ExternalVendorCode/ZeroTierOne
1
0
Fork 0
mirror of https://github.com/zerotier/ZeroTierOne.git synced 2025-06-06 17:31:53 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Upgrade LZ4 to latest version.

Browse Source
This commit is contained in:
Adam Ierymenko 2014-07-30 15:34:15 -07:00
parent 59ebc9d41e
commit 9b93141dd0
5 changed files with 2641 additions and 1956 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1087
ext/lz4/lz4.c
View File

File diff suppressed because it is too large Load Diff

311
ext/lz4/lz4.h
View File

@ -1,7 +1,7 @@
/* /*
LZ4 - Fast LZ compression algorithm LZ4 - Fast LZ compression algorithm
Header File 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) BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
Redistribution and use in source and binary forms, with or without Redistribution and use in source and binary forms, with or without
@ -28,8 +28,8 @@
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
You can contact the author at : 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 source repository : http://code.google.com/p/lz4/
- LZ4 public forum : https://groups.google.com/forum/#!forum/lz4c
*/ */
#pragma once #pragma once
@ -38,20 +38,34 @@ extern "C" {
#endif #endif
//************************************** /**************************************
// Compiler Options Version
//************************************** **************************************/
#if defined(_MSC_VER) && !defined(__cplusplus) // Visual Studio #define LZ4_VERSION_MAJOR 1 /* for major interface/format changes */
# define inline __inline // Visual C is not C99, but supports some kind of inline #define LZ4_VERSION_MINOR 3 /* for minor interface/format changes */
#endif #define LZ4_VERSION_RELEASE 0 /* 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_compress (const char* source, char* dest, int inputSize);
int LZ4_decompress_safe (const char* source, char* dest, int inputSize, int maxOutputSize); int LZ4_decompress_safe (const char* source, char* dest, int compressedSize, int maxDecompressedSize);
/* /*
LZ4_compress() : LZ4_compress() :
@ -59,40 +73,48 @@ LZ4_compress() :
Destination buffer must be already allocated, Destination buffer must be already allocated,
and must be sized to handle worst cases situations (input data not compressible) and must be sized to handle worst cases situations (input data not compressible)
Worst case size evaluation is provided by function LZ4_compressBound() Worst case size evaluation is provided by function LZ4_compressBound()
inputSize : Max supported value is LZ4_MAX_INPUT_VALUE inputSize : Max supported value is LZ4_MAX_INPUT_SIZE
return : the number of bytes written in buffer dest return : the number of bytes written in buffer dest
or 0 if the compression fails or 0 if the compression fails
LZ4_decompress_safe() : LZ4_decompress_safe() :
maxOutputSize : is the size of the destination buffer (which must be already allocated) compressedSize : is obviously the source size
return : the number of bytes decoded in the destination buffer (necessarily <= maxOutputSize) 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. 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 :
it never writes outside of output buffer, and never reads outside of input buffer.
Therefore, it is protected against malicious data packets.
*/ */
//**************************** /*
// Advanced Functions Note :
//**************************** Should you prefer to explicitly allocate compression-table memory using your own allocation method,
#define LZ4_MAX_INPUT_SIZE 0x7E000000 // 2 113 929 216 bytes use the streaming functions provided below, simply reset the memory area between each call to LZ4_compress_continue()
*/
/**************************************
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) #define LZ4_COMPRESSBOUND(isize) ((unsigned int)(isize) > (unsigned int)LZ4_MAX_INPUT_SIZE ? 0 : (isize) + ((isize)/255) + 16)
static inline int LZ4_compressBound(int isize) { return LZ4_COMPRESSBOUND(isize); }
/* /*
LZ4_compressBound() : LZ4_compressBound() :
Provides the maximum size that LZ4 may output in a "worst case" scenario (input data not compressible) 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. 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 stack memory allocation). macro is also provided when result needs to be evaluated at compilation (such as stack memory allocation).
isize : is the input size. Max supported value is LZ4_MAX_INPUT_SIZE isize : is the input size. Max supported value is LZ4_MAX_INPUT_SIZE
return : maximum output size in a "worst case" scenario return : maximum output size in a "worst case" scenario
or 0, if input size is too large ( > LZ4_MAX_INPUT_SIZE) 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() : LZ4_compress_limitedOutput() :
Compress 'inputSize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'. Compress 'inputSize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'.
@ -104,100 +126,199 @@ LZ4_compress_limitedOutput() :
return : the number of bytes written in buffer 'dest' return : the number of bytes written in buffer 'dest'
or 0 if the compression fails or 0 if the compression fails
*/ */
int LZ4_compress_limitedOutput (const char* source, char* dest, int inputSize, 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() : 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) 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. If the source stream is detected malformed, the function will stop decoding and return a negative result.
note : This function is a bit faster than LZ4_decompress_safe() Destination buffer must be already allocated. Its size must be a minimum of 'originalSize' bytes.
This function never writes outside of output buffers, but may read beyond input buffer in case of malicious data packet. note : This function fully respect memory boundaries for properly formed compressed data.
Use this function preferably into a trusted environment (data to decode comes from a trusted source). It is a bit faster than LZ4_decompress_safe().
Destination buffer must be already allocated. Its size must be a minimum of 'outputSize' bytes. However, it does not provide any protection against intentionnally 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() : LZ4_decompress_safe_partial() :
This function decompress a compressed block of size 'inputSize' at position 'source' This function decompress a compressed block of size 'compressedSize' at position 'source'
into output buffer 'dest' of size 'maxOutputSize'. into destination buffer 'dest' of size 'maxDecompressedSize'.
The function tries to stop decompressing operation as soon as 'targetOutputSize' has been reached, The function tries to stop decompressing operation as soon as 'targetOutputSize' has been reached,
reducing decompression time. 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. Note : this number can be < 'targetOutputSize' should the compressed block to decode be smaller.
Always control how many bytes were decoded. Always control how many bytes were decoded.
If the source stream is detected 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 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);
//**************************** /***********************************************
// Stream Functions 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;
/*
* LZ4_resetStream
* Use this function to init a newly allocated LZ4_stream_t structure
* You can also reset an existing LZ4_stream_t structure
*/
void LZ4_resetStream (LZ4_stream_t* LZ4_stream);
/*
* 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 safe 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 : 1 if OK, 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_setDictDecode()
*/
int LZ4_decompress_safe_continue (LZ4_streamDecode_t* LZ4_streamDecode, const char* source, char* dest, int compressedSize, int maxOutputSize);
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()
all together into a single function call.
It doesn't use nor update an LZ4_streamDecode_t structure.
*/
int LZ4_decompress_safe_usingDict (const char* source, char* dest, int compressedSize, int maxOutputSize, 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); void* LZ4_create (const char* inputBuffer);
int LZ4_compress_continue (void* LZ4_Data, const char* source, char* dest, int inputSize); int LZ4_sizeofStreamState(void);
int LZ4_compress_limitedOutput_continue (void* LZ4_Data, const char* source, char* dest, int inputSize, int maxOutputSize); int LZ4_resetStreamState(void* state, const char* inputBuffer);
char* LZ4_slideInputBuffer (void* LZ4_Data); char* LZ4_slideInputBuffer (void* state);
int LZ4_free (void* LZ4_Data);
/*
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 LZ4 Data Structure, thanks to the function :
void* LZ4_create (const char* inputBuffer);
The result of the function is the (void*) pointer on the LZ4 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* 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.
All blocks are expected to lay next to each other within the input buffer, starting from 'inputBuffer'.
To compress each block, use either LZ4_compress_continue() or LZ4_compress_limitedOutput_continue().
Their behavior are identical to LZ4_compress() or LZ4_compress_limitedOutput(),
but require the LZ4 Data Structure as their first argument, and check that each block starts right after the previous one.
If next block does not begin immediately after the previous one, the compression will fail (return 0).
When it's no longer possible to lay the next block after the previous one (not enough space left into input buffer), a call to :
char* LZ4_slideInputBuffer(void* LZ4_Data);
must be performed. It will typically copy the latest 64KB of input at the beginning of input buffer.
Note that, for this function to work properly, minimum size of an input buffer must be 192KB.
==> The memory position where the next input data block must start is provided as the result of the function.
Compression can then resume, using LZ4_compress_continue() or LZ4_compress_limitedOutput_continue(), as usual.
When compression is completed, a call to LZ4_free() will release the memory used by the LZ4 Data Structure.
*/
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);
/*
*_withPrefix64k() :
These decoding functions work the same as their "normal name" versions,
but can use up to 64KB of data in front of 'char* dest'.
These functions are necessary to decode 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); }
/*
These functions are deprecated and should no longer be used.
They are provided here for compatibility with existing user programs.
*/
/* 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) #if defined (__cplusplus)

339
ext/lz4/lz4hc.c
View File

@ -1,6 +1,6 @@
/* /*
LZ4 HC - High Compression Mode of LZ4 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) BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
Redistribution and use in source and binary forms, with or without Redistribution and use in source and binary forms, with or without
@ -31,31 +31,43 @@
- LZ4 source repository : http://code.google.com/p/lz4/ - LZ4 source repository : http://code.google.com/p/lz4/
*/ */
//**************************************
// 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 ALLOCATOR(s) calloc(1,s)
#define FREEMEM free #define FREEMEM free
#include <string.h> // memset, memcpy #include <string.h> /* memset, memcpy */
#define MEM_INIT memset #define MEM_INIT memset
//************************************** /**************************************
// CPU Feature Detection CPU Feature Detection
//************************************** **************************************/
// 32 or 64 bits ? /* 32 or 64 bits ? */
#if (defined(__x86_64__) || defined(_M_X64) || defined(_WIN64) \ #if (defined(__x86_64__) || defined(_M_X64) || defined(_WIN64) \
|| defined(__powerpc64__) || defined(__ppc64__) || defined(__PPC64__) \ || defined(__powerpc64__) || defined(__powerpc64le__) \
|| defined(__64BIT__) || defined(_LP64) || defined(__LP64__) \ || defined(__ppc64__) || defined(__ppc64le__) \
|| defined(__ia64) || defined(__itanium__) || defined(_M_IA64) ) // Detects 64 bits mode || defined(__PPC64__) || defined(__PPC64LE__) \
|| defined(__ia64) || defined(__itanium__) || defined(_M_IA64) ) /* Detects 64 bits mode */
# define LZ4_ARCH64 1 # define LZ4_ARCH64 1
#else #else
# define LZ4_ARCH64 0 # define LZ4_ARCH64 0
#endif #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__) #if defined (__GLIBC__)
# include <endian.h> # include <endian.h>
# if (__BYTE_ORDER == __BIG_ENDIAN) # if (__BYTE_ORDER == __BIG_ENDIAN)
@ -69,43 +81,45 @@
|| defined(_MIPSEB) || defined(__s390__) || defined(_MIPSEB) || defined(__s390__)
# define LZ4_BIG_ENDIAN 1 # define LZ4_BIG_ENDIAN 1
#else #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 #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 * Unaligned memory access is automatically enabled for "common" CPU, such as x86.
// If you know your target CPU supports unaligned memory access, you want to force this option manually to improve performance * 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) #if defined(__ARM_FEATURE_UNALIGNED)
# define LZ4_FORCE_UNALIGNED_ACCESS 1 # define LZ4_FORCE_UNALIGNED_ACCESS 1
#endif #endif
// Define this parameter if your target system or compiler 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 #if defined(_MSC_VER) && defined(_WIN32_WCE) /* Visual Studio for Windows CE does not support Hardware bit count */
# define LZ4_FORCE_SW_BITCOUNT # define LZ4_FORCE_SW_BITCOUNT
#endif #endif
//************************************** /**************************************
// Compiler Options Compiler Options
//************************************** **************************************/
#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L // C99 #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */
/* "restrict" is a known keyword */ /* "restrict" is a known keyword */
#else #else
# define restrict // Disable restrict # define restrict /* Disable restrict */
#endif #endif
#ifdef _MSC_VER // Visual Studio #ifdef _MSC_VER /* Visual Studio */
# define FORCE_INLINE static __forceinline # define FORCE_INLINE static __forceinline
# include <intrin.h> // For Visual 2005 # include <intrin.h> /* For Visual 2005 */
# if LZ4_ARCH64 // 64-bits # if LZ4_ARCH64 /* 64-bits */
# pragma intrinsic(_BitScanForward64) // For Visual 2005 # pragma intrinsic(_BitScanForward64) /* For Visual 2005 */
# pragma intrinsic(_BitScanReverse64) // For Visual 2005 # pragma intrinsic(_BitScanReverse64) /* For Visual 2005 */
# else // 32-bits # else /* 32-bits */
# pragma intrinsic(_BitScanForward) // For Visual 2005 # pragma intrinsic(_BitScanForward) /* For Visual 2005 */
# pragma intrinsic(_BitScanReverse) // For Visual 2005 # pragma intrinsic(_BitScanReverse) /* For Visual 2005 */
# endif # endif
# pragma warning(disable : 4127) // disable: C4127: conditional expression is constant # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */
# pragma warning(disable : 4701) // disable: C4701: potentially uninitialized local variable used # pragma warning(disable : 4701) /* disable: C4701: potentially uninitialized local variable used */
#else #else
# ifdef __GNUC__ # ifdef __GNUC__
# define FORCE_INLINE static inline __attribute__((always_inline)) # define FORCE_INLINE static inline __attribute__((always_inline))
@ -114,24 +128,24 @@
# endif # endif
#endif #endif
#ifdef _MSC_VER // Visual Studio #ifdef _MSC_VER /* Visual Studio */
# define lz4_bswap16(x) _byteswap_ushort(x) # define lz4_bswap16(x) _byteswap_ushort(x)
#else #else
# define lz4_bswap16(x) ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8))) # define lz4_bswap16(x) ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))
#endif #endif
//************************************** /**************************************
// Includes Includes
//************************************** **************************************/
#include "lz4hc.h" #include "lz4hc.h"
#include "lz4.h" #include "lz4.h"
//************************************** /**************************************
// Basic Types Basic Types
//************************************** **************************************/
#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L // C99 #if defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */
# include <stdint.h> # include <stdint.h>
typedef uint8_t BYTE; typedef uint8_t BYTE;
typedef uint16_t U16; typedef uint16_t U16;
@ -173,9 +187,9 @@ typedef struct _U64_S { U64 v; } _PACKED U64_S;
#define A16(x) (((U16_S *)(x))->v) #define A16(x) (((U16_S *)(x))->v)
//************************************** /**************************************
// Constants Constants
//************************************** **************************************/
#define MINMATCH 4 #define MINMATCH 4
#define DICTIONARY_LOGSIZE 16 #define DICTIONARY_LOGSIZE 16
@ -187,8 +201,6 @@ typedef struct _U64_S { U64 v; } _PACKED U64_S;
#define HASHTABLESIZE (1 << HASH_LOG) #define HASHTABLESIZE (1 << HASH_LOG)
#define HASH_MASK (HASHTABLESIZE - 1) #define HASH_MASK (HASHTABLESIZE - 1)
#define MAX_NB_ATTEMPTS 256
#define ML_BITS 4 #define ML_BITS 4
#define ML_MASK (size_t)((1U<<ML_BITS)-1) #define ML_MASK (size_t)((1U<<ML_BITS)-1)
#define RUN_BITS (8-ML_BITS) #define RUN_BITS (8-ML_BITS)
@ -205,25 +217,21 @@ typedef struct _U64_S { U64 v; } _PACKED U64_S;
#define GB *(1U<<30) #define GB *(1U<<30)
//************************************** /**************************************
// Architecture-specific macros Architecture-specific macros
//************************************** **************************************/
#if LZ4_ARCH64 // 64-bit #if LZ4_ARCH64 /* 64-bit */
# define STEPSIZE 8 # define STEPSIZE 8
# define LZ4_COPYSTEP(s,d) A64(d) = A64(s); d+=8; s+=8; # define LZ4_COPYSTEP(s,d) A64(d) = A64(s); d+=8; s+=8;
# define LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d) # define LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d)
# define UARCH U64
# define AARCH A64 # define AARCH A64
# define HTYPE U32 # define HTYPE U32
# define INITBASE(b,s) const BYTE* const b = s # define INITBASE(b,s) const BYTE* const b = s
#else // 32-bit #else /* 32-bit */
# define STEPSIZE 4 # define STEPSIZE 4
# define LZ4_COPYSTEP(s,d) A32(d) = A32(s); d+=4; s+=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 LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d); LZ4_COPYSTEP(s,d);
# define UARCH U32
# define AARCH A32 # define AARCH A32
//# define HTYPE const BYTE*
//# define INITBASE(b,s) const int b = 0
# define HTYPE U32 # define HTYPE U32
# define INITBASE(b,s) const BYTE* const b = s # define INITBASE(b,s) const BYTE* const b = s
#endif #endif
@ -231,15 +239,15 @@ typedef struct _U64_S { U64 v; } _PACKED U64_S;
#if defined(LZ4_BIG_ENDIAN) #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_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; } # 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_READ_LITTLEENDIAN_16(d,s,p) { d = (s) - A16(p); }
# define LZ4_WRITE_LITTLEENDIAN_16(p,v) { A16(p) = v; p+=2; } # define LZ4_WRITE_LITTLEENDIAN_16(p,v) { A16(p) = v; p+=2; }
#endif #endif
//************************************************************ /**************************************
// Local Types Local Types
//************************************************************ **************************************/
typedef struct typedef struct
{ {
const BYTE* inputBuffer; const BYTE* inputBuffer;
@ -251,9 +259,9 @@ typedef struct
} LZ4HC_Data_Structure; } LZ4HC_Data_Structure;
//************************************** /**************************************
// Macros Macros
//************************************** **************************************/
#define LZ4_WILDCOPY(s,d,e) do { LZ4_COPYPACKET(s,d) } while (d<e); #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 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)) #define HASH_FUNCTION(i) (((i) * 2654435761U) >> ((MINMATCH*8)-HASH_LOG))
@ -263,9 +271,9 @@ typedef struct
#define GETNEXT(p) ((p) - (size_t)DELTANEXT(p)) #define GETNEXT(p) ((p) - (size_t)DELTANEXT(p))
//************************************** /**************************************
// Private functions Private functions
//************************************** **************************************/
#if LZ4_ARCH64 #if LZ4_ARCH64
FORCE_INLINE int LZ4_NbCommonBytes (register U64 val) FORCE_INLINE int LZ4_NbCommonBytes (register U64 val)
@ -332,6 +340,11 @@ FORCE_INLINE int LZ4_NbCommonBytes (register U32 val)
#endif #endif
int LZ4_sizeofStreamStateHC()
{
return sizeof(LZ4HC_Data_Structure);
}
FORCE_INLINE void LZ4_initHC (LZ4HC_Data_Structure* hc4, const BYTE* base) FORCE_INLINE void LZ4_initHC (LZ4HC_Data_Structure* hc4, const BYTE* base)
{ {
MEM_INIT((void*)hc4->hashTable, 0, sizeof(hc4->hashTable)); MEM_INIT((void*)hc4->hashTable, 0, sizeof(hc4->hashTable));
@ -342,6 +355,13 @@ FORCE_INLINE void LZ4_initHC (LZ4HC_Data_Structure* hc4, const BYTE* base)
hc4->end = base; hc4->end = base;
} }
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* inputBuffer) void* LZ4_createHC (const char* inputBuffer)
{ {
@ -358,7 +378,7 @@ int LZ4_freeHC (void* LZ4HC_Data)
} }
// Update chains up to ip (excluded) /* Update chains up to ip (excluded) */
FORCE_INLINE void LZ4HC_Insert (LZ4HC_Data_Structure* hc4, const BYTE* ip) FORCE_INLINE void LZ4HC_Insert (LZ4HC_Data_Structure* hc4, const BYTE* ip)
{ {
U16* chainTable = hc4->chainTable; U16* chainTable = hc4->chainTable;
@ -381,12 +401,12 @@ char* LZ4_slideInputBufferHC(void* LZ4HC_Data)
{ {
LZ4HC_Data_Structure* hc4 = (LZ4HC_Data_Structure*)LZ4HC_Data; LZ4HC_Data_Structure* hc4 = (LZ4HC_Data_Structure*)LZ4HC_Data;
U32 distance = (U32)(hc4->end - hc4->inputBuffer) - 64 KB; U32 distance = (U32)(hc4->end - hc4->inputBuffer) - 64 KB;
distance = (distance >> 16) << 16; // Must be a multiple of 64 KB distance = (distance >> 16) << 16; /* Must be a multiple of 64 KB */
LZ4HC_Insert(hc4, hc4->end - MINMATCH); LZ4HC_Insert(hc4, hc4->end - MINMATCH);
memcpy((void*)(hc4->end - 64 KB - distance), (const void*)(hc4->end - 64 KB), 64 KB); memcpy((void*)(hc4->end - 64 KB - distance), (const void*)(hc4->end - 64 KB), 64 KB);
hc4->nextToUpdate -= distance; hc4->nextToUpdate -= distance;
hc4->base -= 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; int i;
hc4->base += 1 GB; hc4->base += 1 GB;
@ -403,7 +423,7 @@ FORCE_INLINE size_t LZ4HC_CommonLength (const BYTE* p1, const BYTE* p2, const BY
while (p1t<matchlimit-(STEPSIZE-1)) 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; } if (!diff) { p1t+=STEPSIZE; p2+=STEPSIZE; continue; }
p1t += LZ4_NbCommonBytes(diff); p1t += LZ4_NbCommonBytes(diff);
return (p1t - p1); return (p1t - p1);
@ -415,26 +435,26 @@ FORCE_INLINE size_t LZ4HC_CommonLength (const BYTE* p1, const BYTE* p2, const BY
} }
FORCE_INLINE 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; U16* const chainTable = hc4->chainTable;
HTYPE* const HashTable = hc4->hashTable; HTYPE* const HashTable = hc4->hashTable;
const BYTE* ref; const BYTE* ref;
INITBASE(base,hc4->base); INITBASE(base,hc4->base);
int nbAttempts=MAX_NB_ATTEMPTS; int nbAttempts=maxNbAttempts;
size_t repl=0, ml=0; 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); LZ4HC_Insert(hc4, ip);
ref = HASH_POINTER(ip); ref = HASH_POINTER(ip);
#define REPEAT_OPTIMIZATION #define REPEAT_OPTIMIZATION
#ifdef REPEAT_OPTIMIZATION #ifdef REPEAT_OPTIMIZATION
// Detect repetitive sequences of length <= 4 /* Detect repetitive sequences of length <= 4 */
if ((U32)(ip-ref) <= 4) // potential repetition if ((U32)(ip-ref) <= 4) /* potential repetition */
{ {
if (A32(ref) == A32(ip)) // confirmed if (A32(ref) == A32(ip)) /* confirmed */
{ {
delta = (U16)(ip-ref); delta = (U16)(ip-ref);
repl = ml = LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit) + MINMATCH; repl = ml = LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit) + MINMATCH;
@ -457,7 +477,7 @@ FORCE_INLINE int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const
} }
#ifdef REPEAT_OPTIMIZATION #ifdef REPEAT_OPTIMIZATION
// Complete table /* Complete table */
if (repl) if (repl)
{ {
const BYTE* ptr = ip; const BYTE* ptr = ip;
@ -466,13 +486,13 @@ FORCE_INLINE int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const
end = ip + repl - (MINMATCH-1); end = ip + repl - (MINMATCH-1);
while(ptr < end-delta) while(ptr < end-delta)
{ {
DELTANEXT(ptr) = delta; // Pre-Load DELTANEXT(ptr) = delta; /* Pre-Load */
ptr++; ptr++;
} }
do do
{ {
DELTANEXT(ptr) = delta; DELTANEXT(ptr) = delta;
HashTable[HASH_VALUE(ptr)] = (HTYPE)((ptr) - base); // Head of chain HashTable[HASH_VALUE(ptr)] = (HTYPE)((ptr) - base); /* Head of chain */
ptr++; ptr++;
} while(ptr < end); } while(ptr < end);
hc4->nextToUpdate = end; hc4->nextToUpdate = end;
@ -483,16 +503,16 @@ FORCE_INLINE int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const
} }
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) 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; U16* const chainTable = hc4->chainTable;
HTYPE* const HashTable = hc4->hashTable; HTYPE* const HashTable = hc4->hashTable;
INITBASE(base,hc4->base); INITBASE(base,hc4->base);
const BYTE* ref; const BYTE* ref;
int nbAttempts = MAX_NB_ATTEMPTS; int nbAttempts = maxNbAttempts;
int delta = (int)(ip-startLimit); int delta = (int)(ip-startLimit);
// First Match /* First Match */
LZ4HC_Insert(hc4, ip); LZ4HC_Insert(hc4, ip);
ref = HASH_POINTER(ip); ref = HASH_POINTER(ip);
@ -509,7 +529,7 @@ FORCE_INLINE int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const
while (ipt<matchlimit-(STEPSIZE-1)) 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; } if (!diff) { ipt+=STEPSIZE; reft+=STEPSIZE; continue; }
ipt += LZ4_NbCommonBytes(diff); ipt += LZ4_NbCommonBytes(diff);
goto _endCount; goto _endCount;
@ -520,7 +540,7 @@ FORCE_INLINE int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const
_endCount: _endCount:
reft = ref; reft = ref;
#else #else
// Easier for code maintenance, but unfortunately slower too /* Easier for code maintenance, but unfortunately slower too */
const BYTE* startt = ip; const BYTE* startt = ip;
const BYTE* reft = ref; const BYTE* reft = ref;
const BYTE* ipt = ip + MINMATCH + LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit); const BYTE* ipt = ip + MINMATCH + LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit);
@ -556,26 +576,26 @@ FORCE_INLINE int LZ4HC_encodeSequence (
int length; int length;
BYTE* token; BYTE* token;
// Encode Literal length /* Encode Literal length */
length = (int)(*ip - *anchor); length = (int)(*ip - *anchor);
token = (*op)++; token = (*op)++;
if ((limitedOutputBuffer) && ((*op + length + (2 + 1 + LASTLITERALS) + (length>>8)) > oend)) return 1; // Check output limit 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; } 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); else *token = (BYTE)(length<<ML_BITS);
// Copy Literals /* Copy Literals */
LZ4_BLINDCOPY(*anchor, *op, length); LZ4_BLINDCOPY(*anchor, *op, length);
// Encode Offset /* Encode Offset */
LZ4_WRITE_LITTLEENDIAN_16(*op,(U16)(*ip-ref)); LZ4_WRITE_LITTLEENDIAN_16(*op,(U16)(*ip-ref));
// Encode MatchLength /* Encode MatchLength */
length = (int)(matchLength-MINMATCH); length = (int)(matchLength-MINMATCH);
if ((limitedOutputBuffer) && (*op + (1 + LASTLITERALS) + (length>>8) > oend)) return 1; // Check output limit 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; } 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); else *token += (BYTE)(length);
// Prepare next loop /* Prepare next loop */
*ip += matchLength; *ip += matchLength;
*anchor = *ip; *anchor = *ip;
@ -583,12 +603,14 @@ FORCE_INLINE int LZ4HC_encodeSequence (
} }
#define MAX_COMPRESSION_LEVEL 16
static int LZ4HC_compress_generic ( static int LZ4HC_compress_generic (
void* ctxvoid, void* ctxvoid,
const char* source, const char* source,
char* dest, char* dest,
int inputSize, int inputSize,
int maxOutputSize, int maxOutputSize,
int compressionLevel,
limitedOutput_directive limit limitedOutput_directive limit
) )
{ {
@ -602,6 +624,7 @@ static int LZ4HC_compress_generic (
BYTE* op = (BYTE*) dest; BYTE* op = (BYTE*) dest;
BYTE* const oend = op + maxOutputSize; 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; int ml, ml2, ml3, ml0;
const BYTE* ref=NULL; const BYTE* ref=NULL;
const BYTE* start2=NULL; const BYTE* start2=NULL;
@ -612,29 +635,29 @@ static int LZ4HC_compress_generic (
const BYTE* ref0; const BYTE* ref0;
// Ensure blocks follow each other /* Ensure blocks follow each other */
if (ip != ctx->end) return 0; if (ip != ctx->end) return 0;
ctx->end += inputSize; ctx->end += inputSize;
ip++; ip++;
// Main Loop /* Main Loop */
while (ip < mflimit) while (ip < mflimit)
{ {
ml = LZ4HC_InsertAndFindBestMatch (ctx, ip, matchlimit, (&ref)); ml = LZ4HC_InsertAndFindBestMatch (ctx, ip, matchlimit, (&ref), maxNbAttempts);
if (!ml) { ip++; continue; } if (!ml) { ip++; continue; }
// saved, in case we would skip too much /* saved, in case we would skip too much */
start0 = ip; start0 = ip;
ref0 = ref; ref0 = ref;
ml0 = ml; ml0 = ml;
_Search2: _Search2:
if (ip+ml < mflimit) if (ip+ml < mflimit)
ml2 = LZ4HC_InsertAndGetWiderMatch(ctx, ip + ml - 2, ip + 1, matchlimit, ml, &ref2, &start2); ml2 = LZ4HC_InsertAndGetWiderMatch(ctx, ip + ml - 2, ip + 1, matchlimit, ml, &ref2, &start2, maxNbAttempts);
else ml2 = ml; else ml2 = ml;
if (ml2 == ml) // No better match if (ml2 == ml) /* No better match */
{ {
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0; if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
continue; continue;
@ -642,7 +665,7 @@ _Search2:
if (start0 < ip) if (start0 < ip)
{ {
if (start2 < ip + ml0) // empirical if (start2 < ip + ml0) /* empirical */
{ {
ip = start0; ip = start0;
ref = ref0; ref = ref0;
@ -650,8 +673,8 @@ _Search2:
} }
} }
// Here, start0==ip /* Here, start0==ip */
if ((start2 - ip) < 3) // First Match too small : removed if ((start2 - ip) < 3) /* First Match too small : removed */
{ {
ml = ml2; ml = ml2;
ip = start2; ip = start2;
@ -660,9 +683,11 @@ _Search2:
} }
_Search3: _Search3:
// Currently we have : /*
// ml2 > ml1, and * Currently we have :
// ip1+3 <= ip2 (usually < ip1+ml1) * ml2 > ml1, and
* ip1+3 <= ip2 (usually < ip1+ml1)
*/
if ((start2 - ip) < OPTIMAL_ML) if ((start2 - ip) < OPTIMAL_ML)
{ {
int correction; int correction;
@ -677,26 +702,26 @@ _Search3:
ml2 -= correction; ml2 -= correction;
} }
} }
// Now, we have start2 = ip+new_ml, with new_ml = min(ml, OPTIMAL_ML=18) /* Now, we have start2 = ip+new_ml, with new_ml = min(ml, OPTIMAL_ML=18) */
if (start2 + ml2 < mflimit) if (start2 + ml2 < mflimit)
ml3 = LZ4HC_InsertAndGetWiderMatch(ctx, start2 + ml2 - 3, start2, matchlimit, ml2, &ref3, &start3); ml3 = LZ4HC_InsertAndGetWiderMatch(ctx, start2 + ml2 - 3, start2, matchlimit, ml2, &ref3, &start3, maxNbAttempts);
else ml3 = ml2; else ml3 = ml2;
if (ml3 == ml2) // No better match : 2 sequences to encode if (ml3 == ml2) /* No better match : 2 sequences to encode */
{ {
// ip & ref are known; Now for ml /* ip & ref are known; Now for ml */
if (start2 < ip+ml) ml = (int)(start2 - ip); if (start2 < ip+ml) ml = (int)(start2 - ip);
// Now, encode 2 sequences /* Now, encode 2 sequences */
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0; if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
ip = start2; ip = start2;
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml2, ref2, limit, oend)) return 0; if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml2, ref2, limit, oend)) return 0;
continue; continue;
} }
if (start3 < ip+ml+3) // Not enough space for match 2 : remove it 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 (start3 >= (ip+ml)) /* can write Seq1 immediately ==> Seq2 is removed, so Seq3 becomes Seq1 */
{ {
if (start2 < ip+ml) if (start2 < ip+ml)
{ {
@ -729,8 +754,10 @@ _Search3:
goto _Search3; goto _Search3;
} }
// OK, now we have 3 ascending matches; let's write at least the first one /*
// ip & ref are known; Now for ml * 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+ml)
{ {
if ((start2 - ip) < (int)ML_MASK) if ((start2 - ip) < (int)ML_MASK)
@ -765,53 +792,101 @@ _Search3:
} }
// Encode Last Literals /* Encode Last Literals */
{ {
int lastRun = (int)(iend - anchor); 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 ((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; } 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); else *op++ = (BYTE)(lastRun<<ML_BITS);
memcpy(op, anchor, iend - anchor); memcpy(op, anchor, iend - anchor);
op += iend-anchor; op += iend-anchor;
} }
// End /* End */
return (int) (((char*)op)-dest); return (int) (((char*)op)-dest);
} }
int LZ4_compressHC(const char* source, char* dest, int inputSize) int LZ4_compressHC2(const char* source, char* dest, int inputSize, int compressionLevel)
{ {
void* ctx = LZ4_createHC(source); void* ctx = LZ4_createHC(source);
int result; int result;
if (ctx==NULL) return 0; if (ctx==NULL) return 0;
result = LZ4HC_compress_generic (ctx, source, dest, inputSize, 0, noLimit); result = LZ4HC_compress_generic (ctx, source, dest, inputSize, 0, compressionLevel, noLimit);
LZ4_freeHC(ctx); LZ4_freeHC(ctx);
return result; return result;
} }
int LZ4_compressHC_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize) int LZ4_compressHC(const char* source, char* dest, int inputSize) { return LZ4_compressHC2(source, dest, inputSize, 0); }
{
return LZ4HC_compress_generic (LZ4HC_Data, source, dest, inputSize, 0, noLimit);
}
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) int LZ4_compressHC_limitedOutput(const char* source, char* dest, int inputSize, int maxOutputSize)
{ {
void* ctx = LZ4_createHC(source); return LZ4_compressHC2_limitedOutput(source, dest, inputSize, maxOutputSize, 0);
int result; }
if (ctx==NULL) return 0;
result = LZ4HC_compress_generic (ctx, source, dest, inputSize, maxOutputSize, limitedOutput);
LZ4_freeHC(ctx); /*****************************
return result; 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) 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, limitedOutput); 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);
}

66
ext/lz4/lz4hc.h
View File

@ -1,7 +1,7 @@
/* /*
LZ4 HC - High Compression Mode of LZ4 LZ4 HC - High Compression Mode of LZ4
Header File 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) BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
Redistribution and use in source and binary forms, with or without Redistribution and use in source and binary forms, with or without
@ -63,19 +63,57 @@ 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 : /* Note :
Decompression functions are provided within LZ4 source code (see "lz4.h") (BSD license) 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);
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); void* LZ4_createHC (const char* inputBuffer);
int LZ4_compressHC_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize); 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); int LZ4_compressHC_limitedOutput_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int maxOutputSize);
char* LZ4_slideInputBufferHC (void* LZ4HC_Data); char* LZ4_slideInputBufferHC (void* LZ4HC_Data);
int LZ4_freeHC (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. 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 : In order to achieve this, it is necessary to start creating the LZ4HC Data Structure, thanks to the function :
@ -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. 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) #if defined (__cplusplus)
} }

2
node/Packet.hpp
View File

@ -993,7 +993,7 @@ public:
if ((compressed())&&(size() >= ZT_PROTO_MIN_PACKET_LENGTH)) { if ((compressed())&&(size() >= ZT_PROTO_MIN_PACKET_LENGTH)) {
if (size() > ZT_PACKET_IDX_PAYLOAD) { if (size() > ZT_PACKET_IDX_PAYLOAD) {
unsigned int compLen = size() - ZT_PACKET_IDX_PAYLOAD; unsigned int compLen = size() - ZT_PACKET_IDX_PAYLOAD;
int ucl = LZ4_uncompress_unknownOutputSize((const char *)field(ZT_PACKET_IDX_PAYLOAD,compLen),(char *)buf,compLen,sizeof(buf)); int ucl = LZ4_decompress_safe((const char *)field(ZT_PACKET_IDX_PAYLOAD,compLen),(char *)buf,compLen,sizeof(buf));
if ((ucl > 0)&&(ucl <= (int)(capacity() - ZT_PACKET_IDX_PAYLOAD))) { if ((ucl > 0)&&(ucl <= (int)(capacity() - ZT_PACKET_IDX_PAYLOAD))) {
setSize((unsigned int)ucl + ZT_PACKET_IDX_PAYLOAD); setSize((unsigned int)ucl + ZT_PACKET_IDX_PAYLOAD);
memcpy(field(ZT_PACKET_IDX_PAYLOAD,(unsigned int)ucl),buf,ucl); memcpy(field(ZT_PACKET_IDX_PAYLOAD,(unsigned int)ucl),buf,ucl);