Richard Purdie [Fri, 13 Feb 2026 12:22:45 +0000 (12:22 +0000)]
pseudo: Update to include an openat2 fix
We're seeing occasional autobuilder failures with tar issues related to openat2.
It appears there are definitions missing on debian 11 and opensuse 15.5 systems
which mean the openat2 syscall intercept isn't compiled in. This then triggers
on systems using the openat2 syscall, such as alma9 where it is used in a tar
CVE fix.
This updates to include the fix from upstream pseudo (along with a compile warning
fix).
This was tested by taking sstate for pseudo-native from a debian 11 system and using
it in a build of "bitbake nativesdk-git -c install" on a alma9 system where that task
failed. After this fix, it completes.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 9 Feb 2026 14:30:33 +0000 (14:30 +0000)]
populate_sdk_ext: Ensure BB_BASHSERBVE_DB_DIR is handled correctly
The new BB_BASHSERBVE_DB_DIR variable needs to be filtered when creating eSDK
local.conf files since we use tricks with the hashserve DB in the eSDK to
create a local cache file as part of eSDK creation.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
devtool: ide-sdk: Support GDB pretty-printing for C++ STL types
GDB requires Python scripts provided by GCC to properly display C++ STL
types. This commit adds support for configuring GDB to use these
pretty-printers in the ide-sdk, covering both the ide_none and ide_code
plugins.
The implementation locates the GCC Python helper scripts in the sysroot
and injects the necessary commands into the GDB initialization files and
IDE debug configurations. This ensures that when debugging C++
applications, STL containers and other complex types are displayed in a
readable format.
Without this:
(gdb) print numbers
$1 = {<std::_Vector_base<int, std::allocator<int> >> = {
_M_impl = {<std::allocator<int>> = {<std::__new_allocator<int>> =
{<No data fields>}, <No data fields>}, <std::_Vector_base<int,
std::allocator<int> >::_Vector_impl_data> =
{_M_start = 0x55555556c370, _M_finish = 0x55555556c37c,
_M_end_of_storage = 0x55555556c37c}, <No data fields>}},
<No data fields>}
With this:
(gdb) print numbers
$1 = std::vector of length 3, capacity 3 = {1, 2, 3}
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add a standard container (std::vector) to the C++ example program to
demonstrate the debugger's capability to inspect and traverse STL
containers during a debugging session. This requires enabling GDB's
pretty-printing feature, which depends on Python scripts shipped with
the compiler.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Improve the reverse mapping for searching the source files for remote
debugging by taking the details from DEBUG_PREFIX_MAP into account when
generating GDB's debug-file-directory mappings. This allows settings
such as DEBUG_PREFIX_MAP = "" for modified recipes to be used to avoid
any path remapping.
Background:
For packaged debug-symbols, the references to the source code need to be
relocated to paths which are valid on the target system. By default,
devtool ide-sdk tries to keep the relocated paths and configures the
debugger to reverse map them back to the original source paths. The goal
is to provide a debug setup which is a close as possible to a regular
build.
Usually this works well, but there are situations where the reverse
mapping is not unambiguous. For example the default DEBUG_PREFIX_MAP
adds two different source paths (${S} and ${B}) to the same target path
(${TARGET_DBGSRC_DIR}). If both source paths contain files with the same
name, the debugger cannot determine which source file to use. For this
example it is usually sufficient to only map ${S} to the target path.
The source files in ${B} are probably a few generated files which are
not that interesting for debugging. But depending on the project, the
files in ${B} might also be relevant for debugging.
Also add a hint to the generated local.conf snippet to use
DEBUG_PREFIX_MAP = "" if the user wants to optimize the build for
debugging.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This adds more test coverage for devtool ide-sdk, with VSCode. The cmake
test case has now a full remote debugging test on Qemu. The test checks
the generated launch.json and tasks.json files, starts gdbserver and
connects to it. The test verifies breakpoints, variables and source file
listing.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Write the ssh command to start gdbserver on target directly into the
tasks.json. This avoids the need to create one more script file on the
host. It also simplifies manual modifications of VSCode's standard
tasks.json which is much more handy than modifying multiple proprietary
scripts used to launch gdbserver.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
devtool: ide-sdk: make install_and_deploy script pass target arg
Previously, the target was hardcoded in the install_and_deploy script,
limiting flexibility. This change allows passing the target as a
command-line argument, enabling IDEs to configure the target dynamically
rather than only at IDE configuration generation time.
This is a first step towards making the target configurable from the IDE.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
devtool: ide-sdk: add gdbserver attach mode support
Enhance remote debugging configuration to support multiple modes
per executable binary. This adds support for gdbserver's attach
mode as an additional debug configuration.
When the binary is detected to run as a systemd service or SysV
init script, an attach debug configuration is generated alongside
the regular configuration that starts the process via gdbserver.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
oe-selftest: devtool: check example services are running
When running the devtool ide-sdk test with qemu, verify that the example
services are actually running on the target by using pgrep to check for
the example executable names.
Also verify that the configuration files in /etc are owned by the proper
user and group, both before and after the install_and_deploy scripts
have run. This is also a check that the install_and_deploy scripts
are working correctly with pseudo.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Extend the C++ example to run as systemd/SysV services
This change adds service capability to the existing C++ example without
modifying its original behavior. The example can now run either as:
- One-shot executables (existing behavior)
- Long-running services via systemd or SysV init
The service runs as an unprivileged user/group, demonstrating security
best practices for service development. This introduces additional
complexity to the build process, particularly around proper pseudo usage
in development builds. The implementation includes:
- Service configuration files (systemd .service and SysV init script)
- Dedicated user/group creation with appropriate permissions
- JSON configuration file for runtime customization, owned by the
service user
- Command-line --endless flag to enable service mode
- Full support for both CMake and Meson build systems
This enhancement enables testing debugger configurations that attach to
running processes, expanding the examples' utility for development tools.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add optional debug logging to all runCmd calls in DevtoolIdeSdkTests
to improve debugging capabilities when tests fail. The logging is only
enabled when the test logger is set to DEBUG level.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
With poky the oe-init-build-env script from the top level directory of
the layer with the higher priority is used to setup the build
environment.
This does no longer work with bitbake-setup. The directory layout
changed and the script is now called init-build-env. Skip the old
implementation if $TOPDIR/init-build-env exists and use it instead.
Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Peter Marko <peter.marko@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Peter Marko [Tue, 10 Feb 2026 08:20:12 +0000 (09:20 +0100)]
mtd-utils: fix ubifs packaging
This change moves fsck.ubifs and mount.ubifs to correct package.
Noticed when lzo was installed to ubi-less image additionally when
upgrading from scarthgap to master as being dependency of fsck.ubifs.
Signed-off-by: Peter Marko <peter.marko@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Barker [Tue, 10 Feb 2026 08:07:57 +0000 (08:07 +0000)]
selftest: Test installation of recipes with complex packaging
When recipes have complex and dynamic packaging, it's easy to
accidentally end up with one or more packages that build but can't be
installed. We've seen this recently with linux-firmware [1].
Skimming through the recipes that set PACKAGESPLITFUNCS or
PACKAGES_DYNAMIC or are otherwise known offenders, I've also picked out
gstreamer1.0-plugins-*, kernel-modules, perl and python3 as recipes
where it would be easy to end up with similar issues in the future.
spdx30_tasks: Exclude 'doc' when exporting PACKAGECONFIG to SPDX
Currently when generating an SBOM, all packages have the 'doc' feature
indicated as disabled. This is in fact *not* a feature that was
declared in the recipe, but instead the documentation of the
PACKAGECONFIG variable.
But to be safe, if somehow a feature is named 'doc' and enabled, do
not exclude it when exporting PACKAGECONFIG features to SPDX.
Signed-off-by: Benjamin Robin (Schneider Electric) <benjamin.robin@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fabio Berton [Mon, 9 Feb 2026 10:53:25 +0000 (10:53 +0000)]
classes/buildhistory: Do not sign buildhistory commits
This change adds the --no-gpg-sign option to prevent buildhistory
commits from failing due to GPG signing issues. Depending on the setup,
buildhistory may fail to create a commit if the user has the
commit.gpgsign option enabled.
For example, if the user creates a signing key that requires a password,
the commit won't be created and will fail with the following error:
/
|error: Enter passphrase: Load key "/home/<user>/.ssh/id_ed25519":
|incorrect passphrase supplied to decrypt private key?
|fatal: failed to write commit object
\
The bitbake command won't fail, but buildhistory won't have a commit.
Also, the commit may silently fail when building inside a container due
to missing packages or issues with accessing the GPG agent.
This is similar to [1], and signing the buildhistory commit
should be avoided to prevent such issues.
Wang Mingyu [Mon, 9 Feb 2026 09:02:44 +0000 (17:02 +0800)]
glslang: upgrade 1.4.335.0 -> 1.4.341.0
Changelog:
===========
- Allow gl_FragDepth identifier to be redeclared
- Replace 'GL_TASK(MESH)_SHADER_NV' with 'GL_TASK(MESH)_SHADER_EXT'
- Scalarize swizzled MeshEXT stores
- Support GL_EXT_long_vector
- Add missing preamble #define for GL_EXT_nontemporal_keyword
- Support GL_NV_push_constant_bank
- Support GL_EXT_descriptor_heap
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Wang Mingyu [Mon, 9 Feb 2026 09:02:46 +0000 (17:02 +0800)]
liburcu: upgrade 0.15.5 -> 0.15.6
Changelog:
===========
* urcu-mb: Add missing TSAN annotation to _urcu_mb_read_lock_update
* lfstack: Coding style cleanup
* urcu-qsbr: Use CMM_SEQ_CST_FENCE for _urcu_qsbr_thread_online
* urcu-mb: Use CMM_SEQ_CST_FENCE for _urcu_mb_read_lock_update
* urcu-qsbr: Use CMM_SEQ_CST_FENCE for quiescent state update and offline
* urcu-mb: Use CMM_SEQ_CST_FENCE for _urcu_mb_read_unlock_update_and_wakeup
* Fix: Only include linux/time_types.h when __NR_futex_time64 is defined
* Use __NR_futex_time64 in futex syscall wrapper
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Wang Mingyu [Mon, 9 Feb 2026 09:02:42 +0000 (17:02 +0800)]
gi-docgen: upgrade 2025.5 -> 2026.1
Changelog:
===========
- Parse method-inline for records, unions and interfaces
- Use dot's `svg_inline` output format
- Fix Red Hat Display italic font-face
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Peter Marko <peter.marko@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Quentin Schulz [Fri, 6 Feb 2026 17:23:57 +0000 (18:23 +0100)]
mesa: remove libxml2-native dependency
Native libxml2 is only needed for xmllint which in turn is only needed
when build-tests meson option is enabled, which defaults to disabled and
isn't enabled by the recipe.
We may need to revisit this if we add ptest support in mesa I assume.
Let's remove it from the DEPENDS for now.
Signed-off-by: Quentin Schulz <quentin.schulz@cherry.de> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Quentin Schulz [Fri, 6 Feb 2026 17:23:56 +0000 (18:23 +0100)]
mesa: make zlib dependency a PACKAGECONFIG
zlib is a mesa meson feature which can be disabled so let's make it a
PACKAGECONFIG.
The motivation actually isn't about disabling it (though we now could as
it's only required for a few things) but rather that this dependency is
a non-native one.
The issue is when a native-only (-native recipe name suffix + inherit
native) recipe includes mesa.inc, it'll include the DEPENDS as well but
will then depend on a non-native recipe because BitBake doesn't add a
suffix to dependencies listed in DEPENDS for native-only recipes like it
does for recipes with BBCLASSEXTEND = native. This is a way to
circumvent the issue by making the dependency a PACKAGECONFIG dependency
such that the listed DEPENDS (third parameter) will be properly modified
by BitBake to have the native variant of the recipe pulled in. See[1]
for more info on why this is needed.
This is in preparation for the addition of a mesa-tools-native
native-only package which will include mesa.inc.
Quentin Schulz [Fri, 6 Feb 2026 17:23:55 +0000 (18:23 +0100)]
mesa: make expat dependency a PACKAGECONFIG
expat is a mesa meson feature which can be disabled so let's make it a
PACKAGECONFIG.
The motivation actually isn't about disabling it (though we now could as
it's only required for a few things) but rather that this dependency is
a non-native one.
The issue is when a native-only (-native recipe name suffix + inherit
native) recipe includes mesa.inc, it'll include the DEPENDS as well but
will then depend on a non-native recipe because BitBake doesn't add a
suffix to dependencies listed in DEPENDS for native-only recipes like it
does for recipes with BBCLASSEXTEND = native. This is a way to
circumvent the issue by making the dependency a PACKAGECONFIG dependency
such that the listed DEPENDS (third parameter) will be properly modified
by BitBake to have the native variant of the recipe pulled in. See[1]
for more info on why this is needed.
Now that expat is a PACKAGECONFIG and that if it's not there expat isn't
autodetected anymore by meson, having xmlconfig PACKAGECONFIG add expat
as a DEPENDS (third parameter) is likely not enough anymore due to expat
feature being explicitly disabled if expat is not in PACKAGECONFIG.
Therefore, remove the third parameter for xmlconfig and add a comment
telling the user there's a PACKAGECONFIG dependency here. Note that
since this commit adds the expat PACKAGECONFIG to all recipes including
mesa.inc, no change in behavior is expected.
This is in preparation for the addition of a mesa-tools-native
native-only package which will include mesa.inc.
Ross Burton [Fri, 6 Feb 2026 14:57:15 +0000 (14:57 +0000)]
jquery: remove recipe
Now that the reproducible selftest doesn't use this recipe, it is no
longer used by any recipes that I can find.
There is an alternative jquery recipe in meta-oe that is slightly
different and has packaging consistent with with javascript recipes in
the same layer.
Signed-off-by: Ross Burton <ross.burton@arm.com> Reviewed-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Fri, 6 Feb 2026 14:57:14 +0000 (14:57 +0000)]
oeqa/selftests/reproducible: use the jQuery CDN instead of jquery-native
Instead of building jquery-native and then copying the file into the
expected place in the build tree, simply tell diffoscope to load jQuery
from their CDN.
In the slim chance that the diffoscope report is viewed offline, the
only regression is that sections cannot be collapsed.
The advantage to this approach is that we no longer need to maintain a
jquery-native recipe simply for the reproducible selftests.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
hosttools: force a --no-rosegment option if host ld supports it
Please see https://bugzilla.yoctoproject.org/show_bug.cgi?id=16087
for the tricky, unpleasant details.
The summary of the issue:
>From Fedora 42 onwards, the --rosegment argument is passed to the
linker by default when building native binaries. This triggers
re-arrangement of sections by patchelf, resulting in binaries which
are loaded incorrectly by older kernel versions. These "contaminated"
native binaries end up in sstate and cause build failures when they
are used on older distros such as AlmaLinux 8 & 9 or Rocky Linux 8 &
9.
[YOCTO #16087]
Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ankur Tyagi [Fri, 6 Feb 2026 10:43:20 +0000 (23:43 +1300)]
mpg123: upgrade 1.33.3 -> 1.33.4
Changelog:
mpg123: In terminal control, ignore 7-bit escape sequences to avoid spurious actions, e.g. when hitting cursor keys.
ports/cmake: Avoid possibly conflicting use of SIZEOF_OFF_T CMake variable when embedding mpg123 with other projects using cmake and different off_t semantics.
Signed-off-by: Ankur Tyagi <ankur.tyagi85@gmail.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* rust-oe-selftest.patch updated for 1.93:
Renamed UI test directive from add-core-stubs to add-minicore as in
https://github.com/rust-lang/rust/commit/3796f7de57f5d4cf4325d1f13aeb07fdaa57983b
* OE-selftests:
- Exclude src/tools/remote-test-client from test suite.
This unit test modifies the TEST_DEVICE_ADDR environment variable,
which breaks the OE test harness that uses the same variable for
QEMU remote testing. Filed a bug with upstream:
https://github.com/rust-lang/rust/issues/151823
- Enable +v8a feature for cortexa57.
Rust 1.93.0 added a regression test (PR #149549) to verify that the
ARMv8-A system register ttbr0_el2 is recognized in inline assembly.
This addresses issue #97724 where LLVM wasn't recognizing ttbr0_el2
as a valid register. Upstream Rust enables +v8a by default for all
AArch64 targets (PR #105026), but Yocto's custom target specification
doesn't inherit this. Adding +v8a ensures LLVM recognizes ARMv8-A
system registers, fixing the test failure.
Signed-off-by: Hemanth Kumar M D <Hemanth.KumarMD@windriver.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mingli Yu [Thu, 5 Feb 2026 08:51:19 +0000 (16:51 +0800)]
lttng-tools: Upgrade 2.14.0 -> 2.14.1
Remove 0001-gen-ust-events-constructor-change-rpath-to-libdir-li.patch
as the logic is included in 2.14.1 [1].
Remove 0001-eventfd.cpp-Remove-the-scope-resolution-operator.patch as
the logic is included in 2.14.1 [2].
Remove 0001-liblttng-ctl-drop-index-allocator-symbols-from-versi.patch
as the logic is included in 2.14.1 [3].
util_event_generator.sh is renamed to util_event_generator.py [4] and
the .py install logic in included as below.
find "${B}/tests/$d" -maxdepth 1 -type f -name "*.py" \
-exec install -t "${D}${PTEST_PATH}/tests/$d" {} +
Ross Burton [Wed, 4 Feb 2026 15:34:22 +0000 (15:34 +0000)]
gawk: trim native build configuration
When we build gawk-native it is only for use in builds where the host
gawk output isn't reproducible across versions[1]. As such it doesn't
need support for readline or mprf, and by removing those from gawk-native
we can get building gawk-native sooner.
[1] oe-core c5bbf0a60b ("gawk: use native gawk when building glibc and grub")
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Unfortunately the update is causing reproducibility issues:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/3317
https://valkyrie.yocto.io/pub/repro-fail/oe-reproducible-20260209-a0dgcw5d/packages/diff-html/
Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 10 Feb 2026 14:49:20 +0000 (14:49 +0000)]
python3-packaging: upgrade 25.0 -> 26.0
Skip the testcase test_pylock.py as this depends on tomli_w, which isn't
currently part of core.
Upstream changelog:
Features:
* PEP 751: support pylock
* PEP 794: import name metadata
* Support for writing metadata to a file
* Support __replace__ on Version
* Support positional pattern matching for Version and SpecifierSet
Behavior adaptations:
* PEP 440 handling of prereleases for Specifier.contains,
SpecifierSet.contains, and SpecifierSet.filter
* Handle PEP 440 edge case in SpecifierSet.filter
* Adjust arbitrary equality intersection preservation in SpecifierSet
* Return False instead of raising for .contains with invalid version
* Support arbitrary equality on arbitrary strings for Specifier and
SpecifierSet's filter and contains method.
* Only try to parse as Version on certain marker keys, return False on
unequal ordered comparisons
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Fri, 6 Feb 2026 20:27:25 +0000 (15:27 -0500)]
linux-yocto/6.18: beaglebone black: boot and graphics fixes
Integrating the following commit(s) to linux-yocto/6.18:
1/2 [
Author: Kevin Hao
Email: haokexin@gmail.com
Subject: net: cpsw_new: Execute ndo_set_rx_mode callback in a work queue
Date: Wed, 4 Feb 2026 14:27:57 +0800
Commit 1767bb2d47b7 ("ipv6: mcast: Don't hold RTNL for
IPV6_ADD_MEMBERSHIP and MCAST_JOIN_GROUP.") removed the RTNL lock for
IPV6_ADD_MEMBERSHIP and MCAST_JOIN_GROUP operations. However, this
change triggered the following call trace on my BeagleBone Black board:
WARNING: net/8021q/vlan_core.c:236 at vlan_for_each+0x120/0x124, CPU#0: rpcbind/496
RTNL: assertion failed at net/8021q/vlan_core.c (236)
Modules linked in:
CPU: 0 UID: 997 PID: 496 Comm: rpcbind Not tainted 6.19.0-rc6-next-20260122-yocto-standard+ #8 PREEMPT
Hardware name: Generic AM33XX (Flattened Device Tree)
Call trace:
unwind_backtrace from show_stack+0x28/0x2c
show_stack from dump_stack_lvl+0x30/0x38
dump_stack_lvl from __warn+0xb8/0x11c
__warn from warn_slowpath_fmt+0x130/0x194
warn_slowpath_fmt from vlan_for_each+0x120/0x124
vlan_for_each from cpsw_add_mc_addr+0x54/0xd8
cpsw_add_mc_addr from __hw_addr_ref_sync_dev+0xc4/0xec
__hw_addr_ref_sync_dev from __dev_mc_add+0x78/0x88
__dev_mc_add from igmp6_group_added+0x84/0xec
igmp6_group_added from __ipv6_dev_mc_inc+0x1fc/0x2f0
__ipv6_dev_mc_inc from __ipv6_sock_mc_join+0x124/0x1b4
__ipv6_sock_mc_join from do_ipv6_setsockopt+0x84c/0x1168
do_ipv6_setsockopt from ipv6_setsockopt+0x88/0xc8
ipv6_setsockopt from do_sock_setsockopt+0xe8/0x19c
do_sock_setsockopt from __sys_setsockopt+0x84/0xac
__sys_setsockopt from ret_fast_syscall+0x0/0x5
This trace occurs because vlan_for_each() is called within
cpsw_ndo_set_rx_mode(), which expects the RTNL lock to be held.
Since modifying vlan_for_each() to operate without the RTNL lock is not
straightforward, and because ndo_set_rx_mode() is invoked both with and
without the RTNL lock across different code paths, simply adding
rtnl_lock() in cpsw_ndo_set_rx_mode() is not a viable solution.
To resolve this issue, we opt to execute the actual processing within
a work queue, following the approach used by the icssg-prueth driver.
Fixes: 1767bb2d47b7 ("ipv6: mcast: Don't hold RTNL for IPV6_ADD_MEMBERSHIP and MCAST_JOIN_GROUP.") Signed-off-by: Kevin Hao <haokexin@gmail.com> Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20260203-bbb-v5-1-ea0ea217a85c@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
]
2/2 [
Author: Kevin Hao
Email: haokexin@gmail.com
Subject: drm/tilcdc: Set preferred depth
Date: Wed, 4 Feb 2026 14:27:58 +0800
The commit c91acda3a380 ("drm/gem: Check for valid formats") adds a
check for valid pixel formats on drm_gem_fb_create(), but this breaks
the X server on the beaglebone black board.
We have set 'DefaultDepth' to 16 in our xorg.conf. In the X modesetting
driver, the drmmode_get_default_bpp() is used to guess the default
depth/bpp. First it tries to get them via DRM_CAP_DUMB_PREFERRED_DEPTH
ioctl, and if it fail, then try to create a FB with 'depth = 24' and
'bpp = 32' to check whether this depth/dpp is a valid combo. Before the
kernel commit c91acda3a380, the FB always can be created successfully.
This will avoid the bpp to be set to 24 forcibly. But after kernel
commit c91acda3a380, the FB will not be created successfully due to the
check of the valid pixel format. Then the bpp is set to 24, but the
'depth = 16' and 'bpp = 24' combo is not a valid pixel format.
Fix this issue by explicitly setting the preferred_depth in this driver.
With this change, the modesetting driver would choose the correct
depth/bpp combo based on our setting in xorg.conf.
Bruce Ashfield [Fri, 6 Feb 2026 20:27:21 +0000 (15:27 -0500)]
linux-yocto/6.12: drop reference recipes
We are going to release the LTS with only 6.18 as the versioned
reference kernel. 6.12 will continue to be updated and used in
previous release branches.
Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The OpenSSL legacy provider supplies algorithms that are either used
infrequently or have been deemed insecure by modern standards. The
Python3 cryptography module can optionally support this provider via the
openssl-ossl-module-legacy package.
Currently, the cryptography module builds with legacy provider support
enabled by default, regardless of whether the legacy modules are
actually included in the system. This patch makes that dependency
explicit by introducing a legacy-openssl packageconfig option that:
* Ensures runtime dependency on openssl-ossl-module-legacy when enabled
* Allows users to disable legacy algorithm support via build configuration
* Aligns python3-cryptography's defaults with OpenSSL's current defaults
The packageconfig option defaults to enabled for consistency with
OpenSSL's current configuration and to avoid breaking existing
deployments. A future security improvement will disable legacy modules
by default in OpenSSL, at which point this packageconfig can also
default to disabled.
Signed-off-by: Colin Pinnell McAllister <colinmca242@gmail.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Max Krummenacher [Mon, 26 Jan 2026 09:36:55 +0000 (10:36 +0100)]
alsa-utils: fix packaging of alsabat
If bat is in PACKAGECONFIG two additional binaries are created which
by the dynamic package split creates two additional packages.
However due to explicitely adding the file alsabat-test.sh to the
alsabat package the alsa-utils-alsabat-test package now is empty
and not created.
Complete commit 4c6d5f96990d ("alsa-utils: fix packaging") which fixed
the same issue for all other affected packages.
Fixes: 26a28ee79350 ("alsa-utils: dynamically split the packages instead of doing it manually") Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Jörg Sommer [Tue, 3 Feb 2026 18:59:54 +0000 (19:59 +0100)]
create-spdx-2.2.bbclass: Add CVE_CHECK_IGNORE to fixed CVEs
The list of CVEs fixed by patches goes to the field *sourceInfo* in the
SBOM. But this list does not contain the CVEs marked for ignoring with the
Bitbake variable *CVE_CHECK_IGNORE*. Many recipes (e.g. openssh, glibc,
python) contain such entries and these are missing in the SBOM. Therefore,
add them to the list.
Signed-off-by: Jörg Sommer <joerg.sommer@navimatix.de> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Peter Marko [Tue, 3 Feb 2026 10:07:13 +0000 (11:07 +0100)]
systemd: properly package new container files
Since last systemd upgrade these files are packaged in systemd package
and thus systemd-container is unconditionally installed in all images as
dependency of systemd package.
Signed-off-by: Peter Marko <peter.marko@siemens.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>