1 /* 2 * xxHash - Fast Hash algorithm 3 * Copyright (c) 2012-2020, Yann Collet, Facebook, Inc. 4 * 5 * You can contact the author at : 6 * - xxHash homepage: http://www.xxhash.com 7 * - xxHash source repository : https://github.com/Cyan4973/xxHash 8 * 9 * This source code is licensed under both the BSD-style license (found in the 10 * LICENSE file in the root directory of this source tree) and the GPLv2 (found 11 * in the COPYING file in the root directory of this source tree). 12 * You may select, at your option, one of the above-listed licenses. 13 */ 14 15 16 /* ************************************* 17 * Tuning parameters 18 ***************************************/ 19 /*!XXH_FORCE_MEMORY_ACCESS : 20 * By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable. 21 * Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal. 22 * The below switch allow to select different access method for improved performance. 23 * Method 0 (default) : use `memcpy()`. Safe and portable. 24 * Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable). 25 * This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`. 26 * Method 2 : direct access. This method doesn't depend on compiler but violate C standard. 27 * It can generate buggy code on targets which do not support unaligned memory accesses. 28 * But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6) 29 * See http://stackoverflow.com/a/32095106/646947 for details. 30 * Prefer these methods in priority order (0 > 1 > 2) 31 */ 32 #ifndef XXH_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ 33 # if defined(__GNUC__) && ( defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) ) 34 # define XXH_FORCE_MEMORY_ACCESS 2 35 # elif (defined(__INTEL_COMPILER) && !defined(WIN32)) || \ 36 (defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) )) || \ 37 defined(__ICCARM__) 38 # define XXH_FORCE_MEMORY_ACCESS 1 39 # endif 40 #endif 41 42 /*!XXH_ACCEPT_NULL_INPUT_POINTER : 43 * If the input pointer is a null pointer, xxHash default behavior is to trigger a memory access error, since it is a bad pointer. 44 * When this option is enabled, xxHash output for null input pointers will be the same as a null-length input. 45 * By default, this option is disabled. To enable it, uncomment below define : 46 */ 47 /* #define XXH_ACCEPT_NULL_INPUT_POINTER 1 */ 48 49 /*!XXH_FORCE_NATIVE_FORMAT : 50 * By default, xxHash library provides endian-independent Hash values, based on little-endian convention. 51 * Results are therefore identical for little-endian and big-endian CPU. 52 * This comes at a performance cost for big-endian CPU, since some swapping is required to emulate little-endian format. 53 * Should endian-independence be of no importance for your application, you may set the #define below to 1, 54 * to improve speed for Big-endian CPU. 55 * This option has no impact on Little_Endian CPU. 56 */ 57 #ifndef XXH_FORCE_NATIVE_FORMAT /* can be defined externally */ 58 # define XXH_FORCE_NATIVE_FORMAT 0 59 #endif 60 61 /*!XXH_FORCE_ALIGN_CHECK : 62 * This is a minor performance trick, only useful with lots of very small keys. 63 * It means : check for aligned/unaligned input. 64 * The check costs one initial branch per hash; set to 0 when the input data 65 * is guaranteed to be aligned. 66 */ 67 #ifndef XXH_FORCE_ALIGN_CHECK /* can be defined externally */ 68 # if defined(__i386) || defined(_M_IX86) || defined(__x86_64__) || defined(_M_X64) 69 # define XXH_FORCE_ALIGN_CHECK 0 70 # else 71 # define XXH_FORCE_ALIGN_CHECK 1 72 # endif 73 #endif 74 75 76 /* ************************************* 77 * Includes & Memory related functions 78 ***************************************/ 79 /* Modify the local functions below should you wish to use some other memory routines */ 80 /* for ZSTD_malloc(), ZSTD_free() */ 81 #define ZSTD_DEPS_NEED_MALLOC 82 #include "zstd_deps.h" /* size_t, ZSTD_malloc, ZSTD_free, ZSTD_memcpy */ 83 static void* XXH_malloc(size_t s) { return ZSTD_malloc(s); } 84 static void XXH_free (void* p) { ZSTD_free(p); } 85 static void* XXH_memcpy(void* dest, const void* src, size_t size) { return ZSTD_memcpy(dest,src,size); } 86 87 #ifndef XXH_STATIC_LINKING_ONLY 88 # define XXH_STATIC_LINKING_ONLY 89 #endif 90 #include "xxhash.h" 91 92 93 /* ************************************* 94 * Compiler Specific Options 95 ***************************************/ 96 #include "compiler.h" 97 98 99 /* ************************************* 100 * Basic Types 101 ***************************************/ 102 #include "mem.h" /* BYTE, U32, U64, size_t */ 103 104 #if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) 105 106 /* Force direct memory access. Only works on CPU which support unaligned memory access in hardware */ 107 static U32 XXH_read32(const void* memPtr) { return *(const U32*) memPtr; } 108 static U64 XXH_read64(const void* memPtr) { return *(const U64*) memPtr; } 109 110 #elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) 111 112 /* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */ 113 /* currently only defined for gcc and icc */ 114 typedef union { U32 u32; U64 u64; } __attribute__((packed)) unalign; 115 116 static U32 XXH_read32(const void* ptr) { return ((const unalign*)ptr)->u32; } 117 static U64 XXH_read64(const void* ptr) { return ((const unalign*)ptr)->u64; } 118 119 #else 120 121 /* portable and safe solution. Generally efficient. 122 * see : http://stackoverflow.com/a/32095106/646947 123 */ 124 125 static U32 XXH_read32(const void* memPtr) 126 { 127 U32 val; 128 ZSTD_memcpy(&val, memPtr, sizeof(val)); 129 return val; 130 } 131 132 static U64 XXH_read64(const void* memPtr) 133 { 134 U64 val; 135 ZSTD_memcpy(&val, memPtr, sizeof(val)); 136 return val; 137 } 138 139 #endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ 140 141 142 /* **************************************** 143 * Compiler-specific Functions and Macros 144 ******************************************/ 145 #define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) 146 147 /* Note : although _rotl exists for minGW (GCC under windows), performance seems poor */ 148 #if defined(_MSC_VER) 149 # define XXH_rotl32(x,r) _rotl(x,r) 150 # define XXH_rotl64(x,r) _rotl64(x,r) 151 #else 152 #if defined(__ICCARM__) 153 # include <intrinsics.h> 154 # define XXH_rotl32(x,r) __ROR(x,(32 - r)) 155 #else 156 # define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r))) 157 #endif 158 # define XXH_rotl64(x,r) ((x << r) | (x >> (64 - r))) 159 #endif 160 161 #if defined(_MSC_VER) /* Visual Studio */ 162 # define XXH_swap32 _byteswap_ulong 163 # define XXH_swap64 _byteswap_uint64 164 #elif GCC_VERSION >= 403 165 # define XXH_swap32 __builtin_bswap32 166 # define XXH_swap64 __builtin_bswap64 167 #else 168 static U32 XXH_swap32 (U32 x) 169 { 170 return ((x << 24) & 0xff000000 ) | 171 ((x << 8) & 0x00ff0000 ) | 172 ((x >> 8) & 0x0000ff00 ) | 173 ((x >> 24) & 0x000000ff ); 174 } 175 static U64 XXH_swap64 (U64 x) 176 { 177 return ((x << 56) & 0xff00000000000000ULL) | 178 ((x << 40) & 0x00ff000000000000ULL) | 179 ((x << 24) & 0x0000ff0000000000ULL) | 180 ((x << 8) & 0x000000ff00000000ULL) | 181 ((x >> 8) & 0x00000000ff000000ULL) | 182 ((x >> 24) & 0x0000000000ff0000ULL) | 183 ((x >> 40) & 0x000000000000ff00ULL) | 184 ((x >> 56) & 0x00000000000000ffULL); 185 } 186 #endif 187 188 189 /* ************************************* 190 * Architecture Macros 191 ***************************************/ 192 typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess; 193 194 /* XXH_CPU_LITTLE_ENDIAN can be defined externally, for example on the compiler command line */ 195 #ifndef XXH_CPU_LITTLE_ENDIAN 196 static const int g_one = 1; 197 # define XXH_CPU_LITTLE_ENDIAN (*(const char*)(&g_one)) 198 #endif 199 200 201 /* *************************** 202 * Memory reads 203 *****************************/ 204 typedef enum { XXH_aligned, XXH_unaligned } XXH_alignment; 205 206 FORCE_INLINE_TEMPLATE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, XXH_alignment align) 207 { 208 if (align==XXH_unaligned) 209 return endian==XXH_littleEndian ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr)); 210 else 211 return endian==XXH_littleEndian ? *(const U32*)ptr : XXH_swap32(*(const U32*)ptr); 212 } 213 214 FORCE_INLINE_TEMPLATE U32 XXH_readLE32(const void* ptr, XXH_endianess endian) 215 { 216 return XXH_readLE32_align(ptr, endian, XXH_unaligned); 217 } 218 219 static U32 XXH_readBE32(const void* ptr) 220 { 221 return XXH_CPU_LITTLE_ENDIAN ? XXH_swap32(XXH_read32(ptr)) : XXH_read32(ptr); 222 } 223 224 FORCE_INLINE_TEMPLATE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, XXH_alignment align) 225 { 226 if (align==XXH_unaligned) 227 return endian==XXH_littleEndian ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr)); 228 else 229 return endian==XXH_littleEndian ? *(const U64*)ptr : XXH_swap64(*(const U64*)ptr); 230 } 231 232 FORCE_INLINE_TEMPLATE U64 XXH_readLE64(const void* ptr, XXH_endianess endian) 233 { 234 return XXH_readLE64_align(ptr, endian, XXH_unaligned); 235 } 236 237 static U64 XXH_readBE64(const void* ptr) 238 { 239 return XXH_CPU_LITTLE_ENDIAN ? XXH_swap64(XXH_read64(ptr)) : XXH_read64(ptr); 240 } 241 242 243 /* ************************************* 244 * Macros 245 ***************************************/ 246 #define XXH_STATIC_ASSERT(c) { enum { XXH_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */ 247 248 249 /* ************************************* 250 * Constants 251 ***************************************/ 252 static const U32 PRIME32_1 = 2654435761U; 253 static const U32 PRIME32_2 = 2246822519U; 254 static const U32 PRIME32_3 = 3266489917U; 255 static const U32 PRIME32_4 = 668265263U; 256 static const U32 PRIME32_5 = 374761393U; 257 258 static const U64 PRIME64_1 = 11400714785074694791ULL; 259 static const U64 PRIME64_2 = 14029467366897019727ULL; 260 static const U64 PRIME64_3 = 1609587929392839161ULL; 261 static const U64 PRIME64_4 = 9650029242287828579ULL; 262 static const U64 PRIME64_5 = 2870177450012600261ULL; 263 264 XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; } 265 266 267 /* ************************** 268 * Utils 269 ****************************/ 270 XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* restrict dstState, const XXH32_state_t* restrict srcState) 271 { 272 ZSTD_memcpy(dstState, srcState, sizeof(*dstState)); 273 } 274 275 XXH_PUBLIC_API void XXH64_copyState(XXH64_state_t* restrict dstState, const XXH64_state_t* restrict srcState) 276 { 277 ZSTD_memcpy(dstState, srcState, sizeof(*dstState)); 278 } 279 280 281 /* *************************** 282 * Simple Hash Functions 283 *****************************/ 284 285 static U32 XXH32_round(U32 seed, U32 input) 286 { 287 seed += input * PRIME32_2; 288 seed = XXH_rotl32(seed, 13); 289 seed *= PRIME32_1; 290 return seed; 291 } 292 293 FORCE_INLINE_TEMPLATE U32 XXH32_endian_align(const void* input, size_t len, U32 seed, XXH_endianess endian, XXH_alignment align) 294 { 295 const BYTE* p = (const BYTE*)input; 296 const BYTE* bEnd = p + len; 297 U32 h32; 298 #define XXH_get32bits(p) XXH_readLE32_align(p, endian, align) 299 300 #ifdef XXH_ACCEPT_NULL_INPUT_POINTER 301 if (p==NULL) { 302 len=0; 303 bEnd=p=(const BYTE*)(size_t)16; 304 } 305 #endif 306 307 if (len>=16) { 308 const BYTE* const limit = bEnd - 16; 309 U32 v1 = seed + PRIME32_1 + PRIME32_2; 310 U32 v2 = seed + PRIME32_2; 311 U32 v3 = seed + 0; 312 U32 v4 = seed - PRIME32_1; 313 314 do { 315 v1 = XXH32_round(v1, XXH_get32bits(p)); p+=4; 316 v2 = XXH32_round(v2, XXH_get32bits(p)); p+=4; 317 v3 = XXH32_round(v3, XXH_get32bits(p)); p+=4; 318 v4 = XXH32_round(v4, XXH_get32bits(p)); p+=4; 319 } while (p<=limit); 320 321 h32 = XXH_rotl32(v1, 1) + XXH_rotl32(v2, 7) + XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18); 322 } else { 323 h32 = seed + PRIME32_5; 324 } 325 326 h32 += (U32) len; 327 328 while (p+4<=bEnd) { 329 h32 += XXH_get32bits(p) * PRIME32_3; 330 h32 = XXH_rotl32(h32, 17) * PRIME32_4 ; 331 p+=4; 332 } 333 334 while (p<bEnd) { 335 h32 += (*p) * PRIME32_5; 336 h32 = XXH_rotl32(h32, 11) * PRIME32_1 ; 337 p++; 338 } 339 340 h32 ^= h32 >> 15; 341 h32 *= PRIME32_2; 342 h32 ^= h32 >> 13; 343 h32 *= PRIME32_3; 344 h32 ^= h32 >> 16; 345 346 return h32; 347 } 348 349 350 XXH_PUBLIC_API unsigned int XXH32 (const void* input, size_t len, unsigned int seed) 351 { 352 #if 0 353 /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ 354 XXH32_CREATESTATE_STATIC(state); 355 XXH32_reset(state, seed); 356 XXH32_update(state, input, len); 357 return XXH32_digest(state); 358 #else 359 XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; 360 361 if (XXH_FORCE_ALIGN_CHECK) { 362 if ((((size_t)input) & 3) == 0) { /* Input is 4-bytes aligned, leverage the speed benefit */ 363 if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) 364 return XXH32_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned); 365 else 366 return XXH32_endian_align(input, len, seed, XXH_bigEndian, XXH_aligned); 367 } } 368 369 if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) 370 return XXH32_endian_align(input, len, seed, XXH_littleEndian, XXH_unaligned); 371 else 372 return XXH32_endian_align(input, len, seed, XXH_bigEndian, XXH_unaligned); 373 #endif 374 } 375 376 377 static U64 XXH64_round(U64 acc, U64 input) 378 { 379 acc += input * PRIME64_2; 380 acc = XXH_rotl64(acc, 31); 381 acc *= PRIME64_1; 382 return acc; 383 } 384 385 static U64 XXH64_mergeRound(U64 acc, U64 val) 386 { 387 val = XXH64_round(0, val); 388 acc ^= val; 389 acc = acc * PRIME64_1 + PRIME64_4; 390 return acc; 391 } 392 393 FORCE_INLINE_TEMPLATE U64 XXH64_endian_align(const void* input, size_t len, U64 seed, XXH_endianess endian, XXH_alignment align) 394 { 395 const BYTE* p = (const BYTE*)input; 396 const BYTE* const bEnd = p + len; 397 U64 h64; 398 #define XXH_get64bits(p) XXH_readLE64_align(p, endian, align) 399 400 #ifdef XXH_ACCEPT_NULL_INPUT_POINTER 401 if (p==NULL) { 402 len=0; 403 bEnd=p=(const BYTE*)(size_t)32; 404 } 405 #endif 406 407 if (len>=32) { 408 const BYTE* const limit = bEnd - 32; 409 U64 v1 = seed + PRIME64_1 + PRIME64_2; 410 U64 v2 = seed + PRIME64_2; 411 U64 v3 = seed + 0; 412 U64 v4 = seed - PRIME64_1; 413 414 do { 415 v1 = XXH64_round(v1, XXH_get64bits(p)); p+=8; 416 v2 = XXH64_round(v2, XXH_get64bits(p)); p+=8; 417 v3 = XXH64_round(v3, XXH_get64bits(p)); p+=8; 418 v4 = XXH64_round(v4, XXH_get64bits(p)); p+=8; 419 } while (p<=limit); 420 421 h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18); 422 h64 = XXH64_mergeRound(h64, v1); 423 h64 = XXH64_mergeRound(h64, v2); 424 h64 = XXH64_mergeRound(h64, v3); 425 h64 = XXH64_mergeRound(h64, v4); 426 427 } else { 428 h64 = seed + PRIME64_5; 429 } 430 431 h64 += (U64) len; 432 433 while (p+8<=bEnd) { 434 U64 const k1 = XXH64_round(0, XXH_get64bits(p)); 435 h64 ^= k1; 436 h64 = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4; 437 p+=8; 438 } 439 440 if (p+4<=bEnd) { 441 h64 ^= (U64)(XXH_get32bits(p)) * PRIME64_1; 442 h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3; 443 p+=4; 444 } 445 446 while (p<bEnd) { 447 h64 ^= (*p) * PRIME64_5; 448 h64 = XXH_rotl64(h64, 11) * PRIME64_1; 449 p++; 450 } 451 452 h64 ^= h64 >> 33; 453 h64 *= PRIME64_2; 454 h64 ^= h64 >> 29; 455 h64 *= PRIME64_3; 456 h64 ^= h64 >> 32; 457 458 return h64; 459 } 460 461 462 XXH_PUBLIC_API unsigned long long XXH64 (const void* input, size_t len, unsigned long long seed) 463 { 464 #if 0 465 /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ 466 XXH64_CREATESTATE_STATIC(state); 467 XXH64_reset(state, seed); 468 XXH64_update(state, input, len); 469 return XXH64_digest(state); 470 #else 471 XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; 472 473 if (XXH_FORCE_ALIGN_CHECK) { 474 if ((((size_t)input) & 7)==0) { /* Input is aligned, let's leverage the speed advantage */ 475 if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) 476 return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned); 477 else 478 return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_aligned); 479 } } 480 481 if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) 482 return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_unaligned); 483 else 484 return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_unaligned); 485 #endif 486 } 487 488 489 /* ************************************************** 490 * Advanced Hash Functions 491 ****************************************************/ 492 493 XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void) 494 { 495 return (XXH32_state_t*)XXH_malloc(sizeof(XXH32_state_t)); 496 } 497 XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr) 498 { 499 XXH_free(statePtr); 500 return XXH_OK; 501 } 502 503 XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void) 504 { 505 return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t)); 506 } 507 XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr) 508 { 509 XXH_free(statePtr); 510 return XXH_OK; 511 } 512 513 514 /*** Hash feed ***/ 515 516 XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, unsigned int seed) 517 { 518 XXH32_state_t state; /* using a local state to memcpy() in order to avoid strict-aliasing warnings */ 519 ZSTD_memset(&state, 0, sizeof(state)-4); /* do not write into reserved, for future removal */ 520 state.v1 = seed + PRIME32_1 + PRIME32_2; 521 state.v2 = seed + PRIME32_2; 522 state.v3 = seed + 0; 523 state.v4 = seed - PRIME32_1; 524 ZSTD_memcpy(statePtr, &state, sizeof(state)); 525 return XXH_OK; 526 } 527 528 529 XXH_PUBLIC_API XXH_errorcode XXH64_reset(XXH64_state_t* statePtr, unsigned long long seed) 530 { 531 XXH64_state_t state; /* using a local state to memcpy() in order to avoid strict-aliasing warnings */ 532 ZSTD_memset(&state, 0, sizeof(state)-8); /* do not write into reserved, for future removal */ 533 state.v1 = seed + PRIME64_1 + PRIME64_2; 534 state.v2 = seed + PRIME64_2; 535 state.v3 = seed + 0; 536 state.v4 = seed - PRIME64_1; 537 ZSTD_memcpy(statePtr, &state, sizeof(state)); 538 return XXH_OK; 539 } 540 541 542 FORCE_INLINE_TEMPLATE XXH_errorcode XXH32_update_endian (XXH32_state_t* state, const void* input, size_t len, XXH_endianess endian) 543 { 544 const BYTE* p = (const BYTE*)input; 545 const BYTE* const bEnd = p + len; 546 547 #ifdef XXH_ACCEPT_NULL_INPUT_POINTER 548 if (input==NULL) return XXH_ERROR; 549 #endif 550 551 state->total_len_32 += (unsigned)len; 552 state->large_len |= (len>=16) | (state->total_len_32>=16); 553 554 if (state->memsize + len < 16) { /* fill in tmp buffer */ 555 XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, len); 556 state->memsize += (unsigned)len; 557 return XXH_OK; 558 } 559 560 if (state->memsize) { /* some data left from previous update */ 561 XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, 16-state->memsize); 562 { const U32* p32 = state->mem32; 563 state->v1 = XXH32_round(state->v1, XXH_readLE32(p32, endian)); p32++; 564 state->v2 = XXH32_round(state->v2, XXH_readLE32(p32, endian)); p32++; 565 state->v3 = XXH32_round(state->v3, XXH_readLE32(p32, endian)); p32++; 566 state->v4 = XXH32_round(state->v4, XXH_readLE32(p32, endian)); p32++; 567 } 568 p += 16-state->memsize; 569 state->memsize = 0; 570 } 571 572 if (p <= bEnd-16) { 573 const BYTE* const limit = bEnd - 16; 574 U32 v1 = state->v1; 575 U32 v2 = state->v2; 576 U32 v3 = state->v3; 577 U32 v4 = state->v4; 578 579 do { 580 v1 = XXH32_round(v1, XXH_readLE32(p, endian)); p+=4; 581 v2 = XXH32_round(v2, XXH_readLE32(p, endian)); p+=4; 582 v3 = XXH32_round(v3, XXH_readLE32(p, endian)); p+=4; 583 v4 = XXH32_round(v4, XXH_readLE32(p, endian)); p+=4; 584 } while (p<=limit); 585 586 state->v1 = v1; 587 state->v2 = v2; 588 state->v3 = v3; 589 state->v4 = v4; 590 } 591 592 if (p < bEnd) { 593 XXH_memcpy(state->mem32, p, (size_t)(bEnd-p)); 594 state->memsize = (unsigned)(bEnd-p); 595 } 596 597 return XXH_OK; 598 } 599 600 XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* state_in, const void* input, size_t len) 601 { 602 XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; 603 604 if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) 605 return XXH32_update_endian(state_in, input, len, XXH_littleEndian); 606 else 607 return XXH32_update_endian(state_in, input, len, XXH_bigEndian); 608 } 609 610 611 612 FORCE_INLINE_TEMPLATE U32 XXH32_digest_endian (const XXH32_state_t* state, XXH_endianess endian) 613 { 614 const BYTE * p = (const BYTE*)state->mem32; 615 const BYTE* const bEnd = (const BYTE*)(state->mem32) + state->memsize; 616 U32 h32; 617 618 if (state->large_len) { 619 h32 = XXH_rotl32(state->v1, 1) + XXH_rotl32(state->v2, 7) + XXH_rotl32(state->v3, 12) + XXH_rotl32(state->v4, 18); 620 } else { 621 h32 = state->v3 /* == seed */ + PRIME32_5; 622 } 623 624 h32 += state->total_len_32; 625 626 while (p+4<=bEnd) { 627 h32 += XXH_readLE32(p, endian) * PRIME32_3; 628 h32 = XXH_rotl32(h32, 17) * PRIME32_4; 629 p+=4; 630 } 631 632 while (p<bEnd) { 633 h32 += (*p) * PRIME32_5; 634 h32 = XXH_rotl32(h32, 11) * PRIME32_1; 635 p++; 636 } 637 638 h32 ^= h32 >> 15; 639 h32 *= PRIME32_2; 640 h32 ^= h32 >> 13; 641 h32 *= PRIME32_3; 642 h32 ^= h32 >> 16; 643 644 return h32; 645 } 646 647 648 XXH_PUBLIC_API unsigned int XXH32_digest (const XXH32_state_t* state_in) 649 { 650 XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; 651 652 if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) 653 return XXH32_digest_endian(state_in, XXH_littleEndian); 654 else 655 return XXH32_digest_endian(state_in, XXH_bigEndian); 656 } 657 658 659 660 /* **** XXH64 **** */ 661 662 FORCE_INLINE_TEMPLATE XXH_errorcode XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_endianess endian) 663 { 664 const BYTE* p = (const BYTE*)input; 665 const BYTE* const bEnd = p + len; 666 667 #ifdef XXH_ACCEPT_NULL_INPUT_POINTER 668 if (input==NULL) return XXH_ERROR; 669 #endif 670 671 state->total_len += len; 672 673 if (state->memsize + len < 32) { /* fill in tmp buffer */ 674 if (input != NULL) { 675 XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, len); 676 } 677 state->memsize += (U32)len; 678 return XXH_OK; 679 } 680 681 if (state->memsize) { /* tmp buffer is full */ 682 XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, 32-state->memsize); 683 state->v1 = XXH64_round(state->v1, XXH_readLE64(state->mem64+0, endian)); 684 state->v2 = XXH64_round(state->v2, XXH_readLE64(state->mem64+1, endian)); 685 state->v3 = XXH64_round(state->v3, XXH_readLE64(state->mem64+2, endian)); 686 state->v4 = XXH64_round(state->v4, XXH_readLE64(state->mem64+3, endian)); 687 p += 32-state->memsize; 688 state->memsize = 0; 689 } 690 691 if (p+32 <= bEnd) { 692 const BYTE* const limit = bEnd - 32; 693 U64 v1 = state->v1; 694 U64 v2 = state->v2; 695 U64 v3 = state->v3; 696 U64 v4 = state->v4; 697 698 do { 699 v1 = XXH64_round(v1, XXH_readLE64(p, endian)); p+=8; 700 v2 = XXH64_round(v2, XXH_readLE64(p, endian)); p+=8; 701 v3 = XXH64_round(v3, XXH_readLE64(p, endian)); p+=8; 702 v4 = XXH64_round(v4, XXH_readLE64(p, endian)); p+=8; 703 } while (p<=limit); 704 705 state->v1 = v1; 706 state->v2 = v2; 707 state->v3 = v3; 708 state->v4 = v4; 709 } 710 711 if (p < bEnd) { 712 XXH_memcpy(state->mem64, p, (size_t)(bEnd-p)); 713 state->memsize = (unsigned)(bEnd-p); 714 } 715 716 return XXH_OK; 717 } 718 719 XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH64_state_t* state_in, const void* input, size_t len) 720 { 721 XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; 722 723 if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) 724 return XXH64_update_endian(state_in, input, len, XXH_littleEndian); 725 else 726 return XXH64_update_endian(state_in, input, len, XXH_bigEndian); 727 } 728 729 730 731 FORCE_INLINE_TEMPLATE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess endian) 732 { 733 const BYTE * p = (const BYTE*)state->mem64; 734 const BYTE* const bEnd = (const BYTE*)state->mem64 + state->memsize; 735 U64 h64; 736 737 if (state->total_len >= 32) { 738 U64 const v1 = state->v1; 739 U64 const v2 = state->v2; 740 U64 const v3 = state->v3; 741 U64 const v4 = state->v4; 742 743 h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18); 744 h64 = XXH64_mergeRound(h64, v1); 745 h64 = XXH64_mergeRound(h64, v2); 746 h64 = XXH64_mergeRound(h64, v3); 747 h64 = XXH64_mergeRound(h64, v4); 748 } else { 749 h64 = state->v3 + PRIME64_5; 750 } 751 752 h64 += (U64) state->total_len; 753 754 while (p+8<=bEnd) { 755 U64 const k1 = XXH64_round(0, XXH_readLE64(p, endian)); 756 h64 ^= k1; 757 h64 = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4; 758 p+=8; 759 } 760 761 if (p+4<=bEnd) { 762 h64 ^= (U64)(XXH_readLE32(p, endian)) * PRIME64_1; 763 h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3; 764 p+=4; 765 } 766 767 while (p<bEnd) { 768 h64 ^= (*p) * PRIME64_5; 769 h64 = XXH_rotl64(h64, 11) * PRIME64_1; 770 p++; 771 } 772 773 h64 ^= h64 >> 33; 774 h64 *= PRIME64_2; 775 h64 ^= h64 >> 29; 776 h64 *= PRIME64_3; 777 h64 ^= h64 >> 32; 778 779 return h64; 780 } 781 782 783 XXH_PUBLIC_API unsigned long long XXH64_digest (const XXH64_state_t* state_in) 784 { 785 XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; 786 787 if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) 788 return XXH64_digest_endian(state_in, XXH_littleEndian); 789 else 790 return XXH64_digest_endian(state_in, XXH_bigEndian); 791 } 792 793 794 /* ************************** 795 * Canonical representation 796 ****************************/ 797 798 /*! Default XXH result types are basic unsigned 32 and 64 bits. 799 * The canonical representation follows human-readable write convention, aka big-endian (large digits first). 800 * These functions allow transformation of hash result into and from its canonical format. 801 * This way, hash values can be written into a file or buffer, and remain comparable across different systems and programs. 802 */ 803 804 XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash) 805 { 806 XXH_STATIC_ASSERT(sizeof(XXH32_canonical_t) == sizeof(XXH32_hash_t)); 807 if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap32(hash); 808 ZSTD_memcpy(dst, &hash, sizeof(*dst)); 809 } 810 811 XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH64_canonical_t* dst, XXH64_hash_t hash) 812 { 813 XXH_STATIC_ASSERT(sizeof(XXH64_canonical_t) == sizeof(XXH64_hash_t)); 814 if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap64(hash); 815 ZSTD_memcpy(dst, &hash, sizeof(*dst)); 816 } 817 818 XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src) 819 { 820 return XXH_readBE32(src); 821 } 822 823 XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(const XXH64_canonical_t* src) 824 { 825 return XXH_readBE64(src); 826 } 827