examples/vm_power_manager/meson.build in DPDK detects the presence of
libvirt:
opt_dep = cc.find_library('virt', required : false)
and then builds some examples or not depending on the availability of
libvirt. Let's make this optional dependency explicit in dpdk.mk, even
if there's no explicit enable/disable option for it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When libvirt is present before DPDK is built, some additional examples
are compiled. One of them fails to build due to a missing <stdlib.h>
include. Let's import a patch from OpenSuse, that we have submitted
upstream, to fix this issue.
We couldn't find any autobuilder failure for this issue, but the
following defconfig allows to reproduce the failure:
BR2_aarch64=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN_AARCH64_GLIBC_STABLE=y
BR2_ROOTFS_DEVICE_CREATION_DYNAMIC_EUDEV=y
BR2_PACKAGE_DPDK=y
BR2_PACKAGE_DPDK_EXAMPLES=y
BR2_PACKAGE_LIBVIRT=y
The problem exists since DPDK v19.11, so it has been in Buildroot
since DPDK was introduced in commit
d17d1b6bde.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Buildroot commit [1] (package/perl: fix build issue with musl)
introduced a patch that is meant to be applied on top of perl-cross,
which is extracted on top of perl only in the target variant.
Since the patch was introduced as a normal package patch, the Buildroot
infra is trying to always apply it, even for the host package variant.
Since perl-cross is not extracted for the host variant, some patched
files are missing. In that case, the host-perl is failing with error:
>>> host-perl 5.42.3 Patching
Applying 0001-configure-keep-_GNU_SOURCE-in-build-flags.patch using patch:
can't find file to patch at input line 46
This commit fixes the issue by moving the package patch in a dedicated
"perl-cross" subdirectory, to make sure it will no longer be applied by
the infra. We apply the patch only for the target package variant using
a _POST_PATCH_HOOKS hook.
Fixes:
- [1]
- https://gitlab.com/buildroot.org/buildroot/-/jobs/16185948610 (TestPerlXMLLibXML)
- ...and few other tests requiring host-perl
[1] d950fff290
Signed-off-by: Julien Olivain <ju.o@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When OpenSSL is enabled, if DES support is enabled, OpenSSH uses
DES_crypt. However, without OpenSSL or its DES support, there is no
available crypt() implementation for OpenSSH libopenbsd-compat xcrypt()
function, resulting in the following error:
.../host/bin/i686-buildroot-linux-gnu-gcc -o sshd-auth sshd-auth.o
auth2-methods.o auth-rhosts.o auth-passwd.o sshpty.o sshlogin.o
servconf.o serverloop.o auth.o auth2.o auth-options.o session.o
auth2-chall.o groupaccess.o auth-bsdauth.o auth2-hostbased.o
auth2-kbdint.o auth2-none.o auth2-passwd.o auth2-pubkey.o
auth2-pubkeyfile.o auth2-gss.o gss-serv.o gss-serv-krb5.o
monitor_wrap.o auth-krb5.o audit.o audit-bsm.o audit-linux.o
platform.o loginrec.o auth-pam.o auth-shadow.o auth-sia.o
sandbox-null.o sandbox-rlimit.o sandbox-darwin.o
sandbox-seccomp-filter.o sandbox-capsicum.o sandbox-solaris.o
sftp-server.o sftp-common.o uidswap.o ssh-pkcs11-client.o
ssh-sk-client.o -L. -Lopenbsd-compat/ -D_LARGEFILE_SOURCE
-D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -O2 -g0
-D_FORTIFY_SOURCE=1 -Wl,-z,relro -Wl,-z,now -Wl,-z,noexecstack
-fstack-protector-strong -pie -lssh -lopenbsd-compat
-L.../host/bin/../i686-buildroot-linux-gnu/sysroot/usr/lib
-lssl -lcrypto -lcrypto -lz
.../host/lib/gcc/i686-buildroot-linux-gnu/15.3.0/../../../../
i686-buildroot-linux-gnu/bin/ld:
openbsd-compat//libopenbsd-compat.a(xcrypt.o): in function `xcrypt':
xcrypt.c:(.text+0x51): undefined reference to `crypt'
.../host/lib/gcc/i686-buildroot-linux-gnu/15.3.0/../../../../
i686-buildroot-linux-gnu/bin/ld:
openbsd-compat//libopenbsd-compat.a(xcrypt.o): in function `xcrypt':
xcrypt.c:(.text+0x51): undefined reference to `crypt' collect2: error:
ld returned 1 exit status make[2]: *** [Makefile:233: sshd-auth] Error
1 make[2]: *** Waiting for unfinished jobs.... collect2: error: ld
returned 1 exit status make[2]: *** [Makefile:230: sshd-session] Error
1 make[1]: *** [package/pkg-generic.mk:273:
.../build/openssh-10.4p1/.stamp_built]
Error 2 make: *** [Makefile:83: _all] Error 2
This commit enables BR2_PACKAGE_LIBXCRYPT with OpenSSH as long as
glibc is used. Since "sshd-auth" is compiled regardless of
BR2_PACKAGE_OPENSSH_SERVER, we need to enable it with BR2_PACKAGE_OPENSSH.
Signed-off-by: Jimmy Durand Wesolowski <jimmy.wesolowski@mobileye.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Buildroot commit 297f6f1921 updated vim.
However README.txt (used as part of the license hash check) has been updated
upstream in [1], without any corresponding hash change in Buildroot, leading
to build failure.
Fixes: https://gitlab.com/buildroot.org/buildroot/-/work_items/189
NB: This also affects 2025.02.x & 2026.05.x, so this patch
should be applied there too.
[1] e7e21018fc
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
patchwork.buildroot.org used to be a redirect to patchwork.ozlabs.org,
but we are now running our own instance, so let's adjust the links in
the website accordingly.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
patchwork.buildroot.org used to be a redirect to patchwork.ozlabs.org,
but we are now running our own instance, so let's adjust the links in
the manual accordingly.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following CVEs:
CVE-2026-19499:
63b53df549
CVE-2026-77117:
6f9b2bfa50
CVE-2026-80489:
cb61572ea3
Added GLIBC_IGNORE_CVES for CVE-2026-19542 which was forgotten in
buildroot commit 58f3137738.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Julien Olivain <ju.o@free.fr>
When BR2_PACKAGE_OPENCV4_WITH_WEBP=y we need to enable mux and demux
support in webp, otherwise the build of OpenCV fails as follows:
CMake Error: The following variables are used in this project,
but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly
in the CMake files:
WEBP_DEMUX_LIBRARY
linked by target "opencv_imgcodecs"
in directory [...]/build/opencv4-4.13.0/modules/imgcodecs
WEBP_MUX_LIBRARY
linked by target "opencv_imgcodecs"
in directory [...]/build/opencv4-4.13.0/modules/imgcodecs
The issue already exists in 2025.02.x.
Fixes:
https://autobuild.buildroot.net/results/d3e0446a87d32469267e241866c4224143170f31/
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Even though it's only documentation, the form of the names of the
post-image.sh and post-build.sh scripts should be consistent with the
names of those scripts used in the code base, using hyphen, not
underscore.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Besides just updating a little terminology, remove the awkward
reference to "his" when referring to developers.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The openscap build can fail with the following error:
In file included from [...]/src/OVAL/probes/unix/linux/systemdunitproperty_probe.c:38:
[...]/src/OVAL/probes/unix/linux/systemdshared.h: In function ‘get_all_systemd_units’:
[...]/src/OVAL/probes/unix/linux/systemdshared.h:188:50: error: implicit declaration of function ‘basename’; did you mean ‘rename’? [-Wimplicit-function-declaration]
188 | char *unit_name_s = oscap_strdup(basename(value.str));
| ^~~~~~~~
| rename
In file included from [...]/src/OVAL/probes/unix/linux/systemdunitdependency_probe.c:37:
[...]/src/OVAL/probes/unix/linux/systemdshared.h: In function ‘get_all_systemd_units’:
[...]/src/OVAL/probes/unix/linux/systemdshared.h:188:50: error: implicit declaration of function ‘basename’; did you mean ‘rename’? [-Wimplicit-function-declaration]
188 | char *unit_name_s = oscap_strdup(basename(value.str));
| ^~~~~~~~
| rename
This error happens when:
- dbus is enabled in the configuration, making openscap build probes
code
- the toolchain uses musl, which does not declare basename() in string.h
the way glibc does
The build error can be reproduced with the following minimal defconfig:
BR2_arm=y
BR2_cortex_a9=y
BR2_ARM_ENABLE_NEON=y
BR2_ARM_ENABLE_VFP=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN_ARMV7_EABIHF_MUSL_BLEEDING_EDGE=y
BR2_PACKAGE_DBUS=y
BR2_PACKAGE_OPENSCAP=y
Backport the upstream fix to allow openscap to build with such
configuration. The custom patch can be removed once openscap is
re-released on its branch 1.3.x.
The issue affects master, 2026.05.x and 2025.02.x
Fixes: https://autobuild.buildroot.org/results/a874d4f34d36fa9f8566be90ad2d5facb99aec24/
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
host-erlang build can fail with the following error:
make[5]: Nothing to be done for 'opt'.
MAKE opt
CC ../priv/bin/x86_64-pc-linux-gnu/odbcserver
/usr/bin/ld: ../priv/obj/x86_64-pc-linux-gnu/odbcserver.o: in function `encode_column_dyn':
odbcserver.c:(.text+0x6b4): undefined reference to `ei_x_encode_tuple_header'
/usr/bin/ld: odbcserver.c:(.text+0x6c2): undefined reference to `ei_x_encode_tuple_header'
/usr/bin/ld: odbcserver.c:(.text+0x6d4): undefined reference to `ei_x_encode_ulong'
/usr/bin/ld: odbcserver.c:(.text+0x6e7): undefined reference to `ei_x_encode_ulong'
/usr/bin/ld: odbcserver.c:(.text+0x6fa): undefined reference to `ei_x_encode_ulong'
/usr/bin/ld: odbcserver.c:(.text+0x708): undefined reference to `ei_x_encode_tuple_header'
/usr/bin/ld: odbcserver.c:(.text+0x71b): undefined reference to `ei_x_encode_ulong'
/usr/bin/ld: odbcserver.c:(.text+0x72e): undefined reference to `ei_x_encode_ulong'
[...]
This can be reproduced with the following minimal defconfig (and
libei.so present on host, see details below):
BR2_x86_64=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_PACKAGE_ERLANG=y
Those missing symbols are part of the erl_interface, exposed by libei.a.
host-erlang builds correctly libei.a _before_ odbcserver.c (it can be
found in lib/erl_interface/obj/x86_64-pc-linux-gnu/libei.a), but the
failure is actually due to the build command generated and used for
odbcserver.c, especially the link arguments:
/usr/bin/gcc \
[...]
-o ../priv/bin/x86_64-pc-linux-gnu/odbcserver \
../priv/obj/x86_64-pc-linux-gnu/odbcserver.o \
-L/usr/lib64 \
-lodbc \
-L/home/alexis/src/buildroot/erlang-master/build/host-erlang-custom/lib/erl_interface/obj/x86_64-pc-linux-gnu \
-lpthread -lei
/usr/lib64 is searched before the path where libei.a has been built, so
if whether a valid libei.a or libei.so is found there, it shadows the
expected libei.a. In the build from which the logs above come, the
notable point is that the host system indeed have a valid libei.so, but
is completely unrelated to erl_interface; it rather exposes the Emulated
Input protocol aimed at Wayland stack; and so it obviously contains none
of the expected ei_* symbols.
Upstream has already identified and fixed the issue, the fix is already
released in versions >= 27.x.y. Erlang 26 (the version currently
packaged in buildroot), isn't supported anymore (only the three latest
releases are supported, see
https://github.com/erlang/otp/blob/master/SECURITY.md), so there won't
be any new minor update that will release this fix.
Pick and backport the fixing patch so that the current version packaged
in buildroot can still build.
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The VERSAL2_CONSOLE variable does not actually exist when building the atf
with plat=versal2. So the current VERSAL2_CONSOLE=cadence1 is not actually
doing anything. For this reason, the atf will print on its default console
which is UART0 or pl011_0.
The correct definition would be CONSOLE=pl011_1 in order to build the atf to
print on UART1 or pl011_1.
However, the git repo xparameters.h of the versal2_plm is not currently
enabling UART1 because the XPAR_XUARTPSV_NUM_INSTANCES is set to 1 including
the address defines only for UART0.
97f2baf7f6/lib/sw_apps/versal_plm/misc/versal_2ve_2vm/xparameters.h (L1519)
The problem with this is that the plm is not configuring UART1, so the atf
will crash at boot time, if it is built with CONSOLE=pl011_1 and the plm has
not already enabled UART1.
For now, we will remove the unnecessary config that is doing nothing. The
versal2_vek385_defconfig is working, but users need to currently open two
console windows to see the boot log because the current config is the
following.
plm - console uart0
asufw - console uart0
atf - console uart0
optee-os - console uart1
u-boot - console uart1
Linux - console uart1
A patch has been submitted to the embeddedsw repo to fix the xparameters.h
file such that the plm will correctly enable UART1. Once this is applied, we
will configure the plm and atf to use UART1 with the versal2_vek385_defconfig.
Signed-off-by: Neal Frager <neal.frager@amd.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
https://github.com/strukturag/libheif/releases/tag/v1.23.2
Fixes the following CVEs:
(CVE numbers will be added upstream when assigned.)
CVE-2026-XXXXX (GHSA-g89c-p67h-r497)
Heap buffer overflow in scale_nearest_neighbor() via duplicate alpha
planes from nested iden/auxl items. (critical)
(GHSA-2jg2-4ch7-h545)
Out-of-bounds read and write in derived-item and pixel-plane handling.
Through iden and auxl item chains, a crafted file could attach pixel
planes whose size differs from the image geometry; crop, scale, and
plane-extraction code then indexed those planes with the wrong size.
A working code-execution exploit was confirmed. Plane sizes are now
validated wherever they are consumed. (critical)
CVE-2026-XXXXX (GHSA-24wx-9w62-c96w)
brotli/zlib decompression of mime metadata and unci image data had no
effective output-size limit, so a decompression bomb could exhaust
memory. Decompressed output is now bounded by the security limits.
(high)
CVE-2026-XXXXX (GHSA-x8xm-cm2c-cfc8)
Chains of derived-image references (grid, iovl, iden) bypassed decode
caching and memory limits, causing CPU and memory amplification. (high)
CVE-2026-XXXXX (GHSA-xw34-mjcp-jqh8)
Sequence sample-timing initialization could produce non-terminating
decode loops and unbounded memory, bypassing max_sequence_frames.
(high)
CVE-2026-XXXXX (GHSA-j264-xvrp-5v7q)
Out-of-bounds write in the unci encoder when
heif_context_add_image_tile() is given a tile whose planes do not match
its declared size. (high)
CVE-2026-XXXXX (GHSA-p58j-h3vm-3fp5)
Heap out-of-bounds read in the inline-mask region API when
mask_data_len does not match the region geometry. (medium)
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since the bump of libxml2 from 2.13.8 to 2.15.0 in Buildroot commit
d81922c1ef, the "virt" plugin of
collectd no longer builds:
src/virt.c:2208:49: error: expected ';', ',' or ')' before 'ATTRIBUTE_UNUSED'
2208 | static void virt_eventloop_timeout_cb(int timer ATTRIBUTE_UNUSED,
| ^~~~~~~~~~~~~~~~
src/virt.c: In function 'register_event_impl':
src/virt.c:2221:26: error: 'virt_eventloop_timeout_cb' undeclared (first use in this function)
2221 | virt_eventloop_timeout_cb, NULL, NULL) < 0) {
| ^~~~~~~~~~~~~~~~~~~~~~~~~
src/virt.c:2221:26: note: each undeclared identifier is reported only once for each function it appears in
This is due to the fact that the virt plugin code was incorrectly
using the ATTRIBUTE_UNUSED define, which was supposed to be an
internal define of libxml2. But it turns out that up to libxml2 2.14.0
and its commit 208f27f9641a59863ce1f7d4992df77f7eb0ea9d, this define
had been made publicly available. It could therefore mistakenly be
used by collectd's virt plugin... until libxml2 was upgraded.
We backport an upstream patch from collectd that fixes the issue.
Fixes:
https://autobuild.buildroot.net/results/4c8463f0372560f4c3a20b0f67854460f0d1c400/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
This commit adds a simple bpftrace test that ensures that not only it
builds fine, but it also runs properly on a minimal test scenario.
Assisted-by: GPT-5.6
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Julien:
- reindent emulator.boot() options
- add a call to "bpftrace --version"
]
Signed-off-by: Julien Olivain <ju.o@free.fr>
Since Buildroot commit 25cb3813e7 which
bumped LLVM from 21.x to 22.x, the build of bpftrace is broken as
bpftrace 0.24.2 only supports LLVM up to 21.
We tried backporting the bpftrace patch that allows using LLVM up to
version 22 but the patch didn't apply cleanly but more importantly it
wasn't clear if this patch was sufficient. Therefore, we opt for
bumping bpftrace entirely to fix the issue.
Packaging changes:
- The new version of bpftrace no longer needs host-bison/host-flex,
because bpftrace is now using a handwritten parser.
- Pass -DUSE_SYSTEM_LIBBPF:BOOL=ON to ensure the system libbpf version
is used, and not the bundled version
- Now depends on kernel headers >= 5.10 because it needs CAP_BPF and
CAP_PERFMON
Upstream changelog:
https://github.com/bpftrace/bpftrace/blob/v0.26.1/CHANGELOG.md
Fixes:
https://autobuild.buildroot.net/results/a3e3fd696685864977c688aa11c2653357cf6207/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Julien: add link to upstream changelog]
Signed-off-by: Julien Olivain <ju.o@free.fr>
bpftrace often needs very recent kernel headers, more recent than the
runtime version actually needed. We already had a workaround in
libbpf making sure that if the kernel headers are older than 6.1, we
would install the libbpf provided headers instead.
As we are about to update bpftrace to a newer version that uses
BPF_TRACE_KPROBE_SESSION, which was introduced in Linux 6.10, we need
to update this workaround accordingly and ensure that the libbpf
header is installed if the kernel headers are older than 6.10.
This is necessary for the update of bpftrace to 0.26.1.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Since commit
566bdcb97f ("package/fluidsynth:
security bump to version 2.5.7"), fluidsynth tries to download some
"gcem" code during its configure step, which not only violates
Buildroot's policies, but also breaks the build if network is not
available during the build.
To fix this, we add an EXTRA_DOWNLOADS to grab gcem and extract it at
the right place. Some minor fix (submitted upstream) is needed to
ensure the FindGCEM.cmake logic properly finds that gcem is already in
the source tree.
Fixes:
https://autobuild.buildroot.net/results/048df28f6ab97a16731e62d7f56c6eba565cda63/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Based on the lighttpd test case. Verify that we can download index.html
from haproxy in front of lighttpd.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
haproxy has a runtime test to verify that it is built with -fwrapv:
haproxy
FATAL ERROR: invalid code detected -- cannot go further, please recompile!
The source code was miscompiled by the compiler, which usually indicates that
some of the CFLAGS needed to work around overzealous compiler optimizations
were overwritten at build time. Please do not force CFLAGS, and read Makefile
and INSTALL files to decide on the best way to pass your local build options.
Build options :
TARGET = custom
CPU = generic
CC = /home/peko/source/buildroot/output-haproxy/host/bin/arm-linux-gcc
CFLAGS = -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -O2 -g0 -D_FORTIFY_SOURCE=1 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -O2 -g0 -D_FORTIFY_SOURCE=1
OPTIONS = USE_THREAD=1 USE_DL=1
DEBUG = -DDEBUG_STRICT -DDEBUG_MEMORY_POOLS
Which comes from:
https://github.com/haproxy/haproxy/blob/v2.6.0/src/haproxy.c#L3008-L3037
So build it with -fwrapv to fix that.
Notice that this message also embeds the build path (through CC), breaking
reproducible builds.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Since upstream commit
e42e3193137d1b21e84b499336e7a8887b8a8689 ("intel: add Jay"), a C23
construct is used in the intel driver code:
enum jay_predication : uint8_t
This causes a build issue with GCC < 13:
In file included from ../src/intel/compiler/jay/jay_builder.h:12,
from ../src/intel/compiler/jay/jay_from_nir.c:23:
../src/intel/compiler/jay/jay_ir.h:582:22: error: expected identifier or ‘(’ before ‘:’ token
582 | enum jay_predication : uint8_t {
| ^
../src/intel/compiler/jay/jay_ir.h:637:25: error: field ‘predication’ has incomplete type
637 | enum jay_predication predication;
| ^~~~~~~~~~~
We fix that by integrating a patch already available in
OpenEmbedded. It changes the code to not use the C23 construct.
This build issue was encountered on host-mesa3d while building an
allyespackageconfig configuration inside our standard Docker
container.
Buildroot commit
c073c97617 ("package/{mesa3d,
mesa3d-headers}: bump version to 26.1.0") that switched to mesa3d
26.1.0, which contains the problematic commit. Therefore 2026.05 is
affected, but not earlier Buildroot versions.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
perl does not build with musl due to memrchr() being unavailable. This
is caused by a perl-cross bug, which does function availability
detection with _GNU_SOURCE defined, but then does the build without
_GNU_SOURCE defined. At least OpenEmbedded and NixOS have faced the
same issue, and worked it around in slightly different ways.
On our side, we create a patch, which was submitted upstream, to solve
the issue.
This issue has been introduced in perl-cross commit b40c560f5d5e,
which was first merged in perl-cross release 1.4.1. From a Buildroot
perspective, we bumped from perl-cross 1.4 to 1.4.1 in commit
8a289667f5, which was merged
2023.05. And indeed the build failure can be reproduced even on our
LTS 2025.02.x, so the fix needs to be backported there.
It should be noted that even if the patch is against perl-cross, we
add it to package/perl/ directly, as patches in perl are applied after
perl has been extracted *and* perl-cross has been extracted on top.
Fixes:
https://autobuild.buildroot.net/results/3e47ade0963642988fd8e1be9a6e8042700619ec/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Since the bump of autoconf to version 2.73 in Buildroot commit, the
build of target bind fails if the host compiler is too old, because
the bind build system tries to use -std=gnu23 when building host tools
which isn't supported by older GCC releases, causing:
checking whether the C compiler works... no
configure: error: in '/home/thomas/autobuild/instance-2/output-1/build/bind-9.20.26':
configure: error: C compiler cannot create executables
See 'config.log' for more details
make: *** [package/pkg-generic.mk:263: /home/thomas/autobuild/instance-2/output-1/build/bind-9.20.26/.stamp_configured] Error 77
To fix this, we backport a number of patches from autoconf-archive, to
fix the m4/ax_prog_cc_for_build.m4 macro file, so that it works with
autoconf 2.73.
OpenEmbedded has a similar fix:
https://git.openembedded.org/openembedded-core/tree/meta/recipes-connectivity/bind/bind/0001-m4-Backport-ax_prog_cc_for_build.m4-macros.patch
but did not backport as carefully the autoconf-archive commits (and
their commit message reverts to sudo).
The patches can be dropped when we update to a newer version of bind
that itself has an updated copy of the m4/ax_prog_cc_for_build.m4
file.
Fixes:
https://autobuild.buildroot.net/results/13e07755101b7cae2f84eef173ef752f92842e71/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
The BR2_PACKAGE_QT6DECLARATIVE_QUICK option has some logic to select
network support in host-qt6base if network support is enabled in
qt6base. However, it turns out that this is actually required at the
top level BR2_PACKAGE_QT6DECLARATIVE option: as soon as network
support is available in qt6base, the qt6declarative build will assume
that qmlprofiler is available... but that requires network support in
host-qt6base.
This fixes the following build failure:
CMake Error at /home/thomas/autobuild/instance-2/output-1/build/qt6base-6.9.1/cmake/QtToolHelpers.cmake:784 (message):
Failed to find the host tool "Qt6::qmlprofiler". It is part of the
Qt6QmlTools package, but the package did not contain the tool. Make sure
that the host module Qml was built with all features enabled (no explicitly
disabled tools).
Call Stack (most recent call first):
/home/thomas/autobuild/instance-2/output-1/build/qt6base-6.9.1/cmake/QtToolHelpers.cmake:83 (qt_internal_find_tool)
tools/qmlprofiler/CMakeLists.txt:11 (qt_internal_add_tool)
Fixes:
https://autobuild.buildroot.net/results/72c956fdf982382d2981c649c456d1edc2c9d6b2/
We did not trace back exactly since when the problem exists, but we
verified that the problem exists in 2025.02.x. It can be reproduced
with the following defconfig:
BR2_aarch64=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN_AARCH64_GLIBC_STABLE=y
BR2_PACKAGE_QT6=y
BR2_PACKAGE_QT6BASE_NETWORK=y
BR2_PACKAGE_QT6DECLARATIVE=y
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
The commit "Enable host test module to ensure that qmltestrunner is
built" in qt6declarative's Config.in feels lonely under
BR2_PACKAGE_QT6DECLARATIVE. It's because it's actually related to a
select done in the sub-option BR2_PACKAGE_QT6DECLARATIVE_QUICK, so
move it there.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Since Firewalld v2.4.3 [1] firewall-cmd added a new log line while
waiting for dbus connection [2].
[BRTEST# firewall-cmd --state
Waiting on dbus connection...
running
The line "Waiting on dbus connection..." is not always printed by
firewall-cmd, so we have to search explicitely for the expected
string to get a reproducible test result.
Update both TestFirewalld accordingly.
This change is required to fix:
https://gitlab.com/buildroot.org/buildroot/-/jobs/16060152329 (TestFirewalldSysVInit)
https://gitlab.com/buildroot.org/buildroot/-/jobs/16060152332 (TestFirewalldSystemd)
[1] 380dd8a348
[2] 5e1c37c966
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
In Glib >= 2.88, GLib.unix_signal_add has been moved to a separate
platform-specific library. This break backward compatibility from
GLib-2.0. A workaround has been applied to pygobject >= 3.55.3
74e4e0f40a
This issue currently break TestFirewalldSysVInit and
TestFirewalldSystemd runtime tests since the bump to glib 2.88.3 [1]:
https://gitlab.com/buildroot.org/buildroot/-/jobs/16060152329 (TestFirewalldSysVInit)
https://gitlab.com/buildroot.org/buildroot/-/jobs/16060152332 (TestFirewalldSystemd)
Break the test TestPythonPy3Gobject now in order to reproduce the same
issue than for Firewalld test.
[1] e313a2d259
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
The mdnsd runtime test can randomly fail on slow runners.
It's hard to reproduce locally (only one failure after a few attempts)
but we can reproduce it easily by removing the while loop entirely.
It means that the "sleep 1" is not used on the Gitlab runner.
The timestamp of the failed job seems to confirm that [1].
07:06:55 [BRTEST# while ! ifconfig eth0 | grep -q 'inet addr'; do sleep 1; done
07:06:55 [BRTEST# echo $?
07:06:55 0
07:06:55 [BRTEST# mquery -T _http._tcp |grep -F buildroot._http._tcp.local
07:06:55 [BRTEST# echo $?
07:06:55 1
So wait a bit for mdnsd to be ready.
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/16060152862
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Buildroot commit [1] (package/apache: use "Debian" filesystem
layout to fix read-only rootfs) changed the filesystem layout.
This had the effect of installing files to different locations
and breaking the test_php runtime test.
This commit fixes the issue by updating the file paths to their
right locations. The "httpd.conf" was updated by following the
same recipe described in the comment (starting from a config
file as installed by the apache Buildroot package).
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/16060152979
[1] 1006666f67
Signed-off-by: Julien Olivain <ju.o@free.fr>
Tested-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Since the bump of autoconf to version 2.73 in Buildroot commit [1], the
build of target flex fails if the host compiler is too old, because
the flex build system tries to use -std=gnu23 which isn't supported by
older GCC releases, causing:
gcc: error: unrecognized command-line option '-std=gnu23'; did you mean '-std=gnu2x'?
gcc: error: unrecognized command-line option '-std=gnu23'; did you mean '-std=gnu2x'?
make[3]: *** [Makefile:1162: stage1flex-buf.o] Error 1
(Indeed the *target* flex package does build some host tools using the
host GCC compiler.)
To fix this issue, we backport an upstream commit that isn't yet in
any flex release.
Fixes:
https://autobuild.buildroot.net/results/aac730b57adb5b54964f1054de781750952ef7d4/
[1] a6e8c07a33
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Julien: add link to commit]
Signed-off-by: Julien Olivain <ju.o@free.fr>
