These functions are useful outside of just the clang recipe, so move
them to a common .inc file so they can be used by other clang-related
recipes.
Also make the function fail if it doesn't recognise the architecture,
instead of returning the empty string and causing mysterious fails later.
(From OE-Core rev: 1d5298533e97dab7636f885ddd740352782395b0)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This variable controls how ccache determines if the cache is valid or not
for a given compiler.
By default, ccache looks at the mtime of the compiler binary. This should
be sufficient for our purposes as the mtime is preserved from sstate
and will only be changed on rebuilds if the binary is actually different.
(From OE-Core rev: 27aec0bab6fcb19c5d5be30763f4038a46922e8f)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Move some environment variables to the configuration file, so there's
less happening in the class.
Max_size was removed so that the per-recipe caches don't grown without
limit. The default cache is 5GB.
(From OE-Core rev: 8c1c38a6b3073d2dd34514f5a90805c12c0f42dd)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We add FIRMWARE_DIR via CFLAGS using escapes for string and clang
option parser flags it
<command line>:4:23: warning: missing terminating '"' character [-Winvalid-pp-token]
4 | #define FIRMWARE_DIR \"/usr/lib/firmware\"
| ^
emitting this warning during configure means some tests fail wrongly. e.g.
test for enable -fPIC -DPIC, which results in textrels in libbbluetooth.so
Pass the string definition using single quotes to avoid the shell
"eating" the quotes and other characters
This works well across gcc and clang
(From OE-Core rev: 23a87f27b1837bd35bea356f0a10816fb35826f2)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The test depends on BSD extension REG_STARTEND which
is not in POSIX, glibc does implement this extention so it
works on glibc systems but musl does not and it needs to
be skipped and LIBC environment variable is used in this
test as well to run it conditionally.
(From OE-Core rev: a74eaca8e1830d09c5ff678c7d361cedbfc6f68a)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Upgrade to U-Boot 2025.07.
Add patches for build race conditions.
Remove the QEMU KVM USB workaround since the issue is fixed upstream.
(From OE-Core rev: bb287fade558d0138f7254876c9029da63e3087f)
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
==========
- Fix usage of typing_extensions.TypedDict nested inside other types
(e.g., typing.Type[typing_extensions.TypedDict]). This is not allowed by the
type system but worked on older versions, so we maintain support.
(From OE-Core rev: a92c8559ea91dc5b4a37e104715d32135341c7cb)
Signed-off-by: Liu Yiding <liuyd.fnst@fujitsu.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
==========
- Fix extent-based deduplication which can cause data corruption if target images are larger than 4GiB
- Switch to on-heap temporary buffers for libzstd and libdeflate to replace on-stack VLAs
- Fix large compressed fragment handling, which could be generated by the `-Eall-fragments` option (though rare) and was rejected by mistake
- Fix corrupted small fragments introduced in erofs-utils 1.8.8
- Fix AUFS whiteout handling
- Properly handle negative GNU mtime
- Fix superblock checksum for small fs block size filesystems
- Fix temporary memory leak from small fragments
- Handle crafted Z_EROFS_COMPRESSION_INTERLACED extents
- Speed up multi-threaded `-Efragments` even further
- Fix DEFLATE due to incorrect maximum Huffman length
- Support `--fsalignblks` to align filesystem sizes
- Support `--vmdk-desc` to generate VMDK for flattened block devices
- Fix image reproducibility of `-E(all-)fragments`
- Support per-segment reaper for multi-threaded compression
- Support multi-threaded fragments
- Support extent-based deduplication for `-Efragments`
- Optimize space allocation performance
(From OE-Core rev: a29c24e540ce6cd70ae7889f69ebe14e024d5507)
Signed-off-by: Liu Yiding <liuyd.fnst@fujitsu.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
==========
- DeviceEventController: update mouse coordinates before sending button events
- Return localized role name for ATSPI_ROLE_EDITBAR
- Revert "Return localized role name for ATSPI_ROLE_EDITBAR"
- Return localized role name for ATSPI_ROLE_EDITBAR
- atspi-device-legacy: Don't crash when XkbGetMap fails
(From OE-Core rev: 9db21ce6bf37541b089751311a4b8781adb0263c)
Signed-off-by: Liu Yiding <liuyd.fnst@fujitsu.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
==========
- Fix bad parsing of lesskey file an env var is a prefix of another env var
- Fix unexpected exit using -K if a key press is received while reading the input file
(From OE-Core rev: 25f101eecdb9921bf42c99b63dd6f036ceff00b5)
Signed-off-by: Liu Yiding <liuyd.fnst@fujitsu.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changelog:
==========
- ecc: Check return value of l_getrandom()
- unit: test-ecdh: Add bytes to l_getrandom wrapper
- unit: Convert UUID test cases to use precheck feature
- unit: Convert PBKDF2 test cases to use precheck feature
- unit: Convert check for random support to use test precheck feature
- unit: Make the failed precheck test case a success
- test: Add flag for indication of inverted precheck results
- unit: Use the precheck handler for testing available key support
- unit: Use the precheck handler for testing available getrandom support
- unit: Use the precheck handler for testing available checksum support
- unit: Use the precheck handler for testing available cipher support
- test: Provide test data also to the precheck function
- unit: Add test cases for precheck of unit tests
- test: Add support for precheck handler for unsupported kernel features
- Add License variable to pkg-config file
(From OE-Core rev: 07b4a4dd2c9e109939cff5a8c58fdf1d01c4d24a)
Signed-off-by: Liu Yiding <liuyd.fnst@fujitsu.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
- v3dv: regression in vkAllocateMemory importing gbm bo
- Vulkan WSI (and zink) use threads on X11 even when the X
connection isn't thread-safe
- sddm-greeter-qt segfault when using nvk+zink
- [regression][bisected] [FirePro W4100]: crashing/rebooting
- Descriptor set layout with binding flags fails due to indices
not matching bindings
- piglit bindless texture tests crash
- [radeonsi] Artifacts in Team Fortress 2 (bisected)
- eglgears_wayland segfault on zink+nvk with PRIME
- vn_renderer_virtgpu.c:13:10: fatal error: 'xf86drm.h' file not
found
- brw: mad instruction printing broken on Gfx11
- radv: RGB9E5 rendering does not ignore alpha write mask
(From OE-Core rev: 0f43573fce7074dfe462e8f6df50404e7dc391c8)
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The version 20250709 is the author date.
* The changes are:
a2287c3 Adjust 'time-stamp' variables to modern Emacs
973e3e6 config.sub: Add Tock OS support
3a71dc1 config.sub: recognize banan_os
f91a544 config.sub: fix a duplicated case pattern
484648c config.sub: Include support for arm64_32 and iOS/tvOS/watchOS simulators
84cc2cc Update copyright years
7f41495 config.sub: add intelgt as a basic machine
9f6e0fe ironclad: change ironclad to ironclad-mlibc
(From OE-Core rev: 28372102300a1c670a6a4fcea3f2f527c73aee56)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This implementation is not part of standard and some implementations
e.g. libc++ have removed it starting with 19.x release
(From OE-Core rev: 66faa33a0a6d046f01c4bff26012cb3b1e11dba6)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When __USE_TIME64_REDIRECTS is defined, glibc redirects struct semid_ds
to a 64-bit time-safe version that omits the sem_otime_high and
sem_ctime_high fields. As a result, the case becomes invalid, leading to
incorrect behavior.
This patch adds a check to skip the test when __USE_TIME64_REDIRECTS is
defined, ensuring the test only runs when the semid_ds structurally
matches semid64_ds and the *_high fields are accessible.
(From OE-Core rev: 07410b54660f75465df8d428844893cbd0c295ea)
Signed-off-by: Jiaying Song <jiaying.song.cn@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For now, using qemuppc64 as MACHINE and building things inside SDK,
we'll get the following error:
ERROR: modpost: "_mcount" undefined
This is because after [1], MPROFILE_KERNEL is using arch/powerpc/tools/
contents to do the check. This in turn affects HAVE_OBJTOOL.
"""
arch/powerpc/Kconfig: select HAVE_OBJTOOL if ARCH_USING_PATCHABLE_FUNCTION_ENTRY || MPROFILE_KERNEL || PPC32
"""
The current MPROFILE_KERNEL for powerpc:
"""
config MPROFILE_KERNEL
depends on PPC64_ELF_ABI_V2 && FUNCTION_TRACER
def_bool $(success,$(srctree)/arch/powerpc/tools/gcc-check-mprofile-kernel.sh $(CC) -mlittle-endian) if CPU_LITTLE_ENDIAN
def_bool $(success,$(srctree)/arch/powerpc/tools/gcc-check-mprofile-kernel.sh $(CC) -mbig-endian) if CPU_BIG_ENDIAN
"""
If MPROFILE_KERNEL is disabled, thus HAVE_OBJTOOL being disabled, then
FTRACE_MCOUNT_USE_RECORDMCOUNT will be enabled, causing this error.
The scripts are listed and copied one by one. Their names and the reasons why
they are copied are as below:
gcc-check-mprofile-kernel.sh: needed by MPROFILE_KERNEL
gcc-check-fpatchable-function-entry.sh: needed by ARCH_USING_PATCHABLE_FUNCTION_ENTRY
head_check.sh/relocs_check.sh/unrel_branch_check.sh: needed by Makefile.postlink
[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=abba759796f9b73eb24df9b734dd063839fc62e0
(From OE-Core rev: df1c7c276422eeaf0eac178aec377a315a249c52)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Seen with opengl-es-cts compilation for qemuarm64
(From OE-Core rev: 79fb0c1a73e659f32bd5d86d84c583233d09e473)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is the patch that got merged into libc finally
(From OE-Core rev: eb79f29c5c2bf8fce13c02c7d7ca362c61e2b0ff)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add clang-cross to sysroot via POPULATESYSROOTDEPS to get llvm tools
e.g. llvm-strip, llvm-nm etc when they are used at default STRIP/NM etc.
(From OE-Core rev: 0d239586ec2f258a42d60f9fb6a058344516218a)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Backport a fix to address random access iterator's expectation of
being const-qualified inside libc++ std::__insertion_sort_unguarded
implementation
(From OE-Core rev: cfd74da11a5290b2e555e2f9f06779f816324ee1)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add traditional-activation PACKAGECONFIG to DBus recipe in order to allow
users to enable or disable traditional DBus service activation.
DBus service activation refers to automatically starting an application
when a DBus message is sent to a service provided by that application.
Traditionally, this is done by the DBus-daemon, thus, the term traditional
here. On systems using systemd, systemd can do this, instead.
On some systems it might be of interest to disable the traditional service
activation in order to ensure that services are always started via systemd.
Per default, traditional service activation is enabled for DBus. Thus, the
traditional-activation PACKAGECONFIG is added to the default PACKAGECONFIG.
Otherwise, we might introduce a breaking change here.
The binary dbus-daemon-launch-helper isn't created in case DBus is built
with traditional service activation disabled, so we change its attributes
only when traditional service activation is turned on.
(From OE-Core rev: 272aafbf0a2a1a158fdff42b6ce295fe861f1327)
Signed-off-by: Weisser, Pascal <pascal.weisser.ext@karlstorz.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There are new CVEs reported for this recipe which are not for this
componene, but for a component with same name from apache.
sqlite> select vendor, product, id, count(*) from products where product like 'orc' group by vendor, product, id;
apache|orc|CVE-2018-8015|1
apache|orc|CVE-2025-47436|4
gstreamer|orc|CVE-2024-40897|1
(From OE-Core rev: 683f19c2e3337784e50134b6a1331b4b9626894d)
Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The fdisk option is not enabled by default when it should be as that was
the existing behaviour.
Neither of the RDEPENDS are needed, as systemd doesn't call the tools
directly and the library dependencies will be generated automatically.
This reverts commit 4ec67113e1e7e1ecea9cde33ef4d3962dad5f2ad.
(From OE-Core rev: fac80672339bce66dc551c659932aa728478ffaf)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
LLDB defaults to adding rpaths into the binaries which are then stripped
by CMake on install.
However, this rpath removal is implemented by editing the binary instead
of relinking at install time, so the final binary will have an entry in
the dynstr section which is all nulls but is as long as the build path.
Obviously this breaks reproducibility, so disable the use of rpaths in
LLDB to remove this problem.
(From OE-Core rev: d96e0458b696a7359d310cbe112c5dc2fc60f97d)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
LLDB links to clang libraries, so it really does need clang and not just
LLVM.
(From OE-Core rev: 1a37bb2f0bbdbd599e1d5967eae1ecde4eff6a80)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Set LLVM_HEADERS_TBLGEN so that we don't build another copy of tblgen.
Also remove LLVM_OPTIMIZED_TABLEGEN, this is redundant as we're not
building a tablegen binary in this recipe anymore.
(From OE-Core rev: 259a3e9a8281c4e9ab73dee82738f359f029b78d)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The cmake.bbclass already does this.
(From OE-Core rev: 85c1aecdf847b97409930d915a90ab850e92bc38)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The GPU targets are incredibly slow to build, so if the DISTRO_FEATURES
doesn't include opengl or vulkan assume that the user will not be using
a GPU and disable them.
Alternatively, a distribution could state that they'll only be using
one of the backends, and set LLVM_TARGETS_GPU explicitly.
On my build machine, disabling the GPU targets reduces the build time of
clang-native from 21m to 16m.
(From OE-Core rev: 2273a0685757421f39541a352d77b67e5ba604d3)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Split the libclc subproject out of the clang recipe and into a dedicated
libclc recipe.
This is useful because libclc is the OpenCL runtime library and as such
isn't target-specific and needs a native clang to build, not a target
libllvm.
Verified that nothing is dropped by adding clang and libclc to an image
and verifying that the file list is the same before and after this
change.
We need to patch the libclc CMakeLists to allow it to use an out-of-tree
prepare_builtins binary, discussion is ongoing with upstream to resolve
this properly.
(From OE-Core rev: 33a8742a1280b4c6779a7aa487c2dd4a713babe6)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When TC_CXX_RUNTIME is set to be 'llvm' then clang looks for
libcxx to provide C++ runtime and unwinding library
(From OE-Core rev: aba2d949818b032222dbaed98e38cd29e3c992b6)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This ensures that default system linker is used correctly based upon
distro features, current default remain same i.e. uses BFD linker
(From OE-Core rev: 33e7c85e92ddbc35989e7afcaf5fe4c14efdefab)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
