diff options
Diffstat (limited to 'thirdparty/zstd/compress')
-rw-r--r-- | thirdparty/zstd/compress/fse_compress.c | 15 | ||||
-rw-r--r-- | thirdparty/zstd/compress/huf_compress.c | 79 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_compress.c | 443 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_compress_internal.h | 56 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_compress_superblock.c | 337 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_cwksp.h | 32 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_double_fast.c | 22 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_double_fast.h | 11 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_fast.c | 20 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_lazy.c | 216 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_lazy.h | 131 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_ldm.c | 10 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_opt.c | 328 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstd_opt.h | 38 | ||||
-rw-r--r-- | thirdparty/zstd/compress/zstdmt_compress.c | 173 |
15 files changed, 1239 insertions, 672 deletions
diff --git a/thirdparty/zstd/compress/fse_compress.c b/thirdparty/zstd/compress/fse_compress.c index 5d3770808d..1ce3cf16ac 100644 --- a/thirdparty/zstd/compress/fse_compress.c +++ b/thirdparty/zstd/compress/fse_compress.c @@ -25,7 +25,7 @@ #include "../common/error_private.h" #define ZSTD_DEPS_NEED_MALLOC #define ZSTD_DEPS_NEED_MATH64 -#include "../common/zstd_deps.h" /* ZSTD_malloc, ZSTD_free, ZSTD_memcpy, ZSTD_memset */ +#include "../common/zstd_deps.h" /* ZSTD_memset */ #include "../common/bits.h" /* ZSTD_highbit32 */ @@ -225,8 +225,8 @@ size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog) size_t const maxHeaderSize = (((maxSymbolValue+1) * tableLog + 4 /* bitCount initialized at 4 */ + 2 /* first two symbols may use one additional bit each */) / 8) - + 1 /* round up to whole nb bytes */ - + 2 /* additional two bytes for bitstream flush */; + + 1 /* round up to whole nb bytes */ + + 2 /* additional two bytes for bitstream flush */; return maxSymbolValue ? maxHeaderSize : FSE_NCOUNTBOUND; /* maxSymbolValue==0 ? use default */ } @@ -255,7 +255,7 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize, /* Init */ remaining = tableSize+1; /* +1 for extra accuracy */ threshold = tableSize; - nbBits = tableLog+1; + nbBits = (int)tableLog+1; while ((symbol < alphabetSize) && (remaining>1)) { /* stops at 1 */ if (previousIs0) { @@ -274,7 +274,7 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize, } while (symbol >= start+3) { start+=3; - bitStream += 3 << bitCount; + bitStream += 3U << bitCount; bitCount += 2; } bitStream += (symbol-start) << bitCount; @@ -294,7 +294,7 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize, count++; /* +1 for extra accuracy */ if (count>=threshold) count += max; /* [0..max[ [max..threshold[ (...) [threshold+max 2*threshold[ */ - bitStream += count << bitCount; + bitStream += (U32)count << bitCount; bitCount += nbBits; bitCount -= (count<max); previousIs0 = (count==1); @@ -322,7 +322,8 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize, out[1] = (BYTE)(bitStream>>8); out+= (bitCount+7) /8; - return (out-ostart); + assert(out >= ostart); + return (size_t)(out-ostart); } diff --git a/thirdparty/zstd/compress/huf_compress.c b/thirdparty/zstd/compress/huf_compress.c index 29871877a7..ea00072320 100644 --- a/thirdparty/zstd/compress/huf_compress.c +++ b/thirdparty/zstd/compress/huf_compress.c @@ -220,6 +220,25 @@ static void HUF_setValue(HUF_CElt* elt, size_t value) } } +HUF_CTableHeader HUF_readCTableHeader(HUF_CElt const* ctable) +{ + HUF_CTableHeader header; + ZSTD_memcpy(&header, ctable, sizeof(header)); + return header; +} + +static void HUF_writeCTableHeader(HUF_CElt* ctable, U32 tableLog, U32 maxSymbolValue) +{ + HUF_CTableHeader header; + HUF_STATIC_ASSERT(sizeof(ctable[0]) == sizeof(header)); + ZSTD_memset(&header, 0, sizeof(header)); + assert(tableLog < 256); + header.tableLog = (BYTE)tableLog; + assert(maxSymbolValue < 256); + header.maxSymbolValue = (BYTE)maxSymbolValue; + ZSTD_memcpy(ctable, &header, sizeof(header)); +} + typedef struct { HUF_CompressWeightsWksp wksp; BYTE bitsToWeight[HUF_TABLELOG_MAX + 1]; /* precomputed conversion table */ @@ -237,6 +256,9 @@ size_t HUF_writeCTable_wksp(void* dst, size_t maxDstSize, HUF_STATIC_ASSERT(HUF_CTABLE_WORKSPACE_SIZE >= sizeof(HUF_WriteCTableWksp)); + assert(HUF_readCTableHeader(CTable).maxSymbolValue == maxSymbolValue); + assert(HUF_readCTableHeader(CTable).tableLog == huffLog); + /* check conditions */ if (workspaceSize < sizeof(HUF_WriteCTableWksp)) return ERROR(GENERIC); if (maxSymbolValue > HUF_SYMBOLVALUE_MAX) return ERROR(maxSymbolValue_tooLarge); @@ -283,7 +305,9 @@ size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void if (tableLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge); if (nbSymbols > *maxSymbolValuePtr+1) return ERROR(maxSymbolValue_tooSmall); - CTable[0] = tableLog; + *maxSymbolValuePtr = nbSymbols - 1; + + HUF_writeCTableHeader(CTable, tableLog, *maxSymbolValuePtr); /* Prepare base value per rank */ { U32 n, nextRankStart = 0; @@ -315,7 +339,6 @@ size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void { U32 n; for (n=0; n<nbSymbols; n++) HUF_setValue(ct + n, valPerRank[HUF_getNbBits(ct[n])]++); } } - *maxSymbolValuePtr = nbSymbols - 1; return readSize; } @@ -323,6 +346,8 @@ U32 HUF_getNbBitsFromCTable(HUF_CElt const* CTable, U32 symbolValue) { const HUF_CElt* const ct = CTable + 1; assert(symbolValue <= HUF_SYMBOLVALUE_MAX); + if (symbolValue > HUF_readCTableHeader(CTable).maxSymbolValue) + return 0; return (U32)HUF_getNbBits(ct[symbolValue]); } @@ -723,7 +748,8 @@ static void HUF_buildCTableFromTree(HUF_CElt* CTable, nodeElt const* huffNode, i HUF_setNbBits(ct + huffNode[n].byte, huffNode[n].nbBits); /* push nbBits per symbol, symbol order */ for (n=0; n<alphabetSize; n++) HUF_setValue(ct + n, valPerRank[HUF_getNbBits(ct[n])]++); /* assign value within rank, symbol order */ - CTable[0] = maxNbBits; + + HUF_writeCTableHeader(CTable, maxNbBits, maxSymbolValue); } size_t @@ -776,13 +802,20 @@ size_t HUF_estimateCompressedSize(const HUF_CElt* CTable, const unsigned* count, } int HUF_validateCTable(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue) { - HUF_CElt const* ct = CTable + 1; - int bad = 0; - int s; - for (s = 0; s <= (int)maxSymbolValue; ++s) { - bad |= (count[s] != 0) & (HUF_getNbBits(ct[s]) == 0); - } - return !bad; + HUF_CTableHeader header = HUF_readCTableHeader(CTable); + HUF_CElt const* ct = CTable + 1; + int bad = 0; + int s; + + assert(header.tableLog <= HUF_TABLELOG_ABSOLUTEMAX); + + if (header.maxSymbolValue < maxSymbolValue) + return 0; + + for (s = 0; s <= (int)maxSymbolValue; ++s) { + bad |= (count[s] != 0) & (HUF_getNbBits(ct[s]) == 0); + } + return !bad; } size_t HUF_compressBound(size_t size) { return HUF_COMPRESSBOUND(size); } @@ -1024,17 +1057,17 @@ HUF_compress1X_usingCTable_internal_body(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable) { - U32 const tableLog = (U32)CTable[0]; + U32 const tableLog = HUF_readCTableHeader(CTable).tableLog; HUF_CElt const* ct = CTable + 1; const BYTE* ip = (const BYTE*) src; BYTE* const ostart = (BYTE*)dst; BYTE* const oend = ostart + dstSize; - BYTE* op = ostart; HUF_CStream_t bitC; /* init */ if (dstSize < 8) return 0; /* not enough space to compress */ - { size_t const initErr = HUF_initCStream(&bitC, op, (size_t)(oend-op)); + { BYTE* op = ostart; + size_t const initErr = HUF_initCStream(&bitC, op, (size_t)(oend-op)); if (HUF_isError(initErr)) return 0; } if (dstSize < HUF_tightCompressBound(srcSize, (size_t)tableLog) || tableLog > 11) @@ -1255,7 +1288,7 @@ unsigned HUF_optimalTableLog( { BYTE* dst = (BYTE*)workSpace + sizeof(HUF_WriteCTableWksp); size_t dstSize = wkspSize - sizeof(HUF_WriteCTableWksp); - size_t maxBits, hSize, newSize; + size_t hSize, newSize; const unsigned symbolCardinality = HUF_cardinality(count, maxSymbolValue); const unsigned minTableLog = HUF_minTableLog(symbolCardinality); size_t optSize = ((size_t) ~0) - 1; @@ -1266,12 +1299,14 @@ unsigned HUF_optimalTableLog( /* Search until size increases */ for (optLogGuess = minTableLog; optLogGuess <= maxTableLog; optLogGuess++) { DEBUGLOG(7, "checking for huffLog=%u", optLogGuess); - maxBits = HUF_buildCTable_wksp(table, count, maxSymbolValue, optLogGuess, workSpace, wkspSize); - if (ERR_isError(maxBits)) continue; - if (maxBits < optLogGuess && optLogGuess > minTableLog) break; + { size_t maxBits = HUF_buildCTable_wksp(table, count, maxSymbolValue, optLogGuess, workSpace, wkspSize); + if (ERR_isError(maxBits)) continue; + + if (maxBits < optLogGuess && optLogGuess > minTableLog) break; - hSize = HUF_writeCTable_wksp(dst, dstSize, table, maxSymbolValue, (U32)maxBits, workSpace, wkspSize); + hSize = HUF_writeCTable_wksp(dst, dstSize, table, maxSymbolValue, (U32)maxBits, workSpace, wkspSize); + } if (ERR_isError(hSize)) continue; @@ -1372,12 +1407,6 @@ HUF_compress_internal (void* dst, size_t dstSize, huffLog = (U32)maxBits; DEBUGLOG(6, "bit distribution completed (%zu symbols)", showCTableBits(table->CTable + 1, maxSymbolValue+1)); } - /* Zero unused symbols in CTable, so we can check it for validity */ - { - size_t const ctableSize = HUF_CTABLE_SIZE_ST(maxSymbolValue); - size_t const unusedSize = sizeof(table->CTable) - ctableSize * sizeof(HUF_CElt); - ZSTD_memset(table->CTable + ctableSize, 0, unusedSize); - } /* Write table description header */ { CHECK_V_F(hSize, HUF_writeCTable_wksp(op, dstSize, table->CTable, maxSymbolValue, huffLog, @@ -1420,7 +1449,7 @@ size_t HUF_compress1X_repeat (void* dst, size_t dstSize, /* HUF_compress4X_repeat(): * compress input using 4 streams. * consider skipping quickly - * re-use an existing huffman compression table */ + * reuse an existing huffman compression table */ size_t HUF_compress4X_repeat (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned huffLog, diff --git a/thirdparty/zstd/compress/zstd_compress.c b/thirdparty/zstd/compress/zstd_compress.c index d6133e70b4..9284e2a480 100644 --- a/thirdparty/zstd/compress/zstd_compress.c +++ b/thirdparty/zstd/compress/zstd_compress.c @@ -178,6 +178,7 @@ static void ZSTD_freeCCtxContent(ZSTD_CCtx* cctx) size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx) { + DEBUGLOG(3, "ZSTD_freeCCtx (address: %p)", (void*)cctx); if (cctx==NULL) return 0; /* support free on NULL */ RETURN_ERROR_IF(cctx->staticSize, memory_allocation, "not compatible with static CCtx"); @@ -649,10 +650,11 @@ static size_t ZSTD_cParam_clampBounds(ZSTD_cParameter cParam, int* value) return 0; } -#define BOUNDCHECK(cParam, val) { \ - RETURN_ERROR_IF(!ZSTD_cParam_withinBounds(cParam,val), \ - parameter_outOfBound, "Param out of bounds"); \ -} +#define BOUNDCHECK(cParam, val) \ + do { \ + RETURN_ERROR_IF(!ZSTD_cParam_withinBounds(cParam,val), \ + parameter_outOfBound, "Param out of bounds"); \ + } while (0) static int ZSTD_isUpdateAuthorized(ZSTD_cParameter param) @@ -868,7 +870,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams, #else FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value), ""); CCtxParams->nbWorkers = value; - return CCtxParams->nbWorkers; + return (size_t)(CCtxParams->nbWorkers); #endif case ZSTD_c_jobSize : @@ -892,7 +894,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams, #else FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value), ""); CCtxParams->overlapLog = value; - return CCtxParams->overlapLog; + return (size_t)CCtxParams->overlapLog; #endif case ZSTD_c_rsyncable : @@ -902,7 +904,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams, #else FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value), ""); CCtxParams->rsyncable = value; - return CCtxParams->rsyncable; + return (size_t)CCtxParams->rsyncable; #endif case ZSTD_c_enableDedicatedDictSearch : @@ -939,8 +941,10 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams, return CCtxParams->ldmParams.hashRateLog; case ZSTD_c_targetCBlockSize : - if (value!=0) /* 0 ==> default */ + if (value!=0) { /* 0 ==> default */ + value = MAX(value, ZSTD_TARGETCBLOCKSIZE_MIN); BOUNDCHECK(ZSTD_c_targetCBlockSize, value); + } CCtxParams->targetCBlockSize = (U32)value; return CCtxParams->targetCBlockSize; @@ -968,7 +972,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams, case ZSTD_c_validateSequences: BOUNDCHECK(ZSTD_c_validateSequences, value); CCtxParams->validateSequences = value; - return CCtxParams->validateSequences; + return (size_t)CCtxParams->validateSequences; case ZSTD_c_useBlockSplitter: BOUNDCHECK(ZSTD_c_useBlockSplitter, value); @@ -983,7 +987,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams, case ZSTD_c_deterministicRefPrefix: BOUNDCHECK(ZSTD_c_deterministicRefPrefix, value); CCtxParams->deterministicRefPrefix = !!value; - return CCtxParams->deterministicRefPrefix; + return (size_t)CCtxParams->deterministicRefPrefix; case ZSTD_c_prefetchCDictTables: BOUNDCHECK(ZSTD_c_prefetchCDictTables, value); @@ -993,7 +997,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams, case ZSTD_c_enableSeqProducerFallback: BOUNDCHECK(ZSTD_c_enableSeqProducerFallback, value); CCtxParams->enableMatchFinderFallback = value; - return CCtxParams->enableMatchFinderFallback; + return (size_t)CCtxParams->enableMatchFinderFallback; case ZSTD_c_maxBlockSize: if (value!=0) /* 0 ==> default */ @@ -1363,7 +1367,6 @@ size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset) RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong, "Reset parameters is only possible during init stage."); ZSTD_clearAllDicts(cctx); - ZSTD_memset(&cctx->externalMatchCtx, 0, sizeof(cctx->externalMatchCtx)); return ZSTD_CCtxParams_reset(&cctx->requestedParams); } return 0; @@ -1391,11 +1394,12 @@ size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams) static ZSTD_compressionParameters ZSTD_clampCParams(ZSTD_compressionParameters cParams) { -# define CLAMP_TYPE(cParam, val, type) { \ - ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); \ - if ((int)val<bounds.lowerBound) val=(type)bounds.lowerBound; \ - else if ((int)val>bounds.upperBound) val=(type)bounds.upperBound; \ - } +# define CLAMP_TYPE(cParam, val, type) \ + do { \ + ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); \ + if ((int)val<bounds.lowerBound) val=(type)bounds.lowerBound; \ + else if ((int)val>bounds.upperBound) val=(type)bounds.upperBound; \ + } while (0) # define CLAMP(cParam, val) CLAMP_TYPE(cParam, val, unsigned) CLAMP(ZSTD_c_windowLog, cParams.windowLog); CLAMP(ZSTD_c_chainLog, cParams.chainLog); @@ -1467,6 +1471,48 @@ ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar, const U64 maxWindowResize = 1ULL << (ZSTD_WINDOWLOG_MAX-1); assert(ZSTD_checkCParams(cPar)==0); + /* Cascade the selected strategy down to the next-highest one built into + * this binary. */ +#ifdef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR + if (cPar.strategy == ZSTD_btultra2) { + cPar.strategy = ZSTD_btultra; + } + if (cPar.strategy == ZSTD_btultra) { + cPar.strategy = ZSTD_btopt; + } +#endif +#ifdef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR + if (cPar.strategy == ZSTD_btopt) { + cPar.strategy = ZSTD_btlazy2; + } +#endif +#ifdef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR + if (cPar.strategy == ZSTD_btlazy2) { + cPar.strategy = ZSTD_lazy2; + } +#endif +#ifdef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR + if (cPar.strategy == ZSTD_lazy2) { + cPar.strategy = ZSTD_lazy; + } +#endif +#ifdef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR + if (cPar.strategy == ZSTD_lazy) { + cPar.strategy = ZSTD_greedy; + } +#endif +#ifdef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR + if (cPar.strategy == ZSTD_greedy) { + cPar.strategy = ZSTD_dfast; + } +#endif +#ifdef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR + if (cPar.strategy == ZSTD_dfast) { + cPar.strategy = ZSTD_fast; + cPar.targetLength = 0; + } +#endif + switch (mode) { case ZSTD_cpm_unknown: case ZSTD_cpm_noAttachDict: @@ -1617,8 +1663,8 @@ ZSTD_sizeof_matchState(const ZSTD_compressionParameters* const cParams, + ZSTD_cwksp_aligned_alloc_size((MaxLL+1) * sizeof(U32)) + ZSTD_cwksp_aligned_alloc_size((MaxOff+1) * sizeof(U32)) + ZSTD_cwksp_aligned_alloc_size((1<<Litbits) * sizeof(U32)) - + ZSTD_cwksp_aligned_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_match_t)) - + ZSTD_cwksp_aligned_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t)); + + ZSTD_cwksp_aligned_alloc_size(ZSTD_OPT_SIZE * sizeof(ZSTD_match_t)) + + ZSTD_cwksp_aligned_alloc_size(ZSTD_OPT_SIZE * sizeof(ZSTD_optimal_t)); size_t const lazyAdditionalSpace = ZSTD_rowMatchFinderUsed(cParams->strategy, useRowMatchFinder) ? ZSTD_cwksp_aligned_alloc_size(hSize) : 0; @@ -1707,7 +1753,7 @@ size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params) * be needed. However, we still allocate two 0-sized buffers, which can * take space under ASAN. */ return ZSTD_estimateCCtxSize_usingCCtxParams_internal( - &cParams, ¶ms->ldmParams, 1, useRowMatchFinder, 0, 0, ZSTD_CONTENTSIZE_UNKNOWN, params->useSequenceProducer, params->maxBlockSize); + &cParams, ¶ms->ldmParams, 1, useRowMatchFinder, 0, 0, ZSTD_CONTENTSIZE_UNKNOWN, ZSTD_hasExtSeqProd(params), params->maxBlockSize); } size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams) @@ -1768,7 +1814,7 @@ size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params) return ZSTD_estimateCCtxSize_usingCCtxParams_internal( &cParams, ¶ms->ldmParams, 1, useRowMatchFinder, inBuffSize, outBuffSize, - ZSTD_CONTENTSIZE_UNKNOWN, params->useSequenceProducer, params->maxBlockSize); + ZSTD_CONTENTSIZE_UNKNOWN, ZSTD_hasExtSeqProd(params), params->maxBlockSize); } } @@ -2001,8 +2047,8 @@ ZSTD_reset_matchState(ZSTD_matchState_t* ms, ms->opt.litLengthFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxLL+1) * sizeof(unsigned)); ms->opt.matchLengthFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxML+1) * sizeof(unsigned)); ms->opt.offCodeFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxOff+1) * sizeof(unsigned)); - ms->opt.matchTable = (ZSTD_match_t*)ZSTD_cwksp_reserve_aligned(ws, (ZSTD_OPT_NUM+1) * sizeof(ZSTD_match_t)); - ms->opt.priceTable = (ZSTD_optimal_t*)ZSTD_cwksp_reserve_aligned(ws, (ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t)); + ms->opt.matchTable = (ZSTD_match_t*)ZSTD_cwksp_reserve_aligned(ws, ZSTD_OPT_SIZE * sizeof(ZSTD_match_t)); + ms->opt.priceTable = (ZSTD_optimal_t*)ZSTD_cwksp_reserve_aligned(ws, ZSTD_OPT_SIZE * sizeof(ZSTD_optimal_t)); } ms->cParams = *cParams; @@ -2074,7 +2120,7 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc, { size_t const windowSize = MAX(1, (size_t)MIN(((U64)1 << params->cParams.windowLog), pledgedSrcSize)); size_t const blockSize = MIN(params->maxBlockSize, windowSize); - size_t const maxNbSeq = ZSTD_maxNbSeq(blockSize, params->cParams.minMatch, params->useSequenceProducer); + size_t const maxNbSeq = ZSTD_maxNbSeq(blockSize, params->cParams.minMatch, ZSTD_hasExtSeqProd(params)); size_t const buffOutSize = (zbuff == ZSTDb_buffered && params->outBufferMode == ZSTD_bm_buffered) ? ZSTD_compressBound(blockSize) + 1 : 0; @@ -2091,8 +2137,7 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc, size_t const neededSpace = ZSTD_estimateCCtxSize_usingCCtxParams_internal( ¶ms->cParams, ¶ms->ldmParams, zc->staticSize != 0, params->useRowMatchFinder, - buffInSize, buffOutSize, pledgedSrcSize, params->useSequenceProducer, params->maxBlockSize); - int resizeWorkspace; + buffInSize, buffOutSize, pledgedSrcSize, ZSTD_hasExtSeqProd(params), params->maxBlockSize); FORWARD_IF_ERROR(neededSpace, "cctx size estimate failed!"); @@ -2101,7 +2146,7 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc, { /* Check if workspace is large enough, alloc a new one if needed */ int const workspaceTooSmall = ZSTD_cwksp_sizeof(ws) < neededSpace; int const workspaceWasteful = ZSTD_cwksp_check_wasteful(ws, neededSpace); - resizeWorkspace = workspaceTooSmall || workspaceWasteful; + int resizeWorkspace = workspaceTooSmall || workspaceWasteful; DEBUGLOG(4, "Need %zu B workspace", neededSpace); DEBUGLOG(4, "windowSize: %zu - blockSize: %zu", windowSize, blockSize); @@ -2176,10 +2221,10 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc, } /* reserve space for block-level external sequences */ - if (params->useSequenceProducer) { + if (ZSTD_hasExtSeqProd(params)) { size_t const maxNbExternalSeq = ZSTD_sequenceBound(blockSize); - zc->externalMatchCtx.seqBufferCapacity = maxNbExternalSeq; - zc->externalMatchCtx.seqBuffer = + zc->extSeqBufCapacity = maxNbExternalSeq; + zc->extSeqBuf = (ZSTD_Sequence*)ZSTD_cwksp_reserve_aligned(ws, maxNbExternalSeq * sizeof(ZSTD_Sequence)); } @@ -2564,7 +2609,7 @@ ZSTD_reduceTable_internal (U32* const table, U32 const size, U32 const reducerVa assert(size < (1U<<31)); /* can be casted to int */ #if ZSTD_MEMORY_SANITIZER && !defined (ZSTD_MSAN_DONT_POISON_WORKSPACE) - /* To validate that the table re-use logic is sound, and that we don't + /* To validate that the table reuse logic is sound, and that we don't * access table space that we haven't cleaned, we re-"poison" the table * space every time we mark it dirty. * @@ -2992,40 +3037,43 @@ ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_paramS static const ZSTD_blockCompressor blockCompressor[4][ZSTD_STRATEGY_MAX+1] = { { ZSTD_compressBlock_fast /* default for 0 */, ZSTD_compressBlock_fast, - ZSTD_compressBlock_doubleFast, - ZSTD_compressBlock_greedy, - ZSTD_compressBlock_lazy, - ZSTD_compressBlock_lazy2, - ZSTD_compressBlock_btlazy2, - ZSTD_compressBlock_btopt, - ZSTD_compressBlock_btultra, - ZSTD_compressBlock_btultra2 }, + ZSTD_COMPRESSBLOCK_DOUBLEFAST, + ZSTD_COMPRESSBLOCK_GREEDY, + ZSTD_COMPRESSBLOCK_LAZY, + ZSTD_COMPRESSBLOCK_LAZY2, + ZSTD_COMPRESSBLOCK_BTLAZY2, + ZSTD_COMPRESSBLOCK_BTOPT, + ZSTD_COMPRESSBLOCK_BTULTRA, + ZSTD_COMPRESSBLOCK_BTULTRA2 + }, { ZSTD_compressBlock_fast_extDict /* default for 0 */, ZSTD_compressBlock_fast_extDict, - ZSTD_compressBlock_doubleFast_extDict, - ZSTD_compressBlock_greedy_extDict, - ZSTD_compressBlock_lazy_extDict, - ZSTD_compressBlock_lazy2_extDict, - ZSTD_compressBlock_btlazy2_extDict, - ZSTD_compressBlock_btopt_extDict, - ZSTD_compressBlock_btultra_extDict, - ZSTD_compressBlock_btultra_extDict }, + ZSTD_COMPRESSBLOCK_DOUBLEFAST_EXTDICT, + ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT, + ZSTD_COMPRESSBLOCK_LAZY_EXTDICT, + ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT, + ZSTD_COMPRESSBLOCK_BTLAZY2_EXTDICT, + ZSTD_COMPRESSBLOCK_BTOPT_EXTDICT, + ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT, + ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT + }, { ZSTD_compressBlock_fast_dictMatchState /* default for 0 */, ZSTD_compressBlock_fast_dictMatchState, - ZSTD_compressBlock_doubleFast_dictMatchState, - ZSTD_compressBlock_greedy_dictMatchState, - ZSTD_compressBlock_lazy_dictMatchState, - ZSTD_compressBlock_lazy2_dictMatchState, - ZSTD_compressBlock_btlazy2_dictMatchState, - ZSTD_compressBlock_btopt_dictMatchState, - ZSTD_compressBlock_btultra_dictMatchState, - ZSTD_compressBlock_btultra_dictMatchState }, + ZSTD_COMPRESSBLOCK_DOUBLEFAST_DICTMATCHSTATE, + ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE, + ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE, + ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE, + ZSTD_COMPRESSBLOCK_BTLAZY2_DICTMATCHSTATE, + ZSTD_COMPRESSBLOCK_BTOPT_DICTMATCHSTATE, + ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE, + ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE + }, { NULL /* default for 0 */, NULL, NULL, - ZSTD_compressBlock_greedy_dedicatedDictSearch, - ZSTD_compressBlock_lazy_dedicatedDictSearch, - ZSTD_compressBlock_lazy2_dedicatedDictSearch, + ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH, + ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH, + ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH, NULL, NULL, NULL, @@ -3038,18 +3086,26 @@ ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_paramS DEBUGLOG(4, "Selected block compressor: dictMode=%d strat=%d rowMatchfinder=%d", (int)dictMode, (int)strat, (int)useRowMatchFinder); if (ZSTD_rowMatchFinderUsed(strat, useRowMatchFinder)) { static const ZSTD_blockCompressor rowBasedBlockCompressors[4][3] = { - { ZSTD_compressBlock_greedy_row, - ZSTD_compressBlock_lazy_row, - ZSTD_compressBlock_lazy2_row }, - { ZSTD_compressBlock_greedy_extDict_row, - ZSTD_compressBlock_lazy_extDict_row, - ZSTD_compressBlock_lazy2_extDict_row }, - { ZSTD_compressBlock_greedy_dictMatchState_row, - ZSTD_compressBlock_lazy_dictMatchState_row, - ZSTD_compressBlock_lazy2_dictMatchState_row }, - { ZSTD_compressBlock_greedy_dedicatedDictSearch_row, - ZSTD_compressBlock_lazy_dedicatedDictSearch_row, - ZSTD_compressBlock_lazy2_dedicatedDictSearch_row } + { + ZSTD_COMPRESSBLOCK_GREEDY_ROW, + ZSTD_COMPRESSBLOCK_LAZY_ROW, + ZSTD_COMPRESSBLOCK_LAZY2_ROW + }, + { + ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT_ROW, + ZSTD_COMPRESSBLOCK_LAZY_EXTDICT_ROW, + ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT_ROW + }, + { + ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE_ROW, + ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE_ROW, + ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE_ROW + }, + { + ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH_ROW, + ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH_ROW, + ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH_ROW + } }; DEBUGLOG(4, "Selecting a row-based matchfinder"); assert(useRowMatchFinder != ZSTD_ps_auto); @@ -3192,7 +3248,7 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize) /* External matchfinder + LDM is technically possible, just not implemented yet. * We need to revisit soon and implement it. */ RETURN_ERROR_IF( - zc->appliedParams.useSequenceProducer, + ZSTD_hasExtSeqProd(&zc->appliedParams), parameter_combination_unsupported, "Long-distance matching with external sequence producer enabled is not currently supported." ); @@ -3211,7 +3267,7 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize) /* External matchfinder + LDM is technically possible, just not implemented yet. * We need to revisit soon and implement it. */ RETURN_ERROR_IF( - zc->appliedParams.useSequenceProducer, + ZSTD_hasExtSeqProd(&zc->appliedParams), parameter_combination_unsupported, "Long-distance matching with external sequence producer enabled is not currently supported." ); @@ -3230,18 +3286,18 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize) zc->appliedParams.useRowMatchFinder, src, srcSize); assert(ldmSeqStore.pos == ldmSeqStore.size); - } else if (zc->appliedParams.useSequenceProducer) { + } else if (ZSTD_hasExtSeqProd(&zc->appliedParams)) { assert( - zc->externalMatchCtx.seqBufferCapacity >= ZSTD_sequenceBound(srcSize) + zc->extSeqBufCapacity >= ZSTD_sequenceBound(srcSize) ); - assert(zc->externalMatchCtx.mFinder != NULL); + assert(zc->appliedParams.extSeqProdFunc != NULL); { U32 const windowSize = (U32)1 << zc->appliedParams.cParams.windowLog; - size_t const nbExternalSeqs = (zc->externalMatchCtx.mFinder)( - zc->externalMatchCtx.mState, - zc->externalMatchCtx.seqBuffer, - zc->externalMatchCtx.seqBufferCapacity, + size_t const nbExternalSeqs = (zc->appliedParams.extSeqProdFunc)( + zc->appliedParams.extSeqProdState, + zc->extSeqBuf, + zc->extSeqBufCapacity, src, srcSize, NULL, 0, /* dict and dictSize, currently not supported */ zc->appliedParams.compressionLevel, @@ -3249,21 +3305,21 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize) ); size_t const nbPostProcessedSeqs = ZSTD_postProcessSequenceProducerResult( - zc->externalMatchCtx.seqBuffer, + zc->extSeqBuf, nbExternalSeqs, - zc->externalMatchCtx.seqBufferCapacity, + zc->extSeqBufCapacity, srcSize ); /* Return early if there is no error, since we don't need to worry about last literals */ if (!ZSTD_isError(nbPostProcessedSeqs)) { ZSTD_sequencePosition seqPos = {0,0,0}; - size_t const seqLenSum = ZSTD_fastSequenceLengthSum(zc->externalMatchCtx.seqBuffer, nbPostProcessedSeqs); + size_t const seqLenSum = ZSTD_fastSequenceLengthSum(zc->extSeqBuf, nbPostProcessedSeqs); RETURN_ERROR_IF(seqLenSum > srcSize, externalSequences_invalid, "External sequences imply too large a block!"); FORWARD_IF_ERROR( ZSTD_copySequencesToSeqStoreExplicitBlockDelim( zc, &seqPos, - zc->externalMatchCtx.seqBuffer, nbPostProcessedSeqs, + zc->extSeqBuf, nbPostProcessedSeqs, src, srcSize, zc->appliedParams.searchForExternalRepcodes ), @@ -3280,9 +3336,11 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize) } /* Fallback to software matchfinder */ - { ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, - zc->appliedParams.useRowMatchFinder, - dictMode); + { ZSTD_blockCompressor const blockCompressor = + ZSTD_selectBlockCompressor( + zc->appliedParams.cParams.strategy, + zc->appliedParams.useRowMatchFinder, + dictMode); ms->ldmSeqStore = NULL; DEBUGLOG( 5, @@ -3292,9 +3350,10 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize) lastLLSize = blockCompressor(ms, &zc->seqStore, zc->blockState.nextCBlock->rep, src, srcSize); } } } else { /* not long range mode and no external matchfinder */ - ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, - zc->appliedParams.useRowMatchFinder, - dictMode); + ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor( + zc->appliedParams.cParams.strategy, + zc->appliedParams.useRowMatchFinder, + dictMode); ms->ldmSeqStore = NULL; lastLLSize = blockCompressor(ms, &zc->seqStore, zc->blockState.nextCBlock->rep, src, srcSize); } @@ -3304,29 +3363,38 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize) return ZSTDbss_compress; } -static void ZSTD_copyBlockSequences(ZSTD_CCtx* zc) +static size_t ZSTD_copyBlockSequences(SeqCollector* seqCollector, const seqStore_t* seqStore, const U32 prevRepcodes[ZSTD_REP_NUM]) { - const seqStore_t* seqStore = ZSTD_getSeqStore(zc); - const seqDef* seqStoreSeqs = seqStore->sequencesStart; - size_t seqStoreSeqSize = seqStore->sequences - seqStoreSeqs; - size_t seqStoreLiteralsSize = (size_t)(seqStore->lit - seqStore->litStart); - size_t literalsRead = 0; - size_t lastLLSize; + const seqDef* inSeqs = seqStore->sequencesStart; + const size_t nbInSequences = seqStore->sequences - inSeqs; + const size_t nbInLiterals = (size_t)(seqStore->lit - seqStore->litStart); - ZSTD_Sequence* outSeqs = &zc->seqCollector.seqStart[zc->seqCollector.seqIndex]; + ZSTD_Sequence* outSeqs = seqCollector->seqIndex == 0 ? seqCollector->seqStart : seqCollector->seqStart + seqCollector->seqIndex; + const size_t nbOutSequences = nbInSequences + 1; + size_t nbOutLiterals = 0; + repcodes_t repcodes; size_t i; - repcodes_t updatedRepcodes; - assert(zc->seqCollector.seqIndex + 1 < zc->seqCollector.maxSequences); - /* Ensure we have enough space for last literals "sequence" */ - assert(zc->seqCollector.maxSequences >= seqStoreSeqSize + 1); - ZSTD_memcpy(updatedRepcodes.rep, zc->blockState.prevCBlock->rep, sizeof(repcodes_t)); - for (i = 0; i < seqStoreSeqSize; ++i) { - U32 rawOffset = seqStoreSeqs[i].offBase - ZSTD_REP_NUM; - outSeqs[i].litLength = seqStoreSeqs[i].litLength; - outSeqs[i].matchLength = seqStoreSeqs[i].mlBase + MINMATCH; + /* Bounds check that we have enough space for every input sequence + * and the block delimiter + */ + assert(seqCollector->seqIndex <= seqCollector->maxSequences); + RETURN_ERROR_IF( + nbOutSequences > (size_t)(seqCollector->maxSequences - seqCollector->seqIndex), + dstSize_tooSmall, + "Not enough space to copy sequences"); + + ZSTD_memcpy(&repcodes, prevRepcodes, sizeof(repcodes)); + for (i = 0; i < nbInSequences; ++i) { + U32 rawOffset; + outSeqs[i].litLength = inSeqs[i].litLength; + outSeqs[i].matchLength = inSeqs[i].mlBase + MINMATCH; outSeqs[i].rep = 0; + /* Handle the possible single length >= 64K + * There can only be one because we add MINMATCH to every match length, + * and blocks are at most 128K. + */ if (i == seqStore->longLengthPos) { if (seqStore->longLengthType == ZSTD_llt_literalLength) { outSeqs[i].litLength += 0x10000; @@ -3335,41 +3403,55 @@ static void ZSTD_copyBlockSequences(ZSTD_CCtx* zc) } } - if (seqStoreSeqs[i].offBase <= ZSTD_REP_NUM) { - /* Derive the correct offset corresponding to a repcode */ - outSeqs[i].rep = seqStoreSeqs[i].offBase; + /* Determine the raw offset given the offBase, which may be a repcode. */ + if (OFFBASE_IS_REPCODE(inSeqs[i].offBase)) { + const U32 repcode = OFFBASE_TO_REPCODE(inSeqs[i].offBase); + assert(repcode > 0); + outSeqs[i].rep = repcode; if (outSeqs[i].litLength != 0) { - rawOffset = updatedRepcodes.rep[outSeqs[i].rep - 1]; + rawOffset = repcodes.rep[repcode - 1]; } else { - if (outSeqs[i].rep == 3) { - rawOffset = updatedRepcodes.rep[0] - 1; + if (repcode == 3) { + assert(repcodes.rep[0] > 1); + rawOffset = repcodes.rep[0] - 1; } else { - rawOffset = updatedRepcodes.rep[outSeqs[i].rep]; + rawOffset = repcodes.rep[repcode]; } } + } else { + rawOffset = OFFBASE_TO_OFFSET(inSeqs[i].offBase); } outSeqs[i].offset = rawOffset; - /* seqStoreSeqs[i].offset == offCode+1, and ZSTD_updateRep() expects offCode - so we provide seqStoreSeqs[i].offset - 1 */ - ZSTD_updateRep(updatedRepcodes.rep, - seqStoreSeqs[i].offBase, - seqStoreSeqs[i].litLength == 0); - literalsRead += outSeqs[i].litLength; + + /* Update repcode history for the sequence */ + ZSTD_updateRep(repcodes.rep, + inSeqs[i].offBase, + inSeqs[i].litLength == 0); + + nbOutLiterals += outSeqs[i].litLength; } /* Insert last literals (if any exist) in the block as a sequence with ml == off == 0. * If there are no last literals, then we'll emit (of: 0, ml: 0, ll: 0), which is a marker * for the block boundary, according to the API. */ - assert(seqStoreLiteralsSize >= literalsRead); - lastLLSize = seqStoreLiteralsSize - literalsRead; - outSeqs[i].litLength = (U32)lastLLSize; - outSeqs[i].matchLength = outSeqs[i].offset = outSeqs[i].rep = 0; - seqStoreSeqSize++; - zc->seqCollector.seqIndex += seqStoreSeqSize; + assert(nbInLiterals >= nbOutLiterals); + { + const size_t lastLLSize = nbInLiterals - nbOutLiterals; + outSeqs[nbInSequences].litLength = (U32)lastLLSize; + outSeqs[nbInSequences].matchLength = 0; + outSeqs[nbInSequences].offset = 0; + assert(nbOutSequences == nbInSequences + 1); + } + seqCollector->seqIndex += nbOutSequences; + assert(seqCollector->seqIndex <= seqCollector->maxSequences); + + return 0; } size_t ZSTD_sequenceBound(size_t srcSize) { - return (srcSize / ZSTD_MINMATCH_MIN) + 1; + const size_t maxNbSeq = (srcSize / ZSTD_MINMATCH_MIN) + 1; + const size_t maxNbDelims = (srcSize / ZSTD_BLOCKSIZE_MAX_MIN) + 1; + return maxNbSeq + maxNbDelims; } size_t ZSTD_generateSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs, @@ -3378,6 +3460,16 @@ size_t ZSTD_generateSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs, const size_t dstCapacity = ZSTD_compressBound(srcSize); void* dst = ZSTD_customMalloc(dstCapacity, ZSTD_defaultCMem); SeqCollector seqCollector; + { + int targetCBlockSize; + FORWARD_IF_ERROR(ZSTD_CCtx_getParameter(zc, ZSTD_c_targetCBlockSize, &targetCBlockSize), ""); + RETURN_ERROR_IF(targetCBlockSize != 0, parameter_unsupported, "targetCBlockSize != 0"); + } + { + int nbWorkers; + FORWARD_IF_ERROR(ZSTD_CCtx_getParameter(zc, ZSTD_c_nbWorkers, &nbWorkers), ""); + RETURN_ERROR_IF(nbWorkers != 0, parameter_unsupported, "nbWorkers != 0"); + } RETURN_ERROR_IF(dst == NULL, memory_allocation, "NULL pointer!"); @@ -3387,8 +3479,12 @@ size_t ZSTD_generateSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs, seqCollector.maxSequences = outSeqsSize; zc->seqCollector = seqCollector; - ZSTD_compress2(zc, dst, dstCapacity, src, srcSize); - ZSTD_customFree(dst, ZSTD_defaultCMem); + { + const size_t ret = ZSTD_compress2(zc, dst, dstCapacity, src, srcSize); + ZSTD_customFree(dst, ZSTD_defaultCMem); + FORWARD_IF_ERROR(ret, "ZSTD_compress2 failed"); + } + assert(zc->seqCollector.seqIndex <= ZSTD_sequenceBound(srcSize)); return zc->seqCollector.seqIndex; } @@ -3981,8 +4077,9 @@ ZSTD_compressSeqStore_singleBlock(ZSTD_CCtx* zc, cSeqsSize = 1; } + /* Sequence collection not supported when block splitting */ if (zc->seqCollector.collectSequences) { - ZSTD_copyBlockSequences(zc); + FORWARD_IF_ERROR(ZSTD_copyBlockSequences(&zc->seqCollector, seqStore, dRepOriginal.rep), "copyBlockSequences failed"); ZSTD_blockState_confirmRepcodesAndEntropyTables(&zc->blockState); return 0; } @@ -4204,6 +4301,7 @@ ZSTD_compressBlock_splitBlock(ZSTD_CCtx* zc, if (bss == ZSTDbss_noCompress) { if (zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid) zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check; + RETURN_ERROR_IF(zc->seqCollector.collectSequences, sequenceProducer_failed, "Uncompressible block"); cSize = ZSTD_noCompressBlock(dst, dstCapacity, src, srcSize, lastBlock); FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed"); DEBUGLOG(4, "ZSTD_compressBlock_splitBlock: Nocompress block"); @@ -4236,11 +4334,15 @@ ZSTD_compressBlock_internal(ZSTD_CCtx* zc, { const size_t bss = ZSTD_buildSeqStore(zc, src, srcSize); FORWARD_IF_ERROR(bss, "ZSTD_buildSeqStore failed"); - if (bss == ZSTDbss_noCompress) { cSize = 0; goto out; } + if (bss == ZSTDbss_noCompress) { + RETURN_ERROR_IF(zc->seqCollector.collectSequences, sequenceProducer_failed, "Uncompressible block"); + cSize = 0; + goto out; + } } if (zc->seqCollector.collectSequences) { - ZSTD_copyBlockSequences(zc); + FORWARD_IF_ERROR(ZSTD_copyBlockSequences(&zc->seqCollector, ZSTD_getSeqStore(zc), zc->blockState.prevCBlock->rep), "copyBlockSequences failed"); ZSTD_blockState_confirmRepcodesAndEntropyTables(&zc->blockState); return 0; } @@ -4553,19 +4655,15 @@ size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity) } } -size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq) +void ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq) { - RETURN_ERROR_IF(cctx->stage != ZSTDcs_init, stage_wrong, - "wrong cctx stage"); - RETURN_ERROR_IF(cctx->appliedParams.ldmParams.enableLdm == ZSTD_ps_enable, - parameter_unsupported, - "incompatible with ldm"); + assert(cctx->stage == ZSTDcs_init); + assert(nbSeq == 0 || cctx->appliedParams.ldmParams.enableLdm != ZSTD_ps_enable); cctx->externSeqStore.seq = seq; cctx->externSeqStore.size = nbSeq; cctx->externSeqStore.capacity = nbSeq; cctx->externSeqStore.pos = 0; cctx->externSeqStore.posInSequence = 0; - return 0; } @@ -4760,12 +4858,19 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms, ZSTD_fillHashTable(ms, iend, dtlm, tfp); break; case ZSTD_dfast: +#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR ZSTD_fillDoubleHashTable(ms, iend, dtlm, tfp); +#else + assert(0); /* shouldn't be called: cparams should've been adjusted. */ +#endif break; case ZSTD_greedy: case ZSTD_lazy: case ZSTD_lazy2: +#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) assert(srcSize >= HASH_READ_SIZE); if (ms->dedicatedDictSearch) { assert(ms->chainTable != NULL); @@ -4782,14 +4887,23 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms, DEBUGLOG(4, "Using chain-based hash table for lazy dict"); } } +#else + assert(0); /* shouldn't be called: cparams should've been adjusted. */ +#endif break; case ZSTD_btlazy2: /* we want the dictionary table fully sorted */ case ZSTD_btopt: case ZSTD_btultra: case ZSTD_btultra2: +#if !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR) assert(srcSize >= HASH_READ_SIZE); ZSTD_updateTree(ms, iend-HASH_READ_SIZE, iend); +#else + assert(0); /* shouldn't be called: cparams should've been adjusted. */ +#endif break; default: @@ -4836,11 +4950,10 @@ size_t ZSTD_loadCEntropy(ZSTD_compressedBlockState_t* bs, void* workspace, /* We only set the loaded table as valid if it contains all non-zero * weights. Otherwise, we set it to check */ - if (!hasZeroWeights) + if (!hasZeroWeights && maxSymbolValue == 255) bs->entropy.huf.repeatMode = HUF_repeat_valid; RETURN_ERROR_IF(HUF_isError(hufHeaderSize), dictionary_corrupted, ""); - RETURN_ERROR_IF(maxSymbolValue < 255, dictionary_corrupted, ""); dictPtr += hufHeaderSize; } @@ -5107,14 +5220,13 @@ static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity) { BYTE* const ostart = (BYTE*)dst; BYTE* op = ostart; - size_t fhSize = 0; DEBUGLOG(4, "ZSTD_writeEpilogue"); RETURN_ERROR_IF(cctx->stage == ZSTDcs_created, stage_wrong, "init missing"); /* special case : empty frame */ if (cctx->stage == ZSTDcs_init) { - fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, &cctx->appliedParams, 0, 0); + size_t fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, &cctx->appliedParams, 0, 0); FORWARD_IF_ERROR(fhSize, "ZSTD_writeFrameHeader failed"); dstCapacity -= fhSize; op += fhSize; @@ -5124,8 +5236,9 @@ static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity) if (cctx->stage != ZSTDcs_ending) { /* write one last empty block, make it the "last" block */ U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1) + 0; - RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall, "no room for epilogue"); - MEM_writeLE32(op, cBlockHeader24); + ZSTD_STATIC_ASSERT(ZSTD_BLOCKHEADERSIZE == 3); + RETURN_ERROR_IF(dstCapacity<3, dstSize_tooSmall, "no room for epilogue"); + MEM_writeLE24(op, cBlockHeader24); op += ZSTD_blockHeaderSize; dstCapacity -= ZSTD_blockHeaderSize; } @@ -5455,7 +5568,7 @@ ZSTD_CDict* ZSTD_createCDict_advanced2( cctxParams.useRowMatchFinder, cctxParams.enableDedicatedDictSearch, customMem); - if (ZSTD_isError( ZSTD_initCDict_internal(cdict, + if (!cdict || ZSTD_isError( ZSTD_initCDict_internal(cdict, dict, dictSize, dictLoadMethod, dictContentType, cctxParams) )) { @@ -5879,7 +5992,7 @@ static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, if (zcs->appliedParams.inBufferMode == ZSTD_bm_stable) { assert(input->pos >= zcs->stableIn_notConsumed); input->pos -= zcs->stableIn_notConsumed; - ip -= zcs->stableIn_notConsumed; + if (ip) ip -= zcs->stableIn_notConsumed; zcs->stableIn_notConsumed = 0; } if (zcs->appliedParams.inBufferMode == ZSTD_bm_buffered) { @@ -6138,7 +6251,7 @@ static size_t ZSTD_CCtx_init_compressStream2(ZSTD_CCtx* cctx, #ifdef ZSTD_MULTITHREAD /* If external matchfinder is enabled, make sure to fail before checking job size (for consistency) */ RETURN_ERROR_IF( - params.useSequenceProducer == 1 && params.nbWorkers >= 1, + ZSTD_hasExtSeqProd(¶ms) && params.nbWorkers >= 1, parameter_combination_unsupported, "External sequence producer isn't supported with nbWorkers >= 1" ); @@ -6430,7 +6543,7 @@ ZSTD_copySequencesToSeqStoreExplicitBlockDelim(ZSTD_CCtx* cctx, if (cctx->appliedParams.validateSequences) { seqPos->posInSrc += litLength + matchLength; FORWARD_IF_ERROR(ZSTD_validateSequence(offBase, matchLength, cctx->appliedParams.cParams.minMatch, seqPos->posInSrc, - cctx->appliedParams.cParams.windowLog, dictSize, cctx->appliedParams.useSequenceProducer), + cctx->appliedParams.cParams.windowLog, dictSize, ZSTD_hasExtSeqProd(&cctx->appliedParams)), "Sequence validation failed"); } RETURN_ERROR_IF(idx - seqPos->idx >= cctx->seqStore.maxNbSeq, externalSequences_invalid, @@ -6568,7 +6681,7 @@ ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* if (cctx->appliedParams.validateSequences) { seqPos->posInSrc += litLength + matchLength; FORWARD_IF_ERROR(ZSTD_validateSequence(offBase, matchLength, cctx->appliedParams.cParams.minMatch, seqPos->posInSrc, - cctx->appliedParams.cParams.windowLog, dictSize, cctx->appliedParams.useSequenceProducer), + cctx->appliedParams.cParams.windowLog, dictSize, ZSTD_hasExtSeqProd(&cctx->appliedParams)), "Sequence validation failed"); } DEBUGLOG(6, "Storing sequence: (of: %u, ml: %u, ll: %u)", offBase, matchLength, litLength); @@ -7014,19 +7127,27 @@ ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSizeH } void ZSTD_registerSequenceProducer( - ZSTD_CCtx* zc, void* mState, - ZSTD_sequenceProducer_F* mFinder + ZSTD_CCtx* zc, + void* extSeqProdState, + ZSTD_sequenceProducer_F extSeqProdFunc +) { + assert(zc != NULL); + ZSTD_CCtxParams_registerSequenceProducer( + &zc->requestedParams, extSeqProdState, extSeqProdFunc + ); +} + +void ZSTD_CCtxParams_registerSequenceProducer( + ZSTD_CCtx_params* params, + void* extSeqProdState, + ZSTD_sequenceProducer_F extSeqProdFunc ) { - if (mFinder != NULL) { - ZSTD_externalMatchCtx emctx; - emctx.mState = mState; - emctx.mFinder = mFinder; - emctx.seqBuffer = NULL; - emctx.seqBufferCapacity = 0; - zc->externalMatchCtx = emctx; - zc->requestedParams.useSequenceProducer = 1; + assert(params != NULL); + if (extSeqProdFunc != NULL) { + params->extSeqProdFunc = extSeqProdFunc; + params->extSeqProdState = extSeqProdState; } else { - ZSTD_memset(&zc->externalMatchCtx, 0, sizeof(zc->externalMatchCtx)); - zc->requestedParams.useSequenceProducer = 0; + params->extSeqProdFunc = NULL; + params->extSeqProdState = NULL; } } diff --git a/thirdparty/zstd/compress/zstd_compress_internal.h b/thirdparty/zstd/compress/zstd_compress_internal.h index 10f68d010e..e41d7b78ec 100644 --- a/thirdparty/zstd/compress/zstd_compress_internal.h +++ b/thirdparty/zstd/compress/zstd_compress_internal.h @@ -39,7 +39,7 @@ extern "C" { It's not a big deal though : candidate will just be sorted again. Additionally, candidate position 1 will be lost. But candidate 1 cannot hide a large tree of candidates, so it's a minimal loss. - The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table re-use with a different strategy. + The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table reuse with a different strategy. This constant is required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */ @@ -159,23 +159,24 @@ typedef struct { UNUSED_ATTR static const rawSeqStore_t kNullRawSeqStore = {NULL, 0, 0, 0, 0}; typedef struct { - int price; - U32 off; - U32 mlen; - U32 litlen; - U32 rep[ZSTD_REP_NUM]; + int price; /* price from beginning of segment to this position */ + U32 off; /* offset of previous match */ + U32 mlen; /* length of previous match */ + U32 litlen; /* nb of literals since previous match */ + U32 rep[ZSTD_REP_NUM]; /* offset history after previous match */ } ZSTD_optimal_t; typedef enum { zop_dynamic=0, zop_predef } ZSTD_OptPrice_e; +#define ZSTD_OPT_SIZE (ZSTD_OPT_NUM+3) typedef struct { /* All tables are allocated inside cctx->workspace by ZSTD_resetCCtx_internal() */ unsigned* litFreq; /* table of literals statistics, of size 256 */ unsigned* litLengthFreq; /* table of litLength statistics, of size (MaxLL+1) */ unsigned* matchLengthFreq; /* table of matchLength statistics, of size (MaxML+1) */ unsigned* offCodeFreq; /* table of offCode statistics, of size (MaxOff+1) */ - ZSTD_match_t* matchTable; /* list of found matches, of size ZSTD_OPT_NUM+1 */ - ZSTD_optimal_t* priceTable; /* All positions tracked by optimal parser, of size ZSTD_OPT_NUM+1 */ + ZSTD_match_t* matchTable; /* list of found matches, of size ZSTD_OPT_SIZE */ + ZSTD_optimal_t* priceTable; /* All positions tracked by optimal parser, of size ZSTD_OPT_SIZE */ U32 litSum; /* nb of literals */ U32 litLengthSum; /* nb of litLength codes */ @@ -228,7 +229,7 @@ struct ZSTD_matchState_t { U32 rowHashLog; /* For row-based matchfinder: Hashlog based on nb of rows in the hashTable.*/ BYTE* tagTable; /* For row-based matchFinder: A row-based table containing the hashes and head index. */ U32 hashCache[ZSTD_ROW_HASH_CACHE_SIZE]; /* For row-based matchFinder: a cache of hashes to improve speed */ - U64 hashSalt; /* For row-based matchFinder: salts the hash for re-use of tag table */ + U64 hashSalt; /* For row-based matchFinder: salts the hash for reuse of tag table */ U32 hashSaltEntropy; /* For row-based matchFinder: collects entropy for salt generation */ U32* hashTable; @@ -360,10 +361,11 @@ struct ZSTD_CCtx_params_s { * if the external matchfinder returns an error code. */ int enableMatchFinderFallback; - /* Indicates whether an external matchfinder has been referenced. - * Users can't set this externally. - * It is set internally in ZSTD_registerSequenceProducer(). */ - int useSequenceProducer; + /* Parameters for the external sequence producer API. + * Users set these parameters through ZSTD_registerSequenceProducer(). + * It is not possible to set these parameters individually through the public API. */ + void* extSeqProdState; + ZSTD_sequenceProducer_F extSeqProdFunc; /* Adjust the max block size*/ size_t maxBlockSize; @@ -401,14 +403,6 @@ typedef struct { ZSTD_entropyCTablesMetadata_t entropyMetadata; } ZSTD_blockSplitCtx; -/* Context for block-level external matchfinder API */ -typedef struct { - void* mState; - ZSTD_sequenceProducer_F* mFinder; - ZSTD_Sequence* seqBuffer; - size_t seqBufferCapacity; -} ZSTD_externalMatchCtx; - struct ZSTD_CCtx_s { ZSTD_compressionStage_e stage; int cParamsChanged; /* == 1 if cParams(except wlog) or compression level are changed in requestedParams. Triggers transmission of new params to ZSTDMT (if available) then reset to 0. */ @@ -479,8 +473,9 @@ struct ZSTD_CCtx_s { /* Workspace for block splitter */ ZSTD_blockSplitCtx blockSplitCtx; - /* Workspace for external matchfinder */ - ZSTD_externalMatchCtx externalMatchCtx; + /* Buffer for output from external sequence producer */ + ZSTD_Sequence* extSeqBuf; + size_t extSeqBufCapacity; }; typedef enum { ZSTD_dtlm_fast, ZSTD_dtlm_full } ZSTD_dictTableLoadMethod_e; @@ -1053,7 +1048,9 @@ MEM_STATIC U32 ZSTD_window_needOverflowCorrection(ZSTD_window_t const window, * The least significant cycleLog bits of the indices must remain the same, * which may be 0. Every index up to maxDist in the past must be valid. */ -MEM_STATIC U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog, +MEM_STATIC +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog, U32 maxDist, void const* src) { /* preemptive overflow correction: @@ -1246,7 +1243,9 @@ MEM_STATIC void ZSTD_window_init(ZSTD_window_t* window) { * forget about the extDict. Handles overlap of the prefix and extDict. * Returns non-zero if the segment is contiguous. */ -MEM_STATIC U32 ZSTD_window_update(ZSTD_window_t* window, +MEM_STATIC +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +U32 ZSTD_window_update(ZSTD_window_t* window, void const* src, size_t srcSize, int forceNonContiguous) { @@ -1467,11 +1466,10 @@ size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity); * This cannot be used when long range matching is enabled. * Zstd will use these sequences, and pass the literals to a secondary block * compressor. - * @return : An error code on failure. * NOTE: seqs are not verified! Invalid sequences can cause out-of-bounds memory * access and data corruption. */ -size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq); +void ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq); /** ZSTD_cycleLog() : * condition for correct operation : hashLog > 1 */ @@ -1509,6 +1507,10 @@ ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* const ZSTD_Sequence* const inSeqs, size_t inSeqsSize, const void* src, size_t blockSize, ZSTD_paramSwitch_e externalRepSearch); +/* Returns 1 if an external sequence producer is registered, otherwise returns 0. */ +MEM_STATIC int ZSTD_hasExtSeqProd(const ZSTD_CCtx_params* params) { + return params->extSeqProdFunc != NULL; +} /* =============================================================== * Deprecated definitions that are still used internally to avoid diff --git a/thirdparty/zstd/compress/zstd_compress_superblock.c b/thirdparty/zstd/compress/zstd_compress_superblock.c index 638c4acbe7..628a2dccd0 100644 --- a/thirdparty/zstd/compress/zstd_compress_superblock.c +++ b/thirdparty/zstd/compress/zstd_compress_superblock.c @@ -76,8 +76,8 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable, } { int const flags = bmi2 ? HUF_flags_bmi2 : 0; - const size_t cSize = singleStream ? HUF_compress1X_usingCTable(op, oend-op, literals, litSize, hufTable, flags) - : HUF_compress4X_usingCTable(op, oend-op, literals, litSize, hufTable, flags); + const size_t cSize = singleStream ? HUF_compress1X_usingCTable(op, (size_t)(oend-op), literals, litSize, hufTable, flags) + : HUF_compress4X_usingCTable(op, (size_t)(oend-op), literals, litSize, hufTable, flags); op += cSize; cLitSize += cSize; if (cSize == 0 || ERR_isError(cSize)) { @@ -102,7 +102,7 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable, switch(lhSize) { case 3: /* 2 - 2 - 10 - 10 */ - { U32 const lhc = hType + ((!singleStream) << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<14); + { U32 const lhc = hType + ((U32)(!singleStream) << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<14); MEM_writeLE24(ostart, lhc); break; } @@ -122,30 +122,30 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable, } *entropyWritten = 1; DEBUGLOG(5, "Compressed literals: %u -> %u", (U32)litSize, (U32)(op-ostart)); - return op-ostart; + return (size_t)(op-ostart); } static size_t ZSTD_seqDecompressedSize(seqStore_t const* seqStore, - const seqDef* sequences, size_t nbSeq, - size_t litSize, int lastSequence) + const seqDef* sequences, size_t nbSeqs, + size_t litSize, int lastSubBlock) { - 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); + size_t n; + for (n=0; n<nbSeqs; n++) { + const ZSTD_sequenceLength seqLen = ZSTD_getSequenceLength(seqStore, sequences+n); litLengthSum += seqLen.litLength; matchLengthSum += seqLen.matchLength; - sp++; } - assert(litLengthSum <= litSize); - if (!lastSequence) { + DEBUGLOG(5, "ZSTD_seqDecompressedSize: %u sequences from %p: %u literals + %u matchlength", + (unsigned)nbSeqs, (const void*)sequences, + (unsigned)litLengthSum, (unsigned)matchLengthSum); + if (!lastSubBlock) assert(litLengthSum == litSize); - } + else + assert(litLengthSum <= litSize); + (void)litLengthSum; return matchLengthSum + litSize; } @@ -180,14 +180,14 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables, /* Sequences Header */ RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/, dstSize_tooSmall, ""); - if (nbSeq < 0x7F) + if (nbSeq < 128) *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; + return (size_t)(op - ostart); } /* seqHead : flags for FSE encoding type */ @@ -209,7 +209,7 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables, } { size_t const bitstreamSize = ZSTD_encodeSequences( - op, oend - op, + op, (size_t)(oend - op), fseTables->matchlengthCTable, mlCode, fseTables->offcodeCTable, ofCode, fseTables->litlengthCTable, llCode, @@ -253,7 +253,7 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables, #endif *entropyWritten = 1; - return op - ostart; + return (size_t)(op - ostart); } /** ZSTD_compressSubBlock() : @@ -279,7 +279,8 @@ static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy, 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); + op, (size_t)(oend-op), + bmi2, writeLitEntropy, litEntropyWritten); FORWARD_IF_ERROR(cLitSize, "ZSTD_compressSubBlock_literal failed"); if (cLitSize == 0) return 0; op += cLitSize; @@ -289,18 +290,18 @@ static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy, sequences, nbSeq, llCode, mlCode, ofCode, cctxParams, - op, oend-op, + op, (size_t)(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; + { size_t cSize = (size_t)(op-ostart) - ZSTD_blockHeaderSize; U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3); MEM_writeLE24(ostart, cBlockHeader24); } - return op-ostart; + return (size_t)(op-ostart); } static size_t ZSTD_estimateSubBlockSize_literal(const BYTE* literals, size_t litSize, @@ -389,7 +390,11 @@ static size_t ZSTD_estimateSubBlockSize_sequences(const BYTE* ofCodeTable, return cSeqSizeEstimate + sequencesSectionHeaderSize; } -static size_t ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize, +typedef struct { + size_t estLitSize; + size_t estBlockSize; +} EstimatedBlockSize; +static EstimatedBlockSize ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize, const BYTE* ofCodeTable, const BYTE* llCodeTable, const BYTE* mlCodeTable, @@ -397,15 +402,17 @@ static size_t ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize, 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, + int writeLitEntropy, int writeSeqEntropy) +{ + EstimatedBlockSize ebs; + ebs.estLitSize = ZSTD_estimateSubBlockSize_literal(literals, litSize, + &entropy->huf, &entropyMetadata->hufMetadata, + workspace, wkspSize, writeLitEntropy); + ebs.estBlockSize = ZSTD_estimateSubBlockSize_sequences(ofCodeTable, llCodeTable, mlCodeTable, nbSeq, &entropy->fse, &entropyMetadata->fseMetadata, workspace, wkspSize, writeSeqEntropy); - return cSizeEstimate + ZSTD_blockHeaderSize; + ebs.estBlockSize += ebs.estLitSize + ZSTD_blockHeaderSize; + return ebs; } static int ZSTD_needSequenceEntropyTables(ZSTD_fseCTablesMetadata_t const* fseMetadata) @@ -419,13 +426,56 @@ static int ZSTD_needSequenceEntropyTables(ZSTD_fseCTablesMetadata_t const* fseMe return 0; } +static size_t countLiterals(seqStore_t const* seqStore, const seqDef* sp, size_t seqCount) +{ + size_t n, total = 0; + assert(sp != NULL); + for (n=0; n<seqCount; n++) { + total += ZSTD_getSequenceLength(seqStore, sp+n).litLength; + } + DEBUGLOG(6, "countLiterals for %zu sequences from %p => %zu bytes", seqCount, (const void*)sp, total); + return total; +} + +#define BYTESCALE 256 + +static size_t sizeBlockSequences(const seqDef* sp, size_t nbSeqs, + size_t targetBudget, size_t avgLitCost, size_t avgSeqCost, + int firstSubBlock) +{ + size_t n, budget = 0, inSize=0; + /* entropy headers */ + size_t const headerSize = (size_t)firstSubBlock * 120 * BYTESCALE; /* generous estimate */ + assert(firstSubBlock==0 || firstSubBlock==1); + budget += headerSize; + + /* first sequence => at least one sequence*/ + budget += sp[0].litLength * avgLitCost + avgSeqCost; + if (budget > targetBudget) return 1; + inSize = sp[0].litLength + (sp[0].mlBase+MINMATCH); + + /* loop over sequences */ + for (n=1; n<nbSeqs; n++) { + size_t currentCost = sp[n].litLength * avgLitCost + avgSeqCost; + budget += currentCost; + inSize += sp[n].litLength + (sp[n].mlBase+MINMATCH); + /* stop when sub-block budget is reached */ + if ( (budget > targetBudget) + /* though continue to expand until the sub-block is deemed compressible */ + && (budget < inSize * BYTESCALE) ) + break; + } + + return n; +} + /** 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. */ + * Entropy will be written into the first block. + * The following blocks use repeat_mode to compress. + * Sub-blocks are all compressed, except the last one when beneficial. + * @return : compressed size of the super block (which features 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, @@ -438,10 +488,12 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr, { const seqDef* const sstart = seqStorePtr->sequencesStart; const seqDef* const send = seqStorePtr->sequences; - const seqDef* sp = sstart; + const seqDef* sp = sstart; /* tracks progresses within seqStorePtr->sequences */ + size_t const nbSeqs = (size_t)(send - sstart); const BYTE* const lstart = seqStorePtr->litStart; const BYTE* const lend = seqStorePtr->lit; const BYTE* lp = lstart; + size_t const nbLiterals = (size_t)(lend - lstart); BYTE const* ip = (BYTE const*)src; BYTE const* const iend = ip + srcSize; BYTE* const ostart = (BYTE*)dst; @@ -450,96 +502,152 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr, 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; + size_t const minTarget = ZSTD_TARGETCBLOCKSIZE_MIN; /* enforce minimum size, to reduce undesirable side effects */ + size_t const targetCBlockSize = MAX(minTarget, cctxParams->targetCBlockSize); + 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); + + DEBUGLOG(5, "ZSTD_compressSubBlock_multi (srcSize=%u, litSize=%u, nbSeq=%u)", + (unsigned)srcSize, (unsigned)(lend-lstart), (unsigned)(send-sstart)); + + /* let's start by a general estimation for the full block */ + if (nbSeqs > 0) { + EstimatedBlockSize const ebs = + ZSTD_estimateSubBlockSize(lp, nbLiterals, + ofCodePtr, llCodePtr, mlCodePtr, nbSeqs, + &nextCBlock->entropy, entropyMetadata, + workspace, wkspSize, + writeLitEntropy, writeSeqEntropy); + /* quick estimation */ + size_t const avgLitCost = nbLiterals ? (ebs.estLitSize * BYTESCALE) / nbLiterals : BYTESCALE; + size_t const avgSeqCost = ((ebs.estBlockSize - ebs.estLitSize) * BYTESCALE) / nbSeqs; + const size_t nbSubBlocks = MAX((ebs.estBlockSize + (targetCBlockSize/2)) / targetCBlockSize, 1); + size_t n, avgBlockBudget, blockBudgetSupp=0; + avgBlockBudget = (ebs.estBlockSize * BYTESCALE) / nbSubBlocks; + DEBUGLOG(5, "estimated fullblock size=%u bytes ; avgLitCost=%.2f ; avgSeqCost=%.2f ; targetCBlockSize=%u, nbSubBlocks=%u ; avgBlockBudget=%.0f bytes", + (unsigned)ebs.estBlockSize, (double)avgLitCost/BYTESCALE, (double)avgSeqCost/BYTESCALE, + (unsigned)targetCBlockSize, (unsigned)nbSubBlocks, (double)avgBlockBudget/BYTESCALE); + /* simplification: if estimates states that the full superblock doesn't compress, just bail out immediately + * this will result in the production of a single uncompressed block covering @srcSize.*/ + if (ebs.estBlockSize > srcSize) return 0; + + /* compress and write sub-blocks */ + assert(nbSubBlocks>0); + for (n=0; n < nbSubBlocks-1; n++) { + /* determine nb of sequences for current sub-block + nbLiterals from next sequence */ + size_t const seqCount = sizeBlockSequences(sp, (size_t)(send-sp), + avgBlockBudget + blockBudgetSupp, avgLitCost, avgSeqCost, n==0); + /* if reached last sequence : break to last sub-block (simplification) */ + assert(seqCount <= (size_t)(send-sp)); + if (sp + seqCount == send) break; + assert(seqCount > 0); + /* compress sub-block */ + { int litEntropyWritten = 0; + int seqEntropyWritten = 0; + size_t litSize = countLiterals(seqStorePtr, sp, seqCount); + const size_t decompressedSize = + ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, 0); + size_t const cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata, + sp, seqCount, + lp, litSize, + llCodePtr, mlCodePtr, ofCodePtr, + cctxParams, + op, (size_t)(oend-op), + bmi2, writeLitEntropy, writeSeqEntropy, + &litEntropyWritten, &seqEntropyWritten, + 0); + FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed"); + + /* check compressibility, update state components */ + if (cSize > 0 && cSize < decompressedSize) { + DEBUGLOG(5, "Committed sub-block compressing %u bytes => %u bytes", + (unsigned)decompressedSize, (unsigned)cSize); + assert(ip + decompressedSize <= iend); + ip += decompressedSize; + lp += litSize; + op += cSize; + llCodePtr += seqCount; + mlCodePtr += seqCount; + ofCodePtr += seqCount; + /* Entropy only needs to be written once */ + if (litEntropyWritten) { + writeLitEntropy = 0; + } + if (seqEntropyWritten) { + writeSeqEntropy = 0; + } + sp += seqCount; + blockBudgetSupp = 0; + } } + /* otherwise : do not compress yet, coalesce current sub-block with following one */ } - /* 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; - } + } /* if (nbSeqs > 0) */ + + /* write last block */ + DEBUGLOG(5, "Generate last sub-block: %u sequences remaining", (unsigned)(send - sp)); + { int litEntropyWritten = 0; + int seqEntropyWritten = 0; + size_t litSize = (size_t)(lend - lp); + size_t seqCount = (size_t)(send - sp); + const size_t decompressedSize = + ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, 1); + size_t const cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata, + sp, seqCount, + lp, litSize, + llCodePtr, mlCodePtr, ofCodePtr, + cctxParams, + op, (size_t)(oend-op), + bmi2, writeLitEntropy, writeSeqEntropy, + &litEntropyWritten, &seqEntropyWritten, + lastBlock); + FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed"); + + /* update pointers, the nb of literals borrowed from next sequence must be preserved */ + if (cSize > 0 && cSize < decompressedSize) { + DEBUGLOG(5, "Last sub-block compressed %u bytes => %u bytes", + (unsigned)decompressedSize, (unsigned)cSize); + assert(ip + decompressedSize <= iend); + ip += decompressedSize; + lp += litSize; + op += cSize; + llCodePtr += seqCount; + mlCodePtr += seqCount; + ofCodePtr += seqCount; + /* Entropy only needs to be written once */ + if (litEntropyWritten) { + writeLitEntropy = 0; + } + if (seqEntropyWritten) { + writeSeqEntropy = 0; } + sp += seqCount; } - } while (!lastSequence); + } + + if (writeLitEntropy) { - DEBUGLOG(5, "ZSTD_compressSubBlock_multi has literal entropy tables unwritten"); + DEBUGLOG(5, "Literal entropy tables were never written"); 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"); + DEBUGLOG(5, "Sequence entropy tables were never written => cancel, emit an uncompressed block"); 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)); + /* some data left : last part of the block sent uncompressed */ + size_t const rSize = (size_t)((iend - ip)); + size_t const cSize = ZSTD_noCompressBlock(op, (size_t)(oend - op), ip, rSize, lastBlock); + DEBUGLOG(5, "Generate last uncompressed sub-block of %u bytes", (unsigned)(rSize)); 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; + const seqDef* seq; repcodes_t rep; ZSTD_memcpy(&rep, prevCBlock->rep, sizeof(rep)); for (seq = sstart; seq < sp; ++seq) { @@ -548,14 +656,17 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr, ZSTD_memcpy(nextCBlock->rep, &rep, sizeof(rep)); } } - DEBUGLOG(5, "ZSTD_compressSubBlock_multi compressed"); - return op-ostart; + + DEBUGLOG(5, "ZSTD_compressSubBlock_multi compressed all subBlocks: total compressed size = %u", + (unsigned)(op-ostart)); + return (size_t)(op-ostart); } size_t ZSTD_compressSuperBlock(ZSTD_CCtx* zc, void* dst, size_t dstCapacity, - void const* src, size_t srcSize, - unsigned lastBlock) { + const void* src, size_t srcSize, + unsigned lastBlock) +{ ZSTD_entropyCTablesMetadata_t entropyMetadata; FORWARD_IF_ERROR(ZSTD_buildBlockEntropyStats(&zc->seqStore, diff --git a/thirdparty/zstd/compress/zstd_cwksp.h b/thirdparty/zstd/compress/zstd_cwksp.h index cc7fb1c715..3eddbd334e 100644 --- a/thirdparty/zstd/compress/zstd_cwksp.h +++ b/thirdparty/zstd/compress/zstd_cwksp.h @@ -192,6 +192,7 @@ MEM_STATIC void ZSTD_cwksp_assert_internal_consistency(ZSTD_cwksp* ws) { { intptr_t const offset = __msan_test_shadow(ws->initOnceStart, (U8*)ZSTD_cwksp_initialAllocStart(ws) - (U8*)ws->initOnceStart); + (void)offset; #if defined(ZSTD_MSAN_PRINT) if(offset!=-1) { __msan_print_shadow((U8*)ws->initOnceStart + offset - 8, 32); @@ -433,7 +434,7 @@ MEM_STATIC void* ZSTD_cwksp_reserve_aligned(ZSTD_cwksp* ws, size_t bytes) /** * Aligned on 64 bytes. These buffers have the special property that - * their values remain constrained, allowing us to re-use them without + * their values remain constrained, allowing us to reuse them without * memset()-ing them. */ MEM_STATIC void* ZSTD_cwksp_reserve_table(ZSTD_cwksp* ws, size_t bytes) @@ -525,7 +526,7 @@ MEM_STATIC void ZSTD_cwksp_mark_tables_dirty(ZSTD_cwksp* ws) DEBUGLOG(4, "cwksp: ZSTD_cwksp_mark_tables_dirty"); #if ZSTD_MEMORY_SANITIZER && !defined (ZSTD_MSAN_DONT_POISON_WORKSPACE) - /* To validate that the table re-use logic is sound, and that we don't + /* To validate that the table reuse logic is sound, and that we don't * access table space that we haven't cleaned, we re-"poison" the table * space every time we mark it dirty. * Since tableValidEnd space and initOnce space may overlap we don't poison @@ -602,9 +603,9 @@ MEM_STATIC void ZSTD_cwksp_clear(ZSTD_cwksp* ws) { DEBUGLOG(4, "cwksp: clearing!"); #if ZSTD_MEMORY_SANITIZER && !defined (ZSTD_MSAN_DONT_POISON_WORKSPACE) - /* To validate that the context re-use logic is sound, and that we don't + /* To validate that the context reuse logic is sound, and that we don't * access stuff that this compression hasn't initialized, we re-"poison" - * the workspace except for the areas in which we expect memory re-use + * the workspace except for the areas in which we expect memory reuse * without initialization (objects, valid tables area and init once * memory). */ { @@ -635,6 +636,15 @@ MEM_STATIC void ZSTD_cwksp_clear(ZSTD_cwksp* ws) { ZSTD_cwksp_assert_internal_consistency(ws); } +MEM_STATIC size_t ZSTD_cwksp_sizeof(const ZSTD_cwksp* ws) { + return (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->workspace); +} + +MEM_STATIC size_t ZSTD_cwksp_used(const ZSTD_cwksp* ws) { + return (size_t)((BYTE*)ws->tableEnd - (BYTE*)ws->workspace) + + (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->allocStart); +} + /** * The provided workspace takes ownership of the buffer [start, start+size). * Any existing values in the workspace are ignored (the previously managed @@ -666,6 +676,11 @@ MEM_STATIC size_t ZSTD_cwksp_create(ZSTD_cwksp* ws, size_t size, ZSTD_customMem MEM_STATIC void ZSTD_cwksp_free(ZSTD_cwksp* ws, ZSTD_customMem customMem) { void *ptr = ws->workspace; DEBUGLOG(4, "cwksp: freeing workspace"); +#if ZSTD_MEMORY_SANITIZER && !defined(ZSTD_MSAN_DONT_POISON_WORKSPACE) + if (ptr != NULL && customMem.customFree != NULL) { + __msan_unpoison(ptr, ZSTD_cwksp_sizeof(ws)); + } +#endif ZSTD_memset(ws, 0, sizeof(ZSTD_cwksp)); ZSTD_customFree(ptr, customMem); } @@ -679,15 +694,6 @@ MEM_STATIC void ZSTD_cwksp_move(ZSTD_cwksp* dst, ZSTD_cwksp* src) { ZSTD_memset(src, 0, sizeof(ZSTD_cwksp)); } -MEM_STATIC size_t ZSTD_cwksp_sizeof(const ZSTD_cwksp* ws) { - return (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->workspace); -} - -MEM_STATIC size_t ZSTD_cwksp_used(const ZSTD_cwksp* ws) { - return (size_t)((BYTE*)ws->tableEnd - (BYTE*)ws->workspace) - + (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->allocStart); -} - MEM_STATIC int ZSTD_cwksp_reserve_failed(const ZSTD_cwksp* ws) { return ws->allocFailed; } diff --git a/thirdparty/zstd/compress/zstd_double_fast.c b/thirdparty/zstd/compress/zstd_double_fast.c index 0ad88ffc7b..a4e9c50d3b 100644 --- a/thirdparty/zstd/compress/zstd_double_fast.c +++ b/thirdparty/zstd/compress/zstd_double_fast.c @@ -11,7 +11,11 @@ #include "zstd_compress_internal.h" #include "zstd_double_fast.h" -static void ZSTD_fillDoubleHashTableForCDict(ZSTD_matchState_t* ms, +#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR + +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_fillDoubleHashTableForCDict(ZSTD_matchState_t* ms, void const* end, ZSTD_dictTableLoadMethod_e dtlm) { const ZSTD_compressionParameters* const cParams = &ms->cParams; @@ -47,7 +51,9 @@ static void ZSTD_fillDoubleHashTableForCDict(ZSTD_matchState_t* ms, } } } -static void ZSTD_fillDoubleHashTableForCCtx(ZSTD_matchState_t* ms, +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_fillDoubleHashTableForCCtx(ZSTD_matchState_t* ms, void const* end, ZSTD_dictTableLoadMethod_e dtlm) { const ZSTD_compressionParameters* const cParams = &ms->cParams; @@ -95,6 +101,7 @@ void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms, FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR size_t ZSTD_compressBlock_doubleFast_noDict_generic( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize, U32 const mls /* template */) @@ -305,6 +312,7 @@ _match_stored: FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize, @@ -348,8 +356,8 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic( if (ms->prefetchCDictTables) { size_t const hashTableBytes = (((size_t)1) << dictCParams->hashLog) * sizeof(U32); size_t const chainTableBytes = (((size_t)1) << dictCParams->chainLog) * sizeof(U32); - PREFETCH_AREA(dictHashLong, hashTableBytes) - PREFETCH_AREA(dictHashSmall, chainTableBytes) + PREFETCH_AREA(dictHashLong, hashTableBytes); + PREFETCH_AREA(dictHashSmall, chainTableBytes); } /* init */ @@ -589,7 +597,9 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState( } -static size_t ZSTD_compressBlock_doubleFast_extDict_generic( +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +size_t ZSTD_compressBlock_doubleFast_extDict_generic( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize, U32 const mls /* template */) @@ -756,3 +766,5 @@ size_t ZSTD_compressBlock_doubleFast_extDict( return ZSTD_compressBlock_doubleFast_extDict_7(ms, seqStore, rep, src, srcSize); } } + +#endif /* ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR */ diff --git a/thirdparty/zstd/compress/zstd_double_fast.h b/thirdparty/zstd/compress/zstd_double_fast.h index 6f0047c4ba..ce6ed8c97f 100644 --- a/thirdparty/zstd/compress/zstd_double_fast.h +++ b/thirdparty/zstd/compress/zstd_double_fast.h @@ -18,9 +18,12 @@ extern "C" { #include "../common/mem.h" /* U32 */ #include "zstd_compress_internal.h" /* ZSTD_CCtx, size_t */ +#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR + void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms, void const* end, ZSTD_dictTableLoadMethod_e dtlm, ZSTD_tableFillPurpose_e tfp); + size_t ZSTD_compressBlock_doubleFast( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); @@ -31,6 +34,14 @@ size_t ZSTD_compressBlock_doubleFast_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); +#define ZSTD_COMPRESSBLOCK_DOUBLEFAST ZSTD_compressBlock_doubleFast +#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_DICTMATCHSTATE ZSTD_compressBlock_doubleFast_dictMatchState +#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_EXTDICT ZSTD_compressBlock_doubleFast_extDict +#else +#define ZSTD_COMPRESSBLOCK_DOUBLEFAST NULL +#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_DICTMATCHSTATE NULL +#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_EXTDICT NULL +#endif /* ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR */ #if defined (__cplusplus) } diff --git a/thirdparty/zstd/compress/zstd_fast.c b/thirdparty/zstd/compress/zstd_fast.c index 5f2c6a2eda..6c4554cfca 100644 --- a/thirdparty/zstd/compress/zstd_fast.c +++ b/thirdparty/zstd/compress/zstd_fast.c @@ -11,7 +11,9 @@ #include "zstd_compress_internal.h" /* ZSTD_hashPtr, ZSTD_count, ZSTD_storeSeq */ #include "zstd_fast.h" -static void ZSTD_fillHashTableForCDict(ZSTD_matchState_t* ms, +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_fillHashTableForCDict(ZSTD_matchState_t* ms, const void* const end, ZSTD_dictTableLoadMethod_e dtlm) { @@ -46,7 +48,9 @@ static void ZSTD_fillHashTableForCDict(ZSTD_matchState_t* ms, } } } } } -static void ZSTD_fillHashTableForCCtx(ZSTD_matchState_t* ms, +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_fillHashTableForCCtx(ZSTD_matchState_t* ms, const void* const end, ZSTD_dictTableLoadMethod_e dtlm) { @@ -139,8 +143,9 @@ void ZSTD_fillHashTable(ZSTD_matchState_t* ms, * * This is also the work we do at the beginning to enter the loop initially. */ -FORCE_INLINE_TEMPLATE size_t -ZSTD_compressBlock_fast_noDict_generic( +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +size_t ZSTD_compressBlock_fast_noDict_generic( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize, U32 const mls, U32 const hasStep) @@ -456,6 +461,7 @@ size_t ZSTD_compressBlock_fast( } FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR size_t ZSTD_compressBlock_fast_dictMatchState_generic( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize, U32 const mls, U32 const hasStep) @@ -502,7 +508,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic( if (ms->prefetchCDictTables) { size_t const hashTableBytes = (((size_t)1) << dictCParams->hashLog) * sizeof(U32); - PREFETCH_AREA(dictHashTable, hashTableBytes) + PREFETCH_AREA(dictHashTable, hashTableBytes); } /* init */ @@ -681,7 +687,9 @@ size_t ZSTD_compressBlock_fast_dictMatchState( } -static size_t ZSTD_compressBlock_fast_extDict_generic( +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +size_t ZSTD_compressBlock_fast_extDict_generic( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize, U32 const mls, U32 const hasStep) { diff --git a/thirdparty/zstd/compress/zstd_lazy.c b/thirdparty/zstd/compress/zstd_lazy.c index 5ba88e8678..67dd55fdb8 100644 --- a/thirdparty/zstd/compress/zstd_lazy.c +++ b/thirdparty/zstd/compress/zstd_lazy.c @@ -12,6 +12,11 @@ #include "zstd_lazy.h" #include "../common/bits.h" /* ZSTD_countTrailingZeros64 */ +#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) + #define kLazySkippingStep 8 @@ -19,8 +24,9 @@ * Binary Tree search ***************************************/ -static void -ZSTD_updateDUBT(ZSTD_matchState_t* ms, +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_updateDUBT(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend, U32 mls) { @@ -63,8 +69,9 @@ ZSTD_updateDUBT(ZSTD_matchState_t* ms, * sort one already inserted but unsorted position * assumption : curr >= btlow == (curr - btmask) * doesn't fail */ -static void -ZSTD_insertDUBT1(const ZSTD_matchState_t* ms, +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_insertDUBT1(const ZSTD_matchState_t* ms, U32 curr, const BYTE* inputEnd, U32 nbCompares, U32 btLow, const ZSTD_dictMode_e dictMode) @@ -152,8 +159,9 @@ ZSTD_insertDUBT1(const ZSTD_matchState_t* ms, } -static size_t -ZSTD_DUBT_findBetterDictMatch ( +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +size_t ZSTD_DUBT_findBetterDictMatch ( const ZSTD_matchState_t* ms, const BYTE* const ip, const BYTE* const iend, size_t* offsetPtr, @@ -230,8 +238,9 @@ ZSTD_DUBT_findBetterDictMatch ( } -static size_t -ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms, +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +size_t ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms, const BYTE* const ip, const BYTE* const iend, size_t* offBasePtr, U32 const mls, @@ -381,8 +390,9 @@ ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms, /** ZSTD_BtFindBestMatch() : Tree updater, providing best match */ -FORCE_INLINE_TEMPLATE size_t -ZSTD_BtFindBestMatch( ZSTD_matchState_t* ms, +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +size_t ZSTD_BtFindBestMatch( ZSTD_matchState_t* ms, const BYTE* const ip, const BYTE* const iLimit, size_t* offBasePtr, const U32 mls /* template */, @@ -617,7 +627,9 @@ size_t ZSTD_dedicatedDictSearch_lazy_search(size_t* offsetPtr, size_t ml, U32 nb /* Update chains up to ip (excluded) Assumption : always within prefix (i.e. not within extDict) */ -FORCE_INLINE_TEMPLATE U32 ZSTD_insertAndFindFirstIndex_internal( +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +U32 ZSTD_insertAndFindFirstIndex_internal( ZSTD_matchState_t* ms, const ZSTD_compressionParameters* const cParams, const BYTE* ip, U32 const mls, U32 const lazySkipping) @@ -651,6 +663,7 @@ U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip) { /* inlining is important to hardwire a hot branch (template emulation) */ FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR size_t ZSTD_HcFindBestMatch( ZSTD_matchState_t* ms, const BYTE* const ip, const BYTE* const iLimit, @@ -819,7 +832,9 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_prefetch(U32 const* hashTable, BYTE const* t * Fill up the hash cache starting at idx, prefetching up to ZSTD_ROW_HASH_CACHE_SIZE entries, * but not beyond iLimit. */ -FORCE_INLINE_TEMPLATE void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const BYTE* base, +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const BYTE* base, U32 const rowLog, U32 const mls, U32 idx, const BYTE* const iLimit) { @@ -845,7 +860,9 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const B * Returns the hash of base + idx, and replaces the hash in the hash cache with the byte at * base + idx + ZSTD_ROW_HASH_CACHE_SIZE. Also prefetches the appropriate rows from hashTable and tagTable. */ -FORCE_INLINE_TEMPLATE U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTable, +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTable, BYTE const* tagTable, BYTE const* base, U32 idx, U32 const hashLog, U32 const rowLog, U32 const mls, @@ -863,10 +880,12 @@ FORCE_INLINE_TEMPLATE U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTab /* ZSTD_row_update_internalImpl(): * Updates the hash table with positions starting from updateStartIdx until updateEndIdx. */ -FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms, - U32 updateStartIdx, U32 const updateEndIdx, - U32 const mls, U32 const rowLog, - U32 const rowMask, U32 const useCache) +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms, + U32 updateStartIdx, U32 const updateEndIdx, + U32 const mls, U32 const rowLog, + U32 const rowMask, U32 const useCache) { U32* const hashTable = ms->hashTable; BYTE* const tagTable = ms->tagTable; @@ -892,9 +911,11 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms, * Inserts the byte at ip into the appropriate position in the hash table, and updates ms->nextToUpdate. * Skips sections of long matches as is necessary. */ -FORCE_INLINE_TEMPLATE void ZSTD_row_update_internal(ZSTD_matchState_t* ms, const BYTE* ip, - U32 const mls, U32 const rowLog, - U32 const rowMask, U32 const useCache) +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_row_update_internal(ZSTD_matchState_t* ms, const BYTE* ip, + U32 const mls, U32 const rowLog, + U32 const rowMask, U32 const useCache) { U32 idx = ms->nextToUpdate; const BYTE* const base = ms->window.base; @@ -1102,20 +1123,21 @@ ZSTD_row_getMatchMask(const BYTE* const tagRow, const BYTE tag, const U32 headGr /* The high-level approach of the SIMD row based match finder is as follows: * - Figure out where to insert the new entry: - * - Generate a hash from a byte along with an additional 1-byte "short hash". The additional byte is our "tag" - * - The hashTable is effectively split into groups or "rows" of 16 or 32 entries of U32, and the hash determines + * - Generate a hash for current input posistion and split it into a one byte of tag and `rowHashLog` bits of index. + * - The hash is salted by a value that changes on every contex reset, so when the same table is used + * we will avoid collisions that would otherwise slow us down by intorducing phantom matches. + * - The hashTable is effectively split into groups or "rows" of 15 or 31 entries of U32, and the index determines * which row to insert into. - * - Determine the correct position within the row to insert the entry into. Each row of 16 or 32 can - * be considered as a circular buffer with a "head" index that resides in the tagTable. - * - Also insert the "tag" into the equivalent row and position in the tagTable. - * - Note: The tagTable has 17 or 33 1-byte entries per row, due to 16 or 32 tags, and 1 "head" entry. - * The 17 or 33 entry rows are spaced out to occur every 32 or 64 bytes, respectively, - * for alignment/performance reasons, leaving some bytes unused. - * - Use SIMD to efficiently compare the tags in the tagTable to the 1-byte "short hash" and + * - Determine the correct position within the row to insert the entry into. Each row of 15 or 31 can + * be considered as a circular buffer with a "head" index that resides in the tagTable (overall 16 or 32 bytes + * per row). + * - Use SIMD to efficiently compare the tags in the tagTable to the 1-byte tag calculated for the position and * generate a bitfield that we can cycle through to check the collisions in the hash table. * - Pick the longest match. + * - Insert the tag into the equivalent row and position in the tagTable. */ FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR size_t ZSTD_RowFindBestMatch( ZSTD_matchState_t* ms, const BYTE* const ip, const BYTE* const iLimit, @@ -1489,8 +1511,9 @@ FORCE_INLINE_TEMPLATE size_t ZSTD_searchMax( * Common parser - lazy strategy *********************************/ -FORCE_INLINE_TEMPLATE size_t -ZSTD_compressBlock_lazy_generic( +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +size_t ZSTD_compressBlock_lazy_generic( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], const void* src, size_t srcSize, @@ -1754,152 +1777,163 @@ _storeSequence: /* Return the last literals size */ return (size_t)(iend - anchor); } +#endif /* build exclusions */ -size_t ZSTD_compressBlock_btlazy2( +#ifndef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_greedy( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_noDict); } -size_t ZSTD_compressBlock_lazy2( +size_t ZSTD_compressBlock_greedy_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dictMatchState); } -size_t ZSTD_compressBlock_lazy( +size_t ZSTD_compressBlock_greedy_dedicatedDictSearch( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dedicatedDictSearch); } -size_t ZSTD_compressBlock_greedy( +size_t ZSTD_compressBlock_greedy_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_noDict); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_noDict); } -size_t ZSTD_compressBlock_btlazy2_dictMatchState( +size_t ZSTD_compressBlock_greedy_dictMatchState_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dictMatchState); } -size_t ZSTD_compressBlock_lazy2_dictMatchState( +size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dedicatedDictSearch); } +#endif -size_t ZSTD_compressBlock_lazy_dictMatchState( +#ifndef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_lazy( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict); } -size_t ZSTD_compressBlock_greedy_dictMatchState( +size_t ZSTD_compressBlock_lazy_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dictMatchState); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState); } - -size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch( +size_t ZSTD_compressBlock_lazy_dedicatedDictSearch( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dedicatedDictSearch); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dedicatedDictSearch); } -size_t ZSTD_compressBlock_lazy_dedicatedDictSearch( +size_t ZSTD_compressBlock_lazy_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dedicatedDictSearch); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_noDict); } -size_t ZSTD_compressBlock_greedy_dedicatedDictSearch( +size_t ZSTD_compressBlock_lazy_dictMatchState_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dedicatedDictSearch); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dictMatchState); } -/* Row-based matchfinder */ -size_t ZSTD_compressBlock_lazy2_row( +size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_noDict); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dedicatedDictSearch); } +#endif -size_t ZSTD_compressBlock_lazy_row( +#ifndef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_lazy2( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_noDict); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict); } -size_t ZSTD_compressBlock_greedy_row( +size_t ZSTD_compressBlock_lazy2_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_noDict); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState); } -size_t ZSTD_compressBlock_lazy2_dictMatchState_row( +size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dictMatchState); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dedicatedDictSearch); } -size_t ZSTD_compressBlock_lazy_dictMatchState_row( +size_t ZSTD_compressBlock_lazy2_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dictMatchState); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_noDict); } -size_t ZSTD_compressBlock_greedy_dictMatchState_row( +size_t ZSTD_compressBlock_lazy2_dictMatchState_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dictMatchState); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dictMatchState); } - size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dedicatedDictSearch); } +#endif -size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row( +#ifndef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_btlazy2( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dedicatedDictSearch); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict); } -size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row( +size_t ZSTD_compressBlock_btlazy2_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dedicatedDictSearch); + return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState); } +#endif +#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR size_t ZSTD_compressBlock_lazy_extDict_generic( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], @@ -2101,8 +2135,9 @@ _storeSequence: /* Return the last literals size */ return (size_t)(iend - anchor); } +#endif /* build exclusions */ - +#ifndef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR size_t ZSTD_compressBlock_greedy_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) @@ -2110,48 +2145,55 @@ size_t ZSTD_compressBlock_greedy_extDict( return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0); } -size_t ZSTD_compressBlock_lazy_extDict( +size_t ZSTD_compressBlock_greedy_extDict_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) - { - return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1); + return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0); } +#endif -size_t ZSTD_compressBlock_lazy2_extDict( +#ifndef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_lazy_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2); + return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1); } -size_t ZSTD_compressBlock_btlazy2_extDict( +size_t ZSTD_compressBlock_lazy_extDict_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) { - return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2); + return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1); } +#endif -size_t ZSTD_compressBlock_greedy_extDict_row( +#ifndef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_lazy2_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) + { - return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0); + return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2); } -size_t ZSTD_compressBlock_lazy_extDict_row( +size_t ZSTD_compressBlock_lazy2_extDict_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) - { - return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1); + return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2); } +#endif -size_t ZSTD_compressBlock_lazy2_extDict_row( +#ifndef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_btlazy2_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize) + { - return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2); + return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2); } +#endif diff --git a/thirdparty/zstd/compress/zstd_lazy.h b/thirdparty/zstd/compress/zstd_lazy.h index 3bde67331e..3635813bdd 100644 --- a/thirdparty/zstd/compress/zstd_lazy.h +++ b/thirdparty/zstd/compress/zstd_lazy.h @@ -27,98 +27,173 @@ extern "C" { #define ZSTD_ROW_HASH_TAG_BITS 8 /* nb bits to use for the tag */ +#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip); void ZSTD_row_update(ZSTD_matchState_t* const ms, const BYTE* ip); void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_matchState_t* ms, const BYTE* const ip); void ZSTD_preserveUnsortedMark (U32* const table, U32 const size, U32 const reducerValue); /*! used in ZSTD_reduceIndex(). preemptively increase value of ZSTD_DUBT_UNSORTED_MARK */ +#endif -size_t ZSTD_compressBlock_btlazy2( +#ifndef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_greedy( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy2( +size_t ZSTD_compressBlock_greedy_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy( +size_t ZSTD_compressBlock_greedy_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_greedy( +size_t ZSTD_compressBlock_greedy_dictMatchState_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy2_row( +size_t ZSTD_compressBlock_greedy_dedicatedDictSearch( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy_row( +size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_greedy_row( +size_t ZSTD_compressBlock_greedy_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); - -size_t ZSTD_compressBlock_btlazy2_dictMatchState( +size_t ZSTD_compressBlock_greedy_extDict_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy2_dictMatchState( + +#define ZSTD_COMPRESSBLOCK_GREEDY ZSTD_compressBlock_greedy +#define ZSTD_COMPRESSBLOCK_GREEDY_ROW ZSTD_compressBlock_greedy_row +#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE ZSTD_compressBlock_greedy_dictMatchState +#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE_ROW ZSTD_compressBlock_greedy_dictMatchState_row +#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH ZSTD_compressBlock_greedy_dedicatedDictSearch +#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH_ROW ZSTD_compressBlock_greedy_dedicatedDictSearch_row +#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT ZSTD_compressBlock_greedy_extDict +#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT_ROW ZSTD_compressBlock_greedy_extDict_row +#else +#define ZSTD_COMPRESSBLOCK_GREEDY NULL +#define ZSTD_COMPRESSBLOCK_GREEDY_ROW NULL +#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE NULL +#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE_ROW NULL +#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH NULL +#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH_ROW NULL +#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT NULL +#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT_ROW NULL +#endif + +#ifndef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_lazy( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy_dictMatchState( +size_t ZSTD_compressBlock_lazy_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_greedy_dictMatchState( +size_t ZSTD_compressBlock_lazy_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy2_dictMatchState_row( +size_t ZSTD_compressBlock_lazy_dictMatchState_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy_dictMatchState_row( +size_t ZSTD_compressBlock_lazy_dedicatedDictSearch( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_greedy_dictMatchState_row( +size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); - -size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch( +size_t ZSTD_compressBlock_lazy_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy_dedicatedDictSearch( +size_t ZSTD_compressBlock_lazy_extDict_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_greedy_dedicatedDictSearch( + +#define ZSTD_COMPRESSBLOCK_LAZY ZSTD_compressBlock_lazy +#define ZSTD_COMPRESSBLOCK_LAZY_ROW ZSTD_compressBlock_lazy_row +#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE ZSTD_compressBlock_lazy_dictMatchState +#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE_ROW ZSTD_compressBlock_lazy_dictMatchState_row +#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH ZSTD_compressBlock_lazy_dedicatedDictSearch +#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH_ROW ZSTD_compressBlock_lazy_dedicatedDictSearch_row +#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT ZSTD_compressBlock_lazy_extDict +#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT_ROW ZSTD_compressBlock_lazy_extDict_row +#else +#define ZSTD_COMPRESSBLOCK_LAZY NULL +#define ZSTD_COMPRESSBLOCK_LAZY_ROW NULL +#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE NULL +#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE_ROW NULL +#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH NULL +#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH_ROW NULL +#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT NULL +#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT_ROW NULL +#endif + +#ifndef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_lazy2( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row( +size_t ZSTD_compressBlock_lazy2_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row( +size_t ZSTD_compressBlock_lazy2_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row( +size_t ZSTD_compressBlock_lazy2_dictMatchState_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); - -size_t ZSTD_compressBlock_greedy_extDict( +size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy_extDict( +size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); size_t ZSTD_compressBlock_lazy2_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_greedy_extDict_row( +size_t ZSTD_compressBlock_lazy2_extDict_row( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy_extDict_row( + +#define ZSTD_COMPRESSBLOCK_LAZY2 ZSTD_compressBlock_lazy2 +#define ZSTD_COMPRESSBLOCK_LAZY2_ROW ZSTD_compressBlock_lazy2_row +#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE ZSTD_compressBlock_lazy2_dictMatchState +#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE_ROW ZSTD_compressBlock_lazy2_dictMatchState_row +#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH ZSTD_compressBlock_lazy2_dedicatedDictSearch +#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH_ROW ZSTD_compressBlock_lazy2_dedicatedDictSearch_row +#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT ZSTD_compressBlock_lazy2_extDict +#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT_ROW ZSTD_compressBlock_lazy2_extDict_row +#else +#define ZSTD_COMPRESSBLOCK_LAZY2 NULL +#define ZSTD_COMPRESSBLOCK_LAZY2_ROW NULL +#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE NULL +#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE_ROW NULL +#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH NULL +#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH_ROW NULL +#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT NULL +#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT_ROW NULL +#endif + +#ifndef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_btlazy2( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_lazy2_extDict_row( +size_t ZSTD_compressBlock_btlazy2_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); size_t ZSTD_compressBlock_btlazy2_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); +#define ZSTD_COMPRESSBLOCK_BTLAZY2 ZSTD_compressBlock_btlazy2 +#define ZSTD_COMPRESSBLOCK_BTLAZY2_DICTMATCHSTATE ZSTD_compressBlock_btlazy2_dictMatchState +#define ZSTD_COMPRESSBLOCK_BTLAZY2_EXTDICT ZSTD_compressBlock_btlazy2_extDict +#else +#define ZSTD_COMPRESSBLOCK_BTLAZY2 NULL +#define ZSTD_COMPRESSBLOCK_BTLAZY2_DICTMATCHSTATE NULL +#define ZSTD_COMPRESSBLOCK_BTLAZY2_EXTDICT NULL +#endif + #if defined (__cplusplus) } diff --git a/thirdparty/zstd/compress/zstd_ldm.c b/thirdparty/zstd/compress/zstd_ldm.c index 3d74ff19e3..17c069fe1d 100644 --- a/thirdparty/zstd/compress/zstd_ldm.c +++ b/thirdparty/zstd/compress/zstd_ldm.c @@ -246,7 +246,11 @@ static size_t ZSTD_ldm_fillFastTables(ZSTD_matchState_t* ms, break; case ZSTD_dfast: +#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR ZSTD_fillDoubleHashTable(ms, iend, ZSTD_dtlm_fast, ZSTD_tfp_forCCtx); +#else + assert(0); /* shouldn't be called: cparams should've been adjusted. */ +#endif break; case ZSTD_greedy: @@ -318,7 +322,9 @@ static void ZSTD_ldm_limitTableUpdate(ZSTD_matchState_t* ms, const BYTE* anchor) } } -static size_t ZSTD_ldm_generateSequences_internal( +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +size_t ZSTD_ldm_generateSequences_internal( ldmState_t* ldmState, rawSeqStore_t* rawSeqStore, ldmParams_t const* params, void const* src, size_t srcSize) { @@ -689,7 +695,6 @@ size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore, /* maybeSplitSequence updates rawSeqStore->pos */ rawSeq const sequence = maybeSplitSequence(rawSeqStore, (U32)(iend - ip), minMatch); - int i; /* End signal */ if (sequence.offset == 0) break; @@ -702,6 +707,7 @@ size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore, /* Run the block compressor */ DEBUGLOG(5, "pos %u : calling block compressor on segment of size %u", (unsigned)(ip-istart), sequence.litLength); { + int i; size_t const newLitLength = blockCompressor(ms, seqStore, rep, ip, sequence.litLength); ip += sequence.litLength; diff --git a/thirdparty/zstd/compress/zstd_opt.c b/thirdparty/zstd/compress/zstd_opt.c index f02a760946..e63073e5a4 100644 --- a/thirdparty/zstd/compress/zstd_opt.c +++ b/thirdparty/zstd/compress/zstd_opt.c @@ -12,6 +12,9 @@ #include "hist.h" #include "zstd_opt.h" +#if !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR) #define ZSTD_LITFREQ_ADD 2 /* scaling factor for litFreq, so that frequencies adapt faster to new stats */ #define ZSTD_MAX_PRICE (1<<30) @@ -264,6 +267,7 @@ static U32 ZSTD_rawLiteralsCost(const BYTE* const literals, U32 const litLength, const optState_t* const optPtr, int optLevel) { + DEBUGLOG(8, "ZSTD_rawLiteralsCost (%u literals)", litLength); if (litLength == 0) return 0; if (!ZSTD_compressedLiterals(optPtr)) @@ -402,9 +406,11 @@ MEM_STATIC U32 ZSTD_readMINMATCH(const void* memPtr, U32 length) /* Update hashTable3 up to ip (excluded) Assumption : always within prefix (i.e. not within extDict) */ -static U32 ZSTD_insertAndFindFirstIndexHash3 (const ZSTD_matchState_t* ms, - U32* nextToUpdate3, - const BYTE* const ip) +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +U32 ZSTD_insertAndFindFirstIndexHash3 (const ZSTD_matchState_t* ms, + U32* nextToUpdate3, + const BYTE* const ip) { U32* const hashTable3 = ms->hashTable3; U32 const hashLog3 = ms->hashLog3; @@ -431,7 +437,9 @@ static U32 ZSTD_insertAndFindFirstIndexHash3 (const ZSTD_matchState_t* ms, * @param ip assumed <= iend-8 . * @param target The target of ZSTD_updateTree_internal() - we are filling to this position * @return : nb of positions added */ -static U32 ZSTD_insertBt1( +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +U32 ZSTD_insertBt1( const ZSTD_matchState_t* ms, const BYTE* const ip, const BYTE* const iend, U32 const target, @@ -550,6 +558,7 @@ static U32 ZSTD_insertBt1( } FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR void ZSTD_updateTree_internal( ZSTD_matchState_t* ms, const BYTE* const ip, const BYTE* const iend, @@ -558,7 +567,7 @@ void ZSTD_updateTree_internal( const BYTE* const base = ms->window.base; U32 const target = (U32)(ip - base); U32 idx = ms->nextToUpdate; - DEBUGLOG(6, "ZSTD_updateTree_internal, from %u to %u (dictMode:%u)", + DEBUGLOG(7, "ZSTD_updateTree_internal, from %u to %u (dictMode:%u)", idx, target, dictMode); while(idx < target) { @@ -575,7 +584,9 @@ void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend) { ZSTD_updateTree_internal(ms, ip, iend, ms->cParams.minMatch, ZSTD_noDict); } -FORCE_INLINE_TEMPLATE U32 +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +U32 ZSTD_insertBtAndGetAllMatches ( ZSTD_match_t* matches, /* store result (found matches) in this table (presumed large enough) */ ZSTD_matchState_t* ms, @@ -816,7 +827,9 @@ typedef U32 (*ZSTD_getAllMatchesFn)( U32 const ll0, U32 const lengthToBeat); -FORCE_INLINE_TEMPLATE U32 ZSTD_btGetAllMatches_internal( +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +U32 ZSTD_btGetAllMatches_internal( ZSTD_match_t* matches, ZSTD_matchState_t* ms, U32* nextToUpdate3, @@ -1035,11 +1048,6 @@ ZSTD_optLdm_processMatchCandidate(ZSTD_optLdm_t* optLdm, * Optimal parser *********************************/ -static U32 ZSTD_totalLen(ZSTD_optimal_t sol) -{ - return sol.litlen + sol.mlen; -} - #if 0 /* debug */ static void @@ -1057,7 +1065,13 @@ listStats(const U32* table, int lastEltID) #endif -FORCE_INLINE_TEMPLATE size_t +#define LIT_PRICE(_p) (int)ZSTD_rawLiteralsCost(_p, 1, optStatePtr, optLevel) +#define LL_PRICE(_l) (int)ZSTD_litLengthPrice(_l, optStatePtr, optLevel) +#define LL_INCPRICE(_l) (LL_PRICE(_l) - LL_PRICE(_l-1)) + +FORCE_INLINE_TEMPLATE +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +size_t ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], @@ -1083,10 +1097,10 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms, ZSTD_optimal_t* const opt = optStatePtr->priceTable; ZSTD_match_t* const matches = optStatePtr->matchTable; - ZSTD_optimal_t lastSequence; + ZSTD_optimal_t lastStretch; ZSTD_optLdm_t optLdm; - ZSTD_memset(&lastSequence, 0, sizeof(ZSTD_optimal_t)); + ZSTD_memset(&lastStretch, 0, sizeof(ZSTD_optimal_t)); optLdm.seqStore = ms->ldmSeqStore ? *ms->ldmSeqStore : kNullRawSeqStore; optLdm.endPosInBlock = optLdm.startPosInBlock = optLdm.offset = 0; @@ -1108,19 +1122,31 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms, U32 const ll0 = !litlen; U32 nbMatches = getAllMatches(matches, ms, &nextToUpdate3, ip, iend, rep, ll0, minMatch); ZSTD_optLdm_processMatchCandidate(&optLdm, matches, &nbMatches, - (U32)(ip-istart), (U32)(iend - ip)); - if (!nbMatches) { ip++; continue; } + (U32)(ip-istart), (U32)(iend-ip)); + if (!nbMatches) { + DEBUGLOG(8, "no match found at cPos %u", (unsigned)(ip-istart)); + ip++; + continue; + } + + /* Match found: let's store this solution, and eventually find more candidates. + * During this forward pass, @opt is used to store stretches, + * defined as "a match followed by N literals". + * Note how this is different from a Sequence, which is "N literals followed by a match". + * Storing stretches allows us to store different match predecessors + * for each literal position part of a literals run. */ /* initialize opt[0] */ - { U32 i ; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; } - opt[0].mlen = 0; /* means is_a_literal */ + opt[0].mlen = 0; /* there are only literals so far */ opt[0].litlen = litlen; - /* We don't need to include the actual price of the literals because - * it is static for the duration of the forward pass, and is included - * in every price. We include the literal length to avoid negative - * prices when we subtract the previous literal length. + /* No need to include the actual price of the literals before the first match + * because it is static for the duration of the forward pass, and is included + * in every subsequent price. But, we include the literal length because + * the cost variation of litlen depends on the value of litlen. */ - opt[0].price = (int)ZSTD_litLengthPrice(litlen, optStatePtr, optLevel); + opt[0].price = LL_PRICE(litlen); + ZSTD_STATIC_ASSERT(sizeof(opt[0].rep[0]) == sizeof(rep[0])); + ZSTD_memcpy(&opt[0].rep, rep, sizeof(opt[0].rep)); /* large match -> immediate encoding */ { U32 const maxML = matches[nbMatches-1].len; @@ -1129,82 +1155,106 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms, nbMatches, maxML, maxOffBase, (U32)(ip-prefixStart)); if (maxML > sufficient_len) { - lastSequence.litlen = litlen; - lastSequence.mlen = maxML; - lastSequence.off = maxOffBase; - DEBUGLOG(6, "large match (%u>%u), immediate encoding", + lastStretch.litlen = 0; + lastStretch.mlen = maxML; + lastStretch.off = maxOffBase; + DEBUGLOG(6, "large match (%u>%u) => immediate encoding", maxML, sufficient_len); cur = 0; - last_pos = ZSTD_totalLen(lastSequence); + last_pos = maxML; goto _shortestPath; } } /* set prices for first matches starting position == 0 */ assert(opt[0].price >= 0); - { U32 const literalsPrice = (U32)opt[0].price + ZSTD_litLengthPrice(0, optStatePtr, optLevel); - U32 pos; + { U32 pos; U32 matchNb; for (pos = 1; pos < minMatch; pos++) { - opt[pos].price = ZSTD_MAX_PRICE; /* mlen, litlen and price will be fixed during forward scanning */ + opt[pos].price = ZSTD_MAX_PRICE; + opt[pos].mlen = 0; + opt[pos].litlen = litlen + pos; } for (matchNb = 0; matchNb < nbMatches; matchNb++) { U32 const offBase = matches[matchNb].off; U32 const end = matches[matchNb].len; for ( ; pos <= end ; pos++ ) { - U32 const matchPrice = ZSTD_getMatchPrice(offBase, pos, optStatePtr, optLevel); - U32 const sequencePrice = literalsPrice + matchPrice; + int const matchPrice = (int)ZSTD_getMatchPrice(offBase, pos, optStatePtr, optLevel); + int const sequencePrice = opt[0].price + matchPrice; DEBUGLOG(7, "rPos:%u => set initial price : %.2f", - pos, ZSTD_fCost((int)sequencePrice)); + pos, ZSTD_fCost(sequencePrice)); opt[pos].mlen = pos; opt[pos].off = offBase; - opt[pos].litlen = litlen; - opt[pos].price = (int)sequencePrice; - } } + opt[pos].litlen = 0; /* end of match */ + opt[pos].price = sequencePrice + LL_PRICE(0); + } + } last_pos = pos-1; + opt[pos].price = ZSTD_MAX_PRICE; } } /* check further positions */ for (cur = 1; cur <= last_pos; cur++) { const BYTE* const inr = ip + cur; - assert(cur < ZSTD_OPT_NUM); - DEBUGLOG(7, "cPos:%zi==rPos:%u", inr-istart, cur) + assert(cur <= ZSTD_OPT_NUM); + DEBUGLOG(7, "cPos:%zi==rPos:%u", inr-istart, cur); /* Fix current position with one literal if cheaper */ - { U32 const litlen = (opt[cur-1].mlen == 0) ? opt[cur-1].litlen + 1 : 1; + { U32 const litlen = opt[cur-1].litlen + 1; int const price = opt[cur-1].price - + (int)ZSTD_rawLiteralsCost(ip+cur-1, 1, optStatePtr, optLevel) - + (int)ZSTD_litLengthPrice(litlen, optStatePtr, optLevel) - - (int)ZSTD_litLengthPrice(litlen-1, optStatePtr, optLevel); + + LIT_PRICE(ip+cur-1) + + LL_INCPRICE(litlen); assert(price < 1000000000); /* overflow check */ if (price <= opt[cur].price) { + ZSTD_optimal_t const prevMatch = opt[cur]; DEBUGLOG(7, "cPos:%zi==rPos:%u : better price (%.2f<=%.2f) using literal (ll==%u) (hist:%u,%u,%u)", inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price), litlen, opt[cur-1].rep[0], opt[cur-1].rep[1], opt[cur-1].rep[2]); - opt[cur].mlen = 0; - opt[cur].off = 0; + opt[cur] = opt[cur-1]; opt[cur].litlen = litlen; opt[cur].price = price; + if ( (optLevel >= 1) /* additional check only for higher modes */ + && (prevMatch.litlen == 0) /* replace a match */ + && (LL_INCPRICE(1) < 0) /* ll1 is cheaper than ll0 */ + && LIKELY(ip + cur < iend) + ) { + /* check next position, in case it would be cheaper */ + int with1literal = prevMatch.price + LIT_PRICE(ip+cur) + LL_INCPRICE(1); + int withMoreLiterals = price + LIT_PRICE(ip+cur) + LL_INCPRICE(litlen+1); + DEBUGLOG(7, "then at next rPos %u : match+1lit %.2f vs %ulits %.2f", + cur+1, ZSTD_fCost(with1literal), litlen+1, ZSTD_fCost(withMoreLiterals)); + if ( (with1literal < withMoreLiterals) + && (with1literal < opt[cur+1].price) ) { + /* update offset history - before it disappears */ + U32 const prev = cur - prevMatch.mlen; + repcodes_t const newReps = ZSTD_newRep(opt[prev].rep, prevMatch.off, opt[prev].litlen==0); + assert(cur >= prevMatch.mlen); + DEBUGLOG(7, "==> match+1lit is cheaper (%.2f < %.2f) (hist:%u,%u,%u) !", + ZSTD_fCost(with1literal), ZSTD_fCost(withMoreLiterals), + newReps.rep[0], newReps.rep[1], newReps.rep[2] ); + opt[cur+1] = prevMatch; /* mlen & offbase */ + ZSTD_memcpy(opt[cur+1].rep, &newReps, sizeof(repcodes_t)); + opt[cur+1].litlen = 1; + opt[cur+1].price = with1literal; + if (last_pos < cur+1) last_pos = cur+1; + } + } } else { - DEBUGLOG(7, "cPos:%zi==rPos:%u : literal would cost more (%.2f>%.2f) (hist:%u,%u,%u)", - inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price), - opt[cur].rep[0], opt[cur].rep[1], opt[cur].rep[2]); + DEBUGLOG(7, "cPos:%zi==rPos:%u : literal would cost more (%.2f>%.2f)", + inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price)); } } - /* Set the repcodes of the current position. We must do it here - * because we rely on the repcodes of the 2nd to last sequence being - * correct to set the next chunks repcodes during the backward - * traversal. + /* Offset history is not updated during match comparison. + * Do it here, now that the match is selected and confirmed. */ ZSTD_STATIC_ASSERT(sizeof(opt[cur].rep) == sizeof(repcodes_t)); assert(cur >= opt[cur].mlen); - if (opt[cur].mlen != 0) { + if (opt[cur].litlen == 0) { + /* just finished a match => alter offset history */ U32 const prev = cur - opt[cur].mlen; - repcodes_t const newReps = ZSTD_newRep(opt[prev].rep, opt[cur].off, opt[cur].litlen==0); + repcodes_t const newReps = ZSTD_newRep(opt[prev].rep, opt[cur].off, opt[prev].litlen==0); ZSTD_memcpy(opt[cur].rep, &newReps, sizeof(repcodes_t)); - } else { - ZSTD_memcpy(opt[cur].rep, opt[cur - 1].rep, sizeof(repcodes_t)); } /* last match must start at a minimum distance of 8 from oend */ @@ -1214,15 +1264,14 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms, if ( (optLevel==0) /*static_test*/ && (opt[cur+1].price <= opt[cur].price + (BITCOST_MULTIPLIER/2)) ) { - DEBUGLOG(7, "move to next rPos:%u : price is <=", cur+1); + DEBUGLOG(7, "skip current position : next rPos(%u) price is cheaper", cur+1); continue; /* skip unpromising positions; about ~+6% speed, -0.01 ratio */ } assert(opt[cur].price >= 0); - { U32 const ll0 = (opt[cur].mlen != 0); - U32 const litlen = (opt[cur].mlen == 0) ? opt[cur].litlen : 0; - U32 const previousPrice = (U32)opt[cur].price; - U32 const basePrice = previousPrice + ZSTD_litLengthPrice(0, optStatePtr, optLevel); + { U32 const ll0 = (opt[cur].litlen == 0); + int const previousPrice = opt[cur].price; + int const basePrice = previousPrice + LL_PRICE(0); U32 nbMatches = getAllMatches(matches, ms, &nextToUpdate3, inr, iend, opt[cur].rep, ll0, minMatch); U32 matchNb; @@ -1234,18 +1283,17 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms, continue; } - { U32 const maxML = matches[nbMatches-1].len; - DEBUGLOG(7, "cPos:%zi==rPos:%u, found %u matches, of maxLength=%u", - inr-istart, cur, nbMatches, maxML); - - if ( (maxML > sufficient_len) - || (cur + maxML >= ZSTD_OPT_NUM) ) { - lastSequence.mlen = maxML; - lastSequence.off = matches[nbMatches-1].off; - lastSequence.litlen = litlen; - cur -= (opt[cur].mlen==0) ? opt[cur].litlen : 0; /* last sequence is actually only literals, fix cur to last match - note : may underflow, in which case, it's first sequence, and it's okay */ - last_pos = cur + ZSTD_totalLen(lastSequence); - if (cur > ZSTD_OPT_NUM) cur = 0; /* underflow => first match */ + { U32 const longestML = matches[nbMatches-1].len; + DEBUGLOG(7, "cPos:%zi==rPos:%u, found %u matches, of longest ML=%u", + inr-istart, cur, nbMatches, longestML); + + if ( (longestML > sufficient_len) + || (cur + longestML >= ZSTD_OPT_NUM) + || (ip + cur + longestML >= iend) ) { + lastStretch.mlen = longestML; + lastStretch.off = matches[nbMatches-1].off; + lastStretch.litlen = 0; + last_pos = cur + longestML; goto _shortestPath; } } @@ -1257,19 +1305,24 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms, U32 mlen; DEBUGLOG(7, "testing match %u => offBase=%4u, mlen=%2u, llen=%2u", - matchNb, matches[matchNb].off, lastML, litlen); + matchNb, matches[matchNb].off, lastML, opt[cur].litlen); for (mlen = lastML; mlen >= startML; mlen--) { /* scan downward */ U32 const pos = cur + mlen; - int const price = (int)basePrice + (int)ZSTD_getMatchPrice(offset, mlen, optStatePtr, optLevel); + int const price = basePrice + (int)ZSTD_getMatchPrice(offset, mlen, optStatePtr, optLevel); if ((pos > last_pos) || (price < opt[pos].price)) { DEBUGLOG(7, "rPos:%u (ml=%2u) => new better price (%.2f<%.2f)", pos, mlen, ZSTD_fCost(price), ZSTD_fCost(opt[pos].price)); - while (last_pos < pos) { opt[last_pos+1].price = ZSTD_MAX_PRICE; last_pos++; } /* fill empty positions */ + while (last_pos < pos) { + /* fill empty positions, for future comparisons */ + last_pos++; + opt[last_pos].price = ZSTD_MAX_PRICE; + opt[last_pos].litlen = !0; /* just needs to be != 0, to mean "not an end of match" */ + } opt[pos].mlen = mlen; opt[pos].off = offset; - opt[pos].litlen = litlen; + opt[pos].litlen = 0; opt[pos].price = price; } else { DEBUGLOG(7, "rPos:%u (ml=%2u) => new price is worse (%.2f>=%.2f)", @@ -1277,47 +1330,81 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms, if (optLevel==0) break; /* early update abort; gets ~+10% speed for about -0.01 ratio loss */ } } } } + opt[last_pos+1].price = ZSTD_MAX_PRICE; } /* for (cur = 1; cur <= last_pos; cur++) */ - lastSequence = opt[last_pos]; - cur = last_pos > ZSTD_totalLen(lastSequence) ? last_pos - ZSTD_totalLen(lastSequence) : 0; /* single sequence, and it starts before `ip` */ - assert(cur < ZSTD_OPT_NUM); /* control overflow*/ + lastStretch = opt[last_pos]; + assert(cur >= lastStretch.mlen); + cur = last_pos - lastStretch.mlen; _shortestPath: /* cur, last_pos, best_mlen, best_off have to be set */ assert(opt[0].mlen == 0); + assert(last_pos >= lastStretch.mlen); + assert(cur == last_pos - lastStretch.mlen); - /* Set the next chunk's repcodes based on the repcodes of the beginning - * of the last match, and the last sequence. This avoids us having to - * update them while traversing the sequences. - */ - if (lastSequence.mlen != 0) { - repcodes_t const reps = ZSTD_newRep(opt[cur].rep, lastSequence.off, lastSequence.litlen==0); - ZSTD_memcpy(rep, &reps, sizeof(reps)); + if (lastStretch.mlen==0) { + /* no solution : all matches have been converted into literals */ + assert(lastStretch.litlen == (ip - anchor) + last_pos); + ip += last_pos; + continue; + } + assert(lastStretch.off > 0); + + /* Update offset history */ + if (lastStretch.litlen == 0) { + /* finishing on a match : update offset history */ + repcodes_t const reps = ZSTD_newRep(opt[cur].rep, lastStretch.off, opt[cur].litlen==0); + ZSTD_memcpy(rep, &reps, sizeof(repcodes_t)); } else { - ZSTD_memcpy(rep, opt[cur].rep, sizeof(repcodes_t)); + ZSTD_memcpy(rep, lastStretch.rep, sizeof(repcodes_t)); + assert(cur >= lastStretch.litlen); + cur -= lastStretch.litlen; } - { U32 const storeEnd = cur + 1; + /* Let's write the shortest path solution. + * It is stored in @opt in reverse order, + * starting from @storeEnd (==cur+2), + * effectively partially @opt overwriting. + * Content is changed too: + * - So far, @opt stored stretches, aka a match followed by literals + * - Now, it will store sequences, aka literals followed by a match + */ + { U32 const storeEnd = cur + 2; U32 storeStart = storeEnd; - U32 seqPos = cur; + U32 stretchPos = cur; DEBUGLOG(6, "start reverse traversal (last_pos:%u, cur:%u)", last_pos, cur); (void)last_pos; - assert(storeEnd < ZSTD_OPT_NUM); - DEBUGLOG(6, "last sequence copied into pos=%u (llen=%u,mlen=%u,ofc=%u)", - storeEnd, lastSequence.litlen, lastSequence.mlen, lastSequence.off); - opt[storeEnd] = lastSequence; - while (seqPos > 0) { - U32 const backDist = ZSTD_totalLen(opt[seqPos]); + assert(storeEnd < ZSTD_OPT_SIZE); + DEBUGLOG(6, "last stretch copied into pos=%u (llen=%u,mlen=%u,ofc=%u)", + storeEnd, lastStretch.litlen, lastStretch.mlen, lastStretch.off); + if (lastStretch.litlen > 0) { + /* last "sequence" is unfinished: just a bunch of literals */ + opt[storeEnd].litlen = lastStretch.litlen; + opt[storeEnd].mlen = 0; + storeStart = storeEnd-1; + opt[storeStart] = lastStretch; + } { + opt[storeEnd] = lastStretch; /* note: litlen will be fixed */ + storeStart = storeEnd; + } + while (1) { + ZSTD_optimal_t nextStretch = opt[stretchPos]; + opt[storeStart].litlen = nextStretch.litlen; + DEBUGLOG(6, "selected sequence (llen=%u,mlen=%u,ofc=%u)", + opt[storeStart].litlen, opt[storeStart].mlen, opt[storeStart].off); + if (nextStretch.mlen == 0) { + /* reaching beginning of segment */ + break; + } storeStart--; - DEBUGLOG(6, "sequence from rPos=%u copied into pos=%u (llen=%u,mlen=%u,ofc=%u)", - seqPos, storeStart, opt[seqPos].litlen, opt[seqPos].mlen, opt[seqPos].off); - opt[storeStart] = opt[seqPos]; - seqPos = (seqPos > backDist) ? seqPos - backDist : 0; + opt[storeStart] = nextStretch; /* note: litlen will be fixed */ + assert(nextStretch.litlen + nextStretch.mlen <= stretchPos); + stretchPos -= nextStretch.litlen + nextStretch.mlen; } /* save sequences */ - DEBUGLOG(6, "sending selected sequences into seqStore") + DEBUGLOG(6, "sending selected sequences into seqStore"); { U32 storePos; for (storePos=storeStart; storePos <= storeEnd; storePos++) { U32 const llen = opt[storePos].litlen; @@ -1339,6 +1426,9 @@ _shortestPath: /* cur, last_pos, best_mlen, best_off have to be set */ anchor += advance; ip = anchor; } } + DEBUGLOG(7, "new offset history : %u, %u, %u", rep[0], rep[1], rep[2]); + + /* update all costs */ ZSTD_setBasePrices(optStatePtr, optLevel); } } /* while (ip < ilimit) */ @@ -1346,21 +1436,27 @@ _shortestPath: /* cur, last_pos, best_mlen, best_off have to be set */ /* Return the last literals size */ return (size_t)(iend - anchor); } +#endif /* build exclusions */ +#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR static size_t ZSTD_compressBlock_opt0( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], const void* src, size_t srcSize, const ZSTD_dictMode_e dictMode) { return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 0 /* optLevel */, dictMode); } +#endif +#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR static size_t ZSTD_compressBlock_opt2( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], const void* src, size_t srcSize, const ZSTD_dictMode_e dictMode) { return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /* optLevel */, dictMode); } +#endif +#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR size_t ZSTD_compressBlock_btopt( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], const void* src, size_t srcSize) @@ -1368,20 +1464,23 @@ size_t ZSTD_compressBlock_btopt( DEBUGLOG(5, "ZSTD_compressBlock_btopt"); return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_noDict); } +#endif +#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR /* ZSTD_initStats_ultra(): * make a first compression pass, just to seed stats with more accurate starting values. * only works on first block, with no dictionary and no ldm. * this function cannot error out, its narrow contract must be respected. */ -static void -ZSTD_initStats_ultra(ZSTD_matchState_t* ms, - seqStore_t* seqStore, - U32 rep[ZSTD_REP_NUM], - const void* src, size_t srcSize) +static +ZSTD_ALLOW_POINTER_OVERFLOW_ATTR +void ZSTD_initStats_ultra(ZSTD_matchState_t* ms, + seqStore_t* seqStore, + U32 rep[ZSTD_REP_NUM], + const void* src, size_t srcSize) { U32 tmpRep[ZSTD_REP_NUM]; /* updated rep codes will sink here */ ZSTD_memcpy(tmpRep, rep, sizeof(tmpRep)); @@ -1425,7 +1524,7 @@ size_t ZSTD_compressBlock_btultra2( * Consequently, this can only work if no data has been previously loaded in tables, * aka, no dictionary, no prefix, no ldm preprocessing. * The compression ratio gain is generally small (~0.5% on first block), - ** the cost is 2x cpu time on first block. */ + * the cost is 2x cpu time on first block. */ assert(srcSize <= ZSTD_BLOCKSIZE_MAX); if ( (ms->opt.litLengthSum==0) /* first block */ && (seqStore->sequences == seqStore->sequencesStart) /* no ldm */ @@ -1438,7 +1537,9 @@ size_t ZSTD_compressBlock_btultra2( return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_noDict); } +#endif +#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR size_t ZSTD_compressBlock_btopt_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], const void* src, size_t srcSize) @@ -1446,18 +1547,20 @@ size_t ZSTD_compressBlock_btopt_dictMatchState( return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_dictMatchState); } -size_t ZSTD_compressBlock_btultra_dictMatchState( +size_t ZSTD_compressBlock_btopt_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], const void* src, size_t srcSize) { - return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_dictMatchState); + return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_extDict); } +#endif -size_t ZSTD_compressBlock_btopt_extDict( +#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_btultra_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], const void* src, size_t srcSize) { - return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_extDict); + return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_dictMatchState); } size_t ZSTD_compressBlock_btultra_extDict( @@ -1466,6 +1569,7 @@ size_t ZSTD_compressBlock_btultra_extDict( { return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_extDict); } +#endif /* note : no btultra2 variant for extDict nor dictMatchState, * because btultra2 is not meant to work with dictionaries diff --git a/thirdparty/zstd/compress/zstd_opt.h b/thirdparty/zstd/compress/zstd_opt.h index 342e5a3112..d4e7113157 100644 --- a/thirdparty/zstd/compress/zstd_opt.h +++ b/thirdparty/zstd/compress/zstd_opt.h @@ -17,30 +17,40 @@ extern "C" { #include "zstd_compress_internal.h" +#if !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR) \ + || !defined(ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR) /* used in ZSTD_loadDictionaryContent() */ void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend); +#endif +#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR size_t ZSTD_compressBlock_btopt( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_btultra( +size_t ZSTD_compressBlock_btopt_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); -size_t ZSTD_compressBlock_btultra2( +size_t ZSTD_compressBlock_btopt_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); +#define ZSTD_COMPRESSBLOCK_BTOPT ZSTD_compressBlock_btopt +#define ZSTD_COMPRESSBLOCK_BTOPT_DICTMATCHSTATE ZSTD_compressBlock_btopt_dictMatchState +#define ZSTD_COMPRESSBLOCK_BTOPT_EXTDICT ZSTD_compressBlock_btopt_extDict +#else +#define ZSTD_COMPRESSBLOCK_BTOPT NULL +#define ZSTD_COMPRESSBLOCK_BTOPT_DICTMATCHSTATE NULL +#define ZSTD_COMPRESSBLOCK_BTOPT_EXTDICT NULL +#endif -size_t ZSTD_compressBlock_btopt_dictMatchState( +#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR +size_t ZSTD_compressBlock_btultra( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); size_t ZSTD_compressBlock_btultra_dictMatchState( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); - -size_t ZSTD_compressBlock_btopt_extDict( - ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], - void const* src, size_t srcSize); size_t ZSTD_compressBlock_btultra_extDict( ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], void const* src, size_t srcSize); @@ -48,6 +58,20 @@ size_t ZSTD_compressBlock_btultra_extDict( /* note : no btultra2 variant for extDict nor dictMatchState, * because btultra2 is not meant to work with dictionaries * and is only specific for the first block (no prefix) */ +size_t ZSTD_compressBlock_btultra2( + ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], + void const* src, size_t srcSize); + +#define ZSTD_COMPRESSBLOCK_BTULTRA ZSTD_compressBlock_btultra +#define ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE ZSTD_compressBlock_btultra_dictMatchState +#define ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT ZSTD_compressBlock_btultra_extDict +#define ZSTD_COMPRESSBLOCK_BTULTRA2 ZSTD_compressBlock_btultra2 +#else +#define ZSTD_COMPRESSBLOCK_BTULTRA NULL +#define ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE NULL +#define ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT NULL +#define ZSTD_COMPRESSBLOCK_BTULTRA2 NULL +#endif #if defined (__cplusplus) } diff --git a/thirdparty/zstd/compress/zstdmt_compress.c b/thirdparty/zstd/compress/zstdmt_compress.c index 6786075569..86ccce3184 100644 --- a/thirdparty/zstd/compress/zstdmt_compress.c +++ b/thirdparty/zstd/compress/zstdmt_compress.c @@ -15,17 +15,13 @@ #endif -/* ====== Constants ====== */ -#define ZSTDMT_OVERLAPLOG_DEFAULT 0 - - /* ====== Dependencies ====== */ -#include "../common/allocations.h" /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */ +#include "../common/allocations.h" /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */ #include "../common/zstd_deps.h" /* ZSTD_memcpy, ZSTD_memset, INT_MAX, UINT_MAX */ #include "../common/mem.h" /* MEM_STATIC */ #include "../common/pool.h" /* threadpool */ #include "../common/threading.h" /* mutex */ -#include "zstd_compress_internal.h" /* MIN, ERROR, ZSTD_*, ZSTD_highbit32 */ +#include "zstd_compress_internal.h" /* MIN, ERROR, ZSTD_*, ZSTD_highbit32 */ #include "zstd_ldm.h" #include "zstdmt_compress.h" @@ -44,12 +40,13 @@ # include <unistd.h> # include <sys/times.h> -# define DEBUG_PRINTHEX(l,p,n) { \ - unsigned debug_u; \ - for (debug_u=0; debug_u<(n); debug_u++) \ - RAWLOG(l, "%02X ", ((const unsigned char*)(p))[debug_u]); \ - RAWLOG(l, " \n"); \ -} +# define DEBUG_PRINTHEX(l,p,n) \ + do { \ + unsigned debug_u; \ + for (debug_u=0; debug_u<(n); debug_u++) \ + RAWLOG(l, "%02X ", ((const unsigned char*)(p))[debug_u]); \ + RAWLOG(l, " \n"); \ + } while (0) static unsigned long long GetCurrentClockTimeMicroseconds(void) { @@ -61,25 +58,28 @@ static unsigned long long GetCurrentClockTimeMicroseconds(void) } } #define MUTEX_WAIT_TIME_DLEVEL 6 -#define ZSTD_PTHREAD_MUTEX_LOCK(mutex) { \ - if (DEBUGLEVEL >= MUTEX_WAIT_TIME_DLEVEL) { \ - unsigned long long const beforeTime = GetCurrentClockTimeMicroseconds(); \ - ZSTD_pthread_mutex_lock(mutex); \ - { unsigned long long const afterTime = GetCurrentClockTimeMicroseconds(); \ - unsigned long long const elapsedTime = (afterTime-beforeTime); \ - if (elapsedTime > 1000) { /* or whatever threshold you like; I'm using 1 millisecond here */ \ - DEBUGLOG(MUTEX_WAIT_TIME_DLEVEL, "Thread took %llu microseconds to acquire mutex %s \n", \ - elapsedTime, #mutex); \ - } } \ - } else { \ - ZSTD_pthread_mutex_lock(mutex); \ - } \ -} +#define ZSTD_PTHREAD_MUTEX_LOCK(mutex) \ + do { \ + if (DEBUGLEVEL >= MUTEX_WAIT_TIME_DLEVEL) { \ + unsigned long long const beforeTime = GetCurrentClockTimeMicroseconds(); \ + ZSTD_pthread_mutex_lock(mutex); \ + { unsigned long long const afterTime = GetCurrentClockTimeMicroseconds(); \ + unsigned long long const elapsedTime = (afterTime-beforeTime); \ + if (elapsedTime > 1000) { \ + /* or whatever threshold you like; I'm using 1 millisecond here */ \ + DEBUGLOG(MUTEX_WAIT_TIME_DLEVEL, \ + "Thread took %llu microseconds to acquire mutex %s \n", \ + elapsedTime, #mutex); \ + } } \ + } else { \ + ZSTD_pthread_mutex_lock(mutex); \ + } \ + } while (0) #else # define ZSTD_PTHREAD_MUTEX_LOCK(m) ZSTD_pthread_mutex_lock(m) -# define DEBUG_PRINTHEX(l,p,n) {} +# define DEBUG_PRINTHEX(l,p,n) do { } while (0) #endif @@ -100,18 +100,39 @@ typedef struct ZSTDMT_bufferPool_s { unsigned totalBuffers; unsigned nbBuffers; ZSTD_customMem cMem; - buffer_t bTable[1]; /* variable size */ + buffer_t* buffers; } ZSTDMT_bufferPool; +static void ZSTDMT_freeBufferPool(ZSTDMT_bufferPool* bufPool) +{ + DEBUGLOG(3, "ZSTDMT_freeBufferPool (address:%08X)", (U32)(size_t)bufPool); + if (!bufPool) return; /* compatibility with free on NULL */ + if (bufPool->buffers) { + unsigned u; + for (u=0; u<bufPool->totalBuffers; u++) { + DEBUGLOG(4, "free buffer %2u (address:%08X)", u, (U32)(size_t)bufPool->buffers[u].start); + ZSTD_customFree(bufPool->buffers[u].start, bufPool->cMem); + } + ZSTD_customFree(bufPool->buffers, bufPool->cMem); + } + ZSTD_pthread_mutex_destroy(&bufPool->poolMutex); + ZSTD_customFree(bufPool, bufPool->cMem); +} + static ZSTDMT_bufferPool* ZSTDMT_createBufferPool(unsigned maxNbBuffers, ZSTD_customMem cMem) { - ZSTDMT_bufferPool* const bufPool = (ZSTDMT_bufferPool*)ZSTD_customCalloc( - sizeof(ZSTDMT_bufferPool) + (maxNbBuffers-1) * sizeof(buffer_t), cMem); + ZSTDMT_bufferPool* const bufPool = + (ZSTDMT_bufferPool*)ZSTD_customCalloc(sizeof(ZSTDMT_bufferPool), cMem); if (bufPool==NULL) return NULL; if (ZSTD_pthread_mutex_init(&bufPool->poolMutex, NULL)) { ZSTD_customFree(bufPool, cMem); return NULL; } + bufPool->buffers = (buffer_t*)ZSTD_customCalloc(maxNbBuffers * sizeof(buffer_t), cMem); + if (bufPool->buffers==NULL) { + ZSTDMT_freeBufferPool(bufPool); + return NULL; + } bufPool->bufferSize = 64 KB; bufPool->totalBuffers = maxNbBuffers; bufPool->nbBuffers = 0; @@ -119,32 +140,19 @@ static ZSTDMT_bufferPool* ZSTDMT_createBufferPool(unsigned maxNbBuffers, ZSTD_cu return bufPool; } -static void ZSTDMT_freeBufferPool(ZSTDMT_bufferPool* bufPool) -{ - unsigned u; - DEBUGLOG(3, "ZSTDMT_freeBufferPool (address:%08X)", (U32)(size_t)bufPool); - if (!bufPool) return; /* compatibility with free on NULL */ - for (u=0; u<bufPool->totalBuffers; u++) { - DEBUGLOG(4, "free buffer %2u (address:%08X)", u, (U32)(size_t)bufPool->bTable[u].start); - ZSTD_customFree(bufPool->bTable[u].start, bufPool->cMem); - } - ZSTD_pthread_mutex_destroy(&bufPool->poolMutex); - ZSTD_customFree(bufPool, bufPool->cMem); -} - /* only works at initialization, not during compression */ static size_t ZSTDMT_sizeof_bufferPool(ZSTDMT_bufferPool* bufPool) { - size_t const poolSize = sizeof(*bufPool) - + (bufPool->totalBuffers - 1) * sizeof(buffer_t); + size_t const poolSize = sizeof(*bufPool); + size_t const arraySize = bufPool->totalBuffers * sizeof(buffer_t); unsigned u; size_t totalBufferSize = 0; ZSTD_pthread_mutex_lock(&bufPool->poolMutex); for (u=0; u<bufPool->totalBuffers; u++) - totalBufferSize += bufPool->bTable[u].capacity; + totalBufferSize += bufPool->buffers[u].capacity; ZSTD_pthread_mutex_unlock(&bufPool->poolMutex); - return poolSize + totalBufferSize; + return poolSize + arraySize + totalBufferSize; } /* ZSTDMT_setBufferSize() : @@ -187,9 +195,9 @@ static buffer_t ZSTDMT_getBuffer(ZSTDMT_bufferPool* bufPool) DEBUGLOG(5, "ZSTDMT_getBuffer: bSize = %u", (U32)bufPool->bufferSize); ZSTD_pthread_mutex_lock(&bufPool->poolMutex); if (bufPool->nbBuffers) { /* try to use an existing buffer */ - buffer_t const buf = bufPool->bTable[--(bufPool->nbBuffers)]; + buffer_t const buf = bufPool->buffers[--(bufPool->nbBuffers)]; size_t const availBufferSize = buf.capacity; - bufPool->bTable[bufPool->nbBuffers] = g_nullBuffer; + bufPool->buffers[bufPool->nbBuffers] = g_nullBuffer; if ((availBufferSize >= bSize) & ((availBufferSize>>3) <= bSize)) { /* large enough, but not too much */ DEBUGLOG(5, "ZSTDMT_getBuffer: provide buffer %u of size %u", @@ -250,14 +258,14 @@ static void ZSTDMT_releaseBuffer(ZSTDMT_bufferPool* bufPool, buffer_t buf) if (buf.start == NULL) return; /* compatible with release on NULL */ ZSTD_pthread_mutex_lock(&bufPool->poolMutex); if (bufPool->nbBuffers < bufPool->totalBuffers) { - bufPool->bTable[bufPool->nbBuffers++] = buf; /* stored for later use */ + bufPool->buffers[bufPool->nbBuffers++] = buf; /* stored for later use */ DEBUGLOG(5, "ZSTDMT_releaseBuffer: stored buffer of size %u in slot %u", (U32)buf.capacity, (U32)(bufPool->nbBuffers-1)); ZSTD_pthread_mutex_unlock(&bufPool->poolMutex); return; } ZSTD_pthread_mutex_unlock(&bufPool->poolMutex); - /* Reached bufferPool capacity (should not happen) */ + /* Reached bufferPool capacity (note: should not happen) */ DEBUGLOG(5, "ZSTDMT_releaseBuffer: pool capacity reached => freeing "); ZSTD_customFree(buf.start, bufPool->cMem); } @@ -350,16 +358,20 @@ typedef struct { int totalCCtx; int availCCtx; ZSTD_customMem cMem; - ZSTD_CCtx* cctx[1]; /* variable size */ + ZSTD_CCtx** cctxs; } ZSTDMT_CCtxPool; -/* note : all CCtx borrowed from the pool should be released back to the pool _before_ freeing the pool */ +/* note : all CCtx borrowed from the pool must be reverted back to the pool _before_ freeing the pool */ static void ZSTDMT_freeCCtxPool(ZSTDMT_CCtxPool* pool) { - int cid; - for (cid=0; cid<pool->totalCCtx; cid++) - ZSTD_freeCCtx(pool->cctx[cid]); /* note : compatible with free on NULL */ + if (!pool) return; ZSTD_pthread_mutex_destroy(&pool->poolMutex); + if (pool->cctxs) { + int cid; + for (cid=0; cid<pool->totalCCtx; cid++) + ZSTD_freeCCtx(pool->cctxs[cid]); /* free compatible with NULL */ + ZSTD_customFree(pool->cctxs, pool->cMem); + } ZSTD_customFree(pool, pool->cMem); } @@ -368,19 +380,24 @@ static void ZSTDMT_freeCCtxPool(ZSTDMT_CCtxPool* pool) static ZSTDMT_CCtxPool* ZSTDMT_createCCtxPool(int nbWorkers, ZSTD_customMem cMem) { - ZSTDMT_CCtxPool* const cctxPool = (ZSTDMT_CCtxPool*) ZSTD_customCalloc( - sizeof(ZSTDMT_CCtxPool) + (nbWorkers-1)*sizeof(ZSTD_CCtx*), cMem); + ZSTDMT_CCtxPool* const cctxPool = + (ZSTDMT_CCtxPool*) ZSTD_customCalloc(sizeof(ZSTDMT_CCtxPool), cMem); assert(nbWorkers > 0); if (!cctxPool) return NULL; if (ZSTD_pthread_mutex_init(&cctxPool->poolMutex, NULL)) { ZSTD_customFree(cctxPool, cMem); return NULL; } - cctxPool->cMem = cMem; cctxPool->totalCCtx = nbWorkers; + cctxPool->cctxs = (ZSTD_CCtx**)ZSTD_customCalloc(nbWorkers * sizeof(ZSTD_CCtx*), cMem); + if (!cctxPool->cctxs) { + ZSTDMT_freeCCtxPool(cctxPool); + return NULL; + } + cctxPool->cMem = cMem; + cctxPool->cctxs[0] = ZSTD_createCCtx_advanced(cMem); + if (!cctxPool->cctxs[0]) { ZSTDMT_freeCCtxPool(cctxPool); return NULL; } cctxPool->availCCtx = 1; /* at least one cctx for single-thread mode */ - cctxPool->cctx[0] = ZSTD_createCCtx_advanced(cMem); - if (!cctxPool->cctx[0]) { ZSTDMT_freeCCtxPool(cctxPool); return NULL; } DEBUGLOG(3, "cctxPool created, with %u workers", nbWorkers); return cctxPool; } @@ -402,16 +419,16 @@ static size_t ZSTDMT_sizeof_CCtxPool(ZSTDMT_CCtxPool* cctxPool) { ZSTD_pthread_mutex_lock(&cctxPool->poolMutex); { unsigned const nbWorkers = cctxPool->totalCCtx; - size_t const poolSize = sizeof(*cctxPool) - + (nbWorkers-1) * sizeof(ZSTD_CCtx*); - unsigned u; + size_t const poolSize = sizeof(*cctxPool); + size_t const arraySize = cctxPool->totalCCtx * sizeof(ZSTD_CCtx*); size_t totalCCtxSize = 0; + unsigned u; for (u=0; u<nbWorkers; u++) { - totalCCtxSize += ZSTD_sizeof_CCtx(cctxPool->cctx[u]); + totalCCtxSize += ZSTD_sizeof_CCtx(cctxPool->cctxs[u]); } ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex); assert(nbWorkers > 0); - return poolSize + totalCCtxSize; + return poolSize + arraySize + totalCCtxSize; } } @@ -421,7 +438,7 @@ static ZSTD_CCtx* ZSTDMT_getCCtx(ZSTDMT_CCtxPool* cctxPool) ZSTD_pthread_mutex_lock(&cctxPool->poolMutex); if (cctxPool->availCCtx) { cctxPool->availCCtx--; - { ZSTD_CCtx* const cctx = cctxPool->cctx[cctxPool->availCCtx]; + { ZSTD_CCtx* const cctx = cctxPool->cctxs[cctxPool->availCCtx]; ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex); return cctx; } } @@ -435,7 +452,7 @@ static void ZSTDMT_releaseCCtx(ZSTDMT_CCtxPool* pool, ZSTD_CCtx* cctx) if (cctx==NULL) return; /* compatibility with release on NULL */ ZSTD_pthread_mutex_lock(&pool->poolMutex); if (pool->availCCtx < pool->totalCCtx) - pool->cctx[pool->availCCtx++] = cctx; + pool->cctxs[pool->availCCtx++] = cctx; else { /* pool overflow : should not happen, since totalCCtx==nbWorkers */ DEBUGLOG(4, "CCtx pool overflow : free cctx"); @@ -601,11 +618,8 @@ static void ZSTDMT_serialState_update(serialState_t* serialState, ZSTD_pthread_mutex_unlock(&serialState->mutex); if (seqStore.size > 0) { - size_t const err = ZSTD_referenceExternalSequences( - jobCCtx, seqStore.seq, seqStore.size); + ZSTD_referenceExternalSequences(jobCCtx, seqStore.seq, seqStore.size); assert(serialState->params.ldmParams.enableLdm == ZSTD_ps_enable); - assert(!ZSTD_isError(err)); - (void)err; } } @@ -657,12 +671,13 @@ typedef struct { unsigned frameChecksumNeeded; /* used only by mtctx */ } ZSTDMT_jobDescription; -#define JOB_ERROR(e) { \ - ZSTD_PTHREAD_MUTEX_LOCK(&job->job_mutex); \ - job->cSize = e; \ - ZSTD_pthread_mutex_unlock(&job->job_mutex); \ - goto _endJob; \ -} +#define JOB_ERROR(e) \ + do { \ + ZSTD_PTHREAD_MUTEX_LOCK(&job->job_mutex); \ + job->cSize = e; \ + ZSTD_pthread_mutex_unlock(&job->job_mutex); \ + goto _endJob; \ + } while (0) /* ZSTDMT_compressionJob() is a POOL_function type */ static void ZSTDMT_compressionJob(void* jobDescription) @@ -1091,7 +1106,7 @@ ZSTD_frameProgression ZSTDMT_getFrameProgression(ZSTDMT_CCtx* mtctx) { unsigned jobNb; unsigned lastJobNb = mtctx->nextJobID + mtctx->jobReady; assert(mtctx->jobReady <= 1); DEBUGLOG(6, "ZSTDMT_getFrameProgression: jobs: from %u to <%u (jobReady:%u)", - mtctx->doneJobID, lastJobNb, mtctx->jobReady) + mtctx->doneJobID, lastJobNb, mtctx->jobReady); for (jobNb = mtctx->doneJobID ; jobNb < lastJobNb ; jobNb++) { unsigned const wJobID = jobNb & mtctx->jobIDMask; ZSTDMT_jobDescription* jobPtr = &mtctx->jobs[wJobID]; |