ddidderr/zstd-rs
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

initial commit

Browse Source
This commit is contained in:
Danielle Rozenblit
2023-01-04 13:01:54 -08:00
parent ef566c8d68
commit 908e812733
328 changed files with 3340 additions and 1426 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/compress/clevels.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/fse_compress.c
+1 -1
View File
@@ -1,6 +1,6 @@
/* ******************************************************************
* FSE : Finite State Entropy encoder
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
*
* You can contact the author at :
* - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy
lib/compress/hist.c
+1 -1
View File
@@ -1,7 +1,7 @@
/* ******************************************************************
* hist : Histogram functions
* part of Finite State Entropy project
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
*
* You can contact the author at :
* - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy
lib/compress/hist.h
+1 -1
View File
@@ -1,7 +1,7 @@
/* ******************************************************************
* hist : Histogram functions
* part of Finite State Entropy project
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
*
* You can contact the author at :
* - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy
lib/compress/huf_compress.c
+36 -18
View File
@@ -1,6 +1,6 @@
/* ******************************************************************
* Huffman encoder, part of New Generation Entropy library
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
*
* You can contact the author at :
* - FSE+HUF source repository : https://github.com/Cyan4973/FiniteStateEntropy
@@ -386,7 +386,7 @@ static U32 HUF_setMaxHeight(nodeElt* huffNode, U32 lastNonNull, U32 targetNbBits
/* renorm totalCost from 2^largestBits to 2^targetNbBits
* note : totalCost is necessarily a multiple of baseCost */
assert((totalCost & (baseCost - 1)) == 0);
assert(((U32)totalCost & (baseCost - 1)) == 0);
totalCost >>= (largestBits - targetNbBits);
assert(totalCost > 0);
@@ -1253,41 +1253,59 @@ unsigned HUF_minTableLog(unsigned symbolCardinality)
return minBitsSymbols;
}
unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, void* workSpace, size_t wkspSize, HUF_CElt* table, const unsigned* count, HUF_depth_mode depthMode)
unsigned HUF_optimalTableLog(
unsigned maxTableLog,
size_t srcSize,
unsigned maxSymbolValue,
void* workSpace, size_t wkspSize,
HUF_CElt* table,
const unsigned* count,
HUF_depth_mode depthMode)
{
unsigned optLog = FSE_optimalTableLog_internal(maxTableLog, srcSize, maxSymbolValue, 1);
assert(srcSize > 1); /* Not supported, RLE should be used instead */
assert(wkspSize >= sizeof(HUF_buildCTable_wksp_tables));
if (depthMode == HUF_depth_optimal) { /** Test valid depths and return optimal **/
BYTE* dst = (BYTE*)workSpace + sizeof(HUF_WriteCTableWksp);
if (depthMode != HUF_depth_optimal) {
/* cheap evaluation, based on FSE */
return FSE_optimalTableLog_internal(maxTableLog, srcSize, maxSymbolValue, 1);
}
{ BYTE* dst = (BYTE*)workSpace + sizeof(HUF_WriteCTableWksp);
size_t dstSize = wkspSize - sizeof(HUF_WriteCTableWksp);
size_t optSize = ((size_t) ~0);
unsigned huffLog;
size_t maxBits, hSize, newSize;
const unsigned symbolCardinality = HUF_cardinality(count, maxSymbolValue);
const unsigned minTableLog = HUF_minTableLog(symbolCardinality);
size_t optSize = ((size_t) ~0) - 1;
unsigned optLog = maxTableLog, optLogGuess;
if (wkspSize < sizeof(HUF_buildCTable_wksp_tables)) return optLog;
DEBUGLOG(6, "HUF_optimalTableLog: probing huf depth (srcSize=%zu)", srcSize);
for (huffLog = HUF_minTableLog(symbolCardinality); huffLog <= maxTableLog; huffLog++) {
maxBits = HUF_buildCTable_wksp(table, count,
maxSymbolValue, huffLog,
workSpace, wkspSize);
/* 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;
hSize = HUF_writeCTable_wksp(dst, dstSize, table, maxSymbolValue, (U32)maxBits,
workSpace, wkspSize);
if (maxBits < optLogGuess && optLogGuess > minTableLog) break;
hSize = HUF_writeCTable_wksp(dst, dstSize, table, maxSymbolValue, (U32)maxBits, workSpace, wkspSize);
if (ERR_isError(hSize)) continue;
newSize = HUF_estimateCompressedSize(table, count, maxSymbolValue) + hSize;
if (newSize > optSize + 1) {
break;
}
if (newSize < optSize) {
optSize = newSize;
optLog = huffLog;
optLog = optLogGuess;
}
}
assert(optLog <= HUF_TABLELOG_MAX);
return optLog;
}
assert(optLog <= HUF_TABLELOG_MAX);
return optLog;
}
/* HUF_compress_internal() :
lib/compress/zstd_compress.c
+478 -238
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
@@ -59,14 +59,17 @@
* Helper functions
***************************************/
/* ZSTD_compressBound()
* Note that the result from this function is only compatible with the "normal"
* full-block strategy.
* When there are a lot of small blocks due to frequent flush in streaming mode
* the overhead of headers can make the compressed data to be larger than the
* return value of ZSTD_compressBound().
* Note that the result from this function is only valid for
* the one-pass compression functions.
* When employing the streaming mode,
* if flushes are frequently altering the size of blocks,
* the overhead from block headers can make the compressed data larger
* than the return value of ZSTD_compressBound().
*/
size_t ZSTD_compressBound(size_t srcSize) {
return ZSTD_COMPRESSBOUND(srcSize);
size_t const r = ZSTD_COMPRESSBOUND(srcSize);
if (r==0) return ERROR(srcSize_wrong);
return r;
}
@@ -275,6 +278,16 @@ static ZSTD_paramSwitch_e ZSTD_resolveEnableLdm(ZSTD_paramSwitch_e mode,
return (cParams->strategy >= ZSTD_btopt && cParams->windowLog >= 27) ? ZSTD_ps_enable : ZSTD_ps_disable;
}
/* Enables validation for external sequences in debug builds. */
static int ZSTD_resolveExternalSequenceValidation(int mode) {
#if defined(DEBUGLEVEL) && (DEBUGLEVEL>=2)
(void)mode;
return 1;
#else
return mode;
#endif
}
/* Returns 1 if compression parameters are such that CDict hashtable and chaintable indices are tagged.
* If so, the tags need to be removed in ZSTD_resetCCtx_byCopyingCDict. */
static int ZSTD_CDictIndicesAreTagged(const ZSTD_compressionParameters* const cParams) {
@@ -298,6 +311,7 @@ static ZSTD_CCtx_params ZSTD_makeCCtxParamsFromCParams(
}
cctxParams.useBlockSplitter = ZSTD_resolveBlockSplitterMode(cctxParams.useBlockSplitter, &cParams);
cctxParams.useRowMatchFinder = ZSTD_resolveRowMatchFinderMode(cctxParams.useRowMatchFinder, &cParams);
cctxParams.validateSequences = ZSTD_resolveExternalSequenceValidation(cctxParams.validateSequences);
assert(!ZSTD_checkCParams(cParams));
return cctxParams;
}
@@ -343,10 +357,13 @@ size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel)
#define ZSTD_NO_CLEVEL 0
/**
* Initializes the cctxParams from params and compressionLevel.
* Initializes `cctxParams` from `params` and `compressionLevel`.
* @param compressionLevel If params are derived from a compression level then that compression level, otherwise ZSTD_NO_CLEVEL.
*/
static void ZSTD_CCtxParams_init_internal(ZSTD_CCtx_params* cctxParams, ZSTD_parameters const* params, int compressionLevel)
static void
ZSTD_CCtxParams_init_internal(ZSTD_CCtx_params* cctxParams,
const ZSTD_parameters* params,
int compressionLevel)
{
assert(!ZSTD_checkCParams(params->cParams));
ZSTD_memset(cctxParams, 0, sizeof(*cctxParams));
@@ -359,6 +376,7 @@ static void ZSTD_CCtxParams_init_internal(ZSTD_CCtx_params* cctxParams, ZSTD_par
cctxParams->useRowMatchFinder = ZSTD_resolveRowMatchFinderMode(cctxParams->useRowMatchFinder, &params->cParams);
cctxParams->useBlockSplitter = ZSTD_resolveBlockSplitterMode(cctxParams->useBlockSplitter, &params->cParams);
cctxParams->ldmParams.enableLdm = ZSTD_resolveEnableLdm(cctxParams->ldmParams.enableLdm, &params->cParams);
cctxParams->validateSequences = ZSTD_resolveExternalSequenceValidation(cctxParams->validateSequences);
DEBUGLOG(4, "ZSTD_CCtxParams_init_internal: useRowMatchFinder=%d, useBlockSplitter=%d ldm=%d",
cctxParams->useRowMatchFinder, cctxParams->useBlockSplitter, cctxParams->ldmParams.enableLdm);
}
@@ -373,7 +391,7 @@ size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_paramete
/**
* Sets cctxParams' cParams and fParams from params, but otherwise leaves them alone.
* @param param Validated zstd parameters.
* @param params Validated zstd parameters.
*/
static void ZSTD_CCtxParams_setZstdParams(
ZSTD_CCtx_params* cctxParams, const ZSTD_parameters* params)
@@ -581,6 +599,16 @@ ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_cParameter param)
bounds.upperBound = (int)ZSTD_ps_disable;
return bounds;
case ZSTD_c_enableMatchFinderFallback:
bounds.lowerBound = 0;
bounds.upperBound = 1;
return bounds;
case ZSTD_c_maxBlockSize:
bounds.lowerBound = 1;
bounds.upperBound = ZSTD_BLOCKSIZE_MAX;
return bounds;
default:
bounds.error = ERROR(parameter_unsupported);
return bounds;
@@ -646,6 +674,8 @@ static int ZSTD_isUpdateAuthorized(ZSTD_cParameter param)
case ZSTD_c_useRowMatchFinder:
case ZSTD_c_deterministicRefPrefix:
case ZSTD_c_prefetchCDictTables:
case ZSTD_c_enableMatchFinderFallback:
case ZSTD_c_maxBlockSize:
default:
return 0;
}
@@ -658,7 +688,7 @@ size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value)
if (ZSTD_isUpdateAuthorized(param)) {
cctx->cParamsChanged = 1;
} else {
RETURN_ERROR(stage_wrong, "can only set params in ctx init stage");
RETURN_ERROR(stage_wrong, "can only set params in cctx init stage");
} }
switch(param)
@@ -702,6 +732,8 @@ size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value)
case ZSTD_c_useRowMatchFinder:
case ZSTD_c_deterministicRefPrefix:
case ZSTD_c_prefetchCDictTables:
case ZSTD_c_enableMatchFinderFallback:
case ZSTD_c_maxBlockSize:
break;
default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
@@ -793,14 +825,14 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
case ZSTD_c_forceAttachDict : {
const ZSTD_dictAttachPref_e pref = (ZSTD_dictAttachPref_e)value;
BOUNDCHECK(ZSTD_c_forceAttachDict, pref);
BOUNDCHECK(ZSTD_c_forceAttachDict, (int)pref);
CCtxParams->attachDictPref = pref;
return CCtxParams->attachDictPref;
}
case ZSTD_c_literalCompressionMode : {
const ZSTD_paramSwitch_e lcm = (ZSTD_paramSwitch_e)value;
BOUNDCHECK(ZSTD_c_literalCompressionMode, lcm);
BOUNDCHECK(ZSTD_c_literalCompressionMode, (int)lcm);
CCtxParams->literalCompressionMode = lcm;
return CCtxParams->literalCompressionMode;
}
@@ -934,6 +966,16 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
CCtxParams->prefetchCDictTables = (ZSTD_paramSwitch_e)value;
return CCtxParams->prefetchCDictTables;
case ZSTD_c_enableMatchFinderFallback:
BOUNDCHECK(ZSTD_c_enableMatchFinderFallback, value);
CCtxParams->enableMatchFinderFallback = value;
return CCtxParams->enableMatchFinderFallback;
case ZSTD_c_maxBlockSize:
BOUNDCHECK(ZSTD_c_maxBlockSize, value);
CCtxParams->maxBlockSize = value;
return CCtxParams->prefetchCDictTables;
default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
}
}
@@ -1069,6 +1111,12 @@ size_t ZSTD_CCtxParams_getParameter(
case ZSTD_c_prefetchCDictTables:
*value = (int)CCtxParams->prefetchCDictTables;
break;
case ZSTD_c_enableMatchFinderFallback:
*value = CCtxParams->enableMatchFinderFallback;
break;
case ZSTD_c_maxBlockSize:
*value = (int)CCtxParams->maxBlockSize;
break;
default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
}
return 0;
@@ -1095,6 +1143,21 @@ size_t ZSTD_CCtx_setParametersUsingCCtxParams(
return 0;
}
size_t ZSTD_CCtx_setCParams(ZSTD_CCtx* cctx, ZSTD_compressionParameters cparams)
{
DEBUGLOG(4, "ZSTD_CCtx_setCParams");
assert(cctx != NULL);
if (cctx->streamStage != zcss_init) {
/* All parameters in @cparams are allowed to be updated during MT compression.
* This must be signaled, so that MT compression picks up the changes */
cctx->cParamsChanged = 1;
}
/* only update if parameters are valid */
FORWARD_IF_ERROR(ZSTD_checkCParams(cparams), "");
cctx->requestedParams.cParams = cparams;
return 0;
}
size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize)
{
DEBUGLOG(4, "ZSTD_CCtx_setPledgedSrcSize to %llu bytes", pledgedSrcSize);
@@ -1240,6 +1303,7 @@ size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset)
RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
"Can't reset parameters only when not in init stage.");
ZSTD_clearAllDicts(cctx);
ZSTD_memset(&cctx->externalMatchCtx, 0, sizeof(cctx->externalMatchCtx));
return ZSTD_CCtxParams_reset(&cctx->requestedParams);
}
return 0;
@@ -1482,6 +1546,13 @@ ZSTD_sizeof_matchState(const ZSTD_compressionParameters* const cParams,
return tableSpace + optSpace + slackSpace + lazyAdditionalSpace;
}
/* Helper function for calculating memory requirements.
* Gives a tighter bound than ZSTD_sequenceBound() by taking minMatch into account. */
static size_t ZSTD_maxNbSeq(size_t blockSize, unsigned minMatch, int useExternalMatchFinder) {
U32 const divider = (minMatch==3 || useExternalMatchFinder) ? 3 : 4;
return blockSize / divider;
}
static size_t ZSTD_estimateCCtxSize_usingCCtxParams_internal(
const ZSTD_compressionParameters* cParams,
const ldmParams_t* ldmParams,
@@ -1489,12 +1560,12 @@ static size_t ZSTD_estimateCCtxSize_usingCCtxParams_internal(
const ZSTD_paramSwitch_e useRowMatchFinder,
const size_t buffInSize,
const size_t buffOutSize,
const U64 pledgedSrcSize)
const U64 pledgedSrcSize,
int useExternalMatchFinder)
{
size_t const windowSize = (size_t) BOUNDED(1ULL, 1ULL << cParams->windowLog, pledgedSrcSize);
size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, windowSize);
U32 const divider = (cParams->minMatch==3) ? 3 : 4;
size_t const maxNbSeq = blockSize / divider;
size_t const maxNbSeq = ZSTD_maxNbSeq(blockSize, cParams->minMatch, useExternalMatchFinder);
size_t const tokenSpace = ZSTD_cwksp_alloc_size(WILDCOPY_OVERLENGTH + blockSize)
+ ZSTD_cwksp_aligned_alloc_size(maxNbSeq * sizeof(seqDef))
+ 3 * ZSTD_cwksp_alloc_size(maxNbSeq * sizeof(BYTE));
@@ -1513,6 +1584,11 @@ static size_t ZSTD_estimateCCtxSize_usingCCtxParams_internal(
size_t const cctxSpace = isStatic ? ZSTD_cwksp_alloc_size(sizeof(ZSTD_CCtx)) : 0;
size_t const maxNbExternalSeq = ZSTD_sequenceBound(blockSize);
size_t const externalSeqSpace = useExternalMatchFinder
? ZSTD_cwksp_alloc_size(maxNbExternalSeq * sizeof(ZSTD_Sequence))
: 0;
size_t const neededSpace =
cctxSpace +
entropySpace +
@@ -1521,7 +1597,8 @@ static size_t ZSTD_estimateCCtxSize_usingCCtxParams_internal(
ldmSeqSpace +
matchStateSize +
tokenSpace +
bufferSpace;
bufferSpace +
externalSeqSpace;
DEBUGLOG(5, "estimate workspace : %u", (U32)neededSpace);
return neededSpace;
@@ -1539,7 +1616,7 @@ 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, &params->ldmParams, 1, useRowMatchFinder, 0, 0, ZSTD_CONTENTSIZE_UNKNOWN);
&cParams, &params->ldmParams, 1, useRowMatchFinder, 0, 0, ZSTD_CONTENTSIZE_UNKNOWN, params->useExternalMatchFinder);
}
size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams)
@@ -1600,7 +1677,7 @@ size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params)
return ZSTD_estimateCCtxSize_usingCCtxParams_internal(
&cParams, &params->ldmParams, 1, useRowMatchFinder, inBuffSize, outBuffSize,
ZSTD_CONTENTSIZE_UNKNOWN);
ZSTD_CONTENTSIZE_UNKNOWN, params->useExternalMatchFinder);
}
}
@@ -1883,8 +1960,7 @@ 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(ZSTD_BLOCKSIZE_MAX, windowSize);
U32 const divider = (params->cParams.minMatch==3) ? 3 : 4;
size_t const maxNbSeq = blockSize / divider;
size_t const maxNbSeq = ZSTD_maxNbSeq(blockSize, params->cParams.minMatch, params->useExternalMatchFinder);
size_t const buffOutSize = (zbuff == ZSTDb_buffered && params->outBufferMode == ZSTD_bm_buffered)
? ZSTD_compressBound(blockSize) + 1
: 0;
@@ -1901,7 +1977,7 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
size_t const neededSpace =
ZSTD_estimateCCtxSize_usingCCtxParams_internal(
&params->cParams, &params->ldmParams, zc->staticSize != 0, params->useRowMatchFinder,
buffInSize, buffOutSize, pledgedSrcSize);
buffInSize, buffOutSize, pledgedSrcSize, params->useExternalMatchFinder);
int resizeWorkspace;
FORWARD_IF_ERROR(neededSpace, "cctx size estimate failed!");
@@ -2014,6 +2090,14 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
zc->ldmState.loadedDictEnd = 0;
}
/* reserve space for block-level external sequences */
if (params->useExternalMatchFinder) {
size_t const maxNbExternalSeq = ZSTD_sequenceBound(blockSize);
zc->externalMatchCtx.seqBufferCapacity = maxNbExternalSeq;
zc->externalMatchCtx.seqBuffer =
(ZSTD_Sequence*)ZSTD_cwksp_reserve_aligned(ws, maxNbExternalSeq * sizeof(ZSTD_Sequence));
}
DEBUGLOG(3, "wksp: finished allocating, %zd bytes remain available", ZSTD_cwksp_available_space(ws));
assert(ZSTD_cwksp_estimated_space_within_bounds(ws, neededSpace, resizeWorkspace));
@@ -2493,11 +2577,13 @@ typedef struct {
* entropyWkspSize must be of size at least ENTROPY_WORKSPACE_SIZE - (MaxSeq + 1)*sizeof(U32)
*/
static ZSTD_symbolEncodingTypeStats_t
ZSTD_buildSequencesStatistics(seqStore_t* seqStorePtr, size_t nbSeq,
const ZSTD_fseCTables_t* prevEntropy, ZSTD_fseCTables_t* nextEntropy,
BYTE* dst, const BYTE* const dstEnd,
ZSTD_strategy strategy, unsigned* countWorkspace,
void* entropyWorkspace, size_t entropyWkspSize) {
ZSTD_buildSequencesStatistics(
const seqStore_t* seqStorePtr, size_t nbSeq,
const ZSTD_fseCTables_t* prevEntropy, ZSTD_fseCTables_t* nextEntropy,
BYTE* dst, const BYTE* const dstEnd,
ZSTD_strategy strategy, unsigned* countWorkspace,
void* entropyWorkspace, size_t entropyWkspSize)
{
BYTE* const ostart = dst;
const BYTE* const oend = dstEnd;
BYTE* op = ostart;
@@ -2616,13 +2702,14 @@ ZSTD_buildSequencesStatistics(seqStore_t* seqStorePtr, size_t nbSeq,
*/
#define SUSPECT_UNCOMPRESSIBLE_LITERAL_RATIO 20
MEM_STATIC size_t
ZSTD_entropyCompressSeqStore_internal(seqStore_t* seqStorePtr,
const ZSTD_entropyCTables_t* prevEntropy,
ZSTD_entropyCTables_t* nextEntropy,
const ZSTD_CCtx_params* cctxParams,
void* dst, size_t dstCapacity,
void* entropyWorkspace, size_t entropyWkspSize,
const int bmi2)
ZSTD_entropyCompressSeqStore_internal(
const seqStore_t* seqStorePtr,
const ZSTD_entropyCTables_t* prevEntropy,
ZSTD_entropyCTables_t* nextEntropy,
const ZSTD_CCtx_params* cctxParams,
void* dst, size_t dstCapacity,
void* entropyWorkspace, size_t entropyWkspSize,
const int bmi2)
{
const int longOffsets = cctxParams->cParams.windowLog >= STREAM_ACCUMULATOR_MIN;
ZSTD_strategy const strategy = cctxParams->cParams.strategy;
@@ -2631,7 +2718,7 @@ ZSTD_entropyCompressSeqStore_internal(seqStore_t* seqStorePtr,
FSE_CTable* CTable_OffsetBits = nextEntropy->fse.offcodeCTable;
FSE_CTable* CTable_MatchLength = nextEntropy->fse.matchlengthCTable;
const seqDef* const sequences = seqStorePtr->sequencesStart;
const size_t nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart;
const size_t nbSeq = (size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
const BYTE* const ofCodeTable = seqStorePtr->ofCode;
const BYTE* const llCodeTable = seqStorePtr->llCode;
const BYTE* const mlCodeTable = seqStorePtr->mlCode;
@@ -2649,21 +2736,20 @@ ZSTD_entropyCompressSeqStore_internal(seqStore_t* seqStorePtr,
/* Compress literals */
{ const BYTE* const literals = seqStorePtr->litStart;
size_t const numSequences = seqStorePtr->sequences - seqStorePtr->sequencesStart;
size_t const numLiterals = seqStorePtr->lit - seqStorePtr->litStart;
size_t const numSequences = (size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
size_t const numLiterals = (size_t)(seqStorePtr->lit - seqStorePtr->litStart);
/* Base suspicion of uncompressibility on ratio of literals to sequences */
unsigned const suspectUncompressible = (numSequences == 0) || (numLiterals / numSequences >= SUSPECT_UNCOMPRESSIBLE_LITERAL_RATIO);
size_t const litSize = (size_t)(seqStorePtr->lit - literals);
HUF_depth_mode depthMode = cctxParams->cParams.strategy >= HUF_OPTIMAL_DEPTH_THRESHOLD ? HUF_depth_optimal : HUF_depth_fast;
size_t const cSize = ZSTD_compressLiterals(
&prevEntropy->huf, &nextEntropy->huf,
cctxParams->cParams.strategy,
ZSTD_literalsCompressionIsDisabled(cctxParams),
op, dstCapacity,
literals, litSize,
entropyWorkspace, entropyWkspSize,
bmi2, suspectUncompressible, depthMode);
&prevEntropy->huf, &nextEntropy->huf,
cctxParams->cParams.strategy,
ZSTD_literalsCompressionIsDisabled(cctxParams),
suspectUncompressible, bmi2);
FORWARD_IF_ERROR(cSize, "ZSTD_compressLiterals failed");
assert(cSize <= dstCapacity);
op += cSize;
@@ -2735,14 +2821,15 @@ ZSTD_entropyCompressSeqStore_internal(seqStore_t* seqStorePtr,
}
MEM_STATIC size_t
ZSTD_entropyCompressSeqStore(seqStore_t* seqStorePtr,
const ZSTD_entropyCTables_t* prevEntropy,
ZSTD_entropyCTables_t* nextEntropy,
const ZSTD_CCtx_params* cctxParams,
void* dst, size_t dstCapacity,
size_t srcSize,
void* entropyWorkspace, size_t entropyWkspSize,
int bmi2)
ZSTD_entropyCompressSeqStore(
const seqStore_t* seqStorePtr,
const ZSTD_entropyCTables_t* prevEntropy,
ZSTD_entropyCTables_t* nextEntropy,
const ZSTD_CCtx_params* cctxParams,
void* dst, size_t dstCapacity,
size_t srcSize,
void* entropyWorkspace, size_t entropyWkspSize,
int bmi2)
{
size_t const cSize = ZSTD_entropyCompressSeqStore_internal(
seqStorePtr, prevEntropy, nextEntropy, cctxParams,
@@ -2763,6 +2850,10 @@ ZSTD_entropyCompressSeqStore(seqStore_t* seqStorePtr,
if (cSize >= maxCSize) return 0; /* block not compressed */
}
DEBUGLOG(5, "ZSTD_entropyCompressSeqStore() cSize: %zu", cSize);
/* libzstd decoder before > v1.5.4 is not compatible with compressed blocks of size ZSTD_BLOCKSIZE_MAX exactly.
* This restriction is indirectly already fulfilled by respecting ZSTD_minGain() condition above.
*/
assert(cSize < ZSTD_BLOCKSIZE_MAX);
return cSize;
}
@@ -2857,6 +2948,55 @@ void ZSTD_resetSeqStore(seqStore_t* ssPtr)
ssPtr->longLengthType = ZSTD_llt_none;
}
/* ZSTD_postProcessExternalMatchFinderResult() :
* Validates and post-processes sequences obtained through the external matchfinder API:
* - Checks whether nbExternalSeqs represents an error condition.
* - Appends a block delimiter to outSeqs if one is not already present.
* See zstd.h for context regarding block delimiters.
* Returns the number of sequences after post-processing, or an error code. */
static size_t ZSTD_postProcessExternalMatchFinderResult(
ZSTD_Sequence* outSeqs, size_t nbExternalSeqs, size_t outSeqsCapacity, size_t srcSize
) {
RETURN_ERROR_IF(
nbExternalSeqs > outSeqsCapacity,
externalMatchFinder_failed,
"External matchfinder returned error code %lu",
(unsigned long)nbExternalSeqs
);
RETURN_ERROR_IF(
nbExternalSeqs == 0 && srcSize > 0,
externalMatchFinder_failed,
"External matchfinder produced zero sequences for a non-empty src buffer!"
);
if (srcSize == 0) {
ZSTD_memset(&outSeqs[0], 0, sizeof(ZSTD_Sequence));
return 1;
}
{
ZSTD_Sequence const lastSeq = outSeqs[nbExternalSeqs - 1];
/* We can return early if lastSeq is already a block delimiter. */
if (lastSeq.offset == 0 && lastSeq.matchLength == 0) {
return nbExternalSeqs;
}
/* This error condition is only possible if the external matchfinder
* produced an invalid parse, by definition of ZSTD_sequenceBound(). */
RETURN_ERROR_IF(
nbExternalSeqs == outSeqsCapacity,
externalMatchFinder_failed,
"nbExternalSeqs == outSeqsCapacity but lastSeq is not a block delimiter!"
);
/* lastSeq is not a block delimiter, so we need to append one. */
ZSTD_memset(&outSeqs[nbExternalSeqs], 0, sizeof(ZSTD_Sequence));
return nbExternalSeqs + 1;
}
}
typedef enum { ZSTDbss_compress, ZSTDbss_noCompress } ZSTD_buildSeqStore_e;
static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
@@ -2904,6 +3044,15 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
}
if (zc->externSeqStore.pos < zc->externSeqStore.size) {
assert(zc->appliedParams.ldmParams.enableLdm == ZSTD_ps_disable);
/* External matchfinder + LDM is technically possible, just not implemented yet.
* We need to revisit soon and implement it. */
RETURN_ERROR_IF(
zc->appliedParams.useExternalMatchFinder,
parameter_combination_unsupported,
"Long-distance matching with external matchfinder enabled is not currently supported."
);
/* Updates ldmSeqStore.pos */
lastLLSize =
ZSTD_ldm_blockCompress(&zc->externSeqStore,
@@ -2915,6 +3064,14 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
} else if (zc->appliedParams.ldmParams.enableLdm == ZSTD_ps_enable) {
rawSeqStore_t ldmSeqStore = kNullRawSeqStore;
/* External matchfinder + LDM is technically possible, just not implemented yet.
* We need to revisit soon and implement it. */
RETURN_ERROR_IF(
zc->appliedParams.useExternalMatchFinder,
parameter_combination_unsupported,
"Long-distance matching with external matchfinder enabled is not currently supported."
);
ldmSeqStore.seq = zc->ldmSequences;
ldmSeqStore.capacity = zc->maxNbLdmSequences;
/* Updates ldmSeqStore.size */
@@ -2929,10 +3086,64 @@ 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 { /* not long range mode */
} else if (zc->appliedParams.useExternalMatchFinder) {
assert(
zc->externalMatchCtx.seqBufferCapacity >= ZSTD_sequenceBound(srcSize)
);
assert(zc->externalMatchCtx.mFinder != 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,
src, srcSize,
NULL, 0, /* dict and dictSize, currently not supported */
zc->appliedParams.compressionLevel,
windowSize
);
size_t const nbPostProcessedSeqs = ZSTD_postProcessExternalMatchFinderResult(
zc->externalMatchCtx.seqBuffer,
nbExternalSeqs,
zc->externalMatchCtx.seqBufferCapacity,
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};
ZSTD_copySequencesToSeqStoreExplicitBlockDelim(
zc, &seqPos, zc->externalMatchCtx.seqBuffer, nbPostProcessedSeqs, src, srcSize
);
ms->ldmSeqStore = NULL;
DEBUGLOG(5, "Copied %lu sequences from external matchfinder to internal seqStore.", (unsigned long)nbExternalSeqs);
return ZSTDbss_compress;
}
/* Propagate the error if fallback is disabled */
if (!zc->appliedParams.enableMatchFinderFallback) {
return nbPostProcessedSeqs;
}
/* Fallback to software matchfinder */
{ ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy,
zc->appliedParams.useRowMatchFinder,
dictMode);
ms->ldmSeqStore = NULL;
DEBUGLOG(
5,
"External matchfinder returned error code %lu. Falling back to internal matchfinder.",
(unsigned long)nbExternalSeqs
);
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);
assert(zc->externalMatchCtx.mFinder == NULL);
ms->ldmSeqStore = NULL;
lastLLSize = blockCompressor(ms, &zc->seqStore, zc->blockState.nextCBlock->rep, src, srcSize);
}
@@ -3055,19 +3266,17 @@ static int ZSTD_isRLE(const BYTE* src, size_t length) {
const size_t unrollMask = unrollSize - 1;
const size_t prefixLength = length & unrollMask;
size_t i;
size_t u;
if (length == 1) return 1;
/* Check if prefix is RLE first before using unrolled loop */
if (prefixLength && ZSTD_count(ip+1, ip, ip+prefixLength) != prefixLength-1) {
return 0;
}
for (i = prefixLength; i != length; i += unrollSize) {
size_t u;
for (u = 0; u < unrollSize; u += sizeof(size_t)) {
if (MEM_readST(ip + i + u) != valueST) {
return 0;
}
}
}
} } }
return 1;
}
@@ -3083,7 +3292,8 @@ static int ZSTD_maybeRLE(seqStore_t const* seqStore)
return nbSeqs < 4 && nbLits < 10;
}
static void ZSTD_blockState_confirmRepcodesAndEntropyTables(ZSTD_blockState_t* const bs)
static void
ZSTD_blockState_confirmRepcodesAndEntropyTables(ZSTD_blockState_t* const bs)
{
ZSTD_compressedBlockState_t* const tmp = bs->prevCBlock;
bs->prevCBlock = bs->nextCBlock;
@@ -3091,7 +3301,9 @@ static void ZSTD_blockState_confirmRepcodesAndEntropyTables(ZSTD_blockState_t* c
}
/* Writes the block header */
static void writeBlockHeader(void* op, size_t cSize, size_t blockSize, U32 lastBlock) {
static void
writeBlockHeader(void* op, size_t cSize, size_t blockSize, U32 lastBlock)
{
U32 const cBlockHeader = cSize == 1 ?
lastBlock + (((U32)bt_rle)<<1) + (U32)(blockSize << 3) :
lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3);
@@ -3104,13 +3316,16 @@ static void writeBlockHeader(void* op, size_t cSize, size_t blockSize, U32 lastB
* Stores literals block type (raw, rle, compressed, repeat) and
* huffman description table to hufMetadata.
* Requires ENTROPY_WORKSPACE_SIZE workspace
* @return : size of huffman description table or error code */
static size_t ZSTD_buildBlockEntropyStats_literals(void* const src, size_t srcSize,
const ZSTD_hufCTables_t* prevHuf,
ZSTD_hufCTables_t* nextHuf,
ZSTD_hufCTablesMetadata_t* hufMetadata,
const int literalsCompressionIsDisabled,
void* workspace, size_t wkspSize, HUF_depth_mode depthMode)
* @return : size of huffman description table, or an error code
*/
static size_t
ZSTD_buildBlockEntropyStats_literals(void* const src, size_t srcSize,
const ZSTD_hufCTables_t* prevHuf,
ZSTD_hufCTables_t* nextHuf,
ZSTD_hufCTablesMetadata_t* hufMetadata,
const int literalsCompressionIsDisabled,
void* workspace, size_t wkspSize,
HUF_depth_mode depthMode)
{
BYTE* const wkspStart = (BYTE*)workspace;
BYTE* const wkspEnd = wkspStart + wkspSize;
@@ -3135,33 +3350,37 @@ static size_t ZSTD_buildBlockEntropyStats_literals(void* const src, size_t srcSi
/* small ? don't even attempt compression (speed opt) */
#ifndef COMPRESS_LITERALS_SIZE_MIN
#define COMPRESS_LITERALS_SIZE_MIN 63
# define COMPRESS_LITERALS_SIZE_MIN 63 /* heuristic */
#endif
{ size_t const minLitSize = (prevHuf->repeatMode == HUF_repeat_valid) ? 6 : COMPRESS_LITERALS_SIZE_MIN;
if (srcSize <= minLitSize) {
DEBUGLOG(5, "set_basic - too small");
hufMetadata->hType = set_basic;
return 0;
}
}
} }
/* Scan input and build symbol stats */
{ size_t const largest = HIST_count_wksp (countWksp, &maxSymbolValue, (const BYTE*)src, srcSize, workspace, wkspSize);
{ size_t const largest =
HIST_count_wksp (countWksp, &maxSymbolValue,
(const BYTE*)src, srcSize,
workspace, wkspSize);
FORWARD_IF_ERROR(largest, "HIST_count_wksp failed");
if (largest == srcSize) {
/* only one literal symbol */
DEBUGLOG(5, "set_rle");
hufMetadata->hType = set_rle;
return 0;
}
if (largest <= (srcSize >> 7)+4) {
/* heuristic: likely not compressible */
DEBUGLOG(5, "set_basic - no gain");
hufMetadata->hType = set_basic;
return 0;
}
}
} }
/* Validate the previous Huffman table */
if (repeat == HUF_repeat_check && !HUF_validateCTable((HUF_CElt const*)prevHuf->CTable, countWksp, maxSymbolValue)) {
if (repeat == HUF_repeat_check
&& !HUF_validateCTable((HUF_CElt const*)prevHuf->CTable, countWksp, maxSymbolValue)) {
repeat = HUF_repeat_none;
}
@@ -3174,35 +3393,34 @@ static size_t ZSTD_buildBlockEntropyStats_literals(void* const src, size_t srcSi
nodeWksp, nodeWkspSize);
FORWARD_IF_ERROR(maxBits, "HUF_buildCTable_wksp");
huffLog = (U32)maxBits;
{ /* Build and write the CTable */
size_t const newCSize = HUF_estimateCompressedSize(
(HUF_CElt*)nextHuf->CTable, countWksp, maxSymbolValue);
size_t const hSize = HUF_writeCTable_wksp(
hufMetadata->hufDesBuffer, sizeof(hufMetadata->hufDesBuffer),
(HUF_CElt*)nextHuf->CTable, maxSymbolValue, huffLog,
nodeWksp, nodeWkspSize);
/* Check against repeating the previous CTable */
if (repeat != HUF_repeat_none) {
size_t const oldCSize = HUF_estimateCompressedSize(
(HUF_CElt const*)prevHuf->CTable, countWksp, maxSymbolValue);
if (oldCSize < srcSize && (oldCSize <= hSize + newCSize || hSize + 12 >= srcSize)) {
DEBUGLOG(5, "set_repeat - smaller");
ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
hufMetadata->hType = set_repeat;
return 0;
}
}
if (newCSize + hSize >= srcSize) {
DEBUGLOG(5, "set_basic - no gains");
}
{ /* Build and write the CTable */
size_t const newCSize = HUF_estimateCompressedSize(
(HUF_CElt*)nextHuf->CTable, countWksp, maxSymbolValue);
size_t const hSize = HUF_writeCTable_wksp(
hufMetadata->hufDesBuffer, sizeof(hufMetadata->hufDesBuffer),
(HUF_CElt*)nextHuf->CTable, maxSymbolValue, huffLog,
nodeWksp, nodeWkspSize);
/* Check against repeating the previous CTable */
if (repeat != HUF_repeat_none) {
size_t const oldCSize = HUF_estimateCompressedSize(
(HUF_CElt const*)prevHuf->CTable, countWksp, maxSymbolValue);
if (oldCSize < srcSize && (oldCSize <= hSize + newCSize || hSize + 12 >= srcSize)) {
DEBUGLOG(5, "set_repeat - smaller");
ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
hufMetadata->hType = set_basic;
hufMetadata->hType = set_repeat;
return 0;
}
DEBUGLOG(5, "set_compressed (hSize=%u)", (U32)hSize);
hufMetadata->hType = set_compressed;
nextHuf->repeatMode = HUF_repeat_check;
return hSize;
} }
if (newCSize + hSize >= srcSize) {
DEBUGLOG(5, "set_basic - no gains");
ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
hufMetadata->hType = set_basic;
return 0;
}
DEBUGLOG(5, "set_compressed (hSize=%u)", (U32)hSize);
hufMetadata->hType = set_compressed;
nextHuf->repeatMode = HUF_repeat_check;
return hSize;
}
}
@@ -3212,7 +3430,8 @@ static size_t ZSTD_buildBlockEntropyStats_literals(void* const src, size_t srcSi
* and updates nextEntropy to the appropriate repeatMode.
*/
static ZSTD_symbolEncodingTypeStats_t
ZSTD_buildDummySequencesStatistics(ZSTD_fseCTables_t* nextEntropy) {
ZSTD_buildDummySequencesStatistics(ZSTD_fseCTables_t* nextEntropy)
{
ZSTD_symbolEncodingTypeStats_t stats = {set_basic, set_basic, set_basic, 0, 0};
nextEntropy->litlength_repeatMode = FSE_repeat_none;
nextEntropy->offcode_repeatMode = FSE_repeat_none;
@@ -3224,16 +3443,18 @@ ZSTD_buildDummySequencesStatistics(ZSTD_fseCTables_t* nextEntropy) {
* Builds entropy for the sequences.
* Stores symbol compression modes and fse table to fseMetadata.
* Requires ENTROPY_WORKSPACE_SIZE wksp.
* @return : size of fse tables or error code */
static size_t ZSTD_buildBlockEntropyStats_sequences(seqStore_t* seqStorePtr,
const ZSTD_fseCTables_t* prevEntropy,
ZSTD_fseCTables_t* nextEntropy,
const ZSTD_CCtx_params* cctxParams,
ZSTD_fseCTablesMetadata_t* fseMetadata,
void* workspace, size_t wkspSize)
* @return : size of fse tables or error code */
static size_t
ZSTD_buildBlockEntropyStats_sequences(
const seqStore_t* seqStorePtr,
const ZSTD_fseCTables_t* prevEntropy,
ZSTD_fseCTables_t* nextEntropy,
const ZSTD_CCtx_params* cctxParams,
ZSTD_fseCTablesMetadata_t* fseMetadata,
void* workspace, size_t wkspSize)
{
ZSTD_strategy const strategy = cctxParams->cParams.strategy;
size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart;
size_t const nbSeq = (size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
BYTE* const ostart = fseMetadata->fseTablesBuffer;
BYTE* const oend = ostart + sizeof(fseMetadata->fseTablesBuffer);
BYTE* op = ostart;
@@ -3260,18 +3481,20 @@ static size_t ZSTD_buildBlockEntropyStats_sequences(seqStore_t* seqStorePtr,
/** ZSTD_buildBlockEntropyStats() :
* Builds entropy for the block.
* Requires workspace size ENTROPY_WORKSPACE_SIZE
*
* @return : 0 on success or error code
* @return : 0 on success, or an error code
* Note : also employed in superblock
*/
size_t ZSTD_buildBlockEntropyStats(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)
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)
{
size_t const litSize = seqStorePtr->lit - seqStorePtr->litStart;
HUF_depth_mode depthMode = cctxParams->cParams.strategy >= HUF_OPTIMAL_DEPTH_THRESHOLD ? HUF_depth_optimal : HUF_depth_fast;
size_t const litSize = (size_t)(seqStorePtr->lit - seqStorePtr->litStart);
int const huf_useOptDepth = (cctxParams->cParams.strategy >= HUF_OPTIMAL_DEPTH_THRESHOLD);
HUF_depth_mode const depthMode = huf_useOptDepth ? HUF_depth_optimal : HUF_depth_fast;
entropyMetadata->hufMetadata.hufDesSize =
ZSTD_buildBlockEntropyStats_literals(seqStorePtr->litStart, litSize,
@@ -3292,11 +3515,12 @@ size_t ZSTD_buildBlockEntropyStats(seqStore_t* seqStorePtr,
}
/* Returns the size estimate for the literals section (header + content) of a block */
static size_t ZSTD_estimateBlockSize_literal(const BYTE* literals, size_t litSize,
const ZSTD_hufCTables_t* huf,
const ZSTD_hufCTablesMetadata_t* hufMetadata,
void* workspace, size_t wkspSize,
int writeEntropy)
static size_t
ZSTD_estimateBlockSize_literal(const BYTE* literals, size_t litSize,
const ZSTD_hufCTables_t* huf,
const ZSTD_hufCTablesMetadata_t* hufMetadata,
void* workspace, size_t wkspSize,
int writeEntropy)
{
unsigned* const countWksp = (unsigned*)workspace;
unsigned maxSymbolValue = HUF_SYMBOLVALUE_MAX;
@@ -3318,12 +3542,13 @@ static size_t ZSTD_estimateBlockSize_literal(const BYTE* literals, size_t litSiz
}
/* Returns the size estimate for the FSE-compressed symbols (of, ml, ll) of a block */
static size_t ZSTD_estimateBlockSize_symbolType(symbolEncodingType_e type,
const BYTE* codeTable, size_t nbSeq, unsigned maxCode,
const FSE_CTable* fseCTable,
const U8* additionalBits,
short const* defaultNorm, U32 defaultNormLog, U32 defaultMax,
void* workspace, size_t wkspSize)
static size_t
ZSTD_estimateBlockSize_symbolType(symbolEncodingType_e type,
const BYTE* codeTable, size_t nbSeq, unsigned maxCode,
const FSE_CTable* fseCTable,
const U8* additionalBits,
short const* defaultNorm, U32 defaultNormLog, U32 defaultMax,
void* workspace, size_t wkspSize)
{
unsigned* const countWksp = (unsigned*)workspace;
const BYTE* ctp = codeTable;
@@ -3355,99 +3580,107 @@ static size_t ZSTD_estimateBlockSize_symbolType(symbolEncodingType_e type,
}
/* Returns the size estimate for the sequences section (header + content) of a block */
static size_t ZSTD_estimateBlockSize_sequences(const BYTE* ofCodeTable,
const BYTE* llCodeTable,
const BYTE* mlCodeTable,
size_t nbSeq,
const ZSTD_fseCTables_t* fseTables,
const ZSTD_fseCTablesMetadata_t* fseMetadata,
void* workspace, size_t wkspSize,
int writeEntropy)
static size_t
ZSTD_estimateBlockSize_sequences(const BYTE* ofCodeTable,
const BYTE* llCodeTable,
const BYTE* mlCodeTable,
size_t nbSeq,
const ZSTD_fseCTables_t* fseTables,
const ZSTD_fseCTablesMetadata_t* fseMetadata,
void* workspace, size_t wkspSize,
int writeEntropy)
{
size_t sequencesSectionHeaderSize = 1 /* seqHead */ + 1 /* min seqSize size */ + (nbSeq >= 128) + (nbSeq >= LONGNBSEQ);
size_t cSeqSizeEstimate = 0;
cSeqSizeEstimate += ZSTD_estimateBlockSize_symbolType(fseMetadata->ofType, ofCodeTable, nbSeq, MaxOff,
fseTables->offcodeCTable, NULL,
OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,
workspace, wkspSize);
fseTables->offcodeCTable, NULL,
OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,
workspace, wkspSize);
cSeqSizeEstimate += ZSTD_estimateBlockSize_symbolType(fseMetadata->llType, llCodeTable, nbSeq, MaxLL,
fseTables->litlengthCTable, LL_bits,
LL_defaultNorm, LL_defaultNormLog, MaxLL,
workspace, wkspSize);
fseTables->litlengthCTable, LL_bits,
LL_defaultNorm, LL_defaultNormLog, MaxLL,
workspace, wkspSize);
cSeqSizeEstimate += ZSTD_estimateBlockSize_symbolType(fseMetadata->mlType, mlCodeTable, nbSeq, MaxML,
fseTables->matchlengthCTable, ML_bits,
ML_defaultNorm, ML_defaultNormLog, MaxML,
workspace, wkspSize);
fseTables->matchlengthCTable, ML_bits,
ML_defaultNorm, ML_defaultNormLog, MaxML,
workspace, wkspSize);
if (writeEntropy) cSeqSizeEstimate += fseMetadata->fseTablesSize;
return cSeqSizeEstimate + sequencesSectionHeaderSize;
}
/* Returns the size estimate for a given stream of literals, of, ll, ml */
static size_t ZSTD_estimateBlockSize(const BYTE* literals, size_t litSize,
const BYTE* ofCodeTable,
const BYTE* llCodeTable,
const BYTE* mlCodeTable,
size_t nbSeq,
const ZSTD_entropyCTables_t* entropy,
const ZSTD_entropyCTablesMetadata_t* entropyMetadata,
void* workspace, size_t wkspSize,
int writeLitEntropy, int writeSeqEntropy) {
static size_t
ZSTD_estimateBlockSize(const BYTE* literals, size_t litSize,
const BYTE* ofCodeTable,
const BYTE* llCodeTable,
const BYTE* mlCodeTable,
size_t nbSeq,
const ZSTD_entropyCTables_t* entropy,
const ZSTD_entropyCTablesMetadata_t* entropyMetadata,
void* workspace, size_t wkspSize,
int writeLitEntropy, int writeSeqEntropy)
{
size_t const literalsSize = ZSTD_estimateBlockSize_literal(literals, litSize,
&entropy->huf, &entropyMetadata->hufMetadata,
workspace, wkspSize, writeLitEntropy);
&entropy->huf, &entropyMetadata->hufMetadata,
workspace, wkspSize, writeLitEntropy);
size_t const seqSize = ZSTD_estimateBlockSize_sequences(ofCodeTable, llCodeTable, mlCodeTable,
nbSeq, &entropy->fse, &entropyMetadata->fseMetadata,
workspace, wkspSize, writeSeqEntropy);
nbSeq, &entropy->fse, &entropyMetadata->fseMetadata,
workspace, wkspSize, writeSeqEntropy);
return seqSize + literalsSize + ZSTD_blockHeaderSize;
}
/* Builds entropy statistics and uses them for blocksize estimation.
*
* Returns the estimated compressed size of the seqStore, or a zstd error.
* @return: estimated compressed size of the seqStore, or a zstd error.
*/
static size_t ZSTD_buildEntropyStatisticsAndEstimateSubBlockSize(seqStore_t* seqStore, ZSTD_CCtx* zc) {
ZSTD_entropyCTablesMetadata_t* entropyMetadata = &zc->blockSplitCtx.entropyMetadata;
static size_t
ZSTD_buildEntropyStatisticsAndEstimateSubBlockSize(seqStore_t* seqStore, ZSTD_CCtx* zc)
{
ZSTD_entropyCTablesMetadata_t* const entropyMetadata = &zc->blockSplitCtx.entropyMetadata;
DEBUGLOG(6, "ZSTD_buildEntropyStatisticsAndEstimateSubBlockSize()");
FORWARD_IF_ERROR(ZSTD_buildBlockEntropyStats(seqStore,
&zc->blockState.prevCBlock->entropy,
&zc->blockState.nextCBlock->entropy,
&zc->appliedParams,
entropyMetadata,
zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE /* statically allocated in resetCCtx */), "");
return ZSTD_estimateBlockSize(seqStore->litStart, (size_t)(seqStore->lit - seqStore->litStart),
zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE), "");
return ZSTD_estimateBlockSize(
seqStore->litStart, (size_t)(seqStore->lit - seqStore->litStart),
seqStore->ofCode, seqStore->llCode, seqStore->mlCode,
(size_t)(seqStore->sequences - seqStore->sequencesStart),
&zc->blockState.nextCBlock->entropy, entropyMetadata, zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE,
&zc->blockState.nextCBlock->entropy,
entropyMetadata,
zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE,
(int)(entropyMetadata->hufMetadata.hType == set_compressed), 1);
}
/* Returns literals bytes represented in a seqStore */
static size_t ZSTD_countSeqStoreLiteralsBytes(const seqStore_t* const seqStore) {
static size_t ZSTD_countSeqStoreLiteralsBytes(const seqStore_t* const seqStore)
{
size_t literalsBytes = 0;
size_t const nbSeqs = seqStore->sequences - seqStore->sequencesStart;
size_t const nbSeqs = (size_t)(seqStore->sequences - seqStore->sequencesStart);
size_t i;
for (i = 0; i < nbSeqs; ++i) {
seqDef seq = seqStore->sequencesStart[i];
seqDef const seq = seqStore->sequencesStart[i];
literalsBytes += seq.litLength;
if (i == seqStore->longLengthPos && seqStore->longLengthType == ZSTD_llt_literalLength) {
literalsBytes += 0x10000;
}
}
} }
return literalsBytes;
}
/* Returns match bytes represented in a seqStore */
static size_t ZSTD_countSeqStoreMatchBytes(const seqStore_t* const seqStore) {
static size_t ZSTD_countSeqStoreMatchBytes(const seqStore_t* const seqStore)
{
size_t matchBytes = 0;
size_t const nbSeqs = seqStore->sequences - seqStore->sequencesStart;
size_t const nbSeqs = (size_t)(seqStore->sequences - seqStore->sequencesStart);
size_t i;
for (i = 0; i < nbSeqs; ++i) {
seqDef seq = seqStore->sequencesStart[i];
matchBytes += seq.mlBase + MINMATCH;
if (i == seqStore->longLengthPos && seqStore->longLengthType == ZSTD_llt_matchLength) {
matchBytes += 0x10000;
}
}
} }
return matchBytes;
}
@@ -3496,6 +3729,7 @@ ZSTD_resolveRepcodeToRawOffset(const U32 rep[ZSTD_REP_NUM], const U32 offBase, c
U32 const adjustedRepCode = OFFBASE_TO_REPCODE(offBase) - 1 + ll0; /* [ 0 - 3 ] */
assert(OFFBASE_IS_REPCODE(offBase));
if (adjustedRepCode == ZSTD_REP_NUM) {
assert(ll0);
/* litlength == 0 and offCode == 2 implies selection of first repcode - 1
* This is only valid if it results in a valid offset value, aka > 0.
* Note : it may happen that `rep[0]==1` in exceptional circumstances.
@@ -3521,14 +3755,16 @@ ZSTD_resolveRepcodeToRawOffset(const U32 rep[ZSTD_REP_NUM], const U32 offBase, c
* 1-3 : repcode 1-3
* 4+ : real_offset+3
*/
static void ZSTD_seqStore_resolveOffCodes(repcodes_t* const dRepcodes, repcodes_t* const cRepcodes,
seqStore_t* const seqStore, U32 const nbSeq) {
static void
ZSTD_seqStore_resolveOffCodes(repcodes_t* const dRepcodes, repcodes_t* const cRepcodes,
const seqStore_t* const seqStore, U32 const nbSeq)
{
U32 idx = 0;
for (; idx < nbSeq; ++idx) {
seqDef* const seq = seqStore->sequencesStart + idx;
U32 const ll0 = (seq->litLength == 0);
U32 const offBase = seq->offBase;
assert(seq->offBase > 0);
assert(offBase > 0);
if (OFFBASE_IS_REPCODE(offBase)) {
U32 const dRawOffset = ZSTD_resolveRepcodeToRawOffset(dRepcodes->rep, offBase, ll0);
U32 const cRawOffset = ZSTD_resolveRepcodeToRawOffset(cRepcodes->rep, offBase, ll0);
@@ -3537,7 +3773,7 @@ static void ZSTD_seqStore_resolveOffCodes(repcodes_t* const dRepcodes, repcodes_
* repcode history.
*/
if (dRawOffset != cRawOffset) {
seq->offBase = cRawOffset + ZSTD_REP_NUM;
seq->offBase = OFFSET_TO_OFFBASE(cRawOffset);
}
}
/* Compression repcode history is always updated with values directly from the unmodified seqStore.
@@ -3554,10 +3790,11 @@ static void ZSTD_seqStore_resolveOffCodes(repcodes_t* const dRepcodes, repcodes_
* Returns the total size of that block (including header) or a ZSTD error code.
*/
static size_t
ZSTD_compressSeqStore_singleBlock(ZSTD_CCtx* zc, seqStore_t* const seqStore,
ZSTD_compressSeqStore_singleBlock(ZSTD_CCtx* zc,
const seqStore_t* const seqStore,
repcodes_t* const dRep, repcodes_t* const cRep,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
const void* src, size_t srcSize,
U32 lastBlock, U32 isPartition)
{
const U32 rleMaxLength = 25;
@@ -3631,10 +3868,11 @@ typedef struct {
/* Helper function to perform the recursive search for block splits.
* Estimates the cost of seqStore prior to split, and estimates the cost of splitting the sequences in half.
* If advantageous to split, then we recurse down the two sub-blocks. If not, or if an error occurred in estimation, then
* we do not recurse.
* If advantageous to split, then we recurse down the two sub-blocks.
* If not, or if an error occurred in estimation, then we do not recurse.
*
* Note: The recursion depth is capped by a heuristic minimum number of sequences, defined by MIN_SEQUENCES_BLOCK_SPLITTING.
* Note: The recursion depth is capped by a heuristic minimum number of sequences,
* defined by MIN_SEQUENCES_BLOCK_SPLITTING.
* In theory, this means the absolute largest recursion depth is 10 == log2(maxNbSeqInBlock/MIN_SEQUENCES_BLOCK_SPLITTING).
* In practice, recursion depth usually doesn't go beyond 4.
*
@@ -3646,19 +3884,20 @@ static void
ZSTD_deriveBlockSplitsHelper(seqStoreSplits* splits, size_t startIdx, size_t endIdx,
ZSTD_CCtx* zc, const seqStore_t* origSeqStore)
{
seqStore_t* fullSeqStoreChunk = &zc->blockSplitCtx.fullSeqStoreChunk;
seqStore_t* firstHalfSeqStore = &zc->blockSplitCtx.firstHalfSeqStore;
seqStore_t* secondHalfSeqStore = &zc->blockSplitCtx.secondHalfSeqStore;
seqStore_t* const fullSeqStoreChunk = &zc->blockSplitCtx.fullSeqStoreChunk;
seqStore_t* const firstHalfSeqStore = &zc->blockSplitCtx.firstHalfSeqStore;
seqStore_t* const secondHalfSeqStore = &zc->blockSplitCtx.secondHalfSeqStore;
size_t estimatedOriginalSize;
size_t estimatedFirstHalfSize;
size_t estimatedSecondHalfSize;
size_t midIdx = (startIdx + endIdx)/2;
DEBUGLOG(5, "ZSTD_deriveBlockSplitsHelper: startIdx=%zu endIdx=%zu", startIdx, endIdx);
assert(endIdx >= startIdx);
if (endIdx - startIdx < MIN_SEQUENCES_BLOCK_SPLITTING || splits->idx >= ZSTD_MAX_NB_BLOCK_SPLITS) {
DEBUGLOG(6, "ZSTD_deriveBlockSplitsHelper: Too few sequences");
DEBUGLOG(6, "ZSTD_deriveBlockSplitsHelper: Too few sequences (%zu)", endIdx - startIdx);
return;
}
DEBUGLOG(5, "ZSTD_deriveBlockSplitsHelper: startIdx=%zu endIdx=%zu", startIdx, endIdx);
ZSTD_deriveSeqStoreChunk(fullSeqStoreChunk, origSeqStore, startIdx, endIdx);
ZSTD_deriveSeqStoreChunk(firstHalfSeqStore, origSeqStore, startIdx, midIdx);
ZSTD_deriveSeqStoreChunk(secondHalfSeqStore, origSeqStore, midIdx, endIdx);
@@ -3679,15 +3918,16 @@ ZSTD_deriveBlockSplitsHelper(seqStoreSplits* splits, size_t startIdx, size_t end
}
}
/* Base recursive function. Populates a table with intra-block partition indices that can improve compression ratio.
/* Base recursive function.
* Populates a table with intra-block partition indices that can improve compression ratio.
*
* Returns the number of splits made (which equals the size of the partition table - 1).
* @return: number of splits made (which equals the size of the partition table - 1).
*/
static size_t ZSTD_deriveBlockSplits(ZSTD_CCtx* zc, U32 partitions[], U32 nbSeq)
{
seqStoreSplits splits = {partitions, 0};
if (nbSeq <= 4) {
DEBUGLOG(5, "ZSTD_deriveBlockSplits: Too few sequences to split");
DEBUGLOG(5, "ZSTD_deriveBlockSplits: Too few sequences to split (%u <= 4)", nbSeq);
/* Refuse to try and split anything with less than 4 sequences */
return 0;
}
@@ -3703,18 +3943,20 @@ static size_t ZSTD_deriveBlockSplits(ZSTD_CCtx* zc, U32 partitions[], U32 nbSeq)
* Returns combined size of all blocks (which includes headers), or a ZSTD error code.
*/
static size_t
ZSTD_compressBlock_splitBlock_internal(ZSTD_CCtx* zc, void* dst, size_t dstCapacity,
const void* src, size_t blockSize, U32 lastBlock, U32 nbSeq)
ZSTD_compressBlock_splitBlock_internal(ZSTD_CCtx* zc,
void* dst, size_t dstCapacity,
const void* src, size_t blockSize,
U32 lastBlock, U32 nbSeq)
{
size_t cSize = 0;
const BYTE* ip = (const BYTE*)src;
BYTE* op = (BYTE*)dst;
size_t i = 0;
size_t srcBytesTotal = 0;
U32* partitions = zc->blockSplitCtx.partitions; /* size == ZSTD_MAX_NB_BLOCK_SPLITS */
seqStore_t* nextSeqStore = &zc->blockSplitCtx.nextSeqStore;
seqStore_t* currSeqStore = &zc->blockSplitCtx.currSeqStore;
size_t numSplits = ZSTD_deriveBlockSplits(zc, partitions, nbSeq);
U32* const partitions = zc->blockSplitCtx.partitions; /* size == ZSTD_MAX_NB_BLOCK_SPLITS */
seqStore_t* const nextSeqStore = &zc->blockSplitCtx.nextSeqStore;
seqStore_t* const currSeqStore = &zc->blockSplitCtx.currSeqStore;
size_t const numSplits = ZSTD_deriveBlockSplits(zc, partitions, nbSeq);
/* If a block is split and some partitions are emitted as RLE/uncompressed, then repcode history
* may become invalid. In order to reconcile potentially invalid repcodes, we keep track of two
@@ -3736,19 +3978,21 @@ ZSTD_compressBlock_splitBlock_internal(ZSTD_CCtx* zc, void* dst, size_t dstCapac
ZSTD_memcpy(cRep.rep, zc->blockState.prevCBlock->rep, sizeof(repcodes_t));
ZSTD_memset(nextSeqStore, 0, sizeof(seqStore_t));
DEBUGLOG(4, "ZSTD_compressBlock_splitBlock_internal (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u)",
DEBUGLOG(5, "ZSTD_compressBlock_splitBlock_internal (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u)",
(unsigned)dstCapacity, (unsigned)zc->blockState.matchState.window.dictLimit,
(unsigned)zc->blockState.matchState.nextToUpdate);
if (numSplits == 0) {
size_t cSizeSingleBlock = ZSTD_compressSeqStore_singleBlock(zc, &zc->seqStore,
&dRep, &cRep,
op, dstCapacity,
ip, blockSize,
lastBlock, 0 /* isPartition */);
size_t cSizeSingleBlock =
ZSTD_compressSeqStore_singleBlock(zc, &zc->seqStore,
&dRep, &cRep,
op, dstCapacity,
ip, blockSize,
lastBlock, 0 /* isPartition */);
FORWARD_IF_ERROR(cSizeSingleBlock, "Compressing single block from splitBlock_internal() failed!");
DEBUGLOG(5, "ZSTD_compressBlock_splitBlock_internal: No splits");
assert(cSizeSingleBlock <= ZSTD_BLOCKSIZE_MAX + ZSTD_blockHeaderSize);
assert(zc->blockSize <= ZSTD_BLOCKSIZE_MAX);
assert(cSizeSingleBlock <= zc->blockSize + ZSTD_blockHeaderSize);
return cSizeSingleBlock;
}
@@ -3773,7 +4017,8 @@ ZSTD_compressBlock_splitBlock_internal(ZSTD_CCtx* zc, void* dst, size_t dstCapac
op, dstCapacity,
ip, srcBytes,
lastBlockEntireSrc, 1 /* isPartition */);
DEBUGLOG(5, "Estimated size: %zu actual size: %zu", ZSTD_buildEntropyStatisticsAndEstimateSubBlockSize(currSeqStore, zc), cSizeChunk);
DEBUGLOG(5, "Estimated size: %zu vs %zu : actual size",
ZSTD_buildEntropyStatisticsAndEstimateSubBlockSize(currSeqStore, zc), cSizeChunk);
FORWARD_IF_ERROR(cSizeChunk, "Compressing chunk failed!");
ip += srcBytes;
@@ -3781,10 +4026,10 @@ ZSTD_compressBlock_splitBlock_internal(ZSTD_CCtx* zc, void* dst, size_t dstCapac
dstCapacity -= cSizeChunk;
cSize += cSizeChunk;
*currSeqStore = *nextSeqStore;
assert(cSizeChunk <= ZSTD_BLOCKSIZE_MAX + ZSTD_blockHeaderSize);
assert(cSizeChunk <= zc->blockSize + ZSTD_blockHeaderSize);
}
/* cRep and dRep may have diverged during the compression. If so, we use the dRep repcodes
* for the next block.
/* cRep and dRep may have diverged during the compression.
* If so, we use the dRep repcodes for the next block.
*/
ZSTD_memcpy(zc->blockState.prevCBlock->rep, dRep.rep, sizeof(repcodes_t));
return cSize;
@@ -3795,8 +4040,6 @@ ZSTD_compressBlock_splitBlock(ZSTD_CCtx* zc,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize, U32 lastBlock)
{
const BYTE* ip = (const BYTE*)src;
BYTE* op = (BYTE*)dst;
U32 nbSeq;
size_t cSize;
DEBUGLOG(4, "ZSTD_compressBlock_splitBlock");
@@ -3807,7 +4050,7 @@ 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;
cSize = ZSTD_noCompressBlock(op, dstCapacity, ip, srcSize, lastBlock);
cSize = ZSTD_noCompressBlock(dst, dstCapacity, src, srcSize, lastBlock);
FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed");
DEBUGLOG(4, "ZSTD_compressBlock_splitBlock: Nocompress block");
return cSize;
@@ -3919,10 +4162,11 @@ static size_t ZSTD_compressBlock_targetCBlockSize_body(ZSTD_CCtx* zc,
* * cSize >= blockBound(srcSize): We have expanded the block too much so
* emit an uncompressed block.
*/
{
size_t const cSize = ZSTD_compressSuperBlock(zc, dst, dstCapacity, src, srcSize, lastBlock);
{ size_t const cSize =
ZSTD_compressSuperBlock(zc, dst, dstCapacity, src, srcSize, lastBlock);
if (cSize != ERROR(dstSize_tooSmall)) {
size_t const maxCSize = srcSize - ZSTD_minGain(srcSize, zc->appliedParams.cParams.strategy);
size_t const maxCSize =
srcSize - ZSTD_minGain(srcSize, zc->appliedParams.cParams.strategy);
FORWARD_IF_ERROR(cSize, "ZSTD_compressSuperBlock failed");
if (cSize != 0 && cSize < maxCSize + ZSTD_blockHeaderSize) {
ZSTD_blockState_confirmRepcodesAndEntropyTables(&zc->blockState);
@@ -3930,7 +4174,7 @@ static size_t ZSTD_compressBlock_targetCBlockSize_body(ZSTD_CCtx* zc,
}
}
}
}
} /* if (bss == ZSTDbss_compress)*/
DEBUGLOG(6, "Resorting to ZSTD_noCompressBlock()");
/* Superblock compression failed, attempt to emit a single no compress block.
@@ -3988,7 +4232,7 @@ static void ZSTD_overflowCorrectIfNeeded(ZSTD_matchState_t* ms,
* All blocks will be terminated, all input will be consumed.
* Function will issue an error if there is not enough `dstCapacity` to hold the compressed content.
* Frame is supposed already started (header already produced)
* @return : compressed size, or an error code
* @return : compressed size, or an error code
*/
static size_t ZSTD_compress_frameChunk(ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity,
@@ -4053,7 +4297,7 @@ static size_t ZSTD_compress_frameChunk(ZSTD_CCtx* cctx,
MEM_writeLE24(op, cBlockHeader);
cSize += ZSTD_blockHeaderSize;
}
}
} /* if (ZSTD_useTargetCBlockSize(&cctx->appliedParams))*/
ip += blockSize;
@@ -5682,6 +5926,7 @@ static size_t ZSTD_CCtx_init_compressStream2(ZSTD_CCtx* cctx,
params.useBlockSplitter = ZSTD_resolveBlockSplitterMode(params.useBlockSplitter, &params.cParams);
params.ldmParams.enableLdm = ZSTD_resolveEnableLdm(params.ldmParams.enableLdm, &params.cParams);
params.useRowMatchFinder = ZSTD_resolveRowMatchFinderMode(params.useRowMatchFinder, &params.cParams);
params.validateSequences = ZSTD_resolveExternalSequenceValidation(params.validateSequences);
#ifdef ZSTD_MULTITHREAD
if ((cctx->pledgedSrcSizePlusOne-1) <= ZSTDMT_JOBSIZE_MIN) {
@@ -5883,12 +6128,6 @@ size_t ZSTD_compress2(ZSTD_CCtx* cctx,
}
}
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;
/* ZSTD_validateSequence() :
* @offCode : is presumed to follow format required by ZSTD_storeSeq()
* @returns a ZSTD error code if sequence is not valid
@@ -5926,10 +6165,7 @@ static U32 ZSTD_finalizeOffBase(U32 rawOffset, const U32 rep[ZSTD_REP_NUM], U32
return offBase;
}
/* 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.
*/
static size_t
size_t
ZSTD_copySequencesToSeqStoreExplicitBlockDelim(ZSTD_CCtx* cctx,
ZSTD_sequencePosition* seqPos,
const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
@@ -5983,19 +6219,7 @@ ZSTD_copySequencesToSeqStoreExplicitBlockDelim(ZSTD_CCtx* cctx,
return 0;
}
/* 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.
*/
static size_t
size_t
ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos,
const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
const void* src, size_t blockSize)
@@ -6528,3 +6752,19 @@ ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSizeH
if (srcSizeHint == 0) srcSizeHint = ZSTD_CONTENTSIZE_UNKNOWN;
return ZSTD_getParams_internal(compressionLevel, srcSizeHint, dictSize, ZSTD_cpm_unknown);
}
void ZSTD_registerExternalMatchFinder(
ZSTD_CCtx* zc, void* mState,
ZSTD_externalMatchFinder_F* mFinder
) {
ZSTD_externalMatchCtx emctx = {
mState,
mFinder,
/* seqBuffer is allocated later (from the cwskp) */
NULL, /* seqBuffer */
0 /* seqBufferCapacity */
};
zc->externalMatchCtx = emctx;
zc->requestedParams.useExternalMatchFinder = 1;
}
lib/compress/zstd_compress_internal.h
+64 -7
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
@@ -118,12 +118,13 @@ typedef struct {
/** ZSTD_buildBlockEntropyStats() :
* Builds entropy for the block.
* @return : 0 on success or error code */
size_t ZSTD_buildBlockEntropyStats(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);
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 *
@@ -149,6 +150,12 @@ typedef struct {
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 {
@@ -339,6 +346,18 @@ struct ZSTD_CCtx_params_s {
/* Controls prefetching in some dictMatchState matchfinders */
ZSTD_paramSwitch_e prefetchCDictTables;
/* Controls whether zstd will fall back to an internal matchfinder
* 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_registerExternalMatchFinder(). */
int useExternalMatchFinder;
/* Adjust the max block size*/
size_t maxBlockSize;
}; /* typedef'd to ZSTD_CCtx_params within "zstd.h" */
#define COMPRESS_SEQUENCES_WORKSPACE_SIZE (sizeof(unsigned) * (MaxSeq + 2))
@@ -370,6 +389,14 @@ typedef struct {
ZSTD_entropyCTablesMetadata_t entropyMetadata;
} ZSTD_blockSplitCtx;
/* Context for block-level external matchfinder API */
typedef struct {
void* mState;
ZSTD_externalMatchFinder_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. */
@@ -439,6 +466,9 @@ struct ZSTD_CCtx_s {
/* Workspace for block splitter */
ZSTD_blockSplitCtx blockSplitCtx;
/* Workspace for external matchfinder */
ZSTD_externalMatchCtx externalMatchCtx;
};
typedef enum { ZSTD_dtlm_fast, ZSTD_dtlm_full } ZSTD_dictTableLoadMethod_e;
@@ -1410,4 +1440,31 @@ U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat);
*/
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);
/* 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);
#endif /* ZSTD_COMPRESS_H */
lib/compress/zstd_compress_literals.c
+43 -19
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
@@ -92,16 +92,37 @@ size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void*
return flSize+1;
}
size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
ZSTD_hufCTables_t* nextHuf,
ZSTD_strategy strategy, int disableLiteralCompression,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
void* entropyWorkspace, size_t entropyWorkspaceSize,
const int bmi2,
unsigned suspectUncompressible, HUF_depth_mode depthMode)
/* ZSTD_minLiteralsToCompress() :
* returns minimal amount of literals
* for literal compression to even be attempted.
* Minimum is made tighter as compression strategy increases.
*/
static size_t
ZSTD_minLiteralsToCompress(ZSTD_strategy strategy, HUF_repeat huf_repeat)
{
assert((int)strategy >= 0);
assert((int)strategy <= 9);
/* btultra2 : min 8 bytes;
* then 2x larger for each successive compression strategy
* max threshold 64 bytes */
{ int const shift = MIN(9-strategy, 3);
size_t const mintc = (huf_repeat == HUF_repeat_valid) ? 6 : 8 << shift;
DEBUGLOG(7, "minLiteralsToCompress = %zu", mintc);
return mintc;
}
}
size_t ZSTD_compressLiterals (
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
void* entropyWorkspace, size_t entropyWorkspaceSize,
const ZSTD_hufCTables_t* prevHuf,
ZSTD_hufCTables_t* nextHuf,
ZSTD_strategy strategy,
int disableLiteralCompression,
int suspectUncompressible,
int bmi2)
{
size_t const minGain = ZSTD_minGain(srcSize, strategy);
size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);
BYTE* const ostart = (BYTE*)dst;
U32 singleStream = srcSize < 256;
@@ -119,15 +140,14 @@ size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
if (disableLiteralCompression)
return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
/* small ? don't even attempt compression (speed opt) */
# define COMPRESS_LITERALS_SIZE_MIN 63
{ size_t const minLitSize = (prevHuf->repeatMode == HUF_repeat_valid) ? 6 : COMPRESS_LITERALS_SIZE_MIN;
if (srcSize <= minLitSize) return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
}
/* if too small, don't even attempt compression (speed opt) */
if (srcSize < ZSTD_minLiteralsToCompress(strategy, prevHuf->repeatMode))
return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
RETURN_ERROR_IF(dstCapacity < lhSize+1, dstSize_tooSmall, "not enough space for compression");
{ HUF_repeat repeat = prevHuf->repeatMode;
int const preferRepeat = (strategy < ZSTD_lazy) ? srcSize <= 1024 : 0;
HUF_depth_mode const depthMode = (strategy >= HUF_OPTIMAL_DEPTH_THRESHOLD) ? HUF_depth_optimal : HUF_depth_fast;
typedef size_t (*huf_compress_f)(void*, size_t, const void*, size_t, unsigned, unsigned, void*, size_t, HUF_CElt*, HUF_repeat*, int, int, unsigned, HUF_depth_mode);
huf_compress_f huf_compress;
if (repeat == HUF_repeat_valid && lhSize == 3) singleStream = 1;
@@ -146,10 +166,11 @@ size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
}
}
if ((cLitSize==0) || (cLitSize >= srcSize - minGain) || ERR_isError(cLitSize)) {
ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
}
{ size_t const minGain = ZSTD_minGain(srcSize, strategy);
if ((cLitSize==0) || (cLitSize >= srcSize - minGain) || ERR_isError(cLitSize)) {
ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
} }
if (cLitSize==1) {
ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize);
@@ -164,16 +185,19 @@ size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
switch(lhSize)
{
case 3: /* 2 - 2 - 10 - 10 */
if (!singleStream) assert(srcSize >= MIN_LITERALS_FOR_4_STREAMS);
{ U32 const lhc = hType + ((U32)(!singleStream) << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<14);
MEM_writeLE24(ostart, lhc);
break;
}
case 4: /* 2 - 2 - 14 - 14 */
assert(srcSize >= MIN_LITERALS_FOR_4_STREAMS);
{ U32 const lhc = hType + (2 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<18);
MEM_writeLE32(ostart, lhc);
break;
}
case 5: /* 2 - 2 - 18 - 18 */
assert(srcSize >= MIN_LITERALS_FOR_4_STREAMS);
{ U32 const lhc = hType + (3 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<22);
MEM_writeLE32(ostart, lhc);
ostart[4] = (BYTE)(cLitSize >> 10);
lib/compress/zstd_compress_literals.h
+12 -8
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
@@ -18,14 +18,18 @@ size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src,
size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize);
/* If suspectUncompressible then some sampling checks will be run to potentially skip huffman coding */
size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
ZSTD_hufCTables_t* nextHuf,
ZSTD_strategy strategy, int disableLiteralCompression,
void* dst, size_t dstCapacity,
/* ZSTD_compressLiterals():
* @entropyWorkspace: must be aligned on 4-bytes boundaries
* @entropyWorkspaceSize : must be >= HUF_WORKSPACE_SIZE
* @suspectUncompressible: sampling checks, to potentially skip huffman coding
*/
size_t ZSTD_compressLiterals (void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
void* entropyWorkspace, size_t entropyWorkspaceSize,
const int bmi2,
unsigned suspectUncompressible, HUF_depth_mode depthMode);
const ZSTD_hufCTables_t* prevHuf,
ZSTD_hufCTables_t* nextHuf,
ZSTD_strategy strategy, int disableLiteralCompression,
int suspectUncompressible,
int bmi2);
#endif /* ZSTD_COMPRESS_LITERALS_H */
lib/compress/zstd_compress_sequences.c
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_compress_sequences.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_compress_superblock.c
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_compress_superblock.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_cwksp.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_double_fast.c
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_double_fast.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_fast.c
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_fast.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_lazy.c
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_lazy.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_ldm.c
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_ldm.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_ldm_geartab.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstd_opt.c
+55 -35
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Przemyslaw Skibinski, Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
@@ -16,7 +16,7 @@
#define ZSTD_LITFREQ_ADD 2 /* scaling factor for litFreq, so that frequencies adapt faster to new stats */
#define ZSTD_MAX_PRICE (1<<30)
#define ZSTD_PREDEF_THRESHOLD 1024 /* if srcSize < ZSTD_PREDEF_THRESHOLD, symbols' cost is assumed static, directly determined by pre-defined distributions */
#define ZSTD_PREDEF_THRESHOLD 8 /* if srcSize < ZSTD_PREDEF_THRESHOLD, symbols' cost is assumed static, directly determined by pre-defined distributions */
/*-*************************************
@@ -26,27 +26,35 @@
#if 0 /* approximation at bit level (for tests) */
# define BITCOST_ACCURACY 0
# define BITCOST_MULTIPLIER (1 << BITCOST_ACCURACY)
# define WEIGHT(stat, opt) ((void)opt, ZSTD_bitWeight(stat))
# define WEIGHT(stat, opt) ((void)(opt), ZSTD_bitWeight(stat))
#elif 0 /* fractional bit accuracy (for tests) */
# define BITCOST_ACCURACY 8
# define BITCOST_MULTIPLIER (1 << BITCOST_ACCURACY)
# define WEIGHT(stat,opt) ((void)opt, ZSTD_fracWeight(stat))
# define WEIGHT(stat,opt) ((void)(opt), ZSTD_fracWeight(stat))
#else /* opt==approx, ultra==accurate */
# define BITCOST_ACCURACY 8
# define BITCOST_MULTIPLIER (1 << BITCOST_ACCURACY)
# define WEIGHT(stat,opt) (opt ? ZSTD_fracWeight(stat) : ZSTD_bitWeight(stat))
# define WEIGHT(stat,opt) ((opt) ? ZSTD_fracWeight(stat) : ZSTD_bitWeight(stat))
#endif
/* ZSTD_bitWeight() :
* provide estimated "cost" of a stat in full bits only */
MEM_STATIC U32 ZSTD_bitWeight(U32 stat)
{
return (ZSTD_highbit32(stat+1) * BITCOST_MULTIPLIER);
}
/* ZSTD_fracWeight() :
* provide fractional-bit "cost" of a stat,
* using linear interpolation approximation */
MEM_STATIC U32 ZSTD_fracWeight(U32 rawStat)
{
U32 const stat = rawStat + 1;
U32 const hb = ZSTD_highbit32(stat);
U32 const BWeight = hb * BITCOST_MULTIPLIER;
/* Fweight was meant for "Fractional weight"
* but it's effectively a value between 1 and 2
* using fixed point arithmetic */
U32 const FWeight = (stat << BITCOST_ACCURACY) >> hb;
U32 const weight = BWeight + FWeight;
assert(hb + BITCOST_ACCURACY < 31);
@@ -88,20 +96,26 @@ static U32 sum_u32(const unsigned table[], size_t nbElts)
return total;
}
static U32 ZSTD_downscaleStats(unsigned* table, U32 lastEltIndex, U32 shift)
typedef enum { base_0possible=0, base_1guaranteed=1 } base_directive_e;
static U32
ZSTD_downscaleStats(unsigned* table, U32 lastEltIndex, U32 shift, base_directive_e base1)
{
U32 s, sum=0;
DEBUGLOG(5, "ZSTD_downscaleStats (nbElts=%u, shift=%u)", (unsigned)lastEltIndex+1, (unsigned)shift);
DEBUGLOG(5, "ZSTD_downscaleStats (nbElts=%u, shift=%u)",
(unsigned)lastEltIndex+1, (unsigned)shift );
assert(shift < 30);
for (s=0; s<lastEltIndex+1; s++) {
table[s] = 1 + (table[s] >> shift);
sum += table[s];
unsigned const base = base1 ? 1 : (table[s]>0);
unsigned const newStat = base + (table[s] >> shift);
sum += newStat;
table[s] = newStat;
}
return sum;
}
/* ZSTD_scaleStats() :
* reduce all elements in table is sum too large
* reduce all elt frequencies in table if sum too large
* return the resulting sum of elements */
static U32 ZSTD_scaleStats(unsigned* table, U32 lastEltIndex, U32 logTarget)
{
@@ -110,7 +124,7 @@ static U32 ZSTD_scaleStats(unsigned* table, U32 lastEltIndex, U32 logTarget)
DEBUGLOG(5, "ZSTD_scaleStats (nbElts=%u, target=%u)", (unsigned)lastEltIndex+1, (unsigned)logTarget);
assert(logTarget < 30);
if (factor <= 1) return prevsum;
return ZSTD_downscaleStats(table, lastEltIndex, ZSTD_highbit32(factor));
return ZSTD_downscaleStats(table, lastEltIndex, ZSTD_highbit32(factor), base_1guaranteed);
}
/* ZSTD_rescaleFreqs() :
@@ -129,18 +143,22 @@ ZSTD_rescaleFreqs(optState_t* const optPtr,
DEBUGLOG(5, "ZSTD_rescaleFreqs (srcSize=%u)", (unsigned)srcSize);
optPtr->priceType = zop_dynamic;
if (optPtr->litLengthSum == 0) { /* first block : init */
if (srcSize <= ZSTD_PREDEF_THRESHOLD) { /* heuristic */
DEBUGLOG(5, "(srcSize <= ZSTD_PREDEF_THRESHOLD) => zop_predef");
if (optPtr->litLengthSum == 0) { /* no literals stats collected -> first block assumed -> init */
/* heuristic: use pre-defined stats for too small inputs */
if (srcSize <= ZSTD_PREDEF_THRESHOLD) {
DEBUGLOG(5, "srcSize <= %i : use predefined stats", ZSTD_PREDEF_THRESHOLD);
optPtr->priceType = zop_predef;
}
assert(optPtr->symbolCosts != NULL);
if (optPtr->symbolCosts->huf.repeatMode == HUF_repeat_valid) {
/* huffman table presumed generated by dictionary */
/* huffman stats covering the full value set : table presumed generated by dictionary */
optPtr->priceType = zop_dynamic;
if (compressedLiterals) {
/* generate literals statistics from huffman table */
unsigned lit;
assert(optPtr->litFreq != NULL);
optPtr->litSum = 0;
@@ -188,13 +206,14 @@ ZSTD_rescaleFreqs(optState_t* const optPtr,
optPtr->offCodeSum += optPtr->offCodeFreq[of];
} }
} else { /* not a dictionary */
} else { /* first block, no dictionary */
assert(optPtr->litFreq != NULL);
if (compressedLiterals) {
/* base initial cost of literals on direct frequency within src */
unsigned lit = MaxLit;
HIST_count_simple(optPtr->litFreq, &lit, src, srcSize); /* use raw first block to init statistics */
optPtr->litSum = ZSTD_downscaleStats(optPtr->litFreq, MaxLit, 8);
optPtr->litSum = ZSTD_downscaleStats(optPtr->litFreq, MaxLit, 8, base_0possible);
}
{ unsigned const baseLLfreqs[MaxLL+1] = {
@@ -224,10 +243,9 @@ ZSTD_rescaleFreqs(optState_t* const optPtr,
optPtr->offCodeSum = sum_u32(baseOFCfreqs, MaxOff+1);
}
}
} else { /* new block : re-use previous statistics, scaled down */
} else { /* new block : scale down accumulated statistics */
if (compressedLiterals)
optPtr->litSum = ZSTD_scaleStats(optPtr->litFreq, MaxLit, 12);
@@ -275,10 +293,11 @@ static U32 ZSTD_litLengthPrice(U32 const litLength, const optState_t* const optP
assert(litLength <= ZSTD_BLOCKSIZE_MAX);
if (optPtr->priceType == zop_predef)
return WEIGHT(litLength, optLevel);
/* We can't compute the litLength price for sizes >= ZSTD_BLOCKSIZE_MAX
* because it isn't representable in the zstd format. So instead just
* call it 1 bit more than ZSTD_BLOCKSIZE_MAX - 1. In this case the block
* would be all literals.
/* ZSTD_LLcode() can't compute litLength price for sizes >= ZSTD_BLOCKSIZE_MAX
* because it isn't representable in the zstd format.
* So instead just pretend it would cost 1 bit more than ZSTD_BLOCKSIZE_MAX - 1.
* In such a case, the block would be all literals.
*/
if (litLength == ZSTD_BLOCKSIZE_MAX)
return BITCOST_MULTIPLIER + ZSTD_litLengthPrice(ZSTD_BLOCKSIZE_MAX - 1, optPtr, optLevel);
@@ -292,7 +311,7 @@ static U32 ZSTD_litLengthPrice(U32 const litLength, const optState_t* const optP
}
/* ZSTD_getMatchPrice() :
* Provides the cost of the match part (offset + matchLength) of a sequence
* Provides the cost of the match part (offset + matchLength) of a sequence.
* Must be combined with ZSTD_fullLiteralsCost() to get the full cost of a sequence.
* @offBase : sumtype, representing an offset or a repcode, and using numeric representation of ZSTD_storeSeq()
* @optLevel: when <2, favors small offset for decompression speed (improved cache efficiency)
@@ -308,8 +327,9 @@ ZSTD_getMatchPrice(U32 const offBase,
U32 const mlBase = matchLength - MINMATCH;
assert(matchLength >= MINMATCH);
if (optPtr->priceType == zop_predef) /* fixed scheme, do not use statistics */
return WEIGHT(mlBase, optLevel) + ((16 + offCode) * BITCOST_MULTIPLIER);
if (optPtr->priceType == zop_predef) /* fixed scheme, does not use statistics */
return WEIGHT(mlBase, optLevel)
+ ((16 + offCode) * BITCOST_MULTIPLIER); /* emulated offset cost */
/* dynamic statistics */
price = (offCode * BITCOST_MULTIPLIER) + (optPtr->offCodeSumBasePrice - WEIGHT(optPtr->offCodeFreq[offCode], optLevel));
@@ -347,7 +367,7 @@ static void ZSTD_updateStats(optState_t* const optPtr,
optPtr->litLengthSum++;
}
/* offset code : expected to follow storeSeq() numeric representation */
/* offset code : follows storeSeq() numeric representation */
{ U32 const offCode = ZSTD_highbit32(offBase);
assert(offCode <= MaxOff);
optPtr->offCodeFreq[offCode]++;
@@ -1352,7 +1372,7 @@ size_t ZSTD_compressBlock_btopt(
/* 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, hence its contract must be respected.
* this function cannot error out, its narrow contract must be respected.
*/
static void
ZSTD_initStats_ultra(ZSTD_matchState_t* ms,
@@ -1371,7 +1391,7 @@ ZSTD_initStats_ultra(ZSTD_matchState_t* ms,
ZSTD_compressBlock_opt2(ms, seqStore, tmpRep, src, srcSize, ZSTD_noDict); /* generate stats into ms->opt*/
/* invalidate first scan from history */
/* invalidate first scan from history, only keep entropy stats */
ZSTD_resetSeqStore(seqStore);
ms->window.base -= srcSize;
ms->window.dictLimit += (U32)srcSize;
@@ -1395,20 +1415,20 @@ size_t ZSTD_compressBlock_btultra2(
U32 const curr = (U32)((const BYTE*)src - ms->window.base);
DEBUGLOG(5, "ZSTD_compressBlock_btultra2 (srcSize=%zu)", srcSize);
/* 2-pass strategy:
/* 2-passes strategy:
* this strategy makes a first pass over first block to collect statistics
* and seed next round's statistics with it.
* After 1st pass, function forgets everything, and starts a new block.
* in order to seed next round's statistics with it.
* After 1st pass, function forgets history, and starts a new block.
* 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 */
&& (ms->window.dictLimit == ms->window.lowLimit) /* no dictionary */
&& (curr == ms->window.dictLimit) /* start of frame, nothing already loaded nor skipped */
&& (srcSize > ZSTD_PREDEF_THRESHOLD)
&& (curr == ms->window.dictLimit) /* start of frame, nothing already loaded nor skipped */
&& (srcSize > ZSTD_PREDEF_THRESHOLD) /* input large enough to not employ default stats */
) {
ZSTD_initStats_ultra(ms, seqStore, rep, src, srcSize);
}
lib/compress/zstd_opt.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstdmt_compress.c
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
lib/compress/zstdmt_compress.h
+1 -1
View File
@@ -1,5 +1,5 @@
/*
* Copyright (c) Yann Collet, Facebook, Inc.
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the