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>
(cherry picked from commit 94aa7f40b5)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
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>
(cherry picked from commit d950fff290)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
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>
(cherry picked from commit 50a1dd2676)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
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>
(cherry picked from commit 29add67666)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
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>
(cherry picked from commit 23fd881bdb)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
In commit 54f96add94 ("package/bind:
security bump version to 9.20.24") the depends on
BR2_TOOLCHAIN_HAS_THREADS_NPTL was incorrectly downgraded to
BR2_TOOLCHAIN_HAS_THREADS:
- depends on BR2_TOOLCHAIN_HAS_THREADS_NPTL # libuv
+ depends on BR2_TOOLCHAIN_HAS_THREADS # liburcu, libuv
This is wrong because libuv depends on
BR2_TOOLCHAIN_HAS_THREADS_NPTL. This causes unmet dependencies:
WARNING: unmet direct dependencies detected for BR2_PACKAGE_LIBUV
Depends on [n]: BR2_TOOLCHAIN_HAS_THREADS_NPTL [=n] && BR2_USE_MMU [=y] && !BR2_STATIC_LIBS [=n] && BR2_TOOLCHAIN_HAS_SYNC_4 [=y] && BR2_TOOLCHAIN_GCC_AT_LEAST_4_9 [=y]
Selected by [y]:
- BR2_PACKAGE_BIND [=y] && BR2_USE_MMU [=y] && BR2_TOOLCHAIN_HAS_SYNC_4 [=y] && BR2_TOOLCHAIN_HAS_THREADS [=y] && BR2_INSTALL_LIBSTDCPP [=y] && !BR2_STATIC_LIBS [=n] && BR2_TOOLCHAIN_GCC_AT_LEAST_4_9 [=y] && BR2_PACKAGE_LIBURCU_ARCH_SUPPORTS [=y]
Fix that by switching back to the BR2_TOOLCHAIN_HAS_THREADS_NPTL
dependency.
Fixes: 54f96add94 ("package/bind: security bump version to 9.20.24")
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit d8dde961bc)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
- CVE-2026-19654:
A unauthenticated remote peer may lead rsyslogd to crash due to a flaw
in the optional imptcp module. A crafted input sequence during
oversize-frame recovery can cause an invalid internal message length
and terminate rsyslogd. No confidentiality or integrity impact,
privilege escalation, or code execution has been identified. imtcp and
the default imptcp framing modes are not affected.
For more information, see:
- https://www.cve.org/CVERecord?id=CVE-2026-19654
- 07b3c40a5a
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 52ae04257a)
[raphael: backport patch for rsyslog 8.2412.0]
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
See the changelog:
- https://nlnetlabs.nl/projects/unbound/download/#unbound-1-25-2
It fixes the following vulnerabilities:
- CVE-2026-14586: Assertion in libngtcp2 when under pressure in high
concurrency DNS-over-QUIC environments.
- CVE-2026-32665: Remote DNS-over-QUIC denial of service due to
`quic-size` budget bypass.
- CVE-2026-40691: Packet of death for DNSCrypt over TCP.
- CVE-2026-41637 Degradation of resolution service from improperly
accounted client-terminated DNS-over-QUIC queries.
- CVE-2026-42955: Extra fix for CVE-2026-40622 to also clamp the TTL of
A/AAAA records disallowing a one-time 'ghost domain' delegation
renewal via glue records.
- CVE-2026-44621: Libunbound applications configured with
'unwanted-reply-threshold' could eventually be abruptly terminated.
- CVE-2026-44687: Off-by-one error in 'harden-below-nxdomain' logic can
shadow a stub/forward zone by a legitimate parent's NXDOMAIN.
- CVE-2026-44690: Cross-zone wildcard cache poisoning via RRSIG.labels
manipulation.
- CVE-2026-46582: A wildcard replay, as another piece of data, triggers
poisoning in the serve expired reply path.
- CVE-2026-50045: 'max-global-quota' reset by DNSSEC validation
restarts.
- CVE-2026-50046: Possible heap use-after-free in an error path when a
DoT forwarded query is jostled out.
- CVE-2026-50243: 'response-ip'/'rpz' can rewrite BOGUS answers instead
of returning SERVFAIL.
- CVE-2026-50248: BOGUS configured primary hostname accepted for XFR in
auth/rpz zones.
- CVE-2026-50251: Attacker supplied `0.0.0.0`/`::` glue triggers
defensive full-cache flush.
- CVE-2026-50252: Possible cache poisoning attack by mapping source port
population per thread.
- CVE-2026-52863: Memory corruption could lead to crash and denial of
service.
- CVE-2026-54478: DNS Cookie bypass when combined with proxy-protocol
use.
- CVE-2026-55708: Privacy/configuration issue when adding local data in
views through 'unbound-control'.
- CVE-2026-55717: 'serve-expired-client-timeout' and 'response-ip' CNAME
redirect could lead to a crash.
- CVE-2026-55973: 'dns-error-reporting: yes' leads to stack buffer
overflow.
- CVE-2026-55990: Packet of death for a DNSCrypt misconfigured Unbound.
- CVE-2026-55991: Remote DNS-over-QUIC (DoQ) flow-control assertion
failure in libngtcp2.
- CVE-2026-56416: Possible heap buffer overflow when validator
canonicalizes RDATA that contains domain name.
- CVE-2026-56444: Degradation of resolution service when
'discard-timeout' and 'serve-expired-client-timeout' are combined in
unusual configuration.
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit d0619dfc6b)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
This release includes a broad round of hardening against malformed and
adversarial input across the Python SDK (bounding allocations and enforcing
decompression limits before trusting size fields read from the input).
See the release notes https://avro.apache.org/blog/2026/08/12/avro-1.12.2/
Also update the download url, because www-eu.apache.org/dist/...
is a redirection to downloads.apache.org/...
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 02d8a41f09)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
In some situations (old Linux <3.17 or glibc <2.25), expat fail at
compilation time with the error:
xmlparse.c:150:4: error: #error You do not have support for any sources of high quality entropy enabled.
For end user security, that is probably not what you want. Your options include:
* Linux >=3.17 + glibc >=2.25 (getrandom): HAVE_GETRANDOM,
* Linux >=3.17 + glibc (including <2.25) (syscall SYS_getrandom): HAVE_SYSCALL_GETRANDOM,
* BSD / macOS >=10.7 / glibc >=2.36 (arc4random_buf): HAVE_ARC4RANDOM_BUF,
* BSD / macOS (including <10.7) / glibc >=2.36 (arc4random): HAVE_ARC4RANDOM,
* BSD / macOS >=10.12 / glibc >=2.25 (getentropy): HAVE_GETENTROPY,
* Linux (including <3.17) / BSD / macOS (including <10.7) / Solaris >=8 (/dev/urandom): XML_DEV_URANDOM,
* Windows >=Vista (rand_s): _WIN32.
If you insist on not using any of these, bypass this error by defining XML_POOR_ENTROPY and be vulnerable to hash flooding;
you have been warned. If you have reasons to patch this detection code away or need changes to the build system, please open a bug. Thank you!
This is caused by the upstream commit [1] "Autotools: Stop using
/dev/urandom by default", first included in expat 2.8.2. The
Buildroot expat package was bumped to that version in commit [2].
But since all Linux systems have /dev/urandom, we can just enable
it by default.
Note: this commit does not globally switch the entropy source to
/dev/urandom. It is rather enabling it in the list of available
sources. On more recent Linux systems (linux >= 3.17, glibc >= 2.25),
other sources will be chosen. The entropy source preference order
amongst the enabled sources is defined in [3].
This commit also changes the _CONF_OPTS to multiline layout to fit
within the 80 characters.
[1] d30eca113a
[2] 6b1f6f7a48
[3] https://github.com/libexpat/libexpat/blob/R_2_8_3/expat/lib/xmlparse.c#L1115-L1142
Signed-off-by: Thomas Devoogdt <thomas.devoogdt@barco.com>
[Julien: add extra info in the commit log]
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit c22fc74f2b)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
Since gdb 9.x, gdb can optionally use the xxhash library. Since we
currently don't do anything about it, it's a potential "silent"
dependency.
In particular, for host-gdb, this means host-gdb might end up being
linked with the system-provided xxhash library if available.
This patch handles this dependency:
- For the target package, by looking at the value of
BR2_PACKAGE_XXHASH
- For the host package, by looking at the value of a newly introduced
BR2_PACKAGE_HOST_GDB_XXHASH
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 9dc567aa78)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
For target gdb, we properly enable/disable lzma support depending on
BR2_PACKAGE_XZ.
However, for host-gdb we don't do anything, which can lead the gdb
configure script to detect and use a system-provided xz library, which
is not desired.
Instead, add an explicit option BR2_PACKAGE_HOST_GDB_LZMA, which when
enabled pulls in host-xz, but also when disabled ensures gdb doesn't
try to use a system-provided xz library.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit c3adba81d1)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
In commit a2380157f6 ("toolchain: enable
libquadmath for PowerPC with VSX"), the definition of
BR2_TOOLCHAIN_HAS_LIBQUADMATH has been extended to also be true when
BR2_POWERPC_CPU_HAS_VSX.
However, practical experiments show that when 64-bit VSX-capable cores
are used in 32-bit mode, libquadmath is not built by GCC, causing
build failures:
cp: cannot stat '/home/autobuild/autobuild/instance-3/output-1/host/powerpc-buildroot-linux-musl/lib*/libquadmath*': No such file or directory
We did an extensive testing, building the 27 combinations of:
- GCC versions: 14, 15, 16
- C library: glibc, uclibc, musl
- PowerPC 32-bit, PowerPC 64-bit, PowerPC 64-bit little endian
This testing provides the following results:
| gcc14 | powerpc64 | glibc | OK |
| gcc14 | powerpc64le | glibc | OK |
| gcc14 | powerpc64le | musl | OK |
| gcc14 | powerpc64le | uclibc | SKIPPED |
| gcc14 | powerpc64 | musl | OK |
| gcc14 | powerpc64 | uclibc | SKIPPED |
| gcc14 | powerpc | glibc | FAILED |
| gcc14 | powerpc | musl | FAILED |
| gcc14 | powerpc | uclibc | FAILED |
| gcc15 | powerpc64 | glibc | OK |
| gcc15 | powerpc64le | glibc | OK |
| gcc15 | powerpc64le | musl | OK |
| gcc15 | powerpc64le | uclibc | SKIPPED |
| gcc15 | powerpc64 | musl | OK |
| gcc15 | powerpc64 | uclibc | SKIPPED |
| gcc15 | powerpc | glibc | FAILED |
| gcc15 | powerpc | musl | FAILED |
| gcc15 | powerpc | uclibc | FAILED |
| gcc16 | powerpc64 | glibc | OK |
| gcc16 | powerpc64le | glibc | OK |
| gcc16 | powerpc64le | musl | OK |
| gcc16 | powerpc64le | uclibc | SKIPPED |
| gcc16 | powerpc64 | musl | OK |
| gcc16 | powerpc64 | uclibc | SKIPPED |
| gcc16 | powerpc | glibc | FAILED |
| gcc16 | powerpc | musl | FAILED |
| gcc16 | powerpc | uclibc | FAILED |
The "SKIPPED" are when the configuration is not possible: uClibc
doesn't support powerpc64 or powerpc64le.
Then, as we can see, the build fails for all "powerpc"
configuration. Our conclusion is therefore that libquadmath is not
supported on PowerPC 32-bit. While we were not able to find direct
evidence in the gcc code base, this practical experiment shows that is
simply doesn't work on PowerPC 32-bit.
So, we take the logical action of adjusting
BR2_TOOLCHAIN_HAS_LIBQUADMATH so that it is true only on
powerpc64/powerpc64le.
Fixes:
https://autobuild.buildroot.org/results/46d435c9f5086a8695f4f6cd4026bb0d194de13c/
Cc: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 06426297c4)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
Fixes a build error caught by the Gitlab pipelines:
/builds/bkuhls/buildroot/br-test-pkg/bootlin-aarch64-glibc-old/build/taglib-2.3/taglib/mpeg/mpegfile.cpp:113:10:
error: expected primary-expression before ‘const’
if(const Header header(&file, headerOffset + i, true); header.isValid()) {
which was introduced by code format changes in upstream commit
dfe2aa5253
which was first released with taglib 2.0, added to buildroot with commit
9cd3464afa.
This "init-statement" C++17 language feature was described in proposal
P0305R1, and according to
https://en.cppreference.com/cpp/compiler_support/17, this feature was
only supported in gcc starting from gcc 7.x.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit d16e4939ca)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
host-libxml-parser-perl compiles XS modules against the system perl
headers, which #include <crypt.h>. On build hosts without libcrypt-dev
installed, the build fails:
.../CORE/reentr.h:126:16: fatal error: crypt.h: No such file or directory
Declaring host-libxcrypt ensures crypt.h is present in the per-package
host sysroot before the build.
This can for example be reproduced on a minimal Debian Forky system,
where libc6-dev no longer pulls libxcrypt-dev.
Signed-off-by: Adam Ford <aford173@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 7e036c739f)
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
The commit [1] decoupled GCC from gcc-bare-metal but by copying
the patches the patch [2] wasn't ignored in gcc-bare-metal.
[1] 367133ac70 package/gcc-bare-metal: decouple from gcc package
[2] 0001-disable-split-stack-for-non-thread-builds.patch lib_patch.Upstream
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
OpenSSL 3.5.8 is a security patch release. The most severe CVE fixed
in this release is Moderate.
This release incorporates the following bug fixes and mitigations:
Fixed QUIC server being able to trigger double free when processing
INITIAL packet.
(CVE-2026-18798)
Fixed heap buffer overflow in CMS key unwrapping.
(CVE-2026-63072)
Fixed invalid pointer dereference in CMP server via crafted protectionAlg.
(CVE-2026-63076)
Fixed unbounded memory growth in QUIC server incoming channel queue.
(CVE-2026-14456)
Fixed RPK server signature algorithm selection being able to dereference
a missing certificate.
(CVE-2026-14457)
Fixed excessive memory use buffering DTLS records for a future epoch.
(CVE-2026-54874)
Fixed untrusted Sender DN being used as a format string in CMP response
validation.
(CVE-2026-63073)
Fixed CMP indefinite cache growth of extraCerts.
(CVE-2026-63074)
Fixed QUIC ACK-only packet retention being able to cause memory exhaustion.
(CVE-2026-63075)
Fixed possibility of AEAD forgeries with empty ciphertext when using
EVP_Cipher().
(CVE-2026-75803)
Fixed checking of authentication tags for empty ciphertexts for AEAD ciphers
in CCM cipher mode.
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
0002-Add-configure-time-check-for-lanl.patch (taken from mosquitto
PR #3358) probes for -lanl with a $(shell ...) that pipes a small C
program into $(CC):
NEED_LIBANL := $(shell printf '#include <stdlib.h>\n#include <netdb.h>\n...
The test program contains literal '#' characters. GNU make strips
comments before expanding functions, and only since 4.3 does it keep
'#' inside a function call argument intact. With older make the line
is truncated at the first '#', so the shell call never sees its
closing parenthesis:
config.mk:322: *** unterminated call to function 'shell': missing ')'. Stop.
That is why the package builds fine locally (make 4.3+) but fails on
the build server, which ships an older GNU make.
Switch to the variant that was actually merged upstream (commit
dfde35a, "Add compile time checks for whether -lanl is required").
It links an empty main() against -lanl instead of compiling a program
that calls getaddrinfo_a(), so the test source needs no #include and
contains no '#' at all, which works on every make version.
Upstream: dfde35aa0f
Signed-off-by: Thomas Devoogdt <thomas.devoogdt@barco.com>
[Titouan: update "Upstream:" tag in the patch]
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Commit 713d63b "package/webkitgtk: add option to enable MiniBrowser",
added support to select BR2_PACKAGE_WEBKITGTK_MINIBROWSER, but forgot
to drop the default -DENABLE_MINIBROWSER=ON entry.
Signed-off-by: Thomas Devoogdt <thomas@devoogdt.com>
Acked-By: Adrian Perez de Castro <aperez@igalia.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 9f292bb7a1)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
In both start() and stop(), ret is only assigned on failure. When
hypervkvpd starts or stops successfully, return "$ret" expands to an
empty string and causes:
/etc/init.d/S10hyperv: return: line 31: Illegal number:
Those double quotes were added in Buildroot commit [1], to fix a
new ShellCheck warning at that time. This was not a complete fix.
Only removing the double quote would reintroduce the ShellCheck
warning. This would also reintroduce a check-package error.
Since a bare return is equivalent to a "return 0", this commit
also initializes with ret=0. Doing so will tell ShellCheck "ret" is
an integer. Therefore, the ShellCheck warning will no longer be
reported.
This commit fixes the invalid return value by removing the double
quotes and initialzing "ret=0".
[1] c4173d8b08
Signed-off-by: Benjamin DeCamp <benjamin8532@protonmail.com>
[Julien:
- add "ret=0" initialization in script to fix check-package error
- add extra info in the commit log
]
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 667335cd18)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Backport the fix for CVE-2026-66035.
The ETM decrypt path does not validate the received packet length before
calculating the decrypt buffer size. A malformed packet can therefore
lead to a heap overflow.
Use Debian's libssh2 1.11.1 backport of the upstream fix.
Signed-off-by: Stefan Müller <stefan.mueller@rey-technology.com>
[Julien: add links to Debian patches]
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 03757abfce)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Backport the fix for CVE-2026-66034.
The publickey subsystem does not sufficiently validate the length of a
server-controlled comment field. A malformed response can therefore
cause an out-of-bounds read.
Use Debian's libssh2 1.11.1 backport of the upstream fix.
Signed-off-by: Stefan Müller <stefan.mueller@rey-technology.com>
[Julien: add links to Debian patches]
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 58581deeca)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Backport the fix for CVE-2026-66033.
The OpenSSL AES-GCM cipher path lacks runtime bounds checks around the
input block size. A malformed packet can therefore lead to an
out-of-bounds read or write.
Use Debian's libssh2 1.11.1 backport of the upstream fix.
Signed-off-by: Stefan Müller <stefan.mueller@rey-technology.com>
[Julien: add links to Debian patches]
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 6755a00cd2)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Backport the fix for CVE-2026-66032.
A SFTP error path can leave a dangling pointer after freeing the
response buffer, which may result in a double free on subsequent error
handling.
Use Debian's libssh2 1.11.1 backport of the upstream fix.
Signed-off-by: Stefan Müller <stefan.mueller@rey-technology.com>
[Julien: add links to Debian patches]
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 05c13e87e9)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Backport the SFTP symlink bounds checking fix for CVE-2025-15661.
The initial fix requires the LIBSSH2_UNCONST compatibility backport on
libssh2 1.11.1. Also include the upstream follow-up fixing
SSH_FXP_STATUS handling introduced by the initial security fix.
The patches are based on the upstream fixes and Debian's libssh2 1.11.1
backports.
Signed-off-by: Stefan Müller <stefan.mueller@rey-technology.com>
[Julien: add links to Debian patches]
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 546fd31c70)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Backport the upstream fix for a heap buffer overflow in
convert_fname() when growing the iconv output buffer.
Backport to: 2025.02.x
Signed-off-by: Stefan Müller <stemu86@gmx.ch>
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit e991fa0716)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Backport the upstream fix for integer overflows while parsing
Content-Range headers, together with the follow-up fix using
strtoll() for wgint values.
Backport to: 2025.02.x
Signed-off-by: Stefan Müller <stemu86@gmx.ch>
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 89485adb29)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Backport the upstream fix for a buffer underflow in
clean_metalink_string(), together with the two required follow-up
fixes for the inverted whitespace check and missing ctype.h include.
Backport to: 2025.02.x
Signed-off-by: Stefan Müller <stemu86@gmx.ch>
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 937e33237e)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
The Arc specific gdb version was removed by commit [1]
but we still have the TestGdbArc that was testing this
version of gdb.
We can now safely remove TestGdbArc.
[1] 0b3d526226
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 2f6b34f851)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
When a package defines $(PKG)_FLAT_STACKSIZE, ELF2FLT_FLAGS contains
-Wl,-elf2flt="-r -s<stack-size>". The embedded quotes are needed to
keep both elf2flt options in single linker argument.
However, many package Makefiles wrap $(TARGET_CFLAGS) in double quotes,
for example:
CFLAGS="$(TARGET_CFLAGS)"
After expansion, the embedded quote terminates the outer CFLAGS quote.
As a result, the shell interprets "-s<stack-size> ..." as a command
instead of passing it to the compiler.
Pass -r and -s<stack-size> in separate -Wl arguments instead. This
avoids embedded quotes; GCC forwards both -elf2flt options to
ld-elf2flt, which collects them before invoking elf2flt.
This got broken by commit
04d7ea4720 ("package: Makefile.in: fix
elf2flt invocation options"), which by adding -r as an elf2flt
argument, did not correctly handle -s$($(PKG)_FLAT_STACKSIZE).
Signed-off-by: Fengwei Tan <tfx2001@outlook.com>
[Thomas: improve commit message]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit e913afbeb1)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
This fixes the following vulnerability:
- CVE-2026-7867:
A flaw was found in udisks2. A local attacker with an active console
session can exploit insufficient authorization checking on the 'as-
user' option in the org.freedesktop.UDisks2.Filesystem.Mount() D-Bus
method. This allows the attacker to spoof the 'as-user' parameter,
mounting filesystems on behalf of arbitrary users, including
privileged accounts. This can lead to local privilege escalation
through mount point injection and manipulation of the mount namespace
visible to privileged users.
https://www.cve.org/CVERecord?id=CVE-2026-7867
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>7
(alternative to commit 93049b2559)
[Titouan:
- Add upstream, SoB, CVE tags in patches
- Add UDISKS_IGNORE_CVES entry
- Add CVE description in the commit message
]
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
wine.mk passes --with-wayland whenever BR2_PACKAGE_WAYLAND is enabled,
but nothing guarantees the rest of what wine's Wayland test needs is in
the configuration. That test is:
WINE_NOTICE_WITH(wayland, [test -z "$WAYLAND_CLIENT_LIBS" \
-o -z "$WAYLAND_SCANNER" -o -z "$XKBCOMMON_LIBS" \
-o -z "$XKBREGISTRY_LIBS" -o "$ac_cv_header_linux_input_h" = "no"], ...)
and because --with-wayland is passed explicitly, WINE_NOTICE_WITH turns
into AC_MSG_ERROR rather than a notice.
So wine needs libxkbcommon, and it needs the libxkbregistry part of it,
which is only built when libxml2 is available. Select both when Wayland
support is enabled, and add libxkbcommon to the build dependencies.
Note that libxml2 is not a direct dependency of wine, it only has to be
in the configuration so that libxkbcommon builds libxkbregistry; the
build ordering is handled by libxkbcommon's own dependency on libxml2.
Signed-off-by: Alsey Coleman Miller <alseycmiller@gmail.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit a5e7f7af9f)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
libxkbregistry is the keyboard layout catalogue half of the library. It
parses the XML layout registry and so needs libxml2, which is presumably
why it was disabled unconditionally rather than wired to a dependency.
wine needs it. Its configure.ac requires XKBREGISTRY_LIBS alongside
wayland-client, wayland-scanner, xkbcommon and linux/input.h before it
will build the Wayland driver, and wine.mk passes --with-wayland for any
build with BR2_PACKAGE_WAYLAND - which turns that notice into a hard
error:
checking for wayland-client.h... yes
checking for wl_display_connect in -lwayland-client... yes
checking for wayland-scanner... .../host/bin/wayland-scanner
checking for xkb_context_new in -lxkbcommon... yes
checking for wayland-egl.h... yes
checking for wl_egl_window_create in -lwayland-egl... yes
configure: error: Wayland development files not found, the Wayland
driver won't be supported.
This is an error since --with-wayland was requested.
Every other term of that test passes; only XKBREGISTRY_LIBS is empty, so
wine and wayland together could not be built on any architecture.
Gated on BR2_PACKAGE_LIBXML2 rather than turned on outright, because
meson.build takes dependency('libxml-2.0') unconditionally once
enable-xkbregistry is set, so a target without libxml2 would fail to
configure.
Regarding since when this is broken, three pieces had to come together:
- libxkbcommon has passed -Denable-xkbregistry=false since commit
1791bc30a5 ("package/libxkbcommon: bump version to 1.0.1", Sep 2020),
i.e. Buildroot 2020.11. libxkbregistry has therefore never been built
in Buildroot.
- wine's configure gained the XKBREGISTRY_LIBS term in its Wayland
test in wine 9.0, with upstream commit d64ea8e4a6c9
("winewayland.drv: Enumerate Xkb layouts and create matching HKL.",
Nov 2023).
- wine.mk started passing --with-wayland in commit 7cb49e7712
("package/wine: bump to version 9.19", Oct 2024), which is what turns
the missing XKBREGISTRY_LIBS from a notice into a hard error.
The breakage therefore dates from Buildroot 2024.11, and every branch
since is affected, including the LTS one: 2025.02.x carries wine 10.0,
whose configure has the XKBREGISTRY_LIBS check, together with
libxkbcommon 1.9.2 built with -Denable-xkbregistry=false, and its wine.mk
passes --with-wayland. 2025.05.x and 2025.08.x are in the same state.
A backport to 2025.02.x is thus needed.
Signed-off-by: Alsey Coleman Miller <alseycmiller@gmail.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 4349b22b91)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
Fixes the following security issues:
- x/mod/sumdb/tlog: fix transparency log tile verification bypass
A malicious GOPROXY was previously capable of forging up to two sumdb
tiles that allow for a requested module to bypass the GOSUMDB check and
persist attacker-controlled module content to a local Go module cache.
This attack allows for a malicious GOPROXY to serve malicious module
content that cannot be detected by evaluating the transparency log.
All tiles are now correctly verified against their parents.
In order to determine if you have been affected:
rm -r go.sum go.work.sum vendor/ && go mod tidy
Thanks to Filippo Valsorda (Geomys) for reporting this issue.
This is CVE-2026-56865 and Go issue https://go.dev/issue/80744.
- x/mod/sumdb: ignore unrelated, unauthenticated hashes in Lookup
A malicious GOSUMDB was capable of serving arbitrary module content not
contained within the transparency log.
This attack allows for a coordinating GOPROXY and GOSUMDB to serve a
client malicious module content that cannot be detected by evaluating
the transparency log.
In order to determine if you have been affected:
rm -r go.sum go.work.sum vendor/ && go mod tidy
Thanks to mundur for reporting this issue.
This is CVE-2026-56864 and Go issue https://go.dev/issue/80745.
- encoding/xml: add recursion depth guard during decode
Previously, DecodeElement would reset the depth counter causing it to
never fire; this could lead to stack exhaustion.
This is CVE-2026-56859 and Go issue https://go.dev/issue/80481.
- net/http: apply ReadHeaderTimeout when doing unencrypted HTTP/2 check
When a server is configured to support unencrypted HTTP/2, it reads a few
bytes from each new connection to see if they contain the HTTP/2 client
preface. Previously, this was being done with no timeout applied.
ReadHeaderTimeout is now applied for this.
This is CVE-2026-56853 and Go issue https://go.dev/issue/80205.
- net/url: avoid quadratic complexity in resolvePath
Previously, resolving relative paths containing parent directory (..)
segments performed string conversions and buffer rewrites on each step,
resulting in quadratic time complexity and high memory allocation
overhead.
Now, path resolution operates on a byte buffer using index-based
backtracking for .. segments, eliminating the quadratic time complexity
and significantly reducing memory allocations.
This is CVE-2026-56860 and Go issue https://go.dev/issue/80494.
- golang.org/x/net/dns/dnsmessage: panic when parsing invalid SVCB record
Parsing an invalid SVCB or HTTPS RR can panic when the size of a
parameter value overflows the message buffer.
Thanks to Mundur (https://github.com/M0nd0R) for reporting this issue.
This is CVE-2026-46600 and Go issue https://go.dev/issue/79795.
- crypto/tls: limit handshake messages we are willing to accept post-handshake
Previously, we always counted handshake messages, such as KeyUpdate, as
state-advancing, regardless of whether a handshake has been completed or
not. As a result, a malicious client can keep sending KeyUpdate messages
to force the server to keep performing key derivation operations
indefinitely.
Thanks to Qi Deng of Aurascape.ai for reporting this issue.
This is CVE-2026-56862 and Go issue https://go.dev/issue/80528.
- html/template: fix Javascript regexp context tracking
Previously, pathological inputs could close an unescaped / early,
allowing for attack-controlled data to inject arbitrary content,
potentially leading to XSS.
Thanks to Ali Sherif for reporting this issue.
This is CVE-2026-56858 and Go issue https://go.dev/issue/80435.
- x/net/idna: failure to reject ASCII-only Punycode-encoded labels
The ToASCII and ToUnicode functions incorrectly accepted Punycode-encoded
labels that decode to an ASCII-only label. For example,
ToUnicode("xn--example-.com") incorrectly returned the name "example.com"
rather than an error.
The idna package implements the processing algorithm from UTS 46. Older
versions of UTS 46 included a specification bug which permitted multiple
ASCII labels to decode to the same Unicode label. UTS 46 revision 33
fixed the specification bug. The idna package now implements the updated
specification.
This behavior can lead to privilege escalation in programs using the idna
package. For example, a program which performs privilege checks on the
ASCII hostname may reject "example.com" but permit "xn--example-.com".
If that program subsequently converts the ASCII hostname to Unicode, it
will inadvertently permits access to the Unicode name "example.com".
Thanks to KC1zs4 (https://github.com/KC1zs4) for reporting this issue.
This is CVE-2026-39821 and Go issue https://go.dev/issue/78760.
- encoding/asn1: enforce maximum recursion depth
Enforce a recursion limit in Unmarshal to prevent stack exhaustion when
parsing deeply-nested, recursive structures.
Thanks to Marwan Atia (marwansamir688@gmail.com) for reporting this issue.
This is CVE-2026-33818 and Go issue https://go.dev/issue/80405.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Julien Olivain <ju.o@free.fr>
(cherry picked from commit 9e9110bf23)
Signed-off-by: Titouan Christophe <titouan.christophe@mind.be>
