Niko Mauno [Tue, 1 Apr 2025 14:04:27 +0000 (14:04 +0000)]
dbus-glib: Revise RDEPENDS:${PN}-tests assignment
Since this recipe has 'dbus' in DEPENDS, assigning dbus-glib-tests
runtime dependency to '${PN}' instead of 'dbus' will induce pulling
the preferred runtime D-Bus package to rootfs, thanks to libdbus'
runtime recommendation of ${VIRTUAL-RUNTIME_dbus}.
Suggested-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Niko Mauno <niko.mauno@vaisala.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 2 Apr 2025 14:45:37 +0000 (15:45 +0100)]
rust-llvm: Disable libedit
This option currently floats and is leading to non-deterministic builds. It
is used for commandline option tab completion which we don't need in our
builds in general, let alone internally within rust's llvm.
This should fix autobuilder deterministic build issues.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Yash Shinde [Mon, 31 Mar 2025 05:54:21 +0000 (22:54 -0700)]
rust: Upgrade 1.84.1->1.85.0
Rust stable version updated to 1.85.0
https://blog.rust-lang.org/2025/02/20/Rust-1.85.0.html
Some of the major updates:
- Update LIC_FILES_CHKSUM in libstd-rs and rust recipes.
License-Update: Unicode license text is updated to Unicode-3.0 License.
https://github.com/rust-lang/rust/commit/6d2a3e9786ec43a0e0af20386e7046328296ac86
[RP: Update LICENSE to reference Unicode-3.0]
- Pass '-Zforce-unstable-if-unmarked' to RUSTFLAGS in libstd-rs.bb
Fix: https://github.com/rust-lang/rust/issues/133857#issuecomment-2526341227
- Downgrade bootstrap cc version causing bootstrap to fail on custom targets. (Backported from v1.85.1)
Fix: https://github.com/rust-lang/rust/pull/137460/commits/e4ca11f87ffca8c63aa56d45b46e62b6acc58bd7
- Rust v1.85.0 tarball doesn't ship gcc tree.
Drop "remove_gcc_directory" postfunc which removed it and prevented the bloat.
Fix: https://github.com/rust-lang/rust/commit/13c3f9b9498013837782b46120085ea19ca75518
Adapted the patch changes with v1.85.0:
repro-issue-fix-with-cc-crate-hashmap.patch
revert-link-std-statically-in-rustc_driver-feature.patch
rust-oe-selftest.patch
rv32-cargo-rustix-0.38.40-fix.patch
Dropped patches:
fix-tidy-check-failure.patch since it's merged with v1.85.0.
Signed-off-by: Yash Shinde <Yash.Shinde@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Yash Shinde [Mon, 31 Mar 2025 05:54:20 +0000 (22:54 -0700)]
rust: Update "do_update_snapshot" task for rust-snapshot.inc
The 'do_update_snapshot' task is failed with below error:
Exception: FileNotFoundError: [Errno 2] No such file or directory: '.../rustc-1.83.0-src/src/stage0.json'
There are changes in use of key-value format in stage0 file in
rust sources and the rust recipe should be apdated for that.
Niko Mauno [Tue, 1 Apr 2025 14:04:30 +0000 (14:04 +0000)]
recipes-connectivity: Drop dbus from RDEPENDS:${PN}
Since each of these recipes already have 'dbus' in their DEPENDS,
explicit runtime dependency declaration to dbus is not stricly
necessary, because dbus-lib has runtime recommendation for
${VIRTUAL-RUNTIME_dbus} which in turn induces pulling also the
preferred runtime D-Bus package to rootfs.
Suggested-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Niko Mauno <niko.mauno@vaisala.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Niko Mauno [Tue, 1 Apr 2025 14:04:29 +0000 (14:04 +0000)]
glib: Support using alternative runtime D-Bus
In order to facilitate oe-core users who wish to use e.g. dbus-broker
as their runtime D-Bus implementation, change the Freedesktop.org
specific D-Bus runtime dependency to honor the new VIRTUAL-RUNTIME_dbus
variable.
Signed-off-by: Niko Mauno <niko.mauno@vaisala.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Niko Mauno [Tue, 1 Apr 2025 14:04:28 +0000 (14:04 +0000)]
systemd: Support using alternative runtime D-Bus
In order to facilitate oe-core users who wish to use e.g. dbus-broker
as their runtime D-Bus implementation, change the Freedesktop.org
specific D-Bus runtime dependency to honor the new VIRTUAL-RUNTIME_dbus
variable.
Signed-off-by: Niko Mauno <niko.mauno@vaisala.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Niko Mauno [Tue, 1 Apr 2025 14:04:26 +0000 (14:04 +0000)]
dbus: Allow using dbus-lib with alternative D-Bus
In order to facilitate oe-core users who wish to use e.g. dbus-broker
as their runtime D-Bus implementation, declare the target specific
dbus-lib runtime recommendation so that it will honor the new
VIRTUAL-RUNTIME_dbus variable.
Signed-off-by: Niko Mauno <niko.mauno@vaisala.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The purpose of the new variable is to facilitate oe-core users who wish
to use an alternative runtime D-Bus implementation instead of the
default Freedesktop.org's dbus, such as dbus-broker, a recipe for which
is currently available under
meta-openembedded/meta-oe/recipes-core/dbus/
While introducing this facilitation the intent is to preserve the
existing functionality, while allowing the user to optionally select an
alternative runtime D-Bus implementation by adding the following line
e.g. to local.conf file:
VIRTUAL-RUNTIME_dbus = "dbus-broker"
As a background, for example the Fedora distribution uses dbus-broker
instead of Freedesktop.org's D-Bus implementation. The following
excerpts from
https://fedoraproject.org/wiki/Changes/DbusBrokerAsTheDefaultDbusImplementation
provide background for their technological decision
This change provides a more scalable and more reliable implementation in place of the reference implementation.
The reference implementation suffers from long-standing issues including potential dead-locks and susceptibility to denial of service attacks, which the replacement implementation does not.
and
No visible changes in behavior are expected, except for different log messages.
Any visibly different behavior to `dbus-daemon` should be reported as a regression.
Ideally, this change should only improve the performance and security of the message bus.
Signed-off-by: Niko Mauno <niko.mauno@vaisala.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
License-Update: GFDL 1.2 -> 1.3, copyright years update
Add patch that fixes a new documentation generation issue in 1.5.0
https://github.com/xiph/flac/pull/824
Remove the check for NASM, no longer used in 1.5.0
Signed-off-by: Michael Opdenacker <michael.opdenacker@rootcommit.com> CC: Peter Marko <peter.marko@siemens.com> CC: Ross Burton <Ross.Burton@arm.com> CC: Quentin Schulz <quentin.schulz@cherry.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Temporarily disable aics tests as they can fail
depending on how the tests are executed. Sometimes they pass,
sometimes they fail. The issue has been observed since BlueZ 5.72 to 5.80
Starting with BlueZ 5.80, the tests began failing when using the
ptest-runner script. This is not a new issue in BlueZ 5.80 which is
why the test is disabled with this commit until a solution is found.
Signed-off-by: Guðni Már Gilbert <gudni.m.g@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This change will prevent these tools from being installed if the image doesn't install bluez5-noinst-tools package.
BlueZ 5.66: tools/mesh-tester
BlueZ 5.66: tools/ioctl-tester
BlueZ 5.65: tools/iso-tester
BlueZ 5.56: tools/btpclientctl
BlueZ 5.51: tools/bcmfw
BlueZ 5.49: tools/rtlfw
BlueZ 5.47: tools/btconfig (not a new tool, but it was moved from bin_PROGRAMS to noinst_PROGRAMS)
Signed-off-by: Guðni Már Gilbert <gudni.m.g@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changes relevant for the build:
* Four patches are dropped because they are included in 5.80.
* Removed changes to etc/bluetooth directory permission bits. They are
now set correctly in 5.80 [1].
* AMP code has been removed [2]. This means "tools/amptest" can be
removed from NOINST_TOOLS_BT
Fix issue with handling address type for all types of keys.
Fix issue with handling maximum number of GATT channels.
Fix issue with handling MTU auto-tuning feature.
Fix issue with handling AVRCP volume in reconfigured transports.
Fix issue with handling VCP volume setting requests.
Fix issue with handling VCP connection management.
Fix issue with handling MAP qualification.
Fix issue with handling PBAP qualification.
Fix issue with handling BNEP qualification.
Add support for PreferredBearer device property.
Add support for SupportedTypes Message Access property.
Add support for HFP, A2DP, AVRCP, AVCTP and MAP latest versions.
Signed-off-by: Guðni Már Gilbert <gudni.m.g@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Mon, 31 Mar 2025 13:26:24 +0000 (14:26 +0100)]
rust-llvm: Compile llvm to use dynamic libraries
Our main llvm recipe uses dynamic linking already but rust-llvm does not.
Enabling this significanly reduces the size of llvm-rust to about a third
of what it was, which is a bettter configuration for us leading to smaller
TMPDIR and sstate objects.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sat, 29 Mar 2025 14:28:51 +0000 (14:28 +0000)]
rust-source: Fix remove_gcc_directory
The patch works for rust-native but not other recipes which change ${S}.
Set it to use the correct path so it works for llvm-rust, cargo and others,
maximising the space saving.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 28 Mar 2025 22:12:44 +0000 (22:12 +0000)]
buildperf: Tweak tooltip positioning for usability
The tooltop contains a key hyperlink to make the graphs usable. Make
it easier to click on the link by removing the space between the data
point and the tooltip.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 28 Mar 2025 22:11:07 +0000 (22:11 +0000)]
buildperf: Avoid step usage in performance line graphs
When using steps in the line chart, you have to look long and hard
to understand if the data is the lower or upper point on the step.
Whilst not as pretty, the sloped line charts are more accurate so
switch to those as the transition points are important and the main
reason for the charts.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Antonin Godard [Thu, 27 Mar 2025 13:16:52 +0000 (14:16 +0100)]
classes/spdx-*.bbclass: move spdx2 specific variables out of spdx-common class
Setting SPDX_ARCHIVE_SOURCES and SPDX_ARCHIVE_PACKAGED makes no
difference when using the spdx-3.0 class. Move these two in the spdx-2.0
class since they are only used there.
Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chen Qi [Thu, 27 Mar 2025 06:21:43 +0000 (14:21 +0800)]
mc: fix buildpath QA regarding unzip
When unzip-native is somehow introduced indirectly into the dependency
chain, then we get buildpath QA issue. Pass a similar value as the ZIP
to fix this issue.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
rust: Fix build failure when multilibs are enabled
When multilibs are enabled, building rust is failing because of the following commit
https://github.com/rust-lang/rust/commit/68034f837a39387e49fc7d7c5b088f5372a1127e
with stage1/rustc cannot able to find dependent *.so files. The issue is
been fixed inrust-master with the following commit by passing the necessary library
paths before executing stage1/rustc
Rust stable version updated to 1.84.1.
https://blog.rust-lang.org/2025/01/30/Rust-1.84.1.html
Renamed and modified the below patch to adapt the new version.
rv32-cargo-rustix-0.38.37-fix.patch->rv32-cargo-rustix-0.38.38-fix.patch
Modified the below patches to adapt the new version.
repro-issue-fix-with-cc-crate-hashmap.patch
revert-link-std-statically-in-rustc_driver-feature.patch
Remove the "src/gcc" directory from the Rust source code to reduce the
filesystem build space.
In Rust 1.83 (and 1.84), a full GCC-14 tree was included, which caused
the tar.xz archive to increase by 128 MB and the unpacked sources to
grow by 1.3 GB. This was an upstream error that has been resolved in
Rust 1.85.
After updating to Rust 1.85 we can revert this removal of "src/gcc"
Signed-off-by: Deepesh Varatharajan <Deepesh.Varatharajan@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 26 Mar 2025 12:11:50 +0000 (12:11 +0000)]
flex: build occasional build failures in test suite compile
GNU Make 4.4 has added --shuffle, which randomly orders build jobs and
is great at exposing missing dependencies. Using this the known problem
with the flex test suite build was found and resolved, so remove the
workaround of doing a non-parallel build and apply a patch.
Also remove redundant Make variables that shouldn't be needed to build
the test suite.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Pavel Zhukov [Tue, 25 Mar 2025 09:40:14 +0000 (10:40 +0100)]
overlayfs-etc: Unmount /sys and /proc before init
/sys filesystem mounted by the preinit script causes shadowing of
/sys/firmware/efi/ by double /sys mounting on systemd enabled systems
[1]. As the result EFI tooling is broken
[1]
sysfs on /sys type sysfs (rw,nosuid,nodev,noexec,relatime)
efivarfs on /sys/firmware/efi/efivars type efivarfs (rw,nosuid,nodev,noexec,relatime)
sysfs on /sys type sysfs (rw,relatime)
Signed-off-by: Pavel Zhukov <pavel@zhukoff.net> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Mon, 24 Mar 2025 21:24:45 +0000 (21:24 +0000)]
acl: improve ptest packaging
As there's a small number of test binaries in acl, instead of
installing large chunks of the build tree we can install just those and
use a boilerplate test runner.
Drop 0001-tests-do-not-hardcode-the-build-path-into-a-helper-l.patch and
replace with an explicit -DBASEDIR= flag passed at build time.
Drop 0001-test-patch-out-failing-bits.patch and delete the tests that
fail entirely as they won't work without a specific user/group setup.
Backport a patch from upstream so that some tests don't use excessive
amounts of memory.
Backport a patch from upstream to cater for both glibc and musl's
behaviour with interleaved stdout/stderr, fixing the tests on musl.
Clean up dependencies now that we're not shipping the build system.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Mon, 24 Mar 2025 19:50:57 +0000 (12:50 -0700)]
man-db: Add missing rdep for col utility
man utility calls col utility internally when formatting is asked for
therefore it expects col to be in rootfs otherwise silently errors with
retcode 3 meaning 'file not found' in this case its due to col not being
found, other distros eg. gets this via bsdextrautils dependency
Add it via packageconfig and keep is disabled by default since its
deprecated and col does not exist on musl
libssh2 ptest mansyntax.sh fails due to this error, which now works
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Mon, 24 Mar 2025 17:20:07 +0000 (17:20 +0000)]
attr: improve ptest packaging
As there's just a few test binaries in attr, instead of installing large
chunks of the build tree we can install just those and use a boilerplate
test runner.
Also add a comment explaining why we have to sed the test suite if musl
is used.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Mon, 24 Mar 2025 16:33:41 +0000 (09:33 -0700)]
syslinux: Use -idirafter to add back path for system stdarg.h
syslinux uses -nostdinc to build freestanding, which makes sense, however
it also tried to latch its own copy of stdarg.h to include system stdarg.h
via "include_next" compiler magic, so it poses to provide own stdarg.h but
then secretly include system version behind the scenes :)
It uses -nostdinc -iwithprefix include
hoping that gcc is uses and gcc has its include-fixed abstraction which
also contains stdarg.h so in the end it will find a version of stdarg.h
from system (even though it is from the compiler install ) and things will
work.
On musl, include-fixed is not expected and system includes are simplified
so that everyone can look into <sysroot>/usr/include to find them. This
can throw syslinux compilation into problems as now it does not find
the header from -iprefix and ends up with errors like
/mnt/b/yoe/master/sources/poky/build/tmp/work/core2-32-poky-linux-musl/syslinux/6.04-pre2/syslinux-6.04-pre2/com32/lib/../include/stdarg.h:9:15: fatal error: stdarg.h: No such file or directory
9 | #include_next <stdarg.h>
| ^~~~~~~~~~
compilation terminated.
Therefore, we use -idirafter to point it into target sysroot as fallback
for system headers if it needs them, its added at the very last in search
order. It also keeps working with glibc based toolchains as usual and
also works with musl toolchains.
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Mon, 24 Mar 2025 16:33:35 +0000 (09:33 -0700)]
ghostscript: Pin to C17 std
The code defines a custom 'bool' type (as an 'int'), which is incompatible
with C23 in which bool is a keyword, and trying to use <stdbool.h> fails
because 'int' and 'bool' are used interchangeably in the code.
Add the flag to CC variable, since CFLAGS is used by both c and c++ compilers
and clang++ is less forgiving when C compiler only option is used on its
cmdline so it complains about -std=gnu17 and bails out.
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Tue, 7 Jan 2025 12:07:43 +0000 (12:07 +0000)]
meta/conf: Start to populate autobuilder config fragments
This populates the fragments directory with:
a) the default set of variables used in all autobuilder builds
b) the default resource related variables used in all autobuilder builds
c) three different multilib test configurations used by the autobuilder
The aim here is to start to make some of the autobuilder configuration more
visable and patchable by users, and to allow some test confiturations to
be user selectable if appropriate and needed for debugging.
The main aautobuilder fragment is probably not directly reusable by most
users, it contains the resource limits as used on the autobuilder itself. I
can see arguments both way for whether this should be included in OE-Core or not
but having an example of how we configure this is probably useful.
Not all configuration in the autobuilder is being moved, this set of variables
is just a basic starting point. Some variables may ultimately make more sense
being migrated elsewhere, perhaps updating the main defaults for poky or nodistro.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Quentin Schulz [Tue, 25 Mar 2025 09:07:54 +0000 (10:07 +0100)]
b4-wrapper-poky.py: output errors to stderr
Print error messages to stderr instead of stdout.
The commands run as part of send-auto-cc-cmd and send-auto-to-cmd b4
commands will make b4 raise an Exception if they return a non-zero
return code and it will only print the content of stderr before that.
Because print defaults to stdout, b4 would raise the exception and not
tell the user why. This commit should now provide a hint to the user
about what went wrong.
Yogita Urade [Tue, 25 Mar 2025 09:07:33 +0000 (09:07 +0000)]
libarchive: upgrade 3.7.7 -> 3.7.8
This upgrade includes fix for CVE-2024-57970, CVE-2025-25724 and
CVE-2025-1632
Changelog:
==========
Libarchive 3.7.8 is a bugfix and security release
Security fixes:
tar reader: Handle truncation in the middle of a GNU long linkname (#2422, CVE-2024-57970)
unzip: fix null pointer dereference (#2532, CVE-2025-1632)
tar reader: fix unchecked return value in list_item_verbose() (#2532, CVE-2025-25724)
Important bugfixes:
7zip reader: add SPARC (#2399) and POWERPC (#2459) filter support for non-LZMA compressors
tar reader: Ignore ustar size when pax size is present (#2405)
tar writer: Fix bug when -s/a/b/ used more than once with b flag (#2435)
cpio: Fix a Y2038 bug on Windows (#2471)
libarchive: Handle ARCHIVE_FILTER_LZOP in archive_read_append_filter (#2519)
libarchive: Adding missing seeker function to archive_read_open_FILE() (#2539)
Signed-off-by: Yogita Urade <yogita.urade@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Martin Jansa [Thu, 20 Mar 2025 16:55:16 +0000 (17:55 +0100)]
cargo.bbclass: show PACKAGECONFIG_CONFARGS in bbnote
* PACKAGECONFIG_CONFARGS was added in:
https://git.openembedded.org/openembedded-core/commit/?id=16745b20452de60ae2474433cc1a2fb1ed9f6a64
but it wasn't added in bbnote above which might lead to confusing errors like I got now:
* it was from recipe where we already use:
RUSTFLAGS:append = " ${PACKAGECONFIG_CONFARGS}"
it will be difficult to use PACKAGECONFIG for RUSTFLAGS and prevent
them to be used here for cargo as well, what about the recipes which
need them to explicitly append them to CARGO_BUILD_FLAGS ?
Signed-off-by: Martin Jansa <martin.jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
projects / thirdparty / openembedded / openembedded-core-contrib.git / log
