diff --git a/lib/legacy/zstd_v03.c b/lib/legacy/zstd_v03.c index 7d82db666..6af6c88c3 100644 --- a/lib/legacy/zstd_v03.c +++ b/lib/legacy/zstd_v03.c @@ -5,3101 +5,16 @@ * This source code is licensed under both the BSD-style license (found in the * LICENSE file in the root directory of this source tree) and the GPLv2 (found * in the COPYING file in the root directory of this source tree). - * You may select, at your option, one of the above-listed licenses. + * You may select, at your option, one or both licenses. */ - -#include /* size_t, ptrdiff_t */ +#include /* size_t */ #include "zstd_v03.h" #include "../common/compiler.h" #include "../common/error_private.h" -/****************************************** -* Compiler-specific -******************************************/ -#if defined(_MSC_VER) /* Visual Studio */ -# include /* _byteswap_ulong */ -# include /* _byteswap_* */ -#endif - - - -/* ****************************************************************** - mem.h - low-level memory access routines - Copyright (C) 2013-2015, Yann Collet. - - BSD 2-Clause License (https://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 */ - - -/**************************************************************** -* Basic Types -*****************************************************************/ -#if defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) -# if defined(_AIX) -# include -# else -# include /* intptr_t */ -# endif - 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_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]; -} - -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 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_readLE24(const void* memPtr) -{ - return MEM_readLE16(memPtr) + (((const BYTE*)memPtr)[2] << 16); -} - -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 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 size_t MEM_readLEST(const void* memPtr) -{ - if (MEM_32bits()) - return (size_t)MEM_readLE32(memPtr); - else - return (size_t)MEM_readLE64(memPtr); -} - - -#if defined (__cplusplus) -} -#endif - -#endif /* MEM_H_MODULE */ - - -/* ****************************************************************** - bitstream - Part of NewGen Entropy library - header file (to include) - Copyright (C) 2013-2015, Yann Collet. - - BSD 2-Clause License (https://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); - - - -/****************************************** -* 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 (U32 val) -{ -# if defined(_MSC_VER) /* Visual */ - unsigned long r; - return _BitScanReverse(&r, val) ? (unsigned)r : 0; -# elif defined(__GNUC__) && (__GNUC__ >= 3) /* Use GCC Intrinsic */ - return __builtin_clz (val) ^ 31; -# 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); - /* fallthrough */ - case 6: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[5]) << (sizeof(size_t)*8 - 24); - /* fallthrough */ - case 5: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[4]) << (sizeof(size_t)*8 - 32); - /* fallthrough */ - case 4: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[3]) << 24; - /* fallthrough */ - case 3: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[2]) << 16; - /* fallthrough */ - case 2: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[1]) << 8; - /* fallthrough */ - 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; -} -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 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; -} - -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 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 */ -/* ****************************************************************** - Error codes and messages - Copyright (C) 2013-2015, Yann Collet - - BSD 2-Clause License (https://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 ERROR_H_MODULE -#define ERROR_H_MODULE - -#if defined (__cplusplus) -extern "C" { -#endif - - -/****************************************** -* Compiler-specific -******************************************/ -#if defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) -# define ERR_STATIC static inline -#elif defined(_MSC_VER) -# define ERR_STATIC static __inline -#elif defined(__GNUC__) -# define ERR_STATIC static __attribute__((unused)) -#else -# define ERR_STATIC static /* this version may generate warnings for unused static functions; disable the relevant warning */ -#endif - - -/****************************************** -* Error Management -******************************************/ -#define PREFIX(name) ZSTD_error_##name - -#define ERROR(name) (size_t)-PREFIX(name) - -#define ERROR_LIST(ITEM) \ - ITEM(PREFIX(No_Error)) ITEM(PREFIX(GENERIC)) \ - ITEM(PREFIX(dstSize_tooSmall)) ITEM(PREFIX(srcSize_wrong)) \ - ITEM(PREFIX(prefix_unknown)) ITEM(PREFIX(corruption_detected)) \ - ITEM(PREFIX(tableLog_tooLarge)) ITEM(PREFIX(maxSymbolValue_tooLarge)) ITEM(PREFIX(maxSymbolValue_tooSmall)) \ - ITEM(PREFIX(maxCode)) - -#define ERROR_GENERATE_ENUM(ENUM) ENUM, -typedef enum { ERROR_LIST(ERROR_GENERATE_ENUM) } ERR_codes; /* enum is exposed, to detect & handle specific errors; compare function result to -enum value */ - -#define ERROR_CONVERTTOSTRING(STRING) #STRING, -#define ERROR_GENERATE_STRING(EXPR) ERROR_CONVERTTOSTRING(EXPR) -static const char* ERR_strings[] = { ERROR_LIST(ERROR_GENERATE_STRING) }; - -ERR_STATIC unsigned ERR_isError(size_t code) { return (code > ERROR(maxCode)); } - -ERR_STATIC const char* ERR_getErrorName(size_t code) -{ - static const char* codeError = "Unspecified error code"; - if (ERR_isError(code)) return ERR_strings[-(int)(code)]; - return codeError; -} - - -#if defined (__cplusplus) -} -#endif - -#endif /* ERROR_H_MODULE */ -/* -Constructor and Destructor of type FSE_CTable - Note that its size depends on 'tableLog' and 'maxSymbolValue' */ -typedef unsigned FSE_CTable; /* don't allocate that. It's just a way to be more restrictive than void* */ -typedef unsigned FSE_DTable; /* don't allocate that. It's just a way to be more restrictive than void* */ - - -/* ****************************************************************** - FSE : Finite State Entropy coder - header file for static linking (only) - Copyright (C) 2013-2015, Yann Collet - - BSD 2-Clause License (https://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 -****************************************************************** */ -#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 */ - -/* You can statically allocate FSE CTable/DTable as a table of unsigned using below macro */ -#define FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) (1 + (1<<(maxTableLog-1)) + ((maxSymbolValue+1)*2)) -#define FSE_DTABLE_SIZE_U32(maxTableLog) (1 + (1<= 1 (otherwise, result will be corrupted) */ - - -/****************************************** -* Implementation of inline 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) -{ - FSE_DTableHeader DTableH; - memcpy(&DTableH, dt, sizeof(DTableH)); - 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 -/* ****************************************************************** - 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 (https://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 -****************************************************************** */ - -#if defined (__cplusplus) -extern "C" { -#endif - -/****************************************** -* Static allocation macros -******************************************/ -/* Huff0 buffer bounds */ -#define HUF_CTABLEBOUND 129 -#define HUF_BLOCKBOUND(size) (size + (size>>8) + 8) /* only true if incompressible pre-filtered with fast heuristic */ -#define HUF_COMPRESSBOUND(size) (HUF_CTABLEBOUND + HUF_BLOCKBOUND(size)) /* Macro version, useful for static allocation */ - -/* static allocation of Huff0's DTable */ -#define HUF_DTABLE_SIZE(maxTableLog) (1 + (1< /* size_t */ - - -/* ************************************* -* Version -***************************************/ -#define ZSTD_VERSION_MAJOR 0 /* for breaking interface changes */ -#define ZSTD_VERSION_MINOR 2 /* for new (non-breaking) interface capabilities */ -#define ZSTD_VERSION_RELEASE 2 /* for tweaks, bug-fixes, or development */ -#define ZSTD_VERSION_NUMBER (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE) - - -/* ************************************* -* Advanced functions -***************************************/ -typedef struct ZSTD_CCtx_s ZSTD_CCtx; /* incomplete type */ - -#if defined (__cplusplus) -} -#endif -/* - zstd - standard compression library - Header File for static linking only - Copyright (C) 2014-2015, Yann Collet. - - BSD 2-Clause License (https://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 -*/ - -/* The objects defined into this file should be considered experimental. - * They are not labelled stable, as their prototype may change in the future. - * You can use them for tests, provide feedback, or if you can endure risk of future changes. - */ - -#if defined (__cplusplus) -extern "C" { -#endif - -/* ************************************* -* Streaming functions -***************************************/ - -typedef struct ZSTDv03_Dctx_s ZSTD_DCtx; - -/* - Use above functions alternatively. - ZSTD_nextSrcSizeToDecompress() tells how much bytes to provide as 'srcSize' to ZSTD_decompressContinue(). - ZSTD_decompressContinue() will use previous data blocks to improve compression if they are located prior to current block. - Result 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. -*/ - -/* ************************************* -* Prefix - version detection -***************************************/ -#define ZSTD_magicNumber 0xFD2FB523 /* v0.3 */ - - -#if defined (__cplusplus) -} -#endif -/* ****************************************************************** - FSE : Finite State Entropy coder - Copyright (C) 2013-2015, Yann Collet. - - BSD 2-Clause License (https://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 - - -/**************************************************************** -* Byte symbol type -****************************************************************/ -#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 -# if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */ -# ifdef __GNUC__ -# define FORCE_INLINE static inline __attribute__((always_inline)) -# else -# define FORCE_INLINE static inline -# endif -# else -# define FORCE_INLINE static -# endif /* __STDC_VERSION__ */ -#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) - - -/* Function templates */ - -#define FSE_DECODE_TYPE FSE_decode_t - -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) -{ - void* ptr = dt+1; - FSE_DTableHeader DTableH; - FSE_DECODE_TYPE* const tableDecode = (FSE_DECODE_TYPE*)ptr; - 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; - FSE_decode_t* const cell = (FSE_decode_t*)(ptr) + 1; - - 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; - FSE_decode_t* const dinfo = (FSE_decode_t*)(ptr) + 1; - 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) -{ - FSE_DTableHeader DTableH; - memcpy(&DTableH, dt, sizeof(DTableH)); - - /* select fast mode (static) */ - if (DTableH.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 - Copyright (C) 2013-2015, Yann Collet. - - BSD 2-Clause License (https://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+Huff0 source repository : https://github.com/Cyan4973/FiniteStateEntropy - - Public forum : https://groups.google.com/forum/#!forum/lz4c -****************************************************************** */ - -/**************************************************************** -* Compiler specifics -****************************************************************/ -#if defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) -/* inline is defined */ -#elif defined(_MSC_VER) -# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ -# define inline __inline -#else -# define inline /* disable inline */ -#endif - - -/**************************************************************** -* Includes -****************************************************************/ -#include /* malloc, free, qsort */ -#include /* memcpy, memset */ -#include /* printf (debug) */ - -/**************************************************************** -* Error Management -****************************************************************/ -#define HUF_STATIC_ASSERT(c) { enum { HUF_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */ - - -/****************************************** -* Helper functions -******************************************/ -static unsigned HUF_isError(size_t code) { return ERR_isError(code); } - -#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 - - - -/********************************************************* -* 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; - size_t oSize; - U32 n; - - if (!srcSize) return ERROR(srcSize_wrong); - iSize = ip[0]; - //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; - } - if (weightTotal == 0) return ERROR(corruption_detected); - - /* 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; - const BYTE* ip = (const BYTE*) src; - size_t iSize = ip[0]; - U32 nbSymbols = 0; - U32 n; - U32 nextRankStart; - void* ptr = DTable+1; - HUF_DEltX2* const dt = (HUF_DEltX2*)(ptr); - - 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* ptr = DTable; - const HUF_DEltX2* const dt = ((const HUF_DEltX2*)ptr) +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]; - const BYTE* ip = (const BYTE*) src; - size_t iSize = ip[0]; - void* ptr = DTable; - HUF_DEltX4* const dt = ((HUF_DEltX4*)ptr) + 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--) - { if (!maxW) return ERROR(GENERIC); } /* necessarily finds a solution before maxW==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* ptr = DTable; - const HUF_DEltX4* const dt = ((const HUF_DEltX4*)ptr) +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); -} - - -/**********************************/ -/* Generic decompression selector */ -/**********************************/ - -typedef struct { U32 tableTime; U32 decode256Time; } algo_time_t; -static const algo_time_t algoTime[16 /* Quantization */][3 /* single, double, quad */] = -{ - /* single, double, quad */ - {{0,0}, {1,1}, {2,2}}, /* Q==0 : impossible */ - {{0,0}, {1,1}, {2,2}}, /* Q==1 : impossible */ - {{ 38,130}, {1313, 74}, {2151, 38}}, /* Q == 2 : 12-18% */ - {{ 448,128}, {1353, 74}, {2238, 41}}, /* Q == 3 : 18-25% */ - {{ 556,128}, {1353, 74}, {2238, 47}}, /* Q == 4 : 25-32% */ - {{ 714,128}, {1418, 74}, {2436, 53}}, /* Q == 5 : 32-38% */ - {{ 883,128}, {1437, 74}, {2464, 61}}, /* Q == 6 : 38-44% */ - {{ 897,128}, {1515, 75}, {2622, 68}}, /* Q == 7 : 44-50% */ - {{ 926,128}, {1613, 75}, {2730, 75}}, /* Q == 8 : 50-56% */ - {{ 947,128}, {1729, 77}, {3359, 77}}, /* Q == 9 : 56-62% */ - {{1107,128}, {2083, 81}, {4006, 84}}, /* Q ==10 : 62-69% */ - {{1177,128}, {2379, 87}, {4785, 88}}, /* Q ==11 : 69-75% */ - {{1242,128}, {2415, 93}, {5155, 84}}, /* Q ==12 : 75-81% */ - {{1349,128}, {2644,106}, {5260,106}}, /* Q ==13 : 81-87% */ - {{1455,128}, {2422,124}, {4174,124}}, /* Q ==14 : 87-93% */ - {{ 722,128}, {1891,145}, {1936,146}}, /* Q ==15 : 93-99% */ -}; - -typedef size_t (*decompressionAlgo)(void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); - -static size_t HUF_decompress (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize) -{ - static const decompressionAlgo decompress[3] = { HUF_decompress4X2, HUF_decompress4X4, NULL }; - /* estimate decompression time */ - U32 Q; - const U32 D256 = (U32)(dstSize >> 8); - U32 Dtime[3]; - U32 algoNb = 0; - int n; - - /* validation checks */ - if (dstSize == 0) return ERROR(dstSize_tooSmall); - if (cSrcSize > dstSize) return ERROR(corruption_detected); /* invalid */ - if (cSrcSize == dstSize) { memcpy(dst, cSrc, dstSize); return dstSize; } /* not compressed */ - if (cSrcSize == 1) { memset(dst, *(const BYTE*)cSrc, dstSize); return dstSize; } /* RLE */ - - /* decoder timing evaluation */ - Q = (U32)(cSrcSize * 16 / dstSize); /* Q < 16 since dstSize > cSrcSize */ - for (n=0; n<3; n++) - Dtime[n] = algoTime[Q][n].tableTime + (algoTime[Q][n].decode256Time * D256); - - Dtime[1] += Dtime[1] >> 4; Dtime[2] += Dtime[2] >> 3; /* advantage to algorithms using less memory, for cache eviction */ - - if (Dtime[1] < Dtime[0]) algoNb = 1; - - return decompress[algoNb](dst, dstSize, cSrc, cSrcSize); - - //return HUF_decompress4X2(dst, dstSize, cSrc, cSrcSize); /* multi-streams single-symbol decoding */ - //return HUF_decompress4X4(dst, dstSize, cSrc, cSrcSize); /* multi-streams double-symbols decoding */ - //return HUF_decompress4X6(dst, dstSize, cSrc, cSrcSize); /* multi-streams quad-symbols decoding */ -} -/* - zstd - standard compression library - Copyright (C) 2014-2015, Yann Collet. - - BSD 2-Clause License (https://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 -*/ - -/* *************************************************************** -* 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 -*/ -#define ZSTD_MEMORY_USAGE 17 - -/*! - * HEAPMODE : - * Select how default compression functions will allocate memory for their hash table, - * in memory stack (0, fastest), or in memory heap (1, requires malloc()) - * Note that compression context is fairly large, as a consequence heap memory is recommended. - */ -#ifndef ZSTD_HEAPMODE -# define ZSTD_HEAPMODE 1 -#endif /* ZSTD_HEAPMODE */ - -/*! -* LEGACY_SUPPORT : -* decompressor can decode older formats (starting from Zstd 0.1+) -*/ -#ifndef ZSTD_LEGACY_SUPPORT -# define ZSTD_LEGACY_SUPPORT 1 -#endif - - -/* ******************************************************* -* Includes -*********************************************************/ -#include /* calloc */ -#include /* memcpy, memmove */ -#include /* debug : printf */ - - -/* ******************************************************* -* Compiler specifics -*********************************************************/ -#ifdef __AVX2__ -# include /* AVX2 intrinsics */ -#endif - -#ifdef _MSC_VER /* Visual Studio */ -# include /* For Visual 2005 */ -# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ -# pragma warning(disable : 4324) /* disable: C4324: padded structure */ -#else -# define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) -#endif - - -/* ******************************************************* -* Constants -*********************************************************/ -#define HASH_LOG (ZSTD_MEMORY_USAGE - 2) -#define HASH_TABLESIZE (1 << HASH_LOG) -#define HASH_MASK (HASH_TABLESIZE - 1) - -#define KNUTH 2654435761 - -#define BIT7 128 -#define BIT6 64 -#define BIT5 32 -#define BIT4 16 -#define BIT1 2 -#define BIT0 1 - -#define KB *(1 <<10) -#define MB *(1 <<20) -#define GB *(1U<<30) - -#define BLOCKSIZE (128 KB) /* define, for static allocation */ -#define MIN_SEQUENCES_SIZE (2 /*seqNb*/ + 2 /*dumps*/ + 3 /*seqTables*/ + 1 /*bitStream*/) -#define MIN_CBLOCK_SIZE (3 /*litCSize*/ + MIN_SEQUENCES_SIZE) -#define IS_RAW BIT0 -#define IS_RLE BIT1 - -#define WORKPLACESIZE (BLOCKSIZE*3) -#define MINMATCH 4 -#define MLbits 7 -#define LLbits 6 -#define Offbits 5 -#define MaxML ((1<blockType = (blockType_t)(headerFlags >> 6); - bpPtr->origSize = (bpPtr->blockType == bt_rle) ? cSize : 0; - - if (bpPtr->blockType == bt_end) return 0; - if (bpPtr->blockType == bt_rle) return 1; - return cSize; -} - -static size_t ZSTD_copyUncompressedBlock(void* dst, size_t maxDstSize, const void* src, size_t srcSize) -{ - if (srcSize > maxDstSize) return ERROR(dstSize_tooSmall); - if (srcSize > 0) { - memcpy(dst, src, srcSize); - } - return srcSize; -} - - -/** ZSTD_decompressLiterals - @return : nb of bytes read from src, or an error code*/ -static size_t ZSTD_decompressLiterals(void* dst, size_t* maxDstSizePtr, - const void* src, size_t srcSize) -{ - const BYTE* ip = (const BYTE*)src; - - const size_t litSize = (MEM_readLE32(src) & 0x1FFFFF) >> 2; /* no buffer issue : srcSize >= MIN_CBLOCK_SIZE */ - const size_t litCSize = (MEM_readLE32(ip+2) & 0xFFFFFF) >> 5; /* no buffer issue : srcSize >= MIN_CBLOCK_SIZE */ - - if (litSize > *maxDstSizePtr) return ERROR(corruption_detected); - if (litCSize + 5 > srcSize) return ERROR(corruption_detected); - - if (HUF_isError(HUF_decompress(dst, litSize, ip+5, litCSize))) return ERROR(corruption_detected); - - *maxDstSizePtr = litSize; - return litCSize + 5; -} - - -/** ZSTD_decodeLiteralsBlock - @return : nb of bytes read from src (< srcSize )*/ -static size_t ZSTD_decodeLiteralsBlock(void* ctx, - const void* src, size_t srcSize) -{ - ZSTD_DCtx* dctx = (ZSTD_DCtx*)ctx; - const BYTE* const istart = (const BYTE* const)src; - - /* any compressed block with literals segment must be at least this size */ - if (srcSize < MIN_CBLOCK_SIZE) return ERROR(corruption_detected); - - switch(*istart & 3) - { - default: - case 0: - { - size_t litSize = BLOCKSIZE; - const size_t readSize = ZSTD_decompressLiterals(dctx->litBuffer, &litSize, src, srcSize); - dctx->litPtr = dctx->litBuffer; - dctx->litSize = litSize; - memset(dctx->litBuffer + dctx->litSize, 0, 8); - return readSize; /* works if it's an error too */ - } - case IS_RAW: - { - const size_t litSize = (MEM_readLE32(istart) & 0xFFFFFF) >> 2; /* no buffer issue : srcSize >= MIN_CBLOCK_SIZE */ - if (litSize > srcSize-11) /* risk of reading too far with wildcopy */ - { - if (litSize > BLOCKSIZE) return ERROR(corruption_detected); - if (litSize > srcSize-3) return ERROR(corruption_detected); - memcpy(dctx->litBuffer, istart, litSize); - dctx->litPtr = dctx->litBuffer; - dctx->litSize = litSize; - memset(dctx->litBuffer + dctx->litSize, 0, 8); - return litSize+3; - } - /* direct reference into compressed stream */ - dctx->litPtr = istart+3; - dctx->litSize = litSize; - return litSize+3; - } - case IS_RLE: - { - const size_t litSize = (MEM_readLE32(istart) & 0xFFFFFF) >> 2; /* no buffer issue : srcSize >= MIN_CBLOCK_SIZE */ - if (litSize > BLOCKSIZE) return ERROR(corruption_detected); - memset(dctx->litBuffer, istart[3], litSize + 8); - dctx->litPtr = dctx->litBuffer; - dctx->litSize = litSize; - return 4; - } - } -} - - -static size_t ZSTD_decodeSeqHeaders(int* nbSeq, const BYTE** dumpsPtr, size_t* dumpsLengthPtr, - FSE_DTable* DTableLL, FSE_DTable* DTableML, FSE_DTable* DTableOffb, - const void* src, size_t srcSize) -{ - const BYTE* const istart = (const BYTE* const)src; - const BYTE* ip = istart; - const BYTE* const iend = istart + srcSize; - U32 LLtype, Offtype, MLtype; - U32 LLlog, Offlog, MLlog; - size_t dumpsLength; - - /* check */ - if (srcSize < 5) return ERROR(srcSize_wrong); - - /* SeqHead */ - *nbSeq = MEM_readLE16(ip); ip+=2; - LLtype = *ip >> 6; - Offtype = (*ip >> 4) & 3; - MLtype = (*ip >> 2) & 3; - if (*ip & 2) - { - dumpsLength = ip[2]; - dumpsLength += ip[1] << 8; - ip += 3; - } - else - { - dumpsLength = ip[1]; - dumpsLength += (ip[0] & 1) << 8; - ip += 2; - } - *dumpsPtr = ip; - ip += dumpsLength; - *dumpsLengthPtr = dumpsLength; - - /* check */ - if (ip > iend-3) return ERROR(srcSize_wrong); /* min : all 3 are "raw", hence no header, but at least xxLog bits per type */ - - /* sequences */ - { - S16 norm[MaxML+1]; /* assumption : MaxML >= MaxLL and MaxOff */ - size_t headerSize; - - /* Build DTables */ - switch(LLtype) - { - case bt_rle : - LLlog = 0; - FSE_buildDTable_rle(DTableLL, *ip++); break; - case bt_raw : - LLlog = LLbits; - FSE_buildDTable_raw(DTableLL, LLbits); break; - default : - { U32 max = MaxLL; - headerSize = FSE_readNCount(norm, &max, &LLlog, ip, iend-ip); - if (FSE_isError(headerSize)) return ERROR(GENERIC); - if (LLlog > LLFSELog) return ERROR(corruption_detected); - ip += headerSize; - FSE_buildDTable(DTableLL, norm, max, LLlog); - } } - - switch(Offtype) - { - case bt_rle : - Offlog = 0; - if (ip > iend-2) return ERROR(srcSize_wrong); /* min : "raw", hence no header, but at least xxLog bits */ - FSE_buildDTable_rle(DTableOffb, *ip++ & MaxOff); /* if *ip > MaxOff, data is corrupted */ - break; - case bt_raw : - Offlog = Offbits; - FSE_buildDTable_raw(DTableOffb, Offbits); break; - default : - { U32 max = MaxOff; - headerSize = FSE_readNCount(norm, &max, &Offlog, ip, iend-ip); - if (FSE_isError(headerSize)) return ERROR(GENERIC); - if (Offlog > OffFSELog) return ERROR(corruption_detected); - ip += headerSize; - FSE_buildDTable(DTableOffb, norm, max, Offlog); - } } - - switch(MLtype) - { - case bt_rle : - MLlog = 0; - if (ip > iend-2) return ERROR(srcSize_wrong); /* min : "raw", hence no header, but at least xxLog bits */ - FSE_buildDTable_rle(DTableML, *ip++); break; - case bt_raw : - MLlog = MLbits; - FSE_buildDTable_raw(DTableML, MLbits); break; - default : - { U32 max = MaxML; - headerSize = FSE_readNCount(norm, &max, &MLlog, ip, iend-ip); - if (FSE_isError(headerSize)) return ERROR(GENERIC); - if (MLlog > MLFSELog) return ERROR(corruption_detected); - ip += headerSize; - FSE_buildDTable(DTableML, norm, max, MLlog); - } } } - - return ip-istart; -} - - -typedef struct { - size_t litLength; - size_t offset; - size_t matchLength; -} seq_t; - -typedef struct { - BIT_DStream_t DStream; - FSE_DState_t stateLL; - FSE_DState_t stateOffb; - FSE_DState_t stateML; - size_t prevOffset; - const BYTE* dumps; - const BYTE* dumpsEnd; -} seqState_t; - - -static void ZSTD_decodeSequence(seq_t* seq, seqState_t* seqState) -{ - size_t litLength; - size_t prevOffset; - size_t offset; - size_t matchLength; - const BYTE* dumps = seqState->dumps; - const BYTE* const de = seqState->dumpsEnd; - - /* Literal length */ - litLength = FSE_decodeSymbol(&(seqState->stateLL), &(seqState->DStream)); - prevOffset = litLength ? seq->offset : seqState->prevOffset; - seqState->prevOffset = seq->offset; - if (litLength == MaxLL) - { - const U32 add = dumps= de) dumps = de-1; /* late correction, to avoid read overflow (data is now corrupted anyway) */ - } - - /* Offset */ - { - static const size_t offsetPrefix[MaxOff+1] = { /* note : size_t faster than U32 */ - 1 /*fake*/, 1, 2, 4, 8, 16, 32, 64, 128, 256, - 512, 1024, 2048, 4096, 8192, 16384, 32768, 65536, 131072, 262144, - 524288, 1048576, 2097152, 4194304, 8388608, 16777216, 33554432, /*fake*/ 1, 1, 1, 1, 1 }; - U32 offsetCode, nbBits; - offsetCode = FSE_decodeSymbol(&(seqState->stateOffb), &(seqState->DStream)); /* <= maxOff, by table construction */ - if (MEM_32bits()) BIT_reloadDStream(&(seqState->DStream)); - nbBits = offsetCode - 1; - if (offsetCode==0) nbBits = 0; /* cmove */ - offset = offsetPrefix[offsetCode] + BIT_readBits(&(seqState->DStream), nbBits); - if (MEM_32bits()) BIT_reloadDStream(&(seqState->DStream)); - if (offsetCode==0) offset = prevOffset; /* cmove */ - } - - /* MatchLength */ - matchLength = FSE_decodeSymbol(&(seqState->stateML), &(seqState->DStream)); - if (matchLength == MaxML) - { - const U32 add = dumps= de) dumps = de-1; /* late correction, to avoid read overflow (data is now corrupted anyway) */ - } - matchLength += MINMATCH; - - /* save result */ - seq->litLength = litLength; - seq->offset = offset; - seq->matchLength = matchLength; - seqState->dumps = dumps; -} - - -static size_t ZSTD_execSequence(BYTE* op, - seq_t sequence, - const BYTE** litPtr, const BYTE* const litLimit, - BYTE* const base, BYTE* const oend) -{ - static const int dec32table[] = {0, 1, 2, 1, 4, 4, 4, 4}; /* added */ - static const int dec64table[] = {8, 8, 8, 7, 8, 9,10,11}; /* subtracted */ - const BYTE* const ostart = op; - BYTE* const oLitEnd = op + sequence.litLength; - BYTE* const oMatchEnd = op + sequence.litLength + sequence.matchLength; /* risk : address space overflow (32-bits) */ - BYTE* const oend_8 = oend-8; - const BYTE* const litEnd = *litPtr + sequence.litLength; - - /* checks */ - size_t const seqLength = sequence.litLength + sequence.matchLength; - - if (seqLength > (size_t)(oend - op)) return ERROR(dstSize_tooSmall); - if (sequence.litLength > (size_t)(litLimit - *litPtr)) return ERROR(corruption_detected); - /* Now we know there are no overflow in literal nor match lengths, can use pointer checks */ - if (oLitEnd > oend_8) return ERROR(dstSize_tooSmall); - if (sequence.offset > (U32)(oLitEnd - base)) return ERROR(corruption_detected); - - if (oMatchEnd > oend) return ERROR(dstSize_tooSmall); /* overwrite beyond dst buffer */ - if (litEnd > litLimit) return ERROR(corruption_detected); /* overRead beyond lit buffer */ - - /* copy Literals */ - ZSTD_wildcopy(op, *litPtr, (ptrdiff_t)sequence.litLength); /* note : oLitEnd <= oend-8 : no risk of overwrite beyond oend */ - op = oLitEnd; - *litPtr = litEnd; /* update for next sequence */ - - /* copy Match */ - { const BYTE* match = op - sequence.offset; - - /* check */ - if (sequence.offset > (size_t)op) return ERROR(corruption_detected); /* address space overflow test (this test seems kept by clang optimizer) */ - //if (match > op) return ERROR(corruption_detected); /* address space overflow test (is clang optimizer removing this test ?) */ - if (match < base) return ERROR(corruption_detected); - - /* close range match, overlap */ - if (sequence.offset < 8) - { - const int dec64 = dec64table[sequence.offset]; - op[0] = match[0]; - op[1] = match[1]; - op[2] = match[2]; - op[3] = match[3]; - match += dec32table[sequence.offset]; - ZSTD_copy4(op+4, match); - match -= dec64; - } - else - { - ZSTD_copy8(op, match); - } - op += 8; match += 8; - - if (oMatchEnd > oend-(16-MINMATCH)) - { - if (op < oend_8) - { - ZSTD_wildcopy(op, match, oend_8 - op); - match += oend_8 - op; - op = oend_8; - } - while (op < oMatchEnd) *op++ = *match++; - } - else - { - ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8); /* works even if matchLength < 8 */ - } - } - - return oMatchEnd - ostart; -} - -static size_t ZSTD_decompressSequences( - void* ctx, - void* dst, size_t maxDstSize, - const void* seqStart, size_t seqSize) -{ - ZSTD_DCtx* dctx = (ZSTD_DCtx*)ctx; - const BYTE* ip = (const BYTE*)seqStart; - const BYTE* const iend = ip + seqSize; - BYTE* const ostart = (BYTE* const)dst; - BYTE* op = ostart; - BYTE* const oend = ostart + maxDstSize; - size_t errorCode, dumpsLength; - const BYTE* litPtr = dctx->litPtr; - const BYTE* const litEnd = litPtr + dctx->litSize; - int nbSeq; - const BYTE* dumps; - U32* DTableLL = dctx->LLTable; - U32* DTableML = dctx->MLTable; - U32* DTableOffb = dctx->OffTable; - BYTE* const base = (BYTE*) (dctx->base); - - /* Build Decoding Tables */ - errorCode = ZSTD_decodeSeqHeaders(&nbSeq, &dumps, &dumpsLength, - DTableLL, DTableML, DTableOffb, - ip, iend-ip); - if (ZSTD_isError(errorCode)) return errorCode; - ip += errorCode; - - /* Regen sequences */ - { - seq_t sequence; - seqState_t seqState; - - memset(&sequence, 0, sizeof(sequence)); - seqState.dumps = dumps; - seqState.dumpsEnd = dumps + dumpsLength; - seqState.prevOffset = sequence.offset = 4; - errorCode = BIT_initDStream(&(seqState.DStream), ip, iend-ip); - if (ERR_isError(errorCode)) return ERROR(corruption_detected); - FSE_initDState(&(seqState.stateLL), &(seqState.DStream), DTableLL); - FSE_initDState(&(seqState.stateOffb), &(seqState.DStream), DTableOffb); - FSE_initDState(&(seqState.stateML), &(seqState.DStream), DTableML); - - for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (nbSeq>0) ; ) - { - size_t oneSeqSize; - nbSeq--; - ZSTD_decodeSequence(&sequence, &seqState); - oneSeqSize = ZSTD_execSequence(op, sequence, &litPtr, litEnd, base, oend); - if (ZSTD_isError(oneSeqSize)) return oneSeqSize; - op += oneSeqSize; - } - - /* check if reached exact end */ - if ( !BIT_endOfDStream(&(seqState.DStream)) ) return ERROR(corruption_detected); /* requested too much : data is corrupted */ - if (nbSeq<0) return ERROR(corruption_detected); /* requested too many sequences : data is corrupted */ - - /* last literal segment */ - { - size_t lastLLSize = litEnd - litPtr; - if (litPtr > litEnd) return ERROR(corruption_detected); - if (op+lastLLSize > oend) return ERROR(dstSize_tooSmall); - if (lastLLSize > 0) { - if (op != litPtr) memmove(op, litPtr, lastLLSize); - op += lastLLSize; - } - } - } - - return op-ostart; -} - - -static size_t ZSTD_decompressBlock( - void* ctx, - void* dst, size_t maxDstSize, - const void* src, size_t srcSize) -{ - /* blockType == blockCompressed */ - const BYTE* ip = (const BYTE*)src; - - /* Decode literals sub-block */ - size_t litCSize = ZSTD_decodeLiteralsBlock(ctx, src, srcSize); - if (ZSTD_isError(litCSize)) return litCSize; - ip += litCSize; - srcSize -= litCSize; - - return ZSTD_decompressSequences(ctx, dst, maxDstSize, ip, srcSize); -} - - -static size_t ZSTD_decompressDCtx(void* ctx, void* dst, size_t maxDstSize, const void* src, size_t srcSize) -{ - const BYTE* ip = (const BYTE*)src; - const BYTE* iend = ip + srcSize; - BYTE* const ostart = (BYTE* const)dst; - BYTE* op = ostart; - BYTE* const oend = ostart + maxDstSize; - size_t remainingSize = srcSize; - U32 magicNumber; - blockProperties_t blockProperties; - - /* Frame Header */ - if (srcSize < ZSTD_frameHeaderSize+ZSTD_blockHeaderSize) return ERROR(srcSize_wrong); - magicNumber = MEM_readLE32(src); - if (magicNumber != ZSTD_magicNumber) return ERROR(prefix_unknown); - ip += ZSTD_frameHeaderSize; remainingSize -= ZSTD_frameHeaderSize; - - /* Loop on each block */ - while (1) - { - size_t decodedSize=0; - size_t cBlockSize = ZSTD_getcBlockSize(ip, iend-ip, &blockProperties); - if (ZSTD_isError(cBlockSize)) return cBlockSize; - - ip += ZSTD_blockHeaderSize; - remainingSize -= ZSTD_blockHeaderSize; - if (cBlockSize > remainingSize) return ERROR(srcSize_wrong); - - switch(blockProperties.blockType) - { - case bt_compressed: - decodedSize = ZSTD_decompressBlock(ctx, op, oend-op, ip, cBlockSize); - break; - case bt_raw : - decodedSize = ZSTD_copyUncompressedBlock(op, oend-op, ip, cBlockSize); - break; - case bt_rle : - return ERROR(GENERIC); /* not yet supported */ - break; - case bt_end : - /* end of frame */ - if (remainingSize) return ERROR(srcSize_wrong); - break; - default: - return ERROR(GENERIC); /* impossible */ - } - if (cBlockSize == 0) break; /* bt_end */ - - if (ZSTD_isError(decodedSize)) return decodedSize; - op += decodedSize; - ip += cBlockSize; - remainingSize -= cBlockSize; - } - - return op-ostart; -} - -static size_t ZSTD_decompress(void* dst, size_t maxDstSize, const void* src, size_t srcSize) -{ - ZSTD_DCtx ctx; - ctx.base = dst; - return ZSTD_decompressDCtx(&ctx, dst, maxDstSize, src, srcSize); -} - -/* ZSTD_errorFrameSizeInfoLegacy() : - assumes `cSize` and `dBound` are _not_ NULL */ -MEM_STATIC void ZSTD_errorFrameSizeInfoLegacy(size_t* cSize, unsigned long long* dBound, size_t ret) -{ - *cSize = ret; - *dBound = ZSTD_CONTENTSIZE_ERROR; -} - -void ZSTDv03_findFrameSizeInfoLegacy(const void *src, size_t srcSize, size_t* cSize, unsigned long long* dBound) -{ - const BYTE* ip = (const BYTE*)src; - size_t remainingSize = srcSize; - size_t nbBlocks = 0; - U32 magicNumber; - blockProperties_t blockProperties; - - /* Frame Header */ - if (srcSize < ZSTD_frameHeaderSize+ZSTD_blockHeaderSize) { - ZSTD_errorFrameSizeInfoLegacy(cSize, dBound, ERROR(srcSize_wrong)); - return; - } - magicNumber = MEM_readLE32(src); - if (magicNumber != ZSTD_magicNumber) { - ZSTD_errorFrameSizeInfoLegacy(cSize, dBound, ERROR(prefix_unknown)); - return; - } - ip += ZSTD_frameHeaderSize; remainingSize -= ZSTD_frameHeaderSize; - - /* Loop on each block */ - while (1) - { - size_t cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties); - if (ZSTD_isError(cBlockSize)) { - ZSTD_errorFrameSizeInfoLegacy(cSize, dBound, cBlockSize); - return; - } - - ip += ZSTD_blockHeaderSize; - remainingSize -= ZSTD_blockHeaderSize; - if (cBlockSize > remainingSize) { - ZSTD_errorFrameSizeInfoLegacy(cSize, dBound, ERROR(srcSize_wrong)); - return; - } - - if (cBlockSize == 0) break; /* bt_end */ - - ip += cBlockSize; - remainingSize -= cBlockSize; - nbBlocks++; - } - - *cSize = ip - (const BYTE*)src; - *dBound = nbBlocks * BLOCKSIZE; -} - - -/******************************* -* Streaming Decompression API -*******************************/ - -static size_t ZSTD_resetDCtx(ZSTD_DCtx* dctx) -{ - dctx->expected = ZSTD_frameHeaderSize; - dctx->phase = 0; - dctx->previousDstEnd = NULL; - dctx->base = NULL; - return 0; -} - -static ZSTD_DCtx* ZSTD_createDCtx(void) -{ - ZSTD_DCtx* dctx = (ZSTD_DCtx*)malloc(sizeof(ZSTD_DCtx)); - if (dctx==NULL) return NULL; - ZSTD_resetDCtx(dctx); - return dctx; -} - -static size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx) -{ - free(dctx); - return 0; -} - -static size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx) -{ - return dctx->expected; -} - -static size_t ZSTD_decompressContinue(ZSTD_DCtx* ctx, void* dst, size_t maxDstSize, const void* src, size_t srcSize) -{ - /* Sanity check */ - if (srcSize != ctx->expected) return ERROR(srcSize_wrong); - if (dst != ctx->previousDstEnd) /* not contiguous */ - ctx->base = dst; - - /* Decompress : frame header */ - if (ctx->phase == 0) - { - /* Check frame magic header */ - U32 magicNumber = MEM_readLE32(src); - if (magicNumber != ZSTD_magicNumber) return ERROR(prefix_unknown); - ctx->phase = 1; - ctx->expected = ZSTD_blockHeaderSize; - return 0; - } - - /* Decompress : block header */ - if (ctx->phase == 1) - { - blockProperties_t bp; - size_t blockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp); - if (ZSTD_isError(blockSize)) return blockSize; - if (bp.blockType == bt_end) - { - ctx->expected = 0; - ctx->phase = 0; - } - else - { - ctx->expected = blockSize; - ctx->bType = bp.blockType; - ctx->phase = 2; - } - - return 0; - } - - /* Decompress : block content */ - { - size_t rSize; - switch(ctx->bType) - { - case bt_compressed: - rSize = ZSTD_decompressBlock(ctx, dst, maxDstSize, src, srcSize); - break; - case bt_raw : - rSize = ZSTD_copyUncompressedBlock(dst, maxDstSize, src, srcSize); - break; - case bt_rle : - return ERROR(GENERIC); /* not yet handled */ - break; - case bt_end : /* should never happen (filtered at phase 1) */ - rSize = 0; - break; - default: - return ERROR(GENERIC); - } - ctx->phase = 1; - ctx->expected = ZSTD_blockHeaderSize; - if (ZSTD_isError(rSize)) return rSize; - ctx->previousDstEnd = (void*)( ((char*)dst) + rSize); - return rSize; - } - -} - - -/* wrapper layer */ - -unsigned ZSTDv03_isError(size_t code) -{ - return ZSTD_isError(code); -} - -size_t ZSTDv03_decompress( void* dst, size_t maxOriginalSize, - const void* src, size_t compressedSize) -{ - return ZSTD_decompress(dst, maxOriginalSize, src, compressedSize); -} - -ZSTDv03_Dctx* ZSTDv03_createDCtx(void) -{ - return (ZSTDv03_Dctx*)ZSTD_createDCtx(); -} - -size_t ZSTDv03_freeDCtx(ZSTDv03_Dctx* dctx) -{ - return ZSTD_freeDCtx((ZSTD_DCtx*)dctx); -} - -size_t ZSTDv03_resetDCtx(ZSTDv03_Dctx* dctx) -{ - return ZSTD_resetDCtx((ZSTD_DCtx*)dctx); -} - -size_t ZSTDv03_nextSrcSizeToDecompress(ZSTDv03_Dctx* dctx) -{ - return ZSTD_nextSrcSizeToDecompress((ZSTD_DCtx*)dctx); -} - -size_t ZSTDv03_decompressContinue(ZSTDv03_Dctx* dctx, void* dst, size_t maxDstSize, const void* src, size_t srcSize) -{ - return ZSTD_decompressContinue((ZSTD_DCtx*)dctx, dst, maxDstSize, src, srcSize); -} +/* Implementation moved to Rust (rust/src/legacy/zstd_v03.rs). + * The frozen v0.3 decoder, including its embedded FSE/Huff0 snapshot and the + * ZSTDv03_Dctx streaming state, now lives entirely in Rust; C code only ever + * holds an opaque ZSTDv03_Dctx pointer. */ diff --git a/rust/src/legacy/mod.rs b/rust/src/legacy/mod.rs index 11d5a136a..3006ae723 100644 --- a/rust/src/legacy/mod.rs +++ b/rust/src/legacy/mod.rs @@ -41,3 +41,6 @@ pub mod zstd_v01; #[cfg(feature = "legacy-v02")] pub mod zstd_v02; + +#[cfg(feature = "legacy-v03")] +pub mod zstd_v03; diff --git a/rust/src/legacy/zstd_v03.rs b/rust/src/legacy/zstd_v03.rs new file mode 100644 index 000000000..a9564ee2c --- /dev/null +++ b/rust/src/legacy/zstd_v03.rs @@ -0,0 +1,1790 @@ +#![allow(non_snake_case)] + +//! Frozen decoder for the zstd v0.3 format. +//! +//! `lib/legacy/zstd_v03.c` is an old, self-contained decoder. This module +//! keeps that boundary: it owns its FSE, bit-stream, and Huffman state rather +//! than depending on the current entropy implementations. The public entry +//! points below retain the C ABI and the context is deliberately malloc/free +//! allocated so C callers can continue to own an opaque `ZSTDv03_Dctx`. + +use crate::errors::{ERR_isError, ZstdErrorCode, ERROR}; +use std::os::raw::{c_uint, c_void}; +use std::ptr; + +const ZSTD_MAGIC_NUMBER: u32 = 0xFD2F_B523; +const ZSTD_CONTENTSIZE_ERROR: u64 = u64::MAX - 1; + +const BLOCKSIZE: usize = 128 * 1024; +const MIN_SEQUENCES_SIZE: usize = 2 + 2 + 3 + 1; +const MIN_CBLOCK_SIZE: usize = 3 + MIN_SEQUENCES_SIZE; +const MINMATCH: usize = 4; + +const ML_BITS: u32 = 7; +const LL_BITS: u32 = 6; +const OFF_BITS: u32 = 5; +const MAX_ML: u32 = (1 << ML_BITS) - 1; +const MAX_LL: u32 = (1 << LL_BITS) - 1; +const MAX_OFF: u32 = 31; +const ML_FSE_LOG: u32 = 10; +const LL_FSE_LOG: u32 = 10; +const OFF_FSE_LOG: u32 = 9; + +const FSE_MAX_MEMORY_USAGE: u32 = 14; +const FSE_MAX_SYMBOL_VALUE: u32 = 255; +const FSE_MAX_TABLELOG: u32 = FSE_MAX_MEMORY_USAGE - 2; +const FSE_MIN_TABLELOG: u32 = 5; +const FSE_TABLELOG_ABSOLUTE_MAX: u32 = 15; + +const HUF_MAX_SYMBOL_VALUE: usize = 255; +const HUF_MAX_TABLELOG: usize = 12; +const HUF_ABSOLUTE_MAX_TABLELOG: usize = 16; + +const BT_COMPRESSED: u32 = 0; +const BT_RAW: u32 = 1; +const BT_RLE: u32 = 2; +const BT_END: u32 = 3; + +const DSTREAM_UNFINISHED: u32 = 0; +const DSTREAM_END_OF_BUFFER: u32 = 1; +const DSTREAM_COMPLETED: u32 = 2; +const DSTREAM_TOO_FAR: u32 = 3; + +const USIZE_BITS: u32 = usize::BITS; + +#[repr(C)] +#[derive(Clone, Copy)] +struct FseDecode { + new_state: u16, + symbol: u8, + nb_bits: u8, +} + +#[repr(C)] +struct FseDTableHeader { + table_log: u16, + fast_mode: u16, +} + +#[derive(Clone, Copy)] +struct DStream { + bit_container: usize, + bits_consumed: u32, + ptr: *const u8, + start: *const u8, +} + +#[derive(Clone, Copy)] +struct FseDState { + state: usize, + table: *const FseDecode, +} + +#[inline] +fn highbit32(value: u32) -> u32 { + value.leading_zeros() ^ 31 +} + +#[inline] +unsafe fn read_le16(ptr: *const u8) -> u16 { + u16::from_le_bytes(std::ptr::read_unaligned(ptr as *const [u8; 2])) +} + +#[inline] +unsafe fn read_le24(ptr: *const u8) -> u32 { + (read_le16(ptr) as u32) | ((*ptr.add(2) as u32) << 16) +} + +#[inline] +unsafe fn read_le32(ptr: *const u8) -> u32 { + u32::from_le_bytes(std::ptr::read_unaligned(ptr as *const [u8; 4])) +} + +#[inline] +unsafe fn read_le_size(ptr: *const u8) -> usize { + if std::mem::size_of::() == 4 { + read_le32(ptr) as usize + } else { + u64::from_le_bytes(std::ptr::read_unaligned(ptr as *const [u8; 8])) as usize + } +} + +/* ****************************************** + * Backward bit stream (the v0.3 snapshot) + ********************************************/ + +unsafe fn init_dstream(stream: &mut DStream, src: *const u8, src_size: usize) -> usize { + let word = std::mem::size_of::(); + if src_size == 0 { + *stream = DStream { + bit_container: 0, + bits_consumed: 0, + ptr: ptr::null(), + start: ptr::null(), + }; + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + + stream.start = src; + if src_size >= word { + stream.ptr = src.add(src_size - word); + stream.bit_container = read_le_size(stream.ptr); + let end_byte = *src.add(src_size - 1) as u32; + if end_byte == 0 { + return ERROR(ZstdErrorCode::Generic); + } + stream.bits_consumed = 8 - highbit32(end_byte); + } else { + stream.ptr = src; + stream.bit_container = *src as usize; + if src_size >= 7 { + stream.bit_container += (*src.add(6) as usize) << (USIZE_BITS as usize - 16); + } + if src_size >= 6 { + stream.bit_container += (*src.add(5) as usize) << (USIZE_BITS as usize - 24); + } + if src_size >= 5 { + stream.bit_container += (*src.add(4) as usize) << (USIZE_BITS as usize - 32); + } + if src_size >= 4 { + stream.bit_container += (*src.add(3) as usize) << 24; + } + if src_size >= 3 { + stream.bit_container += (*src.add(2) as usize) << 16; + } + if src_size >= 2 { + stream.bit_container += (*src.add(1) as usize) << 8; + } + let end_byte = *src.add(src_size - 1) as u32; + if end_byte == 0 { + return ERROR(ZstdErrorCode::Generic); + } + stream.bits_consumed = 8 - highbit32(end_byte); + stream.bits_consumed += ((word - src_size) * 8) as u32; + } + src_size +} + +#[inline] +unsafe fn look_bits(stream: &DStream, nb_bits: u32) -> usize { + let mask = USIZE_BITS - 1; + ((stream.bit_container << (stream.bits_consumed & mask)) >> 1) + >> (mask.wrapping_sub(nb_bits) & mask) +} + +#[inline] +unsafe fn look_bits_fast(stream: &DStream, nb_bits: u32) -> usize { + let mask = USIZE_BITS - 1; + (stream.bit_container << (stream.bits_consumed & mask)) + >> ((mask + 1).wrapping_sub(nb_bits) & mask) +} + +#[inline] +fn skip_bits(stream: &mut DStream, nb_bits: u32) { + stream.bits_consumed = stream.bits_consumed.wrapping_add(nb_bits); +} + +#[inline] +unsafe fn read_bits(stream: &mut DStream, nb_bits: u32) -> usize { + let value = look_bits(stream, nb_bits); + skip_bits(stream, nb_bits); + value +} + +#[inline] +unsafe fn read_bits_fast(stream: &mut DStream, nb_bits: u32) -> usize { + let value = look_bits_fast(stream, nb_bits); + skip_bits(stream, nb_bits); + value +} + +unsafe fn reload_dstream(stream: &mut DStream) -> u32 { + let word = std::mem::size_of::(); + if stream.bits_consumed > (word * 8) as u32 { + return DSTREAM_TOO_FAR; + } + if (stream.ptr as usize) >= (stream.start as usize).wrapping_add(word) { + stream.ptr = stream.ptr.sub((stream.bits_consumed >> 3) as usize); + stream.bits_consumed &= 7; + stream.bit_container = read_le_size(stream.ptr); + return DSTREAM_UNFINISHED; + } + if stream.ptr == stream.start { + if stream.bits_consumed < (word * 8) as u32 { + return DSTREAM_END_OF_BUFFER; + } + return DSTREAM_COMPLETED; + } + + let mut nb_bytes = stream.bits_consumed >> 3; + let mut result = DSTREAM_UNFINISHED; + if (stream.ptr as usize).wrapping_sub(nb_bytes as usize) < stream.start as usize { + nb_bytes = (stream.ptr as usize - stream.start as usize) as u32; + result = DSTREAM_END_OF_BUFFER; + } + stream.ptr = stream.ptr.sub(nb_bytes as usize); + stream.bits_consumed -= nb_bytes * 8; + stream.bit_container = read_le_size(stream.ptr); + result +} + +#[inline] +fn end_of_dstream(stream: &DStream) -> bool { + stream.ptr == stream.start && stream.bits_consumed == USIZE_BITS +} + +/* ****************************************** + * FSE decoding + ********************************************/ + +#[inline] +fn fse_table_step(table_size: u32) -> u32 { + (table_size >> 1) + (table_size >> 3) + 3 +} + +#[allow(clippy::needless_range_loop)] +unsafe fn fse_build_dtable( + dt: &mut [u32], + normalized_counter: &[i16; 256], + max_symbol_value: u32, + table_log: u32, +) -> usize { + if max_symbol_value > FSE_MAX_SYMBOL_VALUE { + return ERROR(ZstdErrorCode::MaxSymbolValueTooLarge); + } + if table_log > FSE_MAX_TABLELOG { + return ERROR(ZstdErrorCode::TableLogTooLarge); + } + + let table_size = 1u32 << table_log; + let table_mask = table_size - 1; + let step = fse_table_step(table_size); + let mut symbol_next = [0u16; 256]; + let table_header = dt.as_mut_ptr() as *mut FseDTableHeader; + let table_decode = dt.as_mut_ptr().add(1) as *mut FseDecode; + let mut position = 0u32; + let mut high_threshold = table_size - 1; + let large_limit = (1i32 << (table_log - 1)) as i16; + let mut no_large = 1u16; + + (*table_header).table_log = table_log as u16; + for symbol in 0..=max_symbol_value as usize { + let count = normalized_counter[symbol]; + if count == -1 { + (*table_decode.add(high_threshold as usize)).symbol = symbol as u8; + high_threshold = high_threshold.wrapping_sub(1); + symbol_next[symbol] = 1; + } else { + if count >= large_limit { + no_large = 0; + } + symbol_next[symbol] = count as u16; + } + } + + for symbol in 0..=max_symbol_value as usize { + let count = normalized_counter[symbol]; + for _ in 0..count.max(0) { + (*table_decode.add(position as usize)).symbol = symbol as u8; + position = (position + step) & table_mask; + while position > high_threshold { + position = (position + step) & table_mask; + } + } + } + if position != 0 { + return ERROR(ZstdErrorCode::Generic); + } + + for index in 0..table_size as usize { + let symbol = (*table_decode.add(index)).symbol as usize; + let next_state = symbol_next[symbol]; + symbol_next[symbol] = symbol_next[symbol].wrapping_add(1); + let nb_bits = (table_log - highbit32(next_state as u32)) as u8; + (*table_decode.add(index)).nb_bits = nb_bits; + (*table_decode.add(index)).new_state = + (((next_state as u32) << nb_bits).wrapping_sub(table_size)) as u16; + } + (*table_header).fast_mode = no_large; + 0 +} + +unsafe fn fse_read_ncount( + normalized_counter: &mut [i16; 256], + max_sv: &mut u32, + table_log: &mut u32, + header: *const u8, + header_size: usize, +) -> usize { + if header_size < 4 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + let start = header as usize; + let end = start.wrapping_add(header_size); + let mut ip = header; + let mut char_num = 0u32; + let mut previous_zero = false; + let mut bit_stream = read_le32(ip); + let mut nb_bits = ((bit_stream & 0xF) + FSE_MIN_TABLELOG) as i32; + if nb_bits > FSE_TABLELOG_ABSOLUTE_MAX as i32 { + return ERROR(ZstdErrorCode::TableLogTooLarge); + } + bit_stream >>= 4; + let mut bit_count = 4i32; + *table_log = nb_bits as u32; + let mut remaining = (1i32 << nb_bits) + 1; + let mut threshold = 1i32 << nb_bits; + nb_bits += 1; + + while remaining > 1 && char_num <= *max_sv { + if previous_zero { + let mut n0 = char_num; + while bit_stream & 0xFFFF == 0xFFFF { + n0 += 24; + if (ip as usize) < end.wrapping_sub(5) { + ip = ip.add(2); + bit_stream = read_le32(ip) >> bit_count; + } else { + bit_stream >>= 16; + bit_count += 16; + } + } + while bit_stream & 3 == 3 { + n0 += 3; + bit_stream >>= 2; + bit_count += 2; + } + n0 += bit_stream & 3; + bit_count += 2; + if n0 > *max_sv { + return ERROR(ZstdErrorCode::MaxSymbolValueTooSmall); + } + while char_num < n0 { + normalized_counter[char_num as usize] = 0; + char_num += 1; + } + if (ip as usize) <= end.wrapping_sub(7) + || (ip as usize).wrapping_add((bit_count >> 3) as usize) <= end.wrapping_sub(4) + { + ip = ip.add((bit_count >> 3) as usize); + bit_count &= 7; + bit_stream = read_le32(ip) >> bit_count; + } else { + bit_stream >>= 2; + } + } + + let max = ((2 * threshold - 1) - remaining) as i16; + let mut count: i16; + if (bit_stream & (threshold - 1) as u32) < max as i32 as u32 { + count = (bit_stream & (threshold - 1) as u32) as u16 as i16; + bit_count += nb_bits - 1; + } else { + count = (bit_stream & (2 * threshold - 1) as u32) as u16 as i16; + if count as i32 >= threshold { + count = (count as i32 - max as i32) as i16; + } + bit_count += nb_bits; + } + count = count.wrapping_sub(1); + remaining -= (count as i32).abs(); + normalized_counter[char_num as usize] = count; + char_num += 1; + previous_zero = count == 0; + while remaining < threshold { + nb_bits -= 1; + threshold >>= 1; + } + + if (ip as usize) <= end.wrapping_sub(7) + || (ip as usize).wrapping_add((bit_count >> 3) as usize) <= end.wrapping_sub(4) + { + ip = ip.add((bit_count >> 3) as usize); + bit_count &= 7; + } else { + bit_count -= (8 * (end.wrapping_sub(4) as isize - ip as usize as isize)) as i32; + ip = (end - 4) as *const u8; + } + bit_stream = read_le32(ip) >> (bit_count & 31); + } + + if remaining != 1 { + return ERROR(ZstdErrorCode::Generic); + } + *max_sv = char_num - 1; + ip = ip.add(((bit_count + 7) >> 3) as usize); + if (ip as usize).wrapping_sub(start) > header_size { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + (ip as usize).wrapping_sub(start) +} + +unsafe fn fse_build_dtable_rle(dt: &mut [u32], symbol: u8) -> usize { + let header = dt.as_mut_ptr() as *mut FseDTableHeader; + let cell = dt.as_mut_ptr().add(1) as *mut FseDecode; + (*header).table_log = 0; + (*header).fast_mode = 0; + (*cell).new_state = 0; + (*cell).symbol = symbol; + (*cell).nb_bits = 0; + 0 +} + +unsafe fn fse_build_dtable_raw(dt: &mut [u32], nb_bits: u32) -> usize { + if nb_bits < 1 { + return ERROR(ZstdErrorCode::Generic); + } + let header = dt.as_mut_ptr() as *mut FseDTableHeader; + let cells = dt.as_mut_ptr().add(1) as *mut FseDecode; + let table_size = 1u32 << nb_bits; + (*header).table_log = nb_bits as u16; + (*header).fast_mode = 1; + for symbol in 0..table_size { + let cell = cells.add(symbol as usize); + (*cell).new_state = 0; + (*cell).symbol = symbol as u8; + (*cell).nb_bits = nb_bits as u8; + } + 0 +} + +unsafe fn fse_init_dstate(state: &mut FseDState, stream: &mut DStream, dt: *const u32) { + let header = dt as *const FseDTableHeader; + state.state = read_bits(stream, (*header).table_log as u32); + reload_dstream(stream); + state.table = dt.add(1) as *const FseDecode; +} + +#[inline] +unsafe fn fse_decode_symbol(state: &mut FseDState, stream: &mut DStream, fast: bool) -> u8 { + let info = *state.table.add(state.state); + let low_bits = if fast { + read_bits_fast(stream, info.nb_bits as u32) + } else { + read_bits(stream, info.nb_bits as u32) + }; + state.state = (info.new_state as usize).wrapping_add(low_bits); + info.symbol +} + +unsafe fn fse_decompress_using_dtable( + dst: *mut u8, + max_dst_size: usize, + c_src: *const u8, + c_src_size: usize, + dt: &[u32], +) -> usize { + let header = &*(dt.as_ptr() as *const FseDTableHeader); + let fast = header.fast_mode != 0; + let start = dst; + let end_addr = (dst as usize).wrapping_add(max_dst_size); + let limit_addr = end_addr.wrapping_sub(3); + let mut op = dst; + let mut stream = DStream { + bit_container: 0, + bits_consumed: 0, + ptr: ptr::null(), + start: ptr::null(), + }; + let mut state1 = FseDState { + state: 0, + table: ptr::null(), + }; + let mut state2 = state1; + let error = init_dstream(&mut stream, c_src, c_src_size); + if ERR_isError(error) { + return error; + } + fse_init_dstate(&mut state1, &mut stream, dt.as_ptr()); + fse_init_dstate(&mut state2, &mut stream, dt.as_ptr()); + + const RELOAD_2: bool = FSE_MAX_TABLELOG * 2 + 7 > USIZE_BITS; + const RELOAD_4: bool = FSE_MAX_TABLELOG * 4 + 7 > USIZE_BITS; + while reload_dstream(&mut stream) == DSTREAM_UNFINISHED && (op as usize) < limit_addr { + *op = fse_decode_symbol(&mut state1, &mut stream, fast); + if RELOAD_2 { + reload_dstream(&mut stream); + } + *op.add(1) = fse_decode_symbol(&mut state2, &mut stream, fast); + if RELOAD_4 && reload_dstream(&mut stream) > DSTREAM_UNFINISHED { + op = op.add(2); + break; + } + *op.add(2) = fse_decode_symbol(&mut state1, &mut stream, fast); + if RELOAD_2 { + reload_dstream(&mut stream); + } + *op.add(3) = fse_decode_symbol(&mut state2, &mut stream, fast); + op = op.add(4); + } + + loop { + if reload_dstream(&mut stream) > DSTREAM_COMPLETED + || op as usize == end_addr + || (end_of_dstream(&stream) && (fast || state1.state == 0)) + { + break; + } + *op = fse_decode_symbol(&mut state1, &mut stream, fast); + op = op.add(1); + if reload_dstream(&mut stream) > DSTREAM_COMPLETED + || op as usize == end_addr + || (end_of_dstream(&stream) && (fast || state2.state == 0)) + { + break; + } + *op = fse_decode_symbol(&mut state2, &mut stream, fast); + op = op.add(1); + } + + if end_of_dstream(&stream) && state1.state == 0 && state2.state == 0 { + return (op as usize) - (start as usize); + } + if op as usize == end_addr { + return ERROR(ZstdErrorCode::DstSizeTooSmall); + } + ERROR(ZstdErrorCode::CorruptionDetected) +} + +unsafe fn fse_decompress( + dst: *mut u8, + max_dst_size: usize, + c_src: *const u8, + c_src_size: usize, +) -> usize { + if c_src_size < 2 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + let mut counters = [0i16; 256]; + let mut max_symbol = FSE_MAX_SYMBOL_VALUE; + let mut table_log = 0; + let header_size = fse_read_ncount( + &mut counters, + &mut max_symbol, + &mut table_log, + c_src, + c_src_size, + ); + if ERR_isError(header_size) { + return header_size; + } + if header_size >= c_src_size { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + let mut table = vec![0u32; 1 + (1usize << FSE_MAX_TABLELOG)]; + let error = fse_build_dtable(&mut table, &counters, max_symbol, table_log); + if ERR_isError(error) { + return error; + } + fse_decompress_using_dtable( + dst, + max_dst_size, + c_src.add(header_size), + c_src_size - header_size, + &table, + ) +} + +/* ****************************************** + * Huffman decoding + ********************************************/ + +#[repr(C)] +#[derive(Clone, Copy)] +struct HufDEltX2 { + byte: u8, + nb_bits: u8, +} + +#[allow(clippy::manual_div_ceil)] +unsafe fn huf_read_stats( + huff_weight: &mut [u8; HUF_MAX_SYMBOL_VALUE + 1], + rank_stats: &mut [u32; HUF_ABSOLUTE_MAX_TABLELOG + 1], + nb_symbols: &mut u32, + table_log: &mut u32, + src: *const u8, + src_size: usize, +) -> usize { + if src_size == 0 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + let mut i_size = *src as usize; + let o_size: usize; + if i_size >= 128 { + if i_size >= 242 { + const RLE_LENGTHS: [usize; 14] = [1, 2, 3, 4, 7, 8, 15, 16, 31, 32, 63, 64, 127, 128]; + let index = i_size - 242; + if index >= RLE_LENGTHS.len() { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + o_size = RLE_LENGTHS[index]; + huff_weight.fill(1); + i_size = 0; + } else { + o_size = i_size - 127; + i_size = (o_size + 1) / 2; + if i_size + 1 > src_size { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + if o_size >= huff_weight.len() { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + let weights = src.add(1); + let mut n = 0; + while n < o_size { + huff_weight[n] = *weights.add(n / 2) >> 4; + if n + 1 < huff_weight.len() { + huff_weight[n + 1] = *weights.add(n / 2) & 15; + } + n += 2; + } + } + } else { + if i_size + 1 > src_size { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + let decoded = fse_decompress( + huff_weight.as_mut_ptr(), + huff_weight.len() - 1, + src.add(1), + i_size, + ); + if ERR_isError(decoded) { + return decoded; + } + o_size = decoded; + } + + rank_stats.fill(0); + let mut weight_total = 0u32; + for &weight in huff_weight.iter().take(o_size) { + if weight as usize >= HUF_ABSOLUTE_MAX_TABLELOG { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + rank_stats[weight as usize] += 1; + weight_total += (1u32 << weight) >> 1; + } + if weight_total == 0 { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + let calculated_log = highbit32(weight_total) + 1; + if calculated_log as usize > HUF_ABSOLUTE_MAX_TABLELOG { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + let total = 1u32 << calculated_log; + let rest = total - weight_total; + if rest == 0 || (1u32 << highbit32(rest)) != rest { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + let last_weight = highbit32(rest) + 1; + huff_weight[o_size] = last_weight as u8; + rank_stats[last_weight as usize] += 1; + if rank_stats[1] < 2 || (rank_stats[1] & 1) != 0 { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + *nb_symbols = (o_size + 1) as u32; + *table_log = calculated_log; + i_size + 1 +} + +#[allow(clippy::needless_range_loop)] +unsafe fn huf_read_dtable_x2( + dtable: &mut [u16; 1 + (1 << HUF_MAX_TABLELOG)], + src: *const u8, + src_size: usize, +) -> usize { + let mut huff_weight = [0u8; HUF_MAX_SYMBOL_VALUE + 1]; + let mut rank_val = [0u32; HUF_ABSOLUTE_MAX_TABLELOG + 1]; + let mut nb_symbols = 0u32; + let mut table_log = 0u32; + let i_size = huf_read_stats( + &mut huff_weight, + &mut rank_val, + &mut nb_symbols, + &mut table_log, + src, + src_size, + ); + if ERR_isError(i_size) { + return i_size; + } + if table_log as usize > dtable[0] as usize { + return ERROR(ZstdErrorCode::TableLogTooLarge); + } + dtable[0] = table_log as u16; + + let mut next_rank_start = 0u32; + for weight in 1..=table_log as usize { + let current = next_rank_start; + next_rank_start += rank_val[weight] << (weight - 1); + rank_val[weight] = current; + } + let cells = dtable.as_mut_ptr().add(1) as *mut HufDEltX2; + for (symbol, &weight) in huff_weight.iter().enumerate().take(nb_symbols as usize) { + let weight = weight as usize; + let length = (1u32 << weight) >> 1; + let entry = HufDEltX2 { + byte: symbol as u8, + nb_bits: (table_log + 1 - weight as u32) as u8, + }; + for index in rank_val[weight]..rank_val[weight] + length { + *cells.add(index as usize) = entry; + } + rank_val[weight] += length; + } + i_size +} + +#[inline] +unsafe fn huf_decode_symbol(stream: &mut DStream, table: *const HufDEltX2, table_log: u32) -> u8 { + let entry = *table.add(look_bits_fast(stream, table_log)); + skip_bits(stream, entry.nb_bits as u32); + entry.byte +} + +unsafe fn huf_decode_stream( + dst: *mut u8, + dst_size: usize, + stream: &mut DStream, + table: *const HufDEltX2, + table_log: u32, +) -> usize { + let start = dst; + let end = dst.add(dst_size); + let mut op = dst; + + /* The v0.3 C decoder uses a 4-symbol unrolled loop followed by a tail. + * Decoding one symbol at a time has the same state transitions and keeps + * the same stop-bit validation while remaining easy to audit. */ + while op < end { + let status = reload_dstream(stream); + if status == DSTREAM_TOO_FAR || (status == DSTREAM_COMPLETED && op < end) { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + *op = huf_decode_symbol(stream, table, table_log); + op = op.add(1); + } + if !end_of_dstream(stream) { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + (op as usize) - (start as usize) +} + +unsafe fn huf_decompress4x2_using_dtable( + dst: *mut u8, + dst_size: usize, + c_src: *const u8, + c_src_size: usize, + dtable: &[u16; 1 + (1 << HUF_MAX_TABLELOG)], +) -> usize { + if c_src_size < 10 { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + let length1 = read_le16(c_src) as usize; + let length2 = read_le16(c_src.add(2)) as usize; + let length3 = read_le16(c_src.add(4)) as usize; + let payload = length1 + .checked_add(length2) + .and_then(|v| v.checked_add(length3)) + .and_then(|v| v.checked_add(6)); + let payload = match payload { + Some(value) if value <= c_src_size => value, + _ => return ERROR(ZstdErrorCode::CorruptionDetected), + }; + let length4 = c_src_size - payload; + let stream1 = c_src.add(6); + let stream2 = stream1.add(length1); + let stream3 = stream2.add(length2); + let stream4 = stream3.add(length3); + let segment = dst_size.div_ceil(4); + let starts = [ + dst, + dst.add(segment), + dst.add(segment * 2), + dst.add(segment * 3), + ]; + let sizes = [ + segment.min(dst_size), + segment.min(dst_size.saturating_sub(segment)), + segment.min(dst_size.saturating_sub(segment * 2)), + dst_size.saturating_sub(segment * 3), + ]; + let lengths = [length1, length2, length3, length4]; + let sources = [stream1, stream2, stream3, stream4]; + let table = dtable.as_ptr().add(1) as *const HufDEltX2; + let table_log = dtable[0] as u32; + for index in 0..4 { + let mut stream = DStream { + bit_container: 0, + bits_consumed: 0, + ptr: ptr::null(), + start: ptr::null(), + }; + let error = init_dstream(&mut stream, sources[index], lengths[index]); + if ERR_isError(error) { + return error; + } + let decoded = huf_decode_stream(starts[index], sizes[index], &mut stream, table, table_log); + if ERR_isError(decoded) { + return decoded; + } + } + dst_size +} + +unsafe fn huf_decompress( + dst: *mut u8, + dst_size: usize, + c_src: *const u8, + c_src_size: usize, +) -> usize { + if dst_size == 0 { + return ERROR(ZstdErrorCode::DstSizeTooSmall); + } + if c_src_size > dst_size { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + if c_src_size == dst_size { + ptr::copy_nonoverlapping(c_src, dst, dst_size); + return dst_size; + } + if c_src_size == 1 { + ptr::write_bytes(dst, *c_src, dst_size); + return dst_size; + } + let mut dtable = [0u16; 1 + (1 << HUF_MAX_TABLELOG)]; + dtable[0] = HUF_MAX_TABLELOG as u16; + let header_size = huf_read_dtable_x2(&mut dtable, c_src, c_src_size); + if ERR_isError(header_size) { + return header_size; + } + if header_size >= c_src_size { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + huf_decompress4x2_using_dtable( + dst, + dst_size, + c_src.add(header_size), + c_src_size - header_size, + &dtable, + ) +} + +/* ****************************************** + * v0.3 frame decoder + ********************************************/ + +const LL_TABLE_SIZE: usize = 1 + (1 << LL_FSE_LOG); +const OFF_TABLE_SIZE: usize = 1 + (1 << OFF_FSE_LOG); +const ML_TABLE_SIZE: usize = 1 + (1 << ML_FSE_LOG); + +#[repr(C)] +pub struct ZSTDv03_Dctx { + ll_table: [u32; LL_TABLE_SIZE], + off_table: [u32; OFF_TABLE_SIZE], + ml_table: [u32; ML_TABLE_SIZE], + previous_dst_end: *mut u8, + base: *mut u8, + expected: usize, + b_type: u32, + phase: u32, + lit_ptr: *const u8, + lit_size: usize, + lit_buffer: [u8; BLOCKSIZE + 8], +} + +#[derive(Clone, Copy)] +struct BlockProperties { + block_type: u32, + orig_size: u32, +} + +#[inline] +unsafe fn get_block_size( + src: *const u8, + src_size: usize, + properties: &mut BlockProperties, +) -> usize { + if src_size < 3 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + let flags = *src; + let c_size = + *src.add(2) as usize | ((*src.add(1) as usize) << 8) | (((*src as usize) & 7) << 16); + properties.block_type = (flags >> 6) as u32; + properties.orig_size = if properties.block_type == BT_RLE { + c_size as u32 + } else { + 0 + }; + if properties.block_type == BT_END { + return 0; + } + if properties.block_type == BT_RLE { + return 1; + } + c_size +} + +unsafe fn copy_uncompressed_block( + dst: *mut u8, + max_dst_size: usize, + src: *const u8, + src_size: usize, +) -> usize { + if src_size > max_dst_size { + return ERROR(ZstdErrorCode::DstSizeTooSmall); + } + if src_size != 0 { + ptr::copy_nonoverlapping(src, dst, src_size); + } + src_size +} + +unsafe fn decompress_literals( + dst: *mut u8, + max_dst_size: &mut usize, + src: *const u8, + src_size: usize, +) -> usize { + let lit_size = ((read_le32(src) & 0x1F_FFFF) >> 2) as usize; + let lit_c_size = ((read_le32(src.add(2)) & 0xFF_FFFF) >> 5) as usize; + if lit_size > *max_dst_size { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + if lit_c_size + 5 > src_size { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + let decoded = huf_decompress(dst, lit_size, src.add(5), lit_c_size); + if ERR_isError(decoded) { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + *max_dst_size = lit_size; + lit_c_size + 5 +} + +unsafe fn decode_literals_block(dctx: &mut ZSTDv03_Dctx, src: *const u8, src_size: usize) -> usize { + if src_size < MIN_CBLOCK_SIZE { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + match *src & 3 { + 0 => { + let mut lit_size = BLOCKSIZE; + let read_size = + decompress_literals(dctx.lit_buffer.as_mut_ptr(), &mut lit_size, src, src_size); + dctx.lit_ptr = dctx.lit_buffer.as_ptr(); + dctx.lit_size = lit_size; + if !ERR_isError(read_size) { + dctx.lit_buffer[lit_size..lit_size + 8].fill(0); + } + read_size + } + 1 => { + let lit_size = ((read_le32(src) & 0xFF_FFFF) >> 2) as usize; + if lit_size > src_size.wrapping_sub(11) { + if lit_size > BLOCKSIZE || lit_size > src_size - 3 { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + ptr::copy_nonoverlapping(src, dctx.lit_buffer.as_mut_ptr(), lit_size); + dctx.lit_ptr = dctx.lit_buffer.as_ptr(); + dctx.lit_size = lit_size; + dctx.lit_buffer[lit_size..lit_size + 8].fill(0); + lit_size + 3 + } else { + dctx.lit_ptr = src.add(3); + dctx.lit_size = lit_size; + lit_size + 3 + } + } + 2 => { + let lit_size = ((read_le32(src) & 0xFF_FFFF) >> 2) as usize; + if lit_size > BLOCKSIZE { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + dctx.lit_buffer[..lit_size + 8].fill(*src.add(3)); + dctx.lit_ptr = dctx.lit_buffer.as_ptr(); + dctx.lit_size = lit_size; + 4 + } + _ => unreachable!(), + } +} + +unsafe fn decode_seq_headers( + dctx: &mut ZSTDv03_Dctx, + nb_seq: &mut i32, + dumps: &mut *const u8, + dumps_length: &mut usize, + src: *const u8, + src_size: usize, +) -> usize { + if src_size < 5 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + let start = src as usize; + let end = start + src_size; + let mut ip = src; + *nb_seq = read_le16(ip) as i32; + ip = ip.add(2); + let types = *ip; + let ll_type = (types >> 6) as u32; + let off_type = ((types >> 4) & 3) as u32; + let ml_type = ((types >> 2) & 3) as u32; + let dump_size: usize; + if types & 2 != 0 { + dump_size = *ip.add(2) as usize | ((*ip.add(1) as usize) << 8); + ip = ip.add(3); + } else { + dump_size = *ip.add(1) as usize | (((*ip as usize) & 1) << 8); + ip = ip.add(2); + } + *dumps = ip; + *dumps_length = dump_size; + if dump_size > end.wrapping_sub(ip as usize) { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + ip = ip.add(dump_size); + if end.wrapping_sub(ip as usize) < 3 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + + let mut norm = [0i16; 256]; + match ll_type { + BT_RLE => { + fse_build_dtable_rle(&mut dctx.ll_table, *ip); + ip = ip.add(1); + } + BT_RAW => { + fse_build_dtable_raw(&mut dctx.ll_table, LL_BITS); + } + _ => { + let mut max = MAX_LL; + let mut log = 0; + let size = fse_read_ncount(&mut norm, &mut max, &mut log, ip, end - ip as usize); + if ERR_isError(size) { + return ERROR(ZstdErrorCode::Generic); + } + if log > LL_FSE_LOG { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + ip = ip.add(size); + let error = fse_build_dtable(&mut dctx.ll_table, &norm, max, log); + if ERR_isError(error) { + return error; + } + } + } + + match off_type { + BT_RLE => { + if end.wrapping_sub(ip as usize) < 2 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + fse_build_dtable_rle(&mut dctx.off_table, *ip & MAX_OFF as u8); + ip = ip.add(1); + } + BT_RAW => { + fse_build_dtable_raw(&mut dctx.off_table, OFF_BITS); + } + _ => { + let mut max = MAX_OFF; + let mut log = 0; + let size = fse_read_ncount(&mut norm, &mut max, &mut log, ip, end - ip as usize); + if ERR_isError(size) { + return ERROR(ZstdErrorCode::Generic); + } + if log > OFF_FSE_LOG { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + ip = ip.add(size); + let error = fse_build_dtable(&mut dctx.off_table, &norm, max, log); + if ERR_isError(error) { + return error; + } + } + } + + match ml_type { + BT_RLE => { + if end.wrapping_sub(ip as usize) < 2 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + fse_build_dtable_rle(&mut dctx.ml_table, *ip); + ip = ip.add(1); + } + BT_RAW => { + fse_build_dtable_raw(&mut dctx.ml_table, ML_BITS); + } + _ => { + let mut max = MAX_ML; + let mut log = 0; + let size = fse_read_ncount(&mut norm, &mut max, &mut log, ip, end - ip as usize); + if ERR_isError(size) { + return ERROR(ZstdErrorCode::Generic); + } + if log > ML_FSE_LOG { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + ip = ip.add(size); + let error = fse_build_dtable(&mut dctx.ml_table, &norm, max, log); + if ERR_isError(error) { + return error; + } + } + } + (ip as usize) - start +} + +#[derive(Clone, Copy)] +struct Sequence { + lit_length: usize, + offset: usize, + match_length: usize, +} + +struct SequenceState { + stream: DStream, + state_ll: FseDState, + state_off: FseDState, + state_ml: FseDState, + prev_offset: usize, + dumps: *const u8, + dumps_end: *const u8, +} + +unsafe fn decode_sequence(sequence: &mut Sequence, state: &mut SequenceState) { + let mut dumps = state.dumps; + let dumps_end = state.dumps_end; + let mut lit_length = fse_decode_symbol(&mut state.state_ll, &mut state.stream, false) as usize; + let previous_offset = if lit_length != 0 { + sequence.offset + } else { + state.prev_offset + }; + state.prev_offset = sequence.offset; + if lit_length == MAX_LL as usize { + let add = if (dumps as usize) < dumps_end as usize { + let value = *dumps as usize; + dumps = dumps.add(1); + value + } else { + 0 + }; + if add < 255 { + lit_length += add; + } else if (dumps as usize) <= (dumps_end as usize).wrapping_sub(3) { + lit_length = read_le24(dumps) as usize; + dumps = dumps.add(3); + } + } + + let offset_code = fse_decode_symbol(&mut state.state_off, &mut state.stream, false) as u32; + let nb_bits = offset_code.saturating_sub(1); + let offset_prefix = [ + 1usize, 1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384, 32768, 65536, + 131072, 262144, 524288, 1048576, 2097152, 4194304, 8388608, 16777216, 33554432, 1, 1, 1, 1, + 1, + ]; + let mut offset = + offset_prefix[offset_code as usize].wrapping_add(read_bits(&mut state.stream, nb_bits)); + if offset_code == 0 { + offset = previous_offset; + } + + let mut match_length = + fse_decode_symbol(&mut state.state_ml, &mut state.stream, false) as usize; + if match_length == MAX_ML as usize { + let add = if (dumps as usize) < dumps_end as usize { + let value = *dumps as usize; + dumps = dumps.add(1); + value + } else { + 0 + }; + if add < 255 { + match_length += add; + } else if (dumps as usize) <= (dumps_end as usize).wrapping_sub(3) { + match_length = read_le24(dumps) as usize; + dumps = dumps.add(3); + } + } + sequence.lit_length = lit_length; + sequence.offset = offset; + sequence.match_length = match_length + MINMATCH; + state.dumps = dumps; +} + +unsafe fn exec_sequence( + op: *mut u8, + sequence: Sequence, + lit_ptr: &mut *const u8, + lit_limit: *const u8, + base: *mut u8, + oend: *mut u8, +) -> usize { + let ostart = op as usize; + let o_lit_end = (op as usize).wrapping_add(sequence.lit_length); + let o_match_end = o_lit_end.wrapping_add(sequence.match_length); + let oend_addr = oend as usize; + let lit_end = (*lit_ptr as usize).wrapping_add(sequence.lit_length); + let sequence_length = sequence.lit_length.wrapping_add(sequence.match_length); + + if sequence_length > oend_addr.wrapping_sub(op as usize) { + return ERROR(ZstdErrorCode::DstSizeTooSmall); + } + if sequence.lit_length > (lit_limit as usize).wrapping_sub(*lit_ptr as usize) { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + if o_lit_end > oend_addr.wrapping_sub(8) { + return ERROR(ZstdErrorCode::DstSizeTooSmall); + } + if sequence.offset > ((o_lit_end.wrapping_sub(base as usize)) as u32) as usize { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + if o_match_end > oend_addr { + return ERROR(ZstdErrorCode::DstSizeTooSmall); + } + if lit_end > lit_limit as usize { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + + if sequence.lit_length != 0 { + ptr::copy(*lit_ptr, op, sequence.lit_length); + } + *lit_ptr = lit_end as *const u8; + let match_start = o_lit_end as *mut u8; + let offset = sequence.offset; + if offset == 0 || offset > o_lit_end.wrapping_sub(base as usize) { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + for index in 0..sequence.match_length { + let source = if index >= offset { + match_start.add(index - offset) + } else { + match_start.sub(offset - index) + }; + *match_start.add(index) = *source; + } + o_match_end - ostart +} + +unsafe fn decompress_sequences( + dctx: &mut ZSTDv03_Dctx, + dst: *mut u8, + max_dst_size: usize, + seq_start: *const u8, + seq_size: usize, +) -> usize { + let mut ip = seq_start; + let seq_end = (seq_start as usize).wrapping_add(seq_size); + let ostart = dst; + let mut op = dst; + let oend = dst.add(max_dst_size); + let mut nb_seq = 0i32; + let mut dumps = ptr::null(); + let mut dumps_length = 0usize; + let header_size = decode_seq_headers( + dctx, + &mut nb_seq, + &mut dumps, + &mut dumps_length, + ip, + seq_size, + ); + if ERR_isError(header_size) { + return header_size; + } + ip = ip.add(header_size); + let mut state = SequenceState { + stream: DStream { + bit_container: 0, + bits_consumed: 0, + ptr: ptr::null(), + start: ptr::null(), + }, + state_ll: FseDState { + state: 0, + table: ptr::null(), + }, + state_off: FseDState { + state: 0, + table: ptr::null(), + }, + state_ml: FseDState { + state: 0, + table: ptr::null(), + }, + prev_offset: 4, + dumps, + dumps_end: dumps.add(dumps_length), + }; + let lit_limit = dctx.lit_ptr.add(dctx.lit_size); + let error = init_dstream(&mut state.stream, ip, seq_end - ip as usize); + if ERR_isError(error) { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + fse_init_dstate( + &mut state.state_ll, + &mut state.stream, + dctx.ll_table.as_ptr(), + ); + fse_init_dstate( + &mut state.state_off, + &mut state.stream, + dctx.off_table.as_ptr(), + ); + fse_init_dstate( + &mut state.state_ml, + &mut state.stream, + dctx.ml_table.as_ptr(), + ); + + let mut sequence = Sequence { + lit_length: 0, + offset: 4, + match_length: 0, + }; + while nb_seq > 0 && reload_dstream(&mut state.stream) <= DSTREAM_COMPLETED { + nb_seq -= 1; + decode_sequence(&mut sequence, &mut state); + let produced = exec_sequence(op, sequence, &mut dctx.lit_ptr, lit_limit, dctx.base, oend); + if ERR_isError(produced) { + return produced; + } + op = op.add(produced); + } + + if !end_of_dstream(&state.stream) || nb_seq < 0 { + return ERROR(ZstdErrorCode::CorruptionDetected); + } + let last_literal_size = (lit_limit as usize).wrapping_sub(dctx.lit_ptr as usize); + if op as usize + last_literal_size > oend as usize { + return ERROR(ZstdErrorCode::DstSizeTooSmall); + } + if last_literal_size != 0 { + if !ptr::eq(op, dctx.lit_ptr) { + ptr::copy(dctx.lit_ptr, op, last_literal_size); + } + op = op.add(last_literal_size); + } + (op as usize) - (ostart as usize) +} + +unsafe fn decompress_block( + dctx: &mut ZSTDv03_Dctx, + dst: *mut u8, + max_dst_size: usize, + src: *const u8, + src_size: usize, +) -> usize { + let literals_size = decode_literals_block(dctx, src, src_size); + if ERR_isError(literals_size) { + return literals_size; + } + decompress_sequences( + dctx, + dst, + max_dst_size, + src.add(literals_size), + src_size - literals_size, + ) +} + +unsafe fn decompress_dctx( + dctx: &mut ZSTDv03_Dctx, + dst: *mut u8, + max_dst_size: usize, + src: *const u8, + src_size: usize, +) -> usize { + if src_size < 4 + 3 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + if read_le32(src) != ZSTD_MAGIC_NUMBER { + return ERROR(ZstdErrorCode::PrefixUnknown); + } + let mut ip = src.add(4); + let mut remaining = src_size - 4; + let ostart = dst; + let mut op = dst; + let oend = dst.add(max_dst_size); + loop { + let mut properties = BlockProperties { + block_type: BT_END, + orig_size: 0, + }; + let block_size = get_block_size(ip, remaining, &mut properties); + if ERR_isError(block_size) { + return block_size; + } + ip = ip.add(3); + remaining -= 3; + if block_size > remaining { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + let decoded = match properties.block_type { + BT_COMPRESSED => { + decompress_block(dctx, op, oend as usize - op as usize, ip, block_size) + } + BT_RAW => copy_uncompressed_block(op, oend as usize - op as usize, ip, block_size), + BT_RLE => return ERROR(ZstdErrorCode::Generic), + BT_END => { + if remaining != 0 { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + 0 + } + _ => return ERROR(ZstdErrorCode::Generic), + }; + if block_size == 0 { + break; + } + if ERR_isError(decoded) { + return decoded; + } + op = op.add(decoded); + ip = ip.add(block_size); + remaining -= block_size; + } + (op as usize) - (ostart as usize) +} + +unsafe fn error_frame_size_info(c_size: *mut usize, d_bound: *mut u64, ret: usize) { + *c_size = ret; + *d_bound = ZSTD_CONTENTSIZE_ERROR; +} + +/* ****************************************** + * Exported C ABI + ********************************************/ + +#[no_mangle] +pub extern "C" fn ZSTDv03_isError(code: usize) -> c_uint { + ERR_isError(code) as c_uint +} + +#[no_mangle] +pub unsafe extern "C" fn ZSTDv03_decompress( + dst: *mut c_void, + max_original_size: usize, + src: *const c_void, + compressed_size: usize, +) -> usize { + let mut dctx = std::mem::MaybeUninit::::zeroed(); + let dctx_ptr = dctx.as_mut_ptr(); + (*dctx_ptr).base = dst as *mut u8; + decompress_dctx( + &mut *dctx_ptr, + dst as *mut u8, + max_original_size, + src as *const u8, + compressed_size, + ) +} + +#[no_mangle] +pub unsafe extern "C" fn ZSTDv03_decompressDCtx( + ctx: *mut c_void, + dst: *mut c_void, + max_original_size: usize, + src: *const c_void, + compressed_size: usize, +) -> usize { + decompress_dctx( + &mut *(ctx as *mut ZSTDv03_Dctx), + dst as *mut u8, + max_original_size, + src as *const u8, + compressed_size, + ) +} + +#[no_mangle] +pub unsafe extern "C" fn ZSTDv03_findFrameSizeInfoLegacy( + src: *const c_void, + src_size: usize, + c_size: *mut usize, + d_bound: *mut u64, +) { + if src_size < 4 + 3 { + error_frame_size_info(c_size, d_bound, ERROR(ZstdErrorCode::SrcSizeWrong)); + return; + } + if read_le32(src as *const u8) != ZSTD_MAGIC_NUMBER { + error_frame_size_info(c_size, d_bound, ERROR(ZstdErrorCode::PrefixUnknown)); + return; + } + let start = src as *const u8; + let mut ip = start.add(4); + let mut remaining = src_size - 4; + let mut nb_blocks = 0usize; + loop { + let mut properties = BlockProperties { + block_type: BT_END, + orig_size: 0, + }; + let block_size = get_block_size(ip, remaining, &mut properties); + if ERR_isError(block_size) { + error_frame_size_info(c_size, d_bound, block_size); + return; + } + ip = ip.add(3); + remaining -= 3; + if block_size > remaining { + error_frame_size_info(c_size, d_bound, ERROR(ZstdErrorCode::SrcSizeWrong)); + return; + } + if block_size == 0 { + break; + } + ip = ip.add(block_size); + remaining -= block_size; + nb_blocks += 1; + } + *c_size = ip as usize - start as usize; + *d_bound = (nb_blocks * BLOCKSIZE) as u64; +} + +#[no_mangle] +pub unsafe extern "C" fn ZSTDv03_resetDCtx(dctx: *mut ZSTDv03_Dctx) -> usize { + (*dctx).expected = 4; + (*dctx).phase = 0; + (*dctx).previous_dst_end = ptr::null_mut(); + (*dctx).base = ptr::null_mut(); + 0 +} + +#[no_mangle] +pub unsafe extern "C" fn ZSTDv03_createDCtx() -> *mut ZSTDv03_Dctx { + let dctx = libc::malloc(std::mem::size_of::()) as *mut ZSTDv03_Dctx; + if dctx.is_null() { + return ptr::null_mut(); + } + ZSTDv03_resetDCtx(dctx); + dctx +} + +#[no_mangle] +pub unsafe extern "C" fn ZSTDv03_freeDCtx(dctx: *mut ZSTDv03_Dctx) -> usize { + libc::free(dctx as *mut c_void); + 0 +} + +#[no_mangle] +pub unsafe extern "C" fn ZSTDv03_nextSrcSizeToDecompress(dctx: *mut ZSTDv03_Dctx) -> usize { + (*dctx).expected +} + +#[no_mangle] +pub unsafe extern "C" fn ZSTDv03_decompressContinue( + dctx: *mut ZSTDv03_Dctx, + dst: *mut c_void, + max_dst_size: usize, + src: *const c_void, + src_size: usize, +) -> usize { + let ctx = &mut *dctx; + let dst = dst as *mut u8; + let src = src as *const u8; + if src_size != ctx.expected { + return ERROR(ZstdErrorCode::SrcSizeWrong); + } + if dst != ctx.previous_dst_end { + ctx.base = dst; + } + if ctx.phase == 0 { + if read_le32(src) != ZSTD_MAGIC_NUMBER { + return ERROR(ZstdErrorCode::PrefixUnknown); + } + ctx.phase = 1; + ctx.expected = 3; + return 0; + } + if ctx.phase == 1 { + let mut properties = BlockProperties { + block_type: BT_END, + orig_size: 0, + }; + let block_size = get_block_size(src, 3, &mut properties); + if ERR_isError(block_size) { + return block_size; + } + if properties.block_type == BT_END { + ctx.expected = 0; + ctx.phase = 0; + } else { + ctx.expected = block_size; + ctx.b_type = properties.block_type; + ctx.phase = 2; + } + return 0; + } + + let produced = match ctx.b_type { + BT_COMPRESSED => decompress_block(ctx, dst, max_dst_size, src, src_size), + BT_RAW => copy_uncompressed_block(dst, max_dst_size, src, src_size), + BT_RLE => return ERROR(ZstdErrorCode::Generic), + BT_END => 0, + _ => return ERROR(ZstdErrorCode::Generic), + }; + ctx.phase = 1; + ctx.expected = 3; + if ERR_isError(produced) { + return produced; + } + ctx.previous_dst_end = dst.add(produced); + produced +} + +#[cfg(test)] +mod tests { + use super::*; + + const RAW_FRAME: &[u8] = &[ + 0x23, 0xB5, 0x2F, 0xFD, 0x40, 0x00, 0x0B, b'r', b'a', b'w', b' ', b'v', b'0', b'.', b'3', + b'!', b'!', b'!', 0xC0, 0x00, 0x00, + ]; + + const COMPRESSED_FRAME: &[u8] = &[ + 0x23, 0xB5, 0x2F, 0xFD, 0x00, 0x00, 0x13, 0x1D, 0x00, 0x00, 0x68, 0x65, 0x6C, 0x6C, 0x6F, + 0x20, 0x0A, 0x01, 0x00, 0x54, 0x00, 0x00, 0x01, 0x30, 0x19, 0x46, 0xC0, 0x00, 0x00, + ]; + + const MULTI_BLOCK_FRAME: &[u8] = &[ + 0x23, 0xB5, 0x2F, 0xFD, 0x00, 0x00, 0x42, 0xC5, 0x00, 0x00, 0x54, 0x68, 0x65, 0x20, 0x71, + 0x75, 0x69, 0x63, 0x6B, 0x20, 0x62, 0x72, 0x6F, 0x77, 0x6E, 0x20, 0x66, 0x6F, 0x78, 0x20, + 0x6A, 0x75, 0x6D, 0x70, 0x73, 0x20, 0x6F, 0x76, 0x65, 0x72, 0x20, 0x74, 0x68, 0x65, 0x20, + 0x6C, 0x61, 0x7A, 0x79, 0x20, 0x64, 0x6F, 0x67, 0x2E, 0x20, 0x65, 0x72, 0x20, 0x74, 0x01, + 0x00, 0x54, 0x04, 0xFF, 0xCB, 0xFF, 0x01, 0x80, 0x06, 0x80, 0xBF, 0x69, 0x03, 0x00, 0x00, + 0x13, 0x05, 0x00, 0x00, 0x20, 0x01, 0x00, 0x54, 0x04, 0xFF, 0xE3, 0x6E, 0x01, 0x00, 0xF0, + 0x7F, 0x00, 0xFC, 0x23, 0x10, 0xC0, 0x00, 0x00, + ]; + + unsafe fn decode(frame: &[u8], capacity: usize) -> Result, usize> { + let mut output = vec![0u8; capacity]; + let size = ZSTDv03_decompress( + output.as_mut_ptr() as *mut c_void, + output.len(), + frame.as_ptr() as *const c_void, + frame.len(), + ); + if ZSTDv03_isError(size) != 0 { + Err(size) + } else { + output.truncate(size); + Ok(output) + } + } + + #[test] + fn decodes_raw_and_compressed_v03_frames() { + unsafe { + assert_eq!(decode(RAW_FRAME, 32).unwrap(), b"raw v0.3!!!"); + assert_eq!( + decode(COMPRESSED_FRAME, 64).unwrap(), + b"hello hello hello hello hello\n" + ); + let expected = b"The quick brown fox jumps over the lazy dog. ".repeat(5000); + assert_eq!(decode(MULTI_BLOCK_FRAME, expected.len()).unwrap(), expected); + } + } + + #[test] + fn reports_frame_size_and_errors() { + unsafe { + let mut c_size = 0; + let mut bound = 0; + ZSTDv03_findFrameSizeInfoLegacy( + RAW_FRAME.as_ptr() as *const c_void, + RAW_FRAME.len(), + &mut c_size, + &mut bound, + ); + assert_eq!(c_size, RAW_FRAME.len()); + assert_eq!(bound, BLOCKSIZE as u64); + + let mut bad = RAW_FRAME.to_vec(); + bad[0] ^= 1; + assert_eq!( + decode(&bad, 32).unwrap_err(), + ERROR(ZstdErrorCode::PrefixUnknown) + ); + assert_eq!( + decode(&RAW_FRAME[..RAW_FRAME.len() - 1], 32).unwrap_err(), + ERROR(ZstdErrorCode::SrcSizeWrong) + ); + assert_eq!( + decode(COMPRESSED_FRAME, 1).unwrap_err(), + ERROR(ZstdErrorCode::DstSizeTooSmall) + ); + } + } + + #[test] + fn streaming_raw_frame_matches_one_shot() { + let expected = b"raw v0.3!!!"; + let mut output = vec![0u8; expected.len()]; + let mut input_offset = 0; + let mut output_offset = 0; + unsafe { + let dctx = ZSTDv03_createDCtx(); + assert!(!dctx.is_null()); + loop { + let needed = ZSTDv03_nextSrcSizeToDecompress(dctx); + if needed == 0 { + break; + } + assert!(input_offset + needed <= RAW_FRAME.len()); + let produced = ZSTDv03_decompressContinue( + dctx, + output.as_mut_ptr().add(output_offset) as *mut c_void, + output.len() - output_offset, + RAW_FRAME.as_ptr().add(input_offset) as *const c_void, + needed, + ); + assert_eq!(ZSTDv03_isError(produced), 0); + input_offset += needed; + output_offset += produced; + } + assert_eq!(ZSTDv03_freeDCtx(dctx), 0); + } + assert_eq!(input_offset, RAW_FRAME.len()); + assert_eq!(output_offset, expected.len()); + assert_eq!(output, expected); + } + + #[test] + fn streaming_context_reset_and_null_free() { + unsafe { + let dctx = ZSTDv03_createDCtx(); + assert!(!dctx.is_null()); + assert_eq!(ZSTDv03_nextSrcSizeToDecompress(dctx), 4); + + let bad_magic = [0u8; 4]; + assert_eq!( + ZSTDv03_decompressContinue( + dctx, + ptr::null_mut(), + 0, + bad_magic.as_ptr() as *const c_void, + bad_magic.len(), + ), + ERROR(ZstdErrorCode::PrefixUnknown) + ); + assert_eq!( + ZSTDv03_decompressContinue( + dctx, + ptr::null_mut(), + 0, + bad_magic.as_ptr() as *const c_void, + bad_magic.len() - 1, + ), + ERROR(ZstdErrorCode::SrcSizeWrong) + ); + assert_eq!(ZSTDv03_resetDCtx(dctx), 0); + assert_eq!(ZSTDv03_nextSrcSizeToDecompress(dctx), 4); + assert_eq!(ZSTDv03_freeDCtx(dctx), 0); + assert_eq!(ZSTDv03_freeDCtx(ptr::null_mut()), 0); + } + } +}