/* * Copyright (c) Meta Platforms, Inc. and affiliates. * All rights reserved. * * This source code is licensed under both the BSD-style license (found in the * LICENSE file in the root directory of this source tree) and the GPLv2 (found * in the COPYING file in the root directory of this source tree). * You may select, at your option, one of the above-listed licenses. */ /* This header contains definitions * that shall **only** be used by modules within lib/compress. */ #ifndef ZSTD_COMPRESS_H #define ZSTD_COMPRESS_H /*-************************************* * Dependencies ***************************************/ #include "../common/zstd_internal.h" #include "zstd_cwksp.h" #ifdef ZSTD_MULTITHREAD # include "zstdmt_compress.h" #endif #include "../common/bits.h" /* ZSTD_highbit32, ZSTD_NbCommonBytes */ #if defined (__cplusplus) extern "C" { #endif /*-************************************* * Constants ***************************************/ #define kSearchStrength 8 #define HASH_READ_SIZE 8 #define ZSTD_DUBT_UNSORTED_MARK 1 /* For btlazy2 strategy, index ZSTD_DUBT_UNSORTED_MARK==1 means "unsorted". It could be confused for a real successor at index "1", if sorted as larger than its predecessor. 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 reuse with a different strategy. This constant is required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */ /*-************************************* * Context memory management ***************************************/ typedef enum { ZSTDcs_created=0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e; typedef enum { zcss_init=0, zcss_load, zcss_flush } ZSTD_cStreamStage; typedef struct ZSTD_prefixDict_s { const void* dict; size_t dictSize; ZSTD_dictContentType_e dictContentType; } ZSTD_prefixDict; typedef struct { void* dictBuffer; void const* dict; size_t dictSize; ZSTD_dictContentType_e dictContentType; ZSTD_CDict* cdict; } ZSTD_localDict; typedef struct { HUF_CElt CTable[HUF_CTABLE_SIZE_ST(255)]; HUF_repeat repeatMode; } ZSTD_hufCTables_t; typedef struct { FSE_CTable offcodeCTable[FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)]; FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)]; FSE_CTable litlengthCTable[FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)]; FSE_repeat offcode_repeatMode; FSE_repeat matchlength_repeatMode; FSE_repeat litlength_repeatMode; } ZSTD_fseCTables_t; typedef struct { ZSTD_hufCTables_t huf; ZSTD_fseCTables_t fse; } ZSTD_entropyCTables_t; /*********************************************** * Entropy buffer statistics structs and funcs * ***********************************************/ /** ZSTD_hufCTablesMetadata_t : * Stores Literals Block Type for a super-block in hType, and * huffman tree description in hufDesBuffer. * hufDesSize refers to the size of huffman tree description in bytes. * This metadata is populated in ZSTD_buildBlockEntropyStats_literals() */ typedef struct { symbolEncodingType_e hType; BYTE hufDesBuffer[ZSTD_MAX_HUF_HEADER_SIZE]; size_t hufDesSize; } ZSTD_hufCTablesMetadata_t; /** ZSTD_fseCTablesMetadata_t : * Stores symbol compression modes for a super-block in {ll, ol, ml}Type, and * fse tables in fseTablesBuffer. * fseTablesSize refers to the size of fse tables in bytes. * This metadata is populated in ZSTD_buildBlockEntropyStats_sequences() */ typedef struct { symbolEncodingType_e llType; symbolEncodingType_e ofType; symbolEncodingType_e mlType; BYTE fseTablesBuffer[ZSTD_MAX_FSE_HEADERS_SIZE]; size_t fseTablesSize; size_t lastCountSize; /* This is to account for bug in 1.3.4. More detail in ZSTD_entropyCompressSeqStore_internal() */ } ZSTD_fseCTablesMetadata_t; typedef struct { ZSTD_hufCTablesMetadata_t hufMetadata; ZSTD_fseCTablesMetadata_t fseMetadata; } ZSTD_entropyCTablesMetadata_t; /** ZSTD_buildBlockEntropyStats() : * Builds entropy for the block. * @return : 0 on success or error code */ size_t ZSTD_buildBlockEntropyStats( const seqStore_t* seqStorePtr, const ZSTD_entropyCTables_t* prevEntropy, ZSTD_entropyCTables_t* nextEntropy, const ZSTD_CCtx_params* cctxParams, ZSTD_entropyCTablesMetadata_t* entropyMetadata, void* workspace, size_t wkspSize); /********************************* * Compression internals structs * *********************************/ typedef struct { U32 off; /* Offset sumtype code for the match, using ZSTD_storeSeq() format */ U32 len; /* Raw length of match */ } ZSTD_match_t; typedef struct { U32 offset; /* Offset of sequence */ U32 litLength; /* Length of literals prior to match */ U32 matchLength; /* Raw length of match */ } rawSeq; typedef struct { rawSeq* seq; /* The start of the sequences */ size_t pos; /* The index in seq where reading stopped. pos <= size. */ size_t posInSequence; /* The position within the sequence at seq[pos] where reading stopped. posInSequence <= seq[pos].litLength + seq[pos].matchLength */ size_t size; /* The number of sequences. <= capacity. */ size_t capacity; /* The capacity starting from `seq` pointer */ } rawSeqStore_t; typedef struct { U32 idx; /* Index in array of ZSTD_Sequence */ U32 posInSequence; /* Position within sequence at idx */ size_t posInSrc; /* Number of bytes given by sequences provided so far */ } ZSTD_sequencePosition; UNUSED_ATTR static const rawSeqStore_t kNullRawSeqStore = {NULL, 0, 0, 0, 0}; typedef struct { 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_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 */ U32 matchLengthSum; /* nb of matchLength codes */ U32 offCodeSum; /* nb of offset codes */ U32 litSumBasePrice; /* to compare to log2(litfreq) */ U32 litLengthSumBasePrice; /* to compare to log2(llfreq) */ U32 matchLengthSumBasePrice;/* to compare to log2(mlfreq) */ U32 offCodeSumBasePrice; /* to compare to log2(offreq) */ ZSTD_OptPrice_e priceType; /* prices can be determined dynamically, or follow a pre-defined cost structure */ const ZSTD_entropyCTables_t* symbolCosts; /* pre-calculated dictionary statistics */ ZSTD_paramSwitch_e literalCompressionMode; } optState_t; typedef struct { ZSTD_entropyCTables_t entropy; U32 rep[ZSTD_REP_NUM]; } ZSTD_compressedBlockState_t; typedef struct { BYTE const* nextSrc; /* next block here to continue on current prefix */ BYTE const* base; /* All regular indexes relative to this position */ BYTE const* dictBase; /* extDict indexes relative to this position */ U32 dictLimit; /* below that point, need extDict */ U32 lowLimit; /* below that point, no more valid data */ U32 nbOverflowCorrections; /* Number of times overflow correction has run since * ZSTD_window_init(). Useful for debugging coredumps * and for ZSTD_WINDOW_OVERFLOW_CORRECT_FREQUENTLY. */ } ZSTD_window_t; #define ZSTD_WINDOW_START_INDEX 2 typedef struct ZSTD_matchState_t ZSTD_matchState_t; #define ZSTD_ROW_HASH_CACHE_SIZE 8 /* Size of prefetching hash cache for row-based matchfinder */ struct ZSTD_matchState_t { ZSTD_window_t window; /* State for window round buffer management */ U32 loadedDictEnd; /* index of end of dictionary, within context's referential. * When loadedDictEnd != 0, a dictionary is in use, and still valid. * This relies on a mechanism to set loadedDictEnd=0 when dictionary is no longer within distance. * Such mechanism is provided within ZSTD_window_enforceMaxDist() and ZSTD_checkDictValidity(). * When dict referential is copied into active context (i.e. not attached), * loadedDictEnd == dictSize, since referential starts from zero. */ U32 nextToUpdate; /* index from which to continue table update */ U32 hashLog3; /* dispatch table for matches of len==3 : larger == faster, more memory */ 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 reuse of tag table */ U32 hashSaltEntropy; /* For row-based matchFinder: collects entropy for salt generation */ U32* hashTable; U32* hashTable3; U32* chainTable; U32 forceNonContiguous; /* Non-zero if we should force non-contiguous load for the next window update. */ int dedicatedDictSearch; /* Indicates whether this matchState is using the * dedicated dictionary search structure. */ optState_t opt; /* optimal parser state */ const ZSTD_matchState_t* dictMatchState; ZSTD_compressionParameters cParams; const rawSeqStore_t* ldmSeqStore; /* Controls prefetching in some dictMatchState matchfinders. * This behavior is controlled from the cctx ms. * This parameter has no effect in the cdict ms. */ int prefetchCDictTables; /* When == 0, lazy match finders insert every position. * When != 0, lazy match finders only insert positions they search. * This allows them to skip much faster over incompressible data, * at a small cost to compression ratio. */ int lazySkipping; }; typedef struct { ZSTD_compressedBlockState_t* prevCBlock; ZSTD_compressedBlockState_t* nextCBlock; ZSTD_matchState_t matchState; } ZSTD_blockState_t; typedef struct { U32 offset; U32 checksum; } ldmEntry_t; typedef struct { BYTE const* split; U32 hash; U32 checksum; ldmEntry_t* bucket; } ldmMatchCandidate_t; #define LDM_BATCH_SIZE 64 typedef struct { ZSTD_window_t window; /* State for the window round buffer management */ ldmEntry_t* hashTable; U32 loadedDictEnd; BYTE* bucketOffsets; /* Next position in bucket to insert entry */ size_t splitIndices[LDM_BATCH_SIZE]; ldmMatchCandidate_t matchCandidates[LDM_BATCH_SIZE]; } ldmState_t; typedef struct { ZSTD_paramSwitch_e enableLdm; /* ZSTD_ps_enable to enable LDM. ZSTD_ps_auto by default */ U32 hashLog; /* Log size of hashTable */ U32 bucketSizeLog; /* Log bucket size for collision resolution, at most 8 */ U32 minMatchLength; /* Minimum match length */ U32 hashRateLog; /* Log number of entries to skip */ U32 windowLog; /* Window log for the LDM */ } ldmParams_t; typedef struct { int collectSequences; ZSTD_Sequence* seqStart; size_t seqIndex; size_t maxSequences; } SeqCollector; struct ZSTD_CCtx_params_s { ZSTD_format_e format; ZSTD_compressionParameters cParams; ZSTD_frameParameters fParams; int compressionLevel; int forceWindow; /* force back-references to respect limit of * 1< 63) ? ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength]; } /* ZSTD_MLcode() : * note : mlBase = matchLength - MINMATCH; * because it's the format it's stored in seqStore->sequences */ MEM_STATIC U32 ZSTD_MLcode(U32 mlBase) { static const BYTE ML_Code[128] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 36, 36, 37, 37, 37, 37, 38, 38, 38, 38, 38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42 }; static const U32 ML_deltaCode = 36; return (mlBase > 127) ? ZSTD_highbit32(mlBase) + ML_deltaCode : ML_Code[mlBase]; } /* ZSTD_cParam_withinBounds: * @return 1 if value is within cParam bounds, * 0 otherwise */ MEM_STATIC int ZSTD_cParam_withinBounds(ZSTD_cParameter cParam, int value) { ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); if (ZSTD_isError(bounds.error)) return 0; if (value < bounds.lowerBound) return 0; if (value > bounds.upperBound) return 0; return 1; } /* ZSTD_noCompressBlock() : * Writes uncompressed block to dst buffer from given src. * Returns the size of the block */ MEM_STATIC size_t ZSTD_noCompressBlock(void* dst, size_t dstCapacity, const void* src, size_t srcSize, U32 lastBlock) { U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw)<<1) + (U32)(srcSize << 3); DEBUGLOG(5, "ZSTD_noCompressBlock (srcSize=%zu, dstCapacity=%zu)", srcSize, dstCapacity); RETURN_ERROR_IF(srcSize + ZSTD_blockHeaderSize > dstCapacity, dstSize_tooSmall, "dst buf too small for uncompressed block"); MEM_writeLE24(dst, cBlockHeader24); ZSTD_memcpy((BYTE*)dst + ZSTD_blockHeaderSize, src, srcSize); return ZSTD_blockHeaderSize + srcSize; } MEM_STATIC size_t ZSTD_rleCompressBlock(void* dst, size_t dstCapacity, BYTE src, size_t srcSize, U32 lastBlock) { BYTE* const op = (BYTE*)dst; U32 const cBlockHeader = lastBlock + (((U32)bt_rle)<<1) + (U32)(srcSize << 3); RETURN_ERROR_IF(dstCapacity < 4, dstSize_tooSmall, ""); MEM_writeLE24(op, cBlockHeader); op[3] = src; return 4; } /* ZSTD_minGain() : * minimum compression required * to generate a compress block or a compressed literals section. * note : use same formula for both situations */ MEM_STATIC size_t ZSTD_minGain(size_t srcSize, ZSTD_strategy strat) { U32 const minlog = (strat>=ZSTD_btultra) ? (U32)(strat) - 1 : 6; ZSTD_STATIC_ASSERT(ZSTD_btultra == 8); assert(ZSTD_cParam_withinBounds(ZSTD_c_strategy, (int)strat)); return (srcSize >> minlog) + 2; } MEM_STATIC int ZSTD_literalsCompressionIsDisabled(const ZSTD_CCtx_params* cctxParams) { switch (cctxParams->literalCompressionMode) { case ZSTD_ps_enable: return 0; case ZSTD_ps_disable: return 1; default: assert(0 /* impossible: pre-validated */); ZSTD_FALLTHROUGH; case ZSTD_ps_auto: return (cctxParams->cParams.strategy == ZSTD_fast) && (cctxParams->cParams.targetLength > 0); } } /*! ZSTD_safecopyLiterals() : * memcpy() function that won't read beyond more than WILDCOPY_OVERLENGTH bytes past ilimit_w. * Only called when the sequence ends past ilimit_w, so it only needs to be optimized for single * large copies. */ static void ZSTD_safecopyLiterals(BYTE* op, BYTE const* ip, BYTE const* const iend, BYTE const* ilimit_w) { assert(iend > ilimit_w); if (ip <= ilimit_w) { ZSTD_wildcopy(op, ip, ilimit_w - ip, ZSTD_no_overlap); op += ilimit_w - ip; ip = ilimit_w; } while (ip < iend) *op++ = *ip++; } #define REPCODE1_TO_OFFBASE REPCODE_TO_OFFBASE(1) #define REPCODE2_TO_OFFBASE REPCODE_TO_OFFBASE(2) #define REPCODE3_TO_OFFBASE REPCODE_TO_OFFBASE(3) #define REPCODE_TO_OFFBASE(r) (assert((r)>=1), assert((r)<=ZSTD_REP_NUM), (r)) /* accepts IDs 1,2,3 */ #define OFFSET_TO_OFFBASE(o) (assert((o)>0), o + ZSTD_REP_NUM) #define OFFBASE_IS_OFFSET(o) ((o) > ZSTD_REP_NUM) #define OFFBASE_IS_REPCODE(o) ( 1 <= (o) && (o) <= ZSTD_REP_NUM) #define OFFBASE_TO_OFFSET(o) (assert(OFFBASE_IS_OFFSET(o)), (o) - ZSTD_REP_NUM) #define OFFBASE_TO_REPCODE(o) (assert(OFFBASE_IS_REPCODE(o)), (o)) /* returns ID 1,2,3 */ /*! ZSTD_storeSeq() : * Store a sequence (litlen, litPtr, offBase and matchLength) into seqStore_t. * @offBase : Users should employ macros REPCODE_TO_OFFBASE() and OFFSET_TO_OFFBASE(). * @matchLength : must be >= MINMATCH * Allowed to over-read literals up to litLimit. */ HINT_INLINE UNUSED_ATTR void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* literals, const BYTE* litLimit, U32 offBase, size_t matchLength) { BYTE const* const litLimit_w = litLimit - WILDCOPY_OVERLENGTH; BYTE const* const litEnd = literals + litLength; #if defined(DEBUGLEVEL) && (DEBUGLEVEL >= 6) static const BYTE* g_start = NULL; if (g_start==NULL) g_start = (const BYTE*)literals; /* note : index only works for compression within a single segment */ { U32 const pos = (U32)((const BYTE*)literals - g_start); DEBUGLOG(6, "Cpos%7u :%3u literals, match%4u bytes at offBase%7u", pos, (U32)litLength, (U32)matchLength, (U32)offBase); } #endif assert((size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart) < seqStorePtr->maxNbSeq); /* copy Literals */ assert(seqStorePtr->maxNbLit <= 128 KB); assert(seqStorePtr->lit + litLength <= seqStorePtr->litStart + seqStorePtr->maxNbLit); assert(literals + litLength <= litLimit); if (litEnd <= litLimit_w) { /* Common case we can use wildcopy. * First copy 16 bytes, because literals are likely short. */ ZSTD_STATIC_ASSERT(WILDCOPY_OVERLENGTH >= 16); ZSTD_copy16(seqStorePtr->lit, literals); if (litLength > 16) { ZSTD_wildcopy(seqStorePtr->lit+16, literals+16, (ptrdiff_t)litLength-16, ZSTD_no_overlap); } } else { ZSTD_safecopyLiterals(seqStorePtr->lit, literals, litEnd, litLimit_w); } seqStorePtr->lit += litLength; /* literal Length */ if (litLength>0xFFFF) { assert(seqStorePtr->longLengthType == ZSTD_llt_none); /* there can only be a single long length */ seqStorePtr->longLengthType = ZSTD_llt_literalLength; seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart); } seqStorePtr->sequences[0].litLength = (U16)litLength; /* match offset */ seqStorePtr->sequences[0].offBase = offBase; /* match Length */ assert(matchLength >= MINMATCH); { size_t const mlBase = matchLength - MINMATCH; if (mlBase>0xFFFF) { assert(seqStorePtr->longLengthType == ZSTD_llt_none); /* there can only be a single long length */ seqStorePtr->longLengthType = ZSTD_llt_matchLength; seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart); } seqStorePtr->sequences[0].mlBase = (U16)mlBase; } seqStorePtr->sequences++; } /* ZSTD_updateRep() : * updates in-place @rep (array of repeat offsets) * @offBase : sum-type, using numeric representation of ZSTD_storeSeq() */ MEM_STATIC void ZSTD_updateRep(U32 rep[ZSTD_REP_NUM], U32 const offBase, U32 const ll0) { if (OFFBASE_IS_OFFSET(offBase)) { /* full offset */ rep[2] = rep[1]; rep[1] = rep[0]; rep[0] = OFFBASE_TO_OFFSET(offBase); } else { /* repcode */ U32 const repCode = OFFBASE_TO_REPCODE(offBase) - 1 + ll0; if (repCode > 0) { /* note : if repCode==0, no change */ U32 const currentOffset = (repCode==ZSTD_REP_NUM) ? (rep[0] - 1) : rep[repCode]; rep[2] = (repCode >= 2) ? rep[1] : rep[2]; rep[1] = rep[0]; rep[0] = currentOffset; } else { /* repCode == 0 */ /* nothing to do */ } } } typedef struct repcodes_s { U32 rep[3]; } repcodes_t; MEM_STATIC repcodes_t ZSTD_newRep(U32 const rep[ZSTD_REP_NUM], U32 const offBase, U32 const ll0) { repcodes_t newReps; ZSTD_memcpy(&newReps, rep, sizeof(newReps)); ZSTD_updateRep(newReps.rep, offBase, ll0); return newReps; } /*-************************************* * Match length counter ***************************************/ MEM_STATIC size_t ZSTD_count(const BYTE* pIn, const BYTE* pMatch, const BYTE* const pInLimit) { const BYTE* const pStart = pIn; const BYTE* const pInLoopLimit = pInLimit - (sizeof(size_t)-1); if (pIn < pInLoopLimit) { { size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn); if (diff) return ZSTD_NbCommonBytes(diff); } pIn+=sizeof(size_t); pMatch+=sizeof(size_t); while (pIn < pInLoopLimit) { size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn); if (!diff) { pIn+=sizeof(size_t); pMatch+=sizeof(size_t); continue; } pIn += ZSTD_NbCommonBytes(diff); return (size_t)(pIn - pStart); } } if (MEM_64bits() && (pIn<(pInLimit-3)) && (MEM_read32(pMatch) == MEM_read32(pIn))) { pIn+=4; pMatch+=4; } if ((pIn<(pInLimit-1)) && (MEM_read16(pMatch) == MEM_read16(pIn))) { pIn+=2; pMatch+=2; } if ((pIn> (32-h) ; } MEM_STATIC size_t ZSTD_hash3Ptr(const void* ptr, U32 h) { return ZSTD_hash3(MEM_readLE32(ptr), h, 0); } /* only in zstd_opt.h */ MEM_STATIC size_t ZSTD_hash3PtrS(const void* ptr, U32 h, U32 s) { return ZSTD_hash3(MEM_readLE32(ptr), h, s); } static const U32 prime4bytes = 2654435761U; static U32 ZSTD_hash4(U32 u, U32 h, U32 s) { assert(h <= 32); return ((u * prime4bytes) ^ s) >> (32-h) ; } static size_t ZSTD_hash4Ptr(const void* ptr, U32 h) { return ZSTD_hash4(MEM_readLE32(ptr), h, 0); } static size_t ZSTD_hash4PtrS(const void* ptr, U32 h, U32 s) { return ZSTD_hash4(MEM_readLE32(ptr), h, s); } static const U64 prime5bytes = 889523592379ULL; static size_t ZSTD_hash5(U64 u, U32 h, U64 s) { assert(h <= 64); return (size_t)((((u << (64-40)) * prime5bytes) ^ s) >> (64-h)) ; } static size_t ZSTD_hash5Ptr(const void* p, U32 h) { return ZSTD_hash5(MEM_readLE64(p), h, 0); } static size_t ZSTD_hash5PtrS(const void* p, U32 h, U64 s) { return ZSTD_hash5(MEM_readLE64(p), h, s); } static const U64 prime6bytes = 227718039650203ULL; static size_t ZSTD_hash6(U64 u, U32 h, U64 s) { assert(h <= 64); return (size_t)((((u << (64-48)) * prime6bytes) ^ s) >> (64-h)) ; } static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h, 0); } static size_t ZSTD_hash6PtrS(const void* p, U32 h, U64 s) { return ZSTD_hash6(MEM_readLE64(p), h, s); } static const U64 prime7bytes = 58295818150454627ULL; static size_t ZSTD_hash7(U64 u, U32 h, U64 s) { assert(h <= 64); return (size_t)((((u << (64-56)) * prime7bytes) ^ s) >> (64-h)) ; } static size_t ZSTD_hash7Ptr(const void* p, U32 h) { return ZSTD_hash7(MEM_readLE64(p), h, 0); } static size_t ZSTD_hash7PtrS(const void* p, U32 h, U64 s) { return ZSTD_hash7(MEM_readLE64(p), h, s); } static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL; static size_t ZSTD_hash8(U64 u, U32 h, U64 s) { assert(h <= 64); return (size_t)((((u) * prime8bytes) ^ s) >> (64-h)) ; } static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h, 0); } static size_t ZSTD_hash8PtrS(const void* p, U32 h, U64 s) { return ZSTD_hash8(MEM_readLE64(p), h, s); } MEM_STATIC FORCE_INLINE_ATTR size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls) { /* Although some of these hashes do support hBits up to 64, some do not. * To be on the safe side, always avoid hBits > 32. */ assert(hBits <= 32); switch(mls) { default: case 4: return ZSTD_hash4Ptr(p, hBits); case 5: return ZSTD_hash5Ptr(p, hBits); case 6: return ZSTD_hash6Ptr(p, hBits); case 7: return ZSTD_hash7Ptr(p, hBits); case 8: return ZSTD_hash8Ptr(p, hBits); } } MEM_STATIC FORCE_INLINE_ATTR size_t ZSTD_hashPtrSalted(const void* p, U32 hBits, U32 mls, const U64 hashSalt) { /* Although some of these hashes do support hBits up to 64, some do not. * To be on the safe side, always avoid hBits > 32. */ assert(hBits <= 32); switch(mls) { default: case 4: return ZSTD_hash4PtrS(p, hBits, (U32)hashSalt); case 5: return ZSTD_hash5PtrS(p, hBits, hashSalt); case 6: return ZSTD_hash6PtrS(p, hBits, hashSalt); case 7: return ZSTD_hash7PtrS(p, hBits, hashSalt); case 8: return ZSTD_hash8PtrS(p, hBits, hashSalt); } } /** ZSTD_ipow() : * Return base^exponent. */ static U64 ZSTD_ipow(U64 base, U64 exponent) { U64 power = 1; while (exponent) { if (exponent & 1) power *= base; exponent >>= 1; base *= base; } return power; } #define ZSTD_ROLL_HASH_CHAR_OFFSET 10 /** ZSTD_rollingHash_append() : * Add the buffer to the hash value. */ static U64 ZSTD_rollingHash_append(U64 hash, void const* buf, size_t size) { BYTE const* istart = (BYTE const*)buf; size_t pos; for (pos = 0; pos < size; ++pos) { hash *= prime8bytes; hash += istart[pos] + ZSTD_ROLL_HASH_CHAR_OFFSET; } return hash; } /** ZSTD_rollingHash_compute() : * Compute the rolling hash value of the buffer. */ MEM_STATIC U64 ZSTD_rollingHash_compute(void const* buf, size_t size) { return ZSTD_rollingHash_append(0, buf, size); } /** ZSTD_rollingHash_primePower() : * Compute the primePower to be passed to ZSTD_rollingHash_rotate() for a hash * over a window of length bytes. */ MEM_STATIC U64 ZSTD_rollingHash_primePower(U32 length) { return ZSTD_ipow(prime8bytes, length - 1); } /** ZSTD_rollingHash_rotate() : * Rotate the rolling hash by one byte. */ MEM_STATIC U64 ZSTD_rollingHash_rotate(U64 hash, BYTE toRemove, BYTE toAdd, U64 primePower) { hash -= (toRemove + ZSTD_ROLL_HASH_CHAR_OFFSET) * primePower; hash *= prime8bytes; hash += toAdd + ZSTD_ROLL_HASH_CHAR_OFFSET; return hash; } /*-************************************* * Round buffer management ***************************************/ #if (ZSTD_WINDOWLOG_MAX_64 > 31) # error "ZSTD_WINDOWLOG_MAX is too large : would overflow ZSTD_CURRENT_MAX" #endif /* Max current allowed */ #define ZSTD_CURRENT_MAX ((3U << 29) + (1U << ZSTD_WINDOWLOG_MAX)) /* Maximum chunk size before overflow correction needs to be called again */ #define ZSTD_CHUNKSIZE_MAX \ ( ((U32)-1) /* Maximum ending current index */ \ - ZSTD_CURRENT_MAX) /* Maximum beginning lowLimit */ /** * ZSTD_window_clear(): * Clears the window containing the history by simply setting it to empty. */ MEM_STATIC void ZSTD_window_clear(ZSTD_window_t* window) { size_t const endT = (size_t)(window->nextSrc - window->base); U32 const end = (U32)endT; window->lowLimit = end; window->dictLimit = end; } MEM_STATIC U32 ZSTD_window_isEmpty(ZSTD_window_t const window) { return window.dictLimit == ZSTD_WINDOW_START_INDEX && window.lowLimit == ZSTD_WINDOW_START_INDEX && (window.nextSrc - window.base) == ZSTD_WINDOW_START_INDEX; } /** * ZSTD_window_hasExtDict(): * Returns non-zero if the window has a non-empty extDict. */ MEM_STATIC U32 ZSTD_window_hasExtDict(ZSTD_window_t const window) { return window.lowLimit < window.dictLimit; } /** * ZSTD_matchState_dictMode(): * Inspects the provided matchState and figures out what dictMode should be * passed to the compressor. */ MEM_STATIC ZSTD_dictMode_e ZSTD_matchState_dictMode(const ZSTD_matchState_t *ms) { return ZSTD_window_hasExtDict(ms->window) ? ZSTD_extDict : ms->dictMatchState != NULL ? (ms->dictMatchState->dedicatedDictSearch ? ZSTD_dedicatedDictSearch : ZSTD_dictMatchState) : ZSTD_noDict; } /* Defining this macro to non-zero tells zstd to run the overflow correction * code much more frequently. This is very inefficient, and should only be * used for tests and fuzzers. */ #ifndef ZSTD_WINDOW_OVERFLOW_CORRECT_FREQUENTLY # ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION # define ZSTD_WINDOW_OVERFLOW_CORRECT_FREQUENTLY 1 # else # define ZSTD_WINDOW_OVERFLOW_CORRECT_FREQUENTLY 0 # endif #endif /** * ZSTD_window_canOverflowCorrect(): * Returns non-zero if the indices are large enough for overflow correction * to work correctly without impacting compression ratio. */ MEM_STATIC U32 ZSTD_window_canOverflowCorrect(ZSTD_window_t const window, U32 cycleLog, U32 maxDist, U32 loadedDictEnd, void const* src) { U32 const cycleSize = 1u << cycleLog; U32 const curr = (U32)((BYTE const*)src - window.base); U32 const minIndexToOverflowCorrect = cycleSize + MAX(maxDist, cycleSize) + ZSTD_WINDOW_START_INDEX; /* Adjust the min index to backoff the overflow correction frequency, * so we don't waste too much CPU in overflow correction. If this * computation overflows we don't really care, we just need to make * sure it is at least minIndexToOverflowCorrect. */ U32 const adjustment = window.nbOverflowCorrections + 1; U32 const adjustedIndex = MAX(minIndexToOverflowCorrect * adjustment, minIndexToOverflowCorrect); U32 const indexLargeEnough = curr > adjustedIndex; /* Only overflow correct early if the dictionary is invalidated already, * so we don't hurt compression ratio. */ U32 const dictionaryInvalidated = curr > maxDist + loadedDictEnd; return indexLargeEnough && dictionaryInvalidated; } /** * ZSTD_window_needOverflowCorrection(): * Returns non-zero if the indices are getting too large and need overflow * protection. */ MEM_STATIC U32 ZSTD_window_needOverflowCorrection(ZSTD_window_t const window, U32 cycleLog, U32 maxDist, U32 loadedDictEnd, void const* src, void const* srcEnd) { U32 const curr = (U32)((BYTE const*)srcEnd - window.base); if (ZSTD_WINDOW_OVERFLOW_CORRECT_FREQUENTLY) { if (ZSTD_window_canOverflowCorrect(window, cycleLog, maxDist, loadedDictEnd, src)) { return 1; } } return curr > ZSTD_CURRENT_MAX; } /** * ZSTD_window_correctOverflow(): * Reduces the indices to protect from index overflow. * Returns the correction made to the indices, which must be applied to every * stored index. * * 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 ZSTD_ALLOW_POINTER_OVERFLOW_ATTR U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog, U32 maxDist, void const* src) { /* preemptive overflow correction: * 1. correction is large enough: * lowLimit > (3<<29) ==> current > 3<<29 + 1< (3<<29 + 1< (3<<29) - (1< (3<<29) - (1<<30) (NOTE: chainLog <= 30) * > 1<<29 * * 2. (ip+ZSTD_CHUNKSIZE_MAX - cctx->base) doesn't overflow: * After correction, current is less than (1<base < 1<<32. * 3. (cctx->lowLimit + 1< 3<<29 + 1<base); U32 const currentCycle = curr & cycleMask; /* Ensure newCurrent - maxDist >= ZSTD_WINDOW_START_INDEX. */ U32 const currentCycleCorrection = currentCycle < ZSTD_WINDOW_START_INDEX ? MAX(cycleSize, ZSTD_WINDOW_START_INDEX) : 0; U32 const newCurrent = currentCycle + currentCycleCorrection + MAX(maxDist, cycleSize); U32 const correction = curr - newCurrent; /* maxDist must be a power of two so that: * (newCurrent & cycleMask) == (curr & cycleMask) * This is required to not corrupt the chains / binary tree. */ assert((maxDist & (maxDist - 1)) == 0); assert((curr & cycleMask) == (newCurrent & cycleMask)); assert(curr > newCurrent); if (!ZSTD_WINDOW_OVERFLOW_CORRECT_FREQUENTLY) { /* Loose bound, should be around 1<<29 (see above) */ assert(correction > 1<<28); } window->base += correction; window->dictBase += correction; if (window->lowLimit < correction + ZSTD_WINDOW_START_INDEX) { window->lowLimit = ZSTD_WINDOW_START_INDEX; } else { window->lowLimit -= correction; } if (window->dictLimit < correction + ZSTD_WINDOW_START_INDEX) { window->dictLimit = ZSTD_WINDOW_START_INDEX; } else { window->dictLimit -= correction; } /* Ensure we can still reference the full window. */ assert(newCurrent >= maxDist); assert(newCurrent - maxDist >= ZSTD_WINDOW_START_INDEX); /* Ensure that lowLimit and dictLimit didn't underflow. */ assert(window->lowLimit <= newCurrent); assert(window->dictLimit <= newCurrent); ++window->nbOverflowCorrections; DEBUGLOG(4, "Correction of 0x%x bytes to lowLimit=0x%x", correction, window->lowLimit); return correction; } /** * ZSTD_window_enforceMaxDist(): * Updates lowLimit so that: * (srcEnd - base) - lowLimit == maxDist + loadedDictEnd * * It ensures index is valid as long as index >= lowLimit. * This must be called before a block compression call. * * loadedDictEnd is only defined if a dictionary is in use for current compression. * As the name implies, loadedDictEnd represents the index at end of dictionary. * The value lies within context's referential, it can be directly compared to blockEndIdx. * * If loadedDictEndPtr is NULL, no dictionary is in use, and we use loadedDictEnd == 0. * If loadedDictEndPtr is not NULL, we set it to zero after updating lowLimit. * This is because dictionaries are allowed to be referenced fully * as long as the last byte of the dictionary is in the window. * Once input has progressed beyond window size, dictionary cannot be referenced anymore. * * In normal dict mode, the dictionary lies between lowLimit and dictLimit. * In dictMatchState mode, lowLimit and dictLimit are the same, * and the dictionary is below them. * forceWindow and dictMatchState are therefore incompatible. */ MEM_STATIC void ZSTD_window_enforceMaxDist(ZSTD_window_t* window, const void* blockEnd, U32 maxDist, U32* loadedDictEndPtr, const ZSTD_matchState_t** dictMatchStatePtr) { U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base); U32 const loadedDictEnd = (loadedDictEndPtr != NULL) ? *loadedDictEndPtr : 0; DEBUGLOG(5, "ZSTD_window_enforceMaxDist: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u", (unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd); /* - When there is no dictionary : loadedDictEnd == 0. In which case, the test (blockEndIdx > maxDist) is merely to avoid overflowing next operation `newLowLimit = blockEndIdx - maxDist`. - When there is a standard dictionary : Index referential is copied from the dictionary, which means it starts from 0. In which case, loadedDictEnd == dictSize, and it makes sense to compare `blockEndIdx > maxDist + dictSize` since `blockEndIdx` also starts from zero. - When there is an attached dictionary : loadedDictEnd is expressed within the referential of the context, so it can be directly compared against blockEndIdx. */ if (blockEndIdx > maxDist + loadedDictEnd) { U32 const newLowLimit = blockEndIdx - maxDist; if (window->lowLimit < newLowLimit) window->lowLimit = newLowLimit; if (window->dictLimit < window->lowLimit) { DEBUGLOG(5, "Update dictLimit to match lowLimit, from %u to %u", (unsigned)window->dictLimit, (unsigned)window->lowLimit); window->dictLimit = window->lowLimit; } /* On reaching window size, dictionaries are invalidated */ if (loadedDictEndPtr) *loadedDictEndPtr = 0; if (dictMatchStatePtr) *dictMatchStatePtr = NULL; } } /* Similar to ZSTD_window_enforceMaxDist(), * but only invalidates dictionary * when input progresses beyond window size. * assumption : loadedDictEndPtr and dictMatchStatePtr are valid (non NULL) * loadedDictEnd uses same referential as window->base * maxDist is the window size */ MEM_STATIC void ZSTD_checkDictValidity(const ZSTD_window_t* window, const void* blockEnd, U32 maxDist, U32* loadedDictEndPtr, const ZSTD_matchState_t** dictMatchStatePtr) { assert(loadedDictEndPtr != NULL); assert(dictMatchStatePtr != NULL); { U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base); U32 const loadedDictEnd = *loadedDictEndPtr; DEBUGLOG(5, "ZSTD_checkDictValidity: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u", (unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd); assert(blockEndIdx >= loadedDictEnd); if (blockEndIdx > loadedDictEnd + maxDist || loadedDictEnd != window->dictLimit) { /* On reaching window size, dictionaries are invalidated. * For simplification, if window size is reached anywhere within next block, * the dictionary is invalidated for the full block. * * We also have to invalidate the dictionary if ZSTD_window_update() has detected * non-contiguous segments, which means that loadedDictEnd != window->dictLimit. * loadedDictEnd may be 0, if forceWindow is true, but in that case we never use * dictMatchState, so setting it to NULL is not a problem. */ DEBUGLOG(6, "invalidating dictionary for current block (distance > windowSize)"); *loadedDictEndPtr = 0; *dictMatchStatePtr = NULL; } else { if (*loadedDictEndPtr != 0) { DEBUGLOG(6, "dictionary considered valid for current block"); } } } } MEM_STATIC void ZSTD_window_init(ZSTD_window_t* window) { ZSTD_memset(window, 0, sizeof(*window)); window->base = (BYTE const*)" "; window->dictBase = (BYTE const*)" "; ZSTD_STATIC_ASSERT(ZSTD_DUBT_UNSORTED_MARK < ZSTD_WINDOW_START_INDEX); /* Start above ZSTD_DUBT_UNSORTED_MARK */ window->dictLimit = ZSTD_WINDOW_START_INDEX; /* start from >0, so that 1st position is valid */ window->lowLimit = ZSTD_WINDOW_START_INDEX; /* it ensures first and later CCtx usages compress the same */ window->nextSrc = window->base + ZSTD_WINDOW_START_INDEX; /* see issue #1241 */ window->nbOverflowCorrections = 0; } /** * ZSTD_window_update(): * Updates the window by appending [src, src + srcSize) to the window. * If it is not contiguous, the current prefix becomes the extDict, and we * forget about the extDict. Handles overlap of the prefix and extDict. * Returns non-zero if the segment is contiguous. */ MEM_STATIC ZSTD_ALLOW_POINTER_OVERFLOW_ATTR U32 ZSTD_window_update(ZSTD_window_t* window, void const* src, size_t srcSize, int forceNonContiguous) { BYTE const* const ip = (BYTE const*)src; U32 contiguous = 1; DEBUGLOG(5, "ZSTD_window_update"); if (srcSize == 0) return contiguous; assert(window->base != NULL); assert(window->dictBase != NULL); /* Check if blocks follow each other */ if (src != window->nextSrc || forceNonContiguous) { /* not contiguous */ size_t const distanceFromBase = (size_t)(window->nextSrc - window->base); DEBUGLOG(5, "Non contiguous blocks, new segment starts at %u", window->dictLimit); window->lowLimit = window->dictLimit; assert(distanceFromBase == (size_t)(U32)distanceFromBase); /* should never overflow */ window->dictLimit = (U32)distanceFromBase; window->dictBase = window->base; window->base = ip - distanceFromBase; /* ms->nextToUpdate = window->dictLimit; */ if (window->dictLimit - window->lowLimit < HASH_READ_SIZE) window->lowLimit = window->dictLimit; /* too small extDict */ contiguous = 0; } window->nextSrc = ip + srcSize; /* if input and dictionary overlap : reduce dictionary (area presumed modified by input) */ if ( (ip+srcSize > window->dictBase + window->lowLimit) & (ip < window->dictBase + window->dictLimit)) { ptrdiff_t const highInputIdx = (ip + srcSize) - window->dictBase; U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)window->dictLimit) ? window->dictLimit : (U32)highInputIdx; window->lowLimit = lowLimitMax; DEBUGLOG(5, "Overlapping extDict and input : new lowLimit = %u", window->lowLimit); } return contiguous; } /** * Returns the lowest allowed match index. It may either be in the ext-dict or the prefix. */ MEM_STATIC U32 ZSTD_getLowestMatchIndex(const ZSTD_matchState_t* ms, U32 curr, unsigned windowLog) { U32 const maxDistance = 1U << windowLog; U32 const lowestValid = ms->window.lowLimit; U32 const withinWindow = (curr - lowestValid > maxDistance) ? curr - maxDistance : lowestValid; U32 const isDictionary = (ms->loadedDictEnd != 0); /* When using a dictionary the entire dictionary is valid if a single byte of the dictionary * is within the window. We invalidate the dictionary (and set loadedDictEnd to 0) when it isn't * valid for the entire block. So this check is sufficient to find the lowest valid match index. */ U32 const matchLowest = isDictionary ? lowestValid : withinWindow; return matchLowest; } /** * Returns the lowest allowed match index in the prefix. */ MEM_STATIC U32 ZSTD_getLowestPrefixIndex(const ZSTD_matchState_t* ms, U32 curr, unsigned windowLog) { U32 const maxDistance = 1U << windowLog; U32 const lowestValid = ms->window.dictLimit; U32 const withinWindow = (curr - lowestValid > maxDistance) ? curr - maxDistance : lowestValid; U32 const isDictionary = (ms->loadedDictEnd != 0); /* When computing the lowest prefix index we need to take the dictionary into account to handle * the edge case where the dictionary and the source are contiguous in memory. */ U32 const matchLowest = isDictionary ? lowestValid : withinWindow; return matchLowest; } /* debug functions */ #if (DEBUGLEVEL>=2) MEM_STATIC double ZSTD_fWeight(U32 rawStat) { U32 const fp_accuracy = 8; U32 const fp_multiplier = (1 << fp_accuracy); U32 const newStat = rawStat + 1; U32 const hb = ZSTD_highbit32(newStat); U32 const BWeight = hb * fp_multiplier; U32 const FWeight = (newStat << fp_accuracy) >> hb; U32 const weight = BWeight + FWeight; assert(hb + fp_accuracy < 31); return (double)weight / fp_multiplier; } /* display a table content, * listing each element, its frequency, and its predicted bit cost */ MEM_STATIC void ZSTD_debugTable(const U32* table, U32 max) { unsigned u, sum; for (u=0, sum=0; u<=max; u++) sum += table[u]; DEBUGLOG(2, "total nb elts: %u", sum); for (u=0; u<=max; u++) { DEBUGLOG(2, "%2u: %5u (%.2f)", u, table[u], ZSTD_fWeight(sum) - ZSTD_fWeight(table[u]) ); } } #endif /* Short Cache */ /* Normally, zstd matchfinders follow this flow: * 1. Compute hash at ip * 2. Load index from hashTable[hash] * 3. Check if *ip == *(base + index) * In dictionary compression, loading *(base + index) is often an L2 or even L3 miss. * * Short cache is an optimization which allows us to avoid step 3 most of the time * when the data doesn't actually match. With short cache, the flow becomes: * 1. Compute (hash, currentTag) at ip. currentTag is an 8-bit independent hash at ip. * 2. Load (index, matchTag) from hashTable[hash]. See ZSTD_writeTaggedIndex to understand how this works. * 3. Only if currentTag == matchTag, check *ip == *(base + index). Otherwise, continue. * * Currently, short cache is only implemented in CDict hashtables. Thus, its use is limited to * dictMatchState matchfinders. */ #define ZSTD_SHORT_CACHE_TAG_BITS 8 #define ZSTD_SHORT_CACHE_TAG_MASK ((1u << ZSTD_SHORT_CACHE_TAG_BITS) - 1) /* Helper function for ZSTD_fillHashTable and ZSTD_fillDoubleHashTable. * Unpacks hashAndTag into (hash, tag), then packs (index, tag) into hashTable[hash]. */ MEM_STATIC void ZSTD_writeTaggedIndex(U32* const hashTable, size_t hashAndTag, U32 index) { size_t const hash = hashAndTag >> ZSTD_SHORT_CACHE_TAG_BITS; U32 const tag = (U32)(hashAndTag & ZSTD_SHORT_CACHE_TAG_MASK); assert(index >> (32 - ZSTD_SHORT_CACHE_TAG_BITS) == 0); hashTable[hash] = (index << ZSTD_SHORT_CACHE_TAG_BITS) | tag; } /* Helper function for short cache matchfinders. * Unpacks tag1 and tag2 from lower bits of packedTag1 and packedTag2, then checks if the tags match. */ MEM_STATIC int ZSTD_comparePackedTags(size_t packedTag1, size_t packedTag2) { U32 const tag1 = packedTag1 & ZSTD_SHORT_CACHE_TAG_MASK; U32 const tag2 = packedTag2 & ZSTD_SHORT_CACHE_TAG_MASK; return tag1 == tag2; } #if defined (__cplusplus) } #endif /* =============================================================== * Shared internal declarations * These prototypes may be called from sources not in lib/compress * =============================================================== */ /* ZSTD_loadCEntropy() : * dict : must point at beginning of a valid zstd dictionary. * return : size of dictionary header (size of magic number + dict ID + entropy tables) * assumptions : magic number supposed already checked * and dictSize >= 8 */ size_t ZSTD_loadCEntropy(ZSTD_compressedBlockState_t* bs, void* workspace, const void* const dict, size_t dictSize); void ZSTD_reset_compressedBlockState(ZSTD_compressedBlockState_t* bs); /* ============================================================== * Private declarations * These prototypes shall only be called from within lib/compress * ============================================================== */ /* ZSTD_getCParamsFromCCtxParams() : * cParams are built depending on compressionLevel, src size hints, * LDM and manually set compression parameters. * Note: srcSizeHint == 0 means 0! */ ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams( const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize, ZSTD_cParamMode_e mode); /*! ZSTD_initCStream_internal() : * Private use only. Init streaming operation. * expects params to be valid. * must receive dict, or cdict, or none, but not both. * @return : 0, or an error code */ size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs, const void* dict, size_t dictSize, const ZSTD_CDict* cdict, const ZSTD_CCtx_params* params, unsigned long long pledgedSrcSize); void ZSTD_resetSeqStore(seqStore_t* ssPtr); /*! ZSTD_getCParamsFromCDict() : * as the name implies */ ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict); /* ZSTD_compressBegin_advanced_internal() : * Private use only. To be called from zstdmt_compress.c. */ size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_dictContentType_e dictContentType, ZSTD_dictTableLoadMethod_e dtlm, const ZSTD_CDict* cdict, const ZSTD_CCtx_params* params, unsigned long long pledgedSrcSize); /* ZSTD_compress_advanced_internal() : * Private use only. To be called from zstdmt_compress.c. */ size_t ZSTD_compress_advanced_internal(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize, const void* dict,size_t dictSize, const ZSTD_CCtx_params* params); /* ZSTD_writeLastEmptyBlock() : * output an empty Block with end-of-frame mark to complete a frame * @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h)) * or an error code if `dstCapacity` is too small ( 1 */ U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat); /** ZSTD_CCtx_trace() : * Trace the end of a compression call. */ void ZSTD_CCtx_trace(ZSTD_CCtx* cctx, size_t extraCSize); /* Returns 0 on success, and a ZSTD_error otherwise. This function scans through an array of * ZSTD_Sequence, storing the sequences it finds, until it reaches a block delimiter. * Note that the block delimiter must include the last literals of the block. */ size_t ZSTD_copySequencesToSeqStoreExplicitBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos, const ZSTD_Sequence* const inSeqs, size_t inSeqsSize, const void* src, size_t blockSize, ZSTD_paramSwitch_e externalRepSearch); /* Returns the number of bytes to move the current read position back by. * Only non-zero if we ended up splitting a sequence. * Otherwise, it may return a ZSTD error if something went wrong. * * This function will attempt to scan through blockSize bytes * represented by the sequences in @inSeqs, * storing any (partial) sequences. * * Occasionally, we may want to change the actual number of bytes we consumed from inSeqs to * avoid splitting a match, or to avoid splitting a match such that it would produce a match * smaller than MINMATCH. In this case, we return the number of bytes that we didn't read from this block. */ size_t ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos, 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 * deprecation warnings. These functions are exactly equivalent to * their public variants, but avoid the deprecation warnings. * =============================================================== */ size_t ZSTD_compressBegin_usingCDict_deprecated(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict); size_t ZSTD_compressContinue_public(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize); size_t ZSTD_compressEnd_public(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize); size_t ZSTD_compressBlock_deprecated(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize); #endif /* ZSTD_COMPRESS_H */