Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame] | 1 | /* crypto/sha/sha512.c */ |
| 2 | /* ==================================================================== |
| 3 | * Copyright (c) 2004 The OpenSSL Project. All rights reserved |
| 4 | * according to the OpenSSL license [found in ../../LICENSE]. |
| 5 | * ==================================================================== |
| 6 | */ |
| 7 | #include <openssl/opensslconf.h> |
| 8 | #if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA512) |
| 9 | /* |
| 10 | * IMPLEMENTATION NOTES. |
| 11 | * |
| 12 | * As you might have noticed 32-bit hash algorithms: |
| 13 | * |
| 14 | * - permit SHA_LONG to be wider than 32-bit (case on CRAY); |
| 15 | * - optimized versions implement two transform functions: one operating |
| 16 | * on [aligned] data in host byte order and one - on data in input |
| 17 | * stream byte order; |
| 18 | * - share common byte-order neutral collector and padding function |
| 19 | * implementations, ../md32_common.h; |
| 20 | * |
| 21 | * Neither of the above applies to this SHA-512 implementations. Reasons |
| 22 | * [in reverse order] are: |
| 23 | * |
| 24 | * - it's the only 64-bit hash algorithm for the moment of this writing, |
| 25 | * there is no need for common collector/padding implementation [yet]; |
| 26 | * - by supporting only one transform function [which operates on |
| 27 | * *aligned* data in input stream byte order, big-endian in this case] |
| 28 | * we minimize burden of maintenance in two ways: a) collector/padding |
| 29 | * function is simpler; b) only one transform function to stare at; |
| 30 | * - SHA_LONG64 is required to be exactly 64-bit in order to be able to |
| 31 | * apply a number of optimizations to mitigate potential performance |
| 32 | * penalties caused by previous design decision; |
| 33 | * |
| 34 | * Caveat lector. |
| 35 | * |
| 36 | * Implementation relies on the fact that "long long" is 64-bit on |
| 37 | * both 32- and 64-bit platforms. If some compiler vendor comes up |
| 38 | * with 128-bit long long, adjustment to sha.h would be required. |
| 39 | * As this implementation relies on 64-bit integer type, it's totally |
| 40 | * inappropriate for platforms which don't support it, most notably |
| 41 | * 16-bit platforms. |
| 42 | * <appro@fy.chalmers.se> |
| 43 | */ |
| 44 | #include <stdlib.h> |
| 45 | #include <string.h> |
| 46 | |
| 47 | #include <openssl/crypto.h> |
| 48 | #include <openssl/sha.h> |
| 49 | #include <openssl/opensslv.h> |
| 50 | |
| 51 | #include "cryptlib.h" |
| 52 | |
| 53 | const char SHA512_version[]="SHA-512" OPENSSL_VERSION_PTEXT; |
| 54 | |
| 55 | #if defined(__i386) || defined(__i386__) || defined(_M_IX86) || \ |
| 56 | defined(__x86_64) || defined(_M_AMD64) || defined(_M_X64) || \ |
| 57 | defined(__s390__) || defined(__s390x__) || \ |
| 58 | defined(SHA512_ASM) |
| 59 | #define SHA512_BLOCK_CAN_MANAGE_UNALIGNED_DATA |
| 60 | #endif |
| 61 | |
Alexandre Savard | 7541067 | 2012-08-08 09:50:01 -0400 | [diff] [blame] | 62 | int SHA384_Init (SHA512_CTX *c) |
Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame] | 63 | { |
Alexandre Savard | 7541067 | 2012-08-08 09:50:01 -0400 | [diff] [blame] | 64 | #if defined(SHA512_ASM) && (defined(__arm__) || defined(__arm)) |
| 65 | /* maintain dword order required by assembler module */ |
| 66 | unsigned int *h = (unsigned int *)c->h; |
| 67 | |
| 68 | h[0] = 0xcbbb9d5d; h[1] = 0xc1059ed8; |
| 69 | h[2] = 0x629a292a; h[3] = 0x367cd507; |
| 70 | h[4] = 0x9159015a; h[5] = 0x3070dd17; |
| 71 | h[6] = 0x152fecd8; h[7] = 0xf70e5939; |
| 72 | h[8] = 0x67332667; h[9] = 0xffc00b31; |
| 73 | h[10] = 0x8eb44a87; h[11] = 0x68581511; |
| 74 | h[12] = 0xdb0c2e0d; h[13] = 0x64f98fa7; |
| 75 | h[14] = 0x47b5481d; h[15] = 0xbefa4fa4; |
| 76 | #else |
Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame] | 77 | c->h[0]=U64(0xcbbb9d5dc1059ed8); |
| 78 | c->h[1]=U64(0x629a292a367cd507); |
| 79 | c->h[2]=U64(0x9159015a3070dd17); |
| 80 | c->h[3]=U64(0x152fecd8f70e5939); |
| 81 | c->h[4]=U64(0x67332667ffc00b31); |
| 82 | c->h[5]=U64(0x8eb44a8768581511); |
| 83 | c->h[6]=U64(0xdb0c2e0d64f98fa7); |
| 84 | c->h[7]=U64(0x47b5481dbefa4fa4); |
Alexandre Savard | 7541067 | 2012-08-08 09:50:01 -0400 | [diff] [blame] | 85 | #endif |
Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame] | 86 | c->Nl=0; c->Nh=0; |
| 87 | c->num=0; c->md_len=SHA384_DIGEST_LENGTH; |
| 88 | return 1; |
| 89 | } |
| 90 | |
Alexandre Savard | 7541067 | 2012-08-08 09:50:01 -0400 | [diff] [blame] | 91 | int SHA512_Init (SHA512_CTX *c) |
Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame] | 92 | { |
Alexandre Savard | 7541067 | 2012-08-08 09:50:01 -0400 | [diff] [blame] | 93 | #if defined(SHA512_ASM) && (defined(__arm__) || defined(__arm)) |
| 94 | /* maintain dword order required by assembler module */ |
| 95 | unsigned int *h = (unsigned int *)c->h; |
| 96 | |
| 97 | h[0] = 0x6a09e667; h[1] = 0xf3bcc908; |
| 98 | h[2] = 0xbb67ae85; h[3] = 0x84caa73b; |
| 99 | h[4] = 0x3c6ef372; h[5] = 0xfe94f82b; |
| 100 | h[6] = 0xa54ff53a; h[7] = 0x5f1d36f1; |
| 101 | h[8] = 0x510e527f; h[9] = 0xade682d1; |
| 102 | h[10] = 0x9b05688c; h[11] = 0x2b3e6c1f; |
| 103 | h[12] = 0x1f83d9ab; h[13] = 0xfb41bd6b; |
| 104 | h[14] = 0x5be0cd19; h[15] = 0x137e2179; |
| 105 | #else |
Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame] | 106 | c->h[0]=U64(0x6a09e667f3bcc908); |
| 107 | c->h[1]=U64(0xbb67ae8584caa73b); |
| 108 | c->h[2]=U64(0x3c6ef372fe94f82b); |
| 109 | c->h[3]=U64(0xa54ff53a5f1d36f1); |
| 110 | c->h[4]=U64(0x510e527fade682d1); |
| 111 | c->h[5]=U64(0x9b05688c2b3e6c1f); |
| 112 | c->h[6]=U64(0x1f83d9abfb41bd6b); |
| 113 | c->h[7]=U64(0x5be0cd19137e2179); |
Alexandre Savard | 7541067 | 2012-08-08 09:50:01 -0400 | [diff] [blame] | 114 | #endif |
Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame] | 115 | c->Nl=0; c->Nh=0; |
| 116 | c->num=0; c->md_len=SHA512_DIGEST_LENGTH; |
| 117 | return 1; |
| 118 | } |
| 119 | |
| 120 | #ifndef SHA512_ASM |
| 121 | static |
| 122 | #endif |
| 123 | void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num); |
| 124 | |
| 125 | int SHA512_Final (unsigned char *md, SHA512_CTX *c) |
| 126 | { |
| 127 | unsigned char *p=(unsigned char *)c->u.p; |
| 128 | size_t n=c->num; |
| 129 | |
| 130 | p[n]=0x80; /* There always is a room for one */ |
| 131 | n++; |
| 132 | if (n > (sizeof(c->u)-16)) |
| 133 | memset (p+n,0,sizeof(c->u)-n), n=0, |
| 134 | sha512_block_data_order (c,p,1); |
| 135 | |
| 136 | memset (p+n,0,sizeof(c->u)-16-n); |
| 137 | #ifdef B_ENDIAN |
| 138 | c->u.d[SHA_LBLOCK-2] = c->Nh; |
| 139 | c->u.d[SHA_LBLOCK-1] = c->Nl; |
| 140 | #else |
| 141 | p[sizeof(c->u)-1] = (unsigned char)(c->Nl); |
| 142 | p[sizeof(c->u)-2] = (unsigned char)(c->Nl>>8); |
| 143 | p[sizeof(c->u)-3] = (unsigned char)(c->Nl>>16); |
| 144 | p[sizeof(c->u)-4] = (unsigned char)(c->Nl>>24); |
| 145 | p[sizeof(c->u)-5] = (unsigned char)(c->Nl>>32); |
| 146 | p[sizeof(c->u)-6] = (unsigned char)(c->Nl>>40); |
| 147 | p[sizeof(c->u)-7] = (unsigned char)(c->Nl>>48); |
| 148 | p[sizeof(c->u)-8] = (unsigned char)(c->Nl>>56); |
| 149 | p[sizeof(c->u)-9] = (unsigned char)(c->Nh); |
| 150 | p[sizeof(c->u)-10] = (unsigned char)(c->Nh>>8); |
| 151 | p[sizeof(c->u)-11] = (unsigned char)(c->Nh>>16); |
| 152 | p[sizeof(c->u)-12] = (unsigned char)(c->Nh>>24); |
| 153 | p[sizeof(c->u)-13] = (unsigned char)(c->Nh>>32); |
| 154 | p[sizeof(c->u)-14] = (unsigned char)(c->Nh>>40); |
| 155 | p[sizeof(c->u)-15] = (unsigned char)(c->Nh>>48); |
| 156 | p[sizeof(c->u)-16] = (unsigned char)(c->Nh>>56); |
| 157 | #endif |
| 158 | |
| 159 | sha512_block_data_order (c,p,1); |
| 160 | |
| 161 | if (md==0) return 0; |
| 162 | |
Alexandre Savard | 7541067 | 2012-08-08 09:50:01 -0400 | [diff] [blame] | 163 | #if defined(SHA512_ASM) && (defined(__arm__) || defined(__arm)) |
| 164 | /* recall assembler dword order... */ |
| 165 | n = c->md_len; |
| 166 | if (n == SHA384_DIGEST_LENGTH || n == SHA512_DIGEST_LENGTH) |
| 167 | { |
| 168 | unsigned int *h = (unsigned int *)c->h, t; |
| 169 | |
| 170 | for (n/=4;n;n--) |
| 171 | { |
| 172 | t = *(h++); |
| 173 | *(md++) = (unsigned char)(t>>24); |
| 174 | *(md++) = (unsigned char)(t>>16); |
| 175 | *(md++) = (unsigned char)(t>>8); |
| 176 | *(md++) = (unsigned char)(t); |
| 177 | } |
| 178 | } |
| 179 | else return 0; |
| 180 | #else |
Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame] | 181 | switch (c->md_len) |
| 182 | { |
| 183 | /* Let compiler decide if it's appropriate to unroll... */ |
| 184 | case SHA384_DIGEST_LENGTH: |
| 185 | for (n=0;n<SHA384_DIGEST_LENGTH/8;n++) |
| 186 | { |
| 187 | SHA_LONG64 t = c->h[n]; |
| 188 | |
| 189 | *(md++) = (unsigned char)(t>>56); |
| 190 | *(md++) = (unsigned char)(t>>48); |
| 191 | *(md++) = (unsigned char)(t>>40); |
| 192 | *(md++) = (unsigned char)(t>>32); |
| 193 | *(md++) = (unsigned char)(t>>24); |
| 194 | *(md++) = (unsigned char)(t>>16); |
| 195 | *(md++) = (unsigned char)(t>>8); |
| 196 | *(md++) = (unsigned char)(t); |
| 197 | } |
| 198 | break; |
| 199 | case SHA512_DIGEST_LENGTH: |
| 200 | for (n=0;n<SHA512_DIGEST_LENGTH/8;n++) |
| 201 | { |
| 202 | SHA_LONG64 t = c->h[n]; |
| 203 | |
| 204 | *(md++) = (unsigned char)(t>>56); |
| 205 | *(md++) = (unsigned char)(t>>48); |
| 206 | *(md++) = (unsigned char)(t>>40); |
| 207 | *(md++) = (unsigned char)(t>>32); |
| 208 | *(md++) = (unsigned char)(t>>24); |
| 209 | *(md++) = (unsigned char)(t>>16); |
| 210 | *(md++) = (unsigned char)(t>>8); |
| 211 | *(md++) = (unsigned char)(t); |
| 212 | } |
| 213 | break; |
| 214 | /* ... as well as make sure md_len is not abused. */ |
| 215 | default: return 0; |
| 216 | } |
Alexandre Savard | 7541067 | 2012-08-08 09:50:01 -0400 | [diff] [blame] | 217 | #endif |
Alexandre Savard | 1b09e31 | 2012-08-07 20:33:29 -0400 | [diff] [blame] | 218 | return 1; |
| 219 | } |
| 220 | |
| 221 | int SHA384_Final (unsigned char *md,SHA512_CTX *c) |
| 222 | { return SHA512_Final (md,c); } |
| 223 | |
| 224 | int SHA512_Update (SHA512_CTX *c, const void *_data, size_t len) |
| 225 | { |
| 226 | SHA_LONG64 l; |
| 227 | unsigned char *p=c->u.p; |
| 228 | const unsigned char *data=(const unsigned char *)_data; |
| 229 | |
| 230 | if (len==0) return 1; |
| 231 | |
| 232 | l = (c->Nl+(((SHA_LONG64)len)<<3))&U64(0xffffffffffffffff); |
| 233 | if (l < c->Nl) c->Nh++; |
| 234 | if (sizeof(len)>=8) c->Nh+=(((SHA_LONG64)len)>>61); |
| 235 | c->Nl=l; |
| 236 | |
| 237 | if (c->num != 0) |
| 238 | { |
| 239 | size_t n = sizeof(c->u) - c->num; |
| 240 | |
| 241 | if (len < n) |
| 242 | { |
| 243 | memcpy (p+c->num,data,len), c->num += (unsigned int)len; |
| 244 | return 1; |
| 245 | } |
| 246 | else { |
| 247 | memcpy (p+c->num,data,n), c->num = 0; |
| 248 | len-=n, data+=n; |
| 249 | sha512_block_data_order (c,p,1); |
| 250 | } |
| 251 | } |
| 252 | |
| 253 | if (len >= sizeof(c->u)) |
| 254 | { |
| 255 | #ifndef SHA512_BLOCK_CAN_MANAGE_UNALIGNED_DATA |
| 256 | if ((size_t)data%sizeof(c->u.d[0]) != 0) |
| 257 | while (len >= sizeof(c->u)) |
| 258 | memcpy (p,data,sizeof(c->u)), |
| 259 | sha512_block_data_order (c,p,1), |
| 260 | len -= sizeof(c->u), |
| 261 | data += sizeof(c->u); |
| 262 | else |
| 263 | #endif |
| 264 | sha512_block_data_order (c,data,len/sizeof(c->u)), |
| 265 | data += len, |
| 266 | len %= sizeof(c->u), |
| 267 | data -= len; |
| 268 | } |
| 269 | |
| 270 | if (len != 0) memcpy (p,data,len), c->num = (int)len; |
| 271 | |
| 272 | return 1; |
| 273 | } |
| 274 | |
| 275 | int SHA384_Update (SHA512_CTX *c, const void *data, size_t len) |
| 276 | { return SHA512_Update (c,data,len); } |
| 277 | |
| 278 | void SHA512_Transform (SHA512_CTX *c, const unsigned char *data) |
| 279 | { sha512_block_data_order (c,data,1); } |
| 280 | |
| 281 | unsigned char *SHA384(const unsigned char *d, size_t n, unsigned char *md) |
| 282 | { |
| 283 | SHA512_CTX c; |
| 284 | static unsigned char m[SHA384_DIGEST_LENGTH]; |
| 285 | |
| 286 | if (md == NULL) md=m; |
| 287 | SHA384_Init(&c); |
| 288 | SHA512_Update(&c,d,n); |
| 289 | SHA512_Final(md,&c); |
| 290 | OPENSSL_cleanse(&c,sizeof(c)); |
| 291 | return(md); |
| 292 | } |
| 293 | |
| 294 | unsigned char *SHA512(const unsigned char *d, size_t n, unsigned char *md) |
| 295 | { |
| 296 | SHA512_CTX c; |
| 297 | static unsigned char m[SHA512_DIGEST_LENGTH]; |
| 298 | |
| 299 | if (md == NULL) md=m; |
| 300 | SHA512_Init(&c); |
| 301 | SHA512_Update(&c,d,n); |
| 302 | SHA512_Final(md,&c); |
| 303 | OPENSSL_cleanse(&c,sizeof(c)); |
| 304 | return(md); |
| 305 | } |
| 306 | |
| 307 | #ifndef SHA512_ASM |
| 308 | static const SHA_LONG64 K512[80] = { |
| 309 | U64(0x428a2f98d728ae22),U64(0x7137449123ef65cd), |
| 310 | U64(0xb5c0fbcfec4d3b2f),U64(0xe9b5dba58189dbbc), |
| 311 | U64(0x3956c25bf348b538),U64(0x59f111f1b605d019), |
| 312 | U64(0x923f82a4af194f9b),U64(0xab1c5ed5da6d8118), |
| 313 | U64(0xd807aa98a3030242),U64(0x12835b0145706fbe), |
| 314 | U64(0x243185be4ee4b28c),U64(0x550c7dc3d5ffb4e2), |
| 315 | U64(0x72be5d74f27b896f),U64(0x80deb1fe3b1696b1), |
| 316 | U64(0x9bdc06a725c71235),U64(0xc19bf174cf692694), |
| 317 | U64(0xe49b69c19ef14ad2),U64(0xefbe4786384f25e3), |
| 318 | U64(0x0fc19dc68b8cd5b5),U64(0x240ca1cc77ac9c65), |
| 319 | U64(0x2de92c6f592b0275),U64(0x4a7484aa6ea6e483), |
| 320 | U64(0x5cb0a9dcbd41fbd4),U64(0x76f988da831153b5), |
| 321 | U64(0x983e5152ee66dfab),U64(0xa831c66d2db43210), |
| 322 | U64(0xb00327c898fb213f),U64(0xbf597fc7beef0ee4), |
| 323 | U64(0xc6e00bf33da88fc2),U64(0xd5a79147930aa725), |
| 324 | U64(0x06ca6351e003826f),U64(0x142929670a0e6e70), |
| 325 | U64(0x27b70a8546d22ffc),U64(0x2e1b21385c26c926), |
| 326 | U64(0x4d2c6dfc5ac42aed),U64(0x53380d139d95b3df), |
| 327 | U64(0x650a73548baf63de),U64(0x766a0abb3c77b2a8), |
| 328 | U64(0x81c2c92e47edaee6),U64(0x92722c851482353b), |
| 329 | U64(0xa2bfe8a14cf10364),U64(0xa81a664bbc423001), |
| 330 | U64(0xc24b8b70d0f89791),U64(0xc76c51a30654be30), |
| 331 | U64(0xd192e819d6ef5218),U64(0xd69906245565a910), |
| 332 | U64(0xf40e35855771202a),U64(0x106aa07032bbd1b8), |
| 333 | U64(0x19a4c116b8d2d0c8),U64(0x1e376c085141ab53), |
| 334 | U64(0x2748774cdf8eeb99),U64(0x34b0bcb5e19b48a8), |
| 335 | U64(0x391c0cb3c5c95a63),U64(0x4ed8aa4ae3418acb), |
| 336 | U64(0x5b9cca4f7763e373),U64(0x682e6ff3d6b2b8a3), |
| 337 | U64(0x748f82ee5defb2fc),U64(0x78a5636f43172f60), |
| 338 | U64(0x84c87814a1f0ab72),U64(0x8cc702081a6439ec), |
| 339 | U64(0x90befffa23631e28),U64(0xa4506cebde82bde9), |
| 340 | U64(0xbef9a3f7b2c67915),U64(0xc67178f2e372532b), |
| 341 | U64(0xca273eceea26619c),U64(0xd186b8c721c0c207), |
| 342 | U64(0xeada7dd6cde0eb1e),U64(0xf57d4f7fee6ed178), |
| 343 | U64(0x06f067aa72176fba),U64(0x0a637dc5a2c898a6), |
| 344 | U64(0x113f9804bef90dae),U64(0x1b710b35131c471b), |
| 345 | U64(0x28db77f523047d84),U64(0x32caab7b40c72493), |
| 346 | U64(0x3c9ebe0a15c9bebc),U64(0x431d67c49c100d4c), |
| 347 | U64(0x4cc5d4becb3e42b6),U64(0x597f299cfc657e2a), |
| 348 | U64(0x5fcb6fab3ad6faec),U64(0x6c44198c4a475817) }; |
| 349 | |
| 350 | #ifndef PEDANTIC |
| 351 | # if defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM) |
| 352 | # if defined(__x86_64) || defined(__x86_64__) |
| 353 | # define ROTR(a,n) ({ SHA_LONG64 ret; \ |
| 354 | asm ("rorq %1,%0" \ |
| 355 | : "=r"(ret) \ |
| 356 | : "J"(n),"0"(a) \ |
| 357 | : "cc"); ret; }) |
| 358 | # if !defined(B_ENDIAN) |
| 359 | # define PULL64(x) ({ SHA_LONG64 ret=*((const SHA_LONG64 *)(&(x))); \ |
| 360 | asm ("bswapq %0" \ |
| 361 | : "=r"(ret) \ |
| 362 | : "0"(ret)); ret; }) |
| 363 | # endif |
| 364 | # elif (defined(__i386) || defined(__i386__)) && !defined(B_ENDIAN) |
| 365 | # if defined(I386_ONLY) |
| 366 | # define PULL64(x) ({ const unsigned int *p=(const unsigned int *)(&(x));\ |
| 367 | unsigned int hi=p[0],lo=p[1]; \ |
| 368 | asm("xchgb %%ah,%%al;xchgb %%dh,%%dl;"\ |
| 369 | "roll $16,%%eax; roll $16,%%edx; "\ |
| 370 | "xchgb %%ah,%%al;xchgb %%dh,%%dl;" \ |
| 371 | : "=a"(lo),"=d"(hi) \ |
| 372 | : "0"(lo),"1"(hi) : "cc"); \ |
| 373 | ((SHA_LONG64)hi)<<32|lo; }) |
| 374 | # else |
| 375 | # define PULL64(x) ({ const unsigned int *p=(const unsigned int *)(&(x));\ |
| 376 | unsigned int hi=p[0],lo=p[1]; \ |
| 377 | asm ("bswapl %0; bswapl %1;" \ |
| 378 | : "=r"(lo),"=r"(hi) \ |
| 379 | : "0"(lo),"1"(hi)); \ |
| 380 | ((SHA_LONG64)hi)<<32|lo; }) |
| 381 | # endif |
| 382 | # elif (defined(_ARCH_PPC) && defined(__64BIT__)) || defined(_ARCH_PPC64) |
| 383 | # define ROTR(a,n) ({ SHA_LONG64 ret; \ |
| 384 | asm ("rotrdi %0,%1,%2" \ |
| 385 | : "=r"(ret) \ |
| 386 | : "r"(a),"K"(n)); ret; }) |
| 387 | # endif |
| 388 | # elif defined(_MSC_VER) |
| 389 | # if defined(_WIN64) /* applies to both IA-64 and AMD64 */ |
| 390 | # pragma intrinsic(_rotr64) |
| 391 | # define ROTR(a,n) _rotr64((a),n) |
| 392 | # endif |
| 393 | # if defined(_M_IX86) && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM) |
| 394 | # if defined(I386_ONLY) |
| 395 | static SHA_LONG64 __fastcall __pull64be(const void *x) |
| 396 | { _asm mov edx, [ecx + 0] |
| 397 | _asm mov eax, [ecx + 4] |
| 398 | _asm xchg dh,dl |
| 399 | _asm xchg ah,al |
| 400 | _asm rol edx,16 |
| 401 | _asm rol eax,16 |
| 402 | _asm xchg dh,dl |
| 403 | _asm xchg ah,al |
| 404 | } |
| 405 | # else |
| 406 | static SHA_LONG64 __fastcall __pull64be(const void *x) |
| 407 | { _asm mov edx, [ecx + 0] |
| 408 | _asm mov eax, [ecx + 4] |
| 409 | _asm bswap edx |
| 410 | _asm bswap eax |
| 411 | } |
| 412 | # endif |
| 413 | # define PULL64(x) __pull64be(&(x)) |
| 414 | # if _MSC_VER<=1200 |
| 415 | # pragma inline_depth(0) |
| 416 | # endif |
| 417 | # endif |
| 418 | # endif |
| 419 | #endif |
| 420 | |
| 421 | #ifndef PULL64 |
| 422 | #define B(x,j) (((SHA_LONG64)(*(((const unsigned char *)(&x))+j)))<<((7-j)*8)) |
| 423 | #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)) |
| 424 | #endif |
| 425 | |
| 426 | #ifndef ROTR |
| 427 | #define ROTR(x,s) (((x)>>s) | (x)<<(64-s)) |
| 428 | #endif |
| 429 | |
| 430 | #define Sigma0(x) (ROTR((x),28) ^ ROTR((x),34) ^ ROTR((x),39)) |
| 431 | #define Sigma1(x) (ROTR((x),14) ^ ROTR((x),18) ^ ROTR((x),41)) |
| 432 | #define sigma0(x) (ROTR((x),1) ^ ROTR((x),8) ^ ((x)>>7)) |
| 433 | #define sigma1(x) (ROTR((x),19) ^ ROTR((x),61) ^ ((x)>>6)) |
| 434 | |
| 435 | #define Ch(x,y,z) (((x) & (y)) ^ ((~(x)) & (z))) |
| 436 | #define Maj(x,y,z) (((x) & (y)) ^ ((x) & (z)) ^ ((y) & (z))) |
| 437 | |
| 438 | |
| 439 | #if defined(__i386) || defined(__i386__) || defined(_M_IX86) |
| 440 | /* |
| 441 | * This code should give better results on 32-bit CPU with less than |
| 442 | * ~24 registers, both size and performance wise... |
| 443 | */ |
| 444 | static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num) |
| 445 | { |
| 446 | const SHA_LONG64 *W=in; |
| 447 | SHA_LONG64 A,E,T; |
| 448 | SHA_LONG64 X[9+80],*F; |
| 449 | int i; |
| 450 | |
| 451 | while (num--) { |
| 452 | |
| 453 | F = X+80; |
| 454 | A = ctx->h[0]; F[1] = ctx->h[1]; |
| 455 | F[2] = ctx->h[2]; F[3] = ctx->h[3]; |
| 456 | E = ctx->h[4]; F[5] = ctx->h[5]; |
| 457 | F[6] = ctx->h[6]; F[7] = ctx->h[7]; |
| 458 | |
| 459 | for (i=0;i<16;i++,F--) |
| 460 | { |
| 461 | #ifdef B_ENDIAN |
| 462 | T = W[i]; |
| 463 | #else |
| 464 | T = PULL64(W[i]); |
| 465 | #endif |
| 466 | F[0] = A; |
| 467 | F[4] = E; |
| 468 | F[8] = T; |
| 469 | T += F[7] + Sigma1(E) + Ch(E,F[5],F[6]) + K512[i]; |
| 470 | E = F[3] + T; |
| 471 | A = T + Sigma0(A) + Maj(A,F[1],F[2]); |
| 472 | } |
| 473 | |
| 474 | for (;i<80;i++,F--) |
| 475 | { |
| 476 | T = sigma0(F[8+16-1]); |
| 477 | T += sigma1(F[8+16-14]); |
| 478 | T += F[8+16] + F[8+16-9]; |
| 479 | |
| 480 | F[0] = A; |
| 481 | F[4] = E; |
| 482 | F[8] = T; |
| 483 | T += F[7] + Sigma1(E) + Ch(E,F[5],F[6]) + K512[i]; |
| 484 | E = F[3] + T; |
| 485 | A = T + Sigma0(A) + Maj(A,F[1],F[2]); |
| 486 | } |
| 487 | |
| 488 | ctx->h[0] += A; ctx->h[1] += F[1]; |
| 489 | ctx->h[2] += F[2]; ctx->h[3] += F[3]; |
| 490 | ctx->h[4] += E; ctx->h[5] += F[5]; |
| 491 | ctx->h[6] += F[6]; ctx->h[7] += F[7]; |
| 492 | |
| 493 | W+=SHA_LBLOCK; |
| 494 | } |
| 495 | } |
| 496 | |
| 497 | #elif defined(OPENSSL_SMALL_FOOTPRINT) |
| 498 | |
| 499 | static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num) |
| 500 | { |
| 501 | const SHA_LONG64 *W=in; |
| 502 | SHA_LONG64 a,b,c,d,e,f,g,h,s0,s1,T1,T2; |
| 503 | SHA_LONG64 X[16]; |
| 504 | int i; |
| 505 | |
| 506 | while (num--) { |
| 507 | |
| 508 | a = ctx->h[0]; b = ctx->h[1]; c = ctx->h[2]; d = ctx->h[3]; |
| 509 | e = ctx->h[4]; f = ctx->h[5]; g = ctx->h[6]; h = ctx->h[7]; |
| 510 | |
| 511 | for (i=0;i<16;i++) |
| 512 | { |
| 513 | #ifdef B_ENDIAN |
| 514 | T1 = X[i] = W[i]; |
| 515 | #else |
| 516 | T1 = X[i] = PULL64(W[i]); |
| 517 | #endif |
| 518 | T1 += h + Sigma1(e) + Ch(e,f,g) + K512[i]; |
| 519 | T2 = Sigma0(a) + Maj(a,b,c); |
| 520 | h = g; g = f; f = e; e = d + T1; |
| 521 | d = c; c = b; b = a; a = T1 + T2; |
| 522 | } |
| 523 | |
| 524 | for (;i<80;i++) |
| 525 | { |
| 526 | s0 = X[(i+1)&0x0f]; s0 = sigma0(s0); |
| 527 | s1 = X[(i+14)&0x0f]; s1 = sigma1(s1); |
| 528 | |
| 529 | T1 = X[i&0xf] += s0 + s1 + X[(i+9)&0xf]; |
| 530 | T1 += h + Sigma1(e) + Ch(e,f,g) + K512[i]; |
| 531 | T2 = Sigma0(a) + Maj(a,b,c); |
| 532 | h = g; g = f; f = e; e = d + T1; |
| 533 | d = c; c = b; b = a; a = T1 + T2; |
| 534 | } |
| 535 | |
| 536 | ctx->h[0] += a; ctx->h[1] += b; ctx->h[2] += c; ctx->h[3] += d; |
| 537 | ctx->h[4] += e; ctx->h[5] += f; ctx->h[6] += g; ctx->h[7] += h; |
| 538 | |
| 539 | W+=SHA_LBLOCK; |
| 540 | } |
| 541 | } |
| 542 | |
| 543 | #else |
| 544 | |
| 545 | #define ROUND_00_15(i,a,b,c,d,e,f,g,h) do { \ |
| 546 | T1 += h + Sigma1(e) + Ch(e,f,g) + K512[i]; \ |
| 547 | h = Sigma0(a) + Maj(a,b,c); \ |
| 548 | d += T1; h += T1; } while (0) |
| 549 | |
| 550 | #define ROUND_16_80(i,j,a,b,c,d,e,f,g,h,X) do { \ |
| 551 | s0 = X[(j+1)&0x0f]; s0 = sigma0(s0); \ |
| 552 | s1 = X[(j+14)&0x0f]; s1 = sigma1(s1); \ |
| 553 | T1 = X[(j)&0x0f] += s0 + s1 + X[(j+9)&0x0f]; \ |
| 554 | ROUND_00_15(i+j,a,b,c,d,e,f,g,h); } while (0) |
| 555 | |
| 556 | static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num) |
| 557 | { |
| 558 | const SHA_LONG64 *W=in; |
| 559 | SHA_LONG64 a,b,c,d,e,f,g,h,s0,s1,T1; |
| 560 | SHA_LONG64 X[16]; |
| 561 | int i; |
| 562 | |
| 563 | while (num--) { |
| 564 | |
| 565 | a = ctx->h[0]; b = ctx->h[1]; c = ctx->h[2]; d = ctx->h[3]; |
| 566 | e = ctx->h[4]; f = ctx->h[5]; g = ctx->h[6]; h = ctx->h[7]; |
| 567 | |
| 568 | #ifdef B_ENDIAN |
| 569 | T1 = X[0] = W[0]; ROUND_00_15(0,a,b,c,d,e,f,g,h); |
| 570 | T1 = X[1] = W[1]; ROUND_00_15(1,h,a,b,c,d,e,f,g); |
| 571 | T1 = X[2] = W[2]; ROUND_00_15(2,g,h,a,b,c,d,e,f); |
| 572 | T1 = X[3] = W[3]; ROUND_00_15(3,f,g,h,a,b,c,d,e); |
| 573 | T1 = X[4] = W[4]; ROUND_00_15(4,e,f,g,h,a,b,c,d); |
| 574 | T1 = X[5] = W[5]; ROUND_00_15(5,d,e,f,g,h,a,b,c); |
| 575 | T1 = X[6] = W[6]; ROUND_00_15(6,c,d,e,f,g,h,a,b); |
| 576 | T1 = X[7] = W[7]; ROUND_00_15(7,b,c,d,e,f,g,h,a); |
| 577 | T1 = X[8] = W[8]; ROUND_00_15(8,a,b,c,d,e,f,g,h); |
| 578 | T1 = X[9] = W[9]; ROUND_00_15(9,h,a,b,c,d,e,f,g); |
| 579 | T1 = X[10] = W[10]; ROUND_00_15(10,g,h,a,b,c,d,e,f); |
| 580 | T1 = X[11] = W[11]; ROUND_00_15(11,f,g,h,a,b,c,d,e); |
| 581 | T1 = X[12] = W[12]; ROUND_00_15(12,e,f,g,h,a,b,c,d); |
| 582 | T1 = X[13] = W[13]; ROUND_00_15(13,d,e,f,g,h,a,b,c); |
| 583 | T1 = X[14] = W[14]; ROUND_00_15(14,c,d,e,f,g,h,a,b); |
| 584 | T1 = X[15] = W[15]; ROUND_00_15(15,b,c,d,e,f,g,h,a); |
| 585 | #else |
| 586 | T1 = X[0] = PULL64(W[0]); ROUND_00_15(0,a,b,c,d,e,f,g,h); |
| 587 | T1 = X[1] = PULL64(W[1]); ROUND_00_15(1,h,a,b,c,d,e,f,g); |
| 588 | T1 = X[2] = PULL64(W[2]); ROUND_00_15(2,g,h,a,b,c,d,e,f); |
| 589 | T1 = X[3] = PULL64(W[3]); ROUND_00_15(3,f,g,h,a,b,c,d,e); |
| 590 | T1 = X[4] = PULL64(W[4]); ROUND_00_15(4,e,f,g,h,a,b,c,d); |
| 591 | T1 = X[5] = PULL64(W[5]); ROUND_00_15(5,d,e,f,g,h,a,b,c); |
| 592 | T1 = X[6] = PULL64(W[6]); ROUND_00_15(6,c,d,e,f,g,h,a,b); |
| 593 | T1 = X[7] = PULL64(W[7]); ROUND_00_15(7,b,c,d,e,f,g,h,a); |
| 594 | T1 = X[8] = PULL64(W[8]); ROUND_00_15(8,a,b,c,d,e,f,g,h); |
| 595 | T1 = X[9] = PULL64(W[9]); ROUND_00_15(9,h,a,b,c,d,e,f,g); |
| 596 | T1 = X[10] = PULL64(W[10]); ROUND_00_15(10,g,h,a,b,c,d,e,f); |
| 597 | T1 = X[11] = PULL64(W[11]); ROUND_00_15(11,f,g,h,a,b,c,d,e); |
| 598 | T1 = X[12] = PULL64(W[12]); ROUND_00_15(12,e,f,g,h,a,b,c,d); |
| 599 | T1 = X[13] = PULL64(W[13]); ROUND_00_15(13,d,e,f,g,h,a,b,c); |
| 600 | T1 = X[14] = PULL64(W[14]); ROUND_00_15(14,c,d,e,f,g,h,a,b); |
| 601 | T1 = X[15] = PULL64(W[15]); ROUND_00_15(15,b,c,d,e,f,g,h,a); |
| 602 | #endif |
| 603 | |
| 604 | for (i=16;i<80;i+=16) |
| 605 | { |
| 606 | ROUND_16_80(i, 0,a,b,c,d,e,f,g,h,X); |
| 607 | ROUND_16_80(i, 1,h,a,b,c,d,e,f,g,X); |
| 608 | ROUND_16_80(i, 2,g,h,a,b,c,d,e,f,X); |
| 609 | ROUND_16_80(i, 3,f,g,h,a,b,c,d,e,X); |
| 610 | ROUND_16_80(i, 4,e,f,g,h,a,b,c,d,X); |
| 611 | ROUND_16_80(i, 5,d,e,f,g,h,a,b,c,X); |
| 612 | ROUND_16_80(i, 6,c,d,e,f,g,h,a,b,X); |
| 613 | ROUND_16_80(i, 7,b,c,d,e,f,g,h,a,X); |
| 614 | ROUND_16_80(i, 8,a,b,c,d,e,f,g,h,X); |
| 615 | ROUND_16_80(i, 9,h,a,b,c,d,e,f,g,X); |
| 616 | ROUND_16_80(i,10,g,h,a,b,c,d,e,f,X); |
| 617 | ROUND_16_80(i,11,f,g,h,a,b,c,d,e,X); |
| 618 | ROUND_16_80(i,12,e,f,g,h,a,b,c,d,X); |
| 619 | ROUND_16_80(i,13,d,e,f,g,h,a,b,c,X); |
| 620 | ROUND_16_80(i,14,c,d,e,f,g,h,a,b,X); |
| 621 | ROUND_16_80(i,15,b,c,d,e,f,g,h,a,X); |
| 622 | } |
| 623 | |
| 624 | ctx->h[0] += a; ctx->h[1] += b; ctx->h[2] += c; ctx->h[3] += d; |
| 625 | ctx->h[4] += e; ctx->h[5] += f; ctx->h[6] += g; ctx->h[7] += h; |
| 626 | |
| 627 | W+=SHA_LBLOCK; |
| 628 | } |
| 629 | } |
| 630 | |
| 631 | #endif |
| 632 | |
| 633 | #endif /* SHA512_ASM */ |
| 634 | |
| 635 | #else /* !OPENSSL_NO_SHA512 */ |
| 636 | |
| 637 | #if defined(PEDANTIC) || defined(__DECC) || defined(OPENSSL_SYS_MACOSX) |
| 638 | static void *dummy=&dummy; |
| 639 | #endif |
| 640 | |
| 641 | #endif /* !OPENSSL_NO_SHA512 */ |