Import OpenSSL 1.0.1k.vendor/openssl/1.0.1k

46 files changed, 616 insertions, 2888 deletions

diff --git a/crypto/Makefile b/crypto/Makefile
index ee5bfbda82bc..2b6397a24b4d 100644
--- a/crypto/Makefile
+++ b/crypto/Makefile
@@ -55,12 +55,7 @@ top:
 all: shared
 
 buildinf.h: ../Makefile
-	( echo "#ifndef MK1MF_BUILD"; \
-	echo '  /* auto-generated by crypto/Makefile for crypto/cversion.c */'; \
-	echo '  #define CFLAGS "$(CC) $(CFLAG)"'; \
-	echo '  #define PLATFORM "$(PLATFORM)"'; \
-	echo "  #define DATE \"`LC_ALL=C LC_TIME=C date`\""; \
-	echo '#endif' ) >buildinf.h
+	$(PERL) $(TOP)/util/mkbuildinf.pl "$(CFLAGS)" "$(PLATFORM)" >buildinf.h
 
 x86cpuid.s:	x86cpuid.pl perlasm/x86asm.pl
 	$(PERL) x86cpuid.pl $(PERLASM_SCHEME) $(CFLAGS) $(PROCESSOR) > $@
diff --git a/crypto/aes/asm/aes-mips.pl b/crypto/aes/asm/aes-mips.pl
index e52395421b3c..537c8d3172b6 100755
--- a/crypto/aes/asm/aes-mips.pl
+++ b/crypto/aes/asm/aes-mips.pl
@@ -70,7 +70,7 @@ $pf = ($flavour =~ /nubi/i) ? $t0 : $t2;
 #
 ######################################################################
 
-$big_endian=(`echo MIPSEL | $ENV{CC} -E -P -`=~/MIPSEL/)?1:0;
+$big_endian=(`echo MIPSEL | $ENV{CC} -E -`=~/MIPSEL/)?1:0 if ($ENV{CC});
 
 for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);	}
 open STDOUT,">$output";
diff --git a/crypto/asn1/a_bitstr.c b/crypto/asn1/a_bitstr.c
index 34179960b87a..4117a67dc36f 100644
--- a/crypto/asn1/a_bitstr.c
+++ b/crypto/asn1/a_bitstr.c
@@ -136,11 +136,16 @@ ASN1_BIT_STRING *c2i_ASN1_BIT_STRING(ASN1_BIT_STRING **a,
 
 	p= *pp;
 	i= *(p++);
+	if (i > 7)
+		{
+		i=ASN1_R_INVALID_BIT_STRING_BITS_LEFT;
+		goto err;
+		}
 	/* We do this to preserve the settings.  If we modify
 	 * the settings, via the _set_bit function, we will recalculate
 	 * on output */
 	ret->flags&= ~(ASN1_STRING_FLAG_BITS_LEFT|0x07); /* clear */
-	ret->flags|=(ASN1_STRING_FLAG_BITS_LEFT|(i&0x07)); /* set */
+	ret->flags|=(ASN1_STRING_FLAG_BITS_LEFT|i); /* set */
 
 	if (len-- > 1) /* using one because of the bits left byte */
 		{
diff --git a/crypto/asn1/a_type.c b/crypto/asn1/a_type.c
index a45d2f9d12dd..5e1bc762d08b 100644
--- a/crypto/asn1/a_type.c
+++ b/crypto/asn1/a_type.c
@@ -113,7 +113,7 @@ 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 ASN1_TYPE_cmp(const ASN1_TYPE *a, const ASN1_TYPE *b)
 	{
 	int result = -1;
 
diff --git a/crypto/asn1/a_verify.c b/crypto/asn1/a_verify.c
index fc84cd3d191b..78dde1d05371 100644
--- a/crypto/asn1/a_verify.c
+++ b/crypto/asn1/a_verify.c
@@ -90,6 +90,12 @@ int ASN1_verify(i2d_of_void *i2d, X509_ALGOR *a, ASN1_BIT_STRING *signature,
 		ASN1err(ASN1_F_ASN1_VERIFY,ASN1_R_UNKNOWN_MESSAGE_DIGEST_ALGORITHM);
 		goto err;
 		}
+
+	if (signature->type == V_ASN1_BIT_STRING && signature->flags & 0x7)
+		{
+		ASN1err(ASN1_F_ASN1_VERIFY, ASN1_R_INVALID_BIT_STRING_BITS_LEFT);
+		goto err;
+		}
 	
 	inl=i2d(data,NULL);
 	buf_in=OPENSSL_malloc((unsigned int)inl);
@@ -146,6 +152,12 @@ int ASN1_item_verify(const ASN1_ITEM *it, X509_ALGOR *a,
 		return -1;
 		}
 
+	if (signature->type == V_ASN1_BIT_STRING && signature->flags & 0x7)
+		{
+		ASN1err(ASN1_F_ASN1_ITEM_VERIFY, ASN1_R_INVALID_BIT_STRING_BITS_LEFT);
+		return -1;
+		}
+
 	EVP_MD_CTX_init(&ctx);
 
 	/* Convert signature OID into digest and public key OIDs */
diff --git a/crypto/asn1/asn1.h b/crypto/asn1/asn1.h
index 220a0c8c637e..3c45d5d03f09 100644
--- a/crypto/asn1/asn1.h
+++ b/crypto/asn1/asn1.h
@@ -776,7 +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);
+int            ASN1_TYPE_cmp(const ASN1_TYPE *a, const ASN1_TYPE *b);
 
 ASN1_OBJECT *	ASN1_OBJECT_new(void );
 void		ASN1_OBJECT_free(ASN1_OBJECT *a);
@@ -1329,6 +1329,7 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_R_ILLEGAL_TIME_VALUE			 184
 #define ASN1_R_INTEGER_NOT_ASCII_FORMAT			 185
 #define ASN1_R_INTEGER_TOO_LARGE_FOR_LONG		 128
+#define ASN1_R_INVALID_BIT_STRING_BITS_LEFT		 220
 #define ASN1_R_INVALID_BMPSTRING_LENGTH			 129
 #define ASN1_R_INVALID_DIGIT				 130
 #define ASN1_R_INVALID_MIME_TYPE			 205
@@ -1378,6 +1379,7 @@ void ERR_load_ASN1_strings(void);
 #define ASN1_R_TIME_NOT_ASCII_FORMAT			 193
 #define ASN1_R_TOO_LONG					 155
 #define ASN1_R_TYPE_NOT_CONSTRUCTED			 156
+#define ASN1_R_TYPE_NOT_PRIMITIVE			 218
 #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
diff --git a/crypto/asn1/asn1_err.c b/crypto/asn1/asn1_err.c
index aa60203ba808..568a8416b811 100644
--- a/crypto/asn1/asn1_err.c
+++ b/crypto/asn1/asn1_err.c
@@ -1,6 +1,6 @@
 /* crypto/asn1/asn1_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2011 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2014 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
@@ -246,6 +246,7 @@ static ERR_STRING_DATA ASN1_str_reasons[]=
 {ERR_REASON(ASN1_R_ILLEGAL_TIME_VALUE)   ,"illegal time value"},
 {ERR_REASON(ASN1_R_INTEGER_NOT_ASCII_FORMAT),"integer not ascii format"},
 {ERR_REASON(ASN1_R_INTEGER_TOO_LARGE_FOR_LONG),"integer too large for long"},
+{ERR_REASON(ASN1_R_INVALID_BIT_STRING_BITS_LEFT),"invalid bit string bits left"},
 {ERR_REASON(ASN1_R_INVALID_BMPSTRING_LENGTH),"invalid bmpstring length"},
 {ERR_REASON(ASN1_R_INVALID_DIGIT)        ,"invalid digit"},
 {ERR_REASON(ASN1_R_INVALID_MIME_TYPE)    ,"invalid mime type"},
@@ -295,6 +296,7 @@ static ERR_STRING_DATA ASN1_str_reasons[]=
 {ERR_REASON(ASN1_R_TIME_NOT_ASCII_FORMAT),"time not ascii format"},
 {ERR_REASON(ASN1_R_TOO_LONG)             ,"too long"},
 {ERR_REASON(ASN1_R_TYPE_NOT_CONSTRUCTED) ,"type not constructed"},
+{ERR_REASON(ASN1_R_TYPE_NOT_PRIMITIVE)   ,"type not primitive"},
 {ERR_REASON(ASN1_R_UNABLE_TO_DECODE_RSA_KEY),"unable to decode rsa key"},
 {ERR_REASON(ASN1_R_UNABLE_TO_DECODE_RSA_PRIVATE_KEY),"unable to decode rsa private key"},
 {ERR_REASON(ASN1_R_UNEXPECTED_EOC)       ,"unexpected eoc"},
diff --git a/crypto/asn1/tasn_dec.c b/crypto/asn1/tasn_dec.c
index 87d7dfdf5c37..2cbfa8147596 100644
--- a/crypto/asn1/tasn_dec.c
+++ b/crypto/asn1/tasn_dec.c
@@ -870,6 +870,14 @@ static int asn1_d2i_ex_primitive(ASN1_VALUE **pval,
 		}
 	else if (cst)
 		{
+		if (utype == V_ASN1_NULL || utype == V_ASN1_BOOLEAN
+			|| utype == V_ASN1_OBJECT || utype == V_ASN1_INTEGER
+			|| utype == V_ASN1_ENUMERATED)
+			{
+			ASN1err(ASN1_F_ASN1_D2I_EX_PRIMITIVE,
+				ASN1_R_TYPE_NOT_PRIMITIVE);
+			return 0;
+			}
 		buf.length = 0;
 		buf.max = 0;
 		buf.data = NULL;
diff --git a/crypto/asn1/x_algor.c b/crypto/asn1/x_algor.c
index 274e456c7382..57cc956c20b2 100644
--- a/crypto/asn1/x_algor.c
+++ b/crypto/asn1/x_algor.c
@@ -142,3 +142,14 @@ void X509_ALGOR_set_md(X509_ALGOR *alg, const EVP_MD *md)
 	X509_ALGOR_set0(alg, OBJ_nid2obj(EVP_MD_type(md)), param_type, NULL);
 
 	}
+
+int X509_ALGOR_cmp(const X509_ALGOR *a, const X509_ALGOR *b)
+	{
+	int rv;
+	rv = OBJ_cmp(a->algorithm, b->algorithm);
+	if (rv)
+		return rv;
+	if (!a->parameter && !b->parameter)
+		return 0;
+	return ASN1_TYPE_cmp(a->parameter, b->parameter);
+	}
diff --git a/crypto/asn1/x_name.c b/crypto/asn1/x_name.c
index d7c2318693f5..22da57040e89 100644
--- a/crypto/asn1/x_name.c
+++ b/crypto/asn1/x_name.c
@@ -350,6 +350,8 @@ static int x509_name_canon(X509_NAME *a)
 			set = entry->set;
 			}
 		tmpentry = X509_NAME_ENTRY_new();
+		if (!tmpentry)
+			goto err;
 		tmpentry->object = OBJ_dup(entry->object);
 		if (!asn1_string_canon(tmpentry->value, entry->value))
 			goto err;
diff --git a/crypto/bio/bio.h b/crypto/bio/bio.h
index 05699ab21210..32eba7148056 100644
--- a/crypto/bio/bio.h
+++ b/crypto/bio/bio.h
@@ -175,6 +175,8 @@ extern "C" {
 #define BIO_CTRL_DGRAM_SET_NEXT_TIMEOUT   45 /* Next DTLS handshake timeout to
                                               * adjust socket timeouts */
 
+#define BIO_CTRL_DGRAM_GET_MTU_OVERHEAD   49
+
 #ifndef OPENSSL_NO_SCTP
 /* SCTP stuff */
 #define BIO_CTRL_DGRAM_SCTP_SET_IN_HANDSHAKE	50
@@ -607,6 +609,8 @@ int BIO_ctrl_reset_read_request(BIO *b);
          (int)BIO_ctrl(b, BIO_CTRL_DGRAM_GET_PEER, 0, (char *)peer)
 #define BIO_dgram_set_peer(b,peer) \
          (int)BIO_ctrl(b, BIO_CTRL_DGRAM_SET_PEER, 0, (char *)peer)
+#define BIO_dgram_get_mtu_overhead(b) \
+         (unsigned int)BIO_ctrl((b), BIO_CTRL_DGRAM_GET_MTU_OVERHEAD, 0, NULL)
 
 /* These two aren't currently implemented */
 /* int BIO_get_ex_num(BIO *bio); */
diff --git a/crypto/bio/bss_dgram.c b/crypto/bio/bss_dgram.c
index d9967e727267..0decf9449b3d 100644
--- a/crypto/bio/bss_dgram.c
+++ b/crypto/bio/bss_dgram.c
@@ -454,6 +454,36 @@ static int dgram_write(BIO *b, const char *in, int inl)
 	return(ret);
 	}
 
+static long dgram_get_mtu_overhead(bio_dgram_data *data)
+	{
+	long ret;
+
+	switch (data->peer.sa.sa_family)
+		{
+		case AF_INET:
+			/* Assume this is UDP - 20 bytes for IP, 8 bytes for UDP */
+			ret = 28;
+			break;
+#if OPENSSL_USE_IPV6
+		case AF_INET6:
+#ifdef IN6_IS_ADDR_V4MAPPED
+			if (IN6_IS_ADDR_V4MAPPED(&data->peer.sa_in6.sin6_addr))
+				/* Assume this is UDP - 20 bytes for IP, 8 bytes for UDP */
+				ret = 28;
+			else
+#endif
+				/* Assume this is UDP - 40 bytes for IP, 8 bytes for UDP */
+				ret = 48;
+			break;
+#endif
+		default:
+			/* We don't know. Go with the historical default */
+			ret = 28;
+			break;
+		}
+	return ret;
+	}
+
 static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 	{
 	long ret=1;
@@ -630,23 +660,24 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 #endif
 		break;
 	case BIO_CTRL_DGRAM_GET_FALLBACK_MTU:
+		ret = -dgram_get_mtu_overhead(data);
 		switch (data->peer.sa.sa_family)
 			{
 			case AF_INET:
-				ret = 576 - 20 - 8;
+				ret += 576;
 				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;
+					ret += 576;
 				else
 #endif
-					ret = 1280 - 40 - 8;
+					ret += 1280;
 				break;
 #endif
 			default:
-				ret = 576 - 20 - 8;
+				ret += 576;
 				break;
 			}
 		break;
@@ -847,6 +878,9 @@ static long dgram_ctrl(BIO *b, int cmd, long num, void *ptr)
 			ret = 0;
 		break;
 #endif
+	case BIO_CTRL_DGRAM_GET_MTU_OVERHEAD:
+		ret = dgram_get_mtu_overhead(data);
+		break;
 	default:
 		ret=0;
 		break;
@@ -893,10 +927,18 @@ BIO *BIO_new_dgram_sctp(int fd, int 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);
+	if (ret < 0)
+		{
+		BIO_vfree(bio);
+		return(NULL);
+		}
 	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);
+	if (ret < 0)
+		{
+		BIO_vfree(bio);
+		return(NULL);
+		}
 
 	/* Test if activation was successful. When using accept(),
 	 * SCTP-AUTH has to be activated for the listening socket
@@ -905,7 +947,13 @@ BIO *BIO_new_dgram_sctp(int fd, int close_flag)
 	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);
+
+	if (ret < 0)
+		{
+		OPENSSL_free(authchunks);
+		BIO_vfree(bio);
+		return(NULL);
+		}
 
 	for (p = (unsigned char*) authchunks->gauth_chunks;
 	     p < (unsigned char*) authchunks + sockopt_len;
@@ -927,16 +975,28 @@ BIO *BIO_new_dgram_sctp(int fd, int close_flag)
 	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);
+	if (ret < 0)
+		{
+		BIO_vfree(bio);
+		return(NULL);
+		}
 #else
 	sockopt_len = (socklen_t) sizeof(struct sctp_event_subscribe);
 	ret = getsockopt(fd, IPPROTO_SCTP, SCTP_EVENTS, &event, &sockopt_len);
-	OPENSSL_assert(ret >= 0);
+	if (ret < 0)
+		{
+		BIO_vfree(bio);
+		return(NULL);
+		}
 
 	event.sctp_authentication_event = 1;
 
 	ret = setsockopt(fd, IPPROTO_SCTP, SCTP_EVENTS, &event, sizeof(struct sctp_event_subscribe));
-	OPENSSL_assert(ret >= 0);
+	if (ret < 0)
+		{
+		BIO_vfree(bio);
+		return(NULL);
+		}
 #endif
 #endif
 
@@ -944,7 +1004,11 @@ BIO *BIO_new_dgram_sctp(int fd, int close_flag)
 	 * 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);
+	if (ret < 0)
+		{
+		BIO_vfree(bio);
+		return(NULL);
+		}
 
 	return(bio);
 	}
@@ -982,7 +1046,12 @@ static int dgram_sctp_free(BIO *a)
 		return 0;
 
 	data = (bio_dgram_sctp_data *)a->ptr;
-	if(data != NULL) OPENSSL_free(data);
+	if(data != NULL)
+		{
+		if(data->saved_message.data != NULL)
+			OPENSSL_free(data->saved_message.data);
+		OPENSSL_free(data);
+		}
 
 	return(1);
 	}
@@ -1034,6 +1103,13 @@ static int dgram_sctp_read(BIO *b, char *out, int outl)
 			msg.msg_flags = 0;
 			n = recvmsg(b->num, &msg, 0);
 
+			if (n <= 0)
+				{
+				if (n < 0)
+					ret = n;
+				break;
+				}
+
 			if (msg.msg_controllen > 0)
 				{
 				for (cmsg = CMSG_FIRSTHDR(&msg); cmsg; cmsg = CMSG_NXTHDR(&msg, cmsg))
@@ -1073,13 +1149,6 @@ static int dgram_sctp_read(BIO *b, char *out, int outl)
 					}
 				}
 
-			if (n <= 0)
-				{
-				if (n < 0)
-					ret = n;
-				break;
-				}
-
 			if (msg.msg_flags & MSG_NOTIFICATION)
 				{
 				snp = (union sctp_notification*) out;
@@ -1099,6 +1168,7 @@ static int dgram_sctp_read(BIO *b, char *out, int outl)
 						dgram_sctp_write(data->saved_message.bio, data->saved_message.data,
 						                 data->saved_message.length);
 						OPENSSL_free(data->saved_message.data);
+						data->saved_message.data = NULL;
 						data->saved_message.length = 0;
 						}
 
@@ -1109,16 +1179,28 @@ static int dgram_sctp_read(BIO *b, char *out, int outl)
 					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);
+					if (i < 0)
+						{
+						ret = i;
+						break;
+						}
 #else
 					eventsize = sizeof(struct sctp_event_subscribe);
 					i = getsockopt(b->num, IPPROTO_SCTP, SCTP_EVENTS, &event, &eventsize);
-					OPENSSL_assert(i >= 0);
+					if (i < 0)
+						{
+						ret = i;
+						break;
+						}
 
 					event.sctp_sender_dry_event = 0;
 
 					i = setsockopt(b->num, IPPROTO_SCTP, SCTP_EVENTS, &event, sizeof(struct sctp_event_subscribe));
-					OPENSSL_assert(i >= 0);
+					if (i < 0)
+						{
+						ret = i;
+						break;
+						}
 #endif
 					}
 
@@ -1151,8 +1233,8 @@ static int dgram_sctp_read(BIO *b, char *out, int outl)
 			 */
 			optlen = (socklen_t) sizeof(int);
 			ret = getsockopt(b->num, SOL_SOCKET, SO_RCVBUF, &optval, &optlen);
-			OPENSSL_assert(ret >= 0);
-			OPENSSL_assert(optval >= 18445);
+			if (ret >= 0)
+				OPENSSL_assert(optval >= 18445);
 
 			/* Test if SCTP doesn't partially deliver below
 			 * max record size (2^14 + 2048 + 13)
@@ -1160,8 +1242,8 @@ static int dgram_sctp_read(BIO *b, char *out, int outl)
 			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);
+			if (ret >= 0)
+				OPENSSL_assert(optval >= 18445);
 
 			/* Partially delivered notification??? Probably a bug.... */
 			OPENSSL_assert(!(msg.msg_flags & MSG_NOTIFICATION));
@@ -1195,15 +1277,15 @@ static int dgram_sctp_read(BIO *b, char *out, int outl)
 			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->gauth_chunks;
-				 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;
-				}
+			if (ii >= 0)
+				for (p = (unsigned char*) authchunks->gauth_chunks;
+				     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);
 
@@ -1258,9 +1340,11 @@ static int dgram_sctp_write(BIO *b, const char *in, int inl)
 	if (data->save_shutdown && !BIO_dgram_sctp_wait_for_dry(b))
 	{
 		data->saved_message.bio = b;
-		data->saved_message.length = inl;
+		if (data->saved_message.data)
+			OPENSSL_free(data->saved_message.data);
 		data->saved_message.data = OPENSSL_malloc(inl);
 		memcpy(data->saved_message.data, in, inl);
+		data->saved_message.length = inl;
 		return inl;
 	}
 
@@ -1367,6 +1451,10 @@ static long dgram_sctp_ctrl(BIO *b, int cmd, long num, void *ptr)
 		 * Returns always 1.
 		 */
 		break;
+	case BIO_CTRL_DGRAM_GET_MTU_OVERHEAD:
+		/* We allow transport protocol fragmentation so this is irrelevant */
+		ret = 0;
+		break;
 	case BIO_CTRL_DGRAM_SCTP_SET_IN_HANDSHAKE:
 		if (num > 0)
 			data->in_handshake = 1;
diff --git a/crypto/bn/asm/mips.pl b/crypto/bn/asm/mips.pl
index d2f3ef7bbf2c..215c9a748325 100755
--- a/crypto/bn/asm/mips.pl
+++ b/crypto/bn/asm/mips.pl
@@ -1872,6 +1872,41 @@ ___
 
 ($a_4,$a_5,$a_6,$a_7)=($b_0,$b_1,$b_2,$b_3);
 
+sub add_c2 () {
+my ($hi,$lo,$c0,$c1,$c2,
+    $warm,      # !$warm denotes first call with specific sequence of
+                # $c_[XYZ] when there is no Z-carry to accumulate yet;
+    $an,$bn     # these two are arguments for multiplication which
+                # result is used in *next* step [which is why it's
+                # commented as "forward multiplication" below];
+    )=@_;
+$code.=<<___;
+	mflo	$lo
+	mfhi	$hi
+	$ADDU	$c0,$lo
+	sltu	$at,$c0,$lo
+	 $MULTU	$an,$bn			# forward multiplication
+	$ADDU	$c0,$lo
+	$ADDU	$at,$hi
+	sltu	$lo,$c0,$lo
+	$ADDU	$c1,$at
+	$ADDU	$hi,$lo
+___
+$code.=<<___	if (!$warm);
+	sltu	$c2,$c1,$at
+	$ADDU	$c1,$hi
+	sltu	$hi,$c1,$hi
+	$ADDU	$c2,$hi
+___
+$code.=<<___	if ($warm);
+	sltu	$at,$c1,$at
+	$ADDU	$c1,$hi
+	$ADDU	$c2,$at
+	sltu	$hi,$c1,$hi
+	$ADDU	$c2,$hi
+___
+}
+
 $code.=<<___;
 
 .align	5
@@ -1920,21 +1955,10 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0,
+		$a_1,$a_1);		# mul_add_c(a[1],b[1],c3,c1,c2);
+$code.=<<___;
 	mflo	$t_1
 	mfhi	$t_2
 	$ADDU	$c_3,$t_1
@@ -1945,67 +1969,19 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0,
+		$a_1,$a_2);		# mul_add_c2(a[1],b[2],c1,c2,c3);
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1,
+		$a_4,$a_0);		# mul_add_c2(a[4],b[0],c2,c3,c1);
+$code.=<<___;
 	$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
+___
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0,
+		$a_3,$a_1);		# mul_add_c2(a[3],b[1],c2,c3,c1);
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1,
+		$a_2,$a_2);		# mul_add_c(a[2],b[2],c2,c3,c1);
+$code.=<<___;
 	mflo	$t_1
 	mfhi	$t_2
 	$ADDU	$c_2,$t_1
@@ -2016,97 +1992,23 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0,
+		$a_1,$a_4);		# mul_add_c2(a[1],b[4],c3,c1,c2);
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1,
+		$a_2,$a_3);		# mul_add_c2(a[2],b[3],c3,c1,c2);
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1,
+		$a_6,$a_0);		# mul_add_c2(a[6],b[0],c1,c2,c3);
+$code.=<<___;
 	$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
+___
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0,
+		$a_5,$a_1);		# mul_add_c2(a[5],b[1],c1,c2,c3);
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1,
+		$a_4,$a_2);		# mul_add_c2(a[4],b[2],c1,c2,c3);
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1,
+		$a_3,$a_3);		# mul_add_c(a[3],b[3],c1,c2,c3);
+$code.=<<___;
 	mflo	$t_1
 	mfhi	$t_2
 	$ADDU	$c_1,$t_1
@@ -2117,112 +2019,25 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0,
+		$a_1,$a_6);		# mul_add_c2(a[1],b[6],c2,c3,c1);
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1,
+		$a_2,$a_5);		# mul_add_c2(a[2],b[5],c2,c3,c1);
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1,
+		$a_3,$a_4);		# mul_add_c2(a[3],b[4],c2,c3,c1);
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1,
+		$a_7,$a_1);		# mul_add_c2(a[7],b[1],c3,c1,c2);
+$code.=<<___;
 	$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
+___
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0,
+		$a_6,$a_2);		# mul_add_c2(a[6],b[2],c3,c1,c2);
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1,
+		$a_5,$a_3);		# mul_add_c2(a[5],b[3],c3,c1,c2);
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1,
+		$a_4,$a_4);		# mul_add_c(a[4],b[4],c3,c1,c2);
+$code.=<<___;
 	mflo	$t_1
 	mfhi	$t_2
 	$ADDU	$c_3,$t_1
@@ -2233,82 +2048,21 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0,
+		$a_3,$a_6);		# mul_add_c2(a[3],b[6],c1,c2,c3);
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1,
+		$a_4,$a_5);		# mul_add_c2(a[4],b[5],c1,c2,c3);
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1,
+		$a_7,$a_3);		# mul_add_c2(a[7],b[3],c2,c3,c1);
+$code.=<<___;
 	$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
+___
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0,
+		$a_6,$a_4);		# mul_add_c2(a[6],b[4],c2,c3,c1);
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1,
+		$a_5,$a_5);		# mul_add_c(a[5],b[5],c2,c3,c1);
+$code.=<<___;
 	mflo	$t_1
 	mfhi	$t_2
 	$ADDU	$c_2,$t_1
@@ -2319,52 +2073,17 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0,
+		$a_5,$a_6);		# mul_add_c2(a[5],b[6],c3,c1,c2);
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1,
+		$a_7,$a_5);		# mul_add_c2(a[7],b[5],c1,c2,c3);
+$code.=<<___;
 	$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
+___
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0,
+		$a_6,$a_6);		# mul_add_c(a[6],b[6],c1,c2,c3);
+$code.=<<___;
 	mflo	$t_1
 	mfhi	$t_2
 	$ADDU	$c_1,$t_1
@@ -2375,21 +2094,10 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0,
+		$a_7,$a_7);		# mul_add_c(a[7],b[7],c3,c1,c2);
+$code.=<<___;
 	$ST	$c_2,13*$BNSZ($a0)
 
 	mflo	$t_1
@@ -2457,21 +2165,10 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0,
+		$a_1,$a_1);		# mul_add_c(a[1],b[1],c3,c1,c2);
+$code.=<<___;
 	mflo	$t_1
 	mfhi	$t_2
 	$ADDU	$c_3,$t_1
@@ -2482,52 +2179,17 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0,
+		$a_1,$a_2);		# mul_add_c2(a2[1],b[2],c1,c2,c3);
+	&add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1,
+		$a_3,$a_1);		# mul_add_c2(a[3],b[1],c2,c3,c1);
+$code.=<<___;
 	$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
+___
+	&add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0,
+		$a_2,$a_2);		# mul_add_c(a[2],b[2],c2,c3,c1);
+$code.=<<___;
 	mflo	$t_1
 	mfhi	$t_2
 	$ADDU	$c_2,$t_1
@@ -2538,21 +2200,10 @@ $code.=<<___;
 	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
+___
+	&add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0,
+		$a_3,$a_3);		# mul_add_c(a[3],b[3],c1,c2,c3);
+$code.=<<___;
 	$ST	$c_3,5*$BNSZ($a0)
 
 	mflo	$t_1
diff --git a/crypto/bn/asm/mips3.s b/crypto/bn/asm/mips3.s
deleted file mode 100644
index dca4105c7db1..000000000000
--- a/crypto/bn/asm/mips3.s
+++ /dev/null
@@ -1,2201 +0,0 @@
-.rdata
-.asciiz	"mips3.s, Version 1.1"
-.asciiz	"MIPS III/IV ISA artwork by Andy Polyakov <appro@fy.chalmers.se>"
-
-/*
- * ====================================================================
- * 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.
- * ====================================================================
- */
-
-/*
- * This is my modest contributon to the OpenSSL project (see
- * http://www.openssl.org/ for more information about it) and is
- * a drop-in MIPS III/IV ISA replacement for crypto/bn/bn_asm.c
- * module. For updates see http://fy.chalmers.se/~appro/hpe/.
- *
- * 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>
- */
-#include <asm.h>
-#include <regdef.h>
-
-#if _MIPS_ISA>=4
-#define	MOVNZ(cond,dst,src)	\
-	movn	dst,src,cond
-#else
-#define	MOVNZ(cond,dst,src)	\
-	.set	noreorder;	\
-	bnezl	cond,.+8;	\
-	move	dst,src;	\
-	.set	reorder
-#endif
-
-.text
-
-.set	noat
-.set	reorder
-
-#define	MINUS4	v1
-
-.align	5
-LEAF(bn_mul_add_words)
-	.set	noreorder
-	bgtzl	a2,.L_bn_mul_add_words_proceed
-	ld	t0,0(a1)
-	jr	ra
-	move	v0,zero
-	.set	reorder
-
-.L_bn_mul_add_words_proceed:
-	li	MINUS4,-4
-	and	ta0,a2,MINUS4
-	move	v0,zero
-	beqz	ta0,.L_bn_mul_add_words_tail
-
-.L_bn_mul_add_words_loop:
-	dmultu	t0,a3
-	ld	t1,0(a0)
-	ld	t2,8(a1)
-	ld	t3,8(a0)
-	ld	ta0,16(a1)
-	ld	ta1,16(a0)
-	daddu	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
-	daddu	t1,AT
-	daddu	v0,t0
-	sltu	AT,t1,AT
-	sd	t1,0(a0)
-	daddu	v0,AT
-
-	dmultu	t2,a3
-	ld	ta2,24(a1)
-	ld	ta3,24(a0)
-	daddu	t3,v0
-	sltu	v0,t3,v0
-	mflo	AT
-	mfhi	t2
-	daddu	t3,AT
-	daddu	v0,t2
-	sltu	AT,t3,AT
-	sd	t3,8(a0)
-	daddu	v0,AT
-
-	dmultu	ta0,a3
-	subu	a2,4
-	PTR_ADD	a0,32
-	PTR_ADD	a1,32
-	daddu	ta1,v0
-	sltu	v0,ta1,v0
-	mflo	AT
-	mfhi	ta0
-	daddu	ta1,AT
-	daddu	v0,ta0
-	sltu	AT,ta1,AT
-	sd	ta1,-16(a0)
-	daddu	v0,AT
-
-
-	dmultu	ta2,a3
-	and	ta0,a2,MINUS4
-	daddu	ta3,v0
-	sltu	v0,ta3,v0
-	mflo	AT
-	mfhi	ta2
-	daddu	ta3,AT
-	daddu	v0,ta2
-	sltu	AT,ta3,AT
-	sd	ta3,-8(a0)
-	daddu	v0,AT
-	.set	noreorder
-	bgtzl	ta0,.L_bn_mul_add_words_loop
-	ld	t0,0(a1)
-
-	bnezl	a2,.L_bn_mul_add_words_tail
-	ld	t0,0(a1)
-	.set	reorder
-
-.L_bn_mul_add_words_return:
-	jr	ra
-
-.L_bn_mul_add_words_tail:
-	dmultu	t0,a3
-	ld	t1,0(a0)
-	subu	a2,1
-	daddu	t1,v0
-	sltu	v0,t1,v0
-	mflo	AT
-	mfhi	t0
-	daddu	t1,AT
-	daddu	v0,t0
-	sltu	AT,t1,AT
-	sd	t1,0(a0)
-	daddu	v0,AT
-	beqz	a2,.L_bn_mul_add_words_return
-
-	ld	t0,8(a1)
-	dmultu	t0,a3
-	ld	t1,8(a0)
-	subu	a2,1
-	daddu	t1,v0
-	sltu	v0,t1,v0
-	mflo	AT
-	mfhi	t0
-	daddu	t1,AT
-	daddu	v0,t0
-	sltu	AT,t1,AT
-	sd	t1,8(a0)
-	daddu	v0,AT
-	beqz	a2,.L_bn_mul_add_words_return
-
-	ld	t0,16(a1)
-	dmultu	t0,a3
-	ld	t1,16(a0)
-	daddu	t1,v0
-	sltu	v0,t1,v0
-	mflo	AT
-	mfhi	t0
-	daddu	t1,AT
-	daddu	v0,t0
-	sltu	AT,t1,AT
-	sd	t1,16(a0)
-	daddu	v0,AT
-	jr	ra
-END(bn_mul_add_words)
-
-.align	5
-LEAF(bn_mul_words)
-	.set	noreorder
-	bgtzl	a2,.L_bn_mul_words_proceed
-	ld	t0,0(a1)
-	jr	ra
-	move	v0,zero
-	.set	reorder
-
-.L_bn_mul_words_proceed:
-	li	MINUS4,-4
-	and	ta0,a2,MINUS4
-	move	v0,zero
-	beqz	ta0,.L_bn_mul_words_tail
-
-.L_bn_mul_words_loop:
-	dmultu	t0,a3
-	ld	t2,8(a1)
-	ld	ta0,16(a1)
-	ld	ta2,24(a1)
-	mflo	AT
-	mfhi	t0
-	daddu	v0,AT
-	sltu	t1,v0,AT
-	sd	v0,0(a0)
-	daddu	v0,t1,t0
-
-	dmultu	t2,a3
-	subu	a2,4
-	PTR_ADD	a0,32
-	PTR_ADD	a1,32
-	mflo	AT
-	mfhi	t2
-	daddu	v0,AT
-	sltu	t3,v0,AT
-	sd	v0,-24(a0)
-	daddu	v0,t3,t2
-
-	dmultu	ta0,a3
-	mflo	AT
-	mfhi	ta0
-	daddu	v0,AT
-	sltu	ta1,v0,AT
-	sd	v0,-16(a0)
-	daddu	v0,ta1,ta0
-
-
-	dmultu	ta2,a3
-	and	ta0,a2,MINUS4
-	mflo	AT
-	mfhi	ta2
-	daddu	v0,AT
-	sltu	ta3,v0,AT
-	sd	v0,-8(a0)
-	daddu	v0,ta3,ta2
-	.set	noreorder
-	bgtzl	ta0,.L_bn_mul_words_loop
-	ld	t0,0(a1)
-
-	bnezl	a2,.L_bn_mul_words_tail
-	ld	t0,0(a1)
-	.set	reorder
-
-.L_bn_mul_words_return:
-	jr	ra
-
-.L_bn_mul_words_tail:
-	dmultu	t0,a3
-	subu	a2,1
-	mflo	AT
-	mfhi	t0
-	daddu	v0,AT
-	sltu	t1,v0,AT
-	sd	v0,0(a0)
-	daddu	v0,t1,t0
-	beqz	a2,.L_bn_mul_words_return
-
-	ld	t0,8(a1)
-	dmultu	t0,a3
-	subu	a2,1
-	mflo	AT
-	mfhi	t0
-	daddu	v0,AT
-	sltu	t1,v0,AT
-	sd	v0,8(a0)
-	daddu	v0,t1,t0
-	beqz	a2,.L_bn_mul_words_return
-
-	ld	t0,16(a1)
-	dmultu	t0,a3
-	mflo	AT
-	mfhi	t0
-	daddu	v0,AT
-	sltu	t1,v0,AT
-	sd	v0,16(a0)
-	daddu	v0,t1,t0
-	jr	ra
-END(bn_mul_words)
-
-.align	5
-LEAF(bn_sqr_words)
-	.set	noreorder
-	bgtzl	a2,.L_bn_sqr_words_proceed
-	ld	t0,0(a1)
-	jr	ra
-	move	v0,zero
-	.set	reorder
-
-.L_bn_sqr_words_proceed:
-	li	MINUS4,-4
-	and	ta0,a2,MINUS4
-	move	v0,zero
-	beqz	ta0,.L_bn_sqr_words_tail
-
-.L_bn_sqr_words_loop:
-	dmultu	t0,t0
-	ld	t2,8(a1)
-	ld	ta0,16(a1)
-	ld	ta2,24(a1)
-	mflo	t1
-	mfhi	t0
-	sd	t1,0(a0)
-	sd	t0,8(a0)
-
-	dmultu	t2,t2
-	subu	a2,4
-	PTR_ADD	a0,64
-	PTR_ADD	a1,32
-	mflo	t3
-	mfhi	t2
-	sd	t3,-48(a0)
-	sd	t2,-40(a0)
-
-	dmultu	ta0,ta0
-	mflo	ta1
-	mfhi	ta0
-	sd	ta1,-32(a0)
-	sd	ta0,-24(a0)
-
-
-	dmultu	ta2,ta2
-	and	ta0,a2,MINUS4
-	mflo	ta3
-	mfhi	ta2
-	sd	ta3,-16(a0)
-	sd	ta2,-8(a0)
-
-	.set	noreorder
-	bgtzl	ta0,.L_bn_sqr_words_loop
-	ld	t0,0(a1)
-
-	bnezl	a2,.L_bn_sqr_words_tail
-	ld	t0,0(a1)
-	.set	reorder
-
-.L_bn_sqr_words_return:
-	move	v0,zero
-	jr	ra
-
-.L_bn_sqr_words_tail:
-	dmultu	t0,t0
-	subu	a2,1
-	mflo	t1
-	mfhi	t0
-	sd	t1,0(a0)
-	sd	t0,8(a0)
-	beqz	a2,.L_bn_sqr_words_return
-
-	ld	t0,8(a1)
-	dmultu	t0,t0
-	subu	a2,1
-	mflo	t1
-	mfhi	t0
-	sd	t1,16(a0)
-	sd	t0,24(a0)
-	beqz	a2,.L_bn_sqr_words_return
-
-	ld	t0,16(a1)
-	dmultu	t0,t0
-	mflo	t1
-	mfhi	t0
-	sd	t1,32(a0)
-	sd	t0,40(a0)
-	jr	ra
-END(bn_sqr_words)
-
-.align	5
-LEAF(bn_add_words)
-	.set	noreorder
-	bgtzl	a3,.L_bn_add_words_proceed
-	ld	t0,0(a1)
-	jr	ra
-	move	v0,zero
-	.set	reorder
-
-.L_bn_add_words_proceed:
-	li	MINUS4,-4
-	and	AT,a3,MINUS4
-	move	v0,zero
-	beqz	AT,.L_bn_add_words_tail
-
-.L_bn_add_words_loop:
-	ld	ta0,0(a2)
-	subu	a3,4
-	ld	t1,8(a1)
-	and	AT,a3,MINUS4
-	ld	t2,16(a1)
-	PTR_ADD	a2,32
-	ld	t3,24(a1)
-	PTR_ADD	a0,32
-	ld	ta1,-24(a2)
-	PTR_ADD	a1,32
-	ld	ta2,-16(a2)
-	ld	ta3,-8(a2)
-	daddu	ta0,t0
-	sltu	t8,ta0,t0
-	daddu	t0,ta0,v0
-	sltu	v0,t0,ta0
-	sd	t0,-32(a0)
-	daddu	v0,t8
-
-	daddu	ta1,t1
-	sltu	t9,ta1,t1
-	daddu	t1,ta1,v0
-	sltu	v0,t1,ta1
-	sd	t1,-24(a0)
-	daddu	v0,t9
-
-	daddu	ta2,t2
-	sltu	t8,ta2,t2
-	daddu	t2,ta2,v0
-	sltu	v0,t2,ta2
-	sd	t2,-16(a0)
-	daddu	v0,t8
-	
-	daddu	ta3,t3
-	sltu	t9,ta3,t3
-	daddu	t3,ta3,v0
-	sltu	v0,t3,ta3
-	sd	t3,-8(a0)
-	daddu	v0,t9
-	
-	.set	noreorder
-	bgtzl	AT,.L_bn_add_words_loop
-	ld	t0,0(a1)
-
-	bnezl	a3,.L_bn_add_words_tail
-	ld	t0,0(a1)
-	.set	reorder
-
-.L_bn_add_words_return:
-	jr	ra
-
-.L_bn_add_words_tail:
-	ld	ta0,0(a2)
-	daddu	ta0,t0
-	subu	a3,1
-	sltu	t8,ta0,t0
-	daddu	t0,ta0,v0
-	sltu	v0,t0,ta0
-	sd	t0,0(a0)
-	daddu	v0,t8
-	beqz	a3,.L_bn_add_words_return
-
-	ld	t1,8(a1)
-	ld	ta1,8(a2)
-	daddu	ta1,t1
-	subu	a3,1
-	sltu	t9,ta1,t1
-	daddu	t1,ta1,v0
-	sltu	v0,t1,ta1
-	sd	t1,8(a0)
-	daddu	v0,t9
-	beqz	a3,.L_bn_add_words_return
-
-	ld	t2,16(a1)
-	ld	ta2,16(a2)
-	daddu	ta2,t2
-	sltu	t8,ta2,t2
-	daddu	t2,ta2,v0
-	sltu	v0,t2,ta2
-	sd	t2,16(a0)
-	daddu	v0,t8
-	jr	ra
-END(bn_add_words)
-
-.align	5
-LEAF(bn_sub_words)
-	.set	noreorder
-	bgtzl	a3,.L_bn_sub_words_proceed
-	ld	t0,0(a1)
-	jr	ra
-	move	v0,zero
-	.set	reorder
-
-.L_bn_sub_words_proceed:
-	li	MINUS4,-4
-	and	AT,a3,MINUS4
-	move	v0,zero
-	beqz	AT,.L_bn_sub_words_tail
-
-.L_bn_sub_words_loop:
-	ld	ta0,0(a2)
-	subu	a3,4
-	ld	t1,8(a1)
-	and	AT,a3,MINUS4
-	ld	t2,16(a1)
-	PTR_ADD	a2,32
-	ld	t3,24(a1)
-	PTR_ADD	a0,32
-	ld	ta1,-24(a2)
-	PTR_ADD	a1,32
-	ld	ta2,-16(a2)
-	ld	ta3,-8(a2)
-	sltu	t8,t0,ta0
-	dsubu	t0,ta0
-	dsubu	ta0,t0,v0
-	sd	ta0,-32(a0)
-	MOVNZ	(t0,v0,t8)
-
-	sltu	t9,t1,ta1
-	dsubu	t1,ta1
-	dsubu	ta1,t1,v0
-	sd	ta1,-24(a0)
-	MOVNZ	(t1,v0,t9)
-
-
-	sltu	t8,t2,ta2
-	dsubu	t2,ta2
-	dsubu	ta2,t2,v0
-	sd	ta2,-16(a0)
-	MOVNZ	(t2,v0,t8)
-
-	sltu	t9,t3,ta3
-	dsubu	t3,ta3
-	dsubu	ta3,t3,v0
-	sd	ta3,-8(a0)
-	MOVNZ	(t3,v0,t9)
-
-	.set	noreorder
-	bgtzl	AT,.L_bn_sub_words_loop
-	ld	t0,0(a1)
-
-	bnezl	a3,.L_bn_sub_words_tail
-	ld	t0,0(a1)
-	.set	reorder
-
-.L_bn_sub_words_return:
-	jr	ra
-
-.L_bn_sub_words_tail:
-	ld	ta0,0(a2)
-	subu	a3,1
-	sltu	t8,t0,ta0
-	dsubu	t0,ta0
-	dsubu	ta0,t0,v0
-	MOVNZ	(t0,v0,t8)
-	sd	ta0,0(a0)
-	beqz	a3,.L_bn_sub_words_return
-
-	ld	t1,8(a1)
-	subu	a3,1
-	ld	ta1,8(a2)
-	sltu	t9,t1,ta1
-	dsubu	t1,ta1
-	dsubu	ta1,t1,v0
-	MOVNZ	(t1,v0,t9)
-	sd	ta1,8(a0)
-	beqz	a3,.L_bn_sub_words_return
-
-	ld	t2,16(a1)
-	ld	ta2,16(a2)
-	sltu	t8,t2,ta2
-	dsubu	t2,ta2
-	dsubu	ta2,t2,v0
-	MOVNZ	(t2,v0,t8)
-	sd	ta2,16(a0)
-	jr	ra
-END(bn_sub_words)
-
-#undef	MINUS4
-
-.align 5
-LEAF(bn_div_3_words)
-	.set	reorder
-	move	a3,a0		/* we know that bn_div_words doesn't
-				 * 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
-	ld	a1,-8(a3)
-	bne	a0,a2,.L_bn_div_3_words_proceed
-	li	v0,-1
-	jr	ra
-.L_bn_div_3_words_proceed:
-	move	ta3,ra
-	bal	bn_div_words
-	move	ra,ta3
-	dmultu	ta2,v0
-	ld	t2,-16(a3)
-	move	ta0,zero
-	mfhi	t1
-	mflo	t0
-	sltu	t8,t1,v1
-.L_bn_div_3_words_inner_loop:
-	bnez	t8,.L_bn_div_3_words_inner_loop_done
-	sgeu	AT,t2,t0
-	seq	t9,t1,v1
-	and	AT,t9
-	sltu	t3,t0,ta2
-	daddu	v1,a2
-	dsubu	t1,t3
-	dsubu	t0,ta2
-	sltu	t8,t1,v1
-	sltu	ta0,v1,a2
-	or	t8,ta0
-	.set	noreorder
-	beqzl	AT,.L_bn_div_3_words_inner_loop
-	dsubu	v0,1
-	.set	reorder
-.L_bn_div_3_words_inner_loop_done:
-	jr	ra
-END(bn_div_3_words)
-
-.align	5
-LEAF(bn_div_words)
-	.set	noreorder
-	bnezl	a2,.L_bn_div_words_proceed
-	move	v1,zero
-	jr	ra
-	li	v0,-1		/* I'd rather signal div-by-zero
-				 * which can be done with 'break 7' */
-
-.L_bn_div_words_proceed:
-	bltz	a2,.L_bn_div_words_body
-	move	t9,v1
-	dsll	a2,1
-	bgtz	a2,.-4
-	addu	t9,1
-
-	.set	reorder
-	negu	t1,t9
-	li	t2,-1
-	dsll	t2,t1
-	and	t2,a0
-	dsrl	AT,a1,t1
-	.set	noreorder
-	bnezl	t2,.+8
-	break	6		/* signal overflow */
-	.set	reorder
-	dsll	a0,t9
-	dsll	a1,t9
-	or	a0,AT
-
-#define	QT	ta0
-#define	HH	ta1
-#define	DH	v1
-.L_bn_div_words_body:
-	dsrl	DH,a2,32
-	sgeu	AT,a0,a2
-	.set	noreorder
-	bnezl	AT,.+8
-	dsubu	a0,a2
-	.set	reorder
-
-	li	QT,-1
-	dsrl	HH,a0,32
-	dsrl	QT,32	/* q=0xffffffff */
-	beq	DH,HH,.L_bn_div_words_skip_div1
-	ddivu	zero,a0,DH
-	mflo	QT
-.L_bn_div_words_skip_div1:
-	dmultu	a2,QT
-	dsll	t3,a0,32
-	dsrl	AT,a1,32
-	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
-	dsubu	t1,v0
-	dsubu	t0,a2
-	b	.L_bn_div_words_inner_loop1
-	dsubu	QT,1
-	.set	reorder
-.L_bn_div_words_inner_loop1_done:
-
-	dsll	a1,32
-	dsubu	a0,t3,t0
-	dsll	v0,QT,32
-
-	li	QT,-1
-	dsrl	HH,a0,32
-	dsrl	QT,32	/* q=0xffffffff */
-	beq	DH,HH,.L_bn_div_words_skip_div2
-	ddivu	zero,a0,DH
-	mflo	QT
-.L_bn_div_words_skip_div2:
-#undef	DH
-	dmultu	a2,QT
-	dsll	t3,a0,32
-	dsrl	AT,a1,32
-	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
-	dsubu	t1,v1
-	dsubu	t0,a2
-	b	.L_bn_div_words_inner_loop2
-	dsubu	QT,1
-	.set	reorder
-.L_bn_div_words_inner_loop2_done:	
-#undef	HH
-
-	dsubu	a0,t3,t0
-	or	v0,QT
-	dsrl	v1,a0,t9	/* v1 contains remainder if anybody wants it */
-	dsrl	a2,t9		/* restore a2 */
-	jr	ra
-#undef	QT
-END(bn_div_words)
-
-#define	a_0	t0
-#define	a_1	t1
-#define	a_2	t2
-#define	a_3	t3
-#define	b_0	ta0
-#define	b_1	ta1
-#define	b_2	ta2
-#define	b_3	ta3
-
-#define	a_4	s0
-#define	a_5	s2
-#define	a_6	s4
-#define	a_7	a1	/* once we load a[7] we don't need a anymore */
-#define	b_4	s1
-#define	b_5	s3
-#define	b_6	s5
-#define	b_7	a2	/* once we load b[7] we don't need b anymore */
-
-#define	t_1	t8
-#define	t_2	t9
-
-#define	c_1	v0
-#define	c_2	v1
-#define	c_3	a3
-
-#define	FRAME_SIZE	48
-
-.align	5
-LEAF(bn_mul_comba8)
-	.set	noreorder
-	PTR_SUB	sp,FRAME_SIZE
-	.frame	sp,64,ra
-	.set	reorder
-	ld	a_0,0(a1)	/* If compiled with -mips3 option on
-				 * R5000 box assembler barks on this
-				 * line with "shouldn't 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_1,8(a1)
-	ld	a_2,16(a1)
-	ld	a_3,24(a1)
-	ld	b_1,8(a2)
-	ld	b_2,16(a2)
-	ld	b_3,24(a2)
-	dmultu	a_0,b_0		/* mul_add_c(a[0],b[0],c1,c2,c3); */
-	sd	s0,0(sp)
-	sd	s1,8(sp)
-	sd	s2,16(sp)
-	sd	s3,24(sp)
-	sd	s4,32(sp)
-	sd	s5,40(sp)
-	mflo	c_1
-	mfhi	c_2
-
-	dmultu	a_0,b_1		/* mul_add_c(a[0],b[1],c2,c3,c1); */
-	ld	a_4,32(a1)
-	ld	a_5,40(a1)
-	ld	a_6,48(a1)
-	ld	a_7,56(a1)
-	ld	b_4,32(a2)
-	ld	b_5,40(a2)
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	c_3,t_2,AT
-	dmultu	a_1,b_0		/* mul_add_c(a[1],b[0],c2,c3,c1); */
-	ld	b_6,48(a2)
-	ld	b_7,56(a2)
-	sd	c_1,0(a0)	/* r[0]=c1; */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	c_1,c_3,t_2
-	sd	c_2,8(a0)	/* r[1]=c2; */
-
-	dmultu	a_2,b_0		/* mul_add_c(a[2],b[0],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	dmultu	a_1,b_1		/* mul_add_c(a[1],b[1],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	c_2,c_1,t_2
-	dmultu	a_0,b_2		/* mul_add_c(a[0],b[2],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,16(a0)	/* r[2]=c3; */
-
-	dmultu	a_0,b_3		/* mul_add_c(a[0],b[3],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	c_3,c_2,t_2
-	dmultu	a_1,b_2		/* mul_add_c(a[1],b[2],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_2,b_1		/* mul_add_c(a[2],b[1],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_3,b_0		/* mul_add_c(a[3],b[0],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,24(a0)	/* r[3]=c1; */
-
-	dmultu	a_4,b_0		/* mul_add_c(a[4],b[0],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	c_1,c_3,t_2
-	dmultu	a_3,b_1		/* mul_add_c(a[3],b[1],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_2,b_2		/* mul_add_c(a[2],b[2],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_1,b_3		/* mul_add_c(a[1],b[3],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_0,b_4		/* mul_add_c(a[0],b[4],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,32(a0)	/* r[4]=c2; */
-
-	dmultu	a_0,b_5		/* mul_add_c(a[0],b[5],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	c_2,c_1,t_2
-	dmultu	a_1,b_4		/* mul_add_c(a[1],b[4],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_2,b_3		/* mul_add_c(a[2],b[3],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_3,b_2		/* mul_add_c(a[3],b[2],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_4,b_1		/* mul_add_c(a[4],b[1],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_5,b_0		/* mul_add_c(a[5],b[0],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,40(a0)	/* r[5]=c3; */
-
-	dmultu	a_6,b_0		/* mul_add_c(a[6],b[0],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	c_3,c_2,t_2
-	dmultu	a_5,b_1		/* mul_add_c(a[5],b[1],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_4,b_2		/* mul_add_c(a[4],b[2],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_3,b_3		/* mul_add_c(a[3],b[3],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_2,b_4		/* mul_add_c(a[2],b[4],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_1,b_5		/* mul_add_c(a[1],b[5],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_0,b_6		/* mul_add_c(a[0],b[6],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,48(a0)	/* r[6]=c1; */
-
-	dmultu	a_0,b_7		/* mul_add_c(a[0],b[7],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	c_1,c_3,t_2
-	dmultu	a_1,b_6		/* mul_add_c(a[1],b[6],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_2,b_5		/* mul_add_c(a[2],b[5],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_3,b_4		/* mul_add_c(a[3],b[4],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_4,b_3		/* mul_add_c(a[4],b[3],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_5,b_2		/* mul_add_c(a[5],b[2],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_6,b_1		/* mul_add_c(a[6],b[1],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_7,b_0		/* mul_add_c(a[7],b[0],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,56(a0)	/* r[7]=c2; */
-
-	dmultu	a_7,b_1		/* mul_add_c(a[7],b[1],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	c_2,c_1,t_2
-	dmultu	a_6,b_2		/* mul_add_c(a[6],b[2],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_5,b_3		/* mul_add_c(a[5],b[3],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_4,b_4		/* mul_add_c(a[4],b[4],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_3,b_5		/* mul_add_c(a[3],b[5],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_2,b_6		/* mul_add_c(a[2],b[6],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_1,b_7		/* mul_add_c(a[1],b[7],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,64(a0)	/* r[8]=c3; */
-
-	dmultu	a_2,b_7		/* mul_add_c(a[2],b[7],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	c_3,c_2,t_2
-	dmultu	a_3,b_6		/* mul_add_c(a[3],b[6],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_4,b_5		/* mul_add_c(a[4],b[5],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_5,b_4		/* mul_add_c(a[5],b[4],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_6,b_3		/* mul_add_c(a[6],b[3],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_7,b_2		/* mul_add_c(a[7],b[2],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,72(a0)	/* r[9]=c1; */
-
-	dmultu	a_7,b_3		/* mul_add_c(a[7],b[3],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	c_1,c_3,t_2
-	dmultu	a_6,b_4		/* mul_add_c(a[6],b[4],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_5,b_5		/* mul_add_c(a[5],b[5],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_4,b_6		/* mul_add_c(a[4],b[6],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_3,b_7		/* mul_add_c(a[3],b[7],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,80(a0)	/* r[10]=c2; */
-
-	dmultu	a_4,b_7		/* mul_add_c(a[4],b[7],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	c_2,c_1,t_2
-	dmultu	a_5,b_6		/* mul_add_c(a[5],b[6],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_6,b_5		/* mul_add_c(a[6],b[5],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_7,b_4		/* mul_add_c(a[7],b[4],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,88(a0)	/* r[11]=c3; */
-
-	dmultu	a_7,b_5		/* mul_add_c(a[7],b[5],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	c_3,c_2,t_2
-	dmultu	a_6,b_6		/* mul_add_c(a[6],b[6],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_5,b_7		/* mul_add_c(a[5],b[7],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,96(a0)	/* r[12]=c1; */
-
-	dmultu	a_6,b_7		/* mul_add_c(a[6],b[7],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	c_1,c_3,t_2
-	dmultu	a_7,b_6		/* mul_add_c(a[7],b[6],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,104(a0)	/* r[13]=c2; */
-
-	dmultu	a_7,b_7		/* mul_add_c(a[7],b[7],c3,c1,c2); */
-	ld	s0,0(sp)
-	ld	s1,8(sp)
-	ld	s2,16(sp)
-	ld	s3,24(sp)
-	ld	s4,32(sp)
-	ld	s5,40(sp)
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sd	c_3,112(a0)	/* r[14]=c3; */
-	sd	c_1,120(a0)	/* r[15]=c1; */
-
-	PTR_ADD	sp,FRAME_SIZE
-
-	jr	ra
-END(bn_mul_comba8)
-
-.align	5
-LEAF(bn_mul_comba4)
-	.set	reorder
-	ld	a_0,0(a1)
-	ld	b_0,0(a2)
-	ld	a_1,8(a1)
-	ld	a_2,16(a1)
-	dmultu	a_0,b_0		/* mul_add_c(a[0],b[0],c1,c2,c3); */
-	ld	a_3,24(a1)
-	ld	b_1,8(a2)
-	ld	b_2,16(a2)
-	ld	b_3,24(a2)
-	mflo	c_1
-	mfhi	c_2
-	sd	c_1,0(a0)
-
-	dmultu	a_0,b_1		/* mul_add_c(a[0],b[1],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	c_3,t_2,AT
-	dmultu	a_1,b_0		/* mul_add_c(a[1],b[0],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	c_1,c_3,t_2
-	sd	c_2,8(a0)
-
-	dmultu	a_2,b_0		/* mul_add_c(a[2],b[0],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	dmultu	a_1,b_1		/* mul_add_c(a[1],b[1],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	c_2,c_1,t_2
-	dmultu	a_0,b_2		/* mul_add_c(a[0],b[2],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,16(a0)
-
-	dmultu	a_0,b_3		/* mul_add_c(a[0],b[3],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	c_3,c_2,t_2
-	dmultu	a_1,b_2		/* mul_add_c(a[1],b[2],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_2,b_1		/* mul_add_c(a[2],b[1],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_3,b_0		/* mul_add_c(a[3],b[0],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,24(a0)
-
-	dmultu	a_3,b_1		/* mul_add_c(a[3],b[1],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	c_1,c_3,t_2
-	dmultu	a_2,b_2		/* mul_add_c(a[2],b[2],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_1,b_3		/* mul_add_c(a[1],b[3],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,32(a0)
-
-	dmultu	a_2,b_3		/* mul_add_c(a[2],b[3],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	c_2,c_1,t_2
-	dmultu	a_3,b_2		/* mul_add_c(a[3],b[2],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,40(a0)
-
-	dmultu	a_3,b_3		/* mul_add_c(a[3],b[3],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sd	c_1,48(a0)
-	sd	c_2,56(a0)
-
-	jr	ra
-END(bn_mul_comba4)
-
-#undef	a_4
-#undef	a_5
-#undef	a_6
-#undef	a_7
-#define	a_4	b_0
-#define	a_5	b_1
-#define	a_6	b_2
-#define	a_7	b_3
-
-.align	5
-LEAF(bn_sqr_comba8)
-	.set	reorder
-	ld	a_0,0(a1)
-	ld	a_1,8(a1)
-	ld	a_2,16(a1)
-	ld	a_3,24(a1)
-
-	dmultu	a_0,a_0		/* mul_add_c(a[0],b[0],c1,c2,c3); */
-	ld	a_4,32(a1)
-	ld	a_5,40(a1)
-	ld	a_6,48(a1)
-	ld	a_7,56(a1)
-	mflo	c_1
-	mfhi	c_2
-	sd	c_1,0(a0)
-
-	dmultu	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
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	c_3,t_2,AT
-	sd	c_2,8(a0)
-
-	dmultu	a_2,a_0		/* mul_add_c2(a[2],b[0],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_2,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_1,a_1		/* mul_add_c(a[1],b[1],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,16(a0)
-
-	dmultu	a_0,a_3		/* mul_add_c2(a[0],b[3],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_3,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_1,a_2		/* mul_add_c2(a[1],b[2],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_3,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,24(a0)
-
-	dmultu	a_4,a_0		/* mul_add_c2(a[4],b[0],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_1,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_3,a_1		/* mul_add_c2(a[3],b[1],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_1,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_2,a_2		/* mul_add_c(a[2],b[2],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,32(a0)
-
-	dmultu	a_0,a_5		/* mul_add_c2(a[0],b[5],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_2,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_1,a_4		/* mul_add_c2(a[1],b[4],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_2,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_2,a_3		/* mul_add_c2(a[2],b[3],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_2,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,40(a0)
-
-	dmultu	a_6,a_0		/* mul_add_c2(a[6],b[0],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_3,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_5,a_1		/* mul_add_c2(a[5],b[1],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_3,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_4,a_2		/* mul_add_c2(a[4],b[2],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_3,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_3,a_3		/* mul_add_c(a[3],b[3],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,48(a0)
-
-	dmultu	a_0,a_7		/* mul_add_c2(a[0],b[7],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_1,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_1,a_6		/* mul_add_c2(a[1],b[6],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_1,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_2,a_5		/* mul_add_c2(a[2],b[5],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_1,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_3,a_4		/* mul_add_c2(a[3],b[4],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_1,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,56(a0)
-
-	dmultu	a_7,a_1		/* mul_add_c2(a[7],b[1],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_2,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_6,a_2		/* mul_add_c2(a[6],b[2],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_2,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_5,a_3		/* mul_add_c2(a[5],b[3],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_2,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_4,a_4		/* mul_add_c(a[4],b[4],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,64(a0)
-
-	dmultu	a_2,a_7		/* mul_add_c2(a[2],b[7],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_3,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_3,a_6		/* mul_add_c2(a[3],b[6],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_3,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_4,a_5		/* mul_add_c2(a[4],b[5],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_3,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,72(a0)
-
-	dmultu	a_7,a_3		/* mul_add_c2(a[7],b[3],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_1,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_6,a_4		/* mul_add_c2(a[6],b[4],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_1,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_5,a_5		/* mul_add_c(a[5],b[5],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,80(a0)
-
-	dmultu	a_4,a_7		/* mul_add_c2(a[4],b[7],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_2,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_5,a_6		/* mul_add_c2(a[5],b[6],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_2,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,88(a0)
-
-	dmultu	a_7,a_5		/* mul_add_c2(a[7],b[5],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_3,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_6,a_6		/* mul_add_c(a[6],b[6],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,96(a0)
-
-	dmultu	a_6,a_7		/* mul_add_c2(a[6],b[7],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_1,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,104(a0)
-
-	dmultu	a_7,a_7		/* mul_add_c(a[7],b[7],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sd	c_3,112(a0)
-	sd	c_1,120(a0)
-
-	jr	ra
-END(bn_sqr_comba8)
-
-.align	5
-LEAF(bn_sqr_comba4)
-	.set	reorder
-	ld	a_0,0(a1)
-	ld	a_1,8(a1)
-	ld	a_2,16(a1)
-	ld	a_3,24(a1)
-	dmultu	a_0,a_0		/* mul_add_c(a[0],b[0],c1,c2,c3); */
-	mflo	c_1
-	mfhi	c_2
-	sd	c_1,0(a0)
-
-	dmultu	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
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	c_3,t_2,AT
-	sd	c_2,8(a0)
-
-	dmultu	a_2,a_0		/* mul_add_c2(a[2],b[0],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_2,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	dmultu	a_1,a_1		/* mul_add_c(a[1],b[1],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,16(a0)
-
-	dmultu	a_0,a_3		/* mul_add_c2(a[0],b[3],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_3,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	dmultu	a_1,a_2		/* mul_add_c(a2[1],b[2],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	slt	AT,t_2,zero
-	daddu	c_3,AT
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sltu	AT,c_2,t_2
-	daddu	c_3,AT
-	sd	c_1,24(a0)
-
-	dmultu	a_3,a_1		/* mul_add_c2(a[3],b[1],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_1,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	dmultu	a_2,a_2		/* mul_add_c(a[2],b[2],c2,c3,c1); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_2,t_1
-	sltu	AT,c_2,t_1
-	daddu	t_2,AT
-	daddu	c_3,t_2
-	sltu	AT,c_3,t_2
-	daddu	c_1,AT
-	sd	c_2,32(a0)
-
-	dmultu	a_2,a_3		/* mul_add_c2(a[2],b[3],c3,c1,c2); */
-	mflo	t_1
-	mfhi	t_2
-	slt	c_2,t_2,zero
-	dsll	t_2,1
-	slt	a2,t_1,zero
-	daddu	t_2,a2
-	dsll	t_1,1
-	daddu	c_3,t_1
-	sltu	AT,c_3,t_1
-	daddu	t_2,AT
-	daddu	c_1,t_2
-	sltu	AT,c_1,t_2
-	daddu	c_2,AT
-	sd	c_3,40(a0)
-
-	dmultu	a_3,a_3		/* mul_add_c(a[3],b[3],c1,c2,c3); */
-	mflo	t_1
-	mfhi	t_2
-	daddu	c_1,t_1
-	sltu	AT,c_1,t_1
-	daddu	t_2,AT
-	daddu	c_2,t_2
-	sd	c_1,48(a0)
-	sd	c_2,56(a0)
-
-	jr	ra
-END(bn_sqr_comba4)
diff --git a/crypto/bn/asm/x86_64-gcc.c b/crypto/bn/asm/x86_64-gcc.c
index 31476abebfb7..2d39407f499e 100644
--- a/crypto/bn/asm/x86_64-gcc.c
+++ b/crypto/bn/asm/x86_64-gcc.c
@@ -273,6 +273,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
 /* sqr_add_c(a,i,c0,c1,c2)  -- c+=a[i]^2 for three word number c=(c2,c1,c0) */
 /* sqr_add_c2(a,i,c0,c1,c2) -- c+=2*a[i]*a[j] for three word number c=(c2,c1,c0) */
 
+/*
+ * Keep in mind that carrying into high part of multiplication result
+ * can not overflow, because it cannot be all-ones.
+ */
 #if 0
 /* original macros are kept for reference purposes */
 #define mul_add_c(a,b,c0,c1,c2) {	\
@@ -287,10 +291,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
 	BN_ULONG ta=(a),tb=(b),t0;	\
 	t1 = BN_UMULT_HIGH(ta,tb);	\
 	t0 = ta * tb;			\
-	t2 = t1+t1; c2 += (t2<t1)?1:0;	\
-	t1 = t0+t0; t2 += (t1<t0)?1:0;	\
-	c0 += t1; t2 += (c0<t1)?1:0;	\
+	c0 += t0; t2 = t1+((c0<t0)?1:0);\
 	c1 += t2; c2 += (c1<t2)?1:0;	\
+	c0 += t0; t1 += (c0<t0)?1:0;	\
+	c1 += t1; c2 += (c1<t1)?1:0;	\
 	}
 #else
 #define mul_add_c(a,b,c0,c1,c2)	do {	\
@@ -328,22 +332,14 @@ BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
 		: "=a"(t1),"=d"(t2)	\
 		: "a"(a),"m"(b)		\
 		: "cc");		\
-	asm ("addq %0,%0; adcq %2,%1"	\
-		: "+d"(t2),"+r"(c2)	\
-		: "g"(0)		\
-		: "cc");		\
-	asm ("addq %0,%0; adcq %2,%1"	\
-		: "+a"(t1),"+d"(t2)	\
-		: "g"(0)		\
-		: "cc");		\
-	asm ("addq %2,%0; adcq %3,%1"	\
-		: "+r"(c0),"+d"(t2)	\
-		: "a"(t1),"g"(0)	\
-		: "cc");		\
-	asm ("addq %2,%0; adcq %3,%1"	\
-		: "+r"(c1),"+r"(c2)	\
-		: "d"(t2),"g"(0)	\
-		: "cc");		\
+	asm ("addq %3,%0; adcq %4,%1; adcq %5,%2"	\
+		: "+r"(c0),"+r"(c1),"+r"(c2)		\
+		: "r"(t1),"r"(t2),"g"(0)		\
+		: "cc");				\
+	asm ("addq %3,%0; adcq %4,%1; adcq %5,%2"	\
+		: "+r"(c0),"+r"(c1),"+r"(c2)		\
+		: "r"(t1),"r"(t2),"g"(0)		\
+		: "cc");				\
 	} while (0)
 #endif
 
diff --git a/crypto/bn/bn.h b/crypto/bn/bn.h
index 21a1a3fe35c0..c4d618522e26 100644
--- a/crypto/bn/bn.h
+++ b/crypto/bn/bn.h
@@ -780,7 +780,9 @@ int RAND_pseudo_bytes(unsigned char *buf,int num);
 #define bn_wcheck_size(bn, words) \
 	do { \
 		const BIGNUM *_bnum2 = (bn); \
-		assert(words <= (_bnum2)->dmax && words >= (_bnum2)->top); \
+		assert((words) <= (_bnum2)->dmax && (words) >= (_bnum2)->top); \
+		/* avoid unused variable warning with NDEBUG */ \
+		(void)(_bnum2); \
 	} while(0)
 
 #else /* !BN_DEBUG */
diff --git a/crypto/bn/bn_asm.c b/crypto/bn/bn_asm.c
index c43c91cc09f4..a33b63411b7f 100644
--- a/crypto/bn/bn_asm.c
+++ b/crypto/bn/bn_asm.c
@@ -438,6 +438,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b, int n)
 /* sqr_add_c(a,i,c0,c1,c2)  -- c+=a[i]^2 for three word number c=(c2,c1,c0) */
 /* sqr_add_c2(a,i,c0,c1,c2) -- c+=2*a[i]*a[j] for three word number c=(c2,c1,c0) */
 
+/*
+ * Keep in mind that carrying into high part of multiplication result
+ * can not overflow, because it cannot be all-ones.
+ */
 #ifdef BN_LLONG
 #define mul_add_c(a,b,c0,c1,c2) \
 	t=(BN_ULLONG)a*b; \
@@ -478,10 +482,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b, int n)
 #define mul_add_c2(a,b,c0,c1,c2) {	\
 	BN_ULONG ta=(a),tb=(b),t0;	\
 	BN_UMULT_LOHI(t0,t1,ta,tb);	\
-	t2 = t1+t1; c2 += (t2<t1)?1:0;	\
-	t1 = t0+t0; t2 += (t1<t0)?1:0;	\
-	c0 += t1; t2 += (c0<t1)?1:0;	\
+	c0 += t0; t2 = t1+((c0<t0)?1:0);\
 	c1 += t2; c2 += (c1<t2)?1:0;	\
+	c0 += t0; t1 += (c0<t0)?1:0;	\
+	c1 += t1; c2 += (c1<t1)?1:0;	\
 	}
 
 #define sqr_add_c(a,i,c0,c1,c2)	{	\
@@ -508,10 +512,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b, int n)
 	BN_ULONG ta=(a),tb=(b),t0;	\
 	t1 = BN_UMULT_HIGH(ta,tb);	\
 	t0 = ta * tb;			\
-	t2 = t1+t1; c2 += (t2<t1)?1:0;	\
-	t1 = t0+t0; t2 += (t1<t0)?1:0;	\
-	c0 += t1; t2 += (c0<t1)?1:0;	\
+	c0 += t0; t2 = t1+((c0<t0)?1:0);\
 	c1 += t2; c2 += (c1<t2)?1:0;	\
+	c0 += t0; t1 += (c0<t0)?1:0;	\
+	c1 += t1; c2 += (c1<t1)?1:0;	\
 	}
 
 #define sqr_add_c(a,i,c0,c1,c2)	{	\
diff --git a/crypto/bn/bn_ctx.c b/crypto/bn/bn_ctx.c
index 3f2256f67575..90aa3aeb9f75 100644
--- a/crypto/bn/bn_ctx.c
+++ b/crypto/bn/bn_ctx.c
@@ -158,7 +158,7 @@ static void ctxdbg(BN_CTX *ctx)
 	unsigned int bnidx = 0, fpidx = 0;
 	BN_POOL_ITEM *item = ctx->pool.head;
 	BN_STACK *stack = &ctx->stack;
-	fprintf(stderr,"(%08x): ", (unsigned int)ctx);
+	fprintf(stderr,"(%16p): ", ctx);
 	while(bnidx < ctx->used)
 		{
 		fprintf(stderr,"%03x ", item->vals[bnidx++ % BN_CTX_POOL_SIZE].dmax);
diff --git a/crypto/bn/bn_div.c b/crypto/bn/bn_div.c
index 7b2403185e62..0ec90e805c47 100644
--- a/crypto/bn/bn_div.c
+++ b/crypto/bn/bn_div.c
@@ -189,15 +189,17 @@ int BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor,
 	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
+	 * so don't just rely on bn_check_top() here
 	 * (bn_check_top() works only for BN_DEBUG builds) */
-	if (num->top > 0 && num->d[num->top - 1] == 0)
+	if ((num->top > 0 && num->d[num->top - 1] == 0) ||
+		(divisor->top > 0 && divisor->d[divisor->top - 1] == 0))
 		{
 		BNerr(BN_F_BN_DIV,BN_R_NOT_INITIALIZED);
 		return 0;
 		}
 
 	bn_check_top(num);
+	bn_check_top(divisor);
 
 	if ((BN_get_flags(num, BN_FLG_CONSTTIME) != 0) || (BN_get_flags(divisor, BN_FLG_CONSTTIME) != 0))
 		{
@@ -207,7 +209,7 @@ int BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor,
 	bn_check_top(dv);
 	bn_check_top(rm);
 	/* bn_check_top(num); */ /* 'num' has been checked already */
-	bn_check_top(divisor);
+	/* bn_check_top(divisor); */ /* 'divisor' has been checked already */
 
 	if (BN_is_zero(divisor))
 		{
diff --git a/crypto/bn/bntest.c b/crypto/bn/bntest.c
index 06f5954acc39..48bc63384c92 100644
--- a/crypto/bn/bntest.c
+++ b/crypto/bn/bntest.c
@@ -107,6 +107,7 @@ int test_mod(BIO *bp,BN_CTX *ctx);
 int test_mod_mul(BIO *bp,BN_CTX *ctx);
 int test_mod_exp(BIO *bp,BN_CTX *ctx);
 int test_mod_exp_mont_consttime(BIO *bp,BN_CTX *ctx);
+int test_mod_exp_mont5(BIO *bp, BN_CTX *ctx);
 int test_exp(BIO *bp,BN_CTX *ctx);
 int test_gf2m_add(BIO *bp);
 int test_gf2m_mod(BIO *bp);
@@ -249,6 +250,7 @@ int main(int argc, char *argv[])
 
 	message(out,"BN_mod_exp_mont_consttime");
 	if (!test_mod_exp_mont_consttime(out,ctx)) goto err;
+	if (!test_mod_exp_mont5(out,ctx)) goto err;
 	(void)BIO_flush(out);
 
 	message(out,"BN_exp");
@@ -676,44 +678,98 @@ int test_mul(BIO *bp)
 
 int test_sqr(BIO *bp, BN_CTX *ctx)
 	{
-	BIGNUM a,c,d,e;
-	int i;
+	BIGNUM *a,*c,*d,*e;
+	int i, ret = 0;
 
-	BN_init(&a);
-	BN_init(&c);
-	BN_init(&d);
-	BN_init(&e);
+	a = BN_new();
+	c = BN_new();
+	d = BN_new();
+	e = BN_new();
+	if (a == NULL || c == NULL || d == NULL || e == NULL)
+		{
+		goto err;
+		}
 
 	for (i=0; i<num0; i++)
 		{
-		BN_bntest_rand(&a,40+i*10,0,0);
-		a.neg=rand_neg();
-		BN_sqr(&c,&a,ctx);
+		BN_bntest_rand(a,40+i*10,0,0);
+		a->neg=rand_neg();
+		BN_sqr(c,a,ctx);
 		if (bp != NULL)
 			{
 			if (!results)
 				{
-				BN_print(bp,&a);
+				BN_print(bp,a);
 				BIO_puts(bp," * ");
-				BN_print(bp,&a);
+				BN_print(bp,a);
 				BIO_puts(bp," - ");
 				}
-			BN_print(bp,&c);
+			BN_print(bp,c);
 			BIO_puts(bp,"\n");
 			}
-		BN_div(&d,&e,&c,&a,ctx);
-		BN_sub(&d,&d,&a);
-		if(!BN_is_zero(&d) || !BN_is_zero(&e))
-		    {
-		    fprintf(stderr,"Square test failed!\n");
-		    return 0;
-		    }
+		BN_div(d,e,c,a,ctx);
+		BN_sub(d,d,a);
+		if(!BN_is_zero(d) || !BN_is_zero(e))
+			{
+			fprintf(stderr,"Square test failed!\n");
+			goto err;
+			}
 		}
-	BN_free(&a);
-	BN_free(&c);
-	BN_free(&d);
-	BN_free(&e);
-	return(1);
+
+	/* Regression test for a BN_sqr overflow bug. */
+	BN_hex2bn(&a,
+		"80000000000000008000000000000001FFFFFFFFFFFFFFFE0000000000000000");
+	BN_sqr(c, a, ctx);
+	if (bp != NULL)
+		{
+		if (!results)
+			{
+			BN_print(bp,a);
+			BIO_puts(bp," * ");
+			BN_print(bp,a);
+			BIO_puts(bp," - ");
+			}
+		BN_print(bp,c);
+		BIO_puts(bp,"\n");
+		}
+	BN_mul(d, a, a, ctx);
+	if (BN_cmp(c, d))
+		{
+		fprintf(stderr, "Square test failed: BN_sqr and BN_mul produce "
+			"different results!\n");
+		goto err;
+		}
+
+	/* Regression test for a BN_sqr overflow bug. */
+	BN_hex2bn(&a,
+		"80000000000000000000000080000001FFFFFFFE000000000000000000000000");
+	BN_sqr(c, a, ctx);
+	if (bp != NULL)
+		{
+		if (!results)
+			{
+			BN_print(bp,a);
+			BIO_puts(bp," * ");
+			BN_print(bp,a);
+			BIO_puts(bp," - ");
+			}
+		BN_print(bp,c);
+		BIO_puts(bp,"\n");
+		}
+	BN_mul(d, a, a, ctx);
+	if (BN_cmp(c, d))
+		{
+		fprintf(stderr, "Square test failed: BN_sqr and BN_mul produce "
+			"different results!\n");
+		goto err;
+		}
+	ret = 1;
+err:
+	if (a != NULL) BN_free(a);
+	if (c != NULL) BN_free(c);
+	if (d != NULL) BN_free(d);
+	if (e != NULL) BN_free(e);
+	return ret;
 	}
 
 int test_mont(BIO *bp, BN_CTX *ctx)
@@ -1012,6 +1068,80 @@ int test_mod_exp_mont_consttime(BIO *bp, BN_CTX *ctx)
 	return(1);
 	}
 
+/* Test constant-time modular exponentiation with 1024-bit inputs,
+ * which on x86_64 cause a different code branch to be taken.
+ */
+int test_mod_exp_mont5(BIO *bp, BN_CTX *ctx)
+	{
+	BIGNUM *a,*p,*m,*d,*e;
+
+	BN_MONT_CTX *mont;
+
+	a=BN_new();
+	p=BN_new();
+	m=BN_new();
+	d=BN_new();
+	e=BN_new();
+
+	mont = BN_MONT_CTX_new();
+
+	BN_bntest_rand(m,1024,0,1); /* must be odd for montgomery */
+	/* Zero exponent */
+	BN_bntest_rand(a,1024,0,0);
+	BN_zero(p);
+	if(!BN_mod_exp_mont_consttime(d,a,p,m,ctx,NULL))
+		return 0;
+	if(!BN_is_one(d))
+		{
+		fprintf(stderr, "Modular exponentiation test failed!\n");
+		return 0;
+		}
+	/* Zero input */
+	BN_bntest_rand(p,1024,0,0);
+	BN_zero(a);
+	if(!BN_mod_exp_mont_consttime(d,a,p,m,ctx,NULL))
+		return 0;
+	if(!BN_is_zero(d))
+		{
+		fprintf(stderr, "Modular exponentiation test failed!\n");
+		return 0;
+		}
+	/* Craft an input whose Montgomery representation is 1,
+	 * i.e., shorter than the modulus m, in order to test
+	 * the const time precomputation scattering/gathering.
+	 */
+	BN_one(a);
+	BN_MONT_CTX_set(mont,m,ctx);
+	if(!BN_from_montgomery(e,a,mont,ctx))
+		return 0;
+	if(!BN_mod_exp_mont_consttime(d,e,p,m,ctx,NULL))
+		return 0;
+	if(!BN_mod_exp_simple(a,e,p,m,ctx))
+		return 0;
+	if(BN_cmp(a,d) != 0)
+		{
+		fprintf(stderr,"Modular exponentiation test failed!\n");
+		return 0;
+		}
+	/* Finally, some regular test vectors. */
+	BN_bntest_rand(e,1024,0,0);
+	if(!BN_mod_exp_mont_consttime(d,e,p,m,ctx,NULL))
+		return 0;
+	if(!BN_mod_exp_simple(a,e,p,m,ctx))
+		return 0;
+	if(BN_cmp(a,d) != 0)
+		{
+		fprintf(stderr,"Modular exponentiation test failed!\n");
+		return 0;
+		}
+	BN_free(a);
+	BN_free(p);
+	BN_free(m);
+	BN_free(d);
+	BN_free(e);
+	return(1);
+	}
+
 int test_exp(BIO *bp, BN_CTX *ctx)
 	{
 	BIGNUM *a,*b,*d,*e,*one;
diff --git a/crypto/constant_time_locl.h b/crypto/constant_time_locl.h
index c0483939fe3d..8af98c168329 100644
--- a/crypto/constant_time_locl.h
+++ b/crypto/constant_time_locl.h
@@ -129,17 +129,12 @@ static inline int constant_time_select_int(unsigned int mask, int a, int b);
 
 static inline unsigned int constant_time_msb(unsigned int a)
 	{
-	return (unsigned int)((int)(a) >> (sizeof(int) * 8 - 1));
+	return 0-(a >> (sizeof(a) * 8 - 1));
 	}
 
 static inline unsigned int constant_time_lt(unsigned int a, unsigned int b)
 	{
-	unsigned int lt;
-	/* Case 1: msb(a) == msb(b). a < b iff the MSB of a - b is set.*/
-	lt = ~(a ^ b) & (a - b);
-	/* Case 2: msb(a) != msb(b). a < b iff the MSB of b is set. */
-	lt |= ~a & b;
-	return constant_time_msb(lt);
+	return constant_time_msb(a^((a^b)|((a-b)^b)));
 	}
 
 static inline unsigned char constant_time_lt_8(unsigned int a, unsigned int b)
@@ -149,12 +144,7 @@ static inline unsigned char constant_time_lt_8(unsigned int a, unsigned int b)
 
 static inline unsigned int constant_time_ge(unsigned int a, unsigned int b)
 	{
-	unsigned int ge;
-	/* Case 1: msb(a) == msb(b). a >= b iff the MSB of a - b is not set.*/
-	ge = ~((a ^ b) | (a - b));
-	/* Case 2: msb(a) != msb(b). a >= b iff the MSB of a is set. */
-	ge |= a & ~b;
-	return constant_time_msb(ge);
+	return ~constant_time_lt(a, b);
 	}
 
 static inline unsigned char constant_time_ge_8(unsigned int a, unsigned int b)
@@ -204,7 +194,7 @@ static inline unsigned char constant_time_select_8(unsigned char mask,
 	return (unsigned char)(constant_time_select(mask, a, b));
 	}
 
-inline int constant_time_select_int(unsigned int mask, int a, int b)
+static inline int constant_time_select_int(unsigned int mask, int a, int b)
 	{
 	return (int)(constant_time_select(mask, (unsigned)(a), (unsigned)(b)));
 	}
diff --git a/crypto/cversion.c b/crypto/cversion.c
index ea9f25fd1666..0336adaa5784 100644
--- a/crypto/cversion.c
+++ b/crypto/cversion.c
@@ -69,10 +69,7 @@ const char *SSLeay_version(int t)
 	if (t == SSLEAY_BUILT_ON)
 		{
 #ifdef DATE
-		static char buf[sizeof(DATE)+11];
-
-		BIO_snprintf(buf,sizeof buf,"built on: %s",DATE);
-		return(buf);
+		return(DATE);
 #else
 		return("built on: date not available");
 #endif
@@ -80,10 +77,7 @@ const char *SSLeay_version(int t)
 	if (t == SSLEAY_CFLAGS)
 		{
 #ifdef CFLAGS
-		static char buf[sizeof(CFLAGS)+11];
-
-		BIO_snprintf(buf,sizeof buf,"compiler: %s",CFLAGS);
-		return(buf);
+		return(cflags);
 #else
 		return("compiler: information not available");
 #endif
@@ -91,10 +85,7 @@ const char *SSLeay_version(int t)
 	if (t == SSLEAY_PLATFORM)
 		{
 #ifdef PLATFORM
-		static char buf[sizeof(PLATFORM)+11];
-
-		BIO_snprintf(buf,sizeof buf,"platform: %s", PLATFORM);
-		return(buf);
+		return(PLATFORM);
 #else
 		return("platform: information not available");
 #endif
diff --git a/crypto/dsa/dsa_asn1.c b/crypto/dsa/dsa_asn1.c
index 60585343746d..473af873e02b 100644
--- a/crypto/dsa/dsa_asn1.c
+++ b/crypto/dsa/dsa_asn1.c
@@ -176,13 +176,25 @@ int DSA_verify(int type, const unsigned char *dgst, int dgst_len,
 	     const unsigned char *sigbuf, int siglen, DSA *dsa)
 	{
 	DSA_SIG *s;
+	const unsigned char *p = sigbuf;
+	unsigned char *der = NULL;
+	int derlen = -1;
 	int ret=-1;
 
 	s = DSA_SIG_new();
 	if (s == NULL) return(ret);
-	if (d2i_DSA_SIG(&s,&sigbuf,siglen) == NULL) goto err;
+	if (d2i_DSA_SIG(&s,&p,siglen) == NULL) goto err;
+	/* Ensure signature uses DER and doesn't have trailing garbage */
+	derlen = i2d_DSA_SIG(s, &der);
+	if (derlen != siglen || memcmp(sigbuf, der, derlen))
+		goto err;
 	ret=DSA_do_verify(dgst,dgst_len,s,dsa);
 err:
+	if (derlen > 0)
+		{
+		OPENSSL_cleanse(der, derlen);
+		OPENSSL_free(der);
+		}
 	DSA_SIG_free(s);
 	return(ret);
 	}
diff --git a/crypto/dso/dso_dlfcn.c b/crypto/dso/dso_dlfcn.c
index 4a56aace0e36..faa9d76b4270 100644
--- a/crypto/dso/dso_dlfcn.c
+++ b/crypto/dso/dso_dlfcn.c
@@ -60,10 +60,8 @@
    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
+#ifndef _GNU_SOURCE
+# define _GNU_SOURCE	/* make sure dladdr is declared */
 #endif
 
 #include <stdio.h>
diff --git a/crypto/ec/ec_lib.c b/crypto/ec/ec_lib.c
index e2c4741b5b77..0992c393b0f4 100644
--- a/crypto/ec/ec_lib.c
+++ b/crypto/ec/ec_lib.c
@@ -68,7 +68,7 @@
 
 #include "ec_lcl.h"
 
-static const char EC_version[] = "EC" OPENSSL_VERSION_PTEXT;
+const char EC_version[] = "EC" OPENSSL_VERSION_PTEXT;
 
 
 /* functions for EC_GROUP objects */
diff --git a/crypto/ec/ec_mult.c b/crypto/ec/ec_mult.c
index 19f21675fbd8..e81200b25535 100644
--- a/crypto/ec/ec_mult.c
+++ b/crypto/ec/ec_mult.c
@@ -445,15 +445,16 @@ int ec_wNAF_mul(const EC_GROUP *group, EC_POINT *r, const BIGNUM *scalar,
 	wNAF_len = OPENSSL_malloc(totalnum * sizeof wNAF_len[0]);
 	wNAF     = OPENSSL_malloc((totalnum + 1) * sizeof wNAF[0]); /* includes space for pivot */
 	val_sub  = OPENSSL_malloc(totalnum * sizeof val_sub[0]);
-		 
+
+	/* Ensure wNAF is initialised in case we end up going to err */
+	if (wNAF) wNAF[0] = NULL;	/* preliminary pivot */
+
 	if (!wsize || !wNAF_len || !wNAF || !val_sub)
 		{
 		ECerr(EC_F_EC_WNAF_MUL, ERR_R_MALLOC_FAILURE);
 		goto err;
 		}
 
-	wNAF[0] = NULL;	/* preliminary pivot */
-
 	/* num_val will be the total number of temporarily precomputed points */
 	num_val = 0;
 
diff --git a/crypto/ec/ec_pmeth.c b/crypto/ec/ec_pmeth.c
index 66ee397d86c4..b62b532cf869 100644
--- a/crypto/ec/ec_pmeth.c
+++ b/crypto/ec/ec_pmeth.c
@@ -167,6 +167,7 @@ static int pkey_ec_verify(EVP_PKEY_CTX *ctx,
 	return ret;
 	}
 
+#ifndef OPENSSL_NO_ECDH
 static int pkey_ec_derive(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen)
 	{
 	int ret;
@@ -200,6 +201,7 @@ static int pkey_ec_derive(EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen)
 	*keylen = ret;
 	return 1;
 	}
+#endif
 
 static int pkey_ec_ctrl(EVP_PKEY_CTX *ctx, int type, int p1, void *p2)
 	{
@@ -333,7 +335,11 @@ const EVP_PKEY_METHOD ec_pkey_meth =
 	0,0,
 
 	0,
+#ifndef OPENSSL_NO_ECDH
 	pkey_ec_derive,
+#else
+	0,
+#endif
 
 	pkey_ec_ctrl,
 	pkey_ec_ctrl_str
diff --git a/crypto/ec/ecp_nistp256.c b/crypto/ec/ecp_nistp256.c
index 4bc0f5dce022..b884f73d31c0 100644
--- a/crypto/ec/ecp_nistp256.c
+++ b/crypto/ec/ecp_nistp256.c
@@ -113,7 +113,6 @@ 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
diff --git a/crypto/ec/ectest.c b/crypto/ec/ectest.c
index d1bf980599c8..8e4154d86549 100644
--- a/crypto/ec/ectest.c
+++ b/crypto/ec/ectest.c
@@ -1366,7 +1366,7 @@ static const struct nistp_test_params nistp_tests_params[] =
 		},
 	};
 
-void nistp_single_test(const struct nistp_test_params *test)
+static void nistp_single_test(const struct nistp_test_params *test)
 	{
 	BN_CTX *ctx;
 	BIGNUM *p, *a, *b, *x, *y, *n, *m, *order;
@@ -1469,7 +1469,7 @@ void nistp_single_test(const struct nistp_test_params *test)
 	BN_CTX_free(ctx);
 	}
 
-void nistp_tests()
+static void nistp_tests()
 	{
 	unsigned i;
 
diff --git a/crypto/ecdsa/Makefile b/crypto/ecdsa/Makefile
index e89e0c010c6b..60c876df1abf 100644
--- a/crypto/ecdsa/Makefile
+++ b/crypto/ecdsa/Makefile
@@ -126,15 +126,16 @@ ecs_sign.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 ecs_sign.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 ecs_sign.o: ../../include/openssl/x509.h ../../include/openssl/x509_vfy.h
 ecs_sign.o: ecs_locl.h ecs_sign.c
-ecs_vrf.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
-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: ../../e_os.h ../../include/openssl/asn1.h
+ecs_vrf.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+ecs_vrf.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+ecs_vrf.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+ecs_vrf.o: ../../include/openssl/ecdsa.h ../../include/openssl/engine.h
+ecs_vrf.o: ../../include/openssl/err.h ../../include/openssl/evp.h
 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
+ecs_vrf.o: ../../include/openssl/x509_vfy.h ../cryptlib.h ecs_locl.h ecs_vrf.c
diff --git a/crypto/ecdsa/ecs_vrf.c b/crypto/ecdsa/ecs_vrf.c
index ef9acf7b6102..2836efe5eff2 100644
--- a/crypto/ecdsa/ecs_vrf.c
+++ b/crypto/ecdsa/ecs_vrf.c
@@ -57,6 +57,7 @@
  */
 
 #include "ecs_locl.h"
+#include "cryptlib.h"
 #ifndef OPENSSL_NO_ENGINE
 #include <openssl/engine.h>
 #endif
@@ -84,13 +85,25 @@ int ECDSA_verify(int type, const unsigned char *dgst, int dgst_len,
 		const unsigned char *sigbuf, int sig_len, EC_KEY *eckey)
  	{
 	ECDSA_SIG *s;
+	const unsigned char *p = sigbuf;
+	unsigned char *der = NULL;
+	int derlen = -1;
 	int ret=-1;
 
 	s = ECDSA_SIG_new();
 	if (s == NULL) return(ret);
-	if (d2i_ECDSA_SIG(&s, &sigbuf, sig_len) == NULL) goto err;
+	if (d2i_ECDSA_SIG(&s, &p, sig_len) == NULL) goto err;
+	/* Ensure signature uses DER and doesn't have trailing garbage */
+	derlen = i2d_ECDSA_SIG(s, &der);
+	if (derlen != sig_len || memcmp(sigbuf, der, derlen))
+		goto err;
 	ret=ECDSA_do_verify(dgst, dgst_len, s, eckey);
 err:
+	if (derlen > 0)
+		{
+		OPENSSL_cleanse(der, derlen);
+		OPENSSL_free(der);
+		}
 	ECDSA_SIG_free(s);
 	return(ret);
 	}
diff --git a/crypto/engine/eng_dyn.c b/crypto/engine/eng_dyn.c
index 807da7a5eb84..01db4556aaa0 100644
--- a/crypto/engine/eng_dyn.c
+++ b/crypto/engine/eng_dyn.c
@@ -114,9 +114,6 @@ static const ENGINE_CMD_DEFN dynamic_cmd_defns[] = {
 		ENGINE_CMD_FLAG_NO_INPUT},
 	{0, NULL, NULL, 0}
 	};
-static const ENGINE_CMD_DEFN dynamic_cmd_defns_empty[] = {
-	{0, NULL, NULL, 0}
-	};
 
 /* Loading code stores state inside the ENGINE structure via the "ex_data"
  * element. We load all our state into a single structure and use that as a
diff --git a/crypto/evp/Makefile b/crypto/evp/Makefile
index fa25bff6b5ca..5d0c6b7db4c9 100644
--- a/crypto/evp/Makefile
+++ b/crypto/evp/Makefile
@@ -383,7 +383,7 @@ 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: ../constant_time_locl.h ../cryptlib.h evp_enc.c evp_locl.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
diff --git a/crypto/evp/e_des3.c b/crypto/evp/e_des3.c
index 8d7b7de292bc..7e1e8b309caa 100644
--- a/crypto/evp/e_des3.c
+++ b/crypto/evp/e_des3.c
@@ -124,12 +124,11 @@ static int des_ede_cbc_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out,
 #ifdef KSSL_DEBUG
 	{
         int i;
-        char *cp;
-	printf("des_ede_cbc_cipher(ctx=%lx, buflen=%d)\n", ctx, ctx->buf_len);
-	printf("\t iv= ");
+	fprintf(stderr,"des_ede_cbc_cipher(ctx=%p, buflen=%d)\n", ctx, ctx->buf_len);
+	fprintf(stderr,"\t iv= ");
         for(i=0;i<8;i++)
-                printf("%02X",ctx->iv[i]);
-	printf("\n");
+                fprintf(stderr,"%02X",ctx->iv[i]);
+	fprintf(stderr,"\n");
 	}
 #endif    /* KSSL_DEBUG */
 	while (inl>=EVP_MAXCHUNK)
@@ -260,11 +259,14 @@ 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", ctx);
-	printf("\tKEY= ");
-        for(i=0;i<24;i++) printf("%02X",key[i]); printf("\n");
-	printf("\t IV= ");
-        for(i=0;i<8;i++) printf("%02X",iv[i]); printf("\n");
+        fprintf(stderr,"des_ede3_init_key(ctx=%p)\n", ctx);
+	fprintf(stderr,"\tKEY= ");
+        for(i=0;i<24;i++) fprintf(stderr,"%02X",key[i]); fprintf(stderr,"\n");
+	if (iv) 
+		{
+		fprintf(stderr,"\t IV= ");
+		for(i=0;i<8;i++) fprintf(stderr,"%02X",iv[i]); fprintf(stderr,"\n");
+		}
 	}
 #endif	/* KSSL_DEBUG */
 
diff --git a/crypto/evp/evp_enc.c b/crypto/evp/evp_enc.c
index 7b1842a52a6d..2e863ac54204 100644
--- a/crypto/evp/evp_enc.c
+++ b/crypto/evp/evp_enc.c
@@ -67,7 +67,6 @@
 #ifdef OPENSSL_FIPS
 #include <openssl/fips.h>
 #endif
-#include "constant_time_locl.h"
 #include "evp_locl.h"
 
 #ifdef OPENSSL_FIPS
@@ -501,21 +500,21 @@ int EVP_DecryptFinal(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl)
 
 int EVP_DecryptFinal_ex(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl)
 	{
-	unsigned int i, b;
-        unsigned char pad, padding_good;
+	int i,n;
+	unsigned int b;
 	*outl=0;
 
 	if (ctx->cipher->flags & EVP_CIPH_FLAG_CUSTOM_CIPHER)
 		{
-		int ret = M_do_cipher(ctx, out, NULL, 0);
-		if (ret < 0)
+		i = M_do_cipher(ctx, out, NULL, 0);
+		if (i < 0)
 			return 0;
 		else
-			*outl = ret;
+			*outl = i;
 		return 1;
 		}
 
-	b=(unsigned int)(ctx->cipher->block_size);
+	b=ctx->cipher->block_size;
 	if (ctx->flags & EVP_CIPH_NO_PADDING)
 		{
 		if(ctx->buf_len)
@@ -534,34 +533,33 @@ int EVP_DecryptFinal_ex(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl)
 			return(0);
 			}
 		OPENSSL_assert(b <= sizeof ctx->final);
-		pad=ctx->final[b-1];
-
-		padding_good = (unsigned char)(~constant_time_is_zero_8(pad));
-		padding_good &= constant_time_ge_8(b, pad);
-
-                for (i = 1; i < b; ++i)
-			{
-			unsigned char is_pad_index = constant_time_lt_8(i, pad);
-			unsigned char pad_byte_good = constant_time_eq_8(ctx->final[b-i-1], pad);
-			padding_good &= constant_time_select_8(is_pad_index, pad_byte_good, 0xff);
-			}
 
 		/*
-		 * At least 1 byte is always padding, so we always write b - 1
-		 * bytes to avoid a timing leak. The caller is required to have |b|
-		 * bytes space in |out| by the API contract.
+		 * The following assumes that the ciphertext has been authenticated.
+		 * Otherwise it provides a padding oracle.
 		 */
-		for (i = 0; i < b - 1; ++i)
-			out[i] = ctx->final[i] & padding_good;
-		/* Safe cast: for a good padding, EVP_MAX_IV_LENGTH >= b >= pad */
-		*outl = padding_good & ((unsigned char)(b - pad));
-		return padding_good & 1;
+		n=ctx->final[b-1];
+		if (n == 0 || n > (int)b)
+			{
+			EVPerr(EVP_F_EVP_DECRYPTFINAL_EX,EVP_R_BAD_DECRYPT);
+			return(0);
+			}
+		for (i=0; i<n; i++)
+			{
+			if (ctx->final[--b] != n)
+				{
+				EVPerr(EVP_F_EVP_DECRYPTFINAL_EX,EVP_R_BAD_DECRYPT);
+				return(0);
+				}
+			}
+		n=ctx->cipher->block_size-n;
+		for (i=0; i<n; i++)
+			out[i]=ctx->final[i];
+		*outl=n;
 		}
 	else
-		{
-		*outl = 0;
-		return 1;
-		}
+		*outl=0;
+	return(1);
 	}
 
 void EVP_CIPHER_CTX_free(EVP_CIPHER_CTX *ctx)
diff --git a/crypto/md32_common.h b/crypto/md32_common.h
index bb7381952a22..a75f136ca540 100644
--- a/crypto/md32_common.h
+++ b/crypto/md32_common.h
@@ -225,8 +225,7 @@
 #define HOST_c2l(c,l)	(l =(((unsigned long)(*((c)++)))<<24),		\
 			 l|=(((unsigned long)(*((c)++)))<<16),		\
 			 l|=(((unsigned long)(*((c)++)))<< 8),		\
-			 l|=(((unsigned long)(*((c)++)))    ),		\
-			 l)
+			 l|=(((unsigned long)(*((c)++)))    )		)
 #endif
 #ifndef HOST_l2c
 #define HOST_l2c(l,c)	(*((c)++)=(unsigned char)(((l)>>24)&0xff),	\
@@ -262,8 +261,7 @@
 #define HOST_c2l(c,l)	(l =(((unsigned long)(*((c)++)))    ),		\
 			 l|=(((unsigned long)(*((c)++)))<< 8),		\
 			 l|=(((unsigned long)(*((c)++)))<<16),		\
-			 l|=(((unsigned long)(*((c)++)))<<24),		\
-			 l)
+			 l|=(((unsigned long)(*((c)++)))<<24)		)
 #endif
 #ifndef HOST_l2c
 #define HOST_l2c(l,c)	(*((c)++)=(unsigned char)(((l)    )&0xff),	\
diff --git a/crypto/mem.c b/crypto/mem.c
index 1cc62eafd122..01052cfa38e9 100644
--- a/crypto/mem.c
+++ b/crypto/mem.c
@@ -255,10 +255,12 @@ void *CRYPTO_malloc_locked(int num, const char *file, int line)
 
 	if (num <= 0) return NULL;
 
-	allow_customize = 0;
+	if(allow_customize)
+		allow_customize = 0;
 	if (malloc_debug_func != NULL)
 		{
-		allow_customize_debug = 0;
+		if(allow_customize_debug)
+			allow_customize_debug = 0;
 		malloc_debug_func(NULL, num, file, line, 0);
 		}
 	ret = malloc_locked_ex_func(num,file,line);
@@ -299,10 +301,12 @@ void *CRYPTO_malloc(int num, const char *file, int line)
 
 	if (num <= 0) return NULL;
 
-	allow_customize = 0;
+	if(allow_customize)
+		allow_customize = 0;
 	if (malloc_debug_func != NULL)
 		{
-		allow_customize_debug = 0;
+		if(allow_customize_debug)
+			allow_customize_debug = 0;
 		malloc_debug_func(NULL, num, file, line, 0);
 		}
 	ret = malloc_ex_func(num,file,line);
diff --git a/crypto/objects/obj_xref.h b/crypto/objects/obj_xref.h
index e23938c29604..33cb24b0708b 100644
--- a/crypto/objects/obj_xref.h
+++ b/crypto/objects/obj_xref.h
@@ -43,9 +43,6 @@ static const nid_triple sigoid_srt[] =
 
 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],
diff --git a/crypto/objects/objxref.pl b/crypto/objects/objxref.pl
index 731d3ae22c27..6c4c832ab205 100755
--- a/crypto/objects/objxref.pl
+++ b/crypto/objects/objxref.pl
@@ -90,7 +90,10 @@ EOF
 
 foreach (@srt2)
 	{
-	my $x = $xref_tbl{$_}[2];
+	my ($p1, $p2, $x) = @{$xref_tbl{$_}};
+	# If digest or signature algorithm is "undef" then the algorithm
+	# needs special handling and is excluded from the cross reference table.
+	next if $p1 eq "undef" || $p2 eq "undef";
 	print "\t\&sigoid_srt\[$x\],\n";
 	}
 
diff --git a/crypto/opensslv.h b/crypto/opensslv.h
index f375967ef913..0aebfd446626 100644
--- a/crypto/opensslv.h
+++ b/crypto/opensslv.h
@@ -29,11 +29,11 @@ extern "C" {
  * (Prior to 0.9.5a beta1, a different scheme was used: MMNNFFRBB for
  *  major minor fix final patch/beta)
  */
-#define OPENSSL_VERSION_NUMBER	0x100010afL
+#define OPENSSL_VERSION_NUMBER	0x100010bfL
 #ifdef OPENSSL_FIPS
-#define OPENSSL_VERSION_TEXT	"OpenSSL 1.0.1j-fips 15 Oct 2014"
+#define OPENSSL_VERSION_TEXT	"OpenSSL 1.0.1k-fips 8 Jan 2015"
 #else
-#define OPENSSL_VERSION_TEXT	"OpenSSL 1.0.1j 15 Oct 2014"
+#define OPENSSL_VERSION_TEXT	"OpenSSL 1.0.1k 8 Jan 2015"
 #endif
 #define OPENSSL_VERSION_PTEXT	" part of " OPENSSL_VERSION_TEXT
 
diff --git a/crypto/sha/asm/sha1-mips.pl b/crypto/sha/asm/sha1-mips.pl
index f1a702f38f53..197bc6b50e95 100755
--- a/crypto/sha/asm/sha1-mips.pl
+++ b/crypto/sha/asm/sha1-mips.pl
@@ -64,7 +64,7 @@ if ($flavour =~ /64|n32/i) {
 #
 ######################################################################
 
-$big_endian=(`echo MIPSEL | $ENV{CC} -E -P -`=~/MIPSEL/)?1:0;
+$big_endian=(`echo MIPSEL | $ENV{CC} -E -`=~/MIPSEL/)?1:0 if ($ENV{CC});
 
 for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);   }
 open STDOUT,">$output";
diff --git a/crypto/sha/asm/sha512-mips.pl b/crypto/sha/asm/sha512-mips.pl
index ffa053bb7d36..6807a2c7220b 100755
--- a/crypto/sha/asm/sha512-mips.pl
+++ b/crypto/sha/asm/sha512-mips.pl
@@ -68,7 +68,7 @@ $pf = ($flavour =~ /nubi/i) ? $t0 : $t2;
 #
 ######################################################################
 
-$big_endian=(`echo MIPSEL | $ENV{CC} -E -P -`=~/MIPSEL/)?1:0;
+$big_endian=(`echo MIPSEL | $ENV{CC} -E -`=~/MIPSEL/)?1:0 if ($ENV{CC});
 
 for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);	}
 open STDOUT,">$output";
diff --git a/crypto/ts/ts_rsp_sign.c b/crypto/ts/ts_rsp_sign.c
index b0f023c9d2f8..2c0e94edda16 100644
--- a/crypto/ts/ts_rsp_sign.c
+++ b/crypto/ts/ts_rsp_sign.c
@@ -977,7 +977,7 @@ TS_RESP_set_genTime_with_precision(ASN1_GENERALIZEDTIME *asn1_time,
 	if (precision > 0)
 	{
 		/* Add fraction of seconds (leave space for dot and null). */
-		BIO_snprintf(p, 2 + precision, ".%ld", usec);
+		BIO_snprintf(p, 2 + precision, ".%06ld", usec);
 		/* We cannot use the snprintf return value, 
 		   because it might have been truncated. */
 		p += strlen(p);
diff --git a/crypto/x509/x509.h b/crypto/x509/x509.h
index 092dd7450d3b..ed767f84dd23 100644
--- a/crypto/x509/x509.h
+++ b/crypto/x509/x509.h
@@ -768,6 +768,7 @@ 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);
+int X509_ALGOR_cmp(const X509_ALGOR *a, const X509_ALGOR *b);
 
 X509_NAME *X509_NAME_dup(X509_NAME *xn);
 X509_NAME_ENTRY *X509_NAME_ENTRY_dup(X509_NAME_ENTRY *ne);
diff --git a/crypto/x509/x509_vpm.c b/crypto/x509/x509_vpm.c
index dfd89d89faf0..ba546bd1fabf 100644
--- a/crypto/x509/x509_vpm.c
+++ b/crypto/x509/x509_vpm.c
@@ -89,6 +89,8 @@ X509_VERIFY_PARAM *X509_VERIFY_PARAM_new(void)
 	{
 	X509_VERIFY_PARAM *param;
 	param = OPENSSL_malloc(sizeof(X509_VERIFY_PARAM));
+	if (!param)
+		return NULL;
 	memset(param, 0, sizeof(X509_VERIFY_PARAM));
 	x509_verify_param_zero(param);
 	return param;
diff --git a/crypto/x509/x_all.c b/crypto/x509/x_all.c
index e06602d65abc..fef55f89d9a0 100644
--- a/crypto/x509/x_all.c
+++ b/crypto/x509/x_all.c
@@ -72,6 +72,8 @@
 
 int X509_verify(X509 *a, EVP_PKEY *r)
 	{
+	if (X509_ALGOR_cmp(a->sig_alg, a->cert_info->signature))
+		return 0;
 	return(ASN1_item_verify(ASN1_ITEM_rptr(X509_CINF),a->sig_alg,
 		a->signature,a->cert_info,r));
 	}