Scott Baker | eee8dd8 | 2019-09-24 12:52:34 -0700 | [diff] [blame] | 1 | /* |
| 2 | * xxHash - Fast Hash algorithm |
| 3 | * Copyright (C) 2012-2016, Yann Collet |
| 4 | * |
| 5 | * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) |
| 6 | * |
| 7 | * Redistribution and use in source and binary forms, with or without |
| 8 | * modification, are permitted provided that the following conditions are |
| 9 | * met: |
| 10 | * |
| 11 | * * Redistributions of source code must retain the above copyright |
| 12 | * notice, this list of conditions and the following disclaimer. |
| 13 | * * Redistributions in binary form must reproduce the above |
| 14 | * copyright notice, this list of conditions and the following disclaimer |
| 15 | * in the documentation and/or other materials provided with the |
| 16 | * distribution. |
| 17 | * |
| 18 | * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS |
| 19 | * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
| 20 | * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR |
| 21 | * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT |
| 22 | * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, |
| 23 | * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT |
| 24 | * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, |
| 25 | * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY |
| 26 | * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
| 27 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
| 28 | * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 29 | * |
| 30 | * You can contact the author at : |
| 31 | * - xxHash homepage: http://www.xxhash.com |
| 32 | * - xxHash source repository : https://github.com/Cyan4973/xxHash |
| 33 | */ |
| 34 | |
| 35 | |
| 36 | /* ************************************* |
| 37 | * Tuning parameters |
| 38 | ***************************************/ |
| 39 | /*!XXH_FORCE_MEMORY_ACCESS : |
| 40 | * By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable. |
| 41 | * Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal. |
| 42 | * The below switch allow to select different access method for improved performance. |
| 43 | * Method 0 (default) : use `memcpy()`. Safe and portable. |
| 44 | * Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable). |
| 45 | * This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`. |
| 46 | * Method 2 : direct access. This method doesn't depend on compiler but violate C standard. |
| 47 | * It can generate buggy code on targets which do not support unaligned memory accesses. |
| 48 | * But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6) |
| 49 | * See http://stackoverflow.com/a/32095106/646947 for details. |
| 50 | * Prefer these methods in priority order (0 > 1 > 2) |
| 51 | */ |
| 52 | #ifndef XXH_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ |
| 53 | # 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__) ) |
| 54 | # define XXH_FORCE_MEMORY_ACCESS 2 |
| 55 | # elif (defined(__INTEL_COMPILER) && !defined(WIN32)) || \ |
| 56 | (defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) )) |
| 57 | # define XXH_FORCE_MEMORY_ACCESS 1 |
| 58 | # endif |
| 59 | #endif |
| 60 | |
| 61 | /*!XXH_ACCEPT_NULL_INPUT_POINTER : |
| 62 | * If the input pointer is a null pointer, xxHash default behavior is to trigger a memory access error, since it is a bad pointer. |
| 63 | * When this option is enabled, xxHash output for null input pointers will be the same as a null-length input. |
| 64 | * By default, this option is disabled. To enable it, uncomment below define : |
| 65 | */ |
| 66 | /* #define XXH_ACCEPT_NULL_INPUT_POINTER 1 */ |
| 67 | |
| 68 | /*!XXH_FORCE_NATIVE_FORMAT : |
Scott Baker | 611f6bd | 2019-10-18 13:45:19 -0700 | [diff] [blame^] | 69 | * By default, xxHash library provides endian-independent Hash values, based on little-endian convention. |
Scott Baker | eee8dd8 | 2019-09-24 12:52:34 -0700 | [diff] [blame] | 70 | * Results are therefore identical for little-endian and big-endian CPU. |
| 71 | * This comes at a performance cost for big-endian CPU, since some swapping is required to emulate little-endian format. |
Scott Baker | 611f6bd | 2019-10-18 13:45:19 -0700 | [diff] [blame^] | 72 | * Should endian-independence be of no importance for your application, you may set the #define below to 1, |
Scott Baker | eee8dd8 | 2019-09-24 12:52:34 -0700 | [diff] [blame] | 73 | * to improve speed for Big-endian CPU. |
| 74 | * This option has no impact on Little_Endian CPU. |
| 75 | */ |
| 76 | #ifndef XXH_FORCE_NATIVE_FORMAT /* can be defined externally */ |
| 77 | # define XXH_FORCE_NATIVE_FORMAT 0 |
| 78 | #endif |
| 79 | |
| 80 | /*!XXH_FORCE_ALIGN_CHECK : |
| 81 | * This is a minor performance trick, only useful with lots of very small keys. |
| 82 | * It means : check for aligned/unaligned input. |
| 83 | * The check costs one initial branch per hash; set to 0 when the input data |
| 84 | * is guaranteed to be aligned. |
| 85 | */ |
| 86 | #ifndef XXH_FORCE_ALIGN_CHECK /* can be defined externally */ |
| 87 | # if defined(__i386) || defined(_M_IX86) || defined(__x86_64__) || defined(_M_X64) |
| 88 | # define XXH_FORCE_ALIGN_CHECK 0 |
| 89 | # else |
| 90 | # define XXH_FORCE_ALIGN_CHECK 1 |
| 91 | # endif |
| 92 | #endif |
| 93 | |
| 94 | |
| 95 | /* ************************************* |
| 96 | * Includes & Memory related functions |
| 97 | ***************************************/ |
| 98 | /* Modify the local functions below should you wish to use some other memory routines */ |
| 99 | /* for malloc(), free() */ |
| 100 | #include <stdlib.h> |
Scott Baker | 611f6bd | 2019-10-18 13:45:19 -0700 | [diff] [blame^] | 101 | #include <stddef.h> /* size_t */ |
Scott Baker | eee8dd8 | 2019-09-24 12:52:34 -0700 | [diff] [blame] | 102 | static void* XXH_malloc(size_t s) { return malloc(s); } |
| 103 | static void XXH_free (void* p) { free(p); } |
| 104 | /* for memcpy() */ |
| 105 | #include <string.h> |
| 106 | static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcpy(dest,src,size); } |
| 107 | |
| 108 | #ifndef XXH_STATIC_LINKING_ONLY |
| 109 | # define XXH_STATIC_LINKING_ONLY |
| 110 | #endif |
| 111 | #include "xxhash.h" |
| 112 | |
| 113 | |
| 114 | /* ************************************* |
| 115 | * Compiler Specific Options |
| 116 | ***************************************/ |
| 117 | #if defined (__GNUC__) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */ |
| 118 | # define INLINE_KEYWORD inline |
| 119 | #else |
| 120 | # define INLINE_KEYWORD |
| 121 | #endif |
| 122 | |
| 123 | #if defined(__GNUC__) |
| 124 | # define FORCE_INLINE_ATTR __attribute__((always_inline)) |
| 125 | #elif defined(_MSC_VER) |
| 126 | # define FORCE_INLINE_ATTR __forceinline |
| 127 | #else |
| 128 | # define FORCE_INLINE_ATTR |
| 129 | #endif |
| 130 | |
| 131 | #define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR |
| 132 | |
| 133 | |
| 134 | #ifdef _MSC_VER |
| 135 | # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ |
| 136 | #endif |
| 137 | |
| 138 | |
| 139 | /* ************************************* |
| 140 | * Basic Types |
| 141 | ***************************************/ |
| 142 | #ifndef MEM_MODULE |
| 143 | # define MEM_MODULE |
| 144 | # if !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) |
| 145 | # include <stdint.h> |
| 146 | typedef uint8_t BYTE; |
| 147 | typedef uint16_t U16; |
| 148 | typedef uint32_t U32; |
| 149 | typedef int32_t S32; |
| 150 | typedef uint64_t U64; |
| 151 | # else |
| 152 | typedef unsigned char BYTE; |
| 153 | typedef unsigned short U16; |
| 154 | typedef unsigned int U32; |
| 155 | typedef signed int S32; |
| 156 | typedef unsigned long long U64; /* if your compiler doesn't support unsigned long long, replace by another 64-bit type here. Note that xxhash.h will also need to be updated. */ |
| 157 | # endif |
| 158 | #endif |
| 159 | |
| 160 | |
| 161 | #if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) |
| 162 | |
| 163 | /* Force direct memory access. Only works on CPU which support unaligned memory access in hardware */ |
| 164 | static U32 XXH_read32(const void* memPtr) { return *(const U32*) memPtr; } |
| 165 | static U64 XXH_read64(const void* memPtr) { return *(const U64*) memPtr; } |
| 166 | |
| 167 | #elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) |
| 168 | |
| 169 | /* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */ |
| 170 | /* currently only defined for gcc and icc */ |
| 171 | typedef union { U32 u32; U64 u64; } __attribute__((packed)) unalign; |
| 172 | |
| 173 | static U32 XXH_read32(const void* ptr) { return ((const unalign*)ptr)->u32; } |
| 174 | static U64 XXH_read64(const void* ptr) { return ((const unalign*)ptr)->u64; } |
| 175 | |
| 176 | #else |
| 177 | |
| 178 | /* portable and safe solution. Generally efficient. |
| 179 | * see : http://stackoverflow.com/a/32095106/646947 |
| 180 | */ |
| 181 | |
| 182 | static U32 XXH_read32(const void* memPtr) |
| 183 | { |
| 184 | U32 val; |
| 185 | memcpy(&val, memPtr, sizeof(val)); |
| 186 | return val; |
| 187 | } |
| 188 | |
| 189 | static U64 XXH_read64(const void* memPtr) |
| 190 | { |
| 191 | U64 val; |
| 192 | memcpy(&val, memPtr, sizeof(val)); |
| 193 | return val; |
| 194 | } |
| 195 | |
| 196 | #endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ |
| 197 | |
| 198 | |
| 199 | /* **************************************** |
| 200 | * Compiler-specific Functions and Macros |
| 201 | ******************************************/ |
| 202 | #define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) |
| 203 | |
| 204 | /* Note : although _rotl exists for minGW (GCC under windows), performance seems poor */ |
| 205 | #if defined(_MSC_VER) |
| 206 | # define XXH_rotl32(x,r) _rotl(x,r) |
| 207 | # define XXH_rotl64(x,r) _rotl64(x,r) |
| 208 | #else |
| 209 | # define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r))) |
| 210 | # define XXH_rotl64(x,r) ((x << r) | (x >> (64 - r))) |
| 211 | #endif |
| 212 | |
| 213 | #if defined(_MSC_VER) /* Visual Studio */ |
| 214 | # define XXH_swap32 _byteswap_ulong |
| 215 | # define XXH_swap64 _byteswap_uint64 |
| 216 | #elif GCC_VERSION >= 403 |
| 217 | # define XXH_swap32 __builtin_bswap32 |
| 218 | # define XXH_swap64 __builtin_bswap64 |
| 219 | #else |
| 220 | static U32 XXH_swap32 (U32 x) |
| 221 | { |
| 222 | return ((x << 24) & 0xff000000 ) | |
| 223 | ((x << 8) & 0x00ff0000 ) | |
| 224 | ((x >> 8) & 0x0000ff00 ) | |
| 225 | ((x >> 24) & 0x000000ff ); |
| 226 | } |
| 227 | static U64 XXH_swap64 (U64 x) |
| 228 | { |
| 229 | return ((x << 56) & 0xff00000000000000ULL) | |
| 230 | ((x << 40) & 0x00ff000000000000ULL) | |
| 231 | ((x << 24) & 0x0000ff0000000000ULL) | |
| 232 | ((x << 8) & 0x000000ff00000000ULL) | |
| 233 | ((x >> 8) & 0x00000000ff000000ULL) | |
| 234 | ((x >> 24) & 0x0000000000ff0000ULL) | |
| 235 | ((x >> 40) & 0x000000000000ff00ULL) | |
| 236 | ((x >> 56) & 0x00000000000000ffULL); |
| 237 | } |
| 238 | #endif |
| 239 | |
| 240 | |
| 241 | /* ************************************* |
| 242 | * Architecture Macros |
| 243 | ***************************************/ |
| 244 | typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess; |
| 245 | |
| 246 | /* XXH_CPU_LITTLE_ENDIAN can be defined externally, for example on the compiler command line */ |
| 247 | #ifndef XXH_CPU_LITTLE_ENDIAN |
| 248 | static const int g_one = 1; |
| 249 | # define XXH_CPU_LITTLE_ENDIAN (*(const char*)(&g_one)) |
| 250 | #endif |
| 251 | |
| 252 | |
| 253 | /* *************************** |
| 254 | * Memory reads |
| 255 | *****************************/ |
| 256 | typedef enum { XXH_aligned, XXH_unaligned } XXH_alignment; |
| 257 | |
| 258 | FORCE_INLINE_TEMPLATE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, XXH_alignment align) |
| 259 | { |
| 260 | if (align==XXH_unaligned) |
| 261 | return endian==XXH_littleEndian ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr)); |
| 262 | else |
| 263 | return endian==XXH_littleEndian ? *(const U32*)ptr : XXH_swap32(*(const U32*)ptr); |
| 264 | } |
| 265 | |
| 266 | FORCE_INLINE_TEMPLATE U32 XXH_readLE32(const void* ptr, XXH_endianess endian) |
| 267 | { |
| 268 | return XXH_readLE32_align(ptr, endian, XXH_unaligned); |
| 269 | } |
| 270 | |
| 271 | static U32 XXH_readBE32(const void* ptr) |
| 272 | { |
| 273 | return XXH_CPU_LITTLE_ENDIAN ? XXH_swap32(XXH_read32(ptr)) : XXH_read32(ptr); |
| 274 | } |
| 275 | |
| 276 | FORCE_INLINE_TEMPLATE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, XXH_alignment align) |
| 277 | { |
| 278 | if (align==XXH_unaligned) |
| 279 | return endian==XXH_littleEndian ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr)); |
| 280 | else |
| 281 | return endian==XXH_littleEndian ? *(const U64*)ptr : XXH_swap64(*(const U64*)ptr); |
| 282 | } |
| 283 | |
| 284 | FORCE_INLINE_TEMPLATE U64 XXH_readLE64(const void* ptr, XXH_endianess endian) |
| 285 | { |
| 286 | return XXH_readLE64_align(ptr, endian, XXH_unaligned); |
| 287 | } |
| 288 | |
| 289 | static U64 XXH_readBE64(const void* ptr) |
| 290 | { |
| 291 | return XXH_CPU_LITTLE_ENDIAN ? XXH_swap64(XXH_read64(ptr)) : XXH_read64(ptr); |
| 292 | } |
| 293 | |
| 294 | |
| 295 | /* ************************************* |
| 296 | * Macros |
| 297 | ***************************************/ |
| 298 | #define XXH_STATIC_ASSERT(c) { enum { XXH_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */ |
| 299 | |
| 300 | |
| 301 | /* ************************************* |
| 302 | * Constants |
| 303 | ***************************************/ |
| 304 | static const U32 PRIME32_1 = 2654435761U; |
| 305 | static const U32 PRIME32_2 = 2246822519U; |
| 306 | static const U32 PRIME32_3 = 3266489917U; |
| 307 | static const U32 PRIME32_4 = 668265263U; |
| 308 | static const U32 PRIME32_5 = 374761393U; |
| 309 | |
| 310 | static const U64 PRIME64_1 = 11400714785074694791ULL; |
| 311 | static const U64 PRIME64_2 = 14029467366897019727ULL; |
| 312 | static const U64 PRIME64_3 = 1609587929392839161ULL; |
| 313 | static const U64 PRIME64_4 = 9650029242287828579ULL; |
| 314 | static const U64 PRIME64_5 = 2870177450012600261ULL; |
| 315 | |
| 316 | XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; } |
| 317 | |
| 318 | |
| 319 | /* ************************** |
| 320 | * Utils |
| 321 | ****************************/ |
| 322 | XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* restrict dstState, const XXH32_state_t* restrict srcState) |
| 323 | { |
| 324 | memcpy(dstState, srcState, sizeof(*dstState)); |
| 325 | } |
| 326 | |
| 327 | XXH_PUBLIC_API void XXH64_copyState(XXH64_state_t* restrict dstState, const XXH64_state_t* restrict srcState) |
| 328 | { |
| 329 | memcpy(dstState, srcState, sizeof(*dstState)); |
| 330 | } |
| 331 | |
| 332 | |
| 333 | /* *************************** |
| 334 | * Simple Hash Functions |
| 335 | *****************************/ |
| 336 | |
| 337 | static U32 XXH32_round(U32 seed, U32 input) |
| 338 | { |
| 339 | seed += input * PRIME32_2; |
| 340 | seed = XXH_rotl32(seed, 13); |
| 341 | seed *= PRIME32_1; |
| 342 | return seed; |
| 343 | } |
| 344 | |
| 345 | FORCE_INLINE_TEMPLATE U32 XXH32_endian_align(const void* input, size_t len, U32 seed, XXH_endianess endian, XXH_alignment align) |
| 346 | { |
| 347 | const BYTE* p = (const BYTE*)input; |
| 348 | const BYTE* bEnd = p + len; |
| 349 | U32 h32; |
| 350 | #define XXH_get32bits(p) XXH_readLE32_align(p, endian, align) |
| 351 | |
| 352 | #ifdef XXH_ACCEPT_NULL_INPUT_POINTER |
| 353 | if (p==NULL) { |
| 354 | len=0; |
| 355 | bEnd=p=(const BYTE*)(size_t)16; |
| 356 | } |
| 357 | #endif |
| 358 | |
| 359 | if (len>=16) { |
| 360 | const BYTE* const limit = bEnd - 16; |
| 361 | U32 v1 = seed + PRIME32_1 + PRIME32_2; |
| 362 | U32 v2 = seed + PRIME32_2; |
| 363 | U32 v3 = seed + 0; |
| 364 | U32 v4 = seed - PRIME32_1; |
| 365 | |
| 366 | do { |
| 367 | v1 = XXH32_round(v1, XXH_get32bits(p)); p+=4; |
| 368 | v2 = XXH32_round(v2, XXH_get32bits(p)); p+=4; |
| 369 | v3 = XXH32_round(v3, XXH_get32bits(p)); p+=4; |
| 370 | v4 = XXH32_round(v4, XXH_get32bits(p)); p+=4; |
| 371 | } while (p<=limit); |
| 372 | |
| 373 | h32 = XXH_rotl32(v1, 1) + XXH_rotl32(v2, 7) + XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18); |
| 374 | } else { |
| 375 | h32 = seed + PRIME32_5; |
| 376 | } |
| 377 | |
| 378 | h32 += (U32) len; |
| 379 | |
| 380 | while (p+4<=bEnd) { |
| 381 | h32 += XXH_get32bits(p) * PRIME32_3; |
| 382 | h32 = XXH_rotl32(h32, 17) * PRIME32_4 ; |
| 383 | p+=4; |
| 384 | } |
| 385 | |
| 386 | while (p<bEnd) { |
| 387 | h32 += (*p) * PRIME32_5; |
| 388 | h32 = XXH_rotl32(h32, 11) * PRIME32_1 ; |
| 389 | p++; |
| 390 | } |
| 391 | |
| 392 | h32 ^= h32 >> 15; |
| 393 | h32 *= PRIME32_2; |
| 394 | h32 ^= h32 >> 13; |
| 395 | h32 *= PRIME32_3; |
| 396 | h32 ^= h32 >> 16; |
| 397 | |
| 398 | return h32; |
| 399 | } |
| 400 | |
| 401 | |
| 402 | XXH_PUBLIC_API unsigned int XXH32 (const void* input, size_t len, unsigned int seed) |
| 403 | { |
| 404 | #if 0 |
| 405 | /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ |
| 406 | XXH32_CREATESTATE_STATIC(state); |
| 407 | XXH32_reset(state, seed); |
| 408 | XXH32_update(state, input, len); |
| 409 | return XXH32_digest(state); |
| 410 | #else |
| 411 | XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; |
| 412 | |
| 413 | if (XXH_FORCE_ALIGN_CHECK) { |
| 414 | if ((((size_t)input) & 3) == 0) { /* Input is 4-bytes aligned, leverage the speed benefit */ |
| 415 | if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) |
| 416 | return XXH32_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned); |
| 417 | else |
| 418 | return XXH32_endian_align(input, len, seed, XXH_bigEndian, XXH_aligned); |
| 419 | } } |
| 420 | |
| 421 | if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) |
| 422 | return XXH32_endian_align(input, len, seed, XXH_littleEndian, XXH_unaligned); |
| 423 | else |
| 424 | return XXH32_endian_align(input, len, seed, XXH_bigEndian, XXH_unaligned); |
| 425 | #endif |
| 426 | } |
| 427 | |
| 428 | |
| 429 | static U64 XXH64_round(U64 acc, U64 input) |
| 430 | { |
| 431 | acc += input * PRIME64_2; |
| 432 | acc = XXH_rotl64(acc, 31); |
| 433 | acc *= PRIME64_1; |
| 434 | return acc; |
| 435 | } |
| 436 | |
| 437 | static U64 XXH64_mergeRound(U64 acc, U64 val) |
| 438 | { |
| 439 | val = XXH64_round(0, val); |
| 440 | acc ^= val; |
| 441 | acc = acc * PRIME64_1 + PRIME64_4; |
| 442 | return acc; |
| 443 | } |
| 444 | |
| 445 | FORCE_INLINE_TEMPLATE U64 XXH64_endian_align(const void* input, size_t len, U64 seed, XXH_endianess endian, XXH_alignment align) |
| 446 | { |
| 447 | const BYTE* p = (const BYTE*)input; |
| 448 | const BYTE* const bEnd = p + len; |
| 449 | U64 h64; |
| 450 | #define XXH_get64bits(p) XXH_readLE64_align(p, endian, align) |
| 451 | |
| 452 | #ifdef XXH_ACCEPT_NULL_INPUT_POINTER |
| 453 | if (p==NULL) { |
| 454 | len=0; |
| 455 | bEnd=p=(const BYTE*)(size_t)32; |
| 456 | } |
| 457 | #endif |
| 458 | |
| 459 | if (len>=32) { |
| 460 | const BYTE* const limit = bEnd - 32; |
| 461 | U64 v1 = seed + PRIME64_1 + PRIME64_2; |
| 462 | U64 v2 = seed + PRIME64_2; |
| 463 | U64 v3 = seed + 0; |
| 464 | U64 v4 = seed - PRIME64_1; |
| 465 | |
| 466 | do { |
| 467 | v1 = XXH64_round(v1, XXH_get64bits(p)); p+=8; |
| 468 | v2 = XXH64_round(v2, XXH_get64bits(p)); p+=8; |
| 469 | v3 = XXH64_round(v3, XXH_get64bits(p)); p+=8; |
| 470 | v4 = XXH64_round(v4, XXH_get64bits(p)); p+=8; |
| 471 | } while (p<=limit); |
| 472 | |
| 473 | h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18); |
| 474 | h64 = XXH64_mergeRound(h64, v1); |
| 475 | h64 = XXH64_mergeRound(h64, v2); |
| 476 | h64 = XXH64_mergeRound(h64, v3); |
| 477 | h64 = XXH64_mergeRound(h64, v4); |
| 478 | |
| 479 | } else { |
| 480 | h64 = seed + PRIME64_5; |
| 481 | } |
| 482 | |
| 483 | h64 += (U64) len; |
| 484 | |
| 485 | while (p+8<=bEnd) { |
| 486 | U64 const k1 = XXH64_round(0, XXH_get64bits(p)); |
| 487 | h64 ^= k1; |
| 488 | h64 = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4; |
| 489 | p+=8; |
| 490 | } |
| 491 | |
| 492 | if (p+4<=bEnd) { |
| 493 | h64 ^= (U64)(XXH_get32bits(p)) * PRIME64_1; |
| 494 | h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3; |
| 495 | p+=4; |
| 496 | } |
| 497 | |
| 498 | while (p<bEnd) { |
| 499 | h64 ^= (*p) * PRIME64_5; |
| 500 | h64 = XXH_rotl64(h64, 11) * PRIME64_1; |
| 501 | p++; |
| 502 | } |
| 503 | |
| 504 | h64 ^= h64 >> 33; |
| 505 | h64 *= PRIME64_2; |
| 506 | h64 ^= h64 >> 29; |
| 507 | h64 *= PRIME64_3; |
| 508 | h64 ^= h64 >> 32; |
| 509 | |
| 510 | return h64; |
| 511 | } |
| 512 | |
| 513 | |
| 514 | XXH_PUBLIC_API unsigned long long XXH64 (const void* input, size_t len, unsigned long long seed) |
| 515 | { |
| 516 | #if 0 |
| 517 | /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ |
| 518 | XXH64_CREATESTATE_STATIC(state); |
| 519 | XXH64_reset(state, seed); |
| 520 | XXH64_update(state, input, len); |
| 521 | return XXH64_digest(state); |
| 522 | #else |
| 523 | XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; |
| 524 | |
| 525 | if (XXH_FORCE_ALIGN_CHECK) { |
| 526 | if ((((size_t)input) & 7)==0) { /* Input is aligned, let's leverage the speed advantage */ |
| 527 | if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) |
| 528 | return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned); |
| 529 | else |
| 530 | return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_aligned); |
| 531 | } } |
| 532 | |
| 533 | if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) |
| 534 | return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_unaligned); |
| 535 | else |
| 536 | return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_unaligned); |
| 537 | #endif |
| 538 | } |
| 539 | |
| 540 | |
| 541 | /* ************************************************** |
| 542 | * Advanced Hash Functions |
| 543 | ****************************************************/ |
| 544 | |
| 545 | XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void) |
| 546 | { |
| 547 | return (XXH32_state_t*)XXH_malloc(sizeof(XXH32_state_t)); |
| 548 | } |
| 549 | XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr) |
| 550 | { |
| 551 | XXH_free(statePtr); |
| 552 | return XXH_OK; |
| 553 | } |
| 554 | |
| 555 | XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void) |
| 556 | { |
| 557 | return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t)); |
| 558 | } |
| 559 | XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr) |
| 560 | { |
| 561 | XXH_free(statePtr); |
| 562 | return XXH_OK; |
| 563 | } |
| 564 | |
| 565 | |
| 566 | /*** Hash feed ***/ |
| 567 | |
| 568 | XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, unsigned int seed) |
| 569 | { |
| 570 | XXH32_state_t state; /* using a local state to memcpy() in order to avoid strict-aliasing warnings */ |
| 571 | memset(&state, 0, sizeof(state)-4); /* do not write into reserved, for future removal */ |
| 572 | state.v1 = seed + PRIME32_1 + PRIME32_2; |
| 573 | state.v2 = seed + PRIME32_2; |
| 574 | state.v3 = seed + 0; |
| 575 | state.v4 = seed - PRIME32_1; |
| 576 | memcpy(statePtr, &state, sizeof(state)); |
| 577 | return XXH_OK; |
| 578 | } |
| 579 | |
| 580 | |
| 581 | XXH_PUBLIC_API XXH_errorcode XXH64_reset(XXH64_state_t* statePtr, unsigned long long seed) |
| 582 | { |
| 583 | XXH64_state_t state; /* using a local state to memcpy() in order to avoid strict-aliasing warnings */ |
| 584 | memset(&state, 0, sizeof(state)-8); /* do not write into reserved, for future removal */ |
| 585 | state.v1 = seed + PRIME64_1 + PRIME64_2; |
| 586 | state.v2 = seed + PRIME64_2; |
| 587 | state.v3 = seed + 0; |
| 588 | state.v4 = seed - PRIME64_1; |
| 589 | memcpy(statePtr, &state, sizeof(state)); |
| 590 | return XXH_OK; |
| 591 | } |
| 592 | |
| 593 | |
| 594 | FORCE_INLINE_TEMPLATE XXH_errorcode XXH32_update_endian (XXH32_state_t* state, const void* input, size_t len, XXH_endianess endian) |
| 595 | { |
| 596 | const BYTE* p = (const BYTE*)input; |
| 597 | const BYTE* const bEnd = p + len; |
| 598 | |
| 599 | #ifdef XXH_ACCEPT_NULL_INPUT_POINTER |
| 600 | if (input==NULL) return XXH_ERROR; |
| 601 | #endif |
| 602 | |
| 603 | state->total_len_32 += (unsigned)len; |
| 604 | state->large_len |= (len>=16) | (state->total_len_32>=16); |
| 605 | |
| 606 | if (state->memsize + len < 16) { /* fill in tmp buffer */ |
| 607 | XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, len); |
| 608 | state->memsize += (unsigned)len; |
| 609 | return XXH_OK; |
| 610 | } |
| 611 | |
| 612 | if (state->memsize) { /* some data left from previous update */ |
| 613 | XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, 16-state->memsize); |
| 614 | { const U32* p32 = state->mem32; |
| 615 | state->v1 = XXH32_round(state->v1, XXH_readLE32(p32, endian)); p32++; |
| 616 | state->v2 = XXH32_round(state->v2, XXH_readLE32(p32, endian)); p32++; |
| 617 | state->v3 = XXH32_round(state->v3, XXH_readLE32(p32, endian)); p32++; |
| 618 | state->v4 = XXH32_round(state->v4, XXH_readLE32(p32, endian)); p32++; |
| 619 | } |
| 620 | p += 16-state->memsize; |
| 621 | state->memsize = 0; |
| 622 | } |
| 623 | |
| 624 | if (p <= bEnd-16) { |
| 625 | const BYTE* const limit = bEnd - 16; |
| 626 | U32 v1 = state->v1; |
| 627 | U32 v2 = state->v2; |
| 628 | U32 v3 = state->v3; |
| 629 | U32 v4 = state->v4; |
| 630 | |
| 631 | do { |
| 632 | v1 = XXH32_round(v1, XXH_readLE32(p, endian)); p+=4; |
| 633 | v2 = XXH32_round(v2, XXH_readLE32(p, endian)); p+=4; |
| 634 | v3 = XXH32_round(v3, XXH_readLE32(p, endian)); p+=4; |
| 635 | v4 = XXH32_round(v4, XXH_readLE32(p, endian)); p+=4; |
| 636 | } while (p<=limit); |
| 637 | |
| 638 | state->v1 = v1; |
| 639 | state->v2 = v2; |
| 640 | state->v3 = v3; |
| 641 | state->v4 = v4; |
| 642 | } |
| 643 | |
| 644 | if (p < bEnd) { |
| 645 | XXH_memcpy(state->mem32, p, (size_t)(bEnd-p)); |
| 646 | state->memsize = (unsigned)(bEnd-p); |
| 647 | } |
| 648 | |
| 649 | return XXH_OK; |
| 650 | } |
| 651 | |
| 652 | XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* state_in, const void* input, size_t len) |
| 653 | { |
| 654 | XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; |
| 655 | |
| 656 | if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) |
| 657 | return XXH32_update_endian(state_in, input, len, XXH_littleEndian); |
| 658 | else |
| 659 | return XXH32_update_endian(state_in, input, len, XXH_bigEndian); |
| 660 | } |
| 661 | |
| 662 | |
| 663 | |
| 664 | FORCE_INLINE_TEMPLATE U32 XXH32_digest_endian (const XXH32_state_t* state, XXH_endianess endian) |
| 665 | { |
| 666 | const BYTE * p = (const BYTE*)state->mem32; |
| 667 | const BYTE* const bEnd = (const BYTE*)(state->mem32) + state->memsize; |
| 668 | U32 h32; |
| 669 | |
| 670 | if (state->large_len) { |
| 671 | h32 = XXH_rotl32(state->v1, 1) + XXH_rotl32(state->v2, 7) + XXH_rotl32(state->v3, 12) + XXH_rotl32(state->v4, 18); |
| 672 | } else { |
| 673 | h32 = state->v3 /* == seed */ + PRIME32_5; |
| 674 | } |
| 675 | |
| 676 | h32 += state->total_len_32; |
| 677 | |
| 678 | while (p+4<=bEnd) { |
| 679 | h32 += XXH_readLE32(p, endian) * PRIME32_3; |
| 680 | h32 = XXH_rotl32(h32, 17) * PRIME32_4; |
| 681 | p+=4; |
| 682 | } |
| 683 | |
| 684 | while (p<bEnd) { |
| 685 | h32 += (*p) * PRIME32_5; |
| 686 | h32 = XXH_rotl32(h32, 11) * PRIME32_1; |
| 687 | p++; |
| 688 | } |
| 689 | |
| 690 | h32 ^= h32 >> 15; |
| 691 | h32 *= PRIME32_2; |
| 692 | h32 ^= h32 >> 13; |
| 693 | h32 *= PRIME32_3; |
| 694 | h32 ^= h32 >> 16; |
| 695 | |
| 696 | return h32; |
| 697 | } |
| 698 | |
| 699 | |
| 700 | XXH_PUBLIC_API unsigned int XXH32_digest (const XXH32_state_t* state_in) |
| 701 | { |
| 702 | XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; |
| 703 | |
| 704 | if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) |
| 705 | return XXH32_digest_endian(state_in, XXH_littleEndian); |
| 706 | else |
| 707 | return XXH32_digest_endian(state_in, XXH_bigEndian); |
| 708 | } |
| 709 | |
| 710 | |
| 711 | |
| 712 | /* **** XXH64 **** */ |
| 713 | |
| 714 | FORCE_INLINE_TEMPLATE XXH_errorcode XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_endianess endian) |
| 715 | { |
| 716 | const BYTE* p = (const BYTE*)input; |
| 717 | const BYTE* const bEnd = p + len; |
| 718 | |
| 719 | #ifdef XXH_ACCEPT_NULL_INPUT_POINTER |
| 720 | if (input==NULL) return XXH_ERROR; |
| 721 | #endif |
| 722 | |
| 723 | state->total_len += len; |
| 724 | |
| 725 | if (state->memsize + len < 32) { /* fill in tmp buffer */ |
| 726 | XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, len); |
| 727 | state->memsize += (U32)len; |
| 728 | return XXH_OK; |
| 729 | } |
| 730 | |
| 731 | if (state->memsize) { /* tmp buffer is full */ |
| 732 | XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, 32-state->memsize); |
| 733 | state->v1 = XXH64_round(state->v1, XXH_readLE64(state->mem64+0, endian)); |
| 734 | state->v2 = XXH64_round(state->v2, XXH_readLE64(state->mem64+1, endian)); |
| 735 | state->v3 = XXH64_round(state->v3, XXH_readLE64(state->mem64+2, endian)); |
| 736 | state->v4 = XXH64_round(state->v4, XXH_readLE64(state->mem64+3, endian)); |
| 737 | p += 32-state->memsize; |
| 738 | state->memsize = 0; |
| 739 | } |
| 740 | |
| 741 | if (p+32 <= bEnd) { |
| 742 | const BYTE* const limit = bEnd - 32; |
| 743 | U64 v1 = state->v1; |
| 744 | U64 v2 = state->v2; |
| 745 | U64 v3 = state->v3; |
| 746 | U64 v4 = state->v4; |
| 747 | |
| 748 | do { |
| 749 | v1 = XXH64_round(v1, XXH_readLE64(p, endian)); p+=8; |
| 750 | v2 = XXH64_round(v2, XXH_readLE64(p, endian)); p+=8; |
| 751 | v3 = XXH64_round(v3, XXH_readLE64(p, endian)); p+=8; |
| 752 | v4 = XXH64_round(v4, XXH_readLE64(p, endian)); p+=8; |
| 753 | } while (p<=limit); |
| 754 | |
| 755 | state->v1 = v1; |
| 756 | state->v2 = v2; |
| 757 | state->v3 = v3; |
| 758 | state->v4 = v4; |
| 759 | } |
| 760 | |
| 761 | if (p < bEnd) { |
| 762 | XXH_memcpy(state->mem64, p, (size_t)(bEnd-p)); |
| 763 | state->memsize = (unsigned)(bEnd-p); |
| 764 | } |
| 765 | |
| 766 | return XXH_OK; |
| 767 | } |
| 768 | |
| 769 | XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH64_state_t* state_in, const void* input, size_t len) |
| 770 | { |
| 771 | XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; |
| 772 | |
| 773 | if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) |
| 774 | return XXH64_update_endian(state_in, input, len, XXH_littleEndian); |
| 775 | else |
| 776 | return XXH64_update_endian(state_in, input, len, XXH_bigEndian); |
| 777 | } |
| 778 | |
| 779 | |
| 780 | |
| 781 | FORCE_INLINE_TEMPLATE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess endian) |
| 782 | { |
| 783 | const BYTE * p = (const BYTE*)state->mem64; |
| 784 | const BYTE* const bEnd = (const BYTE*)state->mem64 + state->memsize; |
| 785 | U64 h64; |
| 786 | |
| 787 | if (state->total_len >= 32) { |
| 788 | U64 const v1 = state->v1; |
| 789 | U64 const v2 = state->v2; |
| 790 | U64 const v3 = state->v3; |
| 791 | U64 const v4 = state->v4; |
| 792 | |
| 793 | h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18); |
| 794 | h64 = XXH64_mergeRound(h64, v1); |
| 795 | h64 = XXH64_mergeRound(h64, v2); |
| 796 | h64 = XXH64_mergeRound(h64, v3); |
| 797 | h64 = XXH64_mergeRound(h64, v4); |
| 798 | } else { |
| 799 | h64 = state->v3 + PRIME64_5; |
| 800 | } |
| 801 | |
| 802 | h64 += (U64) state->total_len; |
| 803 | |
| 804 | while (p+8<=bEnd) { |
| 805 | U64 const k1 = XXH64_round(0, XXH_readLE64(p, endian)); |
| 806 | h64 ^= k1; |
| 807 | h64 = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4; |
| 808 | p+=8; |
| 809 | } |
| 810 | |
| 811 | if (p+4<=bEnd) { |
| 812 | h64 ^= (U64)(XXH_readLE32(p, endian)) * PRIME64_1; |
| 813 | h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3; |
| 814 | p+=4; |
| 815 | } |
| 816 | |
| 817 | while (p<bEnd) { |
| 818 | h64 ^= (*p) * PRIME64_5; |
| 819 | h64 = XXH_rotl64(h64, 11) * PRIME64_1; |
| 820 | p++; |
| 821 | } |
| 822 | |
| 823 | h64 ^= h64 >> 33; |
| 824 | h64 *= PRIME64_2; |
| 825 | h64 ^= h64 >> 29; |
| 826 | h64 *= PRIME64_3; |
| 827 | h64 ^= h64 >> 32; |
| 828 | |
| 829 | return h64; |
| 830 | } |
| 831 | |
| 832 | |
| 833 | XXH_PUBLIC_API unsigned long long XXH64_digest (const XXH64_state_t* state_in) |
| 834 | { |
| 835 | XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; |
| 836 | |
| 837 | if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) |
| 838 | return XXH64_digest_endian(state_in, XXH_littleEndian); |
| 839 | else |
| 840 | return XXH64_digest_endian(state_in, XXH_bigEndian); |
| 841 | } |
| 842 | |
| 843 | |
| 844 | /* ************************** |
| 845 | * Canonical representation |
| 846 | ****************************/ |
| 847 | |
| 848 | /*! Default XXH result types are basic unsigned 32 and 64 bits. |
| 849 | * The canonical representation follows human-readable write convention, aka big-endian (large digits first). |
| 850 | * These functions allow transformation of hash result into and from its canonical format. |
| 851 | * This way, hash values can be written into a file or buffer, and remain comparable across different systems and programs. |
| 852 | */ |
| 853 | |
| 854 | XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash) |
| 855 | { |
| 856 | XXH_STATIC_ASSERT(sizeof(XXH32_canonical_t) == sizeof(XXH32_hash_t)); |
| 857 | if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap32(hash); |
| 858 | memcpy(dst, &hash, sizeof(*dst)); |
| 859 | } |
| 860 | |
| 861 | XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH64_canonical_t* dst, XXH64_hash_t hash) |
| 862 | { |
| 863 | XXH_STATIC_ASSERT(sizeof(XXH64_canonical_t) == sizeof(XXH64_hash_t)); |
| 864 | if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap64(hash); |
| 865 | memcpy(dst, &hash, sizeof(*dst)); |
| 866 | } |
| 867 | |
| 868 | XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src) |
| 869 | { |
| 870 | return XXH_readBE32(src); |
| 871 | } |
| 872 | |
| 873 | XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(const XXH64_canonical_t* src) |
| 874 | { |
| 875 | return XXH_readBE64(src); |
| 876 | } |