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

feat(legacy): port the v0.4 decoder to Rust

Browse Source 
Move the frozen v0.4 frame, entropy, streaming, and dictionary decoder
implementation to Rust while retaining the C translation unit as an ABI
anchor. Register the decoder behind the matching legacy feature and keep its
historical behavior isolated from newer formats.

Test Plan:
- rustfmt +nightly --check --edition 2021 rust/src/legacy/zstd_v04.rs rust/src/legacy/mod.rs
- RUSTC_WRAPPER= CARGO_BUILD_RUSTC_WRAPPER= cargo test --manifest-path rust/Cargo.toml --no-default-features --features decompression,legacy-v04 legacy::zstd_v04
- RUSTC_WRAPPER= CARGO_BUILD_RUSTC_WRAPPER= cargo clippy --manifest-path rust/Cargo.toml --all-targets --no-default-features --features decompression,legacy-v04 -- -D warnings
- C/Rust v0.4 ABI and decompression build checks
- git diff --cached --check
This commit is contained in:
ddidderr
2026-07-12 10:43:32 +02:00
parent b8ccfd7b87
commit 393b2bec45
3 changed files with 2906 additions and 3585 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/legacy/zstd_v04.c
+6 -3585
View File
@@ -5,3594 +5,15 @@
* 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.
*/
/******************************************
* Includes
******************************************/
#include <stddef.h> /* size_t, ptrdiff_t */
#include <string.h> /* memcpy */
#include <stddef.h> /* size_t */
#include "zstd_v04.h"
#include "../common/compiler.h"
#include "../common/error_private.h"
/* ******************************************************************
* mem.h
*******************************************************************/
#ifndef MEM_H_MODULE
#define MEM_H_MODULE
#if defined (__cplusplus)
extern "C" {
#endif
/******************************************
* Compiler-specific
******************************************/
#if defined(_MSC_VER) /* Visual Studio */
# include <stdlib.h> /* _byteswap_ulong */
# include <intrin.h> /* _byteswap_* */
#endif
/****************************************************************
* Basic Types
*****************************************************************/
#if defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
# if defined(_AIX)
# include <inttypes.h>
# else
# include <stdint.h> /* 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
/*-*************************************
* Debug
***************************************/
#include "../common/debug.h"
#ifndef assert
# define assert(condition) ((void)0)
#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 */
/*
zstd - standard compression library
Header File for static linking only
*/
#ifndef ZSTD_STATIC_H
#define ZSTD_STATIC_H
/* *************************************
* Types
***************************************/
#define ZSTD_WINDOWLOG_ABSOLUTEMIN 11
/** from faster to stronger */
typedef enum { ZSTD_fast, ZSTD_greedy, ZSTD_lazy, ZSTD_lazy2, ZSTD_btlazy2 } ZSTD_strategy;
typedef struct
{
U64 srcSize; /* optional : tells how much bytes are present in the frame. Use 0 if not known. */
U32 windowLog; /* largest match distance : larger == more compression, more memory needed during decompression */
U32 contentLog; /* full search segment : larger == more compression, slower, more memory (useless for fast) */
U32 hashLog; /* dispatch table : larger == more memory, faster */
U32 searchLog; /* nb of searches : larger == more compression, slower */
U32 searchLength; /* size of matches : larger == faster decompression, sometimes less compression */
ZSTD_strategy strategy;
} ZSTD_parameters;
typedef ZSTDv04_Dctx ZSTD_DCtx;
/* *************************************
* Advanced functions
***************************************/
/** ZSTD_decompress_usingDict
* Same as ZSTD_decompressDCtx, using a Dictionary content as prefix
* Note : dict can be NULL, in which case, it's equivalent to ZSTD_decompressDCtx() */
static size_t ZSTD_decompress_usingDict(ZSTD_DCtx* ctx,
void* dst, size_t maxDstSize,
const void* src, size_t srcSize,
const void* dict,size_t dictSize);
/* **************************************
* Streaming functions (direct mode)
****************************************/
static size_t ZSTD_resetDCtx(ZSTD_DCtx* dctx);
static size_t ZSTD_getFrameParams(ZSTD_parameters* params, const void* src, size_t srcSize);
static void ZSTD_decompress_insertDictionary(ZSTD_DCtx* ctx, const void* src, size_t srcSize);
static size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx);
static size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize, const void* src, size_t srcSize);
/**
Streaming decompression, bufferless mode
A ZSTD_DCtx object is required to track streaming operations.
Use ZSTD_createDCtx() / ZSTD_freeDCtx() to manage it.
A ZSTD_DCtx object can be re-used multiple times. Use ZSTD_resetDCtx() to return to fresh status.
First operation is to retrieve frame parameters, using ZSTD_getFrameParams().
This function doesn't consume its input. It needs enough input data to properly decode the frame header.
Objective is to retrieve *params.windowlog, to know minimum amount of memory required during decoding.
Result : 0 when successful, it means the ZSTD_parameters structure has been filled.
>0 : means there is not enough data into src. Provides the expected size to successfully decode header.
errorCode, which can be tested using ZSTD_isError() (For example, if it's not a ZSTD header)
Then, you can optionally insert a dictionary.
This operation must mimic the compressor behavior, otherwise decompression will fail or be corrupted.
Then it's possible to start decompression.
Use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() alternatively.
ZSTD_nextSrcSizeToDecompress() tells how much bytes to provide as 'srcSize' to ZSTD_decompressContinue().
ZSTD_decompressContinue() requires this exact amount of bytes, or it will fail.
ZSTD_decompressContinue() needs previous data blocks during decompression, up to (1 << windowlog).
They should preferably be located contiguously, prior to current block. Alternatively, a round buffer is also possible.
@result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst'.
It can be zero, which is not an error; it just means ZSTD_decompressContinue() has decoded some header.
A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero.
Context can then be reset to start a new decompression.
*/
#endif /* ZSTD_STATIC_H */
/*
zstd_internal - common functions to include
Header File for include
*/
#ifndef ZSTD_CCOMMON_H_MODULE
#define ZSTD_CCOMMON_H_MODULE
/* *************************************
* Common macros
***************************************/
#define MIN(a,b) ((a)<(b) ? (a) : (b))
#define MAX(a,b) ((a)>(b) ? (a) : (b))
/* *************************************
* Common constants
***************************************/
#define ZSTD_MAGICNUMBER 0xFD2FB524 /* v0.4 */
#define KB *(1 <<10)
#define MB *(1 <<20)
#define GB *(1U<<30)
#define BLOCKSIZE (128 KB) /* define, for static allocation */
static const size_t ZSTD_blockHeaderSize = 3;
static const size_t ZSTD_frameHeaderSize_min = 5;
#define ZSTD_frameHeaderSize_max 5 /* define, for static allocation */
#define BIT7 128
#define BIT6 64
#define BIT5 32
#define BIT4 16
#define BIT1 2
#define BIT0 1
#define IS_RAW BIT0
#define IS_RLE BIT1
#define MINMATCH 4
#define REPCODE_STARTVALUE 4
#define MLbits 7
#define LLbits 6
#define Offbits 5
#define MaxML ((1<<MLbits) - 1)
#define MaxLL ((1<<LLbits) - 1)
#define MaxOff ((1<<Offbits)- 1)
#define MLFSELog 10
#define LLFSELog 10
#define OffFSELog 9
#define MaxSeq MAX(MaxLL, MaxML)
#define MIN_SEQUENCES_SIZE (2 /*seqNb*/ + 2 /*dumps*/ + 3 /*seqTables*/ + 1 /*bitStream*/)
#define MIN_CBLOCK_SIZE (3 /*litCSize*/ + MIN_SEQUENCES_SIZE)
#define ZSTD_CONTENTSIZE_ERROR (0ULL - 2)
typedef enum { bt_compressed, bt_raw, bt_rle, bt_end } blockType_t;
/* ******************************************
* Shared functions to include for inlining
********************************************/
static void ZSTD_copy8(void* dst, const void* src) { memcpy(dst, src, 8); }
#define COPY8(d,s) { ZSTD_copy8(d,s); d+=8; s+=8; }
/*! ZSTD_wildcopy : custom version of memcpy(), can copy up to 7-8 bytes too many */
static void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length)
{
const BYTE* ip = (const BYTE*)src;
BYTE* op = (BYTE*)dst;
BYTE* const oend = op + length;
do
COPY8(op, ip)
while (op < oend);
}
/* ******************************************************************
FSE : Finite State Entropy coder
header file
****************************************************************** */
#ifndef FSE_H
#define FSE_H
#if defined (__cplusplus)
extern "C" {
#endif
/* *****************************************
* Includes
******************************************/
#include <stddef.h> /* size_t, ptrdiff_t */
/* *****************************************
* FSE simple functions
******************************************/
static size_t FSE_decompress(void* dst, size_t maxDstSize,
const void* cSrc, size_t cSrcSize);
/*!
FSE_decompress():
Decompress FSE data from buffer 'cSrc', of size 'cSrcSize',
into already allocated destination buffer 'dst', of size 'maxDstSize'.
return : size of regenerated data (<= maxDstSize)
or an error code, which can be tested using FSE_isError()
** Important ** : FSE_decompress() doesn't decompress non-compressible nor RLE data !!!
Why ? : making this distinction requires a header.
Header management is intentionally delegated to the user layer, which can better manage special cases.
*/
/* *****************************************
* Tool functions
******************************************/
/* Error Management */
static unsigned FSE_isError(size_t code); /* tells if a return value is an error code */
/* *****************************************
* FSE detailed API
******************************************/
/*!
FSE_compress() does the following:
1. count symbol occurrence from source[] into table count[]
2. normalize counters so that sum(count[]) == Power_of_2 (2^tableLog)
3. save normalized counters to memory buffer using writeNCount()
4. build encoding table 'CTable' from normalized counters
5. encode the data stream using encoding table 'CTable'
FSE_decompress() does the following:
1. read normalized counters with readNCount()
2. build decoding table 'DTable' from normalized counters
3. decode the data stream using decoding table 'DTable'
The following API allows targeting specific sub-functions for advanced tasks.
For example, it's possible to compress several blocks using the same 'CTable',
or to save and provide normalized distribution using external method.
*/
/* *** DECOMPRESSION *** */
/*!
FSE_readNCount():
Read compactly saved 'normalizedCounter' from 'rBuffer'.
return : size read from 'rBuffer'
or an errorCode, which can be tested using FSE_isError()
maxSymbolValuePtr[0] and tableLogPtr[0] will also be updated with their respective values */
static size_t FSE_readNCount (short* normalizedCounter, unsigned* maxSymbolValuePtr, unsigned* tableLogPtr, const void* rBuffer, size_t rBuffSize);
/*!
Constructor and Destructor of type FSE_DTable
Note that its size depends on 'tableLog' */
typedef unsigned FSE_DTable; /* don't allocate that. It's just a way to be more restrictive than void* */
/*!
FSE_buildDTable():
Builds 'dt', which must be already allocated, using FSE_createDTable()
return : 0,
or an errorCode, which can be tested using FSE_isError() */
static size_t FSE_buildDTable ( FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);
/*!
FSE_decompress_usingDTable():
Decompress compressed source 'cSrc' of size 'cSrcSize' using 'dt'
into 'dst' which must be already allocated.
return : size of regenerated data (necessarily <= maxDstSize)
or an errorCode, which can be tested using FSE_isError() */
static size_t FSE_decompress_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const FSE_DTable* dt);
/*!
Tutorial :
----------
(Note : these functions only decompress FSE-compressed blocks.
If block is uncompressed, use memcpy() instead
If block is a single repeated byte, use memset() instead )
The first step is to obtain the normalized frequencies of symbols.
This can be performed by FSE_readNCount() if it was saved using FSE_writeNCount().
'normalizedCounter' must be already allocated, and have at least 'maxSymbolValuePtr[0]+1' cells of signed short.
In practice, that means it's necessary to know 'maxSymbolValue' beforehand,
or size the table to handle worst case situations (typically 256).
FSE_readNCount() will provide 'tableLog' and 'maxSymbolValue'.
The result of FSE_readNCount() is the number of bytes read from 'rBuffer'.
Note that 'rBufferSize' must be at least 4 bytes, even if useful information is less than that.
If there is an error, the function will return an error code, which can be tested using FSE_isError().
The next step is to build the decompression tables 'FSE_DTable' from 'normalizedCounter'.
This is performed by the function FSE_buildDTable().
The space required by 'FSE_DTable' must be already allocated using FSE_createDTable().
If there is an error, the function will return an error code, which can be tested using FSE_isError().
'FSE_DTable' can then be used to decompress 'cSrc', with FSE_decompress_usingDTable().
'cSrcSize' must be strictly correct, otherwise decompression will fail.
FSE_decompress_usingDTable() result will tell how many bytes were regenerated (<=maxDstSize).
If there is an error, the function will return an error code, which can be tested using FSE_isError(). (ex: dst buffer too small)
*/
#if defined (__cplusplus)
}
#endif
#endif /* FSE_H */
/* ******************************************************************
bitstream
Part of NewGen Entropy library
header file (to include)
Copyright (C) 2013-2015, Yann Collet.
BSD 2-Clause License (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);/* fall-through */
case 6: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[5]) << (sizeof(size_t)*8 - 24);/* fall-through */
case 5: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[4]) << (sizeof(size_t)*8 - 32);/* fall-through */
case 4: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[3]) << 24; /* fall-through */
case 3: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[2]) << 16; /* fall-through */
case 2: bitD->bitContainer += (size_t)(((const BYTE*)(bitD->start))[1]) << 8; /* fall-through */
default: break;
}
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 */
/* ******************************************************************
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
****************************************************************** */
#ifndef FSE_STATIC_H
#define FSE_STATIC_H
#if defined (__cplusplus)
extern "C" {
#endif
/* *****************************************
* Static allocation
*******************************************/
/* FSE buffer bounds */
#define FSE_NCOUNTBOUND 512
#define FSE_BLOCKBOUND(size) (size + (size>>7))
#define FSE_COMPRESSBOUND(size) (FSE_NCOUNTBOUND + FSE_BLOCKBOUND(size)) /* Macro version, useful for static allocation */
/* It is possible to statically allocate FSE CTable/DTable as a table of unsigned using below macros */
#define FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) (1 + (1<<(maxTableLog-1)) + ((maxSymbolValue+1)*2))
#define FSE_DTABLE_SIZE_U32(maxTableLog) (1 + (1<<maxTableLog))
/* *****************************************
* FSE advanced API
*******************************************/
static size_t FSE_buildDTable_raw (FSE_DTable* dt, unsigned nbBits);
/* build a fake FSE_DTable, designed to read an uncompressed bitstream where each symbol uses nbBits */
static size_t FSE_buildDTable_rle (FSE_DTable* dt, unsigned char symbolValue);
/* build a fake FSE_DTable, designed to always generate the same symbolValue */
/* *****************************************
* FSE symbol decompression API
*******************************************/
typedef struct
{
size_t state;
const void* table; /* precise table may vary, depending on U16 */
} FSE_DState_t;
static void FSE_initDState(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD, const FSE_DTable* dt);
static unsigned char FSE_decodeSymbol(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD);
static unsigned FSE_endOfDState(const FSE_DState_t* DStatePtr);
/* *****************************************
* FSE unsafe API
*******************************************/
static unsigned char FSE_decodeSymbolFast(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD);
/* faster, but works only if nbBits is always >= 1 (otherwise, result will be corrupted) */
/* *****************************************
* Implementation of inlined functions
*******************************************/
/* decompression */
typedef struct {
U16 tableLog;
U16 fastMode;
} FSE_DTableHeader; /* sizeof U32 */
typedef struct
{
unsigned short newState;
unsigned char symbol;
unsigned char nbBits;
} FSE_decode_t; /* size == U32 */
MEM_STATIC void FSE_initDState(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD, const FSE_DTable* dt)
{
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
#endif /* FSE_STATIC_H */
/* ******************************************************************
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
#define FSE_DECODE_TYPE FSE_decode_t
#endif /* !FSE_COMMONDEFS_ONLY */
/* **************************************************************
* Compiler specifics
****************************************************************/
#ifdef _MSC_VER /* Visual Studio */
# define FORCE_INLINE static __forceinline
# include <intrin.h> /* 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
/* **************************************************************
* Dependencies
****************************************************************/
#include <stdlib.h> /* malloc, free, qsort */
#include <string.h> /* memcpy, memset */
#include <stdio.h> /* printf (debug) */
/* ***************************************************************
* Constants
*****************************************************************/
#define FSE_MAX_TABLELOG (FSE_MAX_MEMORY_USAGE-2)
#define FSE_MAX_TABLESIZE (1U<<FSE_MAX_TABLELOG)
#define FSE_MAXTABLESIZE_MASK (FSE_MAX_TABLESIZE-1)
#define FSE_DEFAULT_TABLELOG (FSE_DEFAULT_MEMORY_USAGE-2)
#define FSE_MIN_TABLELOG 5
#define FSE_TABLELOG_ABSOLUTE_MAX 15
#if FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX
#error "FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX is not supported"
#endif
/* **************************************************************
* Error Management
****************************************************************/
#define FSE_STATIC_ASSERT(c) { enum { FSE_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */
/* **************************************************************
* Complex types
****************************************************************/
typedef U32 DTable_max_t[FSE_DTABLE_SIZE_U32(FSE_MAX_TABLELOG)];
/*-**************************************************************
* Templates
****************************************************************/
/*
designed to be included
for type-specific functions (template emulation in C)
Objective is to write these functions only once, for improved maintenance
*/
/* safety checks */
#ifndef FSE_FUNCTION_EXTENSION
# error "FSE_FUNCTION_EXTENSION must be defined"
#endif
#ifndef FSE_FUNCTION_TYPE
# error "FSE_FUNCTION_TYPE must be defined"
#endif
/* Function names */
#define FSE_CAT(X,Y) X##Y
#define FSE_FUNCTION_NAME(X,Y) FSE_CAT(X,Y)
#define FSE_TYPE_NAME(X,Y) FSE_CAT(X,Y)
static U32 FSE_tableStep(U32 tableSize) { return (tableSize>>1) + (tableSize>>3) + 3; }
static size_t FSE_buildDTable(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog)
{
FSE_DTableHeader DTableH;
void* const tdPtr = dt+1; /* because dt is unsigned, 32-bits aligned on 32-bits */
FSE_DECODE_TYPE* const tableDecode = (FSE_DECODE_TYPE*) (tdPtr);
const U32 tableSize = 1 << tableLog;
const U32 tableMask = tableSize-1;
const U32 step = FSE_tableStep(tableSize);
U16 symbolNext[FSE_MAX_SYMBOL_VALUE+1];
U32 position = 0;
U32 highThreshold = tableSize-1;
const S16 largeLimit= (S16)(1 << (tableLog-1));
U32 noLarge = 1;
U32 s;
/* Sanity Checks */
if (maxSymbolValue > FSE_MAX_SYMBOL_VALUE) return ERROR(maxSymbolValue_tooLarge);
if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge);
/* Init, lay down lowprob symbols */
memset(tableDecode, 0, sizeof(FSE_DECODE_TYPE) * (maxSymbolValue+1) ); /* useless init, but keep static analyzer happy, and we don't need to performance optimize legacy decoders */
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<normalizedCounter[s]; i++)
{
tableDecode[position].symbol = (FSE_FUNCTION_TYPE)s;
position = (position + step) & tableMask;
while (position > 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<tableSize; i++)
{
FSE_FUNCTION_TYPE symbol = (FSE_FUNCTION_TYPE)(tableDecode[i].symbol);
U16 nextState = symbolNext[symbol]++;
tableDecode[i].nbBits = (BYTE) (tableLog - BIT_highbit32 ((U32)nextState) );
tableDecode[i].newState = (U16) ( (nextState << tableDecode[i].nbBits) - tableSize);
}
}
DTableH.fastMode = (U16)noLarge;
memcpy(dt, &DTableH, sizeof(DTableH));
return 0;
}
#ifndef FSE_COMMONDEFS_ONLY
/******************************************
* FSE helper functions
******************************************/
static unsigned FSE_isError(size_t code) { return ERR_isError(code); }
/****************************************************************
* FSE NCount encoding-decoding
****************************************************************/
static short FSE_abs(short a)
{
return a<0 ? -a : a;
}
static size_t FSE_readNCount (short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
const void* headerBuffer, size_t hbSize)
{
const BYTE* const istart = (const BYTE*) headerBuffer;
const BYTE* const iend = istart + hbSize;
const BYTE* ip = istart;
int nbBits;
int remaining;
int threshold;
U32 bitStream;
int bitCount;
unsigned charnum = 0;
int previous0 = 0;
if (hbSize < 4) return ERROR(srcSize_wrong);
bitStream = MEM_readLE32(ip);
nbBits = (bitStream & 0xF) + FSE_MIN_TABLELOG; /* extract tableLog */
if (nbBits > FSE_TABLELOG_ABSOLUTE_MAX) return ERROR(tableLog_tooLarge);
bitStream >>= 4;
bitCount = 4;
*tableLogPtr = nbBits;
remaining = (1<<nbBits)+1;
threshold = 1<<nbBits;
nbBits++;
while ((remaining>1) && (charnum<=*maxSVPtr))
{
if (previous0)
{
unsigned n0 = charnum;
while ((bitStream & 0xFFFF) == 0xFFFF)
{
n0+=24;
if (ip < iend-5)
{
ip+=2;
bitStream = MEM_readLE32(ip) >> bitCount;
}
else
{
bitStream >>= 16;
bitCount+=16;
}
}
while ((bitStream & 3) == 3)
{
n0+=3;
bitStream>>=2;
bitCount+=2;
}
n0 += bitStream & 3;
bitCount += 2;
if (n0 > *maxSVPtr) return ERROR(maxSymbolValue_tooSmall);
while (charnum < n0) normalizedCounter[charnum++] = 0;
if ((ip <= iend-7) || (ip + (bitCount>>3) <= iend-4))
{
ip += bitCount>>3;
bitCount &= 7;
bitStream = MEM_readLE32(ip) >> bitCount;
}
else
bitStream >>= 2;
}
{
const short max = (short)((2*threshold-1)-remaining);
short count;
if ((bitStream & (threshold-1)) < (U32)max)
{
count = (short)(bitStream & (threshold-1));
bitCount += nbBits-1;
}
else
{
count = (short)(bitStream & (2*threshold-1));
if (count >= threshold) count -= max;
bitCount += nbBits;
}
count--; /* extra accuracy */
remaining -= FSE_abs(count);
normalizedCounter[charnum++] = count;
previous0 = !count;
while (remaining < threshold)
{
nbBits--;
threshold >>= 1;
}
{
if ((ip <= iend-7) || (ip + (bitCount>>3) <= iend-4))
{
ip += bitCount>>3;
bitCount &= 7;
}
else
{
bitCount -= (int)(8 * (iend - 4 - ip));
ip = iend - 4;
}
bitStream = MEM_readLE32(ip) >> (bitCount & 31);
}
}
}
if (remaining != 1) return ERROR(GENERIC);
*maxSVPtr = charnum-1;
ip += (bitCount+7)>>3;
if ((size_t)(ip-istart) > hbSize) return ERROR(srcSize_wrong);
return ip-istart;
}
/*********************************************************
* Decompression (Byte symbols)
*********************************************************/
static size_t FSE_buildDTable_rle (FSE_DTable* dt, BYTE symbolValue)
{
void* ptr = dt;
FSE_DTableHeader* const DTableH = (FSE_DTableHeader*)ptr;
void* dPtr = dt + 1;
FSE_decode_t* const cell = (FSE_decode_t*)dPtr;
DTableH->tableLog = 0;
DTableH->fastMode = 0;
cell->newState = 0;
cell->symbol = symbolValue;
cell->nbBits = 0;
return 0;
}
static size_t FSE_buildDTable_raw (FSE_DTable* dt, unsigned nbBits)
{
void* ptr = dt;
FSE_DTableHeader* const DTableH = (FSE_DTableHeader*)ptr;
void* dPtr = dt + 1;
FSE_decode_t* const dinfo = (FSE_decode_t*)dPtr;
const unsigned tableSize = 1 << nbBits;
const unsigned tableMask = tableSize - 1;
const unsigned maxSymbolValue = tableMask;
unsigned s;
/* Sanity checks */
if (nbBits < 1) return ERROR(GENERIC); /* min size */
/* Build Decoding Table */
DTableH->tableLog = (U16)nbBits;
DTableH->fastMode = 1;
for (s=0; s<=maxSymbolValue; s++)
{
dinfo[s].newState = 0;
dinfo[s].symbol = (BYTE)s;
dinfo[s].nbBits = (BYTE)nbBits;
}
return 0;
}
FORCE_INLINE size_t FSE_decompress_usingDTable_generic(
void* dst, size_t maxDstSize,
const void* cSrc, size_t cSrcSize,
const FSE_DTable* dt, const unsigned fast)
{
BYTE* const ostart = (BYTE*) dst;
BYTE* op = ostart;
BYTE* const omax = op + maxDstSize;
BYTE* const olimit = omax-3;
BIT_DStream_t bitD;
FSE_DState_t state1;
FSE_DState_t state2;
size_t errorCode;
/* Init */
errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize); /* replaced last arg by maxCompressed Size */
if (FSE_isError(errorCode)) return errorCode;
FSE_initDState(&state1, &bitD, dt);
FSE_initDState(&state2, &bitD, dt);
#define FSE_GETSYMBOL(statePtr) fast ? FSE_decodeSymbolFast(statePtr, &bitD) : FSE_decodeSymbol(statePtr, &bitD)
/* 4 symbols per loop */
for ( ; (BIT_reloadDStream(&bitD)==BIT_DStream_unfinished) && (op<olimit) ; op+=4)
{
op[0] = FSE_GETSYMBOL(&state1);
if (FSE_MAX_TABLELOG*2+7 > 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;
U32 fastMode;
memcpy(&DTableH, dt, sizeof(DTableH));
fastMode = DTableH.fastMode;
/* select fast mode (static) */
if (fastMode) return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 1);
return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 0);
}
static size_t FSE_decompress(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize)
{
const BYTE* const istart = (const BYTE*)cSrc;
const BYTE* ip = istart;
short counting[FSE_MAX_SYMBOL_VALUE+1];
DTable_max_t dt; /* Static analyzer seems unable to understand this table will be properly initialized later */
unsigned tableLog;
unsigned maxSymbolValue = FSE_MAX_SYMBOL_VALUE;
size_t errorCode;
if (cSrcSize<2) return ERROR(srcSize_wrong); /* too small input size */
/* normal FSE decoding mode */
errorCode = FSE_readNCount (counting, &maxSymbolValue, &tableLog, istart, cSrcSize);
if (FSE_isError(errorCode)) return errorCode;
if (errorCode >= cSrcSize) return ERROR(srcSize_wrong); /* too small input size */
ip += errorCode;
cSrcSize -= errorCode;
errorCode = FSE_buildDTable (dt, counting, maxSymbolValue, tableLog);
if (FSE_isError(errorCode)) return errorCode;
/* always return, even if it is an error code */
return FSE_decompress_usingDTable (dst, maxDstSize, ip, cSrcSize, dt);
}
#endif /* FSE_COMMONDEFS_ONLY */
/* ******************************************************************
Huff0 : Huffman coder, part of New Generation Entropy library
header file
Copyright (C) 2013-2015, Yann Collet.
BSD 2-Clause License (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 HUFF0_H
#define HUFF0_H
#if defined (__cplusplus)
extern "C" {
#endif
/* ****************************************
* Dependency
******************************************/
#include <stddef.h> /* size_t */
/* ****************************************
* Huff0 simple functions
******************************************/
static size_t HUF_decompress(void* dst, size_t dstSize,
const void* cSrc, size_t cSrcSize);
/*!
HUF_decompress():
Decompress Huff0 data from buffer 'cSrc', of size 'cSrcSize',
into already allocated destination buffer 'dst', of size 'dstSize'.
'dstSize' must be the exact size of original (uncompressed) data.
Note : in contrast with FSE, HUF_decompress can regenerate RLE (cSrcSize==1) and uncompressed (cSrcSize==dstSize) data, because it knows size to regenerate.
@return : size of regenerated data (== dstSize)
or an error code, which can be tested using HUF_isError()
*/
/* ****************************************
* Tool functions
******************************************/
/* Error Management */
static unsigned HUF_isError(size_t code); /* tells if a return value is an error code */
#if defined (__cplusplus)
}
#endif
#endif /* HUFF0_H */
/* ******************************************************************
Huff0 : Huffman coder, part of New Generation Entropy library
header file for static linking (only)
Copyright (C) 2013-2015, Yann Collet
BSD 2-Clause License (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 HUFF0_STATIC_H
#define HUFF0_STATIC_H
#if defined (__cplusplus)
extern "C" {
#endif
/* ****************************************
* Static allocation macros
******************************************/
/* static allocation of Huff0's DTable */
#define HUF_DTABLE_SIZE(maxTableLog) (1 + (1<<maxTableLog)) /* nb Cells; use unsigned short for X2, unsigned int for X4 */
#define HUF_CREATE_STATIC_DTABLEX2(DTable, maxTableLog) \
unsigned short DTable[HUF_DTABLE_SIZE(maxTableLog)] = { maxTableLog }
#define HUF_CREATE_STATIC_DTABLEX4(DTable, maxTableLog) \
unsigned int DTable[HUF_DTABLE_SIZE(maxTableLog)] = { maxTableLog }
#define HUF_CREATE_STATIC_DTABLEX6(DTable, maxTableLog) \
unsigned int DTable[HUF_DTABLE_SIZE(maxTableLog) * 3 / 2] = { maxTableLog }
/* ****************************************
* Advanced decompression functions
******************************************/
static size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /* single-symbol decoder */
static size_t HUF_decompress4X4 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /* double-symbols decoder */
/* ****************************************
* Huff0 detailed API
******************************************/
/*!
HUF_decompress() does the following:
1. select the decompression algorithm (X2, X4, X6) based on pre-computed heuristics
2. build Huffman table from save, using HUF_readDTableXn()
3. decode 1 or 4 segments in parallel using HUF_decompressSXn_usingDTable
*/
static size_t HUF_readDTableX2 (unsigned short* DTable, const void* src, size_t srcSize);
static size_t HUF_readDTableX4 (unsigned* DTable, const void* src, size_t srcSize);
static size_t HUF_decompress4X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const unsigned short* DTable);
static size_t HUF_decompress4X4_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const unsigned* DTable);
#if defined (__cplusplus)
}
#endif
#endif /* HUFF0_STATIC_H */
/* ******************************************************************
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
****************************************************************** */
/* **************************************************************
* Compiler specifics
****************************************************************/
#if defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
/* inline is defined */
#elif defined(_MSC_VER)
# define inline __inline
#else
# define inline /* disable inline */
#endif
#ifdef _MSC_VER /* Visual Studio */
# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */
#endif
/* **************************************************************
* Includes
****************************************************************/
#include <stdlib.h> /* malloc, free, qsort */
#include <string.h> /* memcpy, memset */
#include <stdio.h> /* printf (debug) */
/* **************************************************************
* Constants
****************************************************************/
#define HUF_ABSOLUTEMAX_TABLELOG 16 /* absolute limit of HUF_MAX_TABLELOG. Beyond that value, code does not work */
#define HUF_MAX_TABLELOG 12 /* max configured tableLog (for static allocation); can be modified up to HUF_ABSOLUTEMAX_TABLELOG */
#define HUF_DEFAULT_TABLELOG HUF_MAX_TABLELOG /* tableLog by default, when not specified */
#define HUF_MAX_SYMBOL_VALUE 255
#if (HUF_MAX_TABLELOG > HUF_ABSOLUTEMAX_TABLELOG)
# error "HUF_MAX_TABLELOG is too large !"
#endif
/* **************************************************************
* Error Management
****************************************************************/
static unsigned HUF_isError(size_t code) { return ERR_isError(code); }
#define HUF_STATIC_ASSERT(c) { enum { HUF_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */
/*-*******************************************************
* Huff0 : Huffman block decompression
*********************************************************/
typedef struct { BYTE byte; BYTE nbBits; } HUF_DEltX2; /* single-symbol decoding */
typedef struct { U16 sequence; BYTE nbBits; BYTE length; } HUF_DEltX4; /* double-symbols decoding */
typedef struct { BYTE symbol; BYTE weight; } sortedSymbol_t;
/*! HUF_readStats
Read compact Huffman tree, saved by HUF_writeCTable
@huffWeight : destination buffer
@return : size read from `src`
*/
static size_t HUF_readStats(BYTE* huffWeight, size_t hwSize, U32* rankStats,
U32* nbSymbolsPtr, U32* tableLogPtr,
const void* src, size_t srcSize)
{
U32 weightTotal;
U32 tableLog;
const BYTE* ip = (const BYTE*) src;
size_t iSize;
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<oSize; n+=2)
{
huffWeight[n] = ip[n/2] >> 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<oSize; n++)
{
if (huffWeight[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;
size_t iSize;
U32 nbSymbols = 0;
U32 n;
U32 nextRankStart;
void* const dtPtr = DTable + 1;
HUF_DEltX2* const dt = (HUF_DEltX2*)dtPtr;
HUF_STATIC_ASSERT(sizeof(HUF_DEltX2) == sizeof(U16)); /* if compilation fails here, assertion is false */
//memset(huffWeight, 0, sizeof(huffWeight)); /* is not necessary, even though some analyzer complain ... */
iSize = HUF_readStats(huffWeight, HUF_MAX_SYMBOL_VALUE + 1, rankVal, &nbSymbols, &tableLog, src, srcSize);
if (HUF_isError(iSize)) return iSize;
/* check result */
if (tableLog > DTable[0]) return ERROR(tableLog_tooLarge); /* DTable is too small */
DTable[0] = (U16)tableLog; /* maybe should separate sizeof DTable, as allocated, from used size of DTable, in case of DTable re-use */
/* Prepare ranks */
nextRankStart = 0;
for (n=1; n<=tableLog; n++)
{
U32 current = nextRankStart;
nextRankStart += (rankVal[n] << (n-1));
rankVal[n] = current;
}
/* fill DTable */
for (n=0; n<nbSymbols; n++)
{
const U32 w = huffWeight[n];
const U32 length = (1 << w) >> 1;
U32 i;
HUF_DEltX2 D;
D.byte = (BYTE)n; D.nbBits = (BYTE)(tableLog + 1 - w);
for (i = rankVal[w]; i < rankVal[w] + length; i++)
dt[i] = D;
rankVal[w] += length;
}
return iSize;
}
static BYTE HUF_decodeSymbolX2(BIT_DStream_t* Dstream, const HUF_DEltX2* dt, const U32 dtLog)
{
const size_t val = BIT_lookBitsFast(Dstream, dtLog); /* note : dtLog >= 1 */
const BYTE c = dt[val].byte;
BIT_skipBits(Dstream, dt[val].nbBits);
return c;
}
#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) \
*ptr++ = HUF_decodeSymbolX2(DStreamPtr, dt, dtLog)
#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \
if (MEM_64bits() || (HUF_MAX_TABLELOG<=12)) \
HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)
#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \
if (MEM_64bits()) \
HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)
static inline size_t HUF_decodeStreamX2(BYTE* p, BIT_DStream_t* const bitDPtr, BYTE* const pEnd, const HUF_DEltX2* const dt, const U32 dtLog)
{
BYTE* const pStart = p;
/* up to 4 symbols at a time */
while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p <= pEnd-4))
{
HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
HUF_DECODE_SYMBOLX2_1(p, bitDPtr);
HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
}
/* closer to the end */
while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p < pEnd))
HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
/* no more data to retrieve from bitstream, hence no need to reload */
while (p < pEnd)
HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
return pEnd-pStart;
}
static size_t HUF_decompress4X2_usingDTable(
void* dst, size_t dstSize,
const void* cSrc, size_t cSrcSize,
const U16* DTable)
{
if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
{
const BYTE* const istart = (const BYTE*) cSrc;
BYTE* const ostart = (BYTE*) dst;
BYTE* const oend = ostart + dstSize;
const void* const dtPtr = DTable;
const HUF_DEltX2* const dt = ((const HUF_DEltX2*)dtPtr) +1;
const U32 dtLog = DTable[0];
size_t errorCode;
/* Init */
BIT_DStream_t bitD1;
BIT_DStream_t bitD2;
BIT_DStream_t bitD3;
BIT_DStream_t bitD4;
const size_t length1 = MEM_readLE16(istart);
const size_t length2 = MEM_readLE16(istart+2);
const size_t length3 = MEM_readLE16(istart+4);
size_t length4;
const BYTE* const istart1 = istart + 6; /* jumpTable */
const BYTE* const istart2 = istart1 + length1;
const BYTE* const istart3 = istart2 + length2;
const BYTE* const istart4 = istart3 + length3;
const size_t segmentSize = (dstSize+3) / 4;
BYTE* const opStart2 = ostart + segmentSize;
BYTE* const opStart3 = opStart2 + segmentSize;
BYTE* const opStart4 = opStart3 + segmentSize;
BYTE* op1 = ostart;
BYTE* op2 = opStart2;
BYTE* op3 = opStart3;
BYTE* op4 = opStart4;
U32 endSignal;
length4 = cSrcSize - (length1 + length2 + length3 + 6);
if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */
errorCode = BIT_initDStream(&bitD1, istart1, length1);
if (HUF_isError(errorCode)) return errorCode;
errorCode = BIT_initDStream(&bitD2, istart2, length2);
if (HUF_isError(errorCode)) return errorCode;
errorCode = BIT_initDStream(&bitD3, istart3, length3);
if (HUF_isError(errorCode)) return errorCode;
errorCode = BIT_initDStream(&bitD4, istart4, length4);
if (HUF_isError(errorCode)) return errorCode;
/* 16-32 symbols per loop (4-8 symbols per stream) */
endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
for ( ; (endSignal==BIT_DStream_unfinished) && (op4<(oend-7)) ; )
{
HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
HUF_DECODE_SYMBOLX2_1(op1, &bitD1);
HUF_DECODE_SYMBOLX2_1(op2, &bitD2);
HUF_DECODE_SYMBOLX2_1(op3, &bitD3);
HUF_DECODE_SYMBOLX2_1(op4, &bitD4);
HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
HUF_DECODE_SYMBOLX2_0(op1, &bitD1);
HUF_DECODE_SYMBOLX2_0(op2, &bitD2);
HUF_DECODE_SYMBOLX2_0(op3, &bitD3);
HUF_DECODE_SYMBOLX2_0(op4, &bitD4);
endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
}
/* check corruption */
if (op1 > opStart2) return ERROR(corruption_detected);
if (op2 > opStart3) return ERROR(corruption_detected);
if (op3 > opStart4) return ERROR(corruption_detected);
/* note : op4 supposed already verified within main loop */
/* finish bitStreams one by one */
HUF_decodeStreamX2(op1, &bitD1, opStart2, dt, dtLog);
HUF_decodeStreamX2(op2, &bitD2, opStart3, dt, dtLog);
HUF_decodeStreamX2(op3, &bitD3, opStart4, dt, dtLog);
HUF_decodeStreamX2(op4, &bitD4, oend, dt, dtLog);
/* check */
endSignal = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
if (!endSignal) return ERROR(corruption_detected);
/* decoded size */
return dstSize;
}
}
static size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)
{
HUF_CREATE_STATIC_DTABLEX2(DTable, HUF_MAX_TABLELOG);
const BYTE* ip = (const BYTE*) cSrc;
size_t errorCode;
errorCode = HUF_readDTableX2 (DTable, cSrc, cSrcSize);
if (HUF_isError(errorCode)) return errorCode;
if (errorCode >= cSrcSize) return ERROR(srcSize_wrong);
ip += errorCode;
cSrcSize -= errorCode;
return HUF_decompress4X2_usingDTable (dst, dstSize, ip, cSrcSize, DTable);
}
/***************************/
/* double-symbols decoding */
/***************************/
static void HUF_fillDTableX4Level2(HUF_DEltX4* DTable, U32 sizeLog, const U32 consumed,
const U32* rankValOrigin, const int minWeight,
const sortedSymbol_t* sortedSymbols, const U32 sortedListSize,
U32 nbBitsBaseline, U16 baseSeq)
{
HUF_DEltX4 DElt;
U32 rankVal[HUF_ABSOLUTEMAX_TABLELOG + 1];
U32 s;
/* get pre-calculated rankVal */
memcpy(rankVal, rankValOrigin, sizeof(rankVal));
/* fill skipped values */
if (minWeight>1)
{
U32 i, skipSize = rankVal[minWeight];
MEM_writeLE16(&(DElt.sequence), baseSeq);
DElt.nbBits = (BYTE)(consumed);
DElt.length = 1;
for (i = 0; i < skipSize; i++)
DTable[i] = DElt;
}
/* fill DTable */
for (s=0; s<sortedListSize; s++) /* note : sortedSymbols already skipped */
{
const U32 symbol = sortedSymbols[s].symbol;
const U32 weight = sortedSymbols[s].weight;
const U32 nbBits = nbBitsBaseline - weight;
const U32 length = 1 << (sizeLog-nbBits);
const U32 start = rankVal[weight];
U32 i = start;
const U32 end = start + length;
MEM_writeLE16(&(DElt.sequence), (U16)(baseSeq + (symbol << 8)));
DElt.nbBits = (BYTE)(nbBits + consumed);
DElt.length = 2;
do { DTable[i++] = DElt; } while (i<end); /* since length >= 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<sortedListSize; s++)
{
const U16 symbol = sortedList[s].symbol;
const U32 weight = sortedList[s].weight;
const U32 nbBits = nbBitsBaseline - weight;
const U32 start = rankVal[weight];
const U32 length = 1 << (targetLog-nbBits);
if (targetLog-nbBits >= minBits) /* enough room for a second symbol */
{
U32 sortedRank;
int minWeight = nbBits + scaleLog;
if (minWeight < 1) minWeight = 1;
sortedRank = rankStart[minWeight];
HUF_fillDTableX4Level2(DTable+start, targetLog-nbBits, nbBits,
rankValOrigin[nbBits], minWeight,
sortedList+sortedRank, sortedListSize-sortedRank,
nbBitsBaseline, symbol);
}
else
{
U32 i;
const U32 end = start + length;
HUF_DEltX4 DElt;
MEM_writeLE16(&(DElt.sequence), symbol);
DElt.nbBits = (BYTE)(nbBits);
DElt.length = 1;
for (i = start; i < end; i++)
DTable[i] = DElt;
}
rankVal[weight] += length;
}
}
static size_t HUF_readDTableX4 (U32* DTable, const void* src, size_t srcSize)
{
BYTE weightList[HUF_MAX_SYMBOL_VALUE + 1];
sortedSymbol_t sortedSymbol[HUF_MAX_SYMBOL_VALUE + 1];
U32 rankStats[HUF_ABSOLUTEMAX_TABLELOG + 1] = { 0 };
U32 rankStart0[HUF_ABSOLUTEMAX_TABLELOG + 2] = { 0 };
U32* const rankStart = rankStart0+1;
rankVal_t rankVal;
U32 tableLog, maxW, sizeOfSort, nbSymbols;
const U32 memLog = DTable[0];
size_t iSize;
void* dtPtr = DTable;
HUF_DEltX4* const dt = ((HUF_DEltX4*)dtPtr) + 1;
HUF_STATIC_ASSERT(sizeof(HUF_DEltX4) == sizeof(U32)); /* if compilation fails here, assertion is false */
if (memLog > HUF_ABSOLUTEMAX_TABLELOG) return ERROR(tableLog_tooLarge);
//memset(weightList, 0, sizeof(weightList)); /* is not necessary, even though some analyzer complain ... */
iSize = HUF_readStats(weightList, HUF_MAX_SYMBOL_VALUE + 1, rankStats, &nbSymbols, &tableLog, src, srcSize);
if (HUF_isError(iSize)) return iSize;
/* check result */
if (tableLog > memLog) return ERROR(tableLog_tooLarge); /* DTable can't fit code depth */
/* find maxWeight */
for (maxW = tableLog; rankStats[maxW]==0; maxW--)
{ 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<nbSymbols; s++)
{
U32 w = weightList[s];
U32 r = rankStart[w]++;
sortedSymbol[r].symbol = (BYTE)s;
sortedSymbol[r].weight = (BYTE)w;
}
rankStart[0] = 0; /* forget 0w symbols; this is beginning of weight(1) */
}
/* Build rankVal */
{
const U32 minBits = tableLog+1 - maxW;
U32 nextRankVal = 0;
U32 w, consumed;
const int rescale = (memLog-tableLog) - 1; /* tableLog <= memLog */
U32* rankVal0 = rankVal[0];
for (w=1; w<=maxW; w++)
{
U32 current = nextRankVal;
nextRankVal += rankStats[w] << (w+rescale);
rankVal0[w] = current;
}
for (consumed = minBits; consumed <= memLog - minBits; consumed++)
{
U32* rankValPtr = rankVal[consumed];
for (w = 1; w <= maxW; w++)
{
rankValPtr[w] = rankVal0[w] >> consumed;
}
}
}
HUF_fillDTableX4(dt, memLog,
sortedSymbol, sizeOfSort,
rankStart0, rankVal, maxW,
tableLog+1);
return iSize;
}
static U32 HUF_decodeSymbolX4(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog)
{
const size_t val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */
memcpy(op, dt+val, 2);
BIT_skipBits(DStream, dt[val].nbBits);
return dt[val].length;
}
static U32 HUF_decodeLastSymbolX4(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog)
{
const size_t val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */
memcpy(op, dt+val, 1);
if (dt[val].length==1) BIT_skipBits(DStream, dt[val].nbBits);
else
{
if (DStream->bitsConsumed < (sizeof(DStream->bitContainer)*8))
{
BIT_skipBits(DStream, dt[val].nbBits);
if (DStream->bitsConsumed > (sizeof(DStream->bitContainer)*8))
DStream->bitsConsumed = (sizeof(DStream->bitContainer)*8); /* ugly hack; works only because it's the last symbol. Note : can't easily extract nbBits from just this symbol */
}
}
return 1;
}
#define HUF_DECODE_SYMBOLX4_0(ptr, DStreamPtr) \
ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)
#define HUF_DECODE_SYMBOLX4_1(ptr, DStreamPtr) \
if (MEM_64bits() || (HUF_MAX_TABLELOG<=12)) \
ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)
#define HUF_DECODE_SYMBOLX4_2(ptr, DStreamPtr) \
if (MEM_64bits()) \
ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)
static inline size_t HUF_decodeStreamX4(BYTE* p, BIT_DStream_t* bitDPtr, BYTE* const pEnd, const HUF_DEltX4* const dt, const U32 dtLog)
{
BYTE* const pStart = p;
/* up to 8 symbols at a time */
while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p < pEnd-7))
{
HUF_DECODE_SYMBOLX4_2(p, bitDPtr);
HUF_DECODE_SYMBOLX4_1(p, bitDPtr);
HUF_DECODE_SYMBOLX4_2(p, bitDPtr);
HUF_DECODE_SYMBOLX4_0(p, bitDPtr);
}
/* closer to the end */
while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p <= pEnd-2))
HUF_DECODE_SYMBOLX4_0(p, bitDPtr);
while (p <= pEnd-2)
HUF_DECODE_SYMBOLX4_0(p, bitDPtr); /* no need to reload : reached the end of DStream */
if (p < pEnd)
p += HUF_decodeLastSymbolX4(p, bitDPtr, dt, dtLog);
return p-pStart;
}
static size_t HUF_decompress4X4_usingDTable(
void* dst, size_t dstSize,
const void* cSrc, size_t cSrcSize,
const U32* DTable)
{
if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
{
const BYTE* const istart = (const BYTE*) cSrc;
BYTE* const ostart = (BYTE*) dst;
BYTE* const oend = ostart + dstSize;
const void* const dtPtr = DTable;
const HUF_DEltX4* const dt = ((const HUF_DEltX4*)dtPtr) +1;
const U32 dtLog = DTable[0];
size_t errorCode;
/* Init */
BIT_DStream_t bitD1;
BIT_DStream_t bitD2;
BIT_DStream_t bitD3;
BIT_DStream_t bitD4;
const size_t length1 = MEM_readLE16(istart);
const size_t length2 = MEM_readLE16(istart+2);
const size_t length3 = MEM_readLE16(istart+4);
size_t length4;
const BYTE* const istart1 = istart + 6; /* jumpTable */
const BYTE* const istart2 = istart1 + length1;
const BYTE* const istart3 = istart2 + length2;
const BYTE* const istart4 = istart3 + length3;
const size_t segmentSize = (dstSize+3) / 4;
BYTE* const opStart2 = ostart + segmentSize;
BYTE* const opStart3 = opStart2 + segmentSize;
BYTE* const opStart4 = opStart3 + segmentSize;
BYTE* op1 = ostart;
BYTE* op2 = opStart2;
BYTE* op3 = opStart3;
BYTE* op4 = opStart4;
U32 endSignal;
length4 = cSrcSize - (length1 + length2 + length3 + 6);
if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */
errorCode = BIT_initDStream(&bitD1, istart1, length1);
if (HUF_isError(errorCode)) return errorCode;
errorCode = BIT_initDStream(&bitD2, istart2, length2);
if (HUF_isError(errorCode)) return errorCode;
errorCode = BIT_initDStream(&bitD3, istart3, length3);
if (HUF_isError(errorCode)) return errorCode;
errorCode = BIT_initDStream(&bitD4, istart4, length4);
if (HUF_isError(errorCode)) return errorCode;
/* 16-32 symbols per loop (4-8 symbols per stream) */
endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
for ( ; (endSignal==BIT_DStream_unfinished) && (op4<(oend-7)) ; )
{
HUF_DECODE_SYMBOLX4_2(op1, &bitD1);
HUF_DECODE_SYMBOLX4_2(op2, &bitD2);
HUF_DECODE_SYMBOLX4_2(op3, &bitD3);
HUF_DECODE_SYMBOLX4_2(op4, &bitD4);
HUF_DECODE_SYMBOLX4_1(op1, &bitD1);
HUF_DECODE_SYMBOLX4_1(op2, &bitD2);
HUF_DECODE_SYMBOLX4_1(op3, &bitD3);
HUF_DECODE_SYMBOLX4_1(op4, &bitD4);
HUF_DECODE_SYMBOLX4_2(op1, &bitD1);
HUF_DECODE_SYMBOLX4_2(op2, &bitD2);
HUF_DECODE_SYMBOLX4_2(op3, &bitD3);
HUF_DECODE_SYMBOLX4_2(op4, &bitD4);
HUF_DECODE_SYMBOLX4_0(op1, &bitD1);
HUF_DECODE_SYMBOLX4_0(op2, &bitD2);
HUF_DECODE_SYMBOLX4_0(op3, &bitD3);
HUF_DECODE_SYMBOLX4_0(op4, &bitD4);
endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
}
/* check corruption */
if (op1 > opStart2) return ERROR(corruption_detected);
if (op2 > opStart3) return ERROR(corruption_detected);
if (op3 > opStart4) return ERROR(corruption_detected);
/* note : op4 supposed already verified within main loop */
/* finish bitStreams one by one */
HUF_decodeStreamX4(op1, &bitD1, opStart2, dt, dtLog);
HUF_decodeStreamX4(op2, &bitD2, opStart3, dt, dtLog);
HUF_decodeStreamX4(op3, &bitD3, opStart4, dt, dtLog);
HUF_decodeStreamX4(op4, &bitD4, oend, dt, dtLog);
/* check */
endSignal = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
if (!endSignal) return ERROR(corruption_detected);
/* decoded size */
return dstSize;
}
}
static size_t HUF_decompress4X4 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)
{
HUF_CREATE_STATIC_DTABLEX4(DTable, HUF_MAX_TABLELOG);
const BYTE* ip = (const BYTE*) cSrc;
size_t hSize = HUF_readDTableX4 (DTable, cSrc, cSrcSize);
if (HUF_isError(hSize)) return hSize;
if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
ip += hSize;
cSrcSize -= hSize;
return HUF_decompress4X4_usingDTable (dst, dstSize, ip, cSrcSize, DTable);
}
/**********************************/
/* 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 */
}
#endif /* ZSTD_CCOMMON_H_MODULE */
/*
zstd - decompression module fo v0.4 legacy format
Copyright (C) 2015-2016, 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
*****************************************************************/
/*!
* HEAPMODE :
* Select how default decompression function ZSTD_decompress() will allocate memory,
* in memory stack (0), or in memory heap (1, requires malloc())
*/
#ifndef ZSTD_HEAPMODE
# define ZSTD_HEAPMODE 1
#endif
/* *******************************************************
* Includes
*********************************************************/
#include <stdlib.h> /* calloc */
#include <string.h> /* memcpy, memmove */
#include <stdio.h> /* debug : printf */
/* *******************************************************
* Compiler specifics
*********************************************************/
#ifdef _MSC_VER /* Visual Studio */
# include <intrin.h> /* For Visual 2005 */
# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */
# pragma warning(disable : 4324) /* disable: C4324: padded structure */
#endif
/* *************************************
* Local types
***************************************/
typedef struct
{
blockType_t blockType;
U32 origSize;
} blockProperties_t;
/* *******************************************************
* Memory operations
**********************************************************/
static void ZSTD_copy4(void* dst, const void* src) { memcpy(dst, src, 4); }
/* *************************************
* Error Management
***************************************/
/*! ZSTD_isError
* tells if a return value is an error code */
static unsigned ZSTD_isError(size_t code) { return ERR_isError(code); }
/* *************************************************************
* Context management
***************************************************************/
typedef enum { ZSTDds_getFrameHeaderSize, ZSTDds_decodeFrameHeader,
ZSTDds_decodeBlockHeader, ZSTDds_decompressBlock } ZSTD_dStage;
struct ZSTDv04_Dctx_s
{
U32 LLTable[FSE_DTABLE_SIZE_U32(LLFSELog)];
U32 OffTable[FSE_DTABLE_SIZE_U32(OffFSELog)];
U32 MLTable[FSE_DTABLE_SIZE_U32(MLFSELog)];
const void* previousDstEnd;
const void* base;
const void* vBase;
const void* dictEnd;
size_t expected;
size_t headerSize;
ZSTD_parameters params;
blockType_t bType;
ZSTD_dStage stage;
const BYTE* litPtr;
size_t litSize;
BYTE litBuffer[BLOCKSIZE + 8 /* margin for wildcopy */];
BYTE headerBuffer[ZSTD_frameHeaderSize_max];
}; /* typedef'd to ZSTD_DCtx within "zstd_static.h" */
static size_t ZSTD_resetDCtx(ZSTD_DCtx* dctx)
{
dctx->expected = ZSTD_frameHeaderSize_min;
dctx->stage = ZSTDds_getFrameHeaderSize;
dctx->previousDstEnd = NULL;
dctx->base = NULL;
dctx->vBase = NULL;
dctx->dictEnd = 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;
}
/* *************************************************************
* Decompression section
***************************************************************/
/** ZSTD_decodeFrameHeader_Part1
* decode the 1st part of the Frame Header, which tells Frame Header size.
* srcSize must be == ZSTD_frameHeaderSize_min
* @return : the full size of the Frame Header */
static size_t ZSTD_decodeFrameHeader_Part1(ZSTD_DCtx* zc, const void* src, size_t srcSize)
{
U32 magicNumber;
if (srcSize != ZSTD_frameHeaderSize_min) return ERROR(srcSize_wrong);
magicNumber = MEM_readLE32(src);
if (magicNumber != ZSTD_MAGICNUMBER) return ERROR(prefix_unknown);
zc->headerSize = ZSTD_frameHeaderSize_min;
return zc->headerSize;
}
static size_t ZSTD_getFrameParams(ZSTD_parameters* params, const void* src, size_t srcSize)
{
U32 magicNumber;
if (srcSize < ZSTD_frameHeaderSize_min) return ZSTD_frameHeaderSize_max;
magicNumber = MEM_readLE32(src);
if (magicNumber != ZSTD_MAGICNUMBER) return ERROR(prefix_unknown);
memset(params, 0, sizeof(*params));
params->windowLog = (((const BYTE*)src)[4] & 15) + ZSTD_WINDOWLOG_ABSOLUTEMIN;
if ((((const BYTE*)src)[4] >> 4) != 0) return ERROR(frameParameter_unsupported); /* reserved bits */
return 0;
}
/** ZSTD_decodeFrameHeader_Part2
* decode the full Frame Header
* srcSize must be the size provided by ZSTD_decodeFrameHeader_Part1
* @return : 0, or an error code, which can be tested using ZSTD_isError() */
static size_t ZSTD_decodeFrameHeader_Part2(ZSTD_DCtx* zc, const void* src, size_t srcSize)
{
size_t result;
if (srcSize != zc->headerSize) return ERROR(srcSize_wrong);
result = ZSTD_getFrameParams(&(zc->params), src, srcSize);
if ((MEM_32bits()) && (zc->params.windowLog > 25)) return ERROR(frameParameter_unsupported);
return result;
}
static size_t ZSTD_getcBlockSize(const void* src, size_t srcSize, blockProperties_t* bpPtr)
{
const BYTE* const in = (const BYTE* const)src;
BYTE headerFlags;
U32 cSize;
if (srcSize < 3) return ERROR(srcSize_wrong);
headerFlags = *in;
cSize = in[2] + (in[1]<<8) + ((in[0] & 7)<<16);
bpPtr->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_copyRawBlock(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(ZSTD_DCtx* dctx,
const void* src, size_t srcSize) /* note : srcSize < BLOCKSIZE */
{
const BYTE* const istart = (const BYTE*) 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)
{
/* compressed */
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;
}
default:
return ERROR(corruption_detected); /* forbidden nominal case */
}
}
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 >= 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;
if (litLength == MaxLL) {
const U32 add = dumps<de ? *dumps++ : 0;
if (add < 255) litLength += add;
else if (dumps + 3 <= de) {
litLength = MEM_readLE24(dumps);
dumps += 3;
}
if (dumps >= de) { dumps = de-1; } /* late correction, to avoid read overflow (data is now corrupted anyway) */
}
/* Offset */
{ static const U32 offsetPrefix[MaxOff+1] = {
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 */
if (offsetCode | !litLength) seqState->prevOffset = seq->offset; /* cmove */
}
/* MatchLength */
matchLength = FSE_decodeSymbol(&(seqState->stateML), &(seqState->DStream));
if (matchLength == MaxML) {
const U32 add = dumps<de ? *dumps++ : 0;
if (add < 255) matchLength += add;
else if (dumps + 3 <= de){
matchLength = MEM_readLE24(dumps);
dumps += 3;
}
if (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,
BYTE* const oend, seq_t sequence,
const BYTE** litPtr, const BYTE* const litLimit,
const BYTE* const base, const BYTE* const vBase, const BYTE* const dictEnd)
{
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 */
BYTE* const oLitEnd = op + sequence.litLength;
const size_t sequenceLength = sequence.litLength + sequence.matchLength;
BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
BYTE* const oend_8 = oend-8;
const BYTE* const litEnd = *litPtr + sequence.litLength;
const BYTE* match = oLitEnd - sequence.offset;
/* 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 (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 */
if (sequence.offset > (size_t)(oLitEnd - base))
{
/* offset beyond prefix */
if (sequence.offset > (size_t)(oLitEnd - vBase))
return ERROR(corruption_detected);
match = dictEnd - (base-match);
if (match + sequence.matchLength <= dictEnd)
{
memmove(oLitEnd, match, sequence.matchLength);
return sequenceLength;
}
/* span extDict & currentPrefixSegment */
{
size_t length1 = dictEnd - match;
memmove(oLitEnd, match, length1);
op = oLitEnd + length1;
sequence.matchLength -= length1;
match = base;
if (op > oend_8 || sequence.matchLength < MINMATCH) {
while (op < oMatchEnd) *op++ = *match++;
return sequenceLength;
}
}
}
/* Requirement: op <= oend_8 */
/* match within prefix */
if (sequence.offset < 8) {
/* close range match, overlap */
const int sub2 = 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 -= sub2;
} 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, but must be signed */
}
return sequenceLength;
}
static size_t ZSTD_decompressSequences(
ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize)
{
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;
const BYTE* const base = (const BYTE*) (dctx->base);
const BYTE* const vBase = (const BYTE*) (dctx->vBase);
const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
/* 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));
sequence.offset = 4;
seqState.dumps = dumps;
seqState.dumpsEnd = dumps + dumpsLength;
seqState.prevOffset = 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 ; )
{
size_t oneSeqSize;
nbSeq--;
ZSTD_decodeSequence(&sequence, &seqState);
oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, base, vBase, dictEnd);
if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
op += oneSeqSize;
}
/* check if reached exact end */
if ( !BIT_endOfDStream(&(seqState.DStream)) ) return ERROR(corruption_detected); /* DStream should be entirely and exactly consumed; otherwise 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) memcpy(op, litPtr, lastLLSize);
op += lastLLSize;
}
}
}
return op-ostart;
}
static void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst)
{
if (dst != dctx->previousDstEnd) /* not contiguous */
{
dctx->dictEnd = dctx->previousDstEnd;
dctx->vBase = (const char*)dst - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->base));
dctx->base = dst;
dctx->previousDstEnd = dst;
}
}
static size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* src, size_t srcSize)
{
/* blockType == blockCompressed */
const BYTE* ip = (const BYTE*)src;
size_t litCSize;
if (srcSize > BLOCKSIZE) return ERROR(corruption_detected);
/* Decode literals sub-block */
litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize);
if (ZSTD_isError(litCSize)) return litCSize;
ip += litCSize;
srcSize -= litCSize;
return ZSTD_decompressSequences(dctx, dst, maxDstSize, ip, srcSize);
}
static size_t ZSTD_decompress_usingDict(ZSTD_DCtx* ctx,
void* dst, size_t maxDstSize,
const void* src, size_t srcSize,
const void* dict, size_t dictSize)
{
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;
blockProperties_t blockProperties;
/* init */
ZSTD_resetDCtx(ctx);
if (dict)
{
ZSTD_decompress_insertDictionary(ctx, dict, dictSize);
ctx->dictEnd = ctx->previousDstEnd;
ctx->vBase = (const char*)dst - ((const char*)(ctx->previousDstEnd) - (const char*)(ctx->base));
ctx->base = dst;
}
else
{
ctx->vBase = ctx->base = ctx->dictEnd = dst;
}
/* Frame Header */
{
size_t frameHeaderSize;
if (srcSize < ZSTD_frameHeaderSize_min+ZSTD_blockHeaderSize) return ERROR(srcSize_wrong);
frameHeaderSize = ZSTD_decodeFrameHeader_Part1(ctx, src, ZSTD_frameHeaderSize_min);
if (ZSTD_isError(frameHeaderSize)) return frameHeaderSize;
if (srcSize < frameHeaderSize+ZSTD_blockHeaderSize) return ERROR(srcSize_wrong);
ip += frameHeaderSize; remainingSize -= frameHeaderSize;
frameHeaderSize = ZSTD_decodeFrameHeader_Part2(ctx, src, frameHeaderSize);
if (ZSTD_isError(frameHeaderSize)) return 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_internal(ctx, op, oend-op, ip, cBlockSize);
break;
case bt_raw :
decodedSize = ZSTD_copyRawBlock(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;
}
/* ZSTD_errorFrameSizeInfoLegacy() :
assumes `cSize` and `dBound` are _not_ NULL */
static void ZSTD_errorFrameSizeInfoLegacy(size_t* cSize, unsigned long long* dBound, size_t ret)
{
*cSize = ret;
*dBound = ZSTD_CONTENTSIZE_ERROR;
}
void ZSTDv04_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;
blockProperties_t blockProperties;
/* Frame Header */
if (srcSize < ZSTD_frameHeaderSize_min) {
ZSTD_errorFrameSizeInfoLegacy(cSize, dBound, ERROR(srcSize_wrong));
return;
}
if (MEM_readLE32(src) != ZSTD_MAGICNUMBER) {
ZSTD_errorFrameSizeInfoLegacy(cSize, dBound, ERROR(prefix_unknown));
return;
}
ip += ZSTD_frameHeaderSize_min; remainingSize -= ZSTD_frameHeaderSize_min;
/* 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_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);
ZSTD_checkContinuity(ctx, dst);
/* Decompress : frame header; part 1 */
switch (ctx->stage)
{
case ZSTDds_getFrameHeaderSize :
/* get frame header size */
if (srcSize != ZSTD_frameHeaderSize_min) return ERROR(srcSize_wrong); /* impossible */
ctx->headerSize = ZSTD_decodeFrameHeader_Part1(ctx, src, ZSTD_frameHeaderSize_min);
if (ZSTD_isError(ctx->headerSize)) return ctx->headerSize;
memcpy(ctx->headerBuffer, src, ZSTD_frameHeaderSize_min);
if (ctx->headerSize > ZSTD_frameHeaderSize_min) return ERROR(GENERIC); /* impossible */
ctx->expected = 0; /* not necessary to copy more */
/* fallthrough */
case ZSTDds_decodeFrameHeader:
/* get frame header */
{ size_t const result = ZSTD_decodeFrameHeader_Part2(ctx, ctx->headerBuffer, ctx->headerSize);
if (ZSTD_isError(result)) return result;
ctx->expected = ZSTD_blockHeaderSize;
ctx->stage = ZSTDds_decodeBlockHeader;
return 0;
}
case ZSTDds_decodeBlockHeader:
/* Decode block header */
{ blockProperties_t bp;
size_t const blockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp);
if (ZSTD_isError(blockSize)) return blockSize;
if (bp.blockType == bt_end)
{
ctx->expected = 0;
ctx->stage = ZSTDds_getFrameHeaderSize;
}
else
{
ctx->expected = blockSize;
ctx->bType = bp.blockType;
ctx->stage = ZSTDds_decompressBlock;
}
return 0;
}
case ZSTDds_decompressBlock:
{
/* Decompress : block content */
size_t rSize;
switch(ctx->bType)
{
case bt_compressed:
rSize = ZSTD_decompressBlock_internal(ctx, dst, maxDstSize, src, srcSize);
break;
case bt_raw :
rSize = ZSTD_copyRawBlock(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->stage = ZSTDds_decodeBlockHeader;
ctx->expected = ZSTD_blockHeaderSize;
if (ZSTD_isError(rSize)) return rSize;
ctx->previousDstEnd = (char*)dst + rSize;
return rSize;
}
default:
return ERROR(GENERIC); /* impossible */
}
}
static void ZSTD_decompress_insertDictionary(ZSTD_DCtx* ctx, const void* dict, size_t dictSize)
{
ctx->dictEnd = ctx->previousDstEnd;
ctx->vBase = (const char*)dict - ((const char*)(ctx->previousDstEnd) - (const char*)(ctx->base));
ctx->base = dict;
ctx->previousDstEnd = (const char*)dict + dictSize;
}
/*
Buffered version of Zstd compression library
Copyright (C) 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.
*/
/* *************************************
* Includes
***************************************/
#include <stdlib.h>
/** ************************************************
* Streaming decompression
*
* A ZBUFF_DCtx object is required to track streaming operation.
* Use ZBUFF_createDCtx() and ZBUFF_freeDCtx() to create/release resources.
* Use ZBUFF_decompressInit() to start a new decompression operation.
* ZBUFF_DCtx objects can be reused multiple times.
*
* Use ZBUFF_decompressContinue() repetitively to consume your input.
* *srcSizePtr and *maxDstSizePtr can be any size.
* The function will report how many bytes were read or written by modifying *srcSizePtr and *maxDstSizePtr.
* Note that it may not consume the entire input, in which case it's up to the caller to call again the function with remaining input.
* The content of dst will be overwritten (up to *maxDstSizePtr) at each function call, so save its content if it matters or change dst .
* return : a hint to preferred nb of bytes to use as input for next function call (it's only a hint, to improve latency)
* or 0 when a frame is completely decoded
* or an error code, which can be tested using ZBUFF_isError().
*
* Hint : recommended buffer sizes (not compulsory)
* output : 128 KB block size is the internal unit, it ensures it's always possible to write a full block when it's decoded.
* input : just follow indications from ZBUFF_decompressContinue() to minimize latency. It should always be <= 128 KB + 3 .
* **************************************************/
typedef enum { ZBUFFds_init, ZBUFFds_readHeader, ZBUFFds_loadHeader, ZBUFFds_decodeHeader,
ZBUFFds_read, ZBUFFds_load, ZBUFFds_flush } ZBUFF_dStage;
/* *** Resource management *** */
#define ZSTD_frameHeaderSize_max 5 /* too magical, should come from reference */
struct ZBUFFv04_DCtx_s {
ZSTD_DCtx* zc;
ZSTD_parameters params;
char* inBuff;
size_t inBuffSize;
size_t inPos;
char* outBuff;
size_t outBuffSize;
size_t outStart;
size_t outEnd;
size_t hPos;
const char* dict;
size_t dictSize;
ZBUFF_dStage stage;
unsigned char headerBuffer[ZSTD_frameHeaderSize_max];
}; /* typedef'd to ZBUFF_DCtx within "zstd_buffered.h" */
typedef ZBUFFv04_DCtx ZBUFF_DCtx;
static ZBUFF_DCtx* ZBUFF_createDCtx(void)
{
ZBUFF_DCtx* zbc = (ZBUFF_DCtx*)malloc(sizeof(ZBUFF_DCtx));
if (zbc==NULL) return NULL;
memset(zbc, 0, sizeof(*zbc));
zbc->zc = ZSTD_createDCtx();
zbc->stage = ZBUFFds_init;
return zbc;
}
static size_t ZBUFF_freeDCtx(ZBUFF_DCtx* zbc)
{
if (zbc==NULL) return 0; /* support free on null */
ZSTD_freeDCtx(zbc->zc);
free(zbc->inBuff);
free(zbc->outBuff);
free(zbc);
return 0;
}
/* *** Initialization *** */
static size_t ZBUFF_decompressInit(ZBUFF_DCtx* zbc)
{
zbc->stage = ZBUFFds_readHeader;
zbc->hPos = zbc->inPos = zbc->outStart = zbc->outEnd = zbc->dictSize = 0;
return ZSTD_resetDCtx(zbc->zc);
}
static size_t ZBUFF_decompressWithDictionary(ZBUFF_DCtx* zbc, const void* src, size_t srcSize)
{
zbc->dict = (const char*)src;
zbc->dictSize = srcSize;
return 0;
}
static size_t ZBUFF_limitCopy(void* dst, size_t maxDstSize, const void* src, size_t srcSize)
{
size_t length = MIN(maxDstSize, srcSize);
if (length > 0) {
memcpy(dst, src, length);
}
return length;
}
/* *** Decompression *** */
static size_t ZBUFF_decompressContinue(ZBUFF_DCtx* zbc, void* dst, size_t* maxDstSizePtr, const void* src, size_t* srcSizePtr)
{
const char* const istart = (const char*)src;
const char* ip = istart;
const char* const iend = istart + *srcSizePtr;
char* const ostart = (char*)dst;
char* op = ostart;
char* const oend = ostart + *maxDstSizePtr;
U32 notDone = 1;
DEBUGLOG(5, "ZBUFF_decompressContinue");
while (notDone)
{
switch(zbc->stage)
{
case ZBUFFds_init :
DEBUGLOG(5, "ZBUFF_decompressContinue: stage==ZBUFFds_init => ERROR(init_missing)");
return ERROR(init_missing);
case ZBUFFds_readHeader :
/* read header from src */
{ size_t const headerSize = ZSTD_getFrameParams(&(zbc->params), src, *srcSizePtr);
if (ZSTD_isError(headerSize)) return headerSize;
if (headerSize) {
/* not enough input to decode header : tell how many bytes would be necessary */
memcpy(zbc->headerBuffer+zbc->hPos, src, *srcSizePtr);
zbc->hPos += *srcSizePtr;
*maxDstSizePtr = 0;
zbc->stage = ZBUFFds_loadHeader;
return headerSize - zbc->hPos;
}
zbc->stage = ZBUFFds_decodeHeader;
break;
}
case ZBUFFds_loadHeader:
/* complete header from src */
{ size_t headerSize = ZBUFF_limitCopy(
zbc->headerBuffer + zbc->hPos, ZSTD_frameHeaderSize_max - zbc->hPos,
src, *srcSizePtr);
zbc->hPos += headerSize;
ip += headerSize;
headerSize = ZSTD_getFrameParams(&(zbc->params), zbc->headerBuffer, zbc->hPos);
if (ZSTD_isError(headerSize)) return headerSize;
if (headerSize) {
/* not enough input to decode header : tell how many bytes would be necessary */
*maxDstSizePtr = 0;
return headerSize - zbc->hPos;
} }
/* intentional fallthrough */
case ZBUFFds_decodeHeader:
/* apply header to create / resize buffers */
{ size_t const neededOutSize = (size_t)1 << zbc->params.windowLog;
size_t const neededInSize = BLOCKSIZE; /* a block is never > BLOCKSIZE */
if (zbc->inBuffSize < neededInSize) {
free(zbc->inBuff);
zbc->inBuffSize = neededInSize;
zbc->inBuff = (char*)malloc(neededInSize);
if (zbc->inBuff == NULL) return ERROR(memory_allocation);
}
if (zbc->outBuffSize < neededOutSize) {
free(zbc->outBuff);
zbc->outBuffSize = neededOutSize;
zbc->outBuff = (char*)malloc(neededOutSize);
if (zbc->outBuff == NULL) return ERROR(memory_allocation);
} }
if (zbc->dictSize)
ZSTD_decompress_insertDictionary(zbc->zc, zbc->dict, zbc->dictSize);
if (zbc->hPos) {
/* some data already loaded into headerBuffer : transfer into inBuff */
memcpy(zbc->inBuff, zbc->headerBuffer, zbc->hPos);
zbc->inPos = zbc->hPos;
zbc->hPos = 0;
zbc->stage = ZBUFFds_load;
break;
}
zbc->stage = ZBUFFds_read;
/* fall-through */
case ZBUFFds_read:
{
size_t neededInSize = ZSTD_nextSrcSizeToDecompress(zbc->zc);
if (neededInSize==0) /* end of frame */
{
zbc->stage = ZBUFFds_init;
notDone = 0;
break;
}
if ((size_t)(iend-ip) >= neededInSize)
{
/* directly decode from src */
size_t decodedSize = ZSTD_decompressContinue(zbc->zc,
zbc->outBuff + zbc->outStart, zbc->outBuffSize - zbc->outStart,
ip, neededInSize);
if (ZSTD_isError(decodedSize)) return decodedSize;
ip += neededInSize;
if (!decodedSize) break; /* this was just a header */
zbc->outEnd = zbc->outStart + decodedSize;
zbc->stage = ZBUFFds_flush;
break;
}
if (ip==iend) { notDone = 0; break; } /* no more input */
zbc->stage = ZBUFFds_load;
}
/* fall-through */
case ZBUFFds_load:
{
size_t neededInSize = ZSTD_nextSrcSizeToDecompress(zbc->zc);
size_t toLoad = neededInSize - zbc->inPos; /* should always be <= remaining space within inBuff */
size_t loadedSize;
if (toLoad > zbc->inBuffSize - zbc->inPos) return ERROR(corruption_detected); /* should never happen */
loadedSize = ZBUFF_limitCopy(zbc->inBuff + zbc->inPos, toLoad, ip, iend-ip);
ip += loadedSize;
zbc->inPos += loadedSize;
if (loadedSize < toLoad) { notDone = 0; break; } /* not enough input, wait for more */
{
size_t decodedSize = ZSTD_decompressContinue(zbc->zc,
zbc->outBuff + zbc->outStart, zbc->outBuffSize - zbc->outStart,
zbc->inBuff, neededInSize);
if (ZSTD_isError(decodedSize)) return decodedSize;
zbc->inPos = 0; /* input is consumed */
if (!decodedSize) { zbc->stage = ZBUFFds_read; break; } /* this was just a header */
zbc->outEnd = zbc->outStart + decodedSize;
zbc->stage = ZBUFFds_flush;
/* ZBUFFds_flush follows */
}
}
/* fall-through */
case ZBUFFds_flush:
{
size_t toFlushSize = zbc->outEnd - zbc->outStart;
size_t flushedSize = ZBUFF_limitCopy(op, oend-op, zbc->outBuff + zbc->outStart, toFlushSize);
op += flushedSize;
zbc->outStart += flushedSize;
if (flushedSize == toFlushSize)
{
zbc->stage = ZBUFFds_read;
if (zbc->outStart + BLOCKSIZE > zbc->outBuffSize)
zbc->outStart = zbc->outEnd = 0;
break;
}
/* cannot flush everything */
notDone = 0;
break;
}
default: return ERROR(GENERIC); /* impossible */
}
}
*srcSizePtr = ip-istart;
*maxDstSizePtr = op-ostart;
{
size_t nextSrcSizeHint = ZSTD_nextSrcSizeToDecompress(zbc->zc);
if (nextSrcSizeHint > 3) nextSrcSizeHint+= 3; /* get the next block header while at it */
nextSrcSizeHint -= zbc->inPos; /* already loaded*/
return nextSrcSizeHint;
}
}
/* *************************************
* Tool functions
***************************************/
unsigned ZBUFFv04_isError(size_t errorCode) { return ERR_isError(errorCode); }
const char* ZBUFFv04_getErrorName(size_t errorCode) { return ERR_getErrorName(errorCode); }
size_t ZBUFFv04_recommendedDInSize(void) { return BLOCKSIZE + 3; }
size_t ZBUFFv04_recommendedDOutSize(void) { return BLOCKSIZE; }
/*- ========================================================================= -*/
/* final wrapping stage */
size_t ZSTDv04_decompressDCtx(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize, const void* src, size_t srcSize)
{
return ZSTD_decompress_usingDict(dctx, dst, maxDstSize, src, srcSize, NULL, 0);
}
size_t ZSTDv04_decompress(void* dst, size_t maxDstSize, const void* src, size_t srcSize)
{
#if defined(ZSTD_HEAPMODE) && (ZSTD_HEAPMODE==1)
size_t regenSize;
ZSTD_DCtx* dctx = ZSTD_createDCtx();
if (dctx==NULL) return ERROR(memory_allocation);
regenSize = ZSTDv04_decompressDCtx(dctx, dst, maxDstSize, src, srcSize);
ZSTD_freeDCtx(dctx);
return regenSize;
#else
ZSTD_DCtx dctx;
return ZSTDv04_decompressDCtx(&dctx, dst, maxDstSize, src, srcSize);
#endif
}
size_t ZSTDv04_resetDCtx(ZSTDv04_Dctx* dctx) { return ZSTD_resetDCtx(dctx); }
size_t ZSTDv04_nextSrcSizeToDecompress(ZSTDv04_Dctx* dctx)
{
return ZSTD_nextSrcSizeToDecompress(dctx);
}
size_t ZSTDv04_decompressContinue(ZSTDv04_Dctx* dctx, void* dst, size_t maxDstSize, const void* src, size_t srcSize)
{
return ZSTD_decompressContinue(dctx, dst, maxDstSize, src, srcSize);
}
ZBUFFv04_DCtx* ZBUFFv04_createDCtx(void) { return ZBUFF_createDCtx(); }
size_t ZBUFFv04_freeDCtx(ZBUFFv04_DCtx* dctx) { return ZBUFF_freeDCtx(dctx); }
size_t ZBUFFv04_decompressInit(ZBUFFv04_DCtx* dctx) { return ZBUFF_decompressInit(dctx); }
size_t ZBUFFv04_decompressWithDictionary(ZBUFFv04_DCtx* dctx, const void* src, size_t srcSize)
{ return ZBUFF_decompressWithDictionary(dctx, src, srcSize); }
size_t ZBUFFv04_decompressContinue(ZBUFFv04_DCtx* dctx, void* dst, size_t* maxDstSizePtr, const void* src, size_t* srcSizePtr)
{
DEBUGLOG(5, "ZBUFFv04_decompressContinue");
return ZBUFF_decompressContinue(dctx, dst, maxDstSizePtr, src, srcSizePtr);
}
ZSTD_DCtx* ZSTDv04_createDCtx(void) { return ZSTD_createDCtx(); }
size_t ZSTDv04_freeDCtx(ZSTD_DCtx* dctx) { return ZSTD_freeDCtx(dctx); }
/* Implementation moved to Rust (rust/src/legacy/zstd_v04.rs).
* The frozen v0.4 decoder, including its embedded FSE/Huff0 snapshot, frame
* state, and buffered streaming context, now lives entirely in Rust; C code
* only ever holds opaque ZSTDv04_Dctx and ZBUFFv04_DCtx pointers. */
rust/src/legacy/mod.rs
+3
View File
@@ -44,3 +44,6 @@ pub mod zstd_v02;
#[cfg(feature = "legacy-v03")]
pub mod zstd_v03;
#[cfg(feature = "legacy-v04")]
pub mod zstd_v04;
rust/src/legacy/zstd_v04.rs
+2897
View File
@@ -0,0 +1,2897 @@
#![allow(non_snake_case)]
//! Frozen decoder for the zstd v0.4 format.
//!
//! `lib/legacy/zstd_v04.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 `ZSTDv04_Dctx`.
use crate::errors::{ERR_isError, ZstdErrorCode, ERROR};
use std::os::raw::{c_char, c_uint, c_void};
use std::ptr;
const ZSTD_MAGIC_NUMBER: u32 = 0xFD2F_B524;
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 IS_RAW: u8 = 1;
const IS_RLE: u8 = 2;
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 write_le16(dst: *mut u8, value: u16) {
let bytes = value.to_le_bytes();
ptr::copy_nonoverlapping(bytes.as_ptr(), dst, 2);
}
#[inline]
unsafe fn zstd_copy8(dst: *mut u8, src: *const u8) {
ptr::copy(src, dst, 8);
}
#[inline]
unsafe fn zstd_copy4(dst: *mut u8, src: *const u8) {
ptr::copy(src, dst, 4);
}
unsafe fn zstd_wildcopy(dst: *mut u8, src: *const u8, length: isize) {
let mut op = dst;
let mut ip = src;
let end = if length >= 0 {
(dst as usize).wrapping_add(length as usize)
} else {
(dst as usize).wrapping_sub(length.wrapping_neg() as usize)
};
loop {
zstd_copy8(op, ip);
op = op.add(8);
ip = ip.add(8);
if (op as usize) >= end {
break;
}
}
}
#[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::<usize>() == 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 v0.4 snapshot)
********************************************/
unsafe fn init_dstream(stream: &mut DStream, src: *const u8, src_size: usize) -> usize {
let word = std::mem::size_of::<usize>();
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::<usize>();
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 v0.4 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
}
#[repr(C)]
#[derive(Clone, Copy)]
struct HufDEltX4 {
sequence: u16,
nb_bits: u8,
length: u8,
}
#[repr(C)]
#[derive(Clone, Copy)]
struct SortedSymbol {
symbol: u8,
weight: u8,
}
type RankVal = [[u32; HUF_ABSOLUTE_MAX_TABLELOG + 1]; HUF_ABSOLUTE_MAX_TABLELOG];
#[inline]
unsafe fn huf_set_dtable_x4(
dtable: *mut HufDEltX4,
index: usize,
sequence: u16,
nb_bits: u32,
length: u8,
) {
let entry = dtable.add(index);
write_le16(entry.cast::<u8>(), sequence);
(*entry).nb_bits = nb_bits as u8;
(*entry).length = length;
}
#[allow(clippy::too_many_arguments)]
unsafe fn huf_fill_dtable_x4_level2(
dtable: *mut HufDEltX4,
size_log: u32,
consumed: u32,
rank_val_origin: &[u32; HUF_ABSOLUTE_MAX_TABLELOG + 1],
min_weight: usize,
sorted_symbols: &[SortedSymbol; HUF_MAX_SYMBOL_VALUE + 1],
sorted_start: usize,
sorted_list_size: usize,
nb_bits_baseline: u32,
base_seq: u16,
) {
let mut rank_val = *rank_val_origin;
if min_weight > 1 {
let skip_size = rank_val[min_weight] as usize;
for index in 0..skip_size {
huf_set_dtable_x4(dtable, index, base_seq, consumed, 1);
}
}
for item in 0..sorted_list_size {
let sorted = sorted_symbols[sorted_start + item];
let symbol = sorted.symbol as u32;
let weight = sorted.weight as usize;
let nb_bits = nb_bits_baseline - weight as u32;
let length = 1u32 << (size_log - nb_bits);
let start = rank_val[weight] as usize;
let end = start + length as usize;
let sequence = base_seq.wrapping_add((symbol << 8) as u16);
for index in start..end {
huf_set_dtable_x4(dtable, index, sequence, nb_bits + consumed, 2);
}
rank_val[weight] = rank_val[weight].wrapping_add(length);
}
}
#[allow(clippy::too_many_arguments)]
unsafe fn huf_fill_dtable_x4(
dtable: *mut HufDEltX4,
target_log: u32,
sorted_list: &[SortedSymbol; HUF_MAX_SYMBOL_VALUE + 1],
sorted_list_size: usize,
rank_start0: &[u32; HUF_ABSOLUTE_MAX_TABLELOG + 2],
rank_val_origin: &RankVal,
max_weight: u32,
nb_bits_baseline: u32,
) {
let mut rank_val = rank_val_origin[0];
let scale_log = nb_bits_baseline as i32 - target_log as i32;
let min_bits = nb_bits_baseline - max_weight;
for item in 0..sorted_list_size {
let sorted = sorted_list[item];
let symbol = sorted.symbol as u16;
let weight = sorted.weight as usize;
let nb_bits = nb_bits_baseline - weight as u32;
let start = rank_val[weight] as usize;
let length = 1u32 << (target_log - nb_bits);
if target_log - nb_bits >= min_bits {
let mut min_weight = nb_bits as i32 + scale_log;
if min_weight < 1 {
min_weight = 1;
}
let min_weight = min_weight as usize;
let sorted_rank = rank_start0[min_weight + 1] as usize;
huf_fill_dtable_x4_level2(
dtable.add(start),
target_log - nb_bits,
nb_bits,
&rank_val_origin[nb_bits as usize],
min_weight,
sorted_list,
sorted_rank,
sorted_list_size - sorted_rank,
nb_bits_baseline,
symbol,
);
} else {
let end = start + length as usize;
for index in start..end {
huf_set_dtable_x4(dtable, index, symbol, nb_bits, 1);
}
}
rank_val[weight] = rank_val[weight].wrapping_add(length);
}
}
#[allow(clippy::needless_range_loop)]
unsafe fn huf_read_dtable_x4(
dtable: &mut [u32; 1 + (1 << HUF_MAX_TABLELOG)],
src: *const u8,
src_size: usize,
) -> usize {
let mut weight_list = [0u8; HUF_MAX_SYMBOL_VALUE + 1];
let mut sorted_symbols = [SortedSymbol {
symbol: 0,
weight: 0,
}; HUF_MAX_SYMBOL_VALUE + 1];
let mut rank_stats = [0u32; HUF_ABSOLUTE_MAX_TABLELOG + 1];
let mut rank_start0 = [0u32; HUF_ABSOLUTE_MAX_TABLELOG + 2];
let mut rank_val = [[0u32; HUF_ABSOLUTE_MAX_TABLELOG + 1]; HUF_ABSOLUTE_MAX_TABLELOG];
let mut nb_symbols = 0u32;
let mut table_log = 0u32;
let mem_log = dtable[0];
if mem_log > HUF_ABSOLUTE_MAX_TABLELOG as u32 {
return ERROR(ZstdErrorCode::TableLogTooLarge);
}
let i_size = huf_read_stats(
&mut weight_list,
&mut rank_stats,
&mut nb_symbols,
&mut table_log,
src,
src_size,
);
if ERR_isError(i_size) {
return i_size;
}
if table_log > mem_log {
return ERROR(ZstdErrorCode::TableLogTooLarge);
}
let mut max_weight = table_log as usize;
loop {
if rank_stats[max_weight] != 0 {
break;
}
if max_weight == 0 {
return ERROR(ZstdErrorCode::Generic);
}
max_weight -= 1;
}
let mut next_rank_start = 0u32;
for weight in 1..=max_weight {
let current = next_rank_start;
next_rank_start = next_rank_start.wrapping_add(rank_stats[weight]);
rank_start0[weight + 1] = current;
}
rank_start0[0] = next_rank_start;
let size_of_sort = next_rank_start as usize;
for symbol in 0..nb_symbols as usize {
let weight = weight_list[symbol] as usize;
let rank = rank_start0[weight + 1] as usize;
sorted_symbols[rank] = SortedSymbol {
symbol: symbol as u8,
weight: weight as u8,
};
rank_start0[weight + 1] = rank as u32 + 1;
}
rank_start0[1] = 0;
let min_bits = table_log + 1 - max_weight as u32;
let rescale = (mem_log as i32 - table_log as i32) - 1;
let mut next_rank_val = 0u32;
for weight in 1..=max_weight {
let current = next_rank_val;
let shift = (weight as i32 + rescale) as u32;
next_rank_val = next_rank_val.wrapping_add(rank_stats[weight] << shift);
rank_val[0][weight] = current;
}
if min_bits <= mem_log.saturating_sub(min_bits) {
for consumed in min_bits..=mem_log - min_bits {
for weight in 1..=max_weight {
rank_val[consumed as usize][weight] = rank_val[0][weight] >> consumed;
}
}
}
let table = dtable.as_mut_ptr().add(1) as *mut HufDEltX4;
huf_fill_dtable_x4(
table,
mem_log,
&sorted_symbols,
size_of_sort,
&rank_start0,
&rank_val,
max_weight as u32,
table_log + 1,
);
i_size
}
#[inline]
unsafe fn huf_decode_symbol_x4(
op: &mut *mut u8,
stream: &mut DStream,
dtable: *const HufDEltX4,
table_log: u32,
) -> u32 {
let value = look_bits_fast(stream, table_log);
let entry = dtable.add(value);
ptr::copy(entry.cast::<u8>(), *op, 2);
skip_bits(stream, (*entry).nb_bits as u32);
let length = (*entry).length as u32;
*op = (*op).add(length as usize);
length
}
#[inline]
unsafe fn huf_decode_last_symbol_x4(
op: *mut u8,
stream: &mut DStream,
dtable: *const HufDEltX4,
table_log: u32,
) -> u32 {
let value = look_bits_fast(stream, table_log);
let entry = dtable.add(value);
*op = read_le16(entry.cast::<u8>()) as u8;
if (*entry).length == 1 {
skip_bits(stream, (*entry).nb_bits as u32);
} else if stream.bits_consumed < USIZE_BITS {
skip_bits(stream, (*entry).nb_bits as u32);
if stream.bits_consumed > USIZE_BITS {
stream.bits_consumed = USIZE_BITS;
}
}
1
}
#[inline]
unsafe fn huf_decode_symbol_x4_0(
op: &mut *mut u8,
stream: &mut DStream,
dtable: *const HufDEltX4,
table_log: u32,
) {
huf_decode_symbol_x4(op, stream, dtable, table_log);
}
#[inline]
unsafe fn huf_decode_symbol_x4_1(
op: &mut *mut u8,
stream: &mut DStream,
dtable: *const HufDEltX4,
table_log: u32,
) {
if USIZE_BITS == 64 || HUF_MAX_TABLELOG <= 12 {
huf_decode_symbol_x4(op, stream, dtable, table_log);
}
}
#[inline]
unsafe fn huf_decode_symbol_x4_2(
op: &mut *mut u8,
stream: &mut DStream,
dtable: *const HufDEltX4,
table_log: u32,
) {
if USIZE_BITS == 64 {
huf_decode_symbol_x4(op, stream, dtable, table_log);
}
}
unsafe fn huf_decode_stream_x4(
mut p: *mut u8,
stream: &mut DStream,
p_end: *mut u8,
dtable: *const HufDEltX4,
table_log: u32,
) -> usize {
let p_start = p;
while reload_dstream(stream) == DSTREAM_UNFINISHED
&& (p as usize) < (p_end as usize).wrapping_sub(7)
{
huf_decode_symbol_x4_2(&mut p, stream, dtable, table_log);
huf_decode_symbol_x4_1(&mut p, stream, dtable, table_log);
huf_decode_symbol_x4_2(&mut p, stream, dtable, table_log);
huf_decode_symbol_x4_0(&mut p, stream, dtable, table_log);
}
while reload_dstream(stream) == DSTREAM_UNFINISHED
&& (p as usize) <= (p_end as usize).wrapping_sub(2)
{
huf_decode_symbol_x4_0(&mut p, stream, dtable, table_log);
}
while (p as usize) <= (p_end as usize).wrapping_sub(2) {
huf_decode_symbol_x4_0(&mut p, stream, dtable, table_log);
}
if (p as usize) < p_end as usize {
p = p.add(huf_decode_last_symbol_x4(p, stream, dtable, table_log) as usize);
}
(p as usize).wrapping_sub(p_start as usize)
}
unsafe fn huf_decompress4x4_using_dtable(
dst: *mut u8,
dst_size: usize,
c_src: *const u8,
c_src_size: usize,
dtable: &[u32; 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 total = length1
.wrapping_add(length2)
.wrapping_add(length3)
.wrapping_add(6);
let length4 = c_src_size.wrapping_sub(total);
if length4 > c_src_size {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
let streams = [
c_src.add(6),
c_src.add(6 + length1),
c_src.add(6 + length1 + length2),
c_src.add(6 + length1 + length2 + length3),
];
let lengths = [length1, length2, length3, length4];
let segment = dst_size.wrapping_add(3) / 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 table = dtable.as_ptr().add(1) as *const HufDEltX4;
let table_log = dtable[0];
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, streams[index], lengths[index]);
if ERR_isError(error) {
return error;
}
let decoded = huf_decode_stream_x4(
starts[index],
&mut stream,
starts[index].add(sizes[index]),
table,
table_log,
);
if decoded != sizes[index] {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
}
dst_size
}
unsafe fn huf_decompress4x4(
dst: *mut u8,
dst_size: usize,
c_src: *const u8,
c_src_size: usize,
) -> usize {
let mut dtable = [0u32; 1 + (1 << HUF_MAX_TABLELOG)];
dtable[0] = HUF_MAX_TABLELOG as u32;
let header_size = huf_read_dtable_x4(&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_decompress4x4_using_dtable(
dst,
dst_size,
c_src.add(header_size),
c_src_size - header_size,
&dtable,
)
}
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(c_src, dst, dst_size);
return dst_size;
}
if c_src_size == 1 {
ptr::write_bytes(dst, *c_src, dst_size);
return dst_size;
}
const ALGO_TIME: [[[u32; 2]; 3]; 16] = [
[[0, 0], [1, 1], [2, 2]],
[[0, 0], [1, 1], [2, 2]],
[[38, 130], [1313, 74], [2151, 38]],
[[448, 128], [1353, 74], [2238, 41]],
[[556, 128], [1353, 74], [2238, 47]],
[[714, 128], [1418, 74], [2436, 53]],
[[883, 128], [1437, 74], [2464, 61]],
[[897, 128], [1515, 75], [2622, 68]],
[[926, 128], [1613, 75], [2730, 75]],
[[947, 128], [1729, 77], [3359, 77]],
[[1107, 128], [2083, 81], [4006, 84]],
[[1177, 128], [2379, 87], [4785, 88]],
[[1242, 128], [2415, 93], [5155, 84]],
[[1349, 128], [2644, 106], [5260, 106]],
[[1455, 128], [2422, 124], [4174, 124]],
[[722, 128], [1891, 145], [1936, 146]],
];
let q = c_src_size.wrapping_mul(16) / dst_size;
let d256 = (dst_size >> 8) as u32;
let mut dtime = [0u32; 3];
for index in 0..3 {
dtime[index] =
ALGO_TIME[q][index][0].wrapping_add(ALGO_TIME[q][index][1].wrapping_mul(d256));
}
dtime[1] = dtime[1].wrapping_add(dtime[1] >> 4);
dtime[2] = dtime[2].wrapping_add(dtime[2] >> 3);
if dtime[1] < dtime[0] {
huf_decompress4x4(dst, dst_size, c_src, c_src_size)
} else {
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,
)
}
}
/* ******************************************
* v0.4 frame decoder
********************************************/
const FRAME_HEADER_SIZE_MIN: usize = 5;
const FRAME_HEADER_SIZE_MAX: usize = 5;
const BLOCK_HEADER_SIZE: usize = 3;
const WINDOWLOG_ABSOLUTE_MIN: u32 = 11;
const STAGE_GET_FRAME_HEADER_SIZE: u32 = 0;
const STAGE_DECODE_FRAME_HEADER: u32 = 1;
const STAGE_DECODE_BLOCK_HEADER: u32 = 2;
const STAGE_DECOMPRESS_BLOCK: u32 = 3;
#[repr(C)]
#[derive(Clone, Copy)]
struct ZstdParameters {
src_size: u64,
window_log: u32,
content_log: u32,
hash_log: u32,
search_log: u32,
search_length: u32,
strategy: u32,
}
#[repr(C)]
pub struct ZSTDv04_Dctx {
ll_table: [u32; 1 + (1 << LL_FSE_LOG)],
off_table: [u32; 1 + (1 << OFF_FSE_LOG)],
ml_table: [u32; 1 + (1 << ML_FSE_LOG)],
previous_dst_end: *const u8,
base: *const u8,
v_base: *const u8,
dict_end: *const u8,
expected: usize,
header_size: usize,
params: ZstdParameters,
b_type: u32,
stage: u32,
lit_ptr: *const u8,
lit_size: usize,
lit_buffer: [u8; BLOCKSIZE + 8],
header_buffer: [u8; FRAME_HEADER_SIZE_MAX],
}
#[derive(Clone, Copy)]
struct BlockProperties {
block_type: u32,
orig_size: u32,
}
#[inline]
unsafe fn reset_dctx(dctx: &mut ZSTDv04_Dctx) -> usize {
dctx.expected = FRAME_HEADER_SIZE_MIN;
dctx.stage = STAGE_GET_FRAME_HEADER_SIZE;
dctx.previous_dst_end = ptr::null();
dctx.base = ptr::null();
dctx.v_base = ptr::null();
dctx.dict_end = ptr::null();
0
}
#[inline]
unsafe fn create_dctx() -> *mut ZSTDv04_Dctx {
let dctx = libc::malloc(std::mem::size_of::<ZSTDv04_Dctx>()) as *mut ZSTDv04_Dctx;
if dctx.is_null() {
return ptr::null_mut();
}
reset_dctx(&mut *dctx);
dctx
}
#[inline]
unsafe fn free_dctx(dctx: *mut ZSTDv04_Dctx) -> usize {
libc::free(dctx.cast::<c_void>());
0
}
unsafe fn decode_frame_header_part1(
dctx: &mut ZSTDv04_Dctx,
src: *const u8,
src_size: usize,
) -> usize {
if src_size != FRAME_HEADER_SIZE_MIN {
return ERROR(ZstdErrorCode::SrcSizeWrong);
}
if read_le32(src) != ZSTD_MAGIC_NUMBER {
return ERROR(ZstdErrorCode::PrefixUnknown);
}
dctx.header_size = FRAME_HEADER_SIZE_MIN;
dctx.header_size
}
unsafe fn get_frame_params(params: &mut ZstdParameters, src: *const u8, src_size: usize) -> usize {
if src_size < FRAME_HEADER_SIZE_MIN {
return FRAME_HEADER_SIZE_MAX;
}
if read_le32(src) != ZSTD_MAGIC_NUMBER {
return ERROR(ZstdErrorCode::PrefixUnknown);
}
*params = ZstdParameters {
src_size: 0,
window_log: 0,
content_log: 0,
hash_log: 0,
search_log: 0,
search_length: 0,
strategy: 0,
};
let descriptor = *src.add(4);
params.window_log = (descriptor & 15) as u32 + WINDOWLOG_ABSOLUTE_MIN;
if descriptor >> 4 != 0 {
return ERROR(ZstdErrorCode::FrameParameterUnsupported);
}
0
}
unsafe fn decode_frame_header_part2(
dctx: &mut ZSTDv04_Dctx,
src: *const u8,
src_size: usize,
) -> usize {
if src_size != dctx.header_size {
return ERROR(ZstdErrorCode::SrcSizeWrong);
}
let result = get_frame_params(&mut dctx.params, src, src_size);
if USIZE_BITS == 32 && dctx.params.window_log > 25 {
return ERROR(ZstdErrorCode::FrameParameterUnsupported);
}
result
}
unsafe fn get_block_size(
src: *const u8,
src_size: usize,
properties: &mut BlockProperties,
) -> usize {
if src_size < BLOCK_HEADER_SIZE {
return ERROR(ZstdErrorCode::SrcSizeWrong);
}
let header_flags = *src;
let c_size = *src.add(2) as usize
| ((*src.add(1) as usize) << 8)
| (((header_flags as usize) & 7) << 16);
properties.block_type = (header_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_raw_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(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.wrapping_add(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.wrapping_add(5)
}
unsafe fn decode_literals_block(dctx: &mut ZSTDv04_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;
let offset = lit_size.min(BLOCKSIZE);
if offset + 8 <= dctx.lit_buffer.len() {
dctx.lit_buffer[offset..offset + 8].fill(0);
}
read_size
}
IS_RAW => {
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(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);
return lit_size + 3;
}
dctx.lit_ptr = src.add(3);
dctx.lit_size = lit_size;
lit_size + 3
}
IS_RLE => {
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
}
_ => ERROR(ZstdErrorCode::CorruptionDetected),
}
}
unsafe fn decode_seq_headers(
dctx: &mut ZSTDv04_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.wrapping_add(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;
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 | ((types as usize & 1) << 8);
ip = ip.add(2);
}
*dumps = ip;
ip = ip.add(dump_size);
*dumps_length = dump_size;
if (ip as usize) > end.wrapping_sub(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.wrapping_sub(ip as usize),
);
if ERR_isError(size) {
return ERROR(ZstdErrorCode::Generic);
}
if log > LL_FSE_LOG {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
ip = ip.add(size);
fse_build_dtable(&mut dctx.ll_table, &norm, max, log);
}
}
match off_type {
BT_RLE => {
if (ip as usize) > end.wrapping_sub(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.wrapping_sub(ip as usize),
);
if ERR_isError(size) {
return ERROR(ZstdErrorCode::Generic);
}
if log > OFF_FSE_LOG {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
ip = ip.add(size);
fse_build_dtable(&mut dctx.off_table, &norm, max, log);
}
}
match ml_type {
BT_RLE => {
if (ip as usize) > end.wrapping_sub(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.wrapping_sub(ip as usize),
);
if ERR_isError(size) {
return ERROR(ZstdErrorCode::Generic);
}
if log > ML_FSE_LOG {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
ip = ip.add(size);
fse_build_dtable(&mut dctx.ml_table, &norm, max, log);
}
}
(ip as usize).wrapping_sub(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 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
};
let mut lit_length = lit_length;
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);
}
if (dumps as usize) >= dumps_end as usize {
dumps = dumps_end.wrapping_sub(1);
}
}
let offset_code = fse_decode_symbol(&mut state.state_off, &mut state.stream, false) as u32;
if USIZE_BITS == 32 {
reload_dstream(&mut state.stream);
}
let nb_bits = offset_code.wrapping_sub(1);
let nb_bits = if offset_code == 0 { 0 } else { nb_bits };
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 USIZE_BITS == 32 {
reload_dstream(&mut state.stream);
}
if offset_code == 0 {
offset = previous_offset;
}
if offset_code != 0 || lit_length == 0 {
state.prev_offset = sequence.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);
}
if (dumps as usize) >= dumps_end as usize {
dumps = dumps_end.wrapping_sub(1);
}
}
sequence.lit_length = lit_length;
sequence.offset = offset;
sequence.match_length = match_length + MINMATCH;
state.dumps = dumps;
}
#[allow(clippy::too_many_arguments)]
unsafe fn exec_sequence(
mut op: *mut u8,
mut sequence: Sequence,
lit_ptr: &mut *const u8,
lit_limit: *const u8,
base: *const u8,
v_base: *const u8,
dict_end: *const u8,
oend: *mut u8,
) -> usize {
let o_lit_end = (op as usize).wrapping_add(sequence.lit_length);
let sequence_length = sequence.lit_length.wrapping_add(sequence.match_length);
let o_match_end = o_lit_end.wrapping_add(sequence.match_length);
let oend_addr = oend as usize;
let oend_8 = oend_addr.wrapping_sub(8);
let lit_end = (*lit_ptr as usize).wrapping_add(sequence.lit_length);
let mut match_addr = o_lit_end.wrapping_sub(sequence.offset);
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_8 {
return ERROR(ZstdErrorCode::DstSizeTooSmall);
}
if o_match_end > oend_addr {
return ERROR(ZstdErrorCode::DstSizeTooSmall);
}
if lit_end > lit_limit as usize {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
zstd_wildcopy(op, *lit_ptr, sequence.lit_length as isize);
op = o_lit_end as *mut u8;
*lit_ptr = lit_end as *const u8;
if sequence.offset > o_lit_end.wrapping_sub(base as usize) {
if sequence.offset > o_lit_end.wrapping_sub(v_base as usize) {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
match_addr = (dict_end as usize).wrapping_sub((base as usize).wrapping_sub(match_addr));
if match_addr.wrapping_add(sequence.match_length) <= dict_end as usize {
ptr::copy(
match_addr as *const u8,
o_lit_end as *mut u8,
sequence.match_length,
);
return sequence_length;
}
let length1 = (dict_end as usize).wrapping_sub(match_addr);
ptr::copy(match_addr as *const u8, o_lit_end as *mut u8, length1);
op = o_lit_end.wrapping_add(length1) as *mut u8;
sequence.match_length = sequence.match_length.wrapping_sub(length1);
match_addr = base as usize;
if (op as usize) > oend_8 || sequence.match_length < MINMATCH {
let mut out = op;
let mut m = match_addr as *const u8;
while (out as usize) < o_match_end {
*out = *m;
out = out.add(1);
m = m.add(1);
}
return sequence_length;
}
}
if sequence.offset < 8 {
let dec32 = [0usize, 1, 2, 1, 4, 4, 4, 4];
let dec64 = [8usize, 8, 8, 7, 8, 9, 10, 11];
*op = *(match_addr as *const u8);
*op.add(1) = *(match_addr as *const u8).add(1);
*op.add(2) = *(match_addr as *const u8).add(2);
*op.add(3) = *(match_addr as *const u8).add(3);
let match_ptr = (match_addr as *const u8).add(dec32[sequence.offset]);
zstd_copy4(op.add(4), match_ptr);
match_addr = match_addr
.wrapping_add(8)
.wrapping_sub(dec64[sequence.offset]);
} else {
zstd_copy8(op, match_addr as *const u8);
match_addr = match_addr.wrapping_add(8);
}
op = op.add(8);
match_addr = match_addr.wrapping_add(8);
if o_match_end > oend_addr.wrapping_sub(16 - MINMATCH) {
if (op as usize) < oend_8 {
let dist = oend_8 - op as usize;
zstd_wildcopy(op, match_addr as *const u8, dist as isize);
match_addr = match_addr.wrapping_add(dist);
op = oend_8 as *mut u8;
}
while (op as usize) < o_match_end {
*op = *(match_addr as *const u8);
op = op.add(1);
match_addr = match_addr.wrapping_add(1);
}
} else {
zstd_wildcopy(
op,
match_addr as *const u8,
sequence.match_length.wrapping_sub(8) as isize,
);
}
sequence_length
}
unsafe fn decompress_sequences(
dctx: &mut ZSTDv04_Dctx,
dst: *mut u8,
max_dst_size: usize,
seq_start: *const u8,
seq_size: usize,
) -> usize {
let ip = seq_start;
let iend = (seq_start as usize).wrapping_add(seq_size);
let ostart = dst;
let mut op = dst;
let oend = (dst as usize).wrapping_add(max_dst_size) as *mut u8;
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;
}
let 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, iend.wrapping_sub(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 reload_dstream(&mut state.stream) <= DSTREAM_COMPLETED && nb_seq != 0 {
nb_seq -= 1;
decode_sequence(&mut sequence, &mut state);
let produced = exec_sequence(
op,
sequence,
&mut dctx.lit_ptr,
lit_limit,
dctx.base,
dctx.v_base,
dctx.dict_end,
oend,
);
if ERR_isError(produced) {
return produced;
}
op = op.add(produced);
}
if !end_of_dstream(&state.stream) {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
let last_literal_size = (lit_limit as usize).wrapping_sub(dctx.lit_ptr as usize);
if dctx.lit_ptr as usize > lit_limit as usize {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
if (op as usize).wrapping_add(last_literal_size) > oend as usize {
return ERROR(ZstdErrorCode::DstSizeTooSmall);
}
if last_literal_size != 0 {
if !ptr::eq(op.cast_const(), dctx.lit_ptr) {
ptr::copy(dctx.lit_ptr, op, last_literal_size);
}
op = op.add(last_literal_size);
}
(op as usize).wrapping_sub(ostart as usize)
}
unsafe fn check_continuity(dctx: &mut ZSTDv04_Dctx, dst: *const u8) {
if dst != dctx.previous_dst_end {
dctx.dict_end = dctx.previous_dst_end;
let delta = (dctx.previous_dst_end as usize).wrapping_sub(dctx.base as usize);
dctx.v_base = (dst as usize).wrapping_sub(delta) as *const u8;
dctx.base = dst;
dctx.previous_dst_end = dst;
}
}
unsafe fn decompress_block_internal(
dctx: &mut ZSTDv04_Dctx,
dst: *mut u8,
max_dst_size: usize,
src: *const u8,
src_size: usize,
) -> usize {
if src_size > BLOCKSIZE {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
let literal_size = decode_literals_block(dctx, src, src_size);
if ERR_isError(literal_size) {
return literal_size;
}
decompress_sequences(
dctx,
dst,
max_dst_size,
src.add(literal_size),
src_size - literal_size,
)
}
unsafe fn decompress_insert_dictionary(dctx: &mut ZSTDv04_Dctx, dict: *const u8, dict_size: usize) {
dctx.dict_end = dctx.previous_dst_end;
let delta = (dctx.previous_dst_end as usize).wrapping_sub(dctx.base as usize);
dctx.v_base = (dict as usize).wrapping_sub(delta) as *const u8;
dctx.base = dict;
dctx.previous_dst_end = (dict as usize).wrapping_add(dict_size) as *const u8;
}
unsafe fn decompress_using_dict(
ctx: &mut ZSTDv04_Dctx,
dst: *mut u8,
max_dst_size: usize,
src: *const u8,
src_size: usize,
dict: *const u8,
dict_size: usize,
) -> usize {
let mut ip = src;
let ostart = dst;
let mut op = dst;
let oend_addr = (dst as usize).wrapping_add(max_dst_size);
let mut remaining_size = src_size;
let mut block_properties = BlockProperties {
block_type: BT_END,
orig_size: 0,
};
reset_dctx(ctx);
if !dict.is_null() {
decompress_insert_dictionary(ctx, dict, dict_size);
ctx.dict_end = ctx.previous_dst_end;
let delta = (ctx.previous_dst_end as usize).wrapping_sub(ctx.base as usize);
ctx.v_base = (dst as usize).wrapping_sub(delta) as *const u8;
ctx.base = dst;
} else {
ctx.v_base = dst;
ctx.base = dst;
ctx.dict_end = dst;
}
if src_size < FRAME_HEADER_SIZE_MIN + BLOCK_HEADER_SIZE {
return ERROR(ZstdErrorCode::SrcSizeWrong);
}
let header_size = decode_frame_header_part1(ctx, src, FRAME_HEADER_SIZE_MIN);
if ERR_isError(header_size) {
return header_size;
}
if src_size < header_size + BLOCK_HEADER_SIZE {
return ERROR(ZstdErrorCode::SrcSizeWrong);
}
ip = ip.add(header_size);
remaining_size -= header_size;
let header_result = decode_frame_header_part2(ctx, src, header_size);
if ERR_isError(header_result) {
return header_result;
}
loop {
let block_size = get_block_size(ip, remaining_size, &mut block_properties);
if ERR_isError(block_size) {
return block_size;
}
ip = ip.add(BLOCK_HEADER_SIZE);
remaining_size -= BLOCK_HEADER_SIZE;
if block_size > remaining_size {
return ERROR(ZstdErrorCode::SrcSizeWrong);
}
let decoded_size = match block_properties.block_type {
BT_COMPRESSED => decompress_block_internal(
ctx,
op,
oend_addr.wrapping_sub(op as usize),
ip,
block_size,
),
BT_RAW => copy_raw_block(op, oend_addr.wrapping_sub(op as usize), ip, block_size),
BT_RLE => return ERROR(ZstdErrorCode::Generic),
BT_END => {
if remaining_size != 0 {
return ERROR(ZstdErrorCode::SrcSizeWrong);
}
0
}
_ => return ERROR(ZstdErrorCode::Generic),
};
if block_size == 0 {
break;
}
if ERR_isError(decoded_size) {
return decoded_size;
}
op = op.add(decoded_size);
ip = ip.add(block_size);
remaining_size -= block_size;
}
(op as usize).wrapping_sub(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;
}
unsafe fn find_frame_size_info(
src: *const u8,
src_size: usize,
c_size: *mut usize,
d_bound: *mut u64,
) {
let mut ip = src;
let mut remaining_size = src_size;
let mut nb_blocks = 0usize;
let mut block_properties = BlockProperties {
block_type: BT_END,
orig_size: 0,
};
if src_size < FRAME_HEADER_SIZE_MIN {
error_frame_size_info(c_size, d_bound, ERROR(ZstdErrorCode::SrcSizeWrong));
return;
}
if read_le32(src) != ZSTD_MAGIC_NUMBER {
error_frame_size_info(c_size, d_bound, ERROR(ZstdErrorCode::PrefixUnknown));
return;
}
ip = ip.add(FRAME_HEADER_SIZE_MIN);
remaining_size -= FRAME_HEADER_SIZE_MIN;
loop {
let block_size = get_block_size(ip, remaining_size, &mut block_properties);
if ERR_isError(block_size) {
error_frame_size_info(c_size, d_bound, block_size);
return;
}
ip = ip.add(BLOCK_HEADER_SIZE);
remaining_size -= BLOCK_HEADER_SIZE;
if block_size > remaining_size {
error_frame_size_info(c_size, d_bound, ERROR(ZstdErrorCode::SrcSizeWrong));
return;
}
if block_size == 0 {
break;
}
ip = ip.add(block_size);
remaining_size -= block_size;
nb_blocks = nb_blocks.wrapping_add(1);
}
*c_size = (ip as usize).wrapping_sub(src as usize);
*d_bound = (nb_blocks.wrapping_mul(BLOCKSIZE)) as u64;
}
unsafe fn next_src_size_to_decompress(dctx: &ZSTDv04_Dctx) -> usize {
dctx.expected
}
unsafe fn decompress_continue(
ctx: &mut ZSTDv04_Dctx,
dst: *mut u8,
max_dst_size: usize,
src: *const u8,
src_size: usize,
) -> usize {
if src_size != ctx.expected {
return ERROR(ZstdErrorCode::SrcSizeWrong);
}
check_continuity(ctx, dst);
match ctx.stage {
STAGE_GET_FRAME_HEADER_SIZE => {
if src_size != FRAME_HEADER_SIZE_MIN {
return ERROR(ZstdErrorCode::SrcSizeWrong);
}
ctx.header_size = decode_frame_header_part1(ctx, src, FRAME_HEADER_SIZE_MIN);
if ERR_isError(ctx.header_size) {
return ctx.header_size;
}
ptr::copy_nonoverlapping(src, ctx.header_buffer.as_mut_ptr(), FRAME_HEADER_SIZE_MIN);
if ctx.header_size > FRAME_HEADER_SIZE_MIN {
return ERROR(ZstdErrorCode::Generic);
}
ctx.expected = 0;
let result =
decode_frame_header_part2(ctx, ctx.header_buffer.as_ptr(), ctx.header_size);
if ERR_isError(result) {
return result;
}
ctx.expected = BLOCK_HEADER_SIZE;
ctx.stage = STAGE_DECODE_BLOCK_HEADER;
0
}
STAGE_DECODE_FRAME_HEADER => {
let result =
decode_frame_header_part2(ctx, ctx.header_buffer.as_ptr(), ctx.header_size);
if ERR_isError(result) {
return result;
}
ctx.expected = BLOCK_HEADER_SIZE;
ctx.stage = STAGE_DECODE_BLOCK_HEADER;
0
}
STAGE_DECODE_BLOCK_HEADER => {
let mut properties = BlockProperties {
block_type: BT_END,
orig_size: 0,
};
let block_size = get_block_size(src, BLOCK_HEADER_SIZE, &mut properties);
if ERR_isError(block_size) {
return block_size;
}
if properties.block_type == BT_END {
ctx.expected = 0;
ctx.stage = STAGE_GET_FRAME_HEADER_SIZE;
} else {
ctx.expected = block_size;
ctx.b_type = properties.block_type;
ctx.stage = STAGE_DECOMPRESS_BLOCK;
}
0
}
STAGE_DECOMPRESS_BLOCK => {
let result = match ctx.b_type {
BT_COMPRESSED => decompress_block_internal(ctx, dst, max_dst_size, src, src_size),
BT_RAW => copy_raw_block(dst, max_dst_size, src, src_size),
BT_RLE => return ERROR(ZstdErrorCode::Generic),
BT_END => 0,
_ => return ERROR(ZstdErrorCode::Generic),
};
ctx.stage = STAGE_DECODE_BLOCK_HEADER;
ctx.expected = BLOCK_HEADER_SIZE;
if ERR_isError(result) {
return result;
}
ctx.previous_dst_end = (dst as usize).wrapping_add(result) as *const u8;
result
}
_ => ERROR(ZstdErrorCode::Generic),
}
}
#[repr(C)]
pub struct ZBUFFv04_DCtx {
zc: *mut ZSTDv04_Dctx,
params: ZstdParameters,
in_buff: *mut u8,
in_buff_size: usize,
in_pos: usize,
out_buff: *mut u8,
out_buff_size: usize,
out_start: usize,
out_end: usize,
h_pos: usize,
dict: *const u8,
dict_size: usize,
stage: u32,
header_buffer: [u8; FRAME_HEADER_SIZE_MAX],
}
const BUFF_INIT: u32 = 0;
const BUFF_READ_HEADER: u32 = 1;
const BUFF_LOAD_HEADER: u32 = 2;
const BUFF_DECODE_HEADER: u32 = 3;
const BUFF_READ: u32 = 4;
const BUFF_LOAD: u32 = 5;
const BUFF_FLUSH: u32 = 6;
unsafe fn buff_create_dctx() -> *mut ZBUFFv04_DCtx {
let zbc = libc::malloc(std::mem::size_of::<ZBUFFv04_DCtx>()) as *mut ZBUFFv04_DCtx;
if zbc.is_null() {
return ptr::null_mut();
}
ptr::write_bytes(zbc.cast::<u8>(), 0, std::mem::size_of::<ZBUFFv04_DCtx>());
(*zbc).zc = create_dctx();
(*zbc).stage = BUFF_INIT;
zbc
}
unsafe fn buff_free_dctx(zbc: *mut ZBUFFv04_DCtx) -> usize {
if zbc.is_null() {
return 0;
}
free_dctx((*zbc).zc);
libc::free((*zbc).in_buff.cast::<c_void>());
libc::free((*zbc).out_buff.cast::<c_void>());
libc::free(zbc.cast::<c_void>());
0
}
unsafe fn buff_decompress_init(zbc: &mut ZBUFFv04_DCtx) -> usize {
zbc.stage = BUFF_READ_HEADER;
zbc.h_pos = 0;
zbc.in_pos = 0;
zbc.out_start = 0;
zbc.out_end = 0;
zbc.dict_size = 0;
reset_dctx(&mut *zbc.zc)
}
unsafe fn buff_decompress_with_dictionary(
zbc: &mut ZBUFFv04_DCtx,
dict: *const u8,
dict_size: usize,
) -> usize {
zbc.dict = dict;
zbc.dict_size = dict_size;
0
}
unsafe fn buff_limit_copy(
dst: *mut u8,
max_dst_size: usize,
src: *const u8,
src_size: usize,
) -> usize {
let length = max_dst_size.min(src_size);
if length != 0 {
ptr::copy(src, dst, length);
}
length
}
unsafe fn buff_decompress_continue(
zbc: &mut ZBUFFv04_DCtx,
dst: *mut u8,
max_dst_size_ptr: *mut usize,
src: *const u8,
src_size_ptr: *mut usize,
) -> usize {
let istart = src as usize;
let mut ip = istart;
let iend = istart.wrapping_add(*src_size_ptr);
let ostart = dst as usize;
let mut op = ostart;
let oend = ostart.wrapping_add(*max_dst_size_ptr);
let mut not_done = true;
while not_done {
match zbc.stage {
BUFF_INIT => return ERROR(ZstdErrorCode::InitMissing),
BUFF_READ_HEADER => {
let header_size = get_frame_params(&mut zbc.params, src, *src_size_ptr);
if ERR_isError(header_size) {
return header_size;
}
if header_size != 0 {
ptr::copy(
src,
zbc.header_buffer.as_mut_ptr().add(zbc.h_pos),
*src_size_ptr,
);
zbc.h_pos += *src_size_ptr;
*max_dst_size_ptr = 0;
zbc.stage = BUFF_LOAD_HEADER;
return header_size - zbc.h_pos;
}
zbc.stage = BUFF_DECODE_HEADER;
}
BUFF_LOAD_HEADER => {
let header_size = buff_limit_copy(
zbc.header_buffer.as_mut_ptr().add(zbc.h_pos),
FRAME_HEADER_SIZE_MAX - zbc.h_pos,
src,
*src_size_ptr,
);
zbc.h_pos += header_size;
ip = ip.wrapping_add(header_size);
let header_size =
get_frame_params(&mut zbc.params, zbc.header_buffer.as_ptr(), zbc.h_pos);
if ERR_isError(header_size) {
return header_size;
}
if header_size != 0 {
*max_dst_size_ptr = 0;
return header_size - zbc.h_pos;
}
zbc.stage = BUFF_DECODE_HEADER;
}
BUFF_DECODE_HEADER => {
let needed_out_size = 1usize << zbc.params.window_log;
let needed_in_size = BLOCKSIZE;
if zbc.in_buff_size < needed_in_size {
libc::free(zbc.in_buff.cast::<c_void>());
zbc.in_buff_size = needed_in_size;
zbc.in_buff = libc::malloc(needed_in_size) as *mut u8;
if zbc.in_buff.is_null() {
return ERROR(ZstdErrorCode::MemoryAllocation);
}
}
if zbc.out_buff_size < needed_out_size {
libc::free(zbc.out_buff.cast::<c_void>());
zbc.out_buff_size = needed_out_size;
zbc.out_buff = libc::malloc(needed_out_size) as *mut u8;
if zbc.out_buff.is_null() {
return ERROR(ZstdErrorCode::MemoryAllocation);
}
}
if zbc.dict_size != 0 {
decompress_insert_dictionary(&mut *zbc.zc, zbc.dict, zbc.dict_size);
}
if zbc.h_pos != 0 {
ptr::copy(zbc.header_buffer.as_ptr(), zbc.in_buff, zbc.h_pos);
zbc.in_pos = zbc.h_pos;
zbc.h_pos = 0;
zbc.stage = BUFF_LOAD;
} else {
zbc.stage = BUFF_READ;
}
}
BUFF_READ => {
let needed_in_size = next_src_size_to_decompress(&*zbc.zc);
if needed_in_size == 0 {
zbc.stage = BUFF_INIT;
not_done = false;
continue;
}
if iend.wrapping_sub(ip) >= needed_in_size {
let decoded_size = decompress_continue(
&mut *zbc.zc,
zbc.out_buff.add(zbc.out_start),
zbc.out_buff_size - zbc.out_start,
ip as *const u8,
needed_in_size,
);
if ERR_isError(decoded_size) {
return decoded_size;
}
ip = ip.wrapping_add(needed_in_size);
if decoded_size == 0 {
continue;
}
zbc.out_end = zbc.out_start + decoded_size;
zbc.stage = BUFF_FLUSH;
continue;
}
if ip == iend {
not_done = false;
continue;
}
zbc.stage = BUFF_LOAD;
}
BUFF_LOAD => {
let needed_in_size = next_src_size_to_decompress(&*zbc.zc);
let to_load = needed_in_size.wrapping_sub(zbc.in_pos);
if to_load > zbc.in_buff_size.wrapping_sub(zbc.in_pos) {
return ERROR(ZstdErrorCode::CorruptionDetected);
}
let loaded_size = buff_limit_copy(
zbc.in_buff.add(zbc.in_pos),
to_load,
ip as *const u8,
iend.wrapping_sub(ip),
);
ip = ip.wrapping_add(loaded_size);
zbc.in_pos += loaded_size;
if loaded_size < to_load {
not_done = false;
continue;
}
let decoded_size = decompress_continue(
&mut *zbc.zc,
zbc.out_buff.add(zbc.out_start),
zbc.out_buff_size - zbc.out_start,
zbc.in_buff,
needed_in_size,
);
if ERR_isError(decoded_size) {
return decoded_size;
}
zbc.in_pos = 0;
if decoded_size == 0 {
zbc.stage = BUFF_READ;
continue;
}
zbc.out_end = zbc.out_start + decoded_size;
zbc.stage = BUFF_FLUSH;
}
BUFF_FLUSH => {
let to_flush_size = zbc.out_end.wrapping_sub(zbc.out_start);
let flushed_size = buff_limit_copy(
op as *mut u8,
oend.wrapping_sub(op),
zbc.out_buff.add(zbc.out_start),
to_flush_size,
);
op = op.wrapping_add(flushed_size);
zbc.out_start += flushed_size;
if flushed_size == to_flush_size {
zbc.stage = BUFF_READ;
if zbc.out_start + BLOCKSIZE > zbc.out_buff_size {
zbc.out_start = 0;
zbc.out_end = 0;
}
} else {
not_done = false;
}
}
_ => return ERROR(ZstdErrorCode::Generic),
}
}
*src_size_ptr = ip.wrapping_sub(istart);
*max_dst_size_ptr = op.wrapping_sub(ostart);
let mut next_src_size_hint = next_src_size_to_decompress(&*zbc.zc);
if next_src_size_hint > 3 {
next_src_size_hint = next_src_size_hint.wrapping_add(3);
}
next_src_size_hint.wrapping_sub(zbc.in_pos)
}
#[no_mangle]
pub extern "C" fn ZSTDv04_isError(code: usize) -> c_uint {
ERR_isError(code) as c_uint
}
#[no_mangle]
pub unsafe extern "C" fn ZSTDv04_decompress(
dst: *mut c_void,
max_original_size: usize,
src: *const c_void,
compressed_size: usize,
) -> usize {
let dctx = create_dctx();
if dctx.is_null() {
return ERROR(ZstdErrorCode::MemoryAllocation);
}
let result = decompress_using_dict(
&mut *dctx,
dst.cast::<u8>(),
max_original_size,
src.cast::<u8>(),
compressed_size,
ptr::null(),
0,
);
free_dctx(dctx);
result
}
#[no_mangle]
pub unsafe extern "C" fn ZSTDv04_decompressDCtx(
dctx: *mut ZSTDv04_Dctx,
dst: *mut c_void,
max_original_size: usize,
src: *const c_void,
compressed_size: usize,
) -> usize {
decompress_using_dict(
&mut *dctx,
dst.cast::<u8>(),
max_original_size,
src.cast::<u8>(),
compressed_size,
ptr::null(),
0,
)
}
#[no_mangle]
pub unsafe extern "C" fn ZSTDv04_findFrameSizeInfoLegacy(
src: *const c_void,
src_size: usize,
c_size: *mut usize,
d_bound: *mut u64,
) {
find_frame_size_info(src.cast::<u8>(), src_size, c_size, d_bound);
}
#[no_mangle]
pub unsafe extern "C" fn ZSTDv04_resetDCtx(dctx: *mut ZSTDv04_Dctx) -> usize {
reset_dctx(&mut *dctx)
}
#[no_mangle]
pub unsafe extern "C" fn ZSTDv04_createDCtx() -> *mut ZSTDv04_Dctx {
create_dctx()
}
#[no_mangle]
pub unsafe extern "C" fn ZSTDv04_freeDCtx(dctx: *mut ZSTDv04_Dctx) -> usize {
free_dctx(dctx)
}
#[no_mangle]
pub unsafe extern "C" fn ZSTDv04_nextSrcSizeToDecompress(dctx: *mut ZSTDv04_Dctx) -> usize {
next_src_size_to_decompress(&*dctx)
}
#[no_mangle]
pub unsafe extern "C" fn ZSTDv04_decompressContinue(
dctx: *mut ZSTDv04_Dctx,
dst: *mut c_void,
max_dst_size: usize,
src: *const c_void,
src_size: usize,
) -> usize {
decompress_continue(
&mut *dctx,
dst.cast::<u8>(),
max_dst_size,
src.cast::<u8>(),
src_size,
)
}
#[no_mangle]
pub extern "C" fn ZBUFFv04_isError(code: usize) -> c_uint {
ERR_isError(code) as c_uint
}
#[no_mangle]
pub extern "C" fn ZBUFFv04_getErrorName(code: usize) -> *const c_char {
crate::errors::ERR_getErrorName(code)
}
#[no_mangle]
pub extern "C" fn ZBUFFv04_recommendedDInSize() -> usize {
BLOCKSIZE + 3
}
#[no_mangle]
pub extern "C" fn ZBUFFv04_recommendedDOutSize() -> usize {
BLOCKSIZE
}
#[no_mangle]
pub unsafe extern "C" fn ZBUFFv04_createDCtx() -> *mut ZBUFFv04_DCtx {
buff_create_dctx()
}
#[no_mangle]
pub unsafe extern "C" fn ZBUFFv04_freeDCtx(dctx: *mut ZBUFFv04_DCtx) -> usize {
buff_free_dctx(dctx)
}
#[no_mangle]
pub unsafe extern "C" fn ZBUFFv04_decompressInit(dctx: *mut ZBUFFv04_DCtx) -> usize {
buff_decompress_init(&mut *dctx)
}
#[no_mangle]
pub unsafe extern "C" fn ZBUFFv04_decompressWithDictionary(
dctx: *mut ZBUFFv04_DCtx,
dict: *const c_void,
dict_size: usize,
) -> usize {
buff_decompress_with_dictionary(&mut *dctx, dict.cast::<u8>(), dict_size)
}
#[no_mangle]
pub unsafe extern "C" fn ZBUFFv04_decompressContinue(
dctx: *mut ZBUFFv04_DCtx,
dst: *mut c_void,
max_dst_size_ptr: *mut usize,
src: *const c_void,
src_size_ptr: *mut usize,
) -> usize {
buff_decompress_continue(
&mut *dctx,
dst.cast::<u8>(),
max_dst_size_ptr,
src.cast::<u8>(),
src_size_ptr,
)
}
#[cfg(test)]
mod tests {
use super::*;
const RAW_FRAME: &[u8] = &[
0x24, 0xB5, 0x2F, 0xFD, 0x00, 0x40, 0x00, 0x0B, b'r', b'a', b'w', b' ', b'v', b'0', b'.',
b'4', b'!', b'!', b'!', 0xC0, 0x00, 0x00,
];
const COMPRESSED_FRAME: &[u8] = &[
0x24, 0xB5, 0x2F, 0xFD, 0x00, 0x00, 0x00, 0xBB, 0xB0, 0x02, 0xC0, 0x10, 0x00, 0x1E, 0xB0,
0x01, 0x02, 0x00, 0x00, 0x80, 0x00, 0xE8, 0x92, 0x34, 0x12, 0x97, 0xC8, 0xDF, 0xE9, 0xF3,
0xEF, 0x53, 0xEA, 0x1D, 0x27, 0x4F, 0x0C, 0x44, 0x90, 0x0C, 0x8D, 0xF1, 0xB4, 0x89, 0x17,
0x00, 0x18, 0x00, 0x18, 0x00, 0x3F, 0xE6, 0xE2, 0xE3, 0x74, 0xD6, 0xEC, 0xC9, 0x4A, 0xE0,
0x71, 0x71, 0x42, 0x3E, 0x64, 0x4F, 0x6A, 0x45, 0x4E, 0x78, 0xEC, 0x49, 0x03, 0x3F, 0xC6,
0x80, 0xAB, 0x8F, 0x75, 0x5E, 0x6F, 0x2E, 0x3E, 0x7E, 0xC6, 0xDC, 0x45, 0x69, 0x6C, 0xC5,
0xFD, 0xC7, 0x40, 0xB8, 0x84, 0x8A, 0x01, 0xEB, 0xA8, 0xD1, 0x40, 0x39, 0x90, 0x4C, 0x64,
0xF8, 0xEB, 0x53, 0xE6, 0x18, 0x0B, 0x67, 0x12, 0xAD, 0xB8, 0x99, 0xB3, 0x5A, 0x6F, 0x8A,
0x19, 0x03, 0x01, 0x50, 0x67, 0x56, 0xF5, 0x9F, 0x35, 0x84, 0x60, 0xA0, 0x60, 0x91, 0xC9,
0x0A, 0xDC, 0xAB, 0xAB, 0xE0, 0xE2, 0x81, 0xFA, 0xCF, 0xC6, 0xBA, 0x01, 0x0E, 0x00, 0x54,
0x00, 0x00, 0x19, 0x00, 0x00, 0x54, 0x14, 0x00, 0x24, 0x24, 0x04, 0xFE, 0x04, 0x84, 0x4E,
0x41, 0x00, 0x27, 0xE2, 0x02, 0xC4, 0xB1, 0x00, 0xD2, 0x51, 0x00, 0x79, 0x58, 0x41, 0x28,
0x00, 0xE0, 0x0C, 0x01, 0x68, 0x65, 0x00, 0x04, 0x13, 0x0C, 0xDA, 0x0C, 0x80, 0x22, 0x06,
0xC0, 0x00, 0x00,
];
unsafe fn decode(frame: &[u8], capacity: usize) -> Result<Vec<u8>, usize> {
let mut output = vec![0u8; capacity];
let size = ZSTDv04_decompress(
output.as_mut_ptr().cast::<c_void>(),
output.len(),
frame.as_ptr().cast::<c_void>(),
frame.len(),
);
if ZSTDv04_isError(size) != 0 {
Err(size)
} else {
output.truncate(size);
Ok(output)
}
}
#[test]
fn decodes_raw_v04_frames() {
unsafe {
assert_eq!(decode(RAW_FRAME, 32).unwrap(), b"raw v0.4!!!");
}
}
#[test]
fn decodes_repository_compressed_v04_frame() {
let expected =
b"snowden is snowed in / he's now then in his snow den / when does the snow end?\n\
goodbye little dog / you dug some holes in your day / they'll be hard to fill.\n\
when life shuts a door, / just open it. it\xE2\x80\x99s a door. / that is how doors work.\n";
unsafe {
assert_eq!(decode(COMPRESSED_FRAME, expected.len()).unwrap(), expected);
}
}
#[test]
fn reports_frame_size_and_errors() {
unsafe {
let mut c_size = 0;
let mut bound = 0;
ZSTDv04_findFrameSizeInfoLegacy(
RAW_FRAME.as_ptr().cast::<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(RAW_FRAME, 1).unwrap_err(),
ERROR(ZstdErrorCode::DstSizeTooSmall)
);
}
}
#[test]
fn streaming_raw_frame_matches_one_shot() {
let expected = b"raw v0.4!!!";
let mut output = vec![0u8; expected.len()];
let mut input_offset = 0;
let mut output_offset = 0;
unsafe {
let dctx = ZSTDv04_createDCtx();
assert!(!dctx.is_null());
loop {
let needed = ZSTDv04_nextSrcSizeToDecompress(dctx);
if needed == 0 {
break;
}
assert!(input_offset + needed <= RAW_FRAME.len());
let produced = ZSTDv04_decompressContinue(
dctx,
output.as_mut_ptr().add(output_offset).cast::<c_void>(),
output.len() - output_offset,
RAW_FRAME.as_ptr().add(input_offset).cast::<c_void>(),
needed,
);
assert_eq!(ZSTDv04_isError(produced), 0);
input_offset += needed;
output_offset += produced;
}
assert_eq!(ZSTDv04_freeDCtx(dctx), 0);
}
assert_eq!(input_offset, RAW_FRAME.len());
assert_eq!(output_offset, expected.len());
assert_eq!(output, expected);
}
#[test]
fn buffered_streaming_raw_frame_matches_one_shot() {
let expected = b"raw v0.4!!!";
let mut output = vec![0u8; expected.len()];
let mut input_size = RAW_FRAME.len();
let mut output_size = output.len();
unsafe {
let dctx = ZBUFFv04_createDCtx();
assert!(!dctx.is_null());
assert_eq!(ZBUFFv04_decompressInit(dctx), 0);
let hint = ZBUFFv04_decompressContinue(
dctx,
output.as_mut_ptr().cast::<c_void>(),
&mut output_size,
RAW_FRAME.as_ptr().cast::<c_void>(),
&mut input_size,
);
assert_eq!(ZBUFFv04_isError(hint), 0);
assert_eq!(&output[..output_size], expected);
assert_eq!(input_size, RAW_FRAME.len());
assert_eq!(ZBUFFv04_freeDCtx(dctx), 0);
assert_eq!(ZBUFFv04_freeDCtx(ptr::null_mut()), 0);
}
}
}