Lasse Collin [Wed, 28 Sep 2022 08:05:15 +0000 (11:05 +0300)]
liblzma: Stream decoder: Fix restarting after LZMA_MEMLIMIT_ERROR.
If lzma_code() returns LZMA_MEMLIMIT_ERROR it is now possible
to use lzma_memlimit_set() to increase the limit and continue
decoding. This was supposed to work from the beginning but
there was a bug. With other decoders (.lzma or threaded .xz)
this already worked correctly.
Lasse Collin [Mon, 19 Sep 2022 16:34:56 +0000 (19:34 +0300)]
liblzma: Add experimental ARM64 BCJ filter with a temporary Filter ID.
That is, the Filter ID will be changed once the design is final.
The current version will be removed. So files created with the
tempoary Filter ID won't be supported in the future.
test_files.sh now runs xz -l for bad-3-index-uncomp-overflow.xz
because only then the previously-buggy code path gets tested.
Normal decompression doesn't use lzma_index_append() at all.
Instead, lzma_index_hash functions are used and those already
did the overflow check.
Lasse Collin [Fri, 9 Sep 2022 11:12:30 +0000 (14:12 +0300)]
Tests: Silence warnings about unused functions from tuktest.h.
Warnings about unused tuktest_run_test conveniently tell which
test programs haven't been converted to tuktest.h yet but I
silenced that warning too for now anyway.
It is fine to use __attribute__((__unused__)) even when the
function is actually used because the attribute only means
that the function might be unused.
Lasse Collin [Fri, 9 Sep 2022 10:51:57 +0000 (13:51 +0300)]
liblzma: lzma_filters_copy: Keep dest[] unmodified if an error occurs.
lzma_stream_encoder() and lzma_stream_encoder_mt() always assumed
this. Before this patch, failing lzma_filters_copy() could result
in free(invalid_pointer) or invalid memory reads in stream_encoder.c
or stream_encoder_mt.c.
To trigger this, allocating memory for a filter options structure
has to fail. These are tiny allocations so in practice they very
rarely fail.
Certain badness in the filter chain array could also make
lzma_filters_copy() fail but both stream_encoder.c and
stream_encoder_mt.c validate the filter chain before
trying to copy it, so the crash cannot occur this way.
The documentation in src/liblzma/api/lzma/index.h suggests that
both the unpadded (compressed) size and the uncompressed size
are checked for overflow, but only the unpadded size was checked.
The uncompressed check is done first since that is more likely to
occur than the unpadded or index field size overflows.
Lasse Collin [Sun, 4 Sep 2022 20:23:00 +0000 (23:23 +0300)]
liblzma: Vaccinate against an ill patch from RHEL/CentOS 7.
RHEL/CentOS 7 shipped with 5.1.2alpha, including the threaded
encoder that is behind #ifdef LZMA_UNSTABLE in the API headers.
In 5.1.2alpha these symbols are under XZ_5.1.2alpha in liblzma.map.
API/ABI compatibility tracking isn't done between development
releases so newer releases didn't have XZ_5.1.2alpha anymore.
Later RHEL/CentOS 7 updated xz to 5.2.2 but they wanted to keep
the exported symbols compatible with 5.1.2alpha. After checking
the ABI changes it turned out that >= 5.2.0 ABI is backward
compatible with the threaded encoder functions from 5.1.2alpha
(but not vice versa as fixes and extensions to these functions
were made between 5.1.2alpha and 5.2.0).
In RHEL/CentOS 7, XZ Utils 5.2.2 was patched with
xz-5.2.2-compat-libs.patch to modify liblzma.map:
- XZ_5.1.2alpha was added with lzma_stream_encoder_mt and
lzma_stream_encoder_mt_memusage. This matched XZ Utils 5.1.2alpha.
- XZ_5.2 was replaced with XZ_5.2.2. It is clear that this was
an error; the intention was to keep using XZ_5.2 (XZ_5.2.2
has never been used in XZ Utils). So XZ_5.2.2 lists all
symbols that were listed under XZ_5.2 before the patch.
lzma_stream_encoder_mt and _mt_memusage are included too so
they are listed both here and under XZ_5.1.2alpha.
The patch didn't add any __asm__(".symver ...") lines to the .c
files. Thus the resulting liblzma.so exports the threaded encoder
functions under XZ_5.1.2alpha only. Listing the two functions
also under XZ_5.2.2 in liblzma.map has no effect without
matching .symver lines.
The lack of XZ_5.2 in RHEL/CentOS 7 means that binaries linked
against unpatched XZ Utils 5.2.x won't run on RHEL/CentOS 7.
This is unfortunate but this alone isn't too bad as the problem
is contained within RHEL/CentOS 7 and doesn't affect users
of other distributions. It could also be fixed internally in
RHEL/CentOS 7.
The second problem is more serious: In XZ Utils 5.2.2 the API
headers don't have #ifdef LZMA_UNSTABLE for obvious reasons.
This is true in RHEL/CentOS 7 version too. Thus now programs
using new APIs can be compiled without an extra #define. However,
the programs end up depending on symbol version XZ_5.1.2alpha
(and possibly also XZ_5.2.2) instead of XZ_5.2 as they would
with an unpatched XZ Utils 5.2.2. This means that such binaries
won't run on other distributions shipping XZ Utils >= 5.2.0 as
they don't provide XZ_5.1.2alpha or XZ_5.2.2; they only provide
XZ_5.2 (and XZ_5.0). (This includes RHEL/CentOS 8 as the patch
luckily isn't included there anymore with XZ Utils 5.2.4.)
Binaries built by RHEL/CentOS 7 users get distributed and then
people wonder why they don't run on some other distribution.
Seems that people have found out about the patch and been copying
it to some build scripts, seemingly curing the symptoms but
actually spreading the illness further and outside RHEL/CentOS 7.
The ill patch seems to be from late 2016 (RHEL 7.3) and in 2017 it
had spread at least to EasyBuild. I heard about the events only
recently. :-(
This commit splits liblzma.map into two versions: one for
GNU/Linux and another for other OSes that can use symbol versioning
(FreeBSD, Solaris, maybe others). The Linux-specific file and the
matching additions to .c files add full compatibility with binaries
that have been built against a RHEL/CentOS-patched liblzma. Builds
for OSes other than GNU/Linux won't get the vaccine as they should
be immune to the problem (I really hope that no build script uses
the RHEL/CentOS 7 patch outside GNU/Linux).
The RHEL/CentOS compatibility symbols XZ_5.1.2alpha and XZ_5.2.2
are intentionally put *after* XZ_5.2 in liblzma_linux.map. This way
if one forgets to #define HAVE_SYMBOL_VERSIONS_LINUX when building,
the resulting liblzma.so.5 will have lzma_stream_encoder_mt@@XZ_5.2
since XZ_5.2 {...} is the first one that lists that function.
Without HAVE_SYMBOL_VERSIONS_LINUX @XZ_5.1.2alpha and @XZ_5.2.2
will be missing but that's still a minor problem compared to
only having lzma_stream_encoder_mt@@XZ_5.1.2alpha!
The "local: *;" line was moved to XZ_5.0 so that it doesn't need
to be moved around. It doesn't matter where it is put.
Having two similar liblzma_*.map files is a bit silly as it is,
at least for now, easily possible to generate the generic one
from the Linux-specific file. But that adds extra steps and
increases the risk of mistakes when supporting more than one
build system. So I rather maintain two files in parallel and let
validate_map.sh check that they are in sync when "make mydist"
is run.
This adds .symver lines for lzma_stream_encoder_mt@XZ_5.2.2 and
lzma_stream_encoder_mt_memusage@XZ_5.2.2 even though these
weren't exported by RHEL/CentOS 7 (only @@XZ_5.1.2alpha was
for these two). I added these anyway because someone might
misunderstand the RHEL/CentOS 7 patch and think that @XZ_5.2.2
(@@XZ_5.2.2) versions were exported too.
At glance one could suggest using __typeof__ to copy the function
prototypes when making aliases. However, this doesn't work trivially
because __typeof__ won't copy attributes (lzma_nothrow, lzma_pure)
and it won't change symbol visibility from hidden to default (done
by LZMA_API()). Attributes could be copied with __copy__ attribute
but that needs GCC 9 and a fallback method would be needed anyway.
This uses __symver__ attribute with GCC >= 10 and
__asm__(".symver ...") with everything else. The attribute method
is required for LTO (-flto) support with GCC. Using -flto with
GCC older than 10 is now broken on GNU/Linux and will not be fixed
(can silently result in a broken liblzma build that has dangerously
incorrect symbol versions). LTO builds with Clang seem to work
with the traditional __asm__(".symver ...") method.
Thanks to Boud Roukema for reporting the problem and discussing
the details and testing the fix.
Lasse Collin [Wed, 31 Aug 2022 13:42:04 +0000 (16:42 +0300)]
CMake: Add xz symlinks.
These are a minor thing especially since the xz build has
some real problems still like lack of large file support
on 32-bit systems but I'll commit this since the code exists.
Lasse Collin [Tue, 12 Jul 2022 14:59:41 +0000 (17:59 +0300)]
Translations: Add Portuguese translation.
Jia Tan made white-space changes and also changed "Language: pt_BR\n"
to pt. The translator wasn't reached so I'm hoping these changes
are OK and will commit it without translator's approval.
Lasse Collin [Sun, 10 Jul 2022 18:16:40 +0000 (21:16 +0300)]
Translations: Add Serbian translation.
Quite a few white-space changes were made by Jia Tan to make
this look good. Contacting the translator didn't succeed so
I'm committing this without getting translator's approval.
Thanks to Мирослав Николић (Miroslav Nikolic) and Jia Tan.
Lasse Collin [Wed, 11 Mar 2020 20:37:54 +0000 (22:37 +0200)]
Translations: Add hu, zh_CN, and zh_TW.
I made a few white space changes to these without getting them
approved by the translation teams. (I tried to contact the hu and
zh_TW teams but didn't succeed. I didn't contact the zh_CN team.)
Lasse Collin [Wed, 11 Mar 2020 12:18:03 +0000 (14:18 +0200)]
Translations: Add fi and pt_BR, and update de, fr, it, and pl.
The German translation isn't identical to the file in
the Translation Project but the changes (white space changes
only) were approved by the translator Mario Blättermann.
Lasse Collin [Thu, 18 Aug 2022 14:49:16 +0000 (17:49 +0300)]
Build: Include the CMake files in the distribution.
This was supposed to be done in 2020 with 5.2.5 release
already but it was noticed only today. 5.2.5 and 5.2.6
even mention experiemental CMake support in the NEWS entries.
Lasse Collin [Thu, 18 Aug 2022 14:16:49 +0000 (17:16 +0300)]
liblzma: Threaded decoder: Improve LZMA_FAIL_FAST when LZMA_FINISH is used.
It will now return LZMA_DATA_ERROR (not LZMA_OK or LZMA_BUF_ERROR)
if LZMA_FINISH is used and there isn't enough input to finish
decoding the Block Header or the Block. The use of LZMA_DATA_ERROR
is simpler and the less risky than LZMA_BUF_ERROR but this might
be changed before 5.4.0.
Lasse Collin [Mon, 25 Jul 2022 16:11:05 +0000 (19:11 +0300)]
Translations: Change the copyright comment string to use with po4a.
This affects the second line in po4a/xz-man.pot. The man pages of
xzdiff, xzgrep, and xzmore are from GNU gzip and under GNU GPLv2+
while the rest of the man pages are in the public domain.
Lasse Collin [Tue, 19 Jul 2022 20:13:24 +0000 (23:13 +0300)]
xzgrep: Improve error handling, especially signals.
xzgrep wouldn't exit on SIGPIPE or SIGQUIT when it clearly
should have. It's quite possible that it's not perfect still
but at least it's much better.
If multiple exit statuses compete, now it tries to pick
the largest of value.
Some comments were added.
The exit status handling of signals is still broken if the shell
uses values larger than 255 in $? to indicate that a process
died due to a signal ***and*** their "exit" command doesn't take
this into account. This seems to work well with the ksh and yash
versions I tried. However, there is a report in gzip/zgrep that
OpenSolaris 5.11 (not 5.10) has a problem with "exit" truncating
the argument to 8 bits:
Such a bug would break xzgrep but I didn't add a workaround
at least for now. 5.11 is old and I don't know if the problem
exists in modern descendants, or if the problem exists in other
ksh implementations in use.
Lasse Collin [Mon, 18 Jul 2022 21:10:55 +0000 (00:10 +0300)]
xzgrep: Make the fix for ZDI-CAN-16587 more robust.
I don't know if this can make a difference in the real world
but it looked kind of suspicious (what happens with sed
implementations that cannot process very long lines?).
At least this commit shouldn't make it worse.
Lasse Collin [Mon, 18 Jul 2022 18:52:31 +0000 (21:52 +0300)]
xzgrep: Use grep -H --label when available (GNU, *BSDs).
It avoids the use of sed for prefixing filenames to output lines.
Using sed for that is slower and prone to security bugs so now
the sed method is only used as a fallback.
This also fixes an actual bug: When grepping a binary file,
GNU grep nowadays prints its diagnostics to stderr instead of
stdout and thus the sed-method for prefixing the filename doesn't
work. So with this commit grepping binary files gives reasonable
output with GNU grep now.
This was inspired by zgrep but the implementation is different.
Lasse Collin [Sun, 17 Jul 2022 18:36:25 +0000 (21:36 +0300)]
xzgrep: Add more LC_ALL=C to avoid bugs with multibyte characters.
Also replace one use of expr with printf.
The rationale for LC_ALL=C was already mentioned in 69d1b3fc29677af8ade8dc15dba83f0589cb63d6 that fixed a security
issue. However, unrelated uses weren't changed in that commit yet.
POSIX says that with sed and such tools one should use LC_ALL=C
to ensure predictable behavior when strings contain byte sequences
that aren't valid multibyte characters in the current locale. See
under "Application usage" in here:
Lasse Collin [Sun, 17 Jul 2022 17:55:16 +0000 (20:55 +0300)]
xzgrep: Fix parsing of certain options.
Fix handling of "xzgrep -25 foo" (in GNU grep "grep -25 foo" is
an alias for "grep -C25 foo"). xzgrep would treat "foo" as filename
instead of as a pattern. This bug was fixed in zgrep in gzip in 2012.
Add -E, -F, -G, and -P to the "no argument required" list.
Add -X to "argument required" list. It is an
intentionally-undocumented GNU grep option so this isn't
an important option for xzgrep but it seems that other grep
implementations (well, those that I checked) don't support -X
so I hope this change is an improvement still.
grep -d (grep --directories=ACTION) requires an argument. In
contrast to zgrep, I kept -d in the "no argument required" list
because it's not supported in xzgrep (or zgrep). This way
"xzgrep -d" gives an error about option being unsupported instead
of telling that it requires an argument. Both zgrep and xzgrep
tell that it's unsupported if an argument is specified.
Lasse Collin [Wed, 13 Jul 2022 19:24:07 +0000 (22:24 +0300)]
liblzma: Add optional autodetection of LZMA end marker.
Turns out that this is needed for .lzma files as the spec in
LZMA SDK says that end marker may be present even if the size
is stored in the header. Such files are rare but exist in the
real world. The code in liblzma is so old that the spec didn't
exist in LZMA SDK back then and I had understood that such
files weren't possible (the lzma tool in LZMA SDK didn't
create such files).
This modifies the internal API so that LZMA decoder can be told
if EOPM is allowed even when the uncompressed size is known.
It's allowed with .lzma and not with other uses.
Jia Tan [Fri, 1 Jul 2022 13:19:26 +0000 (21:19 +0800)]
Created script to generate code coverage reports.
The script uses lcov and genhtml after running the tests
to show the code coverage statistics. The script will create
a coverage directory where it is run. It can be run both in
and out of the source directory.
Lasse Collin [Tue, 14 Jun 2022 19:21:15 +0000 (22:21 +0300)]
Tests: test_vli: Remove an invalid test-assertion.
lzma_vli is unsigned so trying a signed value results in
a compiler warning from -Wsign-conversion. (lzma_vli)-1
equals to LZMA_VLI_UNKNOWN anyway which is the next assertion.
Jia Tan [Fri, 10 Jun 2022 13:35:18 +0000 (21:35 +0800)]
Tests: Created tests for hardware functions.
Created tests for all API functions exported in
src/liblzma/api/lzma/hardware.h. The tests are fairly trivial
but are helpful because they will inform users if their machines
cannot support these functions. They also improve the code
coverage metrics.
Lasse Collin [Mon, 23 May 2022 18:31:36 +0000 (21:31 +0300)]
Build: Enable Automake's parallel test harness.
It has been the default for quite some time already and
the old serial harness isn't discouraged. The downside is
that with parallel tests one cannot print progress info or
other diagnostics to the terminal; all output from the tests
will be in the log files only. But now that the compression
tests are separated the parallel tests will speed things up.
Lasse Collin [Mon, 23 May 2022 18:17:47 +0000 (21:17 +0300)]
Tests: Split test_compress.sh into separate test unit for each file.
test_compress.sh now takes one command line argument:
a filename to be tested. If it begins with "compress_generated_"
the file will be created with create_compress_files.
This will allow parallel execution of the slow tests.
Lasse Collin [Mon, 23 May 2022 17:32:49 +0000 (20:32 +0300)]
Test: Make create_compress_files.c a little more flexible.
If a command line argument is given, then only the test file
of that type is created. It's quite dumb in sense that unknown
names don't give an error but it's good enough here.
Also use EXIT_FAILURE instead of 1 as exit status for errors.
Lasse Collin [Thu, 14 Apr 2022 11:50:17 +0000 (14:50 +0300)]
xz: Change the cap of the default -T0 memlimit for 32-bit xz.
The SIZE_MAX / 3 was 1365 MiB. 1400 MiB gives little more room
and it looks like a round (artificial) number in --info-memory
once --info-memory is made to display it.
Also, using #if avoids useless code on 64-bit builds.
Lasse Collin [Thu, 14 Apr 2022 11:20:46 +0000 (14:20 +0300)]
xz: Add a default soft memory usage limit for --threads=0.
This is a soft limit in sense that it only affects the number of
threads. It never makes xz fail and it never makes xz change
settings that would affect the compressed output.
The idea is to make -T0 have more reasonable behavior when
the system has very many cores or when a memory-hungry
compression options are used. This also helps with 32-bit xz,
preventing it from running out of address space.
The downside of this commit is that now the number of threads
might become too low compared to what the user expected. I
hope this to be an acceptable compromise as the old behavior
has been a source of well-argued complaints for a long time.
Lasse Collin [Thu, 14 Apr 2022 09:59:09 +0000 (12:59 +0300)]
xz: Make -T0 use multithreaded mode on single-core systems.
The main problem withi the old behavior is that the compressed
output is different on single-core systems vs. multicore systems.
This commit fixes it by making -T0 one thread in multithreaded mode
on single-core systems.
The downside of this is that it uses more memory. However, if
--memlimit-compress is used, xz can (thanks to the previous commit)
drop to the single-threaded mode still.