Commit Graph

1301 Commits

Author SHA1 Message Date
Lasse Collin 882fcfdcd8 Update THANKS (sync with the master branch). 2020-02-06 17:31:55 +02:00
Lasse Collin 134bb77658 Update tests/.gitignore. 2020-02-06 00:01:03 +02:00
Lasse Collin 6912472faf Update m4/.gitignore. 2020-02-06 00:01:03 +02:00
Lasse Collin 68c60735bb Update THANKS. 2020-02-06 00:01:03 +02:00
Lasse Collin e1beaa74bc xz: Comment out annoying sandboxing messages. 2020-02-05 22:00:28 +02:00
Lasse Collin 8238192652 Build: Workaround a POSIX shell detection problem on Solaris.
I don't know if the problem is in gnulib's gl_POSIX_SHELL macro
or if xzgrep does something that isn't in POSIX. The workaround
adds a special case for Solaris: if /usr/xpg4/bin/sh exists and
gl_cv_posix_shell wasn't overriden on the configure command line,
use that shell for xzgrep and other scripts. That shell is known
to work and exists on most Solaris systems.
2020-02-05 22:00:28 +02:00
Lasse Collin 93a1f61e89 Build: Update m4/ax_pthread.m4 from Autoconf Archive. 2020-02-05 22:00:28 +02:00
Lasse Collin d0daa21792 xz: Limit --memlimit-compress to at most 4020 MiB for 32-bit xz.
See the code comment for reasoning. It's far from perfect but
hopefully good enough for certain cases while hopefully doing
nothing bad in other situations.

At presets -5 ... -9, 4020 MiB vs. 4096 MiB makes no difference
on how xz scales down the number of threads.

The limit has to be a few MiB below 4096 MiB because otherwise
things like "xz --lzma2=dict=500MiB" won't scale down the dict
size enough and xz cannot allocate enough memory. With
"ulimit -v $((4096 * 1024))" on x86-64, the limit in xz had
to be no more than 4085 MiB. Some safety margin is good though.

This is hack but it should be useful when running 32-bit xz on
a 64-bit kernel that gives full 4 GiB address space to xz.
Hopefully this is enough to solve this:

https://bugzilla.redhat.com/show_bug.cgi?id=1196786

FreeBSD has a patch that limits the result in tuklib_physmem()
to SIZE_MAX on 32-bit systems. While I think it's not the way
to do it, the results on --memlimit-compress have been good. This
commit should achieve practically identical results for compression
while leaving decompression and tuklib_physmem() and thus
lzma_physmem() unaffected.
2020-02-05 22:00:28 +02:00
Lasse Collin 4433c2dc57 xz: Set the --flush-timeout deadline when the first input byte arrives.
xz --flush-timeout=2000, old version:

  1. xz is started. The next flush will happen after two seconds.
  2. No input for one second.
  3. A burst of a few kilobytes of input.
  4. No input for one second.
  5. Two seconds have passed and flushing starts.

The first second counted towards the flush-timeout even though
there was no pending data. This can cause flushing to occur more
often than needed.

xz --flush-timeout=2000, after this commit:

  1. xz is started.
  2. No input for one second.
  3. A burst of a few kilobytes of input. The next flush will
     happen after two seconds counted from the time when the
     first bytes of the burst were read.
  4. No input for one second.
  5. No input for another second.
  6. Two seconds have passed and flushing starts.
2020-02-05 22:00:28 +02:00
Lasse Collin acc0ef3ac8 xz: Move flush_needed from mytime.h to file_pair struct in file_io.h. 2020-02-05 22:00:28 +02:00
Lasse Collin 4afe69d30b xz: coder.c: Make writing output a separate function.
The same code sequence repeats so it's nicer as a separate function.
Note that in one case there was no test for opt_mode != MODE_TEST,
but that was only because that condition would always be true, so
this commit doesn't change the behavior there.
2020-02-05 22:00:28 +02:00
Lasse Collin ec26f3ace5 xz: Fix semi-busy-waiting in xz --flush-timeout.
When input blocked, xz --flush-timeout=1 would wake up every
millisecond and initiate flushing which would have nothing to
flush and thus would just waste CPU time. The fix disables the
timeout when no input has been seen since the previous flush.
2020-02-05 22:00:28 +02:00
Lasse Collin 3891570324 xz: Refactor io_read() a bit. 2020-02-05 22:00:28 +02:00
Lasse Collin f6d2424534 xz: Update a comment in file_io.h. 2020-02-05 22:00:28 +02:00
Lasse Collin 15b55d5c63 xz: Move the setting of flush_needed in file_io.c to a nicer location. 2020-02-05 22:00:28 +02:00
Lasse Collin 609c706785 xz: Enable Capsicum sandboxing by default if available.
It has been enabled in FreeBSD for a while and reported to work fine.

Thanks to Xin Li.
2020-02-05 20:21:56 +02:00
Lasse Collin 00517d125c Rename unaligned_read32ne to read32ne, and similarly for the others. 2019-12-31 22:41:45 +02:00
Lasse Collin 52d89d8443 Rename read32ne to aligned_read32ne, and similarly for the others.
Using the aligned methods requires more care to ensure that
the address really is aligned, so it's nicer if the aligned
methods are prefixed. The next commit will remove the unaligned_
prefix from the unaligned methods which in liblzma are used in
more places than the aligned ones.
2019-12-31 22:34:34 +02:00
Lasse Collin 850620468b Revise tuklib_integer.h and .m4.
Add a configure option --enable-unsafe-type-punning to get the
old non-conforming memory access methods. It can be useful with
old compilers or in some other less typical situations but
shouldn't normally be used.

Omit the packed struct trick for unaligned access. While it's
best in some cases, this is simpler. If the memcpy trick doesn't
work, one can request unsafe type punning from configure.

Because CRC32/CRC64 code needs fast aligned reads, if no very
safe way to do it is found, type punning is used as a fallback.
This sucks but since it currently works in practice, it seems to
be the least bad option. It's never needed with GCC >= 4.7 or
Clang >= 3.6 since these support __builtin_assume_aligned and
thus fast aligned access can be done with the memcpy trick.

Other things:
  - Support GCC/Clang __builtin_bswapXX
  - Cleaner bswap fallback macros
  - Minor cleanups
2019-12-31 22:34:10 +02:00
Lasse Collin a45badf034 Tests: Hopefully fix test_check.c to work on EBCDIC systems.
Thanks to Daniel Richard G.
2019-12-31 22:31:34 +02:00
Lasse Collin c9a8071e66 Scripts: Put /usr/xpg4/bin to the beginning of PATH on Solaris.
This adds a configure option --enable-path-for-scripts=PREFIX
which defaults to empty except on Solaris it is /usr/xpg4/bin
to make POSIX grep and others available. The Solaris case had
been documented in INSTALL with a manual fix but it's better
to do this automatically since it is needed on most Solaris
systems anyway.

Thanks to Daniel Richard G.
2019-12-31 22:31:30 +02:00
Lasse Collin aba140e2df Fix comment typos in tuklib_mbstr* files. 2019-12-31 22:27:11 +02:00
Lasse Collin 710f5bd769 Add missing include to tuklib_mbstr_width.c.
It didn't matter in XZ Utils because sysdefs.h
includes string.h anyway.
2019-12-31 22:27:11 +02:00
Lasse Collin 0e491aa8cd liblzma: Fix a buggy comment. 2019-12-31 22:26:38 +02:00
Lasse Collin bfc245569f configure.ac: Fix a typo in a comment. 2019-12-31 22:26:38 +02:00
Lasse Collin f18eee9d15 Tests: Silence warnings from clang -Wassign-enum.
Also changed 999 to 99 so it fits even if lzma_check happened
to be 8 bits wide.
2019-12-31 22:26:38 +02:00
Lasse Collin 25f7455472 liblzma: Add a comment. 2019-12-31 22:26:38 +02:00
Lasse Collin 44eb961f2a liblzma: Silence clang -Wmissing-variable-declarations. 2019-12-31 22:26:38 +02:00
Lasse Collin 267afcd995 xz: Silence a warning from clang -Wsign-conversion in main.c. 2019-12-31 22:25:42 +02:00
Lasse Collin 0e3c4002f8 liblzma: Remove incorrect uses of lzma_attribute((__unused__)).
Caught by clang -Wused-but-marked-unused.
2019-12-31 22:25:02 +02:00
Lasse Collin cb708e8fa3 Tests: Silence a warning from -Wsign-conversion. 2019-12-31 22:25:02 +02:00
Lasse Collin c8cace3d6e xz: Fix an integer overflow with 32-bit off_t.
Or any off_t which isn't very big (like signed 64 bit integer
that most system have). A small off_t could overflow if the
file being decompressed had long enough run of zero bytes,
which would result in corrupt output.
2019-12-31 22:25:02 +02:00
Lasse Collin 65a42741e2 Tests: Remove a duplicate branch from tests/tests.h.
The duplication was introduced about eleven years ago and
should have been cleaned up back then already.

This was caught by -Wduplicated-branches.
2019-12-31 22:20:10 +02:00
Lasse Collin 5c4fb60e8d tuklib_mbstr_width: Fix a warning from -Wsign-conversion. 2019-12-31 22:19:18 +02:00
Lasse Collin 37df03ce52 xz: Fix some of the warnings from -Wsign-conversion. 2019-12-31 22:19:18 +02:00
Lasse Collin 7c65ae0f5f tuklib_cpucores: Silence warnings from -Wsign-conversion. 2019-12-31 22:19:18 +02:00
Lasse Collin a502dd1d00 xzdec: Fix warnings from -Wsign-conversion. 2019-12-31 22:19:18 +02:00
Lasse Collin a45d1a5374 liblzma: Fix warnings from -Wsign-conversion.
Also, more parentheses were added to the literal_subcoder
macro in lzma_comon.h (better style but no functional change
in the current usage).
2019-12-31 22:19:18 +02:00
Lasse Collin 4ff87ddf80 tuklib_integer: Silence warnings from -Wsign-conversion. 2019-12-31 22:19:18 +02:00
Lasse Collin ed1a9d3398 tuklib_integer: Fix usage of conv macros.
Use a temporary variable instead of e.g.
conv32le(unaligned_read32ne(buf)) because the macro can
evaluate its argument multiple times.
2019-12-31 22:19:18 +02:00
Lasse Collin 612c88dfc0 Update THANKS. 2019-12-31 22:19:18 +02:00
Lasse Collin 85da31d8b8 liblzma: Fix comments.
Thanks to Bruce Stark.
2019-12-31 22:19:18 +02:00
Lasse Collin 6a73a78895 liblzma: Fix one more unaligned read to use unaligned_read16ne(). 2019-12-31 22:19:18 +02:00
Lasse Collin ce59b34ec9 Update THANKS. 2019-12-31 22:19:18 +02:00
Lasse Collin 94aa3fb568 liblzma: memcmplen: Use ctz32() from tuklib_integer.h.
The same compiler-specific #ifdefs are already in tuklib_integer.h
2019-12-31 22:19:18 +02:00
Lasse Collin 412791486d tuklib_integer: Cleanup MSVC-specific code. 2019-12-31 22:19:18 +02:00
Lasse Collin efbf6e5f09 liblzma: Use unaligned_readXXne functions instead of type punning.
Now gcc -fsanitize=undefined should be clean.

Thanks to Jeffrey Walton.
2019-12-31 22:19:18 +02:00
Lasse Collin 29afef0348 tuklib_integer: Improve unaligned memory access.
Now memcpy() or GNU C packed structs for unaligned access instead
of type punning. See the comment in this commit for details.

Avoiding type punning with unaligned access is needed to
silence gcc -fsanitize=undefined.

New functions: unaliged_readXXne and unaligned_writeXXne where
XX is 16, 32, or 64.
2019-12-31 22:19:12 +02:00
Lasse Collin 596ed3de44 liblzma: Avoid memcpy(NULL, foo, 0) because it is undefined behavior.
I should have always known this but I didn't. Here is an example
as a reminder to myself:

    int mycopy(void *dest, void *src, size_t n)
    {
        memcpy(dest, src, n);
        return dest == NULL;
    }

In the example, a compiler may assume that dest != NULL because
passing NULL to memcpy() would be undefined behavior. Testing
with GCC 8.2.1, mycopy(NULL, NULL, 0) returns 1 with -O0 and -O1.
With -O2 the return value is 0 because the compiler infers that
dest cannot be NULL because it was already used with memcpy()
and thus the test for NULL gets optimized out.

In liblzma, if a null-pointer was passed to memcpy(), there were
no checks for NULL *after* the memcpy() call, so I cautiously
suspect that it shouldn't have caused bad behavior in practice,
but it's hard to be sure, and the problematic cases had to be
fixed anyway.

Thanks to Jeffrey Walton.
2019-07-13 17:56:28 +03:00
Lasse Collin b4b83555c5 Update THANKS. 2019-07-13 17:54:52 +03:00