| 1 |
/* |
| 2 |
* Copyright (c) 2016-2020, Yann Collet, Facebook, Inc. |
| 3 |
* All rights reserved. |
| 4 |
* |
| 5 |
* This source code is licensed under both the BSD-style license (found in the |
| 6 |
* LICENSE file in the root directory of this source tree) and the GPLv2 (found |
| 7 |
* in the COPYING file in the root directory of this source tree). |
| 8 |
* You may select, at your option, one of the above-listed licenses. |
| 9 |
*/ |
| 10 |
|
| 11 |
/*-************************************* |
| 12 |
* Dependencies |
| 13 |
***************************************/ |
| 14 |
#include "zstd_compress_sequences.h" |
| 15 |
|
| 16 |
/** |
| 17 |
* -log2(x / 256) lookup table for x in [0, 256). |
| 18 |
* If x == 0: Return 0 |
| 19 |
* Else: Return floor(-log2(x / 256) * 256) |
| 20 |
*/ |
| 21 |
static unsigned const kInverseProbabilityLog256[256] = { |
| 22 |
0, 2048, 1792, 1642, 1536, 1453, 1386, 1329, 1280, 1236, 1197, 1162, |
| 23 |
1130, 1100, 1073, 1047, 1024, 1001, 980, 960, 941, 923, 906, 889, |
| 24 |
874, 859, 844, 830, 817, 804, 791, 779, 768, 756, 745, 734, |
| 25 |
724, 714, 704, 694, 685, 676, 667, 658, 650, 642, 633, 626, |
| 26 |
618, 610, 603, 595, 588, 581, 574, 567, 561, 554, 548, 542, |
| 27 |
535, 529, 523, 517, 512, 506, 500, 495, 489, 484, 478, 473, |
| 28 |
468, 463, 458, 453, 448, 443, 438, 434, 429, 424, 420, 415, |
| 29 |
411, 407, 402, 398, 394, 390, 386, 382, 377, 373, 370, 366, |
| 30 |
362, 358, 354, 350, 347, 343, 339, 336, 332, 329, 325, 322, |
| 31 |
318, 315, 311, 308, 305, 302, 298, 295, 292, 289, 286, 282, |
| 32 |
279, 276, 273, 270, 267, 264, 261, 258, 256, 253, 250, 247, |
| 33 |
244, 241, 239, 236, 233, 230, 228, 225, 222, 220, 217, 215, |
| 34 |
212, 209, 207, 204, 202, 199, 197, 194, 192, 190, 187, 185, |
| 35 |
182, 180, 178, 175, 173, 171, 168, 166, 164, 162, 159, 157, |
| 36 |
155, 153, 151, 149, 146, 144, 142, 140, 138, 136, 134, 132, |
| 37 |
130, 128, 126, 123, 121, 119, 117, 115, 114, 112, 110, 108, |
| 38 |
106, 104, 102, 100, 98, 96, 94, 93, 91, 89, 87, 85, |
| 39 |
83, 82, 80, 78, 76, 74, 73, 71, 69, 67, 66, 64, |
| 40 |
62, 61, 59, 57, 55, 54, 52, 50, 49, 47, 46, 44, |
| 41 |
42, 41, 39, 37, 36, 34, 33, 31, 30, 28, 26, 25, |
| 42 |
23, 22, 20, 19, 17, 16, 14, 13, 11, 10, 8, 7, |
| 43 |
5, 4, 2, 1, |
| 44 |
}; |
| 45 |
|
| 46 |
static unsigned ZSTD_getFSEMaxSymbolValue(FSE_CTable const* ctable) { |
| 47 |
void const* ptr = ctable; |
| 48 |
U16 const* u16ptr = (U16 const*)ptr; |
| 49 |
U32 const maxSymbolValue = MEM_read16(u16ptr + 1); |
| 50 |
return maxSymbolValue; |
| 51 |
} |
| 52 |
|
| 53 |
/** |
| 54 |
* Returns the cost in bytes of encoding the normalized count header. |
| 55 |
* Returns an error if any of the helper functions return an error. |
| 56 |
*/ |
| 57 |
static size_t ZSTD_NCountCost(unsigned const* count, unsigned const max, |
| 58 |
size_t const nbSeq, unsigned const FSELog) |
| 59 |
{ |
| 60 |
BYTE wksp[FSE_NCOUNTBOUND]; |
| 61 |
S16 norm[MaxSeq + 1]; |
| 62 |
const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max); |
| 63 |
FORWARD_IF_ERROR(FSE_normalizeCount(norm, tableLog, count, nbSeq, max), ""); |
| 64 |
return FSE_writeNCount(wksp, sizeof(wksp), norm, max, tableLog); |
| 65 |
} |
| 66 |
|
| 67 |
/** |
| 68 |
* Returns the cost in bits of encoding the distribution described by count |
| 69 |
* using the entropy bound. |
| 70 |
*/ |
| 71 |
static size_t ZSTD_entropyCost(unsigned const* count, unsigned const max, size_t const total) |
| 72 |
{ |
| 73 |
unsigned cost = 0; |
| 74 |
unsigned s; |
| 75 |
for (s = 0; s <= max; ++s) { |
| 76 |
unsigned norm = (unsigned)((256 * count[s]) / total); |
| 77 |
if (count[s] != 0 && norm == 0) |
| 78 |
norm = 1; |
| 79 |
assert(count[s] < total); |
| 80 |
cost += count[s] * kInverseProbabilityLog256[norm]; |
| 81 |
} |
| 82 |
return cost >> 8; |
| 83 |
} |
| 84 |
|
| 85 |
/** |
| 86 |
* Returns the cost in bits of encoding the distribution in count using ctable. |
| 87 |
* Returns an error if ctable cannot represent all the symbols in count. |
| 88 |
*/ |
| 89 |
size_t ZSTD_fseBitCost( |
| 90 |
FSE_CTable const* ctable, |
| 91 |
unsigned const* count, |
| 92 |
unsigned const max) |
| 93 |
{ |
| 94 |
unsigned const kAccuracyLog = 8; |
| 95 |
size_t cost = 0; |
| 96 |
unsigned s; |
| 97 |
FSE_CState_t cstate; |
| 98 |
FSE_initCState(&cstate, ctable); |
| 99 |
if (ZSTD_getFSEMaxSymbolValue(ctable) < max) { |
| 100 |
DEBUGLOG(5, "Repeat FSE_CTable has maxSymbolValue %u < %u", |
| 101 |
ZSTD_getFSEMaxSymbolValue(ctable), max); |
| 102 |
return ERROR(GENERIC); |
| 103 |
} |
| 104 |
for (s = 0; s <= max; ++s) { |
| 105 |
unsigned const tableLog = cstate.stateLog; |
| 106 |
unsigned const badCost = (tableLog + 1) << kAccuracyLog; |
| 107 |
unsigned const bitCost = FSE_bitCost(cstate.symbolTT, tableLog, s, kAccuracyLog); |
| 108 |
if (count[s] == 0) |
| 109 |
continue; |
| 110 |
if (bitCost >= badCost) { |
| 111 |
DEBUGLOG(5, "Repeat FSE_CTable has Prob[%u] == 0", s); |
| 112 |
return ERROR(GENERIC); |
| 113 |
} |
| 114 |
cost += (size_t)count[s] * bitCost; |
| 115 |
} |
| 116 |
return cost >> kAccuracyLog; |
| 117 |
} |
| 118 |
|
| 119 |
/** |
| 120 |
* Returns the cost in bits of encoding the distribution in count using the |
| 121 |
* table described by norm. The max symbol support by norm is assumed >= max. |
| 122 |
* norm must be valid for every symbol with non-zero probability in count. |
| 123 |
*/ |
| 124 |
size_t ZSTD_crossEntropyCost(short const* norm, unsigned accuracyLog, |
| 125 |
unsigned const* count, unsigned const max) |
| 126 |
{ |
| 127 |
unsigned const shift = 8 - accuracyLog; |
| 128 |
size_t cost = 0; |
| 129 |
unsigned s; |
| 130 |
assert(accuracyLog <= 8); |
| 131 |
for (s = 0; s <= max; ++s) { |
| 132 |
unsigned const normAcc = (norm[s] != -1) ? (unsigned)norm[s] : 1; |
| 133 |
unsigned const norm256 = normAcc << shift; |
| 134 |
assert(norm256 > 0); |
| 135 |
assert(norm256 < 256); |
| 136 |
cost += count[s] * kInverseProbabilityLog256[norm256]; |
| 137 |
} |
| 138 |
return cost >> 8; |
| 139 |
} |
| 140 |
|
| 141 |
symbolEncodingType_e |
| 142 |
ZSTD_selectEncodingType( |
| 143 |
FSE_repeat* repeatMode, unsigned const* count, unsigned const max, |
| 144 |
size_t const mostFrequent, size_t nbSeq, unsigned const FSELog, |
| 145 |
FSE_CTable const* prevCTable, |
| 146 |
short const* defaultNorm, U32 defaultNormLog, |
| 147 |
ZSTD_defaultPolicy_e const isDefaultAllowed, |
| 148 |
ZSTD_strategy const strategy) |
| 149 |
{ |
| 150 |
ZSTD_STATIC_ASSERT(ZSTD_defaultDisallowed == 0 && ZSTD_defaultAllowed != 0); |
| 151 |
if (mostFrequent == nbSeq) { |
| 152 |
*repeatMode = FSE_repeat_none; |
| 153 |
if (isDefaultAllowed && nbSeq <= 2) { |
| 154 |
/* Prefer set_basic over set_rle when there are 2 or less symbols, |
| 155 |
* since RLE uses 1 byte, but set_basic uses 5-6 bits per symbol. |
| 156 |
* If basic encoding isn't possible, always choose RLE. |
| 157 |
*/ |
| 158 |
DEBUGLOG(5, "Selected set_basic"); |
| 159 |
return set_basic; |
| 160 |
} |
| 161 |
DEBUGLOG(5, "Selected set_rle"); |
| 162 |
return set_rle; |
| 163 |
} |
| 164 |
if (strategy < ZSTD_lazy) { |
| 165 |
if (isDefaultAllowed) { |
| 166 |
size_t const staticFse_nbSeq_max = 1000; |
| 167 |
size_t const mult = 10 - strategy; |
| 168 |
size_t const baseLog = 3; |
| 169 |
size_t const dynamicFse_nbSeq_min = (((size_t)1 << defaultNormLog) * mult) >> baseLog; /* 28-36 for offset, 56-72 for lengths */ |
| 170 |
assert(defaultNormLog >= 5 && defaultNormLog <= 6); /* xx_DEFAULTNORMLOG */ |
| 171 |
assert(mult <= 9 && mult >= 7); |
| 172 |
if ( (*repeatMode == FSE_repeat_valid) |
| 173 |
&& (nbSeq < staticFse_nbSeq_max) ) { |
| 174 |
DEBUGLOG(5, "Selected set_repeat"); |
| 175 |
return set_repeat; |
| 176 |
} |
| 177 |
if ( (nbSeq < dynamicFse_nbSeq_min) |
| 178 |
|| (mostFrequent < (nbSeq >> (defaultNormLog-1))) ) { |
| 179 |
DEBUGLOG(5, "Selected set_basic"); |
| 180 |
/* The format allows default tables to be repeated, but it isn't useful. |
| 181 |
* When using simple heuristics to select encoding type, we don't want |
| 182 |
* to confuse these tables with dictionaries. When running more careful |
| 183 |
* analysis, we don't need to waste time checking both repeating tables |
| 184 |
* and default tables. |
| 185 |
*/ |
| 186 |
*repeatMode = FSE_repeat_none; |
| 187 |
return set_basic; |
| 188 |
} |
| 189 |
} |
| 190 |
} else { |
| 191 |
size_t const basicCost = isDefaultAllowed ? ZSTD_crossEntropyCost(defaultNorm, defaultNormLog, count, max) : ERROR(GENERIC); |
| 192 |
size_t const repeatCost = *repeatMode != FSE_repeat_none ? ZSTD_fseBitCost(prevCTable, count, max) : ERROR(GENERIC); |
| 193 |
size_t const NCountCost = ZSTD_NCountCost(count, max, nbSeq, FSELog); |
| 194 |
size_t const compressedCost = (NCountCost << 3) + ZSTD_entropyCost(count, max, nbSeq); |
| 195 |
|
| 196 |
if (isDefaultAllowed) { |
| 197 |
assert(!ZSTD_isError(basicCost)); |
| 198 |
assert(!(*repeatMode == FSE_repeat_valid && ZSTD_isError(repeatCost))); |
| 199 |
} |
| 200 |
assert(!ZSTD_isError(NCountCost)); |
| 201 |
assert(compressedCost < ERROR(maxCode)); |
| 202 |
DEBUGLOG(5, "Estimated bit costs: basic=%u\trepeat=%u\tcompressed=%u", |
| 203 |
(unsigned)basicCost, (unsigned)repeatCost, (unsigned)compressedCost); |
| 204 |
if (basicCost <= repeatCost && basicCost <= compressedCost) { |
| 205 |
DEBUGLOG(5, "Selected set_basic"); |
| 206 |
assert(isDefaultAllowed); |
| 207 |
*repeatMode = FSE_repeat_none; |
| 208 |
return set_basic; |
| 209 |
} |
| 210 |
if (repeatCost <= compressedCost) { |
| 211 |
DEBUGLOG(5, "Selected set_repeat"); |
| 212 |
assert(!ZSTD_isError(repeatCost)); |
| 213 |
return set_repeat; |
| 214 |
} |
| 215 |
assert(compressedCost < basicCost && compressedCost < repeatCost); |
| 216 |
} |
| 217 |
DEBUGLOG(5, "Selected set_compressed"); |
| 218 |
*repeatMode = FSE_repeat_check; |
| 219 |
return set_compressed; |
| 220 |
} |
| 221 |
|
| 222 |
size_t |
| 223 |
ZSTD_buildCTable(void* dst, size_t dstCapacity, |
| 224 |
FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type, |
| 225 |
unsigned* count, U32 max, |
| 226 |
const BYTE* codeTable, size_t nbSeq, |
| 227 |
const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax, |
| 228 |
const FSE_CTable* prevCTable, size_t prevCTableSize, |
| 229 |
void* entropyWorkspace, size_t entropyWorkspaceSize) |
| 230 |
{ |
| 231 |
BYTE* op = (BYTE*)dst; |
| 232 |
const BYTE* const oend = op + dstCapacity; |
| 233 |
DEBUGLOG(6, "ZSTD_buildCTable (dstCapacity=%u)", (unsigned)dstCapacity); |
| 234 |
|
| 235 |
switch (type) { |
| 236 |
case set_rle: |
| 237 |
FORWARD_IF_ERROR(FSE_buildCTable_rle(nextCTable, (BYTE)max), ""); |
| 238 |
RETURN_ERROR_IF(dstCapacity==0, dstSize_tooSmall, "not enough space"); |
| 239 |
*op = codeTable[0]; |
| 240 |
return 1; |
| 241 |
case set_repeat: |
| 242 |
memcpy(nextCTable, prevCTable, prevCTableSize); |
| 243 |
return 0; |
| 244 |
case set_basic: |
| 245 |
FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, defaultNorm, defaultMax, defaultNormLog, entropyWorkspace, entropyWorkspaceSize), ""); /* note : could be pre-calculated */ |
| 246 |
return 0; |
| 247 |
case set_compressed: { |
| 248 |
S16 norm[MaxSeq + 1]; |
| 249 |
size_t nbSeq_1 = nbSeq; |
| 250 |
const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max); |
| 251 |
if (count[codeTable[nbSeq-1]] > 1) { |
| 252 |
count[codeTable[nbSeq-1]]--; |
| 253 |
nbSeq_1--; |
| 254 |
} |
| 255 |
assert(nbSeq_1 > 1); |
| 256 |
FORWARD_IF_ERROR(FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max), ""); |
| 257 |
{ size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog); /* overflow protected */ |
| 258 |
FORWARD_IF_ERROR(NCountSize, "FSE_writeNCount failed"); |
| 259 |
FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, norm, max, tableLog, entropyWorkspace, entropyWorkspaceSize), ""); |
| 260 |
return NCountSize; |
| 261 |
} |
| 262 |
} |
| 263 |
default: assert(0); RETURN_ERROR(GENERIC, "impossible to reach"); |
| 264 |
} |
| 265 |
} |
| 266 |
|
| 267 |
FORCE_INLINE_TEMPLATE size_t |
| 268 |
ZSTD_encodeSequences_body( |
| 269 |
void* dst, size_t dstCapacity, |
| 270 |
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable, |
| 271 |
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable, |
| 272 |
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable, |
| 273 |
seqDef const* sequences, size_t nbSeq, int longOffsets) |
| 274 |
{ |
| 275 |
BIT_CStream_t blockStream; |
| 276 |
FSE_CState_t stateMatchLength; |
| 277 |
FSE_CState_t stateOffsetBits; |
| 278 |
FSE_CState_t stateLitLength; |
| 279 |
|
| 280 |
RETURN_ERROR_IF( |
| 281 |
ERR_isError(BIT_initCStream(&blockStream, dst, dstCapacity)), |
| 282 |
dstSize_tooSmall, "not enough space remaining"); |
| 283 |
DEBUGLOG(6, "available space for bitstream : %i (dstCapacity=%u)", |
| 284 |
(int)(blockStream.endPtr - blockStream.startPtr), |
| 285 |
(unsigned)dstCapacity); |
| 286 |
|
| 287 |
/* first symbols */ |
| 288 |
FSE_initCState2(&stateMatchLength, CTable_MatchLength, mlCodeTable[nbSeq-1]); |
| 289 |
FSE_initCState2(&stateOffsetBits, CTable_OffsetBits, ofCodeTable[nbSeq-1]); |
| 290 |
FSE_initCState2(&stateLitLength, CTable_LitLength, llCodeTable[nbSeq-1]); |
| 291 |
BIT_addBits(&blockStream, sequences[nbSeq-1].litLength, LL_bits[llCodeTable[nbSeq-1]]); |
| 292 |
if (MEM_32bits()) BIT_flushBits(&blockStream); |
| 293 |
BIT_addBits(&blockStream, sequences[nbSeq-1].matchLength, ML_bits[mlCodeTable[nbSeq-1]]); |
| 294 |
if (MEM_32bits()) BIT_flushBits(&blockStream); |
| 295 |
if (longOffsets) { |
| 296 |
U32 const ofBits = ofCodeTable[nbSeq-1]; |
| 297 |
unsigned const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1); |
| 298 |
if (extraBits) { |
| 299 |
BIT_addBits(&blockStream, sequences[nbSeq-1].offset, extraBits); |
| 300 |
BIT_flushBits(&blockStream); |
| 301 |
} |
| 302 |
BIT_addBits(&blockStream, sequences[nbSeq-1].offset >> extraBits, |
| 303 |
ofBits - extraBits); |
| 304 |
} else { |
| 305 |
BIT_addBits(&blockStream, sequences[nbSeq-1].offset, ofCodeTable[nbSeq-1]); |
| 306 |
} |
| 307 |
BIT_flushBits(&blockStream); |
| 308 |
|
| 309 |
{ size_t n; |
| 310 |
for (n=nbSeq-2 ; n<nbSeq ; n--) { /* intentional underflow */ |
| 311 |
BYTE const llCode = llCodeTable[n]; |
| 312 |
BYTE const ofCode = ofCodeTable[n]; |
| 313 |
BYTE const mlCode = mlCodeTable[n]; |
| 314 |
U32 const llBits = LL_bits[llCode]; |
| 315 |
U32 const ofBits = ofCode; |
| 316 |
U32 const mlBits = ML_bits[mlCode]; |
| 317 |
DEBUGLOG(6, "encoding: litlen:%2u - matchlen:%2u - offCode:%7u", |
| 318 |
(unsigned)sequences[n].litLength, |
| 319 |
(unsigned)sequences[n].matchLength + MINMATCH, |
| 320 |
(unsigned)sequences[n].offset); |
| 321 |
/* 32b*/ /* 64b*/ |
| 322 |
/* (7)*/ /* (7)*/ |
| 323 |
FSE_encodeSymbol(&blockStream, &stateOffsetBits, ofCode); /* 15 */ /* 15 */ |
| 324 |
FSE_encodeSymbol(&blockStream, &stateMatchLength, mlCode); /* 24 */ /* 24 */ |
| 325 |
if (MEM_32bits()) BIT_flushBits(&blockStream); /* (7)*/ |
| 326 |
FSE_encodeSymbol(&blockStream, &stateLitLength, llCode); /* 16 */ /* 33 */ |
| 327 |
if (MEM_32bits() || (ofBits+mlBits+llBits >= 64-7-(LLFSELog+MLFSELog+OffFSELog))) |
| 328 |
BIT_flushBits(&blockStream); /* (7)*/ |
| 329 |
BIT_addBits(&blockStream, sequences[n].litLength, llBits); |
| 330 |
if (MEM_32bits() && ((llBits+mlBits)>24)) BIT_flushBits(&blockStream); |
| 331 |
BIT_addBits(&blockStream, sequences[n].matchLength, mlBits); |
| 332 |
if (MEM_32bits() || (ofBits+mlBits+llBits > 56)) BIT_flushBits(&blockStream); |
| 333 |
if (longOffsets) { |
| 334 |
unsigned const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1); |
| 335 |
if (extraBits) { |
| 336 |
BIT_addBits(&blockStream, sequences[n].offset, extraBits); |
| 337 |
BIT_flushBits(&blockStream); /* (7)*/ |
| 338 |
} |
| 339 |
BIT_addBits(&blockStream, sequences[n].offset >> extraBits, |
| 340 |
ofBits - extraBits); /* 31 */ |
| 341 |
} else { |
| 342 |
BIT_addBits(&blockStream, sequences[n].offset, ofBits); /* 31 */ |
| 343 |
} |
| 344 |
BIT_flushBits(&blockStream); /* (7)*/ |
| 345 |
DEBUGLOG(7, "remaining space : %i", (int)(blockStream.endPtr - blockStream.ptr)); |
| 346 |
} } |
| 347 |
|
| 348 |
DEBUGLOG(6, "ZSTD_encodeSequences: flushing ML state with %u bits", stateMatchLength.stateLog); |
| 349 |
FSE_flushCState(&blockStream, &stateMatchLength); |
| 350 |
DEBUGLOG(6, "ZSTD_encodeSequences: flushing Off state with %u bits", stateOffsetBits.stateLog); |
| 351 |
FSE_flushCState(&blockStream, &stateOffsetBits); |
| 352 |
DEBUGLOG(6, "ZSTD_encodeSequences: flushing LL state with %u bits", stateLitLength.stateLog); |
| 353 |
FSE_flushCState(&blockStream, &stateLitLength); |
| 354 |
|
| 355 |
{ size_t const streamSize = BIT_closeCStream(&blockStream); |
| 356 |
RETURN_ERROR_IF(streamSize==0, dstSize_tooSmall, "not enough space"); |
| 357 |
return streamSize; |
| 358 |
} |
| 359 |
} |
| 360 |
|
| 361 |
static size_t |
| 362 |
ZSTD_encodeSequences_default( |
| 363 |
void* dst, size_t dstCapacity, |
| 364 |
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable, |
| 365 |
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable, |
| 366 |
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable, |
| 367 |
seqDef const* sequences, size_t nbSeq, int longOffsets) |
| 368 |
{ |
| 369 |
return ZSTD_encodeSequences_body(dst, dstCapacity, |
| 370 |
CTable_MatchLength, mlCodeTable, |
| 371 |
CTable_OffsetBits, ofCodeTable, |
| 372 |
CTable_LitLength, llCodeTable, |
| 373 |
sequences, nbSeq, longOffsets); |
| 374 |
} |
| 375 |
|
| 376 |
|
| 377 |
#if DYNAMIC_BMI2 |
| 378 |
|
| 379 |
static TARGET_ATTRIBUTE("bmi2") size_t |
| 380 |
ZSTD_encodeSequences_bmi2( |
| 381 |
void* dst, size_t dstCapacity, |
| 382 |
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable, |
| 383 |
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable, |
| 384 |
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable, |
| 385 |
seqDef const* sequences, size_t nbSeq, int longOffsets) |
| 386 |
{ |
| 387 |
return ZSTD_encodeSequences_body(dst, dstCapacity, |
| 388 |
CTable_MatchLength, mlCodeTable, |
| 389 |
CTable_OffsetBits, ofCodeTable, |
| 390 |
CTable_LitLength, llCodeTable, |
| 391 |
sequences, nbSeq, longOffsets); |
| 392 |
} |
| 393 |
|
| 394 |
#endif |
| 395 |
|
| 396 |
size_t ZSTD_encodeSequences( |
| 397 |
void* dst, size_t dstCapacity, |
| 398 |
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable, |
| 399 |
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable, |
| 400 |
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable, |
| 401 |
seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2) |
| 402 |
{ |
| 403 |
DEBUGLOG(5, "ZSTD_encodeSequences: dstCapacity = %u", (unsigned)dstCapacity); |
| 404 |
#if DYNAMIC_BMI2 |
| 405 |
if (bmi2) { |
| 406 |
return ZSTD_encodeSequences_bmi2(dst, dstCapacity, |
| 407 |
CTable_MatchLength, mlCodeTable, |
| 408 |
CTable_OffsetBits, ofCodeTable, |
| 409 |
CTable_LitLength, llCodeTable, |
| 410 |
sequences, nbSeq, longOffsets); |
| 411 |
} |
| 412 |
#endif |
| 413 |
(void)bmi2; |
| 414 |
return ZSTD_encodeSequences_default(dst, dstCapacity, |
| 415 |
CTable_MatchLength, mlCodeTable, |
| 416 |
CTable_OffsetBits, ofCodeTable, |
| 417 |
CTable_LitLength, llCodeTable, |
| 418 |
sequences, nbSeq, longOffsets); |
| 419 |
} |