Commit Graph

840 Commits

Author SHA1 Message Date
Nick Terrell
38c3945de3 [lz4hc] Only allow chain swapping forwards
When the match is very long and found quickly, we can do
matchLength * nbCompares iterations through the chain
swapping, which can really slow down compression.
2019-07-31 10:17:26 -07:00
Nick Terrell
be1738aa46 [lz4hc] Fix pattern detection end of dictionary
The pattern detection in extDict mode could put `matchIndex`
within the last 3 bytes of the dictionary. This would cause
a read out of bounds.
2019-07-31 10:17:21 -07:00
Nick Terrell
58ea585878 [lz4hc] Fix minor pessimization in extDict pattern matching
We should be comparing `matchPtr` not `ip`. This bug just means
that this branch was not taken, so we might miss some of the
forward length.
2019-07-31 10:16:25 -07:00
Nick Terrell
7e97bf377d [lz4hc] Improve pattern detection in ext dict
It is important to continue to look backwards if the current pattern
reaches `lowPrefixPtr`. If the pattern detection doesn't go all the
way to the beginning of the pattern, or the end of the pattern it
slows down the search instead of speeding it up.

The slow unit in `round_trip_stream_fuzzer` used to take 12 seconds
to run with -O3, now it takes 0.2 seconds.

Credit to OSS-Fuzz
2019-07-31 10:16:21 -07:00
Nick Terrell
4c1d4c437d [LZ4HC] Speed up pattern compression with external dictionary
Fixes #761.
2019-07-24 10:59:20 -07:00
Yann Collet
fb8a159436
Merge pull request #763 from terrelln/unused
[lz4frame] Fix unused variable warnings in fuzzing mode
2019-07-19 16:54:01 -07:00
Yann Collet
7a516411d4
Merge pull request #760 from terrelln/destSize
[LZ4_compress_destSize] Fix off-by-one error in fix
2019-07-19 15:22:51 -07:00
Nick Terrell
87e52f7d5d [lz4frame] Fix unused variable warnings in fuzzing mode 2019-07-19 14:44:06 -07:00
Nick Terrell
b487660309 [lz4frame] Skip magic and checksums in fuzzing mode
When `FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION` is defined we skip
magic and checksum checks. This makes it easier to fuzz decompression.
2019-07-18 18:45:32 -07:00
Nick Terrell
1f236e0790 Fix LZ4_attach_dictionary with empty dictionary 2019-07-18 12:29:15 -07:00
Nick Terrell
7c32101c65 [LZ4_compress_destSize] Fix off-by-one error in fix
The next match is looking at the current ip, not the next ip,
so it needs to be cleared as well.

Credit to OSS-Fuzz
2019-07-18 12:20:29 -07:00
W. Felix Handte
40943ba0c9 Unconditionally Clear dictCtx 2019-07-18 13:35:12 -04:00
W. Felix Handte
369fb3900c Fix Data Corruption Bug when Streaming with an Attached Dict in HC Mode
This diff fixes an issue in which we failed to clear the `dictCtx` in HC
compression. The `dictCtx` is not supposed to be used when an `extDict` is
present: matches found in the `dictCtx` do not account for the presence of an
`extDict` segment, and their offsets are therefore miscalculated when one is
present. This can lead to data corruption.

This diff clears the `dictCtx` whenever setting an `extDict`.

This issue was uncovered by @terrelln's fuzzing work.
2019-07-18 12:48:41 -04:00
Nick Terrell
13a2d9e34f [LZ4_compress_destSize] Fix overflow condition 2019-07-17 11:50:47 -07:00
Nick Terrell
6bc6f836a1 [LZ4_compress_destSize] Fix rare data corruption bug 2019-07-17 11:38:38 -07:00
Nick Terrell
690009e2c2 [LZ4_compress_destSize] Allow 2 more bytes of match length 2019-07-17 11:07:24 -07:00
Yann Collet
7654a5a6d2
Merge pull request #752 from terrelln/fuzzers
[ossfuzz] Improve the fuzzers
2019-07-16 11:18:09 -07:00
Nick Terrell
725cb0aafd [lz4] Fix bugs in partial decoding
* Partial decoding could read a few bytes beyond the end of the input
* Partial decoding returned an error with an empty output buffer
2019-07-15 12:21:59 -07:00
Yann Collet
6654c2cd3b ensure conformance with custom LZ4_DISTANCE_MAX
It's now possible to select a custom LZ4_DISTANCE_MAX at compile time,
provided it's <= 65535.

However, in some cases (when compressing in byU16 mode),
the new distance wasn't respected,
as it used to implied that it was necessarily within range.

Added a distance check for this case.
Also : added a new TravisCI test which ensures that
custom LZ4_DISTANCE_MAX compiles correctly
and compresses correctly (relying on `assert()` to find outsized offsets).
2019-07-15 12:11:34 -07:00
Hitatm
8ac954aa71 bugfix: correctly control the offset < LZ4_DISTANCE_MAX,when change the value of LZ4_DISTANCE_MAX, 2019-07-15 22:53:46 +08:00
Sylvestre Ledru
12e5841e76
Remove an useless declaration 2019-07-04 18:13:36 +02:00
Yann Collet
bb5c34a875 bumped version number to v1.9.2
to reduce risks that future bug reports in `dev` branch report `v1.9.1` as the failing version.
2019-07-01 09:01:43 -07:00
Nick Terrell
e72d442300 Fix out-of-bounds read of up to 64 KB in the past 2019-06-28 14:58:35 -07:00
Yann Collet
1d759576b9 precise again that LZ4 decoder needs metadata
and that such metadata must be provided / sent / saved by the application.
2019-06-06 13:20:30 -07:00
Yann Collet
348e107d99 restored FORCE_INLINE 2019-06-04 14:04:49 -07:00
Yann Collet
280fc0856d
Merge pull request #717 from lz4/inplace
Added documentation and macro to support in-place compression and decompression
2019-05-31 12:59:38 -07:00
Yann Collet
5997e139f5 added more details for in-place documentation 2019-05-31 11:56:59 -07:00
Yann Collet
33cb8518ac decompress: changed final memcpy() into memmove()
for compatibility with in-place decompression scenarios.
2019-05-31 11:44:37 -07:00
Chenxi Mao
64b5917736 FAST_DEC_LOOP: only did offset check in specific condition.
When I did FAST_DEC_LOOP performance test, I found the
offset check is much more than v1.8.3

You will see the condition check difference via lzbench with dickens test case.
v1.8.3 34959
v.1.9.x 1055885

After investigate the code, we could see the difference.
v.1.8.3 SKIP the condition check if
if condition is true in:
https://github.com/lz4/lz4/blob/v1.8.3/lib/lz4.c#L1463
AND below condition is true
https://github.com/lz4/lz4/blob/v1.8.3/lib/lz4.c#L1478\
The offset check should be invoked.

v1.9.3
The offset check code will be invoked in every loop which lead to downgrade.
So the fix would be move this check to specific condition
to avoid useless condition check.

After this change, the call number is same as v1.8.3
2019-05-31 08:36:13 +08:00
Yann Collet
676d46df27 updated LZ4_DECOMPRESS_INPLACE_MARGIN
to pass worst case scenario.
Now adds margin proportional to input size to counter local expansion.
2019-05-30 16:19:30 -07:00
Yann Collet
22adbb176a add more doc on in-place (de)compression 2019-05-30 09:45:21 -07:00
Yann Collet
76116495bf some more minor conversion warnings fixes 2019-05-29 13:14:52 -07:00
Yann Collet
444550defa ensure lz4.h can be included with or without LZ4_STATIC_LINKING_ONLY in any order
ensure correct propagation of LZ4_DISTANCE_MAX
2019-05-29 12:21:14 -07:00
Yann Collet
b17f578a91 added comments and macros for in-place (de)compression 2019-05-29 12:06:13 -07:00
Niko Dzhus
2be2fe43a8 fix temporary buffer use when input size hint is respected 2019-05-24 22:08:44 +03:00
Yann Collet
a7151324af
Merge pull request #708 from gabrielstedman/list
Add multiframe report to --list command
2019-05-16 15:56:42 -07:00
gstedman
98a86c8ef6 Add multiframe report to --list command 2019-05-15 21:13:19 +01:00
George Prekas
605d811e6c enable LZ4_FAST_DEC_LOOP build macro on aarch64/GCC by default 2019-05-07 08:36:06 -05:00
Brenden Eng
9e056bc032 Include block checksum in worst case scenario calculation of dstCapacity 2019-04-25 22:37:39 -04:00
Yann Collet
ba99eac4d0 several minor style changes recommended by clang-tidy 2019-04-24 10:03:02 -07:00
Yann Collet
f9781a6643
Merge pull request #697 from ldv-alt/fixes/test-amalgamation
Fix test-amalgamation
2019-04-23 17:20:03 -07:00
Dmitry V. Levin
10726d4c56 test-amalgamation: use a single cat command
Use the list of prerequisites instead of listing those files manually,
this way they will never fall out of sync.

Also update the amalgamation example to use a single cat command.

Fixes: a7e8d394 ("[amalgamation] add test")
Fixes: b192c86b ("[amalgamation] lz4frame.c")
2019-04-23 21:18:11 +00:00
Chongyu Zhu
024216ef73
lib/Makefile: Fix detection of Darwin. 2019-04-24 03:13:19 +08:00
JPeterMugaas
7136489d33 More build imrpvements
Moved a few other tests to Makefiles.inc.  Other things might need to go there.
Made a test for symlink appropriateness.  Windows can NOT handle them the same way Unix-like operating systems do (if at all).  This is mostly the same as the Visual C projects.
embed version info into .dll and .exes that are redistributed.
2019-04-23 07:44:00 -04:00
Yann Collet
c4fe7a2047
Merge pull request #689 from JPeterMugaas/jpm-makefile
Jpm makefile - as described in https://github.com/lz4/lz4/issues/688
2019-04-22 15:29:50 -07:00
JPeterMugaas
2acddd9918 Fix a test for mingw 2019-04-22 16:06:04 -04:00
JPeterMugaas
b48f823f90 Try to put some tests I made in ONE place. I also moved a test for "install" in one place to try to isolate it. 2019-04-22 15:48:48 -04:00
JPeterMugaas
773b66547f Initial commits from diff I submitted earlier 2019-04-22 13:48:59 -04:00
Yann Collet
7a4e3b1fac bumped version number
to v1.9.1
2019-04-19 11:59:49 -07:00
Yann Collet
af35920c81 lz4frame: initializers compatibility with C++
fix #679, reported by @degski
2019-04-19 10:23:50 -07:00