diff options
Diffstat (limited to 'contrib/zstd/zstd_compress.c')
| -rw-r--r-- | contrib/zstd/zstd_compress.c | 4643 |
1 files changed, 2979 insertions, 1664 deletions
diff --git a/contrib/zstd/zstd_compress.c b/contrib/zstd/zstd_compress.c index b0e9195dd..526cb21f6 100644 --- a/contrib/zstd/zstd_compress.c +++ b/contrib/zstd/zstd_compress.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. + * Copyright (c) 2016-2020, Yann Collet, Facebook, Inc. * All rights reserved. * * This source code is licensed under both the BSD-style license (found in the @@ -8,50 +8,41 @@ * You may select, at your option, one of the above-listed licenses. */ - -/*-************************************* -* Tuning parameters -***************************************/ -#ifndef ZSTD_CLEVEL_DEFAULT -# define ZSTD_CLEVEL_DEFAULT 3 -#endif - - /*-************************************* * Dependencies ***************************************/ +#include <limits.h> /* INT_MAX */ #include <string.h> /* memset */ +#include "cpu.h" #include "mem.h" +#include "hist.h" /* HIST_countFast_wksp */ #define FSE_STATIC_LINKING_ONLY /* FSE_encodeSymbol */ #include "fse.h" #define HUF_STATIC_LINKING_ONLY #include "huf.h" -#include "zstd_compress.h" +#include "zstd_compress_internal.h" +#include "zstd_compress_sequences.h" +#include "zstd_compress_literals.h" #include "zstd_fast.h" #include "zstd_double_fast.h" #include "zstd_lazy.h" #include "zstd_opt.h" #include "zstd_ldm.h" +#include "zstd_compress_superblock.h" /*-************************************* * 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(). + */ size_t ZSTD_compressBound(size_t srcSize) { - size_t const lowLimit = 256 KB; - size_t const margin = (srcSize < lowLimit) ? (lowLimit-srcSize) >> 12 : 0; /* from 64 to 0 */ - return srcSize + (srcSize >> 8) + margin; -} - - -/*-************************************* -* Sequence storage -***************************************/ -static void ZSTD_resetSeqStore(seqStore_t* ssPtr) -{ - ssPtr->lit = ssPtr->litStart; - ssPtr->sequences = ssPtr->sequencesStart; - ssPtr->longLengthID = 0; + return ZSTD_COMPRESSBOUND(srcSize); } @@ -59,10 +50,15 @@ static void ZSTD_resetSeqStore(seqStore_t* ssPtr) * Context memory management ***************************************/ struct ZSTD_CDict_s { - void* dictBuffer; const void* dictContent; size_t dictContentSize; - ZSTD_CCtx* refContext; + U32* entropyWorkspace; /* entropy workspace of HUF_WORKSPACE_SIZE bytes */ + ZSTD_cwksp workspace; + ZSTD_matchState_t matchState; + ZSTD_compressedBlockState_t cBlockState; + ZSTD_customMem customMem; + U32 dictID; + int compressionLevel; /* 0 indicates that advanced API was used to select CDict params */ }; /* typedef'd to ZSTD_CDict within "zstd.h" */ ZSTD_CCtx* ZSTD_createCCtx(void) @@ -70,53 +66,97 @@ ZSTD_CCtx* ZSTD_createCCtx(void) return ZSTD_createCCtx_advanced(ZSTD_defaultCMem); } -ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem) +static void ZSTD_initCCtx(ZSTD_CCtx* cctx, ZSTD_customMem memManager) { - ZSTD_CCtx* cctx; - - if (!customMem.customAlloc ^ !customMem.customFree) return NULL; + assert(cctx != NULL); + memset(cctx, 0, sizeof(*cctx)); + cctx->customMem = memManager; + cctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid()); + { size_t const err = ZSTD_CCtx_reset(cctx, ZSTD_reset_parameters); + assert(!ZSTD_isError(err)); + (void)err; + } +} - cctx = (ZSTD_CCtx*) ZSTD_calloc(sizeof(ZSTD_CCtx), customMem); - if (!cctx) return NULL; - cctx->customMem = customMem; - cctx->requestedParams.compressionLevel = ZSTD_CLEVEL_DEFAULT; +ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem) +{ ZSTD_STATIC_ASSERT(zcss_init==0); ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN==(0ULL - 1)); - return cctx; + if (!customMem.customAlloc ^ !customMem.customFree) return NULL; + { ZSTD_CCtx* const cctx = (ZSTD_CCtx*)ZSTD_malloc(sizeof(ZSTD_CCtx), customMem); + if (!cctx) return NULL; + ZSTD_initCCtx(cctx, customMem); + return cctx; + } } -ZSTD_CCtx* ZSTD_initStaticCCtx(void *workspace, size_t workspaceSize) +ZSTD_CCtx* ZSTD_initStaticCCtx(void* workspace, size_t workspaceSize) { - ZSTD_CCtx* const cctx = (ZSTD_CCtx*) workspace; + ZSTD_cwksp ws; + ZSTD_CCtx* cctx; if (workspaceSize <= sizeof(ZSTD_CCtx)) return NULL; /* minimum size */ if ((size_t)workspace & 7) return NULL; /* must be 8-aligned */ - memset(workspace, 0, workspaceSize); /* may be a bit generous, could memset be smaller ? */ - cctx->staticSize = workspaceSize; - cctx->workSpace = (void*)(cctx+1); - cctx->workSpaceSize = workspaceSize - sizeof(ZSTD_CCtx); + ZSTD_cwksp_init(&ws, workspace, workspaceSize); - /* entropy space (never moves) */ - if (cctx->workSpaceSize < sizeof(ZSTD_entropyCTables_t)) return NULL; - assert(((size_t)cctx->workSpace & (sizeof(void*)-1)) == 0); /* ensure correct alignment */ - cctx->entropy = (ZSTD_entropyCTables_t*)cctx->workSpace; + cctx = (ZSTD_CCtx*)ZSTD_cwksp_reserve_object(&ws, sizeof(ZSTD_CCtx)); + if (cctx == NULL) return NULL; + memset(cctx, 0, sizeof(ZSTD_CCtx)); + ZSTD_cwksp_move(&cctx->workspace, &ws); + cctx->staticSize = workspaceSize; + + /* statically sized space. entropyWorkspace never moves (but prev/next block swap places) */ + if (!ZSTD_cwksp_check_available(&cctx->workspace, HUF_WORKSPACE_SIZE + 2 * sizeof(ZSTD_compressedBlockState_t))) return NULL; + cctx->blockState.prevCBlock = (ZSTD_compressedBlockState_t*)ZSTD_cwksp_reserve_object(&cctx->workspace, sizeof(ZSTD_compressedBlockState_t)); + cctx->blockState.nextCBlock = (ZSTD_compressedBlockState_t*)ZSTD_cwksp_reserve_object(&cctx->workspace, sizeof(ZSTD_compressedBlockState_t)); + cctx->entropyWorkspace = (U32*)ZSTD_cwksp_reserve_object(&cctx->workspace, HUF_WORKSPACE_SIZE); + cctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid()); return cctx; } -size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx) +/** + * Clears and frees all of the dictionaries in the CCtx. + */ +static void ZSTD_clearAllDicts(ZSTD_CCtx* cctx) { - if (cctx==NULL) return 0; /* support free on NULL */ - if (cctx->staticSize) return ERROR(memory_allocation); /* not compatible with static CCtx */ - ZSTD_free(cctx->workSpace, cctx->customMem); - cctx->workSpace = NULL; - ZSTD_freeCDict(cctx->cdictLocal); - cctx->cdictLocal = NULL; + ZSTD_free(cctx->localDict.dictBuffer, cctx->customMem); + ZSTD_freeCDict(cctx->localDict.cdict); + memset(&cctx->localDict, 0, sizeof(cctx->localDict)); + memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict)); + cctx->cdict = NULL; +} + +static size_t ZSTD_sizeof_localDict(ZSTD_localDict dict) +{ + size_t const bufferSize = dict.dictBuffer != NULL ? dict.dictSize : 0; + size_t const cdictSize = ZSTD_sizeof_CDict(dict.cdict); + return bufferSize + cdictSize; +} + +static void ZSTD_freeCCtxContent(ZSTD_CCtx* cctx) +{ + assert(cctx != NULL); + assert(cctx->staticSize == 0); + ZSTD_clearAllDicts(cctx); #ifdef ZSTD_MULTITHREAD - ZSTDMT_freeCCtx(cctx->mtctx); - cctx->mtctx = NULL; + ZSTDMT_freeCCtx(cctx->mtctx); cctx->mtctx = NULL; #endif - ZSTD_free(cctx, cctx->customMem); - return 0; /* reserved as a potential error code in the future */ + ZSTD_cwksp_free(&cctx->workspace, cctx->customMem); +} + +size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx) +{ + if (cctx==NULL) return 0; /* support free on NULL */ + RETURN_ERROR_IF(cctx->staticSize, memory_allocation, + "not compatible with static CCtx"); + { + int cctxInWorkspace = ZSTD_cwksp_owns_buffer(&cctx->workspace, cctx); + ZSTD_freeCCtxContent(cctx); + if (!cctxInWorkspace) { + ZSTD_free(cctx, cctx->customMem); + } + } + return 0; } @@ -125,7 +165,7 @@ static size_t ZSTD_sizeof_mtctx(const ZSTD_CCtx* cctx) #ifdef ZSTD_MULTITHREAD return ZSTDMT_sizeof_CCtx(cctx->mtctx); #else - (void) cctx; + (void)cctx; return 0; #endif } @@ -134,12 +174,10 @@ static size_t ZSTD_sizeof_mtctx(const ZSTD_CCtx* cctx) size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx) { if (cctx==NULL) return 0; /* support sizeof on NULL */ - DEBUGLOG(3, "sizeof(*cctx) : %u", (U32)sizeof(*cctx)); - DEBUGLOG(3, "workSpaceSize (including streaming buffers): %u", (U32)cctx->workSpaceSize); - DEBUGLOG(3, "inner cdict : %u", (U32)ZSTD_sizeof_CDict(cctx->cdictLocal)); - DEBUGLOG(3, "inner MTCTX : %u", (U32)ZSTD_sizeof_mtctx(cctx)); - return sizeof(*cctx) + cctx->workSpaceSize - + ZSTD_sizeof_CDict(cctx->cdictLocal) + /* cctx may be in the workspace */ + return (cctx->workspace.workspace == cctx ? 0 : sizeof(*cctx)) + + ZSTD_cwksp_sizeof(&cctx->workspace) + + ZSTD_sizeof_localDict(cctx->localDict) + ZSTD_sizeof_mtctx(cctx); } @@ -151,40 +189,15 @@ size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs) /* private API call, for dictBuilder only */ const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx) { return &(ctx->seqStore); } -#define ZSTD_CLEVEL_CUSTOM 999 - -static ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams( - ZSTD_CCtx_params params, U64 srcSizeHint, size_t dictSize) -{ - return (params.compressionLevel == ZSTD_CLEVEL_CUSTOM ? - params.cParams : - ZSTD_getCParams(params.compressionLevel, srcSizeHint, dictSize)); -} - -static void ZSTD_cLevelToCCtxParams_srcSize(ZSTD_CCtx_params* params, U64 srcSize) -{ - params->cParams = ZSTD_getCParamsFromCCtxParams(*params, srcSize, 0); - params->compressionLevel = ZSTD_CLEVEL_CUSTOM; -} - -static void ZSTD_cLevelToCParams(ZSTD_CCtx* cctx) -{ - ZSTD_cLevelToCCtxParams_srcSize( - &cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1); -} - -static void ZSTD_cLevelToCCtxParams(ZSTD_CCtx_params* params) -{ - ZSTD_cLevelToCCtxParams_srcSize(params, 0); -} - static ZSTD_CCtx_params ZSTD_makeCCtxParamsFromCParams( ZSTD_compressionParameters cParams) { ZSTD_CCtx_params cctxParams; memset(&cctxParams, 0, sizeof(cctxParams)); cctxParams.cParams = cParams; - cctxParams.compressionLevel = ZSTD_CLEVEL_CUSTOM; + cctxParams.compressionLevel = ZSTD_CLEVEL_DEFAULT; /* should not matter, as all cParams are presumed properly defined */ + assert(!ZSTD_checkCParams(cParams)); + cctxParams.fParams.contentSizeFlag = 1; return cctxParams; } @@ -198,6 +211,7 @@ static ZSTD_CCtx_params* ZSTD_createCCtxParams_advanced( if (!params) { return NULL; } params->customMem = customMem; params->compressionLevel = ZSTD_CLEVEL_DEFAULT; + params->fParams.contentSizeFlag = 1; return params; } @@ -213,324 +227,690 @@ size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params) return 0; } -size_t ZSTD_resetCCtxParams(ZSTD_CCtx_params* params) +size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params) { - return ZSTD_initCCtxParams(params, ZSTD_CLEVEL_DEFAULT); + return ZSTD_CCtxParams_init(params, ZSTD_CLEVEL_DEFAULT); } -size_t ZSTD_initCCtxParams(ZSTD_CCtx_params* cctxParams, int compressionLevel) { - if (!cctxParams) { return ERROR(GENERIC); } +size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel) { + RETURN_ERROR_IF(!cctxParams, GENERIC, "NULL pointer!"); memset(cctxParams, 0, sizeof(*cctxParams)); cctxParams->compressionLevel = compressionLevel; + cctxParams->fParams.contentSizeFlag = 1; return 0; } -size_t ZSTD_initCCtxParams_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params) +size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params) { - if (!cctxParams) { return ERROR(GENERIC); } - CHECK_F( ZSTD_checkCParams(params.cParams) ); + RETURN_ERROR_IF(!cctxParams, GENERIC, "NULL pointer!"); + FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) , ""); memset(cctxParams, 0, sizeof(*cctxParams)); + assert(!ZSTD_checkCParams(params.cParams)); cctxParams->cParams = params.cParams; cctxParams->fParams = params.fParams; - cctxParams->compressionLevel = ZSTD_CLEVEL_CUSTOM; + cctxParams->compressionLevel = ZSTD_CLEVEL_DEFAULT; /* should not matter, as all cParams are presumed properly defined */ return 0; } +/* ZSTD_assignParamsToCCtxParams() : + * params is presumed valid at this stage */ static ZSTD_CCtx_params ZSTD_assignParamsToCCtxParams( - ZSTD_CCtx_params cctxParams, ZSTD_parameters params) + const ZSTD_CCtx_params* cctxParams, const ZSTD_parameters* params) { - ZSTD_CCtx_params ret = cctxParams; - ret.cParams = params.cParams; - ret.fParams = params.fParams; + ZSTD_CCtx_params ret = *cctxParams; + assert(!ZSTD_checkCParams(params->cParams)); + ret.cParams = params->cParams; + ret.fParams = params->fParams; + ret.compressionLevel = ZSTD_CLEVEL_DEFAULT; /* should not matter, as all cParams are presumed properly defined */ return ret; } -#define CLAMPCHECK(val,min,max) { \ - if (((val)<(min)) | ((val)>(max))) { \ - return ERROR(parameter_outOfBound); \ -} } - -size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, unsigned value) +ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_cParameter param) { - if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); + ZSTD_bounds bounds = { 0, 0, 0 }; switch(param) { - case ZSTD_p_compressionLevel: - if (value == 0) return 0; /* special value : 0 means "don't change anything" */ - if (cctx->cdict) return ERROR(stage_wrong); - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); - - case ZSTD_p_windowLog: - case ZSTD_p_hashLog: - case ZSTD_p_chainLog: - case ZSTD_p_searchLog: - case ZSTD_p_minMatch: - case ZSTD_p_targetLength: - case ZSTD_p_compressionStrategy: - if (value == 0) return 0; /* special value : 0 means "don't change anything" */ - if (cctx->cdict) return ERROR(stage_wrong); - ZSTD_cLevelToCParams(cctx); /* Can optimize if srcSize is known */ - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); - - case ZSTD_p_contentSizeFlag: - case ZSTD_p_checksumFlag: - case ZSTD_p_dictIDFlag: - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); - - case ZSTD_p_forceMaxWindow : /* Force back-references to remain < windowSize, - * even when referencing into Dictionary content - * default : 0 when using a CDict, 1 when using a Prefix */ - cctx->loadedDictEnd = 0; - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); - - case ZSTD_p_nbThreads: - if (value==0) return 0; - DEBUGLOG(5, " setting nbThreads : %u", value); - if (value > 1 && cctx->staticSize) { - return ERROR(parameter_unsupported); /* MT not compatible with static alloc */ - } - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); - - case ZSTD_p_jobSize: - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); + case ZSTD_c_compressionLevel: + bounds.lowerBound = ZSTD_minCLevel(); + bounds.upperBound = ZSTD_maxCLevel(); + return bounds; + + case ZSTD_c_windowLog: + bounds.lowerBound = ZSTD_WINDOWLOG_MIN; + bounds.upperBound = ZSTD_WINDOWLOG_MAX; + return bounds; + + case ZSTD_c_hashLog: + bounds.lowerBound = ZSTD_HASHLOG_MIN; + bounds.upperBound = ZSTD_HASHLOG_MAX; + return bounds; + + case ZSTD_c_chainLog: + bounds.lowerBound = ZSTD_CHAINLOG_MIN; + bounds.upperBound = ZSTD_CHAINLOG_MAX; + return bounds; + + case ZSTD_c_searchLog: + bounds.lowerBound = ZSTD_SEARCHLOG_MIN; + bounds.upperBound = ZSTD_SEARCHLOG_MAX; + return bounds; + + case ZSTD_c_minMatch: + bounds.lowerBound = ZSTD_MINMATCH_MIN; + bounds.upperBound = ZSTD_MINMATCH_MAX; + return bounds; + + case ZSTD_c_targetLength: + bounds.lowerBound = ZSTD_TARGETLENGTH_MIN; + bounds.upperBound = ZSTD_TARGETLENGTH_MAX; + return bounds; + + case ZSTD_c_strategy: + bounds.lowerBound = ZSTD_STRATEGY_MIN; + bounds.upperBound = ZSTD_STRATEGY_MAX; + return bounds; + + case ZSTD_c_contentSizeFlag: + bounds.lowerBound = 0; + bounds.upperBound = 1; + return bounds; + + case ZSTD_c_checksumFlag: + bounds.lowerBound = 0; + bounds.upperBound = 1; + return bounds; + + case ZSTD_c_dictIDFlag: + bounds.lowerBound = 0; + bounds.upperBound = 1; + return bounds; + + case ZSTD_c_nbWorkers: + bounds.lowerBound = 0; +#ifdef ZSTD_MULTITHREAD + bounds.upperBound = ZSTDMT_NBWORKERS_MAX; +#else + bounds.upperBound = 0; +#endif + return bounds; - case ZSTD_p_overlapSizeLog: - DEBUGLOG(5, " setting overlap with nbThreads == %u", cctx->requestedParams.nbThreads); - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); + case ZSTD_c_jobSize: + bounds.lowerBound = 0; +#ifdef ZSTD_MULTITHREAD + bounds.upperBound = ZSTDMT_JOBSIZE_MAX; +#else + bounds.upperBound = 0; +#endif + return bounds; - case ZSTD_p_enableLongDistanceMatching: - if (cctx->cdict) return ERROR(stage_wrong); - if (value != 0) { - ZSTD_cLevelToCParams(cctx); - } - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); + case ZSTD_c_overlapLog: +#ifdef ZSTD_MULTITHREAD + bounds.lowerBound = ZSTD_OVERLAPLOG_MIN; + bounds.upperBound = ZSTD_OVERLAPLOG_MAX; +#else + bounds.lowerBound = 0; + bounds.upperBound = 0; +#endif + return bounds; + + case ZSTD_c_enableLongDistanceMatching: + bounds.lowerBound = 0; + bounds.upperBound = 1; + return bounds; + + case ZSTD_c_ldmHashLog: + bounds.lowerBound = ZSTD_LDM_HASHLOG_MIN; + bounds.upperBound = ZSTD_LDM_HASHLOG_MAX; + return bounds; + + case ZSTD_c_ldmMinMatch: + bounds.lowerBound = ZSTD_LDM_MINMATCH_MIN; + bounds.upperBound = ZSTD_LDM_MINMATCH_MAX; + return bounds; + + case ZSTD_c_ldmBucketSizeLog: + bounds.lowerBound = ZSTD_LDM_BUCKETSIZELOG_MIN; + bounds.upperBound = ZSTD_LDM_BUCKETSIZELOG_MAX; + return bounds; + + case ZSTD_c_ldmHashRateLog: + bounds.lowerBound = ZSTD_LDM_HASHRATELOG_MIN; + bounds.upperBound = ZSTD_LDM_HASHRATELOG_MAX; + return bounds; + + /* experimental parameters */ + case ZSTD_c_rsyncable: + bounds.lowerBound = 0; + bounds.upperBound = 1; + return bounds; + + case ZSTD_c_forceMaxWindow : + bounds.lowerBound = 0; + bounds.upperBound = 1; + return bounds; + + case ZSTD_c_format: + ZSTD_STATIC_ASSERT(ZSTD_f_zstd1 < ZSTD_f_zstd1_magicless); + bounds.lowerBound = ZSTD_f_zstd1; + bounds.upperBound = ZSTD_f_zstd1_magicless; /* note : how to ensure at compile time that this is the highest value enum ? */ + return bounds; + + case ZSTD_c_forceAttachDict: + ZSTD_STATIC_ASSERT(ZSTD_dictDefaultAttach < ZSTD_dictForceCopy); + bounds.lowerBound = ZSTD_dictDefaultAttach; + bounds.upperBound = ZSTD_dictForceLoad; /* note : how to ensure at compile time that this is the highest value enum ? */ + return bounds; + + case ZSTD_c_literalCompressionMode: + ZSTD_STATIC_ASSERT(ZSTD_lcm_auto < ZSTD_lcm_huffman && ZSTD_lcm_huffman < ZSTD_lcm_uncompressed); + bounds.lowerBound = ZSTD_lcm_auto; + bounds.upperBound = ZSTD_lcm_uncompressed; + return bounds; + + case ZSTD_c_targetCBlockSize: + bounds.lowerBound = ZSTD_TARGETCBLOCKSIZE_MIN; + bounds.upperBound = ZSTD_TARGETCBLOCKSIZE_MAX; + return bounds; + + case ZSTD_c_srcSizeHint: + bounds.lowerBound = ZSTD_SRCSIZEHINT_MIN; + bounds.upperBound = ZSTD_SRCSIZEHINT_MAX; + return bounds; - case ZSTD_p_ldmHashLog: - case ZSTD_p_ldmMinMatch: - if (value == 0) return 0; /* special value : 0 means "don't change anything" */ - if (cctx->cdict) return ERROR(stage_wrong); - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); + default: + bounds.error = ERROR(parameter_unsupported); + return bounds; + } +} - case ZSTD_p_ldmBucketSizeLog: - case ZSTD_p_ldmHashEveryLog: - if (cctx->cdict) return ERROR(stage_wrong); - return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); +/* ZSTD_cParam_clampBounds: + * Clamps the value into the bounded range. + */ +static size_t ZSTD_cParam_clampBounds(ZSTD_cParameter cParam, int* value) +{ + ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); + if (ZSTD_isError(bounds.error)) return bounds.error; + if (*value < bounds.lowerBound) *value = bounds.lowerBound; + if (*value > bounds.upperBound) *value = bounds.upperBound; + return 0; +} - default: return ERROR(parameter_unsupported); - } +#define BOUNDCHECK(cParam, val) { \ + RETURN_ERROR_IF(!ZSTD_cParam_withinBounds(cParam,val), \ + parameter_outOfBound, "Param out of bounds"); \ } -size_t ZSTD_CCtxParam_setParameter( - ZSTD_CCtx_params* params, ZSTD_cParameter param, unsigned value) + +static int ZSTD_isUpdateAuthorized(ZSTD_cParameter param) { switch(param) { - case ZSTD_p_compressionLevel : - if ((int)value > ZSTD_maxCLevel()) value = ZSTD_maxCLevel(); - if (value == 0) return 0; - params->compressionLevel = value; - return 0; - - case ZSTD_p_windowLog : - if (value == 0) return 0; - CLAMPCHECK(value, ZSTD_WINDOWLOG_MIN, ZSTD_WINDOWLOG_MAX); - ZSTD_cLevelToCCtxParams(params); - params->cParams.windowLog = value; - return 0; + case ZSTD_c_compressionLevel: + case ZSTD_c_hashLog: + case ZSTD_c_chainLog: + case ZSTD_c_searchLog: + case ZSTD_c_minMatch: + case ZSTD_c_targetLength: + case ZSTD_c_strategy: + return 1; - case ZSTD_p_hashLog : - if (value == 0) return 0; - CLAMPCHECK(value, ZSTD_HASHLOG_MIN, ZSTD_HASHLOG_MAX); - ZSTD_cLevelToCCtxParams(params); - params->cParams.hashLog = value; + case ZSTD_c_format: + case ZSTD_c_windowLog: + case ZSTD_c_contentSizeFlag: + case ZSTD_c_checksumFlag: + case ZSTD_c_dictIDFlag: + case ZSTD_c_forceMaxWindow : + case ZSTD_c_nbWorkers: + case ZSTD_c_jobSize: + case ZSTD_c_overlapLog: + case ZSTD_c_rsyncable: + case ZSTD_c_enableLongDistanceMatching: + case ZSTD_c_ldmHashLog: + case ZSTD_c_ldmMinMatch: + case ZSTD_c_ldmBucketSizeLog: + case ZSTD_c_ldmHashRateLog: + case ZSTD_c_forceAttachDict: + case ZSTD_c_literalCompressionMode: + case ZSTD_c_targetCBlockSize: + case ZSTD_c_srcSizeHint: + default: return 0; + } +} - case ZSTD_p_chainLog : - if (value == 0) return 0; - CLAMPCHECK(value, ZSTD_CHAINLOG_MIN, ZSTD_CHAINLOG_MAX); - ZSTD_cLevelToCCtxParams(params); - params->cParams.chainLog = value; - return 0; +size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value) +{ + DEBUGLOG(4, "ZSTD_CCtx_setParameter (%i, %i)", (int)param, value); + if (cctx->streamStage != zcss_init) { + if (ZSTD_isUpdateAuthorized(param)) { + cctx->cParamsChanged = 1; + } else { + RETURN_ERROR(stage_wrong, "can only set params in ctx init stage"); + } } - case ZSTD_p_searchLog : - if (value == 0) return 0; - CLAMPCHECK(value, ZSTD_SEARCHLOG_MIN, ZSTD_SEARCHLOG_MAX); - ZSTD_cLevelToCCtxParams(params); - params->cParams.searchLog = value; - return 0; + switch(param) + { + case ZSTD_c_nbWorkers: + RETURN_ERROR_IF((value!=0) && cctx->staticSize, parameter_unsupported, + "MT not compatible with static alloc"); + break; - case ZSTD_p_minMatch : - if (value == 0) return 0; - CLAMPCHECK(value, ZSTD_SEARCHLENGTH_MIN, ZSTD_SEARCHLENGTH_MAX); - ZSTD_cLevelToCCtxParams(params); - params->cParams.searchLength = value; - return 0; + case ZSTD_c_compressionLevel: + case ZSTD_c_windowLog: + case ZSTD_c_hashLog: + case ZSTD_c_chainLog: + case ZSTD_c_searchLog: + case ZSTD_c_minMatch: + case ZSTD_c_targetLength: + case ZSTD_c_strategy: + case ZSTD_c_ldmHashRateLog: + case ZSTD_c_format: + case ZSTD_c_contentSizeFlag: + case ZSTD_c_checksumFlag: + case ZSTD_c_dictIDFlag: + case ZSTD_c_forceMaxWindow: + case ZSTD_c_forceAttachDict: + case ZSTD_c_literalCompressionMode: + case ZSTD_c_jobSize: + case ZSTD_c_overlapLog: + case ZSTD_c_rsyncable: + case ZSTD_c_enableLongDistanceMatching: + case ZSTD_c_ldmHashLog: + case ZSTD_c_ldmMinMatch: + case ZSTD_c_ldmBucketSizeLog: + case ZSTD_c_targetCBlockSize: + case ZSTD_c_srcSizeHint: + break; - case ZSTD_p_targetLength : - if (value == 0) return 0; - CLAMPCHECK(value, ZSTD_TARGETLENGTH_MIN, ZSTD_TARGETLENGTH_MAX); - ZSTD_cLevelToCCtxParams(params); - params->cParams.targetLength = value; - return 0; + default: RETURN_ERROR(parameter_unsupported, "unknown parameter"); + } + return ZSTD_CCtxParams_setParameter(&cctx->requestedParams, param, value); +} - case ZSTD_p_compressionStrategy : - if (value == 0) return 0; - CLAMPCHECK(value, (unsigned)ZSTD_fast, (unsigned)ZSTD_btultra); - ZSTD_cLevelToCCtxParams(params); - params->cParams.strategy = (ZSTD_strategy)value; - return 0; +size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams, + ZSTD_cParameter param, int value) +{ + DEBUGLOG(4, "ZSTD_CCtxParams_setParameter (%i, %i)", (int)param, value); + switch(param) + { + case ZSTD_c_format : + BOUNDCHECK(ZSTD_c_format, value); + CCtxParams->format = (ZSTD_format_e)value; + return (size_t)CCtxParams->format; + + case ZSTD_c_compressionLevel : { + FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value), ""); + if (value) { /* 0 : does not change current level */ + CCtxParams->compressionLevel = value; + } + if (CCtxParams->compressionLevel >= 0) return (size_t)CCtxParams->compressionLevel; + return 0; /* return type (size_t) cannot represent negative values */ + } - case ZSTD_p_contentSizeFlag : + case ZSTD_c_windowLog : + if (value!=0) /* 0 => use default */ + BOUNDCHECK(ZSTD_c_windowLog, value); + CCtxParams->cParams.windowLog = (U32)value; + return CCtxParams->cParams.windowLog; + + case ZSTD_c_hashLog : + if (value!=0) /* 0 => use default */ + BOUNDCHECK(ZSTD_c_hashLog, value); + CCtxParams->cParams.hashLog = (U32)value; + return CCtxParams->cParams.hashLog; + + case ZSTD_c_chainLog : + if (value!=0) /* 0 => use default */ + BOUNDCHECK(ZSTD_c_chainLog, value); + CCtxParams->cParams.chainLog = (U32)value; + return CCtxParams->cParams.chainLog; + + case ZSTD_c_searchLog : + if (value!=0) /* 0 => use default */ + BOUNDCHECK(ZSTD_c_searchLog, value); + CCtxParams->cParams.searchLog = (U32)value; + return (size_t)value; + + case ZSTD_c_minMatch : + if (value!=0) /* 0 => use default */ + BOUNDCHECK(ZSTD_c_minMatch, value); + CCtxParams->cParams.minMatch = value; + return CCtxParams->cParams.minMatch; + + case ZSTD_c_targetLength : + BOUNDCHECK(ZSTD_c_targetLength, value); + CCtxParams->cParams.targetLength = value; + return CCtxParams->cParams.targetLength; + + case ZSTD_c_strategy : + if (value!=0) /* 0 => use default */ + BOUNDCHECK(ZSTD_c_strategy, value); + CCtxParams->cParams.strategy = (ZSTD_strategy)value; + return (size_t)CCtxParams->cParams.strategy; + + case ZSTD_c_contentSizeFlag : /* Content size written in frame header _when known_ (default:1) */ - DEBUGLOG(5, "set content size flag = %u", (value>0)); - params->fParams.contentSizeFlag = value > 0; - return 0; + DEBUGLOG(4, "set content size flag = %u", (value!=0)); + CCtxParams->fParams.contentSizeFlag = value != 0; + return CCtxParams->fParams.contentSizeFlag; - case ZSTD_p_checksumFlag : + case ZSTD_c_checksumFlag : /* A 32-bits content checksum will be calculated and written at end of frame (default:0) */ - params->fParams.checksumFlag = value > 0; - return 0; - - case ZSTD_p_dictIDFlag : /* When applicable, dictionary's dictID is provided in frame header (default:1) */ - DEBUGLOG(5, "set dictIDFlag = %u", (value>0)); - params->fParams.noDictIDFlag = (value == 0); - return 0; + CCtxParams->fParams.checksumFlag = value != 0; + return CCtxParams->fParams.checksumFlag; + + case ZSTD_c_dictIDFlag : /* When applicable, dictionary's dictID is provided in frame header (default:1) */ + DEBUGLOG(4, "set dictIDFlag = %u", (value!=0)); + CCtxParams->fParams.noDictIDFlag = !value; + return !CCtxParams->fParams.noDictIDFlag; + + case ZSTD_c_forceMaxWindow : + CCtxParams->forceWindow = (value != 0); + return CCtxParams->forceWindow; + + case ZSTD_c_forceAttachDict : { + const ZSTD_dictAttachPref_e pref = (ZSTD_dictAttachPref_e)value; + BOUNDCHECK(ZSTD_c_forceAttachDict, pref); + CCtxParams->attachDictPref = pref; + return CCtxParams->attachDictPref; + } - case ZSTD_p_forceMaxWindow : - params->forceWindow = value > 0; - return 0; + case ZSTD_c_literalCompressionMode : { + const ZSTD_literalCompressionMode_e lcm = (ZSTD_literalCompressionMode_e)value; + BOUNDCHECK(ZSTD_c_literalCompressionMode, lcm); + CCtxParams->literalCompressionMode = lcm; + return CCtxParams->literalCompressionMode; + } - case ZSTD_p_nbThreads : - if (value == 0) return 0; + case ZSTD_c_nbWorkers : #ifndef ZSTD_MULTITHREAD - if (value > 1) return ERROR(parameter_unsupported); + RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading"); return 0; #else - return ZSTDMT_initializeCCtxParameters(params, value); + FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value), ""); + CCtxParams->nbWorkers = value; + return CCtxParams->nbWorkers; #endif - case ZSTD_p_jobSize : + case ZSTD_c_jobSize : #ifndef ZSTD_MULTITHREAD - return ERROR(parameter_unsupported); + RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading"); + return 0; #else - if (params->nbThreads <= 1) return ERROR(parameter_unsupported); - return ZSTDMT_CCtxParam_setMTCtxParameter(params, ZSTDMT_p_sectionSize, value); + /* Adjust to the minimum non-default value. */ + if (value != 0 && value < ZSTDMT_JOBSIZE_MIN) + value = ZSTDMT_JOBSIZE_MIN; + FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value), ""); + assert(value >= 0); + CCtxParams->jobSize = value; + return CCtxParams->jobSize; #endif - case ZSTD_p_overlapSizeLog : + case ZSTD_c_overlapLog : #ifndef ZSTD_MULTITHREAD - return ERROR(parameter_unsupported); + RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading"); + return 0; #else - if (params->nbThreads <= 1) return ERROR(parameter_unsupported); - return ZSTDMT_CCtxParam_setMTCtxParameter(params, ZSTDMT_p_overlapSectionLog, value); + FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value), ""); + CCtxParams->overlapLog = value; + return CCtxParams->overlapLog; #endif - case ZSTD_p_enableLongDistanceMatching : - if (value != 0) { - ZSTD_cLevelToCCtxParams(params); - params->cParams.windowLog = ZSTD_LDM_WINDOW_LOG; - } - return ZSTD_ldm_initializeParameters(¶ms->ldmParams, value); - - case ZSTD_p_ldmHashLog : - if (value == 0) return 0; - CLAMPCHECK(value, ZSTD_HASHLOG_MIN, ZSTD_HASHLOG_MAX); - params->ldmParams.hashLog = value; - return 0; - - case ZSTD_p_ldmMinMatch : - if (value == 0) return 0; - CLAMPCHECK(value, ZSTD_LDM_MINMATCH_MIN, ZSTD_LDM_MINMATCH_MAX); - params->ldmParams.minMatchLength = value; + case ZSTD_c_rsyncable : +#ifndef ZSTD_MULTITHREAD + RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading"); return 0; +#else + FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value), ""); + CCtxParams->rsyncable = value; + return CCtxParams->rsyncable; +#endif - case ZSTD_p_ldmBucketSizeLog : - if (value > ZSTD_LDM_BUCKETSIZELOG_MAX) { - return ERROR(parameter_outOfBound); - } - params->ldmParams.bucketSizeLog = value; - return 0; + case ZSTD_c_enableLongDistanceMatching : + CCtxParams->ldmParams.enableLdm = (value!=0); + return CCtxParams->ldmParams.enableLdm; + + case ZSTD_c_ldmHashLog : + if (value!=0) /* 0 ==> auto */ + BOUNDCHECK(ZSTD_c_ldmHashLog, value); + CCtxParams->ldmParams.hashLog = value; + return CCtxParams->ldmParams.hashLog; + + case ZSTD_c_ldmMinMatch : + if (value!=0) /* 0 ==> default */ + BOUNDCHECK(ZSTD_c_ldmMinMatch, value); + CCtxParams->ldmParams.minMatchLength = value; + return CCtxParams->ldmParams.minMatchLength; + + case ZSTD_c_ldmBucketSizeLog : + if (value!=0) /* 0 ==> default */ + BOUNDCHECK(ZSTD_c_ldmBucketSizeLog, value); + CCtxParams->ldmParams.bucketSizeLog = value; + return CCtxParams->ldmParams.bucketSizeLog; + + case ZSTD_c_ldmHashRateLog : + RETURN_ERROR_IF(value > ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN, + parameter_outOfBound, "Param out of bounds!"); + CCtxParams->ldmParams.hashRateLog = value; + return CCtxParams->ldmParams.hashRateLog; + + case ZSTD_c_targetCBlockSize : + if (value!=0) /* 0 ==> default */ + BOUNDCHECK(ZSTD_c_targetCBlockSize, value); + CCtxParams->targetCBlockSize = value; + return CCtxParams->targetCBlockSize; + + case ZSTD_c_srcSizeHint : + if (value!=0) /* 0 ==> default */ + BOUNDCHECK(ZSTD_c_srcSizeHint, value); + CCtxParams->srcSizeHint = value; + return CCtxParams->srcSizeHint; + + default: RETURN_ERROR(parameter_unsupported, "unknown parameter"); + } +} - case ZSTD_p_ldmHashEveryLog : - if (value > ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN) { - return ERROR(parameter_outOfBound); - } - params->ldmParams.hashEveryLog = value; - return 0; +size_t ZSTD_CCtx_getParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value) +{ + return ZSTD_CCtxParams_getParameter(&cctx->requestedParams, param, value); +} - default: return ERROR(parameter_unsupported); +size_t ZSTD_CCtxParams_getParameter( + ZSTD_CCtx_params* CCtxParams, ZSTD_cParameter param, int* value) +{ + switch(param) + { + case ZSTD_c_format : + *value = CCtxParams->format; + break; + case ZSTD_c_compressionLevel : + *value = CCtxParams->compressionLevel; + break; + case ZSTD_c_windowLog : + *value = (int)CCtxParams->cParams.windowLog; + break; + case ZSTD_c_hashLog : + *value = (int)CCtxParams->cParams.hashLog; + break; + case ZSTD_c_chainLog : + *value = (int)CCtxParams->cParams.chainLog; + break; + case ZSTD_c_searchLog : + *value = CCtxParams->cParams.searchLog; + break; + case ZSTD_c_minMatch : + *value = CCtxParams->cParams.minMatch; + break; + case ZSTD_c_targetLength : + *value = CCtxParams->cParams.targetLength; + break; + case ZSTD_c_strategy : + *value = (unsigned)CCtxParams->cParams.strategy; + break; + case ZSTD_c_contentSizeFlag : + *value = CCtxParams->fParams.contentSizeFlag; + break; + case ZSTD_c_checksumFlag : + *value = CCtxParams->fParams.checksumFlag; + break; + case ZSTD_c_dictIDFlag : + *value = !CCtxParams->fParams.noDictIDFlag; + break; + case ZSTD_c_forceMaxWindow : + *value = CCtxParams->forceWindow; + break; + case ZSTD_c_forceAttachDict : + *value = CCtxParams->attachDictPref; + break; + case ZSTD_c_literalCompressionMode : + *value = CCtxParams->literalCompressionMode; + break; + case ZSTD_c_nbWorkers : +#ifndef ZSTD_MULTITHREAD + assert(CCtxParams->nbWorkers == 0); +#endif + *value = CCtxParams->nbWorkers; + break; + case ZSTD_c_jobSize : +#ifndef ZSTD_MULTITHREAD + RETURN_ERROR(parameter_unsupported, "not compiled with multithreading"); +#else + assert(CCtxParams->jobSize <= INT_MAX); + *value = (int)CCtxParams->jobSize; + break; +#endif + case ZSTD_c_overlapLog : +#ifndef ZSTD_MULTITHREAD + RETURN_ERROR(parameter_unsupported, "not compiled with multithreading"); +#else + *value = CCtxParams->overlapLog; + break; +#endif + case ZSTD_c_rsyncable : +#ifndef ZSTD_MULTITHREAD + RETURN_ERROR(parameter_unsupported, "not compiled with multithreading"); +#else + *value = CCtxParams->rsyncable; + break; +#endif + case ZSTD_c_enableLongDistanceMatching : + *value = CCtxParams->ldmParams.enableLdm; + break; + case ZSTD_c_ldmHashLog : + *value = CCtxParams->ldmParams.hashLog; + break; + case ZSTD_c_ldmMinMatch : + *value = CCtxParams->ldmParams.minMatchLength; + break; + case ZSTD_c_ldmBucketSizeLog : + *value = CCtxParams->ldmParams.bucketSizeLog; + break; + case ZSTD_c_ldmHashRateLog : + *value = CCtxParams->ldmParams.hashRateLog; + break; + case ZSTD_c_targetCBlockSize : + *value = (int)CCtxParams->targetCBlockSize; + break; + case ZSTD_c_srcSizeHint : + *value = (int)CCtxParams->srcSizeHint; + break; + default: RETURN_ERROR(parameter_unsupported, "unknown parameter"); } + return 0; } -/** - * This function should be updated whenever ZSTD_CCtx_params is updated. - * Parameters are copied manually before the dictionary is loaded. - * The multithreading parameters jobSize and overlapSizeLog are set only if - * nbThreads > 1. - * - * Pledged srcSize is treated as unknown. +/** ZSTD_CCtx_setParametersUsingCCtxParams() : + * just applies `params` into `cctx` + * no action is performed, parameters are merely stored. + * If ZSTDMT is enabled, parameters are pushed to cctx->mtctx. + * This is possible even if a compression is ongoing. + * In which case, new parameters will be applied on the fly, starting with next compression job. */ size_t ZSTD_CCtx_setParametersUsingCCtxParams( ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params) { - if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); - if (cctx->cdict) return ERROR(stage_wrong); - - /* Assume the compression and frame parameters are validated */ - cctx->requestedParams.cParams = params->cParams; - cctx->requestedParams.fParams = params->fParams; - cctx->requestedParams.compressionLevel = params->compressionLevel; - - /* Set force window explicitly since it sets cctx->loadedDictEnd */ - CHECK_F( ZSTD_CCtx_setParameter( - cctx, ZSTD_p_forceMaxWindow, params->forceWindow) ); - - /* Set multithreading parameters explicitly */ - CHECK_F( ZSTD_CCtx_setParameter(cctx, ZSTD_p_nbThreads, params->nbThreads) ); - if (params->nbThreads > 1) { - CHECK_F( ZSTD_CCtx_setParameter(cctx, ZSTD_p_jobSize, params->jobSize) ); - CHECK_F( ZSTD_CCtx_setParameter( - cctx, ZSTD_p_overlapSizeLog, params->overlapSizeLog) ); - } - - /* Copy long distance matching parameters */ - cctx->requestedParams.ldmParams = params->ldmParams; - - /* customMem is used only for create/free params and can be ignored */ + DEBUGLOG(4, "ZSTD_CCtx_setParametersUsingCCtxParams"); + RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong, + "The context is in the wrong stage!"); + RETURN_ERROR_IF(cctx->cdict, stage_wrong, + "Can't override parameters with cdict attached (some must " + "be inherited from the cdict)."); + + cctx->requestedParams = *params; return 0; } ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize) { - DEBUGLOG(5, " setting pledgedSrcSize to %u", (U32)pledgedSrcSize); - if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); + DEBUGLOG(4, "ZSTD_CCtx_setPledgedSrcSize to %u bytes", (U32)pledgedSrcSize); + RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong, + "Can't set pledgedSrcSize when not in init stage."); cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1; return 0; } +/** + * Initializes the local dict using the requested parameters. + * NOTE: This does not use the pledged src size, because it may be used for more + * than one compression. + */ +static size_t ZSTD_initLocalDict(ZSTD_CCtx* cctx) +{ + ZSTD_localDict* const dl = &cctx->localDict; + ZSTD_compressionParameters const cParams = ZSTD_getCParamsFromCCtxParams( + &cctx->requestedParams, ZSTD_CONTENTSIZE_UNKNOWN, dl->dictSize); + if (dl->dict == NULL) { + /* No local dictionary. */ + assert(dl->dictBuffer == NULL); + assert(dl->cdict == NULL); + assert(dl->dictSize == 0); + return 0; + } + if (dl->cdict != NULL) { + assert(cctx->cdict == dl->cdict); + /* Local dictionary already initialized. */ + return 0; + } + assert(dl->dictSize > 0); + assert(cctx->cdict == NULL); + assert(cctx->prefixDict.dict == NULL); + + dl->cdict = ZSTD_createCDict_advanced( + dl->dict, + dl->dictSize, + ZSTD_dlm_byRef, + dl->dictContentType, + cParams, + cctx->customMem); + RETURN_ERROR_IF(!dl->cdict, memory_allocation, "ZSTD_createCDict_advanced failed"); + cctx->cdict = dl->cdict; + return 0; +} + size_t ZSTD_CCtx_loadDictionary_advanced( ZSTD_CCtx* cctx, const void* dict, size_t dictSize, - ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictMode_e dictMode) -{ - if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); - if (cctx->staticSize) return ERROR(memory_allocation); /* no malloc for static CCtx */ - DEBUGLOG(5, "load dictionary of size %u", (U32)dictSize); - ZSTD_freeCDict(cctx->cdictLocal); /* in case one already exists */ - if (dict==NULL || dictSize==0) { /* no dictionary mode */ - cctx->cdictLocal = NULL; - cctx->cdict = NULL; + ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType) +{ + RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong, + "Can't load a dictionary when ctx is not in init stage."); + RETURN_ERROR_IF(cctx->staticSize, memory_allocation, + "no malloc for static CCtx"); + DEBUGLOG(4, "ZSTD_CCtx_loadDictionary_advanced (size: %u)", (U32)dictSize); + ZSTD_clearAllDicts(cctx); /* in case one already exists */ + if (dict == NULL || dictSize == 0) /* no dictionary mode */ + return 0; + if (dictLoadMethod == ZSTD_dlm_byRef) { + cctx->localDict.dict = dict; } else { - ZSTD_compressionParameters const cParams = - ZSTD_getCParamsFromCCtxParams(cctx->requestedParams, 0, dictSize); - cctx->cdictLocal = ZSTD_createCDict_advanced( - dict, dictSize, - dictLoadMethod, dictMode, - cParams, cctx->customMem); - cctx->cdict = cctx->cdictLocal; - if (cctx->cdictLocal == NULL) - return ERROR(memory_allocation); + void* dictBuffer = ZSTD_malloc(dictSize, cctx->customMem); + RETURN_ERROR_IF(!dictBuffer, memory_allocation, "NULL pointer!"); + memcpy(dictBuffer, dict, dictSize); + cctx->localDict.dictBuffer = dictBuffer; + cctx->localDict.dict = dictBuffer; } + cctx->localDict.dictSize = dictSize; + cctx->localDict.dictContentType = dictContentType; return 0; } @@ -538,432 +918,669 @@ ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference( ZSTD_CCtx* cctx, const void* dict, size_t dictSize) { return ZSTD_CCtx_loadDictionary_advanced( - cctx, dict, dictSize, ZSTD_dlm_byRef, ZSTD_dm_auto); + cctx, dict, dictSize, ZSTD_dlm_byRef, ZSTD_dct_auto); } ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize) { return ZSTD_CCtx_loadDictionary_advanced( - cctx, dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dm_auto); + cctx, dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto); } size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict) { - if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); + RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong, + "Can't ref a dict when ctx not in init stage."); + /* Free the existing local cdict (if any) to save memory. */ + ZSTD_clearAllDicts(cctx); cctx->cdict = cdict; - memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict)); /* exclusive */ return 0; } size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize) { - return ZSTD_CCtx_refPrefix_advanced(cctx, prefix, prefixSize, ZSTD_dm_rawContent); + return ZSTD_CCtx_refPrefix_advanced(cctx, prefix, prefixSize, ZSTD_dct_rawContent); } size_t ZSTD_CCtx_refPrefix_advanced( - ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictMode_e dictMode) + ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType) { - if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); - cctx->cdict = NULL; /* prefix discards any prior cdict */ - cctx->prefixDict.dict = prefix; - cctx->prefixDict.dictSize = prefixSize; - cctx->prefixDict.dictMode = dictMode; + RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong, + "Can't ref a prefix when ctx not in init stage."); + ZSTD_clearAllDicts(cctx); + if (prefix != NULL && prefixSize > 0) { + cctx->prefixDict.dict = prefix; + cctx->prefixDict.dictSize = prefixSize; + cctx->prefixDict.dictContentType = dictContentType; + } return 0; } -static void ZSTD_startNewCompression(ZSTD_CCtx* cctx) -{ - cctx->streamStage = zcss_init; - cctx->pledgedSrcSizePlusOne = 0; -} - /*! ZSTD_CCtx_reset() : * Also dumps dictionary */ -void ZSTD_CCtx_reset(ZSTD_CCtx* cctx) +size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset) { - ZSTD_startNewCompression(cctx); - cctx->cdict = NULL; + if ( (reset == ZSTD_reset_session_only) + || (reset == ZSTD_reset_session_and_parameters) ) { + cctx->streamStage = zcss_init; + cctx->pledgedSrcSizePlusOne = 0; + } + if ( (reset == ZSTD_reset_parameters) + || (reset == ZSTD_reset_session_and_parameters) ) { + RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong, + "Can't reset parameters only when not in init stage."); + ZSTD_clearAllDicts(cctx); + return ZSTD_CCtxParams_reset(&cctx->requestedParams); + } + return 0; } + /** ZSTD_checkCParams() : control CParam values remain within authorized range. @return : 0, or an error code if one value is beyond authorized range */ size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams) { - CLAMPCHECK(cParams.windowLog, ZSTD_WINDOWLOG_MIN, ZSTD_WINDOWLOG_MAX); - CLAMPCHECK(cParams.chainLog, ZSTD_CHAINLOG_MIN, ZSTD_CHAINLOG_MAX); - CLAMPCHECK(cParams.hashLog, ZSTD_HASHLOG_MIN, ZSTD_HASHLOG_MAX); - CLAMPCHECK(cParams.searchLog, ZSTD_SEARCHLOG_MIN, ZSTD_SEARCHLOG_MAX); - CLAMPCHECK(cParams.searchLength, ZSTD_SEARCHLENGTH_MIN, ZSTD_SEARCHLENGTH_MAX); - CLAMPCHECK(cParams.targetLength, ZSTD_TARGETLENGTH_MIN, ZSTD_TARGETLENGTH_MAX); - if ((U32)(cParams.strategy) > (U32)ZSTD_btultra) - return ERROR(parameter_unsupported); + BOUNDCHECK(ZSTD_c_windowLog, (int)cParams.windowLog); + BOUNDCHECK(ZSTD_c_chainLog, (int)cParams.chainLog); + BOUNDCHECK(ZSTD_c_hashLog, (int)cParams.hashLog); + BOUNDCHECK(ZSTD_c_searchLog, (int)cParams.searchLog); + BOUNDCHECK(ZSTD_c_minMatch, (int)cParams.minMatch); + BOUNDCHECK(ZSTD_c_targetLength,(int)cParams.targetLength); + BOUNDCHECK(ZSTD_c_strategy, cParams.strategy); return 0; } /** ZSTD_clampCParams() : * make CParam values within valid range. * @return : valid CParams */ -static ZSTD_compressionParameters ZSTD_clampCParams(ZSTD_compressionParameters cParams) -{ -# define CLAMP(val,min,max) { \ - if (val<min) val=min; \ - else if (val>max) val=max; \ - } - CLAMP(cParams.windowLog, ZSTD_WINDOWLOG_MIN, ZSTD_WINDOWLOG_MAX); - CLAMP(cParams.chainLog, ZSTD_CHAINLOG_MIN, ZSTD_CHAINLOG_MAX); - CLAMP(cParams.hashLog, ZSTD_HASHLOG_MIN, ZSTD_HASHLOG_MAX); - CLAMP(cParams.searchLog, ZSTD_SEARCHLOG_MIN, ZSTD_SEARCHLOG_MAX); - CLAMP(cParams.searchLength, ZSTD_SEARCHLENGTH_MIN, ZSTD_SEARCHLENGTH_MAX); - CLAMP(cParams.targetLength, ZSTD_TARGETLENGTH_MIN, ZSTD_TARGETLENGTH_MAX); - if ((U32)(cParams.strategy) > (U32)ZSTD_btultra) cParams.strategy = ZSTD_btultra; +static ZSTD_compressionParameters +ZSTD_clampCParams(ZSTD_compressionParameters cParams) +{ +# define CLAMP_TYPE(cParam, val, type) { \ + ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); \ + if ((int)val<bounds.lowerBound) val=(type)bounds.lowerBound; \ + else if ((int)val>bounds.upperBound) val=(type)bounds.upperBound; \ + } +# define CLAMP(cParam, val) CLAMP_TYPE(cParam, val, unsigned) + CLAMP(ZSTD_c_windowLog, cParams.windowLog); + CLAMP(ZSTD_c_chainLog, cParams.chainLog); + CLAMP(ZSTD_c_hashLog, cParams.hashLog); + CLAMP(ZSTD_c_searchLog, cParams.searchLog); + CLAMP(ZSTD_c_minMatch, cParams.minMatch); + CLAMP(ZSTD_c_targetLength,cParams.targetLength); + CLAMP_TYPE(ZSTD_c_strategy,cParams.strategy, ZSTD_strategy); return cParams; } /** ZSTD_cycleLog() : * condition for correct operation : hashLog > 1 */ -static U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat) +U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat) { U32 const btScale = ((U32)strat >= (U32)ZSTD_btlazy2); return hashLog - btScale; } /** ZSTD_adjustCParams_internal() : - optimize `cPar` for a given input (`srcSize` and `dictSize`). - mostly downsizing to reduce memory consumption and initialization. - Both `srcSize` and `dictSize` are optional (use 0 if unknown), - but if both are 0, no optimization can be done. - Note : cPar is considered validated at this stage. Use ZSTD_checkParams() to ensure that. */ -ZSTD_compressionParameters ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize) + * optimize `cPar` for a specified input (`srcSize` and `dictSize`). + * mostly downsize to reduce memory consumption and initialization latency. + * `srcSize` can be ZSTD_CONTENTSIZE_UNKNOWN when not known. + * note : `srcSize==0` means 0! + * condition : cPar is presumed validated (can be checked using ZSTD_checkCParams()). */ +static ZSTD_compressionParameters +ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar, + unsigned long long srcSize, + size_t dictSize) { + static const U64 minSrcSize = 513; /* (1<<9) + 1 */ + static const U64 maxWindowResize = 1ULL << (ZSTD_WINDOWLOG_MAX-1); assert(ZSTD_checkCParams(cPar)==0); - if (srcSize+dictSize == 0) return cPar; /* no size information available : no adjustment */ - - /* resize params, to use less memory when necessary */ - { U32 const minSrcSize = (srcSize==0) ? 500 : 0; - U64 const rSize = srcSize + dictSize + minSrcSize; - if (rSize < ((U64)1<<ZSTD_WINDOWLOG_MAX)) { - U32 const srcLog = - MAX(ZSTD_HASHLOG_MIN, (rSize==1) ? 1 : ZSTD_highbit32((U32)(rSize)-1) + 1); - if (cPar.windowLog > srcLog) cPar.windowLog = srcLog; - } } - if (cPar.hashLog > cPar.windowLog) cPar.hashLog = cPar.windowLog; + + if (dictSize && srcSize == ZSTD_CONTENTSIZE_UNKNOWN) + srcSize = minSrcSize; + + /* resize windowLog if input is small enough, to use less memory */ + if ( (srcSize < maxWindowResize) + && (dictSize < maxWindowResize) ) { + U32 const tSize = (U32)(srcSize + dictSize); + static U32 const hashSizeMin = 1 << ZSTD_HASHLOG_MIN; + U32 const srcLog = (tSize < hashSizeMin) ? ZSTD_HASHLOG_MIN : + ZSTD_highbit32(tSize-1) + 1; + if (cPar.windowLog > srcLog) cPar.windowLog = srcLog; + } + if (cPar.hashLog > cPar.windowLog+1) cPar.hashLog = cPar.windowLog+1; { U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy); - if (cycleLog > cPar.windowLog) cPar.chainLog -= (cycleLog - cPar.windowLog); + if (cycleLog > cPar.windowLog) + cPar.chainLog -= (cycleLog - cPar.windowLog); } - if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN) cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN; /* required for frame header */ + if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN) + cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN; /* minimum wlog required for valid frame header */ return cPar; } -ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize) +ZSTD_compressionParameters +ZSTD_adjustCParams(ZSTD_compressionParameters cPar, + unsigned long long srcSize, + size_t dictSize) { - cPar = ZSTD_clampCParams(cPar); + cPar = ZSTD_clampCParams(cPar); /* resulting cPar is necessarily valid (all parameters within range) */ + if (srcSize == 0) srcSize = ZSTD_CONTENTSIZE_UNKNOWN; return ZSTD_adjustCParams_internal(cPar, srcSize, dictSize); } -size_t ZSTD_estimateCCtxSize_advanced_usingCCtxParams(const ZSTD_CCtx_params* params) +static ZSTD_compressionParameters ZSTD_getCParams_internal(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize); +static ZSTD_parameters ZSTD_getParams_internal(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize); + +ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams( + const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize) +{ + ZSTD_compressionParameters cParams; + if (srcSizeHint == ZSTD_CONTENTSIZE_UNKNOWN && CCtxParams->srcSizeHint > 0) { + srcSizeHint = CCtxParams->srcSizeHint; + } + cParams = ZSTD_getCParams_internal(CCtxParams->compressionLevel, srcSizeHint, dictSize); + if (CCtxParams->ldmParams.enableLdm) cParams.windowLog = ZSTD_LDM_DEFAULT_WINDOW_LOG; + if (CCtxParams->cParams.windowLog) cParams.windowLog = CCtxParams->cParams.windowLog; + if (CCtxParams->cParams.hashLog) cParams.hashLog = CCtxParams->cParams.hashLog; + if (CCtxParams->cParams.chainLog) cParams.chainLog = CCtxParams->cParams.chainLog; + if (CCtxParams->cParams.searchLog) cParams.searchLog = CCtxParams->cParams.searchLog; + if (CCtxParams->cParams.minMatch) cParams.minMatch = CCtxParams->cParams.minMatch; + if (CCtxParams->cParams.targetLength) cParams.targetLength = CCtxParams->cParams.targetLength; + if (CCtxParams->cParams.strategy) cParams.strategy = CCtxParams->cParams.strategy; + assert(!ZSTD_checkCParams(cParams)); + /* srcSizeHint == 0 means 0 */ + return ZSTD_adjustCParams_internal(cParams, srcSizeHint, dictSize); +} + +static size_t +ZSTD_sizeof_matchState(const ZSTD_compressionParameters* const cParams, + const U32 forCCtx) { - /* Estimate CCtx size is supported for single-threaded compression only. */ - if (params->nbThreads > 1) { return ERROR(GENERIC); } + size_t const chainSize = (cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams->chainLog); + size_t const hSize = ((size_t)1) << cParams->hashLog; + U32 const hashLog3 = (forCCtx && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0; + size_t const h3Size = hashLog3 ? ((size_t)1) << hashLog3 : 0; + /* We don't use ZSTD_cwksp_alloc_size() here because the tables aren't + * surrounded by redzones in ASAN. */ + size_t const tableSpace = chainSize * sizeof(U32) + + hSize * sizeof(U32) + + h3Size * sizeof(U32); + size_t const optPotentialSpace = + ZSTD_cwksp_alloc_size((MaxML+1) * sizeof(U32)) + + ZSTD_cwksp_alloc_size((MaxLL+1) * sizeof(U32)) + + ZSTD_cwksp_alloc_size((MaxOff+1) * sizeof(U32)) + + ZSTD_cwksp_alloc_size((1<<Litbits) * sizeof(U32)) + + ZSTD_cwksp_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_match_t)) + + ZSTD_cwksp_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t)); + size_t const optSpace = (forCCtx && (cParams->strategy >= ZSTD_btopt)) + ? optPotentialSpace + : 0; + DEBUGLOG(4, "chainSize: %u - hSize: %u - h3Size: %u", + (U32)chainSize, (U32)hSize, (U32)h3Size); + return tableSpace + optSpace; +} + +size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params) +{ + RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only."); { ZSTD_compressionParameters const cParams = - ZSTD_getCParamsFromCCtxParams(*params, 0, 0); + ZSTD_getCParamsFromCCtxParams(params, ZSTD_CONTENTSIZE_UNKNOWN, 0); size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog); - U32 const divider = (cParams.searchLength==3) ? 3 : 4; + U32 const divider = (cParams.minMatch==3) ? 3 : 4; size_t const maxNbSeq = blockSize / divider; - size_t const tokenSpace = blockSize + 11*maxNbSeq; - size_t const chainSize = - (cParams.strategy == ZSTD_fast) ? 0 : (1 << cParams.chainLog); - size_t const hSize = ((size_t)1) << cParams.hashLog; - U32 const hashLog3 = (cParams.searchLength>3) ? - 0 : MIN(ZSTD_HASHLOG3_MAX, cParams.windowLog); - size_t const h3Size = ((size_t)1) << hashLog3; - size_t const entropySpace = sizeof(ZSTD_entropyCTables_t); - size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32); - - size_t const optBudget = - ((MaxML+1) + (MaxLL+1) + (MaxOff+1) + (1<<Litbits))*sizeof(U32) - + (ZSTD_OPT_NUM+1)*(sizeof(ZSTD_match_t) + sizeof(ZSTD_optimal_t)); - size_t const optSpace = ((cParams.strategy == ZSTD_btopt) || (cParams.strategy == ZSTD_btultra)) ? optBudget : 0; - - size_t const ldmSpace = params->ldmParams.enableLdm ? - ZSTD_ldm_getTableSize(params->ldmParams.hashLog, - params->ldmParams.bucketSizeLog) : 0; - - size_t const neededSpace = entropySpace + tableSpace + tokenSpace + - optSpace + ldmSpace; - - DEBUGLOG(5, "sizeof(ZSTD_CCtx) : %u", (U32)sizeof(ZSTD_CCtx)); - DEBUGLOG(5, "estimate workSpace : %u", (U32)neededSpace); - return sizeof(ZSTD_CCtx) + neededSpace; + size_t const tokenSpace = ZSTD_cwksp_alloc_size(WILDCOPY_OVERLENGTH + blockSize) + + ZSTD_cwksp_alloc_size(maxNbSeq * sizeof(seqDef)) + + 3 * ZSTD_cwksp_alloc_size(maxNbSeq * sizeof(BYTE)); + size_t const entropySpace = ZSTD_cwksp_alloc_size(HUF_WORKSPACE_SIZE); + size_t const blockStateSpace = 2 * ZSTD_cwksp_alloc_size(sizeof(ZSTD_compressedBlockState_t)); + size_t const matchStateSize = ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 1); + + size_t const ldmSpace = ZSTD_ldm_getTableSize(params->ldmParams); + size_t const ldmSeqSpace = ZSTD_cwksp_alloc_size(ZSTD_ldm_getMaxNbSeq(params->ldmParams, blockSize) * sizeof(rawSeq)); + + /* estimateCCtxSize is for one-shot compression. So no buffers should + * be needed. However, we still allocate two 0-sized buffers, which can + * take space under ASAN. */ + size_t const bufferSpace = ZSTD_cwksp_alloc_size(0) + + ZSTD_cwksp_alloc_size(0); + + size_t const cctxSpace = ZSTD_cwksp_alloc_size(sizeof(ZSTD_CCtx)); + + size_t const neededSpace = + cctxSpace + + entropySpace + + blockStateSpace + + ldmSpace + + ldmSeqSpace + + matchStateSize + + tokenSpace + + bufferSpace; + + DEBUGLOG(5, "estimate workspace : %u", (U32)neededSpace); + return neededSpace; } } -size_t ZSTD_estimateCCtxSize_advanced_usingCParams(ZSTD_compressionParameters cParams) +size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams) { ZSTD_CCtx_params const params = ZSTD_makeCCtxParamsFromCParams(cParams); - return ZSTD_estimateCCtxSize_advanced_usingCCtxParams(¶ms); + return ZSTD_estimateCCtxSize_usingCCtxParams(¶ms); +} + +static size_t ZSTD_estimateCCtxSize_internal(int compressionLevel) +{ + ZSTD_compressionParameters const cParams = ZSTD_getCParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, 0); + return ZSTD_estimateCCtxSize_usingCParams(cParams); } size_t ZSTD_estimateCCtxSize(int compressionLevel) { - ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, 0, 0); - return ZSTD_estimateCCtxSize_advanced_usingCParams(cParams); + int level; + size_t memBudget = 0; + for (level=MIN(compressionLevel, 1); level<=compressionLevel; level++) { + size_t const newMB = ZSTD_estimateCCtxSize_internal(level); + if (newMB > memBudget) memBudget = newMB; + } + return memBudget; } -size_t ZSTD_estimateCStreamSize_advanced_usingCCtxParams(const ZSTD_CCtx_params* params) +size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params) { - if (params->nbThreads > 1) { return ERROR(GENERIC); } - { size_t const CCtxSize = ZSTD_estimateCCtxSize_advanced_usingCCtxParams(params); - size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << params->cParams.windowLog); - size_t const inBuffSize = ((size_t)1 << params->cParams.windowLog) + blockSize; + RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only."); + { ZSTD_compressionParameters const cParams = + ZSTD_getCParamsFromCCtxParams(params, ZSTD_CONTENTSIZE_UNKNOWN, 0); + size_t const CCtxSize = ZSTD_estimateCCtxSize_usingCCtxParams(params); + size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog); + size_t const inBuffSize = ((size_t)1 << cParams.windowLog) + blockSize; size_t const outBuffSize = ZSTD_compressBound(blockSize) + 1; - size_t const streamingSize = inBuffSize + outBuffSize; + size_t const streamingSize = ZSTD_cwksp_alloc_size(inBuffSize) + + ZSTD_cwksp_alloc_size(outBuffSize); return CCtxSize + streamingSize; } } -size_t ZSTD_estimateCStreamSize_advanced_usingCParams(ZSTD_compressionParameters cParams) +size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams) { ZSTD_CCtx_params const params = ZSTD_makeCCtxParamsFromCParams(cParams); - return ZSTD_estimateCStreamSize_advanced_usingCCtxParams(¶ms); + return ZSTD_estimateCStreamSize_usingCCtxParams(¶ms); } -size_t ZSTD_estimateCStreamSize(int compressionLevel) { - ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, 0, 0); - return ZSTD_estimateCStreamSize_advanced_usingCParams(cParams); +static size_t ZSTD_estimateCStreamSize_internal(int compressionLevel) +{ + ZSTD_compressionParameters const cParams = ZSTD_getCParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, 0); + return ZSTD_estimateCStreamSize_usingCParams(cParams); } -static U32 ZSTD_equivalentCParams(ZSTD_compressionParameters cParams1, - ZSTD_compressionParameters cParams2) +size_t ZSTD_estimateCStreamSize(int compressionLevel) { - U32 bslog1 = MIN(cParams1.windowLog, ZSTD_BLOCKSIZELOG_MAX); - U32 bslog2 = MIN(cParams2.windowLog, ZSTD_BLOCKSIZELOG_MAX); - return (bslog1 == bslog2) /* same block size */ - & (cParams1.hashLog == cParams2.hashLog) - & (cParams1.chainLog == cParams2.chainLog) - & (cParams1.strategy == cParams2.strategy) /* opt parser space */ - & ((cParams1.searchLength==3) == (cParams2.searchLength==3)); /* hashlog3 space */ + int level; + size_t memBudget = 0; + for (level=MIN(compressionLevel, 1); level<=compressionLevel; level++) { + size_t const newMB = ZSTD_estimateCStreamSize_internal(level); + if (newMB > memBudget) memBudget = newMB; + } + return memBudget; } -/** The parameters are equivalent if ldm is not enabled in both sets or - * all the parameters are equivalent. */ -static U32 ZSTD_equivalentLdmParams(ldmParams_t ldmParams1, - ldmParams_t ldmParams2) +/* ZSTD_getFrameProgression(): + * tells how much data has been consumed (input) and produced (output) for current frame. + * able to count progression inside worker threads (non-blocking mode). + */ +ZSTD_frameProgression ZSTD_getFrameProgression(const ZSTD_CCtx* cctx) +{ +#ifdef ZSTD_MULTITHREAD + if (cctx->appliedParams.nbWorkers > 0) { + return ZSTDMT_getFrameProgression(cctx->mtctx); + } +#endif + { ZSTD_frameProgression fp; + size_t const buffered = (cctx->inBuff == NULL) ? 0 : + cctx->inBuffPos - cctx->inToCompress; + if (buffered) assert(cctx->inBuffPos >= cctx->inToCompress); + assert(buffered <= ZSTD_BLOCKSIZE_MAX); + fp.ingested = cctx->consumedSrcSize + buffered; + fp.consumed = cctx->consumedSrcSize; + fp.produced = cctx->producedCSize; + fp.flushed = cctx->producedCSize; /* simplified; some data might still be left within streaming output buffer */ + fp.currentJobID = 0; + fp.nbActiveWorkers = 0; + return fp; +} } + +/*! ZSTD_toFlushNow() + * Only useful for multithreading scenarios currently (nbWorkers >= 1). + */ +size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx) { - return (!ldmParams1.enableLdm && !ldmParams2.enableLdm) || - (ldmParams1.enableLdm == ldmParams2.enableLdm && - ldmParams1.hashLog == ldmParams2.hashLog && - ldmParams1.bucketSizeLog == ldmParams2.bucketSizeLog && - ldmParams1.minMatchLength == ldmParams2.minMatchLength && - ldmParams1.hashEveryLog == ldmParams2.hashEveryLog); +#ifdef ZSTD_MULTITHREAD + if (cctx->appliedParams.nbWorkers > 0) { + return ZSTDMT_toFlushNow(cctx->mtctx); + } +#endif + (void)cctx; + return 0; /* over-simplification; could also check if context is currently running in streaming mode, and in which case, report how many bytes are left to be flushed within output buffer */ } -/** Equivalence for resetCCtx purposes */ -static U32 ZSTD_equivalentParams(ZSTD_CCtx_params params1, - ZSTD_CCtx_params params2) +static void ZSTD_assertEqualCParams(ZSTD_compressionParameters cParams1, + ZSTD_compressionParameters cParams2) +{ + (void)cParams1; + (void)cParams2; + assert(cParams1.windowLog == cParams2.windowLog); + assert(cParams1.chainLog == cParams2.chainLog); + assert(cParams1.hashLog == cParams2.hashLog); + assert(cParams1.searchLog == cParams2.searchLog); + assert(cParams1.minMatch == cParams2.minMatch); + assert(cParams1.targetLength == cParams2.targetLength); + assert(cParams1.strategy == cParams2.strategy); +} + +void ZSTD_reset_compressedBlockState(ZSTD_compressedBlockState_t* bs) { - return ZSTD_equivalentCParams(params1.cParams, params2.cParams) && - ZSTD_equivalentLdmParams(params1.ldmParams, params2.ldmParams); + int i; + for (i = 0; i < ZSTD_REP_NUM; ++i) + bs->rep[i] = repStartValue[i]; + bs->entropy.huf.repeatMode = HUF_repeat_none; + bs->entropy.fse.offcode_repeatMode = FSE_repeat_none; + bs->entropy.fse.matchlength_repeatMode = FSE_repeat_none; + bs->entropy.fse.litlength_repeatMode = FSE_repeat_none; } -/*! ZSTD_continueCCtx() : - * reuse CCtx without reset (note : requires no dictionary) */ -static size_t ZSTD_continueCCtx(ZSTD_CCtx* cctx, ZSTD_CCtx_params params, U64 pledgedSrcSize) +/*! ZSTD_invalidateMatchState() + * Invalidate all the matches in the match finder tables. + * Requires nextSrc and base to be set (can be NULL). + */ +static void ZSTD_invalidateMatchState(ZSTD_matchState_t* ms) { - U32 const end = (U32)(cctx->nextSrc - cctx->base); - DEBUGLOG(5, "continue mode"); - cctx->appliedParams = params; - cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1; - cctx->consumedSrcSize = 0; - if (pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN) - cctx->appliedParams.fParams.contentSizeFlag = 0; - DEBUGLOG(5, "pledged content size : %u ; flag : %u", - (U32)pledgedSrcSize, cctx->appliedParams.fParams.contentSizeFlag); - cctx->lowLimit = end; - cctx->dictLimit = end; - cctx->nextToUpdate = end+1; - cctx->stage = ZSTDcs_init; - cctx->dictID = 0; - cctx->loadedDictEnd = 0; - { int i; for (i=0; i<ZSTD_REP_NUM; i++) cctx->seqStore.rep[i] = repStartValue[i]; } - cctx->optState.litLengthSum = 0; /* force reset of btopt stats */ - XXH64_reset(&cctx->xxhState, 0); + ZSTD_window_clear(&ms->window); + + ms->nextToUpdate = ms->window.dictLimit; + ms->loadedDictEnd = 0; + ms->opt.litLengthSum = 0; /* force reset of btopt stats */ + ms->dictMatchState = NULL; +} + +/** + * Indicates whether this compression proceeds directly from user-provided + * source buffer to user-provided destination buffer (ZSTDb_not_buffered), or + * whether the context needs to buffer the input/output (ZSTDb_buffered). + */ +typedef enum { + ZSTDb_not_buffered, + ZSTDb_buffered +} ZSTD_buffered_policy_e; + +/** + * Controls, for this matchState reset, whether the tables need to be cleared / + * prepared for the coming compression (ZSTDcrp_makeClean), or whether the + * tables can be left unclean (ZSTDcrp_leaveDirty), because we know that a + * subsequent operation will overwrite the table space anyways (e.g., copying + * the matchState contents in from a CDict). + */ +typedef enum { + ZSTDcrp_makeClean, + ZSTDcrp_leaveDirty +} ZSTD_compResetPolicy_e; + +/** + * Controls, for this matchState reset, whether indexing can continue where it + * left off (ZSTDirp_continue), or whether it needs to be restarted from zero + * (ZSTDirp_reset). + */ +typedef enum { + ZSTDirp_continue, + ZSTDirp_reset +} ZSTD_indexResetPolicy_e; + +typedef enum { + ZSTD_resetTarget_CDict, + ZSTD_resetTarget_CCtx +} ZSTD_resetTarget_e; + +static size_t +ZSTD_reset_matchState(ZSTD_matchState_t* ms, + ZSTD_cwksp* ws, + const ZSTD_compressionParameters* cParams, + const ZSTD_compResetPolicy_e crp, + const ZSTD_indexResetPolicy_e forceResetIndex, + const ZSTD_resetTarget_e forWho) +{ + size_t const chainSize = (cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams->chainLog); + size_t const hSize = ((size_t)1) << cParams->hashLog; + U32 const hashLog3 = ((forWho == ZSTD_resetTarget_CCtx) && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0; + size_t const h3Size = hashLog3 ? ((size_t)1) << hashLog3 : 0; + + DEBUGLOG(4, "reset indices : %u", forceResetIndex == ZSTDirp_reset); + if (forceResetIndex == ZSTDirp_reset) { + ZSTD_window_init(&ms->window); + ZSTD_cwksp_mark_tables_dirty(ws); + } + + ms->hashLog3 = hashLog3; + + ZSTD_invalidateMatchState(ms); + + assert(!ZSTD_cwksp_reserve_failed(ws)); /* check that allocation hasn't already failed */ + + ZSTD_cwksp_clear_tables(ws); + + DEBUGLOG(5, "reserving table space"); + /* table Space */ + ms->hashTable = (U32*)ZSTD_cwksp_reserve_table(ws, hSize * sizeof(U32)); + ms->chainTable = (U32*)ZSTD_cwksp_reserve_table(ws, chainSize * sizeof(U32)); + ms->hashTable3 = (U32*)ZSTD_cwksp_reserve_table(ws, h3Size * sizeof(U32)); + RETURN_ERROR_IF(ZSTD_cwksp_reserve_failed(ws), memory_allocation, + "failed a workspace allocation in ZSTD_reset_matchState"); + + DEBUGLOG(4, "reset table : %u", crp!=ZSTDcrp_leaveDirty); + if (crp!=ZSTDcrp_leaveDirty) { + /* reset tables only */ + ZSTD_cwksp_clean_tables(ws); + } + + /* opt parser space */ + if ((forWho == ZSTD_resetTarget_CCtx) && (cParams->strategy >= ZSTD_btopt)) { + DEBUGLOG(4, "reserving optimal parser space"); + ms->opt.litFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (1<<Litbits) * sizeof(unsigned)); + ms->opt.litLengthFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxLL+1) * sizeof(unsigned)); + ms->opt.matchLengthFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxML+1) * sizeof(unsigned)); + ms->opt.offCodeFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxOff+1) * sizeof(unsigned)); + ms->opt.matchTable = (ZSTD_match_t*)ZSTD_cwksp_reserve_aligned(ws, (ZSTD_OPT_NUM+1) * sizeof(ZSTD_match_t)); + ms->opt.priceTable = (ZSTD_optimal_t*)ZSTD_cwksp_reserve_aligned(ws, (ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t)); + } + + ms->cParams = *cParams; + + RETURN_ERROR_IF(ZSTD_cwksp_reserve_failed(ws), memory_allocation, + "failed a workspace allocation in ZSTD_reset_matchState"); + return 0; } -typedef enum { ZSTDcrp_continue, ZSTDcrp_noMemset } ZSTD_compResetPolicy_e; -typedef enum { ZSTDb_not_buffered, ZSTDb_buffered } ZSTD_buffered_policy_e; +/* ZSTD_indexTooCloseToMax() : + * minor optimization : prefer memset() rather than reduceIndex() + * which is measurably slow in some circumstances (reported for Visual Studio). + * Works when re-using a context for a lot of smallish inputs : + * if all inputs are smaller than ZSTD_INDEXOVERFLOW_MARGIN, + * memset() will be triggered before reduceIndex(). + */ +#define ZSTD_INDEXOVERFLOW_MARGIN (16 MB) +static int ZSTD_indexTooCloseToMax(ZSTD_window_t w) +{ + return (size_t)(w.nextSrc - w.base) > (ZSTD_CURRENT_MAX - ZSTD_INDEXOVERFLOW_MARGIN); +} /*! ZSTD_resetCCtx_internal() : note : `params` are assumed fully validated at this stage */ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc, - ZSTD_CCtx_params params, U64 pledgedSrcSize, + ZSTD_CCtx_params params, + U64 const pledgedSrcSize, ZSTD_compResetPolicy_e const crp, ZSTD_buffered_policy_e const zbuff) { + ZSTD_cwksp* const ws = &zc->workspace; + DEBUGLOG(4, "ZSTD_resetCCtx_internal: pledgedSrcSize=%u, wlog=%u", + (U32)pledgedSrcSize, params.cParams.windowLog); assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams))); - if (crp == ZSTDcrp_continue) { - if (ZSTD_equivalentParams(params, zc->appliedParams)) { - DEBUGLOG(5, "ZSTD_equivalentParams()==1"); - assert(!(params.ldmParams.enableLdm && - params.ldmParams.hashEveryLog == ZSTD_LDM_HASHEVERYLOG_NOTSET)); - zc->entropy->hufCTable_repeatMode = HUF_repeat_none; - zc->entropy->offcode_repeatMode = FSE_repeat_none; - zc->entropy->matchlength_repeatMode = FSE_repeat_none; - zc->entropy->litlength_repeatMode = FSE_repeat_none; - return ZSTD_continueCCtx(zc, params, pledgedSrcSize); - } } + zc->isFirstBlock = 1; if (params.ldmParams.enableLdm) { /* Adjust long distance matching parameters */ - ZSTD_ldm_adjustParameters(¶ms.ldmParams, params.cParams.windowLog); + ZSTD_ldm_adjustParameters(¶ms.ldmParams, ¶ms.cParams); assert(params.ldmParams.hashLog >= params.ldmParams.bucketSizeLog); - assert(params.ldmParams.hashEveryLog < 32); - zc->ldmState.hashPower = - ZSTD_ldm_getHashPower(params.ldmParams.minMatchLength); + assert(params.ldmParams.hashRateLog < 32); + zc->ldmState.hashPower = ZSTD_rollingHash_primePower(params.ldmParams.minMatchLength); } - { size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << params.cParams.windowLog); - U32 const divider = (params.cParams.searchLength==3) ? 3 : 4; + { 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 tokenSpace = blockSize + 11*maxNbSeq; - size_t const chainSize = (params.cParams.strategy == ZSTD_fast) ? - 0 : (1 << params.cParams.chainLog); - size_t const hSize = ((size_t)1) << params.cParams.hashLog; - U32 const hashLog3 = (params.cParams.searchLength>3) ? - 0 : MIN(ZSTD_HASHLOG3_MAX, params.cParams.windowLog); - size_t const h3Size = ((size_t)1) << hashLog3; - size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32); + size_t const tokenSpace = ZSTD_cwksp_alloc_size(WILDCOPY_OVERLENGTH + blockSize) + + ZSTD_cwksp_alloc_size(maxNbSeq * sizeof(seqDef)) + + 3 * ZSTD_cwksp_alloc_size(maxNbSeq * sizeof(BYTE)); size_t const buffOutSize = (zbuff==ZSTDb_buffered) ? ZSTD_compressBound(blockSize)+1 : 0; - size_t const buffInSize = (zbuff==ZSTDb_buffered) ? ((size_t)1 << params.cParams.windowLog) + blockSize : 0; - void* ptr; - - /* Check if workSpace is large enough, alloc a new one if needed */ - { size_t const entropySpace = sizeof(ZSTD_entropyCTables_t); - size_t const optPotentialSpace = ((MaxML+1) + (MaxLL+1) + (MaxOff+1) + (1<<Litbits)) * sizeof(U32) - + (ZSTD_OPT_NUM+1) * (sizeof(ZSTD_match_t)+sizeof(ZSTD_optimal_t)); - size_t const optSpace = ( (params.cParams.strategy == ZSTD_btopt) - || (params.cParams.strategy == ZSTD_btultra)) ? - optPotentialSpace : 0; - size_t const bufferSpace = buffInSize + buffOutSize; - size_t const ldmSpace = params.ldmParams.enableLdm - ? ZSTD_ldm_getTableSize(params.ldmParams.hashLog, params.ldmParams.bucketSizeLog) - : 0; - size_t const neededSpace = entropySpace + optSpace + ldmSpace + - tableSpace + tokenSpace + bufferSpace; - - if (zc->workSpaceSize < neededSpace) { /* too small : resize */ - DEBUGLOG(5, "Need to update workSpaceSize from %uK to %uK \n", - (unsigned)zc->workSpaceSize>>10, - (unsigned)neededSpace>>10); - /* static cctx : no resize, error out */ - if (zc->staticSize) return ERROR(memory_allocation); - - zc->workSpaceSize = 0; - ZSTD_free(zc->workSpace, zc->customMem); - zc->workSpace = ZSTD_malloc(neededSpace, zc->customMem); - if (zc->workSpace == NULL) return ERROR(memory_allocation); - zc->workSpaceSize = neededSpace; - ptr = zc->workSpace; - - /* entropy space */ - assert(((size_t)zc->workSpace & 3) == 0); /* ensure correct alignment */ - assert(zc->workSpaceSize >= sizeof(ZSTD_entropyCTables_t)); - zc->entropy = (ZSTD_entropyCTables_t*)zc->workSpace; + size_t const buffInSize = (zbuff==ZSTDb_buffered) ? windowSize + blockSize : 0; + size_t const matchStateSize = ZSTD_sizeof_matchState(¶ms.cParams, /* forCCtx */ 1); + size_t const maxNbLdmSeq = ZSTD_ldm_getMaxNbSeq(params.ldmParams, blockSize); + + ZSTD_indexResetPolicy_e needsIndexReset = zc->initialized ? ZSTDirp_continue : ZSTDirp_reset; + + if (ZSTD_indexTooCloseToMax(zc->blockState.matchState.window)) { + needsIndexReset = ZSTDirp_reset; + } + + if (!zc->staticSize) ZSTD_cwksp_bump_oversized_duration(ws, 0); + + /* Check if workspace is large enough, alloc a new one if needed */ + { size_t const cctxSpace = zc->staticSize ? ZSTD_cwksp_alloc_size(sizeof(ZSTD_CCtx)) : 0; + size_t const entropySpace = ZSTD_cwksp_alloc_size(HUF_WORKSPACE_SIZE); + size_t const blockStateSpace = 2 * ZSTD_cwksp_alloc_size(sizeof(ZSTD_compressedBlockState_t)); + size_t const bufferSpace = ZSTD_cwksp_alloc_size(buffInSize) + ZSTD_cwksp_alloc_size(buffOutSize); + size_t const ldmSpace = ZSTD_ldm_getTableSize(params.ldmParams); + size_t const ldmSeqSpace = ZSTD_cwksp_alloc_size(maxNbLdmSeq * sizeof(rawSeq)); + + size_t const neededSpace = + cctxSpace + + entropySpace + + blockStateSpace + + ldmSpace + + ldmSeqSpace + + matchStateSize + + tokenSpace + + bufferSpace; + + int const workspaceTooSmall = ZSTD_cwksp_sizeof(ws) < neededSpace; + int const workspaceWasteful = ZSTD_cwksp_check_wasteful(ws, neededSpace); + + DEBUGLOG(4, "Need %zuKB workspace, including %zuKB for match state, and %zuKB for buffers", + neededSpace>>10, matchStateSize>>10, bufferSpace>>10); + DEBUGLOG(4, "windowSize: %zu - blockSize: %zu", windowSize, blockSize); + + if (workspaceTooSmall || workspaceWasteful) { + DEBUGLOG(4, "Resize workspaceSize from %zuKB to %zuKB", + ZSTD_cwksp_sizeof(ws) >> 10, + neededSpace >> 10); + + RETURN_ERROR_IF(zc->staticSize, memory_allocation, "static cctx : no resize"); + + needsIndexReset = ZSTDirp_reset; + + ZSTD_cwksp_free(ws, zc->customMem); + FORWARD_IF_ERROR(ZSTD_cwksp_create(ws, neededSpace, zc->customMem), ""); + + DEBUGLOG(5, "reserving object space"); + /* Statically sized space. + * entropyWorkspace never moves, + * though prev/next block swap places */ + assert(ZSTD_cwksp_check_available(ws, 2 * sizeof(ZSTD_compressedBlockState_t))); + zc->blockState.prevCBlock = (ZSTD_compressedBlockState_t*) ZSTD_cwksp_reserve_object(ws, sizeof(ZSTD_compressedBlockState_t)); + RETURN_ERROR_IF(zc->blockState.prevCBlock == NULL, memory_allocation, "couldn't allocate prevCBlock"); + zc->blockState.nextCBlock = (ZSTD_compressedBlockState_t*) ZSTD_cwksp_reserve_object(ws, sizeof(ZSTD_compressedBlockState_t)); + RETURN_ERROR_IF(zc->blockState.nextCBlock == NULL, memory_allocation, "couldn't allocate nextCBlock"); + zc->entropyWorkspace = (U32*) ZSTD_cwksp_reserve_object(ws, HUF_WORKSPACE_SIZE); + RETURN_ERROR_IF(zc->blockState.nextCBlock == NULL, memory_allocation, "couldn't allocate entropyWorkspace"); } } + ZSTD_cwksp_clear(ws); + /* init params */ zc->appliedParams = params; + zc->blockState.matchState.cParams = params.cParams; zc->pledgedSrcSizePlusOne = pledgedSrcSize+1; zc->consumedSrcSize = 0; + zc->producedCSize = 0; if (pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN) zc->appliedParams.fParams.contentSizeFlag = 0; - DEBUGLOG(5, "pledged content size : %u ; flag : %u", - (U32)pledgedSrcSize, zc->appliedParams.fParams.contentSizeFlag); + DEBUGLOG(4, "pledged content size : %u ; flag : %u", + (unsigned)pledgedSrcSize, zc->appliedParams.fParams.contentSizeFlag); zc->blockSize = blockSize; XXH64_reset(&zc->xxhState, 0); zc->stage = ZSTDcs_init; zc->dictID = 0; - zc->loadedDictEnd = 0; - zc->entropy->hufCTable_repeatMode = HUF_repeat_none; - zc->entropy->offcode_repeatMode = FSE_repeat_none; - zc->entropy->matchlength_repeatMode = FSE_repeat_none; - zc->entropy->litlength_repeatMode = FSE_repeat_none; - zc->nextToUpdate = 1; - zc->nextSrc = NULL; - zc->base = NULL; - zc->dictBase = NULL; - zc->dictLimit = 0; - zc->lowLimit = 0; - { int i; for (i=0; i<ZSTD_REP_NUM; i++) zc->seqStore.rep[i] = repStartValue[i]; } - zc->hashLog3 = hashLog3; - zc->optState.litLengthSum = 0; - - ptr = zc->entropy + 1; - - /* opt parser space */ - if ((params.cParams.strategy == ZSTD_btopt) || (params.cParams.strategy == ZSTD_btultra)) { - DEBUGLOG(5, "reserving optimal parser space"); - assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */ - zc->optState.litFreq = (U32*)ptr; - zc->optState.litLengthFreq = zc->optState.litFreq + (1<<Litbits); - zc->optState.matchLengthFreq = zc->optState.litLengthFreq + (MaxLL+1); - zc->optState.offCodeFreq = zc->optState.matchLengthFreq + (MaxML+1); - ptr = zc->optState.offCodeFreq + (MaxOff+1); - zc->optState.matchTable = (ZSTD_match_t*)ptr; - ptr = zc->optState.matchTable + ZSTD_OPT_NUM+1; - zc->optState.priceTable = (ZSTD_optimal_t*)ptr; - ptr = zc->optState.priceTable + ZSTD_OPT_NUM+1; - } - /* ldm hash table */ - /* initialize bucketOffsets table later for pointer alignment */ - if (params.ldmParams.enableLdm) { - size_t const ldmHSize = ((size_t)1) << params.ldmParams.hashLog; - memset(ptr, 0, ldmHSize * sizeof(ldmEntry_t)); - assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */ - zc->ldmState.hashTable = (ldmEntry_t*)ptr; - ptr = zc->ldmState.hashTable + ldmHSize; - } + ZSTD_reset_compressedBlockState(zc->blockState.prevCBlock); - /* table Space */ - if (crp!=ZSTDcrp_noMemset) memset(ptr, 0, tableSpace); /* reset tables only */ - assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */ - zc->hashTable = (U32*)(ptr); - zc->chainTable = zc->hashTable + hSize; - zc->hashTable3 = zc->chainTable + chainSize; - ptr = zc->hashTable3 + h3Size; + /* ZSTD_wildcopy() is used to copy into the literals buffer, + * so we have to oversize the buffer by WILDCOPY_OVERLENGTH bytes. + */ + zc->seqStore.litStart = ZSTD_cwksp_reserve_buffer(ws, blockSize + WILDCOPY_OVERLENGTH); + zc->seqStore.maxNbLit = blockSize; - /* sequences storage */ - zc->seqStore.sequencesStart = (seqDef*)ptr; - ptr = zc->seqStore.sequencesStart + maxNbSeq; - zc->seqStore.llCode = (BYTE*) ptr; - zc->seqStore.mlCode = zc->seqStore.llCode + maxNbSeq; - zc->seqStore.ofCode = zc->seqStore.mlCode + maxNbSeq; - zc->seqStore.litStart = zc->seqStore.ofCode + maxNbSeq; - ptr = zc->seqStore.litStart + blockSize; + /* buffers */ + zc->inBuffSize = buffInSize; + zc->inBuff = (char*)ZSTD_cwksp_reserve_buffer(ws, buffInSize); + zc->outBuffSize = buffOutSize; + zc->outBuff = (char*)ZSTD_cwksp_reserve_buffer(ws, buffOutSize); /* ldm bucketOffsets table */ if (params.ldmParams.enableLdm) { + /* TODO: avoid memset? */ size_t const ldmBucketSize = ((size_t)1) << (params.ldmParams.hashLog - params.ldmParams.bucketSizeLog); - memset(ptr, 0, ldmBucketSize); - zc->ldmState.bucketOffsets = (BYTE*)ptr; - ptr = zc->ldmState.bucketOffsets + ldmBucketSize; + zc->ldmState.bucketOffsets = ZSTD_cwksp_reserve_buffer(ws, ldmBucketSize); + memset(zc->ldmState.bucketOffsets, 0, ldmBucketSize); } - /* buffers */ - zc->inBuffSize = buffInSize; - zc->inBuff = (char*)ptr; - zc->outBuffSize = buffOutSize; - zc->outBuff = zc->inBuff + buffInSize; + /* sequences storage */ + ZSTD_referenceExternalSequences(zc, NULL, 0); + zc->seqStore.maxNbSeq = maxNbSeq; + zc->seqStore.llCode = ZSTD_cwksp_reserve_buffer(ws, maxNbSeq * sizeof(BYTE)); + zc->seqStore.mlCode = ZSTD_cwksp_reserve_buffer(ws, maxNbSeq * sizeof(BYTE)); + zc->seqStore.ofCode = ZSTD_cwksp_reserve_buffer(ws, maxNbSeq * sizeof(BYTE)); + zc->seqStore.sequencesStart = (seqDef*)ZSTD_cwksp_reserve_aligned(ws, maxNbSeq * sizeof(seqDef)); + + FORWARD_IF_ERROR(ZSTD_reset_matchState( + &zc->blockState.matchState, + ws, + ¶ms.cParams, + crp, + needsIndexReset, + ZSTD_resetTarget_CCtx), ""); + + /* ldm hash table */ + if (params.ldmParams.enableLdm) { + /* TODO: avoid memset? */ + size_t const ldmHSize = ((size_t)1) << params.ldmParams.hashLog; + zc->ldmState.hashTable = (ldmEntry_t*)ZSTD_cwksp_reserve_aligned(ws, ldmHSize * sizeof(ldmEntry_t)); + memset(zc->ldmState.hashTable, 0, ldmHSize * sizeof(ldmEntry_t)); + zc->ldmSequences = (rawSeq*)ZSTD_cwksp_reserve_aligned(ws, maxNbLdmSeq * sizeof(rawSeq)); + zc->maxNbLdmSequences = maxNbLdmSeq; + + ZSTD_window_init(&zc->ldmState.window); + ZSTD_window_clear(&zc->ldmState.window); + zc->ldmState.loadedDictEnd = 0; + } + + DEBUGLOG(3, "wksp: finished allocating, %zd bytes remain available", ZSTD_cwksp_available_space(ws)); + zc->initialized = 1; return 0; } @@ -975,25 +1592,188 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc, * do not use with extDict variant ! */ void ZSTD_invalidateRepCodes(ZSTD_CCtx* cctx) { int i; - for (i=0; i<ZSTD_REP_NUM; i++) cctx->seqStore.rep[i] = 0; + for (i=0; i<ZSTD_REP_NUM; i++) cctx->blockState.prevCBlock->rep[i] = 0; + assert(!ZSTD_window_hasExtDict(cctx->blockState.matchState.window)); +} + +/* These are the approximate sizes for each strategy past which copying the + * dictionary tables into the working context is faster than using them + * in-place. + */ +static const size_t attachDictSizeCutoffs[ZSTD_STRATEGY_MAX+1] = { + 8 KB, /* unused */ + 8 KB, /* ZSTD_fast */ + 16 KB, /* ZSTD_dfast */ + 32 KB, /* ZSTD_greedy */ + 32 KB, /* ZSTD_lazy */ + 32 KB, /* ZSTD_lazy2 */ + 32 KB, /* ZSTD_btlazy2 */ + 32 KB, /* ZSTD_btopt */ + 8 KB, /* ZSTD_btultra */ + 8 KB /* ZSTD_btultra2 */ +}; + +static int ZSTD_shouldAttachDict(const ZSTD_CDict* cdict, + const ZSTD_CCtx_params* params, + U64 pledgedSrcSize) +{ + size_t cutoff = attachDictSizeCutoffs[cdict->matchState.cParams.strategy]; + return ( pledgedSrcSize <= cutoff + || pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN + || params->attachDictPref == ZSTD_dictForceAttach ) + && params->attachDictPref != ZSTD_dictForceCopy + && !params->forceWindow; /* dictMatchState isn't correctly + * handled in _enforceMaxDist */ +} + +static size_t +ZSTD_resetCCtx_byAttachingCDict(ZSTD_CCtx* cctx, + const ZSTD_CDict* cdict, + ZSTD_CCtx_params params, + U64 pledgedSrcSize, + ZSTD_buffered_policy_e zbuff) +{ + { const ZSTD_compressionParameters* const cdict_cParams = &cdict->matchState.cParams; + unsigned const windowLog = params.cParams.windowLog; + assert(windowLog != 0); + /* Resize working context table params for input only, since the dict + * has its own tables. */ + /* pledgeSrcSize == 0 means 0! */ + params.cParams = ZSTD_adjustCParams_internal(*cdict_cParams, pledgedSrcSize, 0); + params.cParams.windowLog = windowLog; + FORWARD_IF_ERROR(ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize, + ZSTDcrp_makeClean, zbuff), ""); + assert(cctx->appliedParams.cParams.strategy == cdict_cParams->strategy); + } + + { const U32 cdictEnd = (U32)( cdict->matchState.window.nextSrc + - cdict->matchState.window.base); + const U32 cdictLen = cdictEnd - cdict->matchState.window.dictLimit; + if (cdictLen == 0) { + /* don't even attach dictionaries with no contents */ + DEBUGLOG(4, "skipping attaching empty dictionary"); + } else { + DEBUGLOG(4, "attaching dictionary into context"); + cctx->blockState.matchState.dictMatchState = &cdict->matchState; + + /* prep working match state so dict matches never have negative indices + * when they are translated to the working context's index space. */ + if (cctx->blockState.matchState.window.dictLimit < cdictEnd) { + cctx->blockState.matchState.window.nextSrc = + cctx->blockState.matchState.window.base + cdictEnd; + ZSTD_window_clear(&cctx->blockState.matchState.window); + } + /* loadedDictEnd is expressed within the referential of the active context */ + cctx->blockState.matchState.loadedDictEnd = cctx->blockState.matchState.window.dictLimit; + } } + + cctx->dictID = cdict->dictID; + + /* copy block state */ + memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState)); + + return 0; +} + +static size_t ZSTD_resetCCtx_byCopyingCDict(ZSTD_CCtx* cctx, + const ZSTD_CDict* cdict, + ZSTD_CCtx_params params, + U64 pledgedSrcSize, + ZSTD_buffered_policy_e zbuff) +{ + const ZSTD_compressionParameters *cdict_cParams = &cdict->matchState.cParams; + + DEBUGLOG(4, "copying dictionary into context"); + + { unsigned const windowLog = params.cParams.windowLog; + assert(windowLog != 0); + /* Copy only compression parameters related to tables. */ + params.cParams = *cdict_cParams; + params.cParams.windowLog = windowLog; + FORWARD_IF_ERROR(ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize, + ZSTDcrp_leaveDirty, zbuff), ""); + assert(cctx->appliedParams.cParams.strategy == cdict_cParams->strategy); + assert(cctx->appliedParams.cParams.hashLog == cdict_cParams->hashLog); + assert(cctx->appliedParams.cParams.chainLog == cdict_cParams->chainLog); + } + + ZSTD_cwksp_mark_tables_dirty(&cctx->workspace); + + /* copy tables */ + { size_t const chainSize = (cdict_cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cdict_cParams->chainLog); + size_t const hSize = (size_t)1 << cdict_cParams->hashLog; + + memcpy(cctx->blockState.matchState.hashTable, + cdict->matchState.hashTable, + hSize * sizeof(U32)); + memcpy(cctx->blockState.matchState.chainTable, + cdict->matchState.chainTable, + chainSize * sizeof(U32)); + } + + /* Zero the hashTable3, since the cdict never fills it */ + { int const h3log = cctx->blockState.matchState.hashLog3; + size_t const h3Size = h3log ? ((size_t)1 << h3log) : 0; + assert(cdict->matchState.hashLog3 == 0); + memset(cctx->blockState.matchState.hashTable3, 0, h3Size * sizeof(U32)); + } + + ZSTD_cwksp_mark_tables_clean(&cctx->workspace); + + /* copy dictionary offsets */ + { ZSTD_matchState_t const* srcMatchState = &cdict->matchState; + ZSTD_matchState_t* dstMatchState = &cctx->blockState.matchState; + dstMatchState->window = srcMatchState->window; + dstMatchState->nextToUpdate = srcMatchState->nextToUpdate; + dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd; + } + + cctx->dictID = cdict->dictID; + + /* copy block state */ + memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState)); + + return 0; } +/* We have a choice between copying the dictionary context into the working + * context, or referencing the dictionary context from the working context + * in-place. We decide here which strategy to use. */ +static size_t ZSTD_resetCCtx_usingCDict(ZSTD_CCtx* cctx, + const ZSTD_CDict* cdict, + const ZSTD_CCtx_params* params, + U64 pledgedSrcSize, + ZSTD_buffered_policy_e zbuff) +{ + + DEBUGLOG(4, "ZSTD_resetCCtx_usingCDict (pledgedSrcSize=%u)", + (unsigned)pledgedSrcSize); + + if (ZSTD_shouldAttachDict(cdict, params, pledgedSrcSize)) { + return ZSTD_resetCCtx_byAttachingCDict( + cctx, cdict, *params, pledgedSrcSize, zbuff); + } else { + return ZSTD_resetCCtx_byCopyingCDict( + cctx, cdict, *params, pledgedSrcSize, zbuff); + } +} /*! ZSTD_copyCCtx_internal() : * Duplicate an existing context `srcCCtx` into another one `dstCCtx`. - * The "context", in this case, refers to the hash and chain tables, entropy - * tables, and dictionary offsets. * Only works during stage ZSTDcs_init (i.e. after creation, but before first call to ZSTD_compressContinue()). - * pledgedSrcSize=0 means "empty" if fParams.contentSizeFlag=1 - * @return : 0, or an error code */ + * The "context", in this case, refers to the hash and chain tables, + * entropy tables, and dictionary references. + * `windowLog` value is enforced if != 0, otherwise value is copied from srcCCtx. + * @return : 0, or an error code */ static size_t ZSTD_copyCCtx_internal(ZSTD_CCtx* dstCCtx, const ZSTD_CCtx* srcCCtx, ZSTD_frameParameters fParams, - unsigned long long pledgedSrcSize, + U64 pledgedSrcSize, ZSTD_buffered_policy_e zbuff) { DEBUGLOG(5, "ZSTD_copyCCtx_internal"); - if (srcCCtx->stage!=ZSTDcs_init) return ERROR(stage_wrong); + RETURN_ERROR_IF(srcCCtx->stage!=ZSTDcs_init, stage_wrong, + "Can't copy a ctx that's not in init stage."); memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem)); { ZSTD_CCtx_params params = dstCCtx->requestedParams; @@ -1001,32 +1781,47 @@ static size_t ZSTD_copyCCtx_internal(ZSTD_CCtx* dstCCtx, params.cParams = srcCCtx->appliedParams.cParams; params.fParams = fParams; ZSTD_resetCCtx_internal(dstCCtx, params, pledgedSrcSize, - ZSTDcrp_noMemset, zbuff); + ZSTDcrp_leaveDirty, zbuff); + assert(dstCCtx->appliedParams.cParams.windowLog == srcCCtx->appliedParams.cParams.windowLog); + assert(dstCCtx->appliedParams.cParams.strategy == srcCCtx->appliedParams.cParams.strategy); + assert(dstCCtx->appliedParams.cParams.hashLog == srcCCtx->appliedParams.cParams.hashLog); + assert(dstCCtx->appliedParams.cParams.chainLog == srcCCtx->appliedParams.cParams.chainLog); + assert(dstCCtx->blockState.matchState.hashLog3 == srcCCtx->blockState.matchState.hashLog3); } + ZSTD_cwksp_mark_tables_dirty(&dstCCtx->workspace); + /* copy tables */ - { size_t const chainSize = (srcCCtx->appliedParams.cParams.strategy == ZSTD_fast) ? 0 : (1 << srcCCtx->appliedParams.cParams.chainLog); + { size_t const chainSize = (srcCCtx->appliedParams.cParams.strategy == ZSTD_fast) ? 0 : ((size_t)1 << srcCCtx->appliedParams.cParams.chainLog); size_t const hSize = (size_t)1 << srcCCtx->appliedParams.cParams.hashLog; - size_t const h3Size = (size_t)1 << srcCCtx->hashLog3; - size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32); - assert((U32*)dstCCtx->chainTable == (U32*)dstCCtx->hashTable + hSize); /* chainTable must follow hashTable */ - assert((U32*)dstCCtx->hashTable3 == (U32*)dstCCtx->chainTable + chainSize); - memcpy(dstCCtx->hashTable, srcCCtx->hashTable, tableSpace); /* presumes all tables follow each other */ + int const h3log = srcCCtx->blockState.matchState.hashLog3; + size_t const h3Size = h3log ? ((size_t)1 << h3log) : 0; + + memcpy(dstCCtx->blockState.matchState.hashTable, + srcCCtx->blockState.matchState.hashTable, + hSize * sizeof(U32)); + memcpy(dstCCtx->blockState.matchState.chainTable, + srcCCtx->blockState.matchState.chainTable, + chainSize * sizeof(U32)); + memcpy(dstCCtx->blockState.matchState.hashTable3, + srcCCtx->blockState.matchState.hashTable3, + h3Size * sizeof(U32)); } + ZSTD_cwksp_mark_tables_clean(&dstCCtx->workspace); + /* copy dictionary offsets */ - dstCCtx->nextToUpdate = srcCCtx->nextToUpdate; - dstCCtx->nextToUpdate3= srcCCtx->nextToUpdate3; - dstCCtx->nextSrc = srcCCtx->nextSrc; - dstCCtx->base = srcCCtx->base; - dstCCtx->dictBase = srcCCtx->dictBase; - dstCCtx->dictLimit = srcCCtx->dictLimit; - dstCCtx->lowLimit = srcCCtx->lowLimit; - dstCCtx->loadedDictEnd= srcCCtx->loadedDictEnd; - dstCCtx->dictID = srcCCtx->dictID; - - /* copy entropy tables */ - memcpy(dstCCtx->entropy, srcCCtx->entropy, sizeof(ZSTD_entropyCTables_t)); + { + const ZSTD_matchState_t* srcMatchState = &srcCCtx->blockState.matchState; + ZSTD_matchState_t* dstMatchState = &dstCCtx->blockState.matchState; + dstMatchState->window = srcMatchState->window; + dstMatchState->nextToUpdate = srcMatchState->nextToUpdate; + dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd; + } + dstCCtx->dictID = srcCCtx->dictID; + + /* copy block state */ + memcpy(dstCCtx->blockState.prevCBlock, srcCCtx->blockState.prevCBlock, sizeof(*srcCCtx->blockState.prevCBlock)); return 0; } @@ -1041,52 +1836,86 @@ size_t ZSTD_copyCCtx(ZSTD_CCtx* dstCCtx, const ZSTD_CCtx* srcCCtx, unsigned long ZSTD_frameParameters fParams = { 1 /*content*/, 0 /*checksum*/, 0 /*noDictID*/ }; ZSTD_buffered_policy_e const zbuff = (ZSTD_buffered_policy_e)(srcCCtx->inBuffSize>0); ZSTD_STATIC_ASSERT((U32)ZSTDb_buffered==1); - fParams.contentSizeFlag = pledgedSrcSize>0; + if (pledgedSrcSize==0) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN; + fParams.contentSizeFlag = (pledgedSrcSize != ZSTD_CONTENTSIZE_UNKNOWN); - return ZSTD_copyCCtx_internal(dstCCtx, srcCCtx, fParams, pledgedSrcSize, zbuff); + return ZSTD_copyCCtx_internal(dstCCtx, srcCCtx, + fParams, pledgedSrcSize, + zbuff); } +#define ZSTD_ROWSIZE 16 /*! ZSTD_reduceTable() : - * reduce table indexes by `reducerValue` */ -static void ZSTD_reduceTable (U32* const table, U32 const size, U32 const reducerValue) + * reduce table indexes by `reducerValue`, or squash to zero. + * PreserveMark preserves "unsorted mark" for btlazy2 strategy. + * It must be set to a clear 0/1 value, to remove branch during inlining. + * Presume table size is a multiple of ZSTD_ROWSIZE + * to help auto-vectorization */ +FORCE_INLINE_TEMPLATE void +ZSTD_reduceTable_internal (U32* const table, U32 const size, U32 const reducerValue, int const preserveMark) { - U32 u; - for (u=0 ; u < size ; u++) { - if (table[u] < reducerValue) table[u] = 0; - else table[u] -= reducerValue; - } + int const nbRows = (int)size / ZSTD_ROWSIZE; + int cellNb = 0; + int rowNb; + assert((size & (ZSTD_ROWSIZE-1)) == 0); /* multiple of ZSTD_ROWSIZE */ + assert(size < (1U<<31)); /* can be casted to int */ + +#if defined (MEMORY_SANITIZER) && !defined (ZSTD_MSAN_DONT_POISON_WORKSPACE) + /* To validate that the table re-use logic is sound, and that we don't + * access table space that we haven't cleaned, we re-"poison" the table + * space every time we mark it dirty. + * + * This function however is intended to operate on those dirty tables and + * re-clean them. So when this function is used correctly, we can unpoison + * the memory it operated on. This introduces a blind spot though, since + * if we now try to operate on __actually__ poisoned memory, we will not + * detect that. */ + __msan_unpoison(table, size * sizeof(U32)); +#endif + + for (rowNb=0 ; rowNb < nbRows ; rowNb++) { + int column; + for (column=0; column<ZSTD_ROWSIZE; column++) { + if (preserveMark) { + U32 const adder = (table[cellNb] == ZSTD_DUBT_UNSORTED_MARK) ? reducerValue : 0; + table[cellNb] += adder; + } + if (table[cellNb] < reducerValue) table[cellNb] = 0; + else table[cellNb] -= reducerValue; + cellNb++; + } } } -/*! ZSTD_ldm_reduceTable() : - * reduce table indexes by `reducerValue` */ -static void ZSTD_ldm_reduceTable(ldmEntry_t* const table, U32 const size, - U32 const reducerValue) +static void ZSTD_reduceTable(U32* const table, U32 const size, U32 const reducerValue) { - U32 u; - for (u = 0; u < size; u++) { - if (table[u].offset < reducerValue) table[u].offset = 0; - else table[u].offset -= reducerValue; - } + ZSTD_reduceTable_internal(table, size, reducerValue, 0); +} + +static void ZSTD_reduceTable_btlazy2(U32* const table, U32 const size, U32 const reducerValue) +{ + ZSTD_reduceTable_internal(table, size, reducerValue, 1); } /*! ZSTD_reduceIndex() : * rescale all indexes to avoid future overflow (indexes are U32) */ -static void ZSTD_reduceIndex (ZSTD_CCtx* zc, const U32 reducerValue) +static void ZSTD_reduceIndex (ZSTD_matchState_t* ms, ZSTD_CCtx_params const* params, const U32 reducerValue) { - { U32 const hSize = (U32)1 << zc->appliedParams.cParams.hashLog; - ZSTD_reduceTable(zc->hashTable, hSize, reducerValue); } - - { U32 const chainSize = (zc->appliedParams.cParams.strategy == ZSTD_fast) ? 0 : ((U32)1 << zc->appliedParams.cParams.chainLog); - ZSTD_reduceTable(zc->chainTable, chainSize, reducerValue); } + { U32 const hSize = (U32)1 << params->cParams.hashLog; + ZSTD_reduceTable(ms->hashTable, hSize, reducerValue); + } - { U32 const h3Size = (zc->hashLog3) ? (U32)1 << zc->hashLog3 : 0; - ZSTD_reduceTable(zc->hashTable3, h3Size, reducerValue); } + if (params->cParams.strategy != ZSTD_fast) { + U32 const chainSize = (U32)1 << params->cParams.chainLog; + if (params->cParams.strategy == ZSTD_btlazy2) + ZSTD_reduceTable_btlazy2(ms->chainTable, chainSize, reducerValue); + else + ZSTD_reduceTable(ms->chainTable, chainSize, reducerValue); + } - { if (zc->appliedParams.ldmParams.enableLdm) { - U32 const ldmHSize = (U32)1 << zc->appliedParams.ldmParams.hashLog; - ZSTD_ldm_reduceTable(zc->ldmState.hashTable, ldmHSize, reducerValue); - } + if (ms->hashLog3) { + U32 const h3Size = (U32)1 << ms->hashLog3; + ZSTD_reduceTable(ms->hashTable3, h3Size, reducerValue); } } @@ -1097,152 +1926,21 @@ static void ZSTD_reduceIndex (ZSTD_CCtx* zc, const U32 reducerValue) /* See doc/zstd_compression_format.md for detailed format description */ -size_t ZSTD_noCompressBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize) -{ - if (srcSize + ZSTD_blockHeaderSize > dstCapacity) return ERROR(dstSize_tooSmall); - memcpy((BYTE*)dst + ZSTD_blockHeaderSize, src, srcSize); - MEM_writeLE24(dst, (U32)(srcSize << 2) + (U32)bt_raw); - return ZSTD_blockHeaderSize+srcSize; -} - - -static size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src, size_t srcSize) -{ - BYTE* const ostart = (BYTE* const)dst; - U32 const flSize = 1 + (srcSize>31) + (srcSize>4095); - - if (srcSize + flSize > dstCapacity) return ERROR(dstSize_tooSmall); - - switch(flSize) - { - case 1: /* 2 - 1 - 5 */ - ostart[0] = (BYTE)((U32)set_basic + (srcSize<<3)); - break; - case 2: /* 2 - 2 - 12 */ - MEM_writeLE16(ostart, (U16)((U32)set_basic + (1<<2) + (srcSize<<4))); - break; - case 3: /* 2 - 2 - 20 */ - MEM_writeLE32(ostart, (U32)((U32)set_basic + (3<<2) + (srcSize<<4))); - break; - default: /* not necessary : flSize is {1,2,3} */ - assert(0); - } - - memcpy(ostart + flSize, src, srcSize); - return srcSize + flSize; -} - -static size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize) -{ - BYTE* const ostart = (BYTE* const)dst; - U32 const flSize = 1 + (srcSize>31) + (srcSize>4095); - - (void)dstCapacity; /* dstCapacity already guaranteed to be >=4, hence large enough */ - - switch(flSize) - { - case 1: /* 2 - 1 - 5 */ - ostart[0] = (BYTE)((U32)set_rle + (srcSize<<3)); - break; - case 2: /* 2 - 2 - 12 */ - MEM_writeLE16(ostart, (U16)((U32)set_rle + (1<<2) + (srcSize<<4))); - break; - case 3: /* 2 - 2 - 20 */ - MEM_writeLE32(ostart, (U32)((U32)set_rle + (3<<2) + (srcSize<<4))); - break; - default: /* not necessary : flSize is {1,2,3} */ - assert(0); - } - - ostart[flSize] = *(const BYTE*)src; - return flSize+1; -} - - -static size_t ZSTD_minGain(size_t srcSize) { return (srcSize >> 6) + 2; } - -static size_t ZSTD_compressLiterals (ZSTD_entropyCTables_t * entropy, - ZSTD_strategy strategy, - void* dst, size_t dstCapacity, - const void* src, size_t srcSize) -{ - size_t const minGain = ZSTD_minGain(srcSize); - size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB); - BYTE* const ostart = (BYTE*)dst; - U32 singleStream = srcSize < 256; - symbolEncodingType_e hType = set_compressed; - size_t cLitSize; - - - /* small ? don't even attempt compression (speed opt) */ -# define LITERAL_NOENTROPY 63 - { size_t const minLitSize = entropy->hufCTable_repeatMode == HUF_repeat_valid ? 6 : LITERAL_NOENTROPY; - if (srcSize <= minLitSize) return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize); - } - - if (dstCapacity < lhSize+1) return ERROR(dstSize_tooSmall); /* not enough space for compression */ - { HUF_repeat repeat = entropy->hufCTable_repeatMode; - int const preferRepeat = strategy < ZSTD_lazy ? srcSize <= 1024 : 0; - if (repeat == HUF_repeat_valid && lhSize == 3) singleStream = 1; - cLitSize = singleStream ? HUF_compress1X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11, - entropy->workspace, sizeof(entropy->workspace), (HUF_CElt*)entropy->hufCTable, &repeat, preferRepeat) - : HUF_compress4X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11, - entropy->workspace, sizeof(entropy->workspace), (HUF_CElt*)entropy->hufCTable, &repeat, preferRepeat); - if (repeat != HUF_repeat_none) { hType = set_repeat; } /* reused the existing table */ - else { entropy->hufCTable_repeatMode = HUF_repeat_check; } /* now have a table to reuse */ - } - - if ((cLitSize==0) | (cLitSize >= srcSize - minGain) | ERR_isError(cLitSize)) { - entropy->hufCTable_repeatMode = HUF_repeat_none; - return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize); - } - if (cLitSize==1) { - entropy->hufCTable_repeatMode = HUF_repeat_none; - return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize); - } - - /* Build header */ - switch(lhSize) - { - case 3: /* 2 - 2 - 10 - 10 */ - { U32 const lhc = hType + ((!singleStream) << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<14); - MEM_writeLE24(ostart, lhc); - break; - } - case 4: /* 2 - 2 - 14 - 14 */ - { U32 const lhc = hType + (2 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<18); - MEM_writeLE32(ostart, lhc); - break; - } - case 5: /* 2 - 2 - 18 - 18 */ - { U32 const lhc = hType + (3 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<22); - MEM_writeLE32(ostart, lhc); - ostart[4] = (BYTE)(cLitSize >> 10); - break; - } - default: /* not possible : lhSize is {3,4,5} */ - assert(0); - } - return lhSize+cLitSize; -} - - void ZSTD_seqToCodes(const seqStore_t* seqStorePtr) { - BYTE const LL_deltaCode = 19; - BYTE const ML_deltaCode = 36; const seqDef* const sequences = seqStorePtr->sequencesStart; BYTE* const llCodeTable = seqStorePtr->llCode; BYTE* const ofCodeTable = seqStorePtr->ofCode; BYTE* const mlCodeTable = seqStorePtr->mlCode; U32 const nbSeq = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart); U32 u; + assert(nbSeq <= seqStorePtr->maxNbSeq); for (u=0; u<nbSeq; u++) { U32 const llv = sequences[u].litLength; U32 const mlv = sequences[u].matchLength; - llCodeTable[u] = (llv> 63) ? (BYTE)ZSTD_highbit32(llv) + LL_deltaCode : LL_Code[llv]; + llCodeTable[u] = (BYTE)ZSTD_LLcode(llv); ofCodeTable[u] = (BYTE)ZSTD_highbit32(sequences[u].offset); - mlCodeTable[u] = (mlv>127) ? (BYTE)ZSTD_highbit32(mlv) + ML_deltaCode : ML_Code[mlv]; + mlCodeTable[u] = (BYTE)ZSTD_MLcode(mlv); } if (seqStorePtr->longLengthID==1) llCodeTable[seqStorePtr->longLengthPos] = MaxLL; @@ -1250,154 +1948,33 @@ void ZSTD_seqToCodes(const seqStore_t* seqStorePtr) mlCodeTable[seqStorePtr->longLengthPos] = MaxML; } -MEM_STATIC symbolEncodingType_e ZSTD_selectEncodingType(FSE_repeat* repeatMode, - size_t const mostFrequent, size_t nbSeq, U32 defaultNormLog) -{ -#define MIN_SEQ_FOR_DYNAMIC_FSE 64 -#define MAX_SEQ_FOR_STATIC_FSE 1000 - - if ((mostFrequent == nbSeq) && (nbSeq > 2)) { - *repeatMode = FSE_repeat_check; - return set_rle; - } - if ((*repeatMode == FSE_repeat_valid) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) { - return set_repeat; - } - if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (defaultNormLog-1)))) { - *repeatMode = FSE_repeat_valid; - return set_basic; - } - *repeatMode = FSE_repeat_check; - return set_compressed; -} - -MEM_STATIC size_t ZSTD_buildCTable(void* dst, size_t dstCapacity, - FSE_CTable* CTable, U32 FSELog, symbolEncodingType_e type, - U32* count, U32 max, - BYTE const* codeTable, size_t nbSeq, - S16 const* defaultNorm, U32 defaultNormLog, U32 defaultMax, - void* workspace, size_t workspaceSize) +/* ZSTD_useTargetCBlockSize(): + * Returns if target compressed block size param is being used. + * If used, compression will do best effort to make a compressed block size to be around targetCBlockSize. + * Returns 1 if true, 0 otherwise. */ +static int ZSTD_useTargetCBlockSize(const ZSTD_CCtx_params* cctxParams) { - BYTE* op = (BYTE*)dst; - BYTE const* const oend = op + dstCapacity; - - switch (type) { - case set_rle: - *op = codeTable[0]; - CHECK_F(FSE_buildCTable_rle(CTable, (BYTE)max)); - return 1; - case set_repeat: - return 0; - case set_basic: - CHECK_F(FSE_buildCTable_wksp(CTable, defaultNorm, defaultMax, defaultNormLog, workspace, workspaceSize)); - return 0; - case set_compressed: { - S16 norm[MaxSeq + 1]; - size_t nbSeq_1 = nbSeq; - const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max); - if (count[codeTable[nbSeq-1]] > 1) { - count[codeTable[nbSeq-1]]--; - nbSeq_1--; - } - CHECK_F(FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max)); - { size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog); /* overflow protected */ - if (FSE_isError(NCountSize)) return NCountSize; - CHECK_F(FSE_buildCTable_wksp(CTable, norm, max, tableLog, workspace, workspaceSize)); - return NCountSize; - } - } - default: return assert(0), ERROR(GENERIC); - } -} - -MEM_STATIC size_t ZSTD_encodeSequences(void* dst, size_t dstCapacity, - FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable, - FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable, - FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable, - seqDef const* sequences, size_t nbSeq, int longOffsets) -{ - BIT_CStream_t blockStream; - FSE_CState_t stateMatchLength; - FSE_CState_t stateOffsetBits; - FSE_CState_t stateLitLength; - - CHECK_E(BIT_initCStream(&blockStream, dst, dstCapacity), dstSize_tooSmall); /* not enough space remaining */ - - /* first symbols */ - FSE_initCState2(&stateMatchLength, CTable_MatchLength, mlCodeTable[nbSeq-1]); - FSE_initCState2(&stateOffsetBits, CTable_OffsetBits, ofCodeTable[nbSeq-1]); - FSE_initCState2(&stateLitLength, CTable_LitLength, llCodeTable[nbSeq-1]); - BIT_addBits(&blockStream, sequences[nbSeq-1].litLength, LL_bits[llCodeTable[nbSeq-1]]); - if (MEM_32bits()) BIT_flushBits(&blockStream); - BIT_addBits(&blockStream, sequences[nbSeq-1].matchLength, ML_bits[mlCodeTable[nbSeq-1]]); - if (MEM_32bits()) BIT_flushBits(&blockStream); - if (longOffsets) { - U32 const ofBits = ofCodeTable[nbSeq-1]; - int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1); - if (extraBits) { - BIT_addBits(&blockStream, sequences[nbSeq-1].offset, extraBits); - BIT_flushBits(&blockStream); - } - BIT_addBits(&blockStream, sequences[nbSeq-1].offset >> extraBits, - ofBits - extraBits); - } else { - BIT_addBits(&blockStream, sequences[nbSeq-1].offset, ofCodeTable[nbSeq-1]); - } - BIT_flushBits(&blockStream); - - { size_t n; - for (n=nbSeq-2 ; n<nbSeq ; n--) { /* intentional underflow */ - BYTE const llCode = llCodeTable[n]; - BYTE const ofCode = ofCodeTable[n]; - BYTE const mlCode = mlCodeTable[n]; - U32 const llBits = LL_bits[llCode]; - U32 const ofBits = ofCode; /* 32b*/ /* 64b*/ - U32 const mlBits = ML_bits[mlCode]; - /* (7)*/ /* (7)*/ - FSE_encodeSymbol(&blockStream, &stateOffsetBits, ofCode); /* 15 */ /* 15 */ - FSE_encodeSymbol(&blockStream, &stateMatchLength, mlCode); /* 24 */ /* 24 */ - if (MEM_32bits()) BIT_flushBits(&blockStream); /* (7)*/ - FSE_encodeSymbol(&blockStream, &stateLitLength, llCode); /* 16 */ /* 33 */ - if (MEM_32bits() || (ofBits+mlBits+llBits >= 64-7-(LLFSELog+MLFSELog+OffFSELog))) - BIT_flushBits(&blockStream); /* (7)*/ - BIT_addBits(&blockStream, sequences[n].litLength, llBits); - if (MEM_32bits() && ((llBits+mlBits)>24)) BIT_flushBits(&blockStream); - BIT_addBits(&blockStream, sequences[n].matchLength, mlBits); - if (MEM_32bits() || (ofBits+mlBits+llBits > 56)) BIT_flushBits(&blockStream); - if (longOffsets) { - int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1); - if (extraBits) { - BIT_addBits(&blockStream, sequences[n].offset, extraBits); - BIT_flushBits(&blockStream); /* (7)*/ - } - BIT_addBits(&blockStream, sequences[n].offset >> extraBits, - ofBits - extraBits); /* 31 */ - } else { - BIT_addBits(&blockStream, sequences[n].offset, ofBits); /* 31 */ - } - BIT_flushBits(&blockStream); /* (7)*/ - } } - - FSE_flushCState(&blockStream, &stateMatchLength); - FSE_flushCState(&blockStream, &stateOffsetBits); - FSE_flushCState(&blockStream, &stateLitLength); - - { size_t const streamSize = BIT_closeCStream(&blockStream); - if (streamSize==0) return ERROR(dstSize_tooSmall); /* not enough space */ - return streamSize; - } + DEBUGLOG(5, "ZSTD_useTargetCBlockSize (targetCBlockSize=%zu)", cctxParams->targetCBlockSize); + return (cctxParams->targetCBlockSize != 0); } -MEM_STATIC size_t ZSTD_compressSequences_internal(seqStore_t* seqStorePtr, - ZSTD_entropyCTables_t* entropy, - ZSTD_compressionParameters const* cParams, - void* dst, size_t dstCapacity) +/* ZSTD_compressSequences_internal(): + * actually compresses both literals and sequences */ +MEM_STATIC size_t +ZSTD_compressSequences_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) { - const int longOffsets = cParams->windowLog > STREAM_ACCUMULATOR_MIN; - U32 count[MaxSeq+1]; - FSE_CTable* CTable_LitLength = entropy->litlengthCTable; - FSE_CTable* CTable_OffsetBits = entropy->offcodeCTable; - FSE_CTable* CTable_MatchLength = entropy->matchlengthCTable; + const int longOffsets = cctxParams->cParams.windowLog > STREAM_ACCUMULATOR_MIN; + ZSTD_strategy const strategy = cctxParams->cParams.strategy; + unsigned count[MaxSeq+1]; + FSE_CTable* CTable_LitLength = nextEntropy->fse.litlengthCTable; + FSE_CTable* CTable_OffsetBits = nextEntropy->fse.offcodeCTable; + FSE_CTable* CTable_MatchLength = nextEntropy->fse.matchlengthCTable; U32 LLtype, Offtype, MLtype; /* compressed, raw or rle */ const seqDef* const sequences = seqStorePtr->sequencesStart; const BYTE* const ofCodeTable = seqStorePtr->ofCode; @@ -1406,132 +1983,244 @@ MEM_STATIC size_t ZSTD_compressSequences_internal(seqStore_t* seqStorePtr, BYTE* const ostart = (BYTE*)dst; BYTE* const oend = ostart + dstCapacity; BYTE* op = ostart; - size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart; + size_t const nbSeq = (size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart); BYTE* seqHead; + BYTE* lastNCount = NULL; - ZSTD_STATIC_ASSERT(sizeof(entropy->workspace) >= (1<<MAX(MLFSELog,LLFSELog))); + DEBUGLOG(5, "ZSTD_compressSequences_internal (nbSeq=%zu)", nbSeq); + ZSTD_STATIC_ASSERT(HUF_WORKSPACE_SIZE >= (1<<MAX(MLFSELog,LLFSELog))); /* Compress literals */ { const BYTE* const literals = seqStorePtr->litStart; - size_t const litSize = seqStorePtr->lit - literals; + size_t const litSize = (size_t)(seqStorePtr->lit - literals); size_t const cSize = ZSTD_compressLiterals( - entropy, cParams->strategy, op, dstCapacity, literals, litSize); - if (ZSTD_isError(cSize)) - return cSize; + &prevEntropy->huf, &nextEntropy->huf, + cctxParams->cParams.strategy, + ZSTD_disableLiteralsCompression(cctxParams), + op, dstCapacity, + literals, litSize, + entropyWorkspace, entropyWkspSize, + bmi2); + FORWARD_IF_ERROR(cSize, "ZSTD_compressLiterals failed"); + assert(cSize <= dstCapacity); op += cSize; } /* Sequences Header */ - if ((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead */) return ERROR(dstSize_tooSmall); - if (nbSeq < 0x7F) *op++ = (BYTE)nbSeq; - else if (nbSeq < LONGNBSEQ) op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2; - else op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3; - if (nbSeq==0) return op - ostart; + RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/, + dstSize_tooSmall, "Can't fit seq hdr in output buf!"); + if (nbSeq < 128) { + *op++ = (BYTE)nbSeq; + } else if (nbSeq < LONGNBSEQ) { + op[0] = (BYTE)((nbSeq>>8) + 0x80); + op[1] = (BYTE)nbSeq; + op+=2; + } else { + op[0]=0xFF; + MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)); + op+=3; + } + assert(op <= oend); + if (nbSeq==0) { + /* Copy the old tables over as if we repeated them */ + memcpy(&nextEntropy->fse, &prevEntropy->fse, sizeof(prevEntropy->fse)); + return (size_t)(op - ostart); + } /* seqHead : flags for FSE encoding type */ seqHead = op++; + assert(op <= oend); /* convert length/distances into codes */ ZSTD_seqToCodes(seqStorePtr); - /* CTable for Literal Lengths */ - { U32 max = MaxLL; - size_t const mostFrequent = FSE_countFast_wksp(count, &max, llCodeTable, nbSeq, entropy->workspace); - LLtype = ZSTD_selectEncodingType(&entropy->litlength_repeatMode, mostFrequent, nbSeq, LL_defaultNormLog); - { size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype, - count, max, llCodeTable, nbSeq, LL_defaultNorm, LL_defaultNormLog, MaxLL, - entropy->workspace, sizeof(entropy->workspace)); - if (ZSTD_isError(countSize)) return countSize; + /* build CTable for Literal Lengths */ + { unsigned max = MaxLL; + size_t const mostFrequent = HIST_countFast_wksp(count, &max, llCodeTable, nbSeq, entropyWorkspace, entropyWkspSize); /* can't fail */ + DEBUGLOG(5, "Building LL table"); + nextEntropy->fse.litlength_repeatMode = prevEntropy->fse.litlength_repeatMode; + LLtype = ZSTD_selectEncodingType(&nextEntropy->fse.litlength_repeatMode, + count, max, mostFrequent, nbSeq, + LLFSELog, prevEntropy->fse.litlengthCTable, + LL_defaultNorm, LL_defaultNormLog, + ZSTD_defaultAllowed, strategy); + assert(set_basic < set_compressed && set_rle < set_compressed); + assert(!(LLtype < set_compressed && nextEntropy->fse.litlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */ + { size_t const countSize = ZSTD_buildCTable( + op, (size_t)(oend - op), + CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype, + count, max, llCodeTable, nbSeq, + LL_defaultNorm, LL_defaultNormLog, MaxLL, + prevEntropy->fse.litlengthCTable, + sizeof(prevEntropy->fse.litlengthCTable), + entropyWorkspace, entropyWkspSize); + FORWARD_IF_ERROR(countSize, "ZSTD_buildCTable for LitLens failed"); + if (LLtype == set_compressed) + lastNCount = op; op += countSize; + assert(op <= oend); } } - /* CTable for Offsets */ - { U32 max = MaxOff; - size_t const mostFrequent = FSE_countFast_wksp(count, &max, ofCodeTable, nbSeq, entropy->workspace); - Offtype = ZSTD_selectEncodingType(&entropy->offcode_repeatMode, mostFrequent, nbSeq, OF_defaultNormLog); - { size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype, - count, max, ofCodeTable, nbSeq, OF_defaultNorm, OF_defaultNormLog, MaxOff, - entropy->workspace, sizeof(entropy->workspace)); - if (ZSTD_isError(countSize)) return countSize; + /* build CTable for Offsets */ + { unsigned max = MaxOff; + size_t const mostFrequent = HIST_countFast_wksp( + count, &max, ofCodeTable, nbSeq, entropyWorkspace, entropyWkspSize); /* can't fail */ + /* We can only use the basic table if max <= DefaultMaxOff, otherwise the offsets are too large */ + ZSTD_defaultPolicy_e const defaultPolicy = (max <= DefaultMaxOff) ? ZSTD_defaultAllowed : ZSTD_defaultDisallowed; + DEBUGLOG(5, "Building OF table"); + nextEntropy->fse.offcode_repeatMode = prevEntropy->fse.offcode_repeatMode; + Offtype = ZSTD_selectEncodingType(&nextEntropy->fse.offcode_repeatMode, + count, max, mostFrequent, nbSeq, + OffFSELog, prevEntropy->fse.offcodeCTable, + OF_defaultNorm, OF_defaultNormLog, + defaultPolicy, strategy); + assert(!(Offtype < set_compressed && nextEntropy->fse.offcode_repeatMode != FSE_repeat_none)); /* We don't copy tables */ + { size_t const countSize = ZSTD_buildCTable( + op, (size_t)(oend - op), + CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype, + count, max, ofCodeTable, nbSeq, + OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff, + prevEntropy->fse.offcodeCTable, + sizeof(prevEntropy->fse.offcodeCTable), + entropyWorkspace, entropyWkspSize); + FORWARD_IF_ERROR(countSize, "ZSTD_buildCTable for Offsets failed"); + if (Offtype == set_compressed) + lastNCount = op; op += countSize; + assert(op <= oend); } } - /* CTable for MatchLengths */ - { U32 max = MaxML; - size_t const mostFrequent = FSE_countFast_wksp(count, &max, mlCodeTable, nbSeq, entropy->workspace); - MLtype = ZSTD_selectEncodingType(&entropy->matchlength_repeatMode, mostFrequent, nbSeq, ML_defaultNormLog); - { size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype, - count, max, mlCodeTable, nbSeq, ML_defaultNorm, ML_defaultNormLog, MaxML, - entropy->workspace, sizeof(entropy->workspace)); - if (ZSTD_isError(countSize)) return countSize; + /* build CTable for MatchLengths */ + { unsigned max = MaxML; + size_t const mostFrequent = HIST_countFast_wksp( + count, &max, mlCodeTable, nbSeq, entropyWorkspace, entropyWkspSize); /* can't fail */ + DEBUGLOG(5, "Building ML table (remaining space : %i)", (int)(oend-op)); + nextEntropy->fse.matchlength_repeatMode = prevEntropy->fse.matchlength_repeatMode; + MLtype = ZSTD_selectEncodingType(&nextEntropy->fse.matchlength_repeatMode, + count, max, mostFrequent, nbSeq, + MLFSELog, prevEntropy->fse.matchlengthCTable, + ML_defaultNorm, ML_defaultNormLog, + ZSTD_defaultAllowed, strategy); + assert(!(MLtype < set_compressed && nextEntropy->fse.matchlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */ + { size_t const countSize = ZSTD_buildCTable( + op, (size_t)(oend - op), + CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype, + count, max, mlCodeTable, nbSeq, + ML_defaultNorm, ML_defaultNormLog, MaxML, + prevEntropy->fse.matchlengthCTable, + sizeof(prevEntropy->fse.matchlengthCTable), + entropyWorkspace, entropyWkspSize); + FORWARD_IF_ERROR(countSize, "ZSTD_buildCTable for MatchLengths failed"); + if (MLtype == set_compressed) + lastNCount = op; op += countSize; + assert(op <= oend); } } *seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2)); - { size_t const streamSize = ZSTD_encodeSequences(op, oend - op, - CTable_MatchLength, mlCodeTable, - CTable_OffsetBits, ofCodeTable, - CTable_LitLength, llCodeTable, - sequences, nbSeq, longOffsets); - if (ZSTD_isError(streamSize)) return streamSize; - op += streamSize; + { size_t const bitstreamSize = ZSTD_encodeSequences( + op, (size_t)(oend - op), + CTable_MatchLength, mlCodeTable, + CTable_OffsetBits, ofCodeTable, + CTable_LitLength, llCodeTable, + sequences, nbSeq, + longOffsets, bmi2); + FORWARD_IF_ERROR(bitstreamSize, "ZSTD_encodeSequences failed"); + op += bitstreamSize; + assert(op <= oend); + /* zstd versions <= 1.3.4 mistakenly report corruption when + * FSE_readNCount() receives a buffer < 4 bytes. + * Fixed by https://github.com/facebook/zstd/pull/1146. + * This can happen when the last set_compressed table present is 2 + * bytes and the bitstream is only one byte. + * In this exceedingly rare case, we will simply emit an uncompressed + * block, since it isn't worth optimizing. + */ + if (lastNCount && (op - lastNCount) < 4) { + /* NCountSize >= 2 && bitstreamSize > 0 ==> lastCountSize == 3 */ + assert(op - lastNCount == 3); + DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.3.4 by " + "emitting an uncompressed block."); + return 0; + } } - return op - ostart; + DEBUGLOG(5, "compressed block size : %u", (unsigned)(op - ostart)); + return (size_t)(op - ostart); } -MEM_STATIC size_t ZSTD_compressSequences(seqStore_t* seqStorePtr, - ZSTD_entropyCTables_t* entropy, - ZSTD_compressionParameters const* cParams, - void* dst, size_t dstCapacity, - size_t srcSize) -{ - size_t const cSize = ZSTD_compressSequences_internal(seqStorePtr, entropy, cParams, - dst, dstCapacity); - size_t const minGain = ZSTD_minGain(srcSize); - size_t const maxCSize = srcSize - minGain; - /* If the srcSize <= dstCapacity, then there is enough space to write a - * raw uncompressed block. Since we ran out of space, the block must not - * be compressible, so fall back to a raw uncompressed block. +MEM_STATIC size_t +ZSTD_compressSequences(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_compressSequences_internal( + seqStorePtr, prevEntropy, nextEntropy, cctxParams, + dst, dstCapacity, + entropyWorkspace, entropyWkspSize, bmi2); + if (cSize == 0) return 0; + /* When srcSize <= dstCapacity, there is enough space to write a raw uncompressed block. + * Since we ran out of space, block must be not compressible, so fall back to raw uncompressed block. */ - int const uncompressibleError = cSize == ERROR(dstSize_tooSmall) && srcSize <= dstCapacity; + if ((cSize == ERROR(dstSize_tooSmall)) & (srcSize <= dstCapacity)) + return 0; /* block not compressed */ + FORWARD_IF_ERROR(cSize, "ZSTD_compressSequences_internal failed"); - if (ZSTD_isError(cSize) && !uncompressibleError) - return cSize; /* Check compressibility */ - if (cSize >= maxCSize || uncompressibleError) { - entropy->hufCTable_repeatMode = HUF_repeat_none; - entropy->offcode_repeatMode = FSE_repeat_none; - entropy->matchlength_repeatMode = FSE_repeat_none; - entropy->litlength_repeatMode = FSE_repeat_none; - return 0; + { size_t const maxCSize = srcSize - ZSTD_minGain(srcSize, cctxParams->cParams.strategy); + if (cSize >= maxCSize) return 0; /* block not compressed */ } - assert(!ZSTD_isError(cSize)); - /* confirm repcodes */ - { int i; for (i=0; i<ZSTD_REP_NUM; i++) seqStorePtr->rep[i] = seqStorePtr->repToConfirm[i]; } return cSize; } /* ZSTD_selectBlockCompressor() : * Not static, but internal use only (used by long distance matcher) * assumption : strat is a valid strategy */ -typedef size_t (*ZSTD_blockCompressor) (ZSTD_CCtx* ctx, const void* src, size_t srcSize); -ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, int extDict) +ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_dictMode_e dictMode) { - static const ZSTD_blockCompressor blockCompressor[2][(unsigned)ZSTD_btultra+1] = { + static const ZSTD_blockCompressor blockCompressor[3][ZSTD_STRATEGY_MAX+1] = { { ZSTD_compressBlock_fast /* default for 0 */, - ZSTD_compressBlock_fast, ZSTD_compressBlock_doubleFast, ZSTD_compressBlock_greedy, - ZSTD_compressBlock_lazy, ZSTD_compressBlock_lazy2, ZSTD_compressBlock_btlazy2, - ZSTD_compressBlock_btopt, ZSTD_compressBlock_btultra }, + ZSTD_compressBlock_fast, + ZSTD_compressBlock_doubleFast, + ZSTD_compressBlock_greedy, + ZSTD_compressBlock_lazy, + ZSTD_compressBlock_lazy2, + ZSTD_compressBlock_btlazy2, + ZSTD_compressBlock_btopt, + ZSTD_compressBlock_btultra, + ZSTD_compressBlock_btultra2 }, { ZSTD_compressBlock_fast_extDict /* default for 0 */, - ZSTD_compressBlock_fast_extDict, ZSTD_compressBlock_doubleFast_extDict, ZSTD_compressBlock_greedy_extDict, - ZSTD_compressBlock_lazy_extDict,ZSTD_compressBlock_lazy2_extDict, ZSTD_compressBlock_btlazy2_extDict, - ZSTD_compressBlock_btopt_extDict, ZSTD_compressBlock_btultra_extDict } + ZSTD_compressBlock_fast_extDict, + ZSTD_compressBlock_doubleFast_extDict, + ZSTD_compressBlock_greedy_extDict, + ZSTD_compressBlock_lazy_extDict, + ZSTD_compressBlock_lazy2_extDict, + ZSTD_compressBlock_btlazy2_extDict, + ZSTD_compressBlock_btopt_extDict, + ZSTD_compressBlock_btultra_extDict, + ZSTD_compressBlock_btultra_extDict }, + { ZSTD_compressBlock_fast_dictMatchState /* default for 0 */, + ZSTD_compressBlock_fast_dictMatchState, + ZSTD_compressBlock_doubleFast_dictMatchState, + ZSTD_compressBlock_greedy_dictMatchState, + ZSTD_compressBlock_lazy_dictMatchState, + ZSTD_compressBlock_lazy2_dictMatchState, + ZSTD_compressBlock_btlazy2_dictMatchState, + ZSTD_compressBlock_btopt_dictMatchState, + ZSTD_compressBlock_btultra_dictMatchState, + ZSTD_compressBlock_btultra_dictMatchState } }; + ZSTD_blockCompressor selectedCompressor; ZSTD_STATIC_ASSERT((unsigned)ZSTD_fast == 1); - assert((U32)strat >= (U32)ZSTD_fast); - assert((U32)strat <= (U32)ZSTD_btultra); - return blockCompressor[extDict!=0][(U32)strat]; + assert(ZSTD_cParam_withinBounds(ZSTD_c_strategy, strat)); + selectedCompressor = blockCompressor[(int)dictMode][(int)strat]; + assert(selectedCompressor != NULL); + return selectedCompressor; } static void ZSTD_storeLastLiterals(seqStore_t* seqStorePtr, @@ -1541,33 +2230,347 @@ static void ZSTD_storeLastLiterals(seqStore_t* seqStorePtr, seqStorePtr->lit += lastLLSize; } -static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc, void* dst, size_t dstCapacity, const void* src, size_t srcSize) +void ZSTD_resetSeqStore(seqStore_t* ssPtr) { - const BYTE* const base = zc->base; - const BYTE* const istart = (const BYTE*)src; - const U32 current = (U32)(istart-base); - size_t lastLLSize; - const BYTE* anchor; - U32 const extDict = zc->lowLimit < zc->dictLimit; - const ZSTD_blockCompressor blockCompressor = - zc->appliedParams.ldmParams.enableLdm - ? (extDict ? ZSTD_compressBlock_ldm_extDict : ZSTD_compressBlock_ldm) - : ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, extDict); + ssPtr->lit = ssPtr->litStart; + ssPtr->sequences = ssPtr->sequencesStart; + ssPtr->longLengthID = 0; +} + +typedef enum { ZSTDbss_compress, ZSTDbss_noCompress } ZSTD_buildSeqStore_e; - if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) return 0; /* don't even attempt compression below a certain srcSize */ +static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize) +{ + ZSTD_matchState_t* const ms = &zc->blockState.matchState; + DEBUGLOG(5, "ZSTD_buildSeqStore (srcSize=%zu)", srcSize); + assert(srcSize <= ZSTD_BLOCKSIZE_MAX); + /* Assert that we have correctly flushed the ctx params into the ms's copy */ + ZSTD_assertEqualCParams(zc->appliedParams.cParams, ms->cParams); + if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) { + ZSTD_ldm_skipSequences(&zc->externSeqStore, srcSize, zc->appliedParams.cParams.minMatch); + return ZSTDbss_noCompress; /* don't even attempt compression below a certain srcSize */ + } ZSTD_resetSeqStore(&(zc->seqStore)); - if (current > zc->nextToUpdate + 384) - zc->nextToUpdate = current - MIN(192, (U32)(current - zc->nextToUpdate - 384)); /* limited update after finding a very long match */ + /* required for optimal parser to read stats from dictionary */ + ms->opt.symbolCosts = &zc->blockState.prevCBlock->entropy; + /* tell the optimal parser how we expect to compress literals */ + ms->opt.literalCompressionMode = zc->appliedParams.literalCompressionMode; + /* a gap between an attached dict and the current window is not safe, + * they must remain adjacent, + * and when that stops being the case, the dict must be unset */ + assert(ms->dictMatchState == NULL || ms->loadedDictEnd == ms->window.dictLimit); + + /* limited update after a very long match */ + { const BYTE* const base = ms->window.base; + const BYTE* const istart = (const BYTE*)src; + const U32 current = (U32)(istart-base); + if (sizeof(ptrdiff_t)==8) assert(istart - base < (ptrdiff_t)(U32)(-1)); /* ensure no overflow */ + if (current > ms->nextToUpdate + 384) + ms->nextToUpdate = current - MIN(192, (U32)(current - ms->nextToUpdate - 384)); + } + + /* select and store sequences */ + { ZSTD_dictMode_e const dictMode = ZSTD_matchState_dictMode(ms); + size_t lastLLSize; + { int i; + for (i = 0; i < ZSTD_REP_NUM; ++i) + zc->blockState.nextCBlock->rep[i] = zc->blockState.prevCBlock->rep[i]; + } + if (zc->externSeqStore.pos < zc->externSeqStore.size) { + assert(!zc->appliedParams.ldmParams.enableLdm); + /* Updates ldmSeqStore.pos */ + lastLLSize = + ZSTD_ldm_blockCompress(&zc->externSeqStore, + ms, &zc->seqStore, + zc->blockState.nextCBlock->rep, + src, srcSize); + assert(zc->externSeqStore.pos <= zc->externSeqStore.size); + } else if (zc->appliedParams.ldmParams.enableLdm) { + rawSeqStore_t ldmSeqStore = {NULL, 0, 0, 0}; + + ldmSeqStore.seq = zc->ldmSequences; + ldmSeqStore.capacity = zc->maxNbLdmSequences; + /* Updates ldmSeqStore.size */ + FORWARD_IF_ERROR(ZSTD_ldm_generateSequences(&zc->ldmState, &ldmSeqStore, + &zc->appliedParams.ldmParams, + src, srcSize), ""); + /* Updates ldmSeqStore.pos */ + lastLLSize = + ZSTD_ldm_blockCompress(&ldmSeqStore, + ms, &zc->seqStore, + zc->blockState.nextCBlock->rep, + src, srcSize); + assert(ldmSeqStore.pos == ldmSeqStore.size); + } else { /* not long range mode */ + ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, dictMode); + lastLLSize = blockCompressor(ms, &zc->seqStore, zc->blockState.nextCBlock->rep, src, srcSize); + } + { const BYTE* const lastLiterals = (const BYTE*)src + srcSize - lastLLSize; + ZSTD_storeLastLiterals(&zc->seqStore, lastLiterals, lastLLSize); + } } + return ZSTDbss_compress; +} + +static void ZSTD_copyBlockSequences(ZSTD_CCtx* zc) +{ + const seqStore_t* seqStore = ZSTD_getSeqStore(zc); + const seqDef* seqs = seqStore->sequencesStart; + size_t seqsSize = seqStore->sequences - seqs; + + ZSTD_Sequence* outSeqs = &zc->seqCollector.seqStart[zc->seqCollector.seqIndex]; + size_t i; size_t position; int repIdx; + + assert(zc->seqCollector.seqIndex + 1 < zc->seqCollector.maxSequences); + for (i = 0, position = 0; i < seqsSize; ++i) { + outSeqs[i].offset = seqs[i].offset; + outSeqs[i].litLength = seqs[i].litLength; + outSeqs[i].matchLength = seqs[i].matchLength + MINMATCH; + + if (i == seqStore->longLengthPos) { + if (seqStore->longLengthID == 1) { + outSeqs[i].litLength += 0x10000; + } else if (seqStore->longLengthID == 2) { + outSeqs[i].matchLength += 0x10000; + } + } + + if (outSeqs[i].offset <= ZSTD_REP_NUM) { + outSeqs[i].rep = outSeqs[i].offset; + repIdx = (unsigned int)i - outSeqs[i].offset; + + if (outSeqs[i].litLength == 0) { + if (outSeqs[i].offset < 3) { + --repIdx; + } else { + repIdx = (unsigned int)i - 1; + } + ++outSeqs[i].rep; + } + assert(repIdx >= -3); + outSeqs[i].offset = repIdx >= 0 ? outSeqs[repIdx].offset : repStartValue[-repIdx - 1]; + if (outSeqs[i].rep == 4) { + --outSeqs[i].offset; + } + } else { + outSeqs[i].offset -= ZSTD_REP_NUM; + } + + position += outSeqs[i].litLength; + outSeqs[i].matchPos = (unsigned int)position; + position += outSeqs[i].matchLength; + } + zc->seqCollector.seqIndex += seqsSize; +} + +size_t ZSTD_getSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs, + size_t outSeqsSize, const void* src, size_t srcSize) +{ + const size_t dstCapacity = ZSTD_compressBound(srcSize); + void* dst = ZSTD_malloc(dstCapacity, ZSTD_defaultCMem); + SeqCollector seqCollector; + + RETURN_ERROR_IF(dst == NULL, memory_allocation, "NULL pointer!"); + + seqCollector.collectSequences = 1; + seqCollector.seqStart = outSeqs; + seqCollector.seqIndex = 0; + seqCollector.maxSequences = outSeqsSize; + zc->seqCollector = seqCollector; + + ZSTD_compress2(zc, dst, dstCapacity, src, srcSize); + ZSTD_free(dst, ZSTD_defaultCMem); + return zc->seqCollector.seqIndex; +} + +/* Returns true if the given block is a RLE block */ +static int ZSTD_isRLE(const BYTE *ip, size_t length) { + size_t i; + if (length < 2) return 1; + for (i = 1; i < length; ++i) { + if (ip[0] != ip[i]) return 0; + } + return 1; +} + +/* Returns true if the given block may be RLE. + * This is just a heuristic based on the compressibility. + * It may return both false positives and false negatives. + */ +static int ZSTD_maybeRLE(seqStore_t const* seqStore) +{ + size_t const nbSeqs = (size_t)(seqStore->sequences - seqStore->sequencesStart); + size_t const nbLits = (size_t)(seqStore->lit - seqStore->litStart); + + return nbSeqs < 4 && nbLits < 10; +} - lastLLSize = blockCompressor(zc, src, srcSize); +static void ZSTD_confirmRepcodesAndEntropyTables(ZSTD_CCtx* zc) +{ + ZSTD_compressedBlockState_t* const tmp = zc->blockState.prevCBlock; + zc->blockState.prevCBlock = zc->blockState.nextCBlock; + zc->blockState.nextCBlock = tmp; +} + +static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc, + void* dst, size_t dstCapacity, + const void* src, size_t srcSize, U32 frame) +{ + /* This the upper bound for the length of an rle block. + * This isn't the actual upper bound. Finding the real threshold + * needs further investigation. + */ + const U32 rleMaxLength = 25; + size_t cSize; + const BYTE* ip = (const BYTE*)src; + BYTE* op = (BYTE*)dst; + DEBUGLOG(5, "ZSTD_compressBlock_internal (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u)", + (unsigned)dstCapacity, (unsigned)zc->blockState.matchState.window.dictLimit, + (unsigned)zc->blockState.matchState.nextToUpdate); + + { const size_t bss = ZSTD_buildSeqStore(zc, src, srcSize); + FORWARD_IF_ERROR(bss, "ZSTD_buildSeqStore failed"); + if (bss == ZSTDbss_noCompress) { cSize = 0; goto out; } + } + + if (zc->seqCollector.collectSequences) { + ZSTD_copyBlockSequences(zc); + return 0; + } + + /* encode sequences and literals */ + cSize = ZSTD_compressSequences(&zc->seqStore, + &zc->blockState.prevCBlock->entropy, &zc->blockState.nextCBlock->entropy, + &zc->appliedParams, + dst, dstCapacity, + srcSize, + zc->entropyWorkspace, HUF_WORKSPACE_SIZE /* statically allocated in resetCCtx */, + zc->bmi2); + + if (frame && + /* We don't want to emit our first block as a RLE even if it qualifies because + * doing so will cause the decoder (cli only) to throw a "should consume all input error." + * This is only an issue for zstd <= v1.4.3 + */ + !zc->isFirstBlock && + cSize < rleMaxLength && + ZSTD_isRLE(ip, srcSize)) + { + cSize = 1; + op[0] = ip[0]; + } + +out: + if (!ZSTD_isError(cSize) && cSize > 1) { + ZSTD_confirmRepcodesAndEntropyTables(zc); + } + /* We check that dictionaries have offset codes available for the first + * block. After the first block, the offcode table might not have large + * enough codes to represent the offsets in the data. + */ + if (zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid) + zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check; + + return cSize; +} + +static size_t ZSTD_compressBlock_targetCBlockSize_body(ZSTD_CCtx* zc, + void* dst, size_t dstCapacity, + const void* src, size_t srcSize, + const size_t bss, U32 lastBlock) +{ + DEBUGLOG(6, "Attempting ZSTD_compressSuperBlock()"); + if (bss == ZSTDbss_compress) { + if (/* We don't want to emit our first block as a RLE even if it qualifies because + * doing so will cause the decoder (cli only) to throw a "should consume all input error." + * This is only an issue for zstd <= v1.4.3 + */ + !zc->isFirstBlock && + ZSTD_maybeRLE(&zc->seqStore) && + ZSTD_isRLE((BYTE const*)src, srcSize)) + { + return ZSTD_rleCompressBlock(dst, dstCapacity, *(BYTE const*)src, srcSize, lastBlock); + } + /* Attempt superblock compression. + * + * Note that compressed size of ZSTD_compressSuperBlock() is not bound by the + * standard ZSTD_compressBound(). This is a problem, because even if we have + * space now, taking an extra byte now could cause us to run out of space later + * and violate ZSTD_compressBound(). + * + * Define blockBound(blockSize) = blockSize + ZSTD_blockHeaderSize. + * + * In order to respect ZSTD_compressBound() we must attempt to emit a raw + * uncompressed block in these cases: + * * cSize == 0: Return code for an uncompressed block. + * * cSize == dstSize_tooSmall: We may have expanded beyond blockBound(srcSize). + * ZSTD_noCompressBlock() will return dstSize_tooSmall if we are really out of + * output space. + * * 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); + if (cSize != ERROR(dstSize_tooSmall)) { + 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_confirmRepcodesAndEntropyTables(zc); + return cSize; + } + } + } + } + + DEBUGLOG(6, "Resorting to ZSTD_noCompressBlock()"); + /* Superblock compression failed, attempt to emit a single no compress block. + * The decoder will be able to stream this block since it is uncompressed. + */ + return ZSTD_noCompressBlock(dst, dstCapacity, src, srcSize, lastBlock); +} + +static size_t ZSTD_compressBlock_targetCBlockSize(ZSTD_CCtx* zc, + void* dst, size_t dstCapacity, + const void* src, size_t srcSize, + U32 lastBlock) +{ + size_t cSize = 0; + const size_t bss = ZSTD_buildSeqStore(zc, src, srcSize); + DEBUGLOG(5, "ZSTD_compressBlock_targetCBlockSize (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u, srcSize=%zu)", + (unsigned)dstCapacity, (unsigned)zc->blockState.matchState.window.dictLimit, (unsigned)zc->blockState.matchState.nextToUpdate, srcSize); + FORWARD_IF_ERROR(bss, "ZSTD_buildSeqStore failed"); - /* Last literals */ - anchor = (const BYTE*)src + srcSize - lastLLSize; - ZSTD_storeLastLiterals(&zc->seqStore, anchor, lastLLSize); + cSize = ZSTD_compressBlock_targetCBlockSize_body(zc, dst, dstCapacity, src, srcSize, bss, lastBlock); + FORWARD_IF_ERROR(cSize, "ZSTD_compressBlock_targetCBlockSize_body failed"); - return ZSTD_compressSequences(&zc->seqStore, zc->entropy, &zc->appliedParams.cParams, dst, dstCapacity, srcSize); + if (zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid) + zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check; + + return cSize; } +static void ZSTD_overflowCorrectIfNeeded(ZSTD_matchState_t* ms, + ZSTD_cwksp* ws, + ZSTD_CCtx_params const* params, + void const* ip, + void const* iend) +{ + if (ZSTD_window_needOverflowCorrection(ms->window, iend)) { + U32 const maxDist = (U32)1 << params->cParams.windowLog; + U32 const cycleLog = ZSTD_cycleLog(params->cParams.chainLog, params->cParams.strategy); + U32 const correction = ZSTD_window_correctOverflow(&ms->window, cycleLog, maxDist, ip); + ZSTD_STATIC_ASSERT(ZSTD_CHAINLOG_MAX <= 30); + ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_32 <= 30); + ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX <= 31); + ZSTD_cwksp_mark_tables_dirty(ws); + ZSTD_reduceIndex(ms, params, correction); + ZSTD_cwksp_mark_tables_clean(ws); + if (ms->nextToUpdate < correction) ms->nextToUpdate = 0; + else ms->nextToUpdate -= correction; + /* invalidate dictionaries on overflow correction */ + ms->loadedDictEnd = 0; + ms->dictMatchState = NULL; + } +} /*! ZSTD_compress_frameChunk() : * Compress a chunk of data into one or multiple blocks. @@ -1586,88 +2589,96 @@ static size_t ZSTD_compress_frameChunk (ZSTD_CCtx* cctx, const BYTE* ip = (const BYTE*)src; BYTE* const ostart = (BYTE*)dst; BYTE* op = ostart; - U32 const maxDist = 1 << cctx->appliedParams.cParams.windowLog; + U32 const maxDist = (U32)1 << cctx->appliedParams.cParams.windowLog; + assert(cctx->appliedParams.cParams.windowLog <= ZSTD_WINDOWLOG_MAX); + + DEBUGLOG(5, "ZSTD_compress_frameChunk (blockSize=%u)", (unsigned)blockSize); if (cctx->appliedParams.fParams.checksumFlag && srcSize) XXH64_update(&cctx->xxhState, src, srcSize); while (remaining) { + ZSTD_matchState_t* const ms = &cctx->blockState.matchState; U32 const lastBlock = lastFrameChunk & (blockSize >= remaining); - size_t cSize; - if (dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE) - return ERROR(dstSize_tooSmall); /* not enough space to store compressed block */ + RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE, + dstSize_tooSmall, + "not enough space to store compressed block"); if (remaining < blockSize) blockSize = remaining; - /* preemptive overflow correction */ - if (cctx->lowLimit > (3U<<29)) { - U32 const cycleMask = (1 << ZSTD_cycleLog(cctx->appliedParams.cParams.hashLog, cctx->appliedParams.cParams.strategy)) - 1; - U32 const current = (U32)(ip - cctx->base); - U32 const newCurrent = (current & cycleMask) + (1 << cctx->appliedParams.cParams.windowLog); - U32 const correction = current - newCurrent; - ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_64 <= 30); - ZSTD_reduceIndex(cctx, correction); - cctx->base += correction; - cctx->dictBase += correction; - cctx->lowLimit -= correction; - cctx->dictLimit -= correction; - if (cctx->nextToUpdate < correction) cctx->nextToUpdate = 0; - else cctx->nextToUpdate -= correction; - } + ZSTD_overflowCorrectIfNeeded( + ms, &cctx->workspace, &cctx->appliedParams, ip, ip + blockSize); + ZSTD_checkDictValidity(&ms->window, ip + blockSize, maxDist, &ms->loadedDictEnd, &ms->dictMatchState); - if ((U32)(ip+blockSize - cctx->base) > cctx->loadedDictEnd + maxDist) { - /* enforce maxDist */ - U32 const newLowLimit = (U32)(ip+blockSize - cctx->base) - maxDist; - if (cctx->lowLimit < newLowLimit) cctx->lowLimit = newLowLimit; - if (cctx->dictLimit < cctx->lowLimit) cctx->dictLimit = cctx->lowLimit; - } + /* Ensure hash/chain table insertion resumes no sooner than lowlimit */ + if (ms->nextToUpdate < ms->window.lowLimit) ms->nextToUpdate = ms->window.lowLimit; - cSize = ZSTD_compressBlock_internal(cctx, op+ZSTD_blockHeaderSize, dstCapacity-ZSTD_blockHeaderSize, ip, blockSize); - if (ZSTD_isError(cSize)) return cSize; + { size_t cSize; + if (ZSTD_useTargetCBlockSize(&cctx->appliedParams)) { + cSize = ZSTD_compressBlock_targetCBlockSize(cctx, op, dstCapacity, ip, blockSize, lastBlock); + FORWARD_IF_ERROR(cSize, "ZSTD_compressBlock_targetCBlockSize failed"); + assert(cSize > 0); + assert(cSize <= blockSize + ZSTD_blockHeaderSize); + } else { + cSize = ZSTD_compressBlock_internal(cctx, + op+ZSTD_blockHeaderSize, dstCapacity-ZSTD_blockHeaderSize, + ip, blockSize, 1 /* frame */); + FORWARD_IF_ERROR(cSize, "ZSTD_compressBlock_internal failed"); + + if (cSize == 0) { /* block is not compressible */ + cSize = ZSTD_noCompressBlock(op, dstCapacity, ip, blockSize, lastBlock); + FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed"); + } else { + U32 const cBlockHeader = cSize == 1 ? + lastBlock + (((U32)bt_rle)<<1) + (U32)(blockSize << 3) : + lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3); + MEM_writeLE24(op, cBlockHeader); + cSize += ZSTD_blockHeaderSize; + } + } - if (cSize == 0) { /* block is not compressible */ - U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw)<<1) + (U32)(blockSize << 3); - if (blockSize + ZSTD_blockHeaderSize > dstCapacity) return ERROR(dstSize_tooSmall); - MEM_writeLE32(op, cBlockHeader24); /* no pb, 4th byte will be overwritten */ - memcpy(op + ZSTD_blockHeaderSize, ip, blockSize); - cSize = ZSTD_blockHeaderSize+blockSize; - } else { - U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3); - MEM_writeLE24(op, cBlockHeader24); - cSize += ZSTD_blockHeaderSize; - } - remaining -= blockSize; - dstCapacity -= cSize; - ip += blockSize; - op += cSize; - } + ip += blockSize; + assert(remaining >= blockSize); + remaining -= blockSize; + op += cSize; + assert(dstCapacity >= cSize); + dstCapacity -= cSize; + cctx->isFirstBlock = 0; + DEBUGLOG(5, "ZSTD_compress_frameChunk: adding a block of size %u", + (unsigned)cSize); + } } if (lastFrameChunk && (op>ostart)) cctx->stage = ZSTDcs_ending; - return op-ostart; + return (size_t)(op-ostart); } static size_t ZSTD_writeFrameHeader(void* dst, size_t dstCapacity, - ZSTD_CCtx_params params, U64 pledgedSrcSize, U32 dictID) + const ZSTD_CCtx_params* params, U64 pledgedSrcSize, U32 dictID) { BYTE* const op = (BYTE*)dst; U32 const dictIDSizeCodeLength = (dictID>0) + (dictID>=256) + (dictID>=65536); /* 0-3 */ - U32 const dictIDSizeCode = params.fParams.noDictIDFlag ? 0 : dictIDSizeCodeLength; /* 0-3 */ - U32 const checksumFlag = params.fParams.checksumFlag>0; - U32 const windowSize = 1U << params.cParams.windowLog; - U32 const singleSegment = params.fParams.contentSizeFlag && (windowSize >= pledgedSrcSize); - BYTE const windowLogByte = (BYTE)((params.cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3); - U32 const fcsCode = params.fParams.contentSizeFlag ? + U32 const dictIDSizeCode = params->fParams.noDictIDFlag ? 0 : dictIDSizeCodeLength; /* 0-3 */ + U32 const checksumFlag = params->fParams.checksumFlag>0; + U32 const windowSize = (U32)1 << params->cParams.windowLog; + U32 const singleSegment = params->fParams.contentSizeFlag && (windowSize >= pledgedSrcSize); + BYTE const windowLogByte = (BYTE)((params->cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3); + U32 const fcsCode = params->fParams.contentSizeFlag ? (pledgedSrcSize>=256) + (pledgedSrcSize>=65536+256) + (pledgedSrcSize>=0xFFFFFFFFU) : 0; /* 0-3 */ - BYTE const frameHeaderDecriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) ); - size_t pos; - - if (dstCapacity < ZSTD_frameHeaderSize_max) return ERROR(dstSize_tooSmall); - DEBUGLOG(5, "ZSTD_writeFrameHeader : dictIDFlag : %u ; dictID : %u ; dictIDSizeCode : %u", - !params.fParams.noDictIDFlag, dictID, dictIDSizeCode); - - MEM_writeLE32(dst, ZSTD_MAGICNUMBER); - op[4] = frameHeaderDecriptionByte; pos=5; + BYTE const frameHeaderDescriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) ); + size_t pos=0; + + assert(!(params->fParams.contentSizeFlag && pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN)); + RETURN_ERROR_IF(dstCapacity < ZSTD_FRAMEHEADERSIZE_MAX, dstSize_tooSmall, + "dst buf is too small to fit worst-case frame header size."); + DEBUGLOG(4, "ZSTD_writeFrameHeader : dictIDFlag : %u ; dictID : %u ; dictIDSizeCode : %u", + !params->fParams.noDictIDFlag, (unsigned)dictID, (unsigned)dictIDSizeCode); + + if (params->format == ZSTD_f_zstd1) { + MEM_writeLE32(dst, ZSTD_MAGICNUMBER); + pos = 4; + } + op[pos++] = frameHeaderDescriptionByte; if (!singleSegment) op[pos++] = windowLogByte; switch(dictIDSizeCode) { @@ -1688,129 +2699,189 @@ static size_t ZSTD_writeFrameHeader(void* dst, size_t dstCapacity, return pos; } +/* 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 (<ZSTD_blockHeaderSize) + */ +size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity) +{ + RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize, dstSize_tooSmall, + "dst buf is too small to write frame trailer empty block."); + { U32 const cBlockHeader24 = 1 /*lastBlock*/ + (((U32)bt_raw)<<1); /* 0 size */ + MEM_writeLE24(dst, cBlockHeader24); + return ZSTD_blockHeaderSize; + } +} + +size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq) +{ + RETURN_ERROR_IF(cctx->stage != ZSTDcs_init, stage_wrong, + "wrong cctx stage"); + RETURN_ERROR_IF(cctx->appliedParams.ldmParams.enableLdm, + parameter_unsupported, + "incompatible with ldm"); + cctx->externSeqStore.seq = seq; + cctx->externSeqStore.size = nbSeq; + cctx->externSeqStore.capacity = nbSeq; + cctx->externSeqStore.pos = 0; + return 0; +} + static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize, U32 frame, U32 lastFrameChunk) { - const BYTE* const ip = (const BYTE*) src; + ZSTD_matchState_t* const ms = &cctx->blockState.matchState; size_t fhSize = 0; - DEBUGLOG(5, "ZSTD_compressContinue_internal"); - DEBUGLOG(5, "stage: %u", cctx->stage); - if (cctx->stage==ZSTDcs_created) return ERROR(stage_wrong); /* missing init (ZSTD_compressBegin) */ + DEBUGLOG(5, "ZSTD_compressContinue_internal, stage: %u, srcSize: %u", + cctx->stage, (unsigned)srcSize); + RETURN_ERROR_IF(cctx->stage==ZSTDcs_created, stage_wrong, + "missing init (ZSTD_compressBegin)"); if (frame && (cctx->stage==ZSTDcs_init)) { - fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams, + fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, &cctx->appliedParams, cctx->pledgedSrcSizePlusOne-1, cctx->dictID); - if (ZSTD_isError(fhSize)) return fhSize; + FORWARD_IF_ERROR(fhSize, "ZSTD_writeFrameHeader failed"); + assert(fhSize <= dstCapacity); dstCapacity -= fhSize; dst = (char*)dst + fhSize; cctx->stage = ZSTDcs_ongoing; } - /* Check if blocks follow each other */ - if (src != cctx->nextSrc) { - /* not contiguous */ - ptrdiff_t const delta = cctx->nextSrc - ip; - cctx->lowLimit = cctx->dictLimit; - cctx->dictLimit = (U32)(cctx->nextSrc - cctx->base); - cctx->dictBase = cctx->base; - cctx->base -= delta; - cctx->nextToUpdate = cctx->dictLimit; - if (cctx->dictLimit - cctx->lowLimit < HASH_READ_SIZE) cctx->lowLimit = cctx->dictLimit; /* too small extDict */ - } + if (!srcSize) return fhSize; /* do not generate an empty block if no input */ - /* if input and dictionary overlap : reduce dictionary (area presumed modified by input) */ - if ((ip+srcSize > cctx->dictBase + cctx->lowLimit) & (ip < cctx->dictBase + cctx->dictLimit)) { - ptrdiff_t const highInputIdx = (ip + srcSize) - cctx->dictBase; - U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)cctx->dictLimit) ? cctx->dictLimit : (U32)highInputIdx; - cctx->lowLimit = lowLimitMax; + if (!ZSTD_window_update(&ms->window, src, srcSize)) { + ms->nextToUpdate = ms->window.dictLimit; + } + if (cctx->appliedParams.ldmParams.enableLdm) { + ZSTD_window_update(&cctx->ldmState.window, src, srcSize); } - cctx->nextSrc = ip + srcSize; + if (!frame) { + /* overflow check and correction for block mode */ + ZSTD_overflowCorrectIfNeeded( + ms, &cctx->workspace, &cctx->appliedParams, + src, (BYTE const*)src + srcSize); + } - if (srcSize) { - size_t const cSize = frame ? + DEBUGLOG(5, "ZSTD_compressContinue_internal (blockSize=%u)", (unsigned)cctx->blockSize); + { size_t const cSize = frame ? ZSTD_compress_frameChunk (cctx, dst, dstCapacity, src, srcSize, lastFrameChunk) : - ZSTD_compressBlock_internal (cctx, dst, dstCapacity, src, srcSize); - if (ZSTD_isError(cSize)) return cSize; + ZSTD_compressBlock_internal (cctx, dst, dstCapacity, src, srcSize, 0 /* frame */); + FORWARD_IF_ERROR(cSize, "%s", frame ? "ZSTD_compress_frameChunk failed" : "ZSTD_compressBlock_internal failed"); cctx->consumedSrcSize += srcSize; + cctx->producedCSize += (cSize + fhSize); + assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0)); + if (cctx->pledgedSrcSizePlusOne != 0) { /* control src size */ + ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1); + RETURN_ERROR_IF( + cctx->consumedSrcSize+1 > cctx->pledgedSrcSizePlusOne, + srcSize_wrong, + "error : pledgedSrcSize = %u, while realSrcSize >= %u", + (unsigned)cctx->pledgedSrcSizePlusOne-1, + (unsigned)cctx->consumedSrcSize); + } return cSize + fhSize; - } else - return fhSize; + } } size_t ZSTD_compressContinue (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize) { + DEBUGLOG(5, "ZSTD_compressContinue (srcSize=%u)", (unsigned)srcSize); return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1 /* frame mode */, 0 /* last chunk */); } size_t ZSTD_getBlockSize(const ZSTD_CCtx* cctx) { - ZSTD_compressionParameters const cParams = - ZSTD_getCParamsFromCCtxParams(cctx->appliedParams, 0, 0); - return MIN (ZSTD_BLOCKSIZE_MAX, 1 << cParams.windowLog); + ZSTD_compressionParameters const cParams = cctx->appliedParams.cParams; + assert(!ZSTD_checkCParams(cParams)); + return MIN (ZSTD_BLOCKSIZE_MAX, (U32)1 << cParams.windowLog); } size_t ZSTD_compressBlock(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize) { - size_t const blockSizeMax = ZSTD_getBlockSize(cctx); - if (srcSize > blockSizeMax) return ERROR(srcSize_wrong); + DEBUGLOG(5, "ZSTD_compressBlock: srcSize = %u", (unsigned)srcSize); + { size_t const blockSizeMax = ZSTD_getBlockSize(cctx); + RETURN_ERROR_IF(srcSize > blockSizeMax, srcSize_wrong, "input is larger than a block"); } + return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 0 /* frame mode */, 0 /* last chunk */); } /*! ZSTD_loadDictionaryContent() : * @return : 0, or an error code */ -static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t srcSize) +static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms, + ldmState_t* ls, + ZSTD_cwksp* ws, + ZSTD_CCtx_params const* params, + const void* src, size_t srcSize, + ZSTD_dictTableLoadMethod_e dtlm) { - const BYTE* const ip = (const BYTE*) src; + const BYTE* ip = (const BYTE*) src; const BYTE* const iend = ip + srcSize; - /* input becomes current prefix */ - zc->lowLimit = zc->dictLimit; - zc->dictLimit = (U32)(zc->nextSrc - zc->base); - zc->dictBase = zc->base; - zc->base += ip - zc->nextSrc; - zc->nextToUpdate = zc->dictLimit; - zc->loadedDictEnd = zc->appliedParams.forceWindow ? 0 : (U32)(iend - zc->base); + ZSTD_window_update(&ms->window, src, srcSize); + ms->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ms->window.base); + + if (params->ldmParams.enableLdm && ls != NULL) { + ZSTD_window_update(&ls->window, src, srcSize); + ls->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ls->window.base); + } + + /* Assert that we the ms params match the params we're being given */ + ZSTD_assertEqualCParams(params->cParams, ms->cParams); - zc->nextSrc = iend; if (srcSize <= HASH_READ_SIZE) return 0; - switch(zc->appliedParams.cParams.strategy) - { - case ZSTD_fast: - ZSTD_fillHashTable (zc, iend, zc->appliedParams.cParams.searchLength); - break; - case ZSTD_dfast: - ZSTD_fillDoubleHashTable (zc, iend, zc->appliedParams.cParams.searchLength); - break; + while (iend - ip > HASH_READ_SIZE) { + size_t const remaining = (size_t)(iend - ip); + size_t const chunk = MIN(remaining, ZSTD_CHUNKSIZE_MAX); + const BYTE* const ichunk = ip + chunk; - case ZSTD_greedy: - case ZSTD_lazy: - case ZSTD_lazy2: - if (srcSize >= HASH_READ_SIZE) - ZSTD_insertAndFindFirstIndex(zc, iend-HASH_READ_SIZE, zc->appliedParams.cParams.searchLength); - break; + ZSTD_overflowCorrectIfNeeded(ms, ws, params, ip, ichunk); - case ZSTD_btlazy2: - case ZSTD_btopt: - case ZSTD_btultra: - if (srcSize >= HASH_READ_SIZE) - ZSTD_updateTree(zc, iend-HASH_READ_SIZE, iend, 1 << zc->appliedParams.cParams.searchLog, zc->appliedParams.cParams.searchLength); - break; + if (params->ldmParams.enableLdm && ls != NULL) + ZSTD_ldm_fillHashTable(ls, (const BYTE*)src, (const BYTE*)src + srcSize, ¶ms->ldmParams); - default: - assert(0); /* not possible : not a valid strategy id */ + switch(params->cParams.strategy) + { + case ZSTD_fast: + ZSTD_fillHashTable(ms, ichunk, dtlm); + break; + case ZSTD_dfast: + ZSTD_fillDoubleHashTable(ms, ichunk, dtlm); + break; + + case ZSTD_greedy: + case ZSTD_lazy: + case ZSTD_lazy2: + if (chunk >= HASH_READ_SIZE) + ZSTD_insertAndFindFirstIndex(ms, ichunk-HASH_READ_SIZE); + break; + + case ZSTD_btlazy2: /* we want the dictionary table fully sorted */ + case ZSTD_btopt: + case ZSTD_btultra: + case ZSTD_btultra2: + if (chunk >= HASH_READ_SIZE) + ZSTD_updateTree(ms, ichunk-HASH_READ_SIZE, ichunk); + break; + + default: + assert(0); /* not possible : not a valid strategy id */ + } + + ip = ichunk; } - zc->nextToUpdate = (U32)(iend - zc->base); + ms->nextToUpdate = (U32)(iend - ms->window.base); return 0; } @@ -1821,81 +2892,123 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t NOTE: This behavior is not standard and could be improved in the future. */ static size_t ZSTD_checkDictNCount(short* normalizedCounter, unsigned dictMaxSymbolValue, unsigned maxSymbolValue) { U32 s; - if (dictMaxSymbolValue < maxSymbolValue) return ERROR(dictionary_corrupted); + RETURN_ERROR_IF(dictMaxSymbolValue < maxSymbolValue, dictionary_corrupted, "dict fse tables don't have all symbols"); for (s = 0; s <= maxSymbolValue; ++s) { - if (normalizedCounter[s] == 0) return ERROR(dictionary_corrupted); + RETURN_ERROR_IF(normalizedCounter[s] == 0, dictionary_corrupted, "dict fse tables don't have all symbols"); } return 0; } - -/* Dictionary format : - * See : - * https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#dictionary-format - */ -/*! ZSTD_loadZstdDictionary() : - * @return : 0, or an error code - * assumptions : magic number supposed already checked - * dictSize supposed > 8 - */ -static size_t ZSTD_loadZstdDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize) +size_t ZSTD_loadCEntropy(ZSTD_compressedBlockState_t* bs, void* workspace, + short* offcodeNCount, unsigned* offcodeMaxValue, + const void* const dict, size_t dictSize) { - const BYTE* dictPtr = (const BYTE*)dict; + const BYTE* dictPtr = (const BYTE*)dict; /* skip magic num and dict ID */ const BYTE* const dictEnd = dictPtr + dictSize; - short offcodeNCount[MaxOff+1]; - unsigned offcodeMaxValue = MaxOff; + dictPtr += 8; + bs->entropy.huf.repeatMode = HUF_repeat_check; - ZSTD_STATIC_ASSERT(sizeof(cctx->entropy->workspace) >= (1<<MAX(MLFSELog,LLFSELog))); + { unsigned maxSymbolValue = 255; + unsigned hasZeroWeights = 1; + size_t const hufHeaderSize = HUF_readCTable((HUF_CElt*)bs->entropy.huf.CTable, &maxSymbolValue, dictPtr, + dictEnd-dictPtr, &hasZeroWeights); - dictPtr += 4; /* skip magic number */ - cctx->dictID = cctx->appliedParams.fParams.noDictIDFlag ? 0 : MEM_readLE32(dictPtr); - dictPtr += 4; + /* We only set the loaded table as valid if it contains all non-zero + * weights. Otherwise, we set it to check */ + if (!hasZeroWeights) + bs->entropy.huf.repeatMode = HUF_repeat_valid; - { size_t const hufHeaderSize = HUF_readCTable((HUF_CElt*)cctx->entropy->hufCTable, 255, dictPtr, dictEnd-dictPtr); - if (HUF_isError(hufHeaderSize)) return ERROR(dictionary_corrupted); + RETURN_ERROR_IF(HUF_isError(hufHeaderSize), dictionary_corrupted, ""); + RETURN_ERROR_IF(maxSymbolValue < 255, dictionary_corrupted, ""); dictPtr += hufHeaderSize; } { unsigned offcodeLog; - size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr); - if (FSE_isError(offcodeHeaderSize)) return ERROR(dictionary_corrupted); - if (offcodeLog > OffFSELog) return ERROR(dictionary_corrupted); + size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr); + RETURN_ERROR_IF(FSE_isError(offcodeHeaderSize), dictionary_corrupted, ""); + RETURN_ERROR_IF(offcodeLog > OffFSELog, dictionary_corrupted, ""); /* Defer checking offcodeMaxValue because we need to know the size of the dictionary content */ - CHECK_E( FSE_buildCTable_wksp(cctx->entropy->offcodeCTable, offcodeNCount, offcodeMaxValue, offcodeLog, cctx->entropy->workspace, sizeof(cctx->entropy->workspace)), - dictionary_corrupted); + /* fill all offset symbols to avoid garbage at end of table */ + RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp( + bs->entropy.fse.offcodeCTable, + offcodeNCount, MaxOff, offcodeLog, + workspace, HUF_WORKSPACE_SIZE)), + dictionary_corrupted, ""); dictPtr += offcodeHeaderSize; } { short matchlengthNCount[MaxML+1]; unsigned matchlengthMaxValue = MaxML, matchlengthLog; size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr); - if (FSE_isError(matchlengthHeaderSize)) return ERROR(dictionary_corrupted); - if (matchlengthLog > MLFSELog) return ERROR(dictionary_corrupted); + RETURN_ERROR_IF(FSE_isError(matchlengthHeaderSize), dictionary_corrupted, ""); + RETURN_ERROR_IF(matchlengthLog > MLFSELog, dictionary_corrupted, ""); /* Every match length code must have non-zero probability */ - CHECK_F( ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML)); - CHECK_E( FSE_buildCTable_wksp(cctx->entropy->matchlengthCTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog, cctx->entropy->workspace, sizeof(cctx->entropy->workspace)), - dictionary_corrupted); + FORWARD_IF_ERROR( ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML), ""); + RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp( + bs->entropy.fse.matchlengthCTable, + matchlengthNCount, matchlengthMaxValue, matchlengthLog, + workspace, HUF_WORKSPACE_SIZE)), + dictionary_corrupted, ""); dictPtr += matchlengthHeaderSize; } { short litlengthNCount[MaxLL+1]; unsigned litlengthMaxValue = MaxLL, litlengthLog; size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr); - if (FSE_isError(litlengthHeaderSize)) return ERROR(dictionary_corrupted); - if (litlengthLog > LLFSELog) return ERROR(dictionary_corrupted); + RETURN_ERROR_IF(FSE_isError(litlengthHeaderSize), dictionary_corrupted, ""); + RETURN_ERROR_IF(litlengthLog > LLFSELog, dictionary_corrupted, ""); /* Every literal length code must have non-zero probability */ - CHECK_F( ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL)); - CHECK_E( FSE_buildCTable_wksp(cctx->entropy->litlengthCTable, litlengthNCount, litlengthMaxValue, litlengthLog, cctx->entropy->workspace, sizeof(cctx->entropy->workspace)), - dictionary_corrupted); + FORWARD_IF_ERROR( ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL), ""); + RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp( + bs->entropy.fse.litlengthCTable, + litlengthNCount, litlengthMaxValue, litlengthLog, + workspace, HUF_WORKSPACE_SIZE)), + dictionary_corrupted, ""); dictPtr += litlengthHeaderSize; } - if (dictPtr+12 > dictEnd) return ERROR(dictionary_corrupted); - cctx->seqStore.rep[0] = MEM_readLE32(dictPtr+0); - cctx->seqStore.rep[1] = MEM_readLE32(dictPtr+4); - cctx->seqStore.rep[2] = MEM_readLE32(dictPtr+8); + RETURN_ERROR_IF(dictPtr+12 > dictEnd, dictionary_corrupted, ""); + bs->rep[0] = MEM_readLE32(dictPtr+0); + bs->rep[1] = MEM_readLE32(dictPtr+4); + bs->rep[2] = MEM_readLE32(dictPtr+8); dictPtr += 12; + return dictPtr - (const BYTE*)dict; +} + +/* Dictionary format : + * See : + * https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#dictionary-format + */ +/*! ZSTD_loadZstdDictionary() : + * @return : dictID, or an error code + * assumptions : magic number supposed already checked + * dictSize supposed >= 8 + */ +static size_t ZSTD_loadZstdDictionary(ZSTD_compressedBlockState_t* bs, + ZSTD_matchState_t* ms, + ZSTD_cwksp* ws, + ZSTD_CCtx_params const* params, + const void* dict, size_t dictSize, + ZSTD_dictTableLoadMethod_e dtlm, + void* workspace) +{ + const BYTE* dictPtr = (const BYTE*)dict; + const BYTE* const dictEnd = dictPtr + dictSize; + short offcodeNCount[MaxOff+1]; + unsigned offcodeMaxValue = MaxOff; + size_t dictID; + size_t eSize; + + ZSTD_STATIC_ASSERT(HUF_WORKSPACE_SIZE >= (1<<MAX(MLFSELog,LLFSELog))); + assert(dictSize >= 8); + assert(MEM_readLE32(dictPtr) == ZSTD_MAGIC_DICTIONARY); + + dictID = params->fParams.noDictIDFlag ? 0 : MEM_readLE32(dictPtr + 4 /* skip magic number */ ); + eSize = ZSTD_loadCEntropy(bs, workspace, offcodeNCount, &offcodeMaxValue, dict, dictSize); + FORWARD_IF_ERROR(eSize, "ZSTD_loadCEntropy failed"); + dictPtr += eSize; + { size_t const dictContentSize = (size_t)(dictEnd - dictPtr); U32 offcodeMax = MaxOff; if (dictContentSize <= ((U32)-1) - 128 KB) { @@ -1903,86 +3016,123 @@ static size_t ZSTD_loadZstdDictionary(ZSTD_CCtx* cctx, const void* dict, size_t offcodeMax = ZSTD_highbit32(maxOffset); /* Calculate minimum offset code required to represent maxOffset */ } /* All offset values <= dictContentSize + 128 KB must be representable */ - CHECK_F (ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff))); + FORWARD_IF_ERROR(ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff)), ""); /* All repCodes must be <= dictContentSize and != 0*/ { U32 u; for (u=0; u<3; u++) { - if (cctx->seqStore.rep[u] == 0) return ERROR(dictionary_corrupted); - if (cctx->seqStore.rep[u] > dictContentSize) return ERROR(dictionary_corrupted); + RETURN_ERROR_IF(bs->rep[u] == 0, dictionary_corrupted, ""); + RETURN_ERROR_IF(bs->rep[u] > dictContentSize, dictionary_corrupted, ""); } } - cctx->entropy->hufCTable_repeatMode = HUF_repeat_valid; - cctx->entropy->offcode_repeatMode = FSE_repeat_valid; - cctx->entropy->matchlength_repeatMode = FSE_repeat_valid; - cctx->entropy->litlength_repeatMode = FSE_repeat_valid; - return ZSTD_loadDictionaryContent(cctx, dictPtr, dictContentSize); + bs->entropy.fse.offcode_repeatMode = FSE_repeat_valid; + bs->entropy.fse.matchlength_repeatMode = FSE_repeat_valid; + bs->entropy.fse.litlength_repeatMode = FSE_repeat_valid; + FORWARD_IF_ERROR(ZSTD_loadDictionaryContent( + ms, NULL, ws, params, dictPtr, dictContentSize, dtlm), ""); + return dictID; } } /** ZSTD_compress_insertDictionary() : -* @return : 0, or an error code */ -static size_t ZSTD_compress_insertDictionary(ZSTD_CCtx* cctx, - const void* dict, size_t dictSize, - ZSTD_dictMode_e dictMode) +* @return : dictID, or an error code */ +static size_t +ZSTD_compress_insertDictionary(ZSTD_compressedBlockState_t* bs, + ZSTD_matchState_t* ms, + ldmState_t* ls, + ZSTD_cwksp* ws, + const ZSTD_CCtx_params* params, + const void* dict, size_t dictSize, + ZSTD_dictContentType_e dictContentType, + ZSTD_dictTableLoadMethod_e dtlm, + void* workspace) { - DEBUGLOG(5, "ZSTD_compress_insertDictionary"); - if ((dict==NULL) || (dictSize<=8)) return 0; + DEBUGLOG(4, "ZSTD_compress_insertDictionary (dictSize=%u)", (U32)dictSize); + if ((dict==NULL) || (dictSize<8)) { + RETURN_ERROR_IF(dictContentType == ZSTD_dct_fullDict, dictionary_wrong, ""); + return 0; + } + + ZSTD_reset_compressedBlockState(bs); /* dict restricted modes */ - if (dictMode==ZSTD_dm_rawContent) - return ZSTD_loadDictionaryContent(cctx, dict, dictSize); + if (dictContentType == ZSTD_dct_rawContent) + return ZSTD_loadDictionaryContent(ms, ls, ws, params, dict, dictSize, dtlm); if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) { - if (dictMode == ZSTD_dm_auto) { - DEBUGLOG(5, "raw content dictionary detected"); - return ZSTD_loadDictionaryContent(cctx, dict, dictSize); + if (dictContentType == ZSTD_dct_auto) { + DEBUGLOG(4, "raw content dictionary detected"); + return ZSTD_loadDictionaryContent( + ms, ls, ws, params, dict, dictSize, dtlm); } - if (dictMode == ZSTD_dm_fullDict) - return ERROR(dictionary_wrong); + RETURN_ERROR_IF(dictContentType == ZSTD_dct_fullDict, dictionary_wrong, ""); assert(0); /* impossible */ } /* dict as full zstd dictionary */ - return ZSTD_loadZstdDictionary(cctx, dict, dictSize); + return ZSTD_loadZstdDictionary( + bs, ms, ws, params, dict, dictSize, dtlm, workspace); } +#define ZSTD_USE_CDICT_PARAMS_SRCSIZE_CUTOFF (128 KB) +#define ZSTD_USE_CDICT_PARAMS_DICTSIZE_MULTIPLIER (6) + /*! ZSTD_compressBegin_internal() : * @return : 0, or an error code */ static size_t ZSTD_compressBegin_internal(ZSTD_CCtx* cctx, - const void* dict, size_t dictSize, - ZSTD_dictMode_e dictMode, - const ZSTD_CDict* cdict, - ZSTD_CCtx_params params, U64 pledgedSrcSize, - ZSTD_buffered_policy_e zbuff) -{ - DEBUGLOG(4, "ZSTD_compressBegin_internal"); - DEBUGLOG(4, "dict ? %s", dict ? "dict" : (cdict ? "cdict" : "none")); - DEBUGLOG(4, "dictMode : %u", (U32)dictMode); + const void* dict, size_t dictSize, + ZSTD_dictContentType_e dictContentType, + ZSTD_dictTableLoadMethod_e dtlm, + const ZSTD_CDict* cdict, + const ZSTD_CCtx_params* params, U64 pledgedSrcSize, + ZSTD_buffered_policy_e zbuff) +{ + DEBUGLOG(4, "ZSTD_compressBegin_internal: wlog=%u", params->cParams.windowLog); /* params are supposed to be fully validated at this point */ - assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams))); + assert(!ZSTD_isError(ZSTD_checkCParams(params->cParams))); assert(!((dict) && (cdict))); /* either dict or cdict, not both */ - - if (cdict && cdict->dictContentSize>0) { - return ZSTD_copyCCtx_internal(cctx, cdict->refContext, - params.fParams, pledgedSrcSize, - zbuff); + if ( (cdict) + && (cdict->dictContentSize > 0) + && ( pledgedSrcSize < ZSTD_USE_CDICT_PARAMS_SRCSIZE_CUTOFF + || pledgedSrcSize < cdict->dictContentSize * ZSTD_USE_CDICT_PARAMS_DICTSIZE_MULTIPLIER + || pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN + || cdict->compressionLevel == 0) + && (params->attachDictPref != ZSTD_dictForceLoad) ) { + return ZSTD_resetCCtx_usingCDict(cctx, cdict, params, pledgedSrcSize, zbuff); } - CHECK_F( ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize, - ZSTDcrp_continue, zbuff) ); - return ZSTD_compress_insertDictionary(cctx, dict, dictSize, dictMode); + FORWARD_IF_ERROR( ZSTD_resetCCtx_internal(cctx, *params, pledgedSrcSize, + ZSTDcrp_makeClean, zbuff) , ""); + { size_t const dictID = cdict ? + ZSTD_compress_insertDictionary( + cctx->blockState.prevCBlock, &cctx->blockState.matchState, + &cctx->ldmState, &cctx->workspace, &cctx->appliedParams, cdict->dictContent, + cdict->dictContentSize, dictContentType, dtlm, + cctx->entropyWorkspace) + : ZSTD_compress_insertDictionary( + cctx->blockState.prevCBlock, &cctx->blockState.matchState, + &cctx->ldmState, &cctx->workspace, &cctx->appliedParams, dict, dictSize, + dictContentType, dtlm, cctx->entropyWorkspace); + FORWARD_IF_ERROR(dictID, "ZSTD_compress_insertDictionary failed"); + assert(dictID <= UINT_MAX); + cctx->dictID = (U32)dictID; + } + return 0; } -size_t ZSTD_compressBegin_advanced_internal( - ZSTD_CCtx* cctx, +size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, - ZSTD_dictMode_e dictMode, - ZSTD_CCtx_params params, + ZSTD_dictContentType_e dictContentType, + ZSTD_dictTableLoadMethod_e dtlm, + const ZSTD_CDict* cdict, + const ZSTD_CCtx_params* params, unsigned long long pledgedSrcSize) { + DEBUGLOG(4, "ZSTD_compressBegin_advanced_internal: wlog=%u", params->cParams.windowLog); /* compression parameters verification and optimization */ - CHECK_F( ZSTD_checkCParams(params.cParams) ); - return ZSTD_compressBegin_internal(cctx, dict, dictSize, dictMode, NULL, + FORWARD_IF_ERROR( ZSTD_checkCParams(params->cParams) , ""); + return ZSTD_compressBegin_internal(cctx, + dict, dictSize, dictContentType, dtlm, + cdict, params, pledgedSrcSize, ZSTDb_not_buffered); } @@ -1994,19 +3144,21 @@ size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx, ZSTD_parameters params, unsigned long long pledgedSrcSize) { ZSTD_CCtx_params const cctxParams = - ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params); - return ZSTD_compressBegin_advanced_internal(cctx, dict, dictSize, ZSTD_dm_auto, - cctxParams, - pledgedSrcSize); + ZSTD_assignParamsToCCtxParams(&cctx->requestedParams, ¶ms); + return ZSTD_compressBegin_advanced_internal(cctx, + dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, + NULL /*cdict*/, + &cctxParams, pledgedSrcSize); } size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel) { - ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, dictSize); + ZSTD_parameters const params = ZSTD_getParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize); ZSTD_CCtx_params const cctxParams = - ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params); - return ZSTD_compressBegin_internal(cctx, dict, dictSize, ZSTD_dm_auto, NULL, - cctxParams, 0, ZSTDb_not_buffered); + ZSTD_assignParamsToCCtxParams(&cctx->requestedParams, ¶ms); + DEBUGLOG(4, "ZSTD_compressBegin_usingDict (dictSize=%u)", (unsigned)dictSize); + return ZSTD_compressBegin_internal(cctx, dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL, + &cctxParams, ZSTD_CONTENTSIZE_UNKNOWN, ZSTDb_not_buffered); } size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel) @@ -2024,13 +3176,13 @@ static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity) BYTE* op = ostart; size_t fhSize = 0; - DEBUGLOG(5, "ZSTD_writeEpilogue"); - if (cctx->stage == ZSTDcs_created) return ERROR(stage_wrong); /* init missing */ + DEBUGLOG(4, "ZSTD_writeEpilogue"); + RETURN_ERROR_IF(cctx->stage == ZSTDcs_created, stage_wrong, "init missing"); /* special case : empty frame */ if (cctx->stage == ZSTDcs_init) { - fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams, 0, 0); - if (ZSTD_isError(fhSize)) return fhSize; + fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, &cctx->appliedParams, 0, 0); + FORWARD_IF_ERROR(fhSize, "ZSTD_writeFrameHeader failed"); dstCapacity -= fhSize; op += fhSize; cctx->stage = ZSTDcs_ongoing; @@ -2039,7 +3191,7 @@ static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity) if (cctx->stage != ZSTDcs_ending) { /* write one last empty block, make it the "last" block */ U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1) + 0; - if (dstCapacity<4) return ERROR(dstSize_tooSmall); + RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall, "no room for epilogue"); MEM_writeLE32(op, cBlockHeader24); op += ZSTD_blockHeaderSize; dstCapacity -= ZSTD_blockHeaderSize; @@ -2047,7 +3199,8 @@ static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity) if (cctx->appliedParams.fParams.checksumFlag) { U32 const checksum = (U32) XXH64_digest(&cctx->xxhState); - if (dstCapacity<4) return ERROR(dstSize_tooSmall); + RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall, "no room for checksum"); + DEBUGLOG(4, "ZSTD_writeEpilogue: write checksum : %08X", (unsigned)checksum); MEM_writeLE32(op, checksum); op += 4; } @@ -2056,7 +3209,6 @@ static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity) return op-ostart; } - size_t ZSTD_compressEnd (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize) @@ -2065,43 +3217,53 @@ size_t ZSTD_compressEnd (ZSTD_CCtx* cctx, size_t const cSize = ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1 /* frame mode */, 1 /* last chunk */); - if (ZSTD_isError(cSize)) return cSize; + FORWARD_IF_ERROR(cSize, "ZSTD_compressContinue_internal failed"); endResult = ZSTD_writeEpilogue(cctx, (char*)dst + cSize, dstCapacity-cSize); - if (ZSTD_isError(endResult)) return endResult; - if (cctx->appliedParams.fParams.contentSizeFlag) { /* control src size */ - DEBUGLOG(5, "end of frame : controlling src size"); - if (cctx->pledgedSrcSizePlusOne != cctx->consumedSrcSize+1) { - DEBUGLOG(5, "error : pledgedSrcSize = %u, while realSrcSize = %u", - (U32)cctx->pledgedSrcSizePlusOne-1, (U32)cctx->consumedSrcSize); - return ERROR(srcSize_wrong); - } } + FORWARD_IF_ERROR(endResult, "ZSTD_writeEpilogue failed"); + assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0)); + if (cctx->pledgedSrcSizePlusOne != 0) { /* control src size */ + ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1); + DEBUGLOG(4, "end of frame : controlling src size"); + RETURN_ERROR_IF( + cctx->pledgedSrcSizePlusOne != cctx->consumedSrcSize+1, + srcSize_wrong, + "error : pledgedSrcSize = %u, while realSrcSize = %u", + (unsigned)cctx->pledgedSrcSizePlusOne-1, + (unsigned)cctx->consumedSrcSize); + } return cSize + endResult; } static size_t ZSTD_compress_internal (ZSTD_CCtx* cctx, - void* dst, size_t dstCapacity, - const void* src, size_t srcSize, - const void* dict,size_t dictSize, - ZSTD_parameters params) + void* dst, size_t dstCapacity, + const void* src, size_t srcSize, + const void* dict,size_t dictSize, + const ZSTD_parameters* params) { ZSTD_CCtx_params const cctxParams = - ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params); + ZSTD_assignParamsToCCtxParams(&cctx->requestedParams, params); + DEBUGLOG(4, "ZSTD_compress_internal"); return ZSTD_compress_advanced_internal(cctx, - dst, dstCapacity, - src, srcSize, - dict, dictSize, - cctxParams); + dst, dstCapacity, + src, srcSize, + dict, dictSize, + &cctxParams); } -size_t ZSTD_compress_advanced (ZSTD_CCtx* ctx, +size_t ZSTD_compress_advanced (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize, const void* dict,size_t dictSize, ZSTD_parameters params) { - CHECK_F(ZSTD_checkCParams(params.cParams)); - return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, dict, dictSize, params); + DEBUGLOG(4, "ZSTD_compress_advanced"); + FORWARD_IF_ERROR(ZSTD_checkCParams(params.cParams), ""); + return ZSTD_compress_internal(cctx, + dst, dstCapacity, + src, srcSize, + dict, dictSize, + ¶ms); } /* Internal */ @@ -2110,34 +3272,47 @@ size_t ZSTD_compress_advanced_internal( void* dst, size_t dstCapacity, const void* src, size_t srcSize, const void* dict,size_t dictSize, - ZSTD_CCtx_params params) + const ZSTD_CCtx_params* params) { - CHECK_F( ZSTD_compressBegin_internal(cctx, dict, dictSize, ZSTD_dm_auto, NULL, - params, srcSize, ZSTDb_not_buffered) ); + DEBUGLOG(4, "ZSTD_compress_advanced_internal (srcSize:%u)", (unsigned)srcSize); + FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx, + dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL, + params, srcSize, ZSTDb_not_buffered) , ""); return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize); } -size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize, - const void* dict, size_t dictSize, int compressionLevel) +size_t ZSTD_compress_usingDict(ZSTD_CCtx* cctx, + void* dst, size_t dstCapacity, + const void* src, size_t srcSize, + const void* dict, size_t dictSize, + int compressionLevel) { - ZSTD_parameters params = ZSTD_getParams(compressionLevel, srcSize, dict ? dictSize : 0); - params.fParams.contentSizeFlag = 1; - return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, dict, dictSize, params); + ZSTD_parameters const params = ZSTD_getParams_internal(compressionLevel, srcSize, dict ? dictSize : 0); + ZSTD_CCtx_params cctxParams = ZSTD_assignParamsToCCtxParams(&cctx->requestedParams, ¶ms); + DEBUGLOG(4, "ZSTD_compress_usingDict (srcSize=%u)", (unsigned)srcSize); + assert(params.fParams.contentSizeFlag == 1); + return ZSTD_compress_advanced_internal(cctx, dst, dstCapacity, src, srcSize, dict, dictSize, &cctxParams); } -size_t ZSTD_compressCCtx (ZSTD_CCtx* ctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize, int compressionLevel) +size_t ZSTD_compressCCtx(ZSTD_CCtx* cctx, + void* dst, size_t dstCapacity, + const void* src, size_t srcSize, + int compressionLevel) { - return ZSTD_compress_usingDict(ctx, dst, dstCapacity, src, srcSize, NULL, 0, compressionLevel); + DEBUGLOG(4, "ZSTD_compressCCtx (srcSize=%u)", (unsigned)srcSize); + assert(cctx != NULL); + return ZSTD_compress_usingDict(cctx, dst, dstCapacity, src, srcSize, NULL, 0, compressionLevel); } -size_t ZSTD_compress(void* dst, size_t dstCapacity, const void* src, size_t srcSize, int compressionLevel) +size_t ZSTD_compress(void* dst, size_t dstCapacity, + const void* src, size_t srcSize, + int compressionLevel) { size_t result; ZSTD_CCtx ctxBody; - memset(&ctxBody, 0, sizeof(ctxBody)); - ctxBody.customMem = ZSTD_defaultCMem; + ZSTD_initCCtx(&ctxBody, ZSTD_defaultCMem); result = ZSTD_compressCCtx(&ctxBody, dst, dstCapacity, src, srcSize, compressionLevel); - ZSTD_free(ctxBody.workSpace, ZSTD_defaultCMem); /* can't free ctxBody itself, as it's on stack; free only heap content */ + ZSTD_freeCCtxContent(&ctxBody); /* can't free ctxBody itself, as it's on stack; free only heap content */ return result; } @@ -2150,54 +3325,77 @@ size_t ZSTD_estimateCDictSize_advanced( size_t dictSize, ZSTD_compressionParameters cParams, ZSTD_dictLoadMethod_e dictLoadMethod) { - DEBUGLOG(5, "sizeof(ZSTD_CDict) : %u", (U32)sizeof(ZSTD_CDict)); - DEBUGLOG(5, "CCtx estimate : %u", - (U32)ZSTD_estimateCCtxSize_advanced_usingCParams(cParams)); - return sizeof(ZSTD_CDict) + ZSTD_estimateCCtxSize_advanced_usingCParams(cParams) - + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize); + DEBUGLOG(5, "sizeof(ZSTD_CDict) : %u", (unsigned)sizeof(ZSTD_CDict)); + return ZSTD_cwksp_alloc_size(sizeof(ZSTD_CDict)) + + ZSTD_cwksp_alloc_size(HUF_WORKSPACE_SIZE) + + ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 0) + + (dictLoadMethod == ZSTD_dlm_byRef ? 0 + : ZSTD_cwksp_alloc_size(ZSTD_cwksp_align(dictSize, sizeof(void *)))); } size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel) { - ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, 0, dictSize); + ZSTD_compressionParameters const cParams = ZSTD_getCParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize); return ZSTD_estimateCDictSize_advanced(dictSize, cParams, ZSTD_dlm_byCopy); } size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict) { if (cdict==NULL) return 0; /* support sizeof on NULL */ - DEBUGLOG(5, "sizeof(*cdict) : %u", (U32)sizeof(*cdict)); - DEBUGLOG(5, "ZSTD_sizeof_CCtx : %u", (U32)ZSTD_sizeof_CCtx(cdict->refContext)); - return ZSTD_sizeof_CCtx(cdict->refContext) + (cdict->dictBuffer ? cdict->dictContentSize : 0) + sizeof(*cdict); + DEBUGLOG(5, "sizeof(*cdict) : %u", (unsigned)sizeof(*cdict)); + /* cdict may be in the workspace */ + return (cdict->workspace.workspace == cdict ? 0 : sizeof(*cdict)) + + ZSTD_cwksp_sizeof(&cdict->workspace); } static size_t ZSTD_initCDict_internal( ZSTD_CDict* cdict, const void* dictBuffer, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, - ZSTD_dictMode_e dictMode, + ZSTD_dictContentType_e dictContentType, ZSTD_compressionParameters cParams) { - DEBUGLOG(5, "ZSTD_initCDict_internal, mode %u", (U32)dictMode); + DEBUGLOG(3, "ZSTD_initCDict_internal (dictContentType:%u)", (unsigned)dictContentType); + assert(!ZSTD_checkCParams(cParams)); + cdict->matchState.cParams = cParams; if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dictBuffer) || (!dictSize)) { - cdict->dictBuffer = NULL; cdict->dictContent = dictBuffer; } else { - void* const internalBuffer = ZSTD_malloc(dictSize, cdict->refContext->customMem); - cdict->dictBuffer = internalBuffer; + void *internalBuffer = ZSTD_cwksp_reserve_object(&cdict->workspace, ZSTD_cwksp_align(dictSize, sizeof(void*))); + RETURN_ERROR_IF(!internalBuffer, memory_allocation, "NULL pointer!"); cdict->dictContent = internalBuffer; - if (!internalBuffer) return ERROR(memory_allocation); memcpy(internalBuffer, dictBuffer, dictSize); } cdict->dictContentSize = dictSize; - { ZSTD_CCtx_params cctxParams = cdict->refContext->requestedParams; - cctxParams.cParams = cParams; - CHECK_F( ZSTD_compressBegin_internal(cdict->refContext, - cdict->dictContent, dictSize, dictMode, - NULL, - cctxParams, ZSTD_CONTENTSIZE_UNKNOWN, - ZSTDb_not_buffered) ); + cdict->entropyWorkspace = (U32*)ZSTD_cwksp_reserve_object(&cdict->workspace, HUF_WORKSPACE_SIZE); + + + /* Reset the state to no dictionary */ + ZSTD_reset_compressedBlockState(&cdict->cBlockState); + FORWARD_IF_ERROR(ZSTD_reset_matchState( + &cdict->matchState, + &cdict->workspace, + &cParams, + ZSTDcrp_makeClean, + ZSTDirp_reset, + ZSTD_resetTarget_CDict), ""); + /* (Maybe) load the dictionary + * Skips loading the dictionary if it is < 8 bytes. + */ + { ZSTD_CCtx_params params; + memset(¶ms, 0, sizeof(params)); + params.compressionLevel = ZSTD_CLEVEL_DEFAULT; + params.fParams.contentSizeFlag = 1; + params.cParams = cParams; + { size_t const dictID = ZSTD_compress_insertDictionary( + &cdict->cBlockState, &cdict->matchState, NULL, &cdict->workspace, + ¶ms, cdict->dictContent, cdict->dictContentSize, + dictContentType, ZSTD_dtlm_full, cdict->entropyWorkspace); + FORWARD_IF_ERROR(dictID, "ZSTD_compress_insertDictionary failed"); + assert(dictID <= (size_t)(U32)-1); + cdict->dictID = (U32)dictID; + } } return 0; @@ -2205,24 +3403,38 @@ static size_t ZSTD_initCDict_internal( ZSTD_CDict* ZSTD_createCDict_advanced(const void* dictBuffer, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, - ZSTD_dictMode_e dictMode, + ZSTD_dictContentType_e dictContentType, ZSTD_compressionParameters cParams, ZSTD_customMem customMem) { - DEBUGLOG(5, "ZSTD_createCDict_advanced, mode %u", (U32)dictMode); + DEBUGLOG(3, "ZSTD_createCDict_advanced, mode %u", (unsigned)dictContentType); if (!customMem.customAlloc ^ !customMem.customFree) return NULL; - { ZSTD_CDict* const cdict = (ZSTD_CDict*)ZSTD_malloc(sizeof(ZSTD_CDict), customMem); - ZSTD_CCtx* const cctx = ZSTD_createCCtx_advanced(customMem); - - if (!cdict || !cctx) { - ZSTD_free(cdict, customMem); - ZSTD_freeCCtx(cctx); + { size_t const workspaceSize = + ZSTD_cwksp_alloc_size(sizeof(ZSTD_CDict)) + + ZSTD_cwksp_alloc_size(HUF_WORKSPACE_SIZE) + + ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 0) + + (dictLoadMethod == ZSTD_dlm_byRef ? 0 + : ZSTD_cwksp_alloc_size(ZSTD_cwksp_align(dictSize, sizeof(void*)))); + void* const workspace = ZSTD_malloc(workspaceSize, customMem); + ZSTD_cwksp ws; + ZSTD_CDict* cdict; + + if (!workspace) { + ZSTD_free(workspace, customMem); return NULL; } - cdict->refContext = cctx; + + ZSTD_cwksp_init(&ws, workspace, workspaceSize); + + cdict = (ZSTD_CDict*)ZSTD_cwksp_reserve_object(&ws, sizeof(ZSTD_CDict)); + assert(cdict != NULL); + ZSTD_cwksp_move(&cdict->workspace, &ws); + cdict->customMem = customMem; + cdict->compressionLevel = 0; /* signals advanced API usage */ + if (ZSTD_isError( ZSTD_initCDict_internal(cdict, dictBuffer, dictSize, - dictLoadMethod, dictMode, + dictLoadMethod, dictContentType, cParams) )) { ZSTD_freeCDict(cdict); return NULL; @@ -2234,27 +3446,32 @@ ZSTD_CDict* ZSTD_createCDict_advanced(const void* dictBuffer, size_t dictSize, ZSTD_CDict* ZSTD_createCDict(const void* dict, size_t dictSize, int compressionLevel) { - ZSTD_compressionParameters cParams = ZSTD_getCParams(compressionLevel, 0, dictSize); - return ZSTD_createCDict_advanced(dict, dictSize, - ZSTD_dlm_byCopy, ZSTD_dm_auto, - cParams, ZSTD_defaultCMem); + ZSTD_compressionParameters cParams = ZSTD_getCParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize); + ZSTD_CDict* cdict = ZSTD_createCDict_advanced(dict, dictSize, + ZSTD_dlm_byCopy, ZSTD_dct_auto, + cParams, ZSTD_defaultCMem); + if (cdict) + cdict->compressionLevel = compressionLevel == 0 ? ZSTD_CLEVEL_DEFAULT : compressionLevel; + return cdict; } ZSTD_CDict* ZSTD_createCDict_byReference(const void* dict, size_t dictSize, int compressionLevel) { - ZSTD_compressionParameters cParams = ZSTD_getCParams(compressionLevel, 0, dictSize); + ZSTD_compressionParameters cParams = ZSTD_getCParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize); return ZSTD_createCDict_advanced(dict, dictSize, - ZSTD_dlm_byRef, ZSTD_dm_auto, + ZSTD_dlm_byRef, ZSTD_dct_auto, cParams, ZSTD_defaultCMem); } size_t ZSTD_freeCDict(ZSTD_CDict* cdict) { if (cdict==NULL) return 0; /* support free on NULL */ - { ZSTD_customMem const cMem = cdict->refContext->customMem; - ZSTD_freeCCtx(cdict->refContext); - ZSTD_free(cdict->dictBuffer, cMem); - ZSTD_free(cdict, cMem); + { ZSTD_customMem const cMem = cdict->customMem; + int cdictInWorkspace = ZSTD_cwksp_owns_buffer(&cdict->workspace, cdict); + ZSTD_cwksp_free(&cdict->workspace, cMem); + if (!cdictInWorkspace) { + ZSTD_free(cdict, cMem); + } return 0; } } @@ -2272,43 +3489,48 @@ size_t ZSTD_freeCDict(ZSTD_CDict* cdict) * Note : there is no corresponding "free" function. * Since workspace was allocated externally, it must be freed externally. */ -ZSTD_CDict* ZSTD_initStaticCDict(void* workspace, size_t workspaceSize, +const ZSTD_CDict* ZSTD_initStaticCDict( + void* workspace, size_t workspaceSize, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, - ZSTD_dictMode_e dictMode, + ZSTD_dictContentType_e dictContentType, ZSTD_compressionParameters cParams) { - size_t const cctxSize = ZSTD_estimateCCtxSize_advanced_usingCParams(cParams); - size_t const neededSize = sizeof(ZSTD_CDict) + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize) - + cctxSize; - ZSTD_CDict* const cdict = (ZSTD_CDict*) workspace; - void* ptr; - DEBUGLOG(5, "(size_t)workspace & 7 : %u", (U32)(size_t)workspace & 7); + size_t const matchStateSize = ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 0); + size_t const neededSize = ZSTD_cwksp_alloc_size(sizeof(ZSTD_CDict)) + + (dictLoadMethod == ZSTD_dlm_byRef ? 0 + : ZSTD_cwksp_alloc_size(ZSTD_cwksp_align(dictSize, sizeof(void*)))) + + ZSTD_cwksp_alloc_size(HUF_WORKSPACE_SIZE) + + matchStateSize; + ZSTD_CDict* cdict; + if ((size_t)workspace & 7) return NULL; /* 8-aligned */ - DEBUGLOG(5, "(workspaceSize < neededSize) : (%u < %u) => %u", - (U32)workspaceSize, (U32)neededSize, (U32)(workspaceSize < neededSize)); - if (workspaceSize < neededSize) return NULL; - if (dictLoadMethod == ZSTD_dlm_byCopy) { - memcpy(cdict+1, dict, dictSize); - dict = cdict+1; - ptr = (char*)workspace + sizeof(ZSTD_CDict) + dictSize; - } else { - ptr = cdict+1; + { + ZSTD_cwksp ws; + ZSTD_cwksp_init(&ws, workspace, workspaceSize); + cdict = (ZSTD_CDict*)ZSTD_cwksp_reserve_object(&ws, sizeof(ZSTD_CDict)); + if (cdict == NULL) return NULL; + ZSTD_cwksp_move(&cdict->workspace, &ws); } - cdict->refContext = ZSTD_initStaticCCtx(ptr, cctxSize); + + DEBUGLOG(4, "(workspaceSize < neededSize) : (%u < %u) => %u", + (unsigned)workspaceSize, (unsigned)neededSize, (unsigned)(workspaceSize < neededSize)); + if (workspaceSize < neededSize) return NULL; if (ZSTD_isError( ZSTD_initCDict_internal(cdict, dict, dictSize, - ZSTD_dlm_byRef, dictMode, + dictLoadMethod, dictContentType, cParams) )) return NULL; return cdict; } -ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict) { - return cdict->refContext->appliedParams.cParams; +ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict) +{ + assert(cdict != NULL); + return cdict->matchState.cParams; } /* ZSTD_compressBegin_usingCDict_advanced() : @@ -2317,15 +3539,32 @@ size_t ZSTD_compressBegin_usingCDict_advanced( ZSTD_CCtx* const cctx, const ZSTD_CDict* const cdict, ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize) { - if (cdict==NULL) return ERROR(dictionary_wrong); + DEBUGLOG(4, "ZSTD_compressBegin_usingCDict_advanced"); + RETURN_ERROR_IF(cdict==NULL, dictionary_wrong, "NULL pointer!"); { ZSTD_CCtx_params params = cctx->requestedParams; - params.cParams = ZSTD_getCParamsFromCDict(cdict); + params.cParams = ( pledgedSrcSize < ZSTD_USE_CDICT_PARAMS_SRCSIZE_CUTOFF + || pledgedSrcSize < cdict->dictContentSize * ZSTD_USE_CDICT_PARAMS_DICTSIZE_MULTIPLIER + || pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN + || cdict->compressionLevel == 0 ) + && (params.attachDictPref != ZSTD_dictForceLoad) ? + ZSTD_getCParamsFromCDict(cdict) + : ZSTD_getCParams(cdict->compressionLevel, + pledgedSrcSize, + cdict->dictContentSize); + /* Increase window log to fit the entire dictionary and source if the + * source size is known. Limit the increase to 19, which is the + * window log for compression level 1 with the largest source size. + */ + if (pledgedSrcSize != ZSTD_CONTENTSIZE_UNKNOWN) { + U32 const limitedSrcSize = (U32)MIN(pledgedSrcSize, 1U << 19); + U32 const limitedSrcLog = limitedSrcSize > 1 ? ZSTD_highbit32(limitedSrcSize - 1) + 1 : 1; + params.cParams.windowLog = MAX(params.cParams.windowLog, limitedSrcLog); + } params.fParams = fParams; - DEBUGLOG(5, "ZSTD_compressBegin_usingCDict_advanced"); return ZSTD_compressBegin_internal(cctx, - NULL, 0, ZSTD_dm_auto, + NULL, 0, ZSTD_dct_auto, ZSTD_dtlm_fast, cdict, - params, pledgedSrcSize, + ¶ms, pledgedSrcSize, ZSTDb_not_buffered); } } @@ -2336,8 +3575,8 @@ size_t ZSTD_compressBegin_usingCDict_advanced( size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict) { ZSTD_frameParameters const fParams = { 0 /*content*/, 0 /*checksum*/, 0 /*noDictID*/ }; - DEBUGLOG(5, "ZSTD_compressBegin_usingCDict : dictIDFlag == %u", !fParams.noDictIDFlag); - return ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, 0); + DEBUGLOG(4, "ZSTD_compressBegin_usingCDict : dictIDFlag == %u", !fParams.noDictIDFlag); + return ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, ZSTD_CONTENTSIZE_UNKNOWN); } size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx, @@ -2345,7 +3584,7 @@ size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx, const void* src, size_t srcSize, const ZSTD_CDict* cdict, ZSTD_frameParameters fParams) { - CHECK_F (ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, srcSize)); /* will check if cdict != NULL */ + FORWARD_IF_ERROR(ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, srcSize), ""); /* will check if cdict != NULL */ return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize); } @@ -2371,6 +3610,7 @@ size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx, ZSTD_CStream* ZSTD_createCStream(void) { + DEBUGLOG(3, "ZSTD_createCStream"); return ZSTD_createCStream_advanced(ZSTD_defaultCMem); } @@ -2400,76 +3640,71 @@ size_t ZSTD_CStreamOutSize(void) return ZSTD_compressBound(ZSTD_BLOCKSIZE_MAX) + ZSTD_blockHeaderSize + 4 /* 32-bits hash */ ; } -static size_t ZSTD_resetCStream_internal(ZSTD_CStream* zcs, - const void* dict, size_t dictSize, ZSTD_dictMode_e dictMode, - const ZSTD_CDict* cdict, - const ZSTD_CCtx_params params, unsigned long long pledgedSrcSize) +static size_t ZSTD_resetCStream_internal(ZSTD_CStream* cctx, + const void* const dict, size_t const dictSize, ZSTD_dictContentType_e const dictContentType, + const ZSTD_CDict* const cdict, + ZSTD_CCtx_params params, unsigned long long const pledgedSrcSize) { DEBUGLOG(4, "ZSTD_resetCStream_internal"); + /* Finalize the compression parameters */ + params.cParams = ZSTD_getCParamsFromCCtxParams(¶ms, pledgedSrcSize, dictSize); /* params are supposed to be fully validated at this point */ assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams))); assert(!((dict) && (cdict))); /* either dict or cdict, not both */ - CHECK_F( ZSTD_compressBegin_internal(zcs, - dict, dictSize, dictMode, - cdict, - params, pledgedSrcSize, - ZSTDb_buffered) ); - - zcs->inToCompress = 0; - zcs->inBuffPos = 0; - zcs->inBuffTarget = zcs->blockSize; - zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0; - zcs->streamStage = zcss_load; - zcs->frameEnded = 0; + FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx, + dict, dictSize, dictContentType, ZSTD_dtlm_fast, + cdict, + ¶ms, pledgedSrcSize, + ZSTDb_buffered) , ""); + + cctx->inToCompress = 0; + cctx->inBuffPos = 0; + cctx->inBuffTarget = cctx->blockSize + + (cctx->blockSize == pledgedSrcSize); /* for small input: avoid automatic flush on reaching end of block, since it would require to add a 3-bytes null block to end frame */ + cctx->outBuffContentSize = cctx->outBuffFlushedSize = 0; + cctx->streamStage = zcss_load; + cctx->frameEnded = 0; return 0; /* ready to go */ } -size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize) +/* ZSTD_resetCStream(): + * pledgedSrcSize == 0 means "unknown" */ +size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pss) { - ZSTD_CCtx_params params = zcs->requestedParams; - params.fParams.contentSizeFlag = (pledgedSrcSize > 0); - params.cParams = ZSTD_getCParamsFromCCtxParams(params, pledgedSrcSize, 0); - DEBUGLOG(5, "ZSTD_resetCStream"); - return ZSTD_resetCStream_internal(zcs, NULL, 0, ZSTD_dm_auto, zcs->cdict, params, pledgedSrcSize); + /* temporary : 0 interpreted as "unknown" during transition period. + * Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN. + * 0 will be interpreted as "empty" in the future. + */ + U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss; + DEBUGLOG(4, "ZSTD_resetCStream: pledgedSrcSize = %u", (unsigned)pledgedSrcSize); + FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , ""); + return 0; } /*! ZSTD_initCStream_internal() : - * Note : not static, but hidden (not exposed). Used by zstdmt_compress.c + * Note : for lib/compress only. Used by zstdmt_compress.c. * Assumption 1 : params are valid * Assumption 2 : either dict, or cdict, is defined, not both */ size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs, const void* dict, size_t dictSize, const ZSTD_CDict* cdict, - ZSTD_CCtx_params params, unsigned long long pledgedSrcSize) + const ZSTD_CCtx_params* params, + unsigned long long pledgedSrcSize) { - assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams))); + DEBUGLOG(4, "ZSTD_initCStream_internal"); + FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , ""); + assert(!ZSTD_isError(ZSTD_checkCParams(params->cParams))); + zcs->requestedParams = *params; assert(!((dict) && (cdict))); /* either dict or cdict, not both */ - - if (dict && dictSize >= 8) { - DEBUGLOG(5, "loading dictionary of size %u", (U32)dictSize); - if (zcs->staticSize) { /* static CCtx : never uses malloc */ - /* incompatible with internal cdict creation */ - return ERROR(memory_allocation); - } - ZSTD_freeCDict(zcs->cdictLocal); - zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize, - ZSTD_dlm_byCopy, ZSTD_dm_auto, - params.cParams, zcs->customMem); - zcs->cdict = zcs->cdictLocal; - if (zcs->cdictLocal == NULL) return ERROR(memory_allocation); + if (dict) { + FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) , ""); } else { - if (cdict) { - params.cParams = ZSTD_getCParamsFromCDict(cdict); /* cParams are enforced from cdict */ - } - ZSTD_freeCDict(zcs->cdictLocal); - zcs->cdictLocal = NULL; - zcs->cdict = cdict; + /* Dictionary is cleared if !cdict */ + FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) , ""); } - - params.compressionLevel = ZSTD_CLEVEL_CUSTOM; - zcs->requestedParams = params; - - return ZSTD_resetCStream_internal(zcs, NULL, 0, ZSTD_dm_auto, zcs->cdict, params, pledgedSrcSize); + return 0; } /* ZSTD_initCStream_usingCDict_advanced() : @@ -2478,88 +3713,112 @@ size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs, const ZSTD_CDict* cdict, ZSTD_frameParameters fParams, unsigned long long pledgedSrcSize) -{ /* cannot handle NULL cdict (does not know what to do) */ - if (!cdict) return ERROR(dictionary_wrong); - { ZSTD_CCtx_params params = zcs->requestedParams; - params.cParams = ZSTD_getCParamsFromCDict(cdict); - params.fParams = fParams; - return ZSTD_initCStream_internal(zcs, - NULL, 0, cdict, - params, pledgedSrcSize); - } +{ + DEBUGLOG(4, "ZSTD_initCStream_usingCDict_advanced"); + FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , ""); + zcs->requestedParams.fParams = fParams; + FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) , ""); + return 0; } /* note : cdict must outlive compression session */ size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict) { - ZSTD_frameParameters const fParams = { 0 /* contentSize */, 0 /* checksum */, 0 /* hideDictID */ }; - return ZSTD_initCStream_usingCDict_advanced(zcs, cdict, fParams, 0); /* note : will check that cdict != NULL */ + DEBUGLOG(4, "ZSTD_initCStream_usingCDict"); + FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) , ""); + return 0; } + +/* ZSTD_initCStream_advanced() : + * pledgedSrcSize must be exact. + * if srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN. + * dict is loaded with default parameters ZSTD_dct_auto and ZSTD_dlm_byCopy. */ size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs, const void* dict, size_t dictSize, - ZSTD_parameters params, unsigned long long pledgedSrcSize) + ZSTD_parameters params, unsigned long long pss) { - ZSTD_CCtx_params const cctxParams = - ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params); - CHECK_F( ZSTD_checkCParams(params.cParams) ); - return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, cctxParams, pledgedSrcSize); + /* for compatibility with older programs relying on this behavior. + * Users should now specify ZSTD_CONTENTSIZE_UNKNOWN. + * This line will be removed in the future. + */ + U64 const pledgedSrcSize = (pss==0 && params.fParams.contentSizeFlag==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss; + DEBUGLOG(4, "ZSTD_initCStream_advanced"); + FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , ""); + FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) , ""); + zcs->requestedParams = ZSTD_assignParamsToCCtxParams(&zcs->requestedParams, ¶ms); + FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) , ""); + return 0; } size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel) { - ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, dictSize); - ZSTD_CCtx_params const cctxParams = - ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params); - return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, cctxParams, 0); + DEBUGLOG(4, "ZSTD_initCStream_usingDict"); + FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) , ""); + return 0; } -size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize) +size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pss) { - ZSTD_CCtx_params cctxParams; - ZSTD_parameters const params = ZSTD_getParams(compressionLevel, pledgedSrcSize, 0); - cctxParams = ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params); - cctxParams.fParams.contentSizeFlag = (pledgedSrcSize>0); - return ZSTD_initCStream_internal(zcs, NULL, 0, NULL, cctxParams, pledgedSrcSize); + /* temporary : 0 interpreted as "unknown" during transition period. + * Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN. + * 0 will be interpreted as "empty" in the future. + */ + U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss; + DEBUGLOG(4, "ZSTD_initCStream_srcSize"); + FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, NULL) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , ""); + return 0; } size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel) { - return ZSTD_initCStream_srcSize(zcs, compressionLevel, 0); + DEBUGLOG(4, "ZSTD_initCStream"); + FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, NULL) , ""); + FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) , ""); + return 0; } /*====== Compression ======*/ -MEM_STATIC size_t ZSTD_limitCopy(void* dst, size_t dstCapacity, - const void* src, size_t srcSize) +static size_t ZSTD_nextInputSizeHint(const ZSTD_CCtx* cctx) { - size_t const length = MIN(dstCapacity, srcSize); - if (length) memcpy(dst, src, length); - return length; + size_t hintInSize = cctx->inBuffTarget - cctx->inBuffPos; + if (hintInSize==0) hintInSize = cctx->blockSize; + return hintInSize; } /** ZSTD_compressStream_generic(): - * internal function for all *compressStream*() variants and *compress_generic() + * internal function for all *compressStream*() variants + * non-static, because can be called from zstdmt_compress.c * @return : hint size for next input */ -size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, - ZSTD_outBuffer* output, - ZSTD_inBuffer* input, - ZSTD_EndDirective const flushMode) +static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, + ZSTD_outBuffer* output, + ZSTD_inBuffer* input, + ZSTD_EndDirective const flushMode) { const char* const istart = (const char*)input->src; - const char* const iend = istart + input->size; - const char* ip = istart + input->pos; + const char* const iend = input->size != 0 ? istart + input->size : istart; + const char* ip = input->pos != 0 ? istart + input->pos : istart; char* const ostart = (char*)output->dst; - char* const oend = ostart + output->size; - char* op = ostart + output->pos; + char* const oend = output->size != 0 ? ostart + output->size : ostart; + char* op = output->pos != 0 ? ostart + output->pos : ostart; U32 someMoreWork = 1; /* check expectations */ - DEBUGLOG(5, "ZSTD_compressStream_generic, flush=%u", (U32)flushMode); + DEBUGLOG(5, "ZSTD_compressStream_generic, flush=%u", (unsigned)flushMode); assert(zcs->inBuff != NULL); - assert(zcs->inBuffSize>0); - assert(zcs->outBuff!= NULL); - assert(zcs->outBuffSize>0); + assert(zcs->inBuffSize > 0); + assert(zcs->outBuff != NULL); + assert(zcs->outBuffSize > 0); assert(output->pos <= output->size); assert(input->pos <= input->size); @@ -2567,8 +3826,7 @@ size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, switch(zcs->streamStage) { case zcss_init: - /* call ZSTD_initCStream() first ! */ - return ERROR(init_missing); + RETURN_ERROR(init_missing, "call ZSTD_initCStream() first!"); case zcss_load: if ( (flushMode == ZSTD_e_end) @@ -2577,12 +3835,12 @@ size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, /* shortcut to compression pass directly into output buffer */ size_t const cSize = ZSTD_compressEnd(zcs, op, oend-op, ip, iend-ip); - DEBUGLOG(4, "ZSTD_compressEnd : %u", (U32)cSize); - if (ZSTD_isError(cSize)) return cSize; + DEBUGLOG(4, "ZSTD_compressEnd : cSize=%u", (unsigned)cSize); + FORWARD_IF_ERROR(cSize, "ZSTD_compressEnd failed"); ip = iend; op += cSize; zcs->frameEnded = 1; - ZSTD_startNewCompression(zcs); + ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); someMoreWork = 0; break; } /* complete loading into inBuffer */ @@ -2591,7 +3849,8 @@ size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, zcs->inBuff + zcs->inBuffPos, toLoad, ip, iend-ip); zcs->inBuffPos += loaded; - ip += loaded; + if (loaded != 0) + ip += loaded; if ( (flushMode == ZSTD_e_continue) && (zcs->inBuffPos < zcs->inBuffTarget) ) { /* not enough input to fill full block : stop here */ @@ -2619,14 +3878,14 @@ size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, zcs->inBuff + zcs->inToCompress, iSize) : ZSTD_compressContinue(zcs, cDst, oSize, zcs->inBuff + zcs->inToCompress, iSize); - if (ZSTD_isError(cSize)) return cSize; + FORWARD_IF_ERROR(cSize, "%s", lastBlock ? "ZSTD_compressEnd failed" : "ZSTD_compressContinue failed"); zcs->frameEnded = lastBlock; /* prepare next block */ zcs->inBuffTarget = zcs->inBuffPos + zcs->blockSize; if (zcs->inBuffTarget > zcs->inBuffSize) zcs->inBuffPos = 0, zcs->inBuffTarget = zcs->blockSize; DEBUGLOG(5, "inBuffTarget:%u / inBuffSize:%u", - (U32)zcs->inBuffTarget, (U32)zcs->inBuffSize); + (unsigned)zcs->inBuffTarget, (unsigned)zcs->inBuffSize); if (!lastBlock) assert(zcs->inBuffTarget <= zcs->inBuffSize); zcs->inToCompress = zcs->inBuffPos; @@ -2635,7 +3894,7 @@ size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, if (zcs->frameEnded) { DEBUGLOG(5, "Frame completed directly in outBuffer"); someMoreWork = 0; - ZSTD_startNewCompression(zcs); + ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); } break; } @@ -2647,11 +3906,12 @@ size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, case zcss_flush: DEBUGLOG(5, "flush stage"); { size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize; - size_t const flushed = ZSTD_limitCopy(op, oend-op, + size_t const flushed = ZSTD_limitCopy(op, (size_t)(oend-op), zcs->outBuff + zcs->outBuffFlushedSize, toFlush); DEBUGLOG(5, "toFlush: %u into %u ==> flushed: %u", - (U32)toFlush, (U32)(oend-op), (U32)flushed); - op += flushed; + (unsigned)toFlush, (unsigned)(oend-op), (unsigned)flushed); + if (flushed) + op += flushed; zcs->outBuffFlushedSize += flushed; if (toFlush!=flushed) { /* flush not fully completed, presumably because dst is too small */ @@ -2663,7 +3923,7 @@ size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, if (zcs->frameEnded) { DEBUGLOG(5, "Frame completed on flush"); someMoreWork = 0; - ZSTD_startNewCompression(zcs); + ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); break; } zcs->streamStage = zcss_load; @@ -2678,82 +3938,115 @@ size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, input->pos = ip - istart; output->pos = op - ostart; if (zcs->frameEnded) return 0; - { size_t hintInSize = zcs->inBuffTarget - zcs->inBuffPos; - if (hintInSize==0) hintInSize = zcs->blockSize; - return hintInSize; + return ZSTD_nextInputSizeHint(zcs); +} + +static size_t ZSTD_nextInputSizeHint_MTorST(const ZSTD_CCtx* cctx) +{ +#ifdef ZSTD_MULTITHREAD + if (cctx->appliedParams.nbWorkers >= 1) { + assert(cctx->mtctx != NULL); + return ZSTDMT_nextInputSizeHint(cctx->mtctx); } +#endif + return ZSTD_nextInputSizeHint(cctx); + } size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input) { - /* check conditions */ - if (output->pos > output->size) return ERROR(GENERIC); - if (input->pos > input->size) return ERROR(GENERIC); - - return ZSTD_compressStream_generic(zcs, output, input, ZSTD_e_continue); + FORWARD_IF_ERROR( ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue) , ""); + return ZSTD_nextInputSizeHint_MTorST(zcs); } -size_t ZSTD_compress_generic (ZSTD_CCtx* cctx, - ZSTD_outBuffer* output, - ZSTD_inBuffer* input, - ZSTD_EndDirective endOp) +size_t ZSTD_compressStream2( ZSTD_CCtx* cctx, + ZSTD_outBuffer* output, + ZSTD_inBuffer* input, + ZSTD_EndDirective endOp) { + DEBUGLOG(5, "ZSTD_compressStream2, endOp=%u ", (unsigned)endOp); /* check conditions */ - if (output->pos > output->size) return ERROR(GENERIC); - if (input->pos > input->size) return ERROR(GENERIC); + RETURN_ERROR_IF(output->pos > output->size, GENERIC, "invalid buffer"); + RETURN_ERROR_IF(input->pos > input->size, GENERIC, "invalid buffer"); assert(cctx!=NULL); /* transparent initialization stage */ if (cctx->streamStage == zcss_init) { - ZSTD_prefixDict const prefixDict = cctx->prefixDict; ZSTD_CCtx_params params = cctx->requestedParams; + ZSTD_prefixDict const prefixDict = cctx->prefixDict; + FORWARD_IF_ERROR( ZSTD_initLocalDict(cctx) , ""); /* Init the local dict if present. */ + memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict)); /* single usage */ + assert(prefixDict.dict==NULL || cctx->cdict==NULL); /* only one can be set */ + DEBUGLOG(4, "ZSTD_compressStream2 : transparent init stage"); + if (endOp == ZSTD_e_end) cctx->pledgedSrcSizePlusOne = input->size + 1; /* auto-fix pledgedSrcSize */ params.cParams = ZSTD_getCParamsFromCCtxParams( - cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1, 0 /*dictSize*/); - memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict)); /* single usage */ - assert(prefixDict.dict==NULL || cctx->cdict==NULL); /* only one can be set */ + &cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1, 0 /*dictSize*/); + #ifdef ZSTD_MULTITHREAD - if (params.nbThreads > 1) { - if (cctx->mtctx == NULL || cctx->appliedParams.nbThreads != params.nbThreads) { - ZSTDMT_freeCCtx(cctx->mtctx); - cctx->mtctx = ZSTDMT_createCCtx_advanced(params.nbThreads, cctx->customMem); - if (cctx->mtctx == NULL) return ERROR(memory_allocation); + if ((cctx->pledgedSrcSizePlusOne-1) <= ZSTDMT_JOBSIZE_MIN) { + params.nbWorkers = 0; /* do not invoke multi-threading when src size is too small */ + } + if (params.nbWorkers > 0) { + /* mt context creation */ + if (cctx->mtctx == NULL) { + DEBUGLOG(4, "ZSTD_compressStream2: creating new mtctx for nbWorkers=%u", + params.nbWorkers); + cctx->mtctx = ZSTDMT_createCCtx_advanced((U32)params.nbWorkers, cctx->customMem); + RETURN_ERROR_IF(cctx->mtctx == NULL, memory_allocation, "NULL pointer!"); } - DEBUGLOG(4, "call ZSTDMT_initCStream_internal as nbThreads=%u", params.nbThreads); - CHECK_F( ZSTDMT_initCStream_internal( - cctx->mtctx, - prefixDict.dict, prefixDict.dictSize, ZSTD_dm_rawContent, - cctx->cdict, params, cctx->pledgedSrcSizePlusOne-1) ); + /* mt compression */ + DEBUGLOG(4, "call ZSTDMT_initCStream_internal as nbWorkers=%u", params.nbWorkers); + FORWARD_IF_ERROR( ZSTDMT_initCStream_internal( + cctx->mtctx, + prefixDict.dict, prefixDict.dictSize, prefixDict.dictContentType, + cctx->cdict, params, cctx->pledgedSrcSizePlusOne-1) , ""); cctx->streamStage = zcss_load; - cctx->appliedParams.nbThreads = params.nbThreads; + cctx->appliedParams.nbWorkers = params.nbWorkers; } else #endif - { - CHECK_F( ZSTD_resetCStream_internal( - cctx, prefixDict.dict, prefixDict.dictSize, - prefixDict.dictMode, cctx->cdict, params, - cctx->pledgedSrcSizePlusOne-1) ); + { FORWARD_IF_ERROR( ZSTD_resetCStream_internal(cctx, + prefixDict.dict, prefixDict.dictSize, prefixDict.dictContentType, + cctx->cdict, + params, cctx->pledgedSrcSizePlusOne-1) , ""); + assert(cctx->streamStage == zcss_load); + assert(cctx->appliedParams.nbWorkers == 0); } } + /* end of transparent initialization stage */ /* compression stage */ #ifdef ZSTD_MULTITHREAD - if (cctx->appliedParams.nbThreads > 1) { - size_t const flushMin = ZSTDMT_compressStream_generic(cctx->mtctx, output, input, endOp); - DEBUGLOG(5, "ZSTDMT_compressStream_generic : %u", (U32)flushMin); - if ( ZSTD_isError(flushMin) - || (endOp == ZSTD_e_end && flushMin == 0) ) { /* compression completed */ - ZSTD_startNewCompression(cctx); + if (cctx->appliedParams.nbWorkers > 0) { + int const forceMaxProgress = (endOp == ZSTD_e_flush || endOp == ZSTD_e_end); + size_t flushMin; + assert(forceMaxProgress || endOp == ZSTD_e_continue /* Protection for a new flush type */); + if (cctx->cParamsChanged) { + ZSTDMT_updateCParams_whileCompressing(cctx->mtctx, &cctx->requestedParams); + cctx->cParamsChanged = 0; } + do { + flushMin = ZSTDMT_compressStream_generic(cctx->mtctx, output, input, endOp); + if ( ZSTD_isError(flushMin) + || (endOp == ZSTD_e_end && flushMin == 0) ) { /* compression completed */ + ZSTD_CCtx_reset(cctx, ZSTD_reset_session_only); + } + FORWARD_IF_ERROR(flushMin, "ZSTDMT_compressStream_generic failed"); + } while (forceMaxProgress && flushMin != 0 && output->pos < output->size); + DEBUGLOG(5, "completed ZSTD_compressStream2 delegating to ZSTDMT_compressStream_generic"); + /* Either we don't require maximum forward progress, we've finished the + * flush, or we are out of output space. + */ + assert(!forceMaxProgress || flushMin == 0 || output->pos == output->size); return flushMin; } #endif - CHECK_F( ZSTD_compressStream_generic(cctx, output, input, endOp) ); - DEBUGLOG(5, "completed ZSTD_compress_generic"); + FORWARD_IF_ERROR( ZSTD_compressStream_generic(cctx, output, input, endOp) , ""); + DEBUGLOG(5, "completed ZSTD_compressStream2"); return cctx->outBuffContentSize - cctx->outBuffFlushedSize; /* remaining to flush */ } -size_t ZSTD_compress_generic_simpleArgs ( +size_t ZSTD_compressStream2_simpleArgs ( ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, size_t* dstPos, const void* src, size_t srcSize, size_t* srcPos, @@ -2761,37 +4054,57 @@ size_t ZSTD_compress_generic_simpleArgs ( { ZSTD_outBuffer output = { dst, dstCapacity, *dstPos }; ZSTD_inBuffer input = { src, srcSize, *srcPos }; - /* ZSTD_compress_generic() will check validity of dstPos and srcPos */ - size_t const cErr = ZSTD_compress_generic(cctx, &output, &input, endOp); + /* ZSTD_compressStream2() will check validity of dstPos and srcPos */ + size_t const cErr = ZSTD_compressStream2(cctx, &output, &input, endOp); *dstPos = output.pos; *srcPos = input.pos; return cErr; } +size_t ZSTD_compress2(ZSTD_CCtx* cctx, + void* dst, size_t dstCapacity, + const void* src, size_t srcSize) +{ + DEBUGLOG(4, "ZSTD_compress2 (srcSize=%u)", (unsigned)srcSize); + ZSTD_CCtx_reset(cctx, ZSTD_reset_session_only); + { size_t oPos = 0; + size_t iPos = 0; + size_t const result = ZSTD_compressStream2_simpleArgs(cctx, + dst, dstCapacity, &oPos, + src, srcSize, &iPos, + ZSTD_e_end); + FORWARD_IF_ERROR(result, "ZSTD_compressStream2_simpleArgs failed"); + if (result != 0) { /* compression not completed, due to lack of output space */ + assert(oPos == dstCapacity); + RETURN_ERROR(dstSize_tooSmall, ""); + } + assert(iPos == srcSize); /* all input is expected consumed */ + return oPos; + } +} /*====== Finalize ======*/ /*! ZSTD_flushStream() : -* @return : amount of data remaining to flush */ + * @return : amount of data remaining to flush */ size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output) { ZSTD_inBuffer input = { NULL, 0, 0 }; - if (output->pos > output->size) return ERROR(GENERIC); - CHECK_F( ZSTD_compressStream_generic(zcs, output, &input, ZSTD_e_flush) ); - return zcs->outBuffContentSize - zcs->outBuffFlushedSize; /* remaining to flush */ + return ZSTD_compressStream2(zcs, output, &input, ZSTD_e_flush); } size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output) { ZSTD_inBuffer input = { NULL, 0, 0 }; - if (output->pos > output->size) return ERROR(GENERIC); - CHECK_F( ZSTD_compressStream_generic(zcs, output, &input, ZSTD_e_end) ); + size_t const remainingToFlush = ZSTD_compressStream2(zcs, output, &input, ZSTD_e_end); + FORWARD_IF_ERROR( remainingToFlush , "ZSTD_compressStream2 failed"); + if (zcs->appliedParams.nbWorkers > 0) return remainingToFlush; /* minimal estimation */ + /* single thread mode : attempt to calculate remaining to flush more precisely */ { size_t const lastBlockSize = zcs->frameEnded ? 0 : ZSTD_BLOCKHEADERSIZE; - size_t const checksumSize = zcs->frameEnded ? 0 : zcs->appliedParams.fParams.checksumFlag * 4; - size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize + lastBlockSize + checksumSize; - DEBUGLOG(5, "ZSTD_endStream : remaining to flush : %u", - (unsigned)toFlush); + size_t const checksumSize = (size_t)(zcs->frameEnded ? 0 : zcs->appliedParams.fParams.checksumFlag * 4); + size_t const toFlush = remainingToFlush + lastBlockSize + checksumSize; + DEBUGLOG(4, "ZSTD_endStream : remaining to flush : %u", (unsigned)toFlush); return toFlush; } } @@ -2801,163 +4114,165 @@ size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output) #define ZSTD_MAX_CLEVEL 22 int ZSTD_maxCLevel(void) { return ZSTD_MAX_CLEVEL; } +int ZSTD_minCLevel(void) { return (int)-ZSTD_TARGETLENGTH_MAX; } static const ZSTD_compressionParameters ZSTD_defaultCParameters[4][ZSTD_MAX_CLEVEL+1] = { -{ /* "default" - guarantees a monotonically increasing memory budget */ +{ /* "default" - for any srcSize > 256 KB */ /* W, C, H, S, L, TL, strat */ - { 18, 12, 12, 1, 7, 16, ZSTD_fast }, /* level 0 - never used */ - { 19, 13, 14, 1, 7, 16, ZSTD_fast }, /* level 1 */ - { 19, 15, 16, 1, 6, 16, ZSTD_fast }, /* level 2 */ - { 20, 16, 17, 1, 5, 16, ZSTD_dfast }, /* level 3 */ - { 20, 17, 18, 1, 5, 16, ZSTD_dfast }, /* level 4 */ - { 20, 17, 18, 2, 5, 16, ZSTD_greedy }, /* level 5 */ - { 21, 17, 19, 2, 5, 16, ZSTD_lazy }, /* level 6 */ - { 21, 18, 19, 3, 5, 16, ZSTD_lazy }, /* level 7 */ - { 21, 18, 20, 3, 5, 16, ZSTD_lazy2 }, /* level 8 */ - { 21, 19, 20, 3, 5, 16, ZSTD_lazy2 }, /* level 9 */ - { 21, 19, 21, 4, 5, 16, ZSTD_lazy2 }, /* level 10 */ - { 22, 20, 22, 4, 5, 16, ZSTD_lazy2 }, /* level 11 */ - { 22, 20, 22, 5, 5, 16, ZSTD_lazy2 }, /* level 12 */ - { 22, 21, 22, 5, 5, 16, ZSTD_lazy2 }, /* level 13 */ - { 22, 21, 22, 6, 5, 16, ZSTD_lazy2 }, /* level 14 */ - { 22, 21, 22, 5, 5, 16, ZSTD_btlazy2 }, /* level 15 */ - { 23, 22, 22, 5, 5, 16, ZSTD_btlazy2 }, /* level 16 */ - { 23, 22, 22, 4, 5, 24, ZSTD_btopt }, /* level 17 */ - { 23, 22, 22, 5, 4, 32, ZSTD_btopt }, /* level 18 */ - { 23, 23, 22, 6, 3, 48, ZSTD_btopt }, /* level 19 */ - { 25, 25, 23, 7, 3, 64, ZSTD_btultra }, /* level 20 */ - { 26, 26, 24, 7, 3,256, ZSTD_btultra }, /* level 21 */ - { 27, 27, 25, 9, 3,512, ZSTD_btultra }, /* level 22 */ + { 19, 12, 13, 1, 6, 1, ZSTD_fast }, /* base for negative levels */ + { 19, 13, 14, 1, 7, 0, ZSTD_fast }, /* level 1 */ + { 20, 15, 16, 1, 6, 0, ZSTD_fast }, /* level 2 */ + { 21, 16, 17, 1, 5, 0, ZSTD_dfast }, /* level 3 */ + { 21, 18, 18, 1, 5, 0, ZSTD_dfast }, /* level 4 */ + { 21, 18, 19, 2, 5, 2, ZSTD_greedy }, /* level 5 */ + { 21, 19, 19, 3, 5, 4, ZSTD_greedy }, /* level 6 */ + { 21, 19, 19, 3, 5, 8, ZSTD_lazy }, /* level 7 */ + { 21, 19, 19, 3, 5, 16, ZSTD_lazy2 }, /* level 8 */ + { 21, 19, 20, 4, 5, 16, ZSTD_lazy2 }, /* level 9 */ + { 22, 20, 21, 4, 5, 16, ZSTD_lazy2 }, /* level 10 */ + { 22, 21, 22, 4, 5, 16, ZSTD_lazy2 }, /* level 11 */ + { 22, 21, 22, 5, 5, 16, ZSTD_lazy2 }, /* level 12 */ + { 22, 21, 22, 5, 5, 32, ZSTD_btlazy2 }, /* level 13 */ + { 22, 22, 23, 5, 5, 32, ZSTD_btlazy2 }, /* level 14 */ + { 22, 23, 23, 6, 5, 32, ZSTD_btlazy2 }, /* level 15 */ + { 22, 22, 22, 5, 5, 48, ZSTD_btopt }, /* level 16 */ + { 23, 23, 22, 5, 4, 64, ZSTD_btopt }, /* level 17 */ + { 23, 23, 22, 6, 3, 64, ZSTD_btultra }, /* level 18 */ + { 23, 24, 22, 7, 3,256, ZSTD_btultra2}, /* level 19 */ + { 25, 25, 23, 7, 3,256, ZSTD_btultra2}, /* level 20 */ + { 26, 26, 24, 7, 3,512, ZSTD_btultra2}, /* level 21 */ + { 27, 27, 25, 9, 3,999, ZSTD_btultra2}, /* level 22 */ }, { /* for srcSize <= 256 KB */ /* W, C, H, S, L, T, strat */ - { 0, 0, 0, 0, 0, 0, ZSTD_fast }, /* level 0 - not used */ - { 18, 13, 14, 1, 6, 8, ZSTD_fast }, /* level 1 */ - { 18, 14, 13, 1, 5, 8, ZSTD_dfast }, /* level 2 */ - { 18, 16, 15, 1, 5, 8, ZSTD_dfast }, /* level 3 */ - { 18, 15, 17, 1, 5, 8, ZSTD_greedy }, /* level 4.*/ - { 18, 16, 17, 4, 5, 8, ZSTD_greedy }, /* level 5.*/ - { 18, 16, 17, 3, 5, 8, ZSTD_lazy }, /* level 6.*/ - { 18, 17, 17, 4, 4, 8, ZSTD_lazy }, /* level 7 */ - { 18, 17, 17, 4, 4, 8, ZSTD_lazy2 }, /* level 8 */ - { 18, 17, 17, 5, 4, 8, ZSTD_lazy2 }, /* level 9 */ - { 18, 17, 17, 6, 4, 8, ZSTD_lazy2 }, /* level 10 */ - { 18, 18, 17, 6, 4, 8, ZSTD_lazy2 }, /* level 11.*/ - { 18, 18, 17, 7, 4, 8, ZSTD_lazy2 }, /* level 12.*/ - { 18, 19, 17, 6, 4, 8, ZSTD_btlazy2 }, /* level 13 */ - { 18, 18, 18, 4, 4, 16, ZSTD_btopt }, /* level 14.*/ - { 18, 18, 18, 4, 3, 16, ZSTD_btopt }, /* level 15.*/ - { 18, 19, 18, 6, 3, 32, ZSTD_btopt }, /* level 16.*/ - { 18, 19, 18, 8, 3, 64, ZSTD_btopt }, /* level 17.*/ - { 18, 19, 18, 9, 3,128, ZSTD_btopt }, /* level 18.*/ - { 18, 19, 18, 10, 3,256, ZSTD_btopt }, /* level 19.*/ - { 18, 19, 18, 11, 3,512, ZSTD_btultra }, /* level 20.*/ - { 18, 19, 18, 12, 3,512, ZSTD_btultra }, /* level 21.*/ - { 18, 19, 18, 13, 3,512, ZSTD_btultra }, /* level 22.*/ + { 18, 12, 13, 1, 5, 1, ZSTD_fast }, /* base for negative levels */ + { 18, 13, 14, 1, 6, 0, ZSTD_fast }, /* level 1 */ + { 18, 14, 14, 1, 5, 0, ZSTD_dfast }, /* level 2 */ + { 18, 16, 16, 1, 4, 0, ZSTD_dfast }, /* level 3 */ + { 18, 16, 17, 2, 5, 2, ZSTD_greedy }, /* level 4.*/ + { 18, 18, 18, 3, 5, 2, ZSTD_greedy }, /* level 5.*/ + { 18, 18, 19, 3, 5, 4, ZSTD_lazy }, /* level 6.*/ + { 18, 18, 19, 4, 4, 4, ZSTD_lazy }, /* level 7 */ + { 18, 18, 19, 4, 4, 8, ZSTD_lazy2 }, /* level 8 */ + { 18, 18, 19, 5, 4, 8, ZSTD_lazy2 }, /* level 9 */ + { 18, 18, 19, 6, 4, 8, ZSTD_lazy2 }, /* level 10 */ + { 18, 18, 19, 5, 4, 12, ZSTD_btlazy2 }, /* level 11.*/ + { 18, 19, 19, 7, 4, 12, ZSTD_btlazy2 }, /* level 12.*/ + { 18, 18, 19, 4, 4, 16, ZSTD_btopt }, /* level 13 */ + { 18, 18, 19, 4, 3, 32, ZSTD_btopt }, /* level 14.*/ + { 18, 18, 19, 6, 3,128, ZSTD_btopt }, /* level 15.*/ + { 18, 19, 19, 6, 3,128, ZSTD_btultra }, /* level 16.*/ + { 18, 19, 19, 8, 3,256, ZSTD_btultra }, /* level 17.*/ + { 18, 19, 19, 6, 3,128, ZSTD_btultra2}, /* level 18.*/ + { 18, 19, 19, 8, 3,256, ZSTD_btultra2}, /* level 19.*/ + { 18, 19, 19, 10, 3,512, ZSTD_btultra2}, /* level 20.*/ + { 18, 19, 19, 12, 3,512, ZSTD_btultra2}, /* level 21.*/ + { 18, 19, 19, 13, 3,999, ZSTD_btultra2}, /* level 22.*/ }, { /* for srcSize <= 128 KB */ /* W, C, H, S, L, T, strat */ - { 17, 12, 12, 1, 7, 8, ZSTD_fast }, /* level 0 - not used */ - { 17, 12, 13, 1, 6, 8, ZSTD_fast }, /* level 1 */ - { 17, 13, 16, 1, 5, 8, ZSTD_fast }, /* level 2 */ - { 17, 16, 16, 2, 5, 8, ZSTD_dfast }, /* level 3 */ - { 17, 13, 15, 3, 4, 8, ZSTD_greedy }, /* level 4 */ - { 17, 15, 17, 4, 4, 8, ZSTD_greedy }, /* level 5 */ - { 17, 16, 17, 3, 4, 8, ZSTD_lazy }, /* level 6 */ - { 17, 15, 17, 4, 4, 8, ZSTD_lazy2 }, /* level 7 */ + { 17, 12, 12, 1, 5, 1, ZSTD_fast }, /* base for negative levels */ + { 17, 12, 13, 1, 6, 0, ZSTD_fast }, /* level 1 */ + { 17, 13, 15, 1, 5, 0, ZSTD_fast }, /* level 2 */ + { 17, 15, 16, 2, 5, 0, ZSTD_dfast }, /* level 3 */ + { 17, 17, 17, 2, 4, 0, ZSTD_dfast }, /* level 4 */ + { 17, 16, 17, 3, 4, 2, ZSTD_greedy }, /* level 5 */ + { 17, 17, 17, 3, 4, 4, ZSTD_lazy }, /* level 6 */ + { 17, 17, 17, 3, 4, 8, ZSTD_lazy2 }, /* level 7 */ { 17, 17, 17, 4, 4, 8, ZSTD_lazy2 }, /* level 8 */ { 17, 17, 17, 5, 4, 8, ZSTD_lazy2 }, /* level 9 */ { 17, 17, 17, 6, 4, 8, ZSTD_lazy2 }, /* level 10 */ - { 17, 17, 17, 7, 4, 8, ZSTD_lazy2 }, /* level 11 */ - { 17, 17, 17, 8, 4, 8, ZSTD_lazy2 }, /* level 12 */ - { 17, 18, 17, 6, 4, 8, ZSTD_btlazy2 }, /* level 13.*/ - { 17, 17, 17, 7, 3, 8, ZSTD_btopt }, /* level 14.*/ - { 17, 17, 17, 7, 3, 16, ZSTD_btopt }, /* level 15.*/ - { 17, 18, 17, 7, 3, 32, ZSTD_btopt }, /* level 16.*/ - { 17, 18, 17, 7, 3, 64, ZSTD_btopt }, /* level 17.*/ - { 17, 18, 17, 7, 3,256, ZSTD_btopt }, /* level 18.*/ - { 17, 18, 17, 8, 3,256, ZSTD_btopt }, /* level 19.*/ - { 17, 18, 17, 9, 3,256, ZSTD_btultra }, /* level 20.*/ - { 17, 18, 17, 10, 3,256, ZSTD_btultra }, /* level 21.*/ - { 17, 18, 17, 11, 3,512, ZSTD_btultra }, /* level 22.*/ + { 17, 17, 17, 5, 4, 8, ZSTD_btlazy2 }, /* level 11 */ + { 17, 18, 17, 7, 4, 12, ZSTD_btlazy2 }, /* level 12 */ + { 17, 18, 17, 3, 4, 12, ZSTD_btopt }, /* level 13.*/ + { 17, 18, 17, 4, 3, 32, ZSTD_btopt }, /* level 14.*/ + { 17, 18, 17, 6, 3,256, ZSTD_btopt }, /* level 15.*/ + { 17, 18, 17, 6, 3,128, ZSTD_btultra }, /* level 16.*/ + { 17, 18, 17, 8, 3,256, ZSTD_btultra }, /* level 17.*/ + { 17, 18, 17, 10, 3,512, ZSTD_btultra }, /* level 18.*/ + { 17, 18, 17, 5, 3,256, ZSTD_btultra2}, /* level 19.*/ + { 17, 18, 17, 7, 3,512, ZSTD_btultra2}, /* level 20.*/ + { 17, 18, 17, 9, 3,512, ZSTD_btultra2}, /* level 21.*/ + { 17, 18, 17, 11, 3,999, ZSTD_btultra2}, /* level 22.*/ }, { /* for srcSize <= 16 KB */ /* W, C, H, S, L, T, strat */ - { 14, 12, 12, 1, 7, 6, ZSTD_fast }, /* level 0 - not used */ - { 14, 14, 14, 1, 6, 6, ZSTD_fast }, /* level 1 */ - { 14, 14, 14, 1, 4, 6, ZSTD_fast }, /* level 2 */ - { 14, 14, 14, 1, 4, 6, ZSTD_dfast }, /* level 3.*/ - { 14, 14, 14, 4, 4, 6, ZSTD_greedy }, /* level 4.*/ - { 14, 14, 14, 3, 4, 6, ZSTD_lazy }, /* level 5.*/ - { 14, 14, 14, 4, 4, 6, ZSTD_lazy2 }, /* level 6 */ - { 14, 14, 14, 5, 4, 6, ZSTD_lazy2 }, /* level 7 */ - { 14, 14, 14, 6, 4, 6, ZSTD_lazy2 }, /* level 8.*/ - { 14, 15, 14, 6, 4, 6, ZSTD_btlazy2 }, /* level 9.*/ - { 14, 15, 14, 3, 3, 6, ZSTD_btopt }, /* level 10.*/ - { 14, 15, 14, 6, 3, 8, ZSTD_btopt }, /* level 11.*/ - { 14, 15, 14, 6, 3, 16, ZSTD_btopt }, /* level 12.*/ - { 14, 15, 14, 6, 3, 24, ZSTD_btopt }, /* level 13.*/ - { 14, 15, 15, 6, 3, 48, ZSTD_btopt }, /* level 14.*/ - { 14, 15, 15, 6, 3, 64, ZSTD_btopt }, /* level 15.*/ - { 14, 15, 15, 6, 3, 96, ZSTD_btopt }, /* level 16.*/ - { 14, 15, 15, 6, 3,128, ZSTD_btopt }, /* level 17.*/ - { 14, 15, 15, 6, 3,256, ZSTD_btopt }, /* level 18.*/ - { 14, 15, 15, 7, 3,256, ZSTD_btopt }, /* level 19.*/ - { 14, 15, 15, 8, 3,256, ZSTD_btultra }, /* level 20.*/ - { 14, 15, 15, 9, 3,256, ZSTD_btultra }, /* level 21.*/ - { 14, 15, 15, 10, 3,256, ZSTD_btultra }, /* level 22.*/ + { 14, 12, 13, 1, 5, 1, ZSTD_fast }, /* base for negative levels */ + { 14, 14, 15, 1, 5, 0, ZSTD_fast }, /* level 1 */ + { 14, 14, 15, 1, 4, 0, ZSTD_fast }, /* level 2 */ + { 14, 14, 15, 2, 4, 0, ZSTD_dfast }, /* level 3 */ + { 14, 14, 14, 4, 4, 2, ZSTD_greedy }, /* level 4 */ + { 14, 14, 14, 3, 4, 4, ZSTD_lazy }, /* level 5.*/ + { 14, 14, 14, 4, 4, 8, ZSTD_lazy2 }, /* level 6 */ + { 14, 14, 14, 6, 4, 8, ZSTD_lazy2 }, /* level 7 */ + { 14, 14, 14, 8, 4, 8, ZSTD_lazy2 }, /* level 8.*/ + { 14, 15, 14, 5, 4, 8, ZSTD_btlazy2 }, /* level 9.*/ + { 14, 15, 14, 9, 4, 8, ZSTD_btlazy2 }, /* level 10.*/ + { 14, 15, 14, 3, 4, 12, ZSTD_btopt }, /* level 11.*/ + { 14, 15, 14, 4, 3, 24, ZSTD_btopt }, /* level 12.*/ + { 14, 15, 14, 5, 3, 32, ZSTD_btultra }, /* level 13.*/ + { 14, 15, 15, 6, 3, 64, ZSTD_btultra }, /* level 14.*/ + { 14, 15, 15, 7, 3,256, ZSTD_btultra }, /* level 15.*/ + { 14, 15, 15, 5, 3, 48, ZSTD_btultra2}, /* level 16.*/ + { 14, 15, 15, 6, 3,128, ZSTD_btultra2}, /* level 17.*/ + { 14, 15, 15, 7, 3,256, ZSTD_btultra2}, /* level 18.*/ + { 14, 15, 15, 8, 3,256, ZSTD_btultra2}, /* level 19.*/ + { 14, 15, 15, 8, 3,512, ZSTD_btultra2}, /* level 20.*/ + { 14, 15, 15, 9, 3,512, ZSTD_btultra2}, /* level 21.*/ + { 14, 15, 15, 10, 3,999, ZSTD_btultra2}, /* level 22.*/ }, }; -#if defined(ZSTD_DEBUG) && (ZSTD_DEBUG>=1) -/* This function just controls - * the monotonic memory budget increase of ZSTD_defaultCParameters[0]. - * Run once, on first ZSTD_getCParams() usage, if ZSTD_DEBUG is enabled - */ -MEM_STATIC void ZSTD_check_compressionLevel_monotonicIncrease_memoryBudget(void) +/*! ZSTD_getCParams_internal() : + * @return ZSTD_compressionParameters structure for a selected compression level, srcSize and dictSize. + * Note: srcSizeHint 0 means 0, use ZSTD_CONTENTSIZE_UNKNOWN for unknown. + * Use dictSize == 0 for unknown or unused. */ +static ZSTD_compressionParameters ZSTD_getCParams_internal(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) { - int level; - for (level=1; level<ZSTD_maxCLevel(); level++) { - ZSTD_compressionParameters const c1 = ZSTD_defaultCParameters[0][level]; - ZSTD_compressionParameters const c2 = ZSTD_defaultCParameters[0][level+1]; - assert(c1.windowLog <= c2.windowLog); -# define ZSTD_TABLECOST(h,c) ((1<<(h)) + (1<<(c))) - assert(ZSTD_TABLECOST(c1.hashLog, c1.chainLog) <= ZSTD_TABLECOST(c2.hashLog, c2.chainLog)); + int const unknown = srcSizeHint == ZSTD_CONTENTSIZE_UNKNOWN; + size_t const addedSize = unknown && dictSize > 0 ? 500 : 0; + U64 const rSize = unknown && dictSize == 0 ? ZSTD_CONTENTSIZE_UNKNOWN : srcSizeHint+dictSize+addedSize; + U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB); + int row = compressionLevel; + DEBUGLOG(5, "ZSTD_getCParams_internal (cLevel=%i)", compressionLevel); + if (compressionLevel == 0) row = ZSTD_CLEVEL_DEFAULT; /* 0 == default */ + if (compressionLevel < 0) row = 0; /* entry 0 is baseline for fast mode */ + if (compressionLevel > ZSTD_MAX_CLEVEL) row = ZSTD_MAX_CLEVEL; + { ZSTD_compressionParameters cp = ZSTD_defaultCParameters[tableID][row]; + if (compressionLevel < 0) cp.targetLength = (unsigned)(-compressionLevel); /* acceleration factor */ + /* refine parameters based on srcSize & dictSize */ + return ZSTD_adjustCParams_internal(cp, srcSizeHint, dictSize); } } -#endif /*! ZSTD_getCParams() : -* @return ZSTD_compressionParameters structure for a selected compression level, `srcSize` and `dictSize`. -* Size values are optional, provide 0 if not known or unused */ + * @return ZSTD_compressionParameters structure for a selected compression level, srcSize and dictSize. + * Size values are optional, provide 0 if not known or unused */ ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) { - size_t const addedSize = srcSizeHint ? 0 : 500; - U64 const rSize = srcSizeHint+dictSize ? srcSizeHint+dictSize+addedSize : (U64)-1; - U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB); /* intentional underflow for srcSizeHint == 0 */ - -#if defined(ZSTD_DEBUG) && (ZSTD_DEBUG>=1) - static int g_monotonicTest = 1; - if (g_monotonicTest) { - ZSTD_check_compressionLevel_monotonicIncrease_memoryBudget(); - g_monotonicTest=0; - } -#endif - - if (compressionLevel <= 0) compressionLevel = ZSTD_CLEVEL_DEFAULT; /* 0 == default; no negative compressionLevel yet */ - if (compressionLevel > ZSTD_MAX_CLEVEL) compressionLevel = ZSTD_MAX_CLEVEL; - { ZSTD_compressionParameters const cp = ZSTD_defaultCParameters[tableID][compressionLevel]; - return ZSTD_adjustCParams_internal(cp, srcSizeHint, dictSize); } - + if (srcSizeHint == 0) srcSizeHint = ZSTD_CONTENTSIZE_UNKNOWN; + return ZSTD_getCParams_internal(compressionLevel, srcSizeHint, dictSize); } /*! ZSTD_getParams() : -* same as ZSTD_getCParams(), but @return a `ZSTD_parameters` object (instead of `ZSTD_compressionParameters`). -* All fields of `ZSTD_frameParameters` are set to default (0) */ -ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) { + * same idea as ZSTD_getCParams() + * @return a `ZSTD_parameters` structure (instead of `ZSTD_compressionParameters`). + * Fields of `ZSTD_frameParameters` are set to default values */ +static ZSTD_parameters ZSTD_getParams_internal(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) { ZSTD_parameters params; - ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, srcSizeHint, dictSize); + ZSTD_compressionParameters const cParams = ZSTD_getCParams_internal(compressionLevel, srcSizeHint, dictSize); + DEBUGLOG(5, "ZSTD_getParams (cLevel=%i)", compressionLevel); memset(¶ms, 0, sizeof(params)); params.cParams = cParams; + params.fParams.contentSizeFlag = 1; return params; } + +/*! ZSTD_getParams() : + * same idea as ZSTD_getCParams() + * @return a `ZSTD_parameters` structure (instead of `ZSTD_compressionParameters`). + * Fields of `ZSTD_frameParameters` are set to default values */ +ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) { + if (srcSizeHint == 0) srcSizeHint = ZSTD_CONTENTSIZE_UNKNOWN; + return ZSTD_getParams_internal(compressionLevel, srcSizeHint, dictSize); +} |