When opt_block_size is not used, the Block size for mt encoder is
derived from the minimum of the largest Block specified by
--block-list and the recommended Block size on all filter chains
calculated by lzma_mt_block_size(). This avoids using unnecessary
memory and ensures that all Blocks are large enough for the most memory
needy filter chain.
Previously, only the default filter chain could have its memory usage
adjusted. The filter chains specified with --filtersX were not checked
for memory usage. Now, all used filter chains will be adjusted if
necessary.
The block splitting logic and split_block() function are not needed if
encoders are disabled. This will help slightly reduce the binary size
when built without encoders and allow split_block() to use functions
that require encoders being enabled.
This will only free filter chains created with --filters1-9 since the
default filter chain may be set from a static function variable. The
complexity to free the default filter chain is not worth the burden on
code maintenance.
The new command line options are meant to be combined with --block-list.
They work as an optional extension to --block-list to specify a custom
filter chain for each block listed. The new options allow the creation
of up to 9 reusable filter chains. For instance:
xz --block-list=1:10MiB,3:5MiB,,2:5MiB,1:0 --filters1=delta--lzma2 \
--filters2=x86--lzma2 --filters3=arm64--lzma2
Will create the following blocks:
1. A block of size 10 MiB with filter chain delta, lzma2.
2. A block of size 5 MiB with filter chain arm64, lzma2.
3. A block of size 5 MiB with filter chain arm64, lzma2.
4. A block of size 5 MiB with filter chain x86, lzma2.
5. A block containing the rest of the file contents with filter chain
delta, lzma2.
This is a little cleaner than the previous implementation of
forget_filter_chain(). It is also more consistent since
lzma_str_to_filters() will always terminate the filter chain so there
is no need to terminate it later in coder_set_compression_settings().
The --filters option uses the new lzma_str_to_filters() function
to convert a string into a full filter chain. Using this option
will reset all previous filters set by --preset, --[filter], or
--filters.
Commit 78704f36e7 added an empty
initializer {} to prevent a warning. The empty initializer is a GNU
extension and results in a build failure on MSVC. The -wpedantic flag
warns about empty initializers.
This change only impacts the compiler warning since it was impossible
for the wait_abs struct in stream_encode_mt() to be used before it was
initialized since mythread_condtime_set() will always be called before
mythread_cond_timedwait().
Since the mythread.h code is different between the POSIX and
Windows versions, this warning was only present on Windows builds.
Thanks to Arthur S for reporting the warning and providing an initial
patch.
In lzma_memcmplen(), the <intrin.h> header file is only included if
_MSC_VER and _M_X64 are both defined but _BitScanForward64() was
previously used if _M_X64 was defined. GCC for MSYS2 defines _M_X64 but
not _MSC_VER so _BitScanForward64() was used without including
<intrin.h>.
Now, lzma_memcmplen() will use __builtin_ctzll() for MSYS2 GCC builds as
expected.
The ifunc method avoids indirection via the function pointer
crc64_func. This works on GNU/Linux and probably on FreeBSD too.
The previous __attribute((__constructor__)) method is kept for
compatibility with ELF platforms which do support ifunc.
The ifunc method has some limitations, for example, building
liblzma with -fsanitize=address will result in segfaults.
The configure option --disable-ifunc must be used for such builds.
Thanks to Hans Jansen for the original patch.
Closes: https://github.com/tukaani-project/xz/pull/53
Reword "options required" to "supported options". The previous may have
suggested that the options listed were all required anytime a filter is
used for encoding or decoding. The reword makes this more clear that
adjusting the options is optional.
The lzma_mt_block_size() was previously just an internal function for
the multithreaded .xz encoder. It is used to provide a recommended Block
size for a given filter chain.
This function is helpful to determine the maximum Block size for the
multithreaded .xz encoder when one wants to change the filters between
blocks. Then, this determined Block size can be provided to
lzma_stream_encoder_mt() in the lzma_mt options parameter when
intializing the coder. This requires one to know all the filter chains
they are using before starting to encode (or at least the filter chain
that will need the largest Block size), but that isn't a bad limitation.
Legacy Windows did not need to #include <intrin.h> to use the MSVC
intrinsics. Newer versions likely just issue a warning, but the MSVC
documentation says to include the header file for the intrinsics we use.
GCC and Clang can "pretend" to be MSVC on Windows, so extra checks are
needed in tuklib_integer.h to only include <intrin.h> when it will is
actually needed.
Clang has support for __builtin_clz(), but previously Clang would
fallback to either the MSVC intrinsic or the regular C code. This was
discovered due to a bug where a new version of Clang required the
<intrin.h> header file in order to use the MSVC intrinsics.
Thanks to Anton Kochkov for notifying us about the bug.
The \mainpage command is used in the first block of comments in lzma.h.
This changes the previously nearly empty index.html to use the first
comment block in lzma.h for its contents.
lzma.h is no longer documented separately, but this is for the better
since lzma.h only defined a few macros that users do not need to use.
The individual API header files all have a disclaimer that they should
not be #included directly, so there should be no confusion on the fact
that lzma.h should be the only header used by applications.
Additionally, the note "See ../lzma.h for information about liblzma as
a whole." was removed since lzma.h is now the main page of the
generated HTML and does not have its own page anymore. So it would be
confusing in the HTML version and was only a "nice to have" when
browsing the source files.
Also remove unneeded "sandbox_allowed = false;" as this code
will never be run more than once (making it work with multiple
input files isn't trivial).
The warning causes the exit status to be 2, so this will cause problems
for many scripted use cases for xz. The sandbox usage is already very
limited already, so silently disabling this allows it to be more usable.
If a system has the Capsicum header files but does not actually
implement the system calls, then this would render xz unusable. Instead,
we can check if errno == ENOSYS and not issue a fatal error.
lzma_lzma_preset() does not guarentee that the lzma_options_lzma are
usable in an encoder even if it returns false (success). If liblzma
is built with default configurations, then the options will always be
usable. However if the match finders hc3, hc4, or bt4 are disabled, then
the options may not be usable depending on the preset level requested.
The documentation was updated to reflect this complexity, since this
behavior was unclear before.
All functions now explicitly specify parameter and return values.
The notes and code annotations were moved before the parameter and
return value descriptions for consistency.
Also, the description above lzma_filter_encoder_is_supported() about
not being able to list available filters was removed since
lzma_str_list_filters() will do this.
In the C99 and C17 standards, section 6.5.6 paragraph 8 means that
adding 0 to a null pointer is undefined behavior. As of writing,
"clang -fsanitize=undefined" (Clang 15) diagnoses this. However,
I'm not aware of any compiler that would take advantage of this
when optimizing (Clang 15 included). It's good to avoid this anyway
since compilers might some day infer that pointer arithmetic implies
that the pointer is not NULL. That is, the following foo() would then
unconditionally return 0, even for foo(NULL, 0):
void bar(char *a, char *b);
int foo(char *a, size_t n)
{
bar(a, a + n);
return a == NULL;
}
In contrast to C, C++ explicitly allows null pointer + 0. So if
the above is compiled as C++ then there is no undefined behavior
in the foo(NULL, 0) call.
To me it seems that changing the C standard would be the sane
thing to do (just add one sentence) as it would ensure that a huge
amount of old code won't break in the future. Based on web searches
it seems that a large number of codebases (where null pointer + 0
occurs) are being fixed instead to be future-proof in case compilers
will some day optimize based on it (like making the above foo(NULL, 0)
return 0) which in the worst case will cause security bugs.
Some projects don't plan to change it. For example, gnulib and thus
many GNU tools currently require that null pointer + 0 is defined:
https://lists.gnu.org/archive/html/bug-gnulib/2021-11/msg00000.htmlhttps://www.gnu.org/software/gnulib/manual/html_node/Other-portability-assumptions.html
In XZ Utils null pointer + 0 issue should be fixed after this
commit. This adds a few if-statements and thus branches to avoid
null pointer + 0. These check for size > 0 instead of ptr != NULL
because this way bugs where size > 0 && ptr == NULL will likely
get caught quickly. None of them are in hot spots so it shouldn't
matter for performance.
A little less readable version would be replacing
ptr + offset
with
offset != 0 ? ptr + offset : ptr
or creating a macro for it:
#define my_ptr_add(ptr, offset) \
((offset) != 0 ? ((ptr) + (offset)) : (ptr))
Checking for offset != 0 instead of ptr != NULL allows GCC >= 8.1,
Clang >= 7, and Clang-based ICX to optimize it to the very same code
as ptr + offset. That is, it won't create a branch. So for hot code
this could be a good solution to avoid null pointer + 0. Unfortunately
other compilers like ICC 2021 or MSVC 19.33 (VS2022) will create a
branch from my_ptr_add().
Thanks to Marcin Kowalczyk for reporting the problem:
https://github.com/tukaani-project/xz/issues/36
Standardizing each function to always specify parameters and return
values. Also moved the parameters and return values to the end of each
function description.
Use "member" to refer to struct members as that's the term used
by the C standard.
Use lzma_options_delta.dist and such in docs so that in Doxygen's
HTML output they will link to the doc of the struct member.
Clean up a few trailing white spaces too.
It gives C4146 here since unary minus with unsigned integer
is still unsigned (which is the intention here). Doing it
with substraction makes it clearer and avoids the warning.
Thanks to Nathan Moinvaziri for reporting this.
Standardizing each function to always specify parameters and return
values. Also moved the parameters and return values to the end of each
function description.
A few small things were reworded and long sentences broken up.
All functions now explicitly specify parameter and return values.
Also moved the note about SHA-256 functions not being exported to the
top of the file.
The bug is only a problem in applications that do not properly terminate
the filters[] array with LZMA_VLI_UNKNOWN or have more than
LZMA_FILTERS_MAX filters. This bug does not affect xz.
Added a few sentences to the description for lzma_block_encoder() and
lzma_block_decoder() to highlight that the Block Header must be coded
before calling these functions.
Standardizing each function to always specify params and return values.
Output pointer parameters are also marked with doxygen style [out] to
make it clear. Any note sections were also moved above the parameter and
return sections for consistency.
The flag description for LZMA_STR_NO_VALIDATION was previously confusing
about the treatment for filters than cannot be used with .xz format
(lzma1) without using LZMA_STR_ALL_FILTERS. Now, it is clear that
LZMA_STR_NO_VALIDATION is not a super set of LZMA_STR_ALL_FILTERS.
This way, if xz is stopped the elapsed time and estimated time
remaining won't get confused by the amount of time spent in
the stopped state.
This raises SIGSTOP. It's not clear to me if this is the correct way.
POSIX and glibc docs say that SIGTSTP shouldn't stop the process if
it is orphaned but this commit doesn't attempt to handle that.
Search for SIGTSTP in section 2.4.3:
https://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html
The previous documentation for lzma_str_to_filters() was technically
correct, but misleading. lzma_str_to_filters() returns NULL on success,
which is in practice always defined to 0. This is the same value as
LZMA_OK, but lzma_str_to_filters() does not return lzma_ret so we should
be more clear.
This reverts commit 82e3c968bf.
Macros in the reserved namespace (_foo or __foo) shouldn't be #defined
without a very good reason. Here the alternative would have been
to #define tuklib_has_warning(str) to an approriate value.
Also the tuklib_* files should stay namespace clean if possible.
__has_warning and other __has_foo macros are meant to become
compiler-agnostic so it's not good to check for __clang__ with it.
This also relied on tuklib_common.h for #defining __has_warning
which was confusing as #defining reserved macros is generally
not a good idea.
tuklib_physmem depends on GetProcAddress() for both MSVC and MinGW-w64
to retrieve a function address. The proper way to do this is to cast the
return value to the type of function pointer retrieved. Unfortunately,
this causes a cast-function-type warning, so the best solution is to
simply ignore the warning.
clang supports the __has_warning macro to determine if the version of
clang compiling the code supports a given warning. If we do not define
it for other compilers, it may cause a preprocessor error.
Calling coder_set_compression_settings() in list mode with verbose mode
on caused the filter chain and memory requirements to print. This was
unnecessary since the command results in an error and not consistent
with other formats like lzma and alone.
It doesn't warn on a 64-bit system because truncating
a ptrdiff_t (signed long) to uint32_t is diagnosed under
-Wconversion by GCC and -Wshorten-64-to-32 by Clang.
This is similar to 2ce4f36f17.
The actual initialization of the variables is done inside
mythread_sync() macro. Clang doesn't seem to see that
the initialization code inside the macro is always executed.
clang and gcc differ in how they handle -Wformat-nonliteral. gcc will
allow a non-literal format string as long as the function takes its
format arguments as a va_list.
This affects only 32-bit x86 builds. x86-64 is OK as is.
I still cannot easily test this myself. The reporter has tested
this and it passes the tests included in the CMake build and
performance is good: raw CRC64 is 2-3 times faster than the
C version of the slice-by-four method. (Note that liblzma doesn't
include a MSVC-compatible version of the 32-bit x86 assembly code
for the slice-by-four method.)
Thanks to Iouri Kharon for figuring out a fix, testing, and
benchmarking.
This reverts commit 36edc65ab4.
It was reported that it wasn't a good enough fix and MSVC
still produced (different kind of) bad code when building
for 32-bit x86 if optimizations are enabled.
Thanks to Iouri Kharon.
On some platforms src/xz/suffix.c may need <strings.h> for
strcasecmp() but suffix.c includes the header when it needs it.
Unless there is an old system that otherwise supports enough C99
to build XZ Utils but doesn't have C89/C90-compatible <string.h>,
there should be no need to include <strings.h> in sysdefs.h.
SUSv2 and POSIX.1‐2017 declare only a few functions in <strings.h>.
Of these, strcasecmp() is used on some platforms in suffix.c.
Nothing else in the project needs <strings.h> (at least if
building on a modern system).
sysdefs.h currently includes <strings.h> if HAVE_STRINGS_H is
defined and suffix.c relied on this.
Note that dos/config.h doesn't #define HAVE_STRINGS_H even though
DJGPP does have strings.h. It isn't needed with DJGPP as strcasecmp()
is also in <string.h> in DJGPP.
It quite probably was never needed, that is, any system where memory.h
was required likely couldn't compile XZ Utils for other reasons anyway.
XZ Utils 5.2.6 and later source packages were generated using
Autoconf 2.71 which no longer defines HAVE_MEMORY_H. So the code
being removed is no longer used anyway.
I haven't tested with MSVC myself and there doesn't seem to be
information about the problem online, so I'm relying on the bug report.
Thanks to Iouri Kharon for the bug report and the patch.
common/index.h is needed by liblzma internally and tests. common.h will
include and define many things that are not needed by the tests. Also,
this prevents include order problems because common.h will redefine
LZMA_API resulting in a warning.
5.5.0alpha won't be released, it's just to mark that
the branch is not for stable 5.4.x.
Once again there is no API/ABI stability for new features
in devel versions. The major soname won't be bumped even
if API/ABI of new features breaks between devel releases.
HAVE_DECL_PROGRAM_INVOCATION_NAME is renamed to
HAVE_PROGRAM_INVOCATION_NAME. Previously,
HAVE_DECL_PROGRAM_INVOCATION_NAME was always set when
building with autotools. CMake would only set this when it was 1, and the
dos/config.h did not define it. The new macro definition is consistent
across build systems.
Previously, <sys/time.h> was always included, even if mythread only used
clock_gettime. <time.h> is still needed even if clock_gettime is not used
though because struct timespec is needed for mythread_condtime.
Previously, if threading was enabled HAVE_DECL_CLOCK_MONOTONIC would always
be set to 0 or 1. However, this macro was needed in xz so if xz was not
built with threading and HAVE_DECL_CLOCK_MONOTONIC was not defined but
HAVE_CLOCK_GETTIME was, it caused a warning during build. Now,
HAVE_DECL_CLOCK_MONOTONIC has been renamed to HAVE_CLOCK_MONOTONIC and
will only be set if it is 1.
Using return_if_error on lzma_lzma_lclppb_encode was improper because
return_if_error is expecting an lzma_ret value, but
lzma_lzma_lclppb_encode returns a boolean. This could result in
lzma_microlzma_encoder, which would be misleading for applications.
The code that parses --memlimit options and --block-list modified
the argv[] when parsing the option string from optarg. This was
visible in "ps auxf" and such and could be confusing. I didn't
understand it back in the day when I wrote that code. Now a copy
is allocated when modifiable strings are needed.
The API docs gave an impression that such checks are done
but they actually weren't done. In practice it made little
difference since the calling code has a bug if these are NULL.
Thanks to Jia Tan for the original patch that checked for
block->filters == NULL.
If someone sets up Clang to define __GNUC__ to 10 or greater
then symvers broke. __has_attribute is supported by such GCC
and Clang versions that don't support __symver__ so this should
be much better and simpler way to detect if __symver__ is
actually supported.
Thanks to Tomasz Gajc for the bug report.
It has some complicated downsides and its usefulness is more limited
than I originally thought. So this change is bad for certain very
specific situations but a generic solution that works for other
filters (and is otherwise better too) is planned anyway. And this
way 7-Zip can use the same compatible filter for the .7z format.
This is still marked as experimental with a new temporary Filter ID.
lzma_str_to_filters() uses static error messages which makes
them not very precise. It tells the position in the string
where an error occurred though which helps quite a bit if
applications take advantage of it. Dynamic error messages can
be added later with a new flag if it seems important enough.
Some file formats need support for LZMA1 streams that don't use
the end of payload marker (EOPM) alias end of stream (EOS) marker.
So far liblzma API has supported decompressing such streams via
lzma_alone_decoder() when .lzma header specifies a known
uncompressed size. Encoding support hasn't been available in the API.
Instead of adding a new LZMA1-only API for this purpose, this commit
adds a new filter ID for use with raw encoder and decoder. The main
benefit of this approach is that then also filter chains are possible,
for example, if someone wants to implement support for .7z files that
use the x86 BCJ filter with LZMA1 (not BCJ2 as that isn't supported
in liblzma).
That is, if the specified nice_len is smaller than the minimum
of the match finder, silently use the match finder's minimum value
instead of reporting an error. The old behavior is annoying to users
and it complicates xz options handling too.
It not only makes no sense to put symbol versions into a static library
but it can also cause breakage.
By default Libtool #defines PIC if building a shared library and
doesn't define it for static libraries. This is documented in the
Libtool manual. It can be overriden using --with-pic or --without-pic.
configure.ac detects if --with-pic or --without-pic is used and then
gives an error if neither --disable-shared nor --disable-static was
used at the same time. Thus, in normal situations it works to build
both shared and static library at the same time on GNU/Linux,
only --with-pic or --without-pic requires that only one type of
library is built.
Thanks to John Paul Adrian Glaubitz from Debian for reporting
the problem that occurred on ia64:
https://www.mail-archive.com/xz-devel@tukaani.org/msg00610.html
lzma_filters_free() sets the options to NULL and ids to
LZMA_VLI_UNKNOWN so there is no need to do it by caller;
the filter arrays will always be left in a safe state.
Also use memcpy() instead of a loop to copy a filter chain
when it is known to be safe to copy LZMA_FILTERS_MAX + 1
(even if the elements past the terminator might be uninitialized).
This time it can happen when lzma_stream_encoder_mt() is used
to reinitialize an existing multi-threaded Stream encoder
and one of 1-4 tiny allocations in lzma_filters_copy() fail.
It's very similar to the previous bug
10430fbf38, happening with
an array of lzma_filter structures whose old options are freed
but the replacement never arrives due to a memory allocation
failure in lzma_filters_copy().
The documentation mentions that lzma_block_encoder() supports
LZMA_SYNC_FLUSH but it was never added to supported_actions[]
in the internal structure. Because of this, LZMA_SYNC_FLUSH could
not be used with the Block encoder unless it was the next coder
after something like stream_encoder() or stream_encoder_mt().
The bug was in the single-threaded .xz Stream encoder
in the code that is used for both re-initialization and for
lzma_filters_update(). To trigger it, an application had
to either re-initialize an existing encoder instance with
lzma_stream_encoder() or use lzma_filters_update(), and
then one of the 1-4 tiny allocations in lzma_filters_copy()
(called from stream_encoder_update()) must fail. An error
was correctly reported but the encoder state was corrupted.
This is related to the recent fix in
f8ee61e74e which is good but
it wasn't enough to fix the main problem in stream_encoder.c.
The encoder doesn't support dictionary sizes larger than 1536 MiB.
This is validated, for example, when calculating the memory usage
via lzma_raw_encoder_memusage(). It is also enforced by the LZ
part of the encoder initialization. However, LZMA encoder with
LZMA_MODE_NORMAL did an unsafe calculation with dict_size before
such validation and that results in an infinite loop if dict_size
was 2 << 30 or greater.
This reverts commit 177bdc922c
and also does equivalent change to arm64.c.
Now that ARM64 filter will use lzma_options_bcj, this change
is not needed anymore.
It also works on E2K as it supports these intrinsics.
On x86-64 runtime detection is used so the code keeps working on
older processors too. A CLMUL-only build can be done by using
-msse4.1 -mpclmul in CFLAGS and this will reduce the library
size since the generic implementation and its 8 KiB lookup table
will be omitted.
On 32-bit x86 this isn't used by default for now because by default
on 32-bit x86 the separate assembly file crc64_x86.S is used.
If --disable-assembler is used then this new CLMUL code is used
the same way as on 64-bit x86. However, a CLMUL-only build
(-msse4.1 -mpclmul) won't omit the 8 KiB lookup table on
32-bit x86 due to a currently-missing check for disabled
assembler usage.
The configure.ac check should be such that the code won't be
built if something in the toolchain doesn't support it but
--disable-clmul-crc option can be used to unconditionally
disable this feature.
CLMUL speeds up decompression of files that have compressed very
well (assuming CRC64 is used as a check type). It is know that
the CLMUL code is significantly slower than the generic code for
tiny inputs (especially 1-8 bytes but up to 16 bytes). If that
is a real-world problem then there is already a commented-out
variant that uses the generic version for small inputs.
Thanks to Ilya Kurdyukov for the original patch which was
derived from a white paper from Intel [1] (published in 2009)
and public domain code from [2] (released in 2016).
[1] https://www.intel.com/content/dam/www/public/us/en/documents/white-papers/fast-crc-computation-generic-polynomials-pclmulqdq-paper.pdf
[2] https://github.com/rawrunprotected/crc
This uses it for CRC table initializations when using --disable-small.
It avoids mythread_once() overhead. It also means that then
--disable-small --disable-threads is thread-safe if this attribute
is supported.
__SSE2__ is the correct macro for SSE2 support with GCC, Clang,
and ICC. __SSE2_MATH__ means doing floating point math with SSE2
instead of 387. Often the latter macro is defined if the first
one is but it was still a bug.
In practice this means making the scripts work when
the input files have an unsupported check type which
isn't a problem in practice unless support for
some check types has been disabled at build time.
Modern 32-bit ARM in big endian mode use little endian for
instruction encoding still, so the filters work on such
executables too. It's likely less confusing for users this way.
The --arm64 option hasn't been implemented yet (there is
--experimental-arm64 but it's different). The --arm64 option
is added now anyway because this is the likely result and the
strings need to be ready for translators.
Thanks to Jia Tan.
If configured with --disable-lzip-decoder then --long-help will
still list `lzip' in --format but I left it like that since
due to translations it would be messy to have two help strings.
Features are disabled only in special situations so wrong help
in such a situation shouldn't matter much.
Thanks to Michał Górny for the original patch.
Support for format version 0 was removed from lzip 1.18 for some
reason. .lz format version 0 files are rare (and old) but some
source packages were released in this format, and some people might
have personal files in this format too. It's very little extra code
to support it along side format version 1 so this commits adds
support for both.
The Sync Flush marker extentension to the original .lz format
version 1 isn't supported. It would require changes to the
LZMA decoder itself. Such files are very rare anyway.
See the API doc for lzma_lzip_decoder() for more details about
the .lz format support.
Thanks to Michał Górny for the original patch.
"xz -v < regular_file > out.xz" doesn't display the percentage
and estimated remaining time because it doesn't even try to
check the input file size when input is read from stdin.
This could be improved but for now there's just a comment
to remind about it.
It worked for one input file since the counters are zero when
xz starts but they weren't reset when starting a new file in
passthru mode. For example, if files A, B, and C are one byte each,
then "xz -dcvf A B C" would show file sizes as 1, 2, and 3 bytes
instead of 1, 1, and 1 byte.
This affects lzma_memusage() and lzma_memlimit_set() when used
with the threaded decompressor. Now all allocations are reported
by lzma_memusage() (so it's not misleading) and lzma_memlimit_set()
cannot lower the limit below that value.
The alternative would have been to allow lowering the limit if
doing so is possible by freeing the cached memory but since
the primary use case of lzma_memlimit_set() is to increase
memlimit after LZMA_MEMLIMIT_ERROR this simple approach
was selected.
The cached memory was always included when enforcing
the memory usage limit while decoding.
Thanks to Jia Tan.
Don't call InitOnceComplete() if initialization was already done.
So far mythread_once() has been needed only when building
with --enable-small. windows/build.bash does this together
with --disable-threads so the Vista-specific mythread_once()
is never needed by those builds. VS project files or
CMake-builds don't support HAVE_SMALL builds at all.
Example:
$ xz -dc --single-stream good-0-empty.xz
xz: good-0-empty.xz: Internal error (bug)
The code, that is tries to catch some input file issues early,
didn't anticipate LZMA_STREAM_END which is possible in that
code only when --single-stream is used.
Now files with unsupported check will make xz display
a warning, set the exit status to 2 (unless --no-warn is used),
and then decompress the file normally. This is how it was
supposed to work since the beginning but this was broken by
the commit 231c3c7098, that is,
a little before 5.0.0 was released. The buggy behavior displayed
a message, set exit status 1 (error), and xz didn't attempt to
to decompress the file.
This doesn't matter today except for special builds that disable
CRC64 or SHA-256 at build time (but such builds should be used
in special situations only). The bug matters if new check type
is added in the future and an old xz version is used to decompress
such a file; however, it's likely that such files would use a new
filter too and an old xz wouldn't be able to decompress the file
anyway.
The first hunk in the commit is the actual fix. The second hunk
is a cleanup since LZMA_TELL_ANY_CHECK isn't used in xz.
There is a test file for unsupported check type but it wasn't
used by test_files.sh, perhaps due to different behavior between
xz and the simpler xzdec.
Treating it as a warning (message + exit status 2) matches gzip
and it seems more logical as at that point the output file has
already been successfully closed. When it's a warning it is
possible to suppress it with --no-warn.
On OpenBSD the number of cores online is often less
than what HW_NCPU would return because OpenBSD disables
simultaneous multi-threading (SMT) by default.
Thanks to Christian Weisgerber.
When encoders were disabled and threading enabled, outqueue.c and
outqueue.h were not compiled. The multi threaded decoder required
these files, so compilation failed.
The documentation states LZMA_PROG_ERROR can be returned from
lzma_index_cat. Previously, lzma_index_cat could not return
LZMA_PROG_ERROR. Now, the validation is similar to
lzma_index_append, which does a NULL check on the index
parameter.
The check type of the last Stream in dest was never copied to
dest->checks (the code tried to copy it but it was done too late).
This meant that the value returned by lzma_index_checks() would
only include the check type of the last Stream when multiple
lzma_indexes had been concatenated.
In xz --list this meant that the summary would only list the
check type of the last Stream, so in this sense this was only
a visual bug. However, it's possible that some applications
use this information for purposes other than merely showing
it to the users in an informational message. I'm not aware of
such applications though and it's quite possible that such
applications don't exist.
Regular streamed decompression in xz or any other application
doesn't use lzma_index_cat() and so this bug cannot affect them.
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.
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.
Running the current xzgrep on Slackware 10.1 with GNU bash 3.00.15:
xzgrep: line 231: syntax error near unexpected token `;;'
On SCO OpenServer 5.0.7 with Korn Shell 93r:
syntax error at line 231 : `;;' unexpected
Turns out that some old shells don't like apostrophes (') inside
command substitutions. For example, the following fails:
x=$(echo foo
# asdf'zxcv
echo bar)
printf '%s\n' "$x"
The problem was introduced by commits
69d1b3fc29 (2022-03-29),
bd7b290f3f (2022-07-18), and
a648978b20 (2022-07-19).
5.2.6 is the only stable release that included
this problem.
Thanks to Kevin R. Bulgrien for reporting the problem
on SCO OpenServer 5.0.7 and for providing the fix.
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.
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.
This documents the changes made in commits
6c6da57ae2,
cad299008c, and
898faa9728.
The --info-memory bit hasn't been finished yet
even though it's already mentioned in this commit
under --memlimit-mt-decompress and --threads.
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.
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:
https://debbugs.gnu.org/cgi/bugreport.cgi?bug=22900#25
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.
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.
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.
Also replace one use of expr with printf.
The rationale for LC_ALL=C was already mentioned in
69d1b3fc29 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:
https://pubs.opengroup.org/onlinepubs/9699919799/utilities/sed.html
With GNU sed invalid multibyte strings would work without this;
it's documented in its Texinfo manual. Some other implementations
aren't so forgiving.
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.
Add comments.
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.
Thanks to Karl Beldan for reporting the problem.
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.
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.
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.
In single-threaded mode, --memlimit-compress can make xz scale down
the LZMA2 dictionary size to meet the memory usage limit. This
obviously affects the compressed output. However, if xz was in
threaded mode, --memlimit-compress could make xz reduce the number
of threads but it wouldn't make xz switch from multithreaded mode
to single-threaded mode or scale down the LZMA2 dictionary size.
This seemed illogical and there was even a "FIXME?" about it.
Now --memlimit-compress can make xz switch to single-threaded
mode if one thread in multithreaded mode uses too much memory.
If memory usage is still too high, then the LZMA2 dictionary
size can be scaled down too.
The option --no-adjust was also changed so that it no longer
prevents xz from scaling down the number of threads as that
doesn't affect compressed output (only performance). After
this commit --no-adjust only prevents adjustments that affect
compressed output, that is, with --no-adjust xz won't switch
from multithreaded mode to single-threaded mode and won't
scale down the LZMA2 dictionary size.
The man page wasn't updated yet.
--memlimit-mt-decompress allows specifying the limit for
multithreaded decompression. This matches memlimit_threading in
liblzma. This limit can only affect the number of threads being
used; it will never prevent xz from decompressing a file. The
old --memlimit-decompress option is still used at the same time.
If the value of --memlimit-decompress (the default value or
one specified by the user) is less than the value of
--memlimit-mt-decompress , then --memlimit-mt-decompress is
reduced to match --memlimit-decompress.
Man page wasn't updated yet.
In most cases if the input file is corrupt the application won't
care about the uncompressed content at all. With this new flag
the threaded decoder will return an error as soon as any thread
has detected an error; it won't wait to copy out the data before
the location of the error.
I don't plan to use this in xz to keep the behavior consistent
between single-threaded and multi-threaded modes.
This makes it possible to call lzma_code() in a loop that only
reads new input when lzma_code() didn't fill the output buffer
completely. That isn't the calling style suggested by the
liblzma example program 02_decompress.c so perhaps the usefulness
of this feature is limited.
Also, it is possible to write such a loop so that it works
with the single-threaded decoder but not with the threaded
decoder even after this commit, or so that it works only if
lzma_mt.timeout = 0.
The zlib tutorial <https://zlib.net/zlib_how.html> is a well-known
example of a loop where more input is read only when output isn't
full. Porting this as is to liblzma would work with the
single-threaded decoder (if LZMA_CONCATENATED isn't used) but it
wouldn't work with threaded decoder even after this commit because
the loop assumes that no more output is possible when it cannot
read more input ("if (strm.avail_in == 0) break;"). This cannot
be fixed at liblzma side; the loop has to be modified at least
a little.
I'm adding this in any case because the actual code is simple
and short and should have no harmful side-effects in other
situations.
Malicious filenames can make xzgrep to write to arbitrary files
or (with a GNU sed extension) lead to arbitrary code execution.
xzgrep from XZ Utils versions up to and including 5.2.5 are
affected. 5.3.1alpha and 5.3.2alpha are affected as well.
This patch works for all of them.
This bug was inherited from gzip's zgrep. gzip 1.12 includes
a fix for zgrep.
The issue with the old sed script is that with multiple newlines,
the N-command will read the second line of input, then the
s-commands will be skipped because it's not the end of the
file yet, then a new sed cycle starts and the pattern space
is printed and emptied. So only the last line or two get escaped.
One way to fix this would be to read all lines into the pattern
space first. However, the included fix is even simpler: All lines
except the last line get a backslash appended at the end. To ensure
that shell command substitution doesn't eat a possible trailing
newline, a colon is appended to the filename before escaping.
The colon is later used to separate the filename from the grep
output so it is fine to add it here instead of a few lines later.
The old code also wasn't POSIX compliant as it used \n in the
replacement section of the s-command. Using \<newline> is the
POSIX compatible method.
LC_ALL=C was added to the two critical sed commands. POSIX sed
manual recommends it when using sed to manipulate pathnames
because in other locales invalid multibyte sequences might
cause issues with some sed implementations. In case of GNU sed,
these particular sed scripts wouldn't have such problems but some
other scripts could have, see:
info '(sed)Locale Considerations'
This vulnerability was discovered by:
cleemy desu wayo working with Trend Micro Zero Day Initiative
Thanks to Jim Meyering and Paul Eggert discussing the different
ways to fix this and for coordinating the patch release schedule
with gzip.
If a worker thread has consumed all input so far and it's
waiting on thr->cond and then the main thread enables
partial update for that thread, the code used to deadlock.
This commit allows one dummy decoding pass to occur in this
situation which then also does the partial update.
As part of the fix, this moves thr->progress_* updates to
avoid the second thr->mutex locking.
Thanks to Jia Tan for finding, debugging, and reporting the bug.
LZMA_TIMED_OUT is not an error and thus stopping threads on
LZMA_TIMED_OUT breaks the decoder badly.
Thanks to Jia Tan for finding the bug and for the patch.
If threading support is enabled at build time, this will
use lzma_stream_decoder_mt() even for single-threaded mode.
With memlimit_threading=0 the behavior should be identical.
This needs some work like adding --memlimit-threading=LIMIT.
The original patch from Sebastian Andrzej Siewior included
a method to get currently available RAM on Linux. It might
be one way to go but as it is Linux-only, the available-RAM
approach needs work for portability or using a fallback method
on other OSes.
The man page wasn't updated yet.
Jia Tan
Lasse Collin