Currently the MissingCVEPatch check will complain if another
identifiers than a CVE-YYYY-XXXX is passed after the 'CVE: '
trailer in the patch header introduced in [1].
If a package 'foo' defines the following:
FOO_IGNORE_CVES += GHSA-abcd-1234-efgh
And correctly references it in a patch header:
CVE: GHSA-abcd-1234-efgh
check-package will report the following:
package/foo/foo.mk: patch file '0001-some-patch.patch' is missing 'CVE:' tag
While we could argue that GHSA identifier should not be present in the
'IGNORE_CVES' variable, it's still valuable for users to have a way to
correctly ignore and link a patch for their vulnerability report.
From our point of view it allows us to track GHSA vulnerabilities that
could later be assigned a CVE so we can update the IGNORE_CVES variable.
This commit relaxes the constraint on the 'CVE' trailer to just check
for the presence of any identifier. We still only allow alphanumerical
characters and dashes, to be sure that any tooling that uses the
identifiers won't break.
[1] 1167d0ff3d docs/manual: mention CVE trailer
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
[Arnout:
- keep the regex but allow any alphanumerical string;
- add a test that FOO:123 doesn't match]
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
The `self.pending_patch` is set but unused.
Fixes: b00ac4e346 utils/checkpackagelib: add new check MissingCVEPatch
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
szip is not maintained upstream anymore: 2.1.1 is the last release,
there is no public repository, and the source is only available from
the HDF Group archive. Its license is not a standard open source one,
and the HDF Group itself recommends libaec as a replacement.
The package also fails to build with GCC >= 15 on non-glibc
toolchains, as only linux-gnu hosts get -ansi from szip's configure
and the code relies on pre-C23 empty parameter lists:
rice.c:1323:15: error: too many arguments to function 'winner_ref_function'; expected 0, have 2
libaec provides libsz, a drop-in replacement with the same API and
soname, so select it through the legacy option.
Fixes:
https://autobuild.buildroot.org/results/5e400d0e13480c58e8b1b0178bf7e5537dd31751
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
libaec provides fast lossless compression of 1 up to 32 bit wide
signed or unsigned integers. It implements the Golomb-Rice compression
method under the BSD license and includes libsz, a drop-in replacement
for the SZIP library with the same API (szlib.h) and soname
(libsz.so.2).
https://github.com/Deutsches-Klimarechenzentrum/libaec
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
gclient.py tries to not only read from $HOME/.config, but actually
create some directories/files in there. As $HOME can be inaccessible,
let's override XDG_CONFIG_HOME to tell the depot-tools code to write
its stuff into $(@D)/dl-tmp instead (which we already use as a scratch
directory in this package).
Fixes:
>>> flutter-engine 3.29.2 Downloading
/home/thomas/buildroot/br/output-all/host/share/depot_tools/gclient.py:3367: SyntaxWarning: invalid escape sequence '\s'
dot_git_pattern = re.compile('^(\s*)url(\s*)=.*\.git$')
/home/thomas/buildroot/br/output-all/host/share/depot_tools/gclient_scm.py:436: SyntaxWarning: invalid escape sequence '\.'
tokens = re.search('//(.+).googlesource.com/(.+?)(?:\.git)?$',
Traceback (most recent call last):
File "/home/thomas/buildroot/br/output-all/host/lib/python3.13/pathlib/_local.py", line 722, in mkdir
os.mkdir(self, mode)
~~~~~~~~^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: '/home/br-user/.config/depot_tools'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/thomas/buildroot/br/output-all/host/share/depot_tools/gclient.py", line 108, in <module>
import gclient_scm
File "/home/thomas/buildroot/br/output-all/host/share/depot_tools/gclient_scm.py", line 23, in <module>
import gerrit_util
File "/home/thomas/buildroot/br/output-all/host/share/depot_tools/gerrit_util.py", line 42, in <module>
import metrics
File "/home/thomas/buildroot/br/output-all/host/share/depot_tools/metrics.py", line 23, in <module>
CONFIG_FILE = utils.depot_tools_config_path('metrics.cfg')
File "/home/thomas/buildroot/br/output-all/host/share/depot_tools/utils.py", line 48, in depot_tools_config_path
pathlib.Path(config_dir).mkdir(parents=True, exist_ok=True)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/thomas/buildroot/br/output-all/host/lib/python3.13/pathlib/_local.py", line 726, in mkdir
self.parent.mkdir(parents=True, exist_ok=True)
~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/thomas/buildroot/br/output-all/host/lib/python3.13/pathlib/_local.py", line 722, in mkdir
os.mkdir(self, mode)
~~~~~~~~^^^^^^^^^^^^
PermissionError: [Errno 13] Permission denied: '/home/br-user/.config'
make[1]: *** [package/pkg-generic.mk:181: /home/thomas/buildroot/br/output-all/build/flutter-engine-3.29.2/.stamp_downloaded] Error 1
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The $HOME folder of the br-user inside the container is currently
read-write for the br-user. However, practically speaking, it is only
read-write when the UID of the user running the container from the
host machine is 1000, equal to the UID of the br-user inside the
container. For any other UID on the host machine, the $HOME folder is
in fact already read-only.
Because we do not expect Buildroot to write into $HOME, and in order
to have a consistent behavior regardless of the UID of the user on the
host machine, we change our Docker image to make the $HOME folder of
the br-user entirely non-writable.
Suggested-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This commit introduces a menu comment to users to indicate whether the
`cfdisk` can be built, even when the basic set binaries option is
enabled.
For users wishing to support `fdisks` utilities (cfdisk, fdisk and
sfdisk) in Buildroot, these utilities are enabled through a
`BR2_PACKAGE_UTIL_LINUX_BINARIES` option. This option configures
`all-programs`, which will default enable `fdisks` (along with other
program options that do not have an explicit configuration). While a
user can enable "all programs", there is no guarantee that the
`cfdisk` utility will be built if neither ncurses or slang is
enabled. Even if an explicit Buildroot menu option is added for
`cfdisk` (where we can force select ncurses, if slang is not yet
enabled), there is no specific configuration option in util-linux to
toggle the enablement of `cfdisk`.
Signed-off-by: James Knight <james.d.knight@live.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Both cfdisk and irqtop can either use ncurses or slang. We currently
only support using ncurses, this commit extends that to be able to use
slang as an alternative.
Signed-off-by: James Knight <james.d.knight@live.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
When using /bin/mount to mount filesystems that are not supported
directly by the kernel, it delegates the mount operatin to helper
programs, e.g. mount.fuse.sshfs. Normally /bin/mount searches
these directories in order for a helper: /sbin, /sbin/fs.d, /sbin/fs.
In reality, Buildroot installs e.g. mount.fuse.sshfs into /usr/sbin,
and as a result /bin/mount fails to mount sshfs entries defined in
/etc/fstab.
The patch changes the default search path at configuration.
Signed-off-by: Stephan Henningsen <stephan@asklandd.dk>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Buildroot commit 31af509b4f bumped the
package from 2.40.2 to 2.41.1. This bump includes upstream commit
a3f1255f18
which was added to version 2.41 and includes the usage of
pthread_atfork() causing build errors on non-MMU platforms:
libuuid/src/gen_uuid.c: In function 'uuid_generate_time_generic':
libuuid/src/gen_uuid.c:629:17: error: implicit declaration of function
'pthread_atfork'; did you mean 'pthread_join'?
[-Wimplicit-function-declaration]
629 | pthread_atfork(NULL, NULL, reset_uuidd_cache);
The oldest build error of this kind can be found on the 2025.11.x
branch with version 2.41.1:
https://autobuild.buildroot.net/results/87c/87c74435be14956fd45f730269eac102e70d56da/
The issue occurs because uClibc-ng is not very consistent with its
pthread_atfork() definition: it doesn't provide a prototype for it on
!MMU platforms (causing the build failure), but it does provide a stub
implementation (causing the configure script to believe pthread_atfork
is available). This issue was reported to upstream uClibc-ng at
https://mailman.openadk.org/mailman3/hyperkitty/list/devel@uclibc-ng.org/thread/WEA7IQC3YAWTSFDHCLYRQJZTC4IUIY7J/.
A backport of this patch to the LTS branches should be considered, but
probably not 2025.02.x as it uses version 2.40.x.
Fixes:
https://autobuild.buildroot.org/results/a21e81e64eccc866b575a4f3b1376c50f2ffd837/
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Add test for CycloneDX vulnerability generation.
If multiple packages ignore the same vulnerability but provide different
analyses, the generator must produce multiple vulnerability entries with
the same vulnerability ID but different analysis information.
This situation can occur when generating a CycloneDX SBOM from the
output of 'make show-info-all'. The resulting SBOM includes packages
that exist in multiple versions within Buildroot (gnupg for instance).
In such a case, one package version may ignore a vulnerability because
it is not affected, while another package version may address it through
a patch.
Therefore, multiple vulnerability entries must be generated, each with
its own analysis and reference information.
This adds a test case for this scenario.
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
[Arnout: give next() a None default to return None instead of raise
StopIteration]
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
If a vulnerability is present in multiple components and only one of
them has a patch that addresses that vulnerability, all components will
currently have that vulnerability marked as `resolved_with_pedigree`.
Right now, this issue does not affect any packages, but in the future it
might affect packages that provide multiple version options (e.g.
gnupg and gnupg2).
There is a small chance this happens in a real-world use case, but it
may occur in the context of maintenance when running
`make show-info-all`.
This commit changes how vulnerabilities fixed by a patch are stored.
Instead of placing all vulnerabilities into a single set, it now keeps an
index of the component for which the patch has been applied.
When generating the vulnerability list, the component reference is
checked instead of only the vulnerability ID. If a vulnerability ID has
multiple states for different references, multiple vulnerability
entries are created with distinct analyses.
Consider the hypothetical case where gnupg ignores a vulnerability
because it was introduced in a later version, while gnupg2 has patched
that vulnerability. This would result in the following JSON:
```json
[
{
"id": "CVE-1234-1234",
"analysis": {
"state": "in_triage",
"detail": "The CVE 'CVE-1234-1234' has been marked as ignored by Buildroot"
},
"affects": [
{"ref": "gnupg"}
]
},
{
"id": "CVE-1234-1234",
"analysis": {
"state": "resolved_with_pedigree",
"detail": "The CVE 'CVE-1234-1234' has been marked as ignored by Buildroot"
},
"affects": [
{"ref": "gnupg2"}
]
}
]
```
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
[Arnout:
- update the VULN_WITH_PEDIGREE comment to match the code
- use set() as default instead of {}]
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
Move the code that retrieves the SPDX licenses into a function instead of
doing it top-level.
This allows to pass the CycloneDX version as a function argument instead
of defining it as a top-level constant.
The next commit introduces support for different CycloneDX spec versions.
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
Commit [1] introduced a code block that references
'urllib.error.URLError'. There is no explicit import of 'urllib.error'.
While this is not an issue in practice, ruff complains that the import
is not explicit.
[1] 17c9deb115 utils/generate-cyclonedx: hint at missing Buildroot host package on a specific error
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
The commit [1] fixed a formatting bug with the SCP style git sites
declaration.
Add this style of link to the test case to keep a track of it.
[1] ebcfdb8b0a utils/generate-cyclonedx: fixup scp-style git sites
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
The Linux kernel exposes a CEC API since v4.10. Support only grows the
library with ~3%, so unconditionally enable support for it when building
with new enough kernel headers:
-rwxr-xr-x 1 peko peko 594K Oct 10 18:53 usr/lib/libcec.so.8.1.7.without
-rwxr-xr-x 1 peko peko 614K Oct 10 18:53 usr/lib/libcec.so.8.1.7.with
Reported-by: Stefan Ott <stefan@ott.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
fexecve was added to uClibc in 2019 in commit
858ffad217076227089cc17eb832db0bd1497792.
Also containerd and docker can be used with uClibc.
Runtime tested on a Raspberry PI4 (64 Bit).
System compiled with Bootlin Toolchain.
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
iproute2's configure script enables HAVE_RPC as soon as libtirpc is
found, and misc/ss.c then calls getrpcbynumber() from its RPC service
resolver. Since the bump of libtirpc to version 1.3.7 in Buildroot
commit 3f3d6e43de, which includes upstream
commit:
https://git.linux-nfs.org/?p=steved/libtirpc.git;a=commit;h=7cea8ad66aecc21e6caae330b5d31075af399193
that symbol is only exported when libtirpc is configured with
--enable-rpcdb. On non-glibc toolchains there is no libc fallback
either, so linking ss fails:
ld: ss.o: in function `init_service_resolver':
ss.c:(.text+0x2558): undefined reference to `getrpcbynumber'
Select BR2_PACKAGE_LIBTIRPC_RPCDB when libtirpc is enabled, like lsof
and other users of these functions already do.
Fixes:
https://autobuild.buildroot.org/results/390b7ed00da20c2349fe069ade4a65404834a484/
Fixes: 3f3d6e43de ("package/libtirpc: bump to version 1.3.7")
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Assisted-by: Claude:claude-opus-5
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
iproute2's configure script probes for libtirpc through pkg-config and,
when found, enables HAVE_RPC and links ss against it. libtirpc was never
listed in IPROUTE2_DEPENDENCIES though, so whether the feature ends up
enabled depended on whether libtirpc happened to be built before
iproute2.
Add the missing optional dependency to make the build reproducible.
iproute2's configure script checks for libtirpc availability since
v5.18.0 of iproute2, which Buildroot started using in commit
10990c576b ("package/iproute2: bump version to 5.18.0")
Fixes: 10990c576b ("package/iproute2: bump version to 5.18.0")
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Assisted-by: Claude:claude-opus-5
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Buildroot commit 8dad17ea06 bumped the
package from 0.27.0 to 0.37.2. This bump includes upstream commit
464867a9ce
which added WebSocket support to version 0.33.0 and causes a build error
detected by the Gitlab pipelines for bootlin-sparc-uclibc:
/builds/bkuhls/buildroot/br-test-pkg/bootlin-sparc-uclibc/host/opt/ext-toolchain/bin/../lib/gcc/sparc-buildroot-linux-uclibc/14.3.0/../../../../sparc-buildroot-linux-uclibc/bin/ld:
/tmp/cciyXsAR.ltrans6.ltrans.o: in function
`httplib::ws::WebSocket::close(httplib::ws::CloseStatus,
std::__cxx11::basic_string<char, std::char_traits<char>,
std::allocator<char> > const&)':
<artificial>:(.text+0x6408): undefined reference to `__atomic_exchange_1'
To fix the problem we add -latomic to LDFLAGS when needed.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Users will ultimately need the BR2_LINUX_KERNEL_CUSTOM_DTS_DIR config to
import custom dts files for custom versal2 hardware. Improve the post-image.sh
to support either the default device tree which comes from u-boot or a
custom device tree for custom hardware.
To keep the example post-image script as simple as possible, it assumes that
the first DTB in the LINUX_DTBS list is the one for creating the system.dtb
filename that u-boot will use for booting. Users are free to implement their
own post-image scripts for use cases with multiple DTBs or overlays.
Signed-off-by: Neal Frager <neal.frager@amd.com>
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
By switching to the LINUX_DTBS variable, the post-image.sh script can now
work regardless of whether a user is using BR2_LINUX_KERNEL_INTREE_DTS_NAME
or the BR2_LINUX_KERNEL_CUSTOM_DTS_DIR for custom hardware. This simplifies
the transition users will make when migrating from a versal evaluation board
to a versal custom platform.
To keep the example post-image script as simple as possible, it assumes that
the first DTB in the LINUX_DTBS list is the one for creating the system.dtb
filename that u-boot will use for booting. Users are free to implement their
own post-image scripts for use cases with multiple DTBs.
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>
By switching to the LINUX_DTBS variable, the post-image.sh script can now
work regardless of whether a user is using BR2_LINUX_KERNEL_INTREE_DTS_NAME
or the BR2_LINUX_KERNEL_CUSTOM_DTS_DIR for custom hardware. This simplifies
the transition users will make when migrating from a zynqmp evaluation board
to a zynqmp custom platform.
To keep the example post-image script as simple as possible, it assumes that
the first DTB in the LINUX_DTBS list is the one for creating the system.dtb
filename that u-boot will use for booting. Users are free to implement their
own post-image scripts for use cases with multiple DTBs.
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>
Buildroot commit [1] introduced the check-package error:
boot/edk2/Config.in:19: line contains trailing whitespace
This commit fixes the issue by removing those extra spaces.
[1] ea0f27d64f
Signed-off-by: Julien Olivain <ju.o@free.fr>
When running iperf3 with Linux 6.18.51 on VisionFive 2, one CPU core
stays at 100% load during the test, while network throughput is limited
to approximately 793 Mbits/sec:
[ ID] Interval Transfer Bitrate
[ 5] 0.00-10.01 sec 945 MBytes 793 Mbits/sec receiver
The regression was introduced in Linux 6.14 by commit df542f669307
("net: stmmac: Switch to zero-copy in non-XDP RX path"). It leads to
misaligned access exceptions when parsing packet headers. During the
iperf3 test, perf stat reported approximately 9.6 million alignment
faults over 5 seconds:
$ perf stat -a -e alignment-faults -- sleep 5
Performance counter stats for 'system wide':
9620073 alignment-faults
5.029337000 seconds time elapsed
The kernel must emulate these unaligned accesses, resulting in
significant overhead.
The issue was fixed in Linux 6.18.52 [1]. Bump Linux to 6.18.55, the
latest stable release at the time of writing, to include the fix.
[1] https://lore.kernel.org/all/2026091704-CVE-2026-90059-36e6@gregkh
Fixes: 98aa14a186 ("configs/visionfive2: bump Linux to 6.18.51, U-Boot to 2026.07, and OpenSBI to 1.8.1")
Assisted-by: LLM
Signed-off-by: Fengwei Tan <tfx2001@outlook.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Codeberg has changed the way it generate hashes for at least some of
the generated tarballs (see [1]). Since we cannot update the hash
without changing the file name, we switch to Git download in order to
have a different file name.
[1]: https://codeberg.org/Codeberg/Community/issues/2861
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Since the bump to version 2.13.0 in Buildroot commit [1],
atop fails to build with musl:
drawbar.c:2336:33: error: implicit declaration of function 'waddwstr'; did you mean 'waddstr'? [-Wimplicit-function-declaration]
(Before 2.13.0, the same configurations already failed earlier on
uClibc because _POSIX_C_SOURCE was defined without a value. Upstream
fixed that in 2.13.0 by giving it an explicit value, but the same
pattern remains for _XOPEN_SOURCE.)
ncurses only declares its wide-character API, including waddwstr(),
when NCURSES_WIDECHAR is set. If the application does not set it,
curses.h derives it from the feature test macros:
#if defined(_XOPEN_SOURCE_EXTENDED) || \
(defined(_XOPEN_SOURCE) && (_XOPEN_SOURCE - 0 >= 500))
drawbar.c defines _XOPEN_SOURCE with an empty value, which makes the
check evaluate as 0 >= 500 and fail.
This only breaks with musl because of the C library's <features.h>,
which curses.h pulls in through <stdio.h> before doing the check.
glibc and uClibc both redefine _XOPEN_SOURCE to 700 and define
_XOPEN_SOURCE_EXTENDED whenever _GNU_SOURCE is set, which drawbar.c
does, so the empty definition gets fixed up. musl never modifies the
application's feature test macros, so the empty value reaches
curses.h and the wide-character prototypes are left out.
Add a patch that defines _XOPEN_SOURCE as 700, matching the
_POSIX_C_SOURCE value of 200809L that is already used.
Fixes:
- https://autobuild.buildroot.org/results/914c1d8fa0fc8bbef64e4a64d58e5941d0077a3e
- https://autobuild.buildroot.org/results/b91ea0722202de0d1fdeeaeea167fd1b21b6d5d6
[1] 2110131cab
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Tested-By: Thomas Perale <thomas.perale@mind.be>
[Julien: add link to Buildroot commit introducing the issue]
Signed-off-by: Julien Olivain <ju.o@free.fr>
The last package depending on libplatform, libcec, removed its
dependency:
332bca975a
The maintainer of both libcec and libplatform has put libplatform into
maintenance mode: "Don't use this library in new projects."
97332651dd
so we can safely remove the package from buildroot.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
[Peter: move the COnfig.in.legacy entry under 2026.11]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Add support for pzstd, the multi-threaded implementation of zstd.
The pzstd build requires C++ and threads support. On some architectures
(e.g., Microblaze, SPARC), it also requires linking against libatomic
due to the use of atomic operations.
Ensure that -latomic is passed in LDFLAGS during both the build and
install phases. Consistently passing LDFLAGS is necessary because the
pzstd Makefile re-links the binary during the 'install' target due to
its dependency on the static library.
Fixes build failures reported with bootlin-microblazeel-uclibc and
bootlin-sparc-uclibc toolchains:
Pzstd.cpp:(.text+0x17e8): undefined reference to '__atomic_compare_exchange_1'
Pzstd.cpp:(.text._ZN5pzstd15BufferWorkQueue3popERNS_6BufferE): undefined reference to '__atomic_fetch_sub_4'
For host-zstd, always build pzstd. It just adds a few seconds of
build time.
Signed-off-by: Shubham Chakraborty <chakrabortyshubham66@gmail.com>
Reviewed-by: Joseph Kogut <joseph.kogut@gmail.com>
Tested-by: Joseph Kogut <joseph.kogut@gmail.com>
[Arnout: unconditionally build pzstd for the host]
Signed-off-by: Arnout Vandecappelle <arnout@rnout.be>
This was added in commit 2d4fd5d358 ("pkg-infra: also set PKGDIR for the
download step") for the DOWNLOAD macro, which was reworked in commit
155f752d8a ("core/download: do not hard-code the path to the package hash
file") to instead use <FOO>_HASH_FILE, so drop it.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This was added in commit 604d261fb0 ("infra: remove variable
FOO_DIR_PREFIX") to figure out where to find patches, but this was reworked
in commit 99d4180eda ("package/pkg-generic.mk: factor out patch
directories") to instead use <FOO>_PKGDIR, so drop it.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This was added in commit 7a41cbf938 ("package: fix patching procedure for
host packages") to figure out where to find patches, but this was reworked
in commit 99d4180eda ("package/pkg-generic.mk: factor out patch
directories") to instead use <FOO>_RAWNAME, so drop it.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since the bump to version 1.9.15 in commit
2a28d25d0c ("package/haveged: bump to version 1.9.15"), haveged
unconditionally includes <semaphore.h> and uses named semaphores
(sem_open() and friends) to synchronize command instances. As a
result, the build fails with toolchains without threads support:
haveged.c:37:10: fatal error: semaphore.h: No such file or directory
Add a dependency on BR2_TOOLCHAIN_HAS_THREADS. Since threads are now
always available, unconditionally pass --enable-threads.
Fixes:
https://autobuild.buildroot.org/results/814cb7b5537d8b7b2c23f3d90e37b2b41de7ee3f
Fixes: 2a28d25d0c ("package/haveged: bump to version 1.9.15")
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Assisted-by: Claude:claude-opus-5-5
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
