Loading...
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 | /* * Copyright (c) Yann Collet, Facebook, Inc. * All rights reserved. * * This source code is licensed under both the BSD-style license (found in the * LICENSE file in the root directory of this source tree) and the GPLv2 (found * in the COPYING file in the root directory of this source tree). * You may select, at your option, one of the above-listed licenses. */ /*-************************************* * Dependencies ***************************************/ #include "zstd_compress_superblock.h" #include "../common/zstd_internal.h" /* ZSTD_getSequenceLength */ #include "hist.h" /* HIST_countFast_wksp */ #include "zstd_compress_internal.h" /* ZSTD_[huf|fse|entropy]CTablesMetadata_t */ #include "zstd_compress_sequences.h" #include "zstd_compress_literals.h" /* ZSTD_compressSubBlock_literal() : * Compresses literals section for a sub-block. * When we have to write the Huffman table we will sometimes choose a header * size larger than necessary. This is because we have to pick the header size * before we know the table size + compressed size, so we have a bound on the * table size. If we guessed incorrectly, we fall back to uncompressed literals. * * We write the header when writeEntropy=1 and set entropyWritten=1 when we succeeded * in writing the header, otherwise it is set to 0. * * hufMetadata->hType has literals block type info. * If it is set_basic, all sub-blocks literals section will be Raw_Literals_Block. * If it is set_rle, all sub-blocks literals section will be RLE_Literals_Block. * If it is set_compressed, first sub-block's literals section will be Compressed_Literals_Block * If it is set_compressed, first sub-block's literals section will be Treeless_Literals_Block * and the following sub-blocks' literals sections will be Treeless_Literals_Block. * @return : compressed size of literals section of a sub-block * Or 0 if it unable to compress. * Or error code */ static size_t ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable, const ZSTD_hufCTablesMetadata_t* hufMetadata, const BYTE* literals, size_t litSize, void* dst, size_t dstSize, const int bmi2, int writeEntropy, int* entropyWritten) { size_t const header = writeEntropy ? 200 : 0; size_t const lhSize = 3 + (litSize >= (1 KB - header)) + (litSize >= (16 KB - header)); BYTE* const ostart = (BYTE*)dst; BYTE* const oend = ostart + dstSize; BYTE* op = ostart + lhSize; U32 const singleStream = lhSize == 3; symbolEncodingType_e hType = writeEntropy ? hufMetadata->hType : set_repeat; size_t cLitSize = 0; (void)bmi2; /* TODO bmi2... */ DEBUGLOG(5, "ZSTD_compressSubBlock_literal (litSize=%zu, lhSize=%zu, writeEntropy=%d)", litSize, lhSize, writeEntropy); *entropyWritten = 0; if (litSize == 0 || hufMetadata->hType == set_basic) { DEBUGLOG(5, "ZSTD_compressSubBlock_literal using raw literal"); return ZSTD_noCompressLiterals(dst, dstSize, literals, litSize); } else if (hufMetadata->hType == set_rle) { DEBUGLOG(5, "ZSTD_compressSubBlock_literal using rle literal"); return ZSTD_compressRleLiteralsBlock(dst, dstSize, literals, litSize); } assert(litSize > 0); assert(hufMetadata->hType == set_compressed || hufMetadata->hType == set_repeat); if (writeEntropy && hufMetadata->hType == set_compressed) { ZSTD_memcpy(op, hufMetadata->hufDesBuffer, hufMetadata->hufDesSize); op += hufMetadata->hufDesSize; cLitSize += hufMetadata->hufDesSize; DEBUGLOG(5, "ZSTD_compressSubBlock_literal (hSize=%zu)", hufMetadata->hufDesSize); } /* TODO bmi2 */ { const size_t cSize = singleStream ? HUF_compress1X_usingCTable(op, oend-op, literals, litSize, hufTable) : HUF_compress4X_usingCTable(op, oend-op, literals, litSize, hufTable); op += cSize; cLitSize += cSize; if (cSize == 0 || ERR_isError(cSize)) { DEBUGLOG(5, "Failed to write entropy tables %s", ZSTD_getErrorName(cSize)); return 0; } /* If we expand and we aren't writing a header then emit uncompressed */ if (!writeEntropy && cLitSize >= litSize) { DEBUGLOG(5, "ZSTD_compressSubBlock_literal using raw literal because uncompressible"); return ZSTD_noCompressLiterals(dst, dstSize, literals, litSize); } /* If we are writing headers then allow expansion that doesn't change our header size. */ if (lhSize < (size_t)(3 + (cLitSize >= 1 KB) + (cLitSize >= 16 KB))) { assert(cLitSize > litSize); DEBUGLOG(5, "Literals expanded beyond allowed header size"); return ZSTD_noCompressLiterals(dst, dstSize, literals, litSize); } DEBUGLOG(5, "ZSTD_compressSubBlock_literal (cSize=%zu)", cSize); } /* Build header */ switch(lhSize) { case 3: /* 2 - 2 - 10 - 10 */ { U32 const lhc = hType + ((!singleStream) << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<14); MEM_writeLE24(ostart, lhc); break; } case 4: /* 2 - 2 - 14 - 14 */ { U32 const lhc = hType + (2 << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<18); MEM_writeLE32(ostart, lhc); break; } case 5: /* 2 - 2 - 18 - 18 */ { U32 const lhc = hType + (3 << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<22); MEM_writeLE32(ostart, lhc); ostart[4] = (BYTE)(cLitSize >> 10); break; } default: /* not possible : lhSize is {3,4,5} */ assert(0); } *entropyWritten = 1; DEBUGLOG(5, "Compressed literals: %u -> %u", (U32)litSize, (U32)(op-ostart)); return op-ostart; } static size_t ZSTD_seqDecompressedSize(seqStore_t const* seqStore, const seqDef* sequences, size_t nbSeq, size_t litSize, int lastSequence) { const seqDef* const sstart = sequences; const seqDef* const send = sequences + nbSeq; const seqDef* sp = sstart; size_t matchLengthSum = 0; size_t litLengthSum = 0; (void)(litLengthSum); /* suppress unused variable warning on some environments */ while (send-sp > 0) { ZSTD_sequenceLength const seqLen = ZSTD_getSequenceLength(seqStore, sp); litLengthSum += seqLen.litLength; matchLengthSum += seqLen.matchLength; sp++; } assert(litLengthSum <= litSize); if (!lastSequence) { assert(litLengthSum == litSize); } return matchLengthSum + litSize; } /* ZSTD_compressSubBlock_sequences() : * Compresses sequences section for a sub-block. * fseMetadata->llType, fseMetadata->ofType, and fseMetadata->mlType have * symbol compression modes for the super-block. * The first successfully compressed block will have these in its header. * We set entropyWritten=1 when we succeed in compressing the sequences. * The following sub-blocks will always have repeat mode. * @return : compressed size of sequences section of a sub-block * Or 0 if it is unable to compress * Or error code. */ static size_t ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables, const ZSTD_fseCTablesMetadata_t* fseMetadata, const seqDef* sequences, size_t nbSeq, const BYTE* llCode, const BYTE* mlCode, const BYTE* ofCode, const ZSTD_CCtx_params* cctxParams, void* dst, size_t dstCapacity, const int bmi2, int writeEntropy, int* entropyWritten) { const int longOffsets = cctxParams->cParams.windowLog > STREAM_ACCUMULATOR_MIN; BYTE* const ostart = (BYTE*)dst; BYTE* const oend = ostart + dstCapacity; BYTE* op = ostart; BYTE* seqHead; DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (nbSeq=%zu, writeEntropy=%d, longOffsets=%d)", nbSeq, writeEntropy, longOffsets); *entropyWritten = 0; /* Sequences Header */ RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/, dstSize_tooSmall, ""); if (nbSeq < 0x7F) *op++ = (BYTE)nbSeq; else if (nbSeq < LONGNBSEQ) op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2; else op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3; if (nbSeq==0) { return op - ostart; } /* seqHead : flags for FSE encoding type */ seqHead = op++; DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (seqHeadSize=%u)", (unsigned)(op-ostart)); if (writeEntropy) { const U32 LLtype = fseMetadata->llType; const U32 Offtype = fseMetadata->ofType; const U32 MLtype = fseMetadata->mlType; DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (fseTablesSize=%zu)", fseMetadata->fseTablesSize); *seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2)); ZSTD_memcpy(op, fseMetadata->fseTablesBuffer, fseMetadata->fseTablesSize); op += fseMetadata->fseTablesSize; } else { const U32 repeat = set_repeat; *seqHead = (BYTE)((repeat<<6) + (repeat<<4) + (repeat<<2)); } { size_t const bitstreamSize = ZSTD_encodeSequences( op, oend - op, fseTables->matchlengthCTable, mlCode, fseTables->offcodeCTable, ofCode, fseTables->litlengthCTable, llCode, sequences, nbSeq, longOffsets, bmi2); FORWARD_IF_ERROR(bitstreamSize, "ZSTD_encodeSequences failed"); op += bitstreamSize; /* zstd versions <= 1.3.4 mistakenly report corruption when * FSE_readNCount() receives a buffer < 4 bytes. * Fixed by https://github.com/facebook/zstd/pull/1146. * This can happen when the last set_compressed table present is 2 * bytes and the bitstream is only one byte. * In this exceedingly rare case, we will simply emit an uncompressed * block, since it isn't worth optimizing. */ #ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION if (writeEntropy && fseMetadata->lastCountSize && fseMetadata->lastCountSize + bitstreamSize < 4) { /* NCountSize >= 2 && bitstreamSize > 0 ==> lastCountSize == 3 */ assert(fseMetadata->lastCountSize + bitstreamSize == 3); DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.3.4 by " "emitting an uncompressed block."); return 0; } #endif DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (bitstreamSize=%zu)", bitstreamSize); } /* zstd versions <= 1.4.0 mistakenly report error when * sequences section body size is less than 3 bytes. * Fixed by https://github.com/facebook/zstd/pull/1664. * This can happen when the previous sequences section block is compressed * with rle mode and the current block's sequences section is compressed * with repeat mode where sequences section body size can be 1 byte. */ #ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION if (op-seqHead < 4) { DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.4.0 by emitting " "an uncompressed block when sequences are < 4 bytes"); return 0; } #endif *entropyWritten = 1; return op - ostart; } /* ZSTD_compressSubBlock() : * Compresses a single sub-block. * @return : compressed size of the sub-block * Or 0 if it failed to compress. */ static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy, const ZSTD_entropyCTablesMetadata_t* entropyMetadata, const seqDef* sequences, size_t nbSeq, const BYTE* literals, size_t litSize, const BYTE* llCode, const BYTE* mlCode, const BYTE* ofCode, const ZSTD_CCtx_params* cctxParams, void* dst, size_t dstCapacity, const int bmi2, int writeLitEntropy, int writeSeqEntropy, int* litEntropyWritten, int* seqEntropyWritten, U32 lastBlock) { BYTE* const ostart = (BYTE*)dst; BYTE* const oend = ostart + dstCapacity; BYTE* op = ostart + ZSTD_blockHeaderSize; DEBUGLOG(5, "ZSTD_compressSubBlock (litSize=%zu, nbSeq=%zu, writeLitEntropy=%d, writeSeqEntropy=%d, lastBlock=%d)", litSize, nbSeq, writeLitEntropy, writeSeqEntropy, lastBlock); { size_t cLitSize = ZSTD_compressSubBlock_literal((const HUF_CElt*)entropy->huf.CTable, &entropyMetadata->hufMetadata, literals, litSize, op, oend-op, bmi2, writeLitEntropy, litEntropyWritten); FORWARD_IF_ERROR(cLitSize, "ZSTD_compressSubBlock_literal failed"); if (cLitSize == 0) return 0; op += cLitSize; } { size_t cSeqSize = ZSTD_compressSubBlock_sequences(&entropy->fse, &entropyMetadata->fseMetadata, sequences, nbSeq, llCode, mlCode, ofCode, cctxParams, op, oend-op, bmi2, writeSeqEntropy, seqEntropyWritten); FORWARD_IF_ERROR(cSeqSize, "ZSTD_compressSubBlock_sequences failed"); if (cSeqSize == 0) return 0; op += cSeqSize; } /* Write block header */ { size_t cSize = (op-ostart)-ZSTD_blockHeaderSize; U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3); MEM_writeLE24(ostart, cBlockHeader24); } return op-ostart; } static size_t ZSTD_estimateSubBlockSize_literal(const BYTE* literals, size_t litSize, const ZSTD_hufCTables_t* huf, const ZSTD_hufCTablesMetadata_t* hufMetadata, void* workspace, size_t wkspSize, int writeEntropy) { unsigned* const countWksp = (unsigned*)workspace; unsigned maxSymbolValue = 255; size_t literalSectionHeaderSize = 3; /* Use hard coded size of 3 bytes */ if (hufMetadata->hType == set_basic) return litSize; else if (hufMetadata->hType == set_rle) return 1; else if (hufMetadata->hType == set_compressed || hufMetadata->hType == set_repeat) { size_t const largest = HIST_count_wksp (countWksp, &maxSymbolValue, (const BYTE*)literals, litSize, workspace, wkspSize); if (ZSTD_isError(largest)) return litSize; { size_t cLitSizeEstimate = HUF_estimateCompressedSize((const HUF_CElt*)huf->CTable, countWksp, maxSymbolValue); if (writeEntropy) cLitSizeEstimate += hufMetadata->hufDesSize; return cLitSizeEstimate + literalSectionHeaderSize; } } assert(0); /* impossible */ return 0; } static size_t ZSTD_estimateSubBlockSize_symbolType(symbolEncodingType_e type, const BYTE* codeTable, unsigned maxCode, size_t nbSeq, const FSE_CTable* fseCTable, const U8* additionalBits, short const* defaultNorm, U32 defaultNormLog, U32 defaultMax, void* workspace, size_t wkspSize) { unsigned* const countWksp = (unsigned*)workspace; const BYTE* ctp = codeTable; const BYTE* const ctStart = ctp; const BYTE* const ctEnd = ctStart + nbSeq; size_t cSymbolTypeSizeEstimateInBits = 0; unsigned max = maxCode; HIST_countFast_wksp(countWksp, &max, codeTable, nbSeq, workspace, wkspSize); /* can't fail */ if (type == set_basic) { /* We selected this encoding type, so it must be valid. */ assert(max <= defaultMax); cSymbolTypeSizeEstimateInBits = max <= defaultMax ? ZSTD_crossEntropyCost(defaultNorm, defaultNormLog, countWksp, max) : ERROR(GENERIC); } else if (type == set_rle) { cSymbolTypeSizeEstimateInBits = 0; } else if (type == set_compressed || type == set_repeat) { cSymbolTypeSizeEstimateInBits = ZSTD_fseBitCost(fseCTable, countWksp, max); } if (ZSTD_isError(cSymbolTypeSizeEstimateInBits)) return nbSeq * 10; while (ctp < ctEnd) { if (additionalBits) cSymbolTypeSizeEstimateInBits += additionalBits[*ctp]; else cSymbolTypeSizeEstimateInBits += *ctp; /* for offset, offset code is also the number of additional bits */ ctp++; } return cSymbolTypeSizeEstimateInBits / 8; } static size_t ZSTD_estimateSubBlockSize_sequences(const BYTE* ofCodeTable, const BYTE* llCodeTable, const BYTE* mlCodeTable, size_t nbSeq, const ZSTD_fseCTables_t* fseTables, const ZSTD_fseCTablesMetadata_t* fseMetadata, void* workspace, size_t wkspSize, int writeEntropy) { size_t const sequencesSectionHeaderSize = 3; /* Use hard coded size of 3 bytes */ size_t cSeqSizeEstimate = 0; if (nbSeq == 0) return sequencesSectionHeaderSize; cSeqSizeEstimate += ZSTD_estimateSubBlockSize_symbolType(fseMetadata->ofType, ofCodeTable, MaxOff, nbSeq, fseTables->offcodeCTable, NULL, OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff, workspace, wkspSize); cSeqSizeEstimate += ZSTD_estimateSubBlockSize_symbolType(fseMetadata->llType, llCodeTable, MaxLL, nbSeq, fseTables->litlengthCTable, LL_bits, LL_defaultNorm, LL_defaultNormLog, MaxLL, workspace, wkspSize); cSeqSizeEstimate += ZSTD_estimateSubBlockSize_symbolType(fseMetadata->mlType, mlCodeTable, MaxML, nbSeq, fseTables->matchlengthCTable, ML_bits, ML_defaultNorm, ML_defaultNormLog, MaxML, workspace, wkspSize); if (writeEntropy) cSeqSizeEstimate += fseMetadata->fseTablesSize; return cSeqSizeEstimate + sequencesSectionHeaderSize; } static size_t ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize, const BYTE* ofCodeTable, const BYTE* llCodeTable, const BYTE* mlCodeTable, size_t nbSeq, const ZSTD_entropyCTables_t* entropy, const ZSTD_entropyCTablesMetadata_t* entropyMetadata, void* workspace, size_t wkspSize, int writeLitEntropy, int writeSeqEntropy) { size_t cSizeEstimate = 0; cSizeEstimate += ZSTD_estimateSubBlockSize_literal(literals, litSize, &entropy->huf, &entropyMetadata->hufMetadata, workspace, wkspSize, writeLitEntropy); cSizeEstimate += ZSTD_estimateSubBlockSize_sequences(ofCodeTable, llCodeTable, mlCodeTable, nbSeq, &entropy->fse, &entropyMetadata->fseMetadata, workspace, wkspSize, writeSeqEntropy); return cSizeEstimate + ZSTD_blockHeaderSize; } static int ZSTD_needSequenceEntropyTables(ZSTD_fseCTablesMetadata_t const* fseMetadata) { if (fseMetadata->llType == set_compressed || fseMetadata->llType == set_rle) return 1; if (fseMetadata->mlType == set_compressed || fseMetadata->mlType == set_rle) return 1; if (fseMetadata->ofType == set_compressed || fseMetadata->ofType == set_rle) return 1; return 0; } /* ZSTD_compressSubBlock_multi() : * Breaks super-block into multiple sub-blocks and compresses them. * Entropy will be written to the first block. * The following blocks will use repeat mode to compress. * All sub-blocks are compressed blocks (no raw or rle blocks). * @return : compressed size of the super block (which is multiple ZSTD blocks) * Or 0 if it failed to compress. */ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr, const ZSTD_compressedBlockState_t* prevCBlock, ZSTD_compressedBlockState_t* nextCBlock, const ZSTD_entropyCTablesMetadata_t* entropyMetadata, const ZSTD_CCtx_params* cctxParams, void* dst, size_t dstCapacity, const void* src, size_t srcSize, const int bmi2, U32 lastBlock, void* workspace, size_t wkspSize) { const seqDef* const sstart = seqStorePtr->sequencesStart; const seqDef* const send = seqStorePtr->sequences; const seqDef* sp = sstart; const BYTE* const lstart = seqStorePtr->litStart; const BYTE* const lend = seqStorePtr->lit; const BYTE* lp = lstart; BYTE const* ip = (BYTE const*)src; BYTE const* const iend = ip + srcSize; BYTE* const ostart = (BYTE*)dst; BYTE* const oend = ostart + dstCapacity; BYTE* op = ostart; const BYTE* llCodePtr = seqStorePtr->llCode; const BYTE* mlCodePtr = seqStorePtr->mlCode; const BYTE* ofCodePtr = seqStorePtr->ofCode; size_t targetCBlockSize = cctxParams->targetCBlockSize; size_t litSize, seqCount; int writeLitEntropy = entropyMetadata->hufMetadata.hType == set_compressed; int writeSeqEntropy = 1; int lastSequence = 0; DEBUGLOG(5, "ZSTD_compressSubBlock_multi (litSize=%u, nbSeq=%u)", (unsigned)(lend-lp), (unsigned)(send-sstart)); litSize = 0; seqCount = 0; do { size_t cBlockSizeEstimate = 0; if (sstart == send) { lastSequence = 1; } else { const seqDef* const sequence = sp + seqCount; lastSequence = sequence == send - 1; litSize += ZSTD_getSequenceLength(seqStorePtr, sequence).litLength; seqCount++; } if (lastSequence) { assert(lp <= lend); assert(litSize <= (size_t)(lend - lp)); litSize = (size_t)(lend - lp); } /* I think there is an optimization opportunity here. * Calling ZSTD_estimateSubBlockSize for every sequence can be wasteful * since it recalculates estimate from scratch. * For example, it would recount literal distribution and symbol codes every time. */ cBlockSizeEstimate = ZSTD_estimateSubBlockSize(lp, litSize, ofCodePtr, llCodePtr, mlCodePtr, seqCount, &nextCBlock->entropy, entropyMetadata, workspace, wkspSize, writeLitEntropy, writeSeqEntropy); if (cBlockSizeEstimate > targetCBlockSize || lastSequence) { int litEntropyWritten = 0; int seqEntropyWritten = 0; const size_t decompressedSize = ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, lastSequence); const size_t cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata, sp, seqCount, lp, litSize, llCodePtr, mlCodePtr, ofCodePtr, cctxParams, op, oend-op, bmi2, writeLitEntropy, writeSeqEntropy, &litEntropyWritten, &seqEntropyWritten, lastBlock && lastSequence); FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed"); if (cSize > 0 && cSize < decompressedSize) { DEBUGLOG(5, "Committed the sub-block"); assert(ip + decompressedSize <= iend); ip += decompressedSize; sp += seqCount; lp += litSize; op += cSize; llCodePtr += seqCount; mlCodePtr += seqCount; ofCodePtr += seqCount; litSize = 0; seqCount = 0; /* Entropy only needs to be written once */ if (litEntropyWritten) { writeLitEntropy = 0; } if (seqEntropyWritten) { writeSeqEntropy = 0; } } } } while (!lastSequence); if (writeLitEntropy) { DEBUGLOG(5, "ZSTD_compressSubBlock_multi has literal entropy tables unwritten"); ZSTD_memcpy(&nextCBlock->entropy.huf, &prevCBlock->entropy.huf, sizeof(prevCBlock->entropy.huf)); } if (writeSeqEntropy && ZSTD_needSequenceEntropyTables(&entropyMetadata->fseMetadata)) { /* If we haven't written our entropy tables, then we've violated our contract and * must emit an uncompressed block. */ DEBUGLOG(5, "ZSTD_compressSubBlock_multi has sequence entropy tables unwritten"); return 0; } if (ip < iend) { size_t const cSize = ZSTD_noCompressBlock(op, oend - op, ip, iend - ip, lastBlock); DEBUGLOG(5, "ZSTD_compressSubBlock_multi last sub-block uncompressed, %zu bytes", (size_t)(iend - ip)); FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed"); assert(cSize != 0); op += cSize; /* We have to regenerate the repcodes because we've skipped some sequences */ if (sp < send) { seqDef const* seq; repcodes_t rep; ZSTD_memcpy(&rep, prevCBlock->rep, sizeof(rep)); for (seq = sstart; seq < sp; ++seq) { ZSTD_updateRep(rep.rep, seq->offBase - 1, ZSTD_getSequenceLength(seqStorePtr, seq).litLength == 0); } ZSTD_memcpy(nextCBlock->rep, &rep, sizeof(rep)); } } DEBUGLOG(5, "ZSTD_compressSubBlock_multi compressed"); return op-ostart; } size_t ZSTD_compressSuperBlock(ZSTD_CCtx* zc, void* dst, size_t dstCapacity, void const* src, size_t srcSize, unsigned lastBlock) { ZSTD_entropyCTablesMetadata_t entropyMetadata; FORWARD_IF_ERROR(ZSTD_buildBlockEntropyStats(&zc->seqStore, &zc->blockState.prevCBlock->entropy, &zc->blockState.nextCBlock->entropy, &zc->appliedParams, &entropyMetadata, zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE /* statically allocated in resetCCtx */), ""); return ZSTD_compressSubBlock_multi(&zc->seqStore, zc->blockState.prevCBlock, zc->blockState.nextCBlock, &entropyMetadata, &zc->appliedParams, dst, dstCapacity, src, srcSize, zc->bmi2, lastBlock, zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE /* statically allocated in resetCCtx */); } |