big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

HADOOP-11184. Update Hadoop's lz4 to r123 (cmccabe)

Browse Source
This commit is contained in:
Colin Patrick Mccabe 2014-10-09 17:05:35 -07:00
parent 596702a025
commit d8d628d1e0
12 changed files with 1678 additions and 1197 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
LICENSE.txt
View File

@ -252,13 +252,12 @@ in src/main/native/src/org/apache/hadoop/util:
* BSD-style license that can be found in the LICENSE file.
*/
For src/main/native/src/org/apache/hadoop/io/compress/lz4/{lz4.h,lz4.c,
lz4_encoder.h,lz4hc.h,lz4hc.c,lz4hc_encoder.h},
For src/main/native/src/org/apache/hadoop/io/compress/lz4/{lz4.h,lz4.c,lz4hc.h,lz4hc.c},
/*
LZ4 - Fast LZ compression algorithm
Header File
Copyright (C) 2011-2013, Yann Collet.
Copyright (C) 2011-2014, Yann Collet.
BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
Redistribution and use in source and binary forms, with or without
@ -285,8 +284,8 @@ lz4_encoder.h,lz4hc.h,lz4hc.c,lz4hc_encoder.h},
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
You can contact the author at :
- LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
- LZ4 source repository : http://code.google.com/p/lz4/
- LZ4 public forum : https://groups.google.com/forum/#!forum/lz4c
*/

2
hadoop-common-project/hadoop-common/CHANGES.txt
View File

@ -589,6 +589,8 @@ Release 2.6.0 - UNRELEASED
HADOOP-11174. Delegation token for KMS should only be got once if it
already exists. (Yi Liu via wang)
HADOOP-11184. Update Hadoop's lz4 to version r123. (cmccabe)
OPTIMIZATIONS
HADOOP-10838. Byte array native checksumming. (James Thomas via todd)

1
hadoop-common-project/hadoop-common/pom.xml
View File

@ -490,7 +490,6 @@
<exclude>src/test/resources/kdc/ldif/users.ldif</exclude>
<exclude>src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4.h</exclude>
<exclude>src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4.c</exclude>
<exclude>src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4_encoder.h</exclude>
<exclude>src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4hc.h</exclude>
<exclude>src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4hc.c</exclude>
<exclude>src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4hc_encoder.h</exclude>

1255
hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4.c
View File

File diff suppressed because it is too large Load Diff

304
hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4.h
View File

@ -1,7 +1,7 @@
/*
LZ4 - Fast LZ compression algorithm
Header File
Copyright (C) 2011-2013, Yann Collet.
Copyright (C) 2011-2014, Yann Collet.
BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
Redistribution and use in source and binary forms, with or without
@ -28,150 +28,294 @@
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
You can contact the author at :
- LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
- LZ4 source repository : http://code.google.com/p/lz4/
- LZ4 public forum : https://groups.google.com/forum/#!forum/lz4c
*/
/*
Copied from:
URL: http://lz4.googlecode.com/svn/trunk
Repository Root: http://lz4.googlecode.com/svn
Repository UUID: 650e7d94-2a16-8b24-b05c-7c0b3f6821cd
Revision: 99
Node Kind: directory
Schedule: normal
Last Changed Author: yann.collet.73@gmail.com
Last Changed Rev: 99
Last Changed Date: 2013-07-27 19:19:31 +0800 (Sat, 27 Jul 2013)
*/
#pragma once
#if defined (__cplusplus)
extern "C" {
#endif
/*
* lz4.h provides raw compression format functions, for optimal performance and integration into programs.
* If you need to generate data using an inter-operable format (respecting the framing specification),
* please use lz4frame.h instead.
*/
//**************************************
// Compiler Options
//**************************************
#if defined(_MSC_VER) && !defined(__cplusplus) // Visual Studio
# define inline __inline // Visual C is not C99, but supports some kind of inline
#endif
/**************************************
Version
**************************************/
#define LZ4_VERSION_MAJOR 1 /* for major interface/format changes */
#define LZ4_VERSION_MINOR 3 /* for minor interface/format changes */
#define LZ4_VERSION_RELEASE 1 /* for tweaks, bug-fixes, or development */
#define LZ4_VERSION_NUMBER (LZ4_VERSION_MAJOR *100*100 + LZ4_VERSION_MINOR *100 + LZ4_VERSION_RELEASE)
int LZ4_versionNumber (void);
/**************************************
Tuning parameter
**************************************/
/*
* LZ4_MEMORY_USAGE :
* Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.)
* Increasing memory usage improves compression ratio
* Reduced memory usage can improve speed, due to cache effect
* Default value is 14, for 16KB, which nicely fits into Intel x86 L1 cache
*/
#define LZ4_MEMORY_USAGE 14
//****************************
// Simple Functions
//****************************
/**************************************
Simple Functions
**************************************/
int LZ4_compress (const char* source, char* dest, int inputSize);
int LZ4_decompress_safe (const char* source, char* dest, int inputSize, int maxOutputSize);
int LZ4_compress (const char* source, char* dest, int sourceSize);
int LZ4_decompress_safe (const char* source, char* dest, int compressedSize, int maxDecompressedSize);
/*
LZ4_compress() :
Compresses 'inputSize' bytes from 'source' into 'dest'.
Compresses 'sourceSize' bytes from 'source' into 'dest'.
Destination buffer must be already allocated,
and must be sized to handle worst cases situations (input data not compressible)
Worst case size evaluation is provided by function LZ4_compressBound()
inputSize : Max supported value is ~1.9GB
inputSize : Max supported value is LZ4_MAX_INPUT_SIZE
return : the number of bytes written in buffer dest
or 0 if the compression fails
LZ4_decompress_safe() :
maxOutputSize : is the size of the destination buffer (which must be already allocated)
return : the number of bytes decoded in the destination buffer (necessarily <= maxOutputSize)
compressedSize : is obviously the source size
maxDecompressedSize : is the size of the destination buffer, which must be already allocated.
return : the number of bytes decompressed into the destination buffer (necessarily <= maxDecompressedSize)
If the destination buffer is not large enough, decoding will stop and output an error code (<0).
If the source stream is detected malformed, the function will stop decoding and return a negative result.
This function is protected against buffer overflow exploits (never writes outside of output buffer, and never reads outside of input buffer). Therefore, it is protected against malicious data packets
This function is protected against buffer overflow exploits,
and never writes outside of output buffer, nor reads outside of input buffer.
It is also protected against malicious data packets.
*/
//****************************
// Advanced Functions
//****************************
static inline int LZ4_compressBound(int isize) { return ((isize) + ((isize)/255) + 16); }
#define LZ4_COMPRESSBOUND( isize) ((isize) + ((isize)/255) + 16)
/**************************************
Advanced Functions
**************************************/
#define LZ4_MAX_INPUT_SIZE 0x7E000000 /* 2 113 929 216 bytes */
#define LZ4_COMPRESSBOUND(isize) ((unsigned int)(isize) > (unsigned int)LZ4_MAX_INPUT_SIZE ? 0 : (isize) + ((isize)/255) + 16)
/*
LZ4_compressBound() :
Provides the maximum size that LZ4 may output in a "worst case" scenario (input data not compressible)
primarily useful for memory allocation of output buffer.
inline function is recommended for the general case,
macro is also provided when result needs to be evaluated at compilation (such as table size allocation).
Provides the maximum size that LZ4 compression may output in a "worst case" scenario (input data not compressible)
This function is primarily useful for memory allocation purposes (output buffer size).
Macro LZ4_COMPRESSBOUND() is also provided for compilation-time evaluation (stack memory allocation for example).
isize : is the input size. Max supported value is ~1.9GB
isize : is the input size. Max supported value is LZ4_MAX_INPUT_SIZE
return : maximum output size in a "worst case" scenario
note : this function is limited by "int" range (2^31-1)
or 0, if input size is too large ( > LZ4_MAX_INPUT_SIZE)
*/
int LZ4_compressBound(int isize);
int LZ4_compress_limitedOutput (const char* source, char* dest, int inputSize, int maxOutputSize);
/*
LZ4_compress_limitedOutput() :
Compress 'inputSize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'.
Compress 'sourceSize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'.
If it cannot achieve it, compression will stop, and result of the function will be zero.
This saves time and memory on detecting non-compressible (or barely compressible) data.
This function never writes outside of provided output buffer.
inputSize : Max supported value is ~1.9GB
sourceSize : Max supported value is LZ4_MAX_INPUT_VALUE
maxOutputSize : is the size of the destination buffer (which must be already allocated)
return : the number of bytes written in buffer 'dest'
or 0 if the compression fails
or 0 if compression fails
*/
int LZ4_compress_limitedOutput (const char* source, char* dest, int sourceSize, int maxOutputSize);
int LZ4_decompress_fast (const char* source, char* dest, int outputSize);
/*
LZ4_compress_withState() :
Same compression functions, but using an externally allocated memory space to store compression state.
Use LZ4_sizeofState() to know how much memory must be allocated,
and then, provide it as 'void* state' to compression functions.
*/
int LZ4_sizeofState(void);
int LZ4_compress_withState (void* state, const char* source, char* dest, int inputSize);
int LZ4_compress_limitedOutput_withState (void* state, const char* source, char* dest, int inputSize, int maxOutputSize);
/*
LZ4_decompress_fast() :
outputSize : is the original (uncompressed) size
originalSize : is the original and therefore uncompressed size
return : the number of bytes read from the source buffer (in other words, the compressed size)
If the source stream is malformed, the function will stop decoding and return a negative result.
note : This function is a bit faster than LZ4_decompress_safe()
This function never writes outside of output buffers, and never read before input buffer, but may read beyond input buffer (since it doesn't know its size) in case of malicious data packet.
Use this function preferably into a trusted environment (data to decode comes from a trusted source).
Destination buffer must be already allocated. Its size must be a minimum of 'outputSize' bytes.
If the source stream is detected malformed, the function will stop decoding and return a negative result.
Destination buffer must be already allocated. Its size must be a minimum of 'originalSize' bytes.
note : This function fully respect memory boundaries for properly formed compressed data.
It is a bit faster than LZ4_decompress_safe().
However, it does not provide any protection against intentionally modified data stream (malicious input).
Use this function in trusted environment only (data to decode comes from a trusted source).
*/
int LZ4_decompress_fast (const char* source, char* dest, int originalSize);
int LZ4_decompress_safe_partial (const char* source, char* dest, int inputSize, int targetOutputSize, int maxOutputSize);
/*
LZ4_decompress_safe_partial() :
This function decompress a compressed block of size 'inputSize' at position 'source'
into output buffer 'dest' of size 'maxOutputSize'.
The function stops decompressing operation as soon as 'targetOutputSize' has been reached,
This function decompress a compressed block of size 'compressedSize' at position 'source'
into destination buffer 'dest' of size 'maxDecompressedSize'.
The function tries to stop decompressing operation as soon as 'targetOutputSize' has been reached,
reducing decompression time.
return : the number of bytes decoded in the destination buffer (necessarily <= maxOutputSize)
return : the number of bytes decoded in the destination buffer (necessarily <= maxDecompressedSize)
Note : this number can be < 'targetOutputSize' should the compressed block to decode be smaller.
Always control how many bytes were decoded.
If the source stream is malformed, the function will stop decoding and return a negative result.
If the source stream is detected malformed, the function will stop decoding and return a negative result.
This function never writes outside of output buffer, and never reads outside of input buffer. It is therefore protected against malicious data packets
*/
int LZ4_decompress_safe_partial (const char* source, char* dest, int compressedSize, int targetOutputSize, int maxDecompressedSize);
int LZ4_decompress_safe_withPrefix64k (const char* source, char* dest, int inputSize, int maxOutputSize);
int LZ4_decompress_fast_withPrefix64k (const char* source, char* dest, int outputSize);
/***********************************************
Experimental Streaming Compression Functions
***********************************************/
#define LZ4_STREAMSIZE_U32 ((1 << (LZ4_MEMORY_USAGE-2)) + 8)
#define LZ4_STREAMSIZE (LZ4_STREAMSIZE_U32 * sizeof(unsigned int))
/*
* LZ4_stream_t
* information structure to track an LZ4 stream.
* important : init this structure content before first use !
*/
typedef struct { unsigned int table[LZ4_STREAMSIZE_U32]; } LZ4_stream_t;
/*
*_withPrefix64k() :
These decoding functions work the same as their "normal name" versions,
but will potentially use up to 64KB of data in front of 'char* dest'.
These functions are used for decoding inter-dependant blocks.
*/
//****************************
// Obsolete Functions
//****************************
static inline int LZ4_uncompress (const char* source, char* dest, int outputSize) { return LZ4_decompress_fast(source, dest, outputSize); }
static inline int LZ4_uncompress_unknownOutputSize (const char* source, char* dest, int isize, int maxOutputSize) { return LZ4_decompress_safe(source, dest, isize, maxOutputSize); }
* LZ4_resetStream
* Use this function to init an allocated LZ4_stream_t structure
*/
void LZ4_resetStream (LZ4_stream_t* LZ4_streamPtr);
/*
These functions are deprecated and should no longer be used.
They are provided here for compatibility with existing user programs.
*/
* If you prefer dynamic allocation methods,
* LZ4_createStream will allocate and initialize an LZ4_stream_t structure
* LZ4_freeStream releases its memory.
*/
LZ4_stream_t* LZ4_createStream(void);
int LZ4_freeStream (LZ4_stream_t* LZ4_stream);
/*
* LZ4_loadDict
* Use this function to load a static dictionary into LZ4_stream.
* Any previous data will be forgotten, only 'dictionary' will remain in memory.
* Loading a size of 0 is allowed.
* Return : 1 if OK, 0 if error
*/
int LZ4_loadDict (LZ4_stream_t* LZ4_stream, const char* dictionary, int dictSize);
/*
* LZ4_compress_continue
* Compress data block 'source', using blocks compressed before as dictionary to improve compression ratio
* Previous data blocks are assumed to still be present at their previous location.
*/
int LZ4_compress_continue (LZ4_stream_t* LZ4_stream, const char* source, char* dest, int inputSize);
/*
* LZ4_compress_limitedOutput_continue
* Same as before, but also specify a maximum target compressed size (maxOutputSize)
* If objective cannot be met, compression exits, and returns a zero.
*/
int LZ4_compress_limitedOutput_continue (LZ4_stream_t* LZ4_stream, const char* source, char* dest, int inputSize, int maxOutputSize);
/*
* LZ4_saveDict
* If previously compressed data block is not guaranteed to remain available at its memory location
* save it into a safer place (char* safeBuffer)
* Note : you don't need to call LZ4_loadDict() afterwards,
* dictionary is immediately usable, you can therefore call again LZ4_compress_continue()
* Return : dictionary size in bytes, or 0 if error
* Note : any dictSize > 64 KB will be interpreted as 64KB.
*/
int LZ4_saveDict (LZ4_stream_t* LZ4_stream, char* safeBuffer, int dictSize);
/************************************************
Experimental Streaming Decompression Functions
************************************************/
#define LZ4_STREAMDECODESIZE_U32 4
#define LZ4_STREAMDECODESIZE (LZ4_STREAMDECODESIZE_U32 * sizeof(unsigned int))
/*
* LZ4_streamDecode_t
* information structure to track an LZ4 stream.
* important : init this structure content using LZ4_setStreamDecode or memset() before first use !
*/
typedef struct { unsigned int table[LZ4_STREAMDECODESIZE_U32]; } LZ4_streamDecode_t;
/*
* LZ4_setStreamDecode
* Use this function to instruct where to find the dictionary.
* This function can be used to specify a static dictionary,
* or to instruct where to find some previously decoded data saved into a different memory space.
* Setting a size of 0 is allowed (same effect as no dictionary).
* Return : 1 if OK, 0 if error
*/
int LZ4_setStreamDecode (LZ4_streamDecode_t* LZ4_streamDecode, const char* dictionary, int dictSize);
/*
* If you prefer dynamic allocation methods,
* LZ4_createStreamDecode will allocate and initialize an LZ4_streamDecode_t structure
* LZ4_freeStreamDecode releases its memory.
*/
LZ4_streamDecode_t* LZ4_createStreamDecode(void);
int LZ4_freeStreamDecode (LZ4_streamDecode_t* LZ4_stream);
/*
*_continue() :
These decoding functions allow decompression of multiple blocks in "streaming" mode.
Previously decoded blocks must still be available at the memory position where they were decoded.
If it's not possible, save the relevant part of decoded data into a safe buffer,
and indicate where its new address using LZ4_setStreamDecode()
*/
int LZ4_decompress_safe_continue (LZ4_streamDecode_t* LZ4_streamDecode, const char* source, char* dest, int compressedSize, int maxDecompressedSize);
int LZ4_decompress_fast_continue (LZ4_streamDecode_t* LZ4_streamDecode, const char* source, char* dest, int originalSize);
/*
Advanced decoding functions :
*_usingDict() :
These decoding functions work the same as
a combination of LZ4_setDictDecode() followed by LZ4_decompress_x_continue()
They don't use nor update an LZ4_streamDecode_t structure.
*/
int LZ4_decompress_safe_usingDict (const char* source, char* dest, int compressedSize, int maxDecompressedSize, const char* dictStart, int dictSize);
int LZ4_decompress_fast_usingDict (const char* source, char* dest, int originalSize, const char* dictStart, int dictSize);
/**************************************
Obsolete Functions
**************************************/
/*
Obsolete decompression functions
These function names are deprecated and should no longer be used.
They are only provided here for compatibility with older user programs.
- LZ4_uncompress is the same as LZ4_decompress_fast
- LZ4_uncompress_unknownOutputSize is the same as LZ4_decompress_safe
These function prototypes are now disabled; uncomment them if you really need them.
It is highly recommended to stop using these functions and migrated to newer ones */
/* int LZ4_uncompress (const char* source, char* dest, int outputSize); */
/* int LZ4_uncompress_unknownOutputSize (const char* source, char* dest, int isize, int maxOutputSize); */
/*
* If you prefer dynamic allocation methods,
* LZ4_createStreamDecode()
* provides a pointer (void*) towards an initialized LZ4_streamDecode_t structure.
* LZ4_free just frees it.
*/
/* void* LZ4_createStreamDecode(void); */
/*int LZ4_free (void* LZ4_stream); yes, it's the same one as for compression */
/* Obsolete streaming functions; use new streaming interface whenever possible */
void* LZ4_create (const char* inputBuffer);
int LZ4_sizeofStreamState(void);
int LZ4_resetStreamState(void* state, const char* inputBuffer);
char* LZ4_slideInputBuffer (void* state);
/* Obsolete streaming decoding functions */
int LZ4_decompress_safe_withPrefix64k (const char* source, char* dest, int compressedSize, int maxOutputSize);
int LZ4_decompress_fast_withPrefix64k (const char* source, char* dest, int originalSize);
#if defined (__cplusplus)

258
hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4_encoder.h
View File

@ -1,258 +0,0 @@
/*
LZ4 Encoder - Part of LZ4 compression algorithm
Copyright (C) 2011-2013, Yann Collet.
BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the following disclaimer
in the documentation and/or other materials provided with the
distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
You can contact the author at :
- LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
- LZ4 source repository : http://code.google.com/p/lz4/
*/
/* lz4_encoder.h must be included into lz4.c
The objective of this file is to create a single LZ4 compression function source
which will be instanciated multiple times with minor variations
depending on a set of #define.
*/
//****************************
// Check required defines
//****************************
#ifndef FUNCTION_NAME
# error "FUNTION_NAME is not defined"
#endif
//****************************
// Local definitions
//****************************
#ifdef COMPRESS_64K
# define HASHLOG (MEMORY_USAGE-1)
# define CURRENT_H_TYPE U16
# define CURRENTBASE(base) const BYTE* const base = ip
#else
# define HASHLOG (MEMORY_USAGE-2)
# define CURRENT_H_TYPE HTYPE
# define CURRENTBASE(base) INITBASE(base)
#endif
#define HASHTABLE_NBCELLS (1U<<HASHLOG)
#define LZ4_HASH(i) (((i) * 2654435761U) >> ((MINMATCH*8)-HASHLOG))
#define LZ4_HASHVALUE(p) LZ4_HASH(A32(p))
//****************************
// Function code
//****************************
int FUNCTION_NAME(
#ifdef USE_HEAPMEMORY
void* ctx,
#endif
const char* source,
char* dest,
int inputSize
#ifdef LIMITED_OUTPUT
,int maxOutputSize
#endif
)
{
#ifdef USE_HEAPMEMORY
CURRENT_H_TYPE* HashTable = (CURRENT_H_TYPE*)ctx;
#else
CURRENT_H_TYPE HashTable[HASHTABLE_NBCELLS] = {0};
#endif
const BYTE* ip = (BYTE*) source;
CURRENTBASE(base);
const BYTE* anchor = ip;
const BYTE* const iend = ip + inputSize;
const BYTE* const mflimit = iend - MFLIMIT;
#define matchlimit (iend - LASTLITERALS)
BYTE* op = (BYTE*) dest;
#ifdef LIMITED_OUTPUT
BYTE* const oend = op + maxOutputSize;
#endif
int length;
const int skipStrength = SKIPSTRENGTH;
U32 forwardH;
// Init
if (inputSize<MINLENGTH) goto _last_literals;
#ifdef COMPRESS_64K
if (inputSize>=LZ4_64KLIMIT) return 0; // Size too large (not within 64K limit)
#endif
#ifdef USE_HEAPMEMORY
memset((void*)HashTable, 0, HASHTABLESIZE);
#endif
// First Byte
HashTable[LZ4_HASHVALUE(ip)] = (CURRENT_H_TYPE)(ip - base);
ip++; forwardH = LZ4_HASHVALUE(ip);
// Main Loop
for ( ; ; )
{
int findMatchAttempts = (1U << skipStrength) + 3;
const BYTE* forwardIp = ip;
const BYTE* ref;
BYTE* token;
// Find a match
do {
U32 h = forwardH;
int step = findMatchAttempts++ >> skipStrength;
ip = forwardIp;
forwardIp = ip + step;
if unlikely(forwardIp > mflimit) { goto _last_literals; }
forwardH = LZ4_HASHVALUE(forwardIp);
ref = base + HashTable[h];
HashTable[h] = (CURRENT_H_TYPE)(ip - base);
} while ((ref < ip - MAX_DISTANCE) || (A32(ref) != A32(ip)));
// Catch up
while ((ip>anchor) && (ref>(BYTE*)source) && unlikely(ip[-1]==ref[-1])) { ip--; ref--; }
// Encode Literal length
length = (int)(ip - anchor);
token = op++;
#ifdef LIMITED_OUTPUT
if unlikely(op + length + (2 + 1 + LASTLITERALS) + (length>>8) > oend) return 0; // Check output limit
#endif
if (length>=(int)RUN_MASK)
{
int len = length-RUN_MASK;
*token=(RUN_MASK<<ML_BITS);
for(; len >= 255 ; len-=255) *op++ = 255;
*op++ = (BYTE)len;
}
else *token = (BYTE)(length<<ML_BITS);
// Copy Literals
LZ4_BLINDCOPY(anchor, op, length);
_next_match:
// Encode Offset
LZ4_WRITE_LITTLEENDIAN_16(op,(U16)(ip-ref));
// Start Counting
ip+=MINMATCH; ref+=MINMATCH; // MinMatch already verified
anchor = ip;
while likely(ip<matchlimit-(STEPSIZE-1))
{
size_t diff = AARCH(ref) ^ AARCH(ip);
if (!diff) { ip+=STEPSIZE; ref+=STEPSIZE; continue; }
ip += LZ4_NbCommonBytes(diff);
goto _endCount;
}
if (LZ4_ARCH64) if ((ip<(matchlimit-3)) && (A32(ref) == A32(ip))) { ip+=4; ref+=4; }
if ((ip<(matchlimit-1)) && (A16(ref) == A16(ip))) { ip+=2; ref+=2; }
if ((ip<matchlimit) && (*ref == *ip)) ip++;
_endCount:
// Encode MatchLength
length = (int)(ip - anchor);
#ifdef LIMITED_OUTPUT
if unlikely(op + (1 + LASTLITERALS) + (length>>8) > oend) return 0; // Check output limit
#endif
if (length>=(int)ML_MASK)
{
*token += ML_MASK;
length -= ML_MASK;
for (; length > 509 ; length-=510) { *op++ = 255; *op++ = 255; }
if (length >= 255) { length-=255; *op++ = 255; }
*op++ = (BYTE)length;
}
else *token += (BYTE)(length);
// Test end of chunk
if (ip > mflimit) { anchor = ip; break; }
// Fill table
HashTable[LZ4_HASHVALUE(ip-2)] = (CURRENT_H_TYPE)(ip - 2 - base);
// Test next position
ref = base + HashTable[LZ4_HASHVALUE(ip)];
HashTable[LZ4_HASHVALUE(ip)] = (CURRENT_H_TYPE)(ip - base);
if ((ref >= ip - MAX_DISTANCE) && (A32(ref) == A32(ip))) { token = op++; *token=0; goto _next_match; }
// Prepare next loop
anchor = ip++;
forwardH = LZ4_HASHVALUE(ip);
}
_last_literals:
// Encode Last Literals
{
int lastRun = (int)(iend - anchor);
#ifdef LIMITED_OUTPUT
if (((char*)op - dest) + lastRun + 1 + ((lastRun+255-RUN_MASK)/255) > (U32)maxOutputSize) return 0; // Check output limit
#endif
if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun >= 255 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
else *op++ = (BYTE)(lastRun<<ML_BITS);
memcpy(op, anchor, iend - anchor);
op += iend-anchor;
}
// End
return (int) (((char*)op)-dest);
}
//****************************
// Clean defines
//****************************
// Required defines
#undef FUNCTION_NAME
// Locally Generated
#undef HASHLOG
#undef HASHTABLE_NBCELLS
#undef LZ4_HASH
#undef LZ4_HASHVALUE
#undef CURRENT_H_TYPE
#undef CURRENTBASE
// Optional defines
#ifdef LIMITED_OUTPUT
#undef LIMITED_OUTPUT
#endif
#ifdef USE_HEAPMEMORY
#undef USE_HEAPMEMORY
#endif

567
hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4hc.c
View File

@ -1,6 +1,6 @@
/*
LZ4 HC - High Compression Mode of LZ4
Copyright (C) 2011-2013, Yann Collet.
Copyright (C) 2011-2014, Yann Collet.
BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
Redistribution and use in source and binary forms, with or without
@ -31,36 +31,45 @@
- LZ4 source repository : http://code.google.com/p/lz4/
*/
/*
Note : this source file requires "lz4hc_encoder.h"
*/
//**************************************
// Memory routines
//**************************************
#include <stdlib.h> // calloc, free
/**************************************
Tuning Parameter
**************************************/
#define LZ4HC_DEFAULT_COMPRESSIONLEVEL 8
/**************************************
Memory routines
**************************************/
#include <stdlib.h> /* calloc, free */
#define ALLOCATOR(s) calloc(1,s)
#define FREEMEM free
#include <string.h> // memset, memcpy
#include <string.h> /* memset, memcpy */
#define MEM_INIT memset
//**************************************
// CPU Feature Detection
//**************************************
// 32 or 64 bits ?
/**************************************
CPU Feature Detection
**************************************/
/* 32 or 64 bits ? */
#if (defined(__x86_64__) || defined(_M_X64) || defined(_WIN64) \
|| defined(__powerpc64__) || defined(__ppc64__) || defined(__PPC64__) \
|| defined(__64BIT__) || defined(_LP64) || defined(__LP64__) \
|| defined(__ia64) || defined(__itanium__) || defined(_M_IA64) ) // Detects 64 bits mode
|| defined(__64BIT__) || defined(__mips64) \
|| defined(__powerpc64__) || defined(__powerpc64le__) \
|| defined(__ppc64__) || defined(__ppc64le__) \
|| defined(__PPC64__) || defined(__PPC64LE__) \
|| defined(__ia64) || defined(__itanium__) || defined(_M_IA64) \
|| defined(__s390x__) ) /* Detects 64 bits mode */
# define LZ4_ARCH64 1
#else
# define LZ4_ARCH64 0
#endif
// Little Endian or Big Endian ?
// Overwrite the #define below if you know your architecture endianess
/*
* Little Endian or Big Endian ?
* Overwrite the #define below if you know your architecture endianess
*/
#include <stdlib.h> /* Apparently required to detect endianess */
#if defined (__GLIBC__)
# include <endian.h>
# if (__BYTE_ORDER == __BIG_ENDIAN)
@ -74,69 +83,71 @@ Note : this source file requires "lz4hc_encoder.h"
|| defined(_MIPSEB) || defined(__s390__)
# define LZ4_BIG_ENDIAN 1
#else
// Little Endian assumed. PDP Endian and other very rare endian format are unsupported.
/* Little Endian assumed. PDP Endian and other very rare endian format are unsupported. */
#endif
// Unaligned memory access is automatically enabled for "common" CPU, such as x86.
// For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected
// If you know your target CPU supports unaligned memory access, you want to force this option manually to improve performance
/*
* Unaligned memory access is automatically enabled for "common" CPU, such as x86.
* For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected
* If you know your target CPU supports unaligned memory access, you want to force this option manually to improve performance
*/
#if defined(__ARM_FEATURE_UNALIGNED)
# define LZ4_FORCE_UNALIGNED_ACCESS 1
#endif
// Define this parameter if your target system or compiler does not support hardware bit count
#if defined(_MSC_VER) && defined(_WIN32_WCE) // Visual Studio for Windows CE does not support Hardware bit count
/* Define this parameter if your target system or compiler does not support hardware bit count */
#if defined(_MSC_VER) && defined(_WIN32_WCE) /* Visual Studio for Windows CE does not support Hardware bit count */
# define LZ4_FORCE_SW_BITCOUNT
#endif
//**************************************
// Compiler Options
//**************************************
#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L // C99
/* "restrict" is a known keyword */
/**************************************
Compiler Options
**************************************/
#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */
/* "restrict" is a known keyword */
#else
# define restrict // Disable restrict
# define restrict /* Disable restrict */
#endif
#ifdef _MSC_VER // Visual Studio
# define forceinline static __forceinline
# include <intrin.h> // For Visual 2005
# if LZ4_ARCH64 // 64-bits
# pragma intrinsic(_BitScanForward64) // For Visual 2005
# pragma intrinsic(_BitScanReverse64) // For Visual 2005
# else // 32-bits
# pragma intrinsic(_BitScanForward) // For Visual 2005
# pragma intrinsic(_BitScanReverse) // For Visual 2005
#ifdef _MSC_VER /* Visual Studio */
# define FORCE_INLINE static __forceinline
# include <intrin.h> /* For Visual 2005 */
# if LZ4_ARCH64 /* 64-bits */
# pragma intrinsic(_BitScanForward64) /* For Visual 2005 */
# pragma intrinsic(_BitScanReverse64) /* For Visual 2005 */
# else /* 32-bits */
# pragma intrinsic(_BitScanForward) /* For Visual 2005 */
# pragma intrinsic(_BitScanReverse) /* For Visual 2005 */
# endif
# pragma warning(disable : 4127) // disable: C4127: conditional expression is constant
# pragma warning(disable : 4701) // disable: C4701: potentially uninitialized local variable used
# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */
# pragma warning(disable : 4701) /* disable: C4701: potentially uninitialized local variable used */
#else
# ifdef __GNUC__
# define forceinline static inline __attribute__((always_inline))
# define FORCE_INLINE static inline __attribute__((always_inline))
# else
# define forceinline static inline
# define FORCE_INLINE static inline
# endif
#endif
#ifdef _MSC_VER // Visual Studio
#ifdef _MSC_VER /* Visual Studio */
# define lz4_bswap16(x) _byteswap_ushort(x)
#else
# define lz4_bswap16(x) ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))
#endif
//**************************************
// Includes
//**************************************
/**************************************
Includes
**************************************/
#include "lz4hc.h"
#include "lz4.h"
//**************************************
// Basic Types
//**************************************
#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L // C99
/**************************************
Basic Types
**************************************/
#if defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */
# include <stdint.h>
typedef uint8_t BYTE;
typedef uint16_t U16;
@ -178,9 +189,9 @@ typedef struct _U64_S { U64 v; } _PACKED U64_S;
#define A16(x) (((U16_S *)(x))->v)
//**************************************
// Constants
//**************************************
/**************************************
Constants
**************************************/
#define MINMATCH 4
#define DICTIONARY_LOGSIZE 16
@ -192,8 +203,6 @@ typedef struct _U64_S { U64 v; } _PACKED U64_S;
#define HASHTABLESIZE (1 << HASH_LOG)
#define HASH_MASK (HASHTABLESIZE - 1)
#define MAX_NB_ATTEMPTS 256
#define ML_BITS 4
#define ML_MASK (size_t)((1U<<ML_BITS)-1)
#define RUN_BITS (8-ML_BITS)
@ -210,25 +219,21 @@ typedef struct _U64_S { U64 v; } _PACKED U64_S;
#define GB *(1U<<30)
//**************************************
// Architecture-specific macros
//**************************************
#if LZ4_ARCH64 // 64-bit
/**************************************
Architecture-specific macros
**************************************/
#if LZ4_ARCH64 /* 64-bit */
# define STEPSIZE 8
# define LZ4_COPYSTEP(s,d) A64(d) = A64(s); d+=8; s+=8;
# define LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d)
# define UARCH U64
# define AARCH A64
# define HTYPE U32
# define INITBASE(b,s) const BYTE* const b = s
#else // 32-bit
#else /* 32-bit */
# define STEPSIZE 4
# define LZ4_COPYSTEP(s,d) A32(d) = A32(s); d+=4; s+=4;
# define LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d); LZ4_COPYSTEP(s,d);
# define UARCH U32
# define AARCH A32
//# define HTYPE const BYTE*
//# define INITBASE(b,s) const int b = 0
# define HTYPE U32
# define INITBASE(b,s) const BYTE* const b = s
#endif
@ -236,15 +241,15 @@ typedef struct _U64_S { U64 v; } _PACKED U64_S;
#if defined(LZ4_BIG_ENDIAN)
# define LZ4_READ_LITTLEENDIAN_16(d,s,p) { U16 v = A16(p); v = lz4_bswap16(v); d = (s) - v; }
# define LZ4_WRITE_LITTLEENDIAN_16(p,i) { U16 v = (U16)(i); v = lz4_bswap16(v); A16(p) = v; p+=2; }
#else // Little Endian
#else /* Little Endian */
# define LZ4_READ_LITTLEENDIAN_16(d,s,p) { d = (s) - A16(p); }
# define LZ4_WRITE_LITTLEENDIAN_16(p,v) { A16(p) = v; p+=2; }
#endif
//************************************************************
// Local Types
//************************************************************
/**************************************
Local Types
**************************************/
typedef struct
{
const BYTE* inputBuffer;
@ -256,9 +261,9 @@ typedef struct
} LZ4HC_Data_Structure;
//**************************************
// Macros
//**************************************
/**************************************
Macros
**************************************/
#define LZ4_WILDCOPY(s,d,e) do { LZ4_COPYPACKET(s,d) } while (d<e);
#define LZ4_BLINDCOPY(s,d,l) { BYTE* e=d+l; LZ4_WILDCOPY(s,d,e); d=e; }
#define HASH_FUNCTION(i) (((i) * 2654435761U) >> ((MINMATCH*8)-HASH_LOG))
@ -268,12 +273,12 @@ typedef struct
#define GETNEXT(p) ((p) - (size_t)DELTANEXT(p))
//**************************************
// Private functions
//**************************************
/**************************************
Private functions
**************************************/
#if LZ4_ARCH64
forceinline int LZ4_NbCommonBytes (register U64 val)
FORCE_INLINE int LZ4_NbCommonBytes (register U64 val)
{
#if defined(LZ4_BIG_ENDIAN)
# if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
@ -305,7 +310,7 @@ forceinline int LZ4_NbCommonBytes (register U64 val)
#else
forceinline int LZ4_NbCommonBytes (register U32 val)
FORCE_INLINE int LZ4_NbCommonBytes (register U32 val)
{
#if defined(LZ4_BIG_ENDIAN)
# if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
@ -337,7 +342,12 @@ forceinline int LZ4_NbCommonBytes (register U32 val)
#endif
forceinline int LZ4_InitHC (LZ4HC_Data_Structure* hc4, const BYTE* base)
int LZ4_sizeofStreamStateHC()
{
return sizeof(LZ4HC_Data_Structure);
}
FORCE_INLINE void LZ4_initHC (LZ4HC_Data_Structure* hc4, const BYTE* base)
{
MEM_INIT((void*)hc4->hashTable, 0, sizeof(hc4->hashTable));
MEM_INIT(hc4->chainTable, 0xFF, sizeof(hc4->chainTable));
@ -345,14 +355,20 @@ forceinline int LZ4_InitHC (LZ4HC_Data_Structure* hc4, const BYTE* base)
hc4->base = base;
hc4->inputBuffer = base;
hc4->end = base;
return 1;
}
int LZ4_resetStreamStateHC(void* state, const char* inputBuffer)
{
if ((((size_t)state) & (sizeof(void*)-1)) != 0) return 1; /* Error : pointer is not aligned for pointer (32 or 64 bits) */
LZ4_initHC((LZ4HC_Data_Structure*)state, (const BYTE*)inputBuffer);
return 0;
}
void* LZ4_createHC (const char* slidingInputBuffer)
void* LZ4_createHC (const char* inputBuffer)
{
void* hc4 = ALLOCATOR(sizeof(LZ4HC_Data_Structure));
LZ4_InitHC ((LZ4HC_Data_Structure*)hc4, (const BYTE*)slidingInputBuffer);
LZ4_initHC ((LZ4HC_Data_Structure*)hc4, (const BYTE*)inputBuffer);
return hc4;
}
@ -364,8 +380,8 @@ int LZ4_freeHC (void* LZ4HC_Data)
}
// Update chains up to ip (excluded)
forceinline void LZ4HC_Insert (LZ4HC_Data_Structure* hc4, const BYTE* ip)
/* Update chains up to ip (excluded) */
FORCE_INLINE void LZ4HC_Insert (LZ4HC_Data_Structure* hc4, const BYTE* ip)
{
U16* chainTable = hc4->chainTable;
HTYPE* HashTable = hc4->hashTable;
@ -386,13 +402,16 @@ forceinline void LZ4HC_Insert (LZ4HC_Data_Structure* hc4, const BYTE* ip)
char* LZ4_slideInputBufferHC(void* LZ4HC_Data)
{
LZ4HC_Data_Structure* hc4 = (LZ4HC_Data_Structure*)LZ4HC_Data;
U32 distance = (U32)(hc4->end - hc4->inputBuffer) - 64 KB;
distance = (distance >> 16) << 16; // Must be a multiple of 64 KB
size_t distance = (hc4->end - 64 KB) - hc4->inputBuffer;
if (hc4->end <= hc4->inputBuffer + 64 KB) return (char*)(hc4->end); /* no update : less than 64KB within buffer */
distance = (distance >> 16) << 16; /* Must be a multiple of 64 KB */
LZ4HC_Insert(hc4, hc4->end - MINMATCH);
memcpy((void*)(hc4->end - 64 KB - distance), (const void*)(hc4->end - 64 KB), 64 KB);
hc4->nextToUpdate -= distance;
hc4->base -= distance;
if ((U32)(hc4->inputBuffer - hc4->base) > 1 GB + 64 KB) // Avoid overflow
if ((U32)(hc4->inputBuffer - hc4->base) > 1 GB + 64 KB) /* Avoid overflow */
{
int i;
hc4->base += 1 GB;
@ -403,13 +422,13 @@ char* LZ4_slideInputBufferHC(void* LZ4HC_Data)
}
forceinline size_t LZ4HC_CommonLength (const BYTE* p1, const BYTE* p2, const BYTE* const matchlimit)
FORCE_INLINE size_t LZ4HC_CommonLength (const BYTE* p1, const BYTE* p2, const BYTE* const matchlimit)
{
const BYTE* p1t = p1;
while (p1t<matchlimit-(STEPSIZE-1))
{
UARCH diff = AARCH(p2) ^ AARCH(p1t);
size_t diff = AARCH(p2) ^ AARCH(p1t);
if (!diff) { p1t+=STEPSIZE; p2+=STEPSIZE; continue; }
p1t += LZ4_NbCommonBytes(diff);
return (p1t - p1);
@ -421,26 +440,26 @@ forceinline size_t LZ4HC_CommonLength (const BYTE* p1, const BYTE* p2, const BYT
}
forceinline int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* const matchlimit, const BYTE** matchpos)
FORCE_INLINE int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* const matchlimit, const BYTE** matchpos, const int maxNbAttempts)
{
U16* const chainTable = hc4->chainTable;
HTYPE* const HashTable = hc4->hashTable;
const BYTE* ref;
INITBASE(base,hc4->base);
int nbAttempts=MAX_NB_ATTEMPTS;
int nbAttempts=maxNbAttempts;
size_t repl=0, ml=0;
U16 delta=0; // useless assignment, to remove an uninitialization warning
U16 delta=0; /* useless assignment, to remove an uninitialization warning */
// HC4 match finder
/* HC4 match finder */
LZ4HC_Insert(hc4, ip);
ref = HASH_POINTER(ip);
#define REPEAT_OPTIMIZATION
#ifdef REPEAT_OPTIMIZATION
// Detect repetitive sequences of length <= 4
if ((U32)(ip-ref) <= 4) // potential repetition
/* Detect repetitive sequences of length <= 4 */
if ((U32)(ip-ref) <= 4) /* potential repetition */
{
if (A32(ref) == A32(ip)) // confirmed
if (A32(ref) == A32(ip)) /* confirmed */
{
delta = (U16)(ip-ref);
repl = ml = LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit) + MINMATCH;
@ -463,7 +482,7 @@ forceinline int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const B
}
#ifdef REPEAT_OPTIMIZATION
// Complete table
/* Complete table */
if (repl)
{
const BYTE* ptr = ip;
@ -472,13 +491,13 @@ forceinline int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const B
end = ip + repl - (MINMATCH-1);
while(ptr < end-delta)
{
DELTANEXT(ptr) = delta; // Pre-Load
DELTANEXT(ptr) = delta; /* Pre-Load */
ptr++;
}
do
{
DELTANEXT(ptr) = delta;
HashTable[HASH_VALUE(ptr)] = (HTYPE)((ptr) - base); // Head of chain
HashTable[HASH_VALUE(ptr)] = (HTYPE)((ptr) - base); /* Head of chain */
ptr++;
} while(ptr < end);
hc4->nextToUpdate = end;
@ -489,16 +508,16 @@ forceinline int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const B
}
forceinline int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* startLimit, const BYTE* matchlimit, int longest, const BYTE** matchpos, const BYTE** startpos)
FORCE_INLINE int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* startLimit, const BYTE* matchlimit, int longest, const BYTE** matchpos, const BYTE** startpos, const int maxNbAttempts)
{
U16* const chainTable = hc4->chainTable;
HTYPE* const HashTable = hc4->hashTable;
INITBASE(base,hc4->base);
const BYTE* ref;
int nbAttempts = MAX_NB_ATTEMPTS;
int nbAttempts = maxNbAttempts;
int delta = (int)(ip-startLimit);
// First Match
/* First Match */
LZ4HC_Insert(hc4, ip);
ref = HASH_POINTER(ip);
@ -515,7 +534,7 @@ forceinline int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const B
while (ipt<matchlimit-(STEPSIZE-1))
{
UARCH diff = AARCH(reft) ^ AARCH(ipt);
size_t diff = AARCH(reft) ^ AARCH(ipt);
if (!diff) { ipt+=STEPSIZE; reft+=STEPSIZE; continue; }
ipt += LZ4_NbCommonBytes(diff);
goto _endCount;
@ -526,7 +545,7 @@ forceinline int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const B
_endCount:
reft = ref;
#else
// Easier for code maintenance, but unfortunately slower too
/* Easier for code maintenance, but unfortunately slower too */
const BYTE* startt = ip;
const BYTE* reft = ref;
const BYTE* ipt = ip + MINMATCH + LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit);
@ -548,37 +567,331 @@ _endCount:
}
typedef enum { noLimit = 0, limitedOutput = 1 } limitedOutput_directive;
//**************************************
// Compression functions
//**************************************
FORCE_INLINE int LZ4HC_encodeSequence (
const BYTE** ip,
BYTE** op,
const BYTE** anchor,
int matchLength,
const BYTE* ref,
limitedOutput_directive limitedOutputBuffer,
BYTE* oend)
{
int length;
BYTE* token;
/*
int LZ4_compressHC(
const char* source,
char* dest,
int inputSize)
/* Encode Literal length */
length = (int)(*ip - *anchor);
token = (*op)++;
if ((limitedOutputBuffer) && ((*op + length + (2 + 1 + LASTLITERALS) + (length>>8)) > oend)) return 1; /* Check output limit */
if (length>=(int)RUN_MASK) { int len; *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *(*op)++ = 255; *(*op)++ = (BYTE)len; }
else *token = (BYTE)(length<<ML_BITS);
Compress 'inputSize' bytes from 'source' into an output buffer 'dest'.
Destination buffer must be already allocated, and sized at a minimum of LZ4_compressBound(inputSize).
return : the number of bytes written in buffer 'dest'
*/
#define FUNCTION_NAME LZ4_compressHC
#include "lz4hc_encoder.h"
/* Copy Literals */
LZ4_BLINDCOPY(*anchor, *op, length);
/* Encode Offset */
LZ4_WRITE_LITTLEENDIAN_16(*op,(U16)(*ip-ref));
/* Encode MatchLength */
length = (int)(matchLength-MINMATCH);
if ((limitedOutputBuffer) && (*op + (1 + LASTLITERALS) + (length>>8) > oend)) return 1; /* Check output limit */
if (length>=(int)ML_MASK) { *token+=ML_MASK; length-=ML_MASK; for(; length > 509 ; length-=510) { *(*op)++ = 255; *(*op)++ = 255; } if (length > 254) { length-=255; *(*op)++ = 255; } *(*op)++ = (BYTE)length; }
else *token += (BYTE)(length);
/* Prepare next loop */
*ip += matchLength;
*anchor = *ip;
return 0;
}
/*
int LZ4_compressHC_limitedOutput(
#define MAX_COMPRESSION_LEVEL 16
static int LZ4HC_compress_generic (
void* ctxvoid,
const char* source,
char* dest,
int inputSize,
int maxOutputSize)
int maxOutputSize,
int compressionLevel,
limitedOutput_directive limit
)
{
LZ4HC_Data_Structure* ctx = (LZ4HC_Data_Structure*) ctxvoid;
const BYTE* ip = (const BYTE*) source;
const BYTE* anchor = ip;
const BYTE* const iend = ip + inputSize;
const BYTE* const mflimit = iend - MFLIMIT;
const BYTE* const matchlimit = (iend - LASTLITERALS);
Compress 'inputSize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'.
If it cannot achieve it, compression will stop, and result of the function will be zero.
return : the number of bytes written in buffer 'dest', or 0 if the compression fails
*/
#define FUNCTION_NAME LZ4_compressHC_limitedOutput
#define LIMITED_OUTPUT
#include "lz4hc_encoder.h"
BYTE* op = (BYTE*) dest;
BYTE* const oend = op + maxOutputSize;
const int maxNbAttempts = compressionLevel > MAX_COMPRESSION_LEVEL ? 1 << MAX_COMPRESSION_LEVEL : compressionLevel ? 1<<(compressionLevel-1) : 1<<LZ4HC_DEFAULT_COMPRESSIONLEVEL;
int ml, ml2, ml3, ml0;
const BYTE* ref=NULL;
const BYTE* start2=NULL;
const BYTE* ref2=NULL;
const BYTE* start3=NULL;
const BYTE* ref3=NULL;
const BYTE* start0;
const BYTE* ref0;
/* Ensure blocks follow each other */
if (ip != ctx->end) return 0;
ctx->end += inputSize;
ip++;
/* Main Loop */
while (ip < mflimit)
{
ml = LZ4HC_InsertAndFindBestMatch (ctx, ip, matchlimit, (&ref), maxNbAttempts);
if (!ml) { ip++; continue; }
/* saved, in case we would skip too much */
start0 = ip;
ref0 = ref;
ml0 = ml;
_Search2:
if (ip+ml < mflimit)
ml2 = LZ4HC_InsertAndGetWiderMatch(ctx, ip + ml - 2, ip + 1, matchlimit, ml, &ref2, &start2, maxNbAttempts);
else ml2 = ml;
if (ml2 == ml) /* No better match */
{
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
continue;
}
if (start0 < ip)
{
if (start2 < ip + ml0) /* empirical */
{
ip = start0;
ref = ref0;
ml = ml0;
}
}
/* Here, start0==ip */
if ((start2 - ip) < 3) /* First Match too small : removed */
{
ml = ml2;
ip = start2;
ref =ref2;
goto _Search2;
}
_Search3:
/*
* Currently we have :
* ml2 > ml1, and
* ip1+3 <= ip2 (usually < ip1+ml1)
*/
if ((start2 - ip) < OPTIMAL_ML)
{
int correction;
int new_ml = ml;
if (new_ml > OPTIMAL_ML) new_ml = OPTIMAL_ML;
if (ip+new_ml > start2 + ml2 - MINMATCH) new_ml = (int)(start2 - ip) + ml2 - MINMATCH;
correction = new_ml - (int)(start2 - ip);
if (correction > 0)
{
start2 += correction;
ref2 += correction;
ml2 -= correction;
}
}
/* Now, we have start2 = ip+new_ml, with new_ml = min(ml, OPTIMAL_ML=18) */
if (start2 + ml2 < mflimit)
ml3 = LZ4HC_InsertAndGetWiderMatch(ctx, start2 + ml2 - 3, start2, matchlimit, ml2, &ref3, &start3, maxNbAttempts);
else ml3 = ml2;
if (ml3 == ml2) /* No better match : 2 sequences to encode */
{
/* ip & ref are known; Now for ml */
if (start2 < ip+ml) ml = (int)(start2 - ip);
/* Now, encode 2 sequences */
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
ip = start2;
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml2, ref2, limit, oend)) return 0;
continue;
}
if (start3 < ip+ml+3) /* Not enough space for match 2 : remove it */
{
if (start3 >= (ip+ml)) /* can write Seq1 immediately ==> Seq2 is removed, so Seq3 becomes Seq1 */
{
if (start2 < ip+ml)
{
int correction = (int)(ip+ml - start2);
start2 += correction;
ref2 += correction;
ml2 -= correction;
if (ml2 < MINMATCH)
{
start2 = start3;
ref2 = ref3;
ml2 = ml3;
}
}
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
ip = start3;
ref = ref3;
ml = ml3;
start0 = start2;
ref0 = ref2;
ml0 = ml2;
goto _Search2;
}
start2 = start3;
ref2 = ref3;
ml2 = ml3;
goto _Search3;
}
/*
* OK, now we have 3 ascending matches; let's write at least the first one
* ip & ref are known; Now for ml
*/
if (start2 < ip+ml)
{
if ((start2 - ip) < (int)ML_MASK)
{
int correction;
if (ml > OPTIMAL_ML) ml = OPTIMAL_ML;
if (ip + ml > start2 + ml2 - MINMATCH) ml = (int)(start2 - ip) + ml2 - MINMATCH;
correction = ml - (int)(start2 - ip);
if (correction > 0)
{
start2 += correction;
ref2 += correction;
ml2 -= correction;
}
}
else
{
ml = (int)(start2 - ip);
}
}
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
ip = start2;
ref = ref2;
ml = ml2;
start2 = start3;
ref2 = ref3;
ml2 = ml3;
goto _Search3;
}
/* Encode Last Literals */
{
int lastRun = (int)(iend - anchor);
if ((limit) && (((char*)op - dest) + lastRun + 1 + ((lastRun+255-RUN_MASK)/255) > (U32)maxOutputSize)) return 0; /* Check output limit */
if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
else *op++ = (BYTE)(lastRun<<ML_BITS);
memcpy(op, anchor, iend - anchor);
op += iend-anchor;
}
/* End */
return (int) (((char*)op)-dest);
}
int LZ4_compressHC2(const char* source, char* dest, int inputSize, int compressionLevel)
{
void* ctx = LZ4_createHC(source);
int result;
if (ctx==NULL) return 0;
result = LZ4HC_compress_generic (ctx, source, dest, inputSize, 0, compressionLevel, noLimit);
LZ4_freeHC(ctx);
return result;
}
int LZ4_compressHC(const char* source, char* dest, int inputSize) { return LZ4_compressHC2(source, dest, inputSize, 0); }
int LZ4_compressHC2_limitedOutput(const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel)
{
void* ctx = LZ4_createHC(source);
int result;
if (ctx==NULL) return 0;
result = LZ4HC_compress_generic (ctx, source, dest, inputSize, maxOutputSize, compressionLevel, limitedOutput);
LZ4_freeHC(ctx);
return result;
}
int LZ4_compressHC_limitedOutput(const char* source, char* dest, int inputSize, int maxOutputSize)
{
return LZ4_compressHC2_limitedOutput(source, dest, inputSize, maxOutputSize, 0);
}
/*****************************
Using external allocation
*****************************/
int LZ4_sizeofStateHC() { return sizeof(LZ4HC_Data_Structure); }
int LZ4_compressHC2_withStateHC (void* state, const char* source, char* dest, int inputSize, int compressionLevel)
{
if (((size_t)(state)&(sizeof(void*)-1)) != 0) return 0; /* Error : state is not aligned for pointers (32 or 64 bits) */
LZ4_initHC ((LZ4HC_Data_Structure*)state, (const BYTE*)source);
return LZ4HC_compress_generic (state, source, dest, inputSize, 0, compressionLevel, noLimit);
}
int LZ4_compressHC_withStateHC (void* state, const char* source, char* dest, int inputSize)
{ return LZ4_compressHC2_withStateHC (state, source, dest, inputSize, 0); }
int LZ4_compressHC2_limitedOutput_withStateHC (void* state, const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel)
{
if (((size_t)(state)&(sizeof(void*)-1)) != 0) return 0; /* Error : state is not aligned for pointers (32 or 64 bits) */
LZ4_initHC ((LZ4HC_Data_Structure*)state, (const BYTE*)source);
return LZ4HC_compress_generic (state, source, dest, inputSize, maxOutputSize, compressionLevel, limitedOutput);
}
int LZ4_compressHC_limitedOutput_withStateHC (void* state, const char* source, char* dest, int inputSize, int maxOutputSize)
{ return LZ4_compressHC2_limitedOutput_withStateHC (state, source, dest, inputSize, maxOutputSize, 0); }
/****************************
Stream functions
****************************/
int LZ4_compressHC_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize)
{
return LZ4HC_compress_generic (LZ4HC_Data, source, dest, inputSize, 0, 0, noLimit);
}
int LZ4_compressHC2_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int compressionLevel)
{
return LZ4HC_compress_generic (LZ4HC_Data, source, dest, inputSize, 0, compressionLevel, noLimit);
}
int LZ4_compressHC_limitedOutput_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int maxOutputSize)
{
return LZ4HC_compress_generic (LZ4HC_Data, source, dest, inputSize, maxOutputSize, 0, limitedOutput);
}
int LZ4_compressHC2_limitedOutput_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel)
{
return LZ4HC_compress_generic (LZ4HC_Data, source, dest, inputSize, maxOutputSize, compressionLevel, limitedOutput);
}

76
hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4hc.h
View File

@ -1,7 +1,7 @@
/*
LZ4 HC - High Compression Mode of LZ4
Header File
Copyright (C) 2011-2013, Yann Collet.
Copyright (C) 2011-2014, Yann Collet.
BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
Redistribution and use in source and binary forms, with or without
@ -63,32 +63,70 @@ LZ4_compress_limitedOutput() :
*/
int LZ4_compressHC2 (const char* source, char* dest, int inputSize, int compressionLevel);
int LZ4_compressHC2_limitedOutput (const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel);
/*
Same functions as above, but with programmable 'compressionLevel'.
Recommended values are between 4 and 9, although any value between 0 and 16 will work.
'compressionLevel'==0 means use default 'compressionLevel' value.
Values above 16 behave the same as 16.
Equivalent variants exist for all other compression functions below.
*/
/* Note :
Decompression functions are provided within LZ4 source code (see "lz4.h") (BSD license)
*/
/* Advanced Functions */
/**************************************
Using an external allocation
**************************************/
int LZ4_sizeofStateHC(void);
int LZ4_compressHC_withStateHC (void* state, const char* source, char* dest, int inputSize);
int LZ4_compressHC_limitedOutput_withStateHC (void* state, const char* source, char* dest, int inputSize, int maxOutputSize);
void* LZ4_createHC (const char* slidingInputBuffer);
int LZ4_compressHC2_withStateHC (void* state, const char* source, char* dest, int inputSize, int compressionLevel);
int LZ4_compressHC2_limitedOutput_withStateHC(void* state, const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel);
/*
These functions are provided should you prefer to allocate memory for compression tables with your own allocation methods.
To know how much memory must be allocated for the compression tables, use :
int LZ4_sizeofStateHC();
Note that tables must be aligned for pointer (32 or 64 bits), otherwise compression will fail (return code 0).
The allocated memory can be provided to the compressions functions using 'void* state' parameter.
LZ4_compress_withStateHC() and LZ4_compress_limitedOutput_withStateHC() are equivalent to previously described functions.
They just use the externally allocated memory area instead of allocating their own (on stack, or on heap).
*/
/**************************************
Streaming Functions
**************************************/
/* Note : these streaming functions still follows the older model */
void* LZ4_createHC (const char* inputBuffer);
int LZ4_compressHC_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize);
int LZ4_compressHC_limitedOutput_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int maxOutputSize);
char* LZ4_slideInputBufferHC (void* LZ4HC_Data);
int LZ4_freeHC (void* LZ4HC_Data);
int LZ4_compressHC2_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int compressionLevel);
int LZ4_compressHC2_limitedOutput_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel);
/*
These functions allow the compression of dependent blocks, where each block benefits from prior 64 KB within preceding blocks.
In order to achieve this, it is necessary to start creating the LZ4HC Data Structure, thanks to the function :
void* LZ4_createHC (const char* slidingInputBuffer);
void* LZ4_createHC (const char* inputBuffer);
The result of the function is the (void*) pointer on the LZ4HC Data Structure.
This pointer will be needed in all other functions.
If the pointer returned is NULL, then the allocation has failed, and compression must be aborted.
The only parameter 'const char* slidingInputBuffer' must, obviously, point at the beginning of input buffer.
The only parameter 'const char* inputBuffer' must, obviously, point at the beginning of input buffer.
The input buffer must be already allocated, and size at least 192KB.
'slidingInputBuffer' will also be the 'const char* source' of the first block.
'inputBuffer' will also be the 'const char* source' of the first block.
All blocks are expected to lay next to each other within the input buffer, starting from 'slidingInputBuffer'.
All blocks are expected to lay next to each other within the input buffer, starting from 'inputBuffer'.
To compress each block, use either LZ4_compressHC_continue() or LZ4_compressHC_limitedOutput_continue().
Their behavior are identical to LZ4_compressHC() or LZ4_compressHC_limitedOutput(),
but require the LZ4HC Data Structure as their first argument, and check that each block starts right after the previous one.
@ -105,6 +143,30 @@ Compression can then resume, using LZ4_compressHC_continue() or LZ4_compressHC_l
When compression is completed, a call to LZ4_freeHC() will release the memory used by the LZ4HC Data Structure.
*/
int LZ4_sizeofStreamStateHC(void);
int LZ4_resetStreamStateHC(void* state, const char* inputBuffer);
/*
These functions achieve the same result as :
void* LZ4_createHC (const char* inputBuffer);
They are provided here to allow the user program to allocate memory using its own routines.
To know how much space must be allocated, use LZ4_sizeofStreamStateHC();
Note also that space must be aligned for pointers (32 or 64 bits).
Once space is allocated, you must initialize it using : LZ4_resetStreamStateHC(void* state, const char* inputBuffer);
void* state is a pointer to the space allocated.
It must be aligned for pointers (32 or 64 bits), and be large enough.
The parameter 'const char* inputBuffer' must, obviously, point at the beginning of input buffer.
The input buffer must be already allocated, and size at least 192KB.
'inputBuffer' will also be the 'const char* source' of the first block.
The same space can be re-used multiple times, just by initializing it each time with LZ4_resetStreamState().
return value of LZ4_resetStreamStateHC() must be 0 is OK.
Any other value means there was an error (typically, state is not aligned for pointers (32 or 64 bits)).
*/
#if defined (__cplusplus)
}

349
hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4hc_encoder.h
View File

@ -1,349 +0,0 @@
/*
LZ4 HC Encoder - Part of LZ4 HC algorithm
Copyright (C) 2011-2013, Yann Collet.
BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the following disclaimer
in the documentation and/or other materials provided with the
distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
You can contact the author at :
- LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
- LZ4 source repository : http://code.google.com/p/lz4/
*/
/* lz4hc_encoder.h must be included into lz4hc.c
The objective of this file is to create a single LZ4 compression function source
which will be instanciated multiple times with minor variations
depending on a set of #define.
*/
//****************************
// Check required defines
//****************************
#ifndef FUNCTION_NAME
# error "FUNTION_NAME is not defined"
#endif
//****************************
// Local definitions
//****************************
#define COMBINED_NAME_RAW(n1,n2) n1 ## n2
#define COMBINED_NAME(n1,n2) COMBINED_NAME_RAW(n1,n2)
#define ENCODE_SEQUENCE_NAME COMBINED_NAME(FUNCTION_NAME,_encodeSequence)
#ifdef LIMITED_OUTPUT
# define ENCODE_SEQUENCE(i,o,a,m,r,d) if (ENCODE_SEQUENCE_NAME(i,o,a,m,r,d)) return 0;
#else
# define ENCODE_SEQUENCE(i,o,a,m,r,d) ENCODE_SEQUENCE_NAME(i,o,a,m,r)
#endif
//****************************
// Function code
//****************************
forceinline int ENCODE_SEQUENCE_NAME (
const BYTE** ip,
BYTE** op,
const BYTE** anchor,
int matchLength,
const BYTE* ref
#ifdef LIMITED_OUTPUT
,BYTE* oend
#endif
)
{
int length, len;
BYTE* token;
// Encode Literal length
length = (int)(*ip - *anchor);
token = (*op)++;
#ifdef LIMITED_OUTPUT
if ((*op + length + (2 + 1 + LASTLITERALS) + (length>>8)) > oend) return 1; // Check output limit
#endif
if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *(*op)++ = 255; *(*op)++ = (BYTE)len; }
else *token = (BYTE)(length<<ML_BITS);
// Copy Literals
LZ4_BLINDCOPY(*anchor, *op, length);
// Encode Offset
LZ4_WRITE_LITTLEENDIAN_16(*op,(U16)(*ip-ref));
// Encode MatchLength
length = (int)(matchLength-MINMATCH);
#ifdef LIMITED_OUTPUT
if (*op + (1 + LASTLITERALS) + (length>>8) > oend) return 1; // Check output limit
#endif
if (length>=(int)ML_MASK) { *token+=ML_MASK; length-=ML_MASK; for(; length > 509 ; length-=510) { *(*op)++ = 255; *(*op)++ = 255; } if (length > 254) { length-=255; *(*op)++ = 255; } *(*op)++ = (BYTE)length; }
else *token += (BYTE)(length);
// Prepare next loop
*ip += matchLength;
*anchor = *ip;
return 0;
}
int COMBINED_NAME(FUNCTION_NAME,_continue) (
void* ctxvoid,
const char* source,
char* dest,
int inputSize
#ifdef LIMITED_OUTPUT
,int maxOutputSize
#endif
)
{
LZ4HC_Data_Structure* ctx = (LZ4HC_Data_Structure*) ctxvoid;
const BYTE* ip = (const BYTE*) source;
const BYTE* anchor = ip;
const BYTE* const iend = ip + inputSize;
const BYTE* const mflimit = iend - MFLIMIT;
const BYTE* const matchlimit = (iend - LASTLITERALS);
BYTE* op = (BYTE*) dest;
#ifdef LIMITED_OUTPUT
BYTE* const oend = op + maxOutputSize;
#endif
int ml, ml2, ml3, ml0;
const BYTE* ref=NULL;
const BYTE* start2=NULL;
const BYTE* ref2=NULL;
const BYTE* start3=NULL;
const BYTE* ref3=NULL;
const BYTE* start0;
const BYTE* ref0;
// Ensure blocks follow each other
if (ip != ctx->end) return 0;
ctx->end += inputSize;
ip++;
// Main Loop
while (ip < mflimit)
{
ml = LZ4HC_InsertAndFindBestMatch (ctx, ip, matchlimit, (&ref));
if (!ml) { ip++; continue; }
// saved, in case we would skip too much
start0 = ip;
ref0 = ref;
ml0 = ml;
_Search2:
if (ip+ml < mflimit)
ml2 = LZ4HC_InsertAndGetWiderMatch(ctx, ip + ml - 2, ip + 1, matchlimit, ml, &ref2, &start2);
else ml2 = ml;
if (ml2 == ml) // No better match
{
ENCODE_SEQUENCE(&ip, &op, &anchor, ml, ref, oend);
continue;
}
if (start0 < ip)
{
if (start2 < ip + ml0) // empirical
{
ip = start0;
ref = ref0;
ml = ml0;
}
}
// Here, start0==ip
if ((start2 - ip) < 3) // First Match too small : removed
{
ml = ml2;
ip = start2;
ref =ref2;
goto _Search2;
}
_Search3:
// Currently we have :
// ml2 > ml1, and
// ip1+3 <= ip2 (usually < ip1+ml1)
if ((start2 - ip) < OPTIMAL_ML)
{
int correction;
int new_ml = ml;
if (new_ml > OPTIMAL_ML) new_ml = OPTIMAL_ML;
if (ip+new_ml > start2 + ml2 - MINMATCH) new_ml = (int)(start2 - ip) + ml2 - MINMATCH;
correction = new_ml - (int)(start2 - ip);
if (correction > 0)
{
start2 += correction;
ref2 += correction;
ml2 -= correction;
}
}
// Now, we have start2 = ip+new_ml, with new_ml = min(ml, OPTIMAL_ML=18)
if (start2 + ml2 < mflimit)
ml3 = LZ4HC_InsertAndGetWiderMatch(ctx, start2 + ml2 - 3, start2, matchlimit, ml2, &ref3, &start3);
else ml3 = ml2;
if (ml3 == ml2) // No better match : 2 sequences to encode
{
// ip & ref are known; Now for ml
if (start2 < ip+ml) ml = (int)(start2 - ip);
// Now, encode 2 sequences
ENCODE_SEQUENCE(&ip, &op, &anchor, ml, ref, oend);
ip = start2;
ENCODE_SEQUENCE(&ip, &op, &anchor, ml2, ref2, oend);
continue;
}
if (start3 < ip+ml+3) // Not enough space for match 2 : remove it
{
if (start3 >= (ip+ml)) // can write Seq1 immediately ==> Seq2 is removed, so Seq3 becomes Seq1
{
if (start2 < ip+ml)
{
int correction = (int)(ip+ml - start2);
start2 += correction;
ref2 += correction;
ml2 -= correction;
if (ml2 < MINMATCH)
{
start2 = start3;
ref2 = ref3;
ml2 = ml3;
}
}
ENCODE_SEQUENCE(&ip, &op, &anchor, ml, ref, oend);
ip = start3;
ref = ref3;
ml = ml3;
start0 = start2;
ref0 = ref2;
ml0 = ml2;
goto _Search2;
}
start2 = start3;
ref2 = ref3;
ml2 = ml3;
goto _Search3;
}
// OK, now we have 3 ascending matches; let's write at least the first one
// ip & ref are known; Now for ml
if (start2 < ip+ml)
{
if ((start2 - ip) < (int)ML_MASK)
{
int correction;
if (ml > OPTIMAL_ML) ml = OPTIMAL_ML;
if (ip + ml > start2 + ml2 - MINMATCH) ml = (int)(start2 - ip) + ml2 - MINMATCH;
correction = ml - (int)(start2 - ip);
if (correction > 0)
{
start2 += correction;
ref2 += correction;
ml2 -= correction;
}
}
else
{
ml = (int)(start2 - ip);
}
}
ENCODE_SEQUENCE(&ip, &op, &anchor, ml, ref, oend);
ip = start2;
ref = ref2;
ml = ml2;
start2 = start3;
ref2 = ref3;
ml2 = ml3;
goto _Search3;
}
// Encode Last Literals
{
int lastRun = (int)(iend - anchor);
#ifdef LIMITED_OUTPUT
if (((char*)op - dest) + lastRun + 1 + ((lastRun+255-RUN_MASK)/255) > (U32)maxOutputSize) return 0; // Check output limit
#endif
if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
else *op++ = (BYTE)(lastRun<<ML_BITS);
memcpy(op, anchor, iend - anchor);
op += iend-anchor;
}
// End
return (int) (((char*)op)-dest);
}
int FUNCTION_NAME (const char* source,
char* dest,
int inputSize
#ifdef LIMITED_OUTPUT
,int maxOutputSize
#endif
)
{
void* ctx = LZ4_createHC(source);
int result;
if (ctx==NULL) return 0;
#ifdef LIMITED_OUTPUT
result = COMBINED_NAME(FUNCTION_NAME,_continue) (ctx, source, dest, inputSize, maxOutputSize);
#else
result = COMBINED_NAME(FUNCTION_NAME,_continue) (ctx, source, dest, inputSize);
#endif
LZ4_freeHC(ctx);
return result;
}
//****************************
// Clean defines
//****************************
// Required defines
#undef FUNCTION_NAME
// Locally Generated
#undef ENCODE_SEQUENCE
#undef ENCODE_SEQUENCE_NAME
// Optional defines
#ifdef LIMITED_OUTPUT
#undef LIMITED_OUTPUT
#endif

2
hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-nativetask/pom.xml
View File

@ -172,8 +172,6 @@
<target>
<copy file="${basedir}/../../../hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4.h"
todir="${project.build.directory}/native/" />
<copy file="${basedir}/../../../hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4_encoder.h"
todir="${project.build.directory}/native/" />
<copy file="${basedir}/../../../hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/compress/lz4/lz4.c"
todir="${project.build.directory}/native/" />
<copy todir="${project.build.directory}/native/test/testData"

2
hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-nativetask/src/main/native/src/codec/Lz4Codec.cc
View File

@ -73,7 +73,7 @@ uint32_t Lz4DecompressStream::decompressOneBlock(uint32_t compressedSize, void *
THROW_EXCEPTION(IOException, "readFully reach EOF");
}
_compressedBytesRead += rd;
uint32_t ret = LZ4_uncompress(_tempBuffer, (char*)buff, length);
uint32_t ret = LZ4_decompress_fast(_tempBuffer, (char*)buff, length);
if (ret == compressedSize) {
return length;
} else {

2
hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-nativetask/src/main/native/test/TestCompressions.cc
View File

@ -193,7 +193,7 @@ void MeasureSingleFileLz4(const string & path, CompressResult & total, size_t bl
result.compressTime += endTime - startTime;
startTime = t.now();
for (int i = 0; i < times; i++) {
int osize = LZ4_uncompress(outputBuffer, dest, currentblocksize);
int osize = LZ4_decompress_fast(outputBuffer, dest, currentblocksize);
ASSERT_EQ(currentblocksize, osize);
}
endTime = t.now();