zstd/programs/fileio.c

1768 lines
68 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
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 */
# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */
2016-08-13 23:19:12 +00:00
# pragma warning(disable : 4204) /* non-constant aggregate initializer */
#endif
2016-12-21 12:23:34 +00:00
#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 "platform.h" /* Large Files support, SET_BINARY_MODE */
#include "util.h" /* UTIL_getFileSize */
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
2017-04-06 19:56:40 +00:00
#if defined (_MSC_VER)
# include <sys/stat.h>
# include <io.h>
#endif
2016-12-21 12:23:34 +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"
#ifdef ZSTD_MULTITHREAD
# include "zstdmt_compress.h"
2016-12-05 17:02:40 +00:00
#endif
2017-02-08 15:54:23 +00:00
#if defined(ZSTD_GZCOMPRESS) || defined(ZSTD_GZDECOMPRESS)
# include <zlib.h>
# if !defined(z_const)
# define z_const
# endif
2016-11-30 14:05:54 +00:00
#endif
#if defined(ZSTD_LZMACOMPRESS) || defined(ZSTD_LZMADECOMPRESS)
# include <lzma.h>
#endif
2015-01-24 00:58:16 +00:00
#define LZ4_MAGICNUMBER 0x184D2204
#if defined(ZSTD_LZ4COMPRESS) || defined(ZSTD_LZ4DECOMPRESS)
# include <lz4frame.h>
# include <lz4.h>
#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
* 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 DICTSIZE_MAX (32 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__)
2017-06-15 19:27:32 +00:00
#define DISPLAYOUT(...) fprintf(stdout, __VA_ARGS__)
#define DISPLAYLEVEL(l, ...) { if (g_displayLevel>=l) { DISPLAY(__VA_ARGS__); } }
static int 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(stderr); } } }
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;
2017-04-13 22:35:05 +00:00
#undef MIN
#define MIN(a,b) ((a) < (b) ? (a) : (b))
2016-07-02 23:10:53 +00:00
/* ************************************************************
* Avoid fseek()'s 2GiB barrier with MSVC, MacOS, *BSD, MinGW
***************************************************************/
#if defined(_MSC_VER) && _MSC_VER >= 1400
# define LONG_SEEK _fseeki64
#elif !defined(__64BIT__) && (PLATFORM_POSIX_VERSION >= 200112L) /* No point defining Large file for 64 bit */
2017-02-14 08:52:52 +00:00
# define LONG_SEEK fseeko
#elif defined(__MINGW32__) && !defined(__STRICT_ANSI__) && !defined(__NO_MINGW_LFS) && defined(__MSVCRT__)
# define LONG_SEEK fseeko64
#elif defined(_WIN32) && !defined(__DJGPP__)
# include <windows.h>
static int LONG_SEEK(FILE* file, __int64 offset, int origin) {
LARGE_INTEGER off;
DWORD method;
off.QuadPart = offset;
if (origin == SEEK_END)
method = FILE_END;
else if (origin == SEEK_CUR)
method = FILE_CURRENT;
else
method = FILE_BEGIN;
if (SetFilePointerEx((HANDLE) _get_osfhandle(_fileno(file)), off, NULL, method))
2017-02-13 11:00:59 +00:00
return 0;
else
2017-02-13 11:00:59 +00:00
return -1;
}
#else
# define LONG_SEEK fseek
#endif
/*-*************************************
2016-07-01 23:05:31 +00:00
* Local Parameters - Not thread safe
2015-10-22 15:55:40 +00:00
***************************************/
2017-02-14 08:23:32 +00:00
static FIO_compressionType_t g_compressionType = FIO_zstdCompression;
void FIO_setCompressionType(FIO_compressionType_t compressionType) { g_compressionType = compressionType; }
2015-01-24 00:58:16 +00:00
static U32 g_overwrite = 0;
void FIO_overwriteMode(void) { g_overwrite=1; }
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); }
static U32 g_memLimit = 0;
void FIO_setMemLimit(unsigned memLimit) { g_memLimit = memLimit; }
static U32 g_nbThreads = 1;
void FIO_setNbThreads(unsigned nbThreads) {
#ifndef ZSTD_MULTITHREAD
if (nbThreads > 1) DISPLAYLEVEL(2, "Note : multi-threading is disabled \n");
#endif
g_nbThreads = nbThreads;
}
static U32 g_blockSize = 0;
void FIO_setBlockSize(unsigned blockSize) {
if (blockSize && g_nbThreads==1)
DISPLAYLEVEL(2, "Setting block size is useless in single-thread mode \n");
#ifdef ZSTD_MULTITHREAD
if (blockSize-1 < ZSTDMT_SECTION_SIZE_MIN-1) /* intentional underflow */
DISPLAYLEVEL(2, "Note : minimum block size is %u KB \n", (ZSTDMT_SECTION_SIZE_MIN>>10));
#endif
g_blockSize = blockSize;
}
2017-01-30 22:37:08 +00:00
#define FIO_OVERLAP_LOG_NOTSET 9999
static U32 g_overlapLog = FIO_OVERLAP_LOG_NOTSET;
void FIO_setOverlapLog(unsigned overlapLog){
if (overlapLog && g_nbThreads==1)
DISPLAYLEVEL(2, "Setting overlapLog is useless in single-thread mode \n");
g_overlapLog = overlapLog;
}
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__); \
2016-11-12 01:26:54 +00:00
DISPLAYLEVEL(1, " \n"); \
2015-01-24 00:58:16 +00:00
exit(error); \
}
/*-*************************************
2015-01-24 00:58:16 +00:00
* Functions
2015-10-22 15:55:40 +00:00
***************************************/
2017-04-06 19:56:40 +00:00
/** FIO_remove() :
* @result : Unlink `fileName`, even if it's read-only */
static int FIO_remove(const char* path)
{
#if defined(_WIN32) || defined(WIN32)
/* windows doesn't allow remove read-only files,
* so try to make it writable first */
2017-04-06 19:56:40 +00:00
chmod(path, _S_IWRITE);
#endif
return remove(path);
}
2016-11-12 01:26:54 +00:00
/** FIO_openSrcFile() :
* condition : `dstFileName` must be non-NULL.
* @result : FILE* to `dstFileName`, or NULL if it fails */
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 {
if (!UTIL_isRegFile(srcFileName)) {
2017-01-25 12:11:26 +00:00
DISPLAYLEVEL(1, "zstd: %s is not a regular file -- ignored \n", srcFileName);
return NULL;
}
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;
}
2016-09-15 13:38:44 +00:00
/** FIO_openDstFile() :
* condition : `dstFileName` must be non-NULL.
* @result : FILE* to `dstFileName`, or NULL if it fails */
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 {
2017-03-31 22:20:50 +00:00
if (g_sparseFileSupport == 1) {
g_sparseFileSupport = ZSTD_SPARSE_DEFAULT;
}
2017-04-06 19:56:40 +00:00
if (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);
2017-04-06 19:56:40 +00:00
if (!g_overwrite) {
if (g_displayLevel <= 1) {
/* No interaction possible */
DISPLAY("zstd: %s already exists; not overwritten \n", dstFileName);
2016-07-01 23:05:31 +00:00
return NULL;
2015-12-29 10:57:15 +00:00
}
2017-04-06 19:56:40 +00:00
DISPLAY("zstd: %s already exists; do you wish to overwrite (y/N) ? ", dstFileName);
{ int ch = getchar();
if ((ch!='Y') && (ch!='y')) {
DISPLAY(" not overwritten \n");
return NULL;
}
while ((ch!=EOF) && (ch!='\n')) ch = getchar(); /* flush rest of input line */
}
}
/* need to unlink */
FIO_remove(dstFileName);
} }
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;
}
/*! FIO_createDictBuffer() :
* creates a buffer, pointed by `*bufferPtr`,
* loads `filename` content into it, up to DICTSIZE_MAX bytes.
* @return : loaded size
* if fileName==NULL, returns 0 and a NULL pointer
*/
static size_t FIO_createDictBuffer(void** bufferPtr, const char* fileName)
2015-12-17 19:30:14 +00:00
{
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);
if (fileSize > DICTSIZE_MAX) EXM_THROW(32, "Dictionary file %s is too large (> %u MB)", fileName, DICTSIZE_MAX >> 20); /* avoid extreme cases */
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 {
FILE* srcFile;
FILE* dstFile;
void* srcBuffer;
size_t srcBufferSize;
void* dstBuffer;
size_t dstBufferSize;
#ifdef ZSTD_MULTITHREAD
ZSTDMT_CCtx* cctx;
#else
2016-08-13 21:45:45 +00:00
ZSTD_CStream* cctx;
#endif
} cRess_t;
static cRess_t FIO_createCResources(const char* dictFileName, int cLevel,
U64 srcSize, int srcRegFile,
ZSTD_compressionParameters* comprParams) {
cRess_t ress;
2016-07-13 15:38:39 +00:00
memset(&ress, 0, sizeof(ress));
#ifdef ZSTD_MULTITHREAD
ress.cctx = ZSTDMT_createCCtx(g_nbThreads);
if (ress.cctx == NULL)
EXM_THROW(30, "zstd: allocation error : can't create ZSTD_CStream");
2017-01-30 22:37:08 +00:00
if ((cLevel==ZSTD_maxCLevel()) && (g_overlapLog==FIO_OVERLAP_LOG_NOTSET))
/* use complete window for overlap */
ZSTDMT_setMTCtxParameter(ress.cctx, ZSTDMT_p_overlapSectionLog, 9);
2017-01-30 22:37:08 +00:00
if (g_overlapLog != FIO_OVERLAP_LOG_NOTSET)
ZSTDMT_setMTCtxParameter(ress.cctx, ZSTDMT_p_overlapSectionLog, g_overlapLog);
#else
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");
#endif
2016-08-13 21:45:45 +00:00
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);
if (!ress.srcBuffer || !ress.dstBuffer)
EXM_THROW(31, "zstd: allocation error : not enough memory");
/* dictionary */
2016-09-15 13:38:44 +00:00
{ void* dictBuffer;
size_t const dictBuffSize = FIO_createDictBuffer(&dictBuffer, dictFileName); /* works with dictFileName==NULL */
if (dictFileName && (dictBuffer==NULL))
EXM_THROW(32, "zstd: allocation error : can't create dictBuffer");
2016-09-21 14:46:08 +00:00
{ ZSTD_parameters params = ZSTD_getParams(cLevel, srcSize, dictBuffSize);
params.fParams.contentSizeFlag = srcRegFile;
2016-09-15 13:38:44 +00:00
params.fParams.checksumFlag = g_checksumFlag;
params.fParams.noDictIDFlag = !g_dictIDFlag;
if (comprParams->windowLog) params.cParams.windowLog = comprParams->windowLog;
if (comprParams->chainLog) params.cParams.chainLog = comprParams->chainLog;
if (comprParams->hashLog) params.cParams.hashLog = comprParams->hashLog;
if (comprParams->searchLog) params.cParams.searchLog = comprParams->searchLog;
if (comprParams->searchLength) params.cParams.searchLength = comprParams->searchLength;
if (comprParams->targetLength) params.cParams.targetLength = comprParams->targetLength;
if (comprParams->strategy) params.cParams.strategy = (ZSTD_strategy) comprParams->strategy;
#ifdef ZSTD_MULTITHREAD
{ size_t const errorCode = ZSTDMT_initCStream_advanced(ress.cctx, dictBuffer, dictBuffSize, params, srcSize);
if (ZSTD_isError(errorCode)) EXM_THROW(33, "Error initializing CStream : %s", ZSTD_getErrorName(errorCode));
ZSTDMT_setMTCtxParameter(ress.cctx, ZSTDMT_p_sectionSize, g_blockSize);
#else
2016-09-21 14:46:08 +00:00
{ size_t const errorCode = ZSTD_initCStream_advanced(ress.cctx, dictBuffer, dictBuffSize, params, srcSize);
2016-09-15 13:38:44 +00:00
if (ZSTD_isError(errorCode)) EXM_THROW(33, "Error initializing CStream : %s", ZSTD_getErrorName(errorCode));
#endif
2016-09-15 13:38:44 +00:00
} }
free(dictBuffer);
}
return ress;
}
static void FIO_freeCResources(cRess_t ress)
{
free(ress.srcBuffer);
free(ress.dstBuffer);
#ifdef ZSTD_MULTITHREAD
ZSTDMT_freeCCtx(ress.cctx);
#else
2016-09-15 13:38:44 +00:00
ZSTD_freeCStream(ress.cctx); /* never fails */
#endif
}
2017-02-08 15:54:23 +00:00
#ifdef ZSTD_GZCOMPRESS
static unsigned long long FIO_compressGzFrame(cRess_t* ress,
const char* srcFileName, U64 const srcFileSize,
int compressionLevel, U64* readsize)
2017-02-08 15:54:23 +00:00
{
unsigned long long inFileSize = 0, outFileSize = 0;
z_stream strm;
2017-02-08 16:37:14 +00:00
int ret;
2017-02-08 15:54:23 +00:00
if (compressionLevel > Z_BEST_COMPRESSION)
compressionLevel = Z_BEST_COMPRESSION;
2017-02-08 15:54:23 +00:00
strm.zalloc = Z_NULL;
strm.zfree = Z_NULL;
strm.opaque = Z_NULL;
2017-02-08 16:37:14 +00:00
ret = deflateInit2(&strm, compressionLevel, Z_DEFLATED,
15 /* maxWindowLogSize */ + 16 /* gzip only */,
8, Z_DEFAULT_STRATEGY); /* see http://www.zlib.net/manual.html */
if (ret != Z_OK)
EXM_THROW(71, "zstd: %s: deflateInit2 error %d \n", srcFileName, ret);
2017-02-08 15:54:23 +00:00
2017-02-08 16:37:14 +00:00
strm.next_in = 0;
2017-02-27 12:21:05 +00:00
strm.avail_in = 0;
2017-02-08 15:54:23 +00:00
strm.next_out = (Bytef*)ress->dstBuffer;
strm.avail_out = (uInt)ress->dstBufferSize;
while (1) {
if (strm.avail_in == 0) {
size_t const inSize = fread(ress->srcBuffer, 1, ress->srcBufferSize, ress->srcFile);
if (inSize == 0) break;
inFileSize += inSize;
strm.next_in = (z_const unsigned char*)ress->srcBuffer;
strm.avail_in = (uInt)inSize;
}
ret = deflate(&strm, Z_NO_FLUSH);
if (ret != Z_OK)
EXM_THROW(72, "zstd: %s: deflate error %d \n", srcFileName, ret);
2017-02-08 15:54:23 +00:00
{ size_t const decompBytes = ress->dstBufferSize - strm.avail_out;
if (decompBytes) {
if (fwrite(ress->dstBuffer, 1, decompBytes, ress->dstFile) != decompBytes)
EXM_THROW(73, "Write error : cannot write to output file");
2017-02-08 15:54:23 +00:00
outFileSize += decompBytes;
strm.next_out = (Bytef*)ress->dstBuffer;
strm.avail_out = (uInt)ress->dstBufferSize;
}
}
if (!srcFileSize)
DISPLAYUPDATE(2, "\rRead : %u MB ==> %.2f%%",
(U32)(inFileSize>>20),
(double)outFileSize/inFileSize*100)
else
DISPLAYUPDATE(2, "\rRead : %u / %u MB ==> %.2f%%",
(U32)(inFileSize>>20), (U32)(srcFileSize>>20),
(double)outFileSize/inFileSize*100);
2017-02-08 15:54:23 +00:00
}
while (1) {
2017-02-08 16:37:14 +00:00
ret = deflate(&strm, Z_FINISH);
2017-02-08 15:54:23 +00:00
{ size_t const decompBytes = ress->dstBufferSize - strm.avail_out;
if (decompBytes) {
if (fwrite(ress->dstBuffer, 1, decompBytes, ress->dstFile) != decompBytes)
EXM_THROW(75, "Write error : cannot write to output file");
2017-02-08 15:54:23 +00:00
outFileSize += decompBytes;
strm.next_out = (Bytef*)ress->dstBuffer;
strm.avail_out = (uInt)ress->dstBufferSize;
} }
2017-02-08 15:54:23 +00:00
if (ret == Z_STREAM_END) break;
if (ret != Z_BUF_ERROR)
EXM_THROW(77, "zstd: %s: deflate error %d \n", srcFileName, ret);
2017-02-08 15:54:23 +00:00
}
2017-02-08 16:37:14 +00:00
ret = deflateEnd(&strm);
if (ret != Z_OK)
EXM_THROW(79, "zstd: %s: deflateEnd error %d \n", srcFileName, ret);
2017-02-08 15:54:23 +00:00
*readsize = inFileSize;
return outFileSize;
}
#endif
#ifdef ZSTD_LZMACOMPRESS
static unsigned long long FIO_compressLzmaFrame(cRess_t* ress,
const char* srcFileName, U64 const srcFileSize,
int compressionLevel, U64* readsize, int plain_lzma)
{
unsigned long long inFileSize = 0, outFileSize = 0;
lzma_stream strm = LZMA_STREAM_INIT;
lzma_action action = LZMA_RUN;
lzma_ret ret;
if (compressionLevel < 0) compressionLevel = 0;
if (compressionLevel > 9) compressionLevel = 9;
if (plain_lzma) {
lzma_options_lzma opt_lzma;
if (lzma_lzma_preset(&opt_lzma, compressionLevel))
EXM_THROW(71, "zstd: %s: lzma_lzma_preset error", srcFileName);
ret = lzma_alone_encoder(&strm, &opt_lzma); /* LZMA */
if (ret != LZMA_OK)
EXM_THROW(71, "zstd: %s: lzma_alone_encoder error %d", srcFileName, ret);
} else {
ret = lzma_easy_encoder(&strm, compressionLevel, LZMA_CHECK_CRC64); /* XZ */
if (ret != LZMA_OK)
EXM_THROW(71, "zstd: %s: lzma_easy_encoder error %d", srcFileName, ret);
}
strm.next_in = 0;
strm.avail_in = 0;
strm.next_out = ress->dstBuffer;
strm.avail_out = ress->dstBufferSize;
while (1) {
if (strm.avail_in == 0) {
size_t const inSize = fread(ress->srcBuffer, 1, ress->srcBufferSize, ress->srcFile);
if (inSize == 0) action = LZMA_FINISH;
inFileSize += inSize;
strm.next_in = ress->srcBuffer;
strm.avail_in = inSize;
}
ret = lzma_code(&strm, action);
if (ret != LZMA_OK && ret != LZMA_STREAM_END)
EXM_THROW(72, "zstd: %s: lzma_code encoding error %d", srcFileName, ret);
{ size_t const compBytes = ress->dstBufferSize - strm.avail_out;
if (compBytes) {
if (fwrite(ress->dstBuffer, 1, compBytes, ress->dstFile) != compBytes)
EXM_THROW(73, "Write error : cannot write to output file");
outFileSize += compBytes;
strm.next_out = ress->dstBuffer;
strm.avail_out = ress->dstBufferSize;
} }
if (!srcFileSize)
DISPLAYUPDATE(2, "\rRead : %u MB ==> %.2f%%",
(U32)(inFileSize>>20),
(double)outFileSize/inFileSize*100)
else
DISPLAYUPDATE(2, "\rRead : %u / %u MB ==> %.2f%%",
(U32)(inFileSize>>20), (U32)(srcFileSize>>20),
(double)outFileSize/inFileSize*100);
if (ret == LZMA_STREAM_END) break;
}
lzma_end(&strm);
*readsize = inFileSize;
return outFileSize;
}
#endif
#ifdef ZSTD_LZ4COMPRESS
static int FIO_LZ4_GetBlockSize_FromBlockId (int id) { return (1 << (8 + (2 * id))); }
static unsigned long long FIO_compressLz4Frame(cRess_t* ress,
const char* srcFileName, U64 const srcFileSize,
int compressionLevel, U64* readsize)
{
unsigned long long inFileSize = 0, outFileSize = 0;
LZ4F_preferences_t prefs;
LZ4F_compressionContext_t ctx;
LZ4F_errorCode_t const errorCode = LZ4F_createCompressionContext(&ctx, LZ4F_VERSION);
if (LZ4F_isError(errorCode))
EXM_THROW(31, "zstd: failed to create lz4 compression context");
memset(&prefs, 0, sizeof(prefs));
2017-04-26 17:17:38 +00:00
#if LZ4_VERSION_NUMBER <= 10600
#define LZ4F_blockIndependent blockIndependent
#define LZ4F_max4MB max4MB
#endif
prefs.autoFlush = 1;
prefs.compressionLevel = compressionLevel;
2017-04-26 17:17:38 +00:00
prefs.frameInfo.blockMode = LZ4F_blockIndependent; /* stick to defaults for lz4 cli */
prefs.frameInfo.blockSizeID = LZ4F_max4MB;
prefs.frameInfo.contentChecksumFlag = (contentChecksum_t)g_checksumFlag;
#if LZ4_VERSION_NUMBER >= 10600
prefs.frameInfo.contentSize = srcFileSize;
#endif
{
2017-04-26 17:17:38 +00:00
size_t blockSize = FIO_LZ4_GetBlockSize_FromBlockId(LZ4F_max4MB);
size_t readSize;
size_t headerSize = LZ4F_compressBegin(ctx, ress->dstBuffer, ress->dstBufferSize, &prefs);
if (LZ4F_isError(headerSize))
EXM_THROW(33, "File header generation failed : %s",
LZ4F_getErrorName(headerSize));
{ size_t const sizeCheck = fwrite(ress->dstBuffer, 1, headerSize, ress->dstFile);
if (sizeCheck!=headerSize) EXM_THROW(34, "Write error : cannot write header"); }
outFileSize += headerSize;
/* Read first block */
readSize = fread(ress->srcBuffer, (size_t)1, (size_t)blockSize, ress->srcFile);
inFileSize += readSize;
/* Main Loop */
while (readSize>0) {
size_t outSize;
/* Compress Block */
outSize = LZ4F_compressUpdate(ctx, ress->dstBuffer, ress->dstBufferSize, ress->srcBuffer, readSize, NULL);
if (LZ4F_isError(outSize))
EXM_THROW(35, "zstd: %s: lz4 compression failed : %s",
srcFileName, LZ4F_getErrorName(outSize));
outFileSize += outSize;
if (!srcFileSize)
DISPLAYUPDATE(2, "\rRead : %u MB ==> %.2f%%",
(U32)(inFileSize>>20),
(double)outFileSize/inFileSize*100)
else
DISPLAYUPDATE(2, "\rRead : %u / %u MB ==> %.2f%%",
(U32)(inFileSize>>20), (U32)(srcFileSize>>20),
(double)outFileSize/inFileSize*100);
/* Write Block */
{ size_t const sizeCheck = fwrite(ress->dstBuffer, 1, outSize, ress->dstFile);
if (sizeCheck!=outSize) EXM_THROW(36, "Write error : cannot write compressed block"); }
/* Read next block */
readSize = fread(ress->srcBuffer, (size_t)1, (size_t)blockSize, ress->srcFile);
inFileSize += readSize;
}
if (ferror(ress->srcFile)) EXM_THROW(37, "Error reading %s ", srcFileName);
/* End of Stream mark */
headerSize = LZ4F_compressEnd(ctx, ress->dstBuffer, ress->dstBufferSize, NULL);
if (LZ4F_isError(headerSize))
EXM_THROW(38, "zstd: %s: lz4 end of file generation failed : %s",
srcFileName, LZ4F_getErrorName(headerSize));
{ size_t const sizeCheck = fwrite(ress->dstBuffer, 1, headerSize, ress->dstFile);
if (sizeCheck!=headerSize) EXM_THROW(39, "Write error : cannot write end of stream"); }
outFileSize += headerSize;
}
*readsize = inFileSize;
LZ4F_freeCompressionContext(ctx);
return outFileSize;
}
#endif
/*! 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,
2017-02-08 15:54:23 +00:00
const char* dstFileName, const char* srcFileName, int compressionLevel)
{
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);
switch (g_compressionType) {
case FIO_zstdCompression:
break;
case FIO_gzipCompression:
2017-02-08 16:37:14 +00:00
#ifdef ZSTD_GZCOMPRESS
compressedfilesize = FIO_compressGzFrame(&ress, srcFileName, fileSize, compressionLevel, &readsize);
#else
(void)compressionLevel;
EXM_THROW(20, "zstd: %s: file cannot be compressed as gzip (zstd compiled without ZSTD_GZCOMPRESS) -- ignored \n",
srcFileName);
#endif
goto finish;
case FIO_xzCompression:
case FIO_lzmaCompression:
#ifdef ZSTD_LZMACOMPRESS
compressedfilesize = FIO_compressLzmaFrame(&ress, srcFileName, fileSize, compressionLevel, &readsize, g_compressionType==FIO_lzmaCompression);
2017-02-08 16:37:14 +00:00
#else
(void)compressionLevel;
EXM_THROW(20, "zstd: %s: file cannot be compressed as xz/lzma (zstd compiled without ZSTD_LZMACOMPRESS) -- ignored \n",
srcFileName);
#endif
goto finish;
case FIO_lz4Compression:
#ifdef ZSTD_LZ4COMPRESS
compressedfilesize = FIO_compressLz4Frame(&ress, srcFileName, fileSize, compressionLevel, &readsize);
#else
(void)compressionLevel;
EXM_THROW(20, "zstd: %s: file cannot be compressed as lz4 (zstd compiled without ZSTD_LZ4COMPRESS) -- ignored \n",
srcFileName);
2017-02-08 16:37:14 +00:00
#endif
goto finish;
2017-02-08 15:54:23 +00:00
}
/* init */
#ifdef ZSTD_MULTITHREAD
{ size_t const resetError = ZSTDMT_resetCStream(ress.cctx, fileSize);
#else
2016-09-15 13:38:44 +00:00
{ size_t const resetError = ZSTD_resetCStream(ress.cctx, fileSize);
#endif
if (ZSTD_isError(resetError))
EXM_THROW(21, "Error initializing compression : %s",
ZSTD_getErrorName(resetError));
2016-09-15 13:38:44 +00:00
}
/* 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;
2016-08-16 23:39:22 +00:00
{ ZSTD_inBuffer inBuff = { ress.srcBuffer, inSize, 0 };
while (inBuff.pos != inBuff.size) {
ZSTD_outBuffer outBuff = { ress.dstBuffer, ress.dstBufferSize, 0 };
#ifdef ZSTD_MULTITHREAD
size_t const result = ZSTDMT_compressStream(ress.cctx, &outBuff, &inBuff);
#else
size_t const result = ZSTD_compressStream(ress.cctx, &outBuff, &inBuff);
#endif
if (ZSTD_isError(result))
EXM_THROW(23, "Compression error : %s ", ZSTD_getErrorName(result));
/* Write compressed stream */
if (outBuff.pos) {
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;
} } }
if (g_nbThreads > 1) {
if (!fileSize)
DISPLAYUPDATE(2, "\rRead : %u MB", (U32)(readsize>>20))
else
DISPLAYUPDATE(2, "\rRead : %u / %u MB",
(U32)(readsize>>20), (U32)(fileSize>>20));
} else {
if (!fileSize)
DISPLAYUPDATE(2, "\rRead : %u MB ==> %.2f%%",
(U32)(readsize>>20),
(double)compressedfilesize/readsize*100)
else
DISPLAYUPDATE(2, "\rRead : %u / %u MB ==> %.2f%%",
(U32)(readsize>>20), (U32)(fileSize>>20),
(double)compressedfilesize/readsize*100);
}
}
/* End of Frame */
{ size_t result = 1;
while (result!=0) { /* note : is there any possibility of endless loop ? */
ZSTD_outBuffer outBuff = { ress.dstBuffer, ress.dstBufferSize, 0 };
#ifdef ZSTD_MULTITHREAD
result = ZSTDMT_endStream(ress.cctx, &outBuff);
#else
result = ZSTD_endStream(ress.cctx, &outBuff);
#endif
if (ZSTD_isError(result)) EXM_THROW(26, "Compression error during frame end : %s", ZSTD_getErrorName(result));
{ 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;
}
}
2017-02-08 15:54:23 +00:00
finish:
/* 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 compressionLevel)
{
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 */
2017-02-08 15:54:23 +00:00
result = FIO_compressFilename_internal(ress, dstFileName, srcFileName, compressionLevel);
fclose(ress.srcFile);
if (g_removeSrcFile /* --rm */ && !result && strcmp(srcFileName, stdinmark)) {
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 compressionLevel)
{
int result;
stat_t statbuf;
int stat_result = 0;
ress.dstFile = FIO_openDstFile(dstFileName);
2016-09-15 13:38:44 +00:00
if (ress.dstFile==NULL) return 1; /* could not open dstFileName */
if (strcmp (srcFileName, stdinmark) && UTIL_getFileStat(srcFileName, &statbuf))
stat_result = 1;
2017-02-08 15:54:23 +00:00
result = FIO_compressFilename_srcFile(ress, dstFileName, srcFileName, compressionLevel);
if (fclose(ress.dstFile)) { /* error closing dstFile */
DISPLAYLEVEL(1, "zstd: %s: %s \n", dstFileName, strerror(errno));
result=1;
}
if (result!=0) { /* remove operation artefact */
if (remove(dstFileName))
EXM_THROW(1, "zstd: %s: %s", dstFileName, strerror(errno));
}
else if (strcmp (dstFileName, stdoutmark) && stat_result)
UTIL_setFileStat(dstFileName, &statbuf);
return result;
}
2015-12-17 13:09:55 +00:00
int FIO_compressFilename(const char* dstFileName, const char* srcFileName,
const char* dictFileName, int compressionLevel, ZSTD_compressionParameters* comprParams)
2015-01-24 00:58:16 +00:00
{
clock_t const start = clock();
2016-09-21 14:46:08 +00:00
U64 const srcSize = UTIL_getFileSize(srcFileName);
int const regFile = UTIL_isRegFile(srcFileName);
2015-01-24 00:58:16 +00:00
cRess_t const ress = FIO_createCResources(dictFileName, compressionLevel, srcSize, regFile, comprParams);
2017-02-08 15:54:23 +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
}
typedef struct {
int numActualFrames;
int numSkippableFrames;
2017-06-14 20:26:19 +00:00
unsigned long long decompressedSize;
int decompUnavailable;
2017-06-14 20:26:19 +00:00
unsigned long long compressedSize;
int usesCheck;
} fileInfo_t;
/*
* Reads information from file, stores in *info
* if successful, returns 0, returns 1 for frame analysis error, returns 2 for file not compressed with zstd
* returns 3 for cases in which file could not be opened.
*/
2017-06-15 19:27:32 +00:00
static int getFileInfo(fileInfo_t* info, const char* inFileName){
2017-06-15 23:12:04 +00:00
int detectError = 0;
2017-06-15 19:27:32 +00:00
FILE* const srcFile = FIO_openSrcFile(inFileName);
if (srcFile == NULL) {
DISPLAY("Error: could not open source file %s\n", inFileName);
return 3;
}
2017-06-14 20:26:19 +00:00
info->compressedSize = (unsigned long long)UTIL_getFileSize(inFileName);
/* begin analyzing frame */
2017-06-16 01:08:22 +00:00
for ( ; ; ) {
2017-06-15 21:13:28 +00:00
BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX];
size_t const numBytesRead = fread(headerBuffer, 1, sizeof(headerBuffer), srcFile);
if (numBytesRead < ZSTD_frameHeaderSize_min) {
if (feof(srcFile) && numBytesRead == 0 && info->compressedSize > 0) {
2017-06-15 21:13:28 +00:00
break;
}
else if (feof(srcFile)) {
DISPLAY("Error: reached end of file with incomplete frame\n");
detectError = 2;
break;
}
else {
2017-06-15 21:13:28 +00:00
DISPLAY("Error: did not reach end of file but ran out of frames\n");
detectError = 1;
break;
2017-06-12 23:53:50 +00:00
}
2017-06-15 21:13:28 +00:00
}
2017-06-15 23:12:04 +00:00
{
U32 const magicNumber = MEM_readLE32(headerBuffer);
if (magicNumber == ZSTD_MAGICNUMBER) {
U64 const frameContentSize = ZSTD_getFrameContentSize(headerBuffer, numBytesRead);
if (frameContentSize == ZSTD_CONTENTSIZE_ERROR || frameContentSize == ZSTD_CONTENTSIZE_UNKNOWN) {
info->decompUnavailable = 1;
2017-06-13 17:06:49 +00:00
}
2017-06-15 23:12:04 +00:00
else {
info->decompressedSize += frameContentSize;
}
{
/* move to the end of the frame header */
size_t const headerSize = ZSTD_frameHeaderSize(headerBuffer, numBytesRead);
if (ZSTD_isError(headerSize)) {
DISPLAY("Error: could not determine frame header size\n");
detectError = 1;
break;
}
{
int const ret = fseek(srcFile, ((long)headerSize)-((long)numBytesRead), SEEK_CUR);
if (ret != 0) {
DISPLAY("Error: could not move to end of frame header\n");
detectError = 1;
break;
}
2017-06-15 21:13:28 +00:00
}
2017-06-15 23:12:04 +00:00
}
/* skip the rest of the blocks in the frame */
{
int lastBlock = 0;
2017-06-16 01:08:22 +00:00
do {
2017-06-15 23:12:04 +00:00
BYTE blockHeaderBuffer[3];
U32 blockHeader;
int blockSize;
2017-06-20 18:56:11 +00:00
size_t const readBytes = fread(blockHeaderBuffer, 1, 3, srcFile);
2017-06-15 23:12:04 +00:00
if (readBytes != 3) {
DISPLAY("There was a problem reading the block header\n");
detectError = 1;
break;
}
blockHeader = MEM_readLE24(blockHeaderBuffer);
lastBlock = blockHeader & 1;
2017-06-20 18:58:27 +00:00
blockSize = blockHeader >> 3;
2017-06-15 23:12:04 +00:00
{
int const ret = fseek(srcFile, blockSize, SEEK_CUR);
if (ret != 0) {
DISPLAY("Error: could not skip to end of block\n");
detectError = 1;
break;
}
}
} while (lastBlock != 1);
if (detectError) {
break;
}
}
{
/* check if checksum is used */
BYTE const frameHeaderDescriptor = headerBuffer[4];
int const contentChecksumFlag = (frameHeaderDescriptor & (1 << 2)) >> 2;
if (contentChecksumFlag) {
int const ret = fseek(srcFile, 4, SEEK_CUR);
info->usesCheck = 1;
if (ret != 0) {
2017-06-15 23:12:04 +00:00
DISPLAY("Error: could not skip past checksum\n");
detectError = 1;
break;
}
}
}
2017-06-15 23:12:04 +00:00
info->numActualFrames++;
2017-06-15 21:13:28 +00:00
}
2017-06-15 23:12:04 +00:00
else if (magicNumber == ZSTD_MAGIC_SKIPPABLE_START) {
BYTE frameSizeBuffer[4];
2017-06-16 00:03:59 +00:00
size_t const readBytes = fread(frameSizeBuffer, 1, 4, srcFile);
2017-06-15 23:12:04 +00:00
if (readBytes != 4) {
DISPLAY("There was an error reading skippable frame size");
detectError = 1;
break;
2017-06-13 17:06:49 +00:00
}
2017-06-15 23:12:04 +00:00
{
U32 const frameSize = MEM_readLE32(frameSizeBuffer);
2017-06-20 20:44:05 +00:00
int const ret = LONG_SEEK(srcFile, frameSize, SEEK_CUR);
if (ret != 0) {
2017-06-15 23:12:04 +00:00
DISPLAY("Error: could not find end of skippable frame\n");
detectError = 1;
break;
}
2017-06-13 17:06:49 +00:00
}
2017-06-15 23:12:04 +00:00
info->numSkippableFrames++;
}
else {
detectError = 2;
break;
}
}
}
2017-06-15 19:27:32 +00:00
fclose(srcFile);
return detectError;
}
2017-06-15 22:55:49 +00:00
static void displayInfo(const char* inFileName, fileInfo_t* info, int displayLevel){
2017-06-15 19:27:32 +00:00
double const compressedSizeMB = (double)info->compressedSize/(1 MB);
double const decompressedSizeMB = (double)info->decompressedSize/(1 MB);
2017-06-21 19:27:20 +00:00
double const ratio = (info->compressedSize == 0) ? 0 : ((double)info->decompressedSize)/info->compressedSize;
const char* const checkString = (info->usesCheck ? "XXH64" : "None");
2017-06-16 00:12:41 +00:00
if (displayLevel <= 2) {
if (!info->decompUnavailable) {
2017-06-15 21:13:28 +00:00
DISPLAYOUT("Skippable Non-Skippable Compressed Uncompressed Ratio Check Filename\n");
DISPLAYOUT("%9d %13d %7.2f MB %9.2f MB %5.3f %5s %s\n",
2017-06-13 17:06:49 +00:00
info->numSkippableFrames, info->numActualFrames, compressedSizeMB, decompressedSizeMB,
ratio, checkString, inFileName);
2017-06-13 17:06:49 +00:00
}
2017-06-15 22:25:20 +00:00
else {
DISPLAYOUT("Skippable Non-Skippable Compressed Check Filename\n");
DISPLAYOUT("%9d %13d %7.2f MB %5s %s\n",
2017-06-15 22:25:20 +00:00
info->numSkippableFrames, info->numActualFrames, compressedSizeMB, checkString, inFileName);
2017-06-13 17:06:49 +00:00
}
}
else{
2017-06-15 21:13:28 +00:00
DISPLAYOUT("# Zstandard Frames: %d\n", info->numActualFrames);
DISPLAYOUT("# Skippable Frames: %d\n", info->numSkippableFrames);
DISPLAYOUT("Compressed Size: %.2f MB (%llu B)\n", compressedSizeMB, info->compressedSize);
if (!info->decompUnavailable) {
2017-06-15 21:13:28 +00:00
DISPLAYOUT("Decompressed Size: %.2f MB (%llu B)\n", decompressedSizeMB, info->decompressedSize);
2017-06-21 19:27:20 +00:00
DISPLAYOUT("Ratio: %.4f\n", ratio);
2017-06-13 17:06:49 +00:00
}
DISPLAYOUT("Check: %s\n", checkString);
DISPLAYOUT("\n");
2017-06-13 17:06:49 +00:00
}
}
2017-06-20 20:26:25 +00:00
static int FIO_listFile(const char* inFileName, int displayLevel, unsigned fileNo, unsigned numFiles){
/* initialize info to avoid warnings */
fileInfo_t info;
memset(&info, 0, sizeof(info));
2017-06-20 20:26:25 +00:00
DISPLAYOUT("%s (%u/%u):\n", inFileName, fileNo, numFiles);
{
int const error = getFileInfo(&info, inFileName);
if (error == 1) {
/* display error, but provide output */
DISPLAY("An error occurred with getting file info\n");
}
else if (error == 2) {
DISPLAYOUT("File %s not compressed with zstd\n", inFileName);
if (displayLevel > 2) {
DISPLAYOUT("\n");
}
return 1;
}
else if (error == 3) {
/* error occurred with opening the file */
return 1;
}
displayInfo(inFileName, &info, displayLevel);
return error;
}
}
2015-01-24 00:58:16 +00:00
int FIO_listMultipleFiles(unsigned numFiles, const char** filenameTable, int displayLevel){
if (numFiles == 0) {
DISPLAYOUT("No files given\n");
return 0;
}
DISPLAYOUT("===========================================\n");
DISPLAYOUT("Printing information about compressed files\n");
DISPLAYOUT("===========================================\n");
DISPLAYOUT("Number of files listed: %u\n", numFiles);
{
int error = 0;
unsigned u;
for (u=0; u<numFiles;u++) {
2017-06-20 20:26:25 +00:00
error |= FIO_listFile(filenameTable[u], displayLevel, u+1, numFiles);
}
return error;
}
}
2015-12-17 13:09:55 +00:00
int FIO_compressMultipleFilenames(const char** inFileNamesTable, unsigned nbFiles,
const char* suffix,
const char* dictFileName, int compressionLevel,
ZSTD_compressionParameters* comprParams)
{
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;
2016-09-21 14:46:08 +00:00
U64 const srcSize = (nbFiles != 1) ? 0 : UTIL_getFileSize(inFileNamesTable[0]) ;
int const regFile = (nbFiles != 1) ? 0 : UTIL_isRegFile(inFileNamesTable[0]);
cRess_t ress = FIO_createCResources(dictFileName, compressionLevel, srcSize, regFile, comprParams);
/* init */
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++)
2017-02-08 15:54:23 +00:00
missed_files += FIO_compressFilename_srcFile(ress, stdoutmark, inFileNamesTable[u], compressionLevel);
if (fclose(ress.dstFile))
EXM_THROW(29, "Write error : cannot properly close stdout");
} else {
unsigned u;
for (u=0; u<nbFiles; u++) {
size_t const ifnSize = strlen(inFileNamesTable[u]);
if (dfnSize <= ifnSize+suffixSize+1) { /* resize name buffer */
free(dstFileName);
dfnSize = ifnSize + 20;
dstFileName = (char*)malloc(dfnSize);
if (!dstFileName)
EXM_THROW(30, "zstd: %s", strerror(errno));
}
strcpy(dstFileName, inFileNamesTable[u]);
strcat(dstFileName, suffix);
2017-02-08 15:54:23 +00:00
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;
2016-12-05 16:39:38 +00:00
size_t srcBufferLoaded;
2015-12-17 19:30:14 +00:00
size_t srcBufferSize;
void* dstBuffer;
size_t dstBufferSize;
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");
ZSTD_setDStreamParameter(ress.dctx, DStream_p_maxWindowSize, g_memLimit);
2016-08-13 21:45:45 +00:00
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 */
2016-09-14 15:26:59 +00:00
{ void* dictBuffer;
size_t const dictBufferSize = FIO_createDictBuffer(&dictBuffer, dictFileName);
2016-09-14 15:26:59 +00:00
size_t const initError = ZSTD_initDStream_usingDict(ress.dctx, dictBuffer, dictBufferSize);
if (ZSTD_isError(initError))
EXM_THROW(61, "ZSTD_initDStream_usingDict error : %s", ZSTD_getErrorName(initError));
2016-09-14 15:26:59 +00:00
free(dictBuffer);
}
2015-12-17 19:30:14 +00:00
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);
}
/** 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 = LONG_SEEK(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 = LONG_SEEK(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 = LONG_SEEK(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 = LONG_SEEK(file, storedSkips, SEEK_CUR);
if (seekResult != 0) EXM_THROW(69, "Final skip error (sparse file)");
{ 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");
} }
}
2016-12-05 16:39:38 +00:00
/** FIO_passThrough() : just copy input into output, for compatibility with gzip -df mode
@return : 0 (no error) */
2016-12-05 17:31:14 +00:00
static unsigned FIO_passThrough(FILE* foutput, FILE* finput, void* buffer, size_t bufferSize, size_t alreadyLoaded)
2016-12-05 16:39:38 +00:00
{
size_t const blockSize = MIN(64 KB, bufferSize);
size_t readFromInput = 1;
unsigned storedSkips = 0;
2016-12-05 17:31:14 +00:00
/* assumption : ress->srcBufferLoaded bytes already loaded and stored within buffer */
{ size_t const sizeCheck = fwrite(buffer, 1, alreadyLoaded, foutput);
if (sizeCheck != alreadyLoaded) EXM_THROW(50, "Pass-through write error"); }
2016-12-05 16:39:38 +00:00
while (readFromInput) {
readFromInput = fread(buffer, 1, blockSize, finput);
storedSkips = FIO_fwriteSparse(foutput, buffer, readFromInput, storedSkips);
}
FIO_fwriteSparseEnd(foutput, storedSkips);
return 0;
}
/** FIO_decompressFrame() :
@return : size of decoded frame, or an error code
*/
#define FIO_ERROR_ZSTD_DECODING ((unsigned long long)(-2))
2016-12-05 16:39:38 +00:00
unsigned long long FIO_decompressFrame(dRess_t* ress,
FILE* finput,
const char* srcFileName,
U64 alreadyDecoded)
2015-01-24 00:58:16 +00:00
{
2016-08-28 19:47:17 +00:00
U64 frameSize = 0;
U32 storedSkips = 0;
2015-01-24 00:58:16 +00:00
2016-12-05 16:39:38 +00:00
ZSTD_resetDStream(ress->dctx);
if (strlen(srcFileName)>20) srcFileName += strlen(srcFileName)-20; /* display last 20 characters */
/* Header loading (optional, saves one loop) */
2016-12-05 16:39:38 +00:00
{ size_t const toRead = 9;
if (ress->srcBufferLoaded < toRead)
ress->srcBufferLoaded += fread(((char*)ress->srcBuffer) + ress->srcBufferLoaded, 1, toRead - ress->srcBufferLoaded, finput);
}
/* Main decompression Loop */
2016-02-02 13:36:49 +00:00
while (1) {
2016-12-05 16:39:38 +00:00
ZSTD_inBuffer inBuff = { ress->srcBuffer, ress->srcBufferLoaded, 0 };
ZSTD_outBuffer outBuff= { ress->dstBuffer, ress->dstBufferSize, 0 };
size_t const readSizeHint = ZSTD_decompressStream(ress->dctx, &outBuff, &inBuff);
if (ZSTD_isError(readSizeHint)) {
DISPLAYLEVEL(1, "%s : Decoding error (36) : %s \n",
srcFileName, ZSTD_getErrorName(readSizeHint));
return FIO_ERROR_ZSTD_DECODING;
}
2015-11-25 13:42:45 +00:00
/* Write block */
storedSkips = FIO_fwriteSparse(ress->dstFile, ress->dstBuffer, outBuff.pos, storedSkips);
2016-08-16 23:39:22 +00:00
frameSize += outBuff.pos;
DISPLAYUPDATE(2, "\r%-20.20s : %u MB... ",
srcFileName, (U32)((alreadyDecoded+frameSize)>>20) );
2015-11-25 13:42:45 +00:00
2016-12-05 16:39:38 +00:00
if (inBuff.pos > 0) {
memmove(ress->srcBuffer, (char*)ress->srcBuffer + inBuff.pos, inBuff.size - inBuff.pos);
ress->srcBufferLoaded -= inBuff.pos;
}
2016-09-07 12:54:23 +00:00
if (readSizeHint == 0) break; /* end of frame */
if (inBuff.size != inBuff.pos) {
DISPLAYLEVEL(1, "%s : Decoding error (37) : should consume entire input \n",
srcFileName);
return FIO_ERROR_ZSTD_DECODING;
}
2015-01-24 00:58:16 +00:00
/* Fill input buffer */
2016-12-05 16:39:38 +00:00
{ size_t const toRead = MIN(readSizeHint, ress->srcBufferSize); /* support large skippable frames */
if (ress->srcBufferLoaded < toRead)
ress->srcBufferLoaded += fread((char*)ress->srcBuffer + ress->srcBufferLoaded,
1, toRead - ress->srcBufferLoaded, finput);
if (ress->srcBufferLoaded < toRead) {
DISPLAYLEVEL(1, "%s : Read error (39) : premature end \n",
srcFileName);
return FIO_ERROR_ZSTD_DECODING;
} } }
2015-01-24 00:58:16 +00:00
2016-12-05 16:39:38 +00:00
FIO_fwriteSparseEnd(ress->dstFile, storedSkips);
2015-11-25 13:42:45 +00:00
return frameSize;
2015-09-10 22:26:09 +00:00
}
2016-12-01 10:56:31 +00:00
#ifdef ZSTD_GZDECOMPRESS
static unsigned long long FIO_decompressGzFrame(dRess_t* ress,
FILE* srcFile, const char* srcFileName)
2016-12-02 12:50:29 +00:00
{
2016-12-02 20:40:57 +00:00
unsigned long long outFileSize = 0;
2016-12-02 14:01:31 +00:00
z_stream strm;
int flush = Z_NO_FLUSH;
2017-02-27 12:21:05 +00:00
int ret;
2016-12-02 20:40:57 +00:00
2016-12-02 14:01:31 +00:00
strm.zalloc = Z_NULL;
strm.zfree = Z_NULL;
strm.opaque = Z_NULL;
strm.next_in = 0;
2017-02-27 12:21:05 +00:00
strm.avail_in = 0;
/* see http://www.zlib.net/manual.html */
if (inflateInit2(&strm, 15 /* maxWindowLogSize */ + 16 /* gzip only */) != Z_OK)
return 0;
2016-12-02 20:40:57 +00:00
strm.next_out = (Bytef*)ress->dstBuffer;
strm.avail_out = (uInt)ress->dstBufferSize;
strm.avail_in = (uInt)ress->srcBufferLoaded;
2016-12-05 16:39:38 +00:00
strm.next_in = (z_const unsigned char*)ress->srcBuffer;
2016-12-01 10:56:31 +00:00
2016-12-02 14:01:31 +00:00
for ( ; ; ) {
2016-12-05 16:39:38 +00:00
if (strm.avail_in == 0) {
ress->srcBufferLoaded = fread(ress->srcBuffer, 1, ress->srcBufferSize, srcFile);
if (ress->srcBufferLoaded == 0) flush = Z_FINISH;
2016-12-05 16:39:38 +00:00
strm.next_in = (z_const unsigned char*)ress->srcBuffer;
strm.avail_in = (uInt)ress->srcBufferLoaded;
2016-12-02 12:11:39 +00:00
}
ret = inflate(&strm, flush);
if (ret == Z_BUF_ERROR) EXM_THROW(39, "zstd: %s: premature end", srcFileName);
if (ret != Z_OK && ret != Z_STREAM_END) {
DISPLAY("zstd: %s: inflate error %d \n", srcFileName, ret);
return 0;
}
2016-12-05 16:39:38 +00:00
{ size_t const decompBytes = ress->dstBufferSize - strm.avail_out;
2016-12-02 20:40:57 +00:00
if (decompBytes) {
if (fwrite(ress->dstBuffer, 1, decompBytes, ress->dstFile) != decompBytes)
EXM_THROW(31, "Write error : cannot write to output file");
2016-12-02 20:40:57 +00:00
outFileSize += decompBytes;
strm.next_out = (Bytef*)ress->dstBuffer;
strm.avail_out = (uInt)ress->dstBufferSize;
2016-12-05 16:39:38 +00:00
}
}
if (ret == Z_STREAM_END) break;
}
2016-12-01 10:56:31 +00:00
2016-12-05 16:39:38 +00:00
if (strm.avail_in > 0) memmove(ress->srcBuffer, strm.next_in, strm.avail_in);
ress->srcBufferLoaded = strm.avail_in;
2017-02-27 12:21:05 +00:00
ret = inflateEnd(&strm);
if (ret != Z_OK) EXM_THROW(32, "zstd: %s: inflateEnd error %d", srcFileName, ret);
return outFileSize;
}
#endif
#ifdef ZSTD_LZMADECOMPRESS
static unsigned long long FIO_decompressLzmaFrame(dRess_t* ress, FILE* srcFile, const char* srcFileName, int plain_lzma)
{
unsigned long long outFileSize = 0;
lzma_stream strm = LZMA_STREAM_INIT;
lzma_action action = LZMA_RUN;
lzma_ret ret;
strm.next_in = 0;
strm.avail_in = 0;
if (plain_lzma) {
ret = lzma_alone_decoder(&strm, UINT64_MAX); /* LZMA */
} else {
ret = lzma_stream_decoder(&strm, UINT64_MAX, 0); /* XZ */
}
if (ret != LZMA_OK)
EXM_THROW(71, "zstd: %s: lzma_alone_decoder/lzma_stream_decoder error %d",
srcFileName, ret);
strm.next_out = ress->dstBuffer;
strm.avail_out = ress->dstBufferSize;
strm.avail_in = ress->srcBufferLoaded;
strm.next_in = ress->srcBuffer;
for ( ; ; ) {
if (strm.avail_in == 0) {
ress->srcBufferLoaded = fread(ress->srcBuffer, 1, ress->srcBufferSize, srcFile);
if (ress->srcBufferLoaded == 0) action = LZMA_FINISH;
strm.next_in = ress->srcBuffer;
strm.avail_in = ress->srcBufferLoaded;
}
ret = lzma_code(&strm, action);
if (ret == LZMA_BUF_ERROR)
EXM_THROW(39, "zstd: %s: premature end", srcFileName);
if (ret != LZMA_OK && ret != LZMA_STREAM_END) {
DISPLAY("zstd: %s: lzma_code decoding error %d \n", srcFileName, ret);
return 0;
}
{ size_t const decompBytes = ress->dstBufferSize - strm.avail_out;
if (decompBytes) {
if (fwrite(ress->dstBuffer, 1, decompBytes, ress->dstFile) != decompBytes)
EXM_THROW(31, "Write error : cannot write to output file");
outFileSize += decompBytes;
strm.next_out = ress->dstBuffer;
strm.avail_out = ress->dstBufferSize;
} }
if (ret == LZMA_STREAM_END) break;
}
if (strm.avail_in > 0) memmove(ress->srcBuffer, strm.next_in, strm.avail_in);
ress->srcBufferLoaded = strm.avail_in;
lzma_end(&strm);
2016-12-02 14:01:31 +00:00
return outFileSize;
2016-12-01 10:56:31 +00:00
}
#endif
#ifdef ZSTD_LZ4DECOMPRESS
static unsigned long long FIO_decompressLz4Frame(dRess_t* ress,
FILE* srcFile, const char* srcFileName)
{
unsigned long long filesize = 0;
LZ4F_errorCode_t nextToLoad;
LZ4F_decompressionContext_t dCtx;
LZ4F_errorCode_t const errorCode = LZ4F_createDecompressionContext(&dCtx, LZ4F_VERSION);
if (LZ4F_isError(errorCode))
EXM_THROW(61, "zstd: failed to create lz4 decompression context");
/* Init feed with magic number (already consumed from FILE* sFile) */
{ size_t inSize = 4;
size_t outSize= 0;
MEM_writeLE32(ress->srcBuffer, LZ4_MAGICNUMBER);
nextToLoad = LZ4F_decompress(dCtx, ress->dstBuffer, &outSize, ress->srcBuffer, &inSize, NULL);
if (LZ4F_isError(nextToLoad))
EXM_THROW(62, "zstd: %s: lz4 header error : %s",
srcFileName, LZ4F_getErrorName(nextToLoad));
}
/* Main Loop */
for (;nextToLoad;) {
size_t readSize;
size_t pos = 0;
size_t decodedBytes = ress->dstBufferSize;
/* Read input */
if (nextToLoad > ress->srcBufferSize) nextToLoad = ress->srcBufferSize;
readSize = fread(ress->srcBuffer, 1, nextToLoad, srcFile);
if (!readSize) break; /* reached end of file or stream */
while ((pos < readSize) || (decodedBytes == ress->dstBufferSize)) { /* still to read, or still to flush */
/* Decode Input (at least partially) */
size_t remaining = readSize - pos;
decodedBytes = ress->dstBufferSize;
nextToLoad = LZ4F_decompress(dCtx, ress->dstBuffer, &decodedBytes, (char*)(ress->srcBuffer)+pos, &remaining, NULL);
if (LZ4F_isError(nextToLoad))
EXM_THROW(66, "zstd: %s: decompression error : %s",
srcFileName, LZ4F_getErrorName(nextToLoad));
pos += remaining;
/* Write Block */
if (decodedBytes) {
if (fwrite(ress->dstBuffer, 1, decodedBytes, ress->dstFile) != decodedBytes)
EXM_THROW(63, "Write error : cannot write to output file");
filesize += decodedBytes;
DISPLAYUPDATE(2, "\rDecompressed : %u MB ", (unsigned)(filesize>>20));
}
if (!nextToLoad) break;
}
}
/* can be out because readSize == 0, which could be an fread() error */
if (ferror(srcFile)) EXM_THROW(67, "zstd: %s: read error", srcFileName);
if (nextToLoad!=0) EXM_THROW(68, "zstd: %s: unfinished stream", srcFileName);
LZ4F_freeDecompressionContext(dCtx);
ress->srcBufferLoaded = 0; /* LZ4F will go to the frame boundary */
return filesize;
}
#endif
2016-12-01 10:56:31 +00:00
/** FIO_decompressSrcFile() :
Decompression `srcFileName` into `ress.dstFile`
@return : 0 : OK
1 : operation not started
*/
2016-12-02 23:18:57 +00:00
static int FIO_decompressSrcFile(dRess_t ress, const char* dstFileName, const char* srcFileName)
2015-10-18 21:18:32 +00:00
{
2016-12-02 12:50:29 +00:00
FILE* srcFile;
2016-07-26 14:44:09 +00:00
unsigned readSomething = 0;
2016-12-02 20:40:57 +00:00
unsigned long long filesize = 0;
2016-11-30 12:34:21 +00:00
if (UTIL_isDirectory(srcFileName)) {
DISPLAYLEVEL(1, "zstd: %s is a directory -- ignored \n", srcFileName);
return 1;
}
2016-11-30 12:34:21 +00:00
2016-12-02 12:50:29 +00:00
srcFile = FIO_openSrcFile(srcFileName);
if (srcFile==NULL) return 1;
2016-11-30 14:05:54 +00:00
2016-12-02 14:01:31 +00:00
/* for each frame */
for ( ; ; ) {
/* check magic number -> version */
size_t const toRead = 4;
const BYTE* buf = (const BYTE*)ress.srcBuffer;
2016-12-05 16:39:38 +00:00
if (ress.srcBufferLoaded < toRead)
ress.srcBufferLoaded += fread((char*)ress.srcBuffer + ress.srcBufferLoaded,
(size_t)1, toRead - ress.srcBufferLoaded, srcFile);
2016-12-05 16:39:38 +00:00
if (ress.srcBufferLoaded==0) {
if (readSomething==0) {
DISPLAY("zstd: %s: unexpected end of file \n", srcFileName);
fclose(srcFile);
return 1;
} /* srcFileName is empty */
break; /* no more input */
}
readSomething = 1; /* there is at least >= 4 bytes in srcFile */
if (ress.srcBufferLoaded < toRead) {
DISPLAY("zstd: %s: unknown header \n", srcFileName);
fclose(srcFile);
return 1;
}
if (buf[0] == 31 && buf[1] == 139) { /* gz magic number */
2016-12-02 14:01:31 +00:00
#ifdef ZSTD_GZDECOMPRESS
2016-12-05 16:39:38 +00:00
unsigned long long const result = FIO_decompressGzFrame(&ress, srcFile, srcFileName);
if (result == 0) return 1;
filesize += result;
2016-11-30 14:05:54 +00:00
#else
2016-12-02 20:40:57 +00:00
DISPLAYLEVEL(1, "zstd: %s: gzip file cannot be uncompressed (zstd compiled without ZSTD_GZDECOMPRESS) -- ignored \n", srcFileName);
return 1;
#endif
} else if ((buf[0] == 0xFD && buf[1] == 0x37) /* xz magic number */
|| (buf[0] == 0x5D && buf[1] == 0x00)) { /* lzma header (no magic number) */
#ifdef ZSTD_LZMADECOMPRESS
unsigned long long const result = FIO_decompressLzmaFrame(&ress, srcFile, srcFileName, buf[0] != 0xFD);
if (result == 0) return 1;
filesize += result;
#else
DISPLAYLEVEL(1, "zstd: %s: xz/lzma file cannot be uncompressed (zstd compiled without ZSTD_LZMADECOMPRESS) -- ignored \n", srcFileName);
return 1;
#endif
} else if (MEM_readLE32(buf) == LZ4_MAGICNUMBER) {
#ifdef ZSTD_LZ4DECOMPRESS
unsigned long long const result = FIO_decompressLz4Frame(&ress, srcFile, srcFileName);
if (result == 0) return 1;
filesize += result;
#else
DISPLAYLEVEL(1, "zstd: %s: lz4 file cannot be uncompressed (zstd compiled without ZSTD_LZ4DECOMPRESS) -- ignored \n", srcFileName);
return 1;
2016-11-30 14:05:54 +00:00
#endif
} else {
if (!ZSTD_isFrame(ress.srcBuffer, toRead)) {
2016-12-02 23:18:57 +00:00
if ((g_overwrite) && !strcmp (dstFileName, stdoutmark)) { /* pass-through mode */
unsigned const result = FIO_passThrough(ress.dstFile, srcFile,
ress.srcBuffer, ress.srcBufferSize, ress.srcBufferLoaded);
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);
fclose(srcFile);
return 1;
} }
{ unsigned long long const frameSize = FIO_decompressFrame(&ress, srcFile, srcFileName, filesize);
if (frameSize == FIO_ERROR_ZSTD_DECODING) {
fclose(srcFile);
return 1;
}
filesize += frameSize;
} }
} /* for each frame */
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
2016-12-02 12:50:29 +00:00
/* Close file */
if (fclose(srcFile)) EXM_THROW(33, "zstd: %s close error", srcFileName); /* error should never happen */
if (g_removeSrcFile /* --rm */ && strcmp(srcFileName, stdinmark)) {
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,
2016-12-02 20:40:57 +00:00
const char* dstFileName, const char* srcFileName)
{
int result;
stat_t statbuf;
int stat_result = 0;
2016-11-02 12:08:39 +00:00
ress.dstFile = FIO_openDstFile(dstFileName);
if (ress.dstFile==0) return 1;
2015-12-17 19:30:14 +00:00
if (strcmp (srcFileName, stdinmark) && UTIL_getFileStat(srcFileName, &statbuf))
stat_result = 1;
2016-12-02 23:18:57 +00:00
result = FIO_decompressSrcFile(ress, dstFileName, srcFileName);
if (fclose(ress.dstFile))
EXM_THROW(38, "Write error : cannot properly close %s", dstFileName);
2016-11-02 12:08:39 +00:00
2016-09-07 05:00:08 +00:00
if ( (result != 0)
&& strcmp(dstFileName, nulmark) /* special case : don't remove() /dev/null (#316) */
&& remove(dstFileName) )
2016-09-07 05:01:33 +00:00
result=1; /* don't do anything special if remove() fails */
else if (strcmp (dstFileName, stdoutmark) && stat_result)
UTIL_setFileStat(dstFileName, &statbuf);
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
if (suffix==NULL)
EXM_THROW(70, "zstd: decompression: unknown dst"); /* should never happen */
2016-07-13 17:30:40 +00:00
2016-12-02 23:18:57 +00:00
if (!strcmp(suffix, stdoutmark) || !strcmp(suffix, nulmark)) { /* special cases : -c or -t */
unsigned u;
ress.dstFile = FIO_openDstFile(suffix);
if (ress.dstFile == 0) EXM_THROW(71, "cannot open %s", suffix);
for (u=0; u<nbFiles; u++)
2016-12-02 23:18:57 +00:00
missingFiles += FIO_decompressSrcFile(ress, suffix, srcNamesTable[u]);
if (fclose(ress.dstFile))
EXM_THROW(72, "Write error : cannot properly close stdout");
} else {
size_t suffixSize;
size_t dfnSize = FNSPACE;
unsigned u;
char* dstFileName = (char*)malloc(FNSPACE);
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];
const char* const suffixPtr = strrchr(srcFileName, '.');
size_t const sfnSize = strlen(srcFileName);
if (!suffixPtr) {
DISPLAYLEVEL(1, "zstd: %s: unknown suffix -- ignored \n",
srcFileName);
skippedFiles++;
continue;
}
suffixSize = strlen(suffixPtr);
if (dfnSize+suffixSize <= sfnSize+1) {
free(dstFileName);
dfnSize = sfnSize + 20;
dstFileName = (char*)malloc(dfnSize);
if (dstFileName==NULL)
EXM_THROW(74, "not enough memory for dstFileName");
}
if (sfnSize <= suffixSize
|| (strcmp(suffixPtr, GZ_EXTENSION)
&& strcmp(suffixPtr, XZ_EXTENSION)
&& strcmp(suffixPtr, ZSTD_EXTENSION)
&& strcmp(suffixPtr, LZMA_EXTENSION)
&& strcmp(suffixPtr, LZ4_EXTENSION)) ) {
DISPLAYLEVEL(1, "zstd: %s: unknown suffix (%s/%s/%s/%s/%s expected) -- ignored \n",
srcFileName, GZ_EXTENSION, XZ_EXTENSION, ZSTD_EXTENSION, LZMA_EXTENSION, LZ4_EXTENSION);
skippedFiles++;
continue;
2016-11-30 12:34:21 +00:00
} else {
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 */