Commit Graph

43 Commits

Author SHA1 Message Date
Lasse Collin 120da10ae1 liblzma: Disable branchless C version in range decoder.
Thanks to Sebastian Andrzej Siewior and Sam James for
benchmarking on various systems.
2024-02-22 14:41:29 +02:00
Lasse Collin 746c471643 liblzma: Remove commented-out code. 2024-02-19 11:58:33 +02:00
Lasse Collin 843ddc5f61 liblzma: Wrap a line exceeding 80 chars. 2024-02-17 15:50:21 +02:00
Sebastian Andrzej Siewior e9053c9072 liblzma/rangecoder: Exclude x32 from the x86-64 optimisation.
The x32 port has a x86-64 ABI in term of all registers but uses only
32bit pointer like x86-32. The assembly optimisation fails to compile on
x32. Given the state of x32 I suggest to exclude it from the
optimisation rather than trying to fix it.

Signed-off-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
2024-02-17 15:50:21 +02:00
Jia Tan fb5f6aaf18 Fix typos discovered by codespell. 2024-02-16 22:54:59 +08:00
Lasse Collin a4557bad96 liblzma: Silence warnings in --enable-small build. 2024-02-14 19:21:45 +02:00
Lasse Collin 160b686264 liblzma: Silence a warning. 2024-02-14 19:05:58 +02:00
Lasse Collin 0668907ff7 liblzma: Add comments. 2024-02-14 18:31:16 +02:00
Lasse Collin de55485cb2 liblzma: Choose the range decoder variants using a bitmask macro. 2024-02-14 18:31:16 +02:00
Lasse Collin 3182a330c1 liblzma: Range decoder: Add x86-64 inline assembly.
It's compatible with GCC and Clang.
2024-02-14 18:31:16 +02:00
Lasse Collin cba2edc991 liblzma: Range decoder: Add branchless C code.
It's used only for basic bittrees and fixed-size reverse bittree
because those showed a clear benefit on x86-64 with GCC and Clang.
The other methods were more mixed and thus are commented out but
they should be tested on other archs.
2024-02-14 18:31:16 +02:00
Lasse Collin 5e04706b91 liblzma: LZMA decoder: Optimize loop comparison.
But now it needs one more local variable.
2024-02-14 18:31:16 +02:00
Lasse Collin e0c0ee475c liblzma: LZMA decoder improvements.
This adds macros for bittree decoding which prepares the code
for alternative C versions and inline assembly.
2024-02-14 18:31:16 +02:00
Jia Tan e446ab7a18 liblzma: Creates separate "safe" range decoder mode.
The new "safe" range decoder mode is the same as old range decoder, but
now the default behavior of the range decoder will not check if there is
enough input or output to complete the operation. When the buffers are
close to fully consumed, the "safe" operations must be used instead. This
will improve speed because it will reduce the number of branches needed
for most of the range decoder operations.
2024-02-14 18:31:16 +02:00
Lasse Collin b941549573 liblzma: Include the SPDX license identifier 0BSD to generated files.
Perhaps the generated files aren't even copyrightable but
using the same license for them as for the rest of the liblzma
keeps things more consistent for tools that look for license info.
2024-02-14 18:31:16 +02:00
Lasse Collin 8e4ec79483 liblzma: Fix compilation of price_tablegen.c.
It is built and run only manually so this didn't matter
unless one wanted to regenerate the price_table.c.
2024-02-14 18:31:16 +02:00
Lasse Collin 22af94128b Add SPDX license identifier into 0BSD source code files. 2024-02-14 18:31:16 +02:00
Lasse Collin 689e0228ba Change most public domain parts to 0BSD.
Translations and doc/xz-file-format.txt and doc/lzma-file-format.txt
were not touched.

COPYING.0BSD was added.
2024-02-14 18:31:12 +02:00
Lasse Collin 41113fe30a liblzma: Use lzma_attr_visibility_hidden on private extern declarations.
These variables are internal to liblzma and not exposed in the API.
2023-10-30 18:06:25 +02:00
Jia Tan f6667702bf liblzma: Change quoting style from `...' to '...'.
This was done for both internal and API headers.
2023-09-24 22:09:47 +08:00
Lasse Collin 2523c30705 liblzma: Fix uint64_t vs. size_t confusion.
This broke 32-bit builds due to a pointer type mismatch.

This bug was introduced with the output-size-limited encoding
in 625f4c7c99.

Thanks to huangqinjin for the bug report.
2022-02-06 23:19:32 +02:00
Lasse Collin 421b0aa352 liblzma: Fix missing normalization in rc_encode_dummy().
Without this fix it could attempt to create too much output.
2021-01-14 20:57:11 +02:00
Lasse Collin 625f4c7c99 liblzma: Add rough support for output-size-limited encoding in LZMA1.
With this it is possible to encode LZMA1 data without EOPM so that
the encoder will encode as much input as it can without exceeding
the specified output size limit. The resulting LZMA1 stream will
be a normal LZMA1 stream without EOPM. The actual uncompressed size
will be available to the caller via the uncomp_size pointer.

One missing thing is that the LZMA layer doesn't inform the LZ layer
when the encoding is finished and thus the LZ may read more input
when it won't be used. However, this doesn't matter if encoding is
done with a single call (which is the planned use case for now).
For proper multi-call encoding this should be improved.

This commit only adds the functionality for internal use.
Nothing uses it yet.
2021-01-14 18:58:13 +02:00
Lasse Collin 4572d53e16 liblzma: Fix a comment and RC_SYMBOLS_MAX.
The comment didn't match the value of RC_SYMBOLS_MAX and the value
itself was slightly larger than actually needed. The only harm
about this was that memory usage was a few bytes larger.
2020-03-02 13:54:33 +02:00
Lasse Collin fb6d4f83cb liblzma: Remove incorrect #ifdef from range_common.h.
In most cases it was harmless but it could affect some
custom build systems.

Thanks to Pippijn van Steenhoven.
2018-02-06 18:02:48 +02:00
Lasse Collin 1403707fc6 liblzma: Check that the first byte of range encoded data is 0x00.
It is just to be more pedantic and thus perhaps catch broken
files slightly earlier.
2012-06-28 10:47:49 +03:00
Lasse Collin f42ee98166 Build system fixes
Don't use libtool convenience libraries to avoid recently
discovered long-standing subtle but somewhat severe bugs
in libtool (at least 1.5.22 and 2.2.6 are affected). It
was found when porting XZ Utils to Windows
<http://lists.gnu.org/archive/html/libtool/2009-06/msg00070.html>
but the problem is significant also e.g. on GNU/Linux.

Unless --disable-shared is passed to configure, static
library built from a set of convenience libraries will
contain PIC objects. That is, while libtool builds non-PIC
objects too, only PIC objects will be used from the
convenience libraries. On 32-bit x86 (tested on mobile XP2400+),
using PIC instead of non-PIC makes the decompressor 10 % slower
with the default CFLAGS.

So while xz was linked against static liblzma by default,
it got the slower PIC objects unless --disable-shared was
used. I tend develop and benchmark with --disable-shared
due to faster build time, so I hadn't noticed the problem
in benchmarks earlier.

This commit also adds support for building Windows resources
into liblzma and executables.
2009-06-30 17:09:57 +03:00
Lasse Collin 1c9360b7d1 Fix @variables@ to $(variables) in Makefile.am files.
Fix the ordering of libgnu.a and LTLIBINTL on the linker
command line and added missing LTLIBINTL to tests/Makefile.am.
2009-06-26 14:47:31 +03:00
Lasse Collin 02ddf09bc3 Put the interesting parts of XZ Utils into the public domain.
Some minor documentation cleanups were made at the same time.
2009-04-13 11:27:40 +03:00
Lasse Collin 7ed9d943b3 Remove lzma_init() and other init functions from liblzma API.
Half of developers were already forgetting to use these
functions, which could have caused total breakage in some future
liblzma version or even now if --enable-small was used. Now
liblzma uses pthread_once() to do the initializations unless
it has been built with --disable-threads which make these
initializations thread-unsafe.

When --enable-small isn't used, liblzma currently gets needlessly
linked against libpthread (on systems that have it). While it is
stupid for now, liblzma will need threads in future anyway, so
this stupidity will be temporary only.

When --enable-small is used, different code CRC32 and CRC64 is
now used than without --enable-small. This made the resulting
binary slightly smaller, but the main reason was to clean it up
and to handle the lack of lzma_init_check().

The pkg-config file lzma.pc was renamed to liblzma.pc. I'm not
sure if it works correctly and portably for static linking
(Libs.private includes -pthread or other operating system
specific flags). Hopefully someone complains if it is bad.

lzma_rc_prices[] is now included as a precomputed array even
with --enable-small. It's just 128 bytes now that it uses uint8_t
instead of uint32_t. Smaller array seemed to be at least as fast
as the more bloated uint32_t array on x86; hopefully it's not bad
on other architectures.
2008-12-31 00:30:49 +02:00
Lasse Collin 1dcecfb09b Some API changes, bug fixes, cleanups etc. 2008-09-27 19:09:21 +03:00
Lasse Collin 3b34851de1 Sort of garbage collection commit. :-| Many things are still
broken. API has changed a lot and it will still change a
little more here and there. The command line tool doesn't
have all the required changes to reflect the API changes, so
it's easy to get "internal error" or trigger assertions.
2008-08-28 22:53:15 +03:00
Lasse Collin 369f72fd65 Fix a buffer overflow in the LZMA encoder. It was due to my
misunderstanding of the code. There's no tiny fix for this
problem, so I also cleaned up the code in general.

This reduces the speed of the encoder 2-5 % in the fastest
compression mode ("lzma -1"). High compression modes should
have no noticeable performance difference.

This commit breaks things (especially LZMA_SYNC_FLUSH) but I
will fix them once the new format and LZMA2 has been roughly
implemented. Plain LZMA won't support LZMA_SYNC_FLUSH at all
and won't be supported in the new .lzma format. This may
change still but this is what it looks like now.

Support for known uncompressed size (that is, LZMA or LZMA2
without EOPM) is likely to go away. This means there will
be API changes.
2008-06-01 12:48:17 +03:00
Lasse Collin eb348a60b6 Switch to uint16_t as the type of range coder probabilities. 2008-04-24 19:22:53 +03:00
Lasse Collin 7ab493924e Use 64-bit integer as range encoder's cache size. This fixes a
theoretical data corruption, which should be very hard to trigger
even intentionally.
2008-04-24 17:30:51 +03:00
Lasse Collin 641998c3e1 Replaced the range decoder optimization that used arithmetic
right shift with as fast version that doesn't need
arithmetic right shift. Removed the related check from
configure.ac.
2008-03-24 16:38:40 +02:00
Lasse Collin ad999efd27 Take advantage of arithmetic right shift in range decoder. 2008-03-22 14:39:34 +02:00
Lasse Collin 54ec204f58 Omit invalid space from printf() format string
in price_table_gen.c.
2008-01-15 12:20:41 +02:00
Lasse Collin 9f9b198301 Convert bittree_get_price() and bittree_reverse_get_price()
from macros to inline functions.
2008-01-15 08:36:25 +02:00
Lasse Collin d13d693155 Added precomputed range coder probability price table. 2008-01-15 07:40:21 +02:00
Lasse Collin e22b37968d Major changes to LZ encoder, LZMA encoder, and range encoder.
These changes implement support for LZMA_SYNC_FLUSH in LZMA
encoder, and move the temporary buffer needed by range encoder
from lzma_range_encoder structure to lzma_lz_encoder.
2008-01-14 13:39:54 +02:00
Lasse Collin bbfd1f6ab0 Moved range decoder initialization (reading the first
five input bytes) from LZMA decoder to range decoder
header. Did the same for decoding of direct bits.
2008-01-04 20:45:05 +02:00
Lasse Collin 5d018dc035 Imported to git. 2007-12-09 00:42:33 +02:00