1 /* 2 * Copyright (c) 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 #include "zstd_compress_internal.h" 12 #include "zstd_lazy.h" 13 14 15 /*-************************************* 16 * Binary Tree search 17 ***************************************/ 18 19 static void 20 ZSTD_updateDUBT(ZSTD_matchState_t* ms, 21 const BYTE* ip, const BYTE* iend, 22 U32 mls) 23 { 24 const ZSTD_compressionParameters* const cParams = &ms->cParams; 25 U32* const hashTable = ms->hashTable; 26 U32 const hashLog = cParams->hashLog; 27 28 U32* const bt = ms->chainTable; 29 U32 const btLog = cParams->chainLog - 1; 30 U32 const btMask = (1 << btLog) - 1; 31 32 const BYTE* const base = ms->window.base; 33 U32 const target = (U32)(ip - base); 34 U32 idx = ms->nextToUpdate; 35 36 if (idx != target) 37 DEBUGLOG(7, "ZSTD_updateDUBT, from %u to %u (dictLimit:%u)", 38 idx, target, ms->window.dictLimit); 39 assert(ip + 8 <= iend); /* condition for ZSTD_hashPtr */ 40 (void)iend; 41 42 assert(idx >= ms->window.dictLimit); /* condition for valid base+idx */ 43 for ( ; idx < target ; idx++) { 44 size_t const h = ZSTD_hashPtr(base + idx, hashLog, mls); /* assumption : ip + 8 <= iend */ 45 U32 const matchIndex = hashTable[h]; 46 47 U32* const nextCandidatePtr = bt + 2*(idx&btMask); 48 U32* const sortMarkPtr = nextCandidatePtr + 1; 49 50 DEBUGLOG(8, "ZSTD_updateDUBT: insert %u", idx); 51 hashTable[h] = idx; /* Update Hash Table */ 52 *nextCandidatePtr = matchIndex; /* update BT like a chain */ 53 *sortMarkPtr = ZSTD_DUBT_UNSORTED_MARK; 54 } 55 ms->nextToUpdate = target; 56 } 57 58 59 /** ZSTD_insertDUBT1() : 60 * sort one already inserted but unsorted position 61 * assumption : curr >= btlow == (curr - btmask) 62 * doesn't fail */ 63 static void 64 ZSTD_insertDUBT1(const ZSTD_matchState_t* ms, 65 U32 curr, const BYTE* inputEnd, 66 U32 nbCompares, U32 btLow, 67 const ZSTD_dictMode_e dictMode) 68 { 69 const ZSTD_compressionParameters* const cParams = &ms->cParams; 70 U32* const bt = ms->chainTable; 71 U32 const btLog = cParams->chainLog - 1; 72 U32 const btMask = (1 << btLog) - 1; 73 size_t commonLengthSmaller=0, commonLengthLarger=0; 74 const BYTE* const base = ms->window.base; 75 const BYTE* const dictBase = ms->window.dictBase; 76 const U32 dictLimit = ms->window.dictLimit; 77 const BYTE* const ip = (curr>=dictLimit) ? base + curr : dictBase + curr; 78 const BYTE* const iend = (curr>=dictLimit) ? inputEnd : dictBase + dictLimit; 79 const BYTE* const dictEnd = dictBase + dictLimit; 80 const BYTE* const prefixStart = base + dictLimit; 81 const BYTE* match; 82 U32* smallerPtr = bt + 2*(curr&btMask); 83 U32* largerPtr = smallerPtr + 1; 84 U32 matchIndex = *smallerPtr; /* this candidate is unsorted : next sorted candidate is reached through *smallerPtr, while *largerPtr contains previous unsorted candidate (which is already saved and can be overwritten) */ 85 U32 dummy32; /* to be nullified at the end */ 86 U32 const windowValid = ms->window.lowLimit; 87 U32 const maxDistance = 1U << cParams->windowLog; 88 U32 const windowLow = (curr - windowValid > maxDistance) ? curr - maxDistance : windowValid; 89 90 91 DEBUGLOG(8, "ZSTD_insertDUBT1(%u) (dictLimit=%u, lowLimit=%u)", 92 curr, dictLimit, windowLow); 93 assert(curr >= btLow); 94 assert(ip < iend); /* condition for ZSTD_count */ 95 96 for (; nbCompares && (matchIndex > windowLow); --nbCompares) { 97 U32* const nextPtr = bt + 2*(matchIndex & btMask); 98 size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */ 99 assert(matchIndex < curr); 100 /* note : all candidates are now supposed sorted, 101 * but it's still possible to have nextPtr[1] == ZSTD_DUBT_UNSORTED_MARK 102 * when a real index has the same value as ZSTD_DUBT_UNSORTED_MARK */ 103 104 if ( (dictMode != ZSTD_extDict) 105 || (matchIndex+matchLength >= dictLimit) /* both in current segment*/ 106 || (curr < dictLimit) /* both in extDict */) { 107 const BYTE* const mBase = ( (dictMode != ZSTD_extDict) 108 || (matchIndex+matchLength >= dictLimit)) ? 109 base : dictBase; 110 assert( (matchIndex+matchLength >= dictLimit) /* might be wrong if extDict is incorrectly set to 0 */ 111 || (curr < dictLimit) ); 112 match = mBase + matchIndex; 113 matchLength += ZSTD_count(ip+matchLength, match+matchLength, iend); 114 } else { 115 match = dictBase + matchIndex; 116 matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart); 117 if (matchIndex+matchLength >= dictLimit) 118 match = base + matchIndex; /* preparation for next read of match[matchLength] */ 119 } 120 121 DEBUGLOG(8, "ZSTD_insertDUBT1: comparing %u with %u : found %u common bytes ", 122 curr, matchIndex, (U32)matchLength); 123 124 if (ip+matchLength == iend) { /* equal : no way to know if inf or sup */ 125 break; /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt tree */ 126 } 127 128 if (match[matchLength] < ip[matchLength]) { /* necessarily within buffer */ 129 /* match is smaller than current */ 130 *smallerPtr = matchIndex; /* update smaller idx */ 131 commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */ 132 if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop searching */ 133 DEBUGLOG(8, "ZSTD_insertDUBT1: %u (>btLow=%u) is smaller : next => %u", 134 matchIndex, btLow, nextPtr[1]); 135 smallerPtr = nextPtr+1; /* new "candidate" => larger than match, which was smaller than target */ 136 matchIndex = nextPtr[1]; /* new matchIndex, larger than previous and closer to current */ 137 } else { 138 /* match is larger than current */ 139 *largerPtr = matchIndex; 140 commonLengthLarger = matchLength; 141 if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop searching */ 142 DEBUGLOG(8, "ZSTD_insertDUBT1: %u (>btLow=%u) is larger => %u", 143 matchIndex, btLow, nextPtr[0]); 144 largerPtr = nextPtr; 145 matchIndex = nextPtr[0]; 146 } } 147 148 *smallerPtr = *largerPtr = 0; 149 } 150 151 152 static size_t 153 ZSTD_DUBT_findBetterDictMatch ( 154 const ZSTD_matchState_t* ms, 155 const BYTE* const ip, const BYTE* const iend, 156 size_t* offsetPtr, 157 size_t bestLength, 158 U32 nbCompares, 159 U32 const mls, 160 const ZSTD_dictMode_e dictMode) 161 { 162 const ZSTD_matchState_t * const dms = ms->dictMatchState; 163 const ZSTD_compressionParameters* const dmsCParams = &dms->cParams; 164 const U32 * const dictHashTable = dms->hashTable; 165 U32 const hashLog = dmsCParams->hashLog; 166 size_t const h = ZSTD_hashPtr(ip, hashLog, mls); 167 U32 dictMatchIndex = dictHashTable[h]; 168 169 const BYTE* const base = ms->window.base; 170 const BYTE* const prefixStart = base + ms->window.dictLimit; 171 U32 const curr = (U32)(ip-base); 172 const BYTE* const dictBase = dms->window.base; 173 const BYTE* const dictEnd = dms->window.nextSrc; 174 U32 const dictHighLimit = (U32)(dms->window.nextSrc - dms->window.base); 175 U32 const dictLowLimit = dms->window.lowLimit; 176 U32 const dictIndexDelta = ms->window.lowLimit - dictHighLimit; 177 178 U32* const dictBt = dms->chainTable; 179 U32 const btLog = dmsCParams->chainLog - 1; 180 U32 const btMask = (1 << btLog) - 1; 181 U32 const btLow = (btMask >= dictHighLimit - dictLowLimit) ? dictLowLimit : dictHighLimit - btMask; 182 183 size_t commonLengthSmaller=0, commonLengthLarger=0; 184 185 (void)dictMode; 186 assert(dictMode == ZSTD_dictMatchState); 187 188 for (; nbCompares && (dictMatchIndex > dictLowLimit); --nbCompares) { 189 U32* const nextPtr = dictBt + 2*(dictMatchIndex & btMask); 190 size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */ 191 const BYTE* match = dictBase + dictMatchIndex; 192 matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart); 193 if (dictMatchIndex+matchLength >= dictHighLimit) 194 match = base + dictMatchIndex + dictIndexDelta; /* to prepare for next usage of match[matchLength] */ 195 196 if (matchLength > bestLength) { 197 U32 matchIndex = dictMatchIndex + dictIndexDelta; 198 if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(curr-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) ) { 199 DEBUGLOG(9, "ZSTD_DUBT_findBetterDictMatch(%u) : found better match length %u -> %u and offsetCode %u -> %u (dictMatchIndex %u, matchIndex %u)", 200 curr, (U32)bestLength, (U32)matchLength, (U32)*offsetPtr, STORE_OFFSET(curr - matchIndex), dictMatchIndex, matchIndex); 201 bestLength = matchLength, *offsetPtr = STORE_OFFSET(curr - matchIndex); 202 } 203 if (ip+matchLength == iend) { /* reached end of input : ip[matchLength] is not valid, no way to know if it's larger or smaller than match */ 204 break; /* drop, to guarantee consistency (miss a little bit of compression) */ 205 } 206 } 207 208 if (match[matchLength] < ip[matchLength]) { 209 if (dictMatchIndex <= btLow) { break; } /* beyond tree size, stop the search */ 210 commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */ 211 dictMatchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */ 212 } else { 213 /* match is larger than current */ 214 if (dictMatchIndex <= btLow) { break; } /* beyond tree size, stop the search */ 215 commonLengthLarger = matchLength; 216 dictMatchIndex = nextPtr[0]; 217 } 218 } 219 220 if (bestLength >= MINMATCH) { 221 U32 const mIndex = curr - (U32)STORED_OFFSET(*offsetPtr); (void)mIndex; 222 DEBUGLOG(8, "ZSTD_DUBT_findBetterDictMatch(%u) : found match of length %u and offsetCode %u (pos %u)", 223 curr, (U32)bestLength, (U32)*offsetPtr, mIndex); 224 } 225 return bestLength; 226 227 } 228 229 230 static size_t 231 ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms, 232 const BYTE* const ip, const BYTE* const iend, 233 size_t* offsetPtr, 234 U32 const mls, 235 const ZSTD_dictMode_e dictMode) 236 { 237 const ZSTD_compressionParameters* const cParams = &ms->cParams; 238 U32* const hashTable = ms->hashTable; 239 U32 const hashLog = cParams->hashLog; 240 size_t const h = ZSTD_hashPtr(ip, hashLog, mls); 241 U32 matchIndex = hashTable[h]; 242 243 const BYTE* const base = ms->window.base; 244 U32 const curr = (U32)(ip-base); 245 U32 const windowLow = ZSTD_getLowestMatchIndex(ms, curr, cParams->windowLog); 246 247 U32* const bt = ms->chainTable; 248 U32 const btLog = cParams->chainLog - 1; 249 U32 const btMask = (1 << btLog) - 1; 250 U32 const btLow = (btMask >= curr) ? 0 : curr - btMask; 251 U32 const unsortLimit = MAX(btLow, windowLow); 252 253 U32* nextCandidate = bt + 2*(matchIndex&btMask); 254 U32* unsortedMark = bt + 2*(matchIndex&btMask) + 1; 255 U32 nbCompares = 1U << cParams->searchLog; 256 U32 nbCandidates = nbCompares; 257 U32 previousCandidate = 0; 258 259 DEBUGLOG(7, "ZSTD_DUBT_findBestMatch (%u) ", curr); 260 assert(ip <= iend-8); /* required for h calculation */ 261 assert(dictMode != ZSTD_dedicatedDictSearch); 262 263 /* reach end of unsorted candidates list */ 264 while ( (matchIndex > unsortLimit) 265 && (*unsortedMark == ZSTD_DUBT_UNSORTED_MARK) 266 && (nbCandidates > 1) ) { 267 DEBUGLOG(8, "ZSTD_DUBT_findBestMatch: candidate %u is unsorted", 268 matchIndex); 269 *unsortedMark = previousCandidate; /* the unsortedMark becomes a reversed chain, to move up back to original position */ 270 previousCandidate = matchIndex; 271 matchIndex = *nextCandidate; 272 nextCandidate = bt + 2*(matchIndex&btMask); 273 unsortedMark = bt + 2*(matchIndex&btMask) + 1; 274 nbCandidates --; 275 } 276 277 /* nullify last candidate if it's still unsorted 278 * simplification, detrimental to compression ratio, beneficial for speed */ 279 if ( (matchIndex > unsortLimit) 280 && (*unsortedMark==ZSTD_DUBT_UNSORTED_MARK) ) { 281 DEBUGLOG(7, "ZSTD_DUBT_findBestMatch: nullify last unsorted candidate %u", 282 matchIndex); 283 *nextCandidate = *unsortedMark = 0; 284 } 285 286 /* batch sort stacked candidates */ 287 matchIndex = previousCandidate; 288 while (matchIndex) { /* will end on matchIndex == 0 */ 289 U32* const nextCandidateIdxPtr = bt + 2*(matchIndex&btMask) + 1; 290 U32 const nextCandidateIdx = *nextCandidateIdxPtr; 291 ZSTD_insertDUBT1(ms, matchIndex, iend, 292 nbCandidates, unsortLimit, dictMode); 293 matchIndex = nextCandidateIdx; 294 nbCandidates++; 295 } 296 297 /* find longest match */ 298 { size_t commonLengthSmaller = 0, commonLengthLarger = 0; 299 const BYTE* const dictBase = ms->window.dictBase; 300 const U32 dictLimit = ms->window.dictLimit; 301 const BYTE* const dictEnd = dictBase + dictLimit; 302 const BYTE* const prefixStart = base + dictLimit; 303 U32* smallerPtr = bt + 2*(curr&btMask); 304 U32* largerPtr = bt + 2*(curr&btMask) + 1; 305 U32 matchEndIdx = curr + 8 + 1; 306 U32 dummy32; /* to be nullified at the end */ 307 size_t bestLength = 0; 308 309 matchIndex = hashTable[h]; 310 hashTable[h] = curr; /* Update Hash Table */ 311 312 for (; nbCompares && (matchIndex > windowLow); --nbCompares) { 313 U32* const nextPtr = bt + 2*(matchIndex & btMask); 314 size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */ 315 const BYTE* match; 316 317 if ((dictMode != ZSTD_extDict) || (matchIndex+matchLength >= dictLimit)) { 318 match = base + matchIndex; 319 matchLength += ZSTD_count(ip+matchLength, match+matchLength, iend); 320 } else { 321 match = dictBase + matchIndex; 322 matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart); 323 if (matchIndex+matchLength >= dictLimit) 324 match = base + matchIndex; /* to prepare for next usage of match[matchLength] */ 325 } 326 327 if (matchLength > bestLength) { 328 if (matchLength > matchEndIdx - matchIndex) 329 matchEndIdx = matchIndex + (U32)matchLength; 330 if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(curr-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) ) 331 bestLength = matchLength, *offsetPtr = STORE_OFFSET(curr - matchIndex); 332 if (ip+matchLength == iend) { /* equal : no way to know if inf or sup */ 333 if (dictMode == ZSTD_dictMatchState) { 334 nbCompares = 0; /* in addition to avoiding checking any 335 * further in this loop, make sure we 336 * skip checking in the dictionary. */ 337 } 338 break; /* drop, to guarantee consistency (miss a little bit of compression) */ 339 } 340 } 341 342 if (match[matchLength] < ip[matchLength]) { 343 /* match is smaller than current */ 344 *smallerPtr = matchIndex; /* update smaller idx */ 345 commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */ 346 if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */ 347 smallerPtr = nextPtr+1; /* new "smaller" => larger of match */ 348 matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */ 349 } else { 350 /* match is larger than current */ 351 *largerPtr = matchIndex; 352 commonLengthLarger = matchLength; 353 if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */ 354 largerPtr = nextPtr; 355 matchIndex = nextPtr[0]; 356 } } 357 358 *smallerPtr = *largerPtr = 0; 359 360 assert(nbCompares <= (1U << ZSTD_SEARCHLOG_MAX)); /* Check we haven't underflowed. */ 361 if (dictMode == ZSTD_dictMatchState && nbCompares) { 362 bestLength = ZSTD_DUBT_findBetterDictMatch( 363 ms, ip, iend, 364 offsetPtr, bestLength, nbCompares, 365 mls, dictMode); 366 } 367 368 assert(matchEndIdx > curr+8); /* ensure nextToUpdate is increased */ 369 ms->nextToUpdate = matchEndIdx - 8; /* skip repetitive patterns */ 370 if (bestLength >= MINMATCH) { 371 U32 const mIndex = curr - (U32)STORED_OFFSET(*offsetPtr); (void)mIndex; 372 DEBUGLOG(8, "ZSTD_DUBT_findBestMatch(%u) : found match of length %u and offsetCode %u (pos %u)", 373 curr, (U32)bestLength, (U32)*offsetPtr, mIndex); 374 } 375 return bestLength; 376 } 377 } 378 379 380 /** ZSTD_BtFindBestMatch() : Tree updater, providing best match */ 381 FORCE_INLINE_TEMPLATE size_t 382 ZSTD_BtFindBestMatch( ZSTD_matchState_t* ms, 383 const BYTE* const ip, const BYTE* const iLimit, 384 size_t* offsetPtr, 385 const U32 mls /* template */, 386 const ZSTD_dictMode_e dictMode) 387 { 388 DEBUGLOG(7, "ZSTD_BtFindBestMatch"); 389 if (ip < ms->window.base + ms->nextToUpdate) return 0; /* skipped area */ 390 ZSTD_updateDUBT(ms, ip, iLimit, mls); 391 return ZSTD_DUBT_findBestMatch(ms, ip, iLimit, offsetPtr, mls, dictMode); 392 } 393 394 /*********************************** 395 * Dedicated dict search 396 ***********************************/ 397 398 void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_matchState_t* ms, const BYTE* const ip) 399 { 400 const BYTE* const base = ms->window.base; 401 U32 const target = (U32)(ip - base); 402 U32* const hashTable = ms->hashTable; 403 U32* const chainTable = ms->chainTable; 404 U32 const chainSize = 1 << ms->cParams.chainLog; 405 U32 idx = ms->nextToUpdate; 406 U32 const minChain = chainSize < target - idx ? target - chainSize : idx; 407 U32 const bucketSize = 1 << ZSTD_LAZY_DDSS_BUCKET_LOG; 408 U32 const cacheSize = bucketSize - 1; 409 U32 const chainAttempts = (1 << ms->cParams.searchLog) - cacheSize; 410 U32 const chainLimit = chainAttempts > 255 ? 255 : chainAttempts; 411 412 /* We know the hashtable is oversized by a factor of `bucketSize`. 413 * We are going to temporarily pretend `bucketSize == 1`, keeping only a 414 * single entry. We will use the rest of the space to construct a temporary 415 * chaintable. 416 */ 417 U32 const hashLog = ms->cParams.hashLog - ZSTD_LAZY_DDSS_BUCKET_LOG; 418 U32* const tmpHashTable = hashTable; 419 U32* const tmpChainTable = hashTable + ((size_t)1 << hashLog); 420 U32 const tmpChainSize = (U32)((1 << ZSTD_LAZY_DDSS_BUCKET_LOG) - 1) << hashLog; 421 U32 const tmpMinChain = tmpChainSize < target ? target - tmpChainSize : idx; 422 U32 hashIdx; 423 424 assert(ms->cParams.chainLog <= 24); 425 assert(ms->cParams.hashLog > ms->cParams.chainLog); 426 assert(idx != 0); 427 assert(tmpMinChain <= minChain); 428 429 /* fill conventional hash table and conventional chain table */ 430 for ( ; idx < target; idx++) { 431 U32 const h = (U32)ZSTD_hashPtr(base + idx, hashLog, ms->cParams.minMatch); 432 if (idx >= tmpMinChain) { 433 tmpChainTable[idx - tmpMinChain] = hashTable[h]; 434 } 435 tmpHashTable[h] = idx; 436 } 437 438 /* sort chains into ddss chain table */ 439 { 440 U32 chainPos = 0; 441 for (hashIdx = 0; hashIdx < (1U << hashLog); hashIdx++) { 442 U32 count; 443 U32 countBeyondMinChain = 0; 444 U32 i = tmpHashTable[hashIdx]; 445 for (count = 0; i >= tmpMinChain && count < cacheSize; count++) { 446 /* skip through the chain to the first position that won't be 447 * in the hash cache bucket */ 448 if (i < minChain) { 449 countBeyondMinChain++; 450 } 451 i = tmpChainTable[i - tmpMinChain]; 452 } 453 if (count == cacheSize) { 454 for (count = 0; count < chainLimit;) { 455 if (i < minChain) { 456 if (!i || ++countBeyondMinChain > cacheSize) { 457 /* only allow pulling `cacheSize` number of entries 458 * into the cache or chainTable beyond `minChain`, 459 * to replace the entries pulled out of the 460 * chainTable into the cache. This lets us reach 461 * back further without increasing the total number 462 * of entries in the chainTable, guaranteeing the 463 * DDSS chain table will fit into the space 464 * allocated for the regular one. */ 465 break; 466 } 467 } 468 chainTable[chainPos++] = i; 469 count++; 470 if (i < tmpMinChain) { 471 break; 472 } 473 i = tmpChainTable[i - tmpMinChain]; 474 } 475 } else { 476 count = 0; 477 } 478 if (count) { 479 tmpHashTable[hashIdx] = ((chainPos - count) << 8) + count; 480 } else { 481 tmpHashTable[hashIdx] = 0; 482 } 483 } 484 assert(chainPos <= chainSize); /* I believe this is guaranteed... */ 485 } 486 487 /* move chain pointers into the last entry of each hash bucket */ 488 for (hashIdx = (1 << hashLog); hashIdx; ) { 489 U32 const bucketIdx = --hashIdx << ZSTD_LAZY_DDSS_BUCKET_LOG; 490 U32 const chainPackedPointer = tmpHashTable[hashIdx]; 491 U32 i; 492 for (i = 0; i < cacheSize; i++) { 493 hashTable[bucketIdx + i] = 0; 494 } 495 hashTable[bucketIdx + bucketSize - 1] = chainPackedPointer; 496 } 497 498 /* fill the buckets of the hash table */ 499 for (idx = ms->nextToUpdate; idx < target; idx++) { 500 U32 const h = (U32)ZSTD_hashPtr(base + idx, hashLog, ms->cParams.minMatch) 501 << ZSTD_LAZY_DDSS_BUCKET_LOG; 502 U32 i; 503 /* Shift hash cache down 1. */ 504 for (i = cacheSize - 1; i; i--) 505 hashTable[h + i] = hashTable[h + i - 1]; 506 hashTable[h] = idx; 507 } 508 509 ms->nextToUpdate = target; 510 } 511 512 /* Returns the longest match length found in the dedicated dict search structure. 513 * If none are longer than the argument ml, then ml will be returned. 514 */ 515 FORCE_INLINE_TEMPLATE 516 size_t ZSTD_dedicatedDictSearch_lazy_search(size_t* offsetPtr, size_t ml, U32 nbAttempts, 517 const ZSTD_matchState_t* const dms, 518 const BYTE* const ip, const BYTE* const iLimit, 519 const BYTE* const prefixStart, const U32 curr, 520 const U32 dictLimit, const size_t ddsIdx) { 521 const U32 ddsLowestIndex = dms->window.dictLimit; 522 const BYTE* const ddsBase = dms->window.base; 523 const BYTE* const ddsEnd = dms->window.nextSrc; 524 const U32 ddsSize = (U32)(ddsEnd - ddsBase); 525 const U32 ddsIndexDelta = dictLimit - ddsSize; 526 const U32 bucketSize = (1 << ZSTD_LAZY_DDSS_BUCKET_LOG); 527 const U32 bucketLimit = nbAttempts < bucketSize - 1 ? nbAttempts : bucketSize - 1; 528 U32 ddsAttempt; 529 U32 matchIndex; 530 531 for (ddsAttempt = 0; ddsAttempt < bucketSize - 1; ddsAttempt++) { 532 PREFETCH_L1(ddsBase + dms->hashTable[ddsIdx + ddsAttempt]); 533 } 534 535 { 536 U32 const chainPackedPointer = dms->hashTable[ddsIdx + bucketSize - 1]; 537 U32 const chainIndex = chainPackedPointer >> 8; 538 539 PREFETCH_L1(&dms->chainTable[chainIndex]); 540 } 541 542 for (ddsAttempt = 0; ddsAttempt < bucketLimit; ddsAttempt++) { 543 size_t currentMl=0; 544 const BYTE* match; 545 matchIndex = dms->hashTable[ddsIdx + ddsAttempt]; 546 match = ddsBase + matchIndex; 547 548 if (!matchIndex) { 549 return ml; 550 } 551 552 /* guaranteed by table construction */ 553 (void)ddsLowestIndex; 554 assert(matchIndex >= ddsLowestIndex); 555 assert(match+4 <= ddsEnd); 556 if (MEM_read32(match) == MEM_read32(ip)) { 557 /* assumption : matchIndex <= dictLimit-4 (by table construction) */ 558 currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, ddsEnd, prefixStart) + 4; 559 } 560 561 /* save best solution */ 562 if (currentMl > ml) { 563 ml = currentMl; 564 *offsetPtr = STORE_OFFSET(curr - (matchIndex + ddsIndexDelta)); 565 if (ip+currentMl == iLimit) { 566 /* best possible, avoids read overflow on next attempt */ 567 return ml; 568 } 569 } 570 } 571 572 { 573 U32 const chainPackedPointer = dms->hashTable[ddsIdx + bucketSize - 1]; 574 U32 chainIndex = chainPackedPointer >> 8; 575 U32 const chainLength = chainPackedPointer & 0xFF; 576 U32 const chainAttempts = nbAttempts - ddsAttempt; 577 U32 const chainLimit = chainAttempts > chainLength ? chainLength : chainAttempts; 578 U32 chainAttempt; 579 580 for (chainAttempt = 0 ; chainAttempt < chainLimit; chainAttempt++) { 581 PREFETCH_L1(ddsBase + dms->chainTable[chainIndex + chainAttempt]); 582 } 583 584 for (chainAttempt = 0 ; chainAttempt < chainLimit; chainAttempt++, chainIndex++) { 585 size_t currentMl=0; 586 const BYTE* match; 587 matchIndex = dms->chainTable[chainIndex]; 588 match = ddsBase + matchIndex; 589 590 /* guaranteed by table construction */ 591 assert(matchIndex >= ddsLowestIndex); 592 assert(match+4 <= ddsEnd); 593 if (MEM_read32(match) == MEM_read32(ip)) { 594 /* assumption : matchIndex <= dictLimit-4 (by table construction) */ 595 currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, ddsEnd, prefixStart) + 4; 596 } 597 598 /* save best solution */ 599 if (currentMl > ml) { 600 ml = currentMl; 601 *offsetPtr = STORE_OFFSET(curr - (matchIndex + ddsIndexDelta)); 602 if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */ 603 } 604 } 605 } 606 return ml; 607 } 608 609 610 /* ********************************* 611 * Hash Chain 612 ***********************************/ 613 #define NEXT_IN_CHAIN(d, mask) chainTable[(d) & (mask)] 614 615 /* Update chains up to ip (excluded) 616 Assumption : always within prefix (i.e. not within extDict) */ 617 FORCE_INLINE_TEMPLATE U32 ZSTD_insertAndFindFirstIndex_internal( 618 ZSTD_matchState_t* ms, 619 const ZSTD_compressionParameters* const cParams, 620 const BYTE* ip, U32 const mls) 621 { 622 U32* const hashTable = ms->hashTable; 623 const U32 hashLog = cParams->hashLog; 624 U32* const chainTable = ms->chainTable; 625 const U32 chainMask = (1 << cParams->chainLog) - 1; 626 const BYTE* const base = ms->window.base; 627 const U32 target = (U32)(ip - base); 628 U32 idx = ms->nextToUpdate; 629 630 while(idx < target) { /* catch up */ 631 size_t const h = ZSTD_hashPtr(base+idx, hashLog, mls); 632 NEXT_IN_CHAIN(idx, chainMask) = hashTable[h]; 633 hashTable[h] = idx; 634 idx++; 635 } 636 637 ms->nextToUpdate = target; 638 return hashTable[ZSTD_hashPtr(ip, hashLog, mls)]; 639 } 640 641 U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip) { 642 const ZSTD_compressionParameters* const cParams = &ms->cParams; 643 return ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, ms->cParams.minMatch); 644 } 645 646 /* inlining is important to hardwire a hot branch (template emulation) */ 647 FORCE_INLINE_TEMPLATE 648 size_t ZSTD_HcFindBestMatch( 649 ZSTD_matchState_t* ms, 650 const BYTE* const ip, const BYTE* const iLimit, 651 size_t* offsetPtr, 652 const U32 mls, const ZSTD_dictMode_e dictMode) 653 { 654 const ZSTD_compressionParameters* const cParams = &ms->cParams; 655 U32* const chainTable = ms->chainTable; 656 const U32 chainSize = (1 << cParams->chainLog); 657 const U32 chainMask = chainSize-1; 658 const BYTE* const base = ms->window.base; 659 const BYTE* const dictBase = ms->window.dictBase; 660 const U32 dictLimit = ms->window.dictLimit; 661 const BYTE* const prefixStart = base + dictLimit; 662 const BYTE* const dictEnd = dictBase + dictLimit; 663 const U32 curr = (U32)(ip-base); 664 const U32 maxDistance = 1U << cParams->windowLog; 665 const U32 lowestValid = ms->window.lowLimit; 666 const U32 withinMaxDistance = (curr - lowestValid > maxDistance) ? curr - maxDistance : lowestValid; 667 const U32 isDictionary = (ms->loadedDictEnd != 0); 668 const U32 lowLimit = isDictionary ? lowestValid : withinMaxDistance; 669 const U32 minChain = curr > chainSize ? curr - chainSize : 0; 670 U32 nbAttempts = 1U << cParams->searchLog; 671 size_t ml=4-1; 672 673 const ZSTD_matchState_t* const dms = ms->dictMatchState; 674 const U32 ddsHashLog = dictMode == ZSTD_dedicatedDictSearch 675 ? dms->cParams.hashLog - ZSTD_LAZY_DDSS_BUCKET_LOG : 0; 676 const size_t ddsIdx = dictMode == ZSTD_dedicatedDictSearch 677 ? ZSTD_hashPtr(ip, ddsHashLog, mls) << ZSTD_LAZY_DDSS_BUCKET_LOG : 0; 678 679 U32 matchIndex; 680 681 if (dictMode == ZSTD_dedicatedDictSearch) { 682 const U32* entry = &dms->hashTable[ddsIdx]; 683 PREFETCH_L1(entry); 684 } 685 686 /* HC4 match finder */ 687 matchIndex = ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, mls); 688 689 for ( ; (matchIndex>=lowLimit) & (nbAttempts>0) ; nbAttempts--) { 690 size_t currentMl=0; 691 if ((dictMode != ZSTD_extDict) || matchIndex >= dictLimit) { 692 const BYTE* const match = base + matchIndex; 693 assert(matchIndex >= dictLimit); /* ensures this is true if dictMode != ZSTD_extDict */ 694 if (match[ml] == ip[ml]) /* potentially better */ 695 currentMl = ZSTD_count(ip, match, iLimit); 696 } else { 697 const BYTE* const match = dictBase + matchIndex; 698 assert(match+4 <= dictEnd); 699 if (MEM_read32(match) == MEM_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */ 700 currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dictEnd, prefixStart) + 4; 701 } 702 703 /* save best solution */ 704 if (currentMl > ml) { 705 ml = currentMl; 706 *offsetPtr = STORE_OFFSET(curr - matchIndex); 707 if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */ 708 } 709 710 if (matchIndex <= minChain) break; 711 matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask); 712 } 713 714 assert(nbAttempts <= (1U << ZSTD_SEARCHLOG_MAX)); /* Check we haven't underflowed. */ 715 if (dictMode == ZSTD_dedicatedDictSearch) { 716 ml = ZSTD_dedicatedDictSearch_lazy_search(offsetPtr, ml, nbAttempts, dms, 717 ip, iLimit, prefixStart, curr, dictLimit, ddsIdx); 718 } else if (dictMode == ZSTD_dictMatchState) { 719 const U32* const dmsChainTable = dms->chainTable; 720 const U32 dmsChainSize = (1 << dms->cParams.chainLog); 721 const U32 dmsChainMask = dmsChainSize - 1; 722 const U32 dmsLowestIndex = dms->window.dictLimit; 723 const BYTE* const dmsBase = dms->window.base; 724 const BYTE* const dmsEnd = dms->window.nextSrc; 725 const U32 dmsSize = (U32)(dmsEnd - dmsBase); 726 const U32 dmsIndexDelta = dictLimit - dmsSize; 727 const U32 dmsMinChain = dmsSize > dmsChainSize ? dmsSize - dmsChainSize : 0; 728 729 matchIndex = dms->hashTable[ZSTD_hashPtr(ip, dms->cParams.hashLog, mls)]; 730 731 for ( ; (matchIndex>=dmsLowestIndex) & (nbAttempts>0) ; nbAttempts--) { 732 size_t currentMl=0; 733 const BYTE* const match = dmsBase + matchIndex; 734 assert(match+4 <= dmsEnd); 735 if (MEM_read32(match) == MEM_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */ 736 currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dmsEnd, prefixStart) + 4; 737 738 /* save best solution */ 739 if (currentMl > ml) { 740 ml = currentMl; 741 assert(curr > matchIndex + dmsIndexDelta); 742 *offsetPtr = STORE_OFFSET(curr - (matchIndex + dmsIndexDelta)); 743 if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */ 744 } 745 746 if (matchIndex <= dmsMinChain) break; 747 748 matchIndex = dmsChainTable[matchIndex & dmsChainMask]; 749 } 750 } 751 752 return ml; 753 } 754 755 /* ********************************* 756 * (SIMD) Row-based matchfinder 757 ***********************************/ 758 /* Constants for row-based hash */ 759 #define ZSTD_ROW_HASH_TAG_OFFSET 16 /* byte offset of hashes in the match state's tagTable from the beginning of a row */ 760 #define ZSTD_ROW_HASH_TAG_BITS 8 /* nb bits to use for the tag */ 761 #define ZSTD_ROW_HASH_TAG_MASK ((1u << ZSTD_ROW_HASH_TAG_BITS) - 1) 762 #define ZSTD_ROW_HASH_MAX_ENTRIES 64 /* absolute maximum number of entries per row, for all configurations */ 763 764 #define ZSTD_ROW_HASH_CACHE_MASK (ZSTD_ROW_HASH_CACHE_SIZE - 1) 765 766 typedef U64 ZSTD_VecMask; /* Clarifies when we are interacting with a U64 representing a mask of matches */ 767 768 /* ZSTD_VecMask_next(): 769 * Starting from the LSB, returns the idx of the next non-zero bit. 770 * Basically counting the nb of trailing zeroes. 771 */ 772 static U32 ZSTD_VecMask_next(ZSTD_VecMask val) { 773 assert(val != 0); 774 # if defined(_MSC_VER) && defined(_WIN64) 775 if (val != 0) { 776 unsigned long r; 777 _BitScanForward64(&r, val); 778 return (U32)(r); 779 } else { 780 /* Should not reach this code path */ 781 __assume(0); 782 } 783 # elif (defined(__GNUC__) && ((__GNUC__ > 3) || ((__GNUC__ == 3) && (__GNUC_MINOR__ >= 4)))) 784 if (sizeof(size_t) == 4) { 785 U32 mostSignificantWord = (U32)(val >> 32); 786 U32 leastSignificantWord = (U32)val; 787 if (leastSignificantWord == 0) { 788 return 32 + (U32)__builtin_ctz(mostSignificantWord); 789 } else { 790 return (U32)__builtin_ctz(leastSignificantWord); 791 } 792 } else { 793 return (U32)__builtin_ctzll(val); 794 } 795 # else 796 /* Software ctz version: http://aggregate.org/MAGIC/#Trailing%20Zero%20Count 797 * and: https://stackoverflow.com/questions/2709430/count-number-of-bits-in-a-64-bit-long-big-integer 798 */ 799 val = ~val & (val - 1ULL); /* Lowest set bit mask */ 800 val = val - ((val >> 1) & 0x5555555555555555); 801 val = (val & 0x3333333333333333ULL) + ((val >> 2) & 0x3333333333333333ULL); 802 return (U32)((((val + (val >> 4)) & 0xF0F0F0F0F0F0F0FULL) * 0x101010101010101ULL) >> 56); 803 # endif 804 } 805 806 /* ZSTD_rotateRight_*(): 807 * Rotates a bitfield to the right by "count" bits. 808 * https://en.wikipedia.org/w/index.php?title=Circular_shift&oldid=991635599#Implementing_circular_shifts 809 */ 810 FORCE_INLINE_TEMPLATE 811 U64 ZSTD_rotateRight_U64(U64 const value, U32 count) { 812 assert(count < 64); 813 count &= 0x3F; /* for fickle pattern recognition */ 814 return (value >> count) | (U64)(value << ((0U - count) & 0x3F)); 815 } 816 817 FORCE_INLINE_TEMPLATE 818 U32 ZSTD_rotateRight_U32(U32 const value, U32 count) { 819 assert(count < 32); 820 count &= 0x1F; /* for fickle pattern recognition */ 821 return (value >> count) | (U32)(value << ((0U - count) & 0x1F)); 822 } 823 824 FORCE_INLINE_TEMPLATE 825 U16 ZSTD_rotateRight_U16(U16 const value, U32 count) { 826 assert(count < 16); 827 count &= 0x0F; /* for fickle pattern recognition */ 828 return (value >> count) | (U16)(value << ((0U - count) & 0x0F)); 829 } 830 831 /* ZSTD_row_nextIndex(): 832 * Returns the next index to insert at within a tagTable row, and updates the "head" 833 * value to reflect the update. Essentially cycles backwards from [0, {entries per row}) 834 */ 835 FORCE_INLINE_TEMPLATE U32 ZSTD_row_nextIndex(BYTE* const tagRow, U32 const rowMask) { 836 U32 const next = (*tagRow - 1) & rowMask; 837 *tagRow = (BYTE)next; 838 return next; 839 } 840 841 /* ZSTD_isAligned(): 842 * Checks that a pointer is aligned to "align" bytes which must be a power of 2. 843 */ 844 MEM_STATIC int ZSTD_isAligned(void const* ptr, size_t align) { 845 assert((align & (align - 1)) == 0); 846 return (((size_t)ptr) & (align - 1)) == 0; 847 } 848 849 /* ZSTD_row_prefetch(): 850 * Performs prefetching for the hashTable and tagTable at a given row. 851 */ 852 FORCE_INLINE_TEMPLATE void ZSTD_row_prefetch(U32 const* hashTable, U16 const* tagTable, U32 const relRow, U32 const rowLog) { 853 PREFETCH_L1(hashTable + relRow); 854 if (rowLog >= 5) { 855 PREFETCH_L1(hashTable + relRow + 16); 856 /* Note: prefetching more of the hash table does not appear to be beneficial for 128-entry rows */ 857 } 858 PREFETCH_L1(tagTable + relRow); 859 if (rowLog == 6) { 860 PREFETCH_L1(tagTable + relRow + 32); 861 } 862 assert(rowLog == 4 || rowLog == 5 || rowLog == 6); 863 assert(ZSTD_isAligned(hashTable + relRow, 64)); /* prefetched hash row always 64-byte aligned */ 864 assert(ZSTD_isAligned(tagTable + relRow, (size_t)1 << rowLog)); /* prefetched tagRow sits on correct multiple of bytes (32,64,128) */ 865 } 866 867 /* ZSTD_row_fillHashCache(): 868 * Fill up the hash cache starting at idx, prefetching up to ZSTD_ROW_HASH_CACHE_SIZE entries, 869 * but not beyond iLimit. 870 */ 871 FORCE_INLINE_TEMPLATE void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const BYTE* base, 872 U32 const rowLog, U32 const mls, 873 U32 idx, const BYTE* const iLimit) 874 { 875 U32 const* const hashTable = ms->hashTable; 876 U16 const* const tagTable = ms->tagTable; 877 U32 const hashLog = ms->rowHashLog; 878 U32 const maxElemsToPrefetch = (base + idx) > iLimit ? 0 : (U32)(iLimit - (base + idx) + 1); 879 U32 const lim = idx + MIN(ZSTD_ROW_HASH_CACHE_SIZE, maxElemsToPrefetch); 880 881 for (; idx < lim; ++idx) { 882 U32 const hash = (U32)ZSTD_hashPtr(base + idx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls); 883 U32 const row = (hash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog; 884 ZSTD_row_prefetch(hashTable, tagTable, row, rowLog); 885 ms->hashCache[idx & ZSTD_ROW_HASH_CACHE_MASK] = hash; 886 } 887 888 DEBUGLOG(6, "ZSTD_row_fillHashCache(): [%u %u %u %u %u %u %u %u]", ms->hashCache[0], ms->hashCache[1], 889 ms->hashCache[2], ms->hashCache[3], ms->hashCache[4], 890 ms->hashCache[5], ms->hashCache[6], ms->hashCache[7]); 891 } 892 893 /* ZSTD_row_nextCachedHash(): 894 * Returns the hash of base + idx, and replaces the hash in the hash cache with the byte at 895 * base + idx + ZSTD_ROW_HASH_CACHE_SIZE. Also prefetches the appropriate rows from hashTable and tagTable. 896 */ 897 FORCE_INLINE_TEMPLATE U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTable, 898 U16 const* tagTable, BYTE const* base, 899 U32 idx, U32 const hashLog, 900 U32 const rowLog, U32 const mls) 901 { 902 U32 const newHash = (U32)ZSTD_hashPtr(base+idx+ZSTD_ROW_HASH_CACHE_SIZE, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls); 903 U32 const row = (newHash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog; 904 ZSTD_row_prefetch(hashTable, tagTable, row, rowLog); 905 { U32 const hash = cache[idx & ZSTD_ROW_HASH_CACHE_MASK]; 906 cache[idx & ZSTD_ROW_HASH_CACHE_MASK] = newHash; 907 return hash; 908 } 909 } 910 911 /* ZSTD_row_update_internalImpl(): 912 * Updates the hash table with positions starting from updateStartIdx until updateEndIdx. 913 */ 914 FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms, 915 U32 updateStartIdx, U32 const updateEndIdx, 916 U32 const mls, U32 const rowLog, 917 U32 const rowMask, U32 const useCache) 918 { 919 U32* const hashTable = ms->hashTable; 920 U16* const tagTable = ms->tagTable; 921 U32 const hashLog = ms->rowHashLog; 922 const BYTE* const base = ms->window.base; 923 924 DEBUGLOG(6, "ZSTD_row_update_internalImpl(): updateStartIdx=%u, updateEndIdx=%u", updateStartIdx, updateEndIdx); 925 for (; updateStartIdx < updateEndIdx; ++updateStartIdx) { 926 U32 const hash = useCache ? ZSTD_row_nextCachedHash(ms->hashCache, hashTable, tagTable, base, updateStartIdx, hashLog, rowLog, mls) 927 : (U32)ZSTD_hashPtr(base + updateStartIdx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls); 928 U32 const relRow = (hash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog; 929 U32* const row = hashTable + relRow; 930 BYTE* tagRow = (BYTE*)(tagTable + relRow); /* Though tagTable is laid out as a table of U16, each tag is only 1 byte. 931 Explicit cast allows us to get exact desired position within each row */ 932 U32 const pos = ZSTD_row_nextIndex(tagRow, rowMask); 933 934 assert(hash == ZSTD_hashPtr(base + updateStartIdx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls)); 935 ((BYTE*)tagRow)[pos + ZSTD_ROW_HASH_TAG_OFFSET] = hash & ZSTD_ROW_HASH_TAG_MASK; 936 row[pos] = updateStartIdx; 937 } 938 } 939 940 /* ZSTD_row_update_internal(): 941 * Inserts the byte at ip into the appropriate position in the hash table, and updates ms->nextToUpdate. 942 * Skips sections of long matches as is necessary. 943 */ 944 FORCE_INLINE_TEMPLATE void ZSTD_row_update_internal(ZSTD_matchState_t* ms, const BYTE* ip, 945 U32 const mls, U32 const rowLog, 946 U32 const rowMask, U32 const useCache) 947 { 948 U32 idx = ms->nextToUpdate; 949 const BYTE* const base = ms->window.base; 950 const U32 target = (U32)(ip - base); 951 const U32 kSkipThreshold = 384; 952 const U32 kMaxMatchStartPositionsToUpdate = 96; 953 const U32 kMaxMatchEndPositionsToUpdate = 32; 954 955 if (useCache) { 956 /* Only skip positions when using hash cache, i.e. 957 * if we are loading a dict, don't skip anything. 958 * If we decide to skip, then we only update a set number 959 * of positions at the beginning and end of the match. 960 */ 961 if (UNLIKELY(target - idx > kSkipThreshold)) { 962 U32 const bound = idx + kMaxMatchStartPositionsToUpdate; 963 ZSTD_row_update_internalImpl(ms, idx, bound, mls, rowLog, rowMask, useCache); 964 idx = target - kMaxMatchEndPositionsToUpdate; 965 ZSTD_row_fillHashCache(ms, base, rowLog, mls, idx, ip+1); 966 } 967 } 968 assert(target >= idx); 969 ZSTD_row_update_internalImpl(ms, idx, target, mls, rowLog, rowMask, useCache); 970 ms->nextToUpdate = target; 971 } 972 973 /* ZSTD_row_update(): 974 * External wrapper for ZSTD_row_update_internal(). Used for filling the hashtable during dictionary 975 * processing. 976 */ 977 void ZSTD_row_update(ZSTD_matchState_t* const ms, const BYTE* ip) { 978 const U32 rowLog = BOUNDED(4, ms->cParams.searchLog, 6); 979 const U32 rowMask = (1u << rowLog) - 1; 980 const U32 mls = MIN(ms->cParams.minMatch, 6 /* mls caps out at 6 */); 981 982 DEBUGLOG(5, "ZSTD_row_update(), rowLog=%u", rowLog); 983 ZSTD_row_update_internal(ms, ip, mls, rowLog, rowMask, 0 /* dont use cache */); 984 } 985 986 #if defined(ZSTD_ARCH_X86_SSE2) 987 FORCE_INLINE_TEMPLATE ZSTD_VecMask 988 ZSTD_row_getSSEMask(int nbChunks, const BYTE* const src, const BYTE tag, const U32 head) 989 { 990 const __m128i comparisonMask = _mm_set1_epi8((char)tag); 991 int matches[4] = {0}; 992 int i; 993 assert(nbChunks == 1 || nbChunks == 2 || nbChunks == 4); 994 for (i=0; i<nbChunks; i++) { 995 const __m128i chunk = _mm_loadu_si128((const __m128i*)(const void*)(src + 16*i)); 996 const __m128i equalMask = _mm_cmpeq_epi8(chunk, comparisonMask); 997 matches[i] = _mm_movemask_epi8(equalMask); 998 } 999 if (nbChunks == 1) return ZSTD_rotateRight_U16((U16)matches[0], head); 1000 if (nbChunks == 2) return ZSTD_rotateRight_U32((U32)matches[1] << 16 | (U32)matches[0], head); 1001 assert(nbChunks == 4); 1002 return ZSTD_rotateRight_U64((U64)matches[3] << 48 | (U64)matches[2] << 32 | (U64)matches[1] << 16 | (U64)matches[0], head); 1003 } 1004 #endif 1005 1006 /* Returns a ZSTD_VecMask (U32) that has the nth bit set to 1 if the newly-computed "tag" matches 1007 * the hash at the nth position in a row of the tagTable. 1008 * Each row is a circular buffer beginning at the value of "head". So we must rotate the "matches" bitfield 1009 * to match up with the actual layout of the entries within the hashTable */ 1010 FORCE_INLINE_TEMPLATE ZSTD_VecMask 1011 ZSTD_row_getMatchMask(const BYTE* const tagRow, const BYTE tag, const U32 head, const U32 rowEntries) 1012 { 1013 const BYTE* const src = tagRow + ZSTD_ROW_HASH_TAG_OFFSET; 1014 assert((rowEntries == 16) || (rowEntries == 32) || rowEntries == 64); 1015 assert(rowEntries <= ZSTD_ROW_HASH_MAX_ENTRIES); 1016 1017 #if defined(ZSTD_ARCH_X86_SSE2) 1018 1019 return ZSTD_row_getSSEMask(rowEntries / 16, src, tag, head); 1020 1021 #else /* SW or NEON-LE */ 1022 1023 # if defined(ZSTD_ARCH_ARM_NEON) 1024 /* This NEON path only works for little endian - otherwise use SWAR below */ 1025 if (MEM_isLittleEndian()) { 1026 if (rowEntries == 16) { 1027 const uint8x16_t chunk = vld1q_u8(src); 1028 const uint16x8_t equalMask = vreinterpretq_u16_u8(vceqq_u8(chunk, vdupq_n_u8(tag))); 1029 const uint16x8_t t0 = vshlq_n_u16(equalMask, 7); 1030 const uint32x4_t t1 = vreinterpretq_u32_u16(vsriq_n_u16(t0, t0, 14)); 1031 const uint64x2_t t2 = vreinterpretq_u64_u32(vshrq_n_u32(t1, 14)); 1032 const uint8x16_t t3 = vreinterpretq_u8_u64(vsraq_n_u64(t2, t2, 28)); 1033 const U16 hi = (U16)vgetq_lane_u8(t3, 8); 1034 const U16 lo = (U16)vgetq_lane_u8(t3, 0); 1035 return ZSTD_rotateRight_U16((hi << 8) | lo, head); 1036 } else if (rowEntries == 32) { 1037 const uint16x8x2_t chunk = vld2q_u16((const U16*)(const void*)src); 1038 const uint8x16_t chunk0 = vreinterpretq_u8_u16(chunk.val[0]); 1039 const uint8x16_t chunk1 = vreinterpretq_u8_u16(chunk.val[1]); 1040 const uint8x16_t equalMask0 = vceqq_u8(chunk0, vdupq_n_u8(tag)); 1041 const uint8x16_t equalMask1 = vceqq_u8(chunk1, vdupq_n_u8(tag)); 1042 const int8x8_t pack0 = vqmovn_s16(vreinterpretq_s16_u8(equalMask0)); 1043 const int8x8_t pack1 = vqmovn_s16(vreinterpretq_s16_u8(equalMask1)); 1044 const uint8x8_t t0 = vreinterpret_u8_s8(pack0); 1045 const uint8x8_t t1 = vreinterpret_u8_s8(pack1); 1046 const uint8x8_t t2 = vsri_n_u8(t1, t0, 2); 1047 const uint8x8x2_t t3 = vuzp_u8(t2, t0); 1048 const uint8x8_t t4 = vsri_n_u8(t3.val[1], t3.val[0], 4); 1049 const U32 matches = vget_lane_u32(vreinterpret_u32_u8(t4), 0); 1050 return ZSTD_rotateRight_U32(matches, head); 1051 } else { /* rowEntries == 64 */ 1052 const uint8x16x4_t chunk = vld4q_u8(src); 1053 const uint8x16_t dup = vdupq_n_u8(tag); 1054 const uint8x16_t cmp0 = vceqq_u8(chunk.val[0], dup); 1055 const uint8x16_t cmp1 = vceqq_u8(chunk.val[1], dup); 1056 const uint8x16_t cmp2 = vceqq_u8(chunk.val[2], dup); 1057 const uint8x16_t cmp3 = vceqq_u8(chunk.val[3], dup); 1058 1059 const uint8x16_t t0 = vsriq_n_u8(cmp1, cmp0, 1); 1060 const uint8x16_t t1 = vsriq_n_u8(cmp3, cmp2, 1); 1061 const uint8x16_t t2 = vsriq_n_u8(t1, t0, 2); 1062 const uint8x16_t t3 = vsriq_n_u8(t2, t2, 4); 1063 const uint8x8_t t4 = vshrn_n_u16(vreinterpretq_u16_u8(t3), 4); 1064 const U64 matches = vget_lane_u64(vreinterpret_u64_u8(t4), 0); 1065 return ZSTD_rotateRight_U64(matches, head); 1066 } 1067 } 1068 # endif /* ZSTD_ARCH_ARM_NEON */ 1069 /* SWAR */ 1070 { const size_t chunkSize = sizeof(size_t); 1071 const size_t shiftAmount = ((chunkSize * 8) - chunkSize); 1072 const size_t xFF = ~((size_t)0); 1073 const size_t x01 = xFF / 0xFF; 1074 const size_t x80 = x01 << 7; 1075 const size_t splatChar = tag * x01; 1076 ZSTD_VecMask matches = 0; 1077 int i = rowEntries - chunkSize; 1078 assert((sizeof(size_t) == 4) || (sizeof(size_t) == 8)); 1079 if (MEM_isLittleEndian()) { /* runtime check so have two loops */ 1080 const size_t extractMagic = (xFF / 0x7F) >> chunkSize; 1081 do { 1082 size_t chunk = MEM_readST(&src[i]); 1083 chunk ^= splatChar; 1084 chunk = (((chunk | x80) - x01) | chunk) & x80; 1085 matches <<= chunkSize; 1086 matches |= (chunk * extractMagic) >> shiftAmount; 1087 i -= chunkSize; 1088 } while (i >= 0); 1089 } else { /* big endian: reverse bits during extraction */ 1090 const size_t msb = xFF ^ (xFF >> 1); 1091 const size_t extractMagic = (msb / 0x1FF) | msb; 1092 do { 1093 size_t chunk = MEM_readST(&src[i]); 1094 chunk ^= splatChar; 1095 chunk = (((chunk | x80) - x01) | chunk) & x80; 1096 matches <<= chunkSize; 1097 matches |= ((chunk >> 7) * extractMagic) >> shiftAmount; 1098 i -= chunkSize; 1099 } while (i >= 0); 1100 } 1101 matches = ~matches; 1102 if (rowEntries == 16) { 1103 return ZSTD_rotateRight_U16((U16)matches, head); 1104 } else if (rowEntries == 32) { 1105 return ZSTD_rotateRight_U32((U32)matches, head); 1106 } else { 1107 return ZSTD_rotateRight_U64((U64)matches, head); 1108 } 1109 } 1110 #endif 1111 } 1112 1113 /* The high-level approach of the SIMD row based match finder is as follows: 1114 * - Figure out where to insert the new entry: 1115 * - Generate a hash from a byte along with an additional 1-byte "short hash". The additional byte is our "tag" 1116 * - The hashTable is effectively split into groups or "rows" of 16 or 32 entries of U32, and the hash determines 1117 * which row to insert into. 1118 * - Determine the correct position within the row to insert the entry into. Each row of 16 or 32 can 1119 * be considered as a circular buffer with a "head" index that resides in the tagTable. 1120 * - Also insert the "tag" into the equivalent row and position in the tagTable. 1121 * - Note: The tagTable has 17 or 33 1-byte entries per row, due to 16 or 32 tags, and 1 "head" entry. 1122 * The 17 or 33 entry rows are spaced out to occur every 32 or 64 bytes, respectively, 1123 * for alignment/performance reasons, leaving some bytes unused. 1124 * - Use SIMD to efficiently compare the tags in the tagTable to the 1-byte "short hash" and 1125 * generate a bitfield that we can cycle through to check the collisions in the hash table. 1126 * - Pick the longest match. 1127 */ 1128 FORCE_INLINE_TEMPLATE 1129 size_t ZSTD_RowFindBestMatch( 1130 ZSTD_matchState_t* ms, 1131 const BYTE* const ip, const BYTE* const iLimit, 1132 size_t* offsetPtr, 1133 const U32 mls, const ZSTD_dictMode_e dictMode, 1134 const U32 rowLog) 1135 { 1136 U32* const hashTable = ms->hashTable; 1137 U16* const tagTable = ms->tagTable; 1138 U32* const hashCache = ms->hashCache; 1139 const U32 hashLog = ms->rowHashLog; 1140 const ZSTD_compressionParameters* const cParams = &ms->cParams; 1141 const BYTE* const base = ms->window.base; 1142 const BYTE* const dictBase = ms->window.dictBase; 1143 const U32 dictLimit = ms->window.dictLimit; 1144 const BYTE* const prefixStart = base + dictLimit; 1145 const BYTE* const dictEnd = dictBase + dictLimit; 1146 const U32 curr = (U32)(ip-base); 1147 const U32 maxDistance = 1U << cParams->windowLog; 1148 const U32 lowestValid = ms->window.lowLimit; 1149 const U32 withinMaxDistance = (curr - lowestValid > maxDistance) ? curr - maxDistance : lowestValid; 1150 const U32 isDictionary = (ms->loadedDictEnd != 0); 1151 const U32 lowLimit = isDictionary ? lowestValid : withinMaxDistance; 1152 const U32 rowEntries = (1U << rowLog); 1153 const U32 rowMask = rowEntries - 1; 1154 const U32 cappedSearchLog = MIN(cParams->searchLog, rowLog); /* nb of searches is capped at nb entries per row */ 1155 U32 nbAttempts = 1U << cappedSearchLog; 1156 size_t ml=4-1; 1157 1158 /* DMS/DDS variables that may be referenced laster */ 1159 const ZSTD_matchState_t* const dms = ms->dictMatchState; 1160 1161 /* Initialize the following variables to satisfy static analyzer */ 1162 size_t ddsIdx = 0; 1163 U32 ddsExtraAttempts = 0; /* cctx hash tables are limited in searches, but allow extra searches into DDS */ 1164 U32 dmsTag = 0; 1165 U32* dmsRow = NULL; 1166 BYTE* dmsTagRow = NULL; 1167 1168 if (dictMode == ZSTD_dedicatedDictSearch) { 1169 const U32 ddsHashLog = dms->cParams.hashLog - ZSTD_LAZY_DDSS_BUCKET_LOG; 1170 { /* Prefetch DDS hashtable entry */ 1171 ddsIdx = ZSTD_hashPtr(ip, ddsHashLog, mls) << ZSTD_LAZY_DDSS_BUCKET_LOG; 1172 PREFETCH_L1(&dms->hashTable[ddsIdx]); 1173 } 1174 ddsExtraAttempts = cParams->searchLog > rowLog ? 1U << (cParams->searchLog - rowLog) : 0; 1175 } 1176 1177 if (dictMode == ZSTD_dictMatchState) { 1178 /* Prefetch DMS rows */ 1179 U32* const dmsHashTable = dms->hashTable; 1180 U16* const dmsTagTable = dms->tagTable; 1181 U32 const dmsHash = (U32)ZSTD_hashPtr(ip, dms->rowHashLog + ZSTD_ROW_HASH_TAG_BITS, mls); 1182 U32 const dmsRelRow = (dmsHash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog; 1183 dmsTag = dmsHash & ZSTD_ROW_HASH_TAG_MASK; 1184 dmsTagRow = (BYTE*)(dmsTagTable + dmsRelRow); 1185 dmsRow = dmsHashTable + dmsRelRow; 1186 ZSTD_row_prefetch(dmsHashTable, dmsTagTable, dmsRelRow, rowLog); 1187 } 1188 1189 /* Update the hashTable and tagTable up to (but not including) ip */ 1190 ZSTD_row_update_internal(ms, ip, mls, rowLog, rowMask, 1 /* useCache */); 1191 { /* Get the hash for ip, compute the appropriate row */ 1192 U32 const hash = ZSTD_row_nextCachedHash(hashCache, hashTable, tagTable, base, curr, hashLog, rowLog, mls); 1193 U32 const relRow = (hash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog; 1194 U32 const tag = hash & ZSTD_ROW_HASH_TAG_MASK; 1195 U32* const row = hashTable + relRow; 1196 BYTE* tagRow = (BYTE*)(tagTable + relRow); 1197 U32 const head = *tagRow & rowMask; 1198 U32 matchBuffer[ZSTD_ROW_HASH_MAX_ENTRIES]; 1199 size_t numMatches = 0; 1200 size_t currMatch = 0; 1201 ZSTD_VecMask matches = ZSTD_row_getMatchMask(tagRow, (BYTE)tag, head, rowEntries); 1202 1203 /* Cycle through the matches and prefetch */ 1204 for (; (matches > 0) && (nbAttempts > 0); --nbAttempts, matches &= (matches - 1)) { 1205 U32 const matchPos = (head + ZSTD_VecMask_next(matches)) & rowMask; 1206 U32 const matchIndex = row[matchPos]; 1207 assert(numMatches < rowEntries); 1208 if (matchIndex < lowLimit) 1209 break; 1210 if ((dictMode != ZSTD_extDict) || matchIndex >= dictLimit) { 1211 PREFETCH_L1(base + matchIndex); 1212 } else { 1213 PREFETCH_L1(dictBase + matchIndex); 1214 } 1215 matchBuffer[numMatches++] = matchIndex; 1216 } 1217 1218 /* Speed opt: insert current byte into hashtable too. This allows us to avoid one iteration of the loop 1219 in ZSTD_row_update_internal() at the next search. */ 1220 { 1221 U32 const pos = ZSTD_row_nextIndex(tagRow, rowMask); 1222 tagRow[pos + ZSTD_ROW_HASH_TAG_OFFSET] = (BYTE)tag; 1223 row[pos] = ms->nextToUpdate++; 1224 } 1225 1226 /* Return the longest match */ 1227 for (; currMatch < numMatches; ++currMatch) { 1228 U32 const matchIndex = matchBuffer[currMatch]; 1229 size_t currentMl=0; 1230 assert(matchIndex < curr); 1231 assert(matchIndex >= lowLimit); 1232 1233 if ((dictMode != ZSTD_extDict) || matchIndex >= dictLimit) { 1234 const BYTE* const match = base + matchIndex; 1235 assert(matchIndex >= dictLimit); /* ensures this is true if dictMode != ZSTD_extDict */ 1236 if (match[ml] == ip[ml]) /* potentially better */ 1237 currentMl = ZSTD_count(ip, match, iLimit); 1238 } else { 1239 const BYTE* const match = dictBase + matchIndex; 1240 assert(match+4 <= dictEnd); 1241 if (MEM_read32(match) == MEM_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */ 1242 currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dictEnd, prefixStart) + 4; 1243 } 1244 1245 /* Save best solution */ 1246 if (currentMl > ml) { 1247 ml = currentMl; 1248 *offsetPtr = STORE_OFFSET(curr - matchIndex); 1249 if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */ 1250 } 1251 } 1252 } 1253 1254 assert(nbAttempts <= (1U << ZSTD_SEARCHLOG_MAX)); /* Check we haven't underflowed. */ 1255 if (dictMode == ZSTD_dedicatedDictSearch) { 1256 ml = ZSTD_dedicatedDictSearch_lazy_search(offsetPtr, ml, nbAttempts + ddsExtraAttempts, dms, 1257 ip, iLimit, prefixStart, curr, dictLimit, ddsIdx); 1258 } else if (dictMode == ZSTD_dictMatchState) { 1259 /* TODO: Measure and potentially add prefetching to DMS */ 1260 const U32 dmsLowestIndex = dms->window.dictLimit; 1261 const BYTE* const dmsBase = dms->window.base; 1262 const BYTE* const dmsEnd = dms->window.nextSrc; 1263 const U32 dmsSize = (U32)(dmsEnd - dmsBase); 1264 const U32 dmsIndexDelta = dictLimit - dmsSize; 1265 1266 { U32 const head = *dmsTagRow & rowMask; 1267 U32 matchBuffer[ZSTD_ROW_HASH_MAX_ENTRIES]; 1268 size_t numMatches = 0; 1269 size_t currMatch = 0; 1270 ZSTD_VecMask matches = ZSTD_row_getMatchMask(dmsTagRow, (BYTE)dmsTag, head, rowEntries); 1271 1272 for (; (matches > 0) && (nbAttempts > 0); --nbAttempts, matches &= (matches - 1)) { 1273 U32 const matchPos = (head + ZSTD_VecMask_next(matches)) & rowMask; 1274 U32 const matchIndex = dmsRow[matchPos]; 1275 if (matchIndex < dmsLowestIndex) 1276 break; 1277 PREFETCH_L1(dmsBase + matchIndex); 1278 matchBuffer[numMatches++] = matchIndex; 1279 } 1280 1281 /* Return the longest match */ 1282 for (; currMatch < numMatches; ++currMatch) { 1283 U32 const matchIndex = matchBuffer[currMatch]; 1284 size_t currentMl=0; 1285 assert(matchIndex >= dmsLowestIndex); 1286 assert(matchIndex < curr); 1287 1288 { const BYTE* const match = dmsBase + matchIndex; 1289 assert(match+4 <= dmsEnd); 1290 if (MEM_read32(match) == MEM_read32(ip)) 1291 currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dmsEnd, prefixStart) + 4; 1292 } 1293 1294 if (currentMl > ml) { 1295 ml = currentMl; 1296 assert(curr > matchIndex + dmsIndexDelta); 1297 *offsetPtr = STORE_OFFSET(curr - (matchIndex + dmsIndexDelta)); 1298 if (ip+currentMl == iLimit) break; 1299 } 1300 } 1301 } 1302 } 1303 return ml; 1304 } 1305 1306 1307 typedef size_t (*searchMax_f)( 1308 ZSTD_matchState_t* ms, 1309 const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr); 1310 1311 /** 1312 * This struct contains the functions necessary for lazy to search. 1313 * Currently, that is only searchMax. However, it is still valuable to have the 1314 * VTable because this makes it easier to add more functions to the VTable later. 1315 * 1316 * TODO: The start of the search function involves loading and calculating a 1317 * bunch of constants from the ZSTD_matchState_t. These computations could be 1318 * done in an initialization function, and saved somewhere in the match state. 1319 * Then we could pass a pointer to the saved state instead of the match state, 1320 * and avoid duplicate computations. 1321 * 1322 * TODO: Move the match re-winding into searchMax. This improves compression 1323 * ratio, and unlocks further simplifications with the next TODO. 1324 * 1325 * TODO: Try moving the repcode search into searchMax. After the re-winding 1326 * and repcode search are in searchMax, there is no more logic in the match 1327 * finder loop that requires knowledge about the dictMode. So we should be 1328 * able to avoid force inlining it, and we can join the extDict loop with 1329 * the single segment loop. It should go in searchMax instead of its own 1330 * function to avoid having multiple virtual function calls per search. 1331 */ 1332 typedef struct { 1333 searchMax_f searchMax; 1334 } ZSTD_LazyVTable; 1335 1336 #define GEN_ZSTD_BT_VTABLE(dictMode, mls) \ 1337 static size_t ZSTD_BtFindBestMatch_##dictMode##_##mls( \ 1338 ZSTD_matchState_t* ms, \ 1339 const BYTE* ip, const BYTE* const iLimit, \ 1340 size_t* offsetPtr) \ 1341 { \ 1342 assert(MAX(4, MIN(6, ms->cParams.minMatch)) == mls); \ 1343 return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, mls, ZSTD_##dictMode); \ 1344 } \ 1345 static const ZSTD_LazyVTable ZSTD_BtVTable_##dictMode##_##mls = { \ 1346 ZSTD_BtFindBestMatch_##dictMode##_##mls \ 1347 }; 1348 1349 #define GEN_ZSTD_HC_VTABLE(dictMode, mls) \ 1350 static size_t ZSTD_HcFindBestMatch_##dictMode##_##mls( \ 1351 ZSTD_matchState_t* ms, \ 1352 const BYTE* ip, const BYTE* const iLimit, \ 1353 size_t* offsetPtr) \ 1354 { \ 1355 assert(MAX(4, MIN(6, ms->cParams.minMatch)) == mls); \ 1356 return ZSTD_HcFindBestMatch(ms, ip, iLimit, offsetPtr, mls, ZSTD_##dictMode); \ 1357 } \ 1358 static const ZSTD_LazyVTable ZSTD_HcVTable_##dictMode##_##mls = { \ 1359 ZSTD_HcFindBestMatch_##dictMode##_##mls \ 1360 }; 1361 1362 #define GEN_ZSTD_ROW_VTABLE(dictMode, mls, rowLog) \ 1363 static size_t ZSTD_RowFindBestMatch_##dictMode##_##mls##_##rowLog( \ 1364 ZSTD_matchState_t* ms, \ 1365 const BYTE* ip, const BYTE* const iLimit, \ 1366 size_t* offsetPtr) \ 1367 { \ 1368 assert(MAX(4, MIN(6, ms->cParams.minMatch)) == mls); \ 1369 assert(MAX(4, MIN(6, ms->cParams.searchLog)) == rowLog); \ 1370 return ZSTD_RowFindBestMatch(ms, ip, iLimit, offsetPtr, mls, ZSTD_##dictMode, rowLog); \ 1371 } \ 1372 static const ZSTD_LazyVTable ZSTD_RowVTable_##dictMode##_##mls##_##rowLog = { \ 1373 ZSTD_RowFindBestMatch_##dictMode##_##mls##_##rowLog \ 1374 }; 1375 1376 #define ZSTD_FOR_EACH_ROWLOG(X, dictMode, mls) \ 1377 X(dictMode, mls, 4) \ 1378 X(dictMode, mls, 5) \ 1379 X(dictMode, mls, 6) 1380 1381 #define ZSTD_FOR_EACH_MLS_ROWLOG(X, dictMode) \ 1382 ZSTD_FOR_EACH_ROWLOG(X, dictMode, 4) \ 1383 ZSTD_FOR_EACH_ROWLOG(X, dictMode, 5) \ 1384 ZSTD_FOR_EACH_ROWLOG(X, dictMode, 6) 1385 1386 #define ZSTD_FOR_EACH_MLS(X, dictMode) \ 1387 X(dictMode, 4) \ 1388 X(dictMode, 5) \ 1389 X(dictMode, 6) 1390 1391 #define ZSTD_FOR_EACH_DICT_MODE(X, ...) \ 1392 X(__VA_ARGS__, noDict) \ 1393 X(__VA_ARGS__, extDict) \ 1394 X(__VA_ARGS__, dictMatchState) \ 1395 X(__VA_ARGS__, dedicatedDictSearch) 1396 1397 /* Generate Row VTables for each combination of (dictMode, mls, rowLog) */ 1398 ZSTD_FOR_EACH_DICT_MODE(ZSTD_FOR_EACH_MLS_ROWLOG, GEN_ZSTD_ROW_VTABLE) 1399 /* Generate Binary Tree VTables for each combination of (dictMode, mls) */ 1400 ZSTD_FOR_EACH_DICT_MODE(ZSTD_FOR_EACH_MLS, GEN_ZSTD_BT_VTABLE) 1401 /* Generate Hash Chain VTables for each combination of (dictMode, mls) */ 1402 ZSTD_FOR_EACH_DICT_MODE(ZSTD_FOR_EACH_MLS, GEN_ZSTD_HC_VTABLE) 1403 1404 #define GEN_ZSTD_BT_VTABLE_ARRAY(dictMode) \ 1405 { \ 1406 &ZSTD_BtVTable_##dictMode##_4, \ 1407 &ZSTD_BtVTable_##dictMode##_5, \ 1408 &ZSTD_BtVTable_##dictMode##_6 \ 1409 } 1410 1411 #define GEN_ZSTD_HC_VTABLE_ARRAY(dictMode) \ 1412 { \ 1413 &ZSTD_HcVTable_##dictMode##_4, \ 1414 &ZSTD_HcVTable_##dictMode##_5, \ 1415 &ZSTD_HcVTable_##dictMode##_6 \ 1416 } 1417 1418 #define GEN_ZSTD_ROW_VTABLE_ARRAY_(dictMode, mls) \ 1419 { \ 1420 &ZSTD_RowVTable_##dictMode##_##mls##_4, \ 1421 &ZSTD_RowVTable_##dictMode##_##mls##_5, \ 1422 &ZSTD_RowVTable_##dictMode##_##mls##_6 \ 1423 } 1424 1425 #define GEN_ZSTD_ROW_VTABLE_ARRAY(dictMode) \ 1426 { \ 1427 GEN_ZSTD_ROW_VTABLE_ARRAY_(dictMode, 4), \ 1428 GEN_ZSTD_ROW_VTABLE_ARRAY_(dictMode, 5), \ 1429 GEN_ZSTD_ROW_VTABLE_ARRAY_(dictMode, 6) \ 1430 } 1431 1432 #define GEN_ZSTD_VTABLE_ARRAY(X) \ 1433 { \ 1434 X(noDict), \ 1435 X(extDict), \ 1436 X(dictMatchState), \ 1437 X(dedicatedDictSearch) \ 1438 } 1439 1440 /* ******************************* 1441 * Common parser - lazy strategy 1442 *********************************/ 1443 typedef enum { search_hashChain=0, search_binaryTree=1, search_rowHash=2 } searchMethod_e; 1444 1445 /** 1446 * This table is indexed first by the four ZSTD_dictMode_e values, and then 1447 * by the two searchMethod_e values. NULLs are placed for configurations 1448 * that should never occur (extDict modes go to the other implementation 1449 * below and there is no DDSS for binary tree search yet). 1450 */ 1451 1452 static ZSTD_LazyVTable const* 1453 ZSTD_selectLazyVTable(ZSTD_matchState_t const* ms, searchMethod_e searchMethod, ZSTD_dictMode_e dictMode) 1454 { 1455 /* Fill the Hc/Bt VTable arrays with the right functions for the (dictMode, mls) combination. */ 1456 ZSTD_LazyVTable const* const hcVTables[4][3] = GEN_ZSTD_VTABLE_ARRAY(GEN_ZSTD_HC_VTABLE_ARRAY); 1457 ZSTD_LazyVTable const* const btVTables[4][3] = GEN_ZSTD_VTABLE_ARRAY(GEN_ZSTD_BT_VTABLE_ARRAY); 1458 /* Fill the Row VTable array with the right functions for the (dictMode, mls, rowLog) combination. */ 1459 ZSTD_LazyVTable const* const rowVTables[4][3][3] = GEN_ZSTD_VTABLE_ARRAY(GEN_ZSTD_ROW_VTABLE_ARRAY); 1460 1461 U32 const mls = MAX(4, MIN(6, ms->cParams.minMatch)); 1462 U32 const rowLog = MAX(4, MIN(6, ms->cParams.searchLog)); 1463 switch (searchMethod) { 1464 case search_hashChain: 1465 return hcVTables[dictMode][mls - 4]; 1466 case search_binaryTree: 1467 return btVTables[dictMode][mls - 4]; 1468 case search_rowHash: 1469 return rowVTables[dictMode][mls - 4][rowLog - 4]; 1470 default: 1471 return NULL; 1472 } 1473 } 1474 1475 FORCE_INLINE_TEMPLATE size_t 1476 ZSTD_compressBlock_lazy_generic( 1477 ZSTD_matchState_t* ms, seqStore_t* seqStore, 1478 U32 rep[ZSTD_REP_NUM], 1479 const void* src, size_t srcSize, 1480 const searchMethod_e searchMethod, const U32 depth, 1481 ZSTD_dictMode_e const dictMode) 1482 { 1483 const BYTE* const istart = (const BYTE*)src; 1484 const BYTE* ip = istart; 1485 const BYTE* anchor = istart; 1486 const BYTE* const iend = istart + srcSize; 1487 const BYTE* const ilimit = (searchMethod == search_rowHash) ? iend - 8 - ZSTD_ROW_HASH_CACHE_SIZE : iend - 8; 1488 const BYTE* const base = ms->window.base; 1489 const U32 prefixLowestIndex = ms->window.dictLimit; 1490 const BYTE* const prefixLowest = base + prefixLowestIndex; 1491 1492 searchMax_f const searchMax = ZSTD_selectLazyVTable(ms, searchMethod, dictMode)->searchMax; 1493 U32 offset_1 = rep[0], offset_2 = rep[1], savedOffset=0; 1494 1495 const int isDMS = dictMode == ZSTD_dictMatchState; 1496 const int isDDS = dictMode == ZSTD_dedicatedDictSearch; 1497 const int isDxS = isDMS || isDDS; 1498 const ZSTD_matchState_t* const dms = ms->dictMatchState; 1499 const U32 dictLowestIndex = isDxS ? dms->window.dictLimit : 0; 1500 const BYTE* const dictBase = isDxS ? dms->window.base : NULL; 1501 const BYTE* const dictLowest = isDxS ? dictBase + dictLowestIndex : NULL; 1502 const BYTE* const dictEnd = isDxS ? dms->window.nextSrc : NULL; 1503 const U32 dictIndexDelta = isDxS ? 1504 prefixLowestIndex - (U32)(dictEnd - dictBase) : 1505 0; 1506 const U32 dictAndPrefixLength = (U32)((ip - prefixLowest) + (dictEnd - dictLowest)); 1507 1508 assert(searchMax != NULL); 1509 1510 DEBUGLOG(5, "ZSTD_compressBlock_lazy_generic (dictMode=%u) (searchFunc=%u)", (U32)dictMode, (U32)searchMethod); 1511 ip += (dictAndPrefixLength == 0); 1512 if (dictMode == ZSTD_noDict) { 1513 U32 const curr = (U32)(ip - base); 1514 U32 const windowLow = ZSTD_getLowestPrefixIndex(ms, curr, ms->cParams.windowLog); 1515 U32 const maxRep = curr - windowLow; 1516 if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0; 1517 if (offset_1 > maxRep) savedOffset = offset_1, offset_1 = 0; 1518 } 1519 if (isDxS) { 1520 /* dictMatchState repCode checks don't currently handle repCode == 0 1521 * disabling. */ 1522 assert(offset_1 <= dictAndPrefixLength); 1523 assert(offset_2 <= dictAndPrefixLength); 1524 } 1525 1526 if (searchMethod == search_rowHash) { 1527 const U32 rowLog = MAX(4, MIN(6, ms->cParams.searchLog)); 1528 ZSTD_row_fillHashCache(ms, base, rowLog, 1529 MIN(ms->cParams.minMatch, 6 /* mls caps out at 6 */), 1530 ms->nextToUpdate, ilimit); 1531 } 1532 1533 /* Match Loop */ 1534 #if defined(__GNUC__) && defined(__x86_64__) 1535 /* I've measured random a 5% speed loss on levels 5 & 6 (greedy) when the 1536 * code alignment is perturbed. To fix the instability align the loop on 32-bytes. 1537 */ 1538 __asm__(".p2align 5"); 1539 #endif 1540 while (ip < ilimit) { 1541 size_t matchLength=0; 1542 size_t offcode=STORE_REPCODE_1; 1543 const BYTE* start=ip+1; 1544 DEBUGLOG(7, "search baseline (depth 0)"); 1545 1546 /* check repCode */ 1547 if (isDxS) { 1548 const U32 repIndex = (U32)(ip - base) + 1 - offset_1; 1549 const BYTE* repMatch = ((dictMode == ZSTD_dictMatchState || dictMode == ZSTD_dedicatedDictSearch) 1550 && repIndex < prefixLowestIndex) ? 1551 dictBase + (repIndex - dictIndexDelta) : 1552 base + repIndex; 1553 if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */) 1554 && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) { 1555 const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend; 1556 matchLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4; 1557 if (depth==0) goto _storeSequence; 1558 } 1559 } 1560 if ( dictMode == ZSTD_noDict 1561 && ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1)))) { 1562 matchLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4; 1563 if (depth==0) goto _storeSequence; 1564 } 1565 1566 /* first search (depth 0) */ 1567 { size_t offsetFound = 999999999; 1568 size_t const ml2 = searchMax(ms, ip, iend, &offsetFound); 1569 if (ml2 > matchLength) 1570 matchLength = ml2, start = ip, offcode=offsetFound; 1571 } 1572 1573 if (matchLength < 4) { 1574 ip += ((ip-anchor) >> kSearchStrength) + 1; /* jump faster over incompressible sections */ 1575 continue; 1576 } 1577 1578 /* let's try to find a better solution */ 1579 if (depth>=1) 1580 while (ip<ilimit) { 1581 DEBUGLOG(7, "search depth 1"); 1582 ip ++; 1583 if ( (dictMode == ZSTD_noDict) 1584 && (offcode) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) { 1585 size_t const mlRep = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4; 1586 int const gain2 = (int)(mlRep * 3); 1587 int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 1); 1588 if ((mlRep >= 4) && (gain2 > gain1)) 1589 matchLength = mlRep, offcode = STORE_REPCODE_1, start = ip; 1590 } 1591 if (isDxS) { 1592 const U32 repIndex = (U32)(ip - base) - offset_1; 1593 const BYTE* repMatch = repIndex < prefixLowestIndex ? 1594 dictBase + (repIndex - dictIndexDelta) : 1595 base + repIndex; 1596 if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */) 1597 && (MEM_read32(repMatch) == MEM_read32(ip)) ) { 1598 const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend; 1599 size_t const mlRep = ZSTD_count_2segments(ip+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4; 1600 int const gain2 = (int)(mlRep * 3); 1601 int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 1); 1602 if ((mlRep >= 4) && (gain2 > gain1)) 1603 matchLength = mlRep, offcode = STORE_REPCODE_1, start = ip; 1604 } 1605 } 1606 { size_t offset2=999999999; 1607 size_t const ml2 = searchMax(ms, ip, iend, &offset2); 1608 int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offset2))); /* raw approx */ 1609 int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 4); 1610 if ((ml2 >= 4) && (gain2 > gain1)) { 1611 matchLength = ml2, offcode = offset2, start = ip; 1612 continue; /* search a better one */ 1613 } } 1614 1615 /* let's find an even better one */ 1616 if ((depth==2) && (ip<ilimit)) { 1617 DEBUGLOG(7, "search depth 2"); 1618 ip ++; 1619 if ( (dictMode == ZSTD_noDict) 1620 && (offcode) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) { 1621 size_t const mlRep = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4; 1622 int const gain2 = (int)(mlRep * 4); 1623 int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 1); 1624 if ((mlRep >= 4) && (gain2 > gain1)) 1625 matchLength = mlRep, offcode = STORE_REPCODE_1, start = ip; 1626 } 1627 if (isDxS) { 1628 const U32 repIndex = (U32)(ip - base) - offset_1; 1629 const BYTE* repMatch = repIndex < prefixLowestIndex ? 1630 dictBase + (repIndex - dictIndexDelta) : 1631 base + repIndex; 1632 if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */) 1633 && (MEM_read32(repMatch) == MEM_read32(ip)) ) { 1634 const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend; 1635 size_t const mlRep = ZSTD_count_2segments(ip+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4; 1636 int const gain2 = (int)(mlRep * 4); 1637 int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 1); 1638 if ((mlRep >= 4) && (gain2 > gain1)) 1639 matchLength = mlRep, offcode = STORE_REPCODE_1, start = ip; 1640 } 1641 } 1642 { size_t offset2=999999999; 1643 size_t const ml2 = searchMax(ms, ip, iend, &offset2); 1644 int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offset2))); /* raw approx */ 1645 int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 7); 1646 if ((ml2 >= 4) && (gain2 > gain1)) { 1647 matchLength = ml2, offcode = offset2, start = ip; 1648 continue; 1649 } } } 1650 break; /* nothing found : store previous solution */ 1651 } 1652 1653 /* NOTE: 1654 * Pay attention that `start[-value]` can lead to strange undefined behavior 1655 * notably if `value` is unsigned, resulting in a large positive `-value`. 1656 */ 1657 /* catch up */ 1658 if (STORED_IS_OFFSET(offcode)) { 1659 if (dictMode == ZSTD_noDict) { 1660 while ( ((start > anchor) & (start - STORED_OFFSET(offcode) > prefixLowest)) 1661 && (start[-1] == (start-STORED_OFFSET(offcode))[-1]) ) /* only search for offset within prefix */ 1662 { start--; matchLength++; } 1663 } 1664 if (isDxS) { 1665 U32 const matchIndex = (U32)((size_t)(start-base) - STORED_OFFSET(offcode)); 1666 const BYTE* match = (matchIndex < prefixLowestIndex) ? dictBase + matchIndex - dictIndexDelta : base + matchIndex; 1667 const BYTE* const mStart = (matchIndex < prefixLowestIndex) ? dictLowest : prefixLowest; 1668 while ((start>anchor) && (match>mStart) && (start[-1] == match[-1])) { start--; match--; matchLength++; } /* catch up */ 1669 } 1670 offset_2 = offset_1; offset_1 = (U32)STORED_OFFSET(offcode); 1671 } 1672 /* store sequence */ 1673 _storeSequence: 1674 { size_t const litLength = (size_t)(start - anchor); 1675 ZSTD_storeSeq(seqStore, litLength, anchor, iend, (U32)offcode, matchLength); 1676 anchor = ip = start + matchLength; 1677 } 1678 1679 /* check immediate repcode */ 1680 if (isDxS) { 1681 while (ip <= ilimit) { 1682 U32 const current2 = (U32)(ip-base); 1683 U32 const repIndex = current2 - offset_2; 1684 const BYTE* repMatch = repIndex < prefixLowestIndex ? 1685 dictBase - dictIndexDelta + repIndex : 1686 base + repIndex; 1687 if ( ((U32)((prefixLowestIndex-1) - (U32)repIndex) >= 3 /* intentional overflow */) 1688 && (MEM_read32(repMatch) == MEM_read32(ip)) ) { 1689 const BYTE* const repEnd2 = repIndex < prefixLowestIndex ? dictEnd : iend; 1690 matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd2, prefixLowest) + 4; 1691 offcode = offset_2; offset_2 = offset_1; offset_1 = (U32)offcode; /* swap offset_2 <=> offset_1 */ 1692 ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, matchLength); 1693 ip += matchLength; 1694 anchor = ip; 1695 continue; 1696 } 1697 break; 1698 } 1699 } 1700 1701 if (dictMode == ZSTD_noDict) { 1702 while ( ((ip <= ilimit) & (offset_2>0)) 1703 && (MEM_read32(ip) == MEM_read32(ip - offset_2)) ) { 1704 /* store sequence */ 1705 matchLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4; 1706 offcode = offset_2; offset_2 = offset_1; offset_1 = (U32)offcode; /* swap repcodes */ 1707 ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, matchLength); 1708 ip += matchLength; 1709 anchor = ip; 1710 continue; /* faster when present ... (?) */ 1711 } } } 1712 1713 /* Save reps for next block */ 1714 rep[0] = offset_1 ? offset_1 : savedOffset; 1715 rep[1] = offset_2 ? offset_2 : savedOffset; 1716 1717 /* Return the last literals size */ 1718 return (size_t)(iend - anchor); 1719 } 1720 1721 1722 size_t ZSTD_compressBlock_btlazy2( 1723 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1724 void const* src, size_t srcSize) 1725 { 1726 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict); 1727 } 1728 1729 size_t ZSTD_compressBlock_lazy2( 1730 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1731 void const* src, size_t srcSize) 1732 { 1733 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict); 1734 } 1735 1736 size_t ZSTD_compressBlock_lazy( 1737 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1738 void const* src, size_t srcSize) 1739 { 1740 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict); 1741 } 1742 1743 size_t ZSTD_compressBlock_greedy( 1744 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1745 void const* src, size_t srcSize) 1746 { 1747 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_noDict); 1748 } 1749 1750 size_t ZSTD_compressBlock_btlazy2_dictMatchState( 1751 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1752 void const* src, size_t srcSize) 1753 { 1754 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState); 1755 } 1756 1757 size_t ZSTD_compressBlock_lazy2_dictMatchState( 1758 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1759 void const* src, size_t srcSize) 1760 { 1761 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState); 1762 } 1763 1764 size_t ZSTD_compressBlock_lazy_dictMatchState( 1765 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1766 void const* src, size_t srcSize) 1767 { 1768 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState); 1769 } 1770 1771 size_t ZSTD_compressBlock_greedy_dictMatchState( 1772 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1773 void const* src, size_t srcSize) 1774 { 1775 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dictMatchState); 1776 } 1777 1778 1779 size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch( 1780 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1781 void const* src, size_t srcSize) 1782 { 1783 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dedicatedDictSearch); 1784 } 1785 1786 size_t ZSTD_compressBlock_lazy_dedicatedDictSearch( 1787 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1788 void const* src, size_t srcSize) 1789 { 1790 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dedicatedDictSearch); 1791 } 1792 1793 size_t ZSTD_compressBlock_greedy_dedicatedDictSearch( 1794 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1795 void const* src, size_t srcSize) 1796 { 1797 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dedicatedDictSearch); 1798 } 1799 1800 /* Row-based matchfinder */ 1801 size_t ZSTD_compressBlock_lazy2_row( 1802 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1803 void const* src, size_t srcSize) 1804 { 1805 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_noDict); 1806 } 1807 1808 size_t ZSTD_compressBlock_lazy_row( 1809 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1810 void const* src, size_t srcSize) 1811 { 1812 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_noDict); 1813 } 1814 1815 size_t ZSTD_compressBlock_greedy_row( 1816 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1817 void const* src, size_t srcSize) 1818 { 1819 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_noDict); 1820 } 1821 1822 size_t ZSTD_compressBlock_lazy2_dictMatchState_row( 1823 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1824 void const* src, size_t srcSize) 1825 { 1826 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dictMatchState); 1827 } 1828 1829 size_t ZSTD_compressBlock_lazy_dictMatchState_row( 1830 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1831 void const* src, size_t srcSize) 1832 { 1833 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dictMatchState); 1834 } 1835 1836 size_t ZSTD_compressBlock_greedy_dictMatchState_row( 1837 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1838 void const* src, size_t srcSize) 1839 { 1840 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dictMatchState); 1841 } 1842 1843 1844 size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row( 1845 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1846 void const* src, size_t srcSize) 1847 { 1848 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dedicatedDictSearch); 1849 } 1850 1851 size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row( 1852 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1853 void const* src, size_t srcSize) 1854 { 1855 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dedicatedDictSearch); 1856 } 1857 1858 size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row( 1859 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 1860 void const* src, size_t srcSize) 1861 { 1862 return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dedicatedDictSearch); 1863 } 1864 1865 FORCE_INLINE_TEMPLATE 1866 size_t ZSTD_compressBlock_lazy_extDict_generic( 1867 ZSTD_matchState_t* ms, seqStore_t* seqStore, 1868 U32 rep[ZSTD_REP_NUM], 1869 const void* src, size_t srcSize, 1870 const searchMethod_e searchMethod, const U32 depth) 1871 { 1872 const BYTE* const istart = (const BYTE*)src; 1873 const BYTE* ip = istart; 1874 const BYTE* anchor = istart; 1875 const BYTE* const iend = istart + srcSize; 1876 const BYTE* const ilimit = searchMethod == search_rowHash ? iend - 8 - ZSTD_ROW_HASH_CACHE_SIZE : iend - 8; 1877 const BYTE* const base = ms->window.base; 1878 const U32 dictLimit = ms->window.dictLimit; 1879 const BYTE* const prefixStart = base + dictLimit; 1880 const BYTE* const dictBase = ms->window.dictBase; 1881 const BYTE* const dictEnd = dictBase + dictLimit; 1882 const BYTE* const dictStart = dictBase + ms->window.lowLimit; 1883 const U32 windowLog = ms->cParams.windowLog; 1884 const U32 rowLog = ms->cParams.searchLog < 5 ? 4 : 5; 1885 1886 searchMax_f const searchMax = ZSTD_selectLazyVTable(ms, searchMethod, ZSTD_extDict)->searchMax; 1887 U32 offset_1 = rep[0], offset_2 = rep[1]; 1888 1889 DEBUGLOG(5, "ZSTD_compressBlock_lazy_extDict_generic (searchFunc=%u)", (U32)searchMethod); 1890 1891 /* init */ 1892 ip += (ip == prefixStart); 1893 if (searchMethod == search_rowHash) { 1894 ZSTD_row_fillHashCache(ms, base, rowLog, 1895 MIN(ms->cParams.minMatch, 6 /* mls caps out at 6 */), 1896 ms->nextToUpdate, ilimit); 1897 } 1898 1899 /* Match Loop */ 1900 #if defined(__GNUC__) && defined(__x86_64__) 1901 /* I've measured random a 5% speed loss on levels 5 & 6 (greedy) when the 1902 * code alignment is perturbed. To fix the instability align the loop on 32-bytes. 1903 */ 1904 __asm__(".p2align 5"); 1905 #endif 1906 while (ip < ilimit) { 1907 size_t matchLength=0; 1908 size_t offcode=STORE_REPCODE_1; 1909 const BYTE* start=ip+1; 1910 U32 curr = (U32)(ip-base); 1911 1912 /* check repCode */ 1913 { const U32 windowLow = ZSTD_getLowestMatchIndex(ms, curr+1, windowLog); 1914 const U32 repIndex = (U32)(curr+1 - offset_1); 1915 const BYTE* const repBase = repIndex < dictLimit ? dictBase : base; 1916 const BYTE* const repMatch = repBase + repIndex; 1917 if ( ((U32)((dictLimit-1) - repIndex) >= 3) /* intentional overflow */ 1918 & (offset_1 <= curr+1 - windowLow) ) /* note: we are searching at curr+1 */ 1919 if (MEM_read32(ip+1) == MEM_read32(repMatch)) { 1920 /* repcode detected we should take it */ 1921 const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend; 1922 matchLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repEnd, prefixStart) + 4; 1923 if (depth==0) goto _storeSequence; 1924 } } 1925 1926 /* first search (depth 0) */ 1927 { size_t offsetFound = 999999999; 1928 size_t const ml2 = searchMax(ms, ip, iend, &offsetFound); 1929 if (ml2 > matchLength) 1930 matchLength = ml2, start = ip, offcode=offsetFound; 1931 } 1932 1933 if (matchLength < 4) { 1934 ip += ((ip-anchor) >> kSearchStrength) + 1; /* jump faster over incompressible sections */ 1935 continue; 1936 } 1937 1938 /* let's try to find a better solution */ 1939 if (depth>=1) 1940 while (ip<ilimit) { 1941 ip ++; 1942 curr++; 1943 /* check repCode */ 1944 if (offcode) { 1945 const U32 windowLow = ZSTD_getLowestMatchIndex(ms, curr, windowLog); 1946 const U32 repIndex = (U32)(curr - offset_1); 1947 const BYTE* const repBase = repIndex < dictLimit ? dictBase : base; 1948 const BYTE* const repMatch = repBase + repIndex; 1949 if ( ((U32)((dictLimit-1) - repIndex) >= 3) /* intentional overflow : do not test positions overlapping 2 memory segments */ 1950 & (offset_1 <= curr - windowLow) ) /* equivalent to `curr > repIndex >= windowLow` */ 1951 if (MEM_read32(ip) == MEM_read32(repMatch)) { 1952 /* repcode detected */ 1953 const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend; 1954 size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4; 1955 int const gain2 = (int)(repLength * 3); 1956 int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 1); 1957 if ((repLength >= 4) && (gain2 > gain1)) 1958 matchLength = repLength, offcode = STORE_REPCODE_1, start = ip; 1959 } } 1960 1961 /* search match, depth 1 */ 1962 { size_t offset2=999999999; 1963 size_t const ml2 = searchMax(ms, ip, iend, &offset2); 1964 int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offset2))); /* raw approx */ 1965 int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 4); 1966 if ((ml2 >= 4) && (gain2 > gain1)) { 1967 matchLength = ml2, offcode = offset2, start = ip; 1968 continue; /* search a better one */ 1969 } } 1970 1971 /* let's find an even better one */ 1972 if ((depth==2) && (ip<ilimit)) { 1973 ip ++; 1974 curr++; 1975 /* check repCode */ 1976 if (offcode) { 1977 const U32 windowLow = ZSTD_getLowestMatchIndex(ms, curr, windowLog); 1978 const U32 repIndex = (U32)(curr - offset_1); 1979 const BYTE* const repBase = repIndex < dictLimit ? dictBase : base; 1980 const BYTE* const repMatch = repBase + repIndex; 1981 if ( ((U32)((dictLimit-1) - repIndex) >= 3) /* intentional overflow : do not test positions overlapping 2 memory segments */ 1982 & (offset_1 <= curr - windowLow) ) /* equivalent to `curr > repIndex >= windowLow` */ 1983 if (MEM_read32(ip) == MEM_read32(repMatch)) { 1984 /* repcode detected */ 1985 const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend; 1986 size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4; 1987 int const gain2 = (int)(repLength * 4); 1988 int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 1); 1989 if ((repLength >= 4) && (gain2 > gain1)) 1990 matchLength = repLength, offcode = STORE_REPCODE_1, start = ip; 1991 } } 1992 1993 /* search match, depth 2 */ 1994 { size_t offset2=999999999; 1995 size_t const ml2 = searchMax(ms, ip, iend, &offset2); 1996 int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offset2))); /* raw approx */ 1997 int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)STORED_TO_OFFBASE(offcode)) + 7); 1998 if ((ml2 >= 4) && (gain2 > gain1)) { 1999 matchLength = ml2, offcode = offset2, start = ip; 2000 continue; 2001 } } } 2002 break; /* nothing found : store previous solution */ 2003 } 2004 2005 /* catch up */ 2006 if (STORED_IS_OFFSET(offcode)) { 2007 U32 const matchIndex = (U32)((size_t)(start-base) - STORED_OFFSET(offcode)); 2008 const BYTE* match = (matchIndex < dictLimit) ? dictBase + matchIndex : base + matchIndex; 2009 const BYTE* const mStart = (matchIndex < dictLimit) ? dictStart : prefixStart; 2010 while ((start>anchor) && (match>mStart) && (start[-1] == match[-1])) { start--; match--; matchLength++; } /* catch up */ 2011 offset_2 = offset_1; offset_1 = (U32)STORED_OFFSET(offcode); 2012 } 2013 2014 /* store sequence */ 2015 _storeSequence: 2016 { size_t const litLength = (size_t)(start - anchor); 2017 ZSTD_storeSeq(seqStore, litLength, anchor, iend, (U32)offcode, matchLength); 2018 anchor = ip = start + matchLength; 2019 } 2020 2021 /* check immediate repcode */ 2022 while (ip <= ilimit) { 2023 const U32 repCurrent = (U32)(ip-base); 2024 const U32 windowLow = ZSTD_getLowestMatchIndex(ms, repCurrent, windowLog); 2025 const U32 repIndex = repCurrent - offset_2; 2026 const BYTE* const repBase = repIndex < dictLimit ? dictBase : base; 2027 const BYTE* const repMatch = repBase + repIndex; 2028 if ( ((U32)((dictLimit-1) - repIndex) >= 3) /* intentional overflow : do not test positions overlapping 2 memory segments */ 2029 & (offset_2 <= repCurrent - windowLow) ) /* equivalent to `curr > repIndex >= windowLow` */ 2030 if (MEM_read32(ip) == MEM_read32(repMatch)) { 2031 /* repcode detected we should take it */ 2032 const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend; 2033 matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4; 2034 offcode = offset_2; offset_2 = offset_1; offset_1 = (U32)offcode; /* swap offset history */ 2035 ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, matchLength); 2036 ip += matchLength; 2037 anchor = ip; 2038 continue; /* faster when present ... (?) */ 2039 } 2040 break; 2041 } } 2042 2043 /* Save reps for next block */ 2044 rep[0] = offset_1; 2045 rep[1] = offset_2; 2046 2047 /* Return the last literals size */ 2048 return (size_t)(iend - anchor); 2049 } 2050 2051 2052 size_t ZSTD_compressBlock_greedy_extDict( 2053 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 2054 void const* src, size_t srcSize) 2055 { 2056 return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0); 2057 } 2058 2059 size_t ZSTD_compressBlock_lazy_extDict( 2060 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 2061 void const* src, size_t srcSize) 2062 2063 { 2064 return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1); 2065 } 2066 2067 size_t ZSTD_compressBlock_lazy2_extDict( 2068 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 2069 void const* src, size_t srcSize) 2070 2071 { 2072 return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2); 2073 } 2074 2075 size_t ZSTD_compressBlock_btlazy2_extDict( 2076 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 2077 void const* src, size_t srcSize) 2078 2079 { 2080 return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2); 2081 } 2082 2083 size_t ZSTD_compressBlock_greedy_extDict_row( 2084 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 2085 void const* src, size_t srcSize) 2086 { 2087 return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0); 2088 } 2089 2090 size_t ZSTD_compressBlock_lazy_extDict_row( 2091 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 2092 void const* src, size_t srcSize) 2093 2094 { 2095 return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1); 2096 } 2097 2098 size_t ZSTD_compressBlock_lazy2_extDict_row( 2099 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], 2100 void const* src, size_t srcSize) 2101 2102 { 2103 return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2); 2104 } 2105