Using __attribute__((__no_profile_instrument_function__)) on the ifunc
resolver works around a bug in GCC -fprofile-generate:
it adds profiling code even to ifunc resolvers which can make
the ifunc resolver crash at program startup. This attribute
was not introduced until GCC 7 and Clang 13, so ifunc won't
be used with prior versions of these compilers.
This bug was brought to our attention by:
https://bugs.gentoo.org/925415
And was reported to upstream GCC by:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=11411
The previous Linux Landlock feature test assumed that having the
linux/landlock.h header file was enough. The new feature tests also
requires that prctl() and the required Landlock system calls are
supported.
The original code was good enough for supporting GNU/Linux
and a few others but it wasn't very portable.
CMake doesn't support Solaris Studio's -xldscope=hidden.
If it ever does, things should still work with this commit
as Solaris Studio supports not only its own __global but also
the GNU C __attribute__((visibility("default"))). Support for the
attribute was added in 2007 to Sun Studio 12 compiler version 5.9.
-O3 doesn't seem useful for speed but it makes the code bigger.
CMake makes is difficult for users to simply override the
optimization level: CFLAGS / CMAKE_C_FLAGS aren't helpful because
they go before CMAKE_C_FLAGS_RELEASE. Of course, users can override
CMAKE_C_FLAGS_RELEASE directly but then they have to remember to
add also -DNDEBUG to disable assertions.
This commit changes -O3 to -O2 in CMAKE_C_FLAGS_RELEASE if and only if
CMAKE_C_FLAGS_RELEASE cache variable doesn't already exist. So if
a custom value is passed on the command line (or reconfiguring an
already-configured build), the cache variable won't be modified.
In contrast to Automake, skipping of this test when decoders
are disabled is handled at CMake side instead of test_scripts.sh
because CMake-build doesn't create config.h.
Compared to the Autotools-based build, this has simpler handling
for the shell (@POSIX_SHELL@) and extra PATH entry for the scripts
(configure has --enable-path-for-scripts=PREFIX). The simpler
metho should be enough for non-ancient systems and Solaris.
It helps that cmake_install.cmake doesn't parallelize installation
so symlinks can be created so that the target is always known to
exist (a requirement on Windows in some cases).
This bumps the minimum CMake version from 3.13 to 3.14 to use
file(CREATE_LINK ...). It could be made to work on 3.13 by
calling "cmake -E create_symlink" but it's uglier code and
slower in "make install". 3.14 should be a reasonable version
to require nowadays, especially since the Autotools build
is still the primary build system for most OSes.
If gettext tools are available, the .po files listed in po/LINGUAS
are converted using msgfmt. This allows building with translations
directly from xz.git without Autotools.
If gettext tools aren't available, the Autotools-created .gmo files
in the "po" directory will be used. This allows CMake-based build
to use translations from Autotools-generated tarball.
If translation support is found (Intl_FOUND) but both the
gettext tools and the pre-generated .gmo files are missing,
then "make" will fail.
Landlock is now always used just like pledge(2) is: first in more
permissive mode and later (under certain common conditions) in
a strict mode that doesn't allow opening more files.
I put pledge(2) first in sandbox.c because it's the simplest API
to use and still somewhat fine-grained for basic applications.
So it's the simplest thing to understand for anyone reading sandbox.c.
Even though the proper name for the architecture is aarch64, this
project uses ARM64 throughout. So the rename is for consistency.
Additionally, crc32_arm64.h was slightly refactored for the following
changes:
* Added MSVC, FreeBSD, and macOS support in
is_arch_extension_supported().
* crc32_arch_optimized() now checks the size when aligning the
buffer.
* crc32_arch_optimized() loop conditions were slightly modified to
avoid both decrementing the size and incrementing the buffer
pointer.
* Use the intrinsic wrappers defined in <arm_acle.h> because GCC and
Clang name them differently.
* Minor spacing and comment changes.
The CRC32 instructions in ARM64 can calculate the CRC32 result
for 8 bytes in a single operation, making the use of ARM64
instructions much faster compared to the general CRC32 algorithm.
Optimized CRC32 will be enabled if ARM64 has CRC extension
running on Linux.
Signed-off-by: Chenxi Mao <chenxi.mao2013@gmail.com>
A CLMUL-only build will have the crcxx_clmul() inlined into
lzma_crcxx(). Previously a jump to the extern lzma_crcxx_clmul()
was needed. Notes about shared liblzma on ELF platforms:
- On platforms that support ifunc and -fvisibility=hidden, this
was silly because CLMUL-only build would have that single extra
jump instruction of extra overhead.
- On platforms that support neither -fvisibility=hidden nor linker
version script (liblzma*.map), jumping to lzma_crcxx_clmul()
would go via PLT so a few more instructions of overhead (still
not a big issue but silly nevertheless).
There was a downside with static liblzma too: if an application only
needs lzma_crc64(), static linking would make the linker include the
CLMUL code for both CRC32 and CRC64 from crc_x86_clmul.o even though
the CRC32 code wouldn't be needed, thus increasing code size of the
executable (assuming that -ffunction-sections isn't used).
Also, now compilers are likely to inline crc_simd_body()
even if they don't support the always_inline attribute
(or MSVC's __forceinline). Quite possibly all compilers
that build the code do support such an attribute. But now
it likely isn't a problem even if the attribute wasn't supported.
Now all x86-specific stuff is in crc_x86_clmul.h. If other archs
The other archs can then have their own headers with their own
is_clmul_supported() and crcxx_clmul().
Another bonus is that the build system doesn't need to care if
crc_clmul.c is needed.
is_clmul_supported() stays as inline function as it's not needed
when doing a CLMUL-only build (avoids a warning about unused function).
The sandbox is now enabled for xzdec as well, so it no longer belongs
in just the xz section. xz and xzdec are always built, except for older
MSVC versions, so there isn't a need to conditionally show the sandbox
configuration. CMake will do a little unecessary work on older MSVC
versions that can't build xz or xzdec, but this is a very small
downside.
A CMake option LARGE_FILE_SUPPORT is created if and only if
-D_FILE_OFFSET_BITS=64 affects sizeof(off_t).
This is needed on many 32-bit platforms and even with 64-bit builds
with MinGW-w64 to get support for files larger than 2 GiB.
Autotools based build uses -pthread and thus adds it to Libs.private
in liblzma.pc. CMake doesn't use -pthread at all if pthread functions
are available in libc so Libs.private doesn't get -pthread either.
Now configure will fail if -fsanitize= is found in CFLAGS
and sanitizer-incompatible ifunc or Landlock sandboxing
would be used. These are incompatible with one or more sanitizers.
It's simpler to reject all -fsanitize= uses instead of trying to
pass those that might not cause problems.
CMake-based build was updated similarly. It lets the configuration
finish (SEND_ERROR instead of FATAL_ERROR) so that both error
messages can be seen at once.
Using set(ENABLE_THREADS "posix") is confusing because it sets
a new normal variable and leaves the cache entry with the same
name unchanged. The intent wasn't to change the cache entry so
this switches to a different variable name.
This way typos are caught quickly and compounding error messages
are avoided (a single typo could cause more than one error).
This keeps using SEND_ERROR when the system is lacking a feature
(like threading library or sandboxing method). This way the whole
configuration log will be generated in case someone wishes to
report a problem upstream.
The option is enabled by default, but will only be visible to a user
listing cache variables or using a CMake GUI application if the
immintrin.h header file is found.
This mirrors our Autotools build --disable-clmul-crc functionality.
Both crc32_clmul() and crc64_clmul() are now exported from
crc32_clmul.c as lzma_crc32_clmul() and lzma_crc64_clmul(). This
ensures that is_clmul_supported() (now lzma_is_clmul_supported()) is
not duplicated between crc32_fast.c and crc64_fast.c.
Also, it encapsulates the complexity of the CLMUL implementations into a
single file and reduces the complexity of crc32_fast.c and crc64_fast.c.
Before, CLMUL code was present in crc32_fast.c, crc64_fast.c, and
crc_common.h.
During the conversion, various cleanups were applied to code (thanks to
Lasse Collin) including:
- Require using semicolons with MASK_/L/H/LH macros.
- Variable typing and const handling improvements.
- Improvements to comments.
- Fixes to the pragmas used.
- Removed unneeded variables.
- Whitespace improvements.
- Fixed CRC_USE_GENERIC_FOR_SMALL_INPUTS handling.
- Silenced warnings and removed the need for some #pragmas
Jia Tan
Lasse Collin
Chenxi Mao