Instead of using the setuptoools class but not using setuptools, use the
new python_pdm class.
(From OE-Core rev: 70510190463df42db0424d44ebf96d42f4ba4add)
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 package uses flit, not setuptools.
Also there's no need for it to depend on python3-cryptography.
(From OE-Core rev: 926935eb872408da733103c1d632a901acfd6e69)
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>
There's no need to pull in python_setuptools_build_meta as maturin is a
build system of its own.
(From OE-Core rev: ea639b83816ae89c0da0a0695ac80f4d678b33db)
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>
For the build process to actually work some inherits are always needed:
- python3native for the python that is actually running build.
- setuptools3-base (badly named) defines how to build extensions and
where to put modules.
Instead of inheriting these in every class, inherit them once in the
python_pep517 class.
(From OE-Core rev: 449b226aca6a5db378dc5576dde47499925dfb26)
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>
When crypto is not in tune features then add +nocryto to
-mcpu explicitly. This makes the behavior between clang
and gcc match. Currently -mcpu=cortex-a72 has different
behavior in clang and gcc in terms of what features are
considered default. Clang enables different set of common
features than gcc on other hand. For example clang
enables crypto with default set but gcc
does not, gcc recommends to disable unavailable extensions
in -mcpu [1] explicitly. crypto is optional on cortex-a53
and cortex-a72. This is not as common but Broadcom
SOCs in raspberrypi3/4 have dropped crypto for cost
reasons [2]. This results in illegal instruction
traps [3] [4] when building components e.g. chromium,
qtwebengine, weston etc. with clang using -mcpu=cortex-a72
for rpi4 target.
Adding +nocrypto makes clang behave like gcc does today. We
do have separate tune if crypto enabled cortex-a72 cores
are to be targeted (cortexa72-cryto) as DEFAULTTUNE
They are added to default feature file since crypto extension
is available in multiple arm architecture versions e.g. armv8,
armv9. It is optional extension as per spec [5]
Extensions can be enabled and disabled with -mcpu using the same
syntax as with -march, and have same effect thats why it is
intrumented via TUNE_CCARGS_MARCH_OPTS
[1] https://gcc.gnu.org/onlinedocs/gcc/ARM-Options.html#index-mcpu-2
[2] https://forums.raspberrypi.com/viewtopic.php?f=63&t=207888#p1332960
[3] https://github.com/llvm/llvm-project/issues/85699
[4] https://github.com/llvm/llvm-project/issues/90365
[5] https://developer.arm.com/documentation/109697/2025_06/Feature-descriptions/The-Armv9-0-architecture-extension
(From OE-Core rev: db1b355b2b15ba57bd89c2dfb88c2c667551863e)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Quentin Schulz <quentin.schulz@cherry.de>
Acked-by: Quentin Schulz <quentin.schulz@cherry.de>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When parsing failures occur, trap KeyError to avoid these kinds
of tracebacks (from a parsing error in meta-ti).
bb.parse.ParseError: ParseError at /srv/pokybuild/yocto-worker/check-layer-nightly/build/meta-ti/meta-ti-bsp/recipes-kernel/linux/linux-ti-mainline_git.bb:6: Could not inherit file classes/${KERNEL_BASE_CLASS}.bbclass
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/srv/pokybuild/yocto-worker/check-layer-nightly/build/lib/bb/command.py", line 123, in runAsyncCommand
self.cooker.updateCache()
File "/srv/pokybuild/yocto-worker/check-layer-nightly/build/lib/bb/cooker.py", line 1639, in updateCache
if not self.parser.parse_next():
File "/srv/pokybuild/yocto-worker/check-layer-nightly/build/lib/bb/cooker.py", line 2314, in parse_next
self.shutdown(clean=False, eventmsg=str(exc))
File "/srv/pokybuild/yocto-worker/check-layer-nightly/build/lib/bb/cooker.py", line 2209, in shutdown
read_results()
File "/srv/pokybuild/yocto-worker/check-layer-nightly/build/lib/bb/cooker.py", line 2195, in read_results
self.result_queue.get(timeout=0.25)
File "/usr/lib/python3.10/multiprocessing/queues.py", line 122, in get
return _ForkingPickler.loads(res)
File "/srv/pokybuild/yocto-worker/check-layer-nightly/build/lib/bb/cache.py", line 338, in __setstate__
setattr(self, key, self._restore(state[key], pid))
File "/srv/pokybuild/yocto-worker/check-layer-nightly/build/lib/bb/cache.py", line 318, in _restore
ret[dep] = map[mapnum]
KeyError: 156
(Bitbake rev: 750c68ee7ee3f3d4518348e4c948243504880770)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There was duplication between qa-checks.rst and the doc for the insane
class, so move all of the QA checks in this qa-checks.rst. Remove the
documentation for these checks from classes.rst and link to the
qa-checks.rst document from there.
For QA checks that were only documented only in classes.rst, add them to
qa-checks.rst.
Also sort qa-checks by their names in qa-checks.rst.
(From yocto-docs rev: dc984e5f54c38748603243053fde8b4d078ba586)
Signed-off-by: Antonin Godard <antonin.godard@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This was dropped by commit 452e605b55ad ("cve-check: annotate CVEs
during analysis").
(From yocto-docs rev: 55aedec94dce3e3c0541ffaea666cff02d63a708)
Signed-off-by: Antonin Godard <antonin.godard@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Document the new classes under classes/toolchain as well as
PREFERRED_TOOLCHAIN* and TOOLCHAIN variables, which allow selecting the
toolchain. For now there's "gcc" and "clang" as available toolchain.
(From yocto-docs rev: 69a9568f8bd7ed57efddab507a4294ad3408f4dd)
Signed-off-by: Antonin Godard <antonin.godard@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The tar packaging format was removed some time ago. Also, add some
minor grammatical tweaking.
(From yocto-docs rev: 8e9fd1ca232f3c1e8be51cb881a68b4745ee548a)
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Antonin Godard <antonin.godard@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
After compact vsdisk you have to detach it before exiting
otherwise the vdisk remains attached.
DISKPART> select vdisk file="<path_to_VHDX_file>"
DISKPART> attach vdisk readonly
DISKPART> compact vdisk
DISKPART> detach <------------ new missing command
DISKPART> exit
(From yocto-docs rev: 1cc65ddf1a074f61fe5a63d222f3079b7fcb4c1e)
Signed-off-by: Marco Cavallini <m.cavallini@koansoftware.com>
Signed-off-by: Antonin Godard <antonin.godard@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Quoting Alexander Kanavin [1]:
> 1. BB_HASHSERVE and BB_SIGNATURE_HANDLER do not set up or start a
> server, they set up the hash equivalence client built into bitbake.
>
> 2. The above client configuration (used by poky) also starts a private
> server that is specific to a particular build directory. So hash equiv
> information would not be shared between multiple build directories,
> and this will cause sstate mismatches (as reported in the bug), if
> sstate is shared.
This setup does not shed light on the potential of the hash equivalence
feature. So for now, remove this basic setup, and later rework the
concepts (or create a new) document that explains how to set up a hash
equivalence server shared between builds.
[1]: https://lore.kernel.org/r/CANNYZj_kvLsZG3sgH+nPu9B=pqWBU785w0SGHGdQqB4UW-DtmA@mail.gmail.com
Suggested-by: Alexander Kanavin <alex.kanavin@gmail.com>
(From yocto-docs rev: 231fc9e710ab34db60263f8ae01d4f5970579203)
Signed-off-by: Antonin Godard <antonin.godard@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Provide libunwind based on TC_CXX_RUNTIME
When TC_CXX_RUNTIME is set to be 'llvm' distro wants
to use llvm provided runtime elements therefore provide
libunwind when TC_CXX_RUNTIME = "llvm"
(From OE-Core rev: 61cf402d5bf95094a02d4ea3b1a60cecc04aa0f4)
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>
After upgrading bash to 5.3 to contain commit [1], if SIGINT was trapped in
asynchronous subshell, bash called set_signal_async_ignored to make sure
processes that are created by this asynchronous subshell to ignore SIGINT.
And it caused test case `kill/decode' failed, the signal INT was existed in
both of Blocked and Ignored
root@qemux86-64:~# cd /usr/lib/util-linux/ptest/
root@qemux86-64:/usr/lib/util-linux/ptest# ./run-ptest
...
diff-{{{
|--- /usr/lib/util-linux/ptest/tests/expected/kill/decode 2011-04-05 23:00:00.000000000 +0000
|+++ /usr/lib/util-linux/ptest/tests/output/kill/decode 2025-07-28 12:09:46.279000000 +0000
@@ -16,5 +16,5 @@
Pending (thread): INT ILL
Pending (process): USR1
Blocked: INT ILL ABRT FPE USR1 SEGV TERM
-Ignored: HUP QUIT TRAP PIPE ALRM
+Ignored: HUP INT QUIT TRAP PIPE ALRM
Caught: ILL USR1
}}}-diff
...
Explicitly reset SIGINT to default action after trapping, then the
test case `kill/decode' works on both of old (<5.3) and new (>=5.3) bash
[1] https://cgit.git.savannah.gnu.org/cgit/bash.git/commit/?h=devel&id=a5d2617c7a7e602ace1f4149987cdfd075c4e762
(From OE-Core rev: 175f5a5bba6204cbcfbdeabb6c6b98f0f7453558)
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It's not currently possible to set the build tree to be somewhere we
control, but we know it will always be in the build directory alongside
the pyproject.toml so we can [cleandirs] that.
(From OE-Core rev: 383862cfe4c5acf04124080827c8bc6d00b2e86d)
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>
It's not currently possible to set the build tree to be somewhere we
control, but we know it will always be in the build directory alongside
the setup.py so we can [cleandirs] that.
(From OE-Core rev: f3854f4f60801e3b6788bee3a0a1850fc498d536)
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>
Explicitly call the build command, as by doing this we get to pass -j.
In packages without any C extensions this is mostly a no-op as
bdist_wheel will call it anyway. However, this does make a big
difference to build time for packages with non-trivial C extensions:
python3-cython:do_compile 204.8s -> 70.9s
python3-lxml:do_compile 157.5s -> 92.4s
(From OE-Core rev: 96d13feaa881ae4d275d16f7efca92dbd075241b)
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>
Copy the pyproject.toml check from setuptools3.bbclass so that the legacy
class will also warn if there's a modern build framework available.
(From OE-Core rev: f524589f3d35475600954f5b27b4b74ea079e13c)
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 do builds in a separate directory in this class, so add it to cleandirs
to ensure that it is empty.
(From OE-Core rev: 2575adeceedae72f6359c0a35ec5c5325a4ec363)
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 the CONFIGURE_FILES assignment from python_mesonpy to the common
class, as it isn't specific to mesonpy.
Also extend, so that it doesn't clobber existing settings.
(From OE-Core rev: 049cd6b62853c020b0c85ccee18638ec58887866)
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>
The uninative tarball requires xz to decompress:
WARNING: Disabling uninative as unable to install uninative tarball:
WARNING: tar (child): xz: Cannot exec: No such file or directory
(From OE-Core rev: 9c0bc6edf5e6706597563924f62bbab6fdd38ac2)
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>
* Fix e2fsprogs-native do_compile error with old kernel such as 5.10
(Debian 11):
../../sources/e2fsprogs-1.47.2/debugfs/../misc/create_inode.c:600:20: error: FS_IOC_READ_VERITY_METADATA undeclared (first use in this function)
* And its version should be 1.47.3, not 1.47.2.
(From OE-Core rev: 52cd80f2c88e298d5538a9e426e681b288f39244)
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>
The rk-usb-loader is already enabled in the hosttools_defconfig and
targettools_defconfig and compiled by the recipe.
Add it to the BAREBOX_TOOLS list to actually install it.
(From OE-Core rev: f5e182589b46ebd862721d5e24005d56cdddbe07)
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Rust stable version updated to 1.88.0.
https://blog.rust-lang.org/2025/06/26/Rust-1.88.0/
* Rebase existing patches with v1.88.0
* Exclude tidy and coverage-dump from OE self-tests due to test failures
Although the build completes successfully, OE self-tests fail:
tools/coverage-dump panics during test execution.
tools/tidy fails due to a mismatched GCC submodule commit.
These tests are excluded to allow successful OE self-test runs.
* Two tests from the`codegen` modules now fail only on riscv64.
Enable them on arm32/64 and x86-32/64 targets, while restricting
them on riscv64 via `only-<target_arch>` tags.
Bugzilla link - https://bugzilla.yoctoproject.org/show_bug.cgi?id=15944
Test Results Summary:
+-----------+--------+---------+
| Machine | Passed | Ignored |
+-----------+--------+---------+
| arm-32 | 28,664 | 1,451 |
| arm-64 | 28,748 | 1,396 |
| x86-32 | 28,657 | 1,432 |
| x86-64 | 28,904 | 1,213 |
| riscv-64 | 28,722 | 1,421 |
+-----------+--------+---------+
(From OE-Core rev: 5afc9bdbabfdbcb784ceb637926111e6c7648f41)
Signed-off-by: Jayasurya Maganuru <Maganuru.Jayasurya@windriver.com>
Signed-off-by: Deepesh Varatharajan <Deepesh.Varatharajan@windriver.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
`exit_if_errors` must be called explicitly at the end of the task
execution to ensure that QA issues categorized as errors fail the task.
(From OE-Core rev: acc0c57e455670501433ba4514a977e16dc88c34)
Signed-off-by: Philip Lorenz <philip.lorenz@bmw.de>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The variables checked for are mandatory in all cases so a QA check does
not seem to be the right choice for this. Additionally, the code never
called `exit_if_errors` so the reported QA issue didn't actually fail
the task.
(From OE-Core rev: cb69d7eba80215f9f8b74dafe40ca2289fa294aa)
Signed-off-by: Philip Lorenz <philip.lorenz@bmw.de>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fixes [YOCTO 15917]
When fetching a git repository the .gitattributes file is scanned, checking if LFS
support is required for the repository. This scan checks if the word "lfs" is present
in the file, however the used regex doesn't account for comments, which makes some
repositories[1] be to misidentified as requiring LFS support (which fails fetching, in case
lfs support isn't installed on the build host).
To avoid it, change the used regex to ignore lines starting with "#".
[1]: e.g. https://github.com/MicrochipTech/cryptoauthlib
(Bitbake rev: 7917a758fc328747116c7899e689171bd0efc883)
Signed-off-by: Gyorgy Sarvari <skandigraun@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
