Since Buildroot commit [1] updating qemu defconfigs from Kernel 6.6.32
to 6.12.9, qemu_aarch64_ebbr_defconfig, qemu_aarch64_sbsa_defconfig
and qemu_arm_ebbr_defconfig can fail to build with error:
/bin/sh: 1: python3: not found
make[7]: *** [drivers/gpu/drm/msm/Makefile:182: drivers/gpu/drm/msm/generated/a2xx.xml.h] Error 127
The issue can happen on build hosts without Python, and if the Linux
kernel is built before any package having host-python3 in its
dependencies (e.g. host-qemu, optee-os, ...).
Starting with Linux 6.10, the arm and arm64 architecture defconfigs need
python3 to build. [2]
This commit fixes this issue by adding
BR2_LINUX_KERNEL_NEEDS_HOST_PYTHON3=y in those defconfigs.
[1] 44075c1955
[2] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=0fddd045f88e34d6160785a3a5e506d374566454
Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com>
Cc: Romain Naour <romain.naour@gmail.com>
[Julien:
- reword commit log to add the commit introducing the issue
- fix qemu_aarch64_sbsa_defconfig which has the same issue
]
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 6001601544)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
This patch fixes the CPE ID for both go-bin and go-src packages.
The go-src package did had a CPE ID because GO_SRC_CPE_ID_VENDOR is
defined but the generated CPE ID would be the following:
cpe:2.3:a:golang:go-src:*:*:*:*:*:*:*:*
The CPE product go-src doesn't match to anything and should be 'go'
instead.
For go-bin no CPE_ID variable were defined so no CPE ID would be
generated if using a pre-build toolchain.
This patch duplicate the definition of the CPE_ID variables, I didn't
put the definition in the 'package/go/go.mk' file to avoid generate a
CPE for the go virtual package.
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 1869b4065c)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Fixes: https://autobuild.buildroot.net/results/8d14173f31922a737c326be75aa24c9f9471342a/
The vde2 source code uses empty argument function pointer prototypes like:
typedef ssize_t (* ssize_fun)();
extern ssize_fun vdehist_vderead;
In C17 and earlier, such prototypes means unknown number of arguments,
whereas with C23 it means no arguments, leading to build failures:
libvdehist.c:39:27: error: initialization of 'ssize_fun' {aka 'long int
(*)(void)'} from incompatible pointer type 'ssize_t (*)(int, void *,
size_t)' {aka 'long int (*)(int, void *, long unsigned int)'}
[-Wincompatible-pointer-types]
Workaround it using -std=gnu99. Vde2 is provided both as a host and target
package, so do this unconditionally for both.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
(cherry picked from commit 2218f02155)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Now that the qspi.bin file is included for zynqmp images, the 32M vfat
partition is getting dangerously close to being full.
In the process of bumping Linux kernel 6.12.10 to 6.12.40, the size of the
default Linux kernel Image file for zynqmp products is increasing from 23.4M
to 26.9M. This is leaving less than 1M of free space in the vfat partition
meaning users may unknowingly exceed the 32M size by just adding Linux kernel
configs.
To avoid potential problems, this patch doubles the size of the vfat partition
to 64M, which should be plenty of space for any configuration.
Signed-off-by: Neal Frager <neal.frager@amd.com>
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
(cherry picked from commit c7a45b8302)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
This tool helps to update Linux stable releases.
The script uses the versions found in linux/linux.hash. For each of the
versions it downloads the related hash list and tries to find an updated
stable release. If found it updates all related files and hashes.
Signed-off-by: Gero Schwäricke <gero.schwaericke@grandcentrix.net>
[Arnout: commonalize the sed expression for linux and linux-headers]
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
(cherry picked from commit a87b1800b9)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
This reverts commit fecc337d21,
re-adding optional support for pulseuadio in ffmpeg.
This is possible now that the circular dependency:
pulseaudio -> libsoxr -> ffmpeg -> pulseaudio
has been broken between libsoxr and ffmpeg.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 8db836d292)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Having support for ffmpeg in libsoxr causes several circular
dependencies:
- pulseaudio -> libsoxr -> ffmpeg -> pulseaudio
(This one was worked around in
fecc337d21 by dropping the ffmpeg ->
pulseaudio dependency, but it's annoying as being able to output
audio from ffmpeg into pulseaudio seems useful)
- libopenmpt -> pulseaudio -> libsoxr -> ffmpeg -> libopenmpt
(This is one exists, and is fixed by this commit)
To fix both circular dependencies, we propose to drop the dependency
of libsoxr on ffmpeg, which is used to get some optimized FFT
implementation. While not being useless, this is the point where
breaking the dependency seems the most logical.
Neither Debian nor Fedora build libsoxr with ffmpeg support.
In order to achieve that, we add a new patch making the option
WITH_AVFFT actually effective. Since we no longer build against
ffmpeg, the previously added patch that was fixing the pkg-config file
is no longer needed.
Fixes:
https://autobuild.buildroot.net/results/bbfce69f35cafd56245ebeae79086a8b64592a16/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 4bf2a854f9)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
The new package_dir element contains the path of the package
directory. For Buildroot packages this is a path relative to the main
Buildroot directory, for packages from external trees an absolute
path.
Signed-off-by: Fiona Klute (WIWA) <fiona.klute@gmx.de>
Tested-by: Titouan Christophe <titouan.christophe@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit b10f50399a)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Building the column store plugin (which gets automatically enabled
when Boost with a bunch of submodules is found) currently fails as it
requires a number of runtime checks that obviously cannot work when
cross-compiling.
OpenEmbedded has defined them as such:
-DLSTAT_FOLLOWS_SLASHED_SYMLINK_EXITCODE=0 \
-DSTAT_EMPTY_STRING_BUG_EXITCODE=0 \
-DMASK_LONGDOUBLE_EXITCODE=0 \
but it really isn't clear whether those values are completely correct
for all CPU architectures, especially the last one.
So instead, let's disable the column store plugin until someone
actually needs it and figures out how to set those variables properly.
The issue fixed by this commit can be reproduced using the following
defconfig:
BR2_aarch64=y
BR2_ARM_FPU_VFPV3D16=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN=y
BR2_PACKAGE_MARIADB=y
BR2_PACKAGE_MARIADB_SERVER=y
BR2_PACKAGE_BOOST=y
BR2_PACKAGE_BOOST_DATE_TIME=y
BR2_PACKAGE_BOOST_FILESYSTEM=y
BR2_PACKAGE_BOOST_REGEX=y
BR2_PACKAGE_BOOST_THREAD=y
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 44bba49507)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
The Config.in comment is between the main BR2_PACKAGE_MARIADB option
and its sub-options, causing the sub-options to not appear as nicely
indented in menuconfig.
Before:
[ ] mariadb
[ ] mariadb server
[ ] mariadb embedded server
After:
[ ] mariadb
[ ] mariadb server
[ ] mariadb embedded server
which is what we want.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 8964f0652c)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Recently when trying to reproduce a build error with test-pkg on the toolchain
br-arm-basic, I had to create a new CSV file with that toolchain only to be
able to run the test in that single case.
I believe having a command-line option right away in test-pkg would make this
much easier, as I can now run: `./utils/test-pkg -T br-arm-basic -p redis`
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
(cherry picked from commit f3be464f7a)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
In order to remove thttpd package from Buildroot, we have to replace it
from our testsuite (TestWget and TestLibCurl).
Busybox provide an httpd server applet but it's not enabled in our
default busybox configuration.
For the sake of those tests, add a new busybox option to build and
install the Busybox's httpd server and its init script.
Import S90thttpd from thttpd package to S90httpd but with some changes
following S01syslogd init script as a reference [1].
[1] 3dc8061444
Cc: Julien Olivain <ju.o@free.fr>
Cc: Fiona Klute (WIWA) <fiona.klute@gmx.de>
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Reviewed-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
(cherry picked from commit bcd97e1424)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Since Buildroot commit [1] dc0f7215c6 "package/gcc: switch to 13.x as
default", sipeed_licheepi_zero_defconfig fails when building the Linux
kernel with the error message:
drivers/ata/libahci.c: In function 'ahci_led_store':
./include/linux/compiler.h:350:45: error: call to '__compiletime_assert_1138' declared with attribute error: BUILD_BUG_ON failed: sizeof(_s) > sizeof(long)
350 | _compiletime_assert(condition, msg, __compiletime_assert_, __LINE__)
This issue is fixed by upstream Kernel commit [2].
This commit fixes this issue by updating the Linux Kernel from v5.3.5
(which is not a LTS series) to v6.12.43 (which is a LTS series
including the fixes). With this update,
BR2_LINUX_KERNEL_INTREE_DTS_NAME is updated accordingly to follow the
new DTS organization.
This commit also updates U-Boot from 2022.01 to 2025.07. This new
version now requires BR2_TARGET_UBOOT_NEEDS_GNUTLS=y.
While at it, this commit also switches to a stable glibc Bootlin
external toolchain, to follow the new guidelines from [3].
Finally, this commit adds BR2_DOWNLOAD_FORCE_CHECK_HASHES=y and the
corresponding hash files. The .checkpackageignore is also updated to
reflect that change.
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/11026771001
[1] dc0f7215c6
[2] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f07788079f515ca4a681c5f595bdad19cfbd7b1d
[3] https://elinux.org/Buildroot:DeveloperDaysELCE2024#Rules_for_defconfigs
Cc: Michel Stempin <michel.stempin@wanadoo.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
(cherry picked from commit db951a4c71)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Since glibc 2.28, <sys/sysmacros.h> is no longer implicitly included
by <sys/types.h>, which means major() is only defined if you include
<sys/sysmacros.h>.
This wasn't really visible for a while in loadable plugins built by
lirc-tools as they could happily be built with undefined
symbols... except they would fail to load at runtime.
But then GCC 14.x brought more strict checking of having the function
prototype around, which now detects this issue at build time.
So the actual issue has been around since we switched to glibc 2.28,
but it only started failing to build since gcc 14.x. The first failure
is on August 2024, so this patch should be backported to our stable
branches.
The fix is backport from an upstream commit, post 0.10.2.
Fixes:
https://autobuild.buildroot.net/results/2f64accd430d9467add3f1195f3f59eff8e99a4d/
Minimal reproducer:
BR2_arm=y
BR2_cortex_a9=y
BR2_ARM_ENABLE_VFP=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN_ARMV7_EABIHF_GLIBC_STABLE=y
BR2_INIT_NONE=y
BR2_SYSTEM_BIN_SH_NONE=y
# BR2_PACKAGE_BUSYBOX is not set
BR2_PACKAGE_LIRC_TOOLS=y
# BR2_TARGET_ROOTFS_TAR is not set
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 2313c23242)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Currently, the manual doesn't mention the checkpackageignore file at
all, leaving the user under the assumption that running "make
check-package" and running utils/check-package directly on the package
they have touched would be equivalent, when the former evaluates the
checkpackageignore file while the latter does not. This could cause them
to forget to remove stale entries from that file.
Point out this difference in behavior.
Signed-off-by: Florian Larysch <fl@n621.de>
[Arnout: add that stale lines should be removed from
.checkpackageignore]
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
(cherry picked from commit 47181a164c)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
