Add assembly optimized code for OpenSSL on powerpc, powerpc64 and powerpc64le

Summary:
1. https://github.com/openssl/openssl/commit/34ab13b7d8e3e723adb60be8142e38b7c9cd382a
needs to be merged for ELFv2 support on big-endian.
2. crypto/openssl/crypto/ppccap.c needs to be patched.
Same reason as in https://github.com/openssl/openssl/pull/17082.

Approved by:	jkim, jhibbits
MFC after:	1 month
Differential Revision: https://reviews.freebsd.org/D33076

3 files changed, 272 insertions, 2 deletions

diff --git a/secure/lib/libcrypto/Makefile b/secure/lib/libcrypto/Makefile
index ff9303c9edae..cf9259a19e16 100644
--- a/secure/lib/libcrypto/Makefile
+++ b/secure/lib/libcrypto/Makefile
@@ -28,6 +28,12 @@ SRCS+=	x86_64cpuid.S
 SRCS+=	armv4cpuid.S armcap.c
 .elif defined(ASM_i386)
 SRCS+=	x86cpuid.S
+.elif defined(ASM_powerpc)
+SRCS+=	ppccpuid.S ppccap.c
+.elif defined(ASM_powerpc64)
+SRCS+=	ppccpuid.S ppccap.c
+.elif defined(ASM_powerpc64le)
+SRCS+=	ppccpuid.S ppccap.c
 .else
 SRCS+=	mem_clr.c
 .endif
@@ -44,6 +50,12 @@ SRCS+=	aesni-x86_64.S vpaes-x86_64.S
 SRCS+=	aes-armv4.S aesv8-armx.S bsaes-armv7.S
 .elif defined(ASM_i386)
 SRCS+=	aes_core.c aesni-x86.S vpaes-x86.S
+.elif defined(ASM_powerpc)
+SRCS+=	aes_core.c aes-ppc.S vpaes-ppc.S aesp8-ppc.S
+.elif defined(ASM_powerpc64)
+SRCS+=	aes_core.c aes-ppc.S vpaes-ppc.S aesp8-ppc.S
+.elif defined(ASM_powerpc64le)
+SRCS+=	aes_core.c aes-ppc.S vpaes-ppc.S aesp8-ppc.S
 .else
 SRCS+=	aes_core.c
 .endif
@@ -99,6 +111,12 @@ SRCS+=	x86_64-mont.S x86_64-mont5.S
 SRCS+=	armv4-gf2m.S armv4-mont.S bn_asm.c
 .elif defined(ASM_i386)
 SRCS+=	bn-586.S co-586.S x86-gf2m.S x86-mont.S
+.elif defined(ASM_powerpc)
+SRCS+=	ppc.S ppc-mont.S
+.elif defined(ASM_powerpc64)
+SRCS+=	ppc.S ppc-mont.S
+.elif defined(ASM_powerpc64le)
+SRCS+=	ppc.S ppc-mont.S
 .else
 SRCS+=	bn_asm.c
 .endif
@@ -128,6 +146,12 @@ SRCS+=	chacha-x86_64.S
 SRCS+=	chacha-armv4.S
 .elif defined(ASM_i386)
 SRCS+=	chacha-x86.S
+.elif defined(ASM_powerpc)
+SRCS+=	chacha-ppc.S
+.elif defined(ASM_powerpc64)
+SRCS+=	chacha-ppc.S
+.elif defined(ASM_powerpc64le)
+SRCS+=	chacha-ppc.S
 .else
 SRCS+=	chacha_enc.c
 .endif
@@ -189,6 +213,10 @@ SRCS+=	ecp_nistz256-x86_64.S ecp_nistz256.c x25519-x86_64.S
 SRCS+=	ecp_nistz256-armv4.S ecp_nistz256.c
 .elif defined(ASM_i386)
 SRCS+=	ecp_nistz256-x86.S ecp_nistz256.c
+.elif defined(ASM_powerpc64)
+SRCS+=	ecp_nistz256-ppc64.S ecp_nistz256.c x25519-ppc64.S
+.elif defined(ASM_powerpc64le)
+SRCS+=	ecp_nistz256-ppc64.S ecp_nistz256.c x25519-ppc64.S
 .endif
 
 # engine
@@ -250,6 +278,12 @@ SRCS+=	aesni-gcm-x86_64.S ghash-x86_64.S
 SRCS+=	ghash-armv4.S ghashv8-armx.S
 .elif defined(ASM_i386)
 SRCS+=	ghash-x86.S
+.elif defined(ASM_powerpc)
+SRCS+=	ghashp8-ppc.S
+.elif defined(ASM_powerpc64)
+SRCS+=	ghashp8-ppc.S
+.elif defined(ASM_powerpc64le)
+SRCS+=	ghashp8-ppc.S
 .endif
 
 # objects
@@ -282,6 +316,12 @@ SRCS+=	poly1305-x86_64.S
 SRCS+=	poly1305-armv4.S
 .elif defined(ASM_i386)
 SRCS+=	poly1305-x86.S
+.elif defined(ASM_powerpc)
+SRCS+=	poly1305-ppc.S poly1305-ppcfp.S
+.elif defined(ASM_powerpc64)
+SRCS+=	poly1305-ppc.S poly1305-ppcfp.S
+.elif defined(ASM_powerpc64le)
+SRCS+=	poly1305-ppc.S poly1305-ppcfp.S
 .endif
 
 # rand
@@ -333,6 +373,12 @@ SRCS+=	sha256-mb-x86_64.S sha256-x86_64.S sha512-x86_64.S
 SRCS+=	keccak1600-armv4.S sha1-armv4-large.S sha256-armv4.S sha512-armv4.S
 .elif defined(ASM_i386)
 SRCS+=	keccak1600.c sha1-586.S sha256-586.S sha512-586.S
+.elif defined(ASM_powerpc)
+SRCS+=	keccak1600.c sha1-ppc.S sha256-ppc.S sha512-ppc.S sha256p8-ppc.S sha512p8-ppc.S
+.elif defined(ASM_powerpc64)
+SRCS+=	keccak1600-ppc64.S sha1-ppc.S sha256-ppc.S sha512-ppc.S sha256p8-ppc.S sha512p8-ppc.S
+.elif defined(ASM_powerpc64le)
+SRCS+=	keccak1600-ppc64.S sha1-ppc.S sha256-ppc.S sha512-ppc.S sha256p8-ppc.S sha512p8-ppc.S
 .else
 SRCS+=	keccak1600.c
 .endif
@@ -430,7 +476,7 @@ SRCS+=	buildinf.h
 CLEANDIRS=	openssl
 CLEANFILES=	buildinf.h opensslconf.h opensslconf.h.tmp
 
-.if defined(ASM_${MACHINE_CPUARCH})
+.if defined(ASM_${MACHINE_CPUARCH}) || defined(ASM_${MACHINE_ARCH})
 _cmd1=/%%NO_ASM%%/d
 .else
 _cmd1=s/%%NO_ASM%%//
@@ -471,6 +517,8 @@ PICFLAG+=	-DOPENSSL_PIC
 .if defined(ASM_amd64)
 .PATH:	${LCRYPTO_SRC}/crypto/bn/asm
 .endif
+.elif defined(ASM_${MACHINE_ARCH})
+.PATH:	${SRCTOP}/sys/crypto/openssl/${MACHINE_ARCH}
 .endif
 
 .PATH:	${LCRYPTO_SRC}/crypto \
diff --git a/secure/lib/libcrypto/Makefile.asm b/secure/lib/libcrypto/Makefile.asm
index 9480c3a59833..5e44285de0d9 100644
--- a/secure/lib/libcrypto/Makefile.asm
+++ b/secure/lib/libcrypto/Makefile.asm
@@ -295,6 +295,195 @@ CLEANFILES=	${ASM} ${SRCS:R:S/$/.s/}
 	env CC=cc perl ${PERLPATH} ${.IMPSRC} elf ${CFLAGS} ${.IMPSRC:R:S/$/.s/} ;\
 	cat ${.IMPSRC:R:S/$/.s/} ;\
 	echo '#endif' ) > ${.TARGET}
+
+.elif defined(ASM_powerpc)
+
+.PATH:	${LCRYPTO_SRC}/crypto \
+	${LCRYPTO_SRC}/crypto/aes/asm \
+	${LCRYPTO_SRC}/crypto/bn/asm \
+	${LCRYPTO_SRC}/crypto/chacha/asm \
+	${LCRYPTO_SRC}/crypto/modes/asm \
+	${LCRYPTO_SRC}/crypto/poly1305/asm \
+	${LCRYPTO_SRC}/crypto/sha/asm
+
+PERLPATH=	-I${LCRYPTO_SRC}/crypto/perlasm
+
+#cpuid
+SRCS=	ppccpuid.pl
+
+#bn
+SRCS+=	ppc.pl ppc-mont.pl
+
+#aes
+SRCS+=	aes-ppc.pl vpaes-ppc.pl aesp8-ppc.pl
+
+#sha1
+SRCS+=	sha1-ppc.pl sha512-ppc.pl sha512p8-ppc.pl
+
+#modes
+SRCS+=	ghashp8-ppc.pl
+
+#chacha
+SRCS+=	chacha-ppc.pl
+
+#poly1305
+SRCS+=	poly1305-ppc.pl poly1305-ppcfp.pl
+
+ASM=	${SRCS:R:S/$/.S/} sha256-ppc.S sha256p8-ppc.S
+
+all:	${ASM}
+
+CLEANFILES=	${ASM}
+.SUFFIXES:	.pl
+
+sha256-ppc.S:	sha512-ppc.pl
+	env CC=cc perl ${.ALLSRC} linux32 ${.TARGET:R:S/$/.s/}
+	( echo '/* $$'FreeBSD'$$ */' ;\
+	echo '/* Do not modify. This file is auto-generated from ${.ALLSRC:T:R:S/$/.pl/}. */' ;\
+	cat ${.TARGET:R:S/$/.s/}) > ${.TARGET}
+
+sha256p8-ppc.S:	sha512p8-ppc.pl
+	env CC=cc perl ${.ALLSRC} linux32 ${.TARGET:R:S/$/.s/}
+	( echo '/* $$'FreeBSD'$$ */' ;\
+	echo '/* Do not modify. This file is auto-generated from ${.ALLSRC:T:R:S/$/.pl/}. */' ;\
+	cat ${.TARGET:R:S/$/.s/}) > ${.TARGET}
+
+.pl.S:
+	env CC=cc perl ${.IMPSRC} linux32 ${.TARGET:R:S/$/.s/}
+	( echo '/* $$'FreeBSD'$$ */' ;\
+	echo '/* Do not modify. This file is auto-generated from ${.ALLSRC:T:R:S/$/.pl/}. */' ;\
+	cat ${.TARGET:R:S/$/.s/}) > ${.TARGET}
+
+.elif defined(ASM_powerpc64)
+
+.PATH:	${LCRYPTO_SRC}/crypto \
+	${LCRYPTO_SRC}/crypto/aes/asm \
+	${LCRYPTO_SRC}/crypto/bn/asm \
+	${LCRYPTO_SRC}/crypto/chacha/asm \
+	${LCRYPTO_SRC}/crypto/ec/asm \
+	${LCRYPTO_SRC}/crypto/modes/asm \
+	${LCRYPTO_SRC}/crypto/poly1305/asm \
+	${LCRYPTO_SRC}/crypto/sha/asm
+
+PERLPATH=	-I${LCRYPTO_SRC}/crypto/perlasm
+
+#cpuid
+SRCS=	ppccpuid.pl
+
+#bn
+SRCS+=	ppc.pl ppc-mont.pl
+
+#aes
+SRCS+=	aes-ppc.pl vpaes-ppc.pl aesp8-ppc.pl
+
+#sha1
+SRCS+=	sha1-ppc.pl sha512-ppc.pl sha512p8-ppc.pl
+
+#modes
+SRCS+=	ghashp8-ppc.pl
+
+#chacha
+SRCS+=	chacha-ppc.pl
+
+#poly1305
+SRCS+=	poly1305-ppc.pl poly1305-ppcfp.pl
+
+#ec
+SRCS+=	ecp_nistz256-ppc64.pl x25519-ppc64.pl
+
+#keccak1600
+SRCS+=	keccak1600-ppc64.pl
+
+ASM=	${SRCS:R:S/$/.S/} sha256-ppc.S sha256p8-ppc.S
+
+all:	${ASM}
+
+CLEANFILES=	${ASM}
+.SUFFIXES:	.pl
+
+sha256-ppc.S:	sha512-ppc.pl
+	env CC=cc perl ${.ALLSRC} linux64v2 ${.TARGET:R:S/$/.s/}
+	( echo '/* $$'FreeBSD'$$ */' ;\
+	echo '/* Do not modify. This file is auto-generated from ${.ALLSRC:T:R:S/$/.pl/}. */' ;\
+	cat ${.TARGET:R:S/$/.s/}) > ${.TARGET}
+
+sha256p8-ppc.S:	sha512p8-ppc.pl
+	env CC=cc perl ${.ALLSRC} linux64v2 ${.TARGET:R:S/$/.s/}
+	( echo '/* $$'FreeBSD'$$ */' ;\
+	echo '/* Do not modify. This file is auto-generated from ${.ALLSRC:T:R:S/$/.pl/}. */' ;\
+	cat ${.TARGET:R:S/$/.s/}) > ${.TARGET}
+
+.pl.S:
+	env CC=cc perl ${.IMPSRC} linux64v2 ${.TARGET:R:S/$/.s/}
+	( echo '/* $$'FreeBSD'$$ */' ;\
+	echo '/* Do not modify. This file is auto-generated from ${.ALLSRC:T:R:S/$/.pl/}. */' ;\
+	cat ${.TARGET:R:S/$/.s/}) > ${.TARGET}
+
+.elif defined(ASM_powerpc64le)
+
+.PATH:	${LCRYPTO_SRC}/crypto \
+	${LCRYPTO_SRC}/crypto/aes/asm \
+	${LCRYPTO_SRC}/crypto/bn/asm \
+	${LCRYPTO_SRC}/crypto/chacha/asm \
+	${LCRYPTO_SRC}/crypto/ec/asm \
+	${LCRYPTO_SRC}/crypto/modes/asm \
+	${LCRYPTO_SRC}/crypto/poly1305/asm \
+	${LCRYPTO_SRC}/crypto/sha/asm
+
+PERLPATH=	-I${LCRYPTO_SRC}/crypto/perlasm
+
+#cpuid
+SRCS=	ppccpuid.pl
+
+#bn
+SRCS+=	ppc.pl ppc-mont.pl
+
+#aes
+SRCS+=	aes-ppc.pl vpaes-ppc.pl aesp8-ppc.pl
+
+#sha1
+SRCS+=	sha1-ppc.pl sha512-ppc.pl sha512p8-ppc.pl
+
+#modes
+SRCS+=	ghashp8-ppc.pl
+
+#chacha
+SRCS+=	chacha-ppc.pl
+
+#poly1305
+SRCS+=	poly1305-ppc.pl poly1305-ppcfp.pl
+
+#ec
+SRCS+=	ecp_nistz256-ppc64.pl x25519-ppc64.pl
+
+#keccak1600
+SRCS+=	keccak1600-ppc64.pl
+
+ASM=	${SRCS:R:S/$/.S/} sha256-ppc.S sha256p8-ppc.S
+
+all:	${ASM}
+
+CLEANFILES=	${ASM}
+.SUFFIXES:	.pl
+
+sha256-ppc.S:	sha512-ppc.pl
+	env CC=cc perl ${.ALLSRC} linux64le ${.TARGET:R:S/$/.s/}
+	( echo '/* $$'FreeBSD'$$ */' ;\
+	echo '/* Do not modify. This file is auto-generated from ${.ALLSRC:T:R:S/$/.pl/}. */' ;\
+	cat ${.TARGET:R:S/$/.s/}) > ${.TARGET}
+
+sha256p8-ppc.S:	sha512p8-ppc.pl
+	env CC=cc perl ${.ALLSRC} linux64le ${.TARGET:R:S/$/.s/}
+	( echo '/* $$'FreeBSD'$$ */' ;\
+	echo '/* Do not modify. This file is auto-generated from ${.ALLSRC:T:R:S/$/.pl/}. */' ;\
+	cat ${.TARGET:R:S/$/.s/}) > ${.TARGET}
+
+.pl.S:
+	env CC=cc perl ${.IMPSRC} linux64le ${.TARGET:R:S/$/.s/}
+	( echo '/* $$'FreeBSD'$$ */' ;\
+	echo '/* Do not modify. This file is auto-generated from ${.ALLSRC:T:R:S/$/.pl/}. */' ;\
+	cat ${.TARGET:R:S/$/.s/}) > ${.TARGET}
+
 .endif
 
 .include <bsd.prog.mk>
diff --git a/secure/lib/libcrypto/Makefile.inc b/secure/lib/libcrypto/Makefile.inc
index 7e37eda14cd2..f6221b7f74e0 100644
--- a/secure/lib/libcrypto/Makefile.inc
+++ b/secure/lib/libcrypto/Makefile.inc
@@ -23,9 +23,12 @@ CFLAGS+=	-DB_ENDIAN
 .if ${MACHINE_CPUARCH} == "aarch64" || ${MACHINE_CPUARCH} == "amd64" || \
     ${MACHINE_CPUARCH} == "arm" || ${MACHINE_CPUARCH} == "i386"
 ASM_${MACHINE_CPUARCH}=
+.elif ${MACHINE_ARCH} == "powerpc" || ${MACHINE_ARCH} == "powerpc64" || \
+    ${MACHINE_ARCH} == "powerpc64le"
+ASM_${MACHINE_ARCH}=
 .endif
 
-.if defined(ASM_${MACHINE_CPUARCH})
+.if defined(ASM_${MACHINE_CPUARCH}) || defined(ASM_${MACHINE_ARCH})
 CFLAGS+=	-DOPENSSL_CPUID_OBJ
 .if defined(ASM_aarch64)
 CFLAGS+=	-DOPENSSL_BN_ASM_MONT
@@ -69,6 +72,36 @@ CFLAGS+=	-DGHASH_ASM
 CFLAGS+=	-DECP_NISTZ256_ASM
 CFLAGS+=	-DPADLOCK_ASM
 CFLAGS+=	-DPOLY1305_ASM
+.elif defined(ASM_powerpc)
+CFLAGS+=	-DOPENSSL_BN_ASM_MONT
+CFLAGS+=	-DAES_ASM
+CFLAGS+=	-DVPAES_ASM
+CFLAGS+=	-DSHA1_ASM
+CFLAGS+=	-DSHA256_ASM
+CFLAGS+=	-DSHA512_ASM
+CFLAGS+=	-DPOLY1305_ASM
+.elif defined(ASM_powerpc64)
+CFLAGS+=	-DOPENSSL_BN_ASM_MONT
+CFLAGS+=	-DAES_ASM
+CFLAGS+=	-DVPAES_ASM
+CFLAGS+=	-DSHA1_ASM
+CFLAGS+=	-DSHA256_ASM
+CFLAGS+=	-DSHA512_ASM
+CFLAGS+=	-DPOLY1305_ASM
+CFLAGS+=	-DECP_NISTZ256_ASM
+CFLAGS+=	-DX25519_ASM
+CFLAGS+=	-DKECCAK1600_ASM
+.elif defined(ASM_powerpc64le)
+CFLAGS+=	-DOPENSSL_BN_ASM_MONT
+CFLAGS+=	-DAES_ASM
+CFLAGS+=	-DVPAES_ASM
+CFLAGS+=	-DSHA1_ASM
+CFLAGS+=	-DSHA256_ASM
+CFLAGS+=	-DSHA512_ASM
+CFLAGS+=	-DPOLY1305_ASM
+CFLAGS+=	-DECP_NISTZ256_ASM
+CFLAGS+=	-DX25519_ASM
+CFLAGS+=	-DKECCAK1600_ASM
 .endif
 .endif