName);
+
+		case GEN_DNS:
+		return nc_dns(gen->d.dNSName, base->d.dNSName);
+
+		case GEN_EMAIL:
+		return nc_email(gen->d.rfc822Name, base->d.rfc822Name);
+
+		case GEN_URI:
+		return nc_uri(gen->d.uniformResourceIdentifier,
+					base->d.uniformResourceIdentifier);
+
+		default:
+		return X509_V_ERR_UNSUPPORTED_CONSTRAINT_TYPE;
+		}
+
+	}
+
+/* directoryName name constraint matching.
+ * The canonical encoding of X509_NAME makes this comparison easy. It is
+ * matched if the subtree is a subset of the name.
+ */
+
+static int nc_dn(X509_NAME *nm, X509_NAME *base)
+	{
+	/* Ensure canonical encodings are up to date.  */
+	if (nm->modified && i2d_X509_NAME(nm, NULL) < 0)
+		return X509_V_ERR_OUT_OF_MEM;
+	if (base->modified && i2d_X509_NAME(base, NULL) < 0)
+		return X509_V_ERR_OUT_OF_MEM;
+	if (base->canon_enclen > nm->canon_enclen)
+		return X509_V_ERR_PERMITTED_VIOLATION;
+	if (memcmp(base->canon_enc, nm->canon_enc, base->canon_enclen))
+		return X509_V_ERR_PERMITTED_VIOLATION;
+	return X509_V_OK;
+	}
+
+static int nc_dns(ASN1_IA5STRING *dns, ASN1_IA5STRING *base)
+	{
+	char *baseptr = (char *)base->data;
+	char *dnsptr = (char *)dns->data;
+	/* Empty matches everything */
+	if (!*baseptr)
+		return X509_V_OK;
+	/* Otherwise can add zero or more components on the left so
+	 * compare RHS and if dns is longer and expect '.' as preceding
+	 * character.
+	 */
+	if (dns->length > base->length)
+		{
+		dnsptr += dns->length - base->length;
+		if (dnsptr[-1] != '.')
+			return X509_V_ERR_PERMITTED_VIOLATION;
+		}
+
+	if (strcasecmp(baseptr, dnsptr))
+			return X509_V_ERR_PERMITTED_VIOLATION;
+
+	return X509_V_OK;
+
+	}
+
+static int nc_email(ASN1_IA5STRING *eml, ASN1_IA5STRING *base)
+	{
+	const char *baseptr = (char *)base->data;
+	const char *emlptr = (char *)eml->data;
+
+	const char *baseat = strchr(baseptr, '@');
+	const char *emlat = strchr(emlptr, '@');
+	if (!emlat)
+		return X509_V_ERR_UNSUPPORTED_NAME_SYNTAX;
+	/* Special case: inital '.' is RHS match */
+	if (!baseat && (*baseptr == '.'))
+		{
+		if (eml->length > base->length)
+			{
+			emlptr += eml->length - base->length;
+			if (!strcasecmp(baseptr, emlptr))
+				return X509_V_OK;
+			}
+		return X509_V_ERR_PERMITTED_VIOLATION;
+		}
+
+	/* If we have anything before '@' match local part */
+
+	if (baseat)
+		{
+		if (baseat != baseptr)
+			{
+			if ((baseat - baseptr) != (emlat - emlptr))
+				return X509_V_ERR_PERMITTED_VIOLATION;
+			/* Case sensitive match of local part */
+			if (strncmp(baseptr, emlptr, emlat - emlptr))
+				return X509_V_ERR_PERMITTED_VIOLATION;
+			}
+		/* Position base after '@' */
+		baseptr = baseat + 1;
+		}
+	emlptr = emlat + 1;
+	/* Just have hostname left to match: case insensitive */
+	if (strcasecmp(baseptr, emlptr))
+		return X509_V_ERR_PERMITTED_VIOLATION;
+
+	return X509_V_OK;
+
+	}
+
+static int nc_uri(ASN1_IA5STRING *uri, ASN1_IA5STRING *base)
+	{
+	const char *baseptr = (char *)base->data;
+	const char *hostptr = (char *)uri->data;
+	const char *p = strchr(hostptr, ':');
+	int hostlen;
+	/* Check for foo:// and skip past it */
+	if (!p || (p[1] != '/') || (p[2] != '/'))
+		return X509_V_ERR_UNSUPPORTED_NAME_SYNTAX;
+	hostptr = p + 3;
+
+	/* Determine length of hostname part of URI */
+
+	/* Look for a port indicator as end of hostname first */
+
+	p = strchr(hostptr, ':');
+	/* Otherwise look for trailing slash */
+	if (!p)
+		p = strchr(hostptr, '/');
+
+	if (!p)
+		hostlen = strlen(hostptr);
+	else
+		hostlen = p - hostptr;
+
+	if (hostlen == 0)
+		return X509_V_ERR_UNSUPPORTED_NAME_SYNTAX;
+
+	/* Special case: inital '.' is RHS match */
+	if (*baseptr == '.')
+		{
+		if (hostlen > base->length)
+			{
+			p = hostptr + hostlen - base->length;
+			if (!strncasecmp(p, baseptr, base->length))
+				return X509_V_OK;
+			}
+		return X509_V_ERR_PERMITTED_VIOLATION;
+		}
+
+	if ((base->length != (int)hostlen) || strncasecmp(hostptr, baseptr, hostlen))
+		return X509_V_ERR_PERMITTED_VIOLATION;
+
+	return X509_V_OK;
+
+	}
diff --git a/crypto/x509v3/v3_ocsp.c b/crypto/x509v3/v3_ocsp.c
index 5c19cf4130bb..0c165af3142c 100644
--- a/crypto/x509v3/v3_ocsp.c
+++ b/crypto/x509v3/v3_ocsp.c
@@ -68,19 +68,26 @@
 /* OCSP extensions and a couple of CRL entry extensions
  */
 
-static int i2r_ocsp_crlid(X509V3_EXT_METHOD *method, void *nonce, BIO *out, int indent);
-static int i2r_ocsp_acutoff(X509V3_EXT_METHOD *method, void *nonce, BIO *out, int indent);
-static int i2r_object(X509V3_EXT_METHOD *method, void *obj, BIO *out, int indent);
+static int i2r_ocsp_crlid(const X509V3_EXT_METHOD *method, void *nonce,
+			  BIO *out, int indent);
+static int i2r_ocsp_acutoff(const X509V3_EXT_METHOD *method, void *nonce,
+			    BIO *out, int indent);
+static int i2r_object(const X509V3_EXT_METHOD *method, void *obj, BIO *out,
+		      int indent);
 
 static void *ocsp_nonce_new(void);
 static int i2d_ocsp_nonce(void *a, unsigned char **pp);
 static void *d2i_ocsp_nonce(void *a, const unsigned char **pp, long length);
 static void ocsp_nonce_free(void *a);
-static int i2r_ocsp_nonce(X509V3_EXT_METHOD *method, void *nonce, BIO *out, int indent);
+static int i2r_ocsp_nonce(const X509V3_EXT_METHOD *method, void *nonce,
+			  BIO *out, int indent);
 
-static int i2r_ocsp_nocheck(X509V3_EXT_METHOD *method, void *nocheck, BIO *out, int indent);
-static void *s2i_ocsp_nocheck(X509V3_EXT_METHOD *method, X509V3_CTX *ctx, const char *str);
-static int i2r_ocsp_serviceloc(X509V3_EXT_METHOD *method, void *in, BIO *bp, int ind);
+static int i2r_ocsp_nocheck(const X509V3_EXT_METHOD *method,
+			    void *nocheck, BIO *out, int indent);
+static void *s2i_ocsp_nocheck(const X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
+			      const char *str);
+static int i2r_ocsp_serviceloc(const X509V3_EXT_METHOD *method, void *in,
+			       BIO *bp, int ind);
 
 const X509V3_EXT_METHOD v3_ocsp_crlid = {
 	NID_id_pkix_OCSP_CrlID, 0, ASN1_ITEM_ref(OCSP_CRLID),
@@ -148,7 +155,8 @@ const X509V3_EXT_METHOD v3_ocsp_serviceloc = {
 	NULL
 };
 
-static int i2r_ocsp_crlid(X509V3_EXT_METHOD *method, void *in, BIO *bp, int ind)
+static int i2r_ocsp_crlid(const X509V3_EXT_METHOD *method, void *in, BIO *bp,
+			  int ind)
 {
 	OCSP_CRLID *a = in;
 	if (a->crlUrl)
@@ -174,7 +182,8 @@ static int i2r_ocsp_crlid(X509V3_EXT_METHOD *method, void *in, BIO *bp, int ind)
 	return 0;
 }
 
-static int i2r_ocsp_acutoff(X509V3_EXT_METHOD *method, void *cutoff, BIO *bp, int ind)
+static int i2r_ocsp_acutoff(const X509V3_EXT_METHOD *method, void *cutoff,
+			    BIO *bp, int ind)
 {
 	if (BIO_printf(bp, "%*s", ind, "") <= 0) return 0;
 	if(!ASN1_GENERALIZEDTIME_print(bp, cutoff)) return 0;
@@ -182,7 +191,8 @@ static int i2r_ocsp_acutoff(X509V3_EXT_METHOD *method, void *cutoff, BIO *bp, in
 }
 
 
-static int i2r_object(X509V3_EXT_METHOD *method, void *oid, BIO *bp, int ind)
+static int i2r_object(const X509V3_EXT_METHOD *method, void *oid, BIO *bp,
+		      int ind)
 {
 	if (BIO_printf(bp, "%*s", ind, "") <= 0) return 0;
 	if(i2a_ASN1_OBJECT(bp, oid) <= 0) return 0;
@@ -232,7 +242,8 @@ static void ocsp_nonce_free(void *a)
 	M_ASN1_OCTET_STRING_free(a);
 }
 
-static int i2r_ocsp_nonce(X509V3_EXT_METHOD *method, void *nonce, BIO *out, int indent)
+static int i2r_ocsp_nonce(const X509V3_EXT_METHOD *method, void *nonce,
+			  BIO *out, int indent)
 {
 	if(BIO_printf(out, "%*s", indent, "") <= 0) return 0;
 	if(i2a_ASN1_STRING(out, nonce, V_ASN1_OCTET_STRING) <= 0) return 0;
@@ -241,17 +252,20 @@ static int i2r_ocsp_nonce(X509V3_EXT_METHOD *method, void *nonce, BIO *out, int
 
 /* Nocheck is just a single NULL. Don't print anything and always set it */
 
-static int i2r_ocsp_nocheck(X509V3_EXT_METHOD *method, void *nocheck, BIO *out, int indent)
+static int i2r_ocsp_nocheck(const X509V3_EXT_METHOD *method, void *nocheck,
+			    BIO *out, int indent)
 {
 	return 1;
 }
 
-static void *s2i_ocsp_nocheck(X509V3_EXT_METHOD *method, X509V3_CTX *ctx, const char *str)
+static void *s2i_ocsp_nocheck(const X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
+			      const char *str)
 {
 	return ASN1_NULL_new();
 }
 
-static int i2r_ocsp_serviceloc(X509V3_EXT_METHOD *method, void *in, BIO *bp, int ind)
+static int i2r_ocsp_serviceloc(const X509V3_EXT_METHOD *method, void *in,
+			       BIO *bp, int ind)
         {
 	int i;
 	OCSP_SERVICELOC *a = in;
diff --git a/crypto/x509v3/v3_pci.c b/crypto/x509v3/v3_pci.c
index 823e9afcd296..0dcfa004fe23 100644
--- a/crypto/x509v3/v3_pci.c
+++ b/crypto/x509v3/v3_pci.c
@@ -82,7 +82,7 @@ static int process_pci_value(CONF_VALUE *val,
 		{
 		if (*language)
 			{
-			X509V3err(X509V3_F_PROCESS_PCI_VALUE,X509V3_R_POLICY_LANGUAGE_ALREADTY_DEFINED);
+			X509V3err(X509V3_F_PROCESS_PCI_VALUE,X509V3_R_POLICY_LANGUAGE_ALREADY_DEFINED);
 			X509V3_conf_err(val);
 			return 0;
 			}
@@ -97,7 +97,7 @@ static int process_pci_value(CONF_VALUE *val,
 		{
 		if (*pathlen)
 			{
-			X509V3err(X509V3_F_PROCESS_PCI_VALUE,X509V3_R_POLICY_PATH_LENGTH_ALREADTY_DEFINED);
+			X509V3err(X509V3_F_PROCESS_PCI_VALUE,X509V3_R_POLICY_PATH_LENGTH_ALREADY_DEFINED);
 			X509V3_conf_err(val);
 			return 0;
 			}
diff --git a/crypto/x509v3/v3_pcons.c b/crypto/x509v3/v3_pcons.c
index 86c0ff70e6c0..30ca65235120 100644
--- a/crypto/x509v3/v3_pcons.c
+++ b/crypto/x509v3/v3_pcons.c
@@ -64,10 +64,12 @@
 #include <openssl/conf.h>
 #include <openssl/x509v3.h>
 
-static STACK_OF(CONF_VALUE) *i2v_POLICY_CONSTRAINTS(X509V3_EXT_METHOD *method,
-				void *bcons, STACK_OF(CONF_VALUE) *extlist);
-static void *v2i_POLICY_CONSTRAINTS(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *values);
+static STACK_OF(CONF_VALUE) *
+i2v_POLICY_CONSTRAINTS(const X509V3_EXT_METHOD *method, void *bcons,
+		       STACK_OF(CONF_VALUE) *extlist);
+static void *v2i_POLICY_CONSTRAINTS(const X509V3_EXT_METHOD *method,
+				    X509V3_CTX *ctx,
+				    STACK_OF(CONF_VALUE) *values);
 
 const X509V3_EXT_METHOD v3_policy_constraints = {
 NID_policy_constraints, 0,
@@ -88,8 +90,9 @@ ASN1_SEQUENCE(POLICY_CONSTRAINTS) = {
 IMPLEMENT_ASN1_ALLOC_FUNCTIONS(POLICY_CONSTRAINTS)
 
 
-static STACK_OF(CONF_VALUE) *i2v_POLICY_CONSTRAINTS(X509V3_EXT_METHOD *method,
-	     void *a, STACK_OF(CONF_VALUE) *extlist)
+static STACK_OF(CONF_VALUE) *
+i2v_POLICY_CONSTRAINTS(const X509V3_EXT_METHOD *method, void *a,
+		       STACK_OF(CONF_VALUE) *extlist)
 {
 	POLICY_CONSTRAINTS *pcons = a;
 	X509V3_add_value_int("Require Explicit Policy",
@@ -99,8 +102,9 @@ static STACK_OF(CONF_VALUE) *i2v_POLICY_CONSTRAINTS(X509V3_EXT_METHOD *method,
 	return extlist;
 }
 
-static void *v2i_POLICY_CONSTRAINTS(X509V3_EXT_METHOD *method,
-	     X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *values)
+static void *v2i_POLICY_CONSTRAINTS(const X509V3_EXT_METHOD *method,
+				    X509V3_CTX *ctx,
+				    STACK_OF(CONF_VALUE) *values)
 {
 	POLICY_CONSTRAINTS *pcons=NULL;
 	CONF_VALUE *val;
diff --git a/crypto/x509v3/v3_pmaps.c b/crypto/x509v3/v3_pmaps.c
index da03bbc35d38..865bcd39805a 100644
--- a/crypto/x509v3/v3_pmaps.c
+++ b/crypto/x509v3/v3_pmaps.c
@@ -63,10 +63,11 @@
 #include <openssl/conf.h>
 #include <openssl/x509v3.h>
 
-static void *v2i_POLICY_MAPPINGS(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval);
-static STACK_OF(CONF_VALUE) *i2v_POLICY_MAPPINGS(X509V3_EXT_METHOD *method,
-				void *pmps, STACK_OF(CONF_VALUE) *extlist);
+static void *v2i_POLICY_MAPPINGS(const X509V3_EXT_METHOD *method,
+				 X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval);
+static STACK_OF(CONF_VALUE) *
+i2v_POLICY_MAPPINGS(const X509V3_EXT_METHOD *method, void *pmps,
+		    STACK_OF(CONF_VALUE) *extlist);
 
 const X509V3_EXT_METHOD v3_policy_mappings = {
 	NID_policy_mappings, 0,
@@ -92,8 +93,9 @@ ASN1_ITEM_TEMPLATE_END(POLICY_MAPPINGS)
 IMPLEMENT_ASN1_ALLOC_FUNCTIONS(POLICY_MAPPING)
 
 
-static STACK_OF(CONF_VALUE) *i2v_POLICY_MAPPINGS(X509V3_EXT_METHOD *method,
-		void *a, STACK_OF(CONF_VALUE) *ext_list)
+static STACK_OF(CONF_VALUE) *
+i2v_POLICY_MAPPINGS(const X509V3_EXT_METHOD *method, void *a,
+		    STACK_OF(CONF_VALUE) *ext_list)
 {
 	POLICY_MAPPINGS *pmaps = a;
 	POLICY_MAPPING *pmap;
@@ -109,8 +111,8 @@ static STACK_OF(CONF_VALUE) *i2v_POLICY_MAPPINGS(X509V3_EXT_METHOD *method,
 	return ext_list;
 }
 
-static void *v2i_POLICY_MAPPINGS(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
+static void *v2i_POLICY_MAPPINGS(const X509V3_EXT_METHOD *method,
+				 X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval)
 {
 	POLICY_MAPPINGS *pmaps;
 	POLICY_MAPPING *pmap;
diff --git a/crypto/x509v3/v3_prn.c b/crypto/x509v3/v3_prn.c
index c1bb17f105a1..31462187081b 100644
--- a/crypto/x509v3/v3_prn.c
+++ b/crypto/x509v3/v3_prn.c
@@ -110,7 +110,7 @@ int X509V3_EXT_print(BIO *out, X509_EXTENSION *ext, unsigned long flag, int inde
 	void *ext_str = NULL;
 	char *value = NULL;
 	const unsigned char *p;
-	X509V3_EXT_METHOD *method;	
+	const X509V3_EXT_METHOD *method;	
 	STACK_OF(CONF_VALUE) *nval = NULL;
 	int ok = 1;
 
diff --git a/crypto/x509v3/v3_purp.c b/crypto/x509v3/v3_purp.c
index e18751e01cbc..181bd34979bc 100644
--- a/crypto/x509v3/v3_purp.c
+++ b/crypto/x509v3/v3_purp.c
@@ -71,6 +71,7 @@ static int purpose_smime(const X509 *x, int ca);
 static int check_purpose_smime_sign(const X509_PURPOSE *xp, const X509 *x, int ca);
 static int check_purpose_smime_encrypt(const X509_PURPOSE *xp, const X509 *x, int ca);
 static int check_purpose_crl_sign(const X509_PURPOSE *xp, const X509 *x, int ca);
+static int check_purpose_timestamp_sign(const X509_PURPOSE *xp, const X509 *x, int ca);
 static int no_check(const X509_PURPOSE *xp, const X509 *x, int ca);
 static int ocsp_helper(const X509_PURPOSE *xp, const X509 *x, int ca);
 
@@ -87,6 +88,7 @@ static X509_PURPOSE xstandard[] = {
 	{X509_PURPOSE_CRL_SIGN, X509_TRUST_COMPAT, 0, check_purpose_crl_sign, "CRL signing", "crlsign", NULL},
 	{X509_PURPOSE_ANY, X509_TRUST_DEFAULT, 0, no_check, "Any Purpose", "any", NULL},
 	{X509_PURPOSE_OCSP_HELPER, X509_TRUST_COMPAT, 0, ocsp_helper, "OCSP helper", "ocsphelper", NULL},
+	{X509_PURPOSE_TIMESTAMP_SIGN, X509_TRUST_TSA, 0, check_purpose_timestamp_sign, "Time Stamp signing", "timestampsign", NULL},
 };
 
 #define X509_PURPOSE_COUNT (sizeof(xstandard)/sizeof(X509_PURPOSE))
@@ -265,11 +267,14 @@ int X509_PURPOSE_get_trust(X509_PURPOSE *xp)
 	return xp->trust;
 }
 
-static int nid_cmp(int *a, int *b)
+static int nid_cmp(const int *a, const int *b)
 	{
 	return *a - *b;
 	}
 
+DECLARE_OBJ_BSEARCH_CMP_FN(int, int, nid);
+IMPLEMENT_OBJ_BSEARCH_CMP_FN(int, int, nid);
+
 int X509_supported_extension(X509_EXTENSION *ex)
 	{
 	/* This table is a list of the NIDs of supported extensions:
@@ -280,7 +285,7 @@ int X509_supported_extension(X509_EXTENSION *ex)
 	 * searched using bsearch.
 	 */
 
-	static int supported_nids[] = {
+	static const int supported_nids[] = {
 		NID_netscape_cert_type, /* 71 */
         	NID_key_usage,		/* 83 */
 		NID_subject_alt_name,	/* 85 */
@@ -292,24 +297,62 @@ int X509_supported_extension(X509_EXTENSION *ex)
 		NID_sbgp_autonomousSysNum, /* 291 */
 #endif
 		NID_policy_constraints,	/* 401 */
-		NID_proxyCertInfo,	/* 661 */
+		NID_proxyCertInfo,	/* 663 */
+		NID_name_constraints,	/* 666 */
+		NID_policy_mappings,	/* 747 */
 		NID_inhibit_any_policy	/* 748 */
 	};
 
-	int ex_nid;
-
-	ex_nid = OBJ_obj2nid(X509_EXTENSION_get_object(ex));
+	int ex_nid = OBJ_obj2nid(X509_EXTENSION_get_object(ex));
 
 	if (ex_nid == NID_undef) 
 		return 0;
 
-	if (OBJ_bsearch((char *)&ex_nid, (char *)supported_nids,
-		sizeof(supported_nids)/sizeof(int), sizeof(int),
-		(int (*)(const void *, const void *))nid_cmp))
+	if (OBJ_bsearch_nid(&ex_nid, supported_nids,
+			sizeof(supported_nids)/sizeof(int)))
 		return 1;
 	return 0;
 	}
- 
+
+static void setup_dp(X509 *x, DIST_POINT *dp)
+	{
+	X509_NAME *iname = NULL;
+	int i;
+	if (dp->reasons)
+		{
+		if (dp->reasons->length > 0)
+			dp->dp_reasons = dp->reasons->data[0];
+		if (dp->reasons->length > 1)
+			dp->dp_reasons |= (dp->reasons->data[1] << 8);
+		dp->dp_reasons &= CRLDP_ALL_REASONS;
+		}
+	else
+		dp->dp_reasons = CRLDP_ALL_REASONS;
+	if (!dp->distpoint || (dp->distpoint->type != 1))
+		return;
+	for (i = 0; i < sk_GENERAL_NAME_num(dp->CRLissuer); i++)
+		{
+		GENERAL_NAME *gen = sk_GENERAL_NAME_value(dp->CRLissuer, i);
+		if (gen->type == GEN_DIRNAME)
+			{
+			iname = gen->d.directoryName;
+			break;
+			}
+		}
+	if (!iname)
+		iname = X509_get_issuer_name(x);
+
+	DIST_POINT_set_dpname(dp->distpoint, iname);
+
+	}
+
+static void setup_crldp(X509 *x)
+	{
+	int i;
+	x->crldp = X509_get_ext_d2i(x, NID_crl_distribution_points, NULL, NULL);
+	for (i = 0; i < sk_DIST_POINT_num(x->crldp); i++)
+		setup_dp(x, sk_DIST_POINT_value(x->crldp, i));
+	}
 
 static void x509v3_cache_extensions(X509 *x)
 {
@@ -417,16 +460,25 @@ static void x509v3_cache_extensions(X509 *x)
 	}
 	x->skid =X509_get_ext_d2i(x, NID_subject_key_identifier, NULL, NULL);
 	x->akid =X509_get_ext_d2i(x, NID_authority_key_identifier, NULL, NULL);
+	x->altname = X509_get_ext_d2i(x, NID_subject_alt_name, NULL, NULL);
+	x->nc = X509_get_ext_d2i(x, NID_name_constraints, &i, NULL);
+	if (!x->nc && (i != -1))
+		x->ex_flags |= EXFLAG_INVALID;
+	setup_crldp(x);
+
 #ifndef OPENSSL_NO_RFC3779
-	x->rfc3779_addr =X509_get_ext_d2i(x, NID_sbgp_ipAddrBlock, NULL, NULL);
-	x->rfc3779_asid =X509_get_ext_d2i(x, NID_sbgp_autonomousSysNum,
-					  NULL, NULL);
+ 	x->rfc3779_addr =X509_get_ext_d2i(x, NID_sbgp_ipAddrBlock, NULL, NULL);
+ 	x->rfc3779_asid =X509_get_ext_d2i(x, NID_sbgp_autonomousSysNum,
+ 					  NULL, NULL);
 #endif
 	for (i = 0; i < X509_get_ext_count(x); i++)
 		{
 		ex = X509_get_ext(x, i);
 		if (!X509_EXTENSION_get_critical(ex))
 			continue;
+		if (OBJ_obj2nid(X509_EXTENSION_get_object(ex))
+					== NID_freshest_crl)
+			x->ex_flags |= EXFLAG_FRESHEST;
 		if (!X509_supported_extension(ex))
 			{
 			x->ex_flags |= EXFLAG_CRITICAL;
@@ -594,6 +646,41 @@ static int ocsp_helper(const X509_PURPOSE *xp, const X509 *x, int ca)
 	return 1;
 }
 
+static int check_purpose_timestamp_sign(const X509_PURPOSE *xp, const X509 *x,
+					int ca)
+{
+	int i_ext;
+
+	/* If ca is true we must return if this is a valid CA certificate. */
+	if (ca) return check_ca(x);
+
+	/* 
+	 * Check the optional key usage field:
+	 * if Key Usage is present, it must be one of digitalSignature 
+	 * and/or nonRepudiation (other values are not consistent and shall
+	 * be rejected).
+	 */
+	if ((x->ex_flags & EXFLAG_KUSAGE)
+	    && ((x->ex_kusage & ~(KU_NON_REPUDIATION | KU_DIGITAL_SIGNATURE)) ||
+		!(x->ex_kusage & (KU_NON_REPUDIATION | KU_DIGITAL_SIGNATURE))))
+		return 0;
+
+	/* Only time stamp key usage is permitted and it's required. */
+	if (!(x->ex_flags & EXFLAG_XKUSAGE) || x->ex_xkusage != XKU_TIMESTAMP)
+		return 0;
+
+	/* Extended Key Usage MUST be critical */
+	i_ext = X509_get_ext_by_NID((X509 *) x, NID_ext_key_usage, 0);
+	if (i_ext >= 0)
+		{
+		X509_EXTENSION *ext = X509_get_ext((X509 *) x, i_ext);
+		if (!X509_EXTENSION_get_critical(ext))
+			return 0;
+		}
+
+	return 1;
+}
+
 static int no_check(const X509_PURPOSE *xp, const X509 *x, int ca)
 {
 	return 1;
@@ -618,39 +705,14 @@ int X509_check_issued(X509 *issuer, X509 *subject)
 				return X509_V_ERR_SUBJECT_ISSUER_MISMATCH;
 	x509v3_cache_extensions(issuer);
 	x509v3_cache_extensions(subject);
-	if(subject->akid) {
-		/* Check key ids (if present) */
-		if(subject->akid->keyid && issuer->skid &&
-		 ASN1_OCTET_STRING_cmp(subject->akid->keyid, issuer->skid) )
-				return X509_V_ERR_AKID_SKID_MISMATCH;
-		/* Check serial number */
-		if(subject->akid->serial &&
-			ASN1_INTEGER_cmp(X509_get_serialNumber(issuer),
-						subject->akid->serial))
-				return X509_V_ERR_AKID_ISSUER_SERIAL_MISMATCH;
-		/* Check issuer name */
-		if(subject->akid->issuer) {
-			/* Ugh, for some peculiar reason AKID includes
-			 * SEQUENCE OF GeneralName. So look for a DirName.
-			 * There may be more than one but we only take any
-			 * notice of the first.
-			 */
-			GENERAL_NAMES *gens;
-			GENERAL_NAME *gen;
-			X509_NAME *nm = NULL;
-			int i;
-			gens = subject->akid->issuer;
-			for(i = 0; i < sk_GENERAL_NAME_num(gens); i++) {
-				gen = sk_GENERAL_NAME_value(gens, i);
-				if(gen->type == GEN_DIRNAME) {
-					nm = gen->d.dirn;
-					break;
-				}
-			}
-			if(nm && X509_NAME_cmp(nm, X509_get_issuer_name(issuer)))
-				return X509_V_ERR_AKID_ISSUER_SERIAL_MISMATCH;
+
+	if(subject->akid)
+		{
+		int ret = X509_check_akid(issuer, subject->akid);
+		if (ret != X509_V_OK)
+			return ret;
 		}
-	}
+
 	if(subject->ex_flags & EXFLAG_PROXY)
 		{
 		if(ku_reject(issuer, KU_DIGITAL_SIGNATURE))
@@ -661,3 +723,45 @@ int X509_check_issued(X509 *issuer, X509 *subject)
 	return X509_V_OK;
 }
 
+int X509_check_akid(X509 *issuer, AUTHORITY_KEYID *akid)
+	{
+
+	if(!akid)
+		return X509_V_OK;
+
+	/* Check key ids (if present) */
+	if(akid->keyid && issuer->skid &&
+		 ASN1_OCTET_STRING_cmp(akid->keyid, issuer->skid) )
+				return X509_V_ERR_AKID_SKID_MISMATCH;
+	/* Check serial number */
+	if(akid->serial &&
+		ASN1_INTEGER_cmp(X509_get_serialNumber(issuer), akid->serial))
+				return X509_V_ERR_AKID_ISSUER_SERIAL_MISMATCH;
+	/* Check issuer name */
+	if(akid->issuer)
+		{
+		/* Ugh, for some peculiar reason AKID includes
+		 * SEQUENCE OF GeneralName. So look for a DirName.
+		 * There may be more than one but we only take any
+		 * notice of the first.
+		 */
+		GENERAL_NAMES *gens;
+		GENERAL_NAME *gen;
+		X509_NAME *nm = NULL;
+		int i;
+		gens = akid->issuer;
+		for(i = 0; i < sk_GENERAL_NAME_num(gens); i++)
+			{
+			gen = sk_GENERAL_NAME_value(gens, i);
+			if(gen->type == GEN_DIRNAME)
+				{
+				nm = gen->d.dirn;
+				break;
+				}
+			}
+		if(nm && X509_NAME_cmp(nm, X509_get_issuer_name(issuer)))
+			return X509_V_ERR_AKID_ISSUER_SERIAL_MISMATCH;
+		}
+	return X509_V_OK;
+	}
+
diff --git a/crypto/x509v3/v3_skey.c b/crypto/x509v3/v3_skey.c
index 202c9e489656..0a984fbaa872 100644
--- a/crypto/x509v3/v3_skey.c
+++ b/crypto/x509v3/v3_skey.c
@@ -129,7 +129,8 @@ static ASN1_OCTET_STRING *s2i_skey_id(X509V3_EXT_METHOD *method,
 		goto err;
 	}
 
-	EVP_Digest(pk->data, pk->length, pkey_dig, &diglen, EVP_sha1(), NULL);
+	if (!EVP_Digest(pk->data, pk->length, pkey_dig, &diglen, EVP_sha1(), NULL))
+		goto err;
 
 	if(!M_ASN1_OCTET_STRING_set(oct, pkey_dig, diglen)) {
 		X509V3err(X509V3_F_S2I_SKEY_ID,ERR_R_MALLOC_FAILURE);
diff --git a/crypto/x509v3/v3_utl.c b/crypto/x509v3/v3_utl.c
index 7a45216c007f..e0302345400a 100644
--- a/crypto/x509v3/v3_utl.c
+++ b/crypto/x509v3/v3_utl.c
@@ -67,9 +67,9 @@
 
 static char *strip_spaces(char *name);
 static int sk_strcmp(const char * const *a, const char * const *b);
-static STACK *get_email(X509_NAME *name, GENERAL_NAMES *gens);
-static void str_free(void *str);
-static int append_ia5(STACK **sk, ASN1_IA5STRING *email);
+static STACK_OF(OPENSSL_STRING) *get_email(X509_NAME *name, GENERAL_NAMES *gens);
+static void str_free(OPENSSL_STRING str);
+static int append_ia5(STACK_OF(OPENSSL_STRING) **sk, ASN1_IA5STRING *email);
 
 static int ipv4_from_asc(unsigned char *v4, const char *in);
 static int ipv6_from_asc(unsigned char *v6, const char *in);
@@ -360,10 +360,10 @@ static char *strip_spaces(char *name)
  * @@@ (Contents of buffer are always kept in ASCII, also on EBCDIC machines)
  */
 
-char *hex_to_string(unsigned char *buffer, long len)
+char *hex_to_string(const unsigned char *buffer, long len)
 {
 	char *tmp, *q;
-	unsigned char *p;
+	const unsigned char *p;
 	int i;
 	const static char hexdig[] = "0123456789ABCDEF";
 	if(!buffer || !len) return NULL;
@@ -389,7 +389,7 @@ char *hex_to_string(unsigned char *buffer, long len)
  * a buffer
  */
 
-unsigned char *string_to_hex(char *str, long *len)
+unsigned char *string_to_hex(const char *str, long *len)
 {
 	unsigned char *hexbuf, *q;
 	unsigned char ch, cl, *p;
@@ -463,21 +463,23 @@ static int sk_strcmp(const char * const *a, const char * const *b)
 	return strcmp(*a, *b);
 }
 
-STACK *X509_get1_email(X509 *x)
+STACK_OF(OPENSSL_STRING) *X509_get1_email(X509 *x)
 {
 	GENERAL_NAMES *gens;
-	STACK *ret;
+	STACK_OF(OPENSSL_STRING) *ret;
+
 	gens = X509_get_ext_d2i(x, NID_subject_alt_name, NULL, NULL);
 	ret = get_email(X509_get_subject_name(x), gens);
 	sk_GENERAL_NAME_pop_free(gens, GENERAL_NAME_free);
 	return ret;
 }
 
-STACK *X509_get1_ocsp(X509 *x)
+STACK_OF(OPENSSL_STRING) *X509_get1_ocsp(X509 *x)
 {
 	AUTHORITY_INFO_ACCESS *info;
-	STACK *ret = NULL;
+	STACK_OF(OPENSSL_STRING) *ret = NULL;
 	int i;
+
 	info = X509_get_ext_d2i(x, NID_info_access, NULL, NULL);
 	if (!info)
 		return NULL;
@@ -497,11 +499,12 @@ STACK *X509_get1_ocsp(X509 *x)
 	return ret;
 }
 
-STACK *X509_REQ_get1_email(X509_REQ *x)
+STACK_OF(OPENSSL_STRING) *X509_REQ_get1_email(X509_REQ *x)
 {
 	GENERAL_NAMES *gens;
 	STACK_OF(X509_EXTENSION) *exts;
-	STACK *ret;
+	STACK_OF(OPENSSL_STRING) *ret;
+
 	exts = X509_REQ_get_extensions(x);
 	gens = X509V3_get_d2i(exts, NID_subject_alt_name, NULL, NULL);
 	ret = get_email(X509_REQ_get_subject_name(x), gens);
@@ -511,9 +514,9 @@ STACK *X509_REQ_get1_email(X509_REQ *x)
 }
 
 
-static STACK *get_email(X509_NAME *name, GENERAL_NAMES *gens)
+static STACK_OF(OPENSSL_STRING) *get_email(X509_NAME *name, GENERAL_NAMES *gens)
 {
-	STACK *ret = NULL;
+	STACK_OF(OPENSSL_STRING) *ret = NULL;
 	X509_NAME_ENTRY *ne;
 	ASN1_IA5STRING *email;
 	GENERAL_NAME *gen;
@@ -536,23 +539,23 @@ static STACK *get_email(X509_NAME *name, GENERAL_NAMES *gens)
 	return ret;
 }
 
-static void str_free(void *str)
+static void str_free(OPENSSL_STRING str)
 {
 	OPENSSL_free(str);
 }
 
-static int append_ia5(STACK **sk, ASN1_IA5STRING *email)
+static int append_ia5(STACK_OF(OPENSSL_STRING) **sk, ASN1_IA5STRING *email)
 {
 	char *emtmp;
 	/* First some sanity checks */
 	if(email->type != V_ASN1_IA5STRING) return 1;
 	if(!email->data || !email->length) return 1;
-	if(!*sk) *sk = sk_new(sk_strcmp);
+	if(!*sk) *sk = sk_OPENSSL_STRING_new(sk_strcmp);
 	if(!*sk) return 0;
 	/* Don't add duplicates */
-	if(sk_find(*sk, (char *)email->data) != -1) return 1;
+	if(sk_OPENSSL_STRING_find(*sk, (char *)email->data) != -1) return 1;
 	emtmp = BUF_strdup((char *)email->data);
-	if(!emtmp || !sk_push(*sk, emtmp)) {
+	if(!emtmp || !sk_OPENSSL_STRING_push(*sk, emtmp)) {
 		X509_email_free(*sk);
 		*sk = NULL;
 		return 0;
@@ -560,9 +563,9 @@ static int append_ia5(STACK **sk, ASN1_IA5STRING *email)
 	return 1;
 }
 
-void X509_email_free(STACK *sk)
+void X509_email_free(STACK_OF(OPENSSL_STRING) *sk)
 {
-	sk_pop_free(sk, str_free);
+	sk_OPENSSL_STRING_pop_free(sk, str_free);
 }
 
 /* Convert IP addresses both IPv4 and IPv6 into an 
diff --git a/crypto/x509v3/v3err.c b/crypto/x509v3/v3err.c
index d538ad8b805c..f9f6f1f91f4d 100644
--- a/crypto/x509v3/v3err.c
+++ b/crypto/x509v3/v3err.c
@@ -1,6 +1,6 @@
 /* crypto/x509v3/v3err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,6 +70,7 @@
 
 static ERR_STRING_DATA X509V3_str_functs[]=
 	{
+{ERR_FUNC(X509V3_F_A2I_GENERAL_NAME),	"A2I_GENERAL_NAME"},
 {ERR_FUNC(X509V3_F_ASIDENTIFIERCHOICE_CANONIZE),	"ASIDENTIFIERCHOICE_CANONIZE"},
 {ERR_FUNC(X509V3_F_ASIDENTIFIERCHOICE_IS_CANONICAL),	"ASIDENTIFIERCHOICE_IS_CANONICAL"},
 {ERR_FUNC(X509V3_F_COPY_EMAIL),	"COPY_EMAIL"},
@@ -79,6 +80,7 @@ static ERR_STRING_DATA X509V3_str_functs[]=
 {ERR_FUNC(X509V3_F_DO_EXT_I2D),	"DO_EXT_I2D"},
 {ERR_FUNC(X509V3_F_DO_EXT_NCONF),	"DO_EXT_NCONF"},
 {ERR_FUNC(X509V3_F_DO_I2V_NAME_CONSTRAINTS),	"DO_I2V_NAME_CONSTRAINTS"},
+{ERR_FUNC(X509V3_F_GNAMES_FROM_SECTNAME),	"GNAMES_FROM_SECTNAME"},
 {ERR_FUNC(X509V3_F_HEX_TO_STRING),	"hex_to_string"},
 {ERR_FUNC(X509V3_F_I2S_ASN1_ENUMERATED),	"i2s_ASN1_ENUMERATED"},
 {ERR_FUNC(X509V3_F_I2S_ASN1_IA5STRING),	"I2S_ASN1_IA5STRING"},
@@ -95,6 +97,7 @@ static ERR_STRING_DATA X509V3_str_functs[]=
 {ERR_FUNC(X509V3_F_S2I_ASN1_OCTET_STRING),	"s2i_ASN1_OCTET_STRING"},
 {ERR_FUNC(X509V3_F_S2I_ASN1_SKEY_ID),	"S2I_ASN1_SKEY_ID"},
 {ERR_FUNC(X509V3_F_S2I_SKEY_ID),	"S2I_SKEY_ID"},
+{ERR_FUNC(X509V3_F_SET_DIST_POINT_NAME),	"SET_DIST_POINT_NAME"},
 {ERR_FUNC(X509V3_F_STRING_TO_HEX),	"string_to_hex"},
 {ERR_FUNC(X509V3_F_SXNET_ADD_ID_ASC),	"SXNET_add_id_asc"},
 {ERR_FUNC(X509V3_F_SXNET_ADD_ID_INTEGER),	"SXNET_add_id_INTEGER"},
@@ -110,6 +113,7 @@ static ERR_STRING_DATA X509V3_str_functs[]=
 {ERR_FUNC(X509V3_F_V2I_EXTENDED_KEY_USAGE),	"V2I_EXTENDED_KEY_USAGE"},
 {ERR_FUNC(X509V3_F_V2I_GENERAL_NAMES),	"v2i_GENERAL_NAMES"},
 {ERR_FUNC(X509V3_F_V2I_GENERAL_NAME_EX),	"v2i_GENERAL_NAME_ex"},
+{ERR_FUNC(X509V3_F_V2I_IDP),	"V2I_IDP"},
 {ERR_FUNC(X509V3_F_V2I_IPADDRBLOCKS),	"V2I_IPADDRBLOCKS"},
 {ERR_FUNC(X509V3_F_V2I_ISSUER_ALT),	"V2I_ISSUER_ALT"},
 {ERR_FUNC(X509V3_F_V2I_NAME_CONSTRAINTS),	"V2I_NAME_CONSTRAINTS"},
@@ -141,6 +145,7 @@ static ERR_STRING_DATA X509V3_str_reasons[]=
 {ERR_REASON(X509V3_R_BN_DEC2BN_ERROR)    ,"bn dec2bn error"},
 {ERR_REASON(X509V3_R_BN_TO_ASN1_INTEGER_ERROR),"bn to asn1 integer error"},
 {ERR_REASON(X509V3_R_DIRNAME_ERROR)      ,"dirname error"},
+{ERR_REASON(X509V3_R_DISTPOINT_ALREADY_SET),"distpoint already set"},
 {ERR_REASON(X509V3_R_DUPLICATE_ZONE_ID)  ,"duplicate zone id"},
 {ERR_REASON(X509V3_R_ERROR_CONVERTING_ZONE),"error converting zone"},
 {ERR_REASON(X509V3_R_ERROR_CREATING_EXTENSION),"error creating extension"},
@@ -154,6 +159,7 @@ static ERR_STRING_DATA X509V3_str_reasons[]=
 {ERR_REASON(X509V3_R_ILLEGAL_EMPTY_EXTENSION),"illegal empty extension"},
 {ERR_REASON(X509V3_R_ILLEGAL_HEX_DIGIT)  ,"illegal hex digit"},
 {ERR_REASON(X509V3_R_INCORRECT_POLICY_SYNTAX_TAG),"incorrect policy syntax tag"},
+{ERR_REASON(X509V3_R_INVALID_MULTIPLE_RDNS),"invalid multiple rdns"},
 {ERR_REASON(X509V3_R_INVALID_ASNUMBER)   ,"invalid asnumber"},
 {ERR_REASON(X509V3_R_INVALID_ASRANGE)    ,"invalid asrange"},
 {ERR_REASON(X509V3_R_INVALID_BOOLEAN_STRING),"invalid boolean string"},
@@ -187,9 +193,9 @@ static ERR_STRING_DATA X509V3_str_reasons[]=
 {ERR_REASON(X509V3_R_ODD_NUMBER_OF_DIGITS),"odd number of digits"},
 {ERR_REASON(X509V3_R_OPERATION_NOT_DEFINED),"operation not defined"},
 {ERR_REASON(X509V3_R_OTHERNAME_ERROR)    ,"othername error"},
-{ERR_REASON(X509V3_R_POLICY_LANGUAGE_ALREADTY_DEFINED),"policy language alreadty defined"},
+{ERR_REASON(X509V3_R_POLICY_LANGUAGE_ALREADY_DEFINED),"policy language already defined"},
 {ERR_REASON(X509V3_R_POLICY_PATH_LENGTH) ,"policy path length"},
-{ERR_REASON(X509V3_R_POLICY_PATH_LENGTH_ALREADTY_DEFINED),"policy path length alreadty defined"},
+{ERR_REASON(X509V3_R_POLICY_PATH_LENGTH_ALREADY_DEFINED),"policy path length already defined"},
 {ERR_REASON(X509V3_R_POLICY_SYNTAX_NOT_CURRENTLY_SUPPORTED),"policy syntax not currently supported"},
 {ERR_REASON(X509V3_R_POLICY_WHEN_PROXY_LANGUAGE_REQUIRES_NO_POLICY),"policy when proxy language requires no policy"},
 {ERR_REASON(X509V3_R_SECTION_NOT_FOUND)  ,"section not found"},
@@ -200,6 +206,7 @@ static ERR_STRING_DATA X509V3_str_reasons[]=
 {ERR_REASON(X509V3_R_UNKNOWN_EXTENSION_NAME),"unknown extension name"},
 {ERR_REASON(X509V3_R_UNKNOWN_OPTION)     ,"unknown option"},
 {ERR_REASON(X509V3_R_UNSUPPORTED_OPTION) ,"unsupported option"},
+{ERR_REASON(X509V3_R_UNSUPPORTED_TYPE)   ,"unsupported type"},
 {ERR_REASON(X509V3_R_USER_TOO_LONG)      ,"user too long"},
 {0,NULL}
 	};
diff --git a/crypto/x509v3/x509v3.h b/crypto/x509v3/x509v3.h
index 9ef83da755e4..b308abe7cd47 100644
--- a/crypto/x509v3/x509v3.h
+++ b/crypto/x509v3/x509v3.h
@@ -76,12 +76,19 @@ typedef void * (*X509V3_EXT_NEW)(void);
 typedef void (*X509V3_EXT_FREE)(void *);
 typedef void * (*X509V3_EXT_D2I)(void *, const unsigned char ** , long);
 typedef int (*X509V3_EXT_I2D)(void *, unsigned char **);
-typedef STACK_OF(CONF_VALUE) * (*X509V3_EXT_I2V)(struct v3_ext_method *method, void *ext, STACK_OF(CONF_VALUE) *extlist);
-typedef void * (*X509V3_EXT_V2I)(struct v3_ext_method *method, struct v3_ext_ctx *ctx, STACK_OF(CONF_VALUE) *values);
-typedef char * (*X509V3_EXT_I2S)(struct v3_ext_method *method, void *ext);
-typedef void * (*X509V3_EXT_S2I)(struct v3_ext_method *method, struct v3_ext_ctx *ctx, const char *str);
-typedef int (*X509V3_EXT_I2R)(struct v3_ext_method *method, void *ext, BIO *out, int indent);
-typedef void * (*X509V3_EXT_R2I)(struct v3_ext_method *method, struct v3_ext_ctx *ctx, const char *str);
+typedef STACK_OF(CONF_VALUE) *
+  (*X509V3_EXT_I2V)(const struct v3_ext_method *method, void *ext,
+		    STACK_OF(CONF_VALUE) *extlist);
+typedef void * (*X509V3_EXT_V2I)(const struct v3_ext_method *method,
+				 struct v3_ext_ctx *ctx,
+				 STACK_OF(CONF_VALUE) *values);
+typedef char * (*X509V3_EXT_I2S)(const struct v3_ext_method *method, void *ext);
+typedef void * (*X509V3_EXT_S2I)(const struct v3_ext_method *method,
+				 struct v3_ext_ctx *ctx, const char *str);
+typedef int (*X509V3_EXT_I2R)(const struct v3_ext_method *method, void *ext,
+			      BIO *out, int indent);
+typedef void * (*X509V3_EXT_R2I)(const struct v3_ext_method *method,
+				 struct v3_ext_ctx *ctx, const char *str);
 
 /* V3 extension structure */
 
@@ -220,24 +227,41 @@ union {
 	GENERAL_NAMES *fullname;
 	STACK_OF(X509_NAME_ENTRY) *relativename;
 } name;
+/* If relativename then this contains the full distribution point name */
+X509_NAME *dpname;
 } DIST_POINT_NAME;
-
-typedef struct DIST_POINT_st {
+/* All existing reasons */
+#define CRLDP_ALL_REASONS	0x807f
+
+#define CRL_REASON_NONE				-1
+#define CRL_REASON_UNSPECIFIED			0
+#define CRL_REASON_KEY_COMPROMISE		1
+#define CRL_REASON_CA_COMPROMISE		2
+#define CRL_REASON_AFFILIATION_CHANGED		3
+#define CRL_REASON_SUPERSEDED			4
+#define CRL_REASON_CESSATION_OF_OPERATION	5
+#define CRL_REASON_CERTIFICATE_HOLD		6
+#define CRL_REASON_REMOVE_FROM_CRL		8
+#define CRL_REASON_PRIVILEGE_WITHDRAWN		9
+#define CRL_REASON_AA_COMPROMISE		10
+
+struct DIST_POINT_st {
 DIST_POINT_NAME	*distpoint;
 ASN1_BIT_STRING *reasons;
 GENERAL_NAMES *CRLissuer;
-} DIST_POINT;
+int dp_reasons;
+};
 
 typedef STACK_OF(DIST_POINT) CRL_DIST_POINTS;
 
 DECLARE_STACK_OF(DIST_POINT)
 DECLARE_ASN1_SET_OF(DIST_POINT)
 
-typedef struct AUTHORITY_KEYID_st {
+struct AUTHORITY_KEYID_st {
 ASN1_OCTET_STRING *keyid;
 GENERAL_NAMES *issuer;
 ASN1_INTEGER *serial;
-} AUTHORITY_KEYID;
+};
 
 /* Strong extranet structures */
 
@@ -303,10 +327,10 @@ typedef struct GENERAL_SUBTREE_st {
 
 DECLARE_STACK_OF(GENERAL_SUBTREE)
 
-typedef struct NAME_CONSTRAINTS_st {
+struct NAME_CONSTRAINTS_st {
 	STACK_OF(GENERAL_SUBTREE) *permittedSubtrees;
 	STACK_OF(GENERAL_SUBTREE) *excludedSubtrees;
-} NAME_CONSTRAINTS;
+};
 
 typedef struct POLICY_CONSTRAINTS_st {
 	ASN1_INTEGER *requireExplicitPolicy;
@@ -329,6 +353,31 @@ typedef struct PROXY_CERT_INFO_EXTENSION_st
 DECLARE_ASN1_FUNCTIONS(PROXY_POLICY)
 DECLARE_ASN1_FUNCTIONS(PROXY_CERT_INFO_EXTENSION)
 
+struct ISSUING_DIST_POINT_st
+	{
+	DIST_POINT_NAME *distpoint;
+	int onlyuser;
+	int onlyCA;
+	ASN1_BIT_STRING *onlysomereasons;
+	int indirectCRL;
+	int onlyattr;
+	};
+
+/* Values in idp_flags field */
+/* IDP present */
+#define	IDP_PRESENT	0x1
+/* IDP values inconsistent */
+#define IDP_INVALID	0x2
+/* onlyuser true */
+#define	IDP_ONLYUSER	0x4
+/* onlyCA true */
+#define	IDP_ONLYCA	0x8
+/* onlyattr true */
+#define IDP_ONLYATTR	0x10
+/* indirectCRL true */
+#define IDP_INDIRECT	0x20
+/* onlysomereasons present */
+#define IDP_REASONS	0x40
 
 #define X509V3_conf_err(val) ERR_add_error_data(6, "section:", val->section, \
 ",name:", val->name, ",value:", val->value);
@@ -373,6 +422,7 @@ DECLARE_ASN1_FUNCTIONS(PROXY_CERT_INFO_EXTENSION)
 #define EXFLAG_PROXY		0x400
 
 #define EXFLAG_INVALID_POLICY	0x800
+#define EXFLAG_FRESHEST		0x1000
 
 #define KU_DIGITAL_SIGNATURE	0x0080
 #define KU_NON_REPUDIATION	0x0040
@@ -424,9 +474,10 @@ typedef struct x509_purpose_st {
 #define X509_PURPOSE_CRL_SIGN		6
 #define X509_PURPOSE_ANY		7
 #define X509_PURPOSE_OCSP_HELPER	8
+#define X509_PURPOSE_TIMESTAMP_SIGN	9
 
 #define X509_PURPOSE_MIN		1
-#define X509_PURPOSE_MAX		8
+#define X509_PURPOSE_MAX		9
 
 /* Flags for X509V3_EXT_print() */
 
@@ -471,6 +522,9 @@ DECLARE_ASN1_FUNCTIONS(AUTHORITY_KEYID)
 DECLARE_ASN1_FUNCTIONS(PKEY_USAGE_PERIOD)
 
 DECLARE_ASN1_FUNCTIONS(GENERAL_NAME)
+GENERAL_NAME *GENERAL_NAME_dup(GENERAL_NAME *a);
+int GENERAL_NAME_cmp(GENERAL_NAME *a, GENERAL_NAME *b);
+
 
 
 ASN1_BIT_STRING *v2i_ASN1_BIT_STRING(X509V3_EXT_METHOD *method,
@@ -486,11 +540,18 @@ DECLARE_ASN1_FUNCTIONS(GENERAL_NAMES)
 
 STACK_OF(CONF_VALUE) *i2v_GENERAL_NAMES(X509V3_EXT_METHOD *method,
 		GENERAL_NAMES *gen, STACK_OF(CONF_VALUE) *extlist);
-GENERAL_NAMES *v2i_GENERAL_NAMES(X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval);
+GENERAL_NAMES *v2i_GENERAL_NAMES(const X509V3_EXT_METHOD *method,
+				 X509V3_CTX *ctx, STACK_OF(CONF_VALUE) *nval);
 
 DECLARE_ASN1_FUNCTIONS(OTHERNAME)
 DECLARE_ASN1_FUNCTIONS(EDIPARTYNAME)
+int OTHERNAME_cmp(OTHERNAME *a, OTHERNAME *b);
+void GENERAL_NAME_set0_value(GENERAL_NAME *a, int type, void *value);
+void *GENERAL_NAME_get0_value(GENERAL_NAME *a, int *ptype);
+int GENERAL_NAME_set0_othername(GENERAL_NAME *gen,
+				ASN1_OBJECT *oid, ASN1_TYPE *value);
+int GENERAL_NAME_get0_otherName(GENERAL_NAME *gen, 
+				ASN1_OBJECT **poid, ASN1_TYPE **pvalue);
 
 char *i2s_ASN1_OCTET_STRING(X509V3_EXT_METHOD *method, ASN1_OCTET_STRING *ia5);
 ASN1_OCTET_STRING *s2i_ASN1_OCTET_STRING(X509V3_EXT_METHOD *method, X509V3_CTX *ctx, char *str);
@@ -507,6 +568,11 @@ DECLARE_ASN1_FUNCTIONS(NOTICEREF)
 DECLARE_ASN1_FUNCTIONS(CRL_DIST_POINTS)
 DECLARE_ASN1_FUNCTIONS(DIST_POINT)
 DECLARE_ASN1_FUNCTIONS(DIST_POINT_NAME)
+DECLARE_ASN1_FUNCTIONS(ISSUING_DIST_POINT)
+
+int DIST_POINT_set_dpname(DIST_POINT_NAME *dpn, X509_NAME *iname);
+
+int NAME_CONSTRAINTS_check(X509 *x, NAME_CONSTRAINTS *nc);
 
 DECLARE_ASN1_FUNCTIONS(ACCESS_DESCRIPTION)
 DECLARE_ASN1_FUNCTIONS(AUTHORITY_INFO_ACCESS)
@@ -524,11 +590,16 @@ DECLARE_ASN1_ALLOC_FUNCTIONS(NAME_CONSTRAINTS)
 DECLARE_ASN1_ALLOC_FUNCTIONS(POLICY_CONSTRAINTS)
 DECLARE_ASN1_ITEM(POLICY_CONSTRAINTS)
 
+GENERAL_NAME *a2i_GENERAL_NAME(GENERAL_NAME *out,
+			       const X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
+			       int gen_type, char *value, int is_nc);
+
 #ifdef HEADER_CONF_H
-GENERAL_NAME *v2i_GENERAL_NAME(X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
-							CONF_VALUE *cnf);
-GENERAL_NAME *v2i_GENERAL_NAME_ex(GENERAL_NAME *out, X509V3_EXT_METHOD *method,
-				X509V3_CTX *ctx, CONF_VALUE *cnf, int is_nc);
+GENERAL_NAME *v2i_GENERAL_NAME(const X509V3_EXT_METHOD *method, X509V3_CTX *ctx,
+			       CONF_VALUE *cnf);
+GENERAL_NAME *v2i_GENERAL_NAME_ex(GENERAL_NAME *out,
+				  const X509V3_EXT_METHOD *method,
+				  X509V3_CTX *ctx, CONF_VALUE *cnf, int is_nc);
 void X509V3_conf_free(CONF_VALUE *val);
 
 X509_EXTENSION *X509V3_EXT_nconf_nid(CONF *conf, X509V3_CTX *ctx, int ext_nid, char *value);
@@ -538,18 +609,23 @@ int X509V3_EXT_add_nconf(CONF *conf, X509V3_CTX *ctx, char *section, X509 *cert)
 int X509V3_EXT_REQ_add_nconf(CONF *conf, X509V3_CTX *ctx, char *section, X509_REQ *req);
 int X509V3_EXT_CRL_add_nconf(CONF *conf, X509V3_CTX *ctx, char *section, X509_CRL *crl);
 
-X509_EXTENSION *X509V3_EXT_conf_nid(LHASH *conf, X509V3_CTX *ctx, int ext_nid, char *value);
-X509_EXTENSION *X509V3_EXT_conf(LHASH *conf, X509V3_CTX *ctx, char *name, char *value);
-int X509V3_EXT_add_conf(LHASH *conf, X509V3_CTX *ctx, char *section, X509 *cert);
-int X509V3_EXT_REQ_add_conf(LHASH *conf, X509V3_CTX *ctx, char *section, X509_REQ *req);
-int X509V3_EXT_CRL_add_conf(LHASH *conf, X509V3_CTX *ctx, char *section, X509_CRL *crl);
+X509_EXTENSION *X509V3_EXT_conf_nid(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+				    int ext_nid, char *value);
+X509_EXTENSION *X509V3_EXT_conf(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+				char *name, char *value);
+int X509V3_EXT_add_conf(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+			char *section, X509 *cert);
+int X509V3_EXT_REQ_add_conf(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+			    char *section, X509_REQ *req);
+int X509V3_EXT_CRL_add_conf(LHASH_OF(CONF_VALUE) *conf, X509V3_CTX *ctx,
+			    char *section, X509_CRL *crl);
 
 int X509V3_add_value_bool_nf(char *name, int asn1_bool,
-						STACK_OF(CONF_VALUE) **extlist);
+			     STACK_OF(CONF_VALUE) **extlist);
 int X509V3_get_value_bool(CONF_VALUE *value, int *asn1_bool);
 int X509V3_get_value_int(CONF_VALUE *value, ASN1_INTEGER **aint);
 void X509V3_set_nconf(X509V3_CTX *ctx, CONF *conf);
-void X509V3_set_conf_lhash(X509V3_CTX *ctx, LHASH *lhash);
+void X509V3_set_conf_lhash(X509V3_CTX *ctx, LHASH_OF(CONF_VALUE) *lhash);
 #endif
 
 char * X509V3_get_string(X509V3_CTX *ctx, char *name, char *section);
@@ -576,8 +652,8 @@ int X509V3_EXT_add_list(X509V3_EXT_METHOD *extlist);
 int X509V3_EXT_add_alias(int nid_to, int nid_from);
 void X509V3_EXT_cleanup(void);
 
-X509V3_EXT_METHOD *X509V3_EXT_get(X509_EXTENSION *ext);
-X509V3_EXT_METHOD *X509V3_EXT_get_nid(int nid);
+const X509V3_EXT_METHOD *X509V3_EXT_get(X509_EXTENSION *ext);
+const X509V3_EXT_METHOD *X509V3_EXT_get_nid(int nid);
 int X509V3_add_standard_extensions(void);
 STACK_OF(CONF_VALUE) *X509V3_parse_list(const char *line);
 void *X509V3_EXT_d2i(X509_EXTENSION *ext);
@@ -587,8 +663,8 @@ void *X509V3_get_d2i(STACK_OF(X509_EXTENSION) *x, int nid, int *crit, int *idx);
 X509_EXTENSION *X509V3_EXT_i2d(int ext_nid, int crit, void *ext_struc);
 int X509V3_add1_i2d(STACK_OF(X509_EXTENSION) **x, int nid, void *value, int crit, unsigned long flags);
 
-char *hex_to_string(unsigned char *buffer, long len);
-unsigned char *string_to_hex(char *str, long *len);
+char *hex_to_string(const unsigned char *buffer, long len);
+unsigned char *string_to_hex(const char *str, long *len);
 int name_cmp(const char *name, const char *cmp);
 
 void X509V3_EXT_val_prn(BIO *out, STACK_OF(CONF_VALUE) *val, int indent,
@@ -603,6 +679,7 @@ int X509_check_purpose(X509 *x, int id, int ca);
 int X509_supported_extension(X509_EXTENSION *ex);
 int X509_PURPOSE_set(int *p, int purpose);
 int X509_check_issued(X509 *issuer, X509 *subject);
+int X509_check_akid(X509 *issuer, AUTHORITY_KEYID *akid);
 int X509_PURPOSE_get_count(void);
 X509_PURPOSE * X509_PURPOSE_get0(int idx);
 int X509_PURPOSE_get_by_sname(char *sname);
@@ -616,10 +693,10 @@ int X509_PURPOSE_get_trust(X509_PURPOSE *xp);
 void X509_PURPOSE_cleanup(void);
 int X509_PURPOSE_get_id(X509_PURPOSE *);
 
-STACK *X509_get1_email(X509 *x);
-STACK *X509_REQ_get1_email(X509_REQ *x);
-void X509_email_free(STACK *sk);
-STACK *X509_get1_ocsp(X509 *x);
+STACK_OF(OPENSSL_STRING) *X509_get1_email(X509 *x);
+STACK_OF(OPENSSL_STRING) *X509_REQ_get1_email(X509_REQ *x);
+void X509_email_free(STACK_OF(OPENSSL_STRING) *sk);
+STACK_OF(OPENSSL_STRING) *X509_get1_ocsp(X509 *x);
 
 ASN1_OCTET_STRING *a2i_IPADDRESS(const char *ipasc);
 ASN1_OCTET_STRING *a2i_IPADDRESS_NC(const char *ipasc);
@@ -628,6 +705,7 @@ int X509V3_NAME_from_section(X509_NAME *nm, STACK_OF(CONF_VALUE)*dn_sk,
 						unsigned long chtype);
 
 void X509_POLICY_NODE_print(BIO *out, X509_POLICY_NODE *node, int indent);
+DECLARE_STACK_OF(X509_POLICY_NODE)
 
 #ifndef OPENSSL_NO_RFC3779
 
@@ -787,8 +865,9 @@ void ERR_load_X509V3_strings(void);
 /* Error codes for the X509V3 functions. */
 
 /* Function codes. */
-#define X509V3_F_ASIDENTIFIERCHOICE_CANONIZE		 156
-#define X509V3_F_ASIDENTIFIERCHOICE_IS_CANONICAL	 157
+#define X509V3_F_A2I_GENERAL_NAME			 164
+#define X509V3_F_ASIDENTIFIERCHOICE_CANONIZE		 161
+#define X509V3_F_ASIDENTIFIERCHOICE_IS_CANONICAL	 162
 #define X509V3_F_COPY_EMAIL				 122
 #define X509V3_F_COPY_ISSUER				 123
 #define X509V3_F_DO_DIRNAME				 144
@@ -796,6 +875,7 @@ void ERR_load_X509V3_strings(void);
 #define X509V3_F_DO_EXT_I2D				 135
 #define X509V3_F_DO_EXT_NCONF				 151
 #define X509V3_F_DO_I2V_NAME_CONSTRAINTS		 148
+#define X509V3_F_GNAMES_FROM_SECTNAME			 156
 #define X509V3_F_HEX_TO_STRING				 111
 #define X509V3_F_I2S_ASN1_ENUMERATED			 121
 #define X509V3_F_I2S_ASN1_IA5STRING			 149
@@ -812,13 +892,14 @@ void ERR_load_X509V3_strings(void);
 #define X509V3_F_S2I_ASN1_OCTET_STRING			 112
 #define X509V3_F_S2I_ASN1_SKEY_ID			 114
 #define X509V3_F_S2I_SKEY_ID				 115
+#define X509V3_F_SET_DIST_POINT_NAME			 158
 #define X509V3_F_STRING_TO_HEX				 113
 #define X509V3_F_SXNET_ADD_ID_ASC			 125
 #define X509V3_F_SXNET_ADD_ID_INTEGER			 126
 #define X509V3_F_SXNET_ADD_ID_ULONG			 127
 #define X509V3_F_SXNET_GET_ID_ASC			 128
 #define X509V3_F_SXNET_GET_ID_ULONG			 129
-#define X509V3_F_V2I_ASIDENTIFIERS			 158
+#define X509V3_F_V2I_ASIDENTIFIERS			 163
 #define X509V3_F_V2I_ASN1_BIT_STRING			 101
 #define X509V3_F_V2I_AUTHORITY_INFO_ACCESS		 139
 #define X509V3_F_V2I_AUTHORITY_KEYID			 119
@@ -827,6 +908,7 @@ void ERR_load_X509V3_strings(void);
 #define X509V3_F_V2I_EXTENDED_KEY_USAGE			 103
 #define X509V3_F_V2I_GENERAL_NAMES			 118
 #define X509V3_F_V2I_GENERAL_NAME_EX			 117
+#define X509V3_F_V2I_IDP				 157
 #define X509V3_F_V2I_IPADDRBLOCKS			 159
 #define X509V3_F_V2I_ISSUER_ALT				 153
 #define X509V3_F_V2I_NAME_CONSTRAINTS			 147
@@ -855,6 +937,7 @@ void ERR_load_X509V3_strings(void);
 #define X509V3_R_BN_DEC2BN_ERROR			 100
 #define X509V3_R_BN_TO_ASN1_INTEGER_ERROR		 101
 #define X509V3_R_DIRNAME_ERROR				 149
+#define X509V3_R_DISTPOINT_ALREADY_SET			 160
 #define X509V3_R_DUPLICATE_ZONE_ID			 133
 #define X509V3_R_ERROR_CONVERTING_ZONE			 131
 #define X509V3_R_ERROR_CREATING_EXTENSION		 144
@@ -868,12 +951,13 @@ void ERR_load_X509V3_strings(void);
 #define X509V3_R_ILLEGAL_EMPTY_EXTENSION		 151
 #define X509V3_R_ILLEGAL_HEX_DIGIT			 113
 #define X509V3_R_INCORRECT_POLICY_SYNTAX_TAG		 152
-#define X509V3_R_INVALID_ASNUMBER			 160
-#define X509V3_R_INVALID_ASRANGE			 161
+#define X509V3_R_INVALID_MULTIPLE_RDNS			 161
+#define X509V3_R_INVALID_ASNUMBER			 162
+#define X509V3_R_INVALID_ASRANGE			 163
 #define X509V3_R_INVALID_BOOLEAN_STRING			 104
 #define X509V3_R_INVALID_EXTENSION_STRING		 105
-#define X509V3_R_INVALID_INHERITANCE			 162
-#define X509V3_R_INVALID_IPADDRESS			 163
+#define X509V3_R_INVALID_INHERITANCE			 165
+#define X509V3_R_INVALID_IPADDRESS			 166
 #define X509V3_R_INVALID_NAME				 106
 #define X509V3_R_INVALID_NULL_ARGUMENT			 107
 #define X509V3_R_INVALID_NULL_NAME			 108
@@ -901,9 +985,9 @@ void ERR_load_X509V3_strings(void);
 #define X509V3_R_ODD_NUMBER_OF_DIGITS			 112
 #define X509V3_R_OPERATION_NOT_DEFINED			 148
 #define X509V3_R_OTHERNAME_ERROR			 147
-#define X509V3_R_POLICY_LANGUAGE_ALREADTY_DEFINED	 155
+#define X509V3_R_POLICY_LANGUAGE_ALREADY_DEFINED	 155
 #define X509V3_R_POLICY_PATH_LENGTH			 156
-#define X509V3_R_POLICY_PATH_LENGTH_ALREADTY_DEFINED	 157
+#define X509V3_R_POLICY_PATH_LENGTH_ALREADY_DEFINED	 157
 #define X509V3_R_POLICY_SYNTAX_NOT_CURRENTLY_SUPPORTED	 158
 #define X509V3_R_POLICY_WHEN_PROXY_LANGUAGE_REQUIRES_NO_POLICY 159
 #define X509V3_R_SECTION_NOT_FOUND			 150
@@ -914,6 +998,7 @@ void ERR_load_X509V3_strings(void);
 #define X509V3_R_UNKNOWN_EXTENSION_NAME			 130
 #define X509V3_R_UNKNOWN_OPTION				 120
 #define X509V3_R_UNSUPPORTED_OPTION			 117
+#define X509V3_R_UNSUPPORTED_TYPE			 167
 #define X509V3_R_USER_TOO_LONG				 132
 
 #ifdef  __cplusplus
diff --git a/crypto/x86_64cpuid.pl b/crypto/x86_64cpuid.pl
index 2616a03da699..7b7b93b22322 100644
--- a/crypto/x86_64cpuid.pl
+++ b/crypto/x86_64cpuid.pl
@@ -1,108 +1,46 @@
 #!/usr/bin/env perl
 
-$output=shift;
-$masm=1 if ($output =~ /\.asm/);
-open STDOUT,">$output" || die "can't open $output: $!";
-
-print<<___ if(defined($masm));
-_TEXT	SEGMENT
-PUBLIC	OPENSSL_rdtsc
-
-PUBLIC	OPENSSL_atomic_add
-ALIGN	16
-OPENSSL_atomic_add	PROC
-	mov	eax,DWORD PTR[rcx]
-\$Lspin:	lea	r8,DWORD PTR[rdx+rax]
-lock	cmpxchg	DWORD PTR[rcx],r8d
-	jne	\$Lspin
-	mov	eax,r8d
-	cdqe    
-	ret
-OPENSSL_atomic_add	ENDP
-
-PUBLIC	OPENSSL_wipe_cpu
-ALIGN	16
-OPENSSL_wipe_cpu	PROC
-	pxor	xmm0,xmm0
-	pxor	xmm1,xmm1
-	pxor	xmm2,xmm2
-	pxor	xmm3,xmm3
-	pxor	xmm4,xmm4
-	pxor	xmm5,xmm5
-	xor	rcx,rcx
-	xor	rdx,rdx
-	xor	r8,r8
-	xor	r9,r9
-	xor	r10,r10
-	xor	r11,r11
-	lea	rax,QWORD PTR[rsp+8]
-	ret
-OPENSSL_wipe_cpu	ENDP
-_TEXT	ENDS
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
 
-CRT\$XIU	SEGMENT
-EXTRN	OPENSSL_cpuid_setup:PROC
-DQ	OPENSSL_cpuid_setup
-CRT\$XIU	ENDS
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output";
+
+($arg1,$arg2,$arg3,$arg4)=$win64?("%rcx","%rdx","%r8", "%r9") :	# Win64 order
+				 ("%rdi","%rsi","%rdx","%rcx");	# Unix order
+
+print<<___;
+.extern		OPENSSL_cpuid_setup
+.hidden		OPENSSL_cpuid_setup
+.section	.init
+	call	OPENSSL_cpuid_setup
+
+.hidden	OPENSSL_ia32cap_P
+.comm	OPENSSL_ia32cap_P,8,4
 
-___
-print<<___ if(!defined($masm));
 .text
 
 .globl	OPENSSL_atomic_add
-.type	OPENSSL_atomic_add,\@function
+.type	OPENSSL_atomic_add,\@abi-omnipotent
 .align	16
 OPENSSL_atomic_add:
-	movl	(%rdi),%eax
-.Lspin:	leaq	(%rsi,%rax),%r8
-lock;	cmpxchgl	%r8d,(%rdi)
+	movl	($arg1),%eax
+.Lspin:	leaq	($arg2,%rax),%r8
+	.byte	0xf0		# lock
+	cmpxchgl	%r8d,($arg1)
 	jne	.Lspin
 	movl	%r8d,%eax
-	.byte	0x48,0x98
+	.byte	0x48,0x98	# cltq/cdqe
 	ret
 .size	OPENSSL_atomic_add,.-OPENSSL_atomic_add
 
-.globl	OPENSSL_wipe_cpu
-.type	OPENSSL_wipe_cpu,\@function
-.align	16
-OPENSSL_wipe_cpu:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
-	pxor	%xmm8,%xmm8
-	pxor	%xmm9,%xmm9
-	pxor	%xmm10,%xmm10
-	pxor	%xmm11,%xmm11
-	pxor	%xmm12,%xmm12
-	pxor	%xmm13,%xmm13
-	pxor	%xmm14,%xmm14
-	pxor	%xmm15,%xmm15
-	xorq	%rcx,%rcx
-	xorq	%rdx,%rdx
-	xorq	%rsi,%rsi
-	xorq	%rdi,%rdi
-	xorq	%r8,%r8
-	xorq	%r9,%r9
-	xorq	%r10,%r10
-	xorq	%r11,%r11
-	leaq	8(%rsp),%rax
-	ret
-.size	OPENSSL_wipe_cpu,.-OPENSSL_wipe_cpu
-
-.section	.init
-	call	OPENSSL_cpuid_setup
-
-___
-
-open STDOUT,"| $^X perlasm/x86_64-xlate.pl $output";
-print<<___;
-.text
-
 .globl	OPENSSL_rdtsc
 .type	OPENSSL_rdtsc,\@abi-omnipotent
 .align	16
@@ -117,10 +55,12 @@ OPENSSL_rdtsc:
 .type	OPENSSL_ia32_cpuid,\@abi-omnipotent
 .align	16
 OPENSSL_ia32_cpuid:
-	mov	%rbx,%r8
+	mov	%rbx,%r8		# save %rbx
 
 	xor	%eax,%eax
 	cpuid
+	mov	%eax,%r11d		# max value for standard query level
+
 	xor	%eax,%eax
 	cmp	\$0x756e6547,%ebx	# "Genu"
 	setne	%al
@@ -130,30 +70,214 @@ OPENSSL_ia32_cpuid:
 	or	%eax,%r9d
 	cmp	\$0x6c65746e,%ecx	# "ntel"
 	setne	%al
-	or	%eax,%r9d
+	or	%eax,%r9d		# 0 indicates Intel CPU
+	jz	.Lintel
+
+	cmp	\$0x68747541,%ebx	# "Auth"
+	setne	%al
+	mov	%eax,%r10d
+	cmp	\$0x69746E65,%edx	# "enti"
+	setne	%al
+	or	%eax,%r10d
+	cmp	\$0x444D4163,%ecx	# "cAMD"
+	setne	%al
+	or	%eax,%r10d		# 0 indicates AMD CPU
+	jnz	.Lintel
+
+	# AMD specific
+	mov	\$0x80000000,%eax
+	cpuid
+	cmp	\$0x80000001,%eax
+	jb	.Lintel
+	mov	%eax,%r10d
+	mov	\$0x80000001,%eax
+	cpuid
+	or	%ecx,%r9d
+	and	\$0x00000801,%r9d	# isolate AMD XOP bit, 1<<11
+
+	cmp	\$0x80000008,%r10d
+	jb	.Lintel
+
+	mov	\$0x80000008,%eax
+	cpuid
+	movzb	%cl,%r10		# number of cores - 1
+	inc	%r10			# number of cores
 
 	mov	\$1,%eax
 	cpuid
+	bt	\$28,%edx		# test hyper-threading bit
+	jnc	.Lgeneric
+	shr	\$16,%ebx		# number of logical processors
+	cmp	%r10b,%bl
+	ja	.Lgeneric
+	and	\$0xefffffff,%edx	# ~(1<<28)
+	jmp	.Lgeneric
+
+.Lintel:
+	cmp	\$4,%r11d
+	mov	\$-1,%r10d
+	jb	.Lnocacheinfo
+
+	mov	\$4,%eax
+	mov	\$0,%ecx		# query L1D
+	cpuid
+	mov	%eax,%r10d
+	shr	\$14,%r10d
+	and	\$0xfff,%r10d		# number of cores -1 per L1D
+
+.Lnocacheinfo:
+	mov	\$1,%eax
+	cpuid
+	and	\$0xbfefffff,%edx	# force reserved bits to 0
 	cmp	\$0,%r9d
 	jne	.Lnotintel
-	or	\$0x00100000,%edx	# use reserved 20th bit to engage RC4_CHAR
+	or	\$0x40000000,%edx	# set reserved bit#30 on Intel CPUs
 	and	\$15,%ah
 	cmp	\$15,%ah		# examine Family ID
-	je	.Lnotintel
-	or	\$0x40000000,%edx	# use reserved bit to skip unrolled loop
+	jne	.Lnotintel
+	or	\$0x00100000,%edx	# set reserved bit#20 to engage RC4_CHAR
 .Lnotintel:
 	bt	\$28,%edx		# test hyper-threading bit
-	jnc	.Ldone
+	jnc	.Lgeneric
+	and	\$0xefffffff,%edx	# ~(1<<28)
+	cmp	\$0,%r10d
+	je	.Lgeneric
+
+	or	\$0x10000000,%edx	# 1<<28
 	shr	\$16,%ebx
 	cmp	\$1,%bl			# see if cache is shared
-	ja	.Ldone
+	ja	.Lgeneric
 	and	\$0xefffffff,%edx	# ~(1<<28)
+.Lgeneric:
+	and	\$0x00000800,%r9d	# isolate AMD XOP flag
+	and	\$0xfffff7ff,%ecx
+	or	%ecx,%r9d		# merge AMD XOP flag
+
+	mov	%edx,%r10d		# %r9d:%r10d is copy of %ecx:%edx
+	bt	\$27,%r9d		# check OSXSAVE bit
+	jnc	.Lclear_avx
+	xor	%ecx,%ecx		# XCR0
+	.byte	0x0f,0x01,0xd0		# xgetbv
+	and	\$6,%eax		# isolate XMM and YMM state support
+	cmp	\$6,%eax
+	je	.Ldone
+.Lclear_avx:
+	mov	\$0xefffe7ff,%eax	# ~(1<<28|1<<12|1<<11)
+	and	%eax,%r9d		# clear AVX, FMA and AMD XOP bits
 .Ldone:
-	shl	\$32,%rcx
-	mov	%edx,%eax
-	mov	%r8,%rbx
-	or	%rcx,%rax
+	shl	\$32,%r9
+	mov	%r10d,%eax
+	mov	%r8,%rbx		# restore %rbx
+	or	%r9,%rax
 	ret
 .size	OPENSSL_ia32_cpuid,.-OPENSSL_ia32_cpuid
+
+.globl  OPENSSL_cleanse
+.type   OPENSSL_cleanse,\@abi-omnipotent
+.align  16
+OPENSSL_cleanse:
+	xor	%rax,%rax
+	cmp	\$15,$arg2
+	jae	.Lot
+	cmp	\$0,$arg2
+	je	.Lret
+.Little:
+	mov	%al,($arg1)
+	sub	\$1,$arg2
+	lea	1($arg1),$arg1
+	jnz	.Little
+.Lret:
+	ret
+.align	16
+.Lot:
+	test	\$7,$arg1
+	jz	.Laligned
+	mov	%al,($arg1)
+	lea	-1($arg2),$arg2
+	lea	1($arg1),$arg1
+	jmp	.Lot
+.Laligned:
+	mov	%rax,($arg1)
+	lea	-8($arg2),$arg2
+	test	\$-8,$arg2
+	lea	8($arg1),$arg1
+	jnz	.Laligned
+	cmp	\$0,$arg2
+	jne	.Little
+	ret
+.size	OPENSSL_cleanse,.-OPENSSL_cleanse
 ___
+
+print<<___ if (!$win64);
+.globl	OPENSSL_wipe_cpu
+.type	OPENSSL_wipe_cpu,\@abi-omnipotent
+.align	16
+OPENSSL_wipe_cpu:
+	pxor	%xmm0,%xmm0
+	pxor	%xmm1,%xmm1
+	pxor	%xmm2,%xmm2
+	pxor	%xmm3,%xmm3
+	pxor	%xmm4,%xmm4
+	pxor	%xmm5,%xmm5
+	pxor	%xmm6,%xmm6
+	pxor	%xmm7,%xmm7
+	pxor	%xmm8,%xmm8
+	pxor	%xmm9,%xmm9
+	pxor	%xmm10,%xmm10
+	pxor	%xmm11,%xmm11
+	pxor	%xmm12,%xmm12
+	pxor	%xmm13,%xmm13
+	pxor	%xmm14,%xmm14
+	pxor	%xmm15,%xmm15
+	xorq	%rcx,%rcx
+	xorq	%rdx,%rdx
+	xorq	%rsi,%rsi
+	xorq	%rdi,%rdi
+	xorq	%r8,%r8
+	xorq	%r9,%r9
+	xorq	%r10,%r10
+	xorq	%r11,%r11
+	leaq	8(%rsp),%rax
+	ret
+.size	OPENSSL_wipe_cpu,.-OPENSSL_wipe_cpu
+___
+print<<___ if ($win64);
+.globl	OPENSSL_wipe_cpu
+.type	OPENSSL_wipe_cpu,\@abi-omnipotent
+.align	16
+OPENSSL_wipe_cpu:
+	pxor	%xmm0,%xmm0
+	pxor	%xmm1,%xmm1
+	pxor	%xmm2,%xmm2
+	pxor	%xmm3,%xmm3
+	pxor	%xmm4,%xmm4
+	pxor	%xmm5,%xmm5
+	xorq	%rcx,%rcx
+	xorq	%rdx,%rdx
+	xorq	%r8,%r8
+	xorq	%r9,%r9
+	xorq	%r10,%r10
+	xorq	%r11,%r11
+	leaq	8(%rsp),%rax
+	ret
+.size	OPENSSL_wipe_cpu,.-OPENSSL_wipe_cpu
+___
+
+print<<___;
+.globl	OPENSSL_ia32_rdrand
+.type	OPENSSL_ia32_rdrand,\@abi-omnipotent
+.align	16
+OPENSSL_ia32_rdrand:
+	mov	\$8,%ecx
+.Loop_rdrand:
+	rdrand	%rax
+	jc	.Lbreak_rdrand
+	loop	.Loop_rdrand
+.Lbreak_rdrand:
+	cmp	\$0,%rax
+	cmove	%rcx,%rax
+	ret
+.size	OPENSSL_ia32_rdrand,.-OPENSSL_ia32_rdrand
+___
+
 close STDOUT;	# flush
diff --git a/crypto/x86cpuid.pl b/crypto/x86cpuid.pl
index 4408ef2936ec..39fd8f22931d 100644
--- a/crypto/x86cpuid.pl
+++ b/crypto/x86cpuid.pl
@@ -1,6 +1,7 @@
 #!/usr/bin/env perl
 
-push(@INC,"perlasm");
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC, "${dir}perlasm", "perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],"x86cpuid");
@@ -18,42 +19,125 @@ for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); }
 	&pushf	();
 	&pop	("eax");
 	&xor	("ecx","eax");
-	&bt	("ecx",21);
-	&jnc	(&label("done"));
 	&xor	("eax","eax");
+	&bt	("ecx",21);
+	&jnc	(&label("nocpuid"));
 	&cpuid	();
+	&mov	("edi","eax");		# max value for standard query level
+
 	&xor	("eax","eax");
 	&cmp	("ebx",0x756e6547);	# "Genu"
-	&data_byte(0x0f,0x95,0xc0);	#&setne	(&LB("eax"));
+	&setne	(&LB("eax"));
 	&mov	("ebp","eax");
 	&cmp	("edx",0x49656e69);	# "ineI"
-	&data_byte(0x0f,0x95,0xc0);	#&setne	(&LB("eax"));
+	&setne	(&LB("eax"));
 	&or	("ebp","eax");
 	&cmp	("ecx",0x6c65746e);	# "ntel"
-	&data_byte(0x0f,0x95,0xc0);	#&setne	(&LB("eax"));
-	&or	("ebp","eax");
+	&setne	(&LB("eax"));
+	&or	("ebp","eax");		# 0 indicates Intel CPU
+	&jz	(&label("intel"));
+
+	&cmp	("ebx",0x68747541);	# "Auth"
+	&setne	(&LB("eax"));
+	&mov	("esi","eax");
+	&cmp	("edx",0x69746E65);	# "enti"
+	&setne	(&LB("eax"));
+	&or	("esi","eax");
+	&cmp	("ecx",0x444D4163);	# "cAMD"
+	&setne	(&LB("eax"));
+	&or	("esi","eax");		# 0 indicates AMD CPU
+	&jnz	(&label("intel"));
+
+	# AMD specific
+	&mov	("eax",0x80000000);
+	&cpuid	();
+	&cmp	("eax",0x80000001);
+	&jb	(&label("intel"));
+	&mov	("esi","eax");
+	&mov	("eax",0x80000001);
+	&cpuid	();
+	&or	("ebp","ecx");
+	&and	("ebp",1<<11|1);	# isolate XOP bit
+	&cmp	("esi",0x80000008);
+	&jb	(&label("intel"));
+
+	&mov	("eax",0x80000008);
+	&cpuid	();
+	&movz	("esi",&LB("ecx"));	# number of cores - 1
+	&inc	("esi");		# number of cores
+
 	&mov	("eax",1);
 	&cpuid	();
+	&bt	("edx",28);
+	&jnc	(&label("generic"));
+	&shr	("ebx",16);
+	&and	("ebx",0xff);
+	&cmp	("ebx","esi");
+	&ja	(&label("generic"));
+	&and	("edx",0xefffffff);	# clear hyper-threading bit
+	&jmp	(&label("generic"));
+	
+&set_label("intel");
+	&cmp	("edi",4);
+	&mov	("edi",-1);
+	&jb	(&label("nocacheinfo"));
+
+	&mov	("eax",4);
+	&mov	("ecx",0);		# query L1D
+	&cpuid	();
+	&mov	("edi","eax");
+	&shr	("edi",14);
+	&and	("edi",0xfff);		# number of cores -1 per L1D
+
+&set_label("nocacheinfo");
+	&mov	("eax",1);
+	&cpuid	();
+	&and	("edx",0xbfefffff);	# force reserved bits #20, #30 to 0
 	&cmp	("ebp",0);
-	&jne	(&label("notP4"));
-	&and	("eax",15<<8);		# familiy ID
-	&cmp	("eax",15<<8);		# P4?
-	&jne	(&label("notP4"));
-	&or	("edx",1<<20);		# use reserved bit to engage RC4_CHAR
-&set_label("notP4");
+	&jne	(&label("notintel"));
+	&or	("edx",1<<30);		# set reserved bit#30 on Intel CPUs
+	&and	(&HB("eax"),15);	# familiy ID
+	&cmp	(&HB("eax"),15);	# P4?
+	&jne	(&label("notintel"));
+	&or	("edx",1<<20);		# set reserved bit#20 to engage RC4_CHAR
+&set_label("notintel");
 	&bt	("edx",28);		# test hyper-threading bit
-	&jnc	(&label("done"));
+	&jnc	(&label("generic"));
+	&and	("edx",0xefffffff);
+	&cmp	("edi",0);
+	&je	(&label("generic"));
+
+	&or	("edx",0x10000000);
 	&shr	("ebx",16);
-	&and	("ebx",0xff);
-	&cmp	("ebx",1);		# see if cache is shared(*)
-	&ja	(&label("done"));
+	&cmp	(&LB("ebx"),1);
+	&ja	(&label("generic"));
 	&and	("edx",0xefffffff);	# clear hyper-threading bit if not
+
+&set_label("generic");
+	&and	("ebp",1<<11);		# isolate AMD XOP flag
+	&and	("ecx",0xfffff7ff);	# force 11th bit to 0
+	&mov	("esi","edx");
+	&or	("ebp","ecx");		# merge AMD XOP flag
+
+	&bt	("ecx",27);		# check OSXSAVE bit
+	&jnc	(&label("clear_avx"));
+	&xor	("ecx","ecx");
+	&data_byte(0x0f,0x01,0xd0);	# xgetbv
+	&and	("eax",6);
+	&cmp	("eax",6);
+	&je	(&label("done"));
+	&cmp	("eax",2);
+	&je	(&label("clear_avx"));
+&set_label("clear_xmm");
+	&and	("ebp",0xfdfffffd);	# clear AESNI and PCLMULQDQ bits
+	&and	("esi",0xfeffffff);	# clear FXSR
+&set_label("clear_avx");
+	&and	("ebp",0xefffe7ff);	# clear AVX, FMA and AMD XOP bits
 &set_label("done");
-	&mov	("eax","edx");
-	&mov	("edx","ecx");
+	&mov	("eax","esi");
+	&mov	("edx","ebp");
+&set_label("nocpuid");
 &function_end("OPENSSL_ia32_cpuid");
-# (*)	on Core2 this value is set to 2 denoting the fact that L2
-#	cache is shared between cores.
 
 &external_label("OPENSSL_ia32cap_P");
 
@@ -146,8 +230,9 @@ for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); }
 	&bt	(&DWP(0,"ecx"),1);
 	&jnc	(&label("no_x87"));
 	if ($sse2) {
-		&bt	(&DWP(0,"ecx"),26);
-		&jnc	(&label("no_sse2"));
+		&and	("ecx",1<<26|1<<24);	# check SSE2 and FXSR bits
+		&cmp	("ecx",1<<26|1<<24);
+		&jne	(&label("no_sse2"));
 		&pxor	("xmm0","xmm0");
 		&pxor	("xmm1","xmm1");
 		&pxor	("xmm2","xmm2");
@@ -220,6 +305,52 @@ for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); }
 	}
 &function_end_B("OPENSSL_indirect_call");
 
+&function_begin_B("OPENSSL_cleanse");
+	&mov	("edx",&wparam(0));
+	&mov	("ecx",&wparam(1));
+	&xor	("eax","eax");
+	&cmp	("ecx",7);
+	&jae	(&label("lot"));
+	&cmp	("ecx",0);
+	&je	(&label("ret"));
+&set_label("little");
+	&mov	(&BP(0,"edx"),"al");
+	&sub	("ecx",1);
+	&lea	("edx",&DWP(1,"edx"));
+	&jnz	(&label("little"));
+&set_label("ret");
+	&ret	();
+
+&set_label("lot",16);
+	&test	("edx",3);
+	&jz	(&label("aligned"));
+	&mov	(&BP(0,"edx"),"al");
+	&lea	("ecx",&DWP(-1,"ecx"));
+	&lea	("edx",&DWP(1,"edx"));
+	&jmp	(&label("lot"));
+&set_label("aligned");
+	&mov	(&DWP(0,"edx"),"eax");
+	&lea	("ecx",&DWP(-4,"ecx"));
+	&test	("ecx",-4);
+	&lea	("edx",&DWP(4,"edx"));
+	&jnz	(&label("aligned"));
+	&cmp	("ecx",0);
+	&jne	(&label("little"));
+	&ret	();
+&function_end_B("OPENSSL_cleanse");
+
+&function_begin_B("OPENSSL_ia32_rdrand");
+	&mov	("ecx",8);
+&set_label("loop");
+	&rdrand	("eax");
+	&jc	(&label("break"));
+	&loop	(&label("loop"));
+&set_label("break");
+	&cmp	("eax",0);
+	&cmove	("eax","ecx");
+	&ret	();
+&function_end_B("OPENSSL_ia32_rdrand");
+
 &initseg("OPENSSL_cpuid_setup");
 
 &asm_finish();