libinput: upgrade 1.24.0 -> 1.25.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

iso-codes: upgrade 4.15.0 -> 4.16.0

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

ethtool: upgrade 6.6 -> 6.7

avoid_parallel_tests.patch
refreshed for 6.7

Changelog:
===========
* Feature: support for setting TCP data split
* Fix: fix new gcc14 warning
* Fix: fix SFF-8472 transceiver module identification (-m)
* Misc: code cleanup

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cpio: upgrade 2.14 -> 2.15

Changelog:
==========
* Fix operation of --no-absolute-filenames --make-directories
* Restore access and modification times of symlinks in copy-in
  and copy-pass modes.

0001-configure-Include-needed-header-for-major-minor-macr.patch
revmoed since it's included in 2.15

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

at-spi2-core: upgrade 2.50.0 -> 2.50.1

Changelog:
 atk-adaptor: Fix critical when no table cell array is returned.

Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

eudev: allow for predictable network interface names

Do not suppress renaming by eudev if 'pni-names' is
a distro feature.  The previous behavior was to always
suppress renaming.

eudev is used when sysvinit is the init manager.

Signed-off-by: Joe Slater <joe.slater@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa: Upgrade 23.3.4 -> 23.3.5

Mesa 23.3.5 release notes:

https://lists.freedesktop.org/archives/mesa-announce/2024-February/000748.html

Signed-off-by: Fabio Estevam <festevam@denx.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

numactl: upgrade 2.0.16 -> 2.0.17

Drop two backported patches, refresh others.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pseudo: Update to pull in gcc14 fix and missing statvfs64 intercept

rpm 4.19 now builds with LFS64 support enabled by default,
so it calls statvfs64() to get the space available on the
filesystem it is installing packages into.  This is not
getting caught by pseudo, so rpm is checking the host's
root filesystem, rather than the filesystem where the
build is happening.

Merge in that fix and a gcc14 fix.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

maintainers.inc: Add self for libseccomp and gnutls

Signed-off-by: Simone Weiß <simone.p.weiss@posteo.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rpm: fix dependency for package config imaevm

The dependency `ima-evm-utils` of package config `imaevm` has been
removed during rpm upgrade. Add it back, otherwise it fails to run
do_configure when the package config `imaevm` enabled.

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel-devsrc: RDEPENDS on gawk

In reality all architectures have been depending on awk / gawk for
significant time.

gawk directly since v5.10 as it introduced scripts/check-sysctl-docs

grep for /usr/bin/awk results in

* arch/x86/tools/gen-insn-attr-x86.awk:#!/usr/bin/awk -f
* scripts/ver_linux:#!/usr/bin/awk -f
* tools/arch/x86/tools/gen-insn-attr-86.awk:#!/usr/bin/awk -f

So instead of relying on other recipes adding gawk - RDEPEND on it.

Signed-off-by: Claus Stovgaard <claus.stovgaard@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel-devsrc: fix RDEPENDS for make

Since kernel version 6.6 the debian package rules has been split up into
a separate rules file, resulting in a runtime requirement for make.
See scripts/package/debian/rules

Remove the rules file, for not RDEPENDS on make for something we are
not using for building modules.

For reproducing the issue on a normal qemux86-64 machine, change to the
linux-yocto-dev kernel, add the kernel-devsrc to the toolchain target
task and disable ptest in distro features. (Notice ptest adds make as
RDEPENDS for other packages hiding the issue)

when populating sdk you will see dnf failing.

 Problem: conflicting requests
  - nothing provides /usr/bin/make needed by kernel-devsrc-1.0-r0.qemux86_64 from oe-repo

Signed-off-by: Claus Stovgaard <claus.stovgaard@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tune-cortexr52: Add hard-float variant

Add the ability to specify cortexr52hf to get a version that enables the
hard-float calling convention.

Signed-off-by: Mark Hatle <mark.hatle@amd.com>
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tune-cortexr5: Add hard-float variant

Add the ability to specify cortexr5hf to get a version that enables the
hard-float calling convention.

Signed-off-by: Mark Hatle <mark.hatle@amd.com>
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

feature-arm-vfp.inc: Allow hard-float on newer simd targets

Currently hard-float is dependency upon the special TUNE_CCARGS_MFPU which
contains custom -mfpu= values.  However, newer 32-bit architectures like
cortex-r52 use 'simd' instead.  There is no 'simd' entry for -mfpu=
according to the GCC manual, it's more or less automatic based on the
cpu settings.

Add this as an exception to the TUNE_CCARGS_FLOAT setting of hard or softfp.

Signed-off-by: Mark Hatle <mark.hatle@amd.com>
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tune-cortexa78.inc: Add cortexa78 tune, based on cortexa77

Signed-off-by: Mark Hatle <mark.hatle@amd.com>
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rpm: override curl executable search with just 'curl'

rpm is searching for curl executable at do_configure time
and result is non-deterministic, depending on whether curl
is in native sysroot or not:
https://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20240201-8dcic6ea/packages/diff-html/

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto-dev: update to v6.8

-rc1 is available, so we bump our -dev kernel to track.

Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lttng-modules: fix v6.8+ build

linux-yocto-dev is updating to 6.8 and lttng-modules needs to be updated
to account for upsream kernel fixes.

These patches are all grabbed from the lttng review gerrit and
backported to lttng-modules 2.13.0.

Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

setftest/cdn tests: check for exceptions also in fetcher diagnostics

Otherwise the output was cluttered with irrelevant lines that were
describing missing, but excepted cache objects.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

u-boot: add missing dependency on pyelftools-native

When TF-A is necessary in U-Boot binary, binman requires elftools to be
installed to be able to generate that U-boot ITB image.

TF-A is necessary for at least all Aarch64 machines, so that is a
non-negligible amount of boards that have this requirement.

As a side note, Rockchip-based machines didn't need this until commit
12c3e948eeab ("rockchip: Drop the FIT generator script") (v2023.04-rc1).

This is already in meta-rockchip, c.f.
https://git.yoctoproject.org/meta-rockchip/commit/recipes-bsp/u-boot?id=6127d169acf239a53df989f34a6b825fa182cc0c
but I feel like this makes more sense to be present in OE-Core.

Cc: Quentin Schulz <foss+yocto@0leil.net>
Signed-off-by: Quentin Schulz <quentin.schulz@theobroma-systems.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

opensbi: append LDFLAGS to TARGET_CC_ARCH

To solve the following GNU_HASH error:
ERROR: opensbi-1.4-r0 do_package_qa: QA Issue: File /share/opensbi/lp64/generic/firmware/fw_dynamic.elf in package opensbi doesn't have GNU_HASH (didn't pass LDFLAGS?)
File /share/opensbi/lp64/generic/firmware/fw_payload.elf in package opensbi doesn't have GNU_HASH (didn't pass LDFLAGS?)
File /share/opensbi/lp64/generic/firmware/fw_jump.elf in package opensbi doesn't have GNU_HASH (didn't pass LDFLAGS?) [ldflags]

[YOCTO #15370] -- https://bugzilla.yoctoproject.org/show_bug.cgi?id=15370

Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

openssl: backport fix for CVE-2023-6129

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/package_rpm: correctly escape percent characters

This many characters doesn't work with rpm 4.19 packaging
(as shown by nodejs recipes), and per documentation a single escape
is enough:

https://github.com/rpm-software-management/rpm/blob/rpm-4.19.x/docs/manual/spec.md#shell-globbing

It also should be done in a function, and just before writing out the
corrected filename to .spec, not earlier where the path may still
be needed for file operations (such as gettings file attributes).

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

grub2: ignore CVE-2023-4001, this is Red Hat-specific

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dropbear: remove unnecessary line

This line was added in 703e3faaec8c5a22, however
the default value of FILES:${PN} contains "${bindir}/*", so this directory does
not need to be explicitly added.

Signed-off-by: Etienne Cordonnier <ecordonnier@snap.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa: Upgrade 23.3.3 -> 23.3.4

Mesa 23.3.4 release notes:

https://lists.freedesktop.org/archives/mesa-announce/2024-January/000745.html

Signed-off-by: Fabio Estevam <festevam@denx.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

initscripts: Add custom mount args for /var/lib

Adds bitbake variable to set additional mount flags for the /var/lib
overlayfs or bind mount when using a read-only root filesystem. This
can be used to set additional options like "-o nodev".

Signed-off-by: Colin McAllister <colin.mcallister@garmin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

armv8/armv9: Avoid using -march when -mcpu is chosen

Current include logic goes into generic arm v8/v9 architecture tunes and
adds corresponding -march option after synthesizing it from various tune
fragments, this is fine for a machine which is using armv8/armv9 based
tunes but cortex tunes are intentionally using -mcpu option based on
selected tune value. So when cortex based default tune is selected for a
machine, it will add both -mcpu and -march to the compiler commandline
which can result in invalid combinations for this pair in gcc's own
logic. This can then result in compiler warnings/errors reporting this

e.g.

aarch64-yoe-linux-gcc  -mcpu=cortex-a72.cortex-a53 -march=armv8-a+crc+crypto -mbranch-protection=standard
...
cc1: error: switch '-mcpu=cortex-a72.cortex-a53' conflicts with '-march=armv8-a+crc+crypto' switch and resulted in options '+crc+crypto' being added [-Werror]

This is seen in lot of configure test results in glibc 2.39 and the
warning is promoted to errors by gcc in some of these checks especially
with gcc-14, the logs also show it as warning in other places in
configure checks.

mcpu option will compute relevant march implicitly as it specifies a cpu
implementation and this will be the right value to use, therefore do not
specify -march when -mcpu is already describing the cpu.

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Ross Burton <ross.burton@arm.com>
Cc: Jon Mason <jdmason@kudzu.us>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemu: Allow native and nativesdk versions on Linux older then 4.17

Linux kernel 4.17 introduced two new mmap flags, MAP_FIXED_NOREPLACE and
MAP_SHARED_VALIDATE.  Starting with QEMU 8.1, these flags are now used
and required for proper system operation.  In order to build and run on a
system older then 4.17, we need to emulate this new behavior.

Not having a newer kernel could result in the mmap memory being allocated
in a way that will cause failures without QEMU checking for these
conditions.  Note, memory allocation issues are rare in my experience so
this is more of a 'just-in-case' behavior.

SDK_OLDEST_KERNEL is currently set to 3.2.0, the only way this can claim
that qemu works in an SDK is by checking the return values to emulate
the expected behavior.

Signed-off-by: Mark Hatle <mark.hatle@amd.com>
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

go: update 1.20.12 -> 1.20.13

Upgrade to latest 1.20.x release [1]:

$ git log --oneline go1.20.12..go1.20.13
a95136a88c (tag: go1.20.13) [release-branch.go1.20] go1.20.13
5c38c04957 [release-branch.go1.20] crypto/tls: align FIPS-only mode with BoringSSL policy
9e4abed39b [release-branch.go1.20] crypto/internal/boring: upgrade module to fips-20220613
95afc744a7 [release-branch.go1.20] runtime: put ReadMemStats debug assertions behind a double-check mode
8cb86b5f85 [release-branch.go1.20] os/signal: skip nohup tests on darwin builders
59ffd3b90d [release-branch.go1.20] os/signal: remove go t.Run from TestNohup

[1] https://github.com/golang/go/compare/go1.20.12...go1.20.13

Signed-off-by: Jose Quaresma <jose.quaresma@foundries.io>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

devtool: refactor deploy-target

Make the deploy function independent from d. This allows to call the
function also from Python code not running in bitbake.
This is needed to for the devtool ide plugin which will call the
do_install task and the code from devtool deploy-target independently
from a bitbake server. This allows a much quicker workflow.

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

devtool: refactor deploy to use exec_fakeroot_no_d

This is a step towards a deploy function which can be called without
passing the d variable.

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

devtool: refactor exec_fakeroot

Provide a function exec_fakeroot_no_d which does the same like
exec_fakeroot does, but is usable independenlty from bitbake. This
allows to use the fanction from scripts where the d variable is not
available.

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

devtool: support plugins with plugins

Pass the plugins search path to plugins via context. This allows plugins
to search for their own plugins at the same paths.

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cmake.bbclass: use --install

Since version 3.15 CMake provides a command-line signature to install an
already-generated project binary tree. This may be used after building a
project to run installation without using the generated build system or
the native build tool.

This is a small improvement, for regular bitbake calls. CMake does not
check the dependencies again which is expected to be faster.

The main motivation for this change is using CMake from an SDK context.
With this change it is possible to initiate the compile step from an
IDE and later on initiating the install step via bitbake which runs the
install step on pseudo.

This is also what the meson.bbclass already does with the --no-rebuild
option.

Signed-off-by: Adrian Freihofer <adrian.freihofer@siemens.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnutls: Upgrade 3.8.2 -> 3.8.3

Upgrade version to adress recent CVE findings.

Changelog
=========
** libgnutls: Fix more timing side-channel inside RSA-PSK key exchange
   [GNUTLS-SA-2024-01-14, CVSS: medium] [CVE-2024-0553]

** libgnutls: Fix assertion failure when verifying a certificate chain with a
   cycle of cross signatures
   [GNUTLS-SA-2024-01-09, CVSS: medium] [CVE-2024-0567]

** libgnutls: Fix regression in handling Ed25519 keys stored in PKCS#11 token
   certtool was unable to handle Ed25519 keys generated on PKCS#11
   with pkcs11-tool (OpenSC). This is a regression introduced in 3.8.2.

Signed-off-by: Simone Weiß <simone.p.weiss@posteo.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xwayland: Upgrade 23.2.3 -> 23.2.4

Cumulative Fixes for CVE-2023-6816, CVE-2024-0229, CVE-2024-21885, CVE-2024-21886,
CVE-2024-0408, CVE-2024-0409

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3-yamllint: add missing dependency

yamllint requires pathspec module to be available

Signed-off-by: Konrad Weihmann <kweihmann@outlook.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemu: fix target build with ccache enabled

* with ccache inheritted, the BUILD_CC is 'ccache gcc', but because of missing
  quote it ends passing just ccache to host-cc which gets stripped and then
  it calls compiler[0] on empty compiler variable and breaks meson as shown in:
  http://errors.yoctoproject.org/Errors/Details/751436/

python version: Python 3.12.1
Traceback (most recent call last):
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/mesonmain.py", line 194, in run
    return options.run_func(options)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/msetup.py", line 358, in run
    app.generate()
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/msetup.py", line 181, in generate
    return self._generate(env, capture, vslite_ctx)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/msetup.py", line 203, in _generate
    intr = interpreter.Interpreter(b, user_defined_options=user_defined_options)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreter/interpreter.py", line 331, in __init__
    self.parse_project()
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreterbase/interpreterbase.py", line 132, in parse_project
    self.evaluate_codeblock(self.ast, end=1)
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreterbase/interpreterbase.py", line 198, in evaluate_codeblock
    raise e
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreterbase/interpreterbase.py", line 190, in evaluate_codeblock
    self.evaluate_statement(cur)
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreterbase/interpreterbase.py", line 204, in evaluate_statement
    return self.function_call(cur)
           ^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreterbase/interpreterbase.py", line 530, in function_call
    res = func(node, func_args, kwargs)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreterbase/decorators.py", line 260, in wrapper
    return f(*nargs, **wrapped_kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreterbase/decorators.py", line 579, in wrapper
    return f(*wrapped_args, **wrapped_kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreter/interpreter.py", line 1309, in func_project
    self.add_languages(proj_langs, False, MachineChoice.BUILD)
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreter/interpreter.py", line 1476, in add_languages
    success = self.add_languages_for(args, required, for_machine)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/interpreter/interpreter.py", line 1519, in add_languages_for
    comp = compilers.detect_compiler_for(self.environment, lang, for_machine, skip_sanity_check)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/compilers/detect.py", line 115, in detect_compiler_for
    comp = compiler_from_language(env, lang, for_machine)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/compilers/detect.py", line 112, in compiler_from_language
    return lang_map[lang](env, for_machine) if lang in lang_map else None
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/compilers/detect.py", line 614, in detect_c_compiler
    return _detect_c_or_cpp_compiler(env, 'c', for_machine)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot-native/usr/lib/python3.12/site-packages/mesonbuild/compilers/detect.py", line 286, in _detect_c_or_cpp_compiler
    compiler_name = os.path.basename(compiler[0])
                                     ~~~~~~~~^^^
IndexError: list index out of range

...

Project name: qemu
Project version: 8.2.0
C compiler for the host machine: ccache x86_64-oe-linux-clang -target x86_64-oe-linux -m64 -march=core2 -mtune=core2 -msse3 -mfpmath=sse -mlittle-endian --dyld-prefix=/usr -Qunused-arguments --sysroot=TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot -m64 -mcx16 (clang 17.0.6 "clang version 17.0.6 (https://github.com/llvm/llvm-project 6009708b4367171ccdbf4b5905cb6a803753fe18)")
C linker for the host machine: x86_64-oe-linux-clang -target x86_64-oe-linux -m64 -march=core2 -mtune=core2 -msse3 -mfpmath=sse -mlittle-endian --dyld-prefix=/usr -Qunused-arguments --sysroot=TOPDIR/tmp-glibc/work/core2-64-oe-linux/qemu/8.2.0/recipe-sysroot -m64 -mcx16 ld.lld 17.0.6

../qemu-8.2.0/meson.build:1:0: ERROR: Unhandled python exception

    This is a Meson bug and should be reported!

ERROR: meson setup failed

Signed-off-by: Martin Jansa <martin.jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/qemurunner: Handle rare shutdown race

The pid file can disappear when qemu is shutting down leading to a
file not found race before it is read.

Tweak the code to handle this and fix a rare but annoying race error
case.

[YOCTO #15036]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemu: Replace the basename patch with backport

Backport the patch that got applied to fix this issue upstream

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

piglit: Update to latest revision

Update to the latest piglit revision.

Signed-off-by: Fabio Estevam <festevam@denx.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3: Initialize struct termios before calling tcgetattr()

Fixes ptest failures on python 3.12

    Failed ptests:
    {'python3': ['test_tcgetattr',
                 'test_cfmakeraw',
                 'test_setcbreak',
                 'test_setraw',
                 'python3']}

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lttng-modules: upgrade 2.13.10 -> 2.13.11

Changelog:
https://lwn.net/Articles/957396/

Drop 2 backported patches and 1 implemented differently upstream.

Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

elfutils: Fix build with gcc trunk

Fixes a build issue seen with latest gcc trunk.

Fixes
| ../../elfutils-0.189/tests/elfstrmerge.c: In function 'main':
| ../../elfutils-0.189/tests/elfstrmerge.c:450:32: error: 'calloc' sizes specified with 'sizeof' in the earlier argument and not in the later argument [-Werror=calloc-transposed-args]
|   450 |   newscnbufs = calloc (sizeof (void *), newshnums);
|       |                                ^~~~
| ../../elfutils-0.189/tests/elfstrmerge.c:450:32: note: earlier argument should specify number of elements, later size of each element
| cc1: all warnings being treated as errors

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xz: Add ptest support

There are two types of cases: executables and POSIX shell scripts.
All test cases PASS.
Add xz-ptest to PTESTS_FAST because test duration less than 30s
on qemux86-64.

root@qemux86-64:~# ptest-runner xz
START: ptest-runner
2024-01-26T03:32
BEGIN: /usr/lib/xz/ptest
=== test_bcj_exact_size.c ===
PASS: test_exact_size
PASS: test_empty_block

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gtk4: update 4.12.4 -> 4.12.5

Overview of Changes in 4.12.5, 17-01-2024
=========================================

* GtkColumnView:
 - Fix a crash on dispose

* GtkEmojiChooser:
 - Update to CLDR v44
 - Add more translations

* GtkFileDialog:
 - Return an error if no file is selected
 - Make closing the portal file chooser work

* GtkDropDown:
 - Fix display of the initial checkmark

* GtkShortcutsWindow:
 - Reduce the minimum width

* GDK:
 - Make the png loader safer against overflow

* Windows:
 - Use new clipboard api

* Wayland:
 - Fix cursor handling with graphics tablets

* macOS:
 - Silence secure-restore message
 - Fix build on macOS < 10.13

* Translation updates
 Basque
 British English
 Catalan
 Chinese (China)
 Czech
 Georgian
 German
 Hebrew
 Indonesian
 Lithuanian
 Persian
 Polish
 Romanian
 Russian
 Slovenian
 Swedish
 Turkish
 Ukrainian
 Vietnamese

Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rpm: update 4.18.1 -> 4.19.1

Upstream has replaced autoconf with cmake, which necessitates a rewrite of the
recipe and available options, and a rebase to cmake of
0001-Do-not-hardcode-lib-rpm-as-the-installation-path-for.patch

Correct a mistake in 0001-Do-not-read-config-files-from-HOME.patch :
the patch was removing the NULL marker at the end of function arguments,
and 0002-Add-support-for-prefixing-etc-from-RPM_ETCCONFIGDIR-.patch
was restoring it (in addition to the actual change the patch was making).
Now both patches preserve the NULL terminator.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/runtime/rpm: fail tests if test rpm file cannot be found

Discovery of the test file was happening in a class initializer.
That block of code cannot fail (it's not a test), and so it
falls through to completion even if the needed file could not be found.

Then the tests themselves fail later due to class variables not
being set, but all information as to why is already lost at that point.

This converts the discovery to a helper function called from
the tests, so that the function can fail the tests precisely when the
problems occur.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/package_rpm: set bogus locations for passwd/group files

Since https://github.com/rpm-software-management/rpm/commit/f3eaeeb7341085e1850e914350cf1f33d538320d
rpm does its own parsing of /etc/passwd and /etc/group instead of relying on getpwnam() and friends.

This has an unfortunate effect of leaking build host uid/gid values for users and groups
into the cpio header inside rpm file (where previously those were always zero).

Installation of rpm packages relies on rpm header to set files ownership, and that
is a different structure that is build from .spec information, so we can avoid host
contamination by setting the paths to something bogus.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/package_rpm: use weak user/group dependencies

rpm 4.19 automatically generates provides and depends for user and
groups:
https://github.com/rpm-software-management/rpm/blob/rpm-4.19.x/docs/manual/users_and_groups.md#dependencies

This mechanism relies on sysusers.d for the 'provides'
part, and thus is systemd-only at best. So we need to disable it for now,
otherwise image generation fails with unresolved dependencies.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/package_rpm: write file permissions and ownership explicitly into .spec

Per https://github.com/rpm-software-management/rpm/commit/77d3529c31ca090a40b8d3959a0bcdd721a556d6
rpm 4.19.1+ will not consider actual filesystem permissions and ownership, and will quietly default
to root if not expictly set otherwise in .spec file.

There's also additional diagnostics (printing what is in passwd/group)
when user/group name lookup against the sysroot fails.
That is never supposed to happen, and yet there was one report that it did:
https://autobuilder.yoctoproject.org/typhoon/#/builders/44/builds/8493/steps/23/logs/stdio

Investigating that issue led to the first three commits in this patchset:

sysroot user management postinsts: run with /bin/sh -e to report errors when they happen
classes/multilib: expand PACKAGE_WRITE_DEPS in addition to DEPENDS
classes/staging: capture output of sysroot postinsts into logs

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/staging: capture output of sysroot postinsts into logs

This particularly helps with user management postinsts as otherwise
there's no trace left of what was run, in which order, and what was the output.

Here's an example from the logs:

NOTE: Running postinst /srv/storage/alex/yocto/build-multilib/tmp/work/x86_64-pokymllib64-linux/lib64-quilt/0.67/lib64-recipe-sysroot/usr/bin/postinst-lib64-base-passwd, output:
b'/srv/storage/alex/yocto/build-multilib/tmp/work/x86_64-pokymllib64-linux/lib64-quilt/0.67/recipe-sysroot-native/usr/sbin/useradd\nRunning useradd commands...\nNOTE: lib64-ptest-runner: Performing useradd with [--root /srv/storage/alex/yocto/build-multilib/tmp/work/x86_64-pokymllib64-linux/lib64-quilt/0.67/lib64-recipe-sysroot --system --no-create-home --home / --user-group ptest]\n'
NOTE: Running postinst /srv/storage/alex/yocto/build-multilib/tmp/work/x86_64-pokymllib64-linux/lib64-quilt/0.67/lib64-recipe-sysroot/usr/bin/postinst-useradd-lib64-ptest-runner, output:
b'/srv/storage/alex/yocto/build-multilib/tmp/work/x86_64-pokymllib64-linux/lib64-quilt/0.67/recipe-sysroot-native/usr/sbin/useradd\nRunning useradd commands...\nNOTE: lib64-ptest-runner: Performing useradd with [--root /srv/storage/alex/yocto/build-multilib/tmp/work/x86_64-pokymllib64-linux/lib64-quilt/0.67/lib64-recipe-sysroot --system --no-create-home --home / --user-group ptest]\nNOTE: lib64-ptest-runner: user ptest already exists, not re-creating it\n'

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/multilib: expand PACKAGE_WRITE_DEPS in addition to DEPENDS

Otherwise, PACKAGE_WRITE_DEPS would contain non-multilib variants
of dependencies even when building multilib items, resulting in
sysroots being populated with entirely wrong versions of them.

This hasn't been noticed until now through sheer (bad) luck, I think,
except in the cpio recipe, but the previous commit shows that the issues
did occur, quietly. Every other recipe in oe-core and meta-oe does not
prepend the multilib prefix.

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sysroot user management postinsts: run with /bin/sh -e to report errors when they happen

This exposes the following failure in a multilib setup, when everything
up to do_package_write_rpm is in sstate, but do_package_write_rpm is not
(there's a similar fail for lib64-man-db, and the failures themselves will be fixed separately
in the next commit):

Exception: subprocess.CalledProcessError: Command '/srv/storage/alex/yocto/build-multilib/tmp/work/x86_64-pokymllib64-linux/lib64-quilt/0.67/recipe-sysroot/usr/bin/postinst-base-passwd' returned non-zero exit status 1.

Subprocess output:
install: cannot stat '/srv/storage/alex/yocto/build-multilib/tmp/work/x86_64-pokymllib64-linux/lib64-quilt/0.67/lib64-recipe-sysroot/usr/share/base-passwd/passwd.master': No such file or directory

Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3-numpy: Use Large File Support version of fallocate

This was supposed to always be the case from upstream but was exposed
by the 64-bit-time QA checker when file function scanning was enabled.

Signed-off-by: Ola x Nilsson <olani@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

native.bbclass: base_libdir unique from libdir

Use STAGING_BASE_LIBDIR_NATIVE for the value of base_libdir instead of
STAGING_LIBDIR_NATIVE. This will avoid conflicts between the two
directories.

Signed-off-by: William Hauser <william.hauser@meraki.net>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sanity: require bitbake 2.7.2 for the inherit_defer statement

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs-postcommands.bbclass: ignore comment mismatch in systemd_user_check

The check was forcing every field to be identical, this is too
strict. For example, if the comment differs, there's really no
impact. For example, root user's comment is 'root' in passwd,
and it's 'Super User' in sysusers.d/basic.conf. Such difference is
not worth a warning. In fact, previous codes use 'lower()' to
avoid warning between 'nobody' and 'Nobody', and what's more, another
patch puts its own basic.conf.in in systemd's SRC_URI, but it changes
'Super User' to 'root'. Such changes are all unnecessary. We should
just ignore comment mismatch.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd-boot: upgrade to 255.1

0030-meson-Pass-all-static-pie-args-to-linker.patch is dropped as
it's a backported patch and is now in the new 255.1 version.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: upgrade to 255.1

1. Patch changes:

0004-Move-sysusers.d-sysctl.d-binfmt.d-modules-load.d-to-.patch
is removed because it has no real effect now. The /lib is now
/usr/lib because 'usrmerge' is a required distro feature for systemd.

0002-binfmt-Don-t-install-dependency-links-at-install-tim.patch is
refreshed for the new version to avoid patch-fuzz issue.

2. root user's home directory now defaults to "/root":

The sysuers.d/basic.conf is still modified to respect the ROOT_HOME
value, so if users set ROOT_HOME to "/home/root", the behavior is the
same as before. However, this is only for backward compatibility. With
this patch, The ROOT_HOME value is set to "/root" in init-manager-systemd.inc.
This is because systemd's source codes are hardcoding "/root", and other
values are not officially supported. See the list below.
$ grep -rl '"/root"' src/ | grep -v 'src/test'
src/core/namespace.c
src/basic/user-util.c
src/nss-systemd/nss-systemd.c
src/nspawn/nspawn.c
src/firstboot/firstboot.c
src/shared/userdb.c
src/shared/user-record.c
$ grep -rl /root network/ factory/ sysctl.d/ sysusers.d/ rules.d/ tmpfiles.d/ units/ xorg/ tools/
sysusers.d/basic.conf.in
tmpfiles.d/provision.conf
units/emergency.service.in
units/rescue.service.in
tools/list-discoverable-partitions.py
Previously, the recipe was just substituting sysusers.d/basic.conf.in,
which is not enough to be treated as 'fully support'. I deliberately put
a warning message in do_install to warn users about non "/root" ROOT_HOME
value. Don't remove it until all above places are handled.

3. cgroupv2 is now the default.
cgroupv2 is the default for systemd for many years and it's the default
for distros such as ubuntu and fedora. Let's also use it as the default.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: refresh musl patches for v255.1

The current patches are refreshed. A new patch is added:
0022-avoid-missing-LOCK_EX-declaration.patch

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/selftest: add test case to cover 'devtool modify -n' for a git recipe

Add a test case to ensure the following error does not happen again for
'devtool modify -n'.

Traceback (most recent call last):
  File "/buildarea2/chenqi/poky/scripts/devtool", line 349, in <module>
    ret = main()
  File "/buildarea2/chenqi/poky/scripts/devtool", line 336, in main
    ret = args.func(args, config, basepath, workspace)
  File "/buildarea2/chenqi/poky/scripts/lib/devtool/standard.py", line 924, in modify
    if not initial_revs["."]:
KeyError: '.'

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rust: Re-write RPATHs in the copies llvm-config

Ensure that it can still access the native-sysroot for dependencies,
use ORIGIN to indicate this relative its install location, this also
helps in this not getting into the output of llvm-config which could
otherwise provide incorrect library paths

target rust recipe builds ( cross compile ) calls llvm-config from
target sysroot which works ok as long as C++ runtime it needs is
available on build host e.g. libstdc++ etc. which is commonly the
case, however when using clang and llvm runtime this falters since
it should be using libc++ from native sysroot and if this does not
exist on build machine this fails to find libc++ shared object and
llvm-config fails to run. This ensures that llvm-config version in
use is correctly relocated and can use shared libraries from native
sysroot correctly. Adding ORIGIN to sysroot will look for the .so in
same dir as the binary and there is the libc++.so.1 copied in place

Fixes rust build with clang compiler.

| /mnt/b/yoe/master/build/tmp/work/riscv64-yoe-linux/rust/1.74.1/recipe-sysroot/usr/lib/llvm-rust/bin/llvm-config: error while loading shared libraries: libc++.so.1: cannot open shared object file: No such file or director
y
| thread 'main' panicked at llvm.rs:551:19:
| command did not execute successfully: "/mnt/b/yoe/master/build/tmp/work/riscv64-yoe-linux/rust/1.74.1/recipe-sysroot/usr/lib/llvm-rust/bin/llvm-config" "--version"
| expected success, got: exit status: 127

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: Update to 17.0

Update to pulseaudio 17.0 version.

Since pulseaudio commit e96d278bfc51 ("daemon/meson.build: Install
dbus policy in /usr, not /etc"), pulseaudio-system.conf needs to be passed
in FILES, otherwise the following QA error is seen:

ERROR: pulseaudio-17.0-r0 do_package: QA Issue: pulseaudio: Files/directories were installed but not shipped in any package:
  /usr/share/dbus-1
  /usr/share/dbus-1/system.d
  /usr/share/dbus-1/system.d/pulseaudio-system.conf

Signed-off-by: Fabio Estevam <festevam@denx.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

perl: Fix perl-module-* being ignored via COMPLEMENTARY_GLOB

When perl-modules is added via COMPLEMENTARY_GLOB, all perl-module-*
packages recommended by perl-modules are ignored due to the defined
behavior of COMPLEMENTARY_GLOB.

This patch changes the relationship between perl-modules and all of its
perl-module-* from RRECOMMENDS to RDEPENDS. This makes sense as
perl-modules should represent the collection of all optional
perl-module-* packages. After this patch, perl-modules itself is being
RRECOMMENDED instead of the individual perl-module-* packages
perl-modules represents.

Signed-off-by: William Lyu <William.Lyu@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev-extraconf: fix unmount directories containing octal-escaped chars

USB devices are auto-mounted in a directory named like theirs labels.
Special characters like whitespace are octal-escaped in /proc/mounts
output. Using directly this output file as an argument for umount failed
and the mount directory can't be removed as still busy.
Using printf allows these special characters to be unescaped.

Signed-off-by: Jonathan GUILLOT <jonathan@joggee.fr>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: Update status of CVE-2023-4039

This is fixed via a patch added in gcc-13.2.inc already, but still
reported e.g. for libgcc as it is not defining an own source but use the
shared gcc-source.

Signed-off-by: Simone Weiß <simone.p.weiss@posteo.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3-yamllint: Add recipe

Add recipe for yamllint.  There is an upcoming change in u-boot where
the binman tool is now configured to call yamllint to verify the configs
during compile time.

There was a previous patch a year ago from Trevor Woerner that never
made it into oe-core.  This patch is a reworking of his patch but
pointing to a newer version.

Signed-off-by: Ryan Eatmon <reatmon@ti.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

p11-kit: fix parallel build failures

It fails occasionally with missing generated header files:

| ../git/common/asn1.c:42:10: fatal error: openssl.asn.h: No such file or directory
|    42 | #include "openssl.asn.h"
|       |          ^~~~~~~~~~~~~~~
| compilation terminated.

According to meson manual page:

https://mesonbuild.com/Wrap-best-practices-and-tips.html#declare-generated-headers-explicitly

'asn_h_dep' should be a dependency of static_library target 'libp11_asn1'
to make sure that required header files generated before compile
common/asn1.c.

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nativesdk-cairo: fix build error

It fails to build nativesdk-cairo:

| ../cairo-1.18.0/meson.build:381:13: ERROR: Can not run test
    applications in this cross environment.

Set meson property ipc_rmid_deferred_release in cross-file to fix the
issue which is as same as for target cairo

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemu: Fix segfaults in webkitgtk:do_compile on debian11

Add a work around for segfaults we're seeing with qemu 8.2.0 when
compiling webkitgtk on debian 11 autobuilder workers.

The issue has been reported upstream to work out an official fix.

[YOCTO #15367]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rust: Enable rust oe-selftest.

* Enable rust oe-selftest.

* Include the dependent patches for rust oe-selftest in
  meta/recipes-devtools/rust/rust-source.inc

* Disable rust oe-selftest for mips32 target (Rust upstream has classified it into tier 3 target,
  for which the Rust project does not build or test automatically) as it is unstable with rust tests.
  https://doc.rust-lang.org/nightly/rustc/platform-support.html#tier-3

* The testing is done on arm32, arm64, mips64, x86 and  x86_64 targets on Ubuntu 22.04.

Signed-off-by: Yash Shinde <Yash.Shinde@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rust: Remove the test cases whose parent dir is also present in the exclude list

To avoid overlap and redundancy of rust tests, remove the test cases whose
parent dir is already excluded.

Tests which are failing from below dirs are removed as these dirs are
already present in exclude list
            tests/run-make
            tests/rustdoc
            tests/mir-opt
            tests/ui-fulldeps

Signed-off-by: Yash Shinde <Yash.Shinde@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rust: Add new tests in the exclude list for rust oe-selftest

Add newly failing tests cases in the exclude list for
rust oe-selftest.

Signed-off-by: Yash Shinde <Yash.Shinde@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rust: Fix assertion failure error on oe-selftest

Fixes: thread 'main' panicked at 'assertion failed: `(left == right)`
       left: `x86_64-unknown-linux-gnu`,
       right: `x86_64-poky-linux-gnu`: Cannot obtain compiler for non-native build triple at stage 0', compile.rs:1474:13

Add correct target value for cross-compiled targets on stage1 during rust oe-selfest.

Signed-off-by: Yash Shinde <Yash.Shinde@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rust: Enable RUSTC_BOOTSTRAP to use nightly features during rust oe-selftest.

Fixes: error: the option `Z` is only accepted on the nightly compiler
When rust.channel is set to either beta or stable, we can't use
nightly features on bootstrap without RUSTC_BOOTSTRAP.
Set RUSTC_BOOTSTRAP=1 to use nightly features on stable or beta.

Signed-off-by: Yash Shinde <Yash.Shinde@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rust: detect user-specified custom targets in compiletest

Fixes: thread 'main' panicked at 'failed to gather the target spec
       for '<arch>-unknown-linux-gnu', synthetic_targets.rs:66:9

Detect and fetch custom target configurations when rustc is
bootstrapped in rust oe-selftest.

Signed-off-by: Yash Shinde <Yash.Shinde@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rust: Fetch cargo from rust-snapshot dir.

Fixes: Exception: no cargo executable found at
       `${B}/rustc-1.74.1-src/build/x86_64-unknown-linux-gnu/stage0/bin/cargo`

Fix the cargo binary path error on oe-selftest and path set to rust-snapshot dir.

Patch sent to upstream- https://github.com/rust-lang/rust/pull/120125

Signed-off-by: Yash Shinde <Yash.Shinde@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemu: Upgrade 8.1.2 -> 8.2.0

We can drop the mips workaround patch since there were fixes in 8.2.0.

The build system changed and we should drop cross.patch and replace it with
explicit settings for cross-prefix, and host-cc. To make that work we need
to indicate we don't use a cross pkg-config.

PIE isn't availale for mingw so tweak PACKAGECONFIG accordingly.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3-subunit: Add missing module dependency

In version 1.4.3:
* Subunit now has a dependency on an external iso8601
  module rather than shipping its own.

so add the missing dependency.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/selftest/devtool: add test for recipes with multiple sources in SRC_URI

add a non regression test for devtool modify/build on recipe having
several sources in SRC_URI

Signed-off-by: Julien Stephan <jstephan@baylibre.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

devtool: modify: add support for multiple source in SRC_URI

[YOCTO #15162]

when doing devtool modify, sources are extracted into a devtool
temporary workdir. The main source is moved inside
build/workspace/sources/${BPN}/ and local files are moved inside
build/workspace/sources/${BPN}/oe-local-files. Secondary sources are
currently not handled and are lost.

Here is the output of devtool modify/build on bzip2 recipe:

NOTE: bzip2: compiling from external source tree <...>/build/workspace/sources/bzip2
ERROR: bzip2-1.0.8-r0 do_install_ptest_base: ExecutionError('<...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/temp/run.do_install_ptest_base.3368', 1, None, None)
ERROR: Logfile of failure stored in: <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/temp/log.do_install_ptest_base.3368
Log data follows:
| DEBUG: Executing shell function do_install_ptest_base
| NOTE: make -j 16 DESTDIR=<...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest install-ptest
| sed  -n '/^runtest:/,/^install-ptest:/{/^install-ptest:/!p}' \
|            ../../../../../../workspace/sources/bzip2/Makefile.am      > <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest/Makefile
| cp ../../../../../../workspace/sources/bzip2/sample1.ref      <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest/
| cp ../../../../../../workspace/sources/bzip2/sample2.ref      <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest/
| cp ../../../../../../workspace/sources/bzip2/sample3.ref      <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest/
| cp ../../../../../../workspace/sources/bzip2/sample1.bz2      <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest/
| cp ../../../../../../workspace/sources/bzip2/sample2.bz2      <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest/
| cp ../../../../../../workspace/sources/bzip2/sample3.bz2      <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest/
| ln -s /usr/bin/bzip2          <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest/bzip2
| cp: cannot stat '<...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/git/commons-compress': No such file or directory
| WARNING: <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/temp/run.do_install_ptest_base.3368:189 exit 1 from 'cp -r <...>/build/tmp/work/core2-64-poky-linux/bzip2/
1.0.8/git/commons-compress <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/image/usr/lib/bzip2/ptest/bzip2-tests/commons-compress'
| WARNING: Backtrace (BB generated script):
|       #1: do_install_ptest, <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/temp/run.do_install_ptest_base.3368, line 189
|       #2: do_install_ptest_base, <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/temp/run.do_install_ptest_base.3368, line 158
|       #3: main, <...>/build/tmp/work/core2-64-poky-linux/bzip2/1.0.8/temp/run.do_install_ptest_base.3368, line 226
ERROR: Task (<...>/poky/meta/recipes-extended/bzip2/bzip2_1.0.8.bb:do_install_ptest_base) failed with exit code '1'
NOTE: Tasks Summary: Attempted 776 tasks of which 765 didn't need to be rerun and 1 failed.

Summary: 1 task failed:
  <...>/poky/meta/recipes-extended/bzip2/bzip2_1.0.8.bb:do_install_ptest_base

externalsrc class modify SRC_URI to keep only:
* 'file', 'npmsw' and 'crate' sources
* url with type parameter matching 'kmeta' or 'git-dependency'

So by forcing to add type='git-dependency' on secondary sources, we
ensure that when building the recipe, the secondary sources can be
unpacked into WORKDIR.

This allows recipes containing several sources to be built under a
devtool context, but it has some limitations:
* user would not be able to generate patches for the secondary sources
* type="git-dependency" is added for secondary sources even on non git
  sources, so we may want to rename this parameter

Signed-off-by: Julien Stephan <jstephan@baylibre.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

externalsrc: fix task dependency for do_populate_lic

do_populate_lic dependencies are defined inside license.bbclass such as:

  addtask populate_lic after do_patch before do_build

but externalsrc deletes the do_patch task, so the only dependency left for
do_populate_lic is "before do_build"

On a devtool context, when doing devtool modify, sources are extracted inside
build/workspace/sources/${BPN}/ and local files inside
build/workspace/sources/${BPN}/oe-local-files

When building the recipe after a devtool modify, do_unpack is called again to
unpack (possibly modified) local files from
build/workspace/sources/${BPN}/oe-local-files into ${WORKDIR}.

Since the only left dependency for do_populate_lic is do_build, the
do_populate_lic can be called BEFORE do_unpack. Most of the time this is not a
problem, because license files are generally located inside ${S}, which
corresponds to build/workspace/sources/${BPN} (and is already unpacked),
but this can lead to an issue if recipe sets LIC_FILES_CHKSUM to look for
files in ${WORKDIR} (example from init-ifupdown_1.0.bb):

  LIC_FILES_CHKSUM = "file://${WORKDIR}/copyright;md5=3dd6192d306f582dee7687da3d8748ab"

So devtool modify init-ifupdown && bitbake init-ifupdown gives the following
error:

  WARNING: init-ifupdown-1.0-r0 do_populate_lic: Could not copy license file <...>/build/tmp/work/qemux86_64-poky-linux/init-ifupdown/1.0/copyright to <...>/build/tmp/work/qemux86_64-poky-linux/init-ifupdown/1.0/license-destdir/qemux86_64/init-ifupdown/copyright: [Errno 2] No such file or directory: '<...>/build/tmp/work/qemux86_64-poky-linux/init-ifupdown/1.0/copyright'
  ERROR: init-ifupdown-1.0-r0 do_populate_lic: QA Issue: init-ifupdown: LIC_FILES_CHKSUM points to an invalid file: <...>/build/tmp/work/qemux86_64-poky-linux/init-ifupdown/1.0/copyright [license-checksum]
  ERROR: init-ifupdown-1.0-r0 do_populate_lic: Fatal QA errors were found, failing task.
  ERROR: Logfile of failure stored in: <...>/build/tmp/work/qemux86_64-poky-linux/init-ifupdown/1.0/temp/log.do_populate_lic.838584
  ERROR: Task (<...>/poky/meta/recipes-core/init-ifupdown/init-ifupdown_1.0.bb:do_populate_lic) failed with exit code '1'

Fix this by forcing the do_populate_lic task to run after do_unpack

Signed-off-by: Julien Stephan <jstephan@baylibre.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glibc: Do not enable CET on 32bit x86

Support has been removed from 2.39+ see [1]

[1] https://sourceware.org/git/?p=glibc.git;a=commit;h=25f1e16ef03a6a8fb1701c4647d46c564480d88c

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libssh2: backport fix for CVE-2023-48795

Backport the upstream fix for CVE-2023-48795.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

grub2: upgrade 2.06 -> 2.12

Drop patches that have been upstreamed. Refresh others.

This version dropped extra_deps.lst from the tarball [1] and that leads
to build failures. Restore it in do_configure for now.

[1] https://git.savannah.gnu.org/cgit/grub.git/commit/?id=b835601c7639ed1890f2d3db91900a8506011a8e

Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libpsl: upgrade 0.21.2 -> 0.21.5

License-Update: Copyright dates changed

Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cronie: upgrade 1.7.0 -> 1.7.1

Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bluez5: upgrade 5.71 -> 5.72

Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mdadm: Disable ptests

As agreed upon in the bug triage meeting, disable mdadm ptest for the time being

Related: [Yocto #15181] [Yocto #15159] [Yocto #15308] [Yocto #15309]

Signed-off-by: Pavel Zhukov <pavel@zhukoff.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/6.1: drop recipes

6.8-lts + linux-yocto-dev will the active reference kernels in
the upcoming release. We've moved all of the preferred versions
to 6.8, so we can now drop the recipes for 6.1

6.1 will continue to be maintained and updated for released
and supported branches.

Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libxcomposite: Drop obsolete patch

The patch looks obsolete now we have recipe specific sysroots, drop it.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xserver-xorg: disable xvfb by default

xvfb has limited use, so to mitigate CVE-2023-5574 out of the box we can
disable the xvfb PACKAGECONFIG.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xserver-xorg: add PACKAGECONFIG for xvfb

Xvfb is pretty niche and has outstanding unsolved security issues, so
let people disable it and add a conditional CVE_STATUS to reflect this.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cve_check: cleanup logging

Primarily list the number of patches found, useful when debugging.

Also clean up some bad escaping that caused warnings and use
re.IGNORECASE instead of manually doing case-insenstive rang matches.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cve_check: handle CVE_STATUS being set to the empty string

Handle CVE_STATUS[...] being set to an empty string just as if it was
not set at all.

This is needed for evaluated CVE_STATUS values to work, i.e. when
setting not-applicable-config if a PACKAGECONFIG is disabled.

Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

zlib: ignore CVE-2023-6992

This CVE is for iCPE cloudflare:zlib.

Alternative to ignoring would be to limit CVE_PRODUCT, but
historic CVEs already have two - gnu:zlib and zlib:zlib.
So limiting it could miss future CVEs.

Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>