The FRR package currently only installs systemv services files. The
upstream package also provides a systemd service file and a service
template file. Install those when systemd is selected as an init
manager.
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Marcus Hoffmann <buildroot@bubu1.eu>
And updates the kernel version used for the test_openjdk test to
the last LTS version (6.18.21). instead of 5.10.3, which is quite
old and may not be compatible with the latest OpenJDK versions.
This change was tested by running the openjdk test.
Signed-off-by: Dowan Gullient <dowan.gullient@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Update the Device Tree name for the vexpress platform to match the
new directory structure introduced in Linux 6.5.
Since kernel 6.5, the 'arch/arm/boot/dts/' directory was restructured
to avoid having thousands of files in a single folder. Device Trees
for ARM (32-bit) are now organized into subdirectories. For the
Versatile Express platform, the DTS file moved to the 'arm/'
subdirectory.
The build was failing with:
make[3]: *** No rule to make target 'arch/arm/boot/dts/vexpress-v2p-ca9.dtb'. Stop.
Adjust BR2_LINUX_KERNEL_INTREE_DTS_NAME from "vexpress-v2p-ca9" to
"arm/vexpress-v2p-ca9".
Signed-off-by: Dowan Gullient <dowan.gullient@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Update the Device Tree name for the vexpress platform to match the
new directory structure introduced in Linux 6.5.
Since kernel 6.5, the 'arch/arm/boot/dts/' directory was restructured
to avoid having thousands of files in a single folder. Device Trees
for ARM (32-bit) are now organized into subdirectories. For the
Versatile Express platform, the DTS file moved to the 'arm/'
subdirectory.
The build was failing with:
make[3]: *** No rule to make target 'arch/arm/boot/dts/vexpress-v2p-ca9.dtb'. Stop.
Adjust BR2_LINUX_KERNEL_INTREE_DTS_NAME from "vexpress-v2p-ca9" to
"arm/vexpress-v2p-ca9".
Signed-off-by: Dowan Gullient <dowan.gullient@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
This commit bump the kernel version of this docker test (5.4->6.18), and thus
fixes several issues preventing the Docker Compose runtime test from
succeeding due to this bump:
Kernel Infrastructure:
The Docker daemon failed to start because the 'nat' table could not
be initialized. This was due to missing legacy Netfilter support.
- Enable CONFIG_NETFILTER_XTABLES_LEGACY to support iptables nat table.
- Enable CONFIG_NAMESPACES and CONFIG_USER_NS for container isolation.
- Enable CONFIG_TUN and CONFIG_VETH for container networking.
- Enable CONFIG_IKCONFIG and CONFIG_IKCONFIG_PROC for easier debugging.
Build Fixes:
While building the kernel tools, the libelf header was missing because
objtool was not handled correctly for this kernel version. This resulted
in a fatal error: "gelf.h: No such file or directory".
- Select BR2_LINUX_KERNEL_NEEDS_HOST_LIBELF=y to build host-libelf.
Resources and Stability:
The previous 512MB of RAM was tight for the stack (Kernel +
Docker + Containerd + Python), potentially leading to Out-Of-Memory kills.
Idem for the disk size for Docker images/layers storage.
- Increase QEMU RAM to 1024MB (-m 1024M).
- Increase rootfs size to 1024MB.
- Switch to EXT4 (BR2_TARGET_ROOTFS_EXT2_4=y) for better stability
and modern feature support required by Docker's storage drivers.
Test Script Adjustments:
- Update rootfs path in the test script to point to rootfs.ext4.
Signed-off-by: Dowan Gullient <dowan.gullient@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Update the kernel configuration to support modern storage stacks
required for ISO booting on x86_64.
The transition from 4.19 to 6.18 requires explicit activation of:
- CONFIG_PCI: To discover the emulated IDE/SATA controllers.
- CONFIG_ATA & CONFIG_ATA_PIIX: Modern libATA drivers for QEMU's
chipset (replacing the legacy IDE subsystem).
- CONFIG_SCSI & CONFIG_BLK_DEV_SR: Necessary to handle the CD-ROM
as a SCSI device (/dev/sr0), which is mandatory for ISO9660.
Without these options, the kernel cannot locate or mount the
rootfs from the ISO image.
Signed-off-by: Dowan Gullient <dowan.gullient@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
BR2_LINUX_KERNEL_NEEDS_HOST_LIBELF=y :
While building the kernel tools, libelf header is missing because objtools
was not compiled by default on older kernel versions. This results in the following error:
[...]tools/objtool/include/objtool/elf.h:10:10: fatal error: gelf.h: No such file or directory
10 | #include <gelf.h>
That is why BR2_LINUX_KERNEL_NEEDS_HOST_LIBELF was selected to build host-libelf.
Signed-off-by: Dowan Gullient <dowan.gullient@smile.fr>
[Julien: updated kernel to 6.18.21]
Signed-off-by: Julien Olivain <ju.o@free.fr>
Unlike libgpiod2, this version of libgpiod relies on the deprecated GPIO
ABI version 1. This ABI used to be enabled by default. However, since
Linux 7.1-rc1 (commit 7673e4c7f7f9 "Do not enable the v1 uAPI by
default"), this is not the case anymore. Failing to enable the ABI
results in a non-functional libgpiod. For instance:
# gpioinfo
gpioinfo: error creating line iterator: Invalid argument
Fix this by enabling the ABI v1 in LIBGPIOD_LINUX_CONFIG_FIXUPS.
Signed-off-by: Edgar Bonet <bonet@grenoble.cnrs.fr>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
The __name__ == '__main__' guard allows importing pkg-stats as a
module using importlib, circumventing the normal module filename
requirements. This in turn makes it possible to test/debug individual
functions.
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
The file handle can be iterated over directly and each line is used
exactly once, so the only effect of reading all lines into a list
first was higher memory use and complexity.
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
The filter is supposed to exclude host/target infra from output if the
respective package is not built with the current
configuration.
However, excluding host packages did not work correctly: If keep_host
is False because the host package is not built, the next branch was
checked and included the host infra in output with "target" type if
the target package is built. For a package that support host and
target build, but gets built only for the target, this leads to output
like (Meson example):
meson (target)
host-meson (target)
Skip host infra in the target branch instead. Also include
Package.infra in Package.__str__() result, which was needed for
debugging this bug.
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
[Arnout: add it to the main help text, where it belongs]
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
Use only one of the classes for "error" or "warning" status so they
look different, and format the error/warning text for both. Do not
make the text a link if the URL is None.
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
The previous Config.* glob also caught linux/Config.ext.in and
package/php/Config.ext, as well as some backup files created by
editors (e.g. Config.in~ after editing a Config.in file in Emacs),
leading to wrong results depending on directory listing order.
Also use "with" to automatically close the file when the block is
left, even on error.
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
I don't use maintain x11vnc in buildroot anymore and can't easily test.
Signed-off-by: Martin Kepplinger-Novaković <martink@posteo.de>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Eudev install creates a symlink from $(sbindir)/udevadm to
$(bindir)/udevadm if sbindir and bindir are different [1].
Unfortunately the check considers only the configured values as
strings and does not resolve symlinks. This means that in a merged-bin
setup with sbindir=/sbin (symlink to /usr/sbin, which is a symlink to
/usr/bin) and bindir=/usr/bin the /usr/bin/udevadm binary gets
overwritten with a link to itself.
Set sbindir according to the merge setup to avoid the problem, bindir
defaults to /usr/bin in any case.
[1] 9e7c4e744b/src/udev/Makefile.am (L139-L140)
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Marcus Hoffmann <buildroot@bubu1.eu>
If sending the stop signal fails but the PID file exists the following
shutdown wait turns into an endless loop. Avoid that and return the
error immediately.
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Marcus Hoffmann <buildroot@bubu1.eu>
If sending the stop signal fails but the PID file exists the following
shutdown wait turns into an endless loop. Avoid that and return the
error immediately.
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Marcus Hoffmann <buildroot@bubu1.eu>
With dlopen() enabled, kmod will load compression libraries only if
actually needed (loading compressed modules into the kernel does not
if the kernel can handle decompression, depending on its
configuration), and only the needed one.
Use compressed modules in the kmod test without enabling in-kernel
decompression to verify decompression in kmod works.
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
Signed-off-by: Marcus Hoffmann <buildroot@bubu1.eu>
The upstream changelog for version 34 states that the next major
version will drop the Autotools build, and the 34 series is meant for
distros to switch.
Notable changes in the Buildroot package:
* The Meson build handles symlink install, so we can drop the post
install hooks.
* Simplify the merged-usr and merged-bin symlink install handling.
Only the non merged-usr case needs special handling now for placing
the correct symlink to block busybox placing it's only kmod applet
symlink unto /sbin later.
* The previous point also means that host-kmod now installs the full
set of tools.
* Explicitly disable openssl support in host-kmod tools. It is enabled
by default and the build fails if OpenSSL cannot be found. In the
autotools build it was not explicitly set.
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
[Marcus: explain the merged-bin, merged-usr simplifications in the
commit message]
Signed-off-by: Marcus Hoffmann <buildroot@bubu1.eu>
Patch the parts of Make that use code invalid in C23. It wasn't
included when compiling with glibc but failed to build with musl.
Fixes#176. See also https://savannah.gnu.org/bugs/index.php?68302
Signed-off-by: Daniel Houck <Software@DRHouck.me>
[Marcus: link to savannah bug report also in the patchfile]
Signed-off-by: Marcus Hoffmann <buildroot@bubu1.eu>
OpenSSH checks during ./configure if the default location /var/run
exists on the build host, and defaults to ${sysconfdir} (/etc/ssh in
Buildroot) if not. This can break e.g. in a sandbox with its own mount
namespace where there is no reason to create a /var.
On the target this leads to errors when starting/stopping the service
from the init script, which expects the PID file at /var/run/sshd.pid if
sshd is running. A "stop" (including as part of "restart") or "reload"
will always fail with
no /usr/sbin/sshd found; none killed
from start-stop-daemon, even though sshd is running (because the PID
file doesn't exist where start-stop-daemon tries to read it).
On the other hand, a "start" command will try to start sshd even if it
is already running, because it cannot detect the running instance.
Unfortunately sshd does not fail directly in that situation (so the init
script completes), but something like the following appears in the log:
[May 04 14:48:05] auth sshd[4592]: error: Bind to port 22 on 0.0.0.0
failed: Address in use.
[May 04 14:48:05] auth sshd[4592]: error: Bind to port 22 on :: failed:
Address in use.
[May 04 14:48:05] auth sshd[4592]: fatal: Cannot bind any address.
Explicitly setting --with-pid-dir=/var/run overrides either default,
and only logs a warning if /var/run does not exist.
[1] 5fa60ae0d2/configure.ac (L5568-L5576)
Signed-off-by: Fiona Klute <fiona.klute@gmx.de>
[Marcus: expand commit message after discussion on ML]
Signed-off-by: Marcus Hoffmann <buildroot@bubu1.eu>
Fixes the following security issue:
CVE-2026-4046: iconv crash due to assertion failure with untrusted input
https://sourceware.org/git/?p=glibc.git;a=blob;f=advisories/GLIBC-SA-2026-0007;h=b880fb55449693b879beae443e5d9bc1070b938b;hb=HEAD
git shortlog 2.43-17-gdd9945c0b..2.43-22-g8362e8ce1
Adhemerval Zanella (1):
elf: Use dl-symbol-redir-ifunc.h instead _dl_strlen
Adhemerval Zanella Netto (1):
riscv: Resolve calls to memcpy using memcpy-generic in early startup
Florian Weimer (1):
Use pending character state in IBM1390, IBM1399 character sets (CVE-2026-4046)
Michael Jeanson (1):
tests: fix tst-rseq with Linux 7.0
Xi Ruoyao (1):
elf: parse /proc/self/maps as the last resort to find the gap for tst-link-map-contiguous-ldso
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Bugfix release fixing crash / NULL pointer access issues and gsasl
authentication compatibility with Microsoft servers.
Mutt 2.2.14 was released on February 20, 2025. This is a bug-fix release,
fixing a couple possible crashes, in GPGME and with IMAP when logging out,
and correcting a small issue with unnecessary encoding of "." in attachment
names (2231 encoding).
Mutt 2.2.15 was released on October 2, 2025. This is a bug-fix release,
fixing an IMAP search/limit bug, a compilation issue, and containing a
workaround for gsasl SMTP authentication issues with some MS servers. A few
parts of the documentation were also improved.
Mutt 2.2.16 was released on November 22, 2025. This is a bug-fix release,
fixing a resource leak when compiled with OpenSSL/LibreSSL, which could
eventually result in new connections failing.
http://www.mutt.org/news.html
The 2.2.16 release is not available on bitbucket, so change to the official
ftp.mutt.org location instead.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
Bugfix release with large number of (security) fixes.
For 6.2.26:
- a severe issue was found in the compression library (slz) where
specially crafted patterns with tune.bufsize above 17408 or
tune.maxrewrite below 964 (both non-default) could cause output
buffer overflows due to the overhead exceeding the promised
worst-case growth bound of 5 bytes and reach up to 1/16 of the
input contents. Given that the compression output is hardly
controllable, and the canaries at the end of the pools will catch
this at release time, the risk of exploitation by a hostile server
is close to zero, however it will cause repeated crashes if such a
crafted file is present on a server and regularly downloaded. A
workaround consists in keeping tune.maxrewrite at least 1/16 of
tune.bufsize or just not changing them since the defaults are safe.
A CVE was requested two weeks ago for this one, I'll mention it when
it arrives.
- HTTP/2 incomplete transfer detection was missing for HEADERS frames
carrying END_STREAM. When relayed to an HTTP/1.1 server that
responds before the end of the transfer, this can result in bytes
of the next request over the same connection to be ignored. Most of
the time it will cause the connection to be dropped due to an
unparsable request, but when combined with "http-reuse never", or
on totally idle servers, the client could expect the second request
to reuse the same connection and perform a content smuggling attack
that would allow to pass an unverified request to a server. For
those who can't upgrade, a temporary workaround is to disable
HTTP/2 by specifying "alpn http/1.1" on bind lines and adding
"disable-h2-upgrade" in HTTP frontends. A CVE will be requested for
this one.
- HTTP/1.1 bodyless messages announcing a non-null Content-Length did
not force close mode on the backend, potentially causing
desynchronisation between HAProxy and the server in conjunction
with other bugs.
- FCGI record length truncation with large bufsize (>=65544) could
enable request smuggling into PHP-FPM since the 16-bit
content_length field silently truncated to 65535 bytes.
- an unvalidated SNI name_len field in ClientHello could cause OOB
heap reads of up to 65KB via XXH3, smp_dup(), and log-format leaks
on any TCP frontend using req.ssl_sni, possibly causing crashes when
used.
- ECDSA JWT signatures with ES256/384/512 could cause a heap overflow
of ~14 bytes in the DER conversion before verification.
- Lua's httpclient headers conversion accepted more than 101 headers
without bound checking, causing a stack buffer overflow reachable
from any Lua action/task/service.
- peers dictionary cache updates accepted an unvalidated entry id as
array index, allowing OOB heap writes at attacker-controlled
offsets.
- Lua had a use-after-free of HTTP reason strings managed by Lua's GC
between set_status() and start_response(), potentially leaking
adjacent information from memory.
- the regsub sample function could leak ~9-50KB of stale heap data
when back-reference expansion overflowed the output buffer.
- SPOE decode_varint() had no iteration cap, allowing pointer
arithmetic to wrap and dereference memory ~64KB before the
allocation, causing SIGSEGV or parser confusion.
- in sample expressions, less common HTTP methods (PATCH etc.) are
represented by both an enum and a string. The string part was not
handled correctly in sample duplication functions, resulting in
their contents appearing empty when trying to fetch the method.
- QPACK varint decoding is now also limited to 62-bit, and had a risk
of 1-byte OOB reads on truncated streams, which could cause
incorrect header decoding.
- config: a few argument parsing errors in conditional expressions
used in ".if" could be misreported and even cause a crash during
the parsing. Also, a few keywords relying on warnif_misplaced_*
didn't check the return value and didn't count emitted warnings as
warnings.
For more details, see the announcement:
https://www.mail-archive.com/haproxy@formilux.org/msg47016.html
For 6.2.27:
A major issue were fixed by this release. It was related to the scheme-based
normalization. The presence of commas in Host header and authority was permitted
and would be used to compare the values, which then would differ when read via
hdr(host) which splits them on commas, and under certain circumstances, trigger
crashes (at least it did in the OSS-Fuzz environment when injecting the values
directly at the HTX layer). The issue was fixed. Remains the case of the comma
characters in authorities. Even though the spec permits commas in authorities
(not in domain names), there is currently no use case for this and it causes an
ambiguity with the historical use of hdr(host), so we preferred to just deny
them. The change was performed on the 3.4-dev10 and postponed for the next 3.3
release. It will probably be backported to lower versions too.
An issue in the FCGI multiplexer was fixed. The function responsible to emit
FCGI_PARAM records was not handling cases of full buffer in a consistent
way. The issue was quite limited, but the "http-send-name-header" option could
be silently ignored. The issue was fixed by reworking this function.
The scheme-based normalization was fixed to properly handle case of OPTIONS
requests. As stated in RFC9110#4.2.3, when the scheme-based normalization is
performed, an empty path must be normalized to "/", except for OPTIONS request.
Finally, a memory leak on error path (tools) and other minor issues were also
fixed.
For more details, see the announcement:
https://www.mail-archive.com/haproxy@formilux.org/msg47059.html
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
