wolf SSL / wolfSSL-TLS13-Beta

wolfSSL 3.11.1 for TLS1.3 beta

Fork of wolfSSL by wolf SSL

wolfcrypt/src/sha512.c@7:481bce714567, 2017-05-02 (annotated)

Committer:
wolfSSL
Date:
Tue May 02 08:44:47 2017 +0000
Revision:
7:481bce714567
wolfSSL3.10.2

Who changed what in which revision?

UserRevisionLine numberNew contents of line
wolfSSL 7:481bce714567 1 /* sha512.c
wolfSSL 7:481bce714567 2 *
wolfSSL 7:481bce714567 3 * Copyright (C) 2006-2016 wolfSSL Inc.
wolfSSL 7:481bce714567 4 *
wolfSSL 7:481bce714567 5 * This file is part of wolfSSL.
wolfSSL 7:481bce714567 6 *
wolfSSL 7:481bce714567 7 * wolfSSL is free software; you can redistribute it and/or modify
wolfSSL 7:481bce714567 8 * it under the terms of the GNU General Public License as published by
wolfSSL 7:481bce714567 9 * the Free Software Foundation; either version 2 of the License, or
wolfSSL 7:481bce714567 10 * (at your option) any later version.
wolfSSL 7:481bce714567 11 *
wolfSSL 7:481bce714567 12 * wolfSSL is distributed in the hope that it will be useful,
wolfSSL 7:481bce714567 13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
wolfSSL 7:481bce714567 14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
wolfSSL 7:481bce714567 15 * GNU General Public License for more details.
wolfSSL 7:481bce714567 16 *
wolfSSL 7:481bce714567 17 * You should have received a copy of the GNU General Public License
wolfSSL 7:481bce714567 18 * along with this program; if not, write to the Free Software
wolfSSL 7:481bce714567 19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
wolfSSL 7:481bce714567 20 */
wolfSSL 7:481bce714567 21
wolfSSL 7:481bce714567 22
wolfSSL 7:481bce714567 23 #ifdef HAVE_CONFIG_H
wolfSSL 7:481bce714567 24 #include <config.h>
wolfSSL 7:481bce714567 25 #endif
wolfSSL 7:481bce714567 26
wolfSSL 7:481bce714567 27 #include <wolfssl/wolfcrypt/settings.h>
wolfSSL 7:481bce714567 28 #include <wolfssl/wolfcrypt/sha512.h>
wolfSSL 7:481bce714567 29
wolfSSL 7:481bce714567 30 #ifdef WOLFSSL_SHA512
wolfSSL 7:481bce714567 31
wolfSSL 7:481bce714567 32 #ifdef HAVE_FIPS
wolfSSL 7:481bce714567 33 int wc_InitSha512(Sha512* sha)
wolfSSL 7:481bce714567 34 {
wolfSSL 7:481bce714567 35 return InitSha512_fips(sha);
wolfSSL 7:481bce714567 36 }
wolfSSL 7:481bce714567 37
wolfSSL 7:481bce714567 38
wolfSSL 7:481bce714567 39 int wc_Sha512Update(Sha512* sha, const byte* data, word32 len)
wolfSSL 7:481bce714567 40 {
wolfSSL 7:481bce714567 41 return Sha512Update_fips(sha, data, len);
wolfSSL 7:481bce714567 42 }
wolfSSL 7:481bce714567 43
wolfSSL 7:481bce714567 44
wolfSSL 7:481bce714567 45 int wc_Sha512Final(Sha512* sha, byte* out)
wolfSSL 7:481bce714567 46 {
wolfSSL 7:481bce714567 47 return Sha512Final_fips(sha, out);
wolfSSL 7:481bce714567 48 }
wolfSSL 7:481bce714567 49
wolfSSL 7:481bce714567 50
wolfSSL 7:481bce714567 51 #if defined(WOLFSSL_SHA384) || defined(HAVE_AESGCM)
wolfSSL 7:481bce714567 52
wolfSSL 7:481bce714567 53 int wc_InitSha384(Sha384* sha)
wolfSSL 7:481bce714567 54 {
wolfSSL 7:481bce714567 55 return InitSha384_fips(sha);
wolfSSL 7:481bce714567 56 }
wolfSSL 7:481bce714567 57
wolfSSL 7:481bce714567 58
wolfSSL 7:481bce714567 59 int wc_Sha384Update(Sha384* sha, const byte* data, word32 len)
wolfSSL 7:481bce714567 60 {
wolfSSL 7:481bce714567 61 return Sha384Update_fips(sha, data, len);
wolfSSL 7:481bce714567 62 }
wolfSSL 7:481bce714567 63
wolfSSL 7:481bce714567 64
wolfSSL 7:481bce714567 65 int wc_Sha384Final(Sha384* sha, byte* out)
wolfSSL 7:481bce714567 66 {
wolfSSL 7:481bce714567 67 return Sha384Final_fips(sha, out);
wolfSSL 7:481bce714567 68 }
wolfSSL 7:481bce714567 69
wolfSSL 7:481bce714567 70
wolfSSL 7:481bce714567 71 #endif /* WOLFSSL_SHA384 */
wolfSSL 7:481bce714567 72 #else /* else build without using fips */
wolfSSL 7:481bce714567 73 #include <wolfssl/wolfcrypt/logging.h>
wolfSSL 7:481bce714567 74 #include <wolfssl/wolfcrypt/error-crypt.h>
wolfSSL 7:481bce714567 75
wolfSSL 7:481bce714567 76 #ifdef NO_INLINE
wolfSSL 7:481bce714567 77 #include <wolfssl/wolfcrypt/misc.h>
wolfSSL 7:481bce714567 78 #else
wolfSSL 7:481bce714567 79 #define WOLFSSL_MISC_INCLUDED
wolfSSL 7:481bce714567 80 #include <wolfcrypt/src/misc.c>
wolfSSL 7:481bce714567 81 #endif
wolfSSL 7:481bce714567 82
wolfSSL 7:481bce714567 83
wolfSSL 7:481bce714567 84 #if defined(USE_INTEL_SPEEDUP)
wolfSSL 7:481bce714567 85 #define HAVE_INTEL_AVX1
wolfSSL 7:481bce714567 86 #define HAVE_INTEL_AVX2
wolfSSL 7:481bce714567 87 #endif
wolfSSL 7:481bce714567 88
wolfSSL 7:481bce714567 89 #if defined(HAVE_INTEL_AVX1)
wolfSSL 7:481bce714567 90 /* #define DEBUG_XMM */
wolfSSL 7:481bce714567 91 #endif
wolfSSL 7:481bce714567 92
wolfSSL 7:481bce714567 93 #if defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 94 #define HAVE_INTEL_RORX
wolfSSL 7:481bce714567 95 /* #define DEBUG_YMM */
wolfSSL 7:481bce714567 96 #endif
wolfSSL 7:481bce714567 97
wolfSSL 7:481bce714567 98 /*****
wolfSSL 7:481bce714567 99 Intel AVX1/AVX2 Macro Control Structure
wolfSSL 7:481bce714567 100
wolfSSL 7:481bce714567 101 #if defined(HAVE_INteL_SPEEDUP)
wolfSSL 7:481bce714567 102 #define HAVE_INTEL_AVX1
wolfSSL 7:481bce714567 103 #define HAVE_INTEL_AVX2
wolfSSL 7:481bce714567 104 #endif
wolfSSL 7:481bce714567 105
wolfSSL 7:481bce714567 106 int InitSha512(Sha512* sha512) {
wolfSSL 7:481bce714567 107 Save/Recover XMM, YMM
wolfSSL 7:481bce714567 108 ...
wolfSSL 7:481bce714567 109
wolfSSL 7:481bce714567 110 Check Intel AVX cpuid flags
wolfSSL 7:481bce714567 111 }
wolfSSL 7:481bce714567 112
wolfSSL 7:481bce714567 113 #if defined(HAVE_INTEL_AVX1)|| defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 114 Transform_AVX1() ; # Function prototype
wolfSSL 7:481bce714567 115 Transform_AVX2() ; #
wolfSSL 7:481bce714567 116 #endif
wolfSSL 7:481bce714567 117
wolfSSL 7:481bce714567 118 _Transform() { # Native Transform Function body
wolfSSL 7:481bce714567 119
wolfSSL 7:481bce714567 120 }
wolfSSL 7:481bce714567 121
wolfSSL 7:481bce714567 122 int Sha512Update() {
wolfSSL 7:481bce714567 123 Save/Recover XMM, YMM
wolfSSL 7:481bce714567 124 ...
wolfSSL 7:481bce714567 125 }
wolfSSL 7:481bce714567 126
wolfSSL 7:481bce714567 127 int Sha512Final() {
wolfSSL 7:481bce714567 128 Save/Recover XMM, YMM
wolfSSL 7:481bce714567 129 ...
wolfSSL 7:481bce714567 130 }
wolfSSL 7:481bce714567 131
wolfSSL 7:481bce714567 132
wolfSSL 7:481bce714567 133 #if defined(HAVE_INTEL_AVX1)
wolfSSL 7:481bce714567 134
wolfSSL 7:481bce714567 135 XMM Instructions/INLINE asm Definitions
wolfSSL 7:481bce714567 136
wolfSSL 7:481bce714567 137 #endif
wolfSSL 7:481bce714567 138
wolfSSL 7:481bce714567 139 #if defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 140
wolfSSL 7:481bce714567 141 YMM Instructions/INLINE asm Definitions
wolfSSL 7:481bce714567 142
wolfSSL 7:481bce714567 143 #endif
wolfSSL 7:481bce714567 144
wolfSSL 7:481bce714567 145 #if defnied(HAVE_INTEL_AVX1)
wolfSSL 7:481bce714567 146
wolfSSL 7:481bce714567 147 int Transform_AVX1() {
wolfSSL 7:481bce714567 148 Stitched Message Sched/Round
wolfSSL 7:481bce714567 149 }
wolfSSL 7:481bce714567 150
wolfSSL 7:481bce714567 151 #endif
wolfSSL 7:481bce714567 152
wolfSSL 7:481bce714567 153 #if defnied(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 154
wolfSSL 7:481bce714567 155 int Transform_AVX2() {
wolfSSL 7:481bce714567 156 Stitched Message Sched/Round
wolfSSL 7:481bce714567 157 }
wolfSSL 7:481bce714567 158 #endif
wolfSSL 7:481bce714567 159
wolfSSL 7:481bce714567 160
wolfSSL 7:481bce714567 161 */
wolfSSL 7:481bce714567 162
wolfSSL 7:481bce714567 163 #if defined(HAVE_INTEL_AVX1) || defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 164
wolfSSL 7:481bce714567 165
wolfSSL 7:481bce714567 166 /* Each platform needs to query info type 1 from cpuid to see if aesni is
wolfSSL 7:481bce714567 167 * supported. Also, let's setup a macro for proper linkage w/o ABI conflicts
wolfSSL 7:481bce714567 168 */
wolfSSL 7:481bce714567 169
wolfSSL 7:481bce714567 170 #ifndef _MSC_VER
wolfSSL 7:481bce714567 171 #define cpuid(reg, leaf, sub)\
wolfSSL 7:481bce714567 172 __asm__ __volatile__ ("cpuid":\
wolfSSL 7:481bce714567 173 "=a" (reg[0]), "=b" (reg[1]), "=c" (reg[2]), "=d" (reg[3]) :\
wolfSSL 7:481bce714567 174 "a" (leaf), "c"(sub));
wolfSSL 7:481bce714567 175
wolfSSL 7:481bce714567 176 #define XASM_LINK(f) asm(f)
wolfSSL 7:481bce714567 177 #else
wolfSSL 7:481bce714567 178
wolfSSL 7:481bce714567 179 #include <intrin.h>
wolfSSL 7:481bce714567 180 #define cpuid(a,b) __cpuid((int*)a,b)
wolfSSL 7:481bce714567 181
wolfSSL 7:481bce714567 182 #define XASM_LINK(f)
wolfSSL 7:481bce714567 183
wolfSSL 7:481bce714567 184 #endif /* _MSC_VER */
wolfSSL 7:481bce714567 185
wolfSSL 7:481bce714567 186 #define EAX 0
wolfSSL 7:481bce714567 187 #define EBX 1
wolfSSL 7:481bce714567 188 #define ECX 2
wolfSSL 7:481bce714567 189 #define EDX 3
wolfSSL 7:481bce714567 190
wolfSSL 7:481bce714567 191 #define CPUID_AVX1 0x1
wolfSSL 7:481bce714567 192 #define CPUID_AVX2 0x2
wolfSSL 7:481bce714567 193 #define CPUID_RDRAND 0x4
wolfSSL 7:481bce714567 194 #define CPUID_RDSEED 0x8
wolfSSL 7:481bce714567 195 #define CPUID_BMI2 0x10 /* MULX, RORX */
wolfSSL 7:481bce714567 196
wolfSSL 7:481bce714567 197 #define IS_INTEL_AVX1 (cpuid_flags&CPUID_AVX1)
wolfSSL 7:481bce714567 198 #define IS_INTEL_AVX2 (cpuid_flags&CPUID_AVX2)
wolfSSL 7:481bce714567 199 #define IS_INTEL_BMI2 (cpuid_flags&CPUID_BMI2)
wolfSSL 7:481bce714567 200 #define IS_INTEL_RDRAND (cpuid_flags&CPUID_RDRAND)
wolfSSL 7:481bce714567 201 #define IS_INTEL_RDSEED (cpuid_flags&CPUID_RDSEED)
wolfSSL 7:481bce714567 202
wolfSSL 7:481bce714567 203 static word32 cpuid_check = 0 ;
wolfSSL 7:481bce714567 204 static word32 cpuid_flags = 0 ;
wolfSSL 7:481bce714567 205
wolfSSL 7:481bce714567 206 static word32 cpuid_flag(word32 leaf, word32 sub, word32 num, word32 bit) {
wolfSSL 7:481bce714567 207 int got_intel_cpu=0;
wolfSSL 7:481bce714567 208 unsigned int reg[5];
wolfSSL 7:481bce714567 209
wolfSSL 7:481bce714567 210 reg[4] = '\0' ;
wolfSSL 7:481bce714567 211 cpuid(reg, 0, 0);
wolfSSL 7:481bce714567 212 if(XMEMCMP((char *)&(reg[EBX]), "Genu", 4) == 0 &&
wolfSSL 7:481bce714567 213 XMEMCMP((char *)&(reg[EDX]), "ineI", 4) == 0 &&
wolfSSL 7:481bce714567 214 XMEMCMP((char *)&(reg[ECX]), "ntel", 4) == 0) {
wolfSSL 7:481bce714567 215 got_intel_cpu = 1;
wolfSSL 7:481bce714567 216 }
wolfSSL 7:481bce714567 217 if (got_intel_cpu) {
wolfSSL 7:481bce714567 218 cpuid(reg, leaf, sub);
wolfSSL 7:481bce714567 219 return((reg[num]>>bit)&0x1) ;
wolfSSL 7:481bce714567 220 }
wolfSSL 7:481bce714567 221 return 0 ;
wolfSSL 7:481bce714567 222 }
wolfSSL 7:481bce714567 223
wolfSSL 7:481bce714567 224
wolfSSL 7:481bce714567 225 static int set_cpuid_flags() {
wolfSSL 7:481bce714567 226 if(cpuid_check ==0) {
wolfSSL 7:481bce714567 227 if(cpuid_flag(1, 0, ECX, 28)){ cpuid_flags |= CPUID_AVX1 ;}
wolfSSL 7:481bce714567 228 if(cpuid_flag(7, 0, EBX, 5)){ cpuid_flags |= CPUID_AVX2 ; }
wolfSSL 7:481bce714567 229 if(cpuid_flag(7, 0, EBX, 8)) { cpuid_flags |= CPUID_BMI2 ; }
wolfSSL 7:481bce714567 230 if(cpuid_flag(1, 0, ECX, 30)){ cpuid_flags |= CPUID_RDRAND ; }
wolfSSL 7:481bce714567 231 if(cpuid_flag(7, 0, EBX, 18)){ cpuid_flags |= CPUID_RDSEED ; }
wolfSSL 7:481bce714567 232 cpuid_check = 1 ;
wolfSSL 7:481bce714567 233 return 0 ;
wolfSSL 7:481bce714567 234 }
wolfSSL 7:481bce714567 235 return 1 ;
wolfSSL 7:481bce714567 236 }
wolfSSL 7:481bce714567 237
wolfSSL 7:481bce714567 238
wolfSSL 7:481bce714567 239 /* #if defined(HAVE_INTEL_AVX1/2) at the tail of sha512 */
wolfSSL 7:481bce714567 240
wolfSSL 7:481bce714567 241 #if defined(HAVE_INTEL_AVX1)
wolfSSL 7:481bce714567 242 static int Transform_AVX1(Sha512 *sha512) ;
wolfSSL 7:481bce714567 243 #endif
wolfSSL 7:481bce714567 244
wolfSSL 7:481bce714567 245 #if defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 246 static int Transform_AVX2(Sha512 *sha512) ;
wolfSSL 7:481bce714567 247
wolfSSL 7:481bce714567 248 #if defined(HAVE_INTEL_AVX1) && defined(HAVE_INTEL_AVX2) && defined(HAVE_INTEL_RORX)
wolfSSL 7:481bce714567 249 static int Transform_AVX1_RORX(Sha512 *sha512) ;
wolfSSL 7:481bce714567 250 #endif
wolfSSL 7:481bce714567 251
wolfSSL 7:481bce714567 252 #endif
wolfSSL 7:481bce714567 253
wolfSSL 7:481bce714567 254 static int _Transform(Sha512 *sha512) ;
wolfSSL 7:481bce714567 255
wolfSSL 7:481bce714567 256 static int (*Transform_p)(Sha512* sha512) = _Transform ;
wolfSSL 7:481bce714567 257
wolfSSL 7:481bce714567 258 #define Transform(sha512) (*Transform_p)(sha512)
wolfSSL 7:481bce714567 259
wolfSSL 7:481bce714567 260 static void set_Transform(void) {
wolfSSL 7:481bce714567 261 if(set_cpuid_flags()) return ;
wolfSSL 7:481bce714567 262
wolfSSL 7:481bce714567 263 #if defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 264 if(IS_INTEL_AVX2 && IS_INTEL_BMI2){
wolfSSL 7:481bce714567 265 Transform_p = Transform_AVX1_RORX; return ;
wolfSSL 7:481bce714567 266 Transform_p = Transform_AVX2 ;
wolfSSL 7:481bce714567 267 /* for avoiding warning,"not used" */
wolfSSL 7:481bce714567 268 }
wolfSSL 7:481bce714567 269 #endif
wolfSSL 7:481bce714567 270 #if defined(HAVE_INTEL_AVX1)
wolfSSL 7:481bce714567 271 Transform_p = ((IS_INTEL_AVX1) ? Transform_AVX1 : _Transform) ; return ;
wolfSSL 7:481bce714567 272 #endif
wolfSSL 7:481bce714567 273 Transform_p = _Transform ; return ;
wolfSSL 7:481bce714567 274 }
wolfSSL 7:481bce714567 275
wolfSSL 7:481bce714567 276 #else
wolfSSL 7:481bce714567 277 #define Transform(sha512) _Transform(sha512)
wolfSSL 7:481bce714567 278 #endif
wolfSSL 7:481bce714567 279
wolfSSL 7:481bce714567 280 /* Dummy for saving MM_REGs on behalf of Transform */
wolfSSL 7:481bce714567 281 /* #if defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 282 #define SAVE_XMM_YMM __asm__ volatile("orq %%r8, %%r8":::\
wolfSSL 7:481bce714567 283 "%ymm0","%ymm1","%ymm2","%ymm3","%ymm4","%ymm5","%ymm6","%ymm7","%ymm8","%ymm9","%ymm10","%ymm11",\
wolfSSL 7:481bce714567 284 "%ymm12","%ymm13","%ymm14","%ymm15")
wolfSSL 7:481bce714567 285 */
wolfSSL 7:481bce714567 286 #if defined(HAVE_INTEL_AVX1)
wolfSSL 7:481bce714567 287 #define SAVE_XMM_YMM __asm__ volatile("orq %%r8, %%r8":::\
wolfSSL 7:481bce714567 288 "xmm0","xmm1","xmm2","xmm3","xmm4","xmm5","xmm6","xmm7","xmm8","xmm9","xmm10","xmm11","xmm12","xmm13","xmm14","xmm15")
wolfSSL 7:481bce714567 289 #else
wolfSSL 7:481bce714567 290 #define SAVE_XMM_YMM
wolfSSL 7:481bce714567 291 #endif
wolfSSL 7:481bce714567 292
wolfSSL 7:481bce714567 293 #if defined(HAVE_INTEL_AVX1)|| defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 294
wolfSSL 7:481bce714567 295 #include <string.h>
wolfSSL 7:481bce714567 296
wolfSSL 7:481bce714567 297 #endif /* defined(HAVE_INTEL_AVX1)|| defined(HAVE_INTEL_AVX2) */
wolfSSL 7:481bce714567 298
wolfSSL 7:481bce714567 299
wolfSSL 7:481bce714567 300 #if defined(HAVE_INTEL_RORX)
wolfSSL 7:481bce714567 301 #define ROTR(func, bits, x) \
wolfSSL 7:481bce714567 302 word64 func(word64 x) { word64 ret ;\
wolfSSL 7:481bce714567 303 __asm__ ("rorx $"#bits", %1, %0\n\t":"=r"(ret):"r"(x):) ;\
wolfSSL 7:481bce714567 304 return ret ;\
wolfSSL 7:481bce714567 305 }
wolfSSL 7:481bce714567 306
wolfSSL 7:481bce714567 307 static INLINE ROTR(rotrFixed64_28, 28, x)
wolfSSL 7:481bce714567 308 static INLINE ROTR(rotrFixed64_34, 34, x)
wolfSSL 7:481bce714567 309 static INLINE ROTR(rotrFixed64_39, 39, x)
wolfSSL 7:481bce714567 310 static INLINE ROTR(rotrFixed64_14, 14, x)
wolfSSL 7:481bce714567 311 static INLINE ROTR(rotrFixed64_18, 18, x)
wolfSSL 7:481bce714567 312 static INLINE ROTR(rotrFixed64_41, 41, x)
wolfSSL 7:481bce714567 313
wolfSSL 7:481bce714567 314 #define S0_RORX(x) (rotrFixed64_28(x)^rotrFixed64_34(x)^rotrFixed64_39(x))
wolfSSL 7:481bce714567 315 #define S1_RORX(x) (rotrFixed64_14(x)^rotrFixed64_18(x)^rotrFixed64_41(x))
wolfSSL 7:481bce714567 316 #endif
wolfSSL 7:481bce714567 317
wolfSSL 7:481bce714567 318 #if defined(HAVE_BYTEREVERSE64) && !defined(HAVE_INTEL_AVX1) && !defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 319 #define ByteReverseWords64(out, in, size) ByteReverseWords64_1(out, size)
wolfSSL 7:481bce714567 320 #define ByteReverseWords64_1(buf, size)\
wolfSSL 7:481bce714567 321 { unsigned int i ;\
wolfSSL 7:481bce714567 322 for(i=0; i< size/sizeof(word64); i++){\
wolfSSL 7:481bce714567 323 __asm__ volatile("bswapq %0":"+r"(buf[i])::) ;\
wolfSSL 7:481bce714567 324 }\
wolfSSL 7:481bce714567 325 }
wolfSSL 7:481bce714567 326 #endif
wolfSSL 7:481bce714567 327
wolfSSL 7:481bce714567 328
wolfSSL 7:481bce714567 329 int wc_InitSha512(Sha512* sha512)
wolfSSL 7:481bce714567 330 {
wolfSSL 7:481bce714567 331 sha512->digest[0] = W64LIT(0x6a09e667f3bcc908);
wolfSSL 7:481bce714567 332 sha512->digest[1] = W64LIT(0xbb67ae8584caa73b);
wolfSSL 7:481bce714567 333 sha512->digest[2] = W64LIT(0x3c6ef372fe94f82b);
wolfSSL 7:481bce714567 334 sha512->digest[3] = W64LIT(0xa54ff53a5f1d36f1);
wolfSSL 7:481bce714567 335 sha512->digest[4] = W64LIT(0x510e527fade682d1);
wolfSSL 7:481bce714567 336 sha512->digest[5] = W64LIT(0x9b05688c2b3e6c1f);
wolfSSL 7:481bce714567 337 sha512->digest[6] = W64LIT(0x1f83d9abfb41bd6b);
wolfSSL 7:481bce714567 338 sha512->digest[7] = W64LIT(0x5be0cd19137e2179);
wolfSSL 7:481bce714567 339
wolfSSL 7:481bce714567 340 sha512->buffLen = 0;
wolfSSL 7:481bce714567 341 sha512->loLen = 0;
wolfSSL 7:481bce714567 342 sha512->hiLen = 0;
wolfSSL 7:481bce714567 343
wolfSSL 7:481bce714567 344 #if defined(HAVE_INTEL_AVX1)|| defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 345 set_Transform() ; /* choose best Transform function under this runtime environment */
wolfSSL 7:481bce714567 346 #endif
wolfSSL 7:481bce714567 347
wolfSSL 7:481bce714567 348 return 0 ;
wolfSSL 7:481bce714567 349 }
wolfSSL 7:481bce714567 350
wolfSSL 7:481bce714567 351
wolfSSL 7:481bce714567 352 static const word64 K512[80] = {
wolfSSL 7:481bce714567 353 W64LIT(0x428a2f98d728ae22), W64LIT(0x7137449123ef65cd),
wolfSSL 7:481bce714567 354 W64LIT(0xb5c0fbcfec4d3b2f), W64LIT(0xe9b5dba58189dbbc),
wolfSSL 7:481bce714567 355 W64LIT(0x3956c25bf348b538), W64LIT(0x59f111f1b605d019),
wolfSSL 7:481bce714567 356 W64LIT(0x923f82a4af194f9b), W64LIT(0xab1c5ed5da6d8118),
wolfSSL 7:481bce714567 357 W64LIT(0xd807aa98a3030242), W64LIT(0x12835b0145706fbe),
wolfSSL 7:481bce714567 358 W64LIT(0x243185be4ee4b28c), W64LIT(0x550c7dc3d5ffb4e2),
wolfSSL 7:481bce714567 359 W64LIT(0x72be5d74f27b896f), W64LIT(0x80deb1fe3b1696b1),
wolfSSL 7:481bce714567 360 W64LIT(0x9bdc06a725c71235), W64LIT(0xc19bf174cf692694),
wolfSSL 7:481bce714567 361 W64LIT(0xe49b69c19ef14ad2), W64LIT(0xefbe4786384f25e3),
wolfSSL 7:481bce714567 362 W64LIT(0x0fc19dc68b8cd5b5), W64LIT(0x240ca1cc77ac9c65),
wolfSSL 7:481bce714567 363 W64LIT(0x2de92c6f592b0275), W64LIT(0x4a7484aa6ea6e483),
wolfSSL 7:481bce714567 364 W64LIT(0x5cb0a9dcbd41fbd4), W64LIT(0x76f988da831153b5),
wolfSSL 7:481bce714567 365 W64LIT(0x983e5152ee66dfab), W64LIT(0xa831c66d2db43210),
wolfSSL 7:481bce714567 366 W64LIT(0xb00327c898fb213f), W64LIT(0xbf597fc7beef0ee4),
wolfSSL 7:481bce714567 367 W64LIT(0xc6e00bf33da88fc2), W64LIT(0xd5a79147930aa725),
wolfSSL 7:481bce714567 368 W64LIT(0x06ca6351e003826f), W64LIT(0x142929670a0e6e70),
wolfSSL 7:481bce714567 369 W64LIT(0x27b70a8546d22ffc), W64LIT(0x2e1b21385c26c926),
wolfSSL 7:481bce714567 370 W64LIT(0x4d2c6dfc5ac42aed), W64LIT(0x53380d139d95b3df),
wolfSSL 7:481bce714567 371 W64LIT(0x650a73548baf63de), W64LIT(0x766a0abb3c77b2a8),
wolfSSL 7:481bce714567 372 W64LIT(0x81c2c92e47edaee6), W64LIT(0x92722c851482353b),
wolfSSL 7:481bce714567 373 W64LIT(0xa2bfe8a14cf10364), W64LIT(0xa81a664bbc423001),
wolfSSL 7:481bce714567 374 W64LIT(0xc24b8b70d0f89791), W64LIT(0xc76c51a30654be30),
wolfSSL 7:481bce714567 375 W64LIT(0xd192e819d6ef5218), W64LIT(0xd69906245565a910),
wolfSSL 7:481bce714567 376 W64LIT(0xf40e35855771202a), W64LIT(0x106aa07032bbd1b8),
wolfSSL 7:481bce714567 377 W64LIT(0x19a4c116b8d2d0c8), W64LIT(0x1e376c085141ab53),
wolfSSL 7:481bce714567 378 W64LIT(0x2748774cdf8eeb99), W64LIT(0x34b0bcb5e19b48a8),
wolfSSL 7:481bce714567 379 W64LIT(0x391c0cb3c5c95a63), W64LIT(0x4ed8aa4ae3418acb),
wolfSSL 7:481bce714567 380 W64LIT(0x5b9cca4f7763e373), W64LIT(0x682e6ff3d6b2b8a3),
wolfSSL 7:481bce714567 381 W64LIT(0x748f82ee5defb2fc), W64LIT(0x78a5636f43172f60),
wolfSSL 7:481bce714567 382 W64LIT(0x84c87814a1f0ab72), W64LIT(0x8cc702081a6439ec),
wolfSSL 7:481bce714567 383 W64LIT(0x90befffa23631e28), W64LIT(0xa4506cebde82bde9),
wolfSSL 7:481bce714567 384 W64LIT(0xbef9a3f7b2c67915), W64LIT(0xc67178f2e372532b),
wolfSSL 7:481bce714567 385 W64LIT(0xca273eceea26619c), W64LIT(0xd186b8c721c0c207),
wolfSSL 7:481bce714567 386 W64LIT(0xeada7dd6cde0eb1e), W64LIT(0xf57d4f7fee6ed178),
wolfSSL 7:481bce714567 387 W64LIT(0x06f067aa72176fba), W64LIT(0x0a637dc5a2c898a6),
wolfSSL 7:481bce714567 388 W64LIT(0x113f9804bef90dae), W64LIT(0x1b710b35131c471b),
wolfSSL 7:481bce714567 389 W64LIT(0x28db77f523047d84), W64LIT(0x32caab7b40c72493),
wolfSSL 7:481bce714567 390 W64LIT(0x3c9ebe0a15c9bebc), W64LIT(0x431d67c49c100d4c),
wolfSSL 7:481bce714567 391 W64LIT(0x4cc5d4becb3e42b6), W64LIT(0x597f299cfc657e2a),
wolfSSL 7:481bce714567 392 W64LIT(0x5fcb6fab3ad6faec), W64LIT(0x6c44198c4a475817)
wolfSSL 7:481bce714567 393 };
wolfSSL 7:481bce714567 394
wolfSSL 7:481bce714567 395
wolfSSL 7:481bce714567 396
wolfSSL 7:481bce714567 397 #define blk0(i) (W[i] = sha512->buffer[i])
wolfSSL 7:481bce714567 398
wolfSSL 7:481bce714567 399 #define blk2(i) (W[i&15]+=s1(W[(i-2)&15])+W[(i-7)&15]+s0(W[(i-15)&15]))
wolfSSL 7:481bce714567 400
wolfSSL 7:481bce714567 401 #define Ch(x,y,z) (z^(x&(y^z)))
wolfSSL 7:481bce714567 402 #define Maj(x,y,z) ((x&y)|(z&(x|y)))
wolfSSL 7:481bce714567 403
wolfSSL 7:481bce714567 404 #define a(i) T[(0-i)&7]
wolfSSL 7:481bce714567 405 #define b(i) T[(1-i)&7]
wolfSSL 7:481bce714567 406 #define c(i) T[(2-i)&7]
wolfSSL 7:481bce714567 407 #define d(i) T[(3-i)&7]
wolfSSL 7:481bce714567 408 #define e(i) T[(4-i)&7]
wolfSSL 7:481bce714567 409 #define f(i) T[(5-i)&7]
wolfSSL 7:481bce714567 410 #define g(i) T[(6-i)&7]
wolfSSL 7:481bce714567 411 #define h(i) T[(7-i)&7]
wolfSSL 7:481bce714567 412
wolfSSL 7:481bce714567 413 #define S0(x) (rotrFixed64(x,28)^rotrFixed64(x,34)^rotrFixed64(x,39))
wolfSSL 7:481bce714567 414 #define S1(x) (rotrFixed64(x,14)^rotrFixed64(x,18)^rotrFixed64(x,41))
wolfSSL 7:481bce714567 415 #define s0(x) (rotrFixed64(x,1)^rotrFixed64(x,8)^(x>>7))
wolfSSL 7:481bce714567 416 #define s1(x) (rotrFixed64(x,19)^rotrFixed64(x,61)^(x>>6))
wolfSSL 7:481bce714567 417
wolfSSL 7:481bce714567 418 #define R(i) h(i)+=S1(e(i))+Ch(e(i),f(i),g(i))+K[i+j]+(j?blk2(i):blk0(i));\
wolfSSL 7:481bce714567 419 d(i)+=h(i);h(i)+=S0(a(i))+Maj(a(i),b(i),c(i))
wolfSSL 7:481bce714567 420
wolfSSL 7:481bce714567 421 static int _Transform(Sha512* sha512)
wolfSSL 7:481bce714567 422 {
wolfSSL 7:481bce714567 423 const word64* K = K512;
wolfSSL 7:481bce714567 424
wolfSSL 7:481bce714567 425 word32 j;
wolfSSL 7:481bce714567 426 word64 T[8];
wolfSSL 7:481bce714567 427
wolfSSL 7:481bce714567 428
wolfSSL 7:481bce714567 429 #ifdef WOLFSSL_SMALL_STACK
wolfSSL 7:481bce714567 430 word64* W;
wolfSSL 7:481bce714567 431 W = (word64*) XMALLOC(sizeof(word64) * 16, NULL, DYNAMIC_TYPE_TMP_BUFFER);
wolfSSL 7:481bce714567 432 if (W == NULL)
wolfSSL 7:481bce714567 433 return MEMORY_E;
wolfSSL 7:481bce714567 434 #else
wolfSSL 7:481bce714567 435 word64 W[16];
wolfSSL 7:481bce714567 436 #endif
wolfSSL 7:481bce714567 437
wolfSSL 7:481bce714567 438 /* Copy digest to working vars */
wolfSSL 7:481bce714567 439 XMEMCPY(T, sha512->digest, sizeof(T));
wolfSSL 7:481bce714567 440
wolfSSL 7:481bce714567 441 #ifdef USE_SLOW_SHA2
wolfSSL 7:481bce714567 442 /* over twice as small, but 50% slower */
wolfSSL 7:481bce714567 443 /* 80 operations, not unrolled */
wolfSSL 7:481bce714567 444 for (j = 0; j < 80; j += 16) {
wolfSSL 7:481bce714567 445 int m;
wolfSSL 7:481bce714567 446 for (m = 0; m < 16; m++) { /* braces needed here for macros {} */
wolfSSL 7:481bce714567 447 R(m);
wolfSSL 7:481bce714567 448 }
wolfSSL 7:481bce714567 449 }
wolfSSL 7:481bce714567 450 #else
wolfSSL 7:481bce714567 451 /* 80 operations, partially loop unrolled */
wolfSSL 7:481bce714567 452 for (j = 0; j < 80; j += 16) {
wolfSSL 7:481bce714567 453 R( 0); R( 1); R( 2); R( 3);
wolfSSL 7:481bce714567 454 R( 4); R( 5); R( 6); R( 7);
wolfSSL 7:481bce714567 455 R( 8); R( 9); R(10); R(11);
wolfSSL 7:481bce714567 456 R(12); R(13); R(14); R(15);
wolfSSL 7:481bce714567 457 }
wolfSSL 7:481bce714567 458 #endif /* USE_SLOW_SHA2 */
wolfSSL 7:481bce714567 459
wolfSSL 7:481bce714567 460 /* Add the working vars back into digest */
wolfSSL 7:481bce714567 461
wolfSSL 7:481bce714567 462 sha512->digest[0] += a(0);
wolfSSL 7:481bce714567 463 sha512->digest[1] += b(0);
wolfSSL 7:481bce714567 464 sha512->digest[2] += c(0);
wolfSSL 7:481bce714567 465 sha512->digest[3] += d(0);
wolfSSL 7:481bce714567 466 sha512->digest[4] += e(0);
wolfSSL 7:481bce714567 467 sha512->digest[5] += f(0);
wolfSSL 7:481bce714567 468 sha512->digest[6] += g(0);
wolfSSL 7:481bce714567 469 sha512->digest[7] += h(0);
wolfSSL 7:481bce714567 470
wolfSSL 7:481bce714567 471 /* Wipe variables */
wolfSSL 7:481bce714567 472 ForceZero(W, sizeof(word64) * 16);
wolfSSL 7:481bce714567 473 ForceZero(T, sizeof(T));
wolfSSL 7:481bce714567 474
wolfSSL 7:481bce714567 475 #ifdef WOLFSSL_SMALL_STACK
wolfSSL 7:481bce714567 476 XFREE(W, NULL, DYNAMIC_TYPE_TMP_BUFFER);
wolfSSL 7:481bce714567 477 #endif
wolfSSL 7:481bce714567 478
wolfSSL 7:481bce714567 479 return 0;
wolfSSL 7:481bce714567 480 }
wolfSSL 7:481bce714567 481
wolfSSL 7:481bce714567 482
wolfSSL 7:481bce714567 483 static INLINE void AddLength(Sha512* sha512, word32 len)
wolfSSL 7:481bce714567 484 {
wolfSSL 7:481bce714567 485 word64 tmp = sha512->loLen;
wolfSSL 7:481bce714567 486 if ( (sha512->loLen += len) < tmp)
wolfSSL 7:481bce714567 487 sha512->hiLen++; /* carry low to high */
wolfSSL 7:481bce714567 488 }
wolfSSL 7:481bce714567 489
wolfSSL 7:481bce714567 490 static INLINE int Sha512Update(Sha512* sha512, const byte* data, word32 len)
wolfSSL 7:481bce714567 491 {
wolfSSL 7:481bce714567 492 /* do block size increments */
wolfSSL 7:481bce714567 493 byte* local = (byte*)sha512->buffer;
wolfSSL 7:481bce714567 494 SAVE_XMM_YMM ; /* for Intel AVX */
wolfSSL 7:481bce714567 495
wolfSSL 7:481bce714567 496 while (len) {
wolfSSL 7:481bce714567 497 word32 add = min(len, SHA512_BLOCK_SIZE - sha512->buffLen);
wolfSSL 7:481bce714567 498 XMEMCPY(&local[sha512->buffLen], data, add);
wolfSSL 7:481bce714567 499
wolfSSL 7:481bce714567 500 sha512->buffLen += add;
wolfSSL 7:481bce714567 501 data += add;
wolfSSL 7:481bce714567 502 len -= add;
wolfSSL 7:481bce714567 503
wolfSSL 7:481bce714567 504 if (sha512->buffLen == SHA512_BLOCK_SIZE) {
wolfSSL 7:481bce714567 505 int ret;
wolfSSL 7:481bce714567 506 #if defined(LITTLE_ENDIAN_ORDER)
wolfSSL 7:481bce714567 507 #if defined(HAVE_INTEL_AVX1) || defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 508 if(!IS_INTEL_AVX1 && !IS_INTEL_AVX2)
wolfSSL 7:481bce714567 509 #endif
wolfSSL 7:481bce714567 510 ByteReverseWords64(sha512->buffer, sha512->buffer,
wolfSSL 7:481bce714567 511 SHA512_BLOCK_SIZE);
wolfSSL 7:481bce714567 512 #endif
wolfSSL 7:481bce714567 513 ret = Transform(sha512);
wolfSSL 7:481bce714567 514 if (ret != 0)
wolfSSL 7:481bce714567 515 return ret;
wolfSSL 7:481bce714567 516
wolfSSL 7:481bce714567 517 AddLength(sha512, SHA512_BLOCK_SIZE);
wolfSSL 7:481bce714567 518 sha512->buffLen = 0;
wolfSSL 7:481bce714567 519 }
wolfSSL 7:481bce714567 520 }
wolfSSL 7:481bce714567 521 return 0;
wolfSSL 7:481bce714567 522 }
wolfSSL 7:481bce714567 523
wolfSSL 7:481bce714567 524 int wc_Sha512Update(Sha512* sha512, const byte* data, word32 len)
wolfSSL 7:481bce714567 525 {
wolfSSL 7:481bce714567 526 return Sha512Update(sha512, data, len);
wolfSSL 7:481bce714567 527 }
wolfSSL 7:481bce714567 528
wolfSSL 7:481bce714567 529
wolfSSL 7:481bce714567 530 static INLINE int Sha512Final(Sha512* sha512)
wolfSSL 7:481bce714567 531 {
wolfSSL 7:481bce714567 532 byte* local = (byte*)sha512->buffer;
wolfSSL 7:481bce714567 533 int ret;
wolfSSL 7:481bce714567 534
wolfSSL 7:481bce714567 535 SAVE_XMM_YMM ; /* for Intel AVX */
wolfSSL 7:481bce714567 536 AddLength(sha512, sha512->buffLen); /* before adding pads */
wolfSSL 7:481bce714567 537
wolfSSL 7:481bce714567 538 local[sha512->buffLen++] = 0x80; /* add 1 */
wolfSSL 7:481bce714567 539
wolfSSL 7:481bce714567 540 /* pad with zeros */
wolfSSL 7:481bce714567 541 if (sha512->buffLen > SHA512_PAD_SIZE) {
wolfSSL 7:481bce714567 542 XMEMSET(&local[sha512->buffLen], 0, SHA512_BLOCK_SIZE -sha512->buffLen);
wolfSSL 7:481bce714567 543 sha512->buffLen += SHA512_BLOCK_SIZE - sha512->buffLen;
wolfSSL 7:481bce714567 544 #if defined(LITTLE_ENDIAN_ORDER)
wolfSSL 7:481bce714567 545 #if defined(HAVE_INTEL_AVX1) || defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 546 if(!IS_INTEL_AVX1 && !IS_INTEL_AVX2)
wolfSSL 7:481bce714567 547 #endif
wolfSSL 7:481bce714567 548 ByteReverseWords64(sha512->buffer,sha512->buffer,SHA512_BLOCK_SIZE);
wolfSSL 7:481bce714567 549 #endif
wolfSSL 7:481bce714567 550 ret = Transform(sha512);
wolfSSL 7:481bce714567 551 if (ret != 0)
wolfSSL 7:481bce714567 552 return ret;
wolfSSL 7:481bce714567 553
wolfSSL 7:481bce714567 554 sha512->buffLen = 0;
wolfSSL 7:481bce714567 555 }
wolfSSL 7:481bce714567 556 XMEMSET(&local[sha512->buffLen], 0, SHA512_PAD_SIZE - sha512->buffLen);
wolfSSL 7:481bce714567 557
wolfSSL 7:481bce714567 558 /* put lengths in bits */
wolfSSL 7:481bce714567 559 sha512->hiLen = (sha512->loLen >> (8*sizeof(sha512->loLen) - 3)) +
wolfSSL 7:481bce714567 560 (sha512->hiLen << 3);
wolfSSL 7:481bce714567 561 sha512->loLen = sha512->loLen << 3;
wolfSSL 7:481bce714567 562
wolfSSL 7:481bce714567 563 /* store lengths */
wolfSSL 7:481bce714567 564 #if defined(LITTLE_ENDIAN_ORDER)
wolfSSL 7:481bce714567 565 #if defined(HAVE_INTEL_AVX1) || defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 566 if(!IS_INTEL_AVX1 && !IS_INTEL_AVX2)
wolfSSL 7:481bce714567 567 #endif
wolfSSL 7:481bce714567 568 ByteReverseWords64(sha512->buffer, sha512->buffer, SHA512_PAD_SIZE);
wolfSSL 7:481bce714567 569 #endif
wolfSSL 7:481bce714567 570 /* ! length ordering dependent on digest endian type ! */
wolfSSL 7:481bce714567 571
wolfSSL 7:481bce714567 572 sha512->buffer[SHA512_BLOCK_SIZE / sizeof(word64) - 2] = sha512->hiLen;
wolfSSL 7:481bce714567 573 sha512->buffer[SHA512_BLOCK_SIZE / sizeof(word64) - 1] = sha512->loLen;
wolfSSL 7:481bce714567 574 #if defined(HAVE_INTEL_AVX1) || defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 575 if(IS_INTEL_AVX1 || IS_INTEL_AVX2)
wolfSSL 7:481bce714567 576 ByteReverseWords64(&(sha512->buffer[SHA512_BLOCK_SIZE / sizeof(word64) - 2]),
wolfSSL 7:481bce714567 577 &(sha512->buffer[SHA512_BLOCK_SIZE / sizeof(word64) - 2]),
wolfSSL 7:481bce714567 578 SHA512_BLOCK_SIZE - SHA512_PAD_SIZE);
wolfSSL 7:481bce714567 579 #endif
wolfSSL 7:481bce714567 580 ret = Transform(sha512);
wolfSSL 7:481bce714567 581 if (ret != 0)
wolfSSL 7:481bce714567 582 return ret;
wolfSSL 7:481bce714567 583
wolfSSL 7:481bce714567 584 #ifdef LITTLE_ENDIAN_ORDER
wolfSSL 7:481bce714567 585 ByteReverseWords64(sha512->digest, sha512->digest, SHA512_DIGEST_SIZE);
wolfSSL 7:481bce714567 586 #endif
wolfSSL 7:481bce714567 587
wolfSSL 7:481bce714567 588 return 0;
wolfSSL 7:481bce714567 589 }
wolfSSL 7:481bce714567 590
wolfSSL 7:481bce714567 591 int wc_Sha512Final(Sha512* sha512, byte* hash)
wolfSSL 7:481bce714567 592 {
wolfSSL 7:481bce714567 593 int ret = Sha512Final(sha512);
wolfSSL 7:481bce714567 594 if (ret != 0)
wolfSSL 7:481bce714567 595 return ret;
wolfSSL 7:481bce714567 596
wolfSSL 7:481bce714567 597 XMEMCPY(hash, sha512->digest, SHA512_DIGEST_SIZE);
wolfSSL 7:481bce714567 598
wolfSSL 7:481bce714567 599 return wc_InitSha512(sha512); /* reset state */
wolfSSL 7:481bce714567 600 }
wolfSSL 7:481bce714567 601
wolfSSL 7:481bce714567 602
wolfSSL 7:481bce714567 603 #if defined(HAVE_INTEL_AVX1)
wolfSSL 7:481bce714567 604
wolfSSL 7:481bce714567 605 #define Rx_1(i) h(i)+=S1(e(i))+Ch(e(i),f(i),g(i))+K[i+j] + W_X[i] ;
wolfSSL 7:481bce714567 606 #define Rx_2(i) d(i)+=h(i);
wolfSSL 7:481bce714567 607 #define Rx_3(i) h(i)+=S0(a(i))+Maj(a(i),b(i),c(i));
wolfSSL 7:481bce714567 608
wolfSSL 7:481bce714567 609 #if defined(HAVE_INTEL_RORX)
wolfSSL 7:481bce714567 610 #define Rx_RORX_1(i) h(i)+=S1_RORX(e(i))+Ch(e(i),f(i),g(i))+K[i+j] + W_X[i] ;
wolfSSL 7:481bce714567 611 #define Rx_RORX_2(i) d(i)+=h(i);
wolfSSL 7:481bce714567 612 #define Rx_RORX_3(i) h(i)+=S0_RORX(a(i))+Maj(a(i),b(i),c(i));
wolfSSL 7:481bce714567 613 #endif
wolfSSL 7:481bce714567 614
wolfSSL 7:481bce714567 615 #endif
wolfSSL 7:481bce714567 616
wolfSSL 7:481bce714567 617 #if defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 618 #define Ry_1(i, w) h(i)+=S1(e(i))+Ch(e(i),f(i),g(i))+K[i+j] + w ;
wolfSSL 7:481bce714567 619 #define Ry_2(i, w) d(i)+=h(i);
wolfSSL 7:481bce714567 620 #define Ry_3(i, w) h(i)+=S0(a(i))+Maj(a(i),b(i),c(i));
wolfSSL 7:481bce714567 621 #endif
wolfSSL 7:481bce714567 622
wolfSSL 7:481bce714567 623 #if defined(HAVE_INTEL_AVX1) /* INLINE Assember for Intel AVX1 instructions */
wolfSSL 7:481bce714567 624 #if defined(DEBUG_XMM)
wolfSSL 7:481bce714567 625
wolfSSL 7:481bce714567 626 #define SAVE_REG(i) __asm__ volatile("vmovdqu %%xmm"#i", %0 \n\t":"=m"(reg[i][0])::XMM_REGs);
wolfSSL 7:481bce714567 627 #define RECV_REG(i) __asm__ volatile("vmovdqu %0, %%xmm"#i" \n\t"::"m"(reg[i][0]):XMM_REGs);
wolfSSL 7:481bce714567 628
wolfSSL 7:481bce714567 629 #define _DUMP_REG(REG, name)\
wolfSSL 7:481bce714567 630 { word64 buf[16] ;word64 reg[16][2];int k ;\
wolfSSL 7:481bce714567 631 SAVE_REG(0); SAVE_REG(1); SAVE_REG(2); SAVE_REG(3); SAVE_REG(4); \
wolfSSL 7:481bce714567 632 SAVE_REG(5); SAVE_REG(6); SAVE_REG(7);SAVE_REG(8); SAVE_REG(9); SAVE_REG(10);\
wolfSSL 7:481bce714567 633 SAVE_REG(11); SAVE_REG(12); SAVE_REG(13); SAVE_REG(14); SAVE_REG(15); \
wolfSSL 7:481bce714567 634 __asm__ volatile("vmovdqu %%"#REG", %0 \n\t":"=m"(buf[0])::XMM_REGs);\
wolfSSL 7:481bce714567 635 printf(" "#name":\t") ; for(k=0; k<2; k++) printf("%016lx.", (word64)(buf[k])); printf("\n") ; \
wolfSSL 7:481bce714567 636 RECV_REG(0); RECV_REG(1); RECV_REG(2); RECV_REG(3); RECV_REG(4);\
wolfSSL 7:481bce714567 637 RECV_REG(5); RECV_REG(6); RECV_REG(7); RECV_REG(8); RECV_REG(9);\
wolfSSL 7:481bce714567 638 RECV_REG(10); RECV_REG(11); RECV_REG(12); RECV_REG(13); RECV_REG(14); RECV_REG(15);\
wolfSSL 7:481bce714567 639 }
wolfSSL 7:481bce714567 640
wolfSSL 7:481bce714567 641 #define DUMP_REG(REG) _DUMP_REG(REG, #REG)
wolfSSL 7:481bce714567 642 #define PRINTF(fmt, ...)
wolfSSL 7:481bce714567 643
wolfSSL 7:481bce714567 644 #else
wolfSSL 7:481bce714567 645
wolfSSL 7:481bce714567 646 #define DUMP_REG(REG)
wolfSSL 7:481bce714567 647 #define PRINTF(fmt, ...)
wolfSSL 7:481bce714567 648
wolfSSL 7:481bce714567 649 #endif
wolfSSL 7:481bce714567 650
wolfSSL 7:481bce714567 651 #define _MOVE_to_REG(xymm, mem) __asm__ volatile("vmovdqu %0, %%"#xymm" "\
wolfSSL 7:481bce714567 652 :: "m"(mem):XMM_REGs) ;
wolfSSL 7:481bce714567 653 #define _MOVE_to_MEM(mem,i, xymm) __asm__ volatile("vmovdqu %%"#xymm", %0" :\
wolfSSL 7:481bce714567 654 "=m"(mem[i]),"=m"(mem[i+1]),"=m"(mem[i+2]),"=m"(mem[i+3])::XMM_REGs) ;
wolfSSL 7:481bce714567 655 #define _MOVE(dest, src) __asm__ volatile("vmovdqu %%"#src", %%"\
wolfSSL 7:481bce714567 656 #dest" ":::XMM_REGs) ;
wolfSSL 7:481bce714567 657
wolfSSL 7:481bce714567 658 #define _S_TEMP(dest, src, bits, temp) __asm__ volatile("vpsrlq $"#bits", %%"\
wolfSSL 7:481bce714567 659 #src", %%"#dest"\n\tvpsllq $64-"#bits", %%"#src", %%"#temp"\n\tvpor %%"\
wolfSSL 7:481bce714567 660 #temp",%%"#dest", %%"#dest" ":::XMM_REGs) ;
wolfSSL 7:481bce714567 661 #define _AVX1_R(dest, src, bits) __asm__ volatile("vpsrlq $"#bits", %%"\
wolfSSL 7:481bce714567 662 #src", %%"#dest" ":::XMM_REGs) ;
wolfSSL 7:481bce714567 663 #define _XOR(dest, src1, src2) __asm__ volatile("vpxor %%"#src1", %%"\
wolfSSL 7:481bce714567 664 #src2", %%"#dest" ":::XMM_REGs) ;
wolfSSL 7:481bce714567 665 #define _OR(dest, src1, src2) __asm__ volatile("vpor %%"#src1", %%"\
wolfSSL 7:481bce714567 666 #src2", %%"#dest" ":::XMM_REGs) ;
wolfSSL 7:481bce714567 667 #define _ADD(dest, src1, src2) __asm__ volatile("vpaddq %%"#src1", %%"\
wolfSSL 7:481bce714567 668 #src2", %%"#dest" ":::XMM_REGs) ;
wolfSSL 7:481bce714567 669 #define _ADD_MEM(dest, src1, mem) __asm__ volatile("vpaddq %0, %%"#src1", %%"\
wolfSSL 7:481bce714567 670 #dest" "::"m"(mem):XMM_REGs) ;
wolfSSL 7:481bce714567 671
wolfSSL 7:481bce714567 672 #define MOVE_to_REG(xymm, mem) _MOVE_to_REG(xymm, mem)
wolfSSL 7:481bce714567 673 #define MOVE_to_MEM(mem, i, xymm) _MOVE_to_MEM(mem, i, xymm)
wolfSSL 7:481bce714567 674 #define MOVE(dest, src) _MOVE(dest, src)
wolfSSL 7:481bce714567 675
wolfSSL 7:481bce714567 676 #define XOR(dest, src1, src2) _XOR(dest, src1, src2)
wolfSSL 7:481bce714567 677 #define OR(dest, src1, src2) _OR(dest, src1, src2)
wolfSSL 7:481bce714567 678 #define ADD(dest, src1, src2) _ADD(dest, src1, src2)
wolfSSL 7:481bce714567 679
wolfSSL 7:481bce714567 680 #define S_TMP(dest, src, bits, temp) _S_TEMP(dest, src, bits, temp);
wolfSSL 7:481bce714567 681 #define AVX1_S(dest, src, bits) S_TMP(dest, src, bits, S_TEMP)
wolfSSL 7:481bce714567 682 #define AVX1_R(dest, src, bits) _AVX1_R(dest, src, bits)
wolfSSL 7:481bce714567 683
wolfSSL 7:481bce714567 684 #define Init_Mask(mask) \
wolfSSL 7:481bce714567 685 __asm__ volatile("vmovdqu %0, %%xmm1\n\t"::"m"(mask):"%xmm1") ;
wolfSSL 7:481bce714567 686
wolfSSL 7:481bce714567 687 #define _W_from_buff1(w, buff, xmm) \
wolfSSL 7:481bce714567 688 /* X0..3(xmm4..7), W[0..15] = sha512->buffer[0.15]; */\
wolfSSL 7:481bce714567 689 __asm__ volatile("vmovdqu %1, %%"#xmm"\n\t"\
wolfSSL 7:481bce714567 690 "vpshufb %%xmm1, %%"#xmm", %%"#xmm"\n\t"\
wolfSSL 7:481bce714567 691 "vmovdqu %%"#xmm", %0"\
wolfSSL 7:481bce714567 692 :"=m"(w): "m"(buff):"%xmm0") ;
wolfSSL 7:481bce714567 693
wolfSSL 7:481bce714567 694 #define W_from_buff1(w, buff, xmm) _W_from_buff1(w, buff, xmm)
wolfSSL 7:481bce714567 695
wolfSSL 7:481bce714567 696 #define W_from_buff(w, buff)\
wolfSSL 7:481bce714567 697 Init_Mask(mBYTE_FLIP_MASK[0]) ;\
wolfSSL 7:481bce714567 698 W_from_buff1(w[0], buff[0], W_0);\
wolfSSL 7:481bce714567 699 W_from_buff1(w[2], buff[2], W_2);\
wolfSSL 7:481bce714567 700 W_from_buff1(w[4], buff[4], W_4);\
wolfSSL 7:481bce714567 701 W_from_buff1(w[6], buff[6], W_6);\
wolfSSL 7:481bce714567 702 W_from_buff1(w[8], buff[8], W_8);\
wolfSSL 7:481bce714567 703 W_from_buff1(w[10],buff[10],W_10);\
wolfSSL 7:481bce714567 704 W_from_buff1(w[12],buff[12],W_12);\
wolfSSL 7:481bce714567 705 W_from_buff1(w[14],buff[14],W_14);
wolfSSL 7:481bce714567 706
wolfSSL 7:481bce714567 707 static word64 mBYTE_FLIP_MASK[] = { 0x0001020304050607, 0x08090a0b0c0d0e0f } ;
wolfSSL 7:481bce714567 708
wolfSSL 7:481bce714567 709 #define W_I_15 xmm14
wolfSSL 7:481bce714567 710 #define W_I_7 xmm11
wolfSSL 7:481bce714567 711 #define W_I_2 xmm13
wolfSSL 7:481bce714567 712 #define W_I xmm12
wolfSSL 7:481bce714567 713 #define G_TEMP xmm0
wolfSSL 7:481bce714567 714 #define S_TEMP xmm1
wolfSSL 7:481bce714567 715 #define XMM_TEMP0 xmm2
wolfSSL 7:481bce714567 716
wolfSSL 7:481bce714567 717 #define W_0 xmm12
wolfSSL 7:481bce714567 718 #define W_2 xmm3
wolfSSL 7:481bce714567 719 #define W_4 xmm4
wolfSSL 7:481bce714567 720 #define W_6 xmm5
wolfSSL 7:481bce714567 721 #define W_8 xmm6
wolfSSL 7:481bce714567 722 #define W_10 xmm7
wolfSSL 7:481bce714567 723 #define W_12 xmm8
wolfSSL 7:481bce714567 724 #define W_14 xmm9
wolfSSL 7:481bce714567 725
wolfSSL 7:481bce714567 726 #define XMM_REGs
wolfSSL 7:481bce714567 727
wolfSSL 7:481bce714567 728 #define s0_1(dest, src) AVX1_S(dest, src, 1);
wolfSSL 7:481bce714567 729 #define s0_2(dest, src) AVX1_S(G_TEMP, src, 8); XOR(dest, G_TEMP, dest) ;
wolfSSL 7:481bce714567 730 #define s0_3(dest, src) AVX1_R(G_TEMP, src, 7); XOR(dest, G_TEMP, dest) ;
wolfSSL 7:481bce714567 731
wolfSSL 7:481bce714567 732 #define s1_1(dest, src) AVX1_S(dest, src, 19);
wolfSSL 7:481bce714567 733 #define s1_2(dest, src) AVX1_S(G_TEMP, src, 61); XOR(dest, G_TEMP, dest) ;
wolfSSL 7:481bce714567 734 #define s1_3(dest, src) AVX1_R(G_TEMP, src, 6); XOR(dest, G_TEMP, dest) ;
wolfSSL 7:481bce714567 735
wolfSSL 7:481bce714567 736 #define s0_(dest, src) s0_1(dest, src) ; s0_2(dest, src) ; s0_3(dest, src)
wolfSSL 7:481bce714567 737 #define s1_(dest, src) s1_1(dest, src) ; s1_2(dest, src) ; s1_3(dest, src)
wolfSSL 7:481bce714567 738
wolfSSL 7:481bce714567 739 #define Block_xx_1(i) \
wolfSSL 7:481bce714567 740 MOVE_to_REG(W_I_15, W_X[(i-15)&15]) ;\
wolfSSL 7:481bce714567 741 MOVE_to_REG(W_I_7, W_X[(i- 7)&15]) ;\
wolfSSL 7:481bce714567 742
wolfSSL 7:481bce714567 743 #define Block_xx_2(i) \
wolfSSL 7:481bce714567 744 MOVE_to_REG(W_I_2, W_X[(i- 2)&15]) ;\
wolfSSL 7:481bce714567 745 MOVE_to_REG(W_I, W_X[(i)]) ;\
wolfSSL 7:481bce714567 746
wolfSSL 7:481bce714567 747 #define Block_xx_3(i) \
wolfSSL 7:481bce714567 748 s0_ (XMM_TEMP0, W_I_15) ;\
wolfSSL 7:481bce714567 749
wolfSSL 7:481bce714567 750 #define Block_xx_4(i) \
wolfSSL 7:481bce714567 751 ADD(W_I, W_I, XMM_TEMP0) ;\
wolfSSL 7:481bce714567 752 ADD(W_I, W_I, W_I_7) ;\
wolfSSL 7:481bce714567 753
wolfSSL 7:481bce714567 754 #define Block_xx_5(i) \
wolfSSL 7:481bce714567 755 s1_ (XMM_TEMP0, W_I_2) ;\
wolfSSL 7:481bce714567 756
wolfSSL 7:481bce714567 757 #define Block_xx_6(i) \
wolfSSL 7:481bce714567 758 ADD(W_I, W_I, XMM_TEMP0) ;\
wolfSSL 7:481bce714567 759 MOVE_to_MEM(W_X,i, W_I) ;\
wolfSSL 7:481bce714567 760 if(i==0)\
wolfSSL 7:481bce714567 761 MOVE_to_MEM(W_X,16, W_I) ;\
wolfSSL 7:481bce714567 762
wolfSSL 7:481bce714567 763 #define Block_xx_7(i) \
wolfSSL 7:481bce714567 764 MOVE_to_REG(W_I_15, W_X[(i-15)&15]) ;\
wolfSSL 7:481bce714567 765 MOVE_to_REG(W_I_7, W_X[(i- 7)&15]) ;\
wolfSSL 7:481bce714567 766
wolfSSL 7:481bce714567 767 #define Block_xx_8(i) \
wolfSSL 7:481bce714567 768 MOVE_to_REG(W_I_2, W_X[(i- 2)&15]) ;\
wolfSSL 7:481bce714567 769 MOVE_to_REG(W_I, W_X[(i)]) ;\
wolfSSL 7:481bce714567 770
wolfSSL 7:481bce714567 771 #define Block_xx_9(i) \
wolfSSL 7:481bce714567 772 s0_ (XMM_TEMP0, W_I_15) ;\
wolfSSL 7:481bce714567 773
wolfSSL 7:481bce714567 774 #define Block_xx_10(i) \
wolfSSL 7:481bce714567 775 ADD(W_I, W_I, XMM_TEMP0) ;\
wolfSSL 7:481bce714567 776 ADD(W_I, W_I, W_I_7) ;\
wolfSSL 7:481bce714567 777
wolfSSL 7:481bce714567 778 #define Block_xx_11(i) \
wolfSSL 7:481bce714567 779 s1_ (XMM_TEMP0, W_I_2) ;\
wolfSSL 7:481bce714567 780
wolfSSL 7:481bce714567 781 #define Block_xx_12(i) \
wolfSSL 7:481bce714567 782 ADD(W_I, W_I, XMM_TEMP0) ;\
wolfSSL 7:481bce714567 783 MOVE_to_MEM(W_X,i, W_I) ;\
wolfSSL 7:481bce714567 784 if((i)==0)\
wolfSSL 7:481bce714567 785 MOVE_to_MEM(W_X,16, W_I) ;\
wolfSSL 7:481bce714567 786
wolfSSL 7:481bce714567 787 static INLINE void Block_0_1(word64 *W_X) { Block_xx_1(0) ; }
wolfSSL 7:481bce714567 788 static INLINE void Block_0_2(word64 *W_X) { Block_xx_2(0) ; }
wolfSSL 7:481bce714567 789 static INLINE void Block_0_3(void) { Block_xx_3(0) ; }
wolfSSL 7:481bce714567 790 static INLINE void Block_0_4(void) { Block_xx_4(0) ; }
wolfSSL 7:481bce714567 791 static INLINE void Block_0_5(void) { Block_xx_5(0) ; }
wolfSSL 7:481bce714567 792 static INLINE void Block_0_6(word64 *W_X) { Block_xx_6(0) ; }
wolfSSL 7:481bce714567 793 static INLINE void Block_0_7(word64 *W_X) { Block_xx_7(2) ; }
wolfSSL 7:481bce714567 794 static INLINE void Block_0_8(word64 *W_X) { Block_xx_8(2) ; }
wolfSSL 7:481bce714567 795 static INLINE void Block_0_9(void) { Block_xx_9(2) ; }
wolfSSL 7:481bce714567 796 static INLINE void Block_0_10(void){ Block_xx_10(2) ; }
wolfSSL 7:481bce714567 797 static INLINE void Block_0_11(void){ Block_xx_11(2) ; }
wolfSSL 7:481bce714567 798 static INLINE void Block_0_12(word64 *W_X){ Block_xx_12(2) ; }
wolfSSL 7:481bce714567 799
wolfSSL 7:481bce714567 800 static INLINE void Block_4_1(word64 *W_X) { Block_xx_1(4) ; }
wolfSSL 7:481bce714567 801 static INLINE void Block_4_2(word64 *W_X) { Block_xx_2(4) ; }
wolfSSL 7:481bce714567 802 static INLINE void Block_4_3(void) { Block_xx_3(4) ; }
wolfSSL 7:481bce714567 803 static INLINE void Block_4_4(void) { Block_xx_4(4) ; }
wolfSSL 7:481bce714567 804 static INLINE void Block_4_5(void) { Block_xx_5(4) ; }
wolfSSL 7:481bce714567 805 static INLINE void Block_4_6(word64 *W_X) { Block_xx_6(4) ; }
wolfSSL 7:481bce714567 806 static INLINE void Block_4_7(word64 *W_X) { Block_xx_7(6) ; }
wolfSSL 7:481bce714567 807 static INLINE void Block_4_8(word64 *W_X) { Block_xx_8(6) ; }
wolfSSL 7:481bce714567 808 static INLINE void Block_4_9(void) { Block_xx_9(6) ; }
wolfSSL 7:481bce714567 809 static INLINE void Block_4_10(void){ Block_xx_10(6) ; }
wolfSSL 7:481bce714567 810 static INLINE void Block_4_11(void){ Block_xx_11(6) ; }
wolfSSL 7:481bce714567 811 static INLINE void Block_4_12(word64 *W_X){ Block_xx_12(6) ; }
wolfSSL 7:481bce714567 812
wolfSSL 7:481bce714567 813 static INLINE void Block_8_1(word64 *W_X) { Block_xx_1(8) ; }
wolfSSL 7:481bce714567 814 static INLINE void Block_8_2(word64 *W_X) { Block_xx_2(8) ; }
wolfSSL 7:481bce714567 815 static INLINE void Block_8_3(void) { Block_xx_3(8) ; }
wolfSSL 7:481bce714567 816 static INLINE void Block_8_4(void) { Block_xx_4(8) ; }
wolfSSL 7:481bce714567 817 static INLINE void Block_8_5(void) { Block_xx_5(8) ; }
wolfSSL 7:481bce714567 818 static INLINE void Block_8_6(word64 *W_X) { Block_xx_6(8) ; }
wolfSSL 7:481bce714567 819 static INLINE void Block_8_7(word64 *W_X) { Block_xx_7(10) ; }
wolfSSL 7:481bce714567 820 static INLINE void Block_8_8(word64 *W_X) { Block_xx_8(10) ; }
wolfSSL 7:481bce714567 821 static INLINE void Block_8_9(void) { Block_xx_9(10) ; }
wolfSSL 7:481bce714567 822 static INLINE void Block_8_10(void){ Block_xx_10(10) ; }
wolfSSL 7:481bce714567 823 static INLINE void Block_8_11(void){ Block_xx_11(10) ; }
wolfSSL 7:481bce714567 824 static INLINE void Block_8_12(word64 *W_X){ Block_xx_12(10) ; }
wolfSSL 7:481bce714567 825
wolfSSL 7:481bce714567 826 static INLINE void Block_12_1(word64 *W_X) { Block_xx_1(12) ; }
wolfSSL 7:481bce714567 827 static INLINE void Block_12_2(word64 *W_X) { Block_xx_2(12) ; }
wolfSSL 7:481bce714567 828 static INLINE void Block_12_3(void) { Block_xx_3(12) ; }
wolfSSL 7:481bce714567 829 static INLINE void Block_12_4(void) { Block_xx_4(12) ; }
wolfSSL 7:481bce714567 830 static INLINE void Block_12_5(void) { Block_xx_5(12) ; }
wolfSSL 7:481bce714567 831 static INLINE void Block_12_6(word64 *W_X) { Block_xx_6(12) ; }
wolfSSL 7:481bce714567 832 static INLINE void Block_12_7(word64 *W_X) { Block_xx_7(14) ; }
wolfSSL 7:481bce714567 833 static INLINE void Block_12_8(word64 *W_X) { Block_xx_8(14) ; }
wolfSSL 7:481bce714567 834 static INLINE void Block_12_9(void) { Block_xx_9(14) ; }
wolfSSL 7:481bce714567 835 static INLINE void Block_12_10(void){ Block_xx_10(14) ; }
wolfSSL 7:481bce714567 836 static INLINE void Block_12_11(void){ Block_xx_11(14) ; }
wolfSSL 7:481bce714567 837 static INLINE void Block_12_12(word64 *W_X){ Block_xx_12(14) ; }
wolfSSL 7:481bce714567 838
wolfSSL 7:481bce714567 839 #endif
wolfSSL 7:481bce714567 840
wolfSSL 7:481bce714567 841 #if defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 842 static const unsigned long mBYTE_FLIP_MASK_Y[] =
wolfSSL 7:481bce714567 843 { 0x0001020304050607, 0x08090a0b0c0d0e0f, 0x0001020304050607, 0x08090a0b0c0d0e0f } ;
wolfSSL 7:481bce714567 844
wolfSSL 7:481bce714567 845 #define W_from_buff_Y(buff)\
wolfSSL 7:481bce714567 846 { /* X0..3(ymm9..12), W_X[0..15] = sha512->buffer[0.15]; */\
wolfSSL 7:481bce714567 847 __asm__ volatile("vmovdqu %0, %%ymm8\n\t"::"m"(mBYTE_FLIP_MASK_Y[0]):YMM_REGs) ;\
wolfSSL 7:481bce714567 848 __asm__ volatile("vmovdqu %0, %%ymm12\n\t"\
wolfSSL 7:481bce714567 849 "vmovdqu %1, %%ymm4\n\t"\
wolfSSL 7:481bce714567 850 "vpshufb %%ymm8, %%ymm12, %%ymm12\n\t"\
wolfSSL 7:481bce714567 851 "vpshufb %%ymm8, %%ymm4, %%ymm4\n\t"\
wolfSSL 7:481bce714567 852 :: "m"(buff[0]), "m"(buff[4]):YMM_REGs) ;\
wolfSSL 7:481bce714567 853 __asm__ volatile("vmovdqu %0, %%ymm5\n\t"\
wolfSSL 7:481bce714567 854 "vmovdqu %1, %%ymm6\n\t"\
wolfSSL 7:481bce714567 855 "vpshufb %%ymm8, %%ymm5, %%ymm5\n\t"\
wolfSSL 7:481bce714567 856 "vpshufb %%ymm8, %%ymm6, %%ymm6\n\t"\
wolfSSL 7:481bce714567 857 :: "m"(buff[8]), "m"(buff[12]):YMM_REGs) ;\
wolfSSL 7:481bce714567 858 }
wolfSSL 7:481bce714567 859
wolfSSL 7:481bce714567 860 #if defined(DEBUG_YMM)
wolfSSL 7:481bce714567 861
wolfSSL 7:481bce714567 862 #define SAVE_REG_Y(i) __asm__ volatile("vmovdqu %%ymm"#i", %0 \n\t":"=m"(reg[i-4][0])::YMM_REGs);
wolfSSL 7:481bce714567 863 #define RECV_REG_Y(i) __asm__ volatile("vmovdqu %0, %%ymm"#i" \n\t"::"m"(reg[i-4][0]):YMM_REGs);
wolfSSL 7:481bce714567 864
wolfSSL 7:481bce714567 865 #define _DUMP_REG_Y(REG, name)\
wolfSSL 7:481bce714567 866 { word64 buf[16] ;word64 reg[16][2];int k ;\
wolfSSL 7:481bce714567 867 SAVE_REG_Y(4); SAVE_REG_Y(5); SAVE_REG_Y(6); SAVE_REG_Y(7); \
wolfSSL 7:481bce714567 868 SAVE_REG_Y(8); SAVE_REG_Y(9); SAVE_REG_Y(10); SAVE_REG_Y(11); SAVE_REG_Y(12);\
wolfSSL 7:481bce714567 869 SAVE_REG_Y(13); SAVE_REG_Y(14); SAVE_REG_Y(15); \
wolfSSL 7:481bce714567 870 __asm__ volatile("vmovdqu %%"#REG", %0 \n\t":"=m"(buf[0])::YMM_REGs);\
wolfSSL 7:481bce714567 871 printf(" "#name":\t") ; for(k=0; k<4; k++) printf("%016lx.", (word64)buf[k]) ; printf("\n") ; \
wolfSSL 7:481bce714567 872 RECV_REG_Y(4); RECV_REG_Y(5); RECV_REG_Y(6); RECV_REG_Y(7); \
wolfSSL 7:481bce714567 873 RECV_REG_Y(8); RECV_REG_Y(9); RECV_REG_Y(10); RECV_REG_Y(11); RECV_REG_Y(12); \
wolfSSL 7:481bce714567 874 RECV_REG_Y(13); RECV_REG_Y(14); RECV_REG_Y(15);\
wolfSSL 7:481bce714567 875 }
wolfSSL 7:481bce714567 876
wolfSSL 7:481bce714567 877 #define DUMP_REG_Y(REG) _DUMP_REG_Y(REG, #REG)
wolfSSL 7:481bce714567 878 #define DUMP_REG2_Y(REG) _DUMP_REG_Y(REG, #REG)
wolfSSL 7:481bce714567 879 #define PRINTF_Y(fmt, ...)
wolfSSL 7:481bce714567 880
wolfSSL 7:481bce714567 881 #else
wolfSSL 7:481bce714567 882
wolfSSL 7:481bce714567 883 #define DUMP_REG_Y(REG)
wolfSSL 7:481bce714567 884 #define DUMP_REG2_Y(REG)
wolfSSL 7:481bce714567 885 #define PRINTF_Y(fmt, ...)
wolfSSL 7:481bce714567 886
wolfSSL 7:481bce714567 887 #endif
wolfSSL 7:481bce714567 888
wolfSSL 7:481bce714567 889 #define _MOVE_to_REGy(ymm, mem) __asm__ volatile("vmovdqu %0, %%"#ymm" "\
wolfSSL 7:481bce714567 890 :: "m"(mem):YMM_REGs) ;
wolfSSL 7:481bce714567 891 #define _MOVE_to_MEMy(mem,i, ymm) __asm__ volatile("vmovdqu %%"#ymm", %0" \
wolfSSL 7:481bce714567 892 : "=m"(mem[i]),"=m"(mem[i+1]),"=m"(mem[i+2]),"=m"(mem[i+3])::YMM_REGs) ;
wolfSSL 7:481bce714567 893 #define _MOVE_128y(ymm0, ymm1, ymm2, map) __asm__ volatile("vperm2i128 $"\
wolfSSL 7:481bce714567 894 #map", %%"#ymm2", %%"#ymm1", %%"#ymm0" ":::YMM_REGs) ;
wolfSSL 7:481bce714567 895 #define _S_TEMPy(dest, src, bits, temp) \
wolfSSL 7:481bce714567 896 __asm__ volatile("vpsrlq $"#bits", %%"#src", %%"#dest"\n\tvpsllq $64-"#bits\
wolfSSL 7:481bce714567 897 ", %%"#src", %%"#temp"\n\tvpor %%"#temp",%%"#dest", %%"#dest" ":::YMM_REGs) ;
wolfSSL 7:481bce714567 898 #define _AVX2_R(dest, src, bits) __asm__ volatile("vpsrlq $"#bits", %%"\
wolfSSL 7:481bce714567 899 #src", %%"#dest" ":::YMM_REGs) ;
wolfSSL 7:481bce714567 900 #define _XORy(dest, src1, src2) __asm__ volatile("vpxor %%"#src1", %%"\
wolfSSL 7:481bce714567 901 #src2", %%"#dest" ":::YMM_REGs) ;
wolfSSL 7:481bce714567 902 #define _ADDy(dest, src1, src2) __asm__ volatile("vpaddq %%"#src1", %%"\
wolfSSL 7:481bce714567 903 #src2", %%"#dest" ":::YMM_REGs) ;
wolfSSL 7:481bce714567 904 #define _BLENDy(map, dest, src1, src2) __asm__ volatile("vpblendd $"#map", %%"\
wolfSSL 7:481bce714567 905 #src1", %%"#src2", %%"#dest" ":::YMM_REGs) ;
wolfSSL 7:481bce714567 906 #define _BLENDQy(map, dest, src1, src2) __asm__ volatile("vblendpd $"#map", %%"\
wolfSSL 7:481bce714567 907 #src1", %%"#src2", %%"#dest" ":::YMM_REGs) ;
wolfSSL 7:481bce714567 908 #define _PERMQy(map, dest, src) __asm__ volatile("vpermq $"#map", %%"\
wolfSSL 7:481bce714567 909 #src", %%"#dest" ":::YMM_REGs) ;
wolfSSL 7:481bce714567 910
wolfSSL 7:481bce714567 911 #define MOVE_to_REGy(ymm, mem) _MOVE_to_REGy(ymm, mem)
wolfSSL 7:481bce714567 912 #define MOVE_to_MEMy(mem, i, ymm) _MOVE_to_MEMy(mem, i, ymm)
wolfSSL 7:481bce714567 913
wolfSSL 7:481bce714567 914 #define MOVE_128y(ymm0, ymm1, ymm2, map) _MOVE_128y(ymm0, ymm1, ymm2, map)
wolfSSL 7:481bce714567 915 #define XORy(dest, src1, src2) _XORy(dest, src1, src2)
wolfSSL 7:481bce714567 916 #define ADDy(dest, src1, src2) _ADDy(dest, src1, src2)
wolfSSL 7:481bce714567 917 #define BLENDy(map, dest, src1, src2) _BLENDy(map, dest, src1, src2)
wolfSSL 7:481bce714567 918 #define BLENDQy(map, dest, src1, src2) _BLENDQy(map, dest, src1, src2)
wolfSSL 7:481bce714567 919 #define PERMQy(map, dest, src) _PERMQy(map, dest, src)
wolfSSL 7:481bce714567 920
wolfSSL 7:481bce714567 921
wolfSSL 7:481bce714567 922 #define S_TMPy(dest, src, bits, temp) _S_TEMPy(dest, src, bits, temp);
wolfSSL 7:481bce714567 923 #define AVX2_S(dest, src, bits) S_TMPy(dest, src, bits, S_TEMPy)
wolfSSL 7:481bce714567 924 #define AVX2_R(dest, src, bits) _AVX2_R(dest, src, bits)
wolfSSL 7:481bce714567 925
wolfSSL 7:481bce714567 926
wolfSSL 7:481bce714567 927 #define FEEDBACK1_to_W_I_2(w_i_2, w_i) MOVE_128y(YMM_TEMP0, w_i, w_i, 0x08) ;\
wolfSSL 7:481bce714567 928 BLENDy(0xf0, w_i_2, YMM_TEMP0, w_i_2) ;
wolfSSL 7:481bce714567 929
wolfSSL 7:481bce714567 930 #define MOVE_W_to_W_I_15(w_i_15, w_0, w_4) BLENDQy(0x1, w_i_15, w_4, w_0) ;\
wolfSSL 7:481bce714567 931 PERMQy(0x39, w_i_15, w_i_15) ;
wolfSSL 7:481bce714567 932 #define MOVE_W_to_W_I_7(w_i_7, w_8, w_12) BLENDQy(0x1, w_i_7, w_12, w_8) ;\
wolfSSL 7:481bce714567 933 PERMQy(0x39, w_i_7, w_i_7) ;
wolfSSL 7:481bce714567 934 #define MOVE_W_to_W_I_2(w_i_2, w_12) BLENDQy(0xc, w_i_2, w_12, w_i_2) ;\
wolfSSL 7:481bce714567 935 PERMQy(0x0e, w_i_2, w_i_2) ;
wolfSSL 7:481bce714567 936
wolfSSL 7:481bce714567 937
wolfSSL 7:481bce714567 938 #define W_I_16y ymm8
wolfSSL 7:481bce714567 939 #define W_I_15y ymm9
wolfSSL 7:481bce714567 940 #define W_I_7y ymm10
wolfSSL 7:481bce714567 941 #define W_I_2y ymm11
wolfSSL 7:481bce714567 942 #define W_Iy ymm12
wolfSSL 7:481bce714567 943 #define G_TEMPy ymm13
wolfSSL 7:481bce714567 944 #define S_TEMPy ymm14
wolfSSL 7:481bce714567 945 #define YMM_TEMP0 ymm15
wolfSSL 7:481bce714567 946 #define YMM_TEMP0x xmm15
wolfSSL 7:481bce714567 947 #define W_I_TEMPy ymm7
wolfSSL 7:481bce714567 948 #define W_K_TEMPy ymm15
wolfSSL 7:481bce714567 949 #define W_K_TEMPx xmm15
wolfSSL 7:481bce714567 950 #define W_0y ymm12
wolfSSL 7:481bce714567 951 #define W_4y ymm4
wolfSSL 7:481bce714567 952 #define W_8y ymm5
wolfSSL 7:481bce714567 953 #define W_12y ymm6
wolfSSL 7:481bce714567 954
wolfSSL 7:481bce714567 955 #define YMM_REGs
wolfSSL 7:481bce714567 956 /* Registers are saved in Sha512Update/Final */
wolfSSL 7:481bce714567 957 /* "%ymm7","%ymm8","%ymm9","%ymm10","%ymm11","%ymm12","%ymm13","%ymm14","%ymm15"*/
wolfSSL 7:481bce714567 958
wolfSSL 7:481bce714567 959 #define MOVE_15_to_16(w_i_16, w_i_15, w_i_7)\
wolfSSL 7:481bce714567 960 __asm__ volatile("vperm2i128 $0x01, %%"#w_i_15", %%"#w_i_15", %%"#w_i_15" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 961 __asm__ volatile("vpblendd $0x08, %%"#w_i_15", %%"#w_i_7", %%"#w_i_16" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 962 __asm__ volatile("vperm2i128 $0x01, %%"#w_i_7", %%"#w_i_7", %%"#w_i_15" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 963 __asm__ volatile("vpblendd $0x80, %%"#w_i_15", %%"#w_i_16", %%"#w_i_16" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 964 __asm__ volatile("vpshufd $0x93, %%"#w_i_16", %%"#w_i_16" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 965
wolfSSL 7:481bce714567 966 #define MOVE_7_to_15(w_i_15, w_i_7)\
wolfSSL 7:481bce714567 967 __asm__ volatile("vmovdqu %%"#w_i_7", %%"#w_i_15" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 968
wolfSSL 7:481bce714567 969 #define MOVE_I_to_7(w_i_7, w_i)\
wolfSSL 7:481bce714567 970 __asm__ volatile("vperm2i128 $0x01, %%"#w_i", %%"#w_i", %%"#w_i_7" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 971 __asm__ volatile("vpblendd $0x01, %%"#w_i_7", %%"#w_i", %%"#w_i_7" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 972 __asm__ volatile("vpshufd $0x39, %%"#w_i_7", %%"#w_i_7" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 973
wolfSSL 7:481bce714567 974 #define MOVE_I_to_2(w_i_2, w_i)\
wolfSSL 7:481bce714567 975 __asm__ volatile("vperm2i128 $0x01, %%"#w_i", %%"#w_i", %%"#w_i_2" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 976 __asm__ volatile("vpshufd $0x0e, %%"#w_i_2", %%"#w_i_2" ":::YMM_REGs) ;\
wolfSSL 7:481bce714567 977
wolfSSL 7:481bce714567 978 #endif
wolfSSL 7:481bce714567 979
wolfSSL 7:481bce714567 980
wolfSSL 7:481bce714567 981 /*** Transform Body ***/
wolfSSL 7:481bce714567 982 #if defined(HAVE_INTEL_AVX1)
wolfSSL 7:481bce714567 983
wolfSSL 7:481bce714567 984 static int Transform_AVX1(Sha512* sha512)
wolfSSL 7:481bce714567 985 {
wolfSSL 7:481bce714567 986 const word64* K = K512;
wolfSSL 7:481bce714567 987 word64 W_X[16+4];
wolfSSL 7:481bce714567 988 word32 j;
wolfSSL 7:481bce714567 989 word64 T[8];
wolfSSL 7:481bce714567 990 /* Copy digest to working vars */
wolfSSL 7:481bce714567 991 XMEMCPY(T, sha512->digest, sizeof(T));
wolfSSL 7:481bce714567 992
wolfSSL 7:481bce714567 993 W_from_buff(W_X, sha512->buffer) ;
wolfSSL 7:481bce714567 994 for (j = 0; j < 80; j += 16) {
wolfSSL 7:481bce714567 995 Rx_1( 0); Block_0_1(W_X); Rx_2( 0); Block_0_2(W_X); Rx_3( 0); Block_0_3();
wolfSSL 7:481bce714567 996 Rx_1( 1); Block_0_4(); Rx_2( 1); Block_0_5(); Rx_3( 1); Block_0_6(W_X);
wolfSSL 7:481bce714567 997 Rx_1( 2); Block_0_7(W_X); Rx_2( 2); Block_0_8(W_X); Rx_3( 2); Block_0_9();
wolfSSL 7:481bce714567 998 Rx_1( 3); Block_0_10();Rx_2( 3); Block_0_11();Rx_3( 3); Block_0_12(W_X);
wolfSSL 7:481bce714567 999
wolfSSL 7:481bce714567 1000 Rx_1( 4); Block_4_1(W_X); Rx_2( 4); Block_4_2(W_X); Rx_3( 4); Block_4_3();
wolfSSL 7:481bce714567 1001 Rx_1( 5); Block_4_4(); Rx_2( 5); Block_4_5(); Rx_3( 5); Block_4_6(W_X);
wolfSSL 7:481bce714567 1002 Rx_1( 6); Block_4_7(W_X); Rx_2( 6); Block_4_8(W_X); Rx_3( 6); Block_4_9();
wolfSSL 7:481bce714567 1003 Rx_1( 7); Block_4_10();Rx_2( 7); Block_4_11();Rx_3( 7); Block_4_12(W_X);
wolfSSL 7:481bce714567 1004
wolfSSL 7:481bce714567 1005 Rx_1( 8); Block_8_1(W_X); Rx_2( 8); Block_8_2(W_X); Rx_3( 8); Block_8_3();
wolfSSL 7:481bce714567 1006 Rx_1( 9); Block_8_4(); Rx_2( 9); Block_8_5(); Rx_3( 9); Block_8_6(W_X);
wolfSSL 7:481bce714567 1007 Rx_1(10); Block_8_7(W_X); Rx_2(10); Block_8_8(W_X); Rx_3(10); Block_8_9();
wolfSSL 7:481bce714567 1008 Rx_1(11); Block_8_10();Rx_2(11); Block_8_11();Rx_3(11); Block_8_12(W_X);
wolfSSL 7:481bce714567 1009
wolfSSL 7:481bce714567 1010 Rx_1(12); Block_12_1(W_X); Rx_2(12); Block_12_2(W_X); Rx_3(12); Block_12_3();
wolfSSL 7:481bce714567 1011 Rx_1(13); Block_12_4(); Rx_2(13); Block_12_5(); Rx_3(13); Block_12_6(W_X);
wolfSSL 7:481bce714567 1012 Rx_1(14); Block_12_7(W_X); Rx_2(14); Block_12_8(W_X); Rx_3(14); Block_12_9();
wolfSSL 7:481bce714567 1013 Rx_1(15); Block_12_10();Rx_2(15); Block_12_11();Rx_3(15); Block_12_12(W_X);
wolfSSL 7:481bce714567 1014 }
wolfSSL 7:481bce714567 1015
wolfSSL 7:481bce714567 1016 /* Add the working vars back into digest */
wolfSSL 7:481bce714567 1017
wolfSSL 7:481bce714567 1018 sha512->digest[0] += a(0);
wolfSSL 7:481bce714567 1019 sha512->digest[1] += b(0);
wolfSSL 7:481bce714567 1020 sha512->digest[2] += c(0);
wolfSSL 7:481bce714567 1021 sha512->digest[3] += d(0);
wolfSSL 7:481bce714567 1022 sha512->digest[4] += e(0);
wolfSSL 7:481bce714567 1023 sha512->digest[5] += f(0);
wolfSSL 7:481bce714567 1024 sha512->digest[6] += g(0);
wolfSSL 7:481bce714567 1025 sha512->digest[7] += h(0);
wolfSSL 7:481bce714567 1026
wolfSSL 7:481bce714567 1027 /* Wipe variables */
wolfSSL 7:481bce714567 1028 #if !defined(HAVE_INTEL_AVX1)&&!defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 1029 XMEMSET(W_X, 0, sizeof(word64) * 16);
wolfSSL 7:481bce714567 1030 #endif
wolfSSL 7:481bce714567 1031 XMEMSET(T, 0, sizeof(T));
wolfSSL 7:481bce714567 1032
wolfSSL 7:481bce714567 1033 return 0;
wolfSSL 7:481bce714567 1034 }
wolfSSL 7:481bce714567 1035
wolfSSL 7:481bce714567 1036 #endif
wolfSSL 7:481bce714567 1037
wolfSSL 7:481bce714567 1038 #if defined(HAVE_INTEL_AVX2) && defined(HAVE_INTEL_AVX1) && defined(HAVE_INTEL_RORX)
wolfSSL 7:481bce714567 1039
wolfSSL 7:481bce714567 1040 static int Transform_AVX1_RORX(Sha512* sha512)
wolfSSL 7:481bce714567 1041 {
wolfSSL 7:481bce714567 1042 const word64* K = K512;
wolfSSL 7:481bce714567 1043 word64 W_X[16+4];
wolfSSL 7:481bce714567 1044 word32 j;
wolfSSL 7:481bce714567 1045 word64 T[8];
wolfSSL 7:481bce714567 1046 /* Copy digest to working vars */
wolfSSL 7:481bce714567 1047 XMEMCPY(T, sha512->digest, sizeof(T));
wolfSSL 7:481bce714567 1048
wolfSSL 7:481bce714567 1049 W_from_buff(W_X, sha512->buffer) ;
wolfSSL 7:481bce714567 1050 for (j = 0; j < 80; j += 16) {
wolfSSL 7:481bce714567 1051 Rx_RORX_1( 0); Block_0_1(W_X); Rx_RORX_2( 0); Block_0_2(W_X);
wolfSSL 7:481bce714567 1052 Rx_RORX_3( 0); Block_0_3();
wolfSSL 7:481bce714567 1053 Rx_RORX_1( 1); Block_0_4(); Rx_RORX_2( 1); Block_0_5();
wolfSSL 7:481bce714567 1054 Rx_RORX_3( 1); Block_0_6(W_X);
wolfSSL 7:481bce714567 1055 Rx_RORX_1( 2); Block_0_7(W_X); Rx_RORX_2( 2); Block_0_8(W_X);
wolfSSL 7:481bce714567 1056 Rx_RORX_3( 2); Block_0_9();
wolfSSL 7:481bce714567 1057 Rx_RORX_1( 3); Block_0_10();Rx_RORX_2( 3); Block_0_11();
wolfSSL 7:481bce714567 1058 Rx_RORX_3( 3); Block_0_12(W_X);
wolfSSL 7:481bce714567 1059
wolfSSL 7:481bce714567 1060 Rx_RORX_1( 4); Block_4_1(W_X); Rx_RORX_2( 4); Block_4_2(W_X);
wolfSSL 7:481bce714567 1061 Rx_RORX_3( 4); Block_4_3();
wolfSSL 7:481bce714567 1062 Rx_RORX_1( 5); Block_4_4(); Rx_RORX_2( 5); Block_4_5();
wolfSSL 7:481bce714567 1063 Rx_RORX_3( 5); Block_4_6(W_X);
wolfSSL 7:481bce714567 1064 Rx_RORX_1( 6); Block_4_7(W_X); Rx_RORX_2( 6); Block_4_8(W_X);
wolfSSL 7:481bce714567 1065 Rx_RORX_3( 6); Block_4_9();
wolfSSL 7:481bce714567 1066 Rx_RORX_1( 7); Block_4_10();Rx_RORX_2( 7); Block_4_11();
wolfSSL 7:481bce714567 1067 Rx_RORX_3( 7); Block_4_12(W_X);
wolfSSL 7:481bce714567 1068
wolfSSL 7:481bce714567 1069 Rx_RORX_1( 8); Block_8_1(W_X); Rx_RORX_2( 8); Block_8_2(W_X);
wolfSSL 7:481bce714567 1070 Rx_RORX_3( 8); Block_8_3();
wolfSSL 7:481bce714567 1071 Rx_RORX_1( 9); Block_8_4(); Rx_RORX_2( 9); Block_8_5();
wolfSSL 7:481bce714567 1072 Rx_RORX_3( 9); Block_8_6(W_X);
wolfSSL 7:481bce714567 1073 Rx_RORX_1(10); Block_8_7(W_X); Rx_RORX_2(10); Block_8_8(W_X);
wolfSSL 7:481bce714567 1074 Rx_RORX_3(10); Block_8_9();
wolfSSL 7:481bce714567 1075 Rx_RORX_1(11); Block_8_10();Rx_RORX_2(11); Block_8_11();
wolfSSL 7:481bce714567 1076 Rx_RORX_3(11); Block_8_12(W_X);
wolfSSL 7:481bce714567 1077
wolfSSL 7:481bce714567 1078 Rx_RORX_1(12); Block_12_1(W_X); Rx_RORX_2(12); Block_12_2(W_X);
wolfSSL 7:481bce714567 1079 Rx_RORX_3(12); Block_12_3();
wolfSSL 7:481bce714567 1080 Rx_RORX_1(13); Block_12_4(); Rx_RORX_2(13); Block_12_5();
wolfSSL 7:481bce714567 1081 Rx_RORX_3(13); Block_12_6(W_X);
wolfSSL 7:481bce714567 1082 Rx_RORX_1(14); Block_12_7(W_X); Rx_RORX_2(14); Block_12_8(W_X);
wolfSSL 7:481bce714567 1083 Rx_RORX_3(14); Block_12_9();
wolfSSL 7:481bce714567 1084 Rx_RORX_1(15); Block_12_10();Rx_RORX_2(15); Block_12_11();
wolfSSL 7:481bce714567 1085 Rx_RORX_3(15); Block_12_12(W_X);
wolfSSL 7:481bce714567 1086 }
wolfSSL 7:481bce714567 1087 /* Add the working vars back into digest */
wolfSSL 7:481bce714567 1088
wolfSSL 7:481bce714567 1089 sha512->digest[0] += a(0);
wolfSSL 7:481bce714567 1090 sha512->digest[1] += b(0);
wolfSSL 7:481bce714567 1091 sha512->digest[2] += c(0);
wolfSSL 7:481bce714567 1092 sha512->digest[3] += d(0);
wolfSSL 7:481bce714567 1093 sha512->digest[4] += e(0);
wolfSSL 7:481bce714567 1094 sha512->digest[5] += f(0);
wolfSSL 7:481bce714567 1095 sha512->digest[6] += g(0);
wolfSSL 7:481bce714567 1096 sha512->digest[7] += h(0);
wolfSSL 7:481bce714567 1097
wolfSSL 7:481bce714567 1098 /* Wipe variables */
wolfSSL 7:481bce714567 1099 #if !defined(HAVE_INTEL_AVX1)&&!defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 1100 XMEMSET(W_X, 0, sizeof(word64) * 16);
wolfSSL 7:481bce714567 1101 #endif
wolfSSL 7:481bce714567 1102 XMEMSET(T, 0, sizeof(T));
wolfSSL 7:481bce714567 1103
wolfSSL 7:481bce714567 1104 return 0;
wolfSSL 7:481bce714567 1105 }
wolfSSL 7:481bce714567 1106 #endif
wolfSSL 7:481bce714567 1107
wolfSSL 7:481bce714567 1108 #if defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 1109
wolfSSL 7:481bce714567 1110 #define s0_1y(dest, src) AVX2_S(dest, src, 1);
wolfSSL 7:481bce714567 1111 #define s0_2y(dest, src) AVX2_S(G_TEMPy, src, 8); XORy(dest, G_TEMPy, dest) ;
wolfSSL 7:481bce714567 1112 #define s0_3y(dest, src) AVX2_R(G_TEMPy, src, 7); XORy(dest, G_TEMPy, dest) ;
wolfSSL 7:481bce714567 1113
wolfSSL 7:481bce714567 1114 #define s1_1y(dest, src) AVX2_S(dest, src, 19);
wolfSSL 7:481bce714567 1115 #define s1_2y(dest, src) AVX2_S(G_TEMPy, src, 61); XORy(dest, G_TEMPy, dest) ;
wolfSSL 7:481bce714567 1116 #define s1_3y(dest, src) AVX2_R(G_TEMPy, src, 6); XORy(dest, G_TEMPy, dest) ;
wolfSSL 7:481bce714567 1117
wolfSSL 7:481bce714567 1118 #define s0_y(dest, src) s0_1y(dest, src) ; s0_2y(dest, src) ; s0_3y(dest, src)
wolfSSL 7:481bce714567 1119 #define s1_y(dest, src) s1_1y(dest, src) ; s1_2y(dest, src) ; s1_3y(dest, src)
wolfSSL 7:481bce714567 1120
wolfSSL 7:481bce714567 1121
wolfSSL 7:481bce714567 1122 #define Block_Y_xx_1(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1123 MOVE_W_to_W_I_15(W_I_15y, w_0, w_4) ;\
wolfSSL 7:481bce714567 1124 MOVE_W_to_W_I_7 (W_I_7y, w_8, w_12) ;\
wolfSSL 7:481bce714567 1125 MOVE_W_to_W_I_2 (W_I_2y, w_12) ;\
wolfSSL 7:481bce714567 1126
wolfSSL 7:481bce714567 1127 #define Block_Y_xx_2(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1128 s0_1y (YMM_TEMP0, W_I_15y) ;\
wolfSSL 7:481bce714567 1129
wolfSSL 7:481bce714567 1130 #define Block_Y_xx_3(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1131 s0_2y (YMM_TEMP0, W_I_15y) ;\
wolfSSL 7:481bce714567 1132
wolfSSL 7:481bce714567 1133 #define Block_Y_xx_4(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1134 s0_3y (YMM_TEMP0, W_I_15y) ;\
wolfSSL 7:481bce714567 1135
wolfSSL 7:481bce714567 1136 #define Block_Y_xx_5(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1137 ADDy(W_I_TEMPy, w_0, YMM_TEMP0) ;\
wolfSSL 7:481bce714567 1138
wolfSSL 7:481bce714567 1139 #define Block_Y_xx_6(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1140 ADDy(W_I_TEMPy, W_I_TEMPy, W_I_7y) ;\
wolfSSL 7:481bce714567 1141 s1_1y (YMM_TEMP0, W_I_2y) ;\
wolfSSL 7:481bce714567 1142
wolfSSL 7:481bce714567 1143 #define Block_Y_xx_7(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1144 s1_2y (YMM_TEMP0, W_I_2y) ;\
wolfSSL 7:481bce714567 1145
wolfSSL 7:481bce714567 1146 #define Block_Y_xx_8(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1147 s1_3y (YMM_TEMP0, W_I_2y) ;\
wolfSSL 7:481bce714567 1148 ADDy(w_0, W_I_TEMPy, YMM_TEMP0) ;\
wolfSSL 7:481bce714567 1149
wolfSSL 7:481bce714567 1150 #define Block_Y_xx_9(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1151 FEEDBACK1_to_W_I_2(W_I_2y, w_0) ;\
wolfSSL 7:481bce714567 1152
wolfSSL 7:481bce714567 1153 #define Block_Y_xx_10(i, w_0, w_4, w_8, w_12) \
wolfSSL 7:481bce714567 1154 s1_1y (YMM_TEMP0, W_I_2y) ;\
wolfSSL 7:481bce714567 1155
wolfSSL 7:481bce714567 1156 #define Block_Y_xx_11(i, w_0, w_4, w_8, w_12) \
wolfSSL 7:481bce714567 1157 s1_2y (YMM_TEMP0, W_I_2y) ;\
wolfSSL 7:481bce714567 1158
wolfSSL 7:481bce714567 1159 #define Block_Y_xx_12(i, w_0, w_4, w_8, w_12)\
wolfSSL 7:481bce714567 1160 s1_3y (YMM_TEMP0, W_I_2y) ;\
wolfSSL 7:481bce714567 1161 ADDy(w_0, W_I_TEMPy, YMM_TEMP0) ;\
wolfSSL 7:481bce714567 1162 MOVE_to_MEMy(w,0, w_4) ;\
wolfSSL 7:481bce714567 1163
wolfSSL 7:481bce714567 1164
wolfSSL 7:481bce714567 1165 static INLINE void Block_Y_0_1(void) { Block_Y_xx_1(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1166 static INLINE void Block_Y_0_2(void) { Block_Y_xx_2(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1167 static INLINE void Block_Y_0_3(void) { Block_Y_xx_3(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1168 static INLINE void Block_Y_0_4(void) { Block_Y_xx_4(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1169 static INLINE void Block_Y_0_5(void) { Block_Y_xx_5(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1170 static INLINE void Block_Y_0_6(void) { Block_Y_xx_6(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1171 static INLINE void Block_Y_0_7(void) { Block_Y_xx_7(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1172 static INLINE void Block_Y_0_8(void) { Block_Y_xx_8(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1173 static INLINE void Block_Y_0_9(void) { Block_Y_xx_9(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1174 static INLINE void Block_Y_0_10(void){ Block_Y_xx_10(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1175 static INLINE void Block_Y_0_11(void){ Block_Y_xx_11(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1176 static INLINE void Block_Y_0_12(word64 *w){ Block_Y_xx_12(0, W_0y, W_4y, W_8y, W_12y) ; }
wolfSSL 7:481bce714567 1177
wolfSSL 7:481bce714567 1178 static INLINE void Block_Y_4_1(void) { Block_Y_xx_1(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1179 static INLINE void Block_Y_4_2(void) { Block_Y_xx_2(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1180 static INLINE void Block_Y_4_3(void) { Block_Y_xx_3(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1181 static INLINE void Block_Y_4_4(void) { Block_Y_xx_4(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1182 static INLINE void Block_Y_4_5(void) { Block_Y_xx_5(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1183 static INLINE void Block_Y_4_6(void) { Block_Y_xx_6(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1184 static INLINE void Block_Y_4_7(void) { Block_Y_xx_7(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1185 static INLINE void Block_Y_4_8(void) { Block_Y_xx_8(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1186 static INLINE void Block_Y_4_9(void) { Block_Y_xx_9(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1187 static INLINE void Block_Y_4_10(void) { Block_Y_xx_10(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1188 static INLINE void Block_Y_4_11(void) { Block_Y_xx_11(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1189 static INLINE void Block_Y_4_12(word64 *w) { Block_Y_xx_12(4, W_4y, W_8y, W_12y, W_0y) ; }
wolfSSL 7:481bce714567 1190
wolfSSL 7:481bce714567 1191 static INLINE void Block_Y_8_1(void) { Block_Y_xx_1(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1192 static INLINE void Block_Y_8_2(void) { Block_Y_xx_2(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1193 static INLINE void Block_Y_8_3(void) { Block_Y_xx_3(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1194 static INLINE void Block_Y_8_4(void) { Block_Y_xx_4(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1195 static INLINE void Block_Y_8_5(void) { Block_Y_xx_5(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1196 static INLINE void Block_Y_8_6(void) { Block_Y_xx_6(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1197 static INLINE void Block_Y_8_7(void) { Block_Y_xx_7(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1198 static INLINE void Block_Y_8_8(void) { Block_Y_xx_8(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1199 static INLINE void Block_Y_8_9(void) { Block_Y_xx_9(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1200 static INLINE void Block_Y_8_10(void) { Block_Y_xx_10(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1201 static INLINE void Block_Y_8_11(void) { Block_Y_xx_11(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1202 static INLINE void Block_Y_8_12(word64 *w) { Block_Y_xx_12(8, W_8y, W_12y, W_0y, W_4y) ; }
wolfSSL 7:481bce714567 1203
wolfSSL 7:481bce714567 1204 static INLINE void Block_Y_12_1(void) { Block_Y_xx_1(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1205 static INLINE void Block_Y_12_2(void) { Block_Y_xx_2(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1206 static INLINE void Block_Y_12_3(void) { Block_Y_xx_3(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1207 static INLINE void Block_Y_12_4(void) { Block_Y_xx_4(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1208 static INLINE void Block_Y_12_5(void) { Block_Y_xx_5(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1209 static INLINE void Block_Y_12_6(void) { Block_Y_xx_6(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1210 static INLINE void Block_Y_12_7(void) { Block_Y_xx_7(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1211 static INLINE void Block_Y_12_8(void) { Block_Y_xx_8(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1212 static INLINE void Block_Y_12_9(void) { Block_Y_xx_9(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1213 static INLINE void Block_Y_12_10(void) { Block_Y_xx_10(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1214 static INLINE void Block_Y_12_11(void) { Block_Y_xx_11(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1215 static INLINE void Block_Y_12_12(word64 *w) { Block_Y_xx_12(12, W_12y, W_0y, W_4y, W_8y) ; }
wolfSSL 7:481bce714567 1216
wolfSSL 7:481bce714567 1217
wolfSSL 7:481bce714567 1218 static int Transform_AVX2(Sha512* sha512)
wolfSSL 7:481bce714567 1219 {
wolfSSL 7:481bce714567 1220 const word64* K = K512;
wolfSSL 7:481bce714567 1221 word64 w[4] ;
wolfSSL 7:481bce714567 1222 word32 j /*, k*/;
wolfSSL 7:481bce714567 1223 word64 T[8];
wolfSSL 7:481bce714567 1224 /* Copy digest to working vars */
wolfSSL 7:481bce714567 1225 XMEMCPY(T, sha512->digest, sizeof(T));
wolfSSL 7:481bce714567 1226
wolfSSL 7:481bce714567 1227 W_from_buff_Y(sha512->buffer) ;
wolfSSL 7:481bce714567 1228 MOVE_to_MEMy(w,0, W_0y) ;
wolfSSL 7:481bce714567 1229 for (j = 0; j < 80; j += 16) {
wolfSSL 7:481bce714567 1230 Ry_1( 0, w[0]); Block_Y_0_1(); Ry_2( 0, w[0]); Block_Y_0_2();
wolfSSL 7:481bce714567 1231 Ry_3( 0, w[0]); Block_Y_0_3();
wolfSSL 7:481bce714567 1232 Ry_1( 1, w[1]); Block_Y_0_4(); Ry_2( 1, w[1]); Block_Y_0_5();
wolfSSL 7:481bce714567 1233 Ry_3( 1, w[1]); Block_Y_0_6();
wolfSSL 7:481bce714567 1234 Ry_1( 2, w[2]); Block_Y_0_7(); Ry_2( 2, w[2]); Block_Y_0_8();
wolfSSL 7:481bce714567 1235 Ry_3( 2, w[2]); Block_Y_0_9();
wolfSSL 7:481bce714567 1236 Ry_1( 3, w[3]); Block_Y_0_10();Ry_2( 3, w[3]); Block_Y_0_11();
wolfSSL 7:481bce714567 1237 Ry_3( 3, w[3]); Block_Y_0_12(w);
wolfSSL 7:481bce714567 1238
wolfSSL 7:481bce714567 1239 Ry_1( 4, w[0]); Block_Y_4_1(); Ry_2( 4, w[0]); Block_Y_4_2();
wolfSSL 7:481bce714567 1240 Ry_3( 4, w[0]); Block_Y_4_3();
wolfSSL 7:481bce714567 1241 Ry_1( 5, w[1]); Block_Y_4_4(); Ry_2( 5, w[1]); Block_Y_4_5();
wolfSSL 7:481bce714567 1242 Ry_3( 5, w[1]); Block_Y_4_6();
wolfSSL 7:481bce714567 1243 Ry_1( 6, w[2]); Block_Y_4_7(); Ry_2( 6, w[2]); Block_Y_4_8();
wolfSSL 7:481bce714567 1244 Ry_3( 6, w[2]); Block_Y_4_9();
wolfSSL 7:481bce714567 1245 Ry_1( 7, w[3]); Block_Y_4_10(); Ry_2( 7, w[3]);Block_Y_4_11();
wolfSSL 7:481bce714567 1246 Ry_3( 7, w[3]);Block_Y_4_12(w);
wolfSSL 7:481bce714567 1247
wolfSSL 7:481bce714567 1248 Ry_1( 8, w[0]); Block_Y_8_1(); Ry_2( 8, w[0]); Block_Y_8_2();
wolfSSL 7:481bce714567 1249 Ry_3( 8, w[0]); Block_Y_8_3();
wolfSSL 7:481bce714567 1250 Ry_1( 9, w[1]); Block_Y_8_4(); Ry_2( 9, w[1]); Block_Y_8_5();
wolfSSL 7:481bce714567 1251 Ry_3( 9, w[1]); Block_Y_8_6();
wolfSSL 7:481bce714567 1252 Ry_1(10, w[2]); Block_Y_8_7(); Ry_2(10, w[2]); Block_Y_8_8();
wolfSSL 7:481bce714567 1253 Ry_3(10, w[2]); Block_Y_8_9();
wolfSSL 7:481bce714567 1254 Ry_1(11, w[3]); Block_Y_8_10();Ry_2(11, w[3]); Block_Y_8_11();
wolfSSL 7:481bce714567 1255 Ry_3(11, w[3]); Block_Y_8_12(w);
wolfSSL 7:481bce714567 1256
wolfSSL 7:481bce714567 1257 Ry_1(12, w[0]); Block_Y_12_1(); Ry_2(12, w[0]); Block_Y_12_2();
wolfSSL 7:481bce714567 1258 Ry_3(12, w[0]); Block_Y_12_3();
wolfSSL 7:481bce714567 1259 Ry_1(13, w[1]); Block_Y_12_4(); Ry_2(13, w[1]); Block_Y_12_5();
wolfSSL 7:481bce714567 1260 Ry_3(13, w[1]); Block_Y_12_6();
wolfSSL 7:481bce714567 1261 Ry_1(14, w[2]); Block_Y_12_7(); Ry_2(14, w[2]); Block_Y_12_8();
wolfSSL 7:481bce714567 1262 Ry_3(14, w[2]); Block_Y_12_9();
wolfSSL 7:481bce714567 1263 Ry_1(15, w[3]); Block_Y_12_10();Ry_2(15, w[3]); Block_Y_12_11();
wolfSSL 7:481bce714567 1264 Ry_3(15, w[3]);Block_Y_12_12(w);
wolfSSL 7:481bce714567 1265 }
wolfSSL 7:481bce714567 1266
wolfSSL 7:481bce714567 1267 /* Add the working vars back into digest */
wolfSSL 7:481bce714567 1268
wolfSSL 7:481bce714567 1269 sha512->digest[0] += a(0);
wolfSSL 7:481bce714567 1270 sha512->digest[1] += b(0);
wolfSSL 7:481bce714567 1271 sha512->digest[2] += c(0);
wolfSSL 7:481bce714567 1272 sha512->digest[3] += d(0);
wolfSSL 7:481bce714567 1273 sha512->digest[4] += e(0);
wolfSSL 7:481bce714567 1274 sha512->digest[5] += f(0);
wolfSSL 7:481bce714567 1275 sha512->digest[6] += g(0);
wolfSSL 7:481bce714567 1276 sha512->digest[7] += h(0);
wolfSSL 7:481bce714567 1277
wolfSSL 7:481bce714567 1278 /* Wipe variables */
wolfSSL 7:481bce714567 1279 #if !defined(HAVE_INTEL_AVX1)&&!defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 1280 XMEMSET(W, 0, sizeof(word64) * 16);
wolfSSL 7:481bce714567 1281 #endif
wolfSSL 7:481bce714567 1282 XMEMSET(T, 0, sizeof(T));
wolfSSL 7:481bce714567 1283
wolfSSL 7:481bce714567 1284 return 0;
wolfSSL 7:481bce714567 1285 }
wolfSSL 7:481bce714567 1286
wolfSSL 7:481bce714567 1287 #endif
wolfSSL 7:481bce714567 1288
wolfSSL 7:481bce714567 1289
wolfSSL 7:481bce714567 1290 #ifdef WOLFSSL_SHA384
wolfSSL 7:481bce714567 1291 int wc_InitSha384(Sha384* sha384)
wolfSSL 7:481bce714567 1292 {
wolfSSL 7:481bce714567 1293 sha384->digest[0] = W64LIT(0xcbbb9d5dc1059ed8);
wolfSSL 7:481bce714567 1294 sha384->digest[1] = W64LIT(0x629a292a367cd507);
wolfSSL 7:481bce714567 1295 sha384->digest[2] = W64LIT(0x9159015a3070dd17);
wolfSSL 7:481bce714567 1296 sha384->digest[3] = W64LIT(0x152fecd8f70e5939);
wolfSSL 7:481bce714567 1297 sha384->digest[4] = W64LIT(0x67332667ffc00b31);
wolfSSL 7:481bce714567 1298 sha384->digest[5] = W64LIT(0x8eb44a8768581511);
wolfSSL 7:481bce714567 1299 sha384->digest[6] = W64LIT(0xdb0c2e0d64f98fa7);
wolfSSL 7:481bce714567 1300 sha384->digest[7] = W64LIT(0x47b5481dbefa4fa4);
wolfSSL 7:481bce714567 1301
wolfSSL 7:481bce714567 1302 sha384->buffLen = 0;
wolfSSL 7:481bce714567 1303 sha384->loLen = 0;
wolfSSL 7:481bce714567 1304 sha384->hiLen = 0;
wolfSSL 7:481bce714567 1305
wolfSSL 7:481bce714567 1306 #if defined(HAVE_INTEL_AVX1)|| defined(HAVE_INTEL_AVX2)
wolfSSL 7:481bce714567 1307 set_Transform() ;
wolfSSL 7:481bce714567 1308 #endif
wolfSSL 7:481bce714567 1309
wolfSSL 7:481bce714567 1310 return 0;
wolfSSL 7:481bce714567 1311 }
wolfSSL 7:481bce714567 1312
wolfSSL 7:481bce714567 1313 int wc_Sha384Update(Sha384* sha384, const byte* data, word32 len)
wolfSSL 7:481bce714567 1314 {
wolfSSL 7:481bce714567 1315 return Sha512Update((Sha512 *)sha384, data, len);
wolfSSL 7:481bce714567 1316 }
wolfSSL 7:481bce714567 1317
wolfSSL 7:481bce714567 1318
wolfSSL 7:481bce714567 1319 int wc_Sha384Final(Sha384* sha384, byte* hash)
wolfSSL 7:481bce714567 1320 {
wolfSSL 7:481bce714567 1321 int ret = Sha512Final((Sha512 *)sha384);
wolfSSL 7:481bce714567 1322 if (ret != 0)
wolfSSL 7:481bce714567 1323 return ret;
wolfSSL 7:481bce714567 1324
wolfSSL 7:481bce714567 1325 XMEMCPY(hash, sha384->digest, SHA384_DIGEST_SIZE);
wolfSSL 7:481bce714567 1326
wolfSSL 7:481bce714567 1327 return wc_InitSha384(sha384); /* reset state */
wolfSSL 7:481bce714567 1328 }
wolfSSL 7:481bce714567 1329 #endif /* WOLFSSL_SHA384 */
wolfSSL 7:481bce714567 1330
wolfSSL 7:481bce714567 1331 #endif /* HAVE_FIPS */
wolfSSL 7:481bce714567 1332
wolfSSL 7:481bce714567 1333 #endif /* WOLFSSL_SHA512 */
wolfSSL 7:481bce714567 1334
wolfSSL 7:481bce714567 1335