| 1 | /* |
| 2 | * Copyright (C) 2013 Reimar Döffinger <Reimar.Doeffinger@gmx.de> |
| 3 | * |
| 4 | * This file is part of FFmpeg. |
| 5 | * |
| 6 | * FFmpeg is free software; you can redistribute it and/or |
| 7 | * modify it under the terms of the GNU Lesser General Public |
| 8 | * License as published by the Free Software Foundation; either |
| 9 | * version 2.1 of the License, or (at your option) any later version. |
| 10 | * |
| 11 | * FFmpeg is distributed in the hope that it will be useful, |
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 14 | * Lesser General Public License for more details. |
| 15 | * |
| 16 | * You should have received a copy of the GNU Lesser General Public |
| 17 | * License along with FFmpeg; if not, write to the Free Software |
| 18 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
| 19 | */ |
| 20 | #include <stdint.h> |
| 21 | #include "mem.h" |
| 22 | #include "intreadwrite.h" |
| 23 | #include "murmur3.h" |
| 24 | |
| 25 | typedef struct AVMurMur3 { |
| 26 | uint64_t h1, h2; |
| 27 | uint8_t state[16]; |
| 28 | int state_pos; |
| 29 | uint64_t len; |
| 30 | } AVMurMur3; |
| 31 | |
| 32 | AVMurMur3 *av_murmur3_alloc(void) |
| 33 | { |
| 34 | return av_mallocz(sizeof(AVMurMur3)); |
| 35 | } |
| 36 | |
| 37 | void av_murmur3_init_seeded(AVMurMur3 *c, uint64_t seed) |
| 38 | { |
| 39 | memset(c, 0, sizeof(*c)); |
| 40 | c->h1 = c->h2 = seed; |
| 41 | } |
| 42 | |
| 43 | void av_murmur3_init(AVMurMur3 *c) |
| 44 | { |
| 45 | // arbitrary random number as seed |
| 46 | av_murmur3_init_seeded(c, 0x725acc55daddca55); |
| 47 | } |
| 48 | |
| 49 | static const uint64_t c1 = UINT64_C(0x87c37b91114253d5); |
| 50 | static const uint64_t c2 = UINT64_C(0x4cf5ad432745937f); |
| 51 | |
| 52 | #define ROT(a, b) ((a << b) | (a >> (64 - b))) |
| 53 | |
| 54 | static uint64_t inline get_k1(const uint8_t *src) |
| 55 | { |
| 56 | uint64_t k = AV_RL64(src); |
| 57 | k *= c1; |
| 58 | k = ROT(k, 31); |
| 59 | k *= c2; |
| 60 | return k; |
| 61 | } |
| 62 | |
| 63 | static uint64_t inline get_k2(const uint8_t *src) |
| 64 | { |
| 65 | uint64_t k = AV_RL64(src + 8); |
| 66 | k *= c2; |
| 67 | k = ROT(k, 33); |
| 68 | k *= c1; |
| 69 | return k; |
| 70 | } |
| 71 | |
| 72 | static uint64_t inline update_h1(uint64_t k, uint64_t h1, uint64_t h2) |
| 73 | { |
| 74 | k ^= h1; |
| 75 | k = ROT(k, 27); |
| 76 | k += h2; |
| 77 | k *= 5; |
| 78 | k += 0x52dce729; |
| 79 | return k; |
| 80 | } |
| 81 | |
| 82 | static uint64_t inline update_h2(uint64_t k, uint64_t h1, uint64_t h2) |
| 83 | { |
| 84 | k ^= h2; |
| 85 | k = ROT(k, 31); |
| 86 | k += h1; |
| 87 | k *= 5; |
| 88 | k += 0x38495ab5; |
| 89 | return k; |
| 90 | } |
| 91 | |
| 92 | void av_murmur3_update(AVMurMur3 *c, const uint8_t *src, int len) |
| 93 | { |
| 94 | const uint8_t *end; |
| 95 | uint64_t h1 = c->h1, h2 = c->h2; |
| 96 | uint64_t k1, k2; |
| 97 | if (len <= 0) return; |
| 98 | c->len += len; |
| 99 | if (c->state_pos > 0) { |
| 100 | while (c->state_pos < 16) { |
| 101 | c->state[c->state_pos++] = *src++; |
| 102 | if (--len <= 0) return; |
| 103 | } |
| 104 | c->state_pos = 0; |
| 105 | k1 = get_k1(c->state); |
| 106 | k2 = get_k2(c->state); |
| 107 | h1 = update_h1(k1, h1, h2); |
| 108 | h2 = update_h2(k2, h1, h2); |
| 109 | } |
| 110 | |
| 111 | end = src + (len & ~15); |
| 112 | while (src < end) { |
| 113 | // These could be done sequentially instead |
| 114 | // of interleaved, but like this is over 10% faster |
| 115 | k1 = get_k1(src); |
| 116 | k2 = get_k2(src); |
| 117 | h1 = update_h1(k1, h1, h2); |
| 118 | h2 = update_h2(k2, h1, h2); |
| 119 | src += 16; |
| 120 | } |
| 121 | c->h1 = h1; |
| 122 | c->h2 = h2; |
| 123 | |
| 124 | len &= 15; |
| 125 | if (len > 0) { |
| 126 | memcpy(c->state, src, len); |
| 127 | c->state_pos = len; |
| 128 | } |
| 129 | } |
| 130 | |
| 131 | static inline uint64_t fmix(uint64_t k) |
| 132 | { |
| 133 | k ^= k >> 33; |
| 134 | k *= UINT64_C(0xff51afd7ed558ccd); |
| 135 | k ^= k >> 33; |
| 136 | k *= UINT64_C(0xc4ceb9fe1a85ec53); |
| 137 | k ^= k >> 33; |
| 138 | return k; |
| 139 | } |
| 140 | |
| 141 | void av_murmur3_final(AVMurMur3 *c, uint8_t dst[16]) |
| 142 | { |
| 143 | uint64_t h1 = c->h1, h2 = c->h2; |
| 144 | memset(c->state + c->state_pos, 0, sizeof(c->state) - c->state_pos); |
| 145 | h1 ^= get_k1(c->state) ^ c->len; |
| 146 | h2 ^= get_k2(c->state) ^ c->len; |
| 147 | h1 += h2; |
| 148 | h2 += h1; |
| 149 | h1 = fmix(h1); |
| 150 | h2 = fmix(h2); |
| 151 | h1 += h2; |
| 152 | h2 += h1; |
| 153 | AV_WL64(dst, h1); |
| 154 | AV_WL64(dst + 8, h2); |
| 155 | } |
| 156 | |
| 157 | #ifdef TEST |
| 158 | int main(void) |
| 159 | { |
| 160 | int i; |
| 161 | uint8_t hash_result[16] = {0}; |
| 162 | AVMurMur3 *ctx = av_murmur3_alloc(); |
| 163 | #if 1 |
| 164 | uint8_t in[256] = {0}; |
| 165 | uint8_t *hashes = av_mallocz(256 * 16); |
| 166 | for (i = 0; i < 256; i++) |
| 167 | { |
| 168 | in[i] = i; |
| 169 | av_murmur3_init_seeded(ctx, 256 - i); |
| 170 | // Note: this actually tests hashing 0 bytes |
| 171 | av_murmur3_update(ctx, in, i); |
| 172 | av_murmur3_final(ctx, hashes + 16 * i); |
| 173 | } |
| 174 | av_murmur3_init_seeded(ctx, 0); |
| 175 | av_murmur3_update(ctx, hashes, 256 * 16); |
| 176 | av_murmur3_final(ctx, hash_result); |
| 177 | av_free(hashes); |
| 178 | av_freep(&ctx); |
| 179 | printf("result: 0x%"PRIx64" 0x%"PRIx64"\n", AV_RL64(hash_result), AV_RL64(hash_result + 8)); |
| 180 | // official reference value is 32 bit |
| 181 | return AV_RL32(hash_result) != 0x6384ba69; |
| 182 | #else |
| 183 | uint8_t *in = av_mallocz(512*1024); |
| 184 | av_murmur3_init(ctx); |
| 185 | for (i = 0; i < 40*1024; i++) |
| 186 | av_murmur3_update(ctx, in, 512*1024); |
| 187 | av_murmur3_final(ctx, hash_result); |
| 188 | av_free(in); |
| 189 | return hash_result[0]; |
| 190 | #endif |
| 191 | } |
| 192 | #endif |