From 63077bd40c67eb1d6be31f0df60470c227b6894e Mon Sep 17 00:00:00 2001 From: Andy Polyakov Date: Thu, 20 May 2004 21:24:41 +0000 Subject: [PATCH] SHA-256/-512 update. A bug fix, SHA-512 tune-up for AMD64, hook for SSE2 code, Makefile update. --- crypto/sha/Makefile.ssl | 8 +++-- crypto/sha/sha256.c | 15 +++++---- crypto/sha/sha512.c | 67 +++++++++++++++++++++++++++++++++++++++-- 3 files changed, 78 insertions(+), 12 deletions(-) diff --git a/crypto/sha/Makefile.ssl b/crypto/sha/Makefile.ssl index 91b6cd2bb0..dc5d8c975d 100644 --- a/crypto/sha/Makefile.ssl +++ b/crypto/sha/Makefile.ssl @@ -23,12 +23,12 @@ CFLAGS= $(INCLUDES) $(CFLAG) ASFLAGS= $(INCLUDES) $(ASFLAG) GENERAL=Makefile -TEST=shatest.c sha1test.c +TEST=shatest.c sha1test.c sha256t.c sha512t.c APPS= LIB=$(TOP)/libcrypto.a -LIBSRC=sha_dgst.c sha1dgst.c sha_one.c sha1_one.c -LIBOBJ=sha_dgst.o sha1dgst.o sha_one.o sha1_one.o $(SHA1_ASM_OBJ) +LIBSRC=sha_dgst.c sha1dgst.c sha_one.c sha1_one.c sha256.c sha512.c +LIBOBJ=sha_dgst.o sha1dgst.o sha_one.o sha1_one.o sha256.o sha512.o $(SHA1_ASM_OBJ) SRC= $(LIBSRC) @@ -50,6 +50,8 @@ lib: $(LIBOBJ) # elf asm/sx86-elf.s: asm/sha1-586.pl ../perlasm/x86asm.pl (cd asm; $(PERL) sha1-586.pl elf $(CFLAGS) $(PROCESSOR) > sx86-elf.s) +asm/sse2-elf.s: asm/sha512-sse2.pl ../perlasm/x86asm.pl + (cd asm; $(PERL) sha512-sse2.pl elf $(CFLAGS) $(PROCESSOR) > sse2-elf.s) # a.out asm/sx86-out.o: asm/sx86unix.cpp diff --git a/crypto/sha/sha256.c b/crypto/sha/sha256.c index b5a9da1004..f31eb3e786 100644 --- a/crypto/sha/sha256.c +++ b/crypto/sha/sha256.c @@ -1,6 +1,7 @@ /* crypto/sha/sha256.c */ /* ==================================================================== - * Copyright (c) 2004 The OpenSSL Project. All rights reserved. + * Copyright (c) 2004 The OpenSSL Project. All rights reserved + * according to the OpenSSL license [found in ../../LICENSE]. * ==================================================================== */ #include @@ -131,6 +132,7 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) unsigned MD32_REG_T a,b,c,d,e,f,g,h,s0,s1,T1,T2; SHA_LONG X[16]; int i; + const unsigned char *data=in; while (num--) { @@ -139,7 +141,7 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) if (host) { - const SHA_LONG *W=in; + const SHA_LONG *W=(const SHA_LONG *)data; for (i=0;i<16;i++) { @@ -152,7 +154,6 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) } else { - const unsigned char *data=in; SHA_LONG l; for (i=0;i<16;i++) @@ -180,6 +181,7 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) ctx->h[0] += a; ctx->h[1] += b; ctx->h[2] += c; ctx->h[3] += d; ctx->h[4] += e; ctx->h[5] += f; ctx->h[6] += g; ctx->h[7] += h; + data += SHA256_CBLOCK; } } @@ -193,7 +195,7 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) #define ROUND_16_63(i,a,b,c,d,e,f,g,h,X) do { \ s0 = X[(i+1)&0x0f]; s0 = sigma0(s0); \ s1 = X[(i+14)&0x0f]; s1 = sigma1(s1); \ - T1 = X[i&0x0f] += s0 + s1 + X[(i+9)&0x0f]; \ + T1 = X[(i)&0x0f] += s0 + s1 + X[(i+9)&0x0f]; \ ROUND_00_15(i,a,b,c,d,e,f,g,h); } while (0) static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) @@ -201,6 +203,7 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) unsigned MD32_REG_T a,b,c,d,e,f,g,h,s0,s1,T1; SHA_LONG X[16]; int i; + const unsigned char *data=in; while (num--) { @@ -209,7 +212,7 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) if (host) { - const SHA_LONG *W=in; + const SHA_LONG *W=(const SHA_LONG *)data; T1 = X[0] = W[0]; ROUND_00_15(0,a,b,c,d,e,f,g,h); T1 = X[1] = W[1]; ROUND_00_15(1,h,a,b,c,d,e,f,g); @@ -230,7 +233,6 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) } else { - const unsigned char *data=in; SHA_LONG l; HOST_c2l(data,l); T1 = X[0] = l; ROUND_00_15(0,a,b,c,d,e,f,g,h); @@ -266,6 +268,7 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host) ctx->h[0] += a; ctx->h[1] += b; ctx->h[2] += c; ctx->h[3] += d; ctx->h[4] += e; ctx->h[5] += f; ctx->h[6] += g; ctx->h[7] += h; + data += SHA256_CBLOCK; } } diff --git a/crypto/sha/sha512.c b/crypto/sha/sha512.c index 074ac0ebe9..22d9168497 100644 --- a/crypto/sha/sha512.c +++ b/crypto/sha/sha512.c @@ -1,6 +1,7 @@ /* crypto/sha/sha512.c */ /* ==================================================================== - * Copyright (c) 2004 The OpenSSL Project. All rights reserved. + * Copyright (c) 2004 The OpenSSL Project. All rights reserved + * according to the OpenSSL license [found in ../../LICENSE]. * ==================================================================== */ /* @@ -48,6 +49,10 @@ const char *SHA512_version="SHA-512" OPENSSL_VERSION_PTEXT; +#if defined(_M_IX86) || defined(_M_AMD64) || defined(__i386) || defined(__x86_64) +#define SHA512_BLOCK_CAN_MANAGE_UNALIGNED_DATA +#endif + int SHA384_Init (SHA512_CTX *c) { c->h[0]=U64(0xcbbb9d5dc1059ed8); @@ -167,7 +172,7 @@ int SHA512_Update (SHA512_CTX *c, const void *_data, size_t len) if (len >= sizeof(c->u)) { #ifndef SHA512_BLOCK_CAN_MANAGE_UNALIGNED_DATA - if ((int)data%sizeof(c->u.d[0]) != 0) + if ((size_t)data%sizeof(c->u.d[0]) != 0) while (len >= sizeof(c->u)) memcpy (p,data,sizeof(c->u)), sha512_block (c,p,1), @@ -260,10 +265,47 @@ static const SHA_LONG64 K512[80] = { U64(0x4cc5d4becb3e42b6),U64(0x597f299cfc657e2a), U64(0x5fcb6fab3ad6faec),U64(0x6c44198c4a475817) }; +#ifndef PEDANTIC +# if defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM) +# if defined(__x86_64) || defined(__x86_64__) +# define PULL64(x) ({ SHA_LONG64 ret=*((SHA_LONG64 *)(&(x))); \ + asm ("bswapq %0" \ + : "=r"(ret) \ + : "0"(ret)); ret; }) +# endif +# endif +#endif + +#ifndef PULL64 #define B(x,j) (((SHA_LONG64)(*(((unsigned char *)(&x))+j)))<<((7-j)*8)) #define PULL64(x) (B(x,0)|B(x,1)|B(x,2)|B(x,3)|B(x,4)|B(x,5)|B(x,6)|B(x,7)) +#endif + +#ifndef PEDANTIC +# if defined(_MSC_VER) +# if defined(_WIN64) /* applies to both IA-64 and AMD64 */ +# define ROTR(a,n) _rotr64((a),n) +# endif +# elif defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM) +# if defined(__x86_64) || defined(__x86_64__) +# define ROTR(a,n) ({ unsigned long ret; \ + asm ("rorq %1,%0" \ + : "=r"(ret) \ + : "J"(n),"0"(a) \ + : "cc"); ret; }) +# elif defined(_ARCH_PPC) && defined(__64BIT__) +# define ROTR(a,n) ({ unsigned long ret; \ + asm ("rotrdi %0,%1,%2" \ + : "=r"(ret) \ + : "r"(a),"K"(n)); ret; }) +# endif +# endif +#endif +#ifndef ROTR #define ROTR(x,s) (((x)>>s) | (x)<<(64-s)) +#endif + #define Sigma0(x) (ROTR((x),28) ^ ROTR((x),34) ^ ROTR((x),39)) #define Sigma1(x) (ROTR((x),14) ^ ROTR((x),18) ^ ROTR((x),41)) #define sigma0(x) (ROTR((x),1) ^ ROTR((x),8) ^ ((x)>>7)) @@ -272,6 +314,15 @@ static const SHA_LONG64 K512[80] = { #define Ch(x,y,z) (((x) & (y)) ^ ((~(x)) & (z))) #define Maj(x,y,z) (((x) & (y)) ^ ((x) & (z)) ^ ((y) & (z))) +#if defined(OPENSSL_IA32_SSE2) && !defined(OPENSSL_NO_ASM) +#define GO_FOR_SSE2(ctx,in,num) do { \ + extern int OPENSSL_ia32cap; \ + void sha512_block_sse2(void *,const void *,size_t); \ + if (!(OPENSSL_ia32cap & (1<<26))) break; \ + sha512_block_sse2(ctx->h,in,num); return; \ + } while (0) +#endif + #ifdef OPENSSL_SMALL_FOOTPRINT static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num) @@ -281,6 +332,10 @@ static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num) SHA_LONG64 X[16]; int i; +#ifdef GO_FOR_SSE2 + GO_FOR_SSE2(ctx,in,num); +#endif + while (num--) { a = ctx->h[0]; b = ctx->h[1]; c = ctx->h[2]; d = ctx->h[3]; @@ -314,6 +369,7 @@ static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num) ctx->h[0] += a; ctx->h[1] += b; ctx->h[2] += c; ctx->h[3] += d; ctx->h[4] += e; ctx->h[5] += f; ctx->h[6] += g; ctx->h[7] += h; + W+=SHA_LBLOCK; } } @@ -327,7 +383,7 @@ static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num) #define ROUND_16_80(i,a,b,c,d,e,f,g,h,X) do { \ s0 = X[(i+1)&0x0f]; s0 = sigma0(s0); \ s1 = X[(i+14)&0x0f]; s1 = sigma1(s1); \ - T1 = X[i&0x0f] += s0 + s1 + X[(i+9)&0x0f]; \ + T1 = X[(i)&0x0f] += s0 + s1 + X[(i+9)&0x0f]; \ ROUND_00_15(i,a,b,c,d,e,f,g,h); } while (0) static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num) @@ -337,6 +393,10 @@ static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num) SHA_LONG64 X[16]; int i; +#ifdef GO_FOR_SSE2 + GO_FOR_SSE2(ctx,in,num); +#endif + while (num--) { a = ctx->h[0]; b = ctx->h[1]; c = ctx->h[2]; d = ctx->h[3]; @@ -393,6 +453,7 @@ static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num) ctx->h[0] += a; ctx->h[1] += b; ctx->h[2] += c; ctx->h[3] += d; ctx->h[4] += e; ctx->h[5] += f; ctx->h[6] += g; ctx->h[7] += h; + W+=SHA_LBLOCK; } } -- 2.25.1