]> git.ipfire.org Git - thirdparty/openembedded/openembedded-core.git/log
thirdparty/openembedded/openembedded-core.git
9 hours agoovmf: Fix build with GCC 16 and GLIBC 2.43 master
Zoltán Böszörményi [Thu, 9 Apr 2026 12:22:25 +0000 (14:22 +0200)] 
ovmf: Fix build with GCC 16 and GLIBC 2.43

Backport two patches from master to fix native build with
GCC 16 and GLIBC 2.43

commit 3597306191297b504683b83fe7750e49c6a2e836
Author: Gerd Hoffmann <kraxel@redhat.com>
Date:   Thu Jan 29 09:23:32 2026 +0100

    BaseTools/StringFuncs: fix gcc 16 warning

    StringFuncs.c: In function ‘SplitStringByWhitespace’:
    StringFuncs.c:113:15: error: variable ‘Item’ set but not used [-Werror=unused-but-set-variable=]
      113 |   UINTN       Item;
          |               ^~~~

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Author: Gerd Hoffmann <kraxel@redhat.com>
Date:   Mon Dec 8 10:28:50 2025 +0100

    BaseTools/EfiRom: fix compiler warning

    New warning after updating gcc:

    EfiRom.c: In function ‘main’:
    EfiRom.c:78:17: error: assignment discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]

    The assigned value is not used, so fix the warning by just removing it.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Zoltán Böszörményi <zboszor@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agovulkan-samples: Upgrade to latest tip of trunk
Khem Raj [Sat, 11 Apr 2026 00:29:29 +0000 (17:29 -0700)] 
vulkan-samples: Upgrade to latest tip of trunk

- A new sample demonstrating the rasterization order attachment access extension (#1492).
- Broken documentation link to GitHub repo fixed (#1506).
- Vulkan SDK updated to 1.4.341 (#1477).

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agowebkitgtk: Remove escaping of the plus sign in `CVE_PRODUCT`
Benjamin Robin [Fri, 10 Apr 2026 13:10:46 +0000 (15:10 +0200)] 
webkitgtk: Remove escaping of the plus sign in `CVE_PRODUCT`

The `+` character is now properly escaped by `cve_check.cpe_escape()`

Signed-off-by: Benjamin Robin <benjamin.robin@bootlin.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agogtk+: Remove escaping of the plus sign in `CVE_PRODUCT`
Benjamin Robin [Fri, 10 Apr 2026 13:10:45 +0000 (15:10 +0200)] 
gtk+: Remove escaping of the plus sign in `CVE_PRODUCT`

The `+` character is now properly escaped by `cve_check.cpe_escape()`

Signed-off-by: Benjamin Robin <benjamin.robin@bootlin.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agocve_check: do not break old CVE_PRODUCT with escaped +
Benjamin Robin [Fri, 10 Apr 2026 13:10:44 +0000 (15:10 +0200)] 
cve_check: do not break old CVE_PRODUCT with escaped +

For now, until all layer are fixed, replace already escaped plus (+)
with a simple + before doing the escaping.

Signed-off-by: Benjamin Robin <benjamin.robin@bootlin.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agocve_check: Improve escaping of special characters in CPE 2.3
Benjamin Robin [Fri, 10 Apr 2026 13:10:43 +0000 (15:10 +0200)] 
cve_check: Improve escaping of special characters in CPE 2.3

According to the NISTIR 7695 specification [1], multiple characters
require escaping when using formatted strings (e.g., `cpe:2.3:...`),
which use backslash escaping. In "Figure 6-3. ABNF for Formatted String
Binding"", the characters that need escaping are referenced by "escape",
"special", and "punc". More characters must be escaped than just
`\`, `?`, `*`, `:`, and `+`.

Additionally, use `maketrans()` with `translate()`, which is more
efficient than a simple `replace()`.

[1] https://nvlpubs.nist.gov/nistpubs/legacy/ir/nistir7695.pdf

Signed-off-by: Benjamin Robin <benjamin.robin@bootlin.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agoxserver-org: update CVE_PRODUCT
Peter Marko [Sun, 12 Apr 2026 18:52:00 +0000 (20:52 +0200)] 
xserver-org: update CVE_PRODUCT

In cvelistV5, CVE-2024-21886 uses CPE xorg:xserver ([1]).
Detected because this CVE is shown for xwayland recipe.

[1] https://github.com/CVEProject/cvelistV5/blob/cve_2026-04-12_1800Z/cves/2024/21xxx/CVE-2024-21886.json#L646

Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agolibpng: upgrade 1.6.55 -> 1.6.56
Peter Marko [Sun, 12 Apr 2026 15:03:23 +0000 (17:03 +0200)] 
libpng: upgrade 1.6.55 -> 1.6.56

Release notes [1]:
 * Fixed CVE-2026-33416 (high severity):
   Use-after-free via pointer aliasing in `png_set_tRNS` and `png_set_PLTE`.
   (Reported by Halil Oktay and Ryo Shimada;
   fixed by Halil Oktay and Cosmin Truta.)
 * Fixed CVE-2026-33636 (high severity):
   Out-of-bounds read/write in the palette expansion on ARM Neon.
   (Reported by Taegu Ha; fixed by Taegu Ha and Cosmin Truta.)
 * Fixed uninitialized reads beyond `num_trans` in `trans_alpha` buffers.
   (Contributed by Halil Oktay.)
 * Fixed stale `info_ptr->palette` after in-place gamma and background
   transforms.
 * Fixed wrong channel indices in `png_image_read_and_map` RGB_ALPHA path.
   (Contributed by Yuelin Wang.)
 * Fixed wrong background color in colormap read.
   (Contributed by Yuelin Wang.)
 * Fixed dead loop in sPLT write.
   (Contributed by Yuelin Wang.)
 * Added missing null pointer checks in four public API functions.
   (Contributed by Yuelin Wang.)
 * Validated shift bit depths in `png_set_shift` to prevent infinite loop.
   (Contributed by Yuelin Wang.)
 * Avoided undefined behavior in library and tests.
 * Deprecated the hardly-ever-tested POINTER_INDEXING config option.
 * Added negative-stride test coverage for the simplified API.
 * Fixed memory leaks and API misuse in oss-fuzz.
   (Contributed by Owen Sanzas.)
 * Implemented various fixes and improvements in oss-fuzz.
   (Contributed by Bob Friesenhahn and Philippe Antoine.)
 * Performed various refactorings and cleanups.

[1] https://github.com/pnggroup/libpng/blob/v1.6.56/ANNOUNCE

Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agoapr: remove -std=gnu23 from apr_rules.mk to fix reproducibility
Martin Jansa [Sun, 12 Apr 2026 10:57:50 +0000 (12:57 +0200)] 
apr: remove -std=gnu23 from apr_rules.mk to fix reproducibility

With the updated ax_prog_cc_for_build.m4 from
apr: fix build with autoconf-2.73
it now correctly detects -std=gnu23 usability on host gcc and doesn't
try to use it e.g. with gcc-13 which doesn't support it, but that causes
build reproducibility issue between builds on such hosts as
CC_FOR_BUILD is included in installed /usr/share/build-1/apr_rules.mk
https://valkyrie.yocto.io/pub/repro-fail/oe-reproducible-20260412-a5w4mcrk/packages/diff-html/

Signed-off-by: Martin Jansa <martin.jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agotime: Use upstream applied patches
Khem Raj [Wed, 8 Apr 2026 21:43:33 +0000 (14:43 -0700)] 
time: Use upstream applied patches

time project now has a maintainer and has taken bunch of
patches submitted over years, backport the needed ones for
compiling with GCC-15 and Clang.

Current patchset was not adequate for clang and moreover
the applied patches are slightly different than what we are
carrying

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agouseradd.bbclass: allow inheriting with only USERADD_DEPENDS set
Zk47T [Wed, 8 Apr 2026 15:00:02 +0000 (22:00 +0700)] 
useradd.bbclass: allow inheriting with only USERADD_DEPENDS set

When a recipe inherits useradd and only sets USERADD_DEPENDS (to depend
on users/groups created by another recipe), without creating any
users/groups itself, the parse-time sanity check incorrectly throws a
fatal error about missing USERADD_PACKAGES.

Skip the USERADD_PACKAGES/USERADD_PARAM validation when USERADD_DEPENDS
is set but USERADD_PACKAGES is not, since the recipe only needs build
dependency tracking, not user/group creation.

Fixes [YOCTO #15863]

Signed-off-by: Zk47T <zizuzacker@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agobinutils: Set status for CVE-2025-69649
Harish Sadineni [Wed, 8 Apr 2026 12:15:50 +0000 (05:15 -0700)] 
binutils: Set status for CVE-2025-69649

Set CVE_STATUS for CVE-2025-69649, as this CVE already fixed
with binutils 2.46 version update.

According to the NVD reference [1], the issue is addressed by the upstream commit:
66a3492ce68e1ae45b2489bd9a815c39ea5d7f66

This fix is included in binutils v2.46

[1] https://nvd.nist.gov/vuln/detail/CVE-2025-69649

Signed-off-by: Harish Sadineni <Harish.Sadineni@windriver.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agorsync: Add packageconfig to use system zlib
Khem Raj [Wed, 8 Apr 2026 07:43:41 +0000 (00:43 -0700)] 
rsync: Add packageconfig to use system zlib

Enable using system zlib by default, avoid copies
and duplicate builds

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agogstreamer1.0-plugins-bad: Add fdkaac as a package feature
Peter Kjellerstedt [Wed, 8 Apr 2026 02:24:03 +0000 (04:24 +0200)] 
gstreamer1.0-plugins-bad: Add fdkaac as a package feature

This adds support for building gstfdkaac, which is the GStreamer element
that wraps the fdk-aac AAC encoding library from Fraunhofer. fdk-aac is
available in meta-multimedia.

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agopython3-kirk: upgrade 4.0.0 -> 4.1.0
Daniel Turull [Tue, 7 Apr 2026 14:08:24 +0000 (16:08 +0200)] 
python3-kirk: upgrade 4.0.0 -> 4.1.0

Signed-off-by: Daniel Turull <daniel.turull@ericsson.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agoweston: backport gl-shaders fix to avoid shader compiler crashes
Mahesh Angadi [Tue, 7 Apr 2026 09:47:59 +0000 (15:17 +0530)] 
weston: backport gl-shaders fix to avoid shader compiler crashes

Some GPU shader compilers do not optimize away unused YUV shader
uniforms, even when the associated code paths are unreachable. This can cause
Weston to hit assertion failures at runtime on non-Mesa platforms, despite
working correctly on Mesa-based drivers.

Backport the upstream gl-shaders fix to restore stable behavior across
GPU drivers.

Signed-off-by: Mahesh Angadi <mangadi@qti.qualcomm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agodevtool: ide-sdk: use TOOLCHAIN not TCOVERRIDE
Adrian Freihofer [Mon, 6 Apr 2026 22:10:33 +0000 (00:10 +0200)] 
devtool: ide-sdk: use TOOLCHAIN not TCOVERRIDE

Looks like TOOLCHAIN is the correct variable to determine the toolchain
used by a recipe, not TCOVERRIDE.

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agooe-selftest/cpp-example: fix conf file ownership with static UIDs/GIDs
Adrian Freihofer [Mon, 6 Apr 2026 22:10:32 +0000 (00:10 +0200)] 
oe-selftest/cpp-example: fix conf file ownership with static UIDs/GIDs

test_devtool_ide_sdk_none_qemu builds an image containing both
cmake-example and meson-example, starts a QEMU instance, then uses
devtool ide-sdk + devtool deploy-target to rebuild and redeploy each
recipe in turn. The test verifies that /etc/<recipe>.conf is owned by
the matching user both before and after each deploy cycle.

The test was failing with:

  /etc/meson-example.conf not owned by user meson-example: got cmake-example

Root cause: both recipes call

  install -m 0644 -o ${BPN} -g ${BPN} ... ${D}${sysconfdir}/${BPN}.conf

During do_install, pseudo resolves ${BPN} to a UID by looking up
/etc/passwd in the recipe's own isolated RECIPE_SYSROOT. Since the
sysroots are independent, both cmake-example and meson-example each
see themselves as the first --system user and get the same UID (e.g.
100). Both ${D} trees therefore contain files with UID 100. In the
final rootfs cmake-example is allocated UID 100 and meson-example UID
101. Files packaged for meson-example still carry UID 100, so stat
reports them as owned by cmake-example.

A pkg_postinst chown would fix the rootfs, but devtool deploy-target
is a plain tar pipe over SSH with no package-manager involvement - it
never runs pkg_postinst. Whatever UID is embedded in ${D} is what
lands on the target. Not sure how this could be fixed with dynamic UIDs.

A clean solution is to make every recipe sysroot and the final image
agree on the same UIDs from the start, i.e. static IDs.

Fix:
- Enable USERADDEXTENSION = "useradd-staticids" in _write_bb_config so
  the test builds with static IDs for the duration of the test.
- Add cmake-example (UID/GID 533) and meson-example (UID/GID 534) to
  meta-selftest/files/static-passwd and static-group.
- Expand the comment in cpp-example.inc's do_install to document the
  static-ID requirement so future readers understand why the -o/-g
  flags work correctly only under useradd-staticids.
- Fix a copy-paste error in the in-test comment (said
  "meson-example.conf ... cmake-example user" for the cmake block).

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agooe-selftest: devtool: use assertRegex to match test output for meson
Adrian Freihofer [Mon, 6 Apr 2026 22:10:31 +0000 (00:10 +0200)] 
oe-selftest: devtool: use assertRegex to match test output for meson

Replace strict string matching with assertRegex to allow for flexible
whitespace in the "Fail: 0" output from meson tests. This improves test
robustness against formatting changes.
This issue was discovered with scarthgap.

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agooe-selftest: devtool: GDB breakpoint after std::vector is constructed
Adrian Freihofer [Mon, 6 Apr 2026 22:10:30 +0000 (00:10 +0200)] 
oe-selftest: devtool: GDB breakpoint after std::vector is constructed

Change the GDB breakpoint from line 55 to 56 in cpp-example.cpp so that
the std::vector constructor has already executed when GDB stops. This
ensures that inspecting the vector with GDB works as intended also with
older GDB versions (e.g. on scarthgap).

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
13 hours agopackage: update the comment block explaining 'emit_pkgdata'
Adam Blank [Thu, 2 Apr 2026 15:39:40 +0000 (17:39 +0200)] 
package: update the comment block explaining 'emit_pkgdata'

The comment block mentioned PKGDATA_DIR which is now used
differently.

Signed-off-by: Adam Blank <adam.blank.g@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 days agostrace: Fix build with std=gnu23
Khem Raj [Sat, 11 Apr 2026 00:29:28 +0000 (17:29 -0700)] 
strace: Fix build with std=gnu23

This does not build on hosts with gcc <= 13 which does not have std=gnu23
dialect option e.g. ubuntu 24.04, especially with clang which does not
default to gnu23 yet like gcc-15, so autoconf 2.73 decided to enforce the
standard with clang which then spills into CC_FOR_BUILD as well without
this patch

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 days agobind: Fix build with std=gnu23
Khem Raj [Sat, 11 Apr 2026 00:29:27 +0000 (17:29 -0700)] 
bind: Fix build with std=gnu23

This does not build on hosts with gcc <= 13 which does not have std=gnu23
dialect option e.g. ubuntu 24.04, especially with clang which does not
default to gnu23 yet like gcc-15, so autoconf 2.73 decided to enforce the
standard with clang which then spills into CC_FOR_BUILD as well without
this patch

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 days agosudo: Fix build with std=gnu23
Khem Raj [Sat, 11 Apr 2026 00:29:26 +0000 (17:29 -0700)] 
sudo: Fix build with std=gnu23

This does not build on hosts with gcc <= 13 which does not have std=gnu23
dialect option e.g. ubuntu 24.04, especially with clang which does not
default to gnu23 yet like gcc-15, so autoconf 2.73 decided to enforce the
standard with clang which then spills into CC_FOR_BUILD as well without
this patch

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 days agoflex: Fix build with autoconf-2.74 and gcc <= 13
Khem Raj [Fri, 10 Apr 2026 17:01:17 +0000 (10:01 -0700)] 
flex: Fix build with autoconf-2.74 and gcc <= 13

GCC 13 or below does not support -std=gnu23 but autoconf enforces
it on compilers which do not have C23 turned on by default

Update the autconf m4 file instead of passing cmdline option

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 days agoapr: fix build with autoconf-2.73
Martin Jansa [Fri, 10 Apr 2026 13:49:14 +0000 (15:49 +0200)] 
apr: fix build with autoconf-2.73

On hosts with gcc-13 it was trying to use -std=gnu23 and failing
with unrecognized command-line option (gnu23 needs gcc-14 and newer)

Signed-off-by: Martin Jansa <martin.jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
2 days agodb: Pin to use C99 std
Khem Raj [Wed, 8 Apr 2026 16:27:59 +0000 (09:27 -0700)] 
db: Pin to use C99 std

GCC-15 has switched to using C23 by default, we have been selectively
disabling warnings as errors to get by, however with autoconf 2.72 adding
-std=gnu23 it now gets enabled for every compiler and clang-22 is more
obidient and has dropped support for K&R C completely. db5 code has a lot
of K&R C prototypes and it starts to fail vigorously. We can not keep
working around with out uplifting sources to be compliant with newer C
standard like C23.

Therefore pin the cflags to use C99 standard as this package expects.

Drop the code to disable warnings as errors selectively instead add a
fix for addressing implicit int warnings

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
3 days agopython3: upgrade 3.14.3 -> 3.14.4
Trevor Gamblin [Thu, 9 Apr 2026 17:38:26 +0000 (13:38 -0400)] 
python3: upgrade 3.14.3 -> 3.14.4

Changelog: https://docs.python.org/3/whatsnew/changelog.html#python-3-14-4-final

Drop patches:

- 0001-gh-144787-tests-Allow-TLS-v1.2-to-be-minimum-version.patch since
  it was a backport and is now included in 3.14.4.
- 0001-test_pyrepl-skip-test_pyrepl_eio.patch, the test for which was
  determined flaky and dropped upstream in commit 9743d88334ad.

ptests look OK:

|Ran 353 tests in 74.012s
|
|OK (skipped=43)
|
|== Tests result: SUCCESS ==
|
|28 tests skipped:
|    test.test_asyncio.test_windows_events
|    test.test_asyncio.test_windows_utils test.test_gdb.test_backtrace
|    test.test_gdb.test_cfunction test.test_gdb.test_cfunction_full
|    test.test_gdb.test_misc test.test_gdb.test_pretty_print
|    test_android test_apple test_asdl_parser test_clinic test_devpoll
|    test_free_threading test_generated_cases test_idle test_kqueue
|    test_launcher test_msvcrt test_startfile test_tcl test_tkinter
|    test_ttk test_ttk_textonly test_turtle test_winapi
|    test_winconsoleio test_winreg test_wmi
|
|9 tests skipped (resource denied):
|    test_curses test_peg_generator test_smtpnet test_socketserver
|    test_urllib2net test_urllibnet test_winsound test_xpickle
|    test_zipfile64
|
|455 tests OK.
|
|Total duration: 3 min 42 sec
|Total tests: run=46,926 skipped=2,469
|Total test files: run=483/492 skipped=28 resource_denied=9
|Result: SUCCESS
|DURATION: 223
|END: /usr/lib/python3/ptest
|2026-04-09T14:44
|STOP: ptest-runner
|TOTAL: 1 FAIL: 0
|root@qemux86-64:~#

Signed-off-by: Trevor Gamblin <tgamblin@baylibre.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 days agoclang/llvm: Upgrade to 22.1.3 release
Deepesh Varatharajan [Wed, 8 Apr 2026 06:02:32 +0000 (23:02 -0700)] 
clang/llvm: Upgrade to 22.1.3 release

Brings following fixes

e9846648fd61 [RISCV] Use a vector MemVT when converting store+extractelt into a vector store. (#190107)
c55c1c1140e1 [TargetLowering] Speculative fix for a non-determinism issue between different compilers. (#190219)
151b4a3bafdd [clang-repl] Use canonical types in QualTypeToString (#190528)
0fecbff69a53 [clang] Fix issues with const/pure on varargs function. (#190252)
ae980d3b4df7 [lldb] Iterate over a copy of the ModuleList in SearchFilter (#189009)
307a11d7aa62 [Flang] Fix lowering failure for some constructs inside a CHANGE TEAM (#184342)
cd7f7379a016 [AArch64][llvm] Fix encoding for `stshh` instruction (#189588)
4250a0fc5de9 [RISCV] Fix stackmap shadow trimming NOP size for compressed targets (#189774)
da8a5b950c7d [flang] Update Flang Extension doc to reflect previous change (#188088)
0626016c5b90 [LoongArch] Implement canMergeStoresTo to avoid crash while NoImplicitFloat attribute is set (#188005)
bc44e9b80349 [Hexagon] Use __HVX_IEEE_FP__ to guard protos that need -mhvx-ieee-fp (#184422)
e6e59749af18 [Hexagon] Define __HVX_IEEE_FP__ when -mhvx-ieee-fp is enabled (#183829)
dc9d64aad7a9 [lld][COFF] Restore `lto-embed-bitcode` and `-fembed-bitcode` Bitcode Embedding Features (#188398)
8fef0fb1a2de [clang-repl] Fix C89 incompatible keywords (#189432)
7b6514cd3089 [sanitizer_common] [Darwin] Adopt _dyld_get_dyld_header (#182943)
24b53fbc67d9 [X86] Improve illegal return type handling in FastISel (#186723)
cc87fcfeef54 [AccelTable] Pre-compute entry pool offsets for DW_IDX_parent. NFC (#188915)
561944ebcbd1 [lit] Explicitly unset timer to free thread stack (#188717)
58468775a5e2 [lit] dealloc ApplyResult objects as they're waited on (#188642)
719a040bf13f [lit] Stop holding subprocess objects open in TimeoutHelper (#186712)
b8918e6248a8 [LoongArch] Fix incorrect reciprocal sqrt estimate semantics (#187621)
62b14a88906d [MC,clang] Fix -Wa,--noexecstack not emitting .note.GNU-stack (#187880)
4bae0d33b2a4 [ARM] Fix incorrect post increment from Or (#188036)
d033ffe4eeb4 [ARM] Regenerate vld2.ll test. NFC
5f04300518b5 [AArch64] Use an unknown size for memcpy ops with non-constant sizes. (#187445)
f6198f61b2bd [Hexagon] Fix truncation to boolean vector that need widening (#182528)
240f42d95f51 [Hexagon] Fix 64-bit funnel shift miscompilation with register shift amounts (#183669)
a6cd46b239c7 [Hexagon] Fix B0 macro conflict between hexagon_types.h and termios.h (#184539)
4592b0b6388d [libunwind] Fix building on OpenBSD / FreeBSD aarch64 (#188397)
cf542a3d9c00 [lldb] use the Py_REFCNT() macro instead of directly accessing member (#188161)
d40e60759077 [LoongArch] Fix incorrect indexing of incoming byval arguments in tail call eligibility check (#188006)
b68aafeb0dc1 [clang][AST] Preserve qualifiers in getFullyQualifiedType for AutoType (#187717)
55b442495879 [Clang] Fix a concept subsumption bug when template depths are adjusted (#186735)
7b593cd6c0df Bump version to 22.1.3

Signed-off-by: Deepesh Varatharajan <Deepesh.Varatharajan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 days agoghostscript: Fix build with clang-22
Khem Raj [Wed, 8 Apr 2026 05:59:03 +0000 (22:59 -0700)] 
ghostscript: Fix build with clang-22

It seems clang can't handle floating-point arithmetic in
static initializers at compile time in this context.

Change the static const double declarations to
use #define macros fixes the problem

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 days agoopenssl: upgrade 3.5.5 -> 3.5.6
Peter Marko [Wed, 8 Apr 2026 05:57:30 +0000 (07:57 +0200)] 
openssl: upgrade 3.5.5 -> 3.5.6

Release information [1]:

OpenSSL 3.5.6 is a security patch release. The most severe CVE fixed in this release is Medium.
This release incorporates the following bug fixes and mitigations:
* Fixed incorrect failure handling in RSA KEM RSASVE encapsulation. (CVE-2026-31790)
* Fixed loss of key agreement group tuple structure when the DEFAULT keyword is used in
  the server-side configuration of the key-agreement group list. (CVE-2026-2673)
* Fixed potential use-after-free in DANE client code. (CVE-2026-28387)
* Fixed NULL pointer dereference when processing a delta CRL. (CVE-2026-28388)
* Fixed possible NULL dereference when processing CMS KeyAgreeRecipientInfo. (CVE-2026-28389)
* Fixed possible NULL dereference when processing CMS KeyTransportRecipientInfo. (CVE-2026-28390)
* Fixed heap buffer overflow in hexadecimal conversion. (CVE-2026-31789)

[1] https://github.com/openssl/openssl/blob/openssl-3.5/NEWS.md#major-changes-between-openssl-355-and-openssl-356-7-apr-2026

Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 days agoapr-utils: Fix old style function prototype warning with clang
Khem Raj [Wed, 8 Apr 2026 02:32:46 +0000 (19:32 -0700)] 
apr-utils: Fix old style function prototype warning with clang

Fixes build with autoconf-2.73 + clang-22

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 days agoflex: Keep it at C17 with clang
Khem Raj [Thu, 9 Apr 2026 05:56:17 +0000 (22:56 -0700)] 
flex: Keep it at C17 with clang

autoconf 2.73 enforces C23 and adds -std=gnu23 but it leaks it into
CC_FOR_BUILD as well, which bites on systems with GCC < 14 where this
option does not exist.

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 days agogn: Disable -Werror
Zoltan Boszormenyi [Thu, 9 Apr 2026 04:36:28 +0000 (06:36 +0200)] 
gn: Disable -Werror

With GCC 16 and GLIBC 2.43 on the host, this error is thrown:

/usr/include/c++/16/bits/stl_construct.h:88:9: error: array subscript ‘std::mutex[0]’ is partly outside array bounds of ‘unsigned char [24]’ [-Werror=array-bounds=]
   88 |         __location->~_Tp();
      |         ^~~~~~~~~~

Not even the latest version fixes this upstream.

Add --allow-warnings to the python command in do_configure
to disable -Werror during do_compile as a workaround.

Signed-off-by: Zoltán Böszörményi <zboszor@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 days agoglibc: stable 2.43 branch updates
Hemanth Kumar M D [Thu, 9 Apr 2026 03:43:32 +0000 (20:43 -0700)] 
glibc: stable 2.43 branch updates

$ git log --oneline 856c426a753450b8c6861a5b994a564f4fc16d4b..ce1013a197eb4a3b8ff2b07e0672f4d0b976ce7c

ce1013a197 tests: fix tst-rseq with Linux 7.0
60cabd0464 riscv: Resolve calls to memcpy using memcpy-generic in early startup
02ffd413cf elf: Use dl-symbol-redir-ifunc.h instead _dl_strlen
2695314378 elf: parse /proc/self/maps as the last resort to find the gap for tst-link-map-contiguous-ldso
dd9945c0ba resolv: Check hostname for validity (CVE-2026-4438)
5c6fca0c62 resolv: Count records correctly (CVE-2026-4437)
2be6cf2e75 posix: Run tst-wordexp-reuse-mem test
305ce0b588 aarch64: Tests for locking GCS
2ee41ba6ec aarch64: Lock GCS status at startup
fa4a40c7d4 tests: aarch64: fix makefile dependencies for dlopen tests for BTI
9898ea58b5 malloc: Avoid accessing /sys/kernel/mm files
c3ceb93dc4 Add BZ 33904 entry to NEWS
911bd469f8 debug: Fix build with --enable-fortify-source=1 (BZ 33904)
48f5a05a7a nss: Missing checks in __nss_configure_lookup, __nss_database_get (bug 28940)
d6cb7ce0e9 Linux: In getlogin_r, use utmp fallback only for specific errors
140c760d71 nss: Introduce dedicated struct nss_database_for_fork type

Testing Results:
+--------------+--------+--------+------+
| Result       | Before | After  | Diff |
+--------------+--------+--------+------+
| PASS         | 6770   | 6774   | +4   |
| XPASS        | 4      | 4      |  0   |
| FAIL         | 29     | 28     | -1   |
| XFAIL        | 16     | 16     |  0   |
| UNSUPPORTED  | 489    | 490    | +1   |
+--------------+--------+--------+------+

Changes in testcases:
+------------------------------------------------------+--------+-------------+
| Testcase                                             | Before | After       |
+------------------------------------------------------+--------+-------------+
| elf/tst-tls20                                        | FAIL   | PASS        |
| posix/tst-wordexp-reuse-mem                          | N/A    | PASS        |
| resolv/tst-resolv-invalid-ptr                        | N/A    | PASS        |
| resolv/tst-resolv-dns-section                        | N/A    | PASS        |
| nss/tst-nss-malloc-failure-getlogin_r                | N/A    | UNSUPPORTED |
+------------------------------------------------------+--------+-------------+

Justification:

commit - 2be6cf2e75 posix: Run tst-wordexp-reuse-mem test
Fixes Makefile dependency to ensure the testcase is executed.
Passing new testcase:
+PASS: posix/tst-wordexp-reuse-mem

commit - dd9945c0ba resolv: Check hostname for validity (CVE-2026-4438)
Adds validation for hostname parsing and introduces a regression test.
Passing new testcase:
+PASS: resolv/tst-resolv-invalid-ptr

commit - 5c6fca0c62 resolv: Count records correctly (CVE-2026-4437)
Fixes DNS answer section parsing and adds a regression test.
Passing new testcase:
+PASS: resolv/tst-resolv-dns-section

commit - 48f5a05a7a nss: Missing checks in __nss_configure_lookup, __nss_database_get (bug 28940)
Fixes null pointer dereference and improves NSS handling.
Added testcase:
UNSUPPORTED: nss/tst-nss-malloc-failure-getlogin_r

Signed-off-by: Hemanth Kumar M D <Hemanth.KumarMD@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 days agogo: upgrade 1.26.1 -> 1.26.2
Peter Marko [Wed, 8 Apr 2026 21:42:04 +0000 (23:42 +0200)] 
go: upgrade 1.26.1 -> 1.26.2

Upgrade to latest 1.26.x release [1]:

$ git --no-pager log --oneline go1.26.1..go1.26.2
9c8bf0e72a (tag: go1.26.2) [release-branch.go1.26] go1.26.2
096f21b1c5 [release-branch.go1.26] cmd/go: disallow cgo trust boundary bypass
7cafb4140d [release-branch.go1.26] crypto/tls: prevent deadlock when client sends multiple key update messages
a34b5e4d55 [release-branch.go1.26] archive/tar: limit the number of old GNU sparse format entries
c4b4bd7b3a [release-branch.go1.26] cmd/compile: fix loopbce overflow check logic
3a4577a268 [release-branch.go1.26] cmd/compile: fix mem access overlap detection
ceb4da6626 [release-branch.go1.26] crypto/x509: fix wildcard constraint map case sensitivity
babb1c32c2 [release-branch.go1.26] html/template: properly track JS template literal brace depth across contexts
b6176f459a [release-branch.go1.26] internal/syscall/unix: properly support AT_SYMLINK_NOFOLLOW on Linux
917949cc1d [release-branch.go1.26] net/url: permit colons in the host subcomponent of non-http/https URLs
7b4ed1d7d9 [release-branch.go1.26] crypto/x509: fix signature checking limit
8028a51430 [release-branch.go1.26] crypto/x509: hoist policy pruning out of loop
4ef278779f [release-branch.go1.26] runtime/race: apply LLVM zero-initialization fix
608303dd6e [release-branch.go1.26] cmd/link: skip special symbols for label symbol generation
e27183eff8 [release-branch.go1.26] cmd/compile: fix internal compiler error: bad write barrier type
98d4c11d8f [release-branch.go1.26] cmd/compile: treat all zero-sized values as SSA-able
cce667a91e [release-branch.go1.26] cmd/compile: handle zero-sized values more generally
3206e141fb [release-branch.go1.26] net/url: reattach package doc comment
82b95c14ec [release-branch.go1.26] cmd/compile: ternary rewrite of rewrite should skip, not panic
f14d3a8e91 [release-branch.go1.26] cmd/compile: skip race detector test failure for unsupported VMA
68f42c591d [release-branch.go1.26] cmd/compile: gate instrumentEnterExit on NoRaceFunc check
9fe1db7e7e [release-branch.go1.26] runtime: fix value of ENOSYS on mips from 38 to 89
08edcb8d3f [release-branch.go1.26] test/fixedbugs: remove issue46234 test timeout
895fbaecf5 [release-branch.go1.26] doc: remove template use in godebug.md
26548d4981 [release-branch.go1.26] cmd/fix: pull in CL 755980
12bf7d8cd2 [release-branch.go1.26] builtin: incorporate all feedback into doc string for new
943235a5e5 [release-branch.go1.26] builtin: update new function comment
1a44be4cec [release-branch.go1.26] internal/poll: move rsan to heap on windows
3bf3014fe7 [release-branch.go1.26] test/convert5: skip mipsle
e29f35511d [release-branch.go1.26] cmd/go/internal/cache: update trim timestamp before trimming
2b01132c9f [release-branch.go1.26] runtime/poll: fix race condition in Window's SendFile
ba8c3d07a7 [release-branch.go1.26] net/http: restore accidentally removed package doc
99df9bdb2a [release-branch.go1.26] net: correctly handle non-nil empty cmsg buffer on windows
ace2cf6652 [release-branch.go1.26] cmd/compile/internal/typecheck: simplify tcSliceHeader
1368bc269e [release-branch.go1.26] cmd/compile: fix bloop test
840d86c060 [release-branch.go1.26] runtime: fix printfloat, printcomplex buffer sizes

Fixes CVE-2026-32282, CVE-2026-32289, CVE-2026-33810, CVE-2026-27144,
      CVE-2026-27143, CVE-2026-32288, CVE-2026-32283, CVE-2026-27140,
      CVE-2026-32280 and CVE-2026-32281.

Release information: [2]

[1] https://github.com/golang/go/compare/go1.26.1...go1.26.2
[2] https://groups.google.com/g/golang-announce/c/0uYbvbPZRWU

Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
3 days agoclasses/cve-check: remove class
Ross Burton [Tue, 31 Mar 2026 13:24:29 +0000 (14:24 +0100)] 
classes/cve-check: remove class

It's been long known that the cve-check class in oe-core is not that
usable in the real world, for more details see "Future of CVE scanning
in Yocto"[1].  This mail proposed an alternative direction that included
a CVE scanning tool that can be ran both during the build and afterwards,
so that periodic scans of a previously build image is possible.

Last year, Bootlin wrote sbom-cve-check[2] and I compared this to my
proposal in "Comparing cve-check with sbom-cve-check"[3], concluding
that this is likely the missing piece.

Support for sbom-cve-check has been merged into oe-core, and the
cve-check class is now obsolete. So that we don't have to maintain it for
the four-year lifecycle of the Wrynose release, delete it.

This patch also deletes the database fetcher recipes, and the test cases
that were specific to cve-check.  Note that the oe.cve_check library
still exists as this is used by the SPDX classes.

[1] https://lore.kernel.org/openembedded-core/7D6E419E-A7AE-4324-966C-3552C586E452@arm.com/
[2] https://github.com/bootlin/sbom-cve-check
[3] https://lore.kernel.org/openembedded-core/2CD10DD9-FB2A-4B10-B98A-85918EB6B4B7@arm.com/

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>
4 days agolibgcc: Bump PR/HASHEQUIV_HASH_VERSION after reproducibility issues
Richard Purdie [Thu, 9 Apr 2026 20:16:02 +0000 (21:16 +0100)] 
libgcc: Bump PR/HASHEQUIV_HASH_VERSION after reproducibility issues

Simulate a PR bump by changing DESCRIPTION which will force and output
change. PR isn't a good candidate here due to the shared workdir.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoclasses-global/package: Ensure -src package is handled in signatures
Joshua Watt [Thu, 9 Apr 2026 15:46:43 +0000 (09:46 -0600)] 
classes-global/package: Ensure -src package is handled in signatures

populate_packages() will automatically insert a -src package if
PACKAGE_DEBUG_SPLIT_STYLE is "debug-with-srcpkg", even if there is not
-src package in PACKAGES. However, when the -src package is not in
PACKAGES, it is not accounted for in the task signatures which results
in errors.

[YOCTO #16238]

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoclasses/sbom-cve-check: add variable to control the scope of the CVE scan
Ross Burton [Thu, 9 Apr 2026 16:26:22 +0000 (17:26 +0100)] 
classes/sbom-cve-check: add variable to control the scope of the CVE scan

Unless told otherwise, sbom-cve-check will only report on issues found
in target recipes.  Whilst this is the behaviour we want for image-based
scanning, it's not the designed bevhaviour for recipe-based scanning.

Add a new variable SBOM_CVE_CHECK_SCAN_SCOPE that is used as the value
passed with --export-process-native.  Default this to "target" so the
normal behaviour is unchanged, but set it to "both" in the
sbom-cve-check-recipe class.

[1] https://sbom-cve-check.readthedocs.io/en/latest/export.html#export-options

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agopython3-sbom-cve-check: fix the version attribute in the Yocto-style JSON
Ross Burton [Thu, 9 Apr 2026 16:26:21 +0000 (17:26 +0100)] 
python3-sbom-cve-check: fix the version attribute in the Yocto-style JSON

The type of the version was an integer, but it should be a string.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agopython3-sbom-cve-check: move from devtools/python to devtools/sbom-cve-check
Ross Burton [Thu, 9 Apr 2026 16:26:20 +0000 (17:26 +0100)] 
python3-sbom-cve-check: move from devtools/python to devtools/sbom-cve-check

Move the tool recipe from the catch-all python/ directory to be alongside
the recipes that fetch the databases for the tool.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agooeqa: Drop /git/ from our urls
Richard Purdie [Wed, 8 Apr 2026 07:13:13 +0000 (08:13 +0100)] 
oeqa: Drop /git/ from our urls

Using /git/ in our urls is rather old school and not the preferred format now.
Update the urls to the preferred form even if the other ones still work.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoselftest/scripts: Update old git protocol references
Richard Purdie [Wed, 8 Apr 2026 06:45:09 +0000 (07:45 +0100)] 
selftest/scripts: Update old git protocol references

git protocol accesses to our infrastructure are currently struggling and this
has highlighted a number of places we're making those obsolete access forms.

Update them to use https instead of the git protocol since it is preferred
and more reliable.

The devtool test needed quoting to handle the ';' in the url. The -f option
to devtool also shows a deprecation warning so remove that.

There were internal references to git protocol urls inside the nested git
submodules test report, which means those repos need updating to use
new git revisions.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agorecipetool: Recognise https://git. as git urls
Richard Purdie [Wed, 8 Apr 2026 21:43:17 +0000 (22:43 +0100)] 
recipetool: Recognise https://git. as git urls

If a url has git. in it, assume it is likely to be a git cloneable url
and should be treated as such.

This allows us to switch from https://git.yoctoproject.org/git/XXX urls to
the preferred https://git.yoctoproject.org/XXX form.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agodocbook-xsl-stylesheets: Use str:replace from exslt.org to implement string.subst
Khem Raj [Mon, 6 Apr 2026 16:19:08 +0000 (09:19 -0700)] 
docbook-xsl-stylesheets: Use str:replace from exslt.org to implement string.subst

Developement for docbook stylesheets has moved to github
there are newer releases available too which include this fix, however
for now nearing 6.0 release this is least intrusive way to move forward

This is needed for building latest samba versions

Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agorust: Upgrade 1.94.0 -> 1.94.1
Deepesh Varatharajan [Mon, 6 Apr 2026 12:01:53 +0000 (05:01 -0700)] 
rust: Upgrade 1.94.0 -> 1.94.1

https://blog.rust-lang.org/2026/03/26/1.94.1-release/

Rust Changes:
https://github.com/rust-lang/rust/releases/tag/1.94.1

Cargo Changes:
https://doc.rust-lang.org/nightly/cargo/CHANGELOG.html#cargo-1941-2026-03-26
Fixes: CVE-2026-33055 and CVE-2026-33056
Test results:
No changes

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>
4 days agosystemd: fix packaging of 'run0'
Matt Madison [Sun, 5 Apr 2026 13:40:49 +0000 (06:40 -0700)] 
systemd: fix packaging of 'run0'

The 'run0' command was introduced with systemd v256. Its
implementation is in the 'systemd-run' program, with
${bindir}/run0 installed as a symlink to that program.

However, when the upgrade to v256 was done (in commit
89b75b46371d5e9172cb496b461824d8551a2af5), the run0 symlink
wasn't packaged into systemd-extra-utils (where systemd-run
is packaged), introducing a hard RDEPENDS from systemd ->
systemd-extra-utils due to the symlink.

Fix this, and return to the softer RRECOMMENDS dependency
between systemd and systemd-extra-utils, by packaging run0
into systemd-extra-utils.

Signed-off-by: Matt Madison <matt@madison.systems>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agovim: Fix CVE-2026-33412
Ashish Sharma [Sat, 4 Apr 2026 15:51:28 +0000 (08:51 -0700)] 
vim: Fix CVE-2026-33412

Pick patch from [1], also referenced by NVD [2].

The upstream fix escapes newline in SHELL_SPECIAL to prevent command

injection via glob() shell expansion.

[1] https://github.com/vim/vim/commit/645ed6597d1ea896c712cd7ddbb6edee79577e9a

[2] https://nvd.nist.gov/vuln/detail/CVE-2026-33412

Signed-off-by: Ashish Sharma <pahaditechie@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogst-examples: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:18 +0000 (16:17 +0200)] 
gst-examples: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0:
b2a3b2e: Back to development after 1.28.0
6ec50ef: tutorials/android: bump up ndkVersion to 29
4f6b36b: gst-examples/webrtc: fix error reported with android NDK 29
425d36f: gst-examples: webrtc: Update Rust dependencies
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogst-devtools: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:17 +0000 (16:17 +0200)] 
gst-devtools: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0
b2a3b2e: Back to development after 1.28.0
835da19: modules: Remove NEWS from git which is generated from full
  release notes
201b14e: modules: remove RELEASE from git, will be generated from
  template on dist
6376a84: modules: dist common files from monorepo root
57be895: devtools: dots-viewer: Update Rust dependencies
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-rtsp-server: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:16 +0000 (16:17 +0200)] 
gstreamer1.0-rtsp-server: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0
b2a3b2e: Back to development after 1.28.0
835da19: modules: Remove NEWS from git which is generated from full
  release notes
201b14e: modules: remove RELEASE from git, will be generated from
  template on dist
6376a84: modules: dist common files from monorepo root
012b257: rtspclientsink: don't error out when stream transport
  notifies timeout
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
c2ef75a: rtsp-stream: Clear send_thread when it's freed
20d3e66: rtsp-client: Lock media when unlinking session medias
0882b60: gst: Fix a couple of const correctness bugs around strchr()
  usage
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-python: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:15 +0000 (16:17 +0200)] 
gstreamer1.0-python: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0
b2a3b2e: Back to development after 1.28.0
539aaa1: python: Caps constructor has a single optional argument
18d4781: python: Structure constructor has a single required argument
6f21099: python: Add type hints for MapInfo members
0df78d0: python: Remove duplicated TestDoubleRange
d4af24c: python: Add type hint for public fields
5f4fe17: python: int|float type is redundant and cause linter issue
835da19: modules: Remove NEWS from git which is generated from full
  release notes
201b14e: modules: remove RELEASE from git, will be generated from
  template on dist
7694a7d: modules: remove subproject README.md from git
6376a84: modules: dist common files from monorepo root
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
c43f2b2: python: Add item type annotation to Iterator class
49ca53f: python: Make Gst.Bin iterable
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-libav: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:14 +0000 (16:17 +0200)] 
gstreamer1.0-libav: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0
b2a3b2e: Back to development after 1.28.0
be2f846: avviddec: Allow stride changes for some decoders
f0bfd0c: avviddec: Handle field/order changes in mixed interlace mode
f717c6e: avviddec: Simplify picture_changed helper function
0f77771: gst: fix author name: add missing closing angle bracket
d125725: gst: also adapt author names in the gst_plugins_cache.json
  files
92ba6c2: avviddec: Don't increment the AVCodecContext frame counter
529e7a6: avauddec: Don't increment the AVCodecContext frame counter
aa02005: avviddec: Forward any ready video frames on gap events
0a6950d: gst-libav: avvidcmp: set colorimetry on AVFrame
adc82d2: libav: Don't process lines that won't be outputted
835da19: modules: Remove NEWS from git which is generated from full
  release notes
201b14e: modules: remove RELEASE from git, will be generated from
  template on dist
7694a7d: modules: remove subproject README.md from git
6376a84: modules: dist common files from monorepo root
60c9be9: avviddec: Simplify conditional code in context handling
b6fbaaf: avviddec: Improve renegotiation debug message
d94518f: avviddec: Fix handling of mixed interlaced content
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
80cd0ee: avviddec: Refcount codec frame associated with video frame
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-plugins-ugly: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:13 +0000 (16:17 +0200)] 
gstreamer1.0-plugins-ugly: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0
b2a3b2e: Back to development after 1.28.0
835da19: modules: Remove NEWS from git which is generated from full
  release notes
201b14e: modules: remove RELEASE from git, will be generated from
  template on dist
7694a7d: modules: remove subproject README.md from git
6376a84: modules: dist common files from monorepo root
8f9c0ad: asfdemux: Error out on files with more than 32 streams
cb184da: rmdemux: Check if new video fragment overflows the fragment
  storage before storing it
3bcbda6: rmdemux: Avoid integer overflow when checking if enough data
  is available for video fragment
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
37274be: gst-plugins-ugly: update translations
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-plugins-bad: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:12 +0000 (16:17 +0200)] 
gstreamer1.0-plugins-bad: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0
b2a3b2e: Back to development after 1.28.0
99c68cb: meson: Fix libxml2 not building due to wrong option type
e973899: cudaupload, cudadownload: Fix CUDA/GL interop copy path
83b8417: webrtc: sink floating refs of ICE transports
887f2bf: ajasink: Only allow 6 / 8 / 16 audio channels
a74bd95: mpegtspacketizer: Handle clock change/resets without skew
  correction
956e714: mxfdemux: fix gst_mxf_demux_pad_get_stream_time ()
73bc4b8: tsdemux: Fix Continuity Counter handling
3795ca5: tsdemux: Simplify initial packet handling
150a7e4: mxfdemux: always send a segment before sending eos or
  segment-done
3fdc449: codectimestamper: Fix latency query handling
530a52f: vtdec: Always reset flushing flag in flush()
4dfe11e: vtdec: Fix a very slight race in handle_frame() when
  flushing
0a5e48c: vtdec: Make sure to reset last flow return when flushing
ae04185: vtdec: Implement drain() to allow reverse playback
9e0e66f: webrtcbin: Check the presence of encoding-name fields in
  answer caps
fe28caa: vadisplay: Fix a couple of new const-ness warnings around
  strstr() usage
b88511e: classifiertensordecoder: Fix uninitialized variable compiler
  warning
9924971: gstplay: fix reference counting
f721368: v4l2: Add support for AV1 V4l2 decoder
0fd06de: vulkan/ios: Fix scaling and resizing with UIView
0708627: audiobuffersplit: Correctly calculate adapter start/end
  running time for negative rates
9ae3903: audiobuffersplit: Use the output segment for output related
  calculations
1a9c31f: audiobuffersplit: Don't use floating point divisions
  unnecessarily
5e9ff8c: audiobuffersplit: Log if an input buffer has the discont
  flag set or not
2ded76c: audiobuffersplit: Send any pending segment before
  EOS/SEGMENT_DONE
72b9964: audiobuffersplit: Handle SEGMENT_DONE the same way as EOS
1f43861: audiobuffersplit: Convert gap events to silence buffers
16a2fef: audiobuffersplit: Implement negative rates correctly
6c6aba2: mpeghdec: memory leak fix in MPEG-H Audio decoder plugin
63e8a3b: vtdec: Don't re-create session if only the framerate changed
9e17a28: audiobuffersplit: fix reverse playback
59a7e80: vtdec: Add more debug logging
321c742: vtdec: Fix race condition when negotiating during playback
030cfc1: wayland: Fix CLAMP operation of maxFALL and maxCLL
705376e: objectdetectionoverlay: add support for rotated bounding
  boxes
d2aae83: cea608mux: fix overflow when calculating output PTS
05d821e: siren: remove trailing space from klass
24d4d91: openni2src: remove trailing space from desc
2700371: smoothstreaming: remove trailing space from desc
d5e0ba7: gst-plugins-bad: remove trailing spaces: update
  gst_plugins_cache.json
f7f24fc: zebrastripe: adapt klass to recommendation
d032f3c: fieldanalysis: adapt klass to recommendation
3c06dee: dtls: fix author name: add missing angle brackets
0f77771: gst: fix author name: add missing closing angle bracket
4506913: gst-plugins: fix author name: correct incomplete or wrong
  emails
d125725: gst: also adapt author names in the gst_plugins_cache.json
  files
c20ae61: videocodectestsink: fix typo in klass
b463b8f: unixfdsrc: adapt klass "Src" to "Source"
8aa4eff: avtpsrc: adapt klass "Src" to "Source"
e6348c3: rtpsrc: adapt klass "Src" to "Source"
d21129f: frei0r-src: adapt klass "Src" to "Source"
78ce4b2: tsmux: reduce noise for DEBUG log level
5b40409: openh264enc: skip drain for baseline profile
53a95b2: openh264enc: remove broken drain and simplify handle_frame
fafeaef: ajasinkcombiner: Only forward the segment events from the
  video sinkpad
988666b: meson: Add a subproject for providing the LunarG MoltenVK
  SDK
5084c9b: vulkan: load video function pointers conditionally based on
  codec operation
011a06b: waylandsink: make gst_wl_window_commit_buffer handle NULL
  buffers
f201683: vtdec: Fix CM memory leak due to incorrect unref
930b4aa: zxing: Minimal support for compiling with zxing-cpp 3.x
51f45dc: vabasetransform: copy buffer's metadata at copy when import
  buffer
dbefb70: zxing: Fix version check for zxing-cpp 3.0.1
835da19: modules: Remove NEWS from git which is generated from full
  release notes
201b14e: modules: remove RELEASE from git, will be generated from
  template on dist
7694a7d: modules: remove subproject README.md from git
6376a84: modules: dist common files from monorepo root
2601491: vtdec: Avoid busy looping when queue length is smaller than
  DPB size
b76fe59: ladspa: Fix macOS detection
291e479: meson: Deprecate `system = 'ios'` in cross files, use
  subsystem
4aa405d: applemedia: Stop using HAVE_IOS, use TARGET_OS_* macros
  instead
b4eaad5: avfdeviceprovider: AVCaptureDevice manufacturer is more
  widely available
38655e0: applemedia: Enable EAGL support on tvOS too
25fbfd3: vtdec: VTRegisterSupplementalVideoDecoderIfAvailable is
  widely available
e833ae2: vtdec: Software decoders are now more widely available
082bc62: avfvideosrc: Add support for newer iOS 17+ APIs
190153c: vtenc: Fix build with tvOS
1c213de: applemedia: Disable avfvideosrc and avfdeviceprovider on
  tvOS/watchOS
7807e72: vtdec: Check for AV1 and VP9 support once
9d8e01f: macos: Stop using HAVE_OSX, use built-in macros instead
86b0957: frie0r, ladspa: Stop using G_MODULE_SUFFIX
149cc0e: vtdec: Fix GstVideoCodecState leak
2f5b70e: asiosink: Fill silence when paused
ad31d52: applemedia: elements can now be individually registered with
  gst-full
8074134: h264,h265ccextractor: Fix framerate in initial caps
2efd57c: ccconverter: Reset counters on flush-stop
af04ec5: vkav1dec: fix to set SavedOrderHints properly
a10695e: v4l2codecs: Add short and long term controls in linux
  headers
1721252: h265parser: Store raw short/long term RPS sets
8788a6c: codecs: h265dec: Parse extended SPS information
54994bb: v4l2codecs: Set long and short term RPS controls
c22cdd5: build: codecs: Add AV1 decoder to the GIR includes
81b283e: build: codecs: Small style improvement
d795915: build: codecparsers: Install some newly introduce API
  headers
0fd74d0: build: codecparsers: Create a GIR file needed for since
  marking
3686103: codec: h265decoder: Fix annotation and constify return value
1768b12: build: codecs: Add gir dependency to the new GstCodecParsers
  gir
282fd81: codecparsers: h265/vc1: Add missing namespace to some
  defines
9579389: codecparsers: Fix annotation warnings reported at GIR
  constructions
947d59f: doc: codecparsers: Switch to gi-index
13b57d5: meson: Explicitly use cpp_std=c++11 for decklink
1ff173e: vavp8enc: set color format chroma
c1ee041: tensordecoders: fix wrong dims_order check
cbd357d: onnx: set dims_order on GstTensor
94c24fe: vtdec: Fix return value in flush()
e5f8ddf: vtdec: Make sure VT thread won't get stuck after a
  downstream error
a1c2bb8: vtenc: Make sure draining actually drains all frames, port
  fixes from vtdec, unify naming
3313eab: vtdec: Fix wrong DPB size check in the output loop
ab5cead: vtdec: Calculate DTS offset correctly
673922a: applemedia: Sort sources list, fix indentation
1205741: applemedia: Fix broken HAVE_VIDEOTOOLBOX_10_9_6 define
c1772be: applemedia: Remove outdated HAVE_VIDEOTOOLBOX define
8bcac75: applemedia: Get rid of AVFoundation-related defines
68c0b40: avfvideosrc: Small fixes to the plugin header
1758e23: applemedia: Small simplifications to the vulkan checks
561bfde: dvbsuboverlay: Mark parsed byte array as const
69a2c1c: dvbsuboverlay: Add missing bounds checks to the parser
  everywhere
cfb98fe: dvbsuboverlay: Avoid integer overflows and unreasonably
  large displays/regions
fcd957d: libs: jpegparser: boundary checks before copying it
8a26fa7: h266parser: Validate tile index bounds in picture partition
  parsing
080ac76: h266parser: Fix APS ID bounds check in APS parsing
00a086a: h266parser: Fix out of bounds write when parsing pic_timing
  SEI
5285725: h265parser: Validate num_decoding_units_minus1 in pic_timing
  SEI
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
6a3b3b6: av1parse: Add the GstAV1ParseStreamFormat enum and helper
  functions
5b115b5: av1parse: Split the stream format and alignment logic
3b7e321: svtav1: fix "Level of parallelism" property type
  discrepencies
74e26cf: dashsink: test: use playbin3 for DASH playback verification
b99ef73: dashsink: guard splitmuxsink removal in release_pad during
  dispose
31a80ef: vulkan: Fix libMoltenVK.dylib not found when installed
216d38a: all: GThreadFunc return type fixes
9bd09cc: srtpenc: preserve ROC when master key is updated for an
  ongoing session
7e5161c: av1dec: Enable VIDEO_META and VIDEO_ALIGNMENT for pool
b72d4cd: av1parse: Fix the consumed typo in _read_leb128
47002f8: av1parse: Fix a latent wrong setting of cll in
  update_src_caps
0426e58: decklinkvideosink: fix element leak in decklink callback
16b556c: soundtouch: Only allow up to 192kHz and 16 channels
4520fe5: gdppay: Fix null pointer dereference on duplicated caps
  event
cf7ec94: h264parse: Do not update valid DTS and duration
d850e77: h265parse: Update buffer duration only when it's invalid
eb67339: vtdec: Store supplemental codec support in a global variable
fb821c3: onnx: set default pixel-aspect-ratio
8e2db48: tflite: set PAR to 1:1 by default
b5fa0db: vulkan: Clear mutex when GstVulkanImageMemory is freed
be3a040: h264parser: Fix memory leak in gst_h264_parser_parse_nal()
18dd8d6: tsmux: Fix integer overflow in SCTE35 NULL interval
95b4388: gst-plugins-bad: update translations
07c0ecc: dtls: unregister signal handlers from connection
e8b7cc0: vp9parse, av1parse: Remove segment clipping to let
  downstream handle frame boundaries
259d5e7: vtdec: Do not hold the stream lock when pushing out frames
72afb2c: nice: Fix leak of nice thread
391d843: vulkanvp9dec: Fix case in device-specific factory name
09872a7: wasapi2sink: Ignore device errors from default device
5d165e6: mxfdemux: reject corrupted index entry
a277cb4: mxfdemux: add reversed temporal offset bound check
384e65b: waylandsink: fix waylandsink crash when call window flush
9552c85: wlwindow: fix viewport source outside buffer when play
  resolution change stream
8320b8d: vtdec: Supplemental VideoToolbox decoders now registered via
  vtutil helper
dea92e8: vkswapper/vksink: Don't advertise unsupported formats
c8a404b: shm: fix shmsink exit code 1 on clean shutdown
5dfb388: wayland: display: Add protection when replacing wl_output
0882b60: gst: Fix a couple of const correctness bugs around strchr()
  usage
5ed6a6d: wasapi2: Log target device information
b05db6e: sctp: Set number of outgoing & incoming streams to the same
  value
6e69d47: applemedia/vtdec: handle decoder error status for iOS
cf067cc: applemedia/vtenc: make sure to NULL terminate the level
  string
299ffee: vtenc: restart even if VTCompressionSessionCompleteFrames
  fails
a8d3ce8: vtdec: Also reset session when output_callback gets a
  kVTVideoDecoderMalfunctionErr
63efa3c: waylandsink: Properly reset the tag orientation
701194a: nvcodec: Add capability caching to speed up plugin
  initialization
18851f9: vtdec: Don't leak RGBA64_LE caps during class init
327188d: vtdec: Fix incorrect setting of bitfield / flag
b5af409: vtdec: Invert order of output caps to match our preference
e265144: macos: Set activation policy in vulkansink (MoltenVK)
f9cc731: jp2kdecimator: Fix some possible integer overflows in size
  checks
e900fe1: jp2kdecimator: Avoid integer overflows and divisions by zero
  on invalid tile configurations
ad02aae: h264parse: Remove pointless allocation failure handling
43a89cd: h264parse: Avoid NULL pointer dereferences when freeing
  partially parsed SPS/MVC data
c2690af: h266parser: Avoid integer overflow when parsing profile /
  tier / level
845e9cf: av1parse: Avoid signed 32 bit integer overflow when parsing
  LEB128 values
f9d57db: av1parse: Be more explicit about available data when parsing
  LEB128 values
6e84260: av1parse: Allow G_MAXUINT32 as LEB128 encoded value
00d6368: av1parse: Correctly reject LEB128 values where the 8th byte
  has the high bit set
871ac21: vmncdec: Set cursormask to NULL to prevent double free
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-plugins-good: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:11 +0000 (16:17 +0200)] 
gstreamer1.0-plugins-good: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0
b2a3b2e: Back to development after 1.28.0
99c68cb: meson: Fix libxml2 not building due to wrong option type
8388e74: qtdemux: Improve debug output around seek event handling
db3c0c6: qtdemux: Rename last mention of last_stop to position
19e3c57: qtdemux: Set the segment position to the start on EOS in
  reverse playback mode
5c681d2: v4l2: Add v4l2av1dec stateful decoder support
6438539: v4l2: update gst plugins cache
d3c0283: vpxdec: Support downstream pools with alignment requirements
0edccca: qtdemux: Don't ignore flow return when pushing queued
  buffers downstream
ff3edfc: qtdemux: Make sure to not output the same samples multiple
  times in reverse playback mode
d415a2b: qtdemux: Push raw audio/video buffers downstream in reverse
  order if rate < 0
e1c1979: wavpackparse: Parse 32 bit sample rate and channel masks
74d8469: wavpackparse: Print hexadecimal numbers with leading zeroes
  for easier reading
d296187: wavpackparse: Sync flags with latest wavpack
4ad4418: wavpackparse: Include sample type (int / float / dsd) in the
  caps
6631df5: wavpackparse: Correctly parse and advertise depth vs. width
f4d6909: wavpackdec: Set OPEN_NORMALIZE flag to normalize floating
  point samples into the [-1,1] range
7005bb5: wavpackdec: Allow up to 4096 channels and up to maximum
  sample rate
7514ac2: wavpackdec: Re-create wavpack decoder on caps changes
2cda74a: wavpackdec: Output 24 bit samples as actual 24 bit samples
69e76ae: wavpackdec: Allocate output buffer via the base class
293e350: wavpackdec: Output 18 and 20 bit as their corresponding
  formats
4d3bdcb: wavpackenc: Extend caps a bit
c5143e8: wavpackenc: Map buffer readwrite for channel reordering
3db67a3: wavpackenc: Add support for S8/S16/S18/S20/S24/S32 and F32
  samples
4704dfe: wavpack: Update plugin docs cache
d333aeb: qml6glsrc: Fix rendering of scene with clipped items
ee8e500: qml6glsrc: Fix scraping of QQuickWindow content
2b09aec: rtpsource: Add locking for receive reports table
eaadb4d: gst-plugins-good: fix author name: Kentaro Fukuchi
0f77771: gst: fix author name: add missing closing angle bracket
e92f814: rtph263pay: fix author name: where not separated by ','
4506913: gst-plugins: fix author name: correct incomplete or wrong
  emails
d125725: gst: also adapt author names in the gst_plugins_cache.json
  files
f68c471: rtpptdemux/rtpssrcdemux: adapt klass "Demux" to "Demuxer"
75ae4d7: gstrtspsrc: Set new mki in the encoder upon crypto update
09635fe: rtspsrc: Memory leak in gst_rtspsrc_close() when
  GST_RTSP_EEOF error occurs
835da19: modules: Remove NEWS from git which is generated from full
  release notes
201b14e: modules: remove RELEASE from git, will be generated from
  template on dist
7694a7d: modules: remove subproject README.md from git
6376a84: modules: dist common files from monorepo root
291e479: meson: Deprecate `system = 'ios'` in cross files, use
  subsystem
88febbc: osxaudio: Stop using HAVE_IOS, use TARGET_OS_* macros
  instead
c3b73e9: qtdemux: Fix out-of-bounds read when parsing PlayReady DRM
  UUIDs
b4558a4: rtpqdm2depay: error out if anyone tries to use this element
1171ae8: wavparse: Remove pointless duplicated GST_ROUND_UP_2()
3564405: wavparse: Use unsigned integers for data sizes
c73a1f4: wavparse: Use GST_ROUND_UP_2() in two more places instead of
  a manual implementation
8822ee3: wavparse: Define maximum chunk size in a single place
081484e: wavparse: Avoid integer overflow and out-of-bounds read when
  parsing adtl chunks
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
5ee8c64: rtptwcc: fix feedback packet count wrapping at 255
216d38a: all: GThreadFunc return type fixes
180a877: rtph264depay: fix invalid memory access in
  gst_rtp_h264_finish_fragmentation_unit
d824117: Qt6GLVideoItem: caps update fixed
d6ed0a0: qtdemux: fix invalid WebVTT timestamps
b8436bf: wavparse: Avoid overflow in length when setting
  ignore-length=true
586ff9c: wavparse: Fix parsing of RF64 wave files
b08a64e: rgvolume: don't apply dBSPL reference level compensation for
  LUFS values
9ca0bd6: hlsdemux2: fix seekable range for live HLS streams
100a0e6: qtdemux: Don't immediately push segment after moov in push
  mode for fmp4
95919fa: wavenc: Skip writing empty LIST INFO chunk
e717c43: gst-plugins-good: update translations
1ac03ff: qtdemux: fix handling of in-between fragments without tfdt
457b197: qtdemux: Preserve Metas and Flags when doing row alignment
27a9cc0: qtdemux: Avoid integer overflows when handling transform
  matrices
1279ec9: qtdemux: Don't store 64 bit integers in 32 bit integers to
  avoid overflows
0871bb2: qtdemux: Check that big enough stco/stsz are available when
  parsing sample tables
160bba0: qtdemux: Error out instead of trying to handle a truncated
  stts box
b73c493: audioinvert: fix float truncation in transform_float
bcc8c6e: qmlglsink: Fix for caps tracking on multiple setCaps calls
5ebb94c: qt6: Avoid parsing caps on every buffer
49bab9c: qt5: Avoid parsing caps on every buffer
c4f56c0: rtspsrc: Discard early data in ONVIF mode
86e640a: rtspsrc: Fix const-correctness issue around strchr() usage
48cefc4: flvmux: fix race condition on caps get and check
dbd4cb4: qtdemux: Avoid division by zero if 0 audio channels are
  signalled
10fd1ab: qtdemux: Validate chnl defined layout before using it to
  index the layouts array
3441881: qtdemux: Avoid out-of-bounds reads and writes of 64 item
  audio channel positions array
dc7ab66: qtdemux: Fix bit pattern check for omitted audio channels
  map
bad6721: qtdemux: Add various integer overflow and bounds checks to
  uncompressed video handling
8aed48f: flvdemux: Avoid assertions on corrupted streams
35a905a: wavparse: Fix integer overflow when checking available
  buffer size for reading cues
0d819ce: wavparse: Use prepend+reverse instead of append when
  building the cues list
6db6dd0: matroskademux: Add missing parenthesis when calculating bz2
  buffer sizes
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-plugins-base: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:10 +0000 (16:17 +0200)] 
gstreamer1.0-plugins-base: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0
b2a3b2e: Back to development after 1.28.0
e2c7b85: input-selector: add pad release stress test
6cb0626: input-selector: add test eos for remaining non active pad
c4f4ad2: typefindfunctions: Promote y4m_typefind
62629e5: decodebin3: Ensure input->collection is always coherent
cad6787: decodebin3: Improve handling collection change on existing
  pad
7365663: rtp: Add mappings for H266 and AV1 encoding-names
3d82e94: base: Fix a couple of new const-ness warnings around
  strstr() usage
848766d: gl: Fix explicit enabling of EAGL support on iOS
47576c7: eagl: Auto-resize when the UIView window is resized
4faeb36: audiorate: Handle SEGMENT_DONE the same way as EOS
2a7e50c: videodecoder: Handle recovery from temporary reordered
  output
d6fb546: playsink: unref color balance channels with g_object_unref()
efc09c7: decodebin3: Fix switch to smaller collections
acbe1ed: rtsp: Validate transport parameter parsing in RFC 2326
746553b: glsrcbin: adapt klass "Src" to "Source"
a02c813: audiodecoder: Only forward gap events immediately if no
  frames are pending
045b7a0: videodecoder: Move gap event handling into a helper function
ff28f77: videodecoder: Don't directly forward gap events if the
  decoder is not drained
b1c6307: videofilter: Add VIDEO_ALIGNMENT to downstream pool
835da19: modules: Remove NEWS from git which is generated from full
  release notes
201b14e: modules: remove RELEASE from git, will be generated from
  template on dist
7694a7d: modules: remove subproject README.md from git
6376a84: modules: dist common files from monorepo root
c6ea1f1: compositor: Do copy_metas also for background frame
291e479: meson: Deprecate `system = 'ios'` in cross files, use
  subsystem
d765ba7: eagl: Add support for tvOS
507fa0a: gstgl: Fix incorrect usage of G_MODULE_SUFFIX on macOS
e14eecd: video-converter: Do not transform_metas with 0 width or
  height
133c200: meson: Fix building of examples/gl/qt/qglwtextureshare
377a41d: riff: Correctly check that enough RGB palette data is
  available
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
de557ba: glcolorconvert: Fix NULL pointer dereference on buffers
  without video meta
216d38a: all: GThreadFunc return type fixes
ecbaac1: rtpbuffer: Add validation for CSRC list length
8b1dec6: rtcp: Fix buffer overread in SDES packet parsing
3e5ac12: opusenc: Use correct memcpy() size when copying Vorbis
  channel positions
8b4ea49: video-converter: fix I420/A420/AYUV BGRA/ARGB output on
  big-endian
cad4597: videorate: Fix unrestored caps on backward PTS
ae535bd: video: fix too small default stride for UYVP with odd widths
6c2d5bd: gl: upload: Optimize EGL format check when there is no
  context
8112ede: gl: upload: Fix linking glupload with restrictive caps
  filter
5daccd1: gst-plugins-base: update translations
69d7aed: compositor: move gst_compositor_init_blend() to element
  class_init
2f35652: rtsp: gstrtspurl: Parse URL having user without password
ae8c618: videodecoder: Fix different seqnum error
9e4e4a3: audiodecoder: Fix different seqnum error
13f2638: base: audioencoder: Remove fixed caps from srcpad
9379c4a: exiftag: Unmap buffer if parsing a rational number gives a
  zero denominator
87c7096: playback: Make sure to check for empty/any caps before
  getting the first structure
add1c1f: parsebin: Remove pointless casts
2a437f9: exiftag: Ignore invalid fractions with numerator/denominator
  G_MININT
9949c53: audio-resampler-neon: read array operand by hand
d33fc55: subparse: Replace a `g_strv_length()` in a loop header with
  more efficient code
a41cbb4: subparse: Replace regex string matching / replacing with
  plain C string parsing
3a81eb4: subparse: Fix handling of closing of multi-byte tags
12fc1ed: subparse: Avoid NULL-pointer dereferences in mdvdsub parsing
  code
7617d9e: typefind: Avoid signed 32 bit integer overflow when parsing
  AV1 LEB128 values
b86af52: typefind: Use a byte reader for parsing AV1 bitstreams
d9c7e28: typefind: Allow G_MAXUINT32 as LEB128 encoded value when
  parsing AV1 bitstreams
1f78c53: typefind: Remove dead code in AV1 LEB128 parser
66044db: exiftag: Use g_malloc() instead of malloc()
aeef929: exiftag: Add missing bounds check and integer overflow
  protections in various places
e47f697: exiftag: Parse sensitivity value with the correct number of
  bits and endianness
7f6db3d: subparse: Fix integer overflow when calculating qttext
  timestamp
5e78c44: audio-resampler-neon: read array operand by hand, part 2
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0: Upgrade 1.28.0 -> 1.28.2
Peter Kjellerstedt [Wed, 8 Apr 2026 14:17:09 +0000 (16:17 +0200)] 
gstreamer1.0: Upgrade 1.28.0 -> 1.28.2

Changes since 1.28.0
b2a3b2e: Back to development after 1.28.0
f8b4f2a: input-selector: fix race condition selecting next active pad
193b858: input-selector: gracefully handle active pad disappearance
52f8be4: input-selector: don't consider pad being released checking
  for eos
c711f7c: input-selector: fix regression receiving eos event
3d48df2: input-selector: don't wait for active pad when eos
  (active-segment mode)
583ae1a: typefindhelper: EOS case explained in more details
305fb65: gst: Fix a couple of new const-ness warnings around strstr()
  usage
88cdc0d: ptpclock: Fix potential out-of-bounds read when parsing
  ANNOUNCE messages
443fb64: ptpclock: Fix copy-paste mistake in error messages
c520284: ptpclock: Initialize PTP message to avoid invalid
  uninitialized variable warning
f8f0df1: caps: gst_caps_append_structure should not allow NULL
  structure
d08ede5: aggregator: Various introspection annotations / docs fixes
  for vfuncs
1a98767: filesink: Add debug category to file write helper
1f415d1: filesink: Report write error correctly on Windows
1aa6b48: multiqueue: reverse playback: use segment stop position as
  start time
c7e0e43: aggregator: Handle gap event before segment
816e87b: bitwriter: Steal owned data in reset_and_get_data()
7c2da22: tests: bitwriter: Fix leak in test
835da19: modules: Remove NEWS from git which is generated from full
  release notes
201b14e: modules: remove RELEASE from git, will be generated from
  template on dist
7694a7d: modules: remove subproject README.md from git
6376a84: modules: dist common files from monorepo root
291e479: meson: Deprecate `system = 'ios'` in cross files, use
  subsystem
d609133: gstcheck: Disable fork on tvOS and watchOS
55bea0a: gstreamer: Stop using deprecated G_MODULE_SUFFIX
020a8fd: registry: Skip .dSYM bundles when loading plugins, try 2
d8b40e3: gst-stats: Also allow ANSI colored logs without 0x in front
  of the thread id
c21305a: gst: Add explanatory comment to call_async implementation
dcb37e2: Release 1.28.1
32113a6: Back to development after 1.28.1
216d38a: all: GThreadFunc return type fixes
e76eebd: registry: Skip recursion into and from .dSYM bundles
9a94133: filesink: Fix wrong open() in overwrite mode
749cbe0: baseparse: Preserve upstream buffer duration if possible
f71eb59: bin: iterator is not nullable
1b3939b: baseparse: Fix out_buffer leak in frame_free and missing ref
  in frame_copy
1afd273: gstreamer: update translations
cf7231d: queue: Fix potential use-after-free in log function
43421c2: Release 1.28.2

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:23 +0000 (05:20 +0200)] 
gstreamer1.0: upgrade 1.26.9 -> 1.28.0

0002-tests-add-support-for-install-the-tests.patch
refreshed for 1.28.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-rtsp-server: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:22 +0000 (05:20 +0200)] 
gstreamer1.0-rtsp-server: upgrade 1.26.9 -> 1.28.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-python: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:21 +0000 (05:20 +0200)] 
gstreamer1.0-python: upgrade 1.26.9 -> 1.28.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-plugins-ugly: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:20 +0000 (05:20 +0200)] 
gstreamer1.0-plugins-ugly: upgrade 1.26.9 -> 1.28.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-plugins-good: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:19 +0000 (05:20 +0200)] 
gstreamer1.0-plugins-good: upgrade 1.26.9 -> 1.28.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-plugins-base: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:18 +0000 (05:20 +0200)] 
gstreamer1.0-plugins-base: upgrade 1.26.9 -> 1.28.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-plugins-bad: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:17 +0000 (05:20 +0200)] 
gstreamer1.0-plugins-bad: upgrade 1.26.9 -> 1.28.0

* The closedcaption plugin no longer depends on cairo and pango.
* Explicitly disable new plugins that require external dependencies:
  * hip
  * mpeghdec
  * tflite
  * tflite-edgetpu
  * tflite-vsi
  * vmaf
  * wpe2

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogstreamer1.0-libav: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:16 +0000 (05:20 +0200)] 
gstreamer1.0-libav: upgrade 1.26.9 -> 1.28.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogst-examples: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:15 +0000 (05:20 +0200)] 
gst-examples: upgrade 1.26.9 -> 1.28.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogst-devtools: upgrade 1.26.9 -> 1.28.0
Wang Mingyu [Wed, 8 Apr 2026 03:20:14 +0000 (05:20 +0200)] 
gst-devtools: upgrade 1.26.9 -> 1.28.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogrub: update search parameter
Xiangyu Chen [Thu, 26 Feb 2026 08:19:37 +0000 (16:19 +0800)] 
grub: update search parameter

Grub would report an error message in boot stage as below:

   "error: no such device: ((hd0,gpt1)/EFI/BOOT)/EFI/BOOT/grub.cfg"

Consequently, the root variable is not set, and the intended protection
against cross-device configuration loading (the purpose of the original 2014 commit)
is lost.

The most robust fix is to use the --hint parameter.
This separates the search target from the device hint, avoiding
fragile string concatenation and supporting both prefixed and
non-prefixed $cmdpath formats.

Fixes: 5ce73b6055ac ("grub: add cmdpath to grub configuration file")
Signed-off-by: Xiangyu Chen <xiangyu.chen@windriver.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agobusybox: enable mkfs.vfat via config fragment
Ricardo Salveti [Tue, 31 Mar 2026 22:36:02 +0000 (19:36 -0300)] 
busybox: enable mkfs.vfat via config fragment

Enable busybox’s mkfs.vfat (mkdosfs) implementation using a dedicated
mkdosfs.cfg config fragment.

While dosfstools provides mkfs.vfat and is available through
update-alternatives, it is licensed under GPLv3. Busybox also includes a
lightweight mkfs.vfat implementation that is suitable for minimal images
and systems relying primarily on busybox utilities.

Signed-off-by: Ricardo Salveti <ricardo.salveti@oss.qualcomm.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agobusybox: do not build SUID binary without an applet
Jörg Sommer [Tue, 31 Mar 2026 18:23:56 +0000 (20:23 +0200)] 
busybox: do not build SUID binary without an applet

If the merge of all config snippets leads to a SUID binary without any
applets, do not build and install it to reduce the SUID binaries in the
system.

Signed-off-by: Jörg Sommer <joerg.sommer@navimatix.de>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agobarebox: upgrade v2026.03.0 -> v2026.03.1
Yanis BINARD [Fri, 3 Apr 2026 13:30:58 +0000 (15:30 +0200)] 
barebox: upgrade v2026.03.0 -> v2026.03.1

CVE-2026-33243 has been fixed in v2026.03.0[0] but introduced
a regression. v2026.03.1 contains a single commit to fix it[1].

Changelog : https://github.com/barebox/barebox/compare/v2026.03.0...v2026.03.1
* FIT: handle hashed-nodes property not being in hashed nodes order

Release: https://github.com/barebox/barebox/releases/tag/v2026.03.1

[0]: https://github.com/barebox/barebox/commit/55f25be5223b662c6888d9d7ef07d28e0348d7db
[1]: https://lore.kernel.org/barebox/abljJRMecNdejSD0@pengutronix.de/

Signed-off-by: Yanis BINARD <yanis.binard@smile.fr>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agogroff: upgrade 1.23.0 -> 1.24.0
Zhang Peng [Fri, 3 Apr 2026 11:06:58 +0000 (19:06 +0800)] 
groff: upgrade 1.23.0 -> 1.24.0

Upgrade to latest revision of 1.24.0.
- Drop patches included in this release.
- Add patch to fix test-groff not found in cross-compilation.
- Add patch to use SOURCE_DATE_EPOCH in gropdf for reproducible PDF generation.
- COPYING, LICENSES: Refresh GPLv3 text.

Release Note:[https://lists.gnu.org/r/groff/2026-02/msg00149.html]

Signed-off-by: Zhang Peng <peng.zhang1.cn@windriver.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agowebkitgtk: upgrade 2.50.4 -> 2.50.6
Liu Yiding [Fri, 27 Mar 2026 01:05:40 +0000 (09:05 +0800)] 
webkitgtk: upgrade 2.50.4 -> 2.50.6

================
WebKitGTK 2.50.6
================

What's new in WebKitGTK 2.50.6?

  - Fix sample code included in the documentation of the
    webkit_user_content_manager_register_script_message_handler()
    function.
  - Fix MP4 muxing when using GStreamer 1.28.
  - Fix WebAudio not resuming correctly after using window.alert()
  - Fix WebAudio producing incorrect output in some cases due to
    incorrect sample buffer management.
  - Fix several crashes and rendering issues.

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>
4 days agocreaterepo-c: upgrade 1.2.2 -> 1.2.3
Moritz Haase [Fri, 27 Mar 2026 09:23:39 +0000 (10:23 +0100)] 
createrepo-c: upgrade 1.2.2 -> 1.2.3

Full list of changes in this release is available at [0].

[0]: https://github.com/rpm-software-management/createrepo_c/releases/tag/1.2.3

Signed-off-by: Moritz Haase <Moritz.Haase@bmw.de>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoblueprint-compiler: upgrade 0.20.0 -> 0.20.4
Liu Yiding [Thu, 2 Apr 2026 05:38:15 +0000 (13:38 +0800)] 
blueprint-compiler: upgrade 0.20.0 -> 0.20.4

Changelog:
  git://gitlab.gnome.org/GNOME/blueprint-compiler

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>
4 days agouki.bbclass: fix ukify arguments assembly
Pavel Löbl [Thu, 2 Apr 2026 07:33:15 +0000 (09:33 +0200)] 
uki.bbclass: fix ukify arguments assembly

Missing space before --uname ukify option mangles the command line
string when KERNEL_VERSION is set.

Signed-off-by: Pavel Löbl <pavel@loebl.cz>
Reviewed-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>
4 days agoicu: Workaround buildpaths QA check, if TMPDIR is /build/tmp
Jörg Sommer [Sat, 28 Mar 2026 21:04:18 +0000 (22:04 +0100)] 
icu: Workaround buildpaths QA check, if TMPDIR is /build/tmp

The file udatatst.c really contains the string /build/tmp. When building
with kas-container the build directory is /build and hence the TMPDIR
becomes /build/tmp. This triggers the QA check *buildpaths*. To work around
this, slightly modify the string in the source code.

% grep -n -C2 /build/tmp source/test/cintltst/udatatst.c
136-    char* path=(char*)malloc(sizeof(char) * (strlen(ctest_dataOutDir())
137-                                           + strlen(U_ICUDATA_NAME)
138:                                           + strlen("/build/tmp/..")+1 ) );
139-
140-    char        *icuDataFilePath = 0;

Signed-off-by: Jörg Sommer <joerg.sommer@navimatix.de>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoinitramfs-framework: init: fix kernel cmdline parsing
Michael Opdenacker [Thu, 26 Mar 2026 17:34:39 +0000 (17:34 +0000)] 
initramfs-framework: init: fix kernel cmdline parsing

Fix several issues with double quotes in kernel command line

- Kernel options like 'opt="value"' were breaking the parser,
  causing the whole reminder of the command line to be ignored.
  The code only supported 'opt="word1 word2..."

- Setting variables without removing quotes in the value

- Setting variables to values with spaces without enclosing
  the value with quotes. This caused execution errors evaluating
  expressions like:
  bootparam_opt=word1 word2

The first fix is particularly needed for people using the kernel
"bootconfig" configuration parameters to add options to the kernel
command line:

CONFIG_BOOT_CONFIG=y
CONFIG_BOOT_CONFIG_EMBED=y
CONFIG_BOOT_CONFIG_EMBED_FILE="additional-bootargs.bootconfig"

This mechanism systematically adds quotes around options
with values, for example:
init="/sbin/preinit"

Without the fix, the wrong init program can be started from the
initramfs and debug messages are ignored when "debug" is
present after "init" in the kernel command line.

For readability and performance sake, also use shell variable operators
instead of "sed" to remove leading and trailing quotes.

Tested both on host and target machines.
With the below kernel command line:
rootwait init="/sbin/preinit" debug root=/dev/mmcblk0p2 console=ttymxc0 dyndbg="file drivers/usb/core/hub.c +pltf" quiet

The following variables are set:
bootparam_rootwait="true"
bootparam_init="/sbin/preinit"
bootparam_root="/dev/mmcblk0p2"
bootparam_debug="true"
bootparam_console="ttymxc0"
bootparam_dyndbg="file drivers/usb/core/hub.c +pltf"
bootparam_quiet="true"

Signed-off-by: Michael Opdenacker <michael.opdenacker@rootcommit.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agosystemd: backport fix for tpm2 without efi support
Patrick Wicki [Thu, 26 Mar 2026 16:06:50 +0000 (17:06 +0100)] 
systemd: backport fix for tpm2 without efi support

See https://github.com/systemd/systemd/pull/41231 or the patch commit
message for details. The regression affects v259.1 to 259.5 as well as
v260.1. Requested a backport to v259 but for now a patch is needed.

Signed-off-by: Patrick Wicki <patrick.wicki@siemens.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoautoconf: Raise line length limit for sed/awk
Khem Raj [Fri, 3 Apr 2026 22:22:49 +0000 (23:22 +0100)] 
autoconf: Raise line length limit for sed/awk

Latest autoconf is splitting long cmd lines in config.status
it generates using sed/awk to have compatibility with
platform like HPUX,Solaris,AIX which is understandable given
autotools try to achieve wider portability, however, this does
mean that in OE builds we might get these newlines added in
places dividing commands including buildpaths in them and the
methods deployed to scrub them from output artifacts will fail
since the string is now on two lines resulting in build QA to
pass but the strings do sneak in affecting reproducibility
one such example is e2tools recipe in meta-filesystems and there
perhaps are more

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agorecipes: remove redundant autoreconf excludes
Ross Burton [Mon, 30 Mar 2026 13:28:25 +0000 (14:28 +0100)] 
recipes: remove redundant autoreconf excludes

With autoconf 2.73 autoreconf is better at not forcibly running aclocal
or autoheader if the relevant files are not managed by those tools.

This means we can remove a number of explicit excludes as they are no
longer needed.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoautoconf: upgrade to 2.73
Ross Burton [Mon, 30 Mar 2026 13:28:24 +0000 (14:28 +0100)] 
autoconf: upgrade to 2.73

man-host-perl.patch and autoreconf-exclude.patch are now upstream.

Remove a hunk from autotest-automake-result-format.patch that has no
purpose.

License-Update: postal address changed.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoghostscript: ensure the native compile uses C17
Ross Burton [Thu, 2 Apr 2026 13:41:43 +0000 (14:41 +0100)] 
ghostscript: ensure the native compile uses C17

GCC 15.2 defaults to C23:

  $ aarch64-poky-linux-gcc --version
  aarch64-poky-linux-gcc (GCC) 15.2.0
  $ aarch64-poky-linux-gcc -dM -E - </dev/null | grep STDC_VERSION
  #define __STDC_VERSION__ 202311L

As of ghostscript 10.06.0 the build is good with C23[1] and our cross
compiler is 15.2, so we can stop adding -std=gnu17 to the target CFLAGS.

However, our buildtools are currently using GCC 14.2 which has the
interesting behaviour that it has a -std=gnu23 option but does not then
actually claim to support C23:

  $ gcc --version
  gcc (GCC) 14.2.0
  $ gcc -std=gnu23 -dM -E - </dev/null | grep STDC_VERSION
  #define __STDC_VERSION__ 202000L

This then interacts badly with autoconf 2.73 which tells the compiler to
use C23, breaking stdpre.h which handles a number of different
compilers and C standards.

A bug has been filed with ghostscript[2] to try and handle this this
case in the source, but a workaround is to tell the native compiles to
explicitly use C17.

[1] ghostscript ae9409464 ("Bug 708160: Fix compatibility with C23 compilers")
[2] https://bugs.ghostscript.com/show_bug.cgi?id=709290

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoghostscript: do out-of-tree builds
Ross Burton [Thu, 2 Apr 2026 13:41:42 +0000 (14:41 +0100)] 
ghostscript: do out-of-tree builds

Upstream said that out-of-tree builds were meant to work, so I found the
few instances where path assumptions were incorrect and enabled it.

This gives us more reliable builds as we can always throw away the build
tree when reconfiguring.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agonative/cross: ar wrapper: don't prepend 'D' when @<file> is used
Martin Jansa [Thu, 2 Apr 2026 00:14:58 +0000 (02:14 +0200)] 
native/cross: ar wrapper: don't prepend 'D' when @<file> is used

In some cases bazel-native build calls ar with
@<file>      - read options from <file>
syntax, where the ar params are generated in *.param file by
separate bazel action before ar is called.

Unfortunately this intercept breaks that syntax by prepending D like:

$ ./ar-intercept @foo
argv before: ['./ar-intercept', '@foo']
argv after:  ['./ar-intercept', 'D@foo']
./ar-intercept: invalid option -- '@'
Usage: ./ar-intercept [emulation options] [-]{dmpqrstx}[abcDfilMNoOPsSTuvV] [--plugin <name>] [member-name] [count] archive-file file...
       ./ar-intercept -M [<mri-script]
...

or might trigger 'non-deterministic mode requested' error on false positive:

$ ./ar-intercept @Ufo
argv before: ['./ar-intercept', '@Ufo']
ar: non-deterministic mode requested
argv after:  ['./ar-intercept', '@Ufo']
./ar-intercept: invalid option -- '@'
Usage: ./ar-intercept [emulation options] [-]{dmpqrstx}[abcDfilMNoOPsSTuvV] [--plugin <name>] [member-name] [count] archive-file file...
       ./ar-intercept -M [<mri-script]

Don't try to inject D into first line of @<file>, lets assume the caller
uses D already (bazel uses 'rcsD' by default).

       * Ar flags for combining object files into archives. If this is not set, it
       * defaults to "rcsD".
       * TODO(b/37271982): Remove after blaze with ar action_config release

Signed-off-by: Martin Jansa <martin.jansa@gmail.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agomachine/arch-microblaze: Fix newlib builds
Kory Maincent [Wed, 1 Apr 2026 12:36:00 +0000 (14:36 +0200)] 
machine/arch-microblaze: Fix newlib builds

Fix the following errors for newlib libcs:
ld: unrecognized option '--hash-style=sysv'

Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoclang: Replace patch with solution merged upstream
Deepesh Varatharajan [Wed, 1 Apr 2026 12:07:45 +0000 (05:07 -0700)] 
clang: Replace patch with solution merged upstream

Drop 0039-Clang-Rename-OffloadArch-UNUSED-to-UNUSED_-to-avoid-.patch
Backport the equivalent upstream commit and remove the local patch.

https://github.com/llvm/llvm-project/commit/5bc3d1966b2d0b300c338d85bb38e356f0d12a8c

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>
4 days agolibgcc: Fix license with libgcc-src
Joshua Watt [Tue, 7 Apr 2026 19:06:47 +0000 (13:06 -0600)] 
libgcc: Fix license with libgcc-src

Sets the correct license for the libgcc-src package. Note that even
though the package is not in PACKAGES, it is implicitly added when
PACKAGE_DEBUG_SPLIT_STYLE is  "debug-with-srcpkg" (the default). This
was likely not noticed sooner because the fix to allow the code to
detect this was relatively recent in 1816c5a5e7 ("lib/package: Check
incompatible licenses at packaging time")

Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agobusybox: fix DISTRO_FEATURES detection
Ross Burton [Tue, 7 Apr 2026 16:23:37 +0000 (17:23 +0100)] 
busybox: fix DISTRO_FEATURES detection

Changes to how DISTRO_FEATURES is evaluated meant that busybox was using
a partial DISTRO_FEATURES and thus disabling features it should not be.
This is due to the use of immediate assignments which are evaulated
before the final value is calculated (by anonymous Python in base.bbclass).

Remove entirely DO_IPv4/DO_IPv6 as they're used once, replacing with
inline bb.utils.contains_any().

Remove the immediate assignments to OE_FEATURES/OE_DEL so that they are
evaluated on use. The evaluation is expensive, but it only happens once.

[1] oe-core 159148f4de ("meta: Support opting out of any distro features")

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoclasses/sbom-cve-check-recipe: add recipe-scanning class
Ross Burton [Tue, 7 Apr 2026 21:02:26 +0000 (22:02 +0100)] 
classes/sbom-cve-check-recipe: add recipe-scanning class

Add a second sbom-cve-check class that scans the recipe-SBOM, unlike
sbom-cve-check.bbclass which scans specifically an image's SBOM.

In most cases, image-derived SBOM scanning is preferred as this is able
to analyse the actual content of the image, e.g. reporting just the
kernel issues that have actually been compiled into the image (for
linux-yocto 6.18.19, this reduces the CVE count from 108 to 52).

However, for metrics or other specific purposes it might be useful to
scan a recipe-SPDX, so add a class that can do this.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoclasses/sbom-cve-check: extract common functionality to a -common.bbclass
Ross Burton [Tue, 7 Apr 2026 21:02:25 +0000 (22:02 +0100)] 
classes/sbom-cve-check: extract common functionality to a -common.bbclass

Extract the common configuration variables and run_sbom_cve_check() that
actually invokes sbom-cve-check to a separate class, so that other
classes that are not tied to image generation can use the same logic.

No code changes, just movement.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoclasses/sbom-cve-check: extract the database dependencies to a variable
Ross Burton [Tue, 7 Apr 2026 21:02:24 +0000 (22:02 +0100)] 
classes/sbom-cve-check: extract the database dependencies to a variable

This tidies up the dependencies, allowing for future tasks that will
need the same dependencies.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agoclasses/sbom-cve-check: refactor do_sbom_cve_check
Ross Burton [Tue, 7 Apr 2026 21:02:23 +0000 (22:02 +0100)] 
classes/sbom-cve-check: refactor do_sbom_cve_check

Extract the bulk of the logic to a separate function, so the task just
has to pass a few variables.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agosbom-cve-check-config.inc: move to meta/conf/
Ross Burton [Tue, 7 Apr 2026 21:02:22 +0000 (22:02 +0100)] 
sbom-cve-check-config.inc: move to meta/conf/

This file is used by classes and recipes, so it's best to put it into
conf/ instead of alongside the recipe.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agosbom-cve-check-update-db: Fix unpack removing other databases
Benjamin Robin [Tue, 7 Apr 2026 21:02:21 +0000 (22:02 +0100)] 
sbom-cve-check-update-db: Fix unpack removing other databases

Previously, `UNPACKDIR` was set to `${SBOM_CVE_CHECK_DEPLOY_DB_DIR}`,
which points to the `./databases/` directory and may contain multiple
databases. Since `do_unpack` cleans the `${UNPACKDIR}` directory,
`UNPACKDIR` must only contain a single database.

To address this, set `UNPACKDIR` to
`${SBOM_CVE_CHECK_DEPLOY_DB_DIR}/${SBOM_CVE_CHECK_DB_NAME}` and
configure `destsuffix` to an empty string. However, it is currently not
possible to set `BB_GIT_DEFAULT_DESTSUFFIX` to an empty string, so
`destsuffix` is configured through the SRC_URI instead.

Signed-off-by: Benjamin Robin <benjamin.robin@bootlin.com>
Tested-by: Antonin Godard <antonin.godard@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
4 days agosbom-cve-check-update-db: Fix do_populate_lic failure
Benjamin Robin [Tue, 7 Apr 2026 21:02:20 +0000 (22:02 +0100)] 
sbom-cve-check-update-db: Fix do_populate_lic failure

The `do_populate_lic` task was failing because it runs after `do_patch`,
but the indirect dependency on `do_unpack` was lost when the `do_patch`
task was deleted.

It is safe, and even preferable, to retain the `do_patch` task, as this
allows users to apply patches to their CVE databases.

Follow-up to commit 8ef22ad9e302f86b2da4fa81541a464e95b9ef3c
("sbom-cve-check: Add class for post-build CVE analysis").

Signed-off-by: Benjamin Robin <benjamin.robin@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
6 days agoimage_types_wic: Simplify search path
Richard Purdie [Sun, 5 Apr 2026 07:38:00 +0000 (08:38 +0100)] 
image_types_wic: Simplify search path

Standardise on one location by default for wic/wks files and simplify the
search path accordingly.

wic is now using BBPATH instead of BBLAYERS so this needs to be added to
WICVARS too.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
6 days agowic: move canned *wks files
Trevor Woerner [Fri, 3 Apr 2026 18:35:40 +0000 (14:35 -0400)] 
wic: move canned *wks files

When "wic create ..." is invoked with a bare *wks name (i.e. without the
`.wks` extension), wic calls engine.py:find_canned_images() to find the
fully qualified *wks file. This function searches every directory formed by:
    - permutating all BBLAYERS with `/wic`
    - permutating all BBLAYERS with `/scripts/lib/wic/canned-wks`
    - checking `<scripts_path>/lib/wic/canned-wks`
Where `<scripts_path>` is the directory containing the wic program.

When wic is part of oe-core, the last search path succeeds in finding
the canned *wks files in `<topdir>/scripts/lib/wic/canned-wks` (since
the wic program is found in oe-core's `<topdir>/scripts` directory, and
`<topdir>/scripts` is not a BBLAYER).

However, once wic is removed from oe-core, this algorithm will not find
these bare *wks files in any of the above-mentioned search paths since
the oe-core layer will no longer be the home of the wic program, and the
canned *wks files are not located in any directory relative to BBLAYERS.

Since these *wks files are specific to oe-core's meta layer, they should
stay with this layer. Therefore move the *wks files so they exist in one
of the locations searched relative to oe-core/meta's BBLAYERS.

NOTE: this patch is in preparation for removing wic from oe-core
      the wic oe-selftests work fine with this patch being added now

[RP: Updated meta/wic -> meta/files/wic]
Reviewed-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Reviewed-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>