diff options
author | Dale Rahn <drahn@cvs.openbsd.org> | 2020-06-25 04:29:09 +0000 |
---|---|---|
committer | Dale Rahn <drahn@cvs.openbsd.org> | 2020-06-25 04:29:09 +0000 |
commit | e9bb560baca98b7c35156d261a3ece9e2a61db4d (patch) | |
tree | f6406f4f6824d67d24452cc0bd1293a4324ecd3f /lib | |
parent | 75fc76bf12f473542b7f0fcdfc3baeb6082e601c (diff) |
Intial attempt at powerpc64 libcrypto pieces.
just commit this kettenis@
Diffstat (limited to 'lib')
-rw-r--r-- | lib/libcrypto/arch/powerpc64/Makefile.inc | 88 | ||||
-rw-r--r-- | lib/libcrypto/arch/powerpc64/opensslconf.h | 298 |
2 files changed, 386 insertions, 0 deletions
diff --git a/lib/libcrypto/arch/powerpc64/Makefile.inc b/lib/libcrypto/arch/powerpc64/Makefile.inc new file mode 100644 index 00000000000..cc42f67c5c1 --- /dev/null +++ b/lib/libcrypto/arch/powerpc64/Makefile.inc @@ -0,0 +1,88 @@ +# $OpenBSD: Makefile.inc,v 1.1 2020/06/25 04:29:08 drahn Exp $ + +# powerpc-specific libcrypto build rules + +# aes +SRCS+= aes_core.c aes_cbc.c +# slower than C code +#CFLAGS+= -DAES_ASM +#SSLASM+= aes aes-ppc aes-ppc +# bf +SRCS+= bf_enc.c +# bn +SSLASM+= bn ppc bn-ppc +#SSLASM+= bn ppc-mont ppc-mont # bn_mul_mont_int +SSLASM+= bn ppc64-mont ppc64-mont # bn_mul_mont_fpu64 +CFLAGS+= -DOPENSSL_BN_ASM_MONT +# camellia +SRCS+= camellia.c cmll_cbc.c cmll_misc.c +# des +SRCS+= des_enc.c fcrypt_b.c +# rc4 +SRCS+= rc4_enc.c rc4_skey.c +# sha +CFLAGS+= -DSHA1_ASM +SSLASM+= sha sha1-ppc sha1-ppc +CFLAGS+= -DSHA256_ASM +SSLASM+= sha sha512-ppc sha256-ppc +# whrlpool +SRCS+= wp_block.c + +.for dir src dst in ${SSLASM} +SRCS+= ${dst}.S +GENERATED+=${dst}.S +${dst}.S: ${LCRYPTO_SRC}/${dir}/asm/${src}.pl + /usr/bin/perl \ + ${LCRYPTO_SRC}/${dir}/asm/${src}.pl linux32 ${.TARGET} > ${.TARGET} +.endfor + +#CFLAGS+= -DOPENSSL_CPUID_OBJ # it's commented out in ppccap.c +SRCS+= ppccpuid.S ppccap.c +GENERATED+=ppccpuid.S +ppccpuid.S: ${LCRYPTO_SRC}/ppccpuid.pl + /usr/bin/perl \ + ${LCRYPTO_SRC}/ppccpuid.pl linux32 > ${.TARGET} +# $OpenBSD: Makefile.inc,v 1.1 2020/06/25 04:29:08 drahn Exp $ + +# powerpc-specific libcrypto build rules + +# aes +SRCS+= aes_core.c aes_cbc.c +# slower than C code +#CFLAGS+= -DAES_ASM +#SSLASM+= aes aes-ppc aes-ppc +# bf +SRCS+= bf_enc.c +# bn +SSLASM+= bn ppc bn-ppc +#SSLASM+= bn ppc-mont ppc-mont # bn_mul_mont_int +SSLASM+= bn ppc64-mont ppc64-mont # bn_mul_mont_fpu64 +CFLAGS+= -DOPENSSL_BN_ASM_MONT +# camellia +SRCS+= camellia.c cmll_cbc.c cmll_misc.c +# des +SRCS+= des_enc.c fcrypt_b.c +# rc4 +SRCS+= rc4_enc.c rc4_skey.c +# sha +CFLAGS+= -DSHA1_ASM +SSLASM+= sha sha1-ppc sha1-ppc +CFLAGS+= -DSHA256_ASM +SSLASM+= sha sha512-ppc sha256-ppc +# whrlpool +SRCS+= wp_block.c + +.for dir src dst in ${SSLASM} +SRCS+= ${dst}.S +GENERATED+=${dst}.S +${dst}.S: ${LCRYPTO_SRC}/${dir}/asm/${src}.pl + /usr/bin/perl \ + ${LCRYPTO_SRC}/${dir}/asm/${src}.pl linux32 ${.TARGET} > ${.TARGET} +.endfor + +#CFLAGS+= -DOPENSSL_CPUID_OBJ # it's commented out in ppccap.c +SRCS+= ppccpuid.S ppccap.c +GENERATED+=ppccpuid.S +ppccpuid.S: ${LCRYPTO_SRC}/ppccpuid.pl + /usr/bin/perl \ + ${LCRYPTO_SRC}/ppccpuid.pl linux32 > ${.TARGET} diff --git a/lib/libcrypto/arch/powerpc64/opensslconf.h b/lib/libcrypto/arch/powerpc64/opensslconf.h new file mode 100644 index 00000000000..6fc15643a4b --- /dev/null +++ b/lib/libcrypto/arch/powerpc64/opensslconf.h @@ -0,0 +1,298 @@ +#include <openssl/opensslfeatures.h> +/* crypto/opensslconf.h.in */ + +#if defined(HEADER_CRYPTLIB_H) && !defined(OPENSSLDIR) +#define OPENSSLDIR "/etc/ssl" +#endif + +#undef OPENSSL_UNISTD +#define OPENSSL_UNISTD <unistd.h> + +#undef OPENSSL_EXPORT_VAR_AS_FUNCTION + +#if defined(HEADER_IDEA_H) && !defined(IDEA_INT) +#define IDEA_INT unsigned int +#endif + +#if defined(HEADER_MD2_H) && !defined(MD2_INT) +#define MD2_INT unsigned int +#endif + +#if defined(HEADER_RC2_H) && !defined(RC2_INT) +/* I need to put in a mod for the alpha - eay */ +#define RC2_INT unsigned int +#endif + +#if defined(HEADER_RC4_H) +#if !defined(RC4_INT) +/* using int types make the structure larger but make the code faster + * on most boxes I have tested - up to %20 faster. */ +/* + * I don't know what does "most" mean, but declaring "int" is a must on: + * - Intel P6 because partial register stalls are very expensive; + * - elder Alpha because it lacks byte load/store instructions; + */ +#define RC4_INT unsigned int +#endif +#if !defined(RC4_CHUNK) +/* + * This enables code handling data aligned at natural CPU word + * boundary. See crypto/rc4/rc4_enc.c for further details. + */ +#define RC4_CHUNK unsigned long +#endif +#endif + +#if (defined(HEADER_NEW_DES_H) || defined(HEADER_DES_H)) && !defined(DES_LONG) +/* If this is set to 'unsigned int' on a DEC Alpha, this gives about a + * %20 speed up (longs are 8 bytes, int's are 4). */ +#ifndef DES_LONG +#define DES_LONG unsigned int +#endif +#endif + +#if defined(HEADER_BN_H) && !defined(CONFIG_HEADER_BN_H) +#define CONFIG_HEADER_BN_H +#undef BN_LLONG + +/* Should we define BN_DIV2W here? */ + +/* Only one for the following should be defined */ +#define SIXTY_FOUR_BIT_LONG +#undef SIXTY_FOUR_BIT +#undef THIRTY_TWO_BIT +#endif + +#if defined(HEADER_RC4_LOCL_H) && !defined(CONFIG_HEADER_RC4_LOCL_H) +#define CONFIG_HEADER_RC4_LOCL_H +/* if this is defined data[i] is used instead of *data, this is a %20 + * speedup on x86 */ +#undef RC4_INDEX +#endif + +#if defined(HEADER_BF_LOCL_H) && !defined(CONFIG_HEADER_BF_LOCL_H) +#define CONFIG_HEADER_BF_LOCL_H +#undef BF_PTR +#endif /* HEADER_BF_LOCL_H */ + +#if defined(HEADER_DES_LOCL_H) && !defined(CONFIG_HEADER_DES_LOCL_H) +#define CONFIG_HEADER_DES_LOCL_H +#ifndef DES_DEFAULT_OPTIONS +/* the following is tweaked from a config script, that is why it is a + * protected undef/define */ +#ifndef DES_PTR +#undef DES_PTR +#endif + +/* This helps C compiler generate the correct code for multiple functional + * units. It reduces register dependancies at the expense of 2 more + * registers */ +#ifndef DES_RISC1 +#undef DES_RISC1 +#endif + +#ifndef DES_RISC2 +#undef DES_RISC2 +#endif + +#if defined(DES_RISC1) && defined(DES_RISC2) +YOU SHOULD NOT HAVE BOTH DES_RISC1 AND DES_RISC2 DEFINED!!!!! +#endif + +/* Unroll the inner loop, this sometimes helps, sometimes hinders. + * Very mucy CPU dependant */ +#ifndef DES_UNROLL +#define DES_UNROLL +#endif + +/* These default values were supplied by + * Peter Gutman <pgut001@cs.auckland.ac.nz> + * They are only used if nothing else has been defined */ +#if !defined(DES_PTR) && !defined(DES_RISC1) && !defined(DES_RISC2) && !defined(DES_UNROLL) +/* Special defines which change the way the code is built depending on the + CPU and OS. For SGI machines you can use _MIPS_SZLONG (32 or 64) to find + even newer MIPS CPU's, but at the moment one size fits all for + optimization options. Older Sparc's work better with only UNROLL, but + there's no way to tell at compile time what it is you're running on */ + +#if defined( sun ) /* Newer Sparc's */ +# define DES_PTR +# define DES_RISC1 +# define DES_UNROLL +#elif defined( __ultrix ) /* Older MIPS */ +# define DES_PTR +# define DES_RISC2 +# define DES_UNROLL +#elif defined( __osf1__ ) /* Alpha */ +# define DES_PTR +# define DES_RISC2 +#elif defined ( _AIX ) /* RS6000 */ + /* Unknown */ +#elif defined( __hpux ) /* HP-PA */ + /* Unknown */ +#elif defined( __aux ) /* 68K */ + /* Unknown */ +#elif defined( __dgux ) /* 88K (but P6 in latest boxes) */ +# define DES_UNROLL +#elif defined( __sgi ) /* Newer MIPS */ +# define DES_PTR +# define DES_RISC2 +# define DES_UNROLL +#elif defined(i386) || defined(__i386__) /* x86 boxes, should be gcc */ +# define DES_PTR +# define DES_RISC1 +# define DES_UNROLL +#endif /* Systems-specific speed defines */ +#endif + +#endif /* DES_DEFAULT_OPTIONS */ +#endif /* HEADER_DES_LOCL_H */ +#include <openssl/opensslfeatures.h> +/* crypto/opensslconf.h.in */ + +#if defined(HEADER_CRYPTLIB_H) && !defined(OPENSSLDIR) +#define OPENSSLDIR "/etc/ssl" +#endif + +#undef OPENSSL_UNISTD +#define OPENSSL_UNISTD <unistd.h> + +#undef OPENSSL_EXPORT_VAR_AS_FUNCTION + +#if defined(HEADER_IDEA_H) && !defined(IDEA_INT) +#define IDEA_INT unsigned int +#endif + +#if defined(HEADER_MD2_H) && !defined(MD2_INT) +#define MD2_INT unsigned int +#endif + +#if defined(HEADER_RC2_H) && !defined(RC2_INT) +/* I need to put in a mod for the alpha - eay */ +#define RC2_INT unsigned int +#endif + +#if defined(HEADER_RC4_H) +#if !defined(RC4_INT) +/* using int types make the structure larger but make the code faster + * on most boxes I have tested - up to %20 faster. */ +/* + * I don't know what does "most" mean, but declaring "int" is a must on: + * - Intel P6 because partial register stalls are very expensive; + * - elder Alpha because it lacks byte load/store instructions; + */ +#define RC4_INT unsigned int +#endif +#if !defined(RC4_CHUNK) +/* + * This enables code handling data aligned at natural CPU word + * boundary. See crypto/rc4/rc4_enc.c for further details. + */ +#define RC4_CHUNK unsigned long +#endif +#endif + +#if (defined(HEADER_NEW_DES_H) || defined(HEADER_DES_H)) && !defined(DES_LONG) +/* If this is set to 'unsigned int' on a DEC Alpha, this gives about a + * %20 speed up (longs are 8 bytes, int's are 4). */ +#ifndef DES_LONG +#define DES_LONG unsigned int +#endif +#endif + +#if defined(HEADER_BN_H) && !defined(CONFIG_HEADER_BN_H) +#define CONFIG_HEADER_BN_H +#undef BN_LLONG + +/* Should we define BN_DIV2W here? */ + +/* Only one for the following should be defined */ +#define SIXTY_FOUR_BIT_LONG +#undef SIXTY_FOUR_BIT +#undef THIRTY_TWO_BIT +#endif + +#if defined(HEADER_RC4_LOCL_H) && !defined(CONFIG_HEADER_RC4_LOCL_H) +#define CONFIG_HEADER_RC4_LOCL_H +/* if this is defined data[i] is used instead of *data, this is a %20 + * speedup on x86 */ +#undef RC4_INDEX +#endif + +#if defined(HEADER_BF_LOCL_H) && !defined(CONFIG_HEADER_BF_LOCL_H) +#define CONFIG_HEADER_BF_LOCL_H +#undef BF_PTR +#endif /* HEADER_BF_LOCL_H */ + +#if defined(HEADER_DES_LOCL_H) && !defined(CONFIG_HEADER_DES_LOCL_H) +#define CONFIG_HEADER_DES_LOCL_H +#ifndef DES_DEFAULT_OPTIONS +/* the following is tweaked from a config script, that is why it is a + * protected undef/define */ +#ifndef DES_PTR +#undef DES_PTR +#endif + +/* This helps C compiler generate the correct code for multiple functional + * units. It reduces register dependancies at the expense of 2 more + * registers */ +#ifndef DES_RISC1 +#undef DES_RISC1 +#endif + +#ifndef DES_RISC2 +#undef DES_RISC2 +#endif + +#if defined(DES_RISC1) && defined(DES_RISC2) +YOU SHOULD NOT HAVE BOTH DES_RISC1 AND DES_RISC2 DEFINED!!!!! +#endif + +/* Unroll the inner loop, this sometimes helps, sometimes hinders. + * Very mucy CPU dependant */ +#ifndef DES_UNROLL +#define DES_UNROLL +#endif + +/* These default values were supplied by + * Peter Gutman <pgut001@cs.auckland.ac.nz> + * They are only used if nothing else has been defined */ +#if !defined(DES_PTR) && !defined(DES_RISC1) && !defined(DES_RISC2) && !defined(DES_UNROLL) +/* Special defines which change the way the code is built depending on the + CPU and OS. For SGI machines you can use _MIPS_SZLONG (32 or 64) to find + even newer MIPS CPU's, but at the moment one size fits all for + optimization options. Older Sparc's work better with only UNROLL, but + there's no way to tell at compile time what it is you're running on */ + +#if defined( sun ) /* Newer Sparc's */ +# define DES_PTR +# define DES_RISC1 +# define DES_UNROLL +#elif defined( __ultrix ) /* Older MIPS */ +# define DES_PTR +# define DES_RISC2 +# define DES_UNROLL +#elif defined( __osf1__ ) /* Alpha */ +# define DES_PTR +# define DES_RISC2 +#elif defined ( _AIX ) /* RS6000 */ + /* Unknown */ +#elif defined( __hpux ) /* HP-PA */ + /* Unknown */ +#elif defined( __aux ) /* 68K */ + /* Unknown */ +#elif defined( __dgux ) /* 88K (but P6 in latest boxes) */ +# define DES_UNROLL +#elif defined( __sgi ) /* Newer MIPS */ +# define DES_PTR +# define DES_RISC2 +# define DES_UNROLL +#elif defined(i386) || defined(__i386__) /* x86 boxes, should be gcc */ +# define DES_PTR +# define DES_RISC1 +# define DES_UNROLL +#endif /* Systems-specific speed defines */ +#endif + +#endif /* DES_DEFAULT_OPTIONS */ +#endif /* HEADER_DES_LOCL_H */ |