AuroraMiddleware/lz4
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
2,560 Commits 13 Branches 31 Tags 6.8 MiB
dfc431fb3d
Commit Graph

351 Commits

Author SHA1 Message Date
Yann Collet
dfc431fb3d fix NULL ptr arithmetic at lz4:2299 2021-05-28 01:10:41 -07:00
Yann Collet
539c783c98 fix NULL ptr arithmetic in lz4:1680 
only do arithmetic if offset > 0
 2021-05-28 01:08:18 -07:00
Yann Collet
59273b7127 fix UB lz4:988 and lz4:1178 
ensure `dictBase` is only used
when there is an actual dictionary content.
 2021-05-28 00:56:26 -07:00
Yann Collet
c2c0c47d5f fix NULL ptr arithmetic of lz4:1572 
was blindly adding an offset (0) to `dictionary` which could be `NULL`.
 2021-05-27 23:20:28 -07:00
Jasper Lievisse Adriaanse
8301a21773 Fix potential memory corruption with negative memmove() size 2021-02-26 15:21:20 +01:00
Thomas Waldmann
909aae8260 fix some typos (work by Andrea Gelmini) 2021-01-07 18:39:57 +01:00
Yann Collet
50ff8b3ae8 fix minor header date 2020-11-30 18:16:00 -08:00
Yann Collet
e0f85f2fc8 better visual conformance 
only include <intrin.h> on vs2005+ (#947)
remove some useless #pragma
fix a few minor Visual warnings
 2020-11-14 16:27:57 -08:00
Yann Collet
581c459b4e restrict BitScanForward() to VS2005+ 
suggested by @aqrit in #947
 2020-11-14 14:26:08 -08:00
Yann Collet
a296839802 changed LZ4_calloc() to a 2-arguments signature 
to remain similar to stdlib's calloc().

Updated test to use c++ compiler for stricter signature check.
 2020-11-09 08:44:26 -08:00
Yann Collet
44b13db532 Merge branch 'dev' into customMem 2020-11-08 23:26:19 -08:00
Yann Collet
52646e8d75 first proposal for LZ4_USER_MEMORY_FUNCTIONS 
makes it possible to replace at link time
malloc, calloc and free
by user-provided functions
which must be named LZ4_malloc(), LZ4_calloc() and LZ4_free().

answer #937
 2020-11-08 21:17:32 -08:00
Yann Collet
9cf3f106a8
Merge pull request #944 from lz4/fix874 
fix #874
 2020-11-08 20:48:20 -08:00
Yann Collet
2a2b10f192 fixed remaining ubsan warnings 2020-11-08 18:08:43 -08:00
Yann Collet
2964b8a6f6 fix #874 
coverity reported a warning regarding a memcpy() overwrite.
This is a false positive (the memory area is large enough),
but it's true that it's not trivial to determine (encompassing struct),
and it's proper anyway to only memcpy() just the right amount of data.
 2020-11-08 13:21:58 -08:00
Yann Collet
e251a84025 fix minor UBs 
- check alignment before casting a pointer
- saveDict : don't memmove() on NULL dst
 2020-11-07 19:42:57 -08:00
Yann Collet
1d02141bf8
Merge pull request #941 from lz4/revertinline 
Revert "Replace "static" to "LZ4_FORCE_INLINE" for small functions"
 2020-11-07 17:18:04 -08:00
Yann Collet
f61eeb7793 Revert "Replace "static" to "LZ4_FORCE_INLINE" for small functions" 
This reverts commit 0e3933edd4.
 2020-11-07 11:02:30 -08:00
Yann Collet
d4bfcf8489 fix #935 
minor: identical declaration and prototypes of `LZ4HC_compress_optimal()`

also :
very minor optimization of `LZ4_memcpy_using_offset()`
 2020-11-07 10:06:52 -08:00
Yann Collet
b5e2a4acd9
Merge pull request #936 from lz4/alignTest 
More alignment tests
 2020-11-06 20:27:42 -08:00
Yann Collet
211d653ff8 re-enable alignment test on all targets 2020-11-06 16:43:14 -08:00
Yann Collet
e968a24129 unified alignment test 
across lz4.c and lz4hc.c
 2020-11-06 14:46:48 -08:00
remittor
0e3933edd4 Replace "static" to "LZ4_FORCE_INLINE" for small functions 
The "static" specifier does not guarantee that the function will be inlined.
 2020-10-07 09:52:40 +03:00
remittor
749bd91a06 Replace define LZ4_FORCE_O2_INLINE_GCC_PPC64LE to LZ4_FORCE_INLINE 
There is no reason to separate these two definitions!
 2020-10-07 09:51:08 +03:00
remittor
c4792cdfa9 Fix: The "inline" specifier do not use for LZ4_wildCopy8 and LZ4_wildCopy32 
This problem was reproduced on MSVC 2015 (32-bit). Both functions were called using the operator "call".
 2020-10-06 17:16:43 +03:00
Yann Collet
7d21f761c3 fix conversion warning 2020-09-29 21:53:42 -07:00
Yann Collet
ad2d2764c7 fix minor static analyzer warnings 
detected by scan-build, cppcheck and advanved compilation flags
fix #786
 2020-09-29 17:20:52 -07:00
Yann Collet
78f4fdbb89
Merge pull request #923 from lz4/fix784 
fix efficiency of LZ4_compress_HC_destSize()
 2020-09-28 14:04:56 -07:00
Yann Collet
ab89dda91d improved last literals run on LZ4_compress_destSize 
applying new more accurate formula from LZ4_compress_HC_destSize()

also : fix some minor display issue in tests/frametest
 2020-09-28 11:39:00 -07:00
Yann Collet
89736e4e27 ensure last match not too close to end 
must respect MFLIMIT distance from oend
 2020-09-27 23:59:56 -07:00
Yann Collet
8a362a8ac8
Merge pull request #921 from lz4/doubleNull 
fix compressing into NULL
 2020-09-27 21:09:06 -07:00
Yann Collet
e7fe105ac6 fix efficiency of LZ4_compress_HC_destSize() 
LZ4_compress_HC_destSize() had a tendency
to discard its last match when this match overflowed specified dstBuffer limit.
The impact is generally moderate,
but occasionally huge,
typically when this last match is very large
(such as compressing a bunch of zeroes).

Issue #784 fixed for both Chain and Opt implementations.

Added a unit test suggested by @remittor checking this topic.
 2020-09-27 21:04:40 -07:00
Anton Kochkov
9730d91110 Fix compilation with TinyCC 2020-09-27 17:07:51 +08:00
Yann Collet
ee4f37d284 fix compressing into NULL 
fails properly
bug discovered by oss-fuzz
 2020-09-26 11:31:57 -07:00
Yann Collet
10d2e1c694 fixed lz4frame with blocks of size 1 
properly track history
 2020-09-17 14:43:02 -07:00
Yann Collet
ee01df1271 added the actual code change 2020-09-16 23:46:39 -07:00
Yann Collet
c5d6f8a8be fix #783 
LZ4_decompress_safe_partial()
now also supports a scenario where
nb_bytes_to_generate is <= block_decompressed_size
And
nb_bytes_to_read is >= block_compressed_size.

Previously, the only supported scenario was
nb_bytes_to_read == block_compress_size.

Pay attention that,
if nb_bytes_to_read is > block_compressed_size,
then, necessarily, it requires that
nb_bytes_to_generate is <= block_decompress_size.
If both are larger, it will generate corrupted data.
 2020-08-27 00:17:57 -07:00
Yann Collet
3e3a006c6f Merge branch 'dev' into extraInput 2020-08-26 23:20:28 -07:00
Yann Collet
5243173b23 added documentation about LZ4_FORCE_SW_BITCOUNT 
Also : added memory-frugal software byte count for big endian 64-bit cpus.
Disabled by default.
 2020-08-25 22:17:29 -07:00
Yann Collet
ee0a3cfa0c
Merge pull request #898 from aqrit/aqrit-prefixlen 
rejigger bit counting intrinsics
 2020-08-24 15:13:18 -07:00
Yann Collet
0002563be8 removed LZ4_compress_fast_force() 
which serves no more purpose.

The comment implies that the simple presence of this unused function was affecting performance,
and that's the reason why it was not removed earlier.
This is likely another side effect of instruction alignment.

It's obviously unreliable to rely on it in this way,
meaning that the impact will be different, positive of negative,
with any minor code change, and any minor compiler version change, even parameter change.
 2020-08-21 19:23:49 -07:00
aqrit
e45defa8bd
silence warning 
MSVC debug mode complains
 2020-08-17 17:53:07 -04:00
BellaXlp
ab713923a2
fix issue #783 (#862) 
* fix issue #783
 2020-08-12 14:42:10 -07:00
aqrit
e72897402d
rejigger bit counting intrinsics 
Fix lz4/lz4#867
Optimize software fallback routines.
Delete some faulty (and dead?) MSVC big endian code.
 2020-08-11 21:14:09 -04:00
Yann Collet
b26c140a54
Merge pull request #895 from lz4/hugefast 
Fix #876
 2020-08-10 12:52:32 -07:00
Yann Collet
7b1b078dfc fix #876 
by introducing a max limit acceleration value
 2020-08-10 11:03:27 -07:00
W. Felix Handte
a78235e6ad Fix Enum Casts 
Fixes `-Wsign-compare` issues.
 2020-08-10 13:47:08 -04:00
W. Felix Handte
9af86f0841 Remove dirty Field From LZ4_stream_t 2020-08-06 16:06:40 -04:00
W. Felix Handte
d7399232a4 Remove Extraneous Reset in LZ4_attach_dictionary() 
Nothing internally sets dirty anymore. The only way to get that is if you use
an uninitialized context, in which case your warranty is void anyways.
 2020-08-05 12:46:32 -04:00
Nick Terrell
fe2a1b3707 Call LZ4_memcpy() instead of memcpy() 
`LZ4_memcpy()` uses `__builtin_memcpy()` to ensure that clang/gcc
can inline the `memcpy()` calls in freestanding mode.

This is necessary for decompressing the Linux Kernel with LZ4.
Without an analogous patch decompression ran at 77 MB/s, and with
the patch it ran at 884 MB/s.
 2020-08-03 11:28:02 -07:00