zstd/programs/fileio.c

805 lines
30 KiB
C
Raw Normal View History

2016-08-30 17:04:33 +00:00
/**
* Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
* All rights reserved.
*
* This source code is licensed under the BSD-style license found in the
* LICENSE file in the root directory of this source tree. An additional grant
* of patent rights can be found in the PATENTS file in the same directory.
*/
2015-01-24 00:58:16 +00:00
/*
2016-08-13 18:49:47 +00:00
Note : this file is part of zstd command line, which is not library.
2015-01-24 00:58:16 +00:00
The license of ZSTD library is BSD.
The license of this file is GPLv2.
*/
2015-10-18 21:18:32 +00:00
2016-08-28 17:00:49 +00:00
/* *************************************
* Tuning options
***************************************/
#ifndef ZSTD_LEGACY_SUPPORT
/* LEGACY_SUPPORT :
* decompressor can decode older formats (starting from Zstd 0.1+) */
# define ZSTD_LEGACY_SUPPORT 1
#endif
2015-10-22 15:55:40 +00:00
/* *************************************
2015-01-24 00:58:16 +00:00
* Compiler Options
2015-10-22 15:55:40 +00:00
***************************************/
2016-08-13 23:19:12 +00:00
#ifdef _MSC_VER /* Visual */
# define _CRT_SECURE_NO_WARNINGS /* removes Visual warning on strerror() */
# pragma warning(disable : 4204) /* non-constant aggregate initializer */
#endif
#if defined(__MINGW32__) && !defined(_POSIX_SOURCE)
# define _POSIX_SOURCE 1 /* disable %llu warnings with MinGW on Windows */
#endif
2015-01-24 00:58:16 +00:00
2016-02-12 19:19:48 +00:00
/*-*************************************
2015-01-24 00:58:16 +00:00
* Includes
2015-10-22 15:55:40 +00:00
***************************************/
#include "util.h" /* Compiler options, UTIL_GetFileSize, _LARGEFILE64_SOURCE */
2015-11-09 16:42:17 +00:00
#include <stdio.h> /* fprintf, fopen, fread, _fileno, stdin, stdout */
#include <stdlib.h> /* malloc, free */
#include <string.h> /* strcmp, strlen */
#include <time.h> /* clock */
#include <errno.h> /* errno */
2016-04-28 12:40:45 +00:00
#include "mem.h"
2015-01-24 00:58:16 +00:00
#include "fileio.h"
#define ZSTD_STATIC_LINKING_ONLY /* ZSTD_magicNumber, ZSTD_frameHeaderSize_max */
#include "zstd.h"
2015-01-24 00:58:16 +00:00
2016-08-28 17:00:49 +00:00
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT==1)
# include "zstd_legacy.h" /* ZSTD_isLegacy */
#endif
2015-01-24 00:58:16 +00:00
2016-02-12 19:19:48 +00:00
/*-*************************************
2015-01-24 00:58:16 +00:00
* OS-specific Includes
2015-10-22 15:55:40 +00:00
***************************************/
2015-01-24 00:58:16 +00:00
#if defined(MSDOS) || defined(OS2) || defined(WIN32) || defined(_WIN32) || defined(__CYGWIN__)
# include <fcntl.h> /* _O_BINARY */
# include <io.h> /* _setmode, _isatty */
# define SET_BINARY_MODE(file) { if (_setmode(_fileno(file), _O_BINARY) == -1) perror("Cannot set _O_BINARY"); }
2015-01-24 00:58:16 +00:00
#else
# include <unistd.h> /* isatty */
# define SET_BINARY_MODE(file)
#endif
2016-02-12 19:19:48 +00:00
/*-*************************************
2015-01-24 00:58:16 +00:00
* Constants
2015-10-22 15:55:40 +00:00
***************************************/
2016-07-02 23:10:53 +00:00
#define KB *(1<<10)
#define MB *(1<<20)
#define GB *(1U<<30)
2015-01-24 00:58:16 +00:00
#define _1BIT 0x01
#define _2BITS 0x03
#define _3BITS 0x07
#define _4BITS 0x0F
#define _6BITS 0x3F
#define _8BITS 0xFF
2015-11-25 13:42:45 +00:00
#define BLOCKSIZE (128 KB)
#define ROLLBUFFERSIZE (BLOCKSIZE*8*64)
2015-01-24 00:58:16 +00:00
2016-02-12 19:19:48 +00:00
#define FIO_FRAMEHEADERSIZE 5 /* as a define, because needed to allocated table on stack */
#define FSE_CHECKSUM_SEED 0
2015-01-24 00:58:16 +00:00
#define CACHELINE 64
#define MAX_DICT_SIZE (8 MB) /* protection against large input (attack scenario) */
2016-02-12 19:19:48 +00:00
#define FNSPACE 30
2015-01-24 00:58:16 +00:00
/*-*************************************
2015-01-24 00:58:16 +00:00
* Macros
2015-10-22 15:55:40 +00:00
***************************************/
2015-01-24 00:58:16 +00:00
#define DISPLAY(...) fprintf(stderr, __VA_ARGS__)
#define DISPLAYLEVEL(l, ...) if (g_displayLevel>=l) { DISPLAY(__VA_ARGS__); }
static U32 g_displayLevel = 2; /* 0 : no display; 1: errors; 2 : + result + interaction + warnings; 3 : + progression; 4 : + information */
2016-03-10 20:02:25 +00:00
void FIO_setNotificationLevel(unsigned level) { g_displayLevel=level; }
2015-01-24 00:58:16 +00:00
#define DISPLAYUPDATE(l, ...) if (g_displayLevel>=l) { \
2016-07-01 23:05:31 +00:00
if ((clock() - g_time > refreshRate) || (g_displayLevel>=4)) \
2015-01-24 00:58:16 +00:00
{ g_time = clock(); DISPLAY(__VA_ARGS__); \
if (g_displayLevel>=4) fflush(stdout); } }
2016-07-01 23:05:31 +00:00
static const clock_t refreshRate = CLOCKS_PER_SEC * 15 / 100;
2015-01-24 00:58:16 +00:00
static clock_t g_time = 0;
2016-07-02 23:10:53 +00:00
#define MIN(a,b) ((a) < (b) ? (a) : (b))
/*-*************************************
2016-07-01 23:05:31 +00:00
* Local Parameters - Not thread safe
2015-10-22 15:55:40 +00:00
***************************************/
2015-01-24 00:58:16 +00:00
static U32 g_overwrite = 0;
void FIO_overwriteMode(void) { g_overwrite=1; }
2016-03-10 20:02:25 +00:00
static U32 g_maxWLog = 23;
void FIO_setMaxWLog(unsigned maxWLog) { g_maxWLog = maxWLog; }
static U32 g_sparseFileSupport = 1; /* 0 : no sparse allowed; 1: auto (file yes, stdout no); 2: force sparse */
void FIO_setSparseWrite(unsigned sparse) { g_sparseFileSupport=sparse; }
static U32 g_dictIDFlag = 1;
void FIO_setDictIDFlag(unsigned dictIDFlag) { g_dictIDFlag = dictIDFlag; }
static U32 g_checksumFlag = 1;
void FIO_setChecksumFlag(unsigned checksumFlag) { g_checksumFlag = checksumFlag; }
static U32 g_removeSrcFile = 0;
void FIO_setRemoveSrcFile(unsigned flag) { g_removeSrcFile = (flag>0); }
2015-01-24 00:58:16 +00:00
/*-*************************************
2015-01-24 00:58:16 +00:00
* Exceptions
2015-10-22 15:55:40 +00:00
***************************************/
#ifndef DEBUG
# define DEBUG 0
#endif
2015-01-24 00:58:16 +00:00
#define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__);
#define EXM_THROW(error, ...) \
{ \
DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \
DISPLAYLEVEL(1, "Error %i : ", error); \
DISPLAYLEVEL(1, __VA_ARGS__); \
DISPLAYLEVEL(1, "\n"); \
exit(error); \
}
/*-*************************************
2015-01-24 00:58:16 +00:00
* Functions
2015-10-22 15:55:40 +00:00
***************************************/
static FILE* FIO_openSrcFile(const char* srcFileName)
{
FILE* f;
2016-02-02 13:36:49 +00:00
if (!strcmp (srcFileName, stdinmark)) {
DISPLAYLEVEL(4,"Using stdin for input\n");
f = stdin;
SET_BINARY_MODE(stdin);
2016-02-02 13:36:49 +00:00
} else {
f = fopen(srcFileName, "rb");
2016-08-13 18:49:47 +00:00
if ( f==NULL ) DISPLAYLEVEL(1, "zstd: %s: %s \n", srcFileName, strerror(errno));
}
return f;
}
2016-07-25 22:49:47 +00:00
/* `dstFileName must` be non-NULL */
static FILE* FIO_openDstFile(const char* dstFileName)
{
FILE* f;
2016-02-02 13:36:49 +00:00
if (!strcmp (dstFileName, stdoutmark)) {
DISPLAYLEVEL(4,"Using stdout for output\n");
f = stdout;
SET_BINARY_MODE(stdout);
if (g_sparseFileSupport==1) {
g_sparseFileSupport = 0;
DISPLAYLEVEL(4, "Sparse File Support is automatically disabled on stdout ; try --sparse \n");
}
2016-02-02 13:36:49 +00:00
} else {
if (!g_overwrite && strcmp (dstFileName, nulmark)) { /* Check if destination file already exists */
f = fopen( dstFileName, "rb" );
if (f != 0) { /* dest file exists, prompt for overwrite authorization */
fclose(f);
if (g_displayLevel <= 1) {
2015-12-17 19:30:14 +00:00
/* No interaction possible */
DISPLAY("zstd: %s already exists; not overwritten \n", dstFileName);
2016-07-01 23:05:31 +00:00
return NULL;
2015-12-17 19:30:14 +00:00
}
DISPLAY("zstd: %s already exists; do you wish to overwrite (y/N) ? ", dstFileName);
{ int ch = getchar();
2016-02-02 13:36:49 +00:00
if ((ch!='Y') && (ch!='y')) {
DISPLAY(" not overwritten \n");
2016-07-01 23:05:31 +00:00
return NULL;
2015-12-29 10:57:15 +00:00
}
while ((ch!=EOF) && (ch!='\n')) ch = getchar(); /* flush rest of input line */
2016-02-02 13:36:49 +00:00
} } }
f = fopen( dstFileName, "wb" );
2016-08-13 18:49:47 +00:00
if (f==NULL) DISPLAYLEVEL(1, "zstd: %s: %s\n", dstFileName, strerror(errno));
}
2016-07-01 23:05:31 +00:00
return f;
}
2016-03-10 20:02:25 +00:00
/*! FIO_loadFile() :
* creates a buffer, pointed by `*bufferPtr`,
2016-03-10 20:02:25 +00:00
* loads `filename` content into it,
* up to MAX_DICT_SIZE bytes.
* @return : loaded size
2015-12-17 19:30:14 +00:00
*/
static size_t FIO_loadFile(void** bufferPtr, const char* fileName)
{
FILE* fileHandle;
U64 fileSize;
*bufferPtr = NULL;
2016-02-02 13:36:49 +00:00
if (fileName == NULL) return 0;
2015-12-17 19:30:14 +00:00
DISPLAYLEVEL(4,"Loading %s as dictionary \n", fileName);
fileHandle = fopen(fileName, "rb");
2016-07-01 23:05:31 +00:00
if (fileHandle==0) EXM_THROW(31, "zstd: %s: %s", fileName, strerror(errno));
2016-04-28 10:23:33 +00:00
fileSize = UTIL_getFileSize(fileName);
2016-02-02 13:36:49 +00:00
if (fileSize > MAX_DICT_SIZE) {
2015-12-17 19:30:14 +00:00
int seekResult;
if (fileSize > 1 GB) EXM_THROW(32, "Dictionary file %s is too large", fileName); /* avoid extreme cases */
DISPLAYLEVEL(2,"Dictionary %s is too large : using last %u bytes only \n", fileName, (U32)MAX_DICT_SIZE);
2015-12-17 19:30:14 +00:00
seekResult = fseek(fileHandle, (long int)(fileSize-MAX_DICT_SIZE), SEEK_SET); /* use end of file */
2016-07-01 23:05:31 +00:00
if (seekResult != 0) EXM_THROW(33, "zstd: %s: %s", fileName, strerror(errno));
2015-12-17 19:30:14 +00:00
fileSize = MAX_DICT_SIZE;
}
2016-07-01 23:05:31 +00:00
*bufferPtr = malloc((size_t)fileSize);
2016-07-02 09:14:30 +00:00
if (*bufferPtr==NULL) EXM_THROW(34, "zstd: %s", strerror(errno));
{ size_t const readSize = fread(*bufferPtr, 1, (size_t)fileSize, fileHandle);
if (readSize!=fileSize) EXM_THROW(35, "Error reading dictionary file %s", fileName); }
2015-12-17 19:30:14 +00:00
fclose(fileHandle);
return (size_t)fileSize;
}
#ifndef ZSTD_NOCOMPRESS
2016-03-10 20:02:25 +00:00
/*-**********************************************************************
* Compression
************************************************************************/
typedef struct {
void* srcBuffer;
size_t srcBufferSize;
void* dstBuffer;
size_t dstBufferSize;
void* dictBuffer;
size_t dictBufferSize;
2016-08-13 21:45:45 +00:00
ZSTD_CStream* cctx;
FILE* dstFile;
FILE* srcFile;
} cRess_t;
static cRess_t FIO_createCResources(const char* dictFileName)
{
cRess_t ress;
2016-07-13 15:38:39 +00:00
memset(&ress, 0, sizeof(ress));
2016-08-13 21:45:45 +00:00
ress.cctx = ZSTD_createCStream();
if (ress.cctx == NULL) EXM_THROW(30, "zstd: allocation error : can't create ZSTD_CStream");
ress.srcBufferSize = ZSTD_CStreamInSize();
ress.srcBuffer = malloc(ress.srcBufferSize);
2016-08-13 21:45:45 +00:00
ress.dstBufferSize = ZSTD_CStreamOutSize();
ress.dstBuffer = malloc(ress.dstBufferSize);
2016-07-02 09:14:30 +00:00
if (!ress.srcBuffer || !ress.dstBuffer) EXM_THROW(31, "zstd: allocation error : not enough memory");
/* dictionary */
2015-12-17 19:30:14 +00:00
ress.dictBufferSize = FIO_loadFile(&(ress.dictBuffer), dictFileName);
return ress;
}
static void FIO_freeCResources(cRess_t ress)
{
size_t errorCode;
free(ress.srcBuffer);
free(ress.dstBuffer);
free(ress.dictBuffer);
2016-08-13 21:45:45 +00:00
errorCode = ZSTD_freeCStream(ress.cctx);
if (ZSTD_isError(errorCode)) EXM_THROW(38, "zstd: error : can't release ZSTD_CStream : %s", ZSTD_getErrorName(errorCode));
}
/*! FIO_compressFilename_internal() :
2016-03-10 20:02:25 +00:00
* same as FIO_compressFilename_extRess(), with `ress.desFile` already opened.
* @return : 0 : compression completed correctly,
* 1 : missing or pb opening srcFileName
*/
static int FIO_compressFilename_internal(cRess_t ress,
const char* dstFileName, const char* srcFileName,
int cLevel)
{
FILE* const srcFile = ress.srcFile;
FILE* const dstFile = ress.dstFile;
U64 readsize = 0;
U64 compressedfilesize = 0;
2016-04-28 10:23:33 +00:00
U64 const fileSize = UTIL_getFileSize(srcFileName);
/* init */
2016-06-27 13:28:45 +00:00
{ ZSTD_parameters params = ZSTD_getParams(cLevel, fileSize, ress.dictBufferSize);
params.fParams.contentSizeFlag = 1;
params.fParams.checksumFlag = g_checksumFlag;
params.fParams.noDictIDFlag = !g_dictIDFlag;
if ((g_maxWLog) && (params.cParams.windowLog > g_maxWLog)) {
params.cParams.windowLog = g_maxWLog;
params.cParams = ZSTD_adjustCParams(params.cParams, fileSize, ress.dictBufferSize);
}
2016-08-13 21:45:45 +00:00
{ size_t const errorCode = ZSTD_initCStream_advanced(ress.cctx, ress.dictBuffer, ress.dictBufferSize, params, fileSize);
if (ZSTD_isError(errorCode)) EXM_THROW(21, "Error initializing compression : %s", ZSTD_getErrorName(errorCode));
} }
/* Main compression loop */
2016-01-26 15:31:22 +00:00
while (1) {
/* Fill input Buffer */
size_t const inSize = fread(ress.srcBuffer, (size_t)1, ress.srcBufferSize, srcFile);
if (inSize==0) break;
readsize += inSize;
DISPLAYUPDATE(2, "\rRead : %u MB ", (U32)(readsize>>20));
/* Compress using buffered streaming */
2016-08-16 23:39:22 +00:00
{ ZSTD_inBuffer inBuff = { ress.srcBuffer, inSize, 0 };
ZSTD_outBuffer outBuff= { ress.dstBuffer, ress.dstBufferSize, 0 };
{ size_t const result = ZSTD_compressStream(ress.cctx, &outBuff, &inBuff);
2016-08-13 21:45:45 +00:00
if (ZSTD_isError(result)) EXM_THROW(23, "Compression error : %s ", ZSTD_getErrorName(result)); }
2016-08-16 23:39:22 +00:00
if (inBuff.pos != inBuff.size)
/* inBuff should be entirely consumed since buffer sizes are recommended ones */
EXM_THROW(24, "Compression error : input block not fully consumed");
/* Write cBlock */
2016-08-16 23:39:22 +00:00
{ size_t const sizeCheck = fwrite(ress.dstBuffer, 1, outBuff.pos, dstFile);
if (sizeCheck!=outBuff.pos) EXM_THROW(25, "Write error : cannot write compressed block into %s", dstFileName); }
compressedfilesize += outBuff.pos;
}
DISPLAYUPDATE(2, "\rRead : %u MB ==> %.2f%% ", (U32)(readsize>>20), (double)compressedfilesize/readsize*100);
}
/* End of Frame */
2016-08-16 23:39:22 +00:00
{ ZSTD_outBuffer outBuff = { ress.dstBuffer, ress.dstBufferSize, 0 };
size_t const result = ZSTD_endStream(ress.cctx, &outBuff);
if (result!=0) EXM_THROW(26, "Compression error : cannot create frame end");
2016-08-16 23:39:22 +00:00
{ size_t const sizeCheck = fwrite(ress.dstBuffer, 1, outBuff.pos, dstFile);
if (sizeCheck!=outBuff.pos) EXM_THROW(27, "Write error : cannot write frame end into %s", dstFileName); }
compressedfilesize += outBuff.pos;
}
/* Status */
DISPLAYLEVEL(2, "\r%79s\r", "");
DISPLAYLEVEL(2,"%-20s :%6.2f%% (%6llu => %6llu bytes, %s) \n", srcFileName,
2016-07-13 17:30:40 +00:00
(double)compressedfilesize/(readsize+(!readsize) /* avoid div by zero */ )*100,
(unsigned long long)readsize, (unsigned long long) compressedfilesize,
dstFileName);
return 0;
}
2016-07-01 23:05:31 +00:00
/*! FIO_compressFilename_srcFile() :
* note : ress.destFile already opened
* @return : 0 : compression completed correctly,
* 1 : missing or pb opening srcFileName
*/
static int FIO_compressFilename_srcFile(cRess_t ress,
const char* dstFileName, const char* srcFileName,
int cLevel)
{
int result;
/* File check */
if (UTIL_isDirectory(srcFileName)) {
DISPLAYLEVEL(1, "zstd: %s is a directory -- ignored \n", srcFileName);
return 1;
}
ress.srcFile = FIO_openSrcFile(srcFileName);
if (!ress.srcFile) return 1; /* srcFile could not be opened */
result = FIO_compressFilename_internal(ress, dstFileName, srcFileName, cLevel);
fclose(ress.srcFile);
2016-07-13 17:30:40 +00:00
if ((g_removeSrcFile) && (!result)) { if (remove(srcFileName)) EXM_THROW(1, "zstd: %s: %s", srcFileName, strerror(errno)); }
return result;
}
/*! FIO_compressFilename_dstFile() :
* @return : 0 : compression completed correctly,
* 1 : pb
*/
static int FIO_compressFilename_dstFile(cRess_t ress,
const char* dstFileName, const char* srcFileName,
int cLevel)
{
int result;
ress.dstFile = FIO_openDstFile(dstFileName);
2016-07-01 23:05:31 +00:00
if (ress.dstFile==0) return 1;
result = FIO_compressFilename_srcFile(ress, dstFileName, srcFileName, cLevel);
2016-07-01 23:05:31 +00:00
if (fclose(ress.dstFile)) { DISPLAYLEVEL(1, "zstd: %s: %s \n", dstFileName, strerror(errno)); result=1; }
2016-07-13 17:30:40 +00:00
if (result!=0) { if (remove(dstFileName)) EXM_THROW(1, "zstd: %s: %s", dstFileName, strerror(errno)); } /* remove operation artefact */
return result;
}
2015-12-17 13:09:55 +00:00
int FIO_compressFilename(const char* dstFileName, const char* srcFileName,
const char* dictFileName, int compressionLevel)
2015-01-24 00:58:16 +00:00
{
clock_t const start = clock();
2015-01-24 00:58:16 +00:00
cRess_t const ress = FIO_createCResources(dictFileName);
2016-07-01 23:05:31 +00:00
int const result = FIO_compressFilename_dstFile(ress, dstFileName, srcFileName, compressionLevel);
2015-01-24 00:58:16 +00:00
2016-07-01 23:05:31 +00:00
double const seconds = (double)(clock() - start) / CLOCKS_PER_SEC;
DISPLAYLEVEL(4, "Completed in %.2f sec \n", seconds);
FIO_freeCResources(ress);
return result;
2015-01-24 00:58:16 +00:00
}
2015-12-17 13:09:55 +00:00
int FIO_compressMultipleFilenames(const char** inFileNamesTable, unsigned nbFiles,
const char* suffix,
const char* dictFileName, int compressionLevel)
{
int missed_files = 0;
size_t dfnSize = FNSPACE;
2016-07-13 17:30:40 +00:00
char* dstFileName = (char*)malloc(FNSPACE);
size_t const suffixSize = suffix ? strlen(suffix) : 0;
2016-07-13 17:30:40 +00:00
cRess_t ress = FIO_createCResources(dictFileName);
/* init */
2016-07-13 17:30:40 +00:00
if (dstFileName==NULL) EXM_THROW(27, "FIO_compressMultipleFilenames : allocation error for dstFileName");
if (suffix == NULL) EXM_THROW(28, "FIO_compressMultipleFilenames : dst unknown"); /* should never happen */
/* loop on each file */
if (!strcmp(suffix, stdoutmark)) {
unsigned u;
ress.dstFile = stdout;
SET_BINARY_MODE(stdout);
for (u=0; u<nbFiles; u++)
missed_files += FIO_compressFilename_srcFile(ress, stdoutmark,
2016-03-10 20:02:25 +00:00
inFileNamesTable[u], compressionLevel);
if (fclose(ress.dstFile)) EXM_THROW(29, "Write error : cannot properly close %s", stdoutmark);
} else {
unsigned u;
for (u=0; u<nbFiles; u++) {
size_t ifnSize = strlen(inFileNamesTable[u]);
if (dfnSize <= ifnSize+suffixSize+1) { free(dstFileName); dfnSize = ifnSize + 20; dstFileName = (char*)malloc(dfnSize); }
strcpy(dstFileName, inFileNamesTable[u]);
strcat(dstFileName, suffix);
missed_files += FIO_compressFilename_dstFile(ress, dstFileName,
inFileNamesTable[u], compressionLevel);
} }
/* Close & Free */
FIO_freeCResources(ress);
free(dstFileName);
return missed_files;
}
#endif /* #ifndef ZSTD_NOCOMPRESS */
#ifndef ZSTD_NODECOMPRESS
/* **************************************************************************
* Decompression
****************************************************************************/
2015-12-17 19:30:14 +00:00
typedef struct {
void* srcBuffer;
size_t srcBufferSize;
void* dstBuffer;
size_t dstBufferSize;
void* dictBuffer;
size_t dictBufferSize;
2016-08-13 21:45:45 +00:00
ZSTD_DStream* dctx;
FILE* dstFile;
2015-12-17 19:30:14 +00:00
} dRess_t;
static dRess_t FIO_createDResources(const char* dictFileName)
{
dRess_t ress;
2016-07-13 15:38:39 +00:00
memset(&ress, 0, sizeof(ress));
2015-12-17 19:30:14 +00:00
2016-07-13 15:38:39 +00:00
/* Allocation */
2016-08-13 21:45:45 +00:00
ress.dctx = ZSTD_createDStream();
if (ress.dctx==NULL) EXM_THROW(60, "Can't create ZSTD_DStream");
ress.srcBufferSize = ZSTD_DStreamInSize();
2015-12-17 19:30:14 +00:00
ress.srcBuffer = malloc(ress.srcBufferSize);
2016-08-13 21:45:45 +00:00
ress.dstBufferSize = ZSTD_DStreamOutSize();
2015-12-17 19:30:14 +00:00
ress.dstBuffer = malloc(ress.dstBufferSize);
if (!ress.srcBuffer || !ress.dstBuffer) EXM_THROW(61, "Allocation error : not enough memory");
2015-12-17 19:30:14 +00:00
/* dictionary */
ress.dictBufferSize = FIO_loadFile(&(ress.dictBuffer), dictFileName);
return ress;
}
static void FIO_freeDResources(dRess_t ress)
{
2016-08-13 21:45:45 +00:00
size_t const errorCode = ZSTD_freeDStream(ress.dctx);
if (ZSTD_isError(errorCode)) EXM_THROW(69, "Error : can't free ZSTD_DStream context resource : %s", ZSTD_getErrorName(errorCode));
2015-12-17 19:30:14 +00:00
free(ress.srcBuffer);
free(ress.dstBuffer);
free(ress.dictBuffer);
}
/** FIO_fwriteSparse() :
* @return : storedSkips, to be provided to next call to FIO_fwriteSparse() of LZ4IO_fwriteSparseEnd() */
static unsigned FIO_fwriteSparse(FILE* file, const void* buffer, size_t bufferSize, unsigned storedSkips)
{
const size_t* const bufferT = (const size_t*)buffer; /* Buffer is supposed malloc'ed, hence aligned on size_t */
size_t bufferSizeT = bufferSize / sizeof(size_t);
const size_t* const bufferTEnd = bufferT + bufferSizeT;
const size_t* ptrT = bufferT;
static const size_t segmentSizeT = (32 KB) / sizeof(size_t); /* 0-test re-attempted every 32 KB */
if (!g_sparseFileSupport) { /* normal write */
size_t const sizeCheck = fwrite(buffer, 1, bufferSize, file);
if (sizeCheck != bufferSize) EXM_THROW(70, "Write error : cannot write decoded block");
return 0;
}
/* avoid int overflow */
if (storedSkips > 1 GB) {
int const seekResult = fseek(file, 1 GB, SEEK_CUR);
if (seekResult != 0) EXM_THROW(71, "1 GB skip error (sparse file support)");
storedSkips -= 1 GB;
}
while (ptrT < bufferTEnd) {
size_t seg0SizeT = segmentSizeT;
size_t nb0T;
/* count leading zeros */
if (seg0SizeT > bufferSizeT) seg0SizeT = bufferSizeT;
bufferSizeT -= seg0SizeT;
for (nb0T=0; (nb0T < seg0SizeT) && (ptrT[nb0T] == 0); nb0T++) ;
storedSkips += (unsigned)(nb0T * sizeof(size_t));
if (nb0T != seg0SizeT) { /* not all 0s */
int const seekResult = fseek(file, storedSkips, SEEK_CUR);
if (seekResult) EXM_THROW(72, "Sparse skip error ; try --no-sparse");
storedSkips = 0;
seg0SizeT -= nb0T;
ptrT += nb0T;
{ size_t const sizeCheck = fwrite(ptrT, sizeof(size_t), seg0SizeT, file);
if (sizeCheck != seg0SizeT) EXM_THROW(73, "Write error : cannot write decoded block");
} }
ptrT += seg0SizeT;
}
{ static size_t const maskT = sizeof(size_t)-1;
if (bufferSize & maskT) { /* size not multiple of sizeof(size_t) : implies end of block */
const char* const restStart = (const char*)bufferTEnd;
const char* restPtr = restStart;
size_t restSize = bufferSize & maskT;
const char* const restEnd = restStart + restSize;
for ( ; (restPtr < restEnd) && (*restPtr == 0); restPtr++) ;
storedSkips += (unsigned) (restPtr - restStart);
if (restPtr != restEnd) {
int seekResult = fseek(file, storedSkips, SEEK_CUR);
if (seekResult) EXM_THROW(74, "Sparse skip error ; try --no-sparse");
storedSkips = 0;
{ size_t const sizeCheck = fwrite(restPtr, 1, restEnd - restPtr, file);
if (sizeCheck != (size_t)(restEnd - restPtr)) EXM_THROW(75, "Write error : cannot write decoded end of block");
} } } }
return storedSkips;
}
static void FIO_fwriteSparseEnd(FILE* file, unsigned storedSkips)
{
if (storedSkips-->0) { /* implies g_sparseFileSupport>0 */
int const seekResult = fseek(file, storedSkips, SEEK_CUR);
if (seekResult != 0) EXM_THROW(69, "Final skip error (sparse file)\n");
{ const char lastZeroByte[1] = { 0 };
size_t const sizeCheck = fwrite(lastZeroByte, 1, 1, file);
if (sizeCheck != 1) EXM_THROW(69, "Write error : cannot write last zero\n");
} }
}
/** FIO_decompressFrame() :
@return : size of decoded frame
*/
2015-12-17 19:30:14 +00:00
unsigned long long FIO_decompressFrame(dRess_t ress,
FILE* foutput, FILE* finput, size_t alreadyLoaded)
2015-01-24 00:58:16 +00:00
{
2016-08-28 19:47:17 +00:00
U64 frameSize = 0;
size_t readSize;
U32 storedSkips = 0;
2015-01-24 00:58:16 +00:00
2016-08-13 21:45:45 +00:00
ZSTD_initDStream_usingDict(ress.dctx, ress.dictBuffer, ress.dictBufferSize);
/* Header loading (optional, saves one loop) */
{ size_t const toLoad = 9 - alreadyLoaded; /* assumption : 9 >= alreadyLoaded */
size_t const loadedSize = fread(((char*)ress.srcBuffer) + alreadyLoaded, 1, toLoad, finput);
2016-05-10 03:56:09 +00:00
readSize = alreadyLoaded + loadedSize;
}
/* Main decompression Loop */
2016-02-02 13:36:49 +00:00
while (1) {
2016-08-16 23:39:22 +00:00
ZSTD_inBuffer inBuff = { ress.srcBuffer, readSize, 0 };
ZSTD_outBuffer outBuff= { ress.dstBuffer, ress.dstBufferSize, 0 };
size_t const toRead = ZSTD_decompressStream(ress.dctx, &outBuff, &inBuff );
2016-08-13 21:45:45 +00:00
if (ZSTD_isError(toRead)) EXM_THROW(36, "Decoding error : %s", ZSTD_getErrorName(toRead));
2015-11-25 13:42:45 +00:00
/* Write block */
2016-08-16 23:39:22 +00:00
storedSkips = FIO_fwriteSparse(foutput, ress.dstBuffer, outBuff.pos, storedSkips);
frameSize += outBuff.pos;
2015-11-25 13:42:45 +00:00
DISPLAYUPDATE(2, "\rDecoded : %u MB... ", (U32)(frameSize>>20) );
if (toRead == 0) break; /* end of frame */
2016-08-16 23:39:22 +00:00
if (inBuff.size != inBuff.pos) EXM_THROW(37, "Decoding error : should consume entire input");
2015-01-24 00:58:16 +00:00
/* Fill input buffer */
2016-07-28 17:55:09 +00:00
if (toRead > ress.srcBufferSize) EXM_THROW(38, "too large block");
2015-12-17 19:30:14 +00:00
readSize = fread(ress.srcBuffer, 1, toRead, finput);
2016-07-28 17:55:09 +00:00
if (readSize == 0) EXM_THROW(39, "Read error : premature end");
2015-01-24 00:58:16 +00:00
}
FIO_fwriteSparseEnd(foutput, storedSkips);
2015-11-25 13:42:45 +00:00
return frameSize;
2015-09-10 22:26:09 +00:00
}
/** FIO_passThrough() : just copy input into output, for compatibility with gzip -df mode
@return : 0 (no error) */
2016-05-23 22:52:14 +00:00
static unsigned FIO_passThrough(FILE* foutput, FILE* finput, void* buffer, size_t bufferSize)
{
size_t const blockSize = MIN (64 KB, bufferSize);
2016-05-23 22:52:14 +00:00
size_t readFromInput = 1;
unsigned storedSkips = 0;
/* assumption : first 4 bytes already loaded (magic number detection), and stored within buffer */
{ size_t const sizeCheck = fwrite(buffer, 1, 4, foutput);
if (sizeCheck != 4) EXM_THROW(50, "Pass-through write error"); }
2016-05-23 22:52:14 +00:00
while (readFromInput) {
readFromInput = fread(buffer, 1, blockSize, finput);
storedSkips = FIO_fwriteSparse(foutput, buffer, readFromInput, storedSkips);
}
FIO_fwriteSparseEnd(foutput, storedSkips);
return 0;
}
/** FIO_decompressSrcFile() :
Decompression `srcFileName` into `ress.dstFile`
@return : 0 : OK
1 : operation not started
*/
static int FIO_decompressSrcFile(dRess_t ress, const char* srcFileName)
2015-10-18 21:18:32 +00:00
{
2015-12-17 19:30:14 +00:00
unsigned long long filesize = 0;
FILE* const dstFile = ress.dstFile;
FILE* srcFile;
2016-07-26 14:44:09 +00:00
unsigned readSomething = 0;
if (UTIL_isDirectory(srcFileName)) {
DISPLAYLEVEL(1, "zstd: %s is a directory -- ignored \n", srcFileName);
return 1;
}
srcFile = FIO_openSrcFile(srcFileName);
if (srcFile==0) return 1;
2015-11-25 13:42:45 +00:00
2015-10-18 21:18:32 +00:00
/* for each frame */
2016-02-02 13:36:49 +00:00
for ( ; ; ) {
/* check magic number -> version */
size_t const toRead = 4;
size_t const sizeCheck = fread(ress.srcBuffer, (size_t)1, toRead, srcFile);
if (sizeCheck==0) {
if (readSomething==0) { DISPLAY("zstd: %s: unexpected end of file \n", srcFileName); fclose(srcFile); return 1; } /* srcFileName is empty */
break; /* no more input */
}
2016-07-26 14:44:09 +00:00
readSomething = 1;
if (sizeCheck != toRead) { DISPLAY("zstd: %s: unknown header \n", srcFileName); fclose(srcFile); return 1; } /* srcFileName is empty */
{ U32 const magic = MEM_readLE32(ress.srcBuffer);
2016-08-28 17:00:49 +00:00
if (((magic & 0xFFFFFFF0U) != ZSTD_MAGIC_SKIPPABLE_START) & (magic != ZSTD_MAGICNUMBER)
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
& (!ZSTD_isLegacy(ress.srcBuffer, toRead))
#endif
) {
2016-07-28 17:55:09 +00:00
if ((g_overwrite) && !strcmp (srcFileName, stdinmark)) { /* pass-through mode */
2016-07-14 15:18:20 +00:00
unsigned const result = FIO_passThrough(dstFile, srcFile, ress.srcBuffer, ress.srcBufferSize);
if (fclose(srcFile)) EXM_THROW(32, "zstd: %s close error", srcFileName); /* error should never happen */
return result;
} else {
DISPLAYLEVEL(1, "zstd: %s: not in zstd format \n", srcFileName);
2016-07-13 15:38:39 +00:00
fclose(srcFile);
return 1;
} } }
2015-12-17 19:30:14 +00:00
filesize += FIO_decompressFrame(ress, dstFile, srcFile, toRead);
2015-10-18 21:18:32 +00:00
}
2015-12-17 19:30:14 +00:00
/* Final Status */
2015-10-18 21:18:32 +00:00
DISPLAYLEVEL(2, "\r%79s\r", "");
2016-08-28 19:47:17 +00:00
DISPLAYLEVEL(2, "%-20s: %llu bytes \n", srcFileName, filesize);
2015-10-18 21:18:32 +00:00
2015-12-17 19:30:14 +00:00
/* Close */
2016-07-14 15:18:20 +00:00
if (fclose(srcFile)) EXM_THROW(33, "zstd: %s close error", srcFileName); /* error should never happen */
if (g_removeSrcFile) { if (remove(srcFileName)) EXM_THROW(34, "zstd: %s: %s", srcFileName, strerror(errno)); };
return 0;
}
/** FIO_decompressFile_extRess() :
decompress `srcFileName` into `dstFileName`
@return : 0 : OK
1 : operation aborted (src not available, dst already taken, etc.)
*/
static int FIO_decompressDstFile(dRess_t ress,
const char* dstFileName, const char* srcFileName)
{
int result;
ress.dstFile = FIO_openDstFile(dstFileName);
if (ress.dstFile==0) return 1;
2015-12-17 19:30:14 +00:00
result = FIO_decompressSrcFile(ress, srcFileName);
if (fclose(ress.dstFile)) EXM_THROW(38, "Write error : cannot properly close %s", dstFileName);
2016-07-25 23:26:56 +00:00
if (result != 0) if (remove(dstFileName)) result=1; /* don't do anything if remove fails */
return result;
2015-10-18 21:18:32 +00:00
}
2015-12-17 19:30:14 +00:00
int FIO_decompressFilename(const char* dstFileName, const char* srcFileName,
const char* dictFileName)
{
int missingFiles = 0;
dRess_t ress = FIO_createDResources(dictFileName);
missingFiles += FIO_decompressDstFile(ress, dstFileName, srcFileName);
2015-12-17 19:30:14 +00:00
FIO_freeDResources(ress);
return missingFiles;
}
#define MAXSUFFIXSIZE 8
int FIO_decompressMultipleFilenames(const char** srcNamesTable, unsigned nbFiles,
const char* suffix,
const char* dictFileName)
{
int skippedFiles = 0;
int missingFiles = 0;
dRess_t ress = FIO_createDResources(dictFileName);
2015-12-17 19:30:14 +00:00
2016-07-13 17:30:40 +00:00
if (suffix==NULL) EXM_THROW(70, "zstd: decompression: unknown dst"); /* should never happen */
if (!strcmp(suffix, stdoutmark) || !strcmp(suffix, nulmark)) {
unsigned u;
ress.dstFile = FIO_openDstFile(suffix);
if (ress.dstFile == 0) EXM_THROW(71, "cannot open %s", suffix);
for (u=0; u<nbFiles; u++)
missingFiles += FIO_decompressSrcFile(ress, srcNamesTable[u]);
2016-07-13 17:30:40 +00:00
if (fclose(ress.dstFile)) EXM_THROW(72, "Write error : cannot properly close %s", stdoutmark);
} else {
2016-07-13 17:30:40 +00:00
size_t const suffixSize = strlen(suffix);
size_t dfnSize = FNSPACE;
unsigned u;
char* dstFileName = (char*)malloc(FNSPACE);
2016-07-13 17:30:40 +00:00
if (dstFileName==NULL) EXM_THROW(73, "not enough memory for dstFileName");
for (u=0; u<nbFiles; u++) { /* create dstFileName */
const char* const srcFileName = srcNamesTable[u];
size_t const sfnSize = strlen(srcFileName);
const char* const suffixPtr = srcFileName + sfnSize - suffixSize;
if (dfnSize+suffixSize <= sfnSize+1) {
free(dstFileName);
dfnSize = sfnSize + 20;
dstFileName = (char*)malloc(dfnSize);
2016-07-13 17:30:40 +00:00
if (dstFileName==NULL) EXM_THROW(74, "not enough memory for dstFileName");
}
if (sfnSize <= suffixSize || strcmp(suffixPtr, suffix) != 0) {
DISPLAYLEVEL(1, "zstd: %s: unknown suffix (%4s expected) -- ignored \n", srcFileName, suffix);
skippedFiles++;
continue;
}
memcpy(dstFileName, srcFileName, sfnSize - suffixSize);
dstFileName[sfnSize-suffixSize] = '\0';
missingFiles += FIO_decompressDstFile(ress, dstFileName, srcFileName);
}
free(dstFileName);
}
2015-12-17 19:30:14 +00:00
FIO_freeDResources(ress);
return missingFiles + skippedFiles;
}
#endif /* #ifndef ZSTD_NODECOMPRESS */