From 7682e49d0a4c5d5b725962c82d0762da9e32eba8 Mon Sep 17 00:00:00 2001 From: Yann Collet Date: Sun, 31 Jan 2016 23:45:35 +0100 Subject: [PATCH 01/12] added buffer interface to dictBuilder --- dictBuilder/dibcli.c | 13 ++- dictBuilder/dictBuilder.c | 234 ++++++++++++++++++++++---------------- dictBuilder/dictBuilder.h | 55 +++++++-- 3 files changed, 193 insertions(+), 109 deletions(-) diff --git a/dictBuilder/dibcli.c b/dictBuilder/dibcli.c index f93e9707a..8566c6245 100644 --- a/dictBuilder/dibcli.c +++ b/dictBuilder/dibcli.c @@ -184,7 +184,7 @@ int main(int argCount, const char** argv) if (!strcmp(argument, "--verbose")) { g_displayLevel++; if (g_displayLevel<3) g_displayLevel=3; continue; } if (!strcmp(argument, "--quiet")) { g_displayLevel--; continue; } if (!strcmp(argument, "--maxdict")) { nextArgumentIsMaxDict=1; continue; } - if (!strcmp(argument, "--fast")) { selectionLevel=0; cLevel=1; continue; } + if (!strcmp(argument, "--fast")) { selectionLevel=1; cLevel=1; continue; } /* Decode commands (note : aggregated commands are allowed) */ if (argument[0]=='-') { @@ -247,8 +247,15 @@ int main(int argCount, const char** argv) } /* building ... */ - DiB_setNotificationLevel(g_displayLevel); - operationResult = DiB_trainDictionary(dictFileName, maxDictSize, selectionLevel, cLevel, filenameTable, filenameIdx); + { + DiB_params_t param; + param.selectivityLevel = selectionLevel; + param.compressionLevel = cLevel; + DiB_setNotificationLevel(g_displayLevel); + operationResult = DiB_trainFromFiles(dictFileName, maxDictSize, + filenameTable, filenameIdx, + param); + } if (main_pause) waitEnter(); free((void*)filenameTable); diff --git a/dictBuilder/dictBuilder.c b/dictBuilder/dictBuilder.c index fed9b18df..944853fbb 100644 --- a/dictBuilder/dictBuilder.c +++ b/dictBuilder/dictBuilder.c @@ -43,14 +43,14 @@ /*-************************************* * Includes ***************************************/ -#include /* malloc, free */ -#include /* memset */ -#include /* fprintf, fopen, ftello64 */ -#include /* stat64 */ -#include /* stat64 */ -#include /* clock */ +#include /* malloc, free */ +#include /* memset */ +#include /* fprintf, fopen, ftello64 */ +#include /* stat64 */ +#include /* stat64 */ +#include /* clock */ -#include "mem.h" /* read */ +#include "mem.h" /* read */ #include "error_private.h" #include "divsufsort.h" #include "dictBuilder.h" @@ -58,17 +58,13 @@ #include "huff0_static.h" -/* ************************************* +/*-************************************* * Compiler specifics ***************************************/ #if !defined(S_ISREG) # define S_ISREG(x) (((x) & S_IFMT) == S_IFREG) #endif -#ifdef _MSC_VER -#define snprintf sprintf_s -#endif - /*-************************************* * Constants @@ -87,6 +83,9 @@ static const size_t maxMemory = (sizeof(size_t) == 4) ? (2 GB - 64 MB) : ((size_ #define MINRATIO 4 static const U32 g_compressionLevel_default = 5; +static const U32 g_selectivity_default = 9; +static const size_t g_provision_entropySize = 200; +static const size_t g_min_fast_dictContent = 192; /*-************************************* @@ -146,6 +145,10 @@ static unsigned DiB_GetMilliSpan(clock_t nPrevious) return nSpan; } +unsigned DiB_isError(size_t errorCode) { return ERR_isError(errorCode); } + +const char* DiB_getErrorName(size_t errorCode) { return ERR_getErrorName(errorCode); } + /* ******************************************************** * File related operations @@ -563,9 +566,8 @@ static U32 DiB_dictSize(const dictItem* dictList) static void DiB_trainBuffer(dictItem* dictList, U32 dictListSize, const void* const buffer, const size_t bufferSize, /* buffer must end with noisy guard band */ - const char* displayName, - const size_t* fileSizes, unsigned nbFiles, unsigned maxDictSize, - U32 shiftRatio) + const size_t* fileSizes, unsigned nbFiles, + U32 shiftRatio, unsigned maxDictSize) { saidx_t* const suffix0 = (saidx_t*)malloc((bufferSize+2)*sizeof(*suffix0)); saidx_t* const suffix = suffix0+1; @@ -583,7 +585,7 @@ static void DiB_trainBuffer(dictItem* dictList, U32 dictListSize, memset(doneMarks, 0, bufferSize+16); /* sort */ - DISPLAYLEVEL(2, "sorting %s ...\n", displayName); + DISPLAYLEVEL(2, "sorting %u files of total size %u MB ...\n", nbFiles, (U32)(bufferSize>>20)); errorCode = divsufsort((const sauchar_t*)buffer, suffix, (saidx_t)bufferSize); if (errorCode != 0) EXM_THROW(2, "sort failed"); suffix[bufferSize] = (saidx_t)bufferSize; /* leads into noise */ @@ -699,7 +701,7 @@ static void DiB_countEStats(EStats_ress_t esr, #define OFFCODE_MAX 18 static size_t DiB_analyzeEntropy(void* dstBuffer, size_t maxDstSize, unsigned compressionLevel, - const void* srcBuffer, size_t* fileSizes, unsigned nbFiles, + const void* srcBuffer, const size_t* fileSizes, unsigned nbFiles, const void* dictBuffer, size_t dictBufferSize) { U32 countLit[256]; @@ -793,8 +795,7 @@ static size_t DiB_analyzeEntropy(void* dstBuffer, size_t maxDstSize, static void DiB_saveDict(const char* dictFileName, - const void* buff1, size_t buff1Size, - const void* buff2, size_t buff2Size) + const void* buff, size_t buffSize) { FILE* f; size_t n; @@ -802,11 +803,8 @@ static void DiB_saveDict(const char* dictFileName, f = fopen(dictFileName, "wb"); if (f==NULL) EXM_THROW(3, "cannot open %s ", dictFileName); - n = fwrite(buff1, 1, buff1Size, f); - if (n!=buff1Size) EXM_THROW(4, "%s : write error", dictFileName) - - n = fwrite(buff2, 1, buff2Size, f); - if (n!=buff2Size) EXM_THROW(4, "%s : write error", dictFileName) + n = fwrite(buff, 1, buffSize, f); + if (n!=buffSize) EXM_THROW(4, "%s : write error", dictFileName) n = (size_t)fclose(f); if (n!=0) EXM_THROW(5, "%s : flush error", dictFileName) @@ -853,46 +851,35 @@ static size_t DiB_fastSampling(void* dictBuffer, size_t dictSize, } -int DiB_trainDictionary(const char* dictFileName, unsigned maxDictSize, - unsigned shiftRatio, unsigned compressionLevel, - const char** fileNamesTable, unsigned nbFiles) +static size_t DiB_trainFromBuffer_internal( + void* dictBuffer, size_t maxDictSize, + const void* samplesBuffer, const size_t* sampleSizes, unsigned nbSamples, + DiB_params_t params) { - void* srcBuffer; - size_t benchedSize; - size_t* fileSizes = (size_t*)malloc(nbFiles * sizeof(size_t)); - unsigned long long totalSizeToLoad = DiB_getTotalFileSize(fileNamesTable, nbFiles); - const U32 dictListSize = MAX( MAX(DICTLISTSIZE, nbFiles), maxDictSize/16); + const U32 dictListSize = MAX( MAX(DICTLISTSIZE, nbSamples), maxDictSize/16); dictItem* dictList = (dictItem*)malloc(dictListSize * sizeof(*dictList)); - char mfName[20] = {0}; - const char* displayName = NULL; + unsigned selectivity = params.selectivityLevel; + unsigned compressionLevel = params.compressionLevel; + size_t targetDictSize = maxDictSize - g_provision_entropySize; + size_t sBuffSize; + size_t dictSize = 0; + + /* checks */ + if (maxDictSize <= g_provision_entropySize + g_min_fast_dictContent) return ERROR(dstSize_tooSmall); /* init */ - benchedSize = DiB_findMaxMem(totalSizeToLoad * MEMMULT) / MEMMULT; - if ((unsigned long long)benchedSize > totalSizeToLoad) benchedSize = (size_t)totalSizeToLoad; - if (benchedSize < totalSizeToLoad) - DISPLAYLEVEL(1, "Not enough memory; training on %u MB only...\n", (unsigned)(benchedSize >> 20)); - - /* Memory allocation & restrictions */ - srcBuffer = malloc(benchedSize+NOISELENGTH); /* + noise */ - if ((!fileSizes) || (!srcBuffer) || (!dictList)) EXM_THROW(12, "not enough memory for DiB_trainFiles"); /* should not happen */ + { unsigned u; for (u=0, sBuffSize=0; u0) - { - snprintf (mfName, sizeof(mfName), " %u files", nbFiles); - if (nbFiles > 1) displayName = mfName; - else displayName = fileNamesTable[0]; - + /* select stripes */ + if (selectivity>1) { DiB_trainBuffer(dictList, dictListSize, - srcBuffer, benchedSize, - displayName, - fileSizes, nbFiles, maxDictSize, - shiftRatio); + samplesBuffer, sBuffSize, + sampleSizes, nbSamples, + selectivity, targetDictSize); /* display best matches */ if (g_displayLevel>= 3) { @@ -907,72 +894,127 @@ int DiB_trainDictionary(const char* dictFileName, unsigned maxDictSize, U32 d = MIN(40, l); DISPLAYLEVEL(3, "%3u:%3u bytes at pos %8u, savings %7u bytes |", u, l, p, dictList[u].savings); - DiB_printHex(3, (char*)srcBuffer+p, d); + DiB_printHex(3, (const char*)samplesBuffer+p, d); DISPLAYLEVEL(3, "| \n"); } } } /* create dictionary */ { - void* dictContent; U32 dictContentSize = DiB_dictSize(dictList); - void* dictHeader; - size_t dictHeaderSize, hSize, addedContentLength; + size_t hSize; BYTE* ptr; U32 u; - /* build dict */ - #define EBSIZE (2 KB) - dictHeaderSize = EBSIZE; - dictHeader = malloc(dictHeaderSize); - dictContent = malloc(maxDictSize); - if (!dictHeader || !dictContent) EXM_THROW(2, "not enough memory"); - /* build dict content */ - ptr = (BYTE*)dictContent + maxDictSize; + ptr = (BYTE*)dictBuffer + maxDictSize; for (u=1; upos; u++) { U32 l = dictList[u].length; ptr -= l; - memcpy(ptr, (char*)srcBuffer+dictList[u].pos, l); + if (ptr<(BYTE*)dictBuffer) return ERROR(GENERIC); /* should not happen */ + memcpy(ptr, (const char*)samplesBuffer+dictList[u].pos, l); } /* fast mode dict content */ - if (shiftRatio==0) { /* note could also be used to complete a dictionary, but not necessarily better */ - addedContentLength = ptr-(BYTE*)dictContent; - DISPLAYLEVEL(2, "\r%70s\r", ""); /* clean display line */ - DISPLAYLEVEL(2, "Adding %u KB from fast sampling \n", (U32)(addedContentLength>>10)); - addedContentLength = DiB_fastSampling(dictContent, addedContentLength, srcBuffer, benchedSize); - if (!ERR_isError(addedContentLength)) - ptr -= addedContentLength, dictContentSize += addedContentLength; + if (selectivity==1) { /* note could also be used to complete a dictionary, but not necessarily better */ + DISPLAYLEVEL(3, "\r%70s\r", ""); /* clean display line */ + DISPLAYLEVEL(3, "Adding %u KB with fast sampling \n", (U32)(targetDictSize>>10)); + dictContentSize = DiB_fastSampling((char*)dictBuffer + g_provision_entropySize, + targetDictSize, samplesBuffer, sBuffSize); } - /* dictionary header */ - MEM_writeLE32(dictHeader, ZSTD_DICT_MAGIC); + /* dictionary header */ + MEM_writeLE32(dictBuffer, ZSTD_DICT_MAGIC); hSize = 4; - dictHeaderSize -= 4; /* entropic tables */ + DISPLAYLEVEL(2, "\r%70s\r", ""); /* clean display line */ DISPLAYLEVEL(2, "statistics ... \n"); - hSize += DiB_analyzeEntropy((char*)dictHeader+4, dictHeaderSize, + hSize += DiB_analyzeEntropy((char*)dictBuffer+4, maxDictSize-4, compressionLevel, - srcBuffer, fileSizes, nbFiles, - ptr, dictContentSize); + samplesBuffer, sampleSizes, nbSamples, + (char*)dictBuffer + maxDictSize - dictContentSize, dictContentSize); - /* save dict */ - { - size_t dictSize = hSize + dictContentSize; - DISPLAYLEVEL(2, "Save dictionary of size %u into file %s \n", (U32)dictSize, dictFileName); - DiB_saveDict(dictFileName, dictHeader, hSize, ptr, dictContentSize); - //DiB_saveDict(dictFileName, NULL, 0, dictContent, dictContentSize); // content only - } - /* clean */ - free(dictHeader); - free(dictContent); + if (hSize + dictContentSize < maxDictSize) + memmove((char*)dictBuffer + hSize, (char*)dictBuffer + maxDictSize - dictContentSize, dictContentSize); + dictSize = MIN(maxDictSize, hSize+dictContentSize); } /* clean up */ - free(srcBuffer); - free(fileSizes); free(dictList); - return 0; + return dictSize; } + +/* issue : samplesBuffer need to be followed by a noisy guard band. +* work around : duplicate the buffer, and add the noise ? */ +size_t DiB_trainFromBuffer(void* dictBuffer, size_t maxDictSize, + const void* samplesBuffer, const size_t* sampleSizes, unsigned nbSamples, + DiB_params_t params) +{ + size_t sBuffSize; + void* newBuff; + size_t result; + + { unsigned u; for (u=0, sBuffSize=0; u totalSizeToLoad) benchedSize = (size_t)totalSizeToLoad; + if (benchedSize < totalSizeToLoad) + DISPLAYLEVEL(1, "Not enough memory; training on %u MB only...\n", (unsigned)(benchedSize >> 20)); + + /* Memory allocation & restrictions */ + srcBuffer = malloc(benchedSize+NOISELENGTH); /* + noise */ + if ((!fileSizes) || (!srcBuffer) || (!dictBuffer)) EXM_THROW(12, "not enough memory for DiB_trainFiles"); /* should not happen */ + + /* Load input buffer */ + DiB_loadFiles(srcBuffer, benchedSize, fileSizes, fileNamesTable, nbFiles); + DiB_fillNoise((char*)srcBuffer + benchedSize, NOISELENGTH); /* guard band, for end of buffer condition */ + + /* call buffer version */ + dictSize = DiB_trainFromBuffer_internal(dictBuffer, maxDictSize, + srcBuffer, fileSizes, nbFiles, + params); + if (DiB_isError(dictSize)) + { + DISPLAYLEVEL(1, "dictionary training failed : %s", DiB_getErrorName(dictSize)); /* should not happen */ + result = 1; + goto _cleanup; + } + + /* save dict */ + DISPLAYLEVEL(2, "Save dictionary of size %u into file %s \n", (U32)dictSize, dictFileName); + DiB_saveDict(dictFileName, dictBuffer, dictSize); + + /* clean up */ +_cleanup: + free(srcBuffer); + free(dictBuffer); + free(fileSizes); + return result; +} diff --git a/dictBuilder/dictBuilder.h b/dictBuilder/dictBuilder.h index 3c52c7daa..a022583eb 100644 --- a/dictBuilder/dictBuilder.h +++ b/dictBuilder/dictBuilder.h @@ -26,6 +26,9 @@ /* This library is designed for a single-threaded console application. * It exit() and printf() into stderr when it encounters an error condition. */ +#ifndef DICTBUILDER_H_001 +#define DICTBUILDER_H_001 + /*-************************************* * Version ***************************************/ @@ -36,24 +39,56 @@ unsigned DiB_versionNumber (void); +/*-************************************* +* Public type +***************************************/ +typedef struct { + unsigned selectivityLevel; /* 0 means default; larger => bigger selection => larger dictionary */ + unsigned compressionLevel; /* 0 means default; target a specific zstd compression level */ +} DiB_params_t; + + /*-************************************* * Public functions ***************************************/ -/*! DiB_trainDictionary - Train a dictionary from a set of files provided by @fileNamesTable - Resulting dictionary is written in file @dictFileName. - @selectivityLevel change criteria for insertion into the dictionary (more => bigger selection => larger dictionary) - @compressionLevel can be used to target a specific compression level of zstd. 0 means "default". - @result : 0 == ok +/*! DiB_trainFromBuffer + Train a dictionary from a memory buffer @samplesBuffer + where @nbSamples samples have been stored concatenated. + Each sample size is provided into an orderly table @sampleSizes. + Resulting dictionary will be saved into @dictBuffer. + @parameters is optional and can be provided with 0 values to mean "default". + @result : size of dictionary stored into @dictBuffer (<= @dictBufferSize) + or an error code, which can be tested by DiB_isError(). + note : DiB_trainFromBuffer() will send notifications into stderr if instructed to, using DiB_setNotificationLevel() */ -int DiB_trainDictionary(const char* dictFileName, unsigned maxDictSize, - unsigned selectivityLevel, unsigned compressionLevel, - const char** fileNamesTable, unsigned nbFiles); +size_t DiB_trainFromBuffer(void* dictBuffer, size_t dictBufferSize, + const void* samplesBuffer, const size_t* sampleSizes, unsigned nbSamples, + DiB_params_t parameters); +/*! DiB_trainFromFiles + Train a dictionary from a set of files provided by @fileNamesTable + Resulting dictionary is written into file @dictFileName. + @parameters is optional and can be provided with 0 values. + @result : 0 == ok. Any other : error. +*/ +int DiB_trainFromFiles(const char* dictFileName, unsigned maxDictSize, + const char** fileNamesTable, unsigned nbFiles, + DiB_params_t parameters); + + +/*-************************************* +* Helper functions +***************************************/ +unsigned DiB_isError(size_t errorCode); +const char* DiB_getErrorName(size_t errorCode); + /*! DiB_setNotificationLevel Set amount of notification to be displayed on the console. - 0 = no console notification (default). + default initial value : 0 = no console notification. Note : not thread-safe (use a global constant) */ void DiB_setNotificationLevel(unsigned l); + + +#endif From b35c4648bf73dcedf95a4a2117a97f28b2aefb84 Mon Sep 17 00:00:00 2001 From: Yann Collet Date: Mon, 1 Feb 2016 00:03:10 +0100 Subject: [PATCH 02/12] fixed clang warnings --- dictBuilder/dictBuilder.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/dictBuilder/dictBuilder.c b/dictBuilder/dictBuilder.c index 944853fbb..d8b2bdb30 100644 --- a/dictBuilder/dictBuilder.c +++ b/dictBuilder/dictBuilder.c @@ -856,7 +856,7 @@ static size_t DiB_trainFromBuffer_internal( const void* samplesBuffer, const size_t* sampleSizes, unsigned nbSamples, DiB_params_t params) { - const U32 dictListSize = MAX( MAX(DICTLISTSIZE, nbSamples), maxDictSize/16); + const U32 dictListSize = MAX( MAX(DICTLISTSIZE, nbSamples), (U32)(maxDictSize/16)); dictItem* dictList = (dictItem*)malloc(dictListSize * sizeof(*dictList)); unsigned selectivity = params.selectivityLevel; unsigned compressionLevel = params.compressionLevel; @@ -879,7 +879,7 @@ static size_t DiB_trainFromBuffer_internal( DiB_trainBuffer(dictList, dictListSize, samplesBuffer, sBuffSize, sampleSizes, nbSamples, - selectivity, targetDictSize); + selectivity, (U32)targetDictSize); /* display best matches */ if (g_displayLevel>= 3) { @@ -918,7 +918,7 @@ static size_t DiB_trainFromBuffer_internal( if (selectivity==1) { /* note could also be used to complete a dictionary, but not necessarily better */ DISPLAYLEVEL(3, "\r%70s\r", ""); /* clean display line */ DISPLAYLEVEL(3, "Adding %u KB with fast sampling \n", (U32)(targetDictSize>>10)); - dictContentSize = DiB_fastSampling((char*)dictBuffer + g_provision_entropySize, + dictContentSize = (U32)DiB_fastSampling((char*)dictBuffer + g_provision_entropySize, targetDictSize, samplesBuffer, sBuffSize); } From 82d71fe89f223ce0fe5deb640780533db5bc2781 Mon Sep 17 00:00:00 2001 From: Yann Collet Date: Tue, 2 Feb 2016 00:29:00 +0100 Subject: [PATCH 03/12] added wildcard '*' expansion in DictBuilder under Visual --- visual/2013/dictBuilder/dictBuilder.vcxproj | 1 + 1 file changed, 1 insertion(+) diff --git a/visual/2013/dictBuilder/dictBuilder.vcxproj b/visual/2013/dictBuilder/dictBuilder.vcxproj index ad69449f6..0260102b5 100644 --- a/visual/2013/dictBuilder/dictBuilder.vcxproj +++ b/visual/2013/dictBuilder/dictBuilder.vcxproj @@ -99,6 +99,7 @@ Console true + setargv.obj;kernel32.lib;user32.lib;gdi32.lib;winspool.lib;comdlg32.lib;advapi32.lib;shell32.lib;ole32.lib;oleaut32.lib;uuid.lib;odbc32.lib;odbccp32.lib;%(AdditionalDependencies) From 2ce4923238348cec3535aafaeec687ebb1fa4f30 Mon Sep 17 00:00:00 2001 From: Yann Collet Date: Tue, 2 Feb 2016 14:36:49 +0100 Subject: [PATCH 04/12] modified buffer management rules --- lib/zstd_compress.c | 62 ++++++++++++++++++++------------------------- programs/bench.c | 3 ++- programs/datagen.c | 42 ++++++++++++------------------ programs/fileio.c | 59 +++++++++++++----------------------------- 4 files changed, 63 insertions(+), 103 deletions(-) diff --git a/lib/zstd_compress.c b/lib/zstd_compress.c index 6e3d6ca5a..7252d750b 100644 --- a/lib/zstd_compress.c +++ b/lib/zstd_compress.c @@ -114,6 +114,7 @@ struct ZSTD_CCtx_s U32 dictLimit; /* below that point, need extDict */ U32 lowLimit; /* below that point, no more data */ U32 nextToUpdate; /* index from which to continue dictionary update */ + U32 loadedDictEnd; U32 stage; ZSTD_parameters params; void* workSpace; @@ -217,6 +218,7 @@ static size_t ZSTD_resetCCtx_advanced (ZSTD_CCtx* zc, zc->seqStore.dumpsStart = zc->seqStore.matchLengthStart + (blockSize>>2); zc->hbSize = 0; zc->stage = 0; + zc->loadedDictEnd = 0; return 0; } @@ -906,7 +908,7 @@ static size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls) } -/* ************************************* +/*-************************************* * Fast Scan ***************************************/ #define FILLHASHSTEP 3 @@ -916,7 +918,7 @@ static void ZSTD_fillHashTable (ZSTD_CCtx* zc, const void* end, const U32 mls) const U32 hBits = zc->params.hashLog; const BYTE* const base = zc->base; const BYTE* ip = base + zc->nextToUpdate; - const BYTE* const iend = (const BYTE*) end; + const BYTE* const iend = ((const BYTE*)end) - 8; while(ip <= iend) { hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip - base); @@ -1885,47 +1887,41 @@ static ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, int static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc, void* dst, size_t maxDstSize, const void* src, size_t srcSize) { ZSTD_blockCompressor blockCompressor = ZSTD_selectBlockCompressor(zc->params.strategy, zc->lowLimit < zc->dictLimit); - if (srcSize < MIN_CBLOCK_SIZE+3) return 0; /* don't even attempt compression below a certain srcSize */ + if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) return 0; /* don't even attempt compression below a certain srcSize */ blockCompressor(zc, src, srcSize); return ZSTD_compressSequences(zc, dst, maxDstSize, srcSize); } -static size_t ZSTD_compress_generic (ZSTD_CCtx* ctxPtr, +static size_t ZSTD_compress_generic (ZSTD_CCtx* zc, void* dst, size_t maxDstSize, const void* src, size_t srcSize) { - size_t blockSize = ctxPtr->blockSize; + size_t blockSize = zc->blockSize; size_t remaining = srcSize; const BYTE* ip = (const BYTE*)src; BYTE* const ostart = (BYTE*)dst; BYTE* op = ostart; - const U32 maxDist = 1 << ctxPtr->params.windowLog; + const U32 maxDist = 1 << zc->params.windowLog; - while (remaining) - { + while (remaining) { size_t cSize; - if (maxDstSize < 3 + MIN_CBLOCK_SIZE) return ERROR(dstSize_tooSmall); /* not enough space to store compressed block */ + if (maxDstSize < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE) return ERROR(dstSize_tooSmall); /* not enough space to store compressed block */ if (remaining < blockSize) blockSize = remaining; - if ((U32)(ip+blockSize - (ctxPtr->base + ctxPtr->lowLimit)) > maxDist) - { - /* respect windowLog contract */ - ctxPtr->lowLimit = (U32)(ip+blockSize - ctxPtr->base) - maxDist; - if (ctxPtr->dictLimit < ctxPtr->lowLimit) ctxPtr->dictLimit = ctxPtr->lowLimit; + if ((U32)(ip+blockSize - zc->base) > zc->loadedDictEnd + maxDist) { /* enforce maxDist */ + zc->lowLimit = (U32)(ip+blockSize - zc->base) - maxDist; + if (zc->dictLimit < zc->lowLimit) zc->dictLimit = zc->lowLimit; } - cSize = ZSTD_compressBlock_internal(ctxPtr, op+3, maxDstSize-3, ip, blockSize); + cSize = ZSTD_compressBlock_internal(zc, op+ZSTD_blockHeaderSize, maxDstSize-ZSTD_blockHeaderSize, ip, blockSize); if (ZSTD_isError(cSize)) return cSize; - if (cSize == 0) - { - cSize = ZSTD_noCompressBlock(op, maxDstSize, ip, blockSize); /* block is not compressible */ + if (cSize == 0) { /* block is not compressible */ + cSize = ZSTD_noCompressBlock(op, maxDstSize, ip, blockSize); if (ZSTD_isError(cSize)) return cSize; - } - else - { + } else { op[0] = (BYTE)(cSize>>16); op[1] = (BYTE)(cSize>>8); op[2] = (BYTE)cSize; @@ -1951,8 +1947,7 @@ static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* zc, const BYTE* const ip = (const BYTE*) src; size_t hbSize = 0; - if (frame && (zc->stage==0)) - { + if (frame && (zc->stage==0)) { hbSize = zc->hbSize; if (dstSize <= hbSize) return ERROR(dstSize_tooSmall); zc->stage = 1; @@ -1962,8 +1957,7 @@ static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* zc, } /* Check if blocks follow each other */ - if (src != zc->nextSrc) - { + if (src != zc->nextSrc) { /* not contiguous */ size_t delta = zc->nextSrc - ip; zc->lowLimit = zc->dictLimit; @@ -1975,8 +1969,7 @@ static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* zc, } /* preemptive overflow correction */ - if (zc->lowLimit > (1<<30)) - { + if (zc->lowLimit > (1<<30)) { U32 btplus = (zc->params.strategy == ZSTD_btlazy2); U32 contentMask = (1 << (zc->params.contentLog - btplus)) - 1; U32 newLowLimit = zc->lowLimit & contentMask; /* preserve position % contentSize */ @@ -1991,8 +1984,7 @@ static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* zc, } /* if input and dictionary overlap : reduce dictionary (presumed modified by input) */ - if ((ip+srcSize > zc->dictBase + zc->lowLimit) && (ip < zc->dictBase + zc->dictLimit)) - { + if ((ip+srcSize > zc->dictBase + zc->lowLimit) && (ip < zc->dictBase + zc->dictLimit)) { zc->lowLimit = (U32)(ip + srcSize - zc->dictBase); if (zc->lowLimit > zc->dictLimit) zc->lowLimit = zc->dictLimit; } @@ -2034,6 +2026,7 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t zc->dictBase = zc->base; zc->base += ip - zc->nextSrc; zc->nextToUpdate = zc->dictLimit; + zc->loadedDictEnd = (U32)(iend - zc->base); zc->nextSrc = iend; if (srcSize <= 8) return 0; @@ -2041,7 +2034,7 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t switch(zc->params.strategy) { case ZSTD_fast: - ZSTD_fillHashTable (zc, iend-8, zc->params.searchLength); + ZSTD_fillHashTable (zc, iend, zc->params.searchLength); break; case ZSTD_greedy: @@ -2052,13 +2045,13 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t case ZSTD_btlazy2: ZSTD_updateTree(zc, iend-8, iend, 1 << zc->params.searchLog, zc->params.searchLength); - zc->nextToUpdate = (U32)(iend - zc->base); break; default: return ERROR(GENERIC); /* strategy doesn't exist; impossible */ } + zc->nextToUpdate = zc->loadedDictEnd; return 0; } @@ -2112,8 +2105,7 @@ static size_t ZSTD_loadDictEntropyStats(ZSTD_CCtx* zc, const void* dict, size_t static size_t ZSTD_compress_insertDictionary(ZSTD_CCtx* zc, const void* dict, size_t dictSize) { - if (dict && dictSize) - { + if (dict && (dictSize>4)) { U32 magic = MEM_readLE32(dict); size_t eSize; if (magic != ZSTD_DICT_MAGIC) @@ -2151,7 +2143,7 @@ size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* zc, size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* zc, const void* dict, size_t dictSize, int compressionLevel) { - return ZSTD_compressBegin_advanced(zc, dict, dictSize, ZSTD_getParams(compressionLevel, 0)); + return ZSTD_compressBegin_advanced(zc, dict, dictSize, ZSTD_getParams(compressionLevel, 128 KB)); } size_t ZSTD_compressBegin(ZSTD_CCtx* zc, int compressionLevel) @@ -2235,7 +2227,7 @@ size_t ZSTD_compress_advanced (ZSTD_CCtx* ctx, size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx, void* dst, size_t maxDstSize, const void* src, size_t srcSize, const void* dict, size_t dictSize, int compressionLevel) { - return ZSTD_compress_advanced(ctx, dst, maxDstSize, src, srcSize, dict, dictSize, ZSTD_getParams(compressionLevel, srcSize+dictSize)); + return ZSTD_compress_advanced(ctx, dst, maxDstSize, src, srcSize, dict, dictSize, ZSTD_getParams(compressionLevel, srcSize)); } size_t ZSTD_compressCCtx (ZSTD_CCtx* ctx, void* dst, size_t maxDstSize, const void* src, size_t srcSize, int compressionLevel) diff --git a/programs/bench.c b/programs/bench.c index d24991431..1c776c4e6 100644 --- a/programs/bench.c +++ b/programs/bench.c @@ -212,6 +212,7 @@ typedef struct } blockParam_t; #define MIN(a,b) ((a)<(b) ? (a) : (b)) +#define MAX(a,b) ((a)>(b) ? (a) : (b)) static int BMK_benchMem(const void* srcBuffer, size_t srcSize, const char* displayName, int cLevel, @@ -289,7 +290,7 @@ static int BMK_benchMem(const void* srcBuffer, size_t srcSize, while (BMK_GetMilliStart() == milliTime); milliTime = BMK_GetMilliStart(); while (BMK_GetMilliSpan(milliTime) < TIMELOOP) { - ZSTD_compressBegin_advanced(refCtx, dictBuffer, dictBufferSize, ZSTD_getParams(cLevel, dictBufferSize+largestBlockSize)); + ZSTD_compressBegin_advanced(refCtx, dictBuffer, dictBufferSize, ZSTD_getParams(cLevel, MAX(dictBufferSize, largestBlockSize))); for (blockNb=0; blockNb /* malloc */ @@ -31,7 +31,7 @@ #include /* memcpy */ -/************************************** +/*-************************************ * Basic Types **************************************/ #if defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */ @@ -50,7 +50,7 @@ #endif -/************************************** +/*-************************************ * OS-specific Includes **************************************/ #if defined(MSDOS) || defined(OS2) || defined(WIN32) || defined(_WIN32) || defined(__CYGWIN__) @@ -62,7 +62,7 @@ #endif -/************************************** +/*-************************************ * Constants **************************************/ #define KB *(1 <<10) @@ -71,7 +71,7 @@ #define PRIME2 2246822519U -/************************************** +/*-************************************ * Local types **************************************/ #define LTLOG 13 @@ -81,7 +81,7 @@ typedef BYTE litDistribTable[LTSIZE]; -/********************************************************* +/*-******************************************************* * Local Functions *********************************************************/ #define RDG_rotl32(x,r) ((x << r) | (x >> (32 - r))) @@ -103,14 +103,12 @@ static void RDG_fillLiteralDistrib(litDistribTable lt, double ld) BYTE firstChar = '('; BYTE lastChar = '}'; - if (ld==0.0) - { + if (ld==0.0) { character = 0; firstChar = 0; lastChar =255; } - while (i LTSIZE) weight = LTSIZE-i; @@ -140,13 +138,11 @@ void RDG_genBlock(void* buffer, size_t buffSize, size_t prefixSize, double match U32 prevOffset = 1; /* special case : sparse content */ - while (matchProba >= 1.0) - { + while (matchProba >= 1.0) { size_t size0 = RDG_rand(seed) & 3; size0 = (size_t)1 << (16 + size0 * 2); size0 += RDG_rand(seed) & (size0-1); /* because size0 is power of 2*/ - if (buffSize < pos + size0) - { + if (buffSize < pos + size0) { memset(buffPtr+pos, 0, buffSize-pos); return; } @@ -160,11 +156,9 @@ void RDG_genBlock(void* buffer, size_t buffSize, size_t prefixSize, double match if (pos==0) buffPtr[0] = RDG_genChar(seed, lt), pos=1; /* Generate compressible data */ - while (pos < buffSize) - { + while (pos < buffSize) { /* Select : Literal (char) or Match (within 32K) */ - if (RDG_RAND15BITS < matchProba32) - { + if (RDG_RAND15BITS < matchProba32) { /* Copy (within 32K) */ size_t match; size_t d; @@ -178,17 +172,14 @@ void RDG_genBlock(void* buffer, size_t buffSize, size_t prefixSize, double match d = pos + length; if (d > buffSize) d = buffSize; while (pos < d) buffPtr[pos++] = buffPtr[match++]; /* correctly manages overlaps */ - } - else - { + } else { /* Literal (noise) */ size_t d; size_t length = RDG_RANDLENGTH; d = pos + length; if (d > buffSize) d = buffSize; while (pos < d) buffPtr[pos++] = RDG_genChar(seed, lt); - } - } + } } } @@ -220,8 +211,7 @@ void RDG_genStdout(unsigned long long size, double matchProba, double litProba, RDG_genBlock(buff, RDG_DICTSIZE, 0, matchProba, lt, &seed); /* Generate compressible data */ - while (total < size) - { + while (total < size) { RDG_genBlock(buff, RDG_DICTSIZE+RDG_BLOCKSIZE, RDG_DICTSIZE, matchProba, lt, &seed); if (size-total < RDG_BLOCKSIZE) genBlockSize = (size_t)(size-total); total += genBlockSize; @@ -230,6 +220,6 @@ void RDG_genStdout(unsigned long long size, double matchProba, double litProba, memcpy(buff, buff + RDG_BLOCKSIZE, RDG_DICTSIZE); } - // cleanup + /* cleanup */ free(buff); } diff --git a/programs/fileio.c b/programs/fileio.c index 9ce8d8cfe..ed2a06181 100644 --- a/programs/fileio.c +++ b/programs/fileio.c @@ -193,42 +193,30 @@ static U64 FIO_getFileSize(const char* infilename) static int FIO_getFiles(FILE** fileOutPtr, FILE** fileInPtr, const char* dstFileName, const char* srcFileName) { - if (!strcmp (srcFileName, stdinmark)) - { + if (!strcmp (srcFileName, stdinmark)) { DISPLAYLEVEL(4,"Using stdin for input\n"); *fileInPtr = stdin; SET_BINARY_MODE(stdin); - } - else - { + } else { *fileInPtr = fopen(srcFileName, "rb"); } - if ( *fileInPtr==0 ) - { + if ( *fileInPtr==0 ) { DISPLAYLEVEL(1, "Unable to access file for processing: %s\n", srcFileName); return 1; } - if (!strcmp (dstFileName, stdoutmark)) - { + if (!strcmp (dstFileName, stdoutmark)) { DISPLAYLEVEL(4,"Using stdout for output\n"); *fileOutPtr = stdout; SET_BINARY_MODE(stdout); - } - else - { - /* Check if destination file already exists */ - if (!g_overwrite) - { + } else { + if (!g_overwrite) { /* Check if destination file already exists */ *fileOutPtr = fopen( dstFileName, "rb" ); - if (*fileOutPtr != 0) - { - /* prompt for overwrite authorization */ + if (*fileOutPtr != 0) { /* dest file exists, prompt for overwrite authorization */ fclose(*fileOutPtr); DISPLAY("Warning : %s already exists \n", dstFileName); - if ((g_displayLevel <= 1) || (*fileInPtr == stdin)) - { + if ((g_displayLevel <= 1) || (*fileInPtr == stdin)) { /* No interaction possible */ DISPLAY("Operation aborted : %s already exists \n", dstFileName); return 1; @@ -236,15 +224,12 @@ static int FIO_getFiles(FILE** fileOutPtr, FILE** fileInPtr, DISPLAY("Overwrite ? (y/N) : "); { int ch = getchar(); - if ((ch!='Y') && (ch!='y')) - { + if ((ch!='Y') && (ch!='y')) { DISPLAY("No. Operation aborted : %s already exists \n", dstFileName); return 1; } while ((ch!=EOF) && (ch!='\n')) ch = getchar(); /* flush rest of input line */ - } - } - } + } } } *fileOutPtr = fopen( dstFileName, "wb" ); } @@ -265,15 +250,13 @@ static size_t FIO_loadFile(void** bufferPtr, const char* fileName) U64 fileSize; *bufferPtr = NULL; - if (fileName == NULL) - return 0; + if (fileName == NULL) return 0; DISPLAYLEVEL(4,"Loading %s as dictionary \n", fileName); fileHandle = fopen(fileName, "rb"); if (fileHandle==0) EXM_THROW(31, "Error opening file %s", fileName); fileSize = FIO_getFileSize(fileName); - if (fileSize > MAX_DICT_SIZE) - { + if (fileSize > MAX_DICT_SIZE) { int seekResult; if (fileSize > 1 GB) EXM_THROW(32, "Dictionary file %s is too large", fileName); /* avoid extreme cases */ DISPLAYLEVEL(2,"Dictionary %s is too large : using last %u bytes only \n", fileName, MAX_DICT_SIZE); @@ -455,8 +438,7 @@ int FIO_compressMultipleFilenames(const char** inFileNamesTable, unsigned nbFile ress = FIO_createCResources(dictFileName); /* loop on each file */ - for (u=0; u version */ size_t toRead = 4; @@ -573,8 +553,7 @@ static int FIO_decompressFile_extRess(dRess_t ress, if (sizeCheck==0) break; /* no more input */ if (sizeCheck != toRead) EXM_THROW(31, "Read error : cannot read header"); #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT==1) - if (ZSTD_isLegacy(MEM_readLE32(ress.srcBuffer))) - { + if (ZSTD_isLegacy(MEM_readLE32(ress.srcBuffer))) { filesize += FIO_decompressLegacyFrame(dstFile, srcFile, MEM_readLE32(ress.srcBuffer)); continue; } @@ -624,14 +603,12 @@ int FIO_decompressMultipleFilenames(const char** srcNamesTable, unsigned nbFiles if (dstFileName==NULL) EXM_THROW(70, "not enough memory for dstFileName"); ress = FIO_createDResources(dictFileName); - for (u=0; u Date: Tue, 2 Feb 2016 16:00:50 +0100 Subject: [PATCH 05/12] fixed encoding bugs --- lib/zstd_compress.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/lib/zstd_compress.c b/lib/zstd_compress.c index 7252d750b..3dd865151 100644 --- a/lib/zstd_compress.c +++ b/lib/zstd_compress.c @@ -251,6 +251,7 @@ size_t ZSTD_copyCCtx(ZSTD_CCtx* dstCCtx, const ZSTD_CCtx* srcCCtx) dstCCtx->dictBase = srcCCtx->dictBase; dstCCtx->dictLimit = srcCCtx->dictLimit; dstCCtx->lowLimit = srcCCtx->lowLimit; + dstCCtx->loadedDictEnd = srcCCtx->loadedDictEnd; /* copy entropy tables */ dstCCtx->flagStaticTables = srcCCtx->flagStaticTables; @@ -1911,7 +1912,8 @@ static size_t ZSTD_compress_generic (ZSTD_CCtx* zc, if (remaining < blockSize) blockSize = remaining; if ((U32)(ip+blockSize - zc->base) > zc->loadedDictEnd + maxDist) { /* enforce maxDist */ - zc->lowLimit = (U32)(ip+blockSize - zc->base) - maxDist; + U32 newLowLimit = (U32)(ip+blockSize - zc->base) - maxDist; + if (zc->lowLimit < newLowLimit) zc->lowLimit = newLowLimit; if (zc->dictLimit < zc->lowLimit) zc->dictLimit = zc->lowLimit; } From 7d8e6bddb98a085514982b802fb1c8ce980320fb Mon Sep 17 00:00:00 2001 From: Yann Collet Date: Tue, 2 Feb 2016 17:30:37 +0100 Subject: [PATCH 06/12] changed long length format --- lib/zstd_compress.c | 18 +++++++++++++++--- lib/zstd_decompress.c | 10 ++++++---- 2 files changed, 21 insertions(+), 7 deletions(-) diff --git a/lib/zstd_compress.c b/lib/zstd_compress.c index 3dd865151..c89d5ca1b 100644 --- a/lib/zstd_compress.c +++ b/lib/zstd_compress.c @@ -718,7 +718,7 @@ _check_compressibility: */ MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* literals, size_t offsetCode, size_t matchCode) { -#if 0 +#if 0 /* for debug */ static const BYTE* g_start = NULL; if (g_start==NULL) g_start = literals; //if (literals - g_start == 8695) @@ -737,7 +737,13 @@ MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const B *(seqStorePtr->dumps++) = (BYTE)(litLength - MaxLL); } else { *(seqStorePtr->dumps++) = 255; - MEM_writeLE32(seqStorePtr->dumps, (U32)litLength); seqStorePtr->dumps += 3; + if (litLength < (1<<15)) { + MEM_writeLE16(seqStorePtr->dumps, (U16)(litLength<<1)); + seqStorePtr->dumps += 2; + } else { + MEM_writeLE32(seqStorePtr->dumps, (U32)((litLength<<1)+1)); + seqStorePtr->dumps += 3; + } } } else *(seqStorePtr->litLength++) = (BYTE)litLength; @@ -751,7 +757,13 @@ MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const B *(seqStorePtr->dumps++) = (BYTE)(matchCode - MaxML); } else { *(seqStorePtr->dumps++) = 255; - MEM_writeLE32(seqStorePtr->dumps, (U32)matchCode); seqStorePtr->dumps += 3; + if (matchCode < (1<<15)) { + MEM_writeLE16(seqStorePtr->dumps, (U16)(matchCode<<1)); + seqStorePtr->dumps += 2; + } else { + MEM_writeLE32(seqStorePtr->dumps, (U32)((matchCode<<1)+1)); + seqStorePtr->dumps += 3; + } } } else *(seqStorePtr->matchLength++) = (BYTE)matchCode; } diff --git a/lib/zstd_decompress.c b/lib/zstd_decompress.c index 6061c8741..5d637d2bb 100644 --- a/lib/zstd_decompress.c +++ b/lib/zstd_decompress.c @@ -626,8 +626,9 @@ static void ZSTD_decodeSequence(seq_t* seq, seqState_t* seqState) U32 add = *dumps++; if (add < 255) litLength += add; else { - litLength = MEM_readLE32(dumps) & 0xFFFFFF; /* no pb : dumps is always followed by seq tables > 1 byte */ - dumps += 3; + litLength = MEM_readLE32(dumps) & 0xFFFFFF; /* no risk : dumps is always followed by seq tables > 1 byte */ + if (litLength&1) litLength>>=1, dumps += 3; + else litLength = (U16)(litLength)>>1, dumps += 2; } if (dumps >= de) dumps = de-1; /* late correction, to avoid read overflow (data is now corrupted anyway) */ } @@ -659,7 +660,8 @@ static void ZSTD_decodeSequence(seq_t* seq, seqState_t* seqState) if (add < 255) matchLength += add; else { matchLength = MEM_readLE32(dumps) & 0xFFFFFF; /* no pb : dumps is always followed by seq tables > 1 byte */ - dumps += 3; + if (matchLength&1) matchLength>>=1, dumps += 3; + else matchLength = (U16)(matchLength)>>1, dumps += 2; } if (dumps >= de) dumps = de-1; /* late correction, to avoid read overflow (data is now corrupted anyway) */ } @@ -671,7 +673,7 @@ static void ZSTD_decodeSequence(seq_t* seq, seqState_t* seqState) seq->matchLength = matchLength; seqState->dumps = dumps; -#if 0 +#if 0 /* debug */ { static U64 totalDecoded = 0; printf("pos %6u : %3u literals & match %3u bytes at distance %6u \n", From 464fa996e9589cc20d7bb83d0daa4520f8e37991 Mon Sep 17 00:00:00 2001 From: Yann Collet Date: Wed, 3 Feb 2016 01:09:46 +0100 Subject: [PATCH 07/12] added v0.4 legacy support --- lib/Makefile | 2 +- lib/legacy/zstd_legacy.h | 6 +- lib/legacy/zstd_v04.c | 4431 +++++++++++++++++++++++++++++++ lib/legacy/zstd_v04.h | 148 ++ programs/Makefile | 2 +- programs/fuzzer.c | 2 +- programs/legacy/fileio_legacy.c | 82 + 7 files changed, 4669 insertions(+), 4 deletions(-) create mode 100644 lib/legacy/zstd_v04.c create mode 100644 lib/legacy/zstd_v04.h diff --git a/lib/Makefile b/lib/Makefile index a7172b6f8..e0ca0c626 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -53,7 +53,7 @@ LIBDIR ?= $(PREFIX)/lib INCLUDEDIR=$(PREFIX)/include ZSTD_FILES := zstd_compress.c zstd_decompress.c fse.c huff0.c -ZSTD_LEGACY:= legacy/zstd_v01.c legacy/zstd_v02.c legacy/zstd_v03.c +ZSTD_LEGACY:= legacy/zstd_v01.c legacy/zstd_v02.c legacy/zstd_v03.c legacy/zstd_v04.c ifeq ($(ZSTD_LEGACY_SUPPORT), 0) CPPFLAGS += -DZSTD_LEGACY_SUPPORT=0 diff --git a/lib/legacy/zstd_legacy.h b/lib/legacy/zstd_legacy.h index b3e5eb204..4ae1deb68 100644 --- a/lib/legacy/zstd_legacy.h +++ b/lib/legacy/zstd_legacy.h @@ -45,6 +45,7 @@ extern "C" { #include "zstd_v01.h" #include "zstd_v02.h" #include "zstd_v03.h" +#include "zstd_v04.h" MEM_STATIC unsigned ZSTD_isLegacy (U32 magicNumberLE) { @@ -52,7 +53,8 @@ MEM_STATIC unsigned ZSTD_isLegacy (U32 magicNumberLE) { case ZSTDv01_magicNumberLE : case ZSTDv02_magicNumber : - case ZSTDv03_magicNumber : return 1; + case ZSTDv03_magicNumber : + case ZSTDv04_magicNumber : return 1; default : return 0; } } @@ -71,6 +73,8 @@ MEM_STATIC size_t ZSTD_decompressLegacy( return ZSTDv02_decompress(dst, maxOriginalSize, src, compressedSize); case ZSTDv03_magicNumber : return ZSTDv03_decompress(dst, maxOriginalSize, src, compressedSize); + case ZSTDv04_magicNumber : + return ZSTDv04_decompress(dst, maxOriginalSize, src, compressedSize); default : return ERROR(prefix_unknown); } diff --git a/lib/legacy/zstd_v04.c b/lib/legacy/zstd_v04.c new file mode 100644 index 000000000..45a3f61cc --- /dev/null +++ b/lib/legacy/zstd_v04.c @@ -0,0 +1,4431 @@ +/* ****************************************************************** + zstd_v04.c + Decompression module for ZSTD v0.4 legacy format + Copyright (C) 2016, Yann Collet. + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - Homepage : http://www.zstd.net/ +****************************************************************** */ + +/*- Dependencies -*/ +#include "zstd_v04.h" + + +/* ****************************************************************** + mem.h + low-level memory access routines + Copyright (C) 2013-2015, Yann Collet. + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy + - Public forum : https://groups.google.com/forum/#!forum/lz4c +****************************************************************** */ +#ifndef MEM_H_MODULE +#define MEM_H_MODULE + +#if defined (__cplusplus) +extern "C" { +#endif + +/****************************************** +* Includes +******************************************/ +#include /* size_t, ptrdiff_t */ +#include /* memcpy */ + + +/****************************************** +* Compiler-specific +******************************************/ +#if defined(__GNUC__) +# define MEM_STATIC static __attribute__((unused)) +#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) +# define MEM_STATIC static inline +#elif defined(_MSC_VER) +# define MEM_STATIC static __inline +#else +# define MEM_STATIC static /* this version may generate warnings for unused static functions; disable the relevant warning */ +#endif + + +/**************************************************************** +* Basic Types +*****************************************************************/ +#if defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) +# include + typedef uint8_t BYTE; + typedef uint16_t U16; + typedef int16_t S16; + typedef uint32_t U32; + typedef int32_t S32; + typedef uint64_t U64; + typedef int64_t S64; +#else + typedef unsigned char BYTE; + typedef unsigned short U16; + typedef signed short S16; + typedef unsigned int U32; + typedef signed int S32; + typedef unsigned long long U64; + typedef signed long long S64; +#endif + + +/**************************************************************** +* Memory I/O +*****************************************************************/ +/* MEM_FORCE_MEMORY_ACCESS + * By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable. + * Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal. + * The below switch allow to select different access method for improved performance. + * Method 0 (default) : use `memcpy()`. Safe and portable. + * Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable). + * This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`. + * Method 2 : direct access. This method is portable but violate C standard. + * It can generate buggy code on targets generating assembly depending on alignment. + * But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6) + * See http://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details. + * Prefer these methods in priority order (0 > 1 > 2) + */ +#ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ +# if defined(__GNUC__) && ( defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) ) +# define MEM_FORCE_MEMORY_ACCESS 2 +# elif defined(__INTEL_COMPILER) || \ + (defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) )) +# define MEM_FORCE_MEMORY_ACCESS 1 +# endif +#endif + +MEM_STATIC unsigned MEM_32bits(void) { return sizeof(void*)==4; } +MEM_STATIC unsigned MEM_64bits(void) { return sizeof(void*)==8; } + +MEM_STATIC unsigned MEM_isLittleEndian(void) +{ + const union { U32 u; BYTE c[4]; } one = { 1 }; /* don't use static : performance detrimental */ + return one.c[0]; +} + +#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2) + +/* violates C standard on structure alignment. +Only use if no other choice to achieve best performance on target platform */ +MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; } +MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; } +MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; } + +MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; } +MEM_STATIC void MEM_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; } +MEM_STATIC void MEM_write64(void* memPtr, U64 value) { *(U64*)memPtr = value; } + +#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1) + +/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */ +/* currently only defined for gcc and icc */ +typedef union { U16 u16; U32 u32; U64 u64; } __attribute__((packed)) unalign; + +MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign*)ptr)->u16; } +MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign*)ptr)->u32; } +MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign*)ptr)->u64; } + +MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; } +MEM_STATIC void MEM_write32(void* memPtr, U32 value) { ((unalign*)memPtr)->u32 = value; } +MEM_STATIC void MEM_write64(void* memPtr, U64 value) { ((unalign*)memPtr)->u64 = value; } + +#else + +/* default method, safe and standard. + can sometimes prove slower */ + +MEM_STATIC U16 MEM_read16(const void* memPtr) +{ + U16 val; memcpy(&val, memPtr, sizeof(val)); return val; +} + +MEM_STATIC U32 MEM_read32(const void* memPtr) +{ + U32 val; memcpy(&val, memPtr, sizeof(val)); return val; +} + +MEM_STATIC U64 MEM_read64(const void* memPtr) +{ + U64 val; memcpy(&val, memPtr, sizeof(val)); return val; +} + +MEM_STATIC void MEM_write16(void* memPtr, U16 value) +{ + memcpy(memPtr, &value, sizeof(value)); +} + +MEM_STATIC void MEM_write32(void* memPtr, U32 value) +{ + memcpy(memPtr, &value, sizeof(value)); +} + +MEM_STATIC void MEM_write64(void* memPtr, U64 value) +{ + memcpy(memPtr, &value, sizeof(value)); +} + +#endif // MEM_FORCE_MEMORY_ACCESS + + +MEM_STATIC U16 MEM_readLE16(const void* memPtr) +{ + if (MEM_isLittleEndian()) + return MEM_read16(memPtr); + else + { + const BYTE* p = (const BYTE*)memPtr; + return (U16)(p[0] + (p[1]<<8)); + } +} + +MEM_STATIC void MEM_writeLE16(void* memPtr, U16 val) +{ + if (MEM_isLittleEndian()) + { + MEM_write16(memPtr, val); + } + else + { + BYTE* p = (BYTE*)memPtr; + p[0] = (BYTE)val; + p[1] = (BYTE)(val>>8); + } +} + +MEM_STATIC U32 MEM_readLE32(const void* memPtr) +{ + if (MEM_isLittleEndian()) + return MEM_read32(memPtr); + else + { + const BYTE* p = (const BYTE*)memPtr; + return (U32)((U32)p[0] + ((U32)p[1]<<8) + ((U32)p[2]<<16) + ((U32)p[3]<<24)); + } +} + +MEM_STATIC void MEM_writeLE32(void* memPtr, U32 val32) +{ + if (MEM_isLittleEndian()) + { + MEM_write32(memPtr, val32); + } + else + { + BYTE* p = (BYTE*)memPtr; + p[0] = (BYTE)val32; + p[1] = (BYTE)(val32>>8); + p[2] = (BYTE)(val32>>16); + p[3] = (BYTE)(val32>>24); + } +} + +MEM_STATIC U64 MEM_readLE64(const void* memPtr) +{ + if (MEM_isLittleEndian()) + return MEM_read64(memPtr); + else + { + const BYTE* p = (const BYTE*)memPtr; + return (U64)((U64)p[0] + ((U64)p[1]<<8) + ((U64)p[2]<<16) + ((U64)p[3]<<24) + + ((U64)p[4]<<32) + ((U64)p[5]<<40) + ((U64)p[6]<<48) + ((U64)p[7]<<56)); + } +} + +MEM_STATIC void MEM_writeLE64(void* memPtr, U64 val64) +{ + if (MEM_isLittleEndian()) + { + MEM_write64(memPtr, val64); + } + else + { + BYTE* p = (BYTE*)memPtr; + p[0] = (BYTE)val64; + p[1] = (BYTE)(val64>>8); + p[2] = (BYTE)(val64>>16); + p[3] = (BYTE)(val64>>24); + p[4] = (BYTE)(val64>>32); + p[5] = (BYTE)(val64>>40); + p[6] = (BYTE)(val64>>48); + p[7] = (BYTE)(val64>>56); + } +} + +MEM_STATIC size_t MEM_readLEST(const void* memPtr) +{ + if (MEM_32bits()) + return (size_t)MEM_readLE32(memPtr); + else + return (size_t)MEM_readLE64(memPtr); +} + +MEM_STATIC void MEM_writeLEST(void* memPtr, size_t val) +{ + if (MEM_32bits()) + MEM_writeLE32(memPtr, (U32)val); + else + MEM_writeLE64(memPtr, (U64)val); +} + +#if defined (__cplusplus) +} +#endif + +#endif /* MEM_H_MODULE */ + +/* ****************************************************************** + Error codes list + Copyright (C) 2016, Yann Collet + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - Source repository : https://github.com/Cyan4973/zstd +****************************************************************** */ +#ifndef ERROR_PUBLIC_H_MODULE +#define ERROR_PUBLIC_H_MODULE + +#if defined (__cplusplus) +extern "C" { +#endif + + +/* **************************************** +* error list +******************************************/ +enum { + ZSTD_error_No_Error, + ZSTD_error_GENERIC, + ZSTD_error_prefix_unknown, + ZSTD_error_frameParameter_unsupported, + ZSTD_error_frameParameter_unsupportedBy32bitsImplementation, + ZSTD_error_init_missing, + ZSTD_error_memory_allocation, + ZSTD_error_stage_wrong, + ZSTD_error_dstSize_tooSmall, + ZSTD_error_srcSize_wrong, + ZSTD_error_corruption_detected, + ZSTD_error_tableLog_tooLarge, + ZSTD_error_maxSymbolValue_tooLarge, + ZSTD_error_maxSymbolValue_tooSmall, + ZSTD_error_maxCode +}; + +/* note : functions provide error codes in reverse negative order, + so compare with (size_t)(0-enum) */ + + +#if defined (__cplusplus) +} +#endif + +#endif /* ERROR_PUBLIC_H_MODULE */ + + + +/* + zstd - standard compression library + Header File for static linking only + Copyright (C) 2014-2015, Yann Collet. + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - zstd source repository : https://github.com/Cyan4973/zstd + - ztsd public forum : https://groups.google.com/forum/#!forum/lz4c +*/ +#ifndef ZSTD_STATIC_H +#define ZSTD_STATIC_H + +/* The objects defined into this file shall be considered experimental. + * They are not considered stable, as their prototype may change in the future. + * You can use them for tests, provide feedback, or if you can endure risks of future changes. + */ + +#if defined (__cplusplus) +extern "C" { +#endif + +/* ************************************* +* Types +***************************************/ +#define ZSTD_WINDOWLOG_MAX 26 +#define ZSTD_WINDOWLOG_MIN 18 +#define ZSTD_WINDOWLOG_ABSOLUTEMIN 11 +#define ZSTD_CONTENTLOG_MAX (ZSTD_WINDOWLOG_MAX+1) +#define ZSTD_CONTENTLOG_MIN 4 +#define ZSTD_HASHLOG_MAX 28 +#define ZSTD_HASHLOG_MIN 4 +#define ZSTD_SEARCHLOG_MAX (ZSTD_CONTENTLOG_MAX-1) +#define ZSTD_SEARCHLOG_MIN 1 +#define ZSTD_SEARCHLENGTH_MAX 7 +#define ZSTD_SEARCHLENGTH_MIN 4 + +/** from faster to stronger */ +typedef enum { ZSTD_fast, ZSTD_greedy, ZSTD_lazy, ZSTD_lazy2, ZSTD_btlazy2 } ZSTD_strategy; + +typedef struct +{ + U64 srcSize; /* optional : tells how much bytes are present in the frame. Use 0 if not known. */ + U32 windowLog; /* largest match distance : larger == more compression, more memory needed during decompression */ + U32 contentLog; /* full search segment : larger == more compression, slower, more memory (useless for fast) */ + U32 hashLog; /* dispatch table : larger == more memory, faster */ + U32 searchLog; /* nb of searches : larger == more compression, slower */ + U32 searchLength; /* size of matches : larger == faster decompression, sometimes less compression */ + ZSTD_strategy strategy; +} ZSTD_parameters; + +typedef ZSTDv04_Dctx ZSTD_DCtx; + +/* ************************************* +* Advanced functions +***************************************/ +/** ZSTD_decompress_usingDict +* Same as ZSTD_decompressDCtx, using a Dictionary content as prefix +* Note : dict can be NULL, in which case, it's equivalent to ZSTD_decompressDCtx() */ +static size_t ZSTD_decompress_usingDict(ZSTD_DCtx* ctx, + void* dst, size_t maxDstSize, + const void* src, size_t srcSize, + const void* dict,size_t dictSize); + + +/* ************************************** +* Streaming functions (direct mode) +****************************************/ +static size_t ZSTD_resetDCtx(ZSTD_DCtx* dctx); +static size_t ZSTD_getFrameParams(ZSTD_parameters* params, const void* src, size_t srcSize); +static void ZSTD_decompress_insertDictionary(ZSTD_DCtx* ctx, const void* src, size_t srcSize); + +static size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx); +static size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize, const void* src, size_t srcSize); + +/** + Streaming decompression, bufferless mode + + A ZSTD_DCtx object is required to track streaming operations. + Use ZSTD_createDCtx() / ZSTD_freeDCtx() to manage it. + A ZSTD_DCtx object can be re-used multiple times. Use ZSTD_resetDCtx() to return to fresh status. + + First operation is to retrieve frame parameters, using ZSTD_getFrameParams(). + This function doesn't consume its input. It needs enough input data to properly decode the frame header. + Objective is to retrieve *params.windowlog, to know minimum amount of memory required during decoding. + Result : 0 when successful, it means the ZSTD_parameters structure has been filled. + >0 : means there is not enough data into src. Provides the expected size to successfully decode header. + errorCode, which can be tested using ZSTD_isError() (For example, if it's not a ZSTD header) + + Then, you can optionally insert a dictionary. + This operation must mimic the compressor behavior, otherwise decompression will fail or be corrupted. + + Then it's possible to start decompression. + Use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() alternatively. + ZSTD_nextSrcSizeToDecompress() tells how much bytes to provide as 'srcSize' to ZSTD_decompressContinue(). + ZSTD_decompressContinue() requires this exact amount of bytes, or it will fail. + ZSTD_decompressContinue() needs previous data blocks during decompression, up to (1 << windowlog). + They should preferably be located contiguously, prior to current block. Alternatively, a round buffer is also possible. + + @result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst'. + It can be zero, which is not an error; it just means ZSTD_decompressContinue() has decoded some header. + + A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero. + Context can then be reset to start a new decompression. +*/ + + +#if defined (__cplusplus) +} +#endif + +/* ****************************************************************** + Error codes and messages + Copyright (C) 2013-2016, Yann Collet + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - Source repository : https://github.com/Cyan4973/zstd +****************************************************************** */ +/* Note : this module is expected to remain private, do not expose it */ + +#ifndef ERROR_H_MODULE +#define ERROR_H_MODULE + +#if defined (__cplusplus) +extern "C" { +#endif + + +/* ***************************************** +* Includes +******************************************/ +#include /* size_t, ptrdiff_t */ + + +/* ***************************************** +* Compiler-specific +******************************************/ +#if defined(__GNUC__) +# define ERR_STATIC static __attribute__((unused)) +#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) +# define ERR_STATIC static inline +#elif defined(_MSC_VER) +# define ERR_STATIC static __inline +#else +# define ERR_STATIC static /* this version may generate warnings for unused static functions; disable the relevant warning */ +#endif + + +/* ***************************************** +* Error Codes +******************************************/ +#define PREFIX(name) ZSTD_error_##name + +#ifdef ERROR +# undef ERROR /* reported already defined on VS 2015 by Rich Geldreich */ +#endif +#define ERROR(name) (size_t)-PREFIX(name) + +ERR_STATIC unsigned ERR_isError(size_t code) { return (code > ERROR(maxCode)); } + + +/* ***************************************** +* Error Strings +******************************************/ + +ERR_STATIC const char* ERR_getErrorName(size_t code) +{ + static const char* codeError = "Unspecified error code"; + switch( (size_t)(0-code) ) + { + case ZSTD_error_No_Error: return "No error detected"; + case ZSTD_error_GENERIC: return "Error (generic)"; + case ZSTD_error_prefix_unknown: return "Unknown frame descriptor"; + case ZSTD_error_frameParameter_unsupported: return "Unsupported frame parameter"; + case ZSTD_error_frameParameter_unsupportedBy32bitsImplementation: return "Frame parameter unsupported in 32-bits mode"; + case ZSTD_error_init_missing: return "Context should be init first"; + case ZSTD_error_memory_allocation: return "Allocation error : not enough memory"; + case ZSTD_error_dstSize_tooSmall: return "Destination buffer is too small"; + case ZSTD_error_srcSize_wrong: return "Src size incorrect"; + case ZSTD_error_corruption_detected: return "Corrupted block detected"; + case ZSTD_error_tableLog_tooLarge: return "tableLog requires too much memory"; + case ZSTD_error_maxSymbolValue_tooLarge: return "Unsupported max possible Symbol Value : too large"; + case ZSTD_error_maxSymbolValue_tooSmall: return "Specified maxSymbolValue is too small"; + case ZSTD_error_maxCode: + default: return codeError; + } +} + + +#if defined (__cplusplus) +} +#endif + +#endif /* ERROR_H_MODULE */ + + +#endif /* ZSTD_STATIC_H */ + + +/* + zstd_internal - common functions to include + Header File for include + Copyright (C) 2014-2015, Yann Collet. + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - zstd source repository : https://github.com/Cyan4973/zstd + - ztsd public forum : https://groups.google.com/forum/#!forum/lz4c +*/ +#ifndef ZSTD_CCOMMON_H_MODULE +#define ZSTD_CCOMMON_H_MODULE + +#if defined (__cplusplus) +extern "C" { +#endif + +/* ************************************* +* Common macros +***************************************/ +#define MIN(a,b) ((a)<(b) ? (a) : (b)) +#define MAX(a,b) ((a)>(b) ? (a) : (b)) + + +/* ************************************* +* Common constants +***************************************/ +#define ZSTD_MAGICNUMBER 0xFD2FB524 /* v0.4 */ + +#define KB *(1 <<10) +#define MB *(1 <<20) +#define GB *(1U<<30) + +#define BLOCKSIZE (128 KB) /* define, for static allocation */ + +static const size_t ZSTD_blockHeaderSize = 3; +static const size_t ZSTD_frameHeaderSize_min = 5; +#define ZSTD_frameHeaderSize_max 5 /* define, for static allocation */ + +#define BIT7 128 +#define BIT6 64 +#define BIT5 32 +#define BIT4 16 +#define BIT1 2 +#define BIT0 1 + +#define IS_RAW BIT0 +#define IS_RLE BIT1 + +#define MINMATCH 4 +#define REPCODE_STARTVALUE 4 + +#define MLbits 7 +#define LLbits 6 +#define Offbits 5 +#define MaxML ((1< /* size_t, ptrdiff_t */ + + +/* ***************************************** +* FSE simple functions +******************************************/ +static size_t FSE_decompress(void* dst, size_t maxDstSize, + const void* cSrc, size_t cSrcSize); +/*! +FSE_decompress(): + Decompress FSE data from buffer 'cSrc', of size 'cSrcSize', + into already allocated destination buffer 'dst', of size 'maxDstSize'. + return : size of regenerated data (<= maxDstSize) + or an error code, which can be tested using FSE_isError() + + ** Important ** : FSE_decompress() doesn't decompress non-compressible nor RLE data !!! + Why ? : making this distinction requires a header. + Header management is intentionally delegated to the user layer, which can better manage special cases. +*/ + + +/* ***************************************** +* Tool functions +******************************************/ +/* Error Management */ +static unsigned FSE_isError(size_t code); /* tells if a return value is an error code */ + + + +/* ***************************************** +* FSE detailed API +******************************************/ +/*! +FSE_compress() does the following: +1. count symbol occurrence from source[] into table count[] +2. normalize counters so that sum(count[]) == Power_of_2 (2^tableLog) +3. save normalized counters to memory buffer using writeNCount() +4. build encoding table 'CTable' from normalized counters +5. encode the data stream using encoding table 'CTable' + +FSE_decompress() does the following: +1. read normalized counters with readNCount() +2. build decoding table 'DTable' from normalized counters +3. decode the data stream using decoding table 'DTable' + +The following API allows targeting specific sub-functions for advanced tasks. +For example, it's possible to compress several blocks using the same 'CTable', +or to save and provide normalized distribution using external method. +*/ + + +/* *** DECOMPRESSION *** */ + +/*! +FSE_readNCount(): + Read compactly saved 'normalizedCounter' from 'rBuffer'. + return : size read from 'rBuffer' + or an errorCode, which can be tested using FSE_isError() + maxSymbolValuePtr[0] and tableLogPtr[0] will also be updated with their respective values */ +static size_t FSE_readNCount (short* normalizedCounter, unsigned* maxSymbolValuePtr, unsigned* tableLogPtr, const void* rBuffer, size_t rBuffSize); + +/*! +Constructor and Destructor of type FSE_DTable + Note that its size depends on 'tableLog' */ +typedef unsigned FSE_DTable; /* don't allocate that. It's just a way to be more restrictive than void* */ + +/*! +FSE_buildDTable(): + Builds 'dt', which must be already allocated, using FSE_createDTable() + return : 0, + or an errorCode, which can be tested using FSE_isError() */ +static size_t FSE_buildDTable ( FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog); + +/*! +FSE_decompress_usingDTable(): + Decompress compressed source 'cSrc' of size 'cSrcSize' using 'dt' + into 'dst' which must be already allocated. + return : size of regenerated data (necessarily <= maxDstSize) + or an errorCode, which can be tested using FSE_isError() */ +static size_t FSE_decompress_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const FSE_DTable* dt); + +/*! +Tutorial : +---------- +(Note : these functions only decompress FSE-compressed blocks. + If block is uncompressed, use memcpy() instead + If block is a single repeated byte, use memset() instead ) + +The first step is to obtain the normalized frequencies of symbols. +This can be performed by FSE_readNCount() if it was saved using FSE_writeNCount(). +'normalizedCounter' must be already allocated, and have at least 'maxSymbolValuePtr[0]+1' cells of signed short. +In practice, that means it's necessary to know 'maxSymbolValue' beforehand, +or size the table to handle worst case situations (typically 256). +FSE_readNCount() will provide 'tableLog' and 'maxSymbolValue'. +The result of FSE_readNCount() is the number of bytes read from 'rBuffer'. +Note that 'rBufferSize' must be at least 4 bytes, even if useful information is less than that. +If there is an error, the function will return an error code, which can be tested using FSE_isError(). + +The next step is to build the decompression tables 'FSE_DTable' from 'normalizedCounter'. +This is performed by the function FSE_buildDTable(). +The space required by 'FSE_DTable' must be already allocated using FSE_createDTable(). +If there is an error, the function will return an error code, which can be tested using FSE_isError(). + +'FSE_DTable' can then be used to decompress 'cSrc', with FSE_decompress_usingDTable(). +'cSrcSize' must be strictly correct, otherwise decompression will fail. +FSE_decompress_usingDTable() result will tell how many bytes were regenerated (<=maxDstSize). +If there is an error, the function will return an error code, which can be tested using FSE_isError(). (ex: dst buffer too small) +*/ + + +#if defined (__cplusplus) +} +#endif + +#endif /* FSE_H */ + + +/* ****************************************************************** + bitstream + Part of NewGen Entropy library + header file (to include) + Copyright (C) 2013-2015, Yann Collet. + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - Source repository : https://github.com/Cyan4973/FiniteStateEntropy + - Public forum : https://groups.google.com/forum/#!forum/lz4c +****************************************************************** */ +#ifndef BITSTREAM_H_MODULE +#define BITSTREAM_H_MODULE + +#if defined (__cplusplus) +extern "C" { +#endif + + +/* +* This API consists of small unitary functions, which highly benefit from being inlined. +* Since link-time-optimization is not available for all compilers, +* these functions are defined into a .h to be included. +*/ + +/********************************************** +* bitStream decompression API (read backward) +**********************************************/ +typedef struct +{ + size_t bitContainer; + unsigned bitsConsumed; + const char* ptr; + const char* start; +} BIT_DStream_t; + +typedef enum { BIT_DStream_unfinished = 0, + BIT_DStream_endOfBuffer = 1, + BIT_DStream_completed = 2, + BIT_DStream_overflow = 3 } BIT_DStream_status; /* result of BIT_reloadDStream() */ + /* 1,2,4,8 would be better for bitmap combinations, but slows down performance a bit ... :( */ + +MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize); +MEM_STATIC size_t BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits); +MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD); +MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* bitD); + + +/* +* Start by invoking BIT_initDStream(). +* A chunk of the bitStream is then stored into a local register. +* Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t). +* You can then retrieve bitFields stored into the local register, **in reverse order**. +* Local register is manually filled from memory by the BIT_reloadDStream() method. +* A reload guarantee a minimum of ((8*sizeof(size_t))-7) bits when its result is BIT_DStream_unfinished. +* Otherwise, it can be less than that, so proceed accordingly. +* Checking if DStream has reached its end can be performed with BIT_endOfDStream() +*/ + + +/****************************************** +* unsafe API +******************************************/ +MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits); +/* faster, but works only if nbBits >= 1 */ + + + +/**************************************************************** +* Helper functions +****************************************************************/ +MEM_STATIC unsigned BIT_highbit32 (register U32 val) +{ +# if defined(_MSC_VER) /* Visual */ + unsigned long r=0; + _BitScanReverse ( &r, val ); + return (unsigned) r; +# elif defined(__GNUC__) && (__GNUC__ >= 3) /* Use GCC Intrinsic */ + return 31 - __builtin_clz (val); +# else /* Software version */ + static const unsigned DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, 8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31 }; + U32 v = val; + unsigned r; + v |= v >> 1; + v |= v >> 2; + v |= v >> 4; + v |= v >> 8; + v |= v >> 16; + r = DeBruijnClz[ (U32) (v * 0x07C4ACDDU) >> 27]; + return r; +# endif +} + + +/********************************************************** +* bitStream decoding +**********************************************************/ + +/*!BIT_initDStream +* Initialize a BIT_DStream_t. +* @bitD : a pointer to an already allocated BIT_DStream_t structure +* @srcBuffer must point at the beginning of a bitStream +* @srcSize must be the exact size of the bitStream +* @result : size of stream (== srcSize) or an errorCode if a problem is detected +*/ +MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize) +{ + if (srcSize < 1) { memset(bitD, 0, sizeof(*bitD)); return ERROR(srcSize_wrong); } + + if (srcSize >= sizeof(size_t)) /* normal case */ + { + U32 contain32; + bitD->start = (const char*)srcBuffer; + bitD->ptr = (const char*)srcBuffer + srcSize - sizeof(size_t); + bitD->bitContainer = MEM_readLEST(bitD->ptr); + contain32 = ((const BYTE*)srcBuffer)[srcSize-1]; + if (contain32 == 0) return ERROR(GENERIC); /* endMark not present */ + bitD->bitsConsumed = 8 - BIT_highbit32(contain32); + } + else + { + U32 contain32; + bitD->start = (const char*)srcBuffer; + bitD->ptr = bitD->start; + bitD->bitContainer = *(const BYTE*)(bitD->start); + switch(srcSize) + { + case 7: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[6]) << (sizeof(size_t)*8 - 16); + case 6: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[5]) << (sizeof(size_t)*8 - 24); + case 5: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[4]) << (sizeof(size_t)*8 - 32); + case 4: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[3]) << 24; + case 3: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[2]) << 16; + case 2: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[1]) << 8; + default:; + } + contain32 = ((const BYTE*)srcBuffer)[srcSize-1]; + if (contain32 == 0) return ERROR(GENERIC); /* endMark not present */ + bitD->bitsConsumed = 8 - BIT_highbit32(contain32); + bitD->bitsConsumed += (U32)(sizeof(size_t) - srcSize)*8; + } + + return srcSize; +} + +/*!BIT_lookBits + * Provides next n bits from local register + * local register is not modified (bits are still present for next read/look) + * On 32-bits, maxNbBits==25 + * On 64-bits, maxNbBits==57 + * @return : value extracted + */ +MEM_STATIC size_t BIT_lookBits(BIT_DStream_t* bitD, U32 nbBits) +{ + const U32 bitMask = sizeof(bitD->bitContainer)*8 - 1; + return ((bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> 1) >> ((bitMask-nbBits) & bitMask); +} + +/*! BIT_lookBitsFast : +* unsafe version; only works only if nbBits >= 1 */ +MEM_STATIC size_t BIT_lookBitsFast(BIT_DStream_t* bitD, U32 nbBits) +{ + const U32 bitMask = sizeof(bitD->bitContainer)*8 - 1; + return (bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> (((bitMask+1)-nbBits) & bitMask); +} + +MEM_STATIC void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits) +{ + bitD->bitsConsumed += nbBits; +} + +/*!BIT_readBits + * Read next n bits from local register. + * pay attention to not read more than nbBits contained into local register. + * @return : extracted value. + */ +MEM_STATIC size_t BIT_readBits(BIT_DStream_t* bitD, U32 nbBits) +{ + size_t value = BIT_lookBits(bitD, nbBits); + BIT_skipBits(bitD, nbBits); + return value; +} + +/*!BIT_readBitsFast : +* unsafe version; only works only if nbBits >= 1 */ +MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, U32 nbBits) +{ + size_t value = BIT_lookBitsFast(bitD, nbBits); + BIT_skipBits(bitD, nbBits); + return value; +} + +MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD) +{ + if (bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8)) /* should never happen */ + return BIT_DStream_overflow; + + if (bitD->ptr >= bitD->start + sizeof(bitD->bitContainer)) + { + bitD->ptr -= bitD->bitsConsumed >> 3; + bitD->bitsConsumed &= 7; + bitD->bitContainer = MEM_readLEST(bitD->ptr); + return BIT_DStream_unfinished; + } + if (bitD->ptr == bitD->start) + { + if (bitD->bitsConsumed < sizeof(bitD->bitContainer)*8) return BIT_DStream_endOfBuffer; + return BIT_DStream_completed; + } + { + U32 nbBytes = bitD->bitsConsumed >> 3; + BIT_DStream_status result = BIT_DStream_unfinished; + if (bitD->ptr - nbBytes < bitD->start) + { + nbBytes = (U32)(bitD->ptr - bitD->start); /* ptr > start */ + result = BIT_DStream_endOfBuffer; + } + bitD->ptr -= nbBytes; + bitD->bitsConsumed -= nbBytes*8; + bitD->bitContainer = MEM_readLEST(bitD->ptr); /* reminder : srcSize > sizeof(bitD) */ + return result; + } +} + +/*! BIT_endOfDStream +* @return Tells if DStream has reached its exact end +*/ +MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* DStream) +{ + return ((DStream->ptr == DStream->start) && (DStream->bitsConsumed == sizeof(DStream->bitContainer)*8)); +} + +#if defined (__cplusplus) +} +#endif + +#endif /* BITSTREAM_H_MODULE */ + + + +/* ****************************************************************** + FSE : Finite State Entropy coder + header file for static linking (only) + Copyright (C) 2013-2015, Yann Collet + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - Source repository : https://github.com/Cyan4973/FiniteStateEntropy + - Public forum : https://groups.google.com/forum/#!forum/lz4c +****************************************************************** */ +#ifndef FSE_STATIC_H +#define FSE_STATIC_H + +#if defined (__cplusplus) +extern "C" { +#endif + + +/* ***************************************** +* Static allocation +*******************************************/ +/* FSE buffer bounds */ +#define FSE_NCOUNTBOUND 512 +#define FSE_BLOCKBOUND(size) (size + (size>>7)) +#define FSE_COMPRESSBOUND(size) (FSE_NCOUNTBOUND + FSE_BLOCKBOUND(size)) /* Macro version, useful for static allocation */ + +/* It is possible to statically allocate FSE CTable/DTable as a table of unsigned using below macros */ +#define FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) (1 + (1<<(maxTableLog-1)) + ((maxSymbolValue+1)*2)) +#define FSE_DTABLE_SIZE_U32(maxTableLog) (1 + (1<= BIT_DStream_completed + +When it's done, verify decompression is fully completed, by checking both DStream and the relevant states. +Checking if DStream has reached its end is performed by : + BIT_endOfDStream(&DStream); +Check also the states. There might be some symbols left there, if some high probability ones (>50%) are possible. + FSE_endOfDState(&DState); +*/ + + +/* ***************************************** +* FSE unsafe API +*******************************************/ +static unsigned char FSE_decodeSymbolFast(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD); +/* faster, but works only if nbBits is always >= 1 (otherwise, result will be corrupted) */ + + +/* ***************************************** +* Implementation of inlined functions +*******************************************/ +/* decompression */ + +typedef struct { + U16 tableLog; + U16 fastMode; +} FSE_DTableHeader; /* sizeof U32 */ + +typedef struct +{ + unsigned short newState; + unsigned char symbol; + unsigned char nbBits; +} FSE_decode_t; /* size == U32 */ + +MEM_STATIC void FSE_initDState(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD, const FSE_DTable* dt) +{ + const void* ptr = dt; + const FSE_DTableHeader* const DTableH = (const FSE_DTableHeader*)ptr; + DStatePtr->state = BIT_readBits(bitD, DTableH->tableLog); + BIT_reloadDStream(bitD); + DStatePtr->table = dt + 1; +} + +MEM_STATIC BYTE FSE_decodeSymbol(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD) +{ + const FSE_decode_t DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state]; + const U32 nbBits = DInfo.nbBits; + BYTE symbol = DInfo.symbol; + size_t lowBits = BIT_readBits(bitD, nbBits); + + DStatePtr->state = DInfo.newState + lowBits; + return symbol; +} + +MEM_STATIC BYTE FSE_decodeSymbolFast(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD) +{ + const FSE_decode_t DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state]; + const U32 nbBits = DInfo.nbBits; + BYTE symbol = DInfo.symbol; + size_t lowBits = BIT_readBitsFast(bitD, nbBits); + + DStatePtr->state = DInfo.newState + lowBits; + return symbol; +} + +MEM_STATIC unsigned FSE_endOfDState(const FSE_DState_t* DStatePtr) +{ + return DStatePtr->state == 0; +} + + +#if defined (__cplusplus) +} +#endif + +#endif /* FSE_STATIC_H */ + +/* ****************************************************************** + FSE : Finite State Entropy coder + Copyright (C) 2013-2015, Yann Collet. + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy + - Public forum : https://groups.google.com/forum/#!forum/lz4c +****************************************************************** */ + +#ifndef FSE_COMMONDEFS_ONLY + +/* ************************************************************** +* Tuning parameters +****************************************************************/ +/*!MEMORY_USAGE : +* Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.) +* Increasing memory usage improves compression ratio +* Reduced memory usage can improve speed, due to cache effect +* Recommended max value is 14, for 16KB, which nicely fits into Intel x86 L1 cache */ +#define FSE_MAX_MEMORY_USAGE 14 +#define FSE_DEFAULT_MEMORY_USAGE 13 + +/*!FSE_MAX_SYMBOL_VALUE : +* Maximum symbol value authorized. +* Required for proper stack allocation */ +#define FSE_MAX_SYMBOL_VALUE 255 + + +/* ************************************************************** +* template functions type & suffix +****************************************************************/ +#define FSE_FUNCTION_TYPE BYTE +#define FSE_FUNCTION_EXTENSION +#define FSE_DECODE_TYPE FSE_decode_t + + +#endif /* !FSE_COMMONDEFS_ONLY */ + +/* ************************************************************** +* Compiler specifics +****************************************************************/ +#ifdef _MSC_VER /* Visual Studio */ +# define FORCE_INLINE static __forceinline +# include /* For Visual 2005 */ +# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ +# pragma warning(disable : 4214) /* disable: C4214: non-int bitfields */ +#else +# ifdef __GNUC__ +# define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) +# define FORCE_INLINE static inline __attribute__((always_inline)) +# else +# define FORCE_INLINE static inline +# endif +#endif + + +/* ************************************************************** +* Includes +****************************************************************/ +#include /* malloc, free, qsort */ +#include /* memcpy, memset */ +#include /* printf (debug) */ + + +/* *************************************************************** +* Constants +*****************************************************************/ +#define FSE_MAX_TABLELOG (FSE_MAX_MEMORY_USAGE-2) +#define FSE_MAX_TABLESIZE (1U< FSE_TABLELOG_ABSOLUTE_MAX +#error "FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX is not supported" +#endif + + +/* ************************************************************** +* Error Management +****************************************************************/ +#define FSE_STATIC_ASSERT(c) { enum { FSE_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */ + + +/* ************************************************************** +* Complex types +****************************************************************/ +typedef U32 DTable_max_t[FSE_DTABLE_SIZE_U32(FSE_MAX_TABLELOG)]; + + +/* ************************************************************** +* Templates +****************************************************************/ +/* + designed to be included + for type-specific functions (template emulation in C) + Objective is to write these functions only once, for improved maintenance +*/ + +/* safety checks */ +#ifndef FSE_FUNCTION_EXTENSION +# error "FSE_FUNCTION_EXTENSION must be defined" +#endif +#ifndef FSE_FUNCTION_TYPE +# error "FSE_FUNCTION_TYPE must be defined" +#endif + +/* Function names */ +#define FSE_CAT(X,Y) X##Y +#define FSE_FUNCTION_NAME(X,Y) FSE_CAT(X,Y) +#define FSE_TYPE_NAME(X,Y) FSE_CAT(X,Y) + +static U32 FSE_tableStep(U32 tableSize) { return (tableSize>>1) + (tableSize>>3) + 3; } + + +static size_t FSE_buildDTable(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog) +{ + FSE_DTableHeader DTableH; + void* const tdPtr = dt+1; /* because dt is unsigned, 32-bits aligned on 32-bits */ + FSE_DECODE_TYPE* const tableDecode = (FSE_DECODE_TYPE*) (tdPtr); + const U32 tableSize = 1 << tableLog; + const U32 tableMask = tableSize-1; + const U32 step = FSE_tableStep(tableSize); + U16 symbolNext[FSE_MAX_SYMBOL_VALUE+1]; + U32 position = 0; + U32 highThreshold = tableSize-1; + const S16 largeLimit= (S16)(1 << (tableLog-1)); + U32 noLarge = 1; + U32 s; + + /* Sanity Checks */ + if (maxSymbolValue > FSE_MAX_SYMBOL_VALUE) return ERROR(maxSymbolValue_tooLarge); + if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge); + + /* Init, lay down lowprob symbols */ + DTableH.tableLog = (U16)tableLog; + for (s=0; s<=maxSymbolValue; s++) + { + if (normalizedCounter[s]==-1) + { + tableDecode[highThreshold--].symbol = (FSE_FUNCTION_TYPE)s; + symbolNext[s] = 1; + } + else + { + if (normalizedCounter[s] >= largeLimit) noLarge=0; + symbolNext[s] = normalizedCounter[s]; + } + } + + /* Spread symbols */ + for (s=0; s<=maxSymbolValue; s++) + { + int i; + for (i=0; i highThreshold) position = (position + step) & tableMask; /* lowprob area */ + } + } + + if (position!=0) return ERROR(GENERIC); /* position must reach all cells once, otherwise normalizedCounter is incorrect */ + + /* Build Decoding table */ + { + U32 i; + for (i=0; i FSE_TABLELOG_ABSOLUTE_MAX) return ERROR(tableLog_tooLarge); + bitStream >>= 4; + bitCount = 4; + *tableLogPtr = nbBits; + remaining = (1<1) && (charnum<=*maxSVPtr)) + { + if (previous0) + { + unsigned n0 = charnum; + while ((bitStream & 0xFFFF) == 0xFFFF) + { + n0+=24; + if (ip < iend-5) + { + ip+=2; + bitStream = MEM_readLE32(ip) >> bitCount; + } + else + { + bitStream >>= 16; + bitCount+=16; + } + } + while ((bitStream & 3) == 3) + { + n0+=3; + bitStream>>=2; + bitCount+=2; + } + n0 += bitStream & 3; + bitCount += 2; + if (n0 > *maxSVPtr) return ERROR(maxSymbolValue_tooSmall); + while (charnum < n0) normalizedCounter[charnum++] = 0; + if ((ip <= iend-7) || (ip + (bitCount>>3) <= iend-4)) + { + ip += bitCount>>3; + bitCount &= 7; + bitStream = MEM_readLE32(ip) >> bitCount; + } + else + bitStream >>= 2; + } + { + const short max = (short)((2*threshold-1)-remaining); + short count; + + if ((bitStream & (threshold-1)) < (U32)max) + { + count = (short)(bitStream & (threshold-1)); + bitCount += nbBits-1; + } + else + { + count = (short)(bitStream & (2*threshold-1)); + if (count >= threshold) count -= max; + bitCount += nbBits; + } + + count--; /* extra accuracy */ + remaining -= FSE_abs(count); + normalizedCounter[charnum++] = count; + previous0 = !count; + while (remaining < threshold) + { + nbBits--; + threshold >>= 1; + } + + { + if ((ip <= iend-7) || (ip + (bitCount>>3) <= iend-4)) + { + ip += bitCount>>3; + bitCount &= 7; + } + else + { + bitCount -= (int)(8 * (iend - 4 - ip)); + ip = iend - 4; + } + bitStream = MEM_readLE32(ip) >> (bitCount & 31); + } + } + } + if (remaining != 1) return ERROR(GENERIC); + *maxSVPtr = charnum-1; + + ip += (bitCount+7)>>3; + if ((size_t)(ip-istart) > hbSize) return ERROR(srcSize_wrong); + return ip-istart; +} + + +/********************************************************* +* Decompression (Byte symbols) +*********************************************************/ +static size_t FSE_buildDTable_rle (FSE_DTable* dt, BYTE symbolValue) +{ + void* ptr = dt; + FSE_DTableHeader* const DTableH = (FSE_DTableHeader*)ptr; + void* dPtr = dt + 1; + FSE_decode_t* const cell = (FSE_decode_t*)dPtr; + + DTableH->tableLog = 0; + DTableH->fastMode = 0; + + cell->newState = 0; + cell->symbol = symbolValue; + cell->nbBits = 0; + + return 0; +} + + +static size_t FSE_buildDTable_raw (FSE_DTable* dt, unsigned nbBits) +{ + void* ptr = dt; + FSE_DTableHeader* const DTableH = (FSE_DTableHeader*)ptr; + void* dPtr = dt + 1; + FSE_decode_t* const dinfo = (FSE_decode_t*)dPtr; + const unsigned tableSize = 1 << nbBits; + const unsigned tableMask = tableSize - 1; + const unsigned maxSymbolValue = tableMask; + unsigned s; + + /* Sanity checks */ + if (nbBits < 1) return ERROR(GENERIC); /* min size */ + + /* Build Decoding Table */ + DTableH->tableLog = (U16)nbBits; + DTableH->fastMode = 1; + for (s=0; s<=maxSymbolValue; s++) + { + dinfo[s].newState = 0; + dinfo[s].symbol = (BYTE)s; + dinfo[s].nbBits = (BYTE)nbBits; + } + + return 0; +} + +FORCE_INLINE size_t FSE_decompress_usingDTable_generic( + void* dst, size_t maxDstSize, + const void* cSrc, size_t cSrcSize, + const FSE_DTable* dt, const unsigned fast) +{ + BYTE* const ostart = (BYTE*) dst; + BYTE* op = ostart; + BYTE* const omax = op + maxDstSize; + BYTE* const olimit = omax-3; + + BIT_DStream_t bitD; + FSE_DState_t state1; + FSE_DState_t state2; + size_t errorCode; + + /* Init */ + errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize); /* replaced last arg by maxCompressed Size */ + if (FSE_isError(errorCode)) return errorCode; + + FSE_initDState(&state1, &bitD, dt); + FSE_initDState(&state2, &bitD, dt); + +#define FSE_GETSYMBOL(statePtr) fast ? FSE_decodeSymbolFast(statePtr, &bitD) : FSE_decodeSymbol(statePtr, &bitD) + + /* 4 symbols per loop */ + for ( ; (BIT_reloadDStream(&bitD)==BIT_DStream_unfinished) && (op sizeof(bitD.bitContainer)*8) /* This test must be static */ + BIT_reloadDStream(&bitD); + + op[1] = FSE_GETSYMBOL(&state2); + + if (FSE_MAX_TABLELOG*4+7 > sizeof(bitD.bitContainer)*8) /* This test must be static */ + { if (BIT_reloadDStream(&bitD) > BIT_DStream_unfinished) { op+=2; break; } } + + op[2] = FSE_GETSYMBOL(&state1); + + if (FSE_MAX_TABLELOG*2+7 > sizeof(bitD.bitContainer)*8) /* This test must be static */ + BIT_reloadDStream(&bitD); + + op[3] = FSE_GETSYMBOL(&state2); + } + + /* tail */ + /* note : BIT_reloadDStream(&bitD) >= FSE_DStream_partiallyFilled; Ends at exactly BIT_DStream_completed */ + while (1) + { + if ( (BIT_reloadDStream(&bitD)>BIT_DStream_completed) || (op==omax) || (BIT_endOfDStream(&bitD) && (fast || FSE_endOfDState(&state1))) ) + break; + + *op++ = FSE_GETSYMBOL(&state1); + + if ( (BIT_reloadDStream(&bitD)>BIT_DStream_completed) || (op==omax) || (BIT_endOfDStream(&bitD) && (fast || FSE_endOfDState(&state2))) ) + break; + + *op++ = FSE_GETSYMBOL(&state2); + } + + /* end ? */ + if (BIT_endOfDStream(&bitD) && FSE_endOfDState(&state1) && FSE_endOfDState(&state2)) + return op-ostart; + + if (op==omax) return ERROR(dstSize_tooSmall); /* dst buffer is full, but cSrc unfinished */ + + return ERROR(corruption_detected); +} + + +static size_t FSE_decompress_usingDTable(void* dst, size_t originalSize, + const void* cSrc, size_t cSrcSize, + const FSE_DTable* dt) +{ + const void* ptr = dt; + const FSE_DTableHeader* DTableH = (const FSE_DTableHeader*)ptr; + const U32 fastMode = DTableH->fastMode; + + /* select fast mode (static) */ + if (fastMode) return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 1); + return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 0); +} + + +static size_t FSE_decompress(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize) +{ + const BYTE* const istart = (const BYTE*)cSrc; + const BYTE* ip = istart; + short counting[FSE_MAX_SYMBOL_VALUE+1]; + DTable_max_t dt; /* Static analyzer seems unable to understand this table will be properly initialized later */ + unsigned tableLog; + unsigned maxSymbolValue = FSE_MAX_SYMBOL_VALUE; + size_t errorCode; + + if (cSrcSize<2) return ERROR(srcSize_wrong); /* too small input size */ + + /* normal FSE decoding mode */ + errorCode = FSE_readNCount (counting, &maxSymbolValue, &tableLog, istart, cSrcSize); + if (FSE_isError(errorCode)) return errorCode; + if (errorCode >= cSrcSize) return ERROR(srcSize_wrong); /* too small input size */ + ip += errorCode; + cSrcSize -= errorCode; + + errorCode = FSE_buildDTable (dt, counting, maxSymbolValue, tableLog); + if (FSE_isError(errorCode)) return errorCode; + + /* always return, even if it is an error code */ + return FSE_decompress_usingDTable (dst, maxDstSize, ip, cSrcSize, dt); +} + + + +#endif /* FSE_COMMONDEFS_ONLY */ + + +/* ****************************************************************** + Huff0 : Huffman coder, part of New Generation Entropy library + header file + Copyright (C) 2013-2015, Yann Collet. + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - Source repository : https://github.com/Cyan4973/FiniteStateEntropy + - Public forum : https://groups.google.com/forum/#!forum/lz4c +****************************************************************** */ +#ifndef HUFF0_H +#define HUFF0_H + +#if defined (__cplusplus) +extern "C" { +#endif + + +/* **************************************** +* Dependency +******************************************/ +#include /* size_t */ + + +/* **************************************** +* Huff0 simple functions +******************************************/ +static size_t HUF_decompress(void* dst, size_t dstSize, + const void* cSrc, size_t cSrcSize); +/*! +HUF_decompress(): + Decompress Huff0 data from buffer 'cSrc', of size 'cSrcSize', + into already allocated destination buffer 'dst', of size 'dstSize'. + 'dstSize' must be the exact size of original (uncompressed) data. + Note : in contrast with FSE, HUF_decompress can regenerate RLE (cSrcSize==1) and uncompressed (cSrcSize==dstSize) data, because it knows size to regenerate. + @return : size of regenerated data (== dstSize) + or an error code, which can be tested using HUF_isError() +*/ + + +/* **************************************** +* Tool functions +******************************************/ +/* Error Management */ +static unsigned HUF_isError(size_t code); /* tells if a return value is an error code */ + + +#if defined (__cplusplus) +} +#endif + +#endif /* HUFF0_H */ + + +/* ****************************************************************** + Huff0 : Huffman coder, part of New Generation Entropy library + header file for static linking (only) + Copyright (C) 2013-2015, Yann Collet + + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are + met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following disclaimer + in the documentation and/or other materials provided with the + distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + + You can contact the author at : + - Source repository : https://github.com/Cyan4973/FiniteStateEntropy + - Public forum : https://groups.google.com/forum/#!forum/lz4c +****************************************************************** */ +#ifndef HUFF0_STATIC_H +#define HUFF0_STATIC_H + +#if defined (__cplusplus) +extern "C" { +#endif + + +/* **************************************** +* Dependency +******************************************/ +#include "huff0.h" + + +/* **************************************** +* Static allocation macros +******************************************/ +/* static allocation of Huff0's DTable */ +#define HUF_DTABLE_SIZE(maxTableLog) (1 + (1<= 199901L) /* C99 */) +/* inline is defined */ +#elif defined(_MSC_VER) +# define inline __inline +#else +# define inline /* disable inline */ +#endif + + +#ifdef _MSC_VER /* Visual Studio */ +# define FORCE_INLINE static __forceinline +# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ +#else +# ifdef __GNUC__ +# define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) +# define FORCE_INLINE static inline __attribute__((always_inline)) +# else +# define FORCE_INLINE static inline +# endif +#endif + + +/* ************************************************************** +* Includes +****************************************************************/ +#include /* malloc, free, qsort */ +#include /* memcpy, memset */ +#include /* printf (debug) */ + + +/* ************************************************************** +* Constants +****************************************************************/ +#define HUF_ABSOLUTEMAX_TABLELOG 16 /* absolute limit of HUF_MAX_TABLELOG. Beyond that value, code does not work */ +#define HUF_MAX_TABLELOG 12 /* max configured tableLog (for static allocation); can be modified up to HUF_ABSOLUTEMAX_TABLELOG */ +#define HUF_DEFAULT_TABLELOG HUF_MAX_TABLELOG /* tableLog by default, when not specified */ +#define HUF_MAX_SYMBOL_VALUE 255 +#if (HUF_MAX_TABLELOG > HUF_ABSOLUTEMAX_TABLELOG) +# error "HUF_MAX_TABLELOG is too large !" +#endif + + +/* ************************************************************** +* Error Management +****************************************************************/ +static unsigned HUF_isError(size_t code) { return ERR_isError(code); } +#define HUF_STATIC_ASSERT(c) { enum { HUF_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */ + + + +/*-******************************************************* +* Huff0 : Huffman block decompression +*********************************************************/ +typedef struct { BYTE byte; BYTE nbBits; } HUF_DEltX2; /* single-symbol decoding */ + +typedef struct { U16 sequence; BYTE nbBits; BYTE length; } HUF_DEltX4; /* double-symbols decoding */ + +typedef struct { BYTE symbol; BYTE weight; } sortedSymbol_t; + +/*! HUF_readStats + Read compact Huffman tree, saved by HUF_writeCTable + @huffWeight : destination buffer + @return : size read from `src` +*/ +static size_t HUF_readStats(BYTE* huffWeight, size_t hwSize, U32* rankStats, + U32* nbSymbolsPtr, U32* tableLogPtr, + const void* src, size_t srcSize) +{ + U32 weightTotal; + U32 tableLog; + const BYTE* ip = (const BYTE*) src; + size_t iSize = ip[0]; + size_t oSize; + U32 n; + + //memset(huffWeight, 0, hwSize); /* is not necessary, even though some analyzer complain ... */ + + if (iSize >= 128) /* special header */ + { + if (iSize >= (242)) /* RLE */ + { + static int l[14] = { 1, 2, 3, 4, 7, 8, 15, 16, 31, 32, 63, 64, 127, 128 }; + oSize = l[iSize-242]; + memset(huffWeight, 1, hwSize); + iSize = 0; + } + else /* Incompressible */ + { + oSize = iSize - 127; + iSize = ((oSize+1)/2); + if (iSize+1 > srcSize) return ERROR(srcSize_wrong); + if (oSize >= hwSize) return ERROR(corruption_detected); + ip += 1; + for (n=0; n> 4; + huffWeight[n+1] = ip[n/2] & 15; + } + } + } + else /* header compressed with FSE (normal case) */ + { + if (iSize+1 > srcSize) return ERROR(srcSize_wrong); + oSize = FSE_decompress(huffWeight, hwSize-1, ip+1, iSize); /* max (hwSize-1) values decoded, as last one is implied */ + if (FSE_isError(oSize)) return oSize; + } + + /* collect weight stats */ + memset(rankStats, 0, (HUF_ABSOLUTEMAX_TABLELOG + 1) * sizeof(U32)); + weightTotal = 0; + for (n=0; n= HUF_ABSOLUTEMAX_TABLELOG) return ERROR(corruption_detected); + rankStats[huffWeight[n]]++; + weightTotal += (1 << huffWeight[n]) >> 1; + } + + /* get last non-null symbol weight (implied, total must be 2^n) */ + tableLog = BIT_highbit32(weightTotal) + 1; + if (tableLog > HUF_ABSOLUTEMAX_TABLELOG) return ERROR(corruption_detected); + { + U32 total = 1 << tableLog; + U32 rest = total - weightTotal; + U32 verif = 1 << BIT_highbit32(rest); + U32 lastWeight = BIT_highbit32(rest) + 1; + if (verif != rest) return ERROR(corruption_detected); /* last value must be a clean power of 2 */ + huffWeight[oSize] = (BYTE)lastWeight; + rankStats[lastWeight]++; + } + + /* check tree construction validity */ + if ((rankStats[1] < 2) || (rankStats[1] & 1)) return ERROR(corruption_detected); /* by construction : at least 2 elts of rank 1, must be even */ + + /* results */ + *nbSymbolsPtr = (U32)(oSize+1); + *tableLogPtr = tableLog; + return iSize+1; +} + + +/**************************/ +/* single-symbol decoding */ +/**************************/ + +static size_t HUF_readDTableX2 (U16* DTable, const void* src, size_t srcSize) +{ + BYTE huffWeight[HUF_MAX_SYMBOL_VALUE + 1]; + U32 rankVal[HUF_ABSOLUTEMAX_TABLELOG + 1]; /* large enough for values from 0 to 16 */ + U32 tableLog = 0; + size_t iSize; + U32 nbSymbols = 0; + U32 n; + U32 nextRankStart; + void* const dtPtr = DTable + 1; + HUF_DEltX2* const dt = (HUF_DEltX2*)dtPtr; + + HUF_STATIC_ASSERT(sizeof(HUF_DEltX2) == sizeof(U16)); /* if compilation fails here, assertion is false */ + //memset(huffWeight, 0, sizeof(huffWeight)); /* is not necessary, even though some analyzer complain ... */ + + iSize = HUF_readStats(huffWeight, HUF_MAX_SYMBOL_VALUE + 1, rankVal, &nbSymbols, &tableLog, src, srcSize); + if (HUF_isError(iSize)) return iSize; + + /* check result */ + if (tableLog > DTable[0]) return ERROR(tableLog_tooLarge); /* DTable is too small */ + DTable[0] = (U16)tableLog; /* maybe should separate sizeof DTable, as allocated, from used size of DTable, in case of DTable re-use */ + + /* Prepare ranks */ + nextRankStart = 0; + for (n=1; n<=tableLog; n++) + { + U32 current = nextRankStart; + nextRankStart += (rankVal[n] << (n-1)); + rankVal[n] = current; + } + + /* fill DTable */ + for (n=0; n> 1; + U32 i; + HUF_DEltX2 D; + D.byte = (BYTE)n; D.nbBits = (BYTE)(tableLog + 1 - w); + for (i = rankVal[w]; i < rankVal[w] + length; i++) + dt[i] = D; + rankVal[w] += length; + } + + return iSize; +} + +static BYTE HUF_decodeSymbolX2(BIT_DStream_t* Dstream, const HUF_DEltX2* dt, const U32 dtLog) +{ + const size_t val = BIT_lookBitsFast(Dstream, dtLog); /* note : dtLog >= 1 */ + const BYTE c = dt[val].byte; + BIT_skipBits(Dstream, dt[val].nbBits); + return c; +} + +#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) \ + *ptr++ = HUF_decodeSymbolX2(DStreamPtr, dt, dtLog) + +#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \ + if (MEM_64bits() || (HUF_MAX_TABLELOG<=12)) \ + HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) + +#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \ + if (MEM_64bits()) \ + HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) + +static inline size_t HUF_decodeStreamX2(BYTE* p, BIT_DStream_t* const bitDPtr, BYTE* const pEnd, const HUF_DEltX2* const dt, const U32 dtLog) +{ + BYTE* const pStart = p; + + /* up to 4 symbols at a time */ + while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p <= pEnd-4)) + { + HUF_DECODE_SYMBOLX2_2(p, bitDPtr); + HUF_DECODE_SYMBOLX2_1(p, bitDPtr); + HUF_DECODE_SYMBOLX2_2(p, bitDPtr); + HUF_DECODE_SYMBOLX2_0(p, bitDPtr); + } + + /* closer to the end */ + while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p < pEnd)) + HUF_DECODE_SYMBOLX2_0(p, bitDPtr); + + /* no more data to retrieve from bitstream, hence no need to reload */ + while (p < pEnd) + HUF_DECODE_SYMBOLX2_0(p, bitDPtr); + + return pEnd-pStart; +} + + +static size_t HUF_decompress4X2_usingDTable( + void* dst, size_t dstSize, + const void* cSrc, size_t cSrcSize, + const U16* DTable) +{ + if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */ + + { + const BYTE* const istart = (const BYTE*) cSrc; + BYTE* const ostart = (BYTE*) dst; + BYTE* const oend = ostart + dstSize; + const void* const dtPtr = DTable; + const HUF_DEltX2* const dt = ((const HUF_DEltX2*)dtPtr) +1; + const U32 dtLog = DTable[0]; + size_t errorCode; + + /* Init */ + BIT_DStream_t bitD1; + BIT_DStream_t bitD2; + BIT_DStream_t bitD3; + BIT_DStream_t bitD4; + const size_t length1 = MEM_readLE16(istart); + const size_t length2 = MEM_readLE16(istart+2); + const size_t length3 = MEM_readLE16(istart+4); + size_t length4; + const BYTE* const istart1 = istart + 6; /* jumpTable */ + const BYTE* const istart2 = istart1 + length1; + const BYTE* const istart3 = istart2 + length2; + const BYTE* const istart4 = istart3 + length3; + const size_t segmentSize = (dstSize+3) / 4; + BYTE* const opStart2 = ostart + segmentSize; + BYTE* const opStart3 = opStart2 + segmentSize; + BYTE* const opStart4 = opStart3 + segmentSize; + BYTE* op1 = ostart; + BYTE* op2 = opStart2; + BYTE* op3 = opStart3; + BYTE* op4 = opStart4; + U32 endSignal; + + length4 = cSrcSize - (length1 + length2 + length3 + 6); + if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */ + errorCode = BIT_initDStream(&bitD1, istart1, length1); + if (HUF_isError(errorCode)) return errorCode; + errorCode = BIT_initDStream(&bitD2, istart2, length2); + if (HUF_isError(errorCode)) return errorCode; + errorCode = BIT_initDStream(&bitD3, istart3, length3); + if (HUF_isError(errorCode)) return errorCode; + errorCode = BIT_initDStream(&bitD4, istart4, length4); + if (HUF_isError(errorCode)) return errorCode; + + /* 16-32 symbols per loop (4-8 symbols per stream) */ + endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); + for ( ; (endSignal==BIT_DStream_unfinished) && (op4<(oend-7)) ; ) + { + HUF_DECODE_SYMBOLX2_2(op1, &bitD1); + HUF_DECODE_SYMBOLX2_2(op2, &bitD2); + HUF_DECODE_SYMBOLX2_2(op3, &bitD3); + HUF_DECODE_SYMBOLX2_2(op4, &bitD4); + HUF_DECODE_SYMBOLX2_1(op1, &bitD1); + HUF_DECODE_SYMBOLX2_1(op2, &bitD2); + HUF_DECODE_SYMBOLX2_1(op3, &bitD3); + HUF_DECODE_SYMBOLX2_1(op4, &bitD4); + HUF_DECODE_SYMBOLX2_2(op1, &bitD1); + HUF_DECODE_SYMBOLX2_2(op2, &bitD2); + HUF_DECODE_SYMBOLX2_2(op3, &bitD3); + HUF_DECODE_SYMBOLX2_2(op4, &bitD4); + HUF_DECODE_SYMBOLX2_0(op1, &bitD1); + HUF_DECODE_SYMBOLX2_0(op2, &bitD2); + HUF_DECODE_SYMBOLX2_0(op3, &bitD3); + HUF_DECODE_SYMBOLX2_0(op4, &bitD4); + + endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); + } + + /* check corruption */ + if (op1 > opStart2) return ERROR(corruption_detected); + if (op2 > opStart3) return ERROR(corruption_detected); + if (op3 > opStart4) return ERROR(corruption_detected); + /* note : op4 supposed already verified within main loop */ + + /* finish bitStreams one by one */ + HUF_decodeStreamX2(op1, &bitD1, opStart2, dt, dtLog); + HUF_decodeStreamX2(op2, &bitD2, opStart3, dt, dtLog); + HUF_decodeStreamX2(op3, &bitD3, opStart4, dt, dtLog); + HUF_decodeStreamX2(op4, &bitD4, oend, dt, dtLog); + + /* check */ + endSignal = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4); + if (!endSignal) return ERROR(corruption_detected); + + /* decoded size */ + return dstSize; + } +} + + +static size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize) +{ + HUF_CREATE_STATIC_DTABLEX2(DTable, HUF_MAX_TABLELOG); + const BYTE* ip = (const BYTE*) cSrc; + size_t errorCode; + + errorCode = HUF_readDTableX2 (DTable, cSrc, cSrcSize); + if (HUF_isError(errorCode)) return errorCode; + if (errorCode >= cSrcSize) return ERROR(srcSize_wrong); + ip += errorCode; + cSrcSize -= errorCode; + + return HUF_decompress4X2_usingDTable (dst, dstSize, ip, cSrcSize, DTable); +} + + +/***************************/ +/* double-symbols decoding */ +/***************************/ + +static void HUF_fillDTableX4Level2(HUF_DEltX4* DTable, U32 sizeLog, const U32 consumed, + const U32* rankValOrigin, const int minWeight, + const sortedSymbol_t* sortedSymbols, const U32 sortedListSize, + U32 nbBitsBaseline, U16 baseSeq) +{ + HUF_DEltX4 DElt; + U32 rankVal[HUF_ABSOLUTEMAX_TABLELOG + 1]; + U32 s; + + /* get pre-calculated rankVal */ + memcpy(rankVal, rankValOrigin, sizeof(rankVal)); + + /* fill skipped values */ + if (minWeight>1) + { + U32 i, skipSize = rankVal[minWeight]; + MEM_writeLE16(&(DElt.sequence), baseSeq); + DElt.nbBits = (BYTE)(consumed); + DElt.length = 1; + for (i = 0; i < skipSize; i++) + DTable[i] = DElt; + } + + /* fill DTable */ + for (s=0; s= 1 */ + + rankVal[weight] += length; + } +} + +typedef U32 rankVal_t[HUF_ABSOLUTEMAX_TABLELOG][HUF_ABSOLUTEMAX_TABLELOG + 1]; + +static void HUF_fillDTableX4(HUF_DEltX4* DTable, const U32 targetLog, + const sortedSymbol_t* sortedList, const U32 sortedListSize, + const U32* rankStart, rankVal_t rankValOrigin, const U32 maxWeight, + const U32 nbBitsBaseline) +{ + U32 rankVal[HUF_ABSOLUTEMAX_TABLELOG + 1]; + const int scaleLog = nbBitsBaseline - targetLog; /* note : targetLog >= srcLog, hence scaleLog <= 1 */ + const U32 minBits = nbBitsBaseline - maxWeight; + U32 s; + + memcpy(rankVal, rankValOrigin, sizeof(rankVal)); + + /* fill DTable */ + for (s=0; s= minBits) /* enough room for a second symbol */ + { + U32 sortedRank; + int minWeight = nbBits + scaleLog; + if (minWeight < 1) minWeight = 1; + sortedRank = rankStart[minWeight]; + HUF_fillDTableX4Level2(DTable+start, targetLog-nbBits, nbBits, + rankValOrigin[nbBits], minWeight, + sortedList+sortedRank, sortedListSize-sortedRank, + nbBitsBaseline, symbol); + } + else + { + U32 i; + const U32 end = start + length; + HUF_DEltX4 DElt; + + MEM_writeLE16(&(DElt.sequence), symbol); + DElt.nbBits = (BYTE)(nbBits); + DElt.length = 1; + for (i = start; i < end; i++) + DTable[i] = DElt; + } + rankVal[weight] += length; + } +} + +static size_t HUF_readDTableX4 (U32* DTable, const void* src, size_t srcSize) +{ + BYTE weightList[HUF_MAX_SYMBOL_VALUE + 1]; + sortedSymbol_t sortedSymbol[HUF_MAX_SYMBOL_VALUE + 1]; + U32 rankStats[HUF_ABSOLUTEMAX_TABLELOG + 1] = { 0 }; + U32 rankStart0[HUF_ABSOLUTEMAX_TABLELOG + 2] = { 0 }; + U32* const rankStart = rankStart0+1; + rankVal_t rankVal; + U32 tableLog, maxW, sizeOfSort, nbSymbols; + const U32 memLog = DTable[0]; + size_t iSize; + void* dtPtr = DTable; + HUF_DEltX4* const dt = ((HUF_DEltX4*)dtPtr) + 1; + + HUF_STATIC_ASSERT(sizeof(HUF_DEltX4) == sizeof(U32)); /* if compilation fails here, assertion is false */ + if (memLog > HUF_ABSOLUTEMAX_TABLELOG) return ERROR(tableLog_tooLarge); + //memset(weightList, 0, sizeof(weightList)); /* is not necessary, even though some analyzer complain ... */ + + iSize = HUF_readStats(weightList, HUF_MAX_SYMBOL_VALUE + 1, rankStats, &nbSymbols, &tableLog, src, srcSize); + if (HUF_isError(iSize)) return iSize; + + /* check result */ + if (tableLog > memLog) return ERROR(tableLog_tooLarge); /* DTable can't fit code depth */ + + /* find maxWeight */ + for (maxW = tableLog; rankStats[maxW]==0; maxW--) {} /* necessarily finds a solution before 0 */ + + /* Get start index of each weight */ + { + U32 w, nextRankStart = 0; + for (w=1; w<=maxW; w++) + { + U32 current = nextRankStart; + nextRankStart += rankStats[w]; + rankStart[w] = current; + } + rankStart[0] = nextRankStart; /* put all 0w symbols at the end of sorted list*/ + sizeOfSort = nextRankStart; + } + + /* sort symbols by weight */ + { + U32 s; + for (s=0; s> consumed; + } + } + } + + HUF_fillDTableX4(dt, memLog, + sortedSymbol, sizeOfSort, + rankStart0, rankVal, maxW, + tableLog+1); + + return iSize; +} + + +static U32 HUF_decodeSymbolX4(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog) +{ + const size_t val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */ + memcpy(op, dt+val, 2); + BIT_skipBits(DStream, dt[val].nbBits); + return dt[val].length; +} + +static U32 HUF_decodeLastSymbolX4(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog) +{ + const size_t val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */ + memcpy(op, dt+val, 1); + if (dt[val].length==1) BIT_skipBits(DStream, dt[val].nbBits); + else + { + if (DStream->bitsConsumed < (sizeof(DStream->bitContainer)*8)) + { + BIT_skipBits(DStream, dt[val].nbBits); + if (DStream->bitsConsumed > (sizeof(DStream->bitContainer)*8)) + DStream->bitsConsumed = (sizeof(DStream->bitContainer)*8); /* ugly hack; works only because it's the last symbol. Note : can't easily extract nbBits from just this symbol */ + } + } + return 1; +} + + +#define HUF_DECODE_SYMBOLX4_0(ptr, DStreamPtr) \ + ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog) + +#define HUF_DECODE_SYMBOLX4_1(ptr, DStreamPtr) \ + if (MEM_64bits() || (HUF_MAX_TABLELOG<=12)) \ + ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog) + +#define HUF_DECODE_SYMBOLX4_2(ptr, DStreamPtr) \ + if (MEM_64bits()) \ + ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog) + +static inline size_t HUF_decodeStreamX4(BYTE* p, BIT_DStream_t* bitDPtr, BYTE* const pEnd, const HUF_DEltX4* const dt, const U32 dtLog) +{ + BYTE* const pStart = p; + + /* up to 8 symbols at a time */ + while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p < pEnd-7)) + { + HUF_DECODE_SYMBOLX4_2(p, bitDPtr); + HUF_DECODE_SYMBOLX4_1(p, bitDPtr); + HUF_DECODE_SYMBOLX4_2(p, bitDPtr); + HUF_DECODE_SYMBOLX4_0(p, bitDPtr); + } + + /* closer to the end */ + while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p <= pEnd-2)) + HUF_DECODE_SYMBOLX4_0(p, bitDPtr); + + while (p <= pEnd-2) + HUF_DECODE_SYMBOLX4_0(p, bitDPtr); /* no need to reload : reached the end of DStream */ + + if (p < pEnd) + p += HUF_decodeLastSymbolX4(p, bitDPtr, dt, dtLog); + + return p-pStart; +} + +static size_t HUF_decompress4X4_usingDTable( + void* dst, size_t dstSize, + const void* cSrc, size_t cSrcSize, + const U32* DTable) +{ + if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */ + + { + const BYTE* const istart = (const BYTE*) cSrc; + BYTE* const ostart = (BYTE*) dst; + BYTE* const oend = ostart + dstSize; + const void* const dtPtr = DTable; + const HUF_DEltX4* const dt = ((const HUF_DEltX4*)dtPtr) +1; + const U32 dtLog = DTable[0]; + size_t errorCode; + + /* Init */ + BIT_DStream_t bitD1; + BIT_DStream_t bitD2; + BIT_DStream_t bitD3; + BIT_DStream_t bitD4; + const size_t length1 = MEM_readLE16(istart); + const size_t length2 = MEM_readLE16(istart+2); + const size_t length3 = MEM_readLE16(istart+4); + size_t length4; + const BYTE* const istart1 = istart + 6; /* jumpTable */ + const BYTE* const istart2 = istart1 + length1; + const BYTE* const istart3 = istart2 + length2; + const BYTE* const istart4 = istart3 + length3; + const size_t segmentSize = (dstSize+3) / 4; + BYTE* const opStart2 = ostart + segmentSize; + BYTE* const opStart3 = opStart2 + segmentSize; + BYTE* const opStart4 = opStart3 + segmentSize; + BYTE* op1 = ostart; + BYTE* op2 = opStart2; + BYTE* op3 = opStart3; + BYTE* op4 = opStart4; + U32 endSignal; + + length4 = cSrcSize - (length1 + length2 + length3 + 6); + if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */ + errorCode = BIT_initDStream(&bitD1, istart1, length1); + if (HUF_isError(errorCode)) return errorCode; + errorCode = BIT_initDStream(&bitD2, istart2, length2); + if (HUF_isError(errorCode)) return errorCode; + errorCode = BIT_initDStream(&bitD3, istart3, length3); + if (HUF_isError(errorCode)) return errorCode; + errorCode = BIT_initDStream(&bitD4, istart4, length4); + if (HUF_isError(errorCode)) return errorCode; + + /* 16-32 symbols per loop (4-8 symbols per stream) */ + endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); + for ( ; (endSignal==BIT_DStream_unfinished) && (op4<(oend-7)) ; ) + { + HUF_DECODE_SYMBOLX4_2(op1, &bitD1); + HUF_DECODE_SYMBOLX4_2(op2, &bitD2); + HUF_DECODE_SYMBOLX4_2(op3, &bitD3); + HUF_DECODE_SYMBOLX4_2(op4, &bitD4); + HUF_DECODE_SYMBOLX4_1(op1, &bitD1); + HUF_DECODE_SYMBOLX4_1(op2, &bitD2); + HUF_DECODE_SYMBOLX4_1(op3, &bitD3); + HUF_DECODE_SYMBOLX4_1(op4, &bitD4); + HUF_DECODE_SYMBOLX4_2(op1, &bitD1); + HUF_DECODE_SYMBOLX4_2(op2, &bitD2); + HUF_DECODE_SYMBOLX4_2(op3, &bitD3); + HUF_DECODE_SYMBOLX4_2(op4, &bitD4); + HUF_DECODE_SYMBOLX4_0(op1, &bitD1); + HUF_DECODE_SYMBOLX4_0(op2, &bitD2); + HUF_DECODE_SYMBOLX4_0(op3, &bitD3); + HUF_DECODE_SYMBOLX4_0(op4, &bitD4); + + endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); + } + + /* check corruption */ + if (op1 > opStart2) return ERROR(corruption_detected); + if (op2 > opStart3) return ERROR(corruption_detected); + if (op3 > opStart4) return ERROR(corruption_detected); + /* note : op4 supposed already verified within main loop */ + + /* finish bitStreams one by one */ + HUF_decodeStreamX4(op1, &bitD1, opStart2, dt, dtLog); + HUF_decodeStreamX4(op2, &bitD2, opStart3, dt, dtLog); + HUF_decodeStreamX4(op3, &bitD3, opStart4, dt, dtLog); + HUF_decodeStreamX4(op4, &bitD4, oend, dt, dtLog); + + /* check */ + endSignal = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4); + if (!endSignal) return ERROR(corruption_detected); + + /* decoded size */ + return dstSize; + } +} + + +static size_t HUF_decompress4X4 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize) +{ + HUF_CREATE_STATIC_DTABLEX4(DTable, HUF_MAX_TABLELOG); + const BYTE* ip = (const BYTE*) cSrc; + + size_t hSize = HUF_readDTableX4 (DTable, cSrc, cSrcSize); + if (HUF_isError(hSize)) return hSize; + if (hSize >= cSrcSize) return ERROR(srcSize_wrong); + ip += hSize; + cSrcSize -= hSize; + + return HUF_decompress4X4_usingDTable (dst, dstSize, ip, cSrcSize, DTable); +} + + +/**********************************/ +/* quad-symbol decoding */ +/**********************************/ +typedef struct { BYTE nbBits; BYTE nbBytes; } HUF_DDescX6; +typedef union { BYTE byte[4]; U32 sequence; } HUF_DSeqX6; + +/* recursive, up to level 3; may benefit from